Biomedical Research · us
Why Does CAR-T Efficacy Vary from Person to Person? Inherited Genetic Differences Affect Cell Expansion and Toxicity
A whole-genome analysis of more than 200 patients with aggressive lymphoma linked variants in STXBP2, ADAMTSL3, and PTPN22 to inflammatory toxicity, a protective effect, and CAR-T expansion, respectively. The findings open new directions for cell therapy design but are not yet sufficient to directly predict individual patient risk.
After CAR-T cells enter a patient’s body, they sometimes proliferate extensively and effectively attack tumors, but at other times are accompanied by severe inflammatory reactions. In addition to tumor characteristics, disease burden, and manufacturing differences, a study published in *Science Immunology* indicates that genetic variants patients are born with may also help determine how this “living drug” works.
Teams from the Mass General Brigham Cancer Institute, the Broad Institute, and the Dana-Farber Cancer Institute analyzed whole-genome sequences from more than 200 patients with aggressive lymphoma enrolled in two CAR-T clinical trials, then compared genetic variants with post-treatment toxicity and CAR-T cell expansion. The study focused not on mutations acquired by tumors, but on inherited genetic differences present from birth and also found in patients’ T cells.
In one of the trials, variants that suppress STXBP2 expression were linked to CAR-T-related toxicity. The researchers further engineered donor T cells to lack STXBP2 or carry a variant capable of reducing the gene’s expression, resulting in stronger inflammatory effects and providing functional evidence supporting the clinical observations. However, this association was not reported as having been replicated in both trials, so it still needs confirmation in additional independent patient cohorts.
Two other findings extended across both trials: ADAMTSL3-related variants were associated with lower treatment toxicity, while PTPN22 variants were linked to more pronounced CAR-T cell expansion. Robust CAR-T expansion in the body is generally associated with treatment activity, but it is not itself tumor remission or long-term survival. At this stage, the findings also do not establish that patients carrying specific PTPN22 variants will necessarily achieve better efficacy.
These findings highlight a fundamental difference between CAR-T and conventional drugs: autologous CAR-T products are made from each patient’s own cells, so their genetic background also enters the final product. If the associations are validated on a larger scale, genetic information may eventually help with risk stratification. In the field of allogeneic CAR-T, in which a single donor supplies cells for use in multiple people, it may also inform donor selection and cell engineering.
Several barriers remain before clinical application. The study included only slightly more than 200 people from two lymphoma trials, and some signals appeared in only one trial. Observational genetic associations are also susceptible to influences from population composition, treatment products, and clinical conditions. The team’s cell experiments strengthened the biological plausibility of STXBP2, but do not mean that causal relationships have been established for all three genes. The next step is still to replicate the results across different populations, cancer types, and CAR-T platforms, and to demonstrate that adding genetic information genuinely improves decision-making rather than merely adding a seemingly precise risk marker.