Article Text
PDF
Abstract
Background SHOX, located in the pseudoautosomal region 1 (PAR1) of the sexual chromosomes, encodes a transcription factor implicated in human growth. Defects in SHOX or its enhancers have been observed in ∼60% of Leri-Weill dyschondrosteosis (LWD) patients, a skeletal dysplasia characterised by short stature and/or the characteristic Madelung deformity, and in 2–5% of idiopathic short stature (ISS). To identify the molecular defect in the remaining genetically undiagnosed LWD and ISS patients, this study screened previously unanalysed PAR1 regions in 124 LWD and 576 ISS probands.
Methods PAR1 screening was undertaken by multiplex ligation dependent probe amplification (MLPA). Copy number alterations were subsequently confirmed and delimited by locus-specific custom-designed MLPA, array comparative genomic hybridisation (CGH) and breakpoint junction PCR/sequencing.
Results A recurrent PAR1 deletion downstream of SHOX spanning 47543 bp with identical breakpoints was identified in 19 LWD (15.3%) and 11 ISS (1.9%) probands, from 30 unrelated families. Eight evolutionarily conserved regions (ECRs 1–8) identified within the deleted sequence were evaluated for SHOX regulatory activity by means of chromosome conformation capture (3C) in chicken embryo limbs and luciferase reporter assays in human U2OS osteosarcoma cells. The 3C assay indicated potential SHOX regulatory activity by ECR1, which was subsequently confirmed to act as a SHOX enhancer, operating in an orientation and position independent manner, in human U2OS cells.
Conclusions This study has identified the first recurrent PAR1 deletion in LWD and ISS, which results in the loss of a previously uncharacterised SHOX enhancer. The loss of this enhancer may decrease SHOX transcription, resulting in LWD or ISS due to SHOX haploinsufficiency.
- Deletion
- enhancer
- SHOX
- LWD
- diabetes
- genetics
- developmental
- endocrinology
- molecular genetics
- genetic screening/counselling
Statistics from Altmetric.com
- Deletion
- enhancer
- SHOX
- LWD
- diabetes
- genetics
- developmental
- endocrinology
- molecular genetics
- genetic screening/counselling
Introduction
SHOX (short stature homeobox-containing gene) is located in the pseudoautosomal region 1 (PAR1) of the X and Y chromosomes.1 ,2 Mutations in SHOX or its enhancers have been reported in Léri-Weill dyschondrosteosis (LWD, MIM 127300), Langer mesomelic dysplasia (LMD, MIM 249700), and idiopathic short stature (ISS, MIM 300582).1 ,3–10 LWD is a skeletal dysplasia associated with disproportionate short stature due to mesomelic shortening of the limbs, and a characteristic abnormality of the forearms known as Madelung deformity: bowing of the radius and dorsal dislocation of the distal ulna. Heterozygous alterations in SHOX or its enhancers have been identified in ∼60% of LWD individuals.11–13 LMD is caused by biallelic functional loss due to homozygous or compound heterozygous SHOX/PAR1 mutations resulting in severely disproportionate short stature with pronounced mesomelic and rhizomelic limb shortening. Heterozygous SHOX/PAR1 alterations have also been found in ∼2–5% of ISS cases,13 a common complex trait defined as a height below −2 SD scores (SDS) in the absence of known specific causative disorders.14
SHOX has highly homologous orthologs in various vertebrate species including chimpanzee, chicken, frog and zebrafish, but it is absent in rodents. Sequence conservation analysis of non-coding regions has been shown to be a good predictor for the identification of cis-regulatory elements, predominantly controlling the spatiotemporal expression of developmental genes.15 ,16 Comparative genomic and functional analyses in human cells,17 chicken limb buds6 ,18 and zebrafish19 demonstrated that six evolutionarily conserved regions (ECRs) within the PAR1 show SHOX cis-regulatory activity, three downstream6 ,17 ,19 and three upstream of SHOX.18 Deletions of the downstream enhancers have been frequently observed in LWD and ISS individuals,11 ,13 while only one ISS individual has been reported to date presenting with a deletion of two SHOX upstream enhancers.10
In this study, we identified a recurrent 47543 bp downstream PAR1 deletion that did not include any of the known SHOX enhancer elements in 30 patients with LWD or ISS. All affected carriers present with identical deletion extension and breakpoints—characteristics of a seemingly recurrent event. Conservation analysis of the deleted sequence followed by chromosome capture conformation (3C) in chicken embryo limbs was performed to investigate whether a SHOX transcription regulatory element lay within the deleted sequence. Evidence for a potential SHOX regulatory element was obtained and further substantiated by experiments demonstrating this putative regulatory sequence could act as an enhancer in human osteosarcoma U2OS cells. The identified recurrent PAR1 deletion results in the loss of this previously unreported enhancer which we propose may decrease SHOX transcription, resulting in LWD or ISS due to SHOX haploinsufficiency.
Patient cohort
The cohort consisted of 124 LWD/possible LWD and 576 ISS probands from different countries. All participants provided informed consent for the performed studies and ethical approval was obtained from the respective participating institutions. The LWD and ISS patient samples were referred from endocrinology and genetic clinics. LWD patients were ascertained using the inclusion criteria of the presence of the Madelung deformity and mesomelic shortening of the limbs in the proband or a direct family member. Stature was recorded and standard deviation scores (SDS) were determined according to the population standards for age and gender. ISS patients with stature < −2 SDS were ascertained using the current consensus criteria. Clinical details were obtained for all patients recruited into the study. Whenever possible, these included birth details, anthropometric measurements, actual height and height SDS according to national standards, physical examination of extremities, and x-rays of the lower arm. Family histories were also documented, including parental heights with calculated height SDS according to national standards.
A cohort of 334 controls, obtained from the Spanish National DNA bank (University of Salamanca, Spain), with heights within the normal range for the Spanish population for age and gender (−2 < SDS < +2) was also studied to estimate the allelic frequency of the novel deletion in the normal population.
Methods
Multiplex ligation dependent probe amplification analysis
Initial multiplex ligation dependent probe amplification (MLPA) analysis was carried out using the commercial SHOX/PAR1 MLPA Kit (Salsa P018 MRC Holland, Amsterdam, The Netherlands). Subsequently, a custom designed downstream MLPA assay was utilised to confirm and delimit the extensions of the deletions in the downstream PAR1 enhancer region (supplementary table 1). The MLPA data were analysed as previously described.9
Array comparative genomic hybridisation
We performed fine-tiling Y-chromosome specific array comparative genomic hybridisation (aCGH) (NimbleGen Systems, Madison, Wisconsin, USA) in two samples according to the manufacturer's protocols, using the service provided by Imagenes, Berlin, Germany. The median distance between probes was 20 bp. Two replica arrays with dye-swap were performed for each analysed sample. Fluorescence intensities and log2 intensity ratios for the test versus control sample were calculated and visualised using the Signalmap software (NimbleGen Systems).
Breakpoint PCR characterisation
Primers were designed in regions shown to be present as two copies by aCGH. The sense and antisense oligos were 5′-GTTGCCCAGGCTTCGGTTTGT-3′ and 5′-CCATGCCCAAGATGTAGACGG-3′ respectively. The PCR reactions were performed in a final volume of 25 μl using the 1× Qiagen Hotstart Taq Buffer and 1 U of Hotstart Taq polymerase (Qiagen, Valencia, California, USA), 0.5 μM of each oligonucleotide, 400 nM of each primer, and ∼50 ng gDNA. The cycling conditions were as follows: initial denaturation at 94°C for 15 min and 38 cycles at 94°C for 30 s, at 50°C for 40 s, and 72°C for 1 min, with a final extension at 72°C for 8 min. The breakpoints were determined by comparing the obtained sequence against the normal PAR1 sequence using Sequencher V4.10 (Gene Codes Corporation, Ann Arbor, Michigan, USA) and Ensembl and Blast.
The deletion mechanism was investigated by computational analysis. Homology between the 2 kb of the breakpoint flanking sequences was assessed using LALIGN. The sequences at the breakpoint junctions and flanks were also analysed for repeat elements using RepeatMasker, DNA secondary structures (MFOLD Web Server), homologous recombination hotspot motifs (PRDM9-binding sites),20 cruciform structures, restriction enzyme cut sites (NEBcutter V.2.0), homologous recombination hotspot motifs, various DNA–protein interaction motifs (Human Protein-DNA Interactome, hPDI), RAG (recombination acting genes) recombinase recognition sites,21 and DNA elements from the ENCODE projects.
Breakpoint assay
A PCR was specifically designed to detect the ∼47.5 kb deletion. The forward oligo (5′-CTCTCCACACGTATCTCCCGAT-3′) was located across the 5′ breakpoint, while the reverse oligo (5′-TTGCTGTAGCGTCGGTGCGTTG-3′) was located in the undeleted 3′ flanking sequence. The 379 bp product was only detected in deletion carriers. The reactions were performed in a final volume of 25 μl using the 1× AmpliTaq Gold Buffer and 1 U of AmpliTaq Gold DNA Polymerase (Applied Biosystems), 1.5 mM of MgCl2, 0.5 μM each oligonucleotide, 400 nM of each primer, and ∼50 ng gDNA. The cycling conditions were as follows: initial denaturation at 94°C for 10 min and 38 cycles at 94°C for 30 s, at 60 °C for 30 s, and 72°C for 45 s, with a final extension at 72°C for 8 min.
Haplotype analysis
Haplotype analysis was undertaken to determine if a common ancestor could explain the presence of the recurrent deletion. SNPs and microsatellite markers located in the 5′ and 3′ flanking regions of the deleted area were analysed as previously described.5 ,11 ,22
Comparative genomic analysis
For the ECR analysis of the deleted sequences, the human PAR1 region was analysed using the human genome (GRCh36/hg18) as a reference in the VISTA genome browser and those genome assemblies from dog, chicken, frog and zebrafish (supplementary figure 1). An e-value of 1e-7 (corresponding to a sequence size of 100 bp with a 70% homology between species) was chosen as a threshold to filter out low homology hits between the sequences.
Chromosome capture conformation assay
The experimental chicken procedures have been performed following the protocols approved by the Ethical Committee for Animal Research from Consejo Superior de Investigaciones (CSIC) according to the European Union regulations. The 3C assays were performed as previously described.23 ,24 Fertilised chick embryos were incubated to stage HH26, when SHOX has been shown to be highly expressed in the limbs.6 Fore- and hindlimbs were subsequently dissected and chromatin was fixed. Nuclei were isolated to generate a HindIII-based re-ligation array. DNA from two bacterial artificial chromosomes (BACs) covering the homologous chicken genomic region CH261-166E5 (chr1: 133.556.801-133.759.625, 2.1) (May 2006 release) and CH261-118H23 (chr1:133.740.437-133.937.986) were digested and re-ligated to serve as random references for the 3C study. The orthologous sequence equivalent to the human deleted region extends 33 kb, ∼160 kb downstream of chicken SHOX (NW_001471545.1). In order to obtain high resolution data on the critical region, a total of eight 3C primers were designed to cover this 33 kb interval (R5-R12, supplementary table 2). Four additional 3C primers (R1–R4) were designed between the critical region and the 3′ end of SHOX. As a positive 3C control, a primer was designed (R13) in close proximity to the orthologous region of previously described SHOX enhancer ECS4/CNE9.17 ,18 Two additional primers (R14, R15) downstream of the positive control allowed us to generate a general perspective of the interaction map of the 240 kb downstream of chicken SHOX. To assay the different interactions we tested the region using primers located along 240 kb downstream of SHOX, using a primer 8 kb upstream of the chicken SHOX promoter as the fixed primer.
Luciferase assay
The four ECR1 enhancer reporter constructs contained the ECR, ECR1 in both orientations and in different localisations, upstream and downstream of the human SHOX promoter (−432 to +5 bp, NM_000451).25 Basically, the sequence including ECR1 was PCR amplified using the following primers: 5′-CCTTGTCCATCTGCGTCTAC -3′ and 5′-CTAACGGCTCACATGAAACTG -3′ with the incorporation of restriction sites at the 5′ end depending on the orientation and position of ECR1 in the final luciferase construct. The enzymes utilised were BamHI and SalI for cloning ECR1 downstream of the luciferase gene while KpnI and NheI were utilised for cloning ECR1 upstream of the SHOX promoter. The restriction sites were switched for the reverse orientations. The PCR products were subsequently subcloned into pCR2.1 vector (Invitrogen, Life Technologies Carlsbad, CA), according to the manufacturer's instructions. The human SHOX promoter was cloned using the following primers 5′- CTCGAGGAAAACTGGAGTTTGCTTTTCCTCCG - 3′ and 5′- AAGCTTTCCATGGCTGGGGCCGGGGCTGGC - 3′ which included the XhoI and HindIII restriction sites, respectively (underlined). The SHOX promoter and ECR1 sequences were subsequently subcloned into the luciferase reporter vector, pGL3basic (Promega, Madison, Wisconsin, USA), by first subcloning the human SHOX promoter upstream of the luciferase gene and second, subcloning the ECR1 sequences in both orientations in the two positions, downstream of the luciferase gene or upstream of the SHOX promoter in the pGL3SHOXprom vector. The previously described SHOX enhancers, CNE4, CNE5 and ECS4/CNE9,6 ,17 were also cloned and employed as positive controls while ECS517 was used as a negative control. All clones were sequence verified.
U2OS cells were seeded in 12-well plates at a density of 150 000 cells/well. Enhancer reporter constructs were cotransfected with expression plasmids using FUGENE6 transfection reagent (Roche, Mannheim, Germany) according to manufacturer's instructions and with a 2:1 FUGENE:DNA ratio. DNA included 2000 ng of pGL3SHOXprom reporter plasmid, containing ECR1 or the previously reported SHOX enhancer elements and 2 ng of pRL-SV40 control plasmid. Empty expression plasmid was used to normalise DNA concentrations. Each plasmid combination was transfected in three wells. Twenty-four hours later, cells were lysed and the reporter activity was measured in triplicate using the Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer's instructions and using the microplate spectrometer Infinite 200 Pro (Tecan, Männedorf, Switzerland). Each sample was normalised, first, with respect to the Renilla luciferase activity and then second, to that transfected with the empty reporter plasmid. Statistical analyses were undertaken with SPSS version 15.0 (IBM).
Results
Identification of a recurrent PAR1 deletion downstream of SHOX
In order to identify the molecular defect in the LWD and ISS cases with no known SHOX/PAR1 defect, a total of 124 clinically diagnosed or suspected LWD and 576 ISS probands of varying nationalities were screened or re-evaluated for PAR1 alterations by MLPA. A deletion of three probes, L05099, L05100, L05101, was observed using the commercial MLPA SHOX P018D1 salsa (figure 1) in a total of 30 unrelated patients, 19 LWD and 11 ISS (supplementary table 3, figure 2). The size and genomic extension of the deletion was further confirmed and delimited by means of a custom-designed SHOX downstream MLPA assay which included a higher number and density of probes in the involved PAR1 region (figure 1). Fine-tiling chromosome Y aCGH performed in probands 27 and 29 suggested that both presented with identical deletion limits (∼47.5 kb, from ∼700 500 to ∼747 900 bp; chr Y, GRCh36/hg18, figure 3A). Subsequent amplification and sequencing across the deletion breakpoints enabled precise breakpoint mapping and confirmed that indeed the two patients had identical deletion breakpoints resulting in a deletion of exactly 47 543 bp (figure 3B). We subsequently confirmed that all 30 affected carriers shared the same breakpoints. The presence of the deletion was also analysed in family members using a specific breakpoint PCR assay by which amplification was only observed if the deletion specific breakpoints were present. The novel recurrent PAR1 deletion was not detected in a cohort of 334 Spanish normal stature individuals. No deletion copy number variant (CNV) of our disease associated interval was identified in this control cohort nor in the database of genomic variants (DGVs) or in the 1000 genomes structural variation project. Furthermore, allelic haplotype analysis of the affected probands excluded the possibility of a common ancestor, indicating that it is a seemingly recurrent event (supplementary figure 3).
Schematic representation of the genomic location and approximate extensions of the PAR1 deletion, characterised using the commercial and custom-designed multiplex ligation dependent probe amplifications (MLPAs). The upper panel shows the extensions of the deletion as detected using the commercial SHOX/PAR1 MLPA (P018). Previously reported SHOX enhancers, CNE4, CNE5 and ECS4/CNE9, are indicated by rectangular black boxes in the upper panel. The lower panel shows further characterisation of the deletion using a custom designed downstream SHOX MLPA. The MLPA probes are shown in supplementary table 1. MLPA probes are indicated by grey boxes. Normal peaks were classified as showing a ratio of 0.65―1.35 while deletions and duplications were classified as having a ratio <0.65 or >1.35, respectively. ▲ indicates the presence of a heterozygous deletion. The approximate coordinates are according to chromosome X and Y, GRCh36/hg18. The diagram is not drawn to scale.
Family pedigrees of the 30 Leri-Weill dyschondrosteosis (LWD)/idiopathic short stature (ISS) probands with the recurrent downstream PAR1 deletion. The probands are indicated by the arrow. Black filled symbols indicate the clinical diagnosis of LWD while grey filled symbols indicate the clinical diagnosis of ISS. The question mark symbol indicates unknown clinical diagnosis. NA indicates that no genetic analysis was undertaken. Del indicates the presence of a heterozygous SHOX deletion while the open triangle indicates the presence of the recurrent downstream PAR1 deletion. The 47543 bp deletion cosegregated with the phenotype in nine families (probands 3, 8, 10, 13, 15, 16, 18, 24, 26) while cosegregation analysis was inconclusive or undeterminable in 21 families due to the presence of short stature in both parents (n=5; families 12, 20, 25, 27, 29), lack of parental DNA or clinical data (n=15; families 1, 2, 4―7, 9, 11, 14, 17, 19, 21―23, 28, 30). In family 24, the child presented with psychomotor delay, hypotonia, pectus excavatum, cleft palate, epicanthus, and hypotelorism and was found by array comparative genomic hybridisation (KaryoArray®) to also have a deletion of 22qter.
PAR1 deletion characterisation. (A) Fine-tiling chromosome Y array comparative genomic hybridisation (aCGH): delimitation of the PAR1 deletion boundaries in case 27 who presented with the deletion of SHOX/PAR1 MLPA probes L05099, L05100 and L05101. We performed fine-tiling Y-chromosome specific aCGH (NimbleGen Systems) in two samples. The median distance between probes was 20 bp. Two replica arrays with dye-swap were performed for each analysed sample. The two dye swap experiments are shown for one proband. Fluorescence log2 intensity ratios of test to reference are shown for each oligonucleotide from telomere to centromere across the short arm of the Y chromosome. The log2 ratio data are shown in the NimbleGen's SignalMap data viewer and each point indicates the midpoint of a 300 bp window average which has been calculated from 1 to 15 probes. The deletions are indicated with an asterisk. (B) Breakpoint characterisation of the downstream PAR1 deletion. Chromatogram of the 47543 bp downstream PAR1 deletion is represented with a vertical bar indicating the breakpoints. PAR1 sequence coordinates are taken from the X/Y chromosome assembly (GRCh36/hg18) which corresponds to 780549:828093 for the X chromosome and to 730549:778093 for the Y chromosome assemblies (GRCh37/hg19). While no homologous repeat was observed in the 5′ breakpoint, a LINE1 repetitive element was present in the 3′ breakpoint. No microhomology was present in either breakpoint, but instead a precise ‘blunt-end’ fusion was noted.
Chromatin conformation capture assays suggest an interaction between ECR1 and the SHOX promoter
Analysis of the deleted sequence interval showed that none of the three known SHOX downstream enhancers (CNE4, CNE5, ECS4/CNE9) lay within this region (figure 1). As comparative analysis of genomes has proven to be a successful tool for the identification of regulatory elements,16 we analysed the region encompassing the SHOX open reading frame and the ∼47.5 kb deleted sequence. This ∼0.5 Mb region is highly conserved from human to fish (supplementary table 4). This suggests that the non-coding region harbouring the ∼47.5 kb deletion may contain evolutionary conserved cis-regulatory regions governing SHOX expression. A total of eight ECRs (ECR1―8) lay within the deleted region (supplementary figure 1). Two of these ECRs, ECR1 and ECR7, correspond to CNE7 and CNE8, respectively (supplementary figure 2), previously shown to not contain enhancer activity in in ovo electroporation of chicken limb embryos.6 The eight ECRs were used as candidate target regions for subsequent analyses.
The high degree of conservation and synteny along the critical region allowed us to use chicken embryos to determine the chromatin structure in SHOX-expressing cells using chromatin conformation capture assays (3C).23 Thus, fore- and hindlimbs from chicken embryos at HH26 developmental stage, when SHOX has been shown to be highly expressed in the limbs,6 were used as a template for 3C assays. The results from the 3C assays revealed two robustly interacting regions, one corresponding to the positive control, the fragment mapping close to the previously reported SHOX enhancer ECS4/CNE9.6 ,17 Interestingly, the second interaction peak, located between R5 and R6, mapped within the critical deleted region. This fragment contained the chicken orthologous region corresponding to the human ECR1 (figure 4). The other fragments around ECR's 2―8 did not show any interactions in the 3C assays.
Comparative genomic analysis and 3C analysis of the SHOX downstream regulatory landscape. The upper panel represents the comparative genomic analysis of the deleted PAR1 region taking the Chicken (GalGal3) genome as a reference and plotted in the VISTA genome browser. The orthologous sequence equivalent to the human deleted region extends 33 kb, ∼160 kb downstream of chicken SHOX (NW_001471545.1). Coloured peaks indicate the degree of conservation between the reference and the genome of interest. Blue peaks correspond to coding and untranslated regions, while red peaks represent non-coding regions. The lower panel represents the 3C data, the physical interaction between the anchor primer mapping 8 kb upstream of the chicken SHOX promoter and the 15 different fragments distributed downstream of SHOX. In order to obtain high resolution data on the critical region, a total of eight 3C primers were designed to cover this 33 kb interval (R5–R12, supplementary table 2). Four additional 3C primers (R1–R4) were designed between the critical region and the 3′ end of SHOX. As a positive 3C control, a primer was designed (R13) in close proximity to the orthologous region of previously described SHOX enhancer ECS4/CNE9.17 ,18 Two additional primers (R14, R15) downstream of the positive control allowed us to generate a general perspective of the interaction map of the 240 kb downstream of chicken SHOX. The Y axis represents the crosslinking frequency expressed as the ratio of PCR performed on 3C samples relative to bacterial artificial chromosome (BAC) controls. The hashed region (161.5–190.2 kb: 33 kb), including a total of eight fragments, corresponds to the chicken orthologous region of the deleted region present in Leri-Weill dyschondrosteosis (LWD) and idiopathic short stature (ISS) patients. All data points were generated from an average of three independent experiments. The standard error of the means is indicated. The results demonstrate increased interaction frequency associated to the region harbouring an already described SHOX enhancer taken as positive control (ECS4/CNE9,6 ,17 R13; 213.5 kb) and an additional region mapping to the critical region (between R5 and R6) which contained ECR1 (grey highlighted area).
Luciferase reporter assays in human U2OS confirm that ECR1 has enhancer activity
To confirm the potential enhancer activity of this putative regulatory element, we analysed the regulatory activity of ECR1 in human osteosarcoma cells, U2OS. For this purpose, the ECR1 was cloned in both orientations and in two localisations, downstream of the luciferase gene or upstream of the SHOX promoter in the pGL3SHOXprom vector. The clones were then co-transfected with a renilla luciferase control vector into U2OS cells. SHOX enhancer regions CNE4, CNE5 and ECS4/CNE96 ,17 were utilised as positive controls, while ECS517 was employed as a negative control. Increased luciferase activity was observed for ECR1, in both orientations and positions, and all three positive controls (figure 5), demonstrating that ECR1 acts as a SHOX enhancer in an orientation and position independent enhancer in U20S cells.
Regulatory activity for the analysed evolutionarily conserved regions (ECRs) using the luciferase reporter assay. Luciferase reporter activity of U2OS cells transfected with the pGL3 enhancer reporter plasmids and the renilla luciferase control plasmid. Empty reporter plasmid was used to normalise luciferase activity. Fold increase values were obtained by normalising the relative luciferase units of each sample, first, with respect to the Renilla luciferase activity, and second, to that transfected with the empty reporter plasmid. All values represent the mean and SD of three independent samples, with each sample assayed in triplicate. The SHOX enhancer regions, CNE4, CNE5 and ECS4/CNE9,6 ,17 were utilised as positive controls. ECS5, localised outside of the deleted region and which has previously been shown to not have enhancer activity,17 was utilised as a negative control. Increased luciferase activity was observed for all ECR1 constructs and all positive controls. ECR1 showed enhancer activity in both orientations and in both localisations in the pGL3 vector, downstream of the luciferase reporter gene and upstream of the SHOX promoter. The clone containing ECR1 inverted and localised upstream of the promoter was weaker than the other ECR1 clones. This may be due to the enhancer element within the cloned sequence being extremely close to the SHOX promoter in this orientation, and thus unable to form the tertiary structures between the enhancer and the promoter. In summary, the luciferase experiments demonstrate that ECR1 acts as an orientation and position independent enhancer in U20S cells.
Discussion
A recurrent ∼47.5 kb PAR1 deletion, sharing the same deletion extensions and breakpoints, was identified in a total of 30 LWD or ISS probands of varying nationalities and origins. Allelic haplotype analysis of the affected probands excluded the possibility of a common ancestor. This CNV was not observed in 334 Spanish normal stature individuals and not reported in the DGV Database or in the 1000 genome project, thus suggesting that the deletion is pathogenic.
Deletion mechanism
We investigated how this recurrent deletion may have arisen by examining the flanking sequences at the 5′and 3′ breakpoints and evaluating the characteristics of the breakpoint junction to potentially surmise the rearrangement mechanism resulting in this recombinant deletion CNV product. While no homologous repeat was observed in the 5′ breakpoint, a LINE1 repetitive element was present in the 3′ breakpoint. No microhomology was present in either breakpoint, but instead a precise ‘blunt-end’ fusion was noted (figure 3B), suggesting that non-homologous end joining could be the underlying deletion mechanism. However, this is very unlikely since all of the known rearrangements with ‘blunt end’ breakpoints are usually non-recurrent. To date, blunt ended deletions account for ∼3.4% to 5% of all deletions.26 ,27 Therefore, we can only speculate on the existence in this region of a potential unknown element—for example, a retroviral element/retrosposon such as a LINE1—that predisposes to a recurrent deletion potentially mediated through a precise excision mechanism. Little is known about the extent of variation of individual repetitive sequence positions within personal enmesh, although recent studies suggest this may be far greater than anticipated by a ‘single haploid human genome reference sequence.28–32
Functional characterisation
Analysis of the deleted sequence interval showed that none of the three known SHOX downstream enhancers lay within this region. As comparative analysis of genomes has proven to be a successful tool for the identification of regulatory elements16 we analysed the sequence within the ∼47.5 kb deletion for ECRs. Eight ECRs (ECR1–8) were identified within the deleted region. Interestingly, ECR1 and ECR6 were conserved from human to teleosts, while ECR5 was detected up to chicken. The remaining ECRs were conserved between human and dog, suggesting a more recent appearance during evolution. By means of a 3C assay, a PCR based method that permits the determination of physical in vivo interaction between any chromatin segments in chicken embryo limbs—we showed that the genomic region harbouring ECR1, one of the eight ECRs identified in the region, interacts with a sequence close to or within the SHOX proximal promoter. Luciferase reporter assays in human U2OS cells confirmed that ECR1 operates as an enhancer.
A total of seven enhancers of SHOX transcription, including the novel one described herein, are known to date, four downstream and three upstream of SHOX. The loss of these enhancers has been postulated to result in SHOX haploinsufficiency due to decreased SHOX transcription. Other, as yet unidentified, enhancers may also exist. Indeed, Sabherwal et al6 failed to show any enhancer activity for ECR1 (referred to as CNE7 in their paper) using ovo-electroporation of the chicken embryo limb bud using a green fluorescent protein reporter construct driven by the β-globin promoter. Many forms of transgenesis, including ovo-electroporation, result in mosaic expression, thus only permitting the detection of strong reporter activity. Thus, weaker enhancer expression would not be detected using this in vivo enhancer assay. In contrast, the 3C assay can detect both weak and strong regulatory elements. The luciferase reporter assays confirmed that ECR1, CNE4 and ECS4/CNE9 operate as enhancers. In addition, the 3C assay demonstrated for the first time that both enhancers, ECR1 and ECS4/CNE9, interact with the SHOX promoter. Thus, various animal models and the combination of different enhancer analytical techniques may be required to dissect the complex regulation of SHOX transcription by cis-acting PAR1 elements.
Clinical and pathogenicity evaluation
Further studies were undertaken to estimate the pathogenicity of this novel, recurrent PAR1 alteration. First, cosegregation analysis and further examination of the clinical phenotype of the probands and the family deletion carriers showed that the deletion cosegregated with the phenotype in the families of nine probands (seven LWD, two ISS), while in the remaining cases the analysis was inconclusive or undeterminable (figure 2). In four of these LWD families, all deletion carriers presented with the Madelung deformity and/or mesomelic shortening, but only 4/9 deletion carriers presented with short stature (figure 2). An example of this is shown in family 15 where the mother and her two children, a son (aged 6–7 years) and an older daughter (aged 9–10 years)—all three deletion carriers—presented with mesomelic shortening, yet only the male proband presented with short stature. In the remaining three LWD families, the deletion cosegregated with either mesomelic shortening of the limbs, Madelung deformity or short stature. In the two ISS families, all the deletion carriers presented with short stature.
Subsequently, we analysed the heights of the LWD (n=19) and ISS (n=11) deletion carrier probands. The mean height SDS for the LWD and ISS probands were −2.03 and −2.60, respectively, while the median height SDS were −2.16 (range −0.14 to −4.68) and −2.55 (range −2.1 to −3.26), respectively. The height deficit was greater in the probands with ISS than those with LWD. This difference was not unexpected as the LWD probands were referred because of the presence of mesomelic shortening and Madelung deformity, regardless of their stature, while the ISS probands had to comply with the criteria of stature below −2 SDS for sex, age and ethnic origin. In order to avoid the bias introduced by clinical selection of the probands, we subsequently looked at these characteristics in deletion carrier family members (n=18) who had not been clinically selected. The mean height SDS of LWD and ISS family members were −1.50 and −2.56 while the median height SDS were −1.15 (range −0.03 to −4.10) to −2.85 (range −0.83 to −3.32), respectively. Thus, the height SDS ranges were not so different between LWD and ISS family members. We have not been yet able to assess how this deletion affects the target adult heights. Though some parents had heights within the normal range we do not have the data to estimate whether they reached their expected target heights. This observed clinical variability is not unexpected, as it is well documented that SHOX haploinsufficiency due to SHOX deletion/mutations exerts a widely variable effect on the patient's height (−5.6 to 0.3 SDS) and on the presence or absence of limb mesomelic shortening and the Madelung deformity.9 ,33 ,34
Three families are of particular interest for further discussion. The first family (proband 1) consisted of two clinically diagnosed LWD siblings, both of whom are homozygous for the ∼47.5 kb deletion. The deletion was present in the heterozygous state in the two parents for whom molecular analysis excluded the possibility of unknown consanguinity. Though all four individuals had normal stature and both parents presented with no dysplasic signs, the two children presented with the Madelung deformity, thus suggesting reduced penetrance. In the second family, LWD proband 2 was a compound heterozygote for two different PAR1 deletions, the novel ∼47.5 kb PAR1 deletion on the maternal allele and a SHOX encompassing deletion on the paternal allele. While the father had short stature with mesomelic shortening of the limbs, and the mother was of normal height with no dysplasic features, the adult proband presented with short stature, mesomelic limb shortening and the Madelung deformity. In this family, the ∼47.5 kb deletion appears to be exerting little effect on the proband's phenotype and seems to be non-penetrant in the proband's mother. In the third family the genetic situation is similar to that of the previous case but the phenotypic outcome is different: the proband inherited the ∼47.5 kb PAR1 deletion from the mother with proportionate short stature, and a paternally derived de novo SHOX deletion. Although the proband had a predicted adult height of 164±5 cm (−2.08 SDS), he failed to reach his target height (final height 155 cm, −3.48 SDS) and he presented with severe dysplasic features, which were to a similar degree as those observed in LMD. These dysplasic features cannot be explained solely by the de novo SHOX deletion; thus, in this case both PAR1 deletions seem to be contributing to the proband's phenotype, an intermediate between LWD and LMD.
All together, cosegregation data and the clinical features of these three particular families suggest that although the recurrent ∼47.5 kb PAR1 deletion associates with LWD and/or short stature, it may have a reduced penetrance as compared to that observed in individuals carrying SHOX encompassing mutations or larger downstream PAR1 deletions, which may affect multiple enhancers. Analyses of further generations and families will be essential to clarify this point.
The deletion appears to be not only recurrent in our cohort35 but also in the Dutch population, as we and Losekoot et al36 reported recently in the 2012 European Society of Pediatric Endocrinology meeting. In addition, Nicoletti et al identified the identical deletion in one of their 25 studied Italian LWD patients.37
Interestingly, this PAR1 deletion was not detected by any genome-wide association study (GWAS) signal from the large studies systematically evaluating for genomic regions contributing to height.38–41 This may be the consequence of different coincidental facts:1) array designs often provide ‘poor coverage’ of the pseudoautosomal regions; 2) identifying small (<100 kb) CNV by SNP chips is a challenging task; 3) the recurrent nature of our detected deletion suggests multiple de novo events, yet still unknown, or 4) simply the absence of such events in the population under study by GWAS. Nevertheless, our findings clearly implicate non-coding CNV as potentially contributory factors to the population variation in height. This is of particular interest given the sheer number of GWAS signals mapping to non-coding regions (GWAS Catalogue). Nevertheless, our findings clearly implicate non-coding CNV as potentially contributory factors to the population variation in height. This is of particular interest given the sheer number of GWAS signals mapping to non-coding regions (GWAS Catalogue). Therefore, our findings add to an accumulating body of evidence suggesting that Mendelian disease and common disease genetics may share common underlying ‘genomic/genetic architecture’.42
In summary, this work reports the identification of the first recurrent PAR1 deletion, downstream of SHOX, observed in LWD or ISS individuals. Functional analysis demonstrated that the deleted PAR1 sequence included a novel downstream SHOX enhancer, whose deletion could result in SHOX haploinsufficiency due to impaired SHOX transcription in the LWD and ISS patients presenting with this novel deletion.
Web resources
Online Mendelian Inheritance in Man (OMIM), http://www.omim.org
Ensembl Genome Browser, http://www.ensembl.org/index.html)
RepeatMasker, http://www.repeatmasker.org
MFOLD Web Server, http://mfold.rna.albany.edu/?q=mfold/DNA-Folding-Form
NEBcutter V2.0, http://www.tools.neb.com/NEBcutter2/
Human Protein-DNA Interactome (hPDI), http://bioinfo.wilmer.jhu.edu/PDI/
Database of genomic variants (DGV), http://projects.tcag.ca/variation/
1000 genomes structural variation project, http://www.1000genomes.org/
GWAS catalogue, http://www.genome.gov/gwastudies/
Acknowledgments
We thank Noelia Sanchez for her technical help and all the clinicians and patients who participated in the study.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data Supplement 1 - Online figures and tables
Footnotes
Funding This work was supported by the: Ministerio de Innovación y Ciencia (MICINN SAF2009-08230 to SB-S and KEH; BFU2010-14839 and CSD2007-00008 to JLG-S), Fondo de Investigación Sanitaria (FIS PI08/90270 to SB-S and KEH) and the Andalusian Government (CVI-3488 to JLG-S). Postdoctoral fellowships from CIBERER (SB-S) and Spanish National Research Council (CSIC) (JLR).
Competing interests None.
Patient consent Obtained.
Ethics approval Ethics approval was provide by Hospital Universitario La Paz.
Provenance and peer review Not commissioned; externally peer reviewed.