Parathyroid hormone-related protein (PTHrP) interacts with vascular endothelial growth factor (VEGF) in osteoblasts. Since both PTHrP and VEGF have both proinflammatory and profibrogenic features, we assessed here whether these factors might act in concert to promote fibrogenesis in the obstructed kidney. VEGF receptor (VEGFR)-1 was upregulated, while VEGFR-2 was downregulated (at both mRNA and protein levels) in the mouse kidney within 2–6 days after ureteral obstruction. VEGF protein levels also increased in the obstructed kidney at the latter time. Moreover, this VEGF and VEGFR-1 upregulation was higher in mice overexpressing PTHrP in the proximal tubule than in control littermates. These changes were associated with higher fibronectin mRNA expression and α-smooth muscle actin (α-SMA) and integrin-linked kinase (ILK) immunostaining and lower apoptotic tubulointerstitial cells in the mouse obstructed kidney than in control littermates. Pretreatment with a neutralizing anti-VEGF antibody reversed these responses in the obstructed kidney of both types of mice. In vitro, PTHrP-(1-36) increased (maximal 2-fold vs. basal, at 100 nM) α-SMA and ILK protein expression and decreased E-cadherin protein levels in renal tubuloepithelial mouse cortical tubule and normal rat kidney (NRK) 52E cells. PTHrP-(1-36) also decreased cyclosporine A- and/or osmotic stress-induced apoptosis in these cells and in renal fibroblastic NRK 49F cells. These effects elicited by PTHrP-(1-36) were associated with both VEGF and VEGFR-1 upregulation, and abolished by the anti-VEGF antibody. Collectively, these findings strongly suggest that VEGF acts as an important mediator of PTHrP to promote fibrogenesis in the obstructed kidney.
- growth factors
- obstructive nephropathy
- epithelial-mesenchymal transition
tubulointerstitial fibrosis is a paramount event in the progression of chronic renal pathologies. The fibrogenic process is characterized by tubular atrophy and accumulation of extracellular matrix proteins, including fibronectin, types I, III, IV, and V collagens, and laminin (3, 32). Increasing experimental evidence suggests that the influx of macrophages into the renal interstitium is a key step in the development of chronic inflammation and the subsequent interstitial fibrosis in the damaged kidney (3, 47). Fibrogenesis is also accompanied by an increase in proliferation of renal interstitial cells and their activation to myofibroblasts, the main effector cells in this setting (23, 34, 39). In the damaged kidney, the latter cells can be formed from resident fibroblasts and from circulating bone marrow-derived cells, as well as through epithelial-mesenchymal transition (EMT), a process whereby tubuloepithelial cells become matrix-producing myofibroblasts (17, 27). EMT involves several phenotypic changes in tubuloepithelial cells, including actin cytoskeleton reorganization and dissociation of intercellular unions, and is modulated by various cytokines and growth factors (8, 10, 34, 45, 50).
Parathyroid hormone-related protein (PTHrP) is abundant throughout the kidney (5). This factor exhibits growth-modulatory and proinflammatory properties in a variety of cell systems, including tubuloepithelial cells (13, 35, 38). PTHrP is overexpressed in various tubulointerstitial nephropathies, and its overexpression correlates with the development of proteinuria in both diabetic mice and rats with tubulointerstitial damage after protein overload (16, 21). Moreover, a recent study in mice with folic acid-induced nephrotoxicity suggests an important role for PTHrP to promote renal fibrogenesis (35).
Vascular endothelial growth factor (VEGF) is an ubiquitous factor responsible for the integrity of endothelial and capillary networks (33, 46, 49). VEGF and its receptors, VEGF receptor (VEGFR)-1 and VEGFR-2, are present throughout the kidney (40). VEGF exerts a prosurvival action for renal cells (1, 48); however, it might also contribute to renal tubulointerstitial injury through either inducing the synthesis of proinflammatory factors (growth factors and cytokines, and matrix adhesion molecules) or acting as a mediator of some of these factors (19, 30, 40, 46). On the other hand, in the remnant kidney model, progressive glomerulosclerosis and tubulointerstitial fibrosis are accompanied by loss of VEGF expression (41). Moreover, VEGF administration has proven to promote glomerular capillary repair in a rat model of severe glomerulonephritis (44). Thus the putative role of VEGF as a renal profibrogenic factor remains controversial.
PTHrP stimulates VEGF gene expression in a rapid and transient fashion in osteoblasts (9). Recent studies also suggest that the VEGF system might act as a mediator of at least some of the actions of PTHrP in these cells (7). Moreover, a previous report has shown that age-related changes in PTHrP occur, associated with those of VEGF in human osteoblastic cells (29). However, a possible relationship between PTHrP and VEGF in the injured kidney has not yet been examined. In the present study, we assessed the putative interaction between these factors in the kidney with ureteral obstruction in mice. PTHrP upregulation has recently been shown to have a key role in the pathogenesis of inflammation in this well-characterized animal model (38). We carried out both in vivo and in vitro studies to ascertain whether VEGF might be involved in the mechanisms whereby PTHrP would induce renal fibrosis in this setting.
MATERIALS AND METHODS
PTHrP-overexpressing transgenic mice.
Two types of mice with or without renal PTHrP overexpression were used. Targeted overexpression of PTHrP to the mouse renal proximal tubule has been reported in detail elsewhere (12). In brief, mice containing a human PTHrP cDNA under the control of a tetracycline operator were bred to mice bearing a construct consisting of a γ-glutamyl transpeptidase promoter fragment upstream of a tetracycline transactivator fusion protein to generate PTHrP-overexpressing transgenic (PTHrP-TG) animals. The renal proximal convoluted tubule specificity was conferred by the γ-glutamyl transpeptidase-I promoter, which is expressed almost exclusively in this nephron segment of fetal and adult kidney (12). Transgene-bearing founders were continuously outbred to normal CD-1 mice to generate hemizygotes. Genotyping of these mice was carried out by tail DNA PCR. Results obtained with PTHrP-TG mice were compared with those in their control littermates (those bearing either one of the aforementioned constructs or normal CD-1 mice) (35, 38). PTHrP-TG mice showed no alterations in renal morphology, renal function or calcemia, compared with normal mice; except for a mild hypophosphatemia (12, 35, 38). Circulating PTHrP levels in these mice were undetectable (12). All procedures and experimental protocols were approved by the Institutional Animal Care and Use Committee of the Fundación Jiménez Díaz.
Experimental model of unilateral ureteral obstruction.
Unilateral ureteral obstruction (UUO) was performed under anesthesia in sex-unselected mice, weighing 25–35 g, by ligating the left ureter of each animal with 3-0 silk, at two locations and cutting between the ligatures, through an abdominal incision, as described (38). On different days after UUO, groups of mice were killed and the obstructed kidneys were collected. The kidneys from sham-operated mice which had their ureters manipulated but not ligated served as unobstructed controls. From each mouse, kidney portions were separated for different measurements. One portion was fixed in 4% p-formaldehyde for histological studies, and the other one was used for total RNA and protein isolation. A group of mice was treated daily with a polyclonal anti-mouse VEGF antibody (67 μg/kg ip; BD Biosciences, Franklin Lakes, NJ), starting 1 wk before UUO and for 5 days thereafter. A dose of this antibody in the range of 0.05–0.15 μg/ml was shown to neutralize 50% of the mitogenic effect of mouse VEGF, at 10 ng/ml, in human umbilical vein endothelial cells, in vitro (6). In addition, the selected dose of this antibody in our in vivo protocol has proven to be efficient to block various VEGF effects in the mouse kidney (1).
Cell total RNA was isolated with TRIzol (Invitrogen Groningen, The Netherlands), and gene expression was analyzed by real-time PCR according to a previously described protocol (7, 35, 38). Predeveloped fluorogenic mouse-specific primers for VEGF (Mm00437304_m1), VEGFR1 (Mm00438980_m1), VEGFR2 (Mm00440099_m1), and fibronectin (Mm01256734_m1), and TaqMan MGB probes were obtained by Assay-by-Design (Applied Biosystems, Foster City, CA). cDNA was synthesized using avian myeloblastosis virus reverse transcriptase (Promega, Madison, WI) with random hexamer primers, and real-time PCR was carried out in an ABI PRISM 7500 system (Applied Biosystems), following the manufacturer's instructions. 18S rRNA served as a housekeeping gene and was amplified in parallel with tested genes. The mRNA copy numbers were calculated for each sample by using the Ct (arithmetic fit point analysis for the light cycler) value, and normalized against 18S rRNA (7, 35, 38).
Histology and immunohistochemistry.
Fixed renal tissue sections were dehydrated by graded ethanols and xylene and embedded in paraffin. Histological analysis was routinely performed on serial paraffin-embedded renal tissue sections (2 μm) within the same mouse tissue. Immunohistochemistry was performed using a previously described protocol (35, 38) and the following antibodies (dilution, -fold): mouse monoclonal anti-α-smooth muscle actin (α-SMA; 150, Sigma, St. Louis, MO); rabbit polyclonal anti-integrin linked kinase (ILK; 200, Santa Cruz Biotechnology, Santa Cruz, CA), anti-VEGFR-1 (200, Santa Cruz Biotechnology), and anti-VEGFR-2 (200, Santa Cruz Biotechnology) antibodies; and a monoclonal antibody to F4/80 antigen in murine monocytes/macrophages (200, Serotec, Oxford, UK). The tissue sections were incubated with 1.5% goat serum in PBS for 1 h at 37°C to reduce nonspecific staining. Following incubation overnight at 4°C with the corresponding primary antibodies, sections were subsequently incubated with either anti-IgG biotinylated-conjugated antibody followed by the avidin-biotin-peroxidase complex (α-SMA and F4/80, Dako, Glostrup, Denmark) or a polymer-peroxidase complex (ILK, VEGFR-1, and VEGFR-2, Envision+ System, Dako), and 3,3′-diaminobenzidine as chromogen. The sections were counterstained with hematoxylin. Some tissue samples were incubated without the primary antibody, as negative controls. All stainings were evaluated in at least eight different high-power fields/section in four sections from each experimental mouse. The percentage of α-SMA-, ILK-, VEGFR-1-, and VEGFR-2-stained area was estimated by the following semiquantitative score: 0, no staining; 1, up to 25%; 2, between 25 and 50%; 3, between 50 and 75%; and 4, >75%. The number of infiltrating macrophages was evaluated by counting F4/80-positive interstitial cells per field. Tubulointerstitial apoptotic cell death was assessed by enzymatic in situ labeling of DNA strand breaks using the terminal deoxynucleotidyltransferase-mediated dUTP nick-end labeling (TUNEL) method (In Situ Cell Death Detection Kit, Promega) in mouse kidney samples, as described (35). The number of TUNEL-positive tubulointerstitial cells was evaluated by counting the number of stained cell nuclei per field (×400) in the tubules and the interstitium. All evaluations were performed by two independent observers in a blinded fashion, and the corresponding mean score value was obtained for each mouse.
Cell culture studies.
Mouse cortical tubule (MCT) cells as well as normal rat kidney tubuloepithelial cells (NRK-52E; ATCC CRL 1571) and renal fibroblastic NRK-49F (ATCC CRL 1570) cells, all of which respond to PTHrP, were grown in RPMI 1640 (MCT) or DMEM (NRK-52E and NRK-49F) with 10% FBS and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin) in 5% CO2 at 37°C, as described (35, 38). For all of the experiments described below, subconfluent cells (60.000 cells/cm2) were incubated with either various PTHrP-(1-36) concentrations (0.1–100 nM), placental growth factor (PlGF), or VEGF164 (R&D Systems, Minneapolis, MN), both at 20 ng/ml, in the presence of 2% FBS for different time periods (24–48 h). In some experiments, the aforementioned anti-VEGF antibody, at 1 μg/ml, was added to cell cultures 1 h before PTHrP-(1-36). This antibody dose was found to completely neutralize the stimulatory effect of recombinant mouse VEGF on human umbilical vein endothelial cell proliferation (6). Some cell cultures were pretreated with the transcription inhibitor actinomycin D (Sigma), at 5 μM, for 30 min before stimulation with PTHrP (1-36), at 100 nM, for different time periods up to 6 h. Following stimulations, cell extracts in lysis buffer (50 mM Tris·HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 1 mM PMSF, and 0.8 μM aprotinin, pH 7.4) were obtained for protein analysis, and/or total RNA was isolated as described above.
For viability studies, subconfluent cells were incubated as described above for 48 h with 10 μM cyclosporine, a toxic dose for renal tubuloepithelial cells (1), with or without 100 nM PTHrP-(1-36) and/or the neutralizing anti-VEGF antibody, at 1 μg/ml. Nonadherent cells were collected and pooled with adherent cells after gentle trypsinization, followed by incubation in the dark for 1 h at 4°C in 60 μg/ml RNAse A, 50 μg/ml propidium iodide, and 0.05% Nonidet P-40 in PBS. Analysis was then performed in a FACS Calibur cytometer (BD Biosciences) using LYSIS II software to calculate the percentage of hypodiploid cells, corresponding to apoptotic cells (35).
An in vitro experimental model was also used to mimic the effects of renal tubular distention as occurs in UUO (31). MCT cells were first treated with or without PTHrP-(1-36), in the presence or absence of anti-VEGF antibody, in RPMI 1640 medium with 1% FBS as described above. The corresponding culture medium was then removed, and cells were subsequently exposed or not (basal) to two intermittent 5-min periods of incubation with normal (145 mM NaCl, 5 mM KCl) and low-osmolarity (6 mM NaCl, 5 mM KCl, and 50 mM mannitol) solutions. This was followed by incubation for 6 h with the removed culture medium containing the different agonists as mentioned before (37). Cell viability was then assessed by trypan blue exclusion, after collection of nonadherent and adherent cells to calculate the percentage of total dead cells (35).
Western blot analysis.
Protein content in kidney tissue and cell extracts was determined by the Bradford method (Pierce, Rockford, IL), using bovine serum albumin as a standard. Proteins (30–60 μg/lane) were separated on 5–10% polyacrylamide-SDS gels under reducing conditions. After electrophoresis, samples were transferred onto nitrocellulose membranes (Bio-Rad, Hercules, CA), blocked with 5% defatted milk in 50 mM Tris·HCl, pH 7.5, 150 mM NaCl with 0.05% Tween 20, and incubated overnight at 4°C with the following antibodies (dilution, -fold): mouse monoclonal anti-α-SMA(1,000), and anti-E-cadherin (2,000,Sigma); the aforementioned anti-VEGF antibody (1,000, BD Biosciences); and rabbit polyclonal anti-ILK, anti-VEGFR-1 (1,000), and anti-VEGFR-2 (1,000) antibodies (Santa Cruz Biotechnology). As constitutive control, α-tubulin was also detected with a specific monoclonal mouse antibody (Sigma). After extensive washing, the membranes were incubated with peroxidase-conjugated rabbit anti-mouse or goat anti-rabbit IgG in each case and developed by ECL chemiluminiscence (GE Healthcare, Buckinghamshire, UK). Densitometric values of fluorogram bands were normalized to those of the corresponding α-tubulin band.
All results are expressed as means ± SE. The effect of UUO on different factors in vivo and both time course and dose responses of PTHrP-(1-36) in vitro were evaluated by nonparametric variance analysis (Kruskal-Wallis) followed by Dunn's test. A Mann-Whitney test was performed to analyze the differences between PTHrP-TG mice and their control littermates and the effect of the anti-VEGF antibody both in vivo and in vitro. P < 0.05 was considered significant.
Changes in VEGF and VEGF receptors related to PTHrP overexpression in the obstructed mouse kidney.
In the sham-obstructed kidney, we observed similar VEGF and VEGFR-2 expression, at both mRNA (by real-time PCR) and protein (by Western blotting and immunohistochemistry, respectively) levels, in both control and PTHrP-TG mice (Fig. 1, A, C, D, and F). In contrast, VEGFR-1 gene expression as well as VEGFR-1 immunostaining were higher in this kidney from the latter mice than in that from control littermates (Fig. 1, B and E). Following UUO, there was a significant decrease in VEGF mRNA expression (Fig. 1A) and also of VEGFR-2, at both mRNA and protein levels (Fig. 1, C and F), in the obstructed kidney of both types of mice throughout the study. Interestingly, however, renal VEGF protein levels were increased in control and PTHrP-TG mice, and this increase was of a greater magnitude in the latter animals at day 6 following obstruction (Fig. 1D). Renal VEGFR-1 mRNA levels in control mice increased shortly (at day 2) after UUO to values similar to those in the sham-obstructed kidney of PTHrP-TG mice, decreasing, without normalizing, thereafter (Fig. 1B). In contrast, in the latter mice, a sustained upregulation of this receptor expression occurred in the obstructed kidney up to at least day 6 (Fig. 1B). Furthermore, VEGFR-1 immunostaining increased at the latter time following UUO, and this increase was also higher in PTHrP-TG than in control mice (Fig. 1E).
Increased interstitial α-SMA, ILK, and F4/80 immunostaining, fibronectin mRNA expression, and tubulointerstitial cell survival occur associated with PTHrP upregulation in the obstructed mouse kidney and were inhibited by a neutralizing anti-VEGF antibody.
We observed an increase in renal interstitial immunostaining for α-SMA, a myofibroblast marker (10, 27, 45, 50), and ILK, a protein kinase whose overexpression has been shown to occur related to the EMT process (24, 25, 26), in the obstructed mouse kidney at day 6 after UUO, compared with that in sham-operated mice (Fig. 2, A and B). Moreover, this increase was higher in PTHrP-TG mice than in control littermates (Fig. 2, A and B).
To explore the possible cooperation between PTHrP and VEGF to induce these profibrogenic changes in the mouse obstructed kidney, we administered a neutralizing anti-VEGF antibody to both control and PTHrP-TG mice before and during UUO for 6 days. We found that this antibody (but not a nonimmunogenic IgG) significantly decreased the staining score levels for the aforementioned markers in both types of mice but remained over those in sham-operated mice (Fig. 2, A and B). Gene expression of fibronectin was also significantly higher in the sham-obstructed kidneys of PTHrP-TG mice than in control mice, and it increased dramatically after UUO, mainly in the former mice (Fig. 2C). Moreover, as observed for α-SMA and ILK immunostaining, anti-VEGF antibody administration significantly reduced, but did not normalize, fibronectin mRNA levels in both types of animals (Fig. 2C). In addition, an increased number of F4/80-positive cells was observed in the mouse renal interstitium after 6 days of UUO (Fig. 2D). This was accompanied by an increased number of TUNEL-positive tubulointerstitial cells in the obstructed kidney at this time period (Fig. 3, A and B). Furthermore, consistent with previous findings in PTHrP-TG mice following nephrotoxic renal injury (35), a higher number of F4/80-positive cells and lower number of TUNEL-positive cells were observed in the tubulointerstitium of the obstructed kidney of these mice than in control littermates (Figs. 2D and 3, A and B). Administration of the neutralizing anti-VEGF antibody also reversed these changes in both types of mice (Figs. 2D and 3, A and B).
PTHrP-(1-36) increases VEGF and VEGFR-1 while decreasing VEGFR-2 in renal tubuloepithelial cells in vitro.
To assess further whether the VEGF system might act as a modulator of the profibrogenic and prosurvival effects of PTHrP in the obstructed kidney, in the next series of in vitro experiments we used two renal tubuloepithelial cell lines expressing VEGFR-1 and VEGFR-2 (42, 48). First, we found that treatment with 100 nM PTHrP-(1-36) resulted in an increase of VEGF and VEGFR-1 (both mRNA and protein) expression in MCT cells (Fig. 4, A–D). Gene expression of the latter receptor was found to peak at 24 h, associated with an increase in VEGFR1 protein levels, after stimulation with 100 nM PTHrP-(1-36) (Fig. 4, C and D). In contrast, the maximal increase in VEGF mRNA levels induced by this peptide occurred earlier (6 h), decreasing thereafter, when VEGF protein levels peaked (at 24 h) in these cells (Fig. 4, A and B). Treatment with actinomycin D to prevent gene transcription in both control and PTHrP (1-36)-treated cells showed that this PTHrP peptide failed to affect VEGF mRNA stability for up to 6 h in MCT cells (Fig. 4A). A maximal increase in both VEGF and VEGFR-1 protein levels was also found in rat tubuloepithelial NRK 52E cells following PTHrP-(1-36) treatment for 24 h (Fig. 5, A and B). On the other hand, PTHrP-(1-36) decreased VEGFR-2 protein levels in both types of tubuloepithelial cells within this time period (Figs. 4D and 5C).
PTHrP-(1-36) induction of several EMT markers occurs through VEGF in renal tubuloepithelial cells.
We next evaluated the effect of PTHrP-(1-36) on α-SMA, ILK, and E-cadherin, a cell-cell attachment protein which is downregulated during the EMT process (17, 27, 50), in both tubuloepithelial cell lines. We found that this PTHrP peptide dose dependently increased both α-SMA and ILK protein levels, while it decreased E-cadherin protein expression at 48 h in MCT cells (Fig. 6, A–C). PlGF, a specific VEGFR1 agonist (2, 46), or exogenous VEGF164, each at 20 ng/ml, induced similar alterations as PTHrP-(1-36), at 100 nM, on these EMT markers in these cells (Fig. 7, A–C). On the other hand, pretreatment with the aforementioned neutralizing anti-VEGF antibody, but not nonimmune IgG, at 1 μg/ml, was found to inhibit these effects of PTHrP-(1-36) in MCT cells (Fig. 7, A–C) and also those similarly observed in NRK 52E (Fig. 8, A and B).
VEGF can act as a mediator of the prosurvival effects of PTHrP-(1-36) in renal tubuloepithelial and fibroblastic cells.
We also aimed to confirm in vitro that PTHrP can promote renal tubuloepithelial and fibroblastic cell survival through the VEGF system. We found that apoptosis induced by cyclosporine in both tubuloepithelial cell lines and in renal fibroblasts (Fig. 9, A–C), or by osmotic stress in MCT cells (Fig. 9D), was inhibited by 100 nM PTHrP-(1-36). Pretreatment with the neutralizing anti-VEGF antibody abolished these antiapoptotic effects of PTHrP-(1-36) in all these cell lines (Fig. 9, A–D).
In the present study, we examined whether VEGF might act as a mediator of the profibrogenic effects of PTHrP in the damaged kidney, using a well-characterized model of tubulointerstitial disease in mice with ureteral ligation. As recently found to occur for PTHrP (38), the VEGF system was upregulated in the obstructed mouse kidney. Moreover, this upregulation was of greater magnitude, associated with higher macrophage infiltration and expression of various fibrogenesis-related factors, in PTHrP-TG mice than in control mice. Therefore, upregulation of both PTHrP and VEGF seem to be related to profibrogenic changes after kidney obstruction.
In the damaged kidney, the local generation of various vasoconstrictors may lead to renal ischemia, which would contribute to promote both VEGF and VEGFR-1 expression (14, 22). In this regard, renal VEGF protein was upregulated in mice with subtotal nephrectomy, associated with remodeling of peritubular capillaries (36). Our in vivo and in vitro data herein strongly support that PTHrP contributes to overexpression of both VEGF and the VEGFR-1 in the obstructed mouse kidney. Unexpectedly, the observed increase in VEGF protein expression in this setting was not mirrored by similar changes in VEGF mRNA expression in the time frame of our study. However, the present in vitro studies in mouse tubuloepithelial cells showed an induction of VEGF gene by PTHrP-(1-36) which was transient, and it occurred earlier than that of VEGF protein expression. Moreover, our data using the transcription inhibitor actinomycin D suggest that PTHrP-(1-36) does not affect VEGF mRNA stability in these cells. Thus it is possible that VEGF gene overexpression by PTHrP might have occurred before day 2, the earliest time period after UUO which was included in our study. Interestingly, a previous report in a mouse model of nephrotoxic acute renal failure has shown a decrease in VEGF mRNA expression in areas of tissue hypoxia, associated with reduced hypoxia-inducible factor (HIF)-1α, an effect apparently due to increased von Hippel-Lindau protein, inducing HIF-1α proteasomal degradation (51). It is presently unknown whether such a mechanism might explain in part the decreased VEGF mRNA in the mouse obstructed kidney as found herein at days 2–6 after UUO. In addition, our studies cannot rule out that additional translational mechanisms might contribute to the VEGF protein induction by PTHrP-(1-36) as shown herein, as recently suggested to occur for VEGF stimulation by angiotensin II in tubuloepithelial cells (11).
In contrast to the VEGFR-1 overexpression observed in the mouse obstructed kidney, a striking reduction of renal VEGFR-2 gene expression occurred shortly following UUO in mice. This was somewhat unexpected, considering a recent study using small interfering RNA technology to knockdown VEGFR-1 mRNA, which suggests the important role of this receptor to modulate VEGFR-2 expression in endothelial cells (20). Moreover, VEGF has been shown to upregulate VEGFR-2 mRNA expression in these cells, an effect apparently dependent on VEGF binding and subsequent activation of VEGFR-2 tyrosine kinase (43). Thus our present findings support that alternative VEGF-independent mechanisms would be responsible for the observed VEGFR-2 downregulation in the obstructed mouse kidney. At least one of these mechanisms might involve PTHrP, as suggested by our in vitro data. In this regard, however, a putative role of transforming growth factor-β, which is overexpressed in the kidney after UUO and might downregulate VEGFR-2 (18, 28), cannot be ruled out from the present data.
We recently demonstrated that PTHrP overexpression occurs related to inflammation and fibrosis in the acutely damaged mouse kidney (35, 38). Present findings strongly suggest that the deleterious effects induced by PTHrP in this setting might be mediated, at least in part, by the VEGF system. This hypothesis is supported by our results showing that a neutralizing anti-VEGF antibody decreased, but failed to normalize, the expression of several fibrogenic molecules in the obstructed mouse kidney. Moreover, this antibody abrogated the PTHrP (1-36)-induced changes in the protein expression of some of these molecules in renal tubuloepithelial cells in vitro. The different efficiency of this antibody in vivo and in vitro is not surprising, considering the putative differences in the bioavailability of the doses used of this antibody in both experimental conditions. In addition, as recently suggested (38), PTHrP might target cells other than tubuloepithelial cells to promote fibrogenesis in a VEGF-independent manner in the obstructed mouse kidney. Assessing the relative contribution of any of these mechanisms to the profibrogenic actions of PTHrP in this setting awaits further study.
In previous studies from our group, the sham-obstructed kidney of PTHrP-TG mice showed no morphological alterations, but higher expression of various proinflammatory mediators than their control littermates (38). In fact, VEGFR-1 and fibronectin were also found here to be overexpressed in this kidney from PTHrP-TG mice compared with those in control animals. Thus it seems that constitutive PTHrP overexpression in mice might trigger counterregulatory mechanisms, which would prevent kidney damage in the absence of a renal insult. Such compensatory mechanisms might also account for the absence of changes in the renal expression of some of the factors tested here in sham-operated PTHrP-TG mice compared with control mice, in contrast to what would have been expected based on our in vitro results in well-characterized tubuloepithelial cell lines. In addition, the observed changes in several evaluated factors elicited by PTHrP-(1-36) in the latter cells in vitro could have been overshadowed in vivo, considering the interplay of the complex cell interactions within the obstructed kidney. Elucidation of these hypotheses awaits further study. In any event, the present findings further support the notion that PTHrP overexpression might favor the onset of kidney damage, as recently suggested (35, 38).
Consistent with our previous findings in another mouse model of nephrotoxic injury (35), both renal PTHrP overexpression and fibrosis were associated with a decrease in tubulointerstitial cell apoptosis in the obstructed mouse kidney. Moreover, PTHrP-(1-36) was found to protect both tubular and fibroblastic cells from death induced by either the nephrotoxic cyclosporine or osmotic stress. Collectively, the present data and those in two recent reports (35, 38) strongly suggest that the decrease in tubulointerstitial cell apoptosis by PTHrP in the injured kidney might result from its effects on both macrophage infiltration and tubulointerstitial cell survival. Our findings here indicate that these effects of PTHrP appear to occur, at least in part through the VEGF system in the obstructed kidney.
Some known features of VEGF make it a likely candidate as a proinflammatory and profibrogenic factor. Thus VEGF activates nuclear factor-κB and stimulates monocyte chemoattractant protein-1 expression in retinal endothelial cells (30). VEGFR-1 stimulation can induce procoagulant tissue factor production and chemotaxis in monocytes/macrophages (4). Furthermore, a recent study has shown that VEGF inhibition in part prevented both glomerular hypertrophy and proteinuria in the remnant kidney model (41). Our present data are consistent with, and extend further, these previous findings and strongly support the involvement of the VEGF system in the mechanisms of renal injury following UUO. The present in vitro results also call for a VEGFR-1 role in the mechanisms whereby PTHrP induces at least some of its profibrogenic effects through VEGF in renal tubuloepithelial cells. However, our data cannot rule out the involvement of a putative cooperation between VEGFR-1 and VEGFR-2, as described in other nonrenal scenarios (2, 15), in the mechanisms associated with the effects of PTHrP in these cells.
In summary, we show here for the first time the important role of the VEGF system as a mediator of the profibrogenic action of PTHrP in the kidney of mice with ureteral obstruction. Our in vitro findings indicate that this system appears to be critically involved in the mechanisms whereby PTHrP-(1-36) modulates several EMT markers in tubuloepithelial cells and promotes their and fibroblastic cell viability. Present data also suggest that VEGF might be considered as a potential target to modulate at least some of the deleterious actions of PTHrP in the obstructed kidney.
This work was supported by grants from the Spanish Ministerio de Ciencia y Tecnología (SAF2002-04356-C02-02), Ministerio de Educación y Ciencia (SAF2005-05254), Instituto de Salud Carlos III (PI050117 and RETICEF RD06/0013/1002), and Comunidad Autónoma de Madrid (CAM 08.6/0038.1/2000-2 and GR/SAL/0415/2004). J. A. Ardura and R. Berruguete were supported by the Conchita Rábago Foundation and D. Rámila by “Ikerketa eta Prestakuntza Programa” from the Basque Government. J. A. Ardura was the recipient of a travel stipend from the European Calcified Tissue Society to present portions of this study at the 34th European Symposium on Calcified Tissues in Copenhagen (Denmark), May 5–9, 2007. M. V. Alvarez-Arroyo was the recipient of a research contract from the Instituto de Salud Carlos III.
We thank A. F. Stewart and A. García-Ocaña (Dept. of Endocrinology and Metabolism, Univ. of Pittsburgh School of Medicine, Pittsburgh, PA) for generously supplying human PTHrP-(1-36).
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 © 2008 the American Physiological Society