Mucosal Eosinophil Abundance in Non-Inflamed Colonic Tissue Is Associated with Response to Vedolizumab Induction Therapy in Inflammatory Bowel Disease

Abstract

Vedolizumab is used as a treatment for patients with inflammatory bowel disease (IBD), but induction therapy leads to clinical response and remission in approximately 55% and 30% of patients with IBD, respectively. In this study, we aimed to explore the predictive value of mucosal eosinophils and serum eotaxin-1 regarding response to vedolizumab induction therapy. Eighty-four (84) patients with IBD (37 Crohn’s disease [CD], 47 ulcerative colitis [UC]) were included. For 24 patients with IBD, histopathology was assessed for eosinophil counts in non-inflamed colonic tissue prior to vedolizumab treatment. For 64 patients with IBD, serum eotaxin-1 levels were quantified prior to (baseline) and during vedolizumab treatment. Serum samples of 100 patients with IBD (34 CD, 66 UC) from the GEMINI 1 and 2 trials were used for external validation. Baseline mucosal eosinophil numbers in non-inflamed colonic tissue were significantly higher in responders to vedolizumab induction therapy when compared to primary non-responders (69 [34−138] vs. 24 [18−28] eosinophils/high-power field, respectively, p < 0.01). Baseline serum eotaxin-1 levels in the discovery cohort were significantly elevated in responders, compared to primary non-responders (0.33 [0.23−0.44] vs. 0.20 [0.16−0.29] ng/mL, p < 0.01). Prediction models based on mucosal eosinophil counts and serum eotaxin-1 showed an area under the curve (AUC) of 0.90 and 0.79, respectively. However, the predictive capacity of baseline serum eotaxin-1 levels could not be validated in the GEMINI cohort. Mucosal eosinophil abundance in non-inflamed colonic tissue was associated with response to vedolizumab induction therapy in patients with IBD. Future studies are warranted to further validate the potential value of mucosal eosinophils and serum eotaxin-1 as biomarkers for response to vedolizumab therapy.

Keywords: eosinophil; eotaxin-1; inflammatory bowel disease; vedolizumab.

Figure 1

(A–C). Mucosal eosinophil abundance associates with clinical response to vedolizumab induction therapy in the discovery cohort. (A) Median eosinophils/hpf were significantly higher among patients who eventually responded to vedolizumab induction therapy vs. patients who showed no response. (B) Mucosal eosinophil counts were significantly associated to blood eosinophil concentrations. The blue line was fitted using a robust regression method by minimizing least trimmed squares (LTS), which is not unduly affected by outliers. The blue shade represents the 95% confidence interval (CI). (C) Blood eosinophil concentrations significantly increased after vedolizumab induction therapy. *** p < 0.001. Abbreviations: hpf, high power field.

Figure 2

(A–F) Serum eotaxin-1 levels (ng/mL) among clinical responders and primary non-responders to vedolizumab induction therapy in the discovery cohort. (A) Baseline serum eotaxin-1 levels (ng/mL) were significantly higher in patients with IBD eventually responding to vedolizumab induction therapy versus those who did not. (B,C) Baseline serum eotaxin-1 levels (ng/mL) were significantly elevated in patients with Crohn’s disease (CD) who eventually clinically responded to vedolizumab, whereas in patients with UC, no significant difference was observed when comparing responders to non-responders. (D–F) Serum eotaxin-1 levels (ng/mL) significantly increased in patients during the course of vedolizumab induction therapy, in both responders and non-responders. * p < 0.05; ** p < 0.01; *** p < 0.001.

Figure 3

(A–D) Receiver operating characteristics (ROC) curves for mucosal eosinophil counts/hpf. (A) serum CRP levels (mg/L) (B) serum eotaxin-1 levels (ng/mL) (C) and adjusted serum eotaxin-1 levels (combined predicted probability of multivariable logistic regression model) (D) for discriminating responders and non-responders to vedolizumab among patients with IBD. The best discriminative performance to predict clinical response or remission to vedolizumab induction therapy was demonstrated by mucosal eosinophil counts (although based on a subset of n = 24 patients), while the performance of the multivariable model of serum eotaxin-1 levels (n = 64) also showed reasonable discrimination. Data presented are based on patients from the discovery cohort. Abbreviations: CRP, C-reactive protein; AUC, area under the ROC curve.

Figure 4

(A–D) Serum eotaxin-1 levels at week 0 and week 6in responders and non-responders to vedolizumab induction therapy in patients from the GEMINI cohort. (A,B) Patients with CD showed no differences in serum eotaxin-1 levels at baseline nor after six weeks of therapy. (C,D) Similarly, patients with UC demonstrated no significant differences at baseline or at week 6 of therapy.