Biology Reference
In-Depth Information
the cells isolated from human breast tissues and tumors (Perou et al. 2000). The
s-ribons at AT1, AT2, A-T3 are the “gene expression profile” of 13 RNAs selected
by Perou et al. (2000) as biomarkers of the
normal
human breast tissue. The
s-ribons at B-T2, B-T3, and C-T2 (labeled BE, i.e., “before treating with doxorubi-
cin”) are the levels of the 13 RNAs serving as the biomarkers for human breast
tumors
. The distribution of these different types of s-ribons is consistent with the
conclusion that Type A cells are normal, Type B cells are tumor cells resistant to
drug therapy, and Type C cells are tumors that are completely restored to the normal
cell type by drug therapy.
Perou et al. (2000) applied the hierarchical clustering method to the three human
breast tissues - (1) normal, (2) tumor before (BE) treating with doxorubicin, and (3)
tumor after (AF) the drug therapy. The tumor samples were obtained from 65
surgical specimens of human breast tumors. They measured the RNA levels using
microarrays representing 8,102 human genes. Twenty tumors were sampled twice
before (BE) and after (AF) treating with doxorubicin for ~16 weeks. The Cy5 dye
was used to label the cDNA synthesized from the mRNA isolated from experimen-
tal samples, and the Cy3 dye was used to label the cDNA synthesis from the mRNA
isolated from different cultured cell lines which served as the common reference
samples. A total of 84 cDNA microarray experiments were performed, and 1,753
RNAs (which is 22% of 8,102 RNAs) were analyzed, whose levels varied by at least
fourfold from their median level in this sample set in at least three of the samples.
Out of the 65 tissue samples analyzed, a subset of 496 RNAs (termed “intrinsic gene
subset”) was selected that showed significantly greater variations in abundances
between different tumors than between paired samples from the same tumors.
Perou et al. (2000) found that eight clusters of RNAs, numbering 134 in total,
reflected variations in specific cell types in tumors. These RNAs can be viewed as
“breast cancer-related” RNAs since their levels varied more between different
tumors than within each tumor sample. Since these RNA have been selected out
of 1,753 RNAs, the fraction of the RNAs that is related to cancer may be estimated
to be (134/1,753)
8%.
From the microarray data on 8,102 RNAs isolated from human breast tissues and
tumors published by Perou et al. (2000), 54 RNAs specializing in energy metabo-
lism (i.e., glycolysis, Krebs cycle, oxidative phosphorylation, and fatty acid catab-
olism) were selected and their “gene expression profiles” (i.e., s-ribons) in three
different tissue samples, that is,
normal
, and
tumors
before and
after
treating with
doxorubicin, were plotted as shown in Fig.
19.2
. As evident in the four panels in this
figure, all of the 54 s-ribons showed distinct patterns for the three samples with little
overlap, except for RNAs 9, 13, 23, 25, 33, 34, and 51 (whose peaks coincide).
These constitute about 10% of the total number of RNAs. Therefore, we can
conclude that, about 90% of the time, s-ribons participating in energy metabolism
serve as unique biomarkers for the different states of the cells constituting the three
human breast tissues. This is consistent with the following hypothesis:
The structural ribons of energy metabolizing RNAs (and most likely other RNAs as well)
can be employed as reliable cell-state biomarkers for breast cancer theragnostics.
100
¼
(19.1)
