| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA; and
Department of Health Policy and Management, Harvard School of Public Health, Boston, Massachusetts, USA
1Correspondence: Department of Biology, Rm. 68–577a, M.I.T. Cambridge, MA 02139, USA. E-mail: msfox{at}mit.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL SYSTEMSDISCUSSIONREFERENCES |
|---|
Key Words: epigenetics • carcinogenesis • hypermutability • genetic regulation • prevention
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL SYSTEMSDISCUSSIONREFERENCES |
|---|
. Both of these observations indicate that the formation of a transformant could not have been the result of a direct effect of X-ray exposure. Furthermore, a protease inhibitor applied several cell divisions after exposure was found to markedly reduce the ultimate yield of transformants (2)
, showing that the alteration is reversible, and thereby demonstrating that this effect of X-ray exposure is not a change in DNA sequence. While X-ray exposure can damage DNA and result in mutations at all dose levels, exposure could also result in other dose level-sensitive effects, such as actions affecting regulation of enzyme levels. To account for 1) the failure of the yield to be proportional to the number of cells irradiated, 2) the delay of many generations between exposure and mutation, and 3) a process that is reversible during the period of multiplication, we propose a different causal relationship between exposure and mutation. We suggest that exposure to modest doses of radiation induces epigenetic alterations in some regulatory system of the exposed cells. One or more of these changes, such as the loss of the activity of a DNA mis-match repair system, can make cells hypermutable. An example of evidence for the appearance of transiently hypermutable cells in cell populations is given by Drake et al. (3)
. Such epigenetic changes do not involve mutational alterations in DNA sequence, but are nevertheless transmitted to progeny cells. Recent discussions of some possible mechanisms of epigenetic change in carcinogenesis are given by Chen et al. (4)
, Tlsty et al. (5)
, Feinberg et al. (6)
, and Baylin and Ohm (7)
.
The experiments so far described do not permit making a distinction between mutability and the expression of pre-existing recessive mutations by virtue of enhanced chromosomal loss during division (8)
. For simplicity, we do not distinguish between these possibilities and refer to both as hypermutability. As the altered cells multiply, a large altered cell population accumulates within which a rare transformation event, presumably a mutation, could occur. The reversibility suggests that the alteration conferring hypermutability in the exposed cells’ progeny can, in fact, be eliminated.
Although the concept of an expanding altered cell population arose from observations in cell culture experiments, in human cancer the route to a large altered cell population may be different. Here we propose that exposure to damaging agents is often chronic such that continuing exposure may result in a steadily increasing fraction of a steady-state population of cells with alterations that are epigenetically transmitted. The net effect in both laboratory and human cases is similar: the accumulation of an increasingly large population of altered, hypermutable cells; as that population becomes sufficiently large, a mutational event leading to cancer becomes increasingly likely. In fact, we elsewhere described a model showing that, for humans, the accumulation of an increasing fraction of the cell population that is hypermutable could constitute the rate-limiting step in carcinogenesis and result in a pattern of occurrence that tracks the steep rise of cancer incidence with age (9)
.
To the extent that transformation in cell culture constitutes a valid experimental model for learning about human cancer, the observation that the transformation event can occur many cell divisions after carcinogen exposure, combined with the observation that the alteration can be reversed, suggests a new way to think about cancer prevention and perhaps about treatment as well. If the mutations giving rise to malignancy occur many generations after exposure, we have the opportunity to intervene with chemical measures to reduce the hypermutability during this period in which the epigenetic alteration remains reversible. This reduction in hypermutability would, in turn, lead to a reduction in the primary incidence of cancer and perhaps the frequency of its recurrence.
|
EXPERIMENTAL SYSTEMS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL SYSTEMSDISCUSSIONREFERENCES |
|---|
. The process leads to piled-up colonies that can give rise to fibrosarcomas in mice (11
, 12)
. 2) The second system differs only in that it involves exposure to a chemical carcinogen, 3-methylcholanthrene, rather than X-rays (13)
. 3) The third laboratory system involves grafting of X-ray-exposed rat mammary epithelial cells to the skin of syngeneic rats, which are later screened for malignant tumors (14)
. 4) A fourth system was used to investigate the occurrence of mutations in clones arising on Petri dishes from single irradiated Chinese hamster ovary (CHO) cells (15)
.
1. X-ray experiments
Kennedy et al. examined two questions about the transformation of mouse cells in response to X-ray exposure (1)
. The first question is, how does the number of transformants vary over a range of initial cell numbers, all exposed to the same X-ray dose? The second is, when in the history of a colony growing from an irradiated cell does a cell with the phenotype of transformation appear?
Regarding the first question, one might expect, for example, that if the exposure had a direct effect on DNA, the number of transformants following a given dose of irradiation would be proportional to the number of exposed cells on each plate. To test this expectation, the yield on plates with 100–400 cells was compared with the yields on plates seeded with smaller numbers, 50–100, 10–30, and 
5, with all plates exposed to the same radiation dose of six grays. At this dose, 20% of the cells survive. After confluence many generations later, the yield of transformants was observed to be essentially constant, independent of the number of cells exposed. Rather than the yield being determined by the number of exposed cells, the yield appears to be largely determined by some feature common to all of the different plates. An obvious candidate is the common maximal number of cells (2x106) to which each plated population grows at confluence. In fact, Kennedy and Little reported in 1984 that for the same number of irradiated cells, increasing the plate size on which these cells were grown by a factor of 18 resulted in a similar increase in the yield of transformants at confluence (16)
. The yield is proportional to the number of cells at confluence, not the number of cells surviving irradiation.
When does the cell with the inherited transformant phenotype emerge? Plates with 300 surviving cells were incubated to allow growth to near confluence at 5 x 105 cells (1)
. For one group of plates, the cells were left intact and incubation was continued. For other sets of plates, cells from each plate were dissociated, resuspended, and replated at dilutions of 1:10, 1:30, 1:100, 1:1,000, and 1:10,000. This approach makes it possible to distinguish transformants that arose from the original plates prior to dissociation from those occurring during regrowth after dissociation. Transformants that arose prior to dissociation would have given rise to clones that, upon dissociation and replating, would produce many colonies. By contrast, transformants arising after dissociation would be expected to give rise to a small number of randomly appearing clones. Remarkably, at almost all of these dilutions, including the most dilute, the number of foci was similar to the yield on the undisturbed plates, which displayed small numbers of apparently randomly occurring clones. This observation implies that the cells became transformed long after dissociation and replating; indeed, not until the cell population on each plate had grown to close to its maximum value of 2 x 106, at which point rare events occurring at a rate of one in a million could in fact be observed. In other words, the cells that were dissociated and replated at multiple dilutions upon reaching near confluence were, for the most part, behaving similarly to the initial population of exposed cells with regard to the yield of transformants.
There were exceptions, however. For the 1:10 dilution, a small number of the original plates yielded relatively large numbers of foci, reflecting the emergence of a transformed clone several generations prior to dissociation. Here the individual cells of that initial clone gave rise, after dissociation, to several independent foci. Just as in the Luria and Delbruck experiment, the appearance of such clones with excess numbers is what would be expected from early mutations occurring during the growth of independent cultures of cells (17)
.
These results were interpreted in terms of a two-step process. The first step involved an epigenetic change induced at this dose in most or all of the surviving cells that is efficiently transmitted to their progeny. This change makes the cells more likely to undergo the second step of transformation, by mutation, many generations later. Later confirmation by Kennedy, Cairns, and Little strengthened and elaborated the conclusion that transformation is such a two-step process (16
, 18)
. The authors suggested that the initial step is "some change in the pattern of gene expression that... determines the frequency of subsequent rare genetic events." Furthermore, using a Luria-Delbruck analysis, they concluded that the second step is precisely such a rare event, consistent with a mutation, "occurring in slightly less than one in a million descendants of cells that have taken the first step" (18)
.
The notion of epigenetic changes that are stable during subsequent multiplication is not novel (1
, 4
5
6
7
, 19)
and is, in fact, the prominent feature of embryological development. We might restate the original implications of these experiments: an immediate effect of X-ray exposure (at these doses) is an epigenetic, heritable alteration in all or most of the cells. This alteration causes enhanced mutability, and as the progeny of these cells become numerous, a transformant, presumably by mutation, becomes increasingly likely.
2. Chemical carcinogen experiments
Fernandez et al. studied the yield of transformants in cultured C3H 10T1/2 mouse cells exposed to methlycholanthrene as a function of the number of treated cells (13)
. Similar to the X-ray experiments, they also found far fewer transformants than would be expected if the yield was proportional to the numbers of cells exposed. What they did find was that an increase of approximately two orders of magnitude in the number of cells exposed resulted in only a 6-fold increase in the number of transformants. To explain their finding, they, too, suggested that cells were altered as a consequence of exposure, but that with each new generation produced by cell division, some fraction of the progeny spontaneously revert to their normal state. Since larger numbers of initially plated cells require fewer rounds of replication to reach confluence, they have less opportunity for reversion from the altered state and therefore would be expected to have a somewhat larger of yield of transformants, which is, in fact, what the authors observed.
The authors’ interpretation of these results suggests that the process of loss of alteration can occur spontaneously. Independently, Kennedy and Little had earlier reported that after irradiation, treatment with antipain, a protease inhibitor, resulted in a reduction in the yield of transformants (2)
. In 1982, Kennedy showed that even brief exposure to antipain five generations after X-ray exposure can prevent all or most of the anticipated transformation events (20)
. Furthermore, Kennedy demonstrated that a soybean antiprotease had similar transformation prevention properties (21)
. These results showing reversibility provide strong evidence that the initial alteration is nongenetic.
3. Whole animal experiments
The third system for studying transformation involves transplantation of X-ray-exposed epithelial cells grafted at specific sites in syngeneic rats. Graft sites are observed for the numbers of cancers (14)
. In the experiments conducted by Kamiya and collaborators, dissociated cells from harvested mammary glands were irradiated and serially diluted to the desired range of exposed cells per standard graft volume. Recipient rats were grafted at a single site each in the interscapular white fat pad and the cephalolateral ends of each inguinal pad. Later, histopathologically confirmed tumors 
4 mm diameter in cell grafts were scored as cancers.
As the number of cells per graft site rose by 100-fold, the final cancer incidence, defined as the fractions of sites with cancer, rose by only a factor of five. Here again, the absence of a proportional response leads to the conclusion that the presumed mutation leading to cancer does not occur as a direct effect of X-ray exposure (7 grays). The authors interpret these experiments as suggesting an epigenetic change that, after many generations of growth, may give rise to a tumor.
4. When do induced mutations occur?
A fourth set of observations addresses the question of whether the transformation event is the result of a bona fide mutation. As mentioned, the studies by Kennedy, Cairns, and Little showed that transformation behaves as though it were the result of a mutation (18)
. Experiments performed by Little et al. with CHO cells found evidence for a delay of many generations in the appearance of mutations induced by X-ray exposure (15)
. Mutations occurring soon after exposure of single cells would be expected to give rise to clones many generations of multiplication later that would harbor a substantial mutant fraction. However, none of the mutant-containing clones, identified without selection, were observed to have the prominent presence of mutants that would be expected if the initial exposure resulted directly in their induction. On the basis of the observations in these experiments, all mutants would appear to have occurred 10 to 20 generations after exposure. This pattern of late-occurring mutations is similar to that observed for transformants resulting from irradiation in the experiments of Kennedy et al. (1)
.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL SYSTEMSDISCUSSIONREFERENCES |
|---|
The experimental data summarized here strongly support the hypothesis that the X-ray or carcinogen exposures described induce a reversible epigenetic change that confers hypermutability in many of the cells and their progeny. What we refer to as hypermutability may correspond to the notion of genetic instability advanced by other investigators (22)
. It has been shown that the effect of the X-ray-induced epigenetic alteration may be enhanced by brief exposure to carcinogen promoters such as phorbol esters and reversed by similar exposure to protease inhibitors such as antipain and soybean antiprotease (2
, 20
, 21
, 23)
. Within this framework, the rate-limiting step or steps involve the time it takes to accumulate a sufficiently large number of altered hypermutable cells such that it becomes likely that one of the cells in this large population will experience a mutation leading to malignant transformation. This overall hypothesis is strengthened by the fact that it was constructed from the results of cell cultures experiments in which two different kinds of carcinogenic treatment were used as well as from experiments performed on whole animals.
The ultimate question is what, if any, are the implications of these results for human cancer? The challenges of generalizing from cell culture studies to human cancer are well known, and a great deal of work remains to be done to bridge this gap. More specifically, two of the main features of these experiments—namely, the persistence of an epigenetically altered state and its reversibility by appropriate treatment—remain to be demonstrated in human cell lines.
If the cell culture results can be generalized to human cancers, a number of approaches with potential for high rewards are suggested for both treatment and prevention. In particular, these results suggest enlarging our focus to include studies of the existence and reversibility of epigenetic alterations. As an example, could inflammation be responsible for such alterations?
For years, appreciation of the central role of mutations in cancer has led in an apparently logical fashion to an emphasis on the clones of already mutated cells. One aim of this essay has been to make the case that, in formulating strategies for treating and preventing cancer, understanding how to interrupt the process leading to the formation of mutated cells may be at least of equal importance to targeting already transformed cells. Another and closely related aim is to emphasize the importance of launching efforts to identify appropriate human cell lines and a mode of inquiry that might reveal effects similar to those reported in the rodent experiments.
These human cell lines should display spontaneous transformation whose yield would increase after exposure to X-rays or methylcholanthrene. In fact, O’Reilly et al. reported X-ray-induced transformation results in such a human cell line (24)
, and Gowan et al. demonstrated similar results in mice with either ionizing radiation or the chemical agent hydroquinine (25)
. The availability of such cell lines would make it possible to investigate the yield as a function of the number of treated cells. It would also be possible to determine when during the growth of the clone the irreversible change to transformed phenotype occurs. Thus, we would be able to determine the delay between carcinogen exposure and fixation of irreversible change. It would also be important to examine the extent to which irradiated human cells will, upon growth, resuspension, dilution, and regrowth, reiterate the behavior of their ancestors with regard to the yield of transformants.
It would be of particular interest to examine the effects of subsequent chemical treatment of such cells with candidates such as protease inhibitors and other substances to determine whether the enhanced transformability due to carcinogen exposure can be diminished or eliminated. Indeed, such a system could provide an in vitro screen for inhibitors of the appearance of transformants in altered populations and perhaps also for inhibitors of humans cancer. In addition, there is the possibility that such substances may inhibit the recurrence of cancers in individuals whose cancer has regressed as a consequence of aggressive treatment. Finally, if many spontaneous human cancers occur by the mechanism we have proposed, this realization could open up promising new opportunities for prevention of this disease.
|
ACKNOWLEDGMENTS |
|---|
Received for publication May 15, 2006. Accepted for publication June 12, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL SYSTEMSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. M. Wiseman, O. L. Griffith, S. Deen, A. Rajput, H. Masoudi, B. Gilks, L. Goldstein, A. Gown, and S. J. M. Jones Identification of Molecular Markers Altered During Transformation of Differentiated Into Anaplastic Thyroid Carcinoma Arch Surg, August 1, 2007; 142(8): 717 - 729. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
