If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Harvard Stem Cell Institute, Holyoke Center, Cambridge, MassachusettsDepartment of Stem Cell and Regenerative Biology, Harvard University, Cambridge, MassachusettsDepartment of Molecular and Cellular Biology, Harvard University, Cambridge, Massachusetts
The Hedgehog (Hh) signaling pathway regulates tissue patterning during development, including patterning and growth of limbs and face, but whether Hh signaling plays a role in adult kidney remains undefined. In this study, using a panel of hedgehog-reporter mice, we show that the two Hh ligands (Indian hedgehog and sonic hedgehog ligands) are expressed in tubular epithelial cells. We report that the Hh effectors (Gli1 and Gli2) are expressed exclusively in adjacent platelet-derived growth factor receptor-β-positive interstitial pericytes and perivascular fibroblasts, suggesting a paracrine signaling loop. In two models of renal fibrosis, Indian Hh ligand was upregulated with a dramatic activation of downstream Gli effector expression. Hh-responsive Gli1-positive interstitial cells underwent 11-fold proliferative expansion during fibrosis, and both Gli1- and Gli2-positive cells differentiated into α-smooth muscle actin-positive myofibroblasts. In the pericyte-like cell line 10T1/2, hedgehog ligand triggered cell proliferation, suggesting a possible role for this pathway in the regulation of cell cycle progression of myofibroblast progenitors during the development of renal fibrosis. The hedgehog antagonist IPI-926 abolished Gli1 induction in vivo but did not decrease kidney fibrosis. However, the transcriptional induction of Gli2 was unaffected by IPI-926, suggesting the existence of smoothened-independent Gli activation in this model. This study is the first detailed description of paracrine hedgehog signaling in adult kidney, which indicates a possible role for hedgehog-Gli signaling in fibrotic chronic kidney disease.
The Hedgehog (Hh) signaling pathway plays a crucial role in regulating a diverse range of developmental processes in the mammalian embryo, including ventralization of the neural tube, patterning and growth of limbs and face, the formation of organs (such as the lung and gut), development of hair follicles, and decisions of left-right asymmetry.
In the kidney, sonic hedgehog (Shh) expression in papillary collecting duct and ureteric epithelium regulates adjacent mesenchymal cell proliferation and differentiation, and either germline Shh deletion or deletion of Shh from collecting duct leads to severe renal developmental abnormalities, including renal aplasia or hypoplasia.
These secreted, lipid-modified proteins can act at short or long distances by binding to the membrane receptor Patched1 (Ptch1) on target cells, thereby releasing tonic inhibition by Ptch1 on the transmembrane protein smoothened (Smo). Derepressed Smo translocates to the primary cilium, inhibiting production of the truncated repressor forms of the Gli2 and Gli3 transcription factors and promoting preservation of their full-length activator forms, which induce transcription of Hh target genes, including Gli1 and Ptch1, both of which serve as readouts of Hh pathway activation.
Little is known about a role for the Hh pathway in the adult kidney. In cancer and solid organ injury models, recent evidence suggests that epithelial-derived Hh ligands can be reactivated in pathological states to transmit signals to surrounding mesenchymal cells. For example, in carcinogenesis, Hh ligands from the epithelial tumor act on adjacent stroma to promote a favorable tumor microenvironment.
and because of this, we hypothesized that the Hh pathway would be activated in these cells during renal fibrogenesis. Using complementary techniques, including a variety of genetic reporter mice, we demonstrate that the Hh ligands (Ihh and Shh) are expressed in tubular epithelial cells of the kidney, whereas the Hh effectors (Gli1 and Gli2) are expressed in platelet-derived growth factor receptor-β (PDGFR-β)-expressing interstitial pericytes and perivascular fibroblasts. Both Ihh expression and downstream Hh signaling were substantially activated during renal fibrosis, as Hh-responsive pericytes and perivascular fibroblasts proliferated and differentiated into myofibroblasts. Hh ligand drove cell proliferation in a pericyte-like cell line, suggesting that epithelial-derived Hh ligands might direct mesenchymal cell proliferation during renal fibrosis. Pharmacological inhibition of Smo completely suppressed Gli1 induction, but it did not inhibit fibrosis, suggesting that Gli2, whose induction was not inhibited, may be the more important Gli effector in renal fibrosis.
Materials and Methods
All mouse studies were performed according to the animal experimental guidelines issued by the Animal Care and Use Committee at Harvard University. Wild-type mice were from Charles River Laboratories (Wilmington, MA); FVB/N mice were used for unilateral ureteral obstruction (UUO) and C57BL/6 mice were used for unilateral ischemia reperfusion injury (UIRI) time course experiments and quantitative PCR studies. Ptch1-nLacZ (JAX Stock 003081), Gli1-nLacZ (JAX Stock 008211), Gli2-nLacZ (JAX stock 007922), Shh-GFPCre (JAX Stock 005622), and R26-LacZ knock-in mice were purchased from Jackson Laboratories (Bar Harbor, ME). To create Ihh-nLacZ reporter mice, an Ihh-nLacZ reporter allele was constructed and the Ihh locus targeted in the embryonic stem cells, replacing most of the first exon of Ihh with an NLS-LacZ-pA cassette (see Supplemental Figure S1at http://ajp.amjpathol.org).
Mice of 8 to 12 weeks were anesthetized with pentobarbital sodium (60 mg/kg body weight) before surgery, and body temperatures were controlled at 36.5 to 37.5°C throughout all procedures. Each time point represents three to five mice as indicated. For UUO, the left kidney was exposed through a flank incision and the left ureter tied off at the level of the lower pole with two 4.0 silk ties. Mice were sacrificed 3 to 14 days after obstruction. For UIRI, the left kidney was exposed through a flank incision, and the renal pedicle was clamped with nontraumatic microaneurysm clamps (Roboz Surgical Instrument Co., Gaithersburg, MD), which were removed after 28 minutes. Reperfusion was visually verified. Two hours after surgery, 1 mL of 0.9% NaCl intraperitoneally was administered.
Tissue Preparation and Histology
Mice were anesthetized, euthanized, and immediately perfused via the left ventricle with ice-cold PBS for 1 minute. Kidneys were either snap frozen or fixed in 4% paraformaldehyde on ice for 2 hours, then incubated in 30% sucrose in PBS at 4°C overnight. OCT-embedded (Sakura Finetek, Torrance, CA) kidneys were cryosectioned into 7-μm sections. LacZ activity was measured on paraformaldehyde-fixed frozen sections by standard 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-gal) staining for 1 to 6 days at 37°C, and counterstained with nuclear fast red and mounted. To quantify nLacZ cell number, 100× images were taken of the entire cortex (Gli1 and Ptch1), the inner cortex (Ihh) or cortex and medulla (Gli2) of the midsagittal kidney sections containing papilla from at least four different animals; the number of positive cells were then counted in each 100× image using a manual cell counter from ImageJ (http://rsbweb.nih.gov/ij). Primary antibodies included rabbit anti-β-galactosidase (MP Biomedicals, Solon, OH, Cat. #55976, 1:2000), chicken anti-green fluorescent protein (Aves Labs, Tigard, OR, Cat #GFP-1020, 1: 500), rat anti-PDGFR-β (eBioscience, San Diego, CA, Cat. #14–1402, 1:500), Cy3-conjugated α-smooth muscle actin (α-SMA) (Sigma-Aldrich, St. Louis, MO, Cat. #C6198, 1:500), rat anti-F4/80 (Abcam, Cambridge, MA, Cat. #ab6640, 1:500), fluorescein isothiocyanate (FITC)-Lotus tetragonolobus lectin (Vector Labs, Burlingame, CA, Cat. #FL-1321, 1:500), rabbit anti-CD31 (Abcam, Cat. #ab28364, 1:500), rabbit anti-aquaporin 2 (Abcam, Cat. #ab15116, 1:500), fluorescein isothiocyanate (FITC)-conjugated Dolichos biflorus agglutinin (Vector Labs, Cat. # FL-1031, 1:500), and rabbit anti-NKCC2 (Alpha Diagnostic, San Antonio, TX, Cat. #NKCC21-A, 1:100). Secondary antibodies were either FITC or Cy3-conjugated (Jackson ImmunoResearch, West Grove, PA) incubated for 30 minutes, with DAPI nuclear counterstain followed by mounting in ProlongGold (Invitrogen, Carlsbad, CA). Images were obtained by confocal (Nikon C1 eclipse, Nikon, Melville, NY) or standard microscopy (Nikon eclipse 90i, Nikon). Anti-LacZ antibodies reliably labeled LacZ-expressing interstitial cells, although the autofluorescence in tubular epithelia blunted their sensitivity in tubular epithelial cells. Therefore, in certain situations X-gal staining followed by indirect immunofluorescence was performed with pseudocoloring of the X-gal stain.
Cell Culture Experiments
10T1/2 cells (ATCC) were grown in Basal Media Eagle (Gibco, Billings, MT) with 10% fetal bovine serum supplemented with penicillin and streptomycin and 2 mmol/L glutamine. Shh conditioned media was generated from supernatants of Cos7 cells stably transfected with pcDNA3-N-Shh or pcDNA3 control plasmid. For propidium iodide cell cycle analysis and Bromodeoxyuridine (BrdU) uptake cell proliferation assays, cells were grown on 6 well plates, serum starved by incubating in 0.5% fetal bovine serum for 12 hours, and then stimulated for 24 hours with either Shh conditioned media, Cos7 control media, 500 nmol/L smoothened agonist (SAG) (Santa Cruz, Santa Cruz, CA, Cat. #sc-202814) or water control in 0.5% or 10% fetal bovine serum. For the BrdU uptake assay, the cells were incubated in 10 μm BrdU for 2 hours before harvesting and then stained using the BrdU-FITC flow kit (BD Pharmingen, San Diego, CA). For the cell cycle analysis, cells were fixed in ice-cold 100% ethanol, incubated with propidium iodide (400 μg/mL propidium iodide, 2 mmol/L MgCl2, 100 mg/mL RNase), and subject to fluorescence-activated cell sorting (FACS) analysis.
IPI-926 (5 mg/mL) stock solution (Infinity Pharmaceuticals, Cambridge, MA) was prepared fresh for each experiment by dissolving in hydroxyproplyl-β-cyclodextrin (vehicle) and sonicating. Mice were given 40 mg/kg body weight IPI-926 or vehicle by gastric lavage daily until the day before sacrifice, with Gli1-LacZ mice receiving their first dose the day before UUO surgery and being sacrificed on day 7 of UUO, and BALB/c and C57BL/6 mice receiving their first dose 2 days before UUO surgery and being sacrificed on day 10 of UUO.
To confirm the presence of Shh in conditioned media by Western blot, 5 μL of conditioned media was first separated by 10% polyacrylamide gel electrophoresis. To determine the relative amount of α-SMA protein in kidneys from IPI-926- versus vehicle-treated mice, the lower kidney pole from UUO and contralateral (CLK) kidneys were homogenized in radioimmunoprecipitation assay buffer with protease inhibitors using a handheld rotor, the total protein quantified by Bradford Assay and 25 μg separated by 10% polyacrylamide gel electrophoresis. Proteins were transferred to polyvinylidene difluoride membrane, blocked in 5% milk in phosphate buffered saline tween 20, probed overnight at 4°C with goat anti-Shh-N antibody (Santa Cruz, Cat. #sc-1194, 1:200) or mouse anti-α-SMA (Sigma-Aldrich, Cat. #A2547, 1:5000), or for 1 hour at room temperature with mouse anti-glyceraldehydes-3-phosphate dehydrogenase (Abcam, Cat. #ab9484, 1:5000), washed, probed with anti-goat- or mouse-horseradish peroxidase (Dako, Carpinteria, CA, 1:5000) for 1 hour at room temperature, and the antigen antibody complex was visualized using the ECL detection system (PerkinElmer, Waltham, MA).
RNA was extracted from snap-frozen tissue stored at −80°C or cells using standard techniques (RNeasy, Qiagen, Germantown, MD). Reverse transcription was performed with the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA) producing cDNA. Real-time PCR was performed using iQ-SYBR Green supermix (BioRad) and the iQ5 Multicolor Real-Time PCR Detection system (BioRad) for detection of mRNA levels. Glyceraldehydes-3-phosphate dehydrogenase was used as the internal control.
Statistical analyses were performed using Graph Pad Prism software (version 5.0) (GraphPad Software, Inc., San Diego, CA). Analysis of variance was used to compare data among groups followed by a Tukey's post test to compare all groups to each other or a Dunnett's post test to compare all groups to the control group. A two-tailed Student's t-test was used when only two groups were being compared. All results were repeated at least twice. A P value of less than 0.05 was considered significant. The results are presented as mean ± SEM.
Hh Ligands, Shh, and Ihh are Expressed in Tubular Epithelial Cells
To define the expression pattern of Hh pathway members in renal fibrosis, we used available Ptch1-nLacZ, Gli1-nLacZ, and Gli2-nLacZ reporter mice, and generated Ihh-nLacZ knockin reporter mice. Because Shh-GFPCre reporter mice exhibited unexpectedly low green fluorescent protein fluorescence, historical Shh expression was assessed in Shh-GFPCre; R26-LacZ bigenic mice, in which cytoplasmic LacZ expression marks cells that either actively express Shh or expressed Shh at one time in development (see Supplemental Figure S2 at http://ajp.amjpathol.org). LacZ expression in kidney sections from Shh-GFPCre; R26-LacZ adult mice was present exclusively in the papilla, corresponding to in situ hybridization staining of Shh mRNA in P1 kidney (Figure 1A), as well as ureteral urothelium (see Supplemental Figure S3 at http://ajp.amjpathol.org) as expected.
We generated Ihh-nLacZ knockin mice to report Ihh expression (see Supplemental Figure S1 at http://ajp.amjpathol.org). Ihh was expressed predominantly in the inner cortex and outer medulla at the corticomedullary junction, with reduced expression seen throughout the rest of the medulla (Figure 1A). In situ hybridization in P1 mouse kidneys confirmed staining in the outer medulla (Figure 1A), consistent with previous findings during mouse development.
As with Shh, Ihh was exclusively expressed in tubular epithelial cells (Figure 2B). Most Ihh-nLacZ tubular cells in the inner cortex and outer medulla co-stained with the proximal tubular marker Lotus tetragonolobus lectin (Figure 2B; see also Supplemental Figure S4 at http://ajp.amjpathol.org), consistent with a previous report of Ihh expression in dissected proximal tubules by real-time PCR.
In addition, occasional Ihh-nLacZ was observed in thin limbs of Henle (see Supplemental Figure S5A at http://ajp.amjpathol.org), demonstrating Ihh expression of tubular epithelial cells with squamous morphology lacking brush borders. These cells did not costain with collecting duct markers aquaporin 2 or Dilochus biflorus agglutinin, the thick ascending limb marker Na-K-2Cl cotransporter or the endothelial marker CD31 (see Supplemental Figure S5B at http://ajp.amjpathol.org). Relative mRNA expression as determined by quantitative PCR from dissected kidney cortex, medulla, and papilla confirmed that Shh is the most highly expressed Hh ligand in the papilla, and Ihh is the most highly expressed ligand in the medulla and cortex. Dhh expression was minimal (Figure 1C).
The Hh Receptor Ptch1 and Downstream Effectors Gli1 and Gli2 are Expressed in Interstitial Pericytes and Perivascular Fibroblasts
To define the cell types that respond to Hh ligand, we examined the expression patterns of Ptch1 and Gli effectors in the adult kidney. Ptch1 and Gli1 are readouts of Hh pathway activity, and their expression defines Hh-responsive cells. Gli2 lies directly upstream of Gli1 and other Hh transcriptional targets.
Ptch1 (Figure 1A) and Gli1 (Figure 1B) were expressed strongly at the corticomedullary junction, suggesting that these cells may be responding to Ihh in that region, whereas Gli2 (Figure 1B) was expressed most prominently in the inner medulla and papilla. Ptch1 and a lesser amount of Gli1 expression was observed in the inner medulla and papilla as well, likely in response to Ihh in the inner medulla and Shh in the papilla. In situ studies of Ptch1 in P1 kidney sections were consistent with Ptch1-nLacZ expression in adult mice (Figure 1A) and embryonic kidney.
Ptch1 was also expressed in occasional tubular epithelial cells, glomerular cells, and endothelial cells, in addition to interstitial cells (Figure 2C; see also Supplemental Figure S6 at http://ajp.amjpathol.org). In contrast, Gli1 and Gli2 were exclusively expressed in interstitial cells in the adult kidney (Figure 2C). Though there has been a prior report of Gli1 expression in tubules, especially in the setting of decreased transcriptional repressor Glis2,
we did not observe X-gal staining of tubular epithelial cells using our Gli1-nLacZ reporter mouse, even in kidneys from newborn and 7-day-old mice (Figure 3A). We did, however, observe X-gal staining of epithelial cells in the ureteric bud in the nephrogenic zone in kidneys from Gli2-nLacZ newborn mice that was decreased in kidneys from 7-day-old mice and almost completely absent in kidneys from 14-day-old mice (Figure 3B). A higher density of Ptch1, Gli1, and Gli2-positive interstitial cells were observed closely associated with vessels (Figure 2C, bottom three panels). Quantitative mRNA comparisons confirmed that Ptch1 and Gli2 were most prominently expressed in the medulla and papilla, and Gli1 mRNA was highest in the medulla (Figure 1, D and E). Gli3 was also highest in the medulla and papilla (Figure 1E), and was expressed the highest overall when comparing the three Gli effectors in kidney.
Hh Pathway Upregulation in Renal Fibrosis
To investigate Hh pathway modulation during renal fibrosis, we measured mRNA expression of Hh pathway members in corticomedullary kidney lysates from adult mice after 3, 7, and 14 days of chronic injury by UUO compared to sham controls. Expression of the fibrotic marker Collagen 1α1 (Col1α1) and the myofibroblast marker α-SMA progressively increased relative to sham, confirming fibrosis (Figure 4A). A progressive increase in Gli1 and Gli3 mRNA expression occurred on days 3, 7, and 14 of UUO, and a progressive increase in Gli2 mRNA expression occurred on days 7 and 14 (Figure 4C). Gli1 and Gli3 demonstrated a more robust induction relative to sham with a 13.6 ± 4.3-fold increase in Gli1 and a 15.2 ± 5.7-fold increase in Gli3 on day 14 versus a 3.5 ± 1.9-fold increase in Gli2. Gli1 transcription reflects active Hh signaling, and because of this, the results indicate that the Hh pathway is activated during renal fibrosis. Ptch1 expression also increased (Figure 4B), although only by 2.1 ± 0.4-fold, perhaps reflecting its stronger baseline expression compared to Gli1.
Next, we asked which Hh ligand might account for increased Hh signaling. Shh expression did not change during UUO, although Ihh was induced transcriptionally, peaking at day 3 with a 4.5 ± 0.5 fold increase and remained elevated thereafter (Figure 4D). A similar 3.4 ± 0.8 increase in Ihh mRNA at UUO day 3 was observed in a second independent experiment (N = 5). Dhh was also increased relative to sham at all time points, although the absolute amount of Dhh was extremely low -35.5 ± 3.7-fold lower than Ihh mRNA levels, indicating that Ihh is the primary Hh ligand induced by chronic renal injury. To address the generalizability of these findings, we investigated a second model of renal fibrosis, unilateral ischemia reperfusion injury. UIRI has been validated as a model of renal fibrosis in previous reports
and a dramatic increase in α-SMA immunofluorescent staining in UIRI day 14 kidneys compared to CLK provided further confirmation that a robust fibrotic response was achieved (see Supplemental Figure S7 at http://ajp.amjpathol.org). In this model, Ptch1, Gli1, Gli2, and Gli3 mRNA were all significantly increased in medullary kidney lysates relative to sham, with peak levels observed on day 7 in parallel with the peak increase in expression of Col1α1 and α-SMA (Figure 4, E–G). Medullary lysates did not show an increase in Ihh (data not shown), although Ihh was increased at all time points in cortical lysates with a peak increase of 3.9 ± 0.3 at day 7 (Figure 4H). Hh pathway, therefore, is activated in two separate mouse models of kidney fibrosis.
Myofibroblasts Respond to Hh Signals during Fibrosis
To further define the cells that respond to Hh ligands, we quantitated tubular versus interstitial expression of Gli1, Gli2, and Ptch1 during UUO. Gli1 and Gli2 remained exclusively expressed in the interstitium in UUO kidneys without detectable tubular expression (Figure 5, B and D). Compared to uninjured kidneys, cortical Gli1-nLacZ cells increased by 4.1 ± 1.1-fold at 3 days, 10.5 ± 1.8-fold at 7 days, and 10.7 ± 0.8-fold at 14 days after UUO (Figure 5, A–C). The number of LacZ-expressing cells in Gli2-nLacZ mice increased as well, but to a lesser degree, with only a 1.7 ± 0.3 increase in the cortex and 3.9 ± 0.5 increase in the medulla (Figure 5, D and E; see also Supplemental Figure S8A at http://ajp.amjpathol.org). There was a 1.9 ± 0.5-fold decrease in the number of Ptch1-nLacZ tubular epithelial cells, but there was a 4.1 ± 0.6-fold increase in the number of Ptch1-nLacZ interstitial cells (Figure 5, D and E; see also Supplemental Figure S8B at http://ajp.amjpathol.org). In contrast with the transcriptional induction of Ihh observed during renal fibrosis, there was no increase in the number of Ihh-nLacZ cells in UUO. Ihh-nLacZ expression remained localized to tubular epithelial cells in the inner cortex and outer medulla after UUO (Figure 5, D and E; see also Supplemental Figure S8C at http://ajp.amjpathol.org). Thus, the increase in Ihh mRNA expression was not due to an increase in the number of Ihh expressing cells at the level of sensitivity of the Ihh-nLacZ reporter.
During development, epithelial-derived Hh regulates mesenchymal proliferation and differentiation; we therefore sought to more precisely define the interstitial cell type that was responding to Hh signals and asked whether these cells were proliferating during renal fibrosis. A protocol for detection of nuclear LacZ by immunofluorescence was developed for this purpose. Gli1-nLacZ-positive cells uniformly co-expressed the pericyte and perivascular fibroblast marker PDGFR-β in both uninjured and injured kidneys (Figure 6). In the fibrotic but not uninjured kidney, Gli1-nLacZ-positive cells also acquire the myofibroblast marker α-SMA (Figure 6). Macrophages and endothelial cells were often closely opposed to Gli1-nLacZ-positive cells; there was, however, no overlap in the Gli1 expression domain among either of these cell types (Figure 6). The close association between Gli1-nLacZ-positive cells and endothelial cells is consistent with the possibility that some or all of these cells are pericytes. Gli2-nLacZ and Ptch1-nLacZ also colocalized with PDGFR-β in uninjured and injured kidneys, with the majority of them co-expressing the myofibroblast marker α-SMA during injury, but not the macrophage marker F4/80 or endothelial marker CD31 (Figure 6).
To investigate the correlation between Gli1 expression and cell proliferation in UUO, Gli1-nLacZ expressing cells were costained with the cell cycle marker Ki-67. Ki-67-positive cells were observed in both tubules and in the interstitium on day 3 of UUO. The percentage of Gli1-nLacZ positive cells that were co-stained for Ki-67 was 12.6 ± 1.2% compared to only 1.3 ± 0.4% in uninjured kidneys (Figure 7, A and B). These results indicate that many Hh-responsive cells are proliferating in the early stages of renal fibrosis.
Exogenous Hh Ligand Drives Cell Cycle Progression in Pericyte-Like Cells
Next we asked whether Hh ligand could directly induce proliferation of pericyte-like cells in vitro. The mouse mesenchymal cell line 10T1/2 is hedgehog-responsive and multipotent,
so we reasoned that 10T1/2 cells might be a good model for uninjured kidney pericytes. The presence of Shh in conditioned media from Cos7 cells stably transfected with pcDNA-N-Shh was confirmed by Western blot (see Supplemental Figure S9A at http://ajp.amjpathol.org). Then we confirmed that the media containing Shh activates Gli1 expression in these cells
by 153.9 ± 8.2-fold under our conditions. Consistent with this, the Smo agonist SAG induced a 107.5 ± 6.2-fold increase in Gli1 gene expression (Figure 7C). Gli2 and Gli3 were only minimally affected (see Supplemental Figure S9B at http://ajp.amjpathol.org). Neither platelet-derived growth factor nor transforming growth factor-β, both increased in UUO, induced Gli1 expression (Figure 7C). Although 10T1/2 cells have been used to model Hh-induced differentiation, the effect of Hh agonists on cell proliferation in these cells has not been reported. Hh pathway activation either with Shh or SAG induced proliferation of serum-starved 10T1/2 pericytes, as assessed by cell cycle analysis (Figure 7, D and E). In confirmation of these results, Shh and SAG also stimulated BrdU uptake as quantitated by FACS analysis (Figure 7, F and G). These in vitro results suggested that Hh could drive pericyte proliferation during fibrotic injury and are consistent with prior reports that Hh signaling can regulate proliferation of mouse and human mesenchymal cells in vitro.
Pharmacological Inhibition of Hh Signaling Does Not Attenuate Fibrosis
We next investigated the functional role of kidney Hh signaling in vivo by pharmacological inhibition. Cyclopamine is a well-characterized Smo inhibitor, buts its use in vivo is limited by its short half life
IPI-926 nearly completely abolished Gli1 induction after 7 days of UUO, as reflected by the expression of Gli1-nLacZ (Figure 8, A and B). The efficacy of IPI-926 in inhibiting Hh signaling was further confirmed by quantitative PCR from day 10 UUO corticomedullary kidney extracts from BALB/c mice; the increase in Gli1 mRNA expression seen in UUO kidneys from the vehicle-treated mice was completely suppressed, and a decrease in the CLK controls was also seen. Importantly, the increase in Gli2 mRNA seen in UUO was not suppressed by IPI-926, suggesting that the increase in Gli2 in this setting is not smoothened dependent. Despite complete inhibition of Gli1 by IPI-926, there was no decrease in renal fibrosis, as assessed by change in Col1α1, fibronectin, or α-SMA gene transcription, or α-SMA protein levels by Western blot at UUO day 10 (Figure 9, A and B). In a blinded assessment of interstitial fibrosis/tubular atrophy percentage by trichome stain at UUO day 10, there also showed no difference between IPI-926 and vehicle-treated groups. These experiments establish that Gli1 induction in this model is mediated by Hh ligand, but Gli1 does not mediate renal fibrosis in this model.
Activation of canonical Hh signaling in mesenchymal cells during tissue injury has been recently observed in the bladder, liver, and lung.
supporting the hypothesis tested here that Hh-Gli signaling is reactivated in renal fibrosis and that myofibroblasts and their progenitors responds to Hh ligands. These findings also support the general concept that kidney injury responses often reactivate developmental signaling pathways,
and consistent with this, their expression was strongest in the outer medulla of the adult kidney. Ihh induction drives Ptch1 and Gli1 expression in cortex and medulla during fibrosis, because it is expressed in adjacent tubular epithelium, and because Gli1 induction was completely inhibited by the Smo inhibitor IPI-926. The epithelial localization of both Ihh and Shh in the kidney, combined with our demonstration of stromal expression of Gli1 and Gli2 in renal interstitium, indicates that Hh is acting in a paracrine fashion in kidney fibrosis, as it does during renal development.
We observed transcriptional induction of Ihh in renal fibrosis but nontranscriptional mechanisms may also contribute to Hh pathway activation in target cells. Release of pre-formed Hh ligand has been recently reported to occur from peripheral nerves in skin,
and whether such a mechanism operates in the kidney remains to be tested. Smo inhibition did not decrease fibrosis, although redundant pathways for myofibroblast proliferation may exist in this model. Equally important, although Smo inhibition inhibited Gli1 induction, it did not suppress Gli2 induction. Gli1 and Gli2 can have redundant roles themselves,
Our results, therefore, indicate that Gli2 might be the more important Gli effector in renal fibrosis. Recently, evidence indicates that other signaling pathways may sensitize target cells to Hh ligand
Whether noncanonical, Smo-independent Gli activation occurs in kidney fibrosis, and defining the extent to which other more established pro-fibrotic pathways might modulate Hh-Gli signaling in the adult kidney are critical questions that require further investigation.
The functional role of Hh-Gli signaling in renal pericytes, perivascular fibroblasts, and myofibroblasts in vivo remains to be defined. Our in vitro evidence suggested the hypothesis that Hh signaling might contribute to mesenchymal cell proliferation during injury, consistent with its known role in regulating ureteral stromal cell proliferation during development. Our in vivo data, however, do not support this model. Other roles for Hh signaling in renal injury responses are also possible. Hh can drive pro-angiogenic signaling in mesenchymal cells after injury
Another question raised by these studies is why Gli1, Gli2, and Ptch1 are expressed in only some myofibroblasts. Are the Hh-responsive pericytes and perivascular fibroblasts different from their neighboring stromal cells? A growing literature documents Hh pathway activation in mesenchymal stem cell biology,
In the future it will be important to define possible functional differences between Gli1-positive and Gli1-negative interstitial cells. Finally, strong evidence implicates cortical Gli3 repressor function in regulating ureteric tip gene expression and patterning during renal development.
The activation of Hh signaling in cortex that we report here suggests that the balance of Gli activator and repressor forms may be altered during kidney injury.
In summary, we demonstrate, for the first time, strong activation of the Hh-Gli pathway during renal fibrosis. Chronic injury induces Ihh expression, which acts in a paracrine fashion on interstitial pericytes and perivascular fibroblasts to activate Gli effector expression. These findings define the Hh-Gli pathway as a novel developmental signaling pathway that is strongly upregulated in renal fibrosis. Future studies are required to define the functional roles of Gli effector proteins in kidney fibrosis.
We thank Derek DiRocco for help with the unilateral ischemia reperfusion experiments and Vanesa Bijol for help scoring fibrosis severity.
Generation of an Indian hedgehog (Ihh)-nuclear LacZ (nLacZ) reporter allele. A: To construct a positive/negative targeting vector, a 4.7-kb EcoRI/NcoI fragment containing an upstream noncoding sequence and ending at the Ihh ATG start codon was used as a 5′ region of homology and inserted into the pPGKneo/TK-2 vector, in-frame upstream of an (NLS-LacZ-pA)-LoxP-(PGK–neo–pA)-LoxP cassette. A 4.0-kb SacI/EcoRI fragment was used as a 3′ region of homology and inserted upstream of the MC1–tk–pA cassette of the targeting vector. This fragment contains the end of the first exon and most of the first intron of Ihh. R1 embryonic stem cells were electroporated with SalI-linearized targeting vector and selected in geneticin and fialuridine according to standard protocols. Drug-resistant colonies were screened by Southern blot analysis of HindIII-digested genomic DNA probed with a HindIII/EcoRI fragment located upstream of the 5′ homology region. B: Correct targeting was confirmed by probing an AccI digest of genomic DNA with a fragment located within the coding sequence of exon-3, downstream of the 3′ region of homology. Germ-line chimeras were generated by injection of targeted embryonic stem cells into C57Bl/6 blastocysts. Carriers of the targeted allele were crossed with E2a-Cre transgenic mice to induce removal of the neo cassette by Cre-mediated recombination. The structure of the properly floxed Ihh-nLacZ reporter allele was confirmed by Southern blot hybridization of AccI-digested genomic DNA and by PCR analysis using primers flanking the PGK-neo-pAcassette.
Description of reporter mice. A: Indian hedgehog (Ihh)-nLacZ mice were created using gene-targeting techniques and Patched1 (Ptch1)-nLacZ, Gli1-nLacZ, and Gli2-nLacZ reporter mice were obtained from JAX Mice (Bar Harbor, ME). These reporter mice have a nuclear-localized LacZ reporter allele knocked into the start codon for each respective allele. All mice used in the current study were heterozygotes. Nuclear LacZ expression reports active transcription at the Ihh, Ptc1, Gli1, or Gli2 locus at the time of sacrifice. B: The sonic hedgehog (Shh) reporter mice (Shh-GFPCre; R26-LacZ) express a green fluorescent protein-Cre recombinase fusion protein under control of the Shh locus. Activation of transcription leads to recombination of the R26LacZ reporter allele, leading to continuous and heritable expression of cytoplasmic LacZ. With this model, LacZ expression reflects current or historical Shh expression.
Urothelial expression of sonic hedgehog (Shh) in the lower urinary tract. High-powered view of 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-gal) staining from an adult Shh-GFPCre; R26LacZ ureter reveals staining confined to urothelium within the ureter. Scale bar = 50 μm.
Pseudocoloring epithelial LacZ expression with immunofluorescence images. Although anti-LacZ antibodies reliably labeled LacZ-expressing interstitial cells, they did not reliably label LacZ-expressing tubular epithelial cells due to high epithelial autofluorescence. In certain situations, therefore, the 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-gal) stain was performed first, followed by direct or indirect immunofluorescence. The X-gal stain was converted to a pseudocolor and merged with the fluorescence signal from the same section. Scale bar = 50 μm. Aq2 = aquaporin 2; Ihh = Indian hedgehog; LTL = Lotus tetragonolobus lectin; nLacZ = nuclear LacZ; Shh = sonic Hedgehog.
Indian hedgehog (Ihh)-nuclear LacZ (nLacZ) is expressed in some thin limbs of Henle. A: 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-gal) staining of inner medulla from kidney sections from Ihh-nLacZ mice demonstrates positive staining of tubular epithelial cells with squamous morphology lacking brush borders (arrowheads), consistent with thin limbs of Henle. B: X-gal staining of kidney sections from Ihh-nLacZ mice does not colocalize with the collecting duct markers aquaporin 2 (AQP2) and Dolichos biflorus agglutinin (DBA), the thick ascending limb marker Na-K-2Cl cotransporter (NKCC), or the endothelial marker CD31. Scale bar = 50 μm.
Some endothelial cells express Patched1 (Ptch1)-nuclear LacZ (nLacZ). Kidney sections from Ptch1-nLacZ mice were stained for both LacZ and CD31 expression by indirect immunofluorescence. The arrows denote three CD31-positive endothelial cells that costain for nuclear LacZ in the lumen of a large artery. Nuclear LacZ-expressing perivascular fibroblasts are designated by arrowheads. Scale bar = 25 μm.
Dramatic increase in interstitial α-SMA provides histological confirmation of robust fibrotic response in unilateral ischemic reperfusion injury. Day 14 contralateral (CLK) and unilateral ischemia reperfusion injury (UIRI) kidney sections from C57BL/6 mice were labeled with anti-α-smooth muscle actin (α-SMA)-fluorescein isothiocyanate and counterstained with DAPI. Scale bar = 50 μm.
During unilateral ureteral obstruction (UUO), the number of Gli2- and Patched1 (Ptch1)-expressing cells increases, whereas the number of Indian Hedgehog (Ihh)-expressing cells is unchanged. A: Low-power images of 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-gal) staining of medulla from kidney sections from Gli2-nLacZ mice show a progressive increase in X-gal staining with UUO. B: Lower-power images of X-gal staining of cortex from kidney sections from Ptch1-nLacZ mice shows an increase in X-gal staining at UUO day 7 compared to uninjured day 0 kidneys. C: Lower-power images comprised of inner cortex (c) and outer medulla (m) from X-gal stained kidney sections from Ihh-nLacZ mice shows similar amounts of X-gal staining in the cortex during UUO. Scale bar = 50 μm.
A: The presence of sonic hedgehog (Shh) protein in conditioned media from Cos7 cells stably transfected with pcDNA-N-Shh is confirmed by Western blot. B: mRNA expression of Gli2 by quantitative PCR and normalized to glyceraldehydes-3-phosphate dehydrogenase in 10T1/2 cells increases minimally after stimulation with Shh-conditioned media compared to the 153.9 ± 18.2-fold increase observed for Gli1 (Figure 6) and does not increase after stimulation with 500 nmol smoothened agonist (SAG). Gli3 expression in 10T1/2 cells did not increase after stimulation with either Shh-conditioned media or SAG. N = 3 per condition. *P < 0.05, **P < 0.005, and ***P < 0.001 by two-tailed Student's t-test.
Hedgehog signaling in animal development: paradigms and principles.
Supported by the NIH ( RO1 DK088923 to B.D.H. and T32 training grant DK007527 to S.L.F) and the Harvard Stem Cell Institute (B.D.H).
Disclosures: K.A.W. holds stock options or bond holdings in a for-profit corporation or self-directed pension plan (Infinity Pharmaceuticals) and is also a full-time employee of Infinity Pharmaceuticals. None of the other authors disclosed any conflict of interest.