Archives

  • 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • br Interestingly we observed strong significant associations

    2020-08-18


    Interestingly, we observed strong significant associations of multiple OCM genes with expression of the BRCA1 gene, which is involved in homologous recombination repair (Table 3; Supplementary Table 4). These correlations were positive for DHFR (Pearson r = 0.608), TYMS (r = 0.485), GART (r = 0.427), MTHFD1 (r = 0.408), SHMT1 (r = 0.332), ATIC (r = 0.331), and SLC19A1 (r = 0.318) and mod-erately negative for NNMT (r = −0.352; FDR adjusted p ≤ 4.17 × 10−24 for all genes listed). Positive correlations of many OCM genes with BRCA1 expression are consistent with a previous report that cell line treatment with folic Phorbol 12-myristate 13-acetate resulted in an increased BRCA1 gene expression in many cell lines; however, that study found that this association had no effect on DNA repair or that such effects were transient [101]. The biological impact of positive correlations between expres-sion levels of OCM genes and increased BRCA1 expression on drug sensitivity is unclear, because elevated expression of several OCM genes in the GDSC-CCLE dataset also showed very weak ( r < 0.3) but statistically significant (FDR adjusted p < 0.05) associations with sensitivity to multiple PARP in-hibitors. For example, TYMS had r = −0.273 for correlation with log(IC50) of olaparib and r = −0.261 with ABT-888 (veli-parib), SHMT2 had r = −0.288 with ABT-888, and MTR had r = −0.258 with BMN-673 (talazoparib). In contrast, NNMT showed no association with log(IC50) of PARP inhibitors, whereas increased expression levels of a folate receptor gene, FOLR1, and a folate transporter gene, SLC46A1, were very weakly associated with resistance to PARP inhibitors (e.g., r = 0.278 for correlation of SLC46A1 expression with ABT-888). The opposing directions of these weak correlations of OCM gene transcriptional levels with response to PARP inhibitors suggest the need for further investigation of any possible biological consequences of strong correlations be-tween OCM gene expression and BRCA1 gene expression.  D.-J. Min, S. Vural and J. Krushkal
    OCM gene expression was a significant predictor of cancer cell line response to crizotinib when accounting for genomic alterations affecting crizotinib sensitivity
    Sensitivity to certain kinase inhibitors is known to be strongly associated with, or in a number of cases require the presence of specific genomic alterations including gene amplifications, genome rearrangements, or specific protein-changing mu-tations; in addition, initially sensitive tumors often acquire secondary mutations that result in drug resistance (Sup-plementary Table 5) [51,54,79]. We examined three kinase inhibitors, erlotinib, lapatinib, and crizotinib (Table 1), for which sensitivity and resistance within specific cancer categories have been associated with specific genomic rearrangements, gene amplification, or with DNA and protein sequence changes in their molecular target genes (Supplementary Ta-ble 5). The numbers of cell lines with drug response and gene expression data that also had genomic alterations associated with sensitivity to erlotinib or lapatinib were insufficient (≤1) to account for such genomic changes (Supplementary Table 5). We were able to analyze cell line response to crizotinib across all cancer categories, conditional on the presence of genomic alterations affecting crizotinib sensitivity (ALK fusions, ROS1 fusions, MET amplification, or MET exon 14 skipping mutations) or promoting resistance to that agent (Supplementary Table 5). As discussed above, expression of MAT2B, MTR, SLC46A1, and SHMT2 was correlated with crizotinib response (Table 1). In multiple regression analyses, pre-treatment expression of each of these OCM genes remained a significant predictor of log(IC50) of crizotinib after accounting for the presence of genomic alterations with known roles in crizotinib sensitivity or resistance (FDR adjusted p between 1.38 × 10−6 and 0.0084 for log2 of OCM gene expression; Supplementary Table 6). The presence of genomic alterations affecting crizotinib sensitivity was also significantly associated with crizotinib response (p between 0.0121 and 0.0469; Supplementary Table 6). Overall, the fit of the multiple regression models was poor (Supplementary Table 6), most likely due to the low number of cell lines with relevant genome alterations, which dictated the combined use of data from a variety of cancer categories. None of the four OCM genes, MAT2B, MTR, SLC46A1, and SHMT2, had significant differences in log2 of their expression values between the groups of cell lines with and without genomic alterations known to affect sensitivity to crizotinib (p between 0.3126 and 0.6149 when using the Student’s t-test; data not shown). Therefore, OCM gene expression was a significant predictor of crizotinib response independently from the status of genome alterations known to affect crizotinib sensitivity.