AnaptysBio announces investigational rosnilimab data

AnaptysBio (ANAB) announced that investigational rosnilimab, a depleter and agonist targeting PD-1+ T cells, demonstrated a best-in-disease profile in patients with moderate-to-severe rheumatoid arthritis. In the robust, global 424-patient Phase 2b trial, rosnilimab achieved JAK-like efficacy on multiple clinically meaningful measures, including low disease activity and remission on the Clinical Disease Activity Index, as well as ACR70 response, over a six-month period. Furthermore, responses were then durable for at least two months off drug. Rosnilimab was safe and well tolerated, particularly when compared to standard of care biologics or JAKs. Most rosnilimab patients showed symptomatic and clinical improvement by three months. As previously reported, all three doses of rosnilimab achieved statistically significant reductions in mean change from baseline in DAS-28 CRP, the study’s primary endpoint, as well as for ACR20, at Week 12 versus placebo. During the three-month placebo-controlled period in both b/tsDMARD-naive and b/tsDMARD-experienced patients, rosnilimab demonstrated a rapid onset of ACR20 response, an accepted Phase 3 registrational endpoint, as well as a substantial decrease in objective measures such as C-reactive protein, that were both comparable to the Phase 2b trial results of upadacitinib. Rosnilimab demonstrated JAK-like efficacy with deepening of responses through six months on CDAI LDA, CDAI remission and ACR70. Importantly, this is particularly observed in the b/tsDMARD-experienced patients for the 400mg Q4W and 600mg Q2W doses, showing a dose response relative to the 100mg Q4W dose. At baseline, patients had high disease activity with a mean CDAI of 38 and median CDAI of 36. Of the 318 rosnilimab patients in the intent to treat population, CDAI of 10 was achieved by 45% of patients by Week 12 and increased to 69% of patients across all three doses at Week 14, the timepoint required in the trial to be eligible to remain in the all-active treatment period on rosnilimab through Week 28. This sets a ceiling on the number of CDAI LDA responders at Week 28 when using a non-responder imputation analysis of the ITT population to calculate rosnilimab response rates. Max response rates for rosnilimab have not yet been observed as strict criteria at Week 14 prevented patients with meaningful improvement from continuing treatment in the trial. For example, 12 patients achieved CDAI LDA four or six weeks after their last rosnilimab dose at their first follow-up visit, but were ineligible to continue in the all-active treatment period and imputed as non-responders at Week 28 in the below NRI analysis. If these 12 responding patients had been eligible to continue treatment, CDAI LDA would have been up to 232 responders of 318 rosnilimab patients in the ITT across all three doses. Of the remaining patients ineligible to enter the all-active treatment period, ~50% had achieved ACR20 at Week 14 and were trending toward CDAI LDA. CDAI LDA responders at Week 28 had durable responses for at least two months off drug, supporting the potential for maintenance dosing with extended dosing intervals. As of the March 11, data cutoff date, 83% were still in LDA at Week 34 across all doses. Of the remaining 17% of patients who did not sustain CDAI LDA at Week 34, most remained near the cutoff of CDAI 10, with a median CDAI of 13. Rosnilimab demonstrated clinically meaningful improvements that deepened through Week 28 across multiple validated patient-reported outcomes, including the pain visual analog scale and the health assessment questionnaire, or HAQ-Disability Index, a self-reporting tool to measure function and disability. The Pain VAS scale ranges from zero to 100, where scores of 20 indicate mild pain and above 60 indicate severe pain. The rosnilimab patients who entered the all-active treatment period improved on the Pain VAS from a mean baseline of ~65 to ~15, a ~50-point change where the minimal clinically important difference is a 10-point change from baseline. HAQ-DI scores range from zero to three, where scores of one to two indicate moderate to severe disability. The rosnilimab patients who entered the all-active treatment period with a score of ~1.6 reported a 0.9-point reduction in the HAQ-DI to ~0.7, where the MCID is a 0.22-point change from baseline. Clinical outcomes were further substantiated by compelling and objective translational data. An ~50% reduction in the mean CRP from baseline, an objective measure of inflammation, was observed through Week 28 in rosnilimab patients who entered the all-active period. Additionally, translational blood and synovial biopsy biomarker data showed differentiated and consistent immunological impact with robust, on-target pharmacological activity in rosnilimab patients that was not observed on placebo. In blood, rosnilimab demonstrated rapid, deep and sustained reductions of ~90% in PD-1high T cells and ~50% in PD-1+ T cells, and an increase in total Tregs. Together, this resulted in stable total T cell counts and favorable T cell composition reflective of healthy immune homeostasis. Additionally, synovial biopsies of the most impacted joint taken at baseline and after six weeks showed a deep reduction of ~90% in PD-1+ T cells at the 400mg Q4W and 600mg Q2W doses, showing a dose response relative to the 100mg Q4W dose. Gene expression studies of the synovium demonstrated significant decreases of T cell activation and B cell activation pathways in rosnilimab patients. Similarly, highly significant decreases in additional downstream pathways including those relevant to TNF and IL-6 within the synovium were observed.