Biological and Clinical Significance of Endogenous and Chimeric Antigen Receptor (CAR) T Cell Immune Profiling in Relapsed/Refractory Multiple Myeloma (RRMM) Patients

Background: Anti-BCMA CAR T cells are a new standard of care for the treatment of RRMM patients. Hence, it is paramount to identify determinants of response and resistance for individualized approaches. There is emerging data about the importance of depth of response in prolonged progression-free survival (PFS) and of genetic antigen escape in treatment resistance. By contrast, there is limited understanding of how patient’ outcomes are shaped by immune fitness prior aphaeresis, and virtually no data about CAR T and endogenous profiles post infusion.

Aim: Characterize immune dynamics and investigate its prognostic value in RRMM patients treated with CAR T cells.

Methods: Immune profiling was performed using multidimensional and computational flow cytometry in 285 bone marrow aspirates from 120 RRMM patients enrolled in the KarMMa phase 2 trial. In total, 170 immune clusters were identified and systematically quantified in a pileup analysis of samples collected before and at months 1, 3, 6 and 12 after CAR T infusion. Immune scores were developed using cox regression with 10-fold cross-validation.

Results: At baseline (n = 63), there were significant associations between the bone marrow immune composition and the number of prior lines of treatment, prior autologous transplant and staging. Namely, there were higher proportions of ICOS, PD1, TIGIT and TIM3 positive cells within the CD4+ and CD8+ T cell compartments from heavily pretreated patients and those with R-ISS 3. A baseline immune score defined by the percentages of GrzmB+HLADR+ cytotoxic NK cells, TIGIT+ Tregs, CD4+CD127+ICOS+ and CD4+CD25+CD127+TIGIT+, CD8+TIGIT+, CD8+GrzmB+HLADR+ and CD8+CD27+CD95+CD45RAloCD197lo T cells was associated with PFS. Patients with low vs high risk immune score at baseline had median PFS of 25 vs 6 months, respectively (p < .001). Interestingly, there were no significant associations between the baseline risk score and treatment or staging, suggesting potential complementarity of immune profiling with other prognostic factors.

At month 1 after infusion (n = 55), the presence of >1% vs <1% CAR T cells was associated with longer PFS (medians of 20 vs 8 months, p = .002). In addition to the magnitude of CAR T persistence, it was interesting to observe increased percentages of CD8+ CAR T cells in patients with <5 prior lines of therapy, of CD4+HLADR+ and CD8+TEMRA CAR T cells in patients with R-ISS 1/2, as well as of PD1+TIM3+ CD4+ and CD8+ CAR T cells in patients with R-ISS 3. Moreover, the phenotype of CAR T cells was associated with PFS. Patients with low vs high risk immune score (ie, based on the percentages of CD4+CD57+, CD8+CD27-EM, CD8+PD1+TIM3+ and CD4+CD8+ CAR T cells) had median PFS not reached vs 11 months, respectively (p < .0001). Similarly to the above, there were no significant associations between the CAR T immune risk score and treatment or staging, suggesting potential complementarity of CAR T quantification and phenotyping post-infusion with other patient’ characteristics.

Next, we investigated endogenous T cell dynamics in 51 RRMM patients with paired samples prior and after CAR T infusion. There were significant differences in the frequencies of 30 immune clusters, including CD27+ naïve and EM CD4+ and CD8+ T cells, ICOS+ and TIGIT+ Tregs, CD4+ and CD8+ T cells, as well as of CD25+CD127+ CD4+ and CD8+ T cells. Furthermore, an immune risk score defined at the latest available time point (n = 95) was associated with PFS (p < .0001). Patients with lower levels of GrzmB+ cytotoxic NK cells and of immunoregulatory HLADR- NK cells, as well as of CD4+CD27+CD95+CM, CD4+HLADR+, CD8+PD1+ and CD8+CD27-CD95+EM T cells showed a high risk score and a median PFS of 6 vs 20 months in patients with low risk immune score.

Conclusions: This is one of the largest dynamic studies of immune profiling in a clinical trial investigating anti-BCMA CAR T cells in RRMM. Our results uncovered that patient’ immune status is shaped by prior treatment and staging, and that the distribution of different NK, Treg, CD4+ and CD8+ T cells prior aphaeresis is associated with PFS. Furthermore, the phenotype of CAR T cells at month 1 after infusion may be as prognostic as its percentage. We also showed how treatment with CAR T cells modulates the bone marrow immune composition and how it may contribute to immune escape vs surveillance. Altogether, this study supports immune profiling for improved understanding of response and resistance to CAR T cells in RRMM.