BMP-7 fails to attenuate TGF-β1-induced epithelial-to-mesenchymal transition in human proximal tubule epithelial cells

Transcription

1 NDT Advance Access published December 4, 2008 Nephrol Dial Transplant (2008) 1 of 10 doi: /ndt/gfn662 Original Article BMP-7 fails to attenuate TGF-β1-induced epithelial-to-mesenchymal transition in human proximal tubule epithelial cells Paul L. Dudas, Rochelle L. Argentieri and Francis X. Farrell Department of Immunology Research, Centocor Research and Development, Inc., Radnor, PA 19087, USA Abstract Background. In rodent models of chronic renal disease bone morphogenetic protein-7 (BMP-7) has been shown to halt disease progression and promote recovery. Subsequent studies utilizing immortalized rodent renal cell lines showed that BMP-7 was renoprotective by antagonizing TGF-β1-stimulated epithelial-to-mesenchymal transition (EMT). The present study sought to determine if BMP- 7 prevents TGF-β1-induced EMT in primary (RPTEC) and immortalized (HK-2) human proximal tubule epithelial cells. Methods. EMT was determined by quantitative real-time PCR analysis of e-cadherin, vimentin, CTGF and TGF-β1 transcript expression and immunocytochemical analysis of ZO-1 and α-smooth muscle actin (α-sma) protein expression following TGF-β1 treatment in RPTEC and HK-2 cells. Results. In RPTEC and HK-2 cells, TGF-β1 significantly reduced e-cadherin expression and significantly increased vimentin, CTGF and TGF-β1 expression. TGF-β1 also diminished ZO-1 immunoreactivity and increased α-sma expression in confluent cell monolayers. Co-incubation of TGF-β1 with an anti-tgf-β1 neutralizing antibody substantially reduced the cytokine s effects, which indicated EMT in these cells was inhibitable. Co-administration of BMP-7 over a broad concentration range ( µg/ml) with TGF-β1 failed to attenuate EMT in RPTEC or HK-2 cells, as demonstrated by no inhibition of altered e-cadherin, vimentin, CTGF and TGF-β1 expression and no restoration of ZO-1 immunoreactivity. Furthermore, when BMP-7 was applied to proximal tubule cells alone, it also decreased e-cadherin expression and increased vimentin, CTGF and TGF-β1 expression. Additionally, BMP-7 failed to induce the mesenchymal-to-epithelial transition (MET) in NRK- 49F rat renal fibroblasts. BMP-7 did however prevent TGFβ1-mediated e-cadherin downregulation in TCMK-1 mouse renal tubular epithelial cells. BMP-7 activity was routinely Correspondence and offprint requests to: Paul L. Dudas, Department of Immunology Research, Centocor Research and Development, Inc., 145 King of Prussia Road, R-4-2, Radnor, PA 19087, USA. Tel: ; Fax: ; pdudas@its.jnj.com confirmed by examining BMP-7-induced phosphorylation of SMADs 1/5/8, BMP-7 regulation of BMPR-IA, BMP-7- mediated reduction of IL-6 transcript expression and BMP- 7-mediated reduction of secreted IL-6 and IL-8 proteins. Conclusions. In the present study, despite confirming BMP-7 regulation of receptor expression and induction of downstream signalling events, we were unable to demonstrate BMP-7 inhibition of EMT in either primary or immortalized human proximal tubule cells. Moreover, we were unable to demonstrate BMP-7-stimulated MET in rat renal fibroblasts. A protective effect was however observed at an elevated BMP-7 concentration in mouse renal tubular epithelial cells. Keywords: BMP-7; epithelial-to-mesenchymal transition; proximal tubule; TGF-β1 Introduction Chronic progressive kidney disease can initially be manifested as an inflammatory response present in both the glomerular and tubulo-interstitial compartments and ultimately result in organ fibrosis. Damage to the glomerulus is typically mediated by mesangial cell activation/hyperplasia, pro-inflammatory/pro-fibrotic cytokine/chemokine release by mesangial cells and infiltrating inflammatory cells, podocyte loss and a degradation of the glomerular capillary architecture [1 4]. In the tubulo-interstitial space, particularly in the proximal nephron, there is infiltration of pro-inflammatory cells, activation of resident proximal tubule cells to release pro-inflammatory/pro-fibrotic mediators and apoptosis of the proximal tubule epithelium [5,6]. Concomitantly, a conversion of mature proximal tubule epithelium to a de-differentiated, motile, fibroblast-like phenotype occurs [7,8]. This phenotypic conversion, termed epithelial-to-mesenchymal transition (EMT), is speculated to be a protective mechanism in that it enables cellular migration from the tubular microenvironment to the interstitial space, allowing the cells to escape apoptotic cell death [8,9]. However, studies have revealed that EMT can C The Author [2008]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org

2 2 P. L. Dudas et al. also be a major contributor to the onset and pathogenesis of renal fibrosis in that it leads to an increase in myofibroblast activity and tubular atrophy [8,10]. Bone morphogenetic protein-7 (BMP-7), a member of the transforming growth factor-β superfamily, has been implicated in the maintenance of the normal health of the kidney, skeleton and vascular system [11,12]. It was shown to be drastically downregulated in injured kidneys and later demonstrated that administration of recombinant BMP-7 to rodent models of chronic progressive kidney disease halted the progression of the disease and promoted renal recovery [13 15]. A subsequent study in mice examining the mechanism of action of BMP-7 protection showed that BMP-7 functioned not only to prevent renal injury by inhibiting proximal tubule cell EMT but promoted recovery by stimulating the re-population of injured proximal tubules with healthy cells, partly through inducing mesenchymalto-epithelial transition (MET) [16,17]. The present study sought to examine the in vitro efficacy of BMP-7 in the prevention of human proximal tubule cell EMT. Primary human proximal tubule epithelial cells (RPTEC) and an immortalized human proximal tubule epithelial cell line (HK-2) were evaluated. To assess the progression of EMT and potential attenuation by BMP-7, the pro-epithelial marker e-cadherin and the promesenchymal/pro-fibrotic markers vimentin, CTGF and TGF-β1 were assayed by quantitative real-time PCR. Additionally, immunofluorescence microscopy was utilized to assess the expression of α-smooth muscle actin (α-sma) as a marker of the mesenchymal transition and the tightjunction protein ZO-1 as a marker of epithelial integrity. For comparison to a rodent renal epithelial cell line, BMP-7 inhibition of TGF-β1-mediated e-cadherin downregulation was examined in TCMK-1 mouse tubular epithelial cells. Also, the capacity for BMP-7 to stimulate MET in rat renal fibroblasts (NRK-49F) as another prospective protective mechanism of action was investigated. Materials and methods Cell culture and RNA isolation RPTEC were obtained from Lonza (Walkersville, MD) and passaged in the recommended growth medium: renal epithelial basal medium containing 10 µg/ml recombinant human epidermal growth factor, 5 mg/ml insulin, 0.5 mg/ml hydrocortisone, 0.5% FBS, 0.5 mg/ml epinephrine, 6.5 µg/ml T3, 10 mg/ml transferrin, 10 mg/ml gentamicin and 50 µg/ml amphotericin-b. Cells were used between passages 3 and 9. Immortalized human proximal tubule epithelial cells (HK-2) were obtained from ATCC (Manassas, VA) and passaged in the recommended growth medium, keratinocyte-serum free medium containing 5 ng/ml recombinant human epidermal growth factor and 0.05 mg/ml bovine pituitary extract. Cells were used between passages 3 and 20. Mouse renal tubular epithelial cells (TCMK-1) were obtained from ATCC and passaged in the recommended growth medium: Eagle s MEM with 2mML-glutamine and Earle s BSS, 0.1 mm non-essential amino acids, 1 mm sodium pyruvate and 10% FBS. Cells were used between passages 5 and 10. Rat renal fibroblasts (NRK-49F) were obtained from ATCC and passaged in the recommended growth medium: DMEM with 4 mm L-glutamine, 4.5 g/l glucose, and 110 mg/l Na-pyruvate and 5% FBS. Cells were used between passages 3 and 15. RPTEC, HK-2 or TCMK-1 cells were plated at a density of cells/well in standard six-well tissue culture plates 48 h prior to treatment. NRK-49F cells were plated at a density of cells/well in standard 24-well tissue culture plates 48 h prior to treatment. Twenty-four hours prior to treatment, cells were maintained in medium without additives. TGF-β1 (R&D Systems, Minneapolis, MN) ± anti- TGF-β1 neutralizing antibody (R&D Systems), TGF-β1 ± BMP-7 (Curis, Inc., Cambridge, MA) or BMP-7 alone were then applied to the cells for 24 h. TGF-β1 at 10 ng/ml was utilized with RPTEC and TCMK-1, and TGF-β1 at 3 ng/ml was utilized with HK-2 cells for the course of the study. RNA isolations were accomplished using an RNeasy Plus Mini Kit (Qiagen, Valencia, CA) according to the manufacturer s instructions. RNA concentration was obtained using a Nanodrop ND-1000 spectrophotometer (Nanodrop Technologies, Wilmington, DE). RT-PCR and gene expression analysis cdna was generated by reverse transcribing 1 µg of total RNA in a reaction volume of 100 µl using random hexamers as a priming agent with the Applied Biosystems Taqman Reverse Transcriptase Kit (Applied Biosystems, Carlsbad, CA). Two microlitres (20 ng) of cdna was used as template in 20 µl PCR reactions. Quantitative real-time PCR was performed using an ABI Prism 7900 HT Sequence Detection System (Applied Biosystems) with Taqman Universal Master Mix (Applied Biosystems) and gene-specific expression assays containing two unlabelled PCR primers and one FAM dye-labelled Taqman MGB probe (Applied Biosystems). Reactions were performed in at least triplicate and analysed via relative quantitation with data presented as average fold change compared to no treatment after normalization to GAPDH RNA or 18S rrna. Immunofluorescence labelling of ZO-1 and α-sma RPTEC or HK-2 cells were grown to confluence on sterile cover slips (18 18 mm) under the same conditions described above. Cells were then treated with TGF-β1 ± anti-tgf-β1 neutralizing antibody or TGF-β1 ± BMP-7 for 48 h. For ZO-1 staining, cells were rinsed with PBS (3, 5 min) and fixed in 4% paraformaldehyde (15 min, RT). Next, cells were rinsed in PBS (1, 5 min) and permeabilized with 0.5% Triton X-100 in PBS (8 min). Cells were rinsed in PBS % Tween 20 (PBST) (3, 5 min) and blocked with Powerblock (9 min) (Biogenex, San Ramon, CA). After blocking, the cells were rinsed in PBST (3, 5 min) and incubated in mouse anti-human ZO-1 (18 µg/ml) (Invitrogen, Carlsbad, CA) in PBS % BSA % NaN 3 in a humidified chamber for 2 h, RT. Cells were next rinsed in PBST (3, 5 min) followed by incubation in secondary antibody goat anti-mouse Alexa Fluor 488 or 594 (8 µg/ml) (Molecular Probes, Carlsbad, CA) in the same buffer used for the primary antibody, for 1 h, RT in a humidified chamber in the dark. Cells were rinsed in PBST (3, 5 min) and PBS (1, 5 min). Cover slips were overturned

3 BMP-7 fails to attenuate TGF-β1-induced EMT 3 on a microscope slide containing one drop of SlowFade Gold Antifade reagent with DAPI (Invitrogen). Slides were visualized on a Nikon Eclipse E800 microscope. For α-sma staining, cells were rinsed with PBS (3, 5 min) and fixed in 4% paraformaldehyde with 0.15% picric acid (20 min, RT). Next, cells were rinsed in PBS (2, 5min) and permeabilized and blocked with 0.1% Triton X % BSA + 10% normal donkey serum in PBS (45 min, RT). After permeabilization and blocking, the cells were incubated in mouse anti-human α-sma (10 µg/ml) (R&D Systems) in PBS + 1% BSA (3 h, RT). Next, cells were rinsed in PBS + 0.1% BSA (3, 5 min) followed by incubation in secondary antibody donkey anti-mouse Alexa Fluor 594 (8 µg/ml) (Molecular Probes) in PBS + 1% BSA (1 h, RT) in the dark. Cells were rinsed in PBS + 0.1% BSA (3, 5 min) and PBS (1, 5min).Coverslips were mounted on slides and visualized as described above. SDS PAGE and immunoblotting RPTEC or HK-2 cells were plated in full growth medium in six-well tissue culture dishes at cells/well and allowed to adhere for 24 h. The next day, the full growth medium was replaced with supplement-free medium in which the cells were maintained for 24 h. Next, one well of cells was treated with BMP-7 (200 ng/ml) in supplementfree growth medium for 20 min at 37 C. A corresponding control well had the medium changed to fresh supplementfree medium without BMP-7. Following the 20-min treatments, cells were solubilized on ice for 15 min with 1% Triton X-100 in PBS + protease inhibitors (Complete Protease Inhibitor Tablets, Roche, Indianapolis, IN). The total protein concentration of cell lysate from each treatment was determined by Bradford protein assay (Bio-Rad, Hercules, CA). Ten micrograms of total protein from each treatment was solubilized in SDS sample buffer and subjected to SDS PAGE using 10% polyacrylamide gels. Proteins were subsequently transferred to a polyvinylidene fluoride (PVDF) microporous membrane (Millipore, Billerica, MA). Nonspecific binding was blocked by incubating the PVDF membrane at RT for 1 h in 5% BSA + 0.1% Tween 20 in TBS followed by overnight incubation with polyclonal rabbit anti-human phospho-smad 1/5/8 (1:850) (Cell Signaling Technology, Danvers, MA) in TBST at 4 C. The membrane was washed with TBS + 0.1% Tween 20 (TBST) 3, 5 min and then incubated for 1 h at RT with a 1:8000 dilution of goat anti-rabbit HRP-conjugated secondary antibody (Jackson Immunoresearch) in 5% nonfat dry milk in TBST. Next, the membrane was washed 3, 5 min in TBST and 1, 5 min in PBS. Finally, an enhanced chemiluminescence detection (ECL) system (GE Healthcare, Piscataway, NJ) with Kodak Biomax imaging film (Kodak, Rochester, NY) was used for the visualization of the bound antibodies. Fig. 1. TGF-β1-induced epithelial-to-mesenchymal transition in primary and immortalized human proximal tubule epithelial cells. (A) Treating primary human proximal tubule epithelial cells (RPTEC) with TGF-β1 (10 ng/ml, 24 h) significantly decreased e-cadherin and increased vimentin, CTGF and TGF-β1 transcript expression indicating a transition from an epithelial-to-mesenchymal phenotype. (B) Treating immortalized human proximal tubule epithelial cells (HK-2) with TGF-β1(3 ng/ml,24 h) also induced epithelial-to-mesenchymal transition as determined by a downregulation of e-cadherin and concurrent upregulation of vimentin, CTGF and TGF-β1 transcript expression. Pre-incubation of TGF-β1 with anti-tgf-β1 antibody (120 ng/ml, 1 h) abolished TGF-β1-induced EMT. Values are means ± 1SE. Significantly different from paired controls (P < 0.05). τ Significantly different from TGF-β1 treated (P < 0.05). vesting cellular supernatants. Human IL-6 and IL-8 immunoassays (R&D Systems) were used to quantitatively assay these cytokines in HK-2 cellular supernatants. Briefly, a monoclonal antibody specific for IL-6 or IL-8 was precoated onto a 96-well plate. Any relevant cytokine present in the cellular supernatant was bound by the monoclonal antibody and detected using cytokine-specific enzyme-linked polyclonal antibodies. Colorimetric detection following a substrate reaction was used to determine the amount of bound polyclonal antibody. ELISA HK-2 cells were plated at a density of cells/well in standard 24-well tissue culture plates in complete growth medium (described above) 48 h prior to treatment. Twentyfour hours prior to treatment cells were maintained in growth medium without additives. Cells were incubated with BMP-7 ( ng/ml) for 24 h followed by har- Results To examine the prospective efficacy of BMP-7 in inhibiting TGF-β1-induced EMT, we first verified the ability of TGF-β1 to induce EMT in both primary (RPTEC) and immortalized (HK-2) human proximal tubule epithelial cells. As shown in Figure 1A, following 24 h TGF-β1

4 4 P. L. Dudas et al. Fig. 2. ZO-1 protein expression in anti-tgf-β1 antibody treated primary and immortalized human proximal tubule epithelial cells following TGF-β1- induced epithelial-to-mesenchymal transition. (A) Untreated primary human proximaltubule cells (RPTEC) exhibiting robust ZO-1 immunoreactivity at cell cell junctions. (B) Following treatment with TGF-β1 (10 ng/ml, 48 h), ZO-1 immunoreactivity was substantially diminished indicating the cells were undergoing epithelial-to-mesenchymal transition. (C) Addition of an anti-tgf-β1 antibody (400 ng/ml) concurrently with TGF-β1 prevented the TGFβ1-induced downregulation of ZO-1 in RPTEC. (D) Untreated immortalized human proximal tubule cells (HK-2) exhibiting ZO-1 immunoreactivity at cell cell junctions, similar to RPTEC. (E) Following TGF-β1 treatment (3 ng/ml, 48 h), ZO-1 immunoreactivity was substantially diminished and cells assumed a fibroblast-like morphology indicating the onset of epithelial-to-mesenchymal transition. (F) Treatment with anti-tgf-β1 antibody (120 ng/ml) concurrently with TGF-β1 prevented the downregulation of ZO-1 and morphology change in HK-2 cells. Magnifications: RPTEC 200, HK Bars indicate 35 µm. Fig. 3. TGF-β1 induced α-smooth muscle actin expression in primary and immortalized human proximal tubule epithelial cells. (A) Untreated primary human proximal tubule cells (RPTEC) exhibiting negligible α-sma expression. (B) Following treatment with TGF-β1 (10 ng/ml, 48 h), α-sma immunoreactivity was substantially increased indicating the cells are undergoing epithelial-to-mesenchymal transition. (C) Untreated immortalized human proximal tubule cells (HK-2) exhibiting negligible α- SMA expression. (D) Following treatment with TGF-β1 (3 ng/ml, 48 h), α-sma immunoreactivity increased indicating the cells are undergoing epithelial-to-mesenchymal transition. Magnification: 120. Barsindicate 30 µm. treatment, EMT was readily induced in RPTEC as determined by a downregulation of e-cadherin and upregulation of vimentin, CTGF and TGF-β1 transcript expression. For comparison to another human proximal tubule epithelial cell line, the same experiment described above was performed in HK-2 cells. As shown in Figure 1B, TGF-β1 induced EMT similar to RPTEC. Moreover, in this same experiment, EMT was completely inhibited by anti-tgfβ1 antibody treatment. For further confirmation of TGF-β1-induced EMT in RPTEC and HK-2 cells, we examined ZO-1 protein expression as an indicator of tight-junction integrity and maintenance of epithelial organization and α-sma protein expression as an indicator of transition to a mesenchymal phenotype. As shown in Figure 2B and 2E, treatment of RPTEC and HK-2 cells with TGF-β1 for 48 h resulted in a substantial downregulation of ZO-1 immunoreactivity compared to untreated cells (Figure 2A, D). Concurrent with the decrease in ZO-1 protein expression was an alteration in cell morphology from the characteristic organized cobblestone appearance of differentiated epithelial cell monolayers to a disorganized elongated fibroblast-like phenotype. The morphological transition was most evident in the HK-2 cells (Figure 2E). The changes in ZO-1 protein expression for both RPTEC and HK-2 cells and the morphology change in HK-2 cells were prevented following pre-incubation of TGF-β1 with a TGF-β1-neutralizing antibody (Figure 2C,F). Also, in parallel with diminished epithelial differentiation and organization, α-sma protein

5 BMP-7 fails to attenuate TGF-β1-induced EMT 5 Fig. 4. Effect of BMP-7 on TGF-β1-induced epithelial-to-mesenchymal transition in primary and immortalized human proximal tubule epithelial cells. (A) BMP-7, when administered over the concentration range µg/ml in combination with TGF-β1 for 24 h, failed to attenuate changes in e-cadherin, vimentin, CTGF and TGF-β1 transcript expression associated with TGF-β1-induced epithelial-to-mesenchymal transition in primary cultured human proximal tubule epithelial cells (RPTEC). (B) BMP-7, when administered over the concentration range µg/ml, failed to attenuate changes in e-cadherin, vimentin, CTGF and TGF-β1 transcript expression associated with TGF-β1-induced epithelial-to-mesenchymal transition in an immortalized human proximal tubule epithelial cell line (HK-2). Values are means ± 1SE. Significantly different from paired controls (P < 0.05). expression was increased in both RPTEC (Figure 3B) and HK-2 cells (Figure 3D) compared to untreated controls (Figure 3A, C) following 48 h treatment with TGF-β1. We next performed studies to examine the prospective efficacy of BMP-7 in attenuating TGF-β1-induced EMT as was shown previously in rodent proximal tubule cells [16]. Same as above, treatment of RPTEC (Figure 4A) and HK-2 cells (Figure 4B) with TGF-β1 for 24 h diminished e- cadherin transcript expression while upregulating vimentin, CTGF and TGF-β1 transcripts indicating the onset of EMT. BMP-7, when administered over a broad concentration range in combination with TGF-β1, failed to prevent the TGF-β1 effect on e-cadherin, vimentin, CTGF and TGF-β1 transcript expression in either RPTEC or HK-2 cells, but rather in some instances appeared to exacerbate the TGFβ1 insult. Moreover, the complete removal of TGF-β1 after 24 h treatment and addition of BMP-7 for an additional 24 h failed to alleviate TGF-β1-induced EMT (data not shown). We next examined whether BMP-7 was able to counteract TGF-β1-induced EMT at the level of ZO-1 protein expression. Same as above, 48 h TGF-β1 treatment of RPTEC (Figure 5B) or HK-2 cells (Figure 5E) substantially diminished ZO-1 immunoreactivity and the cell monolayer organization compared to untreated cells (Figure 5A, D). BMP-7, when administered in combination with TGF-β1 for the 48 h period, failed to prevent the TGF-β1 effects on ZO-1 protein expression and cell monolayer integrity (Figure 5C,F). Furthermore, the complete removal of TGFβ1 after 48 h and replacement with BMP-7 for 48 h more failed to restore the normal epithelial cell phenotype and monolayer organization (data not shown).

6 6 P. L. Dudas et al. Fig. 5. ZO-1 protein expression in BMP-7-treated primary and immortalized human proximal tubule epithelial cells following TGF-β1-induced epithelial-to-mesenchymal transition. (A) Untreated primary human proximal tubule cells (RPTEC) exhibiting robust ZO-1 immunoreactivity at cell cell junctions. (B) Treatment with TGF-β1 (10 ng/ml, 48 h) resulted in a downregulation of ZO-1 immunoreactivity and disruption of cell monolayer integrity. (C) Addition of BMP-7 (100 ng/ml) failed to re-establish normal ZO-1 immunoreactivity and normal epithelial organization. (D) Untreated immortalized human proximal tubule cells (HK-2) exhibiting ZO-1 immunoreactivity at cell cell junctions, similar to RPTEC. (E) Following TGF-β1 treatment (3 ng/ml, 48 h), ZO-1 immunoreactivity was substantially diminished, cell organization was disrupted and cells began migrating over the cell monolayer (exhibited by cell ghosts ). (F) BMP-7 treatment failed to re-establish a normal epithelial organization when administered under similar conditions as with RPTEC, but rather appeared to exacerbate the TGF-β1 insult. Magnification: 140. Bars indicate 50 µm. Because BMP-7 augmented TGF-β1-mediated gene expression (Figure 4A, B), we examined the effect of BMP-7 alone on RPTEC (Figure 6A) and HK-2 cells (Figure 6B). In RPTEC, BMP-7 alone significantly upregulated vimentin (1 µg/ml BMP-7), CTGF (1, 10 µg/ml BMP-7) and TGF-β1 (0.1, 1, 10 µg/ml BMP-7) transcript expression with a trend towards decreased e-cadherin transcript expression. In HK-2 cells BMP-7 significantly downregulated e-cadherin transcript expression at all concentrations tested and upregulated vimentin (1, 10 µg/ml BMP-7), CTGF (0.01, 1, 10 µg/ml BMP-7) and TGF-β1 (1, 10 µg/ml BMP-7). We next investigated if BMP-7 was capable of inhibiting TGF-β1-induced e-cadherin downregulation in a mouse tubular epithelial cell line (TCMK-1) under similar conditions as described above for human proximal tubule cells. As shown in Figure 7, 24 h treatment of mouse TCMK-1 cells with TGF-β1 induced a downregulation of e-cadherin similar to human proximal tubule epithelium. BMP-7 counteracted the TGF-β1 effect at 100 µg/ml and furthermore stimulated e-cadherin expression to above the level of untreated cells. We next sought to determine the capacity of BMP-7 to stimulate MET in fibroblasts as demonstrated previously by Zeisberg and colleagues [17]. Treating NRK-49F rat renal fibroblasts for 24 h with a broad concentration range of BMP-7 failed to downregulate vimentin transcript expression or induce the upregulation of e-cadherin transcript expression to detectable levels (Figure 8). Moreover, following treatment with BMP-7, TGF-β1 and type I and IV collagen transcripts were increased in comparison to untreated fibroblasts. Next, immunoblotting for phosphorylated SMAD 1/5/8 protein in RPTEC and HK-2 cells was performed to confirm BMP receptor activation and initiation of corresponding intracellular signalling events following BMP- 7 treatment. In BMP-7-treated RPTEC and HK-2 cells, phospho-smad 1/5/8 protein was highly expressed compared to untreated control cells that exhibited negligible basal expression (Figure 9). As an additional control for BMP-7 activity, BMP-7-mediated BMP receptor expression and BMP-7-mediated cytokine expression and secretion were examined. As shown in Figure 10A, treating RPTEC and HK-2 cells with BMP-7 downregulated IL-6 and upregulated BMPR-IA (ALK3) transcript expression. Additionally, in HK-2 cells BMP-7 inhibited IL-6 (Figure 10B) and IL-8 (Figure 10C) secretion in a dose-dependent manner. These results are supportive of a previous study characterizing BMP-7 modulation of cytokine/chemokine transcript expression in proximal tubule epithelial cells [18]. These data indicate BMP-7-mediated activation of its cognate receptors and initiation of downstream signalling and proper responsiveness to BMP-7 treatment. Discussion In the present study, the prospective efficacy of BMP-7 in attenuating TGF-β1-induced EMT in human proximal

7 BMP-7 fails to attenuate TGF-β1-induced EMT 7 Fig. 6. Effect of BMP-7 alone on gene expression in primary and immortalized human proximal tubule epithelial cells. BMP-7, when administered over the concentration range µg/ml for 24 h, altered e-cadherin, vimentin, CTGF and TGF-β1 transcript expression in both primary cultured RPTEC (A) and immortalized HK-2 (B) human proximal tubule epithelial cells. Values are means ± 1SE. Significantly different from paired controls (P < 0.05). tubule epithelium was examined. It was first confirmed in both the primary RPTEC and immortalized HK-2 cells that TGF-β1 treatment was capable of inducing an epithelialto-mesenchymal phenotypic transition. In both cell lines in response to TGF-β1, there was a robust downregulation of e-cadherin transcript expression indicating loss of the epithelial phenotype concomitant with a substantial upregulation of markers defining a mesenchymal/pro-fibrotic phenotype, including vimentin and CTGF. Additionally, TGF-β1 transcript was upregulated in both cell lines in response to exogenous TGF-β1 treatment, suggesting that autocrine/paracrine TGF-β1 signalling potentiates the phenotypic transition. EMT was readily prevented in HK-2 cells with a neutralizing antibody specific for TGF-β1, indicating that the phenotypic transition was an inhibitable process. In concert with the examination of transcript expression for epithelial or mesenchymal markers, expression of α-sma protein and the tight-junction protein ZO-1 were examined by immunofluorescence microscopy as indicators of the mesenchymal transition and cell cell junction integrity and maintenance of epithelial differentiation, respectively. Following TGF-β1 treatment, α-sma expression was substantially increased in both RPTEC and HK-2 cells indicating a pro-mesenchymal phenotypic transition. Additionally, the robust expression of ZO-1 exhibited by untreated RPTEC and HK-2 cells at cell cell junctions was considerably diminished following TGF-β1 treatment. HK-2 cells also assumed a more spindle-shaped morphology further indicative of transition to a mesenchymal phenotype. Pre-incubating TGF-β1 with its respective neutralizing antibody prevented TGF-β1-induced downregulation of ZO-1 in both RPTEC and HK-2 cells and the associated morphological changes observed in HK-2 cells. The inhibition of ZO-1 downregulation and maintenance of normal epithelial organization reinforce the concept that EMT in RPTEC and HK-2 cells can be inhibited. We next sought to determine if BMP-7 was able to exert a protective effect in human RPTEC and HK-2 cells.

8 8 P. L. Dudas et al. Fig. 9. BMP-7-induced SMAD 1/5/8 phosphorylation in primary and immortalized human proximal tubule epithelial cells. Immunoblot analysis of BMP-7 treated primary (RPTEC) and immortalized (HK-2) human proximal tubule epithelial cells (200 ng/ml, 20 min) versus untreated cells demonstrating an upregulation of phospho-smad 1/5/8 protein in treated cells only, confirming receptor activation and proper downstream signalling in response to BMP-7. Fig. 7. Effect of BMP-7 on TGF-β1-induced e-cadherin downregulation in mouse renal tubular epithelial cells (TCMK-1). Treating mouse TCMK-1 cells with TGF-β1 (10 ng/ml, 24 h) significantly decreased e-cadherin transcript expression. 100 µg/ml BMP-7, when administered in combination with TGF-β1 inhibited TGF-β1-induced e-cadherin downregulation µg/ml BMP-7 failed to attenuate the TGF-β1 effect. Values are means ± 1SE. Significantly different from paired controls (P < 0.05). # Significantly different from TGF-β1 treated cells (P < 0.05). Fig. 8. Effect of BMP-7 on renal fibroblast mesenchymal-to-epithelial transition. Treating rat renal fibroblasts (NRK-49F) with BMP-7 ( µg/ml, 24 h) failed to induce the upregulation of e-cadherin transcript expression and downregulation of vimentin transcript expression. Additionally, TGF-β1, Col Iα1 and Col IVα1 transcript levels were elevated above untreated control cells following BMP-7 treatment. Values are means ± 1SE. Significantly different from paired controls (P < 0.05). TGF-β1 consistently induced EMT as determined by those criteria described above; however, BMP-7 when administered concurrently with TGF-β1 failed to alleviate the process. In the primary RPTEC cells, with increasing concentrations of BMP-7 there was an exacerbation of the TGF-β1-mediated pro-fibrotic effect at all BMP-7 concentrations for all genes examined with the exception of TGF-β1 transcript expression at 1 µg/ml BMP-7. Similarly, in HK-2 cells, BMP-7 augmented TGF-β1-induced EMT with the exception of CTGF expression at 100 µg/ml BMP-7 and TGF-β1 expression at µg/ml BMP-7. For both RPTEC and HK-2 cells, in those treatments where TGF-β1 did not have a significant pro-fibrotic effect, there still existed a trend towards exacerbation of EMT compared to untreated controls. Moreover, there appeared to be no BMP-7-mediated protection from TGF-β1-induced downregulation of ZO-1 protein expression and alterations in cell monolayer integrity. Regardless of whether BMP-7 was administered in combination with TGF-β1 or subsequent to the complete removal of TGF-β1, there was no prevention of or recovery from EMT (data not shown). Of important note, through the course of the present study BMP-7 activity was verified in both RPTEC and HK-2 cells through the confirmation of receptor modulation by BMP- 7 and activation of corresponding downstream signalling events and BMP-7 regulation of cytokine transcript expression and cytokine secretion. In a recent report, Motazed and colleagues demonstrated BMP-7-mediated reduction in TGF-β1-stimulated fibronectin production in immortalized HKC-8 human proximal tubule cells [19]. This finding supports BMP-7 anti-fibrotic activity in human proximal tubule epithelium, contrary to what is presented in the present study. Motazed and colleagues did not, however, examine e-cadherin, vimentin, CTGF or TGF-β1 expression nor did the present study address alterations in fibronectin metabolism. Furthermore, Motazed and colleagues did not address if BMP- 7 exhibited the same inhibitory activity on TGF-β1-induced fibronectin production in the primary human proximal tubule cells that were utilized for other experiments in their study. These factors, in turn, make it difficult to reconcile the protective effect observed in their study with the fibrogenic effects observed in ours. It is possible that HKC-8 cells do not respond to TGF-β1 with the same gene expression profile as RPTEC and HK-2 cells or they potentially do but are not responsive to BMP-7-mediated protection, similar to RPTEC and HK-2 cells. Alternatively, RPTEC and HK-2 cells may exhibit greater sensitivity to BMP- 7 with regard to TGF-β1-induced fibronectin expression. This issue requires further characterization of the different cell types utilized in the two studies.

9 BMP-7 fails to attenuate TGF-β1-induced EMT 9 Fig. 10. BMP-7-mediated cytokine expression/secretion and BMP receptor regulation in primary and immortalized human proximal tubule epithelial cells. (A) BMP-7 treatment ( µg/ml, 24 h) of primary (RPTEC) and immortalized (HK-2) human proximal tubule epithelial cells induced a downregulation of IL-6 transcript expression and upregulation of BMPR-IA (ALK3) transcript expression indicating the cells are properly responsive to BMP receptor activation. Values are means ± 1SE. Significantly different from paired controls (P < 0.05). (B, C) In immortalized human proximal tubule epithelial cells (HK-2) BMP-7 inhibited IL-6 and IL-8 secretion in a dose-dependent manner indicating proper responsiveness to BMP receptor activation. Because BMP-7 appeared to augment TGF-β1-induced EMT in RPTEC and HK-2 cells, the effect of BMP-7 alone on the proximal tubule epithelium was examined. BMP-7 was applied to the cells separately from TGF-β1 to discern if by itself it was able to stimulate an anti-fibrotic (proepithelial) response. Treatment with BMP-7 alone in fact stimulated a fibrogenic response in the epithelium similar to that observed with TGF-β1 and TGF-β1 + BMP-7. In RPTEC, BMP-7 upregulated vimentin, CTGF and TGFβ1 transcript expression. There was additionally a trend towards downregulating e-cadherin transcript at all BMP-7 concentrations although this effect was not significant. In HK-2 cells, BMP-7 downregulated e-cadherin transcript expression and upregulated vimentin, CTGF and TGF-β1. As we were unable to demonstrate a protective effect with regard to the inhibition of TGF-β1-induced EMT in the proximal tubule epithelium and observed the pro-fibrogenic effect with BMP-7 alone, we sought to determine if BMP-7 might facilitate MET in renal fibroblasts as an alternate mechanism for mediating protection. Previously, Zeisberg and colleagues [17] demonstrated BMP-7-induced MET in adult human renal fibroblasts and suggested this as a means by which BMP-7 can mediate renoprotection and repair. Treatment of rat renal fibroblasts with increasing concentrations of BMP-7 did not upregulate e-cadherin, as shown previously in human renal fibroblasts, indicating that there was no attainment of an epithelial phenotype, but rather stimulated an increase in vimentin, TGF-β1, Col Iα1 and Col IVα1 transcript expression indicating that the cells were upregulating their fibrogenic activity. This finding may in fact support an earlier study by Ikeda and colleagues [20] examining BMP-7 efficacy in a rat model of overload proteinuria. It was found that compared to control animals the BSA + BMP-7 group had higher total kidney mrna levels for Col IIIα1 and fibronectin. There was however a parallel non-significant decrease in interstitial fibrosis but ultimately no effect of proteinuria. Another study utilizing cultured human hepatic stellate cells and tissue from patients with chronic liver disease also observed BMP-7 pro-fibrogenic activity [21]. In this study, both systemic and hepatic BMP-7 levels were found to be elevated in patients with chronic liver disease and in vitro

10 10 P. L. Dudas et al. experiments utilizing cultured human hepatic stellate cells demonstrated BMP-7 increased proliferation and increased Col Iα1 and fibronectin synthesis. These data would indicate that BMP-7 might enhance human hepatic stellate cell activity and promote liver fibrogenesis in vivo. Certain factors may be attributable to the differences in findings between previous studies demonstrating BMP-7- mediated inhibition of EMT and stimulation of MET and the present study in which the protective effects were not observed. Under normophysiologic conditions, renal BMP- 7 expression differs between rodents and humans. In rodent kidneys BMP-7 expression has been localized to glomerular visceral epithelial cells (podocytes), outer medullary distal convoluted tubule, thick ascending limb and collecting duct [18]. In humans, BMP-7 and phosphorylated SMAD 1/5/8 proteins were detected in the distal nephron and collecting duct only [22]. Additionally, BMP-7 was immunolocalized to the luminal side of the distal nephron tubules. These data would suggest that in humans there is urinary excretion of BMP-7 and that BMP-7 potentially plays a physiological role in the distal nephron. In rodents BMP-7 may function in both the proximal and distal nephron. Both rodent and human renal epithelia express BMP receptors and are responsive to BMP-7 treatment [23,24], so the lack of BMP- 7-mediated protection in human renal tubular cells compared to rodent may be the result of reduced responsiveness reflecting normal cell function typically observed in the in vivo cellular microenvironment. Supportive of this idea is our observation in TCMK-1 mouse renal tubular cells in which BMP-7 (100 µg/ml) inhibited TGFβ1-induced downregulation of e-cadherin transcript expression when administered under similar conditions as the human proximal tubule epithelium. This and the majority of other studies describing BMP- 7 activity with specific human cell types have done so utilizing isolated cell systems that are not entirely reflective of in vivo physiologic states. Cell systems that can more closely emulate in vivo conditions (transwell, 3D culture) may better demonstrate normal BMP-7 activity with respect to a specific cell type. Additionally, renal and systemic BMP-7 levels and translational studies describing BMP-7 activity on specific nephron cell types in the context of human chronic kidney disease need further investigation. These studies may ultimately aid in clarifying the role of BMP-7 in normophysiologic and pathophysiologic states. Acknowledgements. assistance. The authors thank Helen Gao for excellent technical Conflict of interest statement. All authors are employees of Centocor Research & Development, a Johnson & Johnson company. During the course of this study BMP-7 was in-licensed to Centocor/Johnson & Johnson from Curis, Inc. This licensing agreement is no longer in place. References 1. Mason RM, Wahab NA. Extracellular matrix metabolism in diabetic nephropathy. J Am Soc Nephrol 2003; 14: Wolf G, Ziyadeh FN. Cellular and molecular mechanisms of proteinuria in diabetic nephropathy. Nephron Physiol 2007; 106: Ferenbach D, Kluth DC, Hughes J. Inflammatory cells in renal injury and repair. Semin Nephrol 2007; 27: Fogo AB. Mechanisms of progression of chronic kidney disease. Pediatr Nephrol 2007; 22: Strutz F, Muller GA. Interstitial pathomechanisms underlying progressive tubulointerstitial damage. Kidney Blood Press Res 1999; 22: Kuncio GS, Neilson EG, Haverty T. Mechanisms of tubulointerstitial fibrosis. Kidney Int 1991; 39: Iwano M, Plieth D, Danoff TM et al. Evidence that fibroblasts derive from epithelium during tissue fibrosis. J Clin Invest 2002; 110: Rastaldi MP, Ferrario F, Giardino L et al. Epithelial-mesenchymal transition of tubular epithelial cells in human renal biopsies. Kidney Int 2002; 62: Nicolas FJ, Lehmann K, Warne PH et al. Epithelial to mesenchymal transition in Madin-Darby canine kidney cells is accompanied by down-regulation of Smad3 expression, leading to resistance to transforming growth factor-beta-induced growth arrest. JBiolChem2003; 278: Strutz F, Muller GA. Transdifferentiation comes of age. Nephrol Dial Transplant 2000; 15: Simon M, Maresh JG, Harris SE et al. Expression of bone morphogenetic protein-7 mrna in normal and ischemic adult rat kidney. Am JPhysiol1999; 276(Pt 2): F382 F Jena N, Martin-Seisdedos C, McCue P et al. BMP7 null mutation in mice: developmental defects in skeleton, kidney, and eye. Exp Cell Res 1997; 230: Wang SN, Lapage J, Hirschberg R. Loss of tubular bone morphogenetic protein-7 in diabetic nephropathy. J Am Soc Nephrol 2001; 12: Wang S, Chen Q, Simon TC et al. Bone morphogenic protein-7 (BMP-7), a novel therapy for diabetic nephropathy. Kidney Int 2003; 63: Morrissey J, Hruska K, Guo G et al. Bone morphogenetic protein-7 improves renal fibrosis and accelerates the return of renal function. J Am Soc Nephrol 2002; 13(Suppl 1): S14 S Zeisberg M, Hanai J, Sugimoto H et al. BMP-7 counteracts TGFbeta1-induced epithelial-to-mesenchymal transition and reverses chronic renal injury. Nat Med 2003; 9: Zeisberg M, Shah AA, Kalluri R. Bone morphogenic protein-7 induces mesenchymal to epithelial transition in adult renal fibroblasts and facilitates regeneration of injured kidney. JBiolChem2005; 280: Gould SE, Day M, Jones SS et al. BMP-7 regulates chemokine, cytokine, and hemodynamic gene expression in proximal tubule cells. Kidney Int 2002; 61: Motazed R, Colville-Nash P, Kwan JT et al. BMP-7 and proximal tubule epithelial cells: activation of multiple signaling pathways reveals a novel anti-fibrotic. Mechanism Pharm Res 2008; 25: Ikeda Y, Ok Jung Y, Kim H et al. Exogenous bone morphogenetic protein-7 fails to attenuate renal fibrosis in rats with overload proteinuria. Neprhon Exp Nephrol 2004; 97: e123 e Tacke F, Gabele E, Bataille F et al. Bone morphogenetic protein 7 is elevated in patients with chronic liver disease and exerts fibrogenic effects on human hepatic stellate cells. Dig Dis Sci 2007; 52: Wetzel P, Haag J, Campean V et al. Bone morphogenetic protein-7 expression and activity in the human adult normal kidney is predominantly localized to the distal nephron. Kidney Int 2006; 70: Bosukonda D, Shih MS, Sampath KT et al. Characterization of receptors for osteogenic protein-1/bone morphogenetic protein-7 (OP-1/ BMP-7) in rat kidneys. Kidney Int 2000; 58: Kitten AM, Kreisberg JI, Olson MS. Expression of osteogenic protein-1 mrna in cultured kidney cells. J Cell Physiol 1999; 181: Received for publication: Accepted in revised form: