Tracheo-esophageal fistula (TEF) with or without esophageal atresia (EA) is a relatively common malformation, occurring in approximately 1 in 3500 births (Torfs et al.,¹⁹⁹⁵). Approximately half of patients have accompanying malformations, such as occur in Feingold syndrome, anophthalmia-esophageal-genital syndrome, CHARGE syndrome, vertebral anomalies, anal atresia, cardiovascular anomalies, tracheoesophageal fistula, esophageal atresia, renal or radial anomalies, or limb defects (VACTERL) association, and other, less well-characterized conditions involving multiple malformations. Among these accompanying congenital anomalies, cardiac defects are especially common (reviewed in Shaw–Smith,²⁰⁰⁶; de Jong et al., ²⁰¹⁰).

Although the causes of some of the syndromic conditions that may include TEF/EA have been defined, the etiologies of others are less well-understood. TEF/EA is viewed as a complex, heterogeneous disorder resulting from multiple interacting genetic and environmental factors. Many questions remain regarding the genetic contributions to isolated TEF/EA, although there are hints from biologic models implicating key pathways including the Sonic hedgehog signaling network (reviewed in de Jong et al., ²⁰¹⁰).

Specific chromosomal anomalies, such as trisomy 13, 18, and 21 have long been known to be associated withTEF/EA, as well as with other congenital malformations (Torfs et al.,¹⁹⁹⁵). Chromosomal rearrangements visible on routine karyotype, as well as submicroscopic pathogenic copy number variations (CNVs) of numerous chromosomal loci, have also been implicated as causing TEF/EA. Some of these genomic imbalances have been reported in multiple affected individuals, while others have been described in only a single patient (Felix et al.,²⁰⁰⁷). Overall, these results support a heterogeneous model of causation.

We performed high-density microarrays on a cohort of 20 patients ascertained through our study on VACTERL association, with the hypothesis that this testing modality would reveal potentially explanatory genomic imbalances in some patients. One patient had a submicroscopic de novo deletion of chromosome 20q13.33, part of which did not overlap known regions of copy-number variation. Of note, while this patient met inclusion criteria for our study, most clinicians and researchers would not judge him as having “classic” VACTERL association. However, the deletion found in this patient suggests GTPBP5 as a key gene contributing to the patient's presentation.

Many different genomic imbalances have been reported in patients with TEF/EA (Felix et al.,²⁰⁰⁷). The deletion found in this patient is especially interesting as it suggests only two specific genes as being contributory: among the genes in the patient's deleted region, GTPBP5 and LSM14B are the only ones that have not been reported as copy number variations in normal controls. Relatively little is known about the function of these genes, but interestingly, disturbances of the GTPBP5's function result in abnormal mitochondria (Hirano et al.,²⁰⁰⁶). Mitochondrial dysfunction is a well-documented link to congenital malformations such as those seen in VACTERL association, which this patient has features of, even if he were not deemed to meet strict diagnostic criteria (Damian et al.,¹⁹⁹⁶; von Kleist–Retzow et al.,²⁰⁰³; Thauvin–Robinet et al.,²⁰⁰⁶; Solomon et al.,²⁰¹¹).

From a developmental standpoint, LAMA5 is another interesting candidate gene, although with some strong caveats. LAMA5-null mice die in utero with multiple malformations; hypomorphic animals display anomalies such as renal and pulmonary malformations (Nguyen et al.,²⁰⁰²; Shannon et al.,²⁰⁰⁶). Further, LAMA5 interacts with Sonic Hedgehog, a pathway frequently implicated in conditions that include TEF/EA such as VACTERL association (Kim et al.,²⁰⁰¹; Gao et al.,²⁰⁰⁸). However, the mouse models do not correlate well with the specific malformations seen in this particular patient, and further, deletions of this gene have been reported in normal controls (Jakobsson et al.,²⁰⁰⁸).

The deleted region overlaps that reported in another patient, although this previously reported patient had a much larger deletion, making further comparisons quite speculative. This patient, like the one described here, was nondysmorphic; her phenotype interestingly included cardiac anomalies, as well as hydronephrosis (Traylor et al.,²⁰¹⁰).

The fact that the deletion is de novo offers some evidence of possible pathogenicity. One additional possibility that cannot be ignored involves a recessive model in which the patient had, for example, a deletion of one allele and a point mutation in the other. However, further studies would be required to investigate this possibility.

Finally, it must be emphasized that while the patient was ascertained through a study focusing on VACTERL association, he had only two strict component features of VACTERL association, and thus should not be considered to have this diagnosis (although he also had genitourinary anomalies, which are frequent in patients with VACTERL association; Solomon et al.,^2010b). Nevertheless, we hope that these results may be extrapolated to inspire further research on the causes of TEF/EA through molecular studies of GTPBP5.

The authors would like to express their gratitude to the participating patients and families, and would also like to thank Dr. Maximilian Muenke for his support and mentorship. The authors appreciate the assistance of the NIH Undiagnosed Diseases Program (including Drs. William A. Gahl and Thomas C. Markello) in evaluating the SNP array data. This research was supported by the Division of Intramural Research, National Human Genome Research Institute, National Institutes of Health and Human Services, United States of America.