Background: Intrinsic immune responses to acute myeloid leukemia (AML) are inhibited by a variety of mechanisms, such as T cell exhaustion, expansion of immunoregulatory cells and activation of inhibitory pathways. An in-depth understanding of the immune evasion mechanisms in patients with AML at diagnosis and following chemotherapy is essential for developing strategies to augment anti-leukemia immunity to prevent disease relapse.

Methods: We analyzed lymphoid cell dynamics in peripheral blood (PB) and bone marrow (BM) of 28 AML patients (median age 54, range 29-75 years) at multiple time points before and after chemotherapy. Using multi-parameter flow cytometry, we performed an extensive phenotypic characterization of lymphocyte populations, focusing on T cells and NK cells (CD3-CD56+). Within T cells, we characterized the differentiation status (CD45RA, CCR7, CD95), activation/proliferation status (HLA-DR, Ki-67), and the expression of co-stimulatory (CD27, CD28) and co-inhibitory receptors (PD1, CTLA4, TIM3, CD160, 2B4, BTLA, KLRG1). As a control, we used PB (n=28) and BM (n=10) lymphocytes from healthy controls (HC; median age 40, range 25-71 years). The percentage of cells expressing specific markers over time were log-transformed and modeled as outcomes in linear regression models that included fixed effects for time and a random effect for the patient to account for within-patient correlation of measurements. Differential outcomes over time by clinical response were tested with interaction terms.

Results:At diagnosis, AML patients showed significantly lower median frequency of CD8+ naïve T cells (CD45RA+CCR7+) but increased frequency of terminal differentiated CD8+ effector memory T cells (TEMRA; CD45RA+CCR7-) compared to the PB of HC. Expression of inhibitory molecules PD1 and 2B4 was significantly increased on both CD4+ and CD8+ PB T cells compared to HC. CTLA4 (p<0.0001) and KLRG1 (p<0.01) were only increased on CD8+ but not CD4+ T cells of patients. No differences in the expression of TIM3 and BTLA4 were found. Treg (CD4+CD25+CD127-) percentages were similar in the PB of AML patients and HC, but were significantly increased in the BM of patients, where the percentage of effector Tregs (eTregs, CD4+CD45RA-FOXP3high) was particularly elevated. NK cell frequency was lower in the PB (p<0.0001), but not in the BM of AML patients.

Early lymphocyte recovery (ELR, absolute lymphocyte count>200/mm3 after induction chemotherapy) was characterized by increased proliferation (Ki67+) of PB CD4+ and CD8+ T cells as well as increased frequency of PB CD4+ and CD8+ effector memory (EM, CD45RA-CCR7-) and stem cell memory (CD45RA+CCR7+CD95+) T cells. As previously reported (Kanakry et al. Blood 2011), ELR was marked by an increased percentage of PB Tregs, particularly the eTreg subset which also had increased expression of HLA-DR, suggesting heightened functionality in addition to frequency. Both PB and BM Tregs persisted at elevated levels throughout all time points.

When analyzed according to response to induction chemotherapy, only non-responding (NR) patients had increased frequency of PB CD8+TEMRAs at diagnosis compared to HC. No differences in the percentages of PB or BM Tregs were observed between responders and NR patients at diagnosis or at recovery from chemotherapy. However, the percentage of eTregs expressing CTLA-4 remained higher over time in NR patients than in patients achieving remission. An increased frequency of CD8+CD160+ T cells in PB at diagnosis was observed only in NR patients. At recovery from chemotherapy, significantly decreased frequencies of 2B4, BTLA, and CD160 expressing CD4+ and CD8+ PB T cells compared to pretreatment levels were noted, particularly in responders.

Conclusion: Our study provides a detailed immunophenotypic overview of the dynamics in lymphoid populations in patients with AML at diagnosis and during and after chemotherapy-induced lymphopenia. Notably, we have identified several immunoregulatory mechanisms at play in these patients, including alterations in the T cell differentiation state, co-inhibitory molecule expression, and Treg levels and activation states. Ongoing studies are focused on deciphering the function of these defined lymphoid populations. These data have direct implications for the design of strategies for reversing T cell dysfunction in AML.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.