Epiplakin attenuates experimental mouse liver injury by chaperoning keratin reorganization

Background & Aims Epiplakin is a member of the plakin protein family and exclusively expressed in epithelial tissues where it binds to keratins. Epiplakin-deficient (Eppk1−/−) mice displayed no obvious spontaneous phenotype, but their keratinocytes showed a faster keratin network breakdown in response to stress. The role of epiplakin in the stressed liver remained to be elucidated. Methods Wild-type (WT) and Eppk1−/− mice were subjected to common bile duct ligation (CBDL) or fed with a 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC)-containing diet. The importance of epiplakin during keratin reorganization was assessed in primary hepatocytes. Results Our experiments revealed that epiplakin is expressed in hepatocytes and cholangiocytes, and binds to keratin 8 (K8) and K18 via multiple domains. In several liver stress models epiplakin and K8 genes displayed identical expression patterns and transgenic K8 overexpression resulted in elevated hepatic epiplakin levels. After CBDL and DDC treatment, Eppk1−/− mice developed a more pronounced liver injury and their livers contained larger amounts of hepatocellular keratin granules, indicating impaired disease-induced keratin network reorganization. In line with these findings, primary Eppk1−/− hepatocytes showed increased formation of keratin aggregates after treatment with the phosphatase inhibitor okadaic acid, a phenotype which was rescued by the chemical chaperone trimethylamine N-oxide (TMAO). Finally, transfection experiments revealed that Eppk1−/− primary hepatocytes were less able to tolerate forced K8 overexpression and that TMAO treatment rescued this phenotype. Conclusion Our data indicate that epiplakin plays a protective role during experimental liver injuries by chaperoning disease-induced keratin reorganization.


Supplementary Methods
Mouse injury models. All animals received humane care, were kept under standardized conditions and had free access to water and food. To study the effect of CBDL on WT and Eppk1 ─/─ mice, animals were anaesthetized with ketamine and xylazine. After midline laparotomy, the common bile duct was exposed and double ligated with 6-0 silk sutures. The abdomen was closed in layers, and the animals were allowed to recover on a heat pad. After 5 days, mice were anaesthetized with 150 µl isoflurane and killed by cervical dislocation. Successful bile duct ligation was evaluated by macroscopic evaluation and an increased, dilated gallbladder. To study the effects of DDC-feeding in WT and Eppk1 ─/─ mice, mice were fed with a diet supplemented with 0.1% DDC (#137030, Sigma-Aldrich, St Louis, MO) for 4 weeks.
Control animals were exposed to the corresponding DDC-free diet for the same period. After 4 weeks, animals were sacrificed for collection of blood and liver samples.
To study the effect of an additional liver stress model on epiplakin and keratin upregulation, WT mice were injected intraperitoneally twice a week for 12 weeks with 0.2 ml CCl4/kg mouse weight diluted in olive oil (both from Sigma) and sacrificed 96 hours after the last injection. Mice were killed using CO2 inhalation and livers were removed for RNA isolation.
Serum levels of liver enzymes were measured at Invitro laboratories (Invitro, Vienna, Austria). Livers were isolated and pre-defined lobes were either frozen in liquid nitrogen for biochemical analyses or cryo-sectioning, respectively, or routinely fixed in 4% paraformaldehyde for histology and immunostainings.  Scoring was performed using a scale from 0-4, reflecting none to massive bile duct proliferation (0, none; 1, rare; 2, moderate; 3, frequent; 4, massive). In a second approach, for quantification of the bile duct mass, liver paraffin sections derived from CBDL-and DDC-treated mice were immunofluorescently labeled with antibodies recognizing K19 and subjected to image acquisition. On each liver paraffin section, 4.5 mm 2 were scanned for the presence of K19-positive cells. Subsequently, the K19-positive areas were quantified using Icy software (Quantitative Image Analysis Unit, Institut Pasteur, France). For quantification of hepatocytes comprising keratin aggregates, liver paraffin sections derived from CBDL-and DDC-treated mice were labeled with a K8 antibody and subjected to IFM. On each liver paraffin section, 4.5 mm 2 were scanned for the presence of K8 aggregates. Keratin-enriched fractions were prepared using the high salt extraction method described previously [1]. Briefly, pulverized liver tissue was homogenized with phosphate-buffered saline (PBS) containing 5 mM EDTA and 1% Triton-X 100 followed by centrifugation at 16,000 g. The supernatant, representing the cytosolic fraction, was mixed with SDS-containing sample buffer and later subjected to Coomassie staining and immunoblotting. Protein lysates were separated by SDS-polyacrylamide gel electrophoresis (6−10%). Gels were either stained with Coomassie brilliant blue or proteins were transferred to nitrocellulose membranes.
The membranes were blocked with 5% milk powder in PBS with 0.1% Tween-20 and subsequently incubated with primary antibodies followed by incubation with HRPconjugated secondary antibodies. Detection was carried out using SuperSignal® West Pico Chemiluminescent Substrate (Thermo Scientific, Rockford, IL) and a Fusion FX chemiluminescence system (PEQLAB, Erlangen, Germany).
Quantification of protein bands was performed using QuantiScan version 1.5 software (Biosoft, Cambridge, UK).

Expression and purification of recombinant proteins. C-terminally GST-tagged
PRDs of murine epiplakin were bacterially expressed and affinity-purified as described before [2].

Overexpression of fluorescently tagged proteins in primary hepatocytes.
Primary hepatocytes were seeded on collagen-coated dishes as a mixed culture of Inc., Mountain View, CA) containing individual cDNAs coding for PRDs 1-9 and 16 [2]. To analyze potential colocalization of PRDs with keratin filaments, cells were fixed with methanol 24 hours after transfection and stained with an antibody recognizing K8. Fig. S1. Epiplakin transcript levels are higher in the biliary epithelium than in the liver. qRT-PCR analysis demonstrates significantly higher epiplakin mRNA levels in common bile duct and gall bladder compared to liver. Data are expressed as mean ± SEM; n = 6; *, p < 0.05; n.s., not significant.  Table S1). (