Peroxidase expression is decreased by palmitate in cultured podocytes but increased in podocytes of advanced diabetic nephropathy

High levels of serum free fatty acids (FFAs) are associated with lipotoxicity and type 2 diabetes. Palmitic acid (PA) is the predominant circulating saturated FFA. PA induces mitochondrial superoxide and hydrogen peroxide (H2O 2) generation in cultured podocytes. To elucidate the role of PA in antioxidant defense systems in diabetic nephropathy (DN), cultured podocytes were exposed to 250 μM PA for 1–24 hr, and protein expressions of catalase, peroxiredoxins (Prxs), and glutathione peroxidase (GPx) were examined by western blot analysis. PA induced an early transient increase in the Prx1, Prx2, and GPx1 levels in podocytes, but not catalase. Long‐term exposure of PA to podocytes significantly decreased the protein levels of Prx1, Prx2, GPx1, and catalase. Coincubation of PA‐treated cells with oleic acid, however, restored the expression of these proteins. In advanced human diabetic glomeruli, H2O2 generation was elevated as shown by increased fluorescence of dichlorofluorescein. Strong immunostaining for Prx1, Prx2, GPx1, and catalase was observed in the podocytes of advanced human DN, wherein transforming growth factor‐β1 staining was also positive. These results suggest that podocytes are susceptible to PA‐induced oxidative damage with impaired peroxidase activity and that peroxidases have futile antioxidant effects in the podocytes in the late stages of DN. Given this, PA‐induced podocyte injury via inadequate peroxidase response to H2O2 appears to play an important role in the pathogenesis of DN.

Catalase overexpression in the renal proximal tubular epithelial cells in diabetic animals attenuated ROS generation or the progression of nephropathy (Brezniceanu et al., 2007;Shi et al., 2013). PA increased catalase expression as a line of defense against peroxides in cultured tubular epithelial cells (Ruggiero et al., 2014).
Expression of Prx1, Prx3, and Prx5 was increased in the glomeruli of diabetic mice (Barati et al., 2007). Knockdown of Prx2, but not Prx1, induced significant increases in the intracellular ROS in cultured podocytes (Hsu et al., 2011). Transfection of tubular epithelial cells with Prx2 was protective and mitigated apoptosis (Ruggiero et al., 2014). PA led to a decreased expression of Prx2 protein in cultured tubular epithelial cells (Ruggiero et al., 2014), and so did angiotensin II in cultured podocytes (Hsu et al., 2011).
Of the five GPx isoforms, GPx1 is the only cytosolic enzyme and is mainly present in normal kidneys (de Haan et al., 2005). GP x 1 knockdown or podocyte GPx1 loss in diabetic mice had no increased risk for glomerular damage or oxidative stress as compared with wild-type diabetic mice (Blauwkamp et al., 2008;de Haan et al., 2005). In this regard, GPx1 may not be protective against oxidative renal injury during the development of DN. In cultured podocytes, HG induced no changes in GPx protein expression (Piwkowska et al., 2011).
It is proposed that excess generation of mitochondrial ROS in response to HG (Nishikawa et al., 2000) or PA (E. Lee et al., 2017) plays a central role in the initiation of DN. Indeed, mitochondrial ROS was increased in the glomeruli of living diabetic mice (Galvan et al., 2017), while Dugan et al. (2013) reported opposite findings.
Little is known about the role of peroxidases in PA-induced podocyte injury and pathogenesis of DN. In this regard, we examined the expression of catalase, Prx1, Prx2, and GPx1 proteins in cultured podocytes exposed to PA and OA. Furthermore, H 2 O 2 production and peroxidase expression were examined in the glomeruli of patients with DN.

| Protein extraction and western blot analysis
Differentiated podocytes in collagen-coated six-well plates were serum-starved for 24 hr and treated with 5% BSA, 250 μM PA or OA for 1-24 hr as described previously (E. Lee et al., 2017).
The cells were scraped and purified. Protein quantification was determined using the Lowry method. SDS-PAGE was performed, and the protein was transferred onto polyvinylidene fluoride membranes. The membranes were blocked with 5% BSA and incubated overnight with anti-catalase, anti-Prx1, anti-Prx2, and anti-GPx1. They were then incubated with an HRP-linked secondary antibody for 2 hr and luminescence was created using an enhanced chemiluminescence kit, before imaging and analysis with a chemiluminescence imaging system (Alliance-LD2-87.WL/Auto; Uvitec; Cambridge, UK). Densitometry was performed and processed using the Gen5 software package (Bio-Tek, Winooski, VT). To assess the equality of protein loading, the membrane was reprobed with anti-β-actin.

| Catalase activity assay
Podocytes were exposed to PA for 16 hr. Afterwards, podocytes were lysed on ice in 50 mM potassium phosphate, 0.1 mM ethylenediaminetetraacetic acid (EDTA), 0.1% Triton X-100 (pH 7.8). Extracts were centrifuged at 12,000g for 20 min at 4°C and supernatants were used. The initial rate of disappearance of H 2 O 2 was recorded at a wavelength of 240 nm during 1 min in reaction cell lysates containing 10 mM H 2 O 2 , 50 mM potassium phosphate, 0.1 mM EDTA, pH 7.0 as described by Aebi (1984).

| Detection of glomerular H 2 O 2 production
Frozen kidney biopsy samples of patients with DN and controls were cut into 6-µm-thick sections and placed on a glass slide. Sections were washed with PBS, and incubated with 10 μM carboxy-DCFH DA-AM for 2.5 hr at 37°C in a 5% CO 2 environment. Dichlorofluorescein (DCF) fluorescence was examined (excitation/emission: 490/516 nm) using a Zeiss fluorescence microscope (AXIO Scope A1; Carl Zeiss, Heidenheim, Germany).

| Immunohistochemistry
An avidin-biotin-peroxidase procedure was used for antibody localization. Paraffin-embedded kidney sections (3 µm) were deparaffinized serially. For antigen retrieval, sections were treated with trypsin (Digest-All 2; Invitrogen) for 30 min at 37°C. Endogenous peroxidase activity was quenched with 10% methanol-H 2 O 2 solution for 10 min. Sections were then incubated with primary antibodies against Prx1, Prx2, GPx1, and catalase for 1 hr at room temperature. In addition, they were incubated overnight with rabbit anti-human TGF-β1 at 4°C. Biotinylated goat anti-rabbit IgG or anti-mouse IgG was used as a secondary antibody. Then sections were incubated with streptavidin-conjugated HRP complex, followed by the addition of diaminobenzidine (Sigma-Aldrich) and counterstaining with Mayer's hematoxylin.
Control experiments were performed by omitting the primary antibody or replacing it with the corresponding nonimmune serum.

| Statistical analysis
Data were presented as mean ± SD of three separate experiments.
Results were analyzed by analysis of variance for three groups or by Wilcoxon's rank sum test between two groups. A P value of less than 0.05 was considered significant.

and GPx1 protein expression in cultured podocytes
Incubation of podocytes with 250 µM PA for 1 hr showed no changes in Prx1, Prx2, and GPx1 expression as compared with controls. When cells were exposed to PA for 3 hr, the percentage of Prx1, Prx2, and GPx1 proteins was increased by 300 ± 71%, 375 ± 65%, and 283 ± 24%, respectively, as compared with controls (Figure 1a,b).
3.2 | Long-term incubation of PA decreases the expression of Prx1, Prx2, and GPx1 proteins in podocytes After 6-8-hr incubation with PA, the levels of Prx1, Prx2, and GPx1 proteins in podocytes began to decrease, reaching those of controls.

| H 2 O 2 or HG induces no changes in catalase expression in podocytes
When cultured podocytes were exposed to 3-5 mM H 2 O 2 for 10 min to 1 hr, no significant difference was shown in catalase protein levels as compared to controls (Figure 5a). In addition, incubation of cells with HG (25 mM) for 24-96 hr induced no significant changes in catalase expression, either (Figure 5b).

| ROS-H 2 O 2 production is increased in the glomeruli of human DN
In the controls, no DCF fluorescence was detected (Figure 6a).
In DN, scattered DCF fluorescence appeared in the glomeruli, as shown by fluorescence microscopy (Figure 6b). In advanced DN, particularly, intraglomerular DCF fluorescence was markedly increased, even forming aggregates or clumps (Figure 6c).

| Immunostaining for Prx1, Prx2, GPx1, and catalase is increased in the podocytes of advanced human DN
In the controls, there was no glomerular staining for Prx1, Prx2, GPx1, catalase, and TGF-β1 (Figure 7a In this study, short-term incubation of PA briefly increased the Prx1, Prx2, and GPx1 levels in podocytes. Recently, we showed that PA led to a significant increase in superoxide and H 2 O 2 production in cultured podocytes (E. Lee et al., 2017). Thus, the enhanced expression of these peroxidase proteins could be the most feasible early response of podocytes to the intracellular ROS generation. In experimental DN, urinary excretion of Prx1 was also increased before development of overt histological damage (Korrapati et al., 2012).

Our demonstration of early transient elevation of Prx1 and Prx2
by PA supports the previous notion that Prxs, particularly Prx2, must react sufficiently rapidly with H 2 O 2 to compete with other peroxidases, consuming basal levels of H 2 O 2 (Peskin et al., 2007).
Yet when peroxide production increases, Prxs are inactivated by peroxide-induced hyperoxidation to protect the cell from further oxidative protein damage (Day et al., 2012). Indeed, we found that after 16-hr incubation with PA, protein levels of Prx1 and Prx2 were decreased, which could be related to exhaustion of cytosolic Prx defense.
After Prxs are inactivated, catalase would then play a role to remove excess H 2 O 2 . The concerted role for Prx2 and catalase was demonstrated in tubular epithelial cell protection (Ruggiero et al., 2014). Nonetheless, incubation of podocytes with PA for 16-24 hr rather decreased the catalase levels in this study. Furthermore, neither HG nor 5 mM H 2 O 2 enhanced podocyte catalase expression.
Although Piwkowska et al. (2011) observed a decreased catalase protein expression after a 5-day incubation of HG with podocytes, a maximum 4-day incubation of HG in this study did not induce any changes in catalase expression. We could not extend the incubation time to 5 days because of severe cell apoptosis or lysis, not confirming their results.  (Lu et al., 2015). Interestingly, patients with catalase gene mutation have life-long increased H 2 O 2 concentration, which has cytotoxic effects on pancreatic cells, to be a risk factor for diabetes (Goth, 2008).
We also observed that the decreased levels of Prx1, Prx2, GPx1, and catalase proteins in podocytes after long-term PA exposure were restored by OA, supporting the previous notion that PA suppresses the antioxidative defense, whereas OA preserves it (Sargsyan et al., 2016).
To sum up, long-term exposure of podocytes to PA decreased Prx1, Prx2, GPx1, and catalase levels. Thus, podocytes appear to be susceptible to PA-induced oxidative damage with inadequate peroxidase activity. LEE AND LEE | 9065 F I G U R E 6 Frozen biopsy samples of control (a) and diabetic nephropathy (DN; b,c) exposed to carboxy-DCFH DA-AM for 2.5 hr. In control glomerulus, there is no DCF fluorescence (a). In DN, there is scattered DCF fluorescence in the glomerulus (b). In advanced DN, particularly, intraglomerular DCF fluorescence is markedly increased, forming patchy aggregates or clumps (c). Carboxy-DCFH DA-AM, carboxy-2′,7′dichlorodihydrofluorescein diacetate diacetoxymethyl ester; DCF, dichlorofluorescein F I G U R E 7 Immunostaining for peroxidases and TGF-β1. In controls, there is no staining for Prx1 (a), Prx2 (c), GPx1 (e), catalase (g), and TGF-β1 (i). In the cases of advanced diabetic nephropathy, there is strong immunostaining for Prx1 (b), Prx2 (d), GPx1 (f), catalase (h), and TGF-β1 (j) in the podocytes mainly overlying the lesions of nodular sclerosis (arrows). Magnification: ×200. GPx, glutathione peroxidase; Prx, peroxiredoxin; TGF-β1, transforming growth factor-β1 Another important finding in this study is that the DCF signal is increased in the glomeruli of human DN, which is particularly severe in the late stages.
Mitochondria-derived superoxide is rapidly dismutated to H 2 O 2 through manganese superoxide dismutase. DCFH is used widely to detect intracellular H 2 O 2 , although it can react with a variety of other cellular oxidants besides H 2 O 2 (Rhee, Chang, Jeong, & Kang, 2010). Thus, an enhanced DCF fluorescence signal in the diabetic glomeruli in this study can represent the increased ROS-H 2 O 2 dismutated from the mitochondrial superoxide. In support of this notion, Galvan et al. (2017) demonstrated an increased mitochondrial ROS generation in the glomeruli of living diabetic mice. In addition, urinary H 2 O 2 level is significantly elevated in diabetic mice (Sharma et al., 2008), suggesting that increased H 2 O 2 in diabetic glomeruli might be an important source of H 2 O 2 in urine.
In the current study, there was strong immunostaining for Prx1, Prx2, GPx1, and catalase in the podocytes of patients with late stages of DN, FSGS, and IgAN. These proteins were also focally expressed in the podocytes of nonsclerotic glomeruli in the moderate stages of nephropathies, suggesting that their expression is not confined to the advanced stages of disease. Furthermore, our findings suggest that the occurrence of peroxidases on renal biopsies at the early stages of nephropathies predicts a progressive nature of the disease with sclerotic glomerular lesions, which are not detected by sampling problems. The increased expression of peroxidase proteins in the podocytes of diabetic kidneys appears to contradict our in vitro study, the mechanisms of which are not clear.
TGF-β is produced as latent complexes. Unlike mesangial cells, podocytes do not secrete TGF-β in response to common in vitro fibrogenic stimuli (S. Chen et al., 2005;Iglesias-de la Cruz et al., 2002). As yet mesangial immunostaining for active TGF-β1 is frequently negative in chronic glomerular disease (J. H. Kim, Kim, Moon, Hong, & Lee, 2003;H. W. Kim, Moon, Park, Hong, & Lee, 2002;Wahab et al., 2005), while podocytes covering the sclerotic segments exhibit increased expression of TGF-β1 protein as shown in the current study. In this regard, H. S. Lee and Song (2009) suggested that TGF-β secreted as latent complexes by mesangial cells is stored in mesangial matrix in chronic glomerular disease, from which soluble forms of latent TGF-β are released and localized to the podocyte surface. Podocyte-derived ROS seem to be involved in TGF-β activation in podocytes. In this study, TGF-β1 immunos-

CONFLICTS OF INTEREST
The authors report no conflicts of interest.