Advertisement

Therapeutic Interleukin-6 Trans-signaling Inhibition by Olamkicept (sgp130Fc) in Patients With Active Inflammatory Bowel Disease

Open AccessPublished:March 02, 2021DOI:https://doi.org/10.1053/j.gastro.2021.02.062

      Background & Aims

      A large unmet therapeutic need exists in inflammatory bowel disease (IBD). Inhibition of interleukin (IL)-6 appears to be effective, but the therapeutic benefit of a complete IL6/IL6 receptor (IL6R) blockade is limited by profound immunosuppression. Evidence has emerged that chronic proinflammatory activity of IL6 is mainly mediated by trans-signaling via a complex of IL6 bound to soluble IL6R engaging the gp130 co-receptor without the need for membrane-bound IL6R. We have developed a decoy protein, sgp130Fc, that exclusively blocks IL6 proinflammatory trans-signaling and has shown efficacy in preclinical models of IBD, without signs of immunosuppression.

      Methods

      We present a 12-week, open-label, prospective phase 2a trial (FUTURE) in 16 patients with active IBD treated with the trans-signaling inhibitor olamkicept (sgp130Fc) to assess the molecular mechanisms, safety, and effectiveness of IL6 trans-signaling blockade in vivo. We performed in-depth molecular profiling at various timepoints before and after therapy induction to identify the mechanism of action of olamkicept.

      Results

      Olamkicept was well tolerated and induced clinical response in 44% and clinical remission in 19% of patients. Clinical effectiveness coincided with target inhibition (reduction of phosphorylated STAT3) and marked transcriptional changes in the inflamed mucosa. An olamkicept-specific transcriptional signature, distinguishable from remission signatures of anti–tumor necrosis factor (infliximab) or anti-integrin (vedolizumab) therapies was identified.

      Conclusions

      Our data suggest that blockade of IL6 trans-signaling holds great promise for the therapy of IBD and should undergo full clinical development as a new immunoregulatory therapy for IBD. (EudraCT no., Nu 2016-000205-36)

      Graphical abstract

      Keywords

      Abbreviations used in this paper:

      CD (Crohn’s disease), CDAI (Chrohn’s Disease Activity Index), CRP (C-reactive protein), GO (Gene Ontology), IBD (inflammatory bowel disease), IL (interleukin), IL6R (interleukin 6 receptor), PK (pharmacokinetics), SAE (serious adverse event), SES-CD (Simple Endoscopic Score for Crohn’s disease), sIL6R (soluble interleukin 6 receptor), SNP (single-nucleotide polymorphism), TFBS (transcription factor binding site), TNF (tumor necrosis factor), UC (ulcerative colitis)
      See editorial on page 2247.

       Background and Context

      A large unmet therapeutic need exists in inflammatory bowel disease (IBD). The therapeutic potential of interleukin-6 trans-signaling inhibition was tested in a phase 2a open-label clinical trial.

       New Findings

      The selective trans-signaling inhibitor olamkicept was tested in 16 patients with IBD. Molecular profiling showed that clinical remission was associated with distinct changes in mucosal pSTAT3 levels.

       Limitations

      Clinical efficacy needs to be further investigated in larger placebo-controlled clinical trials.

       Impact

      Interleukin-6 trans-signaling inhibition might open up novel therapeutic avenues for the treatment of IBD.
      Inflammatory bowel disease (IBD) is a chronic relapsing inflammatory disorder of the gut associated with severe symptoms and high health care costs. The pathophysiology of IBD is characterized by an excess of proinflammatory cytokines leading to a dysbalanced innate and adaptive immune homeostasis.
      • Schultze J.L.
      • Rosenstiel P.
      Systems medicine in chronic inflammatory diseases.
      ,
      • Neurath M.F.
      Targeting immune cell circuits and trafficking in inflammatory bowel disease.
      Approved therapies, including monoclonal antibodies against tumor necrosis factor (TNF), interleukin (IL)-12/23 (p40) or adhesion molecules (anti-α4β7 integrin), leave a large unmet need, because full efficacy is seen in only a fraction of patients.
      • Sands B.E.
      • Peyrin-Biroulet L.
      • Loftus E.V.
      • et al.
      Vedolizumab versus adalimumab for moderate-to-severe ulcerative colitis.
      IL6 is a pleiotropic cytokine produced by hematopoietic and nonhematopoietic cells, for example, in response to infection and tissue damage.
      • Aden K.
      • Breuer A.
      • Rehman A.
      • et al.
      Classic IL-6R signalling is dispensable for intestinal epithelial proliferation and repair.
      • Garbers C.
      • Aparicio-Siegmund S.
      • Rose-John S.
      The IL-6/gp130/STAT3 signaling axis: recent advances towards specific inhibition.
      • Garbers C.
      • Heink S.
      • Korn T.
      • et al.
      Interleukin-6: designing specific therapeutics for a complex cytokine.
      • Kang S.
      • Tanaka T.
      • Narazaki M.
      • et al.
      Targeting interleukin-6 signaling in clinic.
      Elevated mucosal levels of IL6 have been found in patients with IBD.
      • Mudter J.
      • Amoussina L.
      • Schenk M.
      • et al.
      The transcription factor IFN regulatory factor–4 controls experimental colitis in mice via T cell–derived IL-6.
      ,
      • Nikolaus S.
      • Waetzig G.H.
      • Butzin S.
      • et al.
      Evaluation of interleukin-6 and its soluble receptor components sIL-6R and sgp130 as markers of inflammation in inflammatory bowel diseases.
      In animal models of colitis as well as in small therapeutic trials in Crohn’s disease (CD), full inhibition of IL6 activity, either by blocking the receptor or the cytokine with monoclonal antibodies, has been effective,
      • Ito H.
      • Takazoe M.
      • Fukuda Y.
      • et al.
      A pilot randomized trial of a human anti-interleukin-6 receptor monoclonal antibody in active Crohn’s disease.
      ,
      • Danese S.
      • Vermeire S.
      • Hellstern P.
      • et al.
      Randomised trial and open-label extension study of an anti-interleukin-6 antibody in Crohn’s disease (ANDANTE I and II).
      albeit at the expense of increased serious adverse events (SAEs) that were mainly caused by excessive immunosuppression and blockade of regenerative responses and that included abscesses, intestinal perforation, and death. Blocking IL6 or its receptor (IL6R) by monoclonal antibodies, respectively, inhibits classic IL6 signaling that is conveyed through the IL6R complex, which is formed by IL6, IL6R, and 2 molecules of the signal transducer gp130. The transmembrane protein gp130 is present on virtually all cells of the body, whereas the membrane-bound IL6R is only expressed together with gp130 on certain cells, including immune cells, hepatocytes, and intestinal epithelial cells. Soluble IL6R (sIL6R), which originates from proteolytic cleavage of membrane-bound IL6R and, to a minor extent, from alternative splicing, can be found in the circulation,
      • Waetzig G.H.
      • Rose-John S.
      Hitting a complex target: an update on interleukin-6 trans-signalling.
      where it can form IL6/sIL6R complexes that activate gp130 on target cells (trans-signaling).
      • Riethmueller S.
      • Somasundaram P.
      • Ehlers J.C.
      • et al.
      Proteolytic origin of the soluble human IL-6R in vivo and a decisive role of N-glycosylation.
      ,
      • Rose-John S.
      • Heinrich P.C.
      Soluble receptors for cytokines and growth factors: generation and biological function.
      Interestingly, soluble variants of gp130 (sgp130) generated by alternative splicing are found in the circulation at amounts exceeding sIL6R levels and neutralizing its activity.
      • Wolf J.
      • Waetzig G.H.
      • Chalaris A.
      • et al.
      Different soluble forms of the interleukin-6 family signal transducer gp130 fine-tune the blockade of interleukin-6 trans-signaling.
      Although classic IL6 signaling is thought to represent a defense mechanism (eg, against pathogens), trans-signaling has been suggested as an important pathomechanism conveying chronic inflammation.
      • Scheller J.
      • Chalaris A.
      • Schmidt-Arras D.
      • et al.
      The pro- and anti-inflammatory properties of the cytokine interleukin-6.
      Active IL6 trans-signaling has been associated with resistance of mucosal effector T cells to apoptosis
      • Atreya R.
      • Mudter J.
      • Finotto S.
      • et al.
      Blockade of interleukin 6 trans signaling suppresses T-cell resistance against apoptosis in chronic intestinal inflammation: evidence in Crohn disease and experimental colitis in vivo.
      ,
      • Mitsuyama K.
      • Matsumoto S.
      • Rose-John S.
      • et al.
      STAT3 activation via interleukin 6 trans-signalling contributes to ileitis in SAMP1/Yit mice.
      and is involved in the differentiation of pathogenic T helper type 17 cells thereby crucially orchestrating inflammatory responses.
      • Garbers C.
      • Aparicio-Siegmund S.
      • Rose-John S.
      The IL-6/gp130/STAT3 signaling axis: recent advances towards specific inhibition.
      ,
      • Rose-John S.
      The soluble interleukin 6 receptor: advanced therapeutic options in inflammation.
      It can thus be hypothesized that a therapeutic approach specifically targeting IL6 trans-signaling (eg, using a decoy gp130Fc molecule) could specifically block chronic inflammation without interfering with physiologic and host defense activities involving classic IL6 signaling. Olamkicept (FE 999301; TJ 301) is a first-in-class gp130 trans-signaling inhibitor and anti-inflammatory biologic compound, in which 2 complete extracellular domains of gp130 were dimerized by fusion to the fragment crystallizable region (Fc region) of human IgG1 (EP1148065B1, EP1873166B1). Olamkicept acts as an IL6/sIL6R trap and does not interact significantly with IL6 or with IL6R individually.
      • Jostock T.
      • Müllberg J.
      • Özbek S.
      • et al.
      Soluble gp130 is the natural inhibitor of soluble interleukin-6 receptor transsignaling responses.
      In animal models, it has succinct anti-inflammatory properties,
      • Garbers C.
      • Heink S.
      • Korn T.
      • et al.
      Interleukin-6: designing specific therapeutics for a complex cytokine.
      without any immune suppression in infection challenge models.
      • Sodenkamp J.
      • Waetzig G.H.
      • Scheller J.
      • et al.
      Therapeutic targeting of interleukin-6 trans-signaling does not affect the outcome of experimental tuberculosis.
      ,
      • Hoge J.
      • Yan I.
      • Jänner N.
      • et al.
      IL-6 controls the innate immune response against Listeria monocytogenes via classical IL-6 signaling.
      In clinical phase 1 trials (EudraCT no. 2012-005142-38 and 2013-004208-20), single doses of up to 750 mg in a total of 64 healthy volunteers and 24 patients with CD as well as multiple doses of up to 600 mg over 4 weeks in a total of 24 healthy volunteers did not induce clinically relevant adverse effects and did not result in systemic immune suppression. The aim of this phase 2a trial was to assess the target engagement/molecular mechanisms, clinical effectiveness, safety, and pharmacokinetics (PK) of 600 mg olamkicept in patients with active IBD after 12 weeks of active treatment and to investigate target engagement in relation to changes in clinical disease activity.

      Methods

       Study Drug

      Olamkicept comprises 2 gp130 extracellular domains dimerized by the Fc part of human IgG1 (sgp130Fc) to trap the complex of IL6 and sIL6R. This leads to inhibition of trans-signaling, mainly affecting IL6-driven chronic inflammation.
      • Waetzig G.H.
      • Rose-John S.
      Hitting a complex target: an update on interleukin-6 trans-signalling.
      We introduced stabilizing modifications (EP1873166B1) into the initial version of sgp130Fc. This modified compound, which allows stable pharmaceutical production, is called olamkicept. Olamkicept was manufactured by Ferring Pharmaceuticals in collaboration with Lonza.

       Patients and Treatments

      This phase 2a trial, conducted at the University Medical Center Schleswig-Holstein in Kiel, Germany, and at the Asklepios Westklinikum in Hamburg, Germany, recruited patients with moderately to severely active ulcerative colitis (UC) and CD aged 21–66 years (Supplementary Figure 1). Disease activity was determined by the Mayo score (maximum, 11; endoscopy, ≥2) and CD Activity Index (CDAI) (220–500; Simple Endoscopic Score for CD [SES-CD], ≥7) accompanied by increased C-reactive protein (CRP) levels (≥5 mg/L). Patients had to have experienced failure with conventional therapies with no more than 2 prior biologics (limited to anti-TNFs and/or vedolizumab). Patient demographic data are summarized in Supplementary Table 1. Glucocorticoids were limited to 20 mg prednisolone equivalent/day and were kept stable over the entire study period. Patients who were receiving an aminosalicylate or an oral immunosuppressant at baseline maintained stable doses throughout the trial.

       Trial Design

      An investigator-initiated, exploratory, open-label trial (EudraCT no. 2016-000205-36) was designed to investigate the target engagement/molecular mechanisms, clinical effectiveness, safety, and PK of trans-signaling inhibition induced by systemic exposure to olamkicept (FE 999301, TJ 301) in patients with active UC or CD. The trial was conducted over 21 months between 2016 and 2019, and 600 mg of olamkicept was administered by intravenous infusion every 2 weeks for 12 weeks (ie, 7 infusions). Clinical disease activity (including endoscopy) was assessed at baseline and weeks 2, 6, and 14. Additional visits for biosampling (blood, sigmoid mucosal biopsies) and PK (blood) were conducted 4 hours and 24 hours after therapy. All endoscopies were centrally read. A continuous paper diary was dispensed to the patients for the self-reported elements of the clinical scores and collected at each visit. Patients were instructed for self-reporting as suggested by the US Food and Drug Administration. Individual patient data for all clinical and biomarker assessments are shown in Supplementary Figure 2.

       Safety Assessments

      Active treatment was used for 12 weeks, with a final safety follow-up visit at weeks 16–18. Adverse events (classified according to the Medical Dictionary for Regulatory Activities, version 21.0), results of laboratory tests and safety assessments, and concomitant medications were recorded throughout the trial.

       Outcome Measures

      The primary molecular outcome was the change of a mucosal proinflammatory gene signature (TNF, IL1A, REG1A, IL8, IL1B, and LILRA) as a composite score from baseline to week 14. The composite score consists of a set of genes that represent the level of mucosal inflammation in mucosal tissue.
      • Häsler R.
      • Sheibani-Tezerji R.
      • Sinha A.
      • et al.
      Uncoupling of mucosal gene regulation, mRNA splicing and adherent microbiota signatures in inflammatory bowel disease.
      The primary assessment was remission followed by the secondary measures of response and endoscopic improvement. Clinical response and remission were assessed at week 14. Clinical remission was defined as Mayo score of ≤2, bleeding score of 0, and endoscopy of ≤1 for UC and CDAI of <150 for CD. Clinical response was defined as a reduction of the Mayo score of ≥3 points and bleeding score ≤1(UC) or as a reduction of the CDAI of >100 (CD). Endoscopic remission was defined as a subscore of 0 or 1 on the Mayo endoscopic component or an SES-CD of ≤4 with no ulcers present, and endoscopic response was a reduction of the Mayo score by 1 point or a reduction of the SES-CD by 50%.

       Trial Oversight and Reporting

      The trial sponsor (University Hospital Schleswig-Holstein, represented by Stefan Schreiber) together with the principal academic investigators provided formal oversight and provided trial drugs through the institutional pharmacy, which manufactured the infusions under Good Manufacturing Practice according to European Union directive 2001/20/E. An independent academic clinical research organization (Center for Clinical Trials of the Medical Faculty of Kiel University) managed the collection of all data, maintained the trial database and performed the data analyses. The trial investigators, the participating institutions, the clinical research organization, and the sponsor maintained data confidentiality. The complete manuscript was written by the academic authors. All authors had full data access, interpreted the data, contributed to subsequent drafts, and made the decision to submit the manuscript for publication.

       RNA Sequencing

      Whole blood and sigmoid mucosal biopsy samples (collected from the most inflamed region within 30 cm of the colon sigmoideum) were collected as a longitudinal data set from which the RNA for transcriptome analyses was isolated. The samples were gathered at baseline (0 hours) and 4 hours, 24 hours, 2 weeks, 6 weeks, and 14 weeks after the first olamkicept infusion. Whole RNA libraries were prepared using TruSeq RNA Library Prep technology (Illumina), and in total, 104 blood and 105 biopsy samples were sequenced using HiSeq 4000 (Illumina). (Both data sets were of paired data with 2 × 75 base pairs). See the Supplementary Materials and Methods for further details.

       Statistical Analysis

      Analyses of clinical effects and of safety included all patients who had received at least 1 dose of olamkicept. Missing values for binary outcomes were imputed as nonresponder, and missing values for continuous outcomes were imputed with the use of the last-observation-carried-forward approach. Reasons for discontinuation included worsening of disease (n = 5) and patient withdrawal of consent without reason (n = 1).

       Olamkicept Pharmacokinetics and Anti-drug Antibodies

      Blood sampling for olamkicept PK was performed at baseline, at 4 hours after the first infusion, and thereafter at trough and during follow-up. Analysis of serum concentrations of olamkicept was performed by the Department of Bioanalysis at Ferring Pharmaceuticals A/S. The presence of antibodies against olamkicept was examined at baseline, week 4, week 8, week 15, and week 18 after olamkicept infusion. (Details are described in Supplementary Materials and Methods.)

       Study Approval

      Conduction of the trial was approved by the higher competent authority (Paul-Ehrlich-Institute, Langen, Germany) and the ethics committee of the Medical Faculty of Kiel University (A 102/16). The trial was executed in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines (Good Clinical Practice ordinance V with 2012 modifications). All patients provided written informed consent before the first study procedures.

      Results

       Patients

      Overall, 34 patients were prescreened, and 20 patients entered screening for eligibility (UC, n = 13; CD, n = 7), from which 16 patients (UC, n = 9; CD, n = 7) were enrolled into the trial. Patients received up to 7 infusions with 600 mg olamkicept every 2 weeks. Six patients did not complete the full observation period, and 10 patients (UC, n = 5; CD, n = 5) completed the 12-week trial (Supplementary Figure 1). Baseline demographics and concomitant medications were similar in patients with UC and CD (Supplementary Table 1).

       Effectiveness

      Based on the assessment of clinical disease activity (Mayo score for UC or CDAI for CD), a total of 3 patients (n = 3/16; 19%)—2 of 9 patients with UC (22%) and 1 of 7 patients with CD (14%)—achieved clinical remission (Figure 1AG). Seven of 16 patients (n = 7/16, 44%) achieved clinical response (UC: n = 5/9, 55%; CD: n = 2/7, 28%) (Figure 1AG). Endoscopic remission was achieved in the 3 patients (1 with CD and 2 with UC) who also achieved clinical remission. Endoscopic response was seen in 6 (37.5%) (CD: 1/7; UC: 5/9) patients. Six (37.5%) patients were in clinical response and endoscopic response at week 14, and 1 (6.2%) patient had clinical response but no endoscopic response.
      Figure thumbnail gr1
      Figure 1Trial design and results of clinical and biomarker assessments. (A) Trial design. Red dots represent the timepoints of collection of whole-blood and mucosal biopsy samples, divided into early (0, 4, and 24 hours after the first olamkicept infusion) and later timepoints (2, 6, and 14 weeks after the first exposure). (B) Integrated charts for the total number of patients with IBD and the numbers and proportions of patients achieving clinical response or remission. (C) Selected colonoscopy pictures from patients with CD and UC at baseline and after 14 weeks of treatment. (D, E) Clinical scores collected throughout the treatment period (0 hours to 14 weeks), with lines representing the mean disease activity scores (± standard deviation) of patients grouped by final status of remission (green) and no remission (orange). Graphs detail the (D) Mayo score for UC and (E) CDAI for CD; dotted lines indicate the disease activity threshold. (F, G) Clinical laboratory markers of disease activity: (F) CRP and (G) fecal calprotectin. (H) Levels of pSTAT3 in response to spiked hyper-IL6 in whole blood drawn after olamkicept infusion, showing target engagement independent of remission status. (I) PK of olamkicept showed no significant differences over time or between remission and nonremission patients. (J, K) Levels of (J) IL6 and (K) sIL6R in patients with and without remission. Each graph is accompanied by the corresponding Mann-Whitney U test results (U score and P value) in the upper right corner.

       Adverse Effects

      The concept of this study does not allow a firm assessment of adverse effects potentially caused by a trans-signaling blockade in humans. Adverse events occurred in 81% of patients, which appears to be high (13/16) (Supplementary Table 2). In general, adverse events were unspecific in nature, not related to drug exposure, and not indicative of severe immune suppression. The most frequently reported adverse events included seasonal upper respiratory tract infections (laryngitis, rhinitis), recurrence of herpes labialis, and skin and subcutaneous disorders such as eczema or erythema. A placebo-controlled, larger clinical study, which is underway (NCT03235752), will be necessary to further investigate whether gp130 trans-signaling blockade does not cause any immune suppression in humans, as was suggested by animal experimentation. No abnormal trends were observed in clinical safety laboratory measurements or vital signs. SAEs were reported in 31% of the patients (5/16) and included, for example, an episode of atrial fibrillation (which was already reported as “intermittent atrial fibrillation” before study drug exposure) or unspecific weakness. Notably, no SAE fell into the category of serious infections, no life-threatening SAEs were reported, and all SAEs were qualified as unlikely to be related to olamkicept (Supplementary Table 2).

       Pharmacokinetics and Pharmacodynamics

      Olamkicept PK and exposure did not change significantly between the first and last dosing events (Figure 1K and Supplementary Figure 3). There was no difference between UC and CD patients in terms of PK. The maximal concentration after each administered dose was reached at the end of infusion, and the terminal half-life was between 3.5 and 7.8 days (Supplementary Figure 3). No statistically significant association of olamkicept PK characteristics with clinical outcome was found (Figure 1G). In this small set of patients, circulating endogenous levels of IL6 and sIL6R were not correlated with clinical outcome (Supplementary Figure 2G and H). Overall reduction of clinical disease activity in the trial population was associated with significant reductions in CRP, fecal calprotectin levels (Figure 1F and G and Supplementary Figure 2), and serum levels of IL6, whereas sIL6R levels were not associated with clinical outcome (Figure 1J and K).
      We stimulated whole-blood samples before and after olamkicept infusion with the recombinant IL6/IL6R fusion protein hyper-IL6, which mimics the physiologic IL6 trans-signaling stimulus ex vivo, to confirm the presence of active olamkicept-related inhibition of trans-signaling in the circulation of treated patients.
      • Fischer M.
      • Goldschmitt J.
      • Peschel C.
      • et al.
      A bioactive designer cytokine for human hematopoietic progenitor cell expansion.
      In all patients, inhibition of STAT3 phosphorylation was already seen 4 hours after the first drug infusion and throughout the entire treatment period (Figure 1H; for primary data, see Supplementary Figure 2I and J). The degree of pSTAT3 inhibition in this assay did not correlate with the clinical outcome (Figure 1H). Production of antibodies directed against olamkicept (anti-drug antibodies) was minimal and transient and was observed in only 3 patients at week 12 and week 15 (Supplementary Table 3). Moreover, no impact of antidrug antibodies on individual PK data or ex vivo STAT3 phosphorylation was seen in these 3 patients.
      sIL6R levels are increased in IBD in remission, and the presence of the coding single-nucleotide polymorphism (SNP) rs2228145 in the IL6R gene (leading to an amino acid substitution of aspartic acid to alanine at amino acid position 358 within the extracellular domain of the IL6R [Asp358Ala]) is strongly associated with high levels of sIL6R. We examined whether olamkicept-treated patients reaching remission carry the rs2228145 SNP, which, via increased sIL6R levels, could amplify the IL6/IL6R neutralization by olamkicept. All patients reaching remission were either homozygous or heterozygous carriers of the rs2228145 SNP (Supplementary Table 4). However, it must be noted that such data are anecdotal, and larger trials should include pharmacogenetic investigation of the rs2228145 SNP and assessment of serum levels of IL6R carrier status.

       Analysis of Early Transcriptome Dynamics and Molecular Target Engagement by Olamkicept

      We assessed the longitudinal dynamics of gene expression using RNA sequencing of peripheral blood samples. We focused on early transcriptome changes, which we hypothesized to be indicative of bona fide IL6 trans-signaling inhibition (target engagement). In total, we identified 861 differentially expressed genes at early timepoints (4 hours and/or 24 hours compared to baseline), with a predominance of down-regulation (715 genes down-regulated vs 146 up-regulated) (Figure 2A). A total of 150 genes were consistently differentially expressed throughout the early stages of treatment at both 4 hours and 24 hours. A heatmap for the 50 most strongly regulated genes (baseline vs 4 hours and 24 hours) is shown in Figure 2B. Notably, we did not find any statistically significant association of early gene expression signatures with overall clinical or laboratory disease parameter outcomes (fecal calprotectin, IL6, CRP, or leukocyte count) (Figure 1IK and Supplementary Figure 2DG) as ex ante indicators of therapy response. To dissect the underlying molecular principle of olamkicept-mediated IL6 trans-signaling blockade, we performed Gene Ontology (GO) term analysis and transcription factor binding site enrichment (TFBS) analysis using the set of 861 genes differentially expressed from early timepoints (4 hours and 24 hours). Among the top GO terms of down-regulated genes were neutrophil degranulation, cytokine-mediated signaling pathway and Fc-gamma receptor signaling pathway, indicating that inhibition of IL6 trans-signaling is associated with a plethora of immune-related pathways (Figure 2C). The top down-regulated TFBS hits were C/EBPalpha and C/EBP-beta, 2 transcription factors that are known to be involved in Ras-dependent downstream IL6 signaling as well as neutrophil differentiation and effector function
      • Ramji D.P.
      • Vitelli A.
      • Tronche F.
      • et al.
      The two C/EBP isoforms, IL6DBP/NFIL6 and CEBP6δ/NFIL63, are induced by IL6β to promote acute phase gene transcription via different mechanisms.
      ,
      • Poli V.
      • Mancini F.P.
      • Cortese R.
      IL-6DBP, a nuclear protein involved in interleukin-6 signal transduction, defines a new family of leucine zipper proteins related to CEBP.
      (Figure 2D). Interestingly, we observed a clear dichotomy of dynamic transcriptional changes toward IL6 trans-signaling inhibition between remission and nonremission patients. IL6 trans-signaling inhibition led to a significantly stronger down-regulation at early timepoints (4 hours and 24 hours) in nonremission patients, whereas remission patients showed a more delayed (weeks 2, 6, and 14) but constant down-regulation of IL6 trans-signaling inhibition in the blood transcriptome data (Figure 2EG). We further investigated whether target engagement signatures were also present in the mucosal transcriptome. In contrast to the peripheral blood data, inhibition of IL6 trans-signaling induced only marginal changes in the mucosal transcriptome at the early timepoints, whereas the predominance of gene expression changes occurred at later timepoints, which is further discussed later in the article (Supplementary Figure 4). Taken together, olamkicept treatment induces rapid transcriptomal changes in circulating immune cells. Pathway inference and the enrichment of C/EBP binding sites in the signature could point to neutrophils as a likely source of the early signal.
      Figure thumbnail gr2
      Figure 2Early transcriptome changes after olamkicept treatment in blood samples. (A) Venn diagrams discriminating the number of significantly down-regulated (blue) and up-regulated (red) genes for early timepoints (4 and 24 hours) after the first olamkicept infusion. (B) Heatmap for the top 50 most differentially expressed genes (based and arranged on P value adjusted for multiple comparisons) when comparing all patients before and at early timepoints (4 and 24 hours) after the first olamkicept infusion, with individual patients ordered by no remission (orange) and remission (green) status. (C, D) Bar graphs depicting (C) GO-enriched terms and (D) enriched TFBS, separately detected for down-regulated (blue) and up-regulated (red) genes at early timepoints (4 and 24 hours) and plotted by the corresponding P value. Terms are accompanied by the number of genes that contributed to the enrichment (in parentheses). (EG) Changes in gene expression discriminated by remission status (green: remission; orange: no remission) for (E) genes enriching C-EBPbeta TFBS and for the differentially expressed genes detected (F) 4 hours and (G) 24 hours after the first treatment with olamkicept. ∗Significant differences in fold-change gene expression between remission and nonremission patients (2-way analysis of variance corrected for multiple comparisons by controlling the false discovery rate, P < .05). DE, differentially expressed; FDR, false discovery rate; rRNA, ribosomal RNA.

       Transcriptional Signatures of the Clinical Response to Olamkicept

      As a next step, we analyzed direct olamkicept-induced changes in the disease-affected intestinal mucosa. As the primary endpoint, we assessed the gene expression changes of TNF, IL1A, REG1A, IL8, IL1B, and LILRA before and at week 14 after therapy induction as a molecular surrogate parameter of mucosal inflammation, as described previously.
      • Häsler R.
      • Sheibani-Tezerji R.
      • Sinha A.
      • et al.
      Uncoupling of mucosal gene regulation, mRNA splicing and adherent microbiota signatures in inflammatory bowel disease.
      We identified that olamkicept treatment led to the reduction of gene expression levels of TNF, IL1A, REG1A, IL8, IL1B, and LILRA in patients achieving clinical remission (Supplementary Figure 5). To identify tissue-specific target engagement in the inflamed mucosa, we assessed STAT3 phosphorylation in sigmoid mucosal biopsy samples using a multipanel immunohistochemistry of pSTAT3 along with markers for intestinal epithelial cells (AE1/3), B cells (CD20), T cells (CD3), macrophages (CD68), and monocytes (CD15). This attempt enabled the differentiation between intestinal epithelial and lamina propria pSTAT3 signaling in response to olamkicept in remission and nonremission patients. We observed that epithelial STAT3 phosphorylation gradually decreased with treatment duration and that this was more pronounced, although not statistically significant, in patients achieving clinical remission (Figure 3A). In line with the reported function of IL6 trans-signaling in immune cells, olamkicept treatment induced a significant decrease of lamina propria STAT3 phosphorylation, which was significantly stronger in patients achieving clinical remission (Figure 3B and C). Although we were not able to discriminate a specific cell type that contributed to decreased lamina propria pSTAT3 signaling (Figure 3D), our overall finding of decreased pSTAT3-positive cells in remission patients (Figure 3E and F) indicates that STAT3 is crucially involved in the mechanism of action of olamkicept. STAT3 has been described to play an essential role in the coordination of epithelial proliferation and regeneration, and blockade of IL6 pathways has been described with increased ulcer formation linked to STAT3 inhibition.
      • Aden K.
      • Breuer A.
      • Rehman A.
      • et al.
      Classic IL-6R signalling is dispensable for intestinal epithelial proliferation and repair.
      ,
      • Mitsuyama K.
      • Matsumoto S.
      • Rose-John S.
      • et al.
      STAT3 activation via interleukin 6 trans-signalling contributes to ileitis in SAMP1/Yit mice.
      ,
      • Atreya R.
      • Billmeier U.
      • Rath T.
      • et al.
      First case report of exacerbated ulcerative colitis after anti-interleukin-6R salvage therapy.
      We therefore assessed the impact of olamkicept on mucosal regeneration measured by Ki-67 staining in the epithelium and lamina propria. We found that olamkicept did not significantly impede mucosal proliferation when assessing Ki67 staining in the context of treatment outcome (remission vs nonremission) or time (baseline vs week 14) (Supplementary Figure 6). To further confirm that reduced STAT3 phosphorylation led to downstream changes in the cellular gene signature, we confirmed mucosal inhibition of STAT3 signaling by assessing the expression levels of genes previously shown to be regulated by STAT3
      • Pickert G.
      • Neufert C.
      • Leppkes M.
      • et al.
      STAT3 links IL-22 signaling in intestinal epithelial cells to mucosal wound healing.
      in the intestinal epithelium (SAA2, STAT3, LRG1, REG3A, SOCS3, PLAG2A, and SAA1) and observed that olamkicept-induced remission was associated with a significant down-regulation of STAT3-dependent genes in the intestinal mucosa (Figure 3G). To understand the underlying transcriptional program of olamkicept beyond known STAT3 gene expression signatures, we followed an unsupervised approach and assessed the gene expression changes of the mucosal transcriptome, comparing patients achieving remission (n = 3) and nonremission patients (n = 13) (Figure 4A). In the mucosal samples, we identified 861 down-regulated genes and 247 up-regulated genes in patients who achieved remission (Supplementary Figure 7). The genes differentially expressed exclusively in patients who achieved remission were sorted by adjusted P value, and the top 50 regulated genes were plotted as a heatmap to illustrate the high number of down-regulated genes in the later stages of treatment (Figure 4B). The heatmap also shows the subtle differences between remission and nonremission patients that can already be detected at baseline. Among the highly down-regulated genes in patients achieving remission were chemokine ligands and receptors (CXCR1 and CSF3R), antimicrobial peptides and receptors (DEFB4API3 and DEFA6), activating receptors of neutrophils (FPR1 and FPR2), and several other regulators of inflammation (s100a8, SAA1, CXCL8, OSM, and TREM1), some of which already have been described to be associated with therapy response to anti-TNF therapy in IBD.
      • West N.R.
      • Hegazy A.N.
      • Owens B.M.J.
      • et al.
      Oncostatin M drives intestinal inflammation and predicts response to tumor necrosis factor–neutralizing therapy in patients with inflammatory bowel disease.
      ,
      • Verstockt B.
      • Verstockt S.
      • Dehairs J.
      • et al.
      Low TREM1 expression in whole blood predicts anti-TNF response in inflammatory bowel disease.
      GO term analysis of the down-regulated genes identified enrichment in the terms neutrophil degranulation, cytokine-mediated signaling pathway, antimicrobial immune response, and protein folding in endoplasmic reticulum (Figure 4C). TFBS enrichment of nuclear factor κB motifs (including RelA and p50 sites) was found in down-regulated transcripts (Supplementary Figure 7). Activation of this pleiotropic transcription factor class is a known component in IBD
      • Rogler G.
      • Brand K.
      • Vogl D.
      • et al.
      Nuclear factor κB is activated in macrophages and epithelial cells of inflamed intestinal mucosa.
      ,
      • Schreiber S.
      • Nikolaus S.
      • Hampe J.
      Activation of nuclear factor κB in inflammatory bowel disease.
      and has been observed downstream of IL6 signaling.
      • Wang L.
      • Walia B.
      • Evans J.
      • et al.
      IL-6 induces NF-κB activation in the intestinal epithelia.
      Notably, when comparing clinical responders to nonresponders, we observed a partial phenocopy of the remission/nonremission transcriptome signature with regard to differentially expressed genes and also common GO terms. This finding indicates that transcriptional changes in clinical response point in the same molecular direction as clinical remission, albeit with a lesser magnitude (Supplementary Figure 8).
      Figure thumbnail gr3
      Figure 3Specific mucosal signaling and transcriptome characteristics of patients achieving clinical remission. (A, B) pSTAT3 quantification in (A) intestinal epithelium and (B) lamina propria collected during treatment (0 hours to 14 weeks), with lines representing the mean score grouped by remission (green) and nonremission (orange) status and error bars calculated based on standard deviation. Each graph is accompanied by the corresponding restricted maximum likelihood estimation results (P value) in the upper right corner. (C) Representative multiplexed immunohistochemistry from sigmoid colon including pSTAT3 (magenta) at 0 hours and 14 weeks after the first olamkicept infusion (scale bar = 100 μm). Colocalization analysis of nuclear pSTAT3 signal with immune cell lineage markers in the lamina propria showed that the trend toward lower pSTAT3 during treatment was driven (D) by a decrease of pSTAT3-positive CD15+ granulocytes and (E) by pSTAT3-positive cells that were not labeled by any of the panel markers (marker negative). Lines represent the mean score grouped by remission (green) and nonremission (orange) status, and error bars were calculated based on standard deviation. Each graph is accompanied by the corresponding restricted maximum likelihood estimation (P value) in the upper right corner. (F) Representative images from a patient who reached remission (0 hours to 14 weeks of olamkicept exposure) confirmed that the overall decrease of nuclear pSTAT3 signal (magenta in upper panels) was largely driven by lower numbers of pSTAT3-positive CD15+ cells and pSTAT3-positive cells not labeled by CD3, CD20, or CD68 (scale bar = 200 μm). (G) Heatmap for the 27 genes of the STAT3 pathway detected in the mucosal biopsy samples of our cohort. Columns are organized by timepoints and average expression levels, and are separated by remission (green, left) and nonremission (orange, right) status. Significant changes are marked with ∗P < .05, and genes names are highlighted in bold.
      Figure thumbnail gr4
      Figure 4Olamkicept-specific transcriptome signature in the intestinal mucosa of patients with IBD achieving clinical remission. (A) Bar graph indicating the number of differentially down-regulated (blue) and up-regulated (red) genes detected in biopsy samples after treatment, separated by remission (left) and nonremission patients (right). (B) Heatmap for the top 50 most differentially expressed genes (base on P value adjusted for multiple comparisons) when comparing remission patients before and after olamkicept treatment. The columns are organized by average patient expression level, sorted by timepoint and separated by remission (green, left) and nonremission (orange, right). (C) Bar graphs depicting GO-enriched terms, separately detected for down-regulated (blue) and up-regulated (red) genes at early timepoints and plotted by the corresponding P value. Terms are accompanied by the number of genes that contributed to the enrichment (in parentheses). (D) MPO quantification in sigmoid mucosa collected during treatment (0 hours to 14 weeks), with lines representing the mean score grouped by remission (green) and nonremission (orange) status and error bars calculated based on standard deviation. The corresponding Mann-Whitney U test results (U score and P value) are shown in the upper right corner. (E) Representative immunohistochemistry results from sigmoid colon stained for MPO at 0 hours and 14 weeks after the first olamkicept infusion. (F) Venn diagrams discriminating the number of significantly up-regulated (red) and down-regulated genes (blue) in mucosal biopsy samples from patients remitting in response to olamkicept, infliximab, or vedolizumab. (G) Bar graphs depicting GO-enriched terms, separately detected for down-regulated (blue) and up-regulated (red) genes at early timepoints and plotted by the corresponding P value. Terms are accompanied by the number of genes that contributed to the enrichment (in parentheses). (H) String analysis depicting 31 genes involved in neutrophil degranulation that are uniquely differentially expressed in mucosal biopsy samples from olamkicept-treated remission patients. DE, differentially expressed; MPO, myeloperoxidase.

       Distinct Transcriptome Signatures of Remission Induced by Olamkicept: Comparison to Infliximab and Vedolizumab

      To investigate whether the identified GO term neutrophil degranulation is a functional process uniquely attributed to olamkicept because of decreased neutrophil abundance in the intestinal mucosa, we enumerated neutrophilic granulocytes using myeloperoxidase staining. To our surprise, we did not identify significant differences in the absolute number of neutrophilic granulocytes in the inflamed mucosa between remission and nonremission patients (Figure 4D and E) at early timepoints. We further investigated whether the observed mucosal transcriptome signature is specific to olamkicept in a comparison with a previously described cohort of IBD patients, from which data before and after induction therapy (week 0 vs week 14) with anti-TNF (infliximab, n = 20) and anti–α4-β7-integrin (vedolizumab, n = 18) were available.
      • Zeissig S.
      • Rosati E.
      • Dowds C.M.
      • et al.
      Vedolizumab is associated with changes in innate rather than adaptive immunity in patients with inflammatory bowel disease.
      These patients have been examined under a very similar clinical protocol including multiple endoscopies (NCT02694588). We specifically compared mucosal transcript signatures associated with clinical remission induced by infliximab (n = 14) and vedolizumab (n = 9) with olamkicept-induced clinical remission (n = 3). In total, we observed 494 genes that were uniquely differentially expressed between week 0 and week 14 only in olamkicept-treated patients with IBD. Of those, 337 genes were down-regulated, and 157 genes were up-regulated (Figure 4F). Within GO terms, we identified a subset of genes that uniquely contributed to the GO term neutrophil degranulation only in olamkicept-induced clinical remission (Figure 4G and H). Hence, this analysis implies that olamkicept uniquely affects functional modalities of neutrophilic granulocytes during disease amelioration (Figure 4H). Altogether, our data show that IL6 trans-signaling inhibition by olamkicept induces a mucosal transcriptome signature of clinical remission that is clearly distinguishable from other established targeted therapies in IBD.

      Discussion

      The FUTURE study was designed as an exploratory phase 2a trial to investigate the target engagement, molecular effects, clinical effectiveness, safety, and PK of the selective IL6 trans-signaling inhibitor olamkicept as a treatment for patients with active IBD. The drug was well tolerated in all 16 treated individuals, similar to the results of the phase I trials (EudraCT no. 2012-005142-38 and 2013-004208-20). In particular, neither signs of significant immunosuppression nor intestinal perforations were observed.
      • Danese S.
      • Vermeire S.
      • Hellstern P.
      • et al.
      Randomised trial and open-label extension study of an anti-interleukin-6 antibody in Crohn’s disease (ANDANTE I and II).
      We confirmed uniform functional activity of olamkicept in all treated patients with IBD through inhibition of hyper-IL6 (IL6/sIL6R fusion protein)–induced trans-signaling in an ex vivo whole-blood assay. The clinical results of the trial suggest effectiveness during short-term exposure. Three patients (1 with CD and 2 with UC) went into remission, as evidenced by complete and fast resolution of clinical symptoms, inflammatory markers, and endoscopic activity. Overall, 7 patients (44%) showed a clinically significant level of response. The majority of the responders (5 patients—ie, a 55% response rate) were found in the UC population. It should be noted that the study was not designed to assess clinical efficacy and that the patient population was heterogenous with regard to disease duration and pretreatment of biologics. Given the comparable PK and a highly uniform pattern of inhibition of STAT3 phosphorylation, differences in biologic activity of olamkicept do not explain the clinical differences in effectiveness. The overlay between clinical improvement, endoscopic response, decrease in inflammation markers, and molecular impact of trans-signaling blockade both preceding and paralleling clinical and endoscopic responses could point to a selected subpopulation of superresponders to the therapeutic principle. However, in this early stage of clinical exploration, patients recruited typically are complex, and treatment success could also represent a lesser degree of complexity.
      Although gp130 trans-signaling has been widely suggested to actively drive chronic intestinal inflammation in animal models of IBD, the involved cell populations and the molecular importance of this pathophysiologic process in humans are unknown. RNA sequencing of whole-blood samples from olamkicept-treated patients showed a strong alteration of the blood transcript signature starting as early as 4 hours after the first exposure. This early signature was not related to changes in cellular composition as assessed by differential blood counts (data not shown) and likely demonstrates target engagement of the active compound. Target engagement of olamkicept-mediated IL6 trans-signaling inhibition appeared to be different from target engagement signatures of conventional IL6R blockade, as suggested by indirect comparison of blood transcriptome signatures from patients with IBD (n = 3) treated outside of this study with anti-IL6R antibody (tocilizumab) (Supplementary Figure 9). These differences may reflect different downstream effects between classic and trans-signaling blockade of the IL6 pathway and may also include inhibitory effects of olamkicept on other gp-130–related trans-signaling events (ie, by IL11).
      To assess whether mucosal target engagement of olamkicept treatment is associated with clinical outcome in IBD, we assayed (epithelial/lamina propria) STAT3 phosphorylation in parallel with mucosal gene transcription dynamics. Lamina propria–specific STAT3 phosphorylation and targeted assessment of mucosal STAT3 target genes showed the strong inhibitory effect of olamkicept treatment on the STAT3 pathway, which was significantly augmented in patients achieving clinical remission and is in line with the postulated effect of IL6 trans-signaling via STAT3 activation predicted from multiple murine studies.
      • Atreya R.
      • Mudter J.
      • Finotto S.
      • et al.
      Blockade of interleukin 6 trans signaling suppresses T-cell resistance against apoptosis in chronic intestinal inflammation: evidence in Crohn disease and experimental colitis in vivo.
      ,
      • Mitsuyama K.
      • Matsumoto S.
      • Rose-John S.
      • et al.
      STAT3 activation via interleukin 6 trans-signalling contributes to ileitis in SAMP1/Yit mice.
      Notably, we did not see suppressive effects of olamkicept on signatures of epithelial cell regeneration, as assessed by Ki67 staining, which is also in line with previous preclinical studies showing no inhibitory effect of olamkicept on epithelial wound healing.
      • Aden K.
      • Breuer A.
      • Rehman A.
      • et al.
      Classic IL-6R signalling is dispensable for intestinal epithelial proliferation and repair.
      To understand the olamkicept-specific mechanism of action, we compared mucosal transcriptome signatures of patients entering remission under olamkicept treatment with remission signatures induced by vedolizumab or infliximab, respectively. We could define a unique set of differentially regulated genes in response to olamkicept with an enrichment of signals relevant to neutrophil degranulation (GO term). Neutrophil degranulation comprises genes involved in the orchestrated mobilization and exocytosis of granules, which contain proteins with a wide spectrum of functions, including s100a8/s100a9 proteins, which are used as fecal biomarker (calprotectin) for disease activity in IBD. Hence, it remains to be elucidated whether the olamkicept-specific remission signatures at the end of therapy indicate a drug-specific pattern of mucosal healing or, rather, describe a different state of a common molecular pathway that causally contributes to mucosal healing in IBD. In this context, further systematic studies are needed to distinguish the unique and overlapping mechanisms of action of targeted therapies used to achieve mucosal healing in IBD.
      Taken together, we show in this phase 2a trial that the biology of trans-signaling inhibition by olamkicept can be translated from mouse models into patients. A clear trace of biologic activity was evident in all patients. Olamkicept has a succinct molecular impact on disease pathophysiology that largely parallels what has been expected from murine models and that is associated with clinical effectiveness in a subset of patients. The observed clinical effectiveness in a subset of patients was closely linked to distinct molecular signatures and therefore strongly supports the further clinical development of the drug. The clinical efficacy of olamkicept is presently investigated in a multicenter, randomized, prospective, placebo-controlled phase 2 trial in patients with moderately to severely active ulcerative colitis (NCT03235752).

      Acknowledgments

      The authors thank K. Greve, M. Rohm, M. Hansen, S. Kock, D. Oelsner, S. Baumgarten, M. Reffelmann, M. Schlapkohl, N. Braun, T. Wesse, M. Basso, Y. Dolschanskaya, X. Yi, C. Lancken, and M. Vollstedt for excellent technical assistance.

      CRediT Authorship Contributions

      Stefan Schreiber, MD (Conceptualization: Lead; Funding acquisition: Lead; Investigation: Lead; Writing – original draft: Lead); Konrad Aden, MD (Data curation: Lead; Formal analysis: Lead; Investigation: Lead; Writing – original draft: Lead); Joana P. Bernardes, MD (Formal analysis: Lead); Claudio Conrad, MD (Formal analysis: Lead); Florian Tran, MD (Data curation: Equal); Hanna Höper, MD student (Formal analysis: Supporting); Valery Volk, PhD (Formal analysis: Equal); Neha Mishra, PhD (Formal analysis: Supporting); Johanna Ira Blase, MSc (Data curation: Supporting); Susanna Nikolaus, MD (Investigation: Supporting); Johannes Bethge, MD (Investigation: Supporting); Tanja Kühbacher, MD (Investigation: Supporting); Christoph Roecken, MD (Formal analysis: Supporting); Minhu Chen, MD (Investigation: Supporting); Ian Cottingham, PhD (Formal analysis: Equal; Methodology: Equal); Niclas Petri, PhD (Formal analysis: Equal; Methodology: Equal); Birgitte B. Rasmussen, PhD (Data curation: Equal; Investigation: Equal; Methodology: Equal); Juliane Lokau, PhD (Formal analysis: Supporting); Lennart Lenk, PhD (Visualization: Equal); Christoph Garbers, PhD (Formal analysis: Equal; Methodology: Equal); Friedrich Feuerhake, MD (Data curation: Lead); Stefan Rose-John, PhD (Investigation: Equal; Methodology: Equal; Supervision: Equal; Writing – original draft: Equal); Georg H. Waetzig, PhD (Conceptualization: Equal; Data curation: Equal; Investigation: Equal; Methodology: Equal; Project administration: Equal); Philip Rosenstiel, MD (Conceptualization: Lead; Investigation: Lead; Supervision: Lead; Writing – original draft: Lead).

      Supplementary Materials and Methods

       Multiplexed Immunohistochemistry

      Multiplexed immunohistochemistry was performed on 3-μm sections cut from the formalin-fixed, paraffin-embedded tissue blocks from diagnostic biopsy samples obtained before, 4 hours, 24 hours, 2 weeks, 6 weeks, and 14 weeks after treatment for n = 16 patients (n = 3 remission and n = 13 nonremission). Assay optimization was performed by following the manufacturer’s instructions (OPAL Multiplex Immunohistochemistry Assay Development Guide, Akoya Bioscience). Slides were deparaffinized, the initial antigen retrieval was performed by microwave cooking in Tris-buffered saline at pH 9, and blocking of unspecific protein binding was performed using Protein Block Serum-free solution (Agilent/Dako). Subsequent antigen retrieval steps, including deactivation of the preceding antibody binding, were performed by microwave cooking either in Tris-buffered saline at pH 9 or citrate buffer at pH 6. Consecutive immunohistochemistry staining using the OPAL 7-plex fluorescence system was performed according to the manufacturer’s instructions with the following primary antibodies: anti-CD3 (DAKO, clone A0452), anti-CD20 (DAKO, clone M0755), anti-Pan-Cytokeratin (DAKO, clone AE1/3), anti-CD68 (DAKO, clone PGM-1), anti-pSTAT3 (Cell Signaling, clone D3A7), anti-CD15 (Thermo Fisher Scientific, clone 4E10), and nuclear staining DAPI (Akoya). The following fluorophores in combination with the tyramide signal amplification system were used for detection of bound antibodies: Opal 520, Opal 540, Opal 570, Opal 620, Opal 650, or Opal 690. Fluoromount-G mounting medium (Thermo Fisher Scientific) was applied to cover slides before imaging.

       Multispectral Image Analysis and Quantitative Evaluation

      Whole-slide image scanning was done at 20× magnification using the Vectra Polaris instrument (Akoya Bioscience). Three-channel fluorescent whole-slide images were used to verify regions of interest (full biopsy excluding processing artifacts) for subsequent scanning of image stacks at 20× magnification across the visible spectrum (420–720 nm) for multispectral imaging. Spectral libraries were generated using single-stained scans of colon tissue for each reagent, and image deconvolution was performed with the inForm Advanced Image Analysis software (inForm, version 2.4.10; Akoya). Training for the machine learning–based segmentation, tissue classification, and cell phenotyping algorithms was performed in batches composed of 2 cases, including all timepoints and pairing all cases with remission with nonremission cases, using the inForm software on a total of >70 representative images covering the full spectrum of expected variability regarding staining intensities and cell densities. Visual quality control of analyzed regions of interests was performed by reviewing all composite images, confirming the plausibility of the quantitative phenotyping result. Results were exported in text format and analyzed using the R packages Phenoptr and PhenoptrReports, using the R package, version 0.2.3, Rstudio (Rstudio, version 1.1.456).

       Data Set for RNA Sequencing

      Whole-blood and mucosa biopsy samples were collected for 16 patients, followed by transcriptome isolation and high-throughput RNA sequencing. Whole-RNA libraries were prepared using TruSeq RNA Library Prep technology (Illumina), and in total, 104 blood and 105 biopsy samples were sequenced using HiSeq 3000 and HiSeq 4000 sequencers (Illumina), respectively. The reads were paired-end reads that were 75 base pairs long (2 × 75 base pairs).

       RNA-Sequencing Pipeline

      Gene expression for each patient throughout the course of treatment was systematically accessed. We used our in-house RNA-sequencing pipeline to map and align the sequenced data (https://github.com/nf-core/rnaseq). The workflow processed the raw data from the sequencer with FastQC, version 0.11.3,
      • Wingett S.W.
      • Andrews S.
      FastQ Screen: a tool for multi-genome mapping and quality control.
      and Trimgalore, version 0.4.4; aligned the reads with STAR, version 2.5.2b
      • Dobin A.
      • Davis C.A.
      • Schlesinger F.
      • et al.
      STAR: ultrafast universal RNA-seq aligner.
      ; and generated gene counts with featurecounts, version 1.5.2,
      • Liao Y.
      • Smyth G.K.
      • Shi W.
      featureCounts: an efficient general purpose program for assigning sequence reads to genomic features.
      and StringTie, version 1.3.3b.
      • Pertea M.
      • Pertea G.M.
      • Antonescu C.M.
      • et al.
      StringTie enables improved reconstruction of a transcriptome from RNA-seq reads.
      Quality control was assessed throughout with RSeqQC
      • Wang L.
      • Wang S.
      • Li W.
      RSeQC: quality control of RNA-seq experiments.
      ; dupRadar
      • Sayols S.
      • Scherzinger D.
      • Klein H.
      dupRadar: a Bioconductor package for the assessment of PCR artifacts in RNA-Seq data.
      ; Preseq
      • Daley T.
      • Smith A.D.
      Predicting the molecular complexity of sequencing libraries.
      ; and MultiQC, version 1.4.
      • Ewels P.
      • Magnusson M.
      • Lundin S.
      • et al.
      MultiQC: summarize analysis results for multiple tools and samples in a single report.
      The pipeline output indicated that all sequenced samples mapped well to the GRCh38 Homo sapiens genome (Genome Reference Consortium Human Build 38 and GenBank Assembly ID: GCA_000001405.27), with the total reads per sample aligning, on average, by 70.87% to the reference genome for blood samples and 91.36% for biopsy samples. In our data set, there were both female and male patients; however, only female patients entered remission. We did not conclude that response to olamkicept therapy might be affected by a patient’s sex. Thus, to avoid potential biases in the differential expression, we decided to remove all reads that mapped to Y chromosome genes.

       RNA-Sequencing Statistical Analysis

      Differentially expressed genes were identified by comparing expression profiles throughout the course of treatment, whether by grouping or not the patients into specific categories (eg, patients under clinical remission). Statistical analysis was performed using R, version 3.6.3, unless stated otherwise.
      We set up the analysis to formally test the data as intention to treat or per protocol. Intention-to-treat analysis takes into account the data available without discarding patients who dropped out of the study, thus using the last available point as the last accessible data. Per-protocol analysis discards the missing values for patients who dropped out of the study, thus using only the data available at the end of the course of treatment for data analysis. For the target engagement analysis, we pooled all patient samples and compared gene expression at baseline and after olamkicept treatment, focusing on early timepoints after the first dosage (4 hours and 24 hours). We expected all patients to share similar expression variations, because they responded to the treatment mechanism of action. Moreover, the variations at early timepoints should be stronger in blood samples because of the intravenous administration of olamkicept, whereas intestinal mucosal tissue would not be immediately affected. To further examine the effect of olamkicept in the treatment of IBD, we separated patients who entered clinical remission from those who did not. We aimed to ascertain not only specific signatures of the treatment but also common changes in expression that may mirror mucosal healing by exploring gene expression changes on mucosal biopsy samples that were unique to remission patients.

       Differentially Expressed Genes

      To identify differentially expressed genes between patients and timepoints, we used the DESeq2 R package, version 1.26.0,
      • Love M.I.
      • Huber W.
      • Anders S.
      Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2.
      and ImpulseDE2 R package, version 1.12.0.
      • Fischer D.S.
      • Theis F.J.
      • Yosef N.
      Impulse model-based differential expression analysis of time course sequencing data.
      DESeq2 was used for executing pairwise comparisons between baseline and timepoints after treatment, whereas ImpulseDE2 was used to identify genes with transient differential expression in the longitudinal data set. Both statistical tools are based on a negative binomial distribution model with dispersion trend smoothing. We also determined the normalized read counts per sample by estimating size factors to control for library size, followed by a log2 transformation of the raw count data using DESeq2.

       Heatmaps

      To visualize the data, we used normalized read counts of the differentially expressed genes of interest to plot heatmaps with the stats R package, version 3.6.3.
      R Core Team. R: A language and environment for statistical computing.
      For the heatmaps, we put in the normalized read counts of the genes of interest and used a Spearman ranked correlation to compute a distance matrix, which was then used to depict changes in gene expression throughout the treatment by the variations in Z-scores.

       Gene Ontology and Transcription Factor Binding Site Enrichment Analysis

      For other genes of interest, we investigated their functional characteristics by assessing GO term enrichment for biological processes (GO terms) and by identifying overrepresented TFBSs. For GO term enrichment analysis, we used the TopGO package for R, version 2.38.1,
      • Alexa A.
      • Rahnenfuhrer J.
      Gene set enrichment analysis with topGO.
      and identified the most significant terms by using the Fisher elimination method. Overrepresented TFBSs were detected by using the InnateDB database and a hypergeometric algorithm.
      • Breuer K.
      • Foroushani A.K.
      • Laird M.R.
      • et al.
      InnateDB: systems biology of innate immunity and beyond—recent updates and continuing curation.
      For genes bound to neutrophil degranulation GO terms, we performed a STRING analysis
      • Pertea M.
      • Pertea G.M.
      • Antonescu C.M.
      • et al.
      StringTie enables improved reconstruction of a transcriptome from RNA-seq reads.
      using the Cytoscape, version 3.7.2, platform.
      • Shannon P.
      • Markiel A.
      • Ozier O.
      • et al.
      Cytoscape: a software environment for integrated models of biomolecular interaction networks.

       Clinical Data Statistical Analysis

      The induction of clinical remission was measured by clinical disease parameters for CD (CDAI of <150) and UC (Mayo score of <2, bleeding score of 0, and endoscopy of <1). The CDAI score comprises measurements of patient well-being, abdominal pain, number of liquid or soft stools, abdominal mass, and other complications. The Mayo score assesses the severity of UC, with measurements of stool frequency, rectal bleeding, endoscopic findings, and overall physical assessment. Other clinical scores and parameters, like the Harvey-Bradshaw index score, fecal calprotectin (mg/g of stool), leukocyte count, CRP (mg/L), and STAT3 phosphorylation were also analyzed. For all scores and parameters, we compared the patients entering clinical remission with patients who did not by using a Mann-Whitney test for nonparametric data and a Tukey range test for multiple comparisons.

       Anti-Drug Antibody Detection

      In a first step, samples were screened for anti-olamkicept antibodies (predefined screening cut point [SCP] of 1.12× blank-enhanced chemiluminescence units [ECLUs]). Samples with a mean ECLU less than the SCP were considered negative. Samples with a mean ECLU greater than the SCP were retested in a confirmatory assay. Samples with a confirmed ECLU greater than the SCP (confirmed positive) were retested for neutralizing capacity in a third assay with and without anti-olamkicept depletion by the addition of olamkicept and titer dilution to estimate anti-olamkicept levels. The confirmatory cutpoint was an ECLU depletion of ≥16.3% by addition of olamkicept. Levels of anti-olamkicept antibodies were reported descriptively.

       Inflammatory Biomarkers

      Blood and feces sampling for the assessment of exploratory biomarkers (IL6, sIL6R, STAT3 phosphorylation, CRP, and fecal calprotectin) was performed at every visit. Serum sIL6R was measured by enzyme-linked immunosorbent assay (ELISA) (R&D Systems) according to the manufacturer’s instructions. STAT3 phosphorylation in peripheral blood mononuclear cells was measured using PathScan ELISAs (Cell Signalling Technology) for phosphorylated and total STAT3 according to the manufacturer’s instructions. The assay was performed before and 4 hours after olamkicept administration, as well as at 4 weeks and 14 weeks (before drug administration). Fecal calprotectin and CRP were measured using standard clinical routine procedures.

       Immunohistochemistry

      Immunohistochemistry in mucosal biopsy samples was carried out using monoclonal antibodies directed against pSTAT3 (dilution 1:50; D3A7, Cell Signaling Technology, no. 13684) and Ki-67 (dilution 1:300; SP-6, Thermo Fisher Scientific, no. RM-9106-S) using the autostainer Bond Max System (Leica Microsystems). Pretreatment was performed with ER2 (Leica Biosystems) for 20 minutes. For visualization of the immunoreaction, the Bond Polymer Refine Detection Kit was used (DS 9800, brown labeling; Novocastra; Leica Microsystems, no. DS9800).
      Figure thumbnail fx2
      Supplementary Figure 1Trial description. In total, 20 patients (UC, n = 13; CD, n = 7) were screened, and 16 patients (UC, n = 9; CD, n = 7) received at least 1 dose of olamkicept and were therefore included in the safety analysis cohort. A total of 10 patients (UC, n = 5; CD, n = 5) received olamkicept over the timeframe of 14 weeks. Six patients withdrew from the study either because of worsening of disease (n = 5) or withdrawal of informed consent (n = 1).
      Figure thumbnail fx3
      Supplementary Figure 2Individual patient data. (AJ) Graphs detailing the clinical scores of individual patients during the clinical trial: 0 hours, 4 hours, 24 hours, 2 weeks, 6 weeks, and 14 weeks after the first drug administration. Bold lines correspond to the average scores and fine lines to each patient’s score, discriminated by remission (green) and nonremission (orange) status. (A) Mayo scores for patients with UC and (B) CDAI and (C) Harvey-Bradshaw index scores for patients with CD; included are activity thresholds depicted as dotted lines. (D) Leukocyte counts, together with (E) blood CRP concentration and (F) fecal calprotectin concentration relative to the time before drug administration. Serum ELISA for (G) IL6 and (H) sIL6R and ex vivo stimulation assay for pSTAT3 treated with (I) hyper-IL6 or (J) control treatment.
      Figure thumbnail fx4
      Supplementary Figure 3PK. Graphs depict (A) the time course of the mean serum concentration of olamkicept and (B) the time courses of serum concentrations of olamkicept in all patients after the first (black, solid line) and the last (blue, broken line) administration.
      Figure thumbnail fx5
      Supplementary Figure 4Target engagement in patients’ mucosal biopsy samples after olamkicept treatment. (A) The number of differentially down-regulated (blue) and up-regulated (red) genes detected after treatment in mucosal samples of all patients (4 hours, 24 hours, 2 weeks, 6 weeks, and 14 weeks after the first drug administration). (B, C) Venn diagrams discriminating the number of significant (B) down-regulated genes in blue and (C) up-regulated genes in red for all timepoints after olamkicept treatment. (D) Heatmap for the top 50 most differentially expressed genes (based on P value adjusted for multiple comparisons) at different timepoints (4 hours, 6 weeks, and 14 weeks after first olamkicept treatment), with individual patients ordered by nonremission (orange) and remission (green) status. (E, F) Bar graphs depict (E) GO-enriched terms and (F) enriched TFBS, detected for down-regulated (blue) and up-regulated (red) genes. Each term is accompanied by the number of genes bound to the enrichment and the corresponding P value corrected for FDR. DE, differentially expressed; FDR, false discovery rate; NF, nuclear factor.
      Figure thumbnail fx6
      Supplementary Figure 5Primary endpoint 6-gene score. Representation of the expression patterns for the 6 genes of interest for patients treated with olamkicept. Remission patients are depicted in green and nonremission patients in orange. Circumference represents the mean for grouped remission or nonremission patients, and the error bars represent the corresponding standard deviations (SDs). (A) The average gene score was found by calculating the mean of the normalized read counts for the 6 genes of interest in the infliximab cohort. The dashed line represents the average minus SD of nonremission patients. (B) Depiction of the average gene score for all patients treated with olamkicept. (C) The ranked gene score was found by ranking the normalized expression of each gene across samples and calculating the rank average for remission and nonremission patients. The dashed line represents the average minus SD of nonremission patients. (D) Depiction of the ranked gene score for all patients treated with olamkicept.
      Figure thumbnail fx7
      Supplementary Figure 6Ki67 staining in remission and nonremission patients. Ki-67 quantification in the (A) intestinal epithelium and (B) lamina propria collected during treatment (0 hours to 14 weeks), with lines representing the mean score grouped by remission (green) and nonremission (orange) status and error bars calculated based on standard deviation. Each graph is accompanied by the corresponding Mann-Whitney U test results (U score and P value) in the upper right corner. (C) Representative immunohistochemistry from sigmoid colon stained against Ki-67 at 0 hours and after the end of olamkicept exposure (14 weeks).
      Figure thumbnail fx8
      Supplementary Figure 7Mucosal signatures in remission patients. (A) Venn diagrams indicating the number of significantly down-regulated and up-regulated genes specific to remission (green) and nonremission (orange). (B) Bar graphs depict enriched TFBS, separately detected for down-regulated (blue) and up-regulated (red) genes at early timepoints and plotted by corresponded P value. Each term is accompanied by the number of genes bound to the enrichment and the corresponding P value corrected for FDR. DE, differentially expressed; FDR, false discovery rate.
      Figure thumbnail fx9
      Supplementary Figure 8Mucosal signatures in responder patients. (A) Bar graph indicate the number of differentially down-regulated (blue) and up-regulated (red) genes detected in biopsy samples after treatment, separated by responder (left) and nonresponder patients (right). (B) Venn diagrams depict the overlap between differentially expressed genes (down-regulated/up-regulated) in the responder and remission populations. (C) Heatmap for the top 50 most differentially expressed genes (based on P value adjusted for multiple comparisons) when comparing responder patients before and after olamkicept treatment. Columns are organized by average patient expression, sorted by timepoint and separated by response (green, left) and no response (orange, right). (D) Bar graphs depict GO-enriched terms, separately detected for down-regulated (blue) and up-regulated (red) genes at early timepoints and plotted by the corresponding P value. Terms are accompanied by the number of genes that contributed to the enrichment (in parentheses). (E) Bar graphs depict TFBS-enriched terms, separately detected for down-regulated (blue) and up-regulated (red) genes at early timepoints and plotted by the corresponding P value. Terms are accompanied by the number of genes that contributed to the enrichment (in parentheses). DE, differentially expressed.
      Figure thumbnail fx10
      Supplementary Figure 9Target engagement signature with olamkicept and tocilizumab. Heatmap for the top 50 differentially expressed genes at early timepoints after olamkicept treatment. The differential expression levels of these top 50 genes were assessed in the blood transcriptome (baseline, +4 hours, and +24 hours) of n = 3 patients with IBD (2 UC, 1 CD) treated with anti-IL6R antibody (tocilizumab).

      Supplementary Material

      • Supplementary Table 4

        Assessment of the Coding SNP rs2228145 Within the IL6R Gene

        NOTE. SNP_A_ratio refers to the ratio of reads (%) with the wild-type (WT) variant (A); SNP_C_mut_ratio refers to the ratio of reads with the mutated variant (C). ReadCounts_A refers to the number of reads with the WT variant (A), and ReadCounts_C_mut refers to the number of reads with the mutated variant (C). The + symbol corresponds to a sample that had the same SNP within 9 nucleotides of the distance (position 154454585).

      References

        • Schultze J.L.
        • Rosenstiel P.
        Systems medicine in chronic inflammatory diseases.
        Immunity. 2018; 48: 608-613
        • Neurath M.F.
        Targeting immune cell circuits and trafficking in inflammatory bowel disease.
        Nat Immunol. 2019; 20: 970-979
        • Sands B.E.
        • Peyrin-Biroulet L.
        • Loftus E.V.
        • et al.
        Vedolizumab versus adalimumab for moderate-to-severe ulcerative colitis.
        N Engl J Med. 2019; 381: 1215-1226
        • Aden K.
        • Breuer A.
        • Rehman A.
        • et al.
        Classic IL-6R signalling is dispensable for intestinal epithelial proliferation and repair.
        Oncogenesis. 2016; 5e270
        • Garbers C.
        • Aparicio-Siegmund S.
        • Rose-John S.
        The IL-6/gp130/STAT3 signaling axis: recent advances towards specific inhibition.
        Curr Opin Immunol. 2015; 34: 75-82
        • Garbers C.
        • Heink S.
        • Korn T.
        • et al.
        Interleukin-6: designing specific therapeutics for a complex cytokine.
        Nat Rev Drug Discov. 2018; 17: 395-412
        • Kang S.
        • Tanaka T.
        • Narazaki M.
        • et al.
        Targeting interleukin-6 signaling in clinic.
        Immunity. 2019; 50: 1007-1023
        • Mudter J.
        • Amoussina L.
        • Schenk M.
        • et al.
        The transcription factor IFN regulatory factor–4 controls experimental colitis in mice via T cell–derived IL-6.
        J Clin Invest. 2008; 118: 2415-2426
        • Nikolaus S.
        • Waetzig G.H.
        • Butzin S.
        • et al.
        Evaluation of interleukin-6 and its soluble receptor components sIL-6R and sgp130 as markers of inflammation in inflammatory bowel diseases.
        Int J Colorectal Dis. 2018; 33: 927-936
        • Ito H.
        • Takazoe M.
        • Fukuda Y.
        • et al.
        A pilot randomized trial of a human anti-interleukin-6 receptor monoclonal antibody in active Crohn’s disease.
        Gastroenterology. 2004; 126: 989-996
        • Danese S.
        • Vermeire S.
        • Hellstern P.
        • et al.
        Randomised trial and open-label extension study of an anti-interleukin-6 antibody in Crohn’s disease (ANDANTE I and II).
        Gut. 2019; 68: 40-48
        • Waetzig G.H.
        • Rose-John S.
        Hitting a complex target: an update on interleukin-6 trans-signalling.
        Expert Opin Ther Targets. 2012; 16: 225-236
        • Riethmueller S.
        • Somasundaram P.
        • Ehlers J.C.
        • et al.
        Proteolytic origin of the soluble human IL-6R in vivo and a decisive role of N-glycosylation.
        PLoS Biol. 2017; 15e2000080
        • Rose-John S.
        • Heinrich P.C.
        Soluble receptors for cytokines and growth factors: generation and biological function.
        Biochem J. 1994; 300: 281-290
        • Wolf J.
        • Waetzig G.H.
        • Chalaris A.
        • et al.
        Different soluble forms of the interleukin-6 family signal transducer gp130 fine-tune the blockade of interleukin-6 trans-signaling.
        J Biol Chem. 2016; 291: 16186-16196
        • Scheller J.
        • Chalaris A.
        • Schmidt-Arras D.
        • et al.
        The pro- and anti-inflammatory properties of the cytokine interleukin-6.
        Biochim Biophys Acta. 2011; 1813: 878-888
        • Atreya R.
        • Mudter J.
        • Finotto S.
        • et al.
        Blockade of interleukin 6 trans signaling suppresses T-cell resistance against apoptosis in chronic intestinal inflammation: evidence in Crohn disease and experimental colitis in vivo.
        Nat Med. 2000; 6: 583-588
        • Mitsuyama K.
        • Matsumoto S.
        • Rose-John S.
        • et al.
        STAT3 activation via interleukin 6 trans-signalling contributes to ileitis in SAMP1/Yit mice.
        Gut. 2006; 55: 1263-1269
        • Rose-John S.
        The soluble interleukin 6 receptor: advanced therapeutic options in inflammation.
        Clin Pharmacol Ther. 2017; 102: 591-598
        • Jostock T.
        • Müllberg J.
        • Özbek S.
        • et al.
        Soluble gp130 is the natural inhibitor of soluble interleukin-6 receptor transsignaling responses.
        Eur J Biochem. 2001; 268: 160-167
        • Sodenkamp J.
        • Waetzig G.H.
        • Scheller J.
        • et al.
        Therapeutic targeting of interleukin-6 trans-signaling does not affect the outcome of experimental tuberculosis.
        Immunobiology. 2012; 217: 996-1004
        • Hoge J.
        • Yan I.
        • Jänner N.
        • et al.
        IL-6 controls the innate immune response against Listeria monocytogenes via classical IL-6 signaling.
        J Immunol. 2013; 190: 703-711
        • US Food and Drug Administration
        Ulcerative colitis: clinical trial endpoints guidance for industry.
        (Published August 2016. Accessed April 6, 2021)
        • Häsler R.
        • Sheibani-Tezerji R.
        • Sinha A.
        • et al.
        Uncoupling of mucosal gene regulation, mRNA splicing and adherent microbiota signatures in inflammatory bowel disease.
        Gut. 2017; 66: 2087-2097
        • Fischer M.
        • Goldschmitt J.
        • Peschel C.
        • et al.
        A bioactive designer cytokine for human hematopoietic progenitor cell expansion.
        Nat Biotechnol. 1997; 15: 142-145
        • Ramji D.P.
        • Vitelli A.
        • Tronche F.
        • et al.
        The two C/EBP isoforms, IL6DBP/NFIL6 and CEBP6δ/NFIL63, are induced by IL6β to promote acute phase gene transcription via different mechanisms.
        Nucleic Acids Res. 1993; 21: 289-294
        • Poli V.
        • Mancini F.P.
        • Cortese R.
        IL-6DBP, a nuclear protein involved in interleukin-6 signal transduction, defines a new family of leucine zipper proteins related to CEBP.
        Cell. 1990; 63: 643-653
        • Atreya R.
        • Billmeier U.
        • Rath T.
        • et al.
        First case report of exacerbated ulcerative colitis after anti-interleukin-6R salvage therapy.
        World J Gastroenterol. 2015; 21: 12963-12969
        • Pickert G.
        • Neufert C.
        • Leppkes M.
        • et al.
        STAT3 links IL-22 signaling in intestinal epithelial cells to mucosal wound healing.
        J Exp Med. 2009; 206: 1465-1472
        • West N.R.
        • Hegazy A.N.
        • Owens B.M.J.
        • et al.
        Oncostatin M drives intestinal inflammation and predicts response to tumor necrosis factor–neutralizing therapy in patients with inflammatory bowel disease.
        Nat Med. 2017; 23: 579-589
        • Verstockt B.
        • Verstockt S.
        • Dehairs J.
        • et al.
        Low TREM1 expression in whole blood predicts anti-TNF response in inflammatory bowel disease.
        EBioMedicine. 2019; 40: 733-742
        • Rogler G.
        • Brand K.
        • Vogl D.
        • et al.
        Nuclear factor κB is activated in macrophages and epithelial cells of inflamed intestinal mucosa.
        Gastroenterology. 1998; 115: 357-369
        • Schreiber S.
        • Nikolaus S.
        • Hampe J.
        Activation of nuclear factor κB in inflammatory bowel disease.
        Gut. 1998; 42: 477-484
        • Wang L.
        • Walia B.
        • Evans J.
        • et al.
        IL-6 induces NF-κB activation in the intestinal epithelia.
        J Immunol. 2003; 171: 3194-3201
        • Zeissig S.
        • Rosati E.
        • Dowds C.M.
        • et al.
        Vedolizumab is associated with changes in innate rather than adaptive immunity in patients with inflammatory bowel disease.
        Gut. 2019; 68: 25-39