γδ T cells are loaded with effector weapons of great potential for cancer immunotherapy (160). Findings in recent years point to important roles for these cells, highlighting them as potential predictive biomarkers, which justifies the current focus of studies on the nature of these cells and the TME (14, 25, 161). It is important to remember that several characteristics discussed here make γδ T cells potential candidates for innovative therapies against tumors and include: (i) activation in a TCR-independent manner; (ii) the ability to recognize Ags regardless of MHC/HLA expression; (iii) effector molecules production and direct and indirect cytotoxicity potentiation against cancer cells; and (iv) their role as antigen-presenting cells (APC) that induce the proliferation of antitumor cells (
Figure 2
).
Given the high responsiveness against LCs and the absence of toxicity or alloreactivity against the host (162, 163), the application of γδ T cells in leukemia treatment may mean a new advance in cancer immunity and immunotherapy. To make this possible, several strategies for γδ T cell handling have been developed and tested and have presented interesting data (
Figure 3
). The following subsections will focus on clinical trials and findings, as well as the activity of these cells in response to applied methods. Afterwards, we will discuss potential therapies that may specifically target γδ T cells and their subtypes, while summarizing the main approaches that are being explored to reach their clinical potential.
Expanding γδ T Cells With pAgs, Drugs, Cytokines and Feeder Cells
Intrinsic synthesis of pAgs in cancer cells can be manipulated through pharmacological blockade mediated by aminobiphosphonates (N-BP), such as zoledronate (ZOL) and pamidronate (PAM), which interfere metabolically in the mevalonate pathway (164). The mechanism involved causes these compounds to block the enzymatic activity of farnesyl pyrophosphate synthase, which is present in this metabolic pathway. N-BP–induced interruption results in the intracellular accumulation of pAgs in cancer cells or APCs with subsequent recognition by γδ T cells and activation after cell-cell interaction (165–168). Cancer cell sensitizing with these compounds increases the tumor’s susceptibility to γδ T cell cytotoxicity, and this also applies in leukemia (132).
Some experimental evaluations took advantage of the γδ T cell recognition mechanism (directed to pAgs) to obtain a better in vitro or in vivo expansion of these lymphocytes and test their therapeutic efficacy. To date, these approaches have focused on ZOL, (E)-4-Hydroxy-3-methyl-but-2-enyl pyrophosphate (HMB-PP) and synthetic pAgs, such as bromohydrin pyrophosphate (BrHPP) (169, 170). These compounds are generally administered in combination with low cytokine doses such as IFNs, IL-2, IL-12, IL-15, IL-18 and IL-21. In addition, these approaches can induce an antitumor phenotype and the pronounced expression of associated receptors (171–176).
Vγ9Vδ2 cell expansion has become more accessible because, in addition to being the most prevalent subtype in PB (55, 56), it can also recognize a diversity of relatively well-defined target molecules (177). When these cells are treated with ZOL + IL-2 + IFN type I, their cytotoxic activity is increased and Vγ9Vδ2 cells may be able to efficiently destroy lymphoid and myeloid lineage LCs, as proposed by Watanabe et al. (171). In their study, γδ T cells were generated in vitro with ZOL + IL-2 for 14 days, and after this period they were activated with IFN type I for up to 3 days. Thus, the resultant γδ T cells were well expanded in the culture and showed a significant expression of CD69, TNF-related apoptosis-inducing ligand (TRAIL), IFN-γ and TNF, which suggests the acquisition of an activated phenotype and antileukemic reactivity.
In the same vein, sensitization with ZOL + Imatinib has also been shown to increase the cytotoxic synapse between Vγ9Vδ2 cells and LCs (178). Initially, Imatinib resistant or sensitive LCs had low susceptibility to γδ T cells, but in vitro treatment with ZOL + Imatinib was able to reverse this situation. The lysis of these LCs was mediated by TCR, NKG2D, TRAIL and perforins. This high cytotoxicity was dependent on ZOL, since it was observed that Vγ9Vδ2 cells exerted low antitumor activity that was slightly increased after sensitization of LCs. To validate these findings, it was further demonstrated that when Vγ9Vδ2 cells, ZOL and IL-2 are infused in a leukemia mouse model, they mediate tumor regression in vivo and confer greater survival in these mice (178).
The ability of N-BPs to invigorate exhausted Vγ9Vδ2 cells has also been reported in other investigations (179, 180) and appears to be a promising alternative for their use, given that a higher exhausted γδ T cell frequency has also observed in leukemia (27, 125). It is important to note that, in this context, these cells exhibit a low expression of CD107a, FcγRIII (CD16), IFN-γ and TNF, while B and T lymphocyte attenuator (BTLA), LAG3 and PD1 proteins are more highly regulated on their cell surface (179, 180). When cultured with allogeneic LCs, these lymphocytes had low cytotoxic activity, while γδ T cells from healthy patients responded efficiently (179). Notably, when Vγ9Vδ2 cells that were considered dysfunctional were cultured ex vivo with mature monocyte-derived dendritic cells (Mo-DC) and N-BPs for 8 days without the presence of LCs, the observed functional impairments could be reversed (179).
Ibrutinib has also been shown to activate γδ T cells against LCs, since Weerdt et al. reported that it was able to induce an antitumor phenotype (180). In their study, allogeneic and autologous γδ T cells were cultured with LCs. As already seen, γδ T cells from patients with leukemia proved to be dysfunctional in terms of cytokine production and cytotoxicity, while those from healthy patients had a strong antitumor activity (179). When Vγ9Vδ2 cells from both cases are treated with Ibrutinib, an effector Th1 phenotype and memory cells are induced. Overall, their antitumor properties can be recovered after ex vivo stimulation and after treatment with Ibrutinib, which binds to the IL-2–inducible T cell kinase molecule and promotes activation against LCs (180).
Other investigations have presented a new alternative: the combination of IL-15 plus N-BPs or pAgs promotes significantly greater expansion, high cytotoxicity and a more pronounced Th1 phenotype in γδ T cells, when compared to expansion methods using only IL-2 (174, 181, 182). IL-15 is a powerful growth factor for γδ T cells (183, 184) and can synergize with other molecules and enhance the antileukemic capacity of these cells (
Figure 2
), as we will highlight below.
The ex vivo tests carried out by Van Acker et al. (174) demonstrated that the administration of IL-15 + isopentenyl pyrophosphate (IPP) is able to improve γδ T cell cytotoxicity against LCs. In contrast, γδ T cells stimulated with IL-2 + IPP were more likely to deviate to a Th2 and Th17-like response phenotype when interacting with LCs (174). It is important to highlight that stimulation by IL-15 promoted a more robust IFN-γ and TNF secretion when compared to IL-2 stimulation. In addition, culturing these lymphocytes with IL-2, IL-15 and ZOL for 14 days critically enhanced the expansion rates to almost 1000-fold the total yield of viable cells, which showed a 590-fold increase in the γδ T cells cultured only with IL-2 + ZOL (174).
Interestingly, when IL-2 + IL-15 and ZOL are administered to γδ T cells isolated from patients with leukemia, during 14 days of culture, they assume different phenotypic states. Most of them may exhibit an effector memory phenotype (CD45RA- CD27-), followed by a central memory phenotype (CD45RA- CD27+) (174). In addition, positive regulation of CD56, CD80 and CD86 is also provided (174), suggesting that, in addition to exerting strong antileukemic activity, these cells may also act as APCs and improve the antitumor responses.
Vγ9Vδ2 cell expansion using IL-2 may not even promote satisfactory proliferative rates; however, it is clear that the synergism between IL-2 and IL-15 confers a substantial increase in an inflammatory profile (174, 181), as these cytokines promote a higher transcription factor T-bet expression (181), which in turn, is related to greater cytotoxicity. In addition, the advantage of γδ T cells expanded with IL-2 + IL-15 can be maintained under one of TME’s hallmarks in vivo, namely hypoxia (181). In fact, a striking feature of the leukemic microenvironment is the low partial pressure of oxygen that favors the tumor-associated immunosuppressive pathways, while at the same time promoting expansion of LCs (185). In this context, the persistence of γδ T cells in hypoxia further demonstrates their clinical importance.
Alternatively, the combination of ZOL, IL-2 + IL-18 also promotes the proliferation of effector cells (186, 187) since IL-18 is an important inducer of IFN-γ secretion (188). Given this, it has been reported that this cytokine indirectly induces the expansion of γδ T cells. Tsuda et al. (186) showed that Vγ9Vδ2 cells are efficiently expanded in response to ZOL, IL-2 + IL-18, but in a CD56bright CD11c+ NK-like cell dependent fashion (187). Many studies have reported that the involvement of NK-like cells in the proliferation of γδ T cells implies greater expansion efficiency when compared to methods using dendritic cells (DC) or monocytes (187, 189–191). These findings suggest an approach targeted at feeder cells that may be responsible for γδ T cell clonal proliferation in different methods in vitro and, perhaps, in vivo.
IL-18 can also directly support γδ T cell expansion, even in the absence of feeder cells (192). When Vγ9Vδ2 cells are treated only with ZOL, there is a delay in their in vitro expansion, as prolonged exposure subjects these cells to acute ZOL toxicity (193). However, when IL-18 combined with geranylgeranyl pyrophosphate (GGPP) is added, the proliferative capacity is restored by inhibiting the toxic effects of ZOL, which allows a substantial expansion of viable γδ T cells to occur. IL-18 + GGPP also were able to activate γδ T cells, exhibiting a central memory or effector memory phenotype and with higher IFN-γ production and CD56 expression (192).
In a subsequent study, treatment with ZOL + IL-2 and culture with Mo-DCs stimulated an activated phenotype in γδ T cells. In this context, immature Mo-DCs have been shown to have a particularly higher capacity to intensify γδ T cell cytotoxicity against LCs, whether in autologous or allogeneic condition (194). Furthermore, IL-15 producing DCs isolated from healthy patients and patients with leukemia (in remission) can potentiate γδ T cell cytotoxicity in vitro (182). These DCs induced NKp30, CD16, CD80 and CD86 expression in γδ T cells in an IL-15 dependent manner. This methodology was able to produce γδ T cells with higher expression of co-stimulatory molecules and low expression of inhibitory proteins. In addition, stimulation with DCs + IPP + allogeneic LCs led to high IFN-γ secretion and strong antitumor activity (182).
Deniger et al. (162) demonstrate a new strategy that involves the use of artificial APCs (aAPCs) derived from the K562 leukemic lineage. These feeder cells were modified to express molecules, such as CD19, CD64, CD86, 4-1BBL and IL-15, on their membrane surface. When γδ T cells are cultured with aAPCs + IL-2 + IL-21, there is a remarkably robust 4900 ± 1700-fold polyclonal expansion (162). Most of these cells expressed different γδ TCR domains. Resultant γδ T cells also were able to kill LCs via TCR, NKG2D and DNAM-1 (162).
In the same vein, Cho et al. (175) used CD80+, CD83L+ and 4-1BBL+ aAPCs. At low IL-2 concentrations, these co-stimulatory molecules promoted a remarkable Vγ9Vδ2 cell expansion that secreted higher levels of IFN-γ and TNF (175). Notwithstanding, there was no significant proliferation rate (106-fold increase) when compared to the hefty increase observed in the previous study (162). Triple co-stimulation with these molecules induced not only the high IFN-γ and TNF production, but also the positive regulation of a range of other molecules such as IL-2, IL-6, perforins, Gzm A and Fas ligand (FasL) (175). Most importantly, the expanded cells exhibited a terminal effector phenotype (CD27low CD45RAhigh), followed by an effector memory phenotype (175).
Unlike most of the investigations discussed above, other studies have focused on γδ T cells that express the Vδ1 TCR chain. Substantial evidence has demonstrated the ability of this subtype to kill LCs, as already reviewed. Unlike the Vγ9Vδ2 subtype, these cells do not show susceptibility to activation-induced cell death (AICD), which has been reported in several experimental trials (125, 195, 196). These cells can also exercise immune surveillance for long periods, favoring the longevity of cancer immunity (197–199). Several unique attributes have been discovered that particularly place Vδ1 cells as attractive targets in antileukemic therapies. So far, a few studies have emerged that have sought to translate the functional role of these lymphocytes and their applicability, as we will highlight below.
Siegers et al. (30) developed an in vitro expansion protocol that enabled the proliferation of γδ T cells isolated from PB after treatment with lectin-based compounds named Concanavalin-A (Con-A). Thus, it was possible to expand Vδ1 cells in a greater proportion than the Vγ9Vδ2 subtype when Con-A was combined with IL-2 + IL-4. The low Vγ9Vδ2 cell proportion was motivated by the period of exposure to Con-A, which induced AICD in these lymphocytes (30). Noteworthy, the resulting Vδ1 cells exerted an efficient cytotoxic activity against LCs through TCR, NKG2D, CD56 and FasL (30).
Subsequently, proof-of-concept studies were performed on leukemia xenograft models using a newly established cell generation protocol called Delta One T (DOT), which was designed by Almeida et al. (31). Specifically, this clinical-grade protocol consists of two steps. First, γδ T cells are isolated from PB of healthy donors or patients with leukemia using magnetic beads and are cultured in vitro for 14 days. During this time, these lymphocytes are expanded using a combination of molecules, such as IFN-γ, IL-1β, IL-4 + IL-21, in association with anti-CD3 antibodies (Abs). Then, the expanded cells are transferred to a new culture medium, where they are restimulated by anti-CD3 combined with IL-15 and IFN-γ for another 7 days (31). Overall, this is a 3-week protocol that involves γδ TCR and cytokine stimulation that can accomplish its goals efficiently.
When γδ T cells were submitted to the DOT protocol, expansion was obtained with rates greater than 1000-fold, thus allowing the viable and efficient proliferation of highly cytotoxic cells. It is noteworthy that, with this cell proportion rate, Vδ1 cells, which are generally less frequent in the blood (55, 56), expand from less than 0.5% of all circulating T cells to more than 70% (25, 31). Notably, Vδ1 cells with high expression of NKp30, NKp44, DNAM-1 and 2B4 are also provided, all well established as key-receptors in antileukemic responses (66, 74). These lymphocytes do not regulate inhibitory proteins on their membrane surface, even after 3 weeks of continuous stimulation. In addition, many cell adhesion molecules and chemokine receptors are positively regulated, while these lymphocytes can kill autologous and allogeneic LCs in vivo, and ignoring normal cells (31).
Finally, the same protocol was tested by Lorenzo et al. (200), in which γδ T cells from PB were reinforced using a range of stimulatory molecules (31). While the previous study sought to mobilize Vδ1 cells against a CLL xenograft model (31), the latter work applied the DOT protocol to an AML xenograft model (200). It is important to highlight that in both cases there was an efficient regression of tumors, and this increased mice survival (31, 200). In addition, γδ T cells avoided systemic metastasis of LCs (31). They exerted their antileukemic activity against AML blasts in a partially TCR-dependent manner, while they depended on the B7-H6 expression (200), which binds to NKp30 (67).
Blocking Immune Checkpoints in γδ T Cells and Leukemic Cells
Although they are potent, γδ T cell antitumor responses can be regulated by immune checkpoints (IC). Many inhibitory proteins, such as PD1, CTLA4, LAG3, BTLA, T cell immunoreceptor with Ig and ITIM domains (TIGIT) and T cell immunoglobulin and mucin domain-containing protein 3 (TIM3), are key mediators in inflammatory regression and cell suppression, in the context of the TME (93, 201, 202). Generally, these molecular interactions can act synergistically with the infiltration of suppressive cells that support tumor evasion through the establishment of a strongly tolerogenic environment (26, 76). However, recent advances in cancer immunotherapy using monoclonal Abs (mAbs) targeting ICs, the immune checkpoint inhibitors (ICI), have shown that combinatorial blocking of proteins, such as PD1 and PD-L1, can restore cellular functions and reestablish antitumor activity (203, 204).
The mechanisms of γδ T cell regulation mediated by ICs are diverse and poorly understood, but seemingly unified by the fact that these receptors functionally complement each other and ensure the adjustment of the immune response. PD1 and BTLA are the most potent ICs shown to suppress γδ T cell cytotoxicity in cancer (205, 206). Although CTLA4 expression has not been consistently assessed, it is known that this molecule is rarely expressed in activated γδ T cells (207, 208). Importantly, the expression of these ICs may vary between γδ T cell subtypes, where, for example, Vδ1 cells exhibit higher PD1 expression than their Vγ9Vδ2 counterpart (209).
Early after activation, when the γδ TCRs find their cognate ligands, γδ T cells begin to rapidly display many of these ICs on the cell surface (205, 207). Collectively, the expression of these proteins is low or stable, but temporary, and is sufficient to reduce cytokine production, proliferation and survival of γδ T cells (205, 206, 208, 210, 211). These changes can also be observed in leukemia, as γδ T cells increase the expression of PD1, CTLA4 and BTLA, while LCs strongly regulate the expression of their ligands, such as PD-L1, CD80 and/or CD86, and herpesvirus-entry mediator (HVEM), respectively (212). This represents an important barrier, as these molecules can prevent the efficient activation of γδ T cells and the associated antitumor response. Blocking the expression of these inhibitory receptors through the use of ICIs may be an interesting alternative to reverse the state of anergy and/or cell exhaustion.
The influence of ICIs on γδ T cells and their potential impact on the associated cytotoxic activity, in the context of the leukemic microenvironment, has not yet been characterized and is, therefore, an open question. Despite this, PD1 has been shown to negatively regulate Vγ9Vδ2 cell responses, while the addition of ZOL + anti–PD-L1 was able to bypass the inhibitory signals and promote γδ T cell reactivation against LCs in a TCR-dependent fashion (205). Therefore, this discovery allows us to suggest that γδ TCR-mediated activation is capable of overcoming the inhibitory effects of the PD1/PD-L1 pathway, since the application of ICIs plus ZOL, which is a strong Vγ9Vδ2 TCR stimulator, apparently synergizes the activation of γδ T cells and restores their tumor reactivity (205).
Notably, Hoeres et al. (213) demonstrated that although PD1 signaling can modulate the production of IFN-γ in leukemia-reactive γδ T cells, its additional blockage and stimulation with ZOL can increase the production of this cytokine. Although it did not show a significant effect on the destruction of LCs by γδ T cells, the action of anti-PD1 + ZOL in these lymphocytes was able to induce high IFN-γ secretion, which is a potent inflammatory and antitumor factor (213). As noted, cytokine secretion, such as IFN-γ, can be negatively regulated, and we can infer from this study that the application of ICIs potentially reverses this suppressive condition and is able to stimulate the triggering of an antitumor response.
In addition to PD1/PD-L1, other inhibitory proteins are highly regulated in LCs (i.e., CTLA4, BTLA, TIGIT, TIM3 and LAG3) and their effects on γδ T cells have not yet been fully investigated (26, 159, 214–216). However, previous studies have shown that some of these receptors have great potential for deregulating their antitumor activity, reflecting in cytokine production (213, 217). Nonetheless, evaluating these components before proceeding to a clinical application is important, since these molecules most likely prevent the efficient killing of LCs. This is one of several mechanisms of tumor escape that are commonly observed in recent and innovative treatment modalities, and which also include the chimeric antigen receptor (CAR) T cell therapy (218).
Focusing on γδ T Cell-Engager Molecules in the Leukemic Microenvironment
Antibodies Direct γδ T Cells Against LCs
As we have shown herein, data from in vitro experiments and mouse models unequivocally demonstrate the potential of γδ T cells against leukemia. Knowledge obtained regarding the many signals that regulate their activation and the tumor resistance underlying γδ T cells offers additional approaches that, in addition to inducing an activated status, a (poly)clonal expansion or a more pronounced Th1 phenotype, may also allow more specific targeting against the tumor. Improving γδ T cell efficiency against LCs, however, requires strategies based on their cytotoxic nature, which include, for example, antibody-dependent cell cytotoxicity (ADCC) (75, 219). Therefore, this implies a role for CD16, mAbs and bispecific antibodies (bsAbs) that bind to their respective target antigens.
CD16-mediated ADCC plays an important role in tumor destruction. For this to occur, CD16 must bind to the constant fraction of Abs IgG, thus constituting an optional axis in target cell killing. γδ T cells constitute the major blood T cell population that expresses CD16 (220, 221), although this expression is variable (222). Given this, the potential engagement of therapeutic Abs with the product of Vγ9Vδ2 cells can provide an efficient alternative against LCs (223). Several studies have shown that γδ T cells mediate leukemic regression via a CD16-dependent pathway (136, 223–226), in particular the Vγ9Vδ2 subtype, which positively regulates CD16 and TNF expression when stimulated with pAgs (227).
mAbs-coated LCs are efficiently destroyed by CD16+ γδ T cells via ADCC and these lymphocytes subsequently exhibit APC functions and activate αβ T cells, apparently through the tumor Ags presentation by MHC class II (228). It has been shown that the application of therapeutic CD20-targeting Abs, such as Rituximab (RTX), improves the antileukemic effect of these lymphocytes through tumor destruction by ADCC in vitro. This leads γδ T cells to secrete high levels of IFN-γ, perforins and CCL5 (219). In addition, BrHPP implementation potentiates the RTX bioactivity and consequently also increases γδ T cell cytotoxicity against CD20+ LCs in vitro and in vivo (75).
When peripheral blood mononuclear cells (PBMC) are stimulated with ZOL + IL-2 ex vivo and then cultured with LCs and Obinutuzumab (anti-CD20), it is observed that γδ T cells perform ADCC more efficiently than NK cells (223). Most importantly, the cytotoxicity of these lymphocytes cultured with Obinutuzumab is more potent compared to other tested mAbs, such as RTX. This view was reinforced when LCs treated with Obinutuzumab were substantially lysed in a CD16-dependent manner (223).
Benyamine et al. (136) demonstrated that BTN3A-targeting mAbs (anti-BTN3A 20.1) sensitize LCs and act indirectly in tumor destruction. This is due to the anti-BTN3A Abs binding in three different target molecules: BTN3A1, BTN3A2 and BTN3A3. The combination of these mAbs with γδ T cells and the subsequent infusion in a leukemia murine model was able to decrease the leukemic load in the PB and BM, increasing survival in these mice (136). Taken together, these data create the expectation that targeting mAbs to BTN proteins can be potentially useful in new therapeutic approaches.
Like most other surface molecules expressed in LCs, CD19 is also a potential target to be considered. When LCs are incubated with γδ T cells and modified anti-CD19 Abs (Ab 4G7SDIE), a significant increase in the degranulation marker CD107a is observed, as well as the strong IFN-γ and TNF production (224). In addition, the adoption of bsAbs targeting CD19/CD16 (bsAbs N19-C16) is also able to increase the expression of these inflammatory molecules (224). Interestingly, bsAbs targeting CD19/CD3 (bsAbs N19-CU) also strongly activated γδ T cells and, unlike the other previously tested Abs, mediated the lysis of LCs (224). It should be noted that the use of Abs modified to have a triple specificity to CD16 and CD19 (triplebody SPM-1) was also able to activate these lymphocytes against CD19+ target cells, which was evidenced by the expression of antitumor mediators (225).
The projection of a bsAbs targeting the Vγ9 TCR chain and CD123 (anti-TRGV9/CD123 engager) was also able to recruit γδ T cells against AML blasts (229). This engagement induced its activation and cytotoxicity against endogenous LCs, as evidenced by CD69, CD25 and Gzm B positive regulation. Interestingly, these activated γδ T cells exhibited a low secretion of IL-6 and IL-10, which are cytokines that are highly related to cytokine release syndrome (CRS) in patients undergoing αβ T cell-based therapies (229–231). The efficacy of this approach is evidenced when anti-Vγ9/CD123 directed γδ T cells were infused into a leukemia mouse model and controlled the leukemic proliferation in different compartments in these mice (229).
Finally, it has been shown that CD1d is also an attractive target. A recent study showed that CD1d specific single domain Abs can guide γδ T cells (226). These engagers were able to mobilize and activate these lymphocytes against autologous LCs from patients with CLL. This allowed γδ T cells to produce many inflammatory molecules and maintain their pAgs reactivity (226). Taken together, the many studies reviewed here allow us to suggest that the therapeutic application of Abs can be improved with the use of N-BPs that enhance γδ T cell activation. However, their therapeutic application against leukemia still needs more detailed investigation.
γδ T Cells Expressing CARs
While the application of therapeutic Abs has significantly increased the effectiveness of leukemia treatments, other approaches are also emerging with promising healing potential. Current advances in genetic engineering enable CAR transduction in NK cells, macrophages and T cells, thus offering new horizons for cell therapy, although this has been primarily focused on conventional αβ T cells (232, 233). In this context, γδ T cells are also undergoing a number of improvements in order to enhance their antitumor capacities.
The fact is that γδ T cells can be redirected with CARs against surface molecules expressed by LCs (234). Their unique innate properties and their high capacity for tumor sensing and killing place them in an interesting position in potential approaches against leukemia. CAR γδ T cells can offer a triple activity because, for example, they can recognize LCs (i) through the direct engagement of γδ TCR to their cognate ligand, (ii) through NKRs and their associated ligands, or (iii) through CAR specificity to the target antigen (
Figure 3
) (234, 235). Besides this, their APC functions (211) may allow the prolongation of immune response in the TME (228), since the CAR acquisition preserves the ability of γδ T cells to present tumor Ags (235).
The applicability of these genetically modified T cells has been established by some of the previous studies that evaluated the viability of viral transduction (236, 237) or electroporation (238) of the CAR. Rischer et al. (236) demonstrated for the first time that Vγ9Vδ2 cells can be efficiently transduced with CAR genes. Their study also showed that γδ T cells expressing anti-CD19 CARs destroy CD19+ LCs and produce high levels of IFN-γ in a target-dependent fashion (236). Subsequently, Deniger et al. (238) showed that the introduction of CAR by electroporation in PB-derived γδ T cells is able to produce polyclonal CAR T cells that express Vδ1, Vδ2 and Vδ3 TCR chains (238). For this to happen, approaches already reviewed here were used (162).
Noteworthy, one study demonstrated that CAR γδ T cells adopt a highly activated, but not exhausted, phenotype, as highlighted by the low regulation of CD57 (238). In addition, these lymphocytes tend to assume distinct phenotypic states of effector memory, while positively regulating homing molecules. Specifically, these homing receptors included CXCR4, a molecule associated with migration to BM, as well as CD62L and CCR7, which are linked to migration to lymph nodes (238). This is encouraging since BM and lymph nodes are sites of high tumor growth in acute and chronic leukemias (1, 4, 26, 78, 239–241).
Surprisingly, it has also been confirmed that CAR γδ T cells recognize and kill LCs in BM regardless of the CD19 target. Rozenbaum et al. (242) recently showed that these modified lymphocytes have high IFN-γ production and reactivity to CD19+/- LCs in vitro, which was even enhanced with the addition of ZOL. To investigate in vivo efficacy, the authors injected CAR γδ T cells in a leukemia mouse model. Although it did not induce a complete remission, the infusion of these cells led to a drastic reduction in the leukemic burden in the BM of these mice, which was even more pronounced when ZOL was administered (242).
These studies demonstrate that the production of CAR γδ T cells is viable and supports the high effectiveness of these lymphocytes against many malignancies, especially in leukemias. In contrast to conventional CAR T cell therapy, approaches based on γδ T cells can overcome several currently reported limitations, such as modulation of tumor antigen expression (242, 243) and CRS (229–231).
How About Molecular Switching of TCRs?
One interesting strategy for targeting lymphocytes against the tumor is to design γδ T cells with αβ TCRs or to design αβ T cells with γδ TCRs (244). This therapeutic concept has great potential for combining some unique γδ T cell properties, such as the rapid responsiveness to the tumor, the expression of individual molecules, and the absence of alloreactivity, with the high proliferative capacity and specific reactivity of conventional αβ T cells. Combining these unique aspects through TCR transduction leads us to expect that the resulting antileukemic responses will be long-lasting and based on immunological memory.
This new concept of modified T cells, named T cells engineered with defined γδ TCRs (TEG), was adopted in some studies that showed that TEGs kill LCs in vitro and in vivo models (245). TEGs tend to deregulate the intrinsic αβ TCR expression in their membrane surface, avoiding the graft-vs-host disease (GvHD) (245, 246). In addition, CD4+ TEGs retain their ability to induce a complete maturation of DCs, and stimulation with PAM can potentiate the cytotoxicity of CD8+ or CD4+ TEGs since it promotes higher production of inflammatory molecules, such as IFN-γ, TNF, and IL- 2, in vivo (245).
Similar results were obtained when TEGs cultured with LCs reduced the tumor in vitro (247). In addition, the infusion of TEGs plus IL-2 + PAM in an AML murine model enabled reactivity directed to LCs without affecting the healthy hematopoietic compartment and without being influenced by the TME, when inserted into mice that expressed IL-3, granulocyte-macrophage colony-stimulating factor (GM-CSF), and stem cell factor (SCF) (247), which are molecules that support tumor growth in vivo (248). Therefore, TEGs demonstrated efficiency in reducing the tumor in xenograft models with minimal alloreactivity, which stimulated the projection of a robust manufacturing procedure of TEGs that were validated under good manufacturing practice (GMP) conditions (244, 249).
Finally, γδ T cells transduced with αβ TCR plus CD4 and CD8 co-receptors showed high antitumor activity against LCs (250). As similarly observed in TEGs, transduction of αβ TCR induced a low expression of endogenous γδ TCR. In addition, modified CD8+ or CD4+ γδ T cells expressed high levels of IFN-γ and IL-4, although IFN-γ production was more pronounced in CD8+ cells. Most importantly, these transduced cells were able to kill LCs in vitro, although CD8+ γδ T cells have shown more efficiency than CD4+ cells (250, 251). This evidence supports the important role of γδ T cells in TCR gene transfer-based approaches while suggesting an improved antileukemic capacity when TCR transduction is combined with co-receptors, in particular, with the CD8 protein.
Converting γδ T Cells Into Living Drugs
Source, Isolation and Pre-Activation
γδ T cells and their subtypes are present in several tissues, but the ideal source for obtaining all these lymphocytes is still being determined. Despite this, therapeutic γδ T cells for infusion can be obtained from peripheral blood (252, 253) or umbilical cord blood (UCB) (254, 255). It is important to note that the frequency of γδ T cells varies between 5-10% of peripheral blood T cells (51, 52), while they constitute <1% of T cells in UCB (254). The functional differences between γδ T cell subtypes in these sources are not yet clear, but it is already established that while the subtype expressing Vγ9Vδ2 TCR predominates in PB (51, 55, 56), polyclonal γδ T cells expressing the Vδ1 TCR domain predominate in UCB (52, 256, 257).
γδ T cell expansion from PB is a well-established method and is usually adopted in clinical and experimental trials. For isolation of these lymphocytes, the starting material is the product of leukapheresis, which can be initially enriched through stimuli with several soluble factors (e.g., cytokines and N-BPs) and later undergoes removal of αβ T cells and CD19+ B cells through the use of magnetic beads, depletion or separation kits (optionally maintaining NK cells) (169, 252, 253). Since increasing the γδ T cell product from leukapheresis can further improve its therapeutic handling, adopting the use of molecules as the granulocyte colony-stimulating factor (G-CSF) may mobilize a large amount of antileukemic γδ T cells for peripheral blood, as shown in several studies (258–262).
Alternatively, physical exercise and the consequent systemic activation of β-adrenergic receptors (β-AR), immediately before PBMC isolation, has been shown to substantially increase mobilization for PB, ex vivo expansion and antitumor capacity. In their study, Baker et al. (263) showed that the practice of physical exercises can predict the expansion potential of γδ T cells, which is mobilized in a β-AR type 2 dependent fashion. Therefore, patients with high levels of physical activity mobilized γδ T cells that expanded ex vivo in much higher percentages compared to blood at rest when stimulated with IL-2 + ZOL for 14 days (263). These cells had higher expression of CD56 and NKG2D and showed high cytotoxicity against LCs in vitro.
On the other hand, γδ T cell isolation from UCB is still poorly investigated and so far, it has not been the target of cell expansion protocols in clinical trials. Berglund et al. (264) showed that it is possible to expand γδ T cells derived from UCB in vitro. The authors developed an expansion protocol based on the application of ZOL + IL-2 in culture for 14 days. This promotes the growth of Vγ9Vδ2 cells that mostly adopt a central memory phenotype and secrete higher levels of IL-1β, IL-2 and IL-8 (264). In general, the acquisition and handling of UCB-derived γδ T cells still need to be investigated more fully. Some factors, such as the low frequency of Vγ9Vδ2 cells (more easily expanded in vitro) in UCB and the poorly defined phenotypic diversity in this environment, make handling more limited (254). The approaches discussed here are viable targets for adoptive cell therapy because they also serve as adequate and economical adjuvants for hematopoietic stem cell transplantation (HSCT) (263, 264).
It is not clear whether pre-activation with ZOL + IL-2 can trigger the total antitumor capacity of γδ T cells. However, many in vitro approaches that use other molecules, such as IL-15, have demonstrated greater potential in stimulating the activation of these lymphocytes. As already reviewed, IL-15 associated with pAgs promotes high cytotoxicity in γδ T cells, which is evidenced by the high T-bet expression (181). In addition, the combined use of IL-2 + IL-15 can provide γδ T cells with antileukemic properties (174, 181, 182) even in hypoxia (181).
A mix of cytokines combined with Abs can also promote a pre-activated state in γδ T cells, as evidenced in studies using the DOT protocol. Notably, the use of IFN-γ, IL-1β, IL-4, IL-15, and IL-21 with anti-CD3 Abs positively regulates many NKRs, while ICs, such as PD1, CTLA4 and CD94/NK group 2 member A (NKG2A), are negatively regulated on the cell surface (31, 200). In addition, many homing receptors, such as signal-regulatory protein alpha (SIRPα), integrin-β7, CD31, CD56, CD96 and intercellular adhesion molecule 1 (ICAM-1), are expressed, as well as chemokine receptors, such as CXCR3, CCR6 and CX3C chemokine receptor 1 (CX3CR1) (31). Noteworthy, the junction of these cytokines promotes γδ T cells with APC functions and a higher potential to migrate and recirculate between blood and tissues (31, 174). Therefore, pre-activation using these approaches may lead to better crosstalk with other cytotoxic cells (e.g., NK) or LCs in different compartments (265).
The HSCT Questions
The functional importance of γδ T cells in HSCT has received enormous attention after many years of research. The fact is that the frequency of these lymphocytes may fluctuate between treated and untreated individuals, either during chemotherapy (27, 266) or after HSCT (267–273), implying relevant roles for γδ T cells in the patient’s recovery (274). Several initial reports have shown that αβ TCR depleted allogeneic HSTC (allo-HSCT) was able to increase disease-free survival (2-5 years) after transplantation (267, 268, 273). Notably, this was correlated with a high γδ T cell frequency circulating in the PB and mediating the graft-vs-leukemia (GvL) effect (267). The Vδ1 subtype represented the highest proportion of these cells in the blood of patients (267, 273, 275).
Given that γδ TCRs are not restricted to HLA expression, the triggering of the GvHD effect is less likely, since tumor detection depends on more ubiquitous targets (273, 276). Therefore, the high frequency of these cells contributes to the restoration of the hematopoietic niche and is related to antileukemic responses (273); although this is not their only contribution to the success of HSCT. Higher γδ T cell percentages and a lower incidence of infection was been observed in many patients after HSCT, indicating protective roles in fungal, bacterial and viral infections (268, 273, 276). This made it possible to increase survival in patients with a high frequency of these cells when compared to patients with low or normal counts (277).
Cytomegalovirus (CMV) infection and its reactivation is a major concern after HSCT and, notably, γδ T cells can be essential effectors in controlling viral expansion. Knight et al. (278) reported for the first time that Vδ1 and Vδ3 cells expand as a result of an active response against CMV in patients after allo-HSCT; although there were previous data that showed that these subtypes expand in CMV infection in immunocompetent individuals (275, 276, 279). Interestingly, CMV reactivation after allo-HSCT mobilized these non-Vδ2 subtypes against infected cells and against LCs in vivo (280). This is intriguing and leads us to infer that the reactivation of CMV after HSCT can benefit patients with leukemia, as it impacts the incidence of disease recurrence (281).
Epstein-Barr virus (EBV) infection is also a problem. Farnaut et al. (282) showed that EBV infection resulted in a significant Vδ1 cell expansion in a patient with ALL transplanted with UCB, which represented more than 80% of the total circulating γδ T cells. One year after transplantation, these cells were highly differentiated and exhibit CD57 and CD8 expression while minimally expressing the BTLA protein (282). These data suggest a strongly adaptive response from Vδ1 and Vδ3 cells that possibly improves the efficacy of allografts (269).
Overall, the graft enriched with γδ T cells provides a lower relapse incidence during immune reconstitution after HSCT (274). This is evidenced when patients with low frequencies of these lymphocytes have a high rate of death from relapse (283). In addition, γδ T cell innate and adaptive responses can also prevent the occurrence of infections after HSCT (269, 284, 285). Finally, their functional plasticity can assist in immunological tolerance to the graft and avoid GvHD, as evidenced in many studies (258, 260). Therefore, the data highlighted here position γδ T cells as potential targets in applications aimed at improving clinical results after HSCT, since they induce a potent GvL effect in the absence of GvHD.