Subacute exposure to di-isononyl phthalate alters the morphology, endocrine function, and immune system in the colon of adult female mice

Di-isononyl phthalate (DiNP), a common plasticizer used in polyvinyl chloride products, exhibits endocrine-disrupting capabilities. It is also toxic to the brain, reproductive system, liver, and kidney. However, little is known about how DiNP impacts the gastrointestinal tract (GIT). It is crucial to understand how DiNP exposure affects the GIT because humans are primarily exposed to DiNP through the GIT. Thus, this study tested the hypothesis that subacute exposure to DiNP dysregulates cellular, endocrine, and immunological aspects in the colon of adult female mice. To test this hypothesis, adult female mice were dosed with vehicle control or DiNP doses ranging from 0.02 to 200 mg/kg for 10–14 days. After the treatment period, mice were euthanized during diestrus, and colon tissue samples were subjected to morphological, biochemical, and hormone assays. DiNP exposure significantly increased histological damage in the colon compared to control. Exposure to DiNP also significantly decreased sICAM-1 levels, increased Tnf expression, decreased a cell cycle regulator (Ccnb1), and increased apoptotic factors (Aifm1 and Bcl2l10) in the colon compared to control. Colon-extracted lipids revealed that DiNP exposure significantly decreased estradiol levels compared to control. Collectively, these data indicate that subacute exposure to DiNP alters colon morphology and physiology in adult female mice.

The doses of DiNP were selected to mimic environmentally relevant exposures or typical toxicological doses that could be compared to other studies that used higher doses. For example, 0.02 mg/kg/day was to used reflect occupational exposure 35 , whereas 0.2 mg/kg/day was selected to mimic exposure in children aged 0-18 months who commonly mouth plastic toys containing DiNP 36 . The other three doses (2,20, and 200 mg/kg/day) were selected because they are typical ranges used in traditional toxicology studies on DiNP 9,27,37,38 . Experimental animals. This experiment was performed in an AAALAC (Association for Assessment and Accreditation of Laboratory Animal Care)-approved animal facility. Specifically, this toxicological study was conducted in the Veterinary Medicine Basic Science Building (University of Illinois at Urbana-Champaign). Female outbred CD-1 mice (approximately 2 months old) were purchased from Charles River (Wilmington, MA), group-housed (3 mice per cage), and were allowed to acclimate to the new facility for 7 days. The facility that housed these mice kept ambient temperatures at 21.1 ± 2.2 °C, humidity at 50 ± 20%, and light:dark cycles for 12 h each cycle. Mice were also given Tekland Rodent Diet (8604) and reverse-osmosis-treated water ad libitum. All animal procedures were approved by the University of Illinois Institutional Animal Care and Use Committee or Illinois IACUC (Protocol No.: 20034 and 19110).
Experimental design. Two-month-old female mice (N = 36 mice total; 6 mice per treatment group) were orally dosed by gently pipetting corn oil vehicle (control) or DiNP (0.02, 0.2, 2, 20, or 200 mg/kg) into the mouth. Dosing occurred once a day in the morning for at least 10 days, and the mice were euthanized one hour after their last dosing. Estrous cycles in female mice last 4-5 days 39 and were monitored by daily vaginal smears. Mice were euthanized by CO 2 asphyxiation and cervical dislocation during diestrus. Mice were continually dosed with either corn oil or DiNP until they were in diestrus, which is why some mice were dosed for more than 10 days. All mice were euthanized by day 14 of the experimental timeline. See Fig. 1 for a depiction of the experimental timeline. All methods were carried out in accordance with relevant guidelines and regulations.
Tissue collection. Colons were trimmed to remove mesenteric tissues and flushed with cold, sterile 1X PBS to remove colonic contents before weighing the mass and measuring the length of the colon. Distal colons were cut into pieces approximately 0.5 cm long and placed into sterile Eppendorf tubes. The Eppendorf tubes were then snap-frozen in liquid nitrogen and stored in − 80 °C for RNA, lipid, and cytokine extractions. Colon weight-to-length ratio was calculated by taking the weight (mg) divided the length of the colon (cm).
Colon sections were randomized and graded without knowledge of treatment group on the following attributes: enterocyte sloughing, focal or cellular inflammation, edema, crypt damage, and aberrant colon walls. The scoring system used for grading histological sections in the colon is summarized in Table 1. In detail, the attributes listed above were graded on a scale of 0 to 3, with 0 being normal, 1 being minimal or mild, 2 being moderate, and 3 being severe. Several sections (2-3 sections) were examined and graded for each colon. The grades for each characteristic (enterocyte sloughing, focal or cellular inflammation, edema, crypt damage, and www.nature.com/scientificreports/ aberrant colon walls) were summed to give each tissue a total score. Then, the average was taken for control and treatment groups.

RNA extraction and gene expression analysis.
Frozen tissues were used for RNA extraction using the RNeasy Micro Kit by Qiagen (Qiagen, Inc., Valencia, CA). RNA extraction was carried out according to the manufacturer's instructions. Once RNA was extracted, samples were eluted in RNase-free water. RNA concentration was quantified at λ = 260 nm using NanoDrop (NanoDrop ND-1000, ThermoScientific, Waltham, MA) and then stored at − 80 °C until complementary DNA (cDNA) synthesis. For cDNA synthesis, RNA (100 ng) was reverse transcribed into DNA using iScript Reverse Transcriptase (Bio-Rad Laboratories, Inc., Hercules, CA) according to the manufacturer's instructions. Each qPCR reaction was done in duplicate. Each replicate contained 2 µL of cDNA, forward and reverse primers (5 pmol) for select genes, SsoFastEvaGreen Supermix, and nuclease-free water to give a final reaction volume of 10 µL. Real-time quantitative polymerase chain reactions (RT-qPCR) were carried out using the CFX96 Real-Time Detection System (Bio-Rad Laboratories) and CFX Manager Software.
Lipid extractions and hormone measurements. Before lipid extraction, colon weights were measured so that hormone levels could be normalized to colon weight. Lipids were extracted from the colon to determine concentrations of estradiol and testosterone. To do this, a mixture of PBS (1x):HCl (0.1 N) (5:2 v/v) was added to each sample and then homogenized. Lipids were extracted with ethyl acetate:isopropanol (1:1 v/v) and then separated with PBS:ethyl acetate (3:2 v/v) mixture. The ethyl acetate phase was collected and evaporated using Speedvac at 30 °C for 2-3 h. Lipid samples were stored at − 80 °C until hormone measurements.
Testosterone and estradiol enzyme-linked immunosorbent assay (ELISA) kits were purchased from DRG and used to measure testosterone and estradiol levels in the colon, respectively. Lypocheck (Bio-Rad Laboratories) was used as a control for the DRG ELISA kits. Testosterone and estradiol assays were carried out according to the manufacturer's protocol. ELISA plate absorbance values were read at 450 nm with the Multiskan Ascent microtiter plate reader (Thermo Electron Corporation).
Cytokine extraction and measurements. Frozen tissues stored in − 80 °C were weighed before cytokine extraction. Colon tissues were resuspended in T-PER Tissue Protein Extraction Reagent (ThermoFisher, Rockford, IL) and antifoam SE-15 (Millipore Sigma, St. Louis, MO). Then, tissues were homogenized using a Tis- . The homogenized solution was centrifuged for 10 min at 6000 rpm at 4 °C. The supernatant was recovered and stored at − 80 °C. A sample of the supernatant was also used to determine the protein concentration using Pierce BCA Protein Assay Kit (ThermoFisher Scientific, Rockford, IL). After cytokine extraction from the colon tissues, the samples were subjected to the mouse sICAM-1/CD54 Quantikine ELISA kit (R&D Systems, Minneapolis, MN) and the mouse TNF alpha uncoated ELISA kit (Invitrogen, Waltham, MA). ELISA protocols were followed according to manufacturer's instructions. The sICAM ELISA assay was read with the Multiskan Ascent microtiter plate reader (Thermo Electron Corporation), and the TNF-α ELISA assay was read with the MQX200 UQuant microplate reader (BioTek).

Statistical analysis.
The data presented in this study were analyzed using SPSS Statistics software (SPSS Inc., Chicago, IL) and expressed as means ± standard error of the means (SEM). Data were assessed for normality using the Shapiro-Wilk test. Data that were normally distributed and that met the assumption for homogeneity of variance were analyzed using one-way analysis of variance (ANOVA). If the ANOVA test reported p < 0.05, we conducted post hoc analysis using Dunnett's 2-sided test. The following data were analyzed using ANOVA: colon length, colon weight, colon weight-to-length ratio, estradiol levels, tight junctions (Cldn-1, Zo-1, Zo-2, and Zo-3), inflammation (sICAM-1, Il5, Il6), and cell health (Ki67, Ccna2, Ccnb1, Ccnd2, Ccne1, Cdk4, Aifm1, and Bcl2). Data that were not normally distributed and that did not meet homogeneity of variance were analyzed using the Mann-Whitney U test. The following data were analyzed using the Mann-Whitney U test: colon histology, testosterone levels, tight junctions (Ocln and Cldn-4), cell cycle regulators (Cdkn1a and Bcl2l10), inflammation (Ifng, Tnf, and TNF-alpha).
Statistical significance was assigned with one or two asterisks and defined as 0.01 ≤ p < 0.05 or p < 0.01, respectively. Borderline significance (^) was defined as 0.05 ≤ p < 0.10. Table 2. Primer sequences of genes assessed in the colon tissues of female CD-1 mice using real-time quantitative PCR analysis.

Results
Gross measurements. In all mice, the colon length ranged from 5.8 to 10.2 cm. DiNP exposure at all doses did not significantly affect colon length compared to control ( Fig. 2A). Similarly, DiNP exposure did not markedly alter colon weight compared to control (Fig. 2B). Exposure to DiNP also did not significantly alter the colon weight-to-length ratio compared to control (Fig. 2C).
Histopathology. Histological analysis revealed that DiNP exposure increased colonic damage compared to control (Fig. 3). Specifically, DiNP doses at 0.02, 0.2, 2, and 200 mg/kg significantly increased colonic damage compared to control (p < 0.05, Fig. 3C). Interestingly, the changes at the low doses of DiNP (0.02 and 0.2 mg/ kg/day) were mainly due to cellular infiltration and aberrant colon walls, whereas the changes at high doses of DiNP (2 and 200 mg/kg/day) were mainly attributed to edema. Further, some enterocyte sloughing occurred in the 0.02-2 mg/kg DiNP treatment groups.
Hormone levels. DiNP exposure at most doses did not significantly alter testosterone levels in the colon compared to control, but 0.2 mg/kg/day DiNP marginally decreased testosterone levels compared to control (p = 0.065; Fig. 4A). DiNP exposure significantly decreased estradiol concentrations at 0.2, 20, and 200 mg/kg/ day compared to control in the colons (p < 0.05). However, DiNP exposure at 0.02 and 2 mg/kg/day did not alter estradiol levels in the colon significantly compared to control (Fig. 4B).
Apoptosis and cell proliferation. Expression of apoptotic factors including Aifm1 and Bcl2l10 was also examined to determine the health of cells. DiNP exposure at an environmentally relevant dose (0.2 mg/kg) significantly increased expression of Aifm1 compared to control (p < 0.05). DiNP exposure (20 mg/kg) also significantly increased Bcl2l10 expression compared to control (p = 0.012). Because DiNP increased expression of two proapoptotic factors, Aifm1 and Bcl2l10, compared to control, we conducted TUNEL staining to examine whether DiNP treatment caused the DNA fragmentation that occurs during apoptosis. Interestingly, DiNP exposure did not affect TUNEL staining compared to control (Fig. 5B).

Figure 2. Gross measurements of the colon include length (A), weight (B), and weigh-to-length ratio (C).
Values represent mean ± SEM, n = 6/group. www.nature.com/scientificreports/ In addition to examining apoptotic factors, cell survival factors including Bcl2 were examined in each treatment group. DiNP exposure did not alter expression of Bcl2 compared to control (Fig. 5A). Further, DiNP exposure did not significantly alter Ki67 expression compared to control at any dose (Fig. 5A).
Inflammation. Expression of the following cytokines was measured from the distal colon: Il4, Il5, Il6, Il13, Il17a, Tnf, and Ifng (Fig. 6). DiNP exposure did not significantly alter expression of Il4, Il5, Il6, Il13, and Il17a compared to control. However, environmentally relevant doses of DiNP (0.02 and 0.2 mg/kg/day) borderline increased the expression of interferon gamma (Ifng). Interestingly, an environmentally relevant dose (0.2 mg/kg/ day) of DiNP exposure significantly increased Tnf expression compared to control (p = 0.047, Fig. 6). Although DiNP exposure significantly increased the expression of Tnf compared to control, it did not alter TNF-α protein Red arrows indicate leukocyte infiltration, white arrows indicate aberrant colon walls, a yellow arrow indicates aberrant crypts, a green arrow represents enterocyte sloughing, and a blue arrow represents edema. Multiple sections of the colon were graded and given histological scores (C). A single asterisk (*) indicates significant differences compared to control (0.01 ≤ p < 0.05), and two asterisks (**) indicate very significant differences compared to control (p < 0.01).

Discussion
In the present study, we showed that DiNP exposure significantly increased histological damage in the colon compared to controls. These results are consistent with previous studies that showed that subacute DBP exposure at 500 mg/kg/day significantly altered intestinal histopathology 41 . Specifically, DBP exposure significantly increased villus height and villus height/crypt depth ratio (V/C ratio) in the duodenum, whereas DBP exposure significantly decreased villus height and V/C ratio in the jejunum compared to control 41 . DBP exposure at 50 mg/ kg/day did not alter histopathology in the small intestine 41 . In our study, we also observed significantly altered histopathological changes due to subacute DiNP exposure; however, these significant changes were observed in the colon at 0.02, 0.2, 2, and 200 mg/kg DiNP. Similar to DBP exposure at 50 mg/kg, DiNP exposure at 20 mg/ kg did not alter histopathology in the colon. These data suggest that phthalates alter intestinal histopathology in a dose-dependent manner.
This study also investigated the impact of DiNP exposure on sex hormones in the colon. We measured the effects of DiNP exposure on hormone levels in the intestines because studies indicate that estradiol can be synthesized in gut-associated lymphoid tissues, such as the Peyer's patches 18 . In a previous study conducted on a different set of mice using the same Illinois IACUC protocol, we investigated the impact of DiNP exposure on circulating sex hormones (i.e., hormones in the sera) and showed that DiNP exposure at 0.02, 0.1, and 200 mg/kg significantly decreased circulating testosterone levels and that DiNP exposure at 0.1 and 200 mg/kg significantly decreased circulating estradiol levels compared to control 9 . In the present study, we expanded these findings by showing that in the colon, DiNP exposure borderline decreased testosterone levels and significantly decreased estradiol levels in a dose-dependent manner compared to control. Thus, our previous study and the current study agree that DiNP exposure significantly decreases estradiol levels compared to control, suggesting that DiNP has antiestrogenic properties. Several studies have reported that DiNP is antiandrogenic at specific doses 9,42,43 . The present study shows that DiNP exposure causes a marginal decrease in testosterone levels in the colon compared  44 disrupt the expression of genes involved in apoptotic and cell cycle pathways in the ovary. Thus, we examined whether DiNP exposure alters the expression of mRNAs for various cell cycle regulators as well as cell proliferation and apoptosis in the colon. We found that although DiNP exposure significantly altered expression of apoptotic factors (Aifm1 and Bcl2l10) compared to control, it did not significantly increase DNA fragmentation, an indicator of apoptosis, in the colon. Thus, it is possible that the colon is able to recover from changes in expression of apoptotic factors after subacute exposure to DiNP likely due to the quick turnover of colonocytes.
The critical role of the epithelial barrier in the immune regulation has been well documented in the colon 40 . Previous studies have shown that exposure to some phthalates increased intestinal inflammation and permeability 45 , but those studies did not examine the effects of DiNP on epithelial barrier function and immune Figure 7. Protein levels of cytokine sICAM-1 (A) and TNF-α (B) in colon tissue of adult female CD-1 mice exposed to corn oil control or DiNP (0.02 -200 mg/kg/day). Values represent mean ± SEM, n = 6/group. A caret (^) symbol indicates borderline differences compared to control (0.05 ≤ p < 0.10). A single asterisk (*) indicates significant differences compared to control (0.01 ≤ p < 0.05).

Figure 8.
Relative expression of tight junctions in colon of adult female CD-1 mice exposed to corn oil control or DiNP (0.02-200 mg/kg/day). Values represent mean ± SEM, n = 6/group. Two asterisks (**) indicate very significant differences compared to control (p < 0.01).
Scientific Reports | (2020) 10:18788 | https://doi.org/10.1038/s41598-020-75882-0 www.nature.com/scientificreports/ function. Our current data indicate that DiNP exposure decreases the expression of Zo-3, which may explain why DiNP exposure causes changes in the gene expression and protein levels of various cytokines compared to control. The downregulation of tight junctions provides an ineffective barrier to luminal pathogens. Alternatively, DiNP exposure alters cytokine expression, which disrupts the intestinal tight junction barrier 46 . As a result, this would allow luminal antigens to penetrate intestinal tissues. The present study examined gene expression and protein levels of various cytokines. DiNP exposure significantly increased Tnf expression at an environmentally relevant dose compared to control, but it did not significantly alter protein levels of TNF-α in all treatment groups compared to control. Although we did not observe significant changes at the protein level, it is interesting to note that the gene expression and protein level of this cytokine follow a similar dose-response curve. Although TNF-α levels were not significantly altered, TNF-α can still play a role in apoptosis and regulating other immune cells. It is unlikely that TNF-α plays a role in DiNPinduced apoptosis because TUNEL staining did not reveal significant cell death with DiNP treatment compared to control. It is possible that the interaction between TNF-α binding to TNF receptor type 1 (TNFR1) induces pro-inflammatory effects because we observed cellular infiltration in the colonic histology. We also examined sICAM-1 levels in the colon and observed a negative association between DiNP doses and sICAM-1 levels. sICAM-1 helps activate the immune system by promoting the interaction between macrophages and T-cells. These data suggest that DiNP-induced damage occurs by different mechanisms depending on the dose of DiNP.
To summarize these results, it is possible that when DiNP damages the colon, it induces local inflammation and apoptosis of colonic cells, and it also dowregulates expression of tight junctions. The downregulation of tight junctions could further exacerbate colonic inflammation because it allows microbes and microbial metabolites to pass the epithelial barrier. We observed changes in the cell cycle regulation, and this makes sense as cell cycle regulators are linked to cellular apoptosis and proliferation. The decrease in estradiol levels in the colon may also be contributing to the altered immune responses.
These results may have potential to be generalized to humans for several reasons. First, we exposed mice to environmentally relevant doses of phthalates that are in the range of exposure for humans of all ages. Second, we orally administered phthalates to mice in the way that humans are most commonly exposed to phthalates (i.e., ingestion). Third, mice have similar gastrointestinal anatomy and physiology to humans and have extensively been used as premier animal models for human gastrointestinal diseases. Humans and mice are similar in that they both have limited post-gastric fermentation because they are non-ruminants. Mice and humans also have a mouth, esophagus, stomach, small intestine, large intestine, and rectum. However, humans have a sacculated colon and fermentation largely takes place in the colon, whereas mice have a tubular colon and fermentation occurs in the cecum. The enlarged cecum in mice is much smaller in humans and is also known as the appendix. Thus, more research is needed to determine whether DiNP exposure affects immune responses, endocrine functions, and cell health in the colon of humans in an identical or similar manner in mice due to anatomical and physiological differences that exist in the GI tract of humans and mice.

Conclusion
In summary, this study shows that acute exposure to DiNP increases colonic damage, decreases estradiol levels, downregulates expression of tight junctions, alters cytokine levels, and dysregulates expression of cell cycle regulators in the adult female mouse colon. Further studies are needed to understand the mechanisms underlying DiNP-induced effects on the colon. One possibility is that DiNP exposure alters the diversity of gut microbes, and this leads to changes in histopathology, cytokine level, immune function, tight junctions, and estradiol levels. Therefore, future studies should explore the impact of adult DiNP exposure on the gut microbiome in female mice.