Background
Cellular immunotherapy for cancer is a rapidly evolving field, endeavoring to deal with common relapse or resistance to conventional treatments. Herewith, most interesting are the γδ T cells, a T cell subset possessing a combination of innate and adaptive immune cell traits [
1]. Activated γδ T cells have strong cytotoxic effector functions, by means of both death receptor/ligand and cytolytic granule pathways. Moreover, they have an immunoregulatory function, producing various cytokines, including the T helper (Th)1-associated cytokines tumor necrosis factor (TNF)-α and interferon (IFN)-γ [
1]. γδ T cells are considered to be important players in cancer immune surveillance evidenced by; (a) the increased incidence of tumors in γδ T cell deficient mice [
2,
3], (b) the overrepresentation of γδ T cells in reactive lymphatic regions associated with neoplasia, including acute myeloid leukemia (AML) [
4], (c) their infiltration into solid tumors [
5], and (d) their potential to kill a variety of tumor cells [
1]. Pointedly, tumor-infiltrating γδ T cells recently emerged as the most significant favorable prognostic immune population across 39 malignancies [
6]. As γδ T cells can kill tumor cells without previous contact and do not induce graft-versus-host disease, the adoptive transfer of γδ T cells could be a promising alternative for stem cell transplantation and adoptive transfer of αβ T cells [
7].
Among the various tumor targets of γδ T cells [
8], we focused on AML, a heterogeneous hematological malignancy involving the clonal expansion of myeloid blasts in the bone marrow and peripheral blood. Although significant improvement in treatment of AML has been made, the unfortunate reality is that currently available treatments are largely ineffective for most AML patients [
9,
10]. In pursuit of new treatment options, several lines of evidence suggest that immunotherapy is an active modality in AML [
11]. This includes the graft-versus-leukemia effect associated with allogeneic hematopoietic stem cell transplantation (HSCT), where it has been demonstrated that elevated γδ T cell immune recovery after HSCT is associated with a better outcome in terms of infections, graft-versus-host disease and overall survival [
12‐
14]. Moreover, while the use of HSCT is restricted to a minority of patients due to, among other, its high transplant-related mortality and morbidity [
10,
15], γδ T cell immune therapy is well-tolerated and safe [
7,
16]. Adoptive transfer of γδ T cells is therefore an interesting alternative to tackle minimal residual disease and the high relapse rate in AML patients, optionally in combination with HSCT [
17] or other (new) therapeutic agents [
18‐
20]. Overall, while γδ T cell therapy holds great promise, clinical results are thus far modest, underscoring the need to further enhance the immunogenicity of the γδ T cell product [
21].
γδ T cells can be expanded, using combinations of cytokines and phosphoantigens (e.g., isopentenyl pyrophosphate (IPP)) or aminobisphosphonates (e.g., zoledronate) [
22]. The inclusion of aminobisphosphonates relies on their inhibition of farnesyl pyrophosphate synthase, a key enzyme of the mevalonate pathway, leading to accumulation of mevalonate metabolites such as IPP [
23]. When peripheral blood mononuclear cells (PBMC) are treated with zoledronate, IPP will selectively accumulate in monocytes, due to the efficient drug uptake by these cells [
24]. Therefore, in the absence of monocytes, addition of aminobisphosphonates is inefficient for induction of γδ T cell expansion, and the use of IPP or related phosphoantigen is required. To date, the standard protocol for the expansion of γδ T cells out of PBMC relies on the combination of zoledronate and interleukin (IL)-2 [
25].
Limitations on the use of IL-2, relating to among other its toxicity, invigorate the investment in exploring other cytokines of the IL-2 family (also called the common γ chain cytokine family; IL-4, IL-7, IL-9, IL-15, and IL-21) [
26]. IL-15 is of interest as it is closely related to IL-2 and has got the top position in the US National Cancer Institute’s ranking of 20 immunotherapeutic drugs with the greatest potential for broad usage in cancer therapy [
27]. Moreover, both cytokines mediate their effects through a heterotrimeric receptor complex consisting of a cytokine specific α-chain (IL-2Rα or IL-15Rα), the IL-2/15Rβ-chain, and the common γ-chain [
28]. IL-2 and IL-15 share many functions in regulating both adaptive (stimulation of T cell proliferation and induction of cytotoxic T lymphocytes) and innate (activating natural killer (NK) cells) immune responses [
29]. In spite of the resemblances between both cytokines, they also have different functions in vivo. IL-15 is specifically known for its role in the maintenance of long-lasting, high-avidity T cell responses, whereas IL-2 may induce activation-induced cell death and provoke maintenance of regulatory T cells [
26,
30]. Therewithal, IL-15 has shown efficacy in murine models of malignancy, even when IL-2 failed [
31,
32]. Clinical trials with IL-15 as monotherapy have recently been initiated, investigating its therapeutic potential. When looking at the lymphocyte subsets in blood, both γδ T cell proliferation and activation were observed after IL-15 administration.
These results stressed the need for a detailed characterization of the sheer effect of IL-15 on untouched and isolated γδ T cells. The latter will be discussed in the first part of this paper, followed by the translation of the results into a γδ T cell expansion protocol for adoptive transfer.
Methods
Ethics statement and cell material
This study was approved by the Ethics Committee of the Antwerp University Hospital (UZA; Edegem, Belgium) under the reference number B300201419756. Experiments were performed using buffy coats derived from healthy volunteer whole blood donations (supplied by the blood bank of the Red Cross, Mechelen, Belgium) and blood samples from patients with AML obtained from the hematological division of the UZA (Table
1). PBMC were isolated by Ficoll density gradient centrifugation (Ficoll-Paque PLUS; GE Healthcare, Diegem, Belgium). Untouched γδ T cells were isolated from PBMC using the EasySep™ Human Gamma/Delta T Cell Isolation Kit (Grenoble, France), according to the manufacturer’s instructions. The purity of the γδ T cells was on average 94 % (min-max; 89–98 %). Isolated γδ T cells (1 × 10
6 cells/mL) were cultured for 5 days in Iscove’s Modified Dulbecco’s Medium (IMDM, Life Technologies, Merelbeke, Belgium), supplemented with 10 % fetal bovine serum (FBS, Life Technologies), IPP (30 μg/mL; tebu-bio, Le-Perray-en-Yvelines, France), and IL-2 (100 IU/mL; Immunotools, Friesoythe, Germany) or IL-15 (12.5 ng/mL; Immunotools). The Burkitt’s lymphoma tumor cell line Daudi was kindly provided to us by the laboratory of Prof. Kris Thielemans (Free University of Brussels, Brussels, Belgium), and the multiple myeloma cell line U266 was a gift from Prof. Wilfred Germeraad (Maastricht University Medical Center, Maastricht, the Netherlands).
Table 1
AML patient characteristics
A | f | 51 | AML-nos | R1 | 6.5 | 16.2 | WT1+
| nl |
B | m | 51 | AML-rga | Dx | 50 | 64.4 | NPM1+, WT1+
| nl |
C | m | 62 | AML-nos | Dx | 2 | 72 | ASXL1+
| Trisomy 8 |
D | m | 61 | AML-nos | Dx CR1 | 47.8 0 | 79.8 0.6 | Negative | Deletion 17p |
E | f | 76 | AML-mds | Evolution from MDS | 18 | ND | ND | ND |
F | m | 52 | AML-rga | Dx | 92.5 | 94.5 | WT1+, NPM1+, FLT3-ITD+
| inv(3)(q21q26) |
G | f | 62 | AML-nos | CR1 | 0 | 0.8 | ASXL1+
| trisomy 8 |
Proliferation assay
To test the ability of IL-2 and IL-15, in combination with IPP, to induce γδ T cell proliferation, a 5,6-carboxyfluorescein diacetate succinimidyl ester (CFSE; Invitrogen, Merelbeke, Belgium) flow cytometry-based proliferation assay was performed with isolated γδ T cells. Unstimulated CFSE-labeled γδ T cells served as negative control. After 5 days, cells were stained with LIVE/DEAD® Fixable Aqua Stain (Life Technologies), CD56-PE (Becton Dickinson (BD); Erembodegem, Belgium), CD3-PerCP-Cy5.5 (BD), and γδ T cell receptor (TCR)-APC (Miltenyi) and analyzed using a FACSAria II cytometer (BD). γδ T cell proliferation was assessed by quantifying the percentage of proliferating (CFSE-diluted) cells within the viable (LIVE/DEAD−) CD3+γδTCR+ gate.
Expansion protocol of γδ T cells (for adoptive transfer)
PBMC were resuspended in Roswell Park Memorial Institute (RPMI) supplemented with 10 % heat-inactivated human AB serum (Invitrogen, Merelbeke, Belgium), zoledronate (5 μM; Sigma-Aldrich, Diegem, Belgium), IL-2 (100 IU/mL), and/or IL-15 (100 IU/mL) at a final concentration of 1 × 106 cells/mL. Cell cultures were maintained at a cell density of 0.5–2 × 106 cells/mL and were replenished every 2 to 3 days by adding IL-2/IL-15-supplemented medium. Phenotypic and functional assays were performed on cells harvested at least 14 days after first stimulations.
Immunophenotyping
Freshly isolated and 5-day proliferated γδ T cells were membrane-stained with the following monoclonal antibodies; γδ TCR-FITC (Miltenyi), CD56-PE (BD), CD69-PE (BD), and HLA-DR-PE (BD). Propidium iodide (PI; Life Technologies) was added to exclude dead cells from phenotypic analysis. Data acquisition was performed on a FACScan multiparametric flow cytometer (BD). Phenotypic characterization of γδ T cells was examined pre- and post-expansion, using CD27-FITC (BD), CD69-FITC (BD), CD56-PE (BD), CD80-PE (BD), CD45RA-PE-Cy7 (BD), CD28-PerCP-Cy5.5 (BD), CD16-PB (BD), CD86-V450 (BD), γδ TCR-APC (Miltenyi), and HLA-DR-APC-H7 (BD). Live/Dead® Fixable Aqua Stain was used to distinguish viable from non-viable cells. Data were acquired on a FACSAria II flow cytometer (BD). Corresponding species- and isotype-matched antibodies were used as controls.
Cytokine production
γδ T cell cultures were set up as described above. After 5 days of proliferation, cell-free supernatants were harvested and stored at −20 °C before analysis. Samples were assessed by using enzyme-linked immunosorbent assay (ELISA) for the presence of TGF-β (eBioscience, Vienna, Austria) and by using electrochemiluminescence immunoassay (ECLIA; Meso Scale Discovery (MSD), Rockville, MD, USA) for the presence of IFN-γ, TNF-α, IL-5, IL-10, and IL-17. Cytokine measurements were also performed on supernatant of γδ T cell cultures stimulated for an additional 4 h with the tumor cell lines Daudi and U266 at an effector-to-target (E:T) ratio of 5:1.
Intracellular staining
After 14 days of γδ T cell expansion, IFN-γ and TNF-α production was measured using a flow cytometric-based intracellular staining assay. Measurements were also performed after an additional hour of stimulation with Daudi or U266 cells (E:T ratio = 5:1). Brefeldin A (Golgi-Plug 1 μL/mL; BD) was added to the different conditions (1 × 106 cells/mL) and incubated for 3 h at 37 °C/5 % CO2. γδ T cells were then washed and incubated with Live/Dead® Fixable Aqua Stain, CD3-PerCP-Cy5.5 (BD) and γδ TCR-APC (Miltenyi) for 30 min at 4 °C. Subsequently, cells were fixed and permeabilized, using the Foxp3/Transcription Factor Staining Buffer Set (eBioscience), according to the manufacturer’s instructions. Intracellular staining antibodies (IFN-γ-FITC and TNF-α-PE-Cy7, BD) or the corresponding isotype control were added and allowed to bind for 1 h at 4 °C.
Cytotoxicity assay
A flow cytometry-based lysis assay was performed in order to determine the killing activity of γδ T cells against the tumor cell lines Daudi and U266. Tumor cells were labeled prior to co-culture with PKH67 Green Fluorescent Cell Linker dye (Sigma-Aldrich), according to the manufacturer’s protocol, and subsequently co-cultured with γδ T cells at different E:T ratios (1:10, 1:5, 1:1, 5:1, and 10:1). After 4 h, cells were acquired on a FACSAria II flow cytometer following staining with annexin V-APC (BD) and PI. Killing was calculated based on the percentages of viable (annexin V−/PI−) cells within the PKH67+ tumor cell population using the following equation: % killing = 100 − [(% viable tumor cells with γδ T cells/% viable tumor cells without γδ T cells) × 100].
Statistics
Flow cytometry data were analyzed using FlowJo (v10; Treestar, Ashland, OR, USA). GraphPad Prism software (v5.0; San Diego, CA, USA) was used for statistical calculations and artwork. Shapiro-Wilk normality test was performed to ascertain the distribution of the data. p values <0.05 were considered statistically significant. All data are depicted as means ± standard error of the mean.
Discussion
To date, adoptive immunotherapy using T lymphocytes has been applied for nearly 30 years [
35]. Since then, vast improvements, including the use of chimeric antigen receptor (CAR) or T cell receptor (TCR) gene-modified killer T cells, have been made. This resulting in proven clinical benefit for end-stage cancer patients for which standard therapy failed and the provision of long-term protection in some cases [
36‐
38]. Notwithstanding this progress in the field, one of the major problems encountered with the adoptive transfer of classical αβ T cells, gene-modified or not, is the occurrence of off- and on-target toxicity [
37]. Here, γδ T cells have received recent attention as an alternative cell source for T cell-mediated anticancer therapy [
7]. Namely, both adoptive transfer and in vivo activation/expansion of γδ T cells are safe therapeutic modalities that can result in objective clinical responses in the treatment of cancer [
7,
16]. A major advantage with using γδ T cells is that they are unlikely to cause graft-versus-host disease, allowing them to be generated from healthy donors and given in an allogeneic setting as an “off-the-shelf” therapeutic [
7,
14]. In this study, we explored the beneficial effect of IL-15 on γδ T cells and its use as stimulatory signal in the ex vivo expansion of γδ T cells for adoptive transfer. The strength of IL-15-mediated activation was further validated in γδ T cells originating from AML patients.
A recent study in rhesus macaques showed that continuous administration of IL-15 is, among other, associated with increased numbers of circulating γδ T cells [
39]. This has now been confirmed in the first-in-human trial of recombinant IL-15, whereby IL-15 administration in cancer patients induced both γδ T cell proliferation and activation [
26]. These in vivo findings corroborate with our in vitro results, showing substantial proliferation of isolated γδ T cells upon stimulation with IL-15+IPP, but not with IL-2+IPP. In line herewith, Viey et al. demonstrated γδ T cell expansion out of tumor-infiltrating lymphocytes of renal cell carcinoma patients with bromohydrin pyrophosphate, a synthetic phosphoantigen, in combination with IL-15, while the combination with IL-2 was inefficient [
5]. However, when we sought to extrapolate these results into a clinical 2-week expansion protocol based on IL-15 and zoledronate, inconsistencies in expansion successes were detected. In our hands, the best results were obtained when IL-15 and IL-2 were combined, suggesting that these cytokines have distinct roles in γδ T cell biology. An advantage of IL-15 during expansion could be that IL-15 mediates homeostatic competition between αβ T cells, NK cells and γδ T cells by a yet unknown indirect mechanism [
40,
41]. While we detect no difference in the final percentage of γδ T cells in expansion cultures of healthy donors with or without IL-15, this is certainly the case with the cultures of AML patients. In the IL-2+zoledronate cultures, we observed a higher proportion of other blood mononuclear cells, of which the vast majority were αβ T cells. Addition of IL-15 may therefore have counteracted the growth of other cell types such as αβ T cells, constraining the growth of the γδ T cells, preventing cluster formation and subsequently successful γδ T cell expansion [
40,
41]. It is possible that this process is more important when there is only a low percentage of γδ T cells in the starting material or when γδ T cells are less responsive, which is often the case in a malignant setting [
42]. Moreover, γδ T cells of AML patients generally exhibited a higher viability after expansion with IL-15, which was not observed with healthy donors. The question remains whether this improved viability is a result or the cause of the enhanced degree of γδ T cell expansion. Nevertheless, it has been shown that IL-15 is very competent in supporting survival of activated γδ T cells, superior to IL-2 [
43], and T cell survival in general [
44]. This does not mean that IL-2 is by definition redundant, as evidenced by our expansion results. The synergism between IL-2 and IL-15 could lie in the fact that γδ T cells, in response to zoledronate and IL-2, upregulate their expression of
inter alia IL-2Rβ and γ
c [
43].
With regard to the effector/memory state of the γδ T cells after expansion, our results correspond with the available literature concerning γδ T cells expanded with IL-2 and zoledronate [
25,
45]. In particular, the majority of expanded γδ T cells from healthy donors and patients showed for both expansion protocols a predominant effector memory type and, to a lesser extent, a central memory phenotype. Hence, the addition of IL-15 had no apparent influence on the effector/memory state. This is encouraging, since increased effector memory γδ T cells are reported to correlate with objective clinical outcomes in patients treated with zoledronate and IL-2 [
46]. On the other hand, adoptive transfer with αβ T cells has shown that it may also be important to preserve some less differentiated T cell subsets within the infused T cell product, to ensure T cell expansion and potentially long-term T cell persistence [
47‐
49]. Given the presence of γδ T cells with a central memory phenotype upon expansion, our culture protocol fulfills the above-mentioned necessity. Moreover, contributing to the rationale of implementing IL-15 in expansion protocols, it has recently been shown that IL-15 instructs the generation of human memory stem T cells, a T cell subset with superior antitumor responses [
50], from naive precursors [
49,
51].
When looking in detail at the phenotype of cultured γδ T cells, two findings stand out from our results. First, incubation with IL-15+IPP led to a significant upregulation of CD56 relative to IL-2+IPP stimulated and unstimulated γδ T cells. In addition to the stringent association of CD56 with NK cells, CD56 has also been detected on other lymphoid cells, including γδ T cells and activated CD8
+ T cells [
52‐
55]. Moreover, CD56 in the human hematopoietic system is not restricted to lymphoid cells. Both CD56
+ plasmacytoid dendritic cells (DCs) and myeloid DCs, including our IL-15 DCs, feature cytotoxic activity, in addition to serving classical DC functions [
56]. In this context, it has been presumed that CD56 is associated with activated/cytotoxic effector immune cells [
52,
54,
57‐
59]. This would implicate that IL-15 stimulation raises stronger cytotoxic γδ T cells, as effectively confirmed by the enhanced killing of tumor cells in our experiments. Moreover, CD56 expression on γδ T cells following expansion was enhanced as well. Secondly, upon expansion, an increased expression of the co-stimulatory molecules CD80 and CD86 was observed. It has been shown that activated γδ T cells are able to acquire a professional antigen-presenting cell function, expressing high levels of co-stimulatory molecules [
60]. This function may further boost the generation of a potent and long-lasting immune response.
In view of the functionality of γδ T cells, our experiments clearly show that IL-15, in comparison with IL-2, has a superior Th1 polarizing effect on γδ T cells and markedly strengthens the γδ T cell cytotoxic capacity. These results are supported by in vivo data where intestinal γδ T cells of IL-15
−/− knockout mice only produced small amounts of IFN-γ upon stimulation and showed significantly lower cytotoxicity against target cells as compared to wild-type mice [
61]. This indicates that IL-15-mediated signals are indeed indispensable for the development of potent antitumor functions. Furthermore, IL-15 has proved itself superior at maintaining effector functions of already expanded and activated γδ T cells, evidenced by a higher IFN-γ production and CD107a expression after stimulation with zoledronate pre-treated Daudi cells [
43]. In addition, within the context of neonatal immunity, it has been shown that expansion of cord blood γδ T cells with alendronate and IL-15 gave rise to γδ T cells capable of strong protective immune responses [
62]. Although the differences between the effects of IL-2 and IL-15 were subtle, higher expression of among other granzyme B and perforin was detected after activation with IL-15 [
62]. All these data therefore point towards the fact that IL-15 is a pivotal signal to generate more powerful effector γδ T cells. This has recently been substantiated by Ribot et al., identifying the MAPK/ERK-mediated IL-2/IL-15 signaling as the major functional differentiation pathway of human γδ T cells towards antitumor (cytotoxic type 1) effector cells [
63].