Lipopolysaccharide (LPS)-Induced Biliary Epithelial Cell NRas Activation Requires Epidermal Growth Factor Receptor (EGFR)

Cholangiocytes (biliary epithelial cells) actively participate in microbe-induced proinflammatory responses in the liver and contribute to inflammatory and infectious cholangiopathies. We previously demonstrated that cholangiocyte TLR-dependent NRas activation contributes to proinflammatory/ proliferative responses. We test the hypothesis that LPS-induced activation of NRas requires the EGFR. SV40-transformed human cholangiocytes (H69 cells), or low passage normal human cholangiocytes (NHC), were treated with LPS in the presence or absence of EGFR or ADAM metallopeptidase domain 17 (TACE) inhibitors. Ras activation assays, quantitative RT-PCR, and proliferation assays were performed in cells cultured with or without inhibitors or an siRNA to Grb2. Immunofluorescence for phospho-EGFR was performed on LPS-treated mouse samples and specimens from patients with primary sclerosing cholangitis, primary biliary cirrhosis, hepatitis C, and normal livers. LPS-treatment induced an association between the TLR/MyD88 and EGFR/Grb2 signaling apparatus, NRas activation, and EGFR phosphorylation. NRas activation was sensitive to EGFR and TACE inhibitors and correlated with EGFR phosphorylation. The TACE inhibitor and Grb2 depletion prevented LPS-induced IL6 expression (p<0.05) and proliferation (p<0.01). Additionally, cholangiocytes from LPS-treated mouse livers and human primary sclerosing cholangitis (PSC) livers exhibited increased phospho-EGFR (p<0.01). Moreover, LPS-induced mouse cholangiocyte proliferation was inhibited by concurrent treatment with the EGFR inhibitor, Erlotinib. Our results suggest that EGFR is essential for LPS-induced, TLR4/MyD88-mediated NRas activation and induction of a robust proinflammatory cholangiocyte response. These findings have implications not only for revealing the signaling potential of TLRs, but also implicate EGFR as an integral component of cholangiocyte TLR-induced proinflammatory processes.


Introduction
Biliary epithelial cells (cholangiocytes) form a simple epithelial layer that separates the intrahepatic bile duct lumen from liver parenchyma [1]. Cholangiocytes perform the essential role of bile modification and transport of biliary and blood constituents, exist in an environment rich in potential mediators of cellular injury, and participate in portal tract repair processes. Cholangiocytes are periodically exposed to potentially pathogenic organisms or products derived from these microbes [2,3,4]. Indeed, the liver is a major organ for bacteria-derived lipopolysaccharide (LPS) clearance, and, while LPS undergoes metabolism in Kupffer cells and hepatocytes, it is excreted in bile in bioactive form [5,6]. Moreover, in cholestatic liver diseases, including the cholangiopathies primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC), cholangiocytes are exposed to increased levels of enteric microbe-derived LPS [6] through the portal venous circulation. How cholangiocytes respond to this cellular insult and the relevance of this response to hepatobiliary diseases remains to be fully explored.
Cholangiocytes express the epidermal growth factor receptor (EGFR), an upstream mediator of Ras activation [14]. EGFR has been implicated in growth and repair processes in a variety of epithelial cells [15,16], including growth of normal and neoplastic cholangiocytes [17,18]. EGFR activation typically occurs in response to EGFR ligands, including EGF, Transforming growth factor alpha (TGFα), Epiregulin (EREG), Amphiregulin (AREG), and heparin-binding EGF-like growth factor (HB-EGF). These ligands are expressed as transmembrane proteins that are cleaved from the membrane into the extracellular space where they function as autocrine or paracrine ligands for the EGFR. LPS has recently been shown to functionally transactivate cholangiocarcinoma cell EGFR through ADAM17 (TACE)-dependent release of TGFα [19]. Therefore, transactivation of the EGFR is an important mechanism by which LPS stimulates cholangiocyte inflammatory, proliferative and repair processes, which likely involve NRas/MAPK activation.
In the present manuscript we interrogate the role of TACE, EGFR, and the downstream molecular adaptor, Grb2, in LPS-induced NRas activation. Our cumulative data suggest that LPS induced NRas activation was sensitive to EGFR and TACE inhibition and correlated with EGFR phosphorylation, suggesting a progressive transactivation of the EGFR. Our data also suggest that ADAM17 (TACE) and the EGFR molecular adaptor protein, Grb2, are required for LPS-induced NRas activation as well as LPS-induced IL6 expression. We further demonstrate that LPS injected into the tail vein of mice promotes robust cholangiocyte EGFR phosphorylation. Finally, we demonstrate that phospho-EGFR is elevated in the chronic cholestatic liver disease, primary sclerosing cholangitis (PSC).

Ethics Statement
The study was conducted in accordance with the code of federal regulations and written informed consent from the donors was obtained. This study was approved by the Mayo Clinic Institutional Review Board (IRB# 10-005896; Pathobiology of Hepatic Epithelia) and Institutional Animal Care and Use Committee (A61313).

Cell Culture and LPS or Inhibitor Treatment
H69 cells, a well-characterized, SV40-transformed human cholangiocyte cell line [20] (a gift from Dr. D. Jefferson, Tufts University, Boston, MA), and normal human cholangiocytes (NHC), low passage human cholangiocytes isolated from normal liver [21] (a gift from Dr. Medina, University of Navarra, Pamplona, Spain), were used. When appropriate, cells were incubated in serum-free and EGF-free media or cultured in the presence or absence of inhibitors to EGFR (CAS 879127-07-8, Millipore), or Adam17/TACE (TAPI 1, Millipore) at concentrations of 10μM and 25 μM, respectively. For siRNA depletion experiments, cells were transfected with a siRNA to Grb2 (Ambion) or a scrambled control siRNA (Ambion) using Fugene HD transfection reagent (Promega) as per manufacturer's instructions. Cells were then treated with 200 ng/ml of LPS (Sigma) for the indicated times.

NRas Activation Assay
NRas activation was determined as reported previously [12]. Briefly, lysates were cleared, and diluted to equal protein concentration as confirmed by Bradford Analysis. Raf-RBD beads (Cytoskeleton) were added to each sample and incubated at 4°C, collected by centrifugation, washed, and resuspended in Laemmli Sample buffer (Bio-Rad). The samples were then run on a Tris-HCl gel, transferred to a nitrocellulose membrane, and immunoblotted with a primary antibody to NRas (Santa Cruz) and appropriate secondary antibody (LiCor). The membrane was visualized using LiCor's Odyssey infrared imaging system. Ponceau Red staining for total GST-RBD was used as loading control.

Immunoprecipitation and Western Blot
Immunoprecipitation was performed on total protein extracted from H69 cells cultured in the presence or absence of LPS. Lysates were pre-cleared with Protein A/G sepharose beads (Santa Cruz) at 4°C and incubated with a MyD88 (Santa Cruz) or Son of sevenless homolog 1 (SOS1; Santa Cruz) antibody overnight at 4°C. Protein A/G sepharose beads were added to the samples, centrifuged, washed in ice-cold PBS, and resuspended in Laemmli sample buffer. Western blots were performed for Grb2 (Santa Cruz) or TLR4 (Imgenex). Western blots were also performed using cell lysates from untreated, siRNA or inhibitor-treated cells that were cultured in the presence or absence of LPS. Samples were run on a 4-15% gradient Tris-HCL gel, transferred to a nitrocellulose membrane, and immunoblotted with primary antibodies for Grb2 (Santa Cruz), EGFR (Cell Signaling), or EGFR pY1068 (Cell Signaling) and appropriate secondary antibody (LiCor). Membranes were visualized on the LiCor Odyssey Infrared imaging system.

RNA Extraction and RT-PCR
RNA was extracted using Trizol (Invitrogen) following manufacturer's instructions. RNA was then reverse transcribed into cDNA using the Superscript III kit (Invitrogen), and used as template to perform quantitative PCR using primers specific for IL6.

Enzyme-linked immunosorbent assay
Culture medium was centrifuged to remove cellular debris, and 100 μl of cleared medium from H69 cells cultured in the presence or absence of LPS was used in the AREG ELISA following the manufacturer's instructions (R&D Systems). ELISAs were read using a 450-nm wavelength against a 630-nm reference wavelength. Concentration of AREG (pg/ml) was calculated using standard curve analyses.

In Vitro Cell Proliferation Quantitation
Cell proliferation was quantitated using a Nexcelom automated cell counter. H69 cells were plated at a starting density of 5,000 cells/well in 96-well plates. All cells were treated with 200 ng/ml of LPS for 0, 24, 48, or 72 hours in the presence or absence of inhibitors. After each respective time-point, cells were trypsinized and counted.

Liver tissues
Sixteen liver tissue specimens consisting of 4 PSC, 4 primary biliary cirrhosis (PBC), 4 chronic hepatitis C (HCV), and 4 normal livers from surgical resection or explant were utilized. Diseased specimens fulfilled clinical, serological, histological, and/or cholangiographic criteria for their respective diagnoses and had cirrhotic stage fibrosis. Liver specimens were fixed in 10% neutral buffered formalin, embedded in paraffin, and sectioned (4 μm). For the mouse studies, LPS (5 mg/kg body weight) or saline (control) was injected into the tail vein of C57-black mice. For the in vivo EGFR inhibition studies, Erlotinib (3 μM) was injected intraperitoneally one day prior to LPS treatment and daily over the course of 3 days. Mice were sacrificed at 1, 2, and 3 days post-LPS injection, and livers were initially perfused with a 0.9% sodium chloride solution, and perfused with a 4% paraformaldehyde solution to preserve liver samples for paraffin embedding, histological sectioning, and immunofluorescence.

Immunofluorescence and FRET
H69 cells were cultured in the presence or absence of LPS for the indicated time points, fixed in a 3% paraformaldehyde solution, washed in PBS, permeabilized with a 0.1% Triton X-100 solution, blocked in 5% serum, and incubated overnight at 4°C with primary polyclonal antibody to EGFR pY1068 (Sigma). Cells were rinsed in PBS, incubated with appropriate Alexa Fluor (Invitrogen) secondary antibody for 1h at room temperature, washed, and mounted in Pro-Long Gold with DAPI (Invitrogen). Tissue sections were deparaffinized, rehydrated in an ethanol gradient, and boiled in antigen unmasking solution (Vector Labs). Slides were incubated in a blocking solution containing 10% FBS, 1% BSA, and 0.1% Triton-X 100 for 1h at room temperature, incubated with primary antibody to EGFR pY1068 or proliferating cell nuclear antigen (PCNA, PC10, Santa Cruz Biotechnology), washed with a 0.1% Tween-20 solution in PBS, incubated with Alexa-Fluor secondary antibody, washed, and mounted with Prolong gold with DAPI. Confocal immunofluorescence microscopy was performed with a Zeiss LSM 510 as previously described [22]. Adobe Photoshop CS3 (Adobe Systems, San Jose, CA) was used to quantitate fluorescence intensity [23].
Acceptor photobleach Förster Resonant Energy Transfer (FRET) analysis was performed on H69 cells grown on glass coverslips transfected with a Myc-tagged Grb2 plasmid in the presence or absence of LPS for the indicated time-point. Cells were fixed in a 3% paraformaldehyde solution, rinsed in PBS, and blocked in a solution containing 5% serum, 5% glycerol, and 0.04% sodium azide for 45 minutes at room temperature. Cells were incubated with an anti-MyD88 antibody and anti-Myc-tag antibody at 4°C, washed and incubated with a Cy3 or Cy5-conjugated secondary antibody for 45 minutes at room temperature. Coverslips were rinsed in PBS and once in dH20. Coverslips were mounted on the slides using dH20. The acceptor fluorophore (Cy5) was photobleached in a region of interest and the intensity of Cy3 (donor) fluorescence was measured. The FRET efficiency for both conditions (estimate of the fraction of energy transfer to the acceptor per donor excitation event) was calculated using the Donor max /Donor min method [24].

Results
We performed NRas activation assays to demonstrate the time-course of LPS-induced NRas activation. LPS induced rapid (within 2-5 minutes) and persistent (60 minutes) NRas activation in both H69 and NHC (Fig 1A and 1B). Given that i) cholangiocytes express the ERBB family members EGFR and ERBB2, and ii) previous reports demonstrated LPS-induced EGFR phosphorylation in cholangiocarcinoma [17,18,25], we performed western blots for phospho-EGFR or phospho-ERBB2 from 0-60 minutes post-LPS treatment. EGFR phosphorylation was below the detection limit at 0-15 minutes; but evident at 30 and 60 minutes post-LPS treatment; ERBB2 phosphorylation was not detected at any of the time points (Fig 1C and 1D). We also performed confocal immunofluorescence microscopy for phospho-EGFR in H69 cells. At 5 minutes post-LPS treatment, a small proportion of cells exhibited phospho-EGFR ( Fig 1E). However, at 30 minutes post-LPS treatment, we observed a robust increase in phospho-EGFR. Quantitation of fluorescence intensity showed a significant increase in phospho-EGFR at 30 minutes post-LPS treatment compared to 0 hour (untreated) sample ( Fig 1F). These results suggest that LPS not only induces NRas activation, but EGFR phosphorylation as well.
We previously demonstrated that LPS-induced activation of NRas is TLR4-dependent, yet TRAF6-independent [12]. We therefore asked whether LPS-induced NRas activation required EGFR or TACE. NRas activation assays demonstrated that both the EGFR inhibitor and TACE inhibitor (TAPI-1) blocked LPS-induced NRas activation at the 30 minute time point (Fig 2A,  2B and 2C). We next asked whether LPS could induce the release of EGFR ligands to the cell culture media. While TGFα was not detected by ELISA in control or LPS treated cells (data not shown), LPS treatment induced, in a TACE sensitive manner, the release of the EGFR ligand, AREG (Fig 2D). TACE inhibition also suppressed LPS-induced IL6 mRNA expression ( Fig  2E). Moreover, both TACE inhibition and EGFR inhibition suppress LPS-induced cholangiocyte proliferation (Fig 2F). These results support that LPS induces NRas activation and IL6 expression through TACE-dependent EGFR activation.
Having demonstrated that inhibition of either EGFR or the known mediator of EGFR activation, TACE, could prevent LPS-induced NRas activation, we next assessed whether depletion of the downstream mediator of EGFR-dependent NRas activation, Grb2 was required for LPSinduced NRas activation. RNA-interference was used to selectively knock-down Grb2 and we again performed an NRas activation assay. LPS-induced NRas activation was blocked in cells depleted of Grb2 ( Fig 3A); moreover, depletion of Grb2 blocked LPS-induced IL6 expression (Fig 3B). Together, these results suggest that the EGFR-Grb2 complex is required for the LPSinduced activation of NRas and LPS-induced cholangiocyte expression of the proinflammatory mediator, IL6. Having demonstrated that depletion of the Grb2 molecular adaptor diminishes LPS-induced NRas activation and IL6 expression, we assessed whether Grb2 interacts with components of the TLR4 signaling apparatus. Initially, we immunoprecipitated the TLR adaptor protein, MyD88, from lysates of cholangiocytes cultured in the presence or absence of LPS; we then performed immunoblots for Grb2 and TLR4. The immunoprecipitations demonstrated that LPS induced an interaction between MyD88 and TLR4 as well as MyD88 and Grb2 ( Fig  3C). Furthermore, we demonstrated that LPS induced the interaction between Grb2 and the well-known Grb2-associated Ras guanine nucleotide exchange factor, son of sevenless homolog 1 (SOS1) (Fig 3C). As a complimentary approach to address protein complex formation, we force-expressed a Grb2-Myc construct in cholangiocytes and performed acceptor photobleach FRET analysis on cultured cholangiocytes in the presence or absence of LPS. Overexpressed Grb2-Myc and endogenous MyD88 were detected by confocal microscopy using primary antibodies against Myc-tag and Myd88 followed by secondary antibodies conjugated to Cy3 (donor) and Cy5 (acceptor), respectively. LPS treatment induced an increase in donor Cy3 fluorescence detection following Cy5 (acceptor) photobleaching, demonstrating the LPS-induced proximity of MyD88 and Grb2 (Fig 3D and 3E). The FRET efficiency for the LPS-treated cells demonstrated energy transfer between the fluorophores (~10%) (Fig 3F); further suggesting the induction of a TLR/MyD88: EGFR/Grb2 signaling platform.
We previously demonstrated that PSC cholangiocytes exhibit increased NRas activation compared to disease control and normal livers [26]. Having demonstrated here, in a cell culture system, that LPS-induced NRas activation requires the EGFR, we determined using confocal immunofluorescence microscopy, whether phospho-EGFR was present in cholangiocytes of livers with hepatobiliary disease. In normal liver, phospho-EGFR was restricted primarily to the apical plasma membrane in cholangiocytes and exhibited weak fluorescence. In contrast, the cholangiocytes from primary sclerosing cholangitis (PSC) livers exhibited elevated phospho-EGFR, with an approximately 5-fold increase vs. normal (NL), primary biliary cirrhosis (PBC) and Hepatitis-C virus (HCV) infected livers (Fig 4A and 4B). We further addressed whether LPS induced cholangiocyte EGFR phosphorylation and proliferation in a mouse model. LPS (5mg/kg body weight) was injected into the tail vein of C57-black mice and the livers were assessed for EGFR phosphorylation and proliferation by confocal immunofluorescence. Control saline injected animals exhibited minimal EGFR phosphorylation, while LPS treated animals exhibited robust EGFR phosphorylation (Fig 5A and 5B). Moreover, control saline injected animals exhibited minimal cholangiocyte proliferation, while proliferation was increased in LPS treated animals. The observed proliferation was blocked when the animals were treated concurrently with the EGFR inhibitor, Erlotinib (Fig 5C and 5D). Collectively, these findings: i) directly demonstrate that the EGFR is significantly more activated in PSC cholangiocytes compared to normal liver and other liver diseases; ii) indirectly support the concept that cholangiocytes may contribute to the biliary inflammation and fibrosis observed in PSC, and iii) demonstrate that LPS can induce EGFR phosphorylation and proliferation in an animal model of hepatic injury.

Discussion
We previously demonstrated that PAMPs initiate NRas activation in a TLR-dependent manner [12] and NRas activation contributed to the proinflammatory phenotype of LPS-treated cholangiocytes in vitro. In the present manuscript we interrogated the mechanisms of NRas activation. The main findings are that: i) LPS induced NRas activation requires EGFR and TACE; ii) TACE inhibition prevents LPS-induced cholangiocyte IL6 expression and proliferation; iii) a downstream mediator of EGFR activation, Grb2, is essential for LPS-induced NRas activation; iv) cholangiocytes from PSC livers exhibit increased EGFR phosphorylation; and v) mice treated with LPS exhibit increased cholangiocyte EGFR phosphorylation and proliferation. The findings reveal the signaling potential of TLRs, and implicate the EGFR as an integral component of cholangiocyte TLR-induced proinflammatory and reparative processes (Fig 6). Moreover, our data support that EGFR is phosphorylated in PSC, a progressive, incurable cholangiopathy characterized by biliary tract inflammation and fibrosis [27].
Not only is the biliary tract periodically colonized by microorganisms [2,3], microbiallyderived molecules are delivered directly to the liver through enterohepatic circulation. Hence, it is not surprising that the cells of the liver, including cholangiocytes, sense and respond to endogenous (e.g. bile acids, nucleotides) and exogenous (e.g. microbially-derived) molecules [6,28] which may be potentially injurious and contribute to chronic hepatobiliary diseases. We propose, and our data suggest, that pathogen recognition initiates a signaling cascade that promotes growth factor receptor activation, and this cascade is essential for the cholangiocyte proinflammatory and proliferative response to PAMPS. While cholangiocyte pathogen recognition-dependent NFkB activation induces secretion of innate immune responseassociated molecules [9,29], our current results extend the repertoire of TLR-dependent responses to include the activation of a strong mediator of cellular growth and repair, the EGFR.
While a healthy intestine prevents the translocation of microbes and microbially derived products, disruption of the intestinal barrier function results in the translocation of large amounts of these products. Intriguingly, PSC is frequently associated with inflammatory bowel disease (IBD). Indeed, IBD (most often ulcerative colitis) is found in approximately 75% of PSC patients [30,31]. The relationship between IBD and PSC remains obscure, yet forms the basis for the "leaky gut" hypothesis. This hypothesis suggests that the PSC-IBD association may be related to enterohepatic circulation of gut-derived molecules and possibly facilitated by increased intestinal permeability [32,33]. Several observations support the importance of enterohepatically circulated microbial molecules and the cholangiocyte response to microbial products in the etiopathogenesis of PSC. Indeed, freshly isolated cholangiocytes from PSC-diseased livers exhibit hypersensitivity to LPS and other PAMPs [34]. Moreover, PSC is associated with portal bacteremia, bacterial colonization of the bile ducts [35], detection of bacterial 16s ribosomal deoxyribonucleic acid in bile [36,37], and cholangiocytes in PSC livers exhibit an accumulation of LPS [38]. Accumulating evidence now suggests a potentially important role for the intestinal microbiota, and enterohepatic circulation of microbial derived molecules, as a putative mechanistic link between PSC and IBD and a central pathobiological driver of PSC.
Transactivation of the EGFR is a well described phenomenon and occurs through the activation of a variety of plasma membrane receptors including G-protein coupled receptors [39,40], tumor necrosis factor receptor [41], and more recently, pathogen recognition receptors [19,42]. Recently, it was demonstrated that LPS induced the transactivation of cholangiocyte EGFR through TACE-dependent mechanisms [19]. Moreover, it was demonstrated that LPS induced a delayed (~6 hour) EGFR phosphorylation event in cholangiocarcinoma cells which was implicated in LPS-induced cell invasion. Similarly, our data suggest that LPS induced NRas activation, likely through the transactivation of the EGFR in a process involving TACE. Moreover, we implicate EGFR and the molecular adaptor Grb2 in NRas activation.
Cholangiocytes are the target of a diverse group of biliary disorders of different etiologies, collectively referred to as the cholangiopathies [27,43]. The cholangiopathies have in common a number of cholangiocyte-associated processes that may contribute to their pathogenesis, including pro-inflammatory and proliferative responses. Importantly, we demonstrated that the EGFR is highly phosphorylated in cholangiocytes of PSC livers compared to normal, PBC, and HCV livers. PSC is a chronic, cholestatic, fibroinflammatory cholangiopathy with unknown etiology and no effective therapeutic approaches [44]. We recently demonstrated that LPS induced cholangiocyte senescence and the senescence-associated secretory phenotype, likely through NRas activation. Moreover, cholangiocytes in PSC livers exhibit increased NRas activation and a high proportion of senescent cells [26]. Whether the EGFR is involved in LPS-induced cholangiocyte senescence remains unknown but is an active area of investigation. Moreover, patients with PSC are at increased risk of developing cholangiocarcinoma (CCA), a cancer arising from cholangiocytes. Several studies have demonstrated increased expression and activating mutations of EGFR in human CCA [45,46,47,48]; however, this is the first report, to our knowledge, to address the activation state of the EGFR in PSC. Moreover, circulating and intrahepatic LPS are frequently elevated in chronic liver disease. Whether patients with PSC have a genetic predisposition to aberrant LPS-induced EGFR signaling cascades, and how the termination of this signaling cascade is regulated, remains to be determined.
In summary, we have extended our previous findings regarding LPS-induced NRas activation and defined a signaling axis involving the TLR complex and growth factor receptor signaling. This signaling axis is required for robust expression of proinflammatory mediators and using the Mann Whitney U test, revealed an increase in phospho-EGFR immunofluorescence in both PBC (p < 0.05) and PSC (p < 0.0001) compared to normal livers. Moreover, cholangiocytes from PSC diseased livers exhibited an increase in phospho-EGFR immunofluorescence compared to all other conditions (p < 0.0001). Data is presented as box plots (N = 4 for each condition; minimum of 5 ducts per sample) with mean (horizontal line between shaded boxes), 25 th and 75 th percentiles, with bars indicating maximum and minimum values. proliferation in response to microbially-derived insult. In line with previously reported work, we propose that the EGFR and NRas signaling axis are a central component of fibroinflammatory signaling. Further work is required to better define the molecular pathways and pathophysiological consequences of the TLR/EGFR/NRas signaling axis.