Background Structural variants (SVs) affecting non-coding cis-regulatory elements are a common cause of congenital limb malformation. Yet, the functional interpretation of these non-coding variants remains challenging. The human Liebenberg syndrome is characterised by a partial transformation of the arms into legs and has been shown to be caused by SVs at the PITX1 locus leading to its misregulation in the forelimb by its native enhancer element Pen. This study aims to elucidate the genetic cause of an unsolved family with a mild form of Liebenberg syndrome and investigate the role of promoters in long-range gene regulation.
Methods Here, we identify SVs by whole genome sequencing (WGS) and use CRISPR-Cas9 genome editing in transgenic mice to assign pathogenicity to the SVs.
Results In this study, we used WGS in a family with three mildly affected individuals with Liebenberg syndrome and identified the smallest deletion described so far including the first non-coding exon of H2AFY. To functionally characterise the variant, we re-engineered the 8.5 kb deletion using CRISPR-Cas9 technology in the mouse and showed that the promoter of the housekeeping gene H2afy insulates the Pen enhancer from Pitx1 in forelimbs; its loss leads to misexpression of Pitx1 by the pan-limb activity of the Pen enhancer causing Liebenberg syndrome.
Conclusion Our data indicate that housekeeping promoters may titrate promiscuous enhancer activity to ensure normal morphogenesis. The deletion of the H2AFY promoter as a cause of Liebenberg syndrome highlights this new mutational mechanism and its role in congenital disease.
- non coding mutations
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Structural variants (SVs) are a common cause of human congenital malformation syndromes and recent studies show that SVs can be pathogenic affecting protein coding genes and reshuffling the complex non-coding cis-regulatory architecture of the genome.1 One particularly interesting example is a rare homeotic transformation syndrome first described by Liebenberg in 1973.2 The Liebenberg syndrome is characterised by a partial homeotic transformation of the arms into legs with the following clinical features: (1) the distal humerus is fused with an ectopic patella that articulates with a dysplastic elbow joint resembling a primitive knee joint; (2) the distal radius and ulna articulate with an elongated wrist that contains partially fused pisiform, triquetrum, lunate and scaphoid bones similar to the talus and calcaneus of the ankle; (3) variable severity of brachydactyly and (4) forearms have reduced flexion and are fixed at a 30° supination with an inability to fully extend the limb. Liebenberg syndrome is inherited in an autosomal-dominant manner with complete penetrance. Several SVs in the regulatory landscape upstream of PITX1 on chromosome 5 have been reported as causative mechanisms including deletions, a translocation, and a duplication . So far, three overlapping deletions have been described ranging from 107 kb to 275 kb with a minimal critical region of 107 kb that includes the housekeeping gene H2AFY.3-5 Knockout mouse models of H2afy do not display any limb phenotype and thus it has been suggested that rather the misregulation of PITX1 is the pathogenic driver of the limb malformation.3 6 PITX1 is a transcription factor expressed solely in the hindlimb during limb development where it ensures proper outgrowth and patterning of this tissue.7 8 Knockout of Pitx1 eliminates hindlimb characteristics including a loss of the patella and reduction of the calcaneus.7 8 On the other hand, ectopic expression of PITX1 in chick and mouse developing forelimbs results in transformation of skeletal structures, muscles and tendons into a leg-like morphology.9
We recently demonstrated that Liebenberg syndrome is caused by SVs in the Pitx1 regulatory landscape allowing ectopic chromatin interaction between Pitx1 and its enhancer element Pen with consequent misexpression in forelimb tissue.3 10 This novel type of pathomechanism whereby a gene is misregulated by its native enhancer is termed regulatory endoactivation, distinct from enhancer adaption where SVs pair genes with foreign enhancers.10
Here, we report on the hereto smallest deletion identified in Liebenberg syndrome. Three affected individuals in a single family with a mild form of this condition turned out to be negative for the known SVs at the PITX1 locus. By whole genome sequencing (WGS), we identified a 8.5 kb deletion that removes the first non-coding exon of H2AFY. CRISPR-Cas9 engineering of the deletion in mice resulted in Pitx1 misexpression in the forelimb, thereby confirming its pathogenic role in the disease.
Materials and methods
Subjects and ethics approval
The study was performed with the approval of the Charité Ethics Committee, Berlin, Germany. Patients were enrolled with written informed consent for participation in the study. The clinical evaluation included medical history interviews, a physical examination and review of medical records. Blood samples or buccal swaps were obtained from each participating individual, and DNA was extracted by standard procedures.
CRISPR-Cas9 engineered allelic series
Mouse mutants with larger deletions or inversions were created using the adapted CRISPR-Cas9 method CRISVar.11 In brief, sgRNA targeting regions were designed using https://benchling.com/ and http://crispr.mit.edu:8079/ and were selected in order to bear a minimal off target score (table 1). The sgRNA was cloned into the px459 vector (Addgene). The cell culture was done according to standard procedure:10 12 Day 1) MEF CD1 feeder cells seeded onto a 6-well plate. Day 2) 400.000 G4 ES cells were seeded for each transfection. Day 3) Two hours before transfection, the embryonic stem cell (ESC) medium without pen/strep was added. For transfection, a DNA mix consisting of 8 µg of each pX459-sgRNA vector was combined with 125 µl Optimem, and a transfection mix consisting of 25 µl FuGene HD (Promega) and 100 µl OptiMEM (Gibco) was combined and incubated at RT for 15 min before being added drop-wise onto the cells. Day 4) Three 6 cm dishes of DR4-puromycin resistant feeders were seeded for each transfection. Day 5) Targeted G4 cells were split onto three DR4 6 cm dishes and a 48 hours selection was initiated by adding puromycin to the ESC medium (final concentration 2 µg/mL). Day 7) Selection was stopped and recovery initiated by using standard ESC medium. The recovery period was of around 4 days. Day 11) Individual clones (ca. 300 per construct) where picked from the plate and transferred into 96-well plates with CD1 feeders. After 3 days of culture, plates were split in triplicates, two for freezing and one for growth and DNA harvesting. Genotyping was performed by PCR and qPCR analyses.
ES and feeder cells were tested for mycoplasma contamination using Mycoalert detection kit (Lonza) and Mycoalert assay control set (Lonza).
Aggregation of mESC
Embryos and live animals were generated from ESCs by diploid or tetraploid complementation, after thawing a frozen ESC phial seeded on CD1 feeders and grown for 2 days (Artus and Hadjantonakis 2011). Female mice of CD-1 strain were used as foster mothers. Several mouse lines were maintained by crossing them with C57BL6/J mice.
All animal procedures were in accordance with institutional, state and government regulations (Berlin:LAGeSo G0247/13).
RNA isolation and qRT-PCR
To quantify mRNA levels in wild type (WT) and mutant mice, E11.5 forelimb or hindlimb buds were microdissected in cold phosphate-buffered saline and immediately snap frozen and stored at −80°C. To isolate RNA, 500 µl TRIzol was added to the tissue and homogenised using a pestle. 100 µl chloroform was then added, and the samples were pulse vortexed for 15 s and centrifuged at 4°C at 10 000 RPM for 15 min. The supernatant was transferred to a new tube and mixed with 1 vol of 70% EtOH, and loaded onto a RNeasy Mini Kit (QIAGEN) column. The rest of the RNA isolation was done according to manufacturer instruction. cDNA was generated using the Superscript III First-Strand Synthesis System (Thermo Fisher Scientific) whereby 300 ng of RNA was reverse transcribed using random hexamer primers. To quantify the relative abundance of transcripts qRT-PCR analyses of 3–8 biological replicates in technical triplicates was done using the GoTaq qPCR Master Mix (Promega).
Whole-mount in situ hybridisation
The Pitx1 mRNA expression in E11.5 mouse embryos was assessed by whole mount in situ hybridisation (WISH) using a digoxigenin-labelled Pitx1 antisense riboprobe transcribed from a cloned Pitx1 probe (PCR DIG Probe Synthesis Kit, Roche). Whole embryos were fixed overnight in 4% paraformaldehyde /PBS. The embryos were washed in PBST (0.1% Tween) and dehydrated stepwise in 25%, 50% and 75% methanol/PBST and finally stored at −20°C in 100% methanol. The WISH protocol was as follows: Day 1) Embryos were rehydrated on ice in reverse methanol/PBST steps, washed in PBST, bleached in 6% H2O2/PBST for 1 hour and washed in PBST. Embryos were then treated in 10 µg/mL Proteinase K/PBST for 3 min, incubated in glycine/PBST, washed in PBST and finally refixed for 20 min with 4% PFA/PBS, 0.2% glutaraldehyde and 0.1% Tween 20. After further washing steps with PBST, embryos were incubated at 68°C in L1 buffer (50% deionised formamide, 5× saline-sodium citrate , 1% sodium dodecyl sulfate , 0.1% Tween 20 in diethyl pyrocarbonate; pH 4.5) for 10 min. Next, embryos were incubated for 2 hours at 68°C in hybridisation buffer 1 (L1 with 0.1% tRNA and 0.05% heparin). Afterwards, embryos were incubated o.n. at 68°C in hybridisation buffer 2 (hybridisation buffer 1 with 0.1% tRNA and 0.05% heparin and 1:500 DIG-Pitx1 probe). Day 2) Removal of unbound probe was done through a series of washing steps 3×30 min each at 68°C: L1, L2 (50% deionised formamide, 2× SSC pH 4.5, 0.1% Tween 20 in DEPC; pH 4.5) and L3 (2× SSC pH 4.5, 0.1% Tween 20 in DEPC; pH 4.5). Subsequently, embryos were treated for 1 hour with RNase solution (0.1 M NaCl, 0.01 M Tris pH 7.5, 0.2% Tween 20, 100 µg/mL RNase A in H2O), followed by washing in TBST 1 (140 mM NaCl, 2.7 mM KCl, 25 mM Tris-HCl, 1% Tween 20; pH 7.5). Next, embryos were blocked for 2 hours at RT in blocking solution (TBST 1 with 2% calf-serum and 0.2% BSA), followed by incubation at 4°C o.n. in blocking solution containing 1:5000 Anti-Digoxigenin-AP. Day 3) Removal of unbound antibody was done through a series of washing steps 8×30 min at RT with TBST 2 (TBST with 0.1% Tween 20, and 0.05% levamisole/tetramisole) and left o.n. at 4°C. Day 4) Staining of the embryos was initiated by washing at RT with alkaline phosphatase buffer (0.02 M NaCl, 0.05 M MgCl2, 0.1% Tween 20, 0.1 M Tris-HCl and 0.05% levamisole/tetramisole in H2O) 3×20 min, followed by staining with BM Purple AP Substrate (Roche). Limb buds from at least three embryos were analysed from each mutant genotype. The stained limb buds were imaged using Zeiss Discovery V.12 microscope and Leica DFC420 digital camera.
The double delta Ct method was used to calculate fold change between WT and mutant samples. Statistical significance was calculated using type two one-tailed Student’s t-test.
Loss of H2AFY promoter region identified by WGS
Following X-ray investigation of upper limbs of a non-consanguineous Italian family, a woman and her two sons were diagnosed with Liebenberg syndrome (figure 1A).13 All three individuals display typical though mild arm-to-leg transformation affecting the lower arms: brachydactyly, camptodactyly of the fifth finger, fusion of proximal scaphoid and lunate, enlarged triquetrum, enlarged and dysplastic distal humerus that articulates with a reduced olecranon and the proximal radius forming a knee-like joint (figure 1B). Array-CGH (20 kb resolution) of individual II:2 was negative. We next performed WGS and identified a 8.5 kb deletion chr5:134 729 406–134 737 909 (hg19) spanning the first non-coding exon of H2AFY (figure 1C). The deletion cosegregates with the disease in all affected members of the family. Subsequent PCR and Sanger sequencing analysis confirmed the breakpoints (figure 1D and E).
CRISPR-Cas9 reengineering of 8.5 kb Liebenberg deletion in mice
In our previous study, we showed that the Pen enhancer and Pitx1 form close three-dimensional chromatin interactions specifically in the hindlimb, whereas they are separated in forelimbs. SVs at this locus cause aberrant chromatin interactions between the Pen and Pitx1 in forelimbs that result in strong misexpression and arm-to-leg transformation. The SVs re-engineered in mice correspond to 99 kb regions in mice, thus the region responsible for insulating Pen from Pitx1 in forelimbs remains unknown.10 Excitingly, the novel 8.5 kb deletion identified significantly narrows down the putative forelimb specific insulator that normally prevents Pen from acting on Pitx1 in forelimbs (figure 1C).
To investigate whether the 8.5 deletion identified in this study causes Liebenberg syndrome, we first employed CRISPR-Cas9 engineering in mice. We used two sgRNA to induce double strand breaks at the corresponding sites in the mouse genome inducing the error prone non-homologous end joining resulting in indels, deletions, inversions and duplications.11 A homozygous deletion clone Pitx1lieb/lieb corresponding to a 10 kb deletion was aggregated and embryos were analysed at E11.5 for mRNA expression using qRT-PCR and whole-mount in situ hybridisation (WISH) (figure 2A and B). Mutant forelimbs showed a fourfold upregulation of Pitx1 compared with wild type mice (figure 2A and B). At 6 weeks, we analysed the mice using micro-CT to check for possible skeletal malformation. However, the mice did not display any skeletal phenotype (not shown).
Next, we enquired whether the relative reduced linear distance between Pen and Pitx1 may determine the activity of Pitx1 in the forelimb or if the H2afy transcription on its own has an insulating effect. To test this hypothesis, we deleted a 58 kb region encompassing the H2afy gene body, leaving the upstream promoter region intact including non-coding exon 1 (Pitx1del3/del3 ). Strikingly, this deletion is six times greater than in Pitx1lieb/lieb , yet E11.5 Pitx1del3/del3 mutants did not display any Pitx1 expression in developing forelimbs (figure 2A and B). Thus, the relative reduced distance between Pen and Pitx1 did not induce Pitx1 misexpression, but rather the loss of H2afy promoter region and consequently its transcription. Finally, we asked if Pen is the sole pan-limb enhancer that drives Pitx1 misexpression in developing forelimb. We engineered a deletion spanning H2afy region to include Pen, thereby removing the enhancer. Indeed, the mutant mice did not exhibit any ectopic expression in forelimb buds (figure 2A and B). This confirms that the endogenous Pen enhancer is the pathogenic driver of Pitx1 expression in forelimbs of Pitx1lieb/lieb mice.
Altogether, these findings lead us to propose the following model (figure 2C): in the wild type condition, the inactive chromatin conformation of the Pitx1 locus in the forelimb prevents Pen from activating Pitx1. However, changes in the linear organisation of regulatory elements lead either the loss of the H2afy promoter region or place Pen 3’ of the gene allowing ectopic chromatin interactions between cognate Pitx1 and Pen. Consequent misexpression in forelimbs results in Liebenberg syndrome both in mice and humans. Nevertheless, how the H2afy promoter insulates Pen specifically in forelimbs remains elusive.
In this study, we used WGS to identify the smallest deletion causing Liebenberg syndrome and studied its pathogenic mechanism by CRISPR-Cas9 engineering in mice. We previously demonstrated that Liebenberg syndrome is a result of regulatory endoactivation of Pitx1 in forelimbs.10 This was demonstrated by misplacing Pen 3’ of H2afy promoter that resulted in Pitx1 misexpression in forelimbs and consequent arm-to-leg transformation in mice, recapitulating the human Liebenberg syndrome.10 In the current study, the re-engineering of the smallest Liebenberg syndrome allele that removes the H2afy promoter region (Pitx1lieb/lieb ) resulted in a molecular phenotype with mild Pitx1 misexpression in forelimb. Moreover, this misexpression was rescued when the deletion extends to include the Pen enhancer (Pitx1del4/del4 ). The absence of a skeletal phenotype in Pitx1lieb/lieb adult mice suggests that a threshold exists whereby only a certain Pitx1 (mis)expression level induces hindlimb morphology. Furthermore, the level of Pitx1 mRNA determines the severity of the phenotype. This is also observed in the human Liebenberg phenotype whereby the closer Pen is placed in a linear relation to PITX1, the more complete the transformation of the arms into legs is.3 5 This is particularly evident in the duplication case that misplaces PITX1 only 10 kb from Pen, rather than the 400 kb linear distance in the wild type genome, creating the most severe arm-to-leg transformation documented.14 In line with this observation, the Pitx1-Pen transgenic mouse shows strong ectopic expression of Pitx1 and severe forelimb malformation leading to loss of a zeugopodal elements and digits and the elbow joint displayed striking similarities to the knee joint).3 5 Similar examples are observed, for example, at the Shh locus where the dosage of gene misexpression correlates with the severity of the phenotype.15 Finally, the absence of bone malformation in mice harbouring human SVs has recently been observed in several studies indicating that mice are more robust and that differences exist between humans and mice.12 16 17
Interestingly, the ectopic interaction between Pitx1 and Pen is not induced by the sole reduction in proximity, as Pitx1del3/del3 mutants with a 58 kb deletion did not induce ectopic expression of Pitx1 in forelimbs, whereas a 10 kb deletion did (Pitx1lieb/lieb ). Thus, not all sequences are equal and how H2afy promoter mechanistically blocks Pen from activating Pitx1 in forelimb tissue remains an open question. It is thus striking that a tissue-unspecific promoter serves as an insulator. One can speculate that either the expression level of H2afy might be different or that the chromatin interactions with Pen alter. However, H2afy expression is the same in all limbs and no differences in chromatin interactions are detected between Pen and H2afy.10 Interestingly, in Pitx1pen/pen mice,10 in which Pen is deleted, H2afy expression is reduced (unpublished data), supporting that Pen regulates H2afy expression. Even so, the active conformation of the Pitx1 locus in hindlimbs, which brings Pen in close 3D proximity with Pitx1 and with the high affinity for its endogenous promoter enables their pairing and robust transcription, while in the forelimb, they remain separated (Model, figure 2C).
In conclusion, we suggest the following mechanisms: (1) H2afy promoter functions as an insulator by scavenging Pen activity in forelimbs; (2) Promoter-promoter competition between Pitx1 and H2afy with Pen whereby the enhancer has a higher affinity for Pitx1. Differential binding of active and repressive TFs at Pen and H2afy promoter regions may alternate between forelimbs versus hindlimbs to the differential Pen activity. The loss of H2afy promoter thus leaves Pen free to interact with the next promoter, the cognate Pitx1 promoter, leading to its misexpression and, consequently, to Liebenberg syndrome. Recently, a similar mechanism was demonstrated in cancer cells where the PVC1 promoter was shown to have a tumour-suppressor function that regulates the release of MYC transcription on the monoallelic chromosome.18
The authors would like to thank the patients and their families for participating in the research study. We thanks colleagues for fruitful discussions and input on the study.
Contributors BKK and MS: designed the study, collected the data and performed the genetic investigations. BKK performed mouse experiments. BKK and MS: analysed the data and wrote the manuscript. FB, MCD, SM and MS: performed the clinical evaluation of the patients. All authors revised and approved the final version.
Funding This study was supported by a grant from the Deutsche Forschungsgemeinschaft (SP1532/2-1). FB received funding from grant GR-2013-02356227.
Competing interests None declared.
Patient consent Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.