SIRT1 reduction causes renal and retinal injury in diabetes through endothelin 1 and transforming growth factor β1

In diabetes, hyperglycaemia causes up-regulation of endothelin 1 (ET-1) and transforming growth factor beta 1 (TGF-β1). Previously we showed glucose reduces sirtuin1 (SIRT1), a class III histone deacetylase. Here, we investigated the regulatory role of SIRT1 on ET-1 and TGF-β1 expression. Human microvascular endothelial cells were examined following incubation with 25 mmol/l glucose (HG) and 5 mmol/l glucose (NG) with or without SIRT1 or histone acetylase p300 overexpression or knockdown. mRNA expressions of ET-1, TGF-β1, SIRT1, p300 and collagen 1α(I) were examined. SIRT1 enzyme activity, ET-1 and TGF-β1 protein levels were measured. Histone acetylation and endothelial permeability were further investigated. Similar analyses were performed in the kidneys and retinas of SIRT1 overexpressing transgenic mice with or without streptozotocin induced diabetes. Renal functions were evaluated. In the endothelial cells (ECs), HG caused increased permeability and escalated production of ET-1, TGF-β1, collagen Iα(I). These cells also showed increased p300 expression, histone acetylation and reduced SIRT1 levels. These changes were rectified in the ECs following p300 silencing or by SIRT1 overexpression, whereas SIRT1 knockdown or p300 overexpression in NG mimicked the effects of HG. High ET-1 and TGF-β1 levels were seen in the kidneys and retinas of diabetic mice along with micro-albuminuria and increased fibronectin protein (marker of glucose-induced cell injury) levels. Interestingly, these detrimental changes were blunted in SIRT1 overexpressing transgenic mice with diabetes. This study showed a novel SIRT1 mediated protection against renal and retinal injury in diabetes, regulated through p300, ET-1 and TGF-β1.


Introduction
Diabetes and its long-term chronic complications are a growing concern worldwide [1] as the number of people being diagnosed with diabetes in North America is increasing in an alarming rate [1]. In diabetes, dysfunction of the vascular endothelium caused by hyperglycaemia is a key initiating factor in the development of all chronic diabetic complications including diabetic nephropathy and retinopathy [2,3]. A structural hallmark of such chronic diabetic complication is vascular basement membrane thickening and renal mesangial expansion. This is caused by an increased production of extracellular matrix (ECM) proteins such as fibronectin (FN) and collagen. Although other cells may take part in such process, endothelial cells (ECs) lining the blood vessels play an important role [2][3][4][5].
SIRTs (silent information regulator proteins) are a group of NAD dependent class III histone deacetylases that regulate epigenetic gene silencing. SIRTs deacetylase not only histones but also many proteins and transcription factors. SIRT1 is a prominent enzyme in the SIRT family residing both in the cytoplasm and nucleus. SIRT1 has been found to play crucial role in cellular oxidative stress response, metabolism, differentiation, longevity and various diseases including diabetes [20][21][22][23]. We have previously shown that glucose induced increased oxidative stress causes reduction of SIRT1 in ECs and tissues of animals with diabetes causing rapid ageing, reduced antioxidants and increased FN production in diabetes [20]. Moreover SIRT1 being a deacetylase also has a regulatory relationship with p300 [20].
We investigated the role of SIRT1 and its relationship with p300 in regulating ET-1 and TGF-b1 and their possible downstream consequences in ECs and tissues of animals affected by diabetes. To this extent, we used human microvascular ECs (HMECs) for in vitro inspections of glucose-induced endothelial alterations. We further expanded the study using SIRT1 overexpressing transgenic mice in which we induced diabetes with streptozotocin (STZ) and examined their renal and retinal tissues.

Cell culture
Dermal-derived HMEC was obtained (Lonza, Walkersville, MD, USA) and grown in endothelial cell basal medium 2 (EBM-2; Lonza). Human embryonic kidney (HEK293A) cells were obtained from ATCC (Manassas, VA, USA). We have previously described the culture conditions of these cells [24][25][26][27]. The glucose concentration in the growth medium was 5 mmol/l. No insulin was present in any media. All cells were maintained in a humidified atmosphere containing 5% CO 2 and at 37°C incubation. Cells were treated with normal glucose (NG, 5 mmol/l) or high glucose (HG, 25 mmol/l D-glucose) or osmotic control (OSM, 25 mmol/l L-glucose). All reagents were obtained from Sigma Chemicals (Sigma-Aldrich, Oakville, ON, Canada) unless specified otherwise. Experiments were conducted with 6-10 biological replicates.

Animal experiments
Wild-type (C57BL/6J) and SIRT1 transgenic mice (C57BL/6-Actb tm3.1 (Sirt1)Npa /J), male, 8 weeks old were obtained (Jackson Laboratory, Bar Harbor, ME, USA) and genotyped with Real Time RT-PCR. Diabetes was induced by single intra-peritoneal injection of STZ (65 mg/kg in citrate buffer). Streptozotocin (Sigma-Aldrich, Burlington, ON, Canada) was prepared in cold 0.1 M citrate buffer (pH 4.5) fresh before injection. Age and sex matched mice were used as controls and received equal volume of citrate buffer. Diabetes was defined in the animals with blood glucose level >20 mmol/l on two consecutive days with Freestyle Freedom Lite blood glucose monitoring system (Abbott Diabetes Care, Saint-Laurent, QC, Canada). The animals were fed on a standard rodent diet with water ad libitum and killed after 2 months following the development of diabetes (n = 10/group). Urinary micro-albumin (Exocell, Philadelphia, PA, USA) was measured as per the kit instructions. Retina and kidney tissues were collected and snap frozen in liquid nitrogen for histology. All tissues were stored at À80°C until further analysis.
All animal experiments were performed in accordance with regulations specified by the Canadian Council of Animal Care. All protocols were approved by the University of Western Ontario Animal Care and Veterinary Service. The investigation was in compliance with the Guide for the Care and Use of Laboratory Animals (NIH publ. no. 85-23, revised 1996).

SIRT1 enzyme activity assay
Whole cell lysate were collected following treatment with RIPA buffer (Millipore, MA) with addition of protease inhibitor (Roche, Laval, QC, Canada). The total protein concentrations of the samples measured using a commercially available BCA protein assay kit (Pierce, Rockford, IL, USA). Enzyme assay for SIRT1 activity was performed as per the manufacturer instructions (Sigma-Aldrich). The plates were read with a fluorescent spectrophotometer (Biotek, Winooski, VT, USA) at excitation 340 nm and emission 430 nm. mRNA extraction and cDNA synthesis RNA from cells was isolated using TRIZOL TM (Invitrogen, Burlington, ON, Canada) reagent as established at our laboratory. Briefly RNA was extracted with chloroform followed by centrifugation to separate the sample into aqueous and organic phases. The RNA was recovered from the aqueous phase by isopropyl alcohol precipitation and suspended in DEPC water. RNA concentration was assessed on a spectrophotometer (Gene Quant-Pharmacia Biotech, Cambridge, MA, USA). First-strand cDNA was made by using High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA, USA) according to the manufacturer instruction. The resulting products were stored at À20°C until further analysis by real time RT-PCR.

mRNA analysis with real time RT-PCR
Real time RT-PCR was performed in LightCycler TM (Roche Diagnostics) to quantify the mRNA expression of SIRT1 and using the Qiagen One Step RTPCR kit (detection platform SYBR Green I). All of the primers were either ordered or custom made from Sigma-Aldrich (Table 1) temperature (T m ) of specific products and primer dimers. According to the T m value of specific products for respective genes, an additional step (signal acquisition step, 2-3°C below T m ) was added following the elongation phase of RT-PCR. The data were analysed using the standard curve method (R 2 > 0.98, amplification efficiency>90%). The data were normalized to housekeeping gene 18sRNA to account for differences in reverse transcription efficiencies and the amount of template in the reaction mixtures. Adenoviral overexpression of SIRT1 and SIRT1 gene silencing SIRT1 adenovirus was obtained (ABM, Richmond, BC, Canada) and amplified in HEK293A cells. HMECs were transfected with the adenovirus as described earlier [24]. To silence SIRT1, transfection of small interfering RNA (siRNA) was performed using N-TER nanoparticle siRNA transfection system (Sigma-Aldrich) as described previously [20,27]. To silence SIRT1 a combination of two different siRNAs from Dharmacon Inc. (Lafayette, CO, USA) and Santa Cruz Biotechnology (Santa Cruz, CA, USA) was used to minimize the off target effects [20,27]. Transfection efficiency was assessed by real time RT-PCR.
p300 gene silencing and p300 overexpression A combination of four different siRNAs prepared in house was used to specifically silence the p300 expression in ECs as described by us pre-viously [19,20]. p300 overexpression was achieved with expression vectors (generously provided by Dr. Joan Boyes, The Institute of Cancer Research, London, UK) containing the human wild-type p300 (pCI-p300) and its HAT-deletion mutant (pCI-p300 HATD1472-1522) as described earlier [15]. Transfection efficiency was assessed by measuring p300 mRNA expression by real time RT-PCR.

ET-1, TGF-b1 & FN ELISA
Human/mice ET-1 ELISAs (Enzo, Farmingdale, NY, USA) and TGF-b1 ELISAs (eBioscience, San Diego, CA, USA) were done on cell and tissue lysates as per the manufacturer instructions. FN ELISAs (Abcam, Toronto, ON, Canada) were done on mice tissue lysates as per the manufacturer instructions. The total protein concentrations were measured by BCA protein assay kit (Pierce). The plate was read at 450 nm using a plate reader (Multiskan, Thermofisher, Toronto, ON, Canada).

Total ROS/RNS assay
Total ROS/RNS (reactive oxygen species/reactive nitrogen species) assay (Cell Biolab, San Diego, CA, USA) on tissue lysate was measured as per the manufacturer instructions. The plates were read with a fluorescent plate reader (Biotek) at excitation 480 nm and emission 530 nm.

Endothelial permeability assay
Human microvascular endothelial cells were seeded onto inserts (1 lm pores) in 24-well plates, with or without incubation with specific reagents for 24 hrs, and were tested for vascular permeability using the In Vitro Vascular Permeability Assay Kit (Millipore, Billerica, MA, USA) according to the manufacturer's instructions. Endothelial cells were seeded at a concentration of 2 9 10 6 cells/ml (200 ll/insert). Cell morphology and confluency was monitored under the microscope and upon formation of monolayer following 72 hrs, cells were treated with specific reagents (Ad-SIRT1, p300 siRNA, p300 plasmid, SIRT1 siRNA) with or without glucose. Following treatment FITC-Dextran is added to the cells and permeability of this fluorescent molecule is measured at several intervals (0.5, 1, and 2 hrs) by measuring the fluorescence of the receiver plate well solution. The plates were read with a fluorescent spectrophotometer (Biotek) at excitation 485 nm and emission 535 nm.

Western blotting
The western blot analysis was conducted according to the standard protocol established at our lab [15] using acetylated H3 lysine (Epigentek, Farmingdale, NY, USA) & b-actin antibody (Santa Cruz Biotechnology).

Statistical analysis
Data are expressed as mean AE SEM, normalized to controls. The statistical significance of the results was analysed by one-way or two-way

SIRT1 regulates ET-1 and TGF-b1 expressions in the ECs in high glucose
As microvascular ECs are the major target of glucose-induced organ damage, we investigated HMECs exposed to 5 (NG) and 25 (HG) mmol/l glucose for 72 hrs. These concentrations are based on our previous dose dependent analysis of SIRT1 levels [20,24]. Quantitative real time RT-PCR analysis of the ECs showed significant up-regulation of ET-1 and TGF-b1 mRNA levels ( Fig. 1A and C) and down-regulation of SIRT1 mRNA levels ( Fig. 1E) with HG compared to NG. No effects were seen following incubation with osmotic control (25 mmol/l of L-glucose, OSM; Fig. 1A, C and E). These findings were further reflected in ET-1 and TGF-b1 protein analysis and SIRT1 enzyme activity analysis, which showed a marked increase of ET-1 and TGF-b1 protein levels ( Fig. 1B and D) and reduced SIRT1 enzyme activity levels (Fig. 1F) in these cells following HG treatment (compared to NG). We then proceeded to explore any relationship, between the observed glucose-induced down-regulation of SIRT1 and upregulation of ET-1 and TGF-b1 expressions in these ECs. For this purpose, we analysed the ECs following transfection with SIRT1 adenovirus (Ad-SIRT1). Transfection efficiency was measured with real time RT-PCR and confirmed with SIRT1 enzyme analysis ( Fig. 1E and F). We noted the Ad-SIRT1 transfection significantly increased the SIRT1 mRNA and enzyme levels in the ECs in HG (Fig. 1E and F and TGF-b1 mRNA and protein levels in these ECs in HG (Fig. 1A-D). Endothelial cells transfected with null vector did not show such change ( Fig. 1A and C).
In addition, SIRT1 knockdown with siRNA (efficiency in Fig. 1E and F) in NG mimicked the effects of HG, causing an increase in ET-1 and TGF-b1 levels in these ECs (Fig. 1A-D). Conversely, ECs transduced with scramble siRNA did not show these changes (Fig. 1A-D). These results demonstrated a SIRT1 mediated regulation of ET-1 and TGF-b1 in the ECs in hyperglycaemia.

SIRT1 regulates ET-1 and TGF-b1 expressions via p300
In order to investigate the mechanism of such SIRT1 mediated ET-1 and TGF-b1 regulation we looked at transcriptional co-activa-tor p300, as we previously have found SIRT1 and p300 balance each other [20]. In keeping with the earlier findings we noted SIRT1 overexpression with adenovirus in 25 mmol/l glucose normalizes glucose-induced augmented p300 mRNA levels ( Fig. 2A). On the contrary, SIRT1 knockdown in NG with siRNA caused an increase in p300 mRNA levels ( Fig. 2A). Furthermore, p300 overexpression in these cells in NG reduced the SIRT1 mRNA levels ( Fig. 2B and C). These changes were not seen with null vector, scramble siRNA and p300 mutant transfections. These results reestablished a regulatory and balancing role of SIRT1 and p300 on each other in ECs.
To confirm the downstream effects of such changes we examined the cells following knockdown of p300 with siRNA (efficiency in Fig. 2B) in HG. Such treatment not only normalized the SIRT1 mRNA levels (Fig. 2C) but also reduced the ET-1 and TGF-b1 mRNA levels in HG ( Fig. 2D and E) compared to scrambled controls. In addition, p300 overexpression in NG lead to an increased mRNA levels for both ET-1 and TGF-b1 mimicking the HG treatments ( Fig. 2D and E). p300 mutant transfected cells did not show such change. ET-1 and TGF-b1 protein levels paralleled the mRNA findings ( Fig. S1A and B). These results further establish that SIRT1 regulates ET-1 and TGF-b1 through p300. We further conducted several rescue experiments to confirm such pathway. We first knocked down SIRT1 with siRNA in NG in these ECs, following which we transfected the cells with p300 siRNA. Transfection efficiency assessed with real time RT-PCR showed reduction of both SIRT1 (88%) and p300 (61%) mRNA levels in the treated cells. We noted such transfection prevented the ET-1 and TGF-b1 mRNA up-regulation in the treated cells (Fig. 2F). Moreover p300 overexpression in HG following treatment with Ad-SIRT1 reversed the beneficiary effects of SIRT1 causing up-reg ulation of ET-1 and TGF-b1 mRNA levels ( Fig. S1C and D). On the other hand p300 knockdown in these cases did not alter the beneficiary effects of SIRT1 normalizing ET-1 and TGF-b1 mRNA levels ( Fig. S1C and D).

SIRT1 overexpression prevents glucose-induced increased endothelial permeability and collagen Ia(I) expression
To investigate the functional consequences of SIRT1 mediated ET-1 and TGF-b1 regulation we conducted trans-endothelial permeability assay as increased endothelial permeability is a characteristic alter- ation in early diabetic microangiopathy. Data from the in vitro permeability test showed that HG significantly increases ECs permeability and Ad-SIRT1 transfection prevented such leakage in these cells ( Fig. 3A and B). p300 siRNA transfection in HG also had similar effect showing reduction in glucose-induced leakage whereas; SIRT1 knockdown or p300 overexpression in NG both significantly increased the ECs permeability mimicking the HG (Fig. 3A and B). In addition, p300 overexpression in HG following treatment with Ad-SIRT1 reversed the beneficiary effects of SIRT1 causing increased EC permeability whereas p300 knockdown in the same scenario did not alter the beneficiary effects of SIRT1 (Fig. S1E).
Previously we have shown SIRT1 overexpression prevented glucose-induced FN up-regulation in the ECs [27]. We further examined collagen Ia(I) expression levels following HG exposure. With Ad-SIRT1 transfection, we observed a normalization of glucose-induced up-regulation of collagen Ia(I) mRNA levels (Fig. 3C). HG induced increased collagen Ia(I) mRNA expression was further prevented by p300 siRNA transfection (Fig. 3C). In parallel, SIRT1 knockdown or p300 forced expression both caused an up-regulation of collagen Ia(I) mRNA levels in NG (Fig. 3C). Moreover, we conducted western blot analysis of acety-lated histone (Ac-H3K9/14) to see the functional consequences of p300 and SIRT1 alteration. Such analysis showed that HGinduced increased Ac-H3K9/14 was reduced with Ad-SIRT1 or p300 siRNA transfection (Fig. 3D). Opposingly, SIRT1 siRNA or p300 forced expression in NG increased such acetylation in the ECs. These experiments together provide further evidence that SIRT1 regulate ET-1 and TGF-b1 through p300.

SIRT1 regulates ET-1 and TGF-b1 expressions in the kidney and retina of diabetic mice
Following the establishment that SIRT1 regulated ET-1 and TGF-b1 in ECs, we expanded our study to investigate whether the mechanisms seen in these cells was important in the development of retinal tissue from these mice following 2 months of uncontrolled diabetes. We have previously shown that diabetes induced increased ECM protein and vasoactive factor expression is established at this time-point [19,20,[24][25][26][27]. mRNA analysis showed that diabetic mice had significant down-regulation of SIRT1 expressions in the kidneys and retinas (Fig. 4A and B). This finding was further reflected in the SIRT1 enzyme activity levels ( Fig. 4C and D). Moreover, these mice showed significant up-regulation of p300 mRNA levels in the renal and retinal tissues which were abrogated in SIRT1 overexpressing diabetic mice ( Fig. 4E and F). Western blot analysis further showed that SIRT1 overexpression caused reduction in histone acetylation (Ac-H3K9/14) in these tissues demonstrating the downstream functional consequences of p300 alteration in these tissues (Fig. 4G). Absence of Rd8 mutation [28] in the transgenic mice were confirmed with genotyping (Fig. S2A). Microscopic examination of renal and retinal tissues showed no morphologically evident lesions (Fig. S2B).
We further looked at ET-1 and TGF-b1 levels in these mice. SIRT1 overexpressing diabetic mice showed significantly reduced level of ET-1 and TGF-b1 mRNA levels compared to diabetic mice in both tissues (Fig. 5A-D). We further confirmed these results using ELISA which showed diabetes-induced increased ET-1 and TGF-b1 protein levels in the renal and retinal tissues are significantly reduced with SIRT1 overexpression in these tissues (Fig. 5E-H).

SIRT1 overexpression prevents diabetes-induced micro-albuminuria and FN up-regulation in tissues
To examine the downstream consequence of SIRT1 mediated alteration of ET-1 and TGF-b1 in diabetes, we investigated urinary micro-albumin levels to assess renal function in these mice. Diabetic mice showed increased micro-albuminuria compared to controls indicating poor kidney function in these animals (Fig. 6A). SIRT1 overexpression significantly reduced such micro-albuminuria in the transgenic diabetic animals (Fig. 6A). Moreover histological analysis of renal tissues in these animals showed reduced mesangial expansion (a feature of diabetic nepthropathy) compared to the wild-type diabetic animals (Fig. S2C). Furthermore in the renal and retinal tissues of the diabetic animals such overexpression reduced diabetes-induced increased FN protein, total ROS/RNS and collagen Ia(I) mRNA expression levels ( Fig. 6B-D). These results suggest that SIRT1 has a protective role against diabetes-induced renal and retinal damages.

Discussion
In this study, we have shown that SIRT1 regulates glucose-induced overexpression of ET-1 and TGF-b1 in the ECs. We have further shown that SIRT1 regulates ET-1 and TGF-b1 levels in the kidneys and retinas of diabetic animals. In addition, we have demonstrated such processes are regulated through transcriptional co-activator p300. Using adenoviral overexpression of SIRT1 and knockdown of SIRT1 with siRNA, we have directly demonstrated that glucoseinduced ET-1 and TGF-b1 up-regulation can be markedly reduced by increasing the availability of SIRT1. In the kidneys and retinas of SIRT1 overexpressing transgenic mice we have further established the existence of such regulation in a STZ-induced type 1 model of diabetes. To be best of our knowledge, such SIRT1 mediated regulation of ET-1 and TGF-b1 in diabetic complications have not been shown earlier.
In diabetes, high oxidative stress due to hyperglycaemia causes DNA damage and activates several transcription and vasoactive factors altering crucial gene expressions [2,6,7]. Here, we describe the role of SIRT1 causing protection against ET-1 and TGF-b1 induced endothelial damage in hyperglycaemia. We initially demonstrated SIRT1 down-regulation in parallel to ET-1 and TGF- b1 up-regulation in glucose exposed ECs. Then, we directly demonstrated its functional significance in terms of endothelial permeability and collagen Ia(I) expression and the relationship with p300 in these cells. In keeping with our earlier findings here we have shown that acting in opposing manner, p300 and SIRT1 regulate each other in hyperglycaemia. We further showed the functional significance of this mechanism in an animal model of diabetic microangiopathy.
Endothelin 1 and TGF-b1 are two important cytokines playing key roles in tissue damage in diabetic complications. It is of interest to note that such SIRT1 induced regulation is mediated through p300. p300 as a transcriptional co-activator is potentially capable of altering multiple transcripts [15,18,19]. This study shows that it is possible to target either SIRT1 or p300 to prevent diabetes-induced alterations. We have previously shown that p300 is regulated by miR-200b and it regulates miR-146a [24,26]. Additionally, we have shown that miR-195 regulates SIRT1 expression [27]. In this assumption we showed a ying-yang relationship between p300 and SIRT1. Hence a complex web connecting multiple such epigenetic processes involving acetylators, deacetylators, miRNAs and other undetermined epigenetic process may ultimately dictate expression of specific transcripts. It is also possible that SIRT1 may influence several other inflammatory cytokines which are of importance in the context of nephropathy and retinopathy and other chronic diabetic complications. The present study also does not exclude possible relationship of SIRT1 with other HATs or genes of anti-fibrotic/anti-oxidative properties. However, such roles of SIRT1 in these processes need to be established in future through specific experiments. Although the possibility of independent action of SIRT1 and p300 cannot be completely eliminated, our current study suggests that SIRT1 may act through p300 in the ECs to produce these transcripts. As ECs are an important target of glucose induced damage, the current study focused on to investigate the effect of hyperglycaemia on ECs. It is possible that similar alteration of SIRT1 may occur in other cell types, this was however beyond the scope of this article.

Supporting information
Additional Supporting Information may be found in the online version of this article: