Ochratoxin A (OTA) is a mycotoxin produced by several fungi of the genera Aspergillus and Penicillium, mainly by P. verrucosum in temperate climates and A. ochraceus in warm regions. In animal feed materials, the toxin is found most commonly in cereals [1]. The feedstuffs that produce the highest risks of contamination are oats and wheat and their by-products, such as bran [2,3]. Indeed, OTA is mainly concentrated in the seed coat of cereals, which is often used for animal feeding [4]. Recently, 42% of horse feed grains were contaminated by OTA at levels ranging from 0.2 to 4 μg/kg [5]. Analysis of biological fluids could provide a useful tool for assessing individual exposure to mycotoxins. Exposure to OTA may be estimated from the analysis of plasma samples because of its binding to serum albumin delays renal excretion, with consequent large interspecies differences in half-lives [1,6,7]. The horse has high economical value owing to its suitability for sport and as a domestic animal. Even though its feed composition is mainly based on cereals, it has not been considered much in studies on mycotoxin exposure and toxicity, with the exception of fumonisins. To date, no data are available on the bioavailability of OTA in the horse. Ochratoxin A is a potent renal toxin in all animal species, with pigs, dogs and poultry being particularly sensitive. Ruminants are less sensitive, due to degradation of OTA to the less toxic compound (ochratoxin α) by the rumen microflora [4]. In fact, ochratoxicosis has rarely been reported in cattle and in small ruminant species, such as sheep and goats, because microbial activity in the gastrointestinal tract can effectively and dramatically reduce OTA absorption thus protecting the animals. Herbivores, such as horses and rabbits that rely on caecal rather than ruminal fermentation, may absorb intact OTA in the small intestine and are likely to be more sensitive than ruminants [1].

In addition, OTA has been classified by IARC as a possible human carcinogen (class 2B). Ochratoxin A induces immunotoxicity, teratogenicity (through placental transfer), and reproductive toxicity, especially in more sensitive monogastric species, such as swine [8,9,10]. Several studies have demonstrated transplacental transfer of OTA in swine, although conflicting reports have been published: in particular, no residues in piglets of sows fed diets containing OTA at 7–16 µg/kg bw per day throughout gestation [11] or no placental transfer after ingestion of 0.38 mg/kg bw by a pregnant sow on days 21–28 [12] have been described. In contrast, Barnikol and Thalmann [13] found OTA transmission to piglets in utero when the sow was fed naturally contaminated feed and the blood concentrations in the newborn piglets were 0.075–0.12 ng/mL whereas, in the sow, it was 0.20 ng/mL. In humans, the OTA concentration in foetal serum was reported to be twice the maternal concentration, suggesting an active placental transfer of OTA [9,14,15]. To date, no data are available on OTA placental transfer in horses.

The present study focuses on the assessment of OTA occurrence and placental transfer using Enzyme-Linked Immunosorbent Assay (ELISA) on serum samples collected from 36 horses of different ages, breeds and production systems. These data were confirmed in some serum samples collected from mares and their foals by HPLC analysis.

ELISA showed good performance and provided accurate results because the coefficients of variation obtained with standard OTA solution and serum samples were very low (2.7% and 8.3%, respectively). Despite the high cross-reactivity of the antibody toward ochratoxin C, the results with ELISA were unaffected because OTA is the major compound found in blood samples. As observed in Table 1, OTA was found in 83% of serum samples collected from horses (n = 36) with a mean level of 169.2 pg/mL. With respect to diet composition, the majority of horses (from 64% to 80%) fed with commercial feed, hay and oats had less OTA in serum (mean value = 150 pg/mL) than horses fed with bran (mean value = 339 pg/mL). Although different kinetic profiles of OTA in rats (and consequent bioavailability) related to gender, age, weight and fasting condition have been described [16,17,18], due to an uneven number of horses for grouping, statistical analysis was not performed on these bases. The occurrence of OTA in serum samples collected from stallions could represent a risk for equine reproductive efficiency, as it may accumulate in seminal plasma and may have a toxic effect on some functional sperm parameters, as observed in vivo in boar [8].

Serum levels of OTA found in horses were similar to those found by other authors in swine (mean levels of 250–640 pg/mL) [19,20] but higher than those observed in chickens (mean levels of 13 pg/mL) [21]. Differences in OTA serum levels between animal species may well be related to the different half-lives of the toxin, which are strictly dependent on the affinity of its binding to serum proteins (retarding elimination and renal filtration) and to entero-hepatic circulation and biliary excretion, which might be responsible for toxin accumulation and prolonged elimination from the body [1,16,22].

Concerning placental transfer, as observed in Table 2, 83% of pregnant mares contained OTA at concentrations ranging from 69.7 to 348 pg/mL in serum samples, whereas only 50% of their foals were exposed to OTA at levels ranging from 69.5 to 252.6 pg/mL. The ratio between mare’s and foal’s OTA serum levels was variable and no correlation (r = −0.07) was found using Pearson’s correlation analysis.

Placental transfer of OTA was confirmed by HPLC analysis of four additional serum samples collected from mares (n = 2) and from the umbilical cords of their foals. The OTA occurrence was found in serum samples collected from the younger mare (7 years old) and her foal at levels of 33 and 19 pg/mL, respectively (Figure 1). The serum sample collected from the older mare (19 years old) contained OTA levels below the detection limit (15 pg/mL). The parallel presence/absence of OTA in serum samples collected from the mares and their foals might confirm the results obtained by ELISA.

The absence of a correlation between OTA levels in umbilical and mare serum samples was in agreement with the results reported in pigs [1] after in vivo exposure, and could be explained by several factors, such as exposure time (gestation period and related placental vascularization) and/or placental structure. Concerning the gestation period, during the early stage of pregnancy, OTA or its metabolites could pass through the placenta into the foetal circulation and accumulate in the foetal tissues, exerting developmental toxicity, as reported by some authors [10,23]. A reduced blood supply to the placenta or foetal tissues during the latter stages of foetal development might prevent or interfere with the access of toxin to umbilical blood [23]. Concerning placental structure, swine and equine placenta are characterized by a chorionic epithelium which could prevent or interfere with the access of toxin to the umbilical blood in vivo and consequent OTA foetal uptake, strictly related also to the developmental stage of the placenta [1,22]. Concerning other animal species, the few data published on OTA placental transfer in ruminants have shown very low penetration by OTA through the placenta; in fact, after intravenous administration of a high dose of OTA (1 mg/kg body weight) to pregnant ewes, OTA was not detected in the amniotic fluid and foetal tissues and levels were 400- to 1000-fold lower than in the maternal blood [24].This animal species has a syndesmochorial placenta, i.e., the epithelium of the chorion is in contact with the connective tissue of the lamina propria of the uterine mucosa, lacking the epithelium on the maternal side. By contrast in humans, the haemochorial placenta allows direct contact between the maternal blood and the foetal chorion and foetal blood is separated from the maternal blood circulation by polarised cells, which possess carrier-mediated transport pathways, similarly renal proximal tubules and intestinal epithelial cells [9]. For this reason, Miraglia et al. [14] showed higher OTA levels in placenta and in funiculum than those found in the corresponding maternal serum and a slight correlation was reported between umbilical and maternal serum OTA levels [25]. Postupolski et al. [15] found a statistically significant difference (p < 0.05) between the mean OTA concentration in maternal blood serum (1.14 ng/mL) and umbilical cord serum (1.96 ng/mL). The mean ratio calculated for OTA maternal blood/foetal blood was 1.96 ± 0.94. As suggested by Postupolski [15], the higher OTA content in foetal blood could be explained as being due to its active transport across the placenta, perhaps as a consequence of its chemical characteristics (such as the similarity between the chemical structures of OTA and phenylalanine, low molar mass, lipophilic character, serum binding property). Consequently, particularly during the first five months of life, the foetus was more sensitive to xenobiotics (such as OTA) mainly due to insufficiently adequate function of the kidneys and hepatic metabolism systems [15].

This study for the first time reports data on the natural occurrence of OTA in serum samples and placental transfer in horses. Levels of OTA found in serum samples could represent an additional risk for this monogastric animal potentially exposed to mycotoxins through cereal feeding.