Effect of neoadjuvant chemotherapy on the immune microenvironment in gastric cancer as determined by multiplex immunofluorescence and T cell receptor repertoire analysis

Abstract

Background: The combination of immune checkpoint blockade and chemotherapy has revolutionized the treatment of advanced gastric cancer (GC). It is crucial to unravel chemotherapy-induced tumor microenvironment (TME) modulation and identify which immunotherapy would improve antitumor effect.

Methods: In this study, tumor-associated immune cells (TAICs) infiltration in residual tumor after neoadjuvant chemotherapy (NAC) together with 1075 cases of treatment-naïve GC patients was analyzed first. Then we performed multiplex fluorescence staining of a panel of immune markers (CD3, CD4, CD8, FOXP3 and PDL1) and T cell receptor β-chain sequencing to phenotype and enumerate T cell subpopulations and clonal expansion in paired GC samples (prechemotherapy and postchemotherapy) from another cohort of 30 cases of stage II/III GC patients.

Results: Infiltration of CD68⁺ macrophages in residual tumors after NAC was significantly decreased compared with treatment-naïve GC patients, while no significant difference observed with respect to other immune markers. In residual tumors, post-NAC CD8 +T cells and CD68+ macrophages levels were significantly associated with chemotherapy response. Post-NAC CD8+ T cell levels remained as an independent predictor for favorable prognosis. Furthermore, when comparing the paired samples before and after NAC from 30 cases of stage II/III GC patients, we found FOXP3+ regulatory T cells proportion significantly decreased after chemotherapy. Pre-NAC FOXP3+ T reg cells level was much richer in the response group and decreased more significantly in the stromal compartment. CD8+ cytotoxic T lymphocytes levels were elevated after chemotherapy, which was more significant in the group treated with XELOX regimen and in patients with better response, consistent with the TCR diversity elevation.

Conclusions: These findings have deepened our understanding of the immune modulating effect of chemotherapy and suggest that the immune profile of specimens after standard chemotherapy should be considered for the personalized immunotherapy to ultimately improve clinical outcome in patients with GC.

Keywords: gastrointestinal neoplasms; tumor biomarkers; tumor microenvironment.

Figure 1

(A) Comparison of immune cell infiltration of gastric cancer in patients treated with and without NAC. (B) Comparison of immune cells infiltration of gastric cancer in patients with different response (TRG 0–1 vs TRG 2–3) to NAC. X-axis represents the immune markers, and Y-axis represents the staining density of each marker. NAC, neoadjuvant chemotherapy.

Figure 2

(A) Multiplex immunofluorescence images of immune cell infiltration in paired pre-NAC and post-NAC specimens of gastric cancer. (B) Comparison of total immune markers in matched pre-NAC and post-NAC group. X-axis represents the single/combined immune markers. Y-axis indicates the immune infiltrating density that is transformed using Log2 (density +1). NAC, neoadjuvant chemotherapy.

Figure 3

(A) Comparison of immune markers in total, tumor and stromal before NAC between the respond (TRG0-1) and non-respond (TRG2-3) group. (B) Comparison of immune markers in total, tumor and stromal after NAC between the respond (TRG0-1) and non-respond (TRG2-3) group. X-axis represents the single/combined immune markers. Y-axis indicates the immune infiltrating density that is transformed using Log2(densitiy +1). (C) Comparison of the alterations in CD8+ immune cells between the respond (TRG0-1) and non-respond (TRG2-3) group. (D) Comparison of the alterations in FOXP3+ immune cells between the respond (TRG0-1) and non-respond (TRG2-3) group. X-axis represents the respond/no-respond group for NAC, and Y-axis indicates the change of immune infiltrating density after NAC, which is represented by density in post-NAC group minus density in pre-NAC group. NAC, neoadjuvant chemotherapy.

Figure 4

T cell receptor repertoire analysis of preneoadjuvant and postneoadjuvant chemotherapy-treated gastric cancer. Comparison of the number of total TCR (A) and unique TCR (B) and the proportion of unique TCR (C) in patients treated with and without NAC. Estimation relative abundance for the groups of top TCR clonotypes in TCR repertoire of all patients (D) and of paired pre-NAC and post-NAC group (E). The D50 index (F) and normalized Shannon entropy (G) of the TCR repertoire before and after NAC for each patient. (H) Comparison of the number of tumor mutation burdens (TMBs) pre-NAC and post-NAC for each patient. NAC, neoadjuvant chemotherapy.

Figure 5

Analysis of different chemotherapy regimens. Comparison of the normalized Shannon entropy (A) and tumor mutation burdens (B) of the TCR repertoire pre-NAC and post-NAC for each patient in different regimens. (C) Comparison of the normalized Shannon entropy and tumor mutation burdens of the TCR repertoire in matched pre-NAC and post-NAC group for patients accepted XELOX regimen by effect.