Suppression of DNA-PKcs enhances FGF-2 dependent human endothelial cell proliferation via negative regulation of Akt

Abstract

Angiogenesis initiation is crucially dependent on endothelial proliferation and can be stimulated by the fibroblast growth factor 2 (FGF-2). The DNA dependent protein kinase (DNA-PK), long known for its importance in repairing DNA double strand breaks, belongs to the phosphatidylinositol-3 kinase (PI3-K) super family and has recently been identified as one of the enzymes phosphorylating and activating Akt. Due to its similarity with PI3- K, we hypothesized that DNA-PK may have similar effects on endothelial angiogenic processes and signalling. We used primary endothelial cells (HUVEC and PAEC) and human microvascular endothelial cells (HMEC) to study the role of DNA-PK in endothelial proliferation and signalling. DNA-PKcs suppression with the compound NU7026 or with siRNA induced basal endothelial cell proliferation as well as enhanced FGF-2 dependent proliferation. This was associated with an increase in phosphorylated Akt. Tube formation was not affected by DNA-PKcs clearly showing that the role of DNA-PK in endothelial processes differs from that of PI3-K. Our findings indicate DNA-PK as an important enzyme maintaining the quiescent endothelial phenotype by actively inhibiting Akt thus restraining endothelial cell proliferation preventing excessive growth.

1. Introduction

In early steps, the process of angiogenesis crucially depends upon proliferation and migration of endothelial cells before they are organized into tube-like structures. The fibroblast growth factor 2 (FGF-2), also called basic fibroblast growth factor (bFGF), is a potent stimulus for endothelial cell proliferation, migration and tube formation [1,2]. FGF-2 activates the phosphatidylinositol-3 kinase (PI3-K) in endothelial cells [3,4], which regulates Akt or protein kinase B (PKB) activity [5], a most critical step for endothelial cell proliferation, migration and tube formation [2,6]. Akt is activated by phosphorylation on Thr308 and Ser473 [4]. While Thr308 is phosphorylated by the phosphatidylinositol dependent protein kinase 1 (PDK1), the DNA dependent protein kinase (DNA-PK) has recently been demonstrated to be responsible for the phosphorylation of Ser473 [7,8]. Interestingly, FGF-2 has previously been demonstrated to increase DNA-PK activity [9]. DNA-PK is a member of the PI3-K super family and consists of the regulatory subunits Ku70/Ku80 and the catalytic subunit DNA-PKcs. In addition to PI3-K, the catalytic subunit exerts serine/threonine kinase activity but unlike PI3-K does not possess lipid kinase activity. DNA-PKcs is activated by free DNA ends but also by phosphorylation by the subunits Ku or by autophosphorylation [10]. Several proteins have been identified that regulate DNA-PKcs activity, such as protein kinase Cδ [11] and the protein phosphatases 1, 2A [12] and 6 [13]. DNA-PK has long been known for its importance in repairing DNA double strand breaks (DSB’s) by nonhomologous end joining (NHEJ) generated during physiological conditions, such as during variable (diversity) joining, V (D)J, and class switch recombination required for antibody diversity but also DSB’s caused during meiosis and by external factors, such as ionizing irradiation or chemotherapeutic agents [14,15], making it a desirable target for anti-cancer treatment [16]. Indeed, DNA-PKcs−/− mice are hypersensitive to ionizing radiation and are also immunode- ficient [17]. DNA-PK seems to be responsible for the phosphorylation of the activating Ser473 on Akt in HEK293 cells when exposed to naked
DNA [7] and in mouse embryonic fibroblasts upon γ-irradiation [8]. These data are strengthened by a recent study showing that DNA-PK is necessary for the phosphorylation of Ser473 on Akt following DNA damage [18]. Thus Akt seems to be one of the major substrates for DNA- PK in the context of DNA damage repair. However, the effect of growth factors on DNA-PK function and signalling is not yet fully understood.

Apart from its previous described role in DNA DSB repair, DNA-PK has also been found to interact with or influence the activity of several nuclear proteins, some of which are important for regulation of cell cycle progression such as the transcription factor E2F [19], c-myc [20], p53 [21], NFkB [22] and FoxO4 [8]. Moreover, DNA-PK dependent Akt activation was found to protect from TNFα induced apoptosis [23], another cellular fate which is also strongly influenced by PI3-K [24]. In addition, newer findings have revealed that DNA-PKcs can exit the nucleus upon rising levels of cyclic AMP in HEK cells [25] and it was found in lipid rafts of the cellular membrane in the glioblastoma derived cell line MO75K [26]. These findings demonstrate that DNA-PK most likely has broader functions than the repair of DSB’s in the nucleus. In fact, these findings suggest that DNA-PK may have similar properties to its relative PI3-K, a possibility, which is often overseen when using pharmacological inhibitors to suppress PI3-K activity. In this study, we were interested in defining the role of DNA-PKcs in the induction of growth factor dependent signalling, proliferation and tube formation in endothelial cells. We also strived to differ between DNA-PK and PI3-K in Akt activation in these processes, thereby defining the role for endothelial DNA-PK in the initiating steps of endothelium dependent angiogenesis.

2. Materials and methods

2.1. Chemicals

Mouse monoclonal DNA-PKcs antibody and goat polyclonal Lamin A/C antibody were from Santa Cruz Biotechnology (Heidelberg, Germany). Rabbit phospho-Akt Ser473 antibody and VEGFR-2 antibody were from Cell Signalling (Frankfurt am Main, Germany). Mouse monoclonal GAPDH antibody was from Chemicon (Hampshire, England). Horseradish peroxidase-conjugated antibodies, pharmacological DNA-PKcs inhibitor (NU7026) and PI3-K inhibitor LY294002 were from Calbiochem (Darm- stadt, Germany). FGF-2 was purchased from Tebu-bio (Offenbach, Germany). All other chemicals were from Sigma Aldrich (Taufkirchen, Germany).

2.2. Cell lines and cell culture

Human umbilical vein endothelial cells (HUVEC) and porcine aortic endothelial cells (PAEC) were isolated and cultured as previously described [27,28]. Human dermal microvascular endothe- lial cells (HMEC) were provided by Ades et al. [29] and cultured in DMEM from Sigma Aldrich (Taufkirchen, Germany) supplemented with 10% FCS (Biochrom, Berlin, Germany), 10% endothelial cell growth media (PromoCell, Heidelberg, Germany) and 1% Penicillin– Streptomycin (Sigma Aldrich, Taufkirchen, Germany).

2.3. siRNA transfections

PAEC were transfected with 30 nM siRNA using the magnetofec- tion method previously described [30,31]. Validated siRNAs for human DNA-PKcs and negative control siRNA were purchased from Qiagen (Hilden, Germany).

2.4. Cell extracts for western blot analysis and cell fractioning

Protein lysates were prepared and protein content quantified as described elsewhere [32]. Lysates were subjected to western blot analysis as previously described [32]. Cell fractioning was performed with the Proteo Extract kit from Calbiochem (Darmstadt, Germany) according to the supplier’s protocol.

2.5. RT-PCR and RNA isolation

Total RNA was isolated from HMEC with the PeqGOLD total RNA kit from Peqlab (Erlangen, Germany). 2 µg RNA was subjected to RT- PCR using the M-MLV reverse transcriptase (Promega, Mannheim, Germany). For PCR 27 cycles were used, as small differences in RNA concentrations could be distinguished at this point. Reactions were performed according to the supplier’s protocol (Promega, Mannheim, Germany). The following primer sequences were used: DNA-PKcs Forward primer 5′GGCCATGACGAGAGGGAACAC 3′, DNA- PKcs Reverse primer 5′TGAGAGCTGGCGAAGTGGGAGC 3′, β-Actin Forward primer 5′TGTCCACCTTCCAGCAGATGTG 3′, and β-Actin Reverse primer 5′AGTCCTCGGCCACATTGTGAAC 3′ (all produced by Biomers.net, Ulm, Germany).

2.6. Immunofluorescence

HMEC or HUVEC cultivated in 8-well microscope slides from IBIDI (Munich, Germany) were treated and washed with PBS+ followed by fixation in 100% Methanol at −20 °C for a minimum of 30 min. Fixed cells were washed once with PBS+ prior to blocking in 5% BSA in PBS + for 1 h. After staining cells with first antibody 1:200 in blocking solution for 1 h, cells were washed with PBS+ and subsequently incubated with secondary antibody 1:400 in blocking solution for 10 min. Cells were washed and kept in PBS+ for microscopy. Images were taken with an Axiovert 200 M microscope with ApoTome (Zeiss, Jena, Germany).

2.7. MTT assay

Cell proliferation of HMEC was measured by the reduction of Methylthiazoletetrazolium, MTT, (0.5 mg/ml, 2 h) as described by Mosmann et al. [33] with minor modifications. As the inhibitor NU7026 was shown to have a significant effect in in vitro experiments first after 4–24 h of treatment [34], the data obtained after 48 h was set in relation to sham treatment for 24 h.

2.8. Capillary-like structure assay

Confluent HUVEC or HMEC were subjected to the capillary-like structure assay as previously described [28]. Pictures were taken with an Axiovert 135M microscope (Zeiss, Jena, Germany).

2.9. Statistical analysis

Data were analysed using the Student’s t-test, one-way ANOVA or a Wilcoxon’s Rank sum test as appropriate. All data are presented as means±SEM. Results were considered significant at an error proba- bility level of p < 0.05.

3. Results

3.1. Expression and localisation of DNA-PKcs in human endothelial cells

DNA-PKcs is expressed in both primary human umbilical vein endothelial cells (HUVEC) and in human microvascular endothelial cells (HMEC) as assessed by western blotting. As a negative control for DNA-PKcs expression protein lysates from platelets, that lack the nucleus, were subjected to western blotting. No DNA-PKcs expression could be detected in these lysates demonstrating its specific expression pattern (n =3, Fig. 1A). Furthermore, immuno- fluorescence staining and visualisation with a Zeiss Axiovert 200 microscope in ApoTome modus was performed to confirm endothelial expression of DNA-PKcs and to determine the sub- cellular localisation of DNA-PKcs. As seen in Fig. 1B, in both endothelial cell types, DNA-PKcs could exclusively be detected in the nucleus (Lamin A/C staining shown in red was used to detect the nuclear membrane) with weak peri-nuclear traces of DNA-PKcs (n =3).

3.2. DNA-PKcs and endothelial tube formation in vitro

We next investigated if DNA-PKcs also influences the formation of tube-like structures from HUVEC in vitro. Whereas treatment with the PI3-K inhibitor LY294002 (10 µM) significantly impaired this process (Fig. 2A+C, p < 0.01, n = 3, HUVEC) as expected, suppression of DNA- PKcs activity using the pharmacological inhibitor NU7026 (0.3 µM) had no significant effect on the formation of capillary-like structures (n = 5, HUVEC). This was specific for DNA-PKcs activity, since raising the concentration of the inhibitor 1000-fold, a concentration where also PI3-K is inhibited by NU7026, could reverse this effect (p < 0.001, n = 9, HUVEC). As we could observe a stronger expression of DNA- PKcs in HMEC we next performed the assay with HMEC to rule out that the lack of impairment of tube-like structures was not due to a too low expression of DNA-PKcs in HUVEC. As seen in Fig. 2B and C the formation of capillary-like structures was not affected by treatment with the DNA-PKcs inhibitor also in HMEC (n = 7). PI3-K inhibition by LY294002 or an increase in the NU7026 concentration significantly reduced the structures formed (p < 0.05, n = 6, HMEC and p < 0.001, n = 6, HMEC respectively). These effects were equally observed in HMECs and HUVECs treated with starvation media (1% FCS or 1% NBCS respectively) and with normal culture media (10% FCS or 20% NBCS containing a supplement of different growth factors respectively) showing no difference in activation status of DNA-PKcs. To increase the specificity of DNA-PKcs inhibition, porcine aortic endothelial cells (PAEC) were transfected with DNA-PKcs siRNA, as this endothelial cell type showed the greatest efficiency in DNA-PKcs knock-down (Fig. 2D, p < 0.05, n = 4). DNA-PKcs siRNA treatment for 48 h did not inhibit the formation of capillary-like structures of cells grown in starvation media (n = 8) or in normal culture media (n = 6) in comparison to control siRNA (Fig. 2E+F, p = 0.062, n = 6, PAEC). These results indicate that DNA-PKcs does not seem to mimic the function of its relative PI3-K in this context.

3.3. DNA-PKcs knock-down enhances basal endothelial proliferation rates

In the next step, we examined the role of DNA-PK in human endothelial cell proliferation by assessing the reduction of MTT 48 h after treatment of HMEC. As the inhibitor NU7026 was shown to have a significant effect in in vitro experiments first after 4–24 h of treatment [34], the data obtained after 48 h were set in relation to sham treatment for 24 h. Treatment of HMEC with culture medium containing 1% FCS did not induce proliferation over time (up to 48 h) as expected. However, addition of NU7026 (0.3 µM) significantly induced the endothelial proliferation by 1.6 fold (±0.2, p < 0.01, n = 27, Fig. 3A) after 48 h of treatment in comparison to sham treatment. To confirm the specificity of this effect, PAEC were transfected with DNA-PKcs siRNA (30 nM) and proliferation mea- sured after 48 h, as we could observe the greatest efficiency in mRNA silencing at this time point. DNA-PKcs siRNA also significantly induced endothelial proliferation by 2.8-fold (±0.5, p < 0.05, n = 12, Fig. 3B) as compared to cells transfected with control siRNA. These findings indicate that DNA-PKcs normally would have a restraining effect on endothelial proliferation.

3.4. DNA-PKcs inhibition enhances growth factor dependent endothelial proliferation

Having observed an increased basal proliferation upon DNA- PKcs inhibition, we next investigated if a suppression of DNA-PKcs could further enhance growth factor dependent proliferation. Application of NU7026 (0.3 µM) increased cellular proliferation induced by HMEC culture medium (HM) containing 10% FCS and 10% endothelial cell growth medium (PromoCell) supplemented with growth factors by 2 ± 0.13 fold (p < 0.001, n = 24, Fig. 4A) in comparison to treatment with only HM. Moreover, this increase could be prohibited by treatment with an anti-FGF-2 antibody (5 µg/ml) (p < 0.01, n = 12, Fig. 4A), whereas treatment with a control antibody showed no impact on proliferation induced by HM, indicating FGF-2 as one important pro-proliferative factor in the HM. Therefore proliferation was next investigated with cells stimulated with FGF-2 (10 ng/ml). FGF-2 application significantly increased endothelial cell proliferation (p <0.05, n = 32) and inhibition of DNA-PKcs activity also further enhanced the FGF-2 dependent proliferation by 1.6 ± 0.17 fold (p < 0.001, n = 27, Fig. 4A). We further investigated if the effect on FGF-2 dependent proliferation upon suppression of DNA-PKcs activity is affected by Akt activity. The enhancement in FGF-2 dependent proliferation after NU7026 application was completely abrogated when adding an Akt inhibitor (10 µM) in comparison to sham treatment (p < 0.001, n = 18). Interestingly, when treating cells with the endothelial survival factor VEGF (10 ng/ml) an inhibition of DNA- PKcs did not affect the proliferation in any way (n = 18). Finally, suppression of DNA-PKcs expression with siRNA (30 nM) further enhanced the FGF-2 dependent endothelial proliferation by 2.9 ± 0.52 fold compared to treatment with a control siRNA (Fig. 4B, p < 0.001, n = 12, 48 h post transfection, PAEC).

3.5. Basal Akt Ser473 phosphorylation is enhanced by DNA-PKcs inhibition in endothelial cells

As we could observe a significant role of DNA-PKcs in both basal and growth factor dependent endothelial cell proliferation, which could be prohibited by an Akt inhibition, we next assessed the Akt activity by detecting its phosphorylation on Ser473. Treatment with the DNA-PK inhibitor NU7026 (0.5 µM, 24 h) or DNA-PKcs siRNA (30 nM) resulted in a significant increase in Akt phosphorylation and thus activation in comparison to sham or control siRNA treatment respectively as detected with western blot using an anti-phospho Akt (Ser473) antibody (p < 0.05, n =6–9, Fig. 5). When stimulating cells with FGF-2 (10 ng/ml, 10 min) no further increase in Akt phosphorylation could be observed upon NU7026 (n =3, HMEC, Fig. 5) or DNA-PKcs siRNA (n =4, PAEC, Fig. 5) treatment in comparison to FGF-2 stimulation alone. This indicates an already maximal activation of Akt.

3.6. FGF-2 enhances endothelial DNA-PKcs protein expression but does not change cellular localisation

To elucidate if FGF-2 itself may regulate DNA-PKcs either by changing its cellular localisation from nucleus to the membrane or by quenching its inhibitory effects to induce proliferation, we investigated whether FGF-2 affects DNA-PKcs cellular localisation and expression in endothe- lial cells. HMEC were treated with FGF-2 (10 ng/ml) and the protein expression and localisation of DNA-PKcs were evaluated with western blot. As seen in Fig. 6A, the strongest DNA-PKcs expression could be found in the nuclear fraction and traces of DNA-PKcs in the membrane fraction. The distribution of DNA-PKcs in the different fractions was not changed after FGF-2 stimulation (n =3). To confirm the localisation of DNA-PKcs immunofluorescence staining and visualisation with a Zeiss Axiovert 200 microscope in ApoTome modus was performed. Whereas the nuclei showed a strong DNA-PKcs staining (Fig. 6B, n =3, HMEC, red staining), only extremely faint Immunofluorescence could be detected in the peri-nuclear space, which did not differ between nonstimulated and FGF-2 stimulated cells. Surprisingly, when investigating the protein expression a raise in DNA-PKcs protein level could be seen already 24 h after treatment, which reached significance 48 h after treatment (p <0.05, n =7, Fig. 6B). Next, total RNA was isolated from HMEC and RT-PCR with 27 PCR cycles was performed to reveal changes in DNA- PKcs mRNA levels after FGF-2 application. A similar pattern was observed with increasing levels of DNA-PKcs mRNA after 24 h and 48 h FGF-2 treatment (n =4, Fig. 6C).

4. Discussion

In this study, we for the first time demonstrate the role of DNA-PK in endothelial dependent angiogenic events and find DNA-PK to be a negative regulator of Akt and the MAPK p42/44 preventing human endothelial cell proliferation. This process is known to activate, among many, the PI3-K, which has been shown to play a key role in endothelial proliferation. One of its relatives is the DNA dependent protein kinase (DNA-PK), which has been demonstrated to regulate the activity of nuclear proteins involved in cell cycle progression [19–22]. In fact, DNA-PK is a member of the PI3-K super family and shares some features with PI3- K [10]. Moreover, DNA-PK has been demonstrated to phosphorylate and activate Akt [7], which is a downstream target of PI3-K suggesting DNA-PK to work in parallel with or maybe even downstream of PI3-K. Furthermore, many pharmacological PI3-K inhibitors also quench the activity of DNA-PK. This made us question if DNA-PK also would be of similar importance as PI3-K to different steps in the angiogenesis process, such as endothelial cell proliferation and tube formation.

So far most studies regarding the function of DNA-PK have been performed in cell lines derived from different tumours, which either
“naturally” harbours DNA-PK mutations (i.e. MO59J [35,36]) or where DNA-PK has been knocked-out by gene targeting or pharmaceutically inhibited, or cells isolated from mice suffering from severe combined immunodeficiency (SCID), which bear a mutation in the DNA-PKcs gene [37]. Some of these cells also have defects in the function of other proteins [38–40] sometimes making the influence of DNA-PKcs somewhat difficult to interpret. In contrast to such studies, we used primary endothelial cells as well as the endothelial cell line HMEC-1 to investigate the importance of DNA-PKcs in human endothelial cell dependent processes important for angiogenesis initiation. In order to achieve our aim we used the novel DNA-PK inhibitor NU7026 [41] with much higher specificity for DNA-PK (EC50: 0.23 µM) than PI3-K (EC50: 13 µM) as well as a siRNA directed against DNA-PKcs mRNA.

In our experiments, treatment with this specific DNA-PK inhibitor (NU7026) did not show any influence on capillary-like structures in vitro, whereas the PI3-K inhibitor LY294002 as well as a 1000-fold increase in the NU7026 concentration, which has been shown to also inhibit PI3-K [41], significantly reduces the amount of structures formed. The amount of DNA-PKcs expression did not seem to have any impact on this process, since HMEC which showed a higher expression of DNA-PKcs did not differ in their response compared to HUVEC with much lower expression levels. This shows that DNA-PKcs is not comparable with PI3-K in this context and does not seem to be involved in endothelial migration and organization into tubes. On the contrary, a suppression of DNA-PKcs enhanced both the basal endothelial proliferation as well as proliferation induced by growth factor rich medium and by FGF-2 alone. Surprisingly, we also found that DNA-PKcs protein expression was enhanced upon FGF-2 treatment. This may suggest that FGF-2 uses DNA-PKcs as a negative feedback to prevent excessive growth and thus tumour formation. The effect observed on endothelial proliferation seems to be specific for FGF-2, as the VEGF dependent proliferation was unaffected. Further, we observed an upregulation of the pro-proliferative kinase Akt activity in endothe- lial cells lacking DNA-PKcs activity although no stimulus was present, indicating an active negative regulation of this kinase by DNA-PKcs under basal conditions. This would fit well with the fact that the endothelium is quiescent with approximately 0.1% replications/day in vivo under normal conditions [42]. These findings demonstrate that DNA-PKcs may be important for maintaining the quiescent phenotype. Moreover, no further increase in Akt phosphorylation could be detected in cells lacking DNA-PKcs activity but treated with FGF-2 compared to control cells treated with FGF-2. These results are in concordance with the data obtained in the proliferation assay, where treatment with FGF-2 and NU7026 or DNA-PKcs siRNA resulted in an increase in proliferating cells in comparison to FGF-2 treatment alone but did not show any further increase compared to cells treated with only NU7026 or DNA- PKcs siRNA. This observation may be explained by an already maximal activation of Akt by FGF-2. Also, this may indicate that DNA-PKcs regulates Akt during basal conditions while when stimulated by FGF-2 other signalling molecules, important for cell cycle regulation, may be controlled by DNA-PKcs in addition to Akt. The ability of the Akt inhibitor to completely abrogate the enhanced endothelial proliferation caused by DNA-PKcs inhibition indicates that Akt activation is responsible for at least part of the response. Taken together, these results are in concordance with another study by Watanabe et al., where cells derived from SCID mice known to lack DNA-PKcs protein, showed a higher proliferative rate in comparison to wild type cells [19]. However, other studies exist reporting the opposite, where human adenocarcinoma DNA-PK knock-out cells are proliferating at much slower rate than wild type cells [43]. Our findings, as well as these controversial reports may be explained by an interesting recent finding by Surucu et al. [8], which highlights the tissue specific involvement of DNA-PK in Akt signalling: in their study no change in phospho-Akt levels could be observed in tissue samples from skeletal muscle, liver, spleen or brain in DNA-PKcs knock-out (−/−) mice, whereas adipose, brown fat and thymus showed increased phospho-Akt suggesting tissue specific DNA- PK dependent Akt activity. Moreover, DNA-PKcs−/− mice developed thymic lymphomas due to a high proliferative rate of thymocytes. DNA- PK has previously been demonstrated to be responsible for Akt phosphorylation upon irradiation induced DNA damage or in the presence of CpG-DNA and even insulin [7,8,44] but not upon IGF stimulation [8]. These data suggest that the DNA-PK dependent Akt regulation is strongly dependent on the stimulus.

Of note, our data are the first to determine the extent of DNA- PKcs expression in human primary endothelial cells (HUVEC) and the human endothelial cell line HMEC-1. Although its main functions seem to be in the nucleus, Lucero et al. reported the presence of DNA-PKcs in lipid rafts of the cellular membrane [26]. In contrast to these findings, we detected strong DNA-PKcs expression in the cell nucleus of endothelial cells with only peri-nuclear traces of DNA- PKcs. When comparing these results with the data obtained from Huston et al. [25], where a strong overall cytoplasmic staining of DNA-PK could be detected upon forskolin treatment, the faint peri- nuclear staining of DNA-PKcs observed in our experiments is so weak that we believe it can be ignored. Nevertheless, this localisation was not changed when stimulating with FGF-2 indicat- ing that the localisation and also possibly function of DNA-PKcs may vary between cell types. Taken together, ours and others’ results indicate that the way of influence of DNA-PKcs on cellular proliferation and signalling is highly cell- and tissue specific. This observation together with our findings suggests that DNA-PKcs may be an attractive target for proangiogenic therapy to promote vessel outgrowth and thus oxygen- and nutritional supply to ischemic areas without resulting in induction of cell proliferation in other tissues promoting tumour growth.

We therefore conclude that in this study, we were able to define the role of DNA-PKcs in endothelial cellular processes important for neo-angiogenesis. We find DNA-PKcs to exert completely differ- ent effects than its relative PI3-K in endothelial cells. DNA-PKcs is demonstrated to have a restraining effect on endothelial cell pro- liferation through negative regulation of Akt, which could potentially make it a promising therapeutic target for proangiogenic therapy at ischemic regions.