Atezolizumab Combined with Venetoclax and Obinutuzumab for Frontline CLL

Background

Dysfunction of immune subsets is common in patients (pts) with CLL. Preclinical studies showed synergy of venetoclax (VEN) with immune checkpoint inhibitors (ICI) (Kohlhapp, Cancer Discovery 2021). We hypothesized that combined VEN, obinutuzumab (OBIN, CD20 mAb) and atezolizumab (ATEZO), a PD-L1 checkpoint inhibitor will be synergistic. We previously reported early results of combined ATEZO, VEN and OBIN (Jain, ASH 2021) in pts with previously untreated CLL showing high remission rate with undetectable measurable residual disease (uMRD) status.

Methods

This is an investigator-initiated phase 2 trial of VEN, OBIN and ATEZO in pts with untreated CLL meeting IWCLL treatment criteria (NCT02846623). Key eligibility criteria include age≥18 years and adequate organ function (total bilirubin≤1.5xULN, ALT/AST≤2.5xULN, creatinine≤1.5xULN). OBIN was given at 100mg IV cycle (C)1 day (D)1, 900 mg C1D2, 1000mg on C1D8 and C1D15 and then 1000mg on C2-9 D1. ATEZO was given at 1680 mg IV (over 2 days) on C1D3-4 and then C2-9D1-2. VEN was initiated in C3 with standard weekly dose-escalation (20mg up to a target dose of 400mg daily) and was continued until end of C14 (total 12 cycles of VEN). All pts stopped therapy at the end of C14. Response assessments were done with CT and bone marrow with flow cytometry assessment (MRD4; sensitivity 10^-4) at the end of C2 (prior to VEN initiation), C6, C9, and C14. Additionally, we performed single-cell RNA sequencing (scRNA-seq) with paired TCR/BCR immune repertoire profiling on 12 bone marrow samples from 6 patients selected based on their response. Paired samples were collected at baseline and end of C2, prior to initiation of VEN to isolate ICI in CLL bone marrow microenvironment. scRNAseq analysis was conducted in R.

Results

From July 2019 to February 2023, 37 pts were enrolled. The median age was 62 years (range, 21-77). A total of 21/37 (68%) had unmutated IGHV gene, 10/37 (27%) had del11q, and no pts had del(17p)/mutated TP53. 59% had a baseline lymph node size >5cm.

The median follow-up is 40 months. Two pts came off study in C1 (1 for hematoma after anticoagulation initiation for DVT and 1 for ICI colitis). 35 pts initiated VEN. 77% pts downgraded their TLS risk prior to starting VEN after 2 cycles of ATEZO/OBIN. An additional 4 pts discontinued study prior to completion: 2 deaths from COVID19, 1 developed second cancer (GIST) needing treatment, and 1 for ICI hepatitis. 31 pts completed C14, of which 30 (97%) were bone marrow uMRD4 and 1 pts was low+ MRD. 5/30 (17%) pts had an MRD4 relapse at a median 12.2 months after therapy completion including 3 progression events, but none have required subsequent CLL therapy to date. 2- and 4-year PFS estimates are 94% (95% CI 86 -100%) and 89% (95% CI 76 -100%), respectively. 2- and 4-year OS estimates are both 94% (95% CI 86 -100%). Atezolizumab was discontinued early in 3 pts due to immune-related adverse events (irAEs). Grade 3-4 neutropenia occurred in 22/37 (59%) pts. Grade 3-4 thrombocytopenia occurred in 12/37 (32%) pts.

We selected 6 pts with paired bone marrow samples, 3 early responding pts (mean C2 MRD level 2.2%) and 3 early non-responding pts (mean C2 CLL MRD level 44.9%) for correlative scRNAseq analyses. Focusing on T and NK cells as the major mediators of immunotherapy response, differential abundance revealed a CD8+ T cell cluster with high expression of the transcription factor NR4A2 enriched in responders, whereas high expression of the canonical exhaustion transcription factor, TOX, marked a CD8 T cell cluster enriched in non-responders (p<0.001). Gene set enrichment analysis identified TNFA signaling pathways strongly enriched in responding T and NK cells. Immune repertoire analysis highlighted increased TCR expansions and transitions in responding patients. Lastly, using cell-cell interaction predictions, we observed higher interactions between CLL cells and immune cells in responders, prominently with monocyte-T/NK and monocyte-CLL interactions with enrichment in antigen presentation pathways in responding monocytes.

Conclusions

Treatment with VEN, OBIN, ATEZO results in high rates of durable uMRD4 remission with 97% bone marrow uMRD4 among evaluable pts at the end of therapy and 4-year PFS rate of 89% in the whole population. 3 pts experienced irAEs requiring early discontinuation of ATEZO. Using single cell transcriptomic profiling, we identified potential mediators of immune checkpoint blockade response in CLL.