• 2019-07
  • 2019-08
  • 2019-09
  • 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • br Multiple strong positive correlations were observed betwe


    Multiple strong positive correlations were observed between 佛波酯 of OCM genes and genes encoding
    Table 3 Strongest Pearson correlations between expression levels of genes encoding drug target pathway components and one-carbon metabolism genes.
    OCM gene Target pathway r FDR adjusted p
    Listed are the strongest correlation results (Pearson r > 0.45 and FDR adjusted p < 0.05) of OCM genes with genes encoding drug tar-gets and target pathway components for agents listed in Table 1 and Supplementary Table 1. The results provided in the table are based on 1036 CCLE cell lines with available gene expression microarray data. Supplementary Table 4 provides the complete set of correlation results between the OCM genes and drug target pathway compo-nents for which microarray transcriptional measures were available from CCLE. r: Pearson correlation coefficient.  31
    kinases, e.g. ABL2, EGFR, CDK1, CDK2, AURKB, CHEK1, ligands, e.g. VEGFC, and activators, e.g. FGF2 (Table 3), which may suggest a potential explanation for associations between OCM gene expression and cancer cell line sensi-tivity to a number of kinase inhibitors (Table 1; Fig. 3). For example, expression of ABL2, an Abelson family kinase, was significantly positively correlated with NNMT expression (r = 0.472; p = 8.18 × 10−56; Table 3). Its product, ABL2, is a target of several kinase inhibitors including imatinib, nilotinib, and dasatinib [97]. Sensitivity to those agents was associated with higher expression levels of OCM genes, including an association of NNMT with dasatinib sensitivity (r = −0.340, p = 3.31 × 10−6; Table 1). Since ABL2 expres-sion increases with tumor progression in solid tumors [97], it may be of potential interest to investigate whether expres-sion levels of OCM genes may also change during cancer progression.
    Expression of several members of the ERBB family of receptor tyrosine kinase genes was significantly correlated with OCM genes, including both positive and negative as-sociations. The strongest correlation was a positive associa-tion between NNMT expression and expression of the epi-dermal growth factor receptor gene, EGFR (r = −0.467,
    p = 1.32 × 10−54; Table 3). In contrast, EGFR expression had a weaker negative correlation with expression of TYMS, MTR, GART, and DHFR (r between −0.373 and −0.325; p ≤ 2.82 × 10−25; Supplementary Table 4). Two other ERBB family members, ERBB2 and ERBB3, were significantly (p ≤ 3.39 × 10−22) associated with expression of the folate transporter SLC46A1 gene (r = 0.356 for ERBB2 and 0.382 for ERBB3), folate receptor FOLR1 gene (r = 0.310 for ERBB2), and MTR (r = −0.305 for ERBB3). Interestingly, only BHMT showed significant moderate correlations with erlotinib and la-patinib which target EGFR and ERBB2 [98]; however, BHMT expression did not correlate with expression levels of either EGFR, ERBB2, or ERBB3 (Supplementary Table 4). There-fore, molecular mechanisms of association between BHMT expression and log(IC50) of erlotinib and lapatinib require fur-ther investigation.
    Increased expression of several OCM genes was signifi-cantly (p ≤ 5.27 × 10−27) positively correlated with transcrip-tional levels of aurora kinase A and B genes. Among them, TYMS, DHFR and SHMT1 were strongly correlated with el-evated expression of AURKB (r = 0.527, 0.487, and 0.457, respectively; Table 3), whereas MTHFD1 expression was to a lesser extent correlated with both AURKB and AURKA (r = 0.404 and 0.336, respectively; Supplementary Table 4). This association may explain modest correlations of higher ex-pression levels of TYMS, MTR, and GART with sensitivity to aurora kinase inhibitors VX-680, GSK1070916, and Genen-tech Cpd 10 (r between −0.300 and −0.372; Table 1). Be-cause Aurora kinases are commonly overexpressed in can-cer [99], positive correlations between OCM gene expression and AURKA or AURKB expression could potentially also arise due to progression in tumorigenesis rather than through direct co-regulation of these pathways.
    We observed multiple significant correlations with tar-gets of epigenetic drugs, which were relatively weak (0.30 < r < 0.36; Supplementary Table 4). Expression of BRD2, one of the targets of the I-BET-762 bromodomain inhibitor, was negatively correlated with NNMT expression (r = −0.319, p = 2.87 × 10−24; Table 1 and Supplementary
    Table 1). Expression of multiple 佛波酯 histone deacetylase (HDAC) genes showed both positive and negative weak correlations with numerous OCM genes ( r between 0.301 and 0.353, p ≤ 4.58 × 10−23). Among positive correlations was expres-sion of HDAC1 with SHMT1, DHFR, and MTHFD1, that of HDAC2 with TYMS, HDAC3 with DHFR, HDAC10 with FTCD, and HDAC11 with SLC46A1. Negative correlations were observed between HDAC5 and MTHFS, and between HDAC11 and GART and TYMS. Products of these HDAC genes are targeted by histone deacetylase inhibitors tubas-tatin A, vorinostat, and VNLG/124. Sensitivity to those agents was significantly (p ≤ 2.14 × 10−12) correlated with increased expression of GART (r = −0.321), NNMT (r = −0.324), and SHMT2 (r = −0.332), respectively (Table 1; Supplementary Table 1). Association of OCM gene expression with sensitivity to HDAC inhibitors may be due to multiple direct and indirect mechanisms that include correlations between OCM gene ex-pression and HDAC genes, as well as multiple metabolic and regulatory links between the OCM reactions and processes targeted by the HDAC inhibitors [7,77]. For example, NNMT directly controls the ratio of SAM to S-adenosyl homocysteine (SAH), and cellular NNMT levels and its activity directly affect histone methylation and the activity of DNA methyltrans-ferases and histone methyltransferases [77]. Additionally, treatment with HDAC inhibitors directly inhibits glucose transport and results in inhibition of glycolysis, which is linked to the OCM cycle via the serine synthesis pathway (SSP) [1,77,100].