IL-15-induced lymphocytes as adjuvant cellular immunotherapy for gastric cancer

Objectives To test the antitumor potential of lymphocytes transferred via adoptive cell therapy (ACT) in a mouse model of human gastric cancer (GC), and to evaluate the clinical efficacy and safety of combining lymphocytes as adjuvant therapy with first-line chemotherapy in patients with GC. Methods We constructed a human GC xenograft model in sublethally irradiated 6–8-week-old male NCG mice. MKN-45 cells (1 × 106 cells/mouse) were subcutaneously injected into mice’s flanks. After tumors had become palpable, we randomized the mice into control, ACTIL−2, and ACTIL−15 groups. Human lymphocytes were then injected into mouse tail veins. In addition, 63 human patients with histologically or cytologically confirmed stage III–IV GC randomly received S-1 + oxaliplatin + ACTIL−15 (combination therapy group) or S-1 + oxaliplatin (chemotherapy group). Results In the mouse study, treatment with ACTIL−15 cells inhibited tumor growth on adoptive transfer, and mice that received ACTIL−15 cells had significantly longer survival rates (p < 0.05, ACTIL−15 vs. ACTIL−2). In the human study, the median survival rate of patients in the combination therapy group was 472 days (95% confidence interval [CI], 276–668 days), whereas that of patients in the chemotherapy group was 266 days (95% CI, 200–332 days; p < 0.05). Eleven percent (7/63) of patients had adverse reactions, but these reactions did not interfere with treatment. Conclusion Adoptive transfer of ACTIL−15 cells in a mouse model of GC and in patients with advanced GC treated with S1 + oxaliplatin improved survival rates in both, with an acceptable safety profile.


Introduction
Gastric cancer (GC) is challenging to treat. Few patients are eligible for extensive surgery, and median survival is 3-5 months for those with unresectable disease [1]. Clinical trials with various treatment regimens have improved the prognosis, but median survival remains < 1 year [2][3][4][5]. No standard third-line therapy is available for patients with advanced GC who have not responded to ≥ 2 lines of chemotherapy. Therefore, a strategy to improve the efficacy of existing chemotherapy regimens without increasing their toxicity is urgently needed.
The oral anticancer drug fluoropyrimidine (S-1), combined with 1 M tegafur, 0.4 M 5-chloro-2, 4-dihydroxypyrimidine, and 1 M potassium oxonate, is effective against gastrointestinal (GI) cancer cells in vivo [6,7]. The standard for first-line treatment of unresectable advanced or metastatic gastric/gastroesophagealjunction cancer is either oral fluoropyrimidine (e.g., capecitabine or S-1) or capecitabine + cisplatin or oxaliplatin [8][9][10]. Apart from the effects of chemotherapy on tumor cell replication, it has been proposed that chemotherapy produces an antitumor effect by modulating the immune system [11,12]. For example, oxaliplatin can induce immunological death of tumor cells and enhance the efficacy of immunological agents [13].
Novel agents that can improve survival in patients with GC are needed. Immunotherapeutic agents targeting the adaptive immune response have shown promising results in several cancers [14]. Adoptive cell therapy (ACT), the administration of ex vivo-expanded autologous or allogeneic T lymphocytes, can help control tumor growth. However, the therapeutic efficacy of ACT in vivo is affected by the number of cells with antitumor properties [15]. The optimal lymphocytes for ACT are central-memory (T CM )-like populations with higher expression of lymphoid-homing molecules than effector memory (T EM )-like populations [16]. A lack of interleukin-15 (IL-15) has been associated with poor proliferation of T CM -like adoptively transferred cells, a result suggesting that IL-15 is critical in this process [17]. In addition, the antitumor effects of IL-15 have been documented in human GC [18], GC metastases in the liver [19], and human GC xenografts [20]. Motivated by these studies, we tested the ability of IL-15 to promote the growth of lymphocytes with the T CM -like phenotype. To that end, we generated lymphocytes cultured in the presence of either IL-15 or the conventional IL-2.
We evaluated the antitumor potential of ACT lymphocytes in tumor-bearing NCG mice that had received grafts of human gastric-carcinoma cells. We then assessed the clinical efficacy and safety of combining lymphocytes as adjuvant therapy with first-line chemotherapy in human patients with GC.

Chemicals, cell lines, and animals
The human cell lines AGS and NCI-N87 (subsequently referred to as N87) are Lauren intestinal-type gastric-adenocarcinoma (GAC) cell lines, and MKN-45 is a Lauren diffuse-type GAC cell line. We obtained all three lines from the America Type Culture Collection (ATCC; Manassas, VA, USA). Cells were maintained in Roswell Park Memorial Institute 1640 (RPMI-1640) medium supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, 100 mg/mL streptomycin, and 2 mmol/L of L-glutamine ("regular media"). Cancer cell lines were passaged within 6 months from time of receipt. We followed the guidelines of the United Kingdom Coordinating Committee on Cancer Research [21].

Patients
Sixty-three patients with histologically or cytologically confirmed stage III-IV GC participated in the study. The Ethical Review Board of the Medical Ethics Committee of the National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences, and Peking Union Medical College (PUMC; Beijing, China) approved the study protocol. All participants gave written informed consent in accordance with the Declaration of Helsinki. After enrollment, all patients provided their complete medical histories and underwent physical examinations. Their demographics and baseline characteristics are summarized in Table 1. Patient selection criteria were (1) age 18-80 years; (2) expected survival ≥ 3 months; (3) Karnofsky performance status score > 40%; (4) peripheral-blood white blood cell (WBC) count ≥ 4 × 10 9 /L; (5) platelet count ≥ 10 × 10 10 /L; (6) serum aspartate aminotransferase (AST)/alanine aminotransferase (ALT) values below the normal upper limit; (7) no cardiac arrhythmias, congestive heart failure, or severe coronary-artery disease; and (8) not pregnant or lactating.

ACT cell generation
Peripheral-blood mononuclear cells were obtained using Ficoll density centrifugation [22]. We resuspended the cells at a density of 3 × 10 6 cells/mL in X-VIVO 15 medium (Lonza, Basel, Switzerland) supplemented with 10% heatinactivated autologous plasma. We primed the cells by adding 1000 U/mL interferon gamma (IFN-γ) on day 0 and 100 ng/mL anti-cluster of differentiation 3 (anti-CD3) antibody (MACS GMP CD3 pure; Miltenyi Biotech, Bergisch Gladbach, Germany) plus 500 U/mL IL-2 on day 1. On day 4 of culture, cell density was adjusted to 1 × 10 6 cells/mL. We added fresh medium with IL-2 (500 U/mL) or IL-15 (50 ng/ mL) every 3 days. On day 14 of culture, ACT cells were harvested and used for analysis. We designated cells cultured in IL-15 and IL-2 as ACT IL−15 and ACT IL−2 cells, respectively.

In vitro cytolytic assays and cytokine release
We performed a europium release assay for in vitro cytotoxicity analysis as previously described [22]. Target cells (MKN-45, AGS, or N87) were co-cultured with ACT cells in triplicate at effector-to-target cell ratios of 50:1, 10:1, 5:1, and 1:1 in 96-well U-bottom culture plates (Nunc A/S, Roskilde, Denmark). We collected the supernatant from each well and co-incubated it with europium solution (PerkinElmer Corp., Turku, Finland) in flat-bottom 96-well plates (Nunc). Fluorescence was measured with a time-resolved fluorometer (1420-018 Victor; PerkinElmer Corp., Waltham, MA, USA). We calculated the percentage of specific cytolysis for each well as previously described [22] and the mean ± standard deviation (SD) for each duplicate or triplicate.

Tumor xenograft model
To establish the human GC xenograft model in mice, we housed 6-8-week old NCG mice (Institute of Zoology, Chinese Academy of Sciences, Shanghai, China) under specific-pathogen-free (SPF) conditions. Mice were irradiated with 200 cGy for 24 h. We subcutaneously injected MKN-45 cells (1 × 10 6 cells/mouse) into their flanks. When tumor nodules became palpable (7 days after injection), mice were randomized into three groups: blank control (physiological saline injected into the tail vein), ACT IL−2 (ACT IL−2 cells injected into the tail vein), and ACT IL−15 (ACT IL−15 cells injected into the tail vein). We euthanized the animals on day 32 after transplantation of malignant cells or monitored them for survival. Mice that showed signs of physical abnormalities or poor health were sacrificed by carbon dioxide asphyxiation followed by cervical dislocation. Toxicity was defined as ≥ 20% body weight (BW) loss or toxic death. We measured BW 1 × / week and calculated survival from time of tumor cell injection until death. Animals were euthanized when they exhibited a lack of a righting reflex or lack of response to stimuli to avoid suffering. Tumor volume was measured every 2-3 days in a blinded fashion and calculated as [length × width 2 ] / 2.

Immunohistochemistry
We harvested and fixed tumors for 24 h with 10% buffered formalin before embedding them in paraffin. Serial Sects. 5 μm thick were cut for histological analysis and stained with hematoxylin and eosin (H&E) in accordance with standard procedures. For immunohistochemistry (IHC), we incubated sections with anti-human CD3 antibody (1:200) or anti-human proliferating cell nuclear antigen (PCNA) antibody (1:50) and then performed detection using a Dako ChemMate EnVision System (Dako ApS, Glostrup, Denmark) for 30 min. Staining was visualized using diaminobenzidine, and sections were counterstained with hematoxylin.

TUNEL staining
We performed terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining using an ApopTag kit (Oncor, Inc., Cambridge, MA, USA). Briefly, sections were deparaffinized and digested in proteinase K (20 µg/ml) for 30 min at room temperature (RT). We blocked endogenous peroxidase with 0.3% hydrogen peroxide for 20 min and then left the sections in terminal deoxynucleotidyl transferase (TdT) buffer for 15 min at RT. After washing, sections were incubated with a peroxidase-conjugated antidigoxigenin antibody for 30 min at RT, and the reaction was detected using 3,3'-diaminobenzidine (DAB) chromogen. Slides were counterstained with Mayer's hematoxylin.

Enzyme-linked immunosorbent assay
We cultured single-cell suspensions from tumors in the X-VIVO 15 medium at 37 °C for 6 h. Secretion of IFN-γ, tumor necrosis factor alpha (TNF-α), IL-4, and IL-10 by ACT cells was measured in the supernatant of the culture media using enzymelinked immunosorbent assay (ELISA) kits (R&D Systems, Inc., Minneapolis, MN, USA) per manufacturer's instructions.

Statistical analysis of animal studies
We analyzed differences in mean relative tumor volume (RTV) between the treated and control groups on day 25 using a twotailed Student's t test. All statistical analyses were performed using SPSS software version 20.0 (IBM Corp., Armonk, NY, USA). p < 0.05 was considered statistically significant.

Human study design
Using an interactive web response system, we randomly assigned participants (1:1) to receive S-1 + oxaliplatin + ACT IL−15 cells (combination therapy group) or S-1 + oxaliplatin (chemotherapy group). Assignment to the treatment groups was balanced according to stratified factors, including Eastern Cooperative Oncology Group (ECOG) performance status and whether the cancer was unresectable or recurrent. A unique random sequence generated before the trial by an independent statistician was sequentially applied to each patient allocation using the biased coin method. We matched the two groups for sex, age at onset of disease, pathology, tumor size, metastases, and stage at the first visit. All patients in both groups received the same chemotherapy component of the treatment, including doses and cycles. S-1 was orally administered at 80 mg/m 2 divided into two daily doses for 12 days, followed by 9 days off. Oxaliplatin was administered intravenously at 130 mg/m 2 over 1-3 h on day 1, and then every 21 days until the disease had progressed or unacceptable toxic effects had developed. We calculated S-1 dose by patient body surface area. Patients in the combination therapy group received ACT IL−15 cell therapy on day 14 and again after the second chemotherapy treatment. Mean lymphocyte count in the ACT IL−15 cell agent was 5.9 × 10 9 cells.

Assessments during treatment and follow-up
We recorded toxicity assessments, compliance with S-1, and blood test results after each treatment cycle. Tumors were assessed after every other cycle. We performed computed tomography (CT) every 8 weeks until the cancer progressed. Therapy was discontinued if the disease progressed, the patient refused, unacceptable toxicity occurred, or the patient died. We assessed tumor status via imaging approximately every 2 months or until death.
Toxicity was evaluated in accordance with the National Cancer Institute (NCI) Common Toxicity Criteria for Adverse Events, version 3.0. We evaluated response to treatment in accordance with the Response Evaluation Criteria in Solid Tumors (www. cancer. gov). Radiographic evidence of response to treatment was independently reviewed. An independent data-monitoring committee oversaw the safety and efficacy of the trial and other aspects of the study conduct.

Statistical analysis of human studies
Statistical analyses were conducted using SPSS version 20.0 and GraphPad Prism version 5 (GraphPad Software, Inc., San Diego, CA, USA). We evaluated differences in demographic and clinical variables between the two groups using the Pearson χ 2 test, while Fisher's exact test was used for categorical variables. The Kaplan-Meier method was used to analyze progression-free survival (PFS) and overall survival (OS). We calculated PFS from time of first treatment to time of first disease progression or last follow-up, and OS from date of first treatment to date of death resulting from any cause or date of last follow-up. All P values were two tailed, and significance was set at p < 0.05.

Phenotypic and functional polarization of cells cultured in IL-15
We evaluated the quality of cell phenotype and compared cell composition at the end of cultivation. No significant differences in cell composition were found between the IL-2 and IL-15 cultures (Fig. 1A). Expression of the activation antigen CD25 on CD3 + CD56 − or CD3 + CD56 + T lymphocytes cultured with IL-15 was significantly higher than on those cultured with IL-2 (Fig. 1B).
ACT IL−15 cells maintained positive staining for the T CM marker. While the T CM markers CD62L and CCR7 were virtually absent on ACT IL−2 cells (0.3%), ACT IL−15 cells (62.1%) retained expression of both markers ( Fig. 2A). We also counted lymphocytes at indicated time  We also assessed the functional activity of the generated lymphocytes. ACT IL−15 and ACT IL−2 cells were added to target cells AGS, N87, and MKN-45 at an effector-totarget ratio of 50:1 and tested in europium release assays. ACT IL−15 cells induced significantly more cytotoxicity against AGS and N87 than did ACT IL−2 cells, whereas the in vitro cytotoxicity of ACT IL−15 cells against MKN-45 cells was not higher than that of ACT IL−2 cells (Fig. 3A). We also evaluated cytokine secretion and found that ACT IL−15 cells significantly increased IFN-γ secretion from AGS and N87 cells but not from MKN-45 cells (Fig. 3B). In sum, ACT IL−15 cells exhibited cytotoxicity against GC cells and had greater cytotoxicity against AGS and N87 than did ACT IL−2 cells.
We considered that ACT IL−15 cells might have differential effects in vivo and in vitro and, therefore, might be effective against MKN-45 cells in vivo. After culture in medium containing IL-2 or IL-15, lymphocytes proliferated extensively and to a similar extent. In three independent experiments, we transferred ACT cells into MKN-45 gastric-carcinoma-bearing mice and then compared lymphocyte tumor infiltration and persistence.
We measured tumor infiltration of adoptively transferred human cells and the secretion of cytokines by ACT cells in tumor sites. ACT IL−15 cells promoted tumor infiltration and increased IFN-γ secretion potential of adoptively transferred lymphocytes (Fig. 4). Moreover, ACT IL−15 cells inhibited the proliferation of MKN-45 cells, as measured by TUNEL staining and PCNA IHC (Supplementary Fig. S1). We next measured tumor volume and survival rate in mice receiving ACT IL−2 or ACT IL−15 cells and compared with these measures in untreated controls. Treatment with ACT IL−15 inhibited tumor growth (Fig. 5A). Moreover, mice that received ACT IL−15 cells had significantly improved survival rates (p = 0.049, ACT IL−15 vs. ACT IL−2 ; Fig. 5B).
Taken together, these data indicated that adoptive transfer of ACT IL−15 cells was preferable to that of ACT IL−2 cells in tumor immunotherapy.

The role of ACT IL−15 in human gastric-cancer immunotherapy
The combination of immunotherapy and chemotherapy has been proposed as a therapeutic strategy with the potential to improve survival rate and prognosis in patients with GC  [23]. Therefore, we assessed the clinical efficacy and safety of ACT IL−15 administered with S-1 + oxaliplatin in patients with advanced GC.

Clinical evaluation of patients with gastric cancer
Seventy-three patients were enrolled at the National Cancer Center/National Clinical Research Center for Cancer/ Cancer Hospital, Chinese Academy of Medical Sciences, and PUMC from November 2014 to February 2019; three were excluded before randomization (Fig. 6). The remaining 70 patients were randomized into the combination therapy (ACT IL−15 + S-1/oxaliplatin) or chemotherapy (S-1/oxaliplatin) group (n = 35 per group). One patient from the combination therapy group and six from the chemotherapy group were removed from analysis due to poor adherence, leaving 34 patients in the combination therapy group and 29 in the chemotherapy group for final analysis (Fig. 6). Characteristics of all patients are detailed in Table 1.

Effector cell treatment and assessment of safety
ACT IL−15 -related self-limiting adverse drug reactions, including pyrexia, chill, myalgia, and fatigue, were reported by 11% (7/63) of patients; however, the reactions did not delay or stop treatment. No patient exhibited pulmonary or   Table S1).
The median PFS of the combination therapy group was 153 days (95% CI, 109-197 days), whereas that of the chemotherapy group was 136 days (95% CI, 103-169 days; p > 0.05; Fig. 7B). The 1-year PFS rate of the combination Fig. 6 The trial scheme. We enrolled 73 patients, of whom 3 were excluded of the remainder, 35 were randomly assigned to each of the combination therapy group and the chemotherapy group. After elimination of patients based on nonadherence, 34 remained in the combination group and 29 in the chemotherapy group GC, gastric cancer, OS overall survival, PFS progression-free survival therapy group (18.2%) was significantly higher than that of the chemotherapy group (14.7%; p < 0.01). These results provided evidence that the addition of ACT IL−15 to a standard chemotherapy regimen improved patient survival.

Discussion
This study found that IL-15 sustained the growth of memory T cells and the proliferation of adoptively transferred cells, as well as that IL-15 was better than IL-2 in rescuing or generating potentially therapeutic cells from peripheral blood. In the model of human GC, mice that received ACT IL−15 cells survived significantly longer than mice that received ACT IL−2 cells. Most importantly, the combination of ACT IL−15 + chemotherapy was superior to chemotherapy alone in patients with advanced GC. Therefore, IL-15 appears to be a versatile cytokine that increases the activity of adoptively transferred antitumor lymphocytes in vivo and might have more therapeutic potential than IL-2. The transfer of ACT IL−15 cells was well tolerated, and treatment was not interrupted by side effects. These findings suggested that ACT IL−15 + S-1/oxaliplatin has potential as a safe and effective treatment for advanced GC.
In addition to the effects of chemotherapy in inhibiting tumor replication, immunotherapy + chemotherapy has been proposed as a comprehensive treatment that might improve outcomes in human GC [15]. Research has suggested that the antitumor effects of chemotherapy occur through regulation of the immune system [12]. There are no reports of welltolerated chemotherapy regimens in patients with advanced GC that yield a median survival time of ≥ 1 year. We believe that median survival time of 1.3 years and median PFS of 5.1 months in patients with GC in our study was the result of synergy between ACT IL−15 and S-1 + oxaliplatin.
S-1 alone or S-1 + oxaliplatin remains the backbone of GC chemotherapeutics, and it has been widely used as firstline therapy in advanced GC [24,25]. Therefore, we selected S-1/oxaliplatin as chemotherapy treatment. Cancer immunotherapy has been found potentially useful in controlling tumor growth and promoting patient survival [26], and ACT combined with chemotherapeutic regimens for treating GC is more productive than treatment with chemotherapy alone [23,27]. Recent studies have shown that chemoresistant cancer cells are sensitive to the cytotoxic effect of ACT lymphocytes [23,28,29], which, theoretically, have potential to eradicate residual tumor cells after chemotherapy. Our findings suggested that the combination of ACT with S-1/ oxaliplatin favorably modulated the immune milieu of the host, which could be beneficial in instances of chemoresistance. Therefore, we hypothesized that patients would benefit from ACT IL−15 along with S-1/oxaliplatin.
Another study [30] has found that the mere generation of large numbers of highly differentiated lymphocytes is insufficient to achieve tumor regression. The effects of ACT on OS and PFS were most notable in patients who received lymphocytic preparations with high levels of CD45RO + T-cells. CD45, also known as leukocyte common antigen, functions as a tyrosine phosphatase in leukocyte signaling [31]. CD45RO is a marker of effector memory T lymphocytes following adoptive transfer [32] and has been demonstrated to closely represent the activation status of T cells [33]. Memory T cells are known to be generated during cell-mediated immune responses and to survive for months, even years after the antigen is eliminated [34,35]. In our study, IL-15, unlike IL-2, uncoupled differentiation of T cells from proliferation to generate a large number of effective, undifferentiated memory lymphocytes. Therefore, our findings suggested that the combination of ACT IL−15 + S-1/ oxaliplatin favorably modulated the immune milieu of the host, which could be beneficial in instances of chemoresistance. A multivariate Cox regression model showed that the combination of ACT IL−15 + S-1/oxaliplatin treatment, female sex, and number of organs with metastases were independent prognostic factors for OS. Together, these results validated the use of ACT IL−15 along with chemotherapy for improving response in patients with GC.
Although this was a randomized trial, it had several limitations. First, experiments on the treatment of patients with GC were not conducted in a blinded manner. Although the randomization of patients to a combination of ACT IL−15 and chemotherapy versus chemotherapy alone was appropriate, the lack of blinding could have introduced bias into assessment of survival outcomes and adverse effects. Other limitations were that the number of GC patients studied (63 divided into two groups) was modest and that this was a single-center study. Therefore, the representativeness of the results needs testing in larger samples from multiple centers.

Conclusions
Adoptive transfer of ACT IL−15 cells into a mouse model of GC and into patients with advanced GC (the latter also treated with S-1 + oxaliplatin) improved survival rates. These findings will support efforts toward improving the treatment of GC patients and evaluating ACT IL−15 cells in the early treatment of patients with advanced disease.