We report the cases of two young adult siblings who experienced epilepsy as the only symptom of MTHFR deficiency during 14 years (including 8 under metabolic treatment) for one and 9 years for the other one. They harbored the stop-loss c.1970G > C mutation that replaces the stop codon by a serine, extending the MTHFR protein by 50 additional amino acids at its C-terminal segment. This stop-loss mutation was previously reported as homozygous in 2 severely affected French patients with early onset disease (< 1 year old) [9]. In our patients the allele with this stop-loss mutation also carried the c.665C > T, p.(Ala222Val) polymorphism in exon 5 (corresponding to the c.677C > T nucleotide change according to [8]), that was previously suggested to worsen MTHFR deficit for patients carrying other MTHFR variants [9]. The other mutation we identified (c.1162C > T) has never been associated with MTHR deficiency. Its mutation frequency in control databases is similar to that already reported for other MTHFR causing mutations [see Additional file 2]. Since our two patients first experienced symptoms in adulthood, it is likely that the c.1162C > T mutation does not severely affect MTHFR enzymatic activity, that was shown to be correlated to the severity of the disease [1]. Since the mutation is located outside the catalytic domain of the protein, lying within the predicted S-adenosyl methionine (SAM)-binding site, the protein could display altered binding to SAM while still retaining some residual enzymatic activity [10].
Froese and al. compared genotypes of patients with MTHFR deficiency according to early onset (< 1 year old; n = 64) versus late onset (> 1 year old; n = 51) [1]. They classified genotypes in seven categories according to the type of mutations (missense/splicing/other) and location of mutations (catalytic domain versus regulatory domain). They found a significant difference only for 2/7 genotype categories: late onset patients had more frequently two missense mutations located in the regulatory domain (14% vs 3%), and less frequently two splicing mutations in the regulatory domain (4% vs 19%). Missense mutations only correlate with a milder phenotype when both located in the regulatory domain, as two missense mutations in the catalytic domain were found equally in early-onset (28%) and late-onset (29%) patients. For the sub-category of very late onset patients (> 10 years old; n = 24) reviewed here, the data fit with those findings: 29% of patients had two missense mutations in the catalytic domain, 17% two missense mutations located in the regulatory domain, and 0% two splice mutations in the regulatory domain. Concerning clinical variability among siblings, it is interesting to observe that patient n°7, who suffered from a gait disorder since 15 years old, had an asymptomatic 37 year-old brother harbouring the same mutations. On the other hand, four other patients we included in our manuscript (n°5, 15, 19, 20) had siblings who had an earlier pediatric onset, however never < 1 year old. Therefore, the same genotype can lead to some limited variability of the clinical expression of the disease.
The review of all 24 patients shows that epilepsy occurs in 50% of adolescent/adult onset MTHFR deficient patients with a highly variable phenotype and a variable response to anti-epileptic drugs. The core symptom was gait disorder (96%) from both central and peripheral etiologies. Mode of onset was also variable, some patients experiencing sub-acute onset of symptoms, sometimes following chronic evolution of symptoms. Thrombotic events were not as frequent (5/24; 21% of patients) as reported in homocystinuriadue to cystathionine beta synthase (CBS) deficiency [11]. Even though almost all patients (21/24) ended up suffering from a combination of neurologic symptoms, at the onset, 76% of them (13/17) suffered from a single symptom. The delay from onset to occurrence of a second symptom could be as long as 9 years (patient n°2). The delay from onset to diagnosis was rather long (mean 5.75 years). Only two patients were diagnosed when only suffering from a single symptom, including our patient n°1 who was tested for homocysteinemia after the diagnosis of his sister. Brain MRI can help to achieve diagnosis, but thewhite matter changes observed arenot a constant or a specific sign.
Metabolic treatment is mainly based on B9, B12 vitamins, and betaine. Almost all patients have received those 3 components simultaneously (15/18). All five epileptic patients with data regarding evolution under metabolic treatment had decreased frequency and intensity of seizures, which allowed a decrease or a discontinuation of antiepileptic treatment. Among our case reports, patient n°1, had his epilepsy resolved and remained free of other symptoms for 8 years, until last follow-up. Interestingly, his older sister (patient n°2), diagnosed at a later age, experienced gait difficulties 9 years after the appearance of epilepsy, suggesting that the early start of metabolic treatment of her brother prevented the worsening of his disease. Among 18 patients for whom evolution under metabolic treatment was reported, all stabilized or improved clinically, while their homocysteinemia levels, although strongly reduced, never completely normalized. However, very few had a complete disappearance of their symptoms due to irreversible neurological damage accumulating over time, highlighting the need for shorter diagnostic delays in MTHFR deficiency.
This could be achieved if homocysteinemia was tested earlier as a screening test for MTHFR deficiency. Values reported in adolescent/adult onset MTHFR deficiency were consistently above 100 μM (4.5 < N < 15), even for very late onset patients, strongly evoking a genetically-sustained metabolic defect, and notvitamins or renal filtration deficiencies, which can also be associated withhyperhomocysteinemia.
Homocysteine is likely to promote thrombotic events, but it is not known why such events are far less frequent in MTHFR deficiency than in classic homocystinuria despite both deficiencies being associated with similar homocysteinemia levels [11]. Hypomethioninemia may decrease global methylation reactions in the central nervous system, hence possibly affecting myelin, as attested by white matter abnormalities often found in cerebral MRIs of MTHFR deficient patients [12].
In conclusion, these two patients broaden the phenotypic spectrum of epilepsy in adolescent/adult onset MTHFR deficiency. The literature review showed that epilepsy, and other isolated neurological symptoms like spastic paraparesis or cognitive decline, may be the unique manifestations of MTHFR deficiency during several years. Even though adolescent/adult onset MTHFR deficiency is a rare disease, it is a treatable one, for which metabolic treatment comprising B9, B12 and betaine can prevent disease progression and promote improvement. Assessment of homocysteinemiashould be performed in selected patients even if suspicion of MTHFR deficiency is low. We suggest that plasma homocysteine levels be tested in case of occurrence of the following symptoms with unknown etiologies:unexplained epilepsy with or without normal brain MRI, spastic paraparesis, motor predominant peripheral nerve disease with central signs, young onset cognitive disorder, encephalopathy, atypical psychosis (with visual hallucinations, cognitive disorder, drowsiness), and young onset thrombosis. In case of hyperhomocysteinemia, metabolic treatment should be started without delay.