Recycled Translation: Repurposing Drugs for Stroke

Abstract

Stroke, which continues to be a leading cause of death and long-term disability worldwide, has often been described as a clinical graveyard. While multiple small molecule therapeutics have undergone clinical trials in stroke, currently only one Food and Drug Administration (FDA)-approved medication exists for the treatment of stroke, the biological, recombinant tissue plasminogen activator (rt-PA). Repurposing of therapeutics which have previously gained FDA approval for alternative indications serves as a prospective option for stroke therapeutic translation. In contrast to de novo drug development, repurposing strategies have patient-centered and economic advantages. These include increased safety, increased chance of approval, decreased time to approval, and decreased capital investment. Presently, 37 active stroke clinical trials utilize repurposed therapeutics with various initial indications and dosing paradigms. The currently studied repurposed therapeutics fall into six mechanistic categories: (1) anticoagulation; (2) vasculature integrity, response, or red blood cell (RBC) alterations; (3) immune system regulation; (4) neurotransmission; and (5) neuroprotection. Directed hypothesis-driven computational investigation utilizing drug databases, in silico drug-protein interaction modeling, genomic data, and consensus methodology can determine if the current mechanistic repurposing categories have the highest chance of translational success or if other mechanistic avenues should be explored. With this increased focus on repurposed therapeutic strategies over de novo strategies, evolution and optimization of regulatory protections are needed to incentivize innovators and investigators.

History of Therapeutic Translation and Clinical Trial Failure in Stroke

Stroke remains a leading cause of death and long-term disability worldwide, affecting 12.2 million people each year and projected to affect an additional 3.4 million U.S. adults by 2030 [1, 2]. Not only will this 20.5% increase in prevalence lead to increased morbidity and mortality from this devastating disease, but it will also incur a substantial financial burden to the United States health care system. Currently, stroke accounts for $49.8 billion in direct and indirect health care costs in the USA and will continue to rise in tandem with incidence [2].

While this monumental human and financial toll continues to increase, so has the timeline and cost of therapeutic development for stroke. Despite increased understanding of stroke’s pathogenesis and prognostic indicators, therapeutic development for this disease has been fraught with translational failure and often described as a clinical graveyard [3]. As endovascular device trials have recently found success in improving patient outcomes following ischemic stroke, small molecule and cytoprotective therapeutic trials have not [4–6]. As it stands, only one Food and Drug Administration (FDA)-approved medication exists for the treatment of stroke, a biological, recombinant tissue plasminogen activator (rt-PA) [7].

Advantages of Drug Repurposing

Drug repurposing has become an attractive option to help overcome the gap between the bench and the bedside in various diseases such as cancer [8, 9], Alzheimer’s disease (AD) [10], and Parkinson’s disease (PD) [11, 12]. It could also prove a valuable option for the translation of effective small molecule therapeutics in stroke. Rather than focusing on de novo drug development, this approach to pharmaceutical translation centers on assessing the efficacy of FDA-approved therapeutics outside of currently approved uses. This strategy has successfully brought many therapeutics to market for conditions outside of the initial clinical indications. One example is the serotonin and norepinephrine reuptake inhibitor (SNRI) duloxetine. While duloxetine was initially FDA-approved for major depression and diabetic neuropathic pain, it has been repurposed for a variety of other conditions, gaining FDA approval for generalized anxiety disorder, fibromyalgia, and musculoskeletal pain. Additionally, it also has gained European regulatory agency approval for stress urinary incontinence [13, 14]. In contrast to traditional drug development, this repurposing approach, as with duloxetine, has patient-centered and economic advantages. Namely, decreased capital investment [15], decreased time to approval [15, 16], increased confidence in therapeutic safety for patients, and increased chance of approval [15] (Fig. 1).

Fig. 1. De novo drug development vs. repurposing pipeline.

Timeline demonstrating the overlap of de novo central nervous system (CNS) drug development timelines, cost, and probability of success with repurposed CNS therapeutics. While de novo drug development takes an average of 14 years and an estimated $60 million, repurposed therapeutics, which can enter in at phase 2, average less than eight years for completion

Regarding capital investment, based on 2019 data, the pre-launch research and development (R&D) costs of a new molecular entity (NME) or new biological entity (NBE) are estimated to range from $161 million to $4.54 billion [17]. Alternatively, studies utilizing data over larger time frames have found more narrow estimates of $2–3 billion per new compound [18, 19]. Some of the variability in these estimates is attributed to differences in average costs within a disease- or condition-specific area. For example, anti-cancer drugs were found to have the highest average in R&D estimates ranging from $944 million to 4.45 billion costs per therapeutic [17]. Similarly, central nervous system (CNS) disease R&D expenditures represent a significant portion of the U.S. total expenditures or an estimated 11.9% of the $21.7 billion, second only to cancer therapeutics at 25.7% [20]. When considering per-stage costs, one study estimated that the average cost for a CNS therapeutic is $3.9 million for phase 1, $13.9 million for phase 2, $19.2 million for phase 3, $2 million for review, and $14.1 million for phase 4. This totals a staggering $53.1 million per drug [21], not even including preclinical characterization studies, which for stroke therapeutics, are estimated to range from $52,000–$135,000 per compound [22].

Although the exact estimates for these per-phase or per-disease clinical trial costs may vary, most experts agree these costs are rising at a rate outpacing that of inflation. Drug repurposing serves as a response to combat rising clinical trial costs without compromising on safety or efficacy. Studies have estimated repurposed trials to cost an average of $40–80 million, a fraction of some de novo trials [23]. While the exact per-phase and per-disease breakdown of these repurposed trial costs are yet to be precisely examined for CNS therapeutics, on the surface, they do present an apparent cost advantage compared to traditional de novo testing.

One of the most significant contributors to the lower cost of repurposed trials is the decreased time to trial completion. De novo drugs usually take 10–17 years per medication [24] to traverse the clinical trial process, whereas repurposed drugs average 3–12 years for a new indication [8, 23]. This time advantage is owed to the fact that repurposed drugs have already completed the 4–8 year preclinical in vivo and in vitro basic safety and efficacy testing and can enter the trial pipeline at a later phase (Fig. 1) [8]. Depending on the intended drug targets and drug dosing information, pharmaceuticals that have previously been FDA approved may enter the trial pipeline at phase 2, as they have already completed early phase 1 safety and dosage testing [8, 25]. This facet also captures one of the most considerable patient-centered benefits of utilizing repurposed medications or increased confidence in therapeutic safety. Clinical trial participants and future patients are arguably exposed to decreased risk with repurposed therapeutics at a similar dosage scheme compared to NMEs as the safety of these entities has previously been assessed.

Lastly, the chance of successful approval and market distribution is another facet in which repurposed drugs have an advantage over traditional de novo NME or NBE approaches. An analysis of 21,143 compounds from 2000 to 2015 found that only 13.8% of all drug development programs eventually led to approval. When analyzed by the therapeutic group, 15.0% of CNS drugs led to approval, the second-lowest behind oncology [26]. This high attrition rate of novel therapeutics through clinical trial processes has been cited as a critical component of the current therapeutic productivity crisis in the USA [27, 28]. Utilizing repurposed drugs is a key to combating this productivity crisis through engineering a greater chance of successful trial completion. A 2004 study found that repurposed drugs have a phase 1 to market rate of approximately 30%, almost 2–3 times higher de novo drug success rates depending on the targeted disease [15]. More work is needed, however, to determine if this approval ratio remains accurate in today’s clinical trial climate and what the disease-specific variations of this rate are.

History of Drug Repurposing in Stroke

While the advantages of drug repurposing are apparent, a closer investigation of the history and applicability in stroke is warranted. Repurposed pharmaceuticals of varying mechanisms have been investigated in stroke clinical trials, including but not limited to antimicrobial agents, such as minocycline [29, 30], serum globular proteins, such as albumin [31], membrane stabilizers, such as citicoline [32, 33], calcium channel inhibitors, such as magnesium sulfate [34], and direct erythropoiesis stimulators, such as erythropoietin (EPO) [35]. As with the majority of historically investigated repurposed therapeutics, most of these repurposed therapeutics in stroke were further examined following fortuitous observations during experiments or clinical testing. For example, two famously successfully repurposed pharmaceuticals, sildenafil, and minoxidil were initially investigated as potential anginal and antihypertensive agents. However, during clinical trials, side effects were serendipitously noted, which revealed their potential for erectile dysfunction and hair loss, respectively [13, 36, 37]. Similarly, too was minocycline identified as a potential stroke therapeutics based on unexpected early observations. Namely, while it was initially utilized for its anti-inflammatory effects, the research uncovered both antiapoptotic and neuroprotective effects carried through multiple pathways. This not only accrued interest in minocycline as a neuroprotective and vasculoprotective agent in stroke but also in a wide variety of degenerative CNS conditions such as amyotrophic lateral sclerosis (ALS) and PD [38].

While fortuitous side effect discovery early in the drug characterization process has proved a helpful strategy to translate agents in other disease states and conditions, such as with minoxidil, it has proven unfruitful in stroke. Out of the abovementioned repurposed therapeutics which have been assessed in stroke—minocycline, citicoline [32, 33], magnesium sulfate, and EPO—none has yet successfully gained FDA approval for the treatment of stroke. Instead, stroke remains an elusive clinical graveyard for the translation of small molecules. This signifies a need for more diversified and active repurposing strategies for stroke translation [39]. Intentional evidence-based repurposing strategies, such as modeling, data mining, and consensus methodology, are needed to select drugs with maximal chances of successful translation in stroke.

Current Drug Repurposing in Stroke

Assessment of the current state of drug repurposing in stroke is a crucial first step to determining which strategies have been most effective for potential translation. To ascertain the scope and status of ongoing repurposed clinical trials, a set of inclusion criteria was developed. Trials were included in the analysis if they were listed on ClinicalTrials.gov, active (enrolling, recruiting, unknown), utilizing a drug intervention previously approved by the FDA for an indication other than stroke, and in the USA (Fig. 2a–b). All trials either containing drugs that had not previously been FDA approved (including unregulated substances) or which were inactive (completed or terminated) were excluded. Of the approximately 489 active interventional trials for stroke (status other than terminated, completed, withdrawn, or suspended), 47.4% or 37 out of 78 drug trials were with repurposed therapeutics (Table 1). Interestingly, 7 of these 37 trials are utilizing the same repurposed therapeutic, such as apixaban or minocycline, in different combinations and paradigms, revealing the multifaceted utility of repurposed therapeutics as well (Fig. 2c–e).

Fig. 2. Active Stroke Clinical Trials with Repurposed Therapeutics.

Trial status novelty designation comparison between de novo (black) and repurposed trials (pink) (A). Intervention type (drug, biological, and dietary supplemental) proportional comparison between de novo trials (black) and repurposed trials (pink) (B). Proportional comparison of repurposed (pink) and de novo (green) drug trials to other trials (device, etc., black) of all active interventional stroke trials (C). Proportional comparison of all active stroke drug trials (drug, biological, dietary supplemental) between repurposed trials (pink) and de novo trials (black). The proportion of unique repurposed trials (black) and overlapping repurposed drug trials (pink) of all active stroke interventional repurposed trials

Table 1. Active clinical trials with repurposed FDA-approved drugs.

Currently, there are 37 active clinical trials for stroke utilizing drugs previously approved for other clinical indications. The table was generated using ClinicalTrials.gov and Drugs@FDA: FDA-approved drugs databases. Timing of intervention based on trial information as binned into four categories: preemptively (before stroke), acutely (first administration within 48 h of stroke), sub-acutely (within 120 h of stroke), and retroactively (first administration > 120 h after stroke, most often with continuation through rehabilitation)

Therapeutic	FDA-approved use and year	Mechanism of Action	Phase: trial name	Condition	Timing
Immune system regulation
Drug Y-2-Edaravone (Radicava®) + Borneol (d-camphanol)	Edaravone 2017: ALS Borneol Synthetic flavoring substance and adjuvant	Edaravone Unknown/possibly antioxidant Borneol Activates p38-COX-2-PGE2 signaling pathway, represses NF-κB	Phase 1: SAFVAGE trial	• Stroke, acute • Intracranial hemorrhages	Acutely (single dose, in healthy subjects)
Glibenclamide (BIIB093, glyburide, CIRARA)	2002: Diabetes mellitus 2	Increase insulin release through inhibition of ATP- sensitive K-channels in beta-cell plasma membranes	Phase 3: CHARM trial	• Brain edema • Stroke, acute	Sub-acutely (single bolus continuously over 72 h)
Infliximab (Remicade®)	1998: Crohn’s disease 1999: Rheumatoid arthritis 2004: Ankylosing spondylitis 2005: Psoriatic arthritis 2006: Plaque psoriasis. 2005: Ulcerative colitis 2011: Pediatric ulcerative colitis	TNF-a inhibitor	Phase 1 and phase 2: Infliximab therapy for dolichoectatic vertebrobasilar (DVB) Aneurysms	• Aneurysm • Stroke • Vasculitis, tumor necrosis factor-alpha (TNF-α)	Preemptively  (intravenously (I.V.) to patients with DVB aneurysms over 12 months)
Minocycline hydrochloride (Arestin microspheres)	2001: Periodontal disease, reduction of root pocket depth	Inhibition of matrix metalloproteinase-9, poly (ADP-ribose) polymerase, and 30S ribosomal subunit	Phase 2: MASH trial	• Aneurysm, ruptured • Vasospasm, intracranial • Delayed cerebral ischemia • Blood-brain barrier defect	Sub-acutely (up to 4 days post-aneurysmal subarachnoid hemorrhage)
Hydrogen (H2) + minocycline	2001: Periodontal disease, reduction of root pocket depth	Inhibition of matrix metalloproteinase-9, poly (ADP- ribose) polymerase, and 30S ribosomal subunit	Phase 2 and phase 3: H2M trial	• Stroke, ischemic	Acutel (I.V. or per os (P.O.) for 3 (hydrogen) and 5 (minocycline) days post-stroke)
Oxygen nanobubbles (RNS60)	2019 Fast track: AFS	Mitochondrial biogenesis and ↑ TRegs	Phase 2: RESCUE trail	• Stroke, ischemic	Acutely (within 30 min of confirmation of endovascular candidacy—48 h)
Anticoagulation/antiplatelet
Apixaban (Eliquis®)	2012: Prophylaxis for stroke and blood clots in patients with non-valvular atrial fibrillation 2014: Prophylaxis for deep vein thrombosis (DVT) and pulmonary embolism (PE) reduce the risk of a blood clot following knee or hip replacement	Reversible direct inhibition of free and clot-bound factor Xa	Phase 3: ACARDIA trial	• Stroke	Retroactively to prevent reoccurrence (P.O. to patients with “recent embolic stroke ESUS” indefinitely)
Apixaban (Eliquis®)	2012: Prophylaxis for stroke and blood clots in patients with non-valvular atrial fibrillation 2014: Prophylaxis for DVT and PE, reduce risk of a blood clot following knee or hip replacement	Reversible direct inhibition of free and clot-bound factor Xa	Phase 3: ASPIRE trial	• Intracerebral hemorrhage • Atrial fibrillation	Retroactively to prevent reoccurrence (P.O. for a patient with “recent” ICH and high-risk non-valvular AF (CHA2DS2-VASc score ≥ 2) indefinitely)
Argatroban Or eptifibatide (Integrilin)	Argatroban 2000: prophylaxis or treatment of thrombosis in adult patients with heparininduced thrombocytopenia (HIT) Eptifibatide 1998: Acute coronary syndrome (ACS) managed medically or with percutaneous coronary intervention (PCI) and treatment of patients undergoing PCI	Argatroban Direct thrombin inhibitor Eptifibatide GP IIb/IIIa receptor inhibitor	Phase 3: MOST trial	• Acute ischemic stroke	Acutely (I.V. within 60 min but no later than 75 min of IV rt-PA or IV TNK (< 3 h of LNW))
Tirofiban hydrochloride (Aggrastat®)	1999: Acute coronary syndrome	Competitive inhibitor of GP Ilb/IIIa receptors	Phase 1 and phase 2: iTREMPT	• Acute ischemic stroke	Acutely (I.V. within 1 h of mechanical thrombectomy and terminated at 24 h)
Vasculature integrity, response, or RBC alterations
Ferumoxytol	2009: Iron deficiency anemia in adult chronic kidney disease patients	Isolates bioactive iron	Phase 1: Treatment targets for inflamed intracranial atherosclerosis on vessel wall MRI	• Stroke, intracranial • Aneurysm, intracranial • Atherosclerosis	Acutely (single dose as a contrast for participants with either atherosclerosis or known aneurysm)
Hydroxyurea	1998: Sickle cell disease	Inhibitor of S-phase DNA synthesis	Phase 2: HU prevent trial	• Sickle cell disease • Stroke	Preemptively (P.O. for 3 years)
Insulin	1982: Type 1 and type 2 diabetes mellitus	Insulin receptor activation and Gut4 receptor mobilization	Phase 1: Intensive insulin therapy with tight glycemic control to improve outcomes after endovascular therapy for acute ischemic stroke	• Ischemic stroke	Acutely (Start I.V. f < 24 h of stroke for 48 h)
Nicardipine + labetalol + hydralazine	Nicardipine 2008: Hypertension Labetalol 1998: Hypertension Hvdralazine 1997: Hypertension	Nicardipine Inhibition of calcium influx in smooth muscle Labetalol Alpha and beta receptor blockage Hydralazine Inhibition of calcium influx in smooth muscle	Phase 2: BEST-II	• Acute stroke • Endovascular thrombectomy	Acutely (“soon” after endovascular treatment, and administered to goal SBP for 24 h post-intervention)
Nicardipine	2008: hypertension	Inhibition of calcium influx in smooth muscle	Phase 4:1.V. double and triple concentrated nicardipine for stroke and ICH	• Hypertension • Stroke • Intracranial hemorrhage (ICH)	Acutely (patients with AIS or ICH and “need for rapid BP control)
Nitric Oxide (Genosyl®, INOmax®)	1999: Extracorporeal membrane oxygenation in term and near-term (> 34-week gestation) neonates with hypoxic respiratory failure with pulmonary hypertension 2020: Hypoxic respiratory failure with pulmonary hypertension		Phase 2: DOMINO trial	• Stroke	Sub-acutely (within 72 h of stroke onset for 35 min)
Omega-3 polyunsaturated fatty acids (Lovaza®)	2004: Triglycerides > 500 mg/dl	Unknown, possibly inhibition of acyl-CoA: 1,2-diacylglycerol acyltransferase, increased mitochondrial and peroxisomal β-oxidation in the liver, decreased lipogenesis in the liver, and increased plasma lipoprotein lipase activity	Early phase 1: Lovaza’s effect on clopidogrel in a neuro population	• Stroke, ischemic • Transient ischemic attack	Acutely or retroactively (P.O. to AIS or TIA patients daily for 1 month)
Pioglitazone	1999: Type 2 diabetes	Selective stimulation of nuclear receptor peroxisome proliferator-activated receptor gamma (PPAR-γ) and PPAR-α)	Phase 1 and phase 2: Pioglitazone treatment for hyperglycemic acute ischemic stroke	• Stroke, acute • Hyperglycemia • Diabetes	Acutely (within 12 h of stroke P.O. for 3 days)
Recombinant Activated Factor VII (rFVIIa, NovoSeven®)	2005: Bleeding in patients with hemophilia A or B with inhibitors to factor VIII or factor IX. 2014: Glanzmann’s thrombasthenia	Tissue factor-dependent or independent activation of the extrinsic coagulation cascade	Phase 3: FASTEST trial	• Intracerebral hemorrhage	Acutely (within 120 min of stroke onset, I.V. over 2 min)
Recombinant human tissue kallikrein	2009: Acute attacks of hereditary angioedema	↑ bradykinin release	Phase 2 and phase 3: ReM- EDy II trial	• Acute stroke, ischemic • Stroke	Acutely and sub-acutely (within 12 h of stroke, one I.V. dose then subcutaneously biweekly for 22 days)
Tenecteplase (TNKase®)	2000: Treatment of acute MI	Increased conversion of plasminogen to plasmin in the presence of fibrin, recombinant tPA	Phase 3: TWIST trial	Ischemic, stroke Stroke, acute	Acutely (within 4.5 h of waking with stroke)
Verapamil	1998: Rapid conversion to sinus rhythm of paroxysmal supraventricular tachycardias, temporary control of rapid ventricular rate in atrial flutter or atrial fibrillation	Inhibits L-type calcium channels	Phase 1: Verapamil for neuroprotection in stroke	• Ischemic stroke	Acutely (intra-arterial following mechanical thrombectomy)
Neurotransmission
Cyproheptadine	1961: Allergic reaction	H1 and 5-HT antagonist	Not applicable: altering activation patterns post-stroke	• Stroke • Muscle spasticity • Hemiparesis	Retroactively (P.O. for 6 weeks during therapy for stroke survivors in “chronic phase”)
Donepezil (Aricept)	2004: Mild, moderate, and severe AD	Acetylcholinesterase inhibitor	Phase 3: Enhance	• Ischemic stroke	Retroactively (P.O. for 12 weeks for cognitively impaired stroke survivors undergoing inpatient therapy)
Donepezil (Aricept)	2004: Mild, moderate, and severe AD	Acetylcholinesterase inhibitor	Phase 1: Aricept to improve functional tasks in vascular dementia	• Stroke • Vascular dementia • Memory deficits	Retroactively (P.O. for 12 weeks, 4 months–5 years post-stroke)
Escitalopram (Lexapro)	2002: Depression	Selective serotonin reuptake inhibitor (SSRI)	Phase 2: ELISA trial	• Aphasia stroke	Retroactively (P.O. daily for 90 (days within 5 days of stroke onset)
Fluoxetine	1999: Major depressive disorder (age eight and older), obsessive-compulsive disorder, panic disorder, bulimia, binge eating disorder, premenstrual dysphoric disorder, and bipolar depression, as well as treatment-resistant depression when used in combination with olanzapine.	SSRI	Early phase 1: Post-stroke depression in hemorrhagic stroke	• Stroke, hemorrhagic • Depression	Retroactively (P.O. for 1 year for patients admitted for SAH from intracerebral aneurysm)
IncobotulinumtoxinA (Xeomin®, NT 201)	2010: Blepharospasm, cervical dystonia 2015: Upper limb spasticity 2018: Chronic sialorrhea	Inhibition of presynaptic acetylcholine	Phase 4: Xeomin® and gait-related mobility after stroke	• Stroke	Retroactively (intramuscular injection (I.M.) with 4–6 FU in patients diagnosed with hemiparesis and spasticity secondary to stroke)
IncobotulinumtoxinA (Xeomin®, NT 201)	2010: Blepharospasm, cervical dystonia 2015: Upper limb spasticity 2018: Chronic sialorrhea	Inhibition of presynaptic acetylcholine	Phase 3: PATTERN trial	• Lower limb or combined lower limb and upper limb spasticity due to stroke or traumatic brain injury	Retroactively (I.M. injection, 1–5 treatment cycles in with lower limb spasticity due to stroke)
Ketamine	1970: Anti-depressant	NMDA receptor antagonist	Phase 1: Use of transmucosal ketamine in post-stroke depression	• Post-stroke depression	Retroactively (2 intramucosal doses once weekly in post-stroke patients)
Ketamine	Anti-depressant	NMDA receptor antagonist	Phase 2 and phase 3: QUEST-KETA trial	• Acute ischemic stroke	Retroactively (I.V. for 24 h in patients with AIS < 24 from LNW)
Levetiracetam (Keppra)	2000: Focal seizures, myoclonic seizures, and primary generalized seizures	Synaptic vesicle protein SV2A binding	Phase 1: Levetiracetam (Keppra) to improve chronic aphasia in post-stroke patients	• Aphasia, stroke	Retroactively (P.O. twice daily for patients with post-stroke aphasia)
Memantine XR	2002: Moderate to severe Alzheimer’s disease	NMDA receptor antagonist	Early phase 1: Memantine for enhanced stroke recovery	• Ischemic stroke • Upper extremity weakness	Acutely (< 24 h-90 days post-stroke)
Methylphenidate (Concerta®, Ritalin)	2012: ADHD	Inhibition of dopamine active transporter (DAT) and noradrenaline transporter (NAT)	Phase 2: Methylphenidate for PTSD and stroke veterans	PTSD Stroke	Retroactively (“recent stroke,” up to 20mg P.O. twice daily)
Phenylephrine (Neo-Synephrine) Or Norepinephrine (Levophed)	Phenvlephrine 2012: Hypotension Norepinephrine 2021: Severe acute hypotension	Phenvlephrine Agonist of α1-adrenoceptors Norepinephrine Agonist of α1-adrenoceptors	Early phase 1: PRESS trial	• Ischemic stroke • Blood pressure	Acutely (post-thrombectomy, dose to target systolic blood pressure)
Nenroprotection
Fingolimod (Gilenya®)	2010: First-line treatment for relapsing forms of multiple sclerosis (MS) 2018: MS in pediatric patients 10 and up	Activates lymphocyte S1P1 via high-affinity receptor binding, reduces auto aggressive lymphocyte infiltration into the CNS	Early phase 1: FITCH trial	• Intracerebral hemorrhage • Cerebral edema • Stroke • Hemorrhagic • Intracerebral hemorrhage, hypertensive • Intracerebral hemorrhage, intraparenchymal	Acutely (1 P.O. dose < 24 h of stroke)
Intra-arterial cold saline + minocycline (Solodyn) + magnesium sulfate	Minocycline 1971: Inflammatory lesions of non-modular acne vulgaris, certain Gram-positive and Gram-negative bacterial infections Mapnesium sulfate 1986: Magnesium deficiency 2006: Seizure prevention in preeclampsia	Minocycline Inhibition of matrix metalloproteinase-9, poly (ADP-ribose) polymerase, and 30S ribosomal subunit Mapnesium sulfate Cerebral vasodilation	Phase 1: Intra-arterial administration of neuroprotective agents and cold saline at time of recanalization for acute ischemic stroke due to large vessel occlusion	• Stroke, ischemic	Acutely (intraarterially immediately following thrombectomy)

The lack of neuroprotective therapies in combination with mechanical thrombectomy led the NIH to launch the Stroke Preclinical Assessment Network (SPAN), a network of 6 testing centers and a central coordinating center (Lyden P, Stroke in press). These testing centers, comprised of John Hopkins University, Massachusetts General Hospital, Medical College of Georgia, University of Iowa, University of Texas Houston, and Yale University, have been tasked with conducting late-stage preclinical studies of putative neuro-protectants combined with reperfusion. Presently, this network is testing several repurposed drugs for stroke, including tocilizumab, fingolimod, fasudil, and veliparib, with the plan to select the best agent or agents for clinical trials by the NIH StrokeNet.

When considering geographical location and trial status, a majority, or 41% of repurposed trials, are led by institutions in the northeast. This is in direct comparison to the general geographic location of interventional therapeutic stroke trials (drug, dietary supplement, and biologicals), which are concentrated in the South (45%) and have the lowest representation in the northeast (14%). The proportional breakdown of the remaining regions is similar between repurposed and total therapeutic stroke trials. This geographic discrepancy is of interest compared to general stroke incidence and mortality, which historically has been highest in the southeastern states referred to as the “stroke belt” [40].

Ultimately, the overall percentage of active stroke trials utilizing repurposed therapeutics is evidence of the recognized potential utility of this strategy for small molecule therapeutics in stroke. Over time, the progression of these current repurposed trials should be closely followed to better understand which types of therapeutics (targets, type, and pathways) are successful in stroke repurposing. This information can then be used iteratively to plan for future repurposed strategies with potential in stroke translation. For example, the currently studied repurposed stroke therapeutics fit into five mechanistic categories: (1) anticoagulation; (2) vasculature integrity, response, or red blood cell (RBC) alterations; (3) immune system regulation, (4) neurotransmission; and (5) neuroprotection (Fig. 3). These categorical and mechanistic groups are interesting to consider as they reveal within-group trends such as the timing of intervention or target molecules. For example, most trials utilizing therapeutics in the vasculature integrity, response, or RBC alterations mechanistic category ascribe to acute-dosing therapeutic paradigms. In contrast, most trials of therapeutics in the neurotransmission category have retroactive interventional dosing strategies (Table 1). The aforementioned hypothesis-driven strategies, such as retrospective analysis and in silico modeling, can be combined with consensus methodology to assess these trends and determine if these mechanistic categories are appropriate and advantageous avenues for continued exploration of repurposed therapeutics in stroke. Alternatively, these purposeful strategies and consensus methodology may identify new mechanistic categories with higher likelihoods of successful clinical translation. A more extensive discussion of these strategies is warranted.

Fig. 3. Mechanistic Categories of Active Stroke Interventional Repurposed Trials.

Mechanistic grouping and proportions of active interventional repurposed drug stroke trials

Employable Strategies for Selecting Therapeutics with Repurposing Potential

Analysis of current clinical trial progression will be an essential tool for planning future repurposing trials. There are, however, employable strategies that can be used to engineer higher chances of successful repurposed therapeutic selection outside of reliance on fortuitously recognized side effects [41]. These hypothesis-driven strategies fit into three overarching categories, data mining (retrospective analysis, drug-focused databases, disease-focused databases), experimental (high throughput in vitro and in vivo studies, binding assays, clinical trials), and computational/in silico (signature matching, molecular docking/dynamics, pathway mapping, and genetic association) [42]. The approaches can be used in isolation or sequentially by investigators to identify individual drugs with repurposing potential (DRP). Collectively, they can be utilized by expert teams to create guidelines or frameworks of DRP in stroke, such as with the Delphi Consensus Methodology [33].

Data Mining

When looking to repurposed strategies in other disease states, the often-cited first approach is data mining to identify compound classes or mechanisms of action with potential efficacy in the disease of interest. Data mining presents an alternative, low-cost, and viable option for identifying compounds with repurposing potential. In the new big data environment, there exist large databases which fall into three categories for data mining purposes: drug-focused databases, disease-focused databases, and drug-disease linkage databases. Drug-focused databases provide information on gene expression profiles and known targets of cataloged therapeutics [42]. For example, the drug repurposing hub, Clue Repurposing, hosted by the Broad Institute, is a next-generation drug repurposing library that contains information on the chemical structure, mechanism of action, targets, disease areas, indication, purity, phase, and vendor ID of over 6500 previously approved drugs [43, 44]. Sequentially, scouting through other compound databases or small molecular repositories, such as the FDA-drug approved library, Drugs@FDA, which is updated daily and includes the majority of FDA-approved compounds since 1939 [45], and the TEVA screening set, which includes FDA and foreign agency approved compounds [42], could assist in the generation of a list of compounds with repurposing potential in stroke. While compounds approved by other foreign agencies may not provide some of the previously discussed advantages of repurposed therapeutics, such as quick clinical trial time and lower cost, they do have the advantage of a reasonable expectation of safety and efficacy depending on the stringency of the approving agency.

Computational Approaches

Another often utilized approach to identify DRP is computational modeling. Recently, many omics-level tools have been developed to predict on-target and off-target effects of DRP [46]. For example, tools such as PREDICT [47], RANKS [48], the multiscale interactome [49], and DSigDB [50] can identify drug-drug, drug-disease, and drug-gene interactions. Stroke researchers can utilize these tools to identify drugs that are most similar to compounds that have already found high levels of success in stroke trials and avoid drugs that have had unpredicted or catastrophic off-target effects. Therefore, these computational means can increase the safety and efficacy of resulting clinical trials, ultimately increasing the chance of successful approval and translation.

Once a stroke researcher has a specific target or therapeutic in mind, molecular docking and drug-protein interaction tools [51] such as DR.PRODIS [52], GIFT [53], DrumPID [54], PreBINDS [55], and PROMISCUOUS 2.0 [56, 57] can be utilized to evaluate input molecules against target libraries or visa versa. These in silico modeling platforms allow researchers to bypass timely and costly animal models, many of which have been criticized for their questionable translational applicability in stroke drug discovery. In this way, researchers can mimic and analyze direct interactions of their target molecule with their target receptors and pathways without species-dependent heterogeneity producing variability in expected outcomes and effects. With even more precision, potential off-target effects can be better predicted with pathway mapping tools such as MANTRA [58, 59], which utilizes network theory and non-parametric statistics of expression data to identify alternative mechanisms of action. When coupled with genetic association data utilizing tools such as DrugSig [60], ksRepo [61], COGENA [62], and NFFinder [63], which integrate gene expression and co-expression and drug datasets from different platforms, the chance of unexpected off-target and on-target effects can be drastically decreased compared de novo experiments.

Experimental Approaches

Experimental approaches are most similar to traditional therapeutic identification processes. High throughput in vitro and in vivo studies can be utilized to identify DRP either from large existing compound libraries, in a non-hypothesis-driven manner, or from a shortlist of identified DRPs from previous data mining and computational approaches. In vitro approaches include binding and activation assays, and in vivo approaches include animal disease models and human clinical trials. The downside of these approaches, when utilized in a non-hypothesis-driven manner, is a lower chance of eventual clinical trial success and a higher chance of unexpected off-target effects per compound.

Collective Consensus Approaches

Ultimately, investigators and invested parties can utilize multiple approaches sequentially or in parallel to identify potential molecules or compounds of interest, assess potential efficacy in silico, and predict potential off-target and side effects based on aggregate patient data. One collective and the sequential hypothesis-driven process is the Delphi Consensus Strategy, which has been utilized to identify DRPs in other CNS diseases, such as Alzheimer’s disease (AD). In AD, an expert panel utilized the Delphi Consensus process to identify five compound classes with potential for drug repurposing, including tetracycline antibiotics, calcium channel blockers, angiotensin receptor blockers (ARBs), glucagon-like peptide 1 (GLP1) analogs, and retinoid therapy [64]. Since that consensus, compounds in all classes have been assessed in clinical trials. Similarly, directed consensus processes can be utilized in stroke to determine classes of compounds or potential mechanisms which should be the next therapeutic targets for clinical investigation.

A careful comparison of the current mechanistic groupings of repurposed therapeutics in stroke (anticoagulation, vasculature integrity, response, or RBC alterations, immune system regulation, neurotransmission, and neuroprotection) to the groupings of de novo drugs investigated in the past is warranted to ascertain fundamental differences in de novo idealistic approaches and repurposed approaches for stroke. Then, the aforementioned hypothesis-driven strategies, such as data mining, retrospective analysis, and in silico modeling, can be combined to determine if these mechanistic categories are appropriate and advantageous avenues for continued exploration of repurposed therapeutics in stroke. From this comparison, an expert panel of stroke physicians, scientists, and industry representatives can outline a consensus of mechanistic areas with the highest potential for trial success and improved patient outcomes as was done in AD. Once these agreed-upon broad categories have been outlined, individual investigators can utilize compound and disease databases to identify DRP which fit into these consensus mechanistic categories to engineer greater chances of translational success. This iterative and cyclical process can be revisited regularly to adjust or reaffirm consensus categories.

Regulatory and Generalizability Considerations

It is important to discuss potential pitfalls or downsides of repurposed therapeutics, one of which being regulatory affairs and intellectual property considerations. Out of three regulatory approval pathways offered by the FDA, repurposed drugs are only eligible to utilize one, Section 505 (b)(2) [65]. From 2009 to 2015, this specific section was utilized in more than 50% of approved therapeutics compared to novel drug pathways [66, 67]. Explicitly, the FDA guidance report dictates that this section is reserved for “1) an application that contains full reports of investigations of safety and effectiveness (section 505(b)(1)); (2) an application that contains full reports of investigations of safety and effectiveness but where at least some of the information required for approval comes from studies not conducted by or for the applicant and for which the applicant has not obtained a right of reference (section 505(b) (2)); and (3) an application that contains information to show that the proposed product is an identical active ingredient, dosage form, strength, route of administration, labeling, quality, performance characteristics, and intended use, among other things, to a previously approved product (section 505(j))” [68]. Therefore, for a compound to take advantage of the safer, speedier, and cheaper repurposed therapeutic pathway, it must have no alterations from the parent previously approved compound. These stipulations limit the exclusivity protections and related benefits new investigators can hope to garner from repurposed therapeutics if approved compared to de novo therapeutics.

To incentivize investigators to conduct these repurposed studies, indication-specific market exclusivities have been generated. However, despite these indication-specific exclusivities, two-thirds of FDA-approved drugs does not officially receive new indications. Furthermore, this chance of receiving a new indication drops to near zero after generic entry [69]. This illustrates that more appropriate and utilized protections are needed to incentivize innovators and investigators to continue looking to repurposed therapeutics rather than focusing solely on novel technologies.

Limitations of translational efficacy and safety of repurposed therapeutics due to sex and age of clinical trial participants should also be discussed. While the repurposed therapeutics discussed in this review have previously achieved FDA approval, the conditions under which they gained their original approval may be divergent from current standards. For example, an FDA guidance report released in 1977 calls for the exclusion of women of childbearing potential from participation in phase I and early phase II clinical trials [70], which is approximately 75.4 million women in the USA [71]. Furthermore, stipulations for subgroup analysis of clinical trial participants, such as pediatric, geriatric, renal failure patients, were not introduced until an FDA requirement in 1985 [70]. Race and gender were not included as a required subgroup for analysis until 1988, and animals of both sexes were not required in preclinical drug safety testing until 1987 [70]. Since that time, there have been a series of FDA trial guidelines encouraging further diversity in regard to gender and race of clinical trial participants, as studies have revealed complications arise in these populations excluded from early safety trials [70, 72]. Pregnant women, however, are mostly still excluded from clinical trial participation [73]. Therefore, many of the repurposed therapeutics currently undergoing clinical trials for stroke may have been originally approved in trials without sufficient sex, race, and comorbidity representation. Therefore, while repurposed therapeutics are generally considered to be safer than de novo therapeutics, as they have previously undergone human testing without significant complications, safety confidence may be decreased in therapeutics which were not tested on diverse patient populations. With the continued push of more inclusive FDA guidelines encouraging diverse clinical trial patient populations, this potential limitation of repurposed therapeutics should decrease with time.

Summary

Ultimately, hypothesis- and data-driven strategies can accelerate the prioritization of therapeutics with the greatest potential for repurposing in stroke. The repurposed therapeutic pathway provides many economic, safety, and timely advantages over traditionally de novo drug development pathways and has already come to be utilized in stroke as well as many other CNS diseases such as AD. This pathway continues to have the potential for robust growth, although better defined indication-specific protections as well as understanding of the generalizability of early approval trails to larger patient populations are needed to encourage widespread utilization.

Data Availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.