Extracellular vesicles promote activation of pro-inflammatory cancer-associated fibroblasts in oral cancer

Abstract

Introduction: Oral squamous cell carcinoma (OSCC) is the most common form of head and neck cancer and has a survival rate of ∼50% over 5 years. New treatment strategies are sorely needed to improve survival rates-and a better understanding of the mechanisms underlying tumorigenesis is needed to develop these strategies. The role of the tumor microenvironment (TME) has increasingly been identified as crucial in tumor progression and metastasis. One of the main constituents of the TME, cancer-associated fibroblasts (CAFs), plays a key role in influencing the biological behavior of tumors. Multiple mechanisms contribute to CAF activation, such as TGFβ signaling, but the role of extracellular vesicles (EVs) in CAF activation in OSCC is poorly understood. Assessing the impact of oral cancer-derived EVs on CAF activation will help to better illuminate OSCC pathophysiology and may drive development of novel treatments options. Methods: EVs were isolated from OSCC cell lines (Cal 27, SCC-9, SCC-25) using differential centrifugation. Nanoparticle tracking analysis was used for EV characterization, and Western blot to confirm the presence of EV protein markers. Oral fibroblasts were co-cultured with enriched EVs, TGFβ, or PBS over 72 h to assess activation. Flow cytometry was used to evaluate CAF markers. RNA collected from fibroblasts was extracted and the transcriptome was sequenced. Conditioned media from the co-cultures was evaluated with cytokine array profiling. Results: OSCC-derived EVs can activate oral fibroblasts into CAFs that are different from those activated by TGFβ, suggesting different mechanisms of activation and different functional properties. Gene set enrichment analysis showed several upregulated inflammatory pathways in those CAFs exposed to OSCC-derived EVs. Marker genes for inflammatory CAF subtypes were also upregulated, but not in CAFs activated by TGFβ. Finally, cytokine array analysis on secreted proteins revealed elevated levels of several pro-inflammatory cytokines from EV-activated CAFs, for instance IL-8 and CXCL5. Discussion: Our results reveal the ability of OSCC-derived EVs to activate fibroblasts into CAFs. These CAFs seem to have unique properties, differing from TGFβ-activated CAFs. Gaining an understanding of the interplay between EVs and stromal cells such as CAFs could lead to further insights into OSCC tumorigenesis and potential novel therapeutics.

Keywords: cancer-associated fibroblast; extracellular vesicles; inflammation; oral squamous cell carcinoma; tumor microenvironment; tumorigenesis.

FIGURE 1

(A) Identification, quantification and characterization of representative EV particles from Cal 27, SCC-9 and SCC-25 cells according to NanoSight. Size distribution in nm, error bars indicate ± 1 SEM. (B) Protein from OSCC cell and EV protein tested for the vesicle markers Alix, HSP70, Flotillin-1, and CD81 together with GRP94 (negative marker for small EVs) with Western blot. (C)–(G) Flow cytometry analysis of percentage of fibroblasts positive for CAF markers upon co-culture with Cal 27, SCC-9, or SCC-25 derived EVs compared to PBS or TGFβ. (C) FAP, (D) PDPN, (E) CD29, (F) PDGFRβ, and (G) αSMA. n = 4. Statistics: Ordinary one-way ANOVA with Tukey’s multiple comparison test.

FIGURE 2

Overview of RNA sequencing of fibroblasts after treatment with OSCC cell line derived EVs or TGFβ. (A) Proportion of genes significantly up and downregulated compared to PBS control cells. Venn diagrams indicating overlap of upregulated (B) and downregulated (C) genes between treatment groups. (D) Volcano plots of differentially expressed genes in the four treatment types compared to PBS control. The top ten genes showing the highest combination of differential expression (taking into account fold change and statistical significance) are indicated. (E) Log2 fold change values for each treatment group for the top ten most highly differentially expressed genes in at least one sample (as shown in Figure 4D).

FIGURE 3

Enrichment map visualization. Commonalities between different enriched gene sets visualized through network analysis. Nodes are gene sets enriched (red circles) or depleted (blue circles). Lines link nodes with overlapping genes (bold line: high significance, thin line: moderate significance). Enrichment map visualization in GSEA was used together with Cytoscape 3.9.1. Enrichment Map Parameters: p-value cutoff: 0.05; FDR Q-value cutoff: 0.1; overlap coefficient: 0.5.

FIGURE 4

Specific marker genes related to pan-CAF subsets (right). Expression high (yellow), middle (white), low (blue), missing data (grey). Marker genes linked to pan-iCAF and pan-iCAF-2 marked with red rectangle. Z-scores calculated from FPKM values in differential expression analysis of fibroblasts co-cultured with TGFβ (TGFb), EVs from Cal 27 (CAL27), SCC-9 (SCC9) or SCC-25 (SCC25) cells, compared to fibroblasts co-cultured with PBS.

FIGURE 5

Cytokine antibody array profiles (A–F) and ELISA (G–H) of conditioned media from HOrF cell culture media upon 72 h of co-culture. (A) Co-culture with PBS. (B) Co-culture with EVs derived from Cal 27 cells. (C) Co-culture with EVs derived from SCC-9 cells. (D) Co-culture with EVs derived from SCC-25 cells. Positive control spots are outlined in green and negative in gray. Spots corresponding to IL-8 are highlighted in blue and CXCL5 in red. (E) Bar graphs indicating increased protein expression of IL-8 and (F) CXCL5. ELISA on (G) IL-8 and (H) CXCL5.