Tumor MICA status predicts the efficacy of immunotherapy with cytokine-induced killer cells for patients with gastric cancer

In this study, we determine the relationship between the expression of major histocompatibility complex class I chain-related gene A (MICA) in gastric cancer tumors after D2 gastrectomy and the clinical outcome of a CIK-containing adjuvant therapy. Ninety-five consecutive patients with gastric cancer after D2 gastrectomy who received adjuvant chemotherapy combined with CIK cell therapy were enrolled. The MICA expression of their tumors was determined by immunohistochemistry (IHC). High expression of MICA protein was documented by IHC in 38 of 95 tumor samples (40.0 %). The MICA status was significantly associated with the age and stage, p = 0.008 and 0.023, respectively. Analysis of NKG2D on in vitro expanded CIK cells showed that the percentages of NKG2D+ in CD3+/CD56+, CD3−/CD56+, and CD3+/CD8+ cells populations were 97.2 ± 1.4, 97.9 ± 1.8, and 95.6 ± 2.1 %, respectively. For patient with high MICA-expressing tumors, the median DFS and OS were longer than for the patients with tumors with low expression of MICA; 46.0 versus 41.0 months (p = 0.027), and 48.0 versus 42.0 months (p = 0.031), respectively. In a multivariate analysis, stage and MICA expression were independent prognostic factors for DFS and OS. Our findings show that adjuvant chemotherapy plus CIK therapy treatment is a promising modality for treating gastric cancer patients after D2 gastrectomy. Especially, those who have tumors with high expression of MICA were more likely to benefit from such a treatment strategy. Subsequent studies in clinical trial cohorts will be required to confirm the clinical utility of these markers.


Introduction
Gastric cancer is the fourth most commonly diagnosed cancer worldwide and is the second leading cause of cancer-related deaths [1]. The incidence of gastric cancer is different throughout the world. In the USA, new diagnoses have declined steadily, whereas the incidence in Asian countries such as China and Japan remains high [1][2][3].
Gastrectomy has been applied as the first choice of therapy for treating gastric cancer [4][5][6][7]. Chemotherapy alone as adjuvant treatment is now considered the standard of care after adequate D2 gastric cancer resection. Radiotherapy added to postoperative treatment is thought to be most appropriate for patients with less than a D1 surgical resection [8][9][10]. Adjuvant treatment results in a modest reduction in the risk of cancer-related death by 25-30 %, translating into an absolute 5-year survival benefit of only 10-15 % [11]. Therefore, development of an effective therapy for preventing cancer relapse after surgery is needed for prolonging the survival of patients with gastric cancer.
In recent years, immune therapy has become the fourth important treatment modality for treating malignant tumors following surgery, radiotherapy, and chemotherapy [12][13][14]. Adoptive immunotherapy, as adjuvant or alternative treatment, is anticipated to be an innovative approach to treat solid tumors. Cytokine-induced killer (CIK) cells meet all the requirements for application in adoptive immunotherapy [15,16].
The CIK cells are a unique population of cytotoxic T lymphocytes with a characteristic CD3?/CD56? phenotype, and they can be generated from cytokine cocktailinduced peripheral blood mononuclear cells (PBMC) [17]. The high proliferation of CD3?/CD56? cells and cytotoxic activity of T cell receptor alpha/beta (TCR-a/b) cells in CIK cell cultures primarily leads to the effective and substantial lytic response of CIK cells toward tumor cells by releasing of cytoplasmic cytotoxic granule content to the extracellular space upon stimulation by susceptible target cells [16,17]. The cytotoxicity of these CD3?/CD56? cells is non-MHCrestricted, perforin mediated and induced via the natural killer group 2 member D (NKG2D) cell-surface receptor [18,19]. The NKG2D ligands, for example, MHC class I-related chain (MIC) A/B and UL-16 binding protein 1-4, are over-expressed on both solid and hematologic tumor cells, turning these cells into favored targets of CIK cells [20,21]. The MICA proteins have a restricted tissue distribution, and they are rarely expressed by normal cells. It has been reported that MICA is constitutively expressed by intestinal epithelial cells and is broadly expressed in a variety of malignancies. Therefore, it is considered to be a tumor-associated antigen (TAA) [22,23].
Previously, Yu showed that the high expression of MICA is one of the indicators of a poor prognosis for advanced non-small cell lung cancer patients [24]. But until now, MICA expression in gastric carcinoma is unclear, whether high MICA expression can be used as a predictive factor of CIK therapy outcome still remains to be determined. In the present study, we explored MICA expression and the effect of CIK therapy to determine whether MICA expression might be a predictive factor of successful CIK therapy for gastric cancer patients.

Ethics statement
All procedures were conducted in accordance with the Helsinki Declaration and approval from the Ethics Committee of Fujian Provincial Cancer Hospital. Written informed consent was obtained from all participants.

Patients
We retrospectively studied all the patients with stage II-III gastric carcinoma diagnosed after surgery who were admitted into the Department of Medical Oncology of Fujian Provincial Cancer Hospital from January 2009 to March 2012. Ninety-five consecutive patients were recruited into the study according to the following criteria: (1) All the patients had R0 gastrectomy with D2 lymphadenectomy, (2) all patients were diagnosed and histologically confirmed with stages IIA, IIB, IIIA, IIIB, or IIIC disease according to the American Joint Committee on Cancer (AJCC) TNM Staging Classification for Carcinoma of the Stomach (7th ed., 2010), (3) Eastern Cooperative Oncology Group performance status between 0 and 2 before adjuvant chemotherapy, (4) complete medical records were available, (5) availability of adequate formalin-fixed, paraffin-embedded (FFPE) tumor tissue for biological marker evaluation, (6) all patients had received at least four cycles of adjuvant chemotherapy based on 5-fluorouracil (5-FU) or capecitabine doublet regimens, (7) patients received at least two cycles of adjuvant immunotherapy with CIK cells, (8) adequate bone marrow function with leukocyte counts 3000-12,000/mm 3 , hemoglobin C8.0 g/dl, and platelet counts C100,000/mm 3 , (9) adequate liver function with total serum bilirubin B2.0 mg/dl and serum transaminases B100/ UI, and (10) adequate renal function with serum creatinine within the upper limit of normal.

CIK cells preparation and treatment
The CIK cells were isolated and cultured according to a standard protocol as described in previous study [24]. As shown in Table 1, 95 patients received adjuvant chemotherapy combined with at least three cycles of autologous CIK cell therapy, the first one of which was given within 2 weeks after surgery, and the others were given once per month starting within 6 weeks after adjuvant chemotherapy. For each cycle, patients were given an infusion of at least 1.0 9 10 10 CIK cells. Patients were grouped for evaluation by the number of cycles of CIK therapy; C5 versus \5 cycles. The patients were eligible for CIK cell therapy until they no longer agreed to continue treatment or until disease recurrence.

Evaluation of toxicity and efficacy
All patients were followed up at the outpatient clinic from the date of surgery until March 31, 2015, or until the time of death. During treatment, patients were evaluated with abdominal computed tomography (CT) scans every 1-2 months according to the Response Evaluation Criteria In Solid Tumors (RECIST) criteria. Disease-free survival (DFS) was measured from the day of surgery to the first evidence of recurrence or death. Overall survival (OS) was defined from the date of surgery to death from any cause.

Immunohistochemistry (IHC)
Formalin-fixed, paraffin-embedded (FFPE) tissue sections of 4 lm thickness were stained for MICA with anti-human MICA rabbit polyclonal antibody (ab62540) on an automated staining platform (Benchmark, Ventana). Immune complexes were detected with the use of an ultraVIEW 3,3 0 -diaminobenzidine (DAB) detection kit with DAB as the chromogen (Ventana). The omission of the primary antibody and its replacement with PBS was used as the negative control, and tumor infiltrating leukocytes were used as internal positive controls for MICA staining. A section of colorectal tissue which had been shown to be positive for MICA was also used as a positive control [24,25].

Evaluation of immunostaining
The sections were scored semi-quantitatively under light microscopy by two pathologists without any prior knowledge about the clinical history of the patients. The MICA membrane expression was scored semi-quantitatively as previously described [24,25]. The intensity of the immunostaining was classified into four categories: 0 = no staining or only a nonspecific background color, 1 = light yellow, 2 = yellow or deep yellow, and 3 = brown or tan. The percentage of positive cells was assessed and classified into five groups: 0 = B 5 % positive cells, 1 = 6-25 % positive cells, 2 = 26-50 % positive cells, 3 = 51-75 % positive cells, and 4 = 76-100 % positive cells. The IHC score of immunoreactivity was obtained by adding the intensity and percentage scores. IHC scores were divided into two groups, cases with scores of 0-4 were defined as the low-expression group, and cases with scores of 5-7 were defined as the high-expression group [24,25].

Quantitative real-time PCR
The resected tissues were immediately placed in RNAlater RNA Stabilization Reagent (Qiagen) and then stored at -80°C. About 20-30 mg of the RNAlater-stabilized tissue samples was disrupted in buffer containing guanidine isothiocyanate and homogenized. Total RNA was extracted using RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions and quantified using an ND-1000 spectrophotometer (NanoDrop, Rockland, DE, USA). The RNA quality was determined by the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA), and the A260/A280 was between 1.8 and 2.0. The primer sequences are as follows: MICA forward primer, TCAG AGTCATTGGCAGACAT and MICA reverse prime, TGTGGCATCCCTGTGGTC.
About 1 lg total RNA was reverse transcribed into cDNA with First-Strand cDNA Synthesis Kit (Roche). The real-time PCR was performed with LightCycler 480 SYBR Green I Master (Roche) using a LightCycler 480 System. The primers used are listed in Table 1. The reaction was performed with 1 cycle of 95°C for 5 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. All reactions were done in triplicate. The relative expression levels of MICA were calculated using 2-DCt method with the b-actin as the internal reference.

Statistical methods
Statistical analysis was performed with SPSS software (version 21.0, SPSS). For all analyses, the significance level was specified as p \ 0.05. Comparisons between the immunotherapy and control groups were analyzed using the v 2 test and the Fisher exact probability test. Student's t test was used to compare the expression difference between high-MICA and low-MICA tissues for quantitative PCR data. p \ 0.05 was considered statistically significant. The OS and DFS variables were estimated by the Kaplan-Meier method, and survival curves were plotted. Two-sided log-rank tests were used to compare survival rates between groups. Multivariate analyses using the Cox proportional hazards regression model were performed to assess the impact of the variables on DFS and OS.

Results
From January 2009 to March 2012, 95 patients with stage II-III gastric carcinoma were enrolled in this study and their tumors were fully evaluated for MICA expression by immunohistochemistry. The patient characteristics are summarized in Table 1.

MICA protein expression
The MICA protein expression in the tumor tissues was broadly homogenous and was detected mainly at the cell membrane and in the cytoplasm (Fig. 1). High expression of MICA protein, IHC scores of 5-7, was documented in 38 of 95 samples (40.0 %). The MICA expression was significantly associated with the age and stage, p = 0.008   Table 2.

mRNA expression of MICA
We collected resected tissues from 53 patients after gastrectomy, which was immediately placed in RNAlater RNA Stabilization Reagent and then stored at -80°C. Twentyfour patients had confirmed high expression of MICA protein, and 29 patients had confirmed low expression of MICA protein. We detected the mRNA expression of MICA in these 53 cases. The mean mRNA expression in the high expression of MICA protein group and low expression of MICA protein group were 0.0023 ± 0.00057 and 0.0003 ± 0.00007, p \ 0.0001 (Fig. 2).

Phenotypic analysis of CIK cells
The phenotype of the cultured cells was determined by flow cytometry (BD FACSCalibur). The cells were labeled with monoclonal antibodies (mAbs) that recognize human CD3, CD8, CD56, and NKG2D. Phenotypic analysis of   (Fig. 4a). In the univariate analysis, stage and MICA status were associated with DFS and OS, while the number of CIK cycles was correlated with DFS, but not with OS (Fig. 4b). The sex, age, histological grade and the regimen of adjuvant chemotherapy were not correlated with DFS and OS ( Table 3). The median DFS and OS of patients with tumors having high MICA expression who received CIK therapy were longer than in the patients with low expression; 46.0 versus 41.0 months (p = 0.027), and 48.0 versus 42.0 months (p = 0.031), respectively (Fig. 4c).
In the multivariate analysis, variables that included sex, age, histological grade, stage, MICA status, CIK cycles, and the regimen of adjuvant chemotherapy were tested to determine their independent effect on DFS and OS. The stage and the MICA status were independent prognostic factors for DFS and OS (Table 4), which indicating a predictive value of the tumor MICA expression level for response to adjuvant chemotherapy plus CIK therapy for patients with gastric cancer after receiving a D2 gastrectomy. This result was consistent with the univariate analysis.

Discussion
In this study, we have characterized high tumor MICA expression as a predictive biomarker that, independent of tumor histology, sex, age, and regimens of chemotherapy, defines patients with gastric cancer who have received a D2 gastrectomy who are most likely to derive a survival benefit from the addition of CIK therapy to adjuvant chemotherapy.
As shown by blocking antibodies and transfection experiments, natural or induced expression of NKG2D ligands, such as MICA/B, markedly enhances the sensitivity of tumor cells to NK cells or CIK cells in vitro [19,[26][27][28]. But tumor cells often express insufficient levels of NKG2D ligands to stimulate tumor cell rejection, either because expression of the ligand is not sufficiently upregulated early in the development of the tumor, or because tumor cells with lower levels of ligand expression are selected by the immune system in vivo as the tumor evolves [29,30]. In the present study, MICA expression using IHC method was detected mainly at the cell membrane and in the cytoplasm, and was significantly associated with the cancer stage, which indicated that the MICA expression may contribute to the invasion and metastasis of gastric cancer. Furthermore, we detected the mRNA expression of MICA in 53 cases which were collected from the resected gastric tissue from 53 patients. The results showed that the MICA mRNA level was significantly correlated with MICA protein expression. The result of the RT-PCR assay proved to be sensitive, and we will test the prognostic ability of the RT-PCR test with a larger sample in the future. The CIK cells are a heterogeneous population of T lymphocytes with the CD3?/CD56? double-positive cells being the most cytotoxic [15,16]. The CIK cells are capable of undergoing degranulation upon stimulation with either agonistic anti-CD3 mAbs or susceptible target tumor cells; the molecules on malignant cells that stimulate the cytolytic activities of CIK cells do not involve a MHC-TCR interaction. Tumor cell-triggered cytolytic activities in CIK cells cannot be inhibited by neutralizing CD3 or HLA class I mAbs [16,31,32]. Several studies had confirmed that CIK cells express high levels of activating NK receptors, including NKG2D, DNAX accessory molecule-1 (DNAM-1). Cell signaling through NKG2D or other activating receptors results in CIK cell activation, leading to effector cell degranulation and cytotoxic effector function [33][34][35]. Neutralization of DNAM-1, NKG2D, and NKp30 by antibodies confirms that these molecules are involved in the TCR-independent tumor cell recognition and killing by CIK cells [33,34]. In this study, it was found that NKG2D was highly expressed in CD8?, CD3?/CD56?, and CD3-/CD56? cells, while further survival analysis shows that the MICA status of the tumors was independent prognostic factor for DFS and OS. This indicates a predictive value of tumor MICA expression level for response to adjuvant chemotherapy plus CIK therapy for patients with gastric cancer who have received a D2 gastrectomy. Meanwhile, we found that patients who received CIK cell treatments for C5 cycles had significantly longer DFS than did those who received \5 cycles. There was also a trend toward longer overall survival in the patients who received C5 cycles of CIK therapy. However, the significance of the number cycles was not confirmed by multivariate analysis, and so, the clinical efficacy of such an approach remained to be evaluated.
Based on our findings, the patients with gastric cancer who are considering the use of CIK therapy would benefit from further tests of MICA expression on their tumors to better define who will respond effectively to the treatment. Patients who have received a D2 gastrectomy whose tumors have high expression of MICA seem to derive a greater clinical benefit from adjuvant chemotherapy plus CIK therapy. In addition, the increasing frequency of CIK therapy, which could improve the total immune system function in patients after gastrectomy, may result in additional benefits.
Our results are limited by the retrospective nature of the analysis. Therefore, a treatment model for using the tumor MICA status as a guide for CIK cell therapy for patients with gastric cancer needs further investigation with an expanded sample size in further prospective clinical studies.
In conclusion, this study demonstrated that adjuvant chemotherapy plus CIK cell therapy treatment is a promising modality for treating gastric cancer patients after D2 gastrectomy. Especially, those with tumors that have a high expression of MICA were more likely to benefit from such a treatment strategy. Enhancing specificity of CIK cell therapy via immunological approaches and identifying biomarkers of response will significantly improve the application of this therapy.