Fibroblast-derived PI16 enhances tumor immune-suppressive microenvironment via inducing Tregs differentiation

Abstract

Purpose

Esophageal squamous cell carcinoma (ESCC) is aggressive with a poor prognosis. The tumor microenvironment (TME) significantly affects tumor progression and therapy resistance. Previous work has shown that fibroblasts in metastatic lymph nodes can confer cisplatin resistance to ESCC cells via PI16 (peptidase inhibitor 16). This study investigates the role of fibroblast-derived PI16 in the ESCC TME.

Methods

Public single-cell RNA sequencing (scRNA-seq) data for ESCC were analyzed. A cell co-culture assay was performed to evaluate regulatory T cells (Tregs) differentiation from naïve CD4⁺ T cells. Immunoprecipitation and mass spectrometry examined PI16’s mechanism in Treg differentiation. In vitro and in vivo assays were conducted to explore fibroblast-derived PI16’s function. Additionally, multiplex fluorescent immunohistochemistry (mfIHC) was performed.

Results

Analyses of the scRNA-seq dataset (GSE203115) reveal that fibroblasts can be classified into PI16 ⁺ and PI16^- subclusters based on PI16 expression levels. PI16 induces Treg differentiation from naïve CD4⁺ T cells through a DOCK2-dependent mechanism. Treatment with a DOCK2 inhibitor significantly inhibits PI16-induced Treg differentiation and increases Teff cell infiltration in vivo. Moreover, upregulation of PI16 in the tumor stroma is associated with poorer long-term survival outcomes in patients with ESCC.

Conclusions

PI16⁺ fibroblasts promote the differentiation of Tregs from naïve CD4⁺ T cells through interaction with DOCK2. Upregulation of PI16 in the tumor stroma is associated with poorer long-term survival outcomes in patients with ESCC. Given the accumulating evidence on the therapeutic impact of targeting the TME, PI16⁺ fibroblasts emerge as a promising novel therapeutic target to overcome tumor immune suppression.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13402-025-01090-5.

Introduction

According to global cancer statistics 2022, esophageal cancer ranks as the seventh leading cause of cancer-related deaths worldwide [1]. Esophageal squamous cell carcinoma (ESCC) is the predominant histological type of esophageal cancer, accounting for approximately 90% of cases [1, 2]. ESCC is prevalent in Eastern Europe and Asia, with the highest regional incidence rates observed in Eastern Asia [1]. Despite advancements in multidisciplinary therapeutic approaches, the prognosis for ESCC remains unfavorable [2].

Cancer development and progression occur in concert with alterations in the surrounding stroma [3]. The tumor microenvironment (TME) is complex and continuously evolving. It comprises stromal cells, cancer-associated fibroblasts (CAFs), vascular endothelial cells, and both innate and adaptive immune cells [3]. Studies have highlighted the significant role of the TME in tumor progression, metastasis, and response to therapy [4]. CAFs can promote ESCC progression by producing and remodeling extracellular matrix (ECM) and secreting chemokines, cytokines, and growth factors [5–7]. However, the depletion of fibroblasts demonstrates that CAFs can also limit tumor progression [8]. These conflicting data indicate that CAFs consist of heterogeneous subtypes with different functions.

In their steady-state mouse fibroblast atlas, Buechler et al. compiled ten fibroblast clusters of which two represent pan-tissue fibroblasts, distinguished by dominant expression of either Pi16⁺ (peptidase inhibitor 16) or Col15a1⁺ (collagen XV), with both subsets also expressing dermatopontin [9]. Their investigation extended to a range of human tissues and diseases, including cancer, infection, and inflammation, revealing an equivalent Pi16-expressing universal human fibroblast subset al.so exists [9]. PI16 is a member of the cysteine-rich secretory proteins, antigen 5, and pathogenesis-related 1 protein (CAP) superfamily [10]. Research on PI16 in cancers is limited. Studies have shown a significant decrease in serum levels of PI16 in prostate cancer patients [11], and it has been utilized as a prognostic marker for prostate cancer recurrence [12]. Single-cell profiling reveals that PI16 is mainly expressed in fibroblasts [13, 14]. Our previous study in ESCC demonstrated that fibroblasts in metastatic lymph nodes confer cisplatin resistance to tumor cells through PI16 [15].

Cross-talk between fibroblasts and other cell types in TME regulates the tumor immune microenvironment and affects tumor therapeutic effectiveness [9, 16]. This interaction is particularly significant in multiple tumors. For instance, antigen-presenting CAFs induce naïve CD4⁺ T cells into regulatory T-cells (Tregs) dependent on the antigen-specific manner in pancreatic cancer [17]. Additionally, a diminished CD68⁺ CAF subset induces Tregs infiltration and is associated with poor prognosis in oral squamous cell carcinoma [18]. Depletion of Tregs results in the differentiation of inflammatory fibroblast subsets in pancreatic cancer [19]. In this study, we combined single-cell RNA sequencing analysis with multiplex fluorescence immunohistochemistry and determined the communications between PI16⁺ fibroblasts and Tregs in ESCC tumor tissues. Our findings demonstrate that PI16 secreted by fibroblasts induces naïve CD4⁺ T cells into Tregs, thereby augmenting the immune-suppressive microenvironment and promoting tumor growth. Taken together, our study elucidates how PI16⁺ fibroblasts may contribute to immune evasion in ESCC and provides insight into strategies to enhance cancer immune therapy.

Methods

Cell lines

293FT and NIH3T3 cells were obtained from Invitrogen (Thermo Fisher Scientific, Waltham, MA). MEC2 is a murine ESCC cell line derived from the primary mouse ESCC tumors which were induced using 4-Nitroquinoline-1-Oxide in C57BL/6 mice [20]. MEC2h with high tumorigenicity was a gift from Dr. F Wang. Cell authentication was performed via STR profiling, confirming the absence of mycoplasma contamination. All cells were cultured in DMEM (Gibco BRL, NY) supplemented with 10% fetal bovine serum (Gibco BRL, NY) and incubated in a humidified atmosphere at 37 °C with 5% CO2.

Human tissues

Tumor tissues from 200 patients with ESCC were collected at Sun-Yat Sen University (SYSUCC). All these patients did not receive any treatment before surgery. The collection of these samples in this study was approved by the Committees for Ethical Review of Research Involving Human Subjects at SYSUCC. Written consents were obtained from the patients and the study was conducted under the ethical guidelines of the Declaration of Helsinki.

Naïve CD4⁺ T cells isolation and culture

The mouse spleen was homogenized and filtrated through a sterile cell strainer to obtain a single-cell suspension. Red blood cells were removed using red blood cell lysis buffer (KeyGEN BioTch, Nanjing, China). After washing with PBS, Naïve CD4⁺ T cells were enriched with a Naïve CD4⁺ T Cells Isolation Kit (Biolegend, CA). The isolated naïve CD4⁺ T cells were cultured in complete medium consisting of RPMI1640 supplemented with 10% FBS, 10 mmol/L HEPES, 1 mmol/L sodium pyruvate, 2 mmol/L L-glutamine, 0.05 mmol/L 2-mercaptoethanol, and streptomycin/penicillin.

Cell co-culture assay

Naïve CD4⁺ T cells co-culture with fibroblasts

Before seeding naïve CD4⁺ T cells, plates were pre-treated overnight at 4 °C with 10 µg/ml anti-CD3 antibody and 2 µg/ml anti-CD28 antibody. A total of 1 × 10⁵ naïve CD4⁺ T cells derived from C57BL/6 mouse spleen were seeded in the bottom chamber of 24-well plates (Corning, NY), while 2 × 10³ NIH3T3 overexpressing PI16 or control cells were seeded in the top chamber. The cells were co-cultured for 5 days.

Naïve CD4⁺ T cells co-culture with CD8⁺ T cells

Prior to co-culture, naïve CD4⁺ T cells and CD8⁺ T cells isolated from C57BL/6 mouse spleens were independently activated with plate-bound anti-CD3 (10 µg/ml) and soluble anti-CD28 (2 µg/ml) for 48 h and 24 h, respectively. For CD8 ⁺ T cells division experiment, CD8⁺ T cells were labelled with 5 µM 5,6-carboxyfluorescein diacetate succinimidy ester (CFSE) for 10 min at 37℃ before co-culture with naïve CD4⁺ T cells. A total of 1 × 10⁵ naïve CD4⁺ T cells were seeded in the top chamber of 24-well plates (Corning, NY), while 1 × 10⁵ CD8⁺ T cells were seeded in the bottom chamber treated with 5 µg/ml recombinant protein PI16 or PBS control for 5 days. To assess intracellular IFN-γ production, brefeldin A (BFA, BioLegend, San Diego, CA) was added during the final 6 h of co-culture.

Cell co-culture with conditioned medium (CM)

Cells were co-cultured with conditioned medium (CM) from NIH3T3 overexpressing PI16 or control cells. Briefly, the cells were cultured in serum-free DMEM for 24 h. The medium was subsequently filtered through 0.45 μm filters and concentrated using a 10-kDa Amicon^® Ultra-15 Centrifugal Filter Device (Millipore, Burlington, MA). Naïve CD4⁺ T cells were cultured in complete medium described above, with CM added at a 9:1 ratio. Naïve CD4⁺ T cells were cultured in the conditioned medium for 5 days.

Mouse model

All animal experiments conducted were approved by the Institutional Animal Care and Use Committee of Sun Yat-sen University Cancer Center (SYSUCC). Male C57BL/6 mice, aged 6–8 weeks, were obtained from Vital River Laboratory Animal Technology (Beijing, China) and housed under specific pathogen-free (SPF) conditions. Mice were confirmed to be SPF and free from common murine pathogens. Upon arrival, mice were acclimatized for at least 7 days prior to experimentation. For the in vivo tumor growth assay, mice were randomly divided into two groups (n = 7 per group). This sample size ensures sufficient statistical power to detect significant differences between groups. C57BL/6 mice were subcutaneously injected with a mixture of 3.3 × 10⁶ MEC2 cells and 3.3 × 10⁵ NIH3T3 cells. NIH3T3 cells were transfected with either an empty vector or a PI16 overexpression plasmid. In the experiment involving the DOCK2 inhibitor CPYPP [21], MEC2h (3.3 × 10⁶ cells/mouse) was used in the experiment. Mice were administered intratumoral injections of CPYPP (20 µg/mouse) or saline, with the initial injection taking place on day 6 post-inoculation when tumor volumes reached 50–100 mm³. Animals were excluded if they exhibited signs of illness or distress during the acclimatization period or if they failed to develop tumors of the required volume range (50–100 mm³) by day 6 post-inoculation. Subsequent injections were given at 3-day intervals for a total of 4 administrations. The experimental unit was a single animal for drug administration. Mice were euthanized by CO₂ inhalation at the study endpoint (day 19). Tumor volume and tumor weight were examined by blinded investigators. Tumor volume was calculated using the formula: Volume (mm³) = length × width² × 0.5. All mice (n = 7 per group) were included in the analysis, as no animals were excluded due to illness, failure to meet tumor volume criteria, or technical issues. A survival curves assay was conducted on C57BL/6 mice (n = 7 per group), with the humane endpoint defined as a tumor length of ≥ 2 cm, in accordance with the NIH Guidelines for Endpoints in Animal Study Proposals.

Mass spectrometry analysis

Naïve CD4⁺ T cells were incubated with 5 µg/ml PI16 recombinant protein at 37 °C for 4 h. The cells were then collected and lysed using an immunoprecipitation (IP) lysis buffer. An IP assay was performed using the Pierce Direct Magnetic IP/Co-IP kit (Thermo Scientific, Rockford, IL) according to the manufacturer’s instructions. Five µg anti-PI16 (Boster, Wuhan, China) and IgG (Cell Signaling Technology, Danvers, MA) were incubated with the lysis solution overnight at 4 °C. The mixture was incubated with magnetic beads and eluted with elution buffer. The pulled-down proteins were analyzed by mass spectrometry using LC-MS/MS (Monitor Helix, Shanghai, China).

Co-immunoprecipitation (co-IP) assay

Co-IP assays were performed using a Pierce Direct Magnetic IP/Co-IP kit (Thermo Scientific, Rockford, IL) following the manufacturer’s protocol. Naïve CD4⁺ T cells were treated with 5 µg/ml PI16 recombinant protein at 37 °C for 4 h. Subsequently, the cells were harvested and lysed with an IP lysis buffer. Lysates were incubated overnight with antibody-coated beads at 4 °C. After washing, the bound proteins were eluted and analyzed by immunoblotting.

RNA sequencing and analyses

Naïve CD4⁺ T cells were isolated from mouse spleens using the MojoSort™ Mouse CD4 Naïve T Cell Isolation Kit (BioLegend, California, USA). The isolated naïve CD4⁺ T cells were labeled with CD127 and CD25 antibodies, and Tregs (CD25⁺CD127⁻) were sorted via flow cytometry. Tregs (CD25⁺CD127⁻) induced by the PI16 recombinant protein were sorted using the identical protocol. Total RNA from Tregs was extracted using TRIzol (Invitrogen, Carlsbad, CA), and RNA integrity was assessed with an Agilent 2100 bioanalyzer (Agilent, Santa Clara, CA). RNA sequencing was performed by Novogene (Beijing, China). Differential expression analysis was conducted using EdgeR, applying thresholds of|Log2(fold change)| ≥ 1 and Padj ≤ 0.05, with Genome assembly GRCm39 as the reference genome for alignment.

Multiplex fluorescent immunohistochemistry (mfIHC) assay

A multiplex fluorescent immunohistochemistry assay was conducted to spatially visualize and quantify cell markers. Formalin-fixed paraffin-embedded ESCC tumor tissues and mouse tumor tissues were analyzed using the Polaris^® system (PerkinElmer, Waltham, MA, USA) in conjunction with the customized Opal 5-color Manual IHC Kit (Panovue, Beijing, China), following the manufacturer’s instructions. Briefly, the slides were deparaffinized and rehydrated using xylene and graded ethanol. Antigen retrieval was performed by incubating the slides in EDTA buffer (pH 8.0) for 10 min at 97 °C. After washing, the slides were blocked with 4% BSA and incubated with the primary antibody at room temperature for 1 h. Subsequently, HRP-labeled secondary antibody and fluorescein tyramide were incubated separately for 15 min at room temperature. This protocol was repeated for each subsequent antibody until all were labeled. Finally, cell nuclei were stained with DAPI (Abcam, Cambridge, UK). Quantification of fluorescence intensity and spatial cell analysis were performed using HALO 3.3 image analysis software (Indica Labs, Albuquerque, NM).

Single-cell RNA sequencing analysis

Single-cell RNA sequencing (scRNA-seq) data from three ESCC patients were obtained from the GEO under accession number GSE203115 [22]. Data analysis was performed in R v.4.2.1 using the Seurat v.4.2.0 package for data preprocessing, UMAP nonlinear dimensionality reduction, and cell clustering. Following Wu’s study [22], the cellular subset co-expressing CD4 and FOXP3 was designated as Treg cells, while the cell population co-expressing FN1 and COL6A1 was classified as fibroblasts. The CellChat v.1.5.0 package was utilized to analyze cell-cell interactions between Tregs and the fibroblasts with high or low expression of PI16.

Statistical methods

GraphPad Prism 9.5 (GraphPad Software, San Diego, CA) was used for data analysis. Unpaired student t-tests were used for all statistical comparisons between the two groups. Two-way ANOVA Sidak’s multiple comparisons were used for multiple comparisons. Kaplan-Meier analysis with the log-rank test was employed to evaluate overall survival. P < 0.05 was deemed statistically significant. All error bars are represented as the Mean ± S.D.

Supplemental materials

Antibodies and reagents can be found in Supplementary information.

Results

PI16 derived from fibroblast induces Tregs differentiation in vitro and in vivo

We first downloaded a single-cell RNA sequencing dataset (GSE203115) comprising 13,280 individual cells from three ESCC tumor tissues from the Gene Expression Omnibus (GEO). Cell clusters were annotated based on the average expression profiles of curated gene sets. A total of 13 distinct cell clusters were identified and visualized employing uniform manifold approximation and projection (UMAP) analysis (Fig. 1A and Supplementary Fig. 1). The co-expression of FN1 and COL6A1 served as markers for fibroblasts. The expression of PI16 was visualized and the majority of PI16 expression was localized within fibroblasts (Fig. 1B). According to the expression levels of PI16, fibroblasts were divided into two subclusters: PI16⁺ and PI16⁻ subclusters (Fig. 1C). Cell-cell interactions were also evaluated, revealing that the interactions between PI16⁺ fibroblasts and Tregs were more prominent than those between PI16⁻ fibroblasts and Tregs (Fig. 1D). We also downloaded bulk RNA-seq data from the esophageal cancer (ESCA) dataset from the Cancer Genome Atlas Program (TCGA) (TCGA-ESCA), and identified a significant correlation between PI16 expression and Treg infiltration levels (R = 0.325; P < 0.001)(Fig. 1E).

Fig. 1.

The expression of PI16 correlates significantly with Tregs infiltration in ESCC tumor tissues. A) Uniform manifold approximation and projection (UMAP) plots of cells from GSE203115 dataset. B) UMAP plots (left) and violin plots (right) showing the expression of PI16 in fibroblasts. C) UMAP plots of cells from GSE203115 dataset and fibroblasts were divided into two subgroups: PI16-positive and PI16-negative. D) The diagram showing cell-cell interactions. The lines reflect the number of ligand-receptor pairs between cells. E) Analysis of TCGA-ESCA dataset revealed a significant correlation between PI16 expression and the level of Tregs infiltration. (Macro/Mono, macrophages/monocytes; SMC, smooth muscle cells)

To investigate the impact of PI16⁺ fibroblasts on Naïve CD4⁺ T cells, we generated PI16-overexpressing NIH3T3 (3T3-PI16) cells (Fig. 2A) and co-cultured these with naïve CD4⁺ T cells isolated from C57BL/6 mouse spleens (Fig. 2B). The proportion of Tregs (CD4⁺CD25⁺Foxp3⁺) was significantly increased in naïve CD4⁺ T cells co-cultured with 3T3-PI16 group compared to those co-cultured with vector control group (Fig. 2C). In addition, naïve CD4⁺ T cells were also supplemented with conditioned medium (CM) from fibroblasts and the percentage of Tregs was upregulated in naïve CD4⁺ T cells treated with CM from PI16-overexpressing fibroblasts (Fig. 2D). To investigate the impact of PI16 on the immunosuppressive activity of Tregs, we co-cultured CD8⁺ T cells with naïve CD4⁺ T cells, along with either recombinant protein PI16 or a PBS control. The proliferation and cytotoxicity of CD8⁺ T cells were assessed. The reulsts revealed a substantial reduction in cell proliferation and diminished IFN-γ secretion by CD8⁺ T cells in the PI16-treated co-culture group relative to the control (Fig. 2E-F).

Fig. 2.

PI16-overexpressing fibroblasts induce Tregs differentiation. A) Western blotting analysis of PI16 expression in NIH3T3 cells with PI16-overexpression. The Tublin was used as a loading control. B) The schematic of T cells co-cultured with PI16-overexpressing NIH3T3 or control cells. C) Representative FACS plots and summary of the percentage of Tregs (CD4⁺CD25⁺Foxp3⁺) in naïve CD4⁺ T cells co-cultured with NIH3T3-PI16 or control cells. D) Representative FACS plots and summary of the percentage of Tregs (CD4⁺CD25⁺Foxp3⁺) in naïve CD4⁺ T cells co-cultured with CM form NIH3T3-PI16 and control cells. E) Representative flow cytometry plots (left) and summary (right) of CFSE-based proliferation of CD8⁺ T cells co-cultured with naïve CD4⁺ T cells, along with either PI16 or a PBS control. F) Representative flow cytometry profiles (left) and summarized frequencies (right) of IFN-ɤ positive CD8⁺ T cells in CD8⁺ T cells co-cultured with naïve CD4⁺ T cells, along with either PI16 or a PBS control. G) The images of tumors formed in C57BL/6 mice by injecting MEC2 cells mixed with NIH3T3-PI16 or vector control cells. H) Tumor volume and tumor weight of tumors formed in (E). I) Representative FACS and summary of the percentage of Tregs (CD4⁺CD25⁺FOXP3⁺) in T cells isolated from tumors formed in (E). (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; ns, not significant)

To assess the effects of PI16 derived from fibroblasts on ESCC cell proliferation, the murine esophageal cancer cell line MEC2 was co-injected subcutaneously with either 3T3-PI16 or 3T3-vector cells into the flanks of C57BL/6 mice (n = 7 for each group). Our findings revealed a significant enhancement of tumor growth in PI16-overexpressing fibroblasts relative to control cells (Fig. 2G-H). The proportion of Tregs was significantly elevated in tumors derived from MEC2 + 3T3-PI16 cells compared to those originating from MEC2 + 3T3-Vec cells (Fig. 2I).

PI16 induces Tregs differentiation through DOCK2

To ascertain the impact of PI16 on Tregs induction by PI16-overexpressing fibroblasts, naïve CD4⁺ T cells were treated with various concentrations of PI16 recombinant protein for 5 days. TGF-β was used as a positive control. The results revealed a significant enhancement in the differentiation of Tregs from naïve CD4⁺ T cells following exposure to the PI16 recombinant protein (Fig. 3A, and Supplementary Fig. 2A). To explore the mechanism underlying PI16-mediated induction of Tregs from Naïve CD4⁺ T cells, these naïve CD4⁺ T cells were cultured with PI16 recombinant protein (5 µg/ml) at 37℃ for 4 h. Subsequently, cell lysates were harvested and subjected to co-immunoprecipitation assays using anti-PI16 antibodies or an IgG isotype control as a negative control. DOCK2 was identified as a PI16-binding target through mass spectrometry analysis (Fig. 3B). To confirm the interaction between PI16 and DOCK2, co-IP assays were conducted using anti-DOCK2 in naïve CD4⁺ T cells treated with PI16. The results indicated that DOCK2 binds to PI16 (Fig. 3B). As a member of the guanine nucleotide exchange factors (GEFs) family, DOCK2 plays a central role in immune surveillance in both humans and mice [23]. As shown in Fig. 3C, DOCK2 is predominantly expressed in T cells and macrophages/monocytes cells. A correlation analysis demonstrated a positive association between DOCK2 and PI16 expression levels (R = 0.495, P < 0.0001), as well as a significant correlation between DOCK2 expression and the infiltration of Tregs (R = 0.733, P < 0.0001) in the TCGA-ESCA dataset (Fig. 3D-E). Survival analysis indicated that patients with high DOCK2 expression exhibited a poorer prognosis in ESCA (Fig. 3F).

Fig. 3.

PI16-overexpressing fibroblasts induce Tregs differentiation via DOCK2. A) Flow cytometry analysis showing the differentiation of Tregs induced by recombinant PI16 protein from naïve CD4⁺ T cells, as indicated by the percentage of FOXP3⁺CD25⁺ cells. TGF-β served as a positive control. B) Venn diagram illustrating the proteins pulled down by anti-PI16 antibody and rabbit IgG control (top). Co-IP experiments were conducted using an anti-DOCK2 antibody to pull down PI16 in PI16-treated naïve CD4⁺ T cells harvested from mouse spleens (bottom). C) UMAP plots illustrating DOCK2 expression in cells from the GSE203115 dataset. D) Correlation between DOCK2 expression and PI16 expression in TCGA-ESCA dataset (http://timer.cistrome.org/). E) Correlation between DOCK2 expression and the infiltration levels of Tregs in TCGA-ESCC dataset (http://timer.cistrome.org/). F) Overall survival curves for patients stratified by DOCK2 expression in TCGA datasets (http://timer.cistrome.org/). G) GSEA identifies Rho GTPase pathway enrichment in PI16-induced Tregs relative to native splenic Tregs. H) Representative FACS and summary of the percentage of Tregs (FOXP3⁺CD25⁺) in T cells cultured with PI16 with/without CPYPP (DOCK2 inhibitor). (I) The image of tumors formed in C57BL/6 mice by injecting MEC2h cells mixed with NIH3T3-PI16 or vector control cells (NIH3T3-Vec). The mice were treated with CPYPP or saline control. (J) Tumor volume and tumor weight of tumors formed in (I). K) Survival curve analyses were conducted on the four animal groups that received either DOCK2 inhibitor CPYPP or a saline control (n = 7 per group). Survival differences between the groups were assessed using the log-rank (Mantel-Cox) test. (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001)

To clarify the specifity of PI16-induced Tregs differentiation, we performed RNA-seq analysis on PI16-induced Tregs and compared them to splenic Tregs from mice. Gene Set Enrichment Analysis (GSEA) demonstrated a significant enrichment of Rho GTPase signaling pathways in PI16-induced Tregs (P < 0.0001,|NES| >1.5; Fig. 3G). This was further corroborated by Gene Ontology (GO) analysis, which identified significant enrichment of genes associated with both Ras and Rho GTPase signaling pathways among the upregulated genes in PI16-induced Tregs (Supplementary Fig. 3). Given DOCK2’s pivotal role in modulating Rho GTPase signaling pathways [24], these findings suggest a crucial function of PI16-DOCK2 axis in PI16-induced Tregs differentiation.

To further validate whether PI16 induces Tregs differentiation through DOCK2, naïve CD4⁺ T cells were co-cultured with recombinant PI16 protein in the presence of CPYPP, a DOCK2 inhibitor. Flow cytometry analysis revealed that the DOCK2 inhibitor significantly reduced the proportion of Tregs induced by PI16 recombinant protein in a concentration-dependent manner (Fig. 3H, and Supplementary Fig. 2B). The PI16-induced Tregs were almost completely abolished at a CPYPP concentration of 2.5 µM (Fig. 3H). Mouse esophageal cancer cells, either mixed with PI16-overexpressing NIH3T3 cells or control cells, were subcutaneously injected into the flanks of C57BL/6 mice. Subsequently, the mice were administered intratumorally with either CPYPP or saline (n = 7 for each group). The findings indicated that the overexpression of PI16 in fibroblasts significantly augmented tumor growth relative to control fibroblasts (Fig. 3I-J). The inhibition of DOCK2 by CPYPP significantly attenuated the tumor growth enhancement conferred by PI16-overexpressing fibroblasts (Fig. 3I-J). The survival analysis demonstrated that PI16-overexpressing fibroblasts significantly reduced overall survival compared to control fibroblasts (Fig. 3K, P < 0.01). Notably, pharmacological inhibition of DOCK2 using CPYPP effectively reversed the PI16-mediated survival disadvantage, resulting in significantly improved survival outcomes compared to the saline group (Fig. 3K, P < 0.001).

Next, we assessed the expression levels of PI16 and the infiltration of Tregs within the tumors depicted in Fig. 3I-J using mfIHC (Fig. 4A). As Fig. 4B showed, the average cell density of PI16 in the stroma was significantly higher in the MEC + 3T3-PI16 groups compared to MEC + 3T3-Vec groups, irrespective of whether they received CPYPP or saline treatment (Fig. 4B). The proportion of Tregs in the stroma increased in the MEC + 3T3-PI16 group, but this increase was diminished when the mice were treated with the DOCK2 inhibitor (Fig. 4C). Consistently, in saline-treated groups, the quantity of Tregs located within 100 μm of PI16⁺ fibroblasts was significantly higher in the MEC + 3T3-PI16 group compared to the MEC + 3T3-Vec group (Fig. 4D). However, the difference was attenuated when both groups were treated with CPYPP (Fig. 4D). Additionally, CPYPP treatment led to a significant reduction in Tregs within the stroma and decreased the number of Tregs situated within a 100 μm proximity to PI16⁺ fibroblasts in the MEC2 + 3T3-PI16 groups (Fig. 4C-D). To further explore the influence of PI16 on the tumor microenvironment, CD8⁺ T effector (Teff) cells were analyzed using mfIHC in murine tumor tissues (Fig. 4E). Our findings showed a decreased mean density of Teff cells within the MEC + 3T3-PI16 group relative to the MEC + 3T3-Vec control group (Fig. 4F). When the mice were treated with CPYPP, the decrease was diminished (Fig. 4F).

Fig. 4.

The DOCK2 inhibitor impedes PI16-induced Tregs differentiation in vivo. A) Representative pictures of mfIHC staining illustrating the expression levels of PI16, pan-CK, CD4, and FOXP3 in tumors formed in Fig. 3I. The nuclei were counterstained by DAPI. (Scale bar = 50 μm) B) The fluorescence intensity of PI16 in the stroma of (A) was summarized (n = 7). C) The percentages of Tregs in the stroma of tumors were summarized (n = 7). D) The number of Tregs within 100 μm of PI16-positive stromal cells was summarized (n = 7). E) Representative pictures of mfIHC staining for CD8⁺ Teff cell markers (CD8⁺, CD44⁺, CD25⁺, CD62L⁻) in tumors formed in Fig. 3I (n = 7). The nuclei were counterstained by DAPI. (Scale bar = 50 μm) F) The average densities of Teff cells were summarized (n = 7). (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant)

PI16 induces T-bet⁺ Tregs in vivo

To elucidate the transcriptional signature of Tregs induced by PI16, RNA-seq data was analyzed and the results revealed a significant upregulation of TBX21 in PI16-induced Tregs compared to those isolated from mouse spleens. Correlation analyses of the TCGA-ESCA dataset further demonstrated that TBX21 expression is positively correlated with both PI16 (R = 0.359; P < 0.0001) and DOCK2 (R = 0.78; P < 0.0001) expression levels (Fig. 5A-B). MfIHC was employed to analyze the DOCK2⁺ Tbet⁺ Tregs (Fig. 5C). We found that DOCK2⁺Tbet⁺ Tregs were more abundant in tumors with PI16-overexpressing fibroblasts compared to those with control cells (Fig. 5D). Furthermore, treatment with a DOCK2 inhibitor significantly reduced the density of DOCK2⁺Tbet⁺ Tregs in tumor tissues (Fig. 5D).

Fig. 5.

Tbet is upregulated in PI16-induced Tregs. A) Correlation between the RNA levels of TBX21 and PI16 in TCGA-ESCA dataset. B) Correlation between the RNA levels of TBX21 and DOCK2 in TCGA-ESCA dataset. C) Representative images of mfIHC staining for DOCK2, Tbet, and Tregs markers (CD4⁺FOXP3⁺) were represented. The nuclei were counterstained by DAPI. (Scale bar = 50 μm). D) The average densities of DOCK2⁺Tbet⁺ Tregs were summarized. (*, P < 0.05)

PI16 upregulation in the stroma is predictive of poor survival in patients with ESCC

To explore the clinical implications of fibroblasts with PI16 upregulation, mfIHC was utilized to assess both PI16 expression in the tumor stroma and the extent of Tregs infiltration within ESCC tissues (Fig. 6A). Patients were categorized into two cohorts based on the quartiles of PI16 fluorescence intensity in the stroma: PI16-upregulation group (upper 75%) and PI16-downregulation (lower 25%). Survival analysis indicated that PI16-upregulation in tumor stroma correlated with poorer overall survival in patients with ESCC (P < 0.05)(Fig. 6B). The median survival time for the PI16-upregulation group is 6.33 years, compared to 10.09 years for the PI16-downregulation group. Additionally, a positive correlation was observed between the average cell intensity of PI16 and Tregs infiltration in the stroma (Fig. 6C). The average density of Tregs was significantly higher in tumor tissues of patients with elevated PI16 levels (Fig. 6D). The infiltration of Tregs was also evaluated in both tumor and stroma tissues. The proportions of Tregs were significantly elevated in the PI16-upregulation groups compared to the PI16-downregulation groups (Fig. 6E). Notably, the number of Tregs within 100 μm of PI16⁺ cells in the stroma was significantly greater in the PI16-upregulation group compared to that in the PI16-downregulation group (Fig. 6F). Additionally, the percentage of Teff cells was examined using mfIHC (Fig. 6G). The results indicated that the average density of CD8⁺ Teff cells was lower in the PI16-upregulation group than in the PI16-downregulation group (Fig. 6H).

Fig. 6.

The overexpression of PI16 in stroma correlates with poor survival in patients with ESCC. A) Representative pictures of mfIHC staining for PI16, pan-CK, and Tregs markers (CD4⁺FOXP3⁺). The nuclei were counterstained by DAPI. (Scale bar = 50 μm) B) Overall survival curves stratified by PI16 expression in fibroblasts (n (PI16 upregulation in stroma) = 150; n(PI16 downregulation in stroma) = 50). C) Correlation between the fluorescence intensity of PI16 in the stroma and the average density of Tregs. D) The average densities of Tregs in the PI16 upregulation and downregulation groups were summarized. E) The percentage of Tregs in tumor tissues (left) and stromal tissues (right) in the PI16 upregulation and downregulation groups. F) The number of Tregs within 100 μm of PI16-positive cells in the stroma was summarized. G) Representative pictures of mfIHC staining for CD8⁺ Teff cell markers (CD8⁺CD25⁺CD62L^-). The nuclei were counterstained by DAPI. (Scale bar = 50 μm) H) The average density of CD8⁺ Teff cells in the PI16 upregulation and downregulation groups (n (PI16 upregulation in stroma) = 112; n(PI16 downregulation in stroma) = 39)

Taken together, our findings indicate that PI16⁺ fibroblasts foster a tumor-suppressive microenvironment through mechanisms involving enhanced Tregs recruitment and concurrent suppression of Teffs infiltration. PI16⁺ fibroblasts can induce Tregs from naïve CD4⁺ T cells by direct interaction with DOCK2. Notably, inhibition of DOCK2 mitigates this Treg-inducing capacity (Fig. 7).

Fig. 7.

The proposed model illustrates that PI16 from fibroblasts induces differentiation of Tregs via DOCK2

Discussion

Tumor development and progression are closely associated with alterations in the surrounding tumor microenvironment (TME). Fibroblasts play critical roles within TME by maintaining tissue architecture through the production of extracellular matrix, interacting with immune cells, and contributing to a wide range of processes in cancer [25]. For instance, cancer-associated fibroblasts (CAFs) secrete a variety of proteins that are pivotal for tumor biology, influencing processes such as extracellular matrix deposition and remodeling in interstitial tissues [26, 27]. With the advances in single-cell RNA sequencing (scRNA-seq), studies have revealed the high heterogeneity of CAFs, allowing their classification into distinct subtypes based on specific biomarkers. Myofibroblasts (myCAFs) are characterized by biomarkers such as αSMA, THY1, and COL12A1 [28, 29], while inflammatory CAFs (iCAFs) are defined by markers like CLEC3B, COL14A1, and LY6C [30]. In colorectal cancer, two distinct subtypes of CAFs, CAF-A, and CAF-B, have been identified based on the expression of specific biomarkers. CAF-A is marked with high expression levels of FAP, MMP2, and DCN, whereas CAF-B expresses αSMA, TAGLN, and PDGFA [31]. Studies have also highlighted the heterogeneity of fibroblasts in ESCC tissues [4, 32]. Distinct subsets of fibroblasts have been identified and validated for their potential biological significance and prognostic values [32]. For example, a tumor-specific subset of CST1⁺ myofiblasts has been associated with poor prognosis in ESCC, and a subset of antigen-presenting fibroblasts expressing MHC class II molecules has been revealed in ESCC [32].

Buechler and colleagues used publicly available mouse and human scRNA-seq datasets to construct a fibroblast transcriptional atlas [9]. They found that the fibroblast lineage cells could be categorized into either context-specific clusters or clusters present across all tissues. These conserved clusters expressed Dpt and could be further characterized by their expression of Pi16 and Clo15a1 [9]. In this study, data-mining results of scRNA-seq dataset (GSE203115) revealed that CAFs could be divided into two groups: PI16⁺ fibroblasts and PI16⁻ fibroblasts. Murine esophageal cancer cells mixed with PI16-overexpressing fibroblasts (NIH3T3-PI16) were injected into the flanks of immune-competent C57BL/6 mice, demonstrating that PI16⁺ fibroblasts significantly enhance EC tumor growth in vivo. This finding contrasts with our previous results obtained in immune-incompetent nude mice [15]. Consequently, we investigated whether PI16⁺ fibroblasts interact with other immune cells.

The tumor microenvironment encompasses both innate immune cells, including macrophages, neutrophils, dendritic cells, innate lymphoid cells, myeloid-derived suppressor cells, and natural killer cells, as well as adaptive immune cells such as T and B cells. Additionally, it comprises stromal cells, fibroblasts, and epithelial cells [3]. PI16-expressing reticular cells (PI16⁺ RC) from human tonsils exhibit the most significant inflammation-associated structural remodeling and regulate T cell activity in specific subepithelial niches [33]. A comprehensive cross-tissue atlas of human fibroblasts suggests that a PI16⁺ fibroblast subset is associated with immune infiltration in perivascular niches [34]. Cell-cell interaction analyses indicate that PI16⁺ fibroblasts interact strongly with Tregs, and bulk RNA-seq analyses suggest that PI16 expression levels correlate with Treg infiltration in the TCGA-ESCA dataset. Both in vitro and in vivo assays demonstrate that fibroblast-derived PI16 promotes the conversion of naïve CD4⁺ T cells into Tregs.

DOCK2, a kinase predominantly expressed in leukocytes, plays a crucial role in lymphocyte migration [35]. High expression of DOCK2 correlates with a favorable prognosis in acute leukemia [36]. Deficiency of DOCK2 in lymphocytes impairs their ability to homing to the spleen and lymph nodes [37]. Furthermore, DOCK2 is indispensable for T-cell receptor (TCR)-mediated Rac activation [38] and modulates effector T-cell exhaustion within the hepatocellular carcinoma microenvironment [39]. In this study, co-IP and mass spectrometry analyses revealed that PI16 co-precipitates with DOCK2. Targeting DOCK2 effectively inhibited the PI16-induced Treg differentiation both in vitro and in vivo. Our findings propose a novel mechanism by which PI16⁺ fibroblasts drive Treg differentiation through DOCK2 in ESCC.

Single-cell sequencing has revealed that Tregs exhibit significant heterogeneity within tumors and correlate with poor prognosis in patients with cancer [40, 41]. To explore the characteristics of Tregs induced by PI16, we performed RNA-seq analyses on both PI16-induced Tregs and Tregs isolated from murine spleens. Our results showed that TBX21, encoding for T-bet, was significantly upregulated in PI16-induced Tregs. T-bet is a transcription factor known for its role in regulating TH₁ cell responses and maintaining the homeostasis and function of Tregs, particularly during type 1 inflammation [42]. Recent studies have highlighted the critical role of T-bet⁺Foxp3⁺ Tregs in the suppression of tumor immunity [43–45]. In animal experiments, Tbet⁺DOCK2⁺ Tregs were elevated in tumors of murine EC cells mixed with the PI16⁺ fibroblasts group, while increases were attenuated when animals were treated with a DOCK2 inhibitor. Conversely, Teff cells exhibited opposite changes. These results indicate that PI16⁺ fibroblasts induce a tumor-suppressive microenvironment by promoting Treg conversion during ESCC progression. Furthermore, clinical samples of ESCC were stratified by PI16 expression levels in the stroma, revealing that PI16 upregulation in the stroma is associated with poor overall survival in patients with ESCC. Consistent with the animal study results, Treg cells increased and Teff cells decreased in ESCC clinical samples.

Conclusions

In conclusion, our work provides novel insights into the role of PI16⁺ fibroblasts in reshaping the tumor microenvironment of ESCC by inducing Tregs. The downregulation of PI16 in the stroma may serve as a prognostic indicator for long-term survival in patients with ESCC. Moreover, given the accumulating evidence on the therapeutic impact of targeting the TME, PI16⁺ fibroblasts could represent a promising new therapeutic target to overcome tumor immune suppression.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

ESCC

Esophageal squamous cell carcinoma

TME

Tumor microenvironment

CAFs

Cancer-associated fibroblasts

ECM

Extracellular matrix

PI16

Peptidase inhibitor 16

Tregs

Regulatory T Cell

CAP

Cysteine-rich secretory proteins, antigen 5, and pathogenesis-related 1 proteins

MEC2

Mouse esophageal cancer cells

CM

Conditioned medium

IP

Immunoprecipitation

mfIHC

Multiplex fluorescent immunohistochemistry

scRNA-seq

Single-cell RNA sequencing

GEO

Gene Expression Omnibus

UMAP

Uniform manifold approximation and projection

TCGA

The Cancer Genome Atlas

ESCA

Esophageal cancer

DOCK2

Dedicator of cytokinesis protein 2

GO

Gene Ontology

iCAFs

Inflammatory CAFs

TCR

T-cell receptor

Author contributions

D.S. performed the experiments, analyzed data and wrote the draft; L.L., Y.Z., S.L. and T.Z. performed the experiments; Z.X. performed bioinformatics analyses; X.Y.G. provided funding support; Y.L. provided funding support, and supervised the study and wrote the manuscript. All authors have read and approved the final version of the manuscript.

Funding

This study was supported by National Natural Science Foundation of China (82473272); Guangdong Esophageal Cancer Institute Science and Technology Program (M202203 and Q202206). Guangdong Basic & Applied Basic Research-Hybribio Joint Fund (2022A1515220091).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The study was approved by the Committees for Ethical Review of Research Involving Human Subjects at SYSUCC with the number. G2023-075. This study was conducted following the ethical guidelines of the Declaration of Helsinki. All animal experiments were approved by the Animal Ethics Committee at SYSUCC with the number. L102012022228M.

Compliance with ARRIVE guidelines

We confirm that all animal experiments were conducted in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines 2.0. A completed ARRIVE checklist has been included as Supplementary File to ensure transparent and comprehensive reporting of our study.

Competing interests

The authors declare no competing interests.