Epithelial Vasopressin Type-2 Receptors Regulate Myofibroblasts by a YAP-CCN2–Dependent Mechanism in Polycystic Kidney Disease

Significance Statement

In autosomal dominant polycystic kidney disease (ADPKD), progressive fibrosis contributes to renal failure, leading to ESKD. The vasopressin type-2 receptor (V2R) helps to regulate renal water homeostasis and stimulates cyst expansion in ADPKD. We discovered a novel pathogenic pathway behind V2R regulation of fibrosis in ADPKD kidneys. Epithelial V2R stimulation activates interstitial myofibroblasts, in a paracrine manner, in Pkd1 gene knockout (KO) mice. Pharmacologic inhibition and gene knockout studies indicated that V2R regulates myofibroblast activation by a yes-associated protein (YAP)– and connective tissue growth factor (CCN2)–dependent mechanism. The V2R-YAP-CCN2 molecular axis may present novel pharmacologic targets for control of fibrosis in ADPKD.

Visual Abstract

Abstract

Background

Fibrosis is a major cause of loss of renal function in autosomal dominant polycystic kidney disease (ADPKD). In this study, we examined whether vasopressin type-2 receptor (V2R) activity in cystic epithelial cells can stimulate interstitial myofibroblasts and fibrosis in ADPKD kidneys.

Methods

We treated Pkd1 gene knockout (Pkd1KO) mice with dDAVP, a V2R agonist, for 3 days and evaluated the effect on myofibroblast deposition of extracellular matrix (ECM). We also analyzed the effects of conditioned media from primary cultures of human ADPKD cystic epithelial cells on myofibroblast activation. Because secretion of the profibrotic connective tissue growth factor (CCN2) increased significantly in dDAVP-treated Pkd1KO mouse kidneys, we examined its role in V2R-dependent fibrosis in ADPKD as well as that of yes-associated protein (YAP).

Results

V2R stimulation using dDAVP increased the renal interstitial myofibroblast population and ECM deposition. Similarly, conditioned media from human ADPKD cystic epithelial cells increased myofibroblast activation in vitro, suggesting a paracrine mechanism. Renal collecting duct–specific gene deletion of CCN2 significantly reduced cyst growth and myofibroblasts in Pkd1KO mouse kidneys. We found that YAP regulates CCN2, and YAP inhibition or gene deletion reduces renal fibrosis in Pkd1KO mouse kidneys. Importantly, YAP inactivation blocks the dDAVP-induced increase in myofibroblasts in Pkd1KO kidneys. Further in vitro studies showed that V2R regulates YAP by an ERK1/2-dependent mechanism in human ADPKD cystic epithelial cells.

Conclusions

Our results demonstrate a novel mechanism by which cystic epithelial cells stimulate myofibroblasts in the pericystic microenvironment, leading to fibrosis in ADPKD. The V2R-YAP-CCN2 cell signaling pathway may present a potential therapeutic target for fibrosis in ADPKD.

Autosomal dominant polycystic kidney disease (ADPKD) is characterized by the growth of fluid-filled cysts and is the most common inherited kidney disorder, affecting over 12.5 million people worldwide.¹ Cysts in ADPKD commonly originate from renal collecting ducts and distal tubular segments.^2,3 Progressive tubulointerstitial fibrosis accompanies cyst expansion in ADPKD and is thought to be a major contributor to the decline in renal function and the final path to ESKD.^4,5 The minimal interstitial space around renal tubules is normally occupied by blood capillaries, small vessels, and interstitial cells and is important for proper kidney function. Interstitial expansion due to excessive extracellular matrix (ECM) production, reduced degradation, and changes in the ECM composition, together with increased inflammatory infiltrates, disrupts this delicate architecture, leading to loss of renal function in ADPKD.^4–6 Myofibroblasts are highly contractile cells that express α-smooth muscle actin (α-SMA), migrate, proliferate, and persistently secrete large amounts of ECM.⁷ Renal myofibroblasts are known to originate by activation and differentiation of resident fibroblasts and pericytes.⁸ As in other CKDs, increased myofibroblast numbers, indicated by high α-SMA expression, are known to occur in human and mouse polycystic kidney disease (PKD) kidneys.⁴ However, it is currently unknown whether the cystic epithelium has a role in myofibroblast activation and in modifying the pericystic microenvironment to promote fibrosis.

In this study, we tested whether an epithelial-specific stimulus from cystic epithelial cells can regulate interstitial myofibroblast activation in ADPKD kidney. This hypothesis was on the basis of our serendipitous finding that short-term stimulation of the vasopressin type-2 receptor (V2R) significantly increased renal fibrosis in ADPKD kidneys. V2R is normally expressed on renal tubular epithelial cells of the collecting ducts, connecting tubules and thick ascending limbs,² and V2R-mediated cell signaling promotes cystic epithelial cell proliferation in PKD.^9,10 It is unknown if V2R plays any role in renal fibrosis in PKD. Yes-associated protein (YAP) of the Hippo signaling pathway regulates the proliferation of cystic epithelial cells in ADPKD mouse kidneys¹¹ and autosomal recessive PKD cystic liver cells in vitro.¹² YAP can be activated by ECM stiffness, which can stimulate production of profibrotic factors and ECM proteins by fibroblasts and increase proliferation of epithelial cells.¹³ However, the role of the Hippo pathway in either V2R-mediated cell signaling or fibrosis in PKD has not been reported.

Methods

ADPKD Mouse Studies

Mouse Models

(1) One model was Pkd1^f/fPkhd1^cre (Pkd1 gene knockout [Pkd1KO]) orthologous ADPKD mouse model with collecting duct–specific - Pkd1 gene deletion.¹⁴ Wild-type (WT) mice were Pkd1^f/f mice without Cre. (2) Another model is collecting duct–specific YAP knockout Pkd1^f/fYap^f/fPkhd1^cre mice (Yap^f/f mouse is Yap1^tm1.1Dupa/J; stock no. 027929; The Jackson Laboratory, Bar Harbor, ME). (3) The third model is collecting duct–specific CCN2 knockout Pkd1^f/fCCN2^f/fPkhd1^cre (Pkd1-CCN2KO). CCN2 exon 2 floxed mice¹⁵ were received from A.L.

Desmopressin Study

Pkd1KO mice were administered desmopressin (dDAVP; Sigma-Aldrich, St. Louis, MO; 1 µg/kg body wt daily intraperitoneally) on postnatal days P18–P20. The treatment time was selected on the basis of the findings that by P18, Pkd1KO mouse kidneys are highly cystic and that cAMP levels are already high.¹⁶ In a separate study, Pkd1KO mice and YAP gene knockout Pkd1^f/fYap^f/fPkhd1^cre (Pkd1-YapKO) mice were administered dDAVP (1 µg/kg body wt intraperitoneally daily from P18 to P21).

Verteporfin Study

Pkd1KO mice were administered vehicle or verteporfin (75 mg/kg body wt intraperitoneally) on P10, P12, P14, and P16 and euthanized on P18. Verteporfin was dissolved in vehicle (50% corn oil +25% methanol +25% DMSO), vortexed vigorously, and sonicated before injection.

Short-Term dDAVP Treatment

On P18, WT and Pkd1KO mice were given two dDAVP injections (1 µg/kg body wt intraperitoneally) at 8 am and 12 pm and euthanized at 4 pm of the same day.

All animal studies were carried out according to the protocols approved by the University of Kansas Medical Center Institutional Animal Care and Use Committee.

Human Tissues and Cells

Primary culture ADPKD and normal human kidney (NHK) cells and kidney tissue from deidentified patients were from the PKD Biomaterials Core at the University of Kansas. NRK-49F (ATCC CRL-1570) cell lines were also used.

Quantification of Cysts and Tissue Fibrosis

Kidney tissue sections (5 µm) were stained with hematoxylin and eosin, and images were captured using a microscope connected to a digital camera (Leica Microsystems, Buffalo Grove, IL). An observer blinded to the sample identity recorded and quantified the number of cysts, cystic area, and total kidney area using ImageJ (Fiji, Madison, WI).

BUN Levels

BUN levels were measured in serum as described previously¹⁴ using the QuantiChrom Urea Assay Kit from BioAssay Systems (Hayward, CA).

Urinary Osmolality

Spot urine samples were collected before euthanizing the mouse, and urinary osmolality was measured using the VAPRO vapor pressure osmometer (Model 5600; ELITech Benelux).

Western Blot

Mouse kidneys were homogenized in SDS Laemmli buffer and loaded onto 10% or 4%–20% gradient SDS-PAGE gels essentially as described before.^14,17 Primary antibodies for YAP (sc-101199; Santa Cruz Biotechnology, Inc.), CCN2 (sc-101586; Santa Cruz Biotechnology, Inc.), Lamin B (sc-365214; Santa Cruz Biotechnology, Inc.), glyceraldehyde-3-phosphate dehydrogenase (sc-32233; Santa Cruz Biotechnology, Inc.), pERK (sc-7383; Santa Cruz Biotechnology, Inc.), extracellular signal–regulated kinase (ERK; sc-514302; Santa Cruz Biotechnology, Inc.), SMAD (9513S; Cell Signaling Technology, Inc.), pSMAD (9520S; Cell Signaling Technology, Inc.), α-SMA (ab5694; Abcam, Cambridge, MA), V2R (V5514; Millipore Sigma, St. Louis, MO), and type 1 collagen (203002; MD Bioproduct, Oakdale, MN) were used. Secondary antibodies, both anti-mouse (P0447) and anti-rabbit (P0448), were purchased from Dako and ECL reagent (PerkinElmer).

Immunohistochemistry/Immunofluorescence

Fixed and paraffin tissue sections were processed as described before.^14,17 The following primary antibodies were used: α-SMA, Type 1 collagen, CCN2, YAP, V2R, E-Cadherin (sc-7870; Santa Cruz Biotechnology, Inc.), and DBA (RL-1032; Vector Laboratories). For immunohistochemistry, secondary antibodies were applied followed by incubation with Streptavidin HRP conjugate (Invitrogen), and slides were developed with DAB (Vector Laboratories), counterstained with Harris Haematoxylin, dehydrated, and mounted with Permount (Fisher Scientific, Fail Lawn, NJ). For immunofluorescence, goat anti-rabbit IgG fluor and goat anti-mouse IgG Texas red (Invitrogen) secondary antibodies were applied, incubated, washed with PBST, and stained with 4′,6-diamidino-2-phenylindole. Slides were mounted with Flour-G (Invitrogen) and sealed with nail polish. All images were captured using a Nikon 80i upright microscope (Tokyo, Japan) in the KUMC imaging center.

Quantitative Real-Time PCR

RT-PCR using cDNA prepared from RNA isolated from whole-kidney and cultured cells was carried out as described previously.¹⁴ Primer sequences are provided in Supplemental Table 1.

Conditioned Media Collection

Primary culture human ADPKD cystic epithelial cells were grown in 100-mm plates in DME/F-12 (1×) media (catalog no. SH30023.01; HyClone). When confluent, the media were replaced with serum-free media. Conditioned media (CM) were collected after 48 hours, centrifuged to remove debris, and used directly for studies.

Fibroblast to Myofibroblast Differentiation

NRK-49F cells were grown in 100-mm plates in DMEM media (ATCC 30–2002). When 50% confluent, NRK-49F cells were exposed to serum-free CM from ADPKD cells for 48 hours. CM were changed every 24 hours. Cells were lysed, and α-SMA levels were measured by immunoblotting.

Wound Closure Assay

NRK-49F or human ADPKD renal myofibroblasts were seeded in six-well plates and grown until confluent. A sterile pipette tip was used to place a scratch (wound) in the cell monolayer followed by washing with PBS to remove dislodged cells. The wounds were then photographed at the same position at 0 hours and at different time intervals to measure wound closure. A separate set of plates with similar treatment was used to assess cell viability to adjust percentage wound closure to cell viability. Cells were grown in 10% FBS containing medium.

To determine the effect of CCN2 on NRK-49F cell differentiation and wound closure, recombinant CCN2 (connective tissue growth factor, full-length peptide, catalog no. SRP4702–20UG; Sigma Aldrich) was used.

Cell Viability Analyses

Briefly, exponentially growing cells were seeded in 24-well plates. When cells were 40%–50% confluent, they were washed with PBS, and media were replaced with serum-free CM for 48 hours. Following this, cells were incubated in 5 mg/ml 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium solution for 2 hours, and intracellular purple formazan was solubilized in DMSO and quantified by spectrophotometry at 540 nM.

Nuclear-Cytoplasmic Extraction

The cytoplasmic and nuclear fractions were prepared using the NE-PER Nuclear Cytoplasmic Extraction Reagent kit (Thermo Fischer Scientific) following the manufacturer’s protocol.

Statistical Analyses

Values were expressed as mean ± SEM and ± SD for in vivo and in vitro studies, respectively. Data were analyzed by two-tailed unpaired t test with Welch correction or one-way ANOVA followed by the Dunnett multiple comparison test using GraphPad Prism software (Version 5.0d). Probability value of P≤0.05 was considered significant.

Results

Epithelial-Specific V2R Stimulation Increases the Interstitial Myofibroblast Population, ECM Deposition, and Renal Fibrosis in ADPKD Kidneys

In human ADPKD kidneys, immunostaining for α-SMA revealed a dense myofibroblast population in the pericystic area, in close proximity to cystic epithelial cells (Figure 1A). V2R expression was localized to tubular epithelial cells, including cyst-lining cells, but not in the α-SMA–expressing cells (Figure 1A). Human ADPKD kidneys also showed significantly higher α-SMA mRNA levels compared with normal control kidneys (Figure 1B). To determine the effect of a tubular epithelial–specific stimulus on interstitial myofibroblast activation and fibrosis, Pkd1^f/fPkhd1^cre (Pkd1KO) mice^14,16 were treated with dDAVP, a V2R-selective agonist,¹⁸ from postnatal day P18 to P20 and euthanized on P21 (Figure 1C). In Pkd1KO mice, dDAVP treatment significantly upregulated the myofibroblast population and expression of ECM proteins, compared with vehicle treatment, as indicated by higher α-SMA, collagen-1a, collagen-IIIa, and fibronectin mRNA levels (Figure 1D). Collagen-1a and α-SMA protein levels were also higher in dDAVP-treated Pkd1KO mice (Figure 1E, Supplemental Figure 1). Immunostaining for collagen-1a and α-SMA was also increased in dDAVP-treated Pkd1KO kidneys (Figure 1F). Compared with vehicle treatment, a 17% increase in kidney-body weight ratio and 40% increase in BUN levels were observed in dDAVP-treated Pkd1KO mice (Figure 1, G and H). As expected, dDAVP treatment significantly reduced urine osmolality in both WT and Pkd1KO mice, supporting its V2R agonistic activity (Supplemental Figure 2). Notably, in WT mice, dDAVP treatment did not affect α-SMA and collagen-1a expression, kidney-body weight ratio, or BUN (Figure 1, D, E, G, and H). The above results demonstrate that epithelial V2R activation can stimulate interstitial myofibroblast activation in PKD kidneys. The results also suggest that renal cystic epithelial cells could regulate a profibrotic response.

Figure 1.

Activation of V2R increases fibrosis in ADPKD kidneys. (A) In human ADPKD kidney tissue sections, immunostaining for α-SMA (green) and 4′,6-diamidino-2-phenylindole (DAPI; blue), V2R (red) and E-cadherin (green), or V2R (red), and α-SMA (green). Scale bar = 50 μm. *Cysts. (B) α-SMA mRNA in human normal control (n=8) and human ADPKD kidneys (n=9) relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA. (C) Scheme of vehicle or dDAVP (1 µg/kg body wt per day intraperitoneally on P18–P20) treatment in WT and Pkd1KO mice. (D) α-SMA, collagen-1a, collagen-IIIa, and fibronectin (FN1) mRNA relative to GAPDH for WT (n=6) and Pkd1KO (n=8) mice. (E) Immunoblot for α-SMA and collagen-1a. (F) Immunostaining. Scale bar = 50 µm. (G) Two kidney-body weight ratio (percentage; n=8). (H) Plasma BUN levels (n=8). *P<0.05 versus control or vehicle by t test; **P<0.01 versus control or vehicle by t test.

Secreted Factors from ADPKD Renal Tubular Epithelial Cells Can Activate Myofibroblasts

To determine whether ADPKD cystic epithelial cells regulate myofibroblasts, we tested the effect of their CM on myofibroblast activation (fibroblast to myofibroblast differentiation), viability, and migration. CM were collected from primary culture human ADPKD and NHK epithelial cells and tested on primary culture myofibroblasts from human ADPKD kidneys and undifferentiated NRK-49F rat renal fibroblasts. Exposure to ADPKD-CM significantly increased cell viability (Figure 2A) and faster wound closure (Figure 2B) of both human ADPKD myofibroblasts and NRK-49F cells compared with NHK-CM. The human myofibroblasts expressed high levels of α-SMA at baseline, which did not further change upon exposure to ADPKD-CM (Supplemental Figure 3A). However, when NRK-49F fibroblasts were exposed to ADPKD-CM, α-SMA expression increased significantly compared with NHK-CM (Figure 2, C and D, Supplemental Figure 3B). These results suggest that secreted factors from ADPKD epithelial cells can stimulate myofibroblast activation, migration, and viability.

Figure 2.

Secreted factors from ADPKD cystic epithelial cells can induce myofibroblast activation. (A) The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium assay in human primary culture ADPKD fibroblasts and NRK-49F rat fibroblasts incubated in NHK-CM or ADPKD-CM (72 hours exposure to CM): n=3 biologic replicates and n=3–4 technical replicates. (B) Wound healing (percentage wound closure in a scratch assay): n=2 biologic replicates and n=4 technical replicates each. (C) Immunoblot for α-SMA in NRK-49F rat renal fibroblasts incubated for 48 hours with CM from primary culture NHK or ADPKD cells. (D) Densitometry of α-SMA relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH): n=3 biologic replicates and n=1–3 technical replicates. (E) mRNA levels of secreted factors relative to GAPDH mRNA in vehicle or dDAVP-treated Pkd1KO mouse kidneys: n=6. (F) CCN2 mRNA in WT and Pkd1KO mice after 8 hours of dDAVP treatment (1 µg/kg body wt intraperitoneally): n=6. (G) CCN2 mRNA in NHK (n=7) and ADPKD (n=10) kidney tissue and in primary culture human NHK (n=8) and ADPKD (n=7) renal epithelial cells. (H) Immunostaining in human ADPKD kidneys for CCN2 (green) and DBA (red; collecting duct). Each data point represents one individual human sample. *P<0.05 versus NHK by one-way ANOVA followed by the Dunnett multiple comparison test for (B) and by t test for all others; **P<0.01 versus NHK by one-way ANOVA followed by the Dunnett multiple comparison test for (B) and by t test for all others; ***P<0.001 versus NHK by one-way ANOVA followed by the Dunnett multiple comparison test for (B) and by t test for all others. Scale bar = 50 µm

To determine how V2R stimulation activates myofibroblasts, we examined the renal mRNA expression of known secreted and myofibroblast-activating factors in Pkd1KO mice treated with dDAVP for 3 days as in Figure 1. Of the 25 secreted profibrotic factors examined (Supplemental Table 1), renal mRNA levels of CCN2, TGFβ, amphiregulin (AREG), PAI-1, TNFα, IL1β, IL6, IL10, CCL2, and CCL3 were significantly higher in the dDAVP treatment group when compared with the vehicle-treated group (Figure 2E). To further test if V2R signaling can directly stimulate expression of these secreted factors, we performed a short-term study in which dDAVP was administered to WT and Pkd1KO mice, and they were euthanized after only 8 hours. Of the above-mentioned ten secreted factors, only CCN2 mRNA levels increased significantly by twofold in Pkd1KO mice (Figure 2F, Supplemental Figure 3, C and D). The CCN2 mRNA levels remained unaffected by dDAVP treatment in WT mice (Figure 2F). Furthermore, recombinant CCN2 protein induced α-SMA expression and cell migration in a dose-dependent fashion in NRK-49F cells (Supplemental Figure 4, A and B), suggesting that CCN2 could be an important V2R-stimulated secreted profibrotic factor in ADPKD kidneys.

Renal Tubular CCN2 Gene Deletion Reduced Interstitial Myofibroblast Activation, ECM, and Cyst Growth in ADPKD Mouse Kidneys

CCN2 is a matricellular protein that is known to be profibrotic in the kidney,^19,20 secreted by myofibroblasts in mouse CKD models,^21–26 and immunolocalize to human cystic tubular epithelial cells in ADPKD.²⁷ It is unknown if CCN2 plays a pathogenic role in cyst growth or fibrosis in PKD. We found eight- and threefold higher CCN2 mRNA levels, respectively, in human ADPKD kidneys and primary culture ADPKD epithelial cells compared with NHK controls (Figure 2G).

Importantly, CCN2 was localized to the cyst-lining epithelial cells in human ADPKD kidneys (Figure 2H), especially in cysts of collecting duct origin, a tubule segment that expresses V2R (Figure 2H). CCN2 was also detected in cyst-lining epithelium in Pkd1KO mice (Supplemental Figure 4C).

To determine the role of cystic epithelial CCN2 in interstitial myofibroblast activation and fibrosis, we generated a collecting duct–specific CCN2 gene knockout Pkd1^f/fCCN2^f/fPkhd1^cre mouse (Pkd1-CCN2KO). CCN2 expression in collecting duct cystic epithelial cells was reduced in Pkd1-CCN2KO kidneys, whereas other pericystic cells still expressed CCN2 (Supplemental Figure 5A). The Pkd1-CCN2KO kidneys were smaller than Pkd1KO kidneys (Figure 3A) and showed reduced kidney-body weight ratio (32% reduction), BUN (22% reduction), and cystic index (23% reduction) (Figure 3, B–D). Cyst numbers were not significantly different between Pkd1-CCN2KO and Pkd1KO kidneys (Supplemental Figure 5B). Compared with Pkd1KO kidneys, the mRNA levels of α-SMA, collagen-1a, and collagen-IIIa were significantly reduced in the Pkd1-CCN2KO kidneys (Figure 3E). The Pkd1-CCN2KO kidneys also showed significant reduction in α-SMA protein levels (Figure 3, F and G) and reduced immunostaining for α-SMA and collagen-1a (Figure 3H, Supplemental Figure 5C) compared with Pkd1KO kidneys. These results suggest that renal tubular epithelial–specific CCN2 is important for the development of interstitial fibrosis in ADPKD.

Figure 3.

Renal tubular epithelial–specific gene deletion of CCN2 reduced cyst growth and myofibroblast population in ADPKD kidneys. (A) In WT, CCN2^f/fPkhd1^cre (CCN2KO), Pkd1KO, and CCN2^f/fPkd1^f/fPkhd1^cre (Pkd1-CCN2KO) mice, hematoxylin and eosin staining is shown. (B) Kidney-body weight (BWt) ratio. (C) BUN levels. (D) Cystic index. (E) mRNA levels of α-SMA and collagen-1a and collagen-IIIa levels relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA. (F) Immunoblot of kidney tissue lysate and (G) quantitation of band density. (H) Immunostaining of kidney for α-SMA (brown) and collagen-1a (green). *P<0.05 versus Pkd1KO by t test; **P<0.01 versus Pkd1KO by t test; ***P<0.001 versus Pkd1KO by t test. Scale bar = 50 µm

Renal Tubular Epithelial YAP Regulates CCN2 and Fibrosis in ADPKD Kidneys

To determine how CCN2 is regulated in the cystic epithelium, we examined the role of YAP, a transcriptional regulator of CCN2.²⁸ Although YAP is known to regulate cystic epithelial cell proliferation in ADPKD mouse models,^11,12 its role in fibrosis in PKD is unknown. Nuclear YAP expression was detected in the cyst-lining epithelium of human ADPKD and Pkd1KO mouse kidneys (Supplemental Figure 6A). We first determined the role of YAP in CCN2 expression by treating human ADPKD and NHK epithelial cells with verteporfin, an inhibitor of YAP transcriptional enhanced associate domain interaction that is shown to reduce cancer cell proliferation.²⁹ Verteporfin treatment for 16 hours significantly reduced mRNA levels of CCN2, PAI-1, AREG, and CCL2, known transcriptional targets of YAP as well as other profibrotic factors in ADPKD cells (Figure 4A). Although some of these factors were also reduced in verteporfin-treated NHK cells, CCN2 mRNA levels were unaffected (Figure 4A).

Figure 4.

YAP regulates CCN2 in PKD cystic epithelium. (A) mRNA levels in human NHK and ADPKD cells treated with vehicle or verteporfin (Vert; 2.5 µM for 16 hours): n=3 biologic replicates and n=2 technical replicates each. WT or Pkd1KO mice were treated with vehicle or Vert (75 mg/kg body wt intraperitoneally on P10, P12, P14, and P16) and euthanized on P18. (B) Immunostaining for α-SMA and collagen-1a. mRNA levels relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) for (C) α-SMA, (D) collagen-1a, (E) collagen-IIIa, and (F) fibronectin (FN1). (G) Quantitative PCR showing mRNA levels of CCN2, AREG, and PAI-1. (H) Immunoblot of kidney tissue lysates. *P<0.05 versus NHK-vehicle by t test for (A) and versus vehicle by t test in all others; **P<0.01 versus NHK-vehicle by t test for (A) and versus vehicle by t test in all others; ***P<0.001 versus NHK-vehicle by t test for (A) and versus vehicle by t test in all others; ^@@P<0.01 versus ADPKD-vehicle; ^@@@P<0.001 versus ADPKD-vehicle. Scale bar = 50 µm

To examine the role of YAP in myofibroblast activation and renal fibrosis in ADPKD mouse kidneys, we tested the effect of pharmacologic YAP inhibition or renal collecting duct–specific YAP gene deletion. WT and Pkd1KO mice were treated with vehicle or verteporfin from P10 to P18 and euthanized. Verteporfin-treated Pkd1KO mouse kidneys showed reduced immunolabeling of α-SMA and collagen-1a (Figure 4B), whereas no notable changes were observed in verteporfin-treated WT mice (Supplemental Figure 6B). Verteporfin treatment significantly reduced mRNA levels of α-SMA and ECM proteins (Figure 4, C–F), YAP-regulated genes (Figure 4G), and proinflammatory and profibrotic factors (Supplemental Figure 6C) compared with vehicle treatment in Pkd1KO mice. Verteporfin treatment also significantly reduced protein levels of α-SMA, collagen-1a, CCN2, and pSMAD-SMAD ratio in Pkd1KO mice (Figure 4H, Supplemental Figure 7).

Similarly, in collecting duct–specific Pkd1-YapKO mouse kidneys, collagen-1a and α-SMA immunostaining (Figure 5A, Supplemental Figure 8A) and mRNA levels of collagen-1a, α-SMA, and CCN2 were significantly reduced when compared with Pkd1KO mouse kidneys (Figure 5B). At P18, the Pkd1-YapKO mice had significantly smaller kidneys (Figure 5C) and showed a 60% reduction in kidney-body weight ratio and a 55% reduction in BUN levels compared with Pkd1KO mice (Figure 5, D and E). Similarly, verteporfin-treated Pkd1KO mice also showed smaller kidneys, reduced kidney-body weight ratio, and BUN (Figure 5, F–H). Compared with Pkd1-CCN2KO mouse kidneys (Figure 3A), more sparing of the renal cortical parenchyma was observed in both verteporfin-treated Pkd1KO and in Pkd1-YapKO mice (Figure 5, C and F). Renal cyst number and cyst index were also significantly reduced in the Pkd1-YapKO kidneys and verteporfin-treated Pkd1KO kidneys compared with the Pkd1KO kidneys (Supplemental Figure 8, B–E). In both of the above studies, neither verteporfin treatment nor renal collecting duct–specific YAP gene knockout in WT mice caused alterations in renal structure, function, or fibrosis.

Figure 5.

Effect of tubular epithelium–specific YAP gene knockout on renal fibrosis in PKD mice. WT, Yap^f/fPkhd1^cre (Yap KO), Pkd1KO, and Yap^f/fPkd1^f/fPkhd1^cre (Pkd1-YapKO) mice. (A) Immunostaining for collagen-1a and α-SMA. (B) mRNA levels of collagen-1a, α-SMA, and CCN2 relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH; n=6). (C) Hematoxylin and eosin staining of WT, Yap^f/fPkhd1^cre (Yap KO), Pkd1KO, and Yap^f/fPkd1^f/fPkhd1^cre (Pkd1-YapKO) mice kidney sections at P18. (D) Kidney-body weight (BWt) ratio (percentage). (E) BUN levels. (F) H&E staining of WT or Pkd1KO mice treated with vehicle or verteporfin (Vert) kidney sections at P18. (G) Kidney-BWt ratio and (H) BUN levels. *P<0.05 by t test for WT or Pkd1KO mice; **P<0.01 by t test for WT or Pkd1KO mice; ***P<0.001 by t test for WT or Pkd1KO mice. Scale bar = 50 µm

V2R Signaling via ERK1/2 Regulates YAP Expression and Activity in the Cystic Epithelium of ADPKD Kidneys

To determine the mechanism by which V2R signaling regulates YAP, we examined YAP expression in the kidneys of Pkd1KO mice treated with dDAVP as in Figure 1C. Pkd1KO kidneys showed high YAP protein levels compared with WT kidneys, which was further significantly increased by dDAVP treatment (Figure 6, A and B). The dDAVP-treated Pkd1KO mice also showed increased activated ERK1/2 (pERK) levels (Figure 6, A and B). V2R-cAMP-protein kinase A (PKA)–mediated ERK1/2 activation is known to promote cystic epithelial cell proliferation and cyst-filling fluid secretion in PKD,^9,10,30 and ERK1/2 regulates YAP expression in tumor cells.^31,32 Hence, we tested if V2R activation stimulates YAP expression and YAP transcriptional activity in PKD kidneys in an ERK1/2-dependent fashion. In vitro, YAP was detected even in confluent monolayers of primary culture human ADPKD cells, and forskolin, an adenylate cyclase agonist, increased YAP nuclear accumulation (Figure 6C). In ADPKD cells, forskolin-induced YAP protein levels were significantly reduced by UO126, an inhibitor of ERK1/2 (Figure 6, D and E, Supplemental Figure 9). Although YAP levels showed a reducing trend in cells treated with H-89, a PKA inhibitor, it was not significant (Figure 6, D and E, Supplemental Figure 9). These results suggest that the V2R-cAMP-ERK–dependent mechanism could regulate YAP expression in ADPKD cystic epithelial cells.

Figure 6.

V2R-mediated myofibroblast activation in ADPKD kidneys is dependent on ERK1/2 and YAP. (A) Immunoblot of kidney tissue of mice described in Figure 1C: n=3 WT and n=4 Pkd1KO. (B) Quantitation of band density of YAP and pERK-ERK ratio relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH). (C) YAP expression (green) in NHK or ADPKD cells serum deprived (16 hours) and treated with vehicle or forskolin (5 µM for 2 hours). (D) Immunoblot and (E) quantitation of band density for YAP in ADPKD cells treated with forskolin (5 µM) and vehicle, H-89 (5 µM), or UO126 (20 µM) for 16 hours. Experiment was repeated three times. (F) Immunoblot of kidney tissue lysate of dDAVP-treated Pkd1KO mice, dDAVP + verteporfin (Vert)–treated Pkd1KO mice, and dDAVP-treated Pkd1-YapKO mice. (G) Quantitation of band density for α-SMA and collagen-1a relative to GAPDH and (H) urine osmolality measurements in spot urine samples: n=5 for WT and n=6–8 for Pkd1KO mice as shown. *P<0.05 versus vehicle by t test; ***P<0.001 versus vehicle by t test.

To further confirm that YAP plays a critical role in V2R-mediated myofibroblast activation, the effect of dDAVP was tested on vehicle or verteporfin-treated Pkd1KO mice and on Yap-Pkd1KO mice. Although dDAVP treatment increased α-SMA and collagen-1a expression in Pkd1KO mice, neither verteporfin-treated Pkd1KO mice nor Pkd1-YapKO mice showed an increase in α-SMA or collagen-1a (Figure 6, F and G). Urine osmolality in Pkd1KO mice was 80% lower than WT mice (Figure 6H). However, neither collecting ducts YAP gene deletion nor YAP systemic inhibition reduced urine osmolality further in Pkd1KO mice (Figure 6H). Overall, these results suggest that YAP inactivation can reduce myofibroblast activation, fibrosis, and cyst expansion in ADPKD kidneys without the complication of further affecting urine concentrating ability.

Discussion

This study demonstrates a novel mechanism by which V2R-YAP-CCN2–mediated cell signaling in tubular epithelium activates interstitial myofibroblasts and fibrosis in ADPKD kidneys. Our novel observations include the finding that (1) V2R activation increases myofibroblast population; (2) CCN2 is an important V2R-stimulated secreted factor that activates myofibroblasts; (3) YAP activity is key to V2R-mediated CCN2 production and interstitial fibrosis in ADPKD; and (4) unlike V2R inhibition, YAP inhibition does not reduce urine concentrating ability in mice.

This study provides the first example of a new pathologic role for V2R in which stimulation of V2R in ADPKD cystic epithelial cells increases the myofibroblast population in the pericystic microenvironment. Normal V2R activity is essential for the urine concentrating function of the kidney collecting ducts. However, V2R plays a key pathogenic role in PKD by upregulating the cAMP-PKA-ERK1/2 MAPK pathway to drive cell proliferation and cyst expansion.^9,14 Arginine vasopressin deficiency or V2R antagonists are known to reduce cyst growth, whereas dDAVP treatment induces or accelerates cyst growth in rodent models of PKD.^33–37 Importantly, tolvaptan, a V2R antagonist, is the only Food and Drug Administration (FDA)–approved drug for the treatment of ADPKD.³⁸

Although V2R activation has a clear mitogenic effect on ADPKD epithelial cells, this study demonstrates an additional novel role in myofibroblast activation and fibrosis. We found dense populations of α-SMA–expressing myofibroblasts around V2R-expressing cystic epithelium in human ADPKD kidneys. V2R agonist significantly increased profibrotic factors, myofibroblast population, and ECM proteins in cystic mouse kidneys. Importantly, conditioned culture media from human ADPKD cystic epithelial cells stimulated activation, migration, and viability of fibroblasts in vitro. Renal fibrosis is recognized as a major pathology linked to loss of renal function in ADPKD, and TGFβ, inflammatory cells, and matricellular proteins are thought to enhance this fibrosis.^4,5,39–43 Activated stromal myofibroblasts, also called cancer-associated fibroblasts in the tumor microenvironment, are known to enhance tumor cell proliferation and metastasis, whereas tumor cells in turn can stimulate myofibroblast activation and fibrosis.⁴⁴ Our study now demonstrates that cystic epithelial cells can regulate myofibroblasts by a V2R-dependent mechanism.

This study also provides evidence that V2R-stimulated myofibroblast activation is mediated by YAP. YAP and its homolog—transcriptional coactivator with PDZ-binding motif—are transcriptional coactivators of transcriptional enhanced associate domain that regulate cell proliferation, differentiation, and apoptosis.⁴⁵ Activation of the Hippo signaling cascade leads to phosphorylation, nuclear exclusion, cytoplasmic sequestration, and proteolytic degradation of YAP and transcriptional coactivator with PDZ-binding motif.^46–48 YAP protein is elevated and promotes fibrosis in mouse CKD models.⁴⁸ YAP was detected in the cystic epithelium of human and mouse PKD kidneys,²⁷ and YAP gene deletion reduced cyst growth in ADPKD mouse models.¹¹ However, the role of YAP in fibrosis in PKD had not been previously examined. We demonstrated that in response to dDAVP treatment, renal YAP levels and YAP target gene expression were increased in PKD mouse models. Moreover, pharmacologic YAP inhibition and renal tubule–specific YAP gene deletion significantly reduced ECM proteins in PKD mouse kidneys, which could not be reversed by dDAVP treatment. These results suggest that (1) V2R and Hippo signaling pathways interact, (2) YAP is a downstream effector of V2R signaling, and (3) cystic epithelial YAP is critical for V2R-stimulated fibrosis in PKD. The importance of V2R and YAP in myofibroblast activation and fibrosis was also shown by reduced α-SMA expression and fibrosis in Pkd1-YapKO mice and the inability of dDAVP to induce significant α-SMA expression and fibrosis in these mice. Use of verteporfin, a YAP inhibitor and FDA-approved drug for age-related neovascular macular degeneration, could be a potentially useful therapeutic strategy to slow or inhibit renal fibrosis in PKD.

In this study, human ADPKD primary culture cells and Pkd1KO mice showed increased YAP expression and transcriptional activity when treated with forskolin or dDAVP. This increase in YAP coincided with increased pERK1/2 levels in Pkd1KO mice and was inhibited by ERK inhibition in ADPKD cells. These findings in ADPKD are in contrast with prior observations that cAMP-PKA signaling stimulated by forskolin increases Hippo signaling, leading to LATS-mediated phosphorylation and inhibition of YAP in breast cancer cell lines, HEK293 cells,⁴⁹ and NIH3T3L1 fibroblasts.⁵⁰ This could be because cAMP-mediated signaling activates the B-Raf/ERK pathway in cultured PKD cells but not in normal cells.³⁰ V2R-cAMP-PKA–mediated ERK1/2 activation promotes cystic epithelial cell proliferation and cyst-filling fluid secretion in PKD,^9,10,30 and ERK1/2 is known to regulate YAP expression in tumor cells.^31,32

Our finding that epithelial cell V2R and YAP can stimulate myofibroblast activation and fibrosis in ADPKD kidneys suggests the involvement of secreted factors in mediating this paracrine epithelial-myofibroblast interaction. Consistently, examination of dDAVP-treated Pkd1KO mouse kidneys showed increased mRNA levels of YAP-regulated secreted profibrotic factors, such as CCN2, AREG, PAI-1, and CCL2. Moreover, these factors could be significantly reduced by verteporfin treatment in ADPKD cells or Pkd1KO mice. CCN2 could be an important V2R-YAP–regulated secreted factor that contributes to myofibroblast activation in ADPKD kidneys. The role of CCN2 in fibrosis^19,20 has been conclusively shown using CCN2 inactivation in a wide variety of systems, including CKD models such as UUO, diabetic nephropathy, GN, and severe AKI.^21–26 Although high-throughput sequencing of human and mouse PKD kidneys has shown CCN2 to be a major RNA transcript,^41,51–53 it is unknown if CCN2 plays a pathogenic role in PKD. We found that CCN2 gene deletion in collecting ducts significantly reduced not just cyst growth but also, the pericystic myofibroblast population in mouse ADPKD kidneys. Although other experimental CKD models studied autocrine regulation of CCN2 by fibroblasts,^{21–23,25,26} we found abundant CCN2 expression in the cystic epithelial cells in human and mouse ADPKD kidneys. CCN2 was also increased by dDAVP treatment in Pkd1KO mice and reduced by YAP inactivation in Pkd1KO mice and ADPKD cells. This suggests that cystic epithelial CCN2 plays an important role in V2R-YAP–mediated stimulation of renal fibrosis in PKD. Although immune cell infiltrates were not examined in this study, the observation that V2R and YAP can regulate multiple proinflammatory factors, including CCN2, suggests that they could be involved in attracting/activating macrophages, T cells, and granulocytes in ADPKD kidneys and importantly, could also be involved in fibrosis.

In conclusion, these findings provide new insights into a novel pathogenic mechanism by which V2R, an epithelial-specific hormone receptor, stimulates fibrosis in ADPKD. We identify YAP as an important mediator of myofibroblast activation and ECM production (Figure 7). Pharmacologic approaches targeting the V2R-YAP-CCN2 molecular axis may have strong implications for control of fibrosis in ADPKD, suggesting new targets for therapy.

Figure 7.

Cystic epithelial cells stimulate myofibroblast activation in the pericystic microenvironment leading to fibrosis in ADPKD. In ADPKD kidneys, V2R mediated cell signaling activates YAP in cystic epithelial cells leading to increased CCN2 expression, paracrine activation of interstitial myofibroblasts and ECM production.

Disclosures

A. Leask reports that he is a shareholder (<5%, <$500,000) in FibroGen. D. Wallace is a consultant for NovaTarg Therapeutics, Chinook Therapeutics, and Vertex Pharmaceuticals. All remaining authors have nothing to disclose.

Funding

This study was supported by National Institute of Diabetes and Digestive and Kidney Diseases grant R01-DK083525 (to R. Rao). N. Dwivedi was supported by American Heart Association postdoctoral fellowship grant 19POST34380932 and a postdoctoral fellowship grant from the University of Kansas Biomedical Research and Training Program.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2020020190/-/DCSupplemental.

Supplemental Figure 1. Quantitation of band density of α-SMA and collagen-1a relative to GAPDH for mice treated with dDAVP.

Supplemental Figure 2. Urinary osmolality in mice treated with dDAVP.

Supplemental Figure 3. Effect of conditioned media from human primary culture NHK or ADPKD renal cystic epithelial cells on human or rodent renal fibroblasts and mRNA expression of secreted factors from ADPKD mouse kidneys.

Supplemental Figure 4. CCN2 induces myofibroblast activation and migration in vitro.

Supplemental Figure 5. Immunostaining and cyst number in Pkd1KO and Pkd1KO-CCN2KO mouse kidneys.

Supplemental Figure 6. Immunostaining for YAP, collagen-1a, and α-SMA and mRNA levels.

Supplemental Figure 7. Quantitation of band density from immunoblot of kidney lysate from WT or Pkd1KO mice treated with vehicle or verteporfin.

Supplemental Figure 8. Immunostaining, cyst number, and cystic index of Pkd1YAPKO mice or Pkd1KO mice treated with verteporfin.

Supplemental Figure 9. YAP expression in primary culture human ADPKD cells treated with sorskolin, H-89, or UO126.

Supplemental Table 1. List of mouse primers used for quantitative RT-PCR and list of human primers used for quantitative RT-PCR.