CAPITULO III. CONDICIONES DE PRE FACTIBILIDAD EN LA
3.3. Acciones programáticas previas a la municipalización de los ODS
3.3.3. Preparación del sistema de monitoreo y evaluación de los ODS
While there are twelve possible single-base changes (C>A; G>T; C>G; G>C; C>T; G>A; T>A; A>T; T>C; A>G; T>G; A>C), these effectively reduce to six types (C:G>A:T; C:G>G:C; C:G>T:A; T:A>A:T; T:A> C:G; T:A>G:C) as, given the double stranded nature of DNA, it is impossible to distinguish on which strand the mutation occurs. The nomenclature used to describe mutations varies considerably. Within this thesis, mutations will be described from the context of the purine base that is mutated, for example, a C:G>A:T mutation will be described as a C>A transversion, despite the fact that this necessarily also defines a G>T transversion.
Historically, the exploration of mutational spectra has been restricted to analysis of targeted sequencing of established cancer genes (DeMarini et al., 2001, Giglia- Mari and Sarasin, 2003, Schwemmle and Pfeifer, 2000). Using the catalogue of mutation documented in the tumour suppressor gene TP53, clear differences in the mutational spectra between lung carcinomas and other cancers were observed, with C>A transversions more prevalent in tumours from smokers (Schwemmle and Pfeifer, 2000). Furthermore, even at codons that are common hotspots across a variety of cancer types, a preponderance of C>A transversions in lung cancers of smokers was identified compared to other tumours and never-smokers. The C>A transversions derived from smokers’ tumours have been found to exhibit a strong transcriptional strand bias with fewer G>T transversions on the transcribed than the non-transcribed strand, likely reflecting the past activity of transcription-coupled nucleotide excision repair on bulky adducts of guanine caused by tobacco carcinogens (Hainaut et al., 2001). Ultraviolet (UV) light associated damage has been shown to induce C>T and CC>TT transitions in melanoma genomes, and also exhibited transcriptional strand bias, with fewer C>T mutations on the transcribed compared to the non-transcribed strand, probably due to the action of transcription-coupled repair on impaired pyrimidines (Pfeifer et al., 2005).
While informative, a limitation of these single gene studies is that they rely on mutations in driver genes, making it difficult to disentangle the effects of selection with the mutational signatures operating. Moreover, aggregate signals from
thousands of samples are required, meaning only strong exposures or dominant repair processes operating across the majority of tumours can be accurately deciphered.
1.6.1.1 Whole-genome and whole-exome analysis of mutational catalogues
Analyses of mutational catalogues obtained from whole-genome sequencing of tumours with high mutation rates can provide a detailed picture of the processes operating. For example, two studies involving sequencing of a malignant melanoma and a single lung cancer were some of the first to illustrate the power of this approach (Pleasance et al., 2010a; Pleasance et al., 2010b). These two studies identified the characteristic mutational spectra of ultraviolet light and tobacco carcinogens respectively within single tumours. Since these seminal studies, further analysis of a large series of sequenced lung tumours has revealed that smokers on average exhibit a 10-fold increase in the burden of somatic mutations in their cancer genomes compared to never-smokers (Govindan et al., 2012; Imielinski et al., 2012). Moreover, consistent with previous findings and experimental evidence, this elevation is mainly due to the increase in the number of C>A transversions.
In depth examination of the mutational spectra within cancers has also highlighted key differences in mutational processes within different tumours. For example, examination of hyper-mutated endometrial and colorectal tumours has revealed preponderance of C>A and C>G mutations at TpCpT sites specifically in tumours harbouring mutations in POLE-E (Palles et al., 2013), demonstrating somatic aberrations can result in the preponderance of a particular mutational signature.
More recently, algorithms have been developed to quantify the number and contributions of mutational signatures operating within cancers at the single- nucleotide level (Alexandrov et al., 2013b, Fischer et al., 2013). These approaches assume recurrent processes operate across different cancers, and make use of mathematical frameworks to de-convolve the different processes operating within each cancer. Application of NMF (non-negative matrix factorization) and model
selection to over 30 cancer types, represented by more than 7000 tumours, identified 20 distinct mutational signatures (Alexandrov et al., 2013a).
In the majority of cancer samples analysed, at least two mutational processes were identified, consistent with an elevated mutation rate in most cancers (Alexandrov et al., 2013a). The most widespread mutational signature, identified in 25 cancer types, was characterized by C>T transitions at CpG sites, probably reflecting deamination of 5-methylcytosines at CpG sites. This signature correlated with patient age (Alexandrov et al., 2013a), consistent with a large proportion of these mutations having been acquired prior to tumourigenesis.
Another pervasive mutational signature, identified in 15 cancer types, was characterized by C>T and C>G mutations at TpC sites. Orthogonal analysis, adopting a simpler method that used the prevalence of mutations with this motif compared to what would be expected given random mutagenesis, also identified the presence of this signature across multiple cancer types (Burns et al., 2013, Roberts et al., 2013). This signature has been linked to the family of apoliprotein B mRNA editing enzyme catalytic polypeptide-like (APOBEC) enzymes, involved in cytosine deamination. Cell line studies have demonstrated that APOBEC1, APOBEC3A, and APOBEC3B are capable of mutating DNA by the deamination of cytosine flanked by a 5’ thymine and thus result in C>T mutations at TpCpN tri- nucleotides (Harris et al., 2002, Hultquist et al., 2011, Suspene et al., 2011, Taylor et al., 2013), consistent with their role in facilitating tumour mutagenesis. Furthermore, it has been shown that the activation of enzymes APOBEC3A and APOBEC3B in yeast can also lead to in C>G at TpCpN tri-nucleotides (Taylor et al., 2013). This mutational pattern was attributed to replication over an abasic site, formed when an APOBEC deaminated cytosine is excised by uracil-DNA glycosylase, which is catalysed by REV1 (Taylor et al., 2013).
Certain mutational signatures were also identified exclusively in specific cancer types. For instance, in oesophageal carcinomas (including both adeno and squamous cell-carcinomas), a signature characterized by T>G and T>C mutations, particularly at CpTpT sites, was identified, which has been linked to gastric acid
exposure (Dulak et al., 2013, Weaver et al., 2014). Similarly, the previously characterized smoking signature, involving C>A transversions, was also identified in lung and head-and-neck cancers.
In combination these studies demonstrate that many of the mutational processes underlying cancer genome evolution are beginning to be elucidated. Importantly, however, the underlying processes responsible for many of the observed signatures still remain unclear, and in general their temporal decomposition is unknown.