Addition of Interleukin-21 for Expansion of T-Cells for Adoptive Immunotherapy of Murine Melanoma

We previously demonstrated that interleukin (IL)-7/15 was superior to IL-2 for expansion of T cells in vitro for adoptive immunotherapy. We sought to ascertain whether IL-21 would further improve yield and therapeutic efficacy of T cells in culture. Naïve T cell receptor (TcR) transgenic splenocytes or antigen-sensitized lymph node cells were harvested from PMEL-1 mice and exposed to bryostatin-1 and ionomycin (B/I) for 18 h. Cells were then cultured in IL-2, IL-21, IL-7/15 or IL-7/15/21 for six days. Harvested cells were analyzed by flow cytometry and used to treat C57Bl/6 mice injected intravenously with B16 melanoma. Lungs were harvested and metastases counted 14 days after treatment. Culturing lymphocytes in IL-7/15/21 increased expansion compared to IL-2 or IL-7/15. IL-21 and IL-7/15/21 increased CD8+ cells compared to IL-2 or IL-7/15. IL-21 preferentially expanded a CD8+CD44−CD62L+ T “naïve” population, whereas IL-7/15/21 increased CD8+CD44+CD62Lhigh central-memory T cells. T cells grown in IL-7/15/21 were more effective at reducing metastases than IL-2. The addition of IL-21 to IL-7/15 induced greater expansion of lymphocytes in culture and increased the yield of CD8+ T central-memory cells vs. IL-7/15 alone. This may have significant impact on future clinical trials of adoptive immunotherapy, particularly for generating adequate numbers of lymphocytes for treatment.


Introduction
Adoptive immunotherapy (AIT), the infusion of ex vivo expanded lymphocytes, has been extensively studied in animals and humans [1][2][3][4][5]. Although this therapy has demonstrated promising results in multiple murine tumor models, a regimen that optimizes both lymphocyte expansion as well as tumor regression for human therapy remains elusive. AIT takes advantage of ex vivo activation and expansion of T cells away from the suppressive in vivo tumor environment and allows for "re-programming" of the immune cells to optimize their functional status. It also allows for additional treatment of the host (e.g., host lymphocyte depletion) prior to the re-introduction of the selected cells, which may decrease immunosuppression, and optimize trafficking and/or proliferation of the infused cells.
We have shown that T cells from both naïve splenocytes and tumor antigen-sensitized draining lymph nodes (DLN) could be expanded with exposure to interleukins (IL)-7 and 15 after activation with bryostatin and ionomycin (B/I) to significantly greater numbers than the current standard approach using IL-2 alone [6]. These T cells were also able to cure melanoma metastases as effectively as, and sometimes better than, T cells grown in IL-2 [6]. Bryostatin-1 is a macrocyclic lactone derived from Bulgula neritina, a marine invertebrate. In culture, bryostatin activates protein kinase C, while ionomycin increases intracellular calcium [7][8][9]. When combined, these mimic signaling through the CD3/TcR complex and promote activation and proliferation of T cells [6,[10][11][12][13][14][15][16]. B/I selectively activates CD62L-T cells, which represent the "sensitized" T cells capable of anti-tumor activity and is unique to B/I activation compared to anti-CD3 activation [7,15]. These T cells are similar to CAR T cells in the sense that all their CD8 T cells carry a transgenic T cell receptor that recognizes a single epitope present on B16 melanoma cells. We have also shown that this method of stimulation, combined with alternate cytokines, re-programs human T cells and NK cells, making them resistant to suppression by myeloid derived suppressor cells [10]. Phenotypically, whereas IL-2 expansion has been shown to lead to the induction of regulatory T cells and cause T cell activation induced cell death, IL-7/15 has been shown by this lab and others to support preferential differentiation of CD8+ T cells towards a central memory (TCM) phenotype [6,14,17]. This central memory phenotype has been suggested in multiple studies to be more effective at inducing tumor regression than terminally differentiated effector cells, which are more likely to be selectively expanded when T cells are grown in IL-2 [17,18]. Although using IL-2 to stimulate the proliferation of anti-tumor T cells in culture remains the standard clinical approach, we were encouraged by our results with IL-7/15 to add or substitute IL-21 to our expansion regimen. IL-21 is the most recently identified member of the family of cytokines that share the common gamma chain cytokine receptor with IL-2, IL-7 and IL-15 [19,20]. Preliminary experiments performed by other groups demonstrated that IL-21 has potent immunomodulatory effects on T cells as well as NK cells [21][22][23][24]. Both IL-15 and IL-21 have been shown in multiple experiments to enhance the in vivo anti-tumor effects of CD8+ T cells and in some cases to potentiate tumor regression [24][25][26][27][28][29][30]. Because of the promising results seen with IL-21 to date, we endeavored to discover whether B/I and IL-21 exposure alone or in combination with IL-7/15 would increase the expansion of naïve or antigen-sensitized T cells, and whether it would increase anti-tumor activity. In addition, the T cell phenotype stimulated by exposure to IL-21 has varied in studies over the last decade, with some demonstrating increase in TCM cells while others claimed inhibition of this phenotype [19,31,32]. Therefore, we also performed flow cytometry analysis of cells expanded in different cytokines to elucidate which phenotypes were preferentially selected for after exposure to bryostatin, ionomycin and various cytokines.

Comparative Analysis of T Cell Expansion
In repeated experiments, expansion of cells from naïve splenocytes in the IL-7/15 and IL-7/15/21 groups was dramatically higher than for either IL-2 or IL-21. Whereas expansion in IL-2 ranged from 1-to 2.8-fold increase on day 6, cells grown in IL-7/15 expanded from 8.9-to 24. , IL-21 (0.9-to 3.3-fold) and IL-2 (3.7-fold). Again, expansion in IL-7/15 and IL-7/15/21 were significantly greater than in IL-2 or IL-21 (all p < 0.0039), but not significantly different from each other. However, there was a trend in favor of IL-7/15/21 expansion (p = 0.13). It is important to note that when cells were cultured for a total of 14 days, lymphocytes grown in IL-2 not only stopped expanding, but also rapidly began to die and therefore could not be included in expansion data, flow cytometry analysis, or treatment groups.

IFN-Gamma Release ELISA
When B/I pulsed and IL-7/15/21 exposed splenocyte T cells were co-cultured with B16 cells for 24 h, the amount of IFN-γ release was significantly higher than in those splenocyte T cells exposed to IL-2 and co-cultured with B16 cells, averaged over three experiments (p = 0.04) ( Figure 5). T cells exposed to IL-7/15 and co-cultured with B16 cells approached significance compared to IL-2 exposed T cells cultured with B16 cells (p = 0.06), a finding we had demonstrated previously [6]. The amount of IFN-γ produced by IL-7/15 exposed and IL-7/15/21 exposed T cells co-cultured with B16 cells were not significantly different from each other (p = 0.85). Two controls were utilized, a control tumor, LLC, was used to demonstrate that the IFN-γ release from the T cells was specific to exposure to B16, and the cytokine exposed T cells alone, without tumor co-culture. Only the IL-7/15 and the IL-7/15/21 exposed T cells when co-cultured with B16 were significantly different from their LLC and alone controls (p < 0.05), whereas the IL-2 and IL-21 exposed T cells when co-cultured with B16 were not significantly different from their controls (p = 0.24-0.99).  Figure 5. Interferon-gamma ELISA assay. B/I pulsed and IL-2, IL-21, IL-7/15 and IL-7/15/21 exposed spleen-derived T cells were co-cultured with B16 melanoma cells, Lewis lung carcinoma (LLC) cells or media alone for 24 h. The amount of IFN-γ release was significantly higher in the group exposed to IL-7/15/21 and approached significance in the group exposed to IL-7/15 than for spleen-derived T cells expanded in IL-2 or IL-21 and co-cultured with B16 cells, † (vs. IL2+B16) 9 = p < 0.04 (IL-7/15/21), p = 0.06 (IL-7/15).

Treatment of B16 Melanoma Metastases with Lymphocytes Exposed to Different Cytokines
Compared to the control group, mice treated with cyclophosphamide alone or cyclophosphamide plus AIT with the standard dose of spleen-derived lymphocytes (10 million cells/mouse) expanded for 7 days in IL-2, IL-7/15 or IL-7/15/21 exhibited significantly fewer metastatic lung nodules (Figure 6a) (p = 0.0369, p = 0.0330, p < 0.0001, p < 0.0001). At this dose of cells, IL-7/15 and IL-7/15/21 expanded cells were significantly better at curing metastases than cyclophosphamide alone (p < 0.0001, p < 0.0001), whereas IL-2 expanded cells were not more effective than CYP alone (p = 0.49). However, only the IL-7/15/21 expanded cells were significantly better than cyclophosphamide alone for curing metastases at a lower dose of T cells (2 million cells per mouse) (p = 0.002). In addition, when compared to AIT with 10 million cells grown in IL-2 (the current standard), IL-7/15 and IL-7/15/21 expanded cells were significantly better at curing metastases at the same cell dose (10 million) (p < 0.0002, p < 0.0007), with the majority sustaining complete cures (zero metastases seen). In our experiments, we found that AIT with cells expanded in IL-21 alone, whether at 10 million or 2 million cells, was not statistically significantly different from the control or CYP groups (p = 0.27, p = 0.59). A similar experiment was performed with DLN-derived lymphocytes (Figure 6b). Both IL-7/15 and IL-7/15/21 expanded lymphocytes were significantly better than control at the lower dose of AIT (two million cells) at curing melanoma metastases (p = 0.002, p = 0.004). However, at the standard dose of ten million cells both IL7/15 an IL7/15/21 expanded cells were capable of curing metastases significantly better than control, CYP only, and both groups at the two million cell dose (p < 0.02).

Discussion
These results demonstrate that culturing splenocytes and DLN lymphocytes for AIT with IL-7/15/21 results in a much higher yield of cells with equal or greater activity against melanoma metastases than those cultured in either IL-2 (the current standard cytokine for expansion) or even the combination of IL-7+15. Many studies to date have demonstrated that therapy with exogenous IL-21 or with lymphocytes exposed to IL-21 alone are superior to various combinations of cytokines [24,29,[32][33][34]. Others have shown an improved effect of IL-21 combined with IL-2, IL-15 or IL-7 alone [2,22,25,31,35], However, to our knowledge this is the first study examining the effect of IL-7, -15 and -21 concurrently on in vitro lymphocyte expansion. Klebanoff et al. [1] recently reviewed which factors were associated with the most effective adoptive immunotherapy. Of the factors examined, absolute number of infused cells was strongly correlated with the magnitude of tumor regression. They also found that the least differentiated cells demonstrated the most robust anti-tumor activity. In their experience, both T central memory (TCM) cells and T memory stem cells (TSCM) led to significant delay in tumor growth compared to untreated controls or mice receiving T effector memory (TEM) cells. They found that TSCM cells caused a more sustained reduction in tumor growth compared to TCM cells. They also found that no single gamma chain cytokine, when administered exogenously, was significantly better at augmenting anti-tumor function than any other single cytokine. They go on to suggest that it is difficult to uncouple the process of cell expansion and differentiation status in that the longer the cells are expanded ex vivo, the more differentiated they become, and therefore, using longer durations of in vitro expansion to increase total cell numbers sacrifices more effective TCM cells in favor of less effective terminally differentiated TE cells. However, our results demonstrate that expanding splenocytes in IL-7/15/21 after activation with B/I results in a dramatic increase in the total numbers of cells, with the actual yield of CD8+ T cells capable of numbering over one billion after only 6 days of expansion from an average starting cell number of 30 million. Of these CD8+ T cells, over half were shown to be TCM cells (~625 million). Splenocytes expanded in IL-7/15/21 were then subsequently found to cure metastases in vivo without vaccine or exogenous cytokine administration, even at low doses of cells. This finding demonstrates that even with longer times in culture and prolific expansion, lymphocytes grown in IL-7/15/21 are capable of achieving high numbers of total cells, the majority of which are TCM cells that have a significant anti-tumor effect in vivo. In addition, this also demonstrates that the TCM cells needn't be sorted to a pure population for effective anti-tumor function. However, if desired, the total number of TCM cells produced by expansion with IL-7/15/21 is such that a significant number could be enriched for therapeutic use. However, not requiring a sorted population renders adoptive immunotherapy more clinically feasible.
We consistently observed that splenocytes expanded in IL-2 or IL-21 alone did not expand well, were not effective at curing B16 metastases compared to controls, and did not produce IFN-γ when exposed to B16 tumor cells to a significantly different level compared to controls. These IL-2 and IL-21 exposed splenocytes, perhaps unsurprisingly, were the least efficacious at curing or slowing tumor growth. Overall, we found that the IL-7/15/21 cultured splenocytes had the greatest expansion, produced the greatest percentage of CD8+ T cells, the greatest number of CD8+ TCM cells, had the highest amount of IFN-γ production when co-cultured with B16 cells and were most efficacious at curing metastases in vivo. The actual yield of CD8+ T cells and CD8+ TCM cells was significantly higher in IL-7/15/21 cultured splenocytes compared to IL-7/15 splenocytes, which we had previously shown to be superior to IL-2 cultured T cells [6].
In multiple papers, it has been suggested that exposure to IL-21 alone causes lymphocytes to remain in or differentiate into a minimally differentiated phenotype (CD62L+CD44−), and that these lymphocytes have greater anti-tumor capacity relative to cells expanded in other gamma chain cytokines [25,34,36].
Gattinoni et al. [37] describe lymphocytes expanded in IL-21 alone, demonstrating a CD62L+CD44− phenotype described as TSCM-like cells. They explain that because of the unique ability of IL-21 among gamma chain cytokines to sustain STAT3 activation, this maintains a "TSCM-like state that is associated with high proliferative potential and long-term T cell survival" [37]. Unfortunately, in our experiments, although lymphocytes expanded in IL-21 indeed were enriched for the CD62L+CD44− phenotype, this did not translate into either higher proliferative capacity in vitro nor greater anti-tumor efficacy in vivo. This may be because our method of activation with B/I does not support development of TSCM-like cells after exposure to IL-21, or because TSCM-like cells produced by exposure to IL-21 in culture are not as effective as has been demonstrated in earlier studies.

Mice
Virus-free C57Bl/6 mice (National Cancer Institute, National Institutes of Health, Bethesda, MD, USA) between 8 and 12 weeks of age, caged in groups of 6 or fewer, were provided with food and water ad libitum. T cell receptor (TcR) transgenic PMEL-1 mice, with T cell receptors (TcR) specific for the peptide KVPRNQDWL, from the shared melanocyte/melanoma differentiation antigen gp100, were produced from breeding pairs obtained from Jackson Laboratories (Bar Harbor, Maine). All guidelines of the Virginia Commonwealth University Institutional Animal Care and Use Committee, which conform to the American Association for Accreditation of Laboratory Animal Care and the U.S. Department of Agriculture recommendations for the care and humane experimental use of animals, were followed.

Lymphocyte Activation and in Vitro Expansion
Draining lymph nodes (DLNs) or spleens were harvested and dispersed into single cell suspensions in complete RPMI media at 1 × 10 6 cells/mL. Splenic mononuclear cells were enriched by Lympholyte-M density gradient centrifugation and residual erythrocytes were removed by suspension in ammonium chloride potassium (ACK). These cells were then activated by incubation with 5 nM bryostatin-1 (provided by the National Cancer Institute, Bethesda, MD, USA) and 1 μM ionomycin (Calbiochem, San Diego, CA, USA) (B/I) and 80 U/mL of rIL-2 (Chiron, Emeryville, VA, USA) at 37 °C for 18 h. Cells were washed three times with warm complete RPMI and re-suspended at 1-2 × 10 6 cells/mL with either 40 U/mL of rIL-2, 5 ng/mL of IL-21 (BD Biosciences, San Jose, CA, USA), 5 ng/mL each of IL-7 (BD Biosciences, San Jose, CA, USA) and IL-15 (BD Biosciences, San Jose, CA, USA), or 5 ng/mL each of IL-7, IL-15 and IL-21. The cells were allowed to proliferate in culture for an additional 6-7 days and were split every 2-3 days in order to maintain 1× 10 6 -2 × 10 6 cells/mL concentration and additional cytokine was added at 40 U/mL or 5 ng/mL with each split.

Adoptive Immunotherapy
For B16 melanoma lung metastases, mice were inoculated with 250,000 B16 melanoma cells intravenously (IV) 4 days prior to AIT. Mice were pretreated on the day prior to AIT with 100 mg/kg of intraperitoneal (IP) cyclophosphamide (CYP, Mead Johnson, Princeton, NJ, USA). AIT consisted of IV infusion of lymphocytes from cultures at either 2 or 10 million cells/mouse harvested 7 days after activation in vitro. No systemic cytokines or vaccines were administered to the recipient mice. In all AIT experiments, mice were euthanized by CO2 inhalation 14 days after tumor cell infusion and lungs were removed, fixed in 10% formaldehyde, and melanoma lung nodules were counted under a dissecting microscope, as previously described [38].

Flow Cytometry
T cells isolated from DLN and spleens were stained with a panel of antibodies on day 0 (immediately after B/I activation) and on day 6 after activation in vitro. These stained cells were analyzed by multicolor flow cytometry for surface markers on a FACSAria Canto flow cytometer. Fluorescently labeled antibodies directed against the following markers were obtained from Biolegend and eBiosciences: CD4, CD8, CD62L, and CD44. Appropriate isotype controls were used in all cases. T cell subsets analyzed were T effector (TE) CD44+CD62L−, T effector memory (TEM) CD44+CD62Llow, T central memory (TCM) CD44+CD62Lhigh and "T naïve" ("TN") CD44-CD62L+, as described in earlier studies and reviews [39][40][41].

IFN-γ ELISA
Splenocytes were cultured with B/I and cytokines for 7 days as described above. The cells were then harvested, washed and cultured in complete medium at a ratio of 10:1 with irradiated B16 tumor cells or Lewis lung carcinoma (LLC) irradiated cells, as a control, for 24 h. Splenocytes alone, tumor cells alone and media alone were also cultured concurrently as controls. Supernatants were then collected and stored at −80 °C until assay. IFN-γ was detected using a Mouse IFN-γ ELISA kit (BD Pharmingen, San Jose, CA, USA) according to the manufacturer's protocol and performed in duplicate.

Statistical Analysis
At least five mice were included in each experiment. Each outcome was summarized with basic descriptive statistics such as mean and standard deviation for each cytokine. Repeated measures analysis of variance through a linear mixed model was used to compare the various cytokine conditions in fold increase cell numbers on day 6 from baseline, in each T cell phenotype (including both percentage and total number of cells of CD4 and CD8, and TE, TEM, TCM, and TN), in number of lung melanoma metastases, and in ELISA. Pairwise comparisons of the various cytokines were made and tested at type I error of 5% by a t-test with the random measurement errors estimated by the linear mixed model. Normality was checked for each analysis. Natural logarithm was used in the data of fold increase of T cell expansion, ELISA, and percentage and total numbers of T cell phenotypes, and square root was used in number of lung melanoma metastases for normality. SAS 9.2 was used for all analyses.

Conclusions
In conclusion, we demonstrate here that lymphocytes expanded in the triple combination of IL-7/15/21 have a large proliferative capacity, significantly greater than the current standard of IL-2 expansion, and are capable of producing billions of CD8+ T cells and hundreds of millions of CD8+ TCM cells from a starting population of only tens of millions of lymphocytes total. These IL-7/15/21 expanded lymphocytes are capable of inducing regression of metastatic B16 melanoma, even at low doses of cells. In addition, these effects are seen when the total expanded lymphocyte population is utilized for therapy, without the need for sorting a specific phenotypic subset of T-cells and without utilizing exogenous cytokine or vaccinations in vivo. These results show great promise for producing large numbers of effective anti-tumor T cells for future human trials. Also, although demonstrating similar phenotypic characteristics to TSCM-like cells, we found that lymphocyte expansion in IL-21alone improved neither their ability to proliferate in vitro nor their ability to cause tumor regression in vivo. It is therefore unclear whether the CD62L+CD44− IL-21 expanded lymphocytes are indeed TSCM-like cells.