M2a macrophages facilitate resolution of chemically-induced colitis in TLR4-SNP mice

ABSTRACT Toll-like receptor 4 (TLR4) is an innate immune receptor responsive to lipopolysaccharide (LPS). Single nucleotide polymorphisms (SNPs) in human TLR4 that encode an A896G transition at SNP rs4986790 (D299G) and a C1196T transition at SNP rs4986791 (T399I) render individuals hyporesponsive to LPS. In humans, these SNPs are also associated with increased susceptibility to inflammatory bowel diseases (IBDs). Using knock-in mice engineered to express the murine homologs of these human TLR4 mutations (“TLR4-SNP” mice), we have shown that TLR4-SNP mice develop significantly more severe colitis induced by dextran sodium sulfate (DSS) than wild-type (WT) mice, similar to IBD in humans expressing these SNPs. Previous studies have provided indirect evidence for “tissue repair” M2 macrophages (Mφ) in the resolution of colitis. Signaling through the IL-4/IL-13 receptor, IL-4Rα, and the transcription factor, peroxisome proliferator-activated receptor (PPARγ), have been shown to be required for induction of M2a Mφ, and our data provide direct evidence for the involvement of both in the repair of DSS-induced colonic damage. In response to DSS, colons of TLR4-SNP mice produced reduced levels of M2a Mφ marker mRNA and protein, including PPARγ, and therapeutic administration of the PPARγ agonist ligand, rosiglitazone, resolved colitis in TLR4-SNP mice, and increased expression of the M2a protein, Ym1. Together, these data indicate that the failure of TLR4-SNP mice to resolve DSS-induced colitis may be secondary to their failure to induce “tissue repair” M2a Mφ. Importance Inflammatory bowel disease (IBD), including Crohn’s disease and ulcerative colitis, impacts millions of individuals worldwide and severely impairs the quality of life for patients. Dysregulation of innate immune signaling pathways reduces barrier function and exacerbates disease progression. Macrophage (Mφ) signaling pathways are potential targets for IBD therapies. While multiple treatments are available for IBD, (i) not all patients respond, (ii) responses may diminish over time, and (iii) treatments often have undesirable side effects. Genetic studies have shown that the inheritance of two co-segregating SNPs expressed in the innate immune receptor, TLR4, is associated with human IBD. Mice expressing homologous SNPs (“TLR4-SNP” mice) exhibited more severe colitis than WT mice in a DSS-induced colonic inflammation/repair model. We identified a critical role for M2a “tissue repair” Mφ in the resolution of colitis. Our findings provide insight into potential development of novel therapies targeting Mφ signaling pathways that aim to alleviate the debilitating symptoms experienced by individuals with IBD.

. Genome-wide association studies have identified IBD susceptibility loci that are related to epithelial barrier function and innate and adaptive immunity (38,53,57), emphasizing the importance of genetic predisposition and the likely polygenic control of IBD, including genetic variation in TLRs (58,59).However, the precise mechanism(s) by which the TLR4 299/399 SNPs increase sensitivity to IBD in humans is unknown.
Herein, we show that, like humans with IBD, the TLR4-SNP mice exhibit much more severe colitis in response to DSS than WT mice.Using IL-4Rα −/− mice and mice with a conditional deletion of PPARγ in myeloid cells (PPARγ cKO ), that fail to respond normally to IL-4 or IL-13 to induce Mφ with an M2a "repair" phenotype, we show that both IL-4Rαand PPARγ-dependent signaling are required for amelioration of DSS-induced colonic tissue injury.DSS-treated TLR4-SNP mice exhibited a reduced capacity to express M2a Mφ markers, and their colitic disease was ameliorated by therapeutic treatment with a PPARγ agonist ligand.
TLR4-SNP mice developed more severe colitis than WT mice throughout the experi mental period as measured by "colonic symptom score" (Fig. 1B).When the responses of male and female TLR4-SNP mice were directly compared, male TLR4-SNP mice exhibited both enhanced symptom scores from Days 0 to 7 and a reduced capacity for recovery from Days 8 to 14 (Fig. 1C, left panel) compared to female TLR4-SNP mice (Fig. 1C, right panel; significant differences between males and females are indicated by #).Human IBD and DSS-induced colitis are associated with increased colonic permeability (55,76,(79)(80)(81)(82)(83)(84)(85)(86).Relative colonic permeability, assessed by detecting serum fluorescence in mice 4 h after oral administration of 4 kDa fluorescein isothiocyanate (FITC)-dextran, revealed no differences in permeability between control WT and TLR4-SNP mice (Fig. 1D).By Day 7, both strains exhibited increased serum FITC-dextran levels, with TLR4-SNP mice trending toward significantly greater colonic permeability compared to WT mice (Fig. 1D).By Day 14, WT and TLR4-SNP mice exhibited reduced serum levels of FITC-dextran compared to Day 7 levels; however, TLR4-SNP mice had significantly higher serum fluorescence than WT mice (Fig. 1D), suggesting that TLR4-SNP mice have an impaired ability to repair colonic damage.In addition, at Days 7 and 14, TLR4-SNP mice exhibited significantly shorter colon lengths, indicative of an increased inflammatory state, compared to WT mice (Fig. 1E, left and center panels); however, no sex differences in colon length were observed (Fig. 1E, right panel).
Disease severity was also assessed in H&E-stained micrographs of the distal 1 cm colon (DC) of WT and TLR4-SNP mice.No significant differences were observed between untreated WT and TLR4-SNP mice (Fig. 2A, top row).On Day 7, DSS-treated WT mice exhibited few inflammatory infiltrates in the mucosa and submucosal regions compared to non-DSS-treated control mice and the crypt structure remained largely intact (Fig. 2A, middle row).In contrast, by Day 7, DSS-treated TLR4-SNP mice displayed disrupted crypt structures with inflammatory infiltrates in the mucosa and submucosal regions, accompanied by epithelial damage (Fig. 2A, middle row).By Day 14, the crypt structures of DSS-treated TLR4-SNP mice were severely disrupted by the extensive inflammatory infiltrates in the mucosa (Fig. 2A, bottom row; red line) and submucosal (Fig. 2A, bottom row; red arrowhead) regions in contrast to WT mice.These histopathological differences between WT and TLR4-SNP mice were more pronounced in male mice (Fig. 2A, left panels) and confirmed by histologic scoring (Fig. 2B) and measurements of the muscula ris externa, submucosal height, and mucosal depth (Fig. 2C).The notable colonic shortening in DSS-treated TLR4-SNP mice that was observed grossly at Day 14 (Fig. 1E) correlated with broadening of the muscularis mucosa (Fig. 2C, left panel).Moreover, inflammatory infiltrates accumulated and contributed to edema in the submucosa (Fig. 2A, bottom row) which correlated with increased submucosal height (Fig. 2C, center panel), while observed crypt regeneration in response to damage was evidenced by increased mucosal depth (Fig. 2C, right panel).At higher magnification, Day 14 DSStreated TLR4-SNP mice exhibited substantially greater cellular infiltration into the colonic mucosa and submucosa than WT mice (Fig. 2D), with infiltrating cells that were predomi nantly monocytic (Fig. 2E).Collectively, our observations of worse colonic symptoms, reduced colon lengths, and more severe histopathology in TLR4-SNP mice indicate an impaired ability to recover from DSS-induced colonic injury.These findings prompted us to investigate possible mechanism(s) underlying the lack of recovery in the TLR4-SNP mice.

IL-4Rα-dependent signaling is necessary for tissue recovery from DSSinduced colonic damage
There are many studies that indicate that the microbiome is important in DSS-induced colitis (87)(88)(89)(90)(91)(92).However, we initially carried out co-housing experiments in which WT and TLR4-SNP mice were co-housed for 3 weeks prior to DSS treatment and throughout the entire experiment to normalize their microbiomes (93,94).Nonetheless, the TLR4-SNP mice remained much more sensitive to DSS (data not shown), indicating that the increased sensitivity of TLR4-SNP mice to DSS was not likely attributable to dysbiosis.
We first confirmed that IL-4Rα −/− mice do not express IL-4Rα (Fig. 3A) by measuring Il4ra mRNA in colons of WT vs IL-4Rα −/− mice.WT mice displayed unaltered Il4ra mRNA levels in response to DSS treatment, while Il4ra mRNA levels in IL-4Rα −/− colons were at or below the limit of detection.Following the tissue injury phase, IL-4Rα −/− mice exhibited an impaired ability to recover from DSS-induced colonic injury compared to WT mice evidenced by statistically increased colonic symptom scores beginning at Day 10 (Fig. 3B, left panel) with males being more sensitive than females (Fig. 3B, right panel).By Day 14, the IL-4Rα −/− mice exhibited significantly shorter colons than WT mice (Fig. 3C) and increased pathology (Fig. 3D), which was confirmed by histologic scoring (Fig. 3E).Few inflammatory infiltrates were seen in the mucosal and submucosal regions in DSStreated WT mice at Day 14 compared to non-DSS treated controls and the crypt structure remained intact in contrast to the IL-4Rα −/− mice (Fig. 3D).At Day 14, DSS-treated IL-4Rα −/− mice exhibited visibly disrupted crypts, with abundant inflammatory infiltrates in the mucosal and submucosal regions and extensive epithelial damage (Fig. 3D).No sex differences were observed for colon length or histology scores (Fig. 3C and E).These findings are the first to directly support the conclusion that IL-4-and/or IL-13-induced signaling via IL-4Rα is necessary for tissue recovery from DSS-induced damage.

TLR4-SNP mice exhibit a reduced capacity to induce M2a Mφ in response to DSS
Since IL-4Rα signaling was required for the repair of DSS-induced colitis, we next sought to determine if the worsened colitis of DSS-treated TLR4-SNP mice was attributable to a decreased capacity for differentiation of M2a Mφ.Since the repair phase occurs between Days 8 and 14, Days 9 and 11 were selected to measure the induction of M2a Mφ markers in the colons of WT and TLR4-SNP mice.We compared the induction of M2a genes in colon sections of control and DSS-treated mice.M2a Mφ markers, Arg1, and Chil3 mRNA levels were significantly reduced in TLR4-SNP colons at Days 9 and 11, while Mrc1 mRNA showed a trend toward reduced M2a gene expression in TLR4-SNP colon sections compared to WT mice, particularly at Day 9 (Fig. 4A).These observations were confirmed by Western analysis showing reduced Arginase 1 (Arg1), Ym1 (Chil3) (Fig. 4B), and Mrc1 (Fig. 4C) protein levels in TLR4-SNP colon homogenates at Day 11.These findings indicate that TLR4-SNP mice have an impaired ability to induce M2a Mφ markers in response to DSS.

The PPARγ agonist ligand, rosiglitazone, improved DSS-induced colitis in TLR4-SNP mice therapeutically
To confirm the role of PPARγ in DSS-induced colitis, we utilized PPARγ conditional knockout (PPARγ cKO ) mice originally derived by crossing homozygous floxed PPARγ mice with a transgenic mouse containing the Cre recombinase gene under the control of the murine Lysozyme M (Lyz2) promoter to delete the Pparg gene in lysozyme-pro ducing cells (i.e., predominantly Mφ and neutrophils) (133,(147)(148)(149).A lack of PPARγ protein expression in medium-or IL-4-stimulated, PPARγ cKO vs WT thioglycollate-elicited peritoneal Mφ was confirmed by Western blot (Fig. 5A).Although DSS-treated PPARγ cKO mice presented on Days 2 and 4 with milder colonic symptoms compared to DSS-trea ted WT mice, their disease severity significantly increased between Days 10 and 14, suggesting an impaired ability to recover (Fig. 5B).By Day 14, PPARγ cKO mice exhibited significantly shorter colon lengths (Fig. 5C) and increased histopathology (Fig. 5D) than WT mice, phenotypes also observed in DSS-treated TLR4-SNP and IL-4Rα −/− mice.No sex differences were observed for any of the parameters measured.Thiazolidinedione (TZD) agents, such as rosiglitazone, are PPARγ agonist ligands shown previously to ameliorate DSS-induced colitis in mice (139, 150-154) and 2,4,6trinitrobenzene sulfonic acid (TNBS)-induced colitis in rats (155).TLR4-SNP mice were administered 3% DSS in their drinking water starting on Day 0 and then treated thera peutically with saline or rosiglitazone (once daily from Day 2 to Day 7).On Days 8 to 14, all mice were given regular drinking water.Compared to DSS + saline-treated TLR4-SNP mice, TLR4-SNP mice treated with DSS + rosiglitazone displayed reduced clinical symptoms (Fig. 6A), increased colon lengths at Day 14 (Fig. 6B), repaired crypt structure, and reduced numbers of infiltrating inflammatory cells (Fig. 6C).Under high power, the infiltrating cells in the DSS + saline-treated TLR4-SNP mice were predominantly mononu clear with a few granulocytic cells (indicated with red arrows; Fig. 6D).Histopathological changes were reflected in the histology scores (Fig. 6E).Consistent with a role for M2a Mϕ in the resolution of DSS-induced colitis, the rosiglitazone-treated TLR4-SNP mice exhibited increased expression of the M2a protein, Ym1 (Chil3), compared to the salinetreated group (Fig. 6F).These findings indicate that therapeutic treatment with the PPARγ agonist ligand improves DSS-induced colitis symptoms in the hypersensitive TLR4-SNP mice.
Since the human D299G/T399I mutations in TLR4 are co-inherited by a significant proportion of individuals (33), "TLR4-SNP" mice were engineered to enable mechanistic studies by which these TLR4 SNPs impact diseases.To date, these mice have been shown to reflect phenotypes in humans, including LPS hyporesponsiveness, increased sensitivity to Gram negative infection and RSV, and increased resistance to influenza infection (14).In addition, the TLR4-SNP mice exhibit reduced allergic inflammation with LPS and OVA but increased inflammation, eosinophilic infiltration, and Th2 cytokines following house dust mite allergen induction, reflecting a significant effect of the environment on their Th2-dependent responses (160).Further, TLR4-SNP mice show reduced antibody responses and a deficiency in class switching (also Th2-dependent) in response to immunization to T-dependent antigen when adjuvanted with a lipid A-based adjuvant (161).
Although DSS is widely used to induce colitic disease in mice, we recognize that this model has limitations (162).DSS is toxic to epithelial cells resulting in loss of barrier integrity by redistribution of the tight junction components, increased permeability, increased epithelial apoptosis and proinflammatory cytokines, and entry of luminal antigens and microorganisms that induce inflammation (162).The pathogenesis of IBD is complex, and many mechanisms have been implicated (51)(52)(53).In healthy individuals, multiple cellular programs coordinate to effectively promote healing of damaged tissues.One cell type that is particularly important for healing is the alternatively activated or repair "M2" Mφ.
Mφ mediate innate immune responses to PAMPs and DAMPs and play a key role in tissue repair and regeneration (163)(164)(165).The presence of cytokines and other environ mental signals promotes Mφ polarization to a variety of phenotypes.Mφ are often described in terms of a "differentiation spectrum, " with "classically activated" "M1" Mφ, induced by potent inflammatory stimuli such as LPS and IFN-γ, at one end of the differentiation spectrum, and "alternatively activated" or "tissue repair" (166) "M2" Mφ at the other end of the spectrum (113).While M2 Mφ differentiation was originally attributed to signaling by IL-4 and IL-13 through a shared IL-4Rα chain to activate M2-associated genes (116,166), the definition of "M2" Mφ is now far more nuanced and subsets have been defined (M2a, M2b, M2c, M2d, regulatory) based on distinct inducing agents and transcriptomic patterns (112,114).
By using IL-4Rα −/− mice, our study has provided the first direct experimental evidence for the role of IL-4/13 signaling in the repair of DSS-induced colitis in WT mice (Fig. 3).IL-4 and IL-13 signaling through a shared receptor, IL-4Rα, leads to activation of specific transcription factors, e.g., STAT-6, EGR-2 (112,167), that, in turn, recruit the transcription factor PPARγ to activate promoters of specific "M2a" genes (112, 130, 167, 168).We extended our findings in the IL-4Rα −/− mice by showing that mice in which PPARγ had been deleted in myeloid cells (133,(147)(148)(149) were also extremely sensitive to DSS-induced colitis (Fig. 5).We recognize that LysMCre deletion is not absolute and that PPARγ on other cell types (e.g., epithelial cells) remains functional and may even compensate to some degree for the lack of PPARγ on myeloid cells.Nonetheless, like DSS-treated IL-4Rα −/− and TLR4-SNP mice, the PPARγ cKO mice phenocopy the impaired repair responses to DSS.When induction of M2a-related mRNA and protein species were measured at Days 9 and 11, colons of DSS-treated TLR4-SNP mice exhibited reduced levels of M2a Mφ RNA and proteins, including PPARγ protein (Fig. 4A through C).These findings are consistent with previous studies showing that in contrast to WT Mφ TLR4 −/− Mφ failed to induce Pparg mRNA in response to either IL-4 stimulation or RSV infection (135).Treatment of male and female, WT and TLR4-SNP thioglycollate-induced, peritoneal Mφ with exogenous IL-4 resulted in no significant difference in the levels of Arg1 and Chil3 mRNA, indicating that TLR4-SNP Mφ do not have an intrinsic defect in the ability to respond through IL-4Rα (data not shown).
PPARγ heterodimerizes with the NHR, RXR, to activate transcription of many M2a genes (e.g., Arg1, Mrc1, Chil3) (112,130,167), and transcriptional activity can be further modulated by endogenous and exogenous agonist ligands that bind to each compo nent of the heterodimer (169).Many studies have investigated the therapeutic effect of ligand activation of the PPARs in IBD (141,150,(170)(171)(172)(173), including human clinical trials (174)(175)(176) and studies (177).TZDs are a class of structurally related PPARγ ligands originally designed as antidiabetic drugs.In particular, rosiglitazone acts as a highly specific ligand agonist of PPARγ (178,179) that enhances M2a Mφ polarization (130,180,181), reduces DSS-induced inflammation in vivo (150), and leads to overall anti-inflammatory effects (182).Rosiglitazone showed promise in a clinical trial for treatment of active UC (175) but is no longer prescribed due to cardiotoxicity (183).Rosiglitazone therapy resolved lung histology and enhanced the expression of M2a Mφ in rodent models of RSV-induced lung injury (135).Another structurally related TZD, pioglitazone, showed similar outcomes in influenza infection (184).We hypothesized that therapeutic administration of rosiglitazone would also contribute to the resolution of DSS-induced colitis in the TLR4-SNP mice.Indeed, rosiglitazone effectively reversed DSS-induced colonic damage in the TLR4-SNP mice (Fig. 6), further supporting the conclusion that PPARγ may be an effective therapeutic target in patients that express the D299G and T399I TLR4 SNPs.Importantly, recent studies suggest that it is possible to develop PPARγ ligands that dissociate one cell-specific PPARγ function from another (185), offering the possibility of being able to identify a new PPARγ agonist ligand that will ameliorate IBD but not exert cardiotoxicity.
Since the microbiome has been strongly implicated in IBD pathogenesis, future studies will be required to determine whether the microbiomes of WT and TLR4-SNP mice differ and whether potential microbial differences affect the observed phenotypic responses to DSS in the two strains.Since β-HB is a metabolite associated with Gramnegative bacteria, it is possible that bacterial subsets may be reduced or missing in the DSS-treated TLR4-SNP mice such that β-HB is produced to a lesser extent, thereby reducing M2a Mφ differentiation.
In conclusion, TLR4-SNP mice develop more severe DSS-colitis symptoms compared to WT mice and exhibit a reduced capacity to repair colonic damage.IL-4Rα and myeloid PPARγ are necessary for induction of M2a Mφ and we show herein that IL-4Rα and PPARγ (on myeloid cells) are necessary to repair DSS-induced colonic damage.DSS-treated TLR4-SNP mice exhibited reduced M2a Mφ markers including PPARγ but responded to rosiglitazone, providing a potential avenue for future therapeutic development.In the context of personalized medicine, we anticipate that genotyping of IBD patients for the inheritance of TLR4 SNPs may ultimately lead to the identification of a patient subset for which M2a-inducing therapies are likely to be most effective.

DSS-induced colitis; measurements of colonic and histologic scores
Control mice received standard hyperchlorinated drinking water from Days 0 to 14. Experimental mice received 3% DSS (CAS 9011-18-1, MP Biomedicals 160110) in drinking water on Days 0-7 (changed on alternate days) (75,77).Each new lot of DSS was titrated to ensure equivalent potency.On Day 8, all mice were provided regular drinking water only to allow a period of recovery through Day 14.A "colonic symptom score" was measured on alternate days by combining the stool consistency score (0 = normal; 1 = soft; 2 = very soft; 3 = diarrhea) and the fecal occult blood score [0 = no blood (−occult); 2 = microscopic blood in stool (+occult); 3 = macroscopic visible blood in stool (+occult); 4 = gross rectal bleeding] (75, 77).Fecal occult blood was measured using the Fisher Healthcare Sure-Vue Fecal Occult Blood Test (Fisher Scientific).At the indicated times, colons were resected, cleared of feces, and measured in centimeters distally from the cecum to just prior to the rectum.For most experiments, the most distal 1 cm colon section from control and experimental mice was opened longitudi nally and fixed in 4% paraformaldehyde for 2 h at room temperature and sent to the University of Maryland School of Medicine's Pathology Histology Core or to the University of Maryland School of Medicine's and Greenebaum Comprehensive Cancer Center's Pathology Biorepository Shared Services (housed in the Center for Innovative Biomedical Research (CIBR)) for sectioning and/or H&E-staining.At least two sections per mouse and >3 areas per section were analyzed using the ECHO Revolve microscope under 10× magnification for histology scoring or 100× magnification for high-power cellular counts.The total histology score represents the sum of epithelial damage (0-3), mucosal inflammatory infiltrate (0-3), submucosal inflammatory infiltrate (0-3), and muscularis externa inflammatory infiltrate (0-3), each multiplied by the respective extent of damage (1-3) (77).Under high power, the number of monocytic cells and the number of granulocytic cells were recorded in >6 high-power fields per mouse.In some experiments, the DC was flash frozen and subsequently homogenized in 0.5 or 1-mL PBS for Western analysis.The next most distal 1 cm of the colon (medial colon, MC) was flash frozen and used for RNA extraction for qRT-PCR.

Measurement of colonic permeability
On Days 6 or 13, control or DSS-treated mice were fasted overnight before administration of 0.6 mg/g FITC-dextran in PBS (CAS 60842-46-8, Sigma-Aldrich) by oral gavage on Days 7 and 14.Mice were bled 4 h after FITC-dextran administration.Fluorescence was measured in the serum of each mouse (diluted 1:2 in 1× PBS) and compared to a standard curve (0-8,000 ng/mL FITC-dextran) in a black well plate using a fluorescent spectrophotometer (Thermo Electron Corporation Fluoroskan Ascent FL) with 485 nm excitation and 527 nm emission (186).

Gene expression analysis
Flash-frozen colonic tissue taken from the MC was homogenized in TriPure (Roche) and RNA extracted using the manufacturer's protocol.Since DSS has been reported to interfere with the reverse transcriptase reaction (187), RNA was purified using the LiCl purification method (188).cDNA was synthesized by reverse transcription reaction from 1 µg RNA per sample using the qScript cDNA synthesis kit (Quantabio).The 7500 Fast Real-Time PCR System and software (Applied Biosystems) was used to perform qRT-PCR (189).Forward and reverse primers included those corresponding to the housekeeping gene, Hprt, and M2a markers, Chil3, Arg1, Mrc1, and Pparg (Sigma) (Table S1).Primers were designed to cross intron-exon boundaries, and sequences were confirmed to target the gene of interest by NCBI BLAST.Sample amplification was verified using a negative control containing no cDNA.mRNA levels were normalized to relative Hprt mRNA levels and were reported using the −ΔCt method (14,190).

Measurement of β-HB
Serum from each mouse was collected at the indicated times and filtered using 10 kDa centrifugal filter units (Millipore).β-HB was measured using the β-HB (ketone body) colorimetric assay kit (Cayman Chemical; 700190).β-HB standards were prepared according to the manufacturer's protocol.Standards and samples (50 µL) were added to the plate in duplicate.Developer solution (50 µL) was added to each well, and the plate was incubated in the dark for 30 min at 25°C.Absorbance was measured at 450 nm using a ELx808 (BioTek) plate reader.

Isolation and cultivation of peritoneal Mφ
Thioglycollate-elicited Mφ (192) were treated for 24 h with medium or IL-4 (5 ng/mL) (184).Supernatants from lysates of the cells were analyzed by Western blot analysis.

Statistical analysis
Mean values ± standard error of the mean are shown.Comparisons among more than two groups were analyzed using a two-way ANOVA with Sidak's multiple comparisons test (GraphPad Prism 9) unless otherwise noted.Comparisons among exactly two groups were analyzed using an unpaired two-tailed Student's t-test (GraphPad Prism 9).*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.The number and sex of mice per treatment are indicated in the figure legends, and individual data points are presented from the indicated number of experiments.Biostatistical services were provided by CIBR (UMB).

FIG 5
FIG 5 PPARγ cKO mice are more sensitive to DSS-induced colitis than WT mice.(A) PPARγ protein expression in thioglycollate-elicited macrophages from WT and PPARγ cKO mice in the absence or presence of IL-4 (24 h) to confirm genotype.(B) Colonic symptom scores (Mean ± SEM) in control and DSS-treated WT vs PPARγ cKO mice (responses of male and female mice combined).Two-way ANOVA with Sidak's post-hoc multiple comparisons test.*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0001.(C) Representative colon lengths at Day 14 in control and DSS-treated WT vs PPARγ cKO mice.Left panel, gross image of representative DSS-treated WT and PPARγ cKO colons on Day 14.Right panel, male and female responses combined.Two-way ANOVA with Sidak's post-hoc multiple comparisons test.****P < 0001.(D) Left panel: representative H&E-stained images of DC sections.Right panel: histology scores of H&E-stained DC sections from control and DSS-treated WT and PPARγ cKO mice on Day 14, males and females combined.Two-way ANOVA with Sidak post-hoc multiple comparisons test.****P < 0001.Results were derived from three independent experiments in which in total control WT (n = 5), control PPARγ cKO (n = 5), 3% DSS WT (n = 16 males), 3% DSS PPARγ cKO (n = 12;8 males).Although the early kinetics of DSS-induced disease were delayed in PPARγ cKO compared to WT mice, we observed deaths in four males and one female, suggesting that myeloid PPARγ plays a key role in tissue repair.