Article Text

Download PDFPDF
Moral obligations in conducting stem cell-based therapy trials for autism spectrum disorder
  1. Nicole Shu Ling Yeo-Teh1,
  2. Bor Luen Tang1,2
  1. 1 Research Compliance and Integrity Office, National University of Singapore, Singapore
  2. 2 Department of Biochemistry, Yong Loo Lin School of Medicine, National University Health System, National University of Singapore, Singapore
  1. Correspondence to Dr Nicole Shu Ling Yeo-Teh, Research Compliance and Integrity Office, National University of Singapore, Singapore; dprysln{at}; Dr Bor Luen Tang; bchtbl{at}


Unregulated patient treatments and approved clinical trials have been conducted with haematopoietic stem cells and mesenchymal stem cells for children with autism spectrum disorder (ASD). While the former direct-to-consumer practice is usually considered rogue and should be legally constrained, regulated clinical trials could also be ethically questionable. Here, we outline principal objections against these trials as they are currently conducted. Notably, these often lack a clear rationale for how transplanted cells may confer a therapeutic benefit in ASD, and thus, have ill-defined therapeutic outcomes. We posit that ambiguous and unsubstantiated descriptions of outcome from such clinical trials may nonetheless appeal to the lay public as being based on authentic scientific findings. These may further fuel caregivers of patients with ASD to pursue unregulated direct-to-consumer treatments, thus exposing them to unnecessary risks. There is, therefore, a moral obligation on the part of those regulating and conducting clinical trials of stem cell-based therapeutic for ASD minors to incorporate clear therapeutic targets, scientific rigour and reporting accuracy in their work. Any further stem cell-based trials for ASD unsupported by significant preclinical advances and particularly sound scientific hypothesis and aims would be ethically indefensible.

  • stem cell research
  • clinical trials
  • research ethics
  • clinical ethics
  • research on special populations

Data availability statement

No data are available.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Experimental therapy approaches involving transplantation with human stem cells have been attempted in many human diseases ranging from cancer to neurodegeneration. Thus far, the only stem cell-based treatments approved by the US Food and Drug Administration (FDA) is the use of cord blood for the treatment of cancers and disorders of the haematopoietic/immune system.1 There are varying sources of stem cells used in stem cell-based therapy, including pluripotent human stem cell lines or those generated via reprogramming induced pluripotency (induced pluripotent stem cells, or iPSCs), haematopoietic stem cells (eg, those from umbilical cord blood or bone marrow) as well as mesenchymal stem cells (MSCs).2 The latter in particular has emerged as a popular source for experimental therapeutics.3 MSCs could be derived from the human umbilical cord or adult tissues such as the bone marrow or fat pads and are thus potentially useful for autologous transplantation.

For neurological disorders, stem cell-based experimental therapy has been applied to a diversity of conditions, including severe or near-terminal neurodegenerative diseases such as amyotrophic lateral sclerosis (ALS)4 and Parkinson’s disease,5 as well as central nervous system injuries.6 In recent years, stem cell-based therapy has also been explored for the treatment of neurodevelopmental disorders such as autism spectrum disorder (ASD).7

ASD is a complex disorder with a spectrum of neurological phenotype in children, characterised by dysfunctions in reciprocal social interaction and communication, as well as an exhibition of repetitive and stereotypical behaviours.8 9 Patients usually display autistic behaviour from a very young age. ASD has a prevalence of 1 in 54 births in the USA,10 and disease pathogenesis are contributed by both genetic and environmental factors.8 Although there is no recognised treatment for ASD, early diagnosis allows timely intervention, which has been shown to reduce symptoms and improve cognitive function. However, a reduction or loss of symptoms in adulthood does not necessarily mean a complete remission. Instead, the alleviation of autistic behaviour might be due to the masking of symptoms, initial misdiagnosis, or due to a change in ASD criteria stipulated in the Diagnostic and Statistical Manual of Mental Disorders -5 of the American Psychiatric Association.

Much of the stem cell-based therapeutics for ASD are unregulated, direct-to-consumer treatment regimens offered by for-profit stem cell clinics.11 12 A growing trend for stem cell tourism has also been observed, where patients travel overseas to receive unapproved treatments at a significant financial cost (prices upwards of US$20 000 per treatment).13 The treatments offered in both regards are often considered scientifically unsound, and the hype on any therapeutic efficacies have been commented on extensively.14–17 The FDA has issued multiple statements cautioning consumers about the dangers of unregulated stem cell therapies, as well as warning treatment centres, healthcare providers and laboratories on unlawful practices involving cord blood.18 One of the said laboratories involved in storing and processing cord blood for the treatment of autism were found to have significantly contravened current good tissue practice and current good manufacturing practice, including inadequate practice of aseptic techniques while handling samples, as well as insufficient screening of donors’ medical records for risk factors of communicable diseases,19 potentially resulting in preventable but fatal consequences.

On the other hand, there are several regulated clinical trials conducted in a handful of different countries, with results published in peer-reviewed journals. These phase I or I/II trials reported, in general, alleviation of safety concerns in the administration of cord blood cells and MSCs to patients with ASD intravenously, as well as some degree of positive behavioural improvement.20–24 These reports are likely to bode well for the stem cell therapy ‘believers’ in the ASD community and may encourage parents to register ASD children for future trials of similar nature, or to seek treatment directly from the stem cell clinics.

However, a closer scrutiny of the literature and these reports continue to surface some worrying issues, including a lack of therapeutic rationale25 and targeted outcome assessment,7 as well as longitudinal safety assessments. These, coupled with the facts that young patients with ASD (1) are not able to directly give informed consent, (2) generally have a near-average life expectancy and could thus suffer from delayed manifestation of treatment risks (such as eventual development of malignancy), and (3) may experience some degree of ASD symptom remission with age following non-invasive behavioural therapy, make stem cell-based clinical trials, as they are currently regulated, conducted and reported, ethically questionable.

We argue below that those who conduct and approve such stem cell-based clinical trials on patients with ASD have moral obligations to observe the utmost scientific rigour since clinical trial reports are key in influencing parents’ decision to take up such treatment for their child. We opine that any such trials should only be conducted based on substantiated, verified and reliable science and when clear benefit over risk could be reasonably projected.

Experimental stem cell therapy for ASD: issues and shortcomings

Literature searches were performed to gauge attempts and outcomes of experimental stem cell therapy for ASD. A search of the clinical trials database ( with the terms ‘Autism’ and ‘Stem cells’ retrieved 21 items. Of these, four have a ‘unknown’ status, two are ‘withdrawn’, two ‘active, not recruiting’, five ‘recruiting’, one ‘not yet recruiting’ and seven items bear a status of ‘completed’. For the latter, one item has a focused on ‘hyperbaric therapy’ and not stem cell therapy. A search of PubMed ( with the search terms ‘autism’, ‘stem cells’ and ‘clinical trial’ gathered 28 entries. Of these, several are case studies with no organised treatment structure. Six of these appear to be phase I or I/II trials and these are summarised in table 1. Four of these bear a registration number. A scrutiny of these six reports allow an assessment of protocol design and trial outcome. These have been critically reviewed and assessed by Price,7 except for the most recently published Duke University trial.21 Granted that these are phase I/II trials, and the focus is supposedly on safety assessment with a small number of subjects. However, the trials, as they are conducted and reported, present some issues and shortcomings, which are elaborated below. Where relevant, discordance with the International Society for Stem Cell Research (ISSCR) guidelines1 for clinical translation of stem cells are also highlighted.

Table 1

Literature documentation of reports associated with stem cell-based experimental therapy of ASD

A general lack of definitive therapeutic rationale or molecular target

What exactly is the rationale or therapeutic basis for stem cell-based therapy for ASD? Unlike the hopeful replacement or survival augmentation for neuronal demise in neurodegenerative diseases, a clear role for stem cells in alleviating ASD symptoms and phenotype has not been elucidated as of now. The molecular basis of ASD is multifactorial, and the lack of a definitive screening target was the exact reason why an application for Preimplantation Genetic Diagnosis for ASD submitted to the Human Fertilisation and Embryology Authority was denied.26

A potential molecular manifestation of ASD is neuroinflammation, which overlaps with a large number of neurodegenerative diseases.27 28 Neuroinflammation may lead to loss of neuronal function, disruption of synaptic connections and thus perturbations of neural transmission circuits. A recent systematic review of post-mortem analysis in patients with ASD by Liao et al concluded that ‘…the majority of studies supported the presence of neuroinflammation in the neurobiological pattern of ASD as shown by activation of astrocytes and microglia together with abnormal levels of cytokines and chemokines…’.27 Transplanted stem cells, such as MSCs, are known to exert immunomodulatory effects28 29 which might attenuate the degree of neuroinflammation. Perhaps the most frequently cited potential benefit for ASD is therefore immunomodulation30 by transplanted stem cells. This is in fact explicitly mentioned in the abstracts of five of the clinical trial reports alluded to above.20–24

The problem with this notion is twofold. First, accompanying neuroinflammation is but one aspect of ASD neuropathology and is not definitively implicated in all ASD cases. Therefore, it cannot be assumed to either persistently underlie or be taken as the main aetiological driver for all cases of ASD. The severity of neuroinflammation varies between individual patients and the extent to which these varying degrees of neuroinflammation may contribute towards the ASD neurobehavioral phenotype is yet uncertain. Additionally, many genetic and epigenetic factors underlie ASD pathology,31 32 and a number of these factors are known to directly affect critical aspects of neurotransmission. These include synaptic connectivity and function,33 which may not be causally or directly related to neuroinflammation. Furthermore, it is also generally acknowledged that ASD stems mainly from neurodevelopmental dysfunctions, and could involve defects in developmental processes in utero, such as aberrations in neuronal differentiation and migration,34 which again are not directly related to neuroinflammation. If so, the rationale or basis of stem cell-based therapy for immunomodulation in ASD appears rather weak. It is much weaker than those for neurodegenerative disorders or neuronal injuries in which neuroinflammation is a prevailing phenotype and a proven therapeutic target,35 and for that matter, also weaker than its recent controversial purported use against the hyperimmune responses and cytokine storm causing mortality in COVID-19 patients.36 37

Second, if a clinical trial with stem cell therapy has immunomodulation as a working hypothesis as well as an intended effect or outcome, recruited subjects should be assessed for any changes in neuroinflammation status before and after treatment. While this is not straightforward without ready immunohistochemical access to the living brain, cytokine profiling of blood or cerebrospinal fluid can be done, while complementary MRI could be revealing.38 The treatment outcome could then be stratified as such. Price raised this critical point in his review of the ASD stem cell therapy trials,7 and highlighted that none of the trials assessed the levels of proinflammatory cytokines in the subjects except for the Riordan et al study.24 Even for the said study, only two chemokines, macrophage‐derived chemokine (MDC) and thymus, and activation‐regulated chemokine (TARC), were measured. The selection of these two chemokines is based on a single study which showed an elevation of MDC and TARC in the serum of autistic children.39 Classical proinflammatory cytokines such as IL-1β, TNF-α, IL-6 and IFN-γ were not assessed. As such, in none of these trials were the stem cells’ immunomodulatory hypothesis directly or fully tested. Furthermore, the lack of molecular target(s) serving as disease biomarkers rendered any assessment of therapeutic efficacy difficult and vague. If, in the future, the molecular basis of ASD is clearly elucidated, with therapeutic targets identified, it may be possible that stem cells would be a tenable treatment, as it is currently with some blood cancers. Many of the stem cell clinical trials, are indeed, as the authors claimed to be, simply ‘exploratory’ in nature, at the expense of the patient.40 As such, these trials would be in discordance with ISSCR Recommendation, which states that ‘Researchers should take measures to maximise the scientific value of early phase trials’.

A lack of longitudinal risk and safety assessment

The aim of phase I and phase I/II trials is to assess safety and tolerance of the therapeutic agent. In this regard, all six trial reports indicated that the respective stem cell transplantation processes were shown to have few side effects and the stem cell infusion regiments were well tolerated. However, transplantation-incurred symptomatic side effects were in fact observed, but these were not emphasised in the reports. As an example, the intrathecal procedure of Sharma et al resulted in 3 cases of epileptic seizure out of 32 subjects (~9%), which the authors indicated could be medically controlled.22 Additionally, several subjects of Sun and colleagues who received allogeneic human cord tissue mesenchymal stromal cells developed new class I antihuman leucocyte antigen (HLA) antibodies, associated with either a specific lot of transplanted cells, or with a partial HLA match with the donor.21 The authors stated that these antibody responses were ‘clinically silent’ and not associated with any clinical manifestations, presumably within the time frame of follow-up analysis of 12 months. The safety of stem cell transplantation is generally borne out by reports on experimental therapeutics in other diseases.41 42 However, one must bear in mind that these adverse effects are usually limited to those that are clinically measurable during treatment in ASD subjects. Compounding this problem is the fact that some patients with ASD suffer from hyposensitivity or hypersensitivity to pain, and since they also exhibit behavioural deficits that might mask or promote under-reporting of discomfort/pain, the true measure of adverse effects is often difficult to gauge.43 In fact, the maximum follow-up period of the trials utilising stem cell-based therapies for ASD ranged from 3 months (two trials), 12 months (three trials) and a variable assessment period of between 5–26 months (one trial). Although these timelines appear sufficient to observe any acute adverse events contributing to morbidity or mortality, they do not rule out the eventual development of long-term side effects such as malignancy or autoimmune disorders. Importantly, although individuals with ASD do experience increased morbidity and a slightly below average life expectancy,44 these timelines are very short in comparison with the projected lifespan of patients with ASD. It is worth noting that the higher mortality rate of patients with ASD is due largely to accidents (asphyxiation, drowning, suffocation), which have a higher occurrence in low-functioning patients, rather than death by comorbidities from ASD.43 45 By comparison, ALS patients have a life expectancy of only 2–5 years on diagnosis with significantly decreased quality of life.46

While conventional behavioural therapy could improve ASD symptoms and there are cases of symptom remission with age,47 48 the FDA-approved treatment for ALS, riluzole, increases patients’ mean survival by a mere 3 months.46 Furthermore, although stem cell administration is considered relatively ‘safe’, there are ample reports in the literature of both acute and delayed negative consequences. For example, MSC infusions have been reported to enhance renal deterioration a mere 1 week post-infusion49 and induce episodes of thromboembolism within 3 days of infusion.50 Moreover, research on the characteristics of MSCs indicate that these could promote oncogenesis or otherwise enhance cancer progression by contributing to the tumour microenvironment and eliciting changes in anti-tumour immune response.51 52 Indeed, in a review by Bauer et al 53 on adverse events in patients who received unproven stem cell-based interventions for other conditions, complications which were classified as ‘neoplastic’ occurred between 6 months54 and 8 years post-transplantation,55 the latter of which far exceeds the duration of the clinical trials reported above. Furthermore, definitive studies into the long-term effect of the treatment have been presented to be void of side effects, likely due to the short duration (<3 years) of their studies.56–59 Given the reasonable life expectancy of ASD stem cell transplantation recipients, longitudinal monitoring and follow-up, both in terms of cancer risks as well as behavioural outcomes, would be warranted and should be built into future trial protocols. A lack of such long-term follow-up would be in discordance with ISSCR Recommendation, which states that ‘Given the potential for transplanted cellular products to persist, and depending on the nature of the experimental stem cell-based intervention, subjects should be advised to undergo long-term health monitoring’.

The consequences of scientifically inapt stem cell therapy trials for ASD

Although safety assessment appears to be the primary aim, all six trials cited above reported moderate degrees of behavioural improvement of test subjects. Furthermore, four of the six trials are open-label studies with no placebo or control groups. The small number of subjects and the mixed reporting of almost any positive measurements by various tests borders on overinterpretation of data and calls for reservations on the reported potential benefit. One open-label study with controls20 did show statistically significant improvements in stem cell-treated groups over control group (which received only conventional rehabilitation therapy, also administered to the stem cell-treated groups). Interestingly, the group treated with a combination of 2 different stem cell types (cord blood mononuclear cells (CBMNCs) and umbilical cord-derived MSCs) saw a greater improvement than the comparison group who given only CBMNCs. On the other hand, the only study which adopted a randomised, blinded, placebo-controlled, cross-over trial did not show statistically significant differences for any endpoints.40 On the whole, the trials reported to date do not unequivocally attest to a measurable benefit in terms of ASD symptoms alleviation conferred by the transplanted cells but were nonetheless optimistically interpreted as showing some positive effects and few side effects.

Scientific publications, or even the mere citation of scientific reports, are important influencers of the laypersons’ opinion and decision. This is perhaps aptly reflected in the field of ASD by the Wakefield case, in which a fraudulent connection made between autism and the measles–mumps–rubella (MMR) vaccination caused fear among many and flamed antivaccination sentiments.60 61 In times and situations of hardship, any glimmer of hope would be much sought after. For the suffering patients and in the case of ASD, the exhausted caregiving guardians, any hint of therapeutic hope could be disproportionately magnified. As clinical trials that are few and far between offer limited sign-up opportunities, the fear therefore is that desperate and misinformed parents, falsely encouraged by any perceived positive results, may decide to engage the services of unregulated stem cell clinics. It was also noted that only informed consent from the parents of patients were obtained, and assent from the patients were not sought in any of the clinical trials, although adult patients were recruited in some of the trials. Patients with ASD, unable to consent to (or refuse participation in) stem cell-based experimental treatments and may suffer long-term pathological consequences, would thus have their autonomy violated and may even develop treatment-related conditions. Stem cell-based therapy is not by any means a last resort for patients with ASD. Therefore, in experimenting stem cell-based therapy for ASD, an ‘anything goes’ mindset or attitude must not ever be adopted.

Moral obligations and scientific rigor of stem cell therapy trial for ASD

We thus posit that those who approve and those who conduct stem cell therapy-based clinical trials have moral obligations to observe proper scientific rigour and research ethics, particularly for ASD which inflicts children that are generally unable to give direct consent. Any such trials should be conducted only when (1) these could be based on very sound science and (2) clear benefit over risk could be reasonably expected. Pertaining to the first point, this would mean that the treatment rationale needs to be well supported and the study design sufficient to test a research hypothesis. It would also mean better stratification of subjects based on ASD types, more well-defined molecular targets and pathways, as well as the use of biomarkers that would offer quantification of treatment outcome beyond behavioural tests. In other words, one should experiment on patients with ASD only when the preclinical science is sufficiently advanced. Due to the complexity and largely idiopathic nature of ASD, good disease models for therapeutic testing has been hard to come by. Animal studies that have relied on the popular Black and Tan BRachyury (BTBR) T+ tf/J mouse model62 offered little in terms of molecular targets and markers. However, other ASD mouse models with more targeted genetic changes based on human ASD susceptibility genes have emerged.63 Also, the advent of human iPSCs64 that could be developed into human in vitro models65 of neurons and brain organoids66 67 may promote further advances in molecular and pathway profiling of patient-specific defects.68 Until preclinical studies are sufficiently advanced to provide handles necessary to assess therapeutic targets and benefits, a lack of definitive scientific questions to ask and answer would render further stem cell-based clinical trials for ASD morally untenable.

With regard to the second point, any further stem cell-based trials in ASD would definitely require significant enhancement in risk assessment and longitudinal risk monitoring, increase in trial oversight, as well as independent trial analysis and reporting. The need for these enhanced measures are in fact based on nothing more than the basic guiding principles for medical research for human subjects as originally demarcated by the Declaration of Helsinki.69 Price has posited that since ‘The purpose of open-labelled phase I/II trials is to demonstrate safety. If that is achieved, then further such studies are redundant and thereby unethical.’.7 In other words, stem cell-based clinical trials in ASD should now really address the science of a therapeutic approach with the hope of optimisation and improving the efficacy of the treatment via generation of knowledge, and not simply be used to conduct more exploratory or experimental therapy.


Experimental stem cell-based therapy in ASD is typically performed on a group of young and vulnerable subjects who are not able to give direct consent, but nonetheless have a near normal life expectancy and a finite possibility of symptom resolution with conventional behavioural therapy and with age. Stem cell-based interventional therapy could thus not be morally defensible unless (1) there are clear research questions and therapeutic targets/outcomes to address that are aptly informed by preclinical work and (2) when a clear benefit over risk could be reasonably expected.

Data availability statement

No data are available.

Ethics statements

Patient consent for publication



  • Correction notice This paper has been updated since first published to update author details for author 'Bor Luen Tang'.

  • Contributors BLT conceived, drafted and reviewed the manuscript. NY drafted and reviewed the manuscript.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • ISSCR guidelines