Dietary fiber and protein: Changes in degradation,

fermentation pattern and metabolic response of lean and

obese pigs

Yetong Xu

PhD Thesis

August 2020

Department of Animal Science

Graduate School of Technical Sciences

Aarhus University, Foulum

Denmark

SUPERVISORS

Senior scientist Helle Nygaard Lærke, Department of Animal Science, Faculty of Technical

Sciences, Aarhus University

Senior scientist Mette Skou Hedemann, Department of Animal Science, Faculty of Technical

Sciences, Aarhus University

Professor Knud Erik Bach Knudsen, Department of Animal Science, Faculty of Technical

Sciences, Aarhus University

ASSESSMENT COMMITTEE

Senior scientist, Ph.D., Ole Højberg (Chairman), Department of Animal Science, Faculty of

Technical Sciences, Aarhus University, Denmark

Senior research scientist, Ph.D., Sergio Polakof, Human Nutrition, French National Institute

for Agricultural Research (INRA), France

Senior Lecturer, Associate Professor, Ph.D., Anne Nilsson, Department of Food Technology,

Engineering and Nutrition, Lund University, Sweden

I

Preface

This thesis is based on research conducted during my enrolment as a PhD student at the Department

of Animal Science, Aarhus University from September 2017 until August 2020. The thesis includes

data from two animal experiments from two separate projects:

1. MERITS (Metabolic changes by carbohydrate and protein quality in the development and

mitigation of metabolic syndrome) funded by Innovation Fund Denmark (Project no. 4105-

00002B) in collaboration with Aarhus University Hospital, University of California, Davis,

Norwegian University of Life Science, Arla Food Ingredients/Arla Food, Lantmännen R&D,

DuPont Nutrition Biosciences and Ellegaard Gӧttingen Minipigs.

2. ELIN (The effects of enterolignans in chronic disease) funded by Innovation Fund Denmark

(Project no. 0603-00580B) in collaboration with Danish Cancer Society Research Center,

Swedish University of Agricultural Sciences, Lantmännen R&D and DuPont Nutrition

Biosciences.

The PhD project included a stay abroad from April 2018 to July 2018 in the lab of Professor Maria

L. Marco, PhD, Department of Food Science & Technology, University of California, Davis, U.S.A.,

founded by the Graduate School of Science and Technology at Aarhus University.

The animal experiment from MERITS project was completed in March 2017. I am responsible for

sample analysis and data handling of short chain fatty acid (SCFA), microbiota, starch and non-starch

polysaccharides (NSP) of gut content, which are included in Paper I and Paper II. The animal trials

and sample analysis of ELIN project have been completed by others before I enrolled. I was

responsible for data analysis of nutrient digestibility, NSP composition in gut content, SCFA in gut

content and blood and correlations between intestinal microbiota and metabolites, which are included

in Paper III.

This PhD project aims to study the detailed mechanisms of how dietary fiber and protein interact

with the microbiota profile, production and absorption of SCFA and metabolic response, thereby

enhance metabolic health using pigs as models for humans.

II

Acknowledgement

There are so many people encouraging and motivating me during my PhD study and I would never

finish my study without your help. I’m very grateful to all the people who have supported me during

my stay in Denmark.

First of all, I would like to thank my main supervisor Helle Nygaard Lærke, for her patient

supervision and valuable suggestions along the way. Thank you for providing me an opportunity to

study in Denmark and work with a group of wonderful people in Foulum. Your deep understanding

in nutrition and physiology always impress me, and I feel so lucky to have you as my role model in

research. More importantly, thanks for your constructive and precise guidance for my study to help

me improve my understanding in my PhD project and writing skills, moreover, support me and lead

me to this scientific road.

Then I would like to thank my co-supervisor, Knud Eric Bach Knudsen, for giving me the possibility

to join in MERITS project. Thanks for being my main-supervisor for a few months and providing

great supervision in my PhD study and valuable comments in my manuscripts. In addition, I would

like to thank my co-supervisor Mette Skou Hedemann for your kind support and encouragement, and

always constructive and detailed input in my PhD study.

Thanks everyone involved in the MERITS project for their contributions and helpful advice. I would

like to extend my gratitude to Professor Maria L. Marco to give me the opportunity to stay abroad

for three months in her lab at University of California, Davis. Special thanks to Zach and Jason for

your help and guidance in analyzing microbiota and data handling and thanks Zhengyao Xue, Jinlin

Guo, Annabelle Yu, Eric Stevens and Dustin Heeney for kind care and making my stay so

memorable.

Many thanks to all my colleagues at the Department of Animal Science especially Winnie Østergaard,

Lisbeth Mӓrcher, Stina Greis Handberg, Kasper Vrangstrup Poulsen and Thomas Rebsdorf for

excellent technical assistance in my lab work and thanks Leslie Foldager for supporting me on data

handling. I also want to thank Mihai Curtasu for providing so much help for me about the animal

experiments and sample analysis.

Last but not least, my deepest thanks go to my Chinese friends and family for always being supportive

and for believing in me. I would never forget the help and kind care from Xiangyu Guo, Yuan Yue,

Pan Zhou, Long Chen, Kun Zhou, Zhi Liang and Qianying Yi, thank you for making my work and

stay in Foulum so enjoyable. Thanks my best friend Kangni Liu who works in Japan but always

support and comfort me when I’m upset, you always bring me a lot of fun and I’m so lucky to have

you as my bestie for 11 years. Heartfelt thanks to my parents and brother who support me to do

whatever I want, and I would never finish my study and work in Denmark without your love.

Yetong Xu,

August 2020

III

Summary

The rising prevalence of obesity, accompanied by increased risk of metabolic syndrome (MetS) is

raising a worldwide alarm. One of the major factors is the western lifestyle including an unhealthy

diet with high fat and refined carbohydrate contents. Dietary fiber (DF) intake has been proved to

have health beneficial properties such as modulating digestion processes, improving microbiota

profile, and short chain fatty acid (SCFA) production, which are linked to a lower risk of MetS

including cardiovascular disease and type 2 diabetes. Some dietary proteins have insulinotropic

effects, especially the abundant branched chain amino acids of whey protein have been shown to be

efficient in promoting postprandial insulin and incretin, therefore improving the glycemic and insulin

responses.

In this PhD thesis, the effects of dietary strategies based on low or high DF meals rich in arabinoxylan

(AX) and low or high protein contents by addition of whey protein hydrolysate on metabolic

responses and fermentation profile were studied using an obese Gӧttingen Minipig model. Before the

DF and protein intervention, forty-three minipigs were fed a high fat high fructose diet for 20 weeks

to induce obesity. After that, the minipigs were assigned to one of four diets for 8-week ad libitum

feeding. Metabolic responses of blood and urine samples taken in the fasting and non-fasting states

were determined, gene expression of liver, muscle and adipose tissues, carbohydrate, SCFA and

microbiota in gut content were analyzed. In another study, the effects of DF ingredients on digestion

and fermentation processes as well as SCFA absorption were investigated by feeding a rye bran (RB)

diet high in AX to conventional pigs (n = 20) to make a comparison with a refined wheat fiber

(Control) diet high in cellulose (n = 10). Half of the pigs fed the RB diets was treated with antibiotics

to study if it would modify the effects of the RB diets on macronutrient digestibility, DF degradation,

SCFA production and absorption.

The first experiment showed that a high DF content reduced weight gain and improved the C-peptide

secretion of obese minipigs in the non-fasting state without alleviating mild tissue inflammation.

However, a high dietary protein content increased weight gain, and unfavorably altered metabolic

biomarkers and gene expression related to carbohydrate metabolism. AX in the high-DF diets was

degraded until the mid colon, stimulated the abundance of butyrogenic genera, slightly increased

intestinal butyrate production and circulating butyrate levels. High dietary protein also contributed

to intestinal SCFA production, decreased circulating succinate levels, but did not show prebiotic

effects, and proteolytic fermentation was attenuated by high DF. Overall, the modulated fermentation

profile could be linked with the potential mechanisms of the effects of DF and protein on metabolic

responses separately, and a combination of high DF and protein did not have a synergistic effect on

IV

metabolic health in this study. The second experiment showed that the RB diets slowed down and

decreased the protein degradation in the gut. AX in the RB diets was mainly degraded in the cecum

and proximal colon, showed higher butyrate production - but not absorption - compared with the

Control diet. Cellulose in the Control diet was slowly degraded along the large intestine and increased

the production and absorption of total SCFA, acetate and propionate, which influenced the plasma

lipid profile. Although the use of antibiotics did not show noticeable change in the degradation

process, it led to a reduction in butyrate production. Collectively, the study demonstrated that the

intestinal degradation and fermentation patterns were closely associated with the DF ingredients,

which resulted in different profiles of SCFA production and absorption.

In conclusion, the results in this PhD study show how DF and protein influence the metabolic

responses by associating them with the intestinal degradation and fermentation patterns, and provide

a greater understanding regarding the mechanisms of dietary strategies to modulate MetS.

V

Dansk resumé

Forekomst af fedme og deraf følgende forekomst af metabolisk syndrom (MetS) stiger alarmerende

verden over. En af de væsentlige faktorer er den vestlige livsstil, som omfatter en usund kost med

højt indhold af fedt og raffinerede kulhydrater. Indtag af kostfibre (DF) har vist sig at have

sundhedsgavnlige effekter såsom ændring af fordøjelsesprocesser, forbedring af mikrofloraens

sammensætning og produktion af kortkædede fedtsyrer (SCFA), hvilket er forbundet med en lavere

risiko for MetS, inklusiv kardiovaskulære sygdomme og type 2 diabetes. Nogle fødevareproteiner

har insulinotrofiske effekter, især den høje forekomst af forgrenede aminosyrer i valleprotein har vist

sig at stimulere det postprandiale insulin- og inkretinrespons og dermed forbedreglukose- og

insulinresponset.

I denne PhD afhandling blev effekter af fodringsstrategier baseret på lav eller høj DF diæter med højt

indhold af arabinoxylan (AX) og lavt eller højt indhold af protein baseret på tilsætning af

valleproteinhydrolysat på det metaboliske respons og fermenteringsprofilen studeret i en fed

Gӧttingen Minigrisemodel. Før interventionen med DF og protein var 43 minigrise blevet fodret med

en diæt med højt indhold af fedt og fruktose i 20 uger for at inducere fedme. Herefter blev grisene

allokeret til en af fire diæter i 8 uger med ad libitum fodring. Metaboliske responser blev målt i blod

taget fastende og ikke-fastende, genekspression blev målt i lever-, muskel- og fedtvæv, og

kulhydrater, fermenteringsprodukter og mikrobiotasammensætning i tarmindhold blev analyseret. I

et andet forsøg blev effekt af fiberingredienser på fordøjelses- og fermenteringsprocesser samt SCFA

absorption undersøgt i konventionelle grise (n = 20), der blev fodret med en rugklidsbaseret diæt

(RB) med højt indhold af AX og sammenlignet med grise fodret med kontroldiæt indeholdende

raffineret hvedefibre med højt celluloseindhold (n = 10). Halvdelen af grisene i RB gruppen blev

behandlet med antibiotika for at undersøge, om dette ville påvirke effekten af RB diæten på

makronæringsstoffernes fordøjelighed, nedbrydning af DF, samt SCFA produktion og absorption.

Det første forsøg viste, at et højt fiberindhold reducerede vægtøgning og forbedrede sekretionen af

C-peptid i de fede minigrise i ikke-fastende tilstand uden at dæmpe den milde vævsinflammation.

Derimod øgede et højt proteinindhold vægtstigningen og ændrede metaboliske biomarkører samt

ekspression af gener relateret til kulhydratomsætning i ugunstig retning. AX i høj-DF diæterne blev

nedbrudt indtil midten af kolon, stimulerede forekomsten af butyrogene bakterieslægter, samt øgede

den intestinale produktion og det cirkulerende niveauer af butyrat (smørsyre) i moderat omfang. Et

højt proteinindhold bidrog også til SCFA produktion i tarmen, reducerede det cirkulerende niveau af

succinat (ravsyre), men havde ingen prebiotiske effekter, og den proteolytiske fermentering blev

dæmpet af et højt DF indhold. Overordnet kunne en ændret fermentationsprofil forbindes med de

VI

potentielle mekanismer for effekt af henholdsvis DF og protein på metabolisk sundhed, og i dette

studie sås ingen synergistisk effekt af at kombinere DF og protein. Det andet forsøg viste, at RB

diæterne forsinkede og reducerede proteinnedbrydningen i tarmen. AX i RB grupperne blev

hovedsagelig nedbrudt i blindtarmen og den forreste del af kolon, og viste samtidig en højere

butyratproduktion - men ikke absorption - sammenlignet med kontroldiæten. Cellulose i

kontroldiæten blev langsomt nedbrudt gennem tyktarmen og øgede produktionen og absorptionen af

total SCFA, acetat og propionat, hvilket påvirkede plasmalipid-profilen. Selvom brugen af

antibiotika ikke viste nævneværdige ændringer i nedbrydningsprocessen, førte det til en reduktion i

butyratproduktionen. Samlet viste studiet, at den intestinale nedbrydning og fermenteringsmønsteret

var tæt forbundet med DF kilderne, hvilket resulterede i forskellige profiler for SCFA produktion og

absorption.

Sammenfattende viser resultaterne i dette PhD studium, hvordan DF og protein påvirker det

metaboliske respons gennem deres association til nedbrydning og fermenteringsmønster i tarmen, og

studiet bidrager til en større forståelse af de mekanismer, hvorved forskellige koststrategier påvirker

MetS.

VII

My own contributions to the thesis

This PhD thesis comprises three manuscripts based on two animal experiments performed at Aarhus

University. The animal experimental work was completed before I enrolled. Therefore, I was not

involved in planning the studies, but I contributed proportionally in conduction the research by

conducting the analytical work, moreover, was responsible for data handling, interpretation and

presentation; drafting, revising and submitting the manuscripts to peer-reviewed journals. In Paper

I, I performed the data handling and statistical analysis of metabolic biomarkers in blood and urine

samples and gene expression of liver, muscle and adipose tissues and wrote the manuscript; In Paper

II, I conducted the sample analysis of carbohydrates (starch and NSP), SCFA and microbiota of gut

content, performed data handling and wrote, revised and submitted the manuscript; For Paper III, I

accomplished data analysis of nutrient digestibility, carbohydrate and SCFA composition in gut

content and SCFA in the blood samples, correlation analysis of intestinal microbiota and metabolites,

drafted, revised and submitted the manuscript.

For the residual parts of the thesis, I have written all of it and my main supervisor Helle Nygaard

Lærke has given feedback especially in linguistic accuracy and appropriateness, data analysis,

scientific perspectives, suggestions about writing structure and coherence of my thesis and the

manuscripts. My co-supervisors Mette Skou Hedemann and Knud Eric Bach Knudsen have given

their feedback mainly to linguistic expressions, scientific perspectives and rhetorical organization of

the manuscripts.

There are no conflicts of interest to declare in this thesis.

VIII

Other scientific contribution

Scientific paper

‘Obesity-related Metabolome and Gut Microbiota Profiles of Juvenile Göttingen Minipigs – Long-

term Intake of Fructose and Resistant Starch’

Mihai V. Curtasu, Valeria Tafintseva, Zachary Bendiks, Maria L. Marco, Achim Kohler, Yetong Xu, Helle

Nygaard Lærke, Knud Erik Bach Knudsen, Mette Skou Hedemann; Manuscript in preparation.

Presentation at conferences

Oral presentation: ‘Does ad libitum access to carbohydrates influence the development of obesity

and MetS in Göttingen Minipigs - and can it be reverted by fiber and protein?’

The 7th Porcine Biomedical Models meeting, 22 November 2018, University of Copenhagen.

Planned presentation: ‘Microbiome alterations in response to dietary fiber and protein in obese

Göttingen Minipigs?’

The 14th Minipig Research Forum, 13-15 May 2020, Lisbon, Portugal. (The meeting was cancelled due to

Covid-19)

Scientific assistance

Assisted the PhD course ‘Carbohydrates emphasis on nutrition and intestinal health of non-ruminant

animals’

Conducted SCFA analysis of faeces in the human study of MERITS project

IX

List of abbreviations

AX Arabinoxylan

AXOS Arabinoxylan oligosaccharides

A:X Arabinose:xylose

BCAA Branched chain amino acid

BCFA Branched chain fatty acid

BW Body weight

BMI Body mass index

BSA Body surface area

CCK Cholecystokinin

CVD Cardiovascular disease

DF Dietary fiber

DM Dry matter

DPP-IV Dipeptidyl peptidase-IV

GIP Glucose-dependent insulinotropic polypeptide

GLP-1 Glucagon-like peptide

GPR G protein coupled receptor

HDL High density lipoprotein

HMW High molecular weight

HOMA Homeostasis model assessment

IFN-γ Interferon gamma

IL Interleukin

IR Insulin resistance

LDL Low density lipoprotein

LMW Low molecular weight

MetS Metabolic syndrome

NEFA Non-esterified fatty acid

NF-κB Nuclear factor kappa B

NSP Non-starch polysaccharides

OTU Operational taxonomy units

PPAR Peroxisome proliferator-activated receptors

PYY Peptide YY

QIIME Quantitative insight into microbial ecology

RB Rye bran

SCFA Short chain fatty acid

T2D Type 2 diabetes

TNF-α Tumor necrosis factor

VLDL Very low density lipoprotein

Table of Contents

Preface ................................................................................................................................................ I

Acknowledgement ............................................................................................................................. II

Summary .......................................................................................................................................... III

Dansk resumé .................................................................................................................................... V

My own contributions to the thesis .............................................................................................. VII

Other scientific contribution ........................................................................................................ VIII

List of abbreviations ....................................................................................................................... IX

1. Introduction................................................................................................................................ 1

2. Background ................................................................................................................................ 2

2.1 Metabolic syndrome ................................................................................................................ 2

2.1.1 Obesity ............................................................................................................................... 2

2.1.2 Dyslipidemia ...................................................................................................................... 3

2.1.3 Insulin resistance .............................................................................................................. 4

2.1.4 Low-grade inflammation .................................................................................................. 5

2.1.5 Gut microbiota change with MetS .................................................................................. 5

2.2 Dietary fiber ............................................................................................................................. 6

2.2.1 Definition, sources and classification .............................................................................. 6

2.2.2 Grain bran ......................................................................................................................... 7

2.2.3 Role of dietary fiber in appetite ....................................................................................... 8

2.2.4 Role of dietary fiber in glycemic control ........................................................................ 9

2.2.5 Role of dietary fiber in lipid metabolism ........................................................................ 9

2.2.6 Role of dietary fiber in attenuating inflammation ....................................................... 10

2.2.7 Dietary fiber degradation and fermentation by microbiota ....................................... 11

2.3 Dietary protein ....................................................................................................................... 14

2.3.1 Sources, composition and properties ............................................................................ 14

2.3.2 Whey protein ................................................................................................................... 14

2.3.3 Role of dietary protein in appetite ................................................................................ 15

2.3.4 Role of dietary protein in glycemic control .................................................................. 15

2.3.5 Role of dietary protein in lipid metabolism .................................................................. 16

2.3.6 Role of dietary protein in low-grade inflammation ..................................................... 17

2.3.7 Dietary protein degradation and fermentation by microbiota ................................... 18

3. Aim and hypotheses ................................................................................................................. 20

4. Methods and methodological considerations ........................................................................ 21

4.1 Experimental diets and design .............................................................................................. 21

4.1.1 The MERITS study......................................................................................................... 21

4.1.2 The ELIN study............................................................................................................... 22

4.2 Pigs as experimental models for humans............................................................................. 23

4.3 Obesity measurement ............................................................................................................ 24

4.4 Insulin sensitivity assessment................................................................................................ 25

4.5 Gut microbiota analysis ........................................................................................................ 25

4.6 SCFA production and circulation ........................................................................................ 26

4.7 Type and administration route of antibiotics ...................................................................... 27

4.8 Statistical methods ................................................................................................................. 27

5. Brief summary of results ............................................................................................................. 28

6. Paper ............................................................................................................................................. 31

6.1 Paper I ..................................................................................................................................... 31

6.2 Paper II ................................................................................................................................... 71

6.3 Paper III ............................................................................................................................... 112

7. Discussion ................................................................................................................................... 147

7.1 Obesity development of minipig models ............................................................................ 147

7.2 Clinical parameters associated with MetS in minipig models ......................................... 148

7.3 Dietary fiber degradation and nutrient digestion ............................................................. 150

7.4 SCFA production and circulation ...................................................................................... 152

7.5 Microbiome profile .............................................................................................................. 154

8. Conclusion and perspectives ..................................................................................................... 157

References ....................................................................................................................................... 160

1

1. Introduction

The prevalence of obesity is rising worldwide and it has become a critical risk factor for the

development of metabolic syndrome (MetS) including type 2 diabetes, cardiovascular disease,

dyslipidemia and some cancers. Easy access to high in fat, calories, and refined sugar food is a critical

factor responsible for the obesity development (1). Excessive nutrient intake stimulates the synthesis

of triglycerides and lead to an abnormal fat accumulation, which is closely linked with insulin

resistance, hypertention and inflammation (2). Therefore, controlling the development of MetS

becomes greatly significant for public health and it is of interest to investigate if dietary interventions

can prevent metabolic alterations and alleviate obesity–related MetS.

Intake of dietary fiber (DF) obtained from cereals, fruits and vegetables has been associated with

lower risk of MetS (3). DF can not be digested by endogenous enzymes in the small intestine and can

be delivered to large intestine and to some extent fermented by the intestinal microbiota into short-

chain fatty acid (SCFA). The profile of microbiota and their SCFA production in the large intestine

vary depending on different type and composition of DF. For instance, arabinoxylan is an efficient

substrate for intestinal butyrate production as well as enrichment of butyrogenic species (4). SCFA

produced in the gut can signal through receptors on enteroendocrine cells, influence the secretion of

gut hormones to increase satiety and insulin sensitivity (5). Additionally, absorbed SCFA can

influence the release of adipokines, which may inhibit lipolysis and ectopic fat storage (6, 7). Another

food component of interest associated with metabolic health is dietary protein of which some have

shown an insulinotropic effect, e.g. whey protein, as a major constituent of milk protein, can raise

the postprandial levels of branched-chain amino acid and acutely improve lipid profile in subjects

with MetS (8, 9). Observational studies indicate that the insulinotropic effect of whey protein also

affects postprandial glucose response in both healthy and obese subjects (10-12).

Both DF and protein have been shown have beneficial effects on metabolic health, the underlying

biochemical mechanisms of action have not been clearly identified. Pigs have similar metabolic

features and can response similarly to dietary interventions compared with humans (13). In this

thesis, we use obese minipigs and conventional pigs as models to investigate how diets changing in

DF and protein composition influence intestinal degradation pattern, microbiota composition, SCFA

production and absorption, thereby associate them with metabolic response.

2

2. Background

2.1 Metabolic syndrome

Metabolic syndrome (MetS) is a combination of risk factors for the development of obesity, type 2

diabetes (T2D), cardiovascular disease (CVD) and many cancers. MetS is defined in various ways

over time by different expert groups and organizations. In 2009, International Diabetes Federation

(IDF) and the American Heart Association/National Heart, Lung, and Blood Institute (AHA/NHLBI)

representatives proposed the common definition for the clinical diagnosis of the MetS (Table 1) (14).

Patients who fulfil the criteria of MetS have a two-fold increased risk of CVD, three-fold increased

risk of diabetes and 1.5-fold increased risk of all-cause mortality (15).

Table 1. Criteria for clinical diagnosis of the metabolic syndrome (14).

Measure Categorical cut points

Elevated waist circumference* Population- and country-specific definitions

Elevated triglycerides 150 mg/dL (1.7 mmol/L)

or drug treatment for elevated triglycerides

Reduced HDL 40 mg/dL (1.0 mmol/L) in males;

50 mg/dL (1.3 mmol/L) in females;

or drug treatment for low HDL

Elevated blood pressure Systolic 130 and/or diastolic 85 mm Hg

or drug treatment for hypertension

Elevated fasting glucose 100 mg/ dL (5.5 mmol/L)

or drug treatment for elevated levels

HDL indicates high-density lipoprotein cholesterol.

* It is recommended that the IDF cut points (waist > 94 cm for men or > 80 cm for women) be used for non-

Europeans and either the IDF or AHA/NHLBI cut points (waist > 102 cm for men or > 88 cm for women)

used for people of European origin until more data are available.

2.1.1 Obesity

Obesity is defined as abnormal or excessive fat accumulation that impair health, and it occurs when

fat amount exceeds the holding ability of subcutaneous depots. Since 1980, the prevalence of obesity

has doubled in more than 70 countries and now it is becoming the major world-wide nutritional

challenge, affecting both children and adults (15). The important cause for the rapid increase in

obesity rates is dietary energy intake exceeds energy expenditure with declined physical activity and

increased unhealthy, energy-dense food intake in the modern life (1). Up to 2015, approximately 604

million adults and 108 million children were considered as obese and the incidence of obesity in most

western countries was over 20% (16). Obesity can be evaluated through several different

3

anthropometric indexes such as body mass index (BMI, the weight in kilograms divided by the square

of the height in meters) and further be evaluated in terms of fat distribution through determining waist

circumference and waist-to-hip ratio. According to the World Health Organization (WHO), BMI is

classified as: normal weight (BMI: 18.5–24.9 kg/m2), overweight (BMI: 25.0–29.9 kg/m2), obesity

(BMI: 30.0–34.9 kg/m2), severe obesity (BMI: 35.0–39.9 kg/m2) and for morbid obesity (BMI ≥40

kg/m2) (17). From 1988 to 2010, average BMI in USA increased by 0.37% per year in both men and

women, while waist circumference increased by 0.37% and 0.27% per year in men and women,

respectively (18). The appearance of the MetS phenotype is provoked by accumulation of body fat

in other body compartments rather than subcutaneous tissue, particularly with an increase in

abdominal obesity mirrored by a large waist circumference (19). Obesity is the crucial factor to

induce MetS by showing increased insulin resistance (IR), postprandial lipidemia and low-grade

inflammation, which cause elevated risk of cardiovascular disease, type 2 diabetes (T2D),

dyslipidemia and hypertention (20). However, obesity is not always connected with MetS and some

metabolically healthy obese individuals who have high insulin sensitivity do not have hyperlipidemia

or other characteristics of MetS (18).

2.1.2 Dyslipidemia

Adipocytes and adipose tissue store the greatest amounts of body lipids and are the important for

endocrine and immune response, therefore, adipocyte and adipose tissue dysfunction caused by

obesity may increase the risk of metabolic disease such as dyslipidemia (21). Dyslipidemia is

characterized by numerous disturbances including increased triglyceride, low density lipoprotein

(LDL) cholesterol and apolipoprotein B levels and depressed high density lipoprotein (HDL)

cholesterol (22). Dyslipidemia is a risk factor of atherosclerotic cardiovascular disease and has a

strong correlation with obesity caused by excessive adipose tissue deposition (23). Approximately

60%-70% of obese people are dyslipidemic while 50%-60% of overweight people are dyslipidemic

(21). Surplus dietary intake of fat and glucose provides excessive sources for fat droplet accumulation

in the liver as well as the overproduction of very low density lipoprotein (VLDL), which can transport

triglycerides and contribute to the elevation in plasma triglyceride levels (24). The VLDL undergoes

enzymatic exchanges with LDL and HDL via cholesteryl ester transfer protein, and these triglyceride-

enriched LDL and HDL can be degraded by lipases into small and dense LDL and HDL, respectively

(21). The filtration of small and dense LDL particles into the arterial wall leads to the ultimate

development of atherosclerosis, therefore, the therapeutic target for dyslipidemia and cardiovascular

risk has focused on lowering LDL over the last two decades (25). HDL has been reported as an

athero-protective molecule and a marker of overall cardiovascular risk, every 0.026 mmol/L increase

in HDL is associated with an approximate 2-4% decrease in the risk of CVD (26). In obese subjects

4

with the fatty liver disease, the small and dense HDL induced by surplus VLDL formation can be

easily cleared by the kidney, causing lower HDL cholesterol levels (21). At the same time, high levels

of fatty acids can be released by the hypertrophic adipocytes in obese individuals due to the increased

basal lipolysis, which not only deteriorate the development of non-alcoholic fatty liver disease

(NAFLD) but also result in insulin resistance in peripheral organs (27, 28).

2.1.3 Insulin resistance

Insulin is an anabolic hormone produced by pancreatic β-cells and plays an important role in

carbohydrate metabolism as well as protein and lipid anabolic regulation. Metabolic dysfunction of

the interplay between glucose and insulin is evident in MetS. In the IR state, the inability of insulin

to promote glucose uptake in the muscle and to suppress gluconeogenesis in the liver leads to

hyperglycemia and subsequent compensatory hyperinsulinemia (29). IR has a pivotal role in the

pathogenesis of T2D, which accounted for more than 90% of 415 million patients with diabetes in

2015 (30). Patients with T2D have a 15% increased risk of all-cause mortality than people who do

not have diabetes (31). IR can manifest itself in many peripheral tissues including the liver, skeletal

muscle and adipose tissue. Among them, the liver plays an important role in determining fasting

hyperglycemia, therefore hepatic IR is a prediabetic state. It forces β cells of the pancreas to

continuously secrete insulin and ultimately results in the development of β cell failure of T2D.

Moreover, hepatic IR occurs when the fatty liver increases some lipid such as triglyceride secretion,

which can impair hepatic insulin signaling and result in reduced insulin activation of hepatic glycogen

synthesis and de novo lipogenesis (32). Similarly, the accumulation of intramyocellular lipid content

in skeletal muscle impairs insulin-stimulated glucose transport and glycogen synthesis, ultimately

leading to IR (29). Moreover, in adipocytes, insulin plays a critical role in suppressing lipolysis and

promoting lipid and glucose uptake (32). Therefore, IR in liver, muscle and adipose tissue can result

in increased lipolysis and more fatty acids delivery to liver for triglyceride synthesis, whereas hepatic

de novo lipogenesis is reduced due to the impaired hepatic insulin signaling (Figure 1).

(A)

5

Figure 1. (A) Insulin sensitive status: insulin regulates lipid metabolism via liver, muscle and adipose insulin

signaling. (B) Insulin resistance status: impaired regulation in lipid metabolism via liver, muscle and adipose

insulin signaling. Dotted lines represent decreased action or deceased flux. DNL, de novo lipogenesis; FA,

fatty acid; TAG, triglycerides; VLDL, very low density lipoprotein. Adapted from Samuel et al. (2016) (32).

2.1.4 Low-grade inflammation

Low-grade inflammation in adipose tissue is a common outcome of the development of obesity.

Increased adipose-derived fatty acids in obese individuals activate inflammatory signaling and

promote activation of macrophages in adipose tissue, which afterward spread to other tissues such as

liver and skeletal muscle (33). In a low grade inflammation state, increased infiltration and activation

of pro-inflammatory macrophages and some other immune cells are observed, resulting in more pro-

inflammatory cytokines and chemokines production and secretion (34). Moreover, low-grade

inflammation in adipose tissue is considered as a critical risk factor for the development of IR and

T2D in obese people. Factors that associated with promotion of IR progression include altered

secretion of adipocytokines and inflammatory cytokines by enlarged adipocytes and microphages

such as adiponectin, leptin, TNF-, interleukins and interferon gamma (IFN-γ), which can disrupt

insulin-receptor signaling and impact fatty acid transcription factors (like PPARγ) (35, 36). These

pro-inflammatory cytokines can activate Nuclear factor kappa B (NF-κB) pathway, which in turn

may exacerbate IR (36). The continuously increased systemic levels of inflammatory cytokines

including TNF, IL-4 and IL-8 is also a risk factor for the progression of inflammatory bowel disease

and colorectal cancer (37, 38).

2.1.5 Gut microbiota change with MetS

The western lifestyle with low DF intake, increased fat and sugar intake and lack of physical activity

may contribute to the development of intestinal disorders (39). Colorectal cancer is predominantly a

disease of western countries, it has become the third most common cancer type among men and the

second most common type among women (40). Moreover, the occurrence of this intestinal disorder

is accompanied by altered gut microbiota composition (41). The human gut is populated by trillions

of microbiota that are important for the host development and physiology such as maturation of

immune system and modulation of host metabolic pathways (42). Species belonging to the phyla

(B)

6

Firmicutes (incl. Clostidium, Eubacterium, Faecalibacterium, Lactobacillus, Roseburia ect.),

Bacteroidetes (i.e., Bacteriodes, Prevotella) and Actinobacteria (i.e., Bifidobacteria) are the most

abundant, which account for approximately 90% of total bacteria in the gut. Notably, the microbiota

has a symbiotic relationship with the host, which plays a crucial role in energy metabolism, metabolic

signaling across organs, gut barrier integrity and immunity (43). A previous study has shown that

compared with lean individuals, obese subjects have different Firmicutes and Bacteroidetes

proportions as well as microbial diversity, which also change with weight loss (44). Moreover,

Turnbaugh et al (2009) also found the obese cohort had higher levels of fecal SCFA which was

associated with a higher Firmicutes to Bacteroidetes ratio than lean cohort (44). However, the higher

Firmicutes to Bacteroidetes ratio is not an eventual biomarker for obesity as many contradictory

results were reported previously. For instance, altered gut microbiota composition occurs in obesity

and T2D showing a reduced abundance of butyrate-producing bacteria including Roseburia and

Faecalibacterium prausnitzii which belong to the phylum Firmicutes (45). An explanation is that the

relative abundance of butyrate-producing bacteria are low in the obese individuals and progressively

replaced by other bacteria belonging to the same phylum (46). Therefore, the changes of microbiota

at family, genus or species level might be more relevant compared with the Firmicutes to

Bacteroidetes ratio.

2.2 Dietary fiber

2.2.1 Definition, sources and classification

Dietary fiber (DF) is defined as a group of carbohydrate oligo- and polymers with a degree of

polymerization of 3-10 or > 10, which are derived from edible plants and cannot be digested by

endogenous enzymes in the small intestine (47). DF has been considered as an important ingredient

of a healthy diet, which can be obtained from cereals, fruits and vegetables ect. The type and

composition of DF vary with source, including cellulose, non-cellulosic polysaccharides (NCP),

resistant starch, lignin and oligosaccharides (48). Generally, fruits contain a large part of digestible

sugars and DF such as pectin, whereas the DF proportion and composition of vegetables can vary a

lot between different plant origins including leaves, stems, roots and tubers (48). In Northern Europe,

cereal grains are a major source of carbohydrate intake. The whole-grain cereals contain abundant

DF components, mainly composed of non-starch polysaccharide (NSP), resistant starch and lignin

(non-carbohydrate plant compounds) (49).

DF can be categorized according to source, solubility, fermentable ability and physiological effects

as shown in Table 2 (50). One of the simple classifications is dividing DF into soluble and insoluble

fiber based on the ability to be fully dispersed when mixed with water (48). Soluble fiber includes β-

7

glucan, guar gum, pectin and arabinoxylan (AX), wherein only a small fraction of AX are water

soluble and highly viscous. The soluble fiber can be further subdivided based on their molecular size

into two subcategories. The polymeric soluble fibers with high molecular weight (HMW) have high

viscosity, whereas low molecular weight (LMW) property makes soluble oligosaccharides not alter

the viscosity or texture of foods but are highly fermentable and could be utilized by the terminal

ileum (51). Insoluble fiber consist of a major part of DF fractions in cereal grains like cellulose, AX

and lignin. As the major component of insoluble DF of plant cell wall, cellulose has a linear polymer

of glucose units linked by β-(1, 4) linkages, which has laxative effect, contributing to regulate

digestive function (50). The insoluble DF can be either rapidly fermented (e.g. from refined flour),

slowly fermented (e.g. resistant starch, wheat bran) or essentially unfermented (e.g. vascular tissue)

(48). Therefore, solubility and fermentability are not always strongly correlated. Notably, DF can

benefit on gut health by acting as prebiotics, which are defined as ‘a non-digestible compound that,

through its metabolization by microorganisms in the gut, modulates composition and/or activity of

the gut microbiota, thus conferring a beneficial physiologic effect on the host’ (52).

Table 2. Properties of some dietary fiber sources (Modified after Bozzetto et al. (2018) (50)).

Fibers Viscosity Solubility Fermentation Physiological

Effects

Cellulose,

heteroxylans,

lignin

Low Low Low Laxative effect

Inulin, dextrin,

oligosaccharides Low High High Prebiotic effect

Pectin, β-glucan,

guar gum High High High

Prebiotic effect,

reduce nutrients

absorption

Psyllium,

methylcellulose High High Low

Laxative effect,

reduce nutrients

absorption

2.2.2 Grain bran

In Northern European countries, cereal grains like rye and wheat are important sources of

carbohydrate and DF. Whole gain of cereals consist of pericarp/testa, aleurone, starchy endosperm

and germ, while the bran of whole grain including pericarp/testa, aleurone layer and varying parts of

endosperm depending on milling efficiency and have a high DF content (49). For instance, the DF

content is 13-17 g/100 g in whole rye grain and 35-49 g/100 g in rye bran (53), while it is 9-17 g/100

g in whole wheat gain and 36-53 g /100 g in wheat bran (49). Moreover, the bran has a complex

structure with variable properties and composition of NSP, but typically approximately 55% AX,

followed by cellulose and mixed linkages β-glucan (54, 55). AX is formed from a linear backbone

8

of xylose residues mainly substituted with arabinose residues to varying degree at the O-2 position,

O-3 position, or both (56). The functional properties of AX are associated with the extent of

substitution and distribution of substituents along the xylan backbone, which is partly reflected by

different ratio of arabinose to xylose (A:X). Recent studies found that AX influenced the digestion

and absorption processes and increased SCFA especially butyrate production in the large intestine (4,

57). Compared with barley and oats, rye and wheat bran has a higher content of soluble and insoluble

DF in the form of AX (58). Most of AX in wheat bran is insoluble (80%) while the proportion of

soluble AX in rye bran is higher and cellulose content is lower than wheat bran (59, 60). Bran can be

treated with enzymes for the transformation into arabinoxylan oligosaccharides (AXOS) particularly

of insoluble AX from the aleurone layer with a low with a lower substitution pattern (61, 62). Besides

AX, cellulose is also found rich in wheat and rye bran and has an ability to bind water, which helps

promote regular bowel movements, reduce gut transit time, and can also be partly fermented by the

colonic microbiota and produce SCFA (63).

2.2.3 Role of dietary fiber in appetite

At present, a high DF intake is accepted as an essential component of a healthy diet. Most countries

recommend a daily intake of dietary fiber of 25–35 g for adults (25–32 g/d for adult women and 30–

35 g/d for adult men) (64). DF has been reported to improve satiation, satiety and reduce food intake,

which is controlled by appetite signaling through a complex system of hormones. A recent systematic

review found different types of DF influenced the outcomes, where DF showing high viscosity could

reduce appetite more often than the less viscous DF (59% vs. 14%) (65). The increased viscosity

induced by high DF intake contributes to an increase in gastric volume and a decrease in gastric

emptying (66). The prolonged presence of nutrients in the gastrointestinal tract increases the

interaction between nutrients and the mucosa of small intestine to stimulate the release of peptides

involved in appetite regulation (67). This is achieved by numerous neurohumoral gut peptides (i.e.,

ghrelin, cholecystokinin (CKK), glucagon-like peptide-1 and 2 (GLP-1, GLP-2) and peptide YY

(PYY)) secreted from enteroendocrine cells, which can in turn modulate gastrointestinal motility by

activating receptors on sensory, vagal and intrinsic afferent neurons (45). However, a previous study

reported that the satiety-enhancing effect of DF might be attributed to a slower absorption of nutrients

rather than gastric emptying (68). The modulating mechanisms are achieved in two ways: one is

through an increased viscosity of digesta caused by soluble fiber which restricts accessibility for the

digestive enzymes and thereby interferes hydrolysis, and also through physical obstruction of

nutrients in the gastrointestinal tract by insoluble fiber; another way is through an increased SCFA

production influencing the secretion of hormones involved in food intake, lipid storage, and energy

homeostasis (69). In agreement with that, a study reported that wheat fiber consumption for one year

9

increased postprandial GLP-1 secretion and plasma butyrate concentration in hyperinsulinemic

humans compared with a low cereal fiber intake (70). Another previous study found that postprandial

GLP-1 and PYY was increased with enzyme-hydrolyzed AXOS from wheat compared with low DF

control without changes in appetite ratings and energy intake, and assigned this to a low viscosity

(71).

2.2.4 Role of dietary fiber in glycemic control

DF is considered as component with a low glycemic index, which is beneficial for diabetic control.

Many studies have reported that DF had favorable impacts on MetS development, which not only

presented by increased satiety, delayed gastric emptying and reduced macronutrients absorption but

also improved insulin sensitivity (72). Moreover, increased GLP-1 secretion induced by DF can

increase pancreatic β-cell growth and improve insulin sensitivity, which can be a therapeutic target

for treating T2D (73). A high fiber rye bread diet reduced insulin and glucose concentrations in

plasma of humans and was associated with a reduced risk of developing T2D (74). Soluble DF has

been found to be very effective at controlling glucose response by reducing rate and extent of nutrient

absorption in the small intestine through forming viscous gels to slower diffusion rate of products

from starch digestion and inhibit enzyme accessibility to them (75). Previous studies from our group

also showed that a high content of DF in the form of AX content reduced acute glucose and insulin

responses in pigs (4, 76) and MetS subjects (57). The mechanisms of these functions are partly

associated with the viscous properties of the fiber by interfering with digestion and absorption,

altering peripheral glucose uptake and gastrointestinal hormone secretion (77). Moreover, a

supplement of 15 g/day of AX-rich fiber derived from wheat was proved to be effective in reducing

2 h-post glucose and insulin concentrations in people with T2D (78). With rye and enzyme-treated

wheat bran in the diet, less insulin was required to clear glucose from the blood stream in pigs (76).

However, the results from intervention studies are inconsistent. For instance, subjects with T2D

consuming 19 g/day DF from wheat bran had no apparent effect on glycemic control after 3 months

(79). Another study found that people who had high total DF or cereal fiber intake had less T2D risk

while the intake of soluble fiber was not associated with diabetes risk (80). Therefore, the

composition and structure of cereal fiber may be an important factor to consider regarding the effect

on glycemic control.

2.2.5 Role of dietary fiber in lipid metabolism

Consumption of adequate DF has been associated with reduced risk of CVD through a regulation of

plasma lipid profile. The intake of whole gain foods was found to reduce triglyceride and LDL levels

whereas a few studies found a significant change in HDL (81). A pooled analysis reported that each

10 g/d increment in total DF intake was associated with 27% of lowered risk of coronary death events

10

(82). Particularly, soluble DF has been found to have a more effectively hypocholesterolemic effect

than insoluble DF (83). According to meta-analyses of clinical studies, European Food Safety

Authority claims ~ 3 g of soluble fiber from oats or barley can reduce cholesterol by 0.13 mmol/L

and reduce LDL by 0.27 mmol/L (84). The potential mechanisms underlying lipid lowering effect of

soluble DF are possibly through many ways: directly reducing absorption of cholesterol induced by

viscosity; decreased enterohepatic pool of bile acid due to increased bile acid excretion in feces,

which can increase bile acid synthesis from cholesterol, and increase hepatic lipid uptake from the

blood; by the synthesis of SCFA to stimulate GLP-1 and PYY secretion and decrease lipid synthesis;

moreover, influencing the release of adipokines such as leptin and TNF-α to reduce the fat uptake in

adipocytes; or reducing glucose absorption and thereby lower insulin secretion, which can attenuate

the activation of enzyme responsible for cholesterol biosynthesis (81). A 3-week high cereal fiber

intake has been reported to reduce fasting plasma cholesterol and LDL in healthy subjects (85).

Compared with a placebo group, the 15 g/d consumption of AX concentrate for 6 weeks lowered the

fasting serum triglycerides in subjects with impaired glucose tolerance (86). For subjects with MetS,

a 12-week whole grain supplementation decreased postprandial triglyceride responses compared with

a refined cereal diet (87). Nevertheless, the results of studies investigating the effects of DF on lipid

profile is conflicting, especially for wheat. Some studies have assessed the effect of cereal fiber with

high wheat bran in the diet on blood lipid profile, showing no significant change in hyperlipidemic

or T2D subjects compared with a low fiber control (79, 88), whereas, surprisingly, a previous study

found that enzyme-treated wheat bran incorporated into bread increased total cholesterol in people

with abdominal obesity after 12 weeks compared with low DF diets (89). Moreover, a previous study

from our group has proven that rye was more efficient in reducing plasma cholesterol than wheat due

to a higher viscosity while wheat showed lower triglyceride levels in fasting plasma than rye (90).

Overall, inconsistent results may be caused by differences in fiber sources and constituents. Hence,

more studies on the long-term specific effects of DF especially from wheat consumption on lipid

profile in obese models are needed.

2.2.6 Role of dietary fiber in attenuating inflammation

DF intake has been found to reduce the risk of inflammatory bowel disease by establishing a healthy

gut environment. A healthy gut microfloral population can positively influence immune responses.

For instance, increased Bifidobacteria promoted by DF can increase IL-10 release and reduce IFN-γ

production by activation of T cells, whereas abundant Faecalibacterium prausnitzii can increase IL-

10 production and decrease IL-12 levels, which has been correlated with protection against

inflammatory bowel disease (91). Importantly, SCFA derived from fiber fermentation by the

intestinal microbiota is important in regulating acute inflammatory responses. Increased acetate and

11

propionate production activates G protein-coupled receptor (GPR) 43 and thereby produces anti-

inflammatory effects by inhibition of NF-κB (92), while butyrate has been thought to be an important

factor in maintaining normal functions of intestinal cells and is generally considered to be protective

against colorectal cancer and colitis (93). Butyrate has been reported to suppress colonic

inflammation in two ways: by inducing T-cell apoptosis to eliminate the source of inflammation, and

by suppressing IFN-γ mediated inflammation (94, 95). A recent study showed that high cereal DF

diets reduced TNF-α in obese adults compared with baseline, but no differences were observed in

inflammatory markers between the low and high DF groups (96). Additionally, a study found whole

grain intake in particular rye was directly linked to the reduction of fasting IL-6 concentration

compared with the refined grain diet (97). Besides systemic inflammation, moderating effects of DF

on intestinal inflammation have been shown and generally accompanied by improved microbiota

profile. Consumption of whole gain diet enriched in AX, β-glucan and cellulose in overweight or

obese subjects resulted in significant decreases in TNF-α which was associated with the change of

microbiota (98). Similarly, another study also found whole grain wheat consumption reduced TNF-

α and increased IL-10 after 8 weeks compared with refined wheat, which was correlated with

increased Bacteroides and Lactobacillus abundance in feces (99). To date, studies on the effects of

cereal fiber on low-grade systemic and intestinal inflammation are still limited, more studies should

be warranted.

2.2.7 Dietary fiber degradation and fermentation by microbiota

The relationship between DF intake, the gastrointestinal tract and host metabolism is complex (Figure

2). The fermentation of DF in the large intestine is a significant part of the beneficial health effects

of DF on the host, which can be used as a target for the MetS management. Beneficial effects of DF

on intestinal functions and on the gut microbiota profile and fermentation are related to the

physicochemical properties of the digesta, including transit time, fermentability, viscosity of the

digesta. Generally, the degradation of DF in the large intestine occurs in a hierarchical way:

oligosaccharides > starch residues > soluble NSP > insoluble NSP (55). Therefore, differences in the

chemical structure of DF can impact the fermentation processes by the gut microbiota. Almost all

undigested sugars, oligosaccharides, starch and soluble DF are degraded in the cecum and proximal

colon while insoluble AX and cellulose are degraded more distally in pigs (55). AXOS belongs to

soluble and non-viscous fiber and can act as prebiotics selectively utilized by microbiota to benefit

on host health [44]. Since the extent of fermentation is dependent on structure and decreases with

increasing structural complexity, AXOS can be fermented more rapidly due to its lower A:X ratio

and molecular weight compared with more substituted long-chain AX (100). After the degradation

of easily fermented AXOS and soluble AX in proximal part of large intestine, AX that has higher

12

A:X ratio and a more complex structure which is difficult to be degraded and persists until the distal

parts of colon. Therefore, rye and wheat milling fractions are different in fermentability due to their

different AX structure and composition. For instance, rye aleurone AX (more soluble, A:X = 0.42)

are fermented more readily and to a greater extent in pigs than pericarp AX (more insoluble, A:X =

1.04) (101), whereas wheat bran has less soluble AX than rye bran and therefore is more difficult to

ferment (59). Enzyme treatment of the bran could efficiently increase the solubility and availability

of some bioactive substances by changing cell wall structures. Enzyme treatment (mix of xylanase,

glucanase and cellulose) can for instance be used for wheat and rye bran to obtain higher AXOS

content, which can make the bran more fermentable (4, 62, 102).

Dietary fiber interventions has been extensively studied in last few decades for their beneficial effects

on maintaining intestinal homeostasis by promoting the growth of beneficial microorganisms and

reducing pathogenic bacteria. In general, Bacteroidetes are the potent producers of acetate and

propionate while butyrate is mainly produced from Firmicutes (46). It has been fund that patients

with diabetes have reduced levels of butyrate-producing species such as Faecalibacterium

prausnitzii, which belong to Firmicutes and have profound anti-inflammatory effects (46). Soluble

AX was reported to be a fermentable fiber source and efficiently increase colonic butyrate production

as well as butyrate-producing microbiota with positive effects on gut health and MetS (103). A

previous study from our group also demonstrated that with consumption of high DF diets, especially

rich in AX, higher number of Faecalibacterium prausnitzii, Roseburia intestinalis, and Lactobacillus

spp. in feces and higher pool size of butyrate were found compared to a western-style control diet in

pig models (4). Additionally, a study in healthy humans indicated that wheat bran enriched in AXOS

had beneficial effects on gut health by increased fecal Bifidobacteria abundance and SCFA

production (104).

Within the large intestine, SCFA is crucial for colonic health as they are involved in energy

homeostasis and enterocytes differentiation, but they also affect nutrient absorption and satiety by

regulating gastrointestinal hormones (105). More than 90% of the SCFA produced from carbohydrate

fermentation is acetate, propionate and butyrate, whereof acetate is the most abundant. In humans,

total SCFA concentration is low in the terminal ileum but high in all regions of the colon with 130

mM in the caecum to with 80 mM in the descending colon (106, 107). In pig models, SCFA

production mostly peaks in the proximal colon and then decline towards the distal colon (4). SCFA

can be utilized by the intestinal epithelium (~90%), and the remainder was either excreted with feces

or absorbed thereby transported to the liver via the portal vein (108). Therefore, increased DF intake

may increase the production and absorption of SCFA with abundant SCFA reaching peripheral

tissues to serve as energy substrates or participate in different host-signaling mechanisms. Butyrate

13

is the preferred energy source for colonocytes and supply 60-70% energy requirements locally which

is important for the prevention of colonic cancer (109). As the principal SCFA fermentation product

in the large intestine, acetate is the most abundant SCFA in the peripheral circulation and, is a

substrate for gluconeogenesis and the synthesis of cholesterol and triglycerides (110). Approximately

90% of absorbed propionate in the portal vein is metabolized by the liver as an efficient

gluconeogenic substrate and thereby only present at low peripheral concentration (111). Additionally,

SCFA have been shown to reduce the NF-κB activity and inhibit histone deacetylases activation by

activating GPR, therefore, profoundly affects inflammatory processes (92).

As the different SCFA play different roles, it can be more appropriate to emphasize the importance

of specific individual SCFA rather than the total SCFA production in relation to obesity. Therefore,

intake of specific DF sources need to increase in order to stimulate the anti-obesogenic SCFA

production. Butyrate, in particular, has received much attention as it contributes to a large part in gut

health and provides further benefits beyond that. Interventions with butyrate or increase butyrate-

producing bacteria abundance have been associated with beneficial effects on insulin sensitivity in

humans and animal models (7). Therefore, increased consumption of AX-enriched fiber such as

cereals and in particular bran could be an effective way to increase butyrate production in the colon

as AX is a butyrogenic substrate (4, 112).

Figure 2. The relationship between dietary fiber intake, the gastrointestinal tract and host metabolism. Solid

lines indicate well-studied effects of dietary fiber, dashed line indicate more controversial findings. SCFA,

short-chain fatty acids; GLP-1, glucagon-like peptide 1; PYY, peptide YY. Adapted from Müller et al. (2018)

(45).

14

2.3 Dietary protein

2.3.1 Sources, composition and properties

A normal daily protein intake accounts for 10-20% of dietary energy intake, with an average daily

amount of 80 g protein per capita from plant and animal sources (113, 114). In Europe, cereals

account for a major portion of plant protein in the form of bread, and the amount of cereal protein

ranges from 8% in rice to 12% in wheat, while legumes have a higher protein content (35%-40%

protein of soy beans) than cereals (115). The animal-based proteins are excellent sources for essential

amino acids and have high digestibility and net protein utilizations (113). Raw meat contains 20-25%

protein, egg proteins comprise ~13% of whole egg content and milk protein, consists of 20% whey

protein and 80% casein (115).

The protein quality is determined by amino acid composition, digestibility and availability, showing

better biological value for animal protein sources than for plant protein sources (116). For instance,

some abundant amino acids in animal-based products such as leucine are inadequate in plant protein,

whereas for cereals, the essential amino acids lysine, isoleucine, threonine and tryptophan are limiting

amino acids, and legumes are a poor source of sulfur amino acids (117). The digestibility of plant

proteins may be decreased if they are not properly processed. The cereal proteins are enclosed in the

cell walls which have complex combination with fiber and contains enzymes including protease

inhibitors, while protein originating from legumes may contain anti-nutritive factors such as trypsin

inhibitors in the soy beans, which can impact the digestion process and affect the nutritional value of

the protein (115). Moreover, a previous study concluded due to the deficiency in certain essential

amino acids of plant protein, other amino acids will not be properly used for protein synthesis in

muscle compared with animal proteins (116). However, there is a debate that processed meat and red

meat may be potential risk factors for chronic disease, whereas the functional properties and health

effects of dairy proteins have been supported by many studies (113).

2.3.2 Whey protein

Whey protein is a side-stream from cheese and casein production by coagulation of milk. The major

components of whey protein are β-lactoglobulin, α-lactalbumin, bovine serum albumin and

immunoglobulins. It is characterized by a high content of essential amino acids particularly branched

chain amino acids (> 20%) and sulfur-containing amino acids (118). There are three different forms

of whey protein: whey protein concentration (30-85% protein), whey protein isolate (> 90% protein)

and whey protein hydrolysates (119), which are composed of 80-90% protein, 0.5-8% fat and 0.5-

10% lactose (120). Whey protein hydrolysates are produced from whey protein by being pre-treated

with heat and proteolytic enzymes to change the complex structure of the polypeptides. As a result,

it can enhance endoprotease access and hydrolysis (121), leading to a faster and more efficient

15

absorption in the gut compared with intact proteins (122). Additionally, the hydrolysis of whey

protein can induce the release of bioactive peptides and amino acids, which have been suggested to

have physiological effects such as antioxidant, antimicrobial and anticancer activities (123).

2.3.3 Role of dietary protein in appetite

High protein diets are commonly used for body weight loss and maintenance due to an improved

satiety sensation than other macronutrients. However, there may be differences in satiating properties

depending on protein sources and digestion processes. Compared with egg and fish proteins, wheat

and pea proteins have been shown to be efficient in stimulating satiety hormones release (CKK and

GLP-1) in duodenal tissue of humans (124). A previous study also found that high protein intake

with whey or soy increased appetite sensation but without improving weight maintenance after

weight loss in humans (125). Whey protein has been found to be efficient in increasing satiety when

given as a preload or incorporated into the diet at normal protein concentration, which is linked with

the increased postprandial amino acid levels in plasma and secretion of satiety hormones compared

with casein or soy (126, 127). The likely mechanism can be that whey protein is digested and

absorbed faster than other proteins, inducing a sudden increase in plasma amino acids, which can

influence the release of satiety hormones (128) (Figure 3). Whey protein has a high BCAA content,

particularly leucine is able to activate mTOR signaling in ghrelin-producing cells leading to reduced

ghrelin secretion after food intake, whereas mTOR activation in intestinal L-cells leads to increased

GLP-1 secretion showing greater satiety and lower hunger scores (129). An acute study where whey

protein was given as a pre-meal to subjects with MetS showed that the pre-meal delayed gastric

emptying (130). Similar results were also observed in a long-term study with whey protein intake for

12 weeks, showing increased satiety in obese subjects (4). As a result, the interdependent relationship

between gastric emptying and secretion of incretin inhibited appetite, and the satiating effect of whey

protein can be used as a therapeutic tool for body weight control and obesity treatment.

2.3.4 Role of dietary protein in glycemic control

With the intake of dietary proteins, postprandial amino acid concentrations in plasma are elevated

and stimulate the secretion of insulin and glucagon and therefore influence the glucose metabolism

in peripheral tissues (115). By increasing the content of protein in the diet (from 15% to 30%) in

place of carbohydrate, a lower blood glucose response was observed in weight-stable persons with

T2D in a 5-week intervention study (131). In contrast, recent meta-analysis studies found no

significant improvement of a high-protein (> 20%) diet in glycemic control in T2D patients (132,

133). Some protein sources may help control blood glucose by influencing the release of gut

hormones. Whey protein has an incretin role, which is associated with pronounced insulin response

equivalent to that induced by eating a same amount of glucose (134). The high solubility and rapid

16

digestion of whey protein results in a rapid increase in plasma amino acid concentrations, especially

BCAA, which exerts a potent insulinotropic effect and plays an important role in glucose homeostasis

by the mTOR pathway as mentioned before (126, 128). Besides the effect of whey protein on insulin,

another mechanism of influencing postprandial glycemia could be that the action of incretin

hormones such as GLP-1 and PYY (135). A pre-meal of whey protein has been demonstrated to

stimulate insulin and glucagon secretion resulting in reduced blood glucose in subjects with MetS

(130) and subjects with or without T2D (136). The efficacy might change in long-term studies as

intake of whey protein for 12 weeks did not affect fasting insulin in overweight and obese subjects

(89). Correspondingly, a recent study suggested that the stimulating effect of high dietary protein on

insulin might be beneficial for insulin-resistant subjects but harmful for healthy subjects since it can

decrease insulin sensitivity in the long term (137).

Figure 3. Mechanism of action of whey protein on hormone release. BCAA, branched-chain fatty acids; GE,

gastric empty; GIP, glucose dependent insulinotropic peptide; GLP-1, glucagon-like polypeptide-1. Adapted

from Adams et al. (2016) (138).

2.3.5 Role of dietary protein in lipid metabolism

The effects of dietary protein on blood lipids are inconclusive. Recent meta-analysis studies have

showed divergent results in plasma lipid profile with a high-protein diet in T2D patients (132, 133).

Moreover, the effects of high protein diets on the regulation of lipid metabolism vary with protein

sources. A previous study reported that CVD risk factors could be reduced with increased intake of

plant protein at the expense of refined carbohydrates and processed meat products (139). Similarly,

a high plant protein, wheat gluten diet was found to induce decreased serum triacylglycerol levels

and reduce cardiovascular disease risk (140). A review concludes that meat products have

inconsistent effects on plasma lipid levels, whereas whey is more efficient in lowering cholesterol

than casein or soy protein (115). Additionally, supplementation of a fat rich mixed meal with whey

protein to individuals with or without T2D lowered postprandial lipemia compared with cod protein

and gluten in acute studies (8, 141), and intake of whey protein in combination with low DF for 12

17

weeks improved the postprandial lipid profile in abdominal obese subjects (89). A recent systematic

review stressed the importance of whey protein on improving multiple CVD risk factors including

blood pressure and cholesterol in overweight and obese subjects (142). The possible underlying

mechanisms of lowering effect of whey protein on lipid profiles seem to occur independent of body

weight change and may be caused by the promoted hepatic lipid metabolism, inhibited absorption of

fatty acids and cholesterol in the intestine induced by some bioactive components such as β-

lactalbumin of whey protein (12) (143, 144). However, studies on whey protein effects on lipid

metabolism have given controversial results. For example, supplementing whey protein to a fat-rich

meal did not affect triglycerides and chylomicron responses in subjects with MetS (130) or T2D

(136). With BCAA upon high fat diet, mice showed reduced body weight at the expense of

nonalcoholic fatty liver (NAFL) disease and injury (145) whereas another study demonstrated the

protective effect of oral whey proteins against NAFL in rats fed on high carbohydrate fat free diet

(146).

2.3.6 Role of dietary protein in low-grade inflammation

Studies of the effects of high protein intake on inflammatory status of obese models are limited.

However, due to the detrimental effects on the gut barrier of protein fermentation products such

hydrogen sulfide and indoles, most studies reveal that high protein consumption is associated with

an increased risk of inflammatory bowel diseases and relapse of ulcerative colitis (147, 148). The

results depend on protein sources; for instance, an energy-restricted high protein diet (30% energy)

specifically with meat protein but not with vegetable or fish protein was associated with higher degree

of inflammation in obese individuals with MetS compared with a control diet (15% energy from

protein) in an 8-week study (149). On other hand, the biologically active constituents of whey protein

such as immunoglobulins and lactoferrin are considered to be involved in maintaining immune

homeostasis and enhancing immune function and antibody synthesis (9, 150). A study found that

whey protein protected against gut inflammation in a rat model, which resulted from the stimulation

of intestinal mucin synthesis and modification of microbiota composition (151). Another study found

that whey protein intake in a high fat diet beneficially altered expression of genes profiles related to

inflammation particularly in adipose tissue of mice (152). Whereas in overweight and obese subjects,

pro-inflammatory markers including IL-6 and TNF-α were not affected by whey protein

supplementation after 12 weeks (153). In general, the effects of high protein content on low-grade

inflammation were inconclusive and dietary interventions of high protein containing whey products

should be further studied in obese models fed with high fat diets in a long-term.

18

2.3.7 Dietary protein degradation and fermentation by microbiota

Dietary proteins that escape digestion in the small intestine are available for protein fermentation and

microbiota growth in the large intestine. The amount of protein entering the colon depends on the

protein content of ingested food and protein digestibility. Generally, digestibility of dairy and animal

proteins exceeds 90%, higher than the digestibility of plant proteins (70-90%) (154). As the chyme

moves through the large intestine, fermentable carbohydrates become deficient in the distal part of

large intestine which results in an increased pH environment and is favorable for protein fermentation

(154). Undigested or endogenous protein in the large intestine can be fermented by many bacteria

such as Bacteroides spp., Popionibacterium spp., Streptococcus and Clostridium (155) into BCFA,

small amount of SCFA and sulfides, amines, ammonia and various phenols and indoles. The BCFA

(iso-butyrate, isocaproic acid and isovaleric acid) are formed by the metabolism of BCAA including

valine, leucine and isoleucine. Some metabolites including BCFA, phenols and indoles can only be

produced from colonic bacteria, considered as a marker for protein fermentation, for other

metabolites, such as ammonia, can be utilized for the metabolism and protein synthesis of

microorganisms (154).

The relative abundance of several species may be altered in high protein diets, for example, a study

found a high protein diet lead to decreased diversity of species of the Firmicutes and increased

diversity of Bacteroidetes species and Proteobacteria in an in vitro model of the proximal colon

(156). Furthermore, this study showed that the ratio of propionate production was increased due to

abundant Bacteroides with a high protein diet compared to a high carbohydrate diet. Another study

found although high protein diets reduced concentrations of butyrate producers in the cecum and

colon of a murine model, butyrate amount did not change due to increased luminal content with high

protein (157). Notably, the intestinal microbiota is sensitive to variations in dietary protein sources,

for instance, the use of highly digestible replacing less digestible protein sources may reduce the

growth of pathogenic species caused by protein fermentation (155). Highly digestible protein sources

such as whey protein can be digested by the hosts’ enzymes in the small intestine and almost not

available for microbial fermentation in the distal part of the large intestine (97% - 100% digestibility)

(158, 159). A previous study demonstrated that intake of whey protein might act as grow factor for

Lactobacillius and Bifidobacterium potentially by an amino acid composition-mediated mechanism,

which played a role in protection against colitis in a rat model (151). Moreover, it specifically

increased Lactobacillus and decreased Clostridium in high fat diet-fed mice (152).

Compared with DF, much less is known about the role of proteolytic fermentation by microbiota in

host health and metabolism. However, it is clear that the effect of protein on gut health is much more

controversial than DF because of the potentially harmful effects of end-products from protein

19

fermentation. Besides the use of high protein in the management of MetS, it is also important to

recognize the potential adverse effects of high protein for long term supplementation. Some protein

fermentation metabolites such as ammonia, phenol and hydrogen sulfide have potential toxicity to

the colonocytes (154). The production of BCFA comprises 5% ~ 10% of the total SCFA due to the

low level of polypeptides and amino acids reaching the colon compared to DF and their effects on

colonic epithelium have been poorly reported (148). With the exception of BCFA, excessive

accumulation of potentially harmful metabolites including ammonia and hydrogen sulfide is linked

with bowel disorders such as colon cancer (160). The excessive protein intake has been found to

stimulate the growth of potential pathogens such as Clostridium perfringens, and to reduce the

beneficial bacteria like Bifidobacteria (155). High protein diets increased BCFA especially high

protein and low carbohydrate diets decreased butyrate and Roseburia levels in obese humans (161).

In spite of insufficient evidence supporting a role of whey protein fermentation in the risk of bowel

disease, balanced diets with whey protein plus pre- or synbiotics are accepted to lessen the risk factors

(154). Slowly fermentable carbohydrates can be used in the diet to shift fermentation process toward

carbohydrate rather than protein and lower the potentially harmful metabolites production (155).

Therefore, abundant supply of DF may be used to influence protein fermentation in the large intestine

and reduce the detrimental effects on gut health (162). For instance, fermentable DF including wheat

bran can improve microbial ecology and reduce putatively toxic metabolites caused by protein

fermentation in pigs (163) and healthy adults (104). In addition, recent studies have found AXOS

consumption may potentially decrease the intestinal protein fermentation (164, 165).

20

3. Aim and hypotheses

This PhD project is part of two large collaborative projects (MERITS and ELIN) which study the

mechanisms of DF and its associated ingredients and dietary protein in mitigating MetS, the risk for

CVD and cancer. The first part of the PhD project was in MERITS to investigate if dietary strategies

based on high DF from AX enriched wheat and high protein from dairy, cereal and fish sources would

ameliorate the risk of MetS in young obese Göttingen Minipigs as a model for humans. In the second

project (ELIN) the aim was to study the effects of DF composition and treatment with antibiotics on

intestinal carbohydrate degradation and fermentation patterns and SCFA absorption in an intact

conventional pig model fed either an AX enriched rye bran based diet with or without antibiotic

treatment compared to a diet at similar DF level but based on cellulose.

Results from previous studies indicate that the beneficial effects of DF and protein on metabolic

abnormalities remain inconclusive and their interactions are still barely known. For a better

understanding of the physiological mechanisms behind the DF and protein on MetS, more studies

should be performed and clearly identified. Therefore, the overall objective of the present PhD project

was to study the underlying mechanisms of how DF associated components and protein affect

metabolic health by investigating intestinal DF degradation, microbiota composition, production and

absorption of SCFA and metabolic biomarkers associated with MetS.

The research hypotheses are:

1. High DF and protein content or a combination hereof can improve plasma insulin

response, lipid profile, inflammatory biomarkers and tissue gene expression associated

with MetS in a young obese Göttingen Minipig model (Paper I)

2. High DF and protein content or a combination hereof can influence the degradation of DF

and fermentation profile in the gut, including selectively stimulate beneficial bacteria

growth, enhance SCFA production especially butyrate and benefit on systemic health in

a young obese Göttingen Minipig model (Paper II)

3. Different DF composition will result in different profiles in nutrient digestion, DF

degradation and fermentation, wherein AX-enriched rye bran diet can reduce nutrient

digestibility, specifically increase butyrate production and absorption in conventional pigs

while antibiotic treatment will interfere with the fermentation pattern (Paper III)

21

4. Methods and methodological considerations

4.1 Experimental diets and design

4.1.1 The MERITS study

The study from the MERITS project described in Paper I and Paper II aimed to investigate the

possible ameliorating effects of DF and protein on MetS biomarkers in a porcine model with juvenile

obesity, which was induced by a long-term western style diet feeding. A total of 60 female Göttingen

Minipigs at 8 weeks of age were received in four separate batches, housed in pairs and fed a standard

chow diet (Special Diet Services, Dietex International, UK) in the first week. In the second week, 45

minipigs were gradually transitioned to a high fat low DF diet containing 20% fructose (LOFLOP),

and 15 minipigs were to a high amylose maize starch diet containing 20% fructose, which were

studied in another dietary intervention (166) and did not constitute the subject of this thesis. The 45

minipigs were fed ad libitum for 20 weeks to induce obesity and 10 minipigs per treatment were

planned for the DF and protein intervention and 5 minipigs were used as backup. Before the DF and

protein intervention, two minipigs died (one by liver biopsy and one by sudden unexpected death)

therefore, 43 minipigs were obtained after obesity development. All the pigs were allocated to one

of four diets for an 8-week intervention in a two by two factorial design to study the effects of low

versus high dietary levels of DF and protein (Figure 3). The four experimental diets were: a diet low

in DF and protein (LOFLOP), a diet low in DF and high in protein (LOFHIP; 6.8% whey protein

hydrolysate added), a diet high in DF and low in protein (HIFLOP; 20% enzyme-treated wheat bran

added) and a diet high in DF and protein (HIFHIP; 6.8% whey protein hydrolysate and 20% enzyme-

treated wheat bran added).

Figure 4. Experimental design and sample collection in the MERITS study. All pigs were fed low fiber low

protein diet (LIFLOP) for 20 weeks to induced obesity then switched to four diets including LOFLOP, low

22

fiber high protein (LOFHIP), high fiber low protein (HIFLOP) and high fiber high protein (HIFHIP) for 8

weeks.

Wheat bran was provided by Lantmӓnnen Cerealia AB (Malmø, Sweden) and was enzymatically

treated with cell wall-degrading enzymes (xylanase, glucanase and cellulase) by DuPont Industrial

Biosciences Aps (Brabrand, Denmark) as described by Nielsen et al (4). The DF composition of

wheat bran before and after the enzymatic treatment is listed in Table 3. A large fraction of AX are

insoluble in native wheat bran while enzymatic treatment highly increases AXOS composition from

insoluble AX.

Table 3. Dietary fiber composition of the wheat bran before and after enzymatic treatment (167).

g/100g DM Native wheat

bran

Enzyme-treated

wheat bran Change (g)

Total dietary fiber 45.0 45.0

HMW AX 22.5 18.1 -4.4

Soluble 4.0 4.4 0.4

Insoluble 18.5 13.7 -4.8

AXOS 1.9 6.8 4.9

Cellulose 8.3 8.6 0.3

β-glucan 3.3 1.2 -2.1

Klason lignin 6.3 8.1 1.8

HMW AX, high molecular weight arabinoxylan; AXOS, arabinoxylan oligosaccharides.

Whey protein hydrolysate (Lacprodan® HYDRO.REBUILD) was provided by Arla Foods

Ingredients Group P/S (Viby J, Denmark) and it contains 84% protein, 1.1% available carbohydrate

and 1.1% fat (167).

By the end of 8th week, anthropometric measurement was performed and backfat thickness was

measured by ultrasound scanning. The pigs were anaesthetized after 16 hours fasting overnight,

fasting blood samples were collected from the jugular vein and urine samples were collected by using

absorbent tampons. The liver, muscle and subcutaneous adipose tissues were collected by using a

biopsy pistol with specialized needle for gene expression analysis. After this 8-week intervention, all

the minipigs were euthanized, blood samples from portal and jugular vein and urine samples from

urinary bladder were collected in a non-fasting state. Then gut contents from ileum, cecum and colon

were collected. Liver, kidney and heart weights were recorded and liver tissues were frozen

immediately for liver fat analysis.

4.1.2 The ELIN study

The animal experiment of the ELIN project described in Paper III included 30 intact female

conventional pigs and was performed over 4 weeks. The animals were fed a standard swine diet

during the first week and a refined wheat fiber diet (Control) in the second week. In week 3, all the

23

animals were allocated to three groups, 10 pigs continued the Control diet and 20 pigs switched to a

rye bran based diet (RB). In week 4, ten of the RB fed pigs had daily intramuscular antibiotic

injections (RB+) with Streptopenprokain Rosco Vet (benzylpenicillinprocain and

dihydrosreptomycin) while the other 10 pigs remained untreated (RB-). The pigs were fed equal

amount of the diet every day based on average body weight. By the end of week 4, all the pigs were

anaesthetized 3 hours after the morning feed and antibiotic injection for RB+, blood samples were

collected from the carotid artery. Then all the pigs were subsequently euthanized, blood samples form

portal and hepatic veins and digesta from ileum, cecum and colon were collected for further analysis.

Figure 5. Experimental design of the study in ELIN project. The pigs (n = 30) were fed standard diet for one

week followed by refined wheat fiber diet (Control) for the second week. In week 3, 10 pigs continued on the

Control diet and 20 pigs were fed rye bran diet (RB) and 10 of them had daily antibiotic injection in week 4.

Adapted from Bolvig Sørensen (2016) (168).

4.2 Pigs as experimental models for humans

It is difficult to perform human studies due to ethical tissues and limitations of invasive procedures

such as strict control of environmental conditions and multiple sample collection. While, under

permission, pigs allow for induction of disease states, nutritional interventions and more invasive

sampling such as gastrointestinal tract. Pigs are omnivorous animal and is considered a suitable

animal model for metabolic and diabetic research due to the similarities in gastrointestinal

physiology, nutritional requirements and metabolism with humans (169). These properties likely

contribute to their comparable digestion and enteric microbiota with humans, which means pigs can

be a good animal model for studying the change of intestinal microbiota with the nutritional

interventions (170, 171). With similar digestive enzyme secretion rate and activity, pigs are often

used in nutrition research such as studies of digestion and absorption (172). However, the distal part

of small intestine of pigs has a much higher microbial density and thus can degrade indigestible

carbohydrates to a higher degree than humans (172). It should also be noticed that pigs have a larger

cecum with extensive fermentation than humans and the total intestinal length of pigs are different

from humans, but the ratio of intestinal length per kilogram body weight (~0.1) and gastrointestinal

fermentation profile (colon fermenter) are actually very similar, which makes the transit time

24

(average 20-30 h for total intestine) (173) and fermentation processes comparable between pigs and

humans (171). In addition, pigs have similar postprandial response profile of satiety-inducing

hormones especially when it comes to the quantitative and qualitative changes of GLP-1 compared

with humans (174). Furthermore, large quantities of digesta and tissue samples as well as multiple

sample collection of blood from different circulation sites of the body can be obtained at slaughter or

by catheterization by using the pig as a model (4, 76), which evidently is not possible in humans.

One of the considerations when using pigs as human models is their bigger size once fully mature,

while the Göttingen Minipig is one type of miniature pigs, which have small body size, are easily

handled and commonly used in biomedical studies such as for the development of metabolic

syndrome (175, 176). Like human beings, pigs are devoid of brown adipose tissue postnatally which

is an important consideration when used as biomedical model for energy metabolism and obesity

(177). Göttingen Minipigs have served as models for biomedical research in the fields of

atherosclerosis, hypertension and obesity, as they have similar metabolic features and cardiovascular

system and proportionally similar organ size with humans (177), and can easily deposit fat tissues

(178). Studies have found that Göttingen Minipigs are more easily to get obese compared with the

conventional pigs when fed ad libitum (179), and female minipigs are more prone to develop more

severe obesity than males (180). Therefore, young Göttingen Minipigs especially the female gender

have been found as a potential model for diet induced childhood/adolescent obesity and metabolic

syndrome by showing increased fat percentage, insulin resistance and plasma lipids (181). However,

it is noteworthy that the primary de novo lipogenesis site of humans is the liver but adipose tissue in

pigs, therefore, pigs do not easily get liver steatosis as humans do (182). Additionally, a high fat meal

does not activate coagulation factor VII in pigs therefore it is hard to develop a pig model with

thrombosis (183).

4.3 Obesity measurement

In humans, BMI is defined as weight divided by height squared (kg/m2). Although BMI is widely

used for obesity classification, it is not linearly related to the percentage of body fat. Skinfold

thickness is the double layer of skin and subcutaneous fat and generally are more highly correlated

with body fatness than BMI (184), whilst waist circumference measurement is linked with the

distribution of body fat and is strongly associated with metabolic risk and increased morbidity and

mortality from metabolic disorders independent of the BMI (185). Pig skin has a fixed subcutaneous

layer and is structurally similar to human (186), and the backfat thickness of pigs can be used to

detect subcutaneous fat deposition which is correlated with metabolic disorders (187). In humans,

the waist circumference is measure at the midpoint between the lower margin of the last palpable rib

and the top of the iliac crest (188). Similarly, anthropometric measurement such as body size (length),

25

abdominal and chest circumferences can be performed in pigs and to calculate porcine obesity index

(POI) based on the three parameters (Paper I) (189). Instead of BMI used in humans, body surface

area (BSA) is considered as an indicator of metabolic mass in pigs, it is calculated on the basis of

BW and height and can be an important parameter for the obesity measurement for miniature pigs

(186).

4.4 Insulin sensitivity assessment

The homeostasis model assessment (HOMA) is a common method used to assess IR (HOMA-IR)

and β-cell function (HOMA-β) on the basis of fasting glucose and insulin concentrations. HOMA is

a relatively simple, invasive and inexpensive assessment that has been shown to strongly predict the

changes in insulin sensitivity after a therapeutic intervention and development of T2D in large-scale

human studies (190, 191). A critical condition and assumption of HOMA is that subjects are strictly

fasting and in a basal steady state and an existing feedback loop between the liver and pancreatic β-

cell response; glucose levels are regulated by insulin-dependent hepatic glucose production and

insulin levels dependent on the pancreatic β-cells’ ability to respond to changes in glucose levels

(192). Therefore, HOMA-IR reflect the suppressive effect of insulin on hepatic glucose production

while HOMA-β reflect the response ability of β-cell to glucose-stimulated insulin secretion, and the

combination of HOMA-IR and HOMA-β represents the glucose-insulin homeostasis state. In the

present study (Paper I), HOMA was used to assess hepatic IR which is considered the major factor

contributing to the pre-diabetic state, impaired fasting glucose (193). However, HOMA can be used

in subjects with normal glucose tolerance, or mild T2D not in subjects with severely impaired β-cell

function as it is unable to secrete sufficient insulin to compensate for IR (194). In addition, secreted

C-peptide is considered to be a reliable representative for determining β-cell secretory function as it

is not extracted by the liver (195) whereas nearly 50% of the insulin secreted by β-cells are extracted

by the liver and has a large insulin inter-assay variation (196).

4.5 Gut microbiota analysis

In order to study the effects of DF and protein on intestinal microbiota profile, digesta samples from

the cecum, mid colon and rectum were collected at slaughter for the microbial analysis (Paper II).

After completing the DNA extractions and purifications of samples, PCR was performed to amplify

the 16S rRNA sequences of the V4 region and sequencing was completed by using Illumina MiSeq.

Operational taxonomic units (OUT) clustering is widely used to identify the bacterial composition of

samples based on the 16S rRNA sequences. In this study, quantitative Insight Into Microbial Ecology

(QIIME) software package was used to categorize microbiota based on OUT picking strategy and

analyze the communities at different taxonomic levels. Alfa diversity rarefaction curves were

generated with the increasing numbers of randomly sampled sequences per sample and beta diversity

26

was analyzed from the calculation of the weighted UniFrac metric in QIIME. Linear discriminant

analysis (LDA) effect size (LEfSe) analysis of microbiota in Paper II was performed with P < 0.05

and LDA > 2.0 using Galaxy (https://huttenhower.sph.harvard.edu/galaxy/). Sequencing of the 16S

rRNA gene is an accurate and faster method widely used for bacterial identification, however, it has

limitations in identifying bacteria to the genus and species levels due to the high sequence similarities

(197). Moreover, it is limited by short read lengths which is the reason that we exclude one sample

from colon when we did the analysis.

4.6 SCFA production and circulation

The SCFA from intestinal fermentation are rapidly absorbed from the gut lumen and a part of them

enters the portal and peripheral circulation, where they can modulate biological responses of the host.

In vitro studies using cell lines or inocula incubation are simple ways determining SCFA production,

however, it does not account for the complicated microbiome and SCFA absorption in the gut thereby

in vivo studies can be more accurate to quantify production and absorption of SCFA (198). Besides

directly measuring intestinal SCFA concentrations in a pig model, pool size are used to estimate the

intestinal production of SCFA which was calculated by the concentration of SCFA and the

corresponding net amount of gut content (Paper II and III). Although this quantification method does

not provide a true reflection of SCFA production since SCFA can be rapidly absorbed and utilized

by the colon epithelial cells, it is less expensed and easily handled than other methods such as the

stable isotope dilution technique which is considered a golden standard measurement (199). To assess

the absorption of SCFA, blood samples from mesenteric artery, portal and hepatic veins were

collected at slaughter to obtain circulating concentrations of SCFA (Paper III). As developing a

multi-catheterized pig model is a very invasive and challenging-handled process, the pigs were not

catheterized and therefore portal plasma blow was not directly determined. Previously, the blood

blow was found independent on dietary treatments and only varied in relation to the time after feeding

(76, 200). Therefore, the quantitative absorption of SCFA in Paper III was calculated by the portal-

arterial differences of SCFA concentrations and a fixed portal plasma flow rate per kg of BW, which

was also used by our previous study (201). The hepatic metabolism of SCFA was calculated by SCFA

concentrations in hepatic vein, portal vein and mesenteric artery, with the corresponded relative

contributions of portal and mesenteric arterial plasma flow to hepatic venous plasma flow (0.86 and

0.14, respectively) (202). The SCFA concentration in the mesenteric artery was assumed identical to

the concentrations in the hepatic artery. However, the plasma flows may be numerically different

among dietary treatments in Paper III, which should be taken into account when interpreting the

outcomes of this study.

27

4.7 Type and administration route of antibiotics

To mimic the influence of antibiotics on plant and enterolignan metabolism in humans, a commonly

used antibiotic in humans and pigs, penicillin was chose (168). Streptopenprokain Rosco Vet used in

Paper III is a cocktail of benzylpenicillinprocain and dihydrosreptomycin, a veterinary medicine

normally used in pigs of Denmark. Although the general administration route of penicillin in humans

is oral, intramuscular injection was performed in this study to ensure the identical dose of antibiotics

of each pigs. Moreover, we detected antibiotic metabolites in gut content which indicated that

intramuscular injection was able to exert its effect on both systemic and intestinal metabolism (203).

4.8 Statistical methods

Before conducting the animal experiment of project MERITS, the sample size was determined based

on power calculations for triglycerides and total cholesterol, and we calculated that 6-8 minipigs per

treatment were expected to give sufficient statistical power of 80% (α < 0.05). Therefore, 10

minipigs per treatment used in this 8-week intervention study was considered sufficient to give

statistical power. For the animal experiment in ELIN, 10 pigs per group were demonstrated to give

sufficient statistical power (α < 0.05; β > 0.80) according to the power calculations for pool size of

total and individual SCFA in the large intestine. In Paper I and Paper II, a MIXED procedure was

applied to assess the treatment effect of DF levels (high vs. low) and protein levels (high vs. low) as

well as the interaction between DF and protein. Moreover, a comparison of AX structure between

diets and ileal digesta was performed by only including the effect of DF levels in the model due to

absence of protein effect and interaction between fiber and protein. The segment effect and

interactions between fiber and/or protein and segment were included in the MIXED model when

analyzing the parameters of different intestinal segments in Paper II. In Paper III, the MIXED

procedure was applied to assess the effects of dietary treatment, intestinal segment or blood sampling

site and their interactions on parameters of gut content and plasma SCFA concentrations.

Furthermore, a simple analysis of variance (ANOVA) was used to analyze the effects of dietary

treatments on ileal carbohydrate composition and intestinal SCFA production in Paper III. When

repeated measurements were performed, pig was included as a random component and modelled with

the covariance structure by using the autoregressive type. Data that were not normally distributed

with homogeneous variance were log-transformed, and results presented as geometric means with

95% confidence intervals. When there was an interaction, pairwise comparison of groups were

performed and adjusted by multiple comparisons of Tukey–Kramer post hoc test. The levels of

significance was reported at P < 0.05 and tendency at P < 0.10. Correlation analysis was performed

by RStudio to study the relations between microbiota and intestinal metabolites including NSP and

SCFA concentrations, pool size and proportions.

28

5. Brief summary of results

Paper I

Dietary fibre and protein do not synergistically influence insulin, metabolic or inflammatory

biomarkers in young obese Göttingen Minipigs. Yetong Xu, Mihai Victor Curtasu, Knud Erik

Bach Knudsen, Mette Skou Hedemann, Peter Kappel Theil and Helle Nygaard Lærke

Manuscript published in British Journal of Nutrition, August 11, 2020.

Unhealthy western lifestyle energy dense diets have been associated with obesity and related

metabolic syndrome, and dietary interventions with healthy ingredients have received increased

attention in recent years. The objective of this study was to evaluate the effects of DF and protein

intake on metabolic responses in a young obese minipig model. A total of 43 obese minipigs (30-

week old) were randomly assigned to one of four diets with low or high levels of DF and protein for

8 weeks with ad libitum feeding. It was hypothesized that increased DF and protein could improve

the glycemic and insulin response, lipid profile and inflammatory status, alleviating the metabolic

abnormalities related to MetS.

In still-growing minipigs, the weight gain was reduced with high DF compared with the low DF

content, and was increased with high dietary protein compared with low protein content. However,

neither high DF nor high protein mitigated the glucose response, lipid profile or inflammatory status

in this obese minipig model. High DF increased C-peptide levels in the non-fasting state which means

that the beta-cell secretory function is improved. Gene expression analysis showed that high protein

upregulated the gene expression of FBP1 related to gluconeogenesis in the liver, whereas high DF

downregulated the gene expression of FASN involved in fatty acid synthesis in adipose tissue and

also upregulated gene expression of LEPR which was associated with leptin sensitivity. Contrast to

our hypothesis, a combination of high DF with high protein did not show any alleviating effects on

markers of metabolic abnormalities. Synergistic effect of high DF and protein was not detected in

the present study, whereas the results confirmed the advantage of high DF as a dietary strategy to

improve the metabolic homeostasis in obesity.

29

Paper II

Effects of dietary fibre and protein content on intestinal fibre degradation, short-chain fatty

acid and microbiota composition in a high-fat fructose-rich diet induced obese Göttingen

Minipig model. Yetong Xu, Mihai Victor Curtasu, Zachary Bendiks, Maria L. Marco, Natalja

Nørskov, Knud Erik Bach Knudsen, Mette Skou Hedemann and Helle Nygaard Lærke

Manuscript submitted to Food & Function. 2020.

The gut environment can be changed with obesity and MetS, and the dietary composition can

modulate metabolic homeostasis by influencing the gut microbiota and its formation of fermentation

products. The present study aimed to investigate the impacts of DF and protein on fiber degradation,

microbiota change, SCFA production as well as their correlations with each other. The study was

performed by using 43 young obese Göttingen Minipigs fed low or high levels of DF and protein ad

libitum for 8 weeks. We hypothesized that DF and dietary protein content can influence the intestinal

DF structure, stimulate the growth of beneficial bacteria and enhance SCFA production especially

butyrate, which can be linked to improved systemic health and attenuated obesity related diseases.

AX in the high DF diets was continuously degraded until the mid colon in this obese Göttingen

Minipig model. In agreement with our hypothesis, we found that high DF intake increased fecal

microbial diversity and the relative abundance of butyrogenic taxa including Blautia and

Faecalibacterium, but reduced SCFA concentrations except for butyrate, and slightly improved

butyrate pool in the large intestine compared to low DF groups. Importantly, high DF content in the

diet attenuated protein fermentation, leading to a reduced proportion of BCFA in gut contents. High

protein intake increased SCFA, acetate, propionate and BCFA concentrations and did not show

prebiotic effects. Moreover, high protein significantly changed circulating SCFA profile associated

with increased gluconeogenesis, whilst high DF increased circulating butyrate concentrations.

According to correlation analysis, AX contents in cecum and colon were positively linked with the

relative abundance of Blautia and body weight of pigs was negatively correlated with the relative

abundance of Ruminococcus. Overall, AX-enriched high DF diets improved intestinal environment

which possibly was linked with modulated metabolic health, and high protein influenced SCFA

profile but did not show beneficial effects on microbial profile.

30

Paper III

The role of rye bran and antibiotics on the digestion, fermentation process and short-chain

fatty acid production and absorption of pigs. Yetong Xu, Anne Katrine Bolvig Sørensen, Brendan

McCarthy-Sinclair, Maria L. Marco, Knud Erik Bach Knudsen, Mette Skou Hedemann and Helle

Nygaard Lærke

Manuscript submitted to British Journal of Nutrition. July, 2020.

SCFA are the main products from microbiota fermentation of DF and have been associated with

improved metabolic health. AX has been demonstrated to be a substrate for butyrate production. The

aim of present study was to explore the changes of nutrient digestion, fermentation pattern and SCFA

absorption in response to differences in DF composition and effects of the use of antibiotics. Thirty

intact conventional pigs were randomly assigned to 3 treatments, which included a cellulose rich

refined wheat fiber diet as Control, and a rye bran (RB) diet with or without antibiotic treatment of

the pigs. It was hypothesized that the RB diet would reduce nutrient digestibility and stimulate

fermentation process, specifically increase butyrate production and absorption compared with the

refined wheat fiber at iso-DF levels in the diets, whereas antibiotic treatment would disrupt this

process.

The AX-rich rye bran based diet reduced protein and fat digestibility, and AX was fermented more

proximally than cellulose in the colon. Furthermore, we found that the production and proportions of

butyrate in cecum and BCFA in colon were increased with AX-rich RB diet without showing

significant change in net portal absorption of butyrate. The refined wheat fiber diet increased total

SCFA, acetate and propionate production as well as absorption, which correlated with the higher

plasma cholesterol and LDL concentrations. In agreement with our hypothesis, the increased butyrate

production and proportion with the RB diet was negated by antibiotic treatment, which was also

reflected by a change of microbiota composition. The results highlights important differences in the

degradation and fermentation pattern of DF depending on sources and structures, which in turn

influenced systematic SCFA circulation and was of importance for plasma lipid profile.

31

6. Paper

6.1 Paper I

Dietary fibre and protein do not synergistically influence insulin, metabolic or inflammatory

biomarkers in young obese Göttingen Minipigs.

Yetong Xu, Mihai Victor Curtasu, Knud Erik Bach Knudsen, Mette Skou Hedemann, Peter Kappel

Theil and Helle Nygaard Lærke

British Journal of Nutrition, August 11, 2020.

In print, DOI: https://doi.org/10.1017/S0007114520003141.

Accepted manuscript

This peer-reviewed article has been accepted for publication but not yet copyedited or

typeset, and so may be subject to change during the production process. The article is

considered published and may be cited using its DOI

10.1017/S0007114520003141

The British Journal of Nutrition is published by Cambridge University Press on behalf of

The Nutrition Society

Dietary fibre and protein do not synergistically influence insulin, metabolic or

inflammatory biomarkers in young obese Göttingen Minipigs

Yetong Xu*, Mihai Victor Curtasu, Knud Erik Bach Knudsen, Mette Skou Hedemann,

Peter Kappel Theil and Helle Nygaard Lærke

Department of Animal Science, Aarhus University, DK-8830 Tjele, Denmark

* Corresponding Author: Yetong Xu; Email: yetongxu@anis.au.dk; Tel.: +4550641816

Short title: Metabolic effects of fibre and protein in obesity

Keywords: wheat bran; whey protein; obesity; metabolic syndrome; miniature pig model

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

32

Accepted manuscript

ABSTRACT

The effects of dietary fibre (DF) and protein on insulin response, lipidemia and

inflammatory biomarkers were studied in a model experiment with juvenile obese

Göttingen Minipigs. After 20 weeks feeding on a high-fat fructose-rich low-DF diet, forty-

three 30-week old minipigs (31.3 ± 4.0 kg body weight) were allocated to low or high DF

and protein diets for 8 weeks in a 2 × 2 factorial design. High DF contents decreased (P =

0.006) while high protein increased (P < 0.001) the daily gain. High protein contents

increased fasting plasma concentrations of glucose (P = 0.008), non-esterified fatty acid

(P =0.015), ghrelin (P = 0.008) and non-fasting LDL:HDL ratios (P = 0.015). High DF

increased ghrelin (P = 0.036) and C-peptide levels (P = 0.011) in the non-fasting state.

High protein increased the gene expression of fructose-bisphosphatase 1 in liver tissue (P

= 0.043), whereas DF decreased fatty acid synthase expression in adipose tissue (P =

0.035). Interactions between DF and protein level were observed in the expression of

leptin receptor in adipose tissue (P = 0.031) and of peroxisome-proliferator activated

receptors-γ in muscle (P = 0.018) and adipose tissue (P = 0.004). In conclusion, high DF

intake reduced weight gain and had potential benefit on beta-cell secretory function, but

without effect on the lipid profile in this young obese model. High dietary protein by

supplementing with whey protein did not improve insulin sensitivity or lipidemia, and

combining high DF with high protein did not alleviate the risk of metabolic abnormalities.

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

33

INTRODUCTION

Globally, the obesity epidemic is increasing at an alarming rate and especially the increase

in childhood obesity is of concern(1)

. Although obesity itself is not a disease, the condition

often causes endocrine and metabolic changes clustered into the so called metabolic

syndrome (MetS) that typically arises in parallel with obesity(2-4)

. MetS represents a

complex pathophysiological cluster characterized by excessive ectopic lipid accumulation

causing inflammation, insulin resistance (IR), hypertension and hyperlipidemia leading to

type 2 diabetes (T2D) and cardiovascular disease(5)

.

Lifestyle modifications, especially through a healthy diet, have attracted great interest to

alleviate MetS development. Increased consumption of dietary fibre (DF) has been proven

to ameliorate postprandial dyslipidemia and insulin responses, which could effectively

modulate T2D and cardiovascular disease(6, 7)

. DF can influence digestion and absorption

processes at all sites of the gastrointestinal tract. Thus, DF, depending on its composition,

can delay gastric emptying(8)

, impede the digestion processes in the small intestine(9)

,

influence the metabolic outcome of the microbiota in the large intestine(10)

, and influence

the release of gastrointestinal satiety hormones(11)

, which alone or in combination will

improve postprandial glycaemia and insulin responses(9)

. Although acute intervention

studies have indicated soluble DF as most efficient in regulating glycemia and cholesterol

levels(12, 13)

, controlled intervention studies have indicated that diets high in insoluble cereal

fibre may improve IR(14)

to a larger extent than what is the case with more soluble DF

sources(12, 15)

. Another dietary constituent that is known to influence MetS is content and

quality of protein(16)

. A recent study showed that a diet high in either animal or plant protein

reduced liver fat, IR and hepatic inflammation(17)

. It has also been found that milk protein,

whey protein in particular, has insulinotropic properties by affecting the release of incretin

hormones and insulinotropic amino acids(18)

. Moreover, whey protein provided as a pre-

meal has been found to delay gastric emptying(19)

and reduce glycaemic response after

consuming a carbohydrate-rich diet(20, 21)

.

Although DF and protein have been the subject of many investigations, the interactive

effects of DF and protein are poorly understood. In the present study, we investigated the

possible ameliorating effects on MetS biomarkers in an obese porcine model for childhood

obesity(22)

by the use of enzyme-treated wheat bran as DF supplement and hydrolysed whey

protein as protein supplement alone or combined. Göttingen Minipigs have a small body

size and have previously been used for the development of MetS models because of the

anatomical and physiological similarities with adolescent and adult humans(23, 24)

. We

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

34

hypothesised that DF and whey protein and their combinations will have a positive effect

on plasma insulin, lipid biomarkers, inflammatory biomarkers and expression of genes

associated with nutrient metabolism in young obese Göttingen Minipigs.

MATERIALS AND METHODS

Animals and handling

The care and housing of animals were done according to Danish laws and regulations

regarding humane care and use of animals in research (The Danish Ministry of Justice, Act

on Animal Experiments no. 474 of May 15 2014, as stipulated in the executive order no. 12

or January 07 2016) and performed under license obtained from the Danish Animal

Experimentation Inspectorate, Ministry of Food, Agriculture and Fisheries. Animal health

was monitored closely by observing behavior (appetite, activity, interactions with humans

and neighboring pigs) and signs of disease.

A total of forty-three female Göttingen Minipigs (Ellegaard Göttingen Minipigs,

Dalmose, Denmark), raised at Aarhus University, Department of Animal Science, Foulum,

Denmark were used in the study. The minipigs were received at the facilities at eight weeks

of age in four separate blocks over 8 months. At arrival, the minipigs were housed in pairs

and fed a standard minipig chow diet (Special Diet Services, Dietex International, UK)

according to breeders’ recommendations for one week followed by separation and a 1-week

gradual transition to a high-fat low DF diet containing 20% fructose (LOFLOP) for the

following 20 weeks of ad libitum feeding to induce obesity. The minipigs were housed in

individual pens (1.5 m × 2.4 m for pigs < 50 kg, and 3 m × 2.4 m > 50 kg). Water was

provided ad libitum from drinking nipples, and the pens were provided with wood shavings

as bedding material for the first 22 weeks. When entering the intervention of the current

experiment the bedding was removed and replaced with a rubber mattress, and toys were

provided to satisfy the minipigs’ rooting behavior.

At an average body weight (BW) of 31.3 ± 4.0 kg (30-weeks old), the minipigs were

transferred to one of four experimental diets over 3 days to gradually reach 100% of the

experimental diet on day 4 using fixed amounts of feed corresponding to 2.6% of BW. The

four experimental diets were: A diet low in DF and protein (LOFLOP), a diet low in DF

and high in protein (LOFHIP; 6.8% whey protein hydrolysate added), a diet high in DF and

low in protein (HIFLOP; 20% enzyme-treated wheat bran added) and a diet high in DF and

protein (HIFHIP; 6.8% whey protein hydrolysate and 20% enzyme-treated wheat bran

added). Wheat bran was delivered by Lantmӓnnen Cerealia AB (Malmø, Sweden) and was

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

35

enzymatically treated with cell wall-degrading enzymes (xylanase, glucanase and cellulase)

by DuPont Industrial Biosciences Aps (Brabrand, Denmark) as described by Nielsen et

al(10)

. Whey protein hydrolysate (Lacprodan®

HYDRO.REBUILD) was provided by Arla

Foods Ingredients Group P/S (Viby J, Denmark). The experimental diets were produced at

the feed mill at Aarhus University and stored at -20 ºC. Ingredients and nutrient

composition of the experimental diets are shown in Table 1.

After the transition to the experimental diets, voluntary feed intake was recorded for a

total of seven weeks (first week, and then for three times every second week ). BW was

measured at baseline, at forth and eighth weeks after starting the intervention to calculate

daily weight gain and weight gain per unit of feed intake. Anthropometric measurements

including length, chest and abdomen circumference (cm), were taken using measuring tape

at the end of the experimental intervention (week 28).

Sample collection

Before (week 20 of ad libitum feeding with LOFLOP) and after the intervention (week 28),

fasting blood and tissue samples were taken as described in detail by Curtasu(22)

. Briefly,

after an overnight fasting (feed removed at 15.00), the minipigs were anesthetised with 0.1

ml/kg body weight of Zolitil-mixture containing 50 mg/mL tiletamine/zolazepam (Vibrac

SA, Carros, France), 2.5 mg/mL butorphanol (Torbugesic® Vet, Scan Vet Animal Health

A/S, Fredensborg, Denmark), 12.5 mg/mL ketamine (Ketaminol Vet, Intervet Denmark,

Skovlunde, Denmark), and 12.5 mg/mL xylazine (Rompun, Bayer Health Care AG,

Leverkusen, Germany) and blood samples were collected from the jugular vein by veno-

puncture. Following, the minipigs were moved into a left recumbent position and a liver

biopsy sample (50 mg) was taken with a biopsy pistol (Pro-MagTM I 2.5, Argon Medical

Devices, Inc.) and a 14G × 10 cm needle (Argon Medical Devices, Inc.) after shaving and

disinfection of the skin with 0.5% chlorhexidine solution in 85% alcohol (Abena A/S,

Denmark) and subcutaneous injection of a local anesthetic (Procamidor VET, 20 mg/ml,

Richter Pharma, AG). Biopsies were supervised by ultrasound scanning by using a 6-18

MHz linear probe (MyLabTM Five VET, Biosound Esaote, Inc.), and the incision closed

with a surgical staple. For sampling of subcutaneous fat and muscle tissue, the right hind

leg of the minipig was cleaned, sterilised and anesthetised locally as described for the liver

biopsy. After a 15-20 mm incision of the skin, approximatively 100 mg of subcutaneous

adipose tissue (sAT) was collected, followed by collection of 50-100 mg muscle tissue from

the semitendinosus muscle by use of the biopsy pistol. The sAT were snap frozen in liquid

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

36

N2 and stored at -80ºC until analysis. Muscle and liver tissues were placed in sterile tubes

containing RNAlater (Sigma-Aldrich Co. LLC).

After sampling, a few drops of Streptocillin® vet. (Boehringer Ingelheim Danmark

A/S) were administered to the incision site and closed with surgical staples. Urine was

collected by attaching an absorbent tampon under the tail (covering the urethra) using

Omniplast adhesive fabric tape (Hartmann, Germany) before the minipigs were transferred

back to their pens for recovery.

Following the 8-week intervention, the animals were weighed and euthanised in the 9th

week over 3 days. At euthanasia at the end of the experiment, the minipigs had not had their

feed removed, so samples collected reflect the non-fasting state. For sampling, the minipigs

were weighed and anaesthetised with Zolitil mixture as described above. The pigs were

then fitted with a catheter in the ear vein for possible supplementary anesthesia and put in a

supine position for sampling from the jugular vein. Following, the abdominal cavity was

opened by a midline incision and a blood sample was quickly taken from the portal vein.

The minipigs were then euthanised with an overdose of sodium pentobarbital followed by

exsanguination. The entire gastrointestinal tract was removed, and the small intestine and

colon sections were tied off to keep contents in place while measuring the length. Liver,

kidney and heart weights were recorded, and tissue samples from the right medial liver lobe

were frozen immediately for fat analysis. Urine samples were collected by removing the

urinary bladder and direct puncture to determine pH at room temperature with a pH meter

and a sample was stored at -80ºC for further analysis. Backfat thickness was measured by

ultrasound scanning (MyLabTM Five VET, Biosound Esaote, Inc.) with a 6-18 MHz linear

probe. Measurements were taken in the area of the longissimus dorsi muscle over the last

rib while the animal was in a hanging position. The distance from the skin to the last layer

of fat was measured at each recording and exported for calculation.

Blood samples taken at fasting and at slaughter were collected in Lithium Heparin

(LiHep), K3 ethylene-diamine-tetra acetic acid (EDTA), and K3 EDTA/Aprotinin inhibitor

(10000 KIU/mL blood, Nordic Pharma Ltd.), centrifuged for 12 min at 3300 rpm (1220 rcf)

at 4ºC, and plasma was aliquoted for analyses and stored at -80ºC.

Analytical methods

All chemical analyses of the diets were performed in duplicate on freeze-dried samples and

analysed as previously described(10)

. Liver fat was determined based on Bligh and Dyer’s

method(25)

. Briefly, liver samples were homogenised in double amount of methanol using

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

37

an Ultra-Turrax homogeniser (IKA Labortechnik) in an ice bath. Homogenate (600 mg)

was mixed with 1.0 ml water, 1.5 ml methanol and 1.0 ml chloroform, the mixture was

shaken for 1 min and 1.0 ml water and 2.0 ml chloroform was added and shaken for 1 min.

The mixture was centrifuged at 3000 rpm (1220 rcf) for 5 min to get phase separation and

2.00 ml of the lower chloroform phase was taken out, dried and the residue was weighed for

determination of fat content.

Concentrations of glucose, fructosamine, lactate, non-esterified fatty acids (NEFA),

high density lipoproteins (HDL), low density lipoproteins (LDL), total cholesterol,

triglycerides, albumin, AST (aspartate transaminase), ALT (alanine transaminase) and GGT

(gamma-glutamyltransferase) in LiHep plasma were analysed using the ADVIA 1650

Chemistry system (Siemens Diagnostics, Tarrytown, NY, USA) according to the

manufacturer’s instructions (Siemens Diagnostics Clinical Methods for ADVIA 1650).

Urinary glucose, creatinine and total protein analyses were conducted with the same

system. Millipore MILLIPLEX MAP Human Metabolic Hormone bead panel kit

(HMHEMAG-34K, Merck Millipore, Merck KGaA, Darmstadt, Germany) was used to

determine insulin, glucagon, C-peptide and ghrelin (active), GIP and total glucagon-like

peptide-1 (GLP-1)) in in K3 EDTA/Aprotinin inhibitor plasma. Cytokines (IFN-γ, IL-2, IL-

4, IL-10, IL-12, IL-18) in K3 EDTA plasma were measured using a Millipore MILLIPLEX

MAP Porcine bead panel kit (PCYTMAG-23K, Merck Millipore, Merck KGaA, Darmstadt,

Germany). Both kits were run according to the manufacturer’s instructions on a Luminex

MAGPIX system (Luminex Corporation, TX, USA).

Real-time reverse transcriptase polymerase chain reaction (RT-PCR) was performed on

liver, muscle and sAT to analyse gene expression (Table S1). In liver tissue, total RNA

extraction was performed using the NucleoSpin RNA Plus kit (Macherey-Nagel GmbH &

Co., KG., Duren, Germany) according to the manufacturer’s instructions. Total RNA

extraction from muscle and sAT was operated using TRI Reagent® Solution (Ambion,

Applied Biosystems, Stockholm, Sweden) based on the manufacturer’s guidelines. RNA

transcription, cDNA synthesis and RT-PCR quantification were conducted as previously

described(26)

. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH), β-actin and

hypoxanthine phosphoribosyl transferase 1 (HPRT1) were tested as potential housekeeping

genes (HKG). Gene expression data was obtained as Ct values (cycle number for which

logarithmic plots cross a calculated threshold) and used to calculate ΔCt values as the

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

38

difference between Ct of the target gene and mean Ct of HKG. Liver GAPDH exhibited

changes with regard to DF and protein interaction whereas sAT GAPDH exhibited changes

with regard to protein effect. As a result, mean of β-actin and HPRT1 was used as HKG for

liver tissue, whereas mean of β-actin and GAPDH was used as HKG for muscle and β-

actin were used as HKG for sAT. Relative gene-expression was determined using the

(1+efficiencies)-ΔΔCt

method, were ΔΔCt = ΔCttreatment – ΔCtLOFLOP. Results were reported

as fold changes. C-reactive protein (CRP) expression in muscle and sAT as well as muscle

leptin receptor (LEPR) expression are not reported due to detection limit.

Calculations

Porcine obesity index (POI) was calculated as previously described(27)

:

POI (L/cm) = (π × (1/3) × BS × (Abr2 + Cr

2 + Ab × Cc))/BS × 10

3, (1)

where BS, Abr, Cr, Ab and Cc are body size (length), abdomen radius, chest radius,

abdomen circumference and chest circumference, respectively.

Body surface area (BSA) was calculated as follows(28)

:

BSA (m2) = 0.121BW

0.575, (2)

Weights of liver, heart and kidney and lengths of small intestine and colon at slaughter

were calculated relative to BW (kg) at slaughter.

Values of fasting blood glucose (mM) and insulin (MIU/L) were used to calculate the

homeostatic model assessment for insulin resistance (HOMA-IR) and beta-cell function

(HOMA-β) as previously described(29)

:

HOMA-IR = (insulin × glucose)/22.5, (3)

HOMA-β = 20 × insulin/(glucose – 3.5). (4)

Statistical analysis

The pig was regarded as the experimental unit. According to the power calculations for

triglycerides and total cholesterol, 6-8 minipigs completing the study were expected to give

sufficient statistical power (α < 0.05; β = 0.80). All data analysis was accomplished using

the MIXED procedure of SAS (SAS Institute, Inc.) based on the normal mixed model:

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

39

Yijk = μ + αi + βj + (αβ)ij + tk + εijk, (5)

where Yijk is the dependent variable; μ is the overall mean; αi and βj are the fixed effects

of DF level (i = LOF or HIF), and protein level (j = LOP or HIP); (αβ)ij is the interaction

among fixed effects; tk is the random effect of block (k = 1, 2, 3 or 4), and εijk is the residual

error. If an interaction was detected, we conducted a pairwise comparison of groups

adjusted by multiple comparisons of Tukey–Kramer post hoc test. Values are presented as

least square means with standard error of the means (SEM). A log-transformation was

performed for HOMA-β, ratios of glucose:creatinine and protein:creatinine in urine at the

end of experiment, urinary ratios of glucose:creatinine and protein:creatinine at baseline to

obtain variance homogeneity and values of back-transformed data are presented with a 95%

confidence interval (CI). Pearson correlation of delta values for glucose in the portal vein

and jugular vein against daily starch intake was performed by using GraphPad Prism 8.0

(GraphPad Software Inc.). Effects are reported significant when P < 0.05 while P < 0.10

was considered as a tendency. For the statistical analysis of anthropometric measurements

block 2 was excluded due to absence of data.

RESULTS

Diets, feed intake, body weight, morphometric and organ indices

The diets contained almost equal amounts of fat and gross energy as shown in Table 1. As

planned, the CP content was 56% higher in the two high protein diets and the DF content

was approximately doubled in the two high-DF diets, it was also reflected in the relative

energy contributions from protein and DF. The increase was in form of NSP, AX, AXOS,

and Klason lignin due to the inclusion of enzyme-treated wheat bran in the high DF diets.

Along with the higher CP and DF, the starch content was lower: 150 g/kg DM in the

HIFHIP diet compared to 344 g/kg DM in the LOFLOP diet.

The 43 minipigs that for 20 weeks had been intervened by an energy dense diet (diet

LOFLOP) with baseline characteristics’ as in Table S2, increased their BW from 31.3 ± 4.0

kg before the intervention to 46.8 ± 5.4 kg after 8 weeks with no difference between the

four diets (data not shown; P > 0.10). The daily feed intake recorded in weeks 22-28

showed a significant interaction between DF and protein (P = 0.038) with the lowest feed

intake observed with HIFLOP (Figure 1). The daily weight gain as well as the gain per unit

of feed intake was lower with the high fibre diets (-20%, P = 0.006; -11%, P = 0.036)

compared to low fibre diets and higher with the high protein diets (+36%, P < 0.001; +31%, a

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

40

P < 0.001) compared to low protein diets (Figure 1). At the end of intervention, there was

no significant effect of dietary treatment on any of the morphometric measurements of the

pigs either in length, chest circumference, abdomen circumference, POI, or BSA, and

backfat thickness was also not significantly affected by either DF or protein levels (Figure

S1).

The relative weight (% of BW) of liver, heart, kidney, and length (m/kg) of small

intestine and colon as well as liver fat content at slaughter is shown in Figure S2. The

relative weight of the kidney was higher in the protein groups (0.20 vs 0.17% of BW, P =

0.010) whereas none of the other responses were affected by dietary treatment.

Clinical parameters in fasting and non-fasting plasma

Plasma concentrations of clinical parameters in the fasting (jugular vein) and non-fasting

state (jugular and portal vein taken at euthanasia) are shown in Table 2. At fasting, there

were significantly higher plasma glucose concentrations with the high protein diets

compared to low protein diets (6.9 vs 5.8 mM, P = 0.008), and an interaction of DF with

protein in fructosamine (P = 0.022), but no difference between groups by pairwise

comparisons. There were also higher NEFA concentrations (241 vs 161 μM, P = 0.015),

higher albumin (44 vs 40 g/L, P = 0.018) and AST concentrations (28 vs 23 U/L, P = 0.043)

with the high protein compared to low protein diets in fasting plasma. In the non-fasting

samples, a borderline significant increase in LDL concentration was observed with high

protein diets in jugular (1.7 vs 1.2 mM, P = 0.054) and portal venous plasma (1.8 vs 1.3

mM, P = 0.049). Moreover, the higher protein content also resulted in significantly higher

LDL:HDL ratios in plasma from both the jugular vein (0.88 vs 0.65, P = 0.015) and the

portal vein (0.92 vs 0.67, P = 0.014) in the non-fasting state as compared with pigs fed the

low protein diets. For the other responses no significant effects of dietary treatment were

seen. Non-fasting plasma levels of glucose and lactate were not significantly changed by

either DF or protein levels, but, surprisingly, glucose for the HIFHIP diet was higher in the

jugular vein than the portal vein. Overall, the difference in glucose concentration between

the portal and jugular vein was positively related to the intake of starch (r = 0.494, P >

0.001, Figure S3).

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

41

Fasting and non-fasting concentrations of plasma hormones and inflammatory

cytokines

Fasting and non-fasting concentrations of hormones and inflammatory cytokines in plasma

from the jugular vein are shown in Table 3. Fasting ghrelin concentrations were

significantly higher with the high dietary protein diets (28 vs 15 pg/ml, P = 0.008) but

higher with the high DF diets in the non-fasting state (21 vs 15 pg/ml, P = 0.036). At

fasting, GIP concentrations were significantly lower (37 vs 60 pg/ml, P = 0.008) by the

high DF diets compared to low fibre diets, whereas an interaction between DF and protein

was observed in the non-fasting state (P = 0.009) with significantly higher values of the

LOFLOP diet than the other three diets. The high DF diets induced a borderline significant

increase in fasting levels of total GLP-1 (P = 0.053), while diets with high protein content

tended to increase it in the non-fasting state (P = 0.055). The C-peptide concentrations in

non-fasting plasma was higher when the minipigs were fed the high DF diets (60 vs 32

pg/ml, P = 0.011) and a tendency (P = 0.076) for higher concentrations with the high

dietary levels of protein. There was no effect of diet on insulin concentrations, HOMA-IR

and HOMA-β in the fasting state, and only a tendency (P = 0.062) for a higher non-fasting

concentration of insulin with high dietary protein content. High DF content gave rise to

higher fasting concentrations of IL-12 (0.75 vs 0.59 ng/ml, P = 0.047) and IFN-γ (5.8 vs

4.3 ng/ml, P = 0.011) compared to low DF diets, whereas an interaction between DF and

protein was seen in fasting concentrations of IL-4 (P = 0.022) but with no differences

between the groups by pairwise comparisons. In the non-fasting state, only IFN-γ levels

were significantly higher with a high content of DF in the diet (2.3 vs 1.6 ng/ml, P = 0.001).

Clinical parameters in urine at fasting and slaughter

In the fasting state, there was a significant interaction between DF and protein level in

creatinine concentrations of the urine (P = 0.036) with LOFLOP having a significantly

higher concentration than the two high DF diets (Table 4). There was no effect of diet on

urinary glucose and protein concentrations and glucose:creatinine and protein:creatinine

ratios (P > 0.10) in the fasting state. At the non-fasting state, where urine was taken

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

42

directly from the bladder at euthanasia, a high dietary protein content resulted in a

significantly higher pH of the urine (5.6 vs 5.1, P < 0.001) and higher protein

concentrations (85 vs 46 mM, P = 0.032) compared to the low dietary protein diets. An

interaction of DF and protein was observed in protein:creatinine ratio (P = 0.048) with

higher values of HIFHIP diet than the three other diets.

Relative gene expression of liver, muscle and subcutaneous adipose tissue

Biopsies were taken from the liver, muscle and subcutaneous adipose tissues to study the

relative expression of genes involved in carbohydrate and lipid metabolism, inflammation

and transcription factors (Table 5). No effects were observed in gene expression in the liver

except for an increase in the expression of FBP1 with a high protein content of the diet

(1.27 vs 0.96, P = 0.043). In muscle tissue, there were interactions between DF and protein

in the expression of acetyl-Coenzyme A carboxylase alpha (ACACA, P = 0.020),

peroxisome proliferator-activated receptor gamma (PPARγ, P = 0.004) and tumor necrosis

factor (TNF, P = 0.026), where the LOFHIP induced significantly higher expression of

PPARγ than the other three diets in the post-hoc analysis. In sAT, high DF content reduced

the expression of fatty acid synthase (FASN) (0.70 vs 1.20, P = 0.035), whereas an

interaction between DF and protein was seen in the expression of adiponectin receptor 1

(ADIPOR1) (P = 0.048), LEPR (P = 0.031) and PPARγ (P = 0.018) without differences

between groups after adjustment for multiple comparisons.

DISCUSSION

The current study intended to investigate the possible ameliorating effects of DF and

protein on MetS biomarkers in an obese porcine model for childhood obesity(22)

. The daily

feed intake for 6-9 month old female minipigs is restricted to 300-400 g in order to prevent

obesity(30)

. Although the low protein diets in this study had slightly lower protein content

than recommended standard diet (11% vs 13%), the daily protein intake of these growing

minipigs was covered with the at libitum feeding pattern (104 g with the low protein diets

vs 39-52 g with the standard diet). Obesity was induced in the model by feeding a high-fat-

high-fructose diet (diet LOFILOP) and after 20 weeks the pigs were on average

approximately twice as heavy as Göttingen Minipigs fed according to manufacturers’

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

43

recommendations (31 vs 16 kg at 7 months of age)(30)

. The heavy fat accumulation created a

preclinical stage with increased fasting glucose, signs of insulin deregulation and mild signs

of liver inflammation (Table S2). Similar effects have been found in other studies with

Göttingen Minipigs(22, 31, 32)

.

In the current 8 weeks intervention period with diets varying in DF and protein, the

minipigs continued to gain weight although at a different rate depending on the dietary

composition. Thus, the pigs fed the diet high in DF and low in protein consumed less feed

and, consequently gained less weight than the pigs on the other diets. The lower feed intake

of the high DF low protein diet and the contrasting effects of DF and protein on weight gain

is most likely related to: firstly, a lower nutrient digestibility, modified nutrient absorption

and influence of DF on satiety(9, 10)

; and secondly, positive influence of high protein on

growth and modest influence on satiety. In earlier epidemiologically studies(33, 34)

, a positive

effect of high protein in particular of protein of animal origin on weight gain(34)

has also

been observed presumably reflecting that the protein quality plays a crucial role. It should

be noted that the low-DF-low-protein diet used during development of obesity(22)

and

continued to be used in one of the dietary treatment (LOFLOP) in the current intervention

study was specifically designed to provide a reduced protein content in order to redirect

energy from lean tissue accretion to adipose tissue storage in this way diminishing muscle

mass for glucose regulation(35)

. These aspects have to be taken into consideration when

evaluating the effects of the high protein diets on several biomarkers. For instance, the

concentrations of albumin and AST in plasma are most likely a reflection of changed

protein catabolism and turnover rates in the liver with a high dietary protein intake(36)

rather

than a marker for MetS(37)

. To support this view, relative organ weight(30)

, liver fat(35)

and

plasma inflammatory cytokines(22)

were found to be in the normal range.

While we can expect the digestibility to be lower with the high DF diets because of the

insoluble nature of DF making it resistant to microbial degradation(10, 38)

, there are also

indications of satiety induced by gut hormones with DF. GIP was significantly reduced by

high DF both at fasting and at non-fasting but with significantly higher levels during non-

fasting than fasting. GIP stimulates insulin secretion in a manner related to the absorption

of especially glucose and fat(39)

. The lower GIP level with high DF diets and higher GIP

level with the low DF low protein diet is without doubt a reflection of the influx of glucose

being higher for the latter than the former diet. The GLP-1 level also showed borderline

increase with the high DF diets at the fasting stage, indicating an effect of DF on satiety

after overnight fasting; the effect most likely being caused by the slower absorption of

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

44

nutrients(40)

and possibly also triggered by fasting levels of short chain fatty acids being

released from slowly fermented DF(41)

. GLP-1 increases rapidly after a meal and plays a

key role in the regulation of insulin secretion and sensitivity thereby reducing post-meal

glucose concentrations and improves beta-cell proliferation(42)

. Overall, although no

significant change was observed in insulin sensitivity with high DF at the fasting state,

GLP-1 secretion induced by a possible slower nutrient absorption could potentially be

protective against β-cell exhaustion. In contrast to GIP and GLP-1, the results on ghrelin

diverges from the general understanding that DF can suppress ghrelin concentrations and

thereby reduce the feeling of hunger(43)

as the non-fasting level ghrelin was higher in the

high DF diets without influencing either feed intake or weight gain. A previous study with

high DF diets also demonstrated that satiety feeling may be unrelated to the ghrelin

response and did not affect the following food intake(6)

. It has also been found that an AX-

enriched meal increases serum ghrelin levels in healthy human subjects with normal

glucose tolerance(44)

. However, given the ad libitum feeding pattern and expected delayed

nutrient absorption with high DF, increased ghrelin levels could also occur earlier in the

low DF groups, and as a result there was no overall influence on either feed intake or

weight gain.

We have previously shown that an AX-rich diet did not change postprandial glucose

responses of pigs but induced a lower postprandial peak in insulin in portal vein, hepatic

vein and mesentery artery compared to a low-DF western-style diet(45)

. In the present study

we did not see any effect of the high DF diets on either glucose or insulin, which is in line

with a study with human subjects with normal glucose tolerance where an AX-rich diet was

found not to influence glucose or insulin responses(44)

. Although we did not observe

significant changes in insulin concentrations, we found higher non-fasting levels of C-

peptide with the high DF diets, which may be an indicator of improved secretory function

of pancreatic beta-cells. However, the higher level of C-peptide occurred without neither

HOMA-IR nor HOMA-β being influenced by DF. Of note, in our comparisons between

species, it needs to be acknowledged that there are indications that C-peptide may have

different functions in minipigs and humans(32, 46)

.

During the progression of obesity in this juvenile model, we did not observe any

relationship between obesity development and IL-12 and IFN-γ(22)

. After 8 weeks of dietary

interventions, however, elevated concentrations of IL-12 in the fasting state and IFN-γ in

the fasting and non-fasting state were observed in pigs fed the high DF diets. These results

are in contrast to our expectations but generally in agreement with a recent study where it

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

45

was concluded that the diet high in DF from enzyme treated wheat bran did not affect low-

grade inflammation(47)

. The reason for lack of effect is most likely that the degree of

inflammation is much less at obesity than under a pathogen infection where DF have been

found to have immunomodulatory effects by reducing IL-12, IFN- γ and TNF-α

production(48)

. Moreover, MetS, especially the systemic inflammation, has been strongly

associated with general and central obesity(49)

, the absence of significant changes in obesity

markers such as obesity index, backfat thickness and liver fat in the current study

corresponded with the lack of effect on most inflammatory biomarkers except IFN-. The

rationale of measuring at fasting and non-fasting is that inflammatory cytokines have been

reported to change in response a meal(50)

.

The two high protein diets caused higher fasting glucose concentrations, a tendency for

higher urinary glucose concentrations, and a tendency for higher non-fasting insulin, GLP-1

and C-peptide concentrations but no differences in fasting insulin. The higher fasting

glucose level is presumably caused by gluconeogenic amino acids including branched-chain

amino acid (BCAA), as indicated by the higher concentration of glucose in the jugular than

the portal vein and higher urine pH at non-fasting. Although high BCAA content of whey

protein has shown insulionotropic effects in a short term study(18)

, the long-term effects are

contradictory(51)

and a positive association between high circulating BCAA and obesity was

found in our previous study(22)

, which may have detrimental effects on glucose and lipid

homeostasis of obese minipigs(52)

. The increased expression of liver fructose-

bisphosphatase 1 (FBP1), the rate-limiting enzyme in gluconeogenesis, on the high protein

diets is probably also related to a higher gluconeogenesis as also observed in young subjects

with newly diagnosed T2D(53)

. Other studies have shown that high-protein diets, despite

their beneficial effects on satiety, weight loss, and blood lipids, under certain conditions

may modulate amino acid metabolic signature and be a factor in IR and T2D

development(51, 54)

. Moreover, a human intervention study with the same type of diets did

not induce a higher GLP-1 response with high-protein diets(55)

, whereas this was the case

when whey protein was provided as a pre-meal in an acute study(19)

. The difference in

response to protein in the current and the 12 weeks human intervention study(55)

on GLP-1

compared to acute studies(19, 56)

could indicate that whey protein elicit responses by different

mechanisms when provided acutely compared to a chronic intake.

Neither whey protein nor DF influenced total cholesterol in plasma but whey protein

increased non-fasting LDL and the LDL:HDL ratio. A higher fasting NEFA concentration

was also observed when feeding the high protein diets. Combined these data indicate an

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

46

unfavourable progression towards dyslipidaemia when feeding the high protein diets. In a

recent 12 week intervention study with human subjects with abdominal obesity, it was

found that whey protein in combination with a low DF diet reduced total fasting cholesterol

whereas high DF had the opposite effect(16)

.

Obesity-driven insulin resistance in white adipose tissue, liver and skeletal muscle is

the primary cause of T2D and is linked to obesity-associated metabolic abnormalities such

as dyslipidaemia and inflammation(57)

. Therefore, we explored how DF and protein

interventions would affect the expression of key regulatory genes involved in carbohydrate

and lipid metabolism, inflammation and transcription factors in liver, muscle and

subcutaneous adipose tissues. We observed an increased relative gene expression of PPARγ

in muscle tissue in the LOFHIP group, which potentially can be linked to impaired insulin

sensitivity. PPARγ plays an important role in regulating insulin action in skeletal muscle(58)

and mRNA and protein expression of PPARγ has been found to be higher in muscle tissue

of humans with severe insulin resistance(58, 59)

. These data also corroborate the higher

fasting glucose concentrations and tendency for higher non-fasting insulin concentrations in

high protein diets. It appears that DF may attenuate the detrimental effect of protein as

increased gene expression of PPARγ was only found in diet with high protein and low DF

content.

The HIFLOP diet induced a significantly elevated expression of LEPR in the adipose

tissue, the signalling pathway through which leptin controls energy balance(60)

. This may

suggest improved responsiveness to leptin and potentially be associated with the reduced

feed intake and weight gain in this particular group. A lower LEPR mRNA abundance in

the subcutaneous adipose tissue has been associated with morbid obesity when compared to

lean human subjects and strongly correlated with insulin sensitivity(61)

. A previous study

also found that LEPR overexpression in adipose tissue of leptin receptor transgenic mice

could reduce body weight and fat deposition(62)

. As a central enzyme involved in lipid

biosynthesis, FASN gene expression in adipose tissue has been associated with visceral fat

accumulation and impaired insulin sensitivity(63)

. In our study, however, the relative

expression of FASN was suppressed significantly by the high DF content, which potentially

can explain the lower weight gain and improved non-fasting C-peptide concentrations in the

high DF groups.

Our study has several strengths but also some weaknesses. Firstly, we performed a

randomized long-term ad libitum intervention study with well characterised and controlled

diets. Secondly, our sample size based on the power calculation was sufficient to allow a

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

47

reliable detection of changes in metabolic biomarkers. Thirdly, we present very

comprehensive clinical parameters both at non-fasting and fasting. Fourthly, we use

innovative ingredients including wheay protein hydrolysate as protein source and enzyme-

reated wheat bran as DF source. The study, however, also has limitations that need to be

acknowledged. Firstly, we can not exclude that the growth of animals may have masked the

effects of DF and protein, thus further studies are needed to clarify whether DF and high

protein can synergistically affect MetS in an adult obese model with stable weight.

Secondly, a lean control group of pigs feeding on regular diet was not included as the

primary purpose of this study was to investigate the effects of DF and protein on obesity.

Thirdly, since whey protein and enzyme-treated wheat bran were the only protein and fiber

source used to increase dietary protein and DF levels, we can not conclude whether the

effects were induced by the specific sources or by the levels of protein and DF, and further

studies on other DF and protein sources are needed. Morever, in this paper, we have only

presented mRNA data and protein expression of the regulatory genes was not analyzed.

Since mRNA expression is not always correlated with protein expression as well as with the

activation or function of corresponding proteins, the gene expression findings need to be

interpreted with caution and warrant further studies.

In conclusion, the present study demonstrates that an eight-week dietary intervention

with DF and protein did not improve glucose and insulin response directly in an obese

minipig model. However, we demonstrated that a diet enriched with DF from enzyme-

treated wheat bran reduced weight gain and had a potential beneficial effect on beta-cell

secretory function but without effects on lipid biomarkers. Diets enriched in whey protein

hydrolysate tended to increase post-meal insulin levels and several markers related to lipid

and carbohydrate metabolism in an unfavourable way. In contrast to our hypothesis, a

combination of high or low DF and protein diets did not show a synergistic effect on insulin

sensitivity, postprandial lipemia, metabolic or inflammatory biomarkers associated with

MetS as it was also found in a recent human intervention study(55)

and concluded in a recent

review on the impact of DF consumption on insulin resistance and the prevention of

T2D(15)

.

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

48

Accepted manuscript

ACKNOWLEDGEMENT

The authors thank Winnie Østergaard, Lisbeth Mӓrcher, Stina Greis Handberg, Kasper

Vrangstrup Poulsen, Søren Krogh Jensen, Anette Møjbæk Pedersen and Elsebeth Lyng

Pedersen and staff at the animal facility for excellent technical assistance. We thank Leslie

Foldager for data consultation. Whey protein hydrolysate was kindly provided by Arla

Foods Ingredients Group P/S. Wheat bran was delivered by Lantmӓnnen and enzymatically

heated by DuPont Industrial Bioscience. The work was financially supported by Innovation

Fund Denmark (4105-00002B) and industrial partners involved in the MERITS (Metabolic

Changes by Carbohydrate and Protein Quality in the Development and Mitigation of

Metabolic Syndrome) project. Y.T.X. acknowledges scholarship from China Scholarship

Community.

FINANCIAL SUPPORT

This work was supported by grants from the Danish Dairy Research Foundation and the

Innovation Fund Denmark—MERITS (4105-00002B).

CONFLICT OF INTEREST

The authors declare no conflict of interest.

AUTHORSHIP

K.E.B.K. H.N.L and M.S.H conceived and designed the research; H.N.L. M.V.C. and

M.S.H performed the animal experiments; H.N.L. and Y.T.X analysed the data; P.K.T.

provided data consultation; Y.T.X. wrote the paper; All co-authors contributed to draft

review.

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

49

Accepted manuscript

REFERENCES

1. Fryar CD, Carroll MD & Ogden CL (2016) Prevalence of Overweight, Obesity, and

Severe Obesity Among Children and Adolescents Aged 2–19 Years: United States,

1963–1965 Through 2015–2016.:National Center for Health Statistics. Available from:

https://www.cdc.gov/nchs/data/hestat/obesity_child_15_16/obesity_child_15_16.htm.

2. Isomaa B, Almgren P, Tuomi T et al. (2001) Cardiovascular morbidity and mortality

associated with the metabolic syndrome. Diabetes Care 24, 683-689.

3. Balkau B, Charles MA, Drivsholm T et al. (2002) Frequency of the WHO metabolic

syndrome in European cohorts, and an alternative definition of an insulin resistance

syndrome. Diabetes Metab 28(5), 364-376.

4. Alberti KGMM, Zimmet P & Shaw J (2006) Metabolic syndrome—a new world-wide

definition. A Consensus Statement from the International Diabetes Federation. Diabet

Med 23, 469–480.

5. Huang PL (2009) A comprehensive definition for metabolic syndrome. Dis Model Mech

2, 231-237.

6. Hartvigsen ML, Gregersen S, Laerke HN et al. (2014) Effects of concentrated

arabinoxylan and beta-glucan compared with refined wheat and whole grain rye on

glucose and appetite in subjects with the metabolic syndrome: a randomized study. Eur

J Clin Nutr 68(1), 84-90.

7. Hu FB, van Dam RM & Liu S (2001) Diet and risk of Type II diabetes: the role of types

of fat and carbohydrate. Diabetologia 44, 805-817.

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

50

8. Wolever TMS, Tosh SM, Spruill SE et al. (2020) Increasing oat beta-glucan viscosity in

a breakfast meal slows gastric emptying and reduces glycemic and insulinemic

responses but has no effect on appetite, food intake, or plasma ghrelin and PYY

responses in healthy humans: a randomized, placebo-controlled, crossover trial. Am J

Clin Nutr 111(2), 319-328.

9. Yu K, Ke MY, Li WH et al. (2014) The impact of soluble dietary fibre on gastric

emptying, postprandial blood glucose and insulin in patients with type 2 diabetes. Asia

Pac J Clin Nutr 23(2), 210-218.

10.Nielsen TS, Laerke HN, Theil PK et al. (2014) Diets high in resistant starch and

arabinoxylan modulate digestion processes and SCFA pool size in the large intestine

and faecal microbial composition in pigs. Br J Nutr 112(11), 1837-1849.

11.Furness JB, Rivera LR, Cho HJ et al. (2013) The gut as a sensory organ. Nat Rev

Gastroenterol Hepatol 10(12), 729-740.

12.Weickert MO & Pfeiffer AFH (2008) Metabolic effects of dietary fiber consumption

and prevention of diabetes. J Nutr 138, 439-442.

13.Russell WR, Baka A, Bjorck I et al. (2016) Impact of Diet Composition on Blood

Glucose Regulation. Crit Rev Food Sci Nutr 56(4), 541-590.

14.Weickert MO, Roden M, Isken F et al. (2011) Effects of supplemented isoenergetic

diets differing in cereal fiber and protein content on insulin sensitivity in overweight

humans. Am J Clin Nutr 94(2), 459-471.

15.Weickert MO & Pfeiffer AFH (2018) Impact of dietary fiber consumption on insulin

resistance and the prevention of type 2 diabetes. J Nutr 148(1), 7-12.

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

51

16.Rakvaag E, Fuglsang-Nielsen R, Bach Knudsen KE et al. (2019) Whey protein

combined with low dietary fiber improves lipid profile in subjects with abdominal

obesity: A randomized, controlled trial. Nutrients 11(9).

17.Markova M, Pivovarova O, Hornemann S et al. (2017) Isocaloric diets high in animal or

plant protein reduce liver fat and inflammation in individuals with type 2 diabetes.

Gastroenterology 152(3):571-585.

18.Mikael Nilsson MS, Anders H Frid, Jens J Holst et al. (2004) Glycemia and insulinemia

in healthy subjects after lactoseequivalent meals of milk and other food proteins: the

role of plasma amino acids and incretins. Am J Clin Nutr 80, 1246–1253.

19.Bjornshave A, Holst JJ & Hermansen K (2018) Pre-meal effect of whey proteins on

metabolic parameters in subjects with and without type 2 diabetes: A randomized,

crossover trial. Nutrients 10(2).

20.Frid AH, Nilsson M, Holst JJ et al. (2005) Effect of whey on blood glucose and insulin

responses to composite breakfast and lunch meals in type 2 diabetic subjects. Am J Clin

Nutr 83, 69-75.

21. Nilsson M, Holst JJ & Björck IME (2007) Metabolic effects of amino acid mixtures

and whey protein in healthy subjects: studies using glucose-equivalent drinks. Am J Clin

Nutr 85, 996-1004.

22.Curtasu MV (2019) Obesity and metabolic syndrome in miniature pigs as models for

human disease – metabolic changes in response to ad libitum feeding of high-fat-high-

carbohydrate diets. PhD Thesis, Aahus University, Denmark.

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

52

23.Bellinger DA, Merricks EP & Nichols TC (2006) Swine models of type 2 diabetes

mellitus: insulin resistance, glucose tolerance, and cardiovascular complications. ILAR J

47(3), 243-258.

24.Hsu MC, Wang ME, Jiang YF et al. (2017) Long-term feeding of high-fat plus high-

fructose diet induces isolated impaired glucose tolerance and skeletal muscle insulin

resistance in miniature pigs. Diabetol Metab Syndr 9:81.

25.Bligh EG & Dyer WJ (1959) A rapid method of total lipid extraction and purification.

Can J Biochem 37(Number 8).

26.Nielsen DSG, Fredborg M, Andersen V et al. (2017) Administration of Protein Kinase

D1 Induces a Protective Effect on Lipopolysaccharide-Induced Intestinal Inflammation

in a Co-Culture Model of Intestinal Epithelial Caco-2 Cells and RAW264.7 Macrophage

Cells. Int J Inflam, 9273640.

27.Sebert SP, Lecannu G, Kozlowski F et al. (2005) Childhood obesity and insulin

resistance in a Yucatan mini-piglet model: putative roles of IGF-1 and muscle PPARs in

adipose tissue activity and development. Int J Obes (Lond) 29(3), 324-333.

28.Swindle MM, Makin A, Herron AJ et al. (2012) Swine as models in biomedical

research and toxicology testing. Vet Pathol 49(2), 344-356.

29.Levy JC, Matthews DR & Hermans MP (1998) Correct homeostasis model assessment

(HOMA) evaluation uses the computer program. Diabetes Care 21(12), 2191-2192.

30. Ellegaard Göttingen Minipigs. Available online: https://minipigsdk/ (accessed on 03

Jul. 2020).

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

53

31.Larsen MO, Rolin B, Wilken M et al. (2002) High-fat high-energy feeding impairs

fasting glucose and increases fasting insulin levels in the Göttingen Minipig. Ann NY

Acad Sci 967, 414–423.

32.Christoffersen B, Golozoubova V, Pacini G et al. (2013) The young göttingen minipig

as a model of childhood and adolescent obesity: Influence of diet and gender. Obesity

21(1), 149-158.

33. Hernandez-Alonso P, Salas-Salvado J, Ruiz-Canela M et al. (2016) High dietary

protein intake is associated with an increased body weight and total death risk. Clin

Nutr 35(2), 496-506.

34. Halkjaer J, Olsen A, Overvad K, et al. (2011) Intake of total, animal and plant protein

and subsequent changes in weight or waist circumference in European men and

women: the Diogenes project. Int J Obes (Lond) 35(8), 1104-1113.

35. Fisher KD, Scheffler TL, Kasten SC et al. (2013) Energy dense, protein restricted diet

increases adiposity and perturbs metabolism in young, genetically lean pigs. PLoS One

8(8), e72320.

36. Thalacker-Mercer AE & Campbell WW (2008) Dietary protein intake affects albumin

fractional synthesis rate in younger and older adults equally. Nutr Rev 66(2), 91-95.

37. Adilah ZN, Liew WPP, Redzwan SM et al. (2018) Effect of high protein diet and

probiotic Lactobacillus casei Shirota supplementation in aflatoxin B1-induced rats.

Biomed Res Int, 9568351.

38. Mudgil D & Barak S (2013) Composition, properties and health benefits of indigestible

carbohydrate polymers as dietary fiber: a review. Int J Biol Macromol 61, 1-6.

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

54

39. Baggio LL & Drucker DJ (2007) Biology of incretins: GLP-1 and GIP.

Gastroenterology 132(6), 2131-2157.

40. Warrilow A, Mellor D, McKune A et al. (2019) Dietary fat, fibre, satiation, and satiety-

a systematic review of acute studies. Eur J Clin Nutr 73(3), 333-344.

41. Müller M, Hernández MAG, Goossens GH, et al. (2019) Circulating but not faecal

short-chain fatty acids are related to insulin sensitivity, lipolysis and GLP-1

concentrations in humans. Scientific Reports 9(1), 12515.

42. Wang X, Liu H, Chen J et al. (2015) Multiple factors related to the secretion of

glucagon-like peptide-1. Int J Endocrinol 2015, 651757.

43. Papathanasopoulos A & Camilleri M (2010) Dietary fiber supplements: effects in

obesity and metabolic syndrome and relationship to gastrointestinal functions.

Gastroenterology 138(1), 65-72 e1-2.

44. Möhlig M, Koebnick C, Weickert MO et al. (2005) Arabinoxylan-enriched meal

increases serum ghrelin levels in healthy humans. Horm Metab Res 37(05), 303-308.

45. Ingerslev AK, Theil PK, Hedemann MS et al. (2014) Resistant starch and arabinoxylan

augment SCFA absorption, but affect postprandial glucose and insulin responses

differently. Br J Nutr 111(9), 1564-1576.

46. Chen C, Zeng Y, Xu J et al. (2016) Therapeutic effects of soluble dietary fiber

consumption on type 2 diabetes mellitus. Exp Ther Med 12(2), 1232-1242.

47. Rakvaag E, Fuglsang-Nielsen R, Bach Knudsen KE et al. (2019) The combination of

whey protein and dietary fiber does not alter low-grade inflammation or adipose tissue

gene expression in adults with abdominal obesity. Rev Diabet Stud 15, 83-93.

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

55

48. Bermudez-Brito M, Faas MM & de Vos P (2016) Modulation of Dendritic-Epithelial

Cell Responses against Sphingomonas Paucimobilis by Dietary Fibers. Sci Rep 6,

30277.

49. Chung ST, Hsia DS, Chacko SK et al. (2015) Increased gluconeogenesis in youth with

newly diagnosed type 2 diabetes. Diabetologia 58(3), 596-603.

50. Polakof S, Rémond D, Rambeau M et al. (2014) Postprandial metabolic events in mini-

pigs: new insights from a combined approach using plasma metabolomics, tissue gene

expression, and enzyme activity. Metabolomics 11(4), 964-979.

51. Rietman A, Schwarz J, Tome D et al. (2014) High dietary protein intake, reducing or

eliciting insulin resistance? Eur J Clin Nutr 68(9), 973-979.

52. Polakof S, Remond D, David J et al. (2018) Time-course changes in circulating

branched-chain amino acid levels and metabolism in obese Yucatan minipig. Nutrition

50, 66-73.

53. Chung ST, Hsia DS, Chacko SK et al. (2015) Increased gluconeogenesis in youth with

newly diagnosed type 2 diabetes. Diabetologia 58(3), 596-603.

54. Hession M, Rolland C, Kulkarni U et al. (2009) Systematic review of randomized

controlled trials of low-carbohydrate vs. low-fat/low-calorie diets in the management of

obesity and its comorbidities. Obes Rev 10(1), 36-50.

55. Nielsen RF (2019) Long-term effects of whey protein and dietary fiber from wheat on

markers of metabolic risk and bone health in subjects with abdominal obesity. PhD

Thesis, Aarhus University Hospital, Denmark.

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

56

56. Veldhorst MAB, Nieuwenhuizen AG, Hochstenbach-Waelen A et al. (2009) Dose-

dependent satiating effect of whey relative to casein or soy. Physiol Behav 96(4):675-

882.

57. Longo M, Zatterale F, Naderi J et al. (2019) Adipose tissue dysfunction as determinant

of obesity-associated metabolic complications. Int J Mol Sci 20(9).

58. Park KS, Ciaraldi TP, Abrams-Carter L et al. (1997) PPAR-y gene expression is

elevated in skeletal muscle of obese and type II diabetic subjects. Diabetes 46, 1230-

1234.

59. Loviscach M, Rehman N, Carter L et al. (2000) Distribution of peroxisome

proliferator-activated receptors (PPARs) in human skeletal muscle and adipose tissue:

relation to insulin action. Diabetologia 43, 304-311.

60. Friedman JM (2019) Leptin and the endocrine control of energy balance. Nat Metab

1(8), 754-764.

61. Seron K, Corset L, Vasseur F et al. (2006) Distinct impaired regulation of SOCS3 and

long and short isoforms of the leptin receptor in visceral and subcutaneous fat of lean

and obese women. Biochem Biophys Res Commun 348(4), 1232-1238.

62. Wang MY, Orci L, Ravazzola M et al. (2005) Fat storage in adipocytes requires

inactivation of leptin’s paracrine activity: Implications for treatment of human obesity.

PNAS 102, 18011–18016.

63. Berndt J, Kovacs P, Ruschke K et al. (2007) Fatty acid synthase gene expression in

human adipose tissue: association with obesity and type 2 diabetes. Diabetologia 50(7),

1472-1480.

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

57

Table 1. Ingredients and chemical composition of the experimental diets.

Items LOFLOP LOFHIP HIFLOP HIFHIP

Ingredients (g/kg as-fed basis)

Wheat starch 233 167 66 0

Whole grain wheat (roller

milled)

150 150 150 150

Wheat bran (finely milled) 125 125 125 125

Enzyme-treated wheat bran - - 200 200

Wheat gluten 65 65 32 32

Fish meal 20 20 20 20

Whey protein hydrolysate - 68 - 68

Fructose 200 200 200 200

Animal fat (Lard) 150 150 150 150

Vitamins and minerals 57 56 57 56

Chemical composition (g/kg DM)

DM (g/kg as-fed basis) 913 913 919 919

Ash 62 65 73 76

Crude protein (N × 6.25) 113 179 114 175

HCL fat 174 180 188 187

Available carbohydrates 577 522 456 387

Sugars

Fructose 225 223 224 221

Glucose 1 1 7 7

Sucrose 7 7 9 8

Starch 344 292 216 150

Dietary fibre1 100 106 191 205

NSP (soluble NSP) 69 (8) 75 (12) 136 (22) 136 (15)

AX (soluble AX) 44 (5) 46 (7) 86 (16) 85 (12)

RS2 2 1 1 1

AXOS 5 3 12 17

Fructans 6 8 11 9

Klason lignin 19 20 30 41

Gross energy (MJ/ kg DM) 20.7 21.5 21.3 21.7

Energy FAO/WHO (MJ/ kg DM)3

18.9 19.3 18.0 17.9

Protein, % 10.2 15.7 10.8 16.6

Fat, % 34.1 34.4 38.7 38.6

Carbohydrates, % 51.9 45.9 43.1 36.7

Dietary fibre, % 3.8 4.0 7.4 8.0

LOFLOP, low fibre low protein diet; LOFHIP, low fibre high protein diet; HIFLOP, high

fibre low protein diet; HIFHIP, high fibre high protein diet. NSP, total non-starch

polysaccharides; AX, arabinoxylan; RS, resistant starch; AXOS,

arabinoxylan‐oligosaccharides. 1Dietary fibre = NSP + fructans + RS + AXOS + Klason lignin.

2Determined by enzymatic resistant starch assay (AOAC method 2002.02).

3Calculated nutrient concentration and energy conversion factors (FAO) for protein (17

kJ/g), fat (37 kJ/g), carbohydrates (17 kJ/g) and total dietary fibre (8 kJ/g).

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

58

Accepted manuscript

Table 2. Concentrations of clinical parameters in fasting (jugular vein) and non-fasting

plasma (jugular and portal vein) in Göttingen Minipigs fed diets low or high in dietary

fibre and protein.

Item Diet

1

SEM P-value

LOFLOP LOFHIP HIFLOP HIFHIP Fibre Protein F×P2

Fasting, jugular vein

Glucose (mM) 5.7 6.8 5.8 6.9 0.42 0.771 0.008 0.917

Fructosamine (µM) 261 282 269 261 8.3 0.308 0.325 0.022

NEFA (μM) 166 197 156 284 36 0.275 0.015 0.162

Lactate (mM) 2.0 1.8 1.5 1.6 0.24 0.154 0.779 0.506

Triglycerides (mM) 0.53 0.50 0.62 0.65 0.07 0.072 0.990 0.660

LDL (mM) 1.6 1.9 1.4 1.7 0.25 0.361 0.265 0.888

HDL (mM) 1.7 1.7 1.7 1.6 0.20 0.900 0.359 0.450

LDL:HDL 1.0 1.2 0.84 1.1 0.16 0.320 0.096 0.773

Total cholesterol (mM) 4.2 4.6 4.3 3.9 0.71 0.519 0.925 0.444

Albumin (g/L) 42 45 39 43 1.5 0.097 0.018 0.779

AST (U/L) 30 23 25 23 2.3 0.305 0.043 0.301

ALT (U/L) 25 24 23 23 1.8 0.253 0.737 0.832

GGT (U/L) 61 70 63 64 6.5 0.728 0.357 0.373

Non-fasting, jugular vein

Glucose (mM) 5.6 6.3 6.3 7.8 0.60 0.062 0.065 0.496

Lactate (mM) 1.9 1.7 1.8 1.7 0.27 0.798 0.567 0.763

Triglycerides (mM) 0.64 0.57 0.62 1.2 0.19 0.098 0.119 0.076

LDL (mM) 1.3 1.9 1.1 1.5 0.28 0.177 0.054 0.608

HDL (mM) 1.9 1.8 1.8 1.7 0.18 0.356 0.246 0.854

LDL:HDL 0.66 0.98 0.63 0.79 0.10 0.255 0.015 0.382

Total cholesterol (mM) 4.1 4.2 3.8 3.7 0.61 0.341 0.932 0.798

Non-fasting, portal vein

Glucose (mM) 7.7 7.7 6.9 7.5 0.82 0.516 0.689 0.697

Lactate (mM) 4.4 4.0 4.6 4.5 0.90 0.656 0.777 0.872

Triglycerides (mM) 1.0 0.82 1.2 1.5 0.35 0.088 0.729 0.240

LDL (mM) 1.4 2.0 1.2 1.6 0.29 0.241 0.049 0.636

HDL (mM) 2.0 1.8 1.9 1.8 0.18 0.442 0.280 0.724

LDL:HDL 0.68 1.0 0.66 0.82 0.10 0.279 0.014 0.332

Total cholesterol (mM) 4.3 4.6 3.8 3.9 0.68 0.240 0.693 0.779

NEFA, non-esterified fatty acids; LDL, low density lipoprotein; HDL, high density

lipoprotein; AST, aspartate transaminase; ALT, alanine transaminase; GGT, gamma-

glutamyl transferase. 1Minipigs were regarded as the experimental units, n = 10 for low fibre low protein diet

(LOFLOP), n = 10 for low fibre high protein diet (LOFHIP), n = 12 for high fibre low

protein diet (HIFLOP) and n = 11 for high fibre high protein diet (HIFHIP). Data were

expressed as means ± standard error of means (SEM). 2F×P, interaction between fibre and protein level.

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

59

Table 3. Fasting and non-fasting concentrations of circulating hormones and inflammatory

cytokines in Göttingen Minipigs fed diets low or high in dietary fibre and protein.

Item Diet

1 SEM P-value

LOFLOP LOFHIP HIFLOP HIFHIP Fibre Protein F×P2

Fasting

Ghrelin (pg/ml) 14 18 16 38 5.6 0.055 0.008 0.073

GIP (pg/ml) 72 47 40 33 8.8 0.008 0.083 0.286

GLP-1 (pg/ml) 228 258 308 317 38 0.053 0.600 0.769

C-peptide(pg/ml) 28 33 35 38 5.8 0.307 0.468 0.928

Glucagon (pg/ml) 245 266 260 208 29 0.435 0.511 0.180

Insulin (pM) 45 55 57 55 16 0.688 0.815 0.663

HOMA-IR 1.9 2.6 2.4 2.4

0.84

0.891 0.639 0.654

HOMA-β3

62 (36-107) 50 (28-89) 56 (33-95) 50 (29-84)

-

0.858 0.553 0.878

IL-2 (ng/ml) 0.20 0.30 0.35 0.28 0.07 0.186 0.763 0.128

IL-4 (ng/ml) 0.36 1.43 1.32 0.79 0.35 0.605 0.473 0.028

IL-10 (ng/ml) 0.36 0.50 0.49 0.47 0.08 0.418 0.343 0.229

IL-12 (ng/ml) 0.59 0.60 0.71 0.78 0.11 0.047 0.575 0.691

IL-18 (ng/ml) 0.75 0.96 1.08 0.97 0.15 0.178 0.757 0.231

IFN- γ (ng/ml) 4.4 4.1 5.3 6.2 0.59 0.011 0.552 0.292

Non-fasting

Ghrelin (pg/ml) 14 17 20 23 3.0 0.036 0.343 0.997

GIP (pg/ml) 589

a 316

b 248

b 295

b 61 0.003 0.111 0.009

GLP-1 (pg/ml) 258 304 292 354 29.3 0.142 0.055 0.771

C-peptide (pg/ml) 23 41 50 69 11 0.011 0.076 0.945

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

60

Glucagon (pg/ml) 234 277 248 264 27 0.977 0.280 0.602

Insulin (pM) 71 76 55 85 9.0 0.668 0.062 0.171

IL-2 (ng/ml) 0.09 0.20 0.22 0.23 0.05 0.131 0.311 0.342

IL-4 (ng/ml) 0.12 0.55 0.58 0.57 0.21 0.248 0.329 0.302

IL-10 (ng/ml) 0.15 0.34 0.31 0.38 0.10 0.264 0.177 0.533

IL-12 (ng/ml) 0.32 0.30 0.35 0.36 0.04 0.278 0.930 0.656

IL-18 (ng/ml) 0.26 0.45 0.46 0.50 0.10 0.209 0.243 0.468

IFN- γ (ng/ml) 1.6 1.6 2.3 2.4 0.21 0.001 0.621 0.732

HOMA-IR, homeostatic model assessment for insulin resistance; HOMA-β, homeostatic

model assessment for beta-cell function; GIP, gastric inhibitory polypeptide; GLP-1,

glucagon-like polypeptide; IL, interleukin; IFN-γ, interferon γ. 1Mnipigs were regarded as the experimental units, n = 10 for low fibre low protein diet

(LOFLOP), n = 10 for low fibre high protein diet (LOFHIP), n = 12 for high fibre low

protein diet (HIFLOP) and n = 11 for high fibre high protein diet (HIFHIP). Data were

expressed as means ± standard error of means (SEM). 2F×P, interaction between fibre and protein level. Different superscript letters in a row are

presented for the significant interaction (P < 0.05) after adjustment for multiple

comparisons by the Tukey–Kramer post hoc test. 3Values are back-transformed after log transformation and are expressed as means with

lower and upper 95% confidence intervals.

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

61

Accepted manuscript

Table 4. Clinical parameter concentrations of urine collected at fasting and non-fasting

state of Göttingen Minipigs fed diets low or high in dietary fibre and protein.

Item Diet

1

SEM P-value

LOFLOP LOFHIP HIFLOP HIFHIP Fibre Protein F×P2

Fasting

Creatinine (mM) 17a 9.1

ab 7.1

b 8.4

b 2.4 0.018 0.160 0.036

Glucose (mM) 2.8 21 13 20 8.1 0.530 0.074 0.379

Protein (mM) 360 753 243 435 193 0.239 0.116 0.583

Glucose:creatinine3

0.13

(0.02-0.81)

0.80

(0.14-4.8)

0.69

(0.13-3.7)

1.0

(0.17-5.9)

- 0.163 0.118 0.279

Protein:creatinine4 14 (6.1-34) 35 (16-79) 24 (12-49) 38 (17-85) - 0.465 0.100 0.597

Non-fasting

pH 5.0 5.6 5.1 5.6 0.13 0.707 <0.001 0.896

Creatinine (mM) 10 10 10 7.6 1.6 0.387 0.327 0.416

Glucose (mM) 0.72 0.46 0.37 0.54 0.17 0.422 0.852 0.179

Protein (mM) 43 84 49 87 19 0.788 0.032 0.945

Glucose:creatinine5

0.05

(0.03-0.07)

0.04

(0.03-0.07)

0.03

(0.02-0.05)

0.06

(0.04-0.10)

- 0.815 0.258 0.137

Protein:creatinine6

3.9b

(2.0-7.7)

7.5b

(3.8-15)

4.0b

(2.1-7.6)

24a

(12-46)

- 0.040 <0.001 0.048

1Minipigs were regarded as the experimental units, n = 10 for low fibre low protein diet

(LOFLOP), n = 10 for low fibre high protein diet (LOFHIP), n = 12 for high fibre low

protein diet (HIFLOP) and n = 11 for high fibre high protein diet (HIFHIP). Data were

expressed as means ± standard error of means (SEM). 2F×P, interaction between fibre and protein level. Different superscript letters in a row are

presented for the significant interaction (P < 0.05) after adjustment for multiple

comparisons by the Tukey–Kramer post hoc test. 3-6

Values are back transformed after log transformation and are expressed as means (95%

confidence interval).

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

62

Accepted manuscript

Table 5. Gene expression of selected genes in liver, skeletal muscle and subcutaneous adipose tissue after dietary intervention in Gottingen

Minipigs

Item Diet1 P-values

Gene3 Pathway LOFLOP LOFHIP HIFLOP HIFHIP Fibre Protein F×P

2

Liver tissue

SLC2A5 Carbohydrate metabolism 1.00 (0.40-2.50) 1.10 (0.44-2.75) 0.94 (0.39-2.26) 0.32 (0.13-0.78) 0.071 0.126 0.092

SLC2A4 Carbohydrate metabolism 1.00 (0.53-1.88) 1.36 (0.72-2.55) 0.85 (0.47-1.55) 0.80 (0.44-1.47) 0.198 0.680 0.476

SLC2A8 Carbohydrate metabolism 1.00 (0.65-1.54) 0.89 (0.58-1.38) 1.21 (0.80-1.84) 0.81 (0.53-1.24) 0.752 0.11 0.381

HK1 Carbohydrate metabolism 1.00 (0.76-1.32) 1.15 (0.87-1.52) 1.08 (0.82-1.41) 0.92 (0.70-1.20) 0.492 0.891 0.179

FBP1 Carbohydrate metabolism 1.00 (0.60-1.67) 1.32 (0.79-2.21) 0.91 (0.55-1.50) 1.22 (0.74-2.02) 0.511 0.043 0.956

ACACA Fatty acid metabolism 1.00 (0.69-1.45) 1.18 (0.82-1.71) 1.12 (0.80-1.57) 1.04 (0.73-1.47) 0.981 0.853 0.502

ACLY Fatty acid metabolism 1.00 (0.74-1.35) 0.80 (0.59-1.08) 1.18 (0.89-1.55) 0.85 (0.64-1.13) 0.454 0.060 0.706

FASN Fatty acid metabolism 1.00 (0.47-2.11) 1.24 (0.59-2.61) 1.49 (0.76-2.95) 0.89 (0.43-1.80) 0.914 0.627 0.324

ADIPOR1 Fatty acid metabolism 1.00 (0.70-1.43) 1.03 (0.742-1.48) 1.10 (0.78-1.55) 1.14 (0.80-1.62) 0.412 0.770 0.964

PPARγ Lipid Transcription Factors 1.00 (0.59-1.70) 1.28 (0.75-2.19) 1.25 (0.75-2.09) 0.86 (0.51-1.44) 0.716 0.738 0.195

CRP Immune System/Inflammation 1.00 (0.18-5.64) 0.97 (0.17-5.50) 0.72 (0.13-3.89) 1.45 (0.26-7.95) 0.949 0.487 0.475

IL6 Immune System/Inflammation 1.00 (0.52-1.91) 1.42 (0.74-2.70) 1.44 (0.80-2.59) 0.96 (0.52-1.78) 0.990 0.872 0.248

TNF Immune System/Inflammation 1.00 (0.69-1.44) 1.12 (0.78-1.62) 1.23 (0.87-1.75) 1.31 (0.92-1.86) 0.319 0.633 0.873

CCL5 Immune System/Inflammation 1.00 (0.59-1.69) 1.51 (0.89-2.56) 1.30 (0.79-2.14) 1.53 (0.92-2.55) 0.481 0.167 0.526

Muscle tissue

SLC2A4 Carbohydrate metabolism 1.00 (0.66-1.52) 1.03 (0.68-1.57) 1.12 (0.76-1.66) 0.89 (0.60-1.33) 0.929 0.554 0.465

PFKM Carbohydrate metabolism 1.00 (0.42-2.39) 0.98 (0.41-2.35) 0.83 (0.35-1.97) 0.89 (0.37-2.11) 0.237 0.818 0.729

ACACA Fatty acid metabolism 1.00 (0.63-1.58) 1.91 (1.20-3.03) 1.72 (1.11-2.65) 1.35 (0.87-2.12) 0.561 0.350 0.020

FASN Fatty acid metabolism 1.00 (0.49-2.05) 3.27 (1.59-6.71) 1.52 (0.78-2.94) 1.41 (0.71-2.81) 0.556 0.133 0.067

ADIPOR1 Fatty acid metabolism 1.00 (0.63-1.59) 0.97 (0.61-1.55) 0.95 (0.60-1.50) 1.05 (0.66-1.66) 0.923 0.728 0.559

PPARγ Lipid Transcription Factors 1.00b (0.69-1.44) 2.39

a (1.66-3.45) 1.70

ab (1.22-2.38) 1.34

ab (0.93-1.93) 0.999 0.117 0.004

IL6 Immune System/Inflammation 1.00 (0.64-1.57) 1.10 (0.70-1.73) 1.55 (1.02-2.35) 0.92 (0.59-1.41) 0.504 0.234 0.114

TNF Immune System/Inflammation 1.00 (0.62-1.62) 1.81 (1.12-2.94) 1.83 (1.15-2.91) 1.52 (0.95-2.44) 0.194 0.300 0.026

https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at

63

CCL5 Immune System/Inflammation 1.00 (0.61-1.63) 1.54 (0.94-2.51) 1.44 (0.91-2.27) 1.49 (0.93-2.39) 0.428 0.310 0.358

Subcutaneous adipose tissue

SLC2A4 Carbohydrate metabolism 1.00 (0.60-1.68) 1.56 (0.93-2.63) 1.27 (0.78-2.07) 0.84 (0.51-1.39) 0.403 0.965 0.051

FASN Fatty acid metabolism 1.00 (0.50-2.00) 1.39 (0.69-2.79) 0.85 (0.44-1.66) 0.56 (0.28-1.11) 0.035 0.768 0.125

ADIPOR1 Fatty acid metabolism 1.00 (0.68-1.48) 1.42 (0.96-2.09) 1.30 (0.91-1.86) 0.84 (0.58-1.22) 0.529 0.718 0.048

LEPR Fatty acid metabolism 1.00b (0.34-2.94) 0.89

b (0.31-2.60) 2.29

a (0.81-6.48) 0.60

b (0.21-1.71) 0.457 0.007 0.031

LEP Fatty acid metabolism 1.00 (0.47-2.14) 0.90 (0.42-1.93) 0.99 (0.48-2.05) 0.98 (0.47-2.06) 0.887 0.837 0.846

ADIPOQ Fatty acid metabolism 1.00 (0.65-1.53) 1.29 (0.84-1.98) 1.44 (0.98-2.14) 0.90 (0.60-1.36) 0.960 0.532 0.092

CIDEC Fatty acid metabolism 1.00 (0.70-1.43) 1.11 (0.78-1.59) 1.20 (0.86-1.66) 0.75 (0.53-1.06) 0.576 0.266 0.115

PPARγ Lipid Transcription Factors 1.00 (0.52-1.93) 1.60 (0.83-3.11) 1.61 (0.85-3.04) 0.94 (0.49-1.80) 0.929 0.748 0.018

IL6 Immune System/Inflammation 1.00 (0.47-2.13) 0.84 (0.39-1.79) 0.46 (0.22-0.95) 0.58 (0.28-1.19) 0.136 0.923 0.594

TNF Immune System/Inflammation 1.00 (0.49-2.03) 1.42 (0.71-2.85) 1.52 (0.78-2.97) 1.21 (0.61-2.40) 0.565 0.876 0.198

CCL5 Immune System/Inflammation 1.00 (0.65-1.54) 1.27 (0.82-1.97) 1.27 (0.85-1.90) 1.22 (0.80-1.85) 0.626 0.653 0.488

SLC2A5, Solute Carrier Family 2 (Facilitated Glucose/Fructose Transporter) Member 5, Glut5; SLC2A4, Solute Carrier Family 2 (Facilitated

Glucose Transporter) Member 4, Glut4; SLC2A8, Solute Carrier Family 2 (Facilitated Glucose Transporter) Member 8, Glut8; HK1, Hexokinase 1;

FBP1, Fructose-Bisphosphatase 1; PFKM, Phosphofructokinase, Muscle; ACACA, Acetyl-Coenzyme A Carboxylase Alpha; ACLY, ATP-citrate

lyase; FASN, Fatty Acid Synthase; CCL5, C-C Motif Chemokine Ligand 5/ encodes RANTES; ADIPOR1, Adiponectin receptor 1; LEPR, Leptin

Receptor; LEP, Leptin; ADIPOQ, Adiponectin; CIDEC, Cell Death Inducing DFFA Like Effector C; PPARγ, Peroxisome proliferator activated

receptor gamma; CRP, C-reactive protein, pentraxin-related; IL6, Interleukin 6; TNF, Tumor Necrosis Factor. 1Minipigs were regarded as the experimental units, n = 10 for low fibre low protein diet (LOFLOP), n = 10 for low fibre high protein diet

(LOFHIP), n = 12 for high fibre low protein diet (HIFLOP) and n = 11 for high fibre high protein diet (HIFHIP). Data are reported relative to

LOFLOP and expressed as means (95% confidence interval). 2F×P, interaction between fibre and protein level. Different superscript letters in a row are presented for the significant interaction (P < 0.05) after

adjustment for multiple comparisons by the Tukey–Kramer post hoc test.

https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at

64

Accepted manuscript

https://ww

w.cam

bridge.org/core/terms . https://doi.org/10.1017/S0007114520003141

Dow

nloaded from https://w

ww

.cambridge.org/core . Aarhus U

niversitets Biblioteker , on 31 Aug 2020 at 14:23:37 , subject to the Cambridge Core term

s of use, available at

65

66

Table S1. RT-PCR gene expression assays

Gene Symbol Gene name Assay ID1

Carbohydrate metabolism

SLC2A5 Solute Carrier Family 2 (Facilitated Glucose/Fructose

Transporter) Member 5, Glut5 Ss03377332_u1

SLC2A4 Solute Carrier Family 2 (Facilitated Glucose Transporter)

Member 4, Glut4 Ss03373325_g1

SLC2A8 Solute Carrier Family 2 (Facilitated Glucose Transporter)

Member 8, Glut8 Ss03374161_m1

HK1 Hexokinase 1 Ss04323453_gH

FBP1 Fructose-Biphosphatase 1 Ss03393179_u1

PFKM Phosphofructokinase, Muscle Ss03380370_u1

Fatty acid metabolism/ De Novo Lipogenesis

ACACA Acetyl-Coenzyme A Carboxylase Alpha Ss03389962_m1

ACLY ATP-citrate lyase Ss03389566_m1

FASN Fatty Acid Synthase Ss03386194_u1

ADIPOR1 Adiponectin receptor 1 Ss03378803_u1

LEPR Leptin Receptor Ss03379257_u1

LEP Leptin Ss03392404_m1

ADIPOQ Adiponectin Ss03384375_u1

CIDEC Cell Death-Inducing DFFA Like Effector C Ss03389757_m1

Lipid Transcription Factors

PPARG Peroxisome proliferator-activated receptor gamma Ss03394829_m1

Immune System/Inflammation

CRP C-reactive protein, pentraxin-related Ss03390889_m1

IL6 Interleukin 6 Ss03384604_u1

TNF Tumour Necrosis Factor Ss03391318_g1

CCL5 C-C Motif Chemokine Ligand 5/ encodes RANTES Ss03648939_m1

Housekeeping genes

GAPDH Glyceraldehyde 3-phosphate dehydrogenase Ss03375629_u1

HPRT1 Hypoxanthine phosphoribosyltransferase 1 Ss03388274_m1

ACTB β-actin Ss03376563_uH 1 TaqMan Gene Expression Assay

67

Table S2. Baseline of body weight and clinical parameters in fasting plasma and urine at week 20.

Characteristics Means or

median

SD or interquartile

range

Body weight (kg) 31* 12

Fasting plasma

Glucose (mM) 6.1* 1.7

Fructosamine (µM) 258* 25

NEFA (μM) 179 94

Lactate (mM) 2.2 1.2

Triglycerides (mM) 0.57 0.20

LDL (mM) 1.4 0.70

HDL (mM) 1.9# 0.42

LDL:HDL 0.75 0.31

Total cholesterol (mM) 4.3 1.6

Albumin (g/L) 37 7.3

AST (U/L) 27 9.7

ALT (U/L) 23 5.4

GGT (U/L) 66 21

Ghrelin (pg/ml) 13 7.8

GIP (pg/ml) 59# 30

GLP-1 (pg/ml) 293 150

C-peptide(pg/ml) 32 19

Glucagon (pg/ml) 278* 103

Insulin (pM) 39* 20

HOMA-IR 1.6* 0.9

HOMA-β 55 (26-60)

IL-2 (ng/ml) 0.18 (0.12-0.28)

IL-4 (ng/ml) 0.33 (0.22-0.95)

IL-10 (ng/ml) 0.44 0.38

IL-12 (ng/ml) 0.67# 0.23

IL-18 (ng/ml) 0.82 0.54

IFN- γ (ng/ml) 4.5 1.9

Fasting urine

Creatinine (mM) 11 5.8

Glucose (mM) 2.8* (0.47-29)

Protein (mM) 141 (101-399)

Glucose:creatinine1 0.40 (0.04-3.5)

Protein:creatinine2 13 (7.9-37)

Minipigs were regarded as the experimental units, n = 43 for total groups. Data were expressed as

means with standard deviation (SD) or median with 25th-75th interquartile range. *means characteristic levels increased with time over the 20 weeks’ high fat high fructose feeding. #means characteristic levels decreased with time over the 20 weeks’ high fat high fructose feeding.

68

Length Chest Abdomen

0

50

100

150

cm

LOFLOPLOFHIPHIFLOP

HIFHIP

(a)

POI, L/cm BSA, m²0.0

0.5

1.0

1.5

LOFLOP

LOFHIP

HIFLOPHIFHIP

(b)

LOFLO

P

LOFH

IP

HIF

LOP

HIF

HIP

0

2

4

6

8

Ba

ckfa

t th

ick

ne

ss (

mm

)

(c)

Figure S1. Length, chest circumference and abdomen circumference (a), pig obesity index and body

surface area (b) and backfat thickness (c) of Göttingen Minipigs after 8-week fibre and protein

intervention. All data were expressed as means ± standard error of means. Pigs were regarded as the

experimental units, n = 10 for low fibre low protein diet (LOFLOP), n = 10 for low fibre high protein

diet (LOFHIP), n = 12 for high fibre low protein diet (HIFLOP) and n = 11 for high fibre high protein

diet (HIFHIP). Only significant P-values were presented in the figure.

69

LOFLO

P

LOFH

IP

HIF

LOP

HIF

HIP

0.0

0.5

1.0

1.5

2.0

2.5

Liv

er

we

igh

t (%

of

BW

)

(a)

LOFLO

P

LOFH

IP

HIF

LOP

HIF

HIP

0.0

0.1

0.2

0.3

0.4

He

art

wei

gh

t (%

of

BW

)

(b)

LOFLO

P

LOFH

IP

HIF

LOP

HIF

HIP

0.00

0.05

0.10

0.15

0.20

0.25

Kid

ne

y w

eig

ht

(% o

f B

W)

Protein: P = 0.008

(c)

LOFLO

P

LOFH

IP

HIF

LOP

HIF

HIP

0.0

0.1

0.2

0.3

Sm

all

in

test

ine

le

ng

th (

m/k

g B

W)

(d)

LOFLO

P

LOFH

IP

HIF

LOP

HIF

HIP

0.00

0.02

0.04

0.06

0.08

Co

lon

len

gth

(m

/kg

BW

)

(e)

LOFLO

P

LOFH

IP

HIF

LOP

HIF

HIP

0.0

0.5

1.0

1.5

Liv

er

fat

(%)

(f)

Figure S2. Relative weight (% of body weight, BW) of liver (a), heart (b) and kidney (c), small

intestine length (m/kg BW) (d) and colon length (m/kg BW) (e) and liver fat percentage (f) at

slaughter. Minipigs were regarded as the experimental units, n = 10 for low fibre low protein diet

(LOFLOP), n = 10 for low fibre high protein diet (LOFHIP), n = 12 for high fibre low protein diet

(HIFLOP) and n = 11 for high fibre high protein diet (HIFHIP). Values are least-squared means with

standard errors represented by vertical bars. Only significant P-values are presented in the figure.

70

100 200 300 400 500

-4

-2

0

2

4

6

De

lta

_g

lu (

mM

)

r = 0.494; P < 0.001

Starch intake (g/d)

Figure S3. Pearson correlation between delta glucose value (portal vein minus jugular vein) and

daily starch intake of Göttingen Minipigs.

71

6.2 Paper II

Effects of dietary fibre and protein content on intestinal fibre degradation, short-chain fatty

acid and microbiota composition in a high-fat fructose-rich diet induced obese Göttingen

Minipig model.

Yetong Xu, Mihai Victor Curtasu, Zachary Bendiks, Maria L. Marco, Natalja Nørskov, Knud Erik

Bach Knudsen, Mette Skou Hedemann and Helle Nygaard Lærke

Manuscript under review in Food & Function. 2020.

72

Effects of dietary fibre and protein content on intestinal fibre degradation, short-chain fatty 1

acid and microbiota composition in a high-fat fructose-rich diet induced obese Göttingen 2

Minipig model 3

Yetong Xu,a, Mihai Victor Curtasu,a Zachary Bendiks,b Maria L. Marco,b Natalja Nørskov,b Knud 4

Erik Bach Knudsen,b Mette Skou Hedemannb and Helle Nygaard Lærkeb 5

a Department of Animal Science, Aarhus University, DK-8830 Tjele, Denmark 6

b Department of Food Science and Technology, University of California, Davis, CA, USA; 7

zbendiks39@yahoo.com; mmarco@ucdavis.edu 8

Electronic supplementary information (ESI) available: Information of LC-MS/MS parameters for

plasma short-chain fatty acid (SCFA), Tables S1; dietary ingredients and chemical composition,

Table S2; pool size of SCFA in caecum and colon, Table S3; representative MRM chromatogram of

plasma SCFA standards, Fig. S1; faecal SCFA concentrations and distribution, Fig. S2; alpha and

belta- diversity of microbiota, Fig, S3-S4; significantly different microbiota altered by the diets, Fig

S5. Correspondence: Department of Animal Science, Aarhus University, DK-8830 Tjele, Denmark.

Email: yetongxu@anis.au.dk; Tel.: +4550641816

73

Abstract: Obesity-related metabolic syndrome has been linked with gut microbiome dysbiosis 9

while dietary fibre (DF) and protein can modify the gut microbial ecosystem and metabolism. After 10

20-weeks’ on a high-fat fructose-rich diet for development of obesity, forty-three 30-week old 11

Göttingen Minipigs (31 ± 4.0 kg body weight) were allocated to one of four diets with low or high 12

DF and protein contents in a two by two factorial design and digesta were collected from the 13

intestinal segments of minipigs after 8 weeks’ at libitum feeding. High DF content increased (P < 14

0.001) while high protein content decreased (P = 0.004) the content of non-starch polysaccharides 15

(NSP) in all intestinal segments. Arabinoxylan (AX) as proportion of NSP was higher with high DF 16

(P < 0.001) but reduced from distal small intestine until to mid colon (P < 0.001). High DF increased 17

the relative abundance of Blautia, Faecalibacterium and Peptococcus in caecum, mid colon and 18

faeces, reduced the intestinal concentrations of total short-chain fatty acids (SCFA) (P = 0.02) and 19

acetate (P = 0.01) but slightly increased butyrate pools in the large intestine compared to low DF. 20

High protein increased SCFA (P = 0.03) and propionate (P = 0.04) concentrations in the gut. High 21

DF induced a lower increase in BCFA concentration and proportion throughout the colon (P < 22

0.001). Butyrate concentrations in plasma from the jugular vein were increased with high DF diets 23

(P = 0.03), whereas propionate concentrations were increased (P < 0.001) and succinate decreased 24

(P = 0.001) with high protein diets compared with low protein. In conclusion, AX in the high DF 25

diets was continuously degraded until mid-colon, associated with enriched butyrate-producing 26

bacteria and slightly improved butyrate production, while protein fermentation was attenuated by 27

high DF and high protein did not show prebiotic effects in this obese minipig model. 28

Keywords: Dietary fibre; Protein; microbiota; SCFA; miniature pig 29

74

1. Introduction

Obesity has become one of the greatest public health challenges, and although not a disease in itself,

it greatly increases metabolic disorders and the risks for premature deaths globally.1 Intestinal

microbiota plays an important role in human health, and recent translational research showed that

unhealthy diets could cause intestinal microbiota dysbiosis, contributing to the development of

obesity and other metabolic disorders.2 It has been shown that dietary interventions intended to

modulate the gut microbiome composition and function, could be a therapeutic approach to alleviate

obesity and the related metabolic syndrome (MetS).3, 4 Importantly, some intestinal bacteria (like

Bifidobacterium and Lactobacillus) and microbiota-derived products such as short-chain fatty acids

(SCFA) have been demonstrated to regulate pathways related to metabolic control and systemic

inflammation.5

Dietary fibre (DF) is the fraction of carbohydrates and lignin that cannot be digested by host

endogenous digestive enzymes.6 Soluble DF and a variable part of insoluble DF are fermented into

SCFA by microbiota mainly harboring the large intestine. The SCFA produced in the large intestine

are mainly acetate, propionate and butyrate, and the ratio of SCFA is dependent on fermentation

substrate, microbiota composition as well as colonic transit time.7 In monogastric species,

carbohydrate fermentation mainly occurs in the caecum and proximal colon, resulting in low amounts

of fermentable substrate in the distal colon.8 Arabinoxylan (AX), a type of DF found in high

concentrations in wheat and rye bran, consist of a linear backbone of xylose units with arabinose

monomers substitution.9 From earlier studies with pigs it is known that fermentation of some fractions

of AX may be relatively slow in the colon10 and with stimulation of butyrate production.11 A dietary

intervention study demonstrated that an AX-enriched diet attenuated insulin secretion which was

thought to be related to increased butyrate or SCFA absorption.12 The underlying mechanisms could

be the influence of SCFA and butyrate on the release of satiety hormones and inhibition of lipolysis

in adipose tissue.13 An AX-rich diet has also shown beneficial effects on modulating gut health and

MetS by increasing the proportion of Bifidobacterium and reducing certain bacteria genera associated

with dysbiotic intestinal communities.14 Consistent with the increased DF fermentation, the microbial

profile in the distal large intestine was altered by diets rich in AX, resulting in promotion of butyrate-

producing bacteria (such as Faecalibacterium prausnitzii).15, 16 Moreover, using cell wall degrading

enzymes during bread making to produce arabinoxylan oligosaccharides (AXOS) resulted in higher

faecal butyrate concentrations and proportions in healthy human subjects.17

75

Fermentation of protein takes place in the more distal part of the colon particularly when

carbohydrate substrates are scarce.18 Whey protein is known as an insulinotropic substance and could

lower gastric emptying in T2D subjects,19 while the role of whey protein on gut microbiota and SCFA

production is poorly studied. A study showed that consumption of whey protein could protect against

colitis by stimulating intestinal mucin secretion and faecal lactobacilli and bifidobacteria growth in a

high fat diet-fed mice model.20 Another study found that intake of whey protein had no influence on

gut microbiota composition in mice fed a high fat diet.21 Moreover, an in vitro study demonstrated

that whey protein supplementation had a potent prebiotic effect and could stimulate SCFA production

in healthy and obese donors, which may contribute to improved intestinal health and reducing

obesity.22 However, a recent review concluded that high protein consumption especially with low

carbohydrate intake, was associated with intestinal inflammation due to the altered fermentation

environment such as increased branched-chain fatty acids (BCFA), ammonia, hydrogen sulfide and

indoles.8 Supplementation with DF, may exert a protective effect against the effect of proteolytic

metabolites by providing more substrate and shifting microbial species abundance.8

We have previously addressed the impacts of DF and whey protein on metabolic biomarkers

related to MetS.23 In the present study, focus is on the effect of DF and protein as well as their

interaction on gut DF structural modification, microbiota and SCFA profiles. We hypothesize that

dietary DF and protein level can influence the intestinal DF structure, selectively stimulate beneficial

bacteria growth, enhance SCFA production, especially butyrate. Göttingen Minipigs were used in this

study as an adolescent obese model for human individuals due to similar physiology and nutritional

requirements.24

2. Materials and methods

Diets, animals and experimental design have been described in detail in our previous paper presenting

the effects of DF and protein on metabolic biomarkers related to MetS.23

2.1. Diets, animals and experimental design

The animal experiment was performed under the Danish laws and regulations regarding humane care

and use of animals in research (The Danish Ministry of Justice, Act on Animal Experiments no 474

of May 15, 2014, as stipulated in the executive order no. 12 or January 07 2016). Animal handling

and experimental procedures were done according to license obtained from the Danish Animal

Experimentation Inspectorate, Ministry of Food, Agriculture, and Fisheries.

76

Four experimental diets were formulated: a diet low in both DF and protein (LOFLOP), a diet

low in DF and high in protein (LOFHIP; 6.8% whey protein hydrolysate added), a diet high in DF

and low in protein (HIFLOP; 20% enzyme-treated wheat bran added) and a diet high in both DF and

protein (HIFHIP; 6.8% whey protein hydrolysate and 20% enzyme-treated wheat bran added).

Ingredients and nutrient composition of the experimental diets can be found in Table S1 and have

been described elsewhere.23 The high DF diets were formulated to contain 20% DF in the form of

NSP, AX, AXOS, fructans and Klason lignin derived from whole grain wheat and enzyme-treated

wheat bran. The high protein diets were formulated to contain 18% protein from whole grain wheat,

wheat gluten, fish meal and whey protein hydrolysate. All of the diets had a high-fat (15% animal

fat) and high-fructose content (20% fructose) and were finely ground. The wheat bran was delivered

by Lantmӓnnen Cerealia AB (Malmø, Sweden) and enzymatically treated with cell wall-degrading

enzymes (xylanase, glucanase, cellulose) by DuPont Industrial Bioscience Aps (Brabrand, Denmark)

as previously described.16 Whey protein hydrolysate (Lacprodan® HYDRO.REBUILD) was provided

by Arla Foods Ingredients Group P/S (Viby J, Denmark). Samples of the experimental diets were

collected after mixing in the feed production unit at Aarhus University and stored at -20 ºC.

A total of forty-three female Göttingen Minipigs (Ellegaard Göttingen Minipigs, Dalmose,

Denmark) at 8 weeks of age were received at the facilities at Aarhus University in 4 separate blocks

over 8 months. The animal handling was described in detail previously,23 briefly, after a one-week

acclimatization time and one-week diet transition time, all the pigs were fed LOFLOP ad libitum for

20 weeks to induce obesity. At 30 weeks of age (body weight 31.3 ± 4.0 kg), the minipigs were

transferred to one of the four experimental diets over 3 days to gradually reach 100% of the

experimental diet on day 4. After that, all the pigs were fed ad libitum for 8 weeks, water was provided

ad libitum from drinking nipples, and toys were provided to satisfy the minipigs’ rooting behavior.

2.2. Sample collection

The anesthesia and euthanasia procedures were performed to collect samples as described before.23,

25 At the end of 8 weeks of intervention, all pigs were fasted overnight and fresh faeces were collected

for SCFA and microbiota analysis in connection with anesthesia.25 During a period of 3 days in the

9th week of intervention, all the pigs were anaesthesised in the non-fasting state, a blood sample was

taken from the jugular vein, the abdominal cavity opened by a midline incision for sampling of blood

from the portal vein and the pigs euthanized immediately after. The blood samples were collected

into LiHep tubes, centrifuged for 12 min at 3300 rpm at 4°C and immediately frozen at -80°C for

77

SCFA analysis. The entire gastrointestinal tract was ligated at the esophagus and rectum, removed

from the carcass and divided into different intestinal segments: The last third of the small intestine

was termed SI3; The colon was divided into three parts of equal length (proximal, mid and distal)

termed Co1, Co2, and Co3, respectively; SI3, Co1, Co2 and Co3 were tied off and the weight of fresh

digesta and pH of gut contents were recorded. Digesta were collected and kept in a -20°C freezer for

dry matter and carbohydrate analysis. Samples of digesta (1-3 g) from SI3, caecum, Co1, Co2 and

Co3 were collected for SCFA analysis and digesta from SI3 and Co2 were obtained and kept at -80°C

for microbiota analysis.

2.3. Analytical methods

Starch and non-starch polysaccharides (NSP) were measured as described by Bach Knudsen26 except

they were hydrolyzed with 2 M H2SO4 for 1 h instead of 1 M for 2 h during NSP analysis. The soluble

and insoluble NSP of diets and SI3 digesta were acid hydrolyzed and determined as alditol acetates

with derivatization by gas-liquid chromatography. Non-digestible carbohydrate (NDC) in diets and

SI3 digesta was determined by direct acid hydrolysis without starch removal and alcohol

precipitation. Total AX (TAX) was calculated as the sum of arabinose and xylose residues from the

NDC analysis, and soluble and insoluble high-molecular weight AX (SAX and IAX, respectively)

was calculated from the NSP analysis. AXOS were calculated as the difference of AX between NDC

and NSP fractions.

Digesta and faecal samples were subjected to an acid–base treatment followed by diethyl ether

extraction and derivatization to analyse SCFA concentrations as previously described27, and were

determined by GC (HP-6890 Series Gas Chromatograph; Hewlett Packard) using an HP-5 column

(30 m × 0.32 mm × 0.25 µm).

Plasma SCFA concentrations were determined by LC-MS/MS according to Han et al. (2015)28

with modifications. Ten µL of 75 % methanol containing stable isotope internal standards (13C2 acetic

acid, 13C1 propionic acid, 13C2 butyric acid, 13C2 succinic acid and 13C3 valeric acid) were mixed with

10 µL of plasma samples, 10 µL of 200 mM 3-nitrophenylhydrazine hydrochloride (3NPH) and 10

µL of 120 mM N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride (EDC). The mixture

was shaken for 45 min at room temperature and 10 µL of 200 mM of Quinic acid was added to quench

derivatization. The mixture was shaken for 15 min at room temperature and 950 µL of 10 % methanol

was added and mixed, centrifuged for 20 min at 4 ˚C 29,700 × g and 700 µL of supernatant was

transferred into HPLC vail. Samples were analysed on microLC 200 from Eksigent (Framingham,

78

MA, USA) using C18 column (10 cm × 1 mm, 1.2 µm particle size) from Waters (UK). The

chromatographic conditions started at 5 % solvent B (acetonitrile), kept for 1 min, the gradient

increased for 4 min to 27 % solvent B, then increased for 5 min to 30 % solvent B and to 37 % solvent

B during 5 min. Solvent A was pure water. Total chromatographic run was 15 min with flow of 50

µL/min, column oven sat to 30 ˚C and injection volume of 5 µL. MicroLC was interfaced to QTrap

5500 from ABSciex (Framingham, MA, USA) using electro spray ionization (ESI) in negative mode.

The flow injection analysis (FIA) was performed to optimize the turbo V source of the instrument,

where curtain gas was set to 30 psig, nebulizer gas (Gas1) 50 psig, heater gas (Gas2) 50 psig,

temperature was 500 ˚C, ionization spray operated at -4500 eV and collision Gas was set to High.

Deprotonated molecules were detected in Multiple Reaction Monitoring (MRM) mode. The

compound-dependent parameters were optimized by syringe infusion of pure standards and shown in

Table S1. Entrance potential was set to -10 volts and time of MRM scan to 30 msec for all the

compounds. The data analysis was performed in Analyst software 1.6.1 from AB Sciex (Framingham,

MA, USA). The representative chromatogram is presented in Fig. S1.

Bacterial genomic DNA was extracted from caecum, Co2 and faecal samples using a QIAamp

Fast DNA Stoll Mini Kit (Qiagen, Hilden, Germany).14 The samples were homogenized prior to DNA

extractions and purification according to the manufacturer’s instructions. DNA concentrations were

measured by using NanoDrop spectrophotometer (Thermo Scientific, Wilmington, DE) and diluted

to 20 ng/µL. PCR was used to amplify the 16S rRNA V4 regions using the barcoded F515 forward

primer and R806, and Equal volumes of DNA extractions from each sample were combined.29, 30 PCR

products were purified and the pooled amplicons was sequenced using an Illumina MiSeq (PE250)

platform (Illumina Inc., San Diego, CA) at the UC Davis Genome Center (http://dnatech.

genomecenter.ucdavis.edu/), University of California, Davis, California, USA. DNA sequence

analysis was performed using the pipeline Quantitative Insights Into Microbial Ecology (QIIME)

software. The alignment of the paired-end sequencing reads was not done, and R1 and R2 files were

concatenated and the reads were treated as single-ended. After extracting barcodes from the reads,

demultiplexing was performed. The Greengenes 13_8 99% operational taxonomic unit (OTU)

database was used as the reference database during open-reference OTU picking.31 To generate alpha

diversity rarefaction curves, the Chao1 diversity index, Faith’s Phylogenetic Diversity Whole Tree

(faith_pd),32 and observed species were determined by increasing numbers of randomly samples

sequences per sample. The curves became asymptotic at 25322 in caecum, 31787 in colon and 25806

in faecal samples. These numbers of randomly sampled DNA sequences were used to rarefy samples

79

prior to calculation of the of weighted UniFrac distances33 in QIIME. The beta diversity between

treatment groups was visualized using principal coordinates analysis (PCoA) on weighted UniFrac

distances.34

2.4. Calculations

SCFA concentrations were calculated as the sum of formic acid, acetic acid, propionic acid, butyric

acid, isobutyric acid, isovaleric acid and valeric acid concentrations, and BCFA as the sum of the

isobutyric acid, isovaleric acid and isocaproic acid. The pool size of SCFA and BCFA in caecal and

colonic segments were determined by multiplying the concentrations of SCFA by the amount of wet

digesta wherein the SCFA pool size of entire colon was the sum of pool in Co1, Co2 and Co3. The

SCFA distribution (%) in each intestinal segment was calculated as the concentration of the individual

SCFA divided by the total SCFA concentration in the wet digesta and multiplied by 100.

2.5. Statistical analysis

The pig was regarded as the experimental unit. Intestinal carbohydrate composition, SCFA

concentration, net amount, pH values and SCFA distribution of gut content in different segments

were analysed by using the MIXED procedure of SAS (SAS Institute, Inc.).

Yijklm = μ + αi + βj + (αβ)ij + γk + (αγ)ik + (βγ)jk + (αβγ)ijk + Ul + Vm + εijklm, (1)

where Yijklm is the dependent variable; μ is the overall mean; αi, βj and γk are the fixed effects of

dietary fibre levels (i = low or high fibre), protein levels (j = low or high protein) and intestinal

segments (k = SI3, Caecum, Co1, Co2 or Co3), respectively; (αβ)ij is the interaction between fibre

and protein; (αγ)ik is the interaction between fibre and segment; (βγ)jk is the interaction between

protein and segment; and (αβγ)ijk is the interaction for all fixed effects; Ul is the random effect of

block (l = 1, 2, 3 or 4); Vm is the random component related to the pig (m = 1, 2,…, 43). Pig was

included as a random component to account for repeated measurements within pig. The covariance

structure of repeated measurements was modelled using autoregressive type and εijklm is the residual

error. The random effect and residuals were assumed to be independent and normally distributed with

zero expectation.

Arabinoxylan distribution and arabinose:xylose (A:X) ratio of SI3 digesta, pool size of SCFA in

caecum and entire colon, SCFA profile of faeces, SCFA concentrations in plasma, predominant phyla

abundance in caecum, mid-colon and faeces were analysed by a MIXED statistical model without

segment factor where γk, (αγ)ik, (βγ)jk, (αβγ)ijk and Vm components were deleted and no repeated

80

measurement was conducted. If an interaction of fibre with protein was detected, we conducted a

pairwise comparison of groups adjusted by multiple comparisons of Tukey–Kramer post hoc test.

For AX distribution and A:X ratios of SI3 digesta, neither protein effect nor interaction among

fibre and protein was detected, hence, protein effect βj and interaction among fibre and protein (αβ)ij

were further omitted in the MIXED model and only the fixed effect of fibre levels αi (i = low or high

fibre) was analysed and presented. To compare the AX structure change between diet and SI3 digesta

in pigs fed with low and high fibre diets, the difference in AX distribution and A:X ratios of the SI3

digesta of each individual pig and the matching diet was calculated and analysed by the mixed model:

Yij = μ + αi + Uj + εij (2)

where Yij is the dependent variable, μ is the overall mean; αi is the fixed effect of fibre levels (i =

low or high fibre); Uj is the random effect of block (j = 1, 2, 3 or 4); εij is the residual error. Values

were presented as least square means with standard error of the mean (SEM). Levels of significance

was reported as being significant when P < 0.05.

One digesta sample from mid colon in LOFLOP group (n = 9) was omitted due to low-quality

reads. Kruskal-Wallis test (pairwise) was used to determine a difference in alpha diversity at a depth

of 25322 sequences in caecum, 31787 sequences in Co2 and 25806 sequences in faecal samples using

Faith’s phylogenetic diversity as index. For beta diversity analysis, the PERMANOVA pseudo F

statistic was used to quantify how much variation could be attributed to diet. In both cases, P < 0.05

was considered significant. Linear discriminant analysis (LDA) effect size (LEfSe) analysis was

performed with P < 0.05 and LDA > 2.0 using Galaxy (https://huttenhower.sph.harvard.edu/galaxy/).

RStudio (R, Version 1.1.453) was used for correlation analysis between taxa and metabolites. The

taxa data were normalized beforehand using rarefaction to account for uneven read numbers. Genera

were required to have at least 1% abundance after rarefaction. Pearson’s test with false discovery rate

P-value correction was used to determine significance.

3. Results

3.1. Carbohydrate composition in diets and intestinal segments

The four experimental diets were formulated to provide equal amounts of fat and energy with different

amounts of DF (10% vs. 20%) and protein (11% vs. 18%). The starch content was lowest in the

HIFHIP diet, because the addition of enzyme-treated wheat bran and whey protein hydrolysate is at

the expense of wheat starch. In high DF diets (HIF) compared with low DF diets (LOF), AXOS (15

81

vs. 3.5 g/kg) and SAX (13 vs. 6.1 g/kg) were 2~4 times higher by the addition of enzyme-treated

wheat bran (Table S1).

IAX was the main contributor to total AX of both diet and SI3 digesta in the LOF or HIF groups

accounting for >70% (Fig. 1A). Compared with LOF, HIF had a higher proportion of AXOS of total

AX both in diet (15% vs. 7.0%) and SI3 digesta (11% vs. 4.7%, P < 0.001), and a lower proportion

of IAX in diet (72% vs. 80%) and SI3 digesta (71% vs. 75%, P = 0.03), while no significant change

was seen for SAX. Compared to the diet, digesta in SI3 contained an insignificantly but numerically

lower proportion of TAX in the form of AXOS (LOF: 4.7% vs. 7.0 %, P = 0.12 and HIF: 11% vs. 15

%, P = 0.08), and a higher proportion of SAX (LOF: 20 % vs. 13 %, P < 0.001 and HIF: 18 % vs.

14%, P < 0.001) but only a lower proportion of IAX for LOF (75% vs. 80%, P < 0.001). In the HIF

groups, the A:X ratio of AXOS in diet (0.40 vs. 0.80) and SI3 digesta (0.38 vs. 0.64, P < 0.001) was

lower than LOF, while the A:X ratio of IAX was higher (diet: 0.70 vs. 0.63; SI3 digesta: 0.71 vs.

0.61, P < 0.001). The A:X ratio of SAX in SI3 did not differ significantly between DF levels (Fig.

1B), but for HIF, the A:X ratio of SAX was significantly higher in SI3 than in the diet (0.59 vs. 0.50,

P = 0.03) without changes in A:X ratios of AXOS and IAX. Compared with the diet, the A:X ratio

of AXOS was lower in digesta from SI3 of the LOF fed pigs (0.64 vs. 0.80, P = 0.01), whereas no

change in A:X ratios was seen for SAX and IAX.

Table 1 shows starch content and NSP content and structure in intestinal segments from SI3 to

Co3. The starch content was not significantly affected by dietary treatments (P > 0.05) and was

gradually reduced from caecum to Co2 (P < 0.001) without further change in Co3 (P = 0.68). The

total NSP and AX contents were higher with the high DF diets than with low DF (P < 0.001), and

slightly lower with high protein than low protein diets (P = 0.004). Total NSP content was reduced

from caecum to Co1 (P < 0.001), while the AX content was gradually reduced from SI3 to Co1 (P <

0.001). Both total NSP and AX contents did not change along the colon (P > 0.10). The contribution

of AX to total NSP declined until Co3 (P < 0.001) and was higher in the high DF groups than low

DF groups (P < 0.001). For A:X ratios, there was a significant interaction between DF and segment,

showing higher ratios in SI3, caecum and Co1 with high DF than with low DF diets (P < 0.001).

Moreover, there was a gradual increase in A:X ratios from SI3 to Co2.

3.2. SCFA profile, pH and net amount of gut content

With the high DF diets, there were small but significantly lower concentrations of SCFA (P = 0.02)

and acetate (P = 0.01) compared with the low DF diets, whereas giving high protein diets to the pigs

82

resulted in higher SCFA (P = 0.03) and propionate concentrations (P = 0.04) and a tendency for

higher acetate (P = 0.06) concentration (Fig. 2). The concentration of SCFA, acetate and propionate

was low in SI3, highest in caecum and decreased subsequently until Co2 (P < 0.001). For butyrate,

higher concentrations were detected in caecum and Co1 (P < 0.001) than in the other segments. There

was a significant interaction between DF and segment in BCFA concentration (P < 0.001), as the

BCFA concentration increased significantly from Co1 to Co3 with the low DF diets (P < 0.001),

whereas it was lower and almost constant with the high DF diets. There was also an interaction

between protein and segment (P = 0.04); for the high protein diets BCFA concentrations gradually

increased from SI3, caecum to Co1 (P < 0.05) with no significant subsequent change in the more

distal parts of the colon. For the low protein diets, BCFA concentrations did not change from caecum

to Co2 but was higher in Co3 compared with the other segments (P < 0.01).

The amount of wet digesta differed between intestinal segments and was dependent on both DF

and dietary protein levels. Co1 had the highest amount of gut content, and HIFHIP lead to

significantly more digesta than the other three diets (P = 0.006, Fig. 3A). Intestinal pH dropped from

6.8 in SI3 to 6.3 in the caecum (P < 0.001) and then gradually increased in Co1 and Co2 (P < 0.01).

A tendency for lower pH values with high DF diets compared with low DF was observed in Co2 (P

= 0.09). In Co3, the pH decreased again (P = 0.05), and high DF diets resulted in a significantly lower

pH than the low DF diets (P = 0.02, Fig. 3B).

In all intestinal segments, acetate was the dominating contributor to SCFA (> 57%), and its

contribution was highest in caecum and declined in the colon (P < 0.001, Table 2). The proportion of

propionate increased steeply from SI3 to caecum (P < 0.001), but was unchanged throughout the

remainder of the large intestine. The low butyrate concentrations with the high DF diets in the distal

small intestine resulted also in lower proportions of butyrate than with the low DF diets (P < 0.001).

The modest change in proportions of butyrate between segments of the large intestine depended on

fibre levels (P < 0.001) where, compared with caecum, the proportion of butyrate was higher in Co2

and Co3 with the high DF diets (P < 0.05) but only significantly higher in Co3 with the low DF diets

(P < 0.001). The proportion of butyrate was slightly increased from caecum to Co3 with a high protein

intake, whereas it was almost constant throughout the large intestine with the low protein diets,

leading to an interaction between protein and segment (P = 0.05). The contribution of BCFA to SCFA

in Co3 was significantly lower for pigs fed the high DF (P = 0.002) compared to the low DF diets.

The pool of butyrate in caecum was higher in pigs with high DF than with low DF diets (P = 0.05,

Table S2). In the colon, there were significant interactions between DF and protein in the pool size

83

of SCFA (P = 0.02) and acetate (P = 0.02), because they were significantly higher for HIFHIP than

the other diets. In addition, the colonic pools of propionate (P = 0.007) and butyrate (P = 0.04) were

significantly higher with the high DF than the low DF diets. The pool size of BCFA was on the other

hand higher in pigs with high compared to low protein intake (P = 0.03).

The SCFA profiles of faeces showed that the high DF diets resulted in a higher concentration (P

= 0.001) and proportion (P < 0.001) of propionate but lead to a lower proportion of BCFA (P < 0.001)

compared with the low DF diets (Fig. S1).

3.3. Microbiota composition

No difference in alpha diversity (faith_pd) was observed either in the caecum or Co2 among the four

diets, whilst the high DF groups showed significantly increased alpha diversity in bacterial

composition in faecal samples compared with low DF (P = 0.02) (Fig. S2). The high DF diets resulted

in significant changes in the beta-diversity of the caecal, colonic and faecal microbiota as shown by

PCoA of the weighted UniFrac distance metric (P = 0.004, P = 0.001 and P = 0.001, respectively)

(Fig. S3).

Firmicutes and Bacteroidetes were the two predominant phyla in the caecum, Co2, and faeces.

Their proportions in the caecum and Co2 were not significantly changed by dietary treatments (Fig.

4). In faecal samples, Bacteroidetes was more abundant (19.3% vs. 8.3%, P < 0.001) while Firmicutes

were less abundant (72% vs. 85%, P < 0.001) in the high DF groups compared with the low DF

groups, resulting in a higher Bacteroidetes:Firmicutes ratio (0.29 vs. 0.11, P < 0.001) in high DF

groups.

A LEfSe analysis was also performed to determine enrichment of taxa in the four dietary groups,

but no significantly enriched taxa were detected in this comparison. However, when looking only at

the main effects, the proportions of numerous taxa were significantly altered by the DF level and also

a few by the protein level (Fig. S4). Fig. 5 shows 10 of the most abundant genera in caecum, mid

colon and faecal samples, including part of taxa that were significantly changed by fiber or protein

levels. Consumption of high DF diets led to increased relative abundance of Blautia,

Faecalibacterium and Peptococcus in all three gut segments, and Prevotella in Co2 and faeces (Fig.

S4 and Fig. 5). Additionally, other taxa including Actinobacteria, Coriobacteriia and Ruminococcus

in Co2, and Phascolarctobacterium, Collinsella, Odoribacter, Mitsuokella, Sutterella and CF231 in

faeces of pigs were found enriched with the high DF diets (Fig. S4). The low DF diets were associated

with higher relative abundance of Turicibacter and Anaerovorax in Co2 and faeces compared to the

84

high DF (Fig. S4 and Fig. 5), some microbiota such as Ruminococcaceae and Streptococcus in faeces

of pigs were also enriched with the intake of low DF diets (Fig. S4). No protein effect was found in

caecal microbiota, but high protein diets were associated with higher relative abundance of

Peptococcaceae in Co2 and Clostridium in faecal samples compared to low protein (Fig. S4 and Fig.

5), while Lachnospira was observed in faecal samples from pigs fed the low protein diets (Fig. S4).

3.4. Correlations of NSP, SCFA and microbiota

Correlation analyses of components of carbohydrates, SCFA and body weight with the relative

abundance of intestinal bacteria were performed for the caecum (Figure 6A) and Co2 (Figure 6B).

Total NSP and AX contents were positively correlated with relative abundance of Blautia both in

caecum and Co2. In addition, significant positive correlations were seen for the contents of uronic

acids, galactose and cellulose in Co2, but were not significant in caecum. Negative correlations were

identified between rhamnose contents and relative abundance of Lactobacillus, and between uronic

acid and Escherichia in caecum, whereas colonic starch contents correlated positively with

Escherichia. Fucose contents correlated positively with Erysipelotrichaceae in the caecum and with

Parabacteroides in Co2. A significant positive correlation was also seen between Bacteriodales and

the proportion of NCP-glucose in NSP in Co2.

The caecal proportion of SCFA in the form of acetate correlated positively with Ruminococcus

(family Ruminococcaceae, phylum Firmicutes) and was inversely associated with the relative

abundance of Megasphaera. The concentrations of propionate showed positive correlations with

Treponema in the caecum, and with Mitsuokella and Megasphaera in mid-colon, but the proportion

of propionate in caecal digesta correlated positively with the relative abundance of Streptococcus and

Lactobacillus, and inversely with Lachnospiraceae. The proportion of butyrate correlated positively

with the relative abundance of Coprococcus both in the caecum and Co2. In caecal content, the

proportion of SCFA in the form of BCFA correlated negatively with Blautia, whereas positive

correlations were seen with Mogibacteriaceae and Parabacteroides. The latter two bacteria were also

positively correlated with the concentration BCFA in Co2. In addition, a negative correlation was

found between BCFA pool size and CF231 in Co2. Ruminococcaceae in caecum showed a positive

correlation with pH, whereas pH and Prevotella correlated negatively in Co2.

Body weight of the animals was negatively correlated with the abundance of Ruminococcus (family

Lachnospiraceae, phylum Firmicutes) in caecum and Co2. No significant correlations between faecal

85

SCFA and microbiota or between intestinal microbiota and plasma metabolites were observed (data

not shown).

3.5. Plasma SCFA concentrations

There was no dietary effect on the concentrations of SCFA in plasma from the portal vein except that

the concentration of BCFA was higher with the high protein diets (P = 0.007, Table 3). In the jugular

vein, propionate concentrations were higher (P < 0.001) and succinate concentrations were lower (P

= 0.001) with high protein compared to the low protein diets. High DF induced higher butyrate

concentrations compared with low DF in the jugular vein (P = 0.03), but was not significantly affected

in the portal vein (P > 0.25).

4. Discussion

A major purpose of the treatment of wheat bran with cell-wall degrading enzymes was to transform

insoluble AX in the aleurone cells into AXOS.35 Since the enzymes used had a preference for

unsubstituted regions of AX, the outcome was the lower substitution of AXOS than the parent AX.35

This is also found in the present study and the reason for the lower A:X ratio of AXOS and higher

A:X ratio of IAX in diet and SI3 digesta of the high compared to the low DF diets. The structural

characteristics of AX can influence the site for microbial degradation.11, 36 The minor reduction in

AXOS in digesta from the distal third of the small intestine in high DF groups indicates that some

has already been utilized by the microbiota in stomach and small intestine.36 It is also likely that the

increase in the proportion of soluble AX in SI3 compared to the diet is a consequence of the digestion

processes in stomach and small intestine, making more AX soluble and ready for fermentation in the

large intestine.37, 38 The functional properties of AX are associated with the extent of substitution and

distribution of substituents such as arabinose unit and ferulic acid along the xylan backbone, which

is partly reflected by different A:X ratios.39 Our finding that the A:X ratios of soluble AX increased

from diets to SI3 digesta in the high DF groups supported that mainly the less-substituted parts of the

xylan chain were cleaved in the small intestine.36 Similar to a previous study,16 we also found that

some AX was fermented already in the caecum and continued to be fermented in proximal colon,

resulting in an increased degree of arabinose substitution and slightly decreased proportion of AX in

NSP along the colon. However, we observed almost similar A:X ratios between mid and distal colon

most likely because the AX with a high degree of substitution was hardly fermented after mid colon

as also found in other studies.40 A similar pattern has been observed with AX from wheat and rye in

pigs and humans.11

86

Fermentation of DF in the large intestine is one of the main functions of the intestinal microbiota.

The concentrations and distribution of SCFA among the intestinal segments followed in general the

changes in NSP. An increase in the SCFA concentration from distal ileum to caecum and proximal

colon was caused by high bacterial growth and fermentation due to sufficient quantities of readily

fermentable substrate.41 It was followed by a relative decline in the more distal parts of the colon

presumably reflecting depletion of substrate and absorption of SCFA.16, 42 The lower SCFA and

acetate concentrations with high DF can in part be caused by dilution of intestinal content caused by

insoluble difficult degradable DF constituents – cellulose, IAX, and lignin – as found in previous

studies.16 The slightly higher butyrate pool size in the large intestine of the pigs fed the high DF diets

support earlier findings that AX is a substrate that can stimulate butyrate production in the large

intestine. 16, 17

BCFA are produced exclusively from fermentation of protein,43 but protein fermentation will also

contribute to SCFA production. This is probably responsible for the increases in SCFA, propionate

and BCFA concentrations along with borderline significant increases in acetate concentrations along

the large intestine with the high protein diets. A noteworthy finding was that the change of BCFA

concentrations and proportions in the large intestine of the pigs depended on DF levels with lower

proportions and smaller increase in protein fermentation in the mid and distal colon in the pigs fed

high DF. Protein fermentation occurs mainly in the distal colon where available carbohydrates are

becoming scarce,8 and the more resistant and slowly fermentable carbohydrates of the high DF diets

groups might suppress protein fermentation in the distal colon. Previously, consumption of an AX

rich diet has been shown to lower p-cresol in the caecum, which indicated a reduced protein

fermentation44. This was confirmed in another study showing that fibre addition to a high protein diet

reduced the concentrations of BCFA and putatively toxic metabolites, and shifted the microbiota

metabolism from fermenting protein to carbohydrate.45 As ammonia and other end-products of

protein fermentation are basic, the lower protein fermentation induced by the high DF diet in distal

colon can be explanatory for the lower pH values, which may be beneficial in inhibiting the growth

of some pathogens, reducing inflammatory response and exert a protective function against colon

cancer.46 Corresponding to the concentrations of total SCFA being highest in caecum, the lowest pH

values was observed here for all diets, which was linked with a high microbiota activity16 and a

relatively high supply of readily fermentable carbohydrates.37

The modulated intestinal SCFA concentrations were, however, not translated to the portal vein

where only a higher BCFA concentration was in line with the higher intestinal concentrations and

87

pool size of BCFA in pigs fed the high protein diets. We speculate that utilization of SCFA in the

epithelium of colon5 partly abolished the significant differences within dietary treatments in the portal

vein. SCFA can directly influence satiety by modulating intestinal hormones13 and a previous study

indicated that butyrate provided orally reduced adiposity and improved insulin sensitivity by

increasing energy expenditure and fat oxidation in high fat diet-fed mice.47 We also found that the

concentration of butyrate was elevated in peripheral blood with the high DF diets, which potentially

may play a role for the increased C-peptide levels and the downregulated fatty acid synthase gene

expression in adipose tissue observed.23 It could also be a contributing factor for the beneficial effects

of DF on body weight gain seen in the present experiment.23

Intriguingly, the propionate concentration in jugular plasma was significantly higher while

succinate was lower with the high than the low protein. Circulating propionate and butyrate

concentrations have been positively associated with the plasma GLP-1 concentrations,48, 49 suggesting

that higher propionate and butyrate concentrations in jugular vein could contribute to a tendency of

increased plasma GLP-1 concentrations seen in this study.23 The reduced concentrations in peripheral

circulation of succinate, an intermediate of the tricarboxylic acid (TCA) cycle, could be a result of

increased TCA cycle activity with high protein intake indicating more substrate used for

gluconeogenesis, which has been associated with insulin resistance.50 This interpretation is supported

by our previous observation of an upregulated gene expression of the rate-limiting enzyme in

gluconeogenesis (fructose -1,6-bisphosphatase 1) in the liver.23

We observed increased Bacteroidetes:Firmicutes ratios in faecal samples of pigs fed the high DF

diets, which is consistent with a study in mice fed enzyme-treated wheat bran.51 Greater

Bacteroidetes:Firmicutes ratios in lean individuals or animals than in obese ones have also been

reported.52 Moreover, a significant association of increased faecal Bacteroidetes proportion with

weight loss has been found.53 However, it should be noticed that a lower Bacteroidetes:Firmicutes

ratio is not a hallmark of obesity and changed microbial diversity may be a more relevant factor.54

The increased faecal Bacteroidetes abundance corroborated with increased faecal propionate levels

in the pigs fed high DF diets, while a higher alpha diversity could be a result of a better supply of

carbohydrate substrate in distal large intestine. Collectively, these outcomes were possibly linked to

the reduced weight gain with high DF diets in this study.23 The increased abundance of butyrate-

producing bacteria according to the LEfSe analysis in pigs fed the high DF diets may explain the

slightly increased butyrate pool size in caecum and colon. In agreement with this study, a previous

study from our group has shown that an AX rich diet resulted in higher relative abundance of butyrate-

88

producing bacteria such as Faecalibacterium spp. and Blautia spp.16 In line with these findings, whole

grain and bran rich diets may promote a select group of SCFA-producing strains thereby reduce the

risk of obesity and T2D.55 Several studies have demonstrated that inflammatory bowel disease

patients and pre-diabetic human subjects have lower abundance of butyrate-producing bacteria in

faeces than healthy subjects.56, 57 Therefore, the higher relative abundance of butyrogenic microbiota

observed when providing this type of high DF diet could be considered as one of the potential health-

promoting effects of DF on gut ecology. We also found that the high DF diets increased relative

abundance of Prevotella in colon and faeces, which previously was found to play a role in AX

breakdown,44 and associated with the lower pH values in the distal colon of pigs fed the high DF

diets. Interestingly, we found a negative correlation between body weight and Ruminococcus

abundance, possibly revealing the close association of body weight with SCFA produced from this

genus since SCFA plays an important role in lipid turnover and energy homeostasis.5

Although no clear significant correlations between faecal microbiota and SCFA profile were

observed, some bacteria abundant in faeces of the pigs fed high DF such as Phascolarctobacterium

spp. was found to be producer of propionate58 which also was increased in faeces of these pigs. The

faecal samples from the low DF groups were enriched in predominant members of the Firmicutes

phylum including Ruminococcaceae, Streptococcus, Turicibacter and Anaerovorax, the latter two

were also found in high abundance in the colon. Streptococcus and Turicibacter has previously been

reported as lactate-producing bacteria and abundant in low DF diets,59 while Ruminococcaceae might

be responsible for converting lactate to SCFA.60

The increased intestinal propionate concentrations with high protein diets were reflected by minor

alterations of microbiota in mid colon, wherein the growth of propionate-oxidizing bacteria

Peptococcaceae61 was stimulated due to the abundant substrates and partly contributed to the

tendency of increased acetate concentrations. A prebiotic effect of whey protein has previously been

reported, showing increased Lactobacillus and decreased Clostridium with the consumption of whey

protein in high fat diet-fed mice.62 In contrast, abundant Clostridium was observed in faecal samples

of pigs fed high protein diets in this study. The mechanism of the distinct changes of bacteria is not

clear, but both the DF and protein levels, sources and structure could be possible modulating factors.

For instance, high DF may inhibit proteolytic bacteria fermentation processes which was further

reflected in the negative correlation between BCFA proportion and Blautia in caecum as well as

BCFA production and CF231 (member of proteolytic family Paraprevotellaceae) in the mid colon.

Moreover, the absence of major changes of bacteria in the large intestine could be explained by

89

efficient absorption of whey proteins in the small intestine.21 Clostridium can ferment carbohydrate

and besides that, amine-producing species belonging to Clostridium have been previously correlated

with BCFA abundance with high protein diets, which may be an additional risk factor for intestinal

infection.8 However, in the current study the high dietary level of protein was obtained from cereal,

whey and fish proteins, which are thought to have less detrimental effects than other protein sources.63

Additionally, the protein level of high protein diets was only 5% higher than the recommended

standard diets for minipigs (18% vs 13%) and no signs of inflammation showed up with the high

protein diets.23 Therefore, the potentially detrimental effects of high dietary protein on intestinal

function reported previously8 should not be exaggerated in the current study especially when it is

supplemented with abundant DF.

5. Conclusion

The present study supports the hypothesis that high DF intake could alter the intestinal carbohydrate

metabolism with a continuous degradation of AX until the mid-colon in this obese Göttingen Minipig

model. AX rich DF also modestly increased butyrate production, which was further supported by the

enriched butyrogenic genera. Proteolytic fermentation was suppressed by high DF and induced minor

changes in the microbiome profile, showing no prebiotic effects. However, high protein significantly

changed circulating SCFA profile associated with increased gluconeogenesis, whilst high DF

increased circulating butyrate concentrations. Overall, the modulated microbial ecosystem and SCFA

profile triggered by AX-enriched DF diets could be the potential mechanism underlying the positive

effects of DF on weight gain regulation, hormone response and metabolic disorders, while high

protein did not show beneficial effects on metabolic health.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgement

The authors thank Winnie Østergaard, Lisbeth Mӓrcher, Stina Greis Handberg, Kasper Vrangstrup

Poulsen and Thomas Rebsdorf for excellent technical assistance. We thank Leslie Foldager for

consultation on data handling. Whey protein hydrolysate was kindly provided by Arla Foods

Ingredients Group P/S. Wheat bran was delivered by Lantmӓnnen and enzymatically treated by

DuPont Industrial Bioscience. The work was financially supported by Innovation Fund Denmark

(4105-00002B) and industrial partners involved in the MERITS (Metabolic Changes by Carbohydrate

and Protein Quality in the Development and Mitigation of Metabolic Syndrome) project. Y.T.X.

90

acknowledges scholarship from China Scholarship Community and Graduate School for Science and

Technology of Aarhus University.

Reference

1. M. Romani-Perez, A. Agusti and Y. Sanz, Innovation in microbiome-based strategies for

promoting metabolic health, Curr Opin Clin Nutr Metab Care, 2017, 20, 484-491.

2. V. K. Ridaura, J. J. Faith, F. E. Rey, J. Cheng, A. E. Duncan, A. L. Kau, N. W. Griffin, V.

Lombard, B. Henrissat, J. R. Bain, M. J. Muehlbauer, O. Ilkayeva, C. F. Semenkovich, K. Funai,

D. K. Hayashi, B. J. Lyle, M. C. Martini, L. K. Ursell, J. C. Clemente, W. Van Treuren, W. A.

Walters, R. Knight, C. B. Newgard, A. C. Heath and J. I. Gordon. Gut microbiota from twins

discordant for obesity modulate metabolism in mice, Science, 2013, 341, 1241214.

3. J. L. Sonnenburg and F. Backhed, Diet-microbiota interactions as moderators of human

metabolism, Nature, 2016, 535, 56-64.

4. D. K. Dahiya, Renuka, M. Puniya, U. K. Shandilya, T. Dhewa, N. Kumar, S. Kumar, A. K.

Puniya and P. Shukla, Gut Microbiota Modulation and Its Relationship with Obesity Using

Prebiotic Fibers and Probiotics: A Review, Front Microbiol, 2017, 8, 563.

5. D. J. Morrison and T. Preston, Formation of short chain fatty acids by the gut microbiota and

their impact on human metabolism. Gut Microbes, 2016, 7, 189-200.

6. G. O. Phillips and S. W. Cui, An introduction: Evolution and finalisation of the regulatory

definition of dietary fibre, Food Hydrocolloids, 2011, 25, 139-143.

7. M. W. J. Wong, R. de Souza, C. W. C. Kendall, A. Emam and D. J. A. Jenkins, Colonic Health:

Fermentation and Short Chain Fatty Acids, J Clin Gastroenterol, 2006, 40, 235–243.

8. N. E. Diether and B. P. Willing, Microbial Fermentation of Dietary Protein: An Important

Factor in Diet(-)Microbe(-)Host Interaction, Microorganisms, 2019, 7.

9. M. S. Izydorczyk and C. G. Biliaderi. Cereal arabinoxylans: advances in structure and

physicochemical properties, Carbohydrate Polymers, 1995, 28, 33-48.

10. K. E. Bach Knudsen, A. Serena, A. K. Bjørnbak Kjær, H. Jørgensen and R. Engberg, Rye Bread

Enhances the Production and Plasma Concentration of Butyrate but Not the Plasma

Concentrations of Glucose and Insulin in Pigs1, J Nutr, 2005, 135, 1696–1704.

11. H. N. Lærke and K. E. Bach Knudsen, Rye Arabinoxylans: Molecular Structure,

Physicochemical Properties and Physiological Effects in the Gastrointestinal Tract, Cereal

Chem, 2010, 87, 353-362.

12. A. K. Ingerslev, P. K. Theil, M. S. Hedemann, H. N. Laerke and K. E. Bach Knudsen, Resistant

starch and arabinoxylan augment SCFA absorption, but affect postprandial glucose and insulin

responses differently, Br J Nutr, 2014, 111, 1564-1576.

13. F. Brighenti, L. Benini, D. Del Rio, C. Casiraghi, N. Pellegrini, F. Scazzina, D. J. A. Jenkins

and I. Vantini. Colonic fermentation of indigestible carbohydrates contributes to the second-

meal effect, Am J Clin Nutr, 2006, 83, 817-822.

14. S. Hald, A. G. Schioldan, M. .E Moore, A. Dige, H. N. Laerke, J. Agnholt, K. E. Bach Knudsen,

K. Hermansen, M. L. Marial, S. Gregersen and J. F. Dahlerup. Effects of Arabinoxylan and

Resistant Starch on Intestinal Microbiota and Short-Chain Fatty Acids in Subjects with

Metabolic Syndrome: A Randomised Crossover Study, PLoS One, 2016, 11, e0159223.

15. B. A. Williams, D. Zhang, A. T. Lisle, D. Mikkelsen, C. S. McSweeney, S. Kang, W. L. Bryden,

M. J. Gidley, Soluble arabinoxylan enhances large intestinal microbial health biomarkers in pigs

fed a red meat-containing diet, Nutrition, 2016, 32, 491-497.

16. T. S. Nielsen, H. N. Laerke, P. K. Theil, J. F. Sorensen, M. Saarinen, S. Forssten, K. E. Bach

Knudsen. Diets high in resistant starch and arabinoxylan modulate digestion processes and

91

SCFA pool size in the large intestine and faecal microbial composition in pigs, Br J Nutr, 2014,

112, 1837-1849.

17. B. Damen, L. Cloetens, W. F. Broekaert, I. Francois, O. Lescroart, I. Trogh, F. Arnaut, G. W.

Welling, J. Wijffels, J. A. Delcour, K. Verbeke, C. M. Courtin, Consumption of breads

containing in situ-produced arabinoxylan oligosaccharides alters gastrointestinal effects in

healthy volunteers, J Nutr, 2012, 142, 470-477.

18. A. C. Ouwehand, M. Derrien, W. de Vos, K. Tiihonen, N. Rautonen, Prebiotics and other

microbial substrates for gut functionality, Curr Opin Biotechnol, 2005, 16, 212-217.

19. A. Bjornshave, J. J. Holst and K. Hermansen, Pre-Meal Effect of Whey Proteins on Metabolic

Parameters in Subjects with and without Type 2 Diabetes: A Randomized, Crossover Trial,

Nutrients, 2018, 10.

20. R. C. Sprong, A. J. Schonewille and R. van der Meer, Dietary cheese whey protein protects rats

against mild dextran sulfate sodium-induced colitis: role of mucin and microbiota, J Dairy Sci,

2010, 93, 1364-1371.

21. B. Tranberg, L. I. Hellgren, J. Lykkesfeldt, K. Sejrsen, A. Jeamet, I. Rune, M. Ellekilde, D. S.

Nielsen and A. K. Hansen, Whey protein reduces early life weight gain in mice fed a high-fat

diet, PLoS One, 2013, 8, e71439.

22. T. Sanchez-Moya, R. Lopez-Nicolas, D. Planes, C. A. Gonzalez-Bermudez, G. Ros-Berruezo,

C. Frontela-Saseta. In vitro modulation of gut microbiota by whey protein to preserve intestinal

health. Food Funct, 2017, 8, 3053-3063.

23. Y. Xu, M. V. Curtasu, K. E. Bach Knudsen, M. S. Hedemann, P. K. Theil and H. N. Lærke,

Dietary fibre and protein do not synergistically influence insulin, metabolic or inflammatory

biomarkers in young obese Göttingen Minipigs, Br J Nutr, 2020, preprint, DOI:

https://doi.org/10.1017/S0007114520003141.

24. E. Roura, S-J. Koopmans, J-P. Lallès, I. Le Huerou-Luron, N. de Jager, T. Schuurman, D. Val-

Laillet, Critical review evaluating the pig as a model for human nutritional physiology, Nutr

Res Rev, 2016, 29, 60-90.

25. M. V. Curtasu, Obesity and metabolic syndrome in miniature pigs as models for human disease

– metabolic changes in response to ad libitum feeding of high-fat-high-carbohydrate diets, PhD

thesis, Aahus University, 2019.

26. K. E. Bach Knudsen, Carbohydrate and lignin contents of plant materials used in animal

feeding, Anim Feed Sci Technol, 1997, 319-338.

27. M. T. Jensen, R. P. Cox, B. B. Jensen, Microbial production of skatole in the hind gut of pigs

given different diets and its relation to skatole deposition in backfat, Anim Sci, 1995, 61, 293-

304.

28. J. Han, K. Lin, C. Sequeira, C. H. Borchers, An isotope-labeled chemical derivatization method

for the quantitation of short-chain fatty acids in human feces by liquid chromatography-tandem

mass spectrometry, Anal Chim Acta, 2015, 854, 86-94.

29. N. A. Bokulich, C. M. L. Joseph, G. Allen, A. K. Benson, D. A. Mills, Next-Generation

Sequencing Reveals Significant Bacterial Diversity of Botrytized Wine, PLOS ONE, 2012, 7,

e36357.

30. E. A. McInnis, K. M. Kalanetra, D. A. Mills, E. A. Maga, Analysis of raw goat milk microbiota:

Impact of stage of lactation and lysozyme on microbial diversity, Food Microbiol, 2015, 46,

121-131.

31. N. A. Bokulich, B. D. Kaehler, J. R. Rideout, M. Dillon, E. Bolyen, R. Knight, G. A. Huttley

and J. G. Caporaso, Optimizing taxonomic classification of marker-gene amplicon sequences

with QIIME 2's q2-feature-classifier plugin, Microbiome, 2018, 6, 90.

32. D. P. Faith, Conservation evaluation and phylogenetic diversity, Biol Conserv, 1992, 61, 1-10.

92

33. C. Lozupone and R. Knight, UniFrac: a new phylogenetic method for comparing microbial

communities, Appl Environ Microbiol, 2005, 71, 8228-8235.

34. C. Lozupone, M. E. Lladser, D. Knights, J. Stombaugh, R. Knight, UniFrac: an effective

distance metric for microbial community comparison, ISME J, 2011, 5, 169-172.

35. C. T. Vangsøe, J. F. Sørensen and K. E. Bach Knudsen, Aleurone cells are the primary

contributor to arabinoxylan oligosaccharide production from wheat bran after treatment with

cell wall‐degrading enzymes, Int J Food Sci Tech, 2019, 54, 2847-2853.

36. M. M. Kasprzak, H. N. Laerke and K. E. Bach Knudsen, Effects of isolated and complex dietary

fiber matrices in breads on carbohydrate digestibility and physicochemical properties of ileal

effluent from pigs, J Agric Food Chem, 2012, 60, 12469-12476.

37. K. E. Bach Knudsen, B. B. Jensen and H. Inge, Oat bran but not a 0-glucan-enriched oat fraction

enhances butyrate production in the large intestine of pigs1, J Nutr, 1993, 123, 1235-1247.

38. K. E. Bach Knudsen and I. Hansen, Gastrointestinal implications in pigs of wheat and oat

fractions1. Digestibility and bulking properties of polysaccharides and other major constituents,

Br. J Nutr, 1991, 65, 217-232.

39. M. S. Izydorczyk and C. G. Biliaderis, Influence of structure on the physicochemical properties

of wheat arabinoxylan, Carbohydr Polym, 1992, 17, 237-247.

40. L. V. Glitsø, H. Gruppen, H. A. Schols, S. Højsgaard, B. Sandstro¨m and K. E. Bach Knudsen,

Degradation of rye arabinoxylans in the large intestine of pigs, J Sci Food Agric, 1999, 79, 961-

969.

41. D. L. Topping and P. M. Clifton, Short-Chain Fatty Acids and Human Colonic Function: Roles

of Resistant Starch and Nonstarch Polysaccharides, Physiol Rev, 2001, 81, 1031-1064.

42. M. Le Gall, A. Serena, H. Jorgensen, P. K. Theil and K. E. Bach Knudsen, The role of whole-

wheat grain and wheat and rye ingredients on the digestion and fermentation processes in the

gut--a model experiment with pigs, Br J Nutr, 2009, 102, 1590-1600.

43. R. Jha and J. F. D. Berrocoso. Dietary fiber and protein fermentation in the intestine of swine

and their interactive effects on gut health and on the environment: A review, Anim Feed Sci

Technol, 2016, 212, 18-26.

44. D. P. Belobrajdic, A. R. Bird, M. A. Conlon, B. A. Williams, S. Kang, C. S. McSweeney, D.

Zhang, W. L. Bryden, M. J. Gidley and D. L. Topping, An arabinoxylan-rich fraction from

wheat enhances caecal fermentation and protects colonocyte DNA against diet-induced damage

in pigs, Br J Nutr, 2011, 107, 1274-1282.

45. R. Pieper, S. Kroger, J. F. Richter, J. Wang, L. Martin, J. Bindelle, J. K. Htoo, D. von Smolinski,

W. Vahjen, J. Zentek and A. G. Van Kessel, Fermentable fiber ameliorates fermentable protein-

induced changes in microbial ecology, but not the mucosal response, in the colon of piglets, J

Nutr, 2012, 142, 661-667.

46. V. Salvador, C. Cherbut, J-L. Barry, D. Bertrand, C. Bonnet, J. Delort-Laval, Sugar composition

of dietary fibre and short-chain fatty acid production during in vitro fermentation by human

bacteria, Br J Nutr, 1993, 70, 189-197.

47. Z. Gao, J. Yin, J. Zhang, R. E. Ward, R. J. Martin, M. Lefevre, . T. Cefalu and J. Ye, Butyrate

improves insulin sensitivity and increases energy expenditure in mice, Diabetes, 2009, 58,

1509-1517.

48. M. Müller, M. A. G. Hernández, G. H. Goossens, D. Reijnders, J. J. Holst, J. W. E. Jocken, H.

van Eijk, E. E. Canfora and E. E. Blaak, Circulating but not faecal short-chain fatty acids are

related to insulin sensitivity, lipolysis and GLP-1 concentrations in humans, Sci Rep, 2019, 9.

49. K. R. Freeland, C. Wilson and T. M. Wolever. Adaptation of colonic fermentation and

glucagon-like peptide-1 secretion with increased wheat fibre intake for 1 year in

hyperinsulinaemic human subjects, Br J Nutr, 2010, 103, 82-90.

93

50. S. Satapati, N. E. Sunny, B. Kucejova, X. Fu, T. T. He, A. Mendez-Lucas, J. M. Shelton, J. C.

Perales, J. D. Browning and S. C. Burgess, Elevated TCA cycle function in the pathology of

diet-induced hepatic insulin resistance and fatty liver, J Lipid Res, 2012, 53, 1080-1092.

51. D. A. Kieffer, B. D. Piccolo, M. L. Marco, E. B. Kim, M. L. Goodson, M. J. Keenan, T. N.

Dunn, K. E. Bach Knudsen, S. H. Adams and R. J. Martin, Obese Mice Fed a Diet Supplemented

with Enzyme-Treated Wheat Bran Display Marked Shifts in the Liver Metabolome Concurrent

with Altered Gut Bacteria, J Nutr, 2016, 146, 2445-2460.

52. A. Rivera-Piza and S-J. Lee, Effects of dietary fibers and prebiotics in adiposity regulation via

modulation of gut microbiota, Appl Biol Chem, 2020, 63.

53. E. Ruth, P. J. T. Ley, K. Samuel and I. J. Gordon, Human gut microbes associated with obesity,

Nature, 2006, 444, 1022-1023.

54. F. Magne, M. Gotteland, L. Gauthier, A. Zazueta, S. Pesoa, P. Navarrete and R. Balamurugan,

The Firmicutes/Bacteroidetes Ratio: A Relevant Marker of Gut Dysbiosis in Obese Patients?

Nutrients, 2020, 12.

55. M. C. W. Myhrstad, H. Tunsjo, C. Charnock, V. H. Telle-Hansen, Dietary Fiber, Gut

Microbiota, and Metabolic Regulation-Current Status in Human Randomized Trials, Nutrients,

2020, 12.

56. X, Zhang, D. Shen, Z. Fang, Z. Jie, X. Qiu, C. Zhang, Y. Chen and L. J, Human gut microbiota

changes reveal the progression of glucose intolerance, PLoS One, 2013, 8, e71108.

57. R. K. Singh, H. W. Chang, D. Yan, K. M. Lee, D. Ucmak, K. Wong, M. Abrouk, B. Farahnik,

M. Nakamura, T. Zhu, T. Bhutani and W. Liao, Influence of diet on the gut microbiome and

implications for human health, J Transl Med, 2017, 15, 73.

58. P. Louis and H. J. Flint, Formation of propionate and butyrate by the human colonic microbiota,

Environ Microbiol, 2017, 19, 29-41.

59. L. F. Gomez-Arango, H. L. Barrett, S. A. Wilkinson, L. K. Callaway, H. D. McIntyre, M.

Morrison and M. D. Nitert, Low dietary fiber intake increases Collinsella abundance in the gut

microbiota of overweight and obese pregnant women, Gut Microbes, 2018, 9, 189-201.

60. H. J. Flint, K. P. Scott, S. H. Duncan, P. Louis and E. Forano, Microbial degradation of complex

carbohydrates in the gut, Gut Microbes, 2012, 3, 289-306.

61. N. Muller, P. Worm, B. Schink, A. J. Stams, C. M. Plugge, Syntrophic butyrate and propionate

oxidation processes: from genomes to reaction mechanisms, Environ Microbiol Rep, 2010, 2,

489-499.

62. L. McAllan, P. Skuse, P. D. Cotter, P. O'Connor, J. F. Cryan, R. P. Ross, G. Fitzgerald, H. M.

Roche and K. N. Nilaweera, Protein quality and the protein to carbohydrate ratio within a high

fat diet influences energy balance and the gut microbiota in C57BL/6J mice, PLoS One, 2014,

9, e88904.

63. P. Lopez-Legarrea, R. de la Iglesia, I. Abete, S. Navas-Carretero, J. A. Martinez, M. A. Zulet,

The protein type within a hypocaloric diet affects obesity-related inflammation: the RESMENA

project, Nutrition, 2014, 30, 424-429.

94

Table 1 Carbohydrate composition of digesta in different intestinal segments of Göttingen Minipigs

fed diets low or high in dietary fibre and protein.

Item Segment Diet1

SEM P-value2

LOFLOP LOFHIP HIFLOP HIFHIP F P F×P Segment

Starch, % of

DM

SI3 2.8 4.1 2.3 2.2 0.54 0.11 0.06 0.56 <0.001

Caecum 2.8 3.9 2.1 3.2

Co1 1.6 2.4 1.5 2.2

Co2 1.0 1.4 0.98 1.2

Co3 0.81 1.1 0.78 0.86

Total NSP, %

of DM

SI3 35 34 43 40 1.5 <0.001 0.004 0.07 <0.001

Caecum 34 32 43 41

Co1 30 29 40 38

Co2 28 29 40 37

Co3 29 30 41 38

AX, % of DM SI3 22 21 27 25 0.96 <0.001 0.02 0.11 <0.001 Caecum 20 19 25 24

Co1 17 16 23 22

Co2 16 16 23 21

Co3 16 17 24 22

A:X3 SI3 0.60 0.62 0.69 0.69 0.02 0.001 0.35 0.20 <0.001 Caecum 0.80 0.86 0.92 0.88

Co1 0.92 0.97 1.0 1.0

Co2 1.1 1.1 1.1 1.1

Co3 1.1 1.1 1.1 1.1

AX:NSP, % SI3 62 61 63 64 0.61 <0.001 0.74 0.75 <0.001 Caecum 58 57 58 59

Co1 57 57 58 59

Co2 56 56 58 58

Co3 56 56 58 57

Abbreviations: NSP, non-starch polysaccharides; AX, arabinoxylan; A:X, arabinose:xylose ratio;

SI3, distal small intestine; Co1, proximal colon; Co2, mid colon; Co3, distal colon. 1The individual minipig was regarded as the experimental unit, n = 10 for low fibre low protein diet

(LOFLOP), n = 10 for low fibre high protein diet (LOFHIP), n = 12 for high fibre low protein diet

(HIFLOP) and n = 11 for high fibre high protein diet (HIFHIP). Data were expressed as means with

standard error of means (SEM). 2 F, fibre effect; P, protein effect; F × P, fibre × protein interaction. Interactions between fibre or/and

protein and segment were not presented in the table and only presented in a footnote. 3A significant fibre × segment interaction (P < 0.001).

95

Table 2 Short chain fatty acid distribution (% of total SCFA) in different intestinal segments of

Göttingen Minipigs fed diets low or high in dietary fibre and protein.

% of total

SCFA Segment

Diet1 SEM

P-value2

LOFLOP LOFHIP HIFLOP HIFHIP F P F×P Segment

Acetate SI3 61 62 64 66 3.0 0.56 0.76 0.73 <0.001 Caecum 68 67 65 68

Co1 65 64 62 62

Co2 63 62 61 59

Co3 63 59 60 57

Propionate SI3 1.2 1.0 0.9 1.4 1.3 0.55 0.12 0.42 <0.001 Caecum 17 18 17 18

Co1 17 18 17 19

Co2 17 18 16 19

Co3 16 17 17 19

Butyrate3 SI3 6.5 7.5 1.4 0.7 2.1 0.71 0.75 0.33 <0.001 Caecum 13 12 15 12

Co1 14 14 17 14

Co2 15 15 17 16

Co3 14 18 17 17

BCFA4 SI3 0.4 0.0 0.2 0.1 0.33 0.002 0.33 0.99 <0.001 Caecum 1.2 1.5 0.9 1.0

Co1 1.6 2.0 1.4 1.6

Co2 2.8 3.3 2.3 2.6

Co3 4.3 4.1 2.7 3.0

Abbreviations: SCFA, short-chain fatty acids; BCFA, branched-chain fatty acids; SI3, distal small

intestine; Co1, proximal colon; Co2, mid colon; Co3, distal colon. 1The individual minipig was regarded as the experimental unit, n = 10 for low fibre low protein diet

(LOFLOP), n = 10 for low fibre high protein diet (LOFHIP), n = 12 for high fibre low protein diet

(HIFLOP) and n = 11 for high fibre high protein diet (HIFHIP). Data were expressed as means with

standard error of means (SEM). 2 F, fibre effect; P, protein effect; F × P, fibre × protein interaction. Interactions between fibre or/and

protein and segment are not presented in the table and only presented in footnotes. 3Significant fibre × segment interaction (P < 0.001) and protein × segment interaction (P = 0.050). 4Significant fibre × segment interaction (P = 0.002).

96

Table 3 Plasma short-chain fatty acid, succinate and branched-chain fatty acid concentrations

(μmol/L) in portal and jugular vein of Göttingen Minipigs fed diets low or high in dietary fibre and

protein.

Item Diet1

SEM P-value2

LOFLOP LOFHIP HIFLOP HIFHIP F P F×P

Portal vein

SCFA 426 415 348 347 66 0.26 0.92 0.93

Acetate 322 281 263 273 38 0.38 0.67 0.51

Propionate 54 62 39 35 18 0.21 0.88 0.70

Butyrate 28 33 27 13 10 0.25 0.64 0.27

Valerate 5.7 4.1 2.6 3.0 1.4 0.13 0.69 0.46

Succinate 38 41 33 39 4.7 0.41 0.25 0.73

BCFA 17 27 16 24 3.2 0.48 0.007 0.66

Jugular vein

SCFA 164 157 164 170 17 0.60 0.99 0.61

Acetate 149 137 149 150 17 0.61 0.71 0.62

Propionate 1.2 2.0 1.2 2.1 0.17 0.81 <0.001 0.91

Butyrate 0.82 0.59 1.3 1.1 0.29 0.03 0.40 0.83

Valerate 0.08 0.08 0.10 0.07 0.02 0.59 0.26 0.38

Succinate 17 14 18 13 1.5 0.71 0.001 0.40

BCFA 14 14 12 17 2.4 0.81 0.20 0.36

Abbreviations: SCFA, short-chain fatty acids; BCFA, branched-chain fatty acids. 1Minipigs were regarded as the experimental units, n = 10 for low fibre low protein diet (LOFLOP),

n = 10 for low fibre high protein diet (LOFHIP), n = 12 for high fibre low protein diet (HIFLOP) and

n = 11 for high fibre high protein diet (HIFHIP). Data were expressed as means with standard error

of means (SEM). 2 F, fibre effect; P, protein effect; F × P, fibre × protein interaction.

97

AXOS SAX IAX AXOS SAX IAX

0

20

40

60

80

100

% o

f to

tal

AX

DietSI3 digesta

LOF HIF

*

*§ §

A

*

AXOS SAX IAX AXOS SAX IAX

0.0

0.2

0.4

0.6

0.8

1.0

Ara

bin

ose

:xy

lose

DietSI3 digesta

LOF HIF

*§

§

B

*

Fig. 1 Arabinoxylan oligosaccharides (AXOS), soluble arabinoxylan (SAX) and insoluble

arabinoxylan (IAX) as percentage of total AX (A), and arabinose:xylose ratios of AXOS, SAX and

IAX (B) in diet and distal small intestine (SI3) digesta of Göttingen Minipigs fed low and high fibre

diets. Values are least-squares means with standard errors represented by vertical bars. The individual

minipig was regarded as the experimental unit, n = 20 for low fibre diets (LOF), n = 23 for high fibre

diet (HIF). * indicates the significant difference for the paired test comparison between diet and SI3

digesta (P < 0.05). § indicates significant difference at P < 0.05 in SI3 digesta between LOF and

HIF.

98

SI3 Cecum Co1 Co2 Co3

0

50

100

150

SC

FA

, mm

ol/

kg

LOFLOPLOFHIPHIFLOPHIFHIP

Protein, P = 0.03

Fiber, P = 0.02

Segment, P < 0.001

A

SI3 Cecum Co1 Co2 Co3

0

20

40

60

80

100

Ace

tate

, mm

ol/

kg

LOFLOPLOFHIPHIFLOPHIFHIP

Fiber, P = 0.02Segment, P < 0.001

B

SI3 Cecum Co1 Co2 Co3

0

5

10

15

20

25

Pro

pio

nat

e, m

mo

l/k

g

LOFLOPLOFHIPHIFLOPHIFHIP

Protein, P = 0.04Segment, P < 0.001

C

SI3 Cecum Co1 Co2 Co3

0

5

10

15

20

Bu

tyra

te,

mm

ol/

kg

LOFLOPLOFHIPHIFLOPHIFHIP

Segment, P < 0.001

D

SI3 Cecum Co1 Co2 Co3

0

1

2

3

BC

FA

, m

mo

l/k

g

LOFLOPLOFHIPHIFLOPHIFHIP

EFiber Segment, P < 0.001

Protein Segment, P = 0.04

Fig. 2 Concentrations of total SCFA (A), acetate (B), propionate (C), butyrate (D) and BCFA (E) in

distal small intestine (SI3), caecum, proximal colon (Co1), mid colon (Co2) and distal colon (Co3)

of Göttingen Minipigs. Values are least-squares means with standard errors represented by vertical

bars. The individual minipig was regarded as the experimental unit, n = 10 for low fibre low protein

diet (LOFLOP), n = 10 for low fibre high protein diet (LOFHIP), n = 12 for high fibre low protein

diet (HIFLOP) and n = 11 for high fibre high protein diet (HIFHIP). Only significant P -values are

presented in the figure, main effects were not shown when there was an interaction.

99

SI3 Cecum Co1 Co2 Co3

0

100

200

300

400

Net

am

ou

nt,

g

LOFLOPLOFHIPHIFLOPHIFHIP

Fiber Protein segment, P = 0.006A

SI3 Cecum Co1 Co2 Co3

6.0

6.5

7.0

7.5

pH

LOFLOPLOFHIPHIFLOPHIFHIP

Fiber Segment, P = 0.02

B

Fig. 3 Net amount of wet gut content (a) and pH value (b) in distal small intestine (SI3), caecum,

proximal colon (Co1), mid colon (Co2) and distal colon (Co3) of Göttingen Minipigs. Values are

least-squares means with standard errors represented by vertical bars. The individual minipig was

regarded as the experimental unit, n = 10 for low fibre low protein diet (LOFLOP), n = 10 for low

fibre high protein diet (LOFHIP), n = 12 for high fibre low protein diet (HIFLOP) and n = 11 for high

fibre high protein diet (HIFHIP). Only significant P -values are presented in the figure, main effects

were not shown when there was an interaction.

100

LOFLO

P

LOFH

IP

HIF

LOP

HIF

HIP

0

50

100

%

FirmicutesABacteroidetesProteobacteriaOthers

LOFLO

P

LOFH

IP

HIF

LOP

HIF

HIP

0

50

100

%

FirmicutesBBacteroidetesProteobacteriaOthers

LOFLO

P

LOFH

IP

HIF

LOP

HIF

HIP

0

50

100

%

FirmicutesBacteroidetesProteobacteriaOthers

C*

*

Fig. 4 Abundance of predominant phyla in caecum (A), mid colon (B) and faeces (C) of Göttingen

Minpigs fed diets low or high in dietary fibre and protein. Values are least-squares means with

standard errors represented by vertical bars. The individual minipig was regarded as the experimental

unit, n = 10 for low fibre low protein diet (LOFLOP), n = 10 for low fibre high protein diet (LOFHIP),

n = 12 for high fibre low protein diet (HIFLOP) and n = 11 for high fibre high protein diet (HIFHIP).

One digesta sample from mid colon in LOFLOP group (n = 9) was omitted due to low-quality reads.

* indicates significant fiber effects at P < 0.001 on the relative abundance of Firmicutes and

Bacteroidetes.

101

LOFLO

P

LOFH

IP

HIF

LOP

HIF

HIP

0

50

100

Rel

ativ

e ab

un

dan

ce,

%

g__Ruminococcus

g__Prevotellag__Turicibacterg__Blautiag__Coprococcusg__Lactobacillusg__Oscillospirag__Clostridiumg__Faecalibacterium

A

g__Roseburia

*

*

LOFLO

P

LOFH

IP

HIF

LOP

HIF

HIP

0

50

100

Rel

ativ

e ab

un

dan

ce,

%

g__Ruminococcusg__Roseburiag__Prevotellag__Turicibacter

g__Blautiag__Coprococcusg__Lactobacillusg__Oscillospirag__Clostridiumg__FaecalibacteriumB

*

*

*

*

*

LOFLO

P

LOFH

IP

HIF

LOP

HIF

HIP

0

50

100

Rel

ativ

e ab

un

dan

ce,

%

g__Ruminococcusg__Roseburia

g__Turicibacter

g__Blautia

g__Coprococcusg__Lactobacillusg__Oscillospirag__Clostridiumg__Faecalibacterium

C

g__Prevotella**

*

*

#

Fig. 5 Rarified taxonomy table displaying the 10 most abundant genera in caecum (A), mid colon (B)

and faecal samples (C) in Göttingen Minipigs fed low or high fibre and protein diets. The individual

minipig was regarded as the experimental unit, n = 10 for low fibre low protein diet (LOFLOP), n =

10 for low fibre high protein diet (LOFHIP), n = 12 for high fibre low protein diet (HIFLOP) and n

= 11 for high fibre high protein diet (HIFHIP). One digesta sample from mid colon in LOFLOP group

(n = 9) was omitted due to low-quality reads. * indicates significant fiber effects at P < 0.01 and #

indicates significant protein effects at P < 0.01 on the relative abundance of taxa based on Linear

Discriminant Analysis Effect Size analysis.

102

103

Fig. 6 Heat map showing correlations between the rarefied relative abundance of microbiota and short

chain fatty acids (SCFA) and carbohydrates in caecum (A) and mid colon (B). The x-axis contains

taxa > 1% abundance and y-axis contains body weight, starch, individual and total non-starch

polysaccharides (NSP) content, relative percentage of glucose (Glu_NSP) and arabinoxylan

(AX_NSP) in total NSP, ratio of arabinose:xylose (A_X ratio), total and individual SCFA and

branched-chain fatty acids (BCFA) concentrations, pool size and relative percentage in SCFA

(Acetate_SCFA, Butyrate_SCFA, Propionate_SCFA and BCFA_SCFA) or the percentage in sum of

acetate, propionate and butyrate (Acetate_APB, Propionate_APB and Butyrate_APB). Yellow

asterisks indicate statistical significance (P < 0.05) by Pearson’s test.

104

Table S1 Compound-dependent LC-MS/MS parameters, declustering potential (DP), entrance

potential (EP), collision energy (CE) and cell exit potential (CXP).

Item Q1 mass Q3 mass DP EP CE CXP

Acetic acid quantifier 193.8 151.9 -80 -15 -15 -15

Acetic acid qualifier 193.8 46.0 -80 -15 -50 -15 13C2 acetic acid quantifier 196.0 152.0 -100 -15 -24 -13 13C2 acetic acid qualifier 196.0 121.8 -100 -15 -24 -10

Propionic acid quantifier 208.0 136.9 -20 -10 -24 -9

Propionic acid qualifier 208.0 46.0 -20 -10 -50 -20 13C1 propionic acid quantifier 209.0 137.0 -20 -10 -25 -9 13C1 propionic acid qualifier 209.0 165.1 -20 -10 -18 -11

Isobutyric acid quantifier 221.9 178.9 -100 -2 -18 -15

Isobutyric acid qualifier 221.9 42.0 -60 -2 -70 -19

Butyric acid quantifier 221.9 178.9 -100 -2 -18 -15

Butyric acid qualifier 221.9 42.0 -60 -2 -70 -19 13C2 butyric acid quantifier 224.0 205.8 -100 -10 -20 -12

13C2 butyric acid qualifier 224.0 180.0 -100 -10 -18 -15

Succinic acid quantifier 387.1 234.1 -30 -10 -25 -12 13C2 succinic acid qualifier 387.1 97.8 -30 -10 -47 -10 13C2 succinic acid quantifier 389.1 99.9 -145 -10 -40 -10

Succinic acid qualifier 389.1 236.1 -145 -10 -24 -15

Isovaleric acid quantifier 235.9 137.0 -15 -15 -30 -15

Isovaleric acid qualifier 235.9 46.0 -15 -15 -40 -15

Valeric acid quantifier 235.9 137.0 -15 -15 -30 -15

Valeric acid qualifier 235.9 46.0 -15 -15 -40 -15 13C3 valeric acid quantifier 239.0 136.9 -120 -10 -27 -15 13C3 valeric acid qualifier 239.0 151.8 -120 -10 -24 -12

105

Table S2 Ingredients and chemical composition of the experimental diets.

Item LOFLOP LOFHIP HIFLOP HIFHIP

Ingredients (g/kg as-fed basis)

Wheat starch 233 167 66 0

Whole grain wheat (roller milled) 150 150 150 150

Wheat bran (finely milled) 125 125 125 125

Enzyme-treated wheat bran - - 200 200

Wheat gluten 65 65 32 32

Fish meal 20 20 20 20

Whey protein hydrolysate - 68 - 68

Fructose 200 200 200 200

Lard 150 150 150 150

Vitamins and minerals 57 56 57 56

Chemical composition (g/kg DM)

DM (g/kg as-fed basis) 913 913 919 919

Ash 62 65 73 76

Crude protein (N × 6.25) 113 179 114 175

HCL fat 174 180 188 187

Available carbohydrates 577 522 456 387

Sugars

Fructose 225 223 224 221

Glucose 1 1 7 7

Sucrose 7 7 9 8

Starch 344 292 216 150

Dietary fibre1 100 106 191 205

NSP (soluble NSP) 69 (8) 75 (12) 136 (22) 136 (15)

AX (SAX) 44 (5) 46 (7) 86 (16) 85 (12)

AXOS 5 3 12 17

Fructans 6 8 11 9

Klason lignin 19 20 30 41

RS2 2 1 1 1

Gross energy (MJ/ kg DM) 20.7 21.5 21.3 21.7

LOFLOP, low fibre low protein diet; LOFHIP, low fibre high protein diet; HIFLOP, high fibre low

protein diet; HIFHIP, high fibre high protein diet. NSP, total non-starch polysaccharides; AX,

arabinoxylan; RS, resistant starch; A:X, arabinose:xylose; AXOS, low molecular weight

arabinoxylan‐oligosaccharides. 1Dietary fibre = NSP + fructans + RS + AXOS + Klason lignin. 2Determined by enzymatic resistant starch assay (AOAC method 2002.02).

106

Table S3 Pool size (mmol) of short chain fatty acid in caecum and entire colon of Göttingen Minipigs

fed diets low or high in dietary fibre and protein.

Item Diet1

SEM P-value2

LOFLOP LOFHIP HIFLOP HIFHIP F P F×P

Caecum

Total SCFA 9.1 8.1 9.7 13 1.6 0.08 0.29 0.13

Acetate 6.2 5.3 6.3 8.8 1.1 0.09 0.29 0.10

Propionate 1.6 1.5 1.4 2.3 0.31 0.29 0.09 0.08

Butyrate 1.1 0.97 1.4 1.6 0.25 0.05 0.80 0.37

BCFA 0.11 0.11 0.08 0.11 0.02 0.28 0.26 0.33

Entire colon

Total SCFA 29b 27b 29b 50a 5.0 0.02 0.05 0.02

Acetate 18b 17b 18b 30a 3.0 0.03 0.06 0.02

Propionate 5.0 4.7 5.9 10 1.1 0.007 0.09 0.06

Butyrate 4.5 4.3 5.7 7.6 1.2 0.04 0.41 0.32

BCFA 0.68 0.73 0.52 0.98 0.12 0.68 0.03 0.07 1The individual minipig was regarded as the experimental unit, n = 10 for low fibre low protein diet

(LOFLOP), n = 10 for low fibre high protein diet (LOFHIP), n = 12 for high fibre low protein diet

(HIFLOP) and n = 11 for high fibre high protein diet (HIFHIP). Data were expressed as means with

standard error of means (SEM). 2F, fibre effect; P, protein effect; F × P, fibre × protein interaction. Different superscript letters in a

row are presented for the significant interaction (P < 0.05) after adjustment for multiple comparisons

by the Tukey–Kramer post hoc test.

107

Fig. S1 The representative MRM chromatogram of SCFA standards.

108

Total S

CFA

Ace

tate

Propio

nate

Butyra

te

BCFA

0

30

60

90C

on

cen

trat

ion

, mm

ol/

kg LOFLOP

LOFHIP

HIFLOP

HIFHIP

Fiber, P = 0.001

A

Acetate Propionate Butyrate BCFA

0

20

40

60

80

% o

f to

al S

CF

A

LOFLOPLOFHIP

HIFLOP

HIFHIP

Fiber, P < 0.001

Fiber, P < 0.001

B

Fig. S2 Concentrations of total SCFA, acetate, propionate, butyrate and BCFA (A) and distribution

of total SCFA (B) in faeces of Göttingen Minipigs. Values are least-squares means with standard

errors represented by vertical bars. The individual minipig was regarded as the experimental unit, n

= 10 for low fibre low protein diet (LOFLOP), n = 10 for low fibre high protein diet (LOFHIP), n =

12 for high fibre low protein diet (HIFLOP) and n = 11 for high fibre high protein diet (HIFHIP).

Only significant P -values are presented in the figure.

109

128

6857

3586

02

1146

9

1433

7

1720

4

2007

1

2293

8

2580

6

0

10

20

30

Rerafraction depth

fait

h_p

d

LOFHIP

LOFLOP

HIFLOP

HIFHIP

Fiber, P = 0.02

C

128

1456

2784

41

1125

4

1406

8

1688

1

1969

5

2250

8

2532

2

0

5

10

15

20

25

Rerafraction depth

fait

h_p

d

LOFLOPLOFHIPHIFLOPHIFHIP

A

135

3270

64

1059

6

1412

8

1765

9

2119

1

2472

3

2825

5

3178

7

0

5

10

15

20

25

Rerafraction depth

fait

h_p

d

LOFLOPLOFHIPHIFLOPHIFHIP

B

Fig. S3 Alpha rarefaction curves for each diet, showing the Faith’s phylogenetic diversity (faith_ph)

(y axis) as a function of sequencing depth (x axis). Samples were rarefied to 25322 sequences for

caecum (A), 31787 sequences for mid colon (B) and 25806 for faecal samples (C), which means the

minimum sequence per sample depth in the dataset. The individual minipig was regarded as the

experimental unit, n = 10 for low fibre low protein diet (LOFLOP), n = 10 for low fibre high protein

diet (LOFHIP), n = 12 for high fibre low protein diet (HIFLOP) and n = 11 for high fibre high protein

diet (HIFHIP). One digesta sample from mid colon in LOFLOP group (n = 9) was omitted due to

low-quality reads. Only significant P -values are presented in the figure.

110

C

BA

Fig. S4 Principal coordinate analysis plot of the weighted Unifrac metric. The grey, yellow, red and

blue line with points represented caecum (A), mid colon (B) and faecal (C) communities of individual

minipig after feeding low fibre low protein diet (LOFLOP, n = 10), low fibre high protein diet

(LOFHIP, n = 10), high fibre low protein diet (HIFLOP, n = 12) and high fibre high protein diet

(HIFHIP, n = 11), respectively. One digesta sample from mid colon in LOFLOP group (n = 9) was

omitted due to low-quality reads. Only significant fibre effect was observed in principal coordinate

analysis plot of caecum, colon and faecal communities (P < 0.01).

111

Fig. S5 Identification of the most differentially abundant genera in caecum (A), mid colon (B) and

faecal samples (C) in Göttingen Minipigs fed low or high fibre and protein diets. The plot was

generated from Linear Discriminant Analysis Effect Size (LEfSe) analysis with CSS-normalized

OTU table and displays taxa with LDA scores above 2.0 and P-values below 0.01. Genera enriched

in the samples with low fibre (LOF) or low protein (LOP) are indicated with green bars, and genera

enriched in the samples with high fibre (HIF) or high protein (HIP) are indicated with a red bars. The

individual minipig was regarded as the experimental unit, n = 20 for LOF, n = 23 for HIF, n = 22 for

LOP and n = 21 for HIP. One digesta sample from mid colon in LOFLOP group (n = 19 for LOF)

was omitted due to low-quality reads.

112

6.3 Paper III

The role of rye bran and antibiotics on the digestion, fermentation process and short-chain

fatty acid production and absorption of pigs.

Yetong Xu, Anne Katrine Bolvig Sørensen, Brendan McCarthy-Sinclair, Maria L. Marco, Knud Erik

Bach Knudsen, Mette Skou Hedemann and Helle Nygaard Lærke

Manuscript under review in British Journal of Nutrition. July, 2020.

113

The role of rye bran and antibiotics on the digestion, fermentation process and short-chain fatty 1

acid production and absorption of pigs 2

Yetong Xu1*, Anne Katrine Bolvig1, Brendan McCarthy-Sinclair2, Maria L. Marco2, Knud Erik Bach 3

Knudsen1, Mette Skou Hedemann1 and Helle Nygaard Lærke1 4

1 Department of Animal Science, Aarhus University, DK-8830 Tjele, Denmark 5

2 Department of Food Science and Technology, University of California, Davis, CA, USA 6

* Corresponding Author: Yetong Xu; Email: yetongxu@anis.au.dk; Tel.: +4550641816 7

Short title: DF and antibiotics on degradation and SCFA 8

Keywords: dietary fibre; antibiotics; degradation; SCFA; pig model 9

114

ABSTRACT 10

The effects of arabinoxylan (AX)-rich rye bran based diet (RB) and antibiotics on digestion, 11

fermentation and short-chain fatty acids (SCFA) absorption were studied compared with an iso-DF 12

cellulose based diet (Control). Thirty female pigs (body weight 72.5 ± 3.9 kg) were fed a standard 13

swine diet in week 1 and the Control as wash-out in week 2, then divided into 3 groups fed either the 14

Control (n=10) or RB (n=20) for 2 weeks, where 10 pigs from RB had daily intramuscular antibiotic 15

injections in week 4. The degradation of AX in the RB groups mainly occurred in caecum and 16

proximal colon (P < 0.01). Cellulose was degraded to a lower degree than AX but to a higher degree 17

in the Control than RB groups (P < 0.01). The colonic concentrations of SCFA, acetate and propionate 18

were lower with the RB diets compared with Control (P < 0.01), whereas the RB diets increased 19

caecal butyrate concentration with a reduction by the antibiotic treatment (P < 0.001). The apparent 20

protein digestibility in the RB groups was lower than Control except for distal colon and gradually 21

increased throughout the large intestine (P < 0.05). Portal plasma concentrations of SCFA and acetate 22

were lower with the RB diet than Control, accompanied by lower daily net absorption (P < 0.01). In 23

conclusion, the RB diet resulted in different DF degradation processes and SCFA production than the 24

Control diet, whereas antibiotic treatment had marginal effects on the intestinal DF structure but 25

inhibited butyrate production.26

115

INTRODUCTION 27

Dietary fibre (DF), the sum of non-digestible carbohydrate and lignin, has been associated with 28

decreased risk for obesity, cardiovascular disease, cancer and type 2 diabetes(1). In most Northern 29

European countries, cereal products are the important sources of DF. Whole grains, especially the 30

bran fractions, are high in DF and other phytochemicals. It has been shown that the presence of DF 31

with different structural composition and physicochemical properties may have different influence 32

on the transit time and nutrient digestibility along the gastrointestinal tract(2). In the small intestine, 33

the luminal viscosity and water binding capacity of digesta are increased by soluble cereal DF, 34

thereby slowing the movement of digesta(3) whereas in the large intestine, DF are fermented by 35

anaerobic microbiota into multiple metabolites such as short chain fatty acids (SCFA)(2, 4). Only about 36

5% of the produced SCFA is excreted in the faeces while some are metabolised within the gut itself 37

but the majority is absorbed and pass via the portal vein to the liver where it is either metabolised or 38

goes into peripheral circulation(5). Butyrate, in particular, is an important energy source for the 39

intestinal epithelium and can improve gut barrier function and alleviate inflammation(6). Besides the 40

beneficial effects of SCFA on intestinal health, increased peripheral SCFA appear to play a critical 41

role in inhibiting lipolysis within adipose tissue(7), improving insulin secretion and glucose 42

homeostasis(8). 43

The main DF constituents of cereals are arabinoxylan (AX), β-glucan, cellulose and lignin present 44

in different proportions in different tissues of the grain(9). Compared to wheat, rye bran (RB) has a 45

higher content of AX, fructans as well as β-glucans(2, 10). The major non-starch polysaccharides (NSP) 46

components in rye bran, AX, consists of xylan backbones with varying degree of substitution with 47

arabinose and ferulic acid residues(11). It has been found that a high degree of substitution makes the 48

polymer less degradable(12). The type and composition of DF affects the composition of intestinal 49

microbiota and profile of metabolites formed during fermentation, especially SCFA(13). It has been 50

demonstrated that AX is a substrate for intestinal butyrate production as well as enrichment of 51

butyrogenic species(14). However, the microbial fermentation processes are likely to be disturbed by 52

antibiotics. The use of antibiotics has been associated with increased metabolic impairments, such as 53

disrupting intestinal homeostasis and integrity of intestinal defenses(15). A previous study has found 54

that antibiotics decreased the caecal concentrations of SCFA accompanied by the change of 55

microbiota composition in mice(16). It also corroborates with an epidemiological study showing that 56

the use of antibiotics reduced the plasma levels of enterolactone, a metabolite of lignans that require 57

116

microbial conversion(17). This was confirmed in our pig study, where we also found that antibiotic 58

treatment affected the microbiota composition and urine metabolome(18). 59

In this paper, we investigated the effects of a rye bran enriched diet high in AX with or without 60

antibiotics on nutrient digestibility, NSP degradation, SCFA production and absorption compared to 61

a diet with similar DF content, but where the DF stems from purified cellulose only. The use of a pig 62

model allowed control of DF intake and antibiotic dosage and to sample at sites not possible in 63

humans(18). We hypothesised that an AX-enriched diet will reduce nutrient digestibility but stimulate 64

fermentation processes, and specifically increase butyrate production and absorption compared with 65

refined wheat fibre, whereas antibiotic treatment will hamper fibre degradation and the production of 66

SCFA. 67

MATERIALS AND METHODS 68

Diets, animals, experimental design and microbiota assessments have been described in detail 69

previously in a paper presenting results on plant- and enterolignan concentrations and LC−MS based 70

metabolomics(18). 71

Experimental diets 72

Two experimental diets with the same DF level were used: a refined wheat fibre added diet (Control) 73

and a rye bran added diet (RB). The Control diet was formulated by white wheat flour supplemented 74

with purified cellulose-rich and lignin-free wheat fibre (Vitacel WF600, J. Rettenmeier and Söhne 75

Gmbh, Rosenberg, Germany), while AX-rich rye bran was used in the RB diet to replace purified 76

wheat fibre and part of the wheat flour (Table 1 and Table S1). The two dietary treatments were 77

balanced with regard to energy, starch, protein, fat and DF, but varied in DF composition. Both diets 78

were added with chromic oxide (Cr2O3), mixed in a feed production unit and cold-pressed into pellets. 79

Animals and handling 80

The animal experiment was performed in accordance with protocols approved by the Danish Animal 81

Experiments Inspectorate, Ministry of Food, Agriculture and Fisheries and in compliance with the 82

guidelines regarding to animal experiments and care of animals under study. Throughout the 83

experimental period, general health of the animals was monitored and no serious illness was observed. 84

Thirty Duroc × Danish Landrace × Yorkshire female pigs (body weight 72.5 ± 3.9 kg) purchased 85

from a local farmer and raised at Aarhus University, Department of Animal Science, Foulum, 86

Denmark were included in this study in two blocks for four consecutive weeks. All pigs were housed 87

individually and fed a standard swine diet during the first week followed by a wash-out period with 88

117

the Control diet for the second week, which aimed to adjust the baseline and to wash out the lignans 89

as required in our previous study(18). In the third and fourth week, 10 pigs continued with the Control 90

diet and 20 pigs switched to the RB diet. In the fourth week, 10 of the RB fed pigs (RB+) had daily 91

antibiotic injections Streptopenprokain Rosco Vet (Boehringer Ingelheim, Copenhagen, Denmark), 92

(benzylpenicillinprocain and dihydrosreptomycin) in connection with the morning meal, while other 93

10 pigs were left untreated (RB-)(18, 19). All pigs had ad libitum access to water and were fed manually 94

three times daily with equal meal size throughout the experimental period, at 9.00, 14.00 and 19.00 95

hours to mimic a typical human meal pattern. The total daily feed allowance for the pigs was 3% of 96

average body weight and adjusted every week based on body weight. 97

Slaughtering and sample collection 98

At the end of experiment, the pigs were anaesthetised with Zolitil mixture 3 h after the morning meal 99

as described before(18, 20). Blood samples from carotid artery were collected using heparinized 100

vacutainers by venipuncture, and the abdominal cavity was opened to collect blood samples from the 101

portal and hepatic veins(18). Immediately after, the pigs were euthanised and the entire gastrointestinal 102

tract was ligated at the esophagus and rectum, removed from the carcass. Both the small intestine and 103

colon were divided into three parts with equal length (from proximal to distal). The distal third of the 104

small intestine (SI3), caecum, proximal (Co1), mid (Co2) and distal colon (Co3) of pigs were tied off 105

and wet digesta were weighted and pH was measured. Blood samples were centrifuged for 12 min at 106

2000 g at 4°C and plasma were obtained and immediately frozen at -80°C for SCFA analysis. 107

Collected digesta were kept in a -80°C freezer for dry matter (DM) and carbohydrate analysis. 108

Samples of digesta (1-3 g) from SI3, caecum, Co1, Co2 and Co3 were collected for the analysis of 109

SCFA. 110

Analytical methods 111

All chemical analysis of the diets and digesta were performed in duplicate on freeze-dried material. 112

DM was determined by drying to constant weight at 103°C, and ash was analysed according to the 113

AOAC method no. 942.05(21). Nitrogen (N) was measured by Dumas(22) and protein was calculated 114

as N × 6.25. Gross energy was analysed on a 6300 Automatic Isoperibol Calorimeter system (Parr 115

Instruments) and fat determined using the Stoldt procedure(23). The dietary content of fructans and 116

sugars (glucose, fructose and sucrose) was determined as described previously(24). The content of 117

starch, low molecular weight (LMW) non-digestible carbohydrate (NDC) and NSP of diets and 118

digesta were measured as described by Bach Knudsen(9) and Kasprzak et al(25). Klason lignin of the 119

118

diets was measured as the sulfuric acid-insoluble residue as described previously(26). Chromic oxide 120

of diets and digesta was determined colorimetrically using the procedure described by Schürch et 121

al(27). 122

For the SCFA analysis, digesta and faecal samples were subjected to an acid–base treatment 123

followed by diethyl ether extraction and derivatisation as previously described(28), and SCFA 124

concentrations analysed by GC (HP-6890 Series Gas Chromatograph; Hewlett Packard) using an HP-125

5 column (30 m × 0.32 mm × 0.25 µm) and a flame ionisation detector(14). Plasma SCFA was 126

measured by GC as described by Brighenti(29) with 2-ethyl butyrate was used as an internal standard 127

instead of isovalerate(8). 128

Calculations 129

The apparent digestibility of nutrients and NSP in the digesta from intestinal segments were 130

calculated relative to Cr2O3 concentrations according to the following equation: 131

Digestibility of X = 1- Cr2O3(Diet) × X(Dig)

Cr2O3(Dig) × X(Diet) , (1)

where X(Diet) and X(Dig) are the concentrations of the macronutrients or NSP component 132

determined in the diets and digesta, respectively, and Cr2O3(Diet) and Cr2O3(Dig) are the Cr2O3 133

concentrations in the diets and digesta. 134

Mean transit time (MTT) in the intestinal segments was calculated by the following equation: 135

MTT = Cr2O3(Dig) × 24

Cr2O3(day) , (2) 136

where Cr2O3(Dig) and Cr2O3(day) is the Cr2O3 concentrations of digesta and Cr2O3 daily intake, 137

respectively. 138

The net absorption of SCFA into the portal vein was calculated by portal-arterial differences and 139

portal flow measurements using Fick’s principle(30) according to the following equation: 140

Q = (cp – ca)F, (3) 141

where cp and ca denotes the concentrations of SCFA in the portal vein and carotid artery, 142

respectively, and F is the plasma flow in the portal vein which was set as a fixed rate of 31.8 mL/kg 143

per min(31). The daily net absorption was calculated by multiplying the minute absorption by 1440 to 144

give 24 hours’ net absorption. 145

Hepatic extraction (HE) was determined as: 146

HE = (-) (1.00 × Ch) - (0.86 × Cp + 0.14 × Ca)

0.86 × Cp + 0.14 × Ca × 100, (4) 147

119

where ch is the SCFA concentration in hepatic vein and ca is the SCFA concentration in the hepatic 148

artery. The values 0.86 and 0.14 were used for the relative contributions of portal and hepatic arterial 149

plasma flow to hepatic venous plasma flow, respectively(32). 150

Total SCFA concentration was calculated as the sum of formic acid, acetic acid, propionic acid, 151

butyric acid, isobutyric acid, isovaleric acid and valeric acid concentrations, and BCFA concentration 152

was calculated as the sum of the isobutyric acid, isovaleric acid and isocaproic acid concentrations. 153

The pool size of SCFA and BCFA in the entire large intestine was calculated as the sum of the 154

concentrations of SCFA multiplied by the amount of wet digesta in each of the intestinal segments. 155

The SCFA distribution (%) in digesta was calculated as the concentration of the individual SCFA 156

divided by the total SCFA concentration in the digesta and multiplied by 100. 157

Statistical analysis 158

According to the power calculations for pool size of total and individual SCFA in the large 159

intestine(14), 10 pigs per group were demonstrated to give sufficient statistical power (α < 0.05; β > 160

0.80). Two types of analysis on data were conducted in the present study. The first type of analysis 161

was performed using the MIXED procedure of SAS (SAS Institute, Inc.) followed by a pairwise 162

comparison of groups adjusted by multiple comparisons of Tukey–Kramer post hoc test. The effects 163

of diet on the apparent digestibility of DM, macronutrients and NSP components, A:X ratio, and net 164

amount, MTT, pH value and SCFA profile of gut content and plasma SCFA concentrations were 165

analysed by the following mixed model: 166

Yijkl = μ + αi + βj + γk + (αγ)ik + vl + εijkl, (4)

where Yijkl is the dependent variable; μ is the overall mean; αi is the treatment (i = Control, RB+ 167

or RB-); βj is the random effect of block (l = 1 or 2); γk is the intestinal segment (k = SI3, caecum, 168

Co1, Co2 or Co3) or blood sampling site (k = portal vein, hepatic vein or carotid artery); (αγ)ik is the 169

interaction between diet and intestinal segment or blood sampling site; vl is the random component 170

related to the pig (m = 1, 2,…, 30). Pig was included as a random component to account for repeated 171

measurements within pigs. The covariance structure of repeated measurements was modelled using 172

autoregressive type and εijkl is the residual error. The random effect and residuals were assumed to be 173

normally distributed and independent, and their expectations were assumed to be zero. 174

The second type of analysis compared the effects of dietary treatments on ileal carbohydrate 175

composition, total net amount, total MTT and pool size of SCFA in the large intestine (caecum + 176

colon). The analysis was accomplished by using a simple ANOVA-based model: 177

120

Yij = μ + αi + βj + εij (5)

where Yij is the dependent variable, μ is the overall mean; αi is the fixed effect of treatment (i = 178

Control, RB+ or RB-); βj is the random effect of block (j = 1 or 2); εij is the residual error. Variables 179

with normal distribution were presented as least square means with standard error of the mean (SEM). 180

A log-transformation was performed for plasma propionate, butyrate and BCFA concentrations to 181

obtain variance homogeneity and results are presented as geometric means with 95% confidence 182

intervals. Level of significance was reported as being significant when P < 0.05. Pearson correlation 183

analyses between the SCFA absorption and plasma lipid concentration were performed by GraphPad 184

Prism 8.0 (GraphPad Software Inc.) using data from this study and our previous publication (Bolvig 185

et al. 2017, S1)(18). Heat map correlation analyses between caecal carbohydrates, SCFA and taxa were 186

performed by R Studio (R, Version 1.1.453) using Operational taxonomic unit (OTU) data from our 187

previous study(18). 188

RESULTS 189

Dietary composition and body weight 190

The two experimental diets were formulated to provide equal amounts of protein, fat, DF and energy 191

but different composition of DF, which was successfully achieved (Table 1). The RB diet contained 192

twice as much soluble NSP (25 vs. 13 g/kg) and total AX (86 vs. 37 g/kg) compared with the Control 193

diet, whereas the cellulose content was much lower (14 vs. 90 g/kg). AX and cellulose contributed 194

with 64% and 63 % of total NSP in RB and Control, respectively. The A:X ratio of total AX was 195

higher in RB (0.55 vs. 0.22) with lower A:X ratio of soluble AX (0.56 vs. 0.95) compared with the 196

Control. The amounts of fructans and lignin were 7.7 and 27 g/kg DM in RB, 0.4 and 19 g/kg DM in 197

Control, respectively. 198

The average body weight of pigs were 95.1 ± 2.8 kg at the end of the experiment with no significant 199

change between dietary treatments (P = 0.621, data not shown). 200

Carbohydrate composition of ileal digesta 201

For pigs fed the RB diet, the concentrations of NDC and insoluble NSP in ileal digesta were 202

significantly lower than for pigs fed the Control diet (P < 0.001), whereas the content of soluble NSP 203

was significantly higher (P < 0.001, Table 2). AX was the main contributor to the total NDC fraction 204

of digesta in the RB groups accounting for 60%, whereas cellulose was the main NDC fraction in 205

ileal digesta of pigs fed Control contributing with 63%. The LMW NDC content in ileal digesta 206

accounted for 7% and 14% of total NDC for the Control and RB diets, respectively, with a 207

121

significantly higher LMW NDC in the RB+ group than the Control (P = 0.03). The xylose content, 208

however, was significantly lower with the RB diet than the Control (P < 0.01). 209

The A:X ratio of total AX in RB (0.48) was not affect by antibiotics (RB+ vs. RB-, P = 0.99), and 210

was similar in the insoluble and soluble fraction (0.53-0.55). This was in contrast to the Control diet, 211

where the A:X ratio was high in the soluble fraction (0.85) and low in the insoluble (0.09), with an 212

A:X ratio of 0.15 in total AX; significantly lower than of the RB groups (P < 0.01). 213

Apparent digestibility of nutrients and fermentation of NSP components 214

RB+ had a significantly lower apparent DM digestibility in SI3 compared to RB- (P = 0.008) and the 215

Control (P = 0.01), whereas no significant differences were seen between the treatments in DM 216

digestibility in the large intestine (P > 0.10, Table 3). The apparent digestibility of crude protein was 217

significantly lower in the RB groups than the Control group in all intestinal segments (P < 0.05) with 218

no effect of antibiotic treatment (P > 0.10) except in Co3 where RB+ had lower digestibility (P = 219

0.03). For the RB diets there was a gradual increase in protein digestibility throughout the colon (P < 220

0.05), whereas in the Control group, the digestibility increased 10 % units from SI3 to ceacum (P < 221

0.001) and then remained constant throughout the colon (P > 0.10). The apparent digestibility of 222

crude fat was also lower in SI3 of the RB fed pigs than the Control (P < 0.05) without significant 223

effect of antibiotics (RB+ vs. RB-, P = 0.75). Starch was highly digested (> 0.97) in SI3 with no 224

difference between the treatments (P > 0.10), and increased to 0.99 in the caecum (P < 0.001). There 225

was no significant effect of antibiotic treatment (RB+ vs. RB-) on the digestibility of NSP or its 226

components, cellulose and AX, in either SI3, the caecum or colon (P > 0.10). The digestibility of 227

NSP of the RB groups was significantly higher than Control in SI3 (P < 0.001), caecum (P < 0.01) 228

and Co1 (P < 0.001), but the difference disappeared in Co2 (P > 0.05) and Co3 (P > 0.10). The 229

digestibility of cellulose was generally lower in RB than in Control, with a significant difference 230

between RB+ and Control in caecum (P = 0.01) and Co2 (P = 0.02), and for both RB groups in Co3 231

(P < 0.001). Within each segment, the digestibility of AX in RB did not deviate significantly from 232

the Control, except for a 15% reduced digestibility in Co3 that was significant for RB+ (RB+ vs. 233

Control, P = 0.004) and a tendency for RB- (RB- vs. Control, P = 0.06). The apparent digestibility of 234

DM, NSP, cellulose and AX remained constant throughout the colon in the RB fed pigs, whereas 235

increased digestibility was shown in Co3 with the Control diet (P < 0.01). In SI3 of the RB fed pigs, 236

the A:X ratio was similar to the ratio in the diets, but increased significantly in caecum and further to 237

> 1.0 in the colon (P < 0.001) without significant effect of antibiotic treatment (P > 0.10). In all 238

122

segments, this was significantly different from the Control (P < 0.001), where the A:X ratio was < 239

0.2 and hardly changed along the large intestine (P > 0.10). 240

Intestinal bulk and mean transit time 241

The net amount of wet digesta in SI3, caecum, Co2 and Co3 was not affected by dietary treatments 242

(P > 0.10), and it did not differ between SI3 and caecum. The amount was significantly higher for the 243

Control than the RB diet in Co1 (P < 0.001), and significantly reduced from Co1 to the remaining 244

colon segments (P < 0.001, Fig. 1(A)). 245

The MTT was on average 2.8 h in SI3 for all treatments, while the total MTT in the entire large 246

intestine (caecum + colon) was 19 h for the RB groups and 23 h for the Control group (P < 0.01, Fig. 247

1(B)) with no effct of antibiotics (RB- vs. RB+, P = 0.99). The MTT of the RB+ group was 248

significantly lower in Co2 compared with the Control (P = 0.02) and both RB groups had a 249

significantly shorter MTT in Co3 (P < 0.01). The MTT in each segment increased gradually along 250

the length of the large intestine of the Control group (P < 0.01), whereas for RB groups the MTT was 251

about the same in the three colon segments (P > 0.10). 252

Intestinal pH and SCFA parameters 253

The pH dropped from SI3 to caecum for all treatments (P < 0.001) but with no difference between 254

treatments (P > 0.10, Fig. 2). In the colon, the RB groups deviated significantly from the Control (P 255

< 0.001) as pH increased along the colon (P < 0.01), whereas pH was not changed throughout the 256

colon for pigs fed the Control diet (P > 0.10). 257

The concentrations of individual and SCFA were affected differently by diets and antibiotic 258

treatment depending on sites of the gastrointestinal tract (Table 4). The concentrations of SCFA, 259

acetate and propionate in the RB fed pigs were significantly lower than the Control feed pigs (P < 260

0.05) without any effect of antibiotic treatment (RB+ vs. RB-, P > 0.10). Moreover, the concentrations 261

increased from SI3 to caecum for all treatments and did not change in the colon with the Control diet 262

but was reduced in the distal colon with the RB diet (P < 0.01). The RB diet significantly increased 263

butyrate concentrations in caecum compared with the Control diet (P < 0.01). Antibiotic treatment 264

significantly reduced the butyrate concentration (RB+ vs. RB-, P < 0.001). The highest butyrate 265

concentration was found in caecum of RB- which was reduced to half in Co3 (P < 0.001), whereas 266

for the Control, no significant change was seen throughout the colon (P > 0.10). In Co3 the butyrate 267

concentration was higher for the Control group than the RB groups (P < 0.001). RB feeding without 268

antibiotic treatment increased the concentration of BCFA compared to the Control diet in Co1 (P < 269

123

0.001), and BCFA concentration in Co2 and Co3 was higher with the RB diet than the Control (P < 270

0.01). For the Control diet, BCFA concentration was the same in caecum and colon (P > 0.10) 271

whereas BCFA concentration in the RB groups was significantly higher in all colonic segments than 272

SI3 and caecum (P < 0.05). The general pattern of the SCFA and BCFA concentrations were basically 273

also reflected in the pool size (Table S2), with lower total SCFA, acetate, propionate and BCFA pool 274

size with the RB diet than the Control diet in all colonic segments and higher butyrate pool size in 275

caecum with the RB- group than RB+ group (P = 0.04). 276

There was a strong interaction between treatment and segment in the contribution of acetate, 277

butyrate and BCFA to SCFA (P < 0.001), whereas the proportion of propionate was only affected by 278

segment (P < 0.001) and not treatment (P > 0.10, Fig. 3). The proportion of acetate did not 279

significantly change along the intestinal segments for the RB groups (P > 0.10, Fig. 3(A)), but was 280

lower in caecum and Co1 than for the Control group (P < 0.01). The proportion of acetate increased 281

significantly from SI3 to the caecum (P < 0.001) of the Control fed pigs and was with a gradual 282

decrease throughout the colon; Co3 was significantly different to the caecum (P < 0.01). The 283

proportion of propionate was very low in SI3 (1.8%), increased to approximate 22% in the caecum 284

(P < 0.001) and then remained almost constant in the colon with no difference between treatments (P 285

> 0.10, Fig. 3(B)). The proportion of butyrate in the intestinal segments differed between the RB 286

groups and the Control with a significant difference in caecum, Co1 and Co2 (P < 0.001). Antibiotic 287

treatment had a negative influence on the proportion of butyrate being lower in caecum (P < 0.001) 288

but with no difference in the colon segments (P > 0.10). The proportion of BCFA increased steadily 289

throughout the large intestine from zero in SI3 to approximately 5% in Co3 in the pigs fed RB diet 290

(P < 0.05) and with significant effects of antibiotics in any of the segments (P > 0.10). For the Control 291

diet the proportion of BCFA was almost constant throughout the colon (P > 0.10), and significantly 292

lower than the RB groups in Co2 and Co3 (P < 0.001). 293

Correlations between carbohydrates, SCFA and microbiota in caecum 294

The concentrations and proportions of butyrate of total SCFA were positively while the acetate 295

concentrations and proportions were negatively correlated with the caecal non-cellulosic (NCP) 296

fraction particularly AX (Fig. 4). Conversely, the acetate concentrations and proportions of total 297

SCFA were positively correlated, while butyrate concentrations and proportions of total SCFA were 298

negatively correlated with the contents of DM, NSP and cellulose and cellulose proportions of NSP 299

in the caecal digesta. The proportions of propionate were positively correlated with mannose which 300

had negative correlations with acetate proportions. Caecal pH and proportions of SCFA as BCFA 301

124

were negatively correlated with starch and NCPglucose contents in the caecum. On the other hand, 302

caecal starch and NCPglucose contents were positively correlated with the SCFA concentrations and 303

NCPglucose was also significantly correlated with the acetate concentrations. 304

Caecal propionate concentrations were positively correlated with Actinobacteria, Bacteroidales, 305

Gammaproteinbacteria, Erysipelotrichaceae, Coriobacteriaceae, Prevotella and Blautia, and the 306

propionate pools were negatively correlated with Coprococcus (Fig. 5). Total SCFA concentrations 307

showed positive correlations with Ruminicoccaceae and Ruminococcus. The NCPglucose contents were 308

positively correlated with Firmucutes, Clostidiales, Clostridiaceae and Ruminicoccaceae, and 309

negatively correlated with Bacilli and Lactobacillales. Caecal NSP, cellulose and DM contents were 310

negatively correlated while NCP fractions, butyrate concentrations and proportions were positively 311

correlated with Lactobacillales, Bacilli, Streptococcaceae, Streptococcus and Faecalibacterium. 312

Among them except Faecalibacterium were also negatively correlated with acetate concentrations 313

and proportions of SCFA that enriched in the Control diet but positively correlated with rhamnose 314

contents and butyrate pools that enriched in the RB diets. Conversely, Peptococcaceae and 315

Peptococcus correlated positively with total NSP, cellulose and DM, but correlated negatively with 316

the NCP fractions, A:X ratios and caecal butyrate concentrations and proportions that were found 317

higher in the RB groups. Butyrate concentrations were also negatively correlated with 318

Proteobacteria, Gammaproteinbacteria, Clostridiales, and Clostridium, and similar taxa including 319

Clostidiales, Clostridiaceae, Ruminicoccaceae, Turicibacter, Ruminococcus and Clostridium showed 320

negative correlations with the proportions of BCFA and butyrate as well as NCP fractions and A:X 321

ratios. On the other hand, Lactobacillales, Streptococcaceae and Streptococcus were positively 322

correlated with BCFA proportions of SCFA and except for Lactobacillales with BCFA pools. 323

Plasma levels, absorption and hepatic extraction of SCFA and relation to plasma lipids 324

The SCFA concentrations at different sampling sites, daily net absorption and hepatic extraction of 325

SCFA are shown in Table 5. Acetate was the dominating plasma SCFA, accounting for 67-74% of 326

total SCFA in the portal vein and 94-97% in the hepatic vein and carotid artery. The concentrations 327

of SCFA and acetate were lower in the hepatic vein and carotid artery than the portal vein, and were 328

significantly different between hepatic and carotid vein for the Control group (P < 0.001). Compared 329

to the RB groups, the Control group showed significantly higher SCFA and acetate concentrations in 330

portal vein (P < 0.001). The propionate and butyrate concentrations at all sampling sites were not 331

affected by treatments (P > 0.10), and the BCFA concentrations were higher only with the RB+ 332

125

compared with Control (P = 0.04). Overall, propionate, butyrate and BCFA concentrations were 333

highest in the portal vein, intermediate in the hepatic vein, and lowest in the carotid artery (P < 0.001). 334

The calculated daily absorption of SCFA and acetate was lower for RB diets compared to the 335

Control diet (P < 0.05), whereas the lower absorption of propionate was only significant in the RB- 336

group (P = 0.02). The absorption of butyrate and hepatic extraction of SCFA and BCFA was not 337

affected by treatments (P > 0.10). 338

Based on previously reported postprandial lipid levels of Bolvig et al.(18) we found a positively 339

correlated (r > 0.44, P < 0.05) between SCFA, acetate and propionate absorption and plasma 340

cholesterol and LDL concentrations (Fig. 6). The correlations of individual SCFA absorption and 341

triglyceride or HDL were insignificant as was the correlation between butyrate absorption and plasma 342

lipid levels (data not shown). 343

DISCUSSION 344

In the present study, the most noticeable difference between the two diets was the intended difference 345

in NSP composition; 64 % of NSP in the RB diet was in the form of soluble and insoluble AX, 346

whereas 63 % of NSP in the Control diet was cellulose deriving primarily from the added Vitacel 347

wheat fibre. The structure of AX was influenced by its origin being causative for the higher 348

substituted of AX in the RB diet than Control diet as indicated by the A:X ratios of rye bran and 349

Vitacel and also shown in other studies(2, 14). 350

The dietary difference in DF composition persisted in ileal digesta (SI3), where AX still 351

represented 63% of NSP in the RB groups and cellulose 69% in the Control group. The LMW 352

arabinose and xylose were released during passage of stomach and small intestine probably deriving 353

from the white wheat flour containing low amounts of arabinoxylan oligosaccharides (0.2%)(14). The 354

functional properties of AX are linked to the extent of substitution as reflected by A:X ratios (11) and 355

distribution of arabinose and ferulic acid along the xylan backbone(33, 34). For the RB groups, total AX 356

content in SI3 was distributed between soluble (21%) and insoluble (68%) fractions with similar A:X 357

ratios, which was almost unchanged compared to the diet. In the large intestine, the A:X ratio in RB 358

diets increased from SI3 to caecum and Co1 due to degradation of mostly unsubstituted AX. As a 359

consequence, the A:X ratio increased reflecting high resistance of highly substituted AX in rye(8, 12). 360

For the Control, the A:X ratio in diet was low in insoluble AX (0.09) and high in soluble AX (0.85) 361

which remained almost unchanged in ileal digesta. In the large intestine, however, the A:X ratio 362

decrease significantly between ileum and caecum and with hardly no change in the A:X ratio beyond 363

the caecum. This is most likely, caused by the AX in the wheat flour with higher A:X ratio being 364

126

degraded almost completely in caecum whereas the less substituted AX in Vitacel is rather resistant 365

to degradation. These data also illustrates that the ratio of A:X is a proxy not well correlated with AX 366

degradation(12, 34) as the ratio does not reveal differences in substitution pattern e.g. by blocks of 367

double substituted xylose units alternating with unsubstituted units or more evenly distributed mono-368

substituted chains of xylose(33, 35). 369

The ileal degradation of cellulose was in line with what has been reported earlier (-47% to 370

56%)(36). In the large intestine the degradation was lower for the RB diets than the Control diet 371

presumably reflecting a tight cross-linking between the unbranched AX segments, the cellulose 372

microfibrils and lignin(2, 4, 37). It is also for note that AX was fermented more proximal and cellulose 373

more distal in the large instestine as found in previous studies showing higher NSP degradation of 374

rye than wheat diets(2, 10). In the current study we found a similar degradation of NSP in the distal 375

colon for the Control group (58%) as found with similar diets types (59%)(2) but a lower NSP 376

degradation (57-60%) for RB diets compared with a rye aleurone flour diet (73%); the lower 377

degradation of NSP in rye bran than rye aleurone is due to the presence of highly lignified 378

pericarp/testa cell walls(12). 379

These protein and fat digestibility at SI3 was lower for the RB diets compared to the Control diet 380

probably because the higher soluble NSP content of the RB diets than the Control diet increased 381

luminal viscosity, which would impede the digestion and absorption of nutrients(38), hinder the 382

emulsification process in lipid digestion and hamper the reabsorption of the bile acids at the end of 383

the small intestine(8, 34). In the large intestine, crude protein was fermented at more distal segments 384

with the RB diets than the Control diet. The most likely explanation is a higher accessibility of the 385

residual dietary protein in the Control diet (residues from white wheat flour with thin fragile cell 386

walls, easily accessible wheat gluten and egg powder) whereas in RB diet, the majority of protein 387

residues come from rye bran, which has low accessibility due to its enclosure in thicker and more 388

rigid cell walls(39). 389

The transit time in the large intestine was significantly longer from Co2 and beyond for the 390

Control diet compared with the RB diets, which may provide more time for fermentation of DF, 391

permit greater SCFA absorption(40) and influence on gut microbiota composition(41). In previous 392

studies, it was concluded that resistant fibres increased faecal weight to a greater extent than more 393

fermentable fibres(42) and longer transit time with high cellulose wheat fiber diet than rye aleurone 394

diet was also reported(2). Although the degradation of NSP of Control and the RB diets was almost 395

127

the same, the higher lignin content and coarser particles of rye bran were the most likely reason for 396

the difference. 397

In the large intestine, the concentrations of total SCFA in the RB groups were lower compared 398

with the Control group and with a gradual decrease throughout all segments of large intestine as also 399

reflected in the pH values. It has earlier been found that AX-rich rye diet induced higher pH values 400

in colon compared with cellulose-rich wheat diet although the SCFA concentrations in colon did not 401

show a significant difference between the two diets(2). We have also previously found that the pH is 402

lowest in the caecum because of high fermentation and rises because of DF depletion throughout the 403

colon with an AX-enriched diet(2). For the Control group, however, the pH profile and SCFA 404

concentration were distinctly different compared with the RB groups. There was also a good 405

agreement between taxa enrichment and SCFA and/or NSP fractions that were significantly altered 406

depending on dietary treatments. For instance, the concentrations and proportions of SCFA in the 407

form of acetate in the caecum were positively correlated with the cellulose contents and abundance 408

of some bacteria (Clostridiales, Ruminococcaceae and Clostridium) that were found in higher 409

abundance with the Control diet(18). The concentrations and proportions of butyrate were positively 410

correlated with features of AX that has been found previously to stimulate butyrate production(14, 34). 411

In line with that, butyrate production and AX fractions in the caecum were positively correlated with 412

the taxa that were significantly enriched in the RB- group (Lactobacillales, Streptococcus and Bacilli) 413

but not in the RB+ group in the prior paper(18), indicating that antibiotic treatment affected butyrate 414

production by affecting the microbiota composition. This was also found by a previous study which 415

demonstrated that antibiotics induced depletion of anaerobic butyrate-producing bacteria, therefore 416

reduced butyrate levels, and was a risk factor for pathogen expansion in mice(43). Although 417

intramuscular injection instead of oral administration of antibiotics was performed in this study, we 418

found antibiotic metabolism also occurred in the gut(18). These observations confirmed our hypothesis 419

that the beneficial effects of AX on intestinal butyrate production would be negatively affected by 420

the use of antibiotics. Overall, the fermentation patterns were in line with earlier findings 421

demonstrating that high loads of cellulose resulted in stimulation of cellulytic bacteria and acetate 422

production(44) whereas fermentation of AX stimulated production of butyrate(10). Cross-feeding from 423

acetate to butyrate(45) probably also occurred, as the bacteria that were positively correlated with 424

butyrate proportions were negatively correlated with acetate proportions. Together with AX in the 425

RB diet, fucose, galactose and rhamnose showed important associations with butyrate production in 426

this study, which has also been found in previous studies(46, 47). 427

128

Although the main substrates for SCFA production in the large intestine are various forms of 428

exogenous carbohydrates, proteins, as well as endogenous sources also contribute(48). The higher 429

concentration and proportion of BCFA in the colon of pigs fed the RB diet compared to the Control 430

diet is well in line with the lower ileal to rectal disappearance of protein. This result confirms previous 431

results showing that consumption of an AX-enriched wheat bran diet increased the portal 432

concentration and flux of BCFA(8) as more protein reached the large intestine compared with a low 433

DF control diet (14). 434

The SCFA concentrations in artery, portal vein and hepatic vein reflect SCFA enrichment from 435

artery to portal vein because of SCFA production in the gut and hepatic clearance from portal vein to 436

hepatic vein(8). Not only the SCFA production in the gut, but also absorbed SCFA is dependent on 437

the fermentation substrate and colonic transit time(49). This is reflected by the higher acetate and 438

SCFA concentrations and productions in the large intestine and higher portal concentration and net 439

absorption when consuming the Control diet compared with the RB diets. The same trend was seen 440

in the hepatic vein and carotid artery although not statistically significant. The differences in SCFA 441

profile between the gut content, portal, hepatic and arterial plasma levels are a result of the different 442

SCFA being metabolised to different degrees at different sites; butyrate in colonocytes and liver, 443

propionate as gluconeogenic substrate primarily in the liver and acetate as substrate for the synthesis 444

of cholesterol in liver and triglycerides in adipose tissue(5). Hence, acetate was the major SCFA found 445

in peripheral circulation with concentrations of several hundred µM while the concentration of 446

propionate and butyrate are very low (below 15 µM)(50). The hepatic extraction of SCFA was not 447

affected by either diets or antibiotics. A constant hepatic extraction of metabolites irrespective of 448

dietary treatment has previously been reported, which was regarded as mass action driven(8). Studies 449

using catheterized pigs have shown that the fermentation of AX is associated with a net absorption 450

of butyrate(8, 10), which has been associated with improved insulin sensitivity in peripheral tissue(8, 51). 451

In contrast, the current study did not find elevated butyrate absorption or concentration in the blood 452

with the RB diet. This may suggest that the enhanced gut production of butyrate has direct effects on 453

the colonocytes as it is the primary energy source for epithelial cells(52) and after hepatic extraction 454

significant differences in the systemic circulation is hard to detect. Acetate has been reported as the 455

primary substrate for cholesterol synthesis(5), and pigs fed the Control diet showed not only increased 456

acetate absorption but also higher plasma concentrations of cholesterol and LDL compared with the 457

RB groups(18). As indicated by the Pearson correlations, increased total SCFA, acetate as well as 458

propionate absorption was associated with higher circulating cholesterol and LDL concentrations, 459

129

whereas the improved lipid profile in plasma with RB diet may partly be attributed to the lower SCFA 460

absorption, indicating an important role not only in gastrointestinal but also systemic health. 461

In conclusion, the present study showed that the AX-rich rye bran diet had a different DF 462

degradation and fermentation profile, and, as a consequence, also intestinal and systematic SCFA 463

profile compared with the wheat diet enriched with purified cellulose. The AX-enriched rye bran diet 464

had lower nutrient digestibility, but was efficient in producing butyrate by rapidly degrading AX in 465

the proximal part of the large intestine, but with no influence on either butyrate absorption nor 466

systematic circulation. The degradation of cellulose in the wheat fibre diet occurred more gradually 467

and distally in the large intestine, providing substrate for acetate production. The enhanced production 468

of total SCFA, acetate and propionate with the refined wheat fibre diet contributed to the increase in 469

net portal absorption, which may be associated with the synthesis of lipids in plasma. Corresponding 470

with our hypothesis, antibiotic treatment disrupted butyrate production plausibly associated with the 471

change of caecal microbiota profile. 472

ACKNOWLEDGEMENT 473

The authors thank Winnie Østergaard, Lisbeth Mӓrcher, Stina Greis Handberg, Thomas Rebsdorf and 474

Kasper Vrangstrup Poulsen for great assistance in planning and executing the animal experiment and 475

for excellent technical assistance in performing the analyses. The authors also thank the staff at the 476

animal facility. The present research was financially funded by Innovation Fund Denmark (Project 477

ELIN: The effects of enterolignans in chronic diseases: 060300580B). Y.X. acknowledges 478

scholarship from China Scholarship Community (201706350111). 479

FINANCIAL SUPPORT 480

This work was supported by grants from the Innovation Fund Denmark (Project ELIN: The effects 481

of enterolignans in chronic diseases: 060300580B). 482

CONFLICT OF INTEREST 483

The authors declare no conflict of interest. 484

AUTHORSHIP 485

K.E.B.K., H.N.L. and M.S.H. conceived and designed the research; A.K.B.S., H.N.L. and M.S.H. 486

performed the animal experiments; B.M.S. and M.L.M. performed metataxonomics analysis; Y. X. 487

analysed the data; Y. X. wrote the paper; All co-authors contributed to draft review. 488

REFERENCE 489

130

1. Ludwig DS, Pereira MA, Kroenke CH et al. (1999) Dietary fiber, weight gain, and cardiovascular 490

disease risk factors in young adults. Jama 282(16), 1539-1546. 491

2. Le Gall M, Serena A, Jorgensen H et al. (2009) The role of whole-wheat grain and wheat and rye 492

ingredients on the digestion and fermentation processes in the gut--a model experiment with pigs. 493

Br J Nutr 102(11), 1590-1600. 494

3. Ellis PR, Roberts FG, Low AG et al. (1995) The effect of high-molecular-weight guar gum on net 495

apparent glucose absorption and net apparent insulin and gastric inhibitory polypeptide production 496

in the growing pig: relationship to rheological changes in jejunal digesta. Br J Nutr 74(4), 539-497

556. 498

4. Glitsø LV, Brunsgaard G, Højsgaard S et al. (1998) Intestinal degradation in pigs of rye dietary 499

fibre with different structural characteristics. Br J Nutr 80(5), 457-468. 500

5. den Besten G, van Eunen K, Groen AK et al. (2013) The role of short-chain fatty acids in the 501

interplay between diet, gut microbiota, and host energy metabolism. J Lipid Res 54(9), 2325-2340. 502

6. Bach Knudsen KE, Laerke HN, Hedemann MS et al. (2018) Impact of Diet-Modulated Butyrate 503

Production on Intestinal Barrier Function and Inflammation. Nutrients 10(10). 504

7. Ferchaud-Roucher V, Pouteau E, Piloquet H et al. (2005) Colonic fermentation from lactulose 505

inhibits lipolysis in overweight subjects. Am J Physiol Endocrinol Metab 289(4), E716-720. 506

8. Ingerslev AK, Theil PK, Hedemann MS et al. (2014) Resistant starch and arabinoxylan augment 507

SCFA absorption, but affect postprandial glucose and insulin responses differently. Br J Nutr 508

111(9), 1564-1576. 509

9. Bach Knudsen KE (1997) Carbohydrate and lignin contents of plant materials used in animal 510

feeding. Anim Feed Sci Technol 67, 319-338. 511

10.Bach Knudsen KE, Serena A, Bjørnbak Kjær AK et al. (2005) Rye Bread Enhances the Production 512

and Plasma Concentration of Butyrate but Not the Plasma Concentrations of Glucose and Insulin 513

in Pigs1. J Nutr 135(7), 1696–1704. 514

11.Izydorczyk MS & Biliaderi CG (1995) Cereal arabinoxylans: advances in structure and 515

physicochemical properties. Carbohydr Polym 28, 33-48. 516

12.Glitsø LV, Gruppen H, Schols HA et al. (1999) Degradation of rye arabinoxylans in the large 517

intestine of pigs. J Sci Food Agr 79, 961-969. 518

13.Bach Knudsen KE (2015) Microbial degradation of whole-grain complex carbohydrates and 519

impact on short-chain fatty acids and health. Adv Nutr 6(2), 206-213. 520

14.Nielsen TS, Laerke HN, Theil PK et al. (2014) Diets high in resistant starch and arabinoxylan 521

modulate digestion processes and SCFA pool size in the large intestine and faecal microbial 522

composition in pigs. Br J Nutr 112(11),1837-1849. 523

15.Wlodarska M, Willing B, Keeney KM et al. (2011) Antibiotic treatment alters the colonic mucus 524

layer and predisposes the host to exacerbated Citrobacter rodentium-induced colitis. Infect Immun 525

79(4), 1536-1545. 526

16.Guinan J, Wang S, Hazbun TR et al. (2019) Antibiotic-induced decreases in the levels of 527

microbial-derived short-chain fatty acids correlate with increased gastrointestinal colonization of 528

Candida albicans. Sci Rep 9(1), 8872. 529

17.Bolvig AK, Kyrø C, Nørskov NP et al. (2016) Use of antibiotics is associated with lower 530

enterolactone plasma concentration. Mol Nutr Food Res 60(12), 2712-2721. 531

18.Bolvig AK, Norskov NP, Hedemann MS et al. (2017) Effect of Antibiotics and Diet on 532

Enterolactone Concentration and Metabolome Studied by Targeted and Nontargeted LC-MS 533

Metabolomics. J Proteome Res 16(6), 2135-2150. 534

19.Bolvig AK, Adlercreutz H, Theil PK et al. Absorption of plant lignans from cereals in an 535

experimental pig model. Br J Nutr 115(10), 1711-1720. 536

131

20.Laerke HN, Mortensen MA, Hedemann MS et al.(2009) Quantitative aspects of the metabolism 537

of lignans in pigs fed fibre-enriched rye and wheat bread. Br J Nutr 102(7), 985-994. 538

21.Association of Official Analytical Chemists (2000) AOAC Official Methods of Analysis of 539

AOAC International, 17th ed. Gaithersburg, MD: AOAC International. 540

22.Hansen B (1989) Determination of Nitrogen as Elementary N, an Alternative to Kjeldahl. Acta 541

Agriculturae Scandinavica 39(2),113-118. 542

23.Stoldt W (1952) Vorschlag zur Vereinheitlichung der Fettbestimmung in Lebensmitteln 543

(Suggestion to standardise the determination of fat in foodstuffs). Fette und Seifen 54(4), 206-207. 544

24.Larsson K & Bengtsson S (1983) Bestlmning av llttilgingeliga kolhydrater i vsxtmaterial 545

(Determination of readily available carbohydrates in plant material). Methods report No. 22. 546

National Laboratory of Agricultural Chemistry, Uppsala. 547

25.Kasprzak MM, Laerke HN & Knudsen KE (2012) Effects of isolated and complex dietary fiber 548

matrices in breads on carbohydrate digestibility and physicochemical properties of ileal effluent 549

from pigs. J Agric Food Chem 60(51), 12469-12476. 550

26.Bach Knudsen KE (1997) Carbohydrate and lignin contents of plant materials used in animal 551

feeding. Anim Feed Sci Technol 67, 319-338. 552

27.Schürch AF, Lloyd LE & Crampton EW (1950) The use of chromic oxide as an index for 553

determining the digestibility of a diet. J Nutr 41(4), 629-636. 554

28.Jensen MT, Cox RP, Jensen BB (1995) Microbial production of skatole in the hind gut of pigs 555

given different diets and its relation to skatole deposition in backfat. Anim Sci 61, 293-304. 556

29.Brighenti F (1998) Summary of the conclusion of the working group on Profibre interlaboratory 557

study on determination of short chain fatty acids in blood. In: Functional Properties of Non-558

digestible Carbohydrates, pp. 150-153. F Gullion R Amadò, M T Amaral-Collaco, H Andersson, 559

N G Asp, K E Bach Knudsen, M Champ, J Mathers, J A Robertson, I Rowland, and J Van Loo, 560

editors. Brussels: European Commission, DG XII, Science, Research and Development, 150–153. 561

30.Rerat AA, Vaissade P & Vaugelade P (1984) Absorption kinetics of some carbohydrates in 562

conscious pigs. 2. Quantitative aspects. Br J Nutr 51(3), 517-529. 563

31.Bach Knudsen KE, Serena A, Jørgensen H et al. (2007) Rye and other natural cereal fibres enhance 564

the production and plasma concentrations of enterolactone and butyrate. Dietary fibre components 565

and functions. Wageningen Academic Publishers, 219-233. 566

32.Kristensen NB, Norgaard JV, Wamberg S et al. (2009) Absorption and metabolism of benzoic 567

acid in growing pigs. J Anim Sci 87(9), 2815-2822. 568

33.Verwimp T, Craeyveld VC, Courtin CM et al. (2007) Variability in the Structure of Rye Flour 569

Alkali-Extractable Arabinoxylans. J Agric Food Chem 55(5), 1985-1992. 570

34.Lærke HN & Bach Knudsen KE (2010) Rye Arabinoxylans: Molecular Structure, 571

Physicochemical Properties and Physiological Effects in the Gastrointestinal Tract. Cereal Chem 572

87, 353-362. 573

35.Cleemput G, van Oort M, Hessing M et al. (1995) Variation in the Degree of D-Xylose 574

Substitution in Arabinoxylans Extracted from a European Wheat Flour. J Cereal Sci 22, 73-84. 575

36.Bach Knudsen KE, Lærke HN & Jørgensen H (2008) The Role of Fibre in Nutrient Utilization 576

and Animal Health. Proceedings of the 29th Western Nutrition Conference. 93. 577

37.Gibeaut NC & Gibeaut DM (1993) Structural models of primary cell walls in flowering plants: 578

consistency of molecular structure with the physical properties of the walls during growth. Plant 579

J 3(1), 1-30. 580

38.Lærke HN, Pedersen C, Mortensen MA et al. (2008) Rye bread reduces plasma cholesterol levels 581

in hypercholesterolaemic pigs when compared to wheat at similar dietary fibre level. J Sci Food 582

Agri 88(8), 1385-1393. 583

132

39.Fardet A (2010) New hypotheses for the health-protective mechanisms of whole-grain cereals: 584

what is beyond fibre? Nutr Res Rev 23(1), 65-134. 585

40.Tannock GW & Liu Y (2019) Guided dietary fibre intake as a means of directing short-chain fatty 586

acid production by the gut microbiota. J Roy Soc New Zeal 1-22. 587

41.Müller M, Canfora EE & Blaak EE (2018) Gastrointestinal Transit Time, Glucose Homeostasis 588

and Metabolic Health: Modulation by Dietary Fibers. Nutrients 10(3). 589

42.de Vries J, Birkett A, Hulshof T et al. (2016) Effects of Cereal, Fruit and Vegetable Fibers on 590

Human Fecal Weight and Transit Time: A Comprehensive Review of Intervention Trials. 591

Nutrients 8(3),130. 592

43.Rivera-Chavez F, Zhang LF, Faber F et al. (2016) Depletion of Butyrate-Producing Clostridia 593

from the Gut Microbiota Drives an Aerobic Luminal Expansion of Salmonella. Cell Host Microbe 594

19(4), 443-454. 595

44.Zhao J, Bai Y, Tao S et al. (2019) Fiber-rich foods affected gut bacterial community and short-596

chain fatty acids production in pig model. J Funct Foods 57, 266-274. 597

45.den Besten G, Lange K, Havinga R et al. (2013) Gut-derived short-chain fatty acids are vividly 598

assimilated into host carbohydrates and lipids. Am J Physiol Gastrointest Liver Physiol 305(12), 599

G900-910. 600

46.Pickard JM & Chervonsky AV (2015) Intestinal fucose as a mediator of host-microbe symbiosis. 601

J Immunol 194(12), 5588-5593. 602

47.Fernández J, Redondo-Blanco S, M. Miguélez E et al.(2015) Healthy effects of prebiotics and 603

their metabolites against intestinal diseases and colorectal cancer. AIMS Microbiol 1(1), 48-71. 604

48.Bach Knudsen KE, Jørgensen H & Canibe N (2000) Quantification of the absorption of nutrients 605

derived from carbohydrate assimilation: model experiment with catheterised pigs fed on wheat- or 606

oat-based rolls. Br J Nutr 84(4), 449-458. 607

49.Wong JMW, de Souza R, Kendall CWC et al. (2006) Colonic Health: Fermentation and Short 608

Chain Fatty Acids. J Clin Gastroenterol 40, 235–243. 609

50.Jakobsdottir G, Jadert C, Holm L et al. (2013) Propionic and butyric acids, formed in the caecum 610

of rats fed highly fermentable dietary fibre, are reflected in portal and aortic serum. Br J Nutr 611

110(9), 1565-1572. 612

51.Nilsson AC, Ostman EM, Knudsen KE et al. (2010) A cereal-based evening meal rich in 613

indigestible carbohydrates increases plasma butyrate the next morning. J Nutr 140(11), 1932-1936. 614

52. Venegas DP, De la Fuente MK, Landskron G et al. (2019) Short Chain Fatty Acids (SCFAs)-615

Mediated Gut Epithelial and Immune Regulation and Its Relevance for Inflammatory Bowel 616

Diseases. Front Immunol 10, 277. 617

133

Table 1. Ingredients and chemical composition of the refined wheat fiber (Control) and rye bran 618

(RB) diets. 619

Items Control RB

Ingredients (g/kg, as-fed basis)

Standard white wheat flour1 534 425

Rye bran1 - 270

Vitacel WF 6002 123 -

Fructose3 100 100

Wheat gluten4 30

Lard5 60 60

Rapeseed oil 15 7

Cholesterol6 5 5

Egg powder7 100 100

Vitamins and minerals8 30 30

Chromic oxide6 3 3

Chemical composition (g/kg DM)

DM (g/kg as-fed basis) 908 910

Ash 40 52

Protein (N × 6.25) 158 160

HCL fat 143 142

Available carbohydrates 676 645

Sugars 122 140

Starch 410 363

Dietary fiber9 162 169

Total NSP (soluble NSP) 143 (13) 134 (25)

NCP glucose 10 22

Cellulose 90 14

AX (soluble AX) 37 (6.6) 86 (18)

A:X (soluble A:X) 0.22 (0.95) 0.55 (0.56)

Fructans 0.4 7.7

Klason lignin 19 27

Gross energy (MJ/ kg DM) 20.5 20.6

NSP, total non-starch polysaccharides; NCP, non-cellulosic polysaccharides; AX, arabinoxylan. 620 1Lantmännen Cerealia, Denmark. 621 2Wheat fibre, J.Rettenmeier and Söhne Gmbh, Rosenberg, Germany. 622 3Th. Geyer GmbH, Denmark. 623 4Lantmännen Reppe, Sweden. 624 5Daka Denmark A/S, Denmark. 625 6Sigma-Aldrich Denmark ApS. 626 7Sanovo Food A/S, Denmark. 627 8VA Vit SL/US antiox, Vestjyllands Andel, Ringkøbing, Denmark. 628 9Dietary fibre counted as the sum of NSP, fructans and Klason lignin. 629

134

Table 2. Carbohydrate composition of ileal digesta of pigs fed with refined wheat fiber diet (Control), 630

rye bran diet with (RB+) and without antibiotic treatment (RB-). 631

% of DM Diet

SEM P-value Control RB- RB+

Total NDC 60a 45b 44b 0.92 <0.001

AX 15b 27a 26a 0.90 <0.001

A:X 0.15b 0.48a 0.48a 0.01 <0.001

Soluble NSP 4.1b 9.3a 8.4a 0.68 <0.001

AX 2.0b 5.9a 5.3a 0.38 <0.001

A:X 0.85a 0.55b 0.55b 0.04 <0.001

NCP glucose 1.0b 2.0a 1.7ab 0.22 0.016

Insoluble NSP 50a 29b 29b 0.99 <0.001

AX 8.9b 18a 18a 0.58 <0.001

A:X 0.09b 0.53a 0.53a 0.02 <0.001

NCP glucose 2.9b 4.3a 3.8ab 0.42 0.015

Cellulose 38a 5.0b 4.5b 0.61 <0.001

LMW NDC 4.2b 6.2ab 6.6a 1.2 0.028

Arabinose 0.41 0.30 0.06 0.30 0.406

Xylose 4.1a 2.8b 1.9b 0.33 <0.001

DM, dry matter; NDC, non-digestible carbohydrate; AX, arabinoxylan; NSP, non-starch 632

polysaccharides; NCP, non-cellulosic polysaccharides; LMW: low molecular weight. 633

Data were expressed as means with standard error of means (SEM) and n = 10 per group. Mean values 634

within a row with unlike superscript letters were significantly different between the treatments (P < 635

0.05).636

135

Table 3. Digestibility of dry matter (DM), nutrients and non-starch polysaccharides (NSP) 637

components and arabinose:xylose (A:X) ratios in the different intestinal segments of pigs fed with 638

refined wheat fiber diet (Control), rye bran diet with (RB+) and without antibiotic treatment (RB-). 639

Item Segment

Diet

SEM

P-value

Control RB- RB+ Diet Segment Diet ×

Segment

DM SI3 0.68f 0.69f 0.65g 0.007 0.004 <0.001 <0.001 Caecum 0.79de 0.76e 0.77e Co1 0.82cd 0.82bcd 0.81cd Co2 0.85ab 0.83ab 0.82bcd Co3 0.86a 0.84b 0.83abc

Protein SI3 0.76ef 0.71gh 0.69h 0.009 <0.001 <0.001 <0.001

Caecum 0.86ab 0.73fg 0.74fg

Co1 0.87ab 0.79de 0.78de

Co2 0.87a 0.81cd 0.80cde

Co3 0.87a 0.84abc 0.83bc

Fat SI3 0.84a 0.78b 0.77b 0.015 0.004 - -

Starch SI3 0.98 0.97 0.98 0.002 0.815 <0.001 0.542

Caecum 0.99 0.99 0.99

Total NSP SI3 -0.15g 0.11ef 0.04f 0.03 <0.001 <0.001 <0.001

Caecum 0.25e 0.44bcd 0.42cd

Co1 0.36d 0.60a 0.59a

Co2 0.47abc 0.60a 0.57ab

Co3 0.58a 0.60a 0.57ab

Cellulose SI3 -0.27g -0.09defg -0.11defg 0.07 <0.001 <0.001 <0.001 Caecum 0.13cde -0.11efg -0.14fg

Co1 0.25bc 0.15bc 0.13cde

Co2 0.38ab 0.16bc 0.12cde

Co3 0.51a 0.14bcd 0.05cdef

AX SI3 0.11f 0.14f 0.04f 0.03 0.030 <0.001 0.010

Caecum 0.50cde 0.49de 0.47e

Co1 0.60bcd 0.63ab 0.63bc

Co2 0.67ab 0.63ab 0.61bc

Co3 0.75a 0.63abc 0.60bcd

A:X ratio SI3 0.17f 0.53e 0.53e 0.02 <0.001 <0.001 <0.001

Caecum 0.10g 0.84d 0.78d

Co1 0.10fg 1.02bc 1.00c

Co2 0.12fg 1.08abc 1.06abc Co3 0.14fg 1.09a 1.08ab

SI3, distal one-third of the small intestine; Co1, proximal one-third of the colon; Co2, mid one-third 640

of the colon; Co3, distal one-third of the colon; AX, arabinoxylan. 641

Data were expressed as means with standard error of means (SEM) and n = 10 per group. Mean values 642

within a row with unlike superscript letters were significantly different (P < 0.05).643

136

Table 4. Concentrations of total, individual short chain fatty acids (SCFA) and total branched chain 644

fatty acids (BCFA) in different intestinal segments of pigs fed with refined wheat fiber diet (Control), 645

rye bran diet with (RB+) and without antibiotic treatment (RB-). 646

Item Segment

Diet

SEM

P-value

Control RB- RB+ Diet Segment Diet ×

Segment

Concentration (mmol/kg)

SCFA SI3 29f 42ef 30f 5.5 <0.001 <0.001 <0.001

Caecum 132ab 115bc 115bc

Co1 145a 106bcd 102cd

Co2 130ab 86de 83de

Co3 144a 72e 70e

Acetate SI3 17f 26ef 19f 3.4 <0.001 <0.001 <0.001 Caecum 94a 64bc 66b

Co1 98a 61bc 60bc

Co2 85a 50bcd 48cd

Co3 87a 42de 41de

Propionate SI3 0.80f 0.63f 0.38f 1.9 <0.001 <0.001 <0.001 Caecum 29abc 25abc 24bcd Co1 31ab 21cde 21cde Co2 28abc 17de 17de Co3 32a 15e 15e Butyrate SI3 1.1f 2.5ef 0.79f 1.5 0.023 <0.001 <0.001 Caecum 7.0de 22a 14bc

Co1 12bcd 17b 16b

Co2 14bc 13bc 12bcd

Co3 17b 9.1cd 8.7cd

BCFA SI3 0.0g 0.0g 0.0g 0.22 <0.001 <0.001 <0.001 Caecum 1.1f 1.4ef 1.2f

Co1 1.4ef 3.0abc 2.3cde

Co2 1.4ef 3.4abc 2.7bcd

Co3 1.9edf 3.9a 3.3ab

SI3, distal one-third of the small intestine; Co1, proximal one-third of the colon; Co2, mid one-third 647

of the colon; Co3, distal one-third of the colon. 648

Data were expressed as means with standard error of means (SEM) and n = 10 per group. Mean values 649

within a row with unlike superscript letters were significantly different (P < 0.05).650

137

Table 5. Plasma concentrations, hepatic extraction and daily absorption of short chain fatty acids (SCFA) and branched chain fatty acids (BCFA) in

pigs fed with refined wheat fiber diet (Control), rye bran diet with (RB+) and without antibiotic treatment (RB-).

Item Site

Diet

SEM

P-value

Control RB- RB+ Diet Site Diet × Site

Concentration (µmol/L)

SCFA PV 2287a 1442b 1457b 99 <0.001 <0.001 <0.001

HV 825c 608cd 536cd

CA 430d 334d 324d

Acetate PV 1713a 1004b 980b 79 <0.001 <0.001 <0.001

HV 800bc 560cd 507cd

CA 414d 317d 308d

Propionate# PV 420 (273-647) 219(142-337) 275(178-423) - 0.320 <0.001 0.060

HV 7.4 (4.8-11) 5.6 (3.6-8.6) 6.5 (4.2-9.9)

CA 2.4 (1.5-3.6) 2.9 (1.9-4.4) 2.5 (1.6-3.9)

Butyrate# PV 109 (82-145) 129 (97-172) 120 (90-160) - 0.254 <0.001 0.100

HV 8.1 (6.1-11) 8.5 (6.4-11) 13 (9.9-18)

CA 4.9 (3.7-6.5) 7.0 (5.2-9.3) 5.8 (4.3-7.7)

BCFA# PV 33 (26-43) 46 (36-60) 50 (38-64) - 0.020 <0.001 0.882

HV 5.5 (4.2-7.2) 6.2 (4.8-8.1) 7.4 (5.7-9.6)

CA 4.3 (3.3-5.5) 5.8 (4.5-7.6) 6.3 (4.8-8.1)

Net absorption (mol/d)

SCFA 8.0a 4.8b 5.0b 0.72 0.005 - -

Acetate 5.4a 3.0b 2.9b 0.48 0.001 - -

Propionate 1.7a 1.1b 1.3ab 0.18 0.022 - -

Butyrate 0.49 0.60 0.55 0.11 0.629 - -

BCFA 0.14 0.19 0.20 0.02 0.209 - -

Hepatic extraction (%)

138

Total SCFA 50 57 53 7.1 0.782 - -

Acetate 33 44 39 9.7 0.706 - -

Propionate 97 98 97 0.68 0.297 - -

Butyrate 90 88 90 2.1 0.801 - -

BCFA 80 80 80 4.1 0.993 - -

PV, portal vein; HV, hepatic vein; CA, carotid artery. #Values are back-transformed after log transformation and are expressed as means (95% confidence interval).

SEM, standard error of means and n = 10 per group. Mean values within a row with unlike superscript letters were significantly different (P < 0.05).

139

SI3 Cecum Co1 Co2 Co3

0

500

1000

1500

Ne

t a

mo

un

t, g

Control

RB-

RB+

cde cdee

(A)

cdebcde

e

a

b bc bcdbcde

bcdecdecdede

SI3 Cecum Co1 Co2 Co3

0

2

4

6

8

Me

an

tra

nsi

t ti

me

, h

Control

RB-

RB+cd

a

cd

bbc

bc

(B)

bcdcdcd

eefde

fg gg

Fig. 1. Net amount (A) of wet digesta and mean transit time (B) in the intestinal segments of pigs fed

with refined wheat fiber (Control), rye bran diet with (RB+) and without antibiotic treatment (RB-).

SI3, distal one-third of the small intestine; Co1, proximal one-third of the colon; Co2, mid one-third

of the colon; Co3, distal one-third of the colon. Values are least-squares means with standard errors

represented by vertical bars and n = 10 per group. Different superscript letters mean significant

difference (P < 0.05). There were significant effects of diet ((B), P = 0.001), segment ((A) and (B),

P < 0.001) and diet segment interaction ((A) and (B), P < 0.001).

140

SI3 Cecum Co1 Co2 Co3

5.0

5.5

6.0

6.5

7.0

7.5

Control

RB-

RB+bc

cd

ef

abc

ab a

ab

pH

ab

abcabc

de

e

efef

f

Fig. 2. The pH of digesta in the intestinal segments of pigs fed with refined wheat fiber (Control), rye

bran diet with (RB+) and without antibiotic treatment (RB-). SI3, distal one-third of the small

intestine; Co1, proximal one-third of the colon; Co2, mid one-third of the colon; Co3, distal one-third

of the colon. Values are least-squares means with standard errors represented by vertical bars and n

= 10 per group. Different superscript letters mean significant difference (P < 0.05). There were

significant effects of diet (P = 0.001), segment (P < 0.001) and diet segment interaction (P < 0.001).

141

SI3 Cecum Co1 Co2 Co3

40

50

60

70

80

Ace

tate

/SC

FA

, %

Control

RB-

RB+a

ab

cd

cd

d

cd

(A)

bcd

bcd

d

abc

cd

cd cd

cd

bcd

SI3 Cecum Co1 Co2 Co3

0

10

20

30

Pro

pio

nat

e/S

CF

A,

%

Control

RB-

RB+

(B)

SI3 Cecum Co1 Co2 Co3

0

5

10

15

20

25

Bu

tyra

te/S

CF

A,

%

Control

RB-

RB+a

de

g ef

abc

ab

e

bcd

bcd

(C)

fg

gg

cdecdede

SI3 Cecum Co1 Co2 Co3

0

2

4

6

8

BC

FA

/SC

FA

, %

Control

RB-

RB+

de

ab

a

bc

cd

(D)

hhh

fghgh

gh gh ghefg

ef

Fig. 3. The acetate (A), propionate (B), butyrate (C) and branched chain fatty acids (D) proportion of

total short chain fatty acids (SCFA) in different intestinal segments of pigs fed with refined wheat

fiber (Control), rye bran diet with (RB+) and without antibiotic treatment (RB-). SI3, distal one-third

of the small intestine; Co1, proximal one-third of the colon; Co2, mid one-third of the colon; Co3,

distal one-third of the colon. Values are least-squares means with standard errors represented by

vertical bars and n = 10 per group. Different letters mean significant difference (P < 0.05). There

were significant effects of diet ((A), (C) and (D); P < 0.001), segment ((B-D), P < 0.001) and diet

segment interaction ((A), (C) and (D); P < 0.001).

142

Fig. 4. Heat map correlations of caecal carbohydrate with short chain fatty acids (SCFA) composition.

The x-axis contains starch, dry matter (DM), individual and total non-starch polysaccharides (NSP),

relative percentage of cellulose, glucose and arabinoxylan in total NSP (cellulose_NSP, glucose_NSP

and AX_NSP) and ratio of arabinose:xylose (AX ratio), and y-axis contains pH, pool size of SCFA

and branched-chain fatty acids (BCFA), and individual SCFA proportion in total SCFA

(Acetate_SCFA, Butyrate_SCFA, Propionate_SCFA and BCFA_SCFA). Yellow asterisks indicate

statistical significance (P < 0.05) by Pearson’s test.

143

Fig. 5. Heat map correlations of the rarefied relative abundance of caecal microbiota with SCFA and

carbohydrate. The x-axis contains taxa > 1% abundance and y-axis contains starch, dry matter (DM),

individual and total non-starch polysaccharides (NSP), relative percentage of cellulose, glucose and

arabinoxylan in total NSP (cellulose_NSP, glucose_NSP and AX_NSP), ratio of arabinose:xylose

(AX ratio), pH, pool size of SCFA and branched-chain fatty acids (BCFA), and individual SCFA

proportion in total SCFA (Acetate_SCFA, Butyrate_SCFA, Propionate_SCFA and BCFA_SCFA).

Yellow asterisks indicate statistical significance (P < 0.05) by Pearson’s test.

144

Fig. 6. Pearson correlations of total short chain fatty acid (SCFA), acetate and propionate absorption

with plasma cholesterol and LDL of pigs.

145

Table S1. Chemical composition of fiber sources in the diets.

Chemical composition (g/kg DM) Wheat flour Rye bran Vitacel WF 600

DM (g/kg as-fed basis) 864 906 938

Ash 2 53 n.m.

Protein (N × 6.25) 139 213 n.m.

HCL fat 34 n.m.

Available carbohydrates 788 187 n.m.

Sugars 20 67 n.m.

Starch 768 120 n.m.

Dietary fiber9 39 537 1000

Total NSP (soluble NSP) 30 (14) 433 (57) 1000

NCP glucose 3 76 71

Cellulose 2 45 717

AX (soluble AX) 18 (9) 283 (42) 168 (n.m.)

A:X (soluble A:X) 0.80 (0.80) 0.52 (0.56) 0.09 (n.m.)

Fructans 0 32 n.m.

Klason lignin 9 72 n.m.

n.m., not measured.

146

Table S2. Pool size of total, individual short chain fatty acids (SCFA) and total branched chain fatty

acids (BCFA) in different intestinal segments and large intestine of pigs fed with refined wheat fiber

diet (Control), rye bran diet with (RB+) and without antibiotic treatment (RB-).

Pool size,

mmol Segment

Diet

SEM

P-value

Control RB- RB+ Diet Segment Diet ×

Segment

SCFA SI3 13g 19fg 15fg 7.8 <0.001 <0.001 <0.001

Caecum 45cdef 54cde 38cdefg

Co1 145a 75bc 70bcd

Co2 84b 51cdef 44cdefg

Co3 69bcd 33defg 29efg

Acetate SI3 7.3f 12ef 9.5ef 4.9 <0.001 <0.001 <0.001

Caecum 31cde 31cdef 21cdef

Co1 98a 43bc 41bcd

Co2 55b 30cdef 25cdef

Co3 40bcd 19def 16ef

Propionate SI3 0.40f 0.31f 0.20f 1.7 <0.001 <0.001 <0.001

Caecum 9.8cde 11bcde 8.5cde

Co1 30a 15bc 14bcd

Co2 18b 11bcde 9.0cde

Co3 14bcde 7.0edf 6.2ef

Butyrate SI3 0.47e 1.14e 0.43e 1.2 0.169 <0.001 <0.001

Caecum 2.3de 9.4a 4.3bcde

Co1 12a 11a 11a

Co2 9.0ab 7.7abc 6.4abcd

Co3 8.2abc 4.2bcde 3.4cde

BCFA SI3 0.0 0.0 0.0 0.23 0.004 <0.001 0.265

Caecum 0.36 0.62 0.42

Co1 1.3 2.1 1.6

Co2 0.91 2.0 1.5

Co3 0.86 1.9 1.0

SCFA LI 364a 203b 180b 18 <0.001 - -

Acetate LI 225a 117b 104b 15 <0.001 - -

Propionate LI 74a 42b 41b 6.3 <0.001 - -

Butyrate LI 33 32 25 4.9 0.217 - -

BCFA LI 3.8b 7.1a 5.2ab 0.8 0.021 - -

SI3, distal one-third of the small intestine; Co1, proximal one-third of the colon; Co2, mid one-third

of the colon; Co3, distal one-third of the colon; LI, large intestine (caecum + colon).

Data were expressed as means with standard error of means (SEM) and n = 10 per group. Mean

values with unlike superscript letters were significantly different (P < 0.05).

147

7. Discussion

7.1 Obesity development of minipig models

Göttingen Minipig is a breed that can easily develop obesity than conventional pigs and has

comparable fat distribution with humans (80% subcuteneous adipose tissue and 20% visceral adipose

tissue) (175), therefore we use it as an model to study obesity and phenotype signatures associated

with MetS. The standard chow of Göttingen Minipigs includes 13.8 MJ/kg gross energy, 2.1% fat,

13% protein and 14.5% DF, and is normally fed limited to minipigs to avoid obesity (300-400 g/d

for 6-9 months minipigs). In order to develop obesity in this study, all the minipigs had been provided

a high fructose, high fat western lifestyle diet (20.5 MJ/kg gross energy, 23.3% sugar and 10.5% fat,

11.3% protein and 10% DF) with ad libitum feeding to mimic the excessive energy intake of obese

humans. This was achieved after 20 weeks as body weight doubled compared with the normal

minipigs at the same age (166, 204). Meanwhile, mild characteristics of MetS had been observed,

including increased BSA and POI indices, increased fasting glucose and insulin levels, decreased

HDL levels over the 20 weeks obesity induction (166).

Therefore, in the present study, we aimed to investigate if supplementation of DF and protein alone

or combined would ameliorate MetS biomarkers in this obese porcine model. Besides general

obesity, abdominal obesity is one of the typical characteristics of patients with MetS and means

increased visceral adipose tissue (19). Instead of BMI and waist circumference in humans, obesity

indices including BSA and POI were used to assess the general obesity whilst abdominal

circumference of minipigs was for abdominal obesity in Paper I. After the 8-week dietary

intervention, the development of obesity indices followed the trend of body weight, but the effect of

dietary treatments did not reach a statistical significance. This was in agreement with a previous

intervention in humans, which showed insignificant change in body weight and body composition

after 12 weeks of whey protein and wheat fiber intake (89). Although the high level of dietary fat

aimed to accelerate the development of non-alcoholic fatty liver, a liver fat content (< 1.5%) within

a normal range (205) was seen in the young Göttingen Minipigs model in Paper I, which further

illustrated that it is difficult to develop NAFL in pigs since the primary de novo lipogenesis site is

adipose tissue rather than the liver (182). Since metabolic biomarkers can be influenced by weight

loss (206), our initial aim was to use a weight stable pig model during the intervention. However, it

was hard to obtain full-grown and weight stable minipigs with obesity as it would be overspent and

require a long experimental time. Therefore, we changed the aim to growing minipigs with similar

weight gain during the dietary intervention period, which was seen in this study that minipigs were

still in juvenile period (sexual maturity time is normally 3.7-6.5 months) (204) and did not stop

growing. Therefore, one of the limitations of our study is that the minipigs were still young and

148

continued growing during the intervention, which might have overshadowed the effects of the DF

and protein. It is possible that more obvious signs of MetS would occur if we had used fully matured

and full-grown obese animals instead of young animals. Moreover, instead of ad libitum feeding,

restricted feeding strategy on the body weight basis could be used to maintain weight stable in a fully

matured model with obesity.

7.2 Clinical parameters associated with MetS in minipig models

Instead of directly affecting insulin response, high DF diets increased non-fasting C-peptide secretion

in Paper I, which demonstrated an improved pancreatic function. Accordingly, this was also partly

reflected by the increased GLP-1 secretion in the fasting state, which has been found to play a role

in pancreatic β-cell growth and improved insulin sensitivity (73). Our results also corresponded with

previous studies showing that high cereal fiber intake improved GLP-1 secretion (70) and insulin

response (74, 76) possibly due to slower nutrient absorption and increased SCFA production with

high DF intake. However, it should be kept in mind that some conflicting results exist in previous

studies. For example, high DF content in the form of rye flakes and enzyme-treated wheat bran did

not influence postprandial GLP-1 secretion but lowered insulin secretion and its concentration in the

portal vein of healthy pigs compared with feeding a low DF western lifestyle diet (76, 198). The

reasons for the discrepancy could be different physiological status (lean or obese) of the animal

models but also differences in DF sources and diet composition.

Whey proteins can act as potent secretagogues of enteroendocrine L-cells and enhance GLP-1

secretion (207). In agreement with that, the incretin hormone GLP-1 in the non-fasting state tended

to increase with high protein, but it was not reflected in the differences in feed intake over the 8

weeks (Paper I). It suggested an acute effect of high protein on appetite hormones and satiety 2-4

hours after a meal intake (208), whereas a long term continued intake of high protein may lead to

metabolic adaptation so that the improved satiety in the acute phase did not consequently lead to an

overall reduced energy intake or weight loss in a long-term intervention (209). Along with the

difference in GLP-1, insulin levels in the non-fasting state tended to be stimulated by high protein

diets, which is in line with previous studies showing that whey protein high in BCAA has

insulinotropic potential (126, 130). Surprisingly, we observed increased glucose concentrations in

fasting plasma with no significant change in insulin response and sensitivity with the high protein

diets, which was contrast with our hypothesis. Improved insulin sensitivity with high protein energy-

restricted diets was found in overweight, obese or T2D subjects and partly dependent on weight loss

(137). Hence, unchanged insulin sensitivity with the high protein diets in the current study was

possibly due to the weight gain of the minipigs, the supplementation of energy dense (high sugar

high fat) rather than energy restricted diets as well as an ad libitum feeding pattern. A recent study

149

has found that increased circulating BCAA caused by whey protein intake for a long-term was

detrimental for glucose and lipid homeostasis (210). In the present study, high protein intake

upregulated the gene expression of hepatic FBP1, suggesting increased gluconeogenesis and possibly

would contribute to the development of IR as reported before (211). Given that mild signs of insulin

resistance were seen at baseline, increased glucose responses in the high protein groups indicates

possible progression of IR if the high protein intervention had continued for a longer period as also

observed previously (137).

Weickert et al. (212) pointed out that high protein intake had a tendency to reduce insulin sensitivity,

while high cereal fiber intake could attenuate this negative effect. However, we did not find

significant amelioration of glucose and insulin responses when combining DF with high protein,

possibly because the minipigs still seemed to respond to insulin. Similar observations were seen in a

parallel study with combination of whey protein and wheat fiber provided to abdominal obese

humans (167). The authors explained that the method of insulin sensitivity measurement and physical

conditions of the subjects could also contribute to the different outcomes of insulin sensitivity. It is

possible that other methods such as OGTT might potentially improve the ability to detect changes in

insulin sensitivity. However, measurement of OGTT was not accomplished in the current study

because we did not manage to setup a protocol that ensured satisfactory intake of a fixed amount

glucose in these ad libitum fed minipigs.

In Paper I, high DF diets did not show any improvement in plasma lipid biomarkers which was also

shown in the parallel study with abdominally obese humans fed the a same type of DF (89, 142).

High DF, especially soluble fiber with high viscosity, is generally considered a critical factor to

reduce lipid response (81, 83). Compared with low DF diets, high DF diets added with enzyme-

treated wheat bran contained double soluble NSP wherein soluble AX contributed to a major

proportion (76%), but it might not increase viscosity to the same extent, therefore, the differences of

viscosity between high and low DF diets might be insignificant. Correspondingly, a previous study

from our group showed that rye was more efficient than wheat in reducing plasma cholesterol in pigs

due to the higher viscosity of higher soluble AX content in rye (90). Although high DF did not

improve plasma lipid profile, downregulated gene expression of fatty acid synthase (FASN) in

adipose tissue was observed and indicated potentially beneficial effects of high DF on modulating

fatty acid and cholesterol synthesis. The high protein diets did not significantly change plasma

triglyceride concentrations either, but elevated LDL concentration and LDL:HDL ratios were

observed in the portal and jugular vein in the non-fasting state. Moreover, in the fasting state, more

NEFA was also released to the blood with high protein diets. Studies have reported that non-fasting

triglyceridemia was much closely associated with cardiovascular risk than fasting triglyceridemia (8,

150

213). Therefore, the altered lipid profile with the high protein diets could be an indicator of a

promoted dyslipidemia process and increased CVD risk. This was in contrast to the human study

where intake of whey protein in combination with low DF for 12 weeks improved the postprandial

lipid profile in abdominal obese subjects (89). Notably, high BCAA contents in the high protein diets

of Paper I might even elicit detrimental effects, which were unexpected but reported in recent studies

that a combination of BCAA with high dietary fat content caused injured liver in mice (145) and

increased BCAA accumulation in the circulation may impair lipid homeostasis of obese minipigs

(210). Therefore, it is likely that BCAA-enriched high protein diets in combination with a high fat

content can induce more severe hyperlipidemia on a longer term.

Although we observed mild signs of liver inflammation and upregulated gene expression of lipid

metabolism over the 20 weeks’ obesity development period, that was not reflected in changes of

plasma inflammatory cytokines (166). Surprisingly, high DF resulted in higher INF-γ and IL-12

levels compared with the low DF diets, which was in contrast to previous findings (98, 99). However,

studies on effects of DF on low-grade inflammation are limited and a review reported that lowering

of concentrations of inflammatory were associated with weight loss (206). Thus, a potential

explanation was that the minipigs were still growing therefore the inflammatory cytokines could be

influenced by the growing status. Taken together, although a diet high in fiber is in general considered

as a healthy choice, their role in reducing inflammation needs further investigation.

7.3 Dietary fiber degradation and nutrient digestion

Soluble fiber is generally degraded in proximal part of the large intestine while insoluble fiber can

be fermented in a more distal part of colon (55). Hence, we chose fiber sources from cereal grains

and brans that containing enriched AX in soluble and insoluble forms, to obtain a gradual and

continuous degradation along the large intestine. For the high DF diets in Paper I+II, we used enzyme

treatment of wheat bran to release AXOS fractions from insoluble AX in aleurone cells (62) and

make the wheat bran more fermentable (55). By comparison of the AX structure between diets and

ileal digesta in Paper II, we found that AX degradation already occurred in the small intestine where

the LMW oligosaccharides could be utilized by microbiota as previously reported (214). Similarly,

AX was also slightly degraded in ileum (Control: 11%; RB-: 14%), with a large extent of degradation

in cecum (~50%) and proximal colon (~60%) due to a lower flow rate and higher microbial density

compared with the ileum (Paper III) (215). Corresponding to this observation, a previous study from

our group showed that the ileal digestibility of AX in pigs fed with dark ground rye bread was 27%

(214) and the digestibility of AX in cecum and colon of pigs feeding rye aleurone-based diet were

62% and 74%, respectively (216). The different extent of AX degradation are largely dependent on

the characteristics of AX structure, which was also reflected in the degradation patterns seen in Paper

151

III. Importantly, the degradation of AX was accompanied with changes in A:X ratios and linked with

the degree of substitution and substitution pattern of AX. For instance, in another study it was shown

that AX in the pericarp has a high A:X ratio (> 1.0) with only 17% un-substituted xylose, > 50%

mono-/double-substituted xylose and is difficult to be degraded, while AX in the aleurone layer is

largely unsubstituted (62%) with a lower A:X ratio (~0.40) and was gradually degraded from the

ileum to a large extent in the distal colon(> 60%) (101). This demonstrates that the structure of AX

affects not only the extent but also the site of degradation. These earlier observations were confirmed

in the present studies wherein A:X ratios of the diets were medium (0.63 with added enzyme-treated

wheat bran in Paper II and 0.55 rye bran in Paper III, respectively) and AX was continuously

degraded until mid colon (> 60%). After the AX degradation in mid colon, the A:X ratio increased

(> 1.0 ) and the remainder of AX was difficult to be fermented due to the high degree of substitution

of the xylose backbone. For AX in the refined wheat fiber diet of Paper III, the structure is almost

linear and contained a high proportion of insoluble AX, therefore showing almost unchanged A:X

ratios along the large intestine.

The natural cellulose abundant in cereal bran forms rigid, insoluble and crystalline microfibrils in the

plant cell wall, and is more resistant to enzymatic hydrolysis of microorganisms than refined cellulose

(217). Cellulose fines are in general small particles removed from natural cellulose fiber during a

refining process and can be degraded in a larger extent than cellulose in its natural matrix (218).

These properties can explain the lower extent of cellulose degradation with the RB diets compared

with the cellulose enriched Control diet in Paper III which confirms previous finding in our lab (200,

216). The different degradation patterns between RB diet and the Control diet in Paper III affected

the pH profiles differently wherein the slower degradation of cellulose in the refined wheat fiber diet

led to lower pH values along the large intestine compared with the RB diets.

In Paper III, the protein and fat digestibility was reduced by 6 and 7 percent units with the RB diets

compared with refined wheat fiber diet. This was in correspondence with the findings that the

increased luminal viscosity by DF can hinder the protein hydrolysis by endogenous enzyme and

disturb lipid emulsification in the small intestine (200, 216), therefore can result in reduced

macronutrient digestion and absorption. A large part of protein was presumably entrapped in rye bran

(49) of RB diet and less accessible for the enzymes, thereby showed a lower and slower degradation

along the intestine than the refined wheat fiber diet. The lower digestibility of fat and protein of pigs

fed with RB diets confirms the findings of previous studies in rye-based high DF diets (4, 216).

The effects of antibiotics on DF degradation or fermentation process have previously been poorly

studied. Our study (Paper III) showed only a marginal effect of antibiotic treatment on nutrient

digestion and fiber degradation processes. Although a muscular injection of the antibiotics in pigs

152

allowed us to control the identical antibiotic dosage for each animal, one of the limitations is the

different administration route compared with human studies wherein antibiotics mostly are

administered orally. Therefore, it is likely that the muscular injection of in our study weakened the

efficacy of the antibiotics and may be one of the reasons for the absence of effects on digestion and

degradation processes. If we instead had oral administration and extended the duration of antibiotic

treatment, the efficacy of antibiotics could be increased and effects on digestion and degradation

processes might be more obvious.

7.4 SCFA production and circulation

One of the main mechanisms of action of DF on systemic health is caused by SCFA produced from

the microbial fermentation in the gut (219). Different SCFA distributional profiles are functionally

important; butyrate in particularly, is the principal source of metabolic energy for the colonocytes

and is considered to play an important role in modulating intestinal ecology (110), moreover, may

subsequently alleviate obesity-related metabolic disorders found in humans and animal models (7).

Therefore, we aimed to improve metabolic health through stimulated butyrate production by

providing AX-enriched high DF diets. In Paper II we saw that high DF diets modestly increased the

butyrate pool in the large intestine and the RB diets in Paper III increased butyrate production and

its proportion of SCFA to a greater extent than the Control diet especially in cecum. These

observations demonstrated that DF enriched in AX could improve butyrate production in the large

intestine which fits well with previous findings (4, 54). Moreover, a close relation between AX

content and butyrate concentration or proportion was demonstrated in the correlation heat map of

cecal digesta (Paper III). Besides AX content, some individual sugars that present in rye bran

including galactose, fucose and uronic acid showed positive correlations with butyrate concentration

and proportion in the cecal digesta. However, the treatment with antibiotics in pigs negated the

beneficial effects of AX on butyrate production by showing a reduced butyrate concentration, pool

size and proportion in cecum compared with the un-treated RB group. Similarly, other studies have

demonstrated that antibiotic treatment can adversely affect the microbiome and decrease the SCFA

production in vivo (220) or in vitro (221), highlighting the potential negative effects of antibiotics on

intestinal fermentation process. Differences in DF composition affected the fermentation products,

consumption of the Control diet high in cellulose significantly increased acetate production in Paper

III, reflected by the strong correlations between cellulose content in cecal digesta and acetate

concentration and proportion as also illustrated in another study in pigs (222).

The changes in SCFA profiles along the intestine corresponded with the fermentation activity which

was high in cecum and proximal colon due to high microbial activity and plentiful fermentable

substrates (215). With the depletion of carbohydrates along the large intestine, proteolytic

153

fermentation increased and correspondingly, pH also increased. It was reported that 30% fermented

protein is converted to SCFA, of which BCFA constitutes between 16% and 23% depending on the

type of fermented proteins (223). Therefore, not only BCFA production was increased with high

protein diets in Paper II, increased total SCFA, acetate and propionate were also observed in the

intestinal segments. However, it should be noticed that high protein fermentation can elicit

detrimental effects by producing metabolites such as hydrogen sulfide, ammonia, and p-cresol, ect.

that promotes intestinal inflammatory disease (148). Interestingly, high DF contents limited

proteolytic fermentation in the distal colon by showing a reduced BCFA proportion in comparison

to the low DF groups. Similar findings was also reported in a recent review (148) that high fiber

intake may alter protein fermentation pathways and provide protective effects against gut

inflammation. The decreased protein fermentation in the distal colon of the high DF groups can be

one of the reasons for a lower pH value compared with the low DF groups, which can prevent the

overgrowth of pathogenic bacteria and reduce the risk of bowel disease [61]. In Paper III, the RB

diet increased the concentration of BCFA and their proportion of SCFA, which corresponded with

the decreased protein digestibility thus increasing the delivery of this nutrient to the gut microbiota.

A wide range of metabolites after intestinal fermentation is absorbed by intestinal epithelium to

varying extent and participates in systemic metabolism via blood circulation which plays a role in

metabolic health (105). In Paper III, the higher SCFA, acetate and propionate absorption could be

attributed to a larger intestinal production and a longer transit time in the large intestine with the

Control diet compared to the RB diets, which would provide more time for fermentation of DF and

thereby permit greater SCFA absorption (215). Both acetate and propionate have been reported to

promote lipogenesis (110, 224) which was further reflected in our study, where acetate and

propionate absorption correlated with plasma cholesterol and LDL concentrations in Paper III,

illustrating the beneficial effects of RB diets on plasma lipid profile (203). However, the butyrate

absorption in Paper III was not aligned with the increased intestinal butyrate production, possibly

because it was largely utilized by the epithelial colonocytes (110) and the differences therefore

difficult to detect in the subsequent systematic circulation. Moreover, the changes of intestinal SCFA

concentrations in Paper II were not totally translated into portal circulation where only BCFA

showed significantly increased levels with high protein diets. The mechanism is not clear, but SCFA

can be utilized by colonocytes (110) which could cause a more even SCFA profile among dietary

treatments in portal vein. High propionate and butyrate levels in the blood can induce the

accumulation of TCA intermediates and affect normal cell metabolism (108), hence, the liver clears

the major part of propionate and butyrate from the portal circulation (110, 225). This was also

illustrated in Paper III, where the hepatic clearance of propionate (97%) and butyrate (89%) was high

154

and intermediate for acetate (~40%), irrespective of diets. High hepatic extraction rates of propionate

(94%) and butyrate (82%) were also reported in a previous study and were neither changed with total

DF levels nor AX content (76).

After hepatic extraction, circulating SCFA may have a different profile compared with portal, and

systemic SCFA are also derived from endogenous sources including metabolism of fat, carbohydrate

and amino acids (226). It is therefore reasonable that the SCFA in peripheral circulation has a

different profile compared with the portal vein and intestine. The increased propionate concentrations

in jugular vein of the high protein groups could be due to higher catabolism of BCAA than the low

protein groups, while the increased butyrate concentrations with high DF diets could be partly due to

endogenous fatty acid oxidation that had been indicated by a study (226). A previous study found

that fasting circulating SCFA levels were related to metabolic parameters including GLP-1, lipid and

insulin sensitivity (227). Therefore, the higher butyrate levels in the jugular vein of pigs fed high DF

may potentially contribute to the beneficial effects of DF on beta-cell function and associated with

the increased GLP-1 concentrations in the fasting plasma found in Paper I. Interestingly, we found

higher propionate and lower succinate concentrations in plasma from jugular vein with high protein

intake. As mentioned above that the porcine liver removes ~90% of the net portal absorption of

propionate for gluconeogenesis and circulating propionate at elevated concentrations probably were

toxic (108). However, the plasma propionate concentrations in jugular vein were in a normal range

compared to other human and pig studies (198, 227). Therefore we speculate that the higher

propionate concentrations with the high protein diets may have contributed to the tendency of

increased GLP-1 concentrations in non-fasting plasma found in Paper I. The reduced jugular

concentrations of succinate with the high protein diets could be explained by more succinate

participating in gluconeogenesis corresponding with the upregulated gene expression of the rate-

limiting enzyme (FBP1) involved in gluconeogenesis in the liver.

7.5 Microbiome profile

Diets have both direct and indirect effects on gastrointestinal function of the host. The indirect effect

are mediated by the influence on the composition and activity of the gut microbiota and gut

environment, and changes in microbiome composition are detectable within 24 hours of initiating

dietary intervention (228). After 8-week dietary intervention, higher alpha diversity of microbiota

and Bacteroidetes:Firmicutes ratios were found in fecal samples of pigs fed high DF diets in Paper

III. Actually, lower alpha diversity and Bacteroidetes:Firmicutes ratios were normally found in the

gut or feces of obese humans and animal models (44, 229, 230). However, a previous study showed

that induction of changes of Firmicutes and Bacteroidetes with high DF diets were very variable

between different studies and it is difficult to link certain phyla to particular diets and obesity (231).

155

Likewise, the genera within each phylum did not always present similar changes as phyla (46). For

instance, the high DF diets induced a relative decrease of Firmicutes but relative enrichment of

butyrate-producing genera belonging to the Firmicutes phylum in Paper II. This could be explained

by a relative decrease of other genera in Firmicutes such as Turicibacter and Streptococcus which

were found to be less abundant with the high DF diets. Therefore, it is more important to analyze the

specific microbiota at genus or specie levels rather than only focusing on phyla.

Substrate availability is a major determinant of microbial community and metabolite profile in the

intestine. For instance, whole plant foods normally have protective effects against gut inflammation,

favoring the growth of beneficial fiber-degrading bacteria in the colon (231). Therefore, specific

dietary interventions can be used to change the type of bacterial groups dominating in the large

intestine. In Paper II, AX-rich diets induced an increase in butyrate production with a concomitant

relative increased abundance of Faecalibacterium, Peptococcus and Blautia genera. This

corresponded with our previous studies (4, 76, 216), and is considered beneficial to maintain a healthy

gut environment and negatively associated with inflammatory bowel disease (215, 232). Additionally,

NSP and the individual constituent sugars clustered together and linked with the same group of

genera Blautia in cecum and colon in Paper II. Of note, the close links between NCP fractions and

concentration of butyrate in caecum (Paper III) explains why they also clustered together and

correlated with the same taxa such as Faecalibacterium. Similar trends were found for NSP and

cellulose that correlated positively with acetate and simultaneously drove them in the same direction

in their association with acetate-producing bacteria such as Ruminococcus.

As mentioned before, antibiotic administration by muscular injection were able to act both

systemically and in the intestines demonstrated by detection of metabolites of antibiotics both in

plasma and gut content (168, 203). Although muscular antibiotic injection did not induce a

pronounced change in degradation pattern, the microbiota composition and SCFA production was

sensitive to antibiotic administration. For instance, the abundance of some bacteria (such as

Streptococcus) that found enriched in the non-antibiotic treated RB group were reduced with

antibiotic treatment (102), which could be linked with the reduced butyrate production. Gut

microbiota plays an important role in maintaining the intestinal immune homeostasis, mediated by

SCFA dependent pathways (220). However, oral antibiotic treatment was reported to disrupt gut

microbiota, decrease SCFA production in the large intestine and increase susceptibility to infections

(220). Moreover, a study showed that penicillin can suppress Gram-positive, anaerobic bacteria

growth such as Streptococcus and Clostridium strains (233). Here we demonstrated that

administration of antibiotics reversed the beneficial effects of DF on intestinal environment by

changing microbial profile and butyrate production.

156

In Paper II, the high protein diets induced minor influence on the microbiota profile wherein some

proteolytic taxa abundance including Peptococcaceae and Clostridium were increased with the high

protein intake and associated with increased propionate (234) and BCFA concentrations in the gut

(148). Other studies have reported that intake of whey protein could improve Lactobacillius (151)

(152) and Bifidobaterium growth in mice models (151), which was not observed in the present

studies. However, it is still inconclusive whether whey protein stimulates growth of beneficial or

pathogenic bacteria in the large intestine of pigs and humans as it can be easily and quickly digested

in the small intestine (158) and is found to induce less genetic damage to colonocytes than meat,

casein or soya (235). We speculate that the remainder of proteins that enter the large intestine were

mainly from cereal protein within the plant cell walls and from fish meals, and they can influence the

gut microbiota in different ways. Individuals consuming plant protein (pea protein extract) exhibited

increased abundance of beneficial taxa Bifidobacterium and Lactobacillus, and decreased abundance

of pathogenic species such as Bacteroides fragilis and Clostridium perfringens (236). However, due

to a high fat content in the diet, the bacteria that were mainly increased in response to a high animal

protein diets were typically bile-tolerant microbiota such as Bacteroides and Clostridia (231). Taken

together, the protein source can be a critical factor for the changes in gut microbiota and the use of

highly digestible protein sources may reduce the growth of pathogenic species. Moreover, it is

possible that a high protein content in the diet combined with a high fat content can elicit different

outcomes compared with a diet with a lower fat content and further studies are still needed.

The effect of combination of high DF and protein intake on microbiota composition is less studied.

In this thesis, the combination of high DF and protein did not show significant changes in microbial

profile compared with other dietary treatments, which was partly in accordance with the absence of

interaction of DF and protein on the intestinal SCFA concentrations. Similar results regarding the

deficiency of interaction between dietary protein concentration and fermentable carbohydrates on the

intestinal microbiota have also been reported by other authors (237). They found that a simultaneous

increase in the amount of fiber into high protein diets resulted in a reduction in the protein

fermentation but with no significant effects on the microbiota counts in feces of pigs. It may indicate

that a high DF content incorporated into a high protein diet possibly could act as a tool to reduce the

risk of detrimental effects of protein fermentation by interfering the formation of deleterious

metabolites and altering microbiota composition.

157

8. Conclusion and perspectives

This PhD thesis presents results that contributes with new knowledge on the effects of DF and protein

on degradation and fermentation processes in the gut and metabolic responses in relation to systemic

health and microbial ecology in the lean and obese pig models. The main findings from the present

thesis are summarized below.

Dietary fibre and protein do not synergistically influence insulin, metabolic or inflammatory

biomarkers in young obese Göttingen Minipigs

- In still-growing minipigs, the weight gain was reduced by high DF diets compared with the

low DF, and was increased with high dietary protein compared with low protein content.

- A high DF content in the diet increased the non-fasting C-peptide concentrations with no

changes in glucose or insulin responses whereas high protein increased fasting glucose levels

with no significant change in insulin.

- High DF diets did not attenuate lipid or inflammatory biomarkers compared to the low DF

while high protein content increased the levels of dyslipidemic markers.

- Feeding high DF diets upregulated adipose expression of gene involved in lipid synthesis

(FASN) while high protein upregulated hepatic expression of gene (FBP1) which is a key

regulator of the gluconeogenesis pathway.

Effects of dietary fibre and protein content on intestinal fibre degradation, short-chain fatty

acid and microbiota composition in a high-fat fructose-rich diet induced obese Göttingen

Minipig model

- AX was partly degraded in the distal small intestine and further gradually fermented until to

mid colon.

- High DF intake reduced the intestinal concentrations of total SCFA and acetate and slightly

increased butyrate production in the large intestine while high protein intake elevated the

concentrations of total SCFA, propionate and BCFA in the gut.

- Proteolytic fermentation was reduced by a high DF content in the diet by presenting a lower

BCFA concentrations and proportions of SCFA in the colon than the low DF diets.

- High DF diets increased alpha diversity of fecal microbiota and the relative abundance of

butyrate-producing bacteria in the cecum, mid colon and feces whereas the protein content

only had limited effect on microbial profile.

- High DF increased circulating butyrate concentration while high protein increased

propionate and decreased succinate concentrations, linked with the effects of DF and protein

on hormone response and metabolic disorders.

158

The role of rye bran and antibiotics on the digestion, fermentation process and short-chain

fatty acid production and absorption of pigs

- Differences in DF composition was the main factor for altered degradation and fermentation

processes in the pigs, wherein AX was largely degraded in the cecum and proximal colon,

whereas the cellulose in the Control diet had a slow and gradual degradation along the large

intestine.

- The RB diets reduced protein and fat digestibility and induced a slower protein degradation

in the colon than the Control diet.

- The cellulose contents in cecal digesta correlated positively with acetate concentrations and

proportions, while AX contents positively associated with butyrate concentrations and

proportions.

- Antibiotic treatment had only a limited effect on nutrient digestion and DF degradation, and

significantly interfered with the butyrate concentrations and proportions in the cecum of the

pigs fed RB diets.

- The AX-enriched NCP fractions that clustered together with butyrate concentrations and

proportions were positively correlated with a group of taxa that coincided with the same taxa

enriched in the RB group without antibiotic treatment, whereas cellulose contents and acetate

concentrations and proportions had the same correlations with taxa that were found enriched

in the Control group.

- The RB diets reduced the absorption of total SCFA, acetate and propionate and was associated

with an improved lipid profile compared with the Control diet.

Overall, the results show that different cereal-based DF sources influence degradation and

fermentation patterns. In our pig models, DF from whole grains and cereal brans have positive effects

on systemic and intestinal environment and is suggested to be consumed on a regular basis to mitigate

obesity-related MetS. On the other hand, a high protein intake from whey, cereal and fish meal did

not show beneficial effects on host health and we found that the combination of high DF and protein

did not have synergistic effects on metabolic biomarkers and intestinal environment in young obese

Göttingen Minipigs. The modulated intestinal degradation and fermentation patterns could be linked

to the potential mechanisms underlying the effects of DF and protein on metabolic health.

The minipigs are still growing due to the ad libitum feeding strategy and also because they are not

full-grown during the dietary intervention. This is one of the limitations of the present thesis that

growing lean and obese pig models were used. This continued growth can mask the effects of dietary

intervention. Further studies conducted in an obese pig model with stable body weight and more

159

severe metabolic disturbances may provide more powerful evidence for the roles of DF and protein

in MetS.

A trend of dyslipidemia and pre-diabetic state development with the high protein feeding was

indicated, contrasting the findings in a parallel human study, where whey protein combined with low

DF improved the lipid profile. Moreover, the upregulated gene expression of FBP1 in the liver with

high protein diets raises the question whether high protein intake will have more adverse effects on

lipid metabolism in a longer term of dietary intervention. However, high fat and high sugar content

need to be taken into account when we interpret the effects of high DF and/or protein. We cannot

exclude that a combination of high fat and high protein consumption generate further detrimental

metabolic disorders compared with a lower fat-based diet. Given that only one source of protein or

fiber was used to increase dietary protein and DF level, it is not clear whether the effects were induced

by their specific source or by the level of DF or protein, which should be identified by future studies.

Changes in intestinal and circulating SCFA concentrations with dietary treatments should be noticed

and they may signal through cell surface receptors, like GPR41, GPR43 and GPR109A, to activate

signaling cascades that control physiological functions (238). It would be interesting to explore the

effects of high DF or protein on the gene expression of SCFA receptors and the gene (glucose-6-

phosphate) related to intestinal gluconeogenesis in cecum and colon. Besides for gluconeogenesis,

the reduced succinate levels in systemic circulation also arises a question that if it can be the results

of increased urinary loss of TCA cycle intermediates with high protein intake. High protein intake

may change plasma amino acid profile such as BCAA which may act as indicators for MetS (210).

Therefore, further studies are needed to explore the role of these metabolites in blood and urine and

confirm if they could act as early biomarkers of MetS. We observed increased BCFA concentrations

and pool size in the colon with the high protein diets, but fewer studies have investigated the role of

BCFA compared with carbohydrate-derived SCFA and further research should focus on elucidating

the mechanisms of action of BCFA in health and disease.

This thesis highlights the importance of the gut microbiota and intestinal metabolites on systemic

health. High DF consumption increased the bacterial diversity and richness of beneficial taxa. In this

way, a new state of ecological homeostasis of the gut microbiota may be achieved with beneficial

implications for host health. As the microbial utilization of DF and protein begins already in the small

intestine, it would be of interest to explore the bacterial profile at this site. Moreover, effects of oral

antibiotics intake on nutrient digestion and fiber degradation are still needed.

160

References 1. Schwartz MW, Seeley RJ, Zeltser LM, Drewnowski A, Ravussin E, Redman LM, et al. Obesity

Pathogenesis: An Endocrine Society Scientific Statement. Endocr Rev. 2017;38(4):267-96.

2. Stanhope KL. Role of fructose-containing sugars in the epidemics of obesity and metabolic

syndrome. Annu Rev Med. 2012;63:329-43.

3. Wei B, Liu Y, Lin X, Fang Y, Cui J, Wan J. Dietary fiber intake and risk of metabolic syndrome:

A meta-analysis of observational studies. Clin Nutr. 2018;37(6 Pt A):1935-42.

4. Nielsen TS, Laerke HN, Theil PK, Sorensen JF, Saarinen M, Forssten S, et al. Diets high in

resistant starch and arabinoxylan modulate digestion processes and SCFA pool size in the large

intestine and faecal microbial composition in pigs. Br J Nutr. 2014;112(11):1837-49.

5. Hur KY, Lee MS. Gut Microbiota and Metabolic Disorders. Diabetes Metab J. 2015;39(3):198-

203.

6. Weyer C, Foley JE, Bogardus C, Tataranni PA, Pratley RE. Enlarged subcutaneous abdominal

adipocyte size, but not obesity itself, predicts Type II diabetes independent of insulin resistance.

Diabetologia 2000;43:1498-506.

7. Brahe LK, Astrup A, Larsen LH. Is butyrate the link between diet, intestinal microbiota and

obesity-related metabolic diseases? Obes Rev. 2013;14(12):950-9.

8. Holmer-Jensen J, Mortensen LS, Astrup A, de Vrese M, Holst JJ, Thomsen C, et al. Acute

differential effects of dietary protein quality on postprandial lipemia in obese non-diabetic

subjects. Nutr Res. 2013;33(1):34-40.

9. Sousa GT, Lira FS, Rosa JC, de Oliveira EP, Oyama LM, Santos RV, et al. Dietary whey protein

lessens several risk factors for metabolic diseases: a review. Lipids Health Dis. 2012;11:67.

10. Mikael Nilsson MS, Anders H Frid, Jens J Holst, and Inger ME Björck. Glycemia and

insulinemia in healthy subjects after lactoseequivalent meals of milk and other food proteins:

the role of plasma amino acids and incretins. Am J Clin Nutr. 2004;80:1246–53.

11. Mikael Nilsson JJH, and Inger ME Björck. Metabolic effects of amino acid mixtures and whey

protein in healthy subjects: studies using glucose-equivalent drinks. Am J Clin Nutr.

2007;85:996-1004.

12. Pal S, Ellis V, Dhaliwal S. Effects of whey protein isolate on body composition, lipids, insulin

and glucose in overweight and obese individuals. Br J Nutr. 2010;104(5):716-23.

13. Nielsen KL, Hartvigsen ML, Hedemann MS, Laerke HN, Hermansen K, Bach Knudsen KE.

Similar metabolic responses in pigs and humans to breads with different contents and

compositions of dietary fibers: a metabolomics study. Am J Clin Nutr. 2014;99(4):941-9.

14. Alberti KG, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato KA, et al. Harmonizing

the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task

Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American

Heart Association; World Heart Federation; International Atherosclerosis Society; and

International Association for the Study of Obesity. Circulation. 2009;120(16):1640-5.

15. GBD 2015 Obesity Collaborators AA, Forouzanfar MH, Reitsma MB, Sur P, Estep K, et al. .

Health Effects of Overweight and Obesity in 195 Countries over 25 Years. The New England

journal of medicine. 2017;377(1):13-27.

16. Afshin A, Forouzanfar MH, Reitsma MB, Sur P, Estep K, Lee A, et al. Health effects of

overweight and obesity in 195 countries over 25 years. New England Journal of Medicine.

2017;377(1):13-27.

17. Bardou M, Barkun AN, Martel M. Obesity and colorectal cancer. Gut. 2013;62(6):933-47.

18. Saklayen MG. The Global Epidemic of the Metabolic Syndrome. Curr Hypertens Rep.

2018;20(2):12.

19. Han TS, Lean ME. A clinical perspective of obesity, metabolic syndrome and cardiovascular

disease. JRSM Cardiovasc Dis. 2016;5:2048004016633371.

161

20. Al-Mashhadi AL, Poulsen CB, von Wachenfeldt K, Robertson AK, Bentzon JF, Nielsen LB, et

al. Diet-Induced Abdominal Obesity, Metabolic Changes, and Atherosclerosis in

Hypercholesterolemic Minipigs. J Diabetes Res. 2018;2018:6823193.

21. Bays HE, Toth PP, Kris-Etherton PM, Abate N, Aronne LJ, Brown WV, et al. Obesity, adiposity,

and dyslipidemia: a consensus statement from the National Lipid Association. J Clin Lipidol.

2013;7(4):304-83.

22. Lazarte J, Hegele RA. Dyslipidemia Management in Adults With Diabetes. Can J Diabetes.

2020;44(1):53-60.

23. Delma J. Nieves MC, Barbara Retzlaff, Carolyn E. Walden, John D. Brunzell, Robert H. Knopp,

and Steven E. Kahn. The Atherogenic Lipoprotein Profile Associated With Obesity and Insulin

Resistance Is Largely Attributable to Intra-Abdominal Fat. Diabetes 2003;52:172–9.

24. Franssen R, Monajemi H, Stroes ES, Kastelein JJ. Obesity and dyslipidemia. Med Clin North

Am. 2011;95(5):893-902.

25. Berman AN, Blankstein R. Optimizing Dyslipidemia Management for the Prevention of

Cardiovascular Disease: a Focus on Risk Assessment and Therapeutic Options. Curr Cardiol

Rep. 2019;21(9):110.

26. Gordon DJ PJ, Garrison RJ, Neaton JD, Castelli WP, Knoke JD, et al. High-density lipoprotein

cholesterol and cardiovascular disease. Four prospective American studies. . Circulation

1989;79(1):8–15.

27. Heilbronn L, Smith SR, Ravussin E. Failure of fat cell proliferation, mitochondrial function and

fat oxidation results in ectopic fat storage, insulin resistance and type II diabetes mellitus. Int J

Obes Relat Metab Disord. 2004;28 Suppl 4:S12-21.

28. van Dam AD, Boon MR, Berbee JFP, Rensen PCN, van Harmelen V. Targeting white, brown

and perivascular adipose tissue in atherosclerosis development. Eur J Pharmacol. 2017;816:82-

92.

29. Akhtar DH, Iqbal U, Vazquez-Montesino LM, Dennis BB, Ahmed A. Pathogenesis of Insulin

Resistance and Atherogenic Dyslipidemia in Nonalcoholic Fatty Liver Disease. J Clin Transl

Hepatol. 2019;7(4):362-70.

30. International Diabetes Federation. IDF Diabetes Atlas tehwdoaO, 2016).

31. Chatterjee S, Khunti K, Davies MJ. Type 2 diabetes. The Lancet. 2017;389(10085):2239-51.

32. Samuel VT, Shulman GI. The pathogenesis of insulin resistance: integrating signaling pathways

and substrate flux. J Clin Invest. 2016;126(1):12-22.

33. Zatterale F, Longo M, Naderi J, Raciti GA, Desiderio A, Miele C, et al. Chronic Adipose Tissue

Inflammation Linking Obesity to Insulin Resistance and Type 2 Diabetes. Front Physiol.

2019;10:1607.

34. Chawla A, Nguyen KD, Goh YP. Macrophage-mediated inflammation in metabolic disease. Nat

Rev Immunol. 2011;11(11):738-49.

35. Tilg H. The role of cytokines in non-alcoholic fatty liver disease. Dig Dis. 2010;28(1):179-85.

36. Tilg H, Moschen AR. Inflammatory mechanisms in the regulation of insulin resistance. Mol

Med. 2008;14(3-4):222-31.

37. Yamaguchi M, Okamura S, Yamaji T, Iwasaki M, Tsugane S, Shetty V, et al. Plasma cytokine

levels and the presence of colorectal cancer. PLoS One. 2019;14(3):e0213602.

38. Younis N, Zarif R, Mahfouz R. Inflammatory bowel disease: between genetics and microbiota.

Mol Biol Rep. 2020.

39. Kirkegaard H, Johnsen NF, Christensen J, Frederiksen K, Overvad K, Tjonneland A.

Association of adherence to lifestyle recommendations and risk of colorectal cancer: a

prospective Danish cohort study. Bmj. 2010;341(oct26 2):c5504-c.

40. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, et al. Cancer incidence and

mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer.

2015;136(5):E359-86.

162

41. Grazioso TP, Brandt M, Djouder N. Diet, Microbiota, and Colorectal Cancer. iScience.

2019;21:168-87.

42. Makki K, Deehan EC, Walter J, Backhed F. The Impact of Dietary Fiber on Gut Microbiota in

Host Health and Disease. Cell Host Microbe. 2018;23(6):705-15.

43. Woting A, Blaut M. The Intestinal Microbiota in Metabolic Disease. Nutrients. 2016;8(4):202.

44. Turnbaugh PJ, Hamady M, Yatsunenko T, Cantarel BL, Duncan A, Ley RE, et al. A core gut

microbiome in obese and lean twins. Nature. 2009;457(7228):480-4.

45. Müller M, Canfora, E. E., Blaak, E. E. Gastrointestinal Transit Time, Glucose Homeostasis and

Metabolic Health: Modulation by Dietary Fibers. Nutrients. 2018;10(3).

46. Magne F, Gotteland M, Gauthier L, Zazueta A, Pesoa S, Navarrete P, et al. The

Firmicutes/Bacteroidetes Ratio: A Relevant Marker of Gut Dysbiosis in Obese Patients?

Nutrients. 2020;12(5).

47. Jones JM. CODEX-aligned dietary fiber definitions help to bridge the ‘fiber gap’. Nutrition

Journal 2014;13:34.

48. Williams BA, Mikkelsen D, Flanagan BM, Gidley MJ. "Dietary fibre": moving beyond the

"soluble/insoluble" classification for monogastric nutrition, with an emphasis on humans and

pigs. J Anim Sci Biotechnol. 2019;10:45.

49. Fardet A. New hypotheses for the health-protective mechanisms of whole-grain cereals: what is

beyond fibre? Nutr Res Rev. 2010;23(1):65-134.

50. Bozzetto L, Costabile G, Della Pepa G, Ciciola P, Vetrani C, Vitale M, et al. Dietary Fibre as a

Unifying Remedy for the Whole Spectrum of Obesity-Associated Cardiovascular Risk.

Nutrients. 2018;10(7).

51. Houdijk JGM, Verstegen MWA, Bosch MW, van Laere KJM. Dietary fructooligosaccharides

and transgalactooligosaccharides can affect fermentation characteristics in gut contents and

portal plasma of growing pigs. Livestock Production Science. 2002;73(2):175-84.

52. Bindels LB, Delzenne NM, Cani PD, Walter J. Towards a more comprehensive concept for

prebiotics. Nat Rev Gastroenterol Hepatol. 2015;12(5):303-10.

53. Karppinen S. Dietary fibre components of rye bran and their fermentation in vitro. Espoo 2003

VTT Publications 500 96 p + app 52 p

https://wwwvttresearchcom/sites/default/files/pdf/publications/2003/P500pdf. 2003.

54. Lærke KEBKaHN. Rye Arabinoxylans: Molecular Structure, Physicochemical Properties and

Physiological Effects in the Gastrointestinal Tract. Cereal Chem. 2010;87:353-62.

55. Bach Knudsen KE, Norskov NP, Bolvig AK, Hedemann MS, Laerke HN. Dietary fibers and

associated phytochemicals in cereals. Mol Nutr Food Res. 2017;61(7).

56. Izydorczyk MS, Biliaderis CG. Cereal arabinoxylans: advances in structure and

physicochemical properties. Carbohydrate Polymers. 1995;28(1):33-48.

57. Hartvigsen ML, Gregersen S, Laerke HN, Holst JJ, Bach Knudsen KE, Hermansen K. Effects

of concentrated arabinoxylan and beta-glucan compared with refined wheat and whole grain rye

on glucose and appetite in subjects with the metabolic syndrome: a randomized study. Eur J Clin

Nutr. 2014;68(1):84-90.

58. Knudsen KEB. Carbohydrate and lignin contents of plant materials used in animal feeding.

Animal Feed Science and Technology 1997:319-38.

59. Knudsen KE. Fiber and nonstarch polysaccharide content and variation in common crops used

in broiler diets. Poult Sci. 2014;93(9):2380-93.

60. Kamal-Eldin A, Laerke HN, Knudsen KE, Lampi AM, Piironen V, Adlercreutz H, et al.

Physical, microscopic and chemical characterisation of industrial rye and wheat brans from the

Nordic countries. Food Nutr Res. 2009;53.

61. Van Craeyveld V, Dornez E, Holopainen U, Selinheimo E, Poutanen K, Delcour JA, and Courtin

CM. Wheat Bran AX Properties and Choice of Xylanase Affect Enzymic Production of Wheat

Bran-Derived Arabinoxylan-Oligosaccharides. Cereal Chem 2010;87(4):283–91.

163

62. Vangsøe CT, Sørensen JF, Bach Knudsen KE. Aleurone cells are the primary contributor to

arabinoxylan oligosaccharide production from wheat bran after treatment with cell wall‐

degrading enzymes. International Journal of Food Science & Technology. 2019;54(10):2847-

53.

63. Mudgil D, Barak S. Composition, properties and health benefits of indigestible carbohydrate

polymers as dietary fiber: a review. Int J Biol Macromol. 2013;61:1-6.

64. Stephen AM, Champ MM, Cloran SJ, Fleith M, van Lieshout L, Mejborn H, et al. Dietary fibre

in Europe: current state of knowledge on definitions, sources, recommendations, intakes and

relationships to health. Nutr Res Rev. 2017;30(2):149-90.

65. Wanders AJ, van den Borne JJ, de Graaf C, Hulshof T, Jonathan MC, Kristensen M, et al. Effects

of dietary fibre on subjective appetite, energy intake and body weight: a systematic review of

randomized controlled trials. Obes Rev. 2011;12(9):724-39.

66. Rebello CJ, O'Neil CE, Greenway FL. Dietary fiber and satiety: the effects of oats on satiety.

Nutr Rev. 2016;74(2):131-47.

67. Kristensen M, Jensen MG. Dietary fibres in the regulation of appetite and food intake.

Importance of viscosity. Appetite. 2011;56(1):65-70.

68. French SJ, Read NW. Effect of guar gum on hunger and satiety after meals of differing fat

content: relationship with gastric emptying. The American Journal of Clinical Nutrition.

1994;59(1):87-91.

69. Warrilow A, Mellor D, McKune A, Pumpa K. Dietary fat, fibre, satiation, and satiety-a

systematic review of acute studies. Eur J Clin Nutr. 2019;73(3):333-44.

70. Freeland KR, Wilson C, Wolever TM. Adaptation of colonic fermentation and glucagon-like

peptide-1 secretion with increased wheat fibre intake for 1 year in hyperinsulinaemic human

subjects. Br J Nutr. 2010;103(1):82-90.

71. Lafond DW, Greaves KA, Maki KC, Leidy HJ, Romsos DR. Effects of two dietary fibers as part

of ready-to-eat cereal (RTEC) breakfasts on perceived appetite and gut hormones in overweight

women. Nutrients. 2015;7(2):1245-66.

72. Weickert MO, Pfeiffer AFH. Impact of Dietary Fiber Consumption on Insulin Resistance and

the Prevention of Type 2 Diabetes. J Nutr. 2018;148(1):7-12.

73. McRorie JW, Fahey GC. A review of gastrointestinal physiology and the mechanisms

underlying the health benefits of dietary fiber: Matching an effective fiber with specific patient

needs. Clinical Nursing Studies. 2013;1(4).

74. Lundin EA, Zhang JX, Lairon D, Tidehag P, Åman P, Adlercreutz H, et al. Effects of meal

frequency and high-fibre rye-bread diet on glucose and lipid metabolism and ileal excretion of

energy and sterols in ileostomy subjects. European Journal Of Clinical Nutrition. 2004;58:1410.

75. Qi X, Al‐Ghazzewi FH, Tester RF. Dietary Fiber, Gastric Emptying, and Carbohydrate

Digestion: A Mini‐Review. Starch - Stärke. 2018;70(9-10).

76. Ingerslev AK, Theil PK, Hedemann MS, Laerke HN, Bach Knudsen KE. Resistant starch and

arabinoxylan augment SCFA absorption, but affect postprandial glucose and insulin responses

differently. Br J Nutr. 2014;111(9):1564-76.

77. Papathanasopoulos A, Camilleri M. Dietary fiber supplements: effects in obesity and metabolic

syndrome and relationship to gastrointestinal functions. Gastroenterology. 2010;138(1):65-72

e1-2.

78. Lu ZX, Walker KZ, Muir JG, O'Dea K. Arabinoxylan fibre improves metabolic control in people

with Type II diabetes. Eur J Clin Nutr. 2004;58(4):621-8.

79. Jenkins DJA, Cyril WCK, Augustin LSA, Martini MC, Axelsen M, Faulkner D, Vidgen E,

Parker T, Lau H, Connelly PW, Teitel J, Singer W, Vandenbroucke AC, Leiter LA, Josse RG.

Effect of Wheat Bran on Glycemic Control and Risk Factors for Cardiovascular Disease in Type

2 Diabetes. Diabetes Care. 2002;25(9):1522–8.

80. Meyer KA, Kushi LH, Jacobs Jr DR, Slavin J, Sellers TA, and Folsom AR. Carbohydrates,

dietary fiber, and incident type 2 diabetes in older women. Am J Clin Nutr. 2000;71:921–30.

164

81. Surampudi P, Enkhmaa B, Anuurad E, Berglund L. Lipid Lowering with Soluble Dietary Fiber.

Curr Atheroscler Rep. 2016;18(12):75.

82. Mark A. Pereira EOR, Katarina Augustsson, Gary E. Fraser, Uri Goldbourt, Berit L. Heitmann,

Goran Hallmans, Paul Knekt, Simin Liu, Pirjo Pietinen, Donna Spiegelman, June Stevens, Jarmo

Virtamo, Walter C. Willett, Alberto Ascherio. Dietary Fiber and Risk of Coronary Heart

Disease: A Pooled Analysis of Cohort Studies. Arch Intern Med 2004;164(4):370-6.

83. Martin O. Weickert aAFHP. Metabolic Effects of Dietary Fiber Consumption and Prevention of

Diabetes. The Journal of Nutrition. 2008;138:439-42.

84. Harland J. 2 - Authorised EU health claims for barley and oat beta-glucans. In: Sadler MJ, editor.

Foods, Nutrients and Food Ingredients with Authorised Eu Health Claims: Woodhead

Publishing; 2014. p. 25-45.

85. Giacco R, Clemente G, Cipriano D, Luongo D, Viscovo D, Patti L, et al. Effects of the regular

consumption of wholemeal wheat foods on cardiovascular risk factors in healthy people.

Nutrition, Metabolism and Cardiovascular Diseases. 2010;20(3):186-94.

86. Garcia AL, Steiniger J, Reich SC, Weickert MO, Harsch I, Machowetz A, et al. Arabinoxylan

fibre consumption improved glucose metabolism, but did not affect serum adipokines in subjects

with impaired glucose tolerance. Horm Metab Res. 2006;38(11):761-6.

87. Giacco R, Costabile G, Della Pepa G, Anniballi G, Griffo E, Mangione A, et al. A whole-grain

cereal-based diet lowers postprandial plasma insulin and triglyceride levels in individuals with

metabolic syndrome. Nutr Metab Cardiovasc Dis. 2014;24(8):837-44.

88. Jenkins DJ, Kendall CW, Vuksan V, Augustin LS, Mehling C, Parker T, et al. Effect of wheat

bran on serum lipids: influence of particle size and wheat protein. J Am Coll Nutr.

1999;18(2):159-65.

89. Nielsen RF. Long-term effects of whey protein and dietary fiber from wheat on markers of

metabolic risk and bone health in subjects with abdominal obesity. PhD Thesis, Aarhus

University Hospital, Denmark. 2019.

90. Lærke HN, Pedersen C, Mortensen MA, Theil PK, Larsen T, Knudsen KEB. Rye bread reduces

plasma cholesterol levels in hypercholesterolaemic pigs when compared to wheat at similar

dietary fibre level. Journal of the Science of Food and Agriculture. 2008;88(8):1385-93.

91. Wong C, Harris PJ, Ferguson LR. Potential Benefits of Dietary Fibre Intervention in

Inflammatory Bowel Disease. Int J Mol Sci. 2016;17(6).

92. Li M, van Esch B, Wagenaar GTM, Garssen J, Folkerts G, Henricks PAJ. Pro- and anti-

inflammatory effects of short chain fatty acids on immune and endothelial cells. Eur J

Pharmacol. 2018;831:52-9.

93. Topping DL, Clifton PM. Short-Chain Fatty Acids and Human Colonic Function: Roles of

Resistant Starch and Nonstarch Polysaccharides. Physiological Reviews. 2001;81(3):1031-64.

94. Zimmerman MA, Singh N, Martin PM, Thangaraju M, Ganapathy V, Waller JL, et al. Butyrate

suppresses colonic inflammation through HDAC1-dependent Fas upregulation and Fas-

mediated apoptosis of T cells. Am J Physiol Gastrointest Liver Physiol. 2012;302(12):G1405-

G15.

95. Klampfer L, Huang J, Sasazuki T, Shirasawa S, Augenlicht L. Inhibition of Interferon γ

Signaling by the Short Chain Fatty Acid Butyrate. Molecular Cancer Research. 2003;1(11):855-

62.

96. Rakvaag E, Fuglsang-Nielsen R, Bach Knudsen KE, Hermansen K, Gregersen S. The

Combination of Whey Protein and Dietary Fiber Does Not Alter Low-Grade Inflammation or

Adipose Tissue Gene Expression in Adults with Abdominal Obesity. Rev Diabet Stud.

2019;15:83-93.

97. Roager HM, Vogt JK, Kristensen M, Hansen LBS, Ibrugger S, Maerkedahl RB, et al. Whole

grain-rich diet reduces body weight and systemic low-grade inflammation without inducing

major changes of the gut microbiome: a randomised cross-over trial. Gut. 2017;0:1-11.

165

98. Kopf JC, Suhr MJ, Clarke J, Eyun SI, Riethoven JM, Ramer-Tait AE, et al. Role of whole grains

versus fruits and vegetables in reducing subclinical inflammation and promoting gastrointestinal

health in individuals affected by overweight and obesity: a randomized controlled trial. Nutr J.

2018;17(1):72.

99. Vitaglione P, Mennella I, Ferracane R, Rivellese AA, Giacco R, Ercolini D, et al. Whole-grain

wheat consumption reduces inflammation in a randomized controlled trial on overweight and

obese subjects with unhealthy dietary and lifestyle behaviors: role of polyphenols bound to

cereal dietary fiber. Am J Clin Nutr. 2015;101(2):251-61.

100. Pollet A, Van Craeyveld V, Van de Wiele T, Verstraete W, Delcour JA, Courtin CM. In vitro

fermentation of arabinoxylan oligosaccharides and low molecular mass arabinoxylans with

different structural properties from wheat (Triticum aestivum L.) bran and psyllium (Plantago

ovata Forsk) seed husk. J Agric Food Chem. 2012;60(4):946-54.

101. Glitsø LV, Gruppen H, Schols HA, Højsgaard S, Sandstro¨m B and Bach Knudsen KE.

Degradation of rye arabinoxylans in the large intestine of pigs. Journal of the Science of Food

and Agriculture. 1999;79:961-9.

102. Bolvig AK, Norskov NP, van Vliet S, Foldager L, Curtasu MV, Hedemann MS, et al. Rye Bran

Modified with Cell Wall-Degrading Enzymes Influences the Kinetics of Plant Lignans but Not

of Enterolignans in Multicatheterized Pigs. J Nutr. 2017;147(12):2220-7.

103. Hald S, Schioldan AG, Moore ME, Dige A, Laerke HN, Agnholt J, et al. Effects of Arabinoxylan

and Resistant Starch on Intestinal Microbiota and Short-Chain Fatty Acids in Subjects with

Metabolic Syndrome: A Randomised Crossover Study. PLoS One. 2016;11(7):e0159223.

104. Francois IE, Lescroart O, Veraverbeke WS, Marzorati M, Possemiers S, Evenepoel P, et al.

Effects of a wheat bran extract containing arabinoxylan oligosaccharides on gastrointestinal

health parameters in healthy adult human volunteers: a double-blind, randomised, placebo-

controlled, cross-over trial. Br J Nutr. 2012;108(12):2229-42.

105. Romani-Perez M, Agusti A, Sanz Y. Innovation in microbiome-based strategies for promoting

metabolic health. Curr Opin Clin Nutr Metab Care. 2017;20(6):484-91.

106. Cummings JH PE, Branch WJ, Naylor CP, Macfarlane GT. . Short chain fatty acids in human

large intestine, portal, hepatic and venous blood. . Gut. 1987;28(10):1221-7.

107. Wong JMW, de Souza R, Kendall CWC, Emam A, and Jenkins DJA. Colonic Health:

Fermentation and Short Chain Fatty Acids. J Clin Gastroenterol. 2006;40:235–43.

108. Kristensen NB, Wu G. Metabolic functions of the porcine liver. Nutritional physiology of pigs

- Online publication ed / Knud Erik Bach Knudsen; Niels Jørgen Kjeldsen; Hanne Damgaard

Poulsen; Bent Borg Jensen Foulum : Videncenter for Svineproduktion. 2012.

109. Robles Alonso V, Guarner F. Linking the gut microbiota to human health. Br J Nutr. 2013;109

Suppl 2:S21-6.

110. den Besten G, van Eunen K, Groen AK, Venema K, Reijngoud DJ, Bakker BM. The role of

short-chain fatty acids in the interplay between diet, gut microbiota, and host energy metabolism.

J Lipid Res. 2013;54(9):2325-40.

111. Illman RJ, Topping DL, McIntosh GH, Trimble RP, Storer GB, Taylor MN, and Cheng BQ.

Hypocholesterolemic effects of dietary propionate: studies in whole animals and perfused rat

liver. Ann Nutr Metab. 1988;32:97-107.

112. Damen B, Cloetens L, Broekaert WF, Francois I, Lescroart O, Trogh I, et al. Consumption of

breads containing in situ-produced arabinoxylan oligosaccharides alters gastrointestinal effects

in healthy volunteers. J Nutr. 2012;142(3):470-7.

113. Henchion M, Hayes M, Mullen AM, Fenelon M, Tiwari B. Future Protein Supply and Demand:

Strategies and Factors Influencing a Sustainable Equilibrium. Foods. 2017;6(7).

114. Pfeiffer AFH, Pedersen E, Schwab U, Riserus U, Aas AM, Uusitupa M, et al. The Effects of

Different Quantities and Qualities of Protein Intake in People with Diabetes Mellitus. Nutrients.

2020;12(2).

166

115. Jahan-Mihan A, Luhovyy BL, El Khoury D, Anderson GH. Dietary proteins as determinants of

metabolic and physiologic functions of the gastrointestinal tract. Nutrients. 2011;3(5):574-603.

116. Berrazaga I, Micard V, Gueugneau M, Walrand S. The Role of the Anabolic Properties of Plant-

versus Animal-Based Protein Sources in Supporting Muscle Mass Maintenance: A Critical

Review. Nutrients. 2019;11(8).

117. Friedman M. Nutritional Value of Proteins from Different Food Sources. A Review. J Agric

Food Chem 1996, 44, 6−29. 1996.

118. Carvalho F, Prazeres AR, Rivas J. Cheese whey wastewater: characterization and treatment. Sci

Total Environ. 2013;445-446:385-96.

119. Bjornshave A, Hermansen K. Effects of dairy protein and fat on the metabolic syndrome and

type 2 diabetes. Rev Diabet Stud. 2014;11(2):153-66.

120. Corrochano AR, Buckin V, Kelly PM, Giblin L. Invited review: Whey proteins as antioxidants

and promoters of cellular antioxidant pathways. J Dairy Sci. 2018;101(6):4747-61.

121. Dullius A, Goettert MI, de Souza CFV. Whey protein hydrolysates as a source of bioactive

peptides for functional foods – Biotechnological facilitation of industrial scale-up. Journal of

Functional Foods. 2018;42:58-74.

122. Foegeding EA, Davis JP. Food protein functionality: A comprehensive approach. Food

Hydrocolloids. 2011;25(8):1853-64.

123. Kareb O, Aider M. Whey and Its Derivatives for Probiotics, Prebiotics, Synbiotics, and

Functional Foods: a Critical Review. Probiotics Antimicrob Proteins. 2019;11(2):348-69.

124. Geraedts MC, Troost FJ, Tinnemans R, Soderholm JD, Brummer RJ, Saris WH. Release of

satiety hormones in response to specific dietary proteins is different between human and murine

small intestinal mucosa. Ann Nutr Metab. 2010;56(4):308-13.

125. Louise Kjølbæk LBS, Nadja Buus Søndertoft, Carrie Klestrup Rasmussen, Janne Kunchel

Lorenzen, Anja Serena, Arne Astrup, and Lesli Hingstrup Larsen. Protein supplements after

weight loss do not improve weight maintenance compared with recommended dietary protein

intake despite beneficial effects on appetite sensation and energy expenditure: a randomized,

controlled, double-blinded trial. Am J Clin Nutr 2017;106:684–97.

126. Hall WL, Millward DJ, Long SJ, Morgan LM. Casein and whey exert different effects on plasma

amino acid profiles, gastrointestinal hormone secretion and appetite. Br J Nutr. 2003;89(2):239-

48.

127. Veldhorst MAB, Nieuwenhuizen AG, Hochstenbach-Waelen A, van Vught AJAH, Westerterp

KR, Engelen MPKJ, et al. Dose-dependent satiating effect of whey relative to casein or soy.

Physiology & Behavior. 2009;96(4):675-82.

128. Jakubowicz D, Froy O. Biochemical and metabolic mechanisms by which dietary whey protein

may combat obesity and Type 2 diabetes. J Nutr Biochem. 2013;24(1):1-5.

129. Jakubowicz D, Wainstein J, Landau Z, Ahren B, Barnea M, Bar-Dayan Y, et al. High-energy

breakfast based on whey protein reduces body weight, postprandial glycemia and HbA1C in

Type 2 diabetes. J Nutr Biochem. 2017;49:1-7.

130. Bjornshave A, Holst JJ, Hermansen K. A pre-meal of whey proteins induces differential effects

on glucose and lipid metabolism in subjects with the metabolic syndrome: a randomised cross-

over trial. Eur J Nutr. 2019;58(2):755-64.

131. Gannon MC, Nuttall FQ, Saeed A, Jordan K, and Hoover H. An increase in dietary protein

improves the blood glucose response in persons with type 2 diabetes. Am J Clin Nutr.

2003;78:734-41.

132. Dong JY, Zhang ZL, Wang PY, Qin LQ. Effects of high-protein diets on body weight, glycaemic

control, blood lipids and blood pressure in type 2 diabetes: meta-analysis of randomised

controlled trials. Br J Nutr. 2013;110(5):781-9.

133. Yu Z, Nan F, Wang LY, Jiang H, Chen W, Jiang Y. Effects of high-protein diet on glycemic

control, insulin resistance and blood pressure in type 2 diabetes: A systematic review and meta-

analysis of randomized controlled trials. Clin Nutr. 2020;39(6):1724-34.

167

134. Frid AH, Nilsson M, Holst JJ, and Björck IME. Effect of whey on blood glucose and insulin

responses to composite breakfast and lunch meals in type 2 diabetic subjects. Am J Clin Nutr.

2005;83:69-75.

135. Akhavan T, Luhovyy BL, Panahi S, Kubant R, Brown PH, Anderson GH. Mechanism of action

of pre-meal consumption of whey protein on glycemic control in young adults. J Nutr Biochem.

2014;25(1):36-43.

136. Bjornshave A, Holst JJ, Hermansen K. Pre-Meal Effect of Whey Proteins on Metabolic

Parameters in Subjects with and without Type 2 Diabetes: A Randomized, Crossover Trial.

Nutrients. 2018;10(2).

137. Rietman A, Schwarz J, Tome D, Kok FJ, Mensink M. High dietary protein intake, reducing or

eliciting insulin resistance? Eur J Clin Nutr. 2014;68(9):973-9.

138. Adams RL, Broughton KS. Insulinotropic Effects of Whey: Mechanisms of Action, Recent

Clinical Trials, and Clinical Applications. Ann Nutr Metab. 2016;69(1):56-63.

139. Richter CK, Skulas-Ray AC, Champagne CM, Kris-Etherton PM. Plant protein and animal

proteins: do they differentially affect cardiovascular disease risk? Adv Nutr. 2015;6(6):712-28.

140. David JA Jenkins CWK, Edward Vidgen, Livia SA Augustin, Marjan van Erk, Anouk Geelen,

Tina Parker, Dorothea Faulkner, Vladimir Vuksan, Robert G Josse, Lawrence A Leiter, and

Philip W Connelly. High-protein diets in hyperlipidemia: effect of wheat gluten on serum lipids,

uric acid, and renal function. Am J Clin Nutr. 2001;74:57–63.

141. Mortensen LS, Hartvigsen ML, Brader LJ, Astrup A, Schrezenmeir J, Holst JJ, et al. Differential

effects of protein quality on postprandial lipemia in response to a fat-rich meal in type 2 diabetes:

comparison of whey, casein, gluten, and cod protein. Am J Clin Nutr. 2009;90(1):41-8.

142. Rakvaag E, Fuglsang-Nielsen R, Bach Knudsen KE, Landberg R, Johannesson Hjelholt A,

Søndergaard E, et al. Whey Protein Combined with Low Dietary Fiber Improves Lipid Profile

in Subjects with Abdominal Obesity: A Randomized, Controlled Trial. Nutrients. 2019;11(9).

143. Zhang JW, Tong X, Wan Z, Wang Y, Qin LQ, Szeto IM. Effect of whey protein on blood lipid

profiles: a meta-analysis of randomized controlled trials. Eur J Clin Nutr. 2016;70(8):879-85.

144. Chen Q, Reimer RA. Dairy protein and leucine alter GLP-1 release and mRNA of genes

involved in intestinal lipid metabolism in vitro. Nutrition. 2009;25(3):340-9.

145. Zhang F, Zhao S, Yan W, Xia Y, Chen X, Wang W, et al. Branched Chain Amino Acids Cause

Liver Injury in Obese/Diabetic Mice by Promoting Adipocyte Lipolysis and Inhibiting Hepatic

Autophagy. EBioMedicine. 2016;13:157-67.

146. Essam M Hamad SHT, Abdel-Gawad I Abou Dawood, Mahmoud Z Sitohy and Mahmoud

Abdel-Hamid. Protective effect of whey proteins against nonalcoholic fatty liver in rats. Lipids

in Health and Disease. 2011;10(57):http://www.lipidworld.com/content/10/1/57.

147. Portune KJ, Beaumont M, Davila A-M, Tomé D, Blachier F, Sanz Y. Gut microbiota role in

dietary protein metabolism and health-related outcomes: The two sides of the coin. Trends in

Food Science & Technology. 2016;57:213-32.

148. Diether NE, Willing BP. Microbial Fermentation of Dietary Protein: An Important Factor in

Diet(-)Microbe(-)Host Interaction. Microorganisms. 2019;7(1).

149. Lopez-Legarrea P, de la Iglesia R, Abete I, Navas-Carretero S, Martinez JA, Zulet MA. The

protein type within a hypocaloric diet affects obesity-related inflammation: the RESMENA

project. Nutrition. 2014;30(4):424-9.

150. Batista MA, Campos NCA, Silvestre MPC, Yildiz F. Whey and protein derivatives:

Applications in food products development, technological properties and functional effects on

child health. Cogent Food & Agriculture. 2018;4(1).

151. Sprong RC, Schonewille AJ, van der Meer R. Dietary cheese whey protein protects rats against

mild dextran sulfate sodium-induced colitis: role of mucin and microbiota. J Dairy Sci.

2010;93(4):1364-71.

168

152. McAllan L, Skuse P, Cotter PD, O'Connor P, Cryan JF, Ross RP, et al. Protein quality and the

protein to carbohydrate ratio within a high fat diet influences energy balance and the gut

microbiota in C57BL/6J mice. PLoS One. 2014;9(2):e88904.

153. Pal S, Ellis V. The chronic effects of whey proteins on blood pressure, vascular function, and

inflammatory markers in overweight individuals. Obesity (Silver Spring). 2010;18(7):1354-9.

154. Windey K, De Preter V, Verbeke K. Relevance of protein fermentation to gut health. Mol Nutr

Food Res. 2012;56(1):184-96.

155. Rist VT, Weiss E, Eklund M, Mosenthin R. Impact of dietary protein on microbiota composition

and activity in the gastrointestinal tract of piglets in relation to gut health: a review. Animal.

2013;7(7):1067-78.

156. Aguirre M, Eck A, Koenen ME, Savelkoul PH, Budding AE, Venema K. Diet drives quick

changes in the metabolic activity and composition of human gut microbiota in a validated in

vitro gut model. Res Microbiol. 2016;167(2):114-25.

157. Liu X, Blouin JM, Santacruz A, Lan A, Andriamihaja M, Wilkanowicz S, et al. High-protein

diet modifies colonic microbiota and luminal environment but not colonocyte metabolism in the

rat model: the increased luminal bulk connection. Am J Physiol Gastrointest Liver Physiol.

2014;307(4):G459-70.

158. Tranberg B, Hellgren LI, Lykkesfeldt J, Sejrsen K, Jeamet A, Rune I, et al. Whey protein reduces

early life weight gain in mice fed a high-fat diet. PLoS One. 2013;8(8):e71439.

159. Gilani GS, Sepehr E. Protein Digestibility and Quality in Products Containing Antinutritional

Factors Are Adversely Affected by Old Age in Rats. The Journal of Nutrition. 2003;133(1):220–

5.

160. Bingham SA. High-meat diets and cancer risk. Proc Nutr Soc. 1999;58(2):243-8.

161. Russell WR, Gratz SW, Duncan SH, Holtrop G, Ince J, Scobbie L, et al. High-protein, reduced-

carbohydrate weight-loss diets promote metabolite profiles likely to be detrimental to colonic

health. Am J Clin Nutr. 2011;93(5):1062-72.

162. George T Macfarlane SM. Bacteria, Colonic Fermentation, and Gastrointestinal Health. Journal

of AOAC INTERNATIONAL. 2012;95(1):50–60.

163. Pieper R, Kroger S, Richter JF, Wang J, Martin L, Bindelle J, et al. Fermentable fiber ameliorates

fermentable protein-induced changes in microbial ecology, but not the mucosal response, in the

colon of piglets. J Nutr. 2012;142(4):661-7.

164. Cloetens L, De Preter V, Swennen K, Broekaert WF, Courtin CM, Delcour JA, et al. Dose-

response effect of arabinoxylooligosaccharides on gastrointestinal motility and on colonic

bacterial metabolism in healthy volunteers. J Am Coll Nutr. 2008;27(4):512-8.

165. Cloetens L, Broekaert WF, Delaedt Y, Ollevier F, Courtin CM, Delcour JA, et al. Tolerance of

arabinoxylan-oligosaccharides and their prebiotic activity in healthy subjects: a randomised,

placebo-controlled cross-over study. Br J Nutr. 2010;103(5):703-13.

166. Curtasu MV. Obesity and metabolic syndrome in miniature pigs as models for human disease –

metabolic changes in response to ad libitum feeding of high-fat-high-carbohydrate diets. PhD

Thesis 2019:Aahus University, Denmark.

167. Rakvaag E. Effects of whey protein and wheat bran on risk factors associated with the metabolic

syndrome in adults with abdominal obesity. PhD Thesis, Aarhus University Hospital, Denmark.

2019.

168. Sørensen AKB. Plant and enterolignans: absorption and kinetics as influenced by dietary factors

and antibiotics. PhD thesis, Aarhus University, Denmark. 2016.

169. Hsu MC, Wang ME, Jiang YF, Liu HC, Chen YC, Chiu CH. Long-term feeding of high-fat plus

high-fructose diet induces isolated impaired glucose tolerance and skeletal muscle insulin

resistance in miniature pigs. Diabetol Metab Syndr. 2017;9:81.

170. Heinritz SN, Mosenthin R, Weiss E. Use of pigs as a potential model for research into dietary

modulation of the human gut microbiota. Nutr Res Rev. 2013;26(2):191-209.

169

171. Gonzalez LM, Moeser AJ, Blikslager AT. Porcine models of digestive disease: the future of

large animal translational research. Transl Res. 2015;166(1):12-27.

172. Guilloteau P, Zabielski R, Hammon HM, Metges CC. Nutritional programming of

gastrointestinal tract development. Is the pig a good model for man? Nutr Res Rev.

2010;23(1):4-22.

173. Kararli TT. COMPARISON OF THE GASTROINTESTINAL ANATOMY, PHYSIOLOGY,

AND BIOCHEMISTRY OF HUMANS AND COMMONLY USED LABORATORY

ANIMALS. BIOPHARMACEUTICS & DRUG DISPOSITION. 1995;16:351-80.

174. Roura E, Koopmans SJ, Lalles JP, Le Huerou-Luron I, de Jager N, Schuurman T, et al. Critical

review evaluating the pig as a model for human nutritional physiology. Nutr Res Rev.

2016;29(1):60-90.

175. Renner S, Blutke A, Dobenecker B, Dhom G, Muller TD, Finan B, et al. Metabolic syndrome

and extensive adipose tissue inflammation in morbidly obese Gottingen minipigs. Mol Metab.

2018;16:180-90.

176. Johansen T, Hansen H, Richelsen B, Malmlöf R. The Obese Gottingen Minipig as a Model of

the Metabolic Syndrome: Dietary Effects on Obesity, Insulin Sensitivity, and Growth Hormone

Profile2001. 150-5 p.

177. Spurlock ME, Gabler NK. The Development of Porcine Models of Obesity and the Metabolic

Syndrome. The Journal of Nutrition. 2008;138:397–402.

178. Thue Johansen HSH, Bjørn Richelsen, and Kjell Malmlöf. The Obese Göttingen Minipig as a

Model of the Metabolic Syndrome: Dietary Effects on Obesity, Insulin Sensitivity, and Growth

Hormone Profile. Comparative Medicine. 2001;51(2):150-5.

179. Kirsten Raun PvV, and Lotte Bjerre Knudsen. Liraglutide, a Once-daily Human Glucagon-like

Peptide-1 Analog, Minimizes Food Intake in Severely Obese Minipigs. OBESITY. 2007;15.

180. Bollen PJA, Madsen LW, Meyer OA, Ritskes-Hoitinga J. Growth differences of male and

female Göttingen minipigs during ad libitum feeding: a pilot study. Laboratory Animals

2005;39:80–93.

181. Christoffersen B, Golozoubova V, Pacini G, Svendsen O, Raun K. The Young Göttingen

Minipig as a Model of Childhood and Adolescent Obesity: Influence of Diet and Gender.

Obesity. 2012.

182. Bergen WG, Mersmann HJ. Comparative Aspects of Lipid Metabolism: Impact on

Contemporary Research and Use of Animal Models. The Journal of Nutrition. 2005;135:2499–

502.

183. Olsen Alstrup AK, Larsen LF, Bladbjerg E-M, Hansen AK, Jespersen J.and Marckmann P. A

high-fat meal does not activate blood coagulation factor VII in minipigs. Blood Coagulation and

Fibrinolysis. 2001;12:117±22.

184. Himes JH. Challenges of accurately measuring and using BMI and other indicators of obesity in

children. Pediatrics. 2009;124 Suppl 1:S3-22.

185. Gurunathan U, Myles PS. Limitations of body mass index as an obesity measure of perioperative

risk. Br J Anaesth. 2016;116(3):319-21.

186. Swindle MM, Makin A, Herron AJ, Clubb FJ, Jr., Frazier KS. Swine as models in biomedical

research and toxicology testing. Vet Pathol. 2012;49(2):344-56.

187. Du ZQ, Fan B, Zhao X, Amoako R, Rothschild MF. Association analyses between type 2

diabetes genes and obesity traits in pigs. Obesity (Silver Spring). 2009;17(2):323-9.

188. Organization. WH. Waist Circumference and Waist-Hip Ratio: Report of a WHO Expert

Consultation Available from

http://apps.who.int/iris/bitstream/10665/44583/1/9789241501491_eng.pdf.

189. Sebert SP, Lecannu G, Kozlowski F, Siliart B, Bard JM, Krempf M, et al. Childhood obesity

and insulin resistance in a Yucatan mini-piglet model: putative roles of IGF-1 and muscle PPARs

in adipose tissue activity and development. Int J Obes (Lond). 2005;29(3):324-33.

170

190. Anwar Borai CLaGAAF. The biochemical assessment of insulin resistanc. Annals of Clinical

Biochemistry. 2007;44:324–42.

191. Vogeser M, Konig D, Frey I, Predel HG, Parhofer KG, Berg A. Fasting serum insulin and the

homeostasis model of insulin resistance (HOMA-IR) in the monitoring of lifestyle interventions

in obese persons. Clin Biochem. 2007;40(13-14):964-8.

192. Muniyappa R, Lee S, Chen H, Quon MJ. Current approaches for assessing insulin sensitivity

and resistance in vivo: advantages, limitations, and appropriate usage. Am J Physiol Endocrinol

Metab. 2008;294(1):E15-26.

193. Anwar Borai CL, Ibrahim Kaddam and Gordon Ferns. Selection of the appropriate method for

the assessment of insulin resistance. Medical Research Methodology 2011;11:158.

194. Chang AM, Smith MJ, Bloem CJ, Galecki AT, Halter JB, Supiano MA. Limitation of the

homeostasis model assessment to predict insulin resistance and beta-cell dysfunction in older

people. J Clin Endocrinol Metab. 2006;91(2):629-34.

195. Saisho Y. Postprandial C-peptide Index: The Best Marker of Beta Cell Function? International

Journal of Diabetes & Clinical Diagnosis. 2014;1(1).

196. Tara M. Wallace JCL, and David R. Matthews. Use and Abuse of HOMA Modeling. Diabetes

Care. 2004;27(6):1487-95.

197. Peker N, Garcia-Croes S, Dijkhuizen B, Wiersma HH, van Zanten E, Wisselink G, et al. A

Comparison of Three Different Bioinformatics Analyses of the 16S-23S rRNA Encoding Region

for Bacterial Identification. Front Microbiol. 2019;10:620.

198. Ingerslev AK. The impact of short-chain fatty acids on metabolic responses - Studies in pigs fed

diets with contrasting sources and levels of dietary fibres. PhD Thesis. 2015.

199. Sakata T. Pitfalls in short-chain fatty acid research: A methodological review. Anim Sci J.

2019;90(1):3-13.

200. Knud Erik Bach Knudsen AS, Anna Kirstin Bjørnbak Kjær, Henry Jørgensen, and Ricarda

Engberg. Rye Bread Enhances the Production and Plasma Concentration of Butyrate but Not the

Plasma Concentrations of Glucose and Insulin in Pigs. J Nutr 2005;135:1696–704.

201. K.E. Bach Knudsen AS, H. Jørgensen, J.L. Peñalvo and H. Adlercreutz. Rye and other natural

cereal fibres enhance the production and plasma concentrations of enterolactone and butyrate.

Dietary fibre components and functions. 2007;Wageningen Academic Publishers:219-33.

202. Kristensen NB, Norgaard JV, Wamberg S, Engbaek M, Fernandez JA, Zacho HD, et al.

Absorption and metabolism of benzoic acid in growing pigs. J Anim Sci. 2009;87(9):2815-22.

203. Bolvig AK, Norskov NP, Hedemann MS, Foldager L, McCarthy-Sinclair B, Marco ML, et al.

Effect of Antibiotics and Diet on Enterolactone Concentration and Metabolome Studied by

Targeted and Nontargeted LC-MS Metabolomics. J Proteome Res. 2017;16(6):2135-50.

204. Peter B, De Rijk EP, Zeltner A, Emmen HH. Sexual Maturation in the Female Gottingen

Minipig. Toxicol Pathol. 2016;44(3):482-5.

205. Fisher KD, Scheffler TL, Kasten SC, Reinholt BM, van Eyk GR, Escobar J, et al. Energy dense,

protein restricted diet increases adiposity and perturbs metabolism in young, genetically lean

pigs. PLoS One. 2013;8(8):e72320.

206. Calder PC, Ahluwalia N, Brouns F, Buetler T, Clement K, Cunningham K, et al. Dietary factors

and low-grade inflammation in relation to overweight and obesity. Br J Nutr. 2011;106 Suppl

3:S5-78.

207. Power-Grant O, Bruen C, Brennan L, Giblin L, Jakeman P, FitzGerald RJ. In vitro bioactive

properties of intact and enzymatically hydrolysed whey protein: targeting the enteroinsular axis.

Food Funct. 2015;6(3):972-80.

208. Bowen J, Noakes M, Clifton PM. Appetite hormones and energy intake in obese men after

consumption of fructose, glucose and whey protein beverages. Int J Obes (Lond).

2007;31(11):1696-703.

171

209. Pal S, Radavelli-Bagatini S, Hagger M, Ellis V. Comparative effects of whey and casein proteins

on satiety in overweight and obese individuals: a randomized controlled trial. Eur J Clin Nutr.

2014;68(9):980-6.

210. Polakof S, Remond D, David J, Dardevet D, Savary-Auzeloux I. Time-course changes in

circulating branched-chain amino acid levels and metabolism in obese Yucatan minipig.

Nutrition. 2018;50:66-73.

211. Chung ST, Hsia DS, Chacko SK, Rodriguez LM, Haymond MW. Increased gluconeogenesis in

youth with newly diagnosed type 2 diabetes. Diabetologia. 2015;58(3):596-603.

212. Weickert MO, Roden M, Isken F, Hoffmann D, Nowotny P, Osterhoff M, et al. Effects of

supplemented isoenergetic diets differing in cereal fiber and protein content on insulin sensitivity

in overweight humans. Am J Clin Nutr. 2011;94(2):459-71.

213. Sandeep Bansal JEB, Nader Rifai, Samia Mora, Frank M. Sacks, Paul M Ridker. Fasting

Compared With Nonfasting Triglycerides and Risk of Cardiovascular Events in Women. JAMA.

2007;298(3):309-16.

214. Kasprzak MM, Laerke HN, Knudsen KE. Effects of isolated and complex dietary fiber matrices

in breads on carbohydrate digestibility and physicochemical properties of ileal effluent from

pigs. J Agric Food Chem. 2012;60(51):12469-76.

215. Tannock GW, Liu Y. Guided dietary fibre intake as a means of directing short-chain fatty acid

production by the gut microbiota. Journal of the Royal Society of New Zealand. 2019:1-22.

216. Le Gall M, Serena A, Jorgensen H, Theil PK, Bach Knudsen KE. The role of whole-wheat grain

and wheat and rye ingredients on the digestion and fermentation processes in the gut--a model

experiment with pigs. Br J Nutr. 2009;102(11):1590-600.

217. Carpita NC, Gibeaut DM. Structural models of primary cell walls in flowering plants:

consistency of molecular structure with the physical properties of the walls during growth. The

Plant Journal. 1993;3(1):1-30.

218. Olejnik K, Skalski B, Stanislawska A, Wysocka-Robak A. Swelling properties and generation

of cellulose fines originating from bleached kraft pulp refined under different operating

conditions. Cellulose. 2017;24(9):3955-67.

219. Xavier-Santos D, Bedani R, Lima ED, Saad SMI. Impact of probiotics and prebiotics targeting

metabolic syndrome. Journal of Functional Foods. 2020;64.

220. Scott NA, Andrusaite A, Andersen P, Lawson M, Alcon-Giner C, Leclaire C, et al. Antibiotics

induce sustained dysregulation of intestinal T cell immunity by perturbing macrophage

homeostasis. Sci Transl Med. 2018;10(464).

221. El Hage R, Hernandez-Sanabria E, Calatayud Arroyo M, Props R, Van de Wiele T. Propionate-

Producing Consortium Restores Antibiotic-Induced Dysbiosis in a Dynamic in vitro Model of

the Human Intestinal Microbial Ecosystem. Front Microbiol. 2019;10:1206.

222. Zhao J, Bai Y, Tao S, Zhang G, Wang J, Liu L, et al. Fiber-rich foods affected gut bacterial

community and short-chain fatty acids production in pig model. Journal of Functional Foods.

2019;57:266-74.

223. G.T. Macfarlane G.R. Gibson E. Beatty J.H. Cummings. Estimation of short-chain fatty acid

production from protein by human intestinal bacteria based on branched-chain fatty acid

measurements. FEMS Microbiology Ecology. 1992;101:81-8.

224. Beaulieu KE, McBurney MI. Changes in Pig Serum Lipids, Nutrient Digestibility and Sterol

Excretion during Cecal Infusion of Propionate. The Journal of Nutrition. 1992;122(2):241-5.

225. Bloemen JG, Venema K, van de Poll MC, Olde Damink SW, Buurman WA, Dejong CH. Short

chain fatty acids exchange across the gut and liver in humans measured at surgery. Clin Nutr.

2009;28(6):657-61.

226. Thomas M.S. Wolever RGJ, Lawrence A. Leiter, and Jean-Louis Chiasson. Time of Day and

Glucose Tolerance Status Affect Serum Short-Chain Fatty Acid Concentrations in Humans.

Metabolism. 1997;46(7):805-11.

172

227. Müller M, Hernández MAG, Goossens GH, Reijnders D, Holst JJ, Jocken JWE, et al. Circulating

but not faecal short-chain fatty acids are related to insulin sensitivity, lipolysis and GLP-1

concentrations in humans. Scientific Reports. 2019;9(1).

228. Wu GD, Chen J, Hoffmann C, Bittinger K, Chen YY, Keilbaugh SA, et al. Linking long-term

dietary patterns with gut microbial enterotypes. Science. 2011;334(6052):105-8.

229. Rivera-Piza A, Lee S-J. Effects of dietary fibers and prebiotics in adiposity regulation via

modulation of gut microbiota. Applied Biological Chemistry. 2020;63(1).

230. Lee CJ, Sears CL, Maruthur N. Gut microbiome and its role in obesity and insulin resistance.

Ann N Y Acad Sci. 2020;1461(1):37-52.

231. Tomova A, Bukovsky I, Rembert E, Yonas W, Alwarith J, Barnard ND, et al. The Effects of

Vegetarian and Vegan Diets on Gut Microbiota. Front Nutr. 2019;6:47.

232. Lavelle A, Sokol H. Gut microbiota-derived metabolites as key actors in inflammatory bowel

disease. Nat Rev Gastroenterol Hepatol. 2020.

233. Brook I, Wexler HM, Goldstein EJ. Antianaerobic antimicrobials: spectrum and susceptibility

testing. Clin Microbiol Rev. 2013;26(3):526-46.

234. Muller N, Worm P, Schink B, Stams AJ, Plugge CM. Syntrophic butyrate and propionate

oxidation processes: from genomes to reaction mechanisms. Environ Microbiol Rep.

2010;2(4):489-99.

235. Toden S, Bird AR, Topping DL, Conlon MA. Differential effects of dietary whey, casein and

soya on colonic DNA damage and large bowel SCFA in rats fed diets low and high in resistant

starch. British Journal of Nutrition. 2007;97(3):535-43.

236. Singh RK, Chang HW, Yan D, Lee KM, Ucmak D, Wong K, et al. Influence of diet on the gut

microbiome and implications for human health. J Transl Med. 2017;15(1):73.

237. Hermes RG, Molist F, Ywazaki M, Nofrarias M, Gomez de Segura A, Gasa J, et al. Effect of

dietary level of protein and fiber on the productive performance and health status of piglets. J

Anim Sci. 2009;87(11):3569-77.

238. Sivaprakasam S, Prasad PD, Singh N. Benefits of short-chain fatty acids and their receptors in

inflammation and carcinogenesis. Pharmacol Ther. 2016;164:144-51.