T lymphocytes play a critical role in anti-tumor immunity. Besides broadly discussed conventional αβ T lymphocytes, γδ T cells are also now recognized in the context of cancer inhibition. In the blood, among peripheral mononuclear cells (PBMC), γδ T cells generally account for 1 to 5% whereas they are predominant in tissues such as skin and intestine (1). Both residents, as well as circulating γδ T cells upon migration to the tumor site, can display an anti-tumor effect. With a structural difference between the γ and δ chains, γδ T cells can be divided into three main groups, Vδ1, Vδ2 and Vδ3 T cells, all of which recognize antigens independently of the major histocompatibility complex (MHC) molecules.

In B-cell malignancies, such as B-cell lymphomas, chronic lymphocytic leukemia (CLL) or multiple myeloma (MM), tumor cells can be found both in peripheral blood (PB) and in lymphoid organs, such as bone marrow (BM) or lymph nodes (LN). Therefore, these malignant cells can interact with other cell types constituting a specific microenvironment, in which infiltrating γδ T cells can play an important role. Vδ1 and Vδ2 T cells have been described to participate in the anti-cancer responses in B-cell malignancies with sometimes different proportions and different modes of action.

In this review, we first described γδ T cell diversity in B-cell lymphomas, CLL and MM. We then focused on γδ T cell activation and finally we presented attractive candidates for immunotherapies (IT) in B-cell malignancies.

Following T cell receptor (TCR) rearrangement, γδ T cells can be categorized into three main groups: the variable Vγ9 chain paired with Vδ2 (Vγ9Vδ2 T cells, also known as Vδ2 cells) (2), the variable Vδ1 chain with different Vγ chains (3) and Vδ3 T cells. Lymphocytes expressing heterodimers of Vδ2 and Vγ9 chains are predominant in the blood where they account for most (50–95%) of the γδ T cells, whereas Vδ1 T cells (paired with various Vγ chains) are more abundant in tissues, including healthy epithelia or solid tumors (4). Vδ3 like Vδ1 T cells were shown as dominant in the intestinal mucosa, skin, and liver (3), and to actively participate in cancer immunobiology.

These lymphocytes can differentiate into different T helper-like cells (Th1-, Th2-, Th9-, and Th17-like cells), producing a wide range of cytokines to fulfill their physiological role (5–7). More precisely, γδ T cells can harbor different phenotypes, such as: naive, central memory (CM), effector memory (EM) or RA⁺ effector memory (TEMRA) (8, 9). Moreover, γδ T cells co-express other functional receptors, including activating natural killer receptors (NKR: NKG2D, NKp30 and NKp44) (10, 11) and various Toll-like receptors (TLRs) (12). However, they can also express inhibitory NKR such as CD94/NKG2A or immune checkpoints (ICP), such as: PD-1, TIM3, LAG3 or CD39. Interestingly, NKG2A⁺ Vδ2 T cells were shown to exert higher anti-tumor potential (13).

Patients with Hodgkin’s Lymphoma (HL) were characterized with a marginally higher level of circulating γδ T cells, compared to healthy donors (14). The tumor escape from immune surveillance by the γδ T cells in these patients could therefore be due to the immunosuppressive profile of these cells plus an increase of soluble MICA derived from its shedding at the surface of lymphoma cells. Interestingly, HIV-infected individuals developing HL were also shown to display a significant expansion of the Vδ1 T cell subset compared to those without HL. To go further, the authors showed a high expression of CD16 and the inhibitory receptor CD158b by these Vδ1 T cells, concomitantly with a low expression of CCR5, CXCR4 and CXCR3, thus decreasing their homing to the tumor site (15). This discrepancy could point to a causal role in the pathogenesis of HL.

On the other hand, in B-cell non-Hodgkin’s lymphomas (B-NHL), the major subtypes of circulating γδ T cells were shown to be Vγ1, Vδ1 and Vδ2 (16). Compared to healthy donors, patients exhibit an absence of Vγ2 TCR subfamily in PB, BM, and LN. This implies a widespread restriction of the Vγ gene expression repertoire that may be a feature in patients with B-NHL. Moreover, the distribution of Vγ and Vδ subfamilies varied between PB, BM, or LN, and this may be due to the distribution or expansion of γδ T cells in different immune organs and to local immune responses (16).

In the case of diffuse large B-cell lymphoma (DLBCL), an aggressive form of B-NHL, γδ T cells represent a substantial population among infiltrating T lymphocytes. Amongst this population, Vδ1 T cells were shown as the major γδ T cell subset in both tumor and PBMC, whilst Vδ2 T cells were the most common subset in PBMC of healthy donors (17). In this study, the Vδ1 T displaying a naive phenotype (whether in blood or in LN) were shown as functional, however the authors did not observe any correlation between the rate of Vδ1 T and well-established prognostic factors, clinical responses or progression-free survival (PFS). Interestingly, the germinal center (GC) subtype of DLBCL was associated with an increase in Vδ1 cells in the tumors, whereas the non-GC subtype was associated with a lower frequency of γδ T cells (17). Activation or reactivation of Vδ1 T cells in DLBCL patients either by using ex-vivo expanded cells or by promoting their expansion in vivo, could represent a therapeutic outcome.

In the case of follicular lymphoma (FL), Braza and collaborators showed that γδ T cells as well as CD8⁺ T lymphocytes were located in the perifollicular zone of the LN of FL patients and not inside follicles. The majority of FL-LN γδ T cells are Vδ2 CCR7⁺ unlike circulating ones, whereas expression of the chemoattractant CCL19 chemokine is lower in FL-LN than in inflamed LN, explaining the low γδ T cell count in FL-LN (18). However, γδ T cells from FL patients displaying good cytolytic properties against lymphoma cells, ex-vivo expansion or promotion of in vivo expansion could be a therapeutic option if expanded γδ T cells can home in to the tumor site. Nevertheless, in this study, the authors considered only the Vδ2 subtype without taking into account the Vδ1 cells, which can counterbalance the decrease of Vδ2 T cells. In another study, the authors compared reactive LN from lymphoma-free individuals with FL-LN. Unlike Braza’s work, they showed no significant difference in the percentage of the cytolytic γδ T cell population between reactive LN and FL-LN (19). The immune microenvironment of LN can therefore have an important impact on the phenotypical and functional characteristics of infiltrated T cells.

Increase in Vδ1 T cells was also observed in CLL and MM patients. The analysis of PB of patients with CLL revealed a general prevalence of the Vδ1 T cell subtype, with an increased cell count in more severe stages of the disease (20). This increasing percentage of Vδ1 cells was also observed as belonging to the CD27^- compartment from controls to advanced stages of CLL patients, in particular in Binet B and C CLL groups, exhibiting a cytotoxic phenotype with the expression of granzyme B (11). Another study showed an increase of Vδ1 cells in the blood of CLL patients with stable disease, which were able to proliferate and produce TNF-α and IFN-γ in response to autologous CLL cells suggesting the potential of Vδ1 cells based therapies in this disease (21). These contradictions could be explained by the expression of some exhaustion markers, not determined in this study which can thwart the cytotoxic efficacy. An increase of exhaustion markers such as PD-1, TIGIT, TIM3 and CD39, expressed by Vδ1 cells was also shown in BM of MM patients. Whilst these patients displayed no difference regarding an overall level of γδ T cells in comparison to healthy donors, they showed a higher proportion of Vδ1 over Vδ2 T cells (22). Elevated percentage of γδ T cells with an exhausted phenotype (PD-1⁺), associated with a decreased expression of genes involved in effector functions was also found in patients with relapsed/refractory MM (23). However, exhaustion problems could be overcome by using ICP inhibitors such as anti-PD-1 antibodies. Additionally, inhibition of the anti-tumor immune response was associated with elevated levels of γδ regulatory T cells in the PB of MM patients with a bad prognosis (24) as well as in CLL patients (25). Besides, the increase of circulating cytotoxic γδ T in PB of MM patients after autologous hematopoietic stem cell transplantation was associated with improved PFS and overall survival (OS) (26).

Finally, γδ T cells were detected in all B-cell malignancies with an exhausted phenotype. Their reactivation or manipulation with immunotherapeutic approaches could represent a promising therapeutic option for patients with these diseases as developed in the section 4.

Depending on their TCR variant, γδ T cells, can respond to a variety of antigens. Vδ1 and Vδ3 cells can recognize, via their bound TCR glycolipids, MHC-related class Ib molecules CD1c for Vδ1 (27) and CD1d for Vδ3 (28). On the other hand, Vδ2 cells recognize non-peptidic antigens in the form of small pyrophosphate molecules called phosphoantigens (PAg), that can be found endogenously, such as hydroxymethyl-butyl-pyrophosphate (HMBPP/HDMAPP) or metabolites of the mevalonate pathway, or synthetic molecules, such as BrHPP (bromohydrine pyrophosphate) (29, 30). Zoledronate, a third generation aminobisphosphonate, a farnesyl pyrophosphate synthase inhibitor, widely used for osteolysis, has been shown to enhance antitumor Vδ2 cell responses, through the overexpression of endogenous PAg by tumor cells (31, 32). MHC-class I molecules are not involved in PAg recognition by Vδ2 T cells, but other molecules including butyrophilins (BTN3A1/BTN2A1), the ABCA1 transporter, the intracellular RHOB or periplakin molecules were shown to be involved in their activation (33–37). Besides TCR involvement, γδ T cells expressing NKR and TLRs, are also able to respond to stress-induced NKR ligands such as the ribonucleic acid export 1 (RAE1), MHC class I-related molecule A or B (MICA/MICB), UL16-binding proteins (ULBPs) (11, 38) amongst other DAMPs or PAMPs (12). Interaction of γδ T cells with cancer cells expressing these molecules, leads to the formation of an immunological synapse (39, 40) resulting in γδ T cell proliferation, cytokine release and tumor cell lysis (41). However, other molecules such as CD226 (DNAX accessory molecule-1), adhesion molecules (ICAM-1), CD3 or CD2, can also be involved in γδ T cell activation and favor their immune responses. These different activation modes are summarized in Figure 1 .

Considering the maturation stages, γδ TEM cells expressing high levels of chemokine receptors, produce large amounts of IFN-γ and TNF-α in response to TCR stimulation and γδ TEMRA cells, which express several NKR but low levels of chemokine receptors, are highly active against tumoral target cells and efficient to ADCC thanks to CD16 expression. These γδ T effector cells express their cytotoxicity through the production of high amounts of perforin and granzyme. Naive and CM cells do not display effector functions but are able to proliferate.

γδ T cells from patients with B-cell malignancies and particularly Vδ2 T cells were evaluated for their functionality, by the in vitro sensitization of cancer cells or cells of the tumor microenvironment with zoledronate (31, 32). In NHL, LN mesenchymal stromal cells were shown to interfere with Vδ2 lymphocyte cytolytic function and differentiation into Th-1 or EM cells but pre-treatment of these immunosuppressive cells with zoledronate can rescue lymphoma cell killing via the TCR and NKG2D (32). Vδ2 T cells from patients with B-cell lymphoma and MM, expanded in vitro by culture with zoledronate and IL-2, displayed enhanced cytotoxic effects towards MM/B-cell lymphoma cell lines and autologous tumor cells, without cytotoxicity against normal cells in these patients (31). However, approximately 50% of untreated MM patients showed Vδ2 T cells that were unable to proliferate upon stimulation with zoledronate and IL-2, but had strong effector properties exhibiting TEM or TEMRA phenotypes (42, 43). Similar results have been described for untreated CLL patients, who were classified as responders and displayed proliferation of zoledronate-stimulated Vδ2 T cells (25). Interestingly, the low-responders showed significantly greater baseline peripheral Vδ2 T cell counts than the responders, ruling out a quantitative defect. Indeed, the low-responder patients showed an accumulation of TEM and TEMRA Vδ2 cells with high effector functions and low capacity to proliferate, whereas naive and CM were preferentially found in responding patients (25, 44). In addition, Coscia and collaborators showed that a low proliferation capacity of Vδ2 cells was correlated with subsets of CLL patients with unmutated immunoglobulin heavy variable (IGHV) genes (U-CLL). This is in agreement with the upregulation of NKG2D on γδ T cells of CLL patients responding to zoledronate (25). Vδ2 T cells isolated from PBMC of MM patients were also shown to upregulate NKG2D upon in vitro expansion with zoledronate. Additionally, a low-dose treatment with bortezomib (proteasome inhibitor typically used in MM) sensitized MM cells to in vitro lysis by Vδ2 cells through NKG2D (45). Vδ1 cells may also be involved in the anti-cancer response towards CLL cells through NKG2D activation. Indeed, Vδ1 cells, which are enriched in PB of CLL patients, can kill neoplastic CLL cell lines transfected with MICA, and blocking anti-NKG2D antibody largely decreases autologous leukemic cell lysis (21). In addition, purified Vδ1 cells isolated from the PB of MM patients can kill MM cell lines, and produce cytokines involving their TCR and NKG2D, DNAM-1 and adhesion molecules (46). Unfortunately, cancer cells are able to express some enzymes such as ADAM 10 and 17 and are able to shed the stress molecules MIC-A and–B and ULBPs from their surface, therefore decreasing the γδ T response through NKG2D (47).

As γδ T cells, whether Vδ1 or Vδ2, can be activated by B-cell lymphoma, CLL or MM cells, these cells represent essential actors in anti-tumor responses against B-cell malignancies and can definitely be good targets for IT in these diseases.

For a long time neglected, γδ T cells have emerged as a key immune cell type in cancer biology, representing very attractive and promising candidates for cancer IT. Their therapeutic potential in solid and hematological cancers have been extensively reviewed elsewhere (48–53) and here we focused on their exploitation in B-cell malignancies.

γδ T cells can be used in several strategies, based on the in vivo activation to potentiate the tumor-targeting or the optimization of in vivo or ex vivo expansion protocols. These approaches consist of: i) combination with therapeutic drugs or antibodies, ii) adoptive cell transfer (ACT) using of autologous or allogenic γδ T cells expanded ex vivo and iii) ACT using genetically modified γδ T cells. The timeline of γδ T cell-based IT is presented in the Figure 2 and the various cell-based IT, ongoing clinical trials, and associated sponsors are summarized in the Table 1 .

Human γδ T cells present several specific characteristics that make them very attractive for their use in anti-cancer therapy in general and anti-lymphomatous in particular. Firstly, in contrast to αβ T cells, their anti-tumoral activity does not depend on mutational burden, thus rendering them efficient against tumors harboring few somatic mutations. Secondly, as they do not act dependently of MHC I-mediated Ag presentation, unlike CD8⁺ αβ T cells, they exhibit an anti-tumoral efficacy against tumors harboring a downregulation of surface MHC class I molecules. This characteristic is particularly well suited for the “off -the-shelf” allogenic cell therapy. Thirdly, they exhibit increased anti-tumoral activity due to their particular activation mechanisms present on both adaptive cells through the TCR signaling and innate cells through NK signaling (NKG2D, DNAM-1, NKp46, NKp44, NKp30). This is amplified by their low expression of killer inhibitor receptor.

Although Vδ2 T cell-based IT exhibited safety and good tolerance in patients, they also demonstrated limited success due to several reasons among which a highly variable capacity to recognize tumoral cells, functional instability, dysfunction or exhaustion of chronically activated Vδ2 T cells. Thus, innovative strategies were developed to improve tumoral cell recognition, promote durable persistence and circumvent exhaustion mechanisms involving Vδ2 but also Vδ1 T cells. In this context, engineered cells such as DOT and CAR-T offer very encouraging perspectives as well as combination of Vδ2 T cells with antibodies targeting ICP or neutralizing inhibitory cytokines to counteract immune suppression TME and exhaustion processes or transduction of selected high affinity Vγ9Vδ2 TCR into αβ T cells in order to induce a durable and memory-based response.

Several challenges remain, among them the difference of efficacy between cell engagers and ACT involving γδ T cells considering their logistical requirements and costs. Another important point to be considered is the justification for the selection of patients to be treated by such IT based on the identification of tumor Ag recognized by γδ T cells. All these issues will help to better understand, use and develop next-generation γδ T cell-based IT.

LR, CF and MD contributed equally to the work of this review. CL participated in the writing. MP and CB wrote the review and supervised the work. All authors contributed to the article and approved the submitted version.