Deletion of p66^Shc Longevity Gene Protects Against Experimental Diabetic Glomerulopathy by Preventing Diabetes-Induced Oxidative Stress

Adult (aged 3 months) p66^shc KO and coeval SV/129 wild-type mice were divided into the following groups: nondiabetic wild type, nondiabetic KO, diabetic wild type, and diabetic KO. Diabetes was induced by injection of 150 mg/kg body wt streptozotocin i.p. (Sigma, St Louis, MO) (24). The animals were housed and cared in accordance with the Principles of Laboratory Animal Care (National Institutes of Health publ. no. 85-23, revised 1985) and national laws and received water and food ad libitum. Four months after initiating the study, the animals were placed into metabolic cages to collect urine samples. The next day, body weights were recorded and then mice were anesthetized with ketamine (60 mg/kg Imalgene i.p.) and xylazine (7.5 mg/kg Rompum i.p.), a longitudinal incision of the abdominal wall was performed, a blood sample was obtained, and both kidneys were quickly removed.

Mesangial cells were isolated from 1-month-old KO and wild-type mice and characterized as previously described (25). Cells between the 3rd and the 10th passage were cultured in Dulbecco’s modified Eagle’s medium (Sigma) supplemented with 17% fetal bovine serum, 2 mmol/l l-glutamine, and antibiotics (Flow Laboratories, Irvine, Scotland, U.K.) at 37°C in a 95% air/5% CO₂–humidified atmosphere with normal (5.5 mmol/l) or high (30 mmol/l) glucose concentrations for various time periods with or without the inhibitors of the cytosolic enzyme NAD(P)H oxidase, apocynin (50 μmol/l), and diphenyliodonium (10 μmol/l), or the inhibitors of the mitochondrial respiratory chain, 2-thenoyltrifluoroacetone (10 μmol/l) and myxothiazol (3 μmol/l). All experiments had a control for osmolarity (24.5 mmol/l mannitol plus normal glucose) (25).

Blood glucose levels were measured weekly with the aid of an automated colorimetric instrument (Glucocard G meter; Menarini, Florence, Italy). HbA_1c (A1C) was assessed by boronate affinity gel chromatography using the Glyco-Test II 100 (Pierce, Rockford, IL) (24).

Serum and urine creatinine levels were measured by high-performance liquid chromatography, total proteinuria by the Bradford method using the Bradford dye binding protein assay kit (Pierce), and albuminuria by enzyme-linked immunosorbent assay (ELISA) using the Mouse Albumin ELISA Quantitation Kit (Bethyl, Montgomery, TX). Values of proteinuria and albuminuria were normalized by the urine creatinine concentration (24,26,27).

A sagittal section of the kidney was fixed in phosphate-buffered 4% formaldehyde solution and embedded in paraffin. Analysis of renal structure was performed by two pathologists blinded to the group assignment of the specimens on multiple 4-μm sections stained with periodic acid Schiff (PAS). Sections were evaluated for glomerular sclerosis and tubulointerstitial damage (TID) by a standard semiquantitative analysis, and results were expressed as glomerular sclerosis index and TID score, as previously reported (26,27). For morphometrical analysis (26,27), the areas of at least 60 glomerular tuft profiles per sample were measured, the harmonic mean of the profile area (mean glomerular area) was obtained and the mean glomerular volume estimated from it. Then, PAS-positive material in each glomerulus was quantified and expressed as percentage of the glomerular tuft area (fractional mesangial area). Finally, the mean mesangial area was calculated by multiplying the fractional mesangial area by the mean glomerular area and dividing by 100.

Glomerular and tubular cell death rates were assessed in paraffin-embedded sections by immunohistochemistry for active caspase-3 using a rabbit polyclonal IgG antibody (Anti-ACTIVE Caspase-3; Promega Italia, Milan, Italy) (26). Glomerular cell type was identified topographically by counterstaining sections with PAS to mark basement membranes. Mesangial cell apoptosis was assessed by the transferase-mediated dUTP nick-end labeling method using a fluorescein-based kit (Boheringer Mannheim, Milan, Italy). Cell viability was assessed by measuring 3-[4,5-dimethylthiazol-2-yl]-5-(3-carboxymethoxyphenyl)-2-(4-sulfphenyl)-2H-tetrazolium reduction using the CellTiter 96 AQ_ueous One Solution Cell Proliferation Assay (Promega, Madison, WI).

Total RNA was extracted from renal cortex and mesangial cells by the guanidine thiocyanate-phenol-chloroform method using TRIzol Reagent (Invitrogen Italia, San Giuliano Milanese, Italy). Transcripts for fibronectin, collagen IV α1 chain, receptor for AGEs (RAGE), and p66^Shc were quantified by competitive RT-PCR, as previously reported (24,26,27). Briefly, 1 μg total RNA was reverse transcribed using the Retroscript kit (Ambion, Austin, TX), and the resulting cDNA was amplified using the primer sequences and experimental conditions reported in supplemental Table 1 (online appendix available at http://diabetes.diabetesjournals.org). Competitive PCR was performed by using increasing amounts of mutants made by creating a deletion in the original PCR product. After electrophoresis of PCR products, the unknown cDNA/mutant ratio was quantified by scanning densitometry using ImageJ software and results expressed as the ratio of each target to β-actin mRNA level.

Kidney cortex fibronectin, collagen IV, and p66^Shc protein expression was assessed by Western blot analysis (24). Samples were separated by SDS-PAGE and transferred by electroblotting. The membranes were probed with a rabbit polyclonal antibody to human fibronectin (Sigma), mouse collagen IV α1 chain (Abcam, Cambridge, U.K.), or the CH2 region of p66^shc protein (18), followed by a peroxidase-conjugated goat anti-rabbit IgG (Dako), and then developed with enhanced chemiluminescence reagent (Amersham, Amersham, UK). Immunocomplexes were revealed by autoradiography and quantified by scanning densitometry. Results were normalized to the signal of β-actin, which was revealed using a goat polyclonal antibody to human β-actin (Santa Cruz). Renal ECM protein distribution was then verified by immunohistochemistry, as previously reported (27), utilizing the above-reported antibodies. RAGE expression was also assessed by immunohistochemistry using a rabbit polyclonal antibody to human RAGE (amino acids 42–59; Abcam). Sections were analyzed using the Optimas 6.5 image analysis system and results expressed as the percent of glomerular area that was positive for each protein. The amount of fibronectin released from confluent mesangial cells during a 24-h incubation in serum-free medium was measured by ELISA, as previously detailed (25). The levels of fibronectin in conditioned media were quantified using a rabbit polyclonal antibody against rat fibronectin (Calbiochem, San Diego, CA) and values normalized to the DNA content of monolayers, as assessed fluorimetrically in 0.5 N NaOH extracts after reaction with 0.6 μmol/l 4,6-diamidino-2-phenylindole (Sigma).

Plasma levels of isoprostane 8-epi-prostaglandin F_2α, an index of systemic oxidative stress (21), were determined by ELISA using a commercial kit (Cayman, Ann Arbor, MI).

The level of AGEs in serum and kidney cortex extracts was assessed by a competitive ELISA technique, using a mouse monoclonal antibody raised against AGE-modified BSA (24,26,27). Renal content of the glycoxidation product Nε-carboxymethyllysine (CML) and the lipoxidation product 4-hydroxy-2-nonenal (HNE) was assessed in paraffin-embedded sections by immunohistochemistry using a biotinylated mouse monoclonal antibody against CML (Wako, Neuss, Germany) or a rabbit antiserum against HNE (Alpha Diagnostic International, San Antonio, TX) (26,27).

ROS levels were assessed by evaluating formation of the intracellular trapped fluorescent compound resulting from the oxidation of 5-(and-6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate (CM-H₂DCFDA) (Molecular Probes, Eugene, OR) by several ROS, including H₂O₂ (28). Briefly, cells were incubated with 5 μmol/l CM-H₂DCFDA in serum-free medium for 30 min at 37°C, fixed in 2% paraformaldehyde, and observed under a fluorescence microscope (Zeiss-Axioplan 2; Carl Zeiss Italy, Arese, Milan, Italy) at 488-nm excitation and 530-nm emission.

Nuclear protein extracts were obtained from kidney tissue and mesangial cell monolayers using the TransFactor Extraction Kit (BD Biosciences Clontech, Palo Alto, CA). The activation of NF-κB/p65 was assessed by ELISA using the Mercury TransFactor NF-κB p65 kit (BD Biosciences Clontech) (26,27).

Results are expressed as means ± SD and percent change in diabetic animals versus nondiabetic controls. Statistical significance was evaluated by one-way ANOVA followed by the Student-Newman-Keuls test for multiple comparisons. All statistical tests were performed on raw data.

Metabolic derangement and growth impairment were similar in diabetic KO and wild-type mice; levels of nonfasting blood glucose and A1C were four- and twofold, respectively, those of nondiabetic mice (Table 1).

While serum creatinine levels did not differ significantly among groups (Table 1), glomerular barrier function was significantly impaired in both diabetic groups, with significantly less pronounced changes in diabetic KO versus wild-type mice (Fig. 1A and B). The protein-to-creatinine and albumin-to-creatinine ratios increased by 67 and 64%, respectively, in diabetic versus nondiabetic KO mice, as compared with the 4.08 × and 3.39 × increments observed in diabetic versus nondiabetic wild-type animals.

Kidney weights increased significantly in both diabetic groups, but the increment in renal size was less marked in KO versus wild-type mice (20 vs. 34%) (Table 1). Morphological evaluation of kidneys from diabetic wild-type mice showed significant glomerular sclerosis, with PAS-positive deposits within the mesangium and thickening of glomerular basement membrane and Bowman’s capsule. These changes were observed only rarely in kidney sections from diabetic KO mice (Fig. 2). The glomerular sclerosis index increased in diabetic animals of both genotypes, with significantly lower increments in KO versus wild-type mice (37 vs. 111%) (Table 1), whereas the TID score was unchanged (not shown).

At morphometric evaluation, mean glomerular area, mean glomerular volume, and mean mesangial area were slightly lower (by 10–15%) in nondiabetic KO compared with wild-type mice and increased significantly less in diabetic KO (23–36%) compared with wild-type (31–74%) mice; fractional mesangial area did not differ between the two nondiabetic groups and increased significantly only in diabetic wild-type mice (Fig. 1C–F).

Glomerular staining for active caspase-3 was very low in nondiabetic animals and increased significantly only in diabetic wild-type mice (87% vs. diabetic KO), with predominant involvement of podocytes, although mesangial cells were not spared (Fig. 3).

Transcripts for fibronectin increased in diabetic mice versus the corresponding nondiabetic controls, although increases were significant in wild-type but not KO animals (52 vs. 16%) (Fig. 3C). Conversely, no significant change in kidney cortex mRNA levels for collagen IV α1 chain was observed in diabetic versus nondiabetic mice and between the two genotypes (not shown). Likewise, based on Western blot analysis, fibronectin (Fig. 3D and E), but not collagen IV (not shown), kidney cortex expression increased significantly with diabetes in both genotypes, although it was lower in KO than in wild-type mice (both nondiabetic [−15%] and particularly diabetic [−33%]). Since glomeruli only account for a small percentage of whole kidney cortex (29), their content of these matrix proteins was verified by immunohistochemistry, which showed higher glomerular levels of both fibronectin (Fig. 4A–D) and collagen IV (Figs. 4E–H) in diabetic versus nondiabetic mice, with significantly less pronounced increases in diabetic KO versus wild-type mice. In fact, fibronectin content increased 3.83-fold in wild-type vs. 2.17-fold in KO and collagen IV 4.54-fold in wild-type vs. 2.56-fold in KO compared the corresponding nondiabetic mice (supplemental Fig. 1).

Plasma isoprostane 8-epi-prostaglandin F_2α levels were slightly, although not significantly, lower in nondiabetic KO than in wild-type mice. They increased significantly in diabetic versus nondiabetic wild-type mice (62%), whereas the increment detected in diabetic versus nondiabetic KO mice (19%) did not achieve statistical significance (Table 1).

Circulating and renal tissue AGE levels increased significantly in diabetic mice from both genotypes. Increments over the corresponding nondiabetic mice were significantly lower in KO than in wild-type mice (3.33 × and 4.31 × vs. 6.36 × and 8.35×, respectively) (Table 1). Staining for CML (Fig. 5A–D) and HNE (Fig. 5E–H) was slightly positive only at the tubular level in nondiabetic mice. Tubular immunoreactivity increased in diabetic animals from both genotypes, although it was much less intense and diffuse in KO than in wild-type mice. At the glomerular level, positivity for both CML and HNE was observed in diabetic wild-type mice, whereas immunostaining for CML was barely detectable, and that for HNE virtually absent, in diabetic KO mice. RAGE mRNA levels increased significantly (42%) in diabetic wild-type but not KO mice (Fig. 6A). Likewise, RAGE immunostaining, which was negligible in glomeruli from nondiabetic animals, was significantly upregulated in diabetic wild-type but not KO mice and was localized mainly in podocytes (Fig. 6B–E).

The activation of the redox-sensitive transcription factor NF-κB/p65 within the kidney increased significantly in diabetic versus nondiabetic wild-type (242%) but not KO mice (Table 1).

The kidney cortex p66^Shc mRNA and protein expression increased significantly (by 70 and 73%, respectively) in diabetic versus nondiabetic wild-type mice (supplemental Fig. 2A–C).

Cell death rate by apoptosis increased significantly upon exposure to high glucose only in mesangial cells from wild-type mice (Fig. 7A–D); this finding was confirmed by the assessment of cell viability (Fig. 7E).

Transcripts for fibronectin and collagen IV were found to increase under high- versus normal-glucose conditions in mesangial cells from both genotypes, although increases were significant in cells from wild-type but not KO animals (53 and 67% vs. 12 and 16%, respectively) (Fig . 7F and G). Likewise, fibronectin release increased significantly upon exposure to high glucose in conditioned media from wild-type but not KO mesangial cells (81 vs. 15%) (Fig. 7H).

ROS levels were slightly lower in KO versus wild-type cells under normal-glucose conditions. When exposed to high glucose, ROS levels increased markedly in cells from wild-type mice, whereas they were only slightly enhanced in those from KO mice (Fig. 7I–L). The high-glucose–induced increase in ROS-dependent fluorescence in wild-type cells was virtually abolished by the use of apocynin, diphenyliodonium, 2-thenoyltrifluoroacetone, or myxothiazol (supplemental Fig. 3). NF-κB/p65 activation increased significantly upon exposure to high glucose in mesangial cells from wild-type (297%) but not KO mice (Fig. 7M). Finally, mRNA expression of p66^Shc increased significantly (46%) in wild-type cells grown under high- versus normal-glucose conditions (supplemental Fig. 2D).

The main finding of this study was that changes in renal function and structure were significantly less pronounced or even absent in diabetic p66^Shc KO animals compared with diabetic wild-type mice. This was evidenced by the smaller increment in proteinuria, albuminuria, and glomerular area and volume and, particularly, by the lack of any significant mesangial expansion, the most prominent structural feature of diabetic glomerular disease (14). These differences in the severity of renal involvement between the two diabetic groups occurred despite similar degrees of metabolic derangement.

Diabetic glomerulopathy in wild-type mice was associated with enhanced ECM protein expression and cell death rate, as well as with markedly increased serum AGE and plasma isoprostane 8-epi-prostaglandin F_2α levels and renal tissue AGE accumulation, RAGE expression, and NF-κB/p65 activation, as previously reported (26,29–34). Diabetic wild-type mice also had increased renal mRNA and protein levels of p66^Shc, in keeping with a recent report in lympho-monocytes from type 2 diabetic patients (35). Conversely, diabetic KO mice showed no increment in glomerular cell death rate and less marked matrix deposition and elevation in circulating levels of AGE and isoprostanes 8-epi-prostaglandin F_2α; renal content of AGEs, CML, and HNE; expression of RAGE; and activation of NF-κB/p65 versus wild-type mice. Likewise, mesangial cells from KO mice showed no or significantly attenuated glucose-induced apoptosis, upregulation of ECM and ROS levels, and activation of NF-κB/p65, thus supporting the concept that deficiency of p66^Shc protein was associated with attenuated changes in matrix and cell turnover (i.e., the two major processes underlying mesangial expansion and glomerular sclerosis) and reduced susceptibility to diabetes-induced oxidative stress and AGE formation. The finding that diabetes-induced changes in certain parameters (e.g., mesangial expansion, apoptosis) were completely prevented, whereas those in others (e.g., proteinuria, absolute glomerular and mesangial areas) were only partially, though significantly, reduced in p66^Shc-deficient mice might reflect the different influence of p66^Shc- and oxidant-independent mechanisms (possibly the hemodynamic changes induced by hyperglycemia) on the various alterations in renal function and structure. This might also explain the differences in the extent of reduction of ECM upregulation between the in vivo and in vitro conditions. The observation that both glomerular and mesangial areas, but not fractional mesangial area, were reduced in nondiabetic KO versus wild-type mice suggest that resistance to oxidative stress might also be associated with less marked age-related kidney changes (27).

Taken together, these results are consistent with the concept that extensive AGE formation and induction of oxidative stress by chronic hyperglycemia play a pivotal role in the pathogenesis of diabetic glomerulopathy, through the activation of redox-sensitive transcription factors, particularly NF-κB (36). This conclusion is consistent with previous studies indicating that treatment with various antioxidants (33,37,38) or overexpression of the antioxidant enzyme superoxide dismutase (39,40) are effective in preventing or attenuating experimental diabetic glomerulopathy. Our data are also consistent with previous reports in this animal model showing attenuation of nondiabetic vascular disease associated with aging and oxidative stress (21–23).

More importantly, these results shed light on the molecular mechanism(s) underlying diabetes-induced oxidative stress and oxidant-mediated renal injury by providing evidence of the involvement of the p66^Shc pathway. It was shown that proapoptotic signals lead to increased cytosolic and mitochondrial p66^Shc pools (19,41), serine phosphorylation of the cytosolic pool (17), and release of the mitochondrial pool from a high–molecular weight complex (42) that could enable p66^Shc to oxidize cytochrome c, thus generating proapoptotic H₂O₂. Diabetes was shown to provide propapoptotic signals via several mechanisms, including hyperglycemia per se, stretching, AGEs, oxidative stress, and NF-κB activation (43). Moreover, it was speculated that the reaction with cytochrome c might occur if an excess of reduced cytochrome c is present. This condition is achieved when electron flow from complex III is increased, such as in diabetes, or the activity of cytochrome c oxidase is decreased, such as during hypoxia (and possibly the “pseudohypoxia” produced by diabetes [3]) or due to inhibition by nitric oxide (20) (which, under certain conditions, is upregulated in the diabetic state [29]). Recently, gene delivery of Tim44, a component of the high–molecular weight complex that inactivates p66^Shc within the mitochondrion, was found to normalize increased ROS generation, altered inner membrane potential, and enhanced proliferation in smooth muscle cells cultured under high glucose and to ameliorate ROS production, inflammatory response, and neointimal proliferation in the balloon-injured carotid artery of diabetic rats (44). In addition to mediating oxidant-induced apoptosis, p66^Shc was found to regulate steady-state ROS levels, which therefore would reflect moderate p66^Shc activation by chronic stress. Also, this action of p66^Shc could be enhanced by the diabetic state via induction of a sustained mitochondrial ROS overproduction or cytosolic NAD(P)H oxidase activation. The prevention of high-glucose–induced ROS upregulation by inhibitors of both NAD(P)H oxidase and mitochondrial respiratory chain in wild-type cells is in keeping with a recent report in cultured podocytes (45) and indicates that both cytosolic and mitochondrial mechanisms participate in ROS generation under hyperglycemic conditions. Thus, the fact that ROS-dependent fluorescence was abolished by p66^Shc ablation in wild-type cells suggests that p66^Shc regulates this process at both levels.

Prevention of ROS generation by p66^Shc deletion would interfere with oxidative stress–dependent apoptosis and matrix accumulation, in keeping with previous reports indicating that both alterations are ameliorated by the use of antioxidants (33,38,46) or overexpression of the antioxidant enzyme superoxide dismutase (39,40). In diabetic glomerulopathy, ECM alterations prevail quantitatively over changes of cell compartment, which is likely reduced due to enhanced cell death by apoptosis (15). Programmed cell death, a normal event in renal tissue homeostasis (47), has been considered a major mechanism not only for resolution of glomerular hypercellularity in glomerulonephritis (43,48) but also for the loss of cellularity and progression to glomerulosclerosis in chronic renal disease (43,47). This is suggested by the enhanced apoptosis rate detected in several forms of human and experimental renal disease (49), including diabetic nephropathy (31), and by its correlation with loss of renal function and structure (49). Abnormal ECM-cell interaction due to mesangial matrix expansion and consequent disruption of survival factors supplied by collagen IV and laminin (but not collagen I and fibronectin) through β1-integrin could also be responsible for the promotion of apoptosis (47,50). Alternatively, glomerular cell loss through apoptosis might trigger ECM deposition and scarring. Our data showing that p66^Shc deletion prevented diabetes-induced increase of glomerular cell apoptosis, in parallel with reduction of ECM deposition, seem to support the concept of a major role of apoptosis in diabetic glomerulopathy, in keeping with previous reports in p66^Shc-deficient mice indicating that protection of vascular tissue from atherogenesis induced by high-fat diet and tissue damage by ischemia and ischemia reperfusion were associated with reduced cell death rate (21,22).

In conclusion, these data show that p66^Shc ablation is associated with almost complete protection from the development of diabetic glomerulopathy. This protection appears to be dependent on a reduced susceptibility to diabetes-induced, p66^Shc-dependent oxidative stress and consequent negative regulation of apoptosis and ECM deposition. In addition, these results suggest that an intervention targeted to p66^Shc might be effective in the prevention and treatment of diabetic glomerulopathy.

This work was supported by grants from the European Foundation for the Study of Diabetes/Servier, the Ministry of University and Research of Italy (40%), and the Diabetes, Endocrinology and Metabolism Foundation (Rome).