Objective
Our objectives were to examine the extent of described sequence variation in the glucose transporter 3 ( GLUT3 ) gene in children with myelomeningocele (MM), identify novel variations in the GLUT3 gene in these children, and determine whether these variations may confer a risk of MM.
Study Design
We sequenced the 10 exons of GLUT3 , including exon-intron boundaries, on 96 children with MM. Sequencing was performed with Sanger methods and results analyzed with deoxyribonucleic acid analysis software. Frequencies of known single-nucleotide polymorphisms were identified, and those differing from the reference sequence (GRCh37/hg19 assembly) were considered variations.
Results
Six novel and 9 previously described, genetic variations were identified in our population. The novel variations included a large, 83 base pair deletion involving the core promoter region and part of exon 1 (1 of 96 children), and a 2 base pair deletion in the coding sequence of exon 4 (1 of 96 children). The remaining novel variations were located in the introns in the proximity of the splice sites. Novel mutations in GLUT3 were observed among 6.25% of our population. Additionally, the frequency of the rare allele for rs17847972 located in a splice donor site is higher ( P < .001) in MM in our population than expected.
Conclusion
We identified previously undescribed deletions and single-nucleotide variations involving the GLUT3 gene that may be associated with increased susceptibility to MM. Of particular interest, the 2 deletions involve both an important core promoter site and a coding region predicted to have a deleterious effect. The functional significance of these findings is under investigation.
Neural tube defects (NTDs) constitute a heterogeneous category of fetal malformations that result from the failure of neural tube closure by the fourth week of embryological development. NTDs are the most common structural central nervous system defect and occur at an incidence of 0.5-2 per 1000 live births worldwide. The majority of NTDs can be classified as anencephaly or spina bifida. The most common NTD associated with survival is myelomeningocele (MM), a severe form of spina bifida that occurs because of the defective closure of the caudal neural tube with herniation of the spinal cord and meninges. Most infants born with MM survive, and these individuals often have multiple disabilities and increased mortality rate.
The etiology of NTDs is not entirely understood but involves both genetic and environmental factors in association with critical timing during embryogenesis. Clustering of NTDs within families, and associations with various genetic syndromes, underlines the importance of identifying the underlying hereditary basis of NTDs. Maternal folate status has been established as an important factor in the development of NTDs. The association between folate deficiency and NTDs led to the mandated fortification of grain products in the United States in the late 1990s. This fortification has been associated with a 20-30% decrease in the NTD rate.
The mechanism by which folate deficiency causes NTDs remains unclear despite an extensive number of investigative studies. It appears that other genetic and environmental influences may contribute to a folate-resistant phenotype of NTDs. Teratogenic exposures to antiepileptic medications are associated with an increased risk of NTDs, although these risks may be mitigated with appropriate folate supplementation based on evidence from animal studies.
Other environmental factors with genetic controls implicated in the development of NTDs include derangements in glucose metabolism and maternal obesity. Mexican American women are particularly interesting in that they have the highest rates of offspring with NTDs, maternal obesity, and type 2 diabetes mellitus in the United States. A recent analysis from the National Birth Defects Prevention Study showed the factors most associated with NTDs were Hispanic ethnicity, maternal obesity, and low dietary folate intake.
Previous animal studies have shown an increased glucose levels during embryogenesis can alter the expression of proteins involved in glucose metabolism and homeostasis. Hyperglycemia during this critical period is associated with increased apoptosis and increased production of reactive oxygen species generation that favor cell death. More recent human studies have further exhibited an association between high maternal dietary glucose intake and the risk of NTDs in nondiabetic women.
Glucose transporter 3 (GLUT3) is a glycoprotein with 12 transmembrane domains that transports glucose across cell membranes and is a member of a superfamily of transport proteins comprised of 14 members. This group of proteins is encoded by the SLC2 family genes that is subclassified into 3 classes based on sequence similarity of which GLUT3 is in class 1. The GLUT3 ( SLC2A3 ) gene is located on chromosome 12p13.3 and is comprised of 10 exons. GLUT3 is expressed predominantly in neuronal tissue functioning as a high-affinity glucose transporter in conditions of low ambient glucose levels surrounding neurons.
Studies in animal models have shown that GLUT3 is critical in the both the developing pre- and postimplantation murine embryo. More recently GLUT3 expression in human placenta tissue has been shown to be present throughout gestation with a predominance in the first-trimester placenta. Increasing deoxyribonucleic acid (DNA) methylation of the GLUT3 gene has been reported to be associated with decreased expression throughout gestation; thus, GLUT3 as a placental transporter may be of greater significance in the first trimester during periods of embryogenesis.
In previous studies we have demonstrated associations between coding single-nucleotide polymorphisms (SNPs) in 3 genes ( HK1 , LEPR , and GLUT1 ) involved in glucose metabolism and NTDs. Additionally, we have described novel and rare variations in the GLUT1 gene associated with MM.
The objective of our study was to examine the relationship between the sequence of the GLUT3 gene and MM. We sought both to study previously identified polymorphisms and potentially identify new variations.
Materials and Methods
Children with MM and their parents were enrolled into the study from 1996 through 2006 from 3 primary sites (Houston, TX; Los Angeles, CA; Toronto, Canada). Study approval by the institutional review board of the University of Texas Health Science Center at Houston, and informed consent was obtained. Eight hundred sixty-four subjects were enrolled in the study, of which 96 were selected to be included in this project.
The characteristics of the study population are described in detail elsewhere. The children were selected based on ethnicity (48 Hispanic; 48 white). As mentioned previously, the Hispanic population was of particular interest because of the increased rates of NTDs, maternal obesity, and type 2 diabetes mellitus. Additionally, we included only MM-affected children born after the 1998 federally mandated folate fortification to select those children most likely to have had adequate dietary folate intake.
Blood samples were obtained from children and parents, and genomic DNA was extracted from blood lymphocytes utilizing the Puragene DNA purification kit (Gentra Systems, Minneapolis, MN). When a parent was not available for blood sampling, a saliva DNA kit was sent to obtain a sample and the DNA (DNA Genotek; Kanata, Ontario, Canada) was subsequently extracted from the sample. DNA was amplified with whole-genome amplification (Genomiphi; GE Healthcare, Piscataway, NJ), and working stocks of 10 ng/μL were prepared for polymerase chain reaction (PCR) and stored at –80°C. DNA from the selected children has been used in previous studies with consistent genotyping results.
PCR and nested-sequencing primers were designed based on the reference genomic sequences of GLUT3 (NM_006931, GRCh37/hg19 Assembly) extracted from the University of Santa Cruz Genome Browser (UCSC Genome Browser; http://genome.ucsc.edu/cgi-bin/hgGateway ). The primers were designed containing 50-100 bases in the flanking regions of the exons to include splice sites. A total of 10 PCR primer pairs were designed (available upon request) and synthesized by Integrated DNA Technologies USA (Commercial Park, Coralville, IA). PCR amplification of exons with MyTaq Hot Start DNA polymerase (Bioline USA Inc, Taunton, MA) was completed using the MJ Research PTC-100 Programmable Thermal Cycler (MJ Research, Waltham, MA). PCR product sizes were confirmed with 1.8% agarose gel electrophoresis.
The amplified exon DNA was treated with T7 exonuclease 1 and shrimp alkaline phosphatase (United States Biochemicals, Affymetrix, Cleveland, OH) to remove the excess nucleotides and primers. Amplicons were sequenced using the BigDyeTerminator (Applied Biosystems, Foster City, CA) reagents with nested sequencing primers. The sequenced product was resolved on the ABI3100 genetic analyzer (Life Technologies Inc, Grand Island, NY).
Our sequence results were analyzed using the DNA Sequencing Analysis Software version 5.1 from Applied Biosystems Inc. To identify variations in the sequences, we compared results manually to the reference sequences (NM _006931) from the UCSC Genome Browser. The white (Caucasian) reference population (CEU) includes whites from Utah included in the Centre d’Etude du Polymorphisme Humain Collection for mapping genetic markers and the National Heart, Lung, and Blood Institute Exome Sequencing Project. The Mexican American (MEX) reference population includes Mexican Americans recruited from Los Angeles, CA, used in the international HapMap project and data from the 1000 Genomes Project in the SNP database (dbSNP). Admixture analysis was performed to confirm the genetic equivalency between our white and Hispanic children and the referenced populations.
Our sample size of 96 children was selected primarily because of the size of the PCR plate (96 wells), allowing for the efficient utilization of DNA and reaction buffers. Additionally, the statistical convenience of utilizing 96 lies in the overall allele count of 192 (2 alleles per child × 96 children), allowing for the detection of rare alleles, described as a frequency of 0.5% or approximately 1 of 192 alleles.
Rare allele frequencies of all known SNPs identified in our population were tallied and compared with ethnically matched frequencies reported in the dbSNP using Fisher exact test. A two-tailed value of P < .05 was considered significant. We corrected for multiple testing using Bonferroni methods. Variations identified in our population and not reported in dbSNP were considered novel. These novel findings were then reverse sequenced with a reverse sequencing primer for confirmatory testing. Variations present on forward and reverse sequencing were considered valid. The parents of children found to have novel variations were sequenced to assess for the inheritance of the mutation.
Novel variations of particular interest were selected for expanded study in our entire cohort population of 864. In addition to our cohort, we added 92 Mexican American controls, recruited from the Houston area, and 93 white controls from the Human Variation Panel-Caucasian Panel of 100 (HD100CAU; Coriell Institute for Medical Research, Camden, NJ). Custom TaqMan SNP genotyping assays with fluorescently labeled probes were utilized for the interrogation. Genotyping results were read utilizing the 7900HT Fast Real-Time PCR System (Life Technologies) and analyzed using the 7900 HT Sequence Detection System Software.
Results
After sequencing the 10 exons of the GLUT3 gene, we found 6 previously undescribed variations including 2 deletions ( Table 1 ). A large 83 bp deletion (exon1: g:8088945_8088862del83bp) was found spanning the core promoter and 5′ untranslated region (UTR) of exon 1 in a single Hispanic child. We performed DNA sequencing of both parents of the child, neither of which had the deletion, thus confirming the variation as de novo.
| Variations | Functional site | A1/A2 | All MM (n = 96) | CEU (n = 48) | MEX (n = 48) | Vertebrate conservation a | miRNA b | TF c |
|---|---|---|---|---|---|---|---|---|
| Exon 1: g:8088945_8088862del83bp | Promoter/ 5′-UTR | 83 bp/– | 0.5% | 0% | 1% | NA | Y | Y |
| Intron 3: g:8085543C>G | Intron | A/G | 0.5% | 1% | 0% | 92% (22/24) | Y | N |
| Exon 4: c.470_471delTC | Coding-frame shift | TC/– | 0.5% | 0% | 1% | 100% (34/34) | Y | N |
| Intron 5: g:8083043T>C | Intron | T/C | 0.5% | 0% | 1% | 33% (8/24) d | N | N |
| Intron 6: g:8082222C>A | Intron | C/A | 0.5% | 0% | 1% | 53% (9/17) d | Y | N |
| Intron 9: g:8074240C>A | Intron | C/A | 0.5% | 0% | 1% | 31% (9/29) d | Y | N |
a Available vertebrate sequence data for the involved site of genetic variation, information listed as a percentage of vertebrate species with conservation of the sequence (University of Santa Cruz Genome Browser Multiz Alignments of 46 vertebrates)
b Predicted miRNA binding location with alteration or loss of binding with change to the variant allele (miRBase analyses, http://www.mirbase.org/search.shtml )
c Predicted transcription factor binding site with loss of binding with change to variant allele (TF Search; http://www.cbrc.jp/research/db/TFSEARCH.html ; and University of Santa Cruz Genome Browser Transcription Factor ChIP-Seq from ENCODE)
A second deletion, in a separate Hispanic child (exon 4: c.470_471delTC), was identified in the coding sequence of exon 4 ( Figure 1 ). The deletion resulted in a frame shift mutation in the coding sequence. DNA sequencing of both parents found the deletion was paternally inherited. Genotyping of 864 additional children using fluorescently labeled probes designed by custom genotyping Taqman assays (Life Technologies) to determine whether the deletion we identified in exon 4 (exon 4: c.470_471delTC) was present in our larger cohort was completed. As expected, the deletion mutation was not identified in any additional subjects, likely confirming the variation as an isolated, inherited mutation.
Four additional novel variations were identified, all occurring in different intronic regions of the gene. Of note, one of the variations (intron 3: g:8085543C>G) is located in a cryptic splice acceptor site and was identified in a single white child with MM ( Figure 2 ). The aligning sequences of this variant in humans with 24 known vertebrate sequences available through the UCSC Genome Browser (available at http://genome.ucsc.edu/index.html ) showed the reference allele to be highly conserved among vertebrates ( Figure 3 ). The 3 remaining variations (intron 5: g:8083043T>C; intron 6: g:8082222C>A; intron 9: g:8074240C>A) were also aligned and compared with vertebrate sequences and were noted to be highly conserved among primates and less highly conserved when compared with other vertebrates (information available upon request).
Nine previously described SNPs within or flanking the exons of GLUT3 were evaluated. Eight of the 9 SNPs evaluated had population frequencies reported both for the Caucasian reference population (CEU) and the Mexican American reference population (MEX). Heterozygosity was demonstrated in all 9 of the SNPs in white children and in all but 1 SNP (rs151144610) in the Mexican American children ( Table 2 ). Hardy-Weinberg equilibrium testing on all SNPs utilizing Haploview 4.2 (Broad Institute, Cambridge, MA) was performed, and all SNPs were in Hardy-Weinberg equilibrium ( P > .05) with the exception of rs17847972 ( P < .001).
| Variable | Caucasian (CEU) | Hispanic (MEX) | CEU/MEX | ||||
|---|---|---|---|---|---|---|---|
| SNP | A/A2 | All MM | MM | Reference | MM | Reference | P value |
| rs2541279 | T/C | 27% (52/192) | 30% (28/96) | 29% (2383/8600) a | 25% (24/96) | 30% (40/132) b | .8/.4 |
| rs17847967 | T/C | 20% (39/192) | 23% (22/96) | 14% (107/758) b | 18% (17/96) | 20% (26/132) b | .02/.7 |
| rs74059377 | G/A | 1% (2/192) | 1% (1/96) | 0% (0/758) b | 1% (1/96) | 0.7% (1/132) b | .1/1 |
| rs17847972 | C/T | 47% (91/192) | 49% (47/96) | 0.1% (1/758) a | 50% (48/96) | 6% (8/132) b | < .001/< .001 |
| rs41438344 | C/G | 1.5% (3/192) | 1% (1/96) | 0.2% (16/8600) a | 2% (2/96) | 8% (11/132) b | .2/.08 |
| rs741361 | A/G | 41% (78/192) | 43% (41/96) | 37% (280/758) b | 39% (37/96) | 36% (48/132) b (42/98) c | .3/.7 |
| rs151144610 | C/T | 0.5% (1/192) | 1% (1/96) | 0.6% (5/758) a | 0% (0/96) | 0% (0/132) b | .7/– |
| rs25684 | G/A | 40% (77/192) | 43% (41/96) | 48% (4119/8600) a | 37.5% (36/96) | 44% (58/132) b | .3/.3 |
| rs112948014 | AG/– | 2% (4/192) | 1% (1/96) | ND | 3% (3/96) | ND | –/– |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
