This is a guest post by Dina Zielinski (@dinazielinski)
Of all the projects in the lab so far, only one has been ascribed a theme song. (At the risk of leaving you with an earworm for the rest of this post, it’s the theme from the movie ‘Rocky’). It’s not that our findings were particularly epic, rather there was more to the story than we could share, even in the supplemental text.
One of my first undertakings involved a familial case of the second most common facial anomaly after cleft lip and palate. The disorder, hemifacial microsomia (HFM) is characterized by highly variable craniofacial malformations, incomplete penetrance, and most cases are sporadic. However, we identified an Ashkenazi Jewish pedigree in Israel with five generations affected by the disorder (the largest pedigree to date), so we were (cautiously) optimistic.
Exome dreams
We started with exome sequencing of the grandmother and two grandchildren. Our group had success in two prior studies with this approach: Hereditary Spastic Paraparesis in Palestinians and Joubert Syndrome in Ashkenazi Jews. But these cases were recessive mutations in consanguineous families that allowed homozygous mapping of the critical regions. This time was definitely going to be more of a challenge.
Exome sequencing and subsequent filtering identified 4 point mutations shared between the grandmother and her 2 grandchildren. These mutations were not in public databases like dbSNP, several hundred Ashkenazi exomes, and were predicted to be deleterious. We also validated these mutations with Sanger sequencing. One particular mutation seemed quite promising: IQSEQ1. There is a GWAS hit in this gene associated with intracranial aneurysms and previous studies suggested that HFM is related to hemorrhage that can disrupt craniofacial development.
Being skeptical geneticists, we were not completely convinced. Soon after we completed the exome analysis we learned that another relative (a cousin of the grandmother) showed features of HFM and was living outside Philadelphia. Communicating with this guy in his 80s who speaks only Russian was luckily only a brief setback. With help from some fellow researchers at the Whitehead Institute (the great Igor Ulitsky and several members of the Page lab), we recruited him to the study.
We first tried sending him some buccal swabs but summer heat and delayed shipping do not make for good quality genomic DNA.
We decided Philly is not that far from Boston (well, considering the rest of the family was in Israel). So I grabbed the lab cooler and some backup DNA collection kits and was on my way. We hired a certified phlebotomist and together met the family. The cousin only spoke Russian so I mostly chatted with his lovely wife who showed me pictures of her happy, healthy grandchildren (don’t worry: informed consent was in Russian). I headed back to Cambridge with the cooler in the back seat and extracted the DNA as soon as I got to the lab. There was no way I wasn’t getting plenty of quality DNA after that adventure.
Hitting a wall
The Philly trip would end up being a blessing in disguise. The next day I Sanger sequenced the newly prepped DNA to check the 4 candidate mutations. The results were a bit discouraging after my first foray into genetics field work (the source of the whole ‘Rocky’ theme in case you haven’t put it together). He was homozygous for the reference allele for all mutations. We sequenced another ~30 candidate mutations that could be pathogenic if we reduced the filtering stringency but he was homozygous reference for all of them. We felt a bit cornered at this point (I’ll refrain from boxing references, but you get the idea).
Even more disheartening, we soon learned that the proband was diagnosed with medulloblastoma. HFM is quite rare and it was not clear if there was a connection between the two (see the case report by our collaborator).
The eye of the tiger
True to the theme song, we rose up to the challenge, even more determined to find the causal mutation. We had already done genome-wide genotyping of the grandmother and two grandchildren as part of the exome sequencing validation and were still working on the analysis when we decided to include the grandmother’s cousin and investigate copy number variants. The analysis (joint calling with genotype data from healthy Ashkenazi controls) resulted in a candidate list of 8 CNVs. All but 2 were also found in the controls. Of the 2 candidates, 1 overlapped a noncoding region and was also present in healthy Asian controls. We were left with a more than 1 Mb duplication of chromosome 14q22.3. We also validated this CNV by qPCR and it was absent in 45 additional Ashkenazi controls.
The duplication harbors 8 annotated genes so we still had some work to do. After using a ranking tool to find the 3 most clinically similar disorders, we compared the 3 respective disease genes to the biological signatures of all genes in the duplicated region. We also analyzed gene expression data in mouse embryos and compared dosage sensitivity of 3 of the duplicated genes. OTX2 was the top hit in all approaches.
OTX2 encodes a transcription factor known to play a critical role in craniofacial development. We were (finally) encouraged and convinced by our findings. Even more interesting was the fact that OTX2 is one of the most commonly amplified genes in pediatric medulloblastoma. Although present in only one case, this raises the possibility of a similar molecular basis between craniofacial anomalies and this type of cancer.
Truly unique to human genetic studies, we were able to video chat with the proband and her mother to explain our findings. I told them how nice it was to ‘meet’ them and to thank them for all their help throughout the study (mostly implied by my limited Hebrew). It was inspiring to see how interested and appreciative they were, a feeling certainly unmatched by publication.
For the full analysis:
http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0096788