Selective activation of epidermal growth factor receptor in renal proximal tubule induces tubulointerstitial fibrosis

Abstract

Epidermal growth factor receptor (EGFR) has been implicated in the pathogenesis of diabetic nephropathy and renal fibrosis; however, the causative role of sustained EGFR activation is unclear. Here, we generated a novel kidney fibrotic mouse model of persistent EGFR activation by selectively expressing the EGFR ligand, human heparin-binding EGF-like growth factor (hHB-EGF), in renal proximal tubule epithelium. hHB-EGF expression increased tyrosine kinase phosphorylation of EGFR and the subsequent activation of downstream signaling pathways, including ERK and AKT, as well as the profibrotic TGF-β1/SMAD pathway. Epithelial-specific activation of EGFR was sufficient to promote spontaneous and progressive renal tubulointerstitial fibrosis, as characterized by increased collagen deposition, immune cell infiltration, and α-smooth muscle actin (α-SMA)–positive myofibroblasts. Tubule-specific EGFR activation promoted epithelial dedifferentiation and cell-cycle arrest. Furthermore, EGFR activation in epithelial cells promoted the proliferation of α-SMA⁺ myofibroblasts in a paracrine manner. Genetic or pharmacologic inhibition of EGFR tyrosine kinase activity or downstream MEK activity attenuated the fibrotic phenotype. This study provides definitive evidence that sustained activation of EGFR in proximal epithelia is sufficient to cause spontaneous, progressive renal tubulointerstitial fibrosis, evident by epithelial dedifferentiation, increased myofibroblasts, immune cell infiltration, and increased matrix deposition.—Overstreet, J. M., Wang, Y., Wang, X., Niu, A., Gewin, L. S., Yao, B., Harris, R. C., Zhang, M.-Z. Selective activation of epidermal growth factor receptor in renal proximal tubule induces tubulointerstitial fibrosis.

Excessive tissue scarring or fibrosis is a critical contributor to chronic diseases in many organs, including lung, liver, skin, and kidney, all of which ultimately lead to organ failure (1, 2). The prevalence of chronic kidney disease (CKD) in the U.S. population is estimated to be more than 20 million people, with levels projected to increase over time as a result of the increasing prevalence of common risk factors, such as diabetes and hypertension (3, 4). Despite the overwhelming contribution of tubulointerstitial fibrosis to renal morbidity and mortality, there are currently no U.S. Food and Drug Administration–approved therapies (4, 5).

Kidney proximal tubule cells are vulnerable to injury but retain the ability to repair and regenerate (6). Repair of the epithelium is achieved primarily via the dedifferentiation, migration, and proliferation of surviving epithelial cells to maintain the functional integrity of the kidney (7); however, in contrast to normal repair, incomplete or maladaptive repair processes can disrupt tissue architecture, which leads to organ fibrosis (8). Regardless of etiology, chronic epithelial injury induces dysregulation of key processes, such as dedifferentiation, signaling activation/repression, proliferation, and secretion of profibrotic factors (9, 10). Epithelial injury/maladaptive repair, fibroblast proliferation, migration and activation, and recruitment of inflammatory cells culminate in fibrotic disease progression (11). Although myofibroblasts have been documented as the major matrix-producing cells in the kidney (12), the mechanisms that underlie the role of proximal tubule epithelial cells in the initiation and maintenance of fibrosis remain less well understood.

Epidermal growth factor receptor (EGFR) has clear pathologic effects in the development of fibrosis in different organs. In fact, we and others have demonstrated that the inhibition of EGFR either genetically or pharmacologically can limit the progression of diabetic nephropathy as well as angiotensin II (Ang II)–infused or unilateral ureteral obstruction (UUO)–induced kidney fibrosis (13–16). Moreover, Ang II and TGF-β—two major pathways involved in renal fibrosis—can transactivate EGFR in proximal tubule cells (13, 14), whereas EGFR activation also leads to increased proximal tubule TGF-β expression (14). The role of sustained EGFR activation in the renal tubule is unknown. The EGF family consists of the members, EGF, heparin-binding EGF-like growth factor (HB-EGF), TGF-α, amphiregulin, betacellulin, and epiregulin, all of which can promote EGFR tyrosine kinase phosphorylation (17). HB-EGF, as with other members of this family, is a membrane-anchored growth factor that is cleaved by metalloproteinase activity, which allows the soluble molecule to bind to EGFR and activate downstream signaling cascades that are necessary for specific cellular phenotypes (17–19).

We created homozygous transgenic human HB-EGF (hHB-EGF^Tg/Tg) mice that express the EGFR ligand, hHB-EGF, selectively in the proximal tubule, which demonstrates that epithelial-specific persistent EGFR activation is sufficient to initiate tubular dysfunction, likely via dedifferentiation and cell cycle arrest. Furthermore, epithelial cell–derived paracrine factors drive epithelial–fibroblast communication, which further exacerbates renal fibrosis. Pharmacologic or genetic inhibition of EGFR tyrosine kinase activity reduced the fibrotic burden in hHB-EGF^Tg/Tg mice. This transgenic model identifies EGFR activation as an integral factor for the development of tubular dysfunction (e.g., dedifferentiation, growth arrest), which leads to the secretion of factors that crosstalk with matrix-producing interstitial cells and provides a novel kidney fibrotic model and platform with which to implement therapeutic interventions in the development and regression of fibrosis.

MATERIALS AND METHODS

Generation of mice with sustained EGFR activation by specifically expressing hHB-EGF in the proximal tubular epithelia

Previously described heterozygous transgenic hHB-EGF (hHB-EGF^+/Tg) mice (20) were bred onto a C57BL/6 background—a relatively fibrosis-resistant mouse strain. The Cre recombinase controlled by the γ-glutamyl transpeptidase promoter (γGT-Cre) allows for the successful targeting of hHB-EGF expression to the renal proximal tubule. Gene expression usually commences 10–14 d postnatally. Male hHB-EGF^Tg/Tg mice were genotyped by using PCR and used for subsequent experiments. For proof of principle studies, hHB-EGF^Tg/Tg mice were crossed with Waved-2 (Wa-2) mice to generate hHB-EGF^Tg/Tg; Wa-2 mice, which have deficient EGFR tyrosine kinase activity.

Animal studies

All animal experiments were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee of Vanderbilt University. For pharmacologic inhibitor experiments, 4-wk-old hHB-EGF^Tg/Tg mice were administered the EGFR tyrosine kinase inhibitor, erlotinib (80 mg/kg; LC Laboratories, Woburn, MA, USA) or the MEK inhibitor, PD 0325901 (50 mg/kg; Cayman Chemical, Ann Arbor, MI, USA) by daily gastric gavage until up to age 14 wk. Control animals were treated with water (for erlotinib) or DMSO (for PD 0325901) alone. Animals were anesthetized with nembutal (70 mg/kg i.p.) and administered heparin (1000 U/kg, i.p.) to minimize coagulation. One kidney was removed for immunoblotting, RT-PCR, and histologic studies, whereas the other kidney was perfused with FPAS (3.7% formaldehyde, 10 mM/ sodium-m-periodate, 40 mM phosphate buffer, and 1% acetic acid) via cannulation of the aortic trunk using the left ventricle. The fixed kidney was dehydrated via a graded series of ethanol, embedded in paraffin, sectioned at 4-µm thickness, and mounted on glass slides.

Cell culture

Human renal proximal tubule epithelial cells (hRPTECs/TERT1 cells) from American Type Culture Collection (CRL-4031; Manassas, VA, USA) were propagated in 10% fetal bovine serum (FBS) DMEM:F12 medium that was supplemented with 5 pM triiodo-l-thyronine,10 ng/ml recombinant human EGF, 3.5 µg/ml ascorbic acid, 5.0 µg/ml human transferrin, 5.0 µg/ml insulin, 25 ng/ml prostaglandin E₁, 25 ng/ml hydrocortisone, 8.65 ng/ml sodium selenite, 0.1 mg/ml G418, and 1.2 g/l sodium bicarbonate. Mouse renal cortical fibroblasts were cultured in 10% FBS DMEM. For HB-EGF treatment, cells were deprived of serum for 24 h in 0.5% FBS DMEM, then exposed to HB-EGF (25 ng/ml; 24 h) in the presence or absence of erlotinib as indicated. For epithelial–fibroblast communication studies, HB-EGF–treated hRPTECs (1 d) were washed with PBS to remove any recombinant HB-EGF and incubated with serum-free DMEM for 1 d, which allowed cells to condition the medium. hRPTEC-derived conditioned medium was added to quiescent fibroblasts for 2 d with subsequent measurements of cell proliferation by using the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay.

Vybrant MTT cell proliferation assay

MTT cell proliferation assay (Invitrogen, Carlsbad, CA, USA) was performed according to the manufacturer’s instructions. In brief, renal fibroblasts were incubated in MTT solution for 4 h at 37°C, then incubated in SDS-HCl solution for 4 h at 37°C. MTT solution in medium only provided the negative control. Samples were measured at an absorbance wavelength 570 nm.

Reagents and Abs

Rabbit anti–phospho-AKT, rabbit anti–phospho-ERK, rabbit anti–phosho-4E-BP, rabbit anti–phospho-EGFR¹⁰⁶⁸, rabbit anti–phospho-Src, rabbit anti–phospho-p38, rabbit anti–phospho-SMAD3, rabbit anti–phospho-SMAD2, and rabbit anti–phospho-retinoblastoma (Rb) were purchased from Cell Signaling Technology (Danvers, MA, USA). Rabbit anti–phospho-EGFR⁸⁴⁵, rabbit anti–EGFR, goat anti–phospho-EGFR¹¹⁷³, rabbit anti-nitrotyrosine, goat anti–connective tissue growth factor (CTGF), rabbit anti-p21, rabbit anti–phosho-p53^Ser15, and rabbit anti–phospho-histone H3^Ser10 were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). Mouse anti–α-smooth muscle actin (α-SMA) and recombinant HB-EGF were purchased from Sigma-Aldrich (St. Louis, MO, USA). Rabbit anti–Ki-67 was purchased from Abcam (Cambridge, United Kingdom). Rabbit anti–TGF-β was purchased from Bio-Rad (Hercules, CA, USA). Rat anti-CD3 and rat anti-F4/80 were purchased from AbD Serotec (Raleigh, NC, USA). Rabbit anti–collagen I was purchased from Rockland Immunochemicals (Pottstown, PA, USA).

Real-time PCR

Total RNA was isolated from the kidney by using Trizol reagent (Thermo Fisher Scientific, Waltham, MA, USA). SuperScript III First-Strand Synthesis System kit (Thermo Fisher Scientific) was used to synthesize cDNA from equal amounts of total RNA from each sample. Real-time PCR using Taqman Master mix and RNA primers for human HB-EGF (Mm00439307) and mouse glyceraldehyde 3-phosphate dehydrogenase (Mm99999915) was performed by using TaqMan real-time PCR (7900HT; Applied Biosystems, Foster City, CA, USA).

Immunoblotting analysis

Whole kidney tissue was homogenized by using lysis buffer (10 mM Tris–HCl, pH 7.4, 50 mM NaCl, 2 mM EGTA, 2 mM EDTA, 0.5% NP-40, 0.1% SDS, 100 μM Na₃VO₄, 100 mM NaF, 0.5% sodium deoxycholate, 10 mM sodium pyrophosphate, 1 mM PMSF, 10 μg/ml aprotinin, and 10 μg/ml leupeptin) and centrifuged at 15,000 g for 20 min at 4°C. A bicinchoninic acid protein assay kit (Thermo Fisher Scientific) was used to measure the protein concentration of each sample. Western blot analysis of cell or tissue lysates was performed as previously described (20).

Immunohistochemical and immunofluorescent staining and quantitative image analysis

Immunostaining of 1-mo-old mouse kidney tissue was performed as previously described (20). For phosphoprotein staining, antigen retrieval was performed by boiling the section in citric acid buffer (pH 6.0; 100 mM) for 3 × 5 min. For mouse primary Abs, the Mouse on Mouse (M.O.M.) kit (Vector Laboratories, Burlingame, CA, USA) was used to reduce endogenous mouse Ig staining. For α-SMA and Ki-67 (ab15580; Abcam) double fluorescent staining, the section was deparaffinized, antigen retrieval was performed, and the section was blocked with M.O.M. blocking solution for 1 h. The section was incubated with mouse anti–α-SMA Ab, followed by FITC-conjugated anti-mouse IgG. After blocking for 1 h by using 10% normal donkey serum, the section was incubated with an additional primary Ab against Ki-67, followed by Cy3-labeled anti-rabbit IgG. For lotus tetragonolobus lectin (LTL) and phospho-histone H3^Ser10 double fluorescent staining, the section was treated with trypsin for 15 min, blocked for 1 h with 10% normal donkey serum, and incubated with biotinylated LTL (a marker of proximal tubule, B-1325), followed by fluorescein streptavidin. After blocking with 10% normal donkey serum, phospho-histone H4^Ser10 Ab was added, followed by Cy3-labeled anti-rabbit IgG. Staining on slides was imaged with a Nikon TE300 fluorescence microscope and a SpotCam Digital Camera (Diagnostic Instruments, Sterling Heights, MI, USA). Immunostaining color and tissue density were quantified by using BioQuant True Color (v.20; R&M Biometrics, Nashville, TN, USA). Data from 1 animal sample are the average of 5 representative fields that were quantified at original magnification (×160). The histogram represents the average staining intensity of at least 3 mice in each condition.

Masson’s trichrome staining and PicroSirius red staining were performed according to the manufacturer’s protocol (Sigma-Aldrich).

Statistical analyses

Data are presented as means ± sem for at least 3 separate experiments, each in triplicate. An unpaired Student’s t test was used for statistical analysis. A value of P < 0.05 vs. controls was considered statistically significant.

RESULTS

Sustained EGFR activation in the renal proximal tubule is sufficient to initiate spontaneous and progressive renal fibrosis

To investigate the causal role of sustained EGFR activation in the epithelium, B6D2 transgenic mice with selective expression of the EGFR ligand, hHB-EGF, in the renal proximal tubule, driven by a γ-glutamyl transpeptidase promoter (20), were backcrossed onto the C57BL/6 background, a strain that is normally resistant to renal injury. Transgenic mice had no apparent developmental defects. Expression of renal hHB-EGF significantly increased at the mRNA transcript (Fig. 1A) and protein (Fig. 1B) levels, as measured by RT-PCR and Western blot analysis, respectively, in transgenic mice that were homozygous for hHB-EGF expression compared with heterozygous and wild-type mice. Immunohistochemical staining of hHB-EGF showed prominent staining in the renal tubule epithelia of hHB-EGF^Tg/Tg mice, but no evident hHB-EGF staining was found in wild-type mice (Fig. 1C). Of interest, persistent EGFR activation in the tubule resulted in the distortion of the kidney architecture, including tubular dilation and interstitial expansion (Fig. 1C). Kidneys from heterozygous mice were normal and comparable to wild-type mice, even with the modest increase in hHB-EGF expression (data not shown).

Figure 1.

Generation of homozygous transgenic mice with sustained EGFR activation by specifically expressing hHB-EGF in the proximal tubular epithelia. A) Histogram represents the relative increased expression of renal hHB-EGF mRNA in hHB-EGF^Tg/Tg mice (n = 3 mice in each group). ***P < 0.001. B) Immunoblotting analysis determined higher levels of hHB-EGF protein expression in hHB-EGF^Tg/Tg mice compared with control and heterozygous counterparts (n = 3 animals per group). β-Actin served as loading control. C) Immunohistochemical staining of 2-mo-old hHB-EGF kidney tissue sections increased in hHB-EGF^Tg/Tg mice compared with wild-type mice. Original magnification, ×160.

Membrane-tethered HB-EGF acts as an EGFR ligand upon cleavage by matrix metalloproteinases (MMPs) (18, 19). Immunoblotting demonstrated increased phosphorylation of EGFR at Y⁸⁴⁵, Y¹¹⁷³, and Y¹⁰⁶⁸ in hHB-EGF^Tg/Tg mice (Fig. 2A), which was indicative of receptor activation, whereas total EGFR levels remained relatively unchanged (Fig. 2B). Immunohistochemical staining determined that phospho-EGFR at Y⁸⁴⁵ and Y¹¹⁷³ was primarily localized to the proximal tubule of hHB-EGF^Tg/Tg kidneys (Fig. 2C). These data indicate that the expression of epithelial hHB-EGF is sufficient to activate EGFR in the proximal tubule of hHB-EGF^Tg/Tg mice.

Figure 2.

Expression of epithelial hHB-EGF is sufficient to activate EGFR and the related downstream pathways in hHB-EGF^Tg/Tg mice. A) Immunoblotting revealed increased phosphorylation of EGFR at multiple tyrosine sites (Y845, Y1173, and Y1068) in hHB-EGF^Tg/Tg mice compared with wild-type mice. β-Actin acted as loading control (n = 3 mice in each group). B) Total EGFR protein expression in hHB-EGF^Tg/Tg mice was comparable to wild-type mice. β-Actin acted as loading control (n = 3 mice in each group). C) Immunohistochemical analysis demonstrated activated EGFR predominantly in the tubules of 2-mo-old hHB-EGF^Tg/Tg kidneys. D) Immunohistochemical analysis demonstrated an increased activation of ERK (phospho-ERK), SMAD2 (phospho-SMAD2), and SMAD3 (phospho-SMAD3) in the tubules of hHB-EGF^Tg/Tg kidneys. E) Increased phosphorylation of ERK, SMAD2, AKT, p38, 4E-BP, and Src^Y416 occurred in hHB-EGF^Tg/Tg mice compared with wild-type mice as assessed by immunoblotting (n = 3 mice in each group). β-Actin acted as loading control. Original magnification, ×160 (C, D)

Phosphorylated EGFR tyrosine residues serve as docking sites for various signaling molecules that activate intracellular signaling cascades (17). ERK activation increased in the proximal tubules of hHB-EGF^Tg/Tg mice as assessed by immunohistochemistry (Fig. 2D) and immunoblotting (Fig. 2E). Sustained activation of EGFR in the proximal tubule promoted the activation of several other downstream pathways, including PI3K/AKT, p38 MAPK, mammalian target of rapamycin, and Src (Fig. 2E). Moreover, EGFR activation promotes TGF-β expression in a Src^Y416 phosphorylation-dependent manner (14). The TGF-β pathway is a major mediator of fibrotic disease in the kidney (21). Tubule-specific sustained EGFR activation promoted the phosphorylation and localization of TGF-β transcription factors, SMAD2 and SMAD3, in the nuclei of tubule epithelial cells, which suggests the activation of the TGF-β pathway (Fig. 2D, E).

Accumulation of extracellular matrix (ECM) is a pathologic hallmark of tubulointerstitial fibrosis (10). Histologic observations of fibrosis were confirmed by using Masson’s Trichrome staining, which revealed a dramatic increase of collagen deposition in hHB-EGF^Tg/Tg kidneys that was not observed in wild-type littermates (Fig. 3A). Sustained EGFR activation in the proximal tubule initiated tubulointerstitial fibrosis as visualized by PicroSirius red and Masson’s trichrome staining in hHB-EGF^Tg/Tg mice as early as age 3 wk, which progressed up to age 6 mo (Fig. 3B). Glomeruli seemed to be morphologically intact. Immunohistochemistry of kidney sections showed a significant accumulation in collagen I expression in the interstitium of hHB-EGF^Tg/Tg mice compared with wild-type mice (Fig. 3C). Furthermore, immunoblotting of whole kidney lysates demonstrated that tubule-specific EGFR activation increased collagen I expression in hHB-EGF^Tg/Tg kidneys (Fig. 3D). CTGF—a cytokine that is implicated in myofibroblast proliferation and activation of downstream profibrotic signaling (22)—was also increased in the hHB-EGF^Tg/Tg kidney compared with wild-type kidney (Fig. 3D). Of interest, CTGF is also reported to be an EGFR ligand, which thus forms a potential positive feedback to further promote the persistent activation of EGFR (23).

Figure 3.

Sustained EGFR activation in the proximal tubule promoted spontaneous and progressive renal fibrosis. A) Two-month-old male hHB-EGF^Tg/Tg mice had more collagen deposition as assessed by Masson’s trichrome staining compared with age-matched counterparts. Original magnification: ×25 (left), ×160 (right). B) Renal fibrosis evaluated by PicroSirius red and Masson’s trichrome staining of kidney sections from hHB-EGF^Tg/Tg mice (at age 3 wk and 3 and 6 mo) revealed progressive fibrosis. Original magnification, ×160. C) Increased immunohistochemical staining of collagen I was deposited in the interstitium of hHB-EGF^Tg/Tg kidneys. Original magnification, ×160. D) Western blot analysis confirmed the increased expression of both collagen I and CTGF in hHB-EGF^Tg/Tg mice compared with wild-type mice (n = 3 in each group). β-Actin served as loading control.

Sustained EGFR activation in the tubule induces epithelial dedifferentiation and cell cycle arrest in hHB-EGF^Tg/Tg mice

Interstitial fibroblasts are generally recognized as the primary matrix-producing cells in renal fibrotic progression (24). It has become increasingly clear that tubular dysfunction contributes to the initiation and progression of fibrosis; however, the underlying mechanism is less well understood. Studies suggest that epithelial plasticity may convert epithelial cells to a dedifferentiated state (10, 25). Immunoblotting of whole kidney lysates of mice with persistent EGFR activation in the epithelium revealed an increase in the mesenchymal marker, vimentin, and decreases in the epithelial marker, E-cadherin, compared with wild-type littermates (Fig. 4A). Snail and Slug, also known as Snai2, are key transcription factors that regulate the epithelial dedifferentiation program by inhibiting E-cadherin (25). Snail and Slug were markedly increased in the tubule epithelium of hHB-EGF^Tg/Tg kidneys compared with wild-type littermates (Fig. 4A, C), which suggests that tubule-specific EGFR activation is critical for the dedifferentiation of tubules.

Figure 4.

Persistent EGFR activation is required for tubular dedifferentiation in hHB-EGF^Tg/Tg mice. A) Increased vimentin, Snail, and Slug expression, as well as decreased E-cadherin expression, in hHB-EGF^Tg/Tg mice compared with wild-type littermates. α-Tubulin served as loading control. B) Expression of vimentin and E-cadherin as measured by immunoblotting in hHB-EGF^Tg/Tg and Wa-2;hHB-EGF^Tg/Tg mice. α-Tubulin served as loading control. C) Immunoreactivity of Snail and Slug expression increased, whereas E-cadherin expression decreased in the tubular epithelium of hHB-EGF^Tg/Tg kidneys, which was reversed in a Wa-2 background. Original magnification, ×160. D) HB-EGF–induced increased activation of EGFR and TGF-β/SMAD signaling pathways, which was prevented by EGFR tyrosine kinase inhibition using erlotinib in human renal proximal tubule epithelial cells. β-actin confirmed equal loading. E) Western blot analysis demonstrated increased mesenchymal markers, N-cadherin, Slug, Snail, and vimentin, in hRTPECs that were treated with HB-EGF. Erlotinib treatment blocked HB-EGF–induced dedifferentiation markers in hRPTECs. Equal loading was confirmed by β-actin expression.

Wa-2 mice have a genetic point mutation in EGFR that leads to a > 90% reduction in EGFR tyrosine kinase activity (26). We crossed Wa-2 mice with hHB-EGF^Tg/Tg mice to further examine the causative role of sustained EGFR tyrosine phosphorylation in renal fibrosis. Of note, Wa-2;hHB-EGF^Tg/Tg mice had much less Snail, Slug, and vimentin expression, which is consistent with a restoration of E-cadherin expression (Fig. 4B, C), highlighting a requirement of tubule-specific EGFR activation persistence in epithelial dedifferentiation.

TGF-β is a classic inducer of epithelial dedifferentiation (25); however, the role of sustained tubule EGFR activation in differentiation is less well known. HB-EGF treatment activated EGFR and induced TGF-β expression and SMAD3 activation in cultured hRPTECs, all of which were inhibited by the EGFR tyrosine kinase inhibitor, erlotinib (Fig. 4D). Furthermore, the activation of EGFR in hRPTECs elevated the levels of dedifferentiation markers, such as N-cadherin, Slug, Snail, and vimentin, which was attenuated with erlotinib pretreament (Fig. 4E).

Recent studies suggest tubular epithelial cells may undergo cell cycle arrest, which leads to a secretory phenotype that releases CTGF and TGF-β and can potentially act as paracrine signals that activate myofibroblast differentiation and proliferation (27). The role of sustained EGFR activation in tubule growth arrest is unclear. Phosphorylation of Rb promotes cell cycle progression (28). Tubule-specific EGFR activation in hHB-EGF^Tg/Tg mice decreased phospho-RB compared with wild-type mice (Fig. 5A), which indicates incomplete cell cycle progression. Dual immunofluorescence staining of the proximal tubule marker, LTL, and phospho-histone H3^Ser10 revealed that LTL⁺ (green) proximal tubule cells colocalized with phospho-histone H3^Ser10 (red; Fig. 5A), which suggests that EGFR activation in the epithelium promotes tubular growth arrest in fibrotic hHB-EGF^Tg/Tg kidneys. Moreover, tubule epithelial cells of hHB-EGF^Tg/Tg mice had increased α-SMA expression, which is consistent with a dedifferentiated state (Fig. 5B). Immunofluorescent detection of interstitial α-SMA⁺ cells also had increased Ki-67⁺ staining (red; Fig. 5B), which highlights a higher proliferative capacity of myofibroblasts in hHB-EGF^Tg/Tg mice compared with wild-type littermates.

Figure 5.

Sustained EGFR activation leads to proximal tubular cell cycle arrest both in vivo and in vitro. A) Immunohistochemical image represents the loss of phospho-Rb in hHB-EGF^Tg/Tg mice compared with wild-type mice. Dual immunofluorescence staining of kidney sections using Abs against LTL (green, a marker of proximal tubule) and phospho-histone H3^Ser10 (red) revealed higher colocalization (red arrows) in hHB-EGF^Tg/Tg kidneys. Nuclei were visualized by DAPI (blue). Graph depicts the number of cells that stained positive for both LTL and phospho-histone H3^Ser10 (n = 3 in each group). B) Colocalization of interstitial staining of α-SMA (green, a marker of myofibroblasts) and Ki-67 (red, a marker of cell proliferation) was visualized in the kidney by using immunofluorescence. Arrows mark colocalization of Ki67 and α-SMA in interstitial cells. Bar graph quantifies cells that stained positive for both α-SMA and Ki-67 (n = 3 in each group). C) Immunoblotting of whole kidney lysates revealed an increase in cell cycle regulators, phospho-p53^Ser15 and phospho-histone H3^Ser10, in hHB-EGF^Tg/Tg kidneys that was abrogated in erlotinib-treated hHB-EGF^Tg/Tg mice (n = 4 animals per group). β-Actin served as loading control. D) Immunoblotting of cell cycle markers, phospho-p53^Ser15, phospho-histone H3^Ser10, and p21, in HB-EGF–stimulated hRPTECs with or without erlotinib pretreatment. β-Actin was used to determine equal loading. Original magnification, ×160. **P < 0.01, ***P < 0.001.

Immunoblotting of whole kidney lysates of hHB-EGF^Tg/Tg mice demonstrated increased the growth arrest markers, phospho-p53^Ser15 and phospho-histone H3^Ser10 (Fig. 5C). Phosphorylation of p53^Ser15 and histone H3^Ser10 was attenuated in hHB-EGF^Tg/Tg mice that were treated with erlotinib (Fig. 5C). Of interest, in vitro sustained activation of EGFR increased the expression of cell cycle regulators, phospho-p53^Ser15, phospho-histone H3^Ser10, and p21 in hRPTECs (Fig. 5D). Moreover, erlotinib pretreatment reduced the phosphorylation of p53 and histone H3 and expression of p21 (Fig. 5D).

Paracrine factors released from tubule epithelial cells promote myofibroblast proliferation and activation

The conversion to a pathologic secretory phenotype is one common outcome of tubule dedifferentiation and cell cycle arrest (27, 29). Identification of paracrine factors that are released from tubular cells that are critical for communication with fibroblasts is a novel approach to prevent continuous fibrotic signaling. To address this possibility, serum-deprived hRPTECs were treated with HB-EGF for 1 d, washed with PBS to remove any recombinant HB-EGF, and incubated for 24 h in serum-free DMEM, which allowed epithelial cells to condition the media before being added to quiescent fibroblasts for 2 d (Fig. 6A). Conditioned medium from hRPTECs with persistent EGFR activation promoted increased fibroblast proliferation compared with fibroblasts that were incubated in conditioned medium from quiescent hRPTECs (Fig. 6B). Moreover, conditioned medium derived from EGFR-activated hRPTECs that were treated with erlotinib was unable to promote increased fibroblast proliferation (Fig. 6B). Sustained EGFR activation in the tubule epithelium of hHB-EGF^Tg/Tg mice had increased immunostaining of α-SMA⁺ myofibroblasts in the interstitium, with a distinct expression surrounding the proximal tubules (Fig. 6C). Furthermore, exacerbated inflammation is a common feature that is associated with renal fibrosis that may perpetuate fibrotic progression (10, 19). hHB-EGF^Tg/Tg mice had increased infiltration of both F4/80⁺ macrophages (Mϕs) and CD3⁺ T cells in the kidney (Fig. 6D). Collectively, these data suggest that sustained EGFR activation in tubules is sufficient to promote an increase in myofibroblasts and inflammatory cells.

Figure 6.

Sustained EGFR activation in tubules promoted myofibroblast proliferation and immune cell infiltration. A) Schematic of hRPTEC and renal fibroblast crosstalk experiment. Serum-starved (1 d) confluent hRPTECs were treated with HB-EGF (25 ng/ml/d) in the presence or absence of erlotinib (1 h) and washed with PBS to remove any residual HB-EGF before addition of fresh 0.5% FBS/DMEM for 24 h. Conditioned medium was transferred to quiescent fibroblasts at a similar density for 2 d. B) Conditioned media from HB-EGF–treated proximal tubule cells promoted fibroblast proliferation as measured by MTT assay, which was abolished by the EGFR tyrosine kinase inhibitor, erlotinib. Graph (means ± sd) represents renal fibroblast proliferation as measured by an optical density at 570 nm for each experimental condition. C) Increased immunohistochemical staining of α-SMA expression in the interstitium of hHB-EGF^Tg/Tg fibrotic kidneys. Original magnification, ×40. D) Immunohistochemical staining of F4/80 and CD3 expression increased in hHB-EGF^Tg/Tg mice compared with wild-type mice. Original magnification, ×160. ***P < 0.001.

Genetic deficiency of EGFR tyrosine kinase activity attenuates renal fibrosis caused by persistent tubule-specific EGFR activation

To investigate whether targeting persistent EGFR activation in tubules could halt renal fibrosis progression, we used hHB-EGF^Tg/Tg mice in a Wa-2 background as proof of principle. Wa-2;hHB-EGF^Tg/Tg mice had decreased activation of ERK and SMAD2/3 signaling in the tubules (Fig. 7A, B), which is consistent with a reduction in EGFR activation. hHB-EGF^Tg/Tg mice that were deficient in EGFR kinase activity also had diminished collagen and ECM deposition as visualized by Masson’s trichrome and PicroSirius red staining compared with hHB-EGF^Tg/Tg littermates (Fig. 7B, D). Moreover, genetic reduction of EGFR tyrosine kinase activity in hHB-EGF^Tg/Tg mice hindered oxidative stress as measured by nitrotyrosine and profibrotic gene expression, such as CTGF, which was consistent with diminished α-SMA⁺ myofibroblast accumulation and infiltration of F4/80⁺ Mϕs and CD3⁺ T cells (Fig. 7C, D), highlighting the integral role of sustained tyrosine kinase activation in tubules as a mediator of fibrotic progression.

Figure 7.

Genetic deficiency of EGFR tyrosine kinase activity (Wa-2 mice) ameliorates renal fibrosis, immune cell infiltration, and profibrotic signaling in hHB-EGF^Tg/Tg mice. A) Western blot analysis indicated reduced phospho-ERK and phospho-SMAD2 in Wa-2;hHB-EGF^Tg/Tg mice compared with hHB-EGF^Tg/Tg mice (n = 4 in each group). α-Tubulin confirmed equal loading. B) Immunoreactivity of phospho-ERK and phospho-SMAD3 in kidney sections was reduced in Wa-2;hHB-EGF^Tg/Tg mice compared with hHB-EGF^Tg/Tg mice. ECM deposition as measured by PicroSirius Red and Masson’s trichrome staining was abrogated in Wa-2;hHB-EGF^Tg/Tg mice. C) Increased oxidative stress (nitrotyrosine) and profibrotic gene expression (CTGF), as well as interstitial cells, including α-SMA⁺ myofibroblasts, F4/80⁺ Mϕs, and CD3⁺ T cells observed in hHB-EGF^Tg/Tg kidneys were hindered in hHB-EGF^Tg/Tg mice on a Wa-2 background. D) Histograms represent the expression of Masson’s trichrome, F4/80, and CD3 immunostaining of tissue sections from at least 3 hHB-EGF^Tg/Tg and hHB-EGF^Tg/Tg;Wa-2 mice and an average of at least 5 fields per section. Original magnification: × 160 (B, C) ***P < 0.001.

Pharmacologic intervention of EGFR tyrosine kinase activation or MEK abrogated renal fibrosis initiated by persistent EGFR activation in tubules of hHB-EGF^Tg/Tg mice

To investigate whether targeting tubule-specific EGFR activation could prevent the development of fibrosis in hHB-EGF^Tg/Tg mice, we analyzed the effect of pharmacologic inhibitors that target EGFR tyrosine kinase activity or the downstream signaling pathway, MEK. Male C57BL/6 hHB-EGF^Tg/Tg mice received vehicle (water) or erlotinib (80 mg/kg/d) by daily gastric gavage for 10 wk (from age 4 to 14 wk) before the removal of the kidney for analysis. A reduction in the activation of ERK (phospho-ERK) confirmed the functionality of the inhibitor (Fig. 8A, C). Furthermore, erlotinib-treated hHB-EGF^Tg/Tg mice had reduced collagen deposition and tubulointerstitial fibrosis compared with vehicle-treated animals (Fig. 8D).

Figure 8.

Treatment with the EGFR tyrosine kinase inhibitor, erlotinib, or the MEK inhibitor, PD 0325901, attenuated renal fibrosis driven by tubule-specific EGFR activation. A) hHB-EGF^Tg/Tg mice were treated with vehicle (water) or erlotinib (80 mg/kg/d) for 10 wk (age 4–14 wk). Erlotinib-treated mice had reduced phospho-ERK as measured by immunoblotting compared with vehicle-treated animals. α-Tubulin expression served as loading control. B) Treatment with vehicle (DMSO) or PD 0325901 (50 mg/kg/d) for 10 wk in hHB-EGF^Tg/Tg mice beginning at age 4 wk. Phosphorylation of ERK was reduced in PD 0325901–treated hHB-EGF^Tg/Tg mice. α-Tubulin expression determined equal loading. C, D) Erlotinib-treated hHB-EGF^Tg/Tg mice had reduced immunostaining of phopho-ERK (C) and ECM deposition as visualized by Masson’s trichrome staining compared with control mice (D). Bar graph in panel D represents the quantification of Masson’s trichrome of at least 3 vehicle-treated or erlotinib-treated mice. E) Increased expression of phospho-SMAD3 and collagen I and fibrosis, as measured by Sirius red, as well as F4/80⁺ Mϕs and α-SMA⁺ myofibroblasts evident in vehicle-treated hHB-EGF^Tg/Tg kidneys were reduced upon treatment with PD 0325901. Original magnification, ×160. F) Histograms are representative of the expression levels of Sirius red, α-SMA, and F4/80 immunostaining of vehicle or PD 0325901 treatment in at least 3 hHB-EGF^Tg/Tg mice. ***P < 0.001.

To determine whether intervention of downstream signaling pathways was a viable target strategy, we employed an MEK inhibitor, PD 0325901. hHB-EGF^Tg/Tg mice received vehicle (DMSO) or PD 0325901 (50 mg/kg/day) by daily gastric gavage for 10 wk (from age 4 to 14 wk) before kidney isolation. MEK inhibitor-treated hHB-EGF^Tg/Tg mice had reduced ERK phosphorylation (Fig. 8B) and SMAD3 activation (Fig. 8E) compared with mice that were treated with vehicle. Blockade of MEK in hHB-EGF^Tg/Tg mice abrogated collagen and ECM deposition, correlating with reduced density of α-SMA⁺ myofibroblasts and F4/80⁺ Mϕs and subsequent renal fibrosis (Fig. 8E, F). Collectively, pharmacologic inhibition of EGFR/MEK in tubules halts fibrotic disease.

DISCUSSION

Major findings in this study include the following: 1) sustained EGFR activation in the renal proximal tubule is sufficient to initiate progressive renal fibrosis; 2) persistent EGFR activation in the tubule induces epithelial dedifferentiation, which contributes to tubulointerstitial fibrosis; and 3) hHB-EGF^Tg/Tg mice provide a novel kidney tubulointerstitial fibrotic model that is characterized by spontaneity, early onset, consistent progression, and a platform to investigate therapeutic intervention in halting disease progression.

Although TGF-β is recognized as a major driver of fibrosis (21), other signaling pathways that are responsible for initiating and maintaining chronic fibrosis are incompletely elucidated. Here, we implicate tubule-specific EGFR activation in the initiation, maintenance, and progression of renal fibrosis. Although hHB-EGF^Tg/Tg mice were bred onto a C57BL/6 background—a strain relatively resistant to renal injury—surprisingly, homozygous hHB-EGF expression, specifically, in the tubule epithelia of C57BL/6 mice was sufficient to induce fibrosis of the kidney. As early as age 3 wk, hHB-EGF^Tg/Tg mice had increased collagen and matrix deposition, which progressed in mice up to age 6 mo. This spontaneous transgenic model recapitulated known hallmarks of renal fibrosis, including tubular dedifferentiation and dilation, interstitial expansion, accumulation of collagen and ECM molecules, increased inflammation, and myofibroblast activation, with proliferation and differentiation providing a novel platform to investigate renal fibrosis progression.

Similar to other EGF family members, HB-EGF requires ectodomain shedding, the proteolytic process that cleaves the extracellular domain of membrane-anchored cytokines and growth factors that release a soluble, bioactive ligand. Studies have identified that membrane-anchored disintegrin and metalloprotease (ADAM) family and secreted MMP family of enzymes can cleave HB-EGF, which allows for the activation of EGFR. Of interest, pharmacologic or genetic inhibition of TACE/ADAM17 reduced renal tubulointerstitial fibrosis induced by Ang II infusion and UUO (13, 30). Consistent with the role of ADAM17 in ligand-dependent EGFR activation, systemic TWEAK administration in mice increase renal tubule EGFR activation via the up-regulation of ADAM17 and the subsequent processing and release of EGFR ligands, HB-EGF and TGF-α (19). Furthermore, pan-inhibition of MMPs (MMP1–3, 8, 9) blocked HB-EGF–dependent EGFR activation in renal epithelial cells (31). Ectodomain shedding of HB-EGF and other EGF family members likely plays a role in the transient and persistent activation of EGFR, which contributes to renal disease progression. In acute renal injury, ischemia/reperfusion in rats promoted a 12- and 8-fold increase of HB-EGF mRNA, which occurred at 90 min and 6 h, respectively (32). In a chronic setting, we observed a 6-fold increase of HB-EGF mRNA in hHB-EGF^Tg/Tg mice and previously reported a 3-fold increase in HB-EGF protein in diabetic enos^−/− db/db mice (33), which suggests that chronic expression as well as the expression level of the EGF family may contribute to the persistent EGFR activation that is sufficient to induce renal tubulointerstitial fibrosis.

Mice with endothelial-specific knockout of HB-EGF had a reduction in Ang-II–mediated endothelial dysfunction, renal fibrosis, and inflammation (34). Furthermore, HB-EGF depletion in endothelial cells decreased Ang-II–dependent proximal tubule HB-EGF expression and subsequent tubulointerstitial fibrosis. This study confirms that HB-EGF expression in proximal tubules is a causative factor in tubulointerstitial fibrosis. In fact, tissue sections from patients with various etiologies of CKD, including diabetic and focal segmental glomerulosclerosis, revealed an increase of HB-EGF in tubules (35), which is consistent with hHB-EGF^Tg/Tg mice. Furthermore, EGFR activation correlated with renal failure in human crescentic glomerulonephritis and noncrescentic nephropathies (35). Moreover, patients with CKD had increased HB-EGF, in addition to other EGFR ligands, such as amphiregulin and TGF-α (30).

The hHB-EGF^Tg/Tg model reveals how renal tubule cells can directly impact the fibrotic response. Chronic stress or repeated injury of epithelial cells promotes dedifferentiation, altered cell cycle profiles, and profibrotic factor secretion (9). Our data suggest that chronic EGFR activation is sufficient to promote tubular dysfunction, which mediates the dedifferentiation and cell cycle arrest of epithelial cells. Epithelial–mesenchymal transition (EMT) is a process in which epithelial cells lose their apical–basal polarity and acquire a migratory and invasive phenotype (25); however, the role of EMT in renal fibrosis is still debated. Transcriptional repressors of E-cadherin, Snail, and Slug were significantly increased in the tubule epithelium of hHB-EGF^Tg/Tg mice, which is consistent with an increase in the mesenchymal marker, vimentin, and a decrease in E-cadherin expression. Such findings suggest that sustained EGFR activation in the tubule allows epithelial cells to undergo partial EMT without complete loss of cell–cell adhesion or the ability to gain a migratory or invasive phenotype. In fact, recent studies demonstrate that UUO-induced renal fibrosis requires Snail to induce partial EMT in epithelial cells that is necessary for sustained inflammation and fibrogenesis (29). Collectively, current data suggest that the partial EMT phenotype contributes to dedifferentiated epithelial cells that can directly relay signals to promote fibrotic progression without undergoing complete EMT and transitioning into the myofibroblast population.

Normal cell cycle arrest checkpoints ensure accurate DNA replication and cell division to maintain the genomic integrity of the cell (28), whereas persistent epithelial dysfunction can lead to aberrant activation of repair processes, which alters their proliferative capacity and subsequent cell cycle profile (9, 10). Phosphorylation of Rb is necessary for cell cycle advancement from the G₀/G₁ to S phase (28). Tubule-specific EGFR activation led to decreased phospho-Rb and increased activation of p53, which suggests incomplete passage through the G₁ checkpoint. This is consistent with findings that the upstream kinase of p53, ATM, promotes epithelial proliferative arrest and subsequent profibrotic gene expression (36). Furthermore, epithelial-specific phosphatase and tensin homologue deleted on chromosome 10 (PTEN)-loss, which increased p53 phosphorylation, induced epithelial growth arrest in G₁, which led to increased profibrotic gene expression (37). Moreover, proximal tubules of hHB-EGF^Tg/Tg mice had increased phospho-histone H3^Ser10 expression compared with wild-type mice, which further suggests that epithelial cells had exited a quiescent state (G₀) and progressed through the cell cycle, likely arresting at G₁ (phospho-p53, phospho-Rb, p21) or G₂/M (phospho-p53, phospho-histone H3, p21) phases. Tubule growth arrest contributes to an increase in profibrotic secretory proteins and crosstalk with interstitial cells.

Dedifferentiated tubular epithelial cells have been demonstrated to release various chemokines and cytokines, including WNT ligands (38), sonic hedgehog ligands (39), CTGF, and TGF-β (27), all previously associated with fibrosis. Here, we demonstrate that paracrine factors that are released from human tubular epithelial cells after EGFR activation promote the proliferation of fibroblasts. Furthermore, sustained EGFR activation in the tubule epithelium of hHB-EGF^Tg/Tg mice had an increase in α-SMA⁺ interstitial cells, which were predominantly concentrated around the proximal tubules. Furthermore, α-SMA⁺ myofibroblasts had increased Ki-67 in hHB-EGF^Tg/Tg mice, which was indicative of a higher proliferative capacity and associated with an expansion of the interstitial space and disruption of the tissue architecture. These data highlight the significance of persistent EGFR activation in tubular injury and its role in driving fibrosis in neighboring cell types. EGF family members may also play a role in this paracrine signaling (17). The contribution of specific tubular-derived secretory factors will be the subject of our future studies.

Tubule-specific EGFR activation promoted an increase in inflammatory cells in hHB-EGF^Tg/Tg. Fibrosis progression is accompanied by increased infiltration of immune cells, including Mϕs and T cells. Previous studies have implicated Mϕs in renal fibrosis (40–43), and a study indicated that UUO-induced renal fibrosis was attenuated in Rag^−/− mice, which are devoid of B or T cells (44). Of interest, the reconstitution of purified CD4⁺ T cells was sufficient to increase collagen deposition and the subsequent interstitial expansion in the UUO model (44).

Organ fibrosis is a critical contributor to many chronic diseases, such as CKD, idiopathic pulmonary fibrosis, and scleroderma (1, 2). In the kidney, tubulointerstitial fibrosis is a prognostic indicator of a poor outcome and represents a common final pathway of CKD regardless of insult (2–4). Current models to study renal fibrosis include UUO, Ang II infusion, and 5/6 nephrectomy (40). UUO is a simple procedure with rapid onset of fibrosis; however, it is aggressive and difficult to intervene once the disease is established (40). In contrast, 5/6 nephrectomy leads to progressive glomerular and interstitial fibrosis similar to human disease, but the surgery is technically difficult and not all mouse strains are susceptible (40).

The rapid onset and progressive nature of renal fibrosis in the hHB-EGF^Tg/Tg model makes it an attractive target to test the efficacy of therapeutic agents. hHB-EGF^Tg/Tg mice had increased activation of EGFR and downstream signaling, such as ERK and AKT, and promoted the activation of the TGF-β/SMAD pathway in tubule epithelial cells. Previous studies demonstrate that pharmacologic inhibition and genetic deficiency of EGFR activation were sufficient to ameliorate Ang II infusion–induced renal fibrosis (14, 16). Furthermore, Ang II promoted the transactivation of EGFR, which activated the TGF-β pathway (14). EGFR seems to be a central factor in the propagation of fibrotic signaling and is causative in various renal pathologies (13–16); therefore, intervention with therapeutic approaches, including an EGFR tyrosine kinase or MEK inhibitor, attenuated fibrosis driven by tubule EGFR activation. In addition, genetic deficiency of EGFR tyrosine kinase activity in hHB-EGF^Tg/Tg mice prevented dedifferentiation and cell cycle arrest in epithelial cells. These data highlight the treatability of renal fibrosis in hHB-EGF^Tg/Tg mice.

In summary, hHB-EGF expression in proximal tubule cells provides a novel model with which to study the mechanisms that underlie renal fibrosis progression, highlights the significance of tubular EGFR activation in the initiation and progression of tubulointerstitial renal fibrosis, and can serve as a platform to identify new targets and treatment strategies to combat fibrogenesis.

Glossary

α-SMA

α-smooth muscle actin

ADAM

anchored disintegrin and metalloprotease

Ang II

angiotensin II

CKD

chronic kidney disease

CTGF

connective tissue growth factor

ECM

extracellular matrix

EGFR

epidermal growth factor receptor

EMT

epithelial–mesenchymal transition

FBS

fetal bovine serum

HB-EGF

heparin-binding EGF-like growth factor

hHB-EGF

human HB-EGF

hHB-EGF^+/Tg

heterozygous transgenic hHB-EGF

hHB-EGF^Tg/Tg

homozygous transgenic hHB-EGF

hRPTEC

human renal proximal tubule epithelial cell

LTL

lotus tetragonolobus lectin

Mϕ

macrophage

MMP

matrix metalloproteinase

MTT

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide

Rb

retinoblastoma

UUO

unilateral ureteral obstruction

Wa-2

Waved-2

AUTHOR CONTRIBUTIONS

J. M. Overstreet, R. C. Harris, and M.-Z. Zhang designed the research and analyzed the data; J. M. Overstreet, Y. Wang, X. Wang, A. Niu, B. Yao, and M.-Z. Zhang performed the research; J. M. Overstreet, R. C. Harris, and M.-Z. Zhang wrote the manuscript; and L. S. Gewin contributed the mouse renal fibroblasts.