HIVAN phenotype: consequence of epithelial mesenchymal transdifferentiation

Abstract

Human immunodeficiency virus (HIV)-1-associated nephropathy (HIVAN) is characterized by proliferation of glomerular and tubular epithelial cells. We studied the role of epithelial mesenchymal transdifferentiation (EMT) in the development of HIVAN phenotype. Renal cortical sections from six FVB/N (control) and six Tg26 (HIVAN) mice were immunolabeled for PCNA, α-smooth muscle actin (α-SMA), fibroblast-specific protein-1 (FSP1), CD3, and F4/80. Since periglomerular cells (PGCs) and peritubular cells (PTCs) did not show any labeling for CD3 and F4/80 but showed labeling for α-SMA or FSP1, it appears that these were myofibroblasts that migrated from either glomerular or tubular sites, respectively. Occurrence of EMT was also supported by diminished expression of E-cadherin by renal epithelial cells in Tg26 mice. Interestingly, Tg26 mice also showed enhanced renal tissue expression of ZEB2; henceforth, it appears that transcription of molecules required for maintenance of de novo renal epithelial cell phenotype was suppressed. To evaluate the role of ANG II, Tg26 mice in groups of three were administered either normal saline or telmisartan (an AT₁ receptor blocker) for 2 wk, followed by evaluation for renal cell EMT. Renal cortical section of Tg26 mice showed a sevenfold increase (P < 0.001) in parietal epithelial cell (PEC)-PGC and a threefold increase (P < 0.01) in tubular cell (TC)-PTC proliferation (PCNA-positive cells). Similarly, both PECs-PGCs and TCs-PTCs in Tg26 mice showed enhanced expression of α-SMA and FSP1. Both PECs and podocytes contributed to the glomerular proliferative phenotype, but the contribution of PECs was much greater. Telmisartan-receiving Tg26 mice (TRM) showed attenuated number of proliferating PECs-PGCs and TCs-PTCs compared with saline-receiving Tg26 mice (SRM). Similarly, TRM showed diminished expression of α-SMA and FSP1 by both PECs-PGCs and TCs-PTCs compared with SRM. We conclude that EMT contributes to the manifestation of the proliferative phenotype in HIVAN mice.

visceral epithelial cells (podocytes) have been considered to play an important role in the pathogenesis of the collapsing variant of focal segmental glomerular sclerosis (FSGS) (20, 27). The podocyte is a terminally differentiated and highly specialized cell that does not replicate. On that account, the contributory role of podocytes in renal lesions with the proliferative phenotype was debated in the past (9, 18). Since cells occupying the glomerular tuft in the collapsing variant of FSGS do not express podocyte markers such as synaptopodin, vascular endothelial growth factor, and Wilm's tumor antigen (WT-1), their source of origin is still far from clear (1, 2, 21, 22). On the other hand, the cells contained in the glomerular tuft express cytokeratin, a marker expressed by parietal epithelial cells and not by podocytes; therefore, other investigators have questioned the podocyte origin of these cells (11, 32).

Human immunodeficiency virus (HIV)-associated nephropathy (HIVAN) is characterized by the collapsing variant of FSGS (9). Barsoni et al. (3) showed that podocytes in HIVAN are dysregulated and dedifferentiated with loss of all podocyte-specific markers. Thus glomerular lesions in HIVAN are morphologically similar to the idiopathic variety of the collapsing variant of FSGS. However, patients with HIVAN also develop proliferative tubular lesions that manifest in the form of microcyst formation (9). Mice transgenic for HIV-1 genes develop proteinuria and renal failure at the age of 4 wk (10, 19). In this murine model of HIVAN (Tg26 mice), renal lesions are characterized by segmental collapse of the tuft with prominent extracapillary proliferation and frequent formation of adhesions (10, 19). The origin of proliferating cells in this murine model also has been considered to be of podocyte origin by some investigators (3). However, this issue is still debated by other investigators (12). Although there is agreement on glomerular proliferative phenotype, the type of cells contributing to this phenotype remains controversial. The present study has been designed not to pinpoint the lineage of the involved cells (contributing to the phenotype), but rather to focus on the involved mechanism for the induction of the proliferative phenotype.

Epithelial mesenchymal transdedifferentiation (EMT) is a process in which renal epithelial cells lose their epithelial phenotype and attain new characteristic features of mesenchymal cells (15). This process is fundamentally linked to generation of the myofibroblasts (matrix-producing fibroblasts) under altered conditions. The presence of EMT in a mouse model of anti-tubular basement membrane was first reported by Strutz et al. (33) using the fibroblast-specific protein (FSP1) as a marker. These investigators demonstrated that tubular epithelial cells were able to express FSP1, a cytoskeleton-associated calcium-binding protein that is normally expressed in fibroblasts but not in epithelia. Subsequently, occurrence of EMT was demonstrated in the unilateral ureteral obstruction model (35). In this model, transdifferentiated renal epithelial cells also lost E-cadherin, a marker of epithelial cells (35). Later on, Lan et al. (23) reported morphological and phenotypic occurrence of EMT in a renal ablation model (23). EMT has been reported to contribute to the progression of renal fibrosis in animal models of anti-glomerular basement membrane glomerulonephritis, diabetic nephropathy, and nephrotoxic serum nephritis (24, 28–30, 37). However, the role of EMT in HIVAN has not been evaluated. Since EMT often manifests in the form of proliferative phenotype, we asked whether renal cell EMT contributed to the development of the HIVAN phenotype. In the present study, we have evaluated the role of EMT in a murine model of HIVAN (Tg26). In addition, we examined the contribution of glomerular and tubular epithelial cells in the development of the HIVAN phenotype.

MATERIALS AND METHODS

HIV Transgenic Mice

We used age- and sex-matched FVB/N (control) and Tg26 mice (on FVBN background). Breeding pairs of FVB/N were obtained from Jackson Laboratories (Bar Harbor, ME). Breeding pairs to develop Tg26 colonies were kindly gifted by Prof. Paul E. Klotman (Department of Medicine, Mount Sinai Medical Center, New York, NY). The Tg26 transgenic animal has the proviral transgene pNL4-3: d1443, which encodes all the HIV-1 genes except gag and pol, and therefore the mice are noninfectious. Mice were housed in groups of four in a laminar flow facility (Small Animal Facility, Long Island Jewish Medical Center, New Hyde Park, NY). We are maintaining colonies of these animals in our animal facility. Since Tg26 mice develop proteinuria and renal lesions at 4 wk, we have used six (4 wk old) male Tg26 and FVB/N mice in the present study. For genotyping of these animals, tail tips were clipped, DNA was isolated, and PCR studies were carried out using the following primers for Tg26: forward primer HIV-F, 5′-ACATGAGCAGTCAGTTCTGCCGCAGAC; and reverse primer HIV-R, 3′-CAAGGACTCTGATGCGCAGGTGTG. The Ethics Review Committee for Animal Experimentation of Long Island Jewish Medical Center approved the experimental protocol.

ANG II Blockade Studies

Six mice in groups of three (3 wk old) were anesthetized by isoflurane and oxygen mixture inhalation anesthesia. The Alzet mini pumps (model 1007D DURECT) containing either saline alone or telmisartan (300 μg · kg⁻¹ · day⁻¹) dissolved in saline were implanted subcutaneously in the interscapular region of mice. Another group of three FVB/N mice (receiving saline) were used as a control for saline-receiving Tg26 mice. After 2 wk of infusion, animals were euthanized, kidneys were isolated, and renal tissues were prepared for paraffin sections.

Immunohistochemical Staining

The immunohistochemistry protocol has been described previously (20). Briefly, the sections were deparaffinized, and antigen retrieval was accomplished by microwave heating for 10 min at maximum output in 10 mM citrate buffer (pH 6.0). The endogenous peroxidase activity was blocked with 0.3% hydrogen peroxide in methanol for 30 min at room temperature. Sections were washed in phosphate-buffered saline (PBS) three times and incubated in blocking serum solution according to the primary antibody for 1 h at room temperature. The primary antibody was applied in different dilutions: PCNA (mouse monoclonal, dilution 1:500; Santa Cruz Biotechnology, Santa Cruz, CA), α-smooth muscle actin (α-SMA; monoclonal mouse anti-human, dilution 1:200; Dako, Glostrup, Denmark), FSP1 (anti S-100A4, rabbit polyclonal, dilution 1:150; Sigma, St. Louis, MO), E-cadherin (mouse monoclonal, dilution 1:400; Santa Cruz Biotechnology), ZEB2 (rabbit polyclonal, dilution 1:200; Abcam, Cambridge, MA), CD3 (mouse monoclonal, dilution 1:400, catalog no. 17A2; Biolegend, San Diego, CA), or F4/80 (rat monoclonal, dilution 1:200; Abcam) and then incubated overnight at 4°C in a humidifying chamber. Each of the sections was washed three times with PBS and incubated in the appropriate secondary antibody at 1:250 dilutions at room temperature for 1 h. After being washed with PBS three times, sections were incubated in ABC reagent (Vector Laboratories, Burlingame, CA) for 30 min. Sections were washed three times in PBS and placed in the Vector NovaRED substrate kit SK-4800 (Vector Laboratories), followed by counterstaining with methyl green. The sections were then dehydrated and mounted with a xylene-free mounting medium (Permount; Fisher Scientific, Fair Lawn, NJ). In all the batches of immunostaining, appropriate positive and negative controls were used. All the immune-stained slides were coded and blindly studied using a semiquantitative grading score. The number of positively stained cells was counted per tubule as well as per glomerulus in 10 random fields in each labeled renal cortical section (under ×20 lens). Positive staining was characterized by distinctly greater nuclear/cytoplasmic staining from the background staining of that cortical section; moreover, positive cellular staining had to be accompanied by distinctly greater nuclear/cytoplasmic staining of cells compared with cortical sections labeled with only secondary antibody (without primary antibody).

Protein Extraction and Western Blotting

Renal cortical tissue were mixed with lysis buffer [1× PBS, pH 7.4, 0.1% SDS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 1.0 mM sodium orthovanadate, 10 μ1 of protease inhibitor cocktail (100×; Calbiochem) per 1 ml of buffer, and 100 μg/ml PMSF], homogenized with a Dounce homogenizer, and then incubated on ice for 30 min. The samples were subjected to centrifugation at 15,000 g for 20 min at 4°C. The collected supernatant was evaluated for protein concentration as determined using a bicinchoninic acid kit (BCA; Pierce, Rockford, IL). The proteins (20–40 μg/lane) were separated by 10 or 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto a nitrocellulose membrane using a Bio-Rad Western blotting apparatus. After transfer, blots were stained with Ponceau S (Sigma) to check for complete protein transfer and equal loading. The blots were blocked with 0.5% BSA and 0.1% Tween 20 in 1× PBS for 60 min at room temperature and then incubated with the PCNA, α-SMA, FSP1, or E-cadherin primary antibody overnight at 4°C. A horseradish peroxidase-conjugated appropriate secondary antibody was applied for 1 h at room temperature. The blots were then developed using a chemiluminescence detection kit (ECL; Amersham, Arlington Heights, IL) and exposed to Kodak X-OMAT AR film. To ensure equal loading of proteins, the blots were stripped and reprobed for actin.

Reverse Transcription PCR Analysis

Four-week-old FVB/N (control, n = 3) and Tg26 mice (n = 3) were studied for renal cell expression of PCNA, α-SMA, and ZEB2. Renal cortical tissue of Tg26 and FVB/N mice were harvested, and RNA was extracted using Trizol (Invitrogen). For cDNA synthesis, 2 μg of the total RNA were preincubated with 2 nmol of random hexamer (Invitrogen) at 65°C for 5 min. Subsequently, 8 μl of the reverse transcription reaction (RT) mixture containing cloned avian myeloblastosis virus (AMV) reverse transcriptase, 0.5 mmol each of the mixed nucleotides, 0.01 mol dithiothreitol, and 1,000 U/ml RNasin (Invitrogen) were incubated at 42°C for 50 min. For a negative control, a reaction mixture without RNA or reverse transcription was used. Samples were subsequently incubated at 85°C for 5 min to inactivate the RT.

Quantitative PCR was carried out in an ABI Prism 7900HT sequence detection system using the following primer sequences: PCNA, forward GGGTTGGTAGTTGTCGCTGT, reverse AGCACCTTCTTCAGGATGGA; SMA, forward CTGACAGAGGCAACCACTGAA, reverse CATCTCCAGAGTCCAGCACA; and ZEB2, forward CGCTTGACATCACTGAAGGA, reverse CTTGCCACACTCTGTGCATT.

SYBR green was used as the detector and ROX as the passive reference gene. Results are means ± SD and represent three animals as described. The data were analyzed using the comparative threshold cycle (C_T) method (ΔΔC_T method). Differences in C_T are used to quantify relative amount of PCR target contained within each well. The data are expressed as relative mRNA expression in reference to control, normalized to the quantity of RNA input by measurements performed on an endogenous reference gene, GAPDH. A representative gel electrophoresis was also carried out with α-tubulin as the housekeeping gene. After agarose gel electrophoresis, a Polaroid camera system was used to capture images. ImageJ software (Research Services Branch, National Institutes of Health, Bethesda, MD) was used to carry out densitometric analysis of RT-PCR gels.

In Vitro Studies

Conditionally immortalized mouse podocytes.

Conditionally immortalized mouse podocytes were a gift from Dr. Mohammad Husain (Renal Molecular laboratory, Long Island Jewish Medical Center). The cells were maintained in RPMI supplemented with 10% fetal bovine serum (FBS), 1× Pen Strep, and 2 mM l-glutamine (Life Technologies, Rockville, MD) at 33°C (permissive temperature) in the presence of 5% CO₂. To permit immortalized growth, the medium was supplemented with 10 U/ml murine recombinant interferon-γ (Life Technologies) to induce the H-2K^b promoter driving synthesis of the temperature-sensitive (tsA58) SV40 T antigen (TAg). At 37°C (nonpermissive temperature), the TAg is inactivated and podocytes show differentiated morphology with multiple foot processes. All experiments were performed in differentiated podocytes.

Mouse tubular cells.

Mouse tubular cells were a gift from Dr. Poornima Upadhya (Long Island Jewish Medical Center). Mouse tubular cells were characterized by their expression for cytokeratin-18, cytokeratin-19, and vimentin.

Production of Pseudotyped Retroviral Supernatant

Replication defective viral supernatants were prepared as described previously (16, 17). The parental construct (pNL4-3: ΔG/P-GFP), was used to produce vesicular stomatitis virus G (VSV.G) pseudotyped viruses to provide pleiotropism and high-titer virus stocks. Infectious viral supernatants were produced by transient transfection of 293T cells using Effectene (Qiagen) according to the manufacturer's instructions. The HIV-1 gag/pol and VSV.G envelope genes were provided in trans using pCMV R8.91 and pMD.G plasmids, respectively (gifts of Dr. Didier Trono, Salk Institute, La Jolla, CA). As a negative control, virus was also produced from pHR-CMV-IRES2-GFP-ΔB plasmid, which contains HIV-1 long terminal repeats and green fluorescence protein (GFP). The viral stocks were titrated by infecting 293T cells with 10-fold serial dilution as reported earlier (16, 17). The reciprocal of the lowest dilution showing expression of GFP was defined as GFP-expressing units (GEU) per milliliter. Viral stocks ranging from 10⁵ to 10⁶ GEU/ml were obtained. Some low-titer viral stocks were further concentrated by ultracentrifugation.

Podocyte and Tubular Cell Transduction

In case of podocytes, cells were plated in 24-well plates at a density of 10,000 cells per well in 1.0 ml of growth medium at the permissive temperature. The cells were first allowed to grow at permissive temperature on a type 1 collagen-coated surface to 90% confluence and were then transferred to 37°C for 2 wk to inactivate temperature-sensitive T antigen. The cells were transduced using a multiplicity of infection of 0.5 GEU for 2 h. Mouse tubular cells were also transduced with either empty vector or NL4-3.

Western Blotting Studies

Control, empty vector, and NL4-3-transduced podocytes and tubular cells were incubated in serum-free media for 48 h. Subsequently, Western blots were prepared as described above and probed for PCNA, α-SMA, and actin.

Statistical Analysis

For comparison of mean values between two groups, the unpaired t-test was used. To compare values between multiple groups, analysis of variance (ANOVA) was applied and a Bonferroni multiple range test was used to calculate a P value. Statistical significance was defined as P < 0.05.

RESULTS

Renal Cells Show Proliferative Phenotype in Tg26 Mice

To determine the proliferation profile of renal cells, we immunolabeled renal cortical sections from six FVB/N and six Tg26 mice for PCNA. Representative microphotographs of cortical sections immunolabeled for PCNA from FVB/N and Tg26 mice are shown in Fig. 1, A and B, respectively. Since PCNA-positive periglomerular (PGCs) and peritubular cells (PTCs), especially in Tg26 mice, were in contiguity with the parietal epithelial (PECs) and tubular cells (TCs), we have counted PGCs and PTCs together with PECs and TCs, respectively. Cumulative data for mean number of proliferating (PCNA positive) PECs-PGCs and TCs-PTCs are shown in Fig. 1, C and D, respectively. As shown in Fig. 1C, Tg26 mice showed a sevenfold increase (P < 0.001) in the numbers of proliferating PECs-PGCs per glomerulus (n = 100) compared with FVB/N mice. Similarly, Tg26 mice also showed a threefold increase (P < 0.01) in the numbers of proliferating TCs-PTCs per tubule (n = 200) compared with FVB/N mice (Fig. 1D). In addition, renal tissue from Tg26 mice showed enhanced mRNA expression for PCNA in RT-PCR (Fig. 2A) and real-time PCR studies (Fig. 2B). Western blots prepared from renal cortical tissue of control and Tg26 mice were also probed for PCNA. As shown in Fig. 3, renal tissue from Tg26 mice showed enhanced expression of PCNA compared with FVB/N mice.

Fig. 1.

Proliferative phenotype of renal cells in Tg26 mice. Renal cortical sections from FVB/N and Tg26 mice (n = 6) were immunolabeled for PCNA. A: representative microphotograph of a cortical section of a FVB/N mouse. PCNA-positive (+ve) cells show dark brown nuclei. B: representative microphotograph of a cortical section of a Tg26 mouse. PCNA+ve cells show dark brown nuclei. Magnification, ×200. C: the number of PCNA+ve cells in 10 random fields was counted, and the mean number of PCNA+ve cells was calculated per glomerulus. Results are means ± SD from 6 mice. D: the number of PCNA+ve cells in 10 random fields was counted, and the mean number of PCNA+ve cells was calculated per tubule. Results are means ± SD from 6 mice.

Fig. 2.

Renal tissue mRNA expression of α-smooth muscle actin (α-SMA), PCNA, and ZEB2 in FVB/N and Tg26 mice. Renal cortical tissue RNA was isolated from 3 control and 3 age- and sex-matched Tg26 mice. RNAs were evaluated for α-SMA, PCNA, ZEB2, and α-tubulin using RT-PCR and real-time PCR. A: representative gels from a control and a Tg26 mouse. Top gels show renal tissue expression of PCNA, α-SMA, and ZEB2 by a control and a Tg26 mouse; bottom gels show renal tissue expression of tubulin under the same conditions. B: real-time PCR data are means ± SD showing relative PCNA, α-SMA, and ZEB2 mRNA expression from 3 control (FVB/N) and 3 Tg26 mice.

Fig. 3.

Western blotting studies. Proteins were extracted from 4 renal cortical tissues of FVB/N (C, C2, C3, and C4) and Tg26 (Tg1, Tg2, Tg3, and Tg4) mice. Western blots were prepared and probed for ZEB2, E-cadherin, PCNA, α-SMA, and actin.

Enhanced Renal Cell Expression of α-SMA in Tg26 Mice

To determine the transformation of epithelial cells to myofibroblasts, we immunolabeled renal cortical sections from FVB/N and Tg26 mice for α-SMA, a marker of myofibroblasts. Representative microphotographs of cortical sections immunolabeled for α-SMA from FVB/N and Tg26 mice are shown in Fig. 4, A and B, respectively. Cumulative data of PECs-PGCs and TCs-PTCs expressing α-SMA are shown in Fig. 4, C and D, respectively. As shown in Fig. 4C, Tg26 mice showed an eightfold increase (P < 0.006) in the numbers of α-SMA-expressing PECs-PGCs per glomerulus compared with FVB/N mice, whereas Tg26 mice showed a fivefold increase (P < 0.002) in the numbers of α-SMA-expressing TCs-PTCs per tubule compared with FVB/N mice (Fig. 4D). In addition, renal tissue from Tg26 mice showed enhanced mRNA expression for α-SMA in RT-PCR (Fig. 2A) and real-time PCR studies (Fig. 2B). Similarly, in Western blotting studies, renal tissue from Tg26 mice showed enhanced expression of α-SMA compared with FVB/N mice (Fig. 3).

Fig. 4.

Enhanced renal cell expression of α-SMA in Tg26 mice. Renal cortical sections from 6 FVB/N and 6 Tg26 mice were immunolabeled for α-SMA. A: representative microphotograph of a cortical section of a FVB/N mouse. α-SMA staining is indicated by dark brown color. A small artery shows staining for α-SMA. Magnification, ×200. B: representative microphotograph of a cortical section of a Tg26 mouse. α-SMA+ve cells show dark cytoplasm in glomerular (parietal epithelial cell, PEC), periglomerular (PGC), tubular (TC), and peritubular cells (PTC). Magnification, ×200. C: the number of α-SMA+ve cells in 10 random fields was counted, and mean number of α-SMA+ve cells was calculated per glomerulus. Results are means ± SD from 6 mice. D: the number of α-SMA+ve cells in 10 random fields was counted, and mean number of α-SMA+ve cells was calculated per tubule. Results are means ± SD from 6 mice.

Enhanced Renal Cell Expression of FSP1 in Tg26 Mice

To determine the transformation of renal epithelial cells to fibroblasts, we immunolabeled renal cortical sections from FVB/N and Tg26 mice for FSP1, a marker of fibroblasts. However, in the tumor microenvironment, FSP1 also has been demonstrated to be expressed by macrophages and activated lymphocytes (7). Representative microphotographs of cortical sections immunolabeled for FSP1 from FVB/N and Tg26 mice are shown in Fig. 5. Cumulative data of PECs-PGCs and TCs-PTCs expressing FSP1 are shown in Fig. 5, C and D, respectively. As shown in Fig. 5C, Tg26 mice showed a fourfold greater (P < 0.02) number of PECs-PGCs per glomerulus expressing FSP1 compared with FVB/N mice, whereas Tg26 mice showed a threefold greater (P < 0.003) number of FSP1-positive TCs-PTCs/tubule compared with FVB/N mice (Fig. 5D).

Fig. 5.

Enhanced renal cell expression of fibroblast-specific protein-1 (FSP1). Renal cortical sections from 6 FVB/N and 6 Tg26 mice were immunolabeled for FSP1. A: representative microphotograph of a cortical section of a FVB/N mouse. Magnification, ×150. B: representative microphotograph of a cortical section of a Tg26 mouse. FSP1+ve cells are indicated by dark brown cytoplasmic staining, seen in glomerular, periglomerular, tubular, and peritubular cells. Magnification, ×150. C: the glomerulus shown in B was further magnified to visualize FSP1 staining. Magnification, ×1,000. D: a glomerulus from another field of the renal cortical section of a Tg26 mouse visualized under an oil-immersion lens. Both podocytes (open arrows) and parietal epithelial cells (filled arrows) show positive staining for FSP1. Magnification, ×1,000. E: the number of FSP1+ve cells in 10 random fields was counted, and mean number of FSP1+ve cells was calculated per glomerulus. Results are means ± SD from 6 mice. pF: the number of FSP1+ve cells in 10 random fields was counted, and mean number of FSP1+ve cells were calculated per tubule. Results are means ± SD from 6 mice.

Renal Cells Show Diminished Expression of E-Cadherin in Tg26 Mice

Diminished or loss of expression of E-cadherin by epithelial cells has been suggested to indicate an alteration in epithelial cell phenotype. To determine an alteration in E-cadherin expression in renal cells of Tg26 mice, we labeled renal cortical sections of FVB/N and Tg26 mice (n = 3) for E-cadherin. In addition, proteins were extracted from renal cortical tissues of FVB/N and Tg26 mice. Western blotting was performed, and cells were probed for E-cadherin. Both glomerular and epithelial cells of Tg26 mice showed attenuated expression of E-cadherin compared with FVB/N mice (Fig. 6A). Cumulative data of PECs-PGCs and TCs-PTCs expressing E-cadherin are shown in Fig. 6, C and D, respectively. As shown in Fig. 6C, Tg26 mice showed a twofold decrease (P < 0.05) in the numbers of E-cadherin-expressing PECs-PGCs per glomerulus compared with FVB/N mice; similarly, Tg26 mice showed a twofold decrease (P < 0.01) in the numbers of E-cadherin expressing TCs-PTCs per tubule compared with FVB/N mice (Fig. 6D). Western blotting studies also showed that renal tissue of FVB/N mice had higher expression of E-cadherin compared with Tg26 mice (Fig. 3).

Fig. 6.

Diminished renal cell expression of E-cadherin in Tg26 mice. A: representative microphotograph of a cortical section of a FVB/N mouse showing moderate E-cadherin expression by glomerular and tubular epithelial cells. Magnification, ×150. B: representative microphotograph of a cortical section of a Tg26 mouse showing a diminished labeling for E-cadherin by both glomerular and tubular cells. Magnification, ×150. C: the number of E-cadherin+ve cells in 10 random fields was counted, and mean number of E-cadherin+ve cells was calculated per glomerulus. Results are means ± SD from 6 mice. D: the number of E-cadherin+ve cells in 10 random fields was counted, and mean number of E-cadherin+ve cells was calculated per tubule. Results are means ± SD from 6 mice.

To determine the involved mechanism for the diminished E-cadherin expression by renal epithelial cells in Tg26 mice, we examined renal cell expression of ZEB2, a transcriptional repressor of E-cadherin. Renal cortical sections of FVB/N and Tg26 mice were immunolabeled for ZEB2. As shown, in Fig. 7, both glomerular (P < 0.01) and tubular cells (P < 0.001) showed greater expression of ZEB2 in Tg26 mice. To further confirm this finding, RNA was isolated from renal cortical tissues of FVB/N and Tg26 mice, and the expression of ZEB2 was determined by RT-PCR and real-time PCR studies. Tg26 mice showed enhanced mRNA expression of ZEB2 (Fig. 2A). Real-time PCR studies further confirmed these findings (Fig. 2B). These findings indicate that diminished renal cell expression of E-cadherin might have been contributed by enhanced renal cell expression of ZEB2 in Tg26 mice.

Fig. 7.

Enhanced renal cell expression of ZEB2 by Tg26 mice. Renal cortical sections from 6 FVB/N and 6 Tg26 mice were immunolabeled for ZEB2. A: representative microphotograph of a cortical section of a FVB/N mouse. ZEB2+ve staining is indicated by dark brown staining (filled arrows). Magnification, ×200. B: representative microphotograph of a cortical section of a Tg26 mouse. ZEB2 staining is indicated by dark brown staining in both parietal (open arrows) and tubular epithelial cells (filled arrows). Magnification, ×200. C: the number of ZEB2+ve cells in 10 random fields was counted, and mean number of ZEB2+ve cells was calculated per glomerulus. Results are means ± SD from 6 mice. D: the number of ZEB2+ve cells in 10 random fields was counted, and mean number of ZEB2+ve cells was calculated per tubule. Results are means ± SD from 6 mice.

Evaluation of Relationship Between Renal Cell Expression of α-SMA and Proliferative Phenotype in Tg26 Mice

To evaluate a relationship between renal cell expression of α-SMA and PCNA, we labeled serial renal cortical sections for either α-SMA or PCNA. Cytoplasmic labeling for α-SMA only partially correlated with nuclear labeling for PCNA in both PECs-PGCs and TCs-PTCs. A higher number of PECs-PGCs showed expression of α-SMA compared with their expression of PCNA (in the same glomerulus and tubule). Representative microphotographs of serial renal cortical sections from a Tg26 mouse labeled for α-SMA and PCNA are shown in Fig. 8.

Fig. 8.

Correlation between renal cell α-SMA expression and the proliferative phenotype. Serial renal cortical sections of a Tg26 mouse were labeled for either α-SMA (A and B) or PCNA (C and D). A and C and B and D are duplicate serial sections. A and B: labeling for α-SMA in the glomerular and tubular cells is indicated by dark brown (A) and purple staining (B). Magnification, ×400. C and D: labeling for PCNA in the glomerular and tubular cells is indicated by darkly stained nuclei. Magnification, ×400. In A and C, not many cells show staining for both antigens; however, in B and D, there are bands of glomerular and periglomerular cells that appear to be positive for both or for one or the other antigen.

Evaluation of the Role of ANG II in EMT in Tg26 Mice

To determine the role of ANG II in the EMT of renal cells in Tg26 mice, mice in groups of three received either normal saline (saline-receiving mice, SRM) or telmisartan (telmisartan-receiving mice, TRM) for 2 wk. Another group of three FVB/N mice (receiving normal saline) were used as a control for saline-receiving Tg26 mice. Subsequently, kidneys were harvested and renal cortical tissues were immunolabeled for α-SMA, FSP1, and PCNA. As shown in Fig. 9A, Tg26 mice showed a threefold greater (P < 0.01) number of α-SMA-positive cells per tubule compared with FVB/N mice; similarly, in Tg26 mice, the number of α-SMA-positive cells per tubule was greater (P < 0.01) than in FVB/N mice (Fig. 10B). Tg26 mice showed a greater number of PECs expressing FSP1 (Fig. 9B) and PCNA (Fig. 9C) compared with FVB/N mice; moreover, tubules in Tg26 mice showed higher numbers of cells expressing PCNA, α-SMA, and FSP1 compared with FVB/N mice (Fig. 10, A–C). Interestingly, telmisartan not only diminished tubular cell expression of PCNA, α-SMA, and FSP1 (Fig. 10, A–C) but also enhanced tubular cell expression of E-cadherin (Fig. 10D). These findings suggest that the activation of AT₁ receptors contributes to the pathogenesis of HIVAN phenotype.

Fig. 9.

Role of ANG II in glomerular cell epithelial-mesenchymal transdifferentiation (EMT). Tg26 mice in groups of 3 received either saline (Tg26) or telmisartan (Tg26+Tel) for 2 wk. Another group of 3 FVB/N mice (receiving saline) were used as a control. Subsequently, kidneys were harvested and renal cortical sections were immunolabeled for α-SMA, FSP1, and PCNA. A: the number of α-SMA+ve PECs in 10 random fields was counted, and mean number of α-SMA+ve cells was calculated per glomerulus. Results are means ± SD from 3 mice. *P < 0.01 compared with all variables. B: the number of FSP1+ve PECs in 10 random fields was counted, and mean number of FSP1+ve cells was calculated per glomerulus. Results are means ± SD from 3 mice. *P < 0.05 vs. FVB/N. **P < 0.05 compared with Tg26 mice. C: the number of PCNA+ve PECs in 10 random fields was counted, and mean number of PCNA+ve cells was calculated per glomerulus. Results are means ± SD from 3 mice. *P < 0.01 vs. FVB/N. **P < 0.001 vs. Tg26 mice.

Fig. 10.

Role of ANG II in tubular cell EMT. Renal cortical sections from the mice described in Fig. 9 were immunolabeled for PCNA, α-SMA, FSP1, and E-cadherin. A: the number of PCNA+ve tubular cells in 10 random fields was counted, and mean number of PCNA+ve cells was calculated per tubule. Results are means ± SD from 3 mice. *P < 0.01 vs. FVB/N. **P < 0.01 compared with Tg26 mice. B: the number of α-SMA+ve tubular cells in 10 random fields was counted, and mean number of α-SMA+ve cells was calculated per tubule. Results are means ± SD from 3 mice. *P < 0.01 vs. FVBN. **P < 0.05 compared with Tg26 mice. C: the number of FSP1+ve tubular cells in 10 random fields was counted, and mean number of FSP1+ve cells was calculated per tubule. Results are means ± SD from 3 mice. *P < 0.01 vs. all other variables. D: the number of E-cadherin+ve cells in 10 random fields was counted, and mean number of E-cadherin+ve cells was calculated per tubule. Results are means ± SD from 3 mice. *P < 0.01 vs. FVB/N. **P < 0.05 compared with Tg26 mice.

Role of Immune Cell Influx in Enhanced Renal Cell Population

To determine the role of influx of macrophages and T cell in renal interstitium, we immunolabeled renal cortical sections of control and Tg26 mice with anti-CD-3 and anti-F4/80 antibodies. There was no difference in the number of renal interstitial CD3-positive and F4/80-positive cells in control and Tg26 mice (data not shown).

Effect of HIV-1 Transduction on Podocyte EMT

To determine the effect of HIV-1 infection on podocytes in vitro, mouse podocytes were transduced with either empty vector or NL4-3 (HIV-1) and then incubated in serum-free media for 96 h. Subsequently, Western blots were prepared and probed for PCNA, α-SMA, and actin. As shown in Fig. 11, NL4-3-transduced podocytes showed enhanced expression of PCNA and α-SMA compared with control and empty vector-transduced podocytes. These findings indicate that HIV-1 infection promotes podocyte EMT.

Fig. 11.

Human immunodeficiency virus (HIV)-1-transduced podocytes show enhanced expression of PCNA and α-SMA. Equal numbers of control podocytes, empty vector (EV)-transduced podocytes, and NL4-3 (HIV)-transduced podocytes were incubated in serum-free media for 96 h. At the end of the incubation period, cells were harvested and Western blots were prepared, followed by probing for PCNA, α-SMA, and actin. Top lane shows podocyte expression of PCNA under the above-described conditions; middle lane shows podocyte α-SMA expression under similar conditions; and bottom lane shows podocyte actin contents under the same conditions.

Effect of HIV-1 Transduction on Tubular Cell EMT

To study the effect of HIV-1 infection on tubular cells in vitro, mouse proximal tubular cells were transduced with either empty vector or NL4-3 and then incubated in serum-free media for 96 h. At the end of the incubation period, Western blots were prepared and probed for PCNA, α-SMA, and actin. As shown in Fig. 12, NL4-3-transduced tubular cells displayed enhanced expression of PCNA compared with control and empty vector-transduced tubular cells. Interestingly, only NL4-3-transduced tubular cell showed expression of α-SMA. These findings indicate that HIV-1 transduction promotes tubular cell EMT.

Fig. 12.

HIV-1-transduced tubular cells show expression of PCNA and α-SMA. Equal numbers of control tubular cells, EV-transduced tubular cells, and NL4-3-transduced tubular cells were incubated in serum-free media for 96 h. At the end of the incubation period, cells were harvested and Western blots were prepared, followed by probing for PCNA, α-SMA, and actin. Top lane shows tubular cell expression of PCNA under the above-described conditions; middle lane shows tubular cell α-SMA expression under similar conditions; and bottom lane shows tubular cell actin contents under the same conditions.

DISCUSSION

In the present study, Tg26 mice showed an enhanced number of proliferating glomerular, periglomerular, peritubular, and tubular cells compared with age- and sex-matched FVB/N mice. Similarly, enhanced expression of α-SMA and FSP1 was also observed in these cells. Although both PECs and podocytes contributed to proliferative glomerular phenotype in Tg26 mice, PECs apparently predominated. Consonant with diminished expression of E-cadherin by renal epithelia, Tg26 mice showed enhanced mRNA expression of ZEB2, raising the possibility that increased ZEB2 expression in these mice might have contributed to the attenuated expression of E-cadherin. TRM showed attenuated numbers of proliferating glomerular, periglomerular, peritubular, and tubular cells compared with SRM. Renal tubular cells in TRM also showed diminished expression of α-SMA, FSP1, and E-cadherin. These findings indicate a close relationship between EMT and proliferative glomerular and tubular phenotype in HIVAN mice; moreover, the activation of AT₁ receptors appears to contribute to the pathogenesis of both renal cell EMT and proliferative phenotype in HIVAN.

Pathogenesis of FSGS has been the intense focus of several recent publications (1, 2, 9, 11, 18, 20–22, 27, 32). In a pioneering study, Kriz et al. (20) proposed that podocyte loss was the key event in the development of FSGS (20). It was hypothesized that the inability of the adjacent podocytes to proliferate and fill the gap leads to denudation of the glomerular basement membrane with resultant adhesion formation to the Bowman's capsule. Evidence that loss of podocytes may play some role in the pathogenesis of FSGS in HIVAN is borne out by our recent demonstration that HIV-1 induces apoptosis in human podocytes in vitro (17). Nonetheless, loss of podocytes with resultant adhesions to Bowman's capsules is more likely to manifest in the development of conventional FSGS and not the proliferative phenotype as seen in HIVAN. It was thus interesting that Husain et al. (16) recently reported that HIV Nef protein had a potential to promote dedifferentiation and proliferation of mouse podocytes, suggesting the contribution of proliferating podocytes in the pathogenesis of HIVAN. It has therefore been proposed that conventional FSGS and the collapsing variant may have different pathogenesis (21, 22). Although the podocyte has been considered the primary proliferating cell in HIVAN (20, 27), there are data to show that the glomerular epithelia in idiopathic collapsing glomerulopathy with an indistinguishable proliferative glomerular phenotype do not express characteristic markers of podocytes (1). Thus the podocyte lineage of the proliferative glomerular phenotype in idiopathic collapsing glomerulopathy may be questionable. Moreover, other investigators have proposed the role of PECs in the development of the proliferative phenotype in the collapsing variant of FSGS on the basis of positive staining for cytokeratin (11, 32). Dejken et al. (12) also evaluated the role of PECs in HIVAN and demonstrated that the proliferating glomerular epithelial cells were positive for PAX2 (a marker of PECs) and negative for synaptopodin (a marker for podocytes).

In the present study, podocytes recognized by their supracapillary location also showed the proliferative phenotype and expression of EMT-specific markers; nevertheless, the proliferative phenotype was contributed predominantly by PECs. We did not, however, pursue the specific markers to differentiate between the two cell types, because this was not the primary aim of our study. It would also be pertinent at this juncture to speculate that FSGS lesions may develop differently in mice compared with humans. In mouse glomeruli, proximal tubular epithelial cells are present within the glomerulus, close to the urinary pole. Thus we cannot rule out the possibility that proximal tubular epithelial cells proliferate and influence the FSGS phenotype in mice.

Since PGCs did not show any labeling for CD3 (a marker of T cells) and F4/80 (a marker for macrophages), although they expressed both α-SMA and FSP1, it appears that these are myofibroblasts that have migrated from the pools of proliferating glomerular or tubular myofibroblasts. Since the present study was not designed to explore the lineage of proliferating cells, we were not able to delineate their exact contribution to HIVAN phenotype. Nevertheless both glomerular and tubular epithelial cells showed EMT and proliferation.

EMT has been proposed to unfold by the following progressive events: 1) loss of adhesion by epithelial cells, 2) expression of α-SMA and reorganization of the actin cytoskeleton, 3) breach in the epithelial basement membrane, and 4) increased motility and migration of transdifferentiated cells to the adjacent areas (36). Because both PECs and TCs are attached to the basement membrane, a breach in the basement membrane will allow passage of the transdifferentiated cells into the interstitium. Moreover, transdifferentiated cells are more motile and thus more likely to find their way out through the breached basement membrane. Since myofibroblasts express α-SMA, they have the ability to contract. The latter property will allow migration of myofibroblasts into the interstitium. The present study demonstrated the presence of significant numbers of myofibroblasts in the periglomerular and peritubular regions. There was a significant amount of intermingling of the myofibroblasts with the glomerular-periglomerular epithelial cells and tubular-peritubular epithelial cells. These findings indicate that transdifferentiated cells move from one compartment to the other.

ANG II has been demonstrated to contribute to the EMT process in several experimental models of renal injury (26, 34). In these models, upregulation of the renal renin-angiotensin system was associated with renal cell myofibroblast activation (26). Since the development of renal lesions in the present study was associated with EMT, we asked whether ANG II was contributing to renal cell EMT in Tg26 mice. Blockade of the AT₁ receptor in both human and murine HIVAN has been demonstrated to attenuate the progression of renal lesions (5, 6). We asked whether blockade of AT₁ receptors in Tg26 mice modulated renal lesions by influencing renal cell EMT. In the present study, telmisartan-receiving Tg26 mice not only showed reduced renal cell EMT but also showed an attenuation of the proliferative phenotype. These findings indicate that ANG II contributed to renal cell EMT as well as to proliferative phenotype in Tg26 mice.

Ectopic overexpression of Snail1 in epithelial cells has been demonstrated to induce EMT concomitantly with the downregulation of E-cadherin gene expression in epithelial cells (31). This role of Snail was further confirmed by the absence of EMT in the development of murine embryos null for Snail1 (31). Snail1 induces the expression of the Zeb1 transcriptional factor, a transcriptional repressor that also binds to E-cadherin promoter E-boxes (13). High levels of Zeb1 are detected in cells with mesenchymal phenotype and also observed after Snail-induced EMT (14). It has been suggested that Zeb1 might be working by extending the repression of E-cadherin initiated by Snail1 (14). Overexpression of Zeb2, another member of the Zeb family, also induces E-cadherin downregulation and EMT (8). Zeb2 transcripts are not generally induced after EMT and do not always correlate with the mesenchymal phenotype (14). However, recently it was reported that the synthesis of Zeb2 is upregulated after Snail-induced EMT (4). In the present study, overexpression of renal cell Zeb2 was associated with downregulation of renal cell E-cadherin and activation of EMT in Tg26 mice. Our findings further support the role of Zeb2 in the modulation of renal cell EMT in Tg26 mice.

We conclude that renal cell EMT plays a role in the development of the proliferative phenotype in HIVAN. Both glomerular and tubular cells contributed to the development of the proliferative HIVAN phenotype.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants R01 DK084910 and R01 DK083931.

DISCLOSURES

No conflicts of interest are declared by the authors.