AUTHORS
Bofei Wang, PhD, Wei Xiao, PhD2, Mahesh Swaminathan, MD3, Jessica Root, MS4, Joyce Breaker5, Mariela Sivina, PhD6, Patrick Reville, MD, MPH7, William G. Wierda, MD, PhD8, Hussein Ali Abbas, MD, PhD9
AFFILIATIONS
1 The University of Texas MD Anderson Cancer Center, Houston, United States
2 The University of Texas MD Anderson Cancer Center, Houston, United States
3 The University of Texas MD Anderson Cancer Center, Houston, United States
4 The University of Texas MD Anderson Cancer Center, Houston, United States
5 The University of Texas MD Anderson Cancer Center, Houston, United States
6 The University of Texas MD Anderson Cancer Center, Houston, United States
7 Northwell Health, Norwalk, United States
8 The University of Texas MD Anderson Cancer Center, Houston, United States
9 The University of Texas MD Anderson Cancer Center, Houston, United States
ABSTRACT
Background: Chronic lymphocytic leukemia (CLL) is characterized by profound immune dysfunction that manifests in dysregulated cytokine signaling that shapes an immunosuppressive tumor microenvironment. We previously reported treatment results with fixed-duration acalabrutinib plus venetoclax (AV) +/- early obinutuzumab (OBIN), which achieved high undetectable measurable residual disease (uMRD4; 10-4 sensitivity) at end of cycle 26 (EOC26) (NCT04169737). Both arms showed similar EOC26 uMRD4 rates in first-line treatment, but delayed OBIN was associated with improved tolerability and reduced OBIN-associated toxicity. Immune mechanisms underlying differential effects to early vs. delayed OBIN remain poorly understood, and the longitudinal immune changes that occur throughout CLL treatment have not been fully characterized.
Methods: We performed longitudinal multimodal single-cell profiling of 50 samples from 17 patients with treatment-naïve CLL enrolled in the randomized trial of AV with early vs. delayed OBIN. CITE-seq simultaneously captured single-cell transcriptomes and surface-protein expression, complemented by paired T-cell and B-cell receptor (TCR and BCR) sequencing. Samples were profiled at baseline (BSL), EOC9 (primary endpoint), and EOC26. Ten patients received early OBIN during C1-C6, whereas 7 initiated AV and received delayed OBIN during C15-C20. Systemic inflammation was interrogated with high throughput blood-based proteomics in 64 pts at BSL with 18 paired samples at EOC26, with26 Healthy donors (HD) as control.
Results: Of the total 65 distinct patients in the full cohort, the median age at BSL was 61 years (range 36-79), with the most frequent mutations being ATM (29%), NOTCH1(25%), and TP53 (18%). Cellular abundance comparison at BSL and EOC9 demonstrated a significant increase of monocytes and T/NK cells with the pronounced B-cell contraction at EOC9 compared to BSL in both arms. Early OBIN was associated with relatively lower residual B lineage proportion at EOC9 compared to those who received late OBIN (p = 0.057), consistent with the superior uMRD4 rate observed (8/10 vs 1/7; p = 0.015). Within the T-cell compartment, 10 distinct subtypes were detected. Inferred trajectory revealed a continuum from naïve to terminally differentiated T-cell states. CD8 memory, CD8 GZMK+ and regulatory T-cells were significantly contracted at EOC9 in early OBIN arm. Diffusion pseudotime suggested a divergent trajectory of CD8 memory cells into two branches, enriching for cells at BSL and EOC26 respectively. TCR analysis revealed overall more clonal expansion in early OBIN group, concentrated in CD8 cytotoxic and exhaustion populations, suggesting an augmented immune response with early OBIN. The hyperexpanded clones in the early OBIN group contracted at EOC9 and expanded at EOC26, which may reflect a dynamic antigen-driven T-cell remodeling over the treatment course.
To complement single-cell analysis, we measured soluble proteins to investigate circulating inflammatory milieu. The target detectability was 97% across 251 cytokines. PCA analysis suggested clear separation between BSL CLL and HD samples, indicating a potentially distinct inflammatory profile of CLL. A random forest classifier built from the top 10 proteins selected with Boruta over 100 iterations achieved an AUC of 0.99 in discriminating CLL from HD. Differential analysis revealed that upregulation of IL13, IL5, and CXADR at BSL was associated with detectable (d)MRD4 at EOC9 in pts receiving AV. While in pts receiving early OBIN, upregulation of CCL8, CXCL9, IL2, MICB, MUC16, and SLAMF1 were associated with dMRD4. Comparison of BSL and EOC26 with HD suggested that treatment partially reprograms the inflammatory proteomic landscape of CLL toward a healthy state in both arms.
Conclusions: Integrated multi-modal analysis revealed that early OBIN drives distinct immune remodeling and T-cell clonal dynamics. Circulating inflammatory signatures further distinguished CLL from HD and identified treatment-associated inflammatory patterns associated with residual disease, highlighting divergent immune mechanisms underlying response to early versus late OBIN. Ongoing B-cell and BCR analysis, interactome prediction, and paired proteomic comparisons will be presented at the annual meeting. These studies aim to characterize potentially immune restorative cellular and cytokine shifts associated with highly effective combined targeted therapy and uMRD4 remission.
DISCLOSURES
Bofei Wang, PhD: No relevant conflicts of interest to declare.