Evolution of Energy Metabolism, Stem Cells and Cancer Stem Cells: How the Warburg and Barker Hypotheses Might Be Linked

Abstract

The evolutionary transition from single cells to the metazoan forced the appearance of adult stem cells and a hypoxic niche, when oxygenation of the environment forced the appearance of oxidative phosphorylation from that of glycolysis. The prevailing paradigm in the cancer field is that cancers start from the “immortalization” or “re-programming” of a normal, differentiated cell with many mitochondria, that metabolize via oxidative phosphorylation. This paradigm has been challenged with one that assumes that the target cell for carcinogenesis is the normal, immortal adult stem cell, with few mitochondria. This adult organ-specific stem cell is blocked from “mortalizing” or from “programming” to be terminally differentiated. Two hypotheses have been offered to explain cancers, namely, the “stem cell theory” and the “de-differentiation” or “re-programming” theory. This Commentary postulates that the paleochemistry of the oceans, which, initially, provided conditions for life’ s energy to arise via glycolysis, changed to oxidative phosphorylation for life’ s processes. In doing so, stem cells evolved, within hypoxic niches, to protect the species germinal and somatic genomes. This Commentary provides support for the “stem cell theory”, in that cancer cells, which, unlike differentiated cells, have few mitochondria and metabolize via glycolysis. The major argument against the “de-differentiation theory” is that, if re-programming of a differentiated cell to an “induced pluri-potent stem cell” happened in an adult, teratomas, rather than carcinomas, should be the result.

• “Nothing in Biology Makes Sense Except in the Light of Evolution” T. Dobzhansky (1) &

“Some would argue that the search for the origin and treatment of this disease will continue over the next quarter century in much the same manner as it has in the recent past, by adding further layers of complexity to a scientific literature that is already complex beyond measure. But we anticipate otherwise: those researching the cancer (or any other disease) problem will be practicing a dramatically different type of science than we have experienced over the last 25 years. Surely much of this change will be apparent on the technical level. But ultimately, the more fundamental change will be conceptual.”

• D. Hanahan & R.A. Weinberg (2) &

“Certainly, looking for simple relations will not be sufficient, but delineating the exact mechanism of cell cycle control and stem cell development in prostate (breast) cancer should be helpful in understanding these early pre-neoplastic lesions and their relations to diet”.

• D. Coffey (3) &

“The hypothetical coordinating principle, which gets deranged in cancer, could belong to a class of genes exemplified by homeotic genes that determine tissue pattern formation in early embryogenesis.”

• A. Van de Houff (4)

Introduction

While most human beings do not believe that the human species evolved from the early primitive form of life, all cancer scientists know that, within the human being, a cancer is the result of a normal cell that eventually evolves into a cell that no longer responds to the body’ s control mechanisms to behave properly. It is for that reason T. Dobzhanky’ s quote is a guiding principle to try to understand the factors that contribute to that evolutionary process of carcinogenesis. Since that singular event when someone recognized that cancer was a distinct disease, multiple hypotheses of the cause(s) of cancer, generated by the explosion of detailed scientific information from almost every human discipline, e.g., genetic-, gender-, developmental stage-, environmental agents-, dietary-, medication-, life style- and cultural- factors, have been shown to influence the cancer risks. Hanahan and Weinberg (2) saw, correctly, that piling on more details, generated by modern technologies, will be unlikely to provide an integrating insight to this complex problem. This insight is beautifully highlighted by S. L. McKnight when he stated:

“From the past 30 years, research in the biological sciences has been dominated by molecular biology. The successes of this approach have shaped our understanding of innumerable domains of biology. But any field that becomes sufficiently muscular can over-shadow other credible approaches to scientific inquiry. One field etiolated by the cloud of molecular biology has been metabolism. The vast majority of discoveries made by molecular biologists over the past several decades required no attention to the metabolic state of a cell. Molecular biologists needed no distracting thoughts about the metabolic state of a cell to discover microRNAs, the reprogramming of somatic cells into pluripotent stem cells, or gene rearrangement as the underlying basis for the generation of antibody diversity (5).”

What will be necessary to gain some insight to the “cause”, prevention and treatment of cancer is a “grand integrated synthesis” of concepts from the virus theory, radiation theory, the nature and nurture theory, the oncogene/ tumor suppressor gene theories, the stem cell/de-differentiation theories, multi-stage/multi-mechanisms hypothesis, mutation/epigenetic theories, and inflammation hypothesis. One can see that each hypothesis is supported by solid scientific data, but that all are “incomplete” to explain this complex disease. At the risk of demonstrating that no one cancer scientist, including these authors, can provide the correct “Grand Synthesis”, an attempt must be started. This brief, but broad-ranging, search is to find out if there might be a conceptual framework to integrate all of these aforementioned disciplinary concepts. To that end, the “Title” of this analysis includes some of the key concepts that will be examined for their relationship to each other and ultimately to understanding the carcinogenic process, so that a practical approach to prevent and to treat cancers can be achieved.

Paleochemistry, anaerobic to aerobic transition, origin of single cell organisms, evolution of metazoans, and warburg metabolism

It has been hypothesized that during a period of the early physical history of the earth, the oceans achieved the necessary conditions for the appearance of life, via the single cell organism, that could live in the ocean devoid of significant amounts of oxygen (6, 7). Life originated and had its expansion controlled by gravity, nutrients, radiation exposure, temperature, pH, and atmospheric gases, interacting with an adaptive genome, consisting of specific self-replicating nucleic acid molecules. Life, as we know it, evolved under conditions where bacteria had to derive energy by a complex set of biochemical reactions which had a low oxygen tolerance (8). The single major adaptive strategy for the survival of the various species of single cell organisms had to have an unlimited proliferative ability with sufficient nutrients, absence of toxicants, appropriate temperature, etc.

With the advent of oxygen accumulation in the atmosphere and oceans from photosynthetic algae, new forms of bacteria that could produce energy via aerobic respiration. Somehow, proto-eukaryotes formed after they symbiotically engulfed these aerobic machines, now referred to as mitochondria (9). Life for these two major forms of life demonstrated that the two forms to generate energy from glucose were glycolysis or oxidative phosphorylation metabolism. This transition to aerobic energy production, with its concurrent generation of potentially lethal metabolic by-products, i.e. reactive oxygen species-ROS, had to be accompanied by both intrinsic/extrinsic antioxidant systems, coupled redox biochemical systems and various macromolecular repair systems to prevent damage to both the genomic and mitochondrial DNA.

Concurrent with this, a genetically-determined ability to control the mutation rate of their genomes had to be acquired. Indeed, if the organism had a near perfect manner of repairing its genomic DNA or of preventing mutations being formed during normal DNA replication, the species would not survive because of the inevitable environmental change, to which their protected genome would not be adaptive. On the other hand, if the organism had a genetically-determined ability to form too many “errors of DNA repair” or “errors of DNA replication”, the species would not be adaptive, even in a relatively stable environment. Therefore, by producing just the right proportion of mutations in its population and in the individual organism when the environment did change, at least a few might survive with a genotype/phenotype required for the perpetuation of its species. Therefore, unlimited replication and an adaptive mutation rate were required for single cell organisms to survive a changing environment.

At the historic moment when cells that were able to adhere to each other (which, in itself, was a very critical biological event), they acquired new adaptive phenotypes. It has been hypothesized that it was the formation of collagen family of molecules, which is required by multi-cellular organisms to form coherent cells and tissues (8). This multi-cellular “glue” requires molecular oxygen for its synthesis. It has been hypothesized that the change of the paleochemistry of the oceans from an oxygen-deficient milieu to one of oxygen- sufficiency, it became toxic to the anaerobic microorganisms (6, 7). While there are no historic ‘ missing” links for the transition to a multi-cellular organism, it can be surmised the multi-cellular organism acquired new adaptive phenotypes, which had to have new genes to code for these new structures and functions. However, recent discovery of potential “contact-dependant molecules” has been shown in bacteria (10). The first phenotype to appear in this “society of adherent cells” had to be “growth control”. For without a means to control cellular replication within this adherent society of cells, the cells could be likened to a tumor. The second adaptive phenotype appears to have been the ability to differentiate, using a shared genome that had genes to be used for specialized cells within the society of cells, either during different phases of development or within different locations of the multi-cellular organism. This implies genes and mechanisms to regulate the differential expression of genes in the total genome. The third phenotype would be the ability of the cells to commit suicide or to apoptose when a particular cell or cells were either damaged or no longer needed for the next phase of development (11). A fourth phenotype was “senescence” or “mortality” of somatic cells and death of the individual organism. However, clearly, the death of the whole organism is not necessarily related to the senescence or the death of all the cells of the body (live cells can be derived from a multi-cellular organism shortly after its death).

In order that there was an evolutionary survival of a multi-cellular species, another VERY IMPORTANT adaptive fifth feature had to appear, namely the creation of germinal and somatic stem cells. One of the fundamental definitions or characteristic of a stem cell is its state of unlimited proliferative potential or its “immortality”. Concomitant with the ability to have germinal and somatic or adult stem cells within this society of differentiated and mortal somatic cells was the need to create a sixth feature of the metazoans, the “stem cell niche” (12-17) or a sequestered micro-environment to protect these stem cells from the factors required for differentiation. The genes are needed for these stem cell microenvironments. When the multi-cellular organism is conceived, some cells are sequestered from the processes of differentiation, apoptosis, and senescence. The appearance of germinal and adult stem cells had to appear. The germinal stem cells were needed to provide an evolutionary means of species survival, while the somatic stem cells provided cells for specific tissue growth, repair/replacement, and highly specialized differentiated functions.

Now, these new phenotypes (growth control, differentiation, apoptosis, cellular senescence and formation of germ and somatic stem cells) had to have new genes that did not exist in the genomes of the unicellular organisms (although precursor genes might have existed that did not have these exact new functions). The multi-cellular stem cells resembled the unicellular organism under one set of conditions, namely when the cell had to proliferate. Unicellular organisms divide symmetrically to product two daughters that are identical to the mother cell. They can divide in an uncontrolled manner, restricted only by temperature, nutrients, etc. They are, in effect, “immortal”. The definition of a stem cell is that they can divide either symmetrically to make two identical stem cells or asymmetrically to produce one stem cell and one progenitor or transit amplifying cell that will differentiate, apoptose or senesce. Therefore, the transition from the unicellular organism to a multi-cellular organism required a gene or genes that controlled the decision to divide symmetrically or asymmetrically Fig. 1.

Fig. 1. These diagrams illustrate two possible means by which stem cells (Oct-4 +) decide to divide by symmetrical or asymmetrical division. In panel A, if a stem cell binds to a specific extracellular matrix, as symbolized by attaching to a coated plastic dish, the signal received combines with signals in the medium (growth factors, Ca++, nutrients, oxygen, etc.) to stimulate genes and gene products to bring about the division plane to be perpendicular to the attachment plane. As a result, both daughter cells continue to have identical signaling as did the maternal stem cell (symmetrical cell division or expansion of the stem cell population). In panel B, the stem cell binds to a different substrate molecule, as represented by a natural extra-cellular molecule, such as laminin or collagen type 4. In this case, the signal this substrate molecule induces a different intracellular signal that interacts with the same signals from the medium, Ca++, oxygen, etc. to stimulate different genes and gene products to cause a division plane within the stem cell to be formed parallel to the attachment plane. In this case, the daughter cell on the bottom will mimic the same intracellular signaling as its maternal stem cell. Important to note that if these stem cells do not have functional gap junctional intercellular communication, then these signals are not transmitted to the other daughter cell because that daughter cell does not interact with the substrate signal. As a result, these daughter cells receive a different set of combined signals that trigger a commitment to become a progenitor and ultimately terminally differentiated progeny (Permission granted from Oxford University Press, 2011).

From bacteria to cancer cells via stem cells

At this stage, a “leap of the imagination” has to take

place. One might start with Hanahan and Weinberg’ s review, “Hallmarks of Cancer” (2), in which they identified six “unifying characteristics” of cancer cells, namely, self-sufficiency in growth signals; insensitivity to growth inhibitory signals; evasion of programmed cell death; limitless replicative potential; sustained angiogenesis, tissue invasion; and metastasis. More recently, due to rapid accumulation of new molecular details, Hanahan and Weinberg (18) expanded three additional organizing concepts, namely, reprogramming of energy metabolism, evading immune destruction and the complex interaction of normal and cancer cells in the "tumor microenvironment". However, of these characteristics, decades of cancer researchers have noted that cancers do not have “growth control” (19); they do not terminally, but can partially, differentiate (20); they exhibit abnormal ability to apoptose; and do not senesce but remain immortal. Not only are these characteristics distinctly different from normal somatic mammalian cells, but they are reminiscent of bacterial cell characteristics. However, it has to be noted that what is meant when one characterizes mammalian cells as normal with those aforementioned characteristic hallmarks are the somatic metazoan cells. There are two kinds of normal mammalian cells: (a) the normal embryonic, germinal, and adult stem cells, and (b) the somatic non-stem cells (progenitortransit amplifying and the terminally differentiated cells).

Here, the definition of a stem cells is one that (a) has unlimited replicative capacity (self renewal) and (b) is capable of differentiating. Implied in this generalization is that the stem cell can divide either symmetrically or asymmetrically, depending on some external factor(s). The toti- potent stem cell (the fertilized egg) can give rise to the 200 plus differentiated cell types in the human being. The “embryonic stem cell” or “pluri-potent” stem cell, which is functionally defined as being able to form teratomas when placed back into an adult organism to form the three germ layers, adult somatic stem cells (multi-potent, bipolar) have a developmentally- restricted capacity to give rise to all cell types under normal developmental conditions. While there is no universal agreement on the characteristics of either these pluripotent stem cells or organ- specific adult stem cells to date, one thing seems obvious, in a classical sense, namely, the unicellular organism could only divide symmetrically, while the normal adult stem cells could divide both symmetrically for self renewal and to divide asymmetrically to differentiate, to apoptose or to senesce.

Cancer cells, which do not terminally differentiate, but normally divide symmetrically (more elaboration on this generalization during the discussion of the heterogeneous nature of cells within a tumor and of the distinction between “cancer stem cells” and “cancer non-stem cells”). In effect, during the process that converts a “normal mammalian cell” to a malignant “cancer cell”, that cell seems to have “de-evolved” within an organism, in a manner that seems to reverse the evolution of a multi-cellular organism from a unicellular organism.

One characteristic of normal mammalian stem cells and cancer stem cells that seems to have been ignored is the important observation that Loewenstein and Kanno made decades ago, namely, cancer cells do not have functional gap junctional intercellular communication (GJIC) (21). Since unicellular organisms do not have the genes that code for this important biological structure and function and that some very early multi-cellular organisms do have this gene (22), it might be intriguing to speculate that the structure/function of the gap junction family of genes (“connexins”) (23, 24), might have something to do with both the evolution of the multi-cellular organism, as well as the evolution of the cancer cell within a multi-cellular organism. This leads to a brief review of the various theories of carcinogenesis before further examination of the unicellular organism to the cancer cell can be developed.

Multiple hypothesis of carcinogenesis

While many exhaustive reviews of individual hypothesis of cancer causation have been published (from the viral theory (25); stem cell (26-30) versus de-differentiation (31) theories, mutation versus epigenetic theories (32) , oncogene- tumor suppressor theories (33, 34), genetic versus environmental or nature versus nurture or nature and nurture theories (35)), few have tried to integrate all of them, since each seems to have strong support, albeit all being incomplete. Therefore, starting with Loewenstein and Kanno’ s observation that all cancers seem to lack functional gap junctional intercellular communication (GJIC) (21), an assumption will be made, namely, the family of gap junction genes, the “connexins”, will be an integration factor to tie all these cancer causation hypotheses together.

First, the gap junction is a membrane channel, found in metazoans, from early species to the human being. This membrane-associated structure consists of six connexin proteins. Once assembled into a hexamer (“connexon”) and transported to the membrane, it can be joined with the connexons of contiguous cells, form a aqueous channel, through which ions and small molecular weight molecules, such as amino acids, c-AMP, nucleosides, etc., below approximately 1000 D, can passively diffuse directly from cytoplasm to cytoplasm (36-41). It has been noted that these GJ channels can act either as a “sink” or “source” of signals within coupled cells (42). The function of these gap junctions has been attributed to synchronizing both electrotonic and metabolic functions in different tissues. They are found in all organs and in most cells of solid tissues, including the lympho-reticular systems at certain times in their development and function (43). In general, GJIC appears to be critical for the control of cell proliferation (38), development and differentiation (44), apoptosis (45). The observation that cancer cells do not have growth control, do not terminally differentiate or apoptose, nor do they have functional GJIC suggests, but does not prove, that gap junctions are a necessary, but insufficient, factor in the carcinogenic process.

As a start to the integration of a comprehensive hypothesis, the stem cell theory has been one of the older theories of cancer. A number of observations seem to support this hypothesis. First, the immortality of cells of a tumor (more to said on this later with regard to cancer stem cells and cancer non-stem cells) seems to be consistent with the idea that, since the stem cell is defined as an immortal cell until it is induced to terminally differentiate or to “mortalize”, the cancer might have originated from an adult organ-specific stem cell, such as breast or prostate stem cell. The idea of “cancer as a stem cell disease” or “disease of differentiation” (26-29) is reflected in this concept, as well as in “oncogeny as partially-blocked ontogeny” (20). On the other hand, the opposing hypothesis is that a somatic differentiated and mortal cell can “de-differentiate” (26), or be “reprogrammed” to become “immortalized”, and then start down the carcinogenic process.

One of the prevailing paradigms in the cancer field is that one must first “immortalize” normal, mortal cells, before one can neoplastically transform these immortalized cells (46). Using in vitro approaches with either rodent or human primary fibroblasts, one can find in the literature scores of papers showing the induction of immortalization of rodent cells that can, subsequently, become neoplastically transformed. On the other hand, induction of immortalization of human fibroblasts, by non-viral means, seems almost impossible, in spite of a few reports of success (47). Only with the successful “immortalization” of primary human fibroblast and epithelial cells with “immortalizing” viruses, such as SV40 or human papillomaviruses (48, 49), did this paradigm seem to have strong support.

On the other hand, an alterative explanation of these “immortalizing” events is possible, while, at the same time, it can lend support to the viral theory of carcinogenesis (25). If the stem cell is naturally immortal until it is induced to terminally differentiate or to “mortalize”, the SV40 or human papilloma virus might infect all cells of a primary culture, which contain both mortal somatic cells and a few normal immortal adult stem cells. The case of the SV40 immortalizing genes, both the p53 and RB gene proteins are inactivated, rendering the few adult stem cells from “mortalizing” (50). These SV40 viruses haven’ t immortalized an already immortal adult stem cell to become immortal but rather, these viruses prevent the natural immortal adult stem cells from “mortalizing”. This seems to be the case when the SV40 large T gene was transfected into a primary culture of normal human breast epithelial cells (51). The few “immortal” cells obtained from this experiment were characterized has having no connexin gene expression or functional GJIC (52), expressing the Oct4A transcription factor (“stemness”) gene (53), and expressing the estrogen receptor gene (54). These SV40 non-mortalized stem cells could then be neoplastically transformed. During this process, they maintained the expression of the Oct4A and estrogen receptor genes, while not expressing the connexin43 gene or exhibiting functional GJIC (53). The observation, that both SV40 and the human papillomaviruses have been associated with human breast cancers or other cancers (55, 56), leads to the question as to what state of the multi-stage, multiple mechanism theory does the cancer virus infection work.

In the mainstream experimental carcinogenesis studies, the multi-stage, multi-mechanism hypothesis seems to have strong support (57, 58), although, generally not considered in molecular oncology studies. This hypothesis is conceptualized by the “initiation”, “promotion”, “progression” stages. It can start to integrate both the stem cell hypothesis and the role of gap junctions, as well as oncogenes and tumor suppressor genes into a large framework. The initiation concept, derived from experimental animal studies, demonstrated that, after exposure of an animal to a low dose or concentration of physical or chemical agents, which did not induce cancers within the life time of the animal, but if followed by another non-carcinogenic agent, an irreversible event had occurred in a single cell of that exposed animal. Initiation was, then, operationally, due to an irreversible event. Mutagenesis, which leads to an irreversible change in the genomic information (gene or chromosomal mutation), has been offered as the mechanistic basis for the initiating event. However, one must keep in mind that a stable epigenetic event might also explain this observation, as well as the selection of a pre-existing spontaneous mutation, caused by either another mutagen or by an error of DNA replication (59).

Promotion, on the other had, is the operational definition of a process to clonally expand a single initiated cell. The promotion process can be brought about by growth, wound healing, inflammatory processes, and non- DNA damaging agents (60) . The expansion of an initiated cell seems to be the result of both the mitogenesis of the initiated cell and the inhibition of the apoptosis of these initiated cells (61). Mechanistically, it has been hypothesized that the inhibition of GJIC is responsible for allowing the initiated cell to escape the suppressing effect of surrounding normal cells (62). Since GJIC has been linked to growth suppression, induction of apoptosis and differentiation, the inhibition of GJIC might be expected to lead to growth stimulation, inhibition of both differentiation and apoptosis, all of which are characteristics of tumor promotion. Conversely, agents that stimulate GJIC or prevent tumor promoter inhibition of GJIC have been associated with chemoprevention (63). Oncogenes, such as Ha-ras, Neu, Raf, Src, Mos, have been shown to inhibit, stably, GJIC, whereas tumor suppressor genes appear to restore GJIC in tumor cells and to reduce tumor growth (41). To provide even more evidence in the importance of gap junctions in the cancer process, anti-sense connexin genes, transfected into normal communicating cells, induces a tumorigenic phenotype, whereas, the transfection of normal connexin genes restores GJIC in non-communicating tumor cells and reduces tumor growth (63).

Stem cell as targets for the initiation process and as the origin of “cancer stem cells”

To put these observations into a bigger context, while cancer cells do not have functional GJIC, there can be two reasons that cells do not communicate. First, if a the target cell for starting the carcinogenic process, i.e., to be the “initiated cell”, is the adult stem cell, then these cells would be presumed not to express their connexin genes. The second kind of cancer cell would be cells that have expressed their connexin genes but the connexin proteins were rendered non-functional by either a mutation or by oncogenes that had posttranslationally modified the connexin proteins, as has been shown by the src, ras, raf and neu oncogenes (64).

The relationship of gap junctions and stem cells is critical for the understanding of both normal differentiation and development, as well as for abnormal development that has been noted during the cancer process. Lo and Gilula have characterized the expression of connexin genes and functional GJIC in the fertilized egg and early embryo (65). The early embryo in several organisms did not express connexins or have functional GJIC until the compaction stage. While embryonic or pluripotent stem cells have been reported to have expressed connexins, depending on the conditions of culturing them, results can be contradictory (see later). All the reported isolated adult stem cells seem to demonstrate that they do not have functional GJIC. There seems to be a simple reason for the observation that embryonic and adult stem cells express Oct4A transcription factor, “stemness” gene, and do not express their connexin genes. These two genes serve two diametrically different functions. One (the Oct4A) is to maintain a “stemness” state or a primitive, total undifferentiated state, or in the case of organ specific adult stem cells, organ-specific “stemness” potential. The other (connexin) genes are required for differentiation. When a stem cell is sequestered in its “niche”, it is there for at least two important reasons. First, it must be kept in a relatively non-dividing state and second, it must not be in direct contact with its differentiated daughter, since differentiated factors could conceivably cause them to differentiate.

While the operational characterization of in vitro transformation is to produce an “anchorage-independent cell”, the history of successful in vitro transformation of rodent cells is plentiful, while that history of normal human fibroblast or epithelial cells being neoplastically transformed is quite sparse (66, 67). Moreover, when one looks at the history of trying to isolate human breast cancer cells, only a few lines were isolated, including the famous MCF-7 cell line. Moreover, determining the presumptive neoplastic feature of any in vitro “transformed cells” was its ability to grow in an immune-compromised rodent when huge numbers were injected (implying that the population of these transformed cells were heterogeneous, some having the characteristics of being immortal and stem- like, while the others were now no longer immortal, albeit being the result of the carcinogenic process and found in the tumor). This could have been the real genesis of the concept of the “cancer stem cell” or “tumor initiating cell” and “cancer non-stem cell” and tumor non-initiating cell”.

In our ability to isolate adult stem cells, our laboratory group made two major initial assumptions. The first was that the target cell was the adult stem cell. Second, we assumed these adult stem cells did not have functional GJIC. We designed an in vitro, “Kiss of Death” assay, on which normal gap junction-coupled, lethally-irradiated normal human fibroblast formed a feeder layer mat, onto which disassociated cells from a normal human organ were placed. We knew the disassociated organ cells would contain three classes of cells, namely , the few adult stem cells, which had no functional gap junctions; the progenitor cells which did have functional gap junctions; and the terminally differentiated cells (which might have or not have functional gap junctions). However, the terminally differentiated cells, by definition, would not proliferate. The progenitor cells, with functional gap junctions, would couple with the dying feeder layer cells and either “contact-inhibit” or die. On the other hand, the only cells from the organ that would survive and proliferate would be cells that had no functional GJIC. Without gap junctional intercellular communication, the cells would not be “contact-inhibited’ (68). After millions of the dissociated organ tissues were placed on this “Kiss of Death” assay, only a few would survive to form viable colonies. These cells were later shown to have stem-like properties, expressed Oct4A and did not have functional GJIC (69). They would later be seen to differentiate into organoids, and specifically, in the case of human breast stem cells, exhibit features of normal human breast tissue (70).

The most important feature of the human breast stem cells was the relative ease by which the Oct4A- and connexin negative-, estrogen receptor positive- cells were blocked from “mortalization” by SV40 and then, their subsequent neoplastic conversion by a combination of ionizing radiation and the erb2/neu oncogene (71). Whereas, under the same experiment conditions, the differentiated breast epithelial cells, derived from the human breast stem cells, were never “immortalized”, let alone neoplastically transformed (71).

At this point, many of the hypotheses of carcinogenesis that have been briefly reviewed seem to be linked to each other, the stem cell being the “target” for initiation of carcinogenesis. Gap junctions, by being involved in the inhibition of cell proliferation, differentiation and apoptosis, can be inhibited by environmental and dietary agents, are involved in the tumor promotion and progression phase of carcinogenesis (64). Both oncogenes and tumor suppressor genes can affect gap junction function (61). Both mutagenic and epigenetic mechanisms can be involved in all phases of the multi-stage process of carcinogenesis.

Cancer prevention and treatment in view of the Barker hypothesis

While, in principle, one would like to prevent the “initiation” of the target cell (presumptive adult stem cell), the fact is that one can never reduce to zero the probability of initiating all adult cells of each organ. One can reduce the probability of initiation by trying to avoid exposure to initiators (i.e., one can limit exposure to UV light), yet even if one lived in a sun-light- minus environment, some skin cancers would appear due to spontaneous mutations caused by errors in replication of skin stem cells. However, the tumor promotion phase of carcinogenesis in non-genetically-predisposed adults takes decades to produce a cancer. Therefore, the tumor promotion phase would be the most efficacious stage for cancer prevention strategies. While, in principle, this might seem simple, the mechanisms, by which an initiated stem cell are promoted, are varied (phorbol esters inhibit GJIC by activating protein kinase C- phosphorylation of gap junction connexin proteins, but phenobarbital, DDT, TCDD, PCB’ s and phthalates work as tumor promoters to inhibit GJIC via different mechanisms. Therefore, there will be no universal dietary or drug anti-tumor promotion intervention strategy (72).

More recently, if one continues to assume that the adult stem cell is the target for the initiation process, then, in principle, one could increase or decrease the adult stem cell pool during embryonic and fetal development in utero by the mother being exposed to various environmental-, dietary-, and medicinal- agents that might increase or decrease, in an organ-specific fashion, the adult stem cell pool. The “Barker Hypothesis” (73) speculates that many chronic diseases, later in life, might be the result of embryo/ fetal exposures during in utero exposure/behavior of the mother. It was then hypothesized that the underlying mechanism for this “Barker Hypothesis” could be the modulation of adult stem cells (74-76). This, then, could increase or decrease the risk of initiation of cancer later in life.

One human example to support this hypothesis would be the role of soy products of the development of breast tissue and breast cancer frequency in Japan in both the non-exposed and exposed atomic bomb Japanese women (74, 77). Soy products, containing a bioactive compound, genistein, can induce differentiation of human adult breast epithelial stem cells (78). If during pregnancy, Japanese women ate large amounts of soy products, as well as being generally, calorically restricted, then the number of adult breast stem cells would be expected to be reduced (79). After birth and during puberty, the female offspring would have few adult breast stem cells for both breast development and/or being targets of initiation. Interestingly, if rodent results are extrapolated to human beings, that same soy diet in these women might act as tumor promoters of any initiated breast stem cell (80).

If, in the progression stage of carcinogenesis, a stable inhibition of GJIC occurred (activation of a mutated oncogene), then anti-promotion strategies will not work, as this would require a chemotherapeutic strategy to restore GJIC in the malignant tumor cells (64). Given two types of cancer cells, those that never express their connexin genes (e.g., HeLa and MCF-7) (81, 82), and whose connexin genes are normal and expressed (e.g., cells with active oncogenes), the strategies would have to be different (64). In the case of human breast cancer, MCF-7 (estrogen-receptor and Oct4A positive and Cx43 negative), seem to share many phenotypes as did the normal human breast stem cells and SV40-immortalized and neoplasticallytransformed human adult breast stem cells (52).

This type of breast cancer cell (estrogen receptor +, connexin43-, Oct4A+) would have to be treated with agents that might act at the transcriptional level to turn off the estrogen receptor (an adult organ-specific stem cell marker) and Oct4A transcription factor gene for stemness and to turn on the connexin43 gene, needed for growth control, differentiation, or apoptosis. On the other hand, a breast cancer cell, expressing the ERB2/neu oncogene, or a prostate cancer that expresses various nonfunctional connexin genes (83), specific inhibitors to the specific oncogene being expressed might restore the function of the connexins, thereby restoring growth control, differentiation or apoptosis of the malignant cell.

Is the origin of cancer stem cells the adult stem cell or the “re-programming” of the adult somatic differentiated cell?

Several recent concepts have been either “re-introduced” (“cancer stems cells” or “tumor-initiating cells”) or recently conceived (Induced pluri-potent stem cells-“iPS” cells). The work of Al-Hajj et al (84) demonstrated that each tumor and tumorigenic cell line (85) were a mixture of “cancer stem” or “tumor initiating” cells and “non-cancer stem cells”. Although, currently, there seems to be no universally-accepted marker for these “cancer stem cells”, operationally, some feel the ability to isolate “side population” cells from tumors, based on their characteristics of excluding fluorescent toxic drugs (an operational or functional definition) and their ability to re-initiate the tumor from which they were derived, are the “”cancer stem cells” (86, 87). Moreover, the generalization that the “cancer stem cells or “side population cells” were rare cells within a tumor has been questioned (see 69).

In spite of the controversies about the role Oct 4 being a marker for stem cell, the fact that it has been shown as a marker for embryonic stemness and as a critical gene needed for the so-called re-programming of somatic differentiated cells to become “embryonic like” pluripotent stem (“iPS”) cells (88). The subsequent reports of this incredible feat of being able to “de-differentiate” or “re-program” somatic differentiated cells bears on the critical hypothesis of the “target” cell for starting the initiation process of carcinogenesis (89-102). It challenges, directly, the stem cell theory of carcinogenesis. Consequently, it will be important to try to resolve this issue.

In the study by Tai et al (53), a series of clonally-derived series of normal differentiated breast epithelial cells, an SV40- derived “immortalized” cell from the normal human breast stem cell, a weakly tumorigenic cell line and a highly tumorigenic-derivative was tested to find if the stem cell or its normal differentiated daughter cell line could give rise to a neoplastic derivative.

These human adult breast stem cells were defined by the absence of the expression of Cx43 and functional GJIC, by the expression of the estrogen receptor, by the expression of the Oct4A genes, by the expression of ABCG2 (drug transporter gene) and by its ability to form a 3-D human breast organoid on Matrigel (69). These cells were able to be blocked from “mortalization” and subsequently, neoplastically- transformed by a combination of SV40 transfection, X-ray treatment and transfection of the ERB2/Neu oncogene. The results clearly demonstrated that only the Oct4A- expressing cells, but not the normal breast epithelial cells, derived from the Oct4A adult stem cells (which were Oct4A- negative), gave rise to cells that did not differentiate after transfected with the SV40 large T gene (51). These cells, while remaining immortal, were not tumorigenic until they were X irradiated (71). Several clones from this population gave rise to slow growing tumors in the nude mouse. The ERB2/Neu oncogene was transfected into these slow growing tumors to give rise to clones of highly tumorigenic cell lines. It should be noted that, from the normal breast stem cell to the blocked mortalized or “immortal” breast cell to the weakly and highly tumorigenic clones, the Oct4A, stemness gene, remained expressed (53) . In other words, the Oct4A gene was not suppressed and then re-expressed or re-programmed during this process. No “iPS” breast stem cell was created from the normal differentiated breast epithelial cells. Rather an adult breast stem cell was prevented from being repressed and remained “stem-like” during the carcinogenic process. If this process involved an “iPS-like cel” first to be produced, then these cells should have produced a teratoma-like tumor rather than a breast carcinoma. In our studies, no normal differentiated breast epithelial cell, derived from differentiated normal adult breast stem cells, could be “immortalized” or “re-programmed” or “de-differentiated” back to a stem-like cell.

In effect, this demonstrated what is defined as an initiated cell, namely a stem cell that is inhibited by a permanent event, which makes the initiated cell unable to divide asymmetrically under normal conditions, but which can divide symmetrically. This implies that the initiating event in carcinogenesis involves mutating or epigenetically (as SV40) altering some gene, irreversibly, rendering the gene unable to regulate the decision to divide via an asymmetric process. Among other gene products, affected by SV40, both p53 and RB proteins are rendered ineffective, preventing these critical genes needed for differentiation to occur. In Fig. 1, one possible functional gene set might be those genes needed to determine if a cell’ s division plane is parallel or vertical to the niche plane. Which combination of external factors (oxygen tension, extracellular matrix, nutrients or environmental chemicals or drugs) might stimulate that signaling pathway? However, these “cancer stem cells” change the very micro-environment as they grow. The altered micro-environment could then act to alter gene expression in these cells to become partially differentiated (i.e., Oct4A gene is suppressed and the cells become “mortal” but not normally terminally differentiated).

Are iPS cells derived from adult stem cells?

In the field of stem cell biology, particularly focused on the potential for stem cell regenerative therapy, the discovery of “iPS” cells provided, in principle, a means to generate embryonic-like stem cells that would circumvent the philosophical, ethical, religious, political and legal issues of using embryonic stem cell for therapy. In spite of remarkable success of generating iPS cells by many different means, there still are some technical issues that still might prevent their use for human stem cell therapy (103-110). However, while the international community has accepted the general explanation that these “iPS” cells are derived from somatic differentiated cells, there is another possible explanation that seems to be ignored or even discredited (111-114). That explanation is based on the fact that organ specific adult stem cells do exist (probably in all organs). Therefore, all the various protocols (using various viral carriers of a number of “stemness”-requiring genes (Oct4, So× 2, Myc, etc.) might be selecting pre-existing adult stem cells that already express Oct4. This hypothesis seems to have been supported by the recent observation by the Dezawa group (115) that a subpopulation of human fibroblast cells, which appear to be preexisting adult stem cells or “multilineage-differentiating stress-enduring (MUSE) cells, whose properties are similar to those of iPS cells , give rise to the iPS cells, whereas the non-multilineage-differentiating cells did not generate iPS.

Recently, a number of critical experiments have been done that have generated unusual findings. In comparing the expressed genes of the normal embryonic stem cells, of the target differentiated somatic cell from which the “iPS” cells are derived and of the “iPS” cells, these studies have shown that, while there is great similarity in the gene expression profiles of the “iPS” cells and the embryonic stem cells, but not of the differentiated target cells, there was always a signature of the organ from which the “iPS” cells were derived. In addition to epigenetic alterations, genomic instability has been noted in these “iPS” cells (102-109). Clearly, this could be interpreted that “re-programming” was nearly complete but it never can go to completion under today’ s protocols. On the other hand, it could be interpreted as meaning the “iPS” protocol only selected pre-existing adult stem cells that had most of its stemness genes expressed (i.e., Oct4), but that during commitment to the differentiation of any particular organ (skin, liver, breast, prostate, etc.) some pluri-potent stemness genes are permanently repressed. This would ensure that a breast, liver, prostate adult stem cell, in its unique organ-specific niche, control the specific differentiated lineages of cells needed for that organ. While, at this time, this hypothesis might seem heretical, it seems that it is testable. It can be done by simply treating a population of pure adult organ-specific stem cells with any protocol to produce “iPS” cells and comparing the frequency of recovered “iPS” with that of treating a primary culture of differentiated cells. The frequency would be predicted to be much higher from a population of pure adult stem cells than from a population of somatic differentiated cells, which would only contain a few rare adult stem cells.

Indeed, as will be related to the last section, a series of reports examining what happens during the creation of “iPS” cells to the mitochondria. This focus will relate to the origin of “iPS” cells, the cancer stem cells, but also the creation of normal stem cells during the evolution of multi-cellular organisms.

Embryonic stem cells have been shown to metabolize by glycolysis rather than by oxidative phosphorylation and apparently have far fewer mitochondria than their differentiated daughters (116). During the proliferative, but finite, life span of the progenitor and differentiated cells, these mitochondria pick up many mutations in the mitochondrial genome. Therefore, the experimental question is: “What happens to the mitochondria during the re-programming of these cells to become ‘ iPS’ cells?”. Several studies studied and compared the state of the mitochondria in embryonic stem cells, the “iPS” cells and the differentiated cells from which the “iPS” cells were obtained (117, 118). The results indicated that the mitochondrial number and quality resembled that of the embryonic, but not the differentiated, cells. In other words, the interpretation of these results is that “reprogramming” caused the loss of the mutated mitochondria and reduced the number. Most importantly, because the low number of mitochondria in embryonic and “iPS” suggests that these cells are surviving via an aerobic glycolysis metabolism, unlike the differentiate somatic cell, which uses an aerobic metabolic means to generate energy. Alternatively, it seems that another interpretation is that the “iPS” cells were derived from adult stem cells, which would be assumed to have a low number of mitochondria and to exist via an anaerobic metabolic means. One can, in principle, “reprogram”, epigenetically, normal DNA expression, but one cannot epigenetically, “reprogram” mutated mitochondrial DNA.

Cancer as a consequence of the de-evolution of multi-cellularity and re-establishment of the warburg effect

It been said, under a different context, that “What goes around, comes around”. Following the admonition of Dobzhansky to keep the principles of evolution in mind, as well as the advice of Hanahan and Weinberg (2), it would be wise to stand back and to consider the broad concepts of the carcinogenic process, and to be careful not to get buried in the mountains of reductionalistic technical details. A broad attempt to understand the complex process of forming cancers has been made in this Commentary. From the evolution concept, the facts of paleochemistry of the earth’ s ocean and atmosphere, which allowed living organisms to generate energy, anaerobically, from carbon molecules, led to a situation in this early ocean where primitive organisms had to adapt to a world with “toxic oxygen”. They did so by selection of genetic factors coding for the ability to metabolize via aerobic system. The combination of historic events of the evolution of a aerobic algae leading to the formation of a symbiotic mitochondria in a new type of multi-cellular organism, which somehow benefitted from the selection of genes that could synthesize the

collagen family of proteins that allowed for cells to attach (7, 8). This attachment forced the conditions for selection of genes needed for growth control, selected expression of genes in a large genome for unique functions, i.e., differentiation, for selected suicide of injured or non-needed cells, for the formation of germinal and somatic stem cells needed for species survival and tissue replacement and for senescence of somatic cells.

While this speculative view of the origin of the stem cells needed for multi-cellular metazoans might seem in need of more rigorous scientific testing, the reactions of single cell organisms, such as bacteria, to either or both changes in oxygenated and heavy metal exposures, which can result in major phenotypic alterations, might seem to support some aspects of this evolutionally transition, as well as the evolution of normal stem cells to “cancer stem cells”. To demonstrate this, Fig. 2, below, demonstrates what happens when anerobic bacteria are exposed to cadmium, which can induce oxidative stress (119). It is also interesting to note that cadmium can transform normal human breast epithelial cells into a basal-like phenotype (120).

Fig. 2. Filamentation of aerobically grown Hpx-mutants of E. coli cells. Cells were grown in Luria broth, anaerobically (A) or aerobically (B). Magnification: × 400. Permission granted by PNAS (118).

In Fig. 3, interestingly, the serendipitous discovery of the cancer therapeutic drug, cisplatin, by Dr. Barnett Rosenberg (121) demonstrated an adaptive response by E. coli of the non-septation of the bacterial cell after DNA replication. In his discovery, he saw that exposures to cisplatin caused the bacteria to replicate its DNA without septation. It should be noted that, while many feel that the anti-cancer effect of cisplatin is via its DNA-damaging potential, the fact that genotoxicity, a random process, would not be expected to cause a uniform response in all E. coli, as would an “epigenetic” mechanism. While it might be stretching an idea, the oxidative stress-induced septation in E. coli could be viewed as a primitive form of “differentiation” or “mortalization”. The oxidative stress-

Fig. 3. E.coli, grown in traditional growth medium, showing normal morphology. However, when E.coli were grown in the same medium but with a submerged platinum electrode, the E.coli had their DNA replicate but they did not septate. This observation led to Dr. Barnett Rosenberg’ s discovery of the anti-cancer drug, cis-platin. Permission grant by: Paul Rosenberg of the Board of Barros Foundation.

induced anti-cancer effect of cisplatin on cancer cells (possibly “cancer stem cells) might have an evolutionary connection.

Philosophically, one wonders if the pathogenesis of septis by a lethal bacteria is much different that lethal metastatic cancer in a human being (122), since both can lead to death of the host organism by being “immortal”. In the case of the cancer cell (immortal “cancer stem cells”), being characterized as metabolizing by the Warburg process of aerobic glycolysis (123-125), and having few mitochondria), it acts as it has de-evolved back to the early primitive anaerobic single cell parental origin.

In their critical review of the Warburg contributions to cancer cell metabolism (123), Koppenol et al., stated that “… Warburg misinterpreted his own observations and promoted the erroneous idea that damaged respiration is the sine qua non that causes increased glucose formation in cancers,… that today, we understand that the relative increase in glycolysis exhibited by cancer cells under aerobic conditions was mistakenly interpreted as evidence for damage to respiration instead of damage to the regulation of glycolysis”. While the authors did note in their review that complicated quantization of cancer cells compared to normal cells has to take into account the micro environmental niches of both tissues, in particular the tumor tissue being heterogeneous due to ineffective vascularization. However, in view of today’ s understanding that both the normal and cancer tissues are also heterogeneous for a number of other factors. One of the important contributors, in this author’ s opinion, is the fact that tumors consist of both “cancer stem cells” and “cancer non-stem cells” (plus invasive and interactive normal stromal and inflammatory cells). Interestingly, Koppenol et al. (123) stated: “As such, to fully appreciate this range of cancer gene mutations, we need to better understand the normal genomic and metabolic profiles of the cancer cells of origin.” Consequently, if the normal adult stem cell is the target cell that ultimately gives rise to the cancer stem cells, and if the normal and cancer stem cell have few mitochondria and metabolize glucose in a similar manner, the interpretation is not that cancer cells “reverted back” by either mutations or “reprogramming”, but rather, they persisted to metabolize via the Warburg pathway during the evolution of the multi-step, multi-mechanism process of carcinogenesis.

Is cancer a new disease in the evolution of human beings?

While this question is not new, it has been brought up for critical re-examination (125). It seems that, while examination of human remains (mummies, bones, DNA from frozen tissues) seems unlikely to be an unequivocal source to answer the question. Yet from what we know of the carcinogenic process today, the likelihood that it is a new disease seems remote. Since carcinogenesis is a multi- step, multi-mechanism process, cancer might have not been as prevalent as it is today, primarily because the median life span of humans is much longer than in the early stages of human history. However, human DNA is involved in all human cancers, either in the “initiation” stage via somatic mutations of genes (p53, RB, BRCA1, DNA repair genes, etc.), or during the “promotion” phase, i.e., via epigenetic alterations), was well as being susceptible to cancer-associated viruses, e.g., SV40, hepatitis, and papilloma viruses), or chromic inflammation, due to many factors. Given that the initiation or DNA damage (i.e., from UV light) can occur anytime, but that the promotion phase in adults might take decades, it seems reasonable that, with a relatively short life span, the frequency of cancers in adults might be low. In addition, since diet is known today to play roles in either the promotion phase or in an anti-promotion or anti-carcinogenesis fashion, caloric restriction or possible eating the “healthy foods”, not the grilled red meats, sugar-laden, caloric-abundant and processed foods that are associated with more recent cultural evolutionary changes, assisted in reducing cancers. Even other cultural factors, such as early pregnancies to reduce breast cancers, would have influenced the reduced risk for these cancers. In today’ s cultures, postponing marriage and pregnancies, together with poor diets, might contribute to the increased risk to breast cancers of the modern age (75, 76).

In addition, since ancient human DNA was not immune to DNA damage or errors in DNA replication, mutations would have occurred in the germinal DNA, as well as the somatic DNA. Therefore, inherited predisposition to cancer, seen in many cases in childhood, would have occurred in the early humans. The leukemias, brain cancers, renal cancers seen in children, especially those associated with heritable cancers which, in general, show up early, would be hard to detect in ancient remains today. In general, soft tissue cancers would not be easily detected in early human remains.

In general, it seems that it would be fair to conclude that cancer is not a new diseases because the fundamental components, DNA, environmental and dietary, and process, multi-stage, multi-mechanism carcinogenesis, existed during the whole period of the biological evolution of human beings. Cancer only became prevalent when cultural evolution of fire, farming, domestication of animals, and availability of excess calories, as well as those cultural factors that helped to extend the life span of humans (127-129). Therefore, it might be more correct to conclude that the cancer frequencies today are higher than those in the past, not because cancer is a new disease, but because human culture has changed.

Synopsis of the broad integration of concepts leading to this perspective of the stem cell origin of carcinogenesis

In a deliberative attempt to step back from the tsunami of data on the factors influencing human cancer from a wide range of disciplines from molecular genetics to epidemiology, this Commentary has tried to bring together important observations that have generated powerful concepts in different fields that might provide a new view of the carcinogenic process. This attempt is based on the assumption that, while prevention and treatment of cancer might result from dumb luck and large amount of solid reductionalistic/empirical studies by many individual disciplines, quicker understanding will come from what Hanahan and Weinberg stated, that major advances will come from new conceptual insights (2).

In this “Commentary”, a focus has been on a unique biological structure, the gap junction, and its demonstrated homeostatic functions in metazoan development. A framework was provided to integrate concepts of evolution, of stem cell biology, of the role of cell-cell communication in species survival and in normal development of the metazoan, as well in various disease processes. It was generated to provide the reader a “road map” of the thinking that speculated a process, by which the evolution of the germ line and somatic stem cells and of the gap junction gene family, provided a means to survive, while at the same time, these mechanisms characterize many of the stem-cell based diseases, such as cancer.

To follow this line of thinking, a summary of the observations used in this Commentary are briefly linked together. Single individual cells are the origin of cancers. While normal progenitor cells have functional gap junctional intercellular communication, functional gap junctions are not found in cancer cells, either because (a) they never express the connexin genes or (b) the connexin proteins are rendered non-functional by a mutation or by post-transcriptional/posttranslational modifications. Tumors consist of mixtures of “cancer stem cells” and “cancer non-stem cells”. Glycolysis is manifested in aerobic bacteria, normal metazoan stem cells and cancer cells. Both normal stem cells and cancer cells are characterized as exhibiting the Warburg phenomenon. The resistance of some cancer cells is due to the expression of drug transporter genes, not by induction of mutations by toxic agents. With exception of teratomas, cancer stem cells are derived via multi-step, multi-mechanism concept of the “initiation”, “promotion” and “progression” processes. By the fact that initiation of any stem cell can be brought about by simple errors in DNA replication, it will be impossible to reduce the risk of initiation to zero. In other words, all human beings, whether they die before any overt diagnosis of cancer is made, are harboring pre-malignant, “initiated” cells in many of our organs. If we lived longer, all of us would get cancer.

The target cells seem to be the adult stem cells for the initiation of both teratomas and adult cancers. If the target cells were the “re-programmed” somatic differentiated cells that become “embryonic-like of “induced pluripotent cells in vivo, then the tumors that arise from them ought to give rise to teratomas, not carcinomas or sarcomas. Stem cells are those cells that can divide either by symmetrical cell division to produce two daughter stem cells or by asymmetrical cell division to produce one daughter stem cell and one progenitor cell destined to terminally differentiate, to senesce or to apoptose. It has been postulated that “initiated stem cells” are the likely progenitors for the “cancer stem cells”. One can functionally define the “initiation” process is the result of preventing a stem cell from asymmetrical cell division under normal in vivo conditions. However, they can “partially” differentiate when the in vivo microenvironment changes during tumor growth, leading to “cancer non-stem cells”. This leaves a tumor with a mixture of “cancer stem cells” which remain immortal and “cancer non-stem cells”, which now become partially differentiated and mortal, unable to perpetuate or sustain the growth of the tumor.

The promotion phase of cancer, which, in most nonchildhood cancers, takes decades to occurs, involves the clonal expansion of the initiated stem cell and the inhibition of the apoptosis of that initiated stem cell. It is possible to interrupt or possibly even reverse the promotion phase. Once a promoted “initiated”, pre-cancer stem cell” becomes autonomous, i.e. metastatic as a “cancer stem cell”, it is characterized has having transitioned to the “progression” phase; as “cancer stem cells” start to grow into tumors, the micro-environment changes. The changed microenvironment can now send epigenetic-inducing signals to cause some stem cells to divide asymmetrically and to partially differentiate into “cancer non-stem cells”. The frequency of “cancer stem cells” to “cancer non-stem cells” in a tumor will be influenced by the endogenous and/or exogenous factors controlling symmetric versus asymmetric cell division of the cancer stem cells. Therefore, in some tumors, the frequency of “cancer stem cells” might be low, whereas in other tumors, the frequency might be quite high.

The strategy for cancer prevention might occur via: (a) reduction of the adult stem cell pools in specific organs during in utero development (which could, of course, have other negative consequences); or (b) by the induction of gap junctional intercellular communication in “pre-malignant cancer stem cells”, that do not express the connexin genes, or by preventing the down regulation of gap junctional intercellular communication (GJIC) by tumor promoters in initiated stems expressing their connexin genes. In view of the existence of “cancer stem cells” in a tumor, cancer therapy must target these cancer-sustaining cells by either transcriptionally expressing the connexin gene(s) in cancer cells, such as teratomas, and HeLa and MCF-7 carcinomas, or by inhibiting specific oncogene signaling pathways that render the connexin proteins non-functional. On the other hand, modulation of the stem cell pools in utero (increasing or decreasing the numbers of adult stem cells in specific organs) could affect the risk (increase or decrease) to any stem cell-based chronic disease later in life (A possible cellular mechanistic basis for the Barker hypothesis).

While many of these assumptions and observations are highly controversial, this summary was designed to be starting point from which the scientific process of hypothesis testing and experimental design to test these hypotheses and assumptions can start.

Potential conflict of interest

The authors have no conflicting financial interest.