Biliverdin Reductase Mediates Hypoxia-Induced EMT Via PI3-Kinase and Akt

JASN Express. Published on January 9, 2008 as doi: 10.1681/ASN.2006111194

CLINICAL RESEARCH www.jasn.org

Biliverdin Reductase Mediates Hypoxia-Induced EMT via PI3-Kinase and Akt

Rui Zeng,* Ying Yao,* Min Han,* Xiaoqin Zhao,* Xiao-Cheng Liu,* Juncheng Wei,† Yun Luo,* Juan Zhang,* Jianfeng Zhou,† Shixuan Wang,† Ding Ma,† Gang Xu*

*Division of Nephrology, Department of Internal Medicine; †Cancer Biology Research Center, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology Wuhan, Hubei, People’s Republic of China

ABSTRACT Chronic hypoxia in the renal parenchyma is thought to induce epithelial-to-mesenchymal transition (EMT), leading to fibrogenesis and ultimately end-stage renal failure. Biliverdin reductase, recently identified as a serine/threonine/tyrosine kinase that may activate phosphatidylinositol 3-kinase (PI3K) and Akt, is upregulated in response to reactive oxygen species that may accompany hypoxia. We investigated this potential role of biliverdin reductase in hypoxia-induced renal tubular EMT. Expression of biliverdin reductase was upregulated in a human proximal tubule cell line (HK-2) cultured in hypoxic conditions (1% O2), and this was accompanied by reduced expression of E-cadherin and increased expression of the mesenchymal marker vimentin. Inhibiting PI3K reversed these changes, consistent with EMT. In normoxic conditions, overexpression of biliverdin reductase promoted similar characteristics of EMT, which were also reversed by inhibiting PI3K. Furthermore, using small interfering RNA (siRNA) to knockdown biliverdin reductase, we demonstrated that the enzyme associates with phosphorylated Akt and mediates the hypoxia-induced EMT phenotype. In vivo, expression of biliverdin reductase increased in the tubular epithelia of 5/6-nephrectomized rats, and immunohistochemistry of serial sections demonstrated similar localization of phosphorylated Akt and biliverdin reductase. In conclusion, biliverdin reductase mediates hypoxia-induced EMT through a PI3K/Akt-dependent pathway.

J Am Soc Nephrol ●●: –, 2008. doi: 10.1681/ASN.2006111194

Biliverdin reductase (BVR) is an evolutionarily transcriptional repression initiates the conversion conserved soluble enzyme found in various spe- of epithelial cells to mesenchymal cells.12,13 Thus, it cies.1,2 Initially, BVR was solely considered a reduc- is intriguing to speculate that BVR may cause epi- tase that converts biliverdin to bilirubin in the heme degradation pathway.3,4 However, recently, BVR was identified as a serine/threonine/tyrosine ki- Received November 3, 2006. Accepted August 31, 2007. nase5,6 that modulates signal transduction path- 2 7–10 Published online ahead of print. Publication date available at ways and regulates gene expression. Through www.jasn.org. analysis of the primary and secondary structures of BVR, Maines et al. determined that it has two po- R.Z. and Y.Y. contributed equally to this work. tential Src homology2 (SH2) protein docking sites Correspondence: Prof. Gang Xu, Tongji Hospital, Huazhong Uni- versity of Science and Technology, 1095 Jiefang Ave., Wuhan, Hu- that could interact with the SH2 domain-contain- bei 430030, People’s Republic of China. Phone: 86-27-83662848; ing protein PI3K.2,6,11 These features suggest that Fax: 86-27-83662681; E-mail: [email protected] or Prof. Ding BVR is a potential activator of PI3K, which may Ma, Cancer Biology Research Center, Tongji Hospital, Tongji Med- ical College, Huazhong University of Science and Technology, 1095 bind PI3K like a Src family member. The down- Jiefang Ave., Wuhan, Hubai 430030, People’s Republic of China. stream effector of PI3K is the oncogenic serine/ Phone: 86-27-83663351; Fax: 86-27-83662681; E-mail: threonine kinase Akt, which represses transcription [email protected] of the cell adhesion molecule E-cadherin.12 This Copyright © 2008 by the American Society of Nephrology

J Am Soc Nephrol ●●: –, 2008 ISSN : 1046-6673/●●00- 1 CLINICAL RESEARCH www.jasn.org thelial-to-mesenchymal transition via the PI3K/Akt pathway. BVR is also an oxidative stress protein that is activated in response to free radicals.5,7,14 Reactive oxygen species, espe- cially hydroxyl radicals, elicit increases in the expression of BVR protein in vivo and in vitro.7,15 Accumulating evidence indicates that the formation of hydroxyl radicals is stimulated by hypoxia,16–18 suggesting that hypoxia can upregulate the expression of BVR. Recent studies demonstrate that chronic hypoxia-induced EMT is a final common pathway to end-stage renal failure.19,20 However, the mechanisms underlying hypoxia-induced renal fibrogenesis are not fully clear. Considering the known features of BVR, we hypothesized that BVR plays an important role in hypoxia-induced renal tubular EMT. In the study presented here, we characterized the effect of BVR on the hypoxia-induced tubular EMT and investigated the BVR pathways that are involved.

Figure 2. (A) Western blot analysis of E-cadherin, vimentin, and hBVR in parental cells, pcDNA3.1/Zeo empty vector-transfected RESULTS cells, and hBVR-transfected cells. A representative blot from four independent experiments is shown. The histogram shows the Hypoxia Increases the Expression of BVR In Vitro average volume density corrected for the loading control, ␤-actin We induced hypoxia in vitro by incubating HK-2 cells in 1% O2 (n ϭ 4). *P Ͻ 0.05 compared with normoxic controls. (B) Morpho- and 5% CO2 from 12 to 48 h. Twelve hours after hypoxic stim- logic changes in cells. Both parental and pcDNA3.1/Zeo empty ulation human BVR (hBVR) was markedly upregulated (Fig- vector-transfected cells showed a typical cuboidal epithelial ure 1). hBVR remained upregulated for more than 48 h and shape, whereas hBVR-transfected cells were elongated in shape was consistent with the increase in hypoxia-inducible factor 1␣ and larger than control cells, consistent with the morphology of ϫ (HIF-1␣; Figure 1). In normoxic control cells there was little myofibroblasts. Magnification, 200. expression of hBVR and HIF-1␣ (Figure 1). shape and larger than control cells (Figure 2B). E-cadherin, an hBVR Induces EMT in hBVR-Transfected HK-2 Cells epithelial marker, was downregulated in hBVR-transfected There was no difference in protein expression between paren- cells compared with parental cells and empty vector-trans- tal cells and cells transfected with pcDNA3.1/Zeo empty vec- fected cells (Figure 2A). In contrast, the protein vimentin, a tor–controls for hBVR-transfected cells (Figure 2A). Expres- mesenchymal marker, was upregulated in hBVR-transfected sion of the hBVR protein was 1.6-fold greater in cells cells as compared with control cells (Figure 2A). Confocal laser transfected with the hBVR gene compared with control cells microscopy demonstrated that E-cadherin was linearly local- (Figure 2A). Light-inverted microscopic examination revealed ized at cell borders in parental cells and empty vector-trans- that parental cells and empty vector-transfected cells formed a fected cells (Figure 3C (I and V)). In hBVR-transfected cells, cobblestone-like monolayer, a typical epithelial shape. hBVR- E-cadherin formed a zipper-like pattern instead of a linear pat- transfected cells also formed an epithelial monolayer, but there tern (Figure 3 C (II)). were larger spaces between cells, and the cells were elongated in In clones of BVR-transfected cells in which hBVR levels were increased more than 1.6-fold, we found more profound effects on EMT (data not shown). However, as shown in Figure 1, in hypoxic HK-2 cells, the expression of hBVR increased together with the duration of hypoxia, but remained approxi- mately 1.5-fold greater at hypoxic time points (12, 24, and 48 h of hypoxia). Hence, we choose the 1.6-fold clone for the fol- lowing experiment.

PI3K Inhibitors Block the EMT Caused by Transfection with hBVR Figure 1. Western blot analysis of hypoxia-inducible factor 1␣ (HIF-1␣) and human biliverdin reductase (hBVR) in HK-2 cells after Recently, hBVR was determined to have tyrosine kinase ac- 12, 24, and 48 h of hypoxia. A representative blot from four tivity and there are 2 potential SH2 protein docking sites for 2,6 independent experiments is shown. The histogram shows the PI3K in hBVR, which suggests that hBVR is a potential average volume density corrected for the loading control, ␤-actin agonist of PI3K. Hence, we detected the phosphorylation (n ϭ 4). *P Ͻ 0.05 compared with normoxic controls. status and kinase activity of Akt in hBVR-transfected cells.

2 Journal of the American Society of Nephrology J Am Soc Nephrol ●●: –, 2008 www.jasn.org CLINICAL RESEARCH

Figure 3. (A) Western blot analysis of the phosphorylation status of Akt in parental cells, hBVR-transfected cells, Ly294002-treated Figure 4. (A) Western blot analysis of phosphorylation status of hBVR-transfected cells, wortmannin-treated hBVR-transfected Akt in normoxic control cells, 48-h hypoxic cells, 48-h hypoxic cells, and pcDNA3.1/Zeo empty vector-transfected cells. (B) cells treated with Ly294002, and 48-h hypoxic cells treated with Western blot analysis of E-cadherin, vimentin, and hBVR in the wortmannin. (B) Western blot analysis of E-cadherin, vimentin, same groups shown in (A). In both panels, a representative blot of and hBVR in the same groups shown in (A). Both panels show a four independent experiments is shown in A and B. The histo- representative blot from four independent experiments. The his- gram shows the average volume density corrected for the loading togram shows the average volume density corrected for the load- ␤ control, -actin (phosphorylated Akt was corrected for the loading control, ␤-actin (phosphorylated Akt was corrected for the ϭ Ͻ ing control, total Akt) (n 4). *P 0.05 compared with control loading control, total Akt) (n ϭ 4). *P Ͻ 0.05 compared with # Ͻ cells. P 0.05 compared with hBVR-transfected cells. (C) Con- control cells. #P Ͻ 0.05 compared with hBVR-transfected cells. (C) focal microscopic images of E-cadherin in monolayer cells (I, Confocal microscopic images of E-cadherin in monolayer cells (I, parental cells; II, hBVR-transfected cells; III, hBVR-transfected cells normoxic control cells; II, 48-h hypoxic cells; III, Ly294002-treated treated with Ly294002 for 48 h; IV, hBVR-transfected cells treated 48-h hypoxic cells; IV, wortmannin-treated 48-h hypoxic cells). with wortmannin for 48 h; V, pcDNA3.1/Zeo empty vector-trans- Cell nuclei were enhanced by staining of cell nuclei with PI. fected cells). Cell nuclei were enhanced by staining with pro- Magnification, ϫ400. pidium iodine (PI). Magnification, ϫ400.

As shown in Figure 3A, hBVR transfection significantly in- Ly294002 and wortmannin. After 48-h hypoxia, the level of creased the phosphorylation of Akt. Treatment with hBVR was also increased (Figure 4A). The protein E-cad- Ly294002 and wortmannin, inhibitors of PI3K,21 decreased herin was significantly reduced compared with levels in nor- phosphorylation of AKT by 36% (Ly294002) and 32% moxic cells, whereas the protein vimentin was upregulated (wortmannin) (Figure 3A). In the treated cells, E-cadherin in hypoxic cells (Figure 4B). Confocal microscopy demon- protein was increased and vimentin protein was decreased strated that E-cadherin changed from a linear pattern in (Figure 3B), and the morphologic transition was blocked. normoxic cells to a zipper-like pattern at the junctions be- After treatment, E-cadherin relocalized to cell junctions tween hypoxic cells (Figure 4C (I and II)). These changes are (Figure 3C (III and IV)). Neither Ly294002 nor wortmannin all characteristic of myofibroblasts. Treatment with changed the hBVR values in hBVR-transfected cells (Figure Ly294002 or wortmannin reversed the increase in E-cad- 3B). Collectively, these results suggest that increased phos- herin to the level present in normoxic cells (Figure 4B). phorylation of Akt in hBVR- transfected cells promotes After treatment, E-cadherin also reappeared at cell borders phenotypic changes associated with EMT. and E-cadherin distribution changed from a zipper-like pattern to a linear pattern (Figure 4C (III and IV)). Expres- BVR Contributes to Hypoxia-Induced EMT via the sion of vimentin was correspondingly reduced in hypoxic PI3K/Akt Pathway cells after treatment with Ly294002 and wortmannin (Fig- As shown in Figure 4A, phosphorylation of Akt was in- ure 4B). However, neither Ly294002 nor wortmannin af- creased after 48-h hypoxia, whereas it was inhibited by fected the hBVR values in hypoxic cells (Figure 4B).

J Am Soc Nephrol ●●: –, 2008 Biliverdin Reductase Accelerates Hypoxia-Mediated Renal Tubular Epithelial to Mesenchymal Transition 3 CLINICAL RESEARCH www.jasn.org

BVR Depletion Results in Decreased Phosphorylation of Akt and Inhibition of Hypoxia-Induced EMT siRNA experiments were performed to decrease the activity of hBVR in cultured HK-2 cells. As shown in Figure 5A, in normoxic cells, silencing of hBVR was evident from 48 to 96 h after transfection and reached 60 to 80% reduction. Similar results were obtained in hypoxic HK-2 cells (Figure 5B). After 48 or 96 h transfection, the amount of hBVR in hypoxic cells was almost below the normal level. To assess the effect of hBVR knockdown on hypoxia-induced renal EMT, 48 h after siRNA transfection the cells were moved to a hypoxic incubator (1%

O2,5%CO2, 37°C) for another 48 h-incubation. The mock- transfected HK-2 cells and the cells transfected with control siRNA at 48 h served as negative controls and were also shifted to hypoxic incubation. As shown in Figure 6, the negative control cells developed Figure 6. (A) Western blot analysis of E-cadherin, vimentin, similar renal EMT after 48-h hypoxia and demonstrated hBVR, and phosphorylated Akt in normoxic cells, normoxic con- downregulated E-cadherin levels and upregulated vimentin trol-siRNA cells, normoxic hBVR-siRNA cells, 48-h hypoxic cells, protein levels. The HK-2 cells transfected with hBVR-specific 48-h hypoxic control-siRNA cells, and 48-h hypoxic hBVR-siRNA siRNA demonstrated resistance to hypoxia-induced EMT. cells. A representative blot from three independent experiments This was evident because increased levels of E-cadherin protein is shown. The histogram shows the volume density of the repre- ␤ (Figure 6A), a relocalization of E-cadherin protein at cell bor- sentative blot corrected for the loading control, -actin (phos- ders (Figure 6B), and reduced vimentin expression compared phorylated Akt was corrected for the loading control, total Akt). with negative controls (Figure 6A). Phosphorylated Akt was (B) Confocal microscopic images of E-cadherin in monolayers of normoxic cells, normoxic control-siRNA cells, normoxic hBVR- decreased after hBVR siRNA, with or without hypoxia, relative siRNA cells, 48-h hypoxic cells, 48-h hypoxic control-siRNA cells, to levels in control cells with unsilenced hBVR, whereas the and 48-h hypoxic hBVR-siRNA cells. Cell nuclei were enhanced by total amount of Akt protein remained unchanged (Figure 6A). staining of cell nuclei with PI. Magnification, ϫ400. This result further confirms that BVR is an upstream activator of Akt. In normoxic HK-2 cells, although hBVR-specific bation. This suggests that hBVR inhibition does not affect the siRNA inhibited the phosphorylation of Akt, the mock trans- physiologic activity of normoxic HK-2 cells. Taken together fected, control siRNA transfected, and hBVR siRNA knock- with the findings from our experiments treating hypoxic cells down cells all demonstrated similar expression of E-cadherin with PI3K inhibitors, these results suggest that BVR-induced and absence of expression of vimentin after the full 96-h incu- EMT depends on its phosphorylation of Akt.

BVR, through the Akt Pathway, Participates in Hypoxia-Induced Renal Fibrosis In Vivo The 5/6 nephrectomy is not considered a standard model of hypoxia, but it is the classic fibrosis model. Hence, we chose to use the 5/6 subtotal nephrectomy model to investigate hypoxia-related renal fibrogenesis. As shown in Table 1, body weight did not differ between sham-operated and 5/6-nephrectomized rats. However, serum urea nitrogen and serum creatinine were significantly elevated in 5/6-nephrectomized rats compared with sham-operated rats. Also, light microscopic examination revealed glomerular sclerosis and interstitial fibrosis in 5/6-nephrectomized rats (data not shown). Immuno- histochemical analysis showed that rat BVR (rBVR) was widely Figure 5. (A) Time-dependence of hBVR silencing in normoxic expressed in the glomeruli, renal tubules, and interstitial cells HK-2 cells. HK-2 cells were transfected with small interfering RNA in 5/6-nephrectomized rats, whereas there was little staining in (siRNA), and Western blot analysis was performed 12, 24, 48, 72, sham-operated rats (Figure 7A). Western blot demonstrated or 96 h later. A representative blot from three independent experiments is shown. (B) Western blot analysis of siRNA-trans- similar results: rBVR protein was significantly elevated in the fected HK-2 cells under hypoxic conditions. The cells were trans- kidney of 5/6-nephrectomized rats compared with shams (Fig- fected with siRNA and incubated in a hypoxic incubator for 24, ure 7B). As demonstrated by Manotham et al. and Zhang et 48, or 96 h. A representative blot from three independent exper- al.,23,24 HIF-1␣ was also significantly elevated in 5/6-nephrec- iments is shown. tomized rats (Figure 7B). This result confirms that chronic

4 Journal of the American Society of Nephrology J Am Soc Nephrol ●●: –, 2008 www.jasn.org CLINICAL RESEARCH

Table 1. Renal function in sham-operated rats and 5/6-nephrectomized rats Groups n Bun (mmol/L) Scr (␮mol/L) BW (g) Sham-operated 6 6.31 Ϯ 1.04 28.54 Ϯ 5.99 412 Ϯ 11 5/6-nephrectomized 7 35.80 Ϯ 16.55a 120.11 Ϯ 63.78a 398 Ϯ 15 aP Ͻ 0.01 compared with sham-operated rats. Bun, serum urea nitrogen; Scr, serum creatinine; BW, bodyweight.

pressed in some renal epithelial tubular cells (Figure 8, A and B). In 5/6-nephrectomized rats, along with an increase in rBVR, staining for phosphorylated Akt was widely distributed in the glomeruli, renal epithelia, and interstitium; and the distribution patterns of phosphorylated Akt and rBVR were highly congruent in the renal tubules (Figure 8, C and D). Taken together with the in vitro findings, this suggests that BVR is involved in hypoxia-induced renal fibrosis via the Akt pathway.

DISCUSSION

Chronic hypoxia is a crucial process in renal fibrogenesis.22,23 Accumulating data suggest that chronic tubulointerstitial hypoxia induced by peritubular capillary loss aggravates the devel- opment of tubulointerstitial fibrosis in remnant kidneys.22–24 Tubular cells exposed to hypoxia undergo EMT and then be- come myofibroblasts, as indicated by expression of mesenchymal markers, morphologic changes, and acquisition of mobil- ity.19 The myofibroblasts transdifferentiated from tubular cells migrate to the tubular interstitium and sequentially induce renal fibrogenesis.25,26 However, the mechanisms of hypoxia- induced tubular EMT are poorly understood. In this study, we further demonstrated the occurrence of chronic hypoxia in

Figure 7. (A) Immunohistochemistry and immunoreactive area analysis for rat biliverdin reductase (rBVR) in the kidney tissue of sham-operated and 5/6-nephrectomized rats. Representative data from six pairs of rats are shown. The average signals are shown as a histogram (n ϭ 6). Magnification, ϫ200. The immunoreactive area was assessed as described in the Concise Meth- ods. (B) Western blot analysis of HIF-1␣ and hBVR in kidney tissue of sham-operated and 5/6-nephrectomized rats. A representative blot from six pairs of rats is shown. The histogram shows the average volume density corrected for the loading control, ␤-actin (n ϭ 6). *P Ͻ 0.05 compared with sham-operated rats. hypoxia occurs with 5/6 subtotal nephrectomy of the rat. In our linear regression analysis of Western blot bands of rBVR and HIF-1␣ in 5/6-nephrectomized rats (n ϭ 7), the Pearson correlation coefficient was 0.879, indicating a positive correlation (P Ͻ 0.01) between expression of HIF-1␣ and rBVR. To further confirm the relationship between hypoxia-induced EMT and activation of Akt phosphorylation by BVR, we Figure 8. Immunohistochemistry of rBVR and phosphorylated immunolocalized rBVR and phosphorylated Akt on serial sec- Akt on serial kidney sections in sham and 5/6-nephrectomized tions in the kidney of both sham and 5/6-nephrectomized rats. rats. (A) Phosphorylated Akt in the kidney of sham rats. (B) rBVR in In sham rats, although expression of phosphorylated Akt was the kidney of sham rats. (C) Phosphorylated Akt in the kidney of somewhat weak, its distribution in the renal epithelia most 5/6-nephrectomized rats. (D) rBVR in the kidney of 5/6-nephrec- closely corresponded to that of rBVR, which was locally ex- tomized rats. Magnification, ϫ100.

J Am Soc Nephrol ●●: –, 2008 Biliverdin Reductase Accelerates Hypoxia-Mediated Renal Tubular Epithelial to Mesenchymal Transition 5 CLINICAL RESEARCH www.jasn.org remnant kidneys in vivo and the occurrence of EMT of renal tubular EMT via the PI3K/Akt pathway. In addition, Mainess tubular cells after hypoxia in vitro and confirmed that chronic et al. confirmed that BVR itself has PKB/Akt-like activity,2 hypoxia can induce the expression of BVR in renal tubular cells which may theoretically be directly involved in BVR-induced and remnant kidneys. Furthermore, we investigated the role of EMT. BVR in hypoxia-induced renal tubular EMT. To clarify the confliction between BVR’s antioxidant ability BVR is a key enzyme in the biliverdin degradation pathway, and role in tubular EMT, we furthermore investigated the and it can be induced by endogenous or exogenous stimuli.7,15 physiologic reaction of BVR to hypoxia. We observed myofi- Bilirubin, the downstream product of BVR, is a potent physi- broblast-like characters in hypoxic HK-2 cells, which is consis- ologic antioxidant;27,28 this action reflects an amplification cy- tent with previous reports.19,30 We also detected the upregula- cle whereby bilirubin, acting as an antioxidant, is itself oxidized tion of BVR after hypoxia. Upon treatment with Ly294002 or to biliverdin and then physiologically regenerated by BVR.10 wortmannin, BVR remained upregulated, but the hypoxia in- Therefore, through BVR, bilirubin can protect cells from a duced tubular EMT was reversed. In pursuit of more evidence, 10 10,000-fold excess of H2O2. In experimental autoimmune we used hBVR-specific siRNA to investigate the cellular path- encephalomyelitis, a free radical-mediated disease, treatment way required for the hypoxia-induced phenotypic effects in with BVR ameliorates both clinical and pathologic forms of the our model. BVR siRNA knockdown inhibited the phosphory- disease more efficiently than do treatments with traditional lation of Akt in both hypoxic or normoxic cells and suppressed antioxidant enzymes, including superoxidase dismutase, cata- the hypoxic mesenchymal phenotype. Furthermore, we dem- lase, glutathione reductase, and heme oxygenase-1.15 The in- onstrated in vivo that the distribution pattern of phosphory- crease in BVR in untreated experimental autoimmune enceph- lated Akt is highly congruent with that of rBVR in normal alomyelitis indicates that it is protectively upregulated to kidney and in hypoxic renal fibrosis, indicating that rBVR ex- defend against reactive oxygen species.15 The significant in- erts its fibrogenic effect through activation of Akt. Thus, it is crease in BVR in hypoxic HK-2 cells and hypoxic remnant plausible that Akt activation occurs downstream of BVR in kidneys, as demonstrated here, may also be due to upregula- hypoxia-induced EMT. Our results suggest that despite the tion to protect renal cells or tissue from oxidative injuries. protective upregulation of BVR as an antioxidant in hypoxic However, using an hBVR-transfected cell culture system, renal tubular epithelial cells, it accelerated hypoxia-induced we found that BVR can directly induce EMT, which was pre- tubular EMT via the PI3K/Akt pathway. Thus, at least in renal viously demonstrated as a prerequisite for renal interstitial fi- tubular epithelial cells, BVR acts as a two-edged sword in af- brogenesis.29 Transfection of BVR into HK-2 cells caused phe- fecting the physiologic behavior of cells. notypic changes associated with EMT. The cobblestone-like Our work demonstrates that hypoxia-induced tubular EMT epithelial cells were transformed into elongated cells in shape. may be at least partially regulated by BVR through a PI3K/Akt- Levels of E-cadherin, a cell-cell adhesion molecule present in dependent pathway. BVR is not only a powerful antioxidant the plasma membrane of most epithelial cells, were reduced but also a potential accelerator of renal fibrosis in remnant and the protein formed zipper-like patterns at cell borders. In kidneys. addition, vimentin, a cytoskeletal protein in many mesenchymal cells, was induced. BVR has two potential SH-2 protein-docking sites, the CONCISE METHODS Y198MKM motif and the Y228LSF motif.6 Base on those molecular docking sites, BVR is presumed to bind PI3K and activate Generation of the recombinant hBVR vector of PKB/Akt.6 Our work supports this possibility and estab- On the basis of the hBVR cDNA sequence in Genebank (X93086), the lishes that BVR regulates the phosphorylation of Akt. In full-length hBVR was amplified by reverse transcription PCR using 2 hBVR-transfected HK-2 cells, there were high levels of phos- primers containing restriction sites (BamHI-sense, 5Ј-CGGGATC- phorylated Akt. The PI3K/Akt pathway is a major arm in EMT CCAGTGACCGAAGGAAGAGACCAA-3Ј; XbaI-antisense, 5Ј- signaling. A critical molecular feature of EMT is downregula- GCTCTAGAGCTGGTGCCATCTTGGAAGTGC-3Ј). After purifica- tion of E-cadherin.13 The active form of Akt induces a tran- tion, the amplified cDNA was subcloned into the pcDNA 3.1/Zeo scription factor, snail, that represses expression of the E-cad- plasmid (Invitrogen, Inc., NY, USA) and verified by sequencing. herin gene.12 Phosphorylation of Akt is associated with a loss of cell-cell adhesion, a decrease in cell-matrix adhesion, a loss of Cell Culture and Protocol apico-basolateral cell polarization, induction of cell motility, Parental cells were HK-2 cells, which were donated by Professor Yu and other characteristics of myofibroblasts.12,13 In PI3K-inhib- Xueqing from Zhongshan University. HK-2 cells were cultured in ited BVR-transfected cells, we further confirmed the direct as- DMEM/F12 (Invitrogen, Inc.) supplemented with 10% FBS. The sociation between Akt and EMT. After treatment of plasmid pcDNA3.1/Zeo/hBVR was transfected into HK-2 cells with BVR-transfected cells with Ly294002 or wortmannin, phos- Lipofectamine (Invitrogen, Inc.). Zeocin (Invitrogen, Inc.) was added phorylation of Akt was decreased, the reduction in E-cadherin at a concentration of 25 ␮g/ml for screening. A single resistant clone was reversed, and E-cadherin relocalized at cell borders. These was picked and separately propagated. The untransfected blank con- observations convince us that BVR is directly involved in trol (parental) and a pcDNA 3.1/Zeo empty vector-transfected con-

6 Journal of the American Society of Nephrology J Am Soc Nephrol ●●: –, 2008 www.jasn.org CLINICAL RESEARCH

trol (pcDNA3.1/Zeo) were used for reference. For determinations of phorylated Akt, and Akt primary antibodies (Cell Signaling Technol- PI3K activity, the hBVR-transfected cells were treated with 10 ␮M ogy, Danvers, MA) at 4 °C overnight. The bound antibody complexes Ly294002 or 1 ␮M wortmannin (Cell Signaling Tech, Danvers, MA) were visualized by enhanced chemiluminescence reaction, and x-ray

for 48 h. For hypoxic culture, cells were placed in a hypoxic (1% O2, film was scanned using a ChemiImager 5500 image analysis system

5% CO2, 37°C) incubator (Galaxy oxygen control incubator, RS Bio- (Alpha Innotech, San Leandro, CA, USA). Quantity One software tech, Irvine, UK) for 12, 24, or 48 h. Some 48-h hypoxic cells were (BioRad) was used to quantify the density of bands. treated with Ly294002 or wortmannin simultaneously with the incubation; control cells (normoxic cells) were incubated for equivalent Immunohistochemistry and Immunocytochemistry

periods under normoxic conditions (21% O2,5%CO2, 37°C). For immunohistochemical analysis, paraffin sections were incubated with primary anti-BVR antibody (Stressgen Bioreagents Corp.) at 4°C siRNA for hBVR overnight. The sections were incubated with biotinylated goat anti- For silencing, RNA primers complementary to nucleotides 446 to 471 rabbit Ig antibody as the secondary antibody, and the antibody reac- of hBVR were used (sense, GCACGAGGAGCAUGUUGAACU- tions were visualized using diamino benzidine (DAKO, Tokyo, Ja- CUUG; antisense, CAAGAGUUCAACAUGCUCCUCGUGC). HK-2 pan). An irrelevant isotype rabbit Ig was used for the negative control. cells in 25-cm2 flasks were transfected with 400 pmol of the annealed Images from the microscope were captured with a Nikon DXM 1200 RNA primer pair and 10 ␮l Lipofectamine 2000 (Invitrogen Inc.). digital camera using Automatic Camera Tamer (ACT-1) software, Cells were analyzed 12 to 96 h after transfection. Preliminary results version 2.63, and analyzed with HPIAs-1000 image analysis software (n ϭ 3) showed that silencing was evident at 48, 72, and 96 h after (Qianping Image Engineering Co. of Tongji Medical School, China). transfection. Therefore, 48 h after transfection, cells were placed into The software enables the operator to set the optical threshold and a hypoxic incubator for another 48 h to observe the effect of hBVR filter combination to value positive stains. The positive stains were silencing on the hypoxic cells. The HK-2 cells transfected with Lipo- automatically discriminated by the computer and measured for opti- fectamine 2000 (mock transfection) or the cells transfected with con- cal intensity and total area. The staining index of the immunoreactive trol siRNA (Stealth RNAi Negative Control Duplexes, Invitrogen areas was calculated by multiplying the optical intensity and total area. Inc.) were cultured in hypoxic conditions and used as negative Four to five sections per rat were assessed for immunoreactive areas controls. (n ϭ 6). Immunohistochemistry of BVR and phosphorylated AKT was performed using serial sections to clarify their colocalization in Animal Model the kidneys. Male Sprague–Dawley rats weighing 150 to 200 g were obtained from For immunocytochemical analysis, HK-2 cells were cultured on the Tongji laboratory animal center (Wuhan, China). The chronic sterile glass coverslips in six-well plates. The slides were incubated renal hypoxia disease model was induced by 5/6 subtotal nephrecto- overnight at 4°C with anti-E-cadherin antibody, followed by incuba- my.22,23 All rats were euthanized at 12 wk after nephrectomy, and tion with FITC-conjugated goat anti-mouse Ig at room temperature serum was collected for determination of creatinine and urea nitro- for 1 h. Finally, slides were counterstained with propidium iodide and gen. Kidneys were immediately excised; some were fixed with 4% analyzed by confocal laser scanning microscopy. paraformaldehyde, and others were frozen in liquid nitrogen until use. Kidney sections were stained by the periodic acid–Schiff method Statistical Analyses and examined by light microscopy. All procedures were performed in Data were analyzed by standard statistical methods, including linear re- accordance with our university’s guidelines for animal care. gression, the t test, and one-way ANOVA using SPSS (version 12.0). Data are expressed as the mean Ϯ SEM. Significance was assessed at P Ͻ 0.05. Western Blotting Kidney tissues and cells were extracted with lysis buffer containing 1% Triton-100, 0.5% Nonidet P-40, 20 mM Tris-HCl, 15 mM NaCl, 1 ACKNOWLEDGMENTS ⅐ mmol/L EDTA, 1 mmol/L egtazic acid, 1 mM Na3VO4 10H2O,2mM ⅐ ␤ NaF, 2 mM Na2P2O4 10H2O, 10 mM -glycerophosphate disodium Grant supports: National Science Foundation of China (No. salt (pH 8.0), and cocktail inhibitor (5 mM phenylmethanesulfonyl 30370657); New century excellent Talents grant (NCET004–0712), fluoride, 5 ␮g/ml leupeptin, 5 ␮g/ml pepstatin, and 5 ␮g/ml aproti- and the “973” Program of China (No. 2002CB513100). nin) for 30 min on ice. Cell debris was removed by centrifugation at 13,000 g for 20 min at 4°C. Protein concentration was determined ␮ using the Bradford method. Total protein (80 g) was subjected to DISCLOSURES SDS-PAGE. The proteins were then transferred to nitrocellulose None. membranes. After transfer, the membranes were blocked with 5% nonfat milk in TBS with 0.1% Tween-20 for1hat37°C, and blotted routinely with BVR (Stressgen Bioreagents Corp., Victoria, British REFERENCES Columbia, Canada), HIF-1␣ (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), E-cadherin (Santa Cruz Biotechnology), vimentin (Santa 1. Beale SI, Cornejo J: Enzymatic heme oxygenase activity in soluble Cruz Biotechnology), ␤-actin (Santa Cruz Biotechnology), phos- extracts of the unicellular red alga, Cynidium caldarium. Arch Biochem

J Am Soc Nephrol ●●: –, 2008 Biliverdin Reductase Accelerates Hypoxia-Mediated Renal Tubular Epithelial to Mesenchymal Transition 7 CLINICAL RESEARCH www.jasn.org

Biophys 235: 371–384, 1984 16. Juurlink BH: Response of glial cells to ischemia: roles of reactive 2. Maines MD: New insights into biliverdin reductase functions: Linking oxygen species and glutathione. Neurosci Biobehav Rev 21: 151–166, heme metabolism to cell signaling. Physiology 20: 382–389, 2005 1997 3. Fakhrai H, Maines MD: Expression and characterization of a cDNA for 17. Vanden Hoek TL, Li C, Shao Z, Schumacker PT, Becker LB: Significant rat kidney biliverdin reductase. Evidence suggesting the liver and levels of oxidants are generated by isolated cardiomyocytes during kidney enzymes are the same transcript product. J Biol Chem 267: ischemia prior to reperfusion. J Mol Cell Cardiol 29: 2571–2583, 1997 4023–4029, 1992 18. Blokhina O, Virolainen E, Fagerstedt KV: Antioxidants, oxidative dam- 4. Maines MD, Polevoda BV, Huang TJ, McCoubrey WK, Jr: Human age and oxygen deprivation stress: A review. Ann Bot (Lond) 91: biliverdin IX alpha reductase is a zinc-metalloprotein. Characterization 179–194, 2003 of purified and Escherichia coli expressed enzymes. Eur J Biochem 19. Manotham K, Tanaka T, Matsumoto M, Ohse T, Inagi R, Miyata T, 235: 372–381, 1996 Kurokawa K, Fujita T, Ingelfinger JR, Nangaku M: Transdifferentiation 5. Salim M, Brown-Kipphut BA, Maines MD: Human biliverdin reductase of cultured tubular cells induced by hypoxia. Kidney Int 65: 871–880, is autophosphorylated, and phosphorylation is required for bilirubin 2004 formation. J Biol Chem 276: 10929–10934, 2001 20. Nangaku M: Chronic hypoxia and tubulointerstitial injury: A final com- 6. Lerner-Marmarosh N, Shen J, Torno MD, Kravets A, Hu Z, Maines MD: mon pathway to end-stage renal failure. J Am Soc Nephrol 17: 17–25, Human biliverdin reductase: A member of the insulin receptor sub- 2006 strate family with serine/threonine/tyrosine kinase activity. Proc Natl 21. Walker EH, Pacold ME, Perisic O, Stephens L, Hawkins PT, Wymann Acad Sci USA 17; 102: 7109–7114, 2005 MP, Williams RL: Structural determinants of phosphoinositide 3-kinase 7. Maines MD, Ewing JF, Huang TJ, Panahian N: Nuclear localization of inhibition by wortmannin, LY294002, quercetin, myricetin, and stau- biliverdin reductase in the rat kidney: response to nephrotoxins that rosporine. Mol Cell 6: 909–919, 2000 induce heme oxygenase-1. J Pharmacol Exp Ther 296: 1091–1097, 22. Manotham K, Tanaka T, Matsumoto M, Ohse T, Miyata T, Inagi R, 2001 Kurokawa K, Fujita T, Nangaku M: Evidence of tubular hypoxia in the 8. Ahmad Z, Salim M, Maines, MD: Human biliverdin reductase is a early phase in the remnant kidney model. J Am Soc Nephrol 15: leucine zipper-like DNA-binding protein and functions in transcrip- 1277–1288, 2004 tional activation of heme oxygenase-1 by oxidative stress. J Biol Chem 23. Zhang B, Liang X, Shi W, Ye Z, He C, Hu X, Liu S: Role of impaired 277: 9226–9232, 2002 peritubular capillary and hypoxia in progressive interstitial fibrosis 9. Kravets A, Hu Z, Miralem T, Torno MD, Maines, MD: Biliverdin reduc- after 5/6 subtotal nephrectomy of rats. Nephrology 10: 351–357, tase, a novel regulator for induction of activating transcription factor-2 2005 and heme oxygenase-1. J Biol Chem, 279: 19916–19923, 2004 24. Kang DH, Joly AH, Oh SW, Hugo C, Kerjaschki D, Gordon KL, Mazzali 10. Kravets A, Hu Z, Miralem T, Torno MD, Maines MD: Biliverdin reduc- M, Jefferson JA, Hughes J, Madsen KM, Schreiner GF, Johnson RJ: tase: a major physiologic cytoprotectant. Proc Natl Acad Sci USA 99: Impaired angiogenesis in the remnant kidney model: I. Potential role 16093–16098, 2002 of vascular endothelial growth factor and thrombospondin-1. JAm 11. Backer JM, Myers MG, Jr., Shoelson SE, Chin DJ, Sun XJ, Miralpeix M, Soc Nephrol 12: 1434–1447, 2001 Hu P, Margolis B, Skolnik EY, Schlessinger J: Phosphatidylinositol 25. Stutz F, Muller Ga: Transdifferentiation comes of age. Nephrol Dial 3Ј-kinase is activated by association with IRS-1 during insulin stimula- Transplant 15: 1729–1731, 2000 tion. EMBO J 11: 3469–3479, 1992 26. Ng YY, Huang TP, Yang WC, Chen ZP, Yang AH, Mu W, Nikolic- 12. Grille SJ, Bellacosa A, Upson J, Klein-Szanto AJ, van Roy F, Lee-Kwon Paterson DJ, Atkins RC, Lan HY: Tubular epithelial myofibroblast W, Donowitz M, Tsichlis PN, Larue L: The protein kinase Akt induces transdifferentiation in progressive tubulointerstitial fibrosis in 5/6 ne- epithelial mesenchymal transition and promotes enhanced motility phrectomized rats. Kidney Int 54: 864–876, 1998 and invasiveness of squamous cell carcinoma lines. Cancer Res 63: 27. Farrera JA, Jauma A, Ribo JM, Peire MA, Parellada PP, Roques-Choua 2172–2178, 2003 S, Bienvenue E, Seta P: The antioxidant role of bile pigments evalu- 13. Larue L, Bellacosa A: Epithelial-mesenchymal transition in develop- ated by chemical tests. Bioorg Med Chem 2: 181–185, 1994 ment and cancer: role of phosphatidylinositol 3Ј kinase/AKT pathways. 28. Neuzil J, Stocker R: Free and albumin-bound bilirubin are efficient Oncogene 24: 7443–7454, 2005 co-antioxidants for alpha-tocopherol, inhibiting plasma and low den- 14. Miralem T, Hu ZB, Torno MD, Lelli KM, Maines MD: Small interference sity lipoprotein lipid peroxidation. J Biol Chem 269: 16712–16719, RNA-mediated gene silencing of human biliverdin reductase, but not 1994 that of heme oxygenase-1, attenuates arsenite-mediated induction of 29. Yang J, Liu Y: Dissection of key events in tubular epithelial to myofi- the oxygenase and increases apoptosis in 293A kidney cells. J Biol broblast transition and its implications in renal interstitial fibrosis. Am J Chem 280: 17084–17092, 2005 Pathol 159: 1465–1475, 2001 15. Liu Y, Liu J, Tetzlaff W, Paty DW, Cynader MS: Biliverdin reductase, a 30. Fan JM, Ng YY, Hill PA, Nikolic-Paterson DJ, Mu W, Atkins RC, Lan HY: major physiologic cytoprotectant, suppresses experimental autoim- Transforming growth factor ␤ regulates tubular epithelial-myofibro- mune encephalomyelitis. Free Radic Biol Med 40: 960–967, 2005 blast transdifferentiation in vitro. Kidney Int 56: 1455–1467, 1999

8 Journal of the American Society of Nephrology J Am Soc Nephrol ●●: –, 2008