Endothelial-Specific Expression of CIDEC Improves High-Fat Diet–Induced Vascular and Metabolic Dysfunction

Vascular endothelial cells (ECs) constitute the innermost lining of blood vessels and play a major role in the regulation of vascular tone, inflammation, thrombosis, and delivery of oxygen and nutrients to tissues (1). Emerging evidence suggests that ECs also have pivotal functions in regulating whole-body metabolic homeostasis (2–4). Endothelial dysfunction has been implicated in the development and progression of insulin resistance, diabetes, cardiovascular disease, and other metabolic conditions, but mechanisms are not well known (4).

Insulin sensitivity in humans has been shown in multiple studies to have a consistent and direct correlation with the expression of cell death-inducing DNA fragmentation factor-α–like effector C (CIDEC), also known as fat-specific protein 27 (FSP27) (5–10). Initially identified as a lipid droplet-associated protein regulating lipolysis and lipid homeostasis in adipocytes (11,12), we recently demonstrated that it is also expressed in ECs and regulates arteriolar vasomotor and angiogenesis in human adipose tissue (6). In addition, we observed that CIDEC is downregulated in association with visceral obesity as an important cardiovascular risk factor.

A major regulator of angiogenesis includes vascular endothelial growth factor A (VEGFA), which binds and activates its tyrosine kinase receptors VEGFR1 and VEGFR2 (13). The primary angiogenic signal is mediated by activation of VEGFR2 (13). Overexpression of VEGFA in adipose tissue has been shown to increase local vascularization and protect mice against high-fat diet (HFD)–induced obesity and insulin resistance (14). Conversely, pharmacologic or genetic depletion of VEGFR2 results in metabolic impairment and insulin resistance (15). VEGFA also activates endothelial nitric oxide synthase (eNOS) and nitric oxide production, which mediate vascular proliferation and vasorelaxation, and we have previously described a regulatory interaction between CIDEC and VEGFA in ex vivo studies involving human adipose tissue biospecimens (6).

The aim of the current study was to characterize the whole-body physiological effects and potential mechanisms of action of EC-specific expression of human CIDEC in the regulation of vascular function and metabolism using in vitro and in vivo models. Mouse and human CIDEC isoforms can differ in their cellular effects with respect to insulin signaling and glucose homeostasis (9). Therefore, for the highest translational impact, we generated a humanized mouse model, E-CIDECtg, to our knowledge not previously described, with expression of the human CIDEC transgene specifically in ECs without altering endogenous mouse CIDEC expression. Using this novel mouse construct, we performed comprehensive physiological and molecular assessments characterizing the role of endothelial CIDEC in the regulation of vascular and metabolic phenotypes.

All animal experiments were performed in accordance with procedures approved by the Institutional Animal Care and Use Committees of Ohio University and the University of Kentucky. For the generation of E-CIDECtg mice, we first generated floxed mice that expressed human CIDEC transgene. These mice were generated with a Rosa26 knock-in system. Human CIDEC was cloned in a ROSA26-cytomegalovirus (CMV)-loxSTOPlox vector, and the construct was sent to the Beth Israel Deaconess Medical Center Transgenic Core Facility for conventional injections of embryonic stem cells into blastocyst hosts to produce F0 generation chimera (16). Floxed mice were backcrossed with C57BL/6 mice for six generations to make them syngeneic for future experiments. The C57BL/6 offspring were genotyped, and mice expressing human CIDEC specifically in ECs were generated by crossing CIDEC-floxed mice with mice expressing Cre recombinase under the control of endothelial-specific receptor tyrosine kinase promoter (B6.Cg-Tg[Tek-cre]1Ywa/J;[Tek/Tie 2 cre]; The Jackson Laboratory).

Animals were housed at the Ohio University Animal Facility (Athens, OH) with access to mouse chow and water ad libitum and an automatic 12-h day/night cycle. At 3 months of age, E-CIDECtg mice and floxed littermate controls (FLCs) were fed a standard/regular chow diet (5% fat, 22% protein; Prolab RMH 3000) or HFD (60% fat, 18% protein; Envigo TD.06414) for 12 weeks. Body weight was measured weekly during this 12-week period. After the 12 weeks, body composition was determined by nuclear magnetic resonance analysis, and the mice were fasted overnight and then pulsed, or not, with insulin (0.75 IU/kg body wt) prior to being sacrificed. Mice were euthanized by CO₂ asphyxiation, followed by cervical dislocation. Blood and organs were collected immediately after euthanasia for further analysis. Metabolic profiling was performed in these mice following the detailed methods provided in Supplementary Methods.

Genomic DNA isolation, genotyping, RNA isolation, RT-PCR, and quantitative (q)PCR analysis are detailed in Supplementary Methods.

Details are provided in Supplementary Methods. Total cholesterol quantification, assessment of serum triglycerides, serum nonesterified fatty acid (NEFA), HDL, LDL estimation, insulin, leptin, and adiponectin measurement are detailed in the Supplementary Methods.

Details are provided in the Supplementary Methods.

Microvascular adipose staining was performed as published previously (17). Mice were perfused with PBS, and epididymal adipose tissue was isolated. The tissue was cut into 2- to 3-mm pieces and fixed in 4% paraformaldehyde for 4 h at 4°C. Tissues were washed with PBS with 0.3% Triton X-100 (PBST) and blocked with 3% BSA in PBST for 1 h at room temperature. Blocking solution was removed, and tissues were incubated with rabbit anti-mouse CD31 (1:200) (Abcam cat no. ab28364) overnight at 4°C. The next day, tissues were washed with PBST three times and incubated with Alexa 488 (1:400) for 2 h at room temperature. The tissues were washed again three times with PBST and mounted on slides with Silly Putty and fluorescent mount mix. The slides were imaged by confocal microscopy; microvasculature was quantified in ImageJ.

Detailed methodology is provided in the Supplementary Methods.

As described previously in detail (18), siRNA-transfected human umbilical vein ECs (HUVECs) were trypsinized, resuspended in EC growth media, and plated (∼13,000 cells in 100 µL) on a prechilled 96-well plate coated with 65 µL of Matrigel (Corning, cat no. 354230). Tube formation was imaged at 0, 1.5, and 3.0 h with a Nikon upright microscope. Tube length was quantified using ImageJ software.

Thoracic aortic rings were isolated and cleaned from perivascular adipose tissue, as previously reported (19), and carefully mounted in a wire myograph (Model 620M, Danish Myo Technology). Briefly, preload tension was fixed to 9.8 mN in Krebs solution (mmol/L: NaCl, 118.4; KCl, 4.7; KH₂PO₄, 1.2; MgSO₄, 1.2; NaHCO₃, 25.0; glucose, 11.6; and CaCl₂, 1.9). Following preconstriction with phenylephrine (PE), cumulative concentration response (CCR) curves to acetylcholine (1 × 10⁻⁹ to 1 × 10⁻⁵ mol/L) were performed. After recovery, the CCR curve to endothelium-independent vasodilator sodium nitroprusside (SNP; 1 × 10⁻¹⁰ mol/L to 1 × 10⁻⁵–1 × 10⁻¹⁰ mol/L) was assessed in the same aortic segments. Vasorelaxation data are presented as percentage relaxation from PE-induced constriction. For all CCR curves, maximum relaxation from baseline was calculated using a nonlinear curve for the log of the concentration versus the response on a variable slope (four parameters analysis).

Details are provided in the legends and Supplementary Methods.

The data generated during the current study are available from the corresponding author upon reasonable request.

While initial studies identified CIDEC as an adipocyte-specific protein that regulates lipolysis and insulin sensitivity (5–10), we discovered that it is highly expressed in human ECs and regulates microvascular function in humans (6). To characterize physiological mechanisms, we generated floxed humanized mice expressing the human CIDEC transgene. To generate human CIDEC-expressing floxed mice, a floxed STOP signal, followed by the human CIDEC transgene expression cassette, was inserted into the mouse ROSA26 locus (see Research Design and Methods). The CIDEC-floxed mice were crossed with Tek-Cre mice to generate EC-specific CIDEC transgenic mice, termed E-CIDECtg mice (Fig. 1A). The progenies were genotyped by PCR using primers that targeted human CIDEC (450 base pairs [bp]), Tek Cre recombinase (100 bp), and endogenous mouse Cidec (250 bp) (Fig. 1B and Supplementary Table 1). Transgenic mice were born in expected allele frequencies and did not show any abnormal phenotype in viability, development, fertility, or breeding. CIDEC gene expression in ECs was measured (Fig. 1C). To confirm that the expression of CIDEC was restricted to ECs, we analyzed CIDEC mRNA levels in multiple cell types (adipocytes, hepatocytes, myoblasts, and ECs) from E-CIDECtg mice (Fig. 1D). Western blot analysis (with antibodies that bind both human and mouse isoforms) revealed an approximately twofold increase in combined CIDEC (human) and Cidec (mouse) protein levels in ECs, suggesting that human CIDEC expression was equivalent to the expression level of endogenous mouse Cidec (Fig. 1E). The littermate-floxed mice expressing CIDEC but not Tek-Cre were used as floxed controls (FLCs).

We studied the effect of HFD on glucose and insulin tolerance in E-CIDECtg mice. At 3 months of age, E-CIDECtg and FLC mice were divided into chow and HFD groups, and the respective diets were fed for 12 weeks. Under each dietary condition, the rate of body weight gain was similar between FLC and E-CIDECtg mice (Fig. 2A). Body composition analysis indicated that E-CIDECtg mice and FLC mice had similar percentages of lean and fat mass (Fig. 2B and C). Likewise, the weights of organs (liver, heart, epididymal fat pads, subcutaneous fat, brown adipose tissue) did not differ between E-CIDECtg mice and FLC mice (data not shown). No food or energy intake differences were observed between FLC and E-CIDECtg mice on either diet (Fig. 2D and E). HFD-fed E-CIDECtg mice were found to have lower fasting blood glucose levels compared with FLCs (Fig. 2F), evidencing a protective effect of EC-expressed CIDEC against HFD-impaired glucose homeostasis. Consistent with these results, E-CIDECtg mice showed higher glucose tolerance compared with FLCs under both chow- and HFD-fed conditions (Fig. 2G and H). Furthermore, E-CIDECtg mice retained significantly lower fasting blood insulin levels in the HFD-fed group (Fig. 2I). Insulin tolerance test (ITT) demonstrated enhanced insulin sensitivity in E-CIDECtg mice (Fig. 2J and K). Interestingly, in HFD-fed E-CIDECtg mice, glucose levels dropped at a faster rate after insulin injection and recovered similarly to chow-fed FLC and E-CIDECtg mice, indicating normal gluconeogenesis. Overall, these data showed that E-CIDECtg mice were protected against HFD-impaired glucose homeostasis.

Compared with FLCs, E-CIDECtg mice exhibited significantly reduced circulating triglycerides under chow and HFD conditions (Fig. 3A). However, NEFA levels remained similar between E-CIDECtg and FLC mice on both diets (Fig. 3B). Total cholesterol and LDL levels in E-CIDECtg mice remained comparable to those of FLC mice under the chow diet condition but were reduced compared with FLCs under the HFD condition (Fig. 3C and D). HDL remained similar between E-CIDECtg and FLC mice in both chow- and HFD-fed groups (Fig. 3E).

ECs regulate vascular homeostasis in adipose tissue (3,14,20), and obesity is characterized by impaired endothelial function and insulin signaling in fat compartments (21). Therefore, we examined the activation of Akt and eNOS as indices of insulin responsiveness in perigonadal adipose tissue. Under HFD conditions, insulin-mediated activation of Akt and eNOS was greater in E-CIDECtg mice compared with FLCs (Fig. 3F–H). Total Akt normalized to GAPDH did not show significant difference in floxed versus E-CIDECtg mice. Similarly, insulin-mediated Akt activation was also higher in the skeletal muscle and liver of HFD-fed E-CIDECtg mice (Supplementary Fig. 1). We further confirmed insulin-stimulated eNOS activation in the muscles, where eNOS-mediated angiogenesis plays a significant role in metabolism, endocrine function, and locomotion (22). Insulin-stimulated eNOS activation was significantly higher in the skeletal muscle of HFD-fed E-CIDECtg mice compared with the FLCs (Supplementary Fig. 2). Adiponectin, a cardioprotective adipocyte-specific glucose regulatory protein, was higher, and leptin, an energy balance-regulating hormone, was lower in E-CIDECtg mice versus FLCs (Fig. 3I and J).

Since CIDEC expression increased endothelial eNOS activation in E-CIDECtg mice, we probed for its molecular mechanism of action. As VEGFA signaling through VEGFR2 is the major pathway regulating endothelial responses (13), we first studied the localization of CIDEC, VEGFA, and VEGFR2 in VEGFA-stimulated HUVECs. Careful analysis of thin 0.2-µm Z sections of confocal images suggested that all three proteins, CIDEC, VEGFA, and VEGFR2, colocalize (Fig. 4A and B). These results complement our prior observation of direct interaction between CIDEC and VEGFA (6). VEGFA-induced autophosphorylation of the tyrosine residue 1175 (Y1175) of VEGFR2 is crucial for VEGFA-dependent downstream signaling (13). We observed that VEGFR2 pY1175 coimmunoprecipitates CIDEC (Fig. 4C), suggesting that even after VEGFA-mediated activation of VEGFR2, CIDEC remained associated with activated VEGFR2. These results were confirmed by reverse immunoprecipitation using CIDEC antibodies and immunolabeling with VEGFR2 pY1175 and VEGFR2 antibodies (Supplementary Fig. 3A and B). Also, VEGFR2 antibodies immunoprecipitated CIDEC (Supplementary Fig. 3C).

We observed that CIDEC expression increased VEGFA mRNA levels but did not alter VEGFR2 transcripts (Fig. 5A); however, protein expression of both VEGFA and VEGFR2 was increased (Fig. 5B). This prompted us to study the role of CIDEC in the stability of these proteins. Our recent study demonstrated the efficacy of recombinant CIDEC (rCIDEC) in regulating vascular function (6), where we showed that rCIDEC enhanced both eNOS and Akt activation in visceral adipose tissue of obese subjects (6). We also showed that the treatment of adipose arterioles isolated from these individuals had a profound effect on restoring insulin-mediated vasodilation (6). Therefore, we used rCIDEC to study the effects of CIDEC on VEGFR2 stability in the presence of the protein synthesis inhibitor cycloheximide (CHX). Consistent with the published 70-min half-life of VEGFR2 (23,24), VEGFR2 was reduced in HUVECs subjected to 25 µg/mL CHX for 1.5 h (Fig. 5C). However, the addition of rCIDEC prior to CHX stabilized VEGFR2 in a dose-dependent manner, with stabilization being achieved following the addition of 10 nmol/L rCIDEC (Fig. 5C). Protein degradation assay for VEGFA identified a significant reduction in VEGFA expression with CHX treatment. Similar to VEGFR2, adding rCIDEC prior to CHX stabilized VEGFA in a dose-dependent manner (Fig. 5D). Similar results were obtained upon adenovirus-mediated CIDEC expression in the above experiments (data not shown).

To validate the physiological relevance of our findings, we isolated ECs from the hearts of E-CIDECtg mice and FLCs. Consistent with our in vitro results, VEGFA and VEGFR2 protein levels were significantly higher in primary ECs from E-CIDECtg mice (Fig. 5E), confirming CIDEC’s role in VEGFA and VEGFR2 stabilization. Furthermore, primary ECs from E-CIDECtg mice showed greater phosphorylation of VEGFR2 and Akt than ECs from FLCs (Fig. 5F).

We hypothesized that elevated eNOS activation and insulin signaling may spur enhanced angiogenic capacity that is known to be impaired in obesity (6,25,26). To test our hypothesis, we performed ex vivo angiogenesis assays (27) with adipose explants from E-CIDECtg mice and FLCs. Perigonadal adipose tissue explants from E-CIDECtg mice showed enhanced capillary sprouting compared with FLC explants (Fig. 6A and B). We thus hypothesized that CIDEC expression may enhance vascular remodeling and, consequently, tissue perfusion in E-CIDECtg mice. CIDEC expression in the adipose tissue lysate was higher in the E-CIDECtg mice (Supplementary Fig. 4). Whole-mount CD31-stained adipose tissue from HFD-fed E-CIDECtg mice exhibited strikingly higher tissue microvascularization compared with FLCs (Fig. 6C and D), affirming proangiogenic effects of CIDEC in adipose microenvironments. Additionally, immunolabeling also demonstrated enhanced CD31 expression in skeletal muscle, white adipose, and heart tissue lysates from HFD-fed E-CIDECtg mice, relative to FLCs (Fig. 6E), consistent with a generalized whole-body effect of endothelial CIDEC in modulating vascularity.

To assess the regulatory role of the CIDEC-VEGFA interplay in angiogenesis, we performed gain- and loss-of-function in vitro assays examining capillary tube formation on a basement membrane matrix (18). siRNA-mediated depletion of CIDEC reduced HUVEC tube formation (Fig. 7A and Supplementary Fig. 5A and B). Analysis of microscopy images revealed that although HUVECs with CIDEC-knockdown showed migration, they did not coalesce into organized tubular structures and some cells remained solitary (Fig. 7A). Similarly, and expectedly, siRNA-mediated knockdown of VEGFA impaired HUVEC tube formation (Fig. 7B and Supplementary Fig. 5C). However, the addition of rCIDEC rescued the angiogenic capacity of VEGFA-depleted HUVECs (Fig. 7B), suggesting that CIDEC can induce capillary-network development even under reduced VEGFA conditions.

We tested the effects of CIDEC expression on arterial vasodilation in chow- and HFD-fed E-CIDECtg and FLC mice. Basal vascular reactivity of E-CIDECtg and FLCs mice fed regular chow were not statistically different. However, HFD feeding impaired acetylcholine-mediated relaxation of aortic rings from FLCs, while rings from E-CIDECtg mice fed HFD displayed significant protection in endothelium-dependent acetylcholine-mediated relaxation (28) compared with those from FLCs (Fig. 7C). In addition, we observed that aortic ring relaxation in response to endothelium-independent vasodilator SNP was identical between E-CIDECtg and FLC groups regardless of dietary conditions (Fig. 7D), demonstrating that the improved vasorelaxation effect was endothelial specific.

The endothelium plays a significant role in the regulation of whole-body glucose homeostasis as glucose is a primary fuel source in ECs (29). During angiogenesis, glucose uptake is increased in ECs to meet the energy demands (30), and VEGFA activation elevates glycolytic flux to furnish increased energy expenditure. We hypothesized that CIDEC-mediated increased EC function would increase energy expenditure in E-CIDECtg mice. Using indirect calorimetry, we detected Vo₂, Vco₂, and energy expenditure in the chow- and HFD-fed FLC and E-CIDECtg mice during light and dark cycles. Examination of metabolic outcomes in individually housed metabolic cages showed that HFD-fed, but not chow-fed, E-CIDECtg mice had greater Vo₂ and Vco₂ (Fig. 8A–F) than FLCs. As expected, energy expenditure was significantly decreased in FLCs upon HFD feeding. However, HFD-fed E-CIDECtg mice showed significantly higher energy expenditure compared with the FLCs (Fig. 8B, D, and F). The expression of various mitochondrial genes associated with energy expression was increased in the adipose tissue depots of E-CIDECtg mice (Supplementary Fig. 6). Note that no food or energy intake differences were observed among the FLC and E-CIDECtg mice on either diet (Fig. 2D and E). These data demonstrate that E-CIDECtg mice exhibit higher catabolic rates.

In the current study, we generated a novel E-CIDECtg mouse model to characterize the physiological regulation and potential mechanisms of CIDEC control of endothelial function and whole-body glucose homeostasis. E-CIDECtg mice expressing human CIDEC specifically in ECs were protected against HFD-induced glucose intolerance and displayed lower serum triglycerides, LDL, and total cholesterol levels. These findings are significant because they suggest that CIDEC may protect against a number of traditional cardiovascular risk factors, particularly in the setting of obesogenic stress. Moreover, our data show that the CIDEC mechanism of action is, in part, driven by VEGFA/VEGFR2 stabilization and enhanced downstream signaling that promotes proangiogenic and vasodilatory responses. Overall, our study provides, to our knowledge for the first time, evidence of a systemic protective effect of EC CIDEC overexpression against obesity-induced vascular and metabolic derangements, which prompts translational recognition of CIDEC as a potential therapeutic target (4,25,31,32).

Human and mouse orthologs of CIDEC show 90% sequence similarity (5). The adipose-specific roles of murine Cidec and human CIDEC have been found be similar in fat metabolism and insulin signaling (33,34). Interestingly, our recent study suggested a potential metabolic paradox in which whole-body Cidec-knockout mice presumed to be metabolically healthy based on glucose utilization and oxidative metabolism are dysfunctional in exercise capacity and muscular performance (9); thus, positively linking human and mouse CIDEC orthologs with metabolically healthier phenotypes. Furthermore, we recently demonstrated that CIDEC is also expressed in ECs and regulates arteriolar vasomotor and angiogenesis in human adipose tissue (6). Surprisingly, our present study shows that the expression of a single allele of human CIDEC in these constructs had a crucial effect on whole-body metabolism. This was evident at the tissue level, with improved insulin-stimulated Akt and eNOS activation in white adipose tissue, liver, and skeletal muscle of HFD-fed E-CIDECtg mice compared with FLCs. Increased adiponectin and reduced leptin levels in HFD-fed E-CIDECtg mice were also consistent with improved adipose tissue metabolism and suggest EC-adipocyte regulatory interplay (35). Although E-CIDECtg mice had lower leptin levels, there was no significant difference in food consumption, body weight, and fat mass in HFD-fed FLCs and E-CIDECtg mice. This suggests that E-CIDECtg mice may be leptin tolerant compared with the HFD-fed control mice, which develop leptin resistance resulting in higher leptin levels.

A notable finding in E-CIDECtg mice was enhanced vascularization and vasodilator function that can improve tissue perfusion and optimize nutrient delivery/exchange in target organs. This may be particularly clinically relevant since relative pseudohypoxia as a function of adipocyte hypertrophy and capillary rarefaction has been implicated in the pathophysiology of insulin resistance and metabolic diseases associated with obesity (36,37). The vascular effects of CIDEC appear to be driven, in part, by mechanisms that stimulate VEGFA signaling. Interestingly, VEGFA expression in white adipose tissue increases expression of thermogenic genes and heat production in mice (38). Our E-CIDECtg mice showing increased expression of thermogenic genes in various adipose depots, and an increased respiratory exchange ratio is consistent with these findings (38). VEGFA stimulation of VEGFR2 phosphorylation at Tyr1175 is an important step required for EC proliferation, migration, and tube formation (39,40). VEGFR2 activation phosphorylates Akt (39), which then activates eNOS to generate nitric oxide, a key regulator of vascular tone and angiogenesis (41). We demonstrated that CIDEC forms a triprotein complex with VEGFA and VEGFR2, and VEGFR2-pY1175 interaction with CIDEC informs us that VEGFR2 activation is likely simultaneous to the aggregation of VEGFA-VEGFR2-CIDEC complex. Additionally, we found that ECs from E-CIDECtg mice displayed enhanced VEGFA-pY1175 and downstream Akt activation, affirming physiological bioaction. In E-CIDECtg mice, we observed enhanced angiogenic capacity and increased vascularization in several tissues, including heart and skeletal muscle.

The in vitro experiments also suggested that CIDEC can rescue tube formation even in cells treated with VEGFA siRNA, perhaps by stabilizing residual VEGFA. Understanding VEGFA bioaction in disease states can be transformative, as defective VEGF signaling has been linked to vascular aging and perturbed microvascular homeostasis that are upstream drivers of multiorgan malfunction (42). Also, VEGF-inhibitor drugs used in oncology have been recently associated with cardiovascular toxicities, including hypertension and thromboembolic events (43).

Another key result of our study is the robust effect of the CIDEC transgene in improving whole-body glucose homeostasis in HFD-fed mice. The endothelium is the largest organ in the body (44) and consumes glucose as its fuel source for glycolysis (29). One possible mechanism to explain our findings may relate to increased glucose utilization owing to capillary growth and angiogenesis that depend on glycosylation (45,46). However, ECs also use fatty acids (FAs) as energy sources (47). VEGFA enhances FA uptake by upregulating the expression of the FA-binding protein (FABP4), a trafficking protein (48). Capillaries in FA-consuming tissues, such as heart and skeletal muscle, express FABP4 and FABP5 for endothelial transport and cellular energy production via FA oxidation. As such, our observation of increased energy expenditure along with a reduction in serum lipids in E-CIDECtg mice may be related to these physiological processes. Additionally, there is growing recognition that endothelial dysfunction itself can cause metabolic dysregulation, supported by several recent publications involving genetic experimental models. For example, manipulations in the expression of endothelial IGF-1 receptors modulate paracrine signaling that controls glucose uptake in muscle and fat primarily via Nox4-derived H₂O₂ (20). Also, deletion of Argonaute 1 in ECs, which has antiangiogenic actions via thrombospondin 1, improved vascularization and insulin sensitivity (3). Similarly, EC-specific deletion of FOXO1 increased adipose vascular density and protected mice from HFD-induced glycemic dysregulation (2). Our present study adds to the growing body of evidence that endothelial plasticity modulates systemic phenotypes and that strategies targeting a dysfunctional endothelium may serve as a potential therapeutic approach to mitigate obesity-related cardiometabolic disorders.

In our study, we found that improving VEGFA-VEGFR2 signaling protected against HFD-induced metabolic disorders. However, we acknowledge that VEGFA hyperstimulation may also associate with pathogenic angiogenesis seen in conditions such as retinopathy and cancer, where the role of CIDEC merits further investigation. Another potential limitation of our study is that TEK/Tie2 promoters, widely used as endothelial-specific promoters, can be expressed in a monocyte subset, which may have influenced the whole-body metabolic results in E-CIDECtg mice. Interestingly, the Tie2-expressing monocytes have been shown to be proangiogenic (49). However, our findings are affirmed by studies in ECs clearly showing the key role of CIDEC in regulating vascular function via its effect on VEGFA and VEGFR2 stabilization. Furthermore, our ex vivo analysis of aortic rings shows improved vasorelaxation in E-CIDECtg mice, again suggesting endothelium-specific actions of CIDEC. Nevertheless, it remains possible that CIDEC regulates EC function in a cell-autonomous manner via monocyte interactions, which merits further investigation.

In conclusion, our findings show that endothelial-specific expression of the human CIDEC transgene in mice showed protection against HFD-induced glucose intolerance and dyslipidemia. Additionally, we identified a novel pathway of action whereby CIDEC regulates the VEGFA-VEGFR2 axis promoting eNOS signaling, angiogenesis, and vasodilation. Our findings add to our understanding of the role of the vascular endothelium in regulating systemic metabolism and prompt recognition for potentially targeting the vasculature for the treatment of obesity-associated cardiometabolic disorders.

Acknowledgments. The authors acknowledge Dr. Shyam Bansal, from Ohio State University, for critical reading of the manuscript.

Funding. This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases grants R01DK124126 (S.M.N. and V.P.), R01HL142650 (N.G.), R01HL140836 (N.G. and V.P.), and R01DK101711 (V.P.); funds from Osteopathic Heritage Foundation’s Vision 2020 to Heritage College of Osteopathic Medicine at Ohio University (V.P.); and American Heart Association Postdoctoral Fellowship grant 19POST34430086 (B.B.).

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

Author Contributions. B.B. contributed to designing the study, performing experiments, analyzing results, and preparing the manuscript. A.G., R.B., V.M.S., M.S., K.G., C.C.B., and A.S.L. contributed to designing the study, performing experiments, and analyzing the data. S.K. and N.G. contributed to designing the study, analyzing the results, and editing the manuscript. H.M. and S.M.N. contributed to performing experiments. V.P. contributed to designing and overseeing the study, performing experiments, analyzing results, and preparing the manuscript. V.P. 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.