Proliferation of cyst-lining epithelial cells is an integral part of autosomal dominant polycystic kidney disease (ADPKD) cyst growth. Cytokines and growth factors within cyst fluids are positioned to induce cyst growth. Vascular endothelial growth factor (VEGF) is a pleiotropic growth factor present in ADPKD liver cyst fluids (human 1,128 ± 78, mouse 2,787 ± 136 pg/ml) and, to a lesser extent, in ADPKD renal cyst fluids (human 294 ± 41, mouse 191 ± 90 pg/ml). Western blotting showed that receptors for VEGF (VEGFR1 and VEGFR2) were present in both normal mouse bile ducts and pkd2(WS25/−) liver cyst epithelial cells. Treatment of pkd2(WS25/−) liver cyst epithelial cells with VEGF (50–50,000 pg/ml) or liver cyst fluid induced a proliferative response. The effect on proliferation of liver cyst fluid was inhibited by SU-5416, a potent VEGF receptor inhibitor. Treatment of pkd2(WS25/−) mice between 4 and 8 mo of age with SU-5416 markedly reduced the cyst volume density of the liver (vehicle 9.9 ± 4.3%, SU-5416 1.8 ± 0.7% of liver). SU-5416 treatment between 4 and 12 mo of age markedly protected against increases in liver weight [pkd2(+/+) 4.8 ± 0.2%, pkd2(WS25/−)-vehicle 10.8 ± 1.9%, pkd2(WS25/−)-SU-5416 4.8 ± 0.4% body wt]. The capacity of VEGF signaling to induce in vitro proliferation of pkd2(WS25/−) liver cyst epithelial cells and inhibition of in vivo VEGF signaling to retard liver cyst growth in pkd2(WS25/−) mice indicates that the VEGF signaling pathway is a potentially important therapeutic target in the treatment of ADPKD liver cyst disease.
- autosomal dominant polycystic kidney disease
- growth factors
autosomal dominant polycystic kidney disease (ADPKD) is a genetic disease that occurs in ∼1 in 800 individuals. The clinical symptoms generally appear after the third decade of life. The most prevalent clinical manifestation of ADPKD is development of renal cysts and loss of renal function. ADPKD accounts for ≥5% of all end-stage renal disease. The most common extrarenal manifestation, accounting for up to 10% of ADPKD morbidity and mortality, is the development of liver cysts. Within a clinically affected liver the parenchymal volume does not change significantly (7), but the liver can have hundreds of distinct cysts, with cysts reaching several centimeters in diameter. In the liver, the clinical manifestations of ADPKD are associated directly with the compression of the surrounding tissues, organs, and vasculature. Consequently, growth of liver cysts is considered a key element in the onset of the clinical features. Medical therapies that block liver cyst growth are predicted to largely eliminate the clinical consequences of ADPKD liver cyst disease.
ADPKD is considered a molecular recessive disease and is linked to mutations in either PKD1 or PKD2. Affected individuals are predicted to have a germ line mutation in one copy of the gene and, during their lifetime, to undergo a somatic mutation in individual cells in the second copy of the gene. In the kidney, cysts emerge from epithelial cells lining the nephron. In the liver, cysts emerge from cholangiocytes, the epithelial cells lining the intrahepatic bile duct. After the gene mutation, individual epithelial cells undergo clonal expansion to initially form an outpocket from the nephron or duct and subsequently detach from the original nephron or duct to form an autonomous, enclosed cyst within the body of the organ. Lined by a single layer of epithelial cells, these cysts continue to grow and emerge from the organs as superficial cysts. In the case of the ADPKD liver, cysts can enlarge to the point where the total liver volume more than doubles.
While cystogenesis is linked to mutations in PKD1 and PKD2, the development of the disease is markedly heterogeneous. Part of this heterogeneity may be influenced by the nature of the germline mutation and the nature of the second, somatic mutation. However, sibship studies and linkage analyses indicate that much of the heterogeneity is likely due to modifier genes and extragenetic factors (9). One potential source of extragenetic influence is autocrine/paracrine signaling by cytokines and growth factors synthesized and released by epithelial cells that line the cysts (12, 24). Vascular endothelial growth factor (VEGF) is among these identified factors. Under physiological conditions, VEGF promotes angiogenesis in developing tissues and helps maintain the status of established endothelial cells (27). VEGF also contributes to the pathophysiology of specific tumors by promoting their neovascularization (25). Accordingly, the present study investigated the contribution of VEGF to the growth of ADPKD cysts. It showed that 1) VEGF accumulates in ADPKD liver cyst fluid to levels that can drive the proliferation of liver cyst epithelial cells in culture and 2) inhibition of the VEGF signaling pathway in vivo markedly inhibits liver cyst growth in a mouse model of ADPKD.
pkd2(WS25/−) mouse model.
C57BL/6 pkd2(WS25/−) mice were developed by Stefan Somlo (Yale University), and breeding pairs were provided for this project (35). The animals were cared for in the Center for Laboratory Animal Research at the University of Colorado Health Sciences Center (UCHSC), and all experimental protocols were approved by the UCHSC Animal Care and Use Committee. C57BL/6 pkd2(+/−) and C57BL/6 pkd2(WS25/+) mice were used as breeding pairs to generate pkd2(WS25/−) mice for the study. Mice were genotyped by Southern blotting (35). Briefly, DNA was isolated from tail snips (Gentra Systems, Minneapolis, MN) and digested overnight with PstI restriction enzyme (Promega, Madison, WI). The cut DNA was run out on a 0.7% agarose gel and transferred onto nylon membranes overnight. A DNA probe specific to the recombinant region was labeled with 32P by random priming (Decaprime II, Ambion, Austin, TX). The probes were incubated with the blots overnight. Labeled blots were washed and developed by autoradiography. pkd2(WS25/−) mice closely model the human condition by having one copy of pkd2 knocked out and having a second, recombinant-sensitive allele (i.e., WS25) that undergoes high rates of recombination to yield knockouts of the second copy of the gene in somatic cells during the life span of the animals. By 4 mo of age, pkd2(WS25/−) mice develop discernible kidney and liver cysts. Over the subsequent 4–8 mo, these cysts grow and emerge from the tissues.
Culture of pkd2(WS25/−) liver cyst epithelial cells.
Liver cyst epithelial cells (termed pkd2 cells) were isolated from pkd2(WS25/−) mice and grown in culture (5). Unless otherwise noted, media and reagents for cell cultures were obtained from GIBCO-Invitrogen (Carlsbad, CA). pkd2(WS25/−) mice were grown to 10–12 mo of age to allow the liver cysts to grow out from the liver. Mice were euthanized, and superficial liver cyst wall tissues were excised and finely minced under clean conditions. The minced tissue was suspended in 20 ml of growth medium [DMEM-F12, 5% fetal bovine serum (FBS; Hyclone, Logan, UT), 2 mM glutamine, 1% nonessential amino acids, 1% lipid concentrate, 1% vitamin solution, 393 ng/ml dexamethasone (Sigma, St. Louis, MO), 25 ng/ml epidermal growth factor (EGF; Upstate Biologics, Lake Placid, NY), 30 μg/ml bovine pituitary extract (Upstate Biologics), 3.4 μg/ml triiodothyronine, 1% insulin-transferrin-selenium, 0.4 μg/ml forskolin (Sigma), 50 μg/ml soybean trypsin inhibitor, pH 7.5] and placed above a 2-mm slab of rat tail collagen. Epithelial cells were allowed to grow out onto the collagen slab for 4–5 days before refeeding. Biochemical, histological, and physiological assessment have shown that these cells have the characteristics of cholangiocytes, the progenitor cell of liver cyst epithelia (5). Cells were passaged by digesting the collagen from beneath the cells [1 mg/ml collagenase (Sigma), 2 mg/ml dispase; 60 min at 37°C], disrupting the monolayer into small epithelial “islands” and reseeding with these islands. Liver cyst epithelial cells were similarly isolated from BALB/c cpk(+/−) mice, a genetic model with mutations in the Cys-1 gene that similarly give rise to liver cysts (5).
Polarized model of liver cyst epithelial cells.
To develop a polarized epithelial model of liver cyst-lining epithelial cells, pkd2 liver cyst epithelial cells were seeded and grown into confluent monolayers on semipermeable supports. After high transepithelial resistances (>1,000 Ω·cm2) developed, the polarity of cytokine and growth factor secretion was determined by exchanging the apical and basal media and sampling the apical and basal media over the following 3 days. The concentration of VEGF was determined by ELISA analysis as described below.
Isolation of intrahepatic bile ducts.
Normal bile ducts were enriched by an extended collagenase digestion of the liver (14). Livers of pkd2(+/+) mice were perfused for 45–60 min with collagenase solution (DMEM-F12, 0.05% type I collagenase; Worthington, Lakewood, NJ; 37°C). The bulk of the hepatocytes were lightly teased away from the biliary tree. The biliary tree was further shaken in the collagenase solution for 15–30 min to dissociate hepatocytes from the biliary tissue. Western blotting for bile salt export protein (hepatocyte marker) and ezrin (cholangiocyte marker) showed the preparations to be enriched a minimum of fivefold for cholangiocytes (data not shown).
Collection of liver and kidney cyst fluids.
Human kidney and liver cyst fluids were collected from individuals undergoing laparoscopic fenestration procedures. Fluids were drawn into a 10-ml syringe, transferred into sterile tubes, frozen in liquid nitrogen, and stored at −80°C before ELISA analysis. Protocols for collection and use of the materials were approved by the Colorado Multiple Institution Review Board, and all patients provided informed consent for participation in the study. Similarly, liver and kidney cyst fluids from euthanized pkd2(WS25/−) mice were drawn into a 1- or 5-ml syringe, transferred, frozen in liquid nitrogen, and stored at −80°C before analysis.
Cell proliferation assays.
The rates of proliferation of the pkd2 liver cyst epithelial cells were measured by Alamar Blue assay (Biosource/Invitrogen, Carlsbad, CA). The proliferative response of these cells was measured in response to treatment with recombinant mouse VEGF (R&D Systems, Minneapolis, MN) or liver cyst fluid. pkd2(WS25/−) liver cyst epithelial cells were passaged onto collagentreated six-well plates (BD Bioscience, San Jose, CA), allowed to seed overnight in growth medium, incubated for 24 h in serum-free DMEM-F12, and then cultured for 24 h in DMEM-F12 with no serum (baseline control; value set at 0%), 10% FBS (Hyclone, Logan, UT; positive control, value set at 100%), serum-free medium containing 0.05–50 ng/ml recombinant mouse VEGF (R&D Systems), or serum-free medium containing 10% pkd2(WS25/−) liver cyst fluid. Subsequently, as described by the manufacturer, serum-free medium containing 10% Alamar Blue was added to the cells for 18 h, the medium was evaluated for absorbance at 540 and 600 nm, and the reduction of Alamar Blue was calculated. The proliferative effect of liver cyst fluid was confirmed by Western blotting for PCNA (data not shown).
Cytokine array and ELISA assays.
The broad pattern of growth factor and cytokine expression in mouse liver cyst fluid was evaluated by cytokine array analysis, as described by the manufacturer (Panomics, Redwood City, CA). The concentrations of VEGF in mouse liver cyst fluids were measured by ELISAs on 96-well VEGF ELISA plates purchased from ELISA Tech (Aurora, CO). Recombinant mouse VEGF was used to calibrate the responses.
Western blotting assays.
Western blots were performed by standard methods (4). Tissue or cell culture samples were solubilized in 5× PAGE sample buffer (5% sodium dodecyl sulfate, 25% sucrose, 50 mM Tris, 5 mM EDTA, 200 mM dithiothreitol, pH 8.0), the protein content was measured by bicinchoninic acid assay (Pierce, Rockford, IL), and 40 μg of protein was loaded per lane. Nitrocellulose blots were probed with commercial antibodies against PCNA (BD Bioscience; 1:5,000), VEGFR1 (Flt-1) (Epitomics, Burlingame, CA; 1:5,000), VEGFR2 (Flk-1/KDR) (Affinity Bioreagents, Golden, CO; 1:2,000), and actin (Calbiochem, La Jolla, CA; 1:5,000). Primary antibodies were detected with horseradish peroxidase-labeled goat anti-rabbit or anti-mouse secondary antibodies (Bio-Rad, Hercules, CA; 1:10,000). The chemiluminescence signals (Pierce) were detected by a UVP Bioimaging System (Upland, CA).
In vivo pkd2(WS25/−) treatment regimens.
Male pkd2(WS25/−) mice were treated in vivo up to 12 mo of age with SU-5416, a lipophilic, long-acting VEGF receptor inhibitor (2). All animals were raised to 4 mo of age without treatment. This avoided potential developmental complications from SU-5416 that can occur in rodents and, more importantly, allowed the process of cystogenesis to occur unabated. Beginning at 4 mo of age, mice were split into two groups, Control and SU-5416, that were treated by subcutaneous injection (300 μl) at 2-wk intervals. The vehicle consisted of 0.5% carboxymethylcellulose, 0.9% NaCl, 0.4% polysorbate 80, and 0.9% benzoyl alcohol. The experimental group received 0.75 mg of SU-5416 in 300 μl of vehicle. Mice were treated and studied at 8 mo (4 mo of treatment) and 12 mo (8 mo of treatment) of age. At the end of the study, animals were euthanized and weighed, liver and kidneys were weighed, and sections of liver and kidney were frozen in liquid nitrogen, fixed in formaldehyde (3% formaldehyde in phosphate-buffered saline; pH 7.4), or fixed in glutaraldehyde (1% glutaraldehyde, 2% sucrose, 100 mM cacodylate; pH 7.4).
Determination of liver and kidney cyst volume densities.
Cyst volume densities were quantified in livers and kidneys. After fixation, slabs were cut from the middle of the main liver lobe and middle of one kidney, sequentially equilibrated into 5%, 10%, and 25% sucrose, immersed in optimal cutting temperature compound (OCT, Tissue Tek; Torrence, CA), frozen in liquid nitrogen, and sectioned (10 μm). Sections were then stained (1% toluidine blue, 1% borax, 30% EtOH) and digitally imaged (UVP Bioimaging System). With NIH Image, the areas of the observable cysts within the tissue as well as the total area of the tissue section were obtained. The values were expressed as cumulative cyst volume densities as a percentage of the total tissue area.
All values are reported as means ± SE. Statistical analyses of the differences in the means of experimental pairs were evaluated by t-test. Studies with three or more groups were evaluated by an analysis of variance (ANOVA). The ANOVA included a Tukey-Kramer multiple-comparison test. P values <0.05 were considered statistically significant.
Liver cyst fluids accumulate VEGF.
To qualitatively identify factors present in mouse liver cyst fluids that might promote liver cyst growth, liver cyst fluids from pkd2(WS25/−) mice (n = 2) were evaluated on cytokine array blots (Fig. 1A). Compared with human liver cyst fluids (24), mouse liver cyst fluids had some distinct factors and some factors in common. For example, the mouse fluids contained monocyte inflammatory protein 1α and -2, while the arrays did not detect these factors in human liver cyst fluids. Both mouse and human fluids contained IL-6. Mice do not express IL-8 but do express KC, a functional IL-8 analog that binds and activates the same receptors as IL-8 and induces similar phenotypic responses. Interestingly, the mouse liver cyst fluid mirrored the expression of IL-8 in human liver cyst fluid by expressing KC in all samples tested. Finally, both human and mouse liver cyst fluids contained VEGF.
Many of the factors found in liver cyst fluid could potentially contribute to the growth of ADPKD liver cysts. VEGF is well known for its capacity to promote transcellular permeability, chemotaxis, cell division, and cell survival. The present study evaluated the contribution of VEGF and the VEGF signaling pathway in promoting proliferation of liver cyst epithelial cells and growth of ADPKD liver cysts. The concentrations of VEGF in human and mouse ADPKD liver cyst fluids were first evaluated and compared with VEGF concentrations measured in renal cyst fluids. Human and mouse cyst fluids had comparable VEGF concentrations (Fig. 1B). In human samples, VEGF concentrations were higher in liver (1,128 ± 78 pg/ml; n = 61) versus renal (294 ± 41 pg/ml; n = 26) cyst fluids. Likewise, VEGF concentrations in mouse samples were higher in liver (2,787 ± 136 pg/ml; n = 64) versus renal (191 ± 90 pg/ml; n = 9) cyst fluids. The higher levels of VEGF in liver cyst fluids versus kidney fluids suggest that VEGF signaling may be more functionally significant in liver cysts.
VEGF is secreted both apically and basally.
The presence of VEGF in liver cyst fluid indicates that VEGF is secreted across the apical membrane. To assess the polarity of VEGF secretion, confluent monolayers of pkd2(WS25/−) liver cyst epithelial cells (transepithelial resistance > 1,000 Ω·cm2) were grown on semipermeable supports, and the concentration of VEGF was measured in media within the apical and basal chambers (Fig. 2). After 72 h the basal media contained 5,069 ± 134 pg/ml VEGF (n = 6), while the apical media contained 507 ± 11 pg/ml (n = 6). Parenthetically, similar results were observed in primary cultures of liver cyst epithelial cells isolated from BALB/c cpk(+/−) mice. After 72 h, the level of VEGF in the basal media (1,902 ± 176 pg/ml; n = 3) was approximately five times greater than the level found in the apical media (428 ± 34 pg/ml; n = 3). pkd2(WS25/−) and cpk(+/−) mice both develop liver cysts but do so as a result of mutations in distinct genes. This suggests that the synthesis and polarity of VEGF release can occur independent of the primary cystic mutation in mouse liver cyst epithelial cells. Together, these observations indicate that VEGF can be secreted across both membranes but is secreted more robustly across the basolateral membrane. While the present study focuses on VEGF in liver cyst fluid and the potential for autocrine/paracrine signaling of the cyst-lining epithelial cells by VEGF in the apical media, the accumulation of VEGF in the basal media indicates that VEGF signaling of liver cyst epithelial cells along the basal membrane as well as signaling of the underlying vasculature may contribute profoundly to any in vivo effect of VEGF.
Liver cyst epithelial cells express VEGF receptors.
For VEGF in the liver cyst fluid to induce proliferation of the liver cyst-lining epithelial cells, VEGF receptors must be present on the cells to bind VEGF and transduce the VEGF signal to the cell interior. There are three characterized VEGF receptors. VEGFR1 and VEGFR2 are the more broadly expressed and well defined of these receptors. Western blotting showed that both VEGFR1 and VEGFR2 are expressed in the kidneys and livers of both pkd2(+/+) and pkd2(WS25/−) mice (Fig. 3A; n = 1). In the ADPKD liver, cysts emerge from the epithelial cells lining the intrahepatic bile duct. Termed cholangiocytes, these cells make up only 2–4% of the liver cell mass. To determine whether VEGFR1 and VEGFR2 are expressed in “normal” cholangiocytes, normal bile ducts were enriched from the liver and evaluated by Western blotting. Similarly, liver cyst epithelial cells were obtained by excising liver cyst wall tissue that had emerged from the body of the liver. To further enrich the “cystic” cholangiocytes, epithelial cells from the superficial cyst walls were cultured. These epithelial cell cultures display a number of molecular, biochemical, and physiological characteristics that demonstrate that they originated from cholangiocytes (5). Compared with whole liver samples (n = 3), Western blotting showed that both normal (n = 3) and cystic (cyst wall: n = 4; pkd2 cells: n = 3) cholangiocytes expressed both VEGFR1 and VEGFR2 (Fig. 3B).
VEGF induces proliferation of liver cyst epithelial cells.
Given the presence of VEGF in liver cyst fluids and VEGF receptors in liver cyst epithelial cells, the capacity of VEGF to induce cell proliferation was measured in pkd2(WS25/−) liver cyst epithelial cells (Fig. 4A). Compared with the proliferative response derived from 10% FBS, an insignificant effect was measured with 50 pg/ml VEGF (14 ± 10%; n = 4), but significant proliferative responses were measured with 500 (61 ± 15%; n = 4), 5,000 (108 ± 4%; n = 4), and 50,000 (93 ± 13%; n = 4) pg/ml VEGF. In a parallel study, Mz-ChA1 cells, a human cholangiocarcinoma cell line that has successfully served as a model of human cholangiocytes, showed a similar proliferative response to VEGF (data not shown). Consequently, the VEGF concentrations present in liver cyst fluid of pkd2(WS25/−) mice (mLCF; 2,787 ± 136 pg/ml) would likely be sufficient to promote proliferation of liver cyst epithelial cells.
To evaluate whether liver cyst fluid can induce proliferation of liver cyst epithelial cells, pkd2 cells were treated with serum-free medium containing 10% mLCF. The contribution of VEGF signaling to this proliferative response was also evaluated by pretreating paired sets of pkd2 cells with 10 μM SU-5416 (Fig. 4B). Cultures treated with mLCF showed a significant increase in cell proliferation (84 ± 3% of FBS response; n = 8, P < 0.05). Interestingly, in paired pkd2 cells receiving aliquots of the same mLCF, pretreatment with SU-5416 significantly reduced the proliferative response to mLCF (37 ± 3% of FBS response; n = 8, P < 0.05).
Long-term SU-5416 treatment does not affect weight gain.
The above-described biochemical and cell culture studies indicate that the VEGF signaling pathway is positioned to contribute significantly to the proliferation of ADPKD liver cyst epithelial cells and growth of ADPKD liver cysts. To test this directly, pkd2(WS25/−) mice were allowed to grow to 4 mo of age and then were treated with SU-5416, a VEGF receptor inhibitor, for an additional 4 and 8 mo. All pkd2(WS25/−) mice have discernible kidney and liver cysts by 4 mo of age (35). Because the mice were raised to 4 mo of age before treatment, the process of cystogenesis had been initiated and the study focused on the contributions of VEGF signaling on the growth of cysts. SU-5416 has some deleterious effects on lung development in juvenile rodents but is well tolerated in adult rodents (33, 32). The potential for detrimental effects of SU-5416 on the mice was screened for by following the rates of increase in body weight in mice treated with SU-5416 versus vehicle (Fig. 5). Compared with 12-mo-old vehicle-treated controls (44.3 ± 2.0 g; n = 5), there was no significant difference in the body weight of mice treated with SU-5416 (42.0 ± 4.7 g; n = 6). Parenthetically, the difference in the absolute values for the weights of the SU-5416-treated mice versus the vehicle-treated controls (2.3 g) may be influenced by the difference in the liver weights between the two groups at 12 mo of age (6.0 g; see below).
SU-5416 inhibits growth of liver cysts.
The effect of SU-5416 treatment on liver cyst growth was evaluated by measuring the cystic area of the liver at 8 mo of age (4 mo of treatment) and liver weights at 8 and 12 mo of age (4 and 8 mo of treatment). In cross sections of the primary liver lobe from 8-mo-old mice, sections from pkd2(+/+) mice (n = 4) showed that the liver parenchyma was homogeneous, with no evidence of liver cysts (Fig. 6A). In vehicle-treated pkd2(WS25/−) mice, however, the body of the liver sections contained a number of cystic areas of varying sizes. The cystic areas comprised 9.9 ± 4.3% (n = 4) of the total parenchymal area (Fig. 6B). In SU-5416-treated animals small putative cysts were observed, but there was an absence of large cysts in the sections examined. Quantification showed that the cystic areas in the sections from SU-5416-treated animals (1.8 ± 0.7%; n = 3) were significantly smaller than the vehicle-treated group.
The liver weights in the 8-mo-old vehicle-treated animals (8.7 ± 2.3% of body wt; n = 4) were also significantly greater than the livers from age-matched pkd2(+/+) animals (4.6 ± 0.1% of body wt; n = 9). These elevated values resulted primarily from the expansion of a few cysts from the surface of the livers, while much of the liver parenchyma did not appear grossly affected. Interestingly, the liver weights of the paired 8-mo-old pkd2(WS25/−) mice treated for 4 mo with SU-5416 were not significantly increased (5.1 ± 0.4% of body wt; n = 3) compared with the livers from pkd2(+/+) mice. Furthermore, liver weights from the 8-mo-old SU-5416-treated mice were significantly lower than those from the vehicle-treated mice. This protective effect of SU-5416 against liver weight gain persisted at 12 mo of age (Fig. 7). Livers from vehicle-treated pkd2(WS25/−) mice had significant cystic involvement including the presence of large superficial cysts extending from the lobes of the liver (Fig. 7A, top). Livers from SU-5416-treated pkd2(WS25/−) mice also had liver cysts present at the surface of the liver, but these cysts were far more modest in size (Fig. 7A, bottom). This was reflected in their liver weights (Fig. 7B). At 12 mo of age, livers from vehicle-treated pkd2(WS25/−) mice (10.8 ± 1.9% of body wt; n = 5) were significantly enlarged compared with pkd2(+/+) mice (4.8 ± 0.2% of body wt; n = 12), while livers from SU-5416-treated pkd2(WS25/−) mice were not significantly different (4.8 ± 0.4% of body wt; n = 6) from livers from wild-type animals. Furthermore, treatment of pkd2(WS25/−) mice with SU-5416 resulted in a significant reduction in liver weights compared with vehicle-treated pkd2(WS25/−) mice. In short, SU-5416 treatment blocked the growth of ADPKD liver cysts.
Effects of SU-5416 on renal cysts are less pronounced.
While the protective effects of SU-5416 against liver cyst growth were profound, the effects of SU-5416 on renal cyst growth are less clear. Qualitatively, the kidneys from SU-5416-treated mice had cyst involvement that was not apparently different from that in vehicle-treated mice. This included kidneys that were grossly cystic despite receiving SU-5416 (Fig. 8A). Efforts to quantify the cystic involvement did not provide a robust analysis. Cyst areas in the analyzed slabs were markedly variable in kidneys from pkd2(WS25/−) mice and were not different between vehicle-treated and SU-5416-treated groups. This was likely influenced by two features. First, there was significant localized variability in cysts within the bodies of the kidneys, and the consistent sampling of sections taken from the middle of the kidneys could have relatively normal kidney areas analyzed despite the presence of cysts elsewhere in the body of the kidneys. Second, unlike the massive enlargement of pkd2(WS25/−) livers, kidneys from vehicle-treated pkd2(WS25/−) mice (1.5 ± 0.1% of body wt; n = 5) were not grossly enlarged so that their weights were not significantly different than kidneys from pkd2(+/+) mice (1.3 ± 0.1% of body wt, n = 12; Fig. 8B). Kidney weights from pkd2(WS25/−) mice treated with SU-5416 for 8 mo (i.e., 12 mo of age) were 2.3 ± 1.1% of body weight (n = 6). Consequently, the methodology to quantitatively evaluate the effects of SU-5416 treatment on renal cyst growth was not performed. Current evidence in the literature is similarly mixed regarding the contribution of VEGF signaling to the growth of renal cysts. Treatment of developing CD-1 mice with antibodies against VEGFR2 results in the development of renal cysts, suggesting that inhibition of VEGF signaling could promote renal cyst growth (22). Conversely, rapamycin treatment of Han:SPRD rats, a model for ADPKD, retarded the growth of renal cysts (31, 34). Downstream effects of rapamycin included both inhibiting the release of VEGF and blunting the sensitivity of cells to VEGF signaling (13). These studies indicate that inhibition of VEGF signaling is a potential therapeutic strategy for the treatment of ADPKD renal cyst disease.
Clinical severity of ADPKD is impacted by extragenetic factors.
While the genesis of cysts in ADPKD is linked to mutations in either PKD1 or PKD2, the clinical presentation of ADPKD is markedly heterogeneous (16, 23, 21) and modifier genes and extragenetic factors have significant impact on cyst progression and disease severity. For example, a comparative analysis of the age of onset in end-stage renal disease showed a larger intraclass correlation in monozygotic twins versus sibships (26). Furthermore, quantitative genetic analysis of specific phenotypic traits in PKD1 families showed that inherited differences in modifier genes accounted for a substantial portion of the variability in the phenotypic expression of ADPKD (9). One potential source of extragenetic influence is the presence of signaling molecules, including growth factors and cytokines, in the microenvironment in and around the cyst wall.
Epithelial cells that line ADPKD kidney and liver cysts differentially express a number of genes, including upregulating the expression of genes for specific growth factors and cytokines (17). ADPKD kidney and liver cyst fluids contain a number of specific cytokines and growth factors, and treatment of cultured renal epithelial cells with ADPKD renal cyst fluid promotes cell proliferation (24, 36). In ADPKD renal cysts, cyst fluids have been shown to contain appreciable levels of EGF, EGF receptors redistribute to the apical membrane of cyst lining epithelia, and EGF receptor signaling promotes the growth of renal cysts (6). The role of specific factors in promoting liver cyst growth is less clear but is of particular interest since liver function persists even in grossly cystic livers. The clinical manifestations of ADPKD liver cyst disease stem from the compression of surrounding tissues, organs, and vasculature by the enlarged livers. Like the human counterparts (24), liver cyst fluids from pkd2(WS25/−) mice accumulate specific cytokines and growth factors including VEGF (Fig. 1). VEGF and other progrowth factors within the cyst fluids represent attractive targets for therapeutic intervention to block the growth of ADPKD livers.
It should be noted that the comparative level of VEGF expression by normal versus cystic cholangiocytes was not determined. The VEGF that accumulated in the liver cyst fluid could have arisen simply from constitutive levels of VEGF expression. VEGF expression is, however, induced by a variety of mechanisms. The regulation of VEGF expression could potentially be directly influenced by the primary mutation or driven by extragenetic stresses incurred by the cells. For example, polycystin-1 forms protein complexes with proteins in the mammalian target of rapamycin (mTOR) pathway, and loss of polycystin-1 could contribute to activation of the mTOR pathway (30). Inhibition of the mTOR pathway by rapamycin can inhibit angiogenesis and growth of tumors and does so, in part, through decreasing VEGF production (13). Interest in the potential contributions of the mTOR-VEGF axis in impacting ADPKD is further heightened by reports in Han:SPRD rats, a non-polycystin genetic rat model of kidney cyst disease, that demonstrate a protective affect of rapamycin in renal cystogenesis and enlargement (31, 34). Elevated VEGF expression may also be independent of the primary mutation leading to cyst formation. VEGF expression is induced by various cell stressors, and such stressors, including hypoxia and cell stretching, may be experienced by the epithelial cells that line the liver cysts (5, 24). The present studies utilized the pkd2(WS25/−) mice to test the hypothesis that the VEGF that is released by the liver cyst epithelial cells promotes the growth of ADPKD liver cysts.
VEGF can promote cholangiocyte proliferation.
Cholangiocytes are epithelial cells that line the intrahepatic bile duct and, when appropriately mutated, give rise to liver cysts. Independent of any specific gene mutation, cholangiocytes retain a robust capacity to undergo net proliferation. For example, cholangiocytes normally comprise 2–4% of the liver cell mass. Two weeks after ligation of the bile duct, cholangiocytes expand in numbers to comprise ∼30% of the liver cell mass (1). For the lining epithelium, the formation of enclosed ADPKD liver cysts has a number of similarities to the bile duct network upstream of the point of duct ligation. In a study of VEGF signaling in proliferation of cholangiocytes from normal and bile duct-ligated rat livers, cholangiocytes expressed both VEGF and VEGF receptors (11). In these rat studies, immunofluorescence showed the presence of VEGFR2 and VEGFR3. VEGFR1 was not detected. Expression of VEGF and the VEGF receptors was upregulated in cholangiocytes from bile duct-ligated rat livers, and in vitro proliferation assays showed that the cholangiocytes had a proliferative response to VEGF (11). In humans, VEGF and other angiogenic factors accumulate to appreciable levels in human ADPKD liver cyst fluids (24), and immunohistology studies showed that VEGFR1 and VEGFR2 were expressed in cholangiocytes from normal livers as well as livers with ADPKD and Caroli disease (8). In mice, Western blotting showed the presence of VEGFR1 and VEGFR2 (Fig. 3), and administration of VEGF to liver cyst epithelial cell cultures promotes cell proliferation (Fig. 4A; Ref. 8). Importantly, the concentrations of VEGF measured in the liver cyst fluids were well within the range of VEGF concentrations needed to induce cell proliferation. Furthermore, treatment of pkd2 cells with mLCF induced a robust proliferative response (Fig. 4B). This proliferative response was inhibited by blocking VEGF receptor activity (Fig. 4B).
SU-5416 suppresses liver cyst growth.
The in vitro studies indicate that the ADPKD liver cyst epithelial cells can synthesize and release VEGF, inducing proliferation of the liver cyst epithelial cells. In vivo studies using SU-5416 were performed to directly assess the significance of the VEGF signaling pathway in promoting ADPKD liver cyst growth. SU-5416 is a potent antiangiogenic factor whose main activities include inhibition of VEGFR2 (Flk-1/KDR) and VEGFR1 (Flt-1) (10, 18). Failure to significantly inhibit the tyrosine kinase activities of EGF and FGF receptors demonstrates a relatively high specificity of SU-5416 for VEGF receptors (10). SU-5416 has been shown to affect PDGF receptor autophosphorylation, but it remains unclear whether SU-5416 impacts the activity of PDGF receptors in intact cells and tissues (10, 29). In addition to the hormone receptor effects, SU-5416 does potentially have a number of additional effects including inhibiting the expression of VEGF, hypoxia-inducible factor-1α, and cyclooxygenase-2 (20, 28, 37). The contribution of these additional effects of SU-5416 was not specifically evaluated in the present in vivo studies. As a sum of these effects, studies on tumor vascularization and development of emphysema demonstrate directly that SU-5416 inhibits angiogenesis and vascular maintenance (10, 15, 19, 29).
The pkd2(WS25/−) mouse model was well suited for the study. First, the mutated pkd2(WS25/−) mouse gene is directly homologous to human PKD2, one of two genes linked to ADPKD in humans. Second, the liver cysts undergo significant and consistent enlargement by 12 mo of life. This gross enlargement parallels the primary clinical determinant in human ADPKD liver cyst disease and permits the effects of SU-5416 treatment on this key parameter to be assessed directly by measuring liver weights. SU-5416 treatment was initiated at 4 mo of age and continued to either 8 mo (4 mo of treatment) or 12 mo (8 mo of treatment) of age. The initial 4-mo period without treatment permitted the initial development of cysts in the kidneys and livers to occur (35) and, consequently, allowed the study to focus on the relative growth of cysts. Importantly, at 8 mo of age, the cystic area of the liver in pkd2(WS25/−) mice was markedly reduced in animals treated with SU-5416 (Fig. 6). This early effect persisted at 12 mo of age, when liver weights of untreated pkd2(WS25/−) mice were twice as large as those of pkd2(WS25/−) mice that were treated with SU-5416 (Fig. 7). The livers of SU-5416-treated pkd2(WS25/−) mice did have cysts present at the surface, but their sizes were comparatively small and there was not a significant increase in liver weights compared with the livers of pkd2(+/+) mice. These studies demonstrate that the VEGF signaling pathway contributes in a pronounced way to the growth of ADPKD liver cysts in mice and supports further investigation into the contribution of VEGF signaling in the growth of mouse and human ADPKD liver cysts.
VEGF signaling may affect both cyst epithelium and endothelium.
The primary focus of the present study was on autocrine/paracrine signaling by VEGF on the cyst-lining epithelial cells. Indeed, VEGF was present at relatively high concentrations in both human and mouse liver cyst fluids (Fig. 1), VEGF receptors were expressed in the cyst-lining epithelium (Fig. 3), and primary liver cyst epithelial cells had increased rates of proliferation when treated with VEGF (Fig. 4). VEGF, however, is best known for its effects on the vasculature and capacity to induce angiogenesis. Angiogenesis occurs in the walls of human ADPKD renal cysts and is likely influenced by VEGF signaling (3). The present study did not characterize the potential effect of SU-5416 treatment on blocking angiogenesis and vascularization of the liver cyst walls but did show that VEGF is likely secreted across the basolateral membrane of the cyst epithelium into the interstitial tissue of the liver cyst wall (Fig. 2). Compared with the apical release, there was substantially greater release of VEGF across the basolateral membrane. This held for liver cyst epithelial cells from both cpk(+/−) and pkd2(WS25/−) mice. Both mouse lines develop liver cysts, but the cysts arise as a consequence of mutations in distinct genes. In cpk(+/−) mice, cys-1, the gene encoding for the protein cystin, is mutated. In pkd2(WS25/−) mice, pkd2, the gene encoding for polycystin-2, is mutated. The parallel results from these two mouse lines suggest that the release of VEGF may not be primarily modulated by the primary gene mutation. Future studies will determine the effect of VEGF on angiogenesis and vascularization of the liver cyst wall.
Therapeutic potential of inhibition of VEGF signaling pathway.
In summary, the present mouse study demonstrates that VEGF is present in appreciable amounts in human and mouse ADPKD liver cyst fluids (Fig. 1), that cyst-lining epithelial cells express VEGF receptors (Fig. 3), and that addition of VEGF at concentrations found in liver cyst fluid induces proliferation of the ADPKD liver cyst-lining epithelial cells in vitro (Fig. 4). Importantly, these in vitro observations were extended in vivo and showed that inhibition of the VEGF signaling pathway significantly blunts the growth of ADPKD liver cysts (Figs. 6 and 7). If inhibiting the VEGF signaling pathway in humans were similarly successful, the VEGF signaling pathway could prove to be an important therapeutic target in the treatment of human ADPKD liver cyst disease.
The present studies were supported by National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Grant DK-34039 and PKD Foundation Grant 109a2r to R. B. Doctor and NIDDK Grant DK-68581 to V. H. Gattone.
Pkd2(WS25/−) mice used in this study were developed and graciously provided by Stefan Somlo (Yale University).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society