Xeroderma Pigmentosum Type C Primary Skin Fibroblasts Overexpress HGF and Promote Squamous Cell Carcinoma Invasion in the Absence of Genotoxic Stress

Abstract

Xeroderma pigmentosum (XP) is a very rare recessive disease caused by the incapacity to resolve ultraviolet-induced DNA lesions through Nucleotide Excision Repair (NER). Most XP patients suffer from aggressive skin carcinoma and melanoma at a very early age (<8). Our previous results showed that primary XP fibroblasts isolated from healthy (non-photo-exposed) skin negatively impact the extracellular matrix and fail to activate the innate immune system. Here, we show for the first time that XP-C fibroblasts also play a major role in cancer cell invasion ex vivo and in vivo through the overexpression of Hepatocyte Growth Factor/Scatter Factor (HGF/SF) in the absence of genotoxic attacks. The use of inhibitors of the activation of the HGF/SF pathway counteracted the effects of XP fibroblasts on the growth of cancer cells, suggesting new perspectives in the care of XP patients.

Keywords: HGF/SF; fibroblasts; squamous cell carcinoma; xeroderma pigmentosum.

Figure 1

XP-C fibroblasts promote cancer cell invasion in organotypic skin cultures. (A) Representative images of H&E staining of sections from organotypic cultures generated from either WT keratinocytes (KTM1) or SCC13 cells together with either WT (FH29) or XP-C (AS433, AS798) fibroblasts embedded in dermal equivalents; bar: 50 µm. (B) Quantification of SSC13 cell invasion. Values are presented as the mean ± SD of three independent areas; *** p < 0.0001. (C) Immunofluorescence staining of K14 keratin (green) and nuclei (blue) show that invading cells are of epithelial origin; bar: 50 µm. (D) Spheroid assays composed of SCC13-GFP cells and either (red) WT (FH29) or XP-C (AS798) fibroblasts; bar: 250 µm. (E) Confocal microscopy images of spheroids showing SCC13-GFP cells and XP-C fibroblasts. Insets show higher magnification of spheroids. Arrowheads indicate the XP-C fibroblasts leading to invading protrusions. Bar: 40 µm.

Figure 2

XP-C fibroblasts increase the rate of scratch closure of cancer cells ex vivo. Effects of either WT or XP-C fibroblast CM on SCC cells’ scratch closure without (A) or with MMC (B). Upper panels show representative micrografts from the time-lapse video microscopy of SCC13 cells’ scratch closure in the presence of either WT or XP-C CM at indicated times. Lower panels show the kinetics of SCC13 cells’ scratch closure in the presence of either 2 strains of XP-C CM (XP-C-AS433, XP-C-AS798) or 1 strain of WT fibroblast CM (WT-FH29). Dash lines indicate merging of the scratch. Note the substantially slower scratch closure kinetics of SCC13 cells in the presence of MMC (please also see Figure S2).

Figure 3

XP-C fibroblasts overexpress HGF/SF, leading to downstream activation of STAT3 and JNK c-MET pathways in SCC cells. (A) Accumulation of HGF mRNA was significantly higher in XP-C (n = 4) (FXP-C) than in WT (n = 2) fibroblasts (FWT). Values are presented as the mean ± SD of 4 independent experiments, *** p < 0.001. (B) HGF levels (ELISA) were significantly higher in XP-C fibroblast CM (n = 3) than in WT fibroblast CM (n = 3); ** p < 0.01. (C) Immunoblot analysis of cMET and p-MET in SCC cells treated with CM of either WT (FH29, FH84) or XP-C (AS148, AS433, AS875, AS798) fibroblasts. (D–G) Western blot analysis of phosphorylated forms of STAT3, JNK1,2, ERK1, 2, and P38 in SCC13 cells cultured in CM of WT (FTM5, FH29) or XP-C (AS148, AS433, AS875, AS798) primary fibroblasts. β-tubulin was used as the loading/transfer control. Right panel quantifies the ratio of phosphorylated to total protein amounts. Values are the mean ± SD of two independent experiments. *** p < 0.001, * p < 0.05.

Figure 4

Blockade of cMET/HGF in culture media conditioned by primary XP-C fibroblasts reduces activation of JUNK and STAT3. (A) Western blot analysis of total cMET and p-MET levels in lysates of SCC cells cultured in XP-C fibroblast CM (AS875, AS798) ± HGF blocking antibody. GAPDH was used as the loading/transfer control. Right panel shows quantification of p-MET accumulation ± anti HGF-blocking antibody. (B) Western blot analysis of p-STAT3 and p-JNK levels in lysates of SCC cells cultured in the presence of XP-C fibroblast CM (AS798) ± HGF blocking antibody. Tubulin was used as the loading/transfer control. Right panel shows the quantification of the ratio of p-STAT3 and p-JNK to the total protein levels under the different conditions. Values are the mean ± SD. *** p < 0.001; * p < 0.05). For Figure 3 and Figure 4, raw images of blots are shown in Figure S4.

Figure 5

Contractile and invasive properties of cancer cells are normalized by HGF antibodies in XP-C CM. (A) Organoids containing either WT (FH29) or XP-C (AS433, AS798) fibroblasts ± HGF blocking antibody; quantitation is showed in the right panel. (B) Organotypic cultures composed of SCC cells and XP-C (AS798) fibroblasts ± HGF blocking antibody. Sections were stained using either H&E, DAPI (blue), or K14 antibody (green). Cancer cell invasion was significantly reduced by the HGF antibody; bar: 50 µm. Values are the mean ± SD. ** p < 0.01, *** p < 0.001.

Figure 6

Contribution of XP-C fibroblasts to carcinoma cell invasion in vivo. (A) Numbers and tumor volume were measured in either absence or presence of XP-C fibroblasts. (B) In contrast to WT primary fibroblasts (FTM1, FTM5, FH84, FH29; representative image in upper panel), XP-C primary fibroblasts (AS 433, AS 875, AS 875, AS 673, AS 202; lower panel) accumulated in the vicinity of epithelial cancer cell invasions (representative image of graft sections from xenografts). Labeling colors are indicated; bar: 100 µm.

Figure 7

XP-C primary fibroblasts accumulate higher ROS levels than WT primary fibroblasts after treatment with either H₂O₂ or UVA, as indicated in the figure. Pool of 2 independent experiments using WT FH 84 primary fibroblasts as controls; WT, n = 2; XP-C, n = 3. Asterisks indicate statistical significance: ** p < 0.01, *** p < 0.001.