Biallelic TMEM260 variants cause truncus arteriosus, with or without renal defects

Only two families have been reported with biallelic TMEM260 variants segregating with structural heart defects and renal anomalies syndrome (SHDRA). With a combination of genome, exome sequencing and RNA studies, we identified eight individuals from five families with biallelic TMEM260 variants. Variants included one multi‐exon deletion, four nonsense/frameshifts, two splicing changes and one missense change. Together with the published cases, analysis of clinical data revealed ventricular septal defects (12/12), mostly secondary to truncus arteriosus (10/12), elevated creatinine levels (6/12), horse‐shoe kidneys (1/12) and renal cysts (1/12) in patients. Three pregnancies were terminated on detection of severe congenital anomalies. Six patients died between the ages of 6 weeks and 5 years. Using a range of stringencies, carrier frequency for SHDRA was estimated at 0.0007–0.007 across ancestries. In conclusion, this study confirms the genetic basis of SHDRA, expands its known mutational spectrum and clarifies its clinical features. We demonstrate that SHDRA is a severe condition associated with substantial mortality in early childhood and characterised by congenital cardiac malformations with a variable renal phenotype.

severe condition associated with substantial mortality in early childhood and characterised by congenital cardiac malformations with a variable renal phenotype.

K E Y W O R D S
exome sequencing, genome sequencing, kidney, phenotypic variability, renal failure, SHDRA, structural heart defects and renal anomalies syndrome, TMEM260, truncus arteriosus 1 | INTRODUCTION TMEM260 is a 79.5 kDa protein with eight transmembrane spans (www.uniprot.org/uniprot/Q9NX78) located mainly in the nucleoplasm and within focal adhesion sites (www.proteinatlas.org). 1 TMEM260 encodes at least four protein-coding transcripts. Of these, two (ENST00000261556. 11 and ENST00000538838.5) are considered to be the main transcripts. They differ in the utilisation of an internal exon as well as the final three exons, which in the short isoform are non-coding.
Five individuals from two families with biallelic truncating TMEM260 variants and brain, cardiac, renal, and digit abnormalities were reported in 2017. 2 The condition is now listed on OMIM as "structural heart defects and renal anomalies syndrome" (SHDRA; MIM #617478). Notably, the variants in both families mapped to the long isoform, raising the possibility of SHDRA being an isoform-specific disorder. Since the original publication, there have been no further reports in the literature. Knowledge about the variant and the clinical spectrum of this condition is therefore limited (Supplementary background). In this study, we describe eight affected individuals from five families, confirming that biallelic TMEM260 loss of function variants cause SHDRA and helping to define its clinical spectrum.

| Compound-heterozygous TMEM260 variants in foetuses with congenital heart anomalies
In F1-II-3 ( Figure 1A), type I truncus arteriosus (TA) with pulmonary stenosis and ventricular septal defect (VSD) were detected on antenatal anomaly scan at 20 weeks gestation ( Table 1). The pregnancy was terminated at 21 weeks. Post-mortem examination confirmed the cardiac anomalies ( Figure S1) and did not reveal any other abnormalities.
In F1-II-4 a large peri-membranous outlet VSD, type I TA with small pulmonary trunk and small pulmonary artery branches were detected antenatally and the pregnancy was terminated at 24 weeks gestation.
Post-mortem examination confirmed the cardiac anomalies and revealed a horseshoe kidney. The placenta was also abnormal with a two-vessel cord and omphalomesenteric duct remnant.
Trio WGS was performed as part of the 100KGP on F1-II-4 and both parents. Although the initial analysis focussing on several panels from PanelApp was negative, a scan for Mendelian inconsistencies highlighted an apparently homozygous NM_017799.4:c.344G > A:p.
Arg115Lys, it involves the last base of exon 3 and results in a drop in the predicted splicing efficiency (MaxEntScan: 9.65 ! 2.69). We, therefore, performed RT-PCR on peripheral blood sample from F1-I-1, which showed the presence of two bands, with only the larger band seen in controls ( Figure 1C). Sanger sequencing confirmed exon 3 skipping ( Figure 1D, Figure S2), resulting in a frameshift of exon 4 (p. Val65AlafsTer32). Similarly, the expected exon 1-4 junction was detected by RT-PCR in the maternal sample ( Figure S3), resulting in p.
(Glu55PhefsTer20). Collectively, the genetic studies, RNA analysis and similarity of the foetuses' phenotypes with features reported previously, 2 strongly suggest these TMEM260 variants are diseasecausing.

| Identification of additional SHDRA patients
To expand the cohort of patients with SHDRA we interrogated the 100KGP database further and identified a homozygous c.1410C > G: p.Tyr470Ter TMEM260 variant in Family 2 ( Figure 1E). F2-II-2 exhibited a common arterial trunk, tricuspid atresia, VSD, partial anomalous pulmonary venous connection, bilateral hearing loss, global developmental delay, protein losing enteropathy and deteriorating renal function from the age of 15 months ( Figure S4). Multi-organ T A B L E 1 Genetic and clinical information for five newly reported families with rare biallelic variants in TMEM260. cDNA and protein coordinates are based on the longer canonical isoform (NM_017799.4)   failure following cardiac surgery led to death at age 5 (Table S1, Supplementary Case Histories).
Through international collaboration, we uncovered three further families ascertained via WES ( Figure 1E

| DISCUSSION
We present eight individuals, from five independent families, with biallelic TMEM260 variants ( Figure 1). In combination with clinical data published previously, 2 our results suggest congenital cardiac malformations to be the most consistent phenotype of SHDRA. All 12 patients are reported to have VSD and 10/12 had TA (Table 1). In most of these patients, VSD is likely to be secondary to TA. Notably, TA is one of the rarest congenital cardiac anomalies with few known genetic associations in NKX2-5, 4 NKX2-6, 5 GATA6 6 as well as TBX1. 7 Interestingly, TMEM260, is predicted to be one of 1442 target genes for GATA6 predicted using known transcription factor binding site motifs from the TRANSFAC dataset. 8 The JASPAR database of transcription factor binding sites predicts a GATA6 binding site within intron 5 of TMEM260 although the functionality of this motif is unknown. 9 Our results show that the renal phenotype of SHDRA is highly variable. Horseshoe kidney and cysts were noted in one patient each.  11,12 The association of cardiac, cerebral and renal malformations is also reminiscent of ciliopathies, although generally the cardiac features linked to these group of disorders do not include TA.
Antenatal detection of severe congenital malformations led to termination of pregnancy in three cases described here. Out of nine live born pregnancies, six patients died within the age ranges of 6 weeks to 5 years. One of the two individuals whom survived to 5-years old (F2-II-2) had developmental delay and hearing loss. However, due to insufficient numbers it is difficult to confidently associate these features with SHDRA. We note that two other individuals in the present study who survived beyond their first year, were cognitively normal. Although facial dysmorphism was reported in 1/4 of the original cohort, that feature was not replicated here.
This study substantially expands the known mutation spectrum of SHDRA. Including the patients presented here, a total of eight different TMEM260 variants in 12 individuals from seven families have now been identified ( Figure 1F). Of these, two variants are stop-gains, two are frameshifts, one is a multi-exon deletion, two disrupt splicing and one is missense. All variants are supported by in silico tools, including CADD scores, which are 28.3-41 ( Table 1). The distribution of the variants confirms that variants affecting only the longer isoform are sufficient to cause SHDRA.
We show that the carrier frequency for SHDRA could be up to 1 in 140 in certain populations ( Figure S8). This analysis also identified a potential founder variant in the African/African-American population that requires further functional studies to validate its "Likely Pathogenic" status in ClinVar. The c.1698_1701del seen in Family 4 and in an Arabic family described previously 2 may also represent a founder mutation.
In conclusion, our description of five families with biallelic TMEM260 variants confirms the genetic basis of SHDRA and helps delineate the mutational/phenotypic spectrum of the condition. The strong association with TA has important implications for genetic counselling, prenatal diagnostics as well as postnatal targeted genetic testing.