Outcomes of bone marrow mononuclear cell transplantation combined with interventional education for autism spectrum disorder

Abstract The aim of this study was to evaluate the safety and efficacy of autologous bone marrow mononuclear cell transplantation combined with educational intervention for children with autism spectrum disorder. An open‐label clinical trial was performed from July 2017 to August 2019 at Vinmec International Hospital, Hanoi, Vietnam. Thirty children who fulfilled the autism criteria of the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, and had Childhood Autism Rating Scale (CARS) scores >37 were selected. Bone marrow was harvested by anterior iliac crest puncture under general anesthesia. The volume collected was as follows: 8 mL/kg for patients under 10 kg (80 mL + [body weight in kg − 10] × 7 mL) for patients above 10 kg. Mononuclear cells were isolated with a Ficoll gradient and then infused intrathecally. The same procedure was repeated 6 months later. After the first transplantation, all patients underwent 8 weeks of educational intervention based on the Early Start Denver Model. There were no severe adverse events associated with transplantation. The severity of autism spectrum disorder (ASD) was significantly reduced, with the median CARS score decreasing from 50 (range 40‐55.5) to 46.5 (range 33.5‐53.5) (P < .05). Adaptive capacity increased, with the median Vineland Adaptive Behavior Scales score rising from 53.5 to 60.5. Social communication, language, and daily skills improved markedly within 18 months after transplantation. Conversely, repetitive behaviors and hyperactivity decreased remarkably. Autologous bone marrow mononuclear cell transplantation in combination with behavioral intervention was safe and well tolerated in children with ASD (Trial registration: ClinicalTrials.gov identifier: NCT03225651).


| INTRODUCTION
Autism spectrum disorder (ASD) is a complex spectrum of disorders characterized by two typical abnormalities: (a) deficits of social communication and interaction; (b) the presence of restricted interests as well as repetitive and stereotypic verbal and nonverbal behaviors. 1,2 Comorbidities including sleep disorders, seizures, and gastrointestinal difficulties are very common in children with ASD. 3 The prevalence of identified ASD is increasing. 4 In 2016, the overall ASD prevalence among children aged 4 years was 15.6 per 1000 (1/64), 5 and the incidence was 18.5 per 1000 (1/54) in 8-year-old children according to Early Autism and Developmental Disabilities Monitoring Network sites. 6 The etiology of ASD is still not well understood. However, many associated factors including genetic mutations, immune dysregulation, hypoperfusion of some parts of the brain, exposure to maternal antibodies during pregnancy, and weak functional connectivity across brain regions are suggested to contribute to the development of ASD. [7][8][9][10][11] Multiple approaches including behavioral therapy, occupational therapy, speech therapy, and medications are required in the management of ASD to ameliorate autistic symptoms. Educational and behavioral interventions have been recognized as crucial for the management of ASD in children. 12 The evidence indicates that young children with ASD benefit from interventions that focus on improving social interaction, communication, and challenging behaviors. 13,14 Unfortunately, many children who receive those treatments remain significantly impaired. 15 In search of better outcomes in the management of ASD, alternative and complementary treatments are being investigated. Recent reports have suggested that stem cell transplantation result in improvements in several different neurological conditions. [16][17][18] The suggested mechanisms of action of mesenchymal stem cells (MSCs) on the nervous system include neuroprotection, neurogenesis, and synaptogenesis. [19][20][21] Stem cell applications were also assessed in animals using the inbred BTBR T+tf/J (BTBR) mouse strain, 22 which has autistic-like symptoms, to explore the potential of stem cells in the management of ASD. In BTBR mice, Segal-Gavish et al showed that transplantation of MSCs resulted in a reduction in stereotypic behaviors, a decrease in cognitive rigidity, and an improvement in social behavior. Moreover, it has been shown that brain-derived neurotrophic factor protein levels as well as neurogenesis increased in the hippocampus following stem cell treatment. 23 Similarly, Perets et al demonstrated that brainderived neurotrophic factors secreted by transplanted MSCs were a key factor in the observed reductions in stereotypic behavior and in the improvement of cognitive flexibility in the BTBR model. 24 Ha et al revealed that transplanted human adipose-derived stem cells improved repetitive behaviors, social interaction, and anxiety in valproic acid-induced ASD model mice. 25 Based on results obtained from animal research, stem cell transplantations have been conducted for children with ASD at several centers. 26 Two distinct approaches have been explored thus far in the application of cell therapy products for the treatment of ASD: culture expanded and nonexpanded. In an early example of non-cultureexpanded cell therapy, Sharma et al described the use of autologous bone marrow mononuclear cell (BMMNC) transplantation infused via intrathecal route in 32 children with ASD. The procedure was reported as being safe with minor adverse events encountered, such as nausea, vomiting, and pain at the site of injection. Improvements were noticed in different aspects, including social relationships and reciprocity, speech and language patterns, and brain metabolism. 27 A further example of concentrated, but not expanded, cell therapy described the use of autologous BMMNC transplantation via the intrathecal route in 10 children with ASD. The results revealed that the maximal treatment effect was observed within the first 12 months with, again, with no safety concerns. 28 A combined culture-expanded/nonexpanded cell therapy study reported on the outcomes of allogenic cord blood mononuclear cell (CBMNC) transplantation vs CBMNC combined with umbilical cord MSC transplantation for children with ASD. 29 In this trial, four stem cell infusions were carried out via intravenous and intrathecal routes. No severe adverse events after cell transplantation were

Lessons learned
• The combination of cell therapy and educational intervention may improve clinical manifestations such as social communication, language, and daily skills in children with ASD.

Significance statement
The combination of cell therapy and educational intervention may improve clinical manifestations, such as social communication, language, and daily skills, in children with autism spectrum disorder. However, additional studies with control groups should be performed in the future to obtain a more comprehensive and accurate conclusion. described for either group. Improved outcomes were noted in the combination group compared with the CBMNC-alone group.
More recently, Dawson et al reported autologous stored cord blood infusion through intravenous route in 25 children with ASD, which resulted in significant improvements in behavior at 6 months after infusion; these improvements were sustained at 12 months. 30 That study was followed by the first randomized, double-blind, placebo-controlled clinical trial comparing outcomes of autologous cord blood infusion vs placebo for children with ASD. As in all previous studies, autologous intravenous cord blood infusion had no serious adverse events and trended toward improvement, especially in socialization, but clinical outcomes were not significantly different between the two groups. 31 In summary, a number of clinical trials have been performed thus far, exploring the application of cell therapy for the treatment of ASD.
Although the trials have been broadly consistent in outcome reporting, disparities remain around cell sources, processing, dosage, and delivery route. The aim of this study was to investigate the safety and clinical outcomes of high-dosage BMMNC transplantation combined with educational intervention for children with ASD.

| Exclusion criteria
The exclusion criteria were epilepsy; hydrocephalus with ventricular drain; coagulation disorders; allergy to anesthetic agents; severe health conditions such as cancer or heart, lung, liver, or kidney failure; and active infections. Patients with Rett syndrome or fragile X syndrome were also excluded from this study.  is classified into three levels: level 1 ("Requiring support"), level 2 ("Requiring substantial support"), and level 3 ("Requiring very substantial support"). 33 CARS consists of 14 domains assessing behaviors associated with ASD, with a 15th domain rating general impressions of ASD. 34 VABS-II is a standardized measure 35 that yields an overall score and subscale standard scores in the four different domains: socialization, communication, daily living skills, and motor skills. CGI is a rating scale that measures symptom severity and treatment response. 36 The severity is categorized into seven levels: (a) not present (no ASD), (b) barely evident ASD symptoms, (c) mild ASD symptoms, (d) moderate ASD symptoms, (e) moderately severe ASD symptoms, (f) severe ASD symptoms, or (g) very severe ASD symptoms. The response of each patient is also divided into seven levels: level 1, very much improved; level 2, much improved; level 3, minimally improved; level 4, no change; level 5, minimally worse; level 6, much worse; and level 7, very much worse. In addition, main indicators including social interaction, eye contact, expressive language, abnormal behaviors, sensory abnormalities, eating and sleeping difficulties, daily skills, and learning capacity before and after transplantation were collected to examine the effects of stem cell therapy in combination with behavioral intervention.

| Study design
A brief video was recorded at baseline and then at 6, 12, and 18 months after the first transplantation. A questionnaire was completed by patients' caregivers 6 and 18 months after the first transplantation to obtain their assessment and satisfaction. A psychologist kept contact with patients' caregivers via telephone throughout the follow-up period to record any unexpected events during the study.

| Laboratory and imaging diagnostics
Routine hematologic and biochemistry examinations were performed at baseline and 6 months later. Diagnostic imaging examinations, including brain magnetic resonance imaging (MRI) and electroencephalography (EEG), were carried out in all patients at baseline to rule out epilepsy and brain malformations.
Positron emission tomography-computed tomography (PET-CT) was carried out before stem cell transplantation and retaken 12 months after the first transplantation to monitor changes of brain metabolism. On post hoc analysis, decreased and increased fluorodeoxyglucose (FDG) metabolism regions in autistic children were evaluated based on a normal distribution curve. If the measured value is lower than one SD from the median value of the standardized uptake value, it is considered to be hypometabolic, whereas a measured value greater than a SD from the median value is considered to be an increase in metabolism. 37 In evaluation of PET-CT images, dark blue brain areas were defined as severely reduced FDG metabolic rate, and light blue brain areas were assessed as moderate metabolism. Green was considered to have mild metabolic reduction. The yellow-orange brain region was assessed to exhibit increased FDG metabolism. 38  Human Build 37 (GRCh37) reference genome. 39 The Genome Analysis Toolkit 40 and SnpEff (an open-source tool that annotates variants and predicts their effects on genes) 41

| Genetic testing
We detected and validated a de novo CNV in the SHANK3 gene from a female proband (proband A27) and 23 different variants in 22 genes from eight probands (  SHANK3, CHD8, ANK2, and GIGYF2 belong to the genes with the strongest evidence of relevance to ASD.

| Brain MRI and EEG
No abnormalities on MRI or EEG were observed in any of the patients.

| FDG changes after BMMNC transplantation
At baseline, manifestation of hypometabolism was found at seven major brain regions: hippocampus, anterior cingulate gyrus, posterior cingulate gyrus, parietal lobe, frontal lobe, temporal lobe, and central sulcus. After the stem cell transplant, 29 children underwent a second PET-CT scan (the parent refused the second PET-CT in one patient).
Improvement in metabolism was observed in some brain regions where severe hypometabolism was noticed before BMMNC transplantation such as parietal lobe, frontal lobe, and anterior cingulate gyrus. However, these changes were not statistically significant (Table 6).

| Adverse events
None of the patients had any severe adverse events during bone marrow aspiration, stem cell infusion, or following transplantation. The procedure was reported as being safe with minor adverse events encountered. There were no procedure-related major adverse events.
Among 96 adverse events that occurred during study period, 46 (48%) mild and moderate adverse events were recorded that may or may not be related to BMMNC transplantation with symptoms including pain, vomiting, and mild fever. All those adverse events were easily managed through appropriate medication (Table 7).

| Clinical outcomes
After BMMNC transplantation, the severity of ASD decreased remarkably. The median CARS scores decreased from 50 (range 40-55.5) points at baseline to 46.5 (range 33.5-53.5) after 18 months.
Patient-specific CARS score analysis for each patient is presented in Figure 1. The result of the mixed-effects analysis suggests that each visit was associated with a decrease of 1.6 in the CARS score and that this change was statistically significant (Table 8).  According to the DSM-5 classification, the number of patients at level 3 was reduced from 28 to 18. There were no patients at level 1 before transplantation, but after transplantation, five children were recategorized at this level.
After transplantation, improvements were observed in various aspects. Social interaction and eye contact increased remarkably from 37% before the first transplantation to 97% and from 23% to 93%, respectively, after 18 months. Expressive language increased from 47% before transplantation to 67% after 6 months, 87% after 12 months, and 93% after 18 months (Table 2). Abnormal behaviors also decreased after transplantation. The children with repetitive behavior decreased from 93% to 87% (Table 3). Sensory abnormalities and sleeping difficulties progressively improved after transplantation, from 97% to 73% and from 53% to 43%, respectively (

| DISCUSSION
In this study, bone marrow was collected from anterior iliac crests and not from posterior crests as in other reports. 17,27 This change facilitated anesthesia and reduced risks related to the prone position of patients. The dosage of transplanted mononuclear cells and CD34 + cells in our study was higher than those in studies reported by Sharma,Dawson,and Chez. 27,30,31 Differences related to the cell transplants compared with other studies are presented in Table 9.
Granulocyte-colony stimulating factor (G-CSF) was not used in this study to mobilize stem cells from bone marrow, with high dosage values nevertheless isolated. This implies that G-CSF may not be The incidence of minor adverse events was low and easily managed through medication or spontaneously resolved themselves.
Although all participants still belonged to severe level at the baseline after receiving behavioral intervention with a mean duration of 3.5 years, this study showed improvements in various aspects after BMMNC transplantation combined with educational intervention.
Positive changes in social communication, eye contact, language, behaviors, and daily skills were observed after BMMNC transplantation. In addition, learning ability also remarkably improved after transplantation. The number of children who could go to school without support increased after transplantation (Table 10).
Hyperactivity of children with ASD is a disorder that severely impairs quality of life for the whole family. In our study, the rate of children with hyperactive disorder decreased by 50% at 18 months after stem cell transplantation.
Positive changes were found in evaluation measures, including severity and adaptive ability. The number of patients at level 3 (requiring very substantial support) according to DSM-5 decreased from 28 to 18 at 18 months after transplantation.
We noticed that the improvements appeared to be influenced by the CARS scores at baseline. Patients with a CARS score ≤49 at baseline showed better improvement than those who had CARS scores >49 points. This would imply that patients with lesser severity had better outcomes after transplantation.
Genetic abnormalities have been reported in many ASD studies.
In our series, genetic variations were found in eight patients (26%).
Two different pathways for administration have been applied in the delivery of cell therapy for ASD: intravenous and intrathecal.
Although both delivery routes are safe, there are concerns regarding cells delivered through the intravenous route, as transplantations in animal models have shown that the transplanted cells have difficulties passing through organs such as the spleen, kidney, and intestine. 57 The intrathecal route does not present this concern.
The number of BMMNCs transplanted also varies between studies, and the number we have used was higher than those from other groups. [27][28][29][30][31] There have been suggestions of correlation between transplanted stem cell dosage and the extent of subsequent clinical improvement. 58 Although we did not perform an escalating dosage study, we noted that a high dosage of stem cells may be used to obtain satisfactory outcomes in children with ASD. Furthermore, we performed two transplantations instead of one transplantation as described elsewhere. 27,30,31 Multiple transplantations have resulted in positive outcomes in both ASD 29 and spinal cord injury. 59 Moreover, the benefits of repeated transplantations vs single transplantation have been identified in animal myocardium infarction models. 60 In our study, all children received 18 months of follow-up after the first transplantation. This follow-up duration exceeds those in other recent reports. 27,[29][30][31] We noticed that the longer the follow-up duration was, the lower the severity of ASD (CARS score reduction) and the better the children's adaptive functioning (VABS score increase). Meanwhile, there was no case in which the results were not improved or even worse compared with baseline, implying that the treatments have a sustainable long-term effect. It is likely that further extended follow-up times will be required to fully assess the responses of children with ASD over time after stem cell transplantation. Herein, we have demonstrated the safety and feasibility of BMMNC transplantation for the treatment of ASD. The lack of a control group who received either stem cell transplantation or educational intervention only is a limitation of this study. However, our study results provide initial evidence to justify conducting a randomized clinical trial with control groups in the future.

| CONCLUSION
We conclude that autologous BMMNC transplantation is safe. The

DATA AVAILABILITY STATEMENT
All data generated or analyzed during this study are included in this published article and its supplementary information files.