Derivative of Bardoxolone Methyl, dh404, in an Inverse Dose-Dependent Manner Lessens Diabetes-Associated Atherosclerosis and Improves Diabetic Kidney Disease

Diabetes is a common risk factor for both chronic kidney disease and atherosclerosis (1,2). Despite the use of conventional therapies that include blood pressure, glucose-lowering, and hyperlipidemic treatments, significant numbers of diabetic patients suffer morbidity or mortality as a consequence of cardiovascular complications and/or progress to end-stage renal failure. Novel interventions are needed to satisfy this unmet clinical need to limit multiple diabetes-associated complications.

Evidence strongly supports a role for oxidative stress and inflammation as underlying pathogenic mechanisms of both diabetic renal and atherogenic complications. With oxidative stress and inflammation now recognized to be inextricably linked (3), approaches that limit oxidative stress are likely to translate to reduced inflammation. Limiting oxidative stress by bolstering antioxidant defenses is a novel alternative approach to antioxidant therapy that may yield more efficacious outcomes than standard vitamin therapy (4). One approach that has attracted significant attention is the augmentation of antioxidant defenses via an upregulation of the NFE2-related factor 2 (Nrf2)/Kelch-like ECH-associated protein 1 (Keap1) pathway (5–9).

Numerous small-molecule activators of the Nrf2/Keap1 pathway have been identified (10,11), including bardoxolone methyl (BM), also known as CDDO-methyl, a potent synthetic triterpenoid inducer of the phase 2 response (12,13). BM and other related antioxidant inflammation modulators (AIMs) such as CDDO-imidazole (RTA 403) have demonstrated structural and functional improvements in several rodent models of renal disease (14–16), along with an increase in protective Nrf2, peroxisome proliferator–activated receptor γ, and heme oxygenase-1 (HO-1) genes (16). Given the known negative effects of oxidants and inflammation on the vasculature and the promising antioxidant, anti-inflammatory and tissue-protective effects of AIM compounds (13), we hypothesized that Nrf2 agonists may additionally protect the diabetic vasculature against accelerated atherosclerosis. Our hypothesis is strengthened by recent data showing an adaptive induction of Nrf2-driven antioxidant genes in response to hyperglycemia in human endothelial cells (17) and the use of the Nrf2 activator sulforaphane to induce Nrf2 activation of downstream target genes in endothelial cells (18).

Interest in BM was further enhanced by two recent phase 2 clinical trials in which this compound showed a dose-dependent increase in the estimated glomerular filtration rate (eGFR) initially after 56 days of treatment (19) and in a second study (BEAM) (20) after 24 weeks of treatment in patients with advanced chronic kidney disease and type 2 diabetes. In the latter study, the positive effects of drug treatment persisted at 52 weeks after removal of the drug at 24 weeks. However, issues of drug toxicity and off-target effects of this drug class have been raised, particularly after the larger phase 3 clinical trial (BEACON) with primary end points of time–to–first occurrence of either end-stage renal disease or cardiovascular death (21) was terminated prematurely due to unexplained adverse cardiovascular events. Recently, the results of BEACON have been published and suggest a potential effect of BM on heart failure, further emphasizing that this drug class needs more extensive investigations with respect to not only renal but also cardiovascular end points (22). Potential problems with toxicity and a worsening of diabetic nephropathy were also raised in a preclinical study examining the effects of BM analogs RTA 405 and dh404 on renal outcomes in diabetic Zucker fat rats (23). The issue of toxicity was subsequently addressed in a follow-up study (24) that showed that RTA 405 and dh404 are well tolerated and did not confirm the previous report of hepatic and renal toxicity.

To further clarify the effects of Nrf2 activation on diabetic renal injury and investigate its potential role in lessening diabetes-associated atherosclerosis, in this study, we treated diabetic apolipoprotein E–knockout (ApoE^−/−) mice with the BM analog dh404 for 18 weeks and assessed parameters associated with renal structural and functional injury as well as atherosclerosis. As rodents metabolize BM to a metabolite that itself is not toxic to higher mammals but is quite toxic to rodents, BM itself cannot be used for chronic rodent studies. Therefore, tool compounds, including dh404 (also known as dh-CDDO-trifluoroethyl amide; see Supplementary Fig. 1), are used to probe rodent pharmacology for this class and inform on the likely clinical behavior of BM. Indeed, BM and dh404 exhibit very similar qualitative biological properties (25) and the same mechanism of action, in which dh404 has been shown to activate Nrf2 by preventing its degradation as a result of interrupting the ability of Keap1 to bind to Nrf2. Therefore, dh404 is an appropriate surrogate for studying this class of compound. We now present evidence, for the first time, that dh404 lowers oxidative stress and proinflammatory mediators and protects against diabetes-associated atherosclerosis in a dose-dependent manner. In addition, dh404 improves renal structural and functional injury while showing no detrimental effect on liver function in this diabetic model.

dh404, solubilized in sesame oil (SO; Spectrum Chemical) at 3, 10, or 20 mg/kg body weight, was administered to mice once per day by oral gavage.

Eight-week-old male C57Bl/J6 ApoE^−/− mice purchased from the Animal Resource Centre (Perth, Australia) were rendered diabetic by two intraperitoneal injections of streptozotocin (STZ; Sigma-Aldrich, St. Louis, MO) at 100 mg/kg/day on 2 consecutive days (26–28).

In a first cohort, 10-week-old diabetic mice were gavaged with either SO or 10 or 20 mg/kg dh404. Additionally, a group of nondiabetic mice received 20 mg/kg dh404. Aortas and kidneys were collected for quantitative RT-PCR (qRT-PCR) after 5 weeks of gavage. A second cohort of diabetic mice was gavaged with SO or 3, 10, or 20 mg/kg dh404 for 18 weeks. Additionally, sham-injected nondiabetic mice were gavaged with 20 mg/kg dh404. Weekly blood glucose, body weight, and blood pressure measurements [using noninvasive tail-cuff plethysmography (29)] were performed. Three days prior to termination, urine was collected in 24-h metabolic cages. Mice were killed by lethal injection of 2,2,2-tribromoethanol (Sigma-Aldrich, St. Louis, MO) and direct puncture of the right ventricle to obtain blood. Plasma was stored at −80°C. Plasma and urine were analyzed for plasma alanine transaminase (ALT), aspartate transaminase (AST), and urine creatinine (Australian Specialized Animal Pathology Laboratory, Mulgrave, Victoria, Australia). Aortas, kidney, and livers were fixed in 10% neutral buffered formalin. Kidney cortex was snap frozen in liquid nitrogen for RNA extraction.

En face analysis of lesions was conducted after staining with Sudan IV Herxheimer’s solution (26–28). Plaque area was calculated as the proportion of aortic intimal surface area occupied by red-stained plaque (Adobe Photoshop v6.0.1; Adobe Systems, New South Wales, Australia).

Lesions within the sinus were visualized after staining with Oil Red O and quantitated as described previously (26–28).

Urinary albumin-to-creatinine ratio was measured using a mouse albumin ELISA kit (Bethyl Laboratories, Montgomery, TX). Plasma cystatin C was measured using a mouse cystatin C ELISA kit (30) (BioVendor Laboratory Medicine, Brno, Czech Republic).

Urinary 8-isoprostane and 8-hydroxy-2-deoxyguanosine (8-OHdG) were assessed using commercially available EIA kits (Cayman Chemical, Ann Arbor, MI, and StressMarq Biosciences Inc, Victoria, BC, Canada, respectively). Measures were expressed relative to urinary creatinine. Plasma was analyzed for diamicron reactive oxygen metabolites (dROMs), particularly hydroperoxides, using a Free Radical Analytical System-4 (H&D srl, Parma, Italy) and expressed in Carratelli units (31).

Kidney sections were stained with periodic acid-Schiff (PAS). Mesangial area was quantitated using Image-Pro Plus v6.0 (Media Cybernetics, Rockville, MD) and expressed as percentage of PAS-stained area per glomerular cross-sectional area. Tubulointerstitial area was assessed using a point-counting technique (32).

The degree of sclerosis in each glomerulus was graded on a scale of 0–4 (33).

Livers were stained with hematoxylin and eosin and scored blinded on a scale of 0–5 to determine hepatocellular cytoplasmic rarification and periportal inflammation (24). Five to eight livers were examined per group. Ten sections were scored and averaged to obtain an overall score for each liver.

Four-micrometer aorta paraffin sections were stained for nitrotyrosine, 4-hydroxynonenal (4-HNE), and vascular cell adhesion molecule-1 (VCAM-1). Four-micrometer kidney sections were stained for collagen IV and nitrotyrosine (27,28).

Sections were examined under light microscopy (Olympus BX-50; Olympus Optical) and digitized with a high-resolution camera. Digital quantifications (Image Pro Plus v6.0; Media Cybernetics, Bethesda, MD) were performed in a blinded manner.

RNA from kidney and aorta was extracted using TRIzol Reagent (Invitrogen) and cDNA synthesized as described previously (27,28). Gene expression was determined after real-time quantitative RT-PCR and analyzed as described previously (27,28). Gene expression was normalized relative to 18S ribosomal RNA. Primers and probes are shown in Supplementary Table 1.

Tubular NRK52E cells (American Type Culture Collection, Rockville, MD) were maintained in DMEM containing 10% serum and 25 mmol/L glucose. Cells were serum starved and pretreated with dh404 at 0.25, 0.5, or 0.75 µmol or vehicle (DMSO) for 4 h before stimulation with 10 ng/mL transforming growth factor-β1 (TGF-β1) for 72 h (34). Treatment at ≥1 µmol dh404 for 48 h resulted in significant cell death (Supplementary Fig. 2). Cells were collected for qRT-PCR or Western blot analysis (28).

Data are expressed as mean ± SEM and analyzed by one-way ANOVA with Newman-Keuls post hoc testing. Statistical analyses were performed using GraphPad Prism version 6.0 (GraphPad, La Jolla, CA). A P value <0.05 was considered statistically significant.

dh404 had no effect on systolic blood pressure in diabetic mice (control mice 129 ± 2 vs. diabetic mice 129 ± 1, 130 ± 1, and 134 ± 1 mmHg for 3, 10, and 20 mg/kg dh404, respectively). Body weights of STZ diabetic mice were significantly lower than nondiabetic controls (Table 1). However, diabetic mice treated with lower doses of dh404 (3 and 10 mg/kg/day) exhibited higher body weights compared with both vehicle-treated diabetic mice and diabetic mice treated with the highest dose of 20 mg/kg dh404. Vehicle-treated diabetic mice also displayed significantly higher kidney–, liver–, and lung–to–body weight ratios compared with nondiabetic mice. Treatment of mice with 3 and 10 mg/kg/day dh404 attenuated the increase in kidney–to–body weight ratio associated with diabetes (Table 1). Heart–to–body weight ratios were unaffected by diabetes and dh404 treatment. All diabetic mice displayed increased food and water intake with significantly increased urine output regardless of treatment (Table 2). Diabetic mice had significant increases in HbA_1c and blood glucose levels compared with nondiabetic mice regardless of treatment (Table 2). Diabetes was associated with increases in total cholesterol, triglycerides, and HDL and LDL levels. All doses of dh404 reduced these parameters, although not to levels seen in the control mice, with the exception of the 20 mg/kg/day treatment, in which triglycerides and LDLs were not affected by dh404.

Diabetes induced a significant increase in total plaque in ApoE^−/− mice, and this increase was observed in the arch, thoracic, and abdominal regions of the aorta. Treatment of diabetic mice with dh404 at 3 and 10 mg/kg/day significantly reduced plaque within these regions. However, treatment with 20 mg/kg/day dh404 failed to reduce atherosclerotic plaque (Fig. 1A–E).

Diabetes significantly increased lesion deposition within the aortic sinus after 20 weeks of diabetes (Fig. 1F and G). This increase was significantly attenuated after 18 weeks of dh404 treatment at 3 and 10 mg/kg/day. Treatment with 20 mg/kg/day dh404 failed to reduce diabetes-associated lesions.

Gene expression of the inflammatory mediators tumor necrosis factor-α (TNF-α) and monocyte chemotactic protein-1 (MCP-1) was significantly upregulated in diabetic aorta. This was significantly attenuated after 5 weeks of dh404 treatment at 10 mg/kg/day but not at 20 mg/kg/day (Fig. 2A and B). No significant differences were observed in CD36, intracellular adhesion molecule-1 (ICAM-1), p65, and VCAM-1 gene expression between vehicle-treated diabetic and nondiabetic mice (Fig. 2C–F). However, 10 mg/kg/day of dh404 reduced the expression of ICAM-1 and VCAM-1 when compared with vehicle-treated counterparts. VCAM-1 protein levels were significantly increased in diabetic plaque (P < 0.05) and showed a trend toward a reduction after 10 and 20 mg/kg/day dh404 treatment (Fig. 2G and H).

Analysis of 4-HNE, a marker of lipid hydroperoxides, within the vessel walls (Supplementary Fig. 3) and plaques (Supplementary Fig. 4) revealed a trend toward an increase in diabetic mice. This was reduced after treatment with 3 mg/kg/day dh404 in the vessel wall, although this did not reach statistical significance. At the highest dose of dh404, this trend was lost in vessel walls of the diabetic cohort. Plaque data showed a trend toward a reduction in 4-HNE staining in diabetic mice treated with 10 and 20 mg/kg/day dh404.

Analysis of nitrotyrosine, a marker of protein oxidative/nitrosative stress, within plaque (Supplementary Fig. 5) revealed a trend toward an increase in the diabetic cohort, as observed previously by us in diabetic mice (27). This trend was lessened after treatment with dh404, with 20 mg/kg/day showing the greatest reduction in nitrotyrosine staining, albeit that this did not reach statistical significance.

Urinary 8-isoprostane, 8-OHdG, and plasma dROMs were all significantly upregulated in vehicle-treated diabetic mice. dh404 significantly attenuated these measures at all doses tested (Fig. 3A–C). Nitrotyrosine levels were assessed in the tubulointerstitial region of the kidney, and the diabetes-associated increase was significantly decreased at all concentrations of dh404 (Fig. 3D and E; P < 0.001 vs. nondiabetic control subjects).

The expression levels of Nrf2-responsive antioxidant genes in kidney cortex, after 5 weeks of dh404 treatment, are shown in Fig. 4A. NAD(P)H quinone oxidoreductase 1 (NQO1) was upregulated by diabetes and after treatment with 10 and 20 mg/kg/day dh404. Similarly, expression of glutathione-S (GSH-S) transferase was increased by diabetes but further elevated after dh404 treatment. Glutathione peroxidase (GPx) 1 expression increased in response to dh404 treatment, reaching significance at 20 mg/kg/day dh404 in diabetic kidneys.

Inflammatory gene expression is shown in Fig. 4B. Despite no change in interleukin-6 (IL-6) expression, MCP-1 gene expression was increased by diabetes and further increased after 10 and 20 mg/kg/day dh404, with significance reached at the highest dose of dh404. The p65 subunit of nuclear factor-κB (NF-κB) increased after dh404 treatment, reaching significance at 20 mg/kg/day dh404.

Induction of diabetes led to a significant increase in the urinary albumin-to-creatinine ratio, which was significantly reduced by all doses of dh404 (Fig. 4C). Elevations in plasma cystatin C are an indication of declining glomerular filtration (30). Plasma cystatin C levels were significantly lower in vehicle-treated diabetic mice compared with nondiabetic controls (Fig. 4D), consistent with renal hyperfiltering (35). Treatment with dh404 did not alter plasma cystatin C levels, implying that dh404 did not worsen renal function at the different doses.

The percentage of PAS-positive material, indicative of mesangial expansion, was significantly increased in diabetic glomeruli (Fig. 4E and F). Furthermore, the glomerulosclerosis index (GSI) of diabetic glomeruli was significantly higher than that of nondiabetic controls (Fig. 4G). Both mesangial expansion and GSI were attenuated in diabetic mice treated with dh404 at 3 mg/kg/day, but this was not observed at other doses of dh404.

Tubulointerstitial area assessment revealed a significant increase in diabetic kidneys that was significantly attenuated at all three doses of dh404 (Fig. 4H), with the most significant reduction observed at 3 mg/kg/day. Staining for collagen IV showed a trend toward a reduction after 3 mg/kg dh404 when compared with diabetic kidneys treated with vehicle only (Supplementary Fig. 6).

Liver function was not affected by dh404, as plasma levels of ALT and AST were not significantly different among treatment groups (Table 2). Histopathological assessment of hepatocellular cytoplasmic rarification, an indicator of accumulated glycogen, showed no significant differences after treatment with dh404, albeit that a slight increase was observed in the nondiabetic (20 mg/kg dh404) as well as diabetic groups receiving 3 and 10 mg/kg dh404 (Supplementary Table 1). Diabetes caused a slight albeit nonsignificant increase in the number of infiltrating inflammatory cells. The 10-mg/kg dose of dh404 caused a statistically significant increase in infiltrating inflammatory cells. However, this was considered mild, with an infiltrate of <50 cells per portal vein. Thus, no significant treatment-related adverse liver pathology was detected after 18 weeks of dh404 treatment (Supplementary Fig. 7).

To complement our in vivo findings, we explored in vitro, in normal rat kidney (NRK) cells in a dose-dependent manner, the potential impact of dh404 on Nrf2-dependent antioxidant genes NQO1, HO-1, and GSH-S transferase (36). Gene expression of these antioxidants was significantly upregulated by dh404, in the presence or absence of TGF-β, with the greatest increase noted at the highest dose of dh404 (Fig. 5A). GPx1 gene expression was increased significantly by dh404 in the presence of TGF-β. dh404 increased GPx2 expression significantly in the absence of TGF-β, and in the presence of TGF-β, this occurred at the highest dose of dh404. GPx3 expression was increased significantly only at the highest dose of dh404 in the presence or absence of TGF-β (Fig. 5B). Changes in gene expression translated into changes at the protein level, as shown for HO-1 and GPx1 after 0.5-µmol dh404 treatment (Fig. 5B).

Despite positive reductions in IL-6 gene expression with increasing doses of dh404 (Fig. 6A), MCP-1 expression levels increased by ∼20-fold at the highest dose of dh404. p65 gene expression was unaffected by dh404 in TGF-β–untreated cells but showed a slight trend toward an increase at the highest dose of dh404 compared with TGF-β–treated controls (Fig. 6A).

Treatment with TGF-β induced a significant increase in fibronectin and collagen I and IV gene expression (Fig. 6B). Importantly, dh404 significantly attenuated the gene expression of these profibrotic proteins, which was mirrored by a similar protein profile for fibronectin. (Fig. 6B and C).

In the current study, the BM derivative dh404, an Nrf2 agonist, has shown beneficial effects in attenuating diabetes-associated atherosclerosis and nephropathy, albeit in an inverse dose-related manner. To date, however, the role of Nrf2 in atherosclerosis remains controversial. Studies have reported that Nrf2 may be proatherogenic using ApoE/Nrf2 double-knockout mice in which plaque and expression of the macrophage scavenger receptor CD36 were significantly reduced (37–39). In contrast, the absence of Nrf2 within the myeloid lineage in LDL receptor^−/− mice aggravated both early and late stages of atherosclerosis (40,41). Prior to the current study, the role of Nrf2 modulation in diabetes-associated atherosclerosis remained an unexplored area.

The results of this study show for the first time that dh404 lessens the progression of diabetes-associated atherosclerosis. This was observed in all regions of the aorta but appeared to be linked to the dose, with the lower doses of 3 and 10 mg/kg significantly lessening the progression of diabetic lesions, while the highest dose (20 mg/kg) failed to offer vascular protection. This was paralleled by similar reductions in proinflammatory gene expression (TNF-α, MCP-1, ICAM-1, and VCAM-1) at lower doses of dh404, while the 20-mg/kg dose failed to afford such protection. Interestingly, markers of oxidative stress such as urinary 8-isoprostane, 8-OHdG, and plasma dROMs were significantly attenuated by all concentrations of dh404. Similarly, oxidative damage to proteins within plaque, as assessed by nitrotyrosine staining, was reduced particularly at the highest dose of 20 mg/kg dh404, suggesting that at this dose, there is no correlation between reductions in oxidative stress and diabetes-associated atherosclerosis.

However, at lower doses, our results suggest that dh404 mediates its diabetes-associated antiatherogenic effects via its anti-inflammatory and antioxidative actions. Triterpenoids have also been shown to regulate lipid metabolism (8,27). Thus, it is possible that the observed reductions are due in part to the action of dh404 on lipids since we observed reductions in LDL, triglyceride and cholesterol after dh404 treatment, albeit that these reductions were rather modest. Additionally, the possible minor improvement in metabolic control of diabetes by dh404 needs to be taken into consideration. However, these rather modest changes are unlikely to have contributed to the overall improvement in atherosclerosis. Furthermore, the lack of an effect on atherosclerosis at the highest dose of dh404 could represent off-target effects of dh404 that may override any potential benefits of improved metabolic parameters in this model.

In addition to these positive effects on the diabetic macrovasculature, we also show significant improvements in renal function after 18 weeks of dh404 treatment. This was reflected by improvements in the urinary albumin-to-creatinine ratio, with the lower doses of dh404 showing the greatest inhibition. It should be noted that the finding of a reduction in urinary albumin-to-creatinine ratio contrasts with the human data of the BEACON trial (22), in which, despite improvements in eGFR, the urinary albumin-to-creatinine ratio significantly increased in the BM-treated group. A number of possibilities could explain these discordant results, such as a species-specific effect, but more likely they relate to the timing of treatment. Indeed, human subjects had advanced kidney disease (stage 4) prior to exposure to BM, while in this study, mice were treated soon after the establishment of diabetes as a preventative strategy. One of the most prominent effects of BM in the BEAM (20) and BEACON (22) trials included an increase in eGFR in subjects with renal impairment. Unfortunately, most animal models of diabetic nephropathy, including the STZ ApoE^−/− mouse, do not have declining GFR but rather are models of hyperfiltration (35). Indeed, in this study, analysis of plasma cystatin C showed that diabetic mouse kidneys were hyperfiltering, a known early phenomenon of kidney disease (30). Thus, this model is unable to specifically test the ability of Nrf2 agents such as dh404 to increase GFR.

Our structural analyses within the diabetic kidney showed significant reductions in tubulointerstitial injury after treatment with dh404, with the lowest dose providing the greatest protection. Similarly, only the lowest dose of dh404 lessened the expression of the fibrosis marker collagen IV within this region of the kidney, albeit that these reductions fell just short of significance. This was also accompanied by significant reductions in oxidative stress within the tubular region of the kidney after dh404 treatment. Additionally, we also observed significant improvements in glomerulosclerosis at the lowest dose of dh404. Our results therefore contrast with the study of Zoja et al. (23), in which it was reported that dh404 failed to show beneficial effects on proteinuria and significantly worsened tubular damage in the diabetic Zucker fat rat. Our study also contrasts with their study in that we did not observe any adverse effects of drug treatment on liver structure and function and failed to detect any renal pseudotumors. Differences in outcome between our study and that of Zoja et al. (23) may reflect subtle disparity in drug dosage as well as differences in drug bioavailability between species. Our findings are, however, more in agreement with the study by Chin et al. (24), in which it was shown that the BM analogs RTA 405 and dh404 were well tolerated in various rodent models of type 2 diabetes. Our study is also in agreement with the notion that activators of Nrf2 protect against diabetic nephropathy (9).

It is of interest that in the current study, the lower doses of dh404 appear to limit end-organ injury, whereas this benefit is lost at higher dosages. Therefore, it is critical to determine whether the higher dose of drug did not work in vivo due to a lack of an increase in antioxidant production or due to an off-target effect. To address this issue, we analyzed several Nrf2-responsive antioxidant genes (NQO1, GSH-S transferase, and GPx1) in kidney cortex. Indeed, the gene expression of NQO1 and GSH-S transferase are further increased in response to 20 mg/kg dh404. In addition, cell-culture experiments confirmed the responsiveness of Nrf2 genes to dh404 in a dose-dependent manner. Taken together, our results indicate that a lack of end-organ protection by the highest dose of dh404 cannot be explained by a lack of antioxidant responsiveness to dh404.

Possible off-target effects of the drug that may account for the loss of protection at higher doses were assessed. Indeed, proinflammatory mediators such as MCP-1 were significantly upregulated in a dose-responsive manner, with the highest dose of dh404 resulting in an ∼20-fold increase in MCP-1 gene expression in NRK cells. Interestingly, the highest dose of dh404 also caused a significant increase in MCP-1 expression in the kidneys of dh404-treated mice. The latter finding was also accompanied by a significant increase in the expression of the p65 subunit of the proinflammatory transcription factor NF-κB in the high dose–treated diabetic kidney. Similarly, higher aortic MCP-1 expression correlated with greater plaque. Thus, our data suggest that dh404 might act to promote inflammation at higher doses.

Our in vitro results suggest that dh404 regulates its antifibrotic effects in the kidney, at least partly, through the inhibition of TGF-β activity, since both TGF-β–stimulated fibronectin and collagen I and IV expression were inhibited by dh404 in rat kidney tubular cells. Our results are therefore consistent with previous studies that demonstrate Nrf2-mediated renal protection through the inhibition of TGF-β promoter activity and its downstream pathways (9,42).

Finally, these findings need to be considered in the clinical context. The systemic exposure of dh404 at the lower doses in this study, corrected for the potency difference between dh404 and BM, was similar to that observed in both the BEAM and BEACON studies. Therefore, protection from murine end-organ injury occurs at clinically relevant exposures, whereas this beneficial effect was lost at higher doses. Thus, as seen in the Nrf2^−/− studies (37–39), there may be a therapeutic window for this drug class, and detailed preclinical and subsequently clinical studies with such agents will need to be tested in a careful dose-dependent manner.

In summary, our data have shown inverse dose-dependent improvements in diabetes-associated atherosclerosis and diabetic nephropathy through the use of an Nrf2 activator, dh404, via reductions in proinflammatory mediators and oxidative stress in our preclinical model of STZ-induced diabetes. The antiatherogenic function of AIM compounds, and dh404 in particular, has not previously been explored. Our study therefore potentially extends the pharmacotherapeutic benefits of this class of compound to the protection against diabetes-associated atherosclerosis, a major complication affecting diabetic patients. Our study also raises the possibility that dh404 functions to lessen both diabetic kidney injury and diabetes-associated atherosclerosis, a highly desirable clinical outcome since cardiovascular and renal disease often occur concurrently (43). Finally, our findings, showing greater efficacy with respect to both cardiovascular and renal outcomes at lower dh404 doses, highlight the importance of determining the effective pharmacokinetic range for this class of drug. Further preclinical characterization of AIM compounds is therefore vital before one can assess the true potential of this drug class in the clinic.

Acknowledgments. The authors thank Katherine Ververis for excellent technical assistance with the NRK cell studies.

Funding. S.M.T. is supported by a Juvenile Diabetes Research Foundation International Postdoctoral Fellowship.

Duality of Interest. S.M.T., A.S., N.S., D.Y.C.Y., M.E.C., T.C.K., and J.B.d.H. report grants and nonfinancial support from Reata Pharmaceuticals, Inc. during the conduct of the study. C.M. and K.W.W. are employees of Reata Pharmaceuticals, Inc. and report personal fees from Reata Pharmaceuticals during the conduct of the study. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. S.M.T. and J.B.d.H. wrote the manuscript, researched data, and analyzed results. A.S., N.S., D.Y.C.Y., and T.C.K. researched data. C.M., K.W.W., and M.E.C. reviewed and edited the manuscript. J.B.d.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.