Integrative analysis of bulk and single-cell gene expression profiles to identify tumor-associated macrophage-derived CCL18 as a therapeutic target of esophageal squamous cell carcinoma

Abstract

Background: Esophageal squamous cell carcinoma (ESCC) is a common gastrointestinal malignancy with poor patient prognosis. Current treatment for ESCC, including immunotherapy, is only beneficial for a small subset of patients. Better characterization of the tumor microenvironment (TME) and the development of novel therapeutic targets are urgently needed.

Methods: In the present study, we hypothesized that integration of single-cell transcriptomic sequencing and large microarray sequencing of ESCC biopsies would reveal the key cell subtypes and therapeutic targets that determine the prognostic and tumorigenesis of ESCC. We characterized the gene expression profiles, gene sets enrichment, and the TME landscape of a microarray cohort including 84 ESCC tumors and their paired peritumor samples. We integrated single-cell transcriptomic sequencing and bulk microarray sequencing of ESCC to reveal key cell subtypes and druggable targets that determine the prognostic and tumorigenesis of ESCC. We then designed and screened a blocking peptide targeting Chemokine C-C motif ligand 18 (CCL18) derived from tumor associated macrophages and validated its potency by MTT assay. The antitumor activity of CCL18 blocking peptide was validated in vivo by using 4-nitroquinoline-1-oxide (4-NQO) induced spontaneous ESCC mouse model.

Results: Comparative gene expression and cell-cell interaction analyses revealed dysregulated chemokine and cytokine pathways during ESCC carcinogenesis. TME deconvolution and cell interaction analyses allow us to identify the chemokine CCL18 secreted by tumor associated macrophages could promote tumor cell proliferation via JAK2/STAT3 signaling pathway and lead to poor prognosis of ESCC. The peptide Pep3 could inhibit the proliferation of EC-109 cells promoted by CCL18 and significantly restrain the tumor progression in 4-NQO-induced spontaneous ESCC mouse model.

Conclusions: For the first time, we discovered and validated that CCL18 blockade could significantly prevent ESCC progression. Our study revealed the comprehensive cell-cell interaction network in the TME of ESCC and provided novel therapeutic targets and strategies to ESCC treatment.

Keywords: CCL18; Cell–cell interaction; Esophageal squamous cell carcinoma; Tumor associated macrophage; Tumor microenvironment.

Fig. 1

Design overview and molecular characteristics of the ESCC microarray cohort. (A) Schematic diagram of the study design. (B) 351 differentially expressed genes with log fold change > 2 and P < 0.01 between paired peri- and tumor samples in ESCC were shown as heatmap with differential degree (D), tumor (T), nodal (N) and stages information. (C) Differences in hallmark pathway activities scored by GSVA between paired peri- and tumor samples in ESCC. Shown are t values from a linear model, corrected for patient of origin. (D) KEGG Pathway enrichment analysis result of the differentially expressed genes in ESCC

Fig. 2

The immune landscape and prognostic significance of ESCC tumor microenvironment. (A) UMAP visualization of 41,237 cells from 22 peri- and tumor samples (PRJNA777911). (B) Heat map of signature matrix genes derived by CIBERSORTx from scRNA-seq data distinguishing 10 cell types. (C) Heatmap of the normalized absolute abundance for each cell type in the ESCC microarray cohort revealed distinct microenvironment compositions between peri- and tumor sites with t-test. (D) Kaplan–Meier curves of overall survival of ESCC patients stratified by the abundance of macrophages (left), fibroblasts (middle) and mast cells (right)

Fig. 3

Cell–Cell communications in ESCC TME. Number of possible interactions between the five major cell types in peri- and tumor sites (A and B). (C) Total number of possible interactions. (D) Differential number of possible interactions between any two cell populations. Red (positive values) and blue (negative values) in the color bar indicate higher number of predicted interactions in peri- and tumor sites, respectively

Fig. 4

Comparative analysis of cell communication identified differentially expressed ligand-receptor pairs in ESCC. (A) Heatmap showing the expression of ligand-receptor pairs highly expressed in ESCC TME. (B) The significantly related ligand–receptor interactions in the ESCC TME inferred by CellChat analysis. Kaplan–Meier curves of overall survival of ESCC microarray cohort stratified by the expression of CCL18 (C) and PITPNM3 (D). (E) A circle plot showing the cell–cell communication network of CCL18-PITPNM3 axis estimated by CellChat. (F) CCL18 visualization by immunofluorescence imaging in conjunction with CD68⁺ M1 macrophages or CD206⁺ M2 macrophages

Fig. 5

CCL18 promotes proliferation of esophageal squamous carcinoma cell lines. (A and B) The mRNA expression level of CCL18 and PITPNM3 (A), and the secretion level of CCL18 (B) in HET-1A and other seven esophageal carcinoma cell lines. (C and D) The cell viability and growth curve of EC-109 cells in the presence or absence of rhCCL18 at increasing concentrations (10–40 ng/mL), as determined by MTT assay (C) or the cell number counting (D). (E) The colony formation of EC-109 cells in the presence or absence of rhCCL18 at increasing concentrations (10–40 ng/mL). (F and G) The cell growth curve and clonal formation of EC-109, EC-109 Vector and EC-109 shPITPNM3 cells in the presence or absence of 40 ng/mL rhCCL18. ***P < 0.001, **P < 0.01, *P < 0.05. (H) EC-109, EC-109 Vector and EC-109 shPITPNM3 cells were treated with or without 40 ng/mL rCCL18 in the presence or absence of STAT3 inhibitor S3I-201. The expression or phosphorylation of JAK2 and STAT3 were determined by western blot. β-Actin was used as a loading control

Fig. 6

Identification of a CCL18 peptide inhibitor. (A) Sequence alignment of human CCL18, human and mouse CCL3 by the online tool ESPript3. The protein structure was labeled according to CCL18. The consensus sequences were listed. The numbers in green represent the cysteine residues to form the disulfide bonds. (B) The structure of human CCL18, human and mouse CCL3 were superposed as the sequence alignment using the software MOE. (C) The information of peptides derived from human CCL18 or CCL3 selected based on the structure. The effects on the proliferation of EC-109 cells transfected with the vector or shPITPNM3 with or without the existence of 20 ng/mL rhCCL18 by Pep1 (D), Pep2 (E), Pep3 (F) and Pep4 (G) in indicated concentrations using the MTT assay. *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 7

Tumor inhibition of the 4-NQO induced ESCC mouse model by a CCL18 peptide inhibitor. (A) Schematic illustration of the establishment and verification of the ESCC model. The ESCC mice were randomly divided into the NS or Pep3 treatment groups, and treated with 10 mg/kg of Pep3 for 8 days. (B) The body weight of the ESCC mice treated with NS or Pep3. (C) Survival curve of the ESCC mice (n = 8 for NS, n = 7 for Pep3). (D) Tumors of represented with the whole esophageal tissues, the numbers of the tumors and the total long diameter of the tumors were recorded (n = 3–4). *P < 0.05, **P < 0.01. (E) The representative flow cytometry plots and summary data of CD8⁺ or CD4⁺ T cells in the spleen of the ESCC mice (n = 3 or 5). ns, no significance, *P < 0.05. (F) The representative flow cytometry plots and summary data of macrophages or M1/M2 ratio in the spleen of the ESCC mice (n = 3 or 5). ns, no significance, *P < 0.05. (G) The H&E staining and immunostaining of the esophageal tissues of ESCC mice after treatment. The representative figures were obtained from serial section slides

Fig. 8

Schematic illustration of the possible antitumor mechanism of CCL18/CCL3 blockade. CCL18 and CCL3 in the ESCC TME are secreted by TAMs and other myeloid cells. Pep3 may elicit antitumor effects in ESCC by preventing infiltration of tumor-associated macrophages (TAM) and proliferation of ESCC cancer cells, via CCL3/CCR1/CCR5 and CCL18/PITPNM3 pathways