Inflammatory bowel disease (IBD) is a group of chronic inflammatory diseases of the intestine and primarily includes ulcerative colitis (UC), Crohn’s disease (CD), and IBD-unclassified. In 2015, it was estimated that around 2.5 million people suffer from IBD in Europe, with the highest prevalence in Scandinavia and Great Britain, and this number is predicted to further increase [1
]. Characterized as a multifactorial disease, the full pathogenesis is not completely understood, but includes a genetic predisposition, environmental factors, and impaired epithelial barrier and immune response [2
]. Changes in chemokine levels in serum have been reported in IBD [6
], however, assessment of circulating chemokine levels following biological treatment is poorly investigated.
More than 40 members of the chemokine family are described in humans divided into the C, CC, CXC, and CXC3 subfamily. Functionally, CC chemokines are mainly responsible for the recruitment of lymphocytes, whereas CXC chemokines have the highest ability to attract neutrophils and monocytes [8
]. Besides the recruitment of effector cells, chemokines are involved in the adhesion and migration of leukocytes across the endothelium [9
]. To reach the site of inflammation, circulating leukocytes migrate through the endothelium in highly controlled processes, where in the interaction between integrins expressed on effector cells and adhesion molecules expressed on endothelial cells is crucial; in such interactions the effector cells migrate with a high degree of specificity. All integrins are formed by large α and small β chains. In total 14 integrin combinations are described to regulate immune cell traffic in humans [11
]. The specificity of the adhesion molecules differs, whereas vascular cell adhesion molecule 1 (VCAM-1) is expressed in the gut and the brain, mucosal addressin cell adhesion molecule-1 (MAdCAM-1) is only expressed in the intestine [14
The use of synthesized antibodies, biological agents, targeting specific components of the innate and adaptive immune system has improved the treatment of IBD. Biologics targeting tumor necrosis factor α (TNF-α), such as infliximab, adalimumab, golimumab, and certolizumab, have greatly improved IBD treatment. Anti-TNF agents aim to block cytokine signaling and to initiate inflammatory clearance through induction of apoptosis, thereby dampening the inflammatory response [15
Up to 40% of patients treated with anti-TNF agents do not initially respond and among those who achieve remission, 30–40% lose response over time [15
]. Therefore, new therapeutic targets have been investigated. Natalizumab, an antagonist for α4
integrin subunit binds both, the α4
molecules expressed on lymphocytes needed for the binding to endothelial cells via VCAM-1 and MAdCAM-1, respectively, to enter the tissue [14
]. Vedolizumab (Entyvio™, Takeda, Deerfield, IL, USA), so far the only gut-specific biological treatment approved for IBD-patients, is a humanized monoclonal antibody that binds the α4
integrin, and thus blocks the interaction with MAdCAM-1, which is expressed on endothelial cells. In 2013, the results from GEMINI I and II, two large phase III randomized controlled trials in UC and CD, respectively, demonstrated the effectiveness of vedolizumab in inducing and maintaining remission compared to placebo [17
]. Further, vedolizumab might restore colonic expression of genes involved in leukocyte migration in UC responders after therapy [19
]. Any potential modulation of systemic cytokine and chemokine levels following vedolizumab treatment has been poorly investigated.
In this open observational pilot study, we assessed the clinical response and circulating chemokine levels following vedolizumab treatment in IBD-patients previously non-responding to anti-TNF agents.
In this study, we investigated the clinical outcome and circulating chemokine levels following vedolizumab treatment in IBD-patients previously non-responding to anti-TNF agents. A proximity extension assay, a highly sensitive technology to quantify low protein concentrations, was applied [20
] and chemokine expression level was related to clinical response to therapy. CCL13 levels increased after treatment with vedolizumab. Further, pronounced changes in the levels of several other chemokines were seen when patients were sub-grouped into responders versus non-responders, indicating that there might be a prognostic value of measuring chemokine levels when starting therapy with vedolizumab. CCL13 levels at baseline showed an AUC of 0.833 (95% CI 0.58–1.00) to predict response to induction therapy, however this result should be interpreted with caution since the sample size is small. To the best of our knowledge, this pilot study is the first study investigating systemic chemokine levels in IBD-patients before and after therapy with vedolizumab and we find support for that analyses of chemokines in serum are relevant to understand systemic changes induced by a biologic drug mainly considered to be “gut-specific”.
Vedolizumab provides clinical benefit to some patients failing anti-TNF-treatment, although the induction of clinical remission becomes evident later when compared to TNF-naïve patients [23
]. From a pathophysiological point of view, CRP correlated negatively with 6 chemokines before, but not after treatment, and fecal calprotectin correlated negatively with 4 chemokines after treatment. Since median calprotectin levels were reduced, albeit not significantly, and most of the chemokine levels did not change after treatment, this could at least partially explain this correlation pattern. Sands and colleagues have shown that vedolizumab was superior to placebo in inducing clinical remission at week 10 in TNF-failure patients [24
]. They found non-significant, modest improvements in CRP at week 6 and 10 after vedolizumab therapy, along with a non-significant difference in fecal calprotectin at week 6. Lower levels of fecal calprotectin were observed in UC patients treated with vedolizumab compared to placebo at week 6 [18
]. Besides controlling the cell traffic, another strategy is to directly target chemokines and/or their receptors. It has been proposed that targeting the interaction between CCL25 and it receptor CCR9 might be a possible alternative to treat IBD-patients [25
When we analyzed chemokine levels in serum before and after vedolizumab therapy, CCL13 increased after therapy. Taking into account that CCL13 induces the expression of adhesion molecules in epithelial cells and we detected an increased expression of CCL13 after therapy [26
] and that CCL13 is described to recruit monocytes, eosinophils, and Th2 lymphocytes [26
] this might explain a possible rescue mechanisms. The increased expression of adhesion molecule might allow cells from the innate immune system e.g., monocytes or neutrophils, to enter the site of inflammation in higher numbers, thus shifting the main players in the inflammatory process. This observation is of interest, since we detected significant increased levels of CCL13 in the group of non-responders as compared to the responders. However, the clinical implication of this observation needs further investigation.
When we analyzed chemokine levels according to clinical response, we found that responding patients exhibited decreased levels of CCL28 after treatment. On the opposite, in addition to elevated levels of CCL13, CXCL8 was also increased in non-responders after therapy. CCL28 is also known as mucosa-associated epithelial chemokine (MEC) [28
], and its expression is associated with epithelial inflammation and recruitment of regulatory T-cells expressing CCR10 [29
]. Besides its chemotactic activity, CCL28 has antimicrobial functions, and is constitutively expressed in colon [30
]. CXCL8 is a potent chemoattractant for neutrophils, which in turn are an important source of CXCL8 in mucosal tissue of IBD patients [7
]. Thus, in mucosal tissue a positive feedback loop of neutrophil attraction and CXCL8 levels occurs [32
]. We therefore propose that in addition to control T-cell trafficking, the control of leukocyte influx into the intestinal mucosa might be important for achieving clinical response to vedolizumab.
A higher serum concentration of CCL20, CCL23, and CXCL1 was found in non-responders compared to responders at week 10. CCL20, described as macrophage inflammatory protein-3α (MIP-3α), binds specifically to the CCR6 receptor and has chemoattractant functions to T-cells, B-cells, and subsets of dendritic cells [31
]. Systemic levels of CCL23 and CXCL1 are increased in IBD patients compared to healthy controls [7
]. CCL23 binds to CCR1, and is chemoattractant for resting T-cells and monocytes [32
]. CXCL1 binds to CCR2 and is chemoattractant to monocytes and neutrophils [35
]. Taken together, we found higher serum levels of chemokines mainly known to be increased in the serum from patients with IBD, moreover, these levels were even higher in non-responders after treatment. Possibly, an increased amount of chemokines in circulation of non-responding patients might be useful to decide which patients would benefit from vedolizumab treatment beyond the induction phase. However, this need to be confirmed in larger cohorts of patients followed for longer time periods.
Comparing our results with studies investigating the chemokine expression in serum after systemic treatment with infliximab, a decreased serum level of CCL4 and CXCL8 after infliximab treatment has been reported by Sato and colleagues [37
]. Further, no significant differences were observed before and after therapy in the systemic expression of CCL2, CCL3, CCL4, CXCL1, and CX3CL1 by Magnusson and colleagues in non-responders, whereas CCL4 was significantly down-regulated in responders [38
Mouse studies showed that blocking α4
molecules alone did not reduce the inflammation, however, the combination of blocking l
-selectin and the binding of α4
molecule to MAdCAM-1 was efficient in reducing inflammation in the intestine [39
]. Taking our data into account one possible reason why the therapy shows side effects might be the altered systemic expression of chemokines, resulting in the recruitment of different immune cells to the site of inflammation as a rescue mechanism. In our study, non-responder patients showed higher levels of CCL13 and CXCL8, both potent chemoattractants for monocytes and neutrophils, respectively [26
Our study has some limitations that should be taken into account when interpreting the findings. Firstly, this is a pilot study with a small sample size that could imply lack of power for some of the comparisons. However, the chemokine levels in this study could be used to calculate adequate sample sizes for future investigations. Secondly, trials are warranted either to compare the clinical and systemic/local changes of vedolizumab to a placebo control or to other drugs. Thirdly, we did not measure the drug concentration in circulation, which could be associated to the response to therapy. Despite these limitations, our study is the first one to evaluate the effects of vedolizumab therapy on the circulating levels of a large number of chemokines.