Objective Interleukin 13 (IL-13) is thought to play a key role as an effector cytokine in UC. Anrukinzumab, a humanised antibody that inhibits human IL-13, was evaluated for the treatment of UC.
Design In a multicentre, randomised, double-blind, placebo-controlled study, patients with active UC (Mayo score ≥4 and <10) were randomised to anrukinzumab 200, 400 or 600 mg or placebo. Patients received five intravenous administrations over 14 weeks. The primary endpoint was fold change from baseline in faecal calprotectin (FC) at Week 14. Secondary endpoints included safety, pharmacokinetics and IL-13 levels.
Results The modified intention-to-treat population included 84 patients (21 patients/arm). Fold change of FC from baseline at Week 14 was not significantly different for any treatment groups compared with the placebo. The study had a high dropout rate, in part, related to lack of efficacy. The exploratory comparisons of each dose were not significantly different from placebo in terms of change from baseline in total Mayo score, clinical response, clinical remission and proportion of subjects with mucosal healing. An increase in serum total IL-13 (free and bound to anrukinzumab) was observed for all anrukinzumab groups but not with placebo. This suggests significant binding of anrukinzumab to IL-13. The safety profile was not different between the anrukinzumab and placebo groups.
Conclusions A statistically significant therapeutic effect of anrukinzumab could not be demonstrated in patients with active UC in spite of binding of anrukinzumab to IL-13.
Trial registration number ClinicalTrials.gov number NCT01284062.
- ULCERATIVE COLITIS
Statistics from Altmetric.com
Significance of this study
What is already known on this subject?
Elevated production of interleukin (IL)-13 in the mucosa of patients with active UC has been demonstrated.
IL-13 has been shown to impair epithelial barrier function by affecting epithelial apoptosis, tight junctions and restitution velocity.
These findings suggest that the blockade of IL-13 could be an effective strategy to treat UC.
What are the new findings?
Treatment of mild and moderate active UC with anrukinzumab, a humanised antibody that binds and inhibits human IL-13, did not result in significant changes in the faecal biomarker calprotectin relative to the placebo treated patients.
Treatment with anrukinzumab was not different from placebo for induction of response, remission or mucosal healing.
No safety signals were detected as a result of treatment with anrukinzumab.
How might it impact on clinical practice in the foreseeable future?
Immunoblockade of IL-13 to IL-4Rα interaction with anrukinzumab does not afford benefit for the treatment of active UC. This should be considered in the selection of targets related to IL-13 signalling pathway.
UC is a chronic, relapsing disease in which an enhanced Th2-mediated response, involving the cytokines interleukin 5, (IL-5), IL-9 and IL-13, is presumed to be an integral part of the disease process. In particular, there is strong support for a central role of IL-13 as an effector cytokine in the pathogenesis of UC.1
Elevated levels of mRNA transcripts of IL-13 have been shown in rectal biopsy specimens of subjects with UC.2 Additional studies demonstrated that in vitro stimulation of lamina propria mononuclear cells from inflamed UC tissue led to secretion of markedly increased amounts of IL-13 compared with cells from both controls and patients with Crohn's disease.3 In addition, it was shown that the critical cell population for IL-13 secretion was one bearing CD161, a marker found on natural killer T cells.3
IL-13 signalling is mediated by a receptor composed of the IL-13α1 and the IL-4α chain.4 Engagement of the IL-4α chain triggers phosphorylation of the transcription factor, STAT6, which mediates downstream responses to the cytokine.5 Intestinal epithelial cells isolated from UC patients have elevated expression of IL-13α1 and IL-4α.6 There is also a high-level expression of IL-13α2, a decoy receptor that may promote fibrosis and be responsible for elimination of IL-13 in vivo.7
The consistent evidence of IL-13 overproduction in active UC, along with its role in dysregulation of intestinal epithelial cells, suggests that the blockade of IL-13 could be an effective strategy for treating UC. Anrukinzumab is a humanised antibody (IgG1) that binds to IL-13 and inhibits attachment of IL-13 to IL-4Rα thus interrupting the IL-13-induced inflammatory signal.8–10 Anrukinzumab does not block IL-13 binding to IL-13α1 or IL-13α2. This study represents the first investigation of an anti-IL-13 therapy in subjects with active UC and was designed to evaluate the effect on a UC-related biomarker, faecal calprotectin (FC), and investigate the therapeutic potential of targeting IL-13 in this disease.
Trial design and intervention
A phase IIa, randomised, double-blind, sponsor-unblinded, placebo-controlled study was conducted at 38 centres in 10 countries (Austria, Bulgaria, Canada, France, Germany, Hungary, The Netherlands, Poland, Spain and the USA). The study protocol was reviewed and approved by the institutional review board and/or independent ethics committee(s) at each of the investigational centres participating in the study. The study was conducted in compliance with the Declaration of Helsinki and with all International Conference on Harmonization Good Clinical Practice Guidelines.
Male and female patients aged 18–65 years were enrolled if they had a diagnosis of UC as confirmed by histopathology as well as active disease defined by a Mayo score11 ≥4 and <10 with an endoscopic subscore of ≥2 points and FC ≥100 mg/kg during the screening period. FC of at least 100 mg/kg was chosen to ensure that change from baseline in FC was measurable in all patients.
Patients with any of the following were excluded: diagnosis of indeterminate colitis or Crohn's disease, enteric infections, significant concurrent medical conditions, HIV, pre-existing demyelinating disorders, abnormal chest X-ray, positive Mantoux tuberculin skin test or a positive interferon gamma release assay during screening or within 12 weeks prior to randomisation. Patients receiving the following therapies within the specified timeframes relative to baseline were also ineligible: azathioprine or 6-mercaptopurine within 30 days, methotrexate within 7 days, oral, intravenous, injectable, or rectal steroids of any type within 30 days, tumour necrosis factor (TNF) inhibitors or other biological therapy within 8 weeks or at least five half-lives of baseline visit, whichever was longer. Permitted concomitant medications included oral and rectal mesalamine provided that doses were stable for 2 weeks prior to screening.
Eligible subjects were stratified based on previous exposure to treatment with immunosuppressives and/or TNF inhibitors (yes/no), and randomised to receive anrukinzumab 200, 400, or 600 mg or placebo as five intravenous administrations (at baseline and Weeks 2, 4, 8 and 12) during a 14-week treatment phase and then entered into an 18-week safety follow-up period.
The primary endpoint was the fold change from baseline in FC at Week 14 which is expressed as: 100×((post treatment/baseline)-1). Calprotectin levels were measured using the EK-CAL commercial ELISA (Bühlmann Laboratories, Switzerland). Five batch validation of the Bühlmann assay demonstrated that it was fit for purpose with a lower limit of quantitation of 14.2 μg/g, % coefficient of variation ≤11.2 and % accuracy of ≥89.8.
Secondary endpoints included fold change from baseline in FC at Weeks 2, 4, 8 and 12, pharmacokinetics, total IL-13, antidrug and neutralising antibodies, as well as safety and tolerability of anrukinzumab.
Serum samples for pharmacokinetic analysis were collected at baseline and postinfusion on Days 1, 2, 4, and 7, preinfusion and postinfusion at Weeks 2, 4, 8, 12, and 14, and at all follow-up visits. Anrukinzumab samples were assayed using a validated ELISA assay. Serum samples for total IL-13 were collected at baseline, on Days 2, 4, and 7 and Weeks 2, 4, 8, 12, 14, 16, 20, 24, 28, and 32. Total IL-13 was measured using a Singulex IL-13 Immunoassay kit (cat# 03–0005-05); both free IL-13 and IL-13 complexed with anrukinzumab can be detected in this assay. Blood samples for antidrug and neutralising antibodies against anrukinzumab were collected prior to dosing at baseline and at Weeks 4, 8, 12 and 14 and then every 4 weeks until Week 32.
Additional predefined exploratory endpoints included change from baseline in total Mayo score and subscores (stool frequency, rectal bleeding, endoscopic subscore and Physician's Global Assessment), clinical response rate at Week 14 (defined as proportion of subjects with a decrease from baseline of ≥3 points in total Mayo score, with at least a 30% change, accompanied by ≥1 point decrease or absolute score of 0 or 1 in rectal bleeding subscore), clinical remission rate at Week 14 (defined as proportion of subjects with a total Mayo score ≤2, with no individual subscore >1) and inhibition of IL-13-induced phosphorylation of STAT6 in cultured HT-29 cells ex vivo as measured by flow cytometry, faecal lactoferrin, faecal YKL-40, serum high sensitivity C reactive protein (hs-CRP) and immunoglobulin (Ig)E. Post hoc analyses were also conducted to determine mucosal healing rate at Week 14 (defined as a Mayo subscore for endoscopy of 0 or 1) and peripheral blood eosinophil counts.
Optional biopsies solely for gene expression analysis were taken at baseline and Week 14 from consenting individuals from inflamed segments of the colon and also from a non-inflamed segment, when existing. Quantitative RT-PCR was performed using TaqMan low density arrays covering six control genes (GAPDH, GUSB, PGK1, RPLPO, ZNF592 and DDX3Y) and 90 disease or IL-13 associated genes.
For the comparison of fold change from baseline in FC at Week 14, it was estimated that 20 subjects per arm (80 subjects in total) would provide ∼84% power to detect a 60% reduction in fold change of FC from baseline compared with placebo, while 16 subjects per arm (64 subjects in total) provide ∼79% power (assuming 20% dropout). A 60% reduction in fold change from baseline corresponds to a difference of −0.916 for change from baseline in natural log-transformed FC. This power calculation was based on the following assumptions: (1) a one-sided alpha of 0.1 and (2) a common SD of 1.256 in the absolute change from baseline of natural log-transformed FC.
Analyses were based on a modified intention-to treat population, defined as all subjects who received at least one dose of randomised treatment. To achieve the primary objective, a superiority analysis (treatment vs placebo) of changes from baseline in natural log-transformed FC (equivalent to the fold change of FC) at Week 14 was based on an analysis of covariance (ANCOVA) model with terms for treatment group and baseline (in log scale). All the tests were conducted at two-sided 0.2 alpha level and 80% CIs which is equivalent to the one-sided 0.1 alpha level used for sample size calculation. ANCOVA analysis with terms treatment group and baseline value (in log scale) was used for weekly changes from baseline in natural log-transformed FC, with additional supportive analysis based on mixed model with log-transformed FC being the dependent variable and treatment group, time (as categorical variable), time by treatment and baseline FC in log scale being the independent variables and random subject effect. Least squares means, ratio and CIs were based on back log transformation of those from the mixed model. The primary analysis was based on a data-as-observed (DAO) approach without explicit imputation to handle absent data. A last observation carried forward (LOCF) approach was used to impute missing values as additional sensitivity analyses.
Other biomarker endpoints such as faecal lactoferrin, hs-CRP, IgE and eosinophils were analysed at each measured time point using the longitudinal data analysis model. Comparisons of clinical response rate and remission rate between treatment group and placebo group were tested by Fisher's exact test at two-sided alpha of 0.2. Analysis of change from baseline in Mayo score at Week 14 was based on an ANCOVA model with terms for treatment and baseline Mayo score. A linear regression model was used to analyse the relationship between FC and total Mayo score. Summary statistics were used for safety analyses.
In the gene expression analysis, a Wilcoxon signed rank test was performed for each gene, with inflamed and non-inflamed baseline samples paired from the same individual. Individual fold change in expression in involved biopsies from Weeks 0 to 14 was calculated. Changes in individual dose groups were compared with placebo using the Wilcoxon rank sum test.
From March 2011 to April 2013, a total of 152 subjects were screened and 84 subjects were randomised to treatment (21 subjects each in the placebo and anrukinzumab 200, 400 and 600 mg groups) (figure 1). Of those randomised, 57 patients completed the 14-week treatment phase and 45 patients the 18-week safety follow-up. The most frequent reasons for discontinuation (n=39) were adverse events (AEs) including worsening of UC (eight), no longer willing to participate (11) and insufficient clinical response (eight). The majority of discontinuations due to AEs (14) were deemed not to be related to study drug (11).
Baseline demographic characteristics and medical history are presented in table 1. Mean baseline total Mayo scores were similar across treatment groups, and mean FC varied across treatment groups but the difference was not significant.
The fold change from baseline in FC at Week 14 was not significantly different relative to placebo for the three dose levels of anrukinzumab evaluated (table 2). At Week 14, there was a 59% reduction from baseline in FC levels among patients receiving placebo, a 71% reduction in the anrukinzumab 200 mg group, a 21% reduction in the anrukinzumab 400 mg group and actually a 24% increase from baseline in FC in the anrukinzumab 600 mg group. Only at the anrukinzumab 600 mg dose the mean fold change in FC compared with the placebo group (p=0.067) was lower than the established alpha level of 0.2, but with an increase of FC levels instead of expected reduction. Similar findings were obtained from additional analyses using LOCF imputation to replace missing values at Week 14 (see online supplementary table S1).
As a secondary endpoint, the time course of change from baseline in FC is presented in figure 2A. The 200 mg group had lower FC levels than placebo at all time points except Week 2. The fold change decrease among patients receiving 200 mg was higher (p<0.2) than that observed in the placebo group at Week 4 (0.47 vs 0.92; p=0.09), Week 8 (0.36 vs 0.78; p=0.16) and Week 12 (0.19 vs 0.64; p=0.03) with all p values being lower than the established alpha level of 0.2. Levels in the placebo group were consistent throughout the study until Week 14 when they decreased. The time course of absolute values of FC for the treatment groups is shown in figure 2B.
Median anrukinzumab concentrations over time profiles by treatment group are shown in online supplementary figure S1. Serum anrukinzumab exposure (Cmax and AUC) increased from 200 to 600 mg with approximately dose proportional increases for both Day 1 and Week 12 (see online supplementary table S2). There was essentially no accumulation for Cmax for any treatment group. The mean terminal t½ on Week 12 were 392, 471 and 362 h, for the 200, 400 and 600 mg groups, respectively.
Increase from baseline in total IL-13 levels (free and bound to anrukinzumab) was observed for all anrukinzumab treatment groups but not in the placebo group (figure 3). Maximal increase was observed at 2 weeks after treatment started. At the end of treatment, total IL-13 levels gradually declined towards baseline as the drug concentration dropped. For some individual subjects, total IL-13 levels decreased during the treatment period while the drug concentration was still high. This occurred most frequently among subjects receiving anrukinzumab 600 mg. No significant correlation was observed between IL-13 levels and FC changes (r=0.17, p=0.23).
Exploratory and post hoc outcomes
There was no statistically significant difference between any of the three dose levels of anrukinzumab and placebo in the change from baseline in total Mayo score. Improvement in total Mayo score was greatest among patients receiving anrukinzumab 200 and 400 mg (table 3). The total Mayo score showed a weak, but significant correlation with FC (p=0.032, r=0.3) (see online supplementary figure S2). The highest rate of mucosal healing and highest proportion of patients with improvement in Mayo endoscopic subscore at Week 14 were observed among patients receiving anrukinzumab 400 mg whereas improvement in rectal bleeding was highest in the 200 mg group (table 3). There were no statistically significant (p<0.2) differences between any of the anrukinzumab groups and the placebo group for clinical response or remission rates (table 3). When missing or incomplete data were analysed using non-response imputation, the same lack of difference among the treatment groups was observed (data not shown).
Levels of faecal lactoferrin (see online supplementary figure S3A), faecal YKL-40 (data not shown) and serum hs-CRP (see online supplementary figure S3B) showed a large variability over time. In most cases, there was no significant difference between any of the three dose-levels of anrukinzumab evaluated and placebo in the fold change from baseline in any of these biomarkers. The only statistically significant changes were: 200 mg in lactoferrin level decreases at Week 12 (p=0.01) and in YKL-40 level decreases at Week 12 (p=0.03); and 600 mg in lactoferrin increases (p=0.04) and YKL-40 increases (p=0.02) at Week 14.
At Week 4, the fold changes from baseline in total IgE levels in the anrukinzumab 400 mg (p=0.012) and 600 mg groups (p=0.002) were significantly lower than in the placebo group but these differences were not significant at Week 14 (see online supplementary figure S3C). A general dose-dependent decrease of IgE change was observed.
Treatment-related eosinophil count increases were observed for the anrukinzumab 200 and 400 mg groups and the increases persist through Week 14. For the 600 mg group, enosinophil count increased transiently at Week 2 (see online supplementary figure S3D).
IL-13-induced phosphorylation of STAT6 in HT29 cells was markedly reduced by diluted serum samples taken at Day 2 and Week 14 from individuals treated with anrukinzumab but not those receiving placebo (see online supplementary figure S4).
Gene expression analysis
Of the genes analysed, 81 were differentially expressed in inflamed versus non-inflamed biopsies at baseline while five of the six control genes were unchanged (the exception being PGK1) between normal and inflamed tissue (see online supplementary table S3). IL-13 was increased by a factor of 20.7 (p=4.47×10−5, false discovery rate p value=7.66×10−5) in inflamed tissue compared with non-inflamed samples. None of the changes in the genes assayed were found to differ significantly from placebo for any dose group.
Similar proportions of subjects in all four treatment groups reported at least one AE. The most frequently reported AEs (≥4 subjects in any group; all causality) were UC flares, anaemia, nasopharyngitis and headache (table 4). Most AEs were mild or moderate in severity. No antidrug antibody was detected in this study.
A total of 14 subjects experienced at least one serious AE (four subjects each in the placebo and 200 and 600 mg groups, and two subjects in the 400 mg group). Of the observed severe AEs, three were considered treatment-related (one case of abdominal pain in the 200 mg group and two cases of UC flare in the 600 mg group). Worsening UC was reported as a serious AE in nine subjects (three subjects in the placebo, two subjects in the 200 mg group, one subject in the 400 mg group and three subjects in the 600 mg group), with almost half of these flares reported during the post-treatment, follow-up phase. No deaths were reported during the study.
A total of 14 subjects exited the study due to treatment-emergent AEs (four subjects on placebo and five subjects each in the 200 and 600 mg groups). Withdrawal of one subject on 200 mg (abdominal pain) and two subjects on 600 mg (bronchopneumonia and UC flare) was considered related to study drug.
In this phase IIa study in patients with mild to moderate UC, three doses of the anti-IL-13 antibody, anrukinzumab, were tested. None of the dose group showed statistically significant changes from baseline in FC levels at Week 14 compared with placebo, although changes in some biomarkers were observed. There did appear to be an inverse dose response with worse outcomes at higher doses. The lowest dose tested, 200 mg, demonstrated lower FC levels relative to placebo at all time points except for Week 14; these changes were statistically significant at Weeks 4, 8 and 12. Consistent with the significant correlation between FC and total Mayo score, this inverse dose response was also demonstrated in change from baseline in total Mayo score, clinical response, clinical remission and rectal bleeding subscore. Only mucosal healing was better in the 400 mg group compared with the 200 mg group. Therefore, FC served as a reasonable surrogate for these clinical endpoints for this early phase study. The study was not powered to detect statistical differences in the clinical endpoints, and so these need to be interpreted with caution.
The current study is the first to our knowledge to use FC as a primary endpoint to investigate the therapeutic potential of a drug in the treatment of UC. While this primary endpoint was not achieved, the effects of anrukinzumab on FC were consistent with clinical scores and effects on other common serum and stool biomarkers and support the potential value of FC in assessing disease activity of UC. Improvements in methodology and transport media for easier stool sampling in the future will decrease the number of missing samples and enhance the usefulness of FC in early proof-of-mechanism clinical trials as well as in clinical practice.
The significant amount of missing data due to discontinuation of subjects had an impact on the power of the study. The DAO approach of the primary analysis may have been compromised to some degree by patient withdrawals and missing stool samples for assay at Week 14. However, analyses using LOCF imputation to replace missing values at Week 14 showed no greater evidence of any significant differences between the anrukinzumab dose groups and the placebo group. Furthermore, secondary analyses at earlier time points, for which data were available for more subjects, similarly showed little evidence of any difference between any of the anrukinzumab and the placebo groups.
When IL-13 was bound by anrukinzumab, the complex has much larger molecular weight than free IL-13 and therefore cannot be cleared by kidney, a common mechanism of clearance for small cytokines. Although the IL-13 anrukinzumab complex can still be cleared through IL-13α2, the lack of renal clearance results in a slight accumulation of total IL-13. Therefore, serum total IL-13 (free and anrukinzumab-bound) levels increased in all three anrukinzumab dose groups but not in the placebo group suggesting that, at least peripherally, there was significant IL-13 binding to anrukinzumab.
In summary, treatment with anrukinzumab (200, 400, and 600 mg) did not demonstrate a statistically significant therapeutic effect in subjects with active UC, as measured by changes in FC and other disease-related biomarkers or clinical parameters. Treatment with anrukinzumab was generally well tolerated and without safety concerns. As our study supports the biological activity of anrukinzumab, the lack of biomarker, clinical or endoscopic response seriously questions the pathogenic role of IL-13 in UC.
Contributors WR: study design, data analysis and interpretation, patient recruitment, data collection. JP: data analysis and interpretation, patient recruitment, data collection. SK and GT: patient recruitment, data collection. FH, GMC, MH, KP, MO and FC: study design, data analysis and interpretation, study supervision. TMM: data analysis and interpretation, study supervision. HZ and YS: statistical analysis. All authors had full access to all of the data in the study and were involved in reviewing the article critically for important intellectual content and revising the article. All authors approved the final version to be published.
Funding This study was funded by Pfizer. Medical writing support was provided by John Bilbruck of Engage Scientific Solutions which was funded by Pfizer. Support for analysis of gene expression data was provided by Janna Hutz and Padma Reddy, employees of Pfizer.
Competing interests WR has served as a speaker, a consultant and/or an advisory board member for Abbott Laboratories, Abbvie, Aesca, Amgen, AM Pharma, Aptalis, Astellas, Astra Zeneca, Bioclinica, Biogen IDEC, Bristol-Myers Squibb, Cellerix, Chemocentryx, Celgene, Centocor, Danone Austria, Elan, Falk Pharma GmbH, Ferring, Galapagos, Genentech, Grünenthal, Janssen, Johnson & Johnson, Kyowa Hakko Kirin Pharma, Lipid Therapeutics, Millenium, Mitsubishi Tanabe Pharma Corporation, MSD, Novartis, Ocera, Otsuka, PDL, Pharmacosmos, Pfizer, Procter & Gamble, Prometheus, Robarts Clinical Trial, Schering-Plough, Setpointmedical, Shire, Takeda, Therakos, Tigenix, UCB, Vifor, Yakult, Zyngenia and 4SC; JP has served as a speaker, a consultant and/or an advisory board member for Abbvie Laboratories, Boehringer-Ingelheim, Bristol-Myers Squibb, Cellerix, Ferring, Galapagos, Genentech, Janssen, MSD, Novo-Nordisk, NSP, Pfizer, Robarts Clinical Trials, Schering-Plough, Shire, Takeda and Tigenix; FH, MH, KP, HZ, YS and FC are employees of Pfizer; GMC and MO were employees of Pfizer at the time of the study; TMM is an employee of InVentiv Health Clinical, paid contractors to Pfizer in connection with this study.
Patient consent Obtained.
Ethics approval Ethics committee of each participating institution.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement All data generated in the conduct of the study are presented in this manuscript.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.