Paired metagenomic and chemical evaluation of aflatoxin-contaminated dog kibble

Introduction

Identification of chemical toxins from complex or highly processed foods can present 'needle in the haystack' challenges for chemists. Metagenomic data can be used to guide chemical toxicity evaluations by providing DNA-based description of the wholistic composition (eukaryotic, bacterial, protozoal, viral, and antimicrobial resistance) of foods suspected to harbor toxins, allergens, or pathogens. This type of information can focus chemistry-based diagnostics, improve hazard characterization and risk assessment, and address data gaps. Additionally, there is increasing recognition that simultaneously co-occurring mycotoxins, either from single or multiple species, can impact dietary toxicity exposure. Metagenomic data provides a way to address data gaps related to co-occurrence of multiple fungal species.

Methods

Paired metagenomic and chemical data were used to evaluate aflatoxin-contaminated kibble with known levels of specific mycotoxins. Kibble was ground to a fine powder for both chemical and molecular analyses. Chemical analyses were performed with Liquid Chromatography Mass Spectrometry (LCMS) and according to the AOAC Official method 2005.08: Aflatoxins in Corn, Raw Peanuts, and Peanut Butter using Liquid Chromatography with Post-Column Photochemical Derivatization. Metagenomes were created from DNA extracted from ground kibble and sequenced on an Illumina NextSeq 2000 with an average sequence depth of 180 million reads per replicate.

Results and discussion

Metagenomic data demonstrated that the abundance of DNA from putative aflatoxigenic Aspergillus spp. correlated with the levels of aflatoxin quantified by LCMS. Metagenomic data also identified an expansive range of co-occurring fungal taxa which may produce additional mycotoxins. DNA data paired with chemical data provides a novel modality to address current data gaps surrounding dietary mycotoxin exposure, toxigenic fungal taxonomy, and mycotoxins of emerging concern.

Retrospective analysis of dog study data from food and color additive petitions

As part of the US FDA CFSAN's efforts to explore alternatives to animal testing, we retrospectively analyzed a sample of food additive (FAP) and color additive petitions (CAP) submitted to the FDA for the utility of dog study data in safety assessment. FAPs and CAPs containing dog studies (161 petitions) were classified as decisive (38%), supportive (27%), supplemental (29%) or undermined (6%) based on the impact the dog study data had on the final safety decision. Petitions classified as decisive were further categorized based on if the dog study data were used to a) address a safety concern (35/61); b) calculate an acceptable daily intake (ADI) (11/61); c) withdraw a petition (4/61); d) the effect was unique to the dog (2/61); or e) unclear (9/61). Of 11 petitions where the dog study was used to set an ADI, 7 contained studies where the points of departure (POD) from the dog studies were within an 8-fold range of the rodent with differences in study design likely contributing to the difference in PODs. Future research should include the development and use of qualified alternative studies to replace the use of animal testing for food and color additive safety assessment while ensuring human safety.

Reassessment of the cadmium toxicological reference value for use in human health assessments of foods

The U.S. Food and Drug Administration (FDA) developed an oral toxicological reference value (TRV) for characterizing potential health concerns from dietary exposure to cadmium (Cd). The development of the TRV leveraged the FDA's previously published research including (1) a systematic review for adverse health effects associated with oral Cd exposure and (2) a human physiological based pharmacokinetic (PBPK) model adapted from Kjellstrom and Nordberg (1978) for use in reverse dosimetry applied to the U.S. population. Adverse effects of Cd on the bone and kidney are associated with similar points of departure (PODs) of approximately 0.50 μg Cd/g creatinine for females aged 50-60 based on available epidemiologic data. We also used the upper bound estimate of the renal cortical concentration (50 μg/g Cd) occurring in the U.S. population at 50 years of age as a POD. Based on the output from our reverse dosimetry PBPK Model, a range of 0.21-0.36 μg/kg bw/day was developed for the TRV. The animal data used for the animal TRV derivation (0.63-1.8 μg/kg bw/day) confirms biological plausibility for both the bone and kidney endpoints.

A systematic review of adverse health effects associated with oral cadmium exposure

Scientific data characterizing the adverse health effects associated with dietary cadmium (Cd) exposure were identified in order to make informed decisions about the most appropriate toxicological reference value (TRV) for use in assessing dietary Cd exposure. Several TRVs are available for Cd and regulatory organizations have used epidemiologic studies to derive these reference values; however, risk of bias (RoB) evaluations were not included in the assessments. We performed a systematic review by conducting a thorough literature search (through January 4, 2020). There were 1714 references identified by the search strings and 328 studies identified in regulatory assessments. After applying the specific inclusion and exclusion criteria, 208 studies (Human: 105, Animal: 103) were considered eligible for further review and data extraction. For the epidemiologic and animal studies, the critical effects identified for oral Cd exposure from the eligible studies were a decrease in bone mineral density (BMD) and renal tubular degeneration. A RoB analysis was completed for 49 studies (30 epidemiological and 19 animal) investigating these endpoints. The studies identified through the SR that were considered high quality and low RoB (2 human and 5 animal) can be used to characterize dose-response relationships and inform the derivation of a Cd TRV.

Cadmium physiologically based pharmacokinetic (PBPK) models for forward and reverse dosimetry: Review, evaluation, and adaptation to the U.S. population

The goal of this study was to assess a cadmium (Cd) physiologically based pharmacokinetic (PBPK) model to evaluate Cd toxicological reference values (e.g. reference dose, tolerable intake, minimum risk level) adapted to the U.S. population. We reviewed and evaluated previously published Cd PBPK models and developed further adaptations to the 1978 Kjellström and Nordberg (KN) model. Specifically, we propose adaptations with updated U.S.-specific bodyweight, kidney weight and creatinine excretion models by using NHANES data as well as a stochastic PBPK model that provides credible intervals of uncertainty around mean populational estimates. We provide our model review and adaptations as well as present estimates from the newly adapted models using observed U.S. urinary Cd values as a function of gender and age and given dietary exposure as evaluated from NHANES/WWEIA and U.S. Total Diet Study data. Results show all newly adapted models provide acceptable mean estimates of urinary Cd in the U.S. The stochastic model provides credible intervals to further inform regulatory decision making. Validation of the estimated K-Cd concentration values was not possible as data for a representative population was not available. We developed a web-based tool implementing these models and other potential adaptations to facilitate PBPK model estimate comparisons.

A review of the evidence to support interim reference level for dietary lead exposure in adults

FDA developed the interim reference level (IRL) for lead of 3 μg/day in children and 12.5 μg/day in women of childbearing age (WOCBA) to better protect the fetus from lead toxicity. These IRLs correspond to a blood lead level (BLL) of 0.5 μg/dL in both populations. The current investigation was performed to determine if the IRL for WOCBA should apply to the general population of adults. A literature review of epidemiological studies was conducted to determine whether a BLL of 0.5 μg/dL is associated with adverse effects in adults. Some studies reported adverse effects over a wide range of BLLs that included 0.5 μg/dL adding uncertainty to conclusions about effects at 0.5 μg/dL; however, no studies clearly identified this BLL as an adverse effect level. Results also showed that the previously developed PTTDI for adults of 75 μg/day lead may not be health protective, supporting use of a lower reference value for lead toxicity in this population group. Use of the 12.5 μg/day IRL as a benchmark for dietary lead intake is one way FDA will ensure that dietary lead intake in adults is reduced.

Persistent behavior deficits, neuroinflammation, and oxidative stress in a rat model of acute organophosphate intoxication

Current medical countermeasures for organophosphate (OP)-induced status epilepticus (SE) are not effective in preventing long-term morbidity and there is an urgent need for improved therapies. Rat models of acute intoxication with the OP, diisopropylfluorophosphate (DFP), are increasingly being used to evaluate therapeutic candidates for efficacy in mitigating the long-term neurologic effects associated with OP-induced SE. Many of these therapeutic candidates target neuroinflammation and oxidative stress because of their implication in the pathogenesis of persistent neurologic deficits associated with OP-induced SE. Critical to these efforts is the rigorous characterization of the rat DFP model with respect to outcomes associated with acute OP intoxication in humans, which include long-term electroencephalographic, neurobehavioral, and neuropathologic effects, and their temporal relationship to neuroinflammation and oxidative stress. To address these needs, we examined a range of outcomes at later times post-exposure than have previously been reported for this model. Adult male Sprague-Dawley rats were given pyridostigmine bromide (0.1 mg/kg, im) 30 min prior to administration of DFP (4 mg/kg, sc), which was immediately followed by atropine sulfate (2 mg/kg, im) and pralidoxime (25 mg/kg, im). This exposure paradigm triggered robust electroencephalographic and behavioral seizures that rapidly progressed to SE lasting several hours in 90% of exposed animals. Animals that survived DFP-induced SE (~70%) exhibited spontaneous recurrent seizures and hyperreactive responses to tactile stimuli over the first 2 months post-exposure. Performance in the elevated plus maze, open field, and Pavlovian fear conditioning tests indicated that acute DFP intoxication reduced anxiety-like behavior and impaired learning and memory at 1 and 2 months post-exposure in the absence of effects on general locomotor behavior. Immunohistochemical analyses revealed significantly increased expression of biomarkers of reactive astrogliosis, microglial activation and oxidative stress in multiple brain regions at 1 and 2 months post-DFP, although there was significant spatiotemporal heterogeneity across these endpoints. Collectively, these data largely support the relevance of the rat model of acute DFP intoxication as a model for acute OP intoxication in the human, and support the hypothesis that neuroinflammation and/or oxidative stress represent potential therapeutic targets for mitigating the long-term neurologic sequelae of acute OP intoxication.

Persistent neuroinflammation and cognitive impairment in a rat model of acute diisopropylfluorophosphate intoxication

Background

Acute intoxication with organophosphorus (OP) cholinesterase inhibitors can trigger convulsions that progress to life-threatening status epilepticus. Survivors face long-term morbidity including mild-to-severe decline in memory. It is posited that neuroinflammation plays a key role in the pathogenesis of OP-induced neuropsychiatric deficits. Rigorous testing of this hypothesis requires preclinical models that recapitulate relevant phenotypic outcomes. Here, we describe a rat model of acute intoxication with the OP diisopropylfluorophosphate (DFP) that exhibits persistent neuroinflammation and cognitive impairment.

Methods

Neuroinflammation, neurodegeneration, and cognitive function were compared in adult male Sprague Dawley rats injected with an acutely toxic dose of DFP vs. vehicle controls at multiple time points up to 36 days post-exposure. Neuroinflammation was quantified using immunohistochemical biomarkers of microglia (ionized calcium-binding adapter molecule 1, IBA1) and activated astrocytes (glial fibrillary acidic protein, GFAP) and positron emission tomography (PET) imaging of [¹¹C]-(R)-PK11195, a ligand for the 18-kDa mitochondrial membrane translocator protein (TSPO). FluoroJade-B staining was used to assess neurodegeneration; Pavlovian conditioning, to assess cognitive function.

Results

Animals exhibited moderate-to-severe seizures within minutes of DFP injection that continued for up to 6 h post-injection. As indicated by IBA1 and GFAP immunoreactivity and by PET imaging of TSPO, acute DFP intoxication triggered neuroinflammation in the hippocampus and cortex during the first 3 days that peaked at 7 days and persisted to 21 days post-exposure in most animals. Neurodegeneration was detected in multiple brain regions from 1 to 14 days post-exposure. All DFP-intoxicated animals exhibited significant deficits in contextual fear conditioning at 9 and 20 days post-exposure compared to vehicle controls. Whole-brain TSPO labeling positively correlated with seizure severity score, but did not correlate with performance in the contextual fear-conditioning task.

Conclusions

We describe a preclinical model in which acute DFP intoxication causes seizures, persistent neuroinflammation, neurodegeneration, and memory impairment. The extent of the neuroinflammatory response is influenced by seizure severity. However, the observation that a subset of animals with moderate seizures and minimal TSPO labeling exhibited cognitive deficits comparable to those of animals with severe seizures and significant TSPO labeling suggests that DFP may impair learning and memory circuitry via mechanisms independent of seizures or neuroinflammation.

Murine Anorectic Response to Deoxynivalenol (Vomitoxin) Is Sex-Dependent

Deoxynivalenol (DON, vomitoxin), a common trichothecene mycotoxin found in cereal foods, dysregulates immune function and maintenance of energy balance. The purpose of this study was to determine if sex differences are similarly evident in DON's anorectic responses in mice. A bioassay for feed refusal, previously developed by our lab, was used to compare acute i.p. exposures of 1 and 5 mg/kg bw DON in C57BL6 mice. Greater anorectic responses were seen in male than female mice. Male mice had higher organ and plasma concentrations of DON upon acute exposure than their female counterparts. A significant increase in IL-6 plasma levels was also observed in males while cholecystokinin response was higher in females. When effects of sex on food intake and body weight changes were compared after subchronic dietary exposure to 1, 2.5, and 10 ppm DON, males were found again to be more sensitive. Demonstration of male predilection to DON-induced changes in food intake and weight gain might an important consideration in future risk assessment of DON and other trichothecenes.

Factors influencing adverse skin responses in rats receiving repeated subcutaneous injections and potential impact on neurobehavior

Repeated subcutaneous (s.c.) injection is a common route of administration in chronic studies of neuroactive compounds. However, in a pilot study we noted a significant incidence of skin abnormalities in adult male Long-Evans rats receiving daily s.c. injections of peanut oil (1.0 ml/kg) in the subscapular region for 21 d. Histopathological analyses of the lesions were consistent with a foreign body reaction. Subsequent studies were conducted to determine factors that influenced the incidence or severity of skin abnormalities, and whether these adverse skin reactions influenced a specific neurobehavioral outcome. Rats injected daily for 21 d with food grade peanut oil had an earlier onset and greater incidence of skin abnormalities relative to rats receiving an equal volume (1.0 ml/kg/d) of reagent grade peanut oil or triglyceride of coconut oil. Skin abnormalities in animals injected daily with peanut oil were increased in animals housed on corncob versus paper bedding. Comparison of animals obtained from different barrier facilities exposed to the same injection paradigm (reagent grade peanut oil, 1.0 ml/kg/d s.c.) revealed significant differences in the severity of skin abnormalities. However, animals from different barrier facilities did not perform differently in a Pavlovian fear conditioning task. Collectively, these data suggest that environmental factors influence the incidence and severity of skin abnormalities following repeated s.c. injections, but that these adverse skin responses do not significantly influence performance in at least one test of learning and memory.

Behavioral assessment of NIH Swiss mice acutely intoxicated with tetramethylenedisulfotetramine

Tetramethylenedisulfotetramine (TETS) is a potent convulsant poison that is thought to trigger seizures by inhibiting the function of the type A gamma-aminobutyric acid receptor (GABAAR). Acute intoxication with TETS can cause vomiting, convulsions, status epilepticus (SE) and even death. Clinical case reports indicate that individuals who survive poisoning may exhibit long-term neuropsychological issues and cognitive deficits. Therefore, the objective of this research was to determine whether a recently described mouse model of acute TETS intoxication exhibits persistent behavioral deficits. Young adult male NIH Swiss mice received a seizure-inducing dose of TETS (0.15mg/kg, ip) and then were rescued from lethality by administration of diazepam (5mg/kg, ip) approximately 20min post-TETS-exposure. TETS-intoxicated mice typically exhibited 2 clonic seizures prior to administration of diazepam with no subsequent seizures post-diazepam injection as assessed using behavioral criteria. Seizures lasted an average of 72s. Locomotor activity, anxiety-like and depression-relevant behaviors and cognition were assessed at 1week, 1month and 2months post-TETS exposure using open field, elevated-plus maze, light↔dark transitions, tail suspension, forced swim and novel object recognition tasks. Interestingly, preliminary validation tests indicated that NIH Swiss mice do not respond to the shock in fear conditioning tasks. Subsequent evaluation of hot plate and tail flick nociception tasks revealed that this strain exhibits significantly decreased pain sensitivity relative to age- and sex-matched C57BL/6J mice, which displayed normal contextual fear conditioning. NIH Swiss mice acutely intoxicated with TETS exhibited no significant anxiety-related, depression-relevant, learning or memory deficits relative to vehicle controls at any of the time points assessed with the exception of significantly increased locomotor activity at 2months post-TETS intoxication. The general absence of long-term behavioral deficits in TETS-intoxicated mice on these six assays suggests that the neurobehavioral consequences of TETS exposure described in human survivors of acute TETS intoxication are likely due to sustained seizure activity, rather than a direct effect of the chemical itself. Future research efforts are directed toward developing an animal model that better recapitulates the SE and seizure duration reported in humans acutely intoxicated with TETS.

Deoxynivalenol-induced weight loss in the diet-induced obese mouse is reversible and PKR-independent

The trichothecene deoxynivalenol (DON), a potent ribotoxic mycotoxin produced by the cereal blight fungus Fusarium graminearum, commonly contaminates grain-based foods. Oral exposure to DON causes decreased food intake, reduced weight gain and body weight loss in experimental animals - effects that have been linked to dysregulation of hormones responsible for mediating satiety at the central nervous system level. When diet-induced obese (DIO) mice are fed DON, they consume less food, eventually achieving body weights of control diet-fed mice. Here, we extended these findings by characterizing: (1) reversibility of DON-induced body weight loss and anorexia in DIO mice and (2) the role of double-stranded RNA-activated protein kinase (PKR) which has been previously linked to initiation of the ribotoxic stress response. The results demonstrated that DON-induced weight loss was reversible in DIO mice and this effect corresponded to initiation of a robust hyperphagic response. When DIO mice deficient in PKR were exposed to DON, they exhibited weight suppression similar to DIO wild-type fed the toxin, suggesting the toxin's weight effects were not dependent on PKR. Taken together, DON's effects on food consumption and body weight are not permanent and, furthermore, PKR is not an essential signaling molecule for DON's anorectic and weight effects.

Comparison of emetic potencies of the 8-ketotrichothecenes deoxynivalenol, 15-acetyldeoxynivalenol, 3-acetyldeoxynivalenol, fusarenon X, and nivalenol

Although the acute toxic effects of trichothecene mycotoxin deoxynivalenol (DON or vomitoxin), a known cause of human food poisoning, have been well characterized in several animal species, much less is known about closely related 8-ketotrichothecenes that similarly occur in cereal grains colonized by toxigenic fusaria. To address this, we compared potencies of DON, 15-acetyldeoxynivalenol (15-ADON), 3-acetyldeoxynivalenol (3-ADON), fusarenon X (FX), and nivalenol (NIV) in the mink emesis model following intraperitoneal (ip) and oral administration. All five congeners dose-dependently induced emesis by both administration methods. With increasing doses, there were marked decreases in latency to emesis with corresponding increases in emesis duration and number of emetic events. The effective doses resulting in emetic events in 50% of the animals for ip exposure to DON, 15-ADON, 3-ADON, FX, and NIV were 80, 170, 180, 70, and 60 µg/kg bw, respectively, and for oral exposure, they were 30, 40, 290, 30, and 250 µg/kg bw, respectively. The emetic potency of DON determined here was comparable to that reported in analogous studies conducted in pigs and dogs, suggesting that the mink is a suitable small animal model for investigating acute trichothecene toxicity. The use of a mouse pica model, based on the consumption of kaolin, was also evaluated as a possible surrogate for studying emesis but was found unsuitable. From a public health perspective, comparative emetic potency data derived from small animal models such as the mink should be useful for establishing toxic equivalency factors for DON and other trichothecenes.

Anorexia induction by the trichothecene deoxynivalenol (vomitoxin) is mediated by the release of the gut satiety hormone peptide YY

Consumption of deoxynivalenol (DON), a trichothecene mycotoxin known to commonly contaminate grain-based foods, suppresses growth of experimental animals, thus raising concerns over its potential to adversely affect young children. Although this growth impairment is believed to result from anorexia, the initiating mechanisms for appetite suppression remain unknown. Here, we tested the hypothesis that DON induces the release of satiety hormones and that this response corresponds to the toxin's anorectic action. Acute ip exposure to DON had no effect on plasma glucagon-like peptide-1, leptin, amylin, pancreatic polypeptide, gastric inhibitory peptide, or ghrelin; however, the toxin was found to robustly elevate peptide YY (PYY) and cholecystokinin (CCK). Specifically, ip exposure to DON at 1 and 5mg/kg bw induced PYY by up to 2.5-fold and CCK by up to 4.1-fold. These responses peaked within 15-120 min and lasted up to 120 min (CCK) and 240 min (PPY), corresponding with depressed rates of food intake. Direct administration of exogenous PYY or CCK similarly caused reduced food intake. Food intake experiments using the NPY2 receptor antagonist BIIE0246 and the CCK1A receptor antagonist devazepide, individually, suggested that PYY mediated DON-induced anorexia but CCK did not. Orolingual exposure to DON induced plasma PYY and CCK elevation and anorexia comparable with that observed for ip exposure. Taken together, these findings suggest that PYY might be one critical mediator of DON-induced anorexia and, ultimately, growth suppression.

Characterization of deoxynivalenol-induced anorexia using mouse bioassay

A short-term mouse model was devised to investigate induction of food refusal by the common foodborne trichothecene deoxynivalenol (DON). DON dose-dependently induced anorexia within 2 h of exposure when administered either by intraperitoneal (ip.) injection or by oral gavage. The no observed adverse effect and lowest observed adverse effect levels in this assay were 0.5 and 1 mg/kg bw for ip. exposure and 1 and 2.5 mg/kg bw for oral exposure, respectively. DON's effects on food intake were transient, lasting up to 3h at 1 mg/kg bw and up to 6 h at 5 mg/kg bw. Interestingly, a dose-dependent orexigenic response was observed in the 14 h following the initial 2h food intake measurement. Toxin-treated mice exhibited partial resistance to feed refusal when exposed to DON subsequently after 2 d, but not after 7 d suggesting that this modest tolerance was reversible. The short-term mouse bioassay described here was useful in characterizing DON-induced anorexia and should be applicable to elucidating mechanisms underlying this adverse nutritional effect.

Body composition and hormonal effects following exposure to mycotoxin deoxynivalenol in the high-fat diet-induced obese mouse

SCOPE

To characterize the effects of ingesting the common foodborne mycotoxin deoxynivalenol (DON) on body weight and composition in the high-fat (HF) diet-induced obese mice, a model of human obesity.

METHODS AND RESULTS

Female B6C3F1 mice were initially fed HF diets containing 45% kcal (HF45) or 60% kcal (HF60) as fat for 94 days to induce obesity. Half of each group was either continued on unamended HF diets or fed HF diets containing 10 mg/kg DON (DON-HF45 or DON-HF60) for another 54 days. Additional control mice were fed a low-fat (LF) diet containing 10%  kcal as fat for the entire 148-day period. DON induced rapid decreases in body weights and fat mass, which stabilized to those of the LF control within 11 days. These effects corresponded closely to a robust transient decrease in food consumption. While lean body mass did not decline in DON-fed groups, further increases were suppressed. DON exposure reduced plasma insulin, leptin, insulin-like growth factor 1, and insulin-like growth factor acid labile subunit as well as increased hypothalamic mRNA level of the orexigenic agouti-related protein.

CONCLUSION

DON-mediated effects on body weight, fat mass, food intake, and hormonal levels in obese mice were consistent with a state of chronic energy restriction.

Effects of deoxynivalenol consumption on body weight and adiposity in the diet-induced obese mouse

The potential for the obese state to alter sensitivity to toxic chemicals is poorly understood. In this study, dose-response effects of the trichothecene deoxynivalenol (DON), a common food-borne mycotoxin, were determined on body weight of diet-induced obese mice. In study 1, the effects of feeding adult female B6C3F1 mice a high-fat diet (HFD; 60% kcal from fat) containing 0, 2, 5, or 10 ppm DON for 10 wk on body weight and adiposity were compared. Mice consuming 5 or 10 ppm DON exhibited a 15 and 24% decrease in weight gain and a 50 and 83% reduction in periuterine fat, respectively. In study 2, mice were fed HFD for 8 wk to induce obesity and the effects of consuming HFD + 0, 2, 5, or 10 ppm DON for 8 wk were then determined. Mice fed 5 or 10 ppm DON exhibited a 16 and 23% weight reduction and a 0 and 40% periuterine fat reduction, respectively. In a follow-up experiment, food consumption was measured prior to and after the transition from HFD to HFD + 10 ppm DON. Exposure to DON was found to lower HFD consumption within 1 d, with significant weight loss in DON-fed mice evident after 6 d. In both studies 1 and 2, consumption of 5 or 10 ppm DON diminished circulating levels of insulin-like growth factor acid-labile subunit. Taken together, DON consumption lowered weight gain and produced weight loss in diet-induced obese mice at higher thresholds than that observed previously in normal B6C3F1 mice.