Brief Report: Autistic Symptoms, Developmental Regression, Mental Retardation, Epilepsy, and Dyskinesias in CNS Folate Deficiency

Introduction

Isolated deficiency of folate in the CSF has been described in rare individuals with progressive postnatal neurological dysfunction of varying presentations and manifestations. Isolated CSF deficiency of folates was originally reported in an 18-year-old male with progressive sensorineural hearing loss, cerebellar ataxia, distal spinal muscular atrophy, and pyramidal tract dysfunction (Wevers et al. 1994). Subsequently, other authors described several pediatric cases. A series of five children showed deceleration of head growth, psychomotor retardation, cerebellar ataxia, dyskinesia, pyramidal signs, and seizures (Ramaekers et al. 2002). These findings were largely confirmed by the same authors in a series of 20 subjects (Ramaekers and Blau 2004). Isolated Rett syndrome was described in four cases (Ramaekers et al. 2003). Three children demonstrated microcephaly, severe developmental delay, dyskinesia or spasticity, seizures, and CNS calcifications and were diagnosed with atypical Aicardi-Goutieres syndrome (Blau et al. 2003). One female was diagnosed with developmental delay, spastic paraparesis, dystonia, and speech difficulties (Hansen and Blau 2005). One female was identified with developmental delay, mental retardation (MR), seizures, dyskinesias, and autism spectrum manifestations (Moretti et al. 2005). An incomplete form of Kearns-Sayre syndrome associated with a mitochondrial DNA deletion was described in one case (Pineda et al. 2006). Finally, one patient was identified with hypomyelination with atrophy of basal ganglia and cerebellum syndrome (Mercimek-Mahmutoglu and Stockler-Ipsiroglu 2007). Some cases demonstrated clinical improvement after treatment with folinic acid (Hansen and Blau 2005; Mercimek-Mahmutoglu and Stockler-Ipsiroglu 2007; Moretti et al. 2005; Ramaekers et al. 2002). This heterogeneity poses a challenge for the clinician and to our understanding of the association between neurological dysfunction and deficiency of folate metabolites in the CSF. Only recently clinicians have reported the presence of autism spectrum manifestations in some cases of CNS folate deficiency (CFD) (Moretti et al. 2005; Ramaekers et al. 2005). We described a 6-year-old girl with developmental delay, psychomotor regression, seizures, MR, and autistic features (Moretti et al. 2005). This child presented in the neonatal period with seizures and was subsequently diagnosed with developmental delay. After the identification of low folate levels in the CSF, the child was treated with folinic acid. This resulted in objective neurological improvements that allowed a standardized cognitive and behavioral assessment using the Bayley Scales of Infant Development, the Vineland Adaptive Behavior Scales, the Autism Diagnostic Observation Schedule (ADOS), and the Autism Diagnostic Interview—Revised (ADI-R). These tests showed that despite improvement of motor skills and parental reports of increased responsivity, her cognitive, language, and socialization skills remained delayed. The child met criteria for MR and exceeded the cutoff criteria for autism in all areas (nonverbal communication, reciprocal social interaction, play, and repetitive behaviors). Even after taking into consideration her cognitive and her language delays, she demonstrated deficits in communication and socialization that mirrored those observed in children with idiopathic autism. Very few clinical details were published concerning the diagnosis of MR and autism in 4 of 28 cases of CSF folate deficiency (Ramaekers et al. 2005). These authors noted improvement of communication skills in two of these subjects in contrast with results of Moretti et al.

Based on our observation (Moretti et al. 2005) and reports from other authors (James et al. 2004; Ramaekers et al. 2005), we hypothesized that autism is a manifestation of CFD. To investigate this association, we performed detailed clinical evaluations in six additional consecutive subjects with CFD. Standardized neurobehavioral assessments were performed in four of these subjects whose cognitive function was adequate enough to allow this evaluation. The two remaining subjects were evaluated clinically, but were severely impaired and non-ambulatory, precluding any type of formal neurobehavioral assessment. In addition, genes encoding proteins that mediate folate transport or metabolism were sequenced in the whole cohort.

Methods

Patients were referred to Baylor College of Medicine because of an ongoing study regarding children with cerebral folate deficiency. The referrals originated from geneticists aware of the existence of this study. The seven subjects described in this report represent seven consecutive cases evaluated by the authors. Cerebrospinal fluid (CSF) analysis was performed by HPLC as previously described (Hyland and Surtees 1992). The normative range for CSF 5-MTHF was developed over a period of 15 years using the following methodology. Cerebrospinal fluid was collected from patients spanning 0 to 79 years of age in a standardized fashion from the first drop and the 7th 0.5 ml fraction was used for the analysis of 5-MTHF. Over 2,500 CSF samples were analyzed from males and females from a wide ethnic origin and with a wide variety of neurological disorders. Values from patients treated with drugs known to interfere with folate metabolism (anticonvulsants, methotrexate) or affected with infections, seizures, movement disorders, inborn errors of folate, cobalamin, serine or biogenic amine metabolism, and disorders leading to altered methylation were excluded. Four hundred and eighty-nine values remained from the total data set and were used to develop age related reference ranges. The reference range used in the current paper was extracted from this set and defines the highest and lowest values obtained from subjects 2 to 15 years of age.

DNA sequencing, chromosome analysis by karyotyping and array-comparative genomic hybridization (CGH) were performed at the Baylor College of Medicine Medical Genetics Laboratories. Analysis of the PCFT gene was performed as previously described (Qiu et al. 2006). Five of the seven patients also received a formal psychological evaluation. The two remaining subjects were too severely impaired neurologically to undergo neuropsychological assessments. The Bayley Scales of Infant Development, Second Edition (BSID-II) was used to assess cognitive and motor skills. Parents were interviewed using the standardized administration of the Vineland Adaptive Behavior Scales—Interview Edition (VABS). All patients were also given the Autism Diagnostic Observation Schedule (ADOS) Module I. Module I was selected based on developmental level and expressive language abilities of the patients, and is appropriate for children who are nonverbal or have phrase speech, and who are functioning developmentally under the age of three. The parents of all patients were also given the ADI-R. Questions from the ADI-R were not only used to substantiate the diagnosis of autism, but were also utilized to document the presence of a true developmental regression. The ADOS and the ADI-R were administered by an expert clinician (SUP) with research reliability for both instruments. Additional research testing was performed after acquiring informed consent approved by the Institutional Review Board for Human Subject Research at Baylor College of Medicine and appropriate institutions.

Results

The clinical features and CSF analysis for the seven unrelated subjects are summarized in Table 1. All children demonstrated developmental delay, regression, dyskinesia and low CSF levels of 5-methyltetrahydrofolate (5-MTHF), the biologically active form of folate. Red blood cell folate, serum folate, total plasma homocysteine, karyotype, chromosomal microarray analysis, fragile X DNA testing, plasma and CSF amino acids, urine organic acids, and evaluations for mitochondrial disease did not show any abnormalities. The results of psychological testing are depicted in Tables 2, 3, and 4. Of the seven consecutive cases, five children were able to undergo formal neuropsychological evaluation. The poor neurological status of the two remaining subjects precluded this type of evaluation. All five patients who received neuropsychological testing were diagnosed with MR and autism or autism spectrum disorders. These children had difficulty with transitions, were more interested in objects than interactions with people, did not compensate for language delays by using gestures, had limited shared enjoyment in interactions, and exhibited repetitive/sensory behaviors. Even after taking into consideration their cognitive and language impairments, they demonstrated deficits in communication, socialization, play and repetitive behaviors as observed in idiopathic autism.

Oral treatment with folinic acid (5-Formyl-tetrahydrofolate) was initiated at 0.5 mg/kg/day and the dose was doubled after approximately 2 weeks. Clinical follow-up and repeat CSF analysis showed variable treatment response and correction of 5-MTHF levels (Table 1). There were improvements of gait, dyskinesia, and hand use in subjects 1–3 and 7. Subject 1 was also able to tolerate oral feeds.

No mutations were found in the exons and intron/exon boundaries of genes encoding proteins involved in folate transport (RFC, FRα, FRβ, PCFT), methylenetetrahydrofolate reductase (MTHFR), dihydrofolate reductase (DHFR), and formiminotransferase cyclodeaminase (FTCD).

Discussion

We have demonstrated the presence of autism spectrum manifestations in five subjects with CFD. All five patients who were formally evaluated achieved cutoff scores that placed them within the autistic (patients 1, 3, and 7) or autism spectrum (patients 2 and 6) ranges on the ADOS-G and the ADI-R. All five patients exhibited features consistent with those observed in children with idiopathic autism. The overlap in symptomatology between patients with idiopathic autism and patients with CFD is more evident in those patients with higher cognitive and language functioning and seems related to the severity of developmental regression and the level of developmental functioning prior to the regression. In those patients with autistic features and lower functioning in language and cognition (patients 1 and 6), we found that their specific difficulties with social communication cannot be fully accounted for by their delays, given that typically developing children with similar functional levels have acquired these social communication milestones.

Limitations of our study include the small sample number and the fact that the neurobehavioral evaluations were not performed in a blinded fashion. However, these patients with CFD clearly differ in their neurological presentations and complexity from those with autism alone since they present with accompanying neurological features such as seizures, hypotonia, movement disorders, and ataxia. Some of these features such as the seizures and hypotonia may become evident in the neonatal period. In addition, the regression observed in these patients with cerebral folate deficiency occurs at different points during development, whereas the regression in autism typically occurs prior to age 3 years.

Although the association of CFD and autism has been anecdotally reported by us and others (Moretti et al. 2005; Ramaekers et al. 2005), it has not previously been the subject of direct study. Our previous identification of one child with CFD and autism (Moretti et al. 2005) together with our current observations clearly establish an expanded phenotypic spectrum of CFD suggesting that children with psychomotor retardation, regression, seizures, dyskinesia and autism may benefit from analysis of CSF folate levels when the initial genetic and metabolic workup fails to identify an etiology. Clinical improvement with folinic acid treatment in a subset of our patients highlights the importance of considering this diagnosis. However, clinical improvement was observed only in cognitive, language, and motor domains with no amelioration of autistic features, which have remained consistent over time despite ongoing treatment with folinic acid. Moreover, it could be said that the autistic features became more prevalent as their level of developmental functioning improved. This suggests that autism is not simply the consequence of significant developmental delays—otherwise the autistic behaviors should decline as development improves.

Interestingly, in contrast to previous reports, four of the seven children described here manifested with neonatal abnormalities, while the remaining three subjects had a period of normal early development. This suggests that normal development in early infancy is not a necessary feature of CFD and that an abnormality of CSF folate may be present early in the neonatal period. Moreover, in our patients we did not observe the progressive deceleration of head growth that had been previously reported in children with CFD.

The variability in clinical history, neurological manifestations, and laboratory findings found in CFD suggests two considerations. First, the spectrum of neurological dysfunction associated with this metabolic abnormality remains to be fully characterized. Second, different pathways/disorders may lead to low CSF 5-MTHF. For instance, in some cases a severe impairment of CNS folate transport or metabolism may lead to early-onset of neurologic dysfunction, as occurs in hereditary folate malabsorption where CSF folate levels are very low (Geller et al. 2002). In contrast, a more subtle defect may allow some degree of compensation in the neonatal period and lead to clinically detectable impairments later in life. However, developmental delay/regression, dyskinesia, seizures, and autism seem to constitute core manifestations of this condition. The presence of isolated CSF folate deficits led us to consider the possibility of defective folate transport into the brain as the pathophysiological basis for CFD. However, we found no mutations in genes regulating folate transport and metabolism (Rosenblatt and Fenton 2000) suggesting either the presence of genetic alterations in regulatory pathways or a potential defect in folate turnover as possible causes for this neurometabolic condition. Although autoantibodies are found in some children with CFD (Ramaekers et al. 2005), their causative role remains to be determined. However, in the author’s opinion the presence of significant neurological deficits in the neonatal period in a subset of patients with CFD decreases the likelihood of an autoimmune etiology in these particular individuals. Regardless of the precise mechanism of CFD, we hypothesize that reduced 5-MTHF levels in the brain, reflected in the low CSF folate levels, lead to an impairment of methylation reactions in the brain (Rosenblatt and Fenton 2000) with potential consequences for protein function (protein methylation) and regulation of gene expression (DNA and/or histone methylation).

These observations also raise the question of the presence of CNS folate abnormalities in children with idiopathic autism and developmental regression. A recent study in 20 children with idiopathic autism identified biochemical abnormalities in the plasma consistent with impaired methylation capacity and increased oxidative stress (James et al. 2004). In a follow-up study, plasma methionine and the ratio of S-adenosylmethionine (SAM) to S-adenosylhomocysteine (SAH), an indicator of methylation capacity, were significantly decreased in 80 autistic children relative to 73 age-matched controls (James et al. 2006). Two of our patients (subjects one and two) exhibited elevated total homocysteine (424 and 138 nM, respectively, with a normal range of 32–113 nM) and increased S-adenosylhomocysteine (SAH) values (88.8 and 34.5 nM, respectively, with a normal range of 8.9–14.1) in CSF pointing toward a decreased capacity for methylation as a possible contributing factor in the development of autism. The elucidation of the role of methylation reactions in the pathogenesis of autism will require additional studies. However, the partial overlap of clinical manifestations between CFD and neurodevelopmental disorders including autism suggests the intriguing possibility that these conditions may share common molecular/cellular substrates or abnormalities of CNS methylation reactions.