Catalase Deficiency Accelerates Diabetic Renal Injury Through Peroxisomal Dysfunction

Reagents for immunohistochemical staining were purchased from DAKO A/S (Glostrup, Denmark), anti–E-cadherin was obtained from BD Biosciences (San Jose, CA), anti-catalase was obtained from Young In Frontier (Seoul, Korea), and other antibodies were obtained from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). FBS was purchased from GIBCO BRL (Gaithersburg, MD), and MitoSOX was obtained from Molecular Probes (Eugene, OR). Other chemicals and tissue culture plates were obtained from Sigma-Aldrich (St. Louis, MO) and Becton Dickinson Labware (Lincoln Park, NJ), respectively, unless otherwise stated.

Male catalase wild-type (WT) and CKO C57BL/6 J mice (17) were used in this study. All animal experiments were approved by Ewha Womans University Institutional Animal Care and Use Committee. Six-week-old mice were divided into four groups: nondiabetic and diabetic WT mice and nondiabetic and diabetic CKO mice. Diabetes was induced by intraperitoneal injection of streptozotocin (STZ; 50 mg/kg/day) for 5 days and mice were kept up to 10 weeks of diabetes (25). Blood pressure was measured every week by CODA High-Throughout Non-Invasive Tail Blood Pressure system (Kent Scientific Corporation, Torrington, CT).

MMCs were cultured as described previously (26). Subconfluent MMCs were transfected with 20 nmol/L siCat (sense, 5′-CCA GAU ACU CCA AGG CAA ATT-3′; antisense, 5′-UUU GCC UUG GAG UAU CUG GTT-3′) using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA) for 24 h. For overexpression experiments, MMCs were transfected with 100 ng p-cytomegalovirus (pCMV)–human catalase (hCat) plasmid (supplied by Professor Sue Goo Rhee, Ewha Womans University) after siRNA transfection for 24 h.

Blood glucose, urinary creatinine and albumin concentrations, and body weight were measured every 2 weeks. Urine was collected during two consecutive 24-h periods, during which each mouse was individually housed in a metabolic cage with free access to food and water. Blood glucose was measured by the glucose oxidase method, plasma creatinine was measured by Detect X Serum Creatinine Detection Kit (Arbor Assays, Ann Arbor, MI), and plasma FFAs were measured by EnzyChrom Free Fatty Acid Assay Kit (BioAssay Systems, Hayward, CA). Urinary protein was analyzed by Bradford assay and SDS-PAGE (27). The levels of urinary lipid peroxide (LPO) were measured by the thiobarbituric acid method as described previously (28).

Mitochondrial ROS were measured using MitoSOX. In brief, after stimulation, the cells were washed and incubated with 2.5 μmol/L MitoSOX for 10 min. The unbound dye was then removed and fluorescence was analyzed using a flow cytometer (BD Biosciences) and visualized by confocal microscopy (Carl Zeiss, Gottingen, Germany).

In each mouse, quantitative analysis of glomerular volume and fractional mesangial area (FMA) in the glomeruli stained with periodic acid Schiff (PAS) reagent was conducted as described previously (25).

Immunohistochemistry for nitrotyrosine and FN ED-A was performed as described previously (29) using anti–N-nitrotyrosine (1:200) and anti–FN ED-A (1:200). For the detection of peroxin (PEX)5 in MMCs, the cells were fixed with 4% paraformaldehyde for 10 min, incubated with serum-free blocking solution for 15 min, and then reacted with anti-PEX5 (1:200). After washing two times with PBS, the samples were incubated with Alexa Fluor 488–tagged goat anti-mouse IgG (H+L) (1:100) for 1 h. Images were obtained by confocal microscopy (Carl Zeiss).

Isolation of total RNA from the whole kidney and MMCs, synthesis of cDNA by reverse transcription, and quantitative real-time PCR were performed as described previously (25). PCR primer sequences are listed in Supplementary Table 1.

E-cadherin, α-smooth muscle actin (α-SMA), and FN protein were measured by a standard Western blot analysis as described previously (25), using anti-human E-cadherin (1:5,000), anti-mouse α-SMA (1:10,000), and anti-human FN (1:2,000).

For Oil Red O staining, fixed tissues and cells were washed and incubated with 0.3% Oil Red O for 10 min. After washing with distilled water, images were obtained by microscopy (Carl Zeiss).

All results are expressed as mean ± SE. ANOVA was used to assess the differences between multiple groups. If the F statistic was significant, the mean values of each group were compared by Fisher least significant difference method. P < 0.05 was considered significant.

After multiple injections of low-dose STZ, both diabetic WT and diabetic CKO mice maintained hyperglycemia up to 10 weeks of diabetes, but there was no difference between the two groups (Table 1). The kidney weight (absolute and adjusted) of diabetic CKO mice was significantly increased when compared with diabetic WT mice at both 4 and 10 weeks of diabetes (Table 1). Mean blood pressure (mmHg) was significantly increased from 5 weeks of diabetes, but there was little difference between diabetic WT and diabetic CKO mice (WT: 108 ± 8; diabetic WT: 135 ± 12; diabetic CKO: 140 ± 4 at 10 weeks of diabetes), suggesting that blood pressure did not play a significant role in accelerated renal injury in diabetic CKO mice under our experimental condition.

Urinary protein excretion was gradually increased up to 10 weeks of diabetes and was significantly increased in diabetic CKO compared with diabetic WT mice at both 4 weeks and 10 weeks of diabetes (Fig. 1A and B). Plasma creatinine levels were also gradually increased in both diabetic groups and were significantly higher in diabetic CKO mice than in diabetic WT mice at 10 weeks of diabetes (Fig. 1C). Glomerular volume in both diabetic groups was gradually increased up to 10 weeks of diabetes, and in diabetic CKO mice, glomerular volume and FMA were higher than that of diabetic WT mice at both 4 and 10 weeks of diabetes (Fig. 1D–F). It is important to note that glomerular volume and FMA in nondiabetic CKO mice were significantly higher than age-matched nondiabetic WT mice (Supplementary Fig. 1A and B). These results suggest that catalase deficiency exacerbates renal injury in early diabetes.

We next measured the pro- and antifibrotic cytokines along with markers of epithelial-to-mesenchymal transition, which is an early event in DN (30). TGF-β1 (Fig. 2A), connective tissue growth factor (CTGF) (Fig. 2B), α-SMA (Fig. 2D) mRNA, and α-SMA protein (Fig. 2F) expression levels were significantly increased in diabetic CKO compared with diabetic WT mice at 4 and 10 weeks of diabetes. CTGF and α-SMA expressions in diabetic CKO kidneys were more significantly increased at 10 weeks of diabetes. Conversely, bone morphogenic protein 7 (BMP7)—a renoprotective antifibrotic cytokine—mRNA (Fig. 2C), E-cadherin mRNA (Fig. 2E), and protein (Fig. 2G) expressions in diabetic CKO kidneys were significantly decreased compared with diabetic WT kidneys at both 4 and 10 weeks of diabetes. In nondiabetic CKO kidneys, the expressions of TGF-β1, CTGF, α-SMA, and FN mRNA were significantly upregulated, while BMP7 and E-cadherin mRNA were significantly downregulated compared with nondiabetic WT kidneys (Supplementary Fig. 1G).

Consistently, FN mRNA and protein expressions in diabetic CKO kidneys were significantly higher than in diabetic WT kidneys. The increase in FN protein levels in both diabetic WT and CKO kidneys were more evident at 10 weeks of diabetes than at 4 weeks (Fig. 3A and B). Immunohistochemical analysis at 4 weeks of diabetes confirmed that FN ED-A was strongly expressed in the glomeruli of diabetic WT mice and in the tubules and glomeruli of diabetic CKO mice (Fig. 3C). Masson trichrome staining showed significantly increased collagen accumulation in diabetic WT kidneys, which was further increased in diabetic CKO kidneys at 10 weeks of diabetes (Fig. 3D). These results show that catalase deficiency accelerates renal fibrosis in early diabetes.

Catalase null state was confirmed by real-time PCR in the CKO kidneys (Supplementary Fig. 2A). In diabetic WT kidneys, the mRNA levels of antioxidant enzymes, including catalase, glutathione peroxidase-1 (GPx1), peroxiredoxin (Prx)1, Prx5, superoxide dismutase (SOD)1, SOD2, and sulfiredoxin-1, were significantly upregulated compared with nondiabetic WT kidneys (Supplementary Fig. 2A–G). Moreover, Prx1 and Prx5 mRNA expressions in nondiabetic CKO kidneys were significantly higher than in nondiabetic WT kidneys. Unexpectedly, upregulated expressions of antioxidant enzymes were blunted in diabetic CKO kidneys. In fact, renal expressions of GPx1, Prx5, SOD1, SOD2, and sulfiredoxin-1 mRNA in diabetic CKO mice were significantly lower than in diabetic WT mice.

Renal Nox4 plays an important role in increased oxidative stress leading to DN (31), and H₂O₂ upregulates Nox4 mRNA expression in MMCs (32). We, therefore, measured Nox4 mRNA expression in addition to urinary LPO and nitrotyrosine accumulation as markers of renal oxidative stress. Renal Nox4 mRNA (Fig. 4A) expression, urinary LPO (Fig. 4B), and renal nitrotyrosine accumulation (Fig. 4C and D) were higher in diabetic CKO mice than in diabetic WT mice at 4 weeks of diabetes and were continuously increased in diabetic CKO mice up to 10 weeks. These results, together with the mRNA expression profile of antioxidant enzymes in diabetic CKO kidneys, demonstrate that renal oxidative stress is greatly increased in diabetes by catalase deficiency.

We investigated peroxisomal biogenesis as the first step in a mechanistic study of the relationships among catalase, peroxisome, oxidative stress, and DN. Plasma FFAs were significantly upregulated in both diabetic WT and diabetic CKO mice compared with control WT at 4 weeks and further increased in diabetic CKO at 10 weeks of diabetes (Fig. 5A). Renal expressions of peroxisomal biogenesis markers including, PEX5, PEX11-α, and ATP-binding cassette member 3 (Abcd3), were significantly higher in diabetic WT mice at 4 weeks but lower at 10 weeks compared with nondiabetic WT mice (Fig. 5B–E). In diabetic CKO kidneys, these peroxisomal biogenesis markers were significantly decreased compared with diabetic WT kidneys at both 4 and 10 weeks (Fig. 5B–E).

To investigate a direct role of endogenous catalase in the progression of DN, we used MMCs transfected with siCat. siCat decreased the endogenous catalase level down to 40% of basal in MMCs and was restored by additional catalase expression using pCMV-hCAT (Fig. 5F). PEX5, PEX11-α, and Abcd3 mRNA expressions were significantly upregulated by 100 μmol/L PhA with or without 30 mmol/L d-glucose (HG) for 4 h in MMCs, but these increases were inhibited by catalase knock-down (Fig. 5G and H and Supplementary Fig. 3B–E). Peroxisomal biogenesis markers inhibited by catalase deficiency were restored by hCAT transfection, confirming the on-target effect of siCat. N-acetylcysteine (NAC) also effectively restored the inhibition of peroxisomal biogenesis induced by catalase deficiency. Decreased PEX5 expression was further confirmed by Western blot and immunocytochemistry (Fig. 5J and I). It is interesting that 100 μmol/L H₂O₂ treatment significantly increased peroxisomal biogenesis, but not at >250 μmol/L (Supplementary Fig. 3A). These results suggest that FFA-induced oxidative stress requires catalase for peroxisomal fitness, but severe oxidative stress, such as H₂O₂ at levels >250 μmol/L or catalase deficiency under diabetic conditions, disrupts adaptive peroxisomal biogenesis.

To investigate whether catalase deficiency induces the accumulation of mitochondrial ROS, we measured mitochondrial superoxide anion levels using MitoSOX. MitoSOX-sensitive ROS were significantly increased by PhA with or without HG treatment and further increased by catalase knock-down (Fig. 6A and B). Accelerated mitochondrial ROS in catalase knock-down cells were effectively prevented by hCAT transfection or by NAC treatment. MitoSOX levels were not significantly different between PhA alone and PhA plus HG treatment.

To determine if catalase deficiency induced impaired mitochondrial fitness, we measured cytochrome b mRNA level, a marker of mitochondrial biogenesis, in MMCs and diabetic animals. As a result of increased mitochondrial ROS, catalase deficiency inhibited PhA with or without HG–induced cytochrome b mRNA expression, which was restored by hCAT transfection or NAC treatment (Fig. 6C). The PhA plus HG–induced cytochrome b mRNA level was higher than PhA treatment alone (Fig. 6D). Similar to peroxisomal biogenesis markers, cytochrome b mRNA level was higher in diabetic WT kidneys at 4 weeks but lower at 10 weeks of diabetes compared with nondiabetic WT kidneys (Fig. 6E). In diabetic CKO kidneys, cytochrome b mRNA level was significantly decreased at both 4 and 10 weeks compared with diabetic WT kidneys (Fig. 6E). These results suggest that catalase deficiency increases mitochondrial ROS, which might play a role in dysregulation of mitochondrial biogenesis.

To determine if the impaired peroxisomal and mitochondrial biogenesis induced by catalase deficiency under the diabetic milieu affected renal lipid metabolism, we measured the peroxisomal and mitochondrial β-oxidation and lipid accumulation. mRNA expressions of acyl-CoA oxidase (ACO), a key enzyme of peroxisomal β-oxidation, and carnitine palmitoyltransferase (CPT)1-α, a regulator of long-chain FA entry to mitochondria, were significantly decreased in diabetic WT compared with control WT kidneys and were further decreased in diabetic CKO kidneys at both 4 and 10 weeks of diabetes (Fig. 7A). Oil Red O staining revealed more lipid droplets in the glomeruli and tubules of diabetic WT mice than in nondiabetic WT mice at 10 weeks of diabetes (Fig. 7B). In diabetic CKO mice, lipid droplets in the glomeruli were significantly higher than in diabetic WT mice (Fig. 7B).

In MMCs, ACO, and CPT1-α, mRNA expressions were significantly upregulated by PhA with or without HG, and these were inhibited by catalase knock-down (Fig. 7C). Again, hCAT transfection or NAC treatment restored the decreased ACO and CPT1-α mRNA expressions in catalase knock-down cells (Fig. 7C). ACO mRNA levels were not different between MMCs treated with PhA alone versus PhA plus HG, but CPT1-α mRNA levels were higher in MMCs treated with PhA alone compared with PhA plus HG (Fig. 7C). Oil Red O staining showed significantly increased lipid droplets in catalase knock-down cells in response to PhA plus HG, and these were effectively inhibited by NAC treatment (Fig. 7D).

FN mRNA and protein expressions were significantly increased in MMCs cultured under PhA plus HG, which was more evident than PhA treatment alone, and were further increased in catalase knock-down cells (Fig. 7E and F). Induction of FN in catalase deficiency was significantly inhibited by hCAT transfection or NAC treatment (Fig. 7E and F). These results suggest that catalase deficiency impairs peroxisomal and mitochondrial β-oxidation, leading to lipid and ECM accumulations.

This study provides experimental evidence that endogenous catalase plays an important role in renal redox regulation and function under diabetic conditions. Despite equivalent hyperglycemia between diabetic WT and diabetic CKO mice, parameters of DN together with plasma FFAs were significantly upregulated in diabetic CKO mice. These accelerations were accompanied by increased urinary LPO and renal nitrotyrosine accumulation in diabetic CKO mice, confirming that increased ROS in diabetic CKO mice are associated with accelerated diabetic renal injury.

It should be noted that the compensatory increase of antioxidant enzymes observed in diabetic WT kidneys was attenuated in diabetic CKO kidneys. Although the underlying mechanism is unclear, catalase deficiency might have inhibited Nrf2 expression in diabetic kidneys, for in chronic obstructive pulmonary disease (33), severe oxidative stress inhibits Nrf2, which is the primary transcription factor responsible for upregulation of oxidative responses. In this connection, diabetic renal injury was enhanced in Nrf2 null mice (34); however, not all endogenous antioxidant enzymes play equal roles in diabetic kidneys. For example, GPx1 deficiency rendered similar diabetic renal injury in WT mice (35).

Therefore, it is important to delineate the pathogenic mechanism involved in accelerated diabetic renal injury by catalase deficiency. Accordingly, we attempted to determine whether catalase deficiency disturbs peroxisomal biogenesis, leading to increased mitochondrial ROS and lipid accumulation observed in DN. Peroxisomes undergo reorganization in response to metabolic demands for FA β-oxidation under increased FFAs (21,36). In our in vivo and in vitro study, PEX5, PEX11-α, and Abcd3 mRNA expressions were upregulated in 4 weeks of diabetic WT kidneys and MMCs cultured under PhA with or without HG but not in diabetic CKO kidneys and catalase knock-down MMCs. We also observed increased peroxisomal biogenesis markers in response to 100 μmol/L H₂O₂, but not >250 μmol/L, suggesting context-dependent ROS effects (37). In the case of mitochondria, under oxidative/metabolic stress, initial adaptive increase in mitochondrial biogenesis takes place, while excessive ROS can result in mitochondrial damage (38). Likewise, H₂O₂ produced during FA β-oxidation may increase peroxisomal biogenesis–related genes to meet its metabolic demand, but H₂O₂ above a threshold by catalase deficiency may result in peroxisomal damage. NAC and additional catalase using hCAT transfection confirmed that ROS inhibition results in restoration of peroxisomal biogenesis in catalase knock-down cells. Measurement of in situ H₂O₂ in kidneys by using recently developed methods (39) would be needed to assess the effects in vivo. In addition, although the selectivity of siCat was confirmed partly by hCAT transfection, use of primary MMCs from CKO mice is needed.

Short- and medium-chain FA derived from the diet are exclusively degraded in the mitochondria, whereas peroxisomes catalyze very long-chain FA or branched-chain FA (40). This specifically implies that there would be a metabolic interconnection between the two organelles. Indeed, mitochondrial abnormalities were described in PEX5-deficient mice (41) and patients with Zellweger syndrome (42). In addition, Elsner et al. (18) showed that accumulated H₂O₂ in the mitochondria was reduced by overexpression of peroxisomal catalase in insulin-producing cells. Consistently, the MitoSOX experiment clearly showed that basal as well as PhA-induced mitochondrial ROS were significantly upregulated in catalase knock-down cells. It is notable that mitochondrial ROS, as well as markers of peroxisomal biogenesis (PEX5, PEX11-α, and Abcd3) and peroxisomal β-oxidation (ACO), were significantly increased by PhA treatment alone, which were not significantly different from PhA plus HG treatment, suggesting that PhA plays an important role in the regulation of mitochondrial ROS, peroxisomal biogenesis, and β-oxidation. We next evaluated whether ROS-induced mitochondrial biogenesis was also impaired by catalase deficiency. As expected, mitochondrial biogenesis estimated by cytochrome b expression (Fig. 6C and D) and mitochondrial DNA and PGC1-α expression (data not shown) were upregulated in MMCs cultured under PhA with or without HG, which were inhibited in catalase knock-down cells. Again, NAC and hCAT transfection restored inhibitory effects of catalase deficiency, suggesting that mitochondria also are damaged at higher levels of ROS. These observations suggest that peroxisomal impairment under catalase deficiency may result in overexpression of mitochondrial ROS and mitochondrial impairment.

Finally, we determined whether peroxisomal and mitochondrial disruption of catalase deficiency influenced lipid metabolism in diabetic kidneys. In contrast to increased peroxisomal/mitochondrial biogenesis at 4 weeks of diabetes, ACO and CPT1-α mRNA expressions were significantly decreased in both diabetic groups, and a greater decrease was observed in diabetic CKO kidneys than in diabetic WT kidneys. Although discrepancies between peroxisomal/mitochondrial biogenesis and β-oxidation at 4 weeks of diabetes are not clear, decreased ACO mRNA in diabetic kidneys has been previously observed (43). It is conceptually supposed that in early diabetes, increased FFAs induce both peroxisomal/mitochondrial biogenesis and β-oxidation. However, when ROS are increased with development of diabetes, β-oxidation is downregulated. We observed that ACO and CPT1-α mRNA expressions were increased in response to PhA with or without HG in control MMCs, but these were inhibited in catalase knock-down cells. We also observed that 100 μmol/L H₂O₂ decreased ACO mRNA expressions but increased peroxisomal biogenesis markers (Supplementary Fig. 3A and F), suggesting that there might be different thresholds for H₂O₂ concentration between biogenesis and β-oxidation. However, further increased ROS in 10 weeks of diabetes inhibited both biogenesis and β-oxidation. Consistent with decreased β-oxidation, lipid droplets were increased in diabetic CKO kidneys and in catalase knock-down cells cultured under PhA plus HG, which were effectively inhibited by NAC. Therefore, increased ROS in catalase deficiency impair peroxisomal/mitochondrial FA oxidation and subsequently increase lipid accumulation, which might induce renal injury, represented by FN expression in MMCs and diabetic kidney. To the best of our knowledge, this is the first study to focus on peroxisomal dysfunction as a cause of redox dysregulation leading to diabetes complications under catalase deficiency. In summary, the present data provide evidence that FFA-induced peroxisomal dysfunction exacerbates diabetic renal injury and that endogenous catalase is an important antioxidant protecting the kidney through maintaining peroxisomal and mitochondrial fitness under diabetic stress. Further studies elucidating the renoprotective pharmacological and molecular methods to restore peroxisomal biogenesis will provide new insight into the therapeutic strategy against diabetic renal injury.

This work was supported by National Research Foundation grants 2011-0006244 and R31-2008-000-10010-0 funded by the Korean Ministry of Education, Science, and Technology; the second stage of the Brain Korea 21 Project; and RP-Grant 2009 (2009-1949-1-1) from Ewha Womans University to J.P.

No potential conflicts of interest relevant to this article were reported.

I.H. researched data and wrote the manuscript. J.L. and J.P. researched data. J.Y.H. and H.B.L. contributed to discussion. Y.-S.H. created the CKO mice. H.H. contributed to discussion and reviewed and edited the manuscript. H.H. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Parts of this study were presented in abstract form at the 44th Annual Meeting of the American Society of Nephrology (abstract TH-PO541), Philadelphia, Pennsylvania, 8–13 November 2011, and at the XLVII ERA-EDTA Congress (abstract 451358), Munich, Germany, 25–28 June 2010.

The authors thank Professors Sue Goo Rhee and Hyun-Ae Woo at Ewha Womans University for supplying CKO mice and pCMV-hCat.