AECs: Amnion Epithelial Cells; HIE: Hypoxic-Ischemic Encephalopathy; HTH: Hypothermia Therapy; MSCs: Mesenchymal Stromal Cells; UCO: Umbilical Cord Occlusion

Neonatal encephalopathy stemming from a hypoxic-ischemic (HI) injury is a significant cause of death and disabillity in the neonatal period [1,2]. For the past decade, cell therapies have been a key focus of preclinical research for the treatment of neonatal HI injury. Recently, I, along with my colleagues, published a systematic review and meta-analysis of animal studies of combined hypothermia and mesenchymal stem cell (MSC) therapy for the treatment of neonatal HIE. In doing so we wanted to raise awareness that the bulk of the research in this area was not clinically relevant. In the process, we frequently noted the admission of various authors that we currently do not know the ideal ranges for key variables of clinical importance, namely, cell dose, route of administration, dose timing, and number of doses. More importantly to consider, is that systematic investigation in pursuit of the perfect combination of these variables would take decades, so any translation to clinical practice will likely be a slightly modified protocol of some successful preclinical study. Acknowledging that our understanding of how these variables interact is incomplete, some trends appeared during the conduct of the review that should be considered when designing future preclinical studies. These are:

1. Higher cell doses appear more effective than low doses.
2. Intranasal appears and intravenous routes, and both are preferable to more invasive routes.
3. Initiating treatment with cells sooner is preferable to later.
4. Multiple doses appear more effective than a single dose.

Furthermore, these trends appear to be consistent between different cell types/preparations (i.e., mesenchymal stem cells, umbilical cord blood mononuclear cells, etc.). The following will discuss the evidence and counter evidence for these claims.

Direct examinations of cell dose suggest higher cell doses up to 10⁸ convey a greater therapeutic effect than smaller doses [3-5]. Though there must be an upper limit of efficacy and there may even be a dose at which we see toxicity; to date, no negative effect of cell therapies in animal models of HI have been attributed to a high dose. Greggio et al. [3] treated HI injured rats with 10⁶ and 10⁷ human umbilical cord blood mononuclear cells (UCBs) administered via the left common carotid artery. The group that received 10⁷ UCBs had significantly improved learning and long-term spatial memory impairments in the Morris water maze evaluated at nine weeks post-HI. De-Paula et al. (2012) [4] examined HI injured rats treated with 10⁶, 10⁷, 10⁸ of human UCBs into the jugular vein one-day post insult. After 8 weeks of transplantation, morris water maze test showed a significant spatial memory recovery at the highest UCB dose compared with vehicle treated rats. Furthermore, the brain atrophy was also significantly lower in the moderate dose (10⁷ UCBs) and high dose (10⁸ UCBs) groups than HI + vehicle animals [4]. Drobyshevsky et al. [5] tested the effect of an intravenous infusion of human UCBs at three doses - 5 × 10⁶, 2.5 × 10⁶, and 1 × 10⁶ in a validated rabbit model of cerebral palsy. Infusion of 5 × 10⁶ UCBs alleviated postural disturbances, restored the righting reflex, improved locomotion, reduced muscular tone, and dystonia. 2.5 × 10⁶ cells showed reduced but still significant improvement. No improvement was seen at 1 × 10⁶ cells. Lastly, Donega et al. examined the effect of two either 1 × 10⁶ or 2 × 10⁶ human mesenchymal stem cells (MSCs) delivered intranasally 10 days post insult in mice injured at postnatal day 9 [6]. They showed that 2 × 10⁶ MSCs decreased lesion volume, improve motor behaviour, reduced scar formation and microglia activity to a greater magnitude than the 1 × 10⁶ cells dose. The key takeaway here is that in every investigation to date examining cell dose, the group receiving the largest dose has had the greatest therapeutic effect, and never has the greatest dose had a detrimental effect. Extrapolating from the observation that 10⁸ human UCBs cells in rats, which weigh approximately 6 grams at birth [7] (approximately 1.6 × 10⁹ cells per kg) was therapeutically favourable, indicates that cell efficacy may continue to increase with doses of human UCBs cells up to doses as high as 3 × 10⁹ cells in the human population.

Intravenous or intranasal are preferable to more invasive routes

Various administration routes have been studied in animal models of HIE, including intranasal, intrathecal, intracardiac, subcutaneous, intraventricular, intravenous, intraarterial, and intraperitoneal administration [8]. Clinical practice favours the use of less invasive administration routes such as intravenous and intranasal administration for pragmatic reasons, and safety. Both routes have been shown to be safe and effective in animal models. Though intravenous administration is associated with low rates of pulmonary micro embolism, it appears that the safety concerns regarding micro embolism can be mitigated with slow infusion velocities (~0.17–0.2 ml/min in Sprague-Dawley and Wistar rats) [9,10]. Despite favourable risk profiles, both routes have issues regarding the biodistribution of cells following delivery. It is known that most of the cells delivered intranasally never cross the cribriform plate and remain trapped in the nasal cavity [11], and that the majority of cells delivered intravenously are trapped in various systemic organs including the liver, spleen and lungs [12,13]. To my knowledge only one study has directly examined the difference in neuroprotection between intravenous and intranasal routes. Robertson et al. directly compared the outcomes of piglets treated with hypothermia therapy (HTH) and hMSC delivered intranasal or intravenous. The group that received the cells intranasal had increased OLIG2 counts in the hippocampus, internal capsule, and periventricular white matter. Furthermore, intranasal HTH+MSC had reduced apoptotic cells in the internal capsule versus HT, compared with no effect in the intravenous group. Intranasal administration has the added benefit of reduced potential for side effects than intravascular administrations such as pulmonary embolism [14,15]. One option not yet investigated is that of multiple administration sites in the same subject, for example one intravenous to target the considerable systemic inflammation seen following global HI and one intranasal to target the neurological injury.

Treating sooner appears more effective than later

Donega et al. analysed the long-term effects of intranasal MSC treatment on lesion size, sensorimotor, and cognitive behaviour and determined the therapeutic window and dose-response relationships. Nine-day-old mice were subjected to unilateral carotid artery occlusion and hypoxia. MSCs were administered intranasally at 3, 10 or 17 days after HI [6]. The motor, cognitive and histological outcome was investigated. They identified 0.5 × 10⁶ MSCs as the minimal effective dose with a therapeutic window of at least 10 days. Similarly, Li and colleagues found that UCBs administered at 12 hours, but not 5 days, after HI injury in preterm fetal sheep was associated with enhanced myelination, in association with reduced numbers of microglia [16]. However, Davidson et al., [17] compared the effects of early and delayed intracerebroventricular administration of human amnion epithelial cells (AECs) in preterm foetal sheep at 0.7 gestation on brain injury induced by complete umbilical cord occlusion (UCO) or sham occlusion. Foetuses received 1 × 10⁶ AECs or vehicle alone, as an infusion over 1 hour, either 2 or 24 hours after UCO. AEC administration at 2 but not 24 hours was associated with improved neuronal survival in the CA1/2 region to sham-sham levels. By contrast, the 24 hour, but not 2-hour, AEC group showed improved neuronal survival in the striatum and thalamus compared to asphyxia-vehicle. AEC infusion at both 2 and 24 hours had dramatic anti-inflammatory and anti-gliotic effects, including significantly attenuating the increase in microglia after UCO in the white and grey matter and the number of astrocytes in the white matter. Both protocols partially improved myelination but had no effect on total or immature/mature numbers of oligodendrocytes. Neuronal survival in the hippocampus was increased by AEC infusion at either 2 or 24 hours, whereas only AECs at 24 hours were associated with improved neuronal survival in the striatum and thalamus. Meta-regression analysis performed by Archambault et al., showed an increased effect size in studies with a single dose <72h post insult compared with >72 h post insult. Thus, the majority of evidence stacks up in favour of treating earlier rather than later.

Multiple doses are more effective than a single dose

Several studies have shown that the multiple administrations of cells are neuroprotective [18-21]. Van Velthoven et al. demonstrated that a treatment with 10⁵ MSCs at 3 days post insult and 10 days post insult was superior to single administrations at day 3 and day 10 post insult in HI injured P9 mice. Zhu et al. showed improved the long-term functional outcomes of rats, increased mature oligodendrocyte counts, and decreased the number of reactive astrocytes and activated microglia numbers after HI-induced damage in the premature brain following intraperitoneal administration of 10⁶ cells for 3 consecutive days in postnatal day 3 rats [20]. Penny et al. [18], showed that treatment with repeated doses of UCBs is more effective than a single dose for reducing tissue damage, improving brain pathology, and restoring behavioural deficits following perinatal brain injury in the P7 rat model. To my knowledge, this is the only direct evaluation of multiple doses vs a single dose for the treatment on neonatal HIE. However, meta-regression analysis performed by Archambault et al. showed multiple treatments improved sensory motor function and cognitive function in studies using multiple doses, compared with a single dose >72h post insult, and a single dose <72 h post insult. For both comparisons, multiple doses had a greater effect size than single doses at either timepoint.

In conclusion, the observations collected here suggests that initiating treatment early, with multiple high cell doses via intra nasal route or intravenous may yield enhanced therapeutic benefit compared with other protocols. While the scientific consensus is far from settled, the observations are consistent between cell types/preparations and very few counter examples exist. Therefore, a protocol with these initial conditions may be a suitable foundation for designing future research into the therapeutic use of cell therapies for neonatal encephalopathy.

EJT wrote the manuscript, proofread, and revised the manuscript.

The authors have no disclosures to make.

Not required.

EJT is supported by an Australian Government Research Training Program Scholarship.