Oncogenomics Group, QIMR Berghofer Medical Research Institute, Brisbane, Australia.
PLoS One. 2018 Apr 11;13(4):e0194098. doi: 10.1371/journal.pone.0194098. eCollection 2018.
While a number of autosomal dominant and autosomal recessive cancer syndromes have an associated spectrum of cancers, the prevalence and variety of cancer predisposition mutations in patients with multiple primary cancers have not been extensively investigated. An understanding of the variants predisposing to more than one cancer type could improve patient care, including screening and genetic counselling, as well as advancing the understanding of tumour development.
A cohort of 57 patients ascertained due to their cutaneous melanoma (CM) diagnosis and with a history of two or more additional non-cutaneous independent primary cancer types were recruited for this study. Patient blood samples were assessed by whole exome or whole genome sequencing. We focussed on variants in 525 pre-selected genes, including 65 autosomal dominant and 31 autosomal recessive cancer predisposition genes, 116 genes involved in the DNA repair pathway, and 313 commonly somatically mutated in cancer. The same genes were analysed in exome sequence data from 1358 control individuals collected as part of non-cancer studies (UK10K). The identified variants were classified for pathogenicity using online databases, literature and in silico prediction tools.
No known pathogenic autosomal dominant or previously described compound heterozygous mutations in autosomal recessive genes were observed in the multiple cancer cohort. Variants typically found somatically in haematological malignancies (in JAK1, JAK2, SF3B1, SRSF2, TET2 and TYK2) were present in lymphocyte DNA of patients with multiple primary cancers, all of whom had a history of haematological malignancy and cutaneous melanoma, as well as colorectal cancer and/or prostate cancer. Other potentially pathogenic variants were discovered in BUB1B, POLE2, ROS1 and DNMT3A. Compared to controls, multiple cancer cases had significantly more likely damaging mutations (nonsense, frameshift ins/del) in tumour suppressor and tyrosine kinase genes and higher overall burden of mutations in all cancer genes.
We identified several pathogenic variants that likely predispose to at least one of the tumours in patients with multiple cancers. We additionally present evidence that there may be a higher burden of variants of unknown significance in 'cancer genes' in patients with multiple cancer types. Further screens of this nature need to be carried out to build evidence to show if the cancers observed in these patients form part of a cancer spectrum associated with single germline variants in these genes, whether multiple layers of susceptibility exist (oligogenic or polygenic), or if the occurrence of multiple different cancers is due to random chance.
虽然许多常染色体显性和常染色体隐性癌症综合征都有相关的癌症谱，但患有多种原发性癌症的患者中癌症易感性突变的患病率和多样性尚未得到广泛研究。了解导致多种癌症的变异可能会改善患者的护理，包括筛查和遗传咨询，以及提高对肿瘤发展的认识。
本研究招募了一组 57 名因皮肤黑色素瘤 (CM) 诊断并具有两种或多种其他非皮肤独立原发性癌症病史的患者。对患者的血液样本进行全外显子或全基因组测序评估。我们重点研究了 525 个预先选择的基因中的变异，包括 65 个常染色体显性和 31 个常染色体隐性癌症易感性基因、116 个参与 DNA 修复途径的基因和 313 个在癌症中常见的体细胞突变基因。在作为非癌症研究 (UK10K) 一部分收集的 1358 名对照个体的外显子组序列数据中分析了相同的基因。使用在线数据库、文献和计算机预测工具对鉴定出的变异进行致病性分类。
在多癌队列中未观察到已知的常染色体显性或先前描述的常染色体隐性基因复合杂合突变。通常在血液恶性肿瘤中发现的体细胞变异 (JAK1、JAK2、SF3B1、SRSF2、TET2 和 TYK2) 存在于患有多种原发性癌症的淋巴细胞 DNA 中，所有这些患者都有血液恶性肿瘤和皮肤黑色素瘤病史，以及结直肠癌和/或前列腺癌。在 BUB1B、POLE2、ROS1 和 DNMT3A 中发现了其他潜在的致病性变异。与对照组相比，多例癌症患者的肿瘤抑制基因和酪氨酸激酶基因中更有可能发生致病变异 (无义、移码插入/缺失)，并且所有癌症基因的突变总负担更高。
我们鉴定出了一些可能导致至少一种肿瘤的致病性变异，这些变异存在于患有多种癌症的患者中。我们还提供了证据，表明患有多种癌症类型的患者的“癌症基因”中可能存在更高的未知意义变异负担。需要进一步进行此类筛查，以建立证据，表明这些患者中观察到的癌症是否属于与这些基因中的单个种系变异相关的癌症谱的一部分，是否存在多层易感性 (寡基因或多基因)，或者多种不同癌症的发生是否是偶然的。
