Phosphodiesterase type 5 Inhibitors (PDE-5I) as a potential therapeutic drug in stroke: A systematic narrative review

Objectives. Stroke is an acute cerebrovascular disease with high morbidity and mortality rate. Many stroke management strategies can improve prognosis and quality of life, but only to a certain extent. Phosphodiesterase inhibitor-5 (PDE-5i) has shown benefits in numerous preclinical studies, but human studies have been inconclusive. Material and Methods. For this narrative review, we conducted a systematic search following the PRISMA statement guideline from inception until November 2022. The search was conducted in PubMed, ScienceDirect, EBSCOhost, and ProQuest. We included all human and animal studies on this topic, including preclinical studies, randomized clinical trials, and case studies. All the excluded studies are reviews or non-English studies. Review. Many PDE-5i have shown benefits in anatomical and functional outcomes in preclinical acute experimental stroke models. PF-03049423 was safe and well tolerated in humans but had no significant impact on neurological or functional outcomes. Sildenafil on acute/subacute was deemed safe to use and demonstrated improvement over baseline, but power remains unknown due to the lack of a control group. Tadalafil use resulted in a reduction in regional cerebral blood flow in subjects with a history of stroke. Conclusions. Despite promising results in preclinical studies, current evidence shows that PDE-5i does not affect clinical or functional recovery in people with acute stroke. The disparity in the intricacy of stroke pathophysiology between human and animal models was crucial. Further research in a larger population with more consistent stroke onset, medicines, and doses is required for a more conclusive result.

stroke management/therapies mainly concentrate on thrombolytic agents, neuroprotective drugs, mechanical thrombectomy, stent, angioplasty, surgical treatment (decompressive craniectomy and carotid endarterectomy), and rehabilitation training [4]. Comprehensive therapy and early thrombolysis/mechanical thrombectomy have been shown to improve prognosis and quality of life but only to a certain extent.
PDE-5Is, such as Sildenafil (Viagra®), are widely available and primarily used to treat erectile dysfunction and pulmonary hypertension (PH). Its use in stroke patients was still debatable [5]. Many preclinical studies, mostly in animals, have suggested that PDE-5I may improve functional outcomes after an ischemic stroke. These studies have shown the potential use of PDE-5I, which exerted neuronal and synapse protection, improved angiogenesis, cerebral blood flow, and overall brain function, and reduced apoptosis, oxidative stress, and neurodegeneration [6][7][8][9].
The study in humans yielded varying results, whereas many animals benefited. Due to the ambiguity of the evidence, the current study aims to compile and compare all relevant studies related to the therapeutic potency of PDE-5i in stroke subjects, both in animals and humans, to reach a more definitive conclusion.

Search strategy and selection criteria
For this narrative review, a structured literature search was conducted from inception until November 2022 to identify published articles on the potential use of PDE-5i in the recovery of stroke patients. The search was conducted in PubMed, ScienceDirect, EBSCOhost, and ProQuest, by using the following MeSH term keywords "Phosphodiesterase 5 Inhibitor", "PDE 5 inhibitor", "Stroke," "Cerebral stroke," "CVA," and "Brain vascular accident." The search terms, especially for PDE 5 inhibitors, were broad to encompass every drug with a PDE 5 inhibitor mechanism. We did a systematic searching method for this narrative review following the PRISMA statement guideline [10], with a predetermined search strategy in which we identified potential studies, screened titles and abstracts, assessed full-text articles, and determined relevant studies.
All studies related to this topic, both in humans and animals, including preclinical studies, randomized clinical trials, and case studies, were included in the present study. We also reviewed each article's references to find other relevant studies or reports not identified by the search. After eliminating duplicates, the authors reviewed all titles and abstracts and excluded those articles with full text that failed to be retrieved. The exclusion criteria included review studies and non-English studies.

Data extraction
The following data were extracted from the studies selected for inclusion: (1) first author; (2) publication year; (3) region; (4) study design; (5) sample size; (6) age; (7) gender; (8) intervention; (9) results. To ensure the accuracy of all the data, the information was extracted by four independent authors, and conflicting data were resolved with consensus among all the authors.
The data retrieved in animal and human studies differed since many animal trials involved brain removal, whereas only clinical assessments were performed in humans. In animals, we obtained anatomical (infarct size, edema size, penumbra enlargement, angiogenesis, cerebral blood flow, vascular density, neuronal density) and functional (modified six-point scoring, beam walking test, rotarod test, adhesive removal test) outcomes. In humans, we retrieved clinical (mRS day 90, BI score, NIHSS score, cerebral perfusion) FIGURE 1. Study selection process following PRISMA guidelines and functional (sensorimotor, box and blocks test, hand-grip strength test, 10-meter walk test, repeatable battery assessment of neuro-psychological status naming and codings subtests, line cancellation test, and recognition memory test) outcomes for analysis.

Cerebral vascular dysregulation in cerebral ischemia
The advancement in technology and techniques enable us to measure regional cerebral blood flow (CBF), providing us with new insights into the alteration of CBF distribution following focal or global cerebral ischemia. Occlusion of a major cerebral artery severely impaired CBF to the area of the brain supplied by this artery. This impairment can be observed more clearly in the center of the ischemic territory, where the CBF reduction is at its greatest. Surrounding the 'core' area of cerebral ischemia is an area with less severe CBF reduction termed ischemic penumbra [11]. The fate of the penumbra then depends on the residual CBF and the duration of the flow reduction. Studies found that CBF in the penumbra ranged between 12 to 22 cc/100 g per minute. A further drop in the CBF below this critical point for a suffi-cient period resulted in tissue damage [12].
Re-establishment of the CBF in ischemic areas following perma-nent/transient arterial occlusion is beneficial to be done as early as possible. On the other hand, reperfusion may also exacerbate brain injury/ reperfusion injury through the development of hemorrhagic transformations or cerebral edema [13]. After reopening the vessel occlusion, reperfusion in the ischemic core typically shows a biphasic pattern in which CBF increases transiently (post-ischemic hyperperfu-sion/luxury perfusion) followed by more sustained reduction (hypoperfusion), confirmed by laser Doppler flowmetry studies. Post-ischemic hyperperfusion has been demonstrated in several animal and human stroke models [14][15][16][17][18]. What is interesting is that this hyperemic phase is not mediated by an increase in oxygen or glucose utilization and can be attributed to abnormal vasodilatation in the ischemic territory. Such abnormal vasodilatation has been observed to cause multiple negative effects, including lactic acidosis secondary to ischemia-induced anaerobic glycolysis [19] and or release of vasoactive mediators from the ischemic brain, including ions, metabolites, and reactive oxygen species [20]. The mechanisms leading to post-ischemic hypoperfusion are still unclear, but the hypothesis circle around the increase in cerebrovascular resistance due to microvascular compression and vasospasm, as suggested by a scanning electron microscopy study in animal models. Microvessels appeared constricted, compressed, and narrowed, presumably by vasoconstrictors released by the ischemic brain [21].

Vasodilatation in stroke
Nitric oxide (NO) was first named endothelium-derived relaxing factor (EDRF) in the late 1970s, a colorless, odorless gas that transmits biological information to relax smooth muscle cells and cause vasodilatation. This signaling molecule's primary functions include maintaining vascular tone, reducing inflammation response, balancing thrombotic-thrombolytic homeostasis, and regulating cell growth [22]. In the acute stage of cerebral ischemia, one of our physiological self-defense mechanisms is to increase focal NO production in the ischemic area, sustained for at least one hour [23,24]. Guanosine 3',5'-cyclic monophosphate (cGMP) is another critical regulatory factor mediating the vasodilatory effect on the cerebral vessels. cGMP is regulated through synthesis by guanylyl cyclase and degradation by phosphodiesterases, especially the phosphodiesterase type 5 (PDE5) enzyme, as it is a particular enzyme for hydrolysis cGMP. This molecule is closely related to the expression of NO, which activates soluble guanylate cyclase. cGMP is formed in response to nitric oxide (NO) by NO-sensitive guanylyl cyclases in two isoforms (NO-GC1 and NO-GC2). The other way to increase the cGMP level without raising the NO level was to block its degradation [25].

Drug profile and mechanism of action
Several mechanisms of action from PDE-5Is had been proposed: (1) PDE-5i increase expression of nitric oxide (NO) synthases. The NO increase will then activate soluble guanylate cyclase in myocytes, which con-verts guanosine triphosphate to cGMP. The increase in cGMP concentrations then reduced smooth muscle tone and caused vasodilations, meaning there was an increase in cerebral blood flow. (2) PDE-5i prevented the cGMP conversion to GMP. Thus, the cGMP accumu-lates and increases phosphokinase G (PKG), improving cerebrovascular perfusion and producing neurogenesis, angiogenesis, and synaptogenesis [26].
PDE-5i, such as Tadalafil and Sildenafil, showed neurogenesis enhancement and increased proliferating neural progenitor cells in the penumbra [29]. Several animal studies have demonstrated the protective effects on neuronal networks and synaptogenesis by reducing neuron loss, enhancing axonal remodeling, and modulating microglial function [6,7]. Further, Sildenafil reduced the capillary density loss, induced angiogenesis, and cerebral blood flow in the ischemic penumbra shown by MRI. These findings were accompanied by reduced neuronal apoptosis by the expression of apoptotic factors [8,9].

Current evidence of PDE-5i in acute stroke
A total of 16 preclinical studies in animal models and three studies in hu-mans were included in the present study. Table 2 summarizes the studies in animals, and Table 3 summarizes the studies in humans.
Cerebral blood flow, angiogenesis, and vascular density were improved in 4 studies, and no changes in 1 study [23]. Among the 4 studies, improvement was ob-served with Sildenafil [8,9,31], and Zaprinast [23]. All three Sildenafil studies used daily subcutaneous 10 mg/kg administration for seven days. No progress was observed in Verdenafil [35] with 10 mg/kg oral administration. Both Sildenafil and Zaprinast were able to cause a selective increase in CBF through the NO-cGMP pathway in the ischemic penumbra side. Due to the mechanism of CBF increase being located downstream of the NO-mediated cascade, tissue damage caused by NO by acting as a highly toxic radical in combination with the superoxide anion can be avoided [23].
Neuronal loss was significantly reduced in three studies with Yonkenafil [6], Sildenafil [7], and Icariside-II [30]. The Yonkenafil and Sildenafil stu dies used single intravenous/intraperitoneal doses ranging from 8 to 32 mg/kg, whereas the Icariside-II study used intravenous 4-16 mg/kg twice daily for seven days. Sildenafil inhibits neuronal loss via several mechanisms: (1) The Nogo-66 receptor (Nogo-R) pathway, as evidenced by increased Nogo-R expression in the cortex and striatum. When Nogo-R is activated, RhoA is activated, activating ROCK, which phosphorylates several substrates involved in the survival pathway. PTEN is a ROCK downstream target that negatively regulates Akt signaling, which is involved in protein synthesis and cell-to-cell modulation of neurite outgrowth and survival. Thus, inhibiting the expression of Nogo-R, RhoA, and p-PTEN expression will increase p-Akt and PI3K levels via a cGMP-dependent pathway, leading to axonal sprouting and decreased neuronal loss [7]. (2) Suppression of S100B and AQP4 co-expression in astrocytic structures, which regulate calcium influxes caused by reactive astrocytes to injury and the regulation of extracellular matrix volume, neuroinflammation, and calcium signaling, res pec ti vely [29]. (3) Increasing the expression of synaptophysin and reducing the expressions of uncoupling PSD-95 and nNOS, resulting in reduced synaptic damage and synapse structure protection in the ischemic brain [7]. (4) The cGMP-PKG activation results in lower levels of intracellular calcium ions (Ca2+), which are a significant initiator and activator of apoptotic pathways [29].
Neurological (modified six-point scoring, modified neurological severity score, sensorimotor) and behavioral functions (beam walking test, rotarod test, adhesive removal test) were improved in 11 studies, and no changes in only 1 study. Among the 11 studies, improvement was observed with Yonkenafil [6], Sildenafil [7,8,29,31,32,[36][37][38], Icariside-II [30], and Tadalafil [39]. Sildenafil is administered intravenous/intraperitoneally with doses ranging from 12 to 32 mg/kg, orally with doses ranging from 0.3 to 5 mg/kg, and subcutaneously with 10 mg/kg all showed improvement. No functional improvement was observed using Verdenafil [35] with 10 mg/kg oral administration. This result was surprising because studies with oral/subcutaneous were unable to reduce infarct area size but were able to improve the functional score significantly. This improvement in neurological outcome was suspected to be related to the significant increase in numbers of BRdU-(bromo deoxyuridine) and TuJ1-(bIII-tubulin) im munoreactive cells in the ischemic brain of rats, which are markers for progenitor cell proliferation in the subventricular zone and dentate gyrus [37,39].
Many experimental studies in rodents demonstrated some benefit of PDE-5i, though the extent of activity of these inhibitors and the specific underly- ing mechanisms remained unknown. Many factors, including the different experimental species and sizes/ages, duration of ischemic induction, the time before reperfusion performed, types of PDE-5i, doses, route of administration (oral, intravenous, or intra-peritoneal), and timing of PDE5 inhibitor administration, were contributed to the inconsistent findings in this study.
In humans, Cesare and colleagues [40] showed the use of PF-03049423, a selective and brain penetrant PDE-5 inhibitor, on acute stroke subjects was safe and well tolerated but did not significantly impact neurological or functional outcome measured by modified Rankin Scale (mRS), Barthel Index (BI), or National Institutes of Health Stroke Scale (NIHSS), along with other secondary assessment. Although the primary efficacy analysis using mRS showed the proportion of subjects with mRS score ≥2 at Day 90 was lower in the PDE-5i group (n = 42.6%, OR 0.74; 0.41 to 1.31 95% CI) compared to the placebo group (n = 46.2%), but the difference is not significant (p-value 0.49). A tiny difference in placebo in the interim was observed; thus, the author decided to terminate early using a futility rule due to a less than 20% probability of showing a statistically significant effect at the end of the study should they have continued. In the preclinical model, the use of PF-03049423 demonstrated a potential impact on functional improvement. Still, in the current clinical trial, the treatment with PF-03049423 did not show any clinical benefits relative to a placebo in a selected population of stroke patients over a limited time window.
Silver and colleagues [41] showed that using Sildenafil on acute-subacute stroke patients was considered safe. This study primarily assessed the safety use of PDE-5i in acute stroke patients but had a secondary outcome of brain function assessment using mRS, BI, and NIHSS. Among the ten subjects observed at day 90 (one death and one suicide due to a history of depression without suicidal attempts), the median NIHSS score was 2, the median BI was 95 (Range 15-100), and the median mRS score was 1.5 (range 0-5) which represented an improvement from baseline in all the survivors. The proportion of subjects with an mRS score ≥2 after Sildenafil use was 50%. Since there is no control group, thus the signifi-cance/power of the study cannot be determined, although this result was higher compared to the previous study by Cesare and colleagues [40] with 42.6% of the population having mRS score ≥2.
Several other studies of PDE-5i on humans had been conducted in non-stroke and healthy individuals. Outcomes assessed included mean blood flow velocity in the middle cerebral artery [43], average maximal velocities in the middle cerebral artery (Vmca) [44], SPECT imaging with Xenon 133 inhala-tion [44], and fMRI assessment [45]. Three studies showed no improvement in resting cerebral blood flow after receiving Sildenafil compared to a placebo. In contrast, a study with 14 healthy male volunteers found that one hour after administration of 100 mg of Sildenafil improves cerebrovascular regulation (CVR) [46]. Impaired CVR was associated with vascular risk factors and vascular abnormalities in general. Impairment of endothelium-dependent CVR was also postulated as a specific risk factor for small cerebral vessel disease [47,48]. A systematic review by Pauls and colleagues [49] concluded that PDE-5i only affects CBF in certain clinical conditions. PDE-5i might improve CVR measurement, but not basal CBF, especially in disorders characterized by an impaired endothelial dilatory response due to deficient nitric oxide-mediated signaling has disproportionate effects on brain microvascular responsiveness compared to the resting state. There is still a possible action of PDE-5i on resting CBF at the level of small arterioles, as this remains untested up to this date. Further studies using blood-flow-specific cross-sectional imaging techniques like ASL MRI are warranted to explore this hypothesis at the arteriolar level.
Stroke recovery is a complex and multidimensional process in which internal (neurobiological and psychological factors) and external factors (specialized medical care) play a role. Unfortunately, it is evident that these endogenous restorative activities using medications, including PDE-5i, are typically insufficient to promote a full stroke recovery in terms of brain structural damage and dysfunction as well as the patient's functional/behavioral abilities.

CONCLUSIONS
The use of PDE-5i in experimental stroke animal models has demonstrated some advantages in anatomical and functional recovery. There was no improvement in clinical or functional recovery in persons with acute stroke. PDE-5i should be administered with caution in persons with chronic stroke since a drop in regional blood flow was observed. The disparity in the intricacy of stroke pathophysiology between human and animal models was crucial. Further research in a larger population with more consistent stroke onset, medicines, and doses is required to reach a more conclusive result.