B cell leukemia co-opts a hematopoietic stem cell-CD4 T cell tolerance program that facilitates relapse [CITE-Seq]

Despite advances in immune therapies, leukemic relapse is still a major obstacle to long-term survival. Using a murine model of B cell leukemia (B-ALL), we show that B-ALL co-opts a hematopoetic stem cell (HSC)-driven program to induce immune-suppressive type-1 regulatory (Tr1) CD4 T cells. The immune-suppressive function of Tr1s that protects healthy HSCs is utilized by B-ALL to evade T-cell surveillance. Using single-cell approaches, we show that B-ALL induces a Tr1 phenotype in all CD4 T cells irrespective of their initial polarization state. Unlike exhausted T cells, Tr1s do not have epigenetic silencing of effector cytokines. Both IL10R- and PDL1-blockade mitigate Tr1 expansion upon treatment with the tyrosine kinase inhibitor nilotinib. However, PDL1-blockade alone expands Th1-memory CD4s and effector CD8s that prevent relapse. We find a similar population of CD4 Tr1s in human B-ALL patients, which correlate with worse prognosis. Thus, B-ALL co-opts an HSC-driven program to induce Tr1s that suppress anti-leukemia T cell responses, which facilitates relapse. Overall design: For the scATAC-seq study, CD45.1+ mice were injected with CD45.2+ CD4 T cells from a leukemia-specific TCR transgenic mouse (HV1 CD4 T cells) via the tail vein, followed by a challenge with LM138 leukemia cells via the tail vein the next day. HV1 CD4 T cells from leukemic CD45.1+ mice at days 9, 13, and 17 post-leukemia challenge were sorted. As a control, HV1 CD4 T cells from naïve leukemia-specific TCR transgenic mice were also sorted. Nucleii from the sorted T cells were isolated and captured on the 10x Chromium, and library preparation was generated using the Chromium Next GEM Single Cell ATAC library kit (ver. 1.1), following the manufacturer's recommendations. For the scRNA-seq study, CD45.1+ mice were injected with CD45.2+ HV1 CD4 T cells via the tail vein, followed by a challenge with LM138 leukemia cells via the tail vein the next day. Some of the CD45.1+ mice were treated with the tyrosine kinase inhibitor nilotinib five days/week for one week starting at day 14 post-leukemia-establishment, as well as treatment with an isotype control or IL10R blocking or PDL1 blocking antibody on days 14, 16 and 18 post-leukemia-establishment. A 50:50 mixture of CD45.2+ HV1 CD4 T cells as well as CD44+ CD8 T cells were sorted from the indicated experimental treatment arms. The sorted T cells were labeled with treatment arm-specific hashtag oligonucleotide (HTO) antibodies, as well as CITE-Seq antibody derived oligonucleotide tag (ADT) antibodies to CD4, CD39, CD4b, PD1, TIM3 and LAG3. Cells were captured and single cell sequencing for gene expression (GEX), sample-specific hashtag and protein feature barcodes, and T-cell receptor (TCR) repertoires was performed using the 10X Genomics Chromium Next GEM Single Cell 5' (ver. 1.1) reagents, resulting in three (GEX, HTO/ADT, and TCR) libraries.