Connecting genomic sequences with clinical phenotypes can help identify mutations associated with rare conditions and will become particularly important for finding the mutations underlying syndromes associated with clinically heterogeneous characteristics. Bohring-Opitz syndrome, which is characterized by severe developmental delays, feeding difficulties, and distinctive abnormal facial features and posture, is such a condition and has previously been linked with mutations in the additional sex combs like 1 (ASXL1) gene, one of three evolutionarily conserved transcriptional regulators in the ASXL gene family.
In a study recently published in Genome Medicine, and covered in a Nature News article, Matthew Bainbridge and Jim Lupski from the Baylor College of Medicine, USA, alongside Hilger Ropers from the Max Planck Institute for Molecular Genetic, Germany, and colleagues used whole genome and whole exome sequencing to identify truncating de novo mutations in the ASXL3 gene in four young patients. The children shared clinical characteristics with each other and with Bohring-Opitz syndrome, but these features are relatively nonspecific, making it difficult to diagnose them without genetic data. We asked Bainbridge to explain further how the study came about.
What is your current line of research?
A great deal of my research and the work done at the Human Genome Sequencing Center at Baylor College of Medicine (BCM) focuses on associating genes with diseases. We do this on a research basis in collaboration with Johns Hopkins (Baylor-Hopkins Center for Mendelian Genomics) and on a clinical basis with the Whole Genome Lab at BCM.
In this particular case we identified a gene, ASXL3, that causes a very severe congenital genetic disorder resembling Bohring-Opitz syndrome. Children with highly deleterious (truncating) de novo mutations in this gene have extreme intellectual disability, motor skill problems, acute feeding problems and increased mortality.
One aspect that distinguishes our recent study from similar ones is that we found collaborators across the US and in Europe with similarly affected children on the basis of the children’s genetic signature, and not on the basis of their phenotype, as was done traditionally. This was critical because this disease is so rare that it is unlikely that a single institution would encounter multiple affected children. Further, the clinical presentation of the disease is not specific enough that you could establish that these children have the same disease without genetics.
What surprised you most when you started looking at the ASXL3 patient data?
The first surprise was how obvious the mutation was once we had the parental data. We had struggled for a few months using only the genetic data from the child we had started with and turned up a few false leads. When we added in the parental data it was possible to zero-in on the few spontaneous mutations in the child.
The next surprise was that the mutation occurred in a gene that was related to another well-known disease-causing gene – ASXL1. Mutations in the ASXL1 gene had been discovered by my good colleague Alex Hoischen (Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands) to cause Bohring-Opitz syndrome, and he was subsequently instrumental in finding other children with this disease.
So, ASXL1 and ASXL3 are part of the same gene family and cause diseases with some overlapping characteristics. From an evolutionary point of view, this makes a great deal of sense! Two genes that are evolutionarily related are probably functionally related, and defects in these genes could cause similar disease. This type of logic (that mutations in related genes can cause related phenotypes) doesn’t always work in the opposite direction (which would be that similar diseases are caused by mutations in evolutionarily related genes), and such relationships can only really be uncovered using genetic data.
Who is going to be interested in, or affected by, this research?
Firstly, this study (along with others) should help convince clinicians, medical researchers and politicians of the value of ‘ubiquitous clinical sequencing’. This study was only possible because of the low cost of sequencing and the availability of clinical sequencing in USA. In many situations physicians are either unaware of clinical sequencing or unwilling to consider its application. In many places, including Europe, it is not widely available. If it were, making these kinds of diagnoses would be common place.
Secondly, there are almost certainly many more children in the world affected by conditions similar to Bohring-Opitz syndrome. Clinicians with patients who have a similar phenotype have now another gene to consider, and these patients can now get a diagnosis. It is psychologically very important for parents to be able to know once and for all what the causes are for their child’s symptoms. It will also allow researchers to concentrate their efforts on potential therapies for this disease to the known functions of ASXL3.
How was the study received by the scientific community?
Clinicians are always very excited to be able to make more diagnoses on their patients and researchers have yet another example as to why clinical sequencing is important and should be funded. This work also helps illustrate the importance of collaborations in science: this work would not have been possible without Hilger Ropers (Max Planck Institute) and Karen Gripp (A.I. duPont Hospital for Children Wilmington, USA).
Your study is somewhat unusual because it defines a syndrome starting at the molecular rather than phenotypic level. Is this likely to become a general trend?
This will absolutely become more common as sequencing becomes less expensive and more ubiquitous. The data are much easier to share, process and query with computers than phenotypic data, and they enable collaborations between institutions even when they are physically distant from one another. Further, phenotypes can be very heterogeneous or non-specific. Children with mutations in ASXL3 have similar presentation to children with a host of other diseases, and these children can have a range of disease severity that makes it challenging to find ‘clinical overlap’. Indeed, we are discovering in the clinic that many of the cases we sequence with a ‘new’ disease are actually patient with mutations in a known disease gene, but they present so atypically that the diagnosis was not possible by phenotype alone!
Genomics as a field has traditionally been very pro ‘open data’. What are your thoughts on data sharing in the clinical genomics world?
I think science, in general (or at least, ideally), tries to be very open with data. This is in contrast with medical professionals who typically place patient confidentially above all else. Where these two fields meet there is conflict. The value of sharing large amounts of genomic data is obvious: we will be able to catalogue and diagnose all, or virtually all, genetic diseases with something akin to, and as simple as, a Google search. As a society we will decide on whether privacy trumps these advantages. Personally, as someone who has been fortunate enough to meet the parents of many of the children whose data we analyzed, my answer echoes theirs: absolutely anyone, anywhere, who can help these children can have access to the data. Allowing people to suffer for a personal, abstract fear of what might happen if some entity had access to my genomic data is utterly ridiculous. These data should be open and free for everyone.