Laminar Shear Stress Inhibits Endothelial Cell Metabolism via KLF2-Mediated Repression of PFKFB3Significance
Objective—Cellular metabolism was recently shown to regulate endothelial cell phenotype profoundly. Whether the atheroprotective biomechanical stimulus elicited by laminar shear stress modulates endothelial cell metabolism is not known.
Approach and Results—Here, we show that laminar flow exposure reduced glucose uptake and mitochondrial content in endothelium. Shear stress–mediated reduction of endothelial metabolism was reversed by silencing the flow-sensitive transcription factor Krüppel-like factor 2 (KLF2). Endothelial-specific deletion of KLF2 in mice induced glucose uptake in endothelial cells of perfused hearts. KLF2 overexpression recapitulates the inhibitory effects on endothelial glycolysis elicited by laminar flow, as measured by Seahorse flux analysis and glucose uptake measurements. RNA sequencing showed that shear stress reduced the expression of key glycolytic enzymes, such as 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase-3 (PFKFB3), phosphofructokinase-1, and hexokinase 2 in a KLF2-dependent manner. Moreover, KLF2 represses PFKFB3 promoter activity. PFKFB3 knockdown reduced glycolysis, and overexpression increased glycolysis and partially reversed the KLF2-mediated reduction in glycolysis. Furthermore, PFKFB3 overexpression reversed KLF2-mediated reduction in angiogenic sprouting and network formation.
Conclusions—Our data demonstrate that shear stress–mediated repression of endothelial cell metabolism via KLF2 and PFKFB3 controls endothelial cell phenotype.
- shear stress down regulated gene-1 protein, human
Endothelial cells form the inner lining of all blood vessels and not only regulate transport of nutrients to the underlying tissue but also coordinate the formation of new blood vessels, a process termed angiogenesis. Therefore, endothelial cells are highly plastic cells that are capable of switching from a resting quiescent state in normal conduit blood vessels to a highly proliferative and migratory state when angiogenesis takes place. Resting quiescent endothelial cells are termed phalanx cells,1 whereas migratory angiogenic endothelial cells are referred to as tip cells, which are followed by proliferating so-called stalk cells.2 Although the mechanisms regulating tip and stalk cell behavior have been extensively studied, relatively little is known about the control of the phalanx state.
Shear stress, the force that laminar blood flow exerts on endothelial cells, is thought to be one of the factors that determine the quiescent state of endothelial cells.3 This biomechanical stimulus induces the expression of the transcription factor Krüppel-like factor 2 (KLF2), which orchestrates a network of genes that elicit a quiescent endothelial cell phenotype.4,5 Among the factors that are upregulated by KLF2 are anti-inflammatory and antithrombotic proteins, whereas proinflammatory and prothrombotic factors are downregulated by KLF2.4 Although not all effects of shear stress on endothelial cells are mediated by KLF2, KLF2 coordinates approximately half of the gene expression programs evoked by shear stress.5,6
See accompanying editorial on page 13
Recent studies have highlighted the importance of cellular metabolism for the control of endothelial cell phenotype.7,8 Particularly, it was shown that angiogenic endothelial cells rely heavily on glycolysis for migration and proliferation.9 The enzyme PFKFB3 is a key regulator of glycolysis in endothelial cells that has been shown to promote angiogenic sprouting.9–11 However, how resting endothelial cells control their metabolic activity and whether this affects the functional properties of the phalanx phenotype is unclear.
Here, we show that the biomechanical signal shear stress, through the upregulation of KLF2, reduces endothelial metabolic activity by repressing PFKFB3 expression, and thereby maintains a metabolic quiescent phenotype reminiscent of phalanx cells.
Materials and Methods
Materials and Methods are available in the online-only Data Supplement.
Shear Stress Reduces Endothelial Glucose Uptake and Mitochondrial Content in a KLF2-Dependent Manner
To assess the effects of laminar shear stress on endothelial cell metabolism, human umbilical vein endothelial cells (HUVECs) were exposed to laminar shear stress (20 dynes/cm2) for 72 hours to achieve steady-state quiescence or left under static conditions. Glucose uptake and mitochondrial content analysis revealed that shear stress reduces basal endothelial cell glucose uptake and the relative quantity of mitochondria per endothelial cell (Figure 1A and 1B). These results were substantiated by fluorescence microscopy–based analysis of glucose uptake of individual HUVECs exposed to shear stress, which showed that cellular alignment to the flow direction inversely correlates with glucose uptake (Figure IA in the online-only Data Supplement). Because the transcription factor KLF2 is known to be responsible for many shear stress–induced effects on endothelial cells, we determined whether reduction in glucose uptake by shear stress is dependent on KLF2. To this end, we transduced HUVECs with a lentiviral short hairpin RNA construct to silence KLF2 and subsequently subjected the cells to laminar shear stress for 72 hours (Figure 1C), which completely abrogates shear stress–mediated induction of KLF2. Silencing of KLF2 abolished shear stress–mediated reduction of glucose uptake, indicating that regulation of glucose uptake by shear stress is KLF2 dependent (Figure 1D).
To substantiate whether KLF2 regulates metabolic activity of endothelial cells in vivo, we analyzed glucose uptake in endothelial cells of mice lacking endothelial KLF2 (Cdh5-CreERT2;KLF2fl/fl) and wild-type controls (KLF2fl/fl and KLF2fl/+; Figure IB in the online-only Data Supplement). Specifically, hearts of these mice were subjected to Langendorff-mediated perfusion with 2-N-7-nitrobenz-2-oxa-1,3-diazol-4-yl-amino-2-deoxyglucose and simultaneous digestion of the extracellular matrix to obtain a single-cell suspension of cardiac cells. Then, endothelial cells were labeled, and glucose uptake in these cells was quantified using flow cytometry (Figure 1E–1G). Endothelial-specific deletion of KLF2 in mice significantly induces glucose uptake by cardiac endothelial cells (Figure 1G).
KLF2 Reduces Endothelial Metabolic Activity
Overexpression of KLF2 in endothelial cells recapitulates many aspects of shear stress stimulation, including induction of cellular quiescence.4,12 Lentiviral overexpression of KLF2 (Figure 2A) in endothelial cells mimics the induction of KLF2 after shear stress stimulation (Figure 1C) and indeed reduces glucose uptake (Figure 2B). Furthermore, using a transwell assay to model glucose availability to underlying tissue in vitro, we measured an increase in available glucose underneath endothelial monolayers that overexpress KLF2 (Figure 2C), indicating that the reduction of glucose consumption mediated by KLF2 results in more bioavailability of glucose underneath the endothelium. Using Seahorse Flux analysis, we determined that the extracellular acidification rate, an indicator of lactate production, is also lower in KLF2-overexpressing endothelial cells (Figure 2D). Not only basal acidification rate but also glucose-induced glycolysis and maximal glycolytic capacity were reduced in KLF2-transduced cells when compared with mock-transduced control cells (Figure 2E).
Next, we determined whether mitochondrial content and function are also regulated by KLF2. KLF2 overexpression reduces mitochondrial content (Figure 2F), metabolic activity (measured using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; Figure 2G), and ATP levels (Figure 2H), which is similar to the effects of shear stress on endothelial cells (Figure 1). Importantly, mitochondrial membrane potential is significantly enhanced after KLF2 overexpression, suggesting that KLF2 reduces mitochondrial activity, rather than affecting mitochondrial integrity (Figure 2I). Assessment of cellular oxygen consumption revealed that KLF2 overexpression also reduces basal mitochondrial respiration and ATP production (Figure 2J and 2K), but does not affect maximal respiration or spare respiratory capacity, further corroborating that KLF2 does not affect mitochondrial integrity. Because both oxygen consumption and lactate production are lower in KLF2-overexpressing cells, the ratio of oxygen consumption over lactate production rate is not altered by KLF2, arguing against a potential KLF2-mediated shift between aerobic respiration and glycolysis (Figure IC in the online-only Data Supplement).
KLF2 Does Not Induce Senescence or Apoptosis
To exclude the possibility that the KLF2-mediated reduction in metabolic activity is because of induction of senescence, we analyzed proliferation (Figure 3A), acidic β-galactosidase activity (Figure 3B), and p21 expression (Figure 3C) in KLF2-transduced and mock control cells. Whereas KLF2 slightly reduces the number of cells in G2/M-phase, KLF2 overexpression reduced β-galactosidase activity and p21 expression, indicating that reduction in proliferation by KLF2 is not because of senescence. Furthermore, KLF2 reduces apoptosis, as measured by caspase 3/7 activation (Figure 3D) and annexin V staining (Figure ID in the online-only Data Supplement).
KLF2-Mediated Suppression of Metabolic Activity Is Not Mediated via AMPK or Nitric Oxide
To delineate how KLF2 reduces endothelial cell metabolic activity, we first performed phospho-kinase proteome profiling (Figure IIA in the online-only Data Supplement), which showed that KLF2 overexpression reduces phosphorylation of the 5′ adenosine monophosphate–activated protein kinase α 1 subunit (Figure IIB in the online-only Data Supplement). However, silencing 5′ adenosine monophosphate–activated protein kinase α 1 (Figure IIC in the online-only Data Supplement) did not recapitulate any of the metabolic effects observed after KLF2 overexpression (Figure IID–IIG in the online-only Data Supplement). Conversely, KLF2 overexpression induces endothelial nitric oxide synthase expression and phosphorylation (Figure IIA in the online-only Data Supplement). Because nitric oxide has been shown to inhibit mitochondrial respiration,13 increased nitric oxide production could potentially account for the inhibition of mitochondrial activity by KLF2. However, inhibition of nitric oxide synthesis did not affect the inhibitory effect of KLF2 on respiration or glycolysis (Figure IIH and III in the online-only Data Supplement).
KLF2 and Shear Stress Inhibit PFKFB3 Expression
To gain a broad unbiased insight into the regulation of gene expression by shear stress, HUVECs were subjected to laminar shear stress for 72 hours or kept under static control conditions, and we performed next-generation sequencing with RNA isolated from these cells (RNAseq). These experiments showed that shear stress stimulation downregulates the expression of many genes involved in glycolysis (Figure 4A), including the genes encoding hexokinase, phosphofructokinase, 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase (PFKFB). Hexokinase converts glucose into glucose 6-phosphate, phosphofructokinase converts fructose 6-phosphate into fructose 1,6-bisphosphate, and PFKFB converts 6-phosphate into fructose 2,6-biphosphate, a strong allosteric activator of phosphofructokinase (Figure 4A). Because the inhibition of expression of these proteins could explain the inhibitory effects of KLF2 on glycolysis and metabolic activity, we next focused on the regulation of expression of these proteins.
To substantiate the observations from our RNAseq experiments (Figure 4A), we confirmed the downregulation of hexokinase-2, PFKFB3, and phosphofructokinase platelet isoform (PFK1) mRNA expression by quantitative real-time polymerase chain reaction after 72-hour laminar shear stress stimulation (Figure 4B). KLF2 overexpression similarly inhibited the expression of hexokinase-2, PFKFB3, and PFK1 mRNA when compared with mock-transduced cells, as measured by quantitative real-time polymerase chain reaction (Figure 4C). Western blotting confirmed that hexokinase-2, PFKFB3, and PFK1 are also significantly inhibited by KLF2 on protein level (Figure 4D–4F). Next, we determined whether the inhibition of PFKFB3 and PFK1 expression by shear stress is dependent on KLF2. To this end, we exposed HUVECs transduced with shKLF2 lentivirus or control lentivirus to shear stress for 72 hours or kept the cells under static conditions and measured the expression of PFKFB3 and PFK1 (Figure 4G). These experiments show that in the absence of KLF2, shear stress stimulation does not repress the expression of PFKFB3 and PFK1. Because KLF2 can either activate or repress transcription by binding to specificity protein 1/KLF sites,14 we examined the PFK1 and PFKFB3 promoters for consensus specificity protein 1/KLF sites and identified a potential KLF2 binding site 14 bp upstream of the PFKFB3 transcriptional start site (Figure III in the online-only Data Supplement). We then placed the 300 bp conserved part upstream of the PFKFB3 promoter in front of a luciferase transporter construct and measured luciferase activity after transfection into HUVECs that were either mock or KLF2 transduced (Figure 4H). KLF2 overexpression markedly reduced PFKFB3 promoter activity and mutation of the specificity protein 1/KLF site in the PFKFB3 promoter abolished the KLF2-mediated repression of promoter activity. Together, these data indicate that shear stress represses the expression of key glycolytic enzymes, in particular, PFKFB3 via direct inhibition of promoter activity by KLF2.
KLF2 Controls Endothelial Cell Glycolysis and Angiogenic Phenotype Partly via PFKFB3 Inhibition
PFKFB3 was recently shown to be an important regulator of endothelial cell glycolysis,9 and we therefore hypothesized that inhibition of PFKFB3 could be the underlying mechanism by which KLF2 reduces endothelial cell glycolysis. First, we set out to confirm the role of PFKFB3 in endothelial cell glycolysis. To this end, we transfected HUVECs with control small interfering RNA or 2 distinct small interfering RNAs targeting PFKFB3, which both significantly reduced PFKFB3 levels (Figure 5A). Seahorse Flux analysis of these cells showed that robust knockdown of PFKFB3 in HUVECs significantly reduces acidification rate in the presence of glucose (glycolysis) and also reduces maximal glycolytic capacity of HUVECs (Figure 5B and 5C). Interestingly, glycolytic reserve capacity (lactate production in the absence of glucose and mitochondrial function) and cellular oxygen consumption (indicative of mitochondrial activity) are not affected by PFKFB3 depletion (Figure 5B–5D), indicating that PFKFB3 preferentially controls glycolysis.
Next, we tested whether a reduction in PFKFB3 levels is the cause of KLF2-mediated inhibition of glycolysis. Hence, we overexpressed KLF2 in combination with lentiviral overexpression of PFKFB3 in endothelial cells, resulting in PFKFB3 levels in KLF2 overexpressing cells that are comparable with mock-transduced HUVECs (Figure IV in the online-only Data Supplement). Seahorse Flux analysis showed that PFKFB3 overexpression in KLF2 overexpressing cells augmented glycolysis but did not affect basal extracellular acidification rate (in the absence of glucose) (Figure 5E–5H). Consistent with PFKFB3 depletion experiments, overexpression of PFKFB3 did not affect glycolytic reserve capacity or oxygen consumption (Figure 5I and 5J). These experiments suggest that KLF2 reduces endothelial cell glycolytic function, in part, via repression of PFKFB3 expression.
KLF2 has been proposed to repress angiogenic behavior of endothelial cells, via its ability to induce endothelial cell quiescence,15 and PFKFB3 is known to be important for angiogenesis.9 We, therefore, determined whether KLF2 reduces endothelial angiogenic function via reduction of PFKFB3. An endothelial spheroid sprouting assay, which allows endothelial cell outgrowth into a 3-dimensional (3D) matrix, showed that KLF2 indeed reduces endothelial cell sprouting (Figure 6A). Interestingly, PFKFB3 overexpression significantly induced sprouting of KLF2 overexpressing endothelial cells. These results were confirmed in a second angiogenesis assay, where endothelial cells form a 2D vessel network (Figure 6B). Here, KLF2 likewise reduced network formation, whereas PFKFB3 overexpression significantly reversed KLF2-mediated inhibition of endothelial network formation. Conversely, endothelial cell–specific deletion of KLF2 induces sprouting angiogenesis in an aortic ring outgrowth assay, which is dependent on PFKFB3, as shown by treatment with the PFKFB3 inhibitor 3-PO (3-(3-pyridinyl)-1-(4-pyridinyl)-2-propen-1-one; Figure 6C). Together, these experiments show that KLF2 represses endothelial cell metabolism, in part, through transcriptional inhibition of PFKFB3, thereby inducing a metabolic quiescent endothelial cell phenotype (Figure 6D).
Laminar blood flow induces a quiescent endothelial cell phenotype. It has recently been shown that endothelial cell metabolism, and in particular glycolysis, is a key determinant of endothelial cell phenotype.7 Here, we show that fluid shear stress induces endothelial cell metabolic quiescence, via induction of the key transcriptional regulator KLF2. Laminar flow reduces the expression of several genes involved in glucose metabolism and, in particular, PFKFB3 via KLF2-mediated repression of its promoter activity. These results provide a mechanism by which a biomechanical stimulus reduces endothelial cell metabolic activity, thereby inducing a quiescent phalanx cell-like phenotype.
It has been suggested that factors that induce endothelial tip cell phenotype, such as hypoxia-inducible factor 1-alpha activation and vascular endothelial growth factor or fibroblast growth factor stimulation induce glycolysis in endothelial cells, dependent on PFKFB3.9,16 Because KLF2 is known to inhibit hypoxia-inducible factor 1-alpha15 and vascular endothelial growth factor signaling,17 these mechanisms may contribute to the effects of KLF2 on glycolysis. Furthermore, KLF2 heterozygous mice, which would putatively display increased PFKFB3 expression, show an increase in capillary density15 that is perhaps driven by an increase in glycolysis.
It has recently been shown that PFKFB3-induced glycolysis is required for proper angiogenic behavior in vitro and in vivo,9,11 and that transient chemical inhibition of PFKFB3 blocks pathological angiogenesis.10 How PFKFB3 expression is physiologically regulated in endothelial cells has not been described before. Furthermore, shear stress and KLF2 also regulate the expression of several other proteins involved in glycolysis, including hexokinase-2 and PFK1. The inhibition of expression of these proteins could additionally contribute to the metabolically quiescent phenotype elicited by shear stress or KLF2 overexpression in endothelial cells. In fact, PFKFB3 overexpression alone does induce endothelial sprouting and tube formation activity but does not fully alleviate the KLF2-mediated inhibition of endothelial cell activation (Figures 5 and 6). Induction of hexokinase-2 and PFK1 is probably not required to overcome the quiescent state induced by KLF2 fully because these 2 proteins do not seem to be required for endothelial glycolysis (Figure V in the online-only Data Supplement). The role of PFKFB3 in shear stress–induced endothelial phenotype in vivo also remains to be established.
Next to repression of glycolytic activity, KLF2 and shear stress also reduce mitochondrial content and activity. Even though endothelial cells do not rely on oxidative phosphorylation for generating ATP,9 mitochondria are known to induce oxidative stress, regulate calcium signaling, and control apoptosis.18 Interestingly, KLF2 and shear stress are known to reduce oxidative stress in endothelial cells19 and inhibit apoptosis (Figure 3). The reduction of mitochondrial content and increase in membrane potential by KLF2 could contribute to these antiapoptotic and atheroprotective cellular effects of shear stress. Even though KLF2 and shear stress both repress mitochondrial metabolism, the shear stress–mediated repression seems to be independent of KLF2 (data not shown), indicating that other shear stress–induced signaling pathways repress mitochondrial content and activity.
The function of the endothelium is highly dynamic. On hypoxia, endothelial cells need to migrate and proliferate to form new blood vessels rapidly, but under homeostatic resting conditions the endothelium needs to maintain a quiescent state. In this resting state, endothelial cells need to facilitate proper transfer of nutrients and oxygen to the underlying tissues. Indeed, KLF2 reduces metabolic activity of endothelial cells, which results in increased levels of glucose underneath the endothelial cells in our transwell experiments (Figure 2C). It is tempting to speculate that the shear stress–mediated reduction in endothelial cell glucose consumption ensures a maximal delivery of nutrients and oxygen to the adjacent tissues. Furthermore, the present data provide a molecular mechanism for the well-known and important switch of endothelial cell activation and angiogenesis back to a quiescent functional endothelium once blood flow is established.
We thank Dr Auwerx (EPFL [École polytechnique Fédérale de Lausanne], Switzerland) for discussions and advice and Denise Berghäuser and Ariane Fischer for technical support.
Sources of Funding
The study was supported by the German Center for Cardiovascular Research DZHK (BMBF) to S. Dimmeler and W. Chen, the LOEWE Center for Cell and Gene Therapy (State of Hessen) to A. Doddaballapur, R.A. Boon, and S. Dimmeler, a VENI grant (91613050) from NWO/ZonMw to R.H.H., the European Research Council (Advanced grant Angiomirs) to S. Dimmeler and the Deutsche Forschungsgemeinschaft (SFB834/B9) to R.A. Boon.
The online-only Data Supplement is available with this article at http://atvb.ahajournals.org/lookup/suppl/doi:10.1161/ATVBAHA.114.304277/-/DC1.
- Nonstandard Abbreviations and Acronyms
- human umbilical vein endothelial cells
- Krüppel-like factor 2
- Received July 7, 2014.
- Accepted October 16, 2014.
- © 2014 American Heart Association, Inc.
- Lee D-Y,
- Lee C-I,
- Lin T-E,
- Lim SH,
- Zhou J,
- Tseng Y-C,
- Chien S,
- Chiu J-J
- Dekker RJ,
- Boon RA,
- Rondaij MG,
- Kragt A,
- Volger OL,
- Elderkamp YW,
- Meijers JC,
- Voorberg J,
- Pannekoek H,
- Horrevoets AJ
- Eichmann A,
- Simons M
- Xu Y,
- An X,
- Guo X,
- et al
- Atkins GB,
- Jain MK
- Conkright MD,
- Wani MA,
- Lingrel JB
- Kawanami D,
- Mahabeleshwar GH,
- Lin Z,
- Atkins GB,
- Hamik A,
- Haldar SM,
- Maemura K,
- Lamanna JC,
- Jain MK
- Obach M,
- Navarro-Sabaté A,
- Caro J,
- Kong X,
- Duran J,
- Gómez M,
- Perales JC,
- Ventura F,
- Rosa JL,
- Bartrons R
- Bhattacharya R,
- Senbanerjee S,
- Lin Z,
- Mir S,
- Hamik A,
- Wang P,
- Mukherjee P,
- Mukhopadhyay D,
- Jain MK
- Kluge MA,
- Fetterman JL,
- Vita JA
- Fledderus JO,
- Boon RA,
- Volger OL,
- Hurttila H,
- Ylä-Herttuala S,
- Pannekoek H,
- Levonen AL,
- Horrevoets AJ
Laminar shear stress regulates endothelial cell phenotype by inducing quiescence, but whether metabolic changes contribute to this process was not known. This study show that laminar shear stress affects endothelial cell glycolytic metabolism, which controls angiogenic behavior of endothelial cells. These findings provide mechanistic insight into how the switch between the resting and the angiogenic state of endothelial cells is regulated by physiological stimuli.