Human CD4⁺ iNKT cell adoptive immunotherapy induces anti-tumor responses against CD1d-negative EBV-driven B lymphoma

Abstract

Invariant natural killer T (iNKT) cells are a conserved population of innate T lymphocytes that are uniquely suitable as off-the-shelf cellular immunotherapies due to their lack of alloreactivity. Two major subpopulations of human iNKT cells have been delineated, a CD4⁻ subset that has a T_H1/cytolytic profile, and a CD4⁺ subset that appears polyfunctional and can produce both regulatory and immunostimulatory cytokines. Whether these two subsets differ in anti-tumor effects is not known. Using live cell imaging, we found that CD4⁻ iNKT cells limited growth of CD1d⁺ EBV-infected B-lymphoblastoid spheroids in vitro, whereas CD4⁺ iNKT cells showed little or no direct anti-tumor activity. However, the effects of the two subsets were reversed when we tested them as adoptive immunotherapies in vivo using a xenograft model of EBV-driven human B cell lymphoma. We found that EBV-infected B cells down-regulated CD1d in vivo, and administering CD4⁻ iNKT cells had no discernable impact on tumor mass. In contrast, xenotransplanted mice bearing lymphomas showed rapid reduction in tumor mass after administering CD4⁺ iNKT cells. Immunotherapeutic CD4⁺ iNKT cells trafficked to both spleen and tumor and were associated with subsequently enhanced responses of xenotransplanted human T cells against EBV. CD4⁺ iNKT cells also had adjuvant-like effects on monocyte-derived DCs and promoted antigen-dependent responses of human T cells in vitro. These results show that allogeneic CD4⁺ iNKT cellular immunotherapy leads to marked anti-tumor activity through indirect pathways that do not require tumor cell CD1d expression and that are associated with enhanced activity of antigen-specific T cells.

Graphical Abstract

Using a human xenograft model, this study demonstrates that CD4⁺ iNKT cell adoptive immunotherapy is associated with enhanced activation of peptide antigen-specific T cells and leads to marked anti-tumor effects against an EBV-driven B cell lymphoma that down-regulates CD1d in vivo. Therefore, adoptively transferred CD4⁺ iNKT cells may be of value to promote anti-tumor activity of patient T cells regardless of tumor cell CD1d expression.

INTRODUCTION

Invariant Natural Killer T (iNKT) cells are a small subset of innate-like T cells that have attracted attention for their immunotherapeutic potential. They are defined by their use of a semi-invariant T cell receptor (TCR), in which the TCRα chain is rearranged in a nearly invariant (“canonical”) manner and TCRβ chain usage is also constrained ^{1, 2}. The TCRs of iNKT cells recognize conserved lipids as antigens presented by CD1d ³, a type of antigen presenting molecule that is almost completely non-polymorphic in human populations ⁴. As a result of these characteristics, iNKT cells do not mediate alloreactivity, and adoptively transferred allogeneic iNKT cells showed no dose-limiting toxicities in a recent phase I/2 clinical trial ⁵. Moreover, a recent analysis provides support that allogeneic iNKT cells can persist in MHC-mismatched recipients in vivo ⁶. iNKT cells thus have potential for use as an off-the-shelf cellular immunotherapy suitable for treatment of genetically unrelated individuals. However, understanding requirements for iNKT-mediated immunotherapeutic activity, and the associated mechanisms of action, will be key to realizing this potential.

Prior analyses have demonstrated that iNKT cells can mediate anti-tumor effects through multiple distinct pathways. iNKT cells can directly kill tumor cells through cell death-inducing pathways including perforin/granzyme, TRAIL, and CD95/CD95L ^7–10. Additionally, iNKT cells can indirectly promote anti-tumor immunity by acting on key antigen presenting cells (APCs) to promote the activation of other effector populations such as NK cells and T cells. For example, expression of CD40L and IFN-γ by iNKT cells induces DCs in draining lymph nodes to upregulate co-stimulatory ligands and produce IL-12, rendering them more potent activators of anti-tumor immune responses ^11–13. Additionally, iNKT cells have been shown to use CD40 and Fas pathways to limit the suppressive activity of myeloid cells within the tumor microenvironment, resulting in improved anti-tumor immunity ^{14, 15}. On the other hand, in many experimental model systems iNKT cells have been shown to instead mediate potent anti-inflammatory effects that limit immune activation (reviewed in ¹⁶). The anti-inflammatory effects of iNKT cells are associated with production of T_H2 or regulatory cytokines such as IL-4, IL-13, and IL-10 ^17–20, and with the activation of regulatory cell types including M2-polarized macrophages, myeloid-derived suppressor cells (MDSCs), and T_regs ^21–28. It remains unclear what determines the seemingly contrasting anti-tumor and anti-inflammatory functions of iNKT cells.

In particular, it is not clear whether these distinct outcomes are due to different subsets within the iNKT cell population. Human iNKT cells can be broadly subdivided into two major subsets based on expression of CD4 ²⁹. CD4⁻ iNKT cells show a predominantly T_H1 cytokine profile and are enriched for perforin expressing cells, whereas CD4⁺ iNKT cells display a polyfunctional cytokine phenotype that includes production of T_H1, T_H2, and regulatory cytokines, and are enriched for CD95L expression ²⁹. Thus, perforin-dependent anti-tumor cytolytic activity by human iNKT cells seems most likely to be mediated mainly by the CD4⁻ subset, but the role of CD4⁺ iNKT cells is less clear. CD4⁺ iNKT cells might mediate anti-tumor activity through alternate mechanisms, or they may actually curtail anti-tumor responses by promoting regulatory pathways.

Another important question concerns the requirements for activation of iNKT cell anti-tumor activity. iNKT cells can be activated to produce IFN-γ through TCR-independent pathways ^30–33, but other cytokines appear to require at least low levels of TCR stimulation ^{31, 34}. Additionally, prior studies have shown that human iNKT cells require comparatively strong TCR stimulation in order to carry out direct cytolysis ^{9, 35, 36}. Physiologically, the strength of TCR stimulation encountered by iNKT cells is determined by the level of CD1d expression on antigen presenting cells (APCs) in combination with the potency and abundance of the presented antigens. iNKT cells can be activated by recognition of lipid antigens derived either from self or from microbes (reviewed in ³⁷). Certain microbially-derived glycolipids such as α-galactosylceramide (α-GalCer) are highly potent agonists that provide sufficient TCR stimulation to induce iNKT cell cytotoxic responses ^{7, 38, 39}. In contrast, presentation of self lipids is able to induce iNKT cell cytokine secretion, but is typically less effective for promoting cytotoxicity ^{34, 40, 41}. Neoplastic transformation may be associated with lipidomic changes that generate more highly potent antigens, however, this may be counterbalanced by downregulation of CD1d molecules or other compensatory changes that reduce the antigenicity of tumor cells for iNKT cells ⁴². Hence, since evasive events are likely to accrue in tumors over time, iNKT targeting of tumor cells for direct cytolysis through TCR-mediated recognition of CD1d may be limited in clinical settings, but alternative anti-tumor pathways may be maintained.

These factors have a critical bearing on iNKT-based immunotherapeutic strategies. While administering potent glycolipid agonists such as α-GalCer can promote human iNKT cell responses in vivo ^43–45, prior studies have shown that iNKT cells of cancer patients often appear reduced in frequency and/or hyporesponsive ^{46, 47}. Additionally, a number of prior findings have raised concerns that repeated administration of potent agonists may exhaust iNKT cells, or bias them towards a regulatory phenotype, or lead to depletion of CD1d⁺ DCs ^48–50. Therefore, attention has shifted towards protocols involving adoptive iNKT cellular immunotherapies administered without an accompanying glycolipid antigen, which have shown promise in recent clinical trials ^51–53. Since advancing these approaches will benefit from further understanding the anti-tumor functionality of human iNKT subsets, in this study we investigated human CD4⁺ and CD4⁻ iNKT cellular immunotherapies using a xenograft model of human B lymphomagenesis driven by Epstein-Barr virus (EBV).

METHODS

Generation of immunotherapeutic iNKT and γδ T cells.

Blood samples were collected with informed consent from healthy adult donors, and used in accordance with UW Minimal Risk IRB protocol #2018–0304. Peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation using Ficoll-paque PLUS (GE Healthcare). CD4⁺ or CD4⁻ iNKT cells were sorted using human CD1d tetramers loaded with a lipid analogue of α-GalCer (provided by the NIH tetramer facility at Emory University). iNKT cells were expanded in the presence of irradiated PBMCs, 20 ng/ml anti-CD3 mAb (clone OKT3) and/or 1 μg/ml phytohemagglutinen (PHA), in “culture medium” [RPMI1640 containing L-glutamine (Corning), 10% heat-inactivated fetal bovine serum (GeminiBio), 5% heat-inactivated bovine calf serum (Hyclone), 3% pooled human AB serum (Atlanta Biologicals), 2 mM L-glutamine, 100 IU/ml Penicillin, 100 μg/ml Streptomycin (all from MediaTech)] supplemented with 200 U/ml recombinant human IL-2 (Peprotech). The studies were performed using 5 different CD4⁺ iNKT lines (2 clonal and 3 polyclonal isolated from 3 different donors), and 2 polyclonal CD4⁻ iNKT lines isolated from different donors. Expanded iNKT cells were regularly tested to confirm staining by α-GalCer loaded CD1d tetramers and invariant TCR expression using the 6B11 mAb. γδ T cells (3 lines, isolated from different donors) were generated by culturing PBMCs in IL-2 supplemented medium for 7–14 days in the presence of 2.5 μM zoledronic acid (Novartis), followed by magnetic sorting to remove contaminating subsets.

Preparation of EBV.

Experiments were performed using the lytic M81 strain of EBV. M81 bacmid expressing the green fluorescent protein (GFP) and a hygromycin B resistance gene was constructed using bacterial artificial chromosome technology as described ⁵⁴. Viral particles were produced from stably infected 293 cells, and infectious titer was determined by assessing green fluorescence after titrated infection of Raji cells, as previously described ⁵⁵.

Generation of EBV-transformed lymphoblastoid cells.

The S156 lymphoblastoid line was generated by M81 EBV infection of primary human splenic B cells from an EBV-negative donor. Human spleen tissue was collected from a cadaveric donor who was enrolled in the UWHealth organ donation program in accordance with UW IRB protocols #2019–0349 and #2021–0075. A single cell suspension was prepared and incubated with 5,000 IU EBV per 1×10⁶ B cells in culture medium at 37 °C and 5% CO₂. Experiments were performed after transformed B cells had grown out and were proliferating stably.

Live cell imaging.

Lymphoblastoid cells were labeled with 5 μM CFSE (Invitrogen), washed, and plated in triplicate in flat-bottom 96-well plates (Corning) in culture medium containing 500 nM propidium iodide (PI). Unlabeled iNKT or γδ T cells were added at a 3:1 ratio, and imaged for 48h using an IncuCyte S3 instrument (Sartorius) in a humidified incubator at 37°C with 5% CO₂. Data analysis was performed using the IncuCyte platform’s software. Cell masking was performed to identify CFSE singly stained (live) and CFSE/PI double-positive (dead) lymphoblastoid cells. Specific lysis was calculated by determining the dead fraction of the total lymphoblastoid cells and subtracting spontaneous lysis. Spheroid size was determined by setting a threshold to exclude single cells and computing the two-dimensional area (μm²) of CFSE-labeled clusters. Cumulative spheroid growth was determined by calculating the percent change in average spheroid size at 2h intervals from 8–48h of culture.

Flow cytometry.

Single cell suspensions in PBS containing 0.1% bovine calf serum, 2mM EDTA, and 0.05% NaN3, were blocked with 20% human AB serum for 15 min, then stained for 30 min at 4°C with fluorescently-labeled antibodies against the indicated markers. Fluorescence was measured using LSRII (Becton Dickinson) or AttuneNxT (Thermo Fisher Scientific) flow cytometers and data analyzed using FlowJo software (Becton Dickinson). For analysis of intracellular IFN-γ, cells were stained for surface antigens then fixed and permeabilized using the Cyto-Fast buffer set (BioLegend) prior to intracellular staining. Antibodies used were as follows (all from BioLegend): anti-human CD1d (clone 51.1); CD3 (clone OKT3); CD4 (clone OKT4); CD8a (clone RPA-T8 or HIT8a); CD14 (M5E2); CD19 (clone HIB19 or SJ25C1); CD20 (clone 2H7); CD40 (clone HB14); pan-CD45 (clone HI30 or 2D1); CD69 (clone FN50); CD80 (clone 2D10); CD83 (clone HB15E); CD86 (clone IT2.2); HLA-ABC (clone W6/32); HLA-DR (clone L243); IFN-γ (clone 45.B3); iNKT TCR (clone 6B11); TCR Vδ2 (clone B6); murine CD45.1 (clone A20); isotype controls mouse IgG1 (clone MOPC-21), IgG2a (MOPC-173), IgG2b (MG2B-57).

In vivo EBV-Lymphoma model.

De-identified human umbilical cord blood samples were acquired from commercial and academic suppliers, and used in accordance with UW Minimal Risk IRB protocol #2017–0870. Cord blood mononuclear cells (CBMCs) were purified by density gradient centrifugation, suspended in culture medium and incubated with 2000 U M81 EBV for 2 hours at 37 °C and 5% CO₂, then washed and resuspended in sterile PBS. 10×10⁶ cells per mouse were injected intraperitoneally into 6–10 week old NOD.Cg-Prkdc^scidIl2rg^tm1Wjl/SzJ (NSG) mice (Jackson Labs). Mice were maintained in a specific pathogen-free facility using microisolator cages, with sterilized bedding, food, and water. Immunotherapeutic cells were administered intravenously through retro-orbital injection to randomly chosen transplanted mice at the indicated time points. Mice were euthanized 29–31 days after injection of EBV-infected CBMCs and macroscopically visible tumor tissue excised and weighed (analysis performed in a non-blinded manner). Since tumor masses appeared to depend in part on the human CBMC sample used for transplantation, to enable comparisons of results that included mice generated from different CBMC samples, normalized tumor mass was calculated by dividing the raw tumor mass for each mouse by the mean tumor mass of the control group from the same CBMC sample.

Activation of adult peripheral blood lymphocytes.

PBMCs from healthy adult donors were incubated at 37 °C and 5% CO₂ for 48h in round-bottom 96-well plates at 2×10⁵ cells per well in culture medium alone, or with immunotherapeutic CD4⁺ iNKT cells added to a final concentration of 0.5%, or with α-GalCer added to a final concentration of 200 ng/ml. Supernatants were collected for analysis of IFN-γ by ELISA, and cells were analyzed by flow cytometry.

Activation of DCs.

DCs were prepared by culturing monocytes isolated from healthy adult PBMCs using CD14 magnetic beads (Miltenyi Biotec) for 3–4 days in culture medium containing 300 U/ml recombinant human GM-CSF and 200 U/ml IL-4 (Peprotech). The resulting immature DCs were combined at a 1:1 ratio with CD4⁺ iNKT cells, or treated with 500 ng/ml glucopyranosyl lipid A (GLA) adjuvant (Avanti Polar Lipids), or kept in medium alone for an additional 24h, then their cell surface phenotype was analyzed by flow cytometry. To assess T cell activation, DCs were exposed to 0.25 Lf/ml Tetanus toxoid (Massachusetts Biological Labs) or mock-treated for 2h, then washed. Autologous T cells were isolated by pan-T cell magnetic bead negative selection (Miltenyi Biosciences) and combined at a 50:1 ratio with DCs in the presence or absence of 0.5% allogeneic CD4⁺ iNKT cells, then incubated for 24–48h in culture medium with brefeldin A added for the last 12h. Cells were stained for CD3 and CD4, and with an iNKT TCR-specific mAb (6B11), then fixed, permeabilized, stained for intracellular IFN-γ or isotype-matched negative control mAb, and analyzed by flow cytometry.

iNKT trafficking in vivo.

Tumor-bearing mice were injected intravenously with 7.5×10⁶ CFSE-labeled CD4⁺ iNKT cells, and spleen and tumor tissue collected after 24h. Tissues were fixed in 4% paraformaldehyde overnight at 4 °C, then placed in a series of sucrose solutions at 4 °C while shaking (10% sucrose for 1h, 20% sucrose for 1h, 30% sucrose overnight). The tissues were placed in cryomolds (TissueTek), covered with the manufacturer’s medium and frozen at −80° prior to sectioning and mounting on slides. Tissues were counterstained with DAPI to identify cell nuclei and analyzed on a Zeiss LSM 800 confocal microscope to quantitate the number of CFSE positive cells in fields of defined unit size.

Analysis of T cell responses ex vivo.

Human T cells from spleens of tumor bearing mice were enriched using pan-T cell magnetic beads (Miltenyi Biotec) and iNKT cells were removed using the 6B11 TCR-specific mAb, then incubated for 72h in culture medium supplemented with 200 U/ml IL-2. T cells were tested for IFN-γ secretion in response to EBV-infected splenocytes from control mice, or to autologous uninfected CBMCs that were mock-treated or pulsed with 0.1 μM synthetic EBV or CMV peptides (Peptivator pools for MHCI and MHCII from Miltenyi Biotec).

Statistical analyses.

Cumulative spheroid growth was analyzed using a Kolmogorov-Smirnov test. For analyses of normalized tumor mass where results included mice transplanted with different CBMC samples, statistical significance was assessed using an unpaired two-tailed non-parametric t-test (Mann-Whitney). Tumor incidence was analyzed using a two-sided, Fisher’s exact t-test. Analysis of experiments where all transplanted mice or assay replicates were derived from the same human CBMC sample was done using an unpaired two-tailed parametric t-test. Paired responses amongst different human samples were analyzed using a two-tailed paired non-parametric t-test (Wilcoxon matched-pairs signed rank).

RESULTS

Cytotoxic activity of CD4⁺ and CD4⁻ iNKT cells in vitro.

We first used a live cell imaging system to investigate the ability of human CD4⁺ and CD4⁻ iNKT cells to kill EBV-transformed B lymphoblastoid cells in vitro. iNKT cells were flow cytometrically sorted from peripheral blood of healthy adults based on staining with α-GalCer loaded CD1d tetramers, anti-CD3, and anti-CD4. Single iNKT cells (clones) or samples containing 1,000–10,000 sorted iNKT cells (polyclonal lines) were expanded in vitro and used for analysis after they had exited from log-phase proliferation. As a positive control we used human γδ T cells expanded from peripheral blood by exposure to zoledronic acid, since we had previously found that these innate effectors showed potent anti-tumor effects against EBV-driven B lymphomas in vivo ⁵⁶.

Using an IncuCyte live cell imaging platform to assess killing of EBV-transformed cells over a 48h period, we observed that the γδ T cells rapidly mediated killing of lymphoblastoid cells, but killing reached a plateau and then did not progress further after 20h (Fig. 1A, left plot). CD4⁻ iNKT cells produced slower but more sustained killing of the lymphoblastoid cells that ultimately reached a similar level to the γδ T cells (Fig. 1A, left plot). In contrast, CD4⁺ iNKT cells produced only minimal specific lysis over the assay period (Fig. 1A, left plot). Microscopic images from this analysis revealed that the lymphoblastoid cells formed spheroids within the first 6h of culture, and showed that γδ T cells and CD4⁻ iNKT cells produced increased death of lymphoblastoid cells in spheroids within 24h, whereas the level of lymphoblastoid death in the presence of CD4⁺ iNKT cells appeared similar to the spontaneous lysis seen in the absence of effectors (Fig. 1A, middle images). Consistent with their sustained killing response, CD4⁻ iNKT cells significantly limited cumulative spheroid growth from 8–48h of culture, while neither γδ T cells nor CD4⁺ iNKT cells achieved this effect (Fig. 1A, right plot). Thus, both CD4⁻ iNKT and γδ T cells were able to kill EBV-transformed B cells, and CD4⁻ iNKT cells deterred tumor spheroid growth over time, whereas CD4⁺ iNKT cells did not show direct anti-tumor effects.

Figure 1. EBV-lymphoma killing by iNKT and γδ T cells in vitro.

A) Live cell imaging was performed to assess in vitro responses of γδ T cells, CD4⁻ and CD4⁺ iNKT cells to EBV-transformed lymphoblastoid cells. Left plot shows relative specific lysis mediated over time, with symbols representing means + SEM of 3 replicates. Middle panels show microscopic images at 4X magnification of lymphoblastoid spheroids at indicated time points, with live lymphoblastoid cells masked in green and dead in red. Right plot shows percent change in spheroid size (mean + SEM) from 8–48h of culture. B) NSG mice were injected with EBV-treated human CBMCs, and after 25 days were given 5–10×10⁶ allogeneic γδ T cells or CD4⁻ iNKT cells or mock-treated (control), then tumor mass was assessed 4–6 days later. C) Flow cytometric analysis of CD1d expression (shaded histograms) compared to isotype control staining (dotted line).

Anti-tumor activity of iNKT cell immunotherapy.

We used a human xenograft model of EBV-driven B lymphoma to assess anti-tumor activity of allogeneic iNKT cellular immunotherapy in vivo. In this model, human umbilical cord blood mononuclear cells (CBMCs) are briefly exposed in vitro to a lytic strain of EBV then injected intraperitoneally into immunodeficient NOD/SCID/γ_c^−/− (NSG) mice. EBV-infected human B cells within the cord blood sample become neoplastically transformed and form aggressive lymphomas in the peritoneal cavity within 3 weeks ⁵⁶. The B-lymphomas in this model are consistently highly infiltrated by autologous human CD4 and CD8 T cells, but due to the expression of checkpoint ligands these fail to control the tumors ⁵⁷. This model therefore enables investigation of the anti-tumor effectiveness of human cellular immunotherapies in an immunosuppressive context.

Allogeneic CD4⁻ iNKT cells were administered at day 25 (a timepoint at which tumors are already present in the peritoneal cavity), and mice were sacrificed six days later (day 31). Tumor mass was compared to that of control mice from the same cohort that did not receive immunotherapy. Surprisingly, CD4⁻ iNKT immunotherapy was not associated with a detectable reduction in tumor mass, although allogeneic γδ T cell immunotherapy did result in significantly lower tumor burden (Fig. 1B). Since prior studies have found that iNKT cell cytotoxicity is dependent on target cell expression of CD1d ⁹, we performed flow cytometric analysis of CD1d levels. There was detectable expression of CD1d on uninfected cord blood B cells and on the EBV-LCL used for our in vitro analyses, whereas CD1d appeared completely down-regulated on human B cells from lymphoma-bearing mice (Fig. 1C). Thus, CD4⁻ iNKT cells may be ineffective against EBV-driven lymphoma due to down-regulated CD1d expression.

In contrast, adoptive transfer of allogeneic CD4⁺ iNKT cells did produce a highly significant anti-tumor effect (Fig. 2). The mean tumor mass for mice given CD4⁺ iNKT cells at day 25 was 30.6% of that for the control mice, and nearly 50% of the mice treated with CD4⁺ iNKT cells did not have any macroscopically visible tumor tissue (Fig. 2A). CD4⁺ iNKT cells appeared less efficient than γδ T cells at low cell doses (1×10⁶ cells), but at higher doses (5–10×10⁶ cells) the anti-tumor effects of CD4⁺ iNKT cells and γδ T cells were similar (Fig. 2B). Anti-tumor effects mediated by CD4⁺ iNKT cells occurred rapidly, within 3 days after administration (Fig. 2C). Thus, despite producing no detectable anti-tumor effect in vitro, in a xenotransplant model where the autologous human T cells have failed to control tumors, allogeneic CD4⁺ iNKT cell immunotherapy led to rapid reduction in lymphoma mass.

Figure 2. In vivo anti-tumor effects of CD4⁺ iNKT immunotherapy.

A) Left plot shows normalized tumor mass of control mice compared to mice given 5–10×10⁶ allogeneic CD4⁺ iNKT cells at day 25, with tumors assessed at day 29–31; right plot shows tumor incidence. B) Dose-response comparison for CD4⁺ iNKT and γδ T cell immunotherapy given at d25. Symbols show means and SEM of results from 3–24 mice with asterisks indicating significance (*=P<0.05; ***=P<0.001) compared to no immunotherapy. C) Mice were given 5×10⁶ CD4⁺ iNKT cells or mock-treated at day 25, and tumors assessed 2–4 days later.

Adjuvant-like effects of CD4⁺ iNKT cells.

We therefore investigated the ability of immunotherapeutic CD4⁺ iNKT cells to activate other lymphocytes. PBMCs from healthy adult subjects were incubated for 24h alone or in the presence of 2% added CD4⁺ iNKT cells (i.e. 10–100 fold higher frequency than the endogenous iNKT cells in the samples). As a positive control, we cultured PBMCs in parallel with α-GalCer to activate endogenous iNKT cells. CD69 levels on lymphocyte populations of the PBMCs were assessed by flow cytometry, and secreted IFN-γ was quantitated by ELISA. Co-incubating PBMCs with immunotherapeutic CD4⁺ iNKT cells resulted in significant upregulation of CD69 on NK cells, T cells, and B cells, and also led to significantly elevated IFN-γ levels (Fig. 3A). Addition of CD4⁺ iNKT cells produced similar or greater lymphocyte activation as addition of α-GalCer to activate endogenous iNKT cells (Fig. 3A). To assess whether these effects were due to alloreactivity to allogeneic iNKT cells, we incubated a clonal CD4⁺ iNKT cell line in parallel with autologous or allogeneic PBMCs. Similar upregulation of CD69 and elevation of secreted IFN-γ was observed in each case (Fig. 3B), demonstrating that alloreactive activation was not required for the effects of the added CD4⁺ iNKT cells.

Figure 3. CD4⁺ iNKT cells mediate adjuvant-like activation of other immune cells.

A) Human PBMCs were incubated alone, or with 2% allogeneic CD4⁺ iNKT cells, or in medium containing 200 ng/ml α-GalCer, and after 48h CD69 expression by B, T, and NK cells from the PBMC sample was determined by flow cytometry (left plot) and levels of secreted IFN-γ were quantitated by ELISA (right plot). B) CD4⁺ iNKT cells were incubated in parallel with autologous or allogeneic PBMCs (filled symbols), or the PBMCs were incubated alone (open symbols). Left axis shows CD69 expression, right axis shows secreted IFN-γ. C) Monocyte-derived DCs were incubated with CD4⁺ iNKT cells, or treated with synthetic GLA, or kept in medium alone, and after 24h were stained for the indicated markers. Top row histograms show DCs that were autologous to the iNKT cells, bottom row DCs were allogeneic. Plot on right shows fold-increase of each marker for DCs incubated with iNKT cells compared to medium alone; each symbol shows results from an independent analysis, with asterisks indicating statistical significance of fold-increase as determined by a two-tailed 1-sample t-test. D) Monocyte-derived DCs were pulsed with tetanus toxoid (TT Ag) or mock-treated (No Ag) and incubated for 24–48h with autologous T cells in the presence or absence of 0.5% allogeneic iNKT cells. IFN-γ expression by CD4⁺ T cells (excluding iNKT cells) was determined by flow cytometry.

To investigate effects of CD4⁺ iNKT cells on DCs, we generated monocyte-derived DCs and incubated them for 24h alone or in the presence of a 1:1 ratio of CD4⁺ iNKT cells. As a positive control, DCs were incubated in parallel with a synthetic adjuvant called glucopyranosyl lipid A (GLA). Compared to DCs cultured alone, DCs co-incubated with iNKT cells showed elevated levels of co-stimulatory molecules such as CD80, CD86, and CD40, and acquired a more mature phenotype as indicated by upregulation of CD83 and increased cell surface expression of MHC class II molecules (Fig. 3C). The effects of iNKT cells on DCs appeared comparable to those induced by GLA, and were similar regardless of whether the iNKT cells were autologous or allogeneic to the DCs (Fig. 3C). These results suggested that the stimulatory capacity of DCs may be enhanced by exposure to the CD4⁺ iNKT cells we used for immunotherapy.

To investigate this, we tested the impact of the CD4⁺ iNKT cells on DC-mediated activation of antigen-specific T cells. Monocyte-derived DCs from tetanus-immunized donors were pulsed with tetanus toxoid (TT) antigen or mock-treated, then combined at a 1:50 ratio with purified autologous T cells. The cell mixture was incubated for 24–48h in the presence or absence of 0.5% added CD4⁺ iNKT cells, and intracellular cytokine staining was performed to assess IFN-γ production by CD4⁺ T cells (iNKT cells were specifically stained and gated out of the analysis). The frequency of IFN-γ⁺ T cells appeared slightly elevated for T cells exposed to autologous DCs pulsed with TT antigen compared to mock-treated DCs, but the difference did not reach statistical significance (p=0.065), suggesting that unstimulated monocyte-derived DCs were inefficient activators of antigen-specific CD4⁺ T cells (Fig. 3D). In contrast, addition of the CD4⁺ iNKT cells to the cultures was associated with significantly increased frequencies of IFN-γ⁺ T cells in response to antigen-pulsed DCs, and also led to slightly elevated T cell activation in response to mock-treated DCs (Fig. 3D). Together, these results suggested that the CD4⁺ iNKT cells enhanced the T cell stimulating capacity of DCs.

Indirect anti-tumor effects of iNKT immunotherapy.

We previously observed that γδ T cells mediated effective early immunosurveillance to prevent tumor outgrowth in our EBV-driven lymphoma model ⁵⁶. However, in contrast to their marked anti-tumor effects at later time points (Fig. 2), CD4⁺ iNKT cells given within the first 1–3 days after EBV-infection did not induce a detectable reduction in tumor mass (Fig. 4A). Delivering CD4⁺ iNKT cells at day 18 after EBV-infection resulted in a partial reduction in tumor mass, and greater anti-tumor efficacy was observed when iNKT cell immunotherapy was given at day 21 or later (Fig. 4B). Thus, CD4⁺ iNKT cell immunotherapy did not prevent the outgrowth of nascently transformed B cells in vivo, yet produced a marked anti-tumor effect when given at later time points after tumors were already established.

Figure 4. CD4⁺ iNKT immunotherapy does not mediate early immunosurveillance.

A) Tumor mass results from control mice or mice given CD4⁺ iNKT cells within the first 3 days after injection of EBV-treated CBMCs. B) Mice were given CD4⁺ iNKT cells at the indicated times after injection of EBV-treated CBMCs and tumor mass was determined at day 29.

Using fluorescently-labeled CD4⁺ iNKT cells, we found that they trafficked to both spleen and tumor tissue (Fig. 5A), suggesting that their immunotherapeutic effects might be due to activity in spleen. We also found that a small fraction (typically <1%) of the human cells in spleens of tumor-bearing mice were CD33⁺ myleoid cells, and were positive for cell surface CD1d (Fig. 5B). Analysis of human cells in murine spleen 8 days after administration of iNKT immunotherapy revealed a detectable increase in the frequency of (non-iNKT) CD4⁺ T cells (Fig. 5B). To assess impact on CD4⁺ T cell effector functioning, human T cells were harvested from spleens of mice given CD4⁺ iNKT cells or mock-treated controls, and cultured in the presence or absence of T cell-depleted splenocytes from EBV-infected control mice. T cells from mice that had received iNKT cell immunotherapy showed significantly higher IFN-γ production than those from control mice (Fig. 5C). To further investigate, we tested responses of T cells from iNKT-treated or control mice to uninfected autologous CBMCs in the presence or absence of synthetic EBV peptides. T cells from mice given CD4⁺ iNKT immunotherapy showed significantly enhanced IFN-γ production in the presence of EBV peptides compared to either vehicle or CMV peptides (Fig. 5D). Thus, CD4⁺ iNKT immunotherapy was associated with enhanced EBV-specific T cell responses in this model.

Figure 5. Impact of CD4⁺ iNKT immunotherapy on splenic T cells.

A) Lymphoma-bearing mice were injected with fluorescently labeled CD4⁺ iNKT cells. Spleen and tumor tissue was collected after 24h. Left images show fluorescence microscopy of spleen and tumor sections at 20X magnification, with iNKT cells shown in green and cell nuclei in blue; right plot shows quantitation of labeled iNKT cells per mm² (means and SEM from 5–6 sections). B) Flow cytometric analysis of splenocytes from lymphoma-bearing mice (29–31 days post-EBV infection) showing identification of a small population of human CD33⁺ myeloid cells that express cell surface CD1d. Samples were first gated by light scatter and DAPI to exclude dead cells, then gated specifically on human cells using antibodies against murine CD45, human CD45, and pan-HLA class I. CD33⁺ cells shown in left plot were further gated to exclude CD3⁺ or CD19⁺ events, then staining for CD1d (filled histogram) was compared to isotype control (dotted line). Plot on right shows paired results from 3 different mice. C) Mice were given CD4⁺ iNKT cells or mock-treated (control) at day 21 after injection of EBV-treated CBMCs. Plot shows frequency of non-iNKT CD4⁺ T cells out of total human cells in spleen 8 days later. D) Human T cells from spleens of iNKT-treated or control mice were depleted of iNKT cells then cultured alone (No stim) or with autologous EBV-infected splenocytes, and secreted IFN-γ was quantitated by ELISA. E) T cells were cultured alone (No stim) or with uninfected autologous CBMCs treated with vehicle or synthetic EBV or CMV peptides.

DISCUSSION

We show here that allogeneic human CD4⁺ iNKT cell immunotherapy can produce marked anti-tumor effects against an aggressive human B cell lymphoma that down-regulates CD1d in vivo, whereas anti-tumor activity by CD4⁻ iNKT cells may require expression of CD1d by tumor cells. B lymphocytes normally constitutively express CD1d ⁵⁸, however B cell lymphomas and virally infected B cells frequently down-regulate CD1d enabling them to escape iNKT recognition ^{59, 60}. Additionally, in some cases (e.g. Mantle Cell lymphoma), tumor cells that retain CD1d have been found to undergo modifications to endogenous lipid production that result in reduced CD1d-mediated activation of iNKT cells ⁶¹. Our results suggest that while CD4⁻ iNKT cells may fail to effectively kill tumor cells that have undergone such CD1d-related evasive pathways, CD4⁺ iNKT cells are not subject to similar limitations because they mediate anti-tumor effects through indirect mechanisms that do not require tumor cell expression of CD1d.

Our analysis indicates that CD4⁺ iNKT cells act in an adjuvant-like manner to enhance anti-tumor responses by antigen-specific T cells. In our xenotransplant model of EBV-driven B-lymphoma, autologous human T lymphocytes derived from the starting CBMC sample are abundantly present ⁶². These T cells infiltrate the lymphomas, but fail to control tumor growth unless checkpoint blockade is administered, suggesting that they become suppressed or exhausted ⁵⁷. The results presented here suggest that CD4⁺ iNKT immunotherapy relieves the suppressed/exhausted status of T cells in this model, since T cells from iNKT-treated mice showed improved effector responses ex vivo. We hypothesize that this is due to iNKT-mediated activation of key human APCs that are also present in the mice. While CD1d appears to be completely down-regulated on B cells from the xenotransplanted mice, we identified myeloid lineage cells that retain CD1d expression. Since we found that CD4⁺ iNKT cells enhanced the expression of co-stimulatory molecules by monocyte-derived DCs and promoted their ability to activate TT antigen-specific T cells in vitro, we hypothesize that iNKT cells may interact similarly with the CD1d⁺ myeloid APCs to promote T cell activation in the xenotransplant model. Alternatively, the CD4⁺ iNKT immunotherapy may be acting to counteract the suppressive effects of CD1d⁺ myeloid cells within the tumor environment, which is an iNKT anti-tumor mechanism recently delineated in other studies ^{14, 15}. It is also possible that the immunotherapeutic activity of CD4⁺ iNKT cells results from CD1d-independent interactions with EBV-infected B cells, since we have previously found that CD4⁺ iNKT cells can be activated in a TCR-independent manner through ligation of their cell-surface LFA-1 by high density ICAM-1 ³³, and a subset of EBV-transformed B cells show elevated ICAM-1 expression ⁶³.

In contrast to γδ T cells, which we previously observed were able to mediate early immunosurveillance to prevent lymphoma outgrowth ⁵⁶, CD4⁺ iNKT immunotherapy was ineffective at early timepoints, yet when given at later time points resulted in rapid reduction of established tumors. This is consistent with our hypothesis that CD4⁺ iNKT cells mediate anti-tumor effects through adjuvant-like activity that requires a period of time for antigen-specific T cells to expand in response to the EBV-infected B cells. Prior studies have shown that antigen-specific CD4⁺ T cells are important cytolytic effectors for controlling EBV-driven lymphoma ^64–66. Consistent with this we observed enhanced proliferation of (non-iNKT) human CD4⁺ T cells after administration of iNKT immunotherapy. It is not clear whether iNKT immunotherapy also activates CD8⁺ T cells or NK cells in our model, although we did observe activation of NK cells after in vitro co-incubation of PBMC samples with immunotherapeutic CD4⁺ iNKT cells, which is similar to findings by others showing that iNKT cells readily transactivate NK cells ^67–69.

A key question for the development of immunotherapies based on CD4⁺ iNKT cells is whether this subset contains subpopulations that differ in capacity to promote anti-tumor immunity. We noted that all 5 of the independently-generated CD4⁺ iNKT cell lines used in these studies promoted marked anti-tumor effects in our in vivo model, suggesting that anti-tumor activity is shared by a substantial fraction of CD4⁺ iNKT cells. Recent studies by others have shown that expansion of human iNKT cells in the presence of the highly potent lipid antigen α-GalCer is associated with enrichment of iNKT cells expressing cell surface CD62L, which have elevated intracellular expression of the LEF1 transcription factor and mediate prolonged anti-tumor activity in vivo ^{70, 71}. We have observed that our immunotherapeutic CD4⁺ iNKT cells, which are expanded without the addition of α-GalCer, are consistently nearly completely negative for CD62L (data not shown), although they nevertheless grow robustly in culture, respond strongly to repeated restimulation, and persist in vivo for at least 3 weeks after transplantation into NSG mice ⁷². Thus, in future studies it will be of interest to further investigate the impact of distinct culture conditions (e.g. presence or absence of added glycolipid antigens, or inclusion of APCs with defined co-stimulatory ligands) and to assess whether specific iNKT phenotypes are associated with stronger immunotherapeutic activity or distinct mechanisms of action.

Similarly, since we hypothesize that CD4⁺ iNKT cell immunotherapy mediates adjuvant-like effects that promote the anti-tumor responses of endogenous T cells, it will be of great interest to determine whether this is synergistic with checkpoint blockade or agonist antibodies aimed at activating endogenous T cells. In this regard, the role of tumor antigens recognized by the endogenous T cells is also a key question. Since it is possible that the EBV-driven lymphoma model we used in this study may provide particularly abundant tumor-associated antigens for MHC-restricted T cells, it will be critical to determine whether adoptive CD4⁺ iNKT immunotherapy provides a benefit in other types of cancer by enhancing anti-tumor immunity by endogenous T cells. Overall, these results highlight the potential value of CD4⁺ iNKT cells as immunotherapeutic agents to enhance anti-tumor immunity.