Search tips
Search criteria 


Logo of hmgLink to Publisher's site
Hum Mol Genet. 2012 March 15; 21(6): 1248–1259.
Published online 2011 November 24. doi:  10.1093/hmg/ddr555
PMCID: PMC3284116

GLUT10 is required for the development of the cardiovascular system and the notochord and connects mitochondrial function to TGFβ signaling


Growth factor signaling results in dramatic phenotypic changes in cells, which require commensurate alterations in cellular metabolism. Mutations in SLC2A10/GLUT10, a member of the facilitative glucose transporter family, are associated with altered transforming growth factor-β (TGFβ) signaling in patients with arterial tortuosity syndrome (ATS). The objective of this work was to test whether SLC2A10/GLUT10 can serve as a link between TGFβ-related transcriptional regulation and metabolism during development. In zebrafish embryos, knockdown of slc2a10 using antisense morpholino oligonucleotide injection caused a wavy notochord and cardiovascular abnormalities with a reduced heart rate and blood flow, which was coupled with an incomplete and irregular vascular patterning. This was phenocopied by treatment with a small-molecule inhibitor of TGFβ receptor (tgfbr1/alk5). Array hybridization showed that the changes at the transcriptome level caused by the two treatments were highly correlated, revealing that a reduced tgfbr1 signaling is a key feature of ATS in early zebrafish development. Interestingly, a large proportion of the genes, which were specifically dysregulated after glut10 depletion gene and not by tgfbr1 inhibition, play a major role in mitochondrial function. Consistent with these results, slc2a10 morphants showed decreased respiration and reduced TGFβ reporter gene activity. Finally, co-injection of antisense morpholinos targeting slc2a10 and smad7 (a TGFβ inhibitor) resulted in a partial rescue of smad7 morphant phenotypes, suggesting scl2a10/glut10 functions downstream of smads. Taken together, glut10 is essential for cardiovascular development by facilitating both mitochondrial respiration and TGFβ signaling.


Growth factor signaling requires a close coupling to metabolism to direct growth and development and to maintain homestasis. For example, the phosphoinositide 3-kinase/protein kinase B/mammalian target of rapamycin (PI3K/AKT/mTOR) pathway integrates proliferative insulin-like growth factor signals, the availability of amino acids, cellular energy and oxidative potential with protein synthesis (1). TGFβ growth factors provide essential signals in cardiac and vascular development (2) and are key initiators of the fibrotic response in a variety of disease states including myocardial infarction (3) and atherosclerosis (4). Typically, TGFβ signals elicit reduced cell proliferation, epithelial–mesenchymal transition, assumption of migratory phenotypes, elevated synthesis of both structural and remodeling components of the extracellular matrix (ECM) and elaboration of a contractile cytoskeleton. We postulate that such dramatic phenotypic changes require metabolic integration.

As a way to define shared molecules in the TGFβ signaling and metabolic pathways, we focused on SLC2A10/GLUT10, a member of the facilitative glucose transporter family. Mutations in SLC2A10 cause arterial tortuosity syndrome (ATS—OMIM#208050), a recessively inherited disorder characterized by elongation, tortuosity, stenosis and aneurysms of the large and medium sized arteries in association with distinct craniofacial and connective tissue manifestations (5,6). Functional analysis of patient tissue samples has shown that a loss of function of GLUT10 may lead to vascular malformations via upregulation of the TGFβ signaling pathway in the arterial wall.

Increased TGFβ signaling in association with arterial aneurysm formation (and tortuosity) is a common finding in aortic aneurysm syndromes such as Marfan syndrome, caused by heterozygous mutations in the FBN1 gene (7), Loeys–Dietz syndrome caused by heterozygous mutations in the genes encoding the TGFβ receptors 1 and 2 (TGFBR1 and TGFBR2) (8) and autosomal—recessive cutis laxa due to fibulin 4 deficiency (9,10). Involvement of TGFβ upregulation in the pathogenesis of Marfan syndrome has been illustrated in a mouse model in which several clinical features, including pulmonary emphysema, aortic root dilatation and skeletal muscle dysfunction, could be prevented and reversed after administration of TGFβ neutralizing antibodies or a pharmacological inhibitor of the pathway, losartan (7,11,12). Despite the association of TGFβ upregulation with ATS, the developmental origin of vascular malformations in these diseases remains unclear, and in some cases, conflicting. For example, the reported elevation of TGFβ signaling in Loeys–Dietz syndrome blood vessels is paradoxical because these patients carry inactivating mutations in either TGFBR1 or TGFBR2 (8).

The exact role of GLUT10 in the TGFβ signaling pathways and metabolism remains to be elucidated. Mice with homozygous missense substitutions in GLUT10 do not show the same severe vascular abnormalities as encountered in human ATS patients and therefore these models are of limited use to investigate the pathogenetic mechanisms underlying human ATS (13,14). Because recent studies suggest that the zebrafish is a very useful organism to study cardiovascular disorders (15), we aimed to establish an ATS zebrafish model by knocking down the zebrafish slc2a10 gene. Using small molecule treatment and gene expression profiling experiments, we show a significant overlap between glut10 function and the TGFβ signaling pathway. Moreover, we find that expression of several genes necessary for cellular respiration are altered by glut10 deficiency. Finally, functional assays indicate that glut10 is required both for mitochondrial respiration and for optimal TGFβ signaling.


Zebrafish slc2a10/glut10 structure

The zebrafish slc2a10 gene is located on chromosome 11. The gene structure is similar to the human homologue and also contains five exons of similar size, although intronic sequences and untranslated regions are shorter (Fig. 1A). The structure of human GLUT10, the protein encoded by SLC2A10, contains 12 hydrophobic transmembrane domains (TMD) with two large hydrophilic exofacial and endofacial loops. Hydropathy analysis reveals an identical structure for zebrafish glut10 (Supplementary Material, Fig. S1). Aligning the amino acid sequence of human and zebrafish GLUT10 shows a homology of 43%, with the major structural differences being a shorter and divergent exofacial loop 9 between TMD 9 and 10 and a divergent endofacial loop 6 between TMD 6 and 7 in glut10 (Fig. 1B). The N334ATG glycosylation motif in the large exofacial loop 9 in GLUT10, which is a hallmark of the class 3 sugar transport facilitators, is replaced by an N341LTL glycosylation motif in the same loop in glut10 (PROSITE analysis). Similar to its human homologue, glut10 retains most of the sugar transporter signatures, which are characteristic of the mammalian glucose transporters in general and of the subfamily of class 3 sugar transport facilitators in particular.

Figure 1.
Evolutionary conservation of the SLC2A10 gene and GLUT10 protein. (A) The structure of the human SLC2A10 and the zebrafish slc2a10 genes with coding (full boxes) and untranslated (empty boxes) regions shown. MO target regions 1 and 2 are depicted underneath ...

Slc2a10 knockdown phenotype

Slc2a10 is provided to the embryo as a maternal transcript and is widely expressed during gastrulation, segmentation and pharyngula periods, suggesting a developmental role for this transporter (16). To reveal the role of the slc2a10 gene in zebrafish development, we performed knockdown experiments with two different antisense morpholino oligonucleotides (MOs). One MO was targeting the slc2a10 start codon (ATG—MO) and is complementary to parts of exons 1 and 2 (Fig. 1A). The second MO was a splice—blocking MO, complementary to the exon 2–intron 2 donor splice site (splice—MO). Injection of 2.5 ng of ATG—MO caused phenotypic abnormalities, showing a relatively broad range of severity in gross morphology. With increasing severity, one could distinguish wild—type embryos, class 1 embryos characterized by a bowed notochord/tail, class 2 embryos with a wavy notochord or tail and class 3 embryos which were very small and showed extensive tissue dysplasia especially in the tail region (Fig. 2A and Table 1).

Table 1.
Phenotype classification of slc2a10—MO—injected embryos at 48 hpf
Figure 2.
Slc2a10 knockdown phenotype. (A) General morphology of slc2a10 morphants at 48 hpf. Besides wild-type embryos, three different embryo classes can be discerned based on general notochord/tail structure. (B) Class 1 and 2 morphants exhibit bowing and kinking ...

Injection of 7.5 ng of splice—MO caused identical phenotypes with almost identical frequencies for the different classes. Injection of a scrambled control—MO, which has no target in the zebrafish genome, did not cause any visible phenotypic abnormalities compared with uninjected embryos. Survival curves also indicated similar, 5–10% reduction of survival in both ATG- and splice—MO-injected embryos compared with control—MO-injected embryos at 120 hours post-fertilization (hpf; Supplementary Material, Fig. S2). This was due to the high mortality of severely affected class 3 embryos which comprised ~10% of the ATG- and splice—MO-injected embryos.

Next, we investigated the effect of the splice—MO on slc2a10 mRNA splicing by performing reverse transcription polymerase chain reaction (RT–PCR) with primers in exons 1 and 3 of the slc2a10 gene. As expected, we found that injection of the splice—MO resulted in the skipping of exon 2 (Fig. 3), which caused a frameshift and a premature termination codon. The resulting transcript was 50% less abundant than slc2a10 mRNA in control embryos, presumably as a consequence of nonsense-mediated decay.

Figure 3.
MO knockdown of slc2a10. (A) Schematic representation of slc2a10 splicing after splice—MO injection. A black bar represents the target position of the splice—MO. Arrows denote the position of the RT–PCR primers in exons 1 and 3. ...

We were not able to check whether the ATG—MO effectively inhibited translation as no antibodies that target the glut10 protein were available. Therefore, we performed the subsequent experiments using the splice—MO instead of the ATG—MO. We also focused on 48 hpf embryos. At this stage, most of the internal organs are developed including a fully functional cardiovascular system (17). Moreover, at 48 hpf, diffusion-mediated gas exchange still suffices for basic metabolic supplies in the embryos, which makes them independent of convective blood circulation, avoiding secondary effects of circulatory abnormalities.

The 48 hpf morphants were significantly smaller with a reduced embryo/yolk sac extension length ratio (Table 2). They had a bowed (class 1) or wavy (class 2) tail with notochord abnormalities (Fig. 2B). Cardiac edema was frequently observed and most of the embryos showed cardiovascular abnormalities with a reduced heart rate and blood flow, incomplete and irregular patterning of the vasculature especially in the tail, eventually causing blood pooling in the sinus venosus of the heart and in the tail region (Fig. 2C and D).

Table 2.
Impaired growth and circulation in slc2a10 splice—MO-injected embryos compared with control—MO-injected embryos at 48 hpf

Inhibition of TGFβ pathway in zebrafish by drug administration

Because abnormal TGFβ signaling has been shown in cells and tissues from human ATS patients, we performed pharmacological studies to evaluate the effect of the slc2a10 knockdown on the TGFβ pathway. We used a TGFβ type 1 receptor kinase inhibitor (ALK5 inhibitor, LY-364947), which specifically targets the TGFBR1 kinase function and which is, in contrast to other TGFBR1 inhibitors, much less potent against related kinases such as TGFBR2 (18,19). Blocking this kinase inhibits phosphorylation of SMAD2 and SMAD3 and downregulates the TGFβ signaling. Alignment of the amino acid sequence of human TGFBR1 with both zebrafish tgfbr1a and tgfbr1b revealed a high level of conservation (77 and 79%, respectively, data not shown) with almost complete conservation of the kinase domain (96%). This provided a strong structural basis for the use of LY-364947 on zebrafish embryos.

We applied an LY-364947 dilution series ranging from 0 to 100 µm to wild-type embryos to assess toxic effects or to detect specific phenotypes (Supplementary Material, Table S1). At 10 µm no abnormalities were observed, whereas at ≥80 µm all embryos died. At a concentration of 40 µm, a specific dysmorphic phenotype could be detected in almost all embryos (Fig. 4A). The observed anomalies were similar to those found in slc2a10 knockdown embryos (compare Fig. 4A, LY-treated embryo, with Fig. 2A, class I embryo). The embryos were significantly smaller, showed bowing of the tail and notochord, low heart rate, vascular abnormalities, no blood flow with blood pooling in the sinus venosus. In Fli1:eGFP fish, treated with 40 µm LY-364947, condensation of the caudal vein plexus, a structure that slowly remodels into a single vascular tube during embryogenesis, was observed (Fig. 4B). Inhibition of TGFβ signaling did not influence the expression of the slc2a10 gene assayed by quantitative RT–PCR (qPCR), indicating that slc2a10 was not subject to feedback regulation by the TGFβ pathway (data not shown). Finally, treatment of splice—MO-injected embryos with 40 µm LY-364947 made the phenotype even more severe, yielding embryos that all belonged to the severe class 3 (Fig. 4C).

Figure 4.
Tgfbr1 inhibitor (LY-364947) treatment. (A) Wild-type fish treated with 0 and 40 µm LY-364947. Blood pooling in the sinus venosus is indicated by a black arrow. (B) Fli1:eGFP fish treated with 0 or 40 µm LY-364947. Condensation of the ...

Transcriptome analysis in slc2a10 knockdown and LY-364947 treated zebrafish

To further elucidate the role of GLUT10 function in the TGFβ and other signaling pathways, we looked for similarities and differences in the global zebrafish transcript profiles caused by reduced slc2a10 expression and by tgfbr1 inhibition. We extracted RNA from 48 hpf embryos and prepared cDNA with a two-color-labeling procedure, which was subsequently loaded on Agilent expression arrays consisting of 43 803 60mer probes. First, we compared gene expression in embryos injected with slc2a10 splice—MO versus control—MO. Secondly, we compared gene expression in embryos treated with LY-364947 versus untreated embryos. To establish gene sets that were overexpressed or underexpressed, we took as a cut-off transcript-level changes of ≥1.5-fold relative to the control samples. We compared the differentially expressed gene sets between both experiments to define specific gene subsets that are common. Of the 519 genes that were over- or underexpressed in the slc2a10 knockdown model, ~50% (245 genes) were also dysregulated by LY-364947 (Fig. 5A). Furthermore, most of the genes that were dysregulated by both treatments were up- or downregulated in the same direction and to the same extent, as indicated by the strong correlation (r = 0.81) between their mean expression ratios (Fig. 5B). Consistent with known functions of TGFβ, many genes important for cardiovascular, cartilage and eye development and neurogenesis were downregulated (Supplementary Material, Table S2). In contrast, genes involved in DNA replication, DNA repair and cell cycle progression were mostly upregulated. Quantitative PCR experiments validated the array data in all five upregulated (acta2, rrm1, pcna, mcm4, mcm5) and four downregulated genes (acta1, versicanb, mtn1, col10a1) that we tested for both slc2a10—MO and LY-364947 treatments (Supplementary Material, Fig. S3).

Figure 5.
Transcriptome analysis in slc2a10 knockdown and tgfbr1 (LY-364947) inhibitor-treated zebrafish. (A) Venn diagram depicting the overlap between differentially expressed gene transcripts (≥1.5-fold expression; P< 0.05) after LY-364947 treatment ...

A relatively large proportion of the genes that showed differential expression in the slc2a10 knockdown model but not in the LY-364947 treated embryos are involved in the pathways supplying energy to the cell (Supplementary Material, Table S3). Several of these genes belong to the oxidative phosphorylation pathway (cyc1, ndufab1), the concomitant reactive oxygen production pathway (sod2, gpx4a, ant, mpx), the Szent–Györgyi–Krebs cycle (mdh1b, got2a and slc13a2), the glycolysis/gluconeogenesis pathway (eno2, pkm2b, tpi1a, pfkm) and glycogen metabolism (gys2, gygl). In addition, several genes involved in calcium binding and homeostasis and the production of heme or hemoproteins in the mitochondria were uniquely downregulated in the slc2a10 knockdown model. Also, several components of the contractile muscle cytoskeleton, the connective tissue and the cardiovascular system showed differential expression patterns specifically upon slc2a10 knockdown. Finally, compared with the LY-364947 treatment, additional genes involved in DNA replication, DNA repair and cell cycle progression were upregulated when slc2a10 was depleted. We selected 6 upregulated (mpx, pfkm, hmbsl, tnnt1, fn1b, mmp13) and 19 downregulated (sod2, tpi1a, ndufab1, cyc, eno2, got2a, pkm2b, gys2, mdh1b, slc13a2, cyc1, gpx4a, slc25a4, gygl, ppox, pvalb2, pvalb5, mybpc3, agt) transcripts to validate slc2a10 knockdown-specific changes by qPCR and in each case the direction and magnitude of change were replicated (Supplementary Material, Fig. S4).

Slc2a10 is required for mitochondrial function

To test whether altered mitochondrial gene expression caused structural anomalies, we examined mitochondria in slc2a10—MO (7.5 ng) and control—MO (5 ng) treated embryos at 52 hpf by transmission electron microscopy. No remarkable morphological change was found in the mitochondria in slc2a10 knockdown embryos (Fig. 6A and B). To investigate whether glut10-deficient mitochondria were functionally impaired, we measured the oxygen consumption of control- and slc2a10—MO-treated embryos using a Seahorse XF24 extracellular flux analyzer. Knockdown of slc2a10 resulted in a significant, 20–25% reduction in oxygen consumption rate (OCR) starting at 4 hpf (Fig. 6C) and lasting through 24 hpf (Fig. 6D and data not shown).

Figure 6.
Knockdown of slc2a10 does not affect mitochondrial morphology but decreases the OCR. Electron micrographs of mitochondria from control—MO-injected (A) and slc2a10—MO-injected (B) embryos at 52 hpf show normal morphology. Magnification ...

Chemicals or endogenous proteins can equalize the electrochemical potential gradient between the matrix and the intermembrane space of mitochondria, thus uncoupling ATP synthesis from the electron transport chain (20). Uncoupling in turn releases the ‘backpressure’ on the proton transporters of the electron transport chain allowing metabolic oxidation to proceed at a maximal rate. To probe the capacity of the electron transport chain, we administered a chemical uncoupler (FCCP, carbonyl cyanide p-trifluoromethoxy-phenylhydrazone) to the embryos. FCCP treatment significantly increased the OCR in control embryos causing a 3-fold increase at 5 hp and a smaller, 20% increase at 22 hpf (Fig. 6C and D). In contrast, slc2a10 morphants showed no increase in respiration in response to FCCP, indicating that both the physiological and the maximal capacity of the electron transport chain were reduced in the absence of glut10. Treatment of embryos with rotenone, an inhibitor of complex I, reduced respiration in both control and slc2a10 morphant embryos to the same baseline level (Fig. 6C and D), providing further proof of the specificity of our assay for mitochondrial respiration.

Slc2a10 is required for TGFβ signaling

Similarities in phenotypes and gene expression profiles between chemical inhibition of TGFβ signaling and slc2a10 knockdown provide a large body of correlative evidence supporting a role for slc2a10 as a facilitator of TGFβ signaling. To obtain more direct evidence and to begin to map the location of glut10 in the TGFβ pathway, we performed two experiments. First, we injected a TGFβ-responsive luciferase reporter construct (p3TP-lux) into embryos. The promoter in this construct contains three TPA (12-O-tetradecanoylphorbol-13-acetate) response elements and a part of the plasminogen activator inhibitor 1 (PAI1) promoter (21), and has previously been shown to be active in zebrafish (22). Simultaneous injection of slc2a10—MO resulted in a 75% reduction of promoter activity compared with p3TP-lux only (Fig. 7A).

Figure 7.
Knockdown of slc2a10 downregulates TGFβ-like transcriptional responses and rescues smad7 morphant phenotypes. (A) A p3TP-lux TGFβ-responsive reporter construct was injected in the presence or in the absence slc2a10—MO. Luciferase ...

If the loss of glut10 function indeed results in depressed TGFβ-dependent transcriptional responses, as our results thus far show, we reasoned that glut10 knockdown might rescue the phenotypic effects of artificially elevated TGFβ activity. To create a state of increased TGFβ signaling in embryos, we knocked down smad7, an endogenous inhibitor of tgfbr1 signaling. Moderate concentrations (5 ng) of a smad7 splice morpholino resulted in a complete arrest of the development of embryos at the mid-blastula transition (3–4 hpf) and death by 8 hpf (data not shown). At lower MO concentrations (1 ng), smad7 knockdown permitted the development of the embryo through gastrulation and organogenesis but resulted in a recognizable pattern of malformations, including small head and eyes, pericardial edema, severely shortened tail, underdeveloped muscle and notochord (Fig. 7B). Injection of slc2a10—MO alone caused the previously described phenotype with bent notochord, shorter embryo and pericardial edema.

Combined injection of smad7 and slc2a10—MO resulted in a phenotype showing a significant improvement over smad7—MO alone in length, head size, cardiac morphology and the overall volume of mesodermal derivatives (Fig. 7B). However, the combination treatment did not result in an improvement over glut10—MO treatment alone. Heart rate, a quantitative measure of cardiovascular function, showed similar changes, with severe (32%) reduction in response to smad7—MO treatment only and a 17% improvement with combined treatment (Fig. 7C). The heart rate of slc2a10 morphants was reduced by 24% relative to control, but combination treatment with smad7—MO did not improve this further.


We present an ATS zebrafish model, generated by MO-based knockdown of the slc2a10 gene, which encodes the glut10 protein. Two MOs, targeting the slc2a10 start codon and the exon 2–intron 2 donor splice site, respectively, produced identical phenotypes underscoring the specificity of the MOs for slc2a10 as the likelihood of both MOs mistargeting the same gene is very small (23). The most prominent features of the morphants were a bowed/wavy appearance of the notochord and tail region and cardiovascular insufficiency. Cardiovascular abnormalities included incomplete and irregular patterning especially of the venous plexus and the intersegmental vessels. The heart rate was significantly reduced and blood pooling frequently observed in the heart and tail regions. These circulatory abnormalities may represent developmental precursors to lesions observed in human ATS: tortuosity and aneurysms of the large blood vessels.

In morphant fish, the notochord appeared bowed, kinked and shortened. The notochord consists of large mesodermal cells, packed within a sheath of connective tissue. It represents a primitive form of cartilage that defines the primitive longitudinal skeletal axis of the embryo that guides the formation of the vertebral column. It also provides key signals to the development of other mesodermal derivatives, including the vasculature (24). Therefore, it is unclear whether the notochord abnormalities contribute to the vascular patterning defects that we observed following slc2a10 knockdown. It has been shown that early curvature of the notochord in zebrafish embryos can result in a scoliotic adult phenotype (25). Similar to our zebrafish model, vertebral column abnormalities, including scoliosis, have been observed in human ATS (5).

Previously, it has been shown that TGFβ signaling is upregulated in vascular smooth muscle cells of ATS patients (6). This might, at least in part, be responsible for the phenotypic abnormalities encountered in ATS patients, especially because a link between elevated TGFβ signaling and connective tissue defects has been shown in related syndromes, including the Marfan, Loeys–Dietz and some cutis laxa syndromes (8,10,12,26). Surprisingly, we found downregulation, rather than upregulation, of total-body TGFβ signaling in slc2a10 knockdown zebrafish embryos based on five lines of evidence. First, treatment of wild-type embryos with a tgfbr1 inhibitor resulted in a phenotype similar to the ATS zebrafish model, with a bowed notochord/tail region and comparable cardiovascular abnormalities. Secondly, treatment of scl2a10 morphants with tgfbr1 inhibitor aggravated the phenotype. Thirdly, transcriptional profiling showed a significant correlation between the mRNA profile of slc2a10 knockdown and tgfbr1 inhibition. The genes affected by both treatments are related to the development of the cardiovascular system, the eye, neurogenesis and cartilage formation. Involvement of TGFβ signaling in these functions has been demonstrated before (2,2729). Fourthly, the TGFβ reporter construct 3TP-lux showed reduced promoter activity in response to slc2a10 knockdown. Fifthly, slc2a10 knockdown partially rescued the deleterious effects of reduced levels of smad7, but not vice versa. This suggests that glut10 exerts its effect on TGFβ signaling downstream of smads.

It is possible that TGFβ downregulation during early embryogenesis causes a compensatory upregulation later in development. Such a mechanism has been shown in a Tgfbr1 (Alk5) knockout mouse model. A compensatory upregulation of the ALK5 downstream pathway was noted in these mice to be mediated by activin/ALK4 signaling (30). Also, reduced TGFβ signaling caused by initially elevated sequestration may be followed by an excessive TGFβ release from a defective ECM later in life. Impaired elastic fiber formation, an important feature of ATS patients (31), results in a higher amount of ‘bare’ microfibrils that can sequester TGFβ in the ECM. As TGFβ activation is dependent on mechanical forces (32), TGFβ release may increase severely once sufficient intravascular pressure exists, a physiological variable that increases through development.

Our expression study also provides new insights into the specific molecular mechanisms involved in the ATS phenotype, as some pathways are altered by slc2a10 knockdown but not by tgfbr1 inhibition. A key finding is the downregulation of major players in cellular respiration, a process that converts glucose to the high-energy compound ATP through sequential steps of glycolysis in the cytoplasm, the Szent–Györgyi–Krebs cycle, and the oxidative phosphorylation in the mitochondria. In addition, specifically affected genes in the slc2a10 knockdown model involve the reactive oxygen species production pathway, heme biosynthesis and Ca2+ homeostasis, all important mitochondrial functions. Thus, the differential expression pattern overall points to a contribution of mitochondrial dysfunction in the phenotype caused by the loss of glut10 function in the zebrafish embryo.

Mitochondrial dysfunction in slc2a10 knockdown embryos was confirmed by our extracellular flux measurements. In spite of relatively preserved mitochondrial morphology, the loss of glut10 caused reduced overall respiration and reduced maximal flux of the electron transport chain in response to uncoupler administration. Reduced electron transport chain activity in slc2a10 knockdown embryos is consistent with reduced gene expression of electron transport chain components NADH-ubiquinone oxidoreductase 1 alpha/beta subcomplex (ndufab1) and cytochrome C1 (cyc1) (Supplementary Material, Table S2).

Mitochondrial dysfunction observed through altered transcriptional profiles in our study is consistent with the recent finding that GLUT10 is required for dehydroascorbic acid (DHA) transport into the mitochondria (33). DHA is converted to the antioxidant ascorbic acid that reduces reactive oxygen species generated as a result of oxidative phosphorylation. Consequently, defective recycling of DHA in the absence of GLUT10 results in increased sensitivity of cells to oxidative damage (33), which is expected to lead to alterations in the expression of genes required for mitochondrial function as observed in our study.

GLUT10 deficiency results in severe cardiovascular and connective tissue manifestations in both humans (5,6) and zebrafish (this study). In contrast, inactivating mutations in mouse Glut10 result in a mild, subclinical phenotype (13,14). Differences in vitamin C metabolism among species may explain these observations. Some vertebrates, including humans and teleost fish but not mice, lack gulonolactone oxidase, a key enzyme in the biosynthetic pathway of vitamin C (34). These organisms depend on dietary vitamin C and efficient intracellular recycling of this antioxidant and thus may be more susceptible to the loss of GLUT10, a DHA transporter.

Because a primary mitochondrial abnormality in our study led to decreased expression of TGFβ target genes, we conclude that at least a part of the TGFβ signaling pathway is dependent on mitochondrial function. Consistent with this notion, several studies highlighted connections between mitochondria, oxidative stress and TGFβ signaling (3538). This may occur through the coupling of intracellular oxidative pathways and TGFβ signaling by the renin–angiotensin pathway. Indeed, the angiotensinogen transcript is downregulated in the slc2a10 knockdown model and this molecule is known to enhance TGFβ signaling and ECM metabolism (39,40).

Mitochondrial dysfunction in relation to oxidative stress has recently been shown to be involved in the pathogenesis of other connective tissue disorders related to ATS. Mutations in the PYCR1 gene encoding Δ-1-pyrroline-5-carboxylate reductase 1, an mitochondrial enzyme involved in proline metabolism, cause autosomal-recessive cutis laxa type IIB, wrinkly skin syndrome and geroderma osteodysplasticum (41). Together with our findings, this illustrates that proper mitochondrial function is essential for the development and maintenance of connective tissues, in part through interactions with the TGFβ signaling pathway.


Zebrafish maintenance and microscopy

Wild-type AB and transgenic Tg(Fli1:EGFP)y1 zebrafish were reared at a constant temperature of 25°C and maintained on a 14-h light, 10-h dark photoperiod. Fish were fed three times daily with both micropellets (Hikari, Hayward, CA, USA) and brine shrimp (Biomarine, Aquafauna Bio-Marine, Hawthorne, CA, USA). After in vitro fertilization, dead embryos were removed at 8 hpf and surviving embryos were treated with 1-phenyl-2-thiourea to inhibit melanin pigmentation, dechorionated with pronase (Sigma, St. Louis, MO, USA) at 24 hpf and examined at 48 hpf. Microinjection procedures were performed using an Olympus SZX7 stereomicroscope. Live embryos were mounted in 2% methylcellulose and imaged using an Olympus MVX 10 (bright field and fluorescent) microscope equipped with Olympus MicroSuite software. For confocal microscopy, live embryos were anesthetized with tricaine, mounted in 1% low-melting agarose and imaged using a laser-scanning Olympus FV500 confocal microscope utilizing a ×10 objective. Statistical analysis was conducted using a non-parametric Kruskal–Wallis test followed by Dunn's multiple comparison post hoc test or χ2 followed by Bonferroni correction.

Morpholino-mediated knockdown

Antisense MOs (Genetools, Philomath, OR, USA) such as ATG—MO (5′-TCAGGAGCAGGACAGAACAACCCAT-3′) and splice—MO (5′-CAAATAAAGTCCACTTACTTGGTCC-3′) were directed against exons 1 and 2 regions spanning the slc2a10 ATG start codon and the exon 2–intron 2 donor splice site of the slc2a10 pre-mRNA, respectively. The MO against smad7 (5′-ATGAAACTTCAACTTACCAGGTGGT-3′) was also directed against the exon 2–intron 2 donor splice site. A standard control—MO (5′-CCTCTTACCTCAGTTACAATTTATA-3′) was used as a control. Routinely, MOs were microinjected in 1–5 nl volume into 1- to 2-cell stage embryos at 2.5 ng for slc2a10 ATG—MO, 7.5 ng for slc2a10 splice—MO and 5 ng for control—MO and 1 ng for smad7—MO. All MOs were dissolved in 0.2% phenol red and 1× Danieu's buffer [58 mm NaCl, 0.7 mm KCl, 0.4 mm MgSO4, 0.6 mm Ca(NO3)2, 5.0 mm HEPES (pH 7.6)].

Real-time qPCR

After homogenization of ten to fifteen 48 hpf zebrafish embryos, total RNA was isolated with TRIzol reagent (Invitrogen) and cDNA synthesis was performed using the SuperScript III First-Strand Synthesis System for RT–PCR with random hexamer primers (Invitrogen) in a total volume of 20 µl. Amplification efficiency (E) for each primer set was determined on the basis of a 6-fold zebrafish cDNA dilution series. Only primer pairs with E ≥ 85% were used for further experiments. PCR mixtures contained ABI SYBR Green PCR Master Mix, 0.25 µm of each forward and reverse primer and 10 ng cDNA. Cycling conditions were as follows: 10 min at 95°C, 40 cycles at 95°C for 15s, 60°C for 60 s. Subsequently, a melting curve (55–95°C) was generated for every amplicon to check PCR specificities. qPCR analysis was performed on Stratagene Mx3005P qPCR system (Agilent Technologies, Santa Clara, CA, USA). All reactions were carried out in triplicate and normalized to the geometric mean of two stable reference genes, bactin I and elfa, using qBasePlus software (42,43). Expression levels were determined in three independent experiments for each RNA extraction. Differential gene expression was considered significant if the means differed by at least 50 and the 95% confidence intervals of the means did not overlap (equivalent to P< 0.05). Oligonucleotide primers used for qPCR are available upon request.

Pharmacologic treatment

Transforming growth factor-β type I receptor kinase inhibitor or ALK5 inhibitor I [3-(pyridin-2-yl)-4-(4-quinonyl)]-1H-pyrazole] (LY-364947 or HTS-466284, #616451, EMD Chemicals, Gibbstown, NJ, USA) was prepared as a 20 mm stock in dimethyl sulfoxide. Working solutions were made in E3 chemical screening medium (44). Embryos were incubated in the compound starting at 8 hpf, dechorionated at 24 hpf and examined at 48 hpf.

Array hybridization

After quantification and quality control, RNA samples were subjected to T7 linear amplification. Amplified RNAs were chemically labeled with either cy3 or cy5 dyes. Labeled RNA samples were quantified, equalized by mass, paired and combined to test treatment effects. Three biological replicates were used. The paired and balanced RNAs were suspended in Agilent 2× Gene Expression buffer (55 μl), Agilent 10× Blocking agent (11 μl) and Kreablock (27.5 μl). The hybridization solutions were applied to Agilent Zebrafish v2 4 × 44K microarrays. Hybridization was carried out at 65°C for 20 h. Washing procedures were carried out according to Agilent gene expression protocols. Slides were scanned on an Axon 4000B scanner to detect Cy3 and Cy5 fluorescence. Laser power was kept constant for Cy3/Cy5 scans and the photomultiplier tube setting (PMT) was varied for each experiment based on the optimal signal intensity with lowest possible background fluorescence. A low PMT setting scan was also performed to recover signals from saturated elements. Gridding and analysis of images was performed using Genepix v6.1 (Axon, Molecular Devices, Sunnyvale, CA, USA). A Partek Genomics Suite (Partek, St. Louis, MO, USA) was used to normalize and statistically analyze the data. The microarray data set has been deposited into the NCBI Gene Expression Omnibus ( and is available through accession number GSE34510.

Electron microscopy

Embryos were dechorionated at 52 hpf, fixed in glutaraldehyde, stained sequentially with OsO4, tannic acid and uranyl acetate, dehydrated and embedded in Epon (45). Thin sections (60 nm) were cut, placed on formvar-coated grids and counterstained with 7% methanolic uranyl acetate and lead citrate. Sections were viewed with a Tecnai 12 transmission electron microscope at 120 kV, and the images were digitally captured.

The OCR measurement

The OCR was measured in developing zebrafish embryos using XF24 Extracellular Flux Analyzer (Seahorse Bioscience, Billerica, MA, USA). Four control or slca2a10—MO-injected embryos were loaded into each of 5–10 wells of XF24 islet capture microplates (Seahorse Bioscience) at 3 or 20 hpf and incubated at 37°C in E3 solution. After an OCR measurement of 2 h, FCCP was added to 1 µm final concentration and incubated for 1 h. Next, rotenone was added to each well to reach a final concentration of 4 µm followed by a 1 h incubation. The OCR measurements were taken every 6 min after a 3 min mixing period.

Luciferase assay

To verify the role of glut10 in facilitating TGFβ signaling, 1-cell stage wild-type embryos were co-injected with 75 pg pgl2-basic 3TP-lux with 7.5 ng slc2a10—MO or 5 ng control—MO. Ten embryos per replicate were harvested at 24 hpf and lysed in reporter lysis buffer (Promega, Madison, WI, USA) using a pestle homogenizer. The lysates were cleared by centrifugation at 13 000 rpm in a microcentrifuge, and a 10-μl aliquot of each of the four biological replicates were analyzed using a luciferase assay system (Promega) and a Genios plate reader (Tecan, Durham, NC, USA) and results were recorded in relative luminescence units (RLUs). An aliquot of the lysates was used to measure the protein concentration. As all protein concentrations were within ±10% of the mean, we did not correct the RLU readings for protein concentration.


This work was supported by the National Institutes of Health (HL084922, HL090648 to Z.U.), by the March of Dimes (#1-FY09-402 to Z.U.), by Ghent University (Methusalem grant BOF08/01M01108 to A.D.P.) and by the Fighting Aneurysmal Disease (EC-FP7 to P.J.C. and A.D.P.). A.W. was a research fellow and B.L.C is a postdoctoral fellow of the Fund for Scientific Research – Flanders. The Genome Technology Access Center in the Department of Genetics at Washington University School of Medicine assisted with genomic analysis. The center is partially supported by NCI Cancer Center Support Grant #P30 CA91842 to the Siteman Cancer Center and by ICTS/CTSA Grant# UL1RR024992 from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH), and NIH Roadmap for Medical Research. This publication is solely the responsibility of the authors and does not necessarily represent the official view of NCRR or NIH.

Supplementary Material

Supplementary Data:


We thank Dr Mary O. Carayannopoulos for advice on the glut family of proteins, Dr Matthew I. Goldsmith for help with zebrafish methods, Michael E. Heinz for technical help with array hybridization and Dr Beth L. Roman for the p3TP-lux plasmid. We appreciate the support of the Washington University Pediatric Zebrafish Facility and the University of Pittsburgh Zebrafish Facility.

Conflict of Interest statement: None declared.


1. Ma X.M., Blenis J. Molecular mechanisms of mTOR-mediated translational control. Nat. Rev. Mol. Cell Biol. 2009;10:307–318. [PubMed]
2. Goumans M.J., Liu Z., ten Dijke P. TGF-beta signaling in vascular biology and dysfunction. Cell Res. 2009;19:116–127. [PubMed]
3. Nian M., Lee P., Khaper N., Liu P. Inflammatory cytokines and postmyocardial infarction remodeling. Circ. Res. 2004;94:1543–1553. [PubMed]
4. Hayden M.R., Reidy M. Many roads lead to atheroma. Nat. Med. 1995;1:22–23. [PubMed]
5. Callewaert B.L., Willaert A., Kerstjens-Frederikse W.S., De Backer J., Devriendt K., Albrecht B., Ramos-Arroyo M.A., Doco-Fenzy M., Hennekam R.C., Pyeritz R.E., et al. Arterial tortuosity syndrome: clinical and molecular findings in 12 newly identified families. Hum. Mutat. 2008;29:150–158. [PubMed]
6. Coucke P.J., Willaert A., Wessels M.W., Callewaert B., Zoppi N., De Backer J., Fox J.E., Mancini G.M., Kambouris M., Gardella R., et al. Mutations in the facilitative glucose transporter GLUT10 alter angiogenesis and cause arterial tortuosity syndrome. Nat. Genet. 2006;38:452–457. [PubMed]
7. Neptune E.R., Frischmeyer P.A., Arking D.E., Myers L., Bunton T.E., Gayraud B., Ramirez F., Sakai L.Y., Dietz H.C. Dysregulation of TGF-beta activation contributes to pathogenesis in Marfan syndrome. Nat. Genet. 2003;33:407–411. [PubMed]
8. Loeys B.L., Chen J., Neptune E.R., Judge D.P., Podowski M., Holm T., Meyers J., Leitch C.C., Katsanis N., Sharifi N., et al. A syndrome of altered cardiovascular, craniofacial, neurocognitive and skeletal development caused by mutations in TGFBR1 or TGFBR2. Nat. Genet. 2005;37:275–281. [PubMed]
9. Hucthagowder V., Sausgruber N., Kim K.H., Angle B., Marmorstein L.Y., Urban Z. Fibulin-4: a novel gene for an autosomal recessive cutis laxa syndrome. Am. J. Hum. Genet. 2006;78:1075–1080. [PubMed]
10. Renard M., Holm T., Veith R., Callewaert B.L., Ades L.C., Baspinar O., Pickart A., Dasouki M., Hoyer J., Rauch A., et al. Altered TGFbeta signaling and cardiovascular manifestations in patients with autosomal recessive cutis laxa type I caused by fibulin-4 deficiency. Eur. J. Hum. Genet. 2010;18:895–901. [PMC free article] [PubMed]
11. Cohn R.D., van Erp C., Habashi J.P., Soleimani A.A., Klein E.C., Lisi M.T., Gamradt M., ap Rhys C.M., Holm T.M., Loeys B.L., et al. Angiotensin II type 1 receptor blockade attenuates TGF-beta-induced failure of muscle regeneration in multiple myopathic states. Nat. Med. 2007;13:204–210. [PMC free article] [PubMed]
12. Habashi J.P., Judge D.P., Holm T.M., Cohn R.D., Loeys B.L., Cooper T.K., Myers L., Klein E.C., Liu G., Calvi C., et al. Losartan, an AT1 antagonist, prevents aortic aneurysm in a mouse model of Marfan syndrome. Science. 2006;312:117–121. [PMC free article] [PubMed]
13. Callewaert B.L., Loeys B.L., Casteleyn C., Willaert A., Dewint P., De Backer J., Sedlmeier R., Simoens P., De Paepe A.M., Coucke P.J. Absence of arterial phenotype in mice with homozygous slc2A10 missense substitutions. Genesis. 2008;46:385–389. [PubMed]
14. Cheng C.H., Kikuchi T., Chen Y.H., Sabbagha N.G., Lee Y.C., Pan H.J., Chang C., Chen Y.T. Mutations in the SLC2A10 gene cause arterial abnormalities in mice. Cardiovasc. Res. 2009;81:381–388. [PubMed]
15. Chico T.J., Ingham P.W., Crossman D.C. Modeling cardiovascular disease in the zebrafish. Trends Cardiovasc. Med. 2008;18:150–155. [PubMed]
16. Chiarelli N., Ritelli M., Zoppi N., Benini A., Borsani G., Barlati S., Colombi M. Characterization and expression pattern analysis of the facilitative glucose transporter 10 gene (slc2a10) in Danio rerio. Int. J. Dev. Biol. 2011;55:229–236. [PubMed]
17. Thisse C., Zon L.I. Organogenesis—heart and blood formation from the zebrafish point of view. Science. 2002;295:457–462. [PubMed]
18. Li H.Y., Wang Y., Heap C.R., King C.H., Mundla S.R., Voss M., Clawson D.K., Yan L., Campbell R.M., Anderson B.D., et al. Dihydropyrrolopyrazole transforming growth factor-beta type I receptor kinase domain inhibitors: a novel benzimidazole series with selectivity versus transforming growth factor-beta type II receptor kinase and mixed lineage kinase-7. J. Med. Chem. 2006;49:2138–2142. [PubMed]
19. Singh J., Chuaqui C.E., Boriack-Sjodin P.A., Lee W.C., Pontz T., Corbley M.J., Cheung H.K., Arduini R.M., Mead J.N., Newman M.N., et al. Successful shape-based virtual screening: the discovery of a potent inhibitor of the type I TGFbeta receptor kinase (TbetaRI) Bioorg. Med. Chem. Lett. 2003;13:4355–4359. [PubMed]
20. Lowell B.B., Spiegelman B.M. Towards a molecular understanding of adaptive thermogenesis. Nature. 2000;404:652–660. [PubMed]
21. Wrana J.L., Attisano L., Carcamo J., Zentella A., Doody J., Laiho M., Wang X.F., Massague J. TGF beta signals through a heteromeric protein kinase receptor complex. Cell. 1992;71:1003–1014. [PubMed]
22. Park S.O., Lee Y.J., Seki T., Hong K.H., Fliess N., Jiang Z., Park A., Wu X., Kaartinen V., Roman B.L., et al. ALK5- and TGFBR2-independent role of ALK1 in the pathogenesis of hereditary hemorrhagic telangiectasia type 2. Blood. 2008;111:633–642. [PubMed]
23. Sumanas S., Larson J.D. Morpholino phosphorodiamidate oligonucleotides in zebrafish: a recipe for functional genomics? Brief Funct. Genomic Proteomic. 2002;1:239–256. [PubMed]
24. Fouquet B., Weinstein B.M., Serluca F.C., Fishman M.C. Vessel patterning in the embryo of the zebrafish: guidance by notochord. Dev. Biol. 1997;183:37–48. [PubMed]
25. Christiansen H.E., Lang M.R., Pace J.M., Parichy D.M. Critical early roles for col27a1a and col27a1b in zebrafish notochord morphogenesis, vertebral mineralization and post-embryonic axial growth. PLoS One. 2009;4:e8481. [PMC free article] [PubMed]
26. Hu Q., Shifren A., Sens C., Choi J., Szabo Z., Starcher B.C., Knutsen R.H., Shipley J.M., Davis E.C., Mecham R.P., et al. Mechanisms of emphysema in autosomal dominant cutis laxa. Matrix Biol. 2010;29:621–628. [PubMed]
27. Gomes F.C., Sousa Vde O., Romao L. Emerging roles for TGF-beta1 in nervous system development. Int. J. Dev. Neurosci. 2005;23:413–424. [PubMed]
28. Lin Z., Willers C., Xu J., Zheng M.H. The chondrocyte: biology and clinical application. Tissue Eng. 2006;12:1971–1984. [PubMed]
29. Saika S. TGFbeta pathobiology in the eye. Lab Invest. 2006;86:106–115. [PubMed]
30. Carvalho R.L., Itoh F., Goumans M.J., Lebrin F., Kato M., Takahashi S., Ema M., Itoh S., van Rooijen M., Bertolino P., et al. Compensatory signalling induced in the yolk sac vasculature by deletion of TGFbeta receptors in mice. J. Cell Sci. 2007;120:4269–4277. [PubMed]
31. Pletcher B.A., Fox J.E., Boxer R.A., Singh S., Blumenthal D., Cohen T., Brunson S., Tafreshi P., Kahn E. Four sibs with arterial tortuosity: description and review of the literature. Am. J. Med. Genet. 1996;66:121–128. [PubMed]
32. Wipff P.J., Rifkin D.B., Meister J.J., Hinz B. Myofibroblast contraction activates latent TGF-beta1 from the extracellular matrix. J. Cell Biol. 2007;179:1311–1323. [PMC free article] [PubMed]
33. Lee Y.C., Huang H.Y., Chang C.J., Cheng C.H., Chen Y.T. Mitochondrial GLUT10 facilitates dehydroascorbic acid import and protects cells against oxidative stress: mechanistic insight into arterial tortuosity syndrome. Hum. Mol. Genet. 2010;19:3721–3733. [PubMed]
34. Dabrowski K. Gulonolactone oxidase is missing in teleost fish. The direct spectrophotometric assay. Biol. Chem. Hoppe Seyler. 1990;371:207–214. [PubMed]
35. de Cavanagh E.M., Ferder M., Inserra F., Ferder L. Angiotensin II, mitochondria, cytoskeletal, and extracellular matrix connections: an integrating viewpoint. Am. J. Physiol. Heart Circ. Physiol. 2009;296:H550–H558. [PubMed]
36. Guo P., Nishiyama A., Rahman M., Nagai Y., Noma T., Namba T., Ishizawa M., Murakami K., Miyatake A., Kimura S., et al. Contribution of reactive oxygen species to the pathogenesis of left ventricular failure in Dahl salt-sensitive hypertensive rats: effects of angiotensin II blockade. J. Hypertens. 2006;24:1097–1104. [PubMed]
37. Mitchell C., Robin M.A., Mayeuf A., Mahrouf-Yorgov M., Mansouri A., Hamard M., Couton D., Fromenty B., Gilgenkrantz H. Protection against hepatocyte mitochondrial dysfunction delays fibrosis progression in mice. Am. J. Pathol. 2009;175:1929–1937. [PubMed]
38. Zhao W., Zhao T., Chen Y., Ahokas R.A., Sun Y. Oxidative stress mediates cardiac fibrosis by enhancing transforming growth factor-beta1 in hypertensive rats. Mol. Cell Biochem. 2008;317:43–50. [PubMed]
39. Campbell S.E., Katwa L.C. Angiotensin II stimulated expression of transforming growth factor-beta1 in cardiac fibroblasts and myofibroblasts. J. Mol. Cell Cardiol. 1997;29:1947–1958. [PubMed]
40. Schultz Jel J., Witt S.A., Glascock B.J., Nieman M.L., Reiser P.J., Nix S.L., Kimball T.R., Doetschman T. TGF-beta1 mediates the hypertrophic cardiomyocyte growth induced by angiotensin II. J. Clin. Invest. 2002;109:787–796. [PMC free article] [PubMed]
41. Reversade B., Escande-Beillard N., Dimopoulou A., Fischer B., Chng S.C., Li Y., Shboul M., Tham P.Y., Kayserili H., Al-Gazali L., et al. Mutations in PYCR1 cause cutis laxa with progeroid features. Nat. Genet. 2009;41:1016–1021. [PubMed]
42. Hellemans J., Mortier G., De Paepe A., Speleman F., Vandesompele J. qBase relative quantification framework and software for management and automated analysis of real-time quantitative PCR data. Genome Biol. 2007;8:R19. [PMC free article] [PubMed]
43. McCurley A.T., Callard G.V. Characterization of housekeeping genes in zebrafish: male–female differences and effects of tissue type, developmental stage and chemical treatment. BMC Mol. Biol. 2008;9:102. [PMC free article] [PubMed]
44. Murphey R.D., Zon L.I. Small molecule screening in the zebrafish. Methods. 2006;39:255–261. [PubMed]
45. Davis E.C. Smooth muscle cell to elastic lamina connections in developing mouse aorta. Role in aortic medial organization. Lab. Invest. 1993;68:89–99. [PubMed]
46. Larkin M.A., Blackshields G., Brown N.P., Chenna R., McGettigan P.A., McWilliam H., Valentin F., Wallace I.M., Wilm A., Lopez R., et al. Clustal W and Clustal X version 2.0. Bioinformatics. 2007;23:2947–2948. [PubMed]
47. Dawson P.A., Mychaleckyj J.C., Fossey S.C., Mihic S.J., Craddock A.L., Bowden D.W. Sequence and functional analysis of GLUT10: a glucose transporter in the Type 2 diabetes-linked region of chromosome 20q12–13.1. Mol. Genet. Metab. 2001;74:186–199. [PubMed]
48. Joost H.G., Thorens B. The extended GLUT-family of sugar/polyol transport facilitators: nomenclature, sequence characteristics, and potential function of its novel members (review) Mol. Membr Biol. 2001;18:247–256. [PubMed]

Articles from Human Molecular Genetics are provided here courtesy of Oxford University Press