Differential Toxicological Outcome of Corn Oil Exposure in Rats and Mice as Assessed by Microbial Composition, Epithelial Permeability, and Ileal Mucosa-Associated Immune Status

Abstract

Studies to evaluate the toxicity of xenobiotics on the human gut microbiome and related health effects require a diligent selection of (1) an appropriate animal model to facilitate toxicity assessment in predicting human exposure, and (2) an appropriate non-interfering vehicle for the administration of water insoluble compounds. In biomedical studies with water insoluble xenobiotics, corn oil is one of the most commonly used nonaqueous vehicles. This study evaluated the suitability of corn oil as a vehicle in adult female Sprague Dawley rats and adult CD-1 mice; the rodent models that are often utilized in toxicological studies. We studied the host response in terms of change in the intestinal microbiome and mRNA expression of intestinal permeability and immune response-related genes when water (control) and corn oil (2 ml/kg) were administered as a vehicle through oral gavage. The results showed that the use of corn oil as a vehicle has no adverse impact in rats for either the immune response or the intestinal microbial population. On the other hand, mice treated with corn oil showed changes in bacterial community adhered to the ileum, as well as changes in the mRNA expression of intestinal permeability-related and ileal mucosa-associated immune response genes. Overall, results of this study suggest that the type of rodent species and vehicle used in toxicological risk assessments of xenobiotics studies should be taken into consideration in the experimental setup and study design.

Vegetable oils are one of the vehicles of choice for water insoluble xenobiotics undergoing toxicological assessment in animal models (Gad et al., 2016). Oils from soybean, coconut, olive, corn, sesame, and peanut have been used as vehicles for FDA-approved drugs (Savla et al., 2017). Different diet types and changes in dietary choices can lead to significant compositional differences or shifts in gut microbiota (Voreades et al., 2014; Zinocker and Lindseth, 2018). Such microbiota shifts have also been accompanied with physiological responses, for example, modulation of membrane permeability, changes in immune response, shift of circadian rhythm, and/or other metabolic effects (Costantini et al., 2017; Fayfman et al., 2019; Molinero et al., 2019). Diverse effects on the host physiology was noted when oils such as corn, palm, canola, safflower, and fish oil, or lard have been used in research investigations as a source of fats and omega-3/omega-6 polyunsaturated fatty acids (PUFA) (Andersen et al., 2011; Caesar et al., 2015; Costantini et al., 2017; Ghosh et al., 2013; Kaliannan et al., 2015; Kubeck et al., 2016; Nielsen et al., 2007; Pu et al., 2016; Yu et al., 2014). Moreover, diet can have different impact on lean and obese subjects. For example, lean dogs fed with different diets (normal baseline diet, high protein low carbohydrate diet, and high carbohydrate low protein diet) did not increase the richness of the intestinal microbial population; however, in obese dogs different dietary compositions play a significant role in intestinal microbial diversity (Li et al., 2017). Variations in gut microbial populations were also observed while using different sources and combinations of oils (Jiang et al., 2017; Caesar et al., 2015; Tamura et al., 2012). Phylum level analysis of microbiota using both colonic and fecal samples collected from fish oil, lard, or soybean oil fed separately to mice showed enrichment of Proteobacteria in fish oil-fed groups, whereas the lard-fed group showed more Verrucomicrobia and Tenericutes. Increased expression of IL-1β, IL-6, IL-17, IL-18, and TNF-α was also observed in colonic tissue of fish oil-fed mice (Jiang et al., 2017). An abundance of Bacteroides, Turicibacter, and Bilophila genera was noticed in the lard-fed mice, whereas the fish oil-fed mice showed an abundance of Bifidobacterium, Lactobacillus, Streptococcus, and Akkermansia in the cecal samples (Caesar et al., 2015). Two independent studies, conducted on mice using corn oil and a corn oil + fish oil blend to assess the effects of omega-6/omega-3 ratio on gut microbial populations, established that an increase in omega-6/omega-3 can increase the pro-inflammatory bacterial populations, whereas lower omega-6/omega-3 ratios increase anti-inflammatory bacterial population in gut (Kaliannan et al. 2015). Enrichment of bacteria from the Enterobacteriaceae family was observed in high omega-6 PUFAs supplemented diet, whereas omega-3 PUFAs enriched the Lactobacillus and Bifidobacteria genera of phylum Firmicutes (Ghosh et al., 2013). The influence of diet on intestinal microbiome composition and functions is well studied in human as well as animals; however, there have been a paucity of reports on the effects of the vehicle like corn oil used in toxicology risk assessments on the intestinal microbial community, intestinal permeability, and immune response.

Multiple factors contribute to the composition, diversity, and function of the gut microbiome (Sutherland et al., 2020). Choosing an inappropriate animal model or vehicle can affect the drug pharmacokinetics and toxicity data. For example, pyraclostrobin is a fungicide used to control plant pathogens which inhibits mitochondrial respiration of both fungal and mammalian cells. Because pyraclostrobin residues in foods can result in potential human exposure, toxicological studies have been conducted. Adverse effects in mice were observed at a much lower dose of pyraclostrobin when dissolved in corn oil versus the aqueous methoxy cellulose or carboxy methoxy cellulose solutions (Tuttle et al., 2019). Corn oil alone has also been shown to modulate gene expression in the thymus in a dose-dependent manner (Geng et al., 2012). Long-term dietary corn oil use promotes the azoxy-methane-induced colon cancer in rat (Wu et al., 2004). Corn oil (as vehicle for gavage) could cause damage to the renal and reproductive system in pregnant and/or lactating Sprague Dawley rats (Sato et al., 2000). The long-term use of corn oil in Mongolian gerbils during gestation resulted in depleted sperm reserve in adult male offspring (Negrin et al., 2018). Most of the above studies used corn oil at a range from 2.0 to 10.0 ml/kg of animal weight for toxicological risk assessment (Fang and Lin, 2008; Geng et al., 2012; Sato et al., 2000; Wu et al., 2004).

A data-mining study to predict the safe dosing level of drug delivery vehicles for in vivo animal studies was published in 2016 (Gad et al., 2016). The inclusion criteria for this published study included maximum tolerated use levels by animal species, route, dose-limiting toxicity, and duration of study. To our knowledge, a systematic experimental study to assess the effects of drug delivery vehicles alone on ileal mucosa-associated microbial population and its impact on intestinal permeability and gut-associated immune status has not been done. Our main objectives were 2-fold: first, to address the effects of corn oil on the ileal mucosa-associated microbiome using both rat and mouse models. Data obtained from water-gavaged rodents are considered baseline for comparison in the current study. Second, determine which rodent model has a more stable baseline intestinal microbiome upon exposure to corn oil. In addition, we also investigated the effect of corn oil on the mRNA expression of intestinal permeability and immune response-related genes, as well as, examined intestinal cytokine profiles.

MATERIALS AND METHODS

 

 

Animals and treatment

All animals were housed in the animal facility at Southern Research, Birmingham, Alabama, an independent, scientific contract research organization that supports fundamental research planned by NIEHS and other government organizations and was approved by the Southern Research Institutional Animal Care and Use Committee.

Female Sprague Dawley (Hsd:Sprague Dawley® SD®) rats were procured from Envigo, Haslett, Michigan and female CD-1 mice were procured from Charles River, Raleigh, North Carolina. Animals were placed on irradiated NIH-07 feed upon arrival and group housed with irradiated bedding (PJ Murphy Forest Products, Bowling Green, Kentucky) with red tunnels used as environmental enrichment. Ten healthy rats and 10 healthy mice (13–14 weeks old) were used in this study. Five animals from each group (rat and mouse) were gavaged with water (5 ml/kg BW and 10 ml/kg BW DI water, respectively) or corn oil (2 ml/kg [obtained from Welch, Holme & Clark Co., Inc., Newark, New Jersey]) daily for 1 week before sacrifice. The water-treated animals served as the control animals for the respective rodent species. Gavaging of animal was conducted in highly controlled environmental conditions and following all precaution, specifically, individual sterile syringe and needle was used for each animal to prevent cross contamination. Before oral gavage, all syringes were filled with corn oil and water to prevent the cross contamination. In addition, gloves were changed to prevent the cross contamination between animals; and all the animals were gavaged in a similar manner.

Ileum tissue harvesting for DNA, RNA, and protein

Animals were sacrificed by exposure to carbon dioxide. Immediately after sacrifice, the outer surface of the skin was cleaned, intestine was flushed with sterile PBS, and ileum samples were collected in a sterile environment and placed into 2 ml external thread self-standing sterile cryovials certified to be free of DNase, RNase, ATP, Pyrogens, and Human DNA by Phenix Research Products. Samples were then flash frozen in liquid nitrogen and placed on dry ice and stored at −80°C.

DNA extraction for mucosa-associated microbial population

Ileum was minced in small pieces using a scalpel under sterile conditions. Tissue lysate was prepared by transferring these small pieces into bead-beating tubes and placing these tubes in a Fast Prep Machine as described previously (Khare et al., 2004; Williams et al., 2015). The lysate was incubated at 65°C for 20 min with proteinase K followed by treatment with RNAase-A (37°C for 15 min). An equal volume of PCI solution (phenol-chloroform-isopropanol) was added to the cell lysate and centrifuged for 30 min at 12 000 × g. The aqueous layer was transferred to a clean tube, and DNA was precipitated by adding sodium acetate, isopropanol, and polyacryl carrier. Pellet was washed twice with 70% ethanol and air dried and suspended in DNase- and RNase-free water. Double-stranded DNA was quantified by Qubit (Thermo Fisher Scientific, Waltham, Massachusetts) and used as the template for 16S rRNA sequencing for microbial population.

16s rRNA gene sequencing of ileal mucosa-associated bacterial population

PCR primers 515/806 were used to amplify 16S rRNA gene as the template for sequencing for microbial population as described earlier (Gokulan et al., 2018). Purified and pooled PCR products were utilized to prepare Illumina DNA library. For 16S rRNA sequencing 2 × 250 bp Miseq was used. QIIME software was used for operational taxonomic units (OTU) detection and clustering (Caporaso et al., 2010). PCR Sequences (96%) were joined and depleted of barcodes, and then sequences of <150 bp and sequences with ambiguous base calls and with homopolymer runs exceeding 6 bp were also removed (Gokulan et al., 2018). Sequences were denoised, OTUs were generated, and chimeras were removed. OTUs were defined by clustering at 97% similarity. Finally, OTUs were taxonomically classified using BLASTN against a curated database derived from the Greengenes version 13.5 (http://greengenes.secondgenome.com/), RDPII (http://rdp.cme.msu.edu), and NCBI (www.ncbi.nlm.nih.gov), and compiled into each taxonomic level by “counts” (actual number of sequences) as well as “percentage” (relative proportion of sequences within each sample).

The sequence data were filtered to remove low quality and uninformative features to improve the statistical analysis and normalized to address the variability in sampling depth. The community profiling and nonparametric Chao1 index was analyzed using MicrobiomeAnalyst and R-package (Chong et al., 2020; Dhariwal et al., 2017; R-Core-Team, 2016; Weiss et al., 2017). Differences in the bacterial community between samples were evaluated through the beta diversity. The beta diversity was analyzed on a Bray-Curtis dissimilarity index using PERMANOVA and was visualized as PCoA plots (Ramette, 2007).

To gain ontology classification from 16S rRNA data, the output was matched to top hit compilation using reference sequence. Moreover, we used the identity analysis where the percent divergence dictates the level of OTU classification. For this, full taxonomic lineage as well as the identity of the match to the BLASTn database is complied. Bit scores were selected over e-values because they are comparable between different searches and databases whereas e-values represent the chance of obtaining a better match in the same database by random chance (http://www.ncbi.nlm.nih.gov/books/NBK21097/#A614). This analysis compile the OTU to the most relevant taxonomic level based upon the percent identity. Genus level classification is based on the identity to reference sequence between 95% and 97%. OTU file is presented as the Supplementary Table 1 with details on the % homology, e-value, and bit score.

RNA extraction from ileum, cDNA synthesis, and genes profiling

For RNA extraction, rat and mouse ileal tissue samples were minced, homogenized, and suspended in 1.0 ml ice-cold Trizol reagent (Ambion, Austin, Texas). RNAs were then extracted following standard Trizol-chloroform extraction and iso-propanol precipitation steps as described earlier (Williams et al., 2015). Extracted RNA samples were treated with DNase (Ambion, Austin, Texas) following the manufacturer’s protocol. cDNAs were prepared using 2.5 µg DNase-treated RNAs in a 10 µl reaction volume using a Superscript VILO IV (Invitrogen) reverse transcription kit following standard protocol. Gene expression analyses were performed using cell junction pathway finder (PAMM 213Z and PARN 213Z) and antimicrobial response profiler (PAMM 148Z and PARN-148Z) kits from Qiagen, for mouse (PAMM) and rat (PARN) samples, respectively. β-actin was used as reference gene for both mouse and rat antimicrobial response data analysis. β-Glucuronidase was used for mouse and hypoxanthine phosphoribosyltransferase-1 genes were used as reference for rat cell junction pathway data analysis. Fold regulations were determined by normalizing the data using 2^(-ΔΔC_t) value. Genes with a mean 2-fold or greater up or down regulation and a p-value ≤.05 were considered for discussion.

Protein extraction from the ileal tissue

Ileal tissue was minced on the dental wax using a sterile surgical knife and transferred in Eppendorf vial and 0.5 ml ice-cold sample lysis buffer (Bio-Rad, Hercules, California) was added. Minced tissue was homogenized using a handheld homogenized for 1 min. The lysate was centrifuged at 4°C for 10 min at 750 rpm and the clear homogenate was transferred to 2 ml Eppendorf tubes and centrifuged at 4°C for 15 min at 12 000 rpm. The clear supernatant was transferred into new vial and stored at −80°C until cytokine profiling. The protein concentration was measured spectrophotometrically using Bio-Rad Protein Assay reagent (Bio-Rad, Hercules, California).

Multiplex cytokine assay

Gut-associated mucosal chemokine and cytokine levels were evaluated in the tissue lysate using the Bio-Plex Mouse and Rat Cytokine 23-Plex (Bio-Rad, Hercules, California) following the manufacturer’s instructions and as described previously using multiplex a Bio-Plex assay (Gokulan et al., 2018).

Statistical analysis

Data for microbial population abundance was calculated as described earlier (Gokulan et al., 2018). In brief, the hierarchical clustering was performed with the hclust function in the stat package as described earlier (Dhariwal et al., 2017). Data integrity was checked, normalized, and the nonparametric Chao1 index analysis performed using MicrobiomeAnalyst, R-packages and significant differences were evaluated using Anova (Dhariwal et al., 2017; R-Core-Team, 2016). Homogeneity of multiple dispersion statistical method was used to determine the beta-diversity using MicrobiomeAnalyst. To assess the overview of bacterial community at the genus level, sequences were rarefying to the minimum library size. Data scaling was conducted by total sum scaling. Similarity analysis was conducted using Euclidean distance and Ward hierarchical clustering algorithm (Ward, 1963). Heatmap is presented as a visual aid to represent similarity analysis. This heatmap shows the alpha and beta diversity in the animals across all the experimental groups. The data for the top microbial population were determined after applying ANOVA. The colored cell in each column corresponds to an individual OTU abundance within specific samples (alpha diversity). Comparison of the data corresponding to all the experimental groups reveals the change in microbial diversity between the experimental groups and thus the beta diversity.

For mRNA gene expression data, 2-fold or greater up or downregulated genes along with a p-value ≤.05 were considered significant. The Bio-Plex data were exported to Bio-Plex Manager 6.2 and comparisons of different treatment groups were performed using the Mann-Whitney test. All statistical analysis was two sided, and p values of .05 or less were considered significant.

RESULTS

 

Alteration in Bacterial Diversity in Mice and Rats During Oil or Water Treatment by Oral Gavage

Ileal mucosa-associated bacterial population diversity was assessed by 16S rRNA sequencing to evaluate if there was a baseline shift in the bacterial community structure due to the administration of corn oil. Five animals were used in each experimental group to assess animal-to-animal variation. To assess the microbial enrichment, total OTU were calculated in each animal; despite a variable microbial enrichment between the animals in each experimental group, the statistically significant (p < .008) lower microbial OTU were generated only from the corn oil-fed mice experimental group (Figure 1A). The nonparametric Chao1 index was used to further analyze the alpha diversity in all 4 experimental groups. The reason behind choosing to calculate alpha diversity by Chao1 was to measure the OTU diversity based on rare abundance. As shown in Figure 1B, the alpha index diversity shows a large variation in mice (Chao1index p-value:.02) treated with corn oil, as compared with water-treated mice. Here, we show the alpha diversity for an individual animal as well as an average of each experimental group (Figure 1B). However, in rats both treatments showed less variation in the alpha diversity (Figure 1B). The global comparison for significant differences among the bacterial communities showed clear clusters within the experimental groups; water-treated groups versus corn oil-treated groups, both in mice and rats (Figures 1C and 1D). The analysis revealed that mice treated with corn oil versus water showed statistically significant differences ([PERMDISP] F-value: 11.077) (p-value: .010). In contrast, in rat the analysis revealed there is no statistical difference between corn oil versus water treat animals ([PERMDISP] F-value: 0.27711; p-value: .61). The principal component analysis (PCA) reveals that PC1 to PC2 together had separation of 69.2% and 75.2% in mice and rats, respectively. The microbial community clustering indicates less difference in intestinal bacterial community structure in rats upon administration of either water or corn oil.

Figure 1.

Microbial enrichment and diversity in mice and rats during oil or water treatment by oral gavage. A, The bar diagram shows total operational taxonomic units (OTU) counts in rodents (n = 5 in each experimental group). B, The box diagram shows the alpha diversity index (Chao1) of each experimental group either treated with corn oil or water. The PCA plot shows greater variability in the bacterial populations of corn oil-treated mice (C) as compared with corn oil-treated rats (D).

Treatment of Corn Oil Differentially Affects Sequence and Bacterial Richness in Mice and Rats

The 16S rRNA sequencing data were used to analyze operational species richness after exposure to corn oil or water in rodents. As described in the Materials and Methods section, the operational species level classification is based on the >97 identity to reference sequence. The y-axis of rarefaction curve data shows the operational species abundance for individual animals (Figure 2). Mice treated with corn oil had less microbial richness (between 300 and 625) (Figure 2). In contrast, mice treated with water had a higher level of bacterial diversity and values fell between 625 and 875. The rarefaction curve data in rats showed higher operational species richness in corn oil-treated rats (625–750), as compared with water-treated rats (500–750). The x-axis of rarefaction curve showed the number of sequences reads of individual animals matched the operational species richness.

Figure 2.

Rarefaction curves from experimental groups. The rarefaction curve analysis shows the operational species abundance at different numbers of reads for individual animals. A, Top left figure shows the mice treated with corn oil. The y-axis shows the operational species richness, which falls between 300 and 625. B, Top right figure shows operational species abundance of mice treated with water, which had high operational species richness (625–875). C, The bottom left figure shows the operational species abundance in rats treated with corn oil that had high operational species richness (625–750), and D, The bottom right figure shows the operational species abundance of rats treated with water with also a high operational species richness (500–750).

Corn Oil Treatment Causes Changes in the Bacterial Abundance at Phyla Level in Mice

Figure 3 (upper panel) shows stacked bar diagram of the 5 most abundant phyla (relative abundance) in the test groups. On an average, rats administered with water had a high relative abundance of Firmicutes (76%), followed by Protobacteria (16%). There was a decrease in the relative abundance of Protobacteria (4%) and increase in the relative abundance of Firmicutes (87%) when rats were given corn oil. Analysis in the mice showed similar phylum distribution when water was used as a vehicle. In contrast, mice receiving corn oil had an increased abundance of Protobacteria (28%) and Actinobacteria (17%) compared with mice receiving only water. Sample to sample variation for the relative abundance of individual animal is provided in Figure 3 (lower panel).

Figure 3.

Stacked bar diagram of relative abundance of top 5 phyla in mice and rats during water or corn oil treatment. The 100% stacked bar diagram shows the bacterial diversity between water and corn oil treatment in mice and rats. The upper panel shows the average relative abundance of five animals in each experimental group (mice treated with water or corn oil and rats treated with water or corn oil). The lower panel shows the relative abundance of top 5 phyla in each animal.

As there was sample to sample variation for the presence of various phyla in the experimental groups, a stringent analysis (R Studio Statistical package) was conducted to quantitatively show presence and abundance of bacterial phyla that were present in all samples of each experimental group in word cloud (Supplementary Figure 1). The word cloud represents the OTU (classified as Phyla in mice and rats: Supplementary Figures 1A and 1B; or Figures 1C and 1D) that were present in all the samples of that experimental group (n = 5 in each group). The 16S rRNA sequencing data reveals that intestinal bacterial community structure was very similar in rats, despite differences in treatment (water and corn oil). As shown in Supplementary Figures 1A and 1B, there was only a slight decrease in the abundance of Protobacteria when rats were given corn oil. Moreover, in the word cloud we can visualize more phyla with their abundance as shown by larger and darker fonts for higher abundance. The results in mice (Supplementary Figures 1C and 1D) revealed there was a change in the bacterial community structure upon corn oil treatment to mice when compared with water-gavaged control mice.

Corn Oil Treatment Modifies the Abundance of Mucosa-Associated Microbial Population in Mice

Overview of bacterial genera in different experimental groups is shown as the heatmap in Supplementary Figure 2 (all genera). Visibly distinct feature that could be noticed from this heatmap is that the relative abundance of bacterial genera differed in the animals in any experimental groups. We further performed heatmap analysis of the top genera that showed a greater variation, both in dominance, as well as shift in the bacterial genus due to vehicle (corn oil) treatment (Figure 4). Despite differences in the relative abundance, a clear pattern was observed for the predominance of certain genera in mice and rats, as well as, during water or corn oil exposure. The heatmap analysis also revealed that in general Acinetobacter, Brevundimonas, Peptoclostridium, Anoxybacillus, and several other genera were less abundant in rats compared with mice. In rats, corn oil treatment did not cause much perturbation in the genera diversity. Higher relative abundance of Candidatus genus was noticed in oil-gavaged rats as compared with water-gavaged rats; whereas Lactobacillus, Clostridium, and Turicibacter were higher in water-gavaged rats as compared with oil-gavaged rats. In mice, the corn oil treatment caused higher abundance of several genera (Acidovorax, Azohydromonas, Halospirulina, Pelobacter, and Delftia) as compared with water-gavaged mice. Classical univariate statistical comparisons were conducted using Mann-Whitney/Kruskal-Wallis test to identify genera that were significantly difference in the experimental groups. Eight genera showed statistically significant differences among the experimental groups (Figure 5).

Figure 4.

Heatmap of microbial genera present in each animal (mice and rats) gavaged with water or corn oil. Hclust function in R package stat was used to generate heatmap. For generating this heatmap, the OTU data were rarefied to minimum library size. Similarity analysis was conducted using Euclidean distance and Ward hierarchical clustering algorithm. Heatmap color (blue to dark red) displays the row-scaled relative abundance of each genus across all samples.

Figure 5.

Plot of genera that were statistically significantly abundant among the rodents during water or corn oil treatment. Classical Univariate statistical comparisons were conducted using Mann-Whitney/Kruskal-Wallis test to compare the relative abundance of genera in each experimental group (n = 5 in each experimental group). The median, and the box delineates the upper and lower quartile. The whiskers show the maximum and minimum values. The statistical significance (p values) within the experimental groups were as follow: Anoxybacillus = .00090066; Candidatus = .0028974; Rothia = .004563; Peptoclostridium = .02; Turicibacter = .03; Streptococcus = .04; Thermoascaceae = .05; Stenotrophomonas = .06.

Corn Oil Treatment Disturbs Cell-Cell Integrity and Permeability by Altering mRNA Gene Expression in Intestinal Epithelial Cells in Mice

The bacterial community plays an essential role in maintaining intestinal homeostasis. Intestinal integrity and permeability are maintained by various proteins that are encoded by group of genes that include gap junction, cell signaling, tight junction, junctional adhesion molecules, desmosomes, claudin, occluding genes, etc. The microbiome analysis data showed differences in the intestinal mucosa-associated bacterial community structure due to corn oil treatment in mice; hence, we evaluated mRNA expression in intestinal epithelial cell permeability-related genes from all experimental groups by a qPCR method. mRNA expression was profiled both in rats and mice. The analysis revealed that fewer genes associated with permeability had altered expressions in rats (21 genes) in comparison with mice (54 genes). We observed 2 different patterns of gene expression: (1) either up or down regulation at different magnitudes for the same gene in both rats and mice, and (2) opposite pattern for mRNA expression of genes with different magnitudes among rats and mice. In most cases, we noticed regulation of genes associated with intestinal permeability was greater than 2-fold higher in both rats and mice in comparison with the respective controls. However, they were not statistically significant in rats.

In mice, 54 genes were altered (31 upregulated and 23 downregulated) but only 7 genes were statistically significantly differentially regulated (Figure 6); four genes were upregulated (Cldn9, Gjd2, Tjp2, and Tjp3) and 3 were downregulated (Cav3, Cldn1, and Itgal1). In rats, intestinal epithelial cell permeability genes profiling revealed 10 out of 21 genes with above 2-fold regulation; however, none meet the p-value ≤.05 criteria to be considered significant.

Figure 6.

mRNA expression of genes involved in the mouse cell-cell junction pathway. Fold change in expression of genes involved in the cell junction pathway is shown as bar graph. Single (*) and double (**) stars signify the p values ≤.05 and ≤.001, respectively. Fold changes shown here are the average of gene expression for 5 mice, each gavaged with corn oil (gray bar) compared with mice gavaged with water (white bar; used as control).

Corn Oil Differentially Modulate mRNA Expression of Ileal Mucosa-Associated Immune Response Genes in Mice

Antibacterial profiler array was used to determine the immune status-related mRNA gene expression in rats and mice. In rats, the results show that both experimental groups (water vs corn oil) had an almost similar pattern of mRNA gene expression. The mRNA expression of only 3 genes was differentially regulated (>2-fold differential regulation), however, none of them were statistically significant (Supplementary Table 2).

In contrast, comparison of mRNA expression of immune response related genes in mice (water gavaged vs corn oil gavaged) showed that a total of 73 genes were differentially regulated (66 upregulated and 7 downregulated). Upon statistical analysis, 30 genes had significantly altered mRNA expression after corn oil treatment compared with water-treated mice (p values ≤.05). Among them, 28 genes were upregulated and 2 were downregulated. As shown in Figure 7, differentially regulated genes belong to Toll-like receptor (TLR; 6 genes; Figure 7A), Nod-like receptor signaling pathway (8 genes; Figure 7B), antibacterial signaling pathways (6genes; Figure 7C), apoptosis (6 genes; Figure 7D), inflammatory response (3 genes; Figure 7E), and cytokine response (1 gene; Figure 7F). This result indicates that corn oil treatment can induce immune responses in mice.

Figure 7.

mRNA expression of genes involved in the signaling pathways. Fold change in expression of genes involved in the immune response pathway is shown as the bar graph. Single (*) and double (**) stars signify the p values ≤.05 and ≤.001, respectively. Fold changes shown here are the average of gene expression for 5 mice, each gavaged with corn oil (gray bar) compared with the mice gavaged with water (white bar; used as control). Genes were grouped together based on their roles in different stages of immune response pathways, as (A) toll-like receptor signaling pathways, (B) nod-like receptor signaling pathways, (C) signaling pathways downstream to toll- and nod-like receptor signaling pathways, (D) apoptotic pathways, (E) genes of the inflammatory response pathway and, (F) the cytokines genes found to be significantly up/downregulated.

Cytokine Production Is Altered in the Ileum of Mice Due to Corn Oil Treatment

To validate mRNA expression results, we evaluated the cytokines from ileal tissue of mice and rats using the multiplex assay. The multiplex assay revealed that corn oil-treated mice had significantly higher levels of proinflammatory cytokines (IL-1α, IL-1β, IL-4, IL-9, and IL-12 [P70]) (Figure 8) compared with water-treated mice. In contrast, the cytokine production in rats was consistent with mRNA expression data indicating that cytokine production was not altered due to the corn oil treatment.

Figure 8.

Intestinal cytokines levels upon treatment of corn oil. The bar diagram shows the observed concentration of different cytokines; * signifies p values ≤.05. Concentrations of cytokines shown here are the average of the cytokine concentrations obtained for 5 mice gavaged with water (white bar) and corn oil (gray bar), respectively. Rats did not show any significant change in the cytokine levels (data not shown).

DISCUSSION

In the present study, we evaluated 2 rodent species (female Sprague Dawley rats and CD-1 mice) that are commonly used in the toxicological risk assessment of xenobiotics. Thirteen to 14 weeks old rodents were used in this study as this age is considered to have increased stability of intestinal homeostasis, including microbial diversity and immune responses (Laukens et al., 2016). We assessed shift of intestinal bacterial community structure, as well as gut-associated mRNA expression profile of key genes involved in gut homeostasis during exposure to water and corn oil. Our study shows that corn oil-gavaged rats did not show significant impact on intestinal bacterial diversity, mRNA expression of intestinal permeability, or immune response. In contrast, mice showed changes in the bacterial community structure in the ileum, mRNA expression of intestinal permeability related genes, and increased the production of cytokines/chemokines involved in the gut-associated immune responses.

In healthy individuals, the composition of commensal bacteria is a stable microbial community which shapes the development of the local immune system in the gut (Ley et al., 2006; Qin et al., 2010). Protobacteria have been shown to increase 3 to 4-fold in a disease state (Shin et al., 2015). Abnormal expansion of Protobacteria and Actinobacteria phylum in mice treated with corn oil is indicative of a shift from a healthy microbial environment to an imbalanced microbial community with an inability to maintain the normal gut homeostasis. Some genera of Actinobacteria are harmful to both animal and human health. For example, in human Mycobacterium, Rhodococcus, Gordonia, Nocardia all belong to the Actinobacteria phylum and are opportunistic pathogens.

To address the concern of corn oil washing away bacteria from the mucosal area, we calculated total OTU obtained from each sample (Figure 1A). Only the corn oil-fed mice showed a statistically significant (p < .008) lower microbial OTU. Similarly, Chao index (Chao1 p-value .02) between the water-exposed mice and corn oil-exposed mice suggested that the operational species richness of the bacteria was affected during corn oil treatment. Bacterial DNA obtained from these animals was also assessed to determine total bacterial population by real-time PCR using universal primer. Total bacterial number of corn oil-gavaged mice were increased significantly (data not shown), which proves less bacterial diversity, despite of having significant increase in total number of bacteria. The rarefaction data show that operational species richness was less in mice treated with corn oil as compared with other experimental groups. Thus, it could be a possibility that the commensal bacterial population is washing away in this particular experimental group, which results in an increase in Protobacteria and Actinobacteria and forming an unstable microbiome in the gut and led to changes in gut homeostasis. Several environmental factors can contribute change in the intestinal microbiome (Ericsson and Franklin, 2015). Keeping these factors in mind, the present study was conducted in well controlled manner to prevent the influence of environmental or operator-induced contamination; thus, it precludes any introduction of pathogenic bacteria during the oral gavage of oil in mice. The increased Proteobacteria only in mice treated with corn oil is also in-line with an earlier study that showed corn oil treatment increased proinflammatory bacteria including Enterobacteriaceae in mice (Ghosh et al., 2013).

The potential to enable accurate classification of individual organisms at very high taxonomic resolution, when using 16S rRNA sequencing, has been a subject of continued debate in the microbial diversity area (Callahan et al., 2016, 2019; Chiodini et al., 2015, 2016; Huebinger et al., 2013; Johnson et al., 2019; Schloss and Handelsman, 2005; Wang et al., 2007). In this study, the perecentage of divergency was used to evaluate the identity, which further dictates the bacterial classification. As described earlier, this classification compiles the OTU to the most relevant taxonomic level based upon the percent identity (Chiodini et al., 2015). Genus level classification is based on the 97% to 95% identity (3% divergence) to reference sequence (Supplementary Table 1). Rats and mice showed significant differences in the predominant bacterial genus level that resides in the intestinal mucosa (Figure 5). Rats showed a prevalence of commensal organisms named Candidatus, a segmented filamentous bacteria (SFB). These SFB are known to bind specifically to the host intestinal epithelium and exert mucosal immune response (Hedblom et al., 2018; Jiang et al., 2001; Talham et al., 1999)^,. On the other hand, mice showed a predominance of a closely related operational species Peptoclostridium difficile that also bind tightly to the host intestinal epithelium (Davis and Savage, 1974). P. difficile (formerly Clostridium difficile) is an anaerobic bacterium that usually resides asymptomatically in the healthy gut (Rupnik et al., 2009). Mice treated with corn oil had less abundance of P. difficile compared with water-gavaged animals. During dysbiosis, this bacteria is also known to be a major cause of infectious diarrhea in humans (Luo et al., 2016). Apart from P. difficile, several operational species of Acinetobacter were also predominant differentiators between rats and mice, and also within corn oil-treated mice group. Members of Acinetobacter genera are emerging nosocomial pathogens (Linde et al., 2002); a genera that was present in higher abundance in the corn oil-treated mice. Members of this genera are associated mainly with bacteremia in preterm infants and pediatric oncologic patients (Muroi and Tanamoto, 2012). However, none of the animals showed signs of distress or changes in the stool consistency over the 7 days of treatment. In the present study, the operational species representing Lactobacillus intestinalis was found only in water-gavaged rats (data not shown); corn oil-fed animals showed low relative abundance of this bacterial species. It has been shown earlier that the high fat diet (for 16 weeks) considerably decreased the abundance of Lactobacillus intestinalis in rats (Lecomte et al., 2015). Moreover, decrease in the L. intestinalis were shown to be indirectly involved in increases in adiposity (Lecomte et al., 2015). It would be interesting to know the long-term effects of corn oil treatment in mice in terms of bacterial dysbiosis and potential appearance of pathological conditions.

Changes in the mucosa-associated microbial population can also trigger the pathogen-associated molecular patterns. Our results show that multiple genes of the TLR signaling and Nod-like receptor signaling pathways were upregulated in corn oil administered mice. Myd88 plays an important role in activation of NFkB-mediated activation of pro-inflammatory and apoptotic genes. Recruitment of Myd88, IL-1R-associated kinase family proteins (IRAK), and TRAF6 by the TLR/IL-1R family members upon ligand binding serves as the first step in NFkB activation. Upregulation of IRAK1 downregulates TRAF6 via proteasomal degradation of IRAK1-TRAF6 complex and therefore, downregulation of NFkB-mediated pathways (Muroi and Tanamoto, 2012). Our results show significant upregulation of Myd88, Irak1, Nfkbia, and several Mapk family genes (Mapk2k1, Map2k3, Mapk1, Mapk14, and Mapk3). However, no significant upregulation of NFκΒ was observed, suggesting that the TLR signaling pathway has been activated through a Myd88 independent pathway. Myd88 is known to stabilize the interferon-gamma-induced transcripts of tumor necrosis factor and interferon-gamma inducible protein10, showing involvement in TLR independent signaling pathway (Sun and Ding, 2006). TAK1 (MPAKKK family kinase) activation occurs through a complex pathway which requires IRAK1-Traf6 complex. Upon activation, TAK1 activates two separate pathways: (1) NFkB pathway and (2) MAPK pathway (Akira et al., 2006; Kawai and Akira, 2010; Kawasaki and Kawai, 2014). Map3k7 is known to activate MAPKKs- and NFkB-dependent pathways (Ajibade et al., 2013; Dai et al., 2012). mRNA expression of several MAPKs (Map2k1, Map2k3, Mapk1, Mapk14, and Mapk3) was highly upregulated in mice fed with corn oil, thus could be responsible for activation of JNK/ERK signaling pathway, and secretion of pro-inflammatory and apoptotic cytokine/chemokine.

Interferon regulatory factors (IRFs) are transcription factors activated in response to TLR stimulation. IRF1, IRF3, IRF5, and IRF7 are known to induce type 1 interferons in response to stimuli (Jefferies, 2019). Increased expression of IFR5 has been observed during differentiation from bone marrow to macrophage in murine (Krausgruber et al., 2011; Schoenemeyer et al., 2005). IRF5 and RelA bind near transcription start site of inflammatory response genes and can acts in synergy to modulate gene expression (Saliba et al., 2014). However, in the absence of RelA, IRF5 cause downregulation of several genes. We observed more than 2-fold downregulation of RelA, whereas IRF5 was highly upregulated. IRF5 functions downstream to Myd88 and induces secretion of TNF, IL-6, IL-12, and IL-23 while simultaneously repressing IL-10 and TGF-α (Dalmas et al., 2015; Krausgruber et al., 2011). Our result shows that IRF5 and IRF7 both are significantly upregulated, thus probably could induce pro-inflammatory cytokines and interleukins expressions (Zhao et al., 2015).

Corn oil treatment in mice showed upregulation of mRNA expression of Tollip, which acts as an inhibitor of Myd88-dependent NFkB pathway. Tollip, along with Rac1, mediates the internalization of uropathogenic E. coli in bladder epithelial cells (Visvikis et al., 2011). Our results show the upregulation of both Tollip and Rac1, suggesting the perturbation and downstream effects of signaling pathways which are activated in response to bacterial internalization. Moreover, this bacterial internalization could be responsible for activation of NLR-signaling pathway (upregulation of Birc3, Card6/9, Hsp90aa1, Naip1, Pycard, and Sugt1) and apoptotic pathway (upregulation of Jun, Pik3ca, Rac1, Ripk1, Ripk2, and Tnfrsf1a). Supplementary Table 3 provides a comprehensive list for the function of genes involved in the perturbation of mRNA in this study. Loss of commensal bacterial population could have resulted in an increase in the pathogenic bacterial population in these animals. Figure 9 shows the schematic representation for the perturbation of mRNA expression and cytokine release in mice fed with corn oil.

Figure 9.

Schematic diagram showing the perturbation of mRNA expression and cytokine release in mice fed with corn oil. Green arrow shows upregulation and red arrow shows downregulation of respective genes mRNA expression in the pathway.

In the corn oil-fed mice, the cytokine surge and microbial breach in the intestinal mucosal barrier probably could result in the differential expression of mRNA responsible for integrity of cell-cell junctions. In fact, we noticed upregulation of Cldn9, Gjd2, Tjp2, and Tjp3 and downregulation of Cldn1, Cav3, and Itgal. Cldn1 is important for maintaining the tight junction between epithelial cells, therefore, downregulation of Cldn1 is likely to cause increased membrane permeability or leakage through the intestinal epithelium (Krause et al., 2008). Cldn9 has been associated with Na⁺ and K⁺ ion regulation between epithelial cells and the paracellular space (Nakano et al., 2009). Cav3 is plasma membrane associated protein which plays an important role in membrane integrity. Deficiency of Cav3 causes disruption of the membrane, tubule network in skeletal muscle, and signaling pathways (Gazzerro et al., 2010). Cav3 is also involved in Ca⁺² channeling across the membrane (Markandeya et al., 2011). Itgal is an integrin expressing gene which is important for intracellular adhesion as well as trans-membrane signaling (Chigaev and Sklar, 2012; Hogg et al., 2011; Weber et al., 1997). The Gjd2 is responsible for cell gap junctions. Upregulation of Gjd2 is associated with high membrane density of gap junctions and increased intercellular exchange (Carvalho et al., 2010). In our study, the upregulation of Cldn9 along with downregulation of Cldn1 suggests that epithelial permeability might have increased significantly, thus facilitating the entry of luminal content (including microbial population or products) and activation of immune response. Lipopolysaccharide (LPS) mediated activation of NFkΒ is known to downregulate the production of tight junction protein Tjd2 in pulmonary epithelial cells (Han et al., 2004). It cannot be ruled out that oral gavage of corn oil in mice could result in the expulsion of some specific microbial population that adhere to the ileal mucosa. This expulsion may result in the dominance of Gram-negative bacteria and may lead to production of LPS, and thus the activation of TLR signaling and immune response (Figure 9).

In conclusion, the present study shows that corn oil, used as a vehicle, had no significant impact either on immune response or intestinal bacterial community structure in rats. On the other hand, corn oil-treated mice showed shifts in the abundance of bacterial community structure in the ileum, changes in the mRNA expression of permeability related genes, and increased production of several cytokines/chemokines involved in the gut-associated immune responses. In spite of the use of limited number of animals, the promising results obtained from this preliminary study suggests to broaden and replicate similar investigations to assess effects of such treatment in larger population including other species of mice and rats. Likewise, other questions remain to be resolved on the long-term effects of sub-chronic and chronic exposure of xenobiotic delivery vehicles (corn oil as well as other commonly used vehicles) in toxicological studies. In addition, it is crucial to address if the changes observed in mice treated with corn oil could make this animal species more susceptible to microbial infection.

SUPPLEMENTARY DATA

Supplementary data are available at Toxicological Sciences online.

 

Disclaimer: The findings and conclusions presented in this article are those of the authors and do not necessarily represent the views of the US Food and Drug Administration.