POS0074 IMMUNOLOGICAL DIFFERENCES BETWEEN PsA PATIENTS WHO ARE TUMOR NECROSIS FACTOR INHIBITOR-NAIVE AND WHO HAVE INADEQUATE RESPONSE TO TUMOR NECROSIS FACTOR INHIBITORS

Abstract

BackgroundA better understanding of the immunological differences between psoriatic arthritis (PsA) patients (pts) who are tumor necrosis factor inhibitor (TNFi)-naïve & who have inadequate response to TNFi (TNFi-IR) may guide treatment choices. In DISCOVER-1, benefit of the IL-23p19 subunit inhibitor guselkumab (GUS) every-four-weeks (Q4W) & Q8W vs placebo (PBO) in improving PsA signs & symptoms was seen in adults with active PsA.1 The Ph3b COSMOS study of GUS Q8W vs PBO in TNFi-IR PsA pts corroborated these findings.2ObjectivesAssess baseline (BL) molecular differences between TNFi-naïve & -IR PsA pts & investigate GUS pharmacodynamic (PD) effect on cytokine expression over time in these cohorts.MethodsSerum samples collected from consenting biomarker substudy pts in DISCOVER-11 (TNFi-naïve [n=101] & -IR [n=17]), DISCOVER-23 (TNFi-naïve [n=150]), & COSMOS2 (TNFi-IR [n=76]) were analyzed for selected serum cytokine levels. TNFi-IR pts in this post-hoc analysis had active PsA & discontinued 1-2 TNFi due to inadequate efficacy; these pts required a TNFi-specific washout period prior to starting GUS. PD effect of GUS Q8W on cytokine levels was assessed. Differential BL cytokine expression, associations between BL cytokine levels & clinical response (Psoriasis [PsO] Area & Severity Index 75% improvement from BL [PASI75] & American College of Rheumatology 20% improvement [ACR20]), & GUS effect on cytokine levels were analyzed with a General linear model & Spearman linear regression.ResultsBL pt demographics, disease characteristics, & conventional synthetic disease-modifying antirheumatic drug (csDMARD) use were comparable between TNFi-naïve (DISCOVER-1 & -2, N=251) & -IR (DISCOVER-1 & COSMOS, N=93) pts, with differences in mean PASI score (8.9 v 12.5), swollen joint count (SJC) (11.7 v 10.3), PsA duration (5.8 v 9.8 yrs), & PsO duration (16.7 v 20.4 yrs; Table 1). BL serum IL-22 & TNFα levels for pooled treatment groups were higher in TNFi-IR than -naïve pts (p<0.05). At W24, GUS reduced IL-22, IL-17A/F, IL-6, C-reactive protein (CRP), & serum amyloid A protein to similar levels in both cohorts (p<0.05; Figure 1). W24 PASI75 responders had higher BL IL-17F levels with GUS in both cohorts (p<0.05) & higher IL-22 levels in TNFi-IR pts only (p<0.05). A trend of upregulated BL IL-22 expression in W24 ACR20 responders was seen for TNFi-IR pts with GUS (p=0.07).Table 1.BL demographics, disease characteristics, & drug use in TNFi-naïve & -IR cohorts with available cytokine data in DISCOVER-1&2 & COSMOS.*TNFi-naïve (N=251)TNFi-IR (N=93)Age [yrs]47.2 (11.3)48.5 (11.1)Female, n (%)132 (52.6)46 (49.5)Body mass index [kg/m2]29.6 (6.1)30.3 (6.4)Median (range) CRP [mg/dL]0.9 (0.0-12.9)1.0 (0.0-13.2)Log2 IL-22 / TNFα [pg/mL]2.0 (1.4) / 1.1 (0.6)2.5 (1.5) / 1.9 (1.2)Log2 IL-17A / F [pg/mL]-0.4 (1.5) / 1.7 (1.5)-0.1 (1.7) / 2.0 (1.6)SJC [0-66]11.7 (7.1)10.3 (8.3)TJC [0-68]20.3 (13.1)20.6 (14.2)PsA duration [yrs]5.8 (5.9)9.8 (8.2)PsO duration [yrs]16.7 (12.8)20.4 (12.0)PsO Body surface area (%)14.8 (18.6)19.1 (21.3)Investigator’s Global Assessment score [0-4]2.3 (0.9)2.3 (1.0)PASI score [0-72]8.9 (10.6)12.5 (12.0)Enthesitis [Y], n (%)160 (63.7)58 (62.4)csDMARD use [Y], n (%)164 (65.3)62 (66.7)Corticosteroid use (Y), n (%)45 (17.9)19 (20.4)Methotrexate use [Y], n (%)136 (54.2)54 (58.1)Data are mean (SD) unless otherwise noted. *Pts with serum CRP level ≥0.3 mg/dL, SJC ≥3, & TJC ≥3 (to mimic D1 inclusion criteria1). TJC= tender joint countConclusionElevated BL IL-22 expression & association between BL IL-22 levels & W24 PASI75 response, & a W24 trend for an association between upregulated BL IL-22 & ACR20 response, in TNFi-IR pts seen in this exploratory analysis may suggest increased involvement of the IL-23 pathway in TNFi-IR pts. GUS showed comparable & significant PD effects for TNFi-naïve & -IR pts, consistent with observed clinical responses.