Stem cells can be obtained from myriad sources. They can differentiate into diverse cell populations. Pluripotent stem cells, such as embryonic stem cells, are more potent with the ability to generate multiple cell types compared to multipotent stem cells that are derived from human tissue and differentiate into only mature cells [23]. However, the pluripotent stem cells have the drawback of tumorigenicity and forming potential teratomas [24]. Stem cells are also produced endogenously in response to tissue injury and can migrate to the ischemic lesions. Preclinical studies by various groups have shown that stem cells can act through multiple mechanisms including but not limited to neuroprotection, neurogenesis, angiogenesis, synaptogenesis, immunomodulation, release of growth, and trophic factors [25, 26].

Leong et al. in an extensive review of preclinical studies using stem cells in rodent models of ischemic stroke (most commonly used species) noted that among different types of stem cells, mesenchymal stem cells (MSCs, also called mesenchymal stem cell) and neuronal stem/progenitor cells reported most significant improvements in functional outcomes [27]. A recent meta-analysis of 46 preclinical studies of MSCs in ischemic stroke by Vu et al. suggested very promising results [28]. In their meta-analysis, they concluded that MSCs showed improved outcomes across various animal species, time of delivery, degree of immunogenicity, total dose/concentration, and in the presence of other comorbidities. Intracerebral (ICR) administration route was better than IA which in turn was better than IV route. ICR administration, however, is an invasive procedure compared to the alternative IA and IV modes of delivery. In our latest preclinical studies, we have shown the efficacy of IA delivery of MSCs and its superiority over the IV route in a rodent model of ischemic stroke [22]. The pros and cons of all routes of administration of stem cells are discussed in subsequent subsections. The meta-analysis by Vu et al. also suggested that MSCs used in primates showed most improvement in effect sizes compared to rodent and murine models [28]. The easy procurability and harvesting of MSCs, and improvement in outcomes after administration in various species and animal models of ischemic stroke, have been supported by various groups [29, 30]. MSCs can be obtained with relative ease; their procurement is not hampered by ethical considerations, and thus they have several advantages over other cell types.

There are various sources for MSCs, including but not limited to bone marrow (BM), umbilical cord blood cells (UCBCs), stem cells/progenitor cells derived from adipose tissue, stem cells derived from placenta [31], and human adult dental pulp [32]. UCBCs are pluripotent and are a source for MSCs. They are considered to have high regenerative potential and their naive immunologic phenotype makes them preferable for transplantation. Human adult dental pulp stem cells (DPSCs) that are derived from molar teeth are multipotent and have the capacity to differentiate into neurons. A preclinical study showed improved neurological outcomes with DPSCs in stroke animal model [32]. Olfactory ensheathing cells and endogenous stem cells from the subventricular zone (SVZ) and hippocampus are other potential therapies shown in preclinical studies for ischemic stroke [27, 33, 34].

Complex signaling cascades and processes like vascular rolling, adhesion, endothelial transmigration, and migration through the extracellular space to the injury site are needed to recruit stem cells to the area of brain ischemia [35, 36]. Myriad chemical messengers are implicated in these processes. The cytokine stroma-derived factor (SDF)-1 from vascular endothelium and activated platelets acts on CXCR4 chemokine receptor on MSCs and influences the MSC mobilization and homing [37–41]. SDF-1 is regulated partially by the transcription modulator, hypoxia-inducible factor (HIF-1α) [42], matrix metalloproteinases (MMP), and nitric oxide (NO) [40, 41, 43–45]. Other chemokines like monocyte chemoattractant protein-1 (MCP-1) are also reported to have a role in the recruitment of MSCs [38]. Growth factors such as vascular endothelial growth factor (VEGF) and granulocyte colony-stimulating factor (G-CSF) are implicated in BM stem cell mobilization. G-CSF also promotes mobilization of hematopoietic stem cells (HSCs) and endothelial progenitor cells (EPCs) by decreasing SDF-1 expression in BM and CXCR4 expression on HSC [46–50].

Preclinical studies in stroke models suggest that MSCs themselves do not differentiate into neurons in vivo, however, promote differentiation of endogenous neural progenitor cells [27]. MSCs can secrete neurotrophic factors such as brain-derived neurotrophic factor (BDNF), basic fibroblast growth factor (bFGF), angiopoietin-2, transforming growth factor (TGF-B), and insulin-like growth factor (IGF-1). These factors promote the remodeling of the neurons, glia, and neurovascular cells, and enhance axonal sprouting and synaptogenesis [11]. MSCs also decrease neuronal apoptosis and activation of astrocytes to minimize scar formation, and promote angiogenesis by secreting VEGF, bFGF, nerve growth factor (NGF), and other angiogenic factors [13, 51]. Furthermore, MSCs also influence immunomodulation via TGF-B, NO and prostaglandin pathways, JAK-STAT signaling pathways and release of interleukin-6 [51]. Thus, stem cells could be considered to have pleiotropic effects by modulating numerous targets through multiple signaling cascades.

Radioimaging studies can be very helpful in tracking and localization of tracer-labeled stem cells which in turn can help in understanding homing mechanisms. Preclinical studies on rodent models [52, 53] as well as canine model [18] showed that bioluminescence imaging and superparamagnetic iron oxide labeling helped to track and localize IA-delivered stem cells to the ischemic areas after stroke. Other groups [16] also demonstrated the feasibility of magnetic resonance imaging (MRI) in canine model of ischemic stroke.

Therapies provided in the early hours after ischemic stroke are aimed at reducing the injury/limiting the infarct size. On the other hand, therapies started weeks to months after stroke are aimed at promoting remodeling and repair. MSCs and other stem cells may have potential for wide-ranging therapeutic time windows [27, 32, 54]. A compromised BBB after ischemic stroke may allow better delivery of the MSCs to the target ischemic lesion. On the other hand, the BBB may be intact up to 24 h after stroke and may deter the engraftment of target tissue. Similarly, in the chronic phase, remnant glial scar tissue is likely to obstruct the passage of stem cells [27].

We showed that the therapeutic window lasts at least 24 h in our preclinical studies using a rodent model of ischemic stroke [22] (Fig. 6.1). One benefit of early infusion of stem cells is that the inflammatory chemokines that are released after ischemic stroke can help in attracting the stem cells. Some researchers suggested that the effect of MSCs may be reduced if provided 7 days after stroke [55]. However, there are other preclinical IA studies that have shown that stem cell treatment showed significant enhancement of functional recovery when given up to 7 days after ischemic stroke [19] and increased angiogenesis [56]. It is understandable that preclinical studies mostly focused on early therapeutic window in order to demonstrate better outcomes. Nonetheless, the preclinical studies demonstrate a wide therapeutic window for stem cells compared to existing standard of care therapies. On the other hand, some preclinical studies have unequivocally demonstrated that the beneficial effects of stem cells extend for a long duration from 4 months [32] up to a year [14] from the time of administration after ischemic stroke. This is very encouraging for clinical trials, if these time frames in rodent models were converted to human years.

Preclinical studies point out the superiority of IA and IV routes of administration over direct implantation with ICR and intracerebroventricular (ICV) modes of delivery. Firstly, IA and IV routes are less invasive than ICR and ICV routes. IA may be considered more invasive than IV, however compared to all other routes (ICR, ICV, IV, intrathecal), the IA route should be able to deliver significantly increased number of cells with more uniform cell distribution in the target ischemic brain tissue [11, 52]. The verdict is unclear with many preclinical studies showing diverse results.

Many groups have also shown that IA and IV are equally effective for improving neurological recovery, decreasing cerebral damage, and promoting protection mechanisms in ischemic stroke [57–59]. Vasconcelos-dos-Santos et al. used bone marrow mononuclear cells (BM-MNCs) via IA and IV routes, and Zhang et al. used human umbilical tissue-derived cells via IA, IV, ICR, ICV, and intrathecal routes. Both groups showed improved outcomes with all routes of delivery of stem cells. This was echoed by Savitz et al. who used autologous BM-MNCs and showed no difference in outcomes between IA and IV routes [58]. These studies suggest that IA delivery of BM-MNCs may not be superior to their IV delivery in ischemic stroke. Also, Gutierrez-Fernandez et al. used allogenic MSCs in a permanent middle cerebral artery occlusion (MCAo) rat model and showed no difference in ICR and IV routes of administration on the functional outcomes [17]. The permanent occlusion of the internal carotid artery (ICA)-MCA on the side of IA administration would block direct delivery of MSCs to the infarct area and may be the crucial reason for the absence of superior of IA over IV route in efficacy in this study. On the other hand, several preclinical direct comparison studies between IV and IA routes have shown that IA delivery provides higher concentration of cells to the target tissue than IV delivery [15, 21, 53, 60, 61]. BM-MNCs when delivered via IA route showed better results than IV route on the size of infarct volume, target cell delivery, and motor function after ischemic stroke [21, 61]. Other showed this in ischemic stroke using human MSCs in rodent model [60] and mouse neural stem cells in a mouse model [53]. Guzman et al. have shown using bioluminescence imaging that IA injection results in significantly higher and more sustained cell delivery to the brain [53]. Du et al. [21] in a direct comparison study showed greater functional recovery and increased angiogenesis with IA route compared to IV route using human BM-MSCs in a rat ischemic model. Lundberg et al. proved this concept in the traumatic brain injury (TBI) model. They showed in a rat TBI model, using various cell populations, human MSCs, human neural progenitor cells, and rat neural progenitor cells, that targeted IA route is more effective than IV administration [61]. The superiority of IA route over IV route may partly be due to circumvention of cell sequestration in the peripheral filtering organs such as lung capillaries encountered with the IV route [62]. We feel that there are multiple critical factors that can influence superiority of IA route over IV route including but not limited to homing of stem cells in response to chemoattractants, cell migration, level of arteriosclerosis and resulting stenosis, surviving arterial arborization, and disruption of BBB after stroke.

IA delivery of cells, however, does pose certain safety hurdles that need to be addressed fully prior to clinical translation. Microvascular occlusions leading to decreased cerebral blood perfusion have been reported in preclinical studies with IA stem cell delivery. Chua et al. reported that micro-occlusions can be prevented by preserving anterograde flow [63]. Whether related to microvascular occlusions or not, IA delivery of neural progenitor cells and MSCs has been reported to worsen ischemia and increased mortality in some preclinical studies [64–66]. We investigated this issue in our preclinical studies using multiple concentrations of MSCs delivered via IA route [22]. Dose de-escalation studies with real-time monitoring of MCA flow showed that de-escalation of IA MSC cells doses to 1 × 10⁵ MSCs mitigated the decrease in MCA Doppler flow signal that was seen at higher doses (Figs. 6.2 and 6.3). This IA maximum tolerated dose (MTD) of 1 × 10⁵ MSCs given at 24 h post rMCAO also resulted in significantly superior attenuation of neurological deficits and reduction in infarct size after ischemic stroke as compared to IA vehicle and IV MSC groups (Figs. 6.1 and 6.4). Moreover, delivering MSCs using IA route, we also showed that the therapeutic window extended to 24 h. In fact, MSCs delivered at 24 h after ischemic stroke showed greater beneficial effects as compared to MSCs administered at 1 h. The findings were very promising because the benefits of arterial delivery of MSCs translated into long-term neurological improvement as well as reduction in infarct size (Figs. 6.1 and 6.4). Considering all published preclinical studies till date, the IA route of stem cells delivery remains promising with a number of clinically attractive advantages over other modes of delivery.

There has been a surge of early clinical trials of stem cells in acute ischemic stroke and interest has spiked across the globe in the last 5 years [34, 67]. It is of paramount importance to first establish clinical safety during the development of novel therapies. Rapid strides have been made over the past decade in establishing the clinical safety of stem cells in ischemic stroke patients. Kondziolka et al. first demonstrated the feasibility and safety of stem cells in ischemic stroke [68]. They directly implanted stem cells derived from a human teratocarcinoma cell line (NT2N, Layton BioScience Inc, Sunnyvale, CA) via stereotactic ICR delivery in ischemic stroke patients. In this phase I, non-randomized, observer-blinded study involving 12 patients with basal ganglia strokes and motor deficits, they showed that ICR transplantation of stem cells (2–6 × 10⁶) was feasible. All patients were immunosuppressed with cyclosporine A. Two isolated events of a generalized seizure in one patient and a distant brain stem stroke in another patient were thought to be unrelated to the stem cell injections. The 6- and 12-month follow-up positron emission tomography (PET) scans showed improved fluorodeoxyglucose uptake at the implant site in seven and three patients, respectively [68, 69]. Postmortem analysis on one patient who died of myocardial infarction 27 months after transplantation showed the presence of neurons derived from the transplanted cells at the implantation site [70]. Kondziolka et al. followed this study with a phase II, randomized study in nine patients with ischemic and hemorrhagic stroke each [71]. They treated seven patients in each group and four patients served as controls. They reported improved “Action Research Arm Test” gross hand-movement scores compared with the control and baseline values.