Cell therapy – the use of human cells to replace or repair damaged tissue or cells – is not new. The most widely practiced form, bone marrow transplantation, in which bone marrow cells from a donor are used to replace the hematopoietic system of the recipient, has been practiced since the 1950s. Over several decades, that technique, now referred to as hematopoietic stem cell transplantation (HSCT), has evolved considerably by expanding and refining donor cell sources to include mobilized peripheral blood and umbilical cord blood, enhancing accessibility and donor cell engraftment, and improving conditioning regimens and supportive care. What began as a therapy indicated only for leukemia has since found utility in immunodeficiencies, hemoglobinopathies, and certain inherited metabolic diseases, and the field continues to advance by utilizing the same cell sources to harness the power of the immune system as yet another strategy to combat disease. Is cell therapy a cure-all for these conditions? No. Though much improved, HSCT carries a risk of transplant-related morbidity and mortality, and as such is typically reserved to treat the toughest cases, those in which no other options exist. Some patients are cured of their disease, others relapse despite transplant. Some experience life-long side effects. Most patients who undergo transplant for metabolic diseases, such as Krabbe disease, metachromatic leukodystrophy, and X-linked adrenoleukodystrophy, are left with residual neurological deficits. Nonetheless, it is an effective – and the only potentially effective – therapy for many, and it is constantly improving to maximize the benefit to those in need.

Cells as a potential therapeutic modality for neurological conditions was first attempted in humans in the 1990s. Initial investigations focused on using donor cells as a replacement for damaged or failing host cells, analogous to bone marrow transplant, such as transplanting dopaminergic cells into the brains of patients with Parkinson disease. Work in this area is ongoing, with the focus shifting from fetal and embryonic stem cells (ESCs) to cell sources that are less controversial and more easily standardized. The field has since evolved to also conceptualize cells as therapeutic tools that can modify the local cellular environment and/or the function of existing host cells without engraftment, as opposed to replacing host cells. Along those lines, trials of cell therapy for adults with stroke began in the 2000s, followed by multiple sclerosis and amyotrophic lateral sclerosis in the 2010s.

This paper reviews the status of cell therapy investigations in two of the most common pediatric neurodevelopmental disorders, cerebral palsy (CP) and autism spectrum disorder (ASD), in which cell-based clinical trials began in 2009. It is not a comprehensive review of all work done in the field, but broadly covers the state of the science, ongoing challenges, and status of potential therapeutics. At its best, human neurodevelopment is an extremely intricate, complex process orchestrated with precision over years, starting from relatively few cells early in gestation and evolving into elaborate interconnected networks throughout childhood, adolescence, and even adulthood. Any kind of disruption to the neuroenvironment or typical sequence of events can create a litany of cascading, long-term effects. The demand for effective treatments for these complicated disorders is significant. Meeting it will require rigorous scientific study, novel approaches, and persistence. Cell therapy is a worthy attempt.

SOURCES OF CELLS AND POTENTIAL MECHANISMS OF ACTION

To be effective in treating neurological conditions, cellular therapies must repair, replace, or prevent further deterioration of damaged or defective neurological tissues/cells. In most scenarios, particularly including CP and ASD, generation of new donor-derived neurons or other cells in the nervous system is considered unlikely, and mechanistic approaches focus on stimulating tissue repair and maintenance via paracrine effects. Numerous different types of cells, each with unique properties and proposed mechanisms of action, have been studied in preclinical and/or clinical studies of neurodevelopmental disorders. Because of their favorable safety profile, easy accessibility, and noncontroversial methods of procurement, umbilical cord blood and mesenchymal stromal cells (MSCs) have been the most frequently studied cell types in clinical trials.¹

CEREBRAL PALSY

CP, a non-progressive insult to the developing brain manifested by physical impairment, has a multitude of causes. Current therapies include non-pharmacological measures such as physical, occupational, and speech therapies, orthotics, optimal nutrition, constraint therapy, and use of adaptive devices; pharmacological interventions such as oral pharmacological agents, botulinum neurotoxin A injections, and oral or intrathecal baclofen; and surgical options including dorsal rhizotomy and various orthopedic procedures focused on maximizing functional abilities and quality of life. Cell therapy as a potential treatment for CP is aimed at addressing CP caused by an acquired brain injury, as opposed to CP due to genetic conditions or congenital brain malformations. The therapeutic rationale is that cells may act through cell-signaling and paracrine mechanisms to decrease inflammation, provide neuroprotection, stimulate angiogenesis and/or synaptogenesis, support endogenous repair mechanisms, and promote the migration and proliferation of existing NSCs to prevent and/or repair injury and enhance neuroplasticity.

Numerous animal studies have demonstrated functional and survival benefits of cell therapy in models of brain injury. Most have been conducted in small animal models of acute injury such as stroke, hypoxia, and intraventricular hemorrhage, but a few larger animal models exist. A variety of cell sources have been administered to animals through various routes at different time points. Neuroprotection, neovascularization, and neuronal regeneration have all been demonstrated after cell administration in various models,^{11, 12} leading to the initiation of human clinical trials.

At least 20 clinical trials of cell therapy in children with CP or brain injuries at risk for developing CP (i.e. hypoxic ischemic encephalopathy) have been reported, and more are ongoing. Nine randomized studies have been published, one conducted in the United States,¹³ three in South Korea,^14-16 and five in China (Table 1).^17-21 There is substantial variability in all aspects of these studies. Cell sources were autologous or allogeneic and included umbilical cord blood, umbilical cord tissue, bone marrow, mobilized peripheral blood, or fetal neural progenitor cells. Cells were given intravenously, intra-arterially, intrathecally, or intraventricularly in a wide range of doses. Treated patients stretched in age from infancy to 35 years with a variety of functional abilities. Specific outcome measures and their timing also varied, though most included an established measure of gross motor function, commonly the Gross Motor Function Measure-66 or -88. A systematic review and meta-analysis of five of the studies detailing study-specific characteristics was recently published, concluding that there is significant gross motor improvement of limited magnitude after treatment with cell therapy.²²

Despite their differences, some trends can be observed across the published clinical trials. Overall, the safety profile of cell therapy treatments in children with CP is encouraging. Delivering cells directly into the central nervous system is more invasive and causes reactions, typically consisting of transient fever, more frequently than peripheral administration. Overall, however, there is a very low incidence of short-term serious adverse events across all cell types and routes. The highest incidence of serious adverse events was observed in the 2013 Chinese study of erythropoietin with or without intravenous allogeneic cord blood.¹⁴ This is the only study that included systemic immunosuppression (cyclosporine), and the observed adverse events, which were distributed evenly among treatment and placebo groups, were mostly infectious in nature. Long-term adverse events have not been reported in published series. While cord blood cells and MSCs are not expected to engraft in the recipient, making the likelihood of long-term side effects low, continued vigilance is warranted. Most studies report greater motor improvement in treated children than in placebo groups, though the differences are modest and variable. Associations have been made between increasing brain connectivity via magnetic resonance imaging and motor gains.²³ A dose effect is also reported in some preclinical and clinical studies, but differences in cell sources and outcome assessments make defining an optimal dose premature at this stage.

There are multiple issues that prevent drawing definitive conclusions regarding the efficacy of cell therapy for CP, but they coalesce around two central themes: variability and knowledge deficits.²⁴ Variations in the etiology, timing, and severity of acquired brain injuries contribute to a very heterogeneous population of children with CP, and each may impact the potential effectiveness of therapies. To date, no biomarkers have been able to distill this inherent variability or reliably predict the ultimate outcome of a child with or at risk for CP, with or without cell therapy. Additional work needs to be done in this area to allow for the identification of children who are likely to benefit from particular interventions, including cell-based ones.

What constitutes a meaningful improvement? As the injury in CP is static, but a child’s developing brain most certainly is not, children with CP make gains over time. Some progress has been made in attempting to predict longitudinal motor change,²⁵ but there is still substantial variability among children, making it difficult to know how much improvement is attributable to any given treatment beyond the natural history of the condition. Although CP is defined as a motor disability, non-motor comorbidities are common and can impact a child’s abilities as much, if not more, than their motor challenges. While the scientific community is increasingly recognizing the need to include multiple domains in intervention assessments, how best to do so is still a work in progress. In their Common Data Elements for CP,²⁶ the National Institute of Neurological Disorders delineates 13 domains of outcomes and endpoints (cognitive and emotional status; executive functioning; functional outcomes; memory; motor function; participation; quality of life; social emotional; spasticity/movement; speech, language, and communication; trunk control/balance; visual-spatial processing; family and environment), and consideration should be given to including recommended assessments from as many domains as relevant and feasible in clinical trials of CP.

There are, of course, also several sources of variability and knowledge gaps specific to cell therapies for CP. First and foremost involves mechanistic understanding. Significant progress has been made in preclinical models of various types of brain injury, and the data overwhelmingly demonstrate that cells operate through trophic and paracrine mechanisms to modify neuroinflammation,²⁷ provide neuroprotection,²⁸ and facilitate angiogenesis, synaptogenesis, and neurogenesis.^{29, 30} Precise, molecular-level mechanisms, not surprisingly, have yet to be fully elucidated. The developing brain, its response to injury, and the multitude of factors involved in its recovery is perhaps one of the most complex unsolved mysteries in modern medicine – a true, comprehensive understanding of that process and how cells can mitigate it will not be achieved quickly. There is ample preclinical efficacy data and early phase human safety data, however, to warrant further clinical trials while mechanistic research continues. As discussed above, consistency in the eligibility and outcome measures is key to making clinical studies interpretable, useful, and reproducible. With that, well-designed clinical trials can answer questions regarding the efficacy of cell therapy, including the variables of optimal cell type, dose, timing, route of administration, and which patients are most likely to benefit.

AUTISM SPECTRUM DISORDER

Unlike CP, in which a causative brain injury can often be identified, the etiology and pathogenesis of ASD is still largely unknown. While it is generally accepted that environmental, epigenetic, and inflammatory factors all contribute to the development of ASD, their specific interactions and the neurobiological mechanisms underlying the disorder are not completely understood. Disruptions in synaptic functioning, microglial activation, and proinflammatory cytokines have all been implicated, and may all play a role either simultaneously or differentially among individuals. Many cytokines and molecules classically associated with immune regulation are now also recognized as playing a role in typical neurodevelopment. This dual functionality may prove to be an important link in the association of immune-related changes and atypical neurodevelopment in ASD. Immune mediated mechanisms in ASD have been suggested by associations with family history of autoimmune conditions,³¹ increased frequency of certain human leukocyte antigen haplotypes in individuals with ASD,^{32, 33} and abnormalities in the number, function, and gene regulation of microglia in models of ASD.^{34, 35} One common hypothesis is that inflammation and/or immune activation may act to modify expression of ASD risk genes or otherwise disrupt the processes of typical neurodevelopment, leading to the development of ASD.

As with other neurological conditions, the biological and pathophysiological complexities of ASD make a cellular approach to therapeutics appealing. Given the potential role of ongoing immune dysregulation and/or inflammation in ASD and the known immunosuppressive and immunomodulatory properties of MSCs, they are receiving attention as a potential therapeutic cell source. In ASD, it is hypothesized that MSCs may act through paracrine and trophic mechanisms to favorably modulate ongoing inflammation or immune pathology in the brain,³⁶ suppress microglial activation,^{37, 38} protect neurons from further damage, and/or allow restoration of functional neuronal circuitry.^{39, 40} Because of shortcomings in our current understanding of ASD and ability to model it, basic science data to support these proposed mechanisms of action is limited.

Modeling ASD, a complex, heterogeneous condition that affects human-specific constructs such as socialization and communication, is incredibly difficult. Mouse models of single gene disorders associated with ASD symptoms, such as Fragile X, Rett, and tuberous sclerosis complex syndromes, have been used to study ASD, but they represent only a small portion of affected individuals and may not be representative of the wider ASD population. Neuroimaging is attractive as a potential biomarker, and human studies suggest altered brain connectivity in individuals with ASD including long-range underconnectivity,⁴¹ but other patterns have also been reported and correlations to ASD phenotypes and severity require additional study. Electroencephalogram (EEG) changes, commonly including decreased alpha power in the frontal areas, have been observed but are not yet sensitive or specific enough to be used for diagnostic purposes. While human brain tissue samples may provide clues to the pathophysiology and cellular roots of ASD, such samples are, for obvious reasons, rare. To circumvent these issues, some investigators have utilized iPSCs to generate neurons, oligodendrocytes, and microglia from individuals with ASD to study their cellular and molecular characteristics. Differences in neuronal morphology, gene expression, and electrophysiology, for example, have been demonstrated in iPSC-derived neurons from children with ASD compared to their unaffected siblings.⁴² While this is an attractive technique, small sample sizes along with differences in culture methods and the underlying heterogeneity of the condition prevent generalizable conclusions at this stage. Thus, though fraught with similar issues of variability and lack of well-established biomarkers, human trials are currently the most reliable means to study therapeutic interventions in ASD.

As of May 2020, 15 studies of cell therapies in ASD were registered on ClinicialTrials.gov in seven different countries. All are phase 1 or 2 studies, mostly single-arm and open-label, with cell sources including umbilical cord blood (n=4),^43-45 MSCs (n=3),^{46, 47} and autologous bone marrow (n=1)⁴⁸ (one study utilized both cord blood and MSCs).⁴⁹ Results are available for seven trials (Table 2).^{43-47, 49, 50} The single-arm studies demonstrate the relative safety of administering cellular products to children with ASD and generally report improvement on some measures of ASD symptoms in at least a portion of the treated participants. Without a control group for comparison, of course, any gains cannot be attributed to the cells themselves. Two randomized controlled trials have been published. In one study,⁴⁹ children were treated with four doses of either umbilical cord blood mononuclear cells given both intravenously and intrathecally (n=14), umbilical cord blood mononuclear cells plus intrathecal cord tissue MSCs (n=9), or standard therapy (n=14). At 6 months post-treatment, the combination group demonstrated greater improvement in the Childhood Autism Rating Scale than placebo or umbilical cord blood mononuclear cells groups, and both treated groups improved more than the placebo group on the Clinical Global Impression scale. In another recently published study,⁴⁵ 180 children aged 2 to 7 years were randomized to treatment with umbilical cord blood (autologous if available, allogeneic unrelated donor if not) or placebo. While there was no difference in improvement of social communication or ASD symptoms between treatment and placebo groups as a whole, children without intellectual disability who were treated with cord blood demonstrated improved communication skills on the Vineland Adaptive Behaviors Scale, increased sustained attention in eye tracking assessments, and increased in alpha and beta power on EEG. Limitations of the trial included a high expectancy effect, exhibited by greater than expected improvements in the placebo group, and difficulty assessing cognitive function, particularly in younger participants, resulting in accrual of a lower number of children without intellectual disability than projected thereby compromising the planned study analysis. These highlight some of the challenges inherent in conducting clinical trials in children with ASD.

PUBLIC AND SCIENTIFIC ANGST

Along with increasing scientific interest in cell therapies has come increasing public interest.⁵¹ Fueled by a combination of slick, sometimes predatory marketing campaigns, sensationalized success stories, incomplete media coverage, enthusiastic patient advocacy groups, and opportunistic capitalism, cell therapy – often, and not always accurately, termed ‘stem cell therapy’ – has been touted as a cure for countless conditions. This has created a stem cell tourism industry in which patients travel the globe, often paying large sums of money and assuming ill-defined risks, to receive unproven cell therapies. Based on public perception of the efficacy and promise of cell therapies, many patients or patients’ parents feel entitled to receive these products regardless of the quality or quantity of supporting scientific data. Contributing to public confusion, some unscrupulous individuals financially profit and put the whole field at undue risk by treating patients with unregulated and unproven cell products. Internationally, regulatory and scientific communities are clear and unified in their guidelines and position statements that unlicensed cell therapy products should only be administered under the authorization of a regulatory body, preferably in the form of a clinical trial.⁵² In the United States, this takes the form of an Investigational New Drug application issued by the Federal Drug Administration.

In conditions for which there are no cures and innovative therapies are desperately needed, time is often perceived as the enemy. Particularly when those conditions affect children, it is easy to become frustrated with the pace of scientific progress. While we are by no means destined to recapitulate history, it can provide valuable context and lessons. In 1970, after a decade of performing bone marrow transplants, a review of 203 cases⁵³ reported that more than 75% of treated patients died, and donor cell engraftment, the indicator of a ‘successful’ transplant, occurred in less than 40% of recipients. Nevertheless, dedicated investigators persisted and, through iterative approaches, advanced the field. Immense progress has been made in subsequent years, and HSCT is now indicated in multiple scenarios, providing a viable treatment option for many who once had none. The most recently approved cell therapy, autologous chimeric antigen receptor T (CAR-T) cells, have dramatically increased survival for certain patients with acute lymphoblastic leukemia and were finally licensed after 30 years of development including multiple failures along the way.

Cell therapy for neurological conditions is in its infancy. Initial human studies have not produced revolutionary results, but that is not the norm in medical advances and should not be expected instantly in such complex conditions. Cell therapy may prove to be unsuccessful for certain conditions or patients, and it will not be a cure-all for every child with a neurodevelopmental disorder. But there could very well be populations of children for whom cell therapy can improve function and quality of life, opening doors to opportunities that their disorders have closed. Answering the yet unsolved questions will require rigorous scientific studies, persistence, and a bit of patience.