Parathyroid Hormone–Related Protein Promotes Epithelial–Mesenchymal Transition

Abstract

Epithelial–mesenchymal transition (EMT) is an important process that contributes to renal fibrogenesis. TGF-β1 and EGF stimulate EMT. Recent studies suggested that parathyroid hormone–related protein (PTHrP) promotes fibrogenesis in the damaged kidney, apparently dependent on its interaction with vascular endothelial growth factor (VEGF), but whether it also interacts with TGF-β and EGF to modulate EMT is unknown. Here, PTHrP(1-36) increased TGF-β1 in cultured tubuloepithelial cells and TGF-β blockade inhibited PTHrP-induced EMT-related changes, including upregulation of α-smooth muscle actin and integrin-linked kinase, nuclear translocation of Snail, and downregulation of E-cadherin and zonula occludens-1. PTHrP(1-36) also induced EGF receptor (EGFR) activation; inhibition of protein kinase C and metalloproteases abrogated this activation. Inhibition of EGFR activation abolished these EMT-related changes, the activation of ERK1/2, and upregulation of TGF-β1 and VEGF by PTHrP(1-36). Moreover, inhibition of ERK1/2 blocked EMT induced by either PTHrP(1-36), TGF-β1, EGF, or VEGF. In vivo, obstruction of mouse kidneys led to changes consistent with EMT and upregulation of TGF-β1 mRNA, p-EGFR protein, and PTHrP. Taken together, these data suggest that PTHrP, TGF-β, EGF, and VEGF might cooperate through activation of ERK1/2 to induce EMT in renal tubuloepithelial cells.

In the injured kidney, interstitial fibroblasts are considered the main cell type responsible for fibrogenesis. During this process, these cells proliferate and can become activated to myofibroblasts, identified by de novo expression of α-smooth muscle actin (α-SMA), a prognostic marker for the progression of fibrogenesis.^1,2 It is currently known that a large proportion of interstitial myofibroblasts in fibrotic kidneys originate from tubuloepithelial cells through epithelial to mesenchymal transition (EMT).^2,3 On the other hand, EMT may also contribute to the repair process of the damaged tubules. Thus, after renal injury, remnant tubuloepithelial cells de-differentiate to mesenchymal cells, which migrate toward the damaged areas, where they proliferate and subsequently differentiate into the original epithelial phenotype to restore tubular integrity.⁴

EMT is a multistep process that requires the integration of multiple extrinsic and intrinsic pathways. Epithelia in transition lose polarity, adherence, and tight junctions and rearrange the F-actin cytoskeleton, associated with upregulation of many genes used as EMT markers.⁵ The latter includes, in addition to α-SMA, which increases cell contractility and motility,⁶ extracellular matrix proteins such as fibronectin and several types of collagens,^3,5 metalloproteases (MMPs) 2 and 9 implicated in the basal layer degradation,^7,8 and integrin linked kinase (ILK), present in focal contacts necessary to cell movement.⁹ In addition, a decrease in the expression of proteins that keep basolateral polarity, namely cytokeratin, and intercellular junctions, including the adherent junction proteins E-cadherin and β-catenin, takes place in the renal tubuloepithelium during EMT.⁵ Activation of different signaling pathways act as EMT intrinsic regulators: mitogen activated kinases (MAPKs), namely extracellular signal-regulated kinases (ERKs)1/2, p38, and N-terminal c-Jun kinase (JNK); the Ras protein family (Ras, Rho, and Rac); Wnt and Smad proteins; and the transcription factors snail and slug.^5,10–12

Different studies support the important role of various profibrogenic factors on the EMT process in the kidney.^8,13–16 The intracellular mechanisms underlying EMT induction by these factors are poorly known but at least in the case of TGF-β, the main factor studied in this regard, they seem to involve activation of MAPKs and Smad proteins.^12,16,17 In addition, activation of EGF receptor (EGFR) tyrosine kinase can trigger various cell responses including EMT in tubular cells, at least in part through MAPKs signaling pathway.^14,17,18 Moreover, induction of EMT by oxidative and osmotic stress and hypoxia—conditions associated with renal injury—can be accounted for in part by EGFR transactivation.¹⁹

Parathyroid hormone (PTH)-related protein (PTHrP) is upregulated in various experimental nephropathies.²⁰ In addition, a preliminary report has shown that PTHrP upregulation occurs in both tubules and glomeruli in patients with diabetic nephropathy.²¹ Recent studies in mice indicate that PTHrP can act as a proinflammatory and profibrogenic factor in the acutely damaged kidney after folic acid injection or unilateral ureteral obstruction.^22–24 Moreover, one of these studies strongly suggests that the profibrogenic action of PTHrP might be accounted for in part by promoting EMT through interaction with vascular endothelial growth factor (VEGF).²⁴ Interestingly in this scenario, it has previously been suggested that TGF-β might act as a modulator of at least some PTHrP actions through the PTH receptor 1 (PTHR1), common to PTH and PTHrP, in its target cells, such as tubuloepithelial cells and osteoblasts. Hence, TGF-β has shown to attenuate the inhibition of phosphate transport elicited by activation of this receptor in proximal tubule cells,²⁵ whereas PTH itself can induce the expression of TGF-β and amplifies its stimulatory effect on type I collagen production in osteoblastic cells.^26,27 Moreover, a recent report has shown that PTHR1 overexpression in a human osteosarcoma cell line resulted in its increased invasive capacity, associated with upregulation of TGF-β1.²⁸ Also of interest in the present context, in both human embryonic kidney cells HEK-293 and osteoblasts, stimulation of PTHR1 may lead to EGFR transactivation.^29,30 Collectively, these findings suggest that PTHrP might interact with TGF-β and/or EGFR to modulate EMT. To confirm this hypothesis, in this study, we aimed to characterize such interactions and their relative contribution to the mechanisms whereby PTHrP may induce EMT in renal tubuloepithelial cells.

Results

PTHrP(1-36) Can Induce EMT Through TGF-β in Mouse Cortical Tubule Cells

Incubation with PTHrP(1-36), at 100 nM, for 48 h causes a conversion of mouse cortical tubule (MCT) cells from an epithelial to a myofibroblast-like phenotype. As depicted by light microscopy, these transdifferentiated cells lost the typical cobblestone pattern of an epithelial monolayer and displayed a spindle-shape, fibroblast-like morphology (Figure 1A). These changes were associated with induction of the mesenchymal marker α-SMA and a decrease in the epithelial marker zonula occludens-1 (ZO-1) in these cells, as shown by confocal microscopy (Figure 1, B and C). Phalloidin staining, which detects F-actin, visualizes the cytoskeleton, showing actin fibers rearrangement (Figure 1C). In addition, PTHrP(1-36) increased snail immunostaining and induced its nuclear translocation (Figure 1D), where it can act as an E-cadherin transcription repressor. Snail also downregulates other epithelial markers but upregulates some mesenchymal markers (namely fibronectin and vitronectin) during EMT.^31,32 These findings further support and extend recent data,^22,24 indicating that PTHrP(1-36) promotes EMT in MCT cells.

Figure 1.

PTHrP(1-36) causes EMT in renal tubuloepithelial mouse cortical tubule (MCT) cells. Cells were stimulated with or without (control) PTHrP(1-36) (100 nM) for 48 h in culture medium. (A) Representative phase-contrast images of control and PTHrP(1-36)–stimulated MCT cells are shown (original magnification, ×200). Confocal microscopy analysis of α-SMA (B), ZO-1 (C), and snail (D) by immunofluorescence was performed using specific primary antibodies and a FITC-labeled secondary IgG. (C) TRITC-conjugated phalloidin (red staining) was used as a cytoskeleton marker to depict cell morphology. (D) Double immunofluorescence staining was assessed, using 4′,6-diamino-2-phenylindole dihydrochloride for the nucleus (blue) and FITC-labeled IgG (green) for snail. The inset shows in detail the presence of a green image in the cell nucleus. The overlaid images in green and blue (merge) yielded a white tone in the nucleus, indicating snail nuclear localization. This represents the results of three independent observations.

As mentioned above, TGF-β is one of the main inducers of EMT.^13–17 We tested here whether TGF-β could be a downstream mediator of PTHrP(1-36)–induced EMT in MCT cells. First, we evaluated whether this peptide would affect TGF-β production by these cells in a similar manner to PTH in osteoblastic cells.²⁶ We found that TGF-β1 mRNA was induced by PTHrP(1-36), at 100 nM, as early as 3 h, and it remained increased up to at least 18 h in MCT cells (Figure 2A). To establish whether the PTHrP(1-36)–induced rise in TGF-β1 mRNA was accompanied by TGF-β1 protein synthesis, its levels were measured in the MCT-conditioned medium. As expected,³³ this factor was mainly secreted as an inactive protein by these cells. In contrast, an increase in TGF-β1 secretion, which consisted of its active form, was elicited by PTHrP(1-36) at 48 h (Figure 2B).

Figure 2.

Early induction of both TGF-β₁ and VEGF by PTHrP(1-36) in MCT cells. MCT cells were stimulated with PTHrP(1-36) (100 nM) for different time periods. In some experiments, cells were pretreated for 1 h with an EGFR inhibitor (tyrphostin AG 1478, 100 nM) or an ERK1/2 inhibitor (U0126, 10 μM) before addition of PTHrP(1-36). Total cell RNA was isolated to assess mRNA levels of TGF-β₁ (A) and VEGF (C) by real-time PCR. (B) After PTHrP(1-36) stimulation for 48 h, total and active (after transient acidification) TGF-β₁ was measured in the cell-conditioned medium by a specific ELISA. Data are expressed as mean ± SEM of four independent experiments. *P < 0.05 versus control; #P < 0.05 versus PTHrP(1-36) alone.

We next used different strategies to block TGF-β at the time of MCT stimulation with PTHrP(1-36) by adding a neutralizing antibody against active TGF-β, which has proven to block angiotensin II–induced extracellular matrix production and EMT,^33,34 an inhibitor of the TGF-β type I receptor (TRI) kinase (TRI-ki), which prevents TGF-β–induced responses through the inactivation of TRI kinase,³⁴ and decorin, a proteoglycan that inhibits active TGF-β, thus acting as an antagonist.³⁵ All of these TGF-β blockers were shown to antagonize several EMT-related changes induced by PTHrP(1-36), as shown by immunofluorescence, in MCT cells (Figure 3A). Moreover, by Western analysis, E-cadherin was downregulated by this PTHrP peptide in these cells, consistent with recent findings,²⁴ an effect that was also abrogated by all these TGF-β blockers (Figure 3B). These data strongly suggest that this factor may act as a mediator of EMT induction by PTHrP(1-36) in tubuloepithelial cells.

Figure 3.

TGF-β acts as a mediator of PTHrP(1-36)–induced EMT in MCT cells. TGF-β was blocked or not (control) by pretreatment of MCT cells for 1 h with a TGF-β neutralizing antibody (α-TGF-β; 10 μg/ml), TRI-ki (10 μM), and the antagonist decorin (100 nM). Cells were stimulated with PTHrP(1-36) (100 nM) for 48 h. TGF-β1 (1 ng/ml) was used as positive control. (A) Detection of α-SMA, snail, and ZO-1 was performed by indirect immunofluorescence using FITC-labeled secondary IgG and confocal microscopy. This represents the results of three independent observations. (B) After PTHrP(1-36) stimulation, total cell protein was isolated to evaluate E-cadherin protein expression by Western blot. A representative autoradiogram is shown (top panel). GAPDH was used as loading control. Experimental values are mean ± SEM from six independent experiments and were expressed as n-fold over control. *P < 0.05 versus control; #P < 0.05 versus PTHrP(1-36) alone.

PTHrP(1-36) Induces EMT-Related Changes by EGFR Transactivation in MCT Cells

Congruent with previous data in an embryonic kidney cell line,³⁰ we found here that PTHrP(1-36), at 100 nM, rapidly (within 5 min) and transiently induced EGFR phosphorylation in MCT cells (Figure 4A). A major mechanism for EGFR transactivation by ligands interacting with G protein–coupled receptors (GPCRs), including PTH1R, involves activation of Gq/PKC (but not G_s) and proteolytic processing of EGFR ligands by MMPs.^29,30 Consistent with this notion, both calphostin C, a protein kinase C inhibitor, and GM6001, a pan-specific inhibitor of metalloproteases, were found to abolish EGFR activation by PTHrP(1-36) in MCT cells (Figure 4B).

Figure 4.

PTHrP(1-36) induces EGFR phosphorylation in MCT cells. Cells were stimulated with PTHrP(1-36) (100 nM) for different time periods (A) or for 5 min (B). EGFR phosphorylation was assesed by EGFR immunoprecipitation and Western analysis with p-tyrosine PY20 antibody as described in the Concise Methods. Total EGFR protein levels were used as phosphorylation control. (B) The following agents were added 1 h before PTHrP(1-36) stimulation: a PKC inhibitor (calphostin C, 250 nM), a pan-specific MMPs inhibitor (GM 6001, 1 μM), and an ERK 1/2 inhibitor (U0126, 10 μM). Values are mean ± SEM from at least three independent experiments by triplicate. *P < 0.05 and **P < 0.01 versus control.

We evaluated whether EGFR transactivation could also be a mediator of the PTHrP(1-36)–induced EMT in MCT cells. Cell preincubation with tyrphostin AG1478, a specific EGFR inhibitor, blocked the changes elicited by this PTHrP peptide on the expression of several EMT-related proteins, as assessed by immunofluorescence and Western blot (Figure 5, A and B). Supporting further that EGFR transactivation seems to be a mechanism whereby PTHrP(1-36) can induce EMT in MCT cells, both calphostin C and GM6001 significantly decreased the alterations triggered by this PTHrP peptide on ILK and E-cadherin protein expression in these cells (Figure 5C). On the other hand, Rp-cAMPS, a PKA inhibitor, was inefficient in this regard (Figure 5C).

Figure 5.

PTHrP can promote EMT by inducing EGFR transactivation in MCT cells. Cells were preincubated for 1 h with different inhibitors: tyrphostin AG1478 (100 nM), calphostin C (250 nM), GM 6001 (1 μM), PP1 (1 μM), and Rp-cAMPS (1 μM). Cells were stimulated with PTHrP(1-36) (100 nM) for 48 h. (A) Confocal microscopy analysis of α-SMA, snail, and ZO-1 immunofluorescence, using specific primary antibodies and a FITC-labeled secondary IgG was done. The inset shows in detail the presence of a green image denoting the tight junctions in nonstimulated tubuloepithelial MCT cells. This represents the results of three independent observations. (B and C) Changes in ILK and E-cadherin protein expression were analyzed by Western blot. Representative autoradiograms are shown. Results are data ± SEM from at least three independent experiments in triplicate. *P < 0.05 versus control; #P < 0.05 versus PTHrP(1-36) alone.

PTHrP Overexpression in the Mouse Obstructed Kidney Undergoing EMT Is Related to Increased TGF-β₁ Expression and EGFR Activation

We recently reported that some EMT-related changes occur associated with PTHrP overexpression in the mouse obstructed kidney.²⁴ Consistent with the aforementioned in vitro findings, we found here that both TGF-β₁ mRNA and p-EGFR protein levels were augmented in the mouse kidney in this in vivo scenario (Figure 6, A and B); associated with an increase of α-SMA and snail gene expression (Figure 6C). Moreover, all of these factors tested were consistently upregulated in the unobstructed kidney from transgenic mice with targeted overexpression of PTHrP to the renal proximal tubule, congruent with our recent observations.²⁴

Figure 6.

TGF-β₁ mRNA and p-EGFR protein expression associated with EMT-related changes occurs in the obstructed kidney from PTHrP transgenic (PTHrP-TG) mice and their control littermates. On day 4 after UUO, we evaluated gene expression of TGF-β₁ (A) and that of α-SMA and snail (by real-time PCR) (C) and p-EGFR phosphorylation (by Western blot, using antibodies against p-EGFR and EGFR). Representative autoradiograms are shown (B). Experimental values are mean ± SEM of three to four mice per group. All values were normalized against corresponding sham control. *P < 0.05 versus corresponding value in control mice; ^aP < 0.05 versus corresponding value in sham-operated mice.

Role of ERKs on the Intracellular Mechanism Underlying PTHrP(1-36) Induction of EMT in MCT Cells

The putative involvement of ERK activation on EMT induction by PTHrP(1-36) in tubuloepithelial cells was next studied. After PTHR1 activation, β-arrestin–dependent PTH1R internalization and/or G_q signaling can transactivate EGFR, leading to ERK1/2 phosphorylation.³⁰ Consistent with this notion, PTHrP(1-36), at 100 nM, was shown to induce ERK1/2 phosphorylation in a manner reminiscent to that observed for EGFR activation in MCT cells.²³ Moreover, the ERK1/2 phosphorylation inhibitor U0126 failed to affect PTHrP(1-36)–induced transactivation of EGFR in these cells (Figure 4B). On the other hand, tyrphostin AG1478, calphostin C, or several MMP inhibitors (GM6001 and an inhibitor of MMP-2 and MMP-9), in contrast to Rp-cAMPS, significantly inhibited ERK1/2 activation (Figure 7, A and B).

Figure 7.

PTHrP(1-36) phosphorylates ERK1/2 by a mechanism depending on EGFR activation in MCT cells. Before PTHrP(1-36) stimulation for 10 min, cells were preincubated for 1 h with different inhibitors: tyrphostin AG1478 (100 nM), calphostin C (250 nM), GM 6001 (1 μM), MMP-2 and 9 inhibitor (1 μM), PP1 (1 μM), and Rp-cAMPS (1 μM). ERK 1/2 phosphorylation was assessed by Western analysis. Total ERK levels were used for control of phosphorylation changes. Values are means ± SEM from at least three independent experiments in triplicate. *P < 0.05 versus control; #P < 0.05 versus PTHrP(1-36) alone.

We found that two ERK1/2 activation inhibitors, PD98059 and U0126, prevented PTHrP(1-36)–induced conversion of the epithelial to a myofibroblastic phenotype in MCT cells, as shown by confocal microscopy and Western blot (Figure 8, A and B). It has been suggested that Src kinases have a critical role in MAPK activation to induce tubuloepithelial cell dedifferentiation and migration—key events during EMT—after cell injury.^19,36 Thus, the possibility that Src kinases could mediate EMT induction by PTHrP(1-36) through ERK1/2 phosphorylation was examined. We observed that a Src kinase inhibitor abolished ILK and E-cadherin alterations (Figure 5C) and ERK1/2 phosphorylation (Figure 7B) triggered by PTHrP(1-36) in MCT cells.

Figure 8.

Role of ERK 1/2 activation on PTHrP(1-36)-induced EMT in MCT cells. Cells were preincubated for 1 h with the ERK1/2 inhibitors U0126 and PD98059 (10 μM) and treated with PTHrP(1-36) (100 nM) for 48 h. (A) Confocal microscopy analysis of α-SMA, snail, and ZO-1 immunofluorescence, using specific primary antibodies and a FITC-labeled secondary IgG. This represents the results of three independent observations. (B) Protein levels of ILK and E-cadherin were analyzed by Western blot. Representative autoradiograms are shown. Results are mean ± SEM from at least five independent experiments. *P < 0.05 versus control; #P < 0.05 versus PTHrP(1-36) alone.

We also aimed to evaluate whether this mechanism of EMT induction by PTHrP(1-36) in MCT cells might occur in human tubuloepithelial cells. We found that PTHrP(1-36) dose-dependently increased the protein expression of the mesenchymal marker vimentin and decreased that of the epithelial marker cytokeratin in HK-2 cells (Figure 9, A and B). Furthermore, these changes induced by this PTHrP peptide were abrogated by the inhibitors tyrphostin AG1478 and U0126, as shown by indirect immunofluorescence (Figure 9A)

Figure 9.

PTHrP(1-36) can promote EMT by interaction with EGFR and ERK1/2 in HK-2 cells. Cells were preincubated for 1 h with or without an ERK1/2 activation inhibitor (U0126, 10 μM) or tyrphostin AG1478 (100 nM) and then were stimulated with or without (control) PTHrP(1-36), at 100 nM (A) or at different concentrations (B), for 48 h in culture medium. (A) Confocal microscopy analysis of cytokeratin and vimentin immunofluorescence was performed using specific primary antibodies and a FITC-labeled secondary IgG. This represents the results of three independent observations. (B) Protein levels of cytokeratin-18 (Ck18) and vimentin were analyzed by Western blot. Representative autoradiograms are shown. Results are mean ± SEM from three independent experiments. *P < 0.05 versus control.

EGF seems to cooperate with TGF-β to induce at least some EMT changes in several renal tubuloepithelial cell preparations.^8,14,15 We found here that TGF-β1 and EGF, alone or in combination, were similarly efficient in inducing a myofibroblast phenotype (Figure 10A) and modulating several EMT-related proteins in MCT cells (Figures 10B and 11); this induction was blocked by an ERK1/2 activation inhibitor (Figures 10, A and B, and 11). In addition, both TGF-β1 and EGF elicited a rapid ERK1/2 phosphorylation that was maintained up to at least 4 h in these cells (Figure 10C). Our recent report suggested that the VEGF system might act as an important mediator of PTHrP(1-36) to induce several profibrogenic actions, including some EMT-related changes, in the obstructed mouse kidney.²⁴ We found here that U0126 also inhibited the changes elicited by VEGF₁₆₄ in α-SMA and ZO-1, as shown by confocal microscopy, in MCT cells (Figure 11). Of interest, inhibiting either EGFR or ERK1/2 activation abrogated the induction of both TGF-β1 and VEGF mRNA by PTHrP(1-36) in these cells (Figure 2, A and C). Collectively, these data support the hypothesis that ERK1/2 activation is a key pathway in the mechanism underlying EMT induction by any of these factors in renal tubuloepithelial cells.

Figure 10.

Role of ERK pathway in EMT-related changes by TGF-β₁ and EGF in MCT cells. Cells were preincubated for 1 h with or without an ERK1/2 activation inhibitor (U0126, 10 μM) and then treated with TGF-β₁ (1 ng/ml) and EGF (20 ng/ml), alone or in combination for 48 h. (A) Representative phase-contrast images of MCT cells treated with these factors or not (control) are shown (original magnification, ×200). (B) Changes in ILK protein levels analyzed by Western blot. (C) Time course of ERK1/2 phosphorylation after stimulation with TGF-β and EGF. Representative Western blots are shown. Results are mean ± SEM of three independent experiments. *P < 0.05 versus control; ##P < 0.05 versus corresponding factor without U0126.

Figure 11.

Inhibition of ERK activation abrogated the induction of α-SMA and ZO-1 by TGF-β₁, EGF, or VEGF₁₆₄ in MCT cells. Cells were preincubated for 1 h with or without an ERK1/2 activation inhibitor (U0126, 10 μM) and then treated or untreated (control) with TGF-β₁ (1 ng/ml), EGF (20 ng/ml), or VEGF164 (20 ng/ml) for 48 h. Confocal microscopy analysis of α-SMA and ZO-1 immunofluorescence was performed using specific primary antibodies and a FITC-labeled secondary IgG. This represents the results of three independent observations.

Discussion

This study further extends our previous studies²⁴ and shows that PTHrP(1-36) induces a variety of phenotypic changes related to EMT in tubuloepithelial cells. Hence, this peptide induced snail overexpression and its nuclear translocation, associated with the loss of ZO-1 and E-cadherin, key proteins in the maintenance of basolateral polarity and intercellular junctions in the renal tubuloepithelium.^3,5,6 PTHrP(1-36) also induced the phenotypic conversion to a fibroblast-like morphology, related to α-SMA and ILK upregulation. In a previous report, we showed that the VEGF system seemed to have an important role at least in some of these alterations induced by PTHrP(1-36) in tubuloepithelial cells.²⁴ These results show for the first time that PTHrP(1-36) can also interact with EGFR activation and TGF-β to elicit EMT in these cells. In addition, our in vivo data in mice suggest that such interaction might also occur in vivo in the damaged kidney.

Current information regarding the interaction between TGF-β and PTHrP in the kidney is scarce. TGF-β has been shown to downregulate the PTHR1 in renal tubuloepithelial cells,³⁷ but there are no available data about a putative effect of PTHrP on TGF-β in these cells. In this study, we showed that PTHrP(1-36) can increase TGF-β, at both gene and protein secretion levels, in MCT cells. TGF-β blockade by different maneuvers, including a neutralizing antibody, antisense oligonucleotides, decorin, and Smad7 overexpression, was found to diminish renal fibrosis in both experimental models of renal damage and cultured renal cells.³⁸ TGF-β is a key modulator of the entire EMT process. In tubuloepithelial cells, antagonizing TGF-β diminishes EMT caused by either angiotensin II or high glucose (two pro-fibrogenic factors in the kidney).^33,38 We show here that different strategies for blocking TGF-β, such as a neutralizing antibody, a TRI-ki, and the antagonist decorin, significantly diminished EMT induction by PTHrP(1-36) in MCT cells, strongly suggesting that TGF-β is a downstream mediator of this PTHrP(1-36) action.

In this study, we also show that PTHrP(1-36) can induce EGFR phosphorylation with a pattern of response similar to that elicited by ligands signaling through various GPCRs, such as the calcium receptor in prostate cancer cells, the angiotensin II receptor in vascular smooth muscle cells, and the PTH1R in murine osteoblasts and human embryonic kidney cells.^29,30,39,40 Two main different EGFR transactivation mechanisms induced by GPCRs have been described thus far.^41,42 One of them involves the proteolytic processing of EGFR ligands, which are synthesized as membrane-anchored precursors [e.g., common proheparin-binding (proHB)-EGF], by GPCR-activated MMPs to release soluble EGFR ligands.^39,40,43 Such a mechanism has been described for EGFR transactivation through the PTH1R in HEK-293 cells, which seems to involve PKC activation as a consequence of Gq signaling but not PKA signaling.³⁰ Alternatively, EGFR can be transactivated by the soluble kinase Src that directly phosphorylates and activates EGFR.⁴⁴ These results suggest that the first mechanism seems to account mainly for EGFR transactivation by PTHrP(1-36) in tubuloepithelial cells. Although the true EGFR ligand involved in PTHrP(1-36)-induced EGFR transactivation in these cells is presently unknown, HB-EGF type of ligands are expressed by proximal tubule cells.^44,45 In addition, as occurs in rabbit proximal tubule cells under stress,¹⁹ the ERK1/2 pathway does not seem to be a cause but a consequence of EGFR activation by PTHrP(1-36) in MCT cells. In addition, these findings indicate that EGFR activation by this PTHrP peptide can promote EMT in tubuloepithelial cells. Also consistent with this notion, PKC activation but not PKA signaling seems to be involved in this action of PTHrP(1-36). Our data suggest that Src activation also contributes to PTHrP(1-36)-induced EMT features in MCT cells. In this regard, the role of Src as a regulating kinase during the EMT process in renal tubuloepithelial cells has recently been reported.¹⁹ Moreover, in renal tubuloepithelial MDCK cells, Src activation induces ERK1/2 phosphorylation to promote cell migration.³⁶ In addition, accumulated evidence suggests a bidirectional association between Src and EGFR, leading to activation of both kinases, which is a mechanism required for many cellular functions including cell migration.⁴⁶ Our present results suggest that a similar mechanism might be functional in MCT cells after stimulation with PTHrP(1-36).

Our data also show that ERK1/2 pathway is essential for PTHrP(1-36) to induce EMT changes through EGFR activation and also by modulating TGF-β₁ and VEGF expression in tubuloepithelial cells. Of interest in this regard, both EGF and TGF-β₁ were shown to increase VEGF expression in several tubuloepithelial cell lines, an effect that, at least in the case of TGF-β₁, occurs through ERK1/2 activation.^47,48 Thus, such a mechanism might contribute to the VEGF upregulation by PTHrP(1-36) in MCT cells. In any event, these findings are consistent with an important role of ERK1/2 signaling pathway in the EMT process.^17,49 The latter pathway may activate snail and block transcription of E-cadherin, clauidin, and ocludin, leading to disruption of intercellular junctions in tubuloepithelial cells, which is the first step of EMT. Moreover, E-cadherin repression increases cytoplasmic β-catenin, a key factor in EMT induction.^5,12,32

These findings further expand those recently reported²⁴ and strongly support that PTHrP(1-36) can be considered as a new EMT modulator in the kidney. Our data show for the first time that, besides VEGF (as shown in our previous work), EGFR and TGF-β can also be important mediators of this process induced by PTHrP(1-36) in the renal tubular epithelium. Moreover, ERK activation seems to be a key downstream event whereby these growth factors might cooperate to induce EMT (Figure 12).

Figure 12.

Proposed mechanism for ERK1/2 activation as a key signaling pathway targeted by PTHrP(1-36) through interaction with EGFR, TGF-β, and VEGF in renal tubuloepithelial cells.

Concise Methods

Cell Cultures

Renal tubuloepithelial MCT cells and the human proximal tubule cell line HK-2, which respond to PTHrP(1-36),^22–24,50 were grown in RPMI 1640 with 10% FBS, supplemented with 1% insulin-transferrin-sodium selenite media (Sigma-Aldrich, St. Louis, MO) and hydrocortisone (36 ng/ml) (HK-2 cells), and antibiotics in 5% CO₂ at 37°C. Subconfluent cells (60,000 cells/cm²) were incubated with PTHrP(1-36) (100 nM), human recombinant TGF-β₁ (1 ng/ml) (Peprotech, Rocky Hill, NJ), and EGF (20 ng/ml) (Sigma-Aldrich), or mouse recombinant VEGF₁₆₄ (20 ng/ml) (R&D Systems, Minneapolis, MN), in this medium without (HK-2) or with 1% FBS (MCT) for different times. In some experiments, the following inhibitors were added 1 h before PTHrP(1-36): PD98059 (Stressgen Bioreagents, Victoria, BC, Canada); U0126 (Promega, Madison, WI) (each at 10 μM); 4-amino-5-(4-methylphenyl)-7-(t-butyl)pyrazolo[3,4-d]-pyrimidine (PP1) (1 μM); calphostin C (250 nM); (2R)-2-[(4-biphenyllylsulfonyl)amino]-3-phenylpropionic acid (MMP-2/MMP-9 inhibitor I] (1 μM); GM6001, a pan-specific inhibitor of MMPs (1 μM) and TRI-ki (10 μM) (Calbiochem; San Diego, CA); Rp-cAMPS (1μM) and decorin (100 nM) (Sigma-Aldrich); and tyrphostin AG 1478 (100 nM) (Alomone Labs, Jerusalem, Israel), an anti-TGF-β–neutralizing antibody that recognizes bovine, mouse, and human TGF-β₁ and β₂ isoforms (1 μg/ml) (R&D, Minneapolis, MN).

Western Blot Analysis

After 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 phenylmethylsulphonylfluoride, 0.8 μM aprotinin, and a phosphatase-inhibitor cocktail (Set II; Calbiochem)] were obtained for protein analysis. Bradford's method (Pierce, Rockford, IL) was used with BSA as standard. The efficacy of protein loading and transfer to membranes was assessed by α-tubulin and GAPDH.

To analyze changes in EGFR phosphorylation in MCT cells, cell protein extracts (200 μg) were incubated with 0.2 μg rabbit IgG for 1 h at 4°C with shaking. Then, 30 μl of protein A-agarose were added to cell extracts (0.5 ml agarose/2 ml PBS) and incubated for 2 h at 4°C with shaking. After centrifugation for 5 min at 1500 × g, supernatants were incubated overnight with a rabbit polyclonal anti-EGFR antibody (Calbiochem; 1:200) at 4 °C with shaking. Protein A-agarose was added, and protein samples were subsequently incubated and centrifuged as described above.

MCT cell protein extracts or the reconstituted immunoprecipitated pellets (for EGFR analysis) (30 to 80 μg/lane) and mouse kidney protein extracts (150 μg/lane), obtained as recently described,²⁴ were separated on 5 to 10% polyacrylamide-SDS gels under reducing conditions. Samples were transferred onto nitrocellulose membranes (Bio-Rad, Hercules, CA), blocked with 5% defatted milk in 50 mM Tris-HCl, pH 7.5, and 150 mM NaCl with 0.05% Tween-20, and incubated overnight at 4 °C with the following antibodies [dilution, -fold]: mouse monoclonal E-cadherin [2000] (BD Transduction Laboratories, San José, CA); p-tyrosine (PY20) [500] and EGFR [250] (Santa Cruz Biotechnology, Santa Cruz, CA); vimentin [10,000] (BD Pharmigen, Franklin Lakes, NJ); and pan-cytokeratin [1000] (Sigma-Aldrich) antibodies; and rabbit polyclonal ILK [1000] (Santa Cruz Biotechnology); ERK 1/2, or phosphorylated (p)-ERK 1/2 (Thr²⁰²/Tyr²⁰⁴) [2000] (Cell Signaling Technology; Beverly, MA); and p-EGFR (p-Tyr¹⁰⁶⁸) [250] (Calbiochem) antibodies. Membranes were subsequently incubated with peroxidase-conjugated IgG and developed by ECL chemiluminescence (GE Healthcare, Buckinghamshire, UK). Densitometric values of fluorogram bands were normalized to those of corresponding α-tubulin or glyceraldehyde 3-phosphate dehydrogenase.

Immunofluorescence

Cells grown on coverslips were stimulated with the agonists, fixed in Merckofix (Merck, Whitehouse Station, NJ), and permeabilized with 0.1% Triton-X100 for 2 min. After blocking with 10% BSA and 10% FBS for 1 h, they were incubated with several primary antibodies [dilution, -fold]: rabbit polyclonal anti-Z0–1[200] (Zymed Laboratories. San Francisco, CA); anti-α-SMA[200] (ABCam, Cambridge, MA);[200] and anti-snail (Sta. Cruz Biotechnology)[200] antibodies; or mouse monoclonal anti-vimentin[200] (BD Pharmigen) and anti-pancytokeratin[200] (Sigma-Aldrich) antibodies for 1 h, followed by a FITC-conjugated secondary antibody [200] or tetramethyl rhodamine isothiocyanate (TRITC)-conjugated phalloidin (Sigma-Aldrich) [1000] for 1 h. Absence of primary antibody was used as negative control. Samples were mounted in Mowiol 40-88 (Sigma-Aldrich) and examined using a Leica DM-IRB confocal microscope.

Analysis of mRNA Expression

Total RNA was isolated from MCT cells and mouse kidney samples with Trizol (Invitrogen, Groningen, The Netherlands). cDNA was synthesized using the high-capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA) using 2 μg of total RNA primed with random hexamer primers, following the manufacturer's instructions. Real-time PCR was performed using mouse fluorogenic TaqMan MGB probes and primers designed by Assay-on-Demand gene expression products (Applied Biosystems): TGF-β₁ (Mm001178819_m1), α-SMA (Mm01546133_m1), snail (Mm00441533_g1), and VEGF (Mm00437304_m1). Data were normalized to 18S eukaryotic ribosomal RNA. The mRNA copy numbers were calculated for each sample by the instrument software using Ct value (“arithmetic fit point analysis for the lightcycler”). Results were expressed in copy numbers, calculated relative to unstimulated cells after normalization against 18S.

TGF β₁ Protein Assay

TGF-β₁ protein was measured in the MCT cell-conditioned medium after treatment with PTHrP(1-36) (100 nM) for 48 h, using a commercial ELISA (BD Sciences, San Diego, CA) following the manufacturer's instructions. Active TGF-β₁ was determined in 100 μl of the cell-conditioned medium (stored at −80°C). Inactive TGF-β₁ was converted to the active form by incubating these cell culture supernatants with 10 N HCl for 10 min, followed by neutralization with 10 N NaOH/0.5 M HEPES. Protein content was determined by the BCA method (Pierce). TGF-β₁ activity was quantified by comparison with a standard curve of human TGF-β₁.

Mouse Model of Unilateral Ureteral Obstruction

Transgenic mice with targeted overexpression of PTHrP to the renal proximal tubule (PTHrP-TG) and their control littermates were used.^22–24 Unilateral ureteral obstruction (UUO) was performed under anesthesia 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 previously reported.^23,24 Four days thereafter, mice were killed, and the obstructed kidneys were collected. Sham-operated mice, which had their ureters manipulated but not ligated, served as controls. Kidney portions were separated from each mouse and stored at −80°C for subsequent analysis. Experimental protocols were approved by the Institutional Animal Care and Use Committee at Fundación Jiménez Díaz.

Statistical Analysis

Results throughout the text are expressed as mean ± SEM. Differences between agonist-treated groups and controls were assessed by one-way ANOVA, followed by the post hoc Bonferroni or Dunnett test, or Mann-Whitney test, as appropriate. P < 0.05 was considered significant. Statistical analysis was conducted using the SPSS statistical software (version 11.0; SPSS, Chicago, IL).

Disclosures

None.