Cholest-4,6-Dien-3-One Promote Epithelial-To-Mesenchymal Transition (EMT) in Biliary Tree Stem/Progenitor Cell Cultures In Vitro

Human biliary tree stem/progenitor cells (hBTSCs), reside in peribiliary glands, are mainly stimulated by primary sclerosing cholangitis (PSC) and cholangiocarcinoma. In these pathologies, hBTSCs displayed epithelial-to-mesenchymal transition (EMT), senescence characteristics, and impaired differentiation. Here, we investigated the effects of cholest-4,6-dien-3-one, an oxysterol involved in cholangiopathies, on hBTSCs biology. hBTSCs were isolated from donor organs, cultured in self-renewal control conditions, differentiated in mature cholangiocytes by specifically tailored medium, or exposed for 10 days to concentration of cholest-4,6-dien-3-one (0.14 mM). Viability, proliferation, senescence, EMT genes expression, telomerase activity, interleukin 6 (IL6) secretion, differentiation capacity, and HDAC6 gene expression were analyzed. Although the effect of cholest-4,6-dien-3-one was not detected on hBTSCs viability, we found a significant increase in cell proliferation, senescence, and IL6 secretion. Interestingly, cholest-4.6-dien-3-one impaired differentiation in mature cholangiocytes and, simultaneously, induced the EMT markers, significantly reduced the telomerase activity, and induced HDAC6 gene expression. Moreover, cholest-4,6-dien-3-one enhanced bone morphogenic protein 4 (Bmp-4) and sonic hedgehog (Shh) pathways in hBTSCs. The same pathways activated by human recombinant proteins induced the expression of EMT markers in hBTSCs. In conclusion, we demonstrated that chronic exposition of cholest-4,6-dien-3-one induced cell proliferation, EMT markers, and senescence in hBTSC, and also impaired the differentiation in mature cholangiocytes.


Introduction
Primary sclerosing cholangitis (PSC) is a rare idiopathic, heterogeneous, cholestatic liver disease characterized by chronic inflammation of bile ducts, leading to duct fibrosis with associated strictures and liver cirrhosis, bearing an exceptionally high malignancy risk, primarily cholangiocarcinoma (CCA) [1][2][3][4][5]. The aetiopathogenesis of PSC is very complex and not yet well defined [4]. Moreover, PSC remains one of the most complex pathological conditions of clinical hepatology because still,

Cell Cultures
Approximately 5.2 × 10 4 cell/cm 2 (EpCAM + hBTSCs) were seeded in multiwell 6 (JET Biofil, Caton, China #TCP011006) with KM supplemented with fetal bovine serum (FBS) 10% for the first day after cell isolation. After 24 h, the medium was replaced with KM without FBS and it was changed every three days. When the cells obtained the optimal conditions considering morphology and colony formation, they were treated with KM (control condition) or KM added with cholest-4,6-dien-3-one (Gentaur, Milan, Italy #TRC-C431420-2.5) at 0.14 mM [1]. Cells were cultured for 10 days in these conditions in order to mimic the chronic exposure which occurs during human PSC. After 10 days, the cell media were collected and stored at −20 • C, cells were detached by using the trypsin (Gibco # 12604) and were stored at −20 • C in the trizol (Invitrogen, Carlsbad, California, USA #15596026) or ripa (Sigma, St. Louis, Missouri, USA #P5726) for gene or protein analysis, respectively. Furthermore, hBTSCs were cultured in hormonally defined medium specifically tailored to induce the differentiation of hBTSCs toward mature cholangiocytes (CM) [19], or CM supplemented with cholest-4,6-dien-3-one at 0.14 mM for 15 days. The medium was changed every three days and cholest-4,6-dien-3-one was added to CM at day 5 in order to perform a chronical exposure of 10 days.
Finally, in vitro signaling pathway effects on hBTSCs were evaluated by supplementing the KM with the activators and inhibitors, as follows: • Hedgehog pathway [25]: KM was used as basal condition. Pathways induction was performed for 15 days, and the medium was changed every three days. At the end of treatment, cells were detached using trypsin (Gibco, Waltham, Massachusetts, USA # 12604) and stored at −20 • C in trizol (Invitrogen, Carlsbad, California, USA #15596026).

Proliferation and Viability Analysis
hBTSCs were detached using trypsin and counted by trypan blue exclusion assay at several time points (1, 3, and 10 days) to count the cell number in cultures. At day 10, cell viability and proliferation index (population doubling (PD)) were measured by data obtained from trypan blue exclusion assay and the formula described in other studies [19,20]. Values are expressed as mean ± standard deviation. Briefly, equations used for cell viability and PD are reported below respectively: Cell viability = Cells viable (cells viable + cells dead) (1) PD = log 10 (N) − log 10 (Ns) log 10 (2) (2) where N is the harvested cell number, and Ns is the number of the seeded cells initially plated cell number.

Cell Senescence
Senescence after cholest-4,6-dien-3-one treatment was analyzed by X-Gal test (Sigma, St. Louis, Missouri, USA #CS0030), as described previously [19]. Briefly, cells were incubated with chromogenic X-Gal substrate for 4 h in incubator at 37 • C in the absence of CO 2 . Blue cells were counted by optical microscopy used 10 random fields. Values are expressed as mean ± standard deviation.

RNA Extraction and Quantitative Reverse-Transcription Polymerase Chain Reaction (RT-qPCR) Analysis
Cell RNA extraction was performed as described by Roibaudo et al. [27]. RT-qPCR was performed as described in previous papers [19], a detailed protocol is described in Supplementary Materials and Methods. The sequences used for the analysis are indicated in Supplementary Table S1.

Protein Extraction and Western Blot
Total protein extraction of cultured hBTSC, in basal condition (KM) and pathologic condition (KM + cholest-4,6-dien-3-one was performed using RIPA Buffer, phosphatase inhibitor cocktail (Sigma, St. Louis, Missouri, USA #P5726) at concentration 1:100 and protease inhibitor cocktail (Sigma, St. Louis, Missouri, USA #P8340) at concentration 1:100). Quantification was carried on through the protein assay dye reagent (Bio-Rad, Hercules, California, USA #500-0006) by spectrophotometry analysis at 595 nm, as described in the previous paper [1]. The total protein extract was subjected to 4%-20% SDS-PAGE (Mini-PROTEAN ® TGX™ Bio-Rad, Hercules, California, USA #4561094). The resolved proteins were transferred to a 0.2 µm pore-size nitrocellulose membrane (Bio-Rad, Hercules, California, USA #1620146). The membrane was blocked using no-fat milk at 5% in TBST buffer at +4 • C overnight. The following day, block buffer was washed three times using TBST and also anti-hTERT (GeneTex, Irvine, California, USA #GTX124242) was incubated at concentration 1:500 for +4 • C overnight. Finally, the membrane was washed by TBST to eliminate the primary antibody and the membrane was incubate with the secondary antibody, anti-rabbit HRP (Cell Signaling, Danvers, Massachusetts, USA #7074) at concentration 1:2000 for 1 h at room temperature. At the end of incubation, the secondary antibody was washed by TBST and the membrane was incubated with Clarity Western ECL Substrate (Bio-Rad, Hercules, California, USA #1705061) for 3 min before the development of membrane in the darkroom. The densitometric quantization of the Western blot bands was conducted through the ImageJ software. Gene bands were normalized to GAPDH (SAB, College Park, Maryland, USA #41549), as housekeeping control, and expressed as relative quantity of protein. Values are expressed as mean ± standard deviation.

ELISA Assay
Cell medium supernatant at day 10 of culture from two conditions (KM and KM + cholest-4,6-dien-3-one) were harvested and stored at −20 • C. Interleukin 6 (IL6) levels present in the media were measured by ELISA (Cambridge, UK #ab46042) according to the protocol provided by the company. Values are expressed as ng/mL. Values are expressed as mean ± standard deviation.

Telomerase Activity Quantification
Telomerase activity was measured by telomerase activity quantification qPCR assay kit (ScienCell, Carlsbad, CA, USA #8928). Approximately 4 × 10 6 cells for each replicate (N = 6) for both cell culture conditions were extracted and analyzed according to the product protocol. Briefly, cells were detached by trypsin, counted by trypan blue exclusion assay, and pelleted. Cell proteins were extracted by cell lysis buffer supplemented with PMSF 0.1 M in isopropanol and β-mercaptoethanol. After cell protein extraction telomerase reaction was performed as described in the product protocol. Finally, qPCR was done to analyze the telomer production by telomerase. ∆Cq was calculated by the following equation: ∆Cq is quantification cycle value obtained from qPCR. The relative telomerase activity was calculated using the formula described above: Relative telomerase activity = 2 −∆Cq (4)

Protein Pathway Analysis
Protein cell extracts were analyzed by RayBiotech Inc. (Peachtree Corners, GA, USA) for Sonic Hedgehog (Shh), bone morphogenic protein 4 (Bmp-4), angiopoietin 2 (Ang-2), Fas pathways for KM and KM supplemented with cholest-4,6-dien-3-one conditions. High-density GS array kits (RayBiotech, Inc., Norcross, GA, USA) allow semi-quantitative determination of the concentrations of human proteins in a single experiment, with detection sensitivity similar to that of a traditional ELISA. The arrays include a 16-well removable gasket that allows the processing of 16 samples on a single slide. The sensitivities of the antibodies in the arrays have been in the order of pg/mL. Values are expressed as mean ± standard deviation.

Statistical Analysis
Values are expressed as the mean ± standard deviation obtained from groups of six samples. statistical analyses were performed by SPSS statistical software (Version 18.0 SPSS Inc., Chicago, IL, USA). Two-tailed Student's t-tests were performed to assess the statistical significance of differences between groups. Statistical significance was set at p < 0.05.

Results
3.1. Viability, Proliferation and Senescence after Chronic Cholest-4,6-Dien-3-One Exposure in hBTSC Cultures 3.1.1. Cell Number in hBTSC Cultures hBTSCs were cultured in KM, basal condition, and KM supplemented with oxysterol (cholest-4,6-dien-3-one) for 10 days in order to mimic the PSC chronic injury. At every time point (1, 3, and 10 days) cells were detached and counted by trypan blue exclusion assay. Cells grew in PSC mimic condition for 10 days showed a significant increase of cell number in culture (1 416 000 ± 105 709.03; N = 6; p < 0.0001) compared to hBTSCs cultured in basal condition (621 000 ± 65 589.63; N = 6) ( Figure 1A). In the early time points (one and three days), no differences were observed between the two culture conditions. This result suggests that in the long period, cholest-4,6-dien-3-one could have a role in cellular proliferation.
3.1.2. Cell Viability hBTSCs were cultured as described previously. After 10 days of culture, cells were detached and counted both viable and dead cells by trypan blue exclusion assay. At day 10, cells grown in PSC mimic condition (93.98% ± 1.87%) and basal condition (95.04% ± 2.53%) did not show any significant difference in cell viability (N = 6; p > 0.05) ( Figure 1B). The result achieved could indicate that the cholest-4,6-dien-3-one does not influence cell viability.
tumor insurgency, inducing EMT pathway activation and a tumor escape strategy by reduction of Fas receptor.

Discussion
We have investigated the role of a specific oxysterol (cholest-4,6-dien-3-one), implicated with PSC and other cholangiopathies pathogenesis, in the modulation of the biological properties of hBTSCs and differentiated cholangiocytes in vitro. The results obtained in this study have led to

Discussion
We have investigated the role of a specific oxysterol (cholest-4,6-dien-3-one), implicated with PSC and other cholangiopathies pathogenesis, in the modulation of the biological properties of hBTSCs and differentiated cholangiocytes in vitro. The results obtained in this study have led to several conclusions: (1) Cholest-4,6-dien-3-one increased cell proliferation of hBTSCs without any effect on the cell viability, (2) at the same time cholest-4,6-dien-3-one induced the senescence in hBTSCs treated for 10 days, reducing telomerase expression of protein and mRNA and decreasing telomerase activity. PSC is one of chronic inflammatory cholangiopathies frequently complicated by CCA. Interestingly, significant proliferation of hBTSCs, expansion of PBGs, and dysplasia were observed in PSC [31]. Further results from our group indicated that all CCAs emerging in PSC patients were mucin-producing tumors characterized by PBG involvement and high expression of stem/progenitor cell markers. CCA cells were characterized by a higher expression of epithelial-to-mesenchymal transition (EMT) traits and the absence of primary cilia compared to bile ducts and PBG cells in controls and patients with PSC [1].
Although etiologic factors associated with PSC are largely unknown, biologically relevant endogenous (e.g., oxysterols) and exogenous (e.g., LPS from bacterial translocation) molecules present in the bile have been tested as stressors to mimic cholangiocyte damage in PSC [32,33]. In general, oxysterols were demonstrated to induce cholangiocyte apoptosis, to perpetuate inflammation in cholangiopathies, and to display mutagenic and carcinogenic properties [32,33].
As observed in patients affected by PSC in previous investigations [1], our results demonstrated that in vitro cholest-4,6-dien-3-one induced a proinflammatory behavior observed as high proliferation rate and, at the same time, hBTSCs underwent in senescence due to the cholest-4,6-dien-3-one exposure [32,33]. Moreover, cholest-4,6-dien-3-one was able to induce the expression of EMT markers like TWIST1, TWISTI2, SNAIL1, SNAIL2, ZEB1, ZEB2, and simultaneously enhanced the histone deacetylase 6 (HDAC6) expression, and impaired the cholangiocyte differentiation inhibiting the primary cilium expression [34]. HDAC6 is an important histone protein responsible for the deacetylation of lysine residues on the N-terminal part of the core histones. Deacetylation of histones is an important process for epigenetic repression and plays a pivotal role in several biological functions, such as transcriptional regulation, cell cycle progression, and developmental events. The overexpression of histone deacetylase 6 was recognized as a sign of migration phenotype and tumor invasion in hepatocellular carcinoma and CCA, while its inhibition proved evident effects on the restoration and expression of the cilium in cholangiocytes and in tumor decay [34]. As demonstrated in a previous paper [1], hBTSCs exposed to LPS or to oxysterols showed a partial loss of primary cilia, while, in human mucinous CCA cells we observed mostly the absence of primary cilia. Notably, hBTSCs exposed to LPS and to oxysterols and CCA cells showed higher expression of LC3 and p62, markers of activated autophagy pathway [1]. Moreover as demonstrated in previous papers [35][36][37], hBTSCs possess self-renewal properties and multi-potential differentiation capabilities. Furthermore, a subpopulation of these cells expresses pluripotent stem cell markers [36]. Notably, several stressors, mainly associated with inflammation, induced differentiation of pluripotent stem cells through the perturbation of the unsaturated metabolome which characterizes stem cells [38]. Here, the reduction of pluripotent stem cell markers and the induction of EMT genes replicate the effect of stressors observed elsewhere [38].
The EMT gene expression and cilium disarrangement are consistent with PBG hyperplasia to tumor development described in patients with PSC [1]. Deciliated hBTSCs could be more susceptible to neoplastic transformation, thus candidate cell of origin for pCCA and dCCA in patients [39]. In severe liver injury, hepatocyte proliferation is an impaired feature of human chronic liver disease [40][41][42][43][44][45]. Telomere shortening and replicative senescence of mature hepatocytes and the arrest of the hepatocyte cell cycle due to specific insults are the main biologic events occurring in the chronic liver disease condition [46][47][48]. Many studies have reported that senescence microenvironment could enhance tumor insurgence [49][50][51]. The telomerase complex comprises TERT (telomerase reverse transcriptase), TERC (telomerase RNA template), dsykerin, NOP10, reptin, pontin, and other factors that are yet to be identified. Telomere lengthening is the canonical function of telomerase. Embryonic stem cells (ESCs) express high levels of telomerase activity that are required to maintain telomere length [52]. Here, we showed that cholest-4,6-dien-3-one reduces telomerase expression and activity in hBTSCs treated for 10 days with oxysterols. Inflammation has been linked to exhaustion of self-renewal in stem cells. Previous works demonstrated that prolonged exposure to oxysterols triggered hBTSC proliferation, increased NF-κB pathway activation and senescence levels, at the same time, a high pIκB-α/pNF-κB expression has been observed in PBGs [1,32], suggesting the acquisition of a senescence-associated secretory phenotype (SASP) and the existence of a proinflammatory loop. The stimulation of hBTSCs with oxysterols mimicked the senescence-associated secretory phenotype observed in PSC patients [32].
Here we showed increased interleukin 6 (IL-6) production and secretion in hBTSCs in vitro. In previous studies, a strong link between IL-6 secretion and senescence was demonstrated [28][29][30]. One of the limitation of our study is the lack of a detailed analysis of the SASP profile of cells treated with cholest-4,6-dien-3-one and untreated ones.
Here, we demonstrated that a putative pro-inflammatory mediator affects telomerase activity which is associated with a senescence-associated secretory phenotype and to proinflammatory loop, leading to the expression of EMT markers, and tumor insurgence. Recently, IL-6 and programmed death-1-ligand 1 (PD-L1) antibody blockade combination therapy decreases pancreatic ductal adenocarcinoma (PDAC) tumor progression and increases the percentage of intratumoral effector T cells [53]. Our in vitro data are consistent with the induction of potentially tumor escape strategy in hBTSCs exposed chronically to oxysterols, since cells showed a decrease of Fas in this condition. Riccio et al. demonstrated previously an unexpected modulation of T-cells by hBTSCs through the Fas/Fas-ligand pathway [54,55].
The simultaneous expression of sonic hedgehog (Shh) and bone morphogenic protein 4 (Bmp-4) pathways was inducted by chronic exposure to cholest-4,6-dien-3-one in hBTSCs. Moreover, the induction of each discrete pathway by specific media supplemented with agonists induced in hBTSCs the expression of EMT markers. Interestingly, the induction of Shh or Bmp-4 pathway by agonists was less effective than cholest-4,6-dien-3-one stimulation, when both pathways are activated simultaneously, suggesting that they could play a synergic role in EMT induction in hBTSCs. Many studies have indicated that tumors altered these pathways [56][57][58]. In previous works, both PSC/CCA and PSC samples were characterized by an extremely high expression of glioma-associated oncogene 1 (Gli-1; the effector of the sonic hedgehog pathway) by PBG cells, and neoplastic PBGs without primary cilia maintained high expression of Gli-1, suggesting a noncanonical activation of the hedgehog pathway [59,60]. Based on these results, we investigated the activation of these pathways in relation with the induction of EMT genes in hBTSCs. In support of our results, many previous studies can be mentioned. In particular, Shh activation modulates EMT of proliferating cholangiocytes of ductular reaction in the course of injuries associated with biliary fibrosis [61]. Similarly, Bmp activation was demonstrated previously to characterize the EMT process in the non-transformed mammary epithelial cell line [62]. However, further analysis should be performed by a specific antagonist to confirm the role of these pathways in the observed effects of oxysterols in hBTSCs.
Moreover, hBTSCs culture in chronic exposure condition showed a decrease of Fas and improving angiopoietin 2 (ANG-2) levels. Fas reductions could suggest that hBTSCs are developing a tumor escape strategy as already analyzed in our previous work in human CCA primary cell lines [54]. This tumor escape strategy has been extensively studied in several tumors [63][64][65][66][67]. Furthermore, the ANG-2 is involved with VEGF to facilitate cell proliferation and migration of endothelial cells [68].
Recently, authors have suggested that ANG-2 could be involved in the regulation of endothelial integrity and inflammation [69].
The major limitation of our study is the lack of mouse models. In fact, to recapitulate the effect of oxysterols on cholangiocytes and hBTSCs, the administration by diet or injection in mouse models may not be appropriate since the concentration in the bile of such molecules is affected by several factors. What has been previously demonstrated by other authors in synthesis is that chronic liver diseases, metabolic alterations associated with obesity, modulate bile composition and are associated with increased levels of biliary oxysterols [70].
Taken together, our results supported the hypothesis that hBTSCs chronically exposed to cholest-4,6-dien-3-one were affected by an inflammatory microenvironment able to provide a permissive milieu for CCA insurgence. Furthermore, the collected data supported the role of hBTSCs in the pathogenesis of PSC and reproduced the changes previously observed in vivo in patients [1].

Conclusions
We demonstrated cell modification due to the chronical exposition of cholest-4,6-dien-3-one in hBTSC cultures regarding cell proliferation, senescence, EMT markers expression and Shh and Bmp-4 pathways activation. Moreover, we demonstrated that cholest-4,6-dien-3-one could impair hBTSC differentiation tailored in mature cholangiocytes. In the human biliary pathology, proliferation, loss of primary cilium, acquisition of EMT features, and senescence are key processes contributing to the transformation and molecule secretion triggering biliary inflammation. Our previous study demonstrated the hBTSCs contribute to PSC insurgency. Based on our data, we suggest that cholest-4,6-dien-3-one could modify the biliary homeostasis by acting on hBTSCs. However, the major limitations of this study is the lack of data obtained from experimental mouse models of cholestasis and from clinical investigations in patients affected by cholangiopaties.

Funding:
The study was supported by research project grant from Sapienza University of Rome #000324_17_H2_ALVARO_H2020 -ALVARO -PROGETTI H2020 2017.