Introduction
Gastric cancer remains one of the most aggressive types of human cancer, ranking among the five leading causes of cancer incidence and mortality worldwide [
1,
2]. The anti-programmed death 1 (PD-1) inhibitor plus chemotherapy combination showed satisfactory toxicity and antitumor activity for advanced gastric cancer according to the results of the CheckMate 649 trial, with an objective response rate of approximately 60% and a median overall survival of 14.3 months [
3]. Nevertheless, the clinical benefit is limited for a significant portion of patients, and treatment resistance frequently occurs [
4]. Immunotherapy can also be associated with immune-related adverse events and high treatment costs. Hence, identifying patients who will benefit from the anti-PD-1 inhibitor plus chemotherapy combination remains a priority.
The biomarkers predictive of anti-PD-1 response in gastric cancer at this stage include PD-L1 expression, the microsatellite instability (MSI)/mismatch repair (MMR) status, Epstein-Barr virus (EBV) infection, and the tumor mutation burden (TMB) [
5,
6]. These biomarkers only focus on the inherent features of tumor cells and miss interactions with other tumor microenvironment (TME) components [
7], which partially accounts for the unsatisfactory predictive efficacy of the anti-PD-1 response [
8‐
10]. As the most abundant cell component in TME, tumor-associated macrophages (TAMs) typically present with an M2-like phenotype [
11] that is well known to induce immunosuppression through various mechanisms [
12‐
14], which promotes cancer progression [
15,
16] and resistance to immunotherapy [
17].
Although tumor biopsy is widely adopted for immunotherapy biomarker identification and characterization, it is challenging to obtain tissues because of limited accessibility and invasiveness, especially for advanced disease, which also hinders the accessibility of these biomarkers. Emerging evidence suggests that the localized antitumor immune response has to be sustained by the continuous communication between the TME and the peripheral blood [
18,
19]. Hence, peripheral blood analyses have the potential for TME immune monitoring. It is vital to develop biomarkers that could help monitor the status of M2 macrophages within the TME through peripheral blood analyses which are thought to be more readily accessible and noninvasive.
In this study, we analyzed the association of all peripheral leukocyte subpopulations prior to treatment with clinical outcomes in advanced gastric cancer. We also performed immunofluorescence (IF) and immunohistochemistry (IHC) evaluations of tumor samples to characterize the basophil-related TME phenotype.
Materials and methods
Study population
We retrospectively analyzed data collected at a tertiary hospital (Guangzhou, China) between November 2019 and December 2021. Patients were included if they had recurrent or metastatic gastric cancer and were treated with the anti-PD-1 inhibitor plus chemotherapy combination. The study population also comprised a control group consisting of patients with advanced gastric cancer treated with chemotherapy alone in the same period. Patients were eligible if they had received at least three cycles of treatment, had measurable (at least one lesion) or evaluable disease per the Response Evaluation Criteria in Solid Tumors (RECIST, version 1.1), underwent peripheral blood examination within 2 weeks before treatment initiation and prior to cycle 3, and underwent regular radiological assessments. Patients were excluded if they had a concomitant hematological malignancy, recent infection or inflammation, allergic diseases, or if they received any other anti-tumor therapy within one week prior to blood sampling for the basophil count assessment.
The most recent complete blood counts obtained before the initiation of cycle 1 (up to two weeks before the first treatment) and prior to cycle 3 were retrieved from electronic medical records. Patients were treated until the occurrence of disease progression, intolerable toxicity, and/or the doctors’ decision to discontinue treatment. Contrast-enhanced CT, MRI, or PET-CT at baseline, at week 12, and every 12 weeks thereafter were conducted to assess tumor response. Clinical responses were determined according to RECIST 1.1 and classified as complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD). Depending on the best overall response, patients were classified as non-responders (SD or PD) and responders (CR or PR). A separate cohort of gastric cancer patients who underwent surgery without any preoperative therapy, as well as patients treated with neoadjuvant anti-PD-1 plus chemotherapy and surgery, were included to validate findings in tumor samples.
The study protocol complied with the principles defined in the Declaration of Helsinki. Written informed consent was given by all the patients in this study before treatment. All blood tests and treatments were performed in accordance with relevant clinical guidelines.
Treatment regimens
PD-1-targeting antibodies included Nivolumab, Pembrolizumab, Sintilimab, Camrelizumab, or Toripalimab every three weeks. Chemotherapy regimens involved XELOX (capecitabine plus oxaliplatin), SOX (S-1 plus oxaliplatin), FOLFOX (leucovorin, fluorouracil, and oxaliplatin), FLOT (fluorouracil, leucovorin, oxaliplatin, and docetaxel), DCF (docetaxel plus fluorouracil), and other combinations. The regimen was based on the patient’s condition and preference.
Assessment of PD-L1 expression
The PD-L1 combined positive score (CPS) was determined by immunohistochemistry using validated anti-PD-L1 antibodies: E1L3N (Cell Signaling Technology, Danvers, Massachusetts, USA) and 22C3 (Dako North America Inc, Carpinteria, California, USA).
Assessment of the MMR status
The mismatch repair (MMR) status was routinely assessed by immunohistochemistry (IHC) staining of four proteins (MLH1, PMS2, MSH2, and MSH6). Tumors with a deficient MMR (dMMR) phenotype were defined as showing a loss of expression of one or more MMR proteins. A proficient MMR (pMMR) phenotype was defined as one showing intact MMR protein expression.
Assessment of the human epidermal growth factor receptor 2 (HER2) expression status
The human HER2 status should be tested by IHC and/or the FISH test. HER2 positivity is defined as the number of tumor cells showing a strong overexpression (3+) exceeding 10% of the total tumor population. If the number of tumor cells displaying moderate HER2 overexpression (2+) exceeds 10% of the total tumor population, the HER2 status is equivocal and negative otherwise. If the initial HER2 result is equivocal by IHC, then a FISH assay should be added to confirm the HER2 status.
Assessment of the EBV infection status
EBV infection was detected by chromogenic in situ hybridization with EBV-encoded small RNA (EBER) using fluorescein-labeled oligonucleotide probes. The positive EBER nuclear expression in tumor cells with negative signals in normal tissues was EBV-positive.
Scoring pathologic response
The tumor regression of gastric cancer after neoadjuvant anti-PD-1 therapy was graded according to a scoring system developed by the American Joint Committee on Cancer to measure the response of rectal cancer to neoadjuvant chemoradiation. The pathologic response ranges from 0 to 3 as follows: 0 (complete response), no viable cancer cells; 1 (marked response), single or small groups of cancer cells; 2 (moderate response), residual cancer outgrown by fibrosis; and 3 (poor or no response), minimal or no tumor kill, extensive residual cancer, or tumor progression. Tumor regression of grades 0–1 was considered as a pathologic response and the remaining were defined as no pathologic response.
Hematoxylin–eosin, immunofluorescence, and immunohistochemistry staining
Detailed information is presented in Additional file
1. Tumor samples were collected from gastric cancer patients who did not receive any preoperative therapy and patients receiving neoadjuvant anti-PD-1 therapy and curative surgery. All staining was conducted on sections of formalin-fixed paraffin-embedded tumor tissues. Basophils were stained using anti-Pro Major Basic Protein 1 (ProMBP1) antibodies (Biolegend, San Diego, USA, catalog number: 346802). M2 macrophages were assessed using an anti-CD163 antibody (Abcam, Cambridge, UK, catalog number: ab182422). For IHC, the slides were incubated with peroxidase-conjugated secondary antibodies, stained using diaminobenzidine (DAB)-H
2O
2, and counterstained with hematoxylin. For IF, fluorescence-labeled secondary antibodies were used and DAPI was counterstained.
Analysis of stained tumor samples
Detailed information is provided in Additional file
1. The IHC results were evaluated by two independent observers who were blinded to the clinical data. The mean number of stained basophils and CD163
+ macrophages was counted in three different areas at 400 × magnification. Observed cell numbers were divided by the evaluated area to obtain an average cell density. Major discrepancies in cell counts were reviewed and reanalyzed together to reach a consensus.
Statistical analysis
Overall survival (OS) was defined as the interval from the first dose of the anti-PD-1 inhibitor plus chemotherapy combination to either death from any cause or the last follow-up. Progression-free survival (PFS) was defined as the interval from the first dose of the anti-PD-1 inhibitor plus chemotherapy combination to disease progression documented by imaging, death, or last follow-up. The cutoff basophil counts for response (CR/PR) were determined using time-dependent receiver operating characteristic (ROC) analysis. Categorical variables were analyzed using the chi-square test or Fisher’s exact test, and variables from the peripheral blood were examined for normality of distribution and compared either with the Mann–Whitney U-test or the two-sample T-test. Kaplan–Meier survival curves were plotted and compared using the log-rank test. Factors associated with clinical response were explored using binary logistic regression analyses. Covariates associated with PFS and OS were evaluated through univariable and multivariate Cox regression analyses. Variables that reached statistical significance at p ≤ 0.1 were allowed to enter into the multivariate analyses. Correlations between two parameters were estimated using the Pearson correlation coefficient. All p-values were two-sided and confidence intervals (CI) were at the 95% level, with significance predefined to be at p < 0.05. All statistical tests were conducted using SPSS version 22.0 (SPSS Inc, Chicago, IL, USA) or GraphPad Prism (version 8.0e, GraphPad Software).
Discussion
In this study, we reported that among recurrent or metastatic gastric cancer patients treated with the anti-PD-1 inhibitor plus chemotherapy combination, high peripheral basophil counts at baseline and increased basophil counts at cycle 3 compared to baseline were prognostic for unfavorable clinical outcomes. In contrast, no correlation was observed in the chemotherapy-only group. Peripheral basophil counts were positively correlated with the abundance of intratumoral basophils. Both peripheral and intratumoral basophil counts were positively correlated with the abundance of tumor-infiltrating M2 macrophages. Increased basophil counts may indicate an immune-evasive TME characterized by more abundant tumor M2 macrophage infiltration.
Emerging evidence suggests that the efficacy of peripheral cellular biomarkers in improving patient stratification for anti-PD-1 therapy looks promising. Indeed, numerous biomarkers captured in peripheral blood analyses or simple parameters quantified from complete blood routine tests are suggested to be associated with the efficacy of immunotherapy or chemotherapy [
22‐
25]. In this study, decreased peripheral basophil counts at baseline were found to be independently associated with improved radiological responses and was prognostic for favorable survival in advanced gastric cancer. The basophil count might be a potential biomarker of anti-PD-1 efficacy. Consistent with our findings, IL4
+ basophils were found to reduce the survival of pancreatic cancer patients by regulating Th2 inflammation [
26]. Gastric cancer-infiltrating basophils were indicative of worse survival and inferior benefits of adjuvant chemotherapy [
27]. In contrast, basophils prolonged the survival of melanoma patients by recruiting CD8
+ T cells and enhancing tumor rejection [
28]. Higher peripheral basophil counts at baseline correlated with longer OS in immune checkpoint inhibitor-treated melanoma [
29]. Different tissue microenvironments may account for this discrepancy as different tumor microenvironments may represent distinct cytokine milieus that could modulate the specific gene signature of resident basophils and alter the basophils towards different phenotypes [
21]. Although we suggest that peripheral basophil counts were prognostic for survival in the anti-PD-1 plus chemotherapy group but not in the chemotherapy-only group, the definitive role of basophils in immunotherapy is far from clear and warrants further investigation.
Our observation that an early increase in the basophil count from cycle 1 to cycle 3 of anti-PD-1 plus chemotherapy combination in patients with low baseline basophil counts was prognostic for unfavorable clinical outcomes may distinguish patients with a greater probability of disease progression on anti-PD-1 therapy from those with a low risk of progression before radiological assessments. For these patients, peripheral basophil surveillance could potentially guide the implementation of alternative treatment solutions in a timely manner.
Numerous peripheral cellular biomarkers have been demonstrated to predict immune checkpoint inhibitor efficacy; however, few of them have attempted to examine how these biomarkers impact the TME [
24,
30]. Basophils are well known to present in the TME and are typically thought to be critical effector cells in allergy and protective immunity after parasite infection [
31‐
33]. In fact, basophils are highly interactive cells. Unique functions unmet by other blood-borne cells, such as immune imprinting, host response to bacteria, autoimmune disease, and allograft fibrosis, have been identified in basophils [
21,
34‐
37]. In contrast to mast cells whose progenitors become fully differentiated in tissues, basophils are thought to complete their development in hematopoietic tissues and keep circulating in the blood until they are cleared or recruited into tissues in pathological conditions [
38]. Therefore, we stained the basophils on tumor sections and found that high peripheral basophil counts were associated with increased tumor-infiltrating basophil counts. Although basophil accumulation has been found in tumoral tissues and the peripheral blood of cancer patients [
39‐
43], to our knowledge, this study is the first to show such an association in the same patient. Nevertheless, since basophils had been observed in tumor-draining lymph nodes where they regulated the intratumoral Th2 inflammation [
26], it was possible that they may also exert biological effects in other sites in addition to the primary tumor.
Furthermore, it has not been examined if patients who display basophilia before treatment and did not respond to immunotherapy have basophil-mediated immunosuppression. TAMs, a specialized phenotype of M2-polarized macrophages, produced more anti-inflammatory cytokines than M1-type macrophages [
11]. These anti-inflammatory cytokines could induce immunosuppression and promote tumor progression and resistance to immunotherapy [
44]. We interrogated the basophil and M2 macrophage infiltrates within the TME and found that they were spatially in proximity to each other. Further, we observed that both the peripheral and tumor-infiltrating basophil counts had positive correlations with the tumor-infiltrating M2 macrophage count that characterized an immune-evasive TME. Consistent with our findings, tumor-infiltrating basophils had been demonstrated to be associated with M2-polarized macrophage infiltration and decreased interferon-γ expression in gastric cancer [
27]. It has also been shown that basophils regulate alveolar macrophage maturation and immunomodulation function [
21]. In line with this observation, our results give some indications that basophils might modulate the local TME and contribute to immune suppression. To further confirm this conclusion in the context of anti-PD-1 plus chemotherapy, primary tumor samples from gastric cancer patients treated with the neoadjuvant anti-PD-1 plus chemotherapy combination were collected and stained against proMBP1 and CD163. Consistent with our conclusion, the responders have decreased peripheral basophil counts and tumor infiltration of basophils and M2 macrophages.
This study has several limitations that could not be neglected. First, its retrospective design has the inherent deficit of being observational or non-experimental. For example, variables such as allergic conditions or medications that potentially affect the circulating basophil count were not fully considered. Second, only M2 macrophages within the TME were investigated in this study. Basophils may have effects on other immune cell subpopulations in addition to M2 macrophages since they could broadly interact with the immune and non-immune compartmentsyy [
21]. The whole immune landscape of gastric cancer needs to be further investigated in further studies. Lastly, the optimal basophil count threshold needs to be prospectively validated in a multicenter randomized controlled trial.
In conclusion, advanced gastric cancer with high peripheral basophil counts at baseline has an M2 macrophage-infiltrating TME and unfavorable clinical outcomes of the anti-PD-1 inhibitor plus chemotherapy combination. Basophil counts may be associated with an immune-evasive tumor microenvironment via an increase in M2 macrophage infiltration, and this could be a potential biomarker of anti-PD-1 efficacy and an immunotherapeutic target for advanced gastric cancer.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.