[CANCER RESEARCH 50, 6769-6771, November 1, 1990]
Perspectives in Cancer Research
The Cell Cycle: Myths and Realities
Renato Baserga1
Department of Pathology and the Fels Institute for Cancer Research and Molecular Biology, Temple University School of Medicine, Philadelphia, Pennsylvania 19140
A number of startling discoveries in the past few years have
given us a picture of the cell cycle that is solidly based on
molecular biology and genetics. Growth factors (both stimula
tory and inhibitory), oncogenes and antioncogenes, the bio
chemistry of DNA replication, the animal homologues of yeast
cell cycle genes, cyclins, the use of antisense strategies, and
many others have opened new vistas to our understanding of
cell cycle controls (1). This perspective, however, is not dedi
cated to a eulogy of these seminal findings; they speak for
themselves. I will, in fact, be the devil's advocate and do the
opposite, i.e., have a critical look, not at the findings (they are
unarguable), but at some of the conclusions that have been
drawn from these findings. I wish to make myself very clear:
there has been a revolution in the past 10 years in our under
standing of the cell cycle, a revolution supported by findings
the importance of which cannot be exaggerated. And most of
us have been guarded in our conclusions. But as it happens
often in science, a timid suggestion at the end of a discussion,
a carefully worded hypothesis, become transformed in the next
paper (often by other authors) into a fact. Thus, a simple
statement like "gene X may play a role in cell proliferation"
becomes in the next paper "gene X is growth regulatory." It is
not surprising, therefore, that this transforming capacity of
words has given rise to a number of interpretations of cell cycle
events, interpretations that are clearly extrapolations of the
facts, and that I have elected, in this perspective, to call myths.
It is some of these myths that I would like to discuss here, not
to pour water on our enthusiasm (which is justified) but for the
purpose of placing our understanding of the cell cycle back to
its proper perspective, which in simple terms, is: "there are
many things that control cell proliferation and they have not
all been sorted out." This may seem obvious; yet some of the
comments we have read recently in scientific journals are far
from following this maxim.
Let us look at some of these myths.
Myth 1: Growth-regulated Genes Are Growth Regulatory
This is one of the most common and persistent myths. It
began several years ago when it was shown that certain protooncogenes were growth regulated, i.e., that the steady state
mRNA levels of some protooncogenes increased when quies
cent cells (usually fibroblasts or lymphocytes) were stimulated
to proliferate by growth factors. The first two protooncogenes
that were shown to be growth regulated were c-myc (2) and cJt s (3), but several others (c-fgr, c-myb, c-ets-l, etc.) have been
added since then (4).
The extrapolation from growth regulated to growth regula
tory was, in some respects, justified. After all, protooncogenes
must have something to do with the control of cellular prolif
eration, since, when the expression or activity of a protoonco
gene is modified by mutations, translocations, amplification, or
Received 5/11/90; accepted 8/2/90.
' To whom requests for reprints should be addressed, at Department of
Pathology, Temple University School of Medicine, 3420 N. Broad Street, Phila
delphia, PA 19140.
simply overexpression, the regulation of cell growth is affected.
Indeed, even now, despite antioncogenes and yeast genes, one
cannot disregard protooncogenes in any rational scheme of
animal cell proliferation. The extent of their role may be
debatable, but they do have a role. However, their putative
growth-regulatory role is based on their relationship to viral
transforming genes and not on the fact that they are growth
regulated.
In the meantime, scores of laboratories (including mine, I
admit) have been picking out, through differential screening of
complementary DNA libraries, an embarrassing number of
growth-regulated genes. At first, genes the expression of which
increases early in G, (after stimulation of G0 cells) were most
popular but, more recently, growth-regulated genes at the d-S
boundary have become fashionable. In many instances, the
announcement of the discovery of a new growth-regulated gene
is accompanied by more or less bold comments about its im
portant role in the control of cell proliferation. It is time that
this myth be put to rest for two reasons: (a) the expression of
some growth-regulated genes and oncogenes is also induced in
situations in which cell proliferation does not occur, or, indeed,
may even be inhibited, for instance, c-fos (5); and (b) more
important, some exquisitely growth-regulated genes do not have
any growth-regulatory function. The best example is the thymidine kinase (TK) gene: growing cells can do without TK;
indeed, there are animals whose cells are all TK~ without,
obviously, affecting their growth (6). Conclusion: a growthregulated gene is a growth-regulated gene. To become growth
regulatory it must meet other criteria.
Myth 2: What's Going On in My Cells Is Universal and Is the
Only Truth
Probably all of us, at one time or another, commit this sin.
Most of us repent quickly; others persist obstinately in their
parochial views. The myth takes many forms. For instance,
some years ago, there was a lot of resistance to the concept that
the 5'-flanking sequence of the TK gene plays a major role in
the control of TK expression in fibroblasts. The opposition
came mostly from people working with muscle cells, where the
regulation of TK expression is totally different. The point is
that, in this instance, both parties were right; it just happens
that fibroblasts and muscle cells have their own particular ways
of regulating TK expression.
There are many other examples: (a) c-myb clearly plays a
role in the proliferation of hemopoietic cells (7), but fibroblasts
completely ignore it, and even v-myb (with its long terminal
repeat) cannot transform fibroblasts; (b) the classical growthregulated genes of fibroblasts and lymphocytes (c-fos, c-myc,
etc.) are not growth regulated in alveolar lung epithelial cells,
or, at least, their mRNA levels are not growth regulated, al
though their proteins are (8); (c) HeLa cells, very popular in
many laboratories, are often presented as model cells for cell
cycle regulation of gene expression. Mind you, there is nothing
wrong with the data. I have no doubts that gene expression in
6769
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1990 American Association for Cancer Research.
THE CELL CYCLE
HeLa cells is regulated as advertised. Perhaps, normal cells,
i.e., cells with growth controls, have different ways of doing
things ... ; (d) if what's going on in HeLa cells is perhaps more
the exception than the rule, can you imagine making the oocyte
a model of cell cycle regulation? A cell already full of all the
RNAs and proteins that are necessary for cell growth? A cell
that, fertilized, can grow up to the stage of gastrula without
synthesizing rRNA? Try to inhibit rRNA synthesis in somatic
cells: they stop cold; they won't even go through one cycle of
division. Again, what has been described in oocytes is wonderful
work, but it applies to oocytes. Its extrapolation to somatic
cells is premature.
Conclusion: for 30 years I have taught my postdoctoral
fellows and graduate students that what we find in our cells
applies only to our cells, under the conditions we used, at least
until otherwise proved. It is a maxim that should have more
general currency.
Myth 3: If Cell Proliferation Is Inhibited by an Antibody or an
Antisense RNA to a Certain Gene Product, That Gene Product
is Growth Regulatory
This statement is either a myth or a truth, depending on how
we define growth regulatory.
By now there are quite a few reports in the literature that
microinjected antibodies, or antisense RNA, or antisense oligodeoxynucleotides can block cellular proliferation if targeted
to an appropriate gene product. A sample of such gene products
includes c-ras, c-myc, c-fos, cdc2, c-myb, and the proliferating
cell nuclear antigen. Let us consider PCNA2 (9). The question
is, "Is PCNA growth regulatory?" The answer is yes and maybe,
depending on our definitions. PCNA is a cofactor of DNA
polymerase 5, which is necessary for cellular DNA synthesis. It
is fair to say that PCNA (and c-fos, and c-myc, etc.) is necessary
for cellular proliferation and, in this respect, it is growth regu
latory. However, under this broad definition of growth regula
tory, many other gene products would be growth regulatory (for
instance, all the genes necessary for the synthesis of nonessential amino acids), indeed, all kinds of things, like ATP and
anything else that is essential for the life processes of the cell.
Lower the intracellular concentration of ATP, and the cell will
not divide. There is nothing wrong in calling growth regulatory
so many genes and molecules, except that the informational
content of the word then becomes almost trivial.
I suggest the following distinction: when we show that an
antibody, or an antisense RNA, or an antisense oligodeoxynucleotide targeted to an appropriate gene product blocks cell
cycle progression, we should say that particular gene is required
for cellular proliferation. We should reserve the term "growth
regulatory" only to those genes that actually induce cellular
proliferation in quiescent cells. Unfortunately, under this defi
nition, the best examples are not cellular genes. Microinjection
of SV40 T-antigen or adenovirus EIA protein cause quiescent
cells to enter DNA synthesis or divide (6), without any other
manipulation (growth factors, for instance). At present, there
is no single cellular gene that, microinjected or transfected into
quiescent cells, causes them to divide. Cyclin does that in
oocytes (10) but, as mentioned above, that is a special case not
applicable to somatic cells. The closest thing to the SV40 Tantigen is c-myc, when microinjected into quiescent 3T3 cells,
it induces DNA synthesis (11) but only if insulin-like growth
factor 1 is also added. It suggests that somatic cells may need
2The abbreviations used are: PCNA, proliferating cell nuclear antigen; EOF.
epidermal growth factor.
more than one growth-regulatory gene, in other words, that Tantigen and the EIA protein unite, in a single gene product,
information that, in the normal cell, is divided between two or
more genes. I would be willing, in the meantime, to bestow on
c-myc the title of (incomplete) growth-regulatory gene. Inciden
tally, \-onc do not count; what we are after here are normal,
cellular genes that regulate normal cell growth.
Where does this leave the oncogenes (besides c-myc)1! As
already mentioned above, protooncogenes, when suitably mod
ified, can alter the regulation of cell growth. If they only are
required, why would a mutation (for example) cause them to
drive cellular proliferation abnormally? An explanation may be
found in the following illustration: 32D myeloid cells do not
have EGF receptors, are not stimulated by EGF, and have an
absolute requirement for interleukin 3. However, if they carry
the \-erb-B gene, which is a truncated, permanently activated
EGF receptor, 32D myeloid cells can grow in the absence of
interleukin 3 or other growth factors (12). It is as if their cell
cycle had been short-circuited. Perhaps, in the case of protoon
cogenes, their constitutive activation causes cells to progress
through the cell cycle in the absence of those gene products
that in normal cells are required for their activation. This seems
to be the case with viral oncogenes: SV40 T-antigen and aden
ovirus EIA are known to activate G.-S boundary genes (like
TK, PCNA, etc.) in the absence of cellular products that are
normally required by serum-stimulated cells; e.g., the products
of early growth-regulated genes (13).
Myth 4: The Time of Appearance in the Cell Cycle of a mRNA
or a Protein Is an Indication of the Time in the Cell Cycle at
Which That Gene Product Is Required
For instance, since ribonucleotide reducÃ-aseactivity reaches
a peak at 50 h after partial hepatectomy, that must be the time
when ribonucleotide reducÃ-aseis mostly needed. Unfortunately
for the myth, ribonucleotide reducÃ-ase is needed for DNA
synthesis, and, by 50 h after parlial hepaleclomy, DNA synthesis in the regenerating liver has ceased (6).
This is a harmless myth, although il may mislead us in our
search for a funclion. Il is especially misleading in ihe case of
mRNA levels. Somelimes, il is proposed lhal cell cycle pro
gression requires Ihe orderly expression of differenl genes, using
mRNA levels as Ihe indicalors of such orderly progression.
There may or ihere may noi be an orderly sequence of evenls,
bul we cannol support it on ihe basis of mRNA levels. Take,
for inslance, c-fos and c-myc. Al Ihe slarling gale of Gu, c-fos
beals c-myc by aboul 1 h; i.e., c-fos mRNA levels increase l h
earlier lhan c-myc mRNA levels. The finding has no heurislic
conlenl: (a) because cycloheximide experimenls unequivocally
show lhat the c-myc gene is a primary responder thai does not
need the c-fos protein or any other de novo synthesized prolein
for aclivalion and (b) because whal we see on Northern blots is
only what is delectable on Northern blots. Who can say that
there is no c-myc mRNA in the first 20 min after a mitogenic
slimulus? All we can say is lhal it's not deteclable under ihe
condilions used. It's possible thai Ihere is enough lo make
sufficienl myc prolein for whalever its function is. The point is
lhal ihe ¡nlensily of a band on a Northern blol depends on
several faclors, one of which is exquisilely biological, i.e., Ihe
number of mRNA copies per cell al any given lime.
Conclusion: sometimes there is a correlation between time of
appearance of a gene product and its function in the cell cycle,
but it is not a strict correlation and, in some cases, it can even
be misleading.
6770
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1990 American Association for Cancer Research.
THE CELL CYCLE
Myth 5: Cell Cycle Progression Is Simply Regulated by the
Degree of Phosphorylation of Certain Key Proteins
This myth has received a lot of publicity by the fact that two
very important genes, the p34cdc2gene (14) and the retinoblastoma (RB) gene, are regulated in that way. In both cases,
amounts of RNA and protein are essentially constant through
out the cell cycle; what varies is the degree of phosphorylation
of the protein. There is no question in anybody's mind that
these two genes are very important. The evidence that the
p34cdc2gene is required from the transition from S to M is
especially convincing (14, 15); it seems that the binding of its
product to other key proteins depends on its degree of phos
phorylation. However, it phosphorylates also at the G,-S
boundary, and the evidence that it is also needed there is poor:
indeed, the same antibody to the p34cdc2 gene product that,
microinjected, inhibits S to M transition has no effect on the
GI to S transition. The RB gene almost certainly plays a role
in retinoblastoma and certain osteosarcomas. There is no evi
dence, however, that it plays any role in the control of cell cycle
progression in, for instance, fibroblasts or lymphocytes, al
though, even in these cells, the degree of phosphorylation varies
with the cell cycle. This latter finding is by no means evidence
that the RB gene controls cell proliferation in fibroblasts or
lymphocytes. Indeed, Bookstein et al. (16), in a seminal paper,
have shown that, in cells carrying a mutant RB gene, the
introduction of a normal one abolishes their tumorigencity
without altering their growth characteristics in culture (inciden
tally, this takes care of another myth, that transformation and
cell proliferation are regulated by the same genes. There are
things in common between the two, but there also differences).
Once could actually be impertinent and remind the reader of
what a professor of biochemistry used to tell his students: that
proteins can be divided into two classes, the ones that are
phosphorylated and those that have not yet been studied. Even
conceding that some proteins are selectively phosphorylated
during the cell cycle, there are many that do so, including DNA
polymerase a, RNA polymerase I and RNA polymerase II.
Clearly, if one uses phosphorylation as the sole criterion, there
are many candidates for a controlling role in the cell cycle. This
caution is shared by Horowitz et al. (17) who clearly state their
doubts that the RB gene may not be critical for growth regula
tion in many of the cell types in which it is expressed. Their
concluding statement is a model of caution and salutary skep
ticism: "Since the regulatory pathways that constrain cell pro
liferation are poorly understood, we are still many years away
from directly addressing these possibilities."
The same comments, of course, apply to any other posttranslational modification of proteins; per se, it is not proof that
protein modifications control cell cycle progression. The only
valid criteria remain those discussed under Myth 3.
Concluding Remarks
I would like to propose that we reserve the term of growthregulatory genes only to those that, like the SV40 T-antigen or
the adenovirus EIA protein, can induce cellular proliferation
in quiescent cells. Cyclin, therefore, is growth regulatory in
oocytes (10). But this brings us promptly to the second recom
mendation, that different cells may utilize different genes. Too
often, investigators forget this diversity in the regulation of
cellular proliferation and, in doing so, we miss not only the
differences but also the commonalities that could serve as a
clue to the identification of the fundamental mechanism(s). It
is possible (but not yet demonstrated) that somatic cells may
need two or more gene products for growth regulation and
many others for growth (the genes I would like to define as
genes required for cell cycle progression). Personally, I picture
the cell cycle as having three distinct critical steps: Step 1, the
transition from G() to G, (c-wyc?); Step 2, the transition from
G! to S (c-mybl); and Step 3, the transition from S to mitosis
(p34"k2, cyclins, c-mos), each of them requiring several gene
products. Clearly, my view of the cell cycle includes Gn, an
arbitrary decision, which is hotly contested these days by some
colleagues, who would like to exclude the G0 state from the
definition of the cell cycle, also an arbitrary decision. The
difference between growth-regulated genes and cell cycle-regu
lated genes, advocated these days by several groups, seems to
me an unnecessary separation of two processes that are too
closely intertwined to be separated. Again, we should not take
HeLa cells as models of cellular behavior. But to return to my
three critical steps: it is, of course, just an opinion. Now, to
avoid the creation of another myth, I have to promise (and the
readers must remember) not to say in the next paper: the cell
cycle is regulated by three critical steps ....
References
1. Pardee, A. B. G, events and regulation of cell proliferation. Science (Wash
ington, DC). 246: 603-608. 1989.
1. Kelly. K.. Cochran. B. H.. Stiles. C. D., and Leder, P. Cell-specific regulation
of the c-myb gene by lymphocyte mitogens and platelet-derived growth factor.
Cell. 35: 603-610, 1983.
3. Greenberg, M. E.. and Ziff. E. B. Stimulation of 3T3 cells induces transcrip
tion of the c-fos proto-oncogene. Nature (Lond.), 311: 433-438, 1984.
4. Studzinski, G. P. Oncogenes, growth and the cell cycle: an overview. Cell
Tissue Kinet., 22: 405-424. 1989.
5. Morgan, J. I., Cohen. D. R.. Hempstead, J. L., and Curran. T. Mapping
patterns of c-fos expression in the central nervous system after seizure.
Science (Washington. DC). 237: 192-197, 1987.
6. Baserga, R. The Biology of Cell Reproduction. 251 pp. Cambridge. MA:
Harvard University Press, 1985.
7. Gewirtz, A. M., and Calabretta, B. A c-myb antisense oligodeoxynucleotide
inhibits normal human hematopoiesis in vitro. Science (Washington, DC),
242: 1303-1306, 1988.
8. Clement, A., Campisi, J.. Farmer, S. R.. and Brody. J. S. Constitutive
expression of growth-related mRNAs in proliferating and non-proliferating
lung epithelial cells in primary culture: evidence for growth-dependent translational control. Proc. Nati. Acad. Sci. USA, 87: 318-322, 1990.
9. Jaskulski. D., deRiel, J. K.. Mercer. W. E., Calabretta, B., and Baserga, R.
Inhibition of cellular proliferation by antisense oligodeoxynucleotides to
PCNA cyclin. Science (Washington, DC). 240: 1544-1546, 1988.
Murray. A. M., and Kirschner. M. W. Dominoes and clocks: the union of
two views of the cell cycle. Science (Washington, DC), 246: 614-621, 1989.
Kaczmarek, L., Hyland, J. K., Watt, R., Rosenberg, M.. and Baserga, R.
Microinjected c-myb as a competence factor. Science (Washington, DC),
228: 1313-1315, 1985.
12. Pierce, J. H., Ruggiero, M., Fleming, T. P., DiFiore. P. P., Greenberger, J.
S., Varticovski, L., Schlessinger, J., Rovera, G., and Aaronson. S. A. Signal
transduction through the EGF receptor transfected in lL-3-dependent hematopoietic cells. Science (Washington. DC). 239: 628-631. 1988.
13. Liu. H. T.. Baserga. R.. and Mercer, W. E. Adenovirus type 2 activates cell
cycle-dependent genes that are subset of those activated by serum. Mol. Cell.
Biol.. J.-2936-2942, 1985.
14. Draetta, G.. Pianica-Worms. H., Morrison, D., Drunker. B.. Roberts, T.,
and Beach, D. Human cdc2 protein kinase is a major cell-cycle regulated
tyrosine kinase substrate. Nature (Lond.), 336: 738-744, 1988.
15. Nurse, P. Universal control mechanism regulating onset of M-phase. Nature
(Lond.), 344: 503-508. 1990.
16. Bookstein, R., Shew. J. Y., Chen, P. L., Scully. P., and Lee, W. H. Suppres
sion of tumorigenicity of human prostate carcinoma cells by replacing a
mutated RB gene. Science (Washington, DC), 343: 712-715, 1990.
17. Horowitz, J. M., Park. S. H.. Bogenmann, E., Cheng, J. C.. Yandell, D. W.,
Kaye, F. J.. Minna, J. D., Dryja, T. P.. and Weinberg, R. A. Frequent
inactivation of the retinoblastoma anti-oncogene is restricted to a subset of
human tumor cells. Proc. Nati. Acad. Sci. USA. 87: 2775-2779. 1990.
6771
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1990 American Association for Cancer Research.
The Cell Cycle: Myths and Realities
Renato Baserga
Cancer Res 1990;50:6769-6771.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/50/21/6769.citation
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1990 American Association for Cancer Research.