Multiparameter Single Cell Analysis to Identify Immune Cell States That May Contribute to Successful Treatment Discontinuation in Patients with Chronic Myeloid Leukemia

Background: Recently, numerous clinical trials have demonstrated that approximately half of CML patients who have achieved deep and durable remissions on TKI therapy can enjoy prolonged treatment-free remissions. A major question in the field concerns why treatment interruption is not universally successful. Two non-mutually exclusive theories have been proposed to explain successful discontinuation: CML stem cell exhaustion, and immune surveillance. Identification of mediators of CML stem cell exhaustion or successful immune surveillance could lead to adjunctive therapeutic approaches. We set out to determine potential immune mediators of disease control through interrogation of clinical samples using emerging technologies.

Results: Consenting patients who discontinued TKI at our institution were grouped into three categories based upon BCR-ABL1 transcript kinetics following treatment discontinuation. Group 1 is characterized by stable undetectable or minimally detectable BCR-ABL1 transcripts. Group 2 exhibited molecular relapse, defined as loss of major molecular response (MMR) requiring treatment re-initiation, typically within the first 3-9 months of treatment discontinuation. Group 3 displays kinetics that are particularly intriguing, whereby the transcript level rises substantially upon treatment cessation, but remains below the MMR threshold despite remaining off TKI therapy. We hypothesize that disease control in group 3, and a subset of patients in Group 1, is mediated by the immune system.

To investigate the immune system difference between these groups, we performed single cell immune profiling in the pre- (~ 1 week) and post- (~ 3 months) discontinuation PBMC samples obtained from 15 patients through CyTOF and 10x single cell RNA sequencing. Group 1 consist of 5 patients, group 2 consist of 8 patients, and group 3 consist of 2 patients, respectively. There were no meaningful differences in the immune cell composition prior to TKI discontinuation. Assessment of the immune phenotype 3 months after discontinuation in each group by CyTOF revealed an increase in the CD4⁺ T cell population in group 3 following TKI discontinuation. Specifically, memory CD4⁺ T cells and CD38⁺/HLA-DR⁺ activated CD4⁺ T cells were increased in group 3. While patients in groups 1 and 2 showed only decrease or moderate change in memory CD4⁺ T cells and activated CD4⁺ T cells, group 3 showed 53% increase of memory CD4⁺ T cells (21% to 32%) and 640% increase of activated CD4⁺ T cells (0.4% to 3.3%) following TKI discontinuation. While CD45RA^-/CD4⁺ T cells were increased in abundance, the CD45RA^- population was unchanged in CD8⁺ T cells. 10x single cell RNA sequencing of these samples is ongoing and will be presented.

Conclusions: We have conducted multiparameter single cell immune profiling in CML patients with sustained deep molecular remission, loss of MMR, or maintenance of MMR despite a sustained rise in BCR-ABL1 transcript level. Our results suggest that an increase in memory and activated CD4+ T cells following TKI discontinuation may correlate with successful immune surveillance of CML. Assessment of a larger cohort of samples is required to confirm our findings.