Recent therapeutic advances utilising immune-checkpoint inhibitors (ICI) have revolutionised the cancer immunotherapy field which was mainly facilitated by the introduction of monoclonal antibodies (mAb) targeting the T cell immune inhibitory molecules PD-1 and CTLA-4. Antibodies targeting these molecules have been used as single agents or in combination for the treatment of many cancers including previously untreatable solid tumours, such as end-stage melanoma (1). However, despite the wide success of ICI, these therapies are only efficacious in a subset of cancer patients (2). Therefore, novel and more effective complementary cancer immunotherapies are needed.

Several cellular therapeutic strategies have taken centre stage, including adoptive T cell and chimeric antigen receptor (CAR) T cell therapy (3). One particular population of T cells, namely the invariant natural killer T cells (iNKT) are the focus of intense research as they can deliver powerful anti-tumour responses (4, 5). iNKT cells are not restricted to polymorphic human leukocyte antigen (HLA) and bind exclusively to the monomorphic CD1d molecule and therefore do not cause graft versus host disease (6–8). Thus, there is an untapped potential for iNKT cells to serve as an off-the-shelf therapy (9–11). To realise the potential of iNKT cells, we must develop a better understanding of human iNKT cell biology to improve subsequent iNKT cell immunotherapies.

iNKT cells are an unconventional subset of αβ T cells that specifically recognise lipids and glycolipids presented by the monomorphic MHC-like molecule CD1d (12–14). As they leave the thymus into the periphery as fully primed and matured T cells, they are considered an ‘innate-like’ subset that can bridge the gap between innate and adaptive immunity (15). They are also described as attractive targets for cancer immunotherapy as they play important roles in tumour immunosurveillance and anti-tumour immunity (5, 11, 16, 17).

iNKT cells are T lymphocytes that display properties of both T cells and NK cells, as defined by specific cell surface markers such as NK1.1 and NKG2D (10, 18–20). Human iNKT cells express a semi-invariant TCR, with the α chain consisting of an invariant Vα24-Jα18 chain paired with a Vβ11 chain (21). In contrast, the chain pairing differs in mice where an invariant Vα14-Jα18 chain pairs with one of three β chains (Vβ8, Vβ7, or Vβ2) (22).

iNKT cells, similarly to conventional αβ T cells, begin their journey in the thymus where they undergo a rigorous selection process and subsequently migrate to peripheral tissues in a matured state able to perform effector functions without priming (15, 23). Human iNKT cells can be identified using mAbs (TCR Vα24 Jα18 clone 6B11), by reactivity to the prototypical synthetic iNKT agonist glycolipid α-galactosylceramide (αGalCer) and by binding to CD1d-αGalCer tetramers (24). Human iNKT cells can be divided into CD4^-CD8^- double negative (DN), CD4⁺, CD8⁺ or CD4⁺CD8⁺ double positive (DP) populations (25, 26). In contrast to humans, murine iNKT cells are more frequent and only occur as either CD4⁺ or DN T cells (27, 28).

In humans, iNKT cells are predominantly located within the thymus, liver, bone marrow, spleen, and peripheral blood (for more information about the distribution of iNKT cells in mouse and human tissues, we refer the reader to an excellent review by Crosby and Kronenberg (27)). Within human peripheral blood, iNKT cells make up 0.01%-0.2% of T cells with no differences between gender; however, they can range from undetectable to over 1% (25, 26). Even when iNKT cells are in their resting state they carry pre-formed mRNA enabling production of T helper 1 (Th1) and T helper 2 (Th2)-type cytokines (29). Upon recognition of CD1d, iNKT cells rapidly secrete copious immunomodulatory cytokines such as interferon gamma (IFN-γ), tumour necrosis factor alpha (TNF-α), interleukin (IL)-4, and IL-17 to instigate and influence downstream immune responses (29, 30). However, iNKT cells have also been shown to secrete IL-2, IL-5, IL-6, IL-10, IL-13, IL-21, TGF-β and GM-CSF, as well as several chemokines (31).

Activation of iNKT cells results in TCR downregulation, proliferation, and prolonged cytokine secretion (32, 33). The repertoire of Th1 and Th2 type cytokines produced by iNKT cells is modulated by the strength of the iNKT cell TCR signalling, as well as by the iNKT agonist and the type of antigen-presenting cells (APCs) presenting iNKT antigens (11, 15). Consequently, iNKT cells serve as a source of cytokines that activate and recruit other cell types including APCs early during immune responses, while activated APCs subsequently direct the ensuing adaptive immune response which help protect from infection and tumour growth (15, 34).

In addition to activation via iNKT cell TCR recognition of CD1d-lipid complexes, iNKT cells may also be activated by cytokines, such as IL-12 and IL-18 (27, 35). iNKT cells are some of the first responders during an immune response. Indeed, activated iNKT cells upregulate their IL-12 receptor and CD40L; through cross talk they induce the maturation of dendritic cells (DCs), and the subsequent production of IL-12 by the DCs (36). IL-12 secretion by DCs increases IFN-γ production by iNKT cells, leading to a positive feedback loop for Th1 immunity (10, 36). The maturation of DCs leads to increased production of IFN-γ by NK cells along with upregulation of MHC class I and II antigen presentation to T cells (37). This feature of iNKT cell biology to “jump-start” the ensuing innate and downstream adaptive immune response is central to exploiting iNKT cells in immunotherapy to promote anti-cancer immunity (10, 15, 37).

While the CD1d-iNKT interaction is relatively well studied in the mouse, the human iNKT system is less well understood. There have been several clinical trial efforts attempting to use the strong iNKT agonist lipid αGalCer to activate this subset of T cells or through using adoptive iNKT cell immunotherapy ( Table 1 ). However, plenty of evidence suggests the existence of a more diverse human iNKT repertoire, leading to investigations of specific subsets which has the potential to improve future iNKT-targeted cell therapies (38).

Mouse studies have demonstrated that iNKT cells exert powerful anti-tumour responses, however, translation into humans has proved difficult (11, 39). It is now unanimously accepted that iNKT cells play a pivotal role in anti-tumour immunity both in mice and humans (11). Primarily there are three mechanisms of action through which iNKT cells elicit a cytotoxic response: (i) direct tumour lysis, (ii) recruitment and activation of other innate and adaptive immune cells by initiating a Th1 cytokine cascade, and (iii) regulation of immunosuppressive cells in the tumour microenvironment (TME) ( Figure 1 ).

Seminal studies in the 1990s demonstrated that αGalCer was a potent activating ligand of mouse and human iNKT cells (13). αGalCer is presented by CD1d expressing APCs to selectively stimulate iNKT cells. However, iNKT cell activation requires costimulatory signals provided by CD40/CD40L and B7/CD28 interactions (13). Several studies have demonstrated that αGalCer exhibits anti-tumour properties against a variety of tumours, including B16 melanoma (68) and 3LL lung carcinoma (69). Therapeutic administration of αGalCer markedly reduces the number of B16 lung metastases in wild type (WT) C57BL/6 mice. This anti-tumour activity was completely abolished in Jα18^−/− (deficient in iNKT cells), CD1d^−/− (deficient in both NKT cell subsets) and RAG-1^−/− (lack mature B and T lymphocytes) mice, strongly suggesting that iNKT cells were responsible for the anti-tumour effects (20). Additionally, αGalCer treatment following chemotherapy (cisplatin) delayed tumour cell proliferation and increased tumour cell death in mesothelioma AB12 laden mice (70). In mice with liver metastases of adenocarcinoma Colon26 cells, administration of the αGalCer synthetic analogue KRN7000, one day after tumour inoculation significantly inhibited tumour growth in the liver with a potency similar to that of IL-12 (71). Even when treatment was given after nodule formation (day 3), tumour regression was observed (71). Anti-tumour activity of KRN7000 in mice with spontaneous liver metastases of reticulum cell sarcoma M5076 tumour cells suppressed the growth of established liver metastases and resulted in the prolongation of survival time (72). An increase in iNKT cell numbers and IL-12 production by hepatic lymphocytes was markedly enhanced in KRN7000-treated mice (72). Together, these results suggest that the in vivo anti-tumour effects of KRN7000 are dependent on iNKT cells and endogenous IL-12 production.

Adoptive transfer of αGalCer-loaded APCs was explored as an alternative approach to stimulate iNKT cells. DCs pulsed with αGalCer can effectively induce potent anti-tumour cytotoxicity by their specific activation of Vα14⁺ iNKT cells, resulting in the inhibition of tumour metastasis in vivo (73). Moreover, a complete inhibition of B16 melanoma metastasis in the liver was observed when αGalCer-pulsed DCs were injected, even seven days after transfer of tumour cells to syngeneic mice, when small but multiple metastatic nodules were already formed (73).

Mice with a deletion of the Trαj18 gene segment do not express the Vα14 Trαj18 TCR and were found to exclusively lack iNKT cells; while they maintained numbers of lymphocytes that were almost identical to WT mice (5). The resulting iNKT deficient mice were no longer able to mediate IL-12-induced rejection of B16 melanoma tumours (5). Adoptive transfer of IL-12-activated Vα14 iNKT cells prevented hepatic metastasis of B16 melanoma in mice (74). This suggested the involvement of direct cytotoxic mechanisms by iNKT cells rather than cytokine-mediated immune responses. Furthermore, adoptive transfer of iNKT cells into iNKT cell-deficient (Jα18^-/-) mice restored tumour surveillance and protected against methylcholanthrene-induced fibrosarcoma in the absence of exogenous stimulatory factors (4). However, the production of IFN-γ by iNKT and other lymphocytes was essential for protection. Using humanised NSG mice, a recent study showed that tumour localised administration of αGalCer can significantly enhance iNKT cell-mediated anti-tumour capacity against solid tumours (75).

The studies described above utilised the Jα18-deficient mice which were described in 1997 (5). Recently, a study revealed that about 60% of TCRα repertoire diversity was lacking in these mice due to the absence of Trαj gene segments upstream of Trαj18 (76). Subsequently this generated concern regarding the validity of the experimental conclusions, such as those of Toura et al., 1999 (73). The Jα18-deficient mouse strain was again called into question when a report showed that Jα18^−/− mice are also defective in mucosal-associated invariant T (MAIT) cells in both the thymus and peripheral organs (77). Therefore, caution is advised when interpreting data from the TCRJα18^−/− mouse strain. Consequently, the group that described the original Jα18^−/− mice generated a novel Trαj18-deficient mouse line by specifically targeting the Trαj18 gene segment. Apart from the absence of Trαj18, these mice had an undisturbed TCRα repertoire. Next generation sequencing detected normal generation of Vα19Jα33 expressing MAITs, whose development was abrogated in the originally described Jα18^−/− mice (78). Using a B16 melanoma liver metastasis model with mice bearing metastatic melanoma nodules in the liver, intravenous administration of αGalCer-pulsed DCs as described previously resulted in the complete eradication of melanoma metastasis in WT but not in Trαj18^-/^- mice (78). Indeed, the tumour growth in the DC-αGalCer treated Trαj18^-/- mice was similar to that in the vehicle-treated WT and Trαj18^-/- mouse groups, demonstrating the absolute requirement for activated iNKT cells in tumour rejection (78).

Taken together, there is now conclusive evidence from preclinical studies to suggests that iNKT cells exhibit powerful anti-tumour activity and are involved in cancer immunosurveillance (10, 16). They likely act as early warning systems to initiate an anti-tumour response which is subsequently performed by dedicated effectors such as NK cells and/or cytotoxic T lymphocytes (15). The potent iNKT cell-derived IFN-γ production and early activation of effector cells such as NK and CD8⁺ T cells suggests that iNKT cells can be rapidly stimulated by glycolipids on the tumour cell or by other stimuli. Effector cells then directly attack the tumour through direct perforin-dependent lysis and indirectly through raising an IFN-γ response (15). Early recognition of the tumour appears to be key, as delayed transfer of iNKT cells provides less protection (4).

Following promising results in preclinical models, many human clinical trials have begun to exploit iNKT cells to harness their anti-tumour potential. Initially, cancer patients with solid tumours were intravenously injected with soluble αGalCer. However, while αGalCer treatment was well tolerated, it failed to initiate an effective clinical response (79). As preclinical studies suggested that DCs loaded with αGalCer confers better immune responses in vivo (73), αGalCer-pulsed DCs were trialled in patients with advanced and recurrent non-small cell lung cancer, head and neck squamous cell carcinoma and myeloma (80–84). Different strategies were exploited to improve this treatment, which included utilising different types of APC or alternative routes of administration. Again, despite good tolerance and an increase in iNKT cell numbers, clinical benefits were limited (80, 81, 83–85).

Since adoptive T cell therapy for cancer treatment has long been established, several groups begun exploiting expanded iNKT cell products. Adoptive transfer of iNKT cells into patients with melanoma, head and neck squamous cell carcinoma, lung cancer and other solid tumours revealed increased iNKT numbers in vivo and increased levels of IFN-γ (86–88). While adoptive transfer was well tolerated, further improvements would be required to achieve a significant clinical response in patients (11).

Studies that exploited iNKT cells in CAR-T cell therapy showed promising results in murine preclinical studies (6, 89). CAR-iNKT cells exhibited significantly better in vivo responses than traditional CAR-T cells in mice when targeting GD2⁺ neuroblastomas and CD19⁺ lymphomas (6, 89). CAR-iNKT induced little graft versus host disease and their efficacy was augmented through their dual targeting ability of CD1d and CD19/GD2. Indeed, CAR-iNKT targeted both GD2⁺ neuroblastoma cells and CD1d⁺ TAMs. Additionally, CD1d⁺ CD19⁺ lymphoma cells have been targeted by CAR-iNKT cells in a dual pronged attack which effectively localised to the tumour site, had potent anti-tumour activity, and significantly improved the long-term survival of treated mice (89). While these studies indicated that iNKT cells are a highly efficient platform for CAR-based immunotherapy in mice, clinical trials are now ongoing for patients with B cell lymphoma, leukaemia and glioblastoma, and promising initial results were reported in humans (90). Therefore, these translational studies open a promising avenue for iNKT targeting cancer therapies in patients, but they lack clinical efficacy at present (91). iNKT immunotherapy will require further improvements to achieve effective clinical outcomes.

A key challenge for iNKT cell cancer immunotherapy in humans is a defective iNKT cell repertoire in human cancer patients both quantitatively and qualitatively. Several studies have demonstrated reduced iNKT cell numbers in the peripheral blood of cancer patients (39, 79, 92). Furthermore, iNKT cells derived from cancer patients release reduced levels of IFN-γ as they tend to exhibit a Th2 phenotype and CD1d expression can be downregulated in tumours, which abrogates the efficacy of the direct iNKT cell-mediated immune response (46). Several mechanisms have been postulated that may explain the suboptimal efficacies of iNKT cell anti-tumour response in clinical trials (9, 11), such as the induction of iNKT cell anergy after αGalCer treatment, the secretion of both Th1 and Th2 cytokines by iNKT cells and immune suppression in the TME. There are several limitations and obstacles to the clinical translation of iNKT cell therapy into humans, and many strategies have been proposed to overcome these limitations (which have been reviewed extensively elsewhere) (9, 11, 90, 91, 93). These strategies include alternative vectors for the delivery of αGalCer (83, 94), the generation of more potent αGalCer analogues and other iNKT cell agonists (95), the generation of induced pluripotent stem cell-derived iNKT cells and improvements to iNKT-CAR based platforms (90, 96).

A further potential strategy proposed here is the selection and utilisation of specific human iNKT cell subsets to achieve improved outcomes in iNKT-based cancer immunotherapy. This proposal is based on studies revealing that iNKT cells exist as subpopulations with a previously unrecognised diversity in function (19, 25, 38, 97). iNKT cells are not a single, uniform class of T cells as they exhibit heterogeneity in both phenotype and function. In mice, iNKT cells seemingly exhibit at least three distinct thymic populations based on the expression of unique sets of transcription factors, namely iNKT1 (T-bet⁺), iNKT2 (GATA-3⁺) and iNKT17 (RORγt⁺) (98). While iNKT1 produce IFN-γ and also some IL-4, iNKT2 produce IL-4 and iNKT17 produce IL-17. BALB/c mice have large proportions of iNKT2/iNKT17 cells but reduced proportions of iNKT1 cells. C57BL/6 mice were highly enriched for iNKT1 but no other subsets, and NOD mice have equal proportions of all three subsets, revealing inter-strain variability (98). Further described subsets are iNKT_FH or follicular helper iNKT that provide cognate help for B cells (99, 100), and iNKT10 that play important roles in maintaining adipose tissue homeostasis (101).

Although most mouse iNKT cells express the canonical Vα14-Jα18 TCR α-chain, they can use different Vβ chains. Combinations of Vβ-, Jβ-, and CDR3β-encoded residues will ultimately determine the type of ligands that iNKT cells recognise (102, 103). Basal activation, proliferation, TCRβ repertoire and antigen specificity are seemingly modulated by their anatomical location (104). Thus, the so-called invariant NKT cell population expresses a variable TCRVβ repertoire that differs in antigen recognition in individual tissues. Importantly, anatomical differences also apply to human iNKT cells. For example, Jimeno et al. found increased frequencies of atypical iNKT cells (Vα24- or Vβ11-) in tonsils vs. blood, while the frequency of CD4⁺ and CD69⁺ iNKT cells was also different in those anatomical locations (104). The diverse iNKT subpopulations thus occupy and utilise unique anatomical and physiological niches to perform their diverse biological functions, likely to be context and tissue specific.

Mature human iNKT cells can be categorised according to numerous characteristics ( Figure 2 ). Broadly, human iNKT cells are characterised by CD4 expression, as they exist as CD4⁺, and CD4^- subsets (25). CD4^- iNKTs are DN or CD8⁺with a very small subset of DP iNKT cells also identified (26, 107). The relative frequencies vary substantially between individuals: on average, CD4⁺ and DN cells are the most frequent subsets, and CD8⁺ iNKT cells have a low frequency (26, 108). There is evidence suggesting that DN human iNKT cells are different from their CD4⁺ iNKT cell counterparts (25, 107). DN iNKT cells are seemingly similar to mouse iNKT1 cells, exhibiting an increased IFN-γ secretion and cytotoxic function when activated (25, 107).

Although iNKT cell subsets share the expression profile for several chemokine receptors, they may also differ with respect to their chemokine receptor expression ( Figure 2A ). CD4⁺ iNKT cells predominantly express CCR4, while CD8 and DN iNKT cell subsets mainly express CCR1, CCR6 and CXCR6 (108). Based on these observations, a differential tissue distribution can be assumed; for example, CXCR6 plays a role in the homeostatic distribution of iNKT cells to the liver and the lung (109, 110). Human CD4⁺ iNKT cells are broadly associated with Th0-type immune responses and are the exclusive producers of IL-4 and IL-13 upon primary stimulation; whereas, DN iNKT cells have a strict Th1 profile and prominently express several NK lineage receptors (107). Additionally, it has been shown that CD244⁺ CXCR6⁺ iNKT cells, which are present in mice and humans, have enhanced cytotoxic properties producing more IFN-γ compared to CD244^- CXCR6^- iNKT cells (38). Intriguingly, this CD244⁺ CXCR6⁺ iNKT cell subset is CD4^-, potentially explaining the enhanced cytotoxic function of the CD4^- iNKT cell subset (38).

In a B cell lymphoma model, Tian and colleagues demonstrated that CD19-specific CAR-iNKT cells expressing CD62L mediated tumour regression (111). CD62L is involved in homing of naïve and central memory T cells to secondary lymphoid organs. The Metelitsa group also showed that these CD62L⁺ iNKT cells have prolonged persistence and anti-tumour activity in vivo (111). Subsequently, IL-21 was demonstrated to preserve the crucial central memory-like iNKT subset and enhance anti-tumour effector functionality. Following antigenic stimulation with αGalCer, CD62L⁺ iNKT cells both expressed IL-21R and secreted IL-21, each at significantly higher levels than CD62L^- cells (111). Although IL-21 alone failed to expand stimulated iNKT cells, combined IL-2/IL-21 treatment produced more iNKT cells and increased the frequency of CD62L⁺ cells versus IL-2 alone. Gene expression analysis of CD62L⁺ and CD62L^- cells revealed that treatment with a combination of IL-2 and IL-21 (but not IL-2 alone) selectively downregulated the proapoptotic protein BIM in CD62L⁺ iNKT cells, thus protecting them from activation-induced cell death (111, 112). While these studies have been conducted in mice, evidence suggests that human iNKT cells can also express CD62L (113). It would be interesting therefore to evaluate the anti-tumour activity and persistence of human CD62L⁺ iNKT cells in future studies.

Based on their cytokine secretion pattern, human iNKT cells have also been classified into Th-cell subsets ( Figure 2B ). Th1-like iNKT cells have been identified in healthy individuals producing Th1-associated cytokines such as IFN-γ and TNF-α upon stimulation (25, 107, 114). These Th1-like iNKT cells are predominantly DN cells and express more NKG2D compared to CD4⁺ iNKT cells. In contrast, Th2-like iNKT cell subsets with regulatory properties tend to be CD4⁺ and they secrete IL-4, IL-13, and IFN-γ upon activation (25, 107). Human Th17-like iNKT cells secrete the proinflammatory cytokines IL-17, IL-21 and IL-22 when activated (115). Finally, FOXP3-expressing Treg-like iNKT cells that secrete the immunosuppressive cytokine IL-10, as well as T_FH-like iNKT cells that secrete IL-21 upon activation have also been described (115, 116). However, iNKT cells exhibit plasticity in terms of their cytokine production and therefore the above-described definitions are not absolute.

Diversity within human and mouse iNKT cells is further observed based on the TCRβ sequence of the iNKT cell TCR. In mice, the iNKT repertoire displays clonal heterogeneity regarding lipid recognition, such as the αGalCer analogue OCH, which can be explained by the differential Vβ family usage in mouse iNKT cell TCRs (22, 117, 118). In contrast, the human iNKT cell TCR is composed of the invariant Vα24 TCRα chain and the semi-variant Vβ11 TCRβ chain, whereby the CDR3β region is the only truly adaptive element ( Figure 2C ). Thus, human iNKT cells in healthy adults express TCRs with widely variable affinities for CD1d, which are directly dependent on structural differences within the CDR3β loop of the iNKT cell TCR (97). Clonal variations resulted in up to a 40-fold difference in TCR affinity to CD1d and could be directly correlated to iNKT cell function (97). iNKT clones bearing high affinity iNKT cell TCRs proliferated more avidly and produced a greater diversity of cytokines in response to CD1d expressing APCs compared to clones expressing a low affinity iNKT cell TCR (97). iNKT cell TCR affinities are seemingly skewed in rheumatoid arthritis (119) and type 1 diabetes (120) which may also contribute to disease pathology in the context of human cancer. A recent study identified three CDR3β amino acid sequence motifs that were seemingly associated with strong autoreactivity: a VD region with two or more acidic amino acids; usage of the Jβ2-5 allele; and a 13 amino acid long CDR3β motif (105). Acidic amino acid composition, J usage, and the CDR3β region amino acid length individually affected the affinities of conventional TCRs (121–123). Additionally, the Hirano group revealed that the presence of a naturally encoded arginine (Arg) or a serine (Ser) in the third position of the CDR3β loop, can effectively modulate antigen recognition by the iNKT TCR (124). In agreement with previous studies, Ser to Arg substitutions influence the structure of the CDR3α loop thus effecting the iNKT cell TCR affinity (124). Furthermore, such differences in iNKT cell TCR affinities also influence iNKT cell function (124).

Studies in mice have revealed that iNKT cells can be exploited as a powerful platform for cancer treatment (4, 5). However, there are limitations to such studies which makes it difficult to reconcile data for an effective clinical translation of iNKT cell immunotherapy from mice to humans. Laboratory mice are inbred strains that lack genetic heterogeneity (125), and therefore conclusions from pre-clinical models might be exaggerated as it does not take into account the wide genetic variability of the human population. This is complicated by the fact that there are far more mouse studies conducted on iNKT cells and therefore, more studies investigating human iNKT cells are urgently needed. Although there are similarities between iNKT cells derived from mice and humans, there are also some key differences that should be taken into consideration when designing studies for clinical translation into humans. Mice have far more iNKT cells than humans and many studies have shown that targeting iNKT cells in mice induce powerful Th1 mediated anti-tumour immunity (4, 5). However, it is well known that iNKT cells are less frequent in healthy humans (126). Furthermore, cancer patients, particularly those with late stage disease, have reduced iNKT cell numbers and those that are present are often defective in their cytokine secretion, suggesting that cancer patients exhibit a profoundly immunocompromised iNKT cell repertoire (39, 49, 127, 128). It is likely that targeting an already exhausted and unresponsive iNKT cell repertoire in these patients may not lead to potent anti-tumour immune responses, potentially explaining the disappointing results in human iNKT cell-based clinical trials. Therefore, a better understanding of the human iNKT cell repertoire in healthy individuals and in cancer patients is needed to fully characterise human iNKT cell heterogeneity and its contribution in health and disease.

In addition, future iNKT cell targeting therapies should consider exploiting the specific iNKT cell subpopulations with respect to their diverse co-receptor expression phenotypes and TCR affinities. We propose that a better understanding of the mechanisms driving anti-tumour functions of diverse human iNKT cell subpopulations will achieve significant advances in their exploitation in cancer therapies which may lead to better clinical efficacy. Emerging multimodal approaches such as single cell RNA and TCR sequencing will be needed to determine the relevance of specific iNKT cell subsets and clonotypes, which will require better in vivo models that can accurately reflect human disease (129, 130). The combination of TCR sequencing data with the transcriptomic information of single antigen specific T cells has had a significant impact on understanding the heterogeneity of T cell populations (131–133). T cells that express the same TCR αβ undergo clonal expansions and can give rise to progeny with diverse functional phenotypes. Sequencing of the TCR repertoire can highlight clonal diversity and the dynamics of antigen specific responses associated with the anti-tumour response (134, 135). On the other hand, the transcriptomic information can reveal distinct functional phenotypes associated with better disease outcomes. Therefore, coupling these datasets together can reveal distinct T cell populations that are most relevant to disease states (134, 135). Indeed, the application of these technologies have advanced our knowledge of CD8⁺ and CD4⁺ T cell heterogeneity in response to Yellow Fever Virus vaccination in humans (132), peanut allergy (136), and in colorectal cancer (133). Advances in sequencing technologies in recent years allows for T cell clones to be tracked and monitored to assess their behaviour during infection or during anti-tumour responses which enables the tracing of the progeny of individual T cells back to their ancestors (131, 137–139). Fate mapping and lineage tracing can therefore be applied to iNKT cell populations to better understand iNKT cell population dynamics, clonal heterogeneity with regards to CDR3β diversity to highlight subpopulations of iNKT cells that are disease relevant. Furthermore, T cell barcoding (131) which allows linage tracing, can be applied to iNKT cell populations to assess inter-clonal differences and to monitor the contribution of individual clonal populations to effector functions during anti-tumour responses. Fate mapping and lineage tracing of iNKT cells may uncover inter-clonal diversity and could provide data on clonal contribution to the anti-tumour immune response. The application of these new technologies to human iNKT cells will provide high dimensional data that will allow a better understanding of human iNKT cell subpopulations in disease states, thus furthering our understanding of the diversity and complexity of the iNKT cell response in humans. Furthermore, it is imperative that such studies should be conducted in reliable preclinical humanised mouse models of cancer that display the human CD1d antigen presentation system, complemented by human iNKT cells (140, 141). Investigating the anti-tumour efficacy of diverse human iNKT cell subpopulations in these models and through utilising human cancer cells will likely provide a deeper understanding of human iNKT cell subsets in preclinical studies. In summary, exploiting human iNKT cells complement current cancer immunotherapies and as such knowledge of human iNKT cell subpopulations in the context of tumour immunology is urgently required to help improve the translational efficacy of future iNKT cell based immunotherapies.

AL, DB, and SM developed the concept and wrote the manuscript. AL, generated all the figures and tables. All authors contributed to the article and approved the submitted version.