The presence of the different copy number variations on chromosome 14q21-22 (Del1, Del2, Del3, Dupl1, and Dupl2) is indicated for each individual. Del1 involves the GCH1 and BMP4 genes, which are considered to be relevant for the phenotype. Individuals in the younger generations are shown as sex-unspecific diamonds to protect confidentiality of the family. Asterisks denote individuals from whom DNA samples were not obtained. The proband is indicated by a black arrowhead (individual III:5). Other abnormalities of eyes indicates myopia, retinal detachment, and/or macular hemorrhages.
aDel2 and Del3 deletion as well as Dupl1 and Dupl2 duplication.
bDel1 (including GCH1 and BMP4) and Del3 deletion and no duplication.
cNo deletion and no duplication.
Metaphases of individual II:2 (A-C) and the unaffected sibling, individual II:3 (D-F), were hybridized with 3 different probes (left panel, bacterial artificial chromosome clone [BAC] RP11-81D1 [RPCIB753D0181Q8]; middle panel, BAC RP11-12P7 [RPCIB753P0712Q8]; right panel, BAC RP11-105H21 [RPCIB753 H21105Q8]) from the chromosome 14q22 region. Chromosomes 14 were hybridized with a telomeric probe (red). For II:2, all 3 probes (green) can only bind to 1 chromosome 14, indicating a heterozygous deletion of all 3 sequences; for II:3, only the BAC clone RP11-81D1 (D) is involved in the deletion. The position of the BACs as well as of the disease-relevant genes GCH1 and BMP4 are shown where the chromosome is presented in red, and the 2 deletions (Del1 and Del2) are indicated as green lines.
Results of the array comparative genomic hybridization of chromosome 14 in 5 family members (individuals I:2, II:2, II:3, III:4, and III:5). Top, schematic representation of chromosome 14. Middle, results from the genome-wide array comparative genomic hybridization for the region of interest on chromosome 14. A ratio (vertical lines) of 0 indicates that the control and the investigated sample have the same gene dosage (2 copies of each sequence). A ratio of approximately −1 indicates a heterozygous deletion (only 1 copy of the respective sequence). A ratio of +1 indicates a heterozygous duplication (3 copies of the respective sequences). The 2 types of deletions can clearly be differentiated. In addition, carriers of the 669–base pair deletion also have a 330-kb duplication. Names and locations of genes in this region are shown. Bottom, results of the customized high-resolution array comparative genomic hybridization clearly indicated the breakpoints and revealed another 13-kb duplication in II:3.
Top, chromosomal rearrangements involving 2 deletions harbored by individuals affected with DRD (individuals II:2, III:4, and III:5). Middle, wild-type chromosome 14. Bottom, complex chromosomal rearrangement harbored by unaffected individuals (individuals I:2 and II:3), triggering the expansion to a larger deletion. Fragment coordinates (hg19): a, centromere (cen.) to 40 377 930; b, 40 377 931 to 40 407 405; c, 40 407 406 to 53 632 860; d, 53 619 511 to 54 146 632; e, 53 819 947 to 56 319 270 (encompasses BMP4 and GCH1); f, 56 319 271 to 56 987 977; g, 56 987 978 to telomere (tel.). The rearrangements are not drawn to scale.
eTable 1. Microsatellite markers and amplicons investigated to elucidate the size and location of the deletions on chromosome 14q.
eTable 2. Whole-genome sequencing metrics of patient II:3.
eTable 3. Copy number variations and overlapping protein coding genes in unaffected and affected members of the family.
eTable 4. Genes disrupted by the breakpoints on chromosome 14q.
eFigure 1. Genotyping results in the family in the chromosome 14q22 region.
eFigure 2. Identified breakpoints in the chromosome 14 region.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Lohmann K, Redin C, Tönnies H, et al. Complex and Dynamic Chromosomal Rearrangements in a Family With Seemingly Non-Mendelian Inheritance of Dopa-Responsive Dystonia. JAMA Neurol. 2017;74(7):806–812. doi:10.1001/jamaneurol.2017.0666
What is the underlying genetic cause in a family with dopa-responsive dystonia as well as skeletal and eye abnormalities?
Using a variety of molecular methods to examine 10 members of a family, including 5 with dopa-responsive dystonia and skeletal and eye abnormalities, we revealed complex chromosomal rearrangements in both affected and unaffected patients comprising a heterozygous deletion of several disease-linked genes (contiguous gene syndrome). Deleted genes included GCH1, which is responsible for the dopa-responsive dystonia, and BMP4, which has been linked to skeletal and eye abnormalities.
These findings alert neurologists to the importance of clinical red flags indicating chromosomal rearrangements, the detection of which requires an interdisciplinary approach in modern clinical-diagnostic care.
Chromosomal rearrangements are increasingly recognized to underlie neurologic disorders and are often accompanied by additional clinical signs beyond the gene-specific phenotypic spectrum.
To elucidate the causal genetic variant in a large US family with co-occurrence of dopa-responsive dystonia as well as skeletal and eye abnormalities (ie, ptosis, myopia, and retina detachment).
Design, Setting, and Participants
We examined 10 members of a family, including 5 patients with dopa-responsive dystonia and skeletal and/or eye abnormalities, from a US tertiary referral center for neurological diseases using multiple conventional molecular methods, including fluorescence in situ hybridization and array comparative genomic hybridization as well as large-insert whole-genome sequencing to survey multiple classes of genomic variations. Of note, there was a seemingly implausible transmission pattern in this family due to a mutation-negative obligate mutation carrier.
Main Outcomes and Measures
Genetic diagnosis in affected family members and insight into the formation of large deletions.
Four members were diagnosed with definite and 1 with probable dopa-responsive dystonia. All 5 affected individuals carried a large heterozygous deletion encompassing all 6 exons of GCH1. Additionally, all mutation carriers had congenital ptosis requiring surgery, 4 had myopia, 2 had retinal detachment, and 2 showed skeletal abnormalities of the hands, ie, polydactyly or syndactyly or missing a hand digit. Two individuals were reported to be free of any disease. Analyses revealed complex chromosomal rearrangements on chromosome 14q21-22 in unaffected individuals that triggered the expansion to a larger deletion segregating with affection status. The expansion occurred recurrently, explaining the seemingly non-mendelian inheritance pattern. These rearrangements included a deletion of GCH1, which likely contributes to the dopa-responsive dystonia, as well as a deletion of BMP4 as a potential cause of digital and eye abnormalities.
Conclusions and Relevance
Our findings alert neurologists to the importance of clinical red flags, ie, unexpected co-occurrence of clinical features that may point to the presence of chromosomal rearrangements as the primary disease cause. The clinical management and diagnostics of such patients requires an interdisciplinary approach in modern clinical-diagnostic care.
While the increasingly widespread use of genetic testing has led to a sharp rise in the number of confirmed molecular diagnoses in patients with previously unresolved neurologic conditions,1 it has also been paralleled by the discovery of an unexpected complexity of phenotype-genotype relationships, which challenge neurologists in clinical practice. With evolving technologies, different classes of DNA mutations that were undetectable by classic sequence analysis have been recognized as important causes of diseases, such as deletions or duplications of 1 or more exons.2 While approximately 5% of all mutations related to monogenic diseases were ascribed to large insertions or deletions in 2001, approximately 10% of reported mutations in the Human Genome Mutation Database (http://www.hgmd.cf.ac.uk/ac/hahaha.php) are currently explained by chromosomal rearrangements. This number is likely still an underestimate, given the low resolution of conventional molecular methods and the fact that quantitative screening methods are not routinely applied.3
Chromosomal rearrangements often encompass several genes and are frequently causative of neurodevelopmental disorders, such as intellectual disability and neuropsychiatric conditions.4,5 Exonic rearrangements also represent a frequent type of mutation, for instance in PRKN, also known as Parkin or PARK2, causing Parkinson disease.6 Likewise, 10% to 40% of mutations in the GTP cyclohydrolase 1 (GCH1) gene causing autosomal dominantly inherited dopa-responsive dystonia (DRD) are deletions of 1 or more exons,7-10 rarely involving all 6 exons.7,11 Dopa-responsive dystonia is typically characterized by early-onset leg dystonia, diurnal fluctuations, and an excellent response to levodopa.12,13
Here, we present a family with DRD and a large deletion encompassing the GCH1 and adjacent genes. Affected family members also had ptosis and skeletal abnormalities. Surprisingly, this large deletion was absent in an obligate carrier. Using cytogenetic, molecular, and novel genomics methods, we revealed highly complex chromosomal rearrangements in the obligate carrier and the mother, explaining the unusual phenotypic features and pattern of transmission that does not follow a mendelian pattern of inheritance.
The index patient (A27674, III:5) of the family (Figure 1) was previously reported to carry a heterozygous deletion of all 6 GCH1 exons.7 Informed consent was obtained from all individuals, and the study was approved by the Ethics Committee at the University of Lübeck. Neurological examination, systematic videotaping,14 and medical record review were possible for 7 individuals. For an additional 3 family members, only medical reports were available. Diagnostic criteria for DRD were definite dystonia with a marked and sustained response to levodopa therapy.15 Whole-blood samples of all 8 available family members were collected for genetic study.
Genotyping was performed on the IR2 DNA sequencer (LI-COR Biosciences) using standard protocols. Paternity testing was carried out using 12 highly polymorphic microsatellite markers on different chromosomes. Haplotype analysis in the 14q22 region comprised a total of 42 (including 15 newly designed, D14Skh1-26) microsatellite markers (eTable 1 in the Supplement).
Large structural abnormalities were examined by G-banding on metaphase chromosomes. Three different bacterial artificial chromosome clones from the German Resource Center for Genome Research (RPCIB753D0181Q8, RPCIB753P0712Q8, RPCIB753H21105Q8; RZPD Berlin) were used for fluorescence in situ hybridization (FISH) on chromosome 14q22.
Quantitative polymerase chain reaction (PCR) was carried out with SYBR Green I on the LightCycler 2.0 (RocheDiagnostics) using primers for 22 amplicons on the genomic level in the breakpoint regions (eTable 1 in the Supplement). Calculated concentrations were compared with β-globin that served as a reference gene.
To further elucidate the DNA rearrangement on chromosome 14, we carried out array comparative genomic hybridization (aCGH) using the Human Genome Microarray 244A platform, which contains 238 381 oligonucleotides (60-mer) distributed across the human genome (Agilent). Array CGH was performed according to standard protocols. For the subsequent analysis, we focused on chromosome 14 and applied the Aberration Detection Method-2 (ADM-2) algorithm of the Agilent CGH Analytics software with a threshold of 6.0. Significant chromosomal gains and deletions were defined by a log2-ratio of 10 neighboring oligonucleotides exceeding 0.4 and corresponding to an average resolution of approximately 100 kb. To obtain a higher resolution of the breakpoint regions, we used a customized high-resolution chip for the region of interest on chromosome 14 (Nimblegen). The tiling region (53.4 Mb to 57.9 Mb) was covered by approximately 39 600 single probes with an average probe size of 56 base pairs, corresponding to 49% of the tiling region.
Whole-genome sequencing (WGS) for individual II:3 was performed using our custom large-insert jumping library protocol.16,17 In brief, DNA was sheared to a target size of 3 kb using a Covaris S2 instrument. Sonicated DNA was end-repaired, ligated to cap adapters containing EcoP15I digestion sites, and circularized to an internal biotinylated stuffer. Circularized biotinylated fragments were EcoP15I-digested, end-repaired, and pulled down using Streptavidin-coated beads. Sequencing libraries were finalized after ligation of sequencing adapters and further PCR amplification. Sequencing was performed using an Illumina HiSeq 2500 and generated more than 153 million of 25–base pair paired-end reads, reaching 80-fold of physical coverage (eTable 2 in the Supplement).
All computational analyses have been described previously.5,16-22 Alignment of reverse-complemented paired-end reads was performed with Burrows-Wheeler aligner.23 Clusters of discordant read-pairs regarding orientation, insert size, and mate mapping location were extracted using Sambamba24 and classified under a specific structural variant category (deletion, insertion, inversion, and translocation) based on their signatures.
Four members were diagnosed with definite and 1 with probable DRD (Figure 1) (Table). All 5 affected individuals carried a large heterozygous deletion encompassing all 6 exons of GCH1. Additionally, all mutation carriers had congenital ptosis requiring surgery, 4 had myopia, 2 had retinal detachment, and 2 showed skeletal abnormalities of the hands, ie, polydactyly or syndactyly or missing a hand digit (Figure 1) (Table). Individuals I:2 and II:3 were reported to be free of any disease.
One obligate carrier (II:3), whose sibling and offspring were both affected, screened negative for the GCH1 deletion despite confirmed paternity. Haplotype analysis on chromosome 14q22 revealed several incompatibilities that could be best explained by large deletions (eFigure 1 in the Supplement). Quantitative PCR confirmed 2 different deletions in members of the family; a larger deletion (approximately 3338 to 3479 kb; Del1) was found in affected family members comprising markers D14Skh25-D14Skh6, including GCH1, whereas in II:3 and the mother I:2, haplotype analysis suggested another smaller deletion (approximately 667 to 671 kb; Del2) from D14Skh16-D14Skh6 located downstream of GCH1 (Figure 1) (eTable 1 in the Supplement).
The presence of 2 different deletions on chromosome 14q22 was confirmed by FISH, where all 3 bacterial artificial chromosome clones were involved in Del1, but only RPCIB753D0181Q8 fell within Del2 (Figure 2). Because the exact breakpoints could not be determined, more complex genomic rearrangements were assumed, but large interchromosomal rearrangements were excluded cytogenetically. However, we identified an approximately 330-kb duplication (Dupl1) by genome-wide aCGH (Figure 3, upper panel). Further, we also found an approximately 13-kb duplication (Dupl2) on chromosome 14q22 in carriers of Del2 by high-resolution aCGH (Figure 3, lower panel). A deletion-spanning PCR revealed that Dupl2 was located at the centromeric breakpoint of both Del1 and also, surprisingly, Del2.
The final and conclusive experiment was WGS of unaffected carrier II:3 using large-insert jumping libraries. This revealed an additional, more distantly located approximately 29-kb deletion (Del3) on chromosome 14q21 more than 13 Mb proximal to Del1. It also revealed a total of 4 breakpoints in individual II:3 in the chromosome 14q21.1-q22.3 region involving 2 large deletions (Del3 [29 kb] and Del2 [669 kb]), 2 large duplications (Dupl2 [13 kb] and Dupl1 [328 kb]), and 2 inverted segments of 2.7 Mb total (Figure 4, lower panel), both of which were lost in affected individuals, resulting in the expansion to Del1 (Figure 4, upper panel). All breakpoints were confirmed by Sanger sequencing and also found in I:2 (eFigure 2 in the Supplement).
Genes within the structural variations are listed in eTable 3 in the Supplement. Within Del1, this included GCH1, mutations in which cause DRD, and bone morphogenic factor 4 (BMP4), a gene previously linked to ptosis, microphthalmia, and brain and digit anomalies.25,26 While there were no genes located in the areas of Dupl1 and Del3, PELI2 was affected by Del1/Del2 and DDHD1 by Dupl2. DDHD1 was the only characterized gene disrupted by a breakpoint (eTable 4 in the Supplement). Of note, biallelic DDHD1 mutations have been reported in spastic paraplegia 28.27
Our molecular findings not only provide an explanation for the additional phenotypic features beyond DRD, ie, a contiguous gene syndrome, but also for the seemingly non-mendelian pattern of inheritance. In this family, a small 669-kb deletion (Del2) expanded independently twice (I:2 to II:2 and II:3 to III:5) to the identical deletion of 3355 kb (Del1) triggered by complex chromosomal rearrangements, including a large 2.7-Mb inversion (Figure 4). Of note, the deletion of the inverted sequence resulted in a de novo deletion of the disease-relevant genes GCH1 and BMP4 in II:2 and III:5. There was no indication for mosaicism in these patients by any of the methods used, including single-cell FISH and WGS analyses, suggesting that it represents a germline expansion.
Importantly, these findings are of immediate translational value, as they directly affect genetic diagnostics and specific counseling of the family, including carriers of Del2. Although the complexity of genomic rearrangements detected in this pedigree is likely exceptional and the series of investigations required to conclusively resolve the underlying mechanism clearly exceed those available in a purely diagnostic setting, it is important to alert neurologists to the occurrence of mutational events beyond those detectable by simple Sanger sequencing. Indeed, we have shown that the co-occurrence of paired duplications by chromosomal microarray often underlie large inversions (termed dupINVdup),20 but this was not the case in this family, in which repetitive or duplicated sequences were not present at the 2 different ends of a deletion breakpoint (eTable 4 in the Supplement). Recent analyses further demonstrated that each human genome harbors an average of 14 large, complex structural rearrangements.22 However, many such rearrangements are only detectable through the use of sequence-based methods and intensive computational analyses. Although, to our knowledge, the structural variations detected in the present family are novel, copy number variations in the chromosome 14q22 region do not seem to be unique events. For instance, DECIPHER, a database focusing on genomic variants, lists approximately 50 of 23 000 patients with diverse phenotypes and copy number variants involving sequences on chromosome 14q22 (https://decipher.sanger.ac.uk/search?q=14q22#consented-patients/results).28 Further, Frias syndrome characterized by mild exophthalmia, palpebral ptosis, and hypertelorism is caused by an even larger deletion of approximately 4 Mb in the chromosome 14q22 region.25
The occurrence of additional phenotypic features in this family, ie, eye (ptosis, myopia, and retinal detachment) and skeletal (polydactyly, syndactyly, and missing digit) abnormalities, pointed to the presence of a larger deletion encompassing not only the GCH1 gene but also neighboring genes, thereby causing the additional phenotypic features. Similar observations have been reported, for example, in the context of myoclonus-dystonia with deletions of both SGCE and the adjacent COL1A3 gene, resulting in additional abnormalities of bone and cartilage.29 Thus, clinically complex cases require a close collaboration of astute clinicians with molecular geneticists. This includes reporting all phenotypic details, especially those that allegedly do not fit the disease spectrum. Among the 16 additional genes within the deleted region in this family, BMP4 is the best candidate to explain both the eye and skeletal abnormalities found in mutation carriers in this family with DRD. Mutations in BMP4 have previously been found in patients with anophthalmia-microphthalmia, ptosis, and developmental anomalies of the digits. Furthermore, BMP4 is also deleted in Frias syndrome.25,26 Haploinsufficiency of the remaining 15 genes do not seem to be related to any disease (eTable 3 in the Supplement). Of note, a breakpoint of Dupl2 affects the DDHD1 gene. However, none of the patients in this family developed symptoms of spastic paraplegia 28, a recessive disorder requiring mutations on both alleles, because they do retain 2 functional copies of DDHD1; one chromosome in this family is fully intact and a second functional copy of DDHD1 is located on the rearranged chromosome as illustrated in Figure 4.
In keeping with our area of expertise, we approached the family from the dystonia phenotype. The importance of complete clinical and molecular characterization is underlined by the story of the blind men and the elephant; if we had focused on the eye and/or skeletal abnormalities in our family, we would have easily missed or misinterpreted the dystonic features that, for this specific form of dystonia, are exquisitely treatable with levodopa. Although the deletion in Frias syndrome also comprises the GCH1 gene,25 dystonia is not mentioned in these patients. This may be because of reduced penetrance or because clinical features of DRD may have been mild and overlooked.
On a technical note, both classic (FISH) and modern (aCGH) cytogenetic methods in conjunction with dense haplotype analysis and quantitative PCR were instrumental in detecting the disease-relevant large deletion (Del1). However, to definitively elucidate the breakpoints and, more importantly, the exact mechanism underlying the recurrent expansion to a 669-kb deletion (Del2), WGS of large-insert libraries16 had to be used to uncover the inversion (structural variation).
Our study had limitations. While we convincingly solved the cause of DRD in this family, the skeletal and eye phenotypes are probably linked to the BMP4 deletion. However, we cannot explain the penetrance and phenotypic variability ranging from myopia to retinal detachment and macular hemorrhages as well as missing or additional digits. Additionally, likely genetic modifiers may influence the spectrum of skeletal and eye abnormalities in carriers of a BMP4 deletion.
In conclusion, with most inherited conditions having a neurologic manifestation and a large proportion of neurologic diseases being attributed to a genetic cause, neurologists should be familiar with clinical red flags, ie, co-occurrence of seemingly unrelated disorders, indicative of contiguous gene syndromes due to copy number variations. Further, neurologists should have a basic understanding of experimental techniques to elucidate the underlying molecular mechanisms, of which this family with DRD serves as an intriguing example.
Corresponding Author: Christine Klein, MD, Institute of Neurogenetics, University of Lübeck, Ratzeburger Allee 160, 23538 Lübeck, Germany (email@example.com).
Accepted for Publication: March 23, 2017.
Published Online: May 30, 2017. doi:10.1001/jamaneurol.2017.0666
Author Contributions: Drs Klein and Lohmann had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Lohmann, Ozelius, Siebert, Talkowski, Saunders-Pullman, Klein.
Acquisition, analysis, or interpretation of data: Lohmann, Redin, Tönnies, Bressman, Martin-Subero, Wiegers, Hinrichs, Hellenbroich, Rakovic, Raymond, Schwinger, Siebert, Talkowski, Saunders-Pullman, Klein.
Drafting of the manuscript: Lohmann, Redin, Talkowski.
Critical revision of the manuscript for important intellectual content: Redin, Tönnies, Bressman, Martin-Subero, Wiegers, Hinrichs, Hellenbroich, Rakovic, Raymond, Ozelius, Schwinger, Siebert, Talkowski, Saunders-Pullman, Klein.
Statistical analysis: Talkowski.
Obtained funding: Lohmann, Saunders-Pullman, Talkowski, Klein.
Administrative, technical, or material support: Lohmann, Tönnies, Hinrichs, Hellenbroich, Raymond, Siebert, Saunders-Pullman, Klein.
Supervision: Lohmann, Siebert, Talkowski, Klein.
Conflict of Interest Disclosures: Dr Klein works as a genetic consultant at Centogene AG. No other disclosures were reported.
Funding/Support: This work was supported by grants from the Edith-Fröhnert-Stiftung, the German Ministry of Education and Research (BMBF; Dystonia Translational Research and Therapy consortium; grant 01GM1514B), the German Research Foundation (DFG; grant FOR2488) (Drs Lohmann and Klein), the Dystonia Medical Research Foundation (Drs Bressman and Saunders-Pullman), and the Hermann and Lilly Schilling Foundation (Dr Klein) as well as by grants K23 NS047256 (Dr Saunders-Pullman), MH095867 (Dr Talkowski), and GM061354 (Dr Talkowski) from the National Institutes of Health. Dr Talkowski was also supported as the Desmond and Ann Heathwood Research Scholar.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.