Abstract
Chimeric antigen receptor T cells (CAR-T) targeting B-cell maturation antigen (BCMA) have transformed the treatment landscape for multiple myeloma. However, despite durable responses in some patients, ~20% develop delayed hematologic recovery after CAR-T. We previously reported that prolonged cytopenias beyond 100 days are associated with persistently elevated inflammatory markers and serve as an early surrogate for reduced long-term survival (Avigan et al., Clin Cancer Res 2025). The endogenous T cell repertoire and recovery following lymphodepletion and CAR-T have similarly been linked to clinical outcomes, potentially serving both as a marker and important mediator of immune dysregulation (Cheloni et al., Nat Commun 2025).We therefore hypothesized that CAR-T hematologic toxicity is associated with a distinct T cell recovery pattern, which may drive both ongoing toxicity and disease response.
We retrospectively analyzed two independent cohorts of cilta-cel treated patients with either bulk T cell receptor β-chain (TCR) next-generation sequencing (NGS cohort: 21 patients, 32 samples) or single-cell CITE-seq with paired TCR sequencing (CITE-seq cohort: 14 patients, 30 samples) from peripheral blood. Samples in the NGS cohort were collected clinically at varying timepoints within 2 years post-CAR-T and grouped into 6-month intervals. Patients in the CITE-seq cohort had samples collected at baseline and at least one sample collected within 6 months of CAR-T. For analysis, patients were divided by hematologic recovery before or after 100 days, with recovery defined as the first of a consecutive 30-day period without grade 3 toxicity or growth-factor support. Expanded TCR clones were defined as >3x the upper limit of the polyclonal background, per our center's validated protocol, and TCR diversity was calculated by Shannon equitability index with the top 10 clones in the NGS cohort and all clones in the CITE-seq cohort. TCR clonality was compared using Fisher's exact test and diversity by Wilcoxon rank-sum test. Multivariable linear regression was used to model diversity over time, adjusting for age and prior lines of therapy.
Patients in the NGS cohort had a median age of 63.5 with a median of 5 prior lines of therapy, and 11 (52%) had early hematologic recovery versus 10 (48%) with delayed recovery beyond day 100. Those with delayed recovery showed a trend toward increased monoclonal TCR populations (60% vs 27%, p=.08) and significantly lower TCR diversity beyond 6 months post-CAR-T (p=.029). Furthermore, a multivariate linear regression of native TCR populations showed stable clonal diversity over the 2 years post-infusion in patients with early hematologic recovery, whereas those with delayed hematologic recovery had a progressive loss of endogenous TCR diversity over this period (p=.032).
We next reviewed patients with available CITE-seq as a validation cohort. These patients had a median age of 61 and a median 5 prior lines of therapy, and 10 (71%) showed early hematologic recovery while 4 (29%) had delayed recovery. Baseline TCR diversity did not significantly differ by hematologic recovery group. However, after lymphodepletion and CAR-T infusion, patients with delayed recovery showed a trend toward reduced diversity within the CD8+, CAR-negative T cell population (p=.07), consistent with the NGS cohort. Delayed recovery patients also had significantly more CD4+ T regulatory cells at baseline (p=.04) and decreased CD4+ memory T cells after infusion (p=.02), though CD4+ TCR diversity did not differ between groups. Finally, TCR sequences across both cohorts were cross-referenced against the VDJdb and IEDB databases to identify potential matches to known TCR-antigen specificities. Despite differences in T cell expansion and clonality, over 88% of functional TCR sequences across both groups had unknown antigenic targets, and our current efforts are aimed at defining the antigenic specificity and functional role of these expanded T cell populations.
We demonstrated in two independent cilta-cel cohorts that prolonged cytopenias after CAR-T are associated with impaired endogenous T cell recovery characterized by clonal expansion and diminished diversity of CAR-negative CD8+ T cells. These findings support the hypothesis that disruptions in native T cell recovery after lymphodepletion and CAR-T infusion may promote hematologic toxicity and serve as an early marker for immune dysregulation associated with poor outcomes.
