Arginase-2 Mediates Diabetic Renal Injury

Experiments were conducted in male D2.B6-Ins2^Akita/MatbJ and their wild-type littermate mice (Dilute Brown Agouti [DBA]/2J background; The Jackson Laboratory, stock number 007562) starting at 5 weeks of age (2 weeks of diabetes) until 14 weeks of age (11 weeks of diabetes). Ins2^Akita mice, recommended by the Animal Models of Diabetes Complications Consortium as an optimal model of DN (18,19), developed hyperglycemia at 3 weeks of age. Additional experiments were conducted in male 6-week-old DBA/2J mice (The Jackson Laboratory, stock number 000671) and in male arginase-2–deficient (Arg2^−/−) mice on C57BL/6J background (provided by B. Lee, Baylor College of Medicine, Houston, TX) using multiple low doses of vehicle (lactated Ringer’s solution) or STZ (Sigma, St. Louis, MO; 50 mg/kg body wt dissolved in lactated Ringer’s solution) via intraperitoneal injection. All animal studies were approved by the Penn State University College of Medicine Institutional Animal Care and Use Committee.

S-(2-Boronoethyl)-l-cysteine (BEC), a potent arginase inhibitor (2.3 mg/kg/day; Cayman Chemical, Ann Arbor, MI), or vehicle (phosphate-buffered saline) was administered by continuous subcutaneous infusion for 9 weeks (in Ins2^Akita experiments) or for 6 weeks (in STZ-DBA mice experiments) via a mini-osmotic pump (Alzet; Durect Corporation, Palo Alto, CA) as described previously (20). The osmotic pump was incubated in phosphate-buffered saline for 60 h at 37°C before implantation and was implanted dorsally between the shoulders. The condition of mice and body weight were monitored daily after the pump implantation.

Systolic blood pressure was measured using Coda blood pressure (Kent Scientific Corp, Torrington, CT) (21). Mice were allowed to rest quietly for 10 min at 26°C. All measurements were performed at the same time for all groups to prevent any diurnal variations.

Kidneys were fixed in 4% paraformaldehyde and embedded in paraffin, and 3-μm sections were cut. Sections were stained with periodic acid Schiff (PAS) stain, and all glomeruli were examined at 400× in a blinded manner. All images were obtained with an Olympus BX51 microscope and DP71 digital camera using MicroSuite Basic 2.6 imaging software. Images were taken with 100× (oil) objective with a total magnification of 1,000×. Semiquantitative scores (0–4+) were assigned based on the masked reading, as previously described (22). In brief, each glomerulus on a single section was graded from 0 to 4+, where 0 represents no lesion and 1, 2, 3, and 4+ represent mesangial matrix expansion or sclerosis, involving ≤25, 25–50, 50–75, or >75% of the glomerular tuft area, respectively.

Immunohistochemistry for macrophages was performed using rat anti-mouse Mac-2 monoclonal antibody (clone M3/38; Cedarlane, Burlington, NC) on paraffin sections. Sections were incubated with primary antibody (50 μg/mL) followed by a biotinylated goat IgG anti-rat (Vector Laboratories, Burlingame, CA) secondary antibody according to the manufacturer's protocol. Sections were viewed using an Olympus BX51 microscope and DP71 digital camera using MicroSuite Basic 2.6 imaging software. Images were taken with 100× (oil) objective with a total magnification of 1,000×. The number of glomerular macrophages was counted in 20 glomeruli per section (number of macrophages in glomeruli divided by the number of glomeruli) in blinded fashion under 40× magnification and averaged.

We used flow cytometry to analyze kidney macrophage content (CD11b⁺F4/80^low) at the end of the experiments as described previously (23,24). In brief, kidneys were extracted, minced, digested, and then passed through a filter and a cotton wool column. Fresh kidney suspensions were incubated with anti-mouse CD45-fluorescein isothiocyanate (30-F11; eBioscience, San Diego, CA) for 30 min on ice. Kidney macrophages were then identified using allophycocyanin-labeled rat anti-mouse F4/80 (BM8; eBioscience) and phycoerythrin-labeled rat anti-mouse CD11b (M1/70; eBioscience). All samples were treated with anti-mouse CD16/CD32 (2.4G2) to block nonspecific FcR binding and 7-AAD to eliminate dead cells (Invitrogen, Carlsbad, CA). Counting beads (Caltag, Carlsbad, CA) were used to determine the total number of CD45⁺ cells per gram of kidney tissue. Subsequent flow cytometry data acquisition was performed on FASCalibur (Becton Dickinson, San Jose, CA). Data were analyzed by Flowjo software 6.4 (Tree Star, Ashland, OR). All the antibodies were purchased from eBioscience.

Total RNA was extracted from kidneys using RNeasy Mini Kit (Qiagen, GmbH, Hilden, Germany). Single-strand cDNA was synthesized using iScript cDNA Synthesis Kits (Bio-Rad, Hercules, CA) for two-step real-time RT-PCR. Gene-specific primers for arginase-1 and arginase-2 were designed using Beacon Designer Probe/Primer Design Software (Premier Biosoft International, Palo Alto, CA). Specificity of the PCR products was verified by melting curve analysis. Quantitative real-time PCR was performed using Bio-Rad CFX96 system. Reactions were performed in duplicate, and threshold cycle numbers were averaged. Samples were calculated with normalization to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (25).

UAE was measured by ELISA using Albuwell M (Exocell, Philadelphia, PA) as described previously (25). Blood urea nitrogen (BUN) was measured using VITROS DT60II chemistry slides (Ortho-Clinical Diagnostics, Rochester, NY). Body composition was determined using LF90 Minispec Time Domain Nuclear Magnetic Resonance Spectrometer (Bruker Optics, Billerica MA).

Kidney lysates were prepared by homogenization in lysis buffer (50 mmol/L Tris-HCl, pH 7.5, 0.1 mmol/L EDTA, and protease inhibitors) at 4°C, followed by centrifugation for 20 min at 14,000g at 4°C. The supernatants of kidney lysates and plasma samples were assayed for arginase activity as previously described (26).

Plasma (5-μL aliquots) was electrophoresed on SDS-polyacrylamide gels (12% Mini-PROTEAN TGX Precast Gel; Bio-Rad) and electroblotted onto polyvinylidene fluoride membranes (Millipore, Billerica, MA). Arginase-1 was visualized by immunoblotting as previously described (12,13).

A midline abdominal incision was made. Renal medullary blood flow was monitored by using a laser flow meter (model ALF-21D; Transonic Systems, Inc., Ithaca, NY).

Comparisons between groups were examined by using the SPSS version 19.0 software for Windows (SPSS, Chicago, IL) program. Data are expressed as mean ± SEM. One-way ANOVA was used when more than two groups were compared, and significance of observed differences among the groups was evaluated with a least significant difference post hoc test. Statistical significance was identified at P < 0.05.

To assess the possible clinical significance of arginases in diabetic mice, we continuously infused the arginase inhibitor BEC or vehicle into Ins2^Akita mice and their wild-type DBA littermates for 9 weeks, beginning at 5 weeks of age. As shown in Table 1, Ins2^Akita vehicle-treated mice had increased blood glucose levels, decreased body weight, increased kidney weight, increased kidney weight/body weight ratio, reduced fat and fluid composition, and increased lean composition compared with control mice. Arginase inhibition of Ins2^Akita mice significantly reduced kidney weight and kidney weight/body weight ratio without affecting other measurements. It is noteworthy that treatment with BEC did not reduce blood glucose levels or blood pressure.

To determine whether arginases contribute to diabetic renal injury, we measured 24-h UAE and BUN as indicators of renal injury in Ins2^Akita with and without BEC treatment. Vehicle-treated Ins2^Akita mice had a significant increase in albuminuria compared with controls at 5 and 14 weeks of age (Fig. 1). Albuminuria was significantly reduced in Ins2^Akita mice treated with BEC at 14 weeks of age. Similarly, BUN was significantly increased in vehicle-treated Ins2^Akita mice compared with other groups (Table 1).

We assessed whether both arginases are expressed in the kidneys under normal and diabetes conditions. Arginase-2, but not arginase-1, mRNA was detectable in the kidney (Fig. 2A). This result supports the view that the effect of BEC in the kidney is mainly mediated via inhibition of arginase-2. Both vehicle-treated and BEC-treated Ins2^Akita mice had increased arginase-2 mRNA levels (Fig. 2B). Thus, diabetes is associated with increased kidney arginase-2 expression. Although BEC inhibits arginase activity, it did not downregulate arginase-2 mRNA in diabetes.

To determine whether arginase is critical for macrophage infiltration in DN, we show the distribution and quantitation of macrophages in kidneys by immunohistochemistry (Mac-2–positive macrophages) (Fig. 3A). Vehicle-treated Ins2^Akita mice showed significant increases in glomerular macrophages (2.95 ± 0.14 macrophages/glomerulus; P < 0.0001) compared with vehicle-treated control (0.34 ± 0.04 macrophages/glomerulus). Both BEC-treated control and Ins2^Akita mice had significantly reduced glomerular macrophage recruitment (0.29 ± 0.03 and 0.9 ± 0.14 macrophages/glomerulus; P < 0.0001), respectively, compared with vehicle-treated Ins2^Akita mice at 14 weeks of age.

Similar results were obtained when kidney macrophages (identified as CD11b⁺F4/80^low) were detected by fluorescence-activated cell sorter (Fig. 3B). Kidneys of vehicle-treated Ins2^Akita mice had a significantly greater number of macrophages compared with vehicle-treated control (3.6 ± 0.1 vs. 2.1 ± 0.3 macrophages/gram kidney tissue; P < 0.005). In contrast, kidneys of BEC-treated Ins2^Akita mice had significantly reduced numbers of macrophages (1.6 ± 0.3 macrophages/gram kidney tissue; P < 0.0001) compared with vehicle-treated Ins2^Akita mice at 14 weeks of age.

To determine whether the effect of arginase inhibition is specific to Ins2^Akita mice, we used another type 1 diabetic mouse model (DBA/2J) using STZ to induce diabetes for a 6-week study period. We measured 24-h UAE as indicators of renal injury with and without BEC treatment. Vehicle-treated, STZ-induced diabetic mice had a significant increase in albuminuria compared with controls after 6 weeks, an effect that was significantly reduced with BEC treatment in diabetic mice despite comparable blood glucose levels (Fig. 4). We also show an increase in kidney macrophage recruitment using Mac-2 staining in vehicle-treated, STZ-induced diabetic mice (1.9 ± 0.1 macrophages/glomerulus; P < 0.0001) compared with control (0.6 ± 0.05 macrophages/glomerulus). In contrast, BEC-treated, STZ-diabetic mice had a significant reduction in glomerular macrophage recruitment (1.2 ± 0.06 macrophages/glomerulus; P < 0.0001) compared with vehicle-treated STZ-diabetic mice.

PAS staining of kidney sections showed increased glomerular cellularity and mesangial expansion (score, 1.9 ± 0.05 vs. 1.1 ± 0.09; P < 0.005) after 6 weeks of diabetes in vehicle-treated, STZ-diabetic mice versus controls (Fig. 5A and B). Inhibition of arginases in STZ-induced diabetic mice exhibited significantly reduced glomerular changes (scores, 1.6 ± 0.08; P < 0.05) compared with vehicle-treated, STZ-induced mice (Fig. 5C). As expected for a period of induced diabetes of only 6 weeks, there was little or no indication of fibrosis by Masson’s trichrome or the periodic acid methenamine silver (aka Jones stain) staining, which was not detectably different between any of the groups (data not shown).

We next assessed whether a genetic lack of arginase-2 will mimic the changes observed in Ins2^Akita mice after pharmacologic arginase inhibition. Wild-type and Arg2^−/− mice were treated with STZ and killed at either 6 or 18 weeks after STZ-induced diabetes. Wild-type mice displayed a significant increase in UAE after 6 weeks and after 18 weeks of STZ-induced diabetes (Fig. 6). In contrast, the increase in UAE was significantly reduced in Arg2^−/− mice in a similar manner to BEC treatment in Ins2^Akita. It is noteworthy that arginase-2 deficiency did not alter hyperglycemia in the diabetic mice (Fig. 6).

Diabetes was associated with a significant increase in kidney arginase activity at 6 and 18 weeks in wild-type mice after STZ-induced diabetes (Fig. 7A). These data were confirmed in vehicle-treated Ins2^Akita mice after 9 weeks of diabetes, with increased kidney arginase activity versus control mice (22 ± 1 vs. 12 ± 1 nmol/min/mg). In contrast, Arg2^−/− mice had minimal kidney arginase activity at all time points, confirming previous studies showing that overall arginase activity in the kidney is due almost entirely to arginase-2 (27). The small amount of kidney arginase activity in Arg2^−/− mice under normal and diabetes conditions may be due to kidney-resident leukocytes that express arginase-1 or to trace arginase-1 contamination from plasma.

Plasma arginase activity, which is due almost exclusively to arginase-1 (28), was increased in both wild-type and Arg2^−/− mice after 18 weeks of STZ-induced diabetes and correlated well with increased arginase-1 protein by Western blot (Fig. 7B). These data suggest a possible minor contribution of elevated plasma arginase-1 to the development and progression of DN.

Because arginases and NOS can compete for l-arginine, we assessed whether changes in levels of arginase-2 would correlate with changes in renal blood flow. Whereas increased kidney arginase activity in diabetic mice (Fig. 7A) was associated with reduced renal medullary blood flow (Fig. 7C), medullary blood flow was not reduced in Arg2^−/− mice, suggesting that renal arginase-2 modulates NO production by catabolizing arginine and demonstrating that elevated plasma arginase-1 had no impact on medullary blood flow.

Arginases have a well-established role to alter endothelial function in cardiovascular diseases, yet their role in diabetic kidney injury has not previously been determined. This study shows that pharmacological blockade or genetic deficiency of arginase-2 mediates renal tissue protection as proven by a reduction in albuminuria, BUN, histopathological changes, and kidney macrophage recruitment during diabetes. We speculate that the effect of arginase inhibition in the kidney is mainly mediated via inhibition of arginase-2 because arginase-1 expression was not detectable in the kidney. These findings reveal an important role for arginase-2 in the pathogenesis of DN and provide evidence for arginase inhibition as potential therapeutic modality for treating diabetic patients.

To examine the role of arginases in DN, we used two approaches (use of a potent arginase inhibitor and use of Arg2^−/− mice) in two different models of type 1 diabetes: Ins2^Akita diabetic mice and STZ-induced diabetic mice. We investigated the effects of inhibiting arginases on kidney dysfunction, glomerular histopathological changes, and macrophage recruitment using BEC in Ins2^Akita mice and the STZ-induced diabetic model. BEC significantly ameliorated diabetic albuminuria and was associated with reduced kidney macrophage infiltration and glomerular pathology. Similar to the BEC treatment, lack of arginase-2 expression ameliorated renal function by normalizing UAE in diabetes, even though these mice had elevated blood glucose levels comparable to diabetic wild-type mice. Taken together, our results provide support for blocking arginase as a therapeutic modality for the treatment of DN.

The expression of arginase-2 rather than arginase-1 in the kidney suggests that BEC effects are largely, but perhaps not entirely, mediated via inhibition of arginase-2. Because BEC inhibits both arginase-1 and -2, it is possible that arginase-1 also may contribute somewhat to DN. Currently, there are no isoform-specific arginase inhibitors. The fact that plasma arginase activity and arginase-1 protein were significantly increased in both diabetic wild-type and Arg2^−/− mice suggests a possible contribution of plasma arginase-1 to reductions in arginine bioavailability and to the development and progression of DN. Additional experiments are needed to explore this possibility.

The renal protective effect of BEC correlates with a significant reduction of kidney macrophage infiltration. Whether the reduction in macrophage recruitment is mediated directly by arginase inhibition or indirectly by reducing diabetic renal injury is not clear at this time. Additional studies are required to elucidate the role of arginases on kidney macrophage infiltration.

Arginase inhibition resulted in less mesangial expansion and glomerular hypercellularity in diabetes, indicating a possible contribution of arginases to the initiation and/or progression of diabetic renal fibrosis. Previous immunohistochemical studies have identified arginase-2 expression in proximal straight tubules and inner medullary collecting duct (29,30). The expression of arginase-2 in glomerular cells, such as podocytes, and/or glomerular endothelial cells will be important to investigate.

Alteration in endothelial function is a common underlying event for vascular abnormalities observed in patients with diabetes, insulin resistance, glucose intolerance, hypertension, and peripheral vascular disease. Endothelial dysfunction, characterized by reduced bioavailability of NO and increased oxidative stress, is a hallmark of diabetes (31) and DN (32). NO is produced from arginine by NOS. Under conditions of low arginine level or hyperglycemia, endothelial NOS can become uncoupled, producing reactive oxygen species in lieu of NO (33,34). Recently, low or lack of endothelial NOS has been shown to exacerbate DN (22,35); thus, elucidating the basis for vascular dysfunction in DN is critical (36). Arginase-2 is constitutively expressed and also inducible in endothelial cells (4–6) as well as in kidney cells (29,30). When elevated, arginase-2 can inhibit NOS activity and expression and thus induce endothelial NOS uncoupling, thereby reducing NO bioavailability and inhibiting the NO/cGMP pathway as well as increasing oxidative stress. Previous studies have shown that vascular dysfunction in rodent models of diabetes (8,10,11) or diabetic patients (9) is associated with increased arginase-1 (9,11) or arginase-2 (8) expression and that vascular function is normalized by inhibition of arginase activity or genetic ablation of arginase-2 expression (8,10). The current study is the first to evaluate a possible role for arginase-2 in vascular dysfunction specifically in the kidney. Our data showing reduced renal medullary blood flow in diabetic wild-type mice and its restoration in arginase-2–deficient mice are consistent with these findings, indicating a possible role of arginase-2 in regulating NO bioavailability in the kidney.

In conclusion, our study demonstrates for the first time that arginase-2 plays an essential role in the development of DN. This conclusion is based on two novel observations. First, we demonstrated a beneficial effect of arginase inhibition in animal models of DN. Second, we showed that deficiency specifically of arginase-2 ameliorates diabetic renal injury. Results of this study may ultimately result in novel therapeutic interventions designed to attenuate arginase activity or signaling that regulates arginase-2 expression in the treatment of diabetic kidney disease

This study was supported by National Institutes of Health (NIH) DK-077444 (to A.S.A.) and NIH GM-57384 (to S.M.M.). S.M.M. has received honoraria from Corridor Pharmaceuticals for advisory board participation. No other potential conflicts of interest relevant to this article were reported.

S.M.M. wrote the manuscript, researched data, and contributed to discussion. T.G. researched data. T.K.C. researched data and reviewed the manuscript. D.K.-L. researched data. A.S.A. wrote the manuscript, researched data, and developed the idea.

The authors thank Dr. Brendan Lee for providing the Arg2^−/− mice used in this study and Drs. W. Brian Reeves and Kent E. Vrana (Penn State Hershey Medical Center, Hershey, PA) for helpful discussions.