BCL-2 levels do not predict azathioprine treatment response in inflammatory bowel disease, but inhibition induces lymphocyte apoptosis and ameliorates colitis in mice

Abstract

In inflammatory bowel disease (IBD), inflammation is sustained by an exaggerated response of lymphocytes. This results from enhanced expression of anti-apoptotic B cell lymphoma (BCL-2) and BCL-XL associated with a diminished turnover. Azathioprine (AZA) directly targets BCL-2 family-mediated apoptosis. We investigated whether the BCL-2 family expression pattern could be used to predict treatment response to AZA and determined whether BCL-2 inhibitor A-1211212 effectively diminishes lymphocytes and ameliorates inflammation in a model of colitis. BCL-2 family expression pattern was determined by next-generation sequencing (NGS). BCL-2 inhibitor was administered orally to Il10-/- mice. Haematological analyses were performed with an ADVIA 2120 and changes in immune cells were investigated using quantitative polymerase chain reaction (qPCR) and fluorescence activated cell sorter (FACS). We determined similar expression levels of BCL-2 family members in patients with remission and patients refractory to treatment, showing that BCL-2 family expression can not predict AZA treatment response. Expression was not correlated with the modified Truelove and Witts activity index (MTWAI). BCL-2 inhibitor initiated cell death in T cells from patients refractory to AZA and reduced lymphocyte count in Il10-/- mice. FACS revealed diminished CD8⁺ T cells upon BCL-2 inhibitor in Il10-/- mice without influencing platelets. Tnf, Il1β, IfnƔ and Mcp-1 were decreased upon BCL-2 inhibitor. A-1211212 positively altered the colonic mucosa and ameliorated inflammation in mice. Pro-apoptotic BCL-2 inhibitor A-1211212 diminishes lymphocytes and ameliorates colitis in Il10-/- mice without inducing thrombocytopenia. BCL-2 inhibition could be a new therapy option for patients refractory to AZA.

Keywords: A-1211212; ABT-737; BCL-2; BIM; IBD; apoptosis; mucosal T cell turnover.

Figure 1

6‐Mercaptopurine (6‐MP) ‐mediated apoptosis is abolished in CD4⁺ T cells from inflammatory bowel disease (IBD) patients upon azathioprine (AZA) at time of medical examination. Flow cytometry and quantitative polymerase chain reaction (qPCR). (a) Propidium iodide (PI)/annexin V staining of human CD4⁺ T cells from healthy controls and IBD patients upon AZA at time of medical examination. Analysis of variance (anova) on ranks, all pairwise multiple comparison procedures (Dunn’s method), n as indicated, error bars = standard deviation (s.d.), *P < 0·05. $ = Statistics considering healthy controls. $$ = Statistics considering patients. # = t‐test, normality test (Shapiro–Wilk) and equal variance test passed. (b) BCL‐XL qPCR. anova = all pairwise multiple comparison procedures (Dunn’s method), error bars = s.d., *P < 0·05.

Figure 2

Inhibition of B cell lymphoma (BCL)‐2 initiates cell death in CD4⁺ T cells from inflammatory bowel disease (IBD) patients refractory to azathioprine (AZA). Flow cytometry from three IBD patients upon AZA treatment at time of medical examination, while a considerable deterioration to treatment was indicated.

Figure 3

A‐1211212 and ABT‐737 ameliorate colitis in Il10 ^‐/‐ mice. (a) Mini‐endoscopy was performed with a limited number of mice: 23 wild‐type (WT) mice and 38 Il10 ^‐/‐ mice, (b) Murine endoscopic index of colitis severity (MEICS) upon A‐1211212 or ABT‐737 treatment in WT mice and Il10 ^‐/‐ mice suffering from spontaneous colitis following 14 days of treatment, n as indicated, error bars = standard deviation. One‐way analysis of variance (anova), all pairwise multiple comparison procedures (Holm–Sidak method) for A‐1211212 and non‐parametric t‐test for ABT‐737, *P < 0·05, **P < 0·01, ***P < 0·001.

Figure 4

A‐1211212 and ABT‐737 ameliorate colitis in Il10 ^‐/‐ mice. (a) Haematoxylin and eosin (H&E), (b) histological score from small bowel of 23 wild‐type (WT) mice and 48 Il10 ^‐/‐ mice suffering from spontaneous colitis following 14 days of treatment with A‐1211212 or ABT‐737. Arrows indicate influx of lymphocytes, error bars = standard deviation. One‐way analysis of variance (anova). All pairwise multiple comparison procedures (Holm–Sidak method) for A‐1211212 and non‐parametric t‐test for ABT‐737, *P < 0·05, **P < 0·01, ***P < 0·001.

Figure 5

A‐1211212 decreases myeloperoxidase (MPO) in Il10 ^‐/‐ mice in a dose‐dependent manner. MPO assay using small bowel (a) and colon tissue samples (b) from 23 wild‐type (WT) mice and 32 Il10 ^‐/‐ mice suffering from spontaneous colitis following 14 days of treatment with A‐1211212, error bars = standard deviation. One‐way analysis of variance (anova). All pairwise multiple comparison procedures (Dunn’s method), *P < 0·05.

Figure 6

A‐1211212 decreases mRNA expression of inflammatory cytokines in Il10 ^‐/‐ mice. Quantitative polymerase chain reaction (qPCR) from small bowel. (a) Tnf, Il1β, Ifnγ and Mcp1, (b) Tgfβ mRNA expression in IL10^–/– mice suffering from spontaneous colitis following 14 days of treatment with A‐1211212, n as indicated, error bars = standard deviation. One‐way analysis of variance (anova). All pairwise multiple comparison procedures (Dunn’s method) for Tnf and (Holm–Sidak method) for Il1β, Ifnγ and Mcp1, *P < 0·05.

Figure 7

A‐1211212 decreases CD8⁺ T cell population in splenocytes in wild‐type (WT) and Il10 ^‐/‐ mice. Flow cytometry analysis of 23 WT and 26 Il10 ^‐/‐ mice suffering from spontaneous colitis following 14 days of treatment with A‐121121. (a) Representative images of flow cytometry analysis, n = total 49. Doublets discrimination performed, viable and dead cells included. Cells shown were gated to CD45⁺ and CD8⁺ or CD4⁺ splenocytes, (b) Dose‐dependent decrease of CD8⁺ splenocytes in WT and Il10 ^‐/‐ mice upon A‐1211212 treatment compared to vehicle as determined by flow cytometry, error bars = standard deviation. One‐way analysis of variance (anova). All pairwise multiple comparison procedures (Dunn’s method), *P < 0·05, **P < 0·01, ***P < 0·001.