Final Report Summary - NEUROXSYS (Genomic Regulatory Systems of Human X-linked neurological diseases)
The human X chromosome comprises 150 Mb of DNA sequence, about 5 % of the total length of the human genome. It contains about 800 genes that code for protein. A disproportionally high number of Mendelian (single gene) diseases has been mapped to the X chromosome since gene mutations on this chromosome tend to be maintained in the population because the reproductive fitness of carrier females is generally not affected, yet the recessive phenotype is evident in their sons if they inherit the mutant chromosome. Several large-scale efforts have focused on disease gene-discovery on the X chromosome, and more than 90 genes (over 10 % of the total number of genes in the X chromosome) are currently implicated in the development of X-linked intellectual disability (XLID). Nevertheless, sequencing of the exons of 718 genes on the X chromosome explains disease only in about half of the families. This means that a large proportion of XLID mutations should be found in non-coding DNA, in sequences that determine the activity of the genes (so-called enhancers). Based on the expertise of a consortium of seven groups in Europe and Australia, the NEUROXSYS project has aimed to visualise the expression of X-linked genes known to be involved in the development of intellectual disability and their regulatory sequences using computational analysis, large-scale transgenesis in the zebrafish, chromatin immunoprecipitation in human and mouse tissues, and analysis of relevant regulatory sequences and gene expression patterns at cellular resolution in the zebrafish brain. Lastly, NEUROXSYS has sequenced a large number of putative regulatory sequences and all coding genes in 48 families with XLID with no known mutation and is in the process of analysing mutations in non-coding DNA in these families.
Project context and objectives:
The main objectives of the project are:
1. Establish a database of the X chromosome containing all highly conserved non-coding elements (HCNEs) on the human X chromosome, with accompanying tools for analysis and visualisation (work package (WP)1) that identify HCNEs in relation to their likely target genes through:
(i) HCNE density plots across the entire X chromosome and
(ii) through mapping of all conserved synteny blocks between the human X chromosome and distant vertebrate genomes: five teleosts, chicken, frog and opossum, but also using multiple mammalian genomes.
These were surveyed for breakpoints, chromosomal aberrations, or point mutations causing neurological disease. Each conserved synteny block will contain one or more genes and will allow the identification of HCNEs and their target genes.
2. Generate a cis-regulatory map of the human X chromosome (WP2) by extracting conserved non-coding sequences from all gene neighborhoods. These conserved non-coding human sequences were assayed in a high throughput approach in transgenic zebrafish. We used known loci for XLMR and mapped all elements around these loci and extract the regulatory content of 440 of these. The results from this screen, in the form of images of expression patterns, are stored in the database and are used by all beneficiaries to evaluate whether they are candidates for further investigation of brain diseases.
3. Visualise neuronal patterns of expression driven by human HCNEs (WP3). We analysed elements that yielded brain or neuronal-specific expression by imaging at the cellular level in embryonic, larval, and adult zebrafish brains. This will be used to evaluate whether the elements are candidates for testing in human patient DNA for deletions, rearrangements, and mutations.
4. Identify human elements mutated or deleted / duplicated in X-linked neurological disease using arrays and sequencing of patient DNA (WP4). Finally, we sequenced all detected HCNEs plus all coding exons (about 13 Mb, or 8 % of the X chromosome) in 48 families that had no known coding mutations. These approaches will also boost the understanding of loci that are already mapped, but where the neurobiological basis for development of disease is currently not understood.
Project results:
Work performed so far and results obtained
1. The development and application of computational method to identify putative enhancers and their target genes on the human X chromosome (WP1)
Partner 5 has identified 102 647 putative enhancers on the X chromosome and has predicted a unique target gene for 38 591 of them, of which 12 592 target a known XLID gene. A bioinformatic method that identifies cases of evolutionary co-segregation between these and specific genes was developed. The underlying working hypothesis is that a rearrangement occurring between an enhancer and its target gene would separate the two elements, causing the target gene to be misregulated. The rearrangement would likely be counter-selected during evolution, leading to a high degree of conservation of the functional enhancer / gene pairs. When comparing modern genomes, this would translate to of a high level of co-segregation, where the enhancer and its target gene would be found in the vicinity of each other in most or all genomes where the two sequences are presents. This is the underlying phenomenon thought to explain the observation of genomic regulatory blocks (GRBs) within vertebrate genomes.
2. The management, integration, and dissemination of results generated during the project
WP1 has generated several web-based tools that allow the analysis of the evolutionary history of individual GRBs, namely Ancora and Genomicus. WP1 has also established a genome browser based on the UCSC browser that incorporates data such as HCNEs (WP1), tested enhancers (WP2) and ChIP data (WP4).
3. Generation of a cis-regulatory map of the human X chromosome by extracting conserved non-coding elements from known disease-gene loci and assay in in transgenic zebrafish (WP2)
The work plan aimed to test 400 sequence elements in total. Two hundred elements by partner 6 and 200 elements by partner 7. In total, partner 6 cloned and functionally analysed 218 (corresponding to 220 Kb) highly conserved sequences located in the vicinity of genes involved in human neurological disorders. Thirty one sequence elements where identified to drive tissue specific expression of the GFP reporter. Twenty four of these elements directed GFP expression to neural tissues (11 %). Three elements showed muscle specific expression, 2 elements where expressed in the notochord and two in the eye. Partner 7 has tested a total of 223 HCNEs with an average length of 1531 bp, corresponding to a total of 340 Kb (0.21 %) of the human X chromosome. Of these, 33 (15 %) acted as reproducible enhancers in the brain (in 3 or more independent insertions each), and 72 (32 %) were found to be partially reproducible in the brain. The remainder were either reproducible, but not in brain (21.9 %), or partially reproducible outside of the brain or inconsistent (43.5 %).
4. Visualise neuronal patterns of expression driven by critical human HCNEs at the cellular level in embryonic, larval and adult zebrafish brains (WP3)
Partner 8 and 3 have characterised 25 transgenic lines at the juvenile stage and 20 lines at the adult stage, and have also established 20 gene expression patterns of zebrafish orthologues of human X chromosome genes. The results establish that not all enhancers are active in the adult brain, but some were found to correspond to regions containing adult neural stem cells. A substantial fraction, however, did not show overlap with the endogenous expression of the supposed target gene while yet others exhibited ectopic expression when compared to the endogenous expression pattern.
5. Identify human elements mutated or deleted / duplicated in X-linked neurological disease using arrays and sequencing of patient DNA (WP4)
Partner 2 has used Chromatin immunoprecipitation (ChIP) with antibodies against H3K4me1, p300 (EP300), CBP (CREBBP) and CTCF to identify potential cis-regulatory elements in chromatin isolated from both human foetal brain and E14.5 E16.5 and P0 developing mouse brain. The resulting DNA was hybridised to oligonucleotide tiling arrays for both the mouse and human X-chromosome. There was a high degree of reproducibility on both technical and biological replicates when either cDNA or immunoprecipitated DNA was hybridised to the oligonucleotide arrays. Following advice from the SAC RNA-seq was used to define the transcriptome during human fetal brain development. This process detected all the known early brain enhancers and has revealed many thousands of novel putative enhancers that are apparently active in developing brain tissue. These data are available on as UCSC custom tracks at http://genome-neuroxsys.genereg.net Further, WP4 analysed over 300 patient samples with Cornelia de Lange syndrome and eye malformations for potential deletions, but none were identified. WP4 also started sequencing of approximately 16Mb encompassing non-coding sequence (from WP1) and all exons in 48 patients with plausible XLID with no known coding mutations. The coverage was approximately 80-fold. Initial analysis of this sequencing has revealed plausible coding region mutations in 5 / 48 cases that had been missed in the previous analysis. All other cases show interesting mutations in highly conserved non-coding elements. Currently, partner 2 is following up eight mutations in elements that appear to control known XLID genes.
Potential impact:
Expected final results, impacts, and use
We expect that the final results of the NEUROXSYS consortium will significantly expand the body of information on XLID as well as other X-linked neurological diseases, and will discover hitherto unknown disease mechanisms in both coding and non-coding sequence. These results should also have a direct impact on families carrying these unknown mutations, as new diagnostic tools become available. As a result, the X chromosome will represent the most advanced arena of human neurological disease-causing mutations, both in terms of coding and non-coding mutations.
NEUROXSYS online tools and database
NEUROXSYS has created a wealth of online tools that help with investigation of mutations on the X chromosome, and the annotation of potential target genes to a large number of conserved non-coding, and potentially regulatory, elements. Further, a total of 560 Kb of human sequence have been tested in transgenic zebrafish, which should aid in understanding the non-coding mutations currently evaluated by NEUROXSYS as well as future mutations on the X chromosome.
The identification of non-coding sequence variants that segregate with XLID will lead to advanced understanding of non-coding variants in monogenic disease. This has been made possible through the large-scale bioinformatic work on non-coding, and regulatory, sequence on the human X chromosome by the NEUROXSYS consortium.
Pathogenesis of X-linked neurological disease
Although we have not yet been able to successfully test mutated and disease-linked non-coding sequence in transgenic zebrafish to evaluate potential differences between normal and mutated sequence, this will be possible with new variants as they are identified. In such cases, we should be able to observe in the zebrafish the affected brain areas and the underlying mechanism.
Project website: http://neuroxsys.net/
Dr. Boris Lenhard
E-mail: b.lenhard@imperial.ac.uk
Tel: +44-(0)20-83838353
Project context and objectives:
The main objectives of the project are:
1. Establish a database of the X chromosome containing all highly conserved non-coding elements (HCNEs) on the human X chromosome, with accompanying tools for analysis and visualisation (work package (WP)1) that identify HCNEs in relation to their likely target genes through:
(i) HCNE density plots across the entire X chromosome and
(ii) through mapping of all conserved synteny blocks between the human X chromosome and distant vertebrate genomes: five teleosts, chicken, frog and opossum, but also using multiple mammalian genomes.
These were surveyed for breakpoints, chromosomal aberrations, or point mutations causing neurological disease. Each conserved synteny block will contain one or more genes and will allow the identification of HCNEs and their target genes.
2. Generate a cis-regulatory map of the human X chromosome (WP2) by extracting conserved non-coding sequences from all gene neighborhoods. These conserved non-coding human sequences were assayed in a high throughput approach in transgenic zebrafish. We used known loci for XLMR and mapped all elements around these loci and extract the regulatory content of 440 of these. The results from this screen, in the form of images of expression patterns, are stored in the database and are used by all beneficiaries to evaluate whether they are candidates for further investigation of brain diseases.
3. Visualise neuronal patterns of expression driven by human HCNEs (WP3). We analysed elements that yielded brain or neuronal-specific expression by imaging at the cellular level in embryonic, larval, and adult zebrafish brains. This will be used to evaluate whether the elements are candidates for testing in human patient DNA for deletions, rearrangements, and mutations.
4. Identify human elements mutated or deleted / duplicated in X-linked neurological disease using arrays and sequencing of patient DNA (WP4). Finally, we sequenced all detected HCNEs plus all coding exons (about 13 Mb, or 8 % of the X chromosome) in 48 families that had no known coding mutations. These approaches will also boost the understanding of loci that are already mapped, but where the neurobiological basis for development of disease is currently not understood.
Project results:
Work performed so far and results obtained
1. The development and application of computational method to identify putative enhancers and their target genes on the human X chromosome (WP1)
Partner 5 has identified 102 647 putative enhancers on the X chromosome and has predicted a unique target gene for 38 591 of them, of which 12 592 target a known XLID gene. A bioinformatic method that identifies cases of evolutionary co-segregation between these and specific genes was developed. The underlying working hypothesis is that a rearrangement occurring between an enhancer and its target gene would separate the two elements, causing the target gene to be misregulated. The rearrangement would likely be counter-selected during evolution, leading to a high degree of conservation of the functional enhancer / gene pairs. When comparing modern genomes, this would translate to of a high level of co-segregation, where the enhancer and its target gene would be found in the vicinity of each other in most or all genomes where the two sequences are presents. This is the underlying phenomenon thought to explain the observation of genomic regulatory blocks (GRBs) within vertebrate genomes.
2. The management, integration, and dissemination of results generated during the project
WP1 has generated several web-based tools that allow the analysis of the evolutionary history of individual GRBs, namely Ancora and Genomicus. WP1 has also established a genome browser based on the UCSC browser that incorporates data such as HCNEs (WP1), tested enhancers (WP2) and ChIP data (WP4).
3. Generation of a cis-regulatory map of the human X chromosome by extracting conserved non-coding elements from known disease-gene loci and assay in in transgenic zebrafish (WP2)
The work plan aimed to test 400 sequence elements in total. Two hundred elements by partner 6 and 200 elements by partner 7. In total, partner 6 cloned and functionally analysed 218 (corresponding to 220 Kb) highly conserved sequences located in the vicinity of genes involved in human neurological disorders. Thirty one sequence elements where identified to drive tissue specific expression of the GFP reporter. Twenty four of these elements directed GFP expression to neural tissues (11 %). Three elements showed muscle specific expression, 2 elements where expressed in the notochord and two in the eye. Partner 7 has tested a total of 223 HCNEs with an average length of 1531 bp, corresponding to a total of 340 Kb (0.21 %) of the human X chromosome. Of these, 33 (15 %) acted as reproducible enhancers in the brain (in 3 or more independent insertions each), and 72 (32 %) were found to be partially reproducible in the brain. The remainder were either reproducible, but not in brain (21.9 %), or partially reproducible outside of the brain or inconsistent (43.5 %).
4. Visualise neuronal patterns of expression driven by critical human HCNEs at the cellular level in embryonic, larval and adult zebrafish brains (WP3)
Partner 8 and 3 have characterised 25 transgenic lines at the juvenile stage and 20 lines at the adult stage, and have also established 20 gene expression patterns of zebrafish orthologues of human X chromosome genes. The results establish that not all enhancers are active in the adult brain, but some were found to correspond to regions containing adult neural stem cells. A substantial fraction, however, did not show overlap with the endogenous expression of the supposed target gene while yet others exhibited ectopic expression when compared to the endogenous expression pattern.
5. Identify human elements mutated or deleted / duplicated in X-linked neurological disease using arrays and sequencing of patient DNA (WP4)
Partner 2 has used Chromatin immunoprecipitation (ChIP) with antibodies against H3K4me1, p300 (EP300), CBP (CREBBP) and CTCF to identify potential cis-regulatory elements in chromatin isolated from both human foetal brain and E14.5 E16.5 and P0 developing mouse brain. The resulting DNA was hybridised to oligonucleotide tiling arrays for both the mouse and human X-chromosome. There was a high degree of reproducibility on both technical and biological replicates when either cDNA or immunoprecipitated DNA was hybridised to the oligonucleotide arrays. Following advice from the SAC RNA-seq was used to define the transcriptome during human fetal brain development. This process detected all the known early brain enhancers and has revealed many thousands of novel putative enhancers that are apparently active in developing brain tissue. These data are available on as UCSC custom tracks at http://genome-neuroxsys.genereg.net Further, WP4 analysed over 300 patient samples with Cornelia de Lange syndrome and eye malformations for potential deletions, but none were identified. WP4 also started sequencing of approximately 16Mb encompassing non-coding sequence (from WP1) and all exons in 48 patients with plausible XLID with no known coding mutations. The coverage was approximately 80-fold. Initial analysis of this sequencing has revealed plausible coding region mutations in 5 / 48 cases that had been missed in the previous analysis. All other cases show interesting mutations in highly conserved non-coding elements. Currently, partner 2 is following up eight mutations in elements that appear to control known XLID genes.
Potential impact:
Expected final results, impacts, and use
We expect that the final results of the NEUROXSYS consortium will significantly expand the body of information on XLID as well as other X-linked neurological diseases, and will discover hitherto unknown disease mechanisms in both coding and non-coding sequence. These results should also have a direct impact on families carrying these unknown mutations, as new diagnostic tools become available. As a result, the X chromosome will represent the most advanced arena of human neurological disease-causing mutations, both in terms of coding and non-coding mutations.
NEUROXSYS online tools and database
NEUROXSYS has created a wealth of online tools that help with investigation of mutations on the X chromosome, and the annotation of potential target genes to a large number of conserved non-coding, and potentially regulatory, elements. Further, a total of 560 Kb of human sequence have been tested in transgenic zebrafish, which should aid in understanding the non-coding mutations currently evaluated by NEUROXSYS as well as future mutations on the X chromosome.
The identification of non-coding sequence variants that segregate with XLID will lead to advanced understanding of non-coding variants in monogenic disease. This has been made possible through the large-scale bioinformatic work on non-coding, and regulatory, sequence on the human X chromosome by the NEUROXSYS consortium.
Pathogenesis of X-linked neurological disease
Although we have not yet been able to successfully test mutated and disease-linked non-coding sequence in transgenic zebrafish to evaluate potential differences between normal and mutated sequence, this will be possible with new variants as they are identified. In such cases, we should be able to observe in the zebrafish the affected brain areas and the underlying mechanism.
Project website: http://neuroxsys.net/
Dr. Boris Lenhard
E-mail: b.lenhard@imperial.ac.uk
Tel: +44-(0)20-83838353