A survey of differences between cancer genomes across patients’ age

Age is the most important risk factor for cancer. However, it is unclear how tumours differ among patients of different age. In this work, we comprehensively characterised age-associated molecular landscape to identify molecular alterations that are more or less common in younger or older patients.

Like Comment
Read the paper

             Age is the single most important risk factor for numerous chronic diseases including most types of cancer. The majority of all cancer cases are diagnosed in patients older than 60 years old. Not only cancer incidence, but the cancer mortality rate also increases exponentially with age. As we are now entering an ageing society, there is an immediate need to understand the relationship between age and cancer. In fact, apart from the raising in cancer incidence and mortality as a function of age, studies have suggested differences in characteristics and treatment outcomes between tumours from patients of different ages. For instance, breast cancer diagnosed in younger patients tend to be more aggressive and these patients have poorer survival than older patients. Indeed, the molecular characteristics, such as mutations, of tumours from younger and older patients often are different. In other words, certain genetic mutations may be found in tumours from younger patients, but rarely be found in the elderly and vice versa. This age-associated difference is evident in breast cancer and several other cancer types, including prostate cancer and gliomas. However, most studies on this topic have focused on single cancer types. Thus, a comprehensive pan-cancer study on the age-associated cancer molecular landscape is lacking.

             In our recent paper1, we have asked a simple question: What are the differences between tumours across patients’ age? To answer this, we employed large-scale data from The Cancer Genome Atlas (TCGA) project to comprehensively characterise age-associated molecular alterations across cancer types. We began by showing that genomic instability levels increased with age in pan-cancer and in several cancer types. In a similar manner, we confirmed previous findings that somatic copy number alteration (SCNA) level and point mutation burden also increased as a function of age. These results are consistent with the fact that, with ageing, spontaneous mutations and environmental exposure could damage the genome and result in genomic instability.

            While SCNA level increased with age in several cancer types, an opposite pattern was found in lung adenocarcinoma. It is likely that the younger patients contain a higher proportion of smoking patients, which leads to an increased SCNA level in younger groups. For somatic mutations, contrary to most cancer types, we found a negative association between mutation burden and age in endometrial cancer, which is related to the presence of hypermutation subtypes of endometrial cancer in younger patients.

             We further explored SCNA events that are presented more often in younger or older patients. Several chromosomal arms, harbouring known oncogenes and tumour suppressor genes, showed an age-related pattern in terms of chromosomal gains and losses. These age-related differences in SCNAs were prominent in gliomas, endometrial and ovarian cancers. Likewise, we also found age-related patterns of somatic point mutations in several cancer driver genes, particularly in gliomas. Indeed, some of these age-associated events could be explained by the presence of age-related subtypes. For instance, IDH-mutant subtype of gliomas associated with IDH1, TP53 and ATRX mutations, occurs more frequently in younger patients. Thus, age-associated subtypes of cancers might be important to study further.

             Finally, while gene expression and DNA methylation changes during normal ageing are well studied, it is not clear how gene expression differs as a function of age in cancer, and whether DNA methylation plays a role in controlling this age-related gene expression in cancer. Cancers that have the highest number of age-related genes in both expression and DNA methylation included low-grade glioma and several female reproductive system cancers like breast, ovarian, and endometrial cancers. A strong opposite relationship of gene expression and DNA methylation suggests that age-associated gene expression changes in cancer are repressed, at least in part, by DNA methylation. Furthermore, functional enrichment analysis reveals numerous biological processes that are linked to age-related gene expression and DNA methylation in cancer, including immune-related processes.

             In summary, we performed a comprehensive analysis of the impact of age on genomic, gene expression and DNA methylation landscapes across cancer types. Our results highlight age as an important factor to include in cancer research. Our paper is concurrent with the paper by Li et al.2, who systematically studied the impact of age on the cancer genome, focusing on mutations and copy-number differences. Indeed, results from the two studies are consistent in several respects. Moreover, Li et al. have explored mutational timing and signatures, which suggested possible underlying mechanisms for age-associated genomic differences.

              An obvious question following our study is how and why these age-associated patterns emerge. In other words, why some genomic alterations occur preferentially in younger or older patients. Accumulating evidence suggests the importance of the ageing microenvironment in cancer initiation and progression. We therefore hypothesise that cancer cells harbouring a particular mutation/SCNA may “fit” better in a particular microenvironmental context. As tissue microenvironment changes during ageing, cancer cells containing “fitted” mutation/SCNA might be able to grow and develop into a tumour. Our findings thus open avenues for future research to better understand the intimate relationship between ageing and cancer, which could have an implication in the development of the strategies to treat patients across different age groups to eventually achieve the goal of personalised medicine.  

Link to the paper

1          Chatsirisupachai, K., Lesluyes, T., Paraoan, L., Van Loo, P. & de Magalhães, J. P. An integrative analysis of the age-associated multi-omic landscape across cancers. Nature Communications 12, 2345, doi:10.1038/s41467-021-22560-y (2021).

2          Li, C. H., Haider, S. & Boutros, P. C. Age Influences on the Molecular Presentation of Tumours. bioRxiv, 2020.2007.2007.192237, doi:10.1101/2020.07.07.192237 (2021).

Kasit Chatsirisupachai

PhD Student, University of Liverpool