Cancer

Perspective Research

The Death of the Cancer Cell

Carlos Sonnenschein and Ana M. Soto

Abstract
For a century, the perception that there are qualitative differences between a normal cell and a cell
belonging to a tumor has dominated discussions aimed at explaining cancer. However, an analysis of the
experimental evidence suggests that individual normal cells and individual cancer cells share the same two
fundamental behavioral properties, namely, proliferation and motility. Each individual cancer cell carries no
recognizable molecules or structures that make them consistently distinguishable from normal cells. Herein,
we argue that the differences between normal and cancerous states are instead identifiable at the tissue level
of biological organization, and therefore, the search for identification of a cancer cell should be abandoned.
Cancer Res; 71(13); 4334–7. Ó2011 AACR.

Introduction which in turn would enhance the neoplastic phenotype of the

epithelial mutated cell.
For the past century, the question "what is the difference In the 2% of germ line–inherited neoplasms, a mutation is
between normal and cancerous tissues and cells?" has been of present in all the cells of the organism suggesting that it plays
interest to experimental biologists, medical practitioners, the a causal role by providing a field effect. Although much can be
media, and the public at large. The implication of this quest said about them, we are not dealing here with this subset of
has been that, indeed, there are objective differences between cancers. Instead, given that about 98% of clinically detected
a normal and a cancer cell whereby the latter has been able to cancers are of the "sporadic" variety, and of these, 90% are
"create" a qualitative novelty distinguishable from that of the categorized as carcinomas, under the aegis of the SMT, the
repertoire of a normal cell. Here, based on a critical review of epithelial cancer cell becomes the cell that is the target of
the cancer research literature, we argue against the merits of carcinogens and therefore the one in which those assumed
keeping the concept of "the cancer cell" alive. differences should eventually become obvious. Thus, discrete
qualitative differences such as its shape, its size, its tinctorial
Background properties, and the way it divides and/or moves should
distinguish it from those of a normal counterpart.
The currently prevalent somatic mutation theory of carci- Despite redoubled efforts to find those differences, they
nogenesis and metastases (SMT) explicitly assumes that can- have remained elusive for over a century now. Shortly after an
cer is a disease centered at the cellular level of biological alleged singularity in a putative cancer cell is proposed, the
organization (1). It further claims that cancer represents a same singularity is found in normal cells during the course of
problem of control of cell proliferation and/or cell differentia- the normal development of the organism to which they
tion. In a recent extended update of their 2000 evaluation of belong. The repeated failure to identify those anticipated
what they called the Hallmarks of Cancer, Hanahan and singularities might be a reflection of the adoption by research-
Weinberg reinforced this interpretation from the selected ers of mistaken premises about the level of biological orga-
data they reviewed (2) while making more explicit the parti- nization at which cancer originates. This is the gist of what, in
cipation of the microenvironment in carcinogenesis. However, 1962, Smithers concluded in a commentary entitled "An attack
in their view, the role of the microenvironment is subservient on cytologism" (3) in which he argued that cancer ought to be
to that of the original mutated cancer cell. In other words, studied as a problem of organismal disorganization and not as
consistent with their previous publications, they still unam- a cell-based disease. Smithers’ criticisms and similar ones that
biguously claim that cancer cells recruit stromal cell types preceded and followed his have been ineffective so far in
discrediting the majority view that cancer is a cell-based
disease. In fact, on the one hand, considerable resources
are being invested to finally link those elusive mutational
Authors' Affiliation: Department of Anatomy and Cellular Biology, Tufts
University School of Medicine, Boston, Massachusetts events in that single cancer cell with cancer phenotypes to
Corresponding Author: Ana M. Soto, Department of Anatomy and Cel- vindicate the SMT, whereas on the other, ad hoc options are
lular Biology, Tufts University, 136 Harrison Ave Boston, MA 02111. being offered to reconcile the SMT with the undeniable role
Phone: (617) 636-6954; Fax: (617) 636-3971; E-mail: ana.soto@tufts.edu played by the microenvironment on the putative single cell
doi: 10.1158/0008-5472.CAN-11-0639 that by accumulating an undetermined number of mutations
Ó2011 American Association for Cancer Research. should morph into the cancer cell.

4334 Cancer Res; 71(13) July 1, 2011

 An Obituary to the Cancer Cell

To circumvent the heterogeneity of tumors, microdissec- engineer but a tinker (14), the implication being that with
tion techniques have been used to collect tumor cells which the emergence of multicellularity, the default state of cells
then were processed for whole-exome sequencing to identify must have remained unaltered and that the novelty in
mutations aimed at aiding in patient prognosis and for controlling cell proliferation was the appearance of cell
prioritizing patients for cancer treatments (4). Others are proliferation inhibitors (11). After all, the cell-cycle machin-
now using ever more sophisticated molecular biology and ery charged with generating 2 daughter cells has remained
computational techniques that allow for the analysis of the virtually unchanged from unicellular eukaryotes to cells in
genome of single cells. From these data, inferences were Metazoa (9). Thus, a plausible conclusion drawn from the
drawn about the type and number of mutations that, puta- available data could surmise that quiescence in Metazoa is
tively, are causally responsible for the carcinogenic devel- actively induced and/or inducible (7, 12, 13)
opment and tumor metastases (5). These contributions Like proliferation, motility is a constitutive property of all
which still argue for a cell-based origin of carcinogenesis cells and, therefore, it can only be inhibited (15). Cells from
deserve a much more detailed, critical analysis than the one the 3 embryonic layers exercise motility during early devel-
offered here. Notwithstanding, these strategies are applied opment. During postnatal life some cells, such as the wan-
to fully developed neoplasms, and thus, they are unlikely to dering cells in the connective tissue also move. Even normal
shed light on the events that initiate carcinogenesis (6). epithelial cells stream at variable speed from their "birth" to
Theodor Boveri, the originator of the SMT, stated almost 1 their "death" locations (16). In addition, cells move following
century ago that we are limited by the impossibility of disruptions in tissue architecture generated by wounds.
observing a neoplasm in statu nascendi (7). This assertion During incipient and advanced states of carcinogenesis,
remains true today. times at which metastases are generated, cells also move.
Thus, motility in cancer cells does not represent a newly
Cell Proliferation and Motility during acquired property but the restoration of an ancestral, intrin-
Development and Carcinogenesis sic cellular condition.
As an alternative to the prevailing cell-based perspective
Despite the crucial importance that the control of cell adopted by the SMT, some researchers have concluded that
proliferation has in normal development and in carcinogen- carcinogenesis is equivalent to development gone awry (17).
esis, its role in these processes remains a controversial topic, From this perspective, normal and cancer development would
probably because of the different meaning that "the control belong to the tissue level of biological organization (18). In the
of cell proliferation" elicits in the mind of researchers as most frequent classes of epithelial neoplasms, that is carci-
reflected in the literature. When we refer to the control of nomas, the parenchyma is an epithelium. The repeated
cell proliferation, we mean to address the question—Why demonstrations of the reversibility of the neoplastic pheno-
would a cell enter the cycle to generate 2 daughter cells? type when parenchymal cells isolated from neoplastic tissues
Instead, the great majority of researchers in this field refer to are placed in normal ones points to the contextuality of
"the control of cell proliferation" by dealing with the ques- cellular phenotypes, and thus to the lack of intrinsic differ-
tion—Which are the myriads of biochemical and biomecha- ences between individual normal and cancer cells (19–22).
nical steps that a cell has to conduct to generate 2 daughter Thus, cancer would be the result of a faulty reciprocal inter-
cells? This latter view would be equivalent to asking the action between the parenchyma and stroma.
question, how does a cell proliferate? The current majority
view as represented in textbooks and commentary posits What Went Wrong?
that proliferation is the default state of prokaryotes and
unicellular eukaryotes but that quiescence is, instead, the During the second half of the 19th century, German pathol-
default state in animal cells (8–10).1 However, data collected ogists diagnosed cancers while using light microscopes, com-
using hormone target cells (11), stem cells (12), and lym- parable to those used today. They then suggested that cancer
phocytes (13), as well as evolutionary theory consistently was a tissue-based disease and, separately, that the default
favor the interpretation that proliferation is the default state state of cells was proliferation (11). At the beginning of the 20th
of all living cells (7). century, the advent of gene-centric research imperceptibly
Historical precedents help us in sorting out who proposed allowed the introduction of a reductionist rationale that
what, when, and whose arguments, regardless of their merit, encouraged finding phenotypic differences between the nor-
prevailed during earlier periods. The narrative of the pio- mal and the cancer cell (23). Ever since, this search was
neers of tissue culture techniques at the beginning of the conducted by adopting a "bottom-up" strategy. The more that
20th century was instrumental in introducing the distorted was learned about the structural complexity of the cell, the
view, later adopted by those siding with the SMT, that more distant the perceived "bottom" became. Thus, while
quiescence was the default state of cells in Metazoa (7). In Theodor Boveri was proposing the theory that cancer was
this context, François Jacob noted that nature is not an due to chromatin rearrangements within a single cell (7),
starting in the 1960s until today, genomic somatic mutations
1
(point mutations, deletions, translocations, etc.) and so-called
By default state, an evolutionarily relevant subject, it is meant the state at
which a cell will proliferate, or remain quiescent, in the presence of excess epigenetic changes (DNA methylation and posttranslational
nutrients. modifications of histones) in that fateful original cancer cell

www.aacrjournals.org Cancer Res; 71(13) July 1, 2011 4335

 Sonnenschein and Soto

became the target of the increasingly elusive explanatory tic, signals, etc.). More often than not, recent tendencies in
"bottom" (6;24;25). this regard require that, in addition to the above-referred
A number of circumstances conspired against accepting SMT premises, carcinogenesis include the heavy involve-
the evolutionary relevant premises implied by the German ment of the tissue components surrounding the cancer cell
pathologists (23). Regardless, these scientific miscues niche (2).
resulted in both positive and negative outcomes. Among Finally, one wonders whether alternative premises and
the positive ones can be cited the staggering amount of research programs to those offered by advocates of the
knowledge accumulated at the cellular level of biological SMT might now be appropriate when dealing with approaches
organization which includes all the biochemical and struc- to cancer research (30) and its applications to the clinic and
tural descriptions of cell types, their organelles, their tran- public health at large (29).
scriptional and translational quirks, ever expanding signal What kind of change can be anticipated if a paradigm
transduction pathways and networks. Explicitly, one of the switch is adopted? The novel approach in explaining carci-
declared aims of this decades-old strategy was the identi- nogenesis is anchored on 2 distinct premises: the first posits
fication of those anticipated differences between the normal that proliferation is the default state of all cells and the second
and the cancer cell. That much can be gleaned from a proposes that cancer is a tissue-based disease. These are the
comment by the famed British scientist John Cairns, who premises of the Tissue Organization Field Theory (7, 11).
stated, What consequences this proposed switch may have on the
overall agenda to unravel the cancer puzzle and on improving
. . .Biology and cancer research have developed together.
the outlook for more rationale management of cancer
Invariably, at each stage, the characteristics of the cancer cell
patients? In the important realm of experimental cancer
have been ascribed to some defect in whatever branch of
research, the proposed change would mean a switch from a
biology happens at the time to be fashionable and exciting;
subcellular, gene-centric approach to a tissue-based organicist
today, it is molecular genetics. (26).
one, in which a combined top-down and bottom-up strategy
Despite the monumental effort already displayed, so far, would include systems biology components (18;31). From a
no qualitative differences have been described between a public health standpoint, one would anticipate a change in
normal and a neoplastic cell, an assertion now extended to cancer public policy that would highlight the underappre-
the presence of somatic mutations (27) and aneuploidy in ciated role of the environment in the causation of cancers
normal cells in vivo (28). Nevertheless, judging by the headings and the importance of prevention of exposure (32). In the long
in the table of contents of most cancer research journals and run, this conceptual and experimental switch might generate
textbooks on the biology of cancer, the search for those a more rewarding return on investment than the one based
distinctive differences between normal and cancer cells con- on the century-old SMT.
tinues unabated (8).
Conclusion
Is It Time to Consider Alternatives?
Taking into account the century-old failure to pragmatically
An animated controversy among cancer researchers is characterize features in a cancer cell that would identify it
raging on whether the premises they adopt to collect and from a normal one, it would be now realistic to finally declare
interpret cancer data determine the soundness of the con- the formal death of "the cancer cell." The interpretation of the
clusions they draw. Supporters of the SMT respond to this staggering amount of data collected during this period is
controversy by highlighting the anticipated promises of consistent with this suggestion which concurs with that
genomic research through a massive sequencing effort of already made about 5 decades ago by Smithers (3) and by
the genome of thousands of cancers and the "mining" of vast others since (7, 17).
data sets (6). This gargantuan effort is overtly aimed at
improving the diagnosis, prognosis, and treatment of can- Disclosure of Potential Conflicts of Interest
cers through what is called translational patient-targeted
research. However, a rigorous analysis of these data reveal No potential conflicts of interest were disclosed.
wide gaps in explaining carcinogenesis and on prospects of
clinical relevance (29). Acknowledgments
Returning to the seminal argument by the SMT of whether
one or many mutations in the genome of a single epithelial The authors are grateful to Cheryl Schaeberle, Michael Askenase, and
cell is responsible for the carcinogenic process, those siding Andrew Tharp for their excellent editorial assistance.
with this theory are admitting either by commission or by
omission that no such mutations have been identified so far. Grant Support
Instead, they have offered ad hoc scenarios (i.e., that muta-
tions are instead lodged in neighboring stromal cells, that This work was supported in part by grants from The Parsemus Foundation,
the Avon Foundation, and the NIH (ES0150182, ES012301, and ES08314).
the microenvironment sends either "wrong" signals to said
cell, or that the prospective cancer cell lacks the ability to Received February 22, 2011; revised April 11, 2011; accepted April 11, 2011;
read correctly those at times stimulatory, at others apopto- published OnlineFirst April 20, 2011.

4336 Cancer Res; 71(13) July 1, 2011 Cancer Research

 An Obituary to the Cancer Cell

References
1. Weinberg RA. One renegade cell: how cancer begins. New York: Basic 19. Illmensee K, Mintz B. Totipotency and normal differentiation of single
Books; 1998. teratocarcinoma cell cloned by injection into blastocysts. Proc Nat
2. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Acad Sci U S A 1976;73:549–53.
Cell 2011;144:646–74. 20. Maffini MV, Calabro JM, Soto AM, Sonnenschein C. Stromal regula-
3. Smithers DW. Cancer: an attack of cytologism. Lancet 1962;1:493–9. tion of neoplastic development: age-dependent normalization of
4. Jiao Y, Shi C, Edil BH, de Wilde RF, Klimstra DS, Maitra A, et al. DAXX/ neoplastic mammary cells by mammary stroma. Am J Pathol
ATRX, MEN1, and mTOR pathway genes are frequently altered in 2005;67:1405–10.
pancreatic neuroendocrine tumors. Science 2011;331:1199–203. 21. Hendrix MJ, Seftor EA, Seftor RE, Kasemeier-Kulesa J, Kulesa PM,
5. Navin N, Kendall J, Troge J, Andrews P, Rodgers L, McIndoo J, et al. Postovit LM. Reprogramming metastatic tumour cells with embryonic
Tumour evolution inferred by single-cell sequencing. Nature microenvironments. Nat Rev Cancer 2007;7:246–55.
2011;472:90–4. 22. Kenny PA, Bissell MJ. Tumor reversion: correction of malignant
6. Stratton MR. Exploring the genomes of cancer cells: progress and behavior by microenvironmental cues. Int J Cancer 2005;113:168–70.
promise. Science 2011;331:1553–8. 23. Moss L. What genes can't do. Cambridge, MA: MIT Press; 2003.
7. Sonnenschein C, Soto AM. The society of cells: cancer and control of 24. Sharma S, Kelly TK, Jones PA. Epigenetics in cancer. Carcinogenesis
cell proliferation. New York: Springer Verlag; 1999. 2010;31:27–36.
8. Weinberg RA. The biology of cancer. New York: Taylor & Francis; 25. Chi P, Allis CD, Wang GG. Covalent histone modifications–miswritten,
2006. misinterpreted and mis-erased in human cancers. Nat Rev Cancer
9. Alberts B, Johnson A, Lewis J, Raff M, Roberts K, Walter P. Molecular 2010;10:457–69.
biology of the cell. 5th ed. London, UK: Garland Science; 2008. 26. Cairns J. Matters of life and death. Princeton, NJ: Princeton University
10. Alberts B. Model organisms and human health. Science 2010; Press; 1997.
330:1724. 27. Martin GM, Ogburn CE, Colgin LM, Gown AM, Edland SD, Monnat RJ
11. Sonnenschein C, Soto AM. Theories of carcinogenesis: an emerging Jr. Somatic mutations are frequent and increase with age in human
perspective. Semin Cancer Biol 2008;18:372–7. kidney epithelial cells. Hum Mol Genet 1996;5:215–21.
12. Ying QL, Wray J, Nichols J, Batlle-Morera L, Doble B, Woodgett J, 28. Rehen SK, McConnell MJ, Kaushal D, Kingsbury MA, Yang AH, Chun
et al. The ground state of embryonic stem cell self-renewal. Nature J. Chromosomal variation in neurons of the developing and adult
2008;453:519–23. mammalian nervous system. Proc Natl Acad Sci U S A 2001;
13. Yusuf I, Fruman DA. Regulation of quiescence in lymphocytes. Trends 98:13361–6.
Immunol 2003;24:380–6. 29. Ioannidis JP. Personalized genetic prediction: too limited, too expen-
14. Jacob F. Evolution and tinkering. Science 1977;196:1161–6. sive, or too soon? Ann Intern Med 2009;150:139–41.
15. Buss LW. The evolution of individuality. Princeton, NJ: Princeton 30. Bizzarri M, Cucina A, Conti F, D’Anselmi F. Beyond the oncogene
University Press; 1987. paradigm: understanding complexity in cancerogenesis. Acta
16. Zajicek G, Oren R, Weinreb M. The streaming liver. Liver 1985;5:293– Biotheor 2008;56:173–96.
300. 31. Soto AM, Sonnenschein C, Miquel P-A. On physicalism and down-
17. Pierce GB, Shikes R, Fink LM. Cancer: a problem of developmental ward causation in developmental and cancer biology. Acta Biotheor
biology. Englewoods Cliffs, NJ: Prentice-Hall; 1978. 2008;56:257–74.
18. Soto AM, Sonnenschein C. Emergentism as a default: cancer as a 32. Chistiani DC. Combating environmental causes of cancer. N Engl J
problem of tissue organization. J Biosci 2005;30:103–18. Med 2011;364:791–3.

www.aacrjournals.org Cancer Res; 71(13) July 1, 2011 4337