Invariant natural killer T cells in hematopoietic stem cell transplantation: killer choice for natural suppression

Abstract

Invariant natural killer T cells (iNKTs) are innate-like lipid-reactive T lymphocytes that express an invariant T-cell receptor (TCR). Following engagement of the iTCR, iNKTs rapidly secrete copious amounts of Th1 and Th2 cytokines and promote the functions of several immune cells including NK, T, B and dendritic cells. Accordingly, iNKTs bridge the innate and adaptive immune responses and modulate susceptibility to autoimmunity, infection, allergy and cancer. Allogeneic hematopoietic stem cell transplantation (HSCT) is one of the most effective treatments for patients with hematologic malignancies. However, the beneficial graft versus leukemia (GvL) effect mediated by the conventional T cells contained within the allograft is often hampered by the concurrent occurrence of graft versus host disease (GvHD). Thus, developing strategies that can dissociate GvHD from GvL remain clinically challenging. Several preclinical and clinical studies demonstrate that iNKTs significantly attenuate GvHD without abrogating the GvL effect. Besides preserving the GvL activity of the donor graft, iNKTs themselves exert antitumor immune responses via direct and indirect mechanisms. Herein, we review the various mechanisms by which iNKTs provide antitumor immunity and discuss their roles in GvHD suppression. We also highlight the opportunities and obstacles in manipulating iNKTs for use in the cellular therapy of hematologic malignancies.

INTRODUCTION

Invariant natural killer T cells (iNKTs) or type I NKTs are unique innate-like T lymphocytes that express an invariant T-cell receptor (TCR)-α chain, Vα14-Jα18, which often pairs with Vβ8.2, Vβ7 or Vβ2 (Vα24- Jα18, Vβ11 in humans).^1,2 This iTCR confers reactivity to glycolipid antigens (Ags), such as the prototypical iNKT agonist, α-galactosylceramide (αGC).³ The less well-characterized, type II NKTs are also CD1d-restricted but exhibit diverse TCRαβ chain usage and recognize sulfatide Ags.^1,4,5 The salient features of human and murine iNKTs and the differences between type I and type II NKTs are summarized in Tables 1 and 2, respectively.^1,6–11 iNKTs develop in the thymus from CD1d-restricted CD4⁺CD8⁺DP thymocytes that progress through four different stages of maturation: CD24^hiCD44^loNK1.1⁻ (Stage 0), CD24^loCD44^loNK1.1⁻ (Stage 1), CD24^loCD44^hiNK1.1⁻ (Stage 2), and finally CD24^loCD44^hiNK1.1⁺ (Stage 3).¹⁰ As they go through these developmental stages, iNKTs upregulate NK cell markers (for example, NKG2D and Ly49 receptors), CD69 and CD122, and acquire distinct effector functions¹² that are tightly regulated by several transcription factors, signaling molecules, surface receptors and cytokines.^1,11,13

Table 1.

Characteristics of mouse and human invariant NKT cells

Features	Mouse	Human
Receptors and ligands
NK cell markers	NK1.1− (immature and mature) NK1.1+ (mature)	CD161− (immature) CD161+ (mature)
TCR αβ chain	Vα14 Jα18 Vβ8.2, Vβ7, Vβ2	Vα24 Jα18 Vβ11
CD1d restricted	Yes	Yes
CD4/CD8 expression	CD4+ or CD4−CD8− (DN)	CD4+, CD4− (DN or CD8+)
αGC reactivity	Yes	Yes
Selecting ligand	Controversial	Controversial
Tissue distribution
Blood	0.2–0.5%	0.008–1.176%
Thymus	~0.5%	<0.1%
Liver	20–30%	~1%
Spleen	~1%	Unknown
Functions
Cytokine production	No clear distinction in cytokine production between CD4+ and DN subsets	CD4+: Th2 (IL-4) CD4−: Th1 (IFN-γ)
Cytotoxicity	CD4+ and DN cells are equally cytotoxic against CD1d+ tumors; DN cells more effective in controlling CD1d− tumors	CD4− cells more cytotoxic than CD4+ subset

Abbreviations: αGC = alpha-galactosylceramide; DN=double negative; IFN-γ =Interferon-γ; IL-4 =interleukin-4; NK =natural killer; NKT =natural killer T cell; TCR =T-cell receptor. From references 1–11,20,24,33,52 and 81–83.

Table 2.

Differences and similarities between conventional T and NKT cells

Features	αβ T	NKT cells
		Type I	Type II
TCR repertoire	Diverse	Vα14 Jα18, Vβ8.2, Vβ7, Vβ2	Diverse, some have Vα3.2 Jα9, Vα8 Jα9, Vβ8
MHC-restriction	MHC I or MHC II	CD1d	CD1d
Selecting cells	TEC	DP	DP
TCR ligands	Peptide antigens	Glycolipids (αGC, β-GlcCer, β-ManCer), diacylglycerol, phosphatidylinositol mannoside	Sulfatide, lyso-sulfatide
Positive selection	Yes	Yes	Yes
CD4/CD8 expression	CD4+ or CD8+	CD4+ or DN	CD4+ or DN
Activated Phenotype	After Ag exposure	Yes	Yes
SLAM-SAP dependent	No	Yes	Yes
PLZF expression	No	Yes	Yes
αGC reactivity	No	Yes	No
Antitumor response	Direct cytotoxicity perforin, Fas-FasL or TRAIL-mediated	Direct cytotoxicity: (perforin and Fas–FasL) indirect cytotoxicity: (IFN-γ, transactivation of NK, CD8+T and DC)	Suppressive (Th2 cytokines), counter-regulate type I antitumor activities
Role in GvHD	Donor T cells mediate and Tregs suppress GvHD	Ameliorate GvHD via IL-4 production and Treg expansion	Attenuate GvHD via both IFN-γ and IL-4 secretion

Abbreviations: αGC =alpha-galactosylceramide; β-GlcCer= β-glycosylceramide; β-ManCer=β-mannosylceramide; DC=dendritic cell; DN=double negative; DP=double positive; IL-4 =interleukin-4; IFN-γ = Interferon-γ; NK=natural killer; NKT =natural killer T; PLZF=promyelocytic zinc finger; SAP =SLAM-associated protein; SLAM =signaling lymphocytic activation molecule; TCR=T-cell receptor; TEC =thymic epithelial cell; TRAIL = TNF-related apoptosis-inducing ligand; Treg = regulatory T cell. Information presented in this table are for murine NKT cells and are from references 1–5, 10–12, 43, 44, 71, 73, 74 and 76.

Following iTCR engagement, iNKTs rapidly secrete cytokines and upregulate co-stimulatory receptors, which activate other immune cells including dendritic cells (DC), macrophages, NK, B and T cells.¹⁰ As a result, activation of iNKTs modulates an array of normal and pathogenic immune responses. Allogeneic hematopoietic stem cell transplantation (HSCT) is a potent curative treatment that is widely used for patients with relapsed or refractory hematological malignancies.¹⁴ The donor T cells in the allograft target the leukemia cells to exert a beneficial graft versus leukemia (GvL) effect.¹⁴ However, dysregulated activation and proliferation of donor T cells in the allograft leads to immune-mediated destruction of host tissues resulting in graft versus host disease (GvHD), a serious complication of allogeneic HSCT.¹⁴ Several studies demonstrate that iNKTs significantly attenuate GvHD,^15–18 whereas preserving the GvL effect.^19–22 In addition, iNKTs themselves mediate anti-leukemia activity via direct and indirect mechanisms²³ and also regulate antiviral, -bacterial and -fungal immune responses.²⁴ Together, these functional properties of iNKTs make them ideal candidates for use in cancer therapy. In this review, we discuss the antitumor activities of iNKTs and their roles in GvHD attenuation.

iNKTS AND ANTITUMOR IMMUNITY

iNKTs protect against tumors

Several studies demonstrate that iNKTs mediate protection from tumors. In mice heterozygous for mutations in the tumor suppressor p53, loss of iNKTs enhances susceptibility to tumors.²⁵ Moreover, treatment of iNKT-deficient CD1d^−/− and Jα18^−/− mice with a carcinogen results in increased incidence and earlier onset of tumors when compared with similarly treated wild-type mice.²⁶ Conversely, reconstitution of iNKTs into Jα18^−/− mice prevents the growth of chemically induced sarcomas.²⁷ Furthermore, treatment with αGC can attenuate the growth of adoptively transferred,^28,29 carcinogen-induced,^27,30 or spontaneous tumors³¹ in mice, largely via an interferon-γ (IFN-γ)-dependent manner. Interestingly, the recently described novel agonist of human and murine iNKTs, β-mannosylceramide, confers protection against tumors via a different mechanism involving production of nitric oxide and tumor necrosis factor-α.³² Recent studies also show that iNKTs protect against B-cell lymphomas in mice³³ and that infection of humanized mice with EBV promotes the generation of cytotoxic iNKTs, which produce IFN-γ and enhance T-cell killing of EBV-positive human tumor cells.³⁴ Finally, generation of induced pluripotent stem cells (iPSCs) from mature iNKTs can activate and expand Ag-specific CD8⁺T cell responses to provide protection against leukemia in mice.³⁵

Data on the antitumor functions of human iNKTs are more indirect. Patients with various types of cancer often exhibit reduced numbers of peripheral blood iNKTs and these cells are often impaired in their functions.^36–38 For instance, patients with progressive multiple myeloma have decreased frequency of iNKTs and marked deficiency in αGC-dependent IFN-γ production.³⁹ Conversely, in patients with neuroblastoma,⁴⁰ colorectal⁴¹ and head and neck carcinoma,⁴² increased numbers of peripheral blood and/or tumor-infiltrating iNKTs are associated with a more favorable response to therapy. Collectively, these data reveal an important role for iNKTs in host immunity against various cancers. In contrast, type II NKTs (Table 2) not only downregulate immune surveillance against tumors and facilitate their growth, but they are also known to counter-regulate the antitumor activity of iNKTs.^43,44

Antitumor mechanisms of iNKTs

iNKTs mediate their antitumor activity via multiple mechanisms. Their dominant mode of action involves the activation of other cytolytic effectors such as CD8⁺T and NK cells.^23,45,46 Indeed, αGC or cytokine-stimulated iNKTs robustly produce IFN-γ and upregulate the expression of CD40 ligand. As a result, they promote DC activation and enhance DC-mediated priming of tumor-specific CD4⁺ and CD8⁺T-cell responses.⁴⁷ iNKT-DC interactions also stimulate DC production of IL-12, which serves to further augment NK- and CD8⁺T cell lysis of tumors²³ (Figure 1a).

Figure 1.

Antitumor mechanisms of iNKTs. (a) Indirect mechanism of iNKT cytotoxicity. The cross-talk between iNKTs and Ag-presenting cells (such as DCs), presenting a tumor-derived glycolipid, leads to the activation of iNKTs, IFN-γ and CD40 stimulation. This iNKT-derived IFN-γ induces DC production of IL-12, which further augments IFN-γ production by iNKT and NK cells, and serves to stimulate CD8⁺ T cell- and NK cell-dependent killing of tumor cells. (b) iNKTs recognize glycolipid Ags presented by CD1d on tumor cells and mount direct cytotoxicity via perforin/granzyme exocytosis or Fas–Fas ligand (Fas L) interactions. (c) iNKTs can also limit tumor growth via their interactions with immunosuppressive cells that promote tumor growth such as TAMs and IL-10-producing neutrophils. Although iNKTs can directly kill TAMs, they alleviate the suppressive effect of the neutrophils via CD40–CD40L interactions.

In addition to their immune-stimulatory functions, iNKTs also function as cytotoxic effectors (Figure 1b). Consistent with this notion, mature iNKTs basally express cytolytic proteins (perforin and granzymes)^12,48,49 and can be induced to upregulate death-promoting molecules such as Fas ligand and TRAIL.^49–51 Others and we observe that human and murine iNKTs mount potent cyotoxic responses to numerous CD1d⁺ tumors in vitro and in vivo.^52–55 Furthermore, our studies demonstrate that iNKT-mediated cytotoxicity is critically dependent on CD1d-mediated lipid Ag presentation, functional TCR signaling, the adaptor protein signaling lymphocytic activation molecule-associated protein (SAP) and the tyrosine kinase Fyn (SAP-binding protein) as well as perforin expression; as loss or interference with any of these factors significantly reduced human and murine iNKT antitumor responses.^52,55 As several hematological malignancies (including AML, juvenile myelomonocytic leukemia, B-cell CLL, pediatric ALL, non-Hodgkin lymphoma) express CD1d,⁵⁶ they serve as potential targets for direct iNKT-cell recognition. However, many normal and transformed cells do not express CD1d and are, therefore, not targeted by such iNKT TCR-dependent mechanisms. In this context, recent studies show that human iNKTs kill target cells expressing NKG2D ligands in an NKG2D-dependent and CD1d-independent manner.⁵⁷ Furthermore, engagement of NKG2D promotes iNKT-cell activation in response to weak TCR agonists, suggesting that NKG2D also functions as a co-stimulatory receptor in these cells.⁵⁷

iNKTs can also impede tumor growth by killing or inhibiting immunosuppressive cells within the tumor microenvironment (TME) that facilitate tumor growth, such as tumor-associated macrophages (TAMs) (Figure 1c).^58,59 Studies have shown that hypoxic signaling within the TME results in sustained activation of TAMs and increased IL-6 production that favors tumor progression.⁶⁰ Although iNKTs colocalize with IL-6-producing TAMs, the hypoxic conditions within the TME inhibit iNKT activation.⁵⁸ Interestingly, recent studies show that IL-15 protects iNKTs from hypoxia, such that iNKTs can directly kill the tumor-associated macrophages in a CD1d-dependent manner and provide antitumor immunity.⁵⁹ Other potential targets of iNKTs within the TME include myeloid derived suppressor cells (MDSCs)⁶¹ and IL-10 producing neutrophils.⁶² iNKTs directly interact with these cells in a CD1d and CD40-dependent manner to reverse their suppressive phenotype and restore specific antiviral⁶¹ or antitumor responses⁶² (Figure 1c).

iNKT-based cancer immunotherapy

Several clinical trials have examined whether administration of α GC⁶³ or αGC-loaded DCs with⁶⁴ or without iNKTs^65–67 might prove beneficial in the treatment of cancer. These studies demonstrated that iNKT therapies are well tolerated and induce an objective clinical response in a subset of patients. As discussed above, maximal tumor-directed iNKT responses require tumor cell expression of CD1d. However, many tumors downregulate CD1d and thus evade iNKT recognition.⁶⁸ Two recent studies demonstrated that iNKTs can mediate antitumor activity in a tumor Ag-specific yet CD1d-independent manner.^69,70 In the first study, systemic administration of αGC-loaded soluble CD1d fused with anti-HER2 single-chain Ab Fv fragment significantly reduced lung metastasis of HER2-expressing B16 melanoma cells.⁶⁹ This antitumor activity of the CD1d–anti-HER2 fusion protein was associated with HER2-specific tumor localization and accumulation of iNKT, NK and T cells at the tumor site. In the second study,⁷⁰ primary human iNKTs were engineered to express a chimeric Ag receptor (CAR) against GD2, a disialoganglioside that is highly expressed by neuroblastoma cells. These CAR.GD2 iNKTs localized at the tumor site and exhibited robust antitumor activity against neuroblastoma cells.⁷⁰ Collectively, these studies exemplify how the antitumor activities of iNKTs can be harnessed for clinical application to treat a wider array of cancers that are currently difficult to cure. Finally, as activated iNKTs promote the antitumor functions of NK, T and B cells, future clinical trials involving the co-transfer of tumor-targeted conventional T cells and iNKTs may induce better clinical responses than those obtained using T cells alone.

ROLE OF iNKTS IN GvHD

iNKTs ameliorate GvHD in preclinical models

Non-myeloablative host-conditioning markedly reduces early toxicity post transplantation as compared with the myeloablative regimens; however, acute and chronic GvHD remain a significant clinical challenge for both these approaches.⁷¹ Early studies demonstrated that non-myeloablative host-conditioning with fractionated total lymphoid irradiation (TLI) and anti-thymocyte globulin (ATG) prior to transfer of allogeneic bone marrow (BM) cells and splenocytes attenuates GvHD via induction of regulatory ‘natural suppressor’ T cells in the host, which were later identified as iNKTs.¹⁶ This increase in host iNKTs is associated with elevated IL-4 secretion and protection against GvHD.¹⁶ Consistent with the IL-4-dependent protective role of host iNKTs, IL-4^−/−, CD1d^−/− and Jα18^−/− transplant recipient mice rapidly succumb to GvHD.^16–18,72 Furthermore, although host iNKTs promote donor chimerism,^16,72 they polarize donor T cells toward a Th2 cytokine pattern and inhibit their early expansion and infiltration into GvHD target organs⁷² (Figure 2). Importantly, these cells retain the GvL activity of the graft, which is dependent upon donor CD8⁺ T cells and their production of perforin.¹⁹

Figure 2.

iNKTs protect from GvHD but retain the GvL activity. Reduced intensity host-conditioning with TLI and ATG allows host iNKT expansion. Both host and donor iNKTs interact with APCs via CD1d to produce IL-4 that in turn skews donor T cells toward a Th2 cytokine bias. Both host and donor iNKTs also promote nTreg expansion in an IL-4-dependent manner that can inhibit donor T-cell proliferation and migration to GvHD target organs. Recently described, third-party CD4⁺ iNKTs also inhibit T-cell proliferation, promote Th2-biased cytokine response as well as expansion of donor MDSCs. These donor MDSCs are crucial for nTreg expansion and protection from GvHD lethality. Importantly, iNKTs (host, donor or third party) do not abrogate donor T-cell anti-leukemia activity. Furthermore, iNKTs can also directly kill the leukemia cells and contribute to the GvL effect. Taken together, attenuation of GvHD without loss of the GvL activity decreases tumor burden and improves overall survival of the host.

Donor iNKTs also attenuate GvHD in an IL-4-dependent manner.¹⁵ Recent elegant studies using luciferase-expressing donor T and iNKTs lend insights into the migration, proliferation and suppressive effect of donor CD4⁺iNKTs in a MHC mismatch preclinical model of GvHD.^20,22 In these studies, donor CD4⁺iNKTs robustly expanded in secondary lymphoid organs and migrated to GvHD target organs, as did the donor T cells.²⁰ However, adoptive transfer of low numbers of donor CD4⁺ iNKTs ameliorated GvHD pathology and prolonged survival by inhibiting donor T-cell proliferation and activation, promoting a Th2-biased cytokine pattern²² and significantly downregulating IFN-γ and TNF-α production by donor T cells.^20,22 Consistent with prior studies,¹⁵ donor CD4⁺ iNKTs attenuated GvHD in an IL-4-dependent manner, without abrogating the GvL effect.²² Other recent studies show that low numbers of ‘third party’ iNKTs also protect from lethal GvHD with the same effectiveness as donor CD4⁺iNKTs without abrogating the GvL effect.²¹ Type II NKTs also has a protective role in GvHD,^73,74 via production of both IFN-γ and IL-4; IFN-γ-producing BM type II NKTs induce Fas-dependent apoptosis of donor CD4⁺ and CD8⁺T cells, whereas the IL-4-producing cells skew the immune response toward a Th2 phenotype.⁷⁴

iNKTs mediate protection from GvHD via Tregs

Minimal intensity conditioning with TLI/ATG promotes IL-10-producing donor T_reg proliferation in wild type (WT) but not NKT cell-deficient hosts.¹⁸ Interestingly, adoptive transfer of WT but not IL-4^−/− NKTs into Jα18^−/− hosts restored Treg proliferation and protection from GvHD,¹⁸ suggesting that host iNKTs augment Treg expansion in an IL-4-dependent manner and this prevents donor T-cell expansion and induction of GvHD (Figure 2). Consistent with their protective role against GvHD, donor CD4⁺iNKTs promotes expansion of functional natural Tregs (nTregs) from the allograft.²² Depletion of Tregs from the graft is associated with abrogation of donor Treg expansion and loss of protection against GvHD, highlighting a critical role of donor Tregs in the regulation of GvHD pathogenesis.²² Besides Tregs, MDSCs also have immunoregulatory roles in allogeneic HSCT.⁷⁵ Interestingly, ‘third party’ CD4⁺iNKTs that protect against GvHD also promote expansion of both donor nTregs and MDSCs.²¹ In the same study, depletion of MDSCs abrogated donor nTreg expansion and protection from GvHD, suggesting that the cross-talk between MDSCs and nTregs is crucial for iNKT cell-mediated protection against GvHD.²¹ Although both iNKTs and Tregs have important regulatory roles in GvHD,^71,76,77 there are certain advantages of using iNKTs over Tregs in allogeneic HSCT as summarized in Table 3.

Table 3.

Advantages of using iNKTs versus Tregs in allogeneic HSCT

• Can be readily expanded from PBMCs or iPSCs to generate very large numbers for infusion during HSCT

• Easily identifiable by surface markers (iTCR, CD1d-tetramer reactivity) as opposed to Tregs that require intracellular staining for nuclear transcription factors (Foxp3 and Ikaros)

• MHC incompatibility not an issue owing to limited polymorphism of CD1d

• Very few iNKTs are required to mediate GvHD suppression; also persist longer than Tregs in vivo

• Both host and donor iNKTs can mediate protection from GvHD, whereas only donor but not host Tregs can do so

• iNKTs promote expansion of the few natural Tregs present in the allograft

• As CD1d is expressed on several hematological malignancies, they can be direct target for iNKT-cell recognition and cytotoxicity

• Although both iNKTs and Tregs can effectively separate the GvHD and GvL effects, only iNKTs have an inherent ability to contribute to the GvL effect via direct and indirect antitumor mechanisms

• iNKTs can provide protection from infections during the post-transplantation recovery period, as they can mediate antiviral, -fungal and -bacterial activities

• iNKTs contribute to hematopoiesis through secretion of GM-CSF and IL-3, as well as via direct recognition of CD1d expressed on hematopoietic progenitors

Abbreviations: HSCT= hematopoietic stem cell transplantation; IL-3= interleukin-3; iPSCs=induced pluripotent stem cells; iNKT= invariant natural killer T; iTCR =invariant T-cell receptor; Treg =regulatory T cell. From references 71, 76 and 77.

Role of human iNKTs in clinical GvHD

Several clinical studies highlight a protective role for human iNKTs against GvHD. First, non-myeloablative conditioning with TLI/ATG prior to HSCT decreases the incidence of acute^78,79 and chronic⁷⁹ GvHD. This lower incidence of GvHD is associated with a higher iNKT/T cell ratio, increased IL-4 production and marked reduction in donor T-cell proliferation.⁷⁸ Importantly, the TLI/ATG regimen does not abrogate the GvL effect, as evidenced by the high incidence of sustained complete remission (CR) among patients with active disease at the time of transplantation.⁷⁹ Second, both CD4⁺ and CD4⁻ subsets of iNKTs (Table 1) are reduced in patients with acute GvHD.⁸⁰ Third, enhanced iNKT reconstitution post transplantation has been shown to be a predictive factor for an improved overall survival associated with reduction in GvHD without abrogation of the GvL effect.⁸¹ Fourth, lower numbers of CD4⁻iNKTs in the donor graft is associated with clinically significant GvHD in patients undergoing HLA-identical allogeneic HSCT.⁸² Last, in a recent study, the frequency of iNKTs in pediatric HSCT patients was significantly reduced in the relapsed but not the non-relapsing patient cohort.⁸³ In the same study, there was no difference in the numbers of CD4⁺ and CD8⁺ T cells between the groups, suggesting that only the frequency of iNKTs correlates with a remission state after HSCT.⁸³ Thus, in a clinical setting, increasing iNKT numbers in donor grafts with very few iNKTs (by adding back in vitro expanded iNKTs^55,84–86 (Figure 3)), or adoptively transferring iNKTs into leukemia patients that fail to reconstitute the iNKT compartment early after allogeneic HSCT might provide an attractive strategy for suppressing GvHD and preventing leukemia relapse.

Figure 3.

In vitro expansion and isolation of human iNKTs. (a) Human PBMC are cultured in complete medium (Aim-V, 10% fetal calf serum; recombinant human (rh) IL-2 (50 U/mL)) and αGC (500 ng/mL)). After 4 days, cultures are supplemented with rhIL-15 (10 ng/mL) and rh IL-2 (10 U/mL) and 4–5 days later, iNKTs are purified by MACS sorting, based on expression of Vα24, the α-chain of the iNKT TCR. Using this approach, we observe that iNKTs can be expanded 500–1000-fold and it is therefore very feasible to obtain the large number of cells within a week. (b) Representative FACS plots demonstrate how iNKTs can be successfully expanded from the blood and isolated to >99% purity. Alternatively, iNKTs can be first isolated from PBMCs by MACS or high-speed cell sorting and then expanded in vitro in the presence of αGC and cytokines (IL-2, IL-7 and IL-15). For long-term cultures, iNKTs can be restimulated every 8–12 days with αGC-pulsed, irradiated autologous PBMC in the presence of cytokines.

Information on the graft composition of iNKTs in BM, PBSCs and CB is limited. There is only one study⁸² that has examined the frequencies of total as well as CD4⁺ and CD4⁻ iNKTs in donor PBSC grafts. In a previous study,⁸⁰ there was a significant difference in the number of reconstituted iNKTs in patients who received BMT and PBSC. The only variable associated with the number of iNKTs was the stem cell source (peripheral blood or BM). In PBSC recipients, the number of iNKTs was in the normal range within 1 month, whereas in the patients who received BMT, the iNKTs were not reconstituted within the first year post transplant. In another study, iNKT cells were reconstituted within a month after umbilical cord blood transplantation (UCBT).⁸⁷ Taken together, these studies indicate that the graft composition of iNKT cells in BM, PBSC and CB are different, which may impact the kinetics of iNKT-cell reconstitution as well as their repertoire in the transplant recipients.

Another factor that may impact iNKT-cell reconstitution and function in patients is the use of the immunosuppressive drug to prevent GvHD. However, studies have shown that the administration of cyclosporine with short-term methotrexate does not significantly affect iNKT numbers or donor chimerism in BMT or PBSC recipients.^80,81 Conversely, UCBT patients that received cyclosporine and mycophenolate mofetil had phenotypically and functionally immature iNKTs initially post engraftment but displayed rapid effector functions, including cytokine production and cytolytic activity within 3–6 months post-UCBT.⁸⁷ Given that iNKT cell profiles are similar in CB and early after UCBT, these observations indicate that administration of the immunosuppressive drugs may only transiently impact iNKT-cell functions in the immediate post-transplant period, if at all.

Specific human iNKT subsets regulate the opposing pro-GvL and anti-GvHD effects

In transplant recipients, CD4⁻ iNKTs reconstitute faster and attain functional maturity more rapidly than their CD4⁺ counterparts.⁸³ Importantly, the CD4⁻ iNKTs have a Th1 bias; they secrete higher amounts of IFN-γ than IL-4 and preferentially express perforin.⁸ Human CD4⁻ (but not CD4⁺) iNKTs express innate immune recognition receptors such NKG2D, CD94 and NKG2A.^8,57 Accordingly, human CD4⁻ iNKTs kill target cells expressing NKG2D ligands in an NKG2D-dependent and CD1d-independent manner.⁵⁷ It is thought that CD4⁻ iNKTs not only promote GvL but also suppress GvHD. In support of this notion, in vitro studies demonstrated that CD4⁻ iNKTs display direct cytotoxicity against CD1d-expressing mature myeloid DCs.⁸² In an allogeneic setting, alloreactivity of iNKTs depends on TCR-CD1d interactions as well as those involving activating killer Ig receptors (KIRs).⁸⁸ Consistently, recent studies demonstrate that human iNKTs express KIRs including KIRDL4, KIR3DL2 and KIR2DL1.⁸⁸ Thus, it is possible that donor CD4⁻ iNKT cells downregulate GvHD by killing of host APCs (cells which perpetuate GvHD in secondary lymphoid organs) in a TCR-CD1d and KIR-dependent manner. On the other hand, CD4⁺ iNKT cells can ameliorate GvHD by their provision of IL-4,^8,81 which can polarize the pathogenic donor T cells toward an anti-inflammatory Th2 response as well promote expansion of the regulatory T cells.^18,89 We therefore favor the interpretation that the opposing pro-GvL and anti-GvHD effects are likely being mediated by distinct iNKT subsets that are each endowed with distinct cytokine profiles, resulting in a collectively beneficial effector response for transplant recipients.

Pharmacological manipulation of iNKTs for use in allogeneic HSCT Given the protective role of iNKTs in allogeneic HSCT, the use of αGC provides an attractive pharmacological approach to effectively separate the GvHD and GvL effects. Indeed, donor iNKTs expanded in vitro by stimulation with αGC attenuate GvHD.^90,91 Furthermore, in vivo injection of αGC or OCH (a homologue of αGC) attenuates GvHD severity via host iNKT production of IL-4 and Th2 polarization of donor T cells.⁸⁹ However, contradictory to these favorable observations, another study⁹² reported that the administration of αGC (but not its N-acyl variant, C20:2) induces hyper-acute GvHD and rapid mortality in mice. Exacerbation of GvHD in this study was associated with profound iNKT activation and IL-12 secretion by host DCs, resulting in NK- and T-cell activation and increased serum levels of pro-inflammatory cytokines.⁹² Studies have shown that aqueous αGC when presented by DCs activates iNKTs, whereas liposomal αGC that is usually presented by B cells triggers IL-10 production by iNKT and B cells, resulting in expansion of tolerogenic DCs and generation of Tregs.^93,94 In line with these observations, administration of the liposomal formulation of αGC (RGI-2001) prevents GvHD in mice (via expansion of nTregs) but retains the GvL effect.⁹⁵ The safety and efficacy of this pharmacological approach is currently under investigation in HSCT patients with hematologic malignancies in a multi-center phase 1/2 clinical trial (NCT013729209). In addition, progenipoietin-1, a chimeric cytokine that stimulates both G-CSF and Flt-3 L, has been shown to suppress GvHD⁹⁶ but promote iNKT-mediated anti-leukemia response.⁹⁷ Interestingly, in a recent study low doses of αGC and β-mannosylceramide acted synergistically to modulate iNKT responses in a preclinical tumor model.³² Whether this strategy can suppress GvHD, remains to be determined. Previously, it was shown that lenalidomide, a thalidomide analog enhances αGC-induced iNKT expansion and IFN-γ production in both healthy donors and patients with MM,⁹⁸ suggesting that combining iNKT ligands with lenalidomide may provide a viable approach to enhance the efficacy of either therapy against human cancer. However, in two separate clinical trials,^99,100 treatment with lenalidomide after allogeneic HSCT induced severe GvHD in the MM patients⁹⁹ and those with advanced MDS or AML.¹⁰⁰ Thus, while pharmacological manipulation of iNKTs holds significant promise, it is critical to understand the mechanisms by which these agents modulate various immune cell functions and to ensure that injections of these agonists into cancer patients will not induce anergy or hyper-activation of iNKT or other immune cells.

CONCLUSIONS AND FUTURE PERSPECTIVES

In conclusion, we document several preclinical and clinical studies that support the development of innovative iNKT-based therapies for the treatment of cancers. Given that several hematological malignancies express CD1d, they can serve as direct targets for iNKT recognition and attack. Furthermore, by virtue of their DC-priming capabilities and robust cytokine production, iNKTs hold great potential to modulate the anti-leukemia effects of other cytolytic effectors and confer protection from infections; advantages not provided by conventional T-cell-based therapies. Importantly, limited polymorphism of the human CD1d gene and lack of CD1d incompatibility between donors and recipients renders transfer of mature iNKTs a more applicable approach than the infusion of conventional T cells. However, rational use of these cells in cancer immunotherapy awaits better understanding of iNKT reconstitution, effector functions and survival properties in humans. Furthermore, given the heterogeneous nature of iNKTs, the challenge remains in understanding how to manipulate the different subsets of human iNKTs to induce favorable clinical response with no or limited adverse side-effects. Thus, future studies to identify the molecular pathways involved in the differentiation of iNKT subsets and their skewing toward specific effector functions as well as their regulation of other immune cells is highly warranted. These studies will hopefully allow us to fully understand and exploit the therapeutic potential of iNKTs for the treatment of cancers or other immune disorders.