Targeting HSP90 Ameliorates Nephropathy and Atherosclerosis Through Suppression of NF-κB and STAT Signaling Pathways in Diabetic Mice

The prevalence of diabetes has reached epidemic proportions. Nephropathy and atherosclerosis are the major diabetes-driven complications, resulting in disability and increased risk of mortality in diabetic patients. In fact, diabetic nephropathy is the leading cause of end-stage renal disease and also a macrovascular disease risk factor, and atherosclerosis is the main reason for impaired life expectancy in diabetic patients (1,2). Therefore, the prevention of micro- and macrovasculature damage associated with diabetes to improve patient quality of life and life expectancy is a major public health concern. Owing to the fact that current treatment of diabetes is insufficient to avert the complications in a significant proportion of patients, diabetes-associated renal and vascular damage might be counteracted by blocking injury mechanisms and promoting protective and reparative factors. Based on both observational clinical studies and functional evidence from animal models, inflammation is considered a key contributor to the onset and progression of diabetic nephropathy and atherosclerosis, with promising therapeutic opportunities (2,3).

Cellular stress response is directed by the conserved family of heat shock proteins (HSPs) to promote cytoprotection through the transcriptional regulation by heat shock transcription factor 1 (HSF1) (4). Given the ubiquity of HSP, it seems unsurprising that these chaperone proteins have been related to many metabolic abnormalities. Evidence from patients and animal models indicate that diabetes differentially modulates the expression of HSP25/27, HSP60, HSP70/72, and HSP90 (5–7), and this may affect the ability of cells to mount an effective cytoprotective response. HSP90, one of the most conserved HSPs, uses the energy generated by ATP binding and hydrolysis to allow folding, maturation, and stability of the so-called “client” proteins involved in cell survival, signal transduction, transcriptional regulation, and innate and adaptive immunity (4). To do so, HSP90 interacts with other cochaperones, forming a multichaperone complex that recognizes client proteins and modulates their activities (4). HSP90 machinery in association with its cochaperone HSP70 is involved in folding and activation of newly synthesized protein kinases, including IκB kinase (IKK) and Janus kinase (JAK) (8,9). IKK and JAK play, respectively, a key role in the activation of nuclear factor-κB (NF-κB) and signal transducers and activators of transcription (STAT), major transcription factors that regulate the expression of many genes involved in inflammation (10,11). This overall prompted the notion of HSP90 as a molecular target of interest for the treatment of several immunological and inflammatory disorders (12,13). So far, most clinical trials of HSP90 inhibitors have focused on cancer therapy (14).

Furthermore, HSP90 inhibitors effectively disrupt HSP90-HSF1 complexes, allowing HSF1 activation and subsequent nuclear translocation to transactivate HSP genes (4,14). Preclinical data implicate HSP90 as a promising anti-inflammatory target in rheumatoid arthritis, systemic lupus erythematosus, uveitis, liver injury, and cardiovascular disease models (12,15–18). In diabetic animals, HSP90 inhibition improved insulin sensitivity (19), high-fat diet–induced renal failure (20), and neurodegeneration (21), but the underlying mechanisms involved in these antidiabetic actions are not well defined. Hence, in this work we investigate whether targeting HSP90 delays the progression of diabetes complications in a diabetes mouse model of concomitant renal and macrovascular disease.

Animal studies were performed according to the Directive 2010/63/EU of the European Parliament and were approved by the Institutional Animal Care and Use Committee (IIS-Fundacion Jimenez Diaz). Experimental diabetes model of insulin deficiency was induced in 10-week-old male apolipoprotein E–deficient (apoE^−/−) mice by two daily intraperitoneal injections of streptozotocin (STZ, 125 mg/kg/day; Sigma-Aldrich, St. Louis, MO) (22–25). Animals maintained on standard diet were monitored every 2–3 days for body weight and nonfasting blood glucose. Severely hyperglycemic mice (blood glucose >29 mmol/L) received insulin (1–1.5 IU) to maintain blood glucose levels within a more tolerable range. Mice with overt diabetes (glucose >19.4 mmol/L) were randomized to receive 10 weeks of treatment with vehicle (200 μL saline i.p., every other day; n = 9) and 17-dimethylaminoethylamino-17-demethoxygeldanamycin (DMAG; InvivoGen, San Diego, CA) at 2 mg/kg/day (n = 9) and 4 mg/kg/day (n = 6). Age-matched apoE^−/− mice were used as nondiabetic controls (vehicle, n = 4; DMAG 4 mg/kg/day, n = 3). At the end of the study, 16 h–fasted mice were anesthetized (100 mg/kg ketamine and 15 mg/kg xylazine), saline perfused, and killed. Blood and urine samples were collected for biochemistry.

Dissected kidneys were snap-frozen for expression studies or stored in 4% paraformaldehyde for histology. Aortas were divided into two parts: the upper aortic root was embedded in Tissue-Tek O.C.T. Compound (Sakura Finetek Europe, the Netherlands) for histological analysis, and the abdominal/thoracic aorta was processed for mRNA analysis.

Paraffin-embedded kidney sections (3 μm) were stained with periodic acid Schiff (PAS). Renal lesions were semiquantitatively graded (0–3 scale) in a blinded manner according to the extent of lesions in glomeruli (hypertrophy, hypercellularity, and mesangial expansion; 30 glomeruli per sample), tubules (atrophy and degeneration), and interstitium (fibrosis and infiltration; 20 fields at ×40 per sample). Glomerular area and PAS⁺ mesangial area were quantified by computerized morphometry. Immunodetection of macrophages (F4/80; AbD Serotec, Oxford, U.K.), T cells (CD3; DAKO, Glostrup, Denmark), HSP70 (Abcam), and tyrosine-phosphorylated STAT1 (P-STAT1; Life Technologies, Rockville, MD) and STAT3 (P-STAT3; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) was assessed by immunoperoxidase. Renal fibrosis was determined by picrosirius red staining. Positive staining (>10 fields at ×20 magnification) was quantified using Image Pro-Plus (Media Cybernetics, Bethesda, MD) and expressed as percentage of total area and number of positive cells per glomerular cross-section or interstitial area measured (mm²).

Atherosclerotic lesions in serial 8-μm aortic sections (covering ∼1,000 μm from valve leaflets) were quantified by morphometry after Oil Red O/hematoxylin staining. Individual lesion area was determined by averaging the maximal values (approximately three sections). Collagen content was determined by picrosirius red staining. Lesional macrophages (MOMA-2; AbD Serotec, Oxford, U.K.), CD3 T cells, vascular smooth muscle cells (VSMCs) (α-actin-Cy3; Sigma-Aldrich), chemokine (C-C motif) ligand (CCL) 2 (Santa Cruz Biotechnology), CCL5 (Antibodies Online), and tumor necrosis factor α (TNF-α; Santa Cruz Biotechnology) were detected by immunoperoxidase or immunofluorescence. Positive staining was expressed as percentage of total plaque area or number of positive cells per lesion area (26,27). Activated NF-κB in kidney and aorta was detected by Southwestern histochemistry using digoxigenin-labeled probes (26,28).

Primary mesangial cells (MCs) from mouse kidneys were cultured in RPMI 1640 with 25 mmol/L HEPES, pH 7.4, supplemented with 10% FBS (22). Murine MC line (SV40 MES 13, CRL-1927; American Type Culture Collection [ATCC], Manassas, VA) was maintained in Dulbecco’s modified Eagle’s medium (DMEM):F12 medium containing 5% FBS. Murine proximal tubuloepithelial MCT line (29) was cultured in RPMI 1640 containing 10% FBS. Mouse vascular endothelial cell line (MILE SVEN 1, CRL-2279; ATCC) was maintained in DMEM with 5% FBS. Murine bone marrow–derived macrophages were cultured for 7 days in RPMI 1640 with 10% FBS containing 10% L929-cell conditioned medium (23). Murine monocyte/macrophage cell line (RAW 264.7, TIB-71; ATCC) was maintained in DMEM with 10% FBS. All culture media were supplemented with 100 units/mL penicillin, 100 g/mL streptomycin, and 2 mmol/L l-glutamine (Life Technologies). Quiescent cells were pretreated for 4 h with DMAG (5–50 nmol/L) before stimulation with high glucose (HG; 30 mmol/L d-glucose; Sigma-Aldrich) or cytokines (interleukin-6 [IL-6] 10² units/mL plus interferon-γ [IFN-γ] 10³ units/mL; PeproTech, Rocky Hill, NJ). Cell viability was assessed by the 1-(4,5-dimethylthiazol2-yl)-3,5-diphenylformazan thiazolyl blue formazan (MTT) method.

Cells grown to 60–70% confluence were transfected with 20–30 nmol/L of small interfering RNA (siRNA) targeting HSP70 or negative control scramble siRNA (Ambion) using Lipofectamine RNAiMAX reagent (Life Technologies). Transfected cells (silencing efficiency 50–75%) were pretreated with DMAG for 4 h before 24 h of cytokine stimulation.

Total RNA from mouse tissues (kidney and aorta) and cultured cells was extracted with TRIzol (Life Technologies) and analyzed by real-time quantitative PCR using Taqman gene expression assays (Applied Biosystem, Foster City, CA). Target gene expression was normalized to housekeeping gene (18S).

Total cell or kidney whole lysate proteins were resolved on SDS-PAGE gels and immunoblotted for fibronectin (Millipore Corporation, Billerica, MA), HSP32, HSP70 (Enzo Life Sciences, Farmingdale, NY), transforming growth factor-β (TGF-β), HSP27, HSP90, IκBα (Santa Cruz Biotechnology), P-STAT1 (Life Technologies), and P-STAT3 (Cell Signaling, Beverly, MA), using β-actin (Santa Cruz Biotechnology) or α-tubulin (Sigma-Aldrich) as loading controls. CCL2 and CCL5 protein levels were measured by ELISA (BD Biosciences; R&D Systems, Minneapolis, MN).

Renal cell chemoattractant capacity was determined using a transwell coculture system (28). In brief, RAW 264.7 cells (1 × 10⁵) were seeded on the upper transwell inserts (8.0-µm pore size; Merck Millipore, Billerica, MA) and then placed onto a 24-well plate containing MC or MCT cells (5 × 10⁴ per well), which were treated for 4 h with DMAG prior to cytokine stimulation (IL-6 + IFN-γ, 20 h). Cocultures were incubated for a further 8 h, nonmigratory cells were removed from the upper surface of the membrane, and migrated cells were fixed (4% paraformaldehyde), stained (0.2% crystal violet), and counted on 10 randomly selected ×200 fields.

Values are expressed as mean ± SEM. Statistical analyses were performed using Prism 5 (GraphPad Software, Inc., La Joya, CA). Differences across groups were considered significant at P < 0.05 using either nonparametric Mann-Whitney U test or one-way ANOVA followed by post hoc Bonferroni pairwise comparison test.

STZ-induced pancreatic injury in apoE^−/− mice represents an insulin-deficient model that combines hyperglycemia and hyperlipidemia. Diabetic apoE^−/− mice develop accelerated renal and vascular injury, with similarities to human diabetic nephropathy and atherosclerosis (24,30). We compared the evolution of diabetes in apoE^−/− mice after treatment with either vehicle or DMAG at two different concentrations (2 and 4 mg/kg) for 10 weeks. Throughout the study, blood glucose curves were similar in the three diabetic groups (Supplementary Fig. 1). At the end of the study, levels of blood glucose, GHbA_1c, total cholesterol, HDL and LDL cholesterol, and triglycerides were all significantly higher in diabetic apoE^−/− mice compared with nondiabetic controls, but they were not affected by DMAG treatment (Table 1). Serum transaminase activities were also similar across the groups (Table 1), indicating preserved liver function. As expected, diabetic mice had evidence of enhanced kidney disease, with significant increases in serum creatinine, kidney-to-body weight ratio, and urinary albumin-to-creatinine ratio (Table 1). Interestingly, DMAG treatment in diabetic animals prevented body weight loss and also significantly decreased serum creatinine, relative kidney weight, and albuminuria (Table 1), reflecting an improved renal function by DMAG.

Histopathological examination of PAS-stained sections revealed that diabetes caused moderate to severe renal damage in apoE^−/− mice, compared with nondiabetic controls (Fig. 1A–C). Furthermore, DMAG treatment dose-dependently ameliorated the renal pathologic changes associated with diabetes, including: 1) glomerular hypertrophy, hypercellularity (mesangial proliferation and infiltrating cells), mesangial matrix expansion, and capillary dilation; 2) tubular atrophy, epithelial dilation, and deposits of glycogen; and 3) interstitial fibrosis and inflammatory infiltrate (Fig. 1A–C and Table 1). DMAG had no significant effect on any parameters measured in nondiabetic controls (Fig. 1A–C and Table 1).

Quantification of F4/80-positive macrophages (Fig. 2A) and CD3-positive T cells (Fig. 2B) demonstrated a marked decrease (∼60%) in the number of infiltrating cells in both glomeruli and interstitium of diabetic mice treated with DMAG, compared with the vehicle-treated group. Consistently, DMAG reduced the gene expression levels of monocyte and T-cell chemokines (CCL2 and CCL5) and proinflammatory cytokine TNF-α in diabetic kidneys, with significant effect at the highest dose (% decrease vs. vehicle: 59 ± 8, 45 ± 12, and 85 ± 3, respectively; P < 0.05) (Fig. 2C). Chemokine protein levels in diabetic kidneys were also found significantly decreased in DMAG groups (Fig. 2D).

To assess the extent of renal fibrosis, we measured the collagen content by picrosirius red staining in kidney sections from nondiabetic and diabetic mice. DMAG effectively reduced the glomerular and tubulointerstitial fibrosis in diabetic apoE^−/− mice without significant effect in nondiabetic controls (Fig. 3A and B). Furthermore, diabetic mice treated with DMAG exhibited significantly lower levels of extracellular matrix proteins (type I collagen and fibronectin) and profibrotic factor TGF-β than the vehicle-treated group (Fig. 3C and D). DMAG also suppressed the renal expression of kidney injury molecule-1 (Kim-1) (Fig. 3C), a sensitive biomarker of tubular injury in several kidney diseases.

To assess the efficiency of DMAG on HSP90 inhibition, we studied the effects on its cochaperone HSP70, which is upregulated as a result of the blockage of the ATP-binding site of HSP90 (14). As expected, DMAG upregulated the mRNA and protein expression of HSP70 (∼2–3.5-fold increase vs. vehicle), but not HSP90, in diabetic kidneys, with a wide distribution in both glomerular and tubular regions of DMAG-treated mice (Supplementary Fig. 2A–C).

Among the different HSP90-associated proteins, we further analyzed whether DMAG modulates renal activation of NF-κB and STAT, key intracellular pathways in the regulation of inflammatory and fibrotic genes (3). In situ Southwestern histochemistry revealed an intense nuclear localization of activated NF-κB in glomerular and tubulointerstitial cells of diabetic mice, and a dose-dependent inhibition by DMAG treatment (Fig. 4A–C). DMAG administration also prevented STAT1 and STAT3 activation in diabetic kidneys, as demonstrated by immunodetection of tyrosine-phosphorylated proteins (Fig. 4A–D).

Quantification of aortic root sections after Oil Red O/hematoxylin staining (Fig. 5A–C) revealed that the induction of diabetes exacerbates atherosclerosis in apoE^−/− mice (2.5-fold increase in area compared with nondiabetic controls) and also demonstrates the atheroprotective effect of DMAG. In fact, DMAG attenuated atherosclerosis in nondiabetic apoE^−/− mice (Fig. 5A and B), as previously reported (17). Remarkably, atherosclerotic plaques of diabetic mice with DMAG treatment exhibited a dose-dependent decrease in the lesion size (Fig. 5B), extension (Fig. 5C), and neutral lipid content (% Oil Red O–positive area: vehicle, 23.9 ± 3.9; DMAG 2 mg/kg, 15.1 ± 1.2; DMAG 4 mg/kg, 4.2 ± 1.2; P < 0.05 vs. vehicle; not shown).

Accumulation of monocytes/macrophages (MOMA-2) (Fig. 5D) and T cells (CD3) (Fig. 5E) within the atherosclerotic plaques of diabetic mice was reduced by DMAG treatment, with significant effect at high dose (% inhibition vs. vehicle: 58 ± 9 and 56 ± 13, respectively; P < 0.02). Furthermore, analysis of collagen (picrosirius red staining) and VSMC (α-actin immunofluorescence) content in atherosclerotic plaques from DMAG-treated diabetic mice showed a more stable phenotype, with significant raises in the relative proportions of collagen to lipids and VSMCs to macrophages when compared with vehicle-treated diabetic mice (Fig. 5F).

The anti-inflammatory effect of DMAG in diabetic atherosclerosis was also evidenced by a marked decrease in the activation of NF-κB (Fig. 6A) and the expression of CCL2 (Fig. 6B), CCL5 (Fig. 6C), and TNF-α (Fig. 6D) in plaques of DMAG-treated compared with vehicle-treated mice. In addition, quantitative real-time PCR analysis on RNA from aorta presented notable reductions in gene expression of CCL2, CCL5, and TNF-α as a result of DMAG treatment (Fig. 6E). Furthermore, DMAG treatment resulted in a dose-dependent induction of HSP70 mRNA expression in mouse aorta (Fig. 6E).

The in vitro effects of DMAG on NF-κB and STAT activation and target gene expression were investigated in renal cells, vascular endothelial cells, and macrophages stimulated under hyperglycemic (HG) or proinflammatory (IL-6 + IFN-γ) conditions, in an attempt to mimic the diabetic milieu.

In the tubuloepithelial MCT cell line and primary macrophages, DMAG prevented HG-induced NF-κB activation, as evidenced by increased protein levels of inhibitory IκBα subunit (Fig. 7A). Similarly, DMAG pretreatment inhibited the degradation of IκBα in cytokine-stimulated MCT (Supplementary Fig. 3A), whereas no effect was observed with DMAG alone. Confocal microscopy experiments further revealed that DMAG reduced the nuclear translocation of the p65 subunit in vascular endothelial cells exposed to HG (Fig. 7B). We also observed that DMAG prevented the phosphorylation of STAT1 and STAT3 proteins in both HG-stimulated macrophages (Fig. 7C) and cytokine-stimulated renal cells (Fig. 7D and Supplementary Fig. 3B). Cell viability remained unaffected under these experimental conditions (Supplementary Fig. 3C).

We next examined the role of HSP90 inhibition on the expression of NF-κB/STAT-dependent genes (e.g., CCL2 and type I collagen) induced by hyperglycemia or inflammation. In cytokine-stimulated renal cells, DMAG pretreatment resulted in a significantly reduced mRNA expression of CCL2 and type I collagen (Fig. 7E). DMAG also prevented CCL2 protein secretion (Supplementary Fig. 3E) and led to a lower macrophage chemoattractant capacity of renal cells, as determined by a migration assay in a coculture system (% inhibition vs. cytokines: 56 ± 2 and 78 ± 4, respectively; P < 0.02) (Fig. 7F). Similarly, DMAG was also able to prevent HG-induced CCL2 expression in MCT, macrophages, and endothelial cells (Supplementary Fig. 3D).

The involvement of HSP70 in DMAG-mediated cellular effects was further explored by genetic silencing in renal cells. First, we observed that DMAG dose-dependently increased both gene (Supplementary Fig. 4A) and protein (Supplementary Fig. 4B) expression of HSP family members, predominantly HSP70, and to a lesser degree HSP32 and HSP27. In addition, cell transfection with HSP70 siRNA markedly reduced HSP70 expression, without affecting HSP90 levels (Supplementary Fig. 5). We also observed that HSP70 silencing increased the cellular susceptibility to inflammatory stimulation, as demonstrated by a 2.5-fold higher Ccl2 mRNA expression in HSP70-silenced cells compared with cells transfected with a control scrambled siRNA (Fig. 7G). Remarkably, DMAG was not able to significantly inhibit the cytokine-dependent Ccl2 mRNA expression in HSP70-silenced cells (Fig. 7G), providing evidence that the beneficial effect associated with HSP90 inhibition is partially due to the induction of HSP70 expression.

Alterations in cellular homeostasis, signaling pathways, and gene expression operate concurrently to develop diabetic vascular derangements, such as nephropathy and atherosclerosis. Diabetes is associated with defects in HSP function mediated by either transcriptional or posttranslational mechanisms that regulate cellular levels of HSP in a tissue-specific manner, and such impairments contribute to diabetes complications, thus resulting in a vicious cycle (31,32). The role of HSP90 extends beyond heat shock response owing to its chaperoning function to fold and stabilize many client proteins involved in cell proliferation, differentiation, apoptosis, and inflammation, and is therefore proposed as an important target in immunity and inflammation (4). In this report, we demonstrate that pharmacological HSP90 inhibition ameliorates diabetes-associated renal damage and atheroprogression through the attenuation of cellular processes regulated by NF-κB and STAT signaling pathways, therefore indicating that HSP90 is a therapeutic target in diabetes complications.

HSP90 inhibitors such as geldanamycin and its derivatives targeting HSP90 N terminus and blocking its ATPase activity have emerged as a clinically relevant strategy in cancer and are also promising drugs for immune and inflammatory diseases, including diabetes (14). Our results in a well-established mouse model of combined hyperglycemia and hyperlipidemia (STZ-induced diabetic apoE^−/− mice) demonstrate that HSP90 inhibition by DMAG effectively improved renal function, as evidenced by reduced serum creatinine and urine albumin-to-creatinine ratio, without overt side effects and toxicity. Unlike a previous study that shows reversed hyperglycemia by HSP90 inhibitors in diet-induced obese mice (19), we observed that DMAG administration did not affect the metabolic severity of diabetes, with no changes in hyperglycemia, lipid profile, and body weight. Of note, DMAG treatment attenuated the structural pathologic changes in the diabetic kidney, including glomerular hyperplasia and mesangial matrix expansion, as well as tubular atrophy, interstitial fibrosis, and inflammation, hallmarks of end-stage renal failure. In line with the recently reported effect of HSP90 inhibitor on high-fat diet–induced renal failure in diabetes (20), our work adds differences in treatment dose and duration and study samples and provides mechanistic details about the renoprotective action of DMAG in diabetic mice.

Diabetes-associated inflammation promotes a progressive accumulation of kidney leukocytes (33) that contribute to the development and progression of diabetic nephropathy either by direct interaction with the renal cells or by releasing cytokines and growth factors involved in cell proliferation, migration, and extracellular matrix production (34,35). Besides renal histological improvement, we also found a marked anti-inflammatory and antifibrotic effect of HSP90 inhibition. In fact, DMAG-treated mice exhibited reduced macrophage and T-cell infiltration along with a decreased expression of proinflammatory genes (TNF-α, CCL2, and CCL5). Previous reports indicated that HSP90 inhibitors prevent inflammatory gene expression in different cell types, including cancer cells (36), leukocytes (17,37), and renal and vascular cells (15,17). In line with this, our studies in cultured mesangial and tubular cells mimicking hyperglycemic and inflammatory diabetic conditions demonstrated that DMAG inhibited the gene expression and protein secretion of CCL2 and also mitigated macrophage chemotaxis, further substantiating the in vivo findings. Moreover, we observed that DMAG protected from the development of renal fibrosis in diabetic mice, with significant reductions in the expression of extracellular matrix proteins and profibrotic factor TGF-β. These observations were further confirmed in vitro by analyzing the expression of an extracellular matrix protein (type I collagen) in renal cells. In agreement with the previously reported ability of DMAG to inhibit TGF-β signaling in vitro (38) and to reduce renal fibrosis in mice (39), our findings point toward HSP90 as an attractive target to attenuate renal inflammation and fibrosis in diabetic nephropathy.

Our data highlight a potential use of HSP90 inhibitors for prevention of diabetes-associated atherosclerosis. Indeed, DMAG administration dose-dependently reduced the size and extension of atherosclerotic lesions in STZ-induced diabetic apoE^−/− mice and also altered plaque composition without any incidence in lipid profile. Increased HSP90 expression has been linked to vulnerability of human advanced atherosclerotic lesions, and therefore it has been suggested as a promising target to stabilize atheroma plaques through the modulation of inflammation and oxidative stress (17,18,40). Our observations in diabetic mice revealed that DMAG treatment promoted a less inflamed, more stable phenotype characterized by lower content of lipids, macrophages, and T cells and higher collagen and VSMC content. Additionally, atherosclerotic lesions of DMAG-treated mice displayed a lower amount of NF-κB–activated cells together with a downregulation in the aortic expression of cytokines and chemokines involved in proatherogenic activation of vessel cells. Our in vitro studies performed in macrophages and endothelial cells, two of the cell types involved in atherosclerosis progression, demonstrated that DMAG impaired inflammatory gene expression in a hyperglycemic environment. These results are in agreement with the recently reported antimigratory effect of HSP90 inhibitor during atherogenesis (41).

Our study provides in vivo and in vitro evidence that DMAG inhibits NF-κB and STAT, two representative HSP90-regulated pathways that control key pathological mechanisms involved in diabetes complications. Owing to its chaperoning function, HSP90 regulates the stability and the activation of NF-κB and STAT signaling pathways by direct association and stabilization of their respective activating kinases, IKK (42) and JAK2 (43). Accordingly, the efficacy of HSP90 inhibitors in animal models has been linked to the disruption of the IKK complex and JAK2 protein stability, further inhibiting downstream transcription factors such as p50/p65 NF-κB (44) and STAT1/STAT3/STAT5 (45). Dysregulated NF-κB and STAT pathways contribute to diabetic nephropathy and atherosclerosis by inducing the transcription of many genes associated with inflammation, renal fibrosis (45–47), and the proatherogenic state (23,48). We have previously tested a number of potential novel approaches (e.g., kinase inhibitors, permeable peptides, and gene therapy) to tackle diabetic nephropathy and atherosclerosis by separately inhibiting the STAT and NF-κB pathways (23,26,28,29). Here, we demonstrated that DMAG treatment resulted in combined inhibition of the NF-κB and STAT pathways and subsequent downregulation of their inducible genes (e.g., cytokines, chemokines, and extracellular matrix proteins) both in diabetic mice and in vitro. Given that HSP90 inhibitors degrade many different client proteins, it is likely that the protective effects of DMAG on diabetic mice may result from inhibition of multiple target proteins in renal and vascular cells.

Another established effect of HSP90 inhibition is the induction of HS70 expression through the activation of HSF1, resulting from HSP90-HSF1 complex disruption (14,19). HSP70 is likely to confer protection against disturbed metabolic homeostasis via multiple modes of action, including reduced inflammation (49) and improved insulin sensitivity (19). In agreement with previous descriptions (38,50), we demonstrate that DMAG treatment dose-dependently induced HSP70 expression in kidneys and aorta of diabetic mice and in cultured cells. Interestingly, we also observed that HSP70 gene silencing enhanced cytokine responsiveness of cells and partially attenuated the inhibitory effect of DMAG, thus suggesting that DMAG-mediated HSP70 induction contributes to cytoprotection and counterbalance of diabetes-induced cell damage.

Collectively, our experimental data provide the first evidence that HSP90 inhibition simultaneously ameliorates diabetes-associated nephropathy and atherosclerosis through the induction of protective HSP70 and the attenuation of the NF-κB and STAT pathways, thus resulting in improved renal inflammation and fibrosis and stabilization of atherosclerotic plaques. Therefore, we propose HSP90 inhibition as a novel intervention to limit diabetes-associated complications.

Acknowledgments. The authors thank Dr. M.D. Sanchez-Niño (Nephrology Department, IIS-Fundacion Jimenez Diaz) for helpful advice on the siRNA experiments.

Funding. This work was supported by grants from the Spanish Ministry of Economy and Competitiveness (SAF2012-38830 and SAF2010-21852), the Spanish Ministry of Health (FIS PI14/00386 and DiabetesCancerConnect PIE13/00051), the Spanish Societies of Nephrology and Atherosclerosis, and the Iñigo Alvarez de Toledo Renal Foundation. I.L. was supported by a postdoctoral fellowship from FIS (Sara Borrell program).

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

Author Contributions. I.L. designed experiments, researched and analyzed data, and wrote the manuscript. A.O. and C.R. researched and analyzed data and critically revised the manuscript. B.M., J.M.-M., and J.B. researched, analyzed, or interpreted data. J.E. and J.-L.M.-V. critically revised the manuscript for important intellectual content. C.G.-G. conceived and designed the study, analyzed data, and drafted and reviewed the manuscript. C.G.-G. 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.