FGF8 induces therapy resistance in neoadjuvantly radiated rectal cancer

Abstract

Purpose: Therapy response to neoadjuvant radiochemotherapy (nRCT) of locally advanced rectal cancer varies widely so that markers predicting response are urgently needed. Fibroblast growth factor (FGF) and FGF receptor (FGFR) signaling is involved in pro-survival signaling and thereby may result in radiation resistance.

Methods: In a cohort of 43 rectal cancer patients, who received nRCT, we analyzed protein levels of FGF 8 and its downstream target Survivin by immunohistochemistry to assess their impact on nRCT response. In vitro resistance models were created by exposing colorectal cancer cell lines to fractionated irradiation and selecting long-term survivors.

Results: Our findings revealed significantly higher FGF8 and Survivin staining scores in pre-treatment biopsies as well as in surgical specimens of non-responsive compared to responsive patients. Functional studies demonstrated dose-dependent induction of FGF8 mRNA expression in mismatch-incompetent DLD1 cells already after one dose of irradiation. Surviving clones after one or two series of radiation were more resistant to an additional radiation fraction than non-irradiated controls and showed a significant increase in expression of the FGF8 receptor FGFR3 and of Survivin on both the RNA and the protein levels.

Conclusion: The results of this study suggest that FGF8 and Survivin contribute to radiation resistance in rectal cancer and may serve as markers to select patients who may not benefit from neoadjuvant radiotherapy.

Keywords: Fibroblast growth factor 8; Neoadjuvant radiochemotherapy; Rectal cancer; Survivin; Therapy response.

Fig. 1

FGF8 and Survivin staining in responders and non-responders. Serial sections of pre-treatment biopsies were stained for FGF8 and Survivin. Patients were classified into responders and non-responders according to Dworak et al. (1997). a, b, d, e All specimens showed positive staining, but staining intensity differed between responders and non-responders. The bar corresponds to 200 µm. c, f Staining intensity was scored as described in “Materials and methods” and compared between responders and non-responders for FGF8 (b) and Survivin (c). g Pearson correlation coefficient was computed for IRS values of FGF8 and Survivin. (p = 0.02; r² = 0.14; y = 5, 13 + 0, 37x). h Specimens were grouped into high FGF8 and low FGF8 expression as described in Table 2 and their Survivin IRSs were analyzed by Students’ t test. *Statistical significance at p < 0.05. ***Statistical significance at p < 0.001

Fig. 2

FGF8 and Survivin levels in surgical specimens of patients with incomplete response. Surgical specimens after nRCT of patients, who had a less than complete response were stained for FGF8 (a–c), Survivin (d–f), the alternative ligand FGF18 (g–i) and the proliferation marker Ki-67 (j–l). IRS was determined as described and compared between responders and non-responders (c, f, i, l). Size bar represents 200 µm. *Statistical significance at p < 0.05. ***Statistical significance at p < 0.001

Fig. 3

FGF8 and Survivin expressions in irradiated cell models. a DLD1 cells were irradiated with doses of γ-radiation of 2, 4 and 6 Gy. 24 h later, RNA was isolated and mRNA levels for FGF8 and FGFR3 were determined by qRT-PCR using TaqMan kits and the ΔΔct method. The house-keeping gene for normalization was GAPDH. Bars represent the mean ± SD of two independent experiments. b, c Logarithmically growing cultures of DLD1 cells were starved for 24 h before 10 ng/ml FGF8 was added to the medium. 24 and 48 h later, lysates were harvested for isolation of RNA and proteins. b Survivin mRNA levels were determined by qRT-PCR using TaqMan kits and the ΔΔct method. The house-keeping gene for normalization was GAPDH. c Survivin protein was determined by western blot. d–i Cultures of DLD1, HCT116, and HT29 cells were exposed to 5 (IR1) or 10 (IR2) doses of 2 Gy irradiation. Long-term-surviving cell clones were collected and clone pools were expanded for analysis. d DLD1 control cells, DLD1-IR1 and DLD1-IR2 were exposed to an additional dose of 2 Gy irradiation and radiation sensitivity was determined by the clonogenic assay. e–h mRNA levels in growing cultures were determined by qRT-PCR using TaqMan kits and the ΔΔct method with GAPDH as the house-keeping gene and non-irradiated cultures of equal density as the control. e Survivin, f FGF8, g FGFR3-IIIb, h FGFR3-IIIc. i Protein lysates were analyzed by western blotting using antibodies to Survivin, FGFR3 and p-FGFR. The figure shows representative panels from two independent experiments