Medical Hypotheses (2008) 71, 126–140
http://intl.elsevierhealth.com/journals/mehy
Defining the steps that lead to cancer: Replicative
telomere erosion, aneuploidy and an epigenetic
maturation arrest of tissue stem cells
Reinhard Stindl *,1
Department of Molecular and Cell Biology, 353 Donner Hall, University of California at Berkeley,
Berkeley, CA 94720-3206, USA
Alpharm GesmbH, Sebastian-Kneippgasse 5-7, 2380 Perchtoldsdorf, Austria
Received 4 January 2008; accepted 7 January 2008
Summary
Recently, an influential sequencing study found that more than 1700 genes had non-silent mutations in either
a breast or colorectal cancer, out of just 11 breast and 11 colorectal tumor samples. This is not surprising given the fact
that genomic instability is the hallmark of cancer cells. The plethora of genomic alterations found in every carcinoma does
not obey the ‘law of genotype–phenotype correlation’, since the same histological subtype of cancer harbors different
gene mutations and chromosomal aberrations in every patient. In an attempt to make sense out of the observed genetic
and chromosomal chaos in cancer, I propose a cascade model. According to this model, tissue regeneration depends on the
proliferation and serial activation of stem cells. Replicative telomere erosion limits the proliferative life span of adult
stem cells and results in the Hayflick limit (M1). However, local tissue exhaustion or old age might promote the activation
of M1-deficient tissue stem cells. Extended proliferation of these cells leads to telomere-driven chromosomal instability
and aneuploidy (abnormal balance of chromosomes and/or chromosome material). Several of the aforementioned steps
have been already described in the literature. However, in contrast to common theories, it is proposed here that the
genomic damage blocks the epigenetic differentiation switch. As a result of aneuploidy, differentiation-specific genes
cannot be activated by modification of methylation patterns. Consequently, the phenotype of cancer tissue is largely
determined by the epigenetic maturation arrest of tissue stem cells, which in addition enables a fraction of cancer cells to
proliferate, invade and metastasize, as normal adult stem cells do. The new model combines genetic and epigenetic
alterations of cancer cells in one causative cascade and offers an explanation for why identical histologic cancer types
harbor a confusing variety of chromosomal and gene aberrations. The Viennese Cascade, as presented here, may end the
debate on if and how ‘tumor-unspecific’ aneuploidy leads to cancer.
c 2008 Elsevier Ltd. All rights reserved.
Introduction
* Address: Department of Molecular and Cell Biology, 353
Donner Hall, University of California at Berkeley, Berkeley, CA
In the US today, the lifetime probability of develop-
94720-3206, USA. Tel.: +1 510 642 6549; fax: +1 510 642 6455.
ing cancer is 46% for men and 38% for women [1],
E-mail address: reinhard_stindl@yahoo.de
1 Tel.: +43 1 8657526; fax: +43 1 8697252.
and the mortality rate from cancer has changed
0306-9877/$ - see front matter 
c 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mehy.2008.01.010

---

The Viennese Cascade
127
very little over the past 50 years [2,3]. Carcinomas,
information necessary for malignant expression
malignant epithelial tumors, account for 80% of
[27]. Yet, certain characteristics of malignancy
cancer-related deaths in the western world [4].
are clearly not physiological (e.g. anaphase bridges
Therefore, the focus of this paper is on epithelial
and atypical mitoses), but can be attributed to crit-
cancer that currently represents an ‘epidemic’ in
ically short telomeres and aneuploidy [28]. Aneu-
the middle-aged and elderly population.
ploidy usually means an abnormal number of
In the past, enthusiasm for tumor cytogenetics
chromosomes, deviating from euploidy. In recent
gradually faded due to a confusing plethora of
literature, this term has expanded to include all
chromosomal aberrations found in human cancers
unbalanced chromosomal aberrations, including
(reviewed in Ref. [5]). As a consequence, the cause
both structural and numerical defects.
of cellular transformation was thought to have a
Almost 80 years ago, Mueller and McClintock pro-
mutational basis. Ironically, molecular geneticists
posed the concept of protective caps at the end of
have been confronted with an even larger variety
eukaryotic chromosomes [29], called telomeres. In
of ‘cancer-causing’ gene mutations [2,6]. This is
1961, Hayflick reported an elegant experiment dem-
not surprising given the fact that genomic instabil-
onstrating that normal human cells are capable of a
ity is the hallmark of cancer cells [7,8]. Thousands
limited number of cell divisions [30]. Subsequently,
of genomic alterations [9] found in every cancer
Olovnikov suggested a biological explanation for this
patient do not obey the ‘law of genotype–pheno-
phenomenon, namely incomplete DNA replication
type correlation’, since the same histological sub-
[31,32]. It later became known as ‘the end-replica-
type of cancer harbors different mutations in
tion problem of linear DNA molecules’ [32]. Based
every patient [10–12]. Tumor cell heterogeneity
on experimental findings, telomeres shorten every
with no two individual cells harboring identical
time a somatic cell divides, and subsequently limit
mutations [13] has further impeded progress in
the proliferation potential of normal human cells,
the field of molecular genetics. The purpose of this
constituting the so-called mitotic clock [33]. Once
article is to present a model in which the observed
a telomere is critically short, it can no longer form
chaotic variety of chromosomal aberrations and
the functional t-loop structure [34]. Human cell pro-
gene mutations leads to similar histologic subtypes
liferation usually stops before telomeres reach this
of cancers (Fig. 1).
critically short length, a phenomenon called the
Hayflick limit, replicative senescence, or M1. If cells
are M1-deficient (e.g. via deregulation of the p53
Historical background
and RB1 pathways), proliferation continues and
short non-functional telomeres lead to chromosome
In 1867, Cohnheim hypothesized that all tissues are
end fusion and to breakage–fusion–bridge cycles
renewed by cells in the blood stream [14] and that
during cell division [29]. This results in crisis-stage,
cancer in adults develops from embryonic cells that
which is characterized by chromosomal instability
are produced in excess and remain in fully mature
and aneuploidy (Fig. 1) [35,36]. In vitro experiments
organs [14]. In retrospect, both predictions are
with simian virus 40 (SV40) genes suggest that spon-
chillingly accurate. A large body of evidence now
taneous telomerase expression may rescue some hu-
supports the idea that adult human stem cells cir-
man cells from crisis by stabilizing the telomeres
culate in the blood stream and participate in tissue
[37,38]. Thus elevated telomerase expression is
regeneration of many, if not all, organs [15–17];
thought to induce a state of cellular immortality.
but the full extent of their regeneration potential
In 1890, von Hansemann concluded that malig-
is still unknown [18]. Following the classic ‘initia-
nant tumor cells are defined by abnormal chroma-
tion–promotion’ experiments [19], many cancers
tin content [39]. More than 20 years later, Boveri
are considered a disease of stem cells, in which
suggested that tumors might arise as a conse-
(cancer) stem cells renew (tumor) tissue [20].
quence of abnormal segregation of chromosomes
The observation that only a minority of leukemia
to daughter cells. He postulated that tumor growth
or tumor cells can initiate cancer [21,22], and that
is based on a particular, incorrect chromosome
these cells display many stem cell characteristics
combination which is the cause of abnormal growth
resurrected the ‘cancer stem cell hypothesis’
characteristics passed on to daughter cells [40,41].
[23,24], originally proposed in 1961 [25,26]. In
Until quite recently, these ideas have received only
the 1970s, Pierce suggested that cancers represent
limited attention, because the primary focus of
a maturation arrest of stem cells. He stated that all
cancer genetics has been mainly on the role of indi-
of the characteristics associated with malignancy
vidual oncogenes and tumor suppressor genes
are expressed during some stage of development,
[42,43]. A somewhat collective disappointment
suggesting that the normal genome contains the
was that more than 40 years of cancer cytogenetic

---

128
Stindl
Figure 1
The Viennese Cascade.
research resulted in just a handful of tumor-spe-
ferentiation can be linked to epigenetic mecha-
cific chromosomal aberrations (reviewed in Ref.
nisms. More than 20 years ago, patterns of
[5]). Among these findings, approximately 40% of
genomic 5-methyl-cytosine levels in tissue were
published recurrent balanced aberrations were
observed [51]. Since then, it has been clearly
found to be biased [44]. A confusing combination
shown that DNA methylation represents a mecha-
of apparently meaningless abnormalities in numer-
nism for the regulation of gene expression in verte-
ous solid tumors were regarded as epiphenomena
brates [52,53]. Demethylation of so-called ‘CpG
incurred during tumor progression [5]. Despite the
islands’ in promoter regions occurs during differen-
fact that chromosomal instability is an early event
tiation and is a required step in the process of tran-
in tumorigenesis [45,46] and that there is abundant
scriptional activation [53–55].
evidence of correlation between increasing aneu-
ploidy and greater malignancy in tumors [47–49].
Indeed human adult stem cells, after acquiring
The author’s quest
aneuploidy in vitro, have been shown to transform
spontaneously [50].
Duesberg proposed that the cancer phenotype is a
Every normal somatic cell contains the same set
direct consequence of aneuploidy changing the
of chromosomes and genes, yet the human body
dosage of thousands of genes [56–58]. According
consists of hundreds of different cell types. Differ-
to him, the phenotype is the result of a geno-
entiation of distinct cell types requires transcrip-
type–phenotype correlation, just as trisomy of
tional activation of differentiation-specific genes
chromosome 21 generates Down syndrome [56].
and the suppression of genes associated with the
As a postdoctoral fellow in his laboratory (1999–
progenitor cell. Consequently, tissue-specific dif-
2002) I was skeptical about aneuploidy and its

---

The Viennese Cascade
129
Figure 2
Grossly different forms of chromosomal imbalance without any impact on the malignant phenotype: (A) A
representative karyotype of the grossly aneuploid cell line HA1ER (=human embryonic kidney cells transfected with
SV40 early region, hTERT and H-ras [64]), and (B) the pseudo-diploid subclone HA1ER-3 [59]. Both cell lines give rise to
similar numbers of colonies in soft agar [59] and result in indistinguishable phenotype and growth characteristics
in vitro. Although this is a highly artificial setting, the same is true for in vivo cancer. Similar histologic subtypes of
cancers exhibit different spectrum of genetic aberrations [11]. (Multicolor-FISH was performed according to the
supplier’s protocol (MetaSystems, Altlussheim, Germany)).
direct effect on the cancer phenotype (Fig. 2A and
esis or a coincidence due to oncogene deregulation
B) [43,59,60]. Aneuploidy is one of the most com-
[62]. In a milestone paper, Weinberg and col-
mon properties of cancers [61], but it has not been
leagues claimed that three defined genetic ele-
entirely clear whether it is essential for tumorigen-
ments (later four [63]) suffice to produce a

---

130
Stindl
human malignant tumor [59,64], providing the ulti-
can be easily recognized by a histopathologist.
mate support to the somatic mutation model. How-
Hepatocellular carcinoma cells almost always
ever, I received the supposedly diploid cells [59]
retain some characteristics of normal liver tis-
from Hahn in January 2001 and based on multi-
sue.
However
‘transdifferentiation’
should
color-FISH analysis I could not find one diploid cell
occur commonly, if the observed 11,000 muta-
in any of those after only 1 week in culture (Fig 2B)
tional hits [9,68] per cancer cell have a direct
[60]. Identical chromosomal rearrangements were
effect on the cellular phenotype. Yet, precan-
present in a large fraction (up to 89%) of the cells
cerous epithelial tissue seemingly dedifferenti-
(my unpublished letter to Cancer Research), which
ates, but it usually does not transdifferentiate.
cannot be explained by the ‘‘outgrow of aneuploid
3. Knudson’s two-hit model is based on an inherited
variants’’ (authors reply [60]).
mutation or loss of a gene (RB1) causing a rare
Ironically, large-T antigen, one of the ‘defined
form of cancer [69]. However, cancer gene muta-
genetic elements’ inserted by the Weinberg group,
tions in germ cells play only a marginal role in the
has been known to induce chromosomal instability
genesis of common cancers [70], and the vast
in human cells [38]. Several years ago, the Bacch-
majority of cancers are sporadic [70]. Yet, if
etti group actually studied large-T antigen and
mutations in a defined number of genes cause
chromosomal instability in the same embryonic kid-
cancer, and if >1% of all human genes are cancer
ney cells (HA1) [38]. Another blow to the ‘defined
genes [2,6], inherited forms of epithelial cancers
genetic elements’ model is that hyperproliferative
(=carcinomas) should occur commonly in young
tumors in transgenic mice (large-T antigen Tet on/
adults. As has been pointed out by Duesberg,
off system [65]) can only be switched off at an
inherited mutations in cancer genes would reduce
early stage. After several months of large-T antigen
the number of additional mutational hits required
(TAg)
activity,
tumors
became
irreversible,
and would significantly reduce the age of onset
although TAg expression was silenced. Most impor-
[56]. However, carcinomas display a characteris-
tantly, tumor cells were found to be polyploid [65].
tic age range with practically no cases arising in
However, a chromosome analysis was not under-
young people (Figure 11.1 of Ref. [4]).
taken in this study published in Science. The
4. Cancer
incidence
increases
exponentially
authors concluded that ‘‘. . .cumulative changes oc-
beyond the age of 40 (Fig. 3A), and this increase
cur. . .that prohibit reversal. . .The nature of these
is caused almost exclusively by epithelial can-
changes remains to be identified.’’ [65].
cers. One out of every two men and one out of
Beside the observation that carcinoma cells are
every three women develop cancer during their
always aneuploid (structurally and/or numeri-
lifetime [1]. This is a remarkably efficient path-
cally), there are several other unresolved issues
ological mechanism, which is difficult to explain
regarding the somatic gene mutation theory. Four
by the requirement of 4–5 collaborating muta-
of them are listed below:
tions in a cell [4] based on somatic mutation
rates in normal cells [8]. Therefore a mutator
1. As Duesberg pointed out in a letter to Science,
phenotype was proposed [8]. However, what
viral promoters that are used in most current
causes these highly elevated mutation rates in
gene transfection experiments result in onco-
all carcinoma patients remains a mystery.
gene expression rates of up to 100 times stron-
ger than cellular promoters [57]. Furthermore,
In an attempt to shed light on these unresolved
the expression rate of a gene under the control
issues, I present here a causative cascade. Most
of a viral promoter is stable and cannot be
of the mechanisms in this cascade have been al-
fine-tuned. Hence, the insertion of oncogenes
ready associated with cancer progression in some
with viral promoters and/or the cultivation of
way. Yet, the combination and series of events de-
human cells might occasionally result in unex-
scribed here is unique and should be considered an
pected effects on carcinogenesis [66], including
important shift in thinking. The challenge was to
induction of aneuploidy [43].
explain how the (observed) chaotic variety of chro-
2. Somatic gene mutations are thought to largely
mosome and gene aberrations [12] can lead to
determine the phenotype and the behavior of
identical cancer phenotypes.
cancer cells (genotype–phenotype correlation),
but it has been found that tumors have diverse
mutational profiles [10,67]. Thousands of muta-
The Viennese Cascade
tions per cancer cell [9,62,68] would make his-
topathology a hopeless endeavor; but this is
The human body undergoes an extensive process of
not the case. Metastatic liver cancer in the lung
tissue regeneration, replacing many billion cells

---

The Viennese Cascade
131
each day. The task is accomplished by highly regu-
CpG islands (compared to their differentiated
lated proliferation [71] and serial activation of tis-
counterparts) [87]. The Viennese Cascade provides
sue stem cells [72]. Senescent tissue stem cells are
an explanation for aberrant methylation patterns
replenished by adult stem cells circulating in the
found in most human cancers [88]. Based on the
bloodstream and residing in the bone marrow,
current view that the differentiation switch modi-
and potentially other organs [73]. Exhaustion of
fies expression patterns of several genes on many
replicative capacity and a limited supply of fresh
chromosomes [89], it is proposed here that the
stem cells may contribute to tissue aging [74]. Tel-
intactness, correct location and correct copy num-
omerase activity measured in adult stem cells is
ber of differentiation-specific genes are critical.
not sufficient to prevent telomere loss and suggests
Hence, genetic damage caused by chromosomal
that cells with insufficient telomere length undergo
instability may impair epigenetic modifications of
senescence [74,75]. This situation is in stark con-
differentiation-specific genes (Fig. 1). Due to
trast to embryonic stem cells that express high lev-
extensive aneuploidy, the demethylation machin-
els of telomerase, exhibit remarkable genomic
ery cannot activate tissue-specific patterns of gene
stability, are capable of unlimited self-renewal,
expression and the committed progenitor cell re-
and are pluripotent [74]. Under physiological con-
mains in its current differentiation state, resulting
ditions, the differentiation potential of adult stem
in epigenetic maturation arrest. Although my pro-
cells is much more limited [76], being multipotent
posed mechanism focuses mainly on methylated
at most [71]. Consequently, cells give rise to one or
DNA modifications, the perturbation of histone
multiple cell lineages of one particular tissue only
methylation/acetylation and other yet unknown
[74,77,78].
Circulating
adult
stem
cells
are
epigenetic mechanisms that are involved in differ-
recruited to a specific organ and reside there as tis-
entiation may also be causative.
sue stem cells. Subsequently, committed progeni-
Depending on the pattern and the extent of
tors are derived from tissue stem cells.
aneuploidy, immature cells might be blocked at
It is proposed here that in the elderly, the num-
different developmental stages, causing variable
ber of adult stem cells is already limited, and
degrees of malignancy. Accordingly, it has been
deregulated tissue stem cells undergo prolifera-
shown that increasing aneuploidy is mirrored by
tion. These cells are M1-deficient (e.g. have dereg-
morphological dedifferentiation [90]. For example,
ulated p53 and RB1 pathways) and continue to
advanced epithelial cancer is characterized by an
multiply beyond their limit [79], causing critically
immature phenotype and severe aneuploidy [47].
short telomeres in proliferating progenitor cells.
It is an essential part of the hypothesis that differ-
Subsequently, non-functional telomeres lead to
ent aneuploidy/mutation patterns give rise to iden-
chromosome fusion. During cell division, fused
tical phenotypes, because differentiation depends
chromosomes eventually get pulled to opposite
on the epigenetic modification of differentiation
poles (Fig. 1), resulting in anaphase bridges [28],
genes, whereas alterations in all other genes would
and eventual chromosome breakage [80]. The pro-
have no direct effect on the maturation arrest.
cess starts again with fusion of the broken parts
Therefore, different patterns of chromosomal and
(breakage–fusion–bridge cycle, Fig. 1). Break-
gene aberrations that have been found in identical
age–fusion–bridge cycles lead to unbalanced
histologic subtypes of human tumors can lead to
translocations of chromosome fragments [28,81].
identical cancer phenotypes. In addition, a partic-
In addition, dysfunctional telomeres result in tetra-
ular combination of genomic alterations may cause
ploidization [82] and losses of whole chromosomes
transformation of a particular tissue or cell type
[28]. These structural and numerical chromosome
only, since differentiation-specific genes are tis-
aberrations (=aneuploidy) affect the integrity,
sue-specific [89].
location, and copy number of hundreds or thou-
A critical step in the cascade involves telome-
sands of genes.
rase reactivation. Increased telomerase activity
Cancer cells are often characterized by a hyper-
stabilizes short telomeres, leading to decreased
methylated state of gene promoters and a de-
chromosomal instability in cancer cells [36]. It is
creased
global
level
of
DNA
methylation
tempting to suggest that the proliferation of stem
compared to normal cells [54,55,83,84]. Global
cells with unstable chromosomes begets high levels
hypomethylation seems to have no direct onco-
of telomerase activity. Telomerase activation may
genic effect [85] and might simply be associated
represent an attempt to heal stem cells with unsta-
with proliferation state [86]. Promoter hyperme-
ble chromosomes. Chromosome healing has been
thylation is thought to inactivate tumor suppressor
described to be restricted to the embryonic devel-
genes; but it is not entirely clear why many gene
opment stages in maize [91], which might explain
promoters in cancer cells contain hypermethylated
why high levels of telomerase expression occur in

---

132
Stindl
‘embryonic-like’ cancer stem cells. Unfortunately,
Discussion
the rescue mechanism sometimes fails and pro-
duces immortal ‘aneuploid monster cells’. Accord-
The Viennese Cascade combines genetic and epige-
ingly, clinical elevation of telomerase activity
netic alterations of cancer cells in one causative
positively correlates with pathological stage and
cascade (Fig. 1) and offers an explanation for how
poor prognosis in many cancer patients [92].
different patterns of gene and chromosomal aber-
Metastases recapitulate the organization of
rations can lead to identical histologic phenotypes.
their primary tumors [93], and the probability of
Several steps in this cascade have already been
metastasis suggests that nongenetic mechanisms
investigated by numerous leaders in the field. Dues-
are involved [94]. My proposed model is not based
berg resurrected the aneuploidy-cancer-theory
on mutated metastasis genes [95], since migration
[57], and subsequent papers dealt with the direct
is inherent to normal adult stem cells [96]. Sev-
effect of aneuploidy on the cancer phenotype
eral studies have suggested that tumor cell migra-
[101]. It has been recently proposed by Vogelstein
tion and metastasis are a consequence of cancer
and others that chromosomal instability (CIN) can
stem cells responding to chemokines from other
initiate tumorigenesis [46,61]. DePinho developed
tissues [97–99], as normal adult stem cells do
a theory of carcinogenesis based on telomere ero-
[78,97,100].
sion and chromosomal instability, in which transfor-
Despite the similarities between cancer stem
mation and metastasis are a consequence of
cells and normal adult stem cells, there are three
additional gene-specific mutations [36]. DePinho’s
fundamental differences:
model does not include the cancer stem cell con-
cept or the differentiation block. Pathak proposed
1. The progeny of cancer stem cells cannot fully
that a telomere erosion/aneuploidy/transforma-
differentiate. Although, in most cancers differ-
tion sequence occurs in tissue-specific stem cells
entiation occurs in a fraction of cells and tumor
[102,103]. However Pathak’s theoretical work does
tissue is characterized by a varying mixture of
not provide a mechanism to explain how aneuploidy
partly- and fully-differentiated cells with some
leads to stem cell transformation, it does not in-
characteristics of the tissue of origin. The lack
clude epigenetic aberrations, and metastasis is
of cellular differentiation is accompanied by
seen as a consequence of a series of genetic muta-
cytological evidence of short telomeres, chro-
tions and amplifications of telomeric DNA [103].
mosomal instability and aneuploidy, including
Matzke and coworkers showed that trisomy in plants
anaphase bridges and nuclear abnormalities.
can lead to increased DNA methylation in regulatory
2. Cancer progenitor cells usually do not respond to
regions and hypothesized that aneuploidy predis-
signals to stop proliferation. This might be a
poses DNA to epigenetic modifications [104]. Yet,
direct consequence of maturation arrest, or a
the Viennese Cascade proposes that aneuploidy re-
consequence of tissue cells that continue to
sults in aberrant methylation patterns because epi-
send regeneration signals (=lack of negative
genetic modifications that are essential for normal
feedback due to the production of undifferenti-
tissue differentiation cannot be performed.
ated progeny). The role of oncogene activity and
Supporters of the somatic mutation cancer theory
impaired intercellular communication might
might argue that there is no need for an alternative
also play a role.
model. However, the facts remain that the age-ad-
3. Cancer stem cells can migrate to and grow in the
justed cancer mortality rate has remained relatively
wrong organs, resulting in metastasis.
stable in the last half century and we have yet to ful-
fill the expectations for a new era of molecularly-
The metastatic potential of a tumor may depend
based cancer research and therapy [2]. Indeed, the
on the nature of the maturation arrest, which itself
utility of any individual gene therapeutic approach
is determined by the pattern and amount of genet-
may be negligible in light of the ever-growing per-
ic damage, as well as the tissue type. It has been
plexing diversity of cancer-causing gene mutations
proposed by others that tumors derived from early
[12]. The fact that every tumor harbors its own un-
tissue stem cells have an increased tendency to
ique set of cancer gene mutations [10,12] is a stag-
metastasize, whereas a maturation arrest of com-
gering
scenario
for
molecular
geneticists.
mitted progenitor cells may only result in local
Currently, there are 23,544 somatic mutations re-
invasive growth [96]. According to my model, mat-
ported in 2000 original publications with respect
uration arrest is neither 100% effective nor is it
to the p53 tumor suppressor gene alone (http://
fixed, but rather is seen as a continuous dynamic
www-p53.iarc.fr/Statistics.html). A recent sequenc-
state, where a fraction of cells is able to escape
ing study, lead by Vogelstein, found that 1718 genes,
and differentiate.

---

The Viennese Cascade
133
representing 9.4% of all known coding genes in the
preferentially occur in already highly methylated
human genome, had at least one non-silent mutation
regions [107].
in either a breast or colorectal cancer (out of 11
breast and 11 colorectal tumor samples). On aver-
Benign tumors – a different entity
age, the number of non-silent gene mutations per tu-
mor was shown to be between 77 and 101 [12].
According to the somatic gene mutation model,
Accordingly, the authors concluded that the genomic
every benign tumor represents an early stage of
landscapes of (breast and colorectal) cancers are
cancer, since those cells already harbor prolifera-
composed of a handful of commonly mutated gene
tion-inducing mutations. Several cycles of prolifer-
‘‘mountains’’ and a much larger number of gene
ation and selection should increase the occurrence
‘‘hills’’ that are mutated at low frequency. The Vien-
of malignant cell formation. However, this is not
nese Cascade should be considered an important
the case, and many types of benign tumors never
shift in thinking as it can explain how this staggering
undergo malignant transformation.
variety of ‘‘hills’’, representing different chromo-
Benign adult tumor cells are characterized by
some and gene aberration patterns, results in identi-
diploid or near-diploid karyotypes [48], intact rep-
cal cancer phenotypes.
licative limits (M1) [108] and functional telomeres
The current scientific knowledge of carcinogen-
[35]. Telomerase activity is absent or low in benign
esis is extensive, and any new theories must ex-
adult tumors [109,110]. Benign tumor cells are
plain experimental results and cancer-related
capable of differentiation, do not grow invasively,
phenomena better than the current somatic gene
do not metastasize and do not show any signs of
mutation model does. This is outlined in the follow-
chromosomal instability (e.g. anaphase bridges or
ing sections.
nuclear anomalies) [35]. This type of benign tumor
rarely transforms into malignant cancer. Accord-
Oncogenes and tumor suppressor genes
ingly, early human colon adenomas are diploid
[111] and the large majority of polyps never be-
The role of somatic gene mutation in carcinogene-
come malignant. Dysplastic lesions are another
sis is a subject of debate [62]. Oncogenes are usu-
group of putative benign tumors, which represent
ally
activated
by
gene
mutations,
gene
the initial stage of premalignant growth. Aneu-
amplifications or balanced chromosomal aberra-
ploidy has been reported in dysplastic lesions
tions. Tumor suppressor genes are deactivated by
[112–114], and cells display variability in nuclear
gene mutations or unbalanced chromosomal aber-
size and shape, increased nucleus-to-cytoplasm
rations such as deletions (loss of heterozygosity).
size ratio, increased mitotic activity, and lack of
In a recent issue of PNAS, the article by Feng
the cytoplasmic features associated with differen-
et al. [105] provides compelling evidence that tu-
tiation. Furthermore, the relative numbers of the
mor suppressors, like p53, remain structurally in-
variety of cell types seen in normal tissue are no
tact but decline in functional activities during
longer observed [4]. Cells of dysplastic lesions
aging. Hence, it is tempting to suggest that nonmu-
and carcinoma in situ do not penetrate the base-
tational declines in tumor suppressor functions
ment membrane. Interestingly, the level of dyspla-
might be intrinsic to adult stem cells at old age.
sia correlates directly with polyp size [115].
It has been shown that normal human mammary
Dysplastic adenomas of the colon are not clonal
epithelial cells spontaneously escape senescence
[116,117] and greater genetic clonal diversity is
and acquire genomic changes [79]. Indeed, the pre-
associated with higher risks for transformation
vailing characteristic of epithelial carcinogenesis is
[118,119]. In accordance with my model, advanced
chromosomal instability [106]. In accordance with
chromosomal instability produces cells with a vari-
the Viennese Cascade, the aberrant activity of
ety of aneuploidy patterns, increasing the probabil-
oncogenes and tumor suppressor genes might
ity of maturation arrest.
either increase the proliferation rate of a particu-
It is known that the heterogeneity characteristic
lar tissue or the proliferation potential of a partic-
of early colon carcinomas is lost in late-stage can-
ular stem cell (by inactivating M1), and eventually
cer [120], which in my view is suggestive of the
drive stem cells into telomere crisis. However,
emergence of an immortal clone with active telo-
most of the gene mutations found in human tumors
merase. In conclusion, benign tumors are the local
[12] might be a consequence of increased prolifer-
result of excessive proliferation of diploid, mortal
ation rates and decreased repair activity in
stem cells, whereas dysplasia and carcinoma
transcriptional silent genes [107]. It has been pro-
in situ represent initial growths of M1-deficient,
posed that gene mutations in somatic cells would
aneuploid stem cells.

---

134
Stindl
Cell fusion experiments and their
sa of e.g. the esophagus. Cytogenetic studies
unexpected results
reveal that most bilateral breast carcinomas arise
independently [124]. The same phenomenon has
In a number of experiments, when tumor cells
been described for thyroid cancer: individual
were fused with normal cells, the resulting tetra-
tumor foci in patients with multifocal papillary
ploid cell hybrids lost the ability to form tumors
thyroid cancer often arise as independent tumors
[4]. These unexpected results led to the concept
[125]. Hence, these clinical observations do not
of tumor suppressor genes. Inactivating mutations
support the argument of somatic mutation very
in these genes seem to be recessive in the pres-
well, since the probability of acquiring 3–6 cancer
ence of an intact, wild-type allele. The introduc-
mutations (the original Vogelstein model) [126] in
tion of wild-type alleles during cell fusion was
several distinct somatic cells is very, very low. A
thought to directly suppress the tumor phenotype.
mutator phenotype [8] might lead to a chaotic
However, an alternative explanation may be that
mixture
of
mutations,
but
cancer-promoting
the new set of intact chromosomes from the nor-
mutations would also be neutralized by cancer-
mal cell can be targeted by the differentiation
inhibiting mutations. Most importantly, these
machinery, and promoters of differentiation-spe-
proposed mechanisms do not reflect the current
cific genes can be demethylated. Once the genes
situation where every second man and every third
are switched on, the (hybrid) cell differentiates
woman are developing cancer [1]. Yet, tumor
and loses its stem cell behavior and ability to form
multifocality can be easily explained by tissue
tumors. Since cell hybrids are not stable and are
exhaustion and multiple recruitment of aneuploid
known to lose chromosomes, the reoccurrence of
stem cells.
malignancy after several rounds of multiplication
[121] can be easily explained by the Viennese
Tissue exhaustion as a consequence of aging
Cascade.
and chronic wounds
The phenomenon of multifocality
Aging is the most prevalent risk factor for cancer.
In accordance with my proposed theory, the aging
The concept of field cancerization [122] refers to
phenotype is largely a result of replicative senes-
the frequent presence of premalignant, genetically
cence and generalized tissue exhaustion. A short-
altered, epithelial cells in the surrounding mucosa
age of stem cells in many aged tissues leads to
of a carcinoma [114]. Fields with dimensions of
the recruitment of M1-deficient and aneuploid
>7 cm in diameter [123] containing multiple pre-
stem cells, and would explain the exponential in-
cancerous lesions have been detected in the muco-
crease of cancer incidence in the elderly.
Table 1
Hallmarks of epithelial cancer and proposed causative mechanisms
Chromosomal instability and aneuploidy
M1-deficiency and extended proliferation of cells with critically
short telomeres leads to breakage–fusion–bridge cycles
Limitless replicative potential
Telomerase up regulation in aneuploid cancer stem cells
Uncontrolled growth
Direct consequence of maturation arrest and/or tissue
regeneration cannot be completed due to the production of
undifferentiated, non-functional cells, eliminating negative
feedback in cancer stem cells
Immature phenotype
Aneuploidy inhibits epigenetic modifications of differentiation-
specific genes
Sustained angiogenesis
Proliferation of cancer stem cells, like normal stem cells,
activates a pro-angiogenic switch in the surrounding connective
tissue
Tissue invasion
Cancer stem cells migrate like normal adult stem cells do
Metastasis
Cancer stem cells respond to regeneration signals (i.e.
chemokines) from other organs

---

The Viennese Cascade
135
Figure 3
Age-association of cancer and aneuploidy syndromes. (A) Age is the most prevalent risk factor for cancer.
Cancer incidence rates increase steadily with advancing age (source: SEER1992-2002). Most intriguingly, there is an
exponential growth in incidence rates leading to an explosion in absolute numbers of newly diagnosed cancers beyond
the age of 40 (in both sexes). (B) Incidence rates of trisomic pregnancies (aneuploidy syndromes, e.g. Morbus Down)
increase exponentially starting at maternal age of 35 (data taken from Ref. [129]), known as the maternal-age effect.
The similarity between both incidences is striking. By definition, aneuploidy syndromes are caused by chromosomal
imbalance and it has been shown that the majority of oocytes in women aged 40 years or older are aneuploid [129–
131]. I propose a similar age-association of aneuploidy in tissue stem cells.
Chronic injury can be associated with an ele-
aneuploidy displays similar kinetics (Fig. 3B). The
vated cancer risk, as in the case of acid reflux dis-
incidence of Down syndrome births and other aneu-
ease
and
esophageal
adenocarcinoma
[127].
ploidy syndromes increases exponentially with the
Dvorak described cancer as wounds that do not
mother’s age, starting at middle age [129]. Oogene-
heal [128]. Others have speculated that cancer rep-
sis may represent the ultimate model of tissue
resents the continuous and unregulated state of tis-
regeneration. A defined number of oocytes are pro-
sue repair [127]. According to the Viennese
duced in the fetal ovaries and the eggs are activated
Cascade, chronic wounds might eventually lead to
in a serial manner, like adult tissue stem cells. It has
tissue exhaustion and tissue stem cells proliferat-
been shown that the majority of oocytes in women
ing beyond their limit (Table 1).
40 years or older are aneuploid [129–131]. In a par-
allel manner, an accumulation of M1-deficient and
Age-association of cancer and aneuploidy
aneuploid stem cells may contribute to age-depen-
syndromes
dent increase of cancer incidence.
As proposed, cancer initiation starts with the activa-
Resistance to therapy
tion of M1-deficient tissue stem cells relatively late
in life, with the exception of local tissue exhaustion
Chemoresistance is considered to be an acquired
(e.g. smoking, alcoholism, chronic inflammation). A
feature of cancer cells. Numerous mutations in
stem cell hierarchy may exist, where healthy, intact
‘chemoresistance genes’ have been described
cells are utilized first, giving way to genetically-
[132], even against the new-age drug Gleevec
damaged cells. This may explain the exponential in-
[133]. Surprisingly, chemoresistance is an inherent
crease of cancer observed in middle-aged humans
feature of normal adult stem cells [134–136]. Pro-
(Fig. 3A). Another pathological condition based on
tection of the most vital, regeneration-competent

---

Document Outline

• Defining the steps that lead to cancer: Replicative telomere erosion, aneuploidy and an epigenetic maturation arrest of tissue stem cells
  • Introduction
  • Historical background
  • The author ' s quest
  • The Viennese Cascade
  • Discussion
    • Oncogenes and tumor suppressor genes
    • Benign tumors - a different entity
    • Cell fusion experiments and their unexpected results
    • The phenomenon of multifocality
    • Tissue exhaustion as a consequence of aging and chronic wounds
    • Age-association of cancer and aneuploidy syndromes
    • Resistance to therapy
  • Implications of the Viennese Cascade for research, prevention and therapy
  • Conclusion
  • Acknowledgements
  • References

---