Cell-based therapy is emerging as a potential new approach to improve functional outcome following stroke, a single leading cause of adult mortality and long-term disability worldwide. At the time of writing, we are aware of at least three active phase I/II clinical trials using adult or fetal human stem/progenitor cells to treat ischemic stroke, and 14 documented as recruiting (http://clinicaltrials.gov; Table 1). This clinical activity is based on a decade of preclinical studies which have reported significant therapeutic benefits following cell-based therapy in rodent models of focal cerebral ischemia, a standard animal model for this insult. This concise review summarizes the current progress, technical variables, and key findings of these studies in the hope of encouraging better protocol design and conduct of future preclinical outcome studies to optimize translational success to the clinic. A PubMed search (see supporting information for keywords used) identified 244 citations, of which 45 met the criteria of this review (see supporting information for inclusion criteria).
In the ongoing quest for the optimum human cell type in regenerative medicine for stroke, mesenchymal stem cells (MSCs) and neural stem/progenitor cells (NSCs) at this point in time appear to offer the greatest clinical potential. 44.4% and 24.4% of studies that used MSCs and NSCs, respectively, all reported significant improvement in functional outcome. Cell-induced benefits, however, were not consistently correlated with a decrease in infarct size, that is, a positive structural outcome. In view of the substantial benefits these cell types offer in animal studies, there have been a small number of phase I/II clinical safety and efficacy studies ([1–3]; Table 1) which are inherently limited in evaluating therapeutic effects; phase III randomized placebo-controlled clinical studies with specific functional outcome measures in large cohorts of stroke patients would be the future steps for realistic clinical translation. Fewer preclinical studies have examined the use of other cell types including amniotic epithelial cells, dental pulp stem cells (DPSCs), and peripheral blood progenitor cells, with their efficacy at this stage yet to be reproduced in multiple laboratories. In order to better understand if there are specific advantages associated with each stem/progenitor cell type, “head-to-head” preclinical studies are required; only two have been undertaken to date.
There are two cell-based approaches to transplantation—allogeneic and autologous. Due to economic bias, autologous transplantation of stem/progenitor cells is often a less investigated approach, although in hematological practice this has been the preferred strategy for many years. Allogeneic transplantation is enticing in its scalability and enabling of “off-the-shelf” use for stroke therapy, and intriguingly, was more effective than the autologous approach in eliciting functional improvement . While the human brain is considered an immune privileged site, it remains uncertain as to the need for immunosuppression with this transplantation regime . Eight studies using human MSCs or NSCs in non-nude rats without immunosuppression reported their nonimmunogenic potential and immunoregulatory effects in the mediation of functional recovery.
Fundamental questions with regard to the optimal cell delivery route, dosage, and timing remain largely unresolved to date. Intravenous (53.3% of studies) and intracerebral (44.4%) administration were the most frequently performed in comparison to intraarterial (6.7%), intracerebroventricular (4.4%), and intrathecal (2.2%) routes. Direct introduction of cells into the brain parenchyma allows for most efficient yet invasive delivery, thus less aggressive and presumably safer delivery procedures via vascular routes are ostensibly more appealing, especially when the direct method of intracerebral cell transplantation did not appear to be superior in therapeutic efficacy over other indirect administration routes. Interestingly, bone marrow mononuclear cells introduced via intravascular infusion promoted similar therapeutic benefits independent of whether they were grafted intravenously or intraarterially, despite significant homing of cells to the brain at 2 hours postadministration in the intraarterial as compared to the i.v. approach . It is unknown to date the potential adverse effects of intravenously infused larger cells such as MSCs (13–19 μm)  entrapped in the pulmonary circulation, which is a significant but likely transient effect, and if this may counter positive therapeutic outcome. We recently found there was a 9% mortality rate with intracerebral transplantation (associated with the procedure) of adult human DPSCs at 24 hours post-stroke , which from a clinical perspective would counter any potential neurological benefits using this route of delivery in the acute stroke setting. The optimal cell dose required for maximum therapeutic efficacy remains to be specifically defined and may vary for different cell types. The total number of stem/progenitor cells engrafted into the brain and their long-term survival may not necessarily be important determinants to post-stroke functional improvement , hence “more may not be better.”
The optimal timing of delivery may relate to the underlying mechanisms of action that different stem/progenitor cell types may possess. It has been argued that unlike pharmacological (neuroprotective) treatment necessitating acute intervention, which has been a translational failure, cell-based (neurorestorative) therapy enables an extended treatment time window (during subacute and/or chronic phases of stroke). To understand how this may translate into realistic clinical treatment options, we propose the following future scenario—when a patient presents outside the 4.5 hour time window for effective thrombolysis, i.v. cell-based therapy may be pursued after 24 hours when the patient is medically stable, from an infective and a pulmonary perspective, so that intervention may be safe. Clearly this hypothetical scenario requires many more years of study to realize.
Intracerebral and i.v. cell delivery at 24 hours (55.6% of studies), 3 days (13.3%), 7 days (20.0%), and up to 4 weeks (4.4%) after experimental stroke reported significant enhancement of functional recovery, except for studies mostly using too low a cell dose (≤0.045 million cells for intracerebral, ≤0.5 million cells for i.v.) or too short an outcome assessment endpoint (7 days post-stroke). While the degree of functional improvement was not significantly correlated with timing of delivery, structural outcome appeared to be dependent on this variable, with 1.5% lower treatment efficacy for each day delayed . This may be attributable to the progressive downregulation of chemoattractant signals from the ischemic tissue to transplanted cells. Nonetheless, previous studies have shown that MSCs  and NSCs  administered at 24 hours post-stroke resulted in greater functional recovery in comparison to transplantation at 7 days. Transplantation of cells within 24 hours of stroke onset may expose them to acute toxicity and cell death, while delivery after 4 weeks in the rodent stroke model (extrapolating this to the human stroke brain this window may be measured in terms of a few months post-stroke) may encounter the significant inhibitory barrier of a peri-infarct glial scar. We would thus propose using cell-based therapy between the acute and chronic periods following stroke onset, although further studies need to be conducted in the chronic period to better understand the challenges to the outer window of therapeutic efficacy.
Another challenge to successful clinical translation and to enable more efficient interpretation of preclinical study results is the use of appropriate treatment controls. The large majority (73.3%) of studies used the vehicle solution in which cells were delivered as control, which is recommended ; however, in a few studies (11.1%) saline or phosphate-buffered saline (PBS) was used, which should be considered weak control groups. One of the earliest studies used PBS as control solution when cells were delivered in saline , and five studies reported using middle cerebral artery occlusion (MCAo)-only animals as control. There is also an argument that a different population of nonfunctional, viable cells such as fibroblasts should be used as treatment controls, with two studies using this.
We strongly propose that the successful clinical translation of cell-based therapy necessitates the use in preclinical research of scientifically relevant and reproducible functional neurological outcome assessments, which are sufficiently sensitive to dissect sensorimotor asymmetry related to the site of ischemia. The rotarod, for instance, permits compensatory behavior using the tail or stroke-unaffected limbs , which could potentially produce false-positive outcomes (11 out of 14 studies reported improvement). Proximal MCAo may not cause damage to the rodent dorsal hippocampus and, therefore, functional tests examining spatial memory/learning such as the Morris watermaze test (used in three studies) may not be apt in this stroke model . Assessment of global motor function, body coordination, or hind limb function alone (24.4% of studies) is evidently inadequate. All preclinical studies to date that used the step test (five studies), vibrissae-evoked forelimb placement (three studies), foot-fault test (four studies), and pellet-reaching task (one study), and the preponderance of studies that used the adhesive tape removal test (16 out of 17 studies), and limb placement (three out of four studies) to assess forelimb sensorimotor asymmetry have yielded positive outcomes. As there has been widespread use of STAIR  and STEPS  guidelines, we would contend that the scientific community also embarks upon discussion toward a consensus on reproducible and reliable functional outcome assessment.
Following a decade of preclinical research in cell-based therapy for ischemic stroke what can be concluded in regard to clinical translation? There is compelling evidence that, regardless of cell type and delivery route, exogenous cells demonstrate targeted migration to the infarct site. Such an intrinsic property of human stem/progenitor cells is highly attractive, as extending the therapeutic time window is the clinical challenge that confronts stroke health professionals and their patients. It has been suggested that the most desired outcomes occur when intravenously transplanted cells cross the blood brain barrier, enter the brain from the periphery, and eventually reach the injured brain site . Besides, the penumbra, with salvageable tissue, represents the neuropathological target for therapeutic intervention.
Stroke is a brain insult involving the loss of neurovascular unit integrity  instead of specific neural cell types and as such cell-based therapy is likely to benefit by a multifactorial approach to restore or enhance recovery following this insult. As demonstrated in this review and by others , there is ample preclinical evidence that cell-based therapy in stroke is beneficial; however, the cellular and molecular mechanisms of action which underlie these benefits are not well understood. There are likely five cellular processes involved: angiogenesis, neurogenesis, neuroprotection, neuroplasticity, and immunomodulation. Of the 26 (out of 45) preclinical studies that investigated the underlying mechanisms of action, 46.2% reported enhanced host angiogenesis (including endothelial cell proliferation, vascularization, and increased blood flow in the peri-infarct region), 34.6% observed increased endogenous neurogenesis (including proliferation of subventricular zone NSCs, migration to the ischemic border, and subsequent differentiation), 30.8% demonstrated neuroprotective effects (reduced host cell apoptosis and in some cases decreased infarct volume), while 19.2% and 7.7%, respectively, reported host white matter changes (including increased fiber density and axonal outgrowth) and neuronal plasticity (increased dendritic length and branching of cortical neurons). In addition, 26.9% demonstrated the immunomodulatory capacity of grafted cells (such as decreased astrocytic and microglial activation in the peri-infarct area) following cell-based therapy. Intriguingly, transplanted human umbilical cord blood cells have been found to mediate post-stroke neuroprotection by means of reduction in spleen size and attenuation of inflammatory cytokines . This type of data suggests that cell-induced recovery may also be whole body phenomena, which requires further investigation, particularly as intravascular routes of delivery means that cells deposit extracranially. It is now recognized that the parsimonious explanation of neural cell replacement leading to restoration of post-stroke function as a sole underlying mechanism of action needs to be extended and encompass the interconnected and interdependent endogenous regenerative processes [20, 21].
The observed cellular mechanisms of action are presumably mediated by a wide spectrum of cell-secreted soluble factors. These molecular factors likely act via a paracrine mechanism, whereby transplanted exogenous stem/progenitor cells reciprocally interact with the local microenvironment to promote functional benefits. A detailed analysis of all the proposed molecules is beyond the scope of this concise review; however, we submit that an advantage of cell-based therapy in comparison to previous single molecular treatment paradigms for ischemic stroke is the complex interplay of multifactorial cellular and molecular processes that underlie stem cell biology. In brief, we may categorize these molecules in relation to their mechanisms of action. With respect to molecules secreted by exogenous stem/progenitor cells per se the following factors have been identified: proangiogenic/vasculogenic (vascular endothelial growth factor and angiopoietin-2), neurotrophic (brain-derived neurotrophic factor, glial-derived neurotrophic factor, nerve growth factor, neurotrophin-3, and insulin-like growth factor-1), chemoattractive (stromal cell-derived factor-1, also known as CXCL12), and immunoregulatory (interleukin-1β, IL-1β, IL-6, and transforming growth factor-β1) factors. Interested readers are referred to other articles reviewing possible mechanisms of action underlying cell-based therapy [20, 22].
On the other hand, human embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs) have future promise in the treatment of neurological dysfunction due to their high proliferative and plastic capacities; however, tumorigenicity and immunogenicity associated with the use of human ESCs/iPSCs and ethical concerns relating to ESCs remain significant challenges to clinical translation. The original approach of genetic manipulation to generate iPSCs  also presents a number of challenging issues: clinical concerns regarding possible deleterious host genomic modifications, low reprogramming efficiency (currently at <1%), variation in the quality of iPSC clones, and epigenetic and genetic differences compared to ESCs . With the 2012 Nobel Prize in Medicine recognizing the reprogramming of mature cells to pluripotency and the worldwide investment into iPSC research, the potential clinical application of this embryonic-like stem cell in the treatment of stroke remains to be determined. At this juncture, the two main advantages of iPSCs appear to be for autologous transplantation and in vitro disease modeling. Patient-derived disease-specific iPSCs will likely be useful in elucidating disease pathogenesis and screening potential drug candidates with respect to a wide spectrum of neuropsychiatric disease such as Parkinson's disease and schizophrenia . As iPSCs are a relatively recent stem cell type in comparison to adult stem/progenitor cells, more extensive studies will be required to establish their relative therapeutic potential.
It is undeniable that the growing body of preclinical research in this field provides promising evidence of the efficacy of adult and fetal human stem/progenitor cells to enhance post-stroke functional recovery; however, it is critical that future studies are more focused to answer fundamental questions required to optimize translational success to the clinical arena. The few clinical trials that are underway may be early and undertaken without full preclinical data to answer key questions but there is a human and economic imperative for progress. We would strongly advocate international consensus recommendations for future preclinical research so that we may together successfully journey through the “valley of death”  to everyday clinical use for stroke victims worldwide.