Immunotherapy of experimental and human stroke with agents approved for multiple sclerosis: a systematic review.

BACKGROUND: 'Thromboinflammation' describes a novel concept in stroke pathophysiology that has opened up the possibility of immunotherapeutic approaches which could become promising strategies for targeted stroke therapies in the future.
METHODS: We reviewed current evidence for agents approved for multiple sclerosis in preclinical and clinical stroke studies. A systematic review was performed in accordance with the PRISMA statement, searching MEDLINE, the Cochrane Central Register of Controlled Trials, and reference lists of articles published until 16 October 2017.
RESULTS: The review included 52 of 629 identified studies, consisting of 5 clinical and 47 preclinical trials. Most of the studies showed beneficial effects of the evaluated immunotherapeutic drugs in terms of reduction in morphological lesion size and improvement in functional outcome. Nevertheless, the significance of these findings is limited due to the high degree of heterogeneity.
CONCLUSIONS: Immunotherapy of stroke might be effective and could become a promising treatment strategy, but larger clinical trials with standardized interventions and outcome measures are needed.

Introduction

The concept of ‘thromboinflammation’ describes the pathophysiological link between thrombus formation, that is, mechanisms related to platelet aggregation and plasmatic coagulation, and inflammation in the development of ischaemic stroke (IS).[1]

With the help of murine stroke models and transgenic mouse strains, it has been demonstrated by independent groups that immunological processes involving the innate or adaptive immune system modulate IS pathophysiology. Histopathological findings demonstrating immune cell infiltration in the ischaemic brain were reported as early as the 1990s, but the pathophysiological relevance remained unclear.[2] Yilmaz and coworkers were the first to show that Rag1 mice, that is, mice lacking lymphocytes, develop smaller cerebral infarctions compared with wild-type animals.[3] This observation paved the way for rodent stroke studies analysing immunomodulating therapies with agents that have already been approved for the treatment of patients with multiple sclerosis. As the majority of rodent studies had positive outcomes, in the next step, the first clinical trials in stroke patients were started.

The aim of our study was to systematically assess the published preclinical and clinical studies that analysed an immunotherapeutic compound approved for multiple sclerosis in stroke.

Methods

We conducted a systematic review and followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines.[4]

Prior to systematically reviewing the literature, the following eligibility criteria were defined: (a) clinical studies: randomized controlled trials (RCTs) or prospective studies with or without a control group if they analysed a compound approved for multiple sclerosis in patients aged 18 years or older with stroke [cerebral ischaemia or spontaneous intracranial haemorrhage (ICH)] or transient ischaemic attack; (b) animal studies that analysed a compound approved for multiple sclerosis in an experimental stroke model. Publications about spinal cord ischaemia and subarachnoid haemorrhages have been excluded due to particular pathophysiological mechanisms.

The main outcome measures that have been considered were mainly stroke volume and functional deficits. In part, additional outcomes, such as brain oedema, local brain inflammation or systemic cytokine levels, have been assessed. Due to space restrictions, it has been necessary to limit Tables 1–6 to principal content. Because of the pronounced heterogeneity of the study design, stroke model, intervention and outcome variables, it was not possible to calculate a meaningful meta-analysis for any of the outcome variables.

Table 1.

Characteristics of preclinical studies analysing glatiramer acetate.

Author	Species	Stroke model	Intervention	Main outcome
Ibarra[6]	Sprague–Dawley rats (age: n.a.; sex: male)	Transient MCAO (120 min)	Injection of 200 µg Cop-I and CFA 30 min after reperfusion versus saline and CFA only	Reduction in stroke volume and functional deficit on day 7 (p < 0.05) but not on day 1 (p > 0.05)
Poittevin[7]	C57BL/6 mice (age: 10–12 weeks; sex: male)	Transient (45 min) and permanent MCAO	Glatiramer acetate 2 mg/200 µl NaCl 0.9% s.c. versus vehicle	Permanent MCAO at day 3, no difference in infarct volume or brain oedema (p > 0.05); transient MCAO no change at day 3 (p > 0.05), no difference in functional outcome at day 7 (p > 0.05)
Kraft[8]	C57BL/6 mice (age: 6–8 weeks; sex: male)	Transient MCAO (60 min)	Glatiramer acetate 3.5 mg/kg i.v. 30 min before ischaemia versus vehicle	No difference in infarct volume (p > 0.05) and neurological outcome (p > 0.05)
Cruz[9]	Sprague–Dawley rats (age: 9 weeks; sex: male)	Transient MCAO (90 min)	Glatiramer acetate 200 µg in saline emulsified in CFA containing 5 mg/ml of Mycobacterium tuberculosis H37Ra in a total volume of 150 μl versus saline and CFA only	Improvement in neurological recovery after MCAO not at day 1 (p > 0.05), but at day 7 (p < 0.05) and between days 14 and 60 (p < 0.01); reduced stroke volume (p < 0.05) 60 days after stroke

CFA, complete Freund’s adjuvant; COP-I, copolymer I; i.v., intravenously; MCAO, middle cerebral artery occlusion; n.a., not applicable; NaCl, sodium chloride; s.c., subcutaneously.

Table 2.

Characteristics of preclinical studies analysing interferon β.

Author	Species	Stroke model	Intervention	Main outcome
Liu[10]	New Zealand white rabbits (age: n.a.; sex: n.a.)	Clot embolus surgically injected into MCA	IFN-β pretreatment with 107 U s.c. 4 h before clot placement and 0.5 × 107 U within 30 min after; IFN-β post-treatment with 107 U s.c. immediately after clot placement and 0.5 × 107 U 4 h later; control group: no IFN-β application	Reduced infarct volume (pretreatment p = 0.003; post-treatment p = 0.004)
Veldhuis[11]	Fischer rats (age: n.a.; sex: male)	Transient MCAO (60 min)	Recombinant rat IFN-β 500,000 U s.c. 2 days prior to surgery, or at reperfusion, or 4 h after stroke onset, or 6 h after stroke onset versus control (saline)	Infarct volume smaller for IFN-β on day 1 (p < 0.01) versus control; on day 1, greater improvement in pretreated group compared with groups treated after stroke (p < 0.05); from day 7 onwards, no difference between the IFN-β groups (p > 0.05)
Veldhuis[12]	Fischer rats (age: 8–12 weeks; sex: male)	Transient MCAO by using a microclip on the MCA through a small cranial burr hole, reperfusion after 60 min	IFN-β 500,000 U (8 µg) s.c. once daily until 7 days after reperfusion versus vehicle; treatment began 2 days before MCAO, on reperfusion, 4 h after stroke onset or 6 h after stroke onset	Reduction in lesion volume in all IFN-β treatment strategies on days 1, 7 and 21 (p < 0.05 to p < 0.001)
Maier[13]	Sprague–Dawley rats (age: n.a.; sex: male)	Transient MCAO (60 min) or permanent MCAO	Rat IFN-β 8 or 16 µg i.v. once daily for 3 or 7 days, or PEG-IFN-β i.v. or s.c. for 1 day	IFN-β and PEG-IFN-β failed to mitigate stroke volume and functional deficits on day 7 (p > 0.05)
Kuo[14]	C57BL/6 and Ifnar1tm1Agt/Mmjax (Ifnar1−/− mice) (age: 8–12 weeks; sex: male)	Transient MCAO (40 min)	Recombinant murine IFN-β 10,000 U i.v. 3 h before MCAO induction or 3 h after reperfusion	Pre- and post-treatment with IFN-β reduced infarct volume (p = 0.001) and functional deficit (p < 0.05) in C57/BL6 mice; no change in infarct volume in Ifnar1−/− mice
Cruz[15]	C57BL/6 mice versus C57BL/6 mice with LysMCre/IRF2BP2flox (ablation of IRF2BP2) (age: 2 months; sex: male)	Photothrombotic stroke	Mouse recombinant IFN-β 10,000 U in 100 µl saline, 30 min after photothrombosis for all animals	Similar infarct volume in wild-type and transgenic mice 1 day after stroke; lesion volumes reduced in control versus knock-out mice on day 4 (p < 0.05); worse functional outcome in transgenic mice versus control animals (p < 0.05)

IFN-β, interferon β; IRF2BP2, interferon regulatory factor 2 binding protein 2; i.v., intravenously; MCA, middle cerebral artery; MCAO, MCA occlusion; n.a., not applicable; PEG, pegylated; s.c., subcutaneously.

Table 3.

Characteristics of preclinical studies analysing fingolimod.

Author	Species	Stroke model	Intervention	Main outcome
Czech[16]	C57BL/6J mice (age: 10 weeks; sex: male)	Transient MCAO (90 min)	FTY720 1 mg/kg i.p. after initiation of anaesthesia	Reduction in stroke volume (p < 0.05) and functional deficits (p < 0.01)
Wacker[17]	Swiss-Webster ND4 mice (age: n.a.; sex: male)	Transient MCAO (60 min)	FTY720 0.24 mg/kg or 1 mg/kg i.p. 30 min before hypoxic preconditioning; for mice not subjected to HPC, FTY720 treatment 48 h before MCAO	Reduction in infarct volume and functional deficits with 1 mg/kg FTY720 (p < 0.05), not 0.24 mg/kg (p > 0.05); even stronger protection from ischaemic stroke in combination with HPC (p < 0.05)
Shichita[18]	C57BL/6 and other mouse strains (age: 9–17 weeks; sex: male)	Transient MCAO (60 min)	FTY720 1 mg/kg 5 min before reperfusion and once daily for 3 days versus H2O	Reduction in infarct volume (p < 0.01)
Hasegawa[19]	Sprague–Dawley rats (age: n.a.; sex: male)	Transient MCAO (120 min)	FTY720 0.25 mg/kg or 1 mg/kg i.p. immediately after reperfusion versus vehicle; in other groups SEW2871 (selective S1P1 agonist) and VPC23019 (S1P1, S1P3 and S1P4 antagonist)	Reduction in infarct volume and functional deficits on days 1 and 3 (p < 0.05)
Pfeilschifter[20]	C57BL/6J mice (age: 10 weeks; sex: male)	Transient MCAO (90 min, 180 min)	FTY720 1 mg/kg i.p. 2 h after vessel occlusion versus vehicle	Smaller lesion size on day 1 after 3 h MCAO (p = 0.001); better neurological performance (p = 0.005); smaller lesion size after 90 min MCAO (p = 0.013), no improvement in functional outcome (p = 0.81)
Pfeilschifter[21]	C57BL/6, SphK1−/− and SphK2−/− mice (age: 10–12 weeks; sex: n.a.)	Transient MCAO (90 min, 180 min)	FTY720 1 mg/kg i.p. 2 h after vessel occlusion versus vehicle	Reduction in stroke volume at day 1 (p = 0.001 and 0.013, respectively)
Liesz[22]	C57BL/6 mice (age: 8–10 weeks; sex: male)	Permanent MCAO, transient MCAO (60 min)	FTY720 1 mg/kg p.o. starting at 48 h before or at 3 h after ischaemia induction versus PBS; single dose of FTY720; FTY720 1 mg/kg i.p. once daily beginning 48 h before MCAO	No difference (p > 0.05) in infarct volume and functional outcome
Wei[23]	C57BL/6 mice, Sprague–Dawley rats (age: n.a.; sex: male)	Transient MCAO in mice (90 min) and rats (2 h); permanent MCAO in mice	FTY720 (1 mg/kg or 0.5 mg/kg) i.p. 30 min after reperfusion versus saline; FTY720 (1 mg/kg) versus saline 1 h before distal MCAO; 3 mg/kg FTY720 2, 24 and 48 h after reperfusion; FTY720 (1 mg/mg, i.p.) 30 min after reperfusion in rats; 1 mg/kg FTY720 2 or 4 h after occlusion in permanent model	Reduced infarct volume on day 2 in all experimental settings (p < 0.05 to p < 0.001); improved neurological function for the 1 mg/kg group (p < 0.05), not in the 0.5 mg/kg group (p > 0.05) on day 2, and in group given 3 mg/kg twice after reperfusion on days 1, 3, 7, 10 and 14 (p < 0.05 to p < 0.001)
Rolland[24]	CD-1 mice (age: 8 weeks; sex: n.a.)	ICH induction by intrastriatal collagenase injection versus needle insertion only (sham operation)	FTY720 1 mg/kg i.p. 1 h after ICH induction versus vehicle	Reduced brain oedema (p < 0.05), better functional outcome on days 1 and 3 (p < 0.05)
Rolland[25]	CD-1 mice and Sprague–Dawley rats (age: n.a.; sex: male)	Experimental ICH (collagenase or autologous blood injection in striatum)	FTY720 1 mg/kg i.p. single dose 1 h after or daily dose 1, 24 and 48 h after ICH versus vehicle	Less brain oedema in FTY720-treated mice versus vehicle group (p < 0.05); better neurological function (p < 0.05)
Kraft[26]	C57BL/6 and Rag1−/− mice (age: 6–8 weeks; sex: male)	Transient MCAO (60 min, 90 min)	FTY720 1 mg/kg i.p. immediately before reperfusion versus vehicle	Reduction in stroke volume at day 1 (p = 0.048); improved neurological function (p = 0.02 to p = 0.03)
Campos[27]	C57BL/6 mice (age: n.a.; sex: male)	Thromboembolic stroke model using mouse-α-thrombin dissolved in 18% glycerol/saline	1. MCAO not treated with rt-PA (permanent occlusion); fingolimod 0.5 mg/kg i.p. versus saline 45 min, 24 and 48 h after occlusion2. MCAO + early rt-PA; rt-PA i.v. 30 min after thrombin injection (transient occlusion); fingolimod versus saline 30 min (together with rt-PA), 24 and 48 h after occlusion3. MCAO + delayed rt-PA, rt-PA i.v. 3 h after thrombin injection (transient occlusion); fingolimod versus saline 3 h (together with rt-PA), 24 and 48 h after occlusion	In absence of rt-PA, fingolimod reduced stroke volumes (p < 0.05) and improved functional outcome (p < 0.05); early rt-PA and fingolimod applications had no impact on stroke volume (p > 0.05) but improved functional outcome (p < 0.05); late rt-PA and fingolimod applications reduced stroke volume (p < 0.05) and improved functional outcome (p < 0.05)
Cai[28]	C57BL/6 mice (age: 10–12 weeks; sex: n.a.)	Transient MCAO (180 min)	FTY720 1 mg/kg i.p. versus vehicle versus rt-PA 10 mg/kg i.v. versus rt-PA 10 mg/kg i.v. + FTY720 1 mg/kg i.p.; all directly before reperfusion	Higher mortality in FTY720 + rt-PA group (61%) versus vehicle (33%), FTY720 (39%) and rt-PA only (44%)
Brunkhorst[29]	C57BL/6J mice (age: 6–12 weeks; sex: male)	Photothrombotic stroke	FTY720 1 mg/kg i.p. twice daily for 5 days, beginning 3 days after photothrombotic stroke versus saline	Improvement in functional outcome on day 7 (p = 0.013 to p = 0.003) and day 31 (p = 0.02 to p = 0.03)
Hasegawa[30]	Sprague–Dawley rats (age: n.a.; sex: male)	Transient MCAO (120 min)	FTY720 0.25 mg/kg in DMSO versus DMSO i.p. directly after reperfusion	Infarct volume reduction on day 1 in FTY720 group (p = 0.05)
Lu[31]	CD-1 mice (age: n.a.; sex: male)	Experimental ICH (collagenase injection in basal ganglia)	FTY720 0.5 mg/kg i.p. 30 min after surgery versus vehicle and once daily in following 2 days	Reduction in brain oedema and haematoma volume after 72 h (p < 0.05). After 3, 7 and 14 days, reduced oedema and brain atrophy (p < 0.05)
Moon[32]	ICR mice (age: 7 weeks; sex: male)	Transient MCAO (60 min, 90 min)	FTY720 3 mg/kg i.p. immediately after reperfusion (90 min MCAO) or 30 min prior to 60 min MCAO versus vehicle	Reduction in stroke volume (p < 0.05)
Schuhmann[33]	C57BL/6 mice (age: 6–8 weeks; sex: male)	Transient MCAO (30 min)	FTY720 1 mg/kg i.p before ischaemia + after 2 days versus vehicle	Smaller lesion size on day 1 (p < 0.05); functional improvement (p < 0.05)
Nazari[34]	Sprague–Dawley rats (age: n.a.; sex: male)	Transient MCAO (60 min)	FTY720 0.5 mg/kg i.p. 24 h before vessel occlusion versus vehicle plus once daily every 2 days	Reduced brain oedema (p < 0.01) and neurological deficit score (p < 0.05) after 24 h, 3 and 7 days
Sun[35]	C57BL/6J and Rag2−/− mice (for the FTY720 part only C57BL/6J mice were used) (age: 7–8 weeks; sex: male)	Experimental ICH (injection of autologous blood)	FTY720 1 mg/kg 30 min after ICH induction versus vehicle versus RP101075 (selective S1PR1 agonist) versus RP101075 + W146 (S1PR1 antagonist)	Reduction in functional deficits on days 1 and 3 (p < 0.01); infarct volumes have not been published for the FTY720 group
Schlunk[36]	CD-1 mice (age: 12–16 weeks; sex: male)	Experimental ICH (injection of collagenase type VII-S)	FTY720 1 mg/kg i.p. 1 h after ICH induction versus vehicle	No change in mortality, functional outcome, haematoma volume and oedema (p > 0.05)
Rolland[37]	Sprague–Dawley rats (age: n.a.; sex: male/female)	Germinal matrix haemorrhage in rat pups	FTY720 0.25 mg/kg or 1.0 mg/kg i.p. 1, 24 and 48 h after surgical intervention versus DMSO in saline	Better functional outcome (p < 0.05) and increased total brain surface area (p < 0.05) in both dosages
Zhang[38]	C57BL/6 and BALB/c nude mice (age: 8–12 weeks; sex: male)	Experimental ICH (autologous blood injection in striatum)	FTY720 1 mg/kg i.p. 1, 24 and 48 h after ICH versus vehicle	Lower BBB leakage and CD4+/CD8+ cells in nude and FTY720 treated wild-type mice versus vehicle group (p < 0.05)

BBB, blood–brain barrier; DMSO, dimethyl sulphoxide; FTY720, fingolimod; HPC, hypoxic preconditioning; ICH, intracerebral haemorrhage; i.p., intraperitoneally; i.v., intravenously; MCAO, middle cerebral artery occlusion; n.a., not applicable; PBS, phosphate-buffered saline; p.o., per os; rt-PA, recombinant-tissue plasminogen activator.

Table 4.

Characteristics of preclinical studies analysing natalizumab.

Author	Species	Stroke model	Intervention	Main outcome
Becker[39]	Lewis rats (age: n.a.; sex: male)	Transient MCAO (180 min)	TA-2 (selective anti-α4 antibody) 2.5 mg/kg i.p. 2 h after vessel occlusion versus isotype control	Improved functional outcome on day 1 (p = 0.006), day 2 (p = 0.011); smaller infarct volume (p = 0.012)
Relton[40]	Wistar rats (age and sex: n.a.)	Transient MCAO (90 min)	TA-2 (selective anti-α4 antibody) 2.5 mg/kg versus isotype control i.v. immediately after reperfusion	Reduced total (p < 0.05) and subcortical stroke volume (p < 0.001), functional improvement after 24 h (p < 0.01)
Relton[41]	SHR (hypertensive) or Sprague–Dawley rats (age: n.a.; sex: male)	Transient MCAO (60 min)	TA-2 (selective anti-α4 antibody) 2.5 mg/kg versus isotype control i.v. 24 h before induction of cerebral ischaemia	Reduced total (p < 0.05) and subcortical stroke volume (p < 0.01) in hypertensive SHR rats; reduced total and subcortical stroke volume (p < 0.001) in normotensive Sprague–Dawley rats
Liesz[42]	C57BL/6J mice (age: 10–12 weeks; sex: male)	Transient (30 or 60 min) or permanent MCAO	CD49d-specific monoclonal antibody 300 μg i.p. 24 h before or 3 h after induction of ischaemia versus isotype control	Reduced infarct volumes at day 7 (p < 0.001), not at day 1 (p > 0.05), for permanent MCAO and 30 min (but not 60 min) transient MCAO; improved functional outcome on days 3 and 7 (p < 0.05)
Langhauser[43]	C57BL/6 mice (age: 6–8 weeks; sex: male)	Transient MCAO (30 min) or permanent MCAO (coagulation model)	CD49d-specific monoclonal antibody 300 μg i.p. 24 h before or 3 h after induction of ischaemia versus isotype control	Independent of prophylactic or therapeutic treatment and stroke model: no change in stroke volume and functional scores on days 1 and 7 (p > 0.05); no change in survival (p > 0.05)
Hammond[44]	C57BL/6J mice (age: 8–12 weeks; sex: male)	Experimental ICH (injection of autologous blood)	Anti-α4 integrin antibody (clone R1-2) 300 µg i.p. 2–6 h before ICH	Improved functional outcome (p < 0.01)
Llovera[45]	C57BL/6J mice (age: 8–10 weeks; sex: male)	Transient MCAO (60 min) or permanent MCAO (coagulation model)	CD49d-specific monoclonal antibody 300 μg i.p. 3 h after induction of ischaemia versus isotype control	Reduced stroke volume in permanent MCAO (day 7, p < 0.05), not for transient MCAO (day 4, p > 0.05); no impact on functional outcome (p > 0.05)
Neumann[46]	LysM-eGFP and CX3CR1-eGFP mice (age: 8–10 weeks; sex: male)	Permanent MCAO (coagulation model) or transient MCAO (45 min)	CD49d-specific monoclonal antibody 150 μg i.v. at beginning of reperfusion (transient MCAO) and 24 h later versus isotype control	Reduced stroke volume and improved functional outcome (p < 0.05)

ICH, intracranial haemorrhage; i.p., intraperitoneally; i.v., intravenously; MCAO, middle cerebral artery occlusion; n.a., not applicable; SHR, spontaneously hypertensive rat.

Table 5.

Characteristics of preclinical studies analysing dimethyl fumarate.

Author	Species	Stroke model	Intervention	Main outcome
Iniaghe[47]	CD-1 mice (age: n.a.; sex: male)	Experimental ICH (injection of collagenase or autologous blood)	DMF 10 or 100 mg/kg i.p. 1 h after ICH versus vehicle; further experimental groups, including siRNA or MAFG siRNA	Low-dose DMF (10 mg/kg) did not improve functional outcome (p > 0.05), high dose (100 mg/kg) reduced functional deficits at days 1 and 3 (p < 0.05); no impact on haematoma volume (p > 0.05)
Zhao[48]	Sprague–Dawley rats, Nrf2+/+ and Nrf2−/− mice (age: n.a.; sex: male)	Experimental ICH (injection of autologous blood)	Rats: DMF 15 mg/kg i.p. 2 h after ICH and then twice daily on days 1–3 versus vehicle; mice: DMF 15 mg/kg i.p. 24 h after ICH and then at days 2 and 3 versus vehicle	Amelioration of neurological deficit in rats at days 1 and 3 after ICH, in wild-type but not Nrf−/− mice (p < 0.05)
Kunze[49]	C57BL/6 and Nrf2−/− mice (age: 8–10 weeks; sex: male)	Transient MCAO (60 min)	DMF 15 mg/kg twice daily for 1, 2 or 3 consecutive days versus vehicle; alternatively, DMF in 0.08% Methocel™ 15 mg/kg twice daily via oral gavage for 1, 2 or 3 consecutive days versus vehicle	Lower BBB leakage and brain oedema (p < 0.01)
Lin[50]	Sprague–Dawley rats (age: n.a.; sex: male)	Transient MCAO (120 min)	DMF 25 or 50 mg/kg p.o. 2–3 h after transient MCAO and twice daily afterwards	Reduced infarct volume and improved neurobehavioural deficits 24 h, 72–84 h, 7 days and 14 days after MCAO (p < 0.05)
Yao[51]	C57BL/6 and Nrf2−/− mice (age: 8–10 weeks; sex: n.a.)	Transient MCAO (60 min)	DMF 30 or 45 mg/kg twice daily p.o. for 7 days, first dose given 15 min before reperfusion; Nrf2−/− mice were treated with 45 mg/kg only; control groups received PBS	Reduced infarct volumes in the 30 mg/kg (p < 0.05) and the 45 mg/kg DMF group (p < 0.01) on days 3 and 7 in C57/BL6 mice, not in Nrf2−/− mice (p > 0.05); functional improvement on days 3 and 7 in both DMF groups (p < 0.05 to p < 0.01), not on day 1 (p > 0.05)
Safari[52]	Sprague–Dawley rats (age: n.a.; sex: male)	Transient MCAO (60 min)	DMF 15 mg/kg diluted in 200 µl 0.08% Methocel™/H2O twice daily p.o. on days 0–14 (first application immediately after MCAO) versus vehicle versus sham treatment	Functional improvement on days 10 and 14 (p < 0.05)

BBB, blood–brain barrier; DMF, dimethyl fumarate; ICH, intracerebral haemorrhage; i.p., intraperitoneally; i.v., intravenously; MCAO, middle cerebral artery occlusion; n.a., not applicable; PBS, phosphate-buffered saline; p.o., per os; MAFG, musculo-aponeurotic fibrosarcoma-G; siRNA, small interfering ribonucleic acid.

Table 6.

Characteristics of clinical studies.

Author	Study design	Population	Stroke type	Intervention	Major end points	Major results
Natalizumab
Elkins[53]	RCT	Intervention = 79 (mean age 70 ± 14 years; sex: 51% male), control = 82 (mean age 72 ± 12 years; sex: 59% male)	First ischaemic stroke	Natalizumab 300 mg i.v. versus placebo up to 9 h after stroke onset	Primary end points: change in infarct volume from baseline to day 5; secondary end points: change in infarct volume 24 h to days 5 and 30, functional scores, others	No difference between natalizumab and control group regarding primary end point (p > 0.05); improvement in some of the functional scores (p < 0.05)
Fingolimod
Fu[54]	Prospective two-arm, evaluator-blinded study	Intervention = 11 (mean age 61 ± 12 years; sex: 36% male), control = 12 (mean age 58 ± 9 years; sex: 91% male)	Primary supratentorial basal-ganglia ICH, volume 5–30 ml, onset less than 72 h prior to admission, GCS ⩾ 6	Fingolimod 0.5 mg p.o. once daily for 3 consecutive days, max. 1 h after baseline CT scan, max. 72 h after symptom onset	GCS and NIHSS on day 7, 14, 30, and 90; haematoma volume and perihaematomal oedema volume on days 7, 14 and 90	Lower NIHSS scores at 7, 14 and 30 days (p = 0.03 to p < 0.001), lower oedema volume on day 7 (p = 0.04)
Fu[55]	Single-centre, open-label, parallel-group, evaluator-blinded, pilot trial	Intervention = 11 (mean age 62 ± 8 years; sex: 73% male), control = 12 (mean age 55 ± 11 years; sex: 82% male)	Acute ischaemic stroke in anterior circulation, NIHSS ⩾ 5, age ⩾ 18 years, symptom onset to admission 4.5–72 h	Standard treatment according to AHA guidelines + fingolimod 0.5 mg p.o. once daily for 3 consecutive days beginning within 1 h after baseline MRI and no later than 72 h after symptom onset versus standard treatment only	NIHSS, mRS, mBI and lesion volume (MRI) at different time points until day 90	Reduced NIHSS at day 30 (p = 0.049) and day 90 (p = 0.019); reduced infarct volume increase until day 7 (p = 0.0003)
Li[56]	Prospective two-arm, evaluator-blinded study	Intervention = 11, control = 12 (age and sex distribution: n.a.)	ICH patients with matched clinical characteristics, haematoma location, and volume	Fingolimod 0.5 mg p.o. once daily for 3 days, the first dose given within 1 h after baseline CT + standard of care for ICH versus standard management only	Changes of lymphocyte subsets, serum cytokines and impact on vascular permeability	Significant reduction in various immune cells and cytokines (p < 0.05 to p < 0.001)
Zhu[57]	Randomized, open-label, evaluator-blinded, multicentre pilot trial	Intervention = 22 (mean age 60 ± 3 years; sex: 59% male), control = 25 (mean age 59 ± 2 years; sex: 68% male)	First-ever hemispheric ischaemic stroke, age 18–80 years and NIHSS > 5	Alteplase (0.9 mg/kg) versus alteplase + fingolimod 0.5 mg p.o. once daily for 3 consecutive days with the first dose being given before alteplase administration	Primary end points: changes in lesion volume from baseline (DWI) to day 1 (FLAIR), the haemorrhage volume (GRE) at day 1 and extent of clinical improvement at day 1 (NIHSS); secondary outcomes: lesion volume growth from days 1 to 7, recovery at day 90	Better outcome regarding all primary end points (p < 0.05) in the fingolimod group; reduced lesion volume growth (p < 0.01); decreased NIHSS scores from days 1 to 7 (p < 0.01) and good recovery (mRS at day 90, p = 0.01)

AHA, American Heart Association; CT, computed tomography; DWI, diffusion-weighted imaging; FLAIR, fluid-attenuated inversion recovery; GCS, Glasgow Coma Scale; GRE, gradient echo sequences; ICH, intracerebral haemorrhage; i.v., intravenously; max., maximum; mBI, modified Barthel Index; MRI, magnetic resonance imaging; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; p.o., per os; RCT, randomized controlled trial.

Literature search and data extraction

A literature search was conducted on 16 October 2017 including MEDLINE (via PubMed) and the Cochrane Central Register of Controlled Trials. In addition, the reference lists of the included studies were reviewed to identify further studies. We continued the literature search until no further publications were identified. Four reviewers (ZM, MD, VP and PK) independently screened each title and abstract. Studies published until 16 October 12017 were considered. In the case of disagreement regarding study eligibility, a consensus meeting was arranged.

The databases were searched combining extensive search strings with the following Boolean operators: (mitoxantrone OR azathioprine OR ‘glatiramer acetate’ OR glatiramer OR ‘interferon beta’ OR ‘Peginterferon beta-1a’ OR ‘pegylated interferon’ OR FTY720 OR fingolimod OR natalizumab OR ‘α4 integrin’ OR daclizumab OR ocrelizumab OR cladribine OR teriflunomide OR ‘dimethyl fumarate’ OR fumarate OR alemtuzumab) AND (stroke OR ‘ischemic stroke’ OR ‘hemorrhagic stroke’ OR ‘experimental stroke’ OR ‘cerebral ischemia’ OR ICH OR ‘brain hemorrhage’ OR ‘cerebral hemorrhage’ OR ‘cerebral infarction’ OR ‘ischemia-reperfusion’). The compounds [mitoxantrone, azathioprine, glatiramer acetate, interferon β (IFN-β), fingolimod, natalizumab, daclizumab, ocrelizumab, alemtuzumab, cladribine, teriflunomide, dimethyl fumarate (DMF)] were chosen based on their approval for treatment of patients with multiple sclerosis in the European Union. FTY720 is used as a synonym for fingolimod. Mouse CD49d-specific antibodies and selective anti-α4-antibodies equate to natalizumab in humans. Monomethyl fumarate, as the main metabolite of DMF, has been evaluated in preclinical stroke studies,[5] but is not approved for multiple sclerosis and, therefore, not part of this review. At the time of the literature search, ocrelizumab has not been approved in the European Union (EU). As approval was expected in the EU soon, we decided to include ocrelizumab in the literature search.

Extracted data included species, stroke model, intervention and major outcome in the rodent studies (Tables 1–5), as well as study design, population, stroke type, intervention, major end points and major results in the clinical trials (Table 6).[6-57]

Results

The database literature search identified 624 papers. Five additional publications were found after screening of the reference lists. Of these 629 publications, 552 papers were excluded after abstract review with regard to inappropriate content. The 77 remaining articles were reviewed on a full-text basis. Further, 25 of them were excluded due to the study design or other violation of inclusion criteria. Finally, 52 studies met our eligibility criteria and were included in the review (Figure 1, Tables 1–6).

Figure 1.

Preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow diagram.

Description of included studies

Of the 52 included studies, 47 were animal studies (Tables 1–5)[6-52] and 5 were clinical trials (Table 6).[53-57] Study characteristics and interventions are summarized in Tables 1–6.

Description of preclinical studies analysing glatiramer acetate

We identified four studies that met our inclusion criteria.[6-9] In all of them, a transient middle cerebral artery occlusion (MCAO) was performed and in one, an additional permanent MCAO[7] was performed. Two studies used mice,[7,8] the other, two rats.[6,9] Glatiramer acetate application (dose, route, time point), read-out times, as well as outcomes differed between the studies. Two of the studies showed stroke volume reduction;[6,9] the other two did not.[7,8] A more detailed synopsis can be found in Table 1.

Description of preclinical studies analysing interferon β

Six studies met our inclusion criteria.[10-15] Different stroke models have been used including transient MCAO,[11-14] permanent MCAO,[13] photothrombotic stroke[15] and a clot embolus model.[10] Two studies used mice,[14,15] three used rats,[11-13] and one used rabbits.[10] IFN-β (dose, route, time point), read-out times, as well as outcomes differed between the studies, with four studies showing a reduction in stroke volume[10-12,14] and one not.[13] Cruz and colleagues provide evidence that the anti-inflammatory and stroke-protective effect of IFN-β is lost in mice lacking interferon regulatory factor 2 binding protein 2 (IRF2BP2).[15] A more detailed synopsis can be found in Table 2.

Description of preclinical studies analysing fingolimod

We identified 23 studies that met our inclusion criteria.[16-38] Sixteen studies analysed ischaemic stroke using transient[16-23,26,28,30,32-34] or permanent MCAO,[22,23] a thromboembolic stroke model[27] or photothrombotic stroke.[29] Seven studies investigated ICH.[24,25,31,35-38] A broad spectrum of mice, as well as Sprague–Dawley rats have been used throughout the studies. Fingolimod (FTY720) treatment varied between the studies regarding dose (0.24–3 mg/kg),[17,23,32] application route and time. The majority of studies evaluating IS described FTY720-related reduction in stroke volumes.[16-21,23,26,27,30,32,33] A more detailed synopsis can be found in Table 3.

Description of preclinical studies analysing natalizumab

We identified eight studies that used different MCAO models[39-43,45,46] or an ICH model[44] and analysed different rat strains[39-41] or mice.[42-46] Five of the IS studies described a reduction in stroke volume associated with antibody-mediated α4 integrin blockade,[39-41,42,46] Langhauser and colleagues did not.[43] Llovera and colleagues found the type of MCAO model used (transient versus permanent) to be crucial for stroke volume reduction.[45] Hammond and coworkers evaluated α4 integrin blockade in an ICH model and presented evidence of improvement in functional outcome.[44] A more detailed synopsis can be found in Table 4.

Description of preclinical studies analysing dimethyl fumarate

We detected six studies that performed an experimental ICH by injection of collagenase[47] or autologous blood,[47,48] or used transient MCAO as a model of IS.[49-52] The intervention (dose, timing and route of administration of DMF) differed between the studies. All studies were positive in at least one outcome variable, including function,[47,48,50-52] stroke volume[50,51] and brain oedema.[49] A more detailed synopsis can be found in Table 5.

Description of clinical studies

In total, five clinical trials have been identified that met our inclusion criteria.[53-57] Of these, only the study of Elkins and colleagues is a double-blinded RCT;[53] the others are single-blinded studies.[54-57] Elkins and coworkers evaluated natalizumab 300 mg intravenously in patients with acute and first IS (n = 161). Despite promising data in most of the preclinical studies (see above), the primary end point remained negative.[53] In contrast, all of the studies that analysed fingolimod in IS[55,57] or ICH[54,56] (n = 23–47) reached their end points, including functional outcome[54,55,57] and reduced infarct volume increase.[55,57] A more detailed synopsis can be found in Table 6.

Discussion

In this systematic review, we found that immunotherapy in preclinical IS and ICH improved clinical and paraclinical outcome variables in most of the studies. As a limitation, the preclinical trials are very heterogeneous in design and used different stroke models, different occlusion times of the MCAO model, different doses of the immunotherapeutic drug, distinct time points of treatment and different application routes. Therefore, the comparability of the studies is very low and calculation of a meta-analysis regarding major outcome variables is not possible. The heterogeneity of the studies can also contribute to discrepant results in preclinical trials, which can be paradigmatically seen in studies regarding the role of regulatory T cells in IS.[58] Very often, the methodological quality of preclinical studies is low compared with clinical trials, and blinding and randomization procedures are not common in every laboratory, potentially leading to biased results. Moreover, only one preclinical study identified in our review analysed female animals that definitely does not represent the typical stroke population.[37] Methodological limitations might be one of the reasons for translational roadblocks, that is, difficulties in confirming positive preclinical results in clinical trials. Standardization of animal studies,[59] adherence to the Animal Research: Reporting of In Vivo Experiments guidelines (available at: https://www.nc3rs.org.uk/arrive-guidelines) and multicentre animal RCTs can improve data quality.[45]

Encouraged by positive preclinical trials, the first clinical trials evaluating natalizumab and fingolimod in stroke patients have been conducted. From a methodological perspective, the RCT by Elkins and coworkers is the best of these studies, but remained negative regarding the primary end point, with slight treatment-associated benefits on functional outcomes (ACTION trial).[53] In contrast, the trials analysing fingolimod were not double blinded and much smaller, but positive regarding major outcome variables.[54-57] The latter trials included mainly Asian patients; therefore, generalizability of data might be limited. In summary, the main limitation of the clinical trials is the heterogeneity of the included studies, the restricted data quality and generalizability, as well as the, in part, very low numbers of patients per study. Study heterogeneity comprises mainly population (IS versus ICH) and outcome variables.

The identified studies analysed immunomodulatory treatment with INF-β, glatiramer acetate, fingolimod, natalizumab and DMF. Until now, it is only incompletely understood what the stroke-specific mechanisms of these agents are. Therefore, the following aspects known from multiple sclerosis treatment might be the most relevant effects, but also other mechanisms might play an important role. INF-β inhibits IFN-γ, induces interleukin 10 expression and reduces the transmigration of lymphocytes and monocytes into the central nervous system (CNS). Glatiramer acetate (among other mechanisms) induces protective TH2 cells that secrete immunomodulating cytokines like interleukin-4, -6 and -10. Fingolimod is a sphingosine-1-phosphate-analogon that inhibits the efflux of lymphocytes out of lymph nodes leading to a profound lymphopenia and thus reduced CNS infiltration. Moreover, fingolimod seems to reduce thromboinflammation and improves cerebral blood flow.[26] The monoclonal antibody natalizumab blocks the adhesion molecule α4-integrin that is relevant for the infiltration of immune cells over the blood–brain barrier into the CNS. Finally, DMF has an antioxidant effect and activates the nuclear factor (erythroid-derived 2)-related factor 2 (Nrf2) pathway.

In conclusion, immunotherapy in stroke instrumentalizes the concept of thromboinflammation and could become a novel treatment option in the future. Despite translational limitations, the available clinical data are promising. Nevertheless, given the heterogeneity and low number of clinical studies, it is too early to reliably judge the novel strategy of immunotherapy in general. Therefore, further well-designed trials are urgently needed and are on the way (e.g. ACTION 2 and FAMTAIS[60]) [ClinicalTrials.gov identifiers: NCT02730455 and NCT02956200].