Signaling mechanisms underlying self-renewal of leukemic stem cells (LSCs) are poorly understood, and identifying pathways specifically active in LSCs could provide opportunities for therapeutic intervention. T-cell immunoglobin mucin-3 (TIM-3) is expressed on the surface of LSCs in many types of human acute myeloid leukemia (AML), but not on hematopoietic stem cells (HSCs). Here, we show that TIM-3 and its ligand, galectin-9 (Gal-9), constitute an autocrine loop critical for LSC self-renewal and development of human AML. Serum Gal-9 levels were significantly elevated in AML patients and in mice xenografted with primary human AML samples, and neutralization of Gal-9 inhibited xenogeneic reconstitution of human AML. Gal-9-mediated stimulation of TIM-3 co-activated NF-κB and β-catenin signaling, pathways known to promote LSC self-renewal. These changes were further associated with leukemic transformation of a variety of pre-leukemic disorders and together highlight that targeting the TIM-3/Gal-9 autocrine loop could be a useful strategy for treating myeloid leukemias.

Classic Hodgkin lymphoma (cHL) responds markedly to PD-1 blockade therapy, and the clinical responses are reportedly dependent on expression of major histocompatibility complex class II (MHC-II). This dependence is different from other solid tumors, in which the MHC class I (MHC-I)/CD8+ T-cell axis plays a critical role. In this study, we investigated the role of the MHC-II/CD4+ T-cell axis in the antitumor effect of PD-1 blockade on cHL. In cHL, MHC-I expression was frequently lost, but MHC-II expression was maintained. CD4+ T cells highly infiltrated the tumor microenvironment of MHC-II–expressing cHL, regardless of MHC-I expression status. Consequently, CD4+ T-cell, but not CD8+ T-cell, infiltration was a good prognostic factor in cHL, and PD-1 blockade showed antitumor efficacy against MHC-II–expressing cHL associated with CD4+ T-cell infiltration. Murine lymphoma and solid tumor models revealed the critical role of antitumor effects mediated by CD4+ T cells: an anti-PD-1 monoclonal antibody exerted antitumor effects on MHC-I−MHC-II+ tumors but not on MHC-I−MHC-II− tumors, in a cytotoxic CD4+ T-cell–dependent manner. Furthermore, LAG-3, which reportedly binds to MHC-II, was highly expressed by tumor-infiltrating CD4+ T cells in MHC-II–expressing tumors. Therefore, the combination of LAG-3 blockade with PD-1 blockade showed a far stronger antitumor immunity compared with either treatment alone. We propose that PD-1 blockade therapies have antitumor effects on MHC-II–expressing tumors such as cHL that are mediated by cytotoxic CD4+ T cells and that LAG-3 could be a candidate for combination therapy with PD-1 blockade.

Regulatory T (T reg ) cells suppress effective antitumor immunity in tumor-bearing hosts, thereby becoming promising targets in cancer immunotherapy. Despite the importance of T reg cells in tumor immunity, little is known about their differentiation process and epigenetic profiles in the tumor microenvironment (TME). Here, we showed that T reg cells in the TME of human lung cancers harbored a completely different open chromatin profile compared with CD8 + T cells, conventional CD4 + T cells in the TME, and peripheral T reg cells. The integrative sequencing analyses including ATAC, single-cell RNA, and single-cell ATAC sequencing revealed that BATF, IRF4, NF-κB, and NR4A were important transcription factors for T reg cell differentiation in the TME. In particular, BATF was identified as a key regulator, which leveraged T reg cell differentiation through epigenetically controlling activation-associated gene expression, resulting in the robustness of T reg cells in the TME. The single-cell sequencing approaches also revealed that tissue-resident and tumor-infiltrating T reg cells followed a common pathway for differentiation and activation in a BATF-dependent manner heading toward T reg cells with the most differentiated and activated phenotypes in tissues and tumors. BATF deficiency in T reg cells remarkably inhibited tumor growth, and high BATF expression was associated with poor prognosis in lung cancer, kidney cancer, and melanoma. These findings indicate one of the specific chromatin remodeling and differentiation programs of T reg cells in the TME, which can be applied in the development of T reg cell–targeted therapies.

Several immune checkpoint molecules and immune targets in leukemic cells have been investigated. Recent studies have suggested the potential clinical benefits of immuno-oncology (IO) therapy against acute myeloid leukemia (AML), especially targeting CD33, CD123, and CLL-1, as well as immune checkpoint inhibitors (e.g., anti-PD (programmed cell death)-1 and anti-CTLA4 (cytotoxic T-lymphocyte-associated protein 4) antibodies) with or without conventional chemotherapy. Early-phase clinical trials of chimeric antigen receptor (CAR)-T or natural killer (NK) cells for relapsed/refractory AML showed complete remission (CR) or marked reduction of marrow blasts in a few enrolled patients. Bi-/tri-specific antibodies (e.g., bispecific T-cell engager (BiTE) and dual-affinity retargeting (DART)) exhibited 11–67% CR rates with 13–78% risk of cytokine-releasing syndrome (CRS). Conventional chemotherapy in combination with anti-PD-1/anti-CTLA4 antibody for relapsed/refractory AML showed 10–36% CR rates with 7–24 month-long median survival. The current advantages of IO therapy in the field of AML are summarized herein. However, although cancer vaccination should be included in the concept of IO therapy, it is not mentioned in this review because of the paucity of relevant evidence.

The cellular properties of leukemia stem cells (LSCs) are achieved at least through Class I and Class II mutations that generate signals for enhanced proliferation and impaired differentiation, respectively. Here we show that in t(8;21) acute myelogenous leukemia (AML), hematopoietic stem cells (HSCs) transform into LSCs via definitively-ordered acquisition of Class II (AML1/ETO) and then Class I (c-KIT mutant) abnormalities. Six t(8;21) AML patients with c-KIT mutants maintaining > 3 years of complete remission were analyzed. At diagnosis, all single LSCs had both AML1/ETO and c-KIT mutations. However, in remission, 16 out of 1,728 CD34(+)CD38(-) HSCs and 89 out of 7,187 single HSC-derived myeloerythroid colonies from these patients had AML1/ETO, whose breakpoints were identical to those found in LSCs. These cells had wild-type c-KIT, which expressed AML1/ETO at a low level, and could differentiate into mature blood cells, suggesting that they may be the persistent preleukemic stem cells. Microarray analysis suggested that mutated c-KIT signaling provides LSCs with enhanced survival and proliferation. Thus, in t(8;21) AML, the acquisition of AML1/ETO is not sufficient, and the subsequent upregulation of AML1/ETO and the additional c-KIT mutant signaling are critical steps for transformation into LSCs.