Dynamic protein palmitoylation cycling: A new pathway impacting peripheral arterial disease?

Peripheral artery disease (PAD) is defined as flow limiting focal atherosclerosis in arteries other than the coronary and intracranial vessels, and commonly affects vessels in the lower extremities.¹ The incidence of PAD is nearly equal to coronary artery disease, with a high prevalence in older individuals and those with co-morbid insulin resistance (IR) and type II diabetes (T2D).^{1, 2} In its most severe form, critical limb ischemia (CLI), patients experience leg pain at rest and/or non-healing foot ulcers due to complete occlusions in one or more of the major arteries of lower limbs.¹ The ability of these patients to walk and avoid limb loss depends on neovascularization, a complex process of angiogenesis (new capillary growth) and arteriogenesis (maturation or de novo growth of collateral conduits), that supports tissue perfusion of ischemic tissues.³ Although managing PAD patients with standard medical therapy reduces the risk of cardiovascular events, these agents do not reduce the risk of major lower limb amputation. Novel therapeutic approaches, such as gene delivery of angiogenic growth factors have shown promise in preclinical models ⁴, but have failed in the majority of clinical trials.⁵ Thus, there is an urgent need to understand the underlying mechanisms of impaired neovascularization in an effort to develop novel drug targets to offset the tremendous socioeconomic burden of PAD.

Insulin resistance and T2D is characterized by elevated circulating glucose and lipid levels, and reports demonstrate that both hyperlipidemia and hyperglycemia can result in an imbalance between endothelial derived anti-inflammatory, vasodilatory factors (e.g, nitric oxide, NO) and pro-inflammatory, vasoconstrictive factors (e.g., endothelin-1) factors.⁶ Emerging evidence in endothelial biology suggest that specific metabolites are capable of directly modulating transcriptional and enzymatic activity.⁷ However, little is known about how alterations in the cellular metabolome perturb vascular remodeling in PAD. Since FFA are elevated by multiple sources during IR and T2D, assessing protein palmitoylation is an obvious target for elucidating how alerted metabolic flux contributes to ineffective neovascularization in PAD.

Palmitoylation (S-acylation) is a post-translational modification that influences membrane association and subcellular trafficking of several proteins. Due to the labile nature of the thioester bond between palmitate and cysteine, this reaction is reversible and can impact protein dynamics within cells.⁸ The addition of palmitate to cysteine residues in proteins is catalyze by a family of palmitoyltransferase enzymes ⁹ and removal of palmitate occurs by the acyl-protein thioesterase 1 (APT1). ¹⁰ Thus, the dynamics of protein palmitoylation are governed by palmitoyltransferases and APT1. In this current issue of, Circulation Research, Wei et al.¹⁶, test the hypothesis that inadequate palmitoylation cycling promotes endothelial instability in PAD by deleting a key depalmitoylation enzyme, acyl-protein thioesterase 1 (APT1). In theory, deletion of APT1 would enhance protein palmitoylation in endothelial cells by reducing the turnover of palmitate on hundreds of proteins, thereby influencing signal transduction.

By using an endothelial specific (Tie2-Cre and VE-Cadherin-Cre) knock-out mouse models of APT1 (APT1 ECKO), these investigators demonstrated that the loss of APT1 in EC did not impact the viability of mice, implying that blocking protein depalmitoylation is not essential for life or that there may be other APT-like enzymes to maintain the levels of protein palmitoylation. However, blood flow recovery following unilateral hind-limb ischemia (HLI) was mildly impaired in APT1 ECKO mice compared to controls. Histological examination revealed that APT1 ECKO mice had fewer microvessels in the post-surgical ischemic zone, less pericyte coverage (via NG2 staining) and greater fibronectin deposition compared to controls. Perturbed fibronectin metabolism was further confirmed in a series of in vitro experiments using mouse and human endothelial cell lines. For example, APT1 deficient EC developed disorganized sprouts in in vitro angiogenesis assays, but only when cultured using fibrin-based gels. In addition, APT1 deficient EC demonstrated inappropriate fibronectin-integrin interactions. Together this data provides evidence that impaired palmitoylation cycling in ECs in the setting of ischemia impairs neovessel growth and maturation, which may be due improper cell-matrix or cell-cell interactions.

Since it is accepted that many signaling proteins can be reversibly palmitoylated, the authors used a proteomics approach to delineate potential APT1 substrates that may be mediating these vascular perturbations. Functional annotation clustering of candidate APT1 substrates identified focal adhesions proteins as the highest scoring pathway. Of the proteins in this cluster, the small GTPase, R-Ras, was of particular interest since this protein has been described as a regulator of vascular differentiation that primarily effects the maturation of remodeling blood vessels during developmental, reparative and pathological angiogenesis.^{11, 12} For instance, studies have shown that tumors in R-Ras KO mice show less pericyte coverage, elevated levels of fibronectin rich provisional matrix and disrupted cell-cell interactions.¹² On the other hand, other studies report that forced R-Ras expression in human umbilical vein endothelial cells (HUVEC) increases survival in low serum conditions and inhibits vessel sprouting and tube formation in vitro angiogenesis assays.¹¹

In the current study, palmitoylation of R-Ras was confirmed in several APT1 deficient cell lines, whereas transfection of APT1 reduced R-Ras palmitoylation levels, thus confirming R-Ras as an APT1 substrate. Culture of HUVEC in high glucose led to increased palmitoylation of R-RAS without changing APT1 mRNA or protein levels, suggesting decreased APT1 enzyme activity or increased palmitate availability via de novo lipogenesis. Although the latter was not explored, APT1 enzymatic activity was measurability decreased in HUVEC. Similarly, APT1 activity was decreased in lung extracts from diabetic (db/db) mice, which coincided with increased palmitoylated R-Ras in skeletal muscle from db/db mice compared to control mice. In subsequent experiments, the mechanism of reduced APT1 activity under hyperglycemic conditions was deemed to be due, at least in part, to reduced acetylation of APT1, as the class I and II HDAC inhibitor, trichostatin A (TSA), partially rescued APT1 activity under these conditions. However, it is unclear how cellular acetylation was perturbed during hyperglycemia, especially since it is expected that cytosolic acetyl CoA concentrations would be increased secondary to elevated glucose carbon flux and mitochondrial citrate transport to the cytosol.¹³ Nonetheless, hyperglycemia recapitulated some phenotypes of APT1 deficiency, such as palmitoylated R-Ras, altered cell-cell contacts and disrupted fibronectin turnover.

Previous studies by our group have demonstrated that palmitoylation is necessary to localize eNOS to caveolae for optimal stimulated NO release and that palmitoylation serves as a kinetic trapping mechanism impacting the subcellular turnover of the enzyme.¹⁴ Similarly, in this study, ATP1 deficiency caused persistent palmitoylation of R-Ras, which prevented trafficking of this enzyme, as demonstrated by confocal imaging of reduced localization to the plasma membrane and cell-cell junctions. In addition, these observations were coupled with reduced speed and displacement of this enzyme during particle tracking experiments. Taken together with evidence from previous studies describing the function of R-Ras in vasculature maturation, it seems that many of the vessel phenotypes observed in endothelial APT1 ECKO coincide with reduced R-Ras activity, likely due to an inability of R-Ras to localize to specific subcellular environments.

The results from the experiments presented in this article provide the first evidence of perturbed palmitoylation cycling as a novel factor driving ischemic and hyperglycemia induced vascular remodeling abnormalities. Nevertheless, it is important to discuss some limitations of these studies. Notably, the temporal expression levels of R-Ras fluctuate during vessel injury, with expression levels being low in the initial reparative phases and increasing over time during the transition from vessel proliferation to quiescence.¹¹ Thus, it is likely that reduced R-Ras activity due to persistent palmitoylation has little effect in the early post ischemic period following hind-limb ischemia. In line with this, perfusion recovery was almost identical between APT1 ECKO and control mice 1 week following surgery. Second, both animal models and cell culture systems of APT1 deficiency do not adequately recapitulate the systemic milieu in severe PAD. Indeed high fat feeding has been shown to greatly impair perfusion recovery following HLI, in part, due to reductions in VEGFR2 activation and impaired AKT/eNOS signaling¹⁵ and it would have been interesting to see how APT1 ECKO mice respond to HLI after such a diet. In this same vein, treating cells with high glucose alone neglects the effects of elevated FFAs on palmitoylation itself or on cellular processes that may influence palmitoylation cycling such as ER stress. Third, some of vascular phenotypes following in APT1 ECKO are at odds with what would be expected due to reduced R-Ras activity¹¹, such as reduced vessel formation (lower CD31+ expression in ischemia zone) and suggest that perturbed palmitoylation cycling influences other important pathways involved in post-ischemia neovascularization due to the several other APT substrates affected.

In summary, the study by Wei et al. provides the first evidence of perturbed palmitoylation cycling as a novel factor driving ischemic and hyperglycemia induced vascular remodeling abnormalities. Despite some limitations, this study highlights the need to assess unexplored mechanisms, such as palmitoylation cycling, that halt or impair neovascularization of ischemia tissue in PAD. As such, these studies provide the groundwork for future experiments that may lead to novel drug targets for a disease with a large socioeconomic burden and without any efficacious therapies that attenuate the risk of lower limb amputation. Future studies should explore more clinically relevant animal models, and also explore other APT1 targets that may be mediating some of the effects of persistent palmitoylation on vascular remodeling.