Response

We would like to clarify several points mentioned in the letter from Cooke and explain why mutations in POT1 are important in causing glioma, specifically oligodendroglioma, in a subset of glioma families that we sequenced.

Glioma is a rare cancer (1), and familial glioma is even rarer, but it is well established that germline mutations can cause familial cancers. However, familial glioma shows genetic heterogeneity, meaning any particular allele will only be responsible for a small fraction of the disease. Thus, it is unlikely that any single, extremely rare allele can be associated with cause in familial glioma. Although single nucleotide polymorphisms have been associated with glioma, we sought to find rare, causative alleles in our familial cohort.

Exome sequencing of 55 families identified two individuals with rare mutations in the POT1 gene, one of which occurred in the functionally important OB region of the protein, and the other was truncating. These same mutations were found in the linking relatives indicating that the mutation was inherited from a previous generation (families A and B [2]). In family A, six members and one obligate carrier were found to harbor the mutation, whereas three of those individuals developed glioma. Similarly, in family B, six individuals carried the mutation and two developed glioma, which suggests incomplete penetrance. In both pedigrees, all patients with glioma and DNA available for sequencing had POT1 mutations. When we sequenced 264 additional familial glioma cases we found another case with a truncated POT1 variant. Truncating mutations are more deleterious than substitution mutations and considerably more rare (http://exac.broadinstitute.org/, accessed March 11, 2015). We established that mutant carriers exhibited statistically significantly higher telomere content when compared with the remainder of the cohort, demonstrating deleterious functional impact of the POT1 mutations on this known negative regulator of telomere length. Other reports also associated variants in telomere genes with increased telomere length and glioma incidence (3). In addition, mutations in POT1 and ACD, which encodes the POT1 binding partner, are associated with increased telomere length and familial melanoma (4–6), thus providing a strong association of mutations in these telomere maintenance genes and telomere length with melanoma, with which glioma has been epidemiologically linked.

As we stated, the association between familial glioma and protein-altering POT1 mutations is strong, but it still needs to be biologically validated by in vivo experiments, and more generalized use of POT1 in genetic testing will require additional replication. In our article we specifically referred only to the small subset of families with oligodendrogliomas or gliomas and melanomas with regard to clinical POT1 analysis. As is true of most genetic testing for cancer susceptibility, we recommend affected individuals within the family be tested first to see if POT1 mutations may be relevant to their cancer risk.

It is challenging to use a purely statistical approach when working on a disease that is both extremely rare and whose causative alleles are heterogeneous and incompletely penetrant. Thus, we must employ biological knowledge and previously described disease associations and mechanism, in addition to statistical analysis, to establish whether genes are associated with familial glioma.