Regulatory T cells and M2 macrophages present diverse prognostic value in gastric cancer patients with different clinicopathologic characteristics and chemotherapy strategies

Background Gastric cancer (GC) remains a refractory cancer worldwide. Currently, exploring the differences of the immune status in GC patients with different subgroups might provide promising immunotherapeutic approaches for the treatment of GC. Methods In this study, a total of 598 surgically resected FFPE primary gastric cancer samples were assessed for FOXP3, CD163, CD3, CD8, and PD-L1 markers. The correlations between the immune markers expression and clinicopathological features and prognosis were investigated retrospectively. Results In general, PD-L1, CD3, and CD8 could be regarded as favorable prognostic factors. Our data demonstrated that high infiltration of FOXP3+ Treg indicates better prognosis in stage I–II patients, while the converse outcome was noted in stage III–IV patients. Our data also confirmed different prognostic value in different pathological classifications, chemotherapy strategies, and locations, with or without lymph node metastasis. Also, M2 macrophages indicated poor prognosis in general. However, high M2 macrophage infiltration suggests a favorable prognosis in signet ring cell carcinoma and mucinous adenocarcinoma. Moreover, the prognostic value of the two indices when they are combined is reported. Conclusions These results suggested that different immune statuses are exhibited in different subgroups of GC, which may direct further understanding of the immune status of GC as well as provide a further theoretical basis and potential targets for GC immunotherapy. Electronic supplementary material The online version of this article (10.1186/s12967-019-1929-9) contains supplementary material, which is available to authorized users.

gastric adenocarcinoma into additional subtypes, which is widely used [6,7]. Some studies have described poor prognosis for the signet ring cell carcinoma and mucinous adenocarcinoma compared with other subtypes of GC [8,9].
With GC progression, the infiltrating status of specific immune cells and their differences in predicting prognosis have not been discussed at length in different pathological classifications of GC. Given that GC is a complex disease with heterogeneity [10], subgroup analysis is necessary to further improve the accuracy of immune indicators for the prognosis of GC patients.
Chemotherapy with platinum and fluoro-pyrimidine is currently the recommended first-line therapy for patients with advanced or metastatic GC who have good performance status [11]. In recent years, immunotherapy with immune checkpoint inhibitors has revolutionized the treatment of advanced cancer and demonstrated quite promising results [12][13][14]. Blockade of the PD-L1/PD-1 pathway remains ineffective in GC [15,16]. Hence, to enhance the efficacy of immunotherapy in GC, identifying some more precise predictive biomarkers in order to select better GC patients who might benefit most from immunotherapy is urgently required. Recently, a positive correlation between the presence of tumor-infiltrating lymphocytes (TILs) and survival of patients with malignancies was reported [5]. High infiltration of CD3 + and CD8 + T lymphocytes is associated with better outcome in GC, demonstrating the critical role of T cell-mediated host immunity in repressing tumor progression [17,18]. Tumor infiltrative forkhead box P3-positive (FOXP3 + ) regulatory T cells (Tregs) might play a crucial role in the immune microenvironment of GC [19][20][21]. However, the prognostic value of Treg infiltration remains controversial in GC patients [22,23]. Also, tumor-associated macrophages (TAMs) are known to directly or indirectly support the tumor growth and metastasis in GC and are positively associated with the depth of invasion and clinical stage. TAMs display high plasticity and perform diverse, supportive functions which were specialized for different tissue compartments [24,25]. M2 macrophages are associated with promoting tumor growth, remodeling tissue, promoting angiogenesis, and suppressing adaptive immunity. CD163, a member of the scavenger receptor cysteine-rich family, is a M2 macrophage marker [26,27].
In this study, we used a tissue microarray (TMA) including 598 GC primary lesion specimens to investigate the expression of PD-L1 and quantify tumor-infiltrating CD3 + , CD8 + T lymphocytes, FOXP3 + Tregs, and CD163 + M2 macrophages density to determine their relationships with clinicopathological features and prognosis in GC patients based on different locations, clinicopathological stages, pathological classifications, chemotherapy strategies, and lymph node metastasis. Compared with previous relevant studies, our research have conducted an in-depth analysis of different subgroups of GC, which have rarely been reported before.

Patients
This study retrospectively evaluated 598 gastric cancer samples from patients who underwent primary tumor resection between June 2005 to December 2015 at the Department of Gastrointestinal Surgery, RenJi Hospital, School of Medicine, Shanghai Jiao Tong University. All the samples were definitely diagnosed as gastric cancer by Department of Pathology. We excluded the following types of patients: (1) patients without complete clinical information, postoperative pathological diagnosis, follow-up data, etc.; (2) patients receiving previous radiotherapy or other neoadjuvant treatment; (3) patients with non-neoplastic resection, such as laparotomy or palliative gastrointestinal bypass surgery, and non-adenocarcinoma patients, such as those with gastrointestinal stromal tumors, etc.; (4) patients who suffered from other primary malignant tumors; (5) patients with perioperative death from surgical complications; (6) primary tumor involving ≥ 2 regional sites at the site of occurrence; (7) unclear pathological types; (8) tissue samples were unavailable for TMA. Overall survival time was defined as the interval between gastrectomy and either patient death or the last follow-up. The final follow-up date was June 25, 2018, for all cases examined [28][29][30].
All patients received standard treatments, such as D2 radical resection and adjuvant chemotherapy according to the NCCN guidelines. The first-line chemotherapy is platinum and fluorouracil. When the patient relapses, the second-line chemotherapy regimen paclitaxel is added. Among them, fluorouracil includes 5-fluorouracil or S-1 (or xeloda); cisplatin includes cisplatin or oxaliplatin; paclitaxel includes paclitaxel and docetaxel.
The patients' tumor staging was in accordance with the American Joint Committee on Cancer (AJCC 8th edition) staging system. For each case, the diagnosis was confirmed by a senior pathologist through a review of H&E-stained slides.
In brief, after tissue sections were deparaffinized and rehydrated with graded ethanol, tissues were incubated with 0.3% hydrogen peroxide for 30 min and then blocked with 10% BSA (Sangon, Shanghai, China). Slides were first incubated using the specific antibodies at 4 °C overnight and then labeled with the HRP secondary antibody (Thermo Scientific, US) at room temperature for 1 h. Positive staining was visualized using diaminobenzidin (DAB) substrate liquid (Gene Tech, Shanghai), and counterstained by hematoxylin.
All the staining samples were visualized using the ZEISS Axio Vert.A1 microscope system. For each sample, 5 visual fields showing the highest infiltrating densities at 40× magnification were chosen first, and cell numbers were counted at 200× magnification. After counting the cells, cell destiny was calculated as mm 2 for further statistics. Immune cells in vessels, submucosal lymphatic areas, and necrosis/necrosis adjacent areas were not counted in this research. The positive cell numbers were independently counted, and clinicopathologic data were blinded. In a two-category immunoscore analysis, patients were dichotomized into the high-and low-density group according to the median number of stained cells. As a result, the cutoff for FOXP3 was 11/ mm 2 , CD163 was 20/mm 2 , CD3 was 47/mm 2 , CD8 was 17/mm 2 . PD-L1 positivity was defined as staining in 1% or more tumor cells [31][32][33]. PD-L1 expression on tumor cells instead of stroma was immunohistochemically analyzed by an experienced pathologist. Intensity and percentage of stained cells were evaluated separately for tumor and immune cells by 2 individuals who were blinded to the clinicopathological data.

Ethics statement
All experiments were approved by the Ethical Committee of the Shanghai Jiao Tong University School of Medicine, Renji Hospital, and written informed consent was obtained from patients. All procedures followed were by the ethical standards of the responsible committee on human experimentation and the Helsinki Declaration of 1964 and later versions. Informed consent or a substitute for it was obtained from all patients included in the study.

Statistical analyses
SPSS 23.0 and GraphPad Prism 6.0 were used for statistics in this research. Overall survival analysis was performed using the Kaplan-Meier method and the long-rank test. Univariate and multivariate prognostic analyses were performed using the Cox proportional hazards regression model. Hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated as estimates of the correlations. p-values less than 0.05 were considered statistically significant.

Clinicopathologic characteristics
A total of 598 surgically resected FFPE primary gastric cancer samples in TMA were assessed for FOXP3, CD163, PD-L1, CD3, and CD8 markers, including 418 males and 180 females (Additional file 1: Fig. S1). The median overall survival time is 48.1 months (range: 0-123 months), and the median age at diagnosis is 62 years (range: 26-88 years). The tissue samples consist of 14 cases of papillary adenocarcinoma (pap), 183 cases of tubular adenocarcinoma (tub), 316 cases of poor-differentiated adenocarcinoma (por), 22 cases of mucinous adenocarcinoma (muc) and 63 cases of signet ring cell (sig) type (pathological classification by Japanese classification). In total, 124 GC lesions occurred in "the upper parts of the stomach", 197 GC tumors occurred in "the middle parts of the stomach", and 277 GC cases occurred in "the lower parts of the stomach" [6].
Regarding all 598 GC patients, univariate survival analysis revealed that age, pathological classification, T stage, N stage, M stage, pathological TNM stage, PD-L1 expression, and CD3, CD8, and CD163 infiltration were all significantly associated with patients prognosis (Additional file 2: Table S1). A strong negative correlation was noted between patients' prognosis and the pathological TNM stages. GC patients with positive PD-L1 expression in tumor cells exhibited better survival (5- 17:192 N stage, M stage, PD-L1 expression, and CD163 + M2 macrophage, CD3 + T lymphocyte, and CD8 + T lymphocyte cell density showed that T stage, N stage, M stage and CD163 + M2 macrophages are independent significant poor prognostic factors. However, CD8 + T lymphocytes are a favorable prognostic factor (Additional file 2: Table S1).

Immune indicators predict different prognosis in stage II-III GC patients treated with different chemotherapy strategies
In this section, we focused on GC patients who are eligible for radical resection. Stage I GC patients did not receive chemotherapy, and stage IV GC patients had distant metastases and were unable to undergo radical resection [34]. Thus, stage II-III GC patients were mainly discussed. First-line chemotherapy for GC is platinum and fluorouracil. When the patient relapses, the secondline chemotherapy regimen paclitaxel is added. Thus, the two main chemotherapy strategies are fluorouracil, cisplatin and paclitaxel (a+b+c) and fluorouracil and cisplatin (a+b). FOXP3 low CD163 low was an independent favorable prognostic factor in the (a+b+c) group.(HR: 0.319, 95% CI 0.161-0.634, p = 0.001). However, in the (a+b) group, N stage is an independent poor prognostic factor. (HR: 2.384, 95% CI 1.379-4.121, p = 0.002). Both FOXP3 and CD8 are independent favorable prognostic factors (HR: 0.563, 95% CI 0.344-0.920, p = 0.022; HR: 0.533, 95% CI 0.340-0.837, p = 0.006, respectively) (Additional file 10: Fig. S5, Additional file 11: Table S6).

Discussion
Immune cells in the tumor microenvironment have a crucial influence on tumor occurrence and development [35]. In this study, we conducted an immunohistochemical evaluation of specific immune indices from 598 GC samples. We mainly focused on specific immune cells and immune indicators that exhibit different prognostic value in different subgroups of GC patients.  17:192 In this research, CD3 + T lymphocytes, CD8 + T lymphocytes, and PD-L1 are generally favorable prognostic indicators. Many studies have exhibited that a large number of Tregs infiltrate into various types of tumors in humans. A high frequency of tumor-infiltrating FOXP3 + Tregs was often significantly negatively correlated with patient survival [36,37]. To date, the function of FOXP3 + Treg in GC patients and the evaluation  17:192 of prognosis have been controversial. Some studies have explained that high expression of Tregs in GC suggests poor prognosis, which is related to lymph node metastasis, late stage, and poor pathological types. Some studies imply that high Treg infiltration is a good predictor of prognosis. However, these studies do not distinguish the pathological stages of GC, which is only discussed in general terms [22,23]. In this study, the GC patients were first divided into earlier and later stages for discussion. In earlier stage GC patients, there was a tendency of high FOXP3 + Treg infiltration, indicating a good prognosis. However, in lated stage GC patients, the reverse trend was noted. In recent years, many studies have also discussed the different functions of various subsets of tumor-infiltrating Tregs [38][39][40]. These findings indicate that more attention to the basic research of FOXP3+ Treg cells should be paid, and different functions of FOXP3+ Treg cells in different stages of GC may be explored. The disparate prognostic implications of FOXP3 + Tregs in GC were based on different pathological classifications and different locations,which further illustrate the complexity of FOXP3 + Treg's function in different GC subgroups.
Regarding the infiltrating status of various immune cells in different subgroups of GC, FOXP3Tregs showed the lowest level in signet ring cell carcinoma and mucinous adenocarcinoma but higher in the lower parts subgroup, compared with the middle and upper parts subgroups. The levels of CD163 + M2 macrophage were low in signet ring cell carcinoma and mucinous adenocarcinoma but ample in poor-differentiated adenocarcinoma (Additional file 12: Fig. S6). M2 macrophage is a factor that could promote tumor progression in many tumors. A meta-analysis showed that the number of infiltrating M2 macrophages and total TAMs might be negative prognostic factors for GC patients, while M1 macrophage infiltration may be associated with a favorable survival rate [41]. However, in some specific subgroups of GC, different prognostic implications of M2 macrophages for patients were rarely reported. This study observed that M2 macrophages might be highly infiltrated in signet ring cell carcinoma and mucinous