KIR2DS1–HLA-C status as a predictive marker for benefit from rituximab: a post-hoc analysis of the RICOVER-60 and CLL8 trials

Abstract

The addition of rituximab to chemotherapy has substantially improved outcomes for patients with B-cell malignancies. The mechanisms of action of rituximab include activation of natural killer cells. Killer-cell immunoglobulin-like receptors (KIRs) mediate natural killer cell function through interaction with HLA. We evaluated the clinical impact of KIR-HLA genotypes on rituximab-containing therapy. For this post-hoc analysis, we used data from the RICOVER-60 trial (NCT00052936) as the discovery cohort and the CLL8 trial (NCT00281918) as the validation cohort. RICOVER-60 included patients aged 61-80 years with aggressive B-cell lymphoma treated with CHOP (cyclophosphamide, doxorubicin, vincristine, and prednisone) with or without rituximab. CLL8 included patients aged 30-81 years with chronic lymphocytic leukaemia treated with chemotherapy (fludarabine and cyclophosphamide; FC) with or without rituximab. We evaluated the KIR and HLA-C status of 519 patients with available blood samples in the RICOVER-60 trial and the KIR2DS1 and HLA-C status of 549 patients with available blood samples in the CLL8 trial, and evaluated their associations with event-free survival (RICOVER-60), progression-free survival, and overall survival (RICOVER-60 and CLL8). In the RICOVER-60 trial, 201 (39%) patients were positive for KIR2DS1, 79 (15%) were homozygous for HLA-C2, and 36 (7%) were positive for KIR2DS1 and homozygous for HLA-C2. In the CLL8 trial, 206 (38%) patients were positive for KIR2DS1, 75 (14%) were homozygous for HLA-C2, and 26 (5%) were positive for KIR2DS1 and homozygous for HLA-C2. In the RICOVER-60 trial, both KIR2DS1 and HLA-C status were identified as independent risk factors for survival. KIR2DS1 positivity, homozygosity for HLA-C2, and subsequent KIR2DS1-HLA-C status were associated with adverse clinical outcome in patients receiving rituximab-containing therapy (event-free survival for patients with KIR2DS1-HLA-C2/C2 vs all other patients, HR 2·6 [95% CI 1·4-4·7], p=0·0015; progression-free survival, 2·7 [1·5-5·1], p=0·0013; overall survival, 2·8 [1·5-5·4], p=0·0016) but not in patients receiving CHOP chemotherapy only (event-free survival, 0·9 [0·5-1·7], p=0·85; progression-free survival, 1·1 [0·6-2·0], p=0·81; overall survival, 1·2 [0·6-2·4], p=0·53). A significant interaction between KIR2DS1-HLA-C status and rituximab was observed (p=0·018 for event-free survival and p=0·034 for progression-free survival). In contrast to all other patients, those positive for KIR2DS1 and homozygous for HLA-C2 did not benefit from adding rituximab to CHOP chemotherapy (event-free survival, 1·9 [0·8-4·6], p=0·16; progression-free survival, 1·4 [0·6-3·4], p=0·48; overall survival, 1·6 [0·6-4·3], p=0·33). In the CLL8 trial, KIR2DS1-HLA-C status was confirmed as a predictive marker for benefit from rituximab therapy (p=0·024 for the interaction of KIR2DS1-HLA-C status and rituximab regarding progression-free survival). In contrast to all other patients, those positive for KIR2DS1 and homozygous for HLA-C2 did not benefit from adding rituximab to FC chemotherapy (progression-free survival, 2·1 [0·9-4·9], p=0·094; overall survival, 2·6 [0·5-12·7], p=0·21). Assessment of KIR2DS1 and HLA-C genotype might identify patients who would not benefit from rituximab, thereby allowing alternative therapies to be given. Further validation of these findings in prospective clinical trials is needed. F Hoffman La Roche.