• Open Access

    Mesenchymal stem cell stroke therapy: current limitations in its clinical translation

    Ylenia Pastorello 1†
    Mark Slevin 2,3†*

    Explor Neurosci. 2023;2:98–105 DOI: https://doi.org/10.37349/en.2023.00015

    Received: February 14, 2023 Accepted: April 26, 2023 Published: June 28, 2023

    Academic Editor: Ertugrul Kilic, Istanbul Medipol University, Turkey

    This article belongs to the special issue Neuroinflammation in the Ageing and the Injured Brain


    For more than a decade now, research studies, proof of concept work, and clinical trials have endeavored to understand how mesenchymal stem cells might be used to help protect, repair, and/or regenerate damaged brain tissue following stroke. To date, the majority of studies have not demonstrated significant improvements in either morbidity or medium-long-term outcome, although safety has been relatively well proven. Limitations are likely to be linked to the pathobiological complexity and seriousness of stroke tissue damage, low efficacy of treatment, and short half-life of bio-active proteins released by stem cells. This article will highlight the heterogeneity and limitation of completed studies and the current status of ongoing work. At the same time, the potential of other combinational type treatments, such as drug-loading and targeting, and the use of hydrogels is discussed.


    Mesenchymal stem cells, stroke, drug-loading, drug targeting


    Various in vivo stroke models have been utilized even going as far back as the early 2000s, particularly in rodent models of ischemic stroke and these initially showed significant promise. For example, Kang et al. [1], using a rat model of middle cerebral artery (MCA) occlusion (MCAO), showed significant improvement in motor function following lateral ventricle transplantation of adipose-derived mesenchymal stem cells (MSCs) after their differentiation into neural-like cells. In 2010, Leu et al. [2] used a similar model of stroke and administered 2 million adipose-derived MSCs intravenously and found that, over a period of 3 weeks, the infarction size in the test group was notably smaller and, in addition, motor function improved, and markers of cell apoptosis and inflammation were reduced. Since then, many others have repeated these promising studies, with no less than 53 articles appearing in the PubMed database in 2022 alone.

    The first human clinical trial result is identified as far back as 2005 when Bang et al. [3] performed a randomized controlled early phase II study during which 5 individuals who had suffered severe MCAO were given 100 million MSCs delivered intravenously approximately 5 weeks after the infarction. Monitoring over the forthcoming year showed improved outcomes measured using the Barthel and Rankin stroke scales with no serious side effects, thus presenting some reasonably compelling early evidence of potential therapeutic benefit. In this article, the most recent clinical studies will be analyzed and shortcomings and limitations to date described.

    Human clinical trial results to date

    Systemic delivery

    To date, still, the largest phase II clinical trial data comes from the MASTERS-1 Athersys Multistem study, which was a randomized double-blind, placebo-controlled investigation of i.v. delivery of either 400 million or 1.2 billion allogeneic bone marrow (BM)-derived stem cells within 48 h of the initial infarction in 33 centers from UK and USA [4]. All patients had received successful recombinant tissue plasminogen activator (tPA) therapy for recanalization.

    There was no significant difference in stroke outcome after 12 months; however, those patients who were initially treated within 36 h, particularly, did show some benefits compared to the placebo group, and a MASTERS-2 phase III clinical trial is currently ongoing (over 200 patients treated to date; no outcome data) [4]. The concept of this type of i.v. treatment is clearly not one of direct involvement of the stem cell mixture in cell replacement, but of the paracrine effects and combination of potentially anti-inflammatory and neuroregeneration-promoting growth factors, and cytokines known to be released from this type of cell that would offer protection, improve blood flow and angiogenesis, and support regeneration of peri-infarcted neural connectivity and plasticity [5].

    A similar ongoing but much smaller trial has been listed by Suda et al. [6], originating and the first of its kind in Japan. In this case, dental pulp stem cells are cultured in vitro and delivered as an allogeneic therapy i.v. at a dosage of either 100 million or 300 million within 48 h of ischemic stroke [National Institutes of Health Stroke Scale (NIHSS) from 5–20 at baseline]. No resulting outcome data is available yet to assess.

    Other completed i.v. delivered trials have been mostly early phase II studies. For example, Law et al. [7] delivered BM-derived autologous MSCs 2 million/kg body weight to 9 patients who suffered severe MCA and had NIHSS scores of between 10–35, within 2 months of the primary infarction. No difference in neurological recovery or functional outcome was seen compared with the placebo control group. Similarly, in the STARTING-2 trial, 39 individuals with severe MCA were treated with autologous MSCs (obtained from the BM) within 3 months of symptoms onset and no difference in recovery was seen at the 3-month end-point between treated and control groups, although neuroimaging revealed some protection against corticospinal tract degeneration possibly supporting motor function recovery [8, 9].

    Most recently, de Celis-Ruiz et al. [10] completed the AMASCIS phase II trial with only 4 patients receiving adipose-derived MSCs 1 million/kg body weight i.v. within 2 weeks of tPA recanalized moderate-severe ischemic stroke (NIHSS 8–20), and showed no improvement compared with placebo group over 24 months.

    It is important to note the differences in each of the trials related to numbers of cells delivered, timing of the delivery, severity of the initial stroke, and successful tPA recanalization (or not) amongst other parameters, making them very difficult to compare or correlate. A major limitation may of course be the ability of the MSCs to penetrate the blood-brain barrier (BBB) and exert a concerted targeted paracrine delivery effect over time at the infarcted region of the brain. Yarygin et al. [11] describe the function of the BBB and the consequences of its disruption including neurovascular unit disruption, oedema, and inflammation following ischemic stroke. Evidence is summarized by the inability of stem cells to pass even the disrupted barrier in vivo after i.v. or i.a. injection although some cells are able to temporarily adhere and remain within cerebral capillaries for up to 72 h [11]. Along the same theme, Bang et al. [12] measured circulating extracellular vesicle (EV) expression up to 3 months following i.v. injection of autologous BM-derived MSCs in 39 patients (versus placebo control group) following ischemic stroke (NIHSS 6–21), in the STARTING-2 trial. Their hypothesis that the EV carries the essential medicinal growth factors and cytokines that may help recovery was partially proven as they showed that circulating EV numbers correlated with improvement in motor function as measured by diffusion tensor imaging and magnetic resonance imaging [MRI; resting state functional MRI (rs-fMRI)]; however, whilst micro-RNAs also increased, trophic factor levels did not. EVs, including those secreted by MSCs, can pass freely through the BBB and hence, are candidates for a targeted drug delivery approach after stroke, but mechanisms to enable continued secretion and focused delivery of their anti-inflammatory and pro-regenerative cargo are yet to be found [13].

    Direct in loco injection

    A more targeted delivery approach involves direct injection or transplant of MSCs into the stroked region of the brain, and this has been attempted primarily in chronic stroke patients in several recent clinical studies [14]. Twenty million human neural stem cells were implanted intra-cerebrally (putamen ipsilaterally to the infarct) by stereotaxic injection into 23 stroke patients between 2–13 months after the original infarct (PISCES-2) [14]. Here, improvements in upper limb function were seen, as measured using the action research arm test (ARAT), remaining to the 12-month end-point of assessment, although, only in those individuals displaying residual function at the start of the study, and of course, a study limitation here is the lack of a placebo or non-treated control group. The first human study was carried out on perinatal arterial ischemic stroke (PAIS) in the PASSIoN open-label intervention study [15]. In this study, approximately 50 million BM-derived MSCs were administered intra-nasally in 10 neonates with PAIS. Without adverse events (AEs), whilst markers of inflammation remained elevated, improvements in pre-Wallerian changes to corticospinal tracts were seen in 60% of the patients at 3 months follow-up by MRI. Dehghani et al. [16] also showed the safety of intraparenchymal injection of allogeneic placenta-derived MSC exosomes in five patients with a mean NIHSS of 17.6 and a follow-up of 3 months. This represents the current clinical trial data available as of January 10, 2023 (Table 1).

    Comparative analysis of key factors in the selected clinical studies

    Delivery methodsStudyStudy typeStudy populationNo. of casesNIHSS score rangeSource of stem cellsNo. of injected cellsTime point of administrationDuration of follow-upImprovement of observation indicatorsExperimental constraints
    Systemic deliveryBang et al. [3]Randomized, controlled, early phase II clinical trial30–75 y.o.; MCA stroke57–14Autologous; BM50 million; two timesAt 32–41 days52 weeksHigher BI; lower mRS and NIHSSSmall sample size
    MASTERS-1 [4]Randomized, double-blind, placebo-controlled, phase II clinical trial18–83 y.o.; AIS678–20Allogeneic; BM400/1,200 millionat 24–48 h12 monthsHigher BI; lower mRS and NIHSSSmall sample size; expansion of time window from 24–36 h to 24–48 h
    MASTERS-2 [4]Randomized, phase III clinical trial≥ 18 y.o.; AISRecruitingData not includedAllogeneic1.2 billionat 18–36 h365 daysOngoingData not included
    J-REPAIR [6]Randomized, double-blind, placebo-controlled, multicentre, early phase II clinical trial≥ 20 y.o.; anterior circulation AIS425–20Allogeneic; dental pulp100/300 millionwithin 48 h366 daysOngoingSmall sample size; proof-of-concept study-further studies required
    Law et al. [7]Randomized, assessor-blinded, controlled, single-center, phase II clinical trial30–75 y.o.; MCA stroke910–35Autologous; BM2 million/kg body weightwithin 2 months12 monthsNo difference with control groupSmall sample size
    STARTING-2 [8, 9, 12]Randomized, prospective, open-label, controlled trial30–75 y.o.; MCA stroke396–21Autologous; BM1 million/kg body weightwithin 90 days3 monthsNo difference with control groupSmall sample size; open-label design; short follow-up duration
    AMASCIS [10]Randomized, double-blind, placebo-controlled, single-center, phase IIa pilot clinical trial≥ 60 y.o.; AIS48–20Autologous; AD1 million/kg body weightwithin 2 weeks24 monthsLess AEs; lower NIHSSSmall sample size
    Direct in loco injectionPISCES-2 [14]Prospective, open-label, single-arm, multicentre study> 40 y.o.; upper limb motor deficit after AIS23Arm score 2–4Allogeneic; neural20 millionat 2–13 months12 monthsImproved ARAT in 7 patients; BI in 8; mRS in 7Small sample size; lack of control group; open-label design
    PASSIoN [15]First-in-human, open-label, single-arm, single-center, early phase II, intervention studyNeonates (full-term) with MCA PAIS10Data not includedAllogeneic; BM45/50 millionwithin 7 days3 monthsImprovements in pre-Wallerian changes to corticospinal tracts in 60% of patientsSmall sample size; lack of control group; short follow-up duration
    Dehghani et al. [16]Randomized, prospective, single-center, pilot clinical trialMalignant MCA stroke; decompressive craniectomy candidates511–25Allogeneic; placenta-derived exosomes2 mL (356 µg/mL)within 48 h3 monthsDecreased mRS and NIHSS in 4 patientsSmall sample size; short follow-up duration
    Display full size

    y.o.: years old; BI: Barthel index; mRS: modified Rankin scale; AIS: acute ischemic stroke; AD: adipose tissue-derived

    Conclusions and future perspectives

    All of the above constitutes primary evidence of the relative safety of various modes of delivery of stem cells or their secreted exosomes to patients who have suffered from ischemic stroke. However, the heterogeneous nature of the individual studies together with the small numbers of individuals treated [with lack of Food and Drug Administration (FDA) or regulatory approval being a factor] does not support a strong hypothesis that such therapy has any significant efficacy. Considering the pathobiological complexity of the cellular and tissue associated events following stroke, it is likely that stratified, tailored, and targeted time sensitive approaches will need to be integrated into these therapies in order to provide meaningful improvements in outcome.

    The concept of drug targeting and delivery to the brain is not new and Bruch et al. [17] summarized the use of liposomes for successful delivery of pH-sensitive drugs through the BBB, maintaining a longer potential half-life than directly delivered substances. Similarly, either MSCs themselves or their secreted vesicles could be magnetized, for example, by charging with iron oxide nanoparticles, and then targeted with magnets to the damaged part of the brain, although so far this has only been tested successfully in rodent models of MCAO [18]. However, to date, the evidence does not suggest significant penetration of MSCs through the BBB following systemic delivery, but paracrine effects from direct secretion, for example, of anti-inflammatory cytokines, in addition to polarized stimulation of the innate immune response, are likely to result in some beneficial modification to the stroke or penumbral micro-environment.

    Regarding the heterogeneity of MSCs derived from different sources, there are small differences in general paracrine secretions, for example, comparing BM-derived, umbilical cord purified, and adipose tissue extracted cells. However, bigger differences are likely due to excessive passaging of cells to create the ‘X’ millions required in treatment protocols resulting in mutated aging cells. Further assessment of final therapeutics should be characterized for secretive patterns as MSCs are entirely capable of the pro-inflammatory phenotypical switch, for example, as seen in the adipose tissue of diabetic individuals [19].

    In addition, MSCs are vehicles that can themselves be modified/primed with drug-loading to create an additional dimension for combinational type therapies, where release times have been shown to be up to 72 h. For example, Paudyal et al. [20] showed that MSCs loaded with a cyclin-dependent kinase 5 (CDK5) inhibitor known as CDK5 inhibitory peptide (p5; an anti-apoptotic 24-residue peptide that blocks p35-CDK5 aberrant phosphorylation), when delivered intracortical adjacent to temporarily induced MCAO, led to significant improvements in spatial learning, memory, and motor function over a period of 4 weeks (determined using the Morris water maze).

    Specifically, whilst the mean infarct volume was not significantly reduced, the treatment also led to improvement in bilateral coordination and sensorimotor function (rotating pole), and asymmetry of forelimb usage (cylinder test). There was no effect on cutaneous sensitivity (adhesive tape removal test). Immunofluorescence staining with human cell-specific antibodies indicated a higher number of surviving transplanted cells in the peri-infarcted area of animals treated with human-adipose-derived MSCs (hADMSCs) + p5 compared with hADMSC only-treated or control animals, with a concomitant reduction in the number of phagocytic, annexin 3-positive cells.

    Other options for longer-term release of pharmacologically active substances in larger amounts may even include micro-fragmented adipose tissue, where, for example, the release of paclitaxel was sufficient to perturb the growth of murine xenografted mesothelioma [21, 22].

    When considering the maintenance/protection or recovery of peri-infarcted regions, neuronal plasticity and connectivity rely heavily on extracellular matrix (ECM) stability, and this is degraded and destabilized by proteinases and other molecules soon after stroke. A key molecule is hyaluronan, which plays the main structural role as a scaffold and is essential for remodeling and synaptic plasticity; hence, its breakdown into pro-inflammatory oligosaccharides by hyaluronidases after stroke must be addressed [23].

    In addition, its potential as a drug delivery natural hydrogel amongst others, suggests it may have an important role in future acute stroke therapy [24, 25]. Jiang et al. [26] recently showed that hyaluronic acid-based hydrogels containing MSC-derived EVs, and implanted into the ischemic mouse brain, significantly enhanced brain focal retention time (stabilizing the vesicles) and improved both angiogenesis and neurobehavioural recovery.

    Thus, in conclusion, even though these possibilities hold promise, there is still a very long way to go before an effective therapy for individuals after stroke can be optimized, and a much more inclusive concerted approach is needed in order to facilitate appropriate larger scale clinical trials.



    blood-brain barrier


    bone marrow


    cyclin-dependent kinase 5


    extracellular vesicle


    middle cerebral artery


    middle cerebral artery occlusion


    magnetic resonance imaging


    mesenchymal stem cells


    National Institutes of Health Stroke Scale


    perinatal arterial ischemic stroke


    tissue plasminogen activator


    Author contributions

    YP and MS equally contributed to: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing—original draft, Writing—review and editing.

    Conflicts of interest

    The authors declare no conflicts of interest.

    Ethical approval

    Not applicable.

    Consent to participate

    Not applicable.

    Consent to publication

    Not applicable.

    Availability of data and materials

    Not applicable.


    Not applicable.


    © The Author(s) 2023.


    Kang SK, Lee DH, Bae YC, Kim HK, Baik SY, Jung JS. Improvement of neurological deficits by intracerebral transplantation of human adipose tissue-derived stromal cells after cerebral ischemia in rats. Exp Neurol. 2003;183:35566. [DOI] [PubMed]
    Leu S, Lin YC, Yuen CM, Yen CH, Kao YH, Sun CK, et al. Adipose-derived mesenchymal stem cells markedly attenuate brain infarct size and improve neurological function in rats. J Transl Med. 2010;8:63. [DOI] [PubMed] [PMC]
    Bang OY, Lee JS, Lee PH, Lee G. Autologous mesenchymal stem cell transplantation in stroke patients. Ann Neurol. 2005;57:87482. [DOI] [PubMed]
    Hess DC, Wechsler LR, Clark WM, Savitz SI, Ford GA, Chiu D, et al. Safety and efficacy of multipotent adult progenitor cells in acute ischaemic stroke (MASTERS): a randomised, double-blind, placebo-controlled, phase 2 trial. Lancet Neurol. 2017;16:3608. [DOI] [PubMed]
    Guo B, Sawkulycz X, Heidari N, Rogers R, Liu D, Slevin M. Characterisation of novel angiogenic and potent anti-inflammatory effects of micro-fragmented adipose tissue. Int J Mol Sci. 2021;22:3271. [DOI] [PubMed] [PMC]
    Suda S, Nito C, Ihara M, Iguchi Y, Urabe T, Matsumaru Y, et al.; J- REPAIR trial group. Randomised placebo-controlled multicentre trial to evaluate the efficacy and safety of JTR-161, allogeneic human dental pulp stem cells, in patients with Acute Ischaemic stRoke (J-REPAIR). BMJ Open. 2022;12:e054269. [DOI] [PubMed] [PMC]
    Law ZK, Tan HJ, Chin SP, Wong CY, Wan Yahya WNN, Muda AS, et al. The effects of intravenous infusion of autologous mesenchymal stromal cells in patients with subacute middle cerebral artery infarct: a phase 2 randomized controlled trial on safety, tolerability and efficacy. Cytotherapy. 2021;23:83340. [DOI] [PubMed]
    Chung JW, Chang WH, Bang OY, Moon GJ, Kim SJ, Kim SK, et al.; STARTING-2 Collaborators. Efficacy and safety of intravenous mesenchymal stem cells for ischemic stroke. Neurology. 2021;96:e101223. [DOI] [PubMed]
    Lee J, Chang WH, Chung JW, Kim SJ, Kim SK, Lee JS, et al.; STARTING-2 Collaborators. Efficacy of intravenous mesenchymal stem cells for motor recovery after ischemic stroke: a neuroimaging study. Stroke. 2022;53:208. [DOI] [PubMed]
    de Celis-Ruiz E, Fuentes B, Alonso de Leciñana M, Gutiérrez-Fernández M, Borobia AM, Gutiérrez-Zúñiga R, et al. Final results of Allogeneic Adipose Tissue-Derived Mesenchymal Stem Cells In Acute Ischemic Stroke (AMASCIS): a phase II, randomized, double-blind, placebo-controlled, single-center, pilot clinical trial. Cell Transplant. 2022;31:09636897221083863. [DOI] [PubMed] [PMC]
    Yarygin KN, Namestnikova DD, Sukhinich KK, Gubskiy IL, Majouga AG, Kholodenko IV. Cell therapy of stroke: do the intra-arterially transplanted mesenchymal stem cells cross the blood-brain barrier? Cells. 2021;10:2997. [DOI] [PubMed] [PMC]
    Bang OY, Kim EH, Cho YH, Oh MJ, Chung JW, Chang WH, et al. Circulating extracellular vesicles in stroke patients treated with mesenchymal stem cells: a biomarker analysis of a randomized trial. Stroke. 2022;53:227686. [DOI] [PubMed]
    Ollen-Bittle N, Roseborough AD, Wang W, Wu JD, Whitehead SN. Mechanisms and biomarker potential of extracellular vesicles in stroke. Biology (Basel). 2022;11:1231. [DOI] [PubMed] [PMC]
    Muir KW, Bulters D, Willmot M, Sprigg N, Dixit A, Ward N, et al. Intracerebral implantation of human neural stem cells and motor recovery after stroke: multicentre prospective single-arm study (PISCES-2). J Neurol Neurosurg Psychiatry. 2020;91:396401. [DOI] [PubMed] [PMC]
    Baak LM, Wagenaar N, van der Aa NE, Groenendaal F, Dudink J, Tataranno ML, et al. Feasibility and safety of intranasally administered mesenchymal stromal cells after perinatal arterial ischaemic stroke in the Netherlands (PASSIoN): a first-in-human, open-label intervention study. Lancet Neurol. 2022;21:52836. [DOI] [PubMed]
    Dehghani L, Khojasteh A, Soleimani M, Oraee-Yazdani S, Keshel SH, Saadatnia M, et al. Safety of intraparenchymal injection of allogenic placenta mesenchymal stem cells derived exosome in patients undergoing decompressive craniectomy following malignant middle cerebral artery infarct, a pilot randomized clinical trial. Int J Prev Med. 2022;13:7. [DOI] [PubMed] [PMC]
    Bruch GE, Fernandes LF, Bassi BLT, Alves MTR, Pereira IO, Frézard F, et al. Liposomes for drug delivery in stroke. Brain Res Bull. 2019;152:24656. [DOI] [PubMed]
    Kim HY, Kim TJ, Kang L, Kim YJ, Kang MK, Kim J, et al. Mesenchymal stem cell-derived magnetic extracellular nanovesicles for targeting and treatment of ischemic stroke. Biomaterials. 2020;243:119942. [DOI] [PubMed]
    Pham DV, Nguyen TK, Park PH. Adipokines at the crossroads of obesity and mesenchymal stem cell therapy. Exp Mol Med. 2023;55:31324. [DOI] [PubMed] [PMC]
    Paudyal A, Ghinea FS, Driga MP, Fang WH, Alessandri G, Combes L, et al. p5 Peptide-loaded human adipose-derived mesenchymal stem cells promote neurological recovery after focal cerebral ischemia in a rat model. Transl Stroke Res. 2021;12:12535. [DOI] [PubMed] [PMC]
    Alessandri G, Coccè V, Pastorino F, Paroni R, Dei Cas M, Restelli F, et al. Microfragmented human fat tissue is a natural scaffold for drug delivery: potential application in cancer chemotherapy. J Control Release. 2019;302:218. [DOI] [PubMed]
    La Monica S, Coccé V, Bonelli M, Alessandri G, Alfieri R, Lagrasta CA, et al. Micro-fragmented fat inhibits the progression of human mesothelioma xenografts in mice. Curr Cancer Drug Targets. 2023;23:6638. [DOI] [PubMed]
    Al’Qteishat A, Gaffney J, Krupinski J, Rubio F, West D, Kumar S, et al. Changes in hyaluronan production and metabolism following ischaemic stroke in man. Brain. 2006;129:215876. [DOI] [PubMed]
    Shahi M, Mohammadnejad D, Karimipour M, Rasta SH, Rahbarghazi R, Abedelahi A. Hyaluronic acid and regenerative medicine: new insights into the stroke therapy. Curr Mol Med. 2020;20:67591. [DOI] [PubMed]
    Yu Q, Jian Z, Yang D, Zhu T. Perspective insights into hydrogels and nanomaterials for ischemic stroke. Front Cell Neurosci. 2023;16:1058753. [DOI] [PubMed] [PMC]
    Jiang Y, Wang R, Wang C, Guo Y, Xu T, Zhang Z, et al. Brain microenvironment responsive and pro-angiogenic extracellular vesicle-hydrogel for promoting neurobehavioral recovery in type 2 diabetic mice after stroke. Adv Healthc Mater. 2022;11:2201150. [DOI] [PubMed]