Jp. Demuth et al., THE GENE-EXPRESSION INDEX C-MYC X E2F-1 P21 IS HIGHLY PREDICTIVE OF MALIGNANT PHENOTYPE IN HUMAN BRONCHIAL EPITHELIAL-CELLS/, American journal of respiratory cell and molecular biology, 19(1), 1998, pp. 18-24
Recent methodological developments allow expression measurement of man
y genes simultaneously, thereby revealing patterns of gene expression
that can be related to phenotype. We hypothesized that through the use
of such methods we could identify patterns of gene expression associa
ted with the malignant phenotype in human bronchial epithelial cells (
BEC). To test this hypothesis, a recently developed quantitative rever
se transcriptase polymerase chain reaction method was used to assess s
imultaneously expression of 15 genes mechanistically associated with c
ell-cycle control (c-myc, E2F-1, p21, rb, PCNA, cyclin D2, cyclin D3,
cyclin E, cdc2, CDK2, CDK4, mad, max p21, max p22, and p53) in normal
cell cultures from five individuals and in nine different malignant BE
C lines. Relative to the mean expression levels in cultured normal cel
l populations, expression of c-myc, E2F-1, PCNA, cyclin E, and CDK4 me
ssenger RNA (mRNA) were significantly increased and expression of p21
and p53 mRNA were significantly decreased in one or two, but not all t
hree subtypes (squamous, adenocarcinoma and small cell) of carcinoma c
ell lines evaluated. No single cell-cycle control gene discriminated a
ll three subtypes from normal cell populations. In contrast, the gene
expression index c-myc X E2F-1/p21 separated all carcinoma cell lines
from all normal cell populations initially evaluated. This malignancy
index was validated in an additional three cultured normal BEC and thr
ee carcinoma cell lines, as well as three pairs of matched primary nor
mal bronchial epithelial and primary bronchogenic carcinoma samples, a
nd three pairs of matched primary normal lung parenchyma and primary b
ronchogenic carcinoma tissue. Again, the c-myc X E2F-1/p21 index succe
ssfully discriminated all cultured and primary normal from malignant s
amples and thereby had a predictive value of 1 (no false positives and
no false negatives). We hypothesize that because of functional mutati
ons in cell-cycle regulatory genes (e.g., p53 and/or rb), cells lose t
he ability to maintain a pat tern of gene expression mechanistically a
ssociated with normal, division-limited homeostatic equilibrium. Becau
se the c-myc x E2F-1/p21 gene expression index has high specificity fo
r malignant tissue, it will allow confirmation that there is a signifi
cant amount of tumor tissue present in small (e.g., fine-needle) biops
y specimens prior to evaluating them for expression of other genes, su
ch as those involved in chemoresistance or radioresistance. In additio
n, the goal of most gene therapy efforts is to alter levels of gene ex
pression quantitatively. This index and others derived in a similar ma
nner may better define potential gene therapy targets as well as respo
nse of targeted genes to therapy.