Molecular Biology of the Cell by Bruce Alberts, Alexander Johnson, Julian Lewis, David Morgan, Martin Raff, Keith Roberts, Peter Walter by by Bruce Alberts, Alexander Johnson, Julian Lewis, David Morg

1112 Chapter 20: Cancer
attributed to chromosome breakage and rejoining. Systematic surveys of the protein-coding
genes in common solid tumors—such as those of the breast, colon,
brain, or pancreas—have revealed that an average of 33 to 66 genes have undergone
somatic mutation affecting the sequence of their protein product. Mutations
in noncoding regions of the genome are much more numerous, as one would
expect from the much larger fraction of the genome that noncoding DNA represents.
But they are considerably more difficult to interpret.
The high frequency of mutations testifies to the genetic instability of many
cancer cells, but it leaves us with a difficult problem. How can we discover which
of the mutations are drivers of cancer—that is, causal factors in the development
of the disease—and which are merely passengers—mutations that happen to
have occurred in the same cell as the driver mutations, thanks to genetic instability,
but are irrelevant to the development of the disease? A simple criterion is
based on frequency of occurrence. Driver mutations affecting a gene that plays a
part in the disease will be seen repeatedly, in many different patients. In contrast,
passenger mutations, occurring at more-or-less random locations in the genome
and conferring no selective advantage on the cancer cell, are unlikely to be found
in the same genes in different patients.
Figure 20–25 shows the results of an analysis of this sort for a large sample
of colorectal cancers. The different sites in the genome are laid out on a two-dimensional
array, with chromosome serial number along one axis and position
within each chromosome along the other. The frequency with which mutations
are encountered is shown by height above this plane, creating a mutation “landscape”
with mountains (sites where mutations are found in a large proportion of
the tumors in the sample), hills (where mutations are found less frequently but
still more often than would be expected for a random scattering over the genome),
and hillocks (sites of occasional mutations, occurring at a frequency no higher
than would be expected for mutations scattered at random in each individual
tumor). The mountains and the hills are strong candidates to be the sites of driver
mutations—in other words, sites of cancer-critical genes; the hillocks are likely to
correspond to passengers. Indeed, many of the mountains and hills turn out to be
sites of known oncogenes or tumor suppressor genes, whereas the hillocks mostly
correspond to genes that have no known or probable role in causation of cancer.
Of course, some hillocks may correspond to genes that are mutated in only a few
rare patients but are nevertheless cancer-critical for them.
About One Percent of the Genes in the Human Genome Are
Cancer-Critical
From studies such as the one just described, it is estimated that the number of
driver mutations for an individual case of cancer (the sum of meaningful epigenetic
and genetic changes in both coding sequences and regulatory regions) is typically
on the order of 10, explaining why cancer progression generally involves an
increase in genetic and/or epigenetic instability that enhances the rate of such
changes.
PIK3CA FBXW7 p53 K-Ras
gene position along chromosome
chromosomes 1 to X
APC
Figure 20–25 The mutation landscape in
colorectal cancer. In this two-dimensional
representation of the human genome,
the green surface depicts the 22 human
autosomes plus the X sex chromosome
as being laid out side-by-side in numerical
order from left to right, with the DNA
sequence of each chromosome running
from back to front. The mountains
represent the locations of genes
mutated with high frequency in different,
independent tumors. As indicated, these
are suspected driver mutations in the
adenomatous polyposis coli (APC), K-Ras,
p53, phosphoinositide 3-kinase (PIK3CA),
and ubiquitin ligase (FBXW7) proteins.
(Adapted from L.D. Wood et al., Science
318:1108–1113, 2007.)