675 Measuring Serum Vedolizumab and Antibodies to Vedolizumab: Comparison of Four Commercially Available Clinical Laboratory Assays to the Assay Used in Vedolizumab Development

Abstract

INTRODUCTION: Vedolizumab (VDZ), a monoclonal antibody targeting the α4β7 integrin, is approved for the treatment of inflammatory bowel disease (IBD). Treatment optimization with VDZ may be informed by measuring VDZ serum levels. However, the assays used in the GEMINI studies are not commercially available and the relationship to results obtained with commercial assays is unknown. Therefore, VDZ serum level and anti-VDZ antibody (AVA) assays from Grifols Diagnostic Solutions, Immundiagnostik AG, Progenika Biopharma and Theradiag were compared to the assays used by Takeda (Reference; Ref). METHODS: The following identical sample sets were provided for serum level analysis: 1) 10 pooled sample groups (in triplicate) from biologic-naïve IBD patients spiked with up to 70 mg/mL of VDZ, 2) 2 pooled sample groups (in duplicate) biologic-naive subjects, both groups had a combination of 10 mg/mL infliximab (IFX) and 5 mg/mL of adalimumab (ADA), one group with 20 mg/mL VDZ and a second group without VDZ. 3) 23 singlet samples from VDZ-treated IBD patients with free drug concentrations between 8-60 mg/mL. 4) For AVA, 8 singlet samples with AVA (0 - ≥ 0.5 mg/mL), 4 with 15 mg/mL of VDZ and 4 without VDZ. For VDZ-spiked groups 1 and 2, results for replicates were averaged and were compared to the Ref assay results. Slopes were calculated by weighted least squares regression (spiked samples) or by linear regression (patient samples). AVA was detected with a qualitative assay, with results reported as AVA positive or AVA negative. RESULTS: All assays were specific with no cross-reactivity with IFX or ADA and were selective, detecting VDZ in samples that were spiked with all three drugs. Overall Vendor analysis of spiked samples correlated well (R2 ≥ 0.90) with the Ref assay but there were differences in the measured values based on differing slopes (range 0.94-1.56). Likewise, vendor patient sample analysis correlated well with the Ref assay (R2 ≥ 0.95) but slopes differed ((range 0.81-1.71); Figure 1. The Ref AVA assay has a sensitivity of 3.9 ng/mL and detects AVA at 0.5 mg/mL in the presence of VDZ levels of ≥50 mg/mL. All 4 vendor assays detected AVA but did not detect AVA in the presence of 15 mg/mL VDZ. CONCLUSION: Despite good specificity and selectivity, there are differences between the Ref assays Takeda has used in its clinical trials and commercial assays. This should be acknowledged when interpreting assays results with vedolizumab commercial assays.