l-Phenylalanine Restores Vascular Function in Spontaneously Hypertensive Rats Through Activation of the GCH1-GFRP Complex

Visual Abstract

C ardiovascular diseases pose a considerable societal and economic burden on health care systems (1). Such diseases are usually associated with functional and structural changes within the vascular network as well as concomitant increases in oxidative stress (2). Endothelial dysfunction is characterized by impaired vasodilation, mainly due to loss of nitric oxide (NO) signaling (3,4). NO biosynthesis in the vasculature is primarily catalyzed by endothelial nitric oxide synthase (eNOS) (5), and tetrahydrobiopterin (BH 4 ) is an essential cofactor for all 3 isoforms of NOS (6,7). When BH 4 bioavailability is limited, NOS may become "uncoupled," producing superoxide at the expense of NO, thereby potentiating oxidative stress (8).
Thus, raising endothelial BH 4 levels has been suggested as a strategy to maintain healthy NO production and bioavailability at the level of the endothelium (9,10). To support this, intra-arterial administration of BH 4 improves endothelial dysfunction in patients with hypertension (11), coronary artery disease (12), and hypercholesterolemia (13).
Therefore, other strategies to raise endogenous BH 4 bioavailability at the level of the endothelium are desirable. BH 4 is synthesized from GTP in a reaction where the committing step is mediated by GTP cyclohydrolase-1 (GCH1) (17). Modulation of GCH1 expression has been shown to regulate BH 4 , NO, and cardiovascular function (18)(19)(20). GCH1 is subject to feed-forward regulation by L-phenylalanine (L-phe), via an allosteric protein interaction with GCH1 feedback regulatory protein (GFRP) (21,22). This GCH1-GFRP complex is operative in humans because oral challenge with L-phe leads to a 3-fold rise in plasma biopterin levels (a correlate of BH 4 )-an effect that is attenuated in patients with a loss-of-function GCH1 mutation (23). Targeting endogenous BH 4 biosynthesis, by activating the GCH1-GFRP axis pharmacologically, thus represents a method to enhance vascular BH 4 levels at the level of the endothelium, circumventing the poor bioavailability following oral BH 4 administration (14,24).
To support this hypothesis, it is known the GCH1-GFRP axis regulates BH 4 and NO in endothelial cells (25). Overexpression of GFRP reduces basal BH 4 levels (26) and attenuates the rise in BH 4 and NO that occurs in response to a proinflammatory stimulus (27).
Additionally, the primary source of BH 4 appears to be derived from GCH1 localized within the vascular endothelium, and GFRP is coexpressed within these cells (28,29). Finally, oral L-phe elicits a rise in vascular BH 4 -an effect that is absent in mice lacking endothelial GCH1 (24).  legend, or in the Supplemental Table 2. As previously described (24,30), the combination of purified recombinant GFRP with GCH1 protein had a higher basal activity than GCH1 alone in vitro. The addition of L-phe (2 mmol$l À1 ) had no effect on purified GCH1 activity alone but caused a significant rise in GCH1 activity when coincubated with GFRP, confirming that L-phe is an allosteric regulator of the GCH1-GFRP complex only ( Figure 1A).
In vivo, oral L-phe (100 mg$kg À1 ) bolus to WKY and SHR significantly increased plasma BH 4 levels within 30 min, and levels returned back to baseline within 4 h ( Figure 1B). Correspondingly, a significant rise in nitrite levels was also detected within 30 min, but whereas this returned to baseline in WKY, it remained elevated in SHR for at least 4 h ( Figure 1C).
Interestingly, there were no statistically significant differences in BH 2 and biopterin in all groups although trend increases were observed ( Figures 1D   and 1E).  Nitrite levels were significantly lower in the heart, lung, and liver, but not the plasma, of SHR compared with WKY ( Figure 2B, Supplemental Figure 2). Both bolus and long-term supplementation of L-phe normalized nitrite levels in SHR to WKY control which was restored to WKY values following short-term bolus and long-term L-phe treatment ( Figure 3A). As anticipated, superoxide levels were higher basally in SHR compared with WKY ( Figure 3B).

EFFECTS OF BOLUS AND LONG-TERM L-PHE
Bolus dose or long-term administration of L-phe significantly reduced superoxide levels in SHR ( Figure 3B). Superoxide dismutase, the positive control, reduced superoxide in all study groups ( Figure 3B). Again, we observed no significant changes in BH 2 or biopterin in aortic tissue AE L-phe administration ( Figures 3C and 3D). Unfortunately, in aortic tissues, nitrite levels fell below the limit of detection and were therefore not quantifiable.  confirming that L-phe was absorbed following oral gavage ( Figure 6A, Supplemental Table 1). We did not detect a significant rise in the L-phe/L-tyr ratio in animals treated long-term with L-phe. This is not surprising because the long-term ad libitum L-phe dosing was at a much lower dose than the short-term bolus dose challenge. Although there were trends of decreased dopamine and increased adrenaline/ noradrenaline between SHR and WKY, these did not reach statistical significance in most tissues. The exception was the heart, where adrenaline/ noradrenaline levels were significantly higher in SHR than WKY basally, but equaled WKY levels following L-phe treatment (Figures 6B and 6C). aortic rings, with intact or denuded endothelium following short-term incubation with 0.5 mmol$l À1 L-phe. Data represent mean AE SEM n ¼ 6 animals (in triplicate/animal) for mesenteric arteries and aortic rings (*p < 0.05 and *** p < 0.001 for the whole curve). other abbreviations as in Figure 1.
Heikal et al. Importantly, this is a proof-of-concept study demonstrating that GCH1-GFRP is a rational therapeutic target for vascular dysfunction. Hence, the development of L-phe mimetics that selectively bind to and enhance this protein complex may be of clinical value. Our data suggest that L-phe could itself be translated into the clinic given the minimal effects observed on catecholamines but should be advanced with caution, given L-phe's diverse biological action and potential for predictable adverse drug reactions.
GCH1 binds to GFRP to form a protein complex that is receptive to allosteric regulation by both L-phe (feed forward) and BH 4 (feedback) regulation (21). Our results have confirmed numerous previous reports that L-phe only enhances GCH1 activity when it is bound to GFRP (24,30).
In addition to its essential cofactor role for NOS activation, BH 4 is also required by phenylalanine hydroxylase to catalyze the conversion of L-phe to L-tyr, which is further converted to dopamine, adrenaline, and noradrenaline (41). L-phe thus regulates its own metabolism via feed-forward activation of GCH1-GFRP with subsequent increases in BH 4 and hence phenylalanine hydroxylase activity. This is important because sustained elevation of L-phe can become neurotoxic (42). Indeed, BH 4 has been successfully used as a treatment for a subset of patients with phenylketonuria (43). Consistent with raised biopterin levels seen in patients after L-phe loading (23) and our previous observations in mice (24), plasma BH 4 levels were significantly increased in control WKY and SHR after 100 mg$kg À1 L-phe bolus oral challenge in the present study.
The SHR is an appropriate model to study endothelial dysfunction because the animals demonstrate reduced NO signaling, reduced endothelialdependent vascular relaxation, enhanced cardiovascular remodeling, and increased oxidative stress (19,44,45). In this study, lung BH 4 levels were lower in SHR than in age-matched WKY, consistent with published reports (46). Following a short-term oral dose (4 h) or long-term daily (8 weeks) L-phe challenge, tissue BH 4 levels in SHR were restored to control WKY levels.
Correspondingly, we observed increased aortic superoxide production in SHR basally and L-phe administration increased aortic BH 4 and concomitantly reduced superoxide levels. These data support the hypothesis that L-phe activates the GCH1-GFRP complex in vivo, raising endogenous BH 4 biosynthesis to support full "coupled" NOS activity, thereby reducing oxidative stress in this model of hypertension.
Limited BH 4 bioavailability is believed to lead to NOS uncoupling, generating superoxide instead of NO (8,47). In SHR, the observed endothelial To verify whether L-phe could activate the GCH1-GFRP axis functionally, a series of studies were carried out using fresh conduit (aortic) and resistance (mesenteric) blood vessels from WKY and SHR.
Consistent with published reports, Ach-mediated vascular relaxation in SHR was significantly impaired in comparison to WKY rats. Short-term L-phe incubation within the organ bath significantly improved vascular relaxation in SHR vessels yet had no effect on WKY. This implies that L-phe, via local elevation of BH 4 within the vasculature, enhances NO bioavailability and endothelial function only in circumstances where the pathway is dysfunctional. This is consistent with the differential effects on plasma nitrite between SHR and WKY discussed in the preceding text. Interestingly, L-phe had a more pronounced effect on vascular relaxation in mesenteric   Thus, there is still much work to be undertaken to improve the efficacy and safety of pharmacotherapies that enhance NO bioavailability, but our present study provides the first proof-of-concept data that the GCH1-GFRP complex is a rational therapeutic target to achieve BH 4 elevation and NO restoration within blood vessels.
To translate these findings further, we propose 2 parallel research strategies. The first clinical development strategy would investigate the impact of Lphe administration on vascular function using flow mediated dilatation, in patients with existing endothelial dysfunction versus healthy controls. It may be predicted that flow-mediated dilatation would be improved in the patient cohort, whereas negligible effect would be seen in the nonpatient controls.
However, L-phe, as a therapy, may have challenges given its diverse biological activity, raising potential safety concerns, and these would need to be concomitantly investigated in trial participants. The second parallel strategy would be focused around