Mutation analysis of multiple pilomatricomas in a patient with myotonic dystrophy type 1 suggests a DM1-associated hypermutation phenotype

Myotonic dystrophy type 1 (DM1) is an inherited neuromuscular disease which results from an expansion of repetitive DNA elements within the 3' untranslated region of the DMPK gene. Some patients develop multiple pilomatricomas as well as malignant tumors in other tissues. Mutations of the catenin-β gene (CTNNB1) could be demonstrated in most non-syndromic pilomatricomas. In order to gain insight into the molecular mechanisms which might be responsible for the occurrence of multiple pilomatricomas and cancers in patients with DM1, we have sequenced the CTNNB1 gene of four pilomatricomas and of one pilomatrical carcinoma which developed in one patient with molecularly proven DM1 within 4 years. We further analyzed the pilomatrical tumors for microsatellite instability as well as by NGS for mutations in 161 cancer-associated genes. Somatic and independent point-mutations were detected at typical hotspot regions of CTNNB1 (S33C, S33F, G34V, T41I) while one mutation within CTNNB1 represented a duplication mutation (G34dup.). Pilomatricoma samples were analyzed for microsatellite instability and expression of mismatch repair proteins but no mutated microsatellites could be detected and expression of mismatch repair proteins MLH1, MSH2, MSH6, PMS2 was not perturbed. NGS analysis only revealed one heterozygous germline mutation c.8494C>T; p.(Arg2832Cys) within the ataxia telangiectasia mutated gene (ATM) which remained heterozygous in the pilomatrical tumors. The detection of different somatic mutations in different pilomatricomas and in the pilomatrical carcinoma as well as the observation that the patient developed multiple pilomatricomas and one pilomatrical carcinoma over a short time period strongly suggest that the patient displays a hypermutation phenotype. This hypermutability seems to be tissue and gene restricted. Simultaneous transcription of the mutated DMPK gene and the CTNNB1 gene in cycling hair follicles might constitute an explanation for the observed tissue and gene specificity of hypermutability observed in DM1 patients. Elucidation of putative mechanisms responsible for hypermutability in DM1 patients requires further research.

In order to share the raw NGS-data we have uploaded the data under: SRA: https://www.ncbi.nlm.nih.gov/sra SubmissionID: SUB6884814 BioProject ID: PRJNA603431 and included this information in the revised manuscript.
2) Reviewer #1: (i) The authors should tone down their conclusions, since samples are obtained from only individual, and focus more on presenting their results rather than confronting previous ideas and or hypothesis.
We agree to the reviewer and have soften author summary, introduction and discussion to this regard.
3) Reviewer #1: (ii) Authors should check whether Wnt pathway is activated in the different tumors checking expression of B-catenin and APC by immunohistochemistry and downstream targets such c-myc or cyclin D among others.
We have introduced immunohistochemistry for expression of CTNNB1, c-myc and cyclin D1 and could demonstrate that CTNNB1 is highly Ruben Artero, Ph.D. Academic Editor PLOS ONE 2/4 expressed in the pilomatricoma which further supports a causative role of CTNNB1 mutations. This has already been described before in nonsyndromic pilomatricoma. Expression of c-myc and cyclin D could be detected as well, but at a lower level. Nevertheless, lower expression of c-myc and cyclin D is consistent with pilomatricoma being a benign and very slowly growing tumor. Absence of expression of c-myc and cyclin D in shadow cells which do not express CTNNB1 further supports the assumption that WNT-CTNNB1-pathway plays a crucial role in pilomatricoma. 4) Reviewer #1: (iii) Results from NGS only found a mutation in ATM gene. I would have expected higher percentage. Do the authors have an explanation for this low number.

Indeed, this is a low number of detected mutations, however, in our experience this is not a totally unexpected finding for a panel that encompasses only about 0.4 Mb cumulative target size. The finding is also in line with published data on large tumor mutational burden studies. E.g. Chalmers et al. analyzed 100.000 human cancer genomes and showed that several cancer types show less than 1 mutation/Mb 1 .
Moreover, our assumption is that there might be a transcriptional bias which favors mutation acquisition in CTNNB1 but does not lead to a high mutation load in the whole genome. This would certainly be a very interesting scientific question that needs to be answered. However, to the best of our knowledge, no NGS data on pilomatricomas have been published to date. 6) Reviewer #2: The CTNNB1 mutations the authors have identified in the cancers of the myotonic dystrophy patients have been shown to sensitize tumor cells to TTK kinase inhibitors (Mol. Cancer Ther. 16 (11) 3/4 2609-17). Do the authors believe that treatment with TTK inhibitors could be a viable option for these patients?
The correlation between overexpression of the assembly checkpoint kinase TTK (Mps1) and CTNNB1 mutations is not understood. It could be that Mps1 overexpression favors the occurrence of CTNNB1 mutations or that CTNNB1 overexpression by activating mutations enhances Mps1 expression. In the first case, we would not expect a role of TTK inhibitors in treating tumors of DM1 patients as we assume that CTNNB1 mutations are due to simultaneous transcription of CTNNB1 and the mutated DMPK gene. In the latter case, we could expect a role in treatment if the tumors of DM1 patients would acquire aneuploidy due to Mps1 overexpression, but this would only be the case in malignant tumors and not in pilomatricomas. We have not included this discussion in the revised manuscript. 7) Reviewer #2: Please provide more technical detail on how the relative fraction of tumor cells was determined in the samples that were sequenced in the Methods section. Please provide more details on the results of the sequence analysis, more specifically, the % of reads that were harboring the mutant vs. the wild-type CTNNB1 gene.
We have included the requested information in the manuscript. 8) Reviewer #3: I would like to know the clinical information of each tumor (size, location, etc.). Because multiple pilomatricoma in DM1 patients sometimes get larger than non-syndromic solitary pilomatricoma.
We have included the requested information in the manuscript. 9) Reviewer #3: Your hypothesis on interaction of toxic RNA from mutated DMPK gene is really interesting. I was just wondering if co-translation of DMPK gene and CTNNB1 gene has already been evidenced in the literature or not? Or is it just a hypothesis? I was not able to find any reference about that in this paper. This is just our hypothesis. We replaced co-translation by simultaneous transcription in the manuscript as the putative interaction should already take place at the stage of transcription as described in figure 3 (2). We did not use the term co-transcription as this term is also used for specific forms of post transcriptional modifications. Nevertheless, co-transcription in the sense of simultaneous transcription has been proposed as a mechanism for gene fusions: Wright RL and Vaughan AT. A systematic description of MLL fusion gene formation.Crit Rev Oncol Hematol 2014 -Review. PMID 24787275. We have included this reference.
With best regards Prof. Dr. med. Albert Rübben Head of the Euregio Skin Cancer Center