Does Radiation Cause Cancer?
The author discusses the multistage
aspects of cancer formation and how acute irradiation may
affect these processes.
|by James E Trosko, RERF Chief of Research
|This article was originally published
in RERF Update 4(1):3-5,
|What an outrageous question! With an
enormous amount of experimental animal and human epidemiological
data showing that a wide variety of cancers appear after exposure
to various kinds of radiation, one could hardly question the
conclusion that radiation does "cause" cancer. But wait! Our
understanding of carcinogenesis today leads us to believe
that it is a multistage, multimechanistic process, involving
the interaction of many external and endogenous factors. Consequently,
it is misleading to assume that any single factor or "carcinogen"--chemical
or physical--"causes" cancer. The key word here is cause.
Carcinogenesis involves many steps and mechanisms, with the
interaction of external determinants such as chemical and
physical pollutants, medication/drugs, mutagenic and epigenetic
agents--as these may occur in the diet or as workplace and
environmental pollutants--and endogenous factors related to
genetic background, sex, developmental stage, number of stem
or progenitor cells that give rise to cancer, DNA repair systems,
hormones, growth factors, oncogenes, tumor suppressor genes,
and antimetastasis genes.
Carcinogenesis as a multistep,
Currently, the multistage model of carcinogenesis, derived
from whole animal experimental studies, seems to be a plausible
model for human carcinogenesis. This model indicates that
the first step in carcinogenesis--the initiation stage--is
irreversible. The observation that mutagens appear to be effective
initiators implicates mutagenesis as the mechanism underlying
the initiation stage. The fact that the initiation process
appears to be irreversible is also consistent with the hypothesis
that mutagenesis is at least one mechanism of initiation.
Stable epigenetic repression or activation of genes may be
Most cancer studies have been consistent with the clonal theory
of cancer, ie, the assumption that cancer arises from changes
initiated in one cell (Figure 1). Therefore,
the second step in carcinogenesis--the promotion stage--appears
to involve the clonal expansion of an initiated stem cell,
which, because it is unable to terminally differentiate, accumulates
as a focus of nonterminally differentiated cells. Examples
of such foci might be papillomas of the skin, enzyme-altered
foci of the liver, polyps of the colon, and nodules of the
breast. Obviously, this process must require stimulation of
cell division (ie, it must be mitogenic), at least with respect
to the initiated cell. As demonstrated in experimental animals,
this stage is potentially interruptable and reversible.
The initiation-promotion-progression model of carcinogenesis.
|Beta1, rate of terminal differentiation and death of
stem cell; Beta2, rate of death, but not of terminal differentiation
of the initiated cell (--||-->); Alpha1, rate of cell division
of stem cells; Alpha2, rate of cell division of initiated
cells; Mu1, rate of the molecular event leading to initiation
(ie, possibly mutation); and Mu2, rate at which the second
event occurs within an initiated cell.
|If one of these promoted, initiated
cells acquires additional genetic alterations (eg, other mutations,
stable epigenetic changes) that allow the cell to become promoter-independent,
invasive, and metastatic, then the third step of carcinogenesis--the
conversion or progression stage--has occurred. This step also
appears to be irreversible. Given the observations that mutagens
appear to affect this stage, mutagenesis as well as stable
epigenetic events could be applicable mechanisms for progression.
How do the different factors fit
If carcinogenesis is a multistep process, with each stage
affected by different mechanisms (eg, there are many ways
to cause mutations; many mechanisms lead to mitogenesis),
how could a single exposure to ionizing radiation "cause"
cancer? For those who harbor the idea that ionizing radiation
"causes" cancer, does it not seem incredible that after an
acute exposure, radiation would not only have to activate
one or more oncogenes, as well as inactivate suppressor genes,
it would also have to initiate, promote, or clonally expand
that cell manyfold and then convert one of those initiated
cells by mutating other genes to have invasive and metastatic
abilities by a series of independent events in a single cell?
Would it not be more informative to ask questions such as:
"Which step(s) of carcinogenesis might be affected by ionizing
radiation?"; "By what mechanisms might ionizing radiation
initiate, promote, or bring about progression of carcinogenesis?";
"Does ionizing radiation activate oncogenes?"; and "Does ionizing
radiation deactivate tumor suppressor genes?" Moreover, does
a linear-no threshold model describe the underlying mechanisms
of the multistage nature of carcinogenesis, especially the
promotion or mitogenic step? A recent review of chemical carcinogenic
studies appears to indicate a no-effect, threshold level for
the role of the mitogenic process, primarily linked to the
promotion or clonal expansion of the initiated cell (S Cohen
and L Ellwein, Cancer Res 52:6493-505, 1991). Serious
examination, as suggested in an accompanying article by VP
Bond (RERF Update 4:7, 1992), must be considered. By rephrasing
the problem, specific testable hypotheses, using these concepts
and molecular technologies, might provide new insights into
how (in our case) exposure to A-bomb radiation could have
contributed to the process of carcinogenesis.
Epidemiologists involved in determining the risks of radiation
exposure use the term "confounding factors" for factors, such
as age at exposure, time elapsed since exposure, sex, reproductive
history, diet, postirradiation therapy, etc, that are known
to be associated with "modifying" the radiation response.
In the context of the carcinogenic process outlined above,
the term "confounding factors" is very misleading. In fact,
the term ought to be "contributing factors," with radiation
being only one of these factors.
Oncogenes, tumor suppressor genes,
and intercellular communication
The initiation/promotion/progression hypothesis of carcinogenesis
is an operational concept derived from whole-animal experiments,
having no implied underlying molecular mechanism. Independently,
the oncogene/tumor suppressor gene hypothesis is a concept
derived empirically from molecular in vivo and in vitro studies.
However, to date no viable cellular mechanism has been offered
as to how the various oncogenes and tumor suppressor genes
might function to convert a contact-inhibitable progenitor
cell to a cancer cell, ie, one that is not contact-inhibitable
and unable to terminally differentiate.
In a multicellular organism, tight regulation of a cell's
ability to proliferate and to differentiate must occur. Various
intercellular communication mechanisms, such as extracellular
communication via growth factors, hormones, or gap-junctional
intercellular communication (GJIC) via ions and small molecules
through gap-junction channels, appear to be directly associated
with the regulation of cell growth and differentiation. Since
the major phenotypic dysfunctions of cancer cells seem to
be the lack of contact inhibition and loss of growth control
and the ability to terminally differentiate, it would be reasonable
to speculate that intercellular communication has been disrupted
during the carcinogenic process. Indeed, many if not all cancer
cells have abnormal homologous or selective communication
characteristics. Many chemical tumor promoters, oncogenes,
and growth factors also inhibit intercellular communication,
whereas the few antitumor agents and anticarcinogens seem
to up-regulate GJIC (JE Trosko et al, Pathobiology
58:265-78, 1990; JE Trosko et al, Radiat Res 123:241-51,
1990). One tumor suppressor gene has been associated with
a gap-junction gene (SW Lee et al, Proc Natl Acad Sci USA
88:2825-9, 1991). These circumstantial, but completely
independent, observations are consistent with the hypothesis
that the oncogene/suppressor gene function modulates GJIC,
which, in turn, modulates a cell's ability to proliferate
The role of ionizing radiation in carcinogenesis
The question now arises as to how radiation might affect one
or more of the mechanisms underlying the initiation, promotion,
and progression phases of carcinogenesis.
Radiation: a weak initiator
The current weight of the evidence indicates that
ionizing radiation is a rather weak point mutagen but a good
clastogen (inducer of chromosome breaks, deletions, and rearrangements).
At high enough doses, this translates into radiation being
a good cytotoxicant since chromosome deletions and many types
of chromosome rearrangements are lethal. In contrast, given
the oncogene/tumor suppressor gene paradigm, where a balance
between proto-oncogenes and tumor suppressor genes is needed
for normal growth control and differentiation (see Figure
2), ionizing radiation might generally be predicted
to be a rather weak "activator" of oncogenes but a strong
"deactivator" of tumor suppressor genes, except possibly by
causing rearrangement of normal proto-oncogenes causing them
to be abnormally and stably expressed.
|Figure 2. The yin/yang model
of oncogenes and tumor suppressor genes in the control of
cell growth depicts the balance between positive factors
that stimulate growth and negative factors that suppress
growth. In the quiescent state of a normal cell that is
contact inhibited (solid tissue) or suppressed by extracellular
regulators (soft tissues), the two factors balance out.
During the initiation-promotion-progression process of carcinogenesis,
activation of oncogenes could occur, followed by clonal
expansion of these cells. The loss of tumor suppressor genes
by mutation or by deletion allows the cell to enter the
progression phase by stimulation of cell growth or by inability
to respond to negative growth control (ie, growth inhibition).
The role of gap junctional intercellular communication (for
cells in solid tissues) is speculated to be down-regulated
by oncogenes and up-regulated by tumor suppressor genes.
Radiation as a promoter
If ionizing radiation is to be a promoter, ie, a stimulant
of the clonal expansion of initiated cells, it must be given
at a high enough dose to cause significant cell killing, which,
in turn, would induce compensatory hyperplasia. If one of
the surviving stem cells were to have been previously initiated
by some other environmental mutagen or initiated by the radiation
itself, then regenerative or compensatory hyperplasia could
be seen as "promotion." If the dose were too high, the ionizing
radiation would start to kill some of the preexisting or newly
induced initiated cells, thereby decreasing the cancer incidence
and increasing the likelihood of early death of the organism.
Radiation as a progressor
If the exposed individual has preexisting initiated and promoted
clones of cells (as we all must: the older we get, the more
of these we should have), then ionizing radiation, as an effective
gene and chromosome deletion mutagen, might be expected to
be a relatively good tumor suppressor gene "deactivator."
Assuming that stem cells are the target cells for carcinogenesis,
then the risk that the initiation step of carcinogenesis occurs
from a single exposure to ionizing radiation is influenced
by the number of these stem cells. Some tissues will have
relatively stable numbers of these cells during aging (eg,
skin and the lining of the GI tract), whereas others, such
as the breast, liver, and brain tissues, may have decreasing
numbers of stem cells (due to many intrinsic and extrinsic
factors that I won't discuss here).
On the other hand, depending on previous exposure to other
point mutagens that might initiate (activate oncogenes) and
promote (age might be an important factor here), ionizing
radiation might be a good "progressor." The interaction of
sunlight exposure and ionizing radiation in skin cells of
older individuals, in sun-exposed and non-sun-exposed areas,
might be a test of this hypothesis, since UV radiation is
a good point mutagen and skin carcinogen (presumed initiator
and activator of oncogenes, promoter via its ability to kill
cells, and progressor by its ability to deactivate tumor suppressor
genes by point mutations).
Does radiation "cause"
Clearly, radiation is a contributing factor. How it contributes
in the multistage process of carcinogenesis is not yet known;
however, radiation does seem to be a weak initiator,
both by its ability to induce chromosome rearrangements (to
activate oncogenes) and by its weak point mutagenic potential.
At higher doses, radiation might act as an indirect promoter
of preexisting initiated stem cells by its ability to induce
regenerative hyperplasia via its cell-killing effects. Finally,
because of its ability to delete genes and chromosomes, it
should be effective as a deactivator of tumor suppressor
genes and therefore should act as a progressor at
the late stages of carcinogenesis.
All this does not imply that ionizing radiation is unimportant
in bringing about cancer. However, only by knowing the mechanisms
by which ionizing radiation influences this complex disease
process can we hope to develop meaningful risk estimates from
studies of populations and individuals exposed to radiation,
especially at low doses where the greatest uncertainties exist.
Clearly, epidemiological and statistical analyses of radiation-exposed
populations are critical and necessary. So are basic and fundamental
studies of radiation effects on molecules, cells, and animals.
In addition, the future demands a greater interaction among
epidemiologists and radiation molecular and cell biologists
for better hypothesis design, testing, and interpretion of
epidemiologic studies. At the same time, epidemiological findings
should stimulate laboratory research to explain the results.
The emerging field of molecular epidemiology may provide the
stimulus for this union (PG Shields and CC Harris, J Am
Med Assoc 266: 681-7, 1991). At RERF, we are developing
plans to move in this direction.
|Note added in proof: Three recent
papers (D Zhu et al, Proc Natl Acad Sci USA 88:1883-7,
1991; PP Mehta et al, J Membrane Biol 124:207-25, 1991;
B Ehlibali et al, Proc Natl Acad Sci USA 88:10701-5,
1991) have demonstrated that noncommunicating cancer cells,
when transfected with expressible gap-junction genes, had
restored cell-to-cell communication and normal growth patterns.