The Impact of T Cell Immunity on Chemotherapy Response in Childhood Acute Lymphoblastic Leukemia

While most childhood acute lymphoblastic leukemia (ALL) can be cured with risk-adapted combination chemotherapy, a significant proportion of patients experience relapse and eventually succumb to leukemia. In fact, ~50% of ALL relapses occur in patients who have excellent early treatment response (i.e., negative minimal residual disease [MRD] at the end of induction therapy), suggesting that long-term remission is determined by factors beyond leukemia cell drug sensitivity. Inherited genetic variants in genes related to host immune functions have been linked to ALL prognosis, pointing to a possible role of immune vigilance against residual leukemia. However, little is known about how and which host immune factors influence the long-term-remission of this cancer.

To this end, we first compared leukemia-free survival of murine BCR-ABL1 B-ALL in response to chemotherapy, in immunocompetent mice (C57BL/6J, "B6") or mice with various T cell deficiencies. Complete loss of all T cells in the host (Tcra ^KO) led to a drastic increase in leukemia relapse after treatment with ABL inhibitor dasatinib (p=0.0014) or cytotoxic chemotherapy (e.g., dexamethasone [p=0.0002], and mercaptopurine [p=0.0106]). Similar results were observed when we specifically depleted CD4 or CD8 T cells in host mice (p=0.0302 and 0.0010 for CD4 and CD8 depletion, respectively). ABL1 mutation analysis showed that the immunocompetent mice were much more likely to relapse with wild-type BCR-ABL1 B-ALL after dasatinib withdrawal (71.4%). By contrast, most immunocompromised mice (60.0%, 75%, and 80% of Tcra^KO, CD4 T cell depletion, and CD8 T cell depletion) relapsed with ABL1-T315I B-ALL. Deep sequencing of BCR-ABL1 mutations during dasatinib therapy indicated that drug resistance mutations emerged early in leukemia regardless of immunocompetency of the host, but the lack of T cell immunity strongly predisposed mice to subsequently develop drug resistance and thus relapse. Furthermore, immunocompetent mice that achieved long-term remission also rejected re-engraftment of BCR-ABL1 ALL, suggesting immune memory against leukemia established during dasatinib treatment.

To explore the molecular basis of T-cell immunity against ALL, we collected peripheral T cells during dasatinib therapy from mice which were eventually cured vs. those experienced relapse and compared their global gene expression pattern. Through expression array-based profiling of bulk T cells, we noted upregulation of memory/effector T cell signature genes in mice achieving long-term cure (especially in CD8 T cells), along with genes associated with TCR signaling and cytokine response (IL12 and IFNG). Similarly, scRNA-seq profiling also confirmed the association of TCR signaling, IL12 and IFNG signaling activation in T cells with the cure of ALL in mice. Single-cell RNA-seq analyses of T cell subpopulation specifically identified differences in memory-like T cells from relapse mice vs. those cured with the same chemotherapy, with the latter showing more robust activation of TCR and cytokine signalings. We thus hypothesized that IL12 and IFNG response signaling are key determinants of successful immune surveillance against ALL. Indeed, our in vivo experiments confirmed that 1) Ifng^KO mice relapsed much more quickly during dasatinib treatment with drug resistance BCR-ABL1 mutations than immunocompetent mice (p<0.0001), and 2) the addition of IL12 to dasatinib therapy significantly improved leukemia-free survival in immunocompetent mice with BCR-ABL1 B-ALL (p=0.0028).

Finally, we validated these findings by evaluating the immune cell composition of peripheral blood samples collected during remission from 105 children with ALL treated on the St. Jude Total Therapy Study XV clinical trial. Using the gene expression-based deconvolution method, we quantified the frequency of 9 types of immune cells. There was a significant association of higher level of total T cell (CD8, CD4 naïve, CD4 memory, follicular helper, regulatory, and gamma delta T cells) with superior event-free survival (p=0.031), whereas other immune cells had no impact on ALL treatment outcomes. Even after adjusting for ALL molecular subtypes and MRD levels, T cell frequency remained prognostic (p=0.034).

In conclusion, our results firmly established the role of T cell immunity in the cure of ALL and also pointed to T cell activation as a means of improving treatment outcomes of ALL.