The Role of T cell PPAR γ in mice with experimental inflammatory bowel disease

PPAR γ flfl; CD4Cre⁺ (CD4cre, n = 34) and Cre-wild-type mice (WT, n = 34) in C57BL/6J background were used for these experiments. The CD4cre mice (kindly provided by Dr. R.B. Clark, University of Connecticut) express a transgenic recombinase under the control of the CD4-Cre promoter [18]. Since all T cell precursors express the CD4 co-receptor at the thymic level (during the double-positive thymocyte stage), the PPAR γ gene is deleted from both CD4⁺ and CD8⁺ T cells. The mice were housed at the animal facilities at Virginia Polytechnic Institute and State University in a room maintained at 75°F, with a 12:12 h light-dark cycle starting from 6:00 AM. All experimental procedures were approved by the Institutional Animal Care and Use Committee of Virginia Polytechnic Institute and State University and met or exceeded requirements of the Public Health Service/National Institutes of Health and the Animal Welfare Act. Mice were challenged with 2.5% dextran sodium sulfate (DSS), 36,000-44,000 molecular weight (ICN Biomedicals, Aurora, OH) in the drinking water. After the DSS challenge mice were weighed on a daily basis and examined for clinical signs of disease associated with colitis (i.e., perianal soiling, rectal bleeding, diarrhea, and piloerection). For the DSS challenge, the disease activity indices and rectal bleeding scores were calculated using a modification of a previously published compounded clinical score [2]. Briefly, disease activity index consisted of a scoring for diarrhea and lethargy (0-3), whereas rectal bleeding consisted of a visual observation of blood in feces and the perianal area (0-4). Mice in the DSS study were euthanized on days 0, 2, and 7 of the DSS challenge by carbon dioxide narcosis followed by secondary thoracotomy and blood was withdrawn from the heart. Spleen and mesenteric lymph nodes (MLN) were scored based on size and macroscopic inflammatory lesions (0-3), excised, and single-cell suspensions were prepared as previously described [17] for flow cytometry.

Colonic sections were fixed in 10% buffered neutral formalin, later embedded in paraffin, and then sectioned (5 μm) and stained with hematoxylin and eosin (H&E) for examination of microscopic lesions and changes in the mucosal architecture. Colons were graded with a compounded histologic score including the extent of (1) leukocyte infiltration, (2) mucosal thickening, and (3) epithelial cell erosion. The sections were graded with a score of 0-4 for each of the previous categories and data were analyzed as a normalized compounded score. Five fields per tissue and 11 mice per time point were examined.

MLN and spleen-derived cells or whole blood were seeded onto 96-well plates, centrifuged at 4°C at 3000 rpm for 4 minutes, and washed with PBS containing 5% serum and 0.09% sodium azide (FACS buffer). To assess differential monocyte/macrophage subsets, the cells were then incubated in the dark at 4°C for 20 minutes in FcBlock (20 μg/ml; BD Pharmingen), and then for an additional 20 minutes with fluorochrome-conjugated primary antibodies anti-F4/80-PE-Cy5 (5 μg/mL, ebioscience) and anti-CD11b-Alexa Fluor 700 (2 μg/mL, eBioscience). For lymphocyte subset assessment, cells were incubated with anti-CD4-Alexa Fluor 700 (2 μg/mL; BD Pharmingen), anti-CD8-PerCp-Cy5.5 (2 μg/mL, eBioscience), CD3 PE-Cy5 (2 μg/mL; BD Pharmingen), anti-FoxP3-PE (2 μg/mL, eBioscience), and anti-IL10-PE as previously shown [19]. Flow results were computed with a BD LSR II flow cytometer and data analyses was performed with FACS Diva software (BD).

After homogenization of tissue, total RNA was extracted and purified using the RNAeasy system according to manufacturer's instructions (Qiagen Valencia, CA). The QIAGEN RNase-free DNase supplement kit was used to ensure that the RNA was free from DNA contamination. RNA was then processed and labeled according to the standard target labeling protocols and the samples were hybridized, stained, and scanned per standard Affymetrix protocols at VBI core laboratory on Mouse 430 2.0 expression arrays (Affymetrix Inc., Santa Clara, CA). All statistical analysis of the data was performed within R statistical environment - Version 2.9.0 [20] using Bioconductor packages [21]. Raw microarray data from CEL files were read with 'affy' package [22] and pre-processed by gcRMA algorithm (GC Robust Multiarray Average) that performs the three steps: (i) adjustment of the gene expression signal against the background caused by optical noise and non-specific binding, (ii) robust multi-array normalization [23], and (iii) summarization of the probes belonging to each probe set. The microarray data (both raw and normalized) have been submitted at the Gene Expression Omnibus. (GEO, http://​www.​ncbi.​nlm.​nih.​gov/​geo/​, Data set: GSE20523).

All pathways listed at Kyoto Encyclopedia for Genes and Genomes (Kyoto Encyclopedia of Genes and Genomes) [24] were selected for analysis. Two approaches were taken to find the KEGG pathways significantly affected by DSS in CD4cre mice. In the first approach, a global search was conducted to find the KEGG pathway most significantly affected under the condition. The degree of differential regulation of a pathway in CD4cre (compared to wild-type) mice after 7 days of DSS treatment was derived from a collective differential regulation of the member genes in that pathway, as described earlier [25]. Briefly, each KEGG pathway was assigned a score depending on its degree and direction of modulation, which represented changes in collective gene expression across most genes in each pathway. Normal-normal plot (QQ-plot) and permutation testing were conducted to find the pathway with a value significantly different from the rest of the pathways. In the second approach, the genes differentially expressed on day 7 of DSS treatment, as derived from pair-wise gene-level comparison between day 7 post-DSS and no DSS in CD4cre mice, were selected for exploring which KEGG pathways were over-represented in this collection of genes. These included both up-and down-regulated genes due to DSS. None of these genes were differentially expressed in the wild-type mice at corresponding time point (i.e., day 7) of DSS treatment. Hypergeometric testing was performed on this collection of genes to discover the over-represented KEGG pathways. This procedure used Fisher's exact test to find association between interesting genes (differentially expressed at day 7 in CD4cre mice) and membership to a KEGG pathway. Results from both approaches were combined to list the KEGG pathways that are significantly modulated by DSS on day 7 in CD4cre mice.

KEGG pathways found to be significantly associated with DSS on day 7 in CD4cre mice were accessed using the bioconductor package KEGGSOAP. Specific gene nodes on each pathway were 'colored' according to the direction of differential expression of that gene in CD4cre mice on day 7 of DSS treatment: red if up-regulated, green if down-regulated.

To validate microarray results we used quantitative, real-time, RT-PCR. Total RNA (1 μg) from colons was used to generate a complementary DNA (cDNA) template using the iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA) using previously described conditions [2]. Each gene amplicon was purified with the MiniElute PCR Purification Kit (Qiagen) and quantitated on an agarose gel by using a DNA mass ladder (Promega). These purified amplicons were used to optimize real-time PCR conditions and to generate standard curves in the real-time PCR assay. Primer concentrations and annealing temperatures were optimized for the iCycler iQ system (Bio-Rad) for each set of primers using the system's gradient protocol. PCR efficiencies were maintained between 92 and 105% and correlation coefficients above 0.98 for each primer set during optimization and also during the real-time PCR of sample DNA.

Complementary DNA (cDNA) concentrations for genes of interest were examined by real-time quantitative PCR using an iCycler IQ System and the iQ SYBR green supermix (Bio-Rad). A standard curve was generated for each gene using 10-fold dilutions of purified amplicons starting at 5 pg of cDNA and used later to calculate the starting amount of target cDNA in the unknown samples. SYBR green I is a general double-stranded DNA intercalating dye and may therefore detect nonspecific products and primer/dimers in addition to the amplicon of interest. In order to determine the number of products synthesized during the real-time PCR, a melting curve analysis was performed on each product. Real-time PCR was used to measure the starting amount of nucleic acid of each unknown sample of cDNA on the same 96-well plate.

Flow cytometry, disease activity, pathology and real-time RT-PCR data were analyzed as a completely randomized design. To determine the statistical significance of the model, analysis of variance (ANOVA) was performed using the general linear model procedure of Statistical Analysis Software (SAS), and probability value (P) < 0.05 was considered to be significant. When the model was significant, ANOVA was followed by Fisher's Protected Least Significant Difference multiple comparison method. Statistical analysis for microarray analyses was performed using the software package "limma" [26], for pairwise comparisons (Day 2 post-DSS Vs. no DSS, Day 7 post-DSS Vs. no DSS). P-values for the comparisons of interest were obtained using empirical Bayes method [27] and adjusting for multiple testing [28]. Venn Diagrams were drawn to show the number of genes differentially expressed due to DSS challenge at different time points (days 2 or 7), in different mouse strains (WT or CD4cre) and the genes commonly induced in both strains. A large number of genes found to be differentially expressed on day 7 of DSS challenge in CD4cre mice were further subjected to hypergeometric testing [29] for identifying the enriched KEGG [30] pathways as described in the Gene Set Enrichment Analysis section.

WT or CD4cre mice were challenged with 2.5% DSS for 7 days to induce experimental IBD and to determine the effect of T cell-specific deletion of PPAR γ on colitis severity, gene expression and immune cell distribution. Differences between the groups first appeared on day 4 of the challenge, where CD4cre mice had significantly greater body weight loss despite no significant difference in disease activity (Figure 1A and 1B). The differences in body weights were transient however, and by day 7 body weights were relatively equal between groups.

To more closely examine the effect of T cell-specific PPAR γ deletion on experimental IBD, colons were histologically examined for the presence of inflammatory lesions and scored (Figure 2). Our analysis indicated that colons recovered from CD4cre mice had significantly more epithelial erosion, leukocyte infiltration, and mucosal thickness on day 7 compared to their WT counterparts. As anticipated from a model that initially targets epithelial cells and macrophages, no histological differences were found between WT and CD4cre mice on day 2. The leukocyte infiltration in the colon was polymorphonuclear and lymphoplasmacytic on day 2, but progressed towards a predominantly lymphoplasmacytic infiltrate on day 7 of DSS colitis.

We next characterized the effect of T cell-specific PPAR γ deletion on the distribution of immune cell subsets in the blood, spleen, and MLN by using flow cytometry. In these tissues the CD4cre mice showed a consistent increase in the percent of CD8⁺ T cells and a decrease in the percent of CD4⁺ T cells (Figure 3). The number of F4/80⁺CD11b⁺ monocytes/macrophages in these tissues was unaffected by the deletion of PPAR γ in T cells (data not shown).

We also examined the percentages of regulatory T cells and those expressing IL-10 in these tissues. On day 7, blood from WT mice had significantly more CD4⁺FoxP3⁺ Tregs in comparison to blood from CD4cre mice. While there were no significant differences in the number of CD4⁺FoxP3⁺ cells in MLN (data not shown), the percent of CD4⁺ T cells expressing IL-10 was significantly lower in MLN of CD4cre mice in comparison to WT mice (Figure 3). These data suggest that the deficiency of PPAR γ in T cells impairs the Treg-mediated immunoregulation.

Pairwise comparisons revealed a transcriptional effect of DSS in the colonic mucosa at both time points (Day 2 and 7), with some differences between the two mouse strains. On day 2, there were more genes affected in WT (202) compared to only 8 genes in CD4cre mice (Additional file 1: Supplemental Figure S1). However, the number of genes transcriptionally altered was reduced from 200 to 39 in the WT mice on day 7. On the other hand, CD4cre mice revealed a steep rise in the number of genes differentially expressed in the later time point; from 8 on day 2 to 3036 on day 7 (Additional file 2: Supplemental Figure S2). Since CD4cre and WT mice primarily differ in the presence of T cell PPAR γ, the massive transcriptional modulation on day 7 in CD4cre mice is likely caused by functional changes in T cells derived from the deletion of PPAR γ. Out of the 3036 genes differentially expressed on day 7 in CD4cre mice, some genes were also affected under other conditions with 2990 genes being uniquely modulated on day 7 in CD4cre mice. Additionally, these 2990 genes included leukocyte extravasation markers and pro-inflammatory cytokines described and validated by real-time RT-PCR.

To determine the effect of T cell-specific PPAR γ-deletion on colonic gene expression microarrays were performed from mice on days 0, 2, and 7 of DSS challenge (Figure 4). Leukocyte extravasation markers were significantly enhanced in both WT and CD4cre mice throughout the DSS challenge, though in comparing the two groups we observed that levels were significantly greater in CD4cre mice on day 7. Expression of integrin alpha V (intgav), integrin beta 2 (intgb2), intracellular adhesion molecule (ICAM-1), vascular cell adhesion molecule (VCAM-1), and P-selectin were all significantly enhanced in CD4cre mice compared to the WT mice.

We also examined the mRNA expression of pro-inflammatory cytokines in colons of WT and CD4cre mice. Expression of interleukin 6 (IL-6) and IL-1β were significantly higher in colons from CD4cre mice at day 7 (Figure 4). Levels of cytokine signaling 3 (SOCS-3), an inhibitory protein shown to increase in response to IL-6 that has been shown to be enhanced in both human IBD and murine models of the disease [31‐33], was also significantly upregulated in colons of CD4cre mice. Microarray findings were validated with real time RT-PCR analyses (Figure 5D-H).

We next examined the effect of T cell-specific PPAR γ deletion using gene set enrichment analysis (GSEA) of KEGG pathways. The following three KEGG pathways were significantly differentially regulated as revealed by hypergeometric testing on the 2990 genes differentially expressed on day 7 of DSS treatment compared with no DSS in CD4cre mice: citrate (TCA cycle) or Krebs cycle, apoptosis and ribosomal pathways (Additional file 3: Supplemental Figure S3). Additionally, ribosomal pathway was found to be the most significantly down-regulated pathway from an unbiased global search using all genes (normal Q-Q plot and permutation testing). While many genes of the Krebs cycle and ribosomal pathway were down-regulated, those of apoptosis were upregulated (Additional files 4, 5, 6: Supplemental Figures S4-6 and Additional file 7: Supplemental Table S1). The Krebs cycle pathway is involved in the maintenance of glucose homeostasis, one of the primary roles of PPAR γ. The apoptosis pathway is critically regulated by NF-κB which is repressed by PPAR γ in IBD. However, the pathway most significantly affected by the deletion of PPAR γ was the ribosome pathway (Additional file 6: Supplemental Figure S6), which controls genes involved in protein translation and synthesis.