**1. Introduction**

Mycotoxins are toxic secondary metabolites produced by many filamentous fungi. The most important fungi causing frequent and problematic contamination of human food and animal feed belong to the fungal genera of *Fusarium*, *Aspergillus*, and *Penicillium*. Maize is considered the most susceptible crop for mycotoxin contamination and rice the least susceptible one [1].

*Fusarium* fungi produce a variety of toxic secondary metabolites, which are not essential to fungal growth but can induce several adverse effects in livestock [2]. The most toxicologically important *Fusarium* toxins are fumonisins (FBs), zearalenone (ZEN), and trichothecenes, such as deoxynivalenol (DON), nivalenol (NIV), diacetoxyscirpenol (DAS), and T-2 toxin [3]. Both in vitro and in vivo studies have demonstrated that toxicokinetics, bioavailability, and the mechanisms of action of these substances vary depending on the species involved [4]. ZEN causes reproductive abnormalities in pigs and ruminants and DON is well known for being a potent feed intake inhibitor in pigs [5,6]. Next to these well-known *Fusarium* mycotoxins, there are also several unregulated, so-called emerging mycotoxins, which frequently occur in agricultural products. One of them is fusaric acid (FA), which is found in several types of cereal grain and mixed feeds. This mycotoxin needs to be further investigated in vitro and in vivo because its neurochemical effects and possible synergistic effects with other mycotoxins, especially DON and FBs, may pose a problem to humans and livestock [7].

Besides aflatoxins (AFs) and ochratoxins, which are not *Fusarium* mycotoxins, FBs, ZEN, and trichothecenes, especially DON, are considered highly important in food safety and public health due to their widespread occurrence and toxicity. In people, chronic exposure to mycotoxins, even at low levels, may lead to adverse effects in different organs, such as the liver, kidneys, and immune system [8,9].

Due to the frequent presence of several different mycotoxins in grain and animal feed, widespread reports of co-contamination are of grea<sup>t</sup> potential significance [10]. A global survey indicated that 72% of samples of feed and feed raw materials are positive for at least one mycotoxin and 38% are co-contaminated [11], whereas several studies in European countries, simultaneously analyzing 20 or more mycotoxins, have shown a remarkable 44–100% of such samples to be co-contaminated with more than one mycotoxin [12]. A recent study that included 524 worldwide finished pig feed samples detected more than 235 different metabolites, including regulated mycotoxins, emerging mycotoxins, and modified/masked mycotoxins. DON was detected in 88% of the samples, mostly from the Northern Hemisphere, with a median concentration of 0.206 mg/kg of feed. All DON-contaminated samples were co-contaminated by other mycotoxins, the second most common being ZEN with a median concentration of 0.018 mg/kg of feed, while FA was not among the 60 most prevalent fungal metabolites [13].

Concomitantly occurring mycotoxins can have antagonistic, additive, or synergistic effects [14], but very little is known about their potential interactive toxic effects [2]. Even though the results from the global survey indicated that the *Fusarium* mycotoxins DON, FBs, and ZEN contaminated 55%, 54%, and 36% of feed and feed ingredients, respectively, most samples complied with even the most rigorous European Union regulations or recommendations on the maximal tolerable concentrations of individual mycotoxins [11]. Currently, the European Commission's recommendation and its amendment on the presence of DON, ZEN, ochratoxin A, T-2 and HT-2, and FBs in products intended for animal feeding sugges<sup>t</sup> that compound feed for piglets and gilts does not exceed 0.9 mg of DON/kg, 0.1 mg of ZEN/kg, 0.05 mg of ochratoxin A/kg and 5 mg of fumonisins B1 + B2/kg [15,16]. It is therefore of grea<sup>t</sup> importance to determine the effects of co-contaminating *Fusarium* mycotoxins, especially at naturally occurring concentrations, as well as concentrations lower than the accepted tolerance concentrations for individual mycotoxins.

Pigs are especially interesting for further research because they are one of the most sensitive animal species for *Fusarium* mycotoxins, especially ZEN and trichothecenes, such as DON and T-2. They are usually fed a cereal-rich diet, which can expose them to higher levels of these mycotoxins. As they are physiologically very similar to humans, they can serve as a good translational animal model, especially due to their similarities in the intestinal tract [17]. The effects of these toxins partly depend on their absorption, distribution, metabolism, and excretion (ADME processes) by the animal species in question. As the ADME processes seem to be qualitatively quite similar between pigs and humans, pigs can be very useful in the risk assessment of mycotoxins and for establishing legal limits of mycotoxins [18].

Research on the effects of feeding pigs with *Fusarium* mycotoxin co-contaminated feed has been ongoing for over 30 years, providing insight in various aspects. These studies often emphasized the zootechnical, hematological, biochemical, toxicological, and immunological parameters [19–23], whereas others also investigated histological changes in various organs with or without the aid of immunohistochemistry [14,24–35] or even examined gene expression profiles [36–38].

Since *Fusarium* mycotoxins are such a common contaminant and clearly have effects on different animal species, it is also of interest whether they have detrimental effects on the offspring. Some studies have analyzed the transfer of single or multiple *Fusarium* mycotoxins from sows to their offspring, implying that these can cause indirect effects via a decreased feed intake and via direct effects of diaplacentar transfer of ingested mycotoxins to the developing fetuses [20,21,29–31,39,40].

The liver is an important site of *Fusarium* mycotoxin metabolism [18]. DON's effects on liver have been investigated by studies evaluating biochemical, functional, histopathological parameters [33,34,41–43] and even gene expression profiles [38].

The aim of this study was therefore to determine whether feed containing naturally occurring concentrations of DON, ZEN, and FA would elicit histopathological changes, a difference in the number of apoptotic cells, and the proliferation index in the liver of gilts and their suckling piglets.

#### **2. Materials and Methods**

#### *2.1. Research Design*

This study was conducted on samples retrieved from the experiment approved by the Veterinary Administration of the Republic of Slovenia and described in detail by Jakovac-Strajn et al. [22]. In summary, the experiment included 10 gilts that were fed an experimental diet containing maize naturally contaminated with *Fusarium* mycotoxins, 10 gilts that were fed a control diet, and the offspring of both groups. The gilts were daily fed 3.5 kg of the diet during gestation and 6 kg of the same diet from the day of farrowing until weaning. The gilts from the experimental group consumed significantly less than the control group, but the average bodyweight was not significantly different even at the end of the experiment. At the start of the experiment, the gilts were at 89 ± 2 days of gestation, and they remained in the experiment for a total of 54 ± 1 day. The farrowing in both groups started 24 to 27 days after the start of the experiment and the piglets were weaned at 21 days of age. No antimicrobials were given to either the gilts or their piglets during the experiment.

The experimental diet contained 5.08 mg DON, 0.09 mg ZEN, and 21.6 mg FA per kg of feed. The control diet contained 0.29 mg DON per kg of feed, whereas ZEN (<0.02 mg/kg) and FA (<0.77 mg/kg) were below their detection limits. The concentrations of aflatoxin B1 (<0.2 μg/kg), 15-ADON, NIV, fusarenon-X, DAS, T-2 toxin, HT-2 toxin (<0.05 μg/kg), ochratoxin A, and fumonisins B1, B2, and B3 (<10 μg/kg) were also measured in both diets, but they were all below their detection limits, these being indicated in parentheses.

In order to collect organs for further examination, a single 7-day-old suckling piglet was randomly selected from each of the 20 litters and killed by lethal injection of T-61 solution (embutramide/mebezonium iodide/tetracaine hydrochloride, Intervet, Unterschleißheim, Germany), whereas all the gilts were killed by captive bolt and exsanguination 5 to 8 days after weaning of the remaining piglets in the litters. Afterwards, liver samples were immediately collected, fixed in 10% phosphate buffered formalin and routinely embedded in paraffin blocks. Liver samples from two killed suckling piglets, one from each group, were inappropriate for further processing.

#### *2.2. Histopathology of the Liver of Gilts and Their Suckling Piglets*

Histopathological examination of 4 μm thick tissue sections of formalin-fixed paraffinembedded (FFPE) liver samples stained with hematoxylin and eosin (H&E) was conducted using light microscopy. Several different histopathological changes were assessed in the liver: irregularity of hepatic cords, fibrosis, sinusoidal leukocytosis with inflammatory infiltrates of hepatic lobules, portal tract inflammatory infiltrates, hepatocytes with vacuolar or granular cytoplasm, hepatocellular necrosis and apoptosis, markedly enlarged hepatocytes (hepatocellular megalocytosis), markedly enlarged hepatocellular nuclei (hepatocellular megakaryosis), biliary hyperplasia, dilatation and thickening of blood vessels, and thrombosis of blood or lymphatic vessels.

Each assessed histopathological change was graded for its intensity and extent. The intensity of the histopathological changes was assigned one of the following scores: 0—not present, 1—mild, 2—moderate, and 3—severe. The extent of the histopathological changes was assigned one of the following scores: 0—not present, 1—minimal (0 to <5% of the tissue section), 2—mild (5 to <15% of the tissue section), 3—moderate (15 to <40% of the tissue section) and 4—severe (40% or more of the tissue section). The assigned intensity and extent score were then multiplied to obtain the final score for each histopathological change in the tissue section of each liver sample from both the gilts and their piglets.

#### *2.3. Detection of Apoptotic Cells in the Liver of Gilts and Their Suckling Piglets*

For the detection of apoptotic cells in 4 μm thick FFPE tissue sections of liver from both the gilts and their piglets, we performed the terminal deoxynucleotidyl transferase dUTP nick-end labelling (TUNEL) assay using a commercial kit (ApopTag® peroxidase in situ apoptosis detection kit; Chemicon, Temecula, CA, USA) according to the manufacturer's instructions. Finally, the tissue sections were counterstained with Mayer's hematoxylin and coverslipped. Tissue sections of porcine kidney incubated with RQ1 RNase-Free DNase (M6101; Promega, Madison, WI, USA) were used as the positive control, and tissue sections of porcine kidney that were only incubated with the label solution (without terminal deoxynucleotidyl transferase) served as the negative control.

Using light microscopy, we counted TUNEL-positive cells in 30 randomly selected high-power fields (HPF), and also noted whether they were apoptotic hepatocytes or apoptotic cells in hepatic sinusoids. For hepatocytes to be considered TUNEL-positive, they had to have a clearly stained nucleus. Similarly, apoptotic cells in hepatic sinusoids had to exhibit moderate to marked nuclear staining.

#### *2.4. Determining the Proliferation Index in the Liver of Gilts and Their Suckling Piglets*

The proliferation activity of hepatocytes was evaluated on 4 μm thick FFPE tissue sections of liver from both the gilts and their piglets using immunohistochemical labelling with the mouse monoclonal antibody raised against human Ki-67 antigen, clone MIB-1 (Dako, Glostrup, Denmark), which was diluted 1:75. The antigen retrieval was performed by microwave treatment at a medium power (550 W) for 15 min in ethylenediaminetetraacetic acid (EDTA) with a pH of 8.0. The tissue sections were then incubated with primary antibodies for 1 hour at room temperature in a humid chamber. Endogenous peroxidase activity was quenched in the peroxidase-blocking solution Dako REALTM (Dako, Glostrup, Denmark) for 30 min at room temperature. The visualization kit Dako RE-ALTM EnVisionTM Detection System Peroxidase/DAB+, Rabbit/Mouse (Dako, Glostrup, Denmark) was applied according to the manufacturer's instructions. Finally, the tissue sections were counterstained with Mayer's hematoxylin and coverslipped. Tissue sections

of porcine spleen were used as the positive control, and tissue sections of porcine liver that were not treated with primary antibodies served as the negative control. From one of the experimental gilts, a tissue section of the liver was not acquired due to lack of adequate FFPE tissue.

The proliferation index of hepatocytes was calculated as the rate of Ki-67-positive nuclei in a total of 1000 counted nuclei in the tissue sections of liver under a light microscope.

#### *2.5. Morphometrical Evaluation of Interlobular Connective Tissue in the Liver of Gilts*

The amount of interlobular connective tissue was measured in 4 μm thick FFPE tissue sections of liver samples only from the gilts. The tissue sections were stained with Goldner's Masson trichrome stain to clearly depict fibrous connective tissues under a light microscope coupled with a digital camera. Using the software program NIS-Elements Basic Research (Nikon Instruments Inc., Tokyo, Japan), five consecutive microphotographs at HPF were made for each tissue section and represented the area of measurement. The microphotographs were then converted into a binary-colored output by marking pixels that belonged to either interlobular connective tissue or parenchyma, thus enabling automated detection of interlobular connective tissue. The amount of interlobular connective tissue was expressed as the area fraction of the corresponding pixels out of the total number of pixels in the area of measurement. When necessary, the automatically detected areas of interlobular connective tissue were corrected manually.

## *2.6. Statistical Analysis*

For statistical analysis, we used the R statistical software, version 3.6.2 (R Foundation for Statistical Computing, Vienna, Austria) [44]. The obtained results for the experimental and control groups of both the gilts and their piglets are presented with basic descriptive statistics. The Shapiro–Wilk test was used to assess the normality of the variables. For both gilts and piglets, the differences between the experimental and control group were analyzed with the two-tailed Mann–Whitney U test because most of the variables had a non-normal distribution. The correlations between histopathological changes, the number of apoptotic cells and the proliferation index were assessed separately for the gilts and their piglets with Spearman's rank correlation coefficients and Holm's adjusted *p*-values. Statistical significance was determined as *p* < 0.05, and 0.05 ≤ *p* < 0.1 was marginally significant.
