Percutaneous coronary intervention (PCI) with stent implantation (1) is still challenged by a 2% annual rate of stent-related adverse events which persist in follow-up (2–4). In-stent restenosis (ISR) due to hyperproliferative neointimal (NI) tissue proliferation, remains a leading cause of stent related adverse events. Despite this, the pathophysiology behind NI proliferation remains incompletely understood, with pathologic smooth muscle cell (SMC) migration and proliferation felt to contribute to some extent (1, 2, 5). ISR remains the focus of ongoing investigation to improve our understanding of the underlying biology and develop novel therapeutics to improve stent-related outcomes (6).

Adenosine (ADO) carries numerous regulatory roles both intracellularly and extracellularly, impacting several cell populations involved in vascular homeostasis, including SMCs (7–9). Adenosine signals primarily via 4 receptors with preclinical work suggesting a potential role for adenosine receptor (ADOR)-A2B in the regulation of the vascular effects observed, particularly NI proliferation (10). Indeed, in vitro and in vivo arterial injury models supports adenosine mitigating NI formation via ADOR-A2B-mediated inhibition of SMC proliferation (11–13). Despite this promising preclinical work, ADO has yet to establish itself as a viable therapeutic approach for ISR, owing in part to its innate limitations, particularly its short half-life (7–9, 14). However, focused ADOR small molecule agents have demonstrated improved utility and are used broadly clinically (15). Hence, evaluation of alternative agents which modulate adenosine signaling, but with improved clinical pragmatism holds promise for potential therapies.

Dipyridamole (DP) is an established, cost-effective, FDA-approved therapy with a broad range of clinical indications (16). DP improves vascular outcomes in several settings and mitigates restenosis rates following revascularization in both preclinical and clinical studies; however, DP's therapeutic effect in the presence of stent implantation remains largely unknown (17). DP functions primarily as a phosphodiesterase (PDE) 5/6 inhibitor, mitigating cAMP/cGMP breakdown, a mechanism which previously demonstrated clinical benefit for ISR reduction (18). However, DP is also known to augment circulating ADO levels primarily via ENT-1 re-uptake inhibition (7, 17). Preclinical studies suggest that DP may enact its vasculo-protective effects via ADOR-A2B-mediated inhibition of SMC proliferation, though this is based upon in vitro and preclinical models with limited data in the context of stent implantation and conventional dual antiplatelet therapy (DAPT) (13).

We have previously reported the utility of a preclinical rabbit stent model for evaluation of stent healing utilizing optical coherence tomography (OCT) (6). Given the hypothesized impact of adenosine biology on vascular remodeling after stenting, we sought to evaluate the efficacy of DP to mitigate ISR in this rabbit stent model, followed by in vitro studies to provide mechanistic insights into the observed effects.

In-stent restenosis stemming from NI tissue proliferation continues to cause adverse events following percutaneous treatment of obstructive CAD. DP is an established therapeutic agent for improving vascular outcomes with purported adenosine-mediated effects. In a translational stent model with conventional DAPT, we demonstrate that dipyridamole mitigates NI tissue formation and improves stent strut healing, though no difference in NI proliferation was noted in the presence of atherosclerosis. DP mediates these effects, in part, from reduced SMC content driven by diminished SMC proliferation and migration while altering SMC phenotypic switching. Selective modulation of ADOR-A2B and siRNA knockdown supports an ADOR-A2B mediated component to these observed effects. Thus, therapeutic efficacy derived through either reducing NI proliferation or optimizing stent healing may be achievable by either repurposing DP and/or development of ADOR-A2B specific agonists.

Contextualizing our preclinical results within the known clinical data is of importance. Despite advancements and conventional DESs, target lesion failure (TLF) continue to occur at rates of 5.0% in the first year, followed by an annual 2% annual rate up to 5 years which does not plateau—a focus of ongoing study (4). Clinically, DP has had mixed results in reducing restenosis with early PTCA RCTs—not employing a vascular scaffold like a stent—demonstrating no difference in ISR rates (33). Subsequently, RCTs assessing hemodialysis graft patency rates have shown improved graft patency rates from 23% to 28% at 1 year with DP (34). Meta-analyses of all-comer human revascularization studies assessing DP demonstrated a reduction in vascular occlusion rates from 31% without to 23.5% with DP (RR 0.77, 95% CI 0.67–0.88), with a consistent effect across medical regimens and vascular beds (17). However, clinical data assessing DP with conventional stent placement and DAPT remains very limited (35). Hence, the reported 16.6% NI burden relative reduction and improved strut healing in the presence of stent placement is promising. However, the use of BMSs in this study and the abrogated NI reduction with atherosclerosis are important, suggesting attenuated therapeutic effect may be observed in the presence of established atherosclerosis and conventional drug-eluting stents. Nonetheless, given the FDA-approved status of DP, rapid translation to randomized clinical evaluation of DP in addition to DAPT following conventional DES implantation could readily be performed to definitively establish the therapeutic efficacy of DP in the context of real-world atherosclerosis and DES implantation.

Dipyridamole's effect on restenosis and vascular SMCs is well described in preclinical models. Meta-analyses of preclinical models assessing DP therapy demonstrated a 47% relative risk reduction in vascular occlusion post intervention (RR 0.53, 95% CI 0.4–0.7, p < 0.00001) in addition to a 13% relative reduction in mean difference of NI burden (standard mean difference −1.13, 95% CI −1.74–−0.53, p = 0.0002) (17). Faxon et al. assessed the utility of ASA and DP in rabbit iliac angioplasty model with diet-induced atherosclerotic lesions utilizing histologic assessment at 4 weeks demonstrating an improvement in luminal diameter with oral ASA + DP therapy compared to control (1.3 ± 0.6 mm vs. 0.7 ± 0.6 mm, p < 0.05) (36). This effect was felt to be, in part, related to early platelet accumulation with notable thrombosis present in the control group (36, 37). Singh et al. subsequently employed a rabbit model of femoral and carotid balloon injury utilizing local DP delivery (to abrogate concerns of binding to serum proteins systemically), demonstrating a 63% inhibition of SMC proliferation with a 20% reduction in NI thickness (p < 0.05) (38). Dubey et al. provided mechanistic insights into these observed DP effects with in vitro human aortic SMCs demonstrating inhibition of SMC proliferation and diminished collagen synthesis/extracellular matrix deposition, a plausible means by which DP mitigates NI proliferation (13). However, these studies did not employ conventional antiplatelet regimens nor did they assess stent implantation—important limitations in translating to current day interventional practice. Hence, our study sought to replicate contemporary medical and device therapy finding reduced NI burden, improved strut healing and reduced NI SMC content in the presence of stent placement. Moreover, the dose-dependent inhibition of SMC migration and proliferation observed in vitro were robust, with apoptosis noted with escalating doses, but without concern of a cytotoxic/apoptotic effects for the 20 uM dose of DP selected for our in vitro experiments. Considering the lessons learned from excessive cellular inhibition with first generation DESs, namely incomplete healing and thrombosis, the appropriate balance of cellular inhibition without toxicity is critical for any stent related therapeutic agent, with dosing varying on whether it is employed systemically or locally (4).

The mechanisms underpinning the purported benefits of DP in mitigating NI proliferation and SMC modulation remain complex, but substantial data supports a connection to adenosine and ADOR-A2B signaling. Preclinical murine wire injury models in ADOR-A2B knockout mice demonstrated augmented NI formation in mice lacking ADOR-A2B (10). Use of 2-chloroadenosine (stable adenosine analog) has been shown to mitigate SMC proliferation, migration and collagen synthesis via activation of ADOR-A2B receptors (11, 13). Dubey et al. documented mechanistic insights centered around cyclin D based on in vitro human coronary SMCs with siRNA induced ADOR-A2B knockdown. Briefly, ADOR-A2B signaling leads to activation of adenylyl cyclase leading to augmentation of cAMP and PKA; PKA subsequently inhibits proliferation by blocking several pathways (ERK1/2, Akt and Skp2) that converge at Cyclin D with the collective result being diminished G1 cyclin expression and reduced cell cycle progression as a result (11). This mechanism is also maintained in circulating progenitor SMCs, another documented source of NI proliferation (5, 12). The role of adenosine and circulating progenitors is also supported by rescue of the observed effects by wild type bone marrow transplant in ADOR-A2B knockout mice (10). Hence, ADO is intricately linked to vascular homeostasis and this mechanism demonstrates multiple potentials for DP modulation. DP enacts many effects via PDE 5/6 inhibition which leads to cAMP/cGMP breakdown, hence DP will augment cAMP levels which would stimulate PKA. Indeed, breast malignancy studies have suggested DP induced inhibition of cyclin D1 expression as a mechanism behind diminished malignant cell proliferation (39). Moreover, DP is known to inhibit adenosine re-uptake, thereby augmenting circulating ADO levels from SMC culture in vitro (7, 13). Additionally, Dubey et al. suggested that the DP augmentation of ADO levels may mitigate SMC proliferation and collagen synthesis via activation of ADOR-A2B receptors (13). Lastly, altered SMC collagen and extracellular matrix generation suggests another mechanistic insight—phenotypic modulation. Vascular SMCs maintain plasticity allowing substantial variations in phenotype varying from differentiated quiescent forms to de-differentiated forms in response to injury. Numerous signaling pathways modulate this process including SM-alpha-actin (ACTA2) and Kruppel-like factor 4 and 5 (KLF4,5). In response to injury, the de-differentiated “synthetic” phenotype prevails being characterized by a loss of contractile markers and augmentation of proliferation, migration and protein synthesis—if this process persists beyond the initial injury repair, pathologic NI tissue formation ensues (40).

Collectively, our results align with the mechanisms previously reported. We demonstrate that DP has beneficial effects in mitigating SMC content, migration and proliferation without augmenting apoptosis. These beneficial effects are likely mediated via ADOR-A2B-related mechanisms with abrogation of the DP effect noted with selective A2B inhibition. Moreover, DP, selective A2B modulation and A2B siRNA knockdown all result in modulation of SMC phenotypic switching factors—supporting the role of DP and ADOR-A2B in this process. This aligns with prior studies noting DP's impact on diminishing collagen synthesis (13), while also being able to reverse the inhibitory effects of DP via administration of PDGF, supporting a pote ntial role of DP and SMC phenotypic switching (38). While we did not note a clear rela tion between DP administration and elevated circulating adenosine levels, this is not unexpected considering the complex processes involved in circulating adenosine levels (28, 29). Moreover, unchanged systemic adenosine levels do not necessarily preclude tissue-level alterations in adenosine biology which may enact the observed effects.Going forward, clinical evaluation of DP is warranted coupled with ongoing preclinical evaluation of A2B specific agents for mitigating restenosis.

Dipyridamole mitigates neointimal formation and improves stent strut healing in a preclinical model of stent implantation with conventional antiplatelet therapy. This therapeutic effect is abrogated in the presence of atherosclerosis. Clinical studies of DP in addition to conventional therapy post stent implantation may be warranted to evaluate for clinical efficacy.