Negative Regulation of TGFβ Signaling by Stem Cell Antigen-1 Protects against Ischemic Acute Kidney Injury

Acute kidney injury, often caused by an ischemic insult, is associated with significant short-term morbidity and mortality, and increased risk of chronic kidney disease. The factors affecting the renal response to injury following ischemia and reperfusion remain to be clarified. We found that the Stem cell antigen-1 (Sca-1), commonly used as a stem cell marker, is heavily expressed in renal tubules of the adult mouse kidney. We evaluated its potential role in the kidney using Sca-1 knockout mice submitted to acute ischemia reperfusion injury (IRI), as well as cultured renal proximal tubular cells in which Sca-1 was stably silenced with shRNA. IRI induced more severe injury in Sca-1 null kidneys, as assessed by increased expression of Kim-1 and Ngal, rise in serum creatinine, abnormal pathology, and increased apoptosis of tubular epithelium, and persistent significant renal injury at day 7 post IRI, when recovery of renal function in control animals was nearly complete. Serum creatinine, Kim-1 and Ngal were slightly but significantly elevated even in uninjured Sca-1-/- kidneys. Sca-1 constitutively bound both TGFβ receptors I and II in cultured normal proximal tubular epithelial cells. Its genetic loss or silencing lead to constitutive TGFβ receptor—mediated activation of canonical Smad signaling even in the absence of ligand and to KIM-1 expression in the silenced cells. These studies demonstrate that by normally repressing TGFβ-mediated canonical Smad signaling, Sca-1 plays an important in renal epithelial cell homeostasis and in recovery of renal function following ischemic acute kidney injury.


Introduction
Hospital-associated acute kidney injury (AKI) remains a significant clinical problem worldwide [1], affecting~15% of all hospitalized patients [2,3]. In the United States, more than 3 million hospitalized patients are at risk of AKI each year [4]. AKI once considered an incident (GTGGGAGTAGTGTGTGAAATA) and C8 (AGGCAGCAGTTATTGTGGATT) displayed the most significant knockdown of Sca-1 (S2 Fig), and were used for further analysis. Sca-1 silenced cell lines were maintained in the above growth media and supplemented with 2μg/ml puromycin (Invitrogen). Cells were treated with recombinant human TGFβ 1 (R&D Systems) at a concentration of 5ng/ml for the indicated times, and with the TβRI inhibitor SB431542 (Selleckchem) at 5μM concentration.

Quantification of kidney injury
Tissue injury was scored on a 1-6 scale by adding parameters for severity and extent of tubular injury, as previously published [25]. The scores were assigned as follows. Severity of tubular injury: 0, no injury; 1, mild injury with mild attenuation of the epithelium and a loss of brush border on PAS stain; 2, moderate injury with marked attenuation of the epithelium but without frank denudation of basement membrane; and 3, severe injury with denudation of basement membrane. Extent of tubular injury: 0, no injury; 1, small isolated foci of injured epithelium; 2, confluent areas of injured epithelium but without uniform confluent involvement of corticomedullary junction; and 3, diffuse injury involving the entire cortico-medullary junction. Intermediate scores were assigned when appropriate, e.g. an isolated tubule with very mild injury would get a score of 0.5 + 0.5 = 1. Whole sagittal sections across the middle section of the entire kidney were from 2 normal and 3 Sca-1 -/animals 7 days post-IRI, as well as from uninjured control and Sca-1 -/animals were scored by the pathologist (A.V.) in a blinded fashion. For quantification of leukocyte infiltration into IRI kidneys, 10 high-powered fields were analyzed across two different kidney sections from each animal.

Real-time PCR
To isolate RNA from kidneys, tissue was minced and incubated with collagenase at 37°C for 30 minutes followed by RNA extraction using RNeasy plus kit (Qiagen). cDNA was produced using ProtoScript II reverse transcriptase (New England Biolabs). Real-time PCR was performed on Stratagene MX4000 using iQ SYBR Green supermix (Biorad).

Statistical Analysis
The significance of the difference between experimental groups was determined by analysis of variance followed by a one-tailed Student's t test. Data are expressed as the mean ± SD, with Pvalues of < 0.05 considered significant.

Expression of Sca-1 in the kidney of adult mice
Although Sca-1 (Ly6a) expression has been detected in the adult murine kidney sometime ago [21,22], a detailed description of its tissue distribution has not been undertaken. To determine which renal cells express Sca-1, we utilized a transgenic mouse with EGFP under the control of the Sca-1 promoter, as well as Sca-1-specific antibody immunofluorescence. In Sca-1-EGFP transgenic mice, EGFP expression was prominently detected in tubular epithelium of proximal tubules (Fig 1A-1C), loop of Henle (S1A-S1C Fig), and distal tubules (S1D-S1F Fig), but not in the collecting duct (S1G-S1I Fig). Similar results were obtained using a Sca-1 antibody ( Fig  1D-1F, and data not shown), which also revealed the apical localization of Sca-1 protein on proximal tubular cells (Fig 1D-1F).

Loss of Sca-1 increased kidney injury following IRI
Kidney development appeared normal in Sca1 -/mice, and kidneys from adult Sca1 -/mice were visually similar to wild-type kidneys (data not shown), suggesting that Sca-1 plays little role in kidney development. To explore the potential role of Sca-1 in the adult kidney, we evaluated renal function following unilateral ischemia/reperfusion injury (IRI) with contralateral nephrectomy. Similar to wild-type animals, Sca-1 -/animals displayed peak serum creatinine levels 24 hours following IRI, and began to recover toward baseline levels by 72 hours post-IRI (Fig 2A). However, serum creatinine levels rose again in Sca-1 -/mice by day 7 post-IRI (0.47 mg/dl ± 0.1 ; Fig 2A), when wild-type mice were showing further recovery (0.26 mg/dl ±0.09, p = 0.002; Fig 2A). Interestingly, even uninjured Sca-1 -/mice displayed a slight but significant elevation in levels of serum creatinine when compared to uninjured wild-type mice (0.170 ± 0.04 mg/dL and 0.093 ± 0.04 mg/dL, respectively; p = 0.0019) (Fig 2A). Sca-1 heterozygotes behaved similar to wild-type mice after IRI (unpublished data).
Both Kim-1 and Ngal mRNAs were highly induced following injury in wild type and Sca-1 -/kidneys (Fig 2B and 2C). However, whereas Kim-1 and Ngal mRNA levels progressively declined in wild-type animals over the next 6 days, high levels of Kim-1 expression persisted in Sca-1 -/mice throughout the first week following injury. By day 7 post-IRI, Sca-1 -/kidneys had a greater than 10-fold increase in Kim-1 expression compared to wild-type kidneys ( Fig  2B), indicating impaired renal recovery. Low but significant levels of Kim-1 and Ngal mRNA were detected in uninjured Sca-1 -/kidneys ( Fig 3A), consistent with elevated serum creatinine (Fig 2A), suggesting that Sca-1 -/kidneys are intrinsically susceptible to IRI. The significant increase in Kim-1 expression in uninjured Sca-1 -/kidneys and 7 days post IRI was confirmed by Western blot (Fig 3B). In wild-type animals, Sca-1 expression was significantly increased by day 7 post-IRI (Fig 3C), coincident with recovery of renal function (Fig 2A) and pathological (Figs 2A and 3D) kidney injury indices. In contrast, Sca-1 -/kidneys displayed increased tubule damage 7 days post injury compared to controls (Fig 3D), consistent with the rise in serum creatinine (Fig 2A). These results suggest that Sca-1 plays an important role in maintenance of epithelial cell function under baseline conditions and in the recovery phase of IRI.
Canonical smad signaling is constitutive in Sca-1 silenced renal proximal tubular cells Minimal phospho-activation of Smad3 (p-Smad3) was detected in serum-starved wild type TKPTS cells in the absence of ligand, but silencing of Sca-1 in D5 and C8 cell lines increased p-Smad3 levels by~5-fold ( Fig 4C). Consistently, Smad2/3 displayed a predominantly cytoplasmic localization in wild type cells but was primarily nuclear in Sca-1-silenced cells (Fig 4D). The increase in p-Smad3 and nuclear localization of Smad2/3 in Sca-1-silenced cells in the absence of ligand, suggests that TGFβ signaling is constitutive in these cells. We then tested if TβRI and TβRII form a complex in serum-starved Sca-1-silenced cells. As expected, immunoprecipitating TβRII resulted in barely detectable TβRI in wild-type TKPTS cells (Fig 4E) in the absence of ligand. In contrast, the TβRI/TβRII complex formed spontaneously in D5 and C8 cells (Fig 3E). Collectively, these data show that knockdown of Sca-1 in mouse proximal tubules results in ligand-independent formation of a signaling TβRI/TβRII complex.

Selective regulation of TGFβ signaling by Sca-1 in proximal tubule cells
We next compared the response of wild type, D5-, and C8 TKPTS cells to the TGFβ 1 ligand. Cells were incubated with recombinant TGFβ 1 for various time periods, and Smad3 phosphorylation analyzed by western blotting. As noted earlier (Fig 4C), p-Smad3 was constitutively expressed in D5 and C8 cells but minimally in wild type TKPTS in the absence of ligand (Fig 5A). TKPTS cells showed a time-dependent increase in p-Smad3 in response to TGFβ 1 (Fig 5A). However, minimal changes in p-Smad3 levels took place in D5 and C8 cells following exposure to TGFβ 1 (Fig 5A). Preincubation of wild type TKPTS cells with the TβRI inhibitor SB431542 [32] blocked TGFβ 1 -dependent phosphorylation of Smad3 as well as ligand-independent Smad3 phosphorylation in Sca-1-silenced cells (Fig 5B), indicating that ectopic activation of Smad3 in absence of Sca-1 was also TβRI-mediated. mRNA and protein levels of plasminogen activator inhibitor-1 (Pai-1) did not change in Sca-1 silenced D5 and C8 proximal tubular epithelial cells (Fig 5C and S3A Fig). TGFβ 1 also induces Pai-1 mRNA expression [33] via a TGFβ 1 -directed ERK1/2 signaling [34]. In response to TGFβ 1 , Pai-1 mRNA rose to equivalent levels in wild type and Sca-1-silenced cells (Fig 5C), suggesting that Sca-1 does not regulate this arm of TGFβ 1 signaling. To assess if Sca-1/TβR also regulates noncanonical TGFβ 1 -directed p38 signaling, activation of p38 MAPK was examined Smad ratios of duplicate samples from TKPTS, D5 and C8 were 0.33, 1.55 and 1.76, respectively. (D) Loss of Sca-1 expression increased Smad2/3 nuclear localization in C8 cells. Immunostaining of control TKPTS and C8 (Sca-1 KD) cells showing Smad2/3 localization (green), with actin detected with Alexa555 phalloidin (red), and nuclei labeled with DAPI (blue). Scale bar = 20μm. (E) A representative experiment, one of two, of a Western blot of TRII immunoprecipitates from serum-starved TKPTS, D5, and C8 cells, detected with anti-TRI antibody. Silencing Sca-1 in D5 and C8 cells resulted in constitutive TRI/TRII complex formation in the absence of ligand. doi:10.1371/journal.pone.0129561.g004 Role of Sca-1 in Kidney Epithelium in Sca-1 -/kidneys before and after IRI as well as in Sca-1-silenced D5 and C8 cells. We found no significant change in phosphorylated p38 MAPK in Sca-1 -/kidneys or Sca-1 silenced proximal tubular epithelial cells (S3B and S3C Fig), suggesting that Sca-1 mainly influences TGFβ 1directed canonical Smad signaling in kidney epithelium.
Smad signaling in kidneys of Sca-1 -/mice after IRI Levels of pSmad3 trended to be higher in uninjured kidneys of Sca1 -/animals vs. controls at baseline (Fig 6A), but the differences did not reach statistical significance (p = 0.06). A similar trend was also observed during the first 72 hours following IRI (data not shown). However, by day 7 post-IRI, kidneys from Sca-1 -/animals had significantly higher levels of phospho-Smad3 compared to wild-type kidneys (Fig 6A), as well as increased mRNA expression of Pai-1 ( Fig  6B). Comparisons between wild type and Sca-1 -/animals at 7 days post-IRI, showed a significant increase in the number of apoptotic (TUNEL-positive) cells in Sca-1 -/tubular epithelium (Fig 6C and 6D; p = 0.0125), but no significant change in cell proliferation (Ki67-positive tubular epithelial cells) (Fig 6E). In addition, there was no significant difference in the number of infiltrating F4/80 positive macrophages or CD3 positive T-cells in the kidneys of Sca-1 -/animals 7 days post-IRI when compared to control kidneys (Fig 6F and 6G). Taken together, these data suggest that aberrant activation of canonical Smad signaling in injured Sca-1 -/kidney epithelium likely accounts for increased epithelial cell apoptosis observed following IRI.

Discussion
TGFβ signaling is a key mediator of renal scarring that ultimately leads to kidney failure [35]. Overexpression of TGFβ in the nephron can lead to cell death in proximal tubules [28] and interstitial fibrosis [36], while blocking TGFβ activity can reduce injury after AKI [37]. TGFβ1 mediates progressive renal fibrosis by inducing cell cycle arrest [38], and by stimulating the synthesis of several key fibrotic genes, such as those encoding Pai-1, collagens, fibronectin, connective tissue growth factor (CTGF), and tissue inhibitor of metalloproteinases, thus enhancing ECM production while inhibiting its degradation [39,40]. TGFβ may also mediate renal fibrosis by inducing the transformation of tubular epithelial cells into myofibroblasts through epithelial-mesenchymal transition (EMT) [41].
In this study, we show that Sca-1 protein is heavily expressed in specific renal tubular epithelial segments, especially of the proximal tubule, a nephron segment highly sensitive to IRI. Sca-1 protein expression was examined using transgenic mice, since Sca-1 antibodies are less than ideal for Western analysis and immunostaining of whole organs due to redundancy among the proteins encoded by Ly6 genes [11]. We find that epithelial Sca-1 plays an important homeostatic function under baseline conditions. In its absence or deficiency, kidney tubular epithelium showed elevated levels of Kim-1 and Ngal, suggesting that it is more susceptible to injury. Indeed, in a model of unilateral IRI and contralateral nephrectomy, Sca-1 -/animals had elevated serum creatinine, increased expression of Kim-1 and Ngal, renal tubular epithelial cell injury and increased apoptosis. We did not find a significant increase in infiltrating macrophages or CD3 + lymphocytes in Sca-1 -/vs. control animals, suggesting that Sca-1 null leukocytes do not play a significant role in renal injury. We traced the mechanism of kidney epithelial cell injury to upregulation of epithelial TGFβ receptor signaling. In cultured renal proximal tubular epithelial cells, Sca-1 interacted with TβRI and TβRII in the absence of ligand. In presence of ligand, Sca-1/TβRI interaction was maintained but that between Sca-1 and TβRII was lost. In Sca-1 deficiency states (in null animals or silenced tubular cells), upregulation of TGFβ receptor signaling was primarily mediated via increased canonical Smad signaling, which paralleled the rise in Kim-1 levels, suggesting a cause and effect relationship. There was no contribution detected from the noncanonical TβR-mediated MEK/ERK or p38 pathway [42].
TGFβ signaling is initiated when this ligand binds the high affinity serine/threonine kinase TβRII, allowing TβRII to complex with and activate the low affinity TβRI, which in turn promotes serine phosphorylation of Smad2/3, their association with Smad4 and translocation to the nucleus, where they trigger transcription of profibrotic genes [40]. Our data show that under baseline conditions, Sca-1 binds to both TβRII and TβRI in the absence of ligand, but only to TβRI in its presence. In contrast to our findings in renal epithelium, Sca-1 did not bind TβRII and bound only to TβRI in the absence of ligand in a mammary adenocarcinoma cell line [30], suggesting that Sca-1 interaction with TGFβ receptors is cell context-specific. Association of TβRI with TβRII is mediated by the ectodomain of each [43], a three-fingered protein domain (TFPD) also found in Sca-1 and other Ly6 protein family members [44]. A distinct site in TβRII ectodomain is used to bind ligand. Our data in renal epithelium suggest that Sca-1 may occupy the same (or an overlapping) TGFβ binding site in TβRII, and is thus displaced in presence of TGFβ. Superimposing the TFPDs of TβRII and TβRI using Chimera [17] show that the region in TβRI corresponding to the TGFβ binding-site in TβRII remains accessible to Sca-1 in the ternary TGFβ /TβRII/TβRI complex.
Functionally, Sca-1 suppressed canonical Smad signaling in uninjured or injured renal epithelium, but did not alter TGFβ-directed noncanonical signaling via ERK1/2 or p38 MAPK. These data suggest the following model: in the basal state, binding of epithelial Sca-1 to TβRI and TβRII keeps the two receptors apart on the cell surface. Following ischemic injury, induced TGFβ displaces Sca-1 from TβRII allowing its phospho-activation perhaps by Src [42]. However, continued occupancy of TβRI by Sca-1 prevents its phospho-activation by TβRII, perhaps caused by steric or allosteric effects, thus preventing TβRI-mediated canonical Smad signaling. TGFβ-bound TβRII can still activate downstream noncanonical MAPK signaling [42]. Taken together, our findings suggest that inhibition of canonical Smad signaling accounts for the role of epithelial Sca-1 in the preservation and recovery of renal function. Consistent with this interpretation is data showing that homozygous knockout of Smad3 protects against ischemic AKI in mice [45], and expression of a constitutively active form of TβRI in the proximal tubule resulted in epithelial cell injury and apoptosis [28].
Sca-1 is located within a cluster of related Ly6 genes on mouse chromosome 15, which is syntenic to human chromosome 8q24.3 [11,46]. The segment containing Sca-1 (Ly6A) was deleted between mouse and rat speciation, thus no obvious Sca-1 homolog is known in humans. Given the important roles Sca-1 plays in mediating stem-and differentiated cell stress responses in mice, it is likely that these roles are assumed by one or a number of the eleven remaining Ly6-related genes on human chromosome 8 [46]. The identity of the functional homolog(s) of Sca-1 in humans and whether it also forms Ly6/TβR regulatory complexes remains to be determined. Western detection of phospho-p38 in normal and Sca1 -/kidneys from three animals in each case before and 7 days post-IRI. p-p38/p38 ratios in NL and Sca-1-/-kidneys before or at 7days post injury were not different (p = 0.329 and p = 0.131, respectively). (C) Western detection of phosho-p38 in TKPTS and the Sca-1 silenced cell lines D5 and C8. p-p38/p38 ratios for wild type, D5 and C8 were 1.44, 1.74 and 1.77, respectively. One of two experiments is shown. (TIF)