Christina Hermanrud’s research while affiliated with Karolinska Institutet and other places

A significant proportion of multiple sclerosis (MS) patients treated with interferon beta-1a (Rebif™) develop anti-drug antibodies (ADA) with a negative impact on treatment efficacy. We hypothesized that high-throughput B-cell receptor (BCR) repertoire analysis could be used to predict and monitor ADA development. To study this we analyzed 228 peripheral blood samples from 68 longitudinally followed patients starting on interferon beta-1a. Our results show that whole blood BCR analysis does not reflect, and does not predict ADA development in MS patients treated with interferon beta-1a. We propose that BCR analysis of phenotypically selected cell subsets or tissues might be more informative.

Background Upon treatment with biopharmaceuticals, the immune system may produce anti-drug antibodies (ADA) that inhibit the therapy. Up to 40% of multiple sclerosis patients treated with interferon β (IFNβ) develop ADA, for which a genetic predisposition exists. Here, we present a genome-wide association study on ADA and predict the occurrence of antibodies in multiple sclerosis patients treated with different interferon β preparations. Methods We analyzed a large sample of 2757 genotyped and imputed patients from two cohorts (Sweden and Germany), split between a discovery and a replication dataset. Binding ADA (bADA) levels were measured by capture-ELISA, neutralizing ADA (nADA) titers using a bioassay. Genome-wide association analyses were conducted stratified by cohort and treatment preparation, followed by fixed-effects meta-analysis. Results Binding ADA levels and nADA titers were correlated and showed a significant heritability (47% and 50%, respectively). The risk factors differed strongly by treatment preparation: The top-associated and replicated variants for nADA presence were the HLA-associated variants rs77278603 in IFNβ-1a s.c.- (odds ratio (OR) = 3.55 (95% confidence interval = 2.81–4.48), p = 2.1 × 10⁻²⁶) and rs28366299 in IFNβ-1b s.c.-treated patients (OR = 3.56 (2.69–4.72), p = 6.6 × 10⁻¹⁹). The rs77278603-correlated HLA haplotype DR15-DQ6 conferred risk specifically for IFNβ-1a s.c. (OR = 2.88 (2.29–3.61), p = 7.4 × 10⁻²⁰) while DR3-DQ2 was protective (OR = 0.37 (0.27–0.52), p = 3.7 × 10⁻⁰⁹). The haplotype DR4-DQ3 was the major risk haplotype for IFNβ-1b s.c. (OR = 7.35 (4.33–12.47), p = 1.5 × 10⁻¹³). These haplotypes exhibit large population-specific frequency differences. The best prediction models were achieved for ADA in IFNβ-1a s.c.-treated patients. Here, the prediction in the Swedish cohort showed AUC = 0.91 (0.85–0.95), sensitivity = 0.78, and specificity = 0.90; patients with the top 30% of genetic risk had, compared to patients in the bottom 30%, an OR = 73.9 (11.8–463.6, p = 4.4 × 10⁻⁶) of developing nADA. In the German cohort, the AUC of the same model was 0.83 (0.71–0.92), sensitivity = 0.80, specificity = 0.76, with an OR = 13.8 (3.0–63.3, p = 7.5 × 10⁻⁴). Conclusions We identified several HLA-associated genetic risk factors for ADA against interferon β, which were specific for treatment preparations and population backgrounds. Genetic prediction models could robustly identify patients at risk for developing ADA and might be used for personalized therapy recommendations and stratified ADA screening in clinical practice. These analyses serve as a roadmap for genetic characterizations of ADA against other biopharmaceutical compounds.

A subgroup of patients treated with infliximab lose response to the treatment and one reason for this is the development of anti-drug antibodies (ADA). If used optimally, measuring drug and ADA level could lead to a more personalized and efficient treatment regime, and enable identification of ADA-positive patients before the underlying disease flares or allergic reactions occur. With the use of a drug-tolerant ADA assay which can detect ADA irrespective of drug levels in the sample, we determined the impact of ADA on treatment failure to infliximab. The aims of this study were to estimate the real-life optimal serum infliximab (sIFX) level and set a clinical threshold value for a drug-tolerant ADA assay. Trough levels of sIFX were measured with ELISA. Free ADA was measured with two drug-sensitive methods (ELISA and a bioassay) and one drug-tolerant method (PandA). Two real-life cohorts treated with infliximab were included; a cross-sectional cohort including patients with inflammatory rheumatic diseases (n = 270) and a prospective cohort of rheumatoid arthritis (RA) patients (n = 73) followed for 1 year. Normal range of sIFX was estimated from the prospective cohort and an arbitrary optimal drug level was set to be between 1 and 6 μg/mL. Using this range, optimal sIFX was found in only 60% (163/270) of the patients in the cross-sectional cohort. These patients had significantly better treatment response than those with a drug level under 1 μg/mL, who had an ADA frequency of 34% (19/56) using the drug-tolerant method. In the prospective cohort, the drug-tolerant assay could identify 34% (53/155 samples) as ADA positive in samples with sIFX level >0.2 μg/mL. ADA were seldom detected in patients with >1 μg/mL sIFX, with three interesting exceptions. A clinically relevant ADA threshold was determined to be >3 RECL as measured with the drug-tolerant assay. In a real-life setting, there was a substantial number of patients with suboptimal drug levels and a proportion of these had ADA. Both too low and too high drug levels correlated with worse disease, but for different reasons. Adding a drug-tolerant assay enabled detection of ADA earlier and regardless of drug level at time of sampling.

Background: Upon treatment with biopharmaceuticals, the immune system may produce anti-drug antibodies (ADA) that inhibit the therapy. Up to 40% of multiple sclerosis patients treated with interferon β (IFNβ) develop ADA, for which a genetic predisposition exists. Here, we present a genome-wide association study on ADA and predict the occurrence of antibodies in multiple sclerosis patients treated with different interferon β preparations. Methods: We analyzed a large sample of 2,757 genotyped and imputed patients from two cohorts, split between a discovery and a replication dataset. Binding ADA (bADA) levels were measured by capture-ELISA, neutralizing ADA (nADA) titers using a bioassay. Genome-wide association analyses were conducted stratified by cohort and treatment preparation, followed by fixed-effects meta-analysis. Results: Binding ADA levels and nADA titers were correlated and showed a significant heritability (47% and 50%, respectively). The risk factors differed strongly by treatment preparation: The top-associated and replicated variants for nADA presence were the HLA-associated variants rs77278603 in IFNβ-1a s.c.- (odds ratio (OR)=3.55 (95% confidence interval=2.81-4.48), p=2.1x10-26) and rs28366299 in IFNβ-1b s.c.-treated patients (OR=3.56 (2.69-4.72), p=6.6x10-19). The rs77278603-correlated HLA haplotype DR15-DQ6 conferred risk specifically for IFNβ-1a s.c. (OR=2.88 (2.29-3.61), p=7.4x10-20) while DR3-DQ2 was protective (OR=0.37 (0.27-0.52), p=3.7x10-09). The haplotype DR4-DQ3 was the major risk haplotype for IFNβ-1b s.c. (OR=7.35 (4.33-12.47), p=1.5x10-13). These haplotypes exhibit large population-specific frequency differences. In a cohort of IFNβ-1a s.c.-treated patients, prediction models for nADA reached an AUC of 0.91 (0.85-0.95), a sensitivity of 0.78, and a specificity of 0.90. Patients with the top 30% of genetic risk had, compared to patients in the bottom 30%, an OR of 73.9 (11.8 463.6, p=4.4x10-06) of developing nADA. Conclusions: We identified several HLA-associated genetic risk factors for ADA against interferon β, which were specific for treatment preparations and population backgrounds. Genetic prediction models could robustly identify patients at risk for developing ADA and might be used for personalized therapy recommendations and stratified ADA screening in clinical practice. These analyses serve as a roadmap for genetic characterizations of ADA against other biopharmaceutical compounds.

Objective: Infliximab-treated patients with rheumatoid arthritis (RA) may respond insufficiently due to low serum infliximab (sIFX) levels, caused by anti-drug antibodies (ADAs). However, monitoring of sIFX and ADAs is not routinely implemented, and levels for optimal outcome have not been validated. We searched for predictors for sIFX < 0.2 μg/mL and ADA development in a randomized setting. Methods: In the SWEFOT trial, of 128 patients randomized to methotrexate + IFX therapy, 101 had serum samples at 3, 9, and 21 months that were analysed for sIFX [enzyme-linked immunosorbent assay (ELISA)] and ADAs [ELISA, and precipitation and acid dissociation (PandA) when sIFX > 0.2 μg/mL]. The primary and secondary outcome measures were low disease activity [LDA = 28-joint Disease Activity Score (DAS28) ≤ 3.2] and remission (DAS28 < 2.6). Baseline characteristics were assessed as potential predictors of sIFX < 0.2 μg/mL or ADA positivity, using logistic regression. Results: Categorization of sIFX levels into < 0.2, 0.2–2.9, 3.0–7.0, and > 7.0 μg/mL showed a dose–response association with LDA (30%, 64%, 67%, and 79%, respectively, p = 0.008) and remission (10%, 45%, 39%, and 66%, p = 0.004) at trial cessation (21 months). Female patients had sIFX < 0.2 μg/mL more often than males (35% vs 7%, p = 0.006), with a similar trend for rheumatoid factor (RF)-positive vs RF-negative patients (34% vs 16%, p = 0.059). ADA positivity showed similar patterns, also after adjustment for potential confounders (female sex: p = 0.050; RF positivity: p = 0.067). PandA captured four highly ADA-reactive patients with sIFX > 0.2 μg/mL, of whom three were ADA positive at other time-points, all with high DAS28 at follow-up. Conclusion: In early RA patients receiving IFX as a second-line agent, sIFX < 0.2 μg/mL and ADA development were associated with treatment failure and were more common in females, with a similar trend for RF positivity. Our findings support the use of therapeutic drug monitoring, and PandA in ADA-negative non-responders. Trial registration: SWEFOT NCT00764725 (https://clinicaltrials.gov/ct2/show/NCT00764725).

Interferon beta (IFNβ) is used as a first-line treatment for multiple sclerosis (MS) and is injected intramuscularly or subcutaneously (s.c.). The subcutaneous route is considered more immunogenic as it is associated with increased antidrug antibody-positive patients. The skin contains dendritic cells (DCs) and it is unclear whether these contribute to immunogenicity. To assess the effect of IFNβ on skin-resident cells, IFNβ was injected intradermally (i.d.) ex vivo using a human skin explant model or s.c. in vivo in MS patients. Ex vivo, intradermal IFNβ injections reduced migration and enhanced surface CD86 expression of dermal DCs, and an increased expression of HLA-DR+ was observed in skin biopsies taken after subcutaneous IFNβ injection (in vivo). In both models, IFNβ elevated the expression of several inflammatory cytokines when compared to the control biopsies. Our results show that 3 different IFNβ preparations, normalized in dose and injection site, induce similar immune responses, suggesting that the differences in immunogenicity are likely due to the route and frequency of administration.

Two validated assays, a bridging ELISA and a luciferase-based bioassay, were compared for detection of anti-drug antibodies (ADA) against interferon-beta (IFN-β) in patients with multiple sclerosis. Serum samples were tested from patients enrolled in a prospective study of 18 months. In contrast to the ELISA, when IFN-β-specific rabbit polyclonal and human monoclonal antibodies were tested, the bioassay was the more sensitive to detect IFN-β ADA in patients' sera. For clinical samples, selection of method of ELISA should be evaluated prior to the use of a multi-tiered approach. A titer threshold value is reported that may be used as a predictor for persistently positive neutralizing ADA.

Background Tumour necrosis factor (TNF) inhibitors, with infliximab (IFX) first on the market, have revolutionised treatment of patients with rheumatoid arthritis (RA). However, in a substantial proportion of patients, they lose efficiency, and up to 44% of patients have been found to develop anti-drug antibodies (ADA), leading to low serum IFX (sIFX) levels. Despite this, sIFX measurement is still rarely used for clinical decision making, and standardised clinical threshold titre levels have not been clearly defined. Objectives In an early RA trial adding IFX to methotrexate (MTX) in patients not achieving low disease activity (LDA=DAS28≤3.2) after 3 months monotherapy, we studied whether sIFX or ADA were associated with treatment outcome, and whether easily available baseline parameters predicted ADA development. Methods Of IFX-treated SWEFOT patients (n=128), 101 had available serum samples at follow-up, which were analysed for sIFX levels at 3, 9 and 21 months (routine ELISA). Samples with undetectable sIFX (<0.2 µg/ml) were analysed further for ADA using direct ELISA with plate-bound TNF. Primary and secondary outcome measures were LDA and remission (DAS28 <2.6) at 21 months. Clinical and demographic characteristics of patients at start of IFX therapy (baseline) were tested as potential predictors of ADA development, using uni- and multivariate logistic regression. Results At 3, 9 and 21 months from IFX add-on to MTX, 15%, 23% and 28% of patients, respectively, had undetectable sIFX, and 34% were ever ADA-positive. Significantly higher proportion of patients achieved LDA among those with detectable sIFX, versus undetectable sIFX and positive ADA (67% vs 26%, p=0.002, figure 1A), with similar difference for remission (47% vs 11%, p=0.004, figure 1B). When sIFX levels were further stratified into <0.2, 0.2–5.0, 5.0–10.0 and >10 µg/ml, there was a significant trend across the groups in achievement of LDA (30%, 65%, 70% and 83% respectively, p=0.008, figure 1C) or remission (10%; 41%, 52% and 67%, respectively, p=0.004, figure 1D). Women had undetectable sIFX at 21 months more often than men (35% vs 7%, p=0.006). In multivariate logistic regression analysis, the following baseline characteristics were significant predictors of ever ADA-positivity: female gender, RF-positivity, higher tender joint count, erythrocyte sedimentation rate and lower health assessment questionnaire score (data not shown). • Download figure • Open in new tab • Download powerpoint Abstract OP0232 – Figure 1 Clinical outcome of patients at 3, 9 and 21 months stratified for sIFX and ADA status at the same time points. Proportion of patients in LDA (A) and remission (B) among patient with detectable sIFX level (blue dots) and ADA positive patients with undetectable sIFX levels (red dots). Proportion of patients in LDA (C) and remission (D) among four strata of patients according to sIFX levels undetectable (<0.2 μg/ml) – blue bars, low (0.2–5.0 μg/ml) – red bars, moderate (>5.0–10.0 μg/ml) – green bars, and high (>10.0 μg/ml) – orange bars. Conclusions In early RA patients receiving add-on IFX therapy, ADA-positivity or lower serum IFX levels were associated with a higher risk of not reaching treatment targets, that is LDA or remission. RF positivity and female gender, factors known to be associated with worse clinical outcomes, predicted development of ADA. Disclosure of Interest K. Hambardzumyan: None declared, C. Hermanrud: None declared, P. Marits: None declared, N. Vivar: None declared, S. Ernestam: None declared, J. Wallman Consultant for: AbbVie, Celgene, Eli Lilly, Novartis, UCB, R. van Vollenhoven Grant/research support from: AbbVie, BMS, GSK, Pfizer, UCB, Consultant for: AbbVie, AstraZeneca, Biotest, BMS, Celgene, GSK, Janssen, Lilly, Novartis, Pfizer, UCB, A. Fogdell-Hahn Grant/research support from: Pfizer, S. Saevarsdottir: None declared

Objective To develop and validate a method for the detection of binding anti-drug antibodies (ADAs) against interferon beta (IFN-β) in human serum as part of a European initiative (ABIRISK) aimed at the prediction and analysis of clinical relevance of anti-biopharmaceutical immunization to minimize the risk. Method A two-tiered bridging enzyme-linked immunosorbent assay (ELISA) format was selected and validated according to current recommendations. Screening assay: ADA in serum samples form complexes with immobilized IFN-β and biotinylated IFN-β, which are then detected using HRP labeled Streptavidin and TMB substrate. Confirmation assay: Screen “putative positive” samples are tested in the presence of excess drug (preincubation of sera with 0.3 µg/mL of soluble IFN-β) and percentage of inhibition is calculated. Results The assay is precise, and the sensitivity of the assay was confirmed to be 26 ng/mL using commercially available polyclonal rabbit antihuman IFN-β in human sera as the positive control. Conclusion An ultrasensitive ELISA for IFN-β-binding ADA testing has been validated. This will form the basis to assess anti-biopharmaceutical immunization toward IFN-β with regards to its clinical relevance and may allow for the development of predictive tools, key aims within the ABIRISK consortium.

Neutralizing anti-drug antibodies (NAbs) against therapeutic interferon beta (IFNβ) in people with multiple sclerosis (MS) are measured with cell-based bioassays. The aim of this study was to re-develop and validate two luciferase reporter gene bioassays, LUC and iLite, using a cut-point approach to identify NAb positive samples. Such an approach is favored by the pharmaceutical industry and governmental regulatory agencies as it has a clear statistical basis and overcomes the limitations of the current assays based on the Kawade principle. The work was conducted following the latest assay guidelines. The assays were re-developed and validated as part of the “Anti-Biopharmaceutical Immunization: Prediction and analysis of clinical relevance to minimize the risk” (ABIRISK) consortium and involved a joint collaboration between four academic laboratories and two pharmaceutical companies. The LUC assay was validated at Innsbruck Medical University (LUCIMU), at Rigshospitalet (LUCRH) Copenhagen, and the iLite assay at Karolinska Institutet, Stockholm. For both assays, the optimal serum sample concentration in relation to sensitivity and recovery was 2.5% (v/v) in assay media. A Shapiro–Wilk test indicated a normal distribution for the majority of runs, allowing a parametric approach for cut-point calculation to be used, where NAb positive samples could be identified with 95% confidence. An analysis of means and variances indicated that a floating cut-point should be used for all assays. The assays demonstrated acceptable sensitivity for being cell-based assays, with a confirmed limit of detection in neat serum of 1519 ng/mL for LUCIMU, 814 ng/mL for LUCRH, and 320 ng/mL for iLite. Use of the validated cut-point assay, in comparison with the previously used Kawade method, identified 14% more NAb positive samples. In conclusion, implementation of the cut-point design resulted in increased sensitivity to detect NAbs. However, the clinical significance of these low positive titers needs to be further evaluated.