CTRP13 Preserves Endothelial Function by Targeting GTP Cyclohydrolase 1 in Diabetes

Diabetes increases the risk of developing cardiovascular diseases, including nephropathy, coronary heart disease, and stroke (1,2). Endothelial dysfunction associated with reduced nitric oxide (NO) bioavailability is an important initiator that leads to multiple diabetic cardiovascular events (3,4). Hyperglycemia-induced uncoupling of endothelial NO synthase (eNOS) in the endothelial cells of patients with diabetes diminishes NO availability and leads to excessive superoxide anion (O₂⁻) production, which impairs endothelium-dependent relaxation (EDR) (5,6). Therefore, approaches to enhance NO availability are regarded as promising strategies to improve endothelial function in diabetes.

Adipokines, which are secreted predominantly from adipocytes, are a group of factors that control energy balance, cardiovascular homeostasis, insulin sensitization, and inflammatory responses (7). Currently, accumulative evidence indicates that multiple adipokines are related to the regulation of endothelial dysfunction (8,9). Adiponectin has emerged as a fat-secreted hormone that protects against endothelial dysfunction by inducing the phosphorylation of eNOS at Ser1177 and NO production (10). In addition, leptin induces endothelial neuronal NOS expression to maintain EDR (11). In contrast, resistin can deteriorate endothelium-dependent and endothelium-independent vasorelaxation by decreasing eNOS expression (12,13). However, the precise roles of adipokines in the regulation of endothelial dysfunction remain incompletely characterized.

C1q/tumor necrosis factor–related proteins (CTRPs) are members of the highly conserved family of adiponectins, which have common structural characteristics and a COOH-terminal globular C1q domain and play pleiotropic roles in modulating the physiology, metabolism, and pathophysiology of endocrine, immune, and cardiovascular systems (14). Among the 15 CTRPs identified to date, CTRP13 is highly conserved evolutionarily, with only one single amino acid change between mouse and human forms. As an adipokine, CTRP13 has been reported to improve glucose metabolism in adipocytes, myotubes, and hepatocytes (15). Our recent study suggests that serum CTRP13 is highly associated with vascular calcification in chronic kidney disease, and ectopic CTRP13 infusion dramatically inhibits the osteogenic differentiation of vascular smooth muscle cells and reduces calcium deposition (16). In addition, we found that CTRP13 supplementation reduces macrophage foam cell formation and inflammation reactions, thus alleviating the development of atherosclerotic plaques (17). However, extensive information on the relationship between CTRP13 and endothelial function and the underlying mechanisms is very limited. In this study, we investigated whether there is an association between CTRP13 and endothelial dysfunction in diabetes and explored potential mechanisms.

The clinical experiment was performed according to the standards of the Declaration of Helsinki. This study was supported by the Ethics Committee of Tongji Medical College at Huazhong University of Science and Technology. Human renal artery specimens were obtained from patients who underwent nephrectomy with hyperglycemia (diabetes) or normal blood glucose levels (no diabetes) after informed consent was obtained (Supplementary Table 1).

All experimental procedures were approved by the Animal Use Subcommittee at Tongji Medical College, Huazhong University of Science and Technology. C57BL/6, db/m, and db/db mice at 8–10 weeks of age were purchased from Beijing HFK Bioscience Co. Ltd. (Beijing, China). Peroxisome proliferator–activated receptor α (PPARα) knockout (PPARα^−/−) mice were kindly provided by Dr. Hongliang Li (Animal Research Center, Cardiovascular Research Institute, Wuhan University School of Medicine, Wuhan, China). Mice were housed at 22–24°C under a 12-h light/dark cycle and had free access to water.

For the streptozotocin (STZ)–induced diabetic model, C57BL/6 mice at 8 weeks old were injected with a low dose of STZ (50 mg/kg/day, 5 consecutive days, i.p.) to induce persistent hyperglycemia as described by the Animal Models of Diabetic Complications Consortium (http://www.amdcc.org). Hyperglycemia was defined as a random blood glucose level of >16.6 mmol/L at 2 weeks after injection, and only animals that tested positive were continued in the study. Diabetic mice were housed for another 6 weeks, randomly divided into several groups at 16 weeks old, and administered with recombinant human CTRP13 protein (Aviscera Bioscience) (50 mg/kg or 200 mg/kg; n = 10, respectively) or saline (n = 10) by intraperitoneal injection every other day for 2 weeks (Supplementary Table 2).

For the db/db diabetic model, male db/m and db/db mice at 12–14 weeks were randomly divided into several groups and administered with recombinant human CTRP13 protein (Aviscera Bioscience) (50 mg/kg or 200 mg/kg; n = 10, respectively) or saline (n = 10) by intraperitoneal injection every other day for 2 weeks (Supplementary Table 3).

After CTRP13 treatment, mice were anesthetized with an overdose intraperitoneal injection of 100 mg/kg pentobarbital. Blood and aorta samples were collected for analysis. Plasma CTRP13 levels were measured with an ELISA kit (Cusabio, Wuhan, China).

After the mice were killed, the thoracic aortae were removed and placed in an oxygenated ice-cold Krebs solution that contained 119 mmol/L NaCl, 25 mmol/L NaHCO₃, 2.5 mmol/L CaCl₂, 4.7 mmol/L KCl, 1.2 mmol/L KH₂PO₄, 1 mmol/L MgCl₂, and 11 mmol/L D-glucose. Changes in the isometric tone of the aortic rings were recorded on the myograph (Danish Myo Technology, Aarhus, Denmark). The rings were stretched to an optimal baseline tension of 3 mN and then allowed to equilibrate for 60 min before the experiment commenced. The rings were first contracted with 60 mmol/L KCl and rinsed in Krebs solution. After several washouts, phenylephrine (1 μmol/L) was used to produce a steady contraction, and acetylcholine (ACh) (10 nmol/L to 10 µmol/L) was added cumulatively to induce EDR. Endothelium-independent relaxation in response to sodium nitroprusside (SNP) was stimulated in the aortic rings, with the endothelium removed by gentle rubbing with fine forceps.

Mouse thoracic aortic rings (2 mm) were dissected in sterile PBS and incubated in DMEM (Gibco) supplemented with 10% FBS (Gibco). Drugs, including 3-deoxy-d-arabinoheptulosonate 7-phosphate (DAHP) (10 mmol/L) (Sigma-Aldrich, St. Louis, MO), H89 compound (10 μmol/L) (Sigma-Aldrich), and recombinant CTRP13 (50 ng/mL or 300 ng/mL), were individually added into the culture medium that bathed the aortic rings. High-glucose (HG) conditions were achieved by the addition of 30 mmol/L glucose, whereas 30 mmol/L mannitol was used as the normal glucose (NG) osmotic control. After the incubation period, the rings were transferred to a chamber filled with fresh Krebs solution and mounted on a myograph for measurement of changes in isometric force. ACh, phenylephrine, and SNP were dissolved in water, and other compounds were dissolved in DMSO.

Endothelial cells were isolated from freshly resected arteries as described previously (18). Briefly, mice were anesthetized with an intraperitoneal injection of pentobarbital sodium (40 mg/kg). The aortae were dissected in DMEM and incubated with collagenase type II for 8 min at 37°C. Detached endothelial cells were collected by centrifugation, resuspended in 20% FBS-DMEM, and then cultured in endothelial basal medium supplemented with bovine brain extract (Lonza, Walkersville, MD) until confluency. CD31 microbeads (magnetic-activated cell sorting [MACS]) (catalog number 130–097–418; Miltenyi Biotec, San Diego, CA) were used to separate endothelial cells. First, the CD31-positive cells were magnetically labeled with CD31 microbeads. Then, the cell suspension was loaded onto an MACS column, which was placed in the magnetic field of an MACS separator. The magnetically labeled CD31-positive cells were retained within the column. The magnetically retained CD31-positive cells were eluted as the positively selected cell fraction. Then, CD31 microbead–labeled cells were cultured in DMEM and 10% FBS and used for expression analysis.

Human umbilical vein endothelial cells (HUVECs) (Clonetics; Lonza) were grown in endothelial basal medium supplemented with 2% FBS. Cultured cells were used for experiments between passages 4 and 8. All cells were incubated in a humidified atmosphere of 5% CO₂ at 37°C. When 70–80% confluent, the cells were treated with different agents.

Intracellular reactive oxygen species (ROS) levels were measured using dihydroethidium (DHE) fluorescence as described previously (19). Briefly, cells were incubated with DHE (10 μmol/L) for 30 min, homogenized, and subjected to methanol extraction. High-performance liquid chromatography (HPLC) was performed using a C-18 column (mobile phase: gradient of acetonitrile and 0.1% trifluoroacetic acid) to separate and quantify oxyethidium (product of DHE and O₂⁻) and ethidium (a product of DHE auto-oxidation). ROS levels were determined by conversion of DHE into oxyethidium.

NO production was detected using the fluorescent probe diaminofluorescein (DAF) (20). Briefly, before the end of treatment, 10 μmol/L DAF was added to the medium and incubated for 30 min at 37°C, followed by two washes with PBS. The DAF fluorescent intensity was recorded by a fluorescent reader at the wavelengths for excitation (485 nm) and emission (545 nm).

The measurements of serum NO and serum superoxide dismutase (SOD) activity were made using commercial kits as recommended by the provider. Commercial kits for the measurements of NO and SOD activity were purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing, China).

Tetrahydrobiopterin (BH4) levels in the cell lysates were measured using a competitive ELISA kit (MyBioSource, San Diego, CA). Briefly, ELISA plates precoated with a BH4-specific antibody were incubated with the samples and test standards together with a fixed amount of biotin-labeled BH4. Excess reagents and sample were washed off, avidin-conjugated horseradish peroxidase was added to each well, and the plate was incubated again. 3,3′,5,5′-Tetramethylbenzidine liquid substrate was then added to each well for 10 min. The enzyme substrate reaction was terminated, and the product was measured spectrophotometrically at 450 nm.

Total RNA was extracted with TRIzol (Takara Holdings, Kyoto, Japan). For mRNA quantification, 1 mg of RNA was reverse-transcribed into cDNA using a PrimeScript RT Reagent Kit (Takara Holdings) followed by real-time quantitative PCR with SYBR Green (Bio-Rad) and a Bio-Rad CFX-96 real-time system. All of the quantitative RT-PCR primers were ordered from the Qiagen RT² qPCR Primer Assays online shop (Qiagen, Germantown, MD).

HUVECs were cultured in the presence of the transcriptional inhibitor actinomycin D (2 mg/mL) (Millipore Sigma) for various durations (0, 1, 2, and 4 h) to measure the rate of decay of GTP cyclohydrolase 1 (GCH1) mRNA after treatment with different concentrations of CTRP13 (50 and 300 ng/mL) or with the vehicle control in HG medium.

pGL3 luciferase reporter vectors (Promega, Madison, WI) were used to construct luciferase vectors. To identify the functional site in the human GCH1 promoter, a series of 5′ GCH1 promoter deletions (pGL-GCH1–114, pGL-GCH1–321, pGL-GCH1–522, pGL-GCH1–810, and pGL-GCH1–1,083) was constructed by amplifying the GCH1 promoter region with different forward primers. The putative binding site of PPARα, which is located −280 bp from the human GCH1 promoter, was deleted by site-directed mutagenesis using a QuikChange II Kit (Stratagene, San Diego, CA) according to the manufacturer’s protocol. Luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega).

Chromatin immunoprecipitation (ChIP) assays were performed using the Pierce Agarose ChIP Kit (Thermo Fisher Scientific, Waltham, MA). Briefly, treated HUVECs were crosslinked, harvested, and sonicated to obtain DNA fragments between 400 and 800 bp in length. The supernatants were incubated with antibodies against IgG or PPARα and rotated overnight at 4°C. After washing, DNA was pulled down and purified using the QIAquick PCR purification kit (Qiagen). Purified DNA was analyzed by real-time PCR with specific primers for the GCH1 promoter.

Western blot was performed as described previously (21,22). HUVECs were lysed with ice-cold RIPA lysis buffer (Thermo Fisher Scientific) supplemented with a complete protease inhibitor cocktail (Thermo Fisher Scientific). After centrifugation at 12,000g for 20 min, the proteins were boiled in Roti-Load Buffer (Carl Roth, Karlsruhe, Germany) at 100°C for 10 min. Equal amounts of proteins were separated on SDS-PAGE. eNOS dimerization was assayed using low-temperature SDS-PAGE. Briefly, protein lysates were mixed with loading buffer (without β-mercaptoethanol) and without boiling before loading. Electrophoresis and blotting were performed at 4°C during the entire procedure. Antibodies against GCH1 (1:1,000) (SAB1410516; Sigma-Aldrich), dihydrofolate reductase (DHFR) (1:1,000) (sc-136246; Santa Cruz Biotechnology), eNOS (1:1,000) (32027, 9586; Cell Signaling Technology), phospho-(Ser/Thr) antibody (1:1,000) (ab17464; Abcam), phosphodiesterase (PDE) 4A (1:1,000) (ab14607; Abcam), PDE4B (1:1,000) (72096; Cell Signaling Technology), β-actin (1:2,000) (ab8226; Abcam), and PPARα (1:1,000) (ab227074; Abcam) were used as primary antibodies. After incubation with the corresponding secondary antibody, chemiluminescence signals were detected with Image Lab statistical software (Bio-Rad). Bands were quantified with ImageJ software (National Institutes of Health, Bethesda, MD), and the results are shown as the ratio of the total protein level to the β-actin level and normalized to the control groups.

Protein kinase A (PKA) activity was measured with an ELISA-based PKA activity kit (Assay Design, Ann Arbor, MI) according to the manufacturer’s instructions. HUVECs were homogenized in lysis buffer supplemented with 0.4 mmol/L 3-isobutyl-methylxanthine. PKA activity in the homogenate was then measured.

HUVECs, preinfected with glutathione S-transferase (GST)–tagged PPARα lentivirus (source), were treated with HG and CTRP13 for 48 h. Extracted proteins from cells were incubated with anti-GST antibody (2624; Cell Signaling Technology) and rotated overnight at 4°C. Magnetic beads were added for another 4 h. Bead complexes were washed with RIPA lysis buffer. Immunoprecipitates were then mixed with SDS loading buffer, boiled for 10 min, and subjected to Western blotting.

The data are presented as the mean ± SEM of the experimental results. Concentration-response curves were analyzed using GraphPad Prism version 6.0 software. Statistical significance was determined by Student t test (two-tailed) or one-way ANOVA followed by the Bonferroni post hoc test when more than two treatments were compared. P < 0.05 indicates a significant difference between groups.

The data that support the findings of this study are available within the article and its Supplementary Data or from the corresponding authors on reasonable request.

The level of ACh-induced EDR in the db/db mouse aortae was significantly worse than that in the db/m mouse aortae (Fig. 1A). Similarly, ACh-induced EDR in the main mesenteric arteries of db/db mice was worse than that in db/m mouse main mesenteric arteries (Fig. 1B). CTRP13 infusion (50 mg/kg every other day or 200 mg/kg every other day for 2 weeks) increased the plasma CTRP13 concentration in db/db mice (Supplementary Fig. 1A and Supplementary Table 2) and restored the impaired EDR in the aortae and main mesenteric arteries from the db/db mice (Fig. 1A and B). SNP-induced relaxations were similar among the different groups (Fig. 1C). In addition, we used STZ injection to generate an experimental model of type 1 diabetes (Supplementary Table 3). Compared with the control group, clear endothelial dysfunction with an impaired vasorelaxation response to ACh was evident in the aortae and main mesenteric arteries of mice injected with STZ (Fig. 1D and E). However, exogenous CTRP13 supplementation, verified by elevated plasma CTRP13 levels in STZ-injected mice (Supplementary Fig. 1B), improved EDR in STZ-injected mouse aortae (Fig. 1D). Likewise, ACh-induced EDR in the mesenteric arteries was also augmented by CTRP13 infusion (Fig. 1E). The levels of endothelium-independent vasorelaxation in response to SNP had no statistic differences in the mesenteric arteries among the four groups (Fig. 1F). These results indicate that hyperglycemia-disrupted EDR can be alleviated by CTRP13 supplementation in vivo.

In addition to in vivo CTRP13 infusion, ex vivo recombinant CTRP13 treatment (50 ng/mL or 300 ng/mL) also restored the impaired EDRs in the isolated db/db mouse aortae without influencing SNP-induced relaxations (Fig. 2A and B). Moreover, a 48-h exposure with HG (30 mmol/L) attenuated EDRs in cultured aortae from C57BL/6 mice, which was conversely improved by cotreatment with CTRP13 (Fig. 2C). Consistently, the levels of relaxation in response to SNP were comparable within each of the different groups (Fig. 2D). These results suggest that CTRP13 treatment can also protect endothelial function ex vivo.

eNOS is responsible for vascular NO production (23). When eNOS becomes uncoupled, ROS are generated instead of NO, and this finally leads to endothelial dysfunction (24). First, we measured NO and ROS levels in HG-treated HUVECs with either vehicle or CTRP13. As shown in Fig. 3A and B, HG treatment markedly decreased NO release and increased ROS production in cultured HUVECs, and this was reversed by recombinant CTRP13 treatment. The effects of CTRP13 treatment on NO production and SOD activity in vivo were also examined. In comparison with those in control mice, the serum levels of NO and SOD activity all decreased in STZ-injected mice, but CTRP13 supplementation rescued these abnormal phenotypes (Fig. 3D and E). Similarly, CTRP13 increased serum NO levels and SOD activity in db/db mice (Fig. 3G and H). eNOS dimerization is necessary for its normal production of NO. Thus, we went on investigating the role of CTRP13 on eNOS dimerization. As depicted in Fig. 3C, a decline of 75% in the ratio of eNOS dimers to monomers, which indicates the uncoupling state, was found in HG-treated HUVECs compared with control cells. However, treatment with CTRP13 ameliorated the HG-decreased ratio of eNOS dimers to monomers in a dose-dependent manner. We also isolated endothelial cells from the whole aorta of STZ-injected mice and db/db mice, which showed a significantly decrease in the eNOS dimer level and dimer/monomer ratio, while treatment with CTRP13 expectedly preserved the eNOS uncoupling (Fig. 3F and I).

BH4 is an essential cofactor for eNOS dimerization and completely couples NADPH oxidation to NO production (25,26). Therefore, we further explored the involved upstream pathways. As depicted in Fig. 4A, HG treatment significantly reduced BH4 production compared with the vehicle control, but treatment with CTRP13 effectively recovered the BH4 level in HUVECs. Consistently, the decrease in BH4 level in aortic arteries from STZ-injected or db/db mice was mostly normalized by CTRP13 treatment in a dose-dependent manner (Fig. 4D and G). GCH1 and DHFR are the two key factors that catalyze the biosynthesis of BH4 and NO (27,28). As depicted in Fig. 4B, HG significantly reduced the protein levels of both GCH1 and DHFR. Treatment of HUVECs with exogenous CTRP13 only reversed GCH1 expression but exerted no marked effects on DHFR expression. Similarly, compared with those of the controls, the expression levels of GCH1 in isolated aortic endothelial cells from STZ-injected and db/db mice were markedly upregulated by CTRP13 treatment, whereas the levels of DHFR were not changed under CTRP13 treatment (Fig. 4E and H), all suggesting that CTRP13 promotes GCH1 expression.

To determine whether inhibition of GCH1 could bypass the positive effects of CTRP13 on endothelial function, we applied the GCH1 inhibitor DAHP in aortic rings ex vivo from C57BL/6 mice under HG treatment or in vivo from STZ-injected and db/db mice. Importantly, the addition of DAHP abrogated the protective effects of CTRP13 on improving ACh-induced EDR (Fig. 4C, F, and I), revealing the importance of GCH1 in mediating the protective effects of CTRP13 on endothelial function.

CTRP13 significantly restored the GCH1 mRNA level that was suppressed by HG in cultured HUVECs, compared with that in the controls (Fig. 5A). In accordance with this observation, forced expression of CTRP13 also dramatically abrogated the decrease in GCH1 expression in isolated aortic endothelial cells from STZ-injected and db/db mice (Fig. 5B and C). Considering that CTRP13 upregulated GCH1 expression at both the protein and mRNA levels, we wondered whether this regulation occurred mainly at the level of mRNA synthesis or mRNA decay. Actinomycin D was used to inhibit endogenous mRNA transcription, and the half-life of GCH1 mRNA was determined. Results showed that CTRP13 did not affect GCH1 mRNA degradation (Fig. 5D), indicating the possibility of the transcriptional manner. Next, we constructed a series of luciferase reporter plasmids containing different lengths of the GCH1 promoter, including pGL3 (control), pGL3-GCH1–114 (−114 to +102 bp), pGL3-GCH1–321 (−321 to +102 bp), pGL3-GCH1–522 (−522 to +102 bp), pGL3-GCH1–810 (−810 to +102 bp), and pGL3-GCH1–1,083 (−1,083 to +102 bp), and performed luciferase activity assays to determine whether CTRP13 acted through the transcriptional regulation. Interestingly, CTRP13 deficiency markedly enhanced the luciferase activity of four plasmids (pGL3-GCH1–321, pGL3-GCH1–522, pGL3-GCH1–810, and pGL3-GCH1–1,083), but not the activity of pGL3 (control) or pGL3-GCH1–114, indicating that the region essential for GCH1 regulation is located between −321 and −114 bp (Fig. 5E). We screened one consensus PPARα-binding element candidate in the region. To confirm the importance of this region, we deleted the putative binding site around pGL-GCH1–321 to generate pGL-GCH1–321-Δ. The luciferase activity of this plasmid was sharply abolished, regardless of coincubation with CTRP13 (Fig. 5F), demonstrating that the PPARα binding site was crucial for CTRP13 regulation on the GCH1 promoter activity. We also tested whether CTRP13 treatment disrupted PPARα binding to this GCH1 promoter region. ChIP assays showed that HG treatment could inhibit PPARα binding directly to this region, but the binding affinity was recovered under CTRP13 treatment in a concentration-dependent manner (Fig. 5G). We also measured the expression of PPARα in HUVECs and found the protein levels of PPARα were comparable between the different groups (Fig. 5H). Collectively, these results indicate that CTRP13 promotes PPARα binding to the GCH1 promoter and accelerates GCH1 transactivation, leading to increased expression of GCH1.

To verify whether PPARα mediated the effect of CTRP13 on EDR, we compared the responses of the thoracic aortae isolated from PPARα^−/− mice and their wild-type (WT) littermates. As depicted in Fig. 6A, in aortae from WT littermates, HG-impaired vasorelaxation was prevented by CTRP13. However, this beneficial effect was lost in the aortae from PPARα^−/− mice. SNP-induced relaxations were similar among the different groups (Supplementary Fig. 2). The effects of PPARα deficiency on CTRP13-recoupled eNOS in vivo were also examined. As indicated in Fig. 6B–E, the increase in BH4 biosynthesis and the ratio of eNOS dimers to monomers, together with elevated serum NO levels and superoxide production, due to CTRP13 supplement, were all reversed on PPARα deletion, further suggesting the central role of PPARα in CTRP13-mediated eNOS and NO production.

How the adipokine CTRP13 regulated the transcriptional factor PPARα activity still needs exploration. Multiple research studies have reported that PPARα activity is increased by PKA phosphorylation directly (29,30). We first detected PKA activity in HG-treated HUVECs with or without CTRP13 treatment. HG treatment dramatically reduced PKA activity, while CTRP13 supplementation recovered the suppression of PKA activity (Fig. 6F). Using coimmunoprecipitation, we also found that CTRP13 could promote the phosphorylation of PPARα. H89, a specific inhibitor of PKA, abrogated the phosphorylation of PPARα by CTRP13 treatment (Fig. 6G). In addition, H89 decreased CTRP13-induced GCH1 promoter activity and BH4 levels in HUVECs (Fig. 6H and I). More interestingly, the beneficial effect on hyperglycemia-impaired ACh-induced EDR by CTRP13 was also abolished by treatment with the H89 compound (Fig. 6J). These data suggested that blocking PKA/PPARα leads to decreased GCH1 expression, BH4 production, and eNOS uncoupling, thus disrupting the beneficial effects of CTRP13.

The level of ACh-induced EDR was dramatically lower in renal arteries from patients with diabetes than in those from subjects without diabetes (Fig. 7A), while the arteries from both groups displayed identical levels of relaxation induced by SNP (Supplementary Fig. 3). We also analyzed the BH4 level in these renal arteries. Compared with subjects without diabetes, patients with diabetes showed decreased BH4 levels, and recombinant CTRP13 incubation restored the BH4 levels in a dose-dependent manner (Fig. 7B). Consistently, CTRP13 treatment reversed the downregulation of GCH1 and the ratio of eNOS dimers to monomers in the renal arteries of patients with diabetes (Fig. 7C).

The current study discovered a critical role for the CTRP13-GCH1-eNOS axis in protecting the endothelial function of db/db and STZ-induced diabetic mice. The supplementation of CTRP13 attenuates oxidative stress and increases NO bioavailability by improving eNOS coupling, thus counteracting the deleterious effect of EDR in the arteries of diabetic mice. Importantly, ex vivo CTRP13 treatment preserves endothelial function in the arteries of patients with diabetes.

CTRP13, an adipose tissue–derived adipokine, has been reported to decrease blood glucose levels and to improve glucose metabolism via activating the AMPK signaling pathway (15). CTRP13 also improves lipid-induced insulin resistance in hepatocytes by inhibiting stress-activated protein kinase/c-Jun N-terminal kinase stress signaling (15). Recently, our group demonstrated that CTRP13 has multiple beneficial effects on cardiovascular disease. In chronic coronary artery disease, serum CTRP13 is significantly downregulated, and ectopic CTRP13 infusion dramatically reduces the atherosclerotic progress, resulting in fewer foam cells, trapped macrophages, and inflammatory reactions (17). Furthermore, in chronic kidney disease, CTRP13 supplementation can inhibit vascular smooth muscle cell osteogenic differentiation and subsequent vascular calcification (15). Interestingly, in the current study, we found that CTRP13 infusion could improve endothelial eNOS coupling and NO production to protect endothelial function in diabetic mice and human patients with diabetes. Moreover, in addition to CTRP13, consistent with previous reports, other members of the CTRP family, such as CTRP3, CTRP5, and CTRP9 (Supplementary Fig. 4), could also cause significant vasorelaxation by endothelium-dependent NO release. Whether a mixture of CTRPs has important therapeutic value in the treatment of endothelial dysfunction caused by multiple clinical diseases is an interesting topic for future investigations.

BH4 biosynthesis is catalyzed by GTP cyclohydrolase I, encoded by GCH1 (28). Increasing evidence suggests that expression of GCH1 is a key determinant of endothelial cell BH4 levels and endothelial NOS regulation (31,32). Transgenic overexpression of GCH1 in mice is able to improve endothelial function in vascular disease states, such as atherosclerosis, diabetes, and pulmonary hypertension (33). Zhao et al. (34) reported that GCH1 protein can be degraded via ubiquitin-dependent proteasomes in endothelial cells and that AMPK activation inhibits proteasome activity in endothelial cells. In addition, Li et al. (35) found that hyperlipidemia- or hyperglycemia-induced ectopic miR-133a expression in the vascular endothelium reduced GCH1 protein and BH4 levels and impaired endothelial function. However, in this study, we further demonstrated that in endothelial cells, GCH1 could be regulated by CTRP13 treatment at the transcriptional level. More importantly, we found that CTRP13 promotes PPARα binding to the GCH1 promoter and accelerates GCH1 transactivation, thus leading to increased expression of GCH1. Further experiments are still needed to explore the regulatory process of GCH1, which could provide more possibilities for therapies for endothelial dysfunction.

PPARα activates the transcription of multiple genes involved in lipid metabolism (36). Recent studies have shown that PPARα activators can upregulate eNOS expression mainly by stabilizing eNOS mRNA to protect endothelial function (37). PPARα-deficient mice display impaired endothelial function caused by decreased NO bioavailability (38). Interestingly, in this study, we discovered that PPARα could mediate CTRP13-induced endothelial eNOS coupling and NO production by enhancing GCH1 transactivation. We also found that CTRP13 activated PKA to increase the activation of the PPARα/GCH1 axis. Previous data have shown that PPAR can be phosphorylated by PKA, in which PKA stabilizes the binding of liganded PPAR to DNA and increases PPAR activity (30). Activation of adenylyl cyclase catalyzes the conversion of ATP to cAMP and leads to the activation of PKA signaling (39). cAMP PDE converts cAMP into AMP by breaking the phosphodiester bond, in turn reducing the cAMP levels. Therefore, we went on detecting the expressions of these key genes involved in cAMP equilibrium, which affects PKA activities at the chronic state. CTRP13 treatment could abrogate HG-induced upregulation of PDE4A (cAMP-specific 3′,5′-cyclic PDE4A) and PDE4B (cAMP-specific 3′,5′-cyclic PDE4B) expression (Supplementary Fig. 5), suggesting the possible role of PDE4A, PDE4B, and cAMP in PKA activation under CRTP13 treatment. Therefore, it is also valuable to study how CTRP13 directly or indirectly alters PDE4 expression in endothelial cells in the future.

HG could inhibit the mRNA level of GCH1 to reduce NO production (35), and GCH1 protein can be degraded via ubiquitin-dependent proteasomes in endothelial cells in diabetes (34). While our results showed CTRP13 promoted GCH1 expression via PKA/PPARα at the transcriptional level, it is unclear whether it is the pathway PKA/PPARα that is attributed to the reduced GCH1 expression under HG. Decreased PKA activity was found under HG treatment. When we applied the activated PKA adenovirus or PPARα agonist fenofibrate, damaged ACh-induced EDRs due to HG were all improved (Supplementary Fig. 6), indicating the potential role of PKA/PPARα in HG conditions on GCH1 expression.

In conclusion, our study provides the first line of evidence for the beneficial effects of CTRP13 on endothelial function in arteries from diabetic mice and patients with diabetes. CTRP13 improves endothelial NO production and normalizes excessive ROS levels by enhancing GCH1-dependent eNOS coupling. Thus, targeting the CTRP13–GCH1 cascade can be an alternative therapeutic strategy for ameliorating diabetic vasculopathy.

Funding. This work was supported by the National Natural Science Foundation of China (81170239, 81900268, and 81970257), the major key technology research project of the Science and Technology Innovation Plan of Hubei Province (2016ACA151), key projects of Huazhong University of Science and Technology (2016JCTD107), and the Ministry of Science and Technology of the People’s Republic of China (2016 YFA0101100).

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

Author Contributions. C.W., Y.C., W.X., M.L., S.D., D.Z., and K.H. researched data and approved the manuscript. C.W. and W.X. designed the study and wrote the manuscript. Y.C. and W.X. analyzed and interpreted the data. M.L., S.D., D.Z., and K.H. contributed to the discussion and reviewed and edited the manuscript. C.W. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.