Aflatoxin exposure during the first 1000 days of life in rural SouthAsia assessed by aflatoxin B1-lysine albumin biomarkersJohn D Groopman 12 Patricia A Egner 1 Kerry J Schulze 2 Lee S-F Wu 2Rebecca Merrill 2 Sucheta Mehra 2 Abu A Shamim 2 Hasmot Ali 2 Saijuddin Shaikh 2Alison Gernand 2 Subarna K Khatry 2 Steven C LeClerq 2 Keith P West Jr 2Parul Christian 2

1 Department of Environmental Health Sciences Department of International Health Bloomberg School of Public Health Johns Hopkins University Baltimore MD 21205 USA2 Center for Human Nutrition Department of International Health Bloomberg School of Public Health Johns Hopkins University Baltimore MD 21205 USA

A R T I C L E I N F O

Article historyReceived 11 August 2014Accepted 22 September 2014Available online 13 October 2014

KeywordsAflatoxinAflatoxin B1-lysine albumin biomarkersChild healthMaternal exposure

A B S T R A C T

Aflatoxin B1 is a potent carcinogen occurring from mold growth that contaminates staple grains in hothumid environments In this investigation aflatoxin B1-lysine albumin biomarkers were measured by massspectrometry in rural South Asian women during the first and third trimester of pregnancy and theirchildren at birth and at two years of age These subjects participated in randomized community trialsof antenatal micronutrient supplementation in Sarlahi District southern Nepal and Gaibandha Districtin northwestern Bangladesh Findings from the Nepal samples demonstrated exposure to aflatoxin with94 detectable samples ranging from 045 to 293930 pg aflatoxin B1-lysinemg albumin during preg-nancy In the Bangladesh samples the range was 156 to 6322 pg aflatoxin B1-lysinemg albumin in thefirst trimester 337 to 728 pg aflatoxin B1-lysinemg albumin in the third trimester 462 to 7669 pg af-latoxin B1-lysinemg albumin at birth and 388 to 8144 pg aflatoxin B1-lysinemg albumin at age two yearsAflatoxin B1-lysine adducts in cord blood samples demonstrated that the fetus had the capacity to convertaflatoxin into toxicologically active compounds and the detection in the same 2-year-old children illus-trates exposure over the first 1000 days of life

copy 2014 Elsevier Ltd All rights reserved

1 Introduction

Nearly 200000000 children worldwide suffer from impairedgrowth and development and nearly 2 million die each year beforeage 5 (Black et al 2013 Smith et al 2012) Many of these chil-dren are from South Asia and these long-term health deficits placean enormous burden on these economically developing countriesThe overall contributing factors driving these health problems stillneeds to be completely identified and better intervention strate-gies developed (Bhutta et al 2013)

Aflatoxin B1 is a very potent toxin and carcinogen for humansand this naturally occurring mycotoxin occurs as a consequence ofspecific mold growth that permeates staple grains legumes and nutsconsumed by people especially in hot humid environments in low-income countries (Wogan et al 2012) Aflatoxin induces many

adverse local and systemic effects that impair normal organ andtissue function resulting in growth retardation inflammation andimmune suppression all of which contribute in multiple ways topoor health (Khlangwiset et al 2011 Wild and Gong 2010) Chronicexposure to aflatoxin also significantly enhances excess risk for livercancer in many human populations (Kensler et al 2011 Wild andGong 2010) Similar to most environmental toxicants aflatoxin re-quires metabolic activation by intrinsic cytochrome P450 enzymesmainly localized in the liver to exert its toxicological effects Thesemetabolic activation pathways have been well described by severalresearch teams (Guengerich 2008 Wogan et al 2012) One of thesemetabolic derivatives the aflatoxin B1-lysine adduct that is iso-lated from serum albumin is a validated biomarker of exposure andrisk (Chen et al 2013 Sabbioni et al 1990 Turner et al 2005 Wanget al 1996 Wild et al 1992) A number of studies have found strongassociations between aflatoxin exposure as measured by aflatoxinndashalbumin biomarkers and significant reductions in growth in childrenas assessed by height for age and weight for age (Gong et al 20022003 2004 Turner et al 2007)

Aflatoxin B1 contamination has been detected in food feed anda limited set of blood and urine samples collected in Nepal and atleast one third of the foodstuffs examined contained elevated levels

Abbreviation AFB1 aflatoxin B1 Corresponding author Dr John D Groopman Johns Hopkins University

Department of Environmental Health Sciences Bloomberg School of Public HealthBaltimore MD 21205 United States Tel + 4109553900 fax + 4109550617

E-mail address jgroopmajhsphedu (JD Groopman)

httpdxdoiorg101016jfct2014090160278-6915copy 2014 Elsevier Ltd All rights reserved

Food and Chemical Toxicology 74 (2014) 184ndash189

Contents lists available at ScienceDirect

Food and Chemical Toxicology

journal homepage wwwelseviercom locate foodchemtox

of this mycotoxin (Koirala et al 2005 Okumura et al 1993) Con-tamination of staple grains and foodstuffs were also found inBangladesh (Begum and Samajpati 2000 Dawlatana et al 2002)Earlier literature in India had related aflatoxin exposure to child-hood cirrhosis and acute aflatoxicosis (Krishnamachari et al 1977Tandon et al 1978 Yadgiri et al 1970) The major goal of thiscurrent study was to quantitatively measure aflatoxin exposure usingvalidated biomarkers in pregnancy early infancy and at ~24 monthsof age spanning a critical developmental period referred to as ldquothefirst 1000 days of liferdquo in rural South Asia (Katz et al 2013) Samplesfor this exploratory analysis were drawn from an extant collectionof biospecimens collected from large samples of pregnant womenand infants living in rural settings in the southern plains or Teraiof Nepal and flood plains of northern Bangladesh Both study areassituated along the northern and eastern flanks of the greater Gan-getic floodplain that has over 750 million people are exposed tostrong seasonality high heat humidity and a general lack of ade-quate food storage facilities that makes aflatoxin contamination muchmore likely

In this investigation womenrsquos serum or plasma samples fromthe first and third trimester of pregnancy in both country settingsand additionally in Bangladesh cord blood at birth and serumsamples from the same children at ~24 months of age were studiedA state-of-the-art isotope dilution quantitative mass spectrom-etry method was used for measuring aflatoxin B1-lysine adducts inalbumin in these samples as biomarkers of aflatoxin exposure Thusthe findings reported here represent an initial attempt to charac-terize the exposure to aflatoxin during pregnancy and early life ofthe children These results reveal almost universal excessive expo-sure to this potent naturally occurring mycotoxin

2 Methods

21 Study subjects serum processing and sample selection

Archived serum or plasma specimens for pregnant women and their childrenwere selected from a cohort of subjects who participated in multiple randomizedantenatal micronutrient supplementation trials carried out in the southern rural plainsof Nepal from 1999 to 2001 and more recently (2008ndash2012) in Bangladesh (Westet al 2013) Details of the design major findings and biospecimen collection pro-cedures for the trials have been previously reported (Christian et al 2003a 2003b)Each trial was approved by institutional review boards in their respective coun-tries the Nepal Health Research Council Kathmandu Nepal and the BangladeshMedical Research Council in Dhaka Bangladesh and the Institutional Review Board(IRB) of the Johns Hopkins Bloomberg School of Public Health Baltimore MD USA

The Nepal and Bangladesh trials enrolled n = 4926 and n = 44567 women re-spectively with subsets of women based on geographic residence recruited for moreintensive investigations that included blood draws in the home following enroll-ment and at approximately 32 weeks of gestation To be eligible for participationwomen were screened at 5-weekly intervals for menstrual status and given a urine-based pregnancy test if menstrual status since the previous visit was negative allowingfor enrollment into the studies prior to a gestational age of 11 weeks on averageIn total n = 1165 baseline (ie first trimester) samples and n = 779 third trimestersamples were collected at the Nepal site and n = 2070 first trimester and n = 1597third trimester samples were collected at the Bangladesh site Thirty (30) of the Nepalparticipants were randomly selected for the initial first and third trimesteranalysis

At the Bangladesh site a smaller subset of women (n = 399) were additionallyenrolled in a study that allowed for the collection of cord blood Finally childrenborn to women at the Bangladesh site were followed to 24 months of age A subsetof these children (n = 177) was enrolled in a study that required blood sample col-lection to evaluate biochemical indicators related to growth In this study there were63 mothers for the first and third trimester samples 63 cord blood samples and thesame 63 two-year olds at follow-up available and analyzed

At both sites blood was processed to serum or plasma which was subse-quently shipped to The Johns Hopkins Center for Human Nutrition and stored at minus80 oCto form a large repository of samples from which the samples selected for this seriesof investigations were drawn

Longitudinal (1st and 3rd trimester) samples from pregnant women from Nepalwere randomly sampled from a larger set of samples in which alpha-1-acid glyco-protein (AGP) a marker of inflammation had previously been assessed along withC-reactive protein (CRP) using methods as previously described (Christian et al 1998)Approximately equal numbers of women were selected based on having elevated

(n = 16 gt1 gL) or normal (n = 14 le1 gL) AGP in early pregnancy and correspond-ing 1st and 3rd trimester samples allowing us to examine aflatoxin exposurelongitudinally through pregnancy and its association with inflammation in early preg-nancy A larger set of early pregnancy samples from the Nepal set was then randomlyselected based on birth outcome (n = 27ndash30 women per group having had normalbirth miscarriage stillbirth preterm or low birth weight for a total of n = 141samples) data from this larger set were used to examine determinants of afla-toxin exposure with natural log-transformations used for analysis of the aflatoxin-adduct values by a variety of factors by ANOVA Data presented as geometric meansand standard deviations From the Bangladesh study a longitudinal cohort of ma-ternal 1st trimester 3rd trimester and cord blood samples and samples fromcorresponding children at 24 months of age was assessed by examining all samplesavailable at the intersection of the main sub-study cord blood collection sub-study and child follow-up at 24 months of age (n = 63)

22 Aflatoxin B1-lysine albumin adduct assay by isotope dilution mass spectrometry

All of the serum or plasma samples were analyzed using a minor variation ofthe method reported by McCoy and colleagues (McCoy et al 2008) The sample(200 μL) was mixed with an internal standard (10 μL times 01 ng AFB1-D4-lysmL) andPronase solution (250 μL 13 mgmL PBS) and incubated for 18 h at 37 degC Solid-phase extractionndashprocessed samples were analyzed by UPLC with mass spectrometricdetection using a ThermoElectron TSQ Vantage operated in the positive electrosprayionization SRM mode The internal standard parent molecular ion [(M + H)+ mz 4613]fragmented to yield an ion at mz 3982 The AFB1-lys molecular ion (mz 4572) frag-mented to yield an ion at mz 3941 A 10-point isotopic dilution standard curve wasgenerated by triplicate injection (10 μL) of AFB1-D4-lys (100 pg) mixed with varyingamounts of AFB1-lys (0ndash04 ng) prepared via 2-fold serial dilutions The data werefitted using the method of least-squares with a 1times weighting factor The isotope di-lution standard curve was linear over the range of 02 pg to 04 ng AFB1-lys injectedonto the column in 10 μL aliquots (R2 = 0995) The limit of quantitation was 02 pgAFB1-lysinemg albumin in these analyses

3 Results

This research was initiated by an analysis of aflatoxin B1-lysinealbumin adducts in a pooled sample representing one thousand 6ndash7year old children in Sarlahi Nepal who were being analyzed forplasma proteome differences related to micronutrient status (Schollet al 2012) That analysis showed 565 pg aflatoxin B1-lysinemgalbumin prompting the more detailed examination of aflatoxin ex-posure during pregnancy and early life described here The aflatoxinB1-lysine albumin adduct biomarker levels in first and third tri-mester samples from thirty Nepalese pregnant women are shownin Fig 1 This study examined the individual tracking of aflatoxinexposure between the first and third trimester As seen in theseresults exposures were quite consistent across these two mea-sures obtained approximately 6 months apart indicating a highexposure to both the fetus and the mother during pregnancy Onlytwo and four of the samples from the first and third trimester hadnon-detectable levels respectively These preliminary data were sug-gestive of limited seasonal variation in aflatoxin exposure

Fig 1 Aflatoxin B1-lysine adduct exposures in 30 Nepal pregnant women trackbetween 1st and 3rd trimesters

185JD Groopman et alFood and Chemical Toxicology 74 (2014) 184ndash189

The aflatoxin biomarker levels from the first trimester samples(shown in Fig 1) were then evaluated by inflammatory status asmeasured by alpha-1-acid glycoprotein (AGP) These findings areshown in Fig 2 and reveal that the aflatoxin-albumin adduct levelswere higher among women with increased levels of this circulat-ing marker of inflammation The findings in Fig 2 are the first toassociate elevated AGP with increasing aflatoxin exposure

Among the larger group of samples obtained in early pregnan-cy among Nepalese women only nine of the n = 141 samples hada non-detectable level of aflatoxin-albumin adducts and the rangeof detectable exposures was from 045 to 293930 pg aflatoxinB1-lysinemg albumin The median level of these samples was2245 pg aflatoxin B1-lysinemg albumin with a geometric mean of2528 pg aflatoxin B1-lysinemg albumin and the 25th to 75th per-centile ranged from 742 to 7794 pg aflatoxin B1-lysinemg albuminDeterminants of aflatoxin exposure are further explored in Table 1Although aflatoxin did not vary by AGP in these analyses it was sig-nificantly higher when another marker of inflammation C-reactiveprotein (CRP) was elevated Further exploration of these data re-vealed differences in exposure related to local ethnicities Aflatoxin-albumin adducts in samples from the ethnic Pahadi people (N = 61)were statistically significantly higher (p lt 0001) than those insamples from the ethnic Madhesi (N = 80) The Pahadi live in thehill region of Nepal and traditionally have a higher maize consum-ing diet The two highest values (23933 and 29393 pg aflatoxin B1-lysinemg albumin) were found in samples obtained during the Juneto August rainy season seasonal variability was significant(p lt 00001) with samples acquired in the September to Novem-ber period having the highest level of adducts compared to the otherthree-month intervals during the year (Table 1)

Depicted in Fig 3 is aflatoxin exposure that occurred amongwomen in Bangladesh across pregnancy from the first through thethird trimester The median levels of the aflatoxin biomarker inthe first and third trimester were 1808 and 2535 pg aflatoxinB1-lysinemg albumin respectively (100 positive) The median levelof aflatoxinndashlysine albumin adducts in cord blood samples was2741 pg aflatoxin B1-lysinemg albumin (100 positive) This dem-onstrated that the fetus had the biochemical capacity to convertaflatoxin that crossed the placenta into toxicologically activecompounds Further the detection of the same biomarkers in2-year-old children born to the mothers who were exposed to af-latoxins during pregnancy has a median level of 1379 pg aflatoxinB1-lysinemg albumin (100 positive) These findings illustrated theprofound exposure for these children across their first 1000 dayslife

4 Discussion

The greater Gangetic floodplain regions occupying southern Nepaland northwestern Bangladesh have hot and humid environmentsconducive to aflatoxin B1 contamination of staple grains (Begum andSamajpati 2000 Dawlatana et al 2002 Koirala et al 2005) Whilesubstantial aflatoxin B1 contamination has been detected in a varietyof foodstuffs a much more limited data set exists relating these pu-tative exposures to specific formation of biologically effective dose

Fig 2 Aflatoxin B1-lysine adduct exposures in first trimester of pregnancy in 141Nepal women associate with elevated inflammation biomarker AGP

Table 1Determinants of aflatoxin-albumin adducts (pg aflatoxin B1-lysinemg albumin) amongNepalese women (n = 141) during early pregnancy

pgmg1 p-Value

Aflatoxin (total samples) 2528 (581)AGPle1 gL 2605 (575)gt1 gL 1611 (717) 046CRPle5 mgL 2310 (564)gt5 mgL 11792 (542) 001Age14ndash19 y 2287 (675)20ndash29 y 2794 (611)ge30 y 2049 (431) 069Parity0 3246 (675)1ndash2 3150 (553)ge3 1746 (531) 014PahadiNo 1473 (516)Yes 5405 (505) lt00001LiterateNo 2579 (593)Yes 2357 (553) 079Ground floor constructionNo walls thatch grass sticks or mud 2658 (599)Wood or brick and cement 1969 (485) 046SchoolingNone 2528 (599)1 or more years 2503 (526) 098SeasonDecemberndashFebruary 2822 (495)MarchndashMay 800 (353)JunendashAugust 4699 (587)SeptemberndashNovember 11330 (332) lt00001BMIlt19 kgm2 2198 (630)ge19 kgm2 2967 (531) 032

1 Data expressed as geometric mean (SD)

Fig 3 Bangladesh aflatoxin exposure studies first third trimester cord blood and2 year olds There were 63 mothers for the first and third trimester samples 63 cordblood samples and the same 63 two-year olds at follow-up

186 JD Groopman et alFood and Chemical Toxicology 74 (2014) 184ndash189

metrics such as the aflatoxin B1-albumin biomarker in people Toaddress this data gap a quantitative isotope dilution mass spec-trometry method was deployed to measure the aflatoxin B1-lysinealbumin adduct at biologically critical times in mothers during preg-nancy and in the first thousand days of life in their children An initialanalysis of aflatoxin-albumin adducts in a pooled sample from onethousand 6ndash7 year old children in Nepal found 565 pg aflatoxin B1-lysinemg albumin a level comparable to that detected in verynutritionally compromised children in West Africa when the dif-ferent method of analyses are considered (Gong et al 2004 McCoyet al 2008) This biomarker observation prompted a more de-tailed examination using samples already analyzed for its plasmaproteome that had collected during the first and third trimester ofpregnancy of women enrolled in a randomized micronutrient in-tervention clinical trial in Sarlahi Nepal (Cole et al 2013 Schollet al 2012)

Aflatoxin B1 is a very potent toxic and carcinogenic agent for avariety of species including humans Contamination of staple grainsoccurs both prior to harvest and during storage of these food-stuffs Levels of contamination can vary widely resulting in humanexposures that can range from nanograms per day to milligrams perday depending upon the specific staple grain and amount con-sumed (Kensler et al 2011) While much of the international focuson this contamination has been framed within its carcinogenic prop-erties many of its other toxic effects can have both short and long-term health consequences In South Asia where both maternal andchild health status is tremendously compromised the potentialeffects of aflatoxin exposure are most likely manifested by these non-carcinogenic endpoints (Bhutta et al 2013 Black et al 2013)However in the future as early life health status improves acrossthis region these potential cancer endpoints could become muchmore consequential for these populations Nonetheless when af-latoxin exposure is determined to be substantial these data shouldbe viewed as being a sentinel for overall poor quality of staple grainsand nutrition (CAST 2003)

The wide heterogeneity in aflatoxin contamination levels in grainsresults in a very difficult task to assess human exposure using dietaryquestionnaires and spot testing of staple foods The developmentof mechanistically based biomarkers such as the aflatoxin B1-lysine albumin adduct has helped to overcome the inherentdifficulties of assessing aflatoxin exposure by measuring dietary con-stituents These biomarkers have the added advantage of being ableto integrate exposures over the course of many weeks or monthsdue to the long half-lives of these biomarkers in plasma (McCoy et al2005 2008) Research from our laboratory has also demonstratedthe long-term stability of aflatoxin albumin biomarkers in storedsamples where we have found these adducts to be stable for up to25 years (Scholl and Groopman 2008) The analytical technologyfor the measurement of these biomarkers has also substantially im-proved over the last several decades and comparative studies ofpaired samples have been done to relate measurements obtainedacross a number of analytical platforms The correlation of afla-toxin albumin adducts measured by mass spectrometry HPLC withfluorescence detection and immunoassays has been done and thispermits composite comparisons of human data obtained from manydiverse molecular epidemiologic studies across the globe (McCoyet al 2008 Scholl et al 2006) As a point of reference the valuesobtained in this comparative study revealed that the ELISA methodused provided aflatoxin adduct levels 5-fold higher than mass spec-trometry but the correlation coefficient between the two techniqueswas 098 (McCoy et al 2008) In the work reported in this studywe have employed a quantitative isotope dilution mass spectrom-etry method that permits the measurement of aflatoxin lysineobtained from serum albumin with the use of an isotopically labeledinternal standard This method is highly sensitive and specific per-mitting the quantitation of this biomarker in less than 100 μL of

plasma obtained in individuals residing in high exposure regionsof the world

Despite the emphasis of many investigations on liver cancer con-sequent to aflatoxin exposure there have been a number of studiesprimarily in West Africa that have raised the potential etiologic con-tribution of this toxin to diminished child growth and development(Gong et al 2002 2003 2004) Stunting is calculated to affect nearly200000000 children worldwide primarily in sub-Saharan Africa andSouth Asia (Black et al 2008 Smith et al 2012) Stunting has pro-found associations with a variety of growth and developmentdeficiencies in these children and to date the overall etiology of thisproblem remains obscure Since aflatoxin B1 has major impacts ongrowth and development in a number of animal species that hasbeen well characterized in the veterinary literature it is reasonablehypothesis that similar growth and developmental effects can occurin people The first step in the exploration of this hypothesis is thecharacterization of specific exposures in high-risk communities Thebiologically effective dose biomarker aflatoxin albumin adducts pro-vides an objective metric for characterizing these exposures forsubsequent follow-up and potential intervention

In an attempt to place the extent of exposures in the popula-tions reported here in Nepal and Bangladesh we compared thesecurrent data to those observed in other populations experiencingacute toxic (Kenya) or chronic (China) exposures (Chen et al 2013McCoy et al 2008) evaluated previously in our laboratory Theresults depicted in Fig 4 provide a composite perspective for theaflatoxin exposures detected in Nepal and Bangladesh in contextwith other documented exposures There was an acute aflatoxicosisoutbreak in Kenya in 2004 and again in 2005 in which nearly 150people died after consuming aflatoxin-contaminated maize Thelevels depicted measured in our laboratory were obtained from 107people living in those affected Kenyan villages (McCoy et al 2008Strosnider et al 2006) To provide perspective the median adductlevel was 10-times higher than the median for the 141 positivesamples of 150 pregnant women in Nepal whose blood was col-lected in 1999 However it is important to point out that many ofthe outlying samples in the Nepal data set overlap values of afla-toxin exposure in individuals from the Kenya toxicity event Thecomparative values in Bangladesh (from Fig 3) are shown nextFinally three sets of data from our studies in eastern China a regionof very high liver cancer incidence are shown in the far right of Fig 4Samples in China that were collected in 1983 1995 and 2012 (100samples for each time point) are depicted (Chen et al 2013) Whilethe levels found in 1983 are comparable to those measured morerecently in Nepal and Bangladesh the decline in China is seen by1995 reflecting a transition in this population from a maize

Fig 4 Aflatoxin exposure levels vary widely across regions and time periods Thedata for 1995 and 2012 in China are from our recent publication (Chen et al 2013)

187JD Groopman et alFood and Chemical Toxicology 74 (2014) 184ndash189

metrics such as the aflatoxin B1-albumin biomarker in people Toaddress this data gap a quantitative isotope dilution mass spec-trometry method was deployed to measure the aflatoxin B1-lysinealbumin adduct at biologically critical times in mothers during preg-nancy and in the first thousand days of life in their children An initialanalysis of aflatoxin-albumin adducts in a pooled sample from onethousand 6ndash7 year old children in Nepal found 565 pg aflatoxin B1-lysinemg albumin a level comparable to that detected in verynutritionally compromised children in West Africa when the dif-ferent method of analyses are considered (Gong et al 2004 McCoyet al 2008) This biomarker observation prompted a more de-tailed examination using samples already analyzed for its plasmaproteome that had collected during the first and third trimester ofpregnancy of women enrolled in a randomized micronutrient in-tervention clinical trial in Sarlahi Nepal (Cole et al 2013 Schollet al 2012)

Aflatoxin B1 is a very potent toxic and carcinogenic agent for avariety of species including humans Contamination of staple grainsoccurs both prior to harvest and during storage of these food-stuffs Levels of contamination can vary widely resulting in humanexposures that can range from nanograms per day to milligrams perday depending upon the specific staple grain and amount con-sumed (Kensler et al 2011) While much of the international focuson this contamination has been framed within its carcinogenic prop-erties many of its other toxic effects can have both short and long-term health consequences In South Asia where both maternal andchild health status is tremendously compromised the potentialeffects of aflatoxin exposure are most likely manifested by these non-carcinogenic endpoints (Bhutta et al 2013 Black et al 2013)However in the future as early life health status improves acrossthis region these potential cancer endpoints could become muchmore consequential for these populations Nonetheless when af-latoxin exposure is determined to be substantial these data shouldbe viewed as being a sentinel for overall poor quality of staple grainsand nutrition (CAST 2003)

The wide heterogeneity in aflatoxin contamination levels in grainsresults in a very difficult task to assess human exposure using dietaryquestionnaires and spot testing of staple foods The developmentof mechanistically based biomarkers such as the aflatoxin B1-lysine albumin adduct has helped to overcome the inherentdifficulties of assessing aflatoxin exposure by measuring dietary con-stituents These biomarkers have the added advantage of being ableto integrate exposures over the course of many weeks or monthsdue to the long half-lives of these biomarkers in plasma (McCoy et al2005 2008) Research from our laboratory has also demonstratedthe long-term stability of aflatoxin albumin biomarkers in storedsamples where we have found these adducts to be stable for up to25 years (Scholl and Groopman 2008) The analytical technologyfor the measurement of these biomarkers has also substantially im-proved over the last several decades and comparative studies ofpaired samples have been done to relate measurements obtainedacross a number of analytical platforms The correlation of afla-toxin albumin adducts measured by mass spectrometry HPLC withfluorescence detection and immunoassays has been done and thispermits composite comparisons of human data obtained from manydiverse molecular epidemiologic studies across the globe (McCoyet al 2008 Scholl et al 2006) As a point of reference the valuesobtained in this comparative study revealed that the ELISA methodused provided aflatoxin adduct levels 5-fold higher than mass spec-trometry but the correlation coefficient between the two techniqueswas 098 (McCoy et al 2008) In the work reported in this studywe have employed a quantitative isotope dilution mass spectrom-etry method that permits the measurement of aflatoxin lysineobtained from serum albumin with the use of an isotopically labeledinternal standard This method is highly sensitive and specific per-mitting the quantitation of this biomarker in less than 100 μL of

plasma obtained in individuals residing in high exposure regionsof the world

Despite the emphasis of many investigations on liver cancer con-sequent to aflatoxin exposure there have been a number of studiesprimarily in West Africa that have raised the potential etiologic con-tribution of this toxin to diminished child growth and development(Gong et al 2002 2003 2004) Stunting is calculated to affect nearly200000000 children worldwide primarily in sub-Saharan Africa andSouth Asia (Black et al 2008 Smith et al 2012) Stunting has pro-found associations with a variety of growth and developmentdeficiencies in these children and to date the overall etiology of thisproblem remains obscure Since aflatoxin B1 has major impacts ongrowth and development in a number of animal species that hasbeen well characterized in the veterinary literature it is reasonablehypothesis that similar growth and developmental effects can occurin people The first step in the exploration of this hypothesis is thecharacterization of specific exposures in high-risk communities Thebiologically effective dose biomarker aflatoxin albumin adducts pro-vides an objective metric for characterizing these exposures forsubsequent follow-up and potential intervention

In an attempt to place the extent of exposures in the popula-tions reported here in Nepal and Bangladesh we compared thesecurrent data to those observed in other populations experiencingacute toxic (Kenya) or chronic (China) exposures (Chen et al 2013McCoy et al 2008) evaluated previously in our laboratory Theresults depicted in Fig 4 provide a composite perspective for theaflatoxin exposures detected in Nepal and Bangladesh in contextwith other documented exposures There was an acute aflatoxicosisoutbreak in Kenya in 2004 and again in 2005 in which nearly 150people died after consuming aflatoxin-contaminated maize Thelevels depicted measured in our laboratory were obtained from 107people living in those affected Kenyan villages (McCoy et al 2008Strosnider et al 2006) To provide perspective the median adductlevel was 10-times higher than the median for the 141 positivesamples of 150 pregnant women in Nepal whose blood was col-lected in 1999 However it is important to point out that many ofthe outlying samples in the Nepal data set overlap values of afla-toxin exposure in individuals from the Kenya toxicity event Thecomparative values in Bangladesh (from Fig 3) are shown nextFinally three sets of data from our studies in eastern China a regionof very high liver cancer incidence are shown in the far right of Fig 4Samples in China that were collected in 1983 1995 and 2012 (100samples for each time point) are depicted (Chen et al 2013) Whilethe levels found in 1983 are comparable to those measured morerecently in Nepal and Bangladesh the decline in China is seen by1995 reflecting a transition in this population from a maize

Fig 4 Aflatoxin exposure levels vary widely across regions and time periods Thedata for 1995 and 2012 in China are from our recent publication (Chen et al 2013)

187JD Groopman et alFood and Chemical Toxicology 74 (2014) 184ndash189

dominated diet to one in which rice quickly became the dominatestaple Maize is always more contaminated with aflatoxin com-pared to rice By 2012 virtually all of the samples collected in QidongChina were non-detectable for aflatoxin biomarkers This outcomedocuments the reduction in aflatoxin exposures that can be ob-tained through primary prevention in this case improved foodquality and diversity Finally we have found that this relatively recentreduction in aflatoxin exposure in Qidong China is associated witha nearly 50 reduction in liver cancer mortality among people underthe age of 40 (but not older) in this population suggesting stronglythat reduction of early life exposure can have a major protective effecton cancer outcomes

Data from several published studies can be utilized to calcu-late approximations of daily aflatoxin exposure in adults and childrenThese data sets include information gleaned from micro-dosing ex-periments in humans and plate food-based analyses of exposurewhere aflatoxinndashalbumin adduct analysis was also determined (Cupidet al 2004 Wild and Gong 2010 Wild et al 1992) Extrapolat-ing previously published findings with the aflatoxin B1 albuminadduct levels reported in this work indicates that daily exposurein women and their children range from 1 to 20 μg per day As-suming average body weight of 50 and 10 kg in women and childrenrespectively the adult exposures range from 10 to 200 ng per ki-logram body weight per day The childrenrsquos exposure wassubstantially higher due to lower body mass resulting in an expo-sure range of 50 to 1000 ng per kilogram body weight per day Theseexposure calculations provide a framework for determining tar-geted strategies to reduce dietary exposure to aflatoxin in thesepopulations

In conclusion in this study we have found that aflatoxin expo-sure is ubiquitous among at least some of the rural populations ofNepal and Bangladesh as assessed by a robust mass spectrometry-based albumin adduct biomarker High levels of aflatoxin B1-lysine adducts are observed consistently across the first and thirdtrimesters of pregnancy in both of these populations Aflatoxin B1-lysine albumin adducts in human cord blood samples demonstratethat the fetus has the capacity to metabolize aflatoxin to levels com-parable to that of the mother Substantial aflatoxin B1-lysine albuminadducts measured in the 2 year-old children equivalent to adultlevels strongly suggests life-time exposures to this potent hepa-totoxic and immunotoxic food contaminant occur These datastrongly suggest a need to further explore the sources of aflatoxincontamination in the staple grains to establish a process for inter-vention to lower these exposures These data strongly indicate theneed to develop interventions similar to those used in other high-risk high exposure settings

Conflict of interest

The authors declare that there are no conflicts of interest

Transparency document

The Transparency document associated with this article can befound in the online version

Acknowledgments

Supported in part by the Bill and Melinda Gates Foundation (GH614 and GCE 1046227 GCE 1046221 proteomics) Seattle WA USAand the US NIH P30 CA006973

References

Begum F Samajpati N 2000 Mycotoxin production on rice pulses and oilseedsNaturwissenschaften 87 275ndash277

Bhutta ZA Das JK Rizvi A Gaffey MF Walker N Horton S et al 2013Evidence-based interventions for improvement of maternal and child nutritionwhat can be done and at what cost Lancet 382 452ndash477

Black RE Allen LH Bhutta ZA Caulfield LE de Onis M Ezzati M et al 2008Maternal and child undernutrition global and regional exposures and healthconsequences Lancet 371 243ndash260

Black RE Victora CG Walker SP Bhutta ZA Christian P de Onis M et al 2013Maternal and child undernutrition and overweight in low-income and middle-income countries Lancet 382 427ndash451

CAST 2003 Mycotoxins risk in plants animals and human systems Council forAgricultural Science and Technology p 217

Chen JG Egner PA Ng D Jacobson LP Munoz A Zhu Y-R et al 2013 Reducedaflatoxin exposure presages decline in liver cancer mortality in an endemic regionof China Cancer Prev Res 6 1038ndash1045

Christian P Schulze K Stoltzfus RJ West KP Jr 1998 Hyporetinolemia illnesssymptoms and acute phase protein response in pregnant women with andwithout night blindness Am J Clin Nutr 67 1237ndash1243

Christian P Khatry SK Katz J Pradhan EK LeClerq SC Shrestha SR et al 2003aEffects of alternative maternal micronutrient supplements on low birth weightin rural Nepal double blind randomised community trial BMJ 326 571

Christian P West KP Jr Khatry SK Leclerq SC Pradhan EK Katz J et al 2003bEffects of maternal micronutrient supplementation on fetal loss and infantmortality a cluster-randomized trial in Nepal Am J Clin Nutr 78 1194ndash1202

Cole RN Ruczinski I Schulze K Christian P Herbrich S Wu L et al 2013 Theplasma proteome identifies expected and novel proteins correlated withmicronutrient status in undernourished Nepalese children J Nutr 143 1540ndash1548 doi 15103945jn1113175018 Epub 172013 Aug 175021

Cupid BC Lightfoot TJ Russell D Gant SJ Turner PC Dingley KH et al 2004The formation of AFB 1-macromolecular adducts in rats and humans at dietarylevels of exposure Food Chem Toxicol 42 559ndash569

Dawlatana M Coker RD Nagler MJ Wild CP Hassan MS Blunden G 2002The occurrence of mycotoxins in key commodities in Bangladesh surveillanceresults from 1993 to 1995 J Nat Toxins 11 379ndash386

Gong Y Hounsa A Egal S Turner PC Sutcliffe AE Hall AJ et al 2004Postweaning exposure to aflatoxin results in impaired child growth a longitudinalstudy in Benin West Africa Environ Health Perspect 112 1334ndash1338

Gong YY Cardwell K Hounsa A Egal S Turner PC Hall AJ et al 2002 Dietaryaflatoxin exposure and impaired growth in young children from Benin and Togocross sectional study BMJ 325 20ndash21

Gong YY Egal S Hounsa A Turner PC Hall AJ Cardwell KF et al 2003Determinants of aflatoxin exposure in young children from Benin and Togo WestAfrica the critical role of weaning Int J Epidemiol 32 556ndash562

Guengerich FP 2008 Cytochrome p450 and chemical toxicology Chem Res Toxicol21 70ndash83

Katz J Lee AC Kozuki N Lawn JE Cousens S Blencowe H et al 2013 Mortalityrisk in preterm and small-for-gestational-age infants in low-income andmiddle-income countries a pooled country analysis Lancet 382 417ndash425

Kensler TW Roebuck BD Wogan GN Groopman JD 2011 Aflatoxin a 50-yearodyssey of mechanistic and translational toxicology Toxicol Sci 120 (Suppl 1)S28ndashS48

Khlangwiset P Shephard GS Wu F 2011 Aflatoxins and growth impairment areview Crit Rev Toxicol 41 740ndash755

Koirala P Kumar S Yadav BK Premarajan KC 2005 Occurrence of aflatoxin insome of the food and feed in Nepal Indian J Med Sci 59 331ndash336

Krishnamachari KA Bhat VR Nagarajan V Tilak TB Tulpule PG 1977 Theproblem of aflatoxic human disease in parts of India-epidemiological andecological aspects Ann Nutr Aliment 31 991ndash996

McCoy LF Scholl PF Schleicher RL Groopman JD Powers CD Pfeiffer CM2005 Analysis of aflatoxin B1-lysine adduct in serum using isotope-dilution liquidchromatographytandem mass spectrometry Rapid Commun Mass Spectrom19 2203ndash2210

McCoy LF Scholl PF Sutcliffe AE Kieszak SM Powers CD Rogers HS et al2008 Human aflatoxin albumin adducts quantitatively compared by ELISA HPLCwith fluorescence detection and HPLC with isotope dilution mass spectrometryCancer Epidemiol Biomarkers Prev 17 1653ndash1657

Okumura H Kawamura O Kishimoto S Hasegawa A Shrestha SM Okuda Ket al 1993 Aflatoxin M1 in Nepalese sera quantified by combination ofmonoclonal antibody immunoaffinity chromatography and enzyme-linkedimmunosorbent assay Carcinogenesis 14 1233ndash1235

Sabbioni G Ambs S Wogan GN Groopman JD 1990 The aflatoxin B1-lysineadduct quantified by high-performance liquid chromatography from humanserum albumin samples Carcinogenesis 11 2063ndash2066

Scholl PF Groopman JD 2008 Long-term stability of human aflatoxin B1 albuminadducts assessed by isotope dilution mass spectrometry and high-performanceliquid chromatography-fluorescence Cancer Epidemiol Biomarkers Prev 171436ndash1439

Scholl PF Turner PC Sutcliffe AE Sylla A Diallo MS Friesen MD et al 2006Quantitative comparison of aflatoxin B1 serum albumin adducts in humans byisotope dilution mass spectrometry and ELISA Cancer Epidemiol BiomarkersPrev 15 823ndash826

Scholl PF Cole RN Ruczinski I Gucek M Diez R Rennie A et al 2012 Maternalserum proteome changes between the first and third trimester of pregnancy inrural southern Nepal Placenta 33 424ndash432

Smith LE Stoltzfus RJ Prendergast A 2012 Food chain mycotoxin exposure guthealth and impaired growth a conceptual framework Adv Nutr 3 526ndash531

188 JD Groopman et alFood and Chemical Toxicology 74 (2014) 184ndash189

Strosnider H Azziz-Baumgartner E Banziger M Bhat RV Breiman R BruneMN et al 2006 Workgroup report public health strategies for reducing aflatoxinexposure in developing countries Environ Health Perspect 114 1898ndash1903

Tandon HD Tandon BN Ramalingaswami V 1978 Epidemic of toxic hepatitisin India of possible mycotoxic origin Arch Pathol Lab Med 102 372ndash376

Turner PC Sylla A Gong YY Diallo MS Sutcliffe AE Hall AJ et al 2005Reduction in exposure to carcinogenic aflatoxins by postharvest interventionmeasures in West Africa a community-based intervention study Lancet 3651950ndash1956

Turner PC Collinson AC Cheung YB Gong Y Hall AJ Prentice AM et al 2007Aflatoxin exposure in utero causes growth faltering in Gambian infants Int JEpidemiol 36 1119ndash1125

Wang LY Hatch M Chen CJ Levin B You SL Lu SN et al 1996 Aflatoxinexposure and risk of hepatocellular carcinoma in Taiwan Int J Cancer 67620ndash625

West KP Shamim AA Labrique AB Ali H Shaikh S Mehra S et al 2013Efficacy of antenatal multiple micronutrient (MM) vs iron-folic acid (ifa)supplementation in improving gestational and postnatal viability in ruralBangladesh the JiVitA-3 Trial FASEB J 27

Wild CP Gong YY 2010 Mycotoxins and human disease a largely ignored globalhealth issue Carcinogenesis 31 71ndash82

Wild CP Hudson GJ Sabbioni G Chapot B Hall AJ Wogan GN et al 1992Dietary intake of aflatoxins and the level of albumin-bound aflatoxin in peripheralblood in The Gambia West Africa Cancer Epidemiol Biomarkers Prev 1 229ndash234

Wogan GN Kensler TW Groopman JD 2012 Present and future directions oftranslational research on aflatoxin and hepatocellular carcinoma a review FoodAddit Contam Part A Chem Anal Control Expo Risk Assess 29 249ndash257

Yadgiri B Reddy V Tulpule PG Srikantia SG Gopalan C 1970 Aflatoxin andIndian childhood cirrhosis Am J Clin Nutr 23 94ndash98

189JD Groopman et alFood and Chemical Toxicology 74 (2014) 184ndash189