Human cancer genomes show a complex pattern of mutations that, upon computational deconvolution, resolves into a systematic series of about 30 ?Mutational Signatures? (the pattern of mutations across all possible trinucleotide sequence contexts). Some signatures are common to many cancers (e.g., CG?TA in CpG sites) whereas others appear in single tumor types (e.g., signatures involving GC?TA mutations that are presumably associated with aflatoxin B1 (AFB1) exposure in hepatocellular carcinoma (HCC)). Current sequencing efforts sometimes provide hints, via examination of the details of mutational signatures, as to cancer mechanistic etiology. The present proposal is a bottom up approach to determine which specific chemical insults to DNA cause mutational patterns that match mutational signatures in end stage disease. Because the spectra of human cancers are very complex, part of our analysis is informed by a growing data set from this laboratory on a mouse model in which a single dose of AFB1 causes HCC over 72 weeks. High- fidelity duplex next-generation sequencing of the mouse genome as it progressed towards HCC revealed a distinctive pattern, dominated by GC?TA transversions. Examination of our AFB1 mutational spectra in 3- base contexts revealed that some 5'-NGN-3' contexts are hypermutated whereas others are not. The objective of this proposal is to use site-specifically modified genomes containing AFB1-DNA adducts in all possible 3- base contexts, replicated in living cells, to determine the mechanisms underlying the complex mutational spectrum of AFB1 in vivo. The adducts will be replicated in cells that differ in single variables, such as the type of repair system or polymerase present. The mutational signatures will be compared by cosine similarity to (a) our in-house murine AFB1-induced mutational spectra and (b) mutational signatures identified in human HCC. The work is critically enabled by the development of a unique benzofuran-containing template that provides synthetic access to AFB1 adducts in all sequence contexts required for this work. Several studies suggest that inflammation plays a role in the multistep process of liver carcinogenesis. A quintessential oxidative stress lesion, 7,8-dihydro-8-oxoguanine (8-oxoG), would be expected to induce a pattern of GC?TA mutations in DNA (the same type of mutation as AFB1). In order to distinguish 8-oxoG and AFB1 mutational contributions to cancer in inflamed tissues, we propose to define in parallel the 3-base context-specific mutational signature of 8-oxoG, in cells that are either wild type or altered in replication or repair status. Collectively, these studies will provide a granular view of distinctive steps in carcinogenesis. They also will provide an experimental framework for studies on the contributions of other lesions to cancer mutational signatures. Identification of discrete steps in carcinogenesis (e.g., the role of inflammation) may eventually inform precision prevention strategies (e.g., antioxidant therapy to suppress the 8-oxoG mutational signature).
Human cancer genome sequencing efforts have delivered an enormous amount of data on the genetic changes that occur during the process of carcinogenesis. Mathematical extraction of the complex data set reveals about 30 distinctive ?mutational signatures? that hint at, but do not definitively prove, the molecular events that convert normal cells into cancer cells. Utilizing unique chemical reagents, this proposal takes a bottom up approach to construct complex mutational signatures out of component parts, each of which is mechanistically well understood.
Showing the most recent 10 out of 38 publications
