Background: Why relapse remains a challenge after CAR T-cell therapy
Chimeric antigen receptor (CAR) T-cell therapy has transformed the treatment of relapsed or refractory B-cell acute lymphoblastic leukemia (B-ALL), especially for patients who have not responded to standard chemotherapy. In CD19/CD22 dual-targeting CAR T-cell therapy, T cells are engineered to recognize two B-cell surface antigens, CD19 and CD22, with the goal of reducing relapse caused by loss of a single target. While this strategy improves coverage compared with single-antigen therapy, relapse still occurs in a meaningful proportion of patients.
This study explored why relapse happens after dual-target CAR T-cell therapy and whether the causes differ between early and late relapse. The findings help explain why some patients lose response despite an initially strong remission and highlight new molecular clues, including the role of PAX5 abnormalities.
Study design and patient population
The researchers analyzed 91 patients with B-ALL treated with CD19/CD22 dual-target CAR T cells in a clinical trial (ChiCTR-OPN-16008526). Among the treated patients, the complete remission rate was 91.9%, showing that the therapy was highly effective at inducing initial response.
Of the responders, 32.9% (26 out of 79) later relapsed, with a median time to relapse of about 7 months. The investigators then compared early relapses and late relapses to identify the dominant mechanisms behind disease recurrence.
Main mechanisms of relapse
The study found two major relapse pathways:
1. CAR T-cell functional insufficiency: The engineered T cells failed to persist long enough or lost activity over time. This was the most common cause overall.
2. Antigen insufficiency: Leukemia cells reduced or lost expression of CD19 and/or CD22, making them harder for CAR T cells to detect and eliminate.
In late relapses, more than 90% were linked to CAR T-cell functional insufficiency. In other words, the leukemia often returned because the CAR T cells did not remain active or durable enough to provide long-term surveillance. Late relapse was also associated with the reappearance or recovery of CD19-positive B cells, suggesting that immune control had weakened.
Early relapses were more diverse. Roughly 30.8% were due to CAR T-cell functional insufficiency, and another 30.8% were due to antigen insufficiency. This means that in early relapse, both failure of the therapy itself and immune escape by the leukemia contributed in nearly equal measure.
Role of CD19/CD22 downregulation
One important finding was that a subset of early relapses showed simultaneous downregulation of both CD19 and CD22. This is clinically significant because dual-targeting strategies are designed specifically to reduce the chance that leukemia escapes by losing one antigen alone.
Despite this strategy, the leukemia in some patients still adapted by lowering both target molecules. Two early relapse cases had this pattern and were found to carry preexisting disruptions in the PAX5 gene, either a deletion or a frameshift insertion. These mutated subclones were already present before treatment, then survived and expanded after therapy.
In one patient, additional CD19 mutations appeared later, suggesting that the leukemia continued to evolve under treatment pressure.
PAX5 as a key molecular driver
PAX5 is a transcription factor essential for normal B-cell development. It helps regulate genes involved in B-cell identity, including CD19 and CD22 expression. When PAX5 is disrupted, B cells can lose their normal program and become less visible to therapies that rely on B-cell surface markers.
To test this mechanism, the investigators performed PAX5 knockout experiments in leukemic cells. Removing PAX5 reduced CD19 and CD22 expression and weakened proximal enhancer activity, which are regulatory elements that help turn these genes on. As a result, the leukemia cells became less sensitive to CAR19/22 T-cell killing in vitro.
This provides a biologically plausible explanation for why PAX5-mutated subclones can drive early relapse: they may already be primed to express lower levels of the very antigens targeted by CAR therapy.
What the findings mean clinically
The study offers several practical lessons for clinicians and researchers:
– CAR T-cell persistence matters greatly. Even when remission is achieved, long-term disease control depends on maintaining functional CAR T cells.
– Monitoring for antigen loss is important. Relapse is not always due to CAR T-cell failure alone; leukemic cells can also escape by altering CD19 and CD22 expression.
– Genomic profiling before and after therapy may identify patients at risk. Detecting PAX5 mutations or related B-cell developmental defects could help predict relapse biology.
– Dual-target therapy is helpful but not foolproof. Targeting two antigens reduces, but does not eliminate, immune escape.
The data suggest that CAR T-cell functional insufficiency was the most frequent driver of relapse overall, accounting for 61.5% of cases, while antigen insufficiency was the second most common cause at 15.4%. A distinct subset of early relapses appeared to be driven by disruptive PAX5 mutations.
Broader implications for future treatment
These findings may guide next-generation CAR T-cell design. Potential strategies include improving CAR T-cell durability, adding mechanisms to sustain T-cell activity, using combination therapies to prevent antigen escape, and incorporating molecular risk stratification before treatment.
For patients with PAX5-mutated disease, researchers may need to consider alternative or additional targets beyond CD19 and CD22. More broadly, this study reinforces a key principle in immunotherapy: cancer can escape not only by hiding its targets, but also by outlasting the immune cells meant to destroy it.
Conclusion
In patients with B-ALL treated with CD19/CD22 dual-target CAR T-cell therapy, relapse remains a significant problem despite high initial remission rates. The main causes are CAR T-cell functional insufficiency and antigen insufficiency. Early relapse can also be driven by preexisting PAX5 mutations that reduce CD19/CD22 expression and promote immune escape. These insights may support more precise risk assessment and smarter CAR T-cell strategies in the future.