Dimethylarginine dimethylaminohydrolase 1 modulates endothelial cell growth through NO and Akt

Abstract

Objective

Dimethylarginine dimethylaminohydrolase 1 (DDAH1) modulates NO production by degrading the endogenous NO synthase (NOS) inhibitors ADMA and L-NMMA. This study examined whether, in addition to degrading ADMA, DDAH1 exerts ADMA independent effects that influence endothelial function.

Methods and Results

Using selective gene silencing of DDAH1 with small interfering RNA and overexpression of DDAH1 in HUVEC, we found that DDAH1 acts to promote endothelial cell proliferation, migration and tube formation both by Akt phosphorylation as well as through the traditional role of degrading ADMA. Incubation of HUVEC with the NOS inhibitors L-NAME or ADMA, the soluble guanylyl cyclase inhibitor ODQ, or the cGMP analog 8-pCPT-cGMP had no effect on p-Akt^Ser473, indicating that the increase of p-Akt^Ser473 produced by DDAH1 was independent of the NO-cGMP signaling pathway. DDAH1 formed a protein complex with Ras, and DDAH1 overexpression increased Ras activity. The Ras inhibitor manumycin-A or dominant-negative Ras significantly attenuated the DDAH1-induced increase of p-Akt^Ser473. Furthermore, DDAH1 knockout impaired endothelial sprouting from cultured aortic rings, and overexpression of constitutively active Akt or DDAH1 rescued endothelial sprouting in the aortic rings from these mice.

Conclusions

DDAH1 exerts a unique role in activating Akt that affects endothelial function independent of degrading endogenous NOS inhibitors.

Introduction

Endogenous asymmetric dimethylarginine (ADMA) competes with arginine at nitric oxide (NO) synthase (NOS) to inhibit NO production (1). Recent reports have demonstrated that accumulation of ADMA is a major risk factor for cardiovascular diseases including hypertension (2, 3), atherosclerosis (4), congestive heart failure (5) and stroke (6). ADMA is cleared principally through metabolism by dimethylarginine dimethylaminohydrolase (DDAH) to L-citrulline and dimethylamine, with a small fraction eliminated by renal excretion (7). DDAH activity was reduced and negatively correlated with ADMA levels in rats with diabetes or with chronic hypoxia-induced pulmonary hypertension, implying that the increased ADMA levels in these conditions might be the result of impaired degradation (8,9).

Two DDAH isoforms have been reported (DDAH1 and DDAH2) that are encoded by different genes located on different chromosomes (1,10,11). In coronary vessels we have found that DDAH1 is predominantly expressed in endothelial cells where eNOS is also highly expressed (12). By using the Cre-LoxP approach, we have generated a mouse strain with DDAH1 gene deletion in vascular endothelial cells (endo-DDAH1 KO). DDAH1 expression was greatly reduced in kidney, lung, brain and liver of the endo-DDDAH1 KO mice, while tissue and plasma ADMA were increased in these mice (13). These data indicate that in these organs DDAH1 is predominantly distributed in endothelial cells (13), and that endothelial DDAH1 plays an important role in degrading ADMA.

Akt is a serine-threonine protein kinase implicated in diverse cell functions including migration, proliferation, survival and metabolism. Three different Akt isoforms (Akt1, 2 and 3) encoded by distinct genes have been reported in mammalian cells (14). Both Akt1 and Akt2 are highly expressed in vascular tissues, with Akt1 accounting for ~75% of total endothelial cell Akt activity (15). Akt1 is important for maintaining normal vascular function. Thus, Sessa and associates found that Akt1 KO abolished hind limb angiogenesis in mice subjected to femoral artery ligation (16), and that the Akt1 KO mice had reduced endothelial progenitor cell (EPC) mobilization in response to ischemia. Furthermore, introduction of wild type EPCs, but not EPCs from Akt1^–/– mice, improved limb blood flow after femoral artery ligation. Endothelial cells from Akt1^–/– mice have diminished eNOS phosphorylation and NO production, in agreement with the concept that Akt phosphorylation of serine 1177 activates eNOS (15, 16).

Here by using selective DDAH1 siRNA and DDAH1 overexpression in primary human umbilical artery endothelial cells (HUVEC), we demonstrate that DDAH1 regulates endothelial cell proliferation, migration and tube formation. These responses were associated with DDAH1 dependent changes in ADMA and NO production, consistent with the conventional NO-cGMP signaling pathway. In addition, we also found that DDAH1 regulates endothelial p-Akt^Ser473 content and Akt activity independently of the NO-cGMP pathway, and that the effect of DDAH1 in promoting tube formation and cell proliferation is Akt dependent. Furthermore, we demonstrated that DDAH1 regulates p-Akt^Ser473 content by causing an increase of Ras activity.

Methods

For a detailed description of methods, please see the supplemental materials (available online at http://atvb.ahajournals.org).

Data analysis

All data are presented as mean ± standard error. Comparison between two groups was performed with the unpaired t-test (2-tailed). For comparisons between more than two groups, one-way analysis of variance was used followed by Fisher’s LSD method. Statistical significance was defined as p< 0.05.

Results

Selective knockdown of DDAH1 with siRNA increases ADMA and decreases NO production

As compared to non transfected cells and control siRNA transfected cells, transfection of DDAH1 siRNA successfully knocked down DDAH1 protein and mRNA at 24, 48 and 72 hours by over 80% (Figure 1A-C). DDAH1 siRNA had no effect on expression of DDAH2 or eNOS as demonstrated by Western blot (Figure 1A, Supplemental Figure I). The decrease in DDAH1 protein after DDAH1 siRNA was associated with a significant increase of ADMA, and decreases of nitrite (end-product of NO) and cGMP produced into the culture medium over the 24 hour period from 48 to 72 hours after transfection (Figure 1D, 1E, 1F).

Figure 1. Selective DDAH1 siRNA in HUVEC resulted in decreased DDAH1 expression, increased ADMA and decreased NOx production.

DDAH1 siRNA in HUVEC resulted in decreases of DDAH1 protein (A,B) and mRNA expression (C), increased ADMA in the medium (D), decreased NOx production (E) and decreased cGMP content (F). DDAH1 siRNA and non-specific control siRNA were transiently transfected into HUVEC and cells were collected for measuring the indicated proteins or mRNA at the indicated times. The results are from four independent experiments. For measurement of ADMA, NO and cGMP production, 24 hours after transfection with siRNA HUVEC were cultured in fresh medium for an additional 24 hours and the medium was collected for assay of ADMA and NO content. *p<0.05 as compared with control.

DDAH1 siRNA decreased endothelial tube formation and cell proliferation

As endothelial cell proliferation and tube formation are essential steps in angiogenesis, we determined the effect of DDAH1 siRNA on HUVEC tube formation and proliferation. Growth of HUVEC on Matrigel resulted in cell migration and elongation to form tube-like structures. In comparison with cells transfected with control siRNA, tube formation was decreased in cells transfected with DDAH1 siRNA (Figure 2A, 2B) (n=8; p<0.05). There was no difference in tube formation between nontransfected cells and cells transfected with control LacZ siRNA (data not shown), indicating that the decreased tube formation in cells transfected with DDAH1 siRNA was not a nonspecific effect of the siRNA procedure.

Figure 2. DDAH1 siRNA in HUVEC attenuated tube formation, proliferation, p-Akt^Ser473 and p-eNOS^Ser1177.

DDAH1 siRNA attenuated HUVEC tube formation on Matrigel (A,B) and cell proliferation (C). 24 hours post transfection with siRNA, HUVEC cell number was counted and layered on Matrigel; tube structures were observed 18 hours later. Tube length was decreased in cells transfected with DDAH1 siRNA compared to control siRNA (n=4/group). D: ADMA (5-100 μM) inhibits HUVEC proliferation in a concentration dependent manner. E: The nonselective NOS inhibitor L-NAME (5mM) and the soluble guanylyl cyclase inhibitor ODQ (1μM) attenuated HUVEC proliferation. F: The permeable cGMP analogue PCMPT (1-50μM) dose dependently rescued the decrease of HUVEC proliferation caused by DDAH1 siRNA. G and H: DDAH1 siRNA decreased DDAH1, p-Akt^Ser473 and p-eNOS^Ser1177 but had no effect on total-Akt or total-eNOS. I: The decrease of HUVEC proliferation caused by DDAH1 siRNA was totally rescued by overexpression of constitutively active Akt (MOI 50)(C). *p<0.05 as compared with control conditions.

In addition, DDAH1 siRNA significantly (p<0.05) attenuated HUVEC proliferation (Figure 2C). Addition of the endogenous NOS inhibitor ADMA in the culture medium caused a dose-dependent decrease of HUVEC proliferation (Figure 2D). Both the non-selective NOS inhibitor L-NAME and guanylate cyclase inhibitor 1H-[1,2,4]oxadiazolo-[4,3–2]quinoxalin-1-one(ODQ) (1μM) decreased HUVEC proliferation (Figure 2E). Addition of the membrane permeable cGMP analog 8-pCPT-cGMP dose-dependently rescued the DDAH1 siRNA induced decrease of HUVEC proliferation (Figure 2F). These data suggest that inhibition of the NO-cGMP pathway contributed to the decreased HUVEC proliferation after DDAH1 knockdown. Addition of L-arginine (1 to 5mM) to the culture medium caused a dose-dependent increase of HUVEC proliferation in both control cells and cells with DDAH1 knockdown (supplemental Figure II), but was not able to fully restore proliferation in the DDAH1 knockdown cells.

DDAH1 siRNA decreased HUVEC Akt ^Ser473 and p-eNOS^Ser1177

As Akt plays an important role in regulating endothelial function in part by modulating eNOS phosphorylation and NO production, and since phosphorylation of Akt at the Ser473 site is associated with its activation and physiological function, we determined p-Akt ^Ser473 after DDAH1 siRNA. DDAH1 siRNA caused a significant decrease of p-Akt ^Ser473 content but had no effect on total-Akt (Figure 2G, 2H). Since Akt1 can activate eNOS by direct phosphorylation at Ser1177 (17), total and p-eNOS^Ser1177 protein content were determined. DDAH1 siRNA significantly decreased p-eNOS^Ser1177 (Figure 2G, 2H).

Constitutively active Akt (caAkt) rescued HUVEC proliferation after DDAH1 knockdown

To determine whether the decreased HUVEC proliferation after DDAH1 knockdown was the result of the decreased p-Akt ^Ser473, HEVEC were infected with adenoviral vector expressing caAkt (MOI 50). Overexpression of caAkt completely prevented the decreased HUVEC proliferation caused by DDAH1 knockdown (Figure 2I). Overexpression of caAkt also significantly increased NO production, with a further increase with the addition of L-arginine (1mM) in both control and DDAH1 knockdown cells (Supplemental Figure III). These data indicate that Akt can regulate NO production in endothelial cells, and that the decrease of p-Akt ^Ser473 after DDAH1 siRNA contributed to the decreased cell proliferation.

DDAH1 overexpression increased HUVEC p-Akt ^Ser473 independent of ADMA, NO and cGMP

By infecting HUVEC with differing amounts of Ad-DDAH1, we found that DDAH1 overexpression caused a dose-dependent increase of p-Akt ^Ser473 (MOI:10, 50, 100), with no effect on total-Akt (Figure 3A-C). DDAH1 overexpression in HUVEC caused a significant decrease of ADMA and increase of cGMP (Figure 3B). Importantly, however, the increase of p-Akt ^Ser473 in response to DDAH1 overexpression was not prevented by treatment with L-NAME, ADMA, ODQ or PCMPT (Figure 3D), indicating that the DDAH1 overexpression-induced increase of p-Akt ^Ser473 was not the result of the decrease of ADMA, or an increase of NO or cGMP. To further demonstrate that the phosphorylation of Akt was not dependent on DDAH enzyme activity, we created a DDAH1 Cys273 to Ser site mutation. This DDAH1^C273S mutant lacks DDAH activity, but was still able to activate Akt (Figure 3E, 3F), indicating that DDAH activity is not required for the activation of Akt.

Figure 3. DDAH1 overexpression increased p-Akt^Ser473 and p-eNOS^Ser1177 independent of its effect on ADMA, NO and cGMP and DDAH1 mutant lack DDAH activity activated Akt.

DDAH1 overexpression caused increases of p-Akt^Ser473 and p-eNOS^Ser1177, but had no effect on total Akt or total eNOS (A,B). DDAH1 overexpression decreased ADMA and cGMP in the medium (B). The increase of p-Akt^Ser473 in response to DDAH1 overexpression was dose dependent. MOI: 10, 50, 100 (C), but the DDAH1-induced increase of p-Akt^Ser473 was not affected by the addition of ADMA (100 μM), L-NAME (5 mM) or the soluble guanylyl cyclase inhibitor ODQ (1μM). DDAH activity assay in HL-1 cells showed DDAH1 C273S mutant lack DDAH activity (3E). DDAH1 mutant increased p-Akt^Ser473 in HUVEC. MOI: 10, 40 (3F).

Inhibition of PI3K abolished the DDAH1-induced increases of tube formation, proliferation and migration

To determine whether the increase of p-Akt ^Ser473 after DDAH1 overexpression contributed to the increased HUVEC tube formation, cell proliferation and migration, we treated HUVEC with the PI3 kinase inhibitor Ly294002 (10μM). Ly294002 effectively attenuated the increases of p-Akt ^Ser473 and p-eNOS^Ser1177 in HUVEC overexpressing DDAH1, as well as basal p-Akt ^Ser473 and p-eNOS^Ser1177 in control cells (Figure 4A). Furthermore, Ly294002 blunted the DDAH1-induced increases in nitrite production, tube formation, proliferation and migration (Figure 4). In control HUVEC, Ly294002 also caused significant decreases of NOx production, tube formation, proliferation and migration, indicating that basal Akt activity is important to these cell functions. The decreases of tube formation, proliferation and migration after Ly294002 indicate that Akt exerts effects beyond the increase of NO production in these cells.

Figure 4. The PI3 kinase inhibitor Ly294002 (10μM) blocked the DDAH1 overexpression-induced increases of p-Akt^Ser473, p-eNOS^Ser1177(A), NO production (B), tube formation (C), proliferation (D) and migration (E) in HUVEC.

*p<0.05 as compared with control conditions; #p<0.05 as compared with cells transfected with adenovirus expressing DDAH1. Data were obtained from at least 4 samples for each group.

The increase of p-Akt ^Ser473 after DDAH1 overexpression is not the result of altered PTEN or protein phosphatase 2A (PP2A) expression

As the PIP3 phosphatase PTEN inhibits Akt activation (18), we determined the expression of PTEN and p-PTEN (the unstable form of PTEN) in HUVEC after DDAH1 overexpression. DDAH1 overexpression had no effect on the expression of either PTEN or p-PTEN (Supplemental Figure IV).

As Akt activity can be regulated by its binding to PP2A (19), we studied the effect of DDAH1 on PP2A expression and its interaction with Akt in HUVEC. We found that DDAH1 had no effect on PP2A expression. Using immunoprecipitation of Akt followed by immunoblot of PP2A, we found no detectable binding of PP2A to Akt. DDAH1 overexpression did not affect the binding between Akt and PP2A (Supplemental Figure IV). This indicates that the increased p-Akt ^Ser473 in response to DDAH1 overexpression was unlikely to be the result of PP2A dephosphorylation.

DDAH1 overexpression increased Ras activity

As previous studies have demonstrated that Ras contributes to activation of the PI3K-Akt pathway (20), we determined the effect of DDAH1 overexpression on Ras activation in HUVEC. Ras activity was significantly increased after DDAH1 overexpression (Figure 5A), while inhibition of Ras by either overexpression of dominant negative Ras (Ad-Ras-DN17, MOI50) or the Ras inhibitor manumycin A (3μM) (Figure 5B) significantly attenuated basal Akt activity and DDAH1-induced Akt activation. DDAH1 was also detected in the RBD/Ras-GTP complex (the pull-down bead complex) from cells with DDAH1 overexpression, while neither GFP nor α-tubulin were detectable in the pull-down bead complex (Figure 5C). These findings indicate that DDAH1 was specifically associated with Ras. Furthermore, when we measured Ras activity in lungs from DDAH1 KO mice, we found decreased Ras activity compared to wild type mice (Supplemental Figure V). We also found that Ras associated with mutated DDAH1 (Supplemental Figure VI) or DDAH1 in lungs from wild type mice (Supplemental Figure VII).

Figure 5. DDAH1 overexpression increased HUVEC Ras activity.

Ras activity (GTP-Ras level) was increased in Ad-DDAH1 infected HUVEC (A). Infection with adenoviral dominant negative Ras (N17) (dnRas, MOI 50) or treatment with the Ras inhibitor manumycin A (3μM) decreased the p-Akt^Ser473 level (B). The pull down studies were performed using agarose beads conjugated with Raf peptide containing Ras binding domain (RBD). DDAH1 was associated with Ras. GFP and α-tubulin were not detected with the pull down bead indicating the specificity of the Ras/DDAH1 binding (C). The results shown are representative of 3 or more independent experiments.

DDAH1 effects on ERK activity

As Raf, MEK (mitogen-activated protein kinase, MAPK) and ERK (extracellular signal-regulated kinase) are important downstream targets of Ras, we determined the effects of DDAH1 on ERK activity. DDAH1 knockdown increased p-ERK levels (Figure 6A, 6B), while overexpressing of DDAH1 decreased p-ERK (Figure 6C, 6D). These changes of p-ERK were opposite to the changes of p-Akt. Previous studies have shown that Akt can regulate ERK activity. Thus, Zimmermann et al (17) reported that in the MCF-7 breast cancer cell line, Akt inhibited activation of the Raf-MEK-ERK signaling pathway and shifted the cellular response from cell cycle arrest to proliferation. Similarly, Rommel (37) reported that Akt activation inhibited the Raf-MEK-ERK pathway in differentiated myotubes. Consequently, we tested whether constitutively active Akt (ca-Akt) or the PI3K inhibitor ly294002 could regulate ERK activation. As shown in Supplemental Figure VIII, ca-Akt caused inhibition of ERK activity, while inhibition of Akt activation with ly294002 increased ERK activity. These data demonstrate that Akt activation inhibits ERK activation in HUVEC. In addition, Akt has been reported to antagonize Raf activity by direct phosphorylation of Ser259 (17). This modification creates a binding site for 14-3-3 protein, a negative regulator of Raf. Therefore, we examined p-Raf ^Ser259 in HUVEC. We found that DDAH1 overexpression increased p-Raf ^Ser259 (Figure 6C, 6D). This finding suggests that DDAH1 overexpression inhibited ERK activation by activating Akt which subsequently phosphorylated Raf at ser 259, thereby inhibiting the Raf-MEK-ERK pathway (17,21).

Figure 6. DDAH1 effects on P-ERK1/2 level in HUVEC.

DDAH1 knockdown increased P-ERK1/2 level (6A, 6B). DDAH1 overexpression decreased p-ERK1/2 level and increased p-Raf S259 in HUVEC (6C, 6D). (n=4, P< 0.05)

Overexpression of DDAH1 or constitutively active Akt rescued endothelial sprouting in aortic rings from global DDAH1 KO mice

We recently generated viable global DDAH1 gene deficient mice (global-DDAH1 KO mice). Comprehensive characterization of this strain has been undertaken and the results will be reported in detail in a separate manuscript. Global-DDAH1 KO significantly reduced aortic p-Akt ^Ser473 and p-eNOS^Ser1177 expression (Figure 7A), indicating that DDAH1 plays a role in regulating p-Akt^Ser473 and p-eNOS^Ser1177 in vivo. Furthermore, global-DDAH1 KO decreased endothelial sprouting in cultured aortic rings. Importantly, overexpression of either DDAH1 or constitutively active Akt rescued endothelial sprouting in aortic rings from global DDAH1 KO mice (Figure 7B). To examine NO and Akt effects on endothelial sprouting capacity, the PI3K inhibitor Ly294002 or the NOS inhibitor L-NAME were added to the aortic rings from wild type mice. As shown in Figure 7C, L-NAME (4mM) inhibited microvessel outgrowth by ~ 30% (p<0.05), thus reconfirming a contribution of NO to angiogenesis (22). However, inhibition of p-Akt ^Ser473 with Ly294002 (10μM) caused a more dramatic inhibition (70 %; p<0.05) of microvessel outgrowth, indicating that p-Akt ^Ser473 also exerts an effect on endothelial sprouting in aortic rings which is independent of NO.

Figure 7. Overexpression of DDAH1 or constitutively active Akt (caAkt) rescued microvessel sprouting in aortic rings from DDAH1 KO mice.

Western blot shows that p-Akt^Ser473 and p-eNOS^Ser1177 were decreased in aortas from global DDAH1 KO mice, while total eNOS and Akt were unchanged. *p<0.05, n=5 (A). Endothelial sprouting was decreased in cultured aortic rings from DDAH1 KO mice, and overexpression of either caAkt or DDAH1 restored the sprouting capacity in aortic rings from the KO mice (B). Both Ly294002 (10μM) and LNAME (4mM) inhibited microvessel sprouting in cultured aortic rings from wild type mice (C). A diagram illustrating pathways by which DDAH1 regulates endothelial function (D). *p<0.05 as compared with control. Data are expressed as mean ± SE from 29-30 aortic rings.

Discussion

Here we demonstrate that DDAH1 caused two separate effects that resulted in increased NO production, HUVEC tube formation, proliferation and migration: an increase of p-Akt^Ser473 and a decrease of ADMA. Selective gene silencing of DDAH1 with siRNA decreased endothelial p-Akt ^Ser473, increased ADMA, decreased NO production, and decreased HUVEC proliferation and tube formation, while overexpression of constitutively active Akt rescued HUVEC from these changes induced by DDAH1 knockdown. Global DDAH1 KO decreased aortic p-Akt^Ser47 content and endothelial sprouting from aortic rings, while overexpression of either constitutively active Akt or DDAH1 rescued endothelial sprouting. Taken together, the data demonstrate that DDAH1 enhances endothelial cell function by regulating both the endogenous NOS inhibitors and Akt activity as summarized in Figure 7D.

The importance of DDAH in ADMA degradation and regulation of NO production has long been recognized (1,10), but the cellular distribution of DDAH1 and its physiological significance have been controversial. An early study reported DDAH1 mRNA distribution similar to that of nNOS, with DDAH2 mRNA distribution similar to eNOS (23). However, we recently found DDAH1 highly expressed in HUVEC and in coronary vascular endothelial cells where eNOS is also highly expressed (12). Wang et al reported that DDAH2 mRNA was 5-fold greater than DDAH1 mRNA in rat mesenteric resistance vessels but, using selective gene silencing approaches, only selective DDAH1 siRNA (not DDAH2 siRNA) caused an increase of plasma ADMA (24). In contrast, they found that DDAH2, but not DDAH1, siRNA abolished acetylcholine-induced vasodilatation (24). The investigators concluded that plasma ADMA is regulated predominantly by DDAH1, whereas EDRF/NO in mesenteric microvessels is regulated predominantly by DDAH2. The latter finding is at odds with a previous report that DDAH1^+/- mice develop pulmonary hypertension and demonstrate endothelial dysfunction (25), and our finding that NO mediated endothelial responses in isolated aortic rings was impaired in endo-DDAH1 KO mice, even though DDAH2 expression was unchanged (13). The differing results may be related to differences in the vascular bed studied or the experimental preparation used.

In the present study, DDAH1 knockdown increased ADMA levels, depressed NO production, decreased cell proliferation, and inhibited tube formation. Conversely, overexpression of DDAH1 decreased ADMA, and increased NO production, cell proliferation, migration and tube formation. The observed DDAH1 effects on NO and ADMA production are consistent with the report from Pope et al (26). Of great interest was our finding that DDAH1 overexpression dose-dependently increased p-Akt ^Ser473 and Akt activity, and that DDAH1 gene silencing decreased p-Akt ^Ser473 and Akt activity. The increase of p-Akt^Ser473 after DDAH1 overexpression was not blocked by ADMA, the nonselective NOS inhibitor L-NAME, or the soluble guanylyl cyclase inhibitor ODQ. Furthermore, the addition of cGMP or the permeable cGMP analogue PCMPT had no effect on the p-Akt ^Ser473 content of the HUVEC. These findings indicate that the alterations of p-Akt ^Ser473 in response to DDAH1 overexpression or DDAH1 knockdown were not mediated through the NO-cGMP signaling pathway. Furthermore, the DDAH1 C273S mutant which lacks DDAH activity was also able to activate Akt.

Since eNOS is a known downstream target of Akt, it was not unexpected that the increase of p-Akt ^Ser473 and Akt activity caused by DDAH1 resulted in an increase of p-eNOS^Ser1177 (27) with increased eNOS activity and NO production. The increase of p-eNOS^Ser1177 that we observed in response to DDAH1 overexpression is consistent with a report that transgenic mice overexpressing DDAH1 had increased myocardial p-eNOS^Ser1177 after ischemia-reperfusion (28). In contrast to these findings, Pope et al reported that DDAH1 siRNA caused no change of eNOS activity in bovine aortic endothelial cells (26), and Wang et al reported that siDDAH-1 caused no change in NO released by isolated rat mesenteric resistance vessels in response to acetylcholine stimulation (24). It is possible that differences in culture conditions and cell type might have contributed to the differing findings between these studies. Wang et al reported a 35-50 % reduction in DDAH1 protein following siDDAH-1, whereas we achieved a greater than 80% reduction; the greater reduction of DDAH1 protein expression in our study also could have contributed to the difference in results.

Since ADMA is a competitive inhibitor of NOS, one would anticipate that addition of L-arginine would restore NO production in siDDAH1 cells. However, we found that arginine only partially restored NO production after DDAH1 knockdown. This finding is consistent with a previous report from Pope et al (26). They suggested that DDAH1-silencing might lead to ADMA accumulation in sites that are not freely exchangeable with L-arginine (26). Our study provides an alternative explanation, namely, that DDAH1 acts to increase NO production not only by removing endogenous NOS inhibitors, but also by causing Akt phosphorylation which in turn increases eNOS activity by increasing p-eNOS^Ser1177. The ability of Akt to regulate NO production by increasing p-eNOS^Ser1177 has been previously reported from Sessa’s group (27,29). Our finding that the PI3 kinase inhibitor Ly294002 attenuated the increased p-Akt ^Ser473 and NO production in DDAH1 overexpressing cells indicates that regulation of NO production by DDAH1 was a downstream effect of Akt activation.

Tokuo reported that DDAH1 binds to neurofibromin1 (NF1) (30), a tumor suppressor with function of a GTPase-activating protein (GAP) which inhibits low molecular weight G proteins such as Ras by stimulating their intrinsic GTPase activity (31,32). Increasing phosphorylation of NF1 by PKA negatively regulates NF1 GAP activity (33) and consequently activates Ras. Their results suggested that DDAH1 does not directly affect PKA kinase activity, but up-regulates PKA phosphorylation accessibility of NF1 (30). In the present study we found increased Ras activation in DDAH1 overexpressing cells. Furthermore, inhibition of Ras by Ad-dnRas infection or the Ras inhibitor manumycin A was able to block the Akt activation, supporting the concept that Ras activation is upstream of Akt and that Akt activation inhibited ERK activity, at least in part, through phosphorylation of Raf at ser259 (34). In addition, we found evidence for DDAH1/Ras interaction, together with DDAH1/NF1 interaction (33). Together, these findings identify a novel pathway by which the DDAH1/Ras/NF1 complex is able to regulate endothelial function through the Ras-PI3-Akt pathway.