Evidence for a Role of Transforming Growth Factor (TGF)-β1 in the Induction of Postglomerular Albuminuria in Diabetic Nephropathy: Amelioration by Soluble TGF-β Type II Receptor

IRPT cells (28) (passage 16–24) were grown in Dulbecco’s modified Eagle’s medium (DMEM) containing 4.5 g/l glucose (25 mmol/l), 10% fetal bovine serum (FBS), and l-glutamine (here on referred to as DMEM) at 37°C with 5% CO₂.

A reporter assay for the CAGA-luciferase response element was carried out as previously described (25). IRPT cells were transiently transfected with the (CAGA)₁₂MPL-Luc reporter construct (27) together with the pRL-TK vector (Promega, Madison, WI) in a 10:1 ratio (to control for transfection efficiency) using LipofectAMINE 2000 (Invitrogen, Carlsbad, CA). Cells were serum starved for 8 h and treated with varying concentrations of TGF-β1 (R&D Systems, Minneapolis, MN) (25–100 pmol/l) with and without varying concentrations of the sTβRII.Fc (75–1,200 pmol/l) for 16 h. Cells were lysed and luciferase activity determined with the Dual Reporter Assay (Promega). Relative light units were calculated as ratios of Firefly (reporter) and Renilla (transfection control) values.

Rat serum albumin (Fraction V) (RSA) (Sigma-Aldrich, St. Louis, MO) was conjugated to 5- (and 6-) carboxytetramethylrhodamine, succinimidyl ester (TAMRA; Molecular Probes, Carlsbad, CA) according to the manufacturer’s instructions. In brief, 10 mg/ml TAMRA reactive probe dissolved in anhydrous dimethylformamide (Sigma-Aldrich) was added to 20 mg/ml RSA dissolved in 0.15 mol/l sodium bicarbonate, pH 8.3, for 1 h at room temperature. To remove free dye, the solution was placed in Slide-A-Lyzer dialysis cassettes (10,000 MW cutoff) (Pierce, Rockford, IL) and dialyzed against multiple changes of 0.9% saline solution. Final protein concentration was determined using a bicinchoninic acid protein assay (BCA) (Pierce, Rockford, IL), and samples were stored at 4°C with 0.1 mol/l cacodylate. Before use, the RSA-TAMRA was processed through a PD-10 column (GE Healthcare Bio-Sciences, Uppsala, Sweden) to remove cacodylate and free dye.

IRPT cells were grown to confluence in six-well plates and treated for 24 h with 50 pmol/l TGF-β1 with and without 600 or 1,200 pmol/l sTβRII.Fc. Medium was removed and replaced with 37°C DMEM (without FBS) containing 10 μg/ml RSA-TAMRA (red) and incubated at 37°C with 5% CO₂ for 30 min. Cells were then placed on ice and washed 10 times in 4°C DMEM (without FBS). Cells were then lysed and resuspended with 150 μl dH₂0 per well and analyzed for fluorescence in duplicate 50-μl aliquots in a 96-well plate using a DTX880 Multimode Detector (Beckman Coulter, Fullerton, CA).

All animal experiments were approved by the Institutional Committee on Research Animal Care in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Male SD rats 200–250 g (Charles River, Wilmington, MA) were randomly assigned to control, diabetic, or diabetic + treatment groups. Before the induction of diabetes, all animals were weighed, blood glucose analyzed, and blood sample and 24-h urine sample collected (via metabolic cage). For STZ (ICN, Solon, OH) diabetes induction, animals were fasted overnight and diabetes induced with a single intravenous injection of STZ 50 mg · kg⁻¹ · body wt⁻¹ dissolved in citrate buffer pH 4.5 as previously described (3). The control group received citrate buffer alone. Diabetic rats were treated with 2 units of Humulin Ultralente insulin (Eli Lilly, Indianapolis, IN) every 2nd day to prevent ketoacidosis and to maintain blood glucose levels and were given free access to standard rat chow and water.

The sTβRII.Fc was produced by stably transfected HEK 293 cells (American Type Culture Collection number CRL-1573), cultured in DMEM supplemented with 10% ultra-low IgG FBS (Invitrogen) and 1 mg/ml G418 (Invitrogen), and purified as previously described (25). Purity of the protein was analyzed by 4–12% SDS-PAGE using precast NuPAGE Novex mini gels (Invitrogen) followed by silver staining (Bio-Rad, Hercules, CA). The amount of protein eluted was quantified by the BCA protein assay (Pierce, Rockford, IL). The sTβRII.Fc was administered to the treatment group via intraperitoneal injection (500 μg · kg⁻¹ · body wt⁻¹) three times per week, in 300 μl volume of 0.9% physiological saline. The nontreated diabetic group received 300 μl 0.9% physiological saline in place of the treatment.

Urinary albumin excretion was analyzed by Nephrat enzyme-linked immunosorbent assay (Exocell, Philadelphia, PA); briefly, 24-h urine samples stored at −20°C were thawed and centrifuged at 10,000 rpm for 15 min to remove debris, and samples were assayed according to the manufacturer guide. A1C was analyzed using erythrocytes (resuspended in physiological saline) by the Primus PDQ method using boronate affinity and high-performance liquid chromatography conducted in the Diabetes Unit at Massachusetts General Hospital.

At 12 weeks post-STZ diabetes induction, rats were anesthetized using pentobarbital (50 mg/kg), and the kidneys were flushed with PBS via the left ventricle to remove blood and then fixed by perfusion of periodate, lysine, and 4% paraformaldehyde solution as previously described (29). Kidneys were removed and immersion fixed in paraformaldehyde solution overnight at 4°C, washed extensively in PBS, pH 7.4, and cryoprotected in 30% sucrose in PBS, pH 7.4, for 4 h at room temperature. Kidneys were immersed in optimal cutting temperature compound (Tissue-Tek, Torrance, CA), rapidly frozen, sectioned at 5 μm, placed on slides, and stored at −20°C.

Sections were rehydrated in PBS for 15 min, treated with 1% SDS as previously described (30), and washed and blocked with 1% BSA in PBS solution. Primary antibody against collagen IV (M3F7 Developmental Studies Hybridoma Bank, generous gift from Dr. Iain Drummond, Renal Unit, Massachusetts General Hospital) and megalin monoclonal antibody (generous gift from Dr. Robert McCluskey, Department of Pathology, Massachusetts General Hospital), which reacts with the megalin ectodomain epitopes in the second cluster of ligand binding repeats (31), were diluted in DAKO Antibody diluent (DAKO, Carpinteria, CA) and applied to the tissue for 60 min at room temperature. Following incubation the tissue was washed in high-salt (2.7% NaCl) PBS (2 × 5 min) and PBS (1 × 5 min), and secondary antibody was also diluted in DAKO Antibody diluent, applied for 45 min, and washed in high-salt PBS (2 × 5 min) and PBS (1 × 5 min). Secondary antibodies used were donkey anti-mouse fluorescein isothiocyanate (FITC) IgG and goat anti-rabbit FITC IgG (Jackson ImmunoResearch Laboratories, West Grove, PA). Slides were cover slipped using Vectashield (Vector Labs, Burlingame, CA), sealed, and analyzed using a Nikon epifluorescence microscope equipped with a Hamamatsu Orca CCD camera.

For quantification, digital images were acquired using the Nikon microscope with a ×40 objective (×20 objective was used for interstitial collagen IV expression). All images from different slides were collected with the same microscope and camera acquisition settings in a blinded fashion. For collagen IV quantitation, in glomeruli and interstitial regions, images were acquired as above and normalized for background. Images were then segmented for fluorescence signal, and the percentage of area exhibiting staining was recorded. Percentage of area in the control groups was standardized to one, and the fold increase in collagen IV expression was calculated. For glomerular capillary tuft collagen IV, the same procedure was carried out except that a line was drawn around the glomerular capillary tuft to allow quantitation only in that area. Ten glomeruli were quantified at random in each of four kidneys from each group (40 glomeruli for each group), and 10 areas of the cortex (without glomeruli) were analyzed in each of four sections for each group (40 areas) for interstitial collagen IV analysis. For megalin quantitation, images were taken of S1 proximal tubules (initial segment of the proximal tubule immediately following the glomerulus) at random in a blinded fashion. A line was drawn around the apical membrane/brush border, and the area was analyzed for mean fluorescence intensity. Ten tubules were quantified at random in each of three kidneys from each group (30 S1 proximal tubules for each group). All computer analyses of collagen IV and megalin expression were performed on a Macintosh computer using IP Lab Spectrum acquisition and image analysis software (BD Biosciences Bioimaging, Rockville, MD).

At 12 weeks post–diabetes induction, the kidney cortex was isolated, homogenized in the presence of complete protease inhibitors (Roche Diagnostics, Indianapolis, IN), stirred at 4°C in 10 mmol/l MgCl₂ for 20 min, and centrifuged at 8,000 rpm for 15 min, and the supernatant was collected and centrifuged at 13,200 rpm for 30 min. The pellet was then resuspended in wash buffer (150 mmol/l KCl, 5 mmol/l Tris-HEPES, pH 7.4), centrifuged at 4,000 rpm for 15 min, and the pellet discarded. The supernatant was then centrifuged at 16,000 rpm for 30 min and the pellet resuspended in wash buffer and assayed for protein concentration using BCA protein analysis kit (Pierce, Rockford, IL) as previously described (32).

All data are presented as means ± SE, where n is the number of measurements. Statistical analysis was carried out using Student’s t test, where P < 0.05 was taken as statistically significant.

Initial studies were carried out in vitro to determine whether TGF-β1 could directly stimulate IRPT cells. IRPT cells that express the regular markers of proximal tubules, including megalin (28), were transfected with a CAGA-luciferase response element (27) together with the Renilla vector to control for transfection efficiency. When TGF-β1 binds to its receptor, Smad2 and Smad3 are phosphorylated and bind specific Smad binding elements and “CAGA” boxes such as those contained in the α2 (1) collagen and plasminogen activator inhibitor-1 gene promoters (33–36). Figure 1A demonstrates that IRPT cells were stimulated by TGF-β1 in a concentration-dependent manner (25–100 pmol/l) as demonstrated by activation of the CAGA-luciferase response element. The stimulation of the CAGA-luciferase response element by TGF-β1 appeared to be a concentration-dependent Smad-mediated event, as demonstrated by increased phospo-Smad2 production following treatment of IRPT cells with TGF-β1 (Fig. 1B). Activation of the CAGA-luciferase response element in transfected IRPT cells by 100 pmol/l TGF-β1 was reduced in a nonlinear fashion by increasing concentrations of the sTβRII.Fc (75–1,200 pmol/l) (Fig. 2). Taken together, these results demonstrate that IRPT cells respond to stimulation by TGF-β1 and that the sTβRII.Fc is able to sequester TGF-β1 and modulate its association with the cell.

To determine the effect of TGF-β1 on the binding and uptake of albumin by IRPT cells, confluent IRPT cell monolayers were pretreated for 24 h with 50 pmol/l TGF-β 1 with and without 600 or 1,200 pmol/l sTβRII.Fc. Results in Fig. 3 revealed that 50 pmol/l TGF-β1 reduced albumin association with cells by ∼30%. This TGF-β1 effect was reduced by 30% with the addition of 600 pmol/l sTβRII.Fc and by 60% with 1,200 pmol/l sTβRII.Fc. This demonstrates the ability of TGF-β1 to directly interact with IRPT cells and disrupt albumin binding/uptake as well as the ability of sTβRII.Fc to sequester TGF-β1, thus reducing its deleterious affects on albumin binding/uptake.

Following the induction of STZ diabetes, diabetic rats had significantly reduced body weight and increased water intake and urine excretion over the 12-week study (Table 1) as previously observed (3). A significant increase in urine volume without any significant increase in H₂O intake or change in body weight between treated and diabetic rats at 4 weeks (Table 1) was also observed possibly due to day-to-day variation in water intake and urine excretion. Treatment with the sTβRII.Fc increased body weight in the treated versus the diabetic group (P < 0.05) (although body weight was still less vs. controls, P < 0.001).

Albuminuria is recognized as an important marker of kidney dysfunction onset and progression (2). STZ-induced diabetes significantly increased urinary albumin excretion at 4 weeks (P < 0.01 for control vs. diabetic), 8 weeks (P < 0.001 for control vs. diabetic), and 12 weeks (P < 0.001 for control vs. diabetic) (Fig. 4). Treatment of diabetic rats with the sTβRII.Fc significantly reduced intact albumin excretion at 4 weeks (P < 0.05 for treated vs. diabetic), 8 weeks (P < 0.001 for treated vs. diabetic), and 12 weeks of sTβRII.Fc treatment (P < 0.01 for treated vs. diabetic) (Fig. 4), demonstrating the effectiveness of this agent in reducing albuminuria. Importantly, these changes occurred independently of reductions in blood glucose as analyzed by A1C (hemoglobin glycation) (Fig. 5), which were similarly high in both the diabetic and diabetic + treatment groups (P < 0.001 vs. age-matched controls) at both 6 and 12 weeks, suggesting that the effectiveness of sTβRII.Fc treatment was not influenced by disparities in blood glucose levels between diabetic and diabetic + treatment groups.

Collagen IV has previously been shown to be upregulated by TGF-β1 (16) and was analyzed in this study using immunocytochemistry to determine the effectiveness of the sTβRII.Fc in inhibiting TGF-β1–induced collagen IV production. Figure 6A reveals that collagen IV expression is increased in the glomerular capillary tuft and peritubular interstitium (see arrows) in diabetic rat kidney and decreased following sTβRII.Fc treatment of diabetic rats. Quantitation of collagen IV expression in the glomerulus and peritubular interstitium revealed that it was significantly increased in diabetes (P < 0.001 for control vs. diabetic) and significantly reduced with sTβRII.Fc treatment (P < 0.001 for treated vs. diabetic) (Fig. 6B and C, respectively). Interestingly, although collagen IV expression is significantly reduced in the peritubular interstitium by sTβRII.Fc treatment, it was still significantly higher than in the control (P < 0.01). This finding is in agreement with the trend observed for albuminuria (Fig. 4).

One of the mechanisms proposed by Gekle et al. (22) for the reduced uptake and binding of albumin in opossum kidney cells in the presence of TGF-β1 is an effect of TGF-β1 on the expression of the albumin binding receptor megalin. Analysis of the renal expression of megalin by both Western blot analysis and immunocytochemistry quantitation in the control, diabetic, and diabetic + treatment groups is shown in Fig. 7. Results reveal that megalin expression is reduced in diabetes, by both immunocytochemistry analysis (Fig. 7A and B) (P < 0.05) and Western blot analysis (Fig. 7C and D) (P < 0.05), a finding consistent with that observed by Tojo et al. (37). Interestingly, we now demonstrate that treatment of STZ-induced diabetic rats with the sTβRII.Fc increases megalin expression, suggesting that changes in megalin expression in diabetes is at least in part modulated by TGF-β1.

Albuminuria is considered one of the most important markers for the onset and progression of renal dysfunction in diabetes, and, more recently, it has also been implicated as a marker of cardiovascular and peripheral vascular disease. Although there is debate as to whether albuminuria itself is deleterious to kidney function, it is evident that there is an important link between the factor(s) causing albuminuria and those leading to diabetic nephropathy progression, suggesting that a better understanding of how albuminuria manifests itself will give important clues about the factors involved in kidney disease progression.

Recent studies have shown that the glomerular permeability to molecules of equivalent hemodynamic radius to albumin does not change in diabetes (3) and have implicated disruptions in postglomerular albumin processing via the degradation and retrieval pathways for albumin as an important contributor to albuminuria in diabetic nephropathy (3,19,38–40). Further, increased TGF-β in diabetes and hypertension is linked to disrupted albumin degradation through the inhibition of renal lysosomal activity (3,19,21,39). This may influence both the uptake and degradation of albumin, resulting in increased excretion of intact albumin or “albuminuria.” Recent studies by Gekle et al. (22) also demonstrated directly in vitro that TGF-β can indeed affect both the uptake and processing of albumin as demonstrated in opossum kidney cells (22).

Here we demonstrate in IRPT cells that TGF-β can directly stimulate a transfected CAGA-luciferase response element in a concentration-dependent manner and that the use of sTβRII.Fc can reduce this stimulation. Further, we show that TGF-β was able to reduce the binding/uptake of albumin by IRPT cells. These findings are in agreement with Gekle et al. (22) and suggest that TGF-β, which is upregulated in diabetes by high glucose (8), may play an important role in the manifestation of albuminuria in diabetic nephropathy independently of its prosclerotic effects on the glomerulus. These studies also implicate sTβRII.Fc as a potential treatment for diabetic nephropathy, as supported by our in vivo studies carried out in the well-characterized STZ model of type 1 diabetes. These studies demonstrate that sTβRII.Fc reduces albuminuria and leads to a decrease in the glomerular and peritubular interstitial deposition of collagen IV in vivo independently of changes in blood glucose levels. These results are consistent with previous studies using TGF-β–lowering treatments (9–12,15,41), all of which demonstrated reduced diabetic nephropathy progression including reductions in fibrosis and proteinuria. Further, we demonstrate that diabetes-induced reductions in proximal tubule megalin expression can be partially prevented by treatment with sTβRII.Fc, suggesting that TGF-β1 plays a role in reducing the expression of megalin in diabetes.

Although it has now been demonstrated that TGF-β can directly alter the postglomerular processing of albumin, it is also possible that other cytokines either acting through the activation of TGF-β or via a TGF-β–independent manner may also disrupt albumin processing in diseases including diabetes and hypertension. These proposed pathways are outlined in Fig. 8. High glucose levels in diabetes seem to be the initiating factor for the eventual cascade of hormone and growth factor upregulation and stimulation. Indeed, strict glycemic control in STZ-induced diabetic rats prevents the manifestation of diabetic nephropathy (42). However, strict management of blood glucose levels in humans is not always achievable due to increased risk of hypoglycemia and noncompliance. Hence, there is a need for the development of other agents to be used in conjunction with insulin treatment to help prevent diabetes-related complications such as diabetic nephropathy. Figure 8 indicates how TGF-β production and its subsequent effect on albumin uptake by proximal tubule cells may be influenced by glucose (8), glycated albumin (43), AngII (44), and changes in TGF-β–regulating proteins such as decreased BMP7 (14) and increased gremlin (45), a protein that decreases BMP7 production (46). High glucose may also influence albumin uptake by proximal tubule cells, which may or may not act via a TGF-β–related mechanism (47).

Renin-angiotensin system (RAS)–affecting drugs, such as ACE inhibitors and angiotensin type 2 receptor blockers used in the treatment of diabetic nephropathy, may affect albuminuria via a pathway converging with TGF-β (Fig. 8). High blood glucose has been shown to increase AngII production (48), and this together with the fact that AngII can increase TGF-β1 and vice versa (44) suggests that RAS-affecting drugs may also modulate TGF-β production. This presents an important linking factor for the renoprotective effects of RAS-affecting drugs as well as the potential renoprotective effects of TGF-β blocking agents in the treatment of diabetes-induced albuminuria.

In conclusion, these studies broaden our knowledge on the direct role that TGF-β might play in the induction of postglomerular albuminuria in diabetes. We demonstrate the effectiveness of the sTβRII.Fc in blocking TGF-β–induced disruptions in albumin processing in proximal tubule cells in vitro. Furthermore, in vivo studies in the STZ model of type 1 diabetes show that the sTβRII.Fc is nontoxic and effective in ameliorating in part the progression of diabetic nephropathy in vivo. Further studies are now needed to dissect the effects of the converging TGF-β and RAS pathways on postglomerular albumin uptake, which is an important marker of diabetic nephropathy onset and progression.

L.M.R. is the recipient of a Postdoctoral Fellowship and an Advanced Postdoctoral Fellowship from the Juvenile Diabetes Research Foundation. The Microscopy Core facility of the MGH Program in Membrane Biology receives additional support from the Boston Area Diabetes and Endocrinology Research Center (DK57521) and the Center for the Study of Inflammatory Bowel Disease (DK43341).