Although γδ T cells represent a relatively small subset within all T lymphocytes (1–5%) (1, 2), they have a unique property to recognize cancer cells in a T cell receptor- (TCR) dependent manner but also in a non- major histocompatibility complex- (MHC) restricted way via their expression of natural killer cell receptors (NKRs) (3). In immuno-oncology, γδ T cells represent a novel attractive effector population for immuno-therapeutic strategies such as chimeric antigen receptor T cells (CAR-T cells) or bispecific T cell engagers (BiTEs).

In humans, γδ T cells can be differentiated into two major subsets by their expression of the Vδ chain. Vδ2 cells constitute the major circulating γδ T cell population in the peripheral blood (PB) whereas the Vδ1 subpopulation is enriched in the peripheral tissue (4, 5). Both γδ subpopulations exhibit cytotoxic capacities mediated by TCR- and natural killer group 2D (NKG2D) receptor signaling via production of the effector cytokines interferon (IFN)-γ, tumor necrosis factor (TNF)-α and soluble mediators such as perforin or granzymes (5–7). Moreover, γδ T cells can induce dendritic cell (DC) maturation via secretion of TNF-α (8, 9). They also have an antigen-presenting capacity by MHC-II loading and expression (10). Recent data also demonstrate a phagocytic potential of γδ T cells through expression of the scavenger receptor CD36 which is dependent on the transcription factor CCAAT-enhancer-binding protein α (C/EBPα) (11).

In oncology, both the Vδ1 and Vδ2 T cells have been described to exert pleiotropic effector functions: as mentioned above the tumor-infiltrating IFNγ-producing γδ T cell fraction has shown cytotoxic effects in solid and hematological malignancies (7, 12, 13), whereas interleukin (IL)-17- and galectin 1-secreting γδ T cells promote tumor growth and the recruitment of immunosuppressive myeloid cells (14, 15). By analyzing molecular profiles of expression signatures of 18.000 tumors from 39 different malignancies, including AML and MM with overall survival data, Gentles et al. identified tumor-infiltrating γδ T cells as the leukocyte subset with the most significant favorable cancer-wide prognostic relevance (16).

First clinical observations of patients with refractory/relapsed AML transplanted with haploidentical γδ T cells showed that γδ T cells can induce clinical anti-tumor effects (17, 18). Furthermore, preclinical data provide support for this contention: AML cells were efficiently killed by Vδ2 T cells in vitro (3, 6, 19). In an AML xenograft model, Vδ2 T cells traffic to the BM and have been shown to slow the progression of the disease (6). Also, in MM it has been recently published that impairment of Vδ2 T cell functions (including decreased proliferation and cytotoxicity) was already detectable in monoclonal gammopathy of undetermined significance (MGUS) (20).

Compared to Vδ2 T cells, Vδ1 T cells account for a small proportion of γδ T cells (1). Although most of the studies investigating the antitumoral effects of γδ T cells focus on Vδ2 T cells, it is becoming increasingly evident that Vδ1 T cells play a critical role in the anti-tumor functionality (21, 22). Enhanced reconstitution of γδ T cells following allogeneic hematopoietic stem cell transplantation (aHSCT) was associated with improved survival in patients with AML (23). Cordova et al. demonstrated that Vδ1 tumor-infiltrating lymphocyte- (TIL-) derived cells outperformed Vδ2 TILs in in vitro tumor cytotoxicity assays in malignant melanoma (24). In AML, increased levels of cytomegalovirus- (CMV) specific Vδ1 cells were associated with a reduced relapse probability. In vitro, these Vδ1 cells demonstrated increased cytotoxicity against AML cells, which could be further enhanced following CMV reactivation (19, 25). Additionally, Knight et al. showed a significant cytotoxicity of Vδ1 T cells for MM by lysis against patients' CD38⁺CD138⁺ BM-derived plasma cells in vitro (26).

First phenotypic analyses of BM-derived Vδ2 T cells from AML patients showed an increased subpopulation with a memory phenotype that was associated with a reduced capacity for expansion and cytotoxicity (6).

This study is focused on programmed cell death protein-1 (PD-1), the novel receptor T cell Ig and ITIM domain (TIGIT), the T cell immunoglobulin and mucin domain-containing protein 3 (TIM-3) and the ectonucleotidases ectonucleoside triphosphate diphosphohydrolase-1 (CD39) and ecto-5'-nucleotidase (CD73). All of these checkpoints are involved in αβ T lymphocyte dysfunction in cancer (27–29). TIGIT, PD-1 and TIM-3 are co-inhibitory receptors highly expressed by exhausted T cells in chronic infections and cancer (27, 30, 31). Inhibition of these receptors have shown increased proliferation and cytotoxic efficiency in vitro (27, 30–32). Recently, inhibition of the adenosine-generating enzymes CD39 and CD73 showed anti-tumor immunity through multiple mechanisms, including enhancement of T cell and natural killer- (NK) cell function (33, 34). It is well-acknowledged that multiple co-expression of inhibitory receptors and suppression of inflammatory cues by CD39 enzymatic overactivity are highly associated with the severity of T cell dysfunction (35, 36), thus representing an important feature of T cell exhaustion (37). Regarding the (co)-expression of these checkpoints on γδ T cells, little is known. Jin et al. showed for the first time that TIGIT is also expressed by regulatory Forkhead-Box-Protein P3 (FOXP3)⁺ γδ T cells in AML (38). Moreover, it has been presented that tumor-infiltrating γδ T cells express the ectonucleotidases CD39 and CD73 and may be dysfunctional via the activation of adenosine-mediated pathways (39, 40).

Since circulating γδ T cells in the peripheral blood have been more elucidated, the present study provides an extensive immunophenotypic characterization of γδ T cells derived from the bone marrow of patients with untreated newly diagnosed acute myeloid leukemia or multiple myeloma. We focus on the comparison of the Vδ1- and Vδ2-cell phenotypes, including differentiation and expression of immune checkpoints and metabolic molecules that may dampen antitumor immunity in patients with acute myeloid leukemia in comparison to patients with myeloma and healthy volunteers.

The present study provides a phenotypic analysis of the BM-resident γδ population in patients with AML and MM. Here, we demonstrate that Vδ1 T cells represent the predominant γδ T-cell population in the BM from these patients whereas Vδ2 T cells were dominant in the PB of HDs. The BM-infiltrating Vδ1 T cells in AML and MM showed an increased TEMRA cell compartment with an aberrant subpopulation of CD27⁻CD45RA⁺⁺ cells in comparison to HDs. Expression analyses of corresponding immune effector cells derived from the BM revealed an increased expression of TIGIT, PD-1, TIM-3, and CD39 on γδ TCR cells in comparison to CD4⁺ cells, which was similar to that on CD8⁺ effector cells in both hematologic malignancies. In contrast to the Vδ2 T cells, the increased frequency of PD-1, TIGIT, TIM-3, and CD39 positive cells was mainly observed on Vδ1 T cells. This upregulated expression of co-regulatory receptors in AML and MM was related to the TEMRA γδ subpopulation and within this population, highest expression was found on the CD27⁻CD45RA⁺⁺ cells. Furthermore, γδ T cells in AML and MM exhibited a further feature of exhaustion, manifested by an increased co-expression of multiple co-inhibitory molecules including PD-1, TIM-3, and CD39 together with TIGIT, which was caused by the increased co-expression on the Vδ1 T-cell population. In contrast, CD73 expression was downregulated by the TIGIT⁺ γδ T cells in AML and MM. Regarding the co-expression of multiple co-inhibitory receptors, we observed similar frequencies of double positive (TIGIT⁺PD-1⁺ / TIGIT⁺TIM-3⁺ or TIGIT⁺CD39⁺) cells in both cancer entities. In contrast, the single expression of TIGIT and HLA-DR was higher in MM, whereas TIM-3 was more frequently expressed by γδ T cells derived from patients with AML. These differences between both entities might be explained by variations in the immune response upon chronic stimulation.

γδ T cells are endowed with two independent recognition systems including the γδ TCR and NK cell receptors to identify tumor cells and initiate anti-cancer effector mechanisms, including cytokine production and cytotoxicity. Therefore, γδ T cells represent ideal effectors with a better safety profile and a more favorable anti-tumor efficacy in comparison to conventional αβ T cells because of their HLA-independent recognition of phosphoantigens that are characteristic for dysregulated metabolism in tumors (43, 44). Furthermore, the HLA- independent recognition system results in reduced graft vs. host disease and diminished target effects (44–47). Provided with these special features, γδ T cells represent attractive effector cells for adoptive T cell strategies. A recently designed γδ T cell construct with a GD2-targeted CAR showed potent responses against Disialoganglioside (GD2)⁺ neuroblastoma (48, 49). Moreover, first clinical studies have been conducted to investigate the safety and efficacy of adoptive transfer of autologous or allogeneic γδ T cells in cancer patients including renal cell carcinoma, lung cancer, hepatocellular carcinoma, breast cancer, prostate cancer, and multiple myeloma (50). These studies confirmed that γδ T cell application is safe, with low levels of adverse events and that the clinical responses are ranging from partial to complete remissions (50). In AML, promising data have been recently presented by Ganesan et al. demonstrating increased in vitro and in vivo γδ T cell-mediated cytotoxicity by a bispecific engager molecule against T cell receptor gamma variable 9 (TRGV9) and CD123 (51). However, there is still potential for improvement including the engagement of co-regulatory checkpoint molecules.

In agreement with other studies, we found that in BM from AML and MM patients, the infiltration of tissue-resident Vδ1 T cells was increased, whereas the frequency of the circulating Vδ2 T cells was lower compared to that in the PB from HDs (3, 52). This inversion of the Vδ1/Vδ2 ratio in different tissues has been recently observed in solid cancer including melanoma, colorectal cancer, and non-small cell lung cancer as well (52). Although a clear correlation with γδ T cell infiltration and prognosis of cancer patients is still missing, some studies demonstrated that a reduction of Vδ2 T cells correlates with an advanced stage. In line with this, enhanced levels of Vδ2 T cells correlated with an early stage of melanoma, the absence of metastasis and a longer 5-year disease-free survival rate of colorectal cancer patients (52).

To exclude that the differences observed between HD-derived PB and patient-derived BM could be compartment-associated, we additionally compared paired BM and PB aspirates from nine AML patients on a random basis. We observed no difference in the frequency of the total γδ T cells, nor in the distribution of Vδ1 and Vδ2 T cells. Moreover, no variations of checkpoint expression on γδ T cells and subpopulations respectively, were observed in these nine AML patients. Similar with our findings, Rossol et al. compared Vδ1 T cells derived from the BM and PB from patients with HIV and observed no significant differences (53). Dean et al.'s studies confirmed these observations, revealing no significant differences of γδ T cells when they compared aspirates from the BM and PB derived from patients with an active hematopoietic malignancy or in remission (54). Our study has the limitation of a relatively small sample size of patients with newly diagnosed AML and MM. Therefore, the observations made in this study should be interpreted with caution until validated in a larger cohort.

Our analyses of the differentiation status revealed a shift toward increased TEMRA differentiation. Migration of TEMRAs into inflammatory sites to perform effector functions has also been observed for chronic infections such as CMV, but also for solid cancer including neuroblastoma, colorectal cancer, and melanoma (5). For AML patients, Gertner-Dardenne et al. first described an increased memory profile in γδ T cells derived from the PB and BM, but in contrast to our study, they observed a differentiation into CD27⁻CD45RA⁻ defined EMs (6). To our knowledge, our study is the first description of the differentiation status in the BM for MM. We also found a unique γδ T cell subpopulation in the BM niche from patients with AML and MM that is characterized by the CD27⁻CD45RA⁺⁺ phenotype. This aberrant subpopulation discovered by Odaira et al. is reported as a predominant subpopulation in different types of cancer and is characterized as “exhausted” subgroup defined by diminished proliferation capacity (41).

We and others have previously described an increased expression of TIGIT, PD-1, TIM-3, and CD39 on αβ T cells in AML and MM (31, 32, 55–57). Expression of these markers on αβ T cells was related to features of exhaustion manifested by transcriptional reprogramming, reduced effector cytokine production, decreased proliferation, and impaired lysis of tumor cells (32, 55, 58). Our comparative analyses of the co-regulatory marker expression on the corresponding CD4⁺, CD8⁺, and γδ T cells discovered that γδ T cells in the BM from patients with AML and MM exhibit a similar expression profile of TIGIT, PD-1, TIM-3, and CD39 to that on CD8⁺ T cells. In MM, γδ T cells showed even higher frequencies of TIGIT⁺ cells than the CD8⁺ population. Moreover, this study illustrated the increased co-expression of PD-1 and TIM-3 together with TIGIT on γδ T cells in AML and MM for the first time, hypothesizing that these cells are functionally “exhausted.”

Although the knowledge of checkpoint expression on γδ T cells is still sparse, our observations of an increased frequency of TIGIT⁺ γδ T cells confirm the data of increased levels of TIGIT⁺ γδ T cells in de novo AML patients published by Jin et al. In their study, upregulation of TIGIT on γδ T cells was associated with a lower overall survival rate for non-M3 AML (38). To our knowledge, there is no publication of the TIGIT expression on γδ T cells in MM.

PD-1 expression on human healthy γδ T cells has been reported upon antigen-stimulation (59). Additionally, the PD-L1/PD-1 signaling in γδ T cells prevented αβ T cell activation via checkpoint receptor ligation in pancreatic adenocarcinoma (60). In AML, pembrolizumab treatment in combination with zoledronate and interleukin-2 (IL-2) leads to increased IFN-γ production in γδ T cells in vitro (61). Also, in MM PD-1 expression increased in γδ T cells after zoledronate stimulation in contrast to HD-derived γδ T cells, suggesting that MM-derived γδ T cells are intrinsically programmed to increase their threshold of refractoriness via PD-1 upregulation. Interestingly, this upregulation was observed especially in the central memory subset of the Vδ2 T cells, which in normal conditions is the subset with the highest proliferative capacity (20).

Similar to TIGIT and PD-1, receptors, TIM-3 expression and function on conventional αβ T cells has been profoundly studied, but characterization on γδ T cells is very limited. Schofield et al. reported an upregulation of TIM-3 on γδ T cells in childhood malaria. This was also mainly observed in the TEMRA population and was regulated by IL-18 and IL-12 (62). In colorectal cancer, the fraction of TIM-3⁺ γδ T cells has been reported to be increased, which correlated with tumor-lymph-node-metastasis (TNM) staging. Moreover, TIM-3 signaling significantly inhibited the killing efficiency of Vδ2 T cells against colon cancer cells and reduced the secretion of perforin and granzyme B (63).

CD39⁺ expression has been reported on tissue-resident γδ T cells (64). Moreover, CD39⁺ γδ T cells suppressed immune responses via the adenosine pathway by recruitment of myeloid-derived suppressor cells in colorectal cancer (39). These CD39⁺ γδ T cells also expressed FOXP3⁺, a marker of regulatory αβ T cells but have more potent immunosuppressive activity than CD4⁺ or CD8⁺ regulatory T cells (Tregs) (39). Moreover, Casetti et al. reported that FOXP3⁺ γδ T cells have the potential to inhibit the proliferation of anti-CD3/anti-CD28 stimulated PBMCs in vitro (65).

Despite Vδ1 and Vδ2 T cells both having cytotoxic capability, it has been shown that these two subsets express distinct chemokine receptors and cell adhesion molecules (1, 12, 15, 66, 67). Our data revealed that increased expression of TIGIT, PD-1, TIM-3, and CD39 was mainly confined to the Vδ1 T cell subpopulation in the BM from AML and MM patients. Recent studies found that the less studied Vδ1 T cells outperform Vδ2 T cells in most in vitro and first in vivo cancer models in terms of cytotoxicity and cell persistence after allogeneic cell transfer (68). Moreover, because of their different chemokine-receptor profile, Vδ1 T cells exhibit an increased ability of tissue penetration in comparison to Vδ2 T cells (69). Those results underline the functional relevance of the observed upregulation of multiple co-regulatory receptors on these cytotoxic effector Vδ1 T cells in AML and MM in this study.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by Declaration of Helsinki and approval by the local ethics board of the Ärztekammer Hamburg (PV3469 and PV5119). The patients/participants provided their written informed consent to participate in this study.

FB designed the research study, analyzed the data, and wrote the manuscript. PW performed the experiments, analyzed the data, and reviewed the manuscript. KW, LL, and BF provided the myeloma aspirates and reviewed the manuscript. CB and JSzW reviewed the manuscript. JW and WF conceived the concept, oversaw the interpretation and presentation of the data, and reviewed the manuscript. All authors read and approved the final manuscript.

FB was supported by the Mildred Scheel Nachwuchszentrum Hamburg and the Hamburg Translational Research in Cancer program at the University Medical Center Hamburg-Eppendorf, Hamburg, Germany. JSzW was funded by the SFB1328 and by the DFG SFB841. The project was funded by the Roggenbuck Stiftung.

FB: Travel grant Daiichi Sankyo, Servier, Novartis; advisory board by Jazz. GmbH, Daiichi Sankyo. WF: Membership on an entity's board of directors or advisory Amgen, ARIAD/Incucyte, Pfizer, Novartis, Jazz Pharmaceuticals, Morphosys, Abbvie, Celgene; patents and royalities: Amgen; other support for meeting attendance Amgen, Gilead, Jazz Pharmaceuticals, Servier, Daiichi Sankyo; research funding Amgen, Pfizer. Travel grant, advisory board and research funding by Amgen Inc., travel grant and advisory board by TEVA GmbH, the advisory board: Ariad/Incucyte Inc., travel grant by Gilead Inc and Jazz. GmbH, research funding by Pfizer Inc. KW: Honoraria: AbbVie, Amgen, Adaptive Biotech, Celgene/BMS, GSK, Janssen, Karyopharm, Novartis, Roche, Takeda, Sanofi; Research funding: Amgen, Celgene, Janssen, Sanofi. LL: Non-financial support from GSK and from Abbvie. CB: Travel grant: Astra Zeneca, Bayer Healthcare, Berlin Chemie, Bristol Myers Squipp, Jansen Cilag, Merck Serono, Merck Sharp Dohme, Novartis, Roche Pharma, Sanofi Aventis; Advisory board: Astra Zeneca, Bayer Healthcare, Berlin Chemie, Bristol Myers Squipp, Jansen Cilag, Merck Serono, Merck Sharp Dohme, Novartis, Roche Pharma, Sanofi Aventis; Invited speaker: AOK Germany, med update, Merck Serono; Honoraria: AOK Germany, Astra Zeneca, Bayer Healthcare, Berlin Chemie, GSO Research Organisation, Jansen Cilag, med update, Merck Serono, Merck Sharp Dohme, Novartis, Roche Pharma, Sanofi Aventis. The remaining 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. The handling editor declared a past co-authorship with the authors JW and WF.

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.