|Home | About | Journals | Submit | Contact Us | Français|
To determine causative mutations and clinical status of seven previously unreported kindreds with TRMA syndrome, (Thiamine Responsive Megaloblastic Anemia, OMIM #249270), a recessive disorder of thiamine transporter Slc19A2.
Genomic DNA was purified from blood, and SLC19A2 mutations were characterized by sequencing PCR-amplified coding regions and intron-exon boundaries of all probands. Compound heterozygotes were further analyzed by sequencing parents, or cloning patient genomic DNA, to ascertain that mutations were in trans.
We detected 9 novel SLC19A2 mutations. Of these, five were missense, three nonsense, and one insertion. Five patients from four kindreds were compound heterozygotes, a finding not reported previously for this disorder, which has mostly been found in consanguineous kindreds.
SLC19A2 mutation sites in TRMA are heterogeneous; with no regional “hot spots”. TRMA can be caused by heterozygous compound mutations; in these cases, the disorder is found in outbred populations. To the extent that heterozygous patients were ascertained at older ages, a plausible explanation is that if one or more allele(s) is not null, partial function might be preserved. Phenotypic variability may lead to underdiagnosis or diagnostic delay, as the average time between the onset of symptoms and diagnosis was 8 years in this cohort.
Thiamine Responsive Megaloblastic Anemia (TRMA, OMIM 249270), also known as Roger’s syndrome1, 2, is characterized by the clinical triad of megaloblastic anemia, non-type I diabetes mellitus and sensorineural deafness. In addition other clinical findings, such as optic atrophy, congenital heart defects, short stature, and strokes have been described in a few cases (see Table 3). The disease can manifest anytime between infancy and adolescence and often not all the cardinal findings are present initially. The bone marrow reveals an unusual combination of ringed sideroblasts and megaloblasts. Anemia typically improves significantly with pharmacological doses of thiamine (25–75 mg/day compared to US DRI of ~1.1 mg/day), although the erythrocytes remain macrocytic. Variable improvement in diabetes is also noted. However, the hearing loss is apparently irreversible, though delay in onset of deafness may be possible.
The SLC19A2 gene, on chromosome 1q23.3, is implicated in all published TRMA patients (see Table 1). The SLC19A2 gene encodes a 497 amino acid high-affinity thiamine transporter. This protein, named by its relation to the previously described reduced folate carrier protein - SLC19A1, is predicted to have twelve transmembrane spanning regions. The structure of the SLC19A2 protein itself has not been determined. Analysis of SLC19A2 mutations revealed significant heterogeneity, although the majority are predicted to be null, with premature translation termination due to nonsense or frameshift mutations. All patients reported thus far have been homozygous for the mutations identified (Table 1). Here we report the first four compound heterozygous mutations and additional novel mutations identified, along with the clinical features of the patients.
All patients were referred to our laboratory for analysis of the SLC19A2 gene. We obtained informed consent from either affected individuals or their parents, as appropriate and the study “Genetic Basis of Thiamine-Responsive Anemia” was approved by the Children’s Hospital Boston Committee on Clinical Investigation.
Whole genomic DNA was isolated from peripheral blood of each subject using the Puregene DNA purification Kit (Gentra Systems), and quantified by spectrophotometry.
Exons and intron-exon junctions were amplified by PCR using the genomic DNA under previously reported conditions and primers.3 The fragments were purified with Qiagen purification kits (Qiagen). We determined the sequence of both DNA strands using fluorescent dye-termination chemistry at the Children’s Hospital Boston Molecular Genetics Core Facility. Sequencing results were analyzed using the program Sequencher 4.8 DNA sequence assembly software (Gene Codes). The sequence was compared with the wild type human SLC19A2 gene (GenBank accession: AJ238413) and mutations were confirmed by independent reamplifications. The numbers for the nucleotide changes are reported in accordance with GenBank entry GeneID: 10560 with the A of the initiator methionine codon, and the lead methionine each numbered as 1.
Furthermore, the absence of the missense variants in healthy normals was confirmed in 50 unrelated individuals (i.e. 100 chromosomes) supporting the likelihood that the variants are pathogenic.
The phase of variants found in the compound heterozygotes was assessed by sequencing the unaffected parents when available. In one case, where the parents were unavailable for testing, the gene fragment containing both of the variants was cloned in competent E Coli using the TOPO TA cloning kit (Invitrogen). The plasmid DNA, from the different colonies was isolated using a miniprep kit (Qiagen) and sequenced to determine whether the two variants were in cis or trans.
In addition to the clinical triad of thiamine responsive megaloblastic anemia, diabetes mellitus and sensorineural deafness, various other features were observed, such as short stature, various forms of visual impairment, developmental and speech delay, petit mal seizures and cryptorchism. The clinical findings of the patients in our series are detailed in Table 2 and and33.
Notably, in one case (Patient 2), a compound heterozygote, the hearing is normal even at 15 years of age. Another compound heterozygote (Patient 5) has only minimal sensory-neural hearing deficit at 30 years of age. In both these cases the diagnosis of TRMA was considered even in the absence of the classical triad of symptoms as the anemia and diabetes responded well to pharmacological doses of thiamine. The diagnosis of TRMA was only considered several years after the patient first became symptomatic. This suggests that some individuals with mutations in the SLC19A2 gene may not immediately be recognized as they do not present with the typical clinical triad and variable severity of phenotype. As already mentioned many of these patients present with additional signs and symptoms, thus the clinical presentation can be sometimes misleading.
By determining the sequence of SLC19A2 in all probands we identified nine novel mutations amongst these seven families. Of these, five were missense mutations (Table 2). Five patients were found to be compound heterozygotes. By sequencing DNA of the unaffected parents, or cloning patient genomic DNA, the variants were confirmed to be on different alleles (“trans”). This is important because if they were cis, the patients would be predicted to have one normal allele.
We identified nine novel mutations and one previously known mutation among seven TRMA kindreds. Six of these were missense mutations, three nonsense and one insertion; these can be added to the prior 19 distinct mutations in SLC19A2 (Table 1). These mutations are spread across the first four exons. Among 28 SLC19A2 mutations now known in about 70 total reported patients, only four variants appear in multiple families and of these only two, 484C>T and 515G>A likely arose independently in different ethnic populations (Figure 1), while others are potentially from single founders or limited geographic regions. The majority of the mutations are null mutations as a consequence of nonsense or frameshift mutations.
Although missense mutations could conceivably yield interesting structure/function relationships, or correspond to variable phenotype, not enough data exist to make these correlations. Only five of the previously described mutations are missense, and some of those five impair proper processing and cellular localization4, and may be “functionally null”. In prior reports, all TRMA kindreds had homozygous mutations in SLC19A2, usually due to consanguinity or ethnic isolation and homozygosity by descent. However four patients in this group are compound heterozygotes (Table 2). Two of these patients present with less severe hearing phenotypes. Patient 2 has normal hearing at age 15, while Patient 5 had a first observation of mild hearing loss at age 30. It remains unclear if these milder phenotypes can be explained by compound heterozygous missense mutations that are not functionally null.
Although the clinical triad of anemia, diabetes and deafness defines the syndrome of TRMA, other clinical signs or symptoms are often observed in TRMA patients (Table 3). These associated findings also add difficulties to attempts to correlate genotype and phenotype: for example, optic atrophy, retinal changes, and stroke (or stroke-like episodes) are certainly features of TRMA, but seen only in a minority of cases, even with homozygous null mutations. Although the clinical presentation of TRMA shares some features with Wolfram Syndrome (DIDMOAD, OMIM # 222300) and a variety of mitochondrial disorders, these are known to be distinct.5
The pathophysiology of TRMA syndrome has been elucidated in metabolic studies in fibroblasts from patients, and in a mouse model deficient in Slc19a2, the murine homolog of SLC19A2: Megaloblastosis is due to defective de novo synthesis of nucleic acids, because limited intracellular thiamine affects the pentose phosphate pathway and production of ribose-5-P.6 Observed ringed sideroblasts and ineffective erythropoiesis in TRMA are most likely due to a decrease in succinyl-CoA, which is needed for heme biosynthesis and generated by the action of a thiamine-requiring mitochondrial enzyme, α-ketoglutarate dehydrogenase.7 The Slc19a2 deficient mice become deaf due to a loss of cochlear inner hair cells.8 Although patients’ cochleae cannot be examined as for the mice, this spatial distribution of transporters as a cause of the particular pathology is both plausible and likely. Most tissues of normal mammals express two high-affinity thiamine transporters, both SLC19A2 and SlC19A3, so that if one is missing, it is supposed that the other protects those tissues (brain, heart, liver, skeletal muscle, intestine, for example). However, SLC19A2 is the sole transporter in marrow, pancreatic beta cells, and a subset of cochlear cells8, 9, and therefore deafness, diabetes and anemia are the consequences of deficiency.
A potential limitation of this survey of mutations is one of ascertainment bias. When the TRMA gene was under study in the late 1990s, investigators intentionally accumulated kindreds from consanguineous backgrounds for homozygosity mapping.2 More recently, clinicians identify TRMA when confronted with the full TRMA syndrome (diabetes, macrocytic/thiamine-responsive anemia, and hearing loss). We predict that more mildly affected patients, either with later onset of manifestations or without the cardinal clinical triad, remain to be discovered. We predict that to some degree, less severely affected patients may be expected to have less drastic mutations in the SLC19A2 gene. Many adult patients are now known, but only a few presented as adults. Our study remains open to evaluate research samples from full blown TRMA, but also from subjects with only two suggestive symptoms (e.g. hearing loss and thiamine-responsive anemia, anemia with diabetes, or diabetes and deafness). It is likely that the age-range and spectrum of clinical findings in SLC19A2-related disorders will continue to widen as we learn more about the disorder.
This work is supported in part by NIH grant K24HL004184 to EJN.
We express our gratitude to the patients and their families. We thank Yue-Hua Huang for technical assistance and Eric Shieh for preparing figure 1.