Abstract
BACKGROUND Infants are a vulnerable population whose nutrition changes as complementary foods are introduced; a process which may modify patterns of exposure to dietary mycotoxins. However, exposure monitoring of dietary mycotoxin mixtures in several biological samples obtained from breastfed and non-exclusively breastfed children is scarce.
OBJECTIVES To examine mycotoxin co-exposure patterns in infants using a multi-specimen, multi-mycotoxin approach.
METHODS Breast milk, complementary food and urine obtained from 65 infants, aged 1–18 months, in Ogun state, Nigeria, were analyzed for mycotoxins using ultra-sensitive LC-MS/MS approaches.
RESULTS Complementary food was contaminated with seven distinct classes of mycotoxins including aflatoxins (9/42 samples; range: 1.0–16.2µg/kg) and fumonisins (14/42; range: 8– 167µg/kg). Aflatoxin M1 was detected in breast milk (4/22), while six other classes of mycotoxins were quantified; including dihydrocitrinone (6/22; range: 14.0–59.7ng/L) and sterigmatocystin (1/22; 1.2ng/L) detected for the first time. Mycotoxins were detected in 64/65 of the urine samples, with seven distinct classes of mycotoxins observed demonstrating ubiquitous exposure. Two aflatoxin metabolites (AFM1 and AFQ1) and FB1 were detected in 6/65, 44/65 and 17/65 samples, respectively. Mixtures of mycotoxin classes were common, including 14/42, 22/22 and 56/65 samples having 2–4, 2–6 or 2–6 mycotoxins present, for complementary food, breast milk and urine, respectively. Aflatoxin and/or fumonisin was detected in 12/14, 4/22 and 46/56 for complimentary foods, breast milk and urine, respectively. Furthermore, the detection frequency, mean concentrations and occurrence of mixtures were typically greater in urine of non-exclusively breastfed compared to breastfed children.
CONCLUSIONS The study provides novel insights into mycotoxin co-exposures in children in a mycotoxin high-risk country without proper food safety measures. Albeit a small sample set, it highlights significant transition to higher levels of infant mycotoxin exposure as complementary foods are introduced, providing impetus to mitigate during this critical early-life period and encourage breastfeeding.
Introduction
Nutrition within the first 1000 days of life is crucial to the growth, development and performance of children in their later years and into adulthood as postulated in the ‘developmental origins of health and disease’ (DOHaD) concept (Barker 2000; Bateson et al. 2004; Mandy and Nyirenda 2018). However, children’s diets within this period may be diverse and complex, and depend on the age of the children. The diet categories can include exclusively consumed breast milk, breast milk combined with complementary foods, and exclusively consumed complementary foods. Human breast milk composition varies widely depending on diet, age, lactation stage, number of pregnancies, and other physiological parameters (Innis 2014; Butts et al. 2018; Wu et al. 2018). Complementary foods are mainly made from single or mixed cereals as a porridge, sometimes in combination with nuts, animal milk and other mashed items. These foods also become more diverse as children get older (Kimanya et al. 2009, 2010, 2014; Alvito et al. 2010; Kamala et al. 2016; Chuisseu Njamen et al. 2018; Ojuri et al. 2018, 2019). Unfortunately, in some parts of the world, such foods may contain a number of fungal metabolites known as mycotoxins (Kimanya et al. 2009; 2010, 2014; Alvito et al. 2010; Juan et al. 2014; Cherkani-Hassani et al. 2016; Kamala et al. 2016; Warth et al. 2016; Chuisseu Njamen et al. 2018; Braun et al. 2018; Ojuri et al. 2018, 2019; Preindl et al. 2019).
Mycotoxins such as aflatoxin (AF), fumonisin (FUM), ochratoxin A (OTA), deoxynivalenol (DON), and zearalenone (ZEN) are considered priority mycotoxins and are regulated in food. AFs are categorized as class 1 human liver carcinogen (IARC 2002, 2012a). AF is additionally implicated in immune system suppression, interference with micronutrient absorption, impairment of gut integrity and reduction in normal child growth (Gong et al. 2002, 2003, 2004, 2012; Turner et al. 2003, 2007; Turner 2013; IARC 2015; Shirima et al. 2015; Watson et al. 2016, 2018). FUM intake has been associated with birth defects (Missmer et al. 2006), poor child growth (Shirima et al. 2015) and the incidence of esophageal cancer (Chu et al. 1994; Yoshizawa et al. 1994; Sun et al. 2007, 2011); OTA is linked to renal toxicity (Heussner and Bingle, 2015), while immune system modulation was reported for trichothecenes such as DON (Pestka 2010). Furthermore, in vitro data suggest interactions of regulated mycotoxins with emerging ones such as alternariol (AOH) and its methylated ether form (alternariol monomethyl ether; AME), beauvericin (BEA), enniatins and moniliformin (MON) (Vejdovszky et al. 2017a, 2017b; Aichinger et al 2019; Woelflingseder 2019).
Complementary foods in Africa can be contaminated with mycotoxins (Chuisseu Djamen et al. 2018; Kimanya et al. 2009, 2010, 2014; Juan et al. 2014; Kolakowski et al. 2016; Cappozzo et al. 2017; Oueslati et al. 2018; Ojuri et al. 2018, 2019), and ingested mycotoxins or their metabolites may be found in various bio-fluids including blood, urine and breast milk (Cherkani-Hassani et al. 2016; Warth et al. 2016; Braun et al. 2018; Ferrufino-Guardia et al. 2019; Polychronaki et al. 2008; Ediage et al. 2013; Ezekiel et al. 2014; Shirima et al. 2015; Gerding et al. 2015; Heyndrickx et al. 2015; Schwartzbord et al. 2016; Ayelign et al. 2017; Chen et al. 2017; Papageorgiou et al. 2018; Sanchez and Diaz 2019, Turner et al. 2012a; Vidal et al. 2018). Only a limited number of studies have assessed multiple mycotoxin exposure in Nigerian children using biomarkers of exposure (Adejumo et al. 2013; Ezekiel et al. 2014, 2018b; Braun et al. 2018; McMillan et al. 2018; Sarkanj et al. 2018), while none in any country has assessed infant’s exposures to multi-mycotoxins using multiple specimens (complementary foods and breast milk) and additionally via urine.
This study comprehensively elucidated mycotoxin co-exposure patterns in infants using three measures (food, breast-milk and urine) all with quantitative multiple mycotoxin analysis to compare exposure levels in exclusively breastfed (EB) with non-exclusively breastfed (NEB) infants.
Methods
Study area and population
The study population was 65 infants aged 1–18 months from Ilishan and Ikenne in Remo land of Ogun state (south-western Nigeria). Both are small sister communities within 1 km distance of each other, with rich diversity of foods available in households. Both communities are semi-urban due to the presence of campuses of tertiary institutions and they share similar characteristics as detailed below. Indigent families in these communities are farmers and traders, and both communities consist of approximately 15,000 residents. The food sources for families are local markets, small stores and own farms. Major foods consumed are cereal-based products (rice, maize, wheat), tubers (cassava and yam), and groundnuts, melon seeds and cowpea. The predominant breastfeeding practice in both communities involves the introduction of complementary foods from the first month after birth; though timing depends on family income, mother’s health and work status, and knowledge on breastfeeding practices. A majority of the families associated with the tertiary institutions practice exclusive breastfeeding from birth till end of the sixth month.
Sampling design and ethical considerations
This pilot, cross-sectional survey was conducted between January and February 2016. Families that participated were identified through local health centres and post-natal clinics in the two communities where mother/infant pairs were registered. The study consisted of two cohorts: exclusively breastfed (EB) children (age: 1–6 months) and non-exclusively breastfed (NEB) children (age: 3–18 months) to which 23 and 42 children (female/male ratio = 1:1) were recruited, respectively. A total of 50 and 15 infants were recruited into the study from Ilishan and Ikenne communities, respectively. The age of infants recruited into the study were ascertained from their health records. Infants with known medical conditions (e.g. jaundice, HIV positive or visibly malnourished) were excluded during the recruitment stage as it was the purpose of this study to investigate typical background exposure levels and compare EB with NEB rather than targeting the impact of exposure on health status. Prior to recruitment or inclusion into the study, the mother of each child was informed in her preferred language (English or Yoruba) on the purpose of the study. Only mothers who signed the informed consent document (and on behalf of their infant) to participate in the research were included into the study.
Complementary foods were obtained as “plate-ready” samples (n = 42; 50 g each), by a trained team member, for the NEB children (n = 42) on the day prior to infant urine sample collection. The food samples collected included industrially-processed infant cereal (n = 6), fermented maize gruel ogi (n = 26) and tombran (home-made pudding from mixed cereal and nut; n = 10). One food sample was collected per infant. Breast milk samples (n = 22; 3–5 mL each) from mothers of EB infants were collected on the day preceding the sampling of infant urine. One mother did not provide breast milk sample due to intake of medications. Other details of sampling and handling of the breast milk samples were described by Braun et al. (2018). In this paper, the breast milk samples have been previously analysed for mycotoxins utilizing a less sensitive method. First morning urine samples (10–40 mL each) were collected in 50 falcon tubes from all 65 infants recruited into the study by the mothers, using their experience and knowledge of the time at which their infant pass urine in the morning. All samples (plate-ready complementary food, breast milk and urine) were immediately frozen at −20 °C and couriered on dry ice to Austria for LC-MS/MS analyses.
Ethical permission for this study was granted by the Babcock University Health Research Ethics Committee under two approval numbers: BUHREC294/16 for the EB children and BUHREC156/15 for the NEB children. In adherence to ethical standards, all samples were blinded and coded to exclude participant information prior to analytical measurements.
Food consumption questionnaires
The mothers received a copy of a well-structured questionnaire for completion prior to food, breast milk and urine sample collection. The questionnaires were administered by trained interviewers and translation to the local language (Yoruba) was available on request of some mothers. The administered questionnaire was designed to elicit information on participant demography, family socio-economic status, individual dietary preferences and food consumption patterns and associated health implications among the infants.
Mycotoxin determination in breast milk samples
Mycotoxins were quantified in breast milk samples as described in Braun et al. (2020). Briefly, each homogenized breast milk sample (1 mL) was extracted using 1 mL acidified ACN (1% formic acid) on a shaker for 3 min, followed by addition of anhydrous magnesium sulfate (0.4 g) and sodium chloride (0.1 g). Sequel to two rounds of cold centrifugation (4,750 x g at 10 °C for 10 min; 14,000 × g, 4 °C for 2 min), a SPE clean up step using SPE column (Oasis PRiME HLB®, Waters, Milford, MA, USA) followed. Mycotoxins were then concentrated on a vacuum concentrator (Labconco, Missouri, USA) prior to reconstitution using 81 µL MeOH/ACN and 9 µL of the internal standard mixture. Mycotoxins in breast milk samples were measured on a Sciex QTrap 6500+ LC-MS/MS system (Foster City, CA, USA) equipped with a Turbo-V™ electrospray ionization (ESI) source coupled to an Agilent 1290 series UHPLC system (Waldbronn, Germany). Analytes were separated on an Acquity HSS T3 column (2.1×100mm; Waters, Vienna, Austria) with 1.8 µm particle size at 40 °C. ESI-MS/MS was performed in scheduled multiple reaction monitoring (MRM) mode using fast polarity switching. For each analyte two individual mass transitions were acquired.
LC-MS/MS-based mycotoxin analysis of food samples
Plate-ready complementary food samples were analysed for the presence of multiple mycotoxins according to the dilute and shoot LC-MS/MS-based method described by Sulyok et al. (2020). This method covers the detection of more than 500 metabolites, inclusive of mycotoxins, in food samples. Briefly, 5 g of each food sample was extracted with 20 mL of acetonitrile/water/acetic acid 79:20:1, (v/v/v) in a 50 mL polypropylene tube (Sarstedt, Nümbrecht, Germany) for 90 min on a GFL 3017 rotary shaker (GFL, Burgwedel, Germany). Afterwards, the extract was diluted 1:1 (v/v) in dilution solvent (acetonitrile/water/acetic acid 20:79:1, v/v/v) prior to measurement on a Sciex QTrap 5500 LC-MS/MS System (Applied Biosystem, Foster City, CA, USA) equipped with TurboIonSpray electrospray ionisation (ESI) source and a 1290 Series HPLC System (Agilent, Waldbronn, Germany). Chromatographic separation was performed at 25°C on a Gemini® C18-column, 150×4.6 mm i.d., 5 μm particle size, equipped with a C18 4×3 mm i.d. security guard cartridge (Phenomenex, Torrance, CA, USA). ESI-MS/MS was performed in the scheduled MRM mode both in positive and negative polarities in two separate chromatographic runs per sample by scanning two fragmentation reactions per analyte. The MRM detection window of each analyte was set to its expected retention time ±20 and ±26 s in the positive and the negative modes, respectively. The identification of each positive analyte was confirmed when two MRMs per analyte was obtained (EC 2002). Further details on the method performance including data obtained through inter-laboratory comparison are described in Sulyok et al. (2020).
Urinary mycotoxin measurements by UPLC-MS-MS
Mycotoxins in the urine samples were determined following a stable-isotope dilution assay-based LC-MS/MS method described in detail by Sarkanj et al. (2018). Briefly, each urine sample was centrifuged for 3 min at 5600 × g, then treated β-glucuronidase from E. coli Type IX-A (Sigma-Aldrich, G7396-2MU) of prior to a SPE cleanup on Oasis PRiMEHLB® SPE columns (Waters, Milford, MA, USA). Extracts were then evaporated under nitrogen, reconstituted with 470 µL dilution solvent (10% acetonitrile, 0.1% glacial acetic acid) and fortified with 30 µL of an IS mixture. Measurement of mycotoxins in urine samples was performed on a Sciex QTrap 6500+ LC-MS/MS system (Foster City, CA) equipped with a Turbo-V™ ESI source coupled to an Agilent 1290 series UHPLC system (Waldbronn, Germany). Analytes were separated on an AtlantisT3 HSS column (2.1×100mm; Waters, Wexford, Ireland) with 1.8 µm particle size at 35 °C. ESI-MS/MS was performed in scheduled MRM mode and with a 180 sec detection window. At least two individual transitions were monitored for each analyte.
Data analysis
The IBM SPSS Statistics v21.0 (SPSS, Inc., Chicago, IL, USA) was applied in the analyses of all data. Boxplots were created for the distribution of urinary mycotoxins in the two cohorts of children. Urinary mycotoxin concentrations were logarithm transformed by the equation y = Log10 (1 + ng/L) to create a normal distribution for the boxplots. For comparison of urinary mycotoxin levels in the population, means were separated by the Duncan’s Multiple Range test (DMRT) and tested for significance by analysis of variance (ANOVA) at a = 0.05. Whereas for the comparison of all other data between the two cohorts of children in the study, the independent sample t-test using the Levene’s test for equality of variances were applied to test for significance at a = 0.05.
Results
Demography and food preference of children
The demographic characteristics of the infants as well as their food consumption data are given in Table 1. The NEB infants were significantly (p<0.01) older (9.0±3.6 months) and heavier (8.0±1.6 Kg; p<0.05) than the EB children (3.7±1.6 months and 6.3±1.0 Kg, respectively). All infants received breast milk at least once daily, with higher consumption frequency of 8–10 times per day in the EB cohort compared to 1–3 times per day in the NEB cohort who received complementary food at least thrice per day. NEB infants received a variety of foods, in most cases (>90%) including infant cereal (mix of animal milk with maize, rice, or wheat), ogi (fermented maize gruel) and tombran (mixed cereal and nut pudding).
Occurrence of mycotoxins in breast milk
Breast milk samples were examined for the presence of 34 mycotoxins (or their metabolites), of which nine were observed. All 22 breast milk samples contained mycotoxins although mostly at very low concentrations (Table 2). The most frequently detected mycotoxins were BEA (incidence: 100%; range: 0.99–12.0 ng/L; mean: 3.0 ng/L) and AME (incidence: 96%; range: 0.50–11.7 ng/L; mean: 3.03 ng/L), whereas DHC (incidence: 27%; range: 14.0–59.7 ng/L; mean: 25.2 ng/L) and OTA (incidence: 64%; range: 2.0–67.6 ng/L; mean: 9.58 ng/L) had higher mean levels than all other mycotoxins found in the samples. AFM1 was the only aflatoxin metabolite detected in breast milk, occurring in 18% of the samples at levels below the LOQ; and thus, the data points were assigned LOQ/2 values of 2.0 ng/L (Table 2). Enniatins (EnnB and EnnB1) occurred in 77% of the samples, with EnnB dominating in 73% of samples at mean concentration of 5.0 ng/L (range: 0.70–10.1 ng/L). Sterigmatocystin (STER) was detected in one sample at 1.20 ng/L. The chromatograms of DHC and STER detected in breast milk are shown in Figure 1 (C–D).
Distribution of mycotoxins in plate-ready complementary food
Sixty-five fungal metabolites, of which 14 were mycotoxins (Table 3), and 25 other secondary metabolites (Table S1) were quantified in the samples of plate-ready complementary meals. Aflatoxin B1, B2 and G1 were detected, but not AFG2. Total AFs contaminated 21% of food samples of which the mean was 6.0 µg/kg (range: 1.0–16.0 µg/kg). The mean level of AFG1 (12 µg/kg) deriving from two ogi samples was three-fold higher than that of AFB1 (3 µg/kg). The incidence of FUM (sum of fumonisins B1 and B2) was 29%, of which the mean was: 49 µg/kg (range 7.9–194 µg/kg). BEA (incidence: 79%; mean: 3.8 µg/kg) was the most frequently detected mycotoxin in the food samples. Enniatins, ZEN, MON and DON also occurred in 19, 10, 7 and 5% of the foods, respectively. With respect to occurrence of the toxins in food types, the highest mean concentrations of BEA (19 µg/kg), enniatins (10 µg/kg) and FUM (63 µg/kg), were found in infant cereal. Of the three food types, ogi contained the highest levels of AF (6.5 µg/kg) and DON (62 µg/kg). AF and MON were not found in infant cereal. Ogi and tombran contained more diverse mycotoxins than infant cereal (Table 3).
Mycotoxins in urines of infants
Overview of the distribution of mycotoxins in infant urines
At least one mycotoxin was detected in 64/65 (98.5%) of the urine samples, with 10 distinct mycotoxins observed, including aflatoxins (AFM1 and AFQ1), fumonisin B1, DON, DHC, NIV, OTA, zearalenones (ZEN, α-ZEL, and β-ZEL), representing seven classes of mycotoxin, see Table 4. The detected aflatoxins were two metabolites: AFM1 (6/65, 9% of samples; mean of detected (MoD): 140 ng/L; range: 20–510 ng/L) and AFQ1 (44/65, 68% of samples; MoD: 210 ng/L; range: 3–900ng/L). Two infants had only AFM1 detected, with no AFQ1, thus in total 46/65 (71%) of the infants had been exposed to dietary aflatoxin. Only one of the fumonisin class of mycotoxins was detected, FB1 (17/65, 26% of samples; MoD: 810 ng/L; range: 40–1590 ng/L). OTA, ZEN, DON and DHC were also frequently detected (Table 4). The chromatograms of AFQ1 and DHC detected in urine are given in Figure 1(A–B). Mycotoxins in urine tended to be higher in females versus males, though no statistically significant differences in frequency or concentration for individual toxins was observed (data not shown).
Mycotoxins in exclusively breastfed (EB) and non-exclusively breastfed (NEB) infants’ urine
Eight of the 10 mycotoxins were observed frequently in NEB infants compared to EB infants: AFM1 (12 vs 4%), DON (55 vs 30%), DHC (74 vs 26%), FB1 (31 vs 17%), OTA (38 vs 26%), ZEN (83 vs 57%), respectively, and NIV and α-ZEL which were only detected in NEB infants’ urine (Table 4). The recorded detection frequencies of the mycotoxins resulted in an 46–300% increase from EB to NEB cohorts (Figure 2). There were however reductions in the frequencies of detection for AFQ1 (53 vs 87%) and β-ZEL (14 vs 30%). The mean concentrations in urine with detectable mycotoxins were greater for seven of the 10 mycotoxins, in NEB versus EB infants: AFM1 (166 vs 23 ng/L), AFQ1 (225 vs 182 ng/L), DON (5,282 vs 3,194 ng/L), DHC (202 vs 135 ng/L), FB1 (454 vs 136 ng/L), NIV (161 ng/L vs non-detect) and α-ZEL (20 ng/L vs non detect), respectively, see Table 4; resulting in 124–721% increase (Figure 2). However, significant higher mean levels were only recorded for AFQ1, DON, DHC and FB1 in favour of NEB infants, and for OTA in favour of EB infants (Table 4 and Figure S1).
Mycotoxin co-occurrence and co-exposures
The numbers of different classes of mycotoxins detected in each of breast milk, complementary food and urine were determined (Table 5). All breast milk samples contained more than one and up to five mycotoxin classes; 41% of samples had four or more different mycotoxins detected, though overall mycotoxin concentrations were low. The primary combinations in the mycotoxin 305 mixtures in breast milk included AME, BEA enniatins and OTA. Aflatoxins were observed as part of 4/22 mixtures of toxins, and fumonisins were not measured in this study.
Mycotoxin mixtures, of 2–5 toxins, were observed in just over half (22/42) of the plate-ready complementary food samples. Mixtures predominantly involved three classes of mycotoxins: BEA in 12 mixtures, FUM in nine mixtures and AF in eight mixtures. Of these, five samples with three or more toxins contained both AF and FUM.
Mixtures of 2–6 mycotoxins were observed in about 86% of the urines, with NEB typically having more urine samples with mixtures (93%, with 2–6 toxins) compared to EB infants (74% with 2–5 toxins). No urine from EB infants had six mycotoxins detected. The most frequently occurring classes of toxins in mixtures were ZEN (45/65), AF (42/65), DHC (41/65) and DON (30/65). AF and FB in the same urine occurred in 13/65 (20%). One infant from the NEB cohort was co-exposed to seven mycotoxins and metabolites belonging to six of the assessed mycotoxin classes except NIV (Figure 3).
Discussion
Infants transition from breastfeeding to complementary food is essential for nutritional requirements, however, it is also important to understand if this transition creates opportunities for differential exposure to diverse classes of fungal toxins and other environmental contaminants (i.e. the exposome). The diversity of mycotoxins in breast milk is still in the exploratory stage, though several mycotoxins have been observed, most notably aflatoxins; however, food measurements are by far more established to date despite several limitations (IARC 2012a, 2012b, 2015). This study provides a measure of multiple mycotoxins in both breast milk and complementary plate ready foods, and then compares patterns of these mycotoxins in urine from infants that are either EB or partially receiving breast milk and complementary foods, NEB.
In this study the analytical sensitivity for measuring mycotoxins in breast milk was extremely low, and thus the frequency of toxin detection was high providing unique insights into chronic background level exposure patterns. All 22 samples had two or more toxins, though overall the concentrations would be regarded as low. The lipophilic aflatoxins such as AFB1 or AFG1 are commonly detected in grains in sub-Saharan Africa (Ezekiel et al. 2018a), but were not observed in breastmilk. The AFB1 metabolite, AFM1, was observed in four samples and at a level below the LOQ (4 ng/L). This data suggests that breast milk is not providing significant exposure to the most carcinogenic of the mycotoxins for participants of this study. AFM1 in breast milk occurred at lower mean concentrations than those previously reported from Sierra Leone (800 ng/L, Jonsyn et al. 1995), Egypt (60 ng/L, Polychronaki et al. 2007), Nigeria (35 ng/L, Adejumo et al. 2013) and Ecuador (45 ng/L, Ortiz et al. 2018), but similar to levels reported in two samples from Brazil (0.3 and 0.8 ng/L, Iha et al. 2014).
Prior to the present study, there are no reports on the occurrence of CIT and its metabolites as well as STER in human breast milk (Cherkani-Hassani et al. 2016; Warth et al. 2016; Ali and Degen 2019). Thus, to the best of our knowledge, we present the first report of DHC and STER occurrence in human breast milk. DHC, mostly considered a detoxification product of CIT but also co-detected with CIT in food samples (Ojuri et al. 2018, 2019; Ezekiel et al. 2020), is indicative of CIT exposure (Follmann et al. 2014; Ali and Degen, 2019). The concentrations of DHC in breast milk samples analyzed in the present study were low, and as much as 10 times lower than the concentrations found in urine. However, the presence of DHC in both biological fluids compel us to hypothesize that CIT intake is common in Nigeria (Ojuri et al. 2018, 2019; Akinfala et al. 2020; Ezekiel et al. 2020). Furthermore, it agrees with the suggestions of Ali and Degen (2019) who retrospectively estimated high CIT exposure assessment/intake above the “level of no concern for nephrotoxicity” of 0.2 μg/kg bwt/day set by EFSA (EFSA 2012; EC 2014) in participants in a previous urinary biomarker study in Nigeria (Sarkanj et al. 2018). Due to sparse data on its occurrence and toxicity, there is no regulation on maximum levels (ML) of STER in food, including infant diets, in most countries, Nigeria inclusive (EC 2006; Oplatowska-Stachowiak et al. 2018). Nonetheless, this mycotoxin may be carcinogenic in humans (IARC 1987; EFSA 2013), as such, its detection in breast milk samples provides a basis for the consideration of more investigations into the occurrence and regulation of this mycotoxin in foods intended for infants and young children.
The mean level of the nephrotoxic OTA in the analyzed breast milk was 50 times lower than the maximum tolerable limit of 500 ng/L set for infant food in Europe (EC 2006). Additionally, the recorded level was four to five times lower than those reported in Bolivia (mean: 53 ng/L, Ferrufino-Guardia et al. 2019) and Chile (mean: 44 ng/L, Munoz et al. 2014) but nearly two times higher than the mean level reported in Italy (mean: 6.01 ng/L, Turconi et al. 2004). The identification of BEA and EnnB in human breast milk agrees with previous reports from Nigeria (Braun et al. 2018) and Spain (Rubert et al. 2014). The concentration level of AME in the present study is, however, lower than those reported in a previous study from Nigeria (Braun et al. 2020). Due to poor analytical recoveries at the outset, fumonisins were not included in the breast milk analysis.
Mycotoxins have been previously reported in complementary foods (Juan et al. 2014; Kolakowski et al. 2016; Kamala et al. 2016; Cappozzo et al. 2017; Chuisseu Njamen et al. 2018; Oueslati et al. 2018; Ojuri et al. 2018, 2019). The detection of more diverse spectra of mycotoxins (AF, BEA, EnnB, FUM, MON and ZEN) in the traditionally processed complementary foods (ogi and tombran) compared to the industrially processed infant cereal agrees with an earlier report from our group on mycotoxins in complementary foods from Nigeria (Ojuri et al. 2018, 2019). AFs were detected in about 20% of the food samples, but notably not in the infant cereal. The mean of these samples with detectable AFs in traditionally processed complementary foods exceeded the 4 µg/kg maximum level for total AF in Nigerian baby food. The presence of AF in locally processed complementary foods but not in commercial infant cereal is in keeping with the report from Cameroon (Chuisseu Njamen et al. 2018). However, the levels found in this present study were lower than those previously reported in locally processed baby foods from Nigeria (mean: 104 µg/kg, Ojuri et al. 2018), Cameroon (mean: 177 ug/kg, Chuisseu Njamen et al. 2018) but similar to Ghana (mean: 8.3 µg/kg, Blankson et al. 2019). Unlike most of the aforementioned reports, herein we analyzed plate-ready meal for a more precise food contamination data because it considers food ingested after final preparatory steps such as dilution and cooking (Chuisseu Njamen et al. 2018; Ezekiel et al. 2019). Clearly, the methods employed at household levels, often characterized by poor food processing practices, influenced the mycotoxin contents in the traditionally processed complementary foods. The fact that we could not detect AF in the infant cereal could be attributed to tailoring of production by industrial processors to meet regulatory standards for AF.
FUM was detected in all the complementary food types. However, the mean concentrations we found were lower than the European Union ML of 200 µg/kg for baby food (EC 2006) and the mean level (2808 µg/kg) reported in Tanzania (Kamala et al. 2016). FUM is not regulated in baby foods in Nigeria. Nevertheless, the observed co-occurrence of FUM with AF in about 20% of the complementary foods is of concern, because these two mycotoxins are suggested to potentially interact to negatively influence linear growth in children (IARC 2015; Shirima et al. 2015). Meanwhile, a potential role of additional toxins in mixtures is unknown.
Several mycotoxins were detected in the children’s urine similar to previous reports from Cameroon (Ediage et al. 2013), Nigeria (Ezekiel et al. 2014), Spain (Rodriguez-Carrasco et al. 2014) and Belgium (Heyndrickx et al. 2015) where more than two mycotoxins were detected. We compared the incidences of commonly found mycotoxins in children urine with data from previous studies. Overall, ZEN (incidence: 74%), a potent estrogenic mycotoxin able to cross the human placenta (Warth et al. 2019), was the most frequently detected mycotoxin in the urine samples. Two ZEN metabolites, α-ZEL and β-ZEL, were detected in the urine samples albeit at lower frequencies (11% and 20%, respectively) compared to the parent compound. This is in contrast to the reports of Heyndrickx et al. (2015) wherein ZEN and its metabolites were not detected in children urine from Belgium. The incidence of ZEN in the present study was, however, higher than the incidence (4%, 9/220) previously reported in Cameroonian children employing a method of lower sensitivity (Ediage et al. 2013). Similarly, DHC was found in more than one half (57%) of the 65 children in the present study. This finding is much higher than the 6% (7/124) and 14% (20/142) incidences reported for DHC in urine of Belgian children (Heyndrickx et al. 2015) and those of adults and children in Haiti (Gerding et al. 2015), respectively. We did not detect the parent toxin CIT in any urine sample, which agrees with the reports from Haiti but negates the Belgian study that found both parent compound (CIT) and its metabolite (DHC). The higher incidence recorded in the present study further suggests that CIT exposure is widespread in Nigerian children. DON (46%) and OTA (34%) occurrences in urine in the present study were lower than those previously reported in adults from France (99%, Turner et al. 2010), The United Kingdom (95–100%, Hepworth et al. 2012; Papageorgiou et al. 2018) and Portugal (93%, Silva et al. 2019), but more similar to that observed in Egypt and Iran (Piekkola et al. 2012; Turner et al. 2012b). Infant data on DON in urine are mostly lacking. Similarly, urinary FB1 incidence of 26%, which we recorded, was lower than the 96% incidence reported in a study comprising 147 Tanzanian children’s urine (Shirima et al. 2013) but was higher than the report involving Nepalese children (incidence: 2% (1/50), Mitchell et al. 2016).
To date, AFM1 has been the most reported urinary AF metabolite and biomarker in children (Polychronaki et al. 2008; Ediage et al. 2013; Ezekiel et al. 2014; Gerding et al. 2015; Schwartzbord et al. 2016; Ayelign et al. 2017; Chen et al. 2017; Sanchez and Diaz 2019). However, in the present study, we found AFQ1 as the major aflatoxin metabolite (both in frequency and concentration), occurring in 44/65 (68%) of the children’s urine samples, whereas AFM1 was detected in 6/65 (9%) of urine samples, four of which had both AFM1 and AFQ1. Previous reports on AFQ1 occurrence in human urine were on adolescents and adults 15–64 years of age in China and The Gambia (Groopman et al. 1992a, 1992b; Wang et al. 2001; Mykkanen et al. 2005). Consequently, to the best of our knowledge, AFQ1 is reported for the first time in the urine of children in any setting. Urinary AFM1 and AF-N7-guanine are established biomarkers for aflatoxin exposure, however these were established in adults (Groopman et al. 1992a). Given the role of CYP3Aa4 in AFQ1 formation (Wild and Turner 2002) and the known significant ontogeny in xenobiotic metabolizing enzymes (Blake et al. 2005), involved in AFB1 biotransformation, it would be useful to better understand the relationship between dose and the various metabolites of aflatoxins in urine of the very young. Nevertheless, the recorded higher incidence of AFQ1 in the urine samples compared with AFM1 informs us that aflatoxin exposure in young children may be underreported.
Comparing the two cohorts of infants, higher mean levels of seven out of the 10 detected mycotoxins were found in the NEB urine compared to the EB urine, with statistically significant levels recorded for four of these mycotoxins including AFQ1, DON, DHC and FB1. Notably, increases in the mean mycotoxin levels in urine moving from EB regime to the NEB regime reached 721 and 334% for AFM1 and FB1, respectively. Additionally, the NEB cohort generally recorded higher diversities of mycotoxin mixtures in urine compared to the EB cohort. Here, we show that transition from EB to NEB results in a large variation of multiple mycotoxin exposures in infants. This primarily owes to the higher intake of complementary foods, which are usually made from mycotoxin-prone cereals, nuts and oil seeds (Kimanya et al. 2009; 2010, 2014; Juan et al. 2014; Okeke et al. 2015; Kamala et al. 2016; Chuisseu Njamen et al. 2018; Ojuri et al. 2018, 2019). These observations are strongly suggestive that exclusive breastfeeding provides a relatively, but not completely protected period from mycotoxin exposure.
The spectra and combinations of mycotoxins detected in the three specimens examined in this study further suggest frequent exposure in Nigerian children, especially in the NEB category. Obviously, urine contained higher number of mycotoxins and their mixtures compared to complementary foods and breast milk. Further, some mycotoxins occurred only in the biological fluids reiterating the importance of preferring biomonitoring approaches over food contamination data for exposure assessment in humans (Turner et al. 2012; Warth et al. 2013; Vidal et al. 2018; Ali and Degen 2019). A typical example is DHC and OTA that were detected only in the biological fluids (breast milk for the EB group and urine from both cohorts). Although CIT/DHC was not detected in the examined complementary foods in the present study it was found in 74% of the urines of EB children. CIT levels in grains (especially maize) and complementary food in Nigeria have recently been shown to be high reaching 16,000 µg/kg (Okeke et al. 2015, 2018; Ojuri et al. 2018, 2019; Ezekiel et al. 2020). Fermented cocoa beans in Nigeria, which could also be processed and used in the preparation of foods for infants and young children, have also recently been reported to contain CIT (Akinfala et al. 2020).
An additional important finding was the detection of BEA and EnnB in complementary food and breast milk but not in urine because the urinary analytical method did not include both mycotoxins. EnnB was recently detected in adult human urine from Italy (Rodriguez-Carrasco et al. 2018); however, there is limited data on urinary BEA and EnnB in children. Considering the widespread detection of BEA and EnnB in foods in Nigeria, albeit at low concentrations (Adetunji et al. 2014; Abdus-Salaam et al. 2015; Ezekiel et al. 2016; Oyedele et al. 2017; Ojuri et al. 2018), and the documented in vitro data suggesting their roles in decreasing immune response to infections (Ficheux et al. 2013) as well as their interactions to cause additive myelotoxicity (Ficheux et al. 2012), we recommend the inclusion of both mycotoxins in future urinary biomonitoring studies, especially for children.
The overall health implication for children in the study region and generally where mycotoxin exposure may be high are enormous suggesting the need for urgent interventions targeted at subsistence farmers and household food processing. In addition to the documented effects of the widely studied aflatoxins on child growth, in vitro studies using human intestinal cell lines revealed that DON selectively modulated intestinal transporters (Maresca et al. 2002) while FB1 decreased intestinal cell viability and proliferation (Minervini et al. 2014). Consequently, these mycotoxins may play interactive roles to affect nutrient absorption (Liew and Mohd-Redzwan 2018) which may cumulatively contribute to growth faltering. Thus, future studies may consider exploring the type and extent of interactive effects the various recorded mixtures may have on this highly vulnerable population and their growth into adulthood.
Conclusively, this study provides a comprehensive account on and novel insights into multiple mycotoxin exposures in EB and NEB infants in a high-risk country, Nigeria, based on examination of three specimens (breast milk, complementary food and urine). First, we show the presence of several uncommon mycotoxins in biological fluids (DHC and STER in breast milk; and AFQ1, being the dominant urinary aflatoxin metabolite). The higher incidence of AFQ1 in infant urine compared with AFM1 suggests that aflatoxin exposure in young children may be underreported. Second, the presence of up to six distinct mycotoxin classes are reported for the first time in breast milk and infant urine. Third, urines of NEB infants contained higher levels of mycotoxins and several mixtures of these toxins more frequently than in urines of EB infants. Thus, exposure to mycotoxins for infants less than 18 months of age is common in the region, with a significant transition in higher frequency and levels of exposure with the introduction of complementary foods. Consequently, it is recommended that mothers adhere strictly to measures with potential to reduce mycotoxin contamination in household diets; including maintaining compliance to WHO recommendations for exclusive breastfeeding for 6 months. Once foods are being introduced, adequate drying and proper storage of grains in airtight containers at household levels, sorting out of moldy, damaged and discolored grains prior to meal preparation, and diet diversification to include less mycotoxin prone food crops such as tubers are advised. In addition, prioritizing partnerships focused on driving interventions to mitigate exposures in this vulnerable population and investigating mycotoxin exposure and toxicity holistically in the context of the exposome paradigm is urgent.
Data Availability
Data in this study are available on request.
Conflict of interest
Authors do not have any conflict to declare.
Acknowledgements
Authors are thankful to Chidinma Akwarandu and Linda Ogu for their assistance during sample collection, and Xiaomin Han, during urine sample preparation. Families who donated samples are deeply appreciated. The authors want to gratefully acknowledge the Mass Spectrometry Centre (MSC) of the Faculty of Chemistry at the University of Vienna and Sciex for providing mass spectrometric instrumentation to BW. This work was performed with the financial support of the University of Vienna, the City of Vienna Jubilee Funds (BOKU Research Funding, project MycoMarker) and the Austrian Science Fund (FWF): P 33188-B.