2.1 Clinical tissue samples

In total, 60 newly diagnosed PC patients at the First People’s Hospital of Jiashan County were recruited, and PC tissues and adjacent normal tissues were collected during surgery. All patients only received GEM-based chemotherapy before surgery, then pathological examinations were carried out, of which 33 samples were resistant and the rest (n = 27) were sensitive. Specimens were stored at -80℃. The clinicopathologic features of PC patients has been provided in Table S1. This experiments was allowed to perform by the ethical committee of the First People’s Hospital of Jiashan County and was carried out according to the guidelines of Declaration of Helsinki. Informed consent was obtained from all subjects.

2.2 Cell culture and treatment

Normal fibroblasts (NFs) and CAFs were isolated using fresh normal and resistant PC tissues as described previously [21]. Then cells were cultured in DMEM (Cat#PM150210, Procell, Wuhan, China) comprising 10% FBS (Cat#164210-50, Procell) at 37 °C with 5% CO2. The medium was changed every 3 days. Cells passed to 2–5 generations were used for functional experiments.

Human pancreatic duct epithelial (HPDE) cell line (Cat#CL-0921, Procell) was cultured in keratinocyte serum-free medium containing epidermal growth factor, bovine pituitary extract (Cat#:17005042, Invitrogen, Carlsbad, CA, USA) and 1% penicillin/streptomycin (Cat#PB180120, Procell) at 37 °C with 5% CO2. PC cell lines PANC-1 (Cat#CL-0184) and BxPC-3 (Cat#CL-0042) were purchased from Procell, and cultured in DMEM (Cat#PM150210, Procell) plus 1% penicillin/streptomycin (Cat#PB180120, Procell) and 10% FBS (Cat#164210-50, Procell) at 37℃with 5% CO2. To establish GEM resistant (GR) PC cells, PANC-1 and BxPC-3 cells of 4–5 passages were exposed to increasing doses of GEM (0.1–10 μM) for about 6 months, named PANC-1/GR and BxPC-3/GR. GEM resistant cells were cultured in same completed DMEM containing 5 μM GEM at 37 °C with 5% CO2 to maintain cell resistant phenotype.

2.3 Cell transfection

COL17A1-specific small interfering RNA (siRNA), named si-COL17A1, and nontargeted siRNA (si-NC) as well as ACTN4 overexpressing plasmid (ACTN4) and scramble plasmid (Vector) were designed and provided by GenePharma (Shanghai, China). Lentiviral particles containing COL17A1-specific short hairpin RNA (shRNA) (sh-COL17A1) or nontargeted shRNA (sh-NC) were provided by Hanbio (Shanghai, China) for stable transfection in CAFs.

2.4 Conditioned medium (CM) collection and treatment

NFs, CAFs or CAFs transfected with si-NC, si-COL17A1, sh-NC or sh-COL17A1, were cultured in 10% FBS-contained DMEM overnight. The culture medium of cells were then collected and filtered using a 0.22 μM filter, followed by refreshing with serum-free medium for 48 h. Finally, the CM was harvested through the centrifugation. GEM-resistant PC cells (2 × 105/well) were seeded in 6-well culture plates and CM was added to achieve a concentration of 10% (total volume per well = 3 mL), followed by incubation for 2 days as the experiment design.

2.5 Detection of the half maximal inhibitory concentration (IC50) values

PC cell lines, and resistant cells with or without indicated transfection and/or assigned CAF incubation were placed into a 96-well plate (5 × 103 cells/well) containing different concentrations of GEM and incubated for 72 h. Then 20 μL MTT (Cat# C0009S, 5 mg/mL) (Beyotime, Beijing, China) was added into each well after remove the medium. 2 h later, the MTT crystals were removed by adding 150 μL DMSO. Finally, the absorbance was detected at 490 nm to calculate IC50 values.

2.6 Western blotting

Total proteins were isolated by using RIPA lysis buffer (Cat#P0038, Beyotime) and then loaded onto SDS-PAGE on 10% gels for separation, followed by shifting onto nitrocellulose membranes. Then membranes were stained with primary antibodies including MDR1 (ab170904, 1:2000), COL17A1 (ab184996, 1:1000), ACTN4 (ab108198, 1:2000) and GAPDH (1:5000, ab181602) all night at 4 °C (Abcam, Cambridge, MA, USA), and HRP-conjugated secondary antibodies (ab6728 or ab6721, Abcam) at 37 °C for 2 h. Protein signals were analyzed using the ECL detection system (Cat#P0018AS, Beyotime). The bands of gray intensity were analyzed by Image J software.

2.7 MTT assay

Assigned resistant cells were seeded into a 96-well plate (5000 cells/well) and incubated with 20 μL MTT (Cat# C0009S, 5 mg/mL) (Beyotime) for 2 h incubation, followed by DMSO treatment. Then the absorbance at 490 nm was examined.

2.8 5-Ethynyl-2ʹ-deoxyuridine (EdU) assay

Assigned resistant cells (2 × 103 cells/well) were incubated with 50 μM EdU (Cat# C10310-2, RiboBio, Guangzhou, China) in serum-free DMEM in a 96-well plate for 2 h. Then cells were fixed with 4% formaldehyde for 30 min and permeabilized by 0.3% Triton X-100 for 10 min, and 100 μL 1 × Apollo reaction cocktail was added and incubated with cells for 30 min. Cell nuclei were counterstained with 100 μL DAPI. EdU-positive cells were determined using a fluorescence microscopy.

2.9 Transwell assay

About 100 µL serum-free media-dissolved Matrigel was pre-added onto the upper chamber of Transwell inserts for 1 h. Assigned resistant cells were placed onto the upper chamber with serum-free media. 600 µL complete medium was added into the basolateral chambers. 24 h later, cells were stained with 0.1% crystal violet (Cat#C0121, Beyotime), and invaded cells in five randomly selected fields were counted.

2.10 Sphere formation assay

GEM resistant cells (1 × 105) with indicated transfection and/or incubation were placed onto a 6-well plate with 2 mL serum-free DMEM containing 20 mg/L hFGF (Cat#GF003, Merck Millipore, Billerica, MA, USA), 20 mg/L EGF (Cat#E5160, Merck Millipore), 4 U/L insulin (Cat#FCP001, Kerafast) and 1% penicillin/streptomycin (Cat#PB180120, Procell). Every other day, the medium was changed. Sphere formation efficiency was calculated.

2.11 Quantitative real-time PCR (qRT-PCR)

TRIzol reagent (Cat#15596018CN, Invitrogen) was utilized for the isolation of total RNAs, which were then reversely transcribed into cDNA using PrimeScript RT Master Mix (Cat#RR036B, TaKaRa, Dalian, China). Thereafter, qRT-PCR amplification was carried out with the SYBR Premix ExTaq (Cat#RR041A, TaKaRa). The fold change was calculated by the 2−ΔΔCt method with GAPDH as the housekeeping gene. Table 1 listed the primers for qRT-PCR.

Table 1 Primers sequences used for qRT-PCR2.12 Co-immunoprecipitation (Co-IP) assay

GEM resistant cells were lysed using cell lysis buffer and then incubated with anti-COL17A1 (ab184996), anti-ACTN4 (ab108198) (0.5 μL per mg lysates) or normal rabbit IgG (ab313801) (Abcam) for 12 h at 4 °C. After incubating with Protein A/G agarose for 4 h (Cat# 20421, Thermo Scientific, Pierce, Waltham, MA, USA), enriched protein complexes were eluted for western blotting.

2.13 Immunofluorescence analysis

PANC-1/GR cells transfected with si-NC or si-COL17A1 were stained with COL17A1 (ab184996) or ACTN4 (ab108198) antibody (Abcam) overnight at 4 °C in the dark, followed by a secondary antibody conjugated with Alexa Fluor 488 (Cat#A-11094, Invitrogen) for 1 h at 37 °C. Cell nucleus was stained with DAPI (Cat#, 62248, 1 mg/mL; Invitrogen) for 5 min away from light. Finally, images were observed by a fluorescence microscope.

2.14 Animal experiments

Lentiviral particles carrying sh-NC or sh-COL17A1 were added into the CAFs with multiplicity of infection (MOI) of 15. The CM of infected CAFs was derived and incubated with PANC-1/GR cells. Then stably infected PANC-1/GR cells (3 × 106) were subcutaneously injected into the left dorsal flank of BALB/c nude mice (Six-week-old, Slaike Jingda Laboratory, Hunan, China). Mice were divided into three group: PANC-1/GR, PANC-1/GR + sh-NC-CAFs or PANC-1/GR + sh-COL17A1-CAFs (n = 5/per group). Tumor size was recorded every 3 days to assess the volume. At days 27, mice were euthanized by deeply anesthetized with pentobarbital sodium (100 mg/kg, i.p.) and then cervical dislocation. Next, xenograft tumors were isolated and weighed. The animal experiments were authorized by the Animal Ethical Committee of First People’s Hospital of Jiashan County. The maximal tumor size/burden permitted by the Ethics Committee is no more than 2000 mm3 in volume. No mouse has violated this standard. The maximal tumor size/burden was not exceeded.

2.15 Immunohistochemistry (IHC) analysis

Paraffin-embedded xenograft tumors were deparaffinized in xylene and hydrated in graded ethanol. Antigen was repaired by microwaving sections for 15 min in EDTA, then sections were treated with 3% hydrogen peroxide to quench endogenous peroxidase activity. Next, sections were stained with primary antibody against COL17A1 (ab184996) or ACTN4 (ab108198) overnight, followed by HRP-conjugated secondary antibodies (ab6728 or ab6721) for 1 h. After being stained with diaminobenzidine (Cat#D8001, Merck Millipore), slices were counterstained with hematoxylin (Cat#H3136, Merck Millipore), and then images were acquired under a microscope and the proportion of tumor cells were analyzed.

2.16 Statistical analysis

Each cell experiment was performed in three biological replicates. The data were manifested as mean ± standard deviation (SD), and normal distribution and similarity of variance was analyzed before comparison. Individual statistics of dependent samples were performed by paired t test, of unpaired samples by unpaired t test, and, if not normally distributed, by Mann–Whitney test. For multiple group comparisons, analysis of variance with Tukey post hoc test was used. P < 0.05 indicated statistically significant.