Neuroradiologic Features in X-linked α-Thalassemia/Mental Retardation Syndrome

Discussion

This is the first report on a comprehensive study of brain CT/MRI findings of patients with ATR-X syndrome, though some case reports or personal accounts mentioned some findings of brain MRI/CT, including mild cerebral atrophy, partial or complete agenesis of the corpus callosum, and hypoplasia of the white matter.^5,6 Our results suggest that brain MRI/CT can present a broader spectrum of abnormalities in both white and gray matter than has been expected in ATR-X syndrome. Some of our patients with ATR-X were referred to our medical center because their MRI/CT scans showed white matter abnormalities, or myelination; and for some, leukodystrophy was suspected as their diagnosis. Therefore, we categorized the MRI/CT findings of 27 patients with ATR-X into 5 types from the point of view of white matter lesions, or myelination.

More patients showed nonspecific brain atrophy (type 1) on brain MRI/CT. This result agrees with the finding that, in a review of 168 patients with ATR-X syndrome, 77% presented with microcephaly.⁷ This brain atrophy in ATR-X syndrome is not progressive, and it seems to be due to the reduced production of neurons or glia in the perinatal period, not to some destructive process. This characterization was supported by a study by Bérubé et al⁸ by using a conditional targeting approach with mice, showing that ATRX is a critical mediator of cell survival during early neuronal differentiation and that increased neuronal loss may contribute to severe mental retardation.⁸

Types 2 and 3 suggest an abnormality of the white matter, or myelination. Type 2 includes white matter abnormalities, especially around the trigones, with increased signal intensity showing on T2WI/FLAIR in the periventricular region. Persistently high signal intensity in this peritrigonal area, or the terminal zones, is seen throughout the first decade of life, and it is sometimes very difficult to differentiate from white matter injury resulting from prematurity, which results in periventricular leukomalacia. Moreover, a layer of myelinated white matter is present between the trigones of the ventricle and the terminal zones in healthy patients.⁹ The findings shown in our patients with ATR-X differ from slow myelination in these areas or terminal zones, being normal variants in their distribution and signal intensity. Because these patients did not present clinically with apparent neonatal asphyxia and their findings do not show other signs suggesting periventricular leukomalacia, these white matter abnormalities may be relatively frequent findings in CT/MRI of ATR-X syndrome.

Especially type 3, with its widespread and scattered white matter abnormalities, strongly suggests that normal expression of the ATRX protein is involved in normal myelination because the patient's mutation is a nucleotide substitution in 5′-UTR and ATRX expression should be reduced, though the ATRX protein structure is normal. Actually, red cells of case 1 had more hemoglobin H inclusions (5%) than those of other patients, and he had hypospadias. Most interesting, his development was much better than that of typical patients with ATR-X, and he can walk and speak a few words. This suggests that the quantity of the ATRX protein can affect α-globin expression, genital development, and myelination, but the quality of the ATRX protein is more important for intellectual ability. Another possibility is that the site of the mutation may be more important for ATRX expression in the oligodendroglia, which is involved in myelination. The delayed myelination in type 4 also supports the idea that ATRX is important for normal myelination.

High signal on T2WI/FLAIR and no alteration on DWI in the white matter of these patients meaning that there is no cytotoxic edema but increased water content in the white matter region, and these findings may suggest the prematurity of the blood-brain barrier in ATR-X syndrome, though our data have limitations in that not all MRI/CT scans include diffusion imaging or contrast enhancement.

Most interesting, 2 patients with ATR-X with partial duplications of the ATRX gene, resulting a severe reduction of ATRX messenger RNA and absence of the ATRX protein, presented with normal brain MR imaging findings and agenesis of the corpus callosum, respectively.¹⁰ The former patient was evaluated before his death at 4.5 months of age, and it was difficult to detect white matter abnormalities. For the latter patient, although the age when the MR imaging was evaluated was 6 months, we cannot specify his age. He could have been too young for evaluation of myelination on MR imaging, and agenesis of the corpus callosum is the ultimate result of white matter abnormality. That finding does not contradict our theory that ATRX is important for normal myelination. However, more data should be forthcoming to clarify the relation between the ATRX protein and white matter or myelination.

Progressive brain atrophy in type 5 is an exceptional finding of CT/MRI in ATR-X syndrome. The patient with type 5 presented with a severe developmental delay and intractable epilepsy following West syndrome, which is rare in ATR-X syndrome. He had no episodes of hypoxic encephalopathy. Although it is possible that he may have had some other pathologic conditions, ATR-X syndrome should be considered in the differential diagnosis of patients with progressive brain atrophy on brain MRI/CT. His mutation was in int 35, or the last intron, resulting in skipping exon 35, which consists of 126 base pairs, and the introduction of a 43-amino-acid-deleted ATRX protein. This truncated protein may be related to his severe intellectual disability or epileptic condition.

A gray matter abnormality seems more frequent in those patients with ATRX mutation in the ADD domain, or exons 8, 9, and 10, and a white matter abnormality in the chromatin-remodeling domain, or exons 18 to 31. These results would indicate that mutations in the ADD domain produce more severe and permanent psychomotor deficiencies than those in chromatin-remodeling domains, though the clear phenotype-genotype correlation remains to be established.¹¹ However, conclusion may be too hasty because we have studied fewer cases with ATRX mutations in the chromatin-remodeling domains.

A number of intellectual disorders have been identified whose gene products regulate chromatin and chromosome architecture, and ATR-X syndrome is a disease of chromatin remodeling, as well as Rett syndrome (MeCP2) and Cornelia de Lange syndrome (SMC1A, SMC3, and NIPBL).^3,12 These 3 diseases share clinical manifestations, including severe intellectual disabilities, and ATRX, MeCP2, and cohesion proteins, respectively, indicating that these syndromes directly or indirectly interact with each other. Although these syndromes share a common final pathway to their pathogenesis, there are no common characteristic findings on brain MRI/CT among these diseases, except for brain hypoplasia or atrophy. These abnormal signals in the white matter on MR imaging seem to be relatively specific findings in ATR-X syndrome.

We propose that these findings should be included as associated (<50%) or supporting features of the diagnostic criteria for ATR-X syndrome. We will evaluate the brain MR imaging/CT of additional patients to validate these findings and update the consensus for the neuroradiologic features as a diagnostic criteria of ATR-X syndrome in the near future.