In the present study, we demonstrated that a homozygous mutation (E320Q) in SLC19A3 is associated with epileptic spasms in early infancy, severe psychomotor retardation and characteristic MRI findings, including progressive brain atrophy and bilateral thalami and basal ganglia lesions. It should be noted that four patients in this family showed quite similar clinical presentations, strongly arguing for the distinct genotype-phenotype correlation associated with the homozygous E320Q mutation. Two apparently unrelated disease phenotypes, BBGD (MIM 607483) [10, 11] and Wernicke's-like encephalopathy (MIM 606152) , have been shown to be associated with mutations of SLC19A3. Patients of BBGD with homozygous mutations (G23V, T422A) in SLC19A3 showed childhood-onset encephalopathy, characterized by epilepsy, confusion, external ophthalmoplegia, dysarthria, dysphagia, dystonia, rigidity and quadriparesis. In these patients, the administration of high doses of biotin (5-10 mg/kg/day) early in the progression of the disorder eliminated their symptoms, whereas thiamin was ineffective [10, 11]. Recently, compound heterozygous mutations (E320Q, K44E) in SLC19A3 were shown to cause Wernicke's-like encephalopathy, which is characterized by acute onset of epilepsy, ataxia, nystagmus and ophthalmoplegia in the second decade of life, and symptoms are effectively alleviated with thiamin treatment .
Despite sharing one mutant allele (E320Q) with the known causative mutations in Wernicke's-like encephalopathy, our patients clearly differed in age at onset, brain MRI findings and symptoms, raising the possibility that the functions altered by the mutant SLC19A3 with K44E are substantially different (Table 1). Although both the BBGD phenotype and our patients showed similar bilateral basal ganglia lesions in the brain and encephalopathy that are fatal unless treated, the clinical presentations of our patients are distinct from the previously reported BBGD phenotype in the following ways: (1) epileptic spasms appeared in infancy in all of our patients, in contrast to the epilepsy onset in childhood for patients with BBGD; (2) MRI findings in our patients are characterized by progressive brain atrophy and additional lesions in the bilateral thalami that are not observed in patients with BBGD; (3) our patients did not display obvious dystonia and cogwheel rigidity reflecting dysfunctions in basal ganglia; (4) administration of a high dose of biotin for one year improved neither the neurological symptoms nor the brain MRI findings in Patient V-6, though it is unclear whether administration of biotin at early stages of the disorder (e.g., when epileptic spasms first appeared) improved the subsequent clinical trajectory (Table 1). Thus our cases broaden the phenotypic spectrum of disorders associated with mutations in SLC19A3.
Recently, two novel mutations, both of which created premature stop codons, were identified in BBGD patients . Transfection studies demonstrated that G23V and T422A mutations identified in the BBGD phenotype lead to mutant SLC19A3 that is nonfunctional for thiamin uptake activity . These findings suggest that loss-of-function mutations in SLC19A3 are associated with the BBGD phenotype. In contrast to these loss-of-function mutations, we demonstrated that mutant SLC19A3 with E320Q possesses some thiamin uptake activity that is approximately 60% of wild-type SLC19A3 in CHO cells  and HEK293 cells (Figure 3C). These findings alone seem difficult to account for severe clinical presentations in our cases. Interestingly, a previous report showed that the negative charge at position E320 was conserved in other transporters (SLC19A1, SLC19A2) and was crucial for the formation of a salt bridge with a conserved, positively-charged residue, K380 . The homozygous mutation of E320 may, therefore, destabilize the protein conformation and bring with more deleterious effect, including gain-of-toxic function and/or the dramatically changed protein structure.
In this study, the efficacy of high doses of biotin and/or thiamin for the clinical phenotypes of our patients has not been determined. This indicates that we have not determined whether our patients are phenotypic variation of "biotin-responsive" basal ganglia disease or "biotin-unresponsive" another type of a disease. Therefore, accumulation of genetic analysis and clinical courses from more patients who undergo biotin and/or thiamin treatment as well as further laboratory studies employing cellular or mouse models will be needed to better characterize the clinical phenotypes and biochemical characteristics of the homozygous E320Q mutation in SLC19A3.