GLUT10 maintains the integrity of major arteries through regulation of redox homeostasis and mitochondrial function

Abstract

Glucose transporter 10 (GLUT10) is a member of the GLUT family of membrane transporters, and mutations in this gene cause arterial tortuosity syndrome (ATS). However, the physiological role and regulation of GLUT10 in arteries remains unclear. To further understand its physiological roles in major arteries, we examined the regulatory mechanisms of GLUT10 in ASMCs and aortic tissues. Interestingly, we find that targeting of GLUT10 to mitochondria is increased in ASMCs under both stress and aging conditions, which enhances dehydroascorbic acid (DHA) uptake and maintains intracellular ascorbic acid (AA) levels. We further demonstrate that the targeting of GLUT10 to mitochondria is important to maintain redox homeostasis, mitochondrial structure and mitochondrial function in ASMCs. A missense mutation of GLUT10 (Glut10^G128E) impairs mitochondrial targeting in ASMCs. Consequently, ASMCs isolated from Glut10^G128E mice exhibit increased reactive oxygen species (ROS) levels, fragmented mitochondria and impaired mitochondrial function, as well as enhanced cell proliferation and migration. In vivo, mitochondrial structure is altered, and ROS levels are heightened in aortic tissues of Glut10^G128E mice. Furthermore, increased number and disorganization of ASMCs, along with progressive arterial wall remodeling were observed in aortic tissues of Glut10^G128E mice. These defects were coincident with elevated systolic blood pressure in aged Glut10^G128E animals. Our results describe a novel mechanism that GLUT10 targeting to mitochondria under stress and aging condition has a critical role in maintaining AA levels, redox homeostasis and mitochondrial structure and function in ASMCs, which is likely to contribute to the maintenance of healthy vascular tissue.

Introduction

During the last two centuries, life expectancy has increased significantly in all industrialized countries, bringing about corresponding increases in age-related metabolic and degenerative diseases. Vascular complications are the most significant risks associated with cardiovascular diseases, and age-related morbidity and mortality (1). Accumulating evidence supports the concept that oxidative stress develops with age and contributes to the progression of vascular diseases (2,3). Reactive oxygen species (ROS) play a major role in mediating pathophysiological effects involved in the initiation and progression of cardiovascular dysfunction (4). On the other hand, ROS also act as essential signaling molecules and have physiological roles in normal vascular physiology (5). Therefore, creating a delicate balance between appropriate redox states and oxidative stress is dependent on tight homeostatic regulation of pro-oxidant and antioxidant activities.

Both enzymatic and non-enzymatic systems have evolved to regulate the redox state of cells (6). Synthetic and natural antioxidants have been intensively investigated to assess potential therapeutic or preventive effects in cardiovascular disease (7); however, the results of such studies have been inconclusive and often contradictory. These mixed results suggest that endogenous homeostatic mechanisms govern the balance of antioxidant activity, and therefore it is important to better understand the regulation of antioxidant mechanisms under physiological and pathological conditions. The transcription factor nuclear factor E2–related factor-2 (NRF2), has been demonstrated to respond to oxidative stress and orchestrate enzymatic antioxidant defenses (8). However, little is known about the regulation of non-enzymatic antioxidant systems and their roles in protecting the vascular system under physiological and pathophysiological conditions.

Mutations in SLC2A10, which encodes the GLUT family member GLUT10, cause arterial tortuosity syndrome (ATS; OMIM #208050) in humans. ATS is a recessive inherited disorder characterized by elongation, tortuosity, stenosis and aneurysms of large and medium sized arteries in association with craniofacial and connective tissue manifestations (9). We have previously demonstrated that GLUT10 is highly expressed in ASMCs and partially localizes to mitochondria. Moreover, we provided evidence showing that GLUT10 facilitates the transport of dehydroascorbic acid (DHA) into mitochondria and increases intracellular and mitochondrial ascorbic acid (AA), which then decreases intracellular ROS levels in ASMCs (10). In this study, we extend these findings and further explore the in vivo function of GLUT10 and the pathogenesis of GLUT10 deficiency.

We have previously generated a mouse model carrying the Glut10^G128E mutation. The main phenotypes, observed in homozygous mutant mice, are structural abnormalities in major arteries (11). An association between the DHA uptake by GLUT10 in ASMCs and GLUT10 deficient phenotypes in human and mouse has been made. Therefore, we used in vitro cell models and in vivo Glut10^G128E mice to further identify the potential underlying regulatory mechanisms of GLUT10 that contributes to maintain the structure integrity and function of aortic tissues.

Results

Increased GLUT10 mitochondrial targeting under oxidative stress and aging conditions

To understand the possible regulatory mechanism by which GLUT10 contributes to maintaining redox homeostasis, we investigated the regulation of GLUT10 under stress and aging conditions. Treating A10 cells, a rat ASMC cell line, with H₂O₂ increased intracellular and mitochondrial ROS levels in a concentration-dependent manner (Supplementary Material, Fig. S1A–C). Since GLUT10 facilitates DHA uptake into mitochondria, we examined GLUT10 mitochondrial localization under stress conditions. To our surprise, the proportion of mitochondria-localized GLUT10 was markedly increased after H₂O₂ treatment (Fig. 1A and C). Interestingly, the mitochondrial targeting ability of Glut10^G128E was diminished under both normal and stress conditions (Fig. 1B and C). Accordingly, the DHA uptake was increased (Fig. 1D) and intracellular AA levels (Fig. 1E) were maintained in A10 cells under stress conditions. These results suggest that increased mitochondrial targeting of GLUT10 under stress conditions may increase DHA uptake to maintain intracellular AA levels. The H₂O₂-induced increase in targeting of GLUT10 to mitochondria was further validated by subcellular fractionation (Supplementary Material, Fig. S2A–C) and western blot analysis (Supplementary Material, Fig. S2E) of Glut10-myc expressing MOVAS cells, a mouse ASMC line. After H₂O₂ treatment, GLUT10 accumulation was increased in the mitochondria enriched fraction of these cells (Supplementary Material, Fig. S2E and F).

Figure 1.

Enhanced GLUT10 targeting to mitochondria under oxidative stress increases DHA uptake and intracellular AA levels in A10 cells. Representative confocal images of (A) Glut10/EGFP, and (B) Glut10^G128E/EGFP-expressing A10 cells treated with 100 μM H₂O₂ or without H₂O₂ (Mock) for 24 h. EGFP signal (green color), MitoTracker Red signal (red color) labeled mitochondria, and merged co-localized signal (yellow color). Inserts show enlarged images of the indicated regions. (C) Quantitative data analysed by Imaris showing the percentage of Glut10/EGFP localized to the mitochondria. (D) DHA uptake abilities and (E) intracellular AA levels. A10 cells were treated with H₂O₂ (H₂O₂) or mock treatment and incubated in cultured medium (medium) or provided with 1 mM DHA for 30 min (DHA). Results are shown as mean ± SD. Two-tailed, unpaired Student’s t-test, *P < 0.05; **P < 0.01, ***P < 0.001.

We further validated the increase of GLUT10 mitochondrial targeting under stress conditions in human ASMCs. Similar to murine cells, targeting of endogenous GLUT10 to mitochondria was increased under H₂O₂ treatment in human ASMCs (Fig. 2A and C). Strikingly, we also observed increased intracellular and mitochondrial ROS levels (Fig. 2B) and increased GLUT10 mitochondrial targeting in aged human ASMCs (Fig. 2A and D). Furthermore, human ASMCs exhibited enhanced DHA uptake (Fig. 2E) and increased intracellular AA levels (Fig. 2F) under both H₂O₂ stress and aging conditions.

Figure 2.

Enhanced GLUT10 targeting to the mitochondria under oxidative stress or aging conditions increases DHA uptake and intracellular AA levels in human ASMCs. (A) Representative confocal images show the immunofluorescence staining of GLUT10 (green color), mitochondria (COXV, red color), and merged signal (yellow) in human ASMCs treated with 100 μM H₂O₂ (H₂O₂) or without H₂O₂ (Mock) for 24 h or at late passage (P10–14, aged). Inserts show enlarged images of the indicated regions. (B) The relative intracellular and mitochondrial ROS levels in H₂O₂ treated and aged human ASMC were determined by DCFH-DA or MitoSox Red, respectively. (n = 3). (C, D) Quantitative results analysed by Imaris showed increased percentage of GLUT10 localized to the mitochondria under H₂O₂ treatment (C) and in aged (D) human ASMCs. n, indicated the cell number analysed. (E, F) DHA uptake (E) and intracellular AA levels (F). Human ASMCs were treated with (H₂O₂) or without (Mock) H₂O₂ and incubated normal culture medium (medium) or provided with 1 mM DHA for 30 min (DHA). n = 3. Results are shown as mean ± SD. Two-tailed, unpaired Student’s t-test, *P < 0.05; **P < 0.01, ***P < 0.001.

Collectively, our results suggest that GLUT10 targeting to mitochondria was increased under stress and aging conditions in ASMCs, which enhanced DHA uptake and maintained intracellular AA levels.

Essential role of GLUT10 in maintaining mitochondrial morphology and function in ASMCs

To better understand the impact of GLUT10 on mitochondria, we first investigated the effects of GLUT10 deficiency on mitochondrial respiration. Primary ASMCs were obtained from Glut10^G128E and WT mice and were evaluated for ROS content, mitochondrial respiration and mitochondrial morphology. Glut10^G128E ASMCs exhibited higher intracellular and mitochondrial ROS levels compared with WT ASMCs. Under H₂O₂ treatment and aging conditions (as modeled by late passage cells), both intracellular and mitochondrial ROS levels were increased in WT ASMCs, and the increased levels were further augmented in Glut10^G128E ASMCs (Fig. 3A). The oxygen consumption rate (OCR) was significantly reduced in Glut10^G128E ASMCs, and further reduced in aged Glut10^G128E ASMCs when compared with the same passage of WT ASMCs (Fig. 3B). We also determined the mitochondrial morphology in WT and Glut10^G128E ASMCs. The percentage of small globe shaped mitochondria was increased in Glut10^G128E ASMCs compared with WT ASMCs (Fig. 3C and D). The mitochondrial morphology was highly fragmented in both WT and Glut10^G128E ASMCs under oxidative stress and aging conditions (Supplementary Material, Fig. S3). We further evaluated mitochondrial morphology and density in aortic tissues of Glut10^G128E and WT mice at 15 months of age by electron microscopy (Fig. 3E). Mitochondria were graded from 1 to 3, on the basis of cristae organization and swelling, with a higher grade representing a higher degree of aberration (Fig. 3F). The percentage of aberrant mitochondria was significantly increased in aortic tissues of 15 month-old Glut10^G128E mice (Fig. 3G). Also, the density of mitochondria was decreased significantly in aortic tissues of Glut10^G128E mice (Table 1). Decreased mitochondrial function and increased percentage of small globe shape mitochondria under oxidative stress conditions were also observed in A10 cells (Supplementary Material, Fig. S4) and in human ASMCs (Supplementary Material, Fig. S5A–C). Additionally, small mitochondria were significantly increased in aged human ASMCs compared with young cells (Supplementary Material, Fig. S5D).

Table 1.

Quantification of mitochondrial morphology in the thoracic aorta from Glut10^G128E mice (G128E) and wild type (WT) littermates at 10 months of age

Measurements	WT	G128E
Total number of mitochondriaa	207	115
Alteration grade	Number (%)
Grade 1	141 (68)	13 (11)
Grade 2	50 (24)	79 (69)
Grade 3	16 (8)	23 (20)

a

Total area measured: 54.0566 μm² surrounding the nuclei from eight different areas.

Figure 3.

GLUT10 mutation induces mitochondrial dysfunction, fragmentation, and structural abnormalities in ASMCs. (A) The relative intracellular and mitochondrial ROS levels in H₂O₂ treated and late passage (P8, aged) ASMCs, isolated from 6–8 week old Glut10^G128E and WT mice, were determined by DCFH-DA or MitoSox Red, respectively (n = 3). (B) Oxygen consumption rates (OCR) representing basal respiration (1), ATP production (2), maximal respiration (3), proton leak (4), and non-mitochondrial oxygen consumption (5) in Glut10^G128E and WT ASMCs were determined at early passage (P3, young) or late passage using the Seahorse XF24 extracellular flux analyzer. Oligomycin inhibits ATP synthase, carbonyl cyanide p-trifluoromethoxy-phenylhydrazone (FCCP) is an uncoupler, and antimycin A is an electron transport chain inhibitor. Significantly reduced basal respiration (1), ATP production (2), and maximal respiration (3) were observed in Glut10^G128E ASMCs compared with WT (n = 5). (C) Representative mitochondrial staining of ASMCs from 3-month-old WT and Glut10^G128Emice and Micro-P analyses of mitochondrial morphology (lower panels). (D) Quantitative results showed that the percentage of small globe mitochondria are increased and large mitochondria, including simple tube, branching tube, twisting tube, large globe, and donut mitochondria were significantly decreased in Glut10^G128E ASMCs (G128E) in compared with WT ASMCs (WT). (E) The ultrastructure of mitochondria from aortic tissues of 15-month-old Glut10^G128E (G128E) and WT mice was analysed by transmission electron microscopy. (F) Examples of three grades of mitochondrial morphology classified according to the structural abnormalities described in Methods. Grade 1: mitochondria with normal structure, Grade 2, mitochondria lack well-defined cristae, Grade 3, mitochondria lack well-defined cristae and exhibit extensive swelling. (G) The percentage of abnormal mitochondria (Grade 2 and Grade 3) is increased in aortic tissue from Glut10^G128E mice compared with WT mice (n = 8). The total area measured was 54.0566 μm² surrounding the nuclei. Results are shown as mean ± SD. Two-tailed, unpaired Student’s t-test, *P < 0.05; **P < 0.01, ***P < 0.001.

Impacts of GLUT10 on mitochondrial function and morphology were further examined in GLUT10-knockdown (GLUT10-KD) and GLUT10-overexpressing (GLUT10-OE) MOVAS cell lines (Supplementary Material, Fig. S6A). Decreasing the expression of GLUT10 in MOVAS cells significantly reduced the mitochondrial membrane potential (Supplementary Material, Fig. S6B) and OCR (Supplementary Material, Fig. S6C), and increased the number of small mitochondria (Supplementary Material, Fig. S6D). In contrast, overexpression of GLUT10 in MOVAS cells significantly increased mitochondrial membrane potential (Supplementary Material, Fig. S6B) and OCR (Supplementary Material, Fig. S6C), and decreased small globe shaped mitochondria (Supplementary Material, Fig. S6E). Taken together, these results suggest that GLUT10 function is important in maintaining mitochondrial morphology and function.

As mitochondrial dynamics regulate mitochondrial morphology and function, we investigated the distribution pattern of mitochondrial fission protein (FIS1) and mitochondrial fusion protein (MFN2) in aged human ASMCs. FIS1 was more associated with mitochondria in aged human ASMCs than in young cells (Fig. 4A and C). The association of MFN2 with mitochondria was not changed in aged human ASMCs (Fig. 4B and D). We also examined the distribution pattern of FIS1 and MFN2 in GLUT10-KD and GLUT10-OE MOVAS cells. The association of FIS1 with mitochondria was significantly increased in GLUT10-KD cells and reduced in GLUT10-OE cells (Supplementary Material, Fig. S7A and C). The association of MFN2 with mitochondria was decreased in both GLUT10-KD and GLUT10-OE MOVAS cells (Supplementary Material, Fig. S7B and D). These results suggest that the mitochondrial fragmentation in aged and GLUT10 deficient ASMCs is most likely accompanied by increased FIS1 association with mitochondria.

Figure 4.

FIS1 mediates mitochondrial shortening in aged human ASMCs. (A,B) Representative confocal images show the immunofluorescent staining of FIS1 (A) or MFN2 (B) (green), mitochondria (COXV, red), and merged colocalized signal (yellow color) in early passage (P4, young) and late passage (P14, aged) human ASMCs. Inserts show enlarged images of the indicated regions. (C,D) The quantification data show the percentage of mitochondria colocalized with FIS1 (C) or MFN2 (D) analysed by Imaris (n = 5). Results are shown as mean ± SD. Two-tailed, unpaired Student’s t-test, *P < 0.05; **P <0.01, ***P < 0.001.

Changes in arterial structure and function in Glut10^G128E mice

To explore the physiological role of GLUT10 mitochondrial targeting in major arteries, the aortic tissue of Glut10^G128E mice were compared with their WT littermates (WT) at 3, 5, 10 and 15 months of age. We observed slow progressive arterial wall changes in Glut10^G128E mice (Fig. 5A) and quantified the progressive thickening of irregular arterial walls (Fig. 5D). We also observed disorganization and random breaks in elastic fibers (Fig. 5B) in the descending aorta. Furthermore, we found an increased number and disorganization of ASMCs, which was initially observed in 5 months old Glut10^G128E mice, and became more apparent at 10 and 15 months (Fig. 5C).

Figure 5.

GLUT10 mutation induces age-dependent changes in arterial structure and vascular function. (A–C) Histopathology of the ascending aorta in Glut10^G128E mice and wild type (WT) littermates at 3, 5, 10, and 15 months of age. (A) Representative photomicrographs show that the ascending aorta from Glut10^G128E mice (Glut10^G128E) appeared with irregular morphology and uneven thickness. (B) Elastic fibers in the lamina interna and media were more disorganized and fragmented, with increased number in Glut10^G128E compared with WT mice. (C) An increased number and disorganization of ASMCs in Glut10^G128E compared with WT mice was observed. A and C were stained with hematoxylin and eosin. B was stained with Verhoeff-van Gieson for elastin. Scale bars are 200 µm in A and 50 µm in B and C. D. Quantification of the arterial wall thickness shows the progressive thickening and variability in arterial walls of Glut10^G128E mice. (E) Systolic blood pressure, (F) diastolic blood pressure, and G. pulse rate were determined in 3- and 15-month-old Glut10^G128 and WT mice. 3-month-old, (WT, n = 3, G128E, n = 5); 15-month-old, (WT, n = 5, G128E, n = 7). Results are shown as mean ± SD. Two-tailed, unpaired Student’s t-test, *P < 0.05 compared with WT.

Impacts of arterial changes on cardiovascular function were further assessed in Glut10^G128Emice. Systolic blood pressure was higher in Glut10^G128E compared with WT mice at 15 months of age (Fig. 5E), although the diastolic blood pressure and heart rate did not exhibit measurable differences between the two genotypes at either 3 or 15 months of age (Fig. 5F and G). Furthermore, no difference was observed at 15 months of age in any parameter measured by electrocardiogram (Supplementary Material, Table S1).

Increased ROS in Glut10^G128E ASMCs and aortic tissues of Glut10^G128E mice

In concert with impaired mitochondrial targeting of Glut10^G128E protein, altered mitochondrial structure, and impaired mitochondrial function of Glut10^G128E ASMCs, intracellular and mitochondrial ROS levels were increased in Glut10^G128E ASMCs under normal, oxidative stress and aging conditions (Fig. 3A and Supplementary Material, Fig. S8A). Furthermore, mitochondrial abnormality was also observed in Glut10^G128E ASMCs in aortic tissues. To better determine the physiological consequences of impaired GLUT10 in vivo, we determined ROS levels in the aorta of Glut10^G128E and WT mice. ROS levels were increased in different regions of Glut10^G128E aorta compared with WT (Fig. 6A and Supplementary Material, Fig. S8B and C). Furthermore, oxidative damage to protein was increased in ASMCs and aortic tissues of Glut10^G128E mice (Fig. 6B) as determined by the presence of protein carbonyls (12). Accordingly, we examined the mRNA expression of cytosolic antioxidant enzymes, SOD1 (Fig. 6C), glutathione peroxidase (GPX) (Fig. 6E), and catalase (CAT) (Fig. 6F), and found that they were not different between the two genotypes. In contrast, the mRNA levels of superoxide dismutase 2 (SOD2; a mitochondrial antioxidant enzyme) were upregulated in Glut10^G128E aortic tissues as a compensatory effect (Fig. 6D). The protein levels of SOD2 were examined by immunostaining and western blot (Fig. 6G and 7K), showing that the protein expression of SOD2 was also increased. We further examined the overall oxidation of plasma proteins and antioxidant activity in plasma. Plasma protein oxidation was increased at 1 month (Supplementary Material, Fig. S8D), and antioxidant activity in plasma was increased at 2 and 3 months of age in Glut10^G128E mice (Supplementary Material, Fig. S8E). ROS can be produced in vascular cells by the increased expression of a number of oxidases, including NAPDH oxidase (NOX) enzymes, endothelial nitric oxide synthase (eNOS), and inducible NOS (iNOS). There was, however, no significant difference in the mRNA expression levels of NOX1, eNOS or iNOS between aortic tissues of Glut10^G128E and WT (Supplementary Material, Fig. S9A–C). These results reveal that GLUT10 deficiency increases oxidative stress in ASMCs, aortic tissues, and plasma of Glut10^G128Emice.

Figure 6.

GLUT10 mutation increases ROS, protein oxidation, and antioxidant defenses in aortic tissues. (A) ROS (red) in outer arch of aortic tissues from Glut10^G128E (G128E) mice were markedly increased compared with WT mice, as detected by in situ dihydroethidium (DHE) fluorescence staining. Green is autofluorescence from elastin fibers. Scale bar: 50 µm. (B) Western blots of protein carbonyls showing that oxidized proteins are increased in Glut10^G128E ASMCs and aortic tissues from 5-month-old Glut10^G128E (G128E) compared with WT mice. mRNA levels of (C), SOD1; (D), SOD2; (E) GPX; and (F), CAT in aortic tissues of 3-, 5-, and 10-month-old Glut10^G128E and WT mice (n = 5). G. SOD2 staining is increased in ascending aortic tissues of Glut10^G128E compared with WT mice. Scale bars = 50 µm. The results are shown as mean ± SD. Two-tailed, unpaired Student’s t-test, *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 7.

GLUT10 deficiency induces ASMC proliferation and migration. (A) The proliferation of Glut10^G128E ASMC is increased compared with WT ASMCs (n = 3). (B) Glut10^G128E ASMCs have enhanced cell migration and wound closure ability compared with WT ASMCs. Supplementation of the culture medium with ascorbic acid (AA) reduces the wound closure ability in both Glut10^G128E and WT ASMCs. (C) PDGF-B mRNA expression is elevated in Glut10^G128E ASMCs compared with WT ASMCs (n = 3). (D) The expression of PDGF-B mRNA is increased in aortic tissue of 5-month-old Glut10^G128E compared with WT mice (n = 5). (E) Cyclin D1 staining is increased in the ascending aorta of Glut10^G128E compared with WT mice at 5 months of age. Nuclei are stained with 4', 6-diamidino-2-phenylindole (DAPI). Scale bars = 50 µm. The mRNA expression levels of MMP2 (F), MMP7 (G), MMP9 (H) and TGF-β (I) in aortic tissues of Glut10^G128E and WT mice at 3-, 5-, and 10-months of age (n = 5). (J) Histopathology confirms that MMP2 expression is increased in the ascending aorta of Glut10^G128E mice at 5 months of age compared with WT mice. Scale bars = 50 µm. (K) Western blots of SOD2, Cyclin D1, MMP2, TGF-β, and β-actin in aortic tissues from 5- and 15-month-old Glut10^G128E and WT mice. (L) The activities of MMP2 were examined by gelatin zymography in aortic tissues from 5-month-old Glut10^G128E and WT mice. Results are shown as mean ± SD. Two-tailed, unpaired Student’s t-test, *P < 0.05; **P < 0.01; ***P < 0.001.

Increased ASMC proliferation/migration and vascular remodeling in Glut10^G128E mice

We further investigated the consequence of decreased intracellular AA and increased ROS levels in ASMCs and aortic tissue of Glut10^G128Emice. ASMCs from Glut10^G128E mice have increased cell proliferation (Fig. 7A) and migration (Fig. 7B) compared with ASMCs from WT. In control experiments using AA supplemented medium, the migration abilities of ASMCs from Glut10^G128E and WT mice were comparable (Fig. 7B). These results suggested that functional GLUT10 deficiency increases migration of Glut10^G128E ASMCs. Furthermore, expression of platelet-derived growth factor B (PDGF-B), a principal regulator of ASMC proliferation and migration (13), was significantly induced in Glut10^G128E ASMCs (Fig. 7C) and aortic tissues at 5 months of age (Fig. 7D). Increased cyclin D1 expression promotes cell proliferation (14), and we observed the increased cyclin D1 protein expression in Glut10^G128E aortic tissues (Fig. 7E and K). Vascular remodeling entails degradation and reorganization of extracellular matrix proteins in the vasculature by matrix metalloproteinases (MMPs), and TGF-β1 signaling is involved in abnormal vessel remodeling and dilatation (15). Both mRNA and protein levels of TGF-β1 (Fig. 7I and K) and MMP2 (Fig. 7F, J, K) were significantly increased in Glut10^G128E aortic tissues. A trend toward increased MMP7 and MMP9 mRNA expression in Glut10^G128E aortic tissues was observed (Fig. 7G and H). Importantly, MMP2 activity was increased in Glut10^G128E aortic tissues as examined by gelatin zymography (Fig. 7L). These results suggest that GLUT10 deficiency stimulates proliferation and migration of ASMCs and induces vascular remodeling through PDGF and TGF-β pathways in Glut10^G128E aortic tissues.

The plasminogen activation (PA) system plays a role in the regulation of ASMC migration, vascular remodeling after injury, and the formation of atherosclerosis (16). Thus, we determined if the PA system might be involved in vascular remodeling that occurs in Glut10^G128E mice. We measured the levels of tissue type-PA (tPA) and PA inhibitor-1 (PAI-1), in Glut10^G128E and WT plasma at 3, 5, 10 and 15 months of age. Interestingly, the active tPA protein levels were decreased in Glut10^G128E plasma (Supplementary Material, Fig. S10A), while there were no differences in total PAI-1 protein levels at any age examined (Supplementary Material, Fig. S10B). PAI-1 activities were slightly increased in Glut10^G128E plasma at 5 months (Supplementary Material, Fig. S10C), however overall, these results suggest that the PA system is not likely to play a role in progressive aortic wall changes in Glut10^G128E mice.

Discussion

Oxidative stress has been implicated as a causative factor in vascular structure changes that are associated with vascular dysfunction and cardiovascular diseases (17,18). However, homeostatic regulatory mechanisms for antioxidants in the vascular system are relatively unknown. We previously demonstrated that GLUT10 function as a DHA transport, and ASMCs from Glut10^G128E mice exhibit decreased DHA uptake, decreased AA accumulation, and increased ROS levels (10). Here, we demonstrate a novel regulatory mechanism that enhanced GLUT10 mitochondrial targeting under stress and aging conditions to maintain ascorbic acid (AA) and redox homeostasis, which confers the direct benefits of mitochondria structure and function. Glut10^G128E mutant protein impaired mitochondrial-targeting ability. ASMCs from Glut10^G128E mice, in consequence of impaired AA and redox homeostasis, exhibit altered mitochondrial structure, reduced mitochondrial function, as well as enhanced ASMC proliferation and migration. In vivo, heightened ROS levels, altered mitochondrial structure, increased number and disorganization of ASMCs, along with progressive arterial wall remodeling were observed in aortic tissues of Glut10^G128E mice. These defects finally culminated in elevated systolic blood pressure in aged Glut10^G128E animals. Overall, our results reveal a novel homeostatic mechanism, whereby GLUT10 may play a critical role in protecting major arteries especially under stress and aging conditions. Understanding adaptive mechanisms that protect the arterial system during physiological and pathological conditions is important for the development of efficient clinical strategies to prevent or treat vascular diseases.

Vascular phenotypes in Glut10^G128E mice

We observed slow, progressive changes in aortic walls and increased systolic blood pressure in aged Glut10^G128E mice. At 15 months of age, WT mice displayed a modest thickening of aortic walls compared with young WT mice. In contrast, Glut10^G128E mice showed accelerated changes in aortic wall structure compared with WT mice. Importantly, increased systolic blood pressure, associated with progressive structural changes in large arteries, is one of the most powerful risk factors for predicting cardiovascular diseases in elderly persons (19). Pathologically, we noticed dramatic increases and disorganization of both elastic fibers and ASMCs in aortic tissue of Glut10^G128E mice. Enhanced proliferation and migration of vascular SMCs has been suggested to contribute to age-related vascular remodeling that is an underlying factor in various vascular diseases (20). Our results suggest that GLUT10 plays an essential role in maintaining the aortic wall structure and systolic blood pressure during aging. Furthermore, impaired GLUT10 function accelerates aortic wall changes and increases systolic blood pressure in aged mice.

GLUT10 deficiency in humans leads to ATS. However, we did not observe arterial tortuosity, stenosis/dilation, aneurysm, or connective tissue manifestations in Glut10^G128E mice. The discrepancies between human and mouse might be due to the mild functional changes caused by missense mutations and the intrinsic differences between mouse and human physiology. Twenty-three GLUT10 mutations have been reported to cause ATS (21). However, the associations between disease-related mutations, protein function, and disease pathogenesis remain unclear. The orthologous Glut10^G128E mutation has not been identified in ATS patients, which might suggest that the mild mutation does not lead to ATS phenotypes in human. Previously, we reported that Glut10^G128E ASMCs exhibited impaired DHA uptake and AA accumulation in mitochondria (10). Here we demonstrate that Glut10^G128E protein has impaired mitochondrial targeting ability, causes intracellular ROS to accumulate, and negatively affects mitochondrial function. Despite these functional defects, connective tissue pathology was not observed in Glut10^G128E mice. A trend toward increased collagen staining, but no significant difference in hydroxyproline and type 1 collagen content in aortas was found between Glut10^G128E and WT mice (Supplementary Material, Fig. S11). Recent studies showed DHA transport is defective in the endomembranes of fibroblasts from ATS patients (22). These results suggest that the mild phenotypes observed in Glut10^G128E mice might represent a functional deficiency of GLUT10 which is restricted to mitochondria, and that the endomembrane-localized GLUT10 function in collagen synthesis may be only slightly compromised, if at all. The DHA transport function in endomembranes of Glut10^G128E mutant ASMCs should be further evaluated. Other reasons for the mild phenotype in mice may be attributed to intrinsic differences between mouse and human physiology. For example, rodents can synthesize AA from glucose. However, due to a mutation in the gulanolactone oxidase (Gulo) gene, humans cannot synthesize AA and are dependent on dietary intake. Also, AA synthesis deficient Gulo^-/- mice do not have significant defects in proline hydroxylation and collagen production, suggesting AA-independent synthesis of collagen in mice (23).

Antioxidant systems in the vascular system

The regulation of non-enzymatic antioxidants in the vascular system is not well described. Our results demonstrate that GLUT10 plays a direct role in maintaining redox homeostasis and mitochondrial function by mediating DHA transport into mitochondria in ASMCs. We provide the molecular mechanisms showing that the increased GLUT10 expression and mitochondrial trafficking are involved in maintaining redox homeostasis and associated with mitochondrial function. We further showed PDGF, TGF-β, and MMPs are important mediators of ROS induced ASMC proliferation and migration, which may explain the progressive arterial wall remodeling we observed in Glut10^G128Emice. Previous studies showed that GLUT10 is required for the development of the cardiovascular system and notochord in zebrafish embryos, and acts via mitochondria and TGF-β mediated signaling (24). Here, we provide evidence showing that GLUT10 is important in maintaining aortic tissue structure in postnatal tissue.

Increased mitochondrial oxidative stress impairs vascular function during aging and pathologic conditions (25). Mitochondrial dynamics are important in maintaining cellular metabolism and physiological function (26). The current study shows that stress conditions increase intracellular ROS levels, increase mitochondrial fragmentation, and decrease mitochondrial function in ASMCs and aortic tissues of Glut10^G128E mice. Moreover, mitochondrial structure was disrupted and mitochondrial number was decreased in aortic tissues of Glut10^G128E mice at 15 months of age. These results suggest that mitochondrial fragmentation may be an early event mediated by ROS that triggers arterial wall remodeling and further augments ROS production in arterial walls of aged Glut10^G128E mice. This concept is supported by the findings that mitochondrial shortening induces ASMC migration and intimal hyperplasia (27). Thus, our study links GLUT10 deficiency to excessive ROS production and disruption of mitochondrial dynamics, finally leading to the aortic abnormalities observed in Glut10^G128E mice.

Furthermore, in addition to increased ROS levels in aortic tissues, we observed increased ROS in plasma of Glut10^G128E mice. The human GLUT10 gene, SLC2A10, is located on chromosome 20q12–13.1 (28), a region that has been shown to be associated with type 2 diabetes (T2D) (29). However, the evidence for an association between mutations or polymorphisms in SLC2A10 gene and T2D remain inconclusive and the physiological function of GLUT10 requires further investigation. We have previously reported that variations in SLC2A10 are associated with susceptibility to peripheral vascular complications in T2D independent of other known risk factors (30). Together with current results, those data suggest that GLUT10 may have a role in maintaining the integrity of peripheral arteries. Thus, similar mechanisms may contribute to peripheral vascular complications in T2D. Glut10^G128E mice provide a unique opportunity to study the association of SLC2A10 variation with vascular complications in T2D.

Mitochondrial ROS and AA

How mitochondrial ROS are regulated dictates the biological outcomes. Mitochondria are a major site of ROS production and mitochondrial function can be compromised by severe and prolonged oxidative stress. Mitochondria are well-equipped with antioxidant defenses, including a large pool of glutathione, GPX, glutathione reductase, SOD2, CAT, the thioredoxin system, and AA (31). Studies have shown that AA can enter mitochondria as DHA through GLUT1 (32) or GLUT10 (10), or through SVCT2 (33) in different cell types, suggesting that AA uptake into mitochondria may occur in a cell type-specific manner.

We propose that GLUT10 facilitates DHA uptake into mitochondria followed by reduction to AA, which is an important antioxidant mechanism that protects ASMCs from oxidative stress. Irrespective of the mechanism by which AA accumulates in mitochondria, GSH-dependent reduction of DHA is a major AA recycling reaction in mammalian mitochondria (34,35). Under oxidizing conditions, DHA that is generated in the extracellular space is absorbed by surrounding cells through GLUTs and reduced to AA (36,37), thereby assisting with quenching free radicals and maintaining redox homeostasis. Keeping low intracellular DHA concentration is important to allow DHA to move along a concentration gradient (38). Therefore, a rapid mitochondrial recycling of DHA can drive the facilitated transport of DHA from the extracellular space through GLUTs. Although the transport kinetics of DHA in mitochondria are not well-described, and the possibility exists that other GLUTs may participate in mitochondrial DHA uptake in ASMCs. One of the most intriguing and novel findings of the current work is that GLUT10 is targeted to mitochondria in ASMCs under stress and aging conditions, presumably resulting in increased DHA recycling and extracellular DHA uptake.

The impacts of GLUT10 deficiency on redox homeostasis, mitochondrial dynamics and function, ASMC proliferation and migration, arterial structure, and arterial functions were studied to characterize the physiological and pathologic function of GLUT10 on aortic tissues. However, the other functions of GLUT10, including solute transport characteristics, functions in other subcellular compartments, and functions in other tissues, remain to be elucidated. In addition to aortic tissue, GLUT10 is highly expressed in white adipose tissue and liver and moderately expressed in other tissues (10,28). Importantly, deficiency of GLUT10 in other tissues might impact aortic tissue through indirect means. Further studies are needed to explore this possibility. The present study demonstrates that GLUT10 is targeted to mitochondria under stress conditions, and participates in maintaining mitochondrial function and redox homeostasis in ASMCs and aortic tissue under physiological and pathological conditions.

Materials and Methods

Additional detailed methods are available in the Supplementary Material.

Mice

The creation of Glut10^G128E mice in a C3HeB/FeJ background was described previously (11). All animal protocols were approved by the Institutional Animal Care and Utilization Committee, Academia Sinica (Protocol #14–12-795). The investigation conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health.

Cell culture

A10 cells (ATCC CRL-1476), rat aortic SMC, and MOVAS (ATCC CRL-2797), mouse aortic SMC, were obtained from ATCC (American Type Culture Collection, ATCC, Manassas, VA, USA). Human ASMCs (354–05a, Cell Application Inc., San Diego, CA, USA) were obtained from a normal human aorta of a 53-year-old Caucasian male single donor. All ASMCs were maintained in smooth muscle cell growth medium following the protocol provided by the manufacturer (supplied with 75 μm AA). We used early passage cells (P2-P3) to represent young human ASMCs and late passage cells (P10-P14) to represent the aged human ASMCs.

Colocalization analysis

Cells were transfected with GLUT10/enhanced green fluorescent protein (EGFP) and stained with MitoTracker Red. Z-Series images were obtained from a Leica TCS-SP5 confocal microscope. Imaris 7.7.2 software was employed for visualization and quantification of colocalization. The percentage of EGFP protein colocalized with MitoTracker red was quantified by dividing the total of the intensity sum per object for the colocalized image by the total of the intensity sum per object for the EGFP image.

ROS analysis

The intracellular ROS levels in ASMCs were determined using 2′, 7′-dichlorodihydrofluorescein diacetate (DCFH-DA; Sigma-Aldrich, St. Louis, MO, USA) as described previously (10). The mitochondrial ROS levels in ASMCs were determined using MitoSox Red (Thermo Fisher, Miami, FL, USA). ROS production in freshly cut frozen aortic sections was analysed by dihydroethidium (DHE) (Invitrogen, Grand Island, NY, USA). Images were obtained using a Leica TCS-SP5 confocal microscope (Leica, Wetzlar, Germany).

DHA uptake and intracellular AA measurement

DHA uptake was analysed using ¹⁴C-labeled DHA as previously described (10). Briefly, cells were incubated with 1 mM ¹⁴C-labeled DHA for 30 min and intracellular DHA levels were determined. The intracellular AA levels were determined using an Ascorbic Acid Assay Kit (Abcam, Cambridge, England, UK), according to the manufacturer’s instructions.

Primary ASMC culture

ASMCs were isolated from single mouse aortas from 6–8 week Glut10^G128E mice or wild-type littermates (WT) as described previously (39). Studies used P2-P3 to represent young ASMCs and P8 to represent old ASMCs.

Oxygen consumption rate (OCR) measurements

OCRs of indicated number of ASMCs isolated from Glut10^G128E and WT were measured using the Seahorse XF24 extracellular flux analyzer, changes of oxygen concentrations dissolved in the media were measured every 14 min by a solid-state sensor probe. After determining basal respiration, oligomycin, carbonylcyanide-p-trifluoromethoxyphenylhydrazone, and antimycin were added sequentially to the media, and the OCRs for each condition were quantified. This allowed for the determination of basal respiration, ATP production, proton leak, maximal respiration, and non-mitochondrial oxygen consumption.

Mitochondrial morphology analyses

Mitochondria in cells were stained with MitoTracker Red (Invitrogen). Cell images were obtained with a Leica TCS-SP5 confocal microscope. Micro-P software was used to measure and sort individual mitochondria according to their area and axial length/width (40).

Transmission electron microscopy

Aortic tissues were fixed and processed for electron microscopy using a standard protocol (41). To assess mitochondrial morphology, measurements were made in eight random photos per group at a magnification of 2550-fold. The reference area was 54 0566 μm² for morphometric analyses. We used a three grade classification system, according to the appearance of mitochondria: grade 1, normal; grade 2, normal size lacking well-defined cristae; grade 3, swollen mitochondria, at least double the normal size and lacking well-defined cristae.

Histopathology, immunohistochemistry, and immunofluorescence

The ascending thoracic aorta was collected, fixed, and sections were stained with hematoxylin and eosin for general pathological examinations. Verhoeff-van Gieson or Masson’s trichrome stain to label elastin or collagen fibers, respectively (11). For immunohistochemical or immunofluorescence staining, sections were incubated with anti-SOD2 (Abcam, Cambridge, MA, USA), anti-cyclin D1 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), or anti-matrix metalloproteinase 2 (MMP2) (Abcam) antibodies. Human ASMCs were fixed and stained with anti-GLUT10 (Alpha Diagnostic Intl inc., SanAntonio, TX, USA), anti-COXV (A21351, Invitrogen, Carlsbad, CA, USA), anti-FIS1 (Proteintech, Rosemont, IL, USA) or anti-MFN2 (Proteintech, Rosemont), which were detected by fluorescence-labeled secondary antibodies.

Blood pressure and pulse rate measurements

Blood pressure and pulse rate were measured using a non-invasive BP-2000 Blood Pressure Analysis System (Visitech Systems, Apex, NC, USA).

ECG recordings

ECG recordings were obtained with a Bio Amp PowerLab System. Wave durations were calculated automatically by the software (PowerLab 8/30; ADInstruments, Castle Hill, Australia).

Detection of oxidized proteins

Oxidized proteins were detected by derivatizing carbonyl groups with 2, 4-dinitrophenylhydrazine. The dinitrophenol tag was probed by western blotting using anti-dinitrophenol antibodies (Chemicon International Inc., Billerica, MA, USA).

ASMC proliferation assay

ASMCs from 6–8 week old Glut10^G128E mice and WT were cultured in 6-well plates and cell number was counted at 24, 48 and 72 h after seeding.

ASMC migration assay

For the in vitro wound healing assay, a 0.5 mm-wide wound was made across a monolayer of ASMCs, and photomicrographs of the wound were taken at various times post-wounding. The areas of cell coverage were determined with image analysis software (MetaMorph, Universal Imaging Corp, West Chester, PA, USA). The percentage of cellular coverage relative to the original wound area was measured to evaluate migration.

Quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) and mitochondrial DNA content analyses

Total RNA from aortic tissues or cultured cells was isolated, converted to cDNA, and subjected to qRT-PCR as described previously (10). Expression data were normalized to β-actin mRNA levels. Mitochondrial DNA content was quantified by determining the ratio of content of the mitochondrial gene cyclooxygenase-2 to a nuclear gene, β-globin, by qPCR. Gene-specific primer sequences are provided in Supplementary Material.

Western blot analyses

Total protein lysates from aortic tissues were used for the analysis. The primary antibodies were SOD2 (Abcam), cyclin D1 (Abcam), MMP2 (Santa Cruz), transforming growth factor (TGF)-β (Abcam), and β-actin (Millipore, Billerica, MA, USA). Proteins were visualized using enhanced chemiluminescence (Millipore).

Statistical analyses

Results are shown as mean ± SD. When only two conditions were compared, the data were evaluated by two-tailed, unpaired Student’s t-test. When multiple conditions were compared, a 1- or 2-way ANOVA was used, followed by Dunnett’s multiple comparison post-hoc tests, if appropriate. P-values less than 0.05 were considered significant. *P < 0.05; **P < 0.01; ***P < 0.001.

Supplementary Material

Supplementary Material is available at HMG online.

Acknowledgements

We thank miss Chia-Ni Tsai, Academia Sinica, for her technical assistance. We thank the Taiwan Mouse Clinic (MOST 104–2325-B-001–011) which is funded by the National Research Program for Biopharmaceuticals (NRPB) at the Ministry of Science and Technology (MOST) of Taiwan for their technical support in blood pressure and ECG records. We would like to thank the EM core facilities from ICOB and IPMB, Academia Sinica for providing the service of transmission electron microscopy. We also thank Dr. Chao-Chun Hsu, imaging core facility, ICOB, Academia Sinica, for her technical assistance in the confocal imaging processing.

Conflict of Interest statement. None declared.

Funding

Academia Sinica, Taiwan (Thematic Research Project, 234T), Ministry of Science and Technology, Taiwan (104–2320-B-001–021-MY3, 102–2320-B-001–028-MY2, 101–2320-B-007–007-, and 100–2314-B-007–002-) to Dr. Yi-Ching Lee.

Author notes

Yu-Wei Syu and Hao-Wen Lai authors contributed equally to this work.