Ameliorating liver disease in an autosomal recessive polycystic kidney disease mouse model

Abstract

Systemic and portal hypertension, liver fibrosis, and hepatomegaly are manifestations associated with autosomal recessive polycystic kidney disease (ARPKD), which is caused by malfunctions of fibrocystin/polyductin (FPC). The goal is to understand how liver pathology occurs and to devise therapeutic strategies to treat it. We injected 5-day-old Pkhd1^{del3-4/del3-4} mice for 1 mo with the cystic fibrosis transmembrane conductance regulator (CFTR) modulator VX-809 designed to rescue processing and trafficking of CFTR folding mutants. We used immunostaining and immunofluorescence techniques to evaluate liver pathology. We assessed protein expression via Western blotting. We detected abnormal biliary ducts consistent with ductal plate abnormalities, as well as a greatly increased proliferation of cholangiocytes in the Pkhd1^{del3-4/del3-4} mice. CFTR was present in the apical membrane of cholangiocytes and increased in the Pkhd1^{del3-4/del3-4} mice, consistent with a role for apically located CFTR in enlarged bile ducts. Interestingly, we also found CFTR in the primary cilium, in association with polycystin (PC2). Localization of CFTR and PC2 and overall length of the cilia were increased in the Pkhd1^{del3-4/del3-4} mice. In addition, several of the heat shock proteins; 27, 70, and 90 were upregulated, suggesting that global changes in protein processing and trafficking had occurred. We found that a deficit of FPC leads to bile duct abnormalities, enhanced cholangiocyte proliferation, and misregulation of heat shock proteins, which all returned toward wild type (WT) values following VX-809 treatment. These data suggest that CFTR correctors can be useful as therapeutics for ARPKD. Given that these drugs are already approved for use in humans, they can be fast-tracked for clinical use.

NEW & NOTEWORTHY ARPKD is a multiorgan genetic disorder resulting in newborn morbidity and mortality. There is a critical need for new therapies to treat this disease. We show that persistent cholangiocytes proliferation occurs in a mouse model of ARPKD along with mislocalized CFTR and misregulated heat shock proteins. We found that VX-809, a CFTR modulator, inhibits proliferation and limits bile duct malformation. The data provide a therapeutic pathway for strategies to treat ADPKD.

INTRODUCTION

Autosomal recessive polycystic kidney disease (ARPKD) is a multiorgan autosomal recessive genetic disorder that occurs in 1 in 20,000 live births in the United States, resulting in newborn morbidity and mortality (1, 2). The disease is associated with systemic and portal hypertension, fibrosis of both the liver and kidney, hepatosplenomegaly (1), and enlarged kidneys, with fusiform dilation of the collecting ducts that progresses to end-stage renal disease (ESRD) (3). ARPKD is caused by a variety of mutations in the PKHD1 gene encoding the protein fibrocystin/polyductin (FPC). Although ARPKD is referred to as a kidney disease, patients with this disorder also have liver disease (Caroli syndrome) that is characterized by congenital fibrosis, portal hypertension, and dilation of the biliary ducts (4). Normally, differentiation of hepatic cysts leads to a cessation of proliferation, but in ARPKD, proliferation persists, leading to the expansion of the hepatic ducts into the cysts (5, 6). In ARPKD, the portal ducts contain collagenous tissue, a condition referred to as congenital hepatic fibrosis. Portal hypertension in ARPKD is thought to occur via compression of the portal vein secondary to the liver fibrosis, or alternately via abnormal development of the portal system (7). Thus, the liver pathology in patients with ARPKD derives from two pathways, abnormal ductal plate development and liver fibrosis. Whether these two are separate pathologies or linked (such that the ductal plate abnormality leads to fibrosis) is unknown.

The cystic fibrosis transmembrane conductance regulator (CFTR) ion channel is well known to play a role in driving chloride across the apical membrane of cysts in autosomal dominant polycystic kidney disease (ADPKD) stimulated by high levels of cAMP (8, 9). However, the role of CFTR in ARPKD is less clear. For example, Nakanishi et al. (10) crossed a bpk mouse (an ARPKD model) with a CFTR-knockout mouse to create a ^{bpk−/−;cftr−/−} double homozygous mouse. The results were surprising because the double mutants developed massively enlarged kidneys and died from renal failure at ∼24 days after birth, a more severe phenotype compared with the bpk mouse alone (11). On the other hand, the result is more consistent with the fundamental differences in the pathophysiology of ARPKD compared with ADPKD (autosomal dominant PKD) with the former having ductal plate abnormalities caused by hyperproliferation and fusiform kidney cysts that may not have an obligate dependence on CFTR-driven fluid secretion.

Liver disease is an associated symptom of cystic fibrosis (CF) that is present in about one-third of patients with CF and is the third leading factor in patient mortality (12, 13). The clinical symptoms include steatosis, biliary cirrhosis, and progressive periportal fibrosis that lead to portal hypertension and liver failure (12). This damage in CF resembles to some extent the damage that occurs in ARPKD, except that patients with CF are not known to have ductal plate abnormalities (14).

Cholangiocytes function in the intrahepatic bile duct to secrete the fluid that comprises ∼40% of the volume of bile (15). Early studies have identified CFTR expression in cholangiocytes but not in hepatocytes (16). However, through the use of more sensitive techniques and more in-depth studies, CFTR has subsequently been identified in hepatocytes, where it participates in regulating endoplasmic reticulum (ER) stress (17).

Although one line of thought holds that cholangitis in patients with CF is caused by biliary obstruction resulting from a deficit of fluid in the hepatic ducts, there is evidence for increased collagen deposition and cell proliferation and an abnormal cholangiocyte morphology that go beyond a simple defect in CFTR-dependent fluid transport (18).

CFTR modulators are defined as potentiators that increase CFTR activity and correctors that rescue processing and trafficking of CFTR (19). In CF-associated liver disease, CFTR modulators improve liver function. For example, treatment of patients with the combination of lumacaftor (VX-809) and ivacaftor (VX-770) reduces hepatic steatosis, another common feature of CF liver disease (20). In another study of patients who underwent ivacaftor/lumacaftor treatment, serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), and γ-glutamyl transferase (GGT) decreased significantly (21). An additional study evaluated the effect of ivacaftor (VX-770) on enterohepatic circulation of bile acids, as assessed by biomarkers of bile acid homeostasis, such as fibroblast growth factor 19 (FGF19), which was lower, and 7α-hydroxy-4-cholesten-3 (C4), which was higher in patients bearing the G551D CFTR mutation (22) than in normal individuals. In the patients, treatment with ivacaftor increased the levels of FGF19 and lowered the C4 levels in the direction of normal values. Both studies suggest an improvement in liver function following modulator therapy. Interestingly, in these two studies, the changes in liver function tests did not directly correlate with changes in the function of other organs, as assessed by changes in sweat Cl⁻ or pulmonary function, suggesting that the response of the liver to modulator treatment may be unique when compared with that of other organs associated with CF disease. One concern with modulator therapy is drug-induced liver toxicity, which does occur, but is considered rare (23).

In view of the beneficial effect of modulator therapy on liver function in CF, the overall goal of our study here is to understand the role of CFTR in ARPKD and uncover therapeutic targets for pharmacological strategies to treat ARPKD.

METHODS

Mouse Strains and Treatment

All animal use complied with the regulations of the Johns Hopkins Animal Care and Use Committee. We conducted experiments in Pkhd1^{del3-4/del3-4} deletion model mice, obtained from the Baltimore PKD Center (24). Five-day-old Pkhd1^{del3-4/del3-4} mice of both sexes were divided into an untreated and a treatment group. Treated animals received, via intraperitoneal injection, VX-809 (30 mg/kg body wt) once every 48 h (i.e., every other day) for 30 days dissolved in dimethyl sulfoxide (DMSO). All wild-type (WT), untreated, and treated Pkhd1^{del3-4/del3-4} animals were injected with equal amounts of DMSO (40 µL). Age-matched and no cyst-induced Pkd1fl/fl;Pax8rtTA;TetO-cre mice, provided by the Baltimore PKD Center, were used as WT animals. In the overall study, there we six WT; eight pkhd1^{Pkhd1del3-4/del3-4} and eight pkhd1^{Pkhd1del3-4/del3-4} animals treated with vehicle or with VX-809 using equal numbers of males and females. It should be mentioned that although this was the starting numbers of animals, not all the assays performed were successfully performed on each animal.

Reagents

VX-809 (No. S1565) was purchased from Selleck Chemicals, Huston, TX; heat shock protein (HSP)27 (SC13132), HSP70 (SC66048), polycystin 2 (PC2) (SC28331) antibodies were purchased from Santa Cruz biotechnology; HSP90 (ADI-SPA-830F) antibody was purchased from Enzo Life Sciences; anti-E-cadherin was purchased from R&D Systems; No. AF748; Na⁺-K⁺-ATPase (No. ab76020) antibody was purchased from Abcam, NY. CFTR antibodies were obtained from the Cystic Fibrosis Antibody distribution program, at the University of North Carolina under the direction of Dr. Martina Gentzsch. FPC antibody was obtained from the Baltimore PKD Center

Antibody Validation

All antibodies were validated in several ways. They were thoroughly checked to verify that the band on the Western blot corresponded with the molecular weight reported in Western blots published in the literature or from the manufacturer’s website and with known posttranslation modifications. For CFTR and other glycosylated proteins, enzymes were used to remove the posttranslational moiety to validate that the pattern of glycosylation was consistent with known properties of the protein (25). For the heat shock proteins, additional siRNA experiments validated a reduction in the band on the Western blot consistent with their molecular weight (26, 27). Lack of staining in knockout animals verified the specificity of CFTR (28) and FPC (see Fig. 6) antibodies for their respective proteins.

Figure 6.

Expression of fibrocystin/polyductin (FPC) and heat shock proteins (HSPs) in Pkhd1^{del3-4/del3-4} mice. A: representative images of Western blots of FPC and HSP27, HSP70, and HSP90 expression in wild-type (WT) and Pkhd1^{del3-4/del3-4} mice. B–E: FPC and HSP27, HSP70, and HSP90 expression. Please note that the Western blots were performed on livers that contain more hepatocytes compared with cholangiocytes (15). Columns represent means ± SE. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. ANOVA with multiple comparisons. Dots, number of animals.

Biochemical Determinations

Blood was obtained by cardiac puncture from mice under 3% to 5% isoflurane-oxygen anesthesia. Blood urea nitrogen (BUN), creatinine, aspartate aminotransferase (AST), alanine transaminase (ALT), and albumin (ALB) analyses were performed by spectrophotometric measurement using a Vet Ace clinical chemistry automated random-access analyzer (Alfa Wasserman Diagnostic Technologies).

Immunoblotting

Tissue lysates were prepared from flash-frozen kidney samples that had been stored in liquid nitrogen and have been published previously (29). Samples were centrifuged, the supernatants were mixed with 2× Laemmli sample buffer (Bio-Rad No.1610737), and protein samples were run on 4% to 15% SDS-PAGE gels (Bio-Rad No. 4561084) before transfer to a polyvinylidene fluoride membrane (Bio-Rad, No. 1704157). Membranes were incubated with primary antibody (1:1,000) diluted in blocking buffer for overnight at 4°C. A horseradish peroxidase-conjugated secondary antibody was incubated with the membrane for 1 h, and then Super Signal West Dura extended-duration ECL Prime (Thermo Fisher No. 34075) was used for detection with an Amersham Imager 600 (Cytiva). Band intensities were analyzed using IQTL 8.2 software. See Ref. 26 for further details.

Immunostaining

Livers were obtained from Pkhd1^{del3-4/del3-4} and WT mice under isoflurane-oxygen anesthesia and fixed in 4% paraformaldehyde for 24 h, then paraffin-embedded and sectioned for immunofluorescent staining and confocal laser scanning microscopy as done previously (30). Images of whole livers were visualized using a Nikon fluorescence dissecting microscope equipped with an Olympus OM-D E-M5 Mark II digital camera (Nikon, Tokyo, Japan). For confocal microscopy, images were captured using a Zeiss LSM 880 confocal microscope (see Refs. 28 and 30 for further details).

Statistics

Data were analyzed using Student’s t test or analysis of variance followed by multiple comparisons where appropriate and as mentioned in the figure legends. Tests of significance were performed using the GraphPad Prism Software. All data are represented as means ± SE.

Reproducibility

Whenever possible, different laboratory members replicated the studies to ensure reproducibility. Laboratory members evaluating the effects of CFTR modulators were not blinded to treatment groups but were unaware of the outcome of the data before replication by another laboratory member. In this study, both male and female mice were used. Experimental Pkhd1^{del3-4/del3-4} animals were age-matched to the WT nondisease mice.

RESULTS

Defective Hepatic Bile Ducts in ARPKD

We injected 5-day-old Pkhd1^{del3-4/del3-4 mice} (hereafter referred to as Del3-4 mice) with VX-809 (30 mg/kg) for 1 mo. Hematoxylin & eosin staining (Fig. 1, A and B) showed that after 35 days, untreated mice already had defective morphology when compared with those of WT mice. Importantly, cytokeratin 19 (CK19), a marker of cholangiocytes (15), confirmed that the liver abnormalities consisted of enlarged and malformed hepatic ducts. Importantly, VX-809 caused a significant reduction in the size of the ducts, such that they resembled those of the WT mice. We chose to study young mice before advanced disease pathology had occurred. This is shown in Fig. 1D, which indicates that measures often used to evaluate liver function (e.g., ALB, AST ALT, and BUN), or kidney function (e.g., creatinine) (see Ref. 31) were all within the normal range in mice either treated or untreated with VX-809, measured 35 days after birth. Previous studies in these mice showed that liver disease consisting of ductal plate abnormalities and hyperproliferation begins earlier than cystic kidney disease, with ∼60% of the mice displaying a liver phenotype by 3 mo of age, as compared with only 20% with a severe cystic kidney phenotype (24). In our case, after 1 mo, the abnormal liver phenotype as evidenced by abnormal bile ducts is prominent, although still mild.

Figure 1.

Abnormal bile ducts after treatment with VX-809. Autosomal recessive polycystic kidney disease (ARPKD) mice were injected every other day with VX-809 (on postnatal [PN] days 5–35, 30 mg/kg). A: images of hematoxylin and eosin (H&E)-stained livers from wild-type (WT); Pkhd1^{del3-4/del3-4} (referred to as Del3-4 in all figures) mice that were either untreated or VX809-treated for one month, beginning at 5 days of age. Black arrowheads indicate the bile ducts, and the green arrowheads indicate the portal veins (PV). Images were taken at ×20. The scale bar is 100 µm. B: summary of the % bile duct area. The area represented by bile duct was measured using imageJ software (1.53n). Initially, the duct vein combined area was measured, and then the veins were subtracted from the combined area. Each dot in a column represents at least 50 µm of area measured. Two to three random areas in each image were measured. Data from six animals in each group. C: CK19 expression in confocal images from WT mice, del3-4 mice, or del3-4 mice treated with VX809. CK19 antibody = (Invitrogen No. PA5-29548) (see Ref. 26 for details). Scale bar is 10 µm. Data were analyzed by ANOVA with multiple comparisons using GraphPad Prism software. ***P < 0.001, ****P < 0.0001. Dots represent the number of measurements. D: liver and kidney function tests. Values measured at the end of the treatment period were: albumin (ALB), aspartate aminotransferase (AST), alanine transaminase (ALT), blood urea nitrogen (BUN), and creatinine. Analyses were performed by spectrophotometric measurement using a Vet Ace clinical chemistry automated random-access analyzer (Alfa Wasserman Diagnostic Technologies). Values are means ± SE. Statistical analysis was performed using a paired t test; ns, not significant. Please note that normal values are listed above the bar graphs. Dots represent the number of animals used in each experimental group.

Proliferation is Increased in ARPKD

Persistent proliferation of cholangiocytes is typical of ARPKD (5, 6). Therefore, we assessed the degree of proliferation by using Ki67, a widely used marker of cell proliferation (32). As shown in Fig. 2, A and B, proliferation of cholangiocytes (CK19 positive cells) was almost absent from WT cholangiocytes. In contrast, there was a multifold increase in the proliferation in CK19-positive cholangiocytes in the del3-4 mouse livers. Importantly, VX-809 treatment reduced the proliferation by ∼70%.

Figure 2.

Ki67 expression (A and B) and cystic fibrosis transmembrane conductance regulator (CFTR) and CK19 colocalization (C and D). A: confocal images of liver paraffin sections from wild-type (WT) or del 3-4 (Pkhd1^{del3-4/del3-4}) mice, stained for Ki67 (red), CK19 (green), and the merged image. B: percentage of cells positive for Ki67. The number of cells analyzed: for WT, n = 196; Pkhd1^{del3-4/del3-4}, n = 324; Pkhd1^{del3-4/del3-4} +VX809, n = 344. Samples were incubated with primary antibody Ki67 (Abcam No. ab15580) and CK19 (Invitrogen No. MA5-15884) 5 µg/mL in blocking buffer for 16 h at 4°C. Goat anti-rabbit Alexa Fluor 594- and anti-mouse Alexa Flour 488-conjugated secondary antibodies were used at a 1/250 dilution for 1 h to label the proteins (26). C: confocal images of liver from WT or del3-4 mice, stained for CFTR (red), CK19 (green), and the merged image. Yellow denotes colocalization. D: Pearson’s correlation coefficient R values. Samples were incubated with primary antibody CFTR (596) and CK19 (Invitrogen No. PA5-29548) 5 µg/mL in blocking buffer for 16 h at 4°C. Goat anti-rabbit Alexa Fluor 594- and anti-mouse Alexa Flour 488-conjugated secondary antibodies were used at a 1/250 dilution for 1 h to label the proteins. Scale bar is 10 µm. White dashed boxes in the merge panel show the enlarged portion of the corresponding image. One-way ANOVA with multiple comparisons. ****P < 0.0001. Dots, number of measurements from six animals in each group.

Expression and Location of CFTR in Cholangiocytes

Figure 2, C and D, shows the location of CFTR in cholangiocytes identified by the presence of CK19. This result is consistent with the work of others reporting that CFTR is present in cholangiocytes (15, 26). We did observe an increase in the colocalization of CFTR with CK19 in del3-4 cholangiocytes and a reduction toward WT values following VX-809 treatment. Cytokeratins are components of the intermediate filaments within the cytoplasm (33), and CK19 is associated with cholangiocytes (34). Since CK19 is an intracellularly located filament, increased colocalization with CFTR might indicate an increased association of CK19 with CFTR in ARPKD cells in the intracellular compartment or an abnormal distribution of CK19. With respect to abnormal distribution, abnormal cytokeratin profiles have been identified in diseases associated with ductal plate abnormalities (34).

Next, we conducted colocalization studies of the apically oriented marker E-cadherin (35) and the basolateral marker, Na⁺-K⁺-ATPase (30) (Fig. 3). CFTR colocalized with E-cadherin in WT cells, consistent with its apical location. In del3-4 cells, the colocalization increased by approximately twofold, suggesting a role for CFTR in cysts formed from cholangiocytes. VX-809 lowered the colocalization toward WT levels. There was some colocalization of CFTR with the basolateral membrane in WT cholangiocytes, and the degree of this colocalization was reduced in del3-4 cholangiocytes when compared with WT cells and restored toward normal in cholangiocytes that had been treated with VX-809. Interestingly, CFTR had been previously identified in both the apical and basolateral membrane of cholangiocytes (36).

Figure 3.

Colocalization of cystic fibrosis transmembrane conductance regulator (CFTR) with E-cadherin (A and B) and CFTR with Na⁺-K⁺-ATPase (C and D). A: confocal images of liver sections from wild-type (WT) or del3-4 (Pkhd1^{del3-4/del3-4}) mice, stained for CFTR (red), E-cadherin (green), and the merged image. Yellow denotes colocalization. B: Pearson’s correlation coefficient R values. Samples were incubated with primary antibody CFTR (596) and E-cadherin (R&D No. AF748) 5 µg/mL in blocking buffer for 16 h at 4°C. Goat anti-mouse Alexa Fluor 594- and anti-goat Alexa Flour 488-conjugated secondary antibodies were used at a 1/250 dilution for 1 h to label the proteins. Scale bar is 5 µm. C: confocal images of liver sections from WT or del3-4 mice, stained for CFTR (red), Na⁺-K⁺-ATPase (green), and the merged image. Yellow denotes colocalization. D: Pearson’s correlation coefficient R values. Samples were incubated with primary antibody CFTR (596) and Na-K-ATPase (Abcam No. ab76020) 5 µg/mL in blocking buffer for 16 h at 4°C. Goat anti-rabbit Alexa Fluor 594- and anti-mouse Alexa Flour 488-conjugated secondary antibodies were used at a 1/250 dilution for 1 h to label the proteins. Scale bar is 10 µm. One-way ANOVA with multiple comparisons. *P < 0.05, **P < 0.01, ****P < 0.0001. Dots, number of measurements from six animals in each group.

Both CFTR and PC2 Are Located within the Cilia

FPC and PC2 are well known to be located in the primary cilium (37), but CFTR has not been looked for in that location. Thus, to evaluate whether both proteins are found in the cilia, we costained with an antibody against acetylated α-tubulin, a known marker of the primary cilium (38). Figure 4 shows that both CFTR and PC2 were present in the primary cilium, with significantly higher levels found in del3-4 cholangiocytes that had been restored by VX-809. We determined whether the cilia length was altered and found (Fig. 5) a small but significant increase in primary cilia length in the del3-4 cholangiocytes, which was again restored in the direction of normal length by treatment with VX-809.

Figure 4.

Cystic fibrosis transmembrane conductance regulator (CFTR) localization to the primary cilium (A and B) and polycystin 2 (PC2) localization (C and D). A: confocal images of liver sections from wild-type (WT) and del 3-4 (Pkhd1^{del3-4/del3-4}) mice, stained for CFTR (red), acetylated α-tubulin (green), and the merged image. Yellow denotes colocalization. B: Pearson’s correlation coefficient R values. Samples were incubated with primary antibody CFTR and ac.α-tubulin, 5 µg/mL in blocking buffer for 16 h at 4°C. Goat anti-rabbit Alexa Fluor 594 and anti-mouse Alexa Fluor 488 secondary antibodies were used at a 1/250 dilution for 1 h to label the proteins. The box outlined by white dotted lines indicates the enlarged part of a representative image. Scale bar is 10 µm. C: confocal images of livers from wild-type (WT) or Pkhd1^{del3-4/del3-4} mice, stained for PC2 (red), acetylated α-tubulin (green), and the merged image. Yellow denotes colocalization. D: Pearson’s correlation coefficient R values. Samples were incubated with primary antibody PC2 and ac.α-Tubulin, 5 µg/mL in blocking buffer for 16 h at 4°C. Goat anti-rabbit Alexa Fluor 594 and anti-mouse Alexa Fluor 488 secondary antibodies were used at a 1/250 dilution for 1 h to label the proteins. Scale bar is 10 µm. One-way ANOVA with multiple comparisons. **P < 0.01; ***P < 0.001; ****P < 0.0001. Dots, number of measurements from six animals in each group.

Figure 5.

Length of primary cilia in bile duct epithelial cells. A: confocal images of livers from wild-type (WT) or del3-4 (Pkhd1^{del3-4/del3-4}) mice, stained with acetylated α-tubulin (green). B: summary of ciliary length. Cilia measurements were made from at least five individual images taken from three animals from each group; 8 to 10 cilia were randomly selected for each image. Graph shows individual values for cilia length from each group. One-way ANOVA with multiple comparisons. Dots represent individual measurements. Scale bar is 10 µm. **P < 0.01, ***P < 0.001.

Protein Levels Are Altered in ARPKD

Figure 6 shows that, consistent with del3-4 being a knockout of FPC (24), no protein was detected in the livers isolated from those mice. It is well known that the expression of key cellular proteins is altered in ADPKD, particularly the heat shock proteins (HSPs) (28, 39) thus we decided to evaluate the levels of selected heat shock proteins in del3-4 mouse livers. Figure 6 also shows that HSP27, HSP70, and HSP90 were all profoundly upregulated in the FPC knockout animals. The changes noted here are unaffected by VX-809. Changes in the protein levels of these particular HSPs have been associated with altered protein processing and trafficking, particularly with regard to CFTR (40). Indeed, we noted previously in cholangiocytes isolated from FPC-null mice that HSP70 and 90 are increased similar to what we observed here but Hsp27 is decreased. In the isolated cholangiocytes, VX-809 is effective in restoring the HSPs toward WT. We mention that the proteins measured here were from cell lysates of liver that contain both hepatocytes and cholangiocytes (15) compared with our previous experiments, which involved only cholangiocytes (26).

DISCUSSION

Ductal Plate Abnormalities

A common feature of liver disease in ARPKD is ductal plate malformation leading to malformed, dilated bile ducts (7). As mentioned earlier, normally, differentiation of hepatic cysts leads to a cessation of proliferation, but in ARPKD, proliferation persists, leading to the expansion of the hepatic ducts into the cysts (5, 6). We have demonstrated here in a mouse model of ARPKD that FPC malfunction in cholangiocytes results in enlarged cyst-like bile ducts, which are associated with large increases in cell proliferation when compared with WT ducts. Despite the occurrence of these enlarged cyst-like bile ducts, we did not observe significant changes in measures of either liver or kidney function in the mice. We attribute this lack of change in laboratory parameters to our use of young mice for our experiments. The results do suggest, however, that early treatment of the abnormally proliferating cholangiocytes is warranted before major destruction of the liver, which has been shown to occur in older mice that have lacked functional FPC for longer periods of time (24).

Proliferation

We found that proliferation of the bile duct is enhanced in ARPKD (26). One question is whether CFTR could be involved in this process. Patients with CF are at increased risk of cancers of the digestive tract (41). Although the development of intestinal cancer may have been caused by chronic intestinal inflammation, Than et al. (42) presented evidence that CFTR acts as a tumor suppressor. The question is whether CFTR’s role in cancer is causally connected to enhanced proliferation. Along these lines, Gallagher and Gottlieb (43) have shown an increased cell migration of intestinal cells from the crypts to the villi in CFTR-null mice as a result of the enhanced cell proliferation. Strubberg et al. (44) have taken this further, showing that the loss of CFTR stabilizes the Wnt transducer Dvl, facilitating Wnt/β-catenin signaling. Polycystin-1 (PC1), whose mutations are associated with ADPKD (45), is known to (26) associate with β-catenin and inhibit Wnt signaling via its C-terminal tail (46), perhaps leading to upregulated Wnt signaling in ADPKD. Interestingly, a transgenic mouse engineered to contain an activated form of β-catenin has been shown to display enhanced proliferation and polycystic lesions in renal tubules (47), a key experiment cementing its role in proliferation. We have shown previously that CFTR protein levels in whole cell lysates as measured by Western blotting in cholangiocytes derived from an FPC-null mouse are dramatically lower than those in WT cholangiocytes (26). Thus, it is possible that low levels of CFTR protein in ARPKD cholangiocytes contribute to a proliferative environment.

CFTR in the Liver

We have found that in WT cholangiocytes, CFTR is located principally toward the apical membrane. These findings in WT cholangiocytes are in contrast to what we observed in the kidneys in ADPKD mouse models, in which CFTR had a more basolateral location (30). However, a major function of the nephron is to absorb fluid (48), and thus a basolateral location for CFTR is appropriate. Cholangiocytes are involved in fluid secretion with the majority of CFTR normally located intracellularly, in a specialized vesicular compartment that inserts into the plasma membrane upon stimulation by cAMP-dependent agonists (49).

CFTR’s apical location increases in FPC-null mice. This increase in apical localization parallels what we have observed in cysts in ADPKD, which are known to secrete fluid via a cAMP-dependent mechanism (8, 30). CFTR also colocalizes with the basolateral membrane marker Na⁺-K⁺ ATPase, albeit less strongly.

We have shown previously that cholangiocytes derived from FPC-null mice can form cysts in vitro and that cyst growth is inhibited by VX-809 (26). As mentioned earlier, we found that CFTR protein measured in the cholangiocyte lysate is ∼50% lower in cholangiocytes derived from the FPC-null, del4 mice compared with wild-type mice, whereas its localization to the apical membrane increases. In an elegant and important publication, Banales et al. (36) dissected bile ducts from the PCK rat, a model for ARPKD (50). They showed that steady-state levels of CFTR are higher in cholangiocytes dissected from the rats, whereas we previously recorded a decrease in CFTR protein levels in a cholangiocyte cell line that originated from a mouse model of ARPKD (26). The differences between the two data sets might be attributed to differences between the rat and mouse models or culture versus isolated tissues. Although there are differences in some of their results compared with our previous publication, other key findings are the same. For example, both studies found an increase in apically located CFTR in ARPKD cholangiocytes. Strikingly, as we have observed here, they showed that CFTR is present in the basolateral membrane of normal cholangiocytes. They showed that CFTR in the basolateral plays an important role in fluid transport in ARPKD. Both their study and ours conclude that ARPKD is associated with abnormal expression and location of ion transport proteins (26, 36).

Does the Primary Cilium Play a Role?

Cilia are prominent in cholangiocytes but inherently they are absent in hepatocytes (51). They play a role in cholangiocyte signal transduction as mechanosensors, osmosensors, and chemosensors; sensing both the flow and composition of bile (51). FPC is present in the primary cilium (37), although its functional role at that location is not entirely clear (52). For example, some studies have observed shorter and fewer cilia in the kidneys of FPC-null mice and posited that FPC plays a role in ciliogenesis and renal tubule formation (53). In the PCK rat model, Masyuk et al. (54) found that FPC is present in the primary cilium of cholangiocytes isolated from intrahepatic bile ducts of normal rat livers and become shorter when pkd1 is silenced using siRNA. They also showed that cilia length was diminished in the PCK rat cholangiocytes compared with normal rats. However, others have shown that silencing of FPC in renal cells does not affect either the cilia number or length (37). The latter is most consistent with the data reported here, which indicated only minimal differences in cilia length between WT FPC-containing and FPC-null cholangiocytes. Furthermore, others have shown that ablating cilia in a kif3a^−/−, pkhd1^del4/del4 double mutant mouse has no additional effect when compared with pkhd1^del4/del4 alone. This finding suggests that liver pathology in ARPKD is not influenced by the primary cilia. This situation is in sharp contrast to the role of the cilia in ADPKD, in which ablation of the cilia has a profound effect in moderating disease, particularly if the cilia are ablated after birth (55).

FPC binds to PC2 and colocalizes with it in the primary cilium (37), suggesting a functional role for both proteins in the primary cilium. Studies of PC2 currents in bilayers show that FPC increases the opening of PC2-generated ion channel currents, suggesting that FPC is a positive regulator of PC2 channel activity. In addition to regulating channel activity, FPC also regulates the expression of PC2 (54). We have found here that in del3-4 mice, the level of PC2 in the cilia of cholangiocytes is increased in del3-4 mice.

The primary cilium is a privileged compartment, with entry allowed only for a limited number of proteins such as the PCs (56). FPC is located throughout the components of the primary cilium including the basal body, the transition zone, and the intraflagellar transport system (see Ref. 57 for a review). Given the role of the transition zone as part of the ciliary gate that determines which proteins enter the cilium, an increase in PC2 in the cilia could indicate that the ciliary gate is defective in FPC mutant cells.

We also found that CFTR is present in the primary cilia of cholangiocytes and its levels increased in FPC mutant cells. It is possible that CFTR is playing a role in transporting Cl⁻ across the plasma membrane of the cilium, acting as a counterion to support the movement of Ca²⁺ via PC2 or the function of other Ca²⁺ channels within the ciliary membrane.

Importantly, VX-809 restores the levels of PC2 and CFTR in the cilia toward normal values.

ER-Associated Protein Degradation

CFTR is present in both cholangiocytes, where it drives fluid secretion, and in hepatocytes, where it plays a role in regulating ER stress (15, 17, 26). Regarding the latter effect, we show here that HSP27, HSP70, and HSP90 are elevated in the livers of del3-4 mice. The hepatocyte is the major cell type in the liver, and thus the changes we have noted here are likely the result of an alteration occurring primarily in the hepatocytes. We have observed previously that several HSPs are misregulated in FPC-null cholangiocytes (26). Thus, misregulation of HSP27, HSP70, and HSP90 appears to be common to both hepatocytes and cholangiocytes and both together are likely to contribute to liver disease in ARPKD and as such could be a potential target for therapeutic approaches for both diseases.

What is the consequence of the misregulation of HSPs? HSP27, HSP70, and HSP90 play an active role in the processing and trafficking of mutant CFTR (26, 29, 40, 58) and are most likely involved in similar processes in the mouse liver. Indeed, we have shown previously that silencing either HSP70 or HSP90 inhibits cyst growth in FPC-null cholangiocytes grown in three-dimensional (3-D) culture, suggesting that these HSPs are drivers of cyst growth in the liver. Misregulated HSPs, primarily HSP90, have a direct role in the etiology of fibrosis in several organs, including the liver (59). Interestingly, Yang and coworkers have shown in a homocysteine-induced liver injury model that mouse fed a high-methionine diet develop ER stress and reduced the levels of CFTR in their hepatocytes. Interestingly, treating the hepatic cells with the potentiator ivacaftor (VX-770) alleviates the ER stress. To take this one step further, they treated hepatic cells with the CFTR inhibitor inh¹⁷² (60), which exacerbated ER stress. These results point to a direct involvement of CFTR in the process of ER stress development in the liver. Thus, these findings and our data presented here highlight a role for CFTR modulators as a potential therapy for liver disease in ARPKD.

Conclusions

A deficit of FPC in a mouse model of ARPKD leads to bile duct abnormalities, enhanced cholangiocyte proliferation, and misregulated HSPs. These altered processes are reduced toward normal by the CFTR modulator, VX-809 suggesting that CFTR correctors could be useful therapeutics for ARPKD. Given that they are already approved for use in humans (61), and with further experimentation on liver fibrosis and kidney disease, they can potentially be fast-tracked for clinical use for ADPKD.

DATA AVAILABILITY

All data are contained within the manuscript. Original Western blots and confocal images are available upon request.

GRANTS

This study was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK125272 (to L.C.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

L.C. conceived and designed research; M.K.Y. and A.Z. performed experiments; M.K.Y., A.Z., and L.C. analyzed data; M.K.Y. and L.C. interpreted results of experiments; M.K.Y., A.Z., and L.C. prepared figures; L.C. drafted manuscript; L.C. edited and revised manuscript; M.K.Y. and L.C. approved final version of manuscript.