T-cell immunoglobulin and mucin domain-containing molecule 3 (TIM-3) is a membrane protein which was discovered initially on CD4⁺ T helper 1 (Th1) and CD8⁺ T cytotoxic 1 cells (1), and later on a variety of other cell types including regulatory T (Treg) cells (2), dendritic cells (3), natural killer (NK) cells (4), myeloid cells (5) and mast cells (6). TIM-3 is generally considered as a negative regulator of immune system (7), and has been investigated as a blockade target for immunotherapy of cancer in pre-clinical and clinical settings (8–10).

Interestingly, TIM-3 also expresses on leukemic stem and progenitor populations of non-M3 acute myeloid leukemia (AML) and associated with leukemic transformation of preleukemic diseases, such as myelodysplastic syndrome (MDS) and myeloproliferative neoplasm (11). TIM-3 and its ligand, galectin-9 (Gal-9), constitute an autocrine loop which drives the self-renewal of AML stem cells by activating the nuclear factor-κB (NF-κB) and β-catenin pathways (11). Besides, both TIM-3 and Gal-9 can be released in a free soluble form and involved in the immune escape of AML cells (12).

TIM-3 expression on immune cells has been reported to be associated with disease activity and prognosis in multiple types of cancers, e.g. AML (13), pediatric B-precursor acute lymphoid leukemia (14), lung cancer (2, 4). In general, higher TIM-3 expression was related to poorer prognosis. Additionally, the relation between TIM-3 expression on AML blasts and prognosis of AML was also investigated and contradictory data were published in the literature. Darwish et al. and Kamal et al. both reported TIM-3 as a biomarker of poor prognosis in AML, while Xu et al. reported that increased TIM-3 expression correlated with low-risk group and higher complete remission rate in newly diagnosed non-M3 AML patients (15–17).

According to the data mentioned above, TIM-3 expression on T cells and AML blasts are both related to prognosis of AML, and there might be an association between TIM-3 expression on T cells and AML blasts. To our knowledge, this issue has not been well investigated and published in the literature. In this study, we assessed simultaneously the TIM-3 expression on T cells and AML blasts in patients with non-M3 de novo AML using flow cytometry, and investigated its correlation with clinical outcomes, as well as other clinical parameters, including French-American-British (FAB) classifications, genetic abnormalities and risk stratifications. Furthermore, the Cancer Genome Atlas (TCGA) data for AML were obtained and analyzed to validate the impact of TIM-3 expression on prognosis of non-M3 AML patients.

In the present study, we simultaneously assessed the TIM-3 expression level of leukemic blasts and T lymphocytes in the bone marrow of de novo AML patients using flow cytometry, and found that the TIM-3 expression level of leukemic blasts correlated positively with T lymphocytes. Besides, TIM-3 expression of leukemic blasts correlated positively with the percentage of peripheral CD8⁺ T lymphocytes. When analyzed with clinical data, higher TIM-3 expression of leukemic blasts was shown to significantly correlate with the presence of CBF translocations but not with the survival of patients, and similar results were also obtained with TCGA data.

TIM-3 is a member of the TIM family of immunoregulatory proteins. It was originally discovered in T cells and has been considered as an important checkpoint receptor. So far, four ligands, which are Gal-9, phosphatidylserine, high-mobility group protein B1 and carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM1), have been identified to bind TIM-3. Among them, Gal-9 and CEACAM1 are involved in TIM-3 mediating the inhibition and apoptosis of T cells (7). TIM-3 is a glycoprotein containing both N-linked and O-linked glycans, and glycosylation is required for Gal-9 binding (21, 22).

Besides T cells, TIM-3 also expresses on AML cells, rather than normal hematopoietic stem cells, and is involved in the self-renewal of LSCs. Autocrine Gal-9 binds to TIM-3 and subsequently activates NF-κB and β-catenin signaling, promoting LSC self-renewal (11). In addition, TIM-3 is currently thought to play a crucial role in immune escapes of AML cells. TIM-3, together with its ligand Gal-9, can be released from AML cell surface in a free soluble form through proteolysis, and impair the killing activity of NK cells and IL-2 production of T cells (12). TIM-3 also have a higher expression, together with other checkpoint receptors, like PD-1, on T cells of AML patients as compared to healthy adults (23), and this is believed to be associated with severe T cell exhaustion and disease progression (10).

Our data showed that TIM-3 expression levels on leukemic blasts varied among AML patients, which were in accordance with those of Kikushige et al. and Xu et al. (15, 24) Due to the important role of TIM-3 in AML, it would be intriguing to determine whether varied expression levels of TIM-3 on AML cells influence the immune status and prognosis of AML patients.

Our results revealed a correlation of TIM-3 expression of leukemic blasts with TIM-3 expression of both CD8⁺ and CD4⁺ T lymphocytes and the proportion of CD8⁺ T lymphocytes, indicating that TIM-3 expression of leukemic blasts might alter adaptive immunity of AML patients. TIM-3 on T lymphocytes is generally considered as a marker of T cell exhaustion, especially when coexpressed with other checkpoint receptors (25). For example, TIM-3⁺PD-1⁺CD8⁺ T cells in mice with advanced AML exhibited severe exhausted phenotype as defined by failure to produce IL-2, tumor necrosis factor and interferon-γ, and combined targeting of TIM-3 and PD-1 pathways was more effective in controlling the leukemia burden (10). TIM-3 was also reported to coexpress with two other checkpoint receptors LAG-3 and PD-1 on severe exhausted tumor-infiltrating lymphocytes of patients with glioblastoma (26). Therefore, we consider that our data support the notion that TIM-3 onleukemic blasts can promote an inhibitory immune environment and facilitate the immune escape of AML cells (27). In the future, it would be worthwhile to assess, together with the TIM-3 expression, the expression of other checkpoint receptors and perform the function assays of lymphocytes in AML patients, in order to further determine the effect of TIM-3 expression of leukemic blasts on adaptive immunity.

The relation of TIM-3 with AML prognosis is currently controversial. In our study, the TIM-3 expression level of leukemic blasts was neither correlated with the CR after the first induction chemotherapy nor the survival of AML patients. This data are contradictory to those of Darwish et al., Kamal et al. and Xu et al. In studies of Darwish et al. and Kamal et al., all AML subtypes including M3 were enrolled. M3 patients usually have good prognosis and hardly express TIM-3 on their leukemic blasts. This might partly explain different conclusions they got from those of our study. In the study of Xu et al., AML patients with high TIM-3 expression level achieved a higher CR rate than patients with low TIM-3 expression level (91% versus 67%, p=0.01), while the survival data were not reported. Our data, which include both treatment response after the induction chemotherapy and the survival of patients, are supposed to be more reliable. Moreover, TCGA data of 157 non-M3 AML patients were subsequently analyzed and showed that TIM-3 expression was not associated with the OS of patients, either in the whole cohort or in any ELN risk group, consistent with results of our cohort. Interestingly, Wang et al. recently reported that high TIM-3 expression levels of T and NK cells as a whole and CD34⁺CD38^- cells were significantly associated with a high 2-year cumulative incidence of relapse in t(8;21) AML patients (28). Unfortunately, it is impossible to confirm this finding with our data, as only few patients in our cohort have t(8;21). Further studies are needed to verify this finding and explore the underlying molecular mechanisms.

It should be noted that we assessed TIM-3 expression level on the surface of leukemic blasts in our cohort, while in TCGA data, TIM-3 mRNA expression level in the peripheral blood samples was used for analysis. In the peripheral blood of AML patients, TIM-3 expresses not only in leukemic blasts, but also in multiple other cell types, e.g. T lymphocytes, NK cells, monocytes, etc. However, leukemic blasts are the main cell population (average percentage of 39.6% in non-M3 AML patients of TCGA data), and our data also showed that TIM-3 expression of T cells, another big cell population, correlated positively with that of leukemic blasts, therefore TIM-3 mRNA expression in the peripheral blood is supposed to represent that of leukemic blasts.

CBF translocations include t(8;21)(q22;q22) and inv (16)(p13q22)/t(16;16)(p13;q22), which lead to the formation of fusion genes AML1/ETO and CBFβ/MYH11, respectively. These fusion genes disrupt the signaling of heterodimeric CBF complex in a dominant negative manner, resulting in impaired hematopoietic differentiation (29). AML with CBF translocations accounts for approximately 15% of adult AML cases and is stratified as favorable risk (30). Our results showed that CBF translocations were associated with higher TIM-3 expression on the surface of leukemic blasts. Similar results have been reported by Jan et al. and Xu et al., and it is thought that mutations in CBF may either directly regulate TIM-3 transcription or arrest leukemic cells in a stage of differentiation with high TIM-3 expression (15, 31). Silva et al. reported that latrophilin1/protein kinase C/mammalian target of rapamycin pathway was involved in the expression of TIM-3 and its ligand Gal-9 in AML blasts, but all their experiments were performed in the cell line of THP-1, which carries KMT2A-MLLT3 fusion gene rather than CBF translocations (12). Further studies are needed to elucidate the molecular mechanism by which CBF translocations control TIM-3 expression.

Since AML with CBF translocations are considered as favorable risk, it seems that higher TIM-3 expression should be associated with good prognosis, which is not the case in our study. When we analyzed the data carefully, we found that other low-risk genetic alterations, e.g. NPM1 mutation with wild-type FLT3-ITD, usually had low TIM-3 expression level, and some intermediate-risk AML patients also had high TIM-3 expression level. Eventually, TIM-3 expression level was not related with the clinical outcome of AML patients.

Currently, TIM-3 is considered as a potential target for the treatment of myeloid malignancies. Since the autocrine Gal-9 binding to TIM-3 on the surface of leukemic blasts drives the self-renewal of AML stem cells (11), it is a potential therapeutic strategy to block TIM-3 using anti-TIM-3 mAbs. Several monoclonal antibodies (mAbs) against TIM-3, e.g. MBG453 (NCT03066648) and SHR-1702 (NCT04443751), are investigated in ongoing clinical trials and their clinical efficacy is unknown (32, 33). According to our study, it seems that TIM-3 expression on leukemic blasts does not influence the outcome of AML patients, or traditional chemotherapy can overcome the adverse impact of TIM-3 on outcomes of AML patients. Therefore, we should be cautious to expect that anti-TIM-3 mAbs further improve the clinical outcome of AML patients. However, anti-TIM-3 mAbs bound to leukemic blasts may facilitate antibody-dependent cellular phagocytosis by myeloid cells/macrophages, and they may also block TIM-3 on lymphocytes and other immune cells and potentially enhance the immune response against leukemic blasts (32). Therefore, it remains intriguing to see whether anti-TIM-3 mAbs have a therapeutic effect on myeloid malignancies.

Our study has limitations. First, the cohort size is relatively small, and further studies with larger sample size are still required to confirm these findings. Second, immunophenotyping of multiple checkpoint receptors, other than TIM-3, and function assays of lymphocytes were not performed in our study. These experiments are needed to assess adaptive immune status. Third, flow cytometry can only detect the protein of TIM-3 on the cell surface. It would be better to simultaneously assess HAVCR2 mRNA level using RT-PCR to confirm the results. Forth, western blot should be performed to assess the glycosylation status of TIM-3 in these AML samples, as glycosylation may influence the function of TIM-3, like the binding with Gal-9 (21).

In conclusion, TIM-3 expression of AML blasts correlated positively with TIM-3 expression of T lymphocytes and the proportion of CD8⁺ T lymphocyte, indicating TIM-3 might alter adaptive immunity of AML patients. In addition, TIM-3 expression of AML blasts correlated with CBF translocations rather than the survival of patients. Therefore, TIM-3 might not be a good biomarker for non-M3 AML prognosis under current treatment modalities.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by The Ethical Committee of the First Affiliated Hospital of Anhui Medical University. The ethics committee waived the requirement of written informed consent for participation.

MY and YW conceived and designed the study. JH and LX performed the experiments, analyzed the data and wrote the manuscript. JH and XY performed TCGA data collection and analysis. JH, LX, ZH, XL, JD, ZW, LL, MR, ZL, XC, XWC, JN, JG, QL, QZ and RX performed clinical sample and data collection. All authors contributed to the article and approved the final version.

The study was supported by Research Fund of Anhui Institute of Translational Medicine (Grant No. 2021zhyx-C70), Clinical Medicine Discipline Construction Project of Anhui Medical University (Grant No. 2020lcxk034) and Young Scholar Program of The First Affiliated Hospital of Anhui Medical University (Grant No. 2019kj20).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.