Lymphoma response to chimeric antigen receptor T cell therapy is driven by endogenous tumor-reactive CD8⁺ T cells

CD19-targeted CAR-T cell therapy has become an important treatment for relapsed or refractory large B-cell lymphoma. In the pivotal trial of axi-cel, the objective response rate was 82% and the complete response rate was 54%^[2], while the best overall response rate of tisagenlecleucel in adults with diffuse large B-cell lymphoma reached 52%^[3]. Long-term follow-up data indicate that a minority of these patients can achieve durable, and potentially curable, responses^[1]. The problem, however, lies precisely here: most patients either have primary nonresponse or eventually relapse, and the academic community has consistently lacked a systematic explanation of the biological basis for differences in response durability. Previous understanding of CAR-T focused primarily on CAR-T cells as standalone cytotoxic killing tools. The quality, expansion, and persistence of the infused product were regarded as key variables, whereas the role of the host’s own immune system remained unclear.

This new study directly addresses this gap. The investigators integrated whole-genome and transcriptomic profiling from 54 patients and combined functional experiments, syngeneic mouse models, and mathematical modeling to propose a previously underrecognized mechanism: durable CAR-T efficacy requires collaboration between CAR-T cells and recruited endogenous CD8⁺ T cells (eCD8) participating in the antitumor response. Specifically, tumors carrying biallelic loss of MHC class I molecules or CD58 exhibited high genomic complexity, increased neoantigen burden, and resistance to CAR-T therapy. In immunocompetent hosts, the response to MHC-I-deficient lymphoma was inferior to that in wild-type tumors. The mouse experiments were more persuasive: sustained remission occurred only when eCD8 cells were present, and all mice treated with CD8⁺ cell depletion relapsed after the initial tumor regression mediated by CAR-T. ELISPOT assays of patient samples and mice confirmed enhanced activation of endogenous T cells after CAR-T infusion.

Relative to existing understanding, the advance of this work is conceptual: CAR-T therapy should not be viewed merely as a cytotoxic approach, but also as a catalyst of the host’s endogenous antitumor immunity. This framework expands the role of CAR-T from a “direct killer” to an “immune response igniter,” echoing the clinical difficulty of fully predicting long-term outcomes based only on the intrinsic properties of CAR-T cells. It also provides mechanistic clues for understanding how immune escape alterations, such as MHC-I or CD58 loss, affect CAR-T efficacy. Interestingly, tumor MHC-I loss does not impair CAR recognition of antigens, yet it reduces CAR-T efficacy in hosts with an intact immune system. This observation itself indicates that the endogenous T-cell response is an indispensable component of the efficacy equation.

This study also has its limitations. A sample size of 54 remains limited given the high heterogeneity of lymphoma. The gap between mouse models and the authentic immune ecosystem in humans, the simplifying assumptions of mathematical modeling, and the question of how to translate eCD8 status into actionable treatment stratification or combination strategies in clinical practice—for example, synergy with immune checkpoint inhibition or other approaches—all remain unresolved. Identifying which patients can be “catalyzed” into mounting a sufficiently strong endogenous response, and determining whether this can be predicted and intervened upon before infusion, will be the critical step in translating this discovery into clinical practice.