Clinical and biological effects of combined CD27 and CD20 antibody therapy in relapsed/refractory B-cell lymphoma: the Riva trial

Abstract

Background: CD27 antibody stimulation of T cells has been shown to activate and promote myeloid cell infiltration leading to enhanced antibody-dependent cellular phagocytosis (ADCP) by anti-CD20 in lymphoma preclinical models (Turaj et al Cancer Cell 2017). In this phase IIa study (RiVa NCT03307746), the safety and efficacy of rituximab (ritux) and varlilumab (varli, anti-CD27) was tested in relapsed/refractory B-cell lymphoma. Transcriptional effects of therapy were investigated by RNA sequencing of pre- and post-treatment tumour biopsies and single cell RNA sequencing of in vitro peripheral blood mononuclear cells (PBMC).

Methods: eligible patients with relapsed/refractory CD20 + B-cell lymphoma were randomized 1:1 to arms A: cycle 1 day 1 ritux, day 2 varli; or B: cycle 1 day 1 ritux, day 8 varli. Cycles 2-6 are identical in both arms (day 1 ritux (cycles 2-6); day 2 varli (cycles 3 and 5); 2-weekly cycles. The primary endpoints were overall response and safety. RiVa was funded by CRUK (CRUKD/17/008) and Celldex Therapeutics Inc.

Intratumoral biopsies were taken pre-treatment and on cycle 1 day 7/8 post treatment, i.e. post- ritux and varli in arm A, and post-ritux alone in arm B, and subjected to RNA sequencing, deconvolution by CIBERSORTx and Gene Set Enrichment Analysis (GSEA). To gain further insight into how anti-CD27 stimulated T cells activate myeloid cells, PBMC from healthy donors were treated with varli or an isotype control for 48 h and analysed by CITE-seq (10x Genomics).

Results: twenty-seven patients were randomised; 15 indolent B-cell non-Hodgkin lymphoma cases (NHL) (1 mantle cell lymphoma and 14 follicular lymphoma (FL) grade 1, 2 or 3a) and 12 aggressive B-NHL (9 diffuse large B-cell lymphoma, 2 FL grade 3b and one transformed FL). Median age was 71 (range 49-87), median lines of previous treatment were 4 (range 1-13) and 22% were refractory to the last line of treatment. Thirteen patients completed all 6 cycles of treatment, 1 patient was withdrawn after 1 cycle due to a new cancer diagnosis, 3 patient/investigator withdrawals occurred and 10 progressed on treatment. The overall response rate at the end of treatment was 26.9% (7/26; 95% CI, 13.4-44.7) (indolent B-NHL 26.7% (4/15); aggressive B-NHL 27.3% (3/11)). Within responders, 4 had partial response (PR) (3 aggressive B-NHL, 1 indolent B-NHL) and 3 had stable disease (SD) (all indolent B-NHL), with a duration of response between 2 months to >1 year. Amongst the responders with aggressive B-NHL, one had stage III disease, 3 previous lines of treatment and remained in remission >1 year. Another had stage IV disease, 6 previous lines of treatment including CAR-T cells, and was in remission at last follow up (>16 days). Thirty-three percent (9/27) of patients experienced at least one adverse event graded ≥3 with the commonest being infection (11%, 3 cases).

GSEA of post- vs pre-treatment biopsies in ritux/varli-treated (arm A) patients showed pathways enriched in T-cell signalling (normalised enrichment score (NES) 2.47, q-value=0.001) and Fc gamma receptor-dependent phagocytosis (2.05, q=0.001), which were absent in ritux-treated (arm B) patients. CIBERSORTx analysis showed >100-fold increased CD4 T cell infiltration in partial responders compared to those with progressive disease (p=0.027) and a strong correlation between intratumoral B-cell depletion and macrophage infiltration (Fig 1). Responders were enriched in T-cell signalling and Th1 signatures, pre-treatment (NES 2.52, q=0.02).

PBMC cultures treated with varli induced CD4 + and CD8 + T-cell effector and memory activation (IFNG, TNFA, NFKB), and monocyte differentiation from classical (CD14 ++CD16 -) into intermediate (CD14 ++CD16 +) and dendritic cell-like (CD14 -CD16 +CD83 +CD86 +) phenotypes. In comparison to isotype control data, receptor-ligand interaction analysis predicted that varli-treated T cells may activate monocytes through the MIF signalling pathway (CD74, CD44, CXCR4).

Conclusion: combined rituximab and varlilumab administration is safe in relapsed/refractory B-cell lymphomas and demonstrates efficacy in patients with T-cell activated tumours. Transcriptomic analysis of pre- and post-treatment samples confirms that varlilumab has in vivo agonistic activity. B-cell depletion by rituximab is dependent on intratumoral macrophages and varlilumab induces myeloid cell activation via a T-cell dependent, MIF signalling pathway

More information

Identifiers

ORCID for Cherish Boxall: orcid.org/0000-0002-7850-233X

ORCID for Louise Stanton: orcid.org/0000-0001-8181-840X

ORCID for Gareth Griffiths: orcid.org/0000-0002-9579-8021

Catalogue record

Export record

Altmetrics