Non-synonymous variants in Pre B-cell leukemia homeobox (PBX) genes are associated with congenital heart defects

Abstract

Congenital cardiac malformations are one of the most common birth defects and most are believed to be multigenic/multifactorial in nature. Recently mice lacking Pre-B cell leukemia transcription homeobox (PBX) genes were created and found to have a range of ventricular outflow tract (OFT) malformations. Therefore, we screened 95 patients with congenital heart defects, including OFT malformations, for variants in genes encoding PBX proteins, as well as interacting proteins. The coding exons of PBX1-4, PKNOX1, PKNOX2, MEIS1-3, and PBXIP1 were amplified by polymerase chain reaction and the products analyzed on a Lightscanner. Samples with abnormal melting profiles were analyzed by DNA sequencing. Seven non-synonymous variants (6 novel and 1 SNP) were identified in 5 proteins (Pbx3, Pbx4, Meis1, Meis3 and Pknox1). One Pbx3 variant, p.A136V, is located in a highly conserved polyalanine tract and predicted to be deleterious. This variant was present in 5.2% of heart defect patients compared with 1.3% of 380 race- and ethnicity-matched controls (P<0.05). None of the other variants were predicted to be damaging. In conclusion, our results support the Pbx3 Ala136Val variant as a modifier or risk allele for congenital heart defects and implicate PBX-related genes as candidates for CHD, especially those affecting the cardiac outflow tract.

Introduction

Congenital heart defects (CHD) constitute a major portion of clinically significant developmental abnormalities with an incidence of 3–6% [1–3]. For many forms of CHD the etiology is multifactorial, with environmental and genetic factors playing important roles in determining the expressivity within a genetically predisposed family.

Pre-B cell leukemia homeobox (Pbx) proteins belong to the Pre-B cell leukemia (PBC) transcription factor family and TALE superfamily. These proteins are characterized by a conserved PBC motif and a three-amino-acid loop extension (TALE) in their homeodomains. To date four members of the PBC family, Pbx1-4, have been identified in mammals [4]. Numerous other protein partners have also been identified, including other homeodomain containing proteins such as the MEIS/PREP family, and non-homeodomain containing proteins, such as Foxc1, hematopoietic Pbx1 interacting protein and zinc finger interacting protein [4]. Several studies have described the effects of knocking out Pbx genes in mice. In a recent study of Pbx1 deficient mice it was shown that Pbx1 plays a significant role in patterning of branchial arch arteries and formation of the ventricular outflow tract (OFT), leading to numerous vascular abnormalities and the absence of OFT septation [5]. In a follow-up study the same group examined the cardiac phenotypes of all possible combinations of null alleles of the Pbx1-3 genes. Embryos with different combinations of Pbx mutations displayed a spectrum of cardiac malformations in the OFT that ranged from relatively mild defects such as bicuspid aortic valve (BAV), to more severe defects including tetralogy of fallot (TOF), double outlet right ventricle (DORV), and persistent truncus arteriousus (PTA). In addition, knocking out Meis1 resulted in a mild form of TOF [6].

Based on these results, we hypothesized that defects in the PBX gene family might be found in patients with congenital heart defects, particularly those with OFT malformations, and a screen of our CHD patient population was undertaken.

Materials and Methods

With University of Utah Institutional Review Board approval, children and young adults diagnosed with congenital cardiac malformations were enrolled in the University of Utah Pediatric Cardiology Genotype-Phenotype Core and blood samples collected for DNA isolation. Medical records were reviewed for demographic and clinical data.

DNA was isolated from peripheral blood samples using a Gentra Autopure LS (Qiagen, Valencia, CA). PCR primers were designed to amplify the coding exons of the PBC associated genes using the Exon Primer utility (http://ihg.gsf.de/ihg/ExonPrimer.html). All patient DNA samples were amplified by PCR in duplicate, along with negative (water) controls, using Platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA); primer sequences and amplification conditions are available on request. After amplification the PCR products were analyzed on a Lightscanner, as previously described [7]. Samples giving abnormal curves were purified and sequenced, as previously described [7]. The allelic frequencies of novel non-synonymous variants in the population were determined by screening a panel of ethnicity- and race-matched controls and from the NHLBI exome sequencing project database (http://evs.gs.washington.edu/EVS/). Frequencies were compared between groups using both Chi-square and Fisher’s two-tailed exact tests to determine statistical significance. In addition, in silico analysis of the variants on protein structure/function was investigated using SIFT [8], Panther [9] and Polyphen2 [10].

Results and Discussion

Ninety five Caucasian patients with a spectrum of congenital heart lesions including BAV, aortic stenosis (AS), coarctation of the aorta (COA), hypoplastic left heart syndrome (HLHS), DORV, PTA, transposition of the great arteries (TGA), TOF and atrioventricular septal defect (AVSD) were screened for mutations in the following genes: PBX1-4, PKNOX1, PKNOX2, MEIS1-3, and PBXIP1. Seven non-synonymous variants were identified (Table1). None of the patients harboring these variants had parents or siblings with clinically apparent cardiac malformations even though all of the variants were inherited.

Table 1.

Non-synonymous variants identified in the patients with OFT malformations and in controls

Gene	Exon	Number of Patients (Allelic Frequency)	Phenotypes	Variant	SIFT Prediction (Score)	Panther Prediction (Score)	Polyphen2 Prediction (Score)	Allelic Frequency in Controls	ESP Database Allelic Frequency*
PBX3	8	5 (2.6%)	TOF AVSD PTA BAV/COA HLHS	c.407C>T (p.Ala136Val)	Tolerated (0.10)	Deleterious (−4.68688)	Probably Damaging (0.980)	5/760 (0.66%)	18/2484 (0.72%)
PBX4	1	1 (0.5%)	AVSD	c.19C>T (p.Pro7Ser)	Tolerated (0.43)	No prediction	Unknown	0/760 (0%)	0/2700 (0%)
	3	1 (0.5%)	COA	c.308G>A (p.Gly103Glu)	Tolerated (0.34)	Benign (−1.18809)	Benign (0.034)	0/760 (0%)	0/2700 (0%)
	4	2 (1.1%)	COA/TGA HLHS	c.461C>T (p.Thr154Met)	Tolerated (0.08)	Benign (−2.56188)	Benign (0.104)	3/190 (1.58%)	8/2696 (0.30%)
MEIS1	8	1 (0.5%)	DORV	rs61752693 (p.Arg272His)	Tolerated (0.08)	Benign (−2.49403)	Benign (0.006)	0/760 (0%)	0/2100 (0%)
MEIS3	4	1 (0.5%)	AS	c.330G>T (p.Arg117Leu)	Damaging (0.00)	Benign (−2.9254)	Possibly Damaging (0.802)	1/760 (0.13%)	0/2650 (0%)
PKNOX1	11	1 (0.5%)	DORV	c.1238A>G (p.Glu413Gly)	Tolerated (0.22)	Benign (−1.83073)	Benign (0.000)	0/760 (0%)	0/2650 (0%)

^*Allele frequency for the European/American Population. For variants not detected the approximate number of exome sequence analyses for that gene is used as the denominator.

TOF: Tetralogy of Fallot; AVSD: Atrioventricular Septal Defect; PTA: Persistent Truncus Arteriosus; BAV: Bicuspid Aortic Valve; COA: Coarctation of the Aorta; HLHS: Hypoplastic Left Heart Syndrome; TGA: Transposition of the Great Arteries; AS: Aortic Stenosis; DORV: Double Outlet Right Ventricle.

In Pbx3 we identified a missense variant (p.Ala136Val) in 5 patients with diagnoses including BAV/COA, HLHS, TOF, AVSD and PTA (allelic frequency 0.026). This variant was also identified in 5 of 380 Caucasian control subjects (allelic frequency = 0.0066), which was significantly less frequent than in the patient cohort (P=0.047 by Chi-square test; P=0.032 by Fisher’s exact test). In addition, the allelic frequency of this variant in 1242 individuals in the ESP database is similar to our controls (0.0072: Table 1) but once again significantly less than in our patient cohort (P=0.018 by Chi-square test; P=0.020 by Fisher’s exact test). Alanine 136 is the seventh alanine in a nine-alanine motif in Pbx3 that is highly conserved between species and among different PBC proteins (Figure 1). In silico analysis of the effect of this substitution on protein structure or function gave a high probability that this substitution is deleterious (Table 1).

Figure 1.

The alignments of Pbx proteins in the region of Pbx 3Ala136, which is underlined and bolded.

Expansions of polyalanine repeats have been described as the cause of several diseases, such as holoprosencephaly (in Zic2), hand – foot – genital syndrome (Hox-A13) and synpolydactyly (Hox-D13) [11]. There has only been a single report of a missense variation in a polyalanine tract [12]. This variation in MeCp2 (p.Ala2Val) was identified in a girl with classical Rett syndrome, suggesting this substitution was deleterious to protein function. The function of polyalanine tracts is unclear but has been speculated to have a role in transcription factor repression or to act as a hinge providing the correct conformation for binding DNA or other proteins in the transcription complex [13].

Other novel non-synonymous variants in PBX-related genes were found in our CHD cohort, however none of these variants were predicted to be damaging with high confidence (Table 1). Since few Pbx4 orthologs have been identified in other species and it shares little sequence conservation with Pbx1, 2 or 3, the alignment of protein sequences and analysis using SIFT, Panther and Polyphen2 are of limited utility. However, two of the variants (Pro7Ser and Gly103Glu) were not detected in the control cohort. Interestingly the Pbx4 Pro7Ser was identified in a patient with AVSD and this variant was inherited from her phenotypically normal mother. The proband also harbored the Pbx3 Ala136Val variant, which was inherited from her father who also had no recognizable cardiac malformation. This combination of variants may point to a role for compound mutations in related genes as a cause of CHD similar to the results of mouse studies where multiple pbx alleles were required for development of CHD.

In conclusion, we report several novel non-synonymous PBC pathway variants that are associated with the development of congenital cardiac defects, particularly OFT malformations. In particular the variant in Pbx3 (p.Ala136Val), which is significantly over-represented in the CHD population, may represent an important risk allele for CHD acting as a modifier of important pathways involved in cardiac development. Indeed mice lacking Pbx3 do not have a cardiac phenotype but knocking out one copy of Pbx3 in mice carrying other Pbx mutations resulted in significantly more severe cardiac phenotypes [6]. For example, Pbx1^+/−;Pbx2^+/− mice only developed BAV but after crossing with Pbx3^+/− mice the offspring developed TOF, characterized by right ventricular (RV) outflow tract obstruction, RV hypertrophy, overriding aorta and ventricular septal defects [6]. Because CHD malformations occur in such a large proportion of the population (3–6%) and are likely the manifestation of multiple low-penetrance susceptibilities, the presence of candidate variants in a significant number of ‘control’ individuals does not preclude a CHD-modifier role in patients. Based on these findings and results from studies in mice [5,6], the role of the PBC protein pathway in human cardiac development should be investigated further.

Highlights.

• We screened patients with congenital heart defects for variants in genes encoding lacking Pre-B cell leukemia transcription homeobox (PBX) proteins, as well as interacting proteins.

• One Pbx3 variant, p.A136V, is located in a highly conserved polyalanine tract and predicted to be deleterious.

• This variant was present significantly more frequently in the heart defect patients compared with controls.

• This variant may act as a modifier or risk allele for congenital heart defects.