Leukemia inhibitory factor protects cholangiocarcinoma cells from drug-induced apoptosis via a PI3K/AKT-dependent Mcl-1 activation.

Cholangiocarcinoma is an aggressive, strongly chemoresistant liver malignancy. Leukemia inhibitory factor (LIF), an IL-6 family cytokine, promotes progression of various carcinomas. To investigate the role of LIF in cholangiocarcinoma, we evaluated the expression of LIF and its receptor (LIFR) in human samples. LIF secretion and LIFR expression were assessed in established and primary human cholangiocarcinoma cell lines. In cholangiocarcinoma cells, we tested LIF effects on proliferation, invasion, stem cell-like phenotype, chemotherapy-induced apoptosis (gemcitabine+cisplatin), expression levels of pro-apoptotic (Bax) and anti-apoptotic (Mcl-1) proteins, with/without PI3K inhibition, and of pSTAT3, pERK1/2, pAKT. LIF effect on chemotherapy-induced apoptosis was evaluated after LIFR silencing and Mcl-1 inactivation.Results show that LIF and LIFR expression were higher in neoplastic than in control cholangiocytes; LIF was also expressed by tumor stromal cells. LIF had no effects on cholangiocarcinoma cell proliferation, invasion, and stemness signatures, whilst it counteracted drug-induced apoptosis. Upon LIF stimulation, decreased apoptosis was associated with Mcl-1 and pAKT up-regulation and abolished by PI3K inhibition. LIFR silencing and Mcl-1 blockade restored drug-induced apoptosis.In conclusion, autocrine and paracrine LIF signaling promote chemoresistance in cholangiocarcinoma by up-regulating Mcl-1 via a novel STAT3- and MAPK-independent, PI3K/AKT-dependent pathway. Targeting LIF signaling may increase CCA responsiveness to chemotherapy.


Immunocytochemistry
After fixation with 4% paraformaldehyde (Carlo Erba), cells were incubated overnight at 4°C with primary antibodies against LIF (1:100), and LIFR (1:100). After washing with PBS-T, the cells were incubated for 30 mins at room temperature with the appropriate Alexa Fluor 488 secondary antibody (1:500, Life Technologies) and then mounted with Vectastain + DAPI.

ELISA for LIF quantification
Cells were seeded into a 24-well plate at 5 × 10 4 per well. After 24 h, supernatants were harvested, stored at -80°C and then the ELISA was developed according to the supplier (RayBiotech). A calibration curve was generated for each experiment.

Stem cell-like phenotyping
Briefly, untreated and rhLIF-treated (100 ng/mL for 24 hours) HuCCT-1 and TFK-1 cells were homogenized in 1 mL TRIzol ® Reagent (Life Technologies). Template complementary DNA was obtained by reverse transcription using 0.5 μg of total RNA, Superscript II reverse transcriptase (Life Technologies), random hexamers (50 pmol), and oligo-dT primers (100 pmol) (Promega). Relative transcript levels were quantified using Taqman gene expression probes for human Nanog and Oct4 (Life Technologies) and the real-time PCR was performed on an ABI 7500 thermocycler (Applied Biosystems). The relative expression of each gene was normalized against that of GAPDH.

Gene expression of LIF and LIFR
Using the same approach described for stem-cell phenotyping, we quantified the relative expression levels of LIF and LIFR in established and primary CCA cell

Cell invasiveness
Briefly, 5 × 10 4 cells were re-suspended in serumfree medium and seeded over a polyvinylpyrrolidonefree polycarbonate, 8 μm-pore membrane (Transwell, Costar) coated with 50 μg/mL Matrigel, within a Boyden microchamber. The lower chambers contained serum-free medium with/without rhLIF (10, 100 ng/ mL). After 48 hours, cells on the upper surface of the membrane were removed with a cotton swab whilst cells that adhered to the lower surface were fixed and stained using a Diff-Quick Staining Set (Medion Diagnostics); ten random fields of each membrane were photographed to count the number of clearly discernible nuclei.

Silencing of LIFR
Gene silencing of LIFR was performed using commercially available siRNAs against LIFR; scramble RNA served as a control (both Life Technologies). HuCCT-1 and TFK-1 cell lines were transfected using 20 pM of siRNA and Lipofectamine 2000 transfection reagent (Life Technologies) 24 h after plating. Transfection efficiency was assessed by WB and real-time PCR for LIFR. Figure 1: Extensiveness of LIF and LIFR expression in 'ductular-like' and 'mucin-producing' areas of CCA. LIF expression was more extensive in 'ductular-like' than in 'mucin-producing' CCA bile ducts A., whereas no significant differences in LIFR staining were observed between the two CCA phenotypes B. Representative micrographs illustrating LIF and LIFR staining in sequential sections of 'mucin-producing' (upper panels) and 'ductular-like' (lower panels) CCA areas are shown C. siRNA2, and siRNA3) to suppress LIFR expression was evaluated by both WB and real-time PCR in HuCCT-1 A. C. and TFK-1 B. D. cells. All siRNAs induced a pronounced reduction in protein (A, B) and transcript levels (C, D) of LIFR. Therefore, siRNA1 and siRNA2 were selected for the experiments on cell proliferation (shown in Supplementary Figure 4C, 4D) and drug-induced cytotoxicity (shown in Figure 3C, 3D).