Intra-arterial neuroprotective therapy as an adjunct to endovascular intervention in acute ischemic stroke: A review of the literature and future directions

Abstract

Mechanical thrombectomy for acute ischemic stroke due to large vessel occlusion has been shown to significantly improve outcomes. However, despite efficient rates of recanalization (60–90%), the rates of functional independence remain suboptimal (14–58%), most likely due to pathways of cell death in the brain that have already committed despite successful reperfusion. Pharmacologic neuroprotection provides a potential means of preventing this inevitable damage through targeting excitotoxicity, reactive oxygen species, cellular apoptosis, and inflammation. Numerous clinical trials using various neuroprotective agents have failed, but the majority of these trials did not include endovascular reperfusion, and thus the drugs were not reaching the therapeutic target. Intra-arterial delivery of neuroprotective agents via the guide catheter already in place for mechanical thrombectomy could provide a way to deliver high doses directly to the affected territory while limiting systemic exposure. Agents that have shown promise via the intra-arterial route in preclinical as well as some clinical models include magnesium sulfate, verapamil, cold saline, stem cells, and various combined approaches. Targeted hypothermia, achieved with intra-carotid infusion of cold saline, may provide an effective means of achieving hypothermia of the ischemic tissue while avoiding the systemic effects that have limited its use previously. Combination therapy of targeted hypothermia and a cocktail of drugs that provide anti-excitotoxic, anti-oxidant, anti-apopototic, and anti-inflammatory effects may provide an ideal approach that deserves further study in clinical trials.

Introduction

Stroke is the fifth leading causing of death and the leading cause of long-term disability in the United States, affecting over 800,000 patients per year.¹,² In 2015, several randomized controlled clinical trials demonstrated significantly improved clinical outcomes in patients who underwent endovascular intervention for emergent large vessel occlusion (LVO).^3–6 While the results of these trials were encouraging, despite best interventional practice, there is a mismatch between technical success with thrombectomy and clinical outcomes. Despite efficient rates of recanalization (60–90%), the rates of functional independence remain suboptimal (14–58%).⁷,⁸

This discrepancy is most likely due to the fact that even with quick and effective reperfusion to the occluded territory, many patients still experience stroke because the tissue is either already infarcted or committed down a path of cell death. This presents an opportunity for adjunctive therapy such as pharmacologic neuroprotection to prevent or reverse this devastating cascade. It is generally accepted that there are four primary mechanisms of ischemic injury that can be targeted with drug therapy.^9–11

• Excitotoxicity: During ischemia, the lack of adenosine triphosphate (ATP) leads to breakdown of the sodium/potassium (Na/K) transporter and thus cell depolarization and a surge in intracellular calcium. There is accumulation of glutamate and activation of N-methyl-D-aspartate (NMDA) and alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors. Drugs that target calcium channels, therefore, to prevent this surge in intracellular calcium, and drugs that block NMDA and AMPA receptors have shown promise in neuroprotection.

• Reactive oxygen species: Oxidative and nitrosative stress and loss of energy-dependent scavenger enzymes lead to build up of damaging free radicals. Natural free radical scavengers and antioxidants have thus shown some promise in neuroprotection.

• Cellular apoptosis: Signaling cascades are initiated due to the ischemic conditions that lead to downstream messengers that carry out programmed cell death mechanisms. Anti-apoptotic drugs including caspase inhibitors have shown some promise in inhibiting this cascade.

• Inflammation: Inflammatory cytokines and chemokines attract microglia and other infiltrating immune cells which disrupt the blood brain barrier and lead to cytotoxic edema and gliosis. Complement is deposited. Anti-inflammatory drugs and immune suppressants have thus shown some promise in neuroprotection particularly over the subacute to chronic phase.

The optimal strategy for pharmacologic neuroprotection is likely a combination of drugs that target all of these pathways. There are hundreds of drugs that have shown effectiveness in animal models of stroke, but nearly all have been disappointing failures in clinical trials. There are several likely reasons for this failure of translation: (1) animal models of stroke have poor translatability to humans; (2) ischemic injury is a multi-faceted cascade and treating with only one drug that targets only a select portion of the cascade is ineffective; (3) the vast majority of patients in the clinical trials did not receive endovascular intervention, and thus the drug is not reaching its therapeutic target.

The latter point, taken with the findings of the recent clinical trials demonstrating clinical benefit of endovascular thrombectomy, suggests that many of the promising neuroprotective agents deserve to be reevaluated in trials that combine them with endovascular reperfusion.

The case for intra-arterial delivery of neuroprotective agents

The endovascular technique in stroke intervention provides direct catheter access to the vessel(s) harboring the ischemic territory. Thus, without any additional manipulation, the potential for infusion of drugs directly into the anterior or posterior circulation is available. In addition to convenience, there are several potential advantages to delivering neuroprotective therapy via the intra-arterial route.

In cancer therapy, the pharmacokinetic advantage of administering drugs through an intra-arterial (IA) route is that the first pass of the drug is through the tumor bed, thus avoiding the first-pass systemic metabolism of orally or intravenously delivered chemotherapeutic agents. The concentration of chemotherapy in the tumor is thus very high, and there is less systemic exposure and toxicity. Comparative pharmacokinetics of gemcitabine during IA and intravenous (IV) delivery in patients with unresectable pancreatic cancer showed significantly lower plasma concentrations of gemcitabine and its deaminated metabolite in the systemic circulation after IA delivery, suggestive of a major advantage of IA delivery in terms of reduced systemic exposure and toxicity.¹²

A prominent example of the effectiveness of intra-arterial chemotherapy is in the treatment of liver malignancies, where it is used as first-line treatment for unresectable hepatic metastases from colorectal cancer, as neoadjuvant therapy prior to orthotropic liver transplantation, and as palliative treatment for unresectable hepatocellular carcinoma.^13–15 IA chemotherapy enables chemotherapeutic drug concentrations within liver tumors that are up to 100 times greater than those achievable with systemic chemotherapy.^16–18

In the central nervous system, IA delivery of technetium 99m hexylmethylpropylene amineoxine into human cerebral arteries achieved 50 times higher concentration in brain tissue compared to intravenous injection of the same substance.¹⁹ The use of IA chemotherapy has dramatically improved outcomes in intraocular retinoblastoma. In a recent review of nine years of experience at one institution including 226 eyes in children with advanced stage retinoblastoma, ocular survival at five years was 70.2%, compared to previous reports in the literature of 25–30% with systemic chemotherapy.²⁰

The intra-arterial delivery of verapamil is used commonly in subarachnoid hemorrhage for symptomatic cerebral vasospasm that has progressed despite medical and systemic therapies.²¹,²² For this indication, 5–10 mg of verapamil in 10–20 ml normal saline is infused over 10–20 min directly into either the internal carotid artery or selectively into affected intracranial branches such as the anterior cerebral artery, middle cerebral artery (MCA), or basilar artery. This procedure is well tolerated with a very low complication rate.²¹,²²

Therefore, in stroke cases, providing neuroprotective therapy via the IA route could provide high concentration of drug directly to the target ischemic territory while avoiding first-pass metabolism and limiting systemic toxicity. As in chemotherapy and vasospasm treatment, this has the potential to significantly improve the effectiveness of therapy compared to intravenous treatment.

Intra-arterial neuroprotective therapy in stroke

Only a fraction of the neuroprotective agents tested via the intravenous, oral, or intraperitoneal route have been tested via intra-arterial delivery. Here, we provide a review of the substances that have been used via IA delivery in either animal models of acute ischemic stroke or human clinical trials. They are also summarized in Table 1 and Table 2, respectively.

Table 1.

Pre-clinical in vivo models of large vessel occlusion acute ischemic stroke using intra-arterial neuroprotective agents. The findings of each study are briefly described.

Study	Neuroprotective agent	Model	IA delivery	Findings
Marinov et al.26	Mg	Rat MCAo	90 mg/kg, 30 mg/kg	Reduction in infarct volume 43.1% and 59.8%, dose-dependent
Lee et al.27	Mg	Rat MCAo	750 µmol/kg	After 72 h, SSEPs improved 23–39%; reduced infarct sizes 36–42%; neurobehavioral outcomes improved 24–34%
Maniskas et al.29	Verapamil	Mouse MCAo	0.15 mg/kg in 10 µl saline at 2.5 µl/min	Reduced infarct size from 17.46 to 0.64 mm3, improved neurobehavioral outcome
Chen et al.40	Cold saline	Rat MCAo	0°C saline 2.5 ml over 10 min at 0.25 ml/min	Reduction in brain temp to 30.7–30.9°C, reduced infarct volume (61.7 to 40.6%), reduced cerebral edema (15.4 to 7.8%)
Mattingly et al.42	Cold saline	Swine tandem CCA and MCA occlusion	3 h cold saline infusion after reperfusion	Reduction in ipsilateral brain temp to 26°C, reduced infarct volume (0.050 to 0.005 on MRI, by 58% on histology)
Kamiya et al.45	Autologous bone marrow mononuclear cells	Rat MCAo	1 × 107 cells in 1 ml PBS over 5 min	Reduction in infarct volume (8.19 ± 9.65 mm3) compared to controls (85.6 ± 34.15 mm3), and improved motor function scores
Oh et al. 201546	Human adipose- derived mesenchymal stromal cells	Rat MCAo	5 × 105 cells in 5 µl over 5 min	Reduction in infarct size (25.2% ± 2.9%) compared to controls (35.6% ± 3.3%). Effects are through a paracrine mechanism, as cell replacement effects were limited.
Huang et al.47	Adipose-derived stem cells	Rat MCAo	2 × 106 cells in 50 µl over 10 min	Improvement in neurological recovery at days 7 and 14, regeneration of neuronal fibers and blood vessels in peri-infarct cortex
Khan et al.52	Norcatharidin (MMP-9 inhibitor)	Rat MCAo	0.5 mg/kg in DMSO, 0.10 ml/min over 5–7 min	Reduction in infarct volumes (148.1 ± 10.2 vs. 178.5 ± 9.9 mm3) and improvement in neurologic assessment scales
Wang et al.53	Erythropoietin	Rat MCAo	800 U/kg at 0.1 ml/min over 5 min	Reduction in infarct volume (36.2%) and edema (31.8%), better performance on neurologic testing
Song et al.54	Cold saline + Mg	Rat MCAo	120 mg/kg MgSO4 in 15°C or 37°C saline at 0.4 ml/min over 20 min	IA cold Mg had largest reduction in infarct volume by 65%, improvement in neurologic outcome
Chen et al.55	Cold saline + albumin	Rat MCAo	0.5 g/kg or 1.5 g/kg human albumin in 0°C or 37°C saline	IA cold albumin had largest reduction in infarct volume (67%) and improvement in neurologic function assessments

IA: intra-arterial; MCAo: Middle cerebral artery occlusion model; SSEPs: Somatosensory evoked potentials; CCA: Common carotid artery; MRI: Magnetic resonance imaging; MMP-9: matrix metalloproteinase-9.

Table 2.

Human clinical trials using intra-arterial delivery of neuroprotective agents during mechanical thrombectomy procedure for acute ischemic stroke due to large vessel occlusion.

Clinical Trial	Neuroprotective agent	IA delivery	Primary outcome	Findings
Mack, Saver et al.	Mg	0.75–1.5 g	Serum Mg	Recruiting
Fraser et al.30 (SAVER-I)	Verapamil	10 mg in 20 ml saline over 20 min	Presence of ICH on 24 h scan	IA verapamil following thrombectomy is safe and feasible
Choi et al.43	Cold saline	15°C and 7°C saline in ICA at 33 ml/min for 10 min	Adverse events	Jugular bulb temps decreased 0.84°C compared to 0.15°C bladder temps. 3/18 patients experienced shivering, no other adverse events, change in hemodynamics or lab values
Chen et al.44	Cold saline	4°C saline at 30 ml/min for 10 min	Adverse events	No significant change in vitals, electrolytes or hematocrit in 26 patients undergoing thrombectomy for stroke
Battistella et al.49	Autologous bone marrow mononuclear cells	1–5 × 108 cells in 10 ml over 10 min into M1	Adverse events	IA bone marrow mononuclear cell transplantation is safe and feasible in MCA non-acute ischemic strokes
Friedrich et al.50	Autologous bone marrow mononuclear cells	15 ml over 30 min	Adverse events	No adverse events, 40% good clinical outcome, 30% satisfactory clinical improvement in 90 days
Fraser et al.30	Verapamil and Mg	10 mg verapamil in 10 ml saline, 1g MgSO4 in 20 ml saline over 20 min	Number of patients with symptomatic ICH	Recruiting

IA: intra-arterial; ICH: Intracerebral hemorrhage; ICA: internal carotid artery.

Magnesium sulfate

In the brain, magnesium ions provide a physiologic noncompetitive voltage-dependent block of the NMDA receptor ion channel.²³ Additional neuroprotective effects include promoting vasodilation and thus increasing cerebral blood flow, inhibiting presynaptic release of excitatory neurotransmitters, blocking other voltage-gated calcium channels, antagonizing endothelin-1 and other vasoconstrictors, and replenishing an ischemia-induced Mg-deficient state.²⁴ Multiple in vitro models have shown that increasing extracellular magnesium concentration leads to enhanced NMDA receptor blockade and thus less intracellular calcium.²³ Magnesium sulfate (MgSO₄) has been shown to be neuroprotective in multiple preclinical models of cerebral and spinal cord ischemia, excitotoxic injury, and head trauma.²⁴ It has been shown to be efficacious in more than eight rodent models of both reversible and permanent focal cerebral ischemia, with decreases in infarct size ranging from 26 to 61%.²⁴

In addition to its promising neuroprotective properties, magnesium is inexpensive, readily available, easy to administer and has a long established safety and tolerability profile in myocardial infarction and eclampsia. There have been suggestions of efficacy based on human clinical studies in newborn hypoxic-ischemic injury, global cerebral ischemia after cardiac arrest, and ischemic injury during cardiac and cerebrovascular surgical procedures.²⁴ A prominent stroke trial utilizing magnesium was the FAST-MAG trial, in which a 4 g loading dose of MgSO₄ was administered in the field to stroke patients by paramedics followed by a 16 g maintenance infusion in the hospital.²⁵ The 4 g loading dose was administered in 54 ml normal saline over 15 min. This trial demonstrated feasibility with a mean serum Mg level of 3.6 mEq/L, and safety with no observed drug-related major adverse events. There was no demonstration of efficacy, but only 35.9% of ischemic stroke patients received tissue plasminogen activator (tPA), and the percentage that underwent thrombectomy was not reported.²⁵

An early study by Marinov et al. showed that intra-carotid delivery of MgSO₄ prior to MCA occlusion was protective compared to controls, and the degree of protection was dose dependent.²⁶ Rats given 90 mg/kg MgSO₄ prior to 1.5 h of ischemia showed a 59.8% reduction in infarct volume compared to controls (p < 0.001), and a 43.1% reduction compared to rats given 30 mg/kg (p < 0.001).²⁶ Lee et al. subjected adult rats to 90 min of MCA occlusion followed by reperfusion and intra-carotid administration of either 750 µmol/kg of MgSO₄ or normal saline as a control.²⁷ After 72 h, amplitudes of somatosensory-evoked potentials from ischemic fore- and hindpaw cortical fields were significantly improved by 23% (p < 0.005) and 39% (p < 0.001), respectively, in rats treated with Mg compared to controls. Furthermore, Mg-treated rats had reduced cortical (p < 0.05) and striatal (p < 0.05) infarct sizes by 42% and 36%, and sensory and motor neurobehavioral outcomes were improved by 34% (p < 0.01) and 24% (p < 0.05), respectively.²⁷

There is one currently ongoing phase I human clinical trial in which the investigators aim to deliver IA magnesium sulfate in patients undergoing endovascular stroke intervention (NCT01502761).²⁸ The study design includes four groups of five patients. The first two groups will receive regional delivery of Mg via the guide catheter in the common carotid artery (CCA) or internal carotid artery (ICA) with total dosage of 0.75 g and 1.5 g. The second two groups will be a combination of regional and distal Mg (some delivered via the guide catheter and some via a microcatheter selectively into the previously occluded vessel), with total dosage of 1.5 g split 75/25% and 50/50%. Primary outcome is serum magnesium concentration, and secondary outcome is the number of patients with serious procedure-related adverse events.²⁸

Verapamil

Similar to magnesium, verapamil, an L-type calcium channel blocker, is thought to have neuroprotective effects by reducing calcium influx into cells and thus limiting excitotoxicity. As mentioned above, there is extensive experience in the literature of administering verapamil via IA delivery into the ICA for cerebral vasospasm, making it an attractive potential agent due to its established safety profile.²¹,²²

Maniskas et al. performed a recent experiment including three groups of mice: no intervention, MCAo surgery followed by IA infusion of 10 µl saline at 2.5 µl/min into the ICA, and MCAo surgery followed by IA infusion of 10 µl saline + 0.15 mg/kg verapamil at 2.5 µl/min.²⁹ The authors reported significant improvement in mice treated with IA verapamil in behavioral testing via forced motor (rotor rod) movement and free roam motor (open field) movement, and a significant reduction in mean infarct volume measured on post-stroke day 7 from 17.46 to 0.64 mm³ (p < 0.001). They also found no significant change in heart rate and pulse distention in mice treated with IA verapamil, which they point out is an important finding as IV verapamil would be expected to cause bradycardia and hypotension which could exacerbate ischemic injury.²⁹

Fraser et al. then followed these experiments up with a phase I clinical trial entitled Superselective Administration of VErapamil During Recanalization in Acute Ischemic Stroke (SAVER-I) (NCT02235558).³⁰ Subjects undergoing mechanical thrombectomy with thrombolysis in cerebral infarction (TICI) 2a or better recanalization received 10 mg verapamil in 20 cc of normal saline over 20 min via microcatheter injection directly into the previously occluded artery. Primary endpoint was the presence of significant intracranial hemorrhage on 24 h post-intervention imaging study, and secondary objectives included systemic side effects of verapamil, technical feasibility of IA delivery, and modified Rankin score (mRS) at 90 days. Eleven patients were included, and IA delivery was technically successful in all. No patients met with primary endpoint of significant intracerebral hemorrhage (ICH), and there were two minor complications (a desaturation shortly following drug delivery, and a single post-procedural seizure). Rate of independence (mRS 0–2) at 90 days was 44.4%. The authors concluded that the IA delivery of verapamil immediately following thrombectomy is safe and feasible.³⁰

Cold saline

Hypothermia has been suggested to be the most potent neuroprotective strategy due to its ability to target multiple pathways simultaneously by several mechanisms within the ischemic and reperfusion cascades.³¹ To date, it is the only neuroprotective treatment that has proven efficacy and been introduced into clinical practice to reduce hypoxic brain injury after cardiac arrest.

The mechanisms by which therapeutic hypothermia provides neuroprotection are manifold. Cerebral metabolic rate of glucose and oxygen consumption has been shown to decrease by 5% for each 1°C reduction in body temperature during cooling.³² Cooling of ischemic rat brains to 33°C has been shown to suppress almost all glutamate release, and some studies have shown reduced expression of AMPA and NMDA receptors on hippocampal neurons after global ischemia leading to reduced infarct size.³³,³⁴ Anti-inflammatory effects of therapeutic hypothermia include reduction in levels of interleukin 1β, tumor necrosis factor α, interleukin 6, matrix metalloproteinase-2 (MMP-2) and matrix metalloproteinase-9 (MMP-9), and nuclear factor κβ.³¹ Anti-apoptotic effects include reduced concentrations of second mitochondrial-derived activators of caspases, reduced activity of pro-death signaling substances such as p53 and NAD, reduction of cytochrome c release and thus upregulation of anti-apoptotic proteins in the Bcl-2 family, and downregulation of the tumor necrosis factor pathway via reduction of Fas ligand and caspase 8 expression.³¹ Finally, there is evidence that generation of damaging free radicals is temperature dependent, and thus can be dampened by hypothermia.³⁵

There have been numerous preclinical animal models that have demonstrated neuroprotection from hypothermia. A 2007 systemic review and meta-analysis of animal models of focal cerebral ischemia showed an overall 44% reduction in infarct volume in animals treated with hypothermia compared to controls, for both temporary and permanent ischemia models.³⁶ However, translating these results to humans has been difficult. This is due to multiple challenges such as shivering, achieving cooling within the brief time window within which it could be effective, and complications related to the various methods used to achieve therapeutic hypothermia.

Surface cooling methods are limited by increased shivering and skin necrosis.³¹ Devices such as cooling helmets and nasal cooling devices are limited by slow rates of temperature reduction and local cooling. More invasive measures such as endovascular catheters inserted into the femoral or subclavian veins have the advantage of precise temperature monitoring and fast onset, but are limited by the risks of infection, bleeding, and thrombosis. Regardless of the method, studies have linked systemic hypothermia to a multitude of complications such as cardiovascular compromise, coagulopathy and platelet dysfunction, hypokalemia, and infection.³¹

The 2010 Intravascular Cooling in the Treatment of Stroke – Longer tPA Window (ICTuS-L) study showed a statistically increased rate of pneumonia in the hypothermia group compared with controls.³⁷ The 2016 ICTuS 2 trial utilized a femoral venous intravascular cooling device placed within 2 h of tPA administration, with a cold saline infusion used to achieve a target temperature of 33°C for 24 h along with a medical anti-shivering regimen, followed by 12 h of rewarming.³⁸ The trial demonstrated safety and feasibility of achieving therapeutic hypothermia with endovascular venous delivery of cold saline, but changes implemented to reduce the risk of pneumonia failed.³⁸

Given the complications related to systemic hypothermia described above, such as pneumonia and coagulopathy, there has been interest in intra-carotid infusion of cold saline in order to produce a targeted, local hypothermia of the brain with less overall systemic temperature reduction. A mathematical model developed in 2007 showed that an intra-carotid infusion of cold saline at a rate of 30 ml/min can reduce brain temperature to 33–34°C in 10 min.³⁹ The degree of temperature drop was proportional to the rate of infusion.³⁹

The feasibility, safety, and effectiveness of intra-carotid infusion of cold saline for neuroprotection with targeted hypothermia have been demonstrated in several animal models. Chen et al., using the standard rat middle cerebral artery occlusion (MCAo) model, infused 0°C 0.9% saline through a microcatheter into the ICA and achieved rapid reduction of brain temperature in the cerebral cortex from 37.1 ± 0.3 to 30.7 ± 0.4°C, and in the striatum from 37.5 ± 0.3 to 30.9 ± 0.5°C.⁴⁰ Cold saline was administered in a total volume of 2.5 ml over 10 min at a rate of 0.25 ml/min. The authors reported improvement in neurological testing in rats that received IA cold saline as well as reduction in infarct volume (40.6 ± 5.3 vs. 61.7 ± 8.6%, p < 0·01) and cerebral edema (7.8 ± 2.6 vs. 15.4 ± 3.2%, p < 0·01) 48 h following stroke.⁴⁰

Wang et al. performed a feasibility study using this technique on rhesus monkeys, although no stroke was induced.⁴¹ Microcatheter injection of 100 ml 0°C lactated ringers (LR) solution over 20 min into the MCA of eight monkeys produced a reduction in ipsilateral MCA brain territory to <35°C within 10 min, and the effect lasted for 20 min post-infusion. This local effect was not observed in four control monkeys that received systemic cooling; however, there was no difference in rectal temperature decrease between groups (0.5 ± 0.2 and 0.5 ± 0.3°C, respectively). There was no observed cerebral edema on MRI or vasospasm on angiography in monkeys that received IA cold LR.⁴¹

Mattingly et al. investigated the use of IA cold saline in 28 swine.⁴² Tandem CCA and MCA occlusion was performed for 3 h using a dual lumen balloon catheter in the CCA and craniotomy for MCA occlusion, followed by reperfusion. Half of the animals received 3 h of cold saline infusion via the CCA catheter calibrated to temperature which was continuously monitored along with hemodynamics and blood gas. Animals were sacrificed following the additional 3 h of reperfusion ± cold saline infusion. The authors reported reduction in mean ipsilateral hemisphere temperature from 38 to 26°C and in the contralateral hemisphere to 31.6°C. Stroke volume measured on MRI was significantly reduced from 0.050 ± 0.059 to 0.005 ± 0.011 (p = 0.046), and on histology by 58%, although this did not reach statistical significance due to wide variability (p = 0.256). No significant hemodynamic changes were observed, but there was a reduction in arterial pH from 7.44 ± 0.03 to 7.34 ± 0.18.⁴²

This technique has been attempted in two small scale phase I human clinical trials. A 2010 study performed in the U.S. delivered saline at 15°C and 7°C into the internal carotid artery (ICA) of 18 patients undergoing follow-up routine angiograms at 33 ml/min for 10 min.⁴³ Jugular bulb temperatures demonstrated a drop of 0.84°C compared to 0.15°C for bladder temperatures. Three of the patients experienced shivering, but there was no significant change in vitals, neurologic exam, discomfort, or laboratory values such as hematocrit.⁴³ In a phase I study done in 2016 in China, the authors administered 4°C saline at 30 ml/min for 10 min into the ICA of 26 stroke patients undergoing thrombectomy.⁴⁴ Rectal temperature decreased by 0.1°C, and vitals, electrolytes and hematocrit did not significantly change during or after treatment. The authors thus concluded that intra-arterial infusion of 4°C saline following thrombectomy for LVO patients is safe and feasible.⁴⁴

Stem cells

Several studies have investigated the use of stem cells as neuroprotective agents in ischemic stroke. It has been demonstrated that transplanted cells may have a protective effect via a paracrine mechanism through production of trophic factors, growth factors, and various cytokines that have anti-apoptotic and pro-restorative effects such as angiogenesis, synaptogenesis, and neurogenesis.⁴⁵,⁴⁶ Kamiya et al. compared groups of rats that received 1 × 10⁷ autologous bone marrow mononuclear cells (BMMCs) immediately after reperfusion in an MCAo model via intra-arterial delivery into the ICA vs. into the femoral vein.⁴⁵ They found that cortical infarct volume was significantly reduced in rats that received IA BMMCs (8.19 ± 9.65 mm³), but not in rats that received them intravenously (50.9 ± 41.2 mm³) compared to controls (85.6 ± 34.15 mm³) (p = 0.002). Assessment of motor function via rotarod testing showed significantly improved scores compared to controls 24 h after reperfusion in rats that received IA BMMCs (p = 0.008), while there was no difference in rats that received them intravenously. Fluorescence microscopy of brain sections showed a significantly greater density of transplanted cells (25 cells/mm²) in IA groups than in IV groups (24 cells/mm²) (p = 0.046).⁴⁵

Oh et al. showed that rats that received intra-carotid delivery of human adipose-derived mesenchymal stromal cells (ADMSCs) had significant improvement in ratarod and mNSS tests and that this effect was more pronounced during the early phase of ischemia (within the first three weeks) compared with later phases.⁴⁶ Rats treated with IA ADMSCs had significantly reduced infarct size (25.2% ± 2.9%) compared to controls (35.6% ±3.3%) (p = 0.02). They concluded that the main effects of human AD-MSC transplantation were the early amelioration of neuroinflammation, promotion of subventricular zone neurogenesis, and reduction of astrogliosis in the host brain.⁴⁶ Huang et al. infused adipose-derived stem cells (ADSCs) into the ICA of rats 24 h after transient middle cerebral artery occlusion.⁴⁷ They demonstrated that ADSCs facilitated the regeneration of neuronal fibers and blood vessels in the peri-infarct cortex, improving neurological function.⁴⁷

In addition to these animal models, there have been several early phase clinical trials investigating whether the intra-arterial delivery of stem cells is a safe and feasible treatment in ischemic stroke.⁴⁸ Battistella et al. infused autologous bone marrow mononuclear cells (BM-MNCs) in six patients with non-acute MCA stroke within 90 days of symptom onset.⁴⁹ The stem cells were infused for approximately 10 min into the M1 portion of the MCA. The authors concluded that this treatment was feasible and safe.⁴⁹ Friedrich et al. conducted another trial using intra-arterial BM-MNCs in patients with moderate to severe acute ischemic stroke.⁵⁰ The trial demonstrated 40% rate of good clinical outcome (mRS score ≤2 AT 90 days) and 30% satisfactory clinical improvement at 90 days (mRS score of 0 in patients with NIHSS score <8, mRS score of 0–1 in patients with NIHSS scores of 8–14, or mRS scores of 0–2 in patients with NIHSS score >14).⁵⁰

Others

MMP-9 has been implicated in blood brain barrier disruption during cerebral ischemia and reperfusion through disruption of fibronectin, elastin, vitronectin, and type IV collagen.⁵¹ Norcatharidin (NCTD) is a MMP-9 inhibitor that has shown particular promise in neuroprotection with reduction in hemorrhagic transformation and cerebral edema.⁵¹ Khan et al. administered 0.5 mg/kg of NCTD dissolved in dimethyl sulfoxide and diluted in normal saline (ratio of 1:5) into the ICA of spontaneously hypertensive rats immediately following 90 min of MCAo at a rate of 0.10 ml/min over 5–7 min.⁵² Neurological assessment was performed using the ladder rung walking test at days 1, 3, and 7 following stroke, and also daily using the Garcia Neurological Test. MRI was performed at 24 h to assess infarct volumes and blood brain barrier integrity. Mice treated with NCTD performed significantly better on both neurological assessment tests at all time points, and had significantly smaller infarct volumes (148.1 ± 10.2 vs. 178.5 ± 9.9 mm³, p < 0.05) and lower percentage of hemispheric contrast enhancement (0.3 ± 0.01% vs. 2.5 ± 0.80%, p < 0.05).⁵²

Administration of exogenous erythropoietin (EPO) in ischemic stroke has been shown to have anti-apoptotic, anti-oxidant, neurogenic, and neurotrophic effects in animal models.⁵³ Clinical translation has been limited, however, due to systemic effects such as red cell aplasia and thrombosis. Researchers have thus suggested IA delivery of EPO in order to utilize lower dosage and reduce systemic exposure. Wang et al. utilized a rat MCAo model with 2 h occlusion and compared delivery of 800 U/kg of EPO (0.1 ml/min over 5 min) directly into the ICA immediately after reperfusion to a control group which received IA saline at the same rate and volume.⁵³ Neurologic function was evaluated using the Ludmila Belayev test, blood brain barrier permeability was evaluated using Evans blue dye, and infarct volumes and edema were assessed with triphenyltetrazolium chloride (TTC) staining 24 h post-stroke. Mice that were treated with IA EPO had significant reduction in infarct volume and edema compared to controls (36.2% and 31.8%, respectively), performed significantly better on neurologic testing, and exhibited a smaller amount of dye extravasation (6.74 ± 0.96 ng/g vs. 11.92 ± 1.34 ng/g, p < 0.05). Furthermore, immunohistochemistry was used to demonstrate that IA EPO prevented degradation of Claudin-5 and Occludin, and reduced expression of MMP-2 and MMP-9.⁵³

Combined approaches

Since the ischemic and reperfusion injury cascade is complex and involves many pathways, logic dictates that an ideal therapeutic approach would be to target it with multiple therapeutic agents rather than one. There have been a few attempts at combining IA neuroprotective agents in the literature.

Song et al. sought to combine the neuroptotective effects of IA cold saline and magnesium sulfate with the goal of achieving a synergistic increase in efficacy.⁵⁴ Their experiment consisted of 68 male rats divided into the following six groups: sham surgery, MCAo control, MCAo + 15°C IA magnesium, MCAo + 15°C IA saline, MCAo + normothermic (37°C) IA magnesium, and MCAo + 37°C IA saline. Occlusion time was 30 min, and either saline or 120 mg/kg MgSO₄ was infused at a rate of 0.4 ml/min over 20 min immediately after reperfusion. Neurological function was assessed 48 h after stroke using the modified neurological severity score, and infarct volumes were assessed 48 h after stroke using TTC staining. Animals that received local cold saline infusion experienced a reduction in ipsilateral MCA territory brain temperature to 33–34°C within 5–10 min, without significant change in body temperature. No significant differences were observed in blood pressure, arterial pH, PaO₂, and serum glucose level, although there was a non-significant trend towards a reduction in hematocrit with local infusion. There was a significant reduction in brain infarct volume of 48% in mice that received IA cold saline compared to controls, and 65% in mice that received IA cold magnesium. IA cold magnesium resulted in a significantly lower stroke volume than IA saline alone as well as significantly greater improvement in neurological outcome.⁵⁴

Chen et al. sought to determine whether IA delivery of human albumin solution combined with cold saline could produce neuroprotective effects using a lower dose than what had been used intravenously in prior clinical trials that resulted in severe dose-related adverse effects.⁵⁵ The MCAo model was used in adult rats for 2 h followed by reperfusion. Animals were again divided into six groups: sham surgery, MCAo + IA 0°C saline, MCAo + IA 0°C 0.5 g/kg human albumin, MCAo + IA 37°C 0.5 g/kg human albumin, MACo + systemic (femoral artery) low dose (0.5 g/kg) 37°C human albumin, and MCAo +systemic high dose (1.5 g/kg) 37°C human albumin. Rats that received cold saline or albumin experienced a reduction in ipsilateral brain temperature (37.2 ±0.2°C to 30.5 ± 0.4°C in the cortex, and 37.8 ± 0.1°C to 30.8 ± 0.4°C in the striatum) without a significant difference in body temperature. Significant reductions in infarct volumes were observed in mice treated with high-dose systemic albumin (37% reduction), IA low-dose albumin (43%), IA cold saline (32%), and IA cold albumin (67%). IA cold albumin resulted in the greatest reduction in infarct volume and was significantly lower than the other experimental groups (p < 0.01). Similar significant findings were observed in neurologic function assessments.⁵⁵

The Magnesium And Verapamil After Recanalization in Ischemia of the Cerebrum (MAVARIC) aims to evaluate the safety and feasibility of IA delivery of verapamil and magnesium at the time of recanalization and is currently recruiting patients (NCT02912663).⁵⁶ This is a randomized trial with a placebo group as a control and will administer 10 mg of verapamil in 10 ml of normal saline and 1 g of MgSO₄ in 20 ml of normal saline over a total of 20 min (1 ml/min), through a microcatheter directly into the previously occluded artery. Primary outcome is the number of patients with symptomatic ICH within 48 h of intervention, and secondary outcome is the number of patients with systemic side effects from the drugs.⁵⁶

Future directions

The results of the recent stroke intervention trials are encouraging, but the discrepancy in procedural success with long-term functional outcomes suggests that more can be done to improve clinical outcomes. Pharmacologic neuroprotective strategies have long been investigated as a tool to improve outcomes, but despite promising results in preclinical studies, the results of human clinical trials have been disappointing. Several factors that may explain these failures include lack of significant numbers of patients that underwent successful reperfusion in most of these trials, systemic administration of drugs and toxicity, delay in administration after ischemic injury, and only a single agent being used in a process that involves multiple pathologic pathways. Intra-arterial delivery of a cocktail of neuroprotective agents may help address these issues.

As described here, there have been several preclinical models of large vessel occlusion ischemic stroke that have utilized IA delivery of neuroprotective substances and demonstrated significant neuroprotective effects. Many of these demonstrate the advantage of IA delivery, including higher local concentration of drug, avoidance of first-pass metabolism, and lower systemic exposure. Some even provide examples of drugs that were promising from a neuroprotective standpoint, but clinical translation was limited due to systemic toxicity which could potentially be avoided with IA delivery. This includes hypothermia, which has frustrated researchers for decades due to systemic complications. Targeted ipsilateral brain hypothermia with IA cold saline provides a means to achieve rapid onset of the beneficial neuroprotective effects of hypothermia while avoiding the burden of systemic complications.

The future of neuroprotection in stroke will likely involve combination therapy utilizing multiple agents that can attack the ischemic injury and reperfusion cascade from multiple angles. Some of the preclinical models described here demonstrate the synergistic effect of cold saline with neuroprotective agents such as magnesium sulfate and human albumin. The ideal approach would likely involve a combination of targeted hypothermia and several drugs that provide anti-excitotoxic, anti-oxidant, anti-apoptotic, and anti-inflammatory effects.

One phase I clinical trial (SAVER-I) has already been completed which demonstrated the safety and feasibility of IA delivery of neuroprotective drugs immediately after reperfusion in large vessel occlusion ischemic stroke patients.³⁰ Two additional clinical trials are ongoing.²⁸,⁵⁶ Two human clinical pilot studies have also been published demonstrating the safety and feasibility of infusing cold saline directly into the carotid artery in patients undergoing angiograms and in stroke patients.⁴³,⁴⁴ As additional clinical trials are published, the safety profile of these techniques will become established and the ideal cocktail for neuroprotection will need to be investigated.

Conclusions

Many of the promising neuroprotective agents studied previously for ischemic stroke deserve to be revisited in the setting of high rates of successful reperfusion that enables the agents to actually reach their target territory. IA delivery of such substances presents numerous attractive advantages such as convenience (a catheter is already in the carotid artery), immediate effect, high local concentration, lower systemic exposure and toxicity, and avoidance of first-pass metabolism. Promising substances such as magnesium sulfate, verapamil, stem cells, human albumin, erythropoietin, cold saline, and combinational therapies have already been studied in preclinical and some even in clinical scenarios. The ideal neuroprotective cocktail going forward will likely consist of a combination of targeted hypothermia with cold saline and several drugs that can attack the ischemic injury and reperfusion injury cascade in a synergistic manner, delivered immediately after reperfusion in stroke intervention directly to the target ischemic territory. The endovascular community should investigate these treatments as a way to potentially improve outcomes of stroke intervention.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Thomas W Link https://orcid.org/0000-0003-1325-8847