An informatics approach identifying markers of chemosensitivity in human cancer cell lines

Citation
Sa. Amundson et al., An informatics approach identifying markers of chemosensitivity in human cancer cell lines, CANCER RES, 60(21), 2000, pp. 6101-6110
Citations number
51
Categorie Soggetti
Oncology,"Onconogenesis & Cancer Research
Journal title
CANCER RESEARCH
ISSN journal
00085472 → ACNP
Volume
60
Issue
21
Year of publication
2000
Pages
6101 - 6110
Database
ISI
SICI code
0008-5472(20001101)60:21<6101:AIAIMO>2.0.ZU;2-J
Abstract
We have used a sensitive and reproducible method of measuring mRNA expressi on to compare basal levels of 10 transcripts in the 60 cell lines of the Na tional Cancer Institute's ill vitro anticancer drug screen (NCI-ACDS) under conditions of exponential growth. The strongest correlation among these ta rget genes was between levels of CIP1/WAF1 and BAX, Levels of the three maj or growth arrest and DNA damage-inducible gene transcripts, (GADD34, GADD45 , and GADD153), which are coordinately regulated in response to many stress es, were also correlated across the 60 cell lines. Although the stress indu ction of several of the transcripts studied here has been shown to be depen dent on wild-type p53 status, basal levels of only CIP1/WAF1 and BAX were f ound to correlate with p53 status. As expected, basal expression of O-6 alk yl guanine alkyl-transferase correlated well with resistance to OG-alkylati ng agents (r = -0.44) but not with resistance to alkylators with different mechanisms of action (r = -0.04), When basal expression levels of the 10 ge nes across the NCI-ACDS panel were compared with sensitivities to a panel o f 122 standard chemotherapy agents, the most striking relationship was a st rong negative correlation (r = -0.3) between basal BCL-X levels and sensiti vity to drugs in all of the mechanistic classes except one class of antimet abolites. Sensitivities to a maximally diverse sample of 1200 from 70,000 c ompounds tested in the NCI-ACDS of agents were also negatively correlated w ith BCL-X levels, A novel application of factor analysis revealed that the newly discovered associations were independent of previously demonstrated s ensitivity factors such as p53 mutation status and native population doubli ng time. A similar pattern of correlation was seen for Bcl-X-L protein leve ls. Conversely, BAX and BCL2, two other genes associated with regulation of apoptosis, showed no overall correlation with drug sensitivities. This sug gests that BCL-X may play a unique role in general resistance to cytotoxic agents, with the cell lines demonstrating relative resistance to 70,000 cyt otoxic agents in the NCI-ACDS being characterized by high BCLX expression.