Dual targeting of lymphocyte homing and retention through α4β7 and αEβ7 inhibition in inflammatory bowel disease

Abstract

Anti-integrins are therapeutically effective for inflammatory bowel disease, yet the relative contribution of α4β7 and αEβ7 to gut lymphocyte trafficking is not fully elucidated. Here, we evaluate the effect of α4β7 and αEβ7 blockade using a combination of murine models of gut trafficking and longitudinal gene expression analysis in etrolizumab-treated patients with Crohn's disease (CD). Dual blockade of α4β7 and αEβ7 reduces CD8+ T cell accumulation in the gut to a greater extent than blockade of either integrin alone. Anti-αEβ7 reduces epithelial:T cell interactions and promotes egress of activated T cells from the mucosa into lymphatics. Inflammatory gene expression is greater in human intestinal αEβ7+ T cells. Etrolizumab-treated patients with CD display a treatment-specific reduction in inflammatory and cytotoxic intraepithelial lymphocytes (IEL) genes. Concurrent blockade of α4β7 and αEβ7 promotes reduction of cytotoxic IELs and inflammatory T cells in the gut mucosa through a stepwise inhibition of intestinal tissue entry and retention.

Trial registration: ClinicalTrials.gov NCT02394028.

Keywords: T cell entry and retention; etrolizumab; inflammatory bowel disease; α4β7; αEβ7.

Conflict of interest statement

All authors, except J.G., A.H., M.D., M.S., and W.A.F., are current or past employees of Genentech, a member of the Roche group, and may hold Roche stock or stock options.

© 2021 The Authors.

Figures

Graphical abstract

Figure 1

Blockade of β7 or α4β7 and αEβ7 reduces CD8+ T cell accumulation in the small intestine and colon over α4β7 alone in the KikGR transgenic mouse model (A) Lymphoid and gut tissues from control or FTY720-treated KikGR transgenic mice were analyzed 16 h after mLN photoconversion. Gated CD45+TCRβ+ cells photoconverted in the mLN can be identified by their expression of KikR red fluorescent protein. (B–D) Anti-integrin or isotype control antibodies were administered to KikGR transgenic mice before surgical photoconversion of mLN. Colonic CD4+ (top) and CD8+ (bottom) T cell accumulation was reduced by anti-integrin antibody treatment. Data are from a representative sample of six to eight mice analyzed (B). Reduction in KikR+ CD45+TCRβ+CD8+ T cells from colon (C) and small intestine (D) after anti-β7 treatment in comparison with anti-α4β7 treatment. (E and F) Isotype control, anti-αE, and/or anti-α4β7 integrin antibodies were administered to KikGR mice before surgical photoconversion of mLN. The percentage of photoconverted KikR+ cells among CD45+TCRβ+CD8+ T cells in the colon (E) and small intestine (F) are shown. Bar graph shows means ± SEM. Data are from six to eight animals combined from two to three independent experiments. ∗p 

Figure 2

Reduction in both lamina propria and intraepithelial antigen-specific CD8+ T cells with blockade of α4β7 and αEβ7 is increased over α4β7 blockade alone KikG+ green fluorescent ovalbumin-specific OT-1 TCR transgenic CD8+ T cells were adoptively transferred into WT recipients that were then orally challenged with cholera toxin:OVA the following day. Three days after immunization, recipient mice were treated with isotype or anti-integrin antibodies 16 h before isolation of cells from the small intestine. The number of KikG+ OT-1 cells in the LP (A) and intraepithelial (B) compartments after antibody treatment are shown. Bar graph shows means ± SEM of six to eight mice combined from two independent experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 by one-way ANOVA with Tukey’s multiple-comparison post-test.

Figure 3

αEβ7 or E-cadherin blockade inhibits T cell interactions with intestinal epithelium Adoptive transfer of tdTomato+ OT-1 cells into WT or E-cadherin-CFP recipients was followed by oral challenge with cholera toxin:OVA the following day. Three days after immunization, recipient mice were treated with isotype or anti-αE or anti-E-cadherin antibodies before analysis. (A) OT-1 cell number in the small intestine lamina propria. Bar graph shows means ± SEM of six to eight mice combined from two independent experiments. (B) Confocal microscopy image of intestinal tissue sections stained with anti-E-cadherin (green) and anti-laminin (blue). White arrows point to tdTomato (red) OT-1 T cells in close proximity to E-cadherin on basolateral epithelium. Scale bar: 100 μm. Data are representative of three independent experiments. (C) Representative movie of tdTomato (red) OT-1 T cells interacting with intestinal epithelium (green) in control or anti-αE-treated mice. The basement membrane is indicated by a dashed white line. Scale bar: 20 μm. (D and E) Maximal track speed (D) and track length (E) were quantified for individual cells. Bar graph shows means ± SEM, and each dot represents an individual cell migration event. Data are representative of five independent experiments. One-way ANOVA with Dunnett’s post-test or unpaired Student‘s t test was used to calculate statistical significance. ∗p 

Figure 4

αE inhibition increases intestinal T cell egress from small intestine into mLN (A and B) KikGR+ OT-1 cells were adoptively transferred into WT mice that were then orally challenged with cholera toxin:OVA. Three days after immunization, mice were treated with isotype or anti-αE integrin antibodies or FTY720, followed by a surgical photoconversion of the small intestine. Cells in the small intestine and mLN were evaluated 16 h after photoconversion by flow cytometry. Number of KikR+ OT-1 cells in small intestine (A) and mLN (B). Bar graphs show means ± (SEM) of six to eight animals combined from two to three independent experiments. (C and D) tdTomato+ OT-1 cells were adoptively transferred into WT mice before oral challenge and antibody treatment as in (A) and (B). (C) Representative confocal microscopy image of intestinal tissue sections stained with anti-LYVE-1 (green) and anti-Epcam-1 (blue). White arrows show tdTomato (red) OT-1 T cells within lymphatic vessels. Scale bar: 100 μm. (D) The small intestine was sectioned, and tdTomato OT-1 T cells were counted in ∼100 villus per mouse to determine the frequency of OT-1 cells within lymphatic vessels. Bar graph shows means ± SEM of three animals. Data are representative of three independent experiments. One-way ANOVA with Tukey’s multiple-comparison post-test was used to calculate statistical significance. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001

Figure 5

αEβ7 is highly expressed on human colonic CD8+ T cells and proinflammatory CD4+ T cells (A) Percentage of αE+ cells among CD4+ and CD8+ T cells from healthy or IBD colonic biopsies. (B and C) Colonic αE+ CD4+ T cell and αE−CD4+ T cell production of IL-17 and IFN-γ (B) and percentage of αE+ cells (C) among CD4+Foxp3+CD25+CD127low cells (Treg) and CD4+ T cells. Each dot represents an individual patient sample, and data are shown as means ± SEM. (D) Single-cell RNA-seq data from colonic biopsies from healthy subjects are shown as t-stochastic neighborhood embedding (t-SNE) of CD45+ cells colored by cell subsets (left). Right, expression of αE in the corresponding cell population. (E) T cell subsets and their expression of indicated genes. Dot size indicates the percentage of gene expression, whereas dot color signifies the mean expression level of selected marker genes across cell subsets. (F and G) Deep sequencing of sorted αE+CD8+TCRβ+ cells and αE−CD8+TCRβ+ cells from a diverticulitis (non-IBD) resection sample at single-cell resolution. (F) Heat plot of expression of selected genes. (G) Reads per kilobase of transcript per million mapped reads (RPKM) of granzyme A (left), CCL4 (middle), and CCL5 (right) in αE+ and αE− populations. p < 1.0 × 10−8 for all three genes. (H) Quantification of selected markers in colonic αE+ and αE− CD8+ T cells by flow cytometry. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 by unpaired Student’s t test. Human samples are from Stanford (A and H), UCSF (C–E), or the Mayo clinic (B, F, and G). See also Table S1.

Figure 6

Etrolizumab treatment significantly reduces expression of genes associated with CD8+ cytotoxic IELs Ileal or colonic biopsies were taken before treatment and at 14 weeks after treatment in a randomized placebo-controlled trial of etrolizumab (anti-β7 integrin) in patients with moderately to severely active CD. (A) Fold change of integrin genes at 14 weeks after etrolizumab or placebo treatment. Each dot is an individual patient sample pair. Points are colored by sampling location (blue, colon; green, ileum). Boxes show the standard error, with the middle bar showing the group mean value. (B and C) Volcano plots showing the log2-fold change and p value of each gene, comparing baseline to week-14 samples from etrolizumab-treated patients (B) or placebo-treated patients (C). Each gene is represented by a point, with genes shown in red undergoing a >1.5-fold change at a false-discovery rate (FDR) < 0.05. (D and E) Scatterplots showing the correlation of the log2-fold changes of individual genes in endoscopically active, versus inactive, disease and the log2-fold changes observed between screening and week-14 samples from (D) etrolizumab-treated patients or (E) placebo-treated patients. Each point represents an individual gene. (F and G) Fold change of expression of the signature gene sets of indicated T cell subtype gene sets (F) and the cytotoxic IEL-specific genes (G) at week 14 in patients treated with etrolizumab or placebo. Each dot represents an individual patient sample pair. Points are colored by sampling location (blue, colon; green, ileum). Boxes show the standard error, with the middle bar showing the group mean value. ∗p