Blocking integrin α₄β₇-mediated CD4 T cell recruitment to the intestine and liver protects mice from western diet-induced non-alcoholic steatohepatitis

Abstract

Background & Aims:

The heterodimeric integrin receptor α₄β₇ regulates CD4 T cell recruitment to inflamed tissues, but its role in the pathogenesis of nonalcoholic steatohepatitis (NASH) is unknown. Here we examined the role of α₄β₇-mediated recruitment of CD4 T cells to the intestine and liver in NASH.

Methods:

Male littermate F11r^+/+ (control) and junctional adhesion molecule A knockout F11r^–/– mice were fed a normal diet or a Western diet (WD) for eight weeks. Liver and intestinal tissues were analyzed by histology, qRT-PCR, 16s rRNA sequencing and flow cytometry. Colonic mucosa-associated microbiota was analyzed using 16s rRNA sequencing. Liver biopsies from NASH patients were analyzed by confocal imaging and qRT-PCR.

Results:

WD-fed knockout mice developed NASH and had increased hepatic and intestinal α₄β₇⁺ CD4 T cells relative to control mice which developed mild hepatic steatosis. The increase in α₄β₇⁺ CD4 T cells was associated with markedly higher expression of the α₄β₇ ligand mucosal addressin cell adhesion molecule 1 (MAdCAM-1) in the colonic mucosa and livers of WD-fed knockout mice. Elevated MAdCAM-1 expression correlated with increased mucosa-associated Proteobacteria in the WD-fed knockout mice. Antibiotics reduced MAdCAM-1 expression indicating that the diet-altered microbiota promoted colonic and hepatic MAdCAM-1 expression. α₄β₇ blockade in WD-fed knockout mice significantly decreased α₄β₇⁺ CD4 T cell recruitment to the intestine and liver, attenuated hepatic inflammation and fibrosis, and improved metabolic indices. MAdCAM-1 blockade also reduced hepatic inflammation and fibrosis in WD-fed knockout mice. Hepatic MAdCAM-1 expression was elevated in NASH patients and correlated with higher expression of α₄ and β₇ integrins.

Conclusions:

These findings establish α₄β₇/MAdCAM-1 as a critical axis regulating NASH development through colonic and hepatic CD4 T cell recruitment.

Lay Summary:

Nonalcoholic steatohepatitis (NASH) is an advanced and progressive form of nonalcoholic fatty liver disease (NAFLD), and despite its growing incidence no therapies currently exist to halt NAFLD progression. Here, we show that blocking integrin receptor α₄β₇-mediated recruitment of CD4 T cells to the intestine and liver not only attenuates hepatic inflammation and fibrosis, but also improves metabolic derangements associated with NASH. These findings provide evidence for potential therapeutic application of α₄β₇ antibody in the treatment of human NASH.

Graphical Abstract

INTRODUCTION

Nonalcoholic steatohepatitis (NASH) is an advanced and progressive form of nonalcoholic fatty liver disease (NAFLD), afflicting 50 million people worldwide and over 16 million people in North America [1]. An estimated one in six patients with NASH progress to cirrhosis and recent evidence links NASH to a higher incidence of hepatocellular carcinoma [2]. While a Western diet (WD) is likely a major risk factor for NASH development, the progressive nature of the disease suggests that other ongoing biological insults or “second hits” play a significant and complex role in NASH etiology. Increased intestinal permeability has been reported in human NAFLD [3–5], and we have demonstrated that humans with NAFLD and no known inflammatory bowel disease have mucosal inflammation and decreased expression of the gene F11r encoding the intestinal tight junction (TJ) protein junctional adhesion molecule A in the colonic mucosa [6]. In the F11r knockout mouse model, feeding studies revealed development of severe NASH with significant fibrosis within eight weeks of initiating a WD. WD consumption further potentiated intestinal epithelial permeability in F11r^−/− mice secondary to mucosal inflammation triggered by gut dysbiosis [6]. Loss of the intestinal epithelial barrier in WD-fed F11r^−/− mice resulted in the translocation of gut bacterial products which facilitated rapid progression of NAFLD by driving hepatic inflammation and fibrosis [6]. Depletion of gut microbiota using a broad-spectrum antibiotic cocktail not only attenuated diet-induced hepatic inflammation and fibrosis but also improved metabolic derangements demonstrating a key role for gut microbiota in NASH development [6]. These data and other published work in animal models as well as human studies underscore the importance of intestinal epithelial permeability in NASH development and implicate diet-induced mucosal inflammation as a major contributor to disease progression [3–6]. However, the mechanisms underlying immune mediated mucosal inflammation in NASH have not yet been clearly defined.

The recruitment of immune cells to tissues is a highly regulated process that is essential for immune homeostasis and the resolution of inflammation following injury or infection. Immune cell recruitment is orchestrated by adhesion molecules expressed on homing lymphocytes and their corresponding ligands expressed by endothelial cells [7]. The heterodimeric integrin receptor α₄β₇ regulates intestinal T cell recruitment via interaction with its ligand MAdCAM-1 constitutively expressed by high endothelial venules and the intestinal lamina propria [7]. This process is essential for maintaining intestinal epithelial barrier function, as dysregulated CD4 T cell recruitment to the intestine can underlie mucosal inflammation leading to loss of intestinal barrier function [7–9]. While the α₄β₇/MAdCAM-1 axis is implicated in promoting hepatic inflammation in other chronic liver diseases [10, 11], the contribution of this axis in mucosal and hepatic inflammation in NASH is not known.

In this study, using our recently characterized mouse model of NASH, WD-fed F11r^−/− mice [6], we provide evidence for the first time that α₄β₇-mediated homing of CD4 T cells to the intestine and liver promote NASH development. We demonstrate that WD promotes a systemic increase in α₄β₇⁺ CD4 T cells. These cells are recruited to the intestine and liver by a microbiota-driven increase in colonic and hepatic MAdCAM-1 expression. We show that blocking α₄β₇⁺ CD4 T cell recruitment to the intestine and liver attenuates hepatic inflammation and fibrosis and improves indices of the metabolic syndrome (MetS). Analysis of liver tissue from NASH patients revealed higher expression of MAdCAM-1 which correlated with elevated expression of integrins α₄, ITGA4 and β₇, ITGB7. Together, our results indicate that blockade of α₄β₇ or MAdCAM-1 may represent a novel therapy for treating NASH patients.

METHODS

Mice.

Junctional adhesion molecule A (JAM-A) knock out mice (F11r^−/−) were generated as previously described [6]. Mice were bred and maintained at Emory University and the University of Pittsburgh Division of Animal Resources. All animal studies were approved by Institutional Animal Care and Use Committee.

Human tissue.

Human liver tissues were obtained from Sekisui-XenoTech, LLC (Kansas City, KS) and Biospecimen Processing and Repository Core at Pittsburgh Liver Research Centre. This study was institutional review board exempt as the human tissues used in this study were de-identified, fixed human tissue.

NASH diet:

The western diet (WD) consisted of 0.2% cholesterol, 20% protein, 43% CHO, 23% fat (6.6% trans-fat), and 2.31% fructose (TD.130885; Harlan Laboratories) [6]. The normal diet (ND) is the standard mouse chow that contains 16% protein, 61% carbohydrate and 7.2% fat. Five to six-week-old male mice were fed the WD ad libitum for eight-weeks to induce NASH. Littermate F11r^+/+ mice served as controls for all experiments.

Statistical analysis.

Statistical differences between multiple groups were analyzed by ANOVA and post hoc analysis for multiple group comparison. For α₄β₇ and MAdCAM-1 blocking experiments, statistical differences between groups were analyzed by Fisher’s exact test (two-sided) with post hoc analysis for pairwise comparison. A p value < 0.05 was considered statistically significant. Data shown are representative of three independent experiments. All statistics were performed using GraphPad Prism 8.0 (GraphPad Software) or R software.

RESULTS

WD consumption increases “gut tropic” CD4 T cells in the peripheral blood.

As previously reported [6], WD-fed F11r^−/− mice developed NASH compared to findings of only mild steatosis in control F11r^+/+ mice (Supp. Fig. 1). Increased CD4 T cells in blood can reflect heightened CD4 T cell homing to inflamed tissues [12], and increased proinflammatory CD4 T cells have been reported in the peripheral blood and liver of NASH patients [13–16]. We therefore first determined if higher numbers of CD4 T cells were present in the peripheral blood of WD-fed mice. Flow cytometric analysis of the peripheral blood mononuclear cells (PBMC) from control F11r^+/+ and F11r^−/− mice following eight-weeks of ND or WD-feeding revealed a significant increase in total CD4 T cells in the absence of changes to relative CD4 T cell frequencies in the WD fed F11r^−/− mice (Fig. 1A–B; Supp. Fig. 2A). Since the integrin receptor α₄β₇ directs CD4 T cell trafficking to the gut [7], we next examined peripheral blood CD4 T cells for expression of integrin receptor α₄β₇. As shown in Fig. 1C and Supp. Fig. 2B, a significantly higher percentage and total number of CD4 T cells in WD-fed mice were positive for the expression of α₄β₇. The frequency and the total number of α₄β₇⁺ CD4 T cells were higher in the WD-fed F11r^−/− mice, compared to controls, but the difference in frequency did not reach statistical significance. Similarly, increased frequency of α₄β₇⁺ CD4 T cells, in the absence of changes in relative CD4 T cell numbers, was observed in the spleen of WD fed controls and F11r^−/− mice (Supp. Fig. 3A–D). Together, these findings suggest that WD consumption results in a systemic increase in gut-tropic α₄β₇⁺ CD4 T cells implicating a role of α₄β₇⁺ CD4 T cells in NAFLD.

Figure 1. Western diet increases integrin α₄β₇⁺ CD4 T cells in the peripheral blood.

(A) Gating strategy for the isolation of CD4 T cells. Plots progressively gated on lymphocytes, singlets, live, CD3+, CD8+ T cells. (B) Representative flow plots show percent of CD4 T cells, and scatter plots show total number and percentage of CD4 T cells in peripheral blood of F11r^+/+ and F11r^−/− mice fed a normal diet (ND) or a western diet (WD) for eight-weeks (n = 5 – 7 mice per group). (C) Representative flow plots show percent of α₄β₇⁺ CD4 T cells, and scatter plots show total number and percentage of α₄β₇⁺ CD4 T cells in peripheral blood (n = 5 – 7 mice per group). Data are presented as mean ± SEM. Asterisks indicate significant differences (p < 0.05) between F11r^+/+ and F11r^−/− mice fed an identical diet. Hashtags indicate significant differences (p < 0.05) between ND or WD-fed F11r^+/+ and F11r^−/− mice.

WD consumption increases intestinal recruitment of α₄β₇⁺ CD4 T cells.

Having established that α₄β₇⁺ CD4 T cells are increased in the peripheral blood, we next determined whether WD consumption similarly increased intestinal α₄β₇⁺ CD4 T cells. No differences in the total number and the frequency of CD4 T cells in Peyer’s Patches (PP) were observed between ND-fed control and F11r^−/− mice (Fig. 2A). However, WD feeding significantly increased the frequency and total number of CD4 T cells in PP of both control and F11r^−/− mice with significantly higher numbers of CD4 T cells in F11r^−/− mice (Fig. 2A). In agreement with the PBMC data, WD increased the total number and percentage of α₄β₇⁺ CD4 T cells in the PP of both control and F11r^−/− mice, but the total number of α₄β₇⁺ CD4 T cells were significantly higher in the F11r^−/− mice as compared to control mice (Fig. 2B). WD feeding significantly increased the frequency and total number of CD4 T cells in the colonic lamina propria (CLP) of control and F11r^−/− mice with significantly higher numbers of CD4 T cells in F11r^−/− mice (Fig. 2C). A significantly higher frequency and the total number of CD4 T cells in the CLP of WD-fed F11r^−/− mice relative to control mice are α₄β₇⁺ (Fig. 2D). WD feeding also increased the total number, but not the frequency of α₄β₇⁺ CD4 T cells in the CLP of control mice relative to ND fed mice (Fig. 2D). Increase infiltration of α₄β₇⁺ CD4 T cells in the colonic mucosa was further confirmed by laser confocal imaging demonstrating higher numbers of α₄β₇ and CD4 double positive cells in the colonic mucosa of WD-fed F11r^−/− mice relative to controls (Fig. 2E). Together these data imply that WD promotes recruitment of α₄β₇⁺ CD4 T cells to the intestine.

Figure 2. Western diet increases α₄β₇⁺ CD4 T cell recruitment to the intestine.

(A-B) Scatter plots show total number and percentage of (A) CD4 T cells (n = 5 −12 mice per group) and (B) α₄β₇⁺ CD4 T cells (n = 5 – 8 mice per group) in Peyer’s patches (PP) of F11r^+/+ and F11r^−/− mice fed a normal diet (ND) or western diet (WD) for eight-weeks. (C-D) Scatter plots show total number and percentage of (C) CD4 T cells and (D) α₄β₇⁺ CD4 T cells in the colonic lamina propria (LP; n = 4 – 6 mice per group). Data are presented as mean ± SEM. Asterisks indicate significant differences (p < 0.05) between F11r^+/+ and F11r^−/− mice fed an identical diet. Hashtags indicate significant differences (p < 0.05) between ND or WD-fed F11r^+/+ and F11r^−/− mice. (E) Representative confocal images of CD4 (green) and α₄β₇ (red) immunofluorescence in the colonic mucosa of F11r^−/− mice fed a WD for eight weeks (n = 5 mice per group). Nuclei are stained blue. Scale bar 100 μm.

WD consumption increases microbiota-dependent colonic MAdCAM-1 expression.

While WD feeding resulted in a systemic increase in α₄β₇⁺ CD4 T cells in the control and F11r^−/− mice, recruitment of these cells was higher in the colonic mucosa of WD-fed F11r^−/− mice. Since endothelial MAdCAM-1 expression drives α₄β₇ mediated homing of T cells to the colonic mucosa [7], we examined colonic MAdCAM-1 expression in controls and F11r^−/− mice fed the WD. No difference in colonic MAdCAM-1 expression was detected between ND-fed controls and F11r^−/− mice (Supp. Fig. 4A). In contrast, WD feeding increased colonic MAdCAM-1 expression in control and F11r^−/− mice, but expression was markedly higher in the F11r^−/− mice (Fig. 3A, C). Higher colonic MAdCAM-1 expression in the WD-fed F11r^−/− mice correlated with increased infiltration of α₄β₇⁺ CD4 T cells (Fig. 2) suggesting that α₄β₇/MAdCAM-1 dependent intestinal recruitment of CD4 T cells may promote mucosal inflammation and subsequent loss of barrier function in WD-fed F11/r^−/− mice [6]. Since antibiotic treatment attenuates hepatic inflammation and fibrosis secondary to a marked reduction in colonic inflammation in WD-fed F11r^−/− mice [6], we examined MAdCAM-1 expression in the colonic tissue from antibiotic treated WD-fed F11r^−/− mice. Antibiotic treatment markedly reduced colonic MAdCAM-1 expression suggesting a role of microbiota in regulating colonic MAdCAM-1 expression in NASH (Fig. 3B–C).

Figure 3. Western diet promotes microbiota-dependent colonic MAdCAM-1 expression.

(A-B) Representative confocal images of MAdCAM-1 (red) immunofluorescence in the colonic mucosa of F11r^+/+ and F11r^−/− mice fed a western diet (WD) or WD plus antibiotics for eight-weeks (n = 5 mice per group). Nuclei are stained blue. White arrowheads, MAdCAM-1 expression. Scale bar 100 μm. (C) Quantification of MAdCAM-1 expression in the colon of F11r^+/+ and F11r^−/− mice fed a WD or WD plus antibiotics for eight weeks (n = 5 mice per group). Data are presented as mean ± SEM. Asterisks indicate significant differences (p < 0.05) between F11r^+/+ and F11r^−/− mice fed a WD. Hashtags indicate significant differences (p < 0.05) between WD or WD plus antibiotics fed F11r^−/− mice. (D-E) Cecum mucosa-associated microbiota from F11r^+/+ and F11r^−/− mice fed a ND or WD for eight-weeks were analyzed using 16S rRNA sequencing followed by phylogenetic analysis and a comparison of the microbial community structure using the unweighted UniFrac algorithm (n = 4 – 5 mice per group). (D) Jackknifed principal coordinate analysis (PCoA) of the un-weighted UniFrac distance matrix of the mucosa-associated microbiota. The ovals represent clustering by treatment groups. (E) Relative abundance of mucosa-associated bacterial phyla in F11r^+/+ and F11r^−/− mice. Data are presented as mean ± SEM. Asterisks indicate significant differences (p < 0.05) between F11r^+/+ and F11r^−/− mice fed an identical diet. Hashtags indicate significant differences (p < 0.05) between the ND or WD-fed F11r^+/+ and F11r^−/− mice.

To determine whether changes in the microbiota composition were associated with increased colonic MAdCAM-1 expression, the cecal mucosa-associated microbiota were analyzed using 16S rRNA sequencing followed by phylogenetic analyses and a comparison of the microbial community structure using the unweighted UniFrac algorithm as described previously [6]. We analyzed mucosa-associated microbiota as they have a greater influence on mucosal homeostasis than the luminal microbiota due to mucosal proximity [17]. The mucosa-associated microbial compositions of both control and F11r^−/− mice fed the ND were similar with higher abundance of Proteobacteria, followed by Bacteroidetes, Firmicutes, Tenericutes and Deferribacteres (Fig. 3D–E). WD consumption significantly increased Proteobacteria but decreased Bacteroidetes in the cecal mucosa of F11r^−/− mice, compared to WD-fed controls and the ND-fed mice. In contrast, WD consumption did not alter the abundance of Bacteroidetes and Proteobacteria in the controls. In both controls and F11r^−/− mice, WD consumption decreased Tenericutes. Since expansion of proteobacteria is most commonly associated with gut dysbiosis in intestinal inflammatory GI disorders [18, 19], our data suggest a link between Proteobacteria abundance in the colonic mucosa of WD-fed F11r^−/− mice and increased colonic MAdCAM-1 expression.

Hepatic MAdCAM-1 expression is higher in WD-fed mice.

To evaluate the involvement of α₄β₇⁺/MAdCAM-1 mediated recruitment of CD4 T cells to the NASH liver, we assessed hepatic MAdCAM-1 expression by laser confocal microscopy (Fig. 4A–B). In agreement with a previous study showing increased MAdCAM-1 expression in the liver of mice fed a WD for 24 weeks [20], confocal images of liver tissue sections stained with MAdCAM-1 revealed marked increase in hepatic MAdCAM-1 expression in the WD-fed F11r^−/− mice relative to controls (Fig. 4A–B). No difference in hepatic MAdCAM-1 expression was detected in ND-fed controls and F11r^−/− mice (Supp. Fig. 4B). Treatment with broad-spectrum antibiotics, which attenuates WD-induced hepatic inflammation and fibrosis in F11r^−/− mice [6], reduced hepatic MAdCAM-1 expression, strongly suggesting a role for microbiota in regulating diet-induced hepatic MAdCAM-1 expression (Fig. 4A–B). Further analysis revealed that hepatic MAdCAM-1 staining colocalized with alpha smooth muscle actin (αSMA), a key hepatic stellate cell (HSC) activation marker [21], suggesting that activated HSCs may be the primary cells expressing hepatic MAdCAM-1 in the WD-fed mice (Fig. 4C). Together, these data suggest that WD consumption induces microbiota-dependent hepatic MAdCAM-1 expression.

Figure 4. Western diet-induced increase in hepatic MAdCAM-1 expression is associated with increased recruitment of α₄β₇⁺ CD4 T cells to the liver.

(A) Representative confocal images of MAdCAM-1 (red) immunofluorescence in the liver of F11r^+/+ and F11r^−/− mice fed a normal (ND) or western diet (WD), or WD plus antibiotics for eight-weeks (n = 5 mice per group). Nuclei are stained blue. White arrowheads, MAdCAM-1 expression. Scale bar 100 μm. (B) Quantification of MAdCAM-1 expression in the liver of F11r^+/+ and F11r^−/− mice fed a WD or WD plus antibiotics for eight weeks (n = 5 mice per group). (C) Representative confocal images of MAdCAM-1 (red) and αSMA (green) immunofluorescence in the liver of F11r^−/− mice fed a WD for eight weeks (n = 5 mice per group). Nuclei are stained blue. Scale bar 100 μm. (D-E) Scatter plots show total number and percentage of (D) α₄β₇⁺ CD4 T cells (n = 5 – 8 mice per group) and (E) CD4 T cells (n = 5 – 11 mice per group) in the liver of F11r^+/+ and F11r^−/− mice fed a ND or WD for 8 weeks. (F) Representative confocal images of CD4 (green) and α₄β₇ (red) immunofluorescence in the liver of F11r^−/− mice fed a WD for eight weeks (n = 5 mice per group). Nuclei are stained blue. Scale bar 100 μm. (G-H) Expression of (E) integrin α4, Itga4 and (F) integrin β7, Itgb7 in the liver of F11r^−/− mice fed a WD for eight weeks (n = 5 mice per group). Data are presented as means ± SEM. Asterisks indicate significant differences (p < 0.05) between F11r^+/+ and F11r^−/− mice fed an identical diet. Hashtags indicate significant differences (p < 0.05) between ND or WD-fed control (F11r^+/+) and F11r^−/− mice.

Intrahepatic α₄β₇⁺ CD4 T cells are higher in WD-fed mice.

Knowing that hepatic MAdCAM-1 expression was elevated in F11r^−/− mice fed the WD, we next investigated whether there was a corresponding increase in hepatic α₄β₇⁺ CD4 T cell recruitment. As shown in Fig. 4D, WD feeding significantly increased the frequency and the total number of intrahepatic α₄β₇⁺ CD4 T cells in the F11r^−/− mice, compared to controls. Increased hepatic α₄β₇⁺ CD4 T cell recruitment correlated with a significant increase in the total number of hepatic CD4 T cells relative to controls (Fig. 4E). However, no difference in the frequency of CD4 T cells was observed between ND and WD fed mice (Fig. 4E). WD consumption also increased CD4 T cells and α₄β₇⁺ CD4 T cells in the control mice, compared with the ND-fed mice, however their numbers were significantly lower in control mice and correlated with lower hepatic MAdCAM-1 expression (Fig. 4A–E). Increased infiltration of α₄β₇⁺ CD4 T cells in the liver was further confirmed by laser confocal imaging demonstrating higher numbers of α₄β₇ and CD4 double positive cells in the liver of WD-fed F11r^−/− mice. The transcript levels of integrins α₄ and β₇ were significantly higher in the WD-fed F11r^−/− mice compared to controls (Fig. 4G–H). Further corroborating a role of gut microbiota in regulating α₄β₇/MAdCAM-1-mediated recruitment of immune cells in the NASH liver, transcript levels of integrins α₄ and β₇ were significantly reduced in the liver of antibiotic-treated WD-fed F11r^−/− mice (Fig. 4G–H). Together, these data suggest that WD consumption promotes hepatic recruitment of α₄β₇⁺ CD4 T cells.

Integrin α₄β₇ blockade decreases recruitment of α₄β₇⁺ CD4 T cells to the intestine of WD-fed mice.

To confirm that integrin α₄β₇ is involved in the recruitment of CD4 T cells to the intestine of WD-fed mice, we treated mice with a highly specific neutralizing monoclonal antibody (mAb) against α₄β₇, which is effective in blocking migration of α₄β₇⁺ CD4 T cells to the intestinal mucosa [22]. Mice treated with IgG isotype served as controls. α₄β₇ blockade significantly decreased the total number and frequency of α₄β₇⁺ CD4 T cells in the PP of WD-fed F11r^−/− mice (Fig. 5A). α₄β₇ mAb treatment reduced total CD4 T cell numbers in the PP without altering the frequency of CD4 T cells (Fig. 5B). α₄β₇ mAb treatment also reduced the total number of α₄β₇⁺ CD4 T cells in the CLP of WD-fed F11r^−/− mice, but did not alter the frequency of α₄β₇⁺ CD4 T cells (Fig. 5C–D). Decreased infiltration of α₄β₇⁺ CD4 T cells in the CLP of α₄β₇ mAb-treated mice correlated with a significant reduction in the transcript levels of MAdCAM-1 as well as integrins α₄ and β₇ compared with IgG controls (Fig. 5E–G). These data correlated with histological analysis of the colonic tissue showing decreased immune cell infiltration in the α₄β₇ mAb-treated mice, relative to IgG treated controls (Supp. Fig. 5B). Improvement in colonic inflammation in the α₄β₇ mAb-treated mice correlated with a significant increase in the transcript levels of tight junction proteins occludin and zonula occludens (ZO)-1 in the colon suggesting that α₄β₇ mAb treatment improved colonic epithelial barrier function (Fig. 5H–I). Together, these data demonstrate that α₄β₇ mAb treatment not only reduced mucosal inflammation in WD-fed F11r^−/− mice, but also improved colonic epithelial barrier function.

Figure 5. α₄β₇ blockade decreases α₄β₇⁺ CD4 T cell recruitment to the intestine.

(A-B) Scatter plots show total number and percentage of (A) α₄β₇⁺ CD4 T cells and (B) CD4 T cells in the Peyer’s patches (n = 4 – 6 mice per group). (C-D) Scatter plots show total number and percentage of (C) α₄β₇⁺ CD4 T cells and (D) CD4 T cells in the colonic lamina propria (LP) (n = 3 – 4 mice per group). (E-I) Expression of (E) Madcam1, (F) integrin α4, Itga4, (G) integrin β7, Itga7, (H) occludin and (I) Zo1 in the colonic mucosa (n = 5 mice per group). F11r^−/− mice fed a western diet (WD) for eight-weeks were treated with IgG or α₄β₇ mAb for four weeks starting at week four after the initiation of WD. Data are presented as mean ± SEM. Asterisks indicate significant differences (p < 0.05) between IgG and α₄β₇ mAb treatment.

Integrin α₄β₇ blockade decreases hepatic recruitment of α₄β₇⁺ CD4 T cells in WD-fed mice.

To determine whether α₄β₇ blockade decreased α₄β₇⁺ CD4 T cell recruitment to the liver, we analyzed hepatic CD4 T cells from IgG or α₄β₇ mAb treated F11r^−/− mice fed the WD (Fig. 6A). As shown in Fig 6A, four weeks of α₄β₇ mAb treatment significantly reduced both the percentage and total number of intrahepatic α₄β₇⁺ CD4 T cells in the WD-fed F11r^−/− mice suggesting that α₄β₇ mAb treatment effectively blocked WD-induced recruitment of α₄β₇⁺ CD4 T cells to the liver. α₄β₇ mAb treatment also reduced the total number of hepatic CD4 T cells, but did not affect the frequency of CD4 T cells (Fig. 6B). Decreased infiltration of α₄β₇⁺ CD4 T cells in the liver of α₄β₇ mAb treated mice correlated with a significant reduction in the transcript levels of hepatic MAdCAM-1 as well as integrins α₄ and β₇ compared with IgG controls (Fig. 6C–E). Together, these data demonstrate that integrin α₄β₇ promotes recruitment of CD4 T cells to the NASH liver.

Figure 6. α₄β₇ blockade attenuates hepatic inflammation and fibrosis.

(A-B) Scatter plots show total number and percentage of (A) α₄β₇⁺ CD4 T cells and (B) CD4 T cells in the liver (n = 7 – 8 mice per group). F11r^−/− mice fed a WD for eight weeks were treated with IgG or α₄β₇ mAb for four weeks starting at week four after initiation of the WD (n = 10 mice per group). (C-E) Expression of (C) Madcam1, (D) integrin α₄, Itga4 and (E) integrin β_7, Itgb7 in the liver (n = 5 mice per group). (F) Representative photomicrographs of Hematoxylin and Eosin (H&E) stained liver tissue sections. Black arrowheads, immune cells. (G-H) Serum ALT and AST levels. (I-K) Expression of key molecules associated with hepatic inflammation (n = 10 mice per group). (L) Representative photomicrographs of Sirius Red-stained liver tissue sections. Blue arrowheads, collagen deposition. (M-P) Expression of hepatic stellate cell activation markers and markers of fibrosis in the liver (n = 10 mice per group). Data are presented as mean ± SEM. Asterisks indicate significant differences (p < 0.05) between IgG and α₄β₇ mAb treatment.

Integrin a4b7 blockade attenuates WD-induced hepatic inflammation and fibrosis.

Next, to determine whether α₄β₇ blockade also ameliorated hepatic inflammation and fibrosis in WD-fed F11r^−/− mice, we performed histological analysis of liver tissue sections. As shown in Fig. 6F, hepatic steatosis and inflammation, assessed by H&E staining was ameliorated within four weeks of α₄β₇ mAb treatment in WD-fed F11r^−/− mice. Improvement in hepatic inflammation in α₄β₇ mAb treated WD-fed F11r^−/− mice was confirmed by a significant reduction in serum ALT (Fig. 6G) and AST (Fig. 6H) levels, as well as significantly decreased transcript levels of proinflammatory cytokines TNFα, interleukin (IL)-6 and IL-1β in the liver (Fig. 6I–K). α₄β₇ mAb treatment resulted in a marked decrease in hepatic fibrosis determined by Sirius Red staining (Fig. 6L). Expression of key molecules associated with hepatic fibrogenesis, αSMA, tissue inhibitor of metalloproteinase 1 (TIMP-1), and collagen type I [α1 and α2] were also significantly lower in α₄β₇ mAb treated WD-fed F11r^−/− mice relative to IgG control demonstrating a significant reduction in HSC activation and collagen deposition (Fig. 6M–P). Further confirming a role of α₄β₇/MAdCAM-1 axis in promoting hepatic inflammation and fibrosis in NASH, MAdCAM-1 blockade also reduced hepatic inflammation and fibrosis in WD-fed F11r^−/− mice (Supp. Fig. 6A–L).

Integrin a4b7 blockade improves metabolic syndrome.

In addition to attenuating histological features of steatohepatitis, four weeks of α₄β₇ mAb treatment also significantly improved metabolic parameters including body weight (Fig. 7A), liver (Fig. 7B), and visceral fat (Fig. 7C) weight, improved glucose tolerance (Fig. 7D) and insulin sensitivity (Fig. 7E), and reduced hepatic steatosis (Fig. 7F) in the WD-fed F11r^−/− mice. To gain additional insights into the α₄β₇ mAb induced changes in gene expression profile in WD-fed F11r^−/− mice, we performed RNA-sequencing (RNA-Seq) analysis. Based on FDR < 0.05 and absolute log2 fold change greater than 1.5 for upregulated genes and 2/3 for downregulated gene1, 86 genes were upregulated and 93 genes were down-regulated in the liver of α₄β₇ mAb treated WD-fed F11r^−/− mice (Fig. 7G–I). Pathway enrichment analysis of differentially expressed genes (DEGs) by Ingenuity Pathway Analysis revealed 17 pathways that were significantly enriched (p-value ≤ 0.01) in the α₄β₇ mAb treated WD-fed F11r^−/− mice compared with IgG controls. Genes associated with PXR/RXR activation, LPS/IL-1 mediated inhibition of RXR function, FXR/RXR activation, xenobiotic metabolism, Aryl hydrocarbon receptor signaling and glycogen degradation III were repressed, whereas EIF2 signaling and mTOR signaling were upregulated in α₄β₇ mAb treated WD-fed F11r^−/− mice (Fig. 6I). Collectively, these data suggest that α₄β₇ mAb treatment not only reverses hepatic inflammation and fibrosis but also improves metabolic parameters associated with NASH.

Figure 7. α₄β₇ blocking improves metabolic syndrome.

Changes in (A) body, and (B) liver and (C) visceral fat weight reported as percent of body weight. F11r^−/− mice fed a WD for eight-weeks were treated with IgG or α₄β₇ mAb for four weeks starting at week four following initiation of the WD (n = 10 mice per group). (D) Glucose and (E) insulin tolerance after four weeks of IgG or α₄β₇ mAb treatment (n = 5 mice per group). (F) Confocal microscopic images of BODIPY (lipids) stained liver tissue sections. Scale bar 20 μm. (G) Volcano plot showing differentially regulated genes following α₄β₇ mAb treatment. (H) Heatmap for the differentially expressed genes. (I) Ingenuity Pathway Analysis of most significantly enriched signaling pathways. Data are presented as mean ± SEM. Asterisks indicate significant differences (p < 0.05) between IgG and α₄β₇ mAb treatment.

NASH patients have higher MAdCAM-1 expression in the liver.

To corroborate the human relevance of our findings we analyzed liver biopsies from NASH patients to determine the expression of MAdCAM-1 by confocal microscopy and RT-PCR. Subjects were diagnosed as having NASH if they had a body mass index (BMI) ≥ 25kg/m², ≥ 50% hepatic steatosis, the presence of inflammatory immune cell infiltrates and fibrosis, elevated serum AST and ALT levels, NASH-CRN score >3 and fibrosis score >4 (Fig. 8A–C). We defined controls as patients with a normal BMI, and the absence of hepatic steatosis, inflammatory immune cell infiltrates and fibrosis (Fig. 8A–C). As shown in representative micrographs, hepatic MAdCAM-1 expression was higher in NASH patients relative to the controls (Fig. 8B), and correlated with significantly higher levels of MAdCAM1 transcripts (Fig. 8C). The transcript levels of αSMA (Fig. 8D), and integrins α₄ (Fig. 8E) and β₇ (Fig. 8F) were significantly higher in the liver of NASH patients. Together, these findings document a previously unappreciated role for α₄β₇/MAdCAM-1 in the recruitment of immune cells to the liver and driving hepatic inflammation and fibrosis in NASH.

Figure 8. MAdCAM-1 expression is higher in the liver of NASH patients and correlates with increased integrin a4 and b7 expression in the liver.

(A) Representative photomicrographs of Hematoxylin and Eosin (H&E) stained human liver tissue biopsies obtained from NASH patients (n = 15) and subjects without NASH (controls; n = 8). Scale bar 100 μm. (B-C) Serum AST and ALT levels. (D) Representative confocal images of MAdCAM-1 (green) immunofluorescence in the liver of NASH patients and subjects without NASH. Nuclei are stained blue. Scale bar 100 μm. (E-H) Expression of (E) MADCAM1, (F) ACTA2, (G) integrin α₄, ITGA4 and (H) integrin β_7, ITGB7 in the liver (n = 5). Data are presented as mean ± SEM. Asterisks indicate significant differences (p < 0.05) between controls and NASH patients.

DISCUSSION

Building upon our understanding of the gut-liver axis in NASH pathogenesis, the present findings support the novel observation that α₄β₇⁺ CD4 T cells not only contribute to the loss of intestinal epithelial barrier in the WD-fed F11r^−/− mice, but also promote hepatic inflammation and fibrosis. Our results demonstrate that consumption of a WD results in a systemic increase in α₄β₇⁺ CD4 T cells which are actively recruited to the gut and liver by diet-induced, microbiota-dependent intestinal and hepatic MAdCAM-1 expression. Blocking α₄β₇ decreased both intestinal and hepatic recruitment of α₄β₇⁺ CD4 T cells, substantially reducing mucosal inflammation and hepatic inflammation and fibrosis in WD-fed F11r^−/− mice. Furthermore, we demonstrate that blocking MAdCAM-1 also results in similar attenuation of hepatic inflammation and fibrosis in WD-fed F11r^−/− mice. These observations suggest a previously unappreciated dual role for α₄β₇⁺ CD4 T cells in regulating both mucosal and hepatic inflammation in NASH.

The intestinal epithelial barrier is an important defense against luminal microbes, and the gut mucosal immune system maintains the integrity of this barrier [23]. Our data demonstrate that WD dysregulates mucosal immune homeostasis by increasing α₄β₇⁺ CD4 T cell recruitment to the intestinal mucosa. In concert with our observations suggesting an association between diet-induced, microbiota-dependent increase in colonic MAdCAM-1 expression and increased recruitment of α₄β₇⁺ CD4 T cells to the intestine, a recent study showed that whole-body MAdCAM-1 knockout mice are protected from diet-induced NASH [20]. In contrast, the same study reported more severe NASH in β₇ knockout mice, which was associated with increased hepatic MAdCAM-1 expression, but a decrease in the hepatic CD4/CD8 ratio as well as the percentage of Foxp3+ T regulatory cells [20]. The study did not take into account the established defects in the gut associated lymphoid tissue development in β₇ knockout mice [24, 25], which complicate interpretation of the data. These confounding results emphasize the inherent limitations of whole-body knockout mice in functional studies as nonspecific phenotypes in whole-body knockouts are ignored. The α₄β₇ mAb used in our study is highly specific and binds a conformational epitope accessible only in the heterodimer α₄β₇ [26–28]. Its specificity has been tested in multiple experimental and clinical settings, and the humanized α₄β₇ mAb is an established therapy for inflammatory bowel disease [27–30]. The role of integrin receptor α₄β₇⁺ CD4 T cells in promoting diet-induced mucosal inflammation in NASH is of particular interest as a similar mechanism has been reported in inflammatory bowel disease where α₄β₇ blockade, which reduces mucosal inflammation by decreasing recruitment of CD4 T cells to the mucosa, has become an effective therapy [27, 29, 30]. It should be noted that WD also increased recruitment of α₄β₇⁺ CD4 T cells to the intestine of control mice, which is associated with tight junction disruption as indicated by redistribution of the tight junction protein occludin [6]. However, we did not detect increased intestinal permeability by FITC-dextran in these mice which do not develop NASH [6], suggesting that additional second hits such as gut dysbiosis, are required for disease progression.

Little is known about the potential role of adaptive immune cells as mediators of chronic hepatic inflammation and fibrosis in NASH. Recent advances in our understanding of diseases associated with MetS including NAFLD have demonstrated that the cells of the adaptive immune system have the potential to play a significant role in NASH development and disease progression. The role of T lymphocytes may be especially important, particularly CD4 T cells that produce pro-inflammatory mediators that can regulate activation of the innate immune system [31]. Our studies demonstrated a significant increase in α₄β₇⁺ CD4 T cells in the liver of WD-fed F11r^−/− mice which was associated with more robust hepatic inflammation and fibrosis. Our studies for the first time demonstrate the mechanistic involvement of these cells in NASH as α₄β₇ mAb treatment, which reduced hepatic α₄β₇⁺ CD4 T cell infiltration, attenuated hepatic inflammation and fibrosis indicating α₄β₇ mediated recruitment of CD4 T cells to the liver in NASH. Our data demonstrating increased hepatic MAdCAM-1 expression and increased hepatic infiltration of α4β7+ CD4 T cells of WD-fed F11r^−/− mice strongly support a role for α4β7+ CD4 T cells in promoting hepatic inflammation in NASH. Increased recruitment of α4β7+ CD4 T cells to the liver is also reported in human chronic inflammatory liver diseases [10, 11]. While further large-scale human studies are needed, increased MAdCAM-1 as well as integrins α4 and β7 expression in the liver of NASH patients suggests that the α4β7/MAdCAM-1 axis may also play a role in the recruitment of immune cells to the liver in human NASH.

In summary, our findings demonstrate for the first time provide novel mechanistic insights into the α₄β₇/MAdCAM-1 axis in immune-mediated intestinal epithelial barrier disruption and hepatic inflammation in NASH. Finally, our findings provide a logical framework for targeted therapy using α₄β₇ blockade for the treatment of NASH.

Highlights.

Western diet consumption increases gut homing integrin receptor α₄β₇⁺ CD4 T cells in the peripheral blood.

α₄β₇⁺ CD4 T cells are recruited to the gut and liver by diet-induced, microbiota-dependent expression of MAdCAM-1, the ligand for α₄β₇.

α₄β₇ blockade attenuates hepatic inflammation and fibrosis, and improves metabolic derangements associated with NASH.

Hepatic MAdCAM-1 expression is elevated in NASH patients and correlates with higher expression of integrins α₄ and β₇ in the liver.

α₄β₇/MAdCAM-1 is a critical axis regulating NASH development, and blockade of α₄β₇ may represent a novel therapy for treating NASH patients