Bone Marrow Cells Reduce Fibrogenesis and Enhance Regeneration in Fibrotic Rat Liver

Journal of Surgical Research 169, e15–e26 (2011)doi:10.1016/j.jss.2010.03.023

Bone Marrow Cells Reduce Fibrogenesis and Enhance Regeneration

in Fibrotic Rat Liver

Cheuk-Kwan Sun, M.D., Ph.D.,*,2,3 Chih-Hung Chen, M.D.,{,2 Ying-Hsien Kao, Ph.D.,** Chun-Man Yuen, M.D.,†Jiunn-Jye Sheu, M.D.,‡ Fan-Yen Lee, M.D.,‡ Yen-Ta Chen, M.D.,§ Chia-Te Kung, M.D.,k

and Hon-Kan Yip, M.D.#,1

*Division of General Surgery; †Division of Neurosurgery; ‡Division of Cardiovascular Surgery; §Division of Urology, Department ofSurgery; kDepartment of Emergency Medicine; {Division of Gastroenterology; #Division of Cardiology, Department of Internal Medicine,Chang Gung Memorial Hospital, Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung Hsien, Taiwan,

R.O.C; and **Department of Medical Research, E-DA Hospital, I-Shou University, Kaohsiung Hsien, Taiwan, Republic of China

Submitted for publication December 9, 2009

Background. This study aimed at investigating thecellular and molecular impacts of bone marrow-derived mononuclear cells (BMCs) on regeneration offibrotic liver in rats.

Materials and Methods. Thirty male adult Fisherrats were randomized into three groups: group 1 (nor-mal controls, n [ 10); group 2 (carbon tetrachloride-induced liver fibrosis, n [ 10), and group 3 (liver fibrosiswith portal venous infusion of autologous BMCs, 1 3 106,n [ 10).After7-dculturing,BMCswerecharacterizedbyflow cytometry.Groups2 and3 received BMC aspirationthrough bilateral femurs 5 d before hepatectomy. Allanimals received 70% hepatectomy, whereas only group3 receivedabolus of intra-portal BMC infusion48 h afterhepatectomy. Liver-to-body weight ratio, degree offibrosis (Masson trichrome staining), oxidative stress,peroxisome proliferator activated receptor-g coactiva-tor-1a (PGC-1a), collagen I, tumor necrosis factor-a (TNF-a), and transforming growth factor- b (TGF-b)expressions were analyzed 14 d after hepatectomy.Immunohistochemical staining for albumin, a-smoothmuscle actin, and CD31 was performed for tracingcellular differentiation.

Results. Cellular phenotypes typical of hepatocyte,hepatic stellate cell (HSC), and endothelial cell wereidentified in the engrafted BMCs. Liver-to-body weightratio was enhanced with PGC-1a significantly pre-

1 To whom correspondence and reprint requests should beaddressed at Division of Cardiology, Department of Internal Medicine,Chang Gung Memorial Hospital, 123 Ta-Pei Road, Niao Sung Hsiang,Kaohsiung Hsien, 83301, Taiwan. E-mail: [email protected]

2 These two authors contributed equally to this work.3 Present address: Department of Emergency Medicine, E-Da Hos-

pital, I-Shou University, Kaohsiung, Taiwan.

e15

served, whereas oxidative index, collagen I, a-SMA,TNF-a, and TGF-b expressions were all decreased ingroup 3 compared with group 2 (all P < 0.05).

Conclusions. This study demonstrated a positiveimpact of intra-portal BMC infusion in enhancingregeneration and reducing fibrosis of the regeneratingfibrotic liver in rats through suppressing HSC activa-tion and inflammatory cytokine expressions, preserv-ing mitochondrial function and reducing oxidativestress. � 2011 Elsevier Inc. All rights reserved.

Key Words: bone marrow cells; hepatic fibrosis;regeneration; oxidative stress; mitochondrial function.

INTRODUCTION

The participation of bone marrow-derived stem cellsin liver regeneration remains controversial. Whileearly studies supported a positive role of bone marrowcells in hepatic regeneration [1, 2], a recent studyusing bone marrow-irradiated animals demonstrateda lack of tangible in vivo evidence for the differentiationof bone marrow-derived cellular elements into ovalcells, known progenitors of hepatocytes [3]. Similarcontroversies exist in the therapeutic role of bonemarrow cells in chronic liver injury models of cirrhosis.Although some studies failed to demonstrate a benefi-cial role of bone marrow cells in the process of fibrogen-esis [4, 5], there is a body of recent evidence showing thetherapeutic potential of bone marrow-derived mesen-chymal stem cell (MSC) treatment in animal modelsof liver fibrosis [6, 7]. Moreover, promising resultsfrom recent random clinical trials have also arousedthe interest in the feasibility of applying autologous

0022-4804/$36.00� 2011 Elsevier Inc. All rights reserved.

mailto:[email protected]

JOURNAL OF SURGICAL RESEARCH: VOL. 169, NO. 1, JULY 2011e16

stem cells in the treatment of patients with hepaticinsufficiency [8–11]. The mechanisms underlying theobservations in those studies, however, remain unclear.

By administrating autologous bone marrow-derivedmononuclear cells (BMCs) through the portal vein ofthe carbon tetrachloride-induced hepatic fibrosis after70% hepatectomy in the rat, the present study aimedat addressing the following key issues on the participa-tion of BMCs in the regeneration process of the fibroticliver: (1) the fates of the infused cells; (2) the impact ofBMCs infusion on the rate of regeneration; (3) the effectof BMC infusion on the extent of fibrosis in the regener-ating liver; (4) the change in oxidative stress andmitochondrial function after BMC administration;(5) the impact of BMC infusion on hepatocyte integrity;and (6) the associated changes in cytokine profile thatmay contribute to the observations.

MATERIALS AND METHODS

Ethics

All experimental animal procedures were approved by the Instituteof Animal Care and Use Committee at our hospital and performed inaccordance with the Guide for the Care and Use of LaboratoryAnimals (NIH publication No. 85-23, National Academy Press,Washington, DC, revised 1996).

Animal Model of Liver Fibrosis

For the induction of fibrosis, adult male Fisher rats (Charles RiverTechnology, BioLASCO Co., Ltd., Taiwan) (n ¼ 20) were given carbontetrachloride (mixed with an equal volume of olive oil) [12] subcutane-ously every 72 h for 4 wk at a dose of 1 mL/kg each time. Significant liverfibrosis was confirmed by pathology in all animals after treatment.

Experiment Protocol and Animal Grouping

Liver fibrosis was induced through regular subcutaneous carbon tet-rachlorideadministration for 4wkin20adultmaleFisher ratsweighing250 to 300 g, which were then randomly divided into two groups: group 2(liver fibrosis with 70% hepatectomy and portal venous infusion of 1 mLculture medium not previously used, n¼ 10), and group 3 (liver fibrosiswith 70% hepatectomy and portal venous infusion of BMCs in 1 mLculture medium, n ¼ 10). Ten healthy adult male Fisher rats weighing250 to 300 g receiving equal volume of olive oil (Sigma-Aldrich, St. Louis,MO, USA) [12, 13] through subcutaneous injection at the samefrequency for the same duration (i.e., 4 wk) followed by 70%hepatectomy served as normal controls (group 1). BMCs wereharvested in all animals 5 d before hepatectomy and were infusedback to the same animals 48 h after hepatectomy through the portalvein only in group 3, whereas groups 1 and 2 animals received portalvenous infusion of equal volume of unused culture medium instead.All animals were sacrificed 2 wk after portal venous infusions withthe liver-to-body weight ratio calculated for comparison and the hepatictissue harvested for histologic and biochemical analyses. Plasma sam-ples were also obtained for determining the activities of alanine amino-transferase (ALT).

Preparation of BMCs for Flow Cytometric Analysis and Infusion

Experimental procedures were performed in pathogen-free animaloperation room. All groups of animals were anesthetized with chloralhydrate (35 mg/kg i.p.) 3 wk following fibrosis induction and placed insupine position. After carefully separating the ligament from thepatella, a 0.2 mm diameter electric rotablator was used to drilldirectly into the femoral bone from the distal end. A sterile 22-gaugeneedle syringe was then used to aspirate the bone marrow. Bone mar-row cells from each rat were buffered in 10 mL M199 medium (Gibco,Carlsbad, CA, USA), digested for 40 min with 0.01% collagenase B andDNase1, and filtered through a 30 mm nylon mesh. The BMCs werethen isolated by Ficoll-paque (Amersham Biosciences, Piscataway,NJ, USA) density-gradient centrifugation. Finally, the interphase ofBMCs was collected. These cells were washed twice with PBS, fol-lowed by centrifugation at 400 g for 5 min. The BMCs were then cul-tured in DMEM culture medium containing 10% FBS. Approximately1.2 to 2.0 3 106 BMCs were obtained from each rat (both femoralbones) via this method.

After 1 wk of cultivation, samples of BMCs from group 2 animalswere used to analyze the cell phenotype by flow cytometry. Briefly,the cells were trypsinized, washed twice with PBS and immuno-stained for 30 min on ice with monoclonal antibodies against primaryantibodies CD31 (serotec), KDR (NeoMarkers), CD271 (serotec),CD45 (serotec), CD29 (serotec), and CD90 (serotec). Secondary detec-tion was done using appropriate Alexa Fluor 488 (Molecular Probes,Eugene, OR). Isotype-identical antibodies (IgG) served as controls.After staining, the cells were fixed with 1% paraformaldehyde. Flowcytometric analyses were performed by utilizing a fluorescence-activated cell sorter (Beckman Coulter FC500 flow cytometer, Miami,FL, USA). Each analysis included 100,000 cells per sample. Assess-ment in each sample was performed in duplicate, with the mean levelreported. LMD files were exported and analyzed using the CXPsoftware.

For group 3 animals, autologous BMCs were infused through theportal vein 48 h after hepatectomy. To prepare BMCs for infusion,the cells were trypsinized with the cell density being adjusted to2.0 3 106 cells/mL. Thirty minutes before cell infusion, CM-1,10-dioctadecyl-3,3,30-tetramethylindo-carbocyanine perchloride (CM-DiI;Vybrant DiI cell-labeling solution, Molecular Probes, Inc.) (50 mg/mL)was added to the culture medium. This highly lipophilic carbocyaninedye, which has properties of low cytotoxicity and high resistance to inter-cellular transfer, can be added directly to normal culture media touniformly label suspended or attached culture cells for their visibilityinan implantedarea due to its distinctive red fluorescence. After cultiva-tion at room temperature for 30 min, the labeling process was quenchedby adding an equal volume of FBS. The DiI-labeled cells were washedthrice with culture medium, and adjusted to a concentration of 2 3 106

cells/mL in culture media. A drop of cell suspension was used to mounton slides and to confirm positive labeling under fluorescent microscope.

Hepatectomy

To simulate an acute critical surgical insult to the fibrotic liver, 70%hepatectomy including resection of the median and left lobes was per-formed in group 2 and group 3 animals after 4 wk of carbon tetrachlo-ride subcutaneous administration and equal volume of olive oil wasgiven in group 1. Briefly, an upper median incision was made on theabdomen with the animals in supine position under general anesthe-sia (chloral hydrate 35 mg/kg i.p.), followed by mobilization of the me-dian and left lobes of the liver through careful dissection on thesurrounding ligaments. The median and left lobes were then ligatedat the base and resected. The harvested hepatic tissue was thendivided into several portions (labeled as samples on postoperatived 0, POD 0) for histologic studies, protein, and molecular analysesto compare with the tissue obtained 2 wk later when the animalswere sacrificed (labeled as samples on postoperative d 14, POD 14).Immediately after the procedure, the animals were given subcutane-ous 5% Dextrose saline and 10% glucose solution in drinking water till

SUN ET AL.: BONE MARROW CELLS ALLEVIATE LIVER FIBROSIS IN RATS e17

the end of the third postoperative day to correct hypoglycemia andminimize mortality after the procedure [14, 15].

Portal Venous BMC Infusion

Forty-eight hours after bone marrow cell harvesting and cultur-ing, all animals were anesthetized with inhalational isoflurane(Abbott Laboratories Ltd., Queenborough, Kent, UK) and placed insupine position on a warming pad at 37 �C. The previous laparotomywound was reopened, followed by identification of the portal veinthrough which the harvested BMCs were slowly infused back tothe original animals in group 3 and equal volume of culture mediumnot previously used was given to group 1 and group 2 animals usinga 27 gauge needle attached to a 1 mL syringe. After withdrawal ofthe needle, the puncture site was gently compressed with a cottonswab for at least 3 min for hemostasis.

Immunohistochemical Phenotyping of Engrafted

Bone Marrow Cells

Engraftment of infused BMCs presenting with phenotypes ofhepatic resident cell types including hepatocyte, sinusoidal endothe-lial cells, and activated hepatic stellate cells (HSCs) were identifiedby immunohistochemical staining using primary antibodies againstrat albumin (Bethyl Lab Inc., Montgomery, TX), CD31 (JC/70A,DAKO, Glostrup, Denmark), and a-smooth muscle actin (a-SMA)(Sigma-Aldrich), respectively. Cy2-labeled anti-sheep IgG (JacksonIR Laboratories, West Grove, PA) or FITC-conjugated anti-mouse sec-ondary antibody (Sigma) was then applied, followed by incubation for30 min at room temperature. Isotype-matched IgG antibodies wereused as controls and nuclei were counterstained with 40,6-diami-dino-2-phenylindole (DAPI).

Quantitative Image Analysis of Fibrosis Area

Masson trichrome staining was used for studying fibrosis of theliver. Six serial sections of each liver specimen were prepared at4 mm thickness from paraffin blocks for fibrosis quantification. The in-tegrated area (mm2) of fibrotic spots in the tissue sections was calcu-lated using Image Tool 3 (IT3) image analysis software (Universityof Texas, Health Science Center, San Antonio, TX; Image Tool forWindows, ver. 3.0) as described previously [16]. Three selected sec-tions were quantified for each animal. Three randomly selectedhigh-power fields (HPFs) (4003) were analyzed in each section. Afterdetermining the number of pixels in the fibrotic area from each HPF,the numbers of pixels obtained from the three HPFs were summated.The procedure was repeated in two other sections for each animal.The mean pixel number per HPF for each animal was then deter-mined by summating all pixel numbers and dividing by 9. The meanarea of fibrosis per HPF was obtained using a conversion factor of19.24 (1 mm2 represented 19.24 pixels).

Western Blot Analysis for Collagen I and a-SMA Expression

Equal amounts (10–30 mg) of protein extracts from the liverspecimens were loaded and separated by SDS-PAGE using 8%–10%acrylamide gel. Following electrophoresis, the separated proteinswere transferred electrophoretically to a polyvinylidene difluoride(PVDF) membrane (Amersham Biosciences). Nonspecific proteinswere blocked by incubating the membrane in blocking buffer (5% non-fat dry milk in PBS-T containing 0.1% Tween 20) overnight. Themembranes were incubated with the indicated primary antibodies(collagen I and a-SMA from Sigma; GAPDH from Chemicon (Teme-cula, CA) overnight at 4�C. Horseradish peroxidase-conjugated anti-mouse IgG (1:3000, Amersham Biosciences) was applied as the secondantibody for 1 hour at room temperature. The immunoreactive bandswere visualized by enhanced chemiluminescence (ECL) (Amersham

Biosciences) after exposure to Biomax L film (Kodak, Rochester,NY, USA). For quantification, digitized ECL signals were analyzedusing Labwork UVP software (Massachusetts, MA, USA).

Quantitative Real-Time Polymerase Chain Reaction (RT-PCR)

Analysis for Tissue Necrosis Factor-a (TNF-a), Transforming

Growth Factor-b (TGF-b), and Peroxisome Proliferator

Activated Receptor-g Coactivator-1a (PGC-1a)

Liver tissues were grinded with a rotor-stator homogenizer and to-tal RNA was extracted in TriZol reagent (Invitrogen, Gaithersburg,MD) according to manufacturer’s instruction. After being spectropho-tometrically quantified, 2 mg of total RNA was used for reverse tran-scription reaction with Superscriptase II (Invitrogen) using oligo-dTand random primers. One-twentieth of complementary DNA gener-ated was used as template for real time Q-PCR analysis. Amplificationand detection was done by a LightCycler DNA Master SYBR Green Ikit (Roche Applied Science, Mannheim, Germany) in the LightCyclerDetection System (Roche Applied Science). PCR reaction was carriedout as follows: one cycle of 95�C for 10 min, 45 cycles of 95�C for 5 s,60�C for 5 s, and 72�C for 20 s. After amplification, a final meltingcurve protocol was performed to determine the specificity of PCRreaction. The primer sequences were: for GAPDH gene, forward 50-AGC TGG TCA TCA ATG GGA AA-30, reverse 50-ATTTGATGTTAGCGG GATCG-30 (60�C, 63 bp); for TGF-b1, forward 50-ATGGTGGACCGCAAC AAC-30 and reverse 50-ACAGCAATGGGGGTTCTG -30

(60 �C, 115 bp); for TNF-a, forward 50-GTCTACTGAACTTCGGGGTGA -30 and reverse 50-ATGAGAGGG AGCCCATTTG-30 (60 �C, 67bp); PGC-1a, forward 50- AATTTTTCAAGT CTAACTATGCAGACC-30 and 50-AAAATCCAGAGA GTCATACTTGCTC-30 (60 �C, 92 bp).

Oxidative Stress Reaction of Liver

The principle of using the OxyBlot Protein Oxidation Detection Kit(S7150, purchased from Chemicon) in quantifying oxidative stress isbased on the immunoblot detection of carbonyl groups producedthrough protein oxidation as previously described [17]. The carbonylgroups in the protein side chains are derivatized to 2,4-dinitrophenyl-hydrazone (DNP-hydrazone) by reaction with 2,4-dinitrophenylhy-drazine (DNPH). The DNP-derivatized protein samples are thenseparated by polyacrylamide gel electrophoresis followed by Westernblotting. In brief, DNPH derivatization was carried out on 6 mg of pro-tein for 10 min, followed by one-dimensional SDS-PAGE electrophore-sis resolved on 12% polyacrylamide gel. Proteins were transferred toPVDF membranes, which were then incubated in the primary anti-body solution (anti-DNP 1:150) for 2 h, followed by incubation withsecondary antibody solution (1:300) for 1 h at room temperature.The washing procedure was repeated eight times within 40 min.Immunoreactive bands were visualized by ECL (Amersham Biosci-ences), which was then exposed to Biomax L film (Kodak). For quan-tification, ECL signals were digitized using Labwork software (UVI).Gels were scanned for densitometric analysis using an NIH imagesystem. Oxidative index, which was used for comparison betweendifferent samples, was defined as the ratio between densitometricvalues of the oxyblot bands and those stained with Red Ponceau. Oneach gel, a standard control sample was loaded.

Assessment of Tissue Necrosis and Impairment

in Hepatocyte Integrity

Tissue sections from all animals were hematoxylin and eosin (HE)-stained and examined for the extent of necrosis. Determination ofplasma ALT activity, which is a conventional marker of hepatocellu-lar injury, was performed by standard spectrophotometric proceduresusing a BM/Hitachi 704 Autoanalysor (Hitachi, Tokyo, Japan).


Statistical Analysis

Data were expressed as mean values (mean 6 SD). The significanceof group differences was evaluated with t-test. Continuous variablesamong the three groups were compared by Kruskal-Wallis test fol-lowed by multiple comparison procedure using Wilcoxon rank sumtest and Bonferroni correction. Statistical analyses were performedusing SAS statistical software for Windows ver. 8.2 (SAS Institute,Cary, NC). A probability value < 0.05 was considered statisticallysignificant.

RESULTS

Immunofluorescence and Flow Cytometric Analysis of Cultured

BMCs (Fig. 1)

Morphological features of 1-wk cultured BMCs areshown in Fig. 1 (left upper corner). Some of the cultur-ing cells exhibited cobble stone-like appearance typicalof endothelial cells, whereas other cells were spindle-

FIG. 1. Morphological features of harvested bone marrow mononucleance of endothelial progenitor cells and spindle shape typical of mese200 mm). Identification of surface markers of endothelial progenitor ccytometric analysis.

shaped typical of MSCs. Flow cytometric analysisidentified the phenotypic presentations of endothelialprogenitor cells (CD31, KDR) and MSCs (CD29,CD45, CD90, CD271) after 1 wk cell culturing.

Identification of Engrafted BMCs and Phenotypic

Presentations (Fig. 2)

Engraftment of the BMCs was confirmed by identifi-cation of their characteristic red fluorescence fromDiI-staining on the liver sections 2 wk after portal infu-sion. Differentiations of the implanted BMCs weretraced using double staining with DiI and differentcell-specific antibodies, including albumin, a-SMA,and CD31 for the identification of the phenotypes ofhepatocytes, HSCs, and endothelial cells, respectively.The results are shown in Fig. 2, demonstrating engraft-ment and subsequent differentiation of BMCs after

ar cells after 1 wk of culturing, showing typical ‘‘cobble-stone’’ appear-nchymal stem cells (MSCs) (left upper corner) (scale bar representsells (CD31, KDR) and MSCs (CD29, CD45, CD90, CD271) on flow

FIG. 2. Figures illustrating the phenotypic presentations typical of different hepatic cellular elements. (A) DiI-stained bone marrow-derived mononuclear cells (red) identified; (B) albumin (Alb) positivity noted in the same cells on immunohistochemical staining; (C) cell nucleicounterstained with 40,6-diamidino-2-phenylindole (DAPI) (blue); (D) merged figures showing engrafted cells with locations along fibrous septaand in liver parenchyma surrounding vasculature. Note the brightness of the background caused by positive albumin staining of surroundinghepatocytes. (E)-(H) a-Smooth muscle actin staining (a-SMA) staining of engrafted bone marrow cells, marker of activated hepatic stellate cells(HSCs), showing predominant localization of the cells over fibrous septa. (I)-(L) Peri-vascular location of the engrafted CD31-positive cells,compatible with the phenotype of endothelial cells. Scale bars in right lower corner represent 100 mm.


portal venous administration. Examination of the loca-tions of the engrafted cells showed that albumin-positive cells were mainly around fibrous septae andin liver parenchyma surrounding hepatic vasculature,whereas those positively-stained for a-SMA and CD-31 were found inside pericentral fibrous septae andaround intrahepatic blood vessels, respectively.

Examination of the locations of the engrafted cellsshowed that albumin-positive cells were mainly locatedalong fibrous septae and in liver parenchyma surround-ing hepatic vasculature, whereas the engrafted cellspositively-stained for a-SMA were found inside pericen-tral fibrous septae.DiI/CD31-positive cells were observedlining and surrounding intrahepatic blood vessels.

Impact of BMC Implantation on the Rate of Regeneration

of the Fibrotic Liver (Fig. 3A and B)

Fourteen days after hepatectomy, the liver-to-bodyweight ratio showed a significant increase after BMCinfusion in group 3 compared with that of group 1(P ¼ 0.04), whereas there was no significant differencebetween group 1 and group 2. The increased ratio ingroup 3 cannot be explained by changes in body weightthat showed no significant difference among the three

groups before (i.e., postoperative d 0, POD 0) and after(i.e., postoperative d 14, POD 14) hepatectomy. Theresults suggest that single portal venous infusion ofBMCs may have a positive impact on the regenerationof fibrotic liver in the rats.

Effects of Portal Venous Infusion of BMCs on Collagen I

Protein Expression (Fig. 3C)

Western blot analysis on liver samples from the threegroups of animals revealed a significant increase incollagen I protein expression after fibrosis induction(Fig. 3C). Significant difference in collagen I expressionbetween POD 0 and POD 14, however, was noted onlyin group 3 (P ¼ 0.02), suggesting a beneficial effect ofBMC administration on reducing collagen I depositionin the fibrotic liver during regeneration in the rat.

Influence of BMC Infusion on a-SMA Protein

Expression (Fig. 3D)

The protein expression ofa-SMA was assessed withwestern blotting on liver homogenates on POD 0 andPOD 14 after hepatectomy, demonstrating a significantdecrease in the fibrotic liver of the BMC administration

FIG. 3. (A) Mean liver-to-body weight ratio (%) 14 d after hepatectomy in the three groups of animals (n ¼ 10 per group). BMC: bonemarrow-derived mononuclear cell: * versus y, P < 0.05. (B) No significant change in body weight among the three groups before (postoperatived 0, POD 0) and after (postoperative d 14, POD 14) hepatectomy (n ¼ 10 per group). (C) Collagen I protein expression in normal (i.e., group 1),fibrotic liver without (i.e., group 2), and with (i.e., group 3) bone marrow-derived mononuclear cell (BMC) treatment before (POD 0) and after(POD 14) hepatectomy (n ¼ 10 per group). POD 0: * versus y, P < 0.05; POD 14: z versus x, P < 0.05. Significant difference between POD 0 andPOD 14 noted only in group 3 (P ¼ 0.02). (D) a-Smooth muscle actin (a-SMA) protein expression in normal and fibrotic liver with and withoutBMC infusion before (POD 0) and after (POD 14) hepatectomy, showing increased expression after fibrosis induction with a significant reduc-tion after BMC administration (n ¼ 10 per group). POD 0: * versus y, P < 0.05; POD 14: z versus x, P < 0.05.


group after hepatectomy (group 3) compared with theirpretreatment values (P ¼ 0.02) and with the fibroticcontrols without BMC treatment (group 2) (P < 0.05)(Fig. 3D). The findings imply a reduction in HSC activa-tion during regeneration of the fibrotic liver in the ratafter BMC administration.

Histologic Analysis and Quantification of Fibrosis (Fig. 4)

Liver specimens of each animal harvested on d 0 ofhepatectomy and that on d 14 were stained withMasson trichrome (Fig. 4A–F). The degrees of fibrosiswere compared using Image Tool 3 (IT3) image analysissoftware (Fig. 4G). The results showed a remarkableincrease in fibrosis after carbon tetrachloride adminis-tration in group 2 and group 3 compared with group 1(i.e., normal controls) on d 0 of hepatectomy. Themean number of pixels showed a significant decrease

in both groups 2 and 3 14 d after hepatectomy, implyinga regression in the degree of fibrosis after withdrawal ofthe noxious stimuli (e.g., carbon tetrachloride) duringthe process of regeneration. On the other hand, thepixel number was significantly lower in group 3compared with that in group 2 on d 14 (P < 0.05), sug-gesting an enhancement in the alleviation of fibrosisin the regenerating liver after BMC infusion.

Oxidative Stress and Mitochondrial Function of the Regenerating

Fibrotic Liver (Fig. 5)

Oxyblot analysis revealed a significant reduction inoxidative stress in group 3 animals 14 d after hepatec-tomy compared with their respective levels beforeBMC administration (i.e., POD 0) (Fig. 5A). The rela-tively high oxidative index in the normal controls maybe attributable to the predominance of inactive

FIG. 4. Masson trichrome staining showing fibrosis (blue) in liver sections from three groups of animals. Normal controls (group 1) before(postoperative d 0, POD 0) (A) and after (postoperative d 14, POD 14) hepatectomy (B), showing no notable fibrosis. (C) and (D) illustratesignificant fibrosis in carbon tetrachloride (CCl4)-treated rats (group 2) before and after hepatectomy, respectively. (E) and (F) show a signifi-cant reduction in the degree of fibrosis 14 d after bone marrow-derived mononuclear cell (BMC) therapy (F) compared with the pretreatmenthistology (E) in group 3. Note the resolving fibrosis (green arrows) in (F). (G) Assessment of degree of fibrosis in liver sections using Image Tool(IT) software, demonstrating a significant rise in the degree of fibrosis after CCl4 administration. Note the significant reduction in mean pixelnumber in the two CCl4-treated groups 14 d after hepatectomy (n ¼ 10 per group). POD 0: * versus y, P < 0.05; POD 14: z versus x, P < 0.05;z versus {, P < 0.05; x versus {, P < 0.05. Scale bars in right lower corner represent 200 mm.



extracellular matrix proteins (e.g., collagen and proteo-glycan) in the fibrotic livers [18] (i.e., groups 2 and 3)that ‘‘diluted’’ the active cellular protein ingredientsin the latter. On the other hand, there was a notablesuppression in mRNA expression of PGC-1a after fibro-sis induction (all P < 0.05) (Fig. 5B) with the exceptionof group 3, 2-wk after hepatectomy (P > 0.3 versusgroup 1). The finding suggests significant preservationof mitochondrial metabolic function after BMC infusionin the regenerating fibrotic liver.

The Degree of Necrosis and Impairment in Hepatocyte

Integrity (Fig. 6A)

The extent of necrotic cell death of all animals wasassessed using HE staining (pictures not shown) andserum ALT activities. Although fibrosis was prominentin groups 2 and 3, there was no remarkable necrosis inthe liver sections. The changes in mean plasma ALTactivities are summarized in Fig. 6A. The resultsshowed that hepatectomy on the normal liver causedsignificant impairment in hepatocyte integrity even at2 wk after hepatectomy as reflected in the elevatedplasma ALT (P ¼ 0.04). Moreover, chronic administra-tion of carbon tetrachloride led to significant hepatocyteinjuries, which tended to recover upon cessation of thenoxious stimuli. Nevertheless, significant reduction inplasma ALT was noted only in group 3 animals 2 wkafter hepatectomy (P ¼ 0.02), suggesting a positiverole of BMC in the recovering process.

FIG. 5. (A) Changes in oxidative index in normal and fibrotic liverwith and without bone marrow-derived mononuclear cell (BMC)treatment before (postoperative d 0, POD 0) and after (postoperatived 14, POD 14) hepatectomy, demonstrating significant reduction inoxidative stress in the BMC-treated animals (group 3) 14 d afterhepatectomy (n¼ 10 per group). Positive control: mixture of standardproteins with attached DNP residues. (B) mRNA expressions of perox-isome proliferator activated receptor-g coactivator 1a (PGC-1a) innormal and fibrotic liver with and without BMC treatment before(POD 0) and after (POD 14) hepatectomy, showing significantsuppression in PGC-1a expression after induction of fibrosis but rela-tively well-preserved expression after BMC administration (n ¼ 10per group). POD 0: * versus y, P < 0.05; POD 14: z versus x, P < 0.05.

Changes in mRNA Expressions of Pro-Inflammatory

Cytokines, TNF-a and TGF-b (Fig. 6B and C)

PCR analyses demonstrated a significant up-regulation in the pro-inflammatory cytokine, TNF-a,after fibrosis induction in both group 2 and group 3(Fig. 6B). Significant reduction in TNF-a expression,however, was noted only in group 3 animals 14 d afterhepatectomy compared with their pre-hepatectomy level(P ¼ 0.02) and with that of group 2 after hepatectomy(P ¼ 0.02). The findings suggest an anti-inflammatoryaction of BMC therapy in the fibrotic liver during regen-eration. On the other hand, TGF-b, known mediator ofHSC activation, was remarkably up-regulated in group2 and group 3 after carbon tetrachloride administration.A significant decrease was noted in the fibrotic liver ofthe BMC administration group after hepatectomy(group 3) compared with their pre-treatment values(P < 0.05) and with the fibrotic controls without BMCtreatment (group 2) (P < 0.05) (Fig. 6C). The resultsimply an anti-inflammatory role of BMC administrationin reducing TGF-b expression and hence HSC activationduring regeneration of the fibrotic liver in the rat.

DISCUSSION

Although previous in vitro studies have showna hepatic anti-fibrotic effect of MSC through suppressingproliferation and enhancing apoptosis of activated HSCs[19], several in vivo studies have failed to show anytherapeutic advantage of bone marrow-derived MSCinfusion either in improving hepatic function or reducingfibrosis in chronic cirrhotic animal models [4, 5]. Thedifference in results may be due to the fact that during

FIG. 6. (A) Changes in plasma alanine amino-transferase (ALT)activities in normal and fibrotic liver with and without bonemarrow-derived mononuclear cell (BMC) treatment before (postoper-ative d 0, POD 0) and after (postoperative d 14, POD 14) hepatectomy.Increased plasma ALT activities after CCl4 administration thatsignificantly decreased after BMC therapy (n ¼ 10 per group), *versus y, P < 0.05; * versus z, P < 0.05; z versus x, z versus {, x

versus {, y versus x, P > 0.05. (B) mRNA expressions of tumor necrosisfactor-a (TNF-a) with and without bone marrow-derived mononuclearcell (BMC) treatment before (postoperative d 0, POD 0) and after(postoperative d 14, POD 14) hepatectomy, demonstrating significantelevations in expression after induction of fibrosis that showed signif-icant reduction after BMC infusion (n ¼ 10 per group). POD 0: *versus y, P < 0.05; POD 14: z versus x, P < 0.05; z versus {, P < 0.05;x versus {, P < 0.05. (C) mRNA expressions of transforming growthfactor-b (TGF-b) in normal and fibrotic liver with and without BMCtreatment on POD 0 and POD 14 after hepatectomy, showingsignificant increase in expression after fibrosis induction that de-creased after BMC administration. POD 0: * versus y, P < 0.05;POD 14: z versus x, P < 0.05; z versus {, P< 0.05; x versus {, P < 0.05.


the course of fibrosis and the process of regeneration inthe normal liver, stem cells are not the usualparticipants because of the remarkable replicativecapacity of the organ [20]. Indeed, it has been shownthat this replicative ability can be compromised onlywhen a threshold of 50% loss of hepatocytes is reached,together with a significant decrease in proliferativeability of the remaining mature hepatocytes [21]. Undercritical circumstances such as lethal irradiation, fulmi-nant hepatic failure, or 90% hepatectomy, however, theuse of bone marrow-derived MSCs has been demon-strated to produce dramatic therapeutic effects in exper-imental studies [2, 22, 23]. Consistently, a recent clinicalstudy demonstrated the successful use of autologoushematopoietic stem cells in the rescue of a patient withhepatic failure, highlighting the therapeutic potentialof stem cells in a critical setting of acute drug-induced in-jury on chronic alcoholic liver disease [11]. In terms of he-patic pathogenesis, the typical repairing process afterliver injury is known to involve two distinct phases(i.e., regeneration and fibroplasias) that tend to becomepathogenic with accumulation of extracellular matrixcomponents in case of improper control [24]. The ratio-nale of using a hepatectomized liver fibrosis model inthis study was to mimic the critical clinical condition ofhepatectomy in cirrhotic patients with malignant livertumors who have limited hepatic functional reserve.More importantly, this model attempted to create a situ-ation in which the regenerative phase may outweigh thefibroplastic phase through triggering a regenerativestimulus by hepatectomy while at the same time impos-ing an anti-fibrotic signal [19] via BMC administration.

The role of bone marrow-derived cellular elements inthe process of liver regeneration remains controversial.The results of the current study showed that hepaticengraftment of BMCs with their expression of differentphenotypes of hepatic cellular elements did occur in thefibrotic liver 14 d after portal venous administration inthe rat. Moreover, positivity of cellular marking withanti-rat albumin antibodies not only demonstrated thedevelopment of hepatocyte phenotype in some of theinfused BMCs, but it may also imply their role as func-tioning hepatocytes. The identification of albumin-expressing BMCs is also consistent with the finding ofother previous studies that demonstrated differentia-tion of bone marrow-derived [22] and liver-derived [25]stem cells into functional hepatocytes. Furthermore,in concert with the result of a previous study [25], the lo-cation of differentiated albumin-expressing BMCs werefound near hepatic vascular structure in this study.

However, the scarcity of engrafted cells in this studydoes not support an enhanced hepatic regenerationthrough proliferation of the infused cellular elements.Moreover, bone marrow-derived MSCs have beenreported to account for less than 0.01% of all nucleated


cells in human bone marrow [26]. Another experimentalstudy has also shown that bone marrow-derived cellularelements are unlikely natural sources of hepatic ovalcells in vivo [3]. In this way, our results are consistentwith that of a previous study, which demonstrated thatMSC-conditioned medium itself can provide trophic sup-port to the injured liver by suppressing hepatocellulardeath and stimulating regeneration [27]. Therefore, ourresults endorse the proposal that the therapeutic poten-tials of BMCs may be due to BMC-associated moleculesrather than the engrafted cellular elements per se [28].

The results of the present study demonstrated thatBMC infusion significantly alleviated hepatic fibrosisin vivo as reflected in the histologic improvement andthe reduction in collagen I expression which, at leastpartly, may be attributed to the suppression of HSCactivation as reflected in the significant decreases ina-SMA expression. The findings in this study are in con-cert with those from a previous in vitro investigationdemonstrating an inhibition of HSC proliferation andreduction in collagen synthesis probably through theparacrine effects of MSCs [19].

In this study, a decrease in oxidative index was notedin the BMC-treated animals 14 d after hepatectomy onthe fibrotic liver, suggesting a possible role of BMC asan antioxidant. The finding is consistent with that ofa previous study that showed an antioxidizing action ofbone marrow-derived MSCs in carbon tetrachloride-induced fulminant hepatic failure in a mouse model[23]. Consistently, RT-PCR analysis of the mRNAexpression of PGC-1a, which is a transcriptionalcoactivator and a major upstream regulator of lipidcatabolism, oxidative metabolism, mitochondrialmetabolism, and biogenesis [29], showed a significantreduction in the fibrotic liver in the present study.The finding reflects a malfunctioning of mitochondrion,which is a crucial platform for energy transduction [30].The relatively well preserved PGC-1a expression in thefibrotic liver after BMC administration suggests a possi-ble mitochondria-protective function mediated by BMCthat may at least partly account for the decrease inoxidative stress in liver parenchyma 14 d after BMCadministration in the current study.

Although no remarkable necrosis was noted in thehepatic fibrosis model of the present study, significantelevations in plasma ALT activities were noted afterfibrosis induction. The results suggest that chronic ad-ministration of carbon tetrachloride causes a significantdamage to hepatocyte integrity, which, however, tendsto recover upon cessation of the noxious stimulus [31].In this study, significant reduction was only noted inplasma ALT 2 wk after hepatectomy in the BMC-treated animals, implying a hepatocyte-protective ac-tion through BMC infusion. Accordingly, the observedreduction in oxidative stress and preservation of

mitochondrial function after BMC administrationmay at least partly account for the finding.

Consistent with the findings of this study, a previousin vitro investigation has demonstrated significant im-munomodulating effects of MSCs on activated HSCs,including an inhibition of HSC proliferation, inductionof HSC apoptosis, and a suppression of collagen synthe-sis [19]. In contrast with that study, which showed anenhanced TNF-a secretion by MSCs in vitro, the resultsof the present study revealed a remarkable fibrosis-associated TNF-a over-expression in vivo, which wassignificantly suppressed by MSC administration. Thediscrepancy in results may be due to the interactionsamong different hepatic cellular elements (e.g., Kupffercell and HSC) that give rise to the in vivo cytokineprofile after BMC administration. In the present study,the mRNA expression of TGF-b, a key mediator of HSCactivation, was remarkably up-regulated in the fibroticliver, indicating an ongoing fibrogenic process. Thesignificant down-regulation in TGF-b expression inthe fibrotic liver after BMC therapy, therefore, suggestssuppression in HSC activation, which is consistent withthe reduction in protein expression of a-SMA in thisstudy and also compatible with the results of previousin vitro studies [19]. Taken together, our resultssuggest an overall anti-inflammatory effect throughBMC administration. The finding is again consistentwith that of our previous study [32] that demonstratedsignificant anti-inflammatory actions of BMC ina dilated cardiomyopathy setting.

Proposed mechanisms underlying the potential thera-peutic benefits of portal administration of BMCs in theregenerating fibrotic liver of the rat are summarized inFig. 7. Using autologous BMCs from adult animals, thecurrent study demonstrated that adult stem cells maybe an ideal source for hepatic cell therapies as previouslyproposed by other authors [25]. Moreover, instead of fo-cusing on MSCs that normally constitute only a minorportion of the bone marrow cell population and cannotbe satisfactorily purified without prolonged in vitro cul-turing, labeling, and sorting, this study emphasizes theeffects of unsorted BMCs on the marginally functioningliver to simulate the critical clinical situation that re-quires immediate management through BMC aspirationand direct autologous infusion after 1 wk. This strategyhas been adapted in a recent clinical study that showedencouraging results [9]. Another clinical implication ofthe current study is that with the administration of au-tologous BMCs, patients with concomitant parenchymalliver disease and malignant neoplasms may be able toreceive hepatectomy despite limited hepatic functionalreserve, which is the most important clinical concern indetermining the resectability of tumors and, hence, theprognosis of the patients. Moreover, patients with life-threatening liver failure may benefit from portal

FIG. 7. Summary of the proposed mechanisms underlying the potential therapeutic benefits of portal administration of bone marrow-derived mononuclear cells (BMCs) in the regenerating fibrotic liver of the rat. HSC: hepatic stellate cell; a-SMA: a-Smooth muscleactin; TGF-b: transforming growth factor- b; PGC-1a: peroxisome proliferator activated receptor-g coactivator-1a; TNF-a: Tumor necrosisfactor-a; ALT: alanine aminotransferase.


infusionof unsorted BMCs asa life-savingand a bridgingprocedure for subsequent liver transplantation. A recentclinical study has successfully applied autologous bonemarrow-derived stem cells in the treatment of patientswith hepatic insufficiency, achieving improvement in se-rum parameters [10]. Further experimental and clinicalelucidations may be required to validate the liberal use ofBMCs in the clinical setting of hepatic failure.

In conclusion, the findings of the present study dem-onstrated that BMC infusion significantly enhancedhepatic regeneration and alleviated hepatic fibrosis ina regenerating fibrotic liver model. The underlyingmechanisms may involve suppression of HSC activa-tion, reduction of collagen I expression and oxidativestress, preservation of mitochondrial function, as wellas down-regulation of pro-inflammatory cytokines.

ACKNOWLEDGMENT

The authors gratefully acknowledge financial support from theGang Gung Medical Research Fund (grant no. CMRPG 860211).

REFERENCES

1. Czeizel E, Vaczo G, Kertai P. Effect of bone-marrow on regener-ation of liver of x-irradiated rats. Nature 1962;196:240.

2. Petersen BE, Bowen WC, Patrene KD, et al. Bone marrow asa potential source of hepatic oval cells. Science 1999;284:1168.

3. Menthena A, Deb N, Oertel M, et al. Bone marrow progenitorsare not the source of expanding oval cells in injured liver. StemCells 2004;22:1049.

4. Quintanilha LF, Mannheimer EG, Carvalho AB, et al. Bonemarrow cell transplant does not prevent or reverse murine livercirrhosis. Cell Transplant 2008;17:943.

5. Carvalho AB, Quintanilha LF, Dias JV, et al. Bone marrowmultipotent mesenchymal stromal cells do not reduce fibrosisor improve function in a rat model of severe chronic liver injury.Stem Cells 2008;26:1307.

6. Hardjo M, Miyazaki M, Sakaguchi M, et al. Suppression of car-bon tetrachloride-induced liver fibrosis by transplantation ofa clonal mesenchymal stem cell line derived from rat bonemarrow. Cell Transplant 2009;18:89.

7. Abdel Aziz MT, Atta HM, Mahfouz S, et al. Therapeutic potentialof bone marrow-derived mesenchymal stem cells on experimen-tal liver fibrosis. Clin Biochem 2007;40:893.

8. Kharaziha P, Hellstrom PM, Noorinayer B, et al. Improvementof liver function in liver cirrhosis patients after autologousmesenchymal stem cell injection: A phase I-II clinical trial. EurJ Gastroenterol Hepatol 2009;21:1199.

9. Terai S, Ishikawa T, Omori K, et al. Improved liver function inpatients with liver cirrhosis after autologous bone marrow cellinfusion therapy. Stem Cells 2006;24:2292.

10. Khan AA, Parveen N, Mahaboob VS, et al. Safety and efficacy ofautologous bone marrow stem cell transplantation throughhepatic artery for the treatment of chronic liver failure:A preliminary study. Transplant Proc 2008;40:1140.

11. Gasbarrini A, Rapaccini GL, Rutella S, et al. Rescue therapy byportal infusion of autologous stem cells in a case of drug-inducedhepatitis. Dig Liver Dis 2007;39:878.

12. Parsons CJ, Bradford BU, Pan CQ, et al. Antifibrotic effects ofa tissue inhibitor of metalloproteinase-1 antibody on establishedliver fibrosis in rats. Hepatology 2004;40:1106.

13. Jung KH, Shin HP, Lee S, et al. Effect of human umbilical cordblood-derived mesenchymal stem cells in a cirrhotic rat model.Liver Int 2009;29:898.

14. Sunby CR, Anthony W, Ciechoski MP, et al. Enhanced survivalrates after extended hepatectomy by the early administration ofexogenous glucose. Curr Surg 1987;44:105.

15. Gaub J, Iversen J. Rat liver regeneration after 90% partialhepatectomy. Hepatology 1984;4:902.

16. Sheu JJ, Chang LT, Chiang CH, et al. Impact of diabetes oncardiomyocyte apoptosis and connexin43 gap junction integ-rity: Role of pharmacological modulation. Int Heart J 2007;48:233.

17. Dalla Libera L, Ravara B, Gobbo V, et al. Skeletal musclemyofibrillar protein oxidation in heart failure and the protectiveeffect of Carvedilol. J Mol Cell Cardiol 2005;38:803.

18. Bataller R, Brenner DA. Liver fibrosis. J Clin Invest 2005;115:209.

19. Parekkadan B, van Poll D, Megeed Z, et al. Immunomodulationof activated hepatic stellate cells by mesenchymal stem cells.Biochem Biophys Res Commun 2007;363:247.


20. Fausto N. Liver regeneration and repair: Hepatocytes, progeni-tor cells, and stem cells. Hepatology 2004;39:1477.

21. Katoonizadeh A, Nevens F, Verslype C, et al. Liver regenerationin acute severe liver impairment: A clinicopathological correla-tion study. Liver Int 2006;26:1225.

22. Miyazaki M, Hardjo M, Masaka T, et al. Isolation of a bonemarrow-derived stem cell line with high proliferation potentialand its application for preventing acute fatal liver failure.Stem Cells 2007;25:2855.

23. Kuo TK, Hung SP, Chuang CH, et al. Stem cell therapy for liverdisease: Parameters governing the success of using bone marrowmesenchymal stem cells. Gastroenterology 2008;134:2111.

24. Wynn TA. Cellular and molecular mechanisms of fibrosis.J Pathol 2008;214:199.

25. Najimi M, Khuu DN, Lysy PA, et al. Adult-derived human livermesenchymal-like cells as a potential progenitor reservoir ofhepatocytes? Cell Transplant 2007;16:717.

26. Pittenger MF, Mackay AM, Beck SC, et al. Multilineage poten-tial of adult human mesenchymal stem cells. Science 1999;284:143.

27. van Poll D, Parekkadan B, Cho CH, et al. Mesenchymal stemcell-derived molecules directly modulate hepatocellular deathand regeneration in vitro and in vivo. Hepatology 2008;47:1634.

28. Yip HK, Chang LT, Wu CJ, et al. Autologous bone marrow-derived mononuclear cell therapy prevents the damage of viablemyocardium and improves rat heart function following acuteanterior myocardial infarction. Circ J 2008;72:1336.

29. Garnier A, Fortin D, Delomenie C, et al. Depressed mitochon-drial transcription factors and oxidative capacity in rat failingcardiac and skeletal muscles. J Physiol 2003;551:491.

30. Huss JM, Kelly DP. Mitochondrial energy metabolism in heartfailure: A question of balance. J Clin Invest 2005;115:547.

31. Issa R, Zhou X, Constandinou CM, et al. Spontaneous recoveryfrom micronodular cirrhosis: Evidence for incomplete resolutionassociated with matrix cross-linking. Gastroenterology 2004;126:1795.

32. Sun CK, Chang LT, Sheu JJ, et al. Bone marrow-derived mono-nuclear cell therapy alleviates left ventricular remodeling andimproves heart function in rat-dilated cardiomyopathy. CritCare Med 2009;37:1197.

Bone Marrow Cells Reduce Fibrogenesis and Enhance Regeneration in Fibrotic Rat Liver

Documents

Transcript of Bone Marrow Cells Reduce Fibrogenesis and Enhance Regeneration in Fibrotic Rat Liver