• 2019-07
  • 2019-08
  • 2019-09
  • 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • br Discussion br Here we identified di


    4. Discussion
    Here, we identified 335 differentially expressed transcripts that re-flect the degree of genome instability in breast carcinomas. By fitting random forest models, these transcripts were further reduced to 17 genes responsible for the separation between genomically stable and unstable tumours. The 17-marker panel displayed a high predictive power for clinical outcome and outperformed the 12-gene signature for genomic instability and the clinically validated Oncotype Dx.
    Genome stability was defined using G2I on DNA copy number profiles that separated the cohort into genomically stable (G2I-1, G2I-2) and unstable tumours (G2I-3). The independent prognostic significance of G2I was in agreement with previous studies [17]. Genomic instability was associated with DNA-related clinical parameters (DNA index, S-phase fraction, and ploidy) as well as survival-related parameters (survival status, and follow-up time). These findings are in line with previous studies applying the stemline scatter index (SSI), which is based on the S-phase fraction among other factors, as an independent prognostic indicator of clinical outcome [8,9]. The correlation of genomic instability and the S-phase-fraction indicates that the statisti-cally significant differences in survival outcome are not solely based on genome stability but also overlap with the proliferative activity of  Genomics xxx (xxxx) xxx–xxx
    cancer cells. This is not unexpected, as genomic instability can produce complex aberrations that may lead to selective growth advantages, such as increased proliferation. The risk groups based on the linear predictor of the 17-marker panel were additionally associated with known prognostic factors, such as molecular subtype, lymph node ratio, ER, PR and HER2 status. This overlap with known prognostic factors indicated that one of the downstream effects of genomic instability is to benefit cancer Imipenem with increased proliferative activity. The pathways asso-ciated with the 17-marker panel proposed components of chromosomal organisation, such as kinetochores, centromeric regions etc., as the source of increased proliferation.
    There was overlap between a number of the 335 significantly dif-ferentially expressed transcripts and other gene signatures for genomic instability. The CIN70 signature overlapped with 12 genes (AURKA, CCNB1, CDCA3, CEP55, MAD2L1, PCNA, RAD21, RAD51AP1, RFC4, TPX2, TRIP13, ZWILCH) [12], while the 12-gene signature showed one overlap (AURKA) [11]. The CINSARC signature overlapped with 17 genes (ANLN, AURKA, BUB1B, CCNB1, CDCA3, CENPA, CENPE, CEP55, KIF11, KIF14, KIF18A, MAD2L1, NUF2, RAD51AP1, SPAG5, TRIP13, TPX2) [14], the CUX1-signature in five genes (BUB3, CENPA, CENPE, MAD2L1, ZWILCH) [15], and the Oncotype Dx signature overlapped in two genes (AURKA, CCNB1). The 17-marker panel, however, had only one gene in common with the CIN70 signature (CDCA3) and three genes with the CUX1-signautre (BUB1B, CDCA3, KIF14). Conclusively, different approaches of biomarker discovery resulted in a relatively constant set of genes that appears to govern genomic instability. Hence, the underlying biological processes (mostly chromosome organisation, cell cycle regulation, and RNA processing) and the molecular functions (mainly DNA and RNA binding) of the 335 transcripts offer options for drug development and treatment tailoring.
    Possible causes of genomic instability include defects in DNA da-mage repair, DNA replication, chromosome segregation, chromosome cohesion, and telomere stability, which is in line with the pathways associated with the 17 markers [35]. Furthermore, incorrect repair of DNA double-strand breaks (DSBs) may cause structural variations, such as chromosome translocations, inversions, duplications, and deletions, which can result in the activation of oncogenes and inactivation of tumour suppressor genes [35]. Overexpression of mitotic checkpoint genes (e.g. BUB1B, CDCA3, SPDL1) can lead to checkpoint hyper-activation and prolonged mitosis causing “mitotic slippage” [12,35,36]. Due to the strong association of the 17 markers with centromere- and kinetochore-related pathways, centromere integrity forms a major compound of this signature. Centromeres orchestrate several compo-nents needed for correct chromosome separation and segregation. Therefore, an intrinsic fragility of the centromeres may lead to DNA breaks and abnormal chromosomal rearrangements in cancer cells [37]. Furthermore, the formation of neocentromeres could facilitate to maintain chromosome segregation in cancer cells and also serve as a potential source of genomic rearrangements [37].