Induction of iPS Cells and of Cancer Stem Cells: The Stem Cell or Reprogramming Hypothesis of Cancer?


  • James E. Trosko

    Corresponding author
    1. Department of Pediatrics/Human Development, College of Human Medicine, Michigan State University, East Lansing, Michigan
    • Correspondence to: James E. Trosko, PhD, Department of Pediatrics/Human Development, College of Human Medicine, Michigan State University, East Lansing, MI 48824. E-mail:

    Search for more papers by this author


This article as designed to examine whether the “stoichiometric” or “elite models” of the origin of the “induced pluripotent stem” (iPS) cells fits some experiment facts from the developmental biology of adult stem cells and from the field of cancer research. In brief, since the evidence presented to support the stoichiometric model failed to recognize the factual existence of adult organ specific stem cells, the model has not been rigorously tested. In addition, the demonstration of a subset of cells (MUSE cells) in normal primary in vitro cultures of human fibroblasts (the usual source of iPS cells) seems to be the origin of the iPS cells. Moreover, from the field of carcinogenesis, the “stem cell” versus “de-differentiation” or “reprogramming” hypotheses were examined. Again, using the role of glycolysis, known to be associated with the Warburg effect in cancer cells, a list of experiments showing that (a) normal stem cells, which have few mitochondria, metabolize via glycolysis; (b) the stem cells are targets for “initiation” or “immortalization” or the blockage of differentiation and apoptosis of the stem cells by “immortalizing viruses”; (c) Lactate dehydrogenase A (LDHA), when expressed, is associated with glycolysis and therefore, must be expressed in normal adult stem cells, as well as in cancer cells; and (d) p53, depleted or rendered dysfunctional by SV40 Large T antigen, is associated with the reduction of mitochondrial function and mass and is associated with the Warburg effect. Together, these observations from the iPS and “cancer stem cell” fields support the idea that both iPS cells and cancer stem cell are derived from adult organ-specific stem cells that do not restore or switch their metabolism of glucose from oxidative metabolism to glycolysis but, rather, in both cases, the adult stem cell, which metabolizes by glycolysis, is prevented from differentiation or from metabolizing by oxidative phosphorylation. Anat Rec, 297:161–173. 2014. © 2013 Wiley Periodicals, Inc.

“…The first embryonic cells, blastomeres, of mice and other mammals are all totipotent. During cleavage and early morphogenesis these cells come to occupy different positions in the three-dimensional embryo. Some cells are on the outside, some inside. The different environments of these cells cause the cells to express different patterns of metabolism in accordance with their own developing programs of gene function. These patterns of metabolism create new chemical environments for nearby cells and these changed environments induce yet new programs of gene function in responding cells. Thus a progressive series of reciprocal interactions is established between the cellular environment and genome of each cell. These interactions drive the cell along a specific path of differentiation until a stable equilibrium is reached in the adult. Thereafter little change occurs in the specialized cells and they become remarkably refractory to changes in the environment. They seem stably locked into the terminal patterns of gene function characteristic of adult cells. The genome seems no longer responsible to the signals that were effective earlier in development.” (Markert CL, Devel Genet 1984;4:267–279).

Since the report of the cloning of the sheep, Dolly (Wilmut et al., 1997), the public and the scientific community were made aware of the existence of the stem cell and some of its potentials. Clearly, the concept of stem cells existed in developmental biology, embryology and even cancer biology for many decades. Nevertheless, the idea that cells, taken from an adult animal, such as sheep, could be used in such a manner to “re-create” the exact genetic individual suggested that the differentiated cell could reverse all the developmental molecular processes that converted a primitive fertilized egg to a specialized differentiated cell. To most, Dolly represented the possibility that “time's arrow” could be re-directed or “reprogramming” in midair to return to the bow. This led to an incredible re-focusing scientific research in almost all the life science disciplines (and many other scientific disciplines, such as engineering).

With a tsunami of scientific discoveries in the last decade, such as the isolation of human embryonic stem cells (Shamblott et al., 1998; Thomson et al., 1998), philosophical, religious, legal, political, scientific, and medical issues arose, primarily around the fact that, to obtain these cells, destruction of the embryos had to occur. Nevertheless, even with some available human embryonic stem cells, many scientific and practical medical issues arouse (Eggleson, 2012). Because of the perceived potential use of human embryonic stem cells, new approaches were sought. Some of the potential uses included (a) the basic research on how genes were regulated to generate all the differentiated cell types in the human body and how they were homeostatically regulated to form a functional adult human being; (b) the repair and regeneration of diseased, damaged and aged tissue (regenerative therapy); (c) screening for new drugs; (d) assessing the toxicity of physical, chemical and biological agents; (e) alleviating some genetic diseases; (f) studying the cancer process and treating cancers; and (g) studying how to treat aging and the diseases of aging.

With the exciting prospect of using human stem cells for any of this potential uses, and limitations of the use of human embryonic stem cells, there was an incentive to find alternative methods of generating human embryonic-like cells. Techniques like somatic nuclear transfer (Eggleson, 2012) were attempted. Nevertheless, the biggest technical breakthrough came from the discovery by Yamanaka and his collaborators (2006), when they recovered embryonic-like cells, after exposing primary human primary cultures to a number of genes that had been previously shown to be expressed in embryonic stem cells, such as Oct4, Sox 2, KLF4, and c-MYC. Thereafter, with multiple mixes and matches of this class of genes, and various cocktails and conditions, similar recoveries of these “induced pluripotent stem” (iPS) cells have been published in all kinds of laboratories and commercial firms. Therefore, the reality that cells, with the main properties of embryonic stem cells, could be isolated, demonstrated that the common isolation of iPS cells was validated.

Therefore, this article will not address the fact that these rare cells cannot be isolated, but will address the origin of these cells. Unlike most articles on the topic of “iPS cells” and their potentials and weaknesses, this analysis comes from a different perspective. That perspective comes from a lifetime of studying the origin and processes of human carcinogenesis. What follows, ultimately, comes out of old observations, linking development and cancer, and the early ideas that “cancer was a stem cell disease” (Till, 1982), a “disease of differentiation” (Markert, 1968; Pierce, 1974), and of “oncogeny as partially blocked ontogeny” (Potter, 1978). Indeed, one of the alternative views of the origin of cancers proceeded the current idea of iPS (Takahashi and Yamanaka, 2006), in that cancer and “cancer stem cells” were the result of “de-differentiation” or reprogramming of somatic differentiated cells (Sell, 1993). Moreover, the concept of the evolution of multicellular organisms, stem cells and cancer, which seems to have led to the “Warburg metabolism” of cancer cells, seems to bring something to the understanding of the origin of human iPS cells (Trosko, 2008B).


To begin, it must be stated that there is no universal consensus on how cancers are derived from normal human cells. Nevertheless, there are some commonly agreed upon facts from (a) experimental animal experiments, (b) in vitro studies, (c) genetic analyses of human syndromes that predispose individuals to cancers, and (d) epidemiological studies. In general, aside from a distinct class of cancers (teratomas, more on this later), carcinomas and sarcomas are generated via multiple steps, and multiple mechanisms underlying each step (Weinstein et al., 1984; Pitot and Dragon, 1991). The first step, the “initiation” event appears to be irreversible. The irreversible event on the molecular level could be explained as the result of a gene or chromosomal mutation. However, it must not be discounted that a stable “epigenetic” event in critical genes could also result in the initiation event in a single cell (Trosko, 2008A).

Nevertheless, what might be the biological function of initiation? In the attempt of decades of trying to study carcinogenesis in vitro, since one cannot always extrapolate from animal carcinogenesis studies to human carcinogenesis, nor easily extract mechanistic information from whole animal studies, there has been little or no success in the study of normal human fibroblasts or epithelial in vitro studies (DiPaolo, 1983; Rhim, 1993; Kuroki and Huh, 1993). It was that observation that seemed to puzzle most in the field of in vitro carcinogenesis. Why was it so difficult to neoplastically transform human cells in vitro, yet neoplastically transforming rodent cells appeared to be relatively easy?

The observation by Tsao and his coworkers (Nakano et al., 1985) performed experiments to answer the question: “Why was there no consistency in the neoplastic transformation of Syrian hamster embryo (SHE) cells? Ultimately, it seemed that the answer resided in the fact that, in some cultures, in which one could see “contact-insensitive” cells, were the ones that would give rise to neoplastically transformed cells, whereas, in cultures that did not have clones of contact-insensitive cells, one never found neoplastic transformed cells. That phrase, “contact-sensitive,” immediately led to the early observation that normal cells could communicate with each other via gap junctions to “contact-inhibit” their growth (Eagle, 1965), whereas, cancer cells did not have growth control (or the ability to terminally differentiate or apoptosis or senesce) or contact-inhibit each other (Borek and Sachs, 1966). They, also, lacked functional gap junctions (Loewenstein and Kanno, 1966). This led to the idea that possible the contact-insensitive cells in the primary cultures of the Syrian hamster embryo cells might be stem cells that did not have functional gap junctions (Trosko et al., 2000).

Even with the recent success in getting normal human epithelial cells to grow, with limited lifetimes, in vitro, no success was obtained in neoplastically converting these normal cells. When it was discovered that exposing these normal, primary cultures of human cells to “immortalizing” viruses, such as SV40 or HPV (Bryan and Reddel, 1994; Viallet et al., 1994), one could obtain rare “immortal,” but not tumorigenic, human cells. This occurred after exposure to these viruses or their oncogenic genes (e.g., large T antigen gene or E6–E7 genes), after the bulk of the exposed cells went through “crisis” and only a few cells survived and eventually grow out as “immortalized” clones. It seemed obvious that normal cells were mortal, after it was shown that human fibroblasts consistently had a limited proliferative capacity (Hayflick, 1965), whereas cancer cells were deemed immortal, at least in vitro. These immortalized clones could, then, be subsequently exposed to other “carcinogenic” agents (radiation, chemicals or other “oncogenes” to become neoplastically transformed).

Land et al. (1983) showed that one must first immortalize a normal rodent primary culture of cells with an oncogene, such as myc, then, after cloning out one of the immortalized cells, by transfecting it with the ras oncogene, the cells, subsequently, became neoplastically transformed. These kinds of studies only reinforced the prevailing paradigm in the field of carcinogenesis. The paradigm stated that the first step in the multistep process involved the “immortalization” of a normal, “mortal” cell, where upon, that immortalized cell could then proliferate enough times to accrue all the changes it needed to invade and metastasize. Those were the “hallmarks” of cancer (Hanahan and Weinberg, 2000, 2011).

Taking together the Nakano et al. (1985) and Land et al. (1983) observations, it was thought that, possibly, the reason one could not neoplastically transformed human primary fibroblast or epithelial cells in vitro was because these cultures lost their few adult stem cells. In other words, if carcinogenesis had to have a few adult stem cells in the culture, yet the manner in which these cultures were grown [high oxygen tension (Csete, 2005; Pervaiz et al., 2009; Mohyeldin et al., 2010), high calcium levels (Kolly et al., 2005)], and on plastic surfaces, these adult stem cells might have been eliminated. If there are no target adult stem cells in culture, there would be no initiated cells.

Since it is assumed that adult tissues, such as the skin, intestine, etc., had to have adult stem cells, the question is: “How to isolate them and what might be the characteristics that allow them to remain in their tissues, surrounded by their differentiated daughter cells, known to have functional gap junctions?” It was this line of thinking that led to the design of the “Kiss of Death” assay to select out of normal tissue, containing a few stem cells, many progenitor, finite life-span cells and the terminally differentiated cells (Chang et al., 1987).

When this strategy was implemented, a few clones, which did not have function gap junctional intercellular communication (GJIC) (Trosko et al., 2000), and, which could differentiate into cells of the organ from which they were originated, were subsequently shown to be adult stem cells, expressing both stemness genes, such as OCT4A and organ-specific markers, such as estrogen receptors in the breast stem cells (Tai et al., 2005). More importantly, in a series of human breast stem cells, it was clearly demonstrated that the OCT4A-expressing and connexin43 (Cx43) nonexpressing cells could be prevented from differentiating (or “mortalizing”) into normal beast epithelial cells (non-OCT4A expressing, Cx43 expressing) when transfected with the SV40 large T (Kao et al., 1995). Nevertheless, when the OCT4A-expressing and Cx43 nonexpressing cells were X irradiated, a few weakly tumorigenic clones were obtained, which still expressed OCT4A and did not express Cx43 (Kang et al., 1998). These weakly tumorigenic cells, when transfected with the Neu/Erb2 gene, became highly tumorigenic (Fig. 1).

Figure 1.

The composite of these three figures illustrated that (a) the clonally derived normal human adult breast stem cells expressed Oct4 or Type 1 (a) via immunohistochemical use of fluorescent antibodies to Oct4 in the normal breast stem cells, but not in the differentiated breast cells (c), still expressed in SV40 immortalized normal breast stem cells which were not tumorigenic (d), still expressed in the irradiated and weakly tumorigenic breast stem cells (e) and expressed in the neu/ErB-2 highly transformed clone (f). The young Type 2 cells were partially differentiated normal human breast stem cells. No RT-PCR was performed on these young Type 2 cells (g). In (g), the RT-PCR data correlated with the immunohistochemical data of Oct4 in this series of human breast adult stem cells. In (h), the data clearly show that the normal breast adult stem cells (Type 1) cells have no functional GJIC as measured by the fluorescent scrape loading/dye transfer technique. Type 2 or differentiated Type1 cells were derived from Type 1 or adult human breast stem cells after the Type 1 cells were exposed to cholera toxin (Kao et al., 1995). Permission granted by Karger Publishers.

Importantly, these cells still expressed their Oct4 and did not express their Cx genes. Throughout the tumorigenic process, the original breast adult stem cells, their derivative blocked mortalized cells and their weakly and highly tumorigenic clones, never transcriptionally shut down the OCT4A gene. In effect, these highly tumorigenic cells did not “de-differentiate” or reprogram their OCT4A gene. What happened was that the OCT4A expressing adult and immortal stem cells were blocked from “mortalizing.” This seems to be the functional basis for the first step, or initiation step of carcinogenesis. In other terms, initiation is the inhibition of the asymmetrical step of a stem cell's repertoire of cell division (Fig. 2). In other words, once stimulated to proliferate, it can only proliferate symmetrically to increase the number of “initiated” cells.

Figure 2.

In this diagram, a normal adult stem cell is shown dividing asymmetrically to form one daughter that is committed to ultimately terminally differentiate. The other daughter is designated to be identical to its mother adult stem cell (Oct4 +). If that adult stem cell is exposed to some condition that prevents asymmetrical cell division but does not suppress the Oct4 expression, it is operationally an initiated cell. That is, if mitotically stimulated to divide, it divides symmetrically to form two initiated, nonterminally differentiated cell. Initiation is, then, defined as the process that prevents an immortal normal adult stem cell to terminally differentiate or become “mortal.” These adult initiated stem cells are still Oct-4 positive or benign cancer stem cells. As these initiated Oct4 + cells are stimulated to proliferate and resist apoptosis, the growing benign tumor microenvironment changes, some of these initiated Oct4 + cells can partially differentiate into cancer nonstem cells (Oct4 negative). Eventually, additional stable mutational or epigenetic events occur, providing the benign Oct4 + cancer stem cells to become invasive, metastatic cancer stem cells. Permission granted by Nova Science Publishers, Inc.

This, then, brings us to the original question as to whether the cancer cell was derived from the immortal, normal adult stem cell or from the de-differentiation or reprogramming of a somatic differentiated cell. From Fig. 1, it was clear that SV40 could not immortalize these mortal, normal epithelial nonstem cells, as no immortalized clones were isolated from these experiments. Subsequently, without any clones that had re-expressed OCT4A or de-repressed Cx43, no clone regained the phenotype of an immortalized, “embryonic like” stem cell.

From a different line of reasoning, since human cancers do exists in adults as carcinomas or sarcomas, and since iPS cells, derived in vitro from primary human cultures, by definition, form teratomas when injected back into an adult organism, it seems illogical to assume, in vivo, the carcinogenic process of initiation starts from an adult somatic differentiated cell. That is, also, one reason that the therapeutic use of differentiated cells, derived from embryonic or iPS cells, has this precaution, since, if just one ES or iPS cell is injected with their derived differentiated cells, a tumor could result. That would suggest the reprogramming of a somatic differentiated fibroblast or epithelial cell had to suppress the connexin genes, needed for growth control and differentiation/apoptosis (Trosko, 2009), and to derepress the Oct4A gene needed for “stemness.” If that did happen in vivo, that reprogrammed cell would have formed a teratoma, not a carcinoma or sarcoma.

Lastly, given that, to date, the efficacy of generating iPS cells is rather low, this would imply the process to reprogram the exact set of genes to restore the original ES gene expression is extremely complex or that the cells that represent the iPS clones were derived from the low frequency of adult stem cells in adult organs/tissue. It is well known that adult stem cells do exist in organs, such as skin (Tumbar et al., 2004), intestine (Barker et al., 2010), breast (Chang et al., 2001), and blood (Till, 1982). The fact the iPS cells retain a molecular fingerprint of the tissue from they were obtained (Hochedlinger, 2010; Kim et al., 2010), also, suggests that they were derived from committed organ-specific adult stem cells. This was also, demonstrated when intestinal adult crypt stem cells were the origin of intestinal cancers (Barker et al., 2009). Therefore, by accepting the origin of iPS as being the result of reprogramming, while ignoring the observations that adult organ-specific stem cells exist and that these organ-specific stem cells can give rise to cancers, one could perpetuate a false paradigm.

In summary, of the two opposing hypotheses concerning the origin of the cancer cell, it seems that the adult stem cell hypothesis is supported by the fact that organ-specific adult stem cells exist in most organs. The fact these adult stem cells remain immortal until they are induced to terminally differentiate, the first step in the multistep, multimechanism process of carcinogenesis would appear to be the blockage of differentiation or “mortalization” of the stem cell or the blockage of the stem cell's ability to divide asymmetrically. In effect, this interpretation challenges the prevailing idea that the first step of carcinogenesis is the immortalization of a normal, mortal somatic differentiated cell. This, then, suggests that initiation of carcinogenesis is not the result of reprogramming of a somatic differentiated cell. If reprogramming of a somatic differentiated cell, in vivo, occurred, then that iPS cell would, by definition, have to produce a teratoma, rather than a carcinoma or sarcoma. Consequently, while iPS cells can be produced, in vitro, by several means, it remains to be shown that iPS cells can be produced in vivo. Further, because the origin of iPS cells has been shown to originate from a subset of rare “MUSE” cells in skin fibroblasts, and that adult stem cells do exist in skin fibroblasts (Wakao et al., 2011), the likelihood is high that adult, organ-specific stem cells can be a target to initiate the cancer process.


While it can be argued that trying to test the Spontaneous or Elite hypotheses of the origin of iPS cells or embryonic-like stem cells makes no sense by using the study of the origin of cancer, then another indirect approach might provide another insight. To do this, the possible conditions during the evolution of life on a planet, during the changing physical conditions, might provide clues to how the multicellular human being, with its germinal and somatic stem cells, were ultimately derived from a single cell organism. How did these stem cells, with their unique biological characteristics, allowed them to maintain the species and to give rise to the multicellular human, who does age and exhibit diseases of aging, such as cancer?

To set the framework of the reasoning that the evolution of earth's atmospheric and aqueous gases played a large part in the acquisition of adaptive properties of the human stem cells that led to the survival of the species, as well as the development of the multicellular properties of the individual and his/her aging and diseases of aging processes (Trosko, 2008AA), the philosophical questions to shape this inquiry are:

“Can an uncontrolled proliferation of the earliest form of a single cell organism, in a microenvironment of adequate nutrients, temperature, etc., be likened to the uncontrolled growth of a cancer cell in a living human being?” and “Is there any evolutionary link, and if so, what might that link be?” In other words, “Is cancer an evolutionary throw-back to the properties of the single cell organism?”

From the paleochemistry of the early earth's air and water, it seems that a hostile environment existed, to which most current life forms would not have survived. Clearly, temperature, available chemicals (as potential nutrients), radiation levels, gravity, and atmospheric gases, set limits to which modern living systems had to find compatible to the origin of life. However, as the physical aspects of the earth started to change, the evolution of genes and energy metabolism for life of the single cell organism started to emerge (Saul, 2008). The relative short range of extremes in temperature, specific ratios of ambient gases, the change in seasons, and diurnal cycles of light selected specific biological evolutionary mechanisms, that led to the generation of energy from available nutrients, for life to begin.

When the oceans achieved the right conditions, the first single cell organism was born that metabolized sugars via glycolysis in an anaerobic environment (Saul and Schwartz, 2007).After the appearance of phyto-organisms, which helped to change the paleochemistry of the oceans by producing oxygen, and the symbiotic union of the mitochondrion with the first multicellular organism, dramatic adaptive features appeared. These simple organisms could now produce collagen (needed for cells to stick together), since oxygen is needed for its synthesis.

Within this newly oxygenated aqueous environment, the first multicellular organism, a social collection of mitochondrial-containing cells, now had a means, i.e., oxidative phosphorylation, to generate energy from sugars. This required the co-evolution of genes to cope with the negative side effects of oxidative phosphorylation, namely, the generation of a number of reactive oxygen species (ROS's)/reactive nitrogen species (RNS) (Pervaiz et al., 2009) and to utilize them to act as adaptive signal transducers and gene regulators (Upham and Trosko, 2009; Brigelius-Flohe and Flohe, 2011; Rigoulet et al., 2011). Since the biological evolution of both the individual and the species depended on maintaining the integrity of genomic and mitochondrial DNA, genes had to be selected that protected these DNA's from the highly reactive ROS's (Dang, 2012). Consequently, it seems evident that a series of genes had to be co-evolved (a) to protect the DNA's from ROS-induced macromolecular damage (endogenous antioxidants), and (b) to repair the inevitable damage that might occur (DNA repair mechanisms). Extracellular matrices and the “niche” in this simple multicellular organism, needed to sequester the unique specialized cells, the germinal and somatic stem cells (Nursall, 1959; Saul, 2008; Mohyeldin et al., 2010; Ozbek et al., 2010) were selected for to provide a microenvironment to help maintain the conditions of “stemness.”

Protecting and repairing genomic and mitochondrial DNA were, and are, a critical balance for both the survival of the individual and the sustaining of a species. This is seen when the protective system for the DNA's integrity is close to being perfect, the chance for survival would be small, because inevitable changes in the environment would create nonadaptive conditions for a nonadaptive genome of the species. There has to be the chance of a few adaptive mutations in the genome, as well as some epigenetic alterations in gene expression in a few individuals of the species population. Obviously, if the coded genetic protective mechanisms and DNA repair mechanisms allowed too much DNA damage and too many mutations to be formed, nonadaptive functioning at both the individual and species levels would occur. The selection of both the protective and repair systems in the early multicelled organism allowed the frequency of germinal and somatic mutations to be sufficient for the individual organism to survive long enough to reproduce and to allow for the offspring to survive to reproduce in an ever-changing physical environment. Individuals, who had mutation frequencies which were too high, would incur mutation-related diseases, which would jeopardize his/her ability to reproduce and to maintain the survival of the species. It has to be emphasized here that mutations also occurs without DNA damage and faulty DNA repair, but also via “errors” in replication

The transition of the single cell organism to the first multicell organism had to be assisted by a number of newly selected adaptive phenotypes. The first had to be a means to control cell proliferation. While the presence or absence of nutrients proper temperature, appropriate atmospheric gases, radiation levels, etc., which controlled the growth of single cell organisms, in a multicell organism, uncontrolled growth within an individual would ostensibly end up as a tumor. That first phenotype was “contact inhibition” (Eagle, 1965), which allowed two cells in direct contact with each other to send signals to control their cell proliferation. The second phenotype, which appeared, was a means to regulate sets of genes of the total genome to bring about the specialization of some cells carrying the same genomic information of all cells, but having the ability to express only those genes needed to generate unique functions (muscles, neurons, hepatocytes, germ cells). This phenotype was differentiation. The third new phenotype was the ability to remove, selectively, damaged or nonadaptive differentiated cells during specific periods of development (apoptosis). Senescence of cells/tissues/organs was the fourth new phenotype that led to the finite life span of each species. Each species' life span is different; however the life span needed to be long enough for the individual to become sexually mature and to allow survival of the offspring to reach sexual maturity. Most importantly, for the purpose of this article, was the formation of the fifth phenotype, namely the formation of germinal and somatic stem cells (Trosko, 2011). While this unique cell type is characterized as sharing one form of cell division with its evolutionary single cell precursor, namely to divide symmetrically, to form two daughter cells who are identical to their mother cell, the stem cell could also divide asymmetrically, under certain signals, to produce one daughter identical to the mother cell and one daughter that is destined to differentiate into a mortal, specialized cell. These two stem cell types (germinal and somatic) were selected to pass on adaptive expressed genes that allowed the individual to survive the prevailing environment, so that it could reach the reproductive stage and to survive long enough to protect its offspring and species to reach reproductive age. To provide cells needed for growth, cell replacement and wound healing, and differentiation of specific specialized cells/tissues, somatic stem cells were needed. Lastly, to establish a unique microenvironment to sequester these stem cells to keep them in their “stemness” stage, the sixth phenotype was needed. That unique home for the stem cells was the stem cell “niche” (Fuchs et al., 2004; Tumbar et al., 2004). This unique microenvironment for both the germinal and somatic adult stem cells had to provide a hypoxic in situ environment to keep the stem cell in its stemness state (Kang and Trosko, 2011) (More will be discussed on this critical point.).


Can the Warburg metabolism of cancer tell us something about the origin of the iPS cells or the resolution of the Spontaneous and Elite hypotheses of the formation of iPS cells? Since this article is trying to resolve this fundamental issue via the study of the stem cell theory of cancer, rather than from a strict developmental approach. One of the early observations on the nature of cancer tissues was that it metabolized sugar differently than normal tissue (Warburg, 1956a). Given the general description of those commenting on the appearance of the Warburg metabolism when normal tissues become cancerous, one usually sees the phrase, “the cancer cell reverted from oxidative phosphorylation to glycolysis…,” as though the normal state of tissues is oxidative phosphorylation. Therein lays the problem. Neither the “normal” tissue nor the “normal” tumor tissue consists of a homogeneous population of cell types. Within each normal tissue, there exist multiple differentiated cell types, progenitor cells and a few adult-organ-specific stem cells. Each manifests different expressions of genes from their shared genomes; however, more importantly, each can metabolize sugars differentially, as will be documented below.

Today, we accept that there exists in normal tissues, a few stem cells, and also, we generally assume that cancer cells originated from a single cell (Fialkow, 1976; Nowell, 1976). If the single normal cell in the tissue, that gave rise to the cancer, which is a mixture of cancer stem cells and “cancer nonstem cells” (as well as some normal invasive macrophages, normal stromal cells), it is important that we determine the nature of that cell, not only for academic reasons, but for very important practical reasons. Therefore, in measuring the metabolism of glucose in either normal or cancer tissue, one is measuring a mixture of both types of glucose metabolism. The predominance of one type over the other would reflect the ratio of the two types of cancer cells in tumors (which might have a predominance of the “cancer cells” over “noncancer cells”), but in the normal tissues, the metabolism in the few normal adult stem cells would be masked by the oxidative phosphorylation of the predominant normal differentiated cells.

An attempt will be made to assess how (a) the evolution of the single cell organism to a multicell organism; (b) the transition came about from an hypoxic metabolism of sugars via glycolysis to a mitochondrial-dependent oxidative metabolism; (c) the forced selection of germinal and somatic stem cells occurred; (d) stem cells were developed with few mitochondria to generate energy via glycolysis, whereas their differentiated offspring have many mitochondria and metabolize via oxidative phosphorylation, and (e) tumors came to metabolize via the Warburg metabolic scheme.

Starting from the early forms of life in the form of single cells, that metabolized sugars for energy for life via glycolysis in an oxygen-deficient environment, the emergence of mitochondria-containing organism in an oxygenated environment afforded a means to generate energy from sugars, together with the production of ROS's. Whatever made multicellularity adaptable to increase survival in an ever changing environment, the adhesion of single cells into a social community, the appearance of differentiation ultimately led to the metazoan. This forced the emergence of both germinal stem cells and somatic organ-specific stem cells to protect the species and the individual organism, as well as the protective niche microenvironments from the oxygen tension needed by the progenitor and differentiated cells and from the production of ROS's.

One simple evolutionary-strategy was for the stem cells, which are defined by their ability to divide both symmetrically and asymmetrically, to control the metabolic mechanism by which means to divide. In other words, during development, the amplification of stem cells was needed for growth of a tissue, symmetrical cell division was regulated. On the other hand, when differentiation was required, a shift in cell division was made to asymmetrical cell division. Because of the possibility of errors of DNA repair or of errors of DNA replication, the stem cells normally would pass on the responsibility of symmetrical cell division to the progenitor or “transit amplifying cells” (Juckett, 1987), which are characterized by their finite life span in an oxygenated environment. This would reduce the chance that detrimental mutations in the stem cells for the species and for the individual. In effect, as many have argued (Krtolica et al., 2001; Krtolica and Campsi, 2002; Campsi, 2003), the progenitor cells, by accruing mutations, were a way to minimize the lethal consequences of mutations. Clearly, mutations do occur in both germinal and somatic stem cells. These can and do generate both some adaptive genes for both the survival of the species and the individual. However, too many mutations in the stem cells would be detrimental to both the species and individual.

Clues to the origin of the cancer and to the origin of the iPS cells might now come from examining the metabolic characterization of normal stem cells and cancer stem cells. To put the fundamental biochemical issue up front, glucose can be oxidized and be converted to pyruvate via either glycolysis, followed by fermentation, to become lactate or by complete oxidation with mitochondrial respiration (Gatenby and Gilles, 2004). To re-iterate, it has been shown that normal embryonic stem cells have far fewer mitochondria than their differentiated daughter cells (Nesti et al., 2007) and that mesenchymal stem cells metabolized glucose by anaerobic glycolysis than by oxidative phosphorylation (Chen, 2008). Moreover, in a glucose restriction model, it was observed increased proliferation ability, increased antioxidant defense ability, and increased aerobic metabolism in mesenchymal stem cells (Lo et al., 2011). The limited life span progenitor cells metabolize glucose for energy via the tricarboxylic acid cycle and oxidative phosphorylation. This means, as during the finite life span of the proliferative progenitor cells, their mitochondria pick up many mutations in their mitochondrial genome. It matters little, in terms of cancer, as these cells will not become immortal and they senesce or die.

Consequently, the critical question is: “What happens to the mitochondria during the reprogramming of the differentiated somatic cells in the iPS cells. Interestingly, two studies were designed to analyze this question in the embryonic stem cells, the iPS cells, and the differentiated cells from which the iPS cells were obtained (Nesti et al., 2007; Armstrong et al., 2010). The results of these studies, which examined both the quantity and quality of the mitochondria in embryonic stem cells, the iPS cells and the original fibroblast population from which the iPS were derived, showed that the quantity and quality of the embryonic and iPS cells' mitochondria were almost identical, but they were different from the original differentiated somatic fibroblasts.

Further, since embryonic stem and iPS cells had few mitochondria, this suggested that both of these stem cells survived via an aerobic glycolysis, unlike the differentiated somatic fibroblasts, which had many mitochondria and utilized aerobic oxidative phosphorylation. At this point, normal embryonic stem cells and the normal iPS cells share a common means to generate energy via glycolysis. Moreover, the mitochondria of the embryonic and iPS cells had both few mitochondria with few mitochondrial mutations, compared to the many mitochondria with many mutations.

Interpretation of these results is critical. For if the iPS cell was reprogrammed at the genomic DNA level by epigenetic means, and it lost mitochondrial numbers during this reprogramming process, it might suggest that reprogramming at the genomic level caused the genomic control over the numbers of mitochondria. It also might suggest that only the unmutated mitochondria survived this reprogramming process. If one accepts this interpretation, one would have to invoke this interpretation, because one cannot “reprogram” mutations in the mitochondria. Back mutations are the only way one could restore the mitochondrial DNA to the unmutated state. That would be impossible for all the different kinds of mutations in the many mitochondria in the somatic fibroblasts. Nevertheless, there is another interpretation, in that, if the iPS cells were derived from the rare adult stem cells in the primary cultures, oxidative phosphorylation-induced ROS's would be minimal from their few mitochondrial DNA. It seems that, using Occam's razor, this later hypothesis would be more likely correct. In principle, it might be possible, epigenetically, to reprogram normal genomic DNA expression, but one cannot, epigenetically, reprogram mutated mitochondrial DNA.

To test this hypothesis that the iPS cells were derived from rare adult stem cells in adult tissues, the prediction would be to show that the frequency of iPS cells would be higher if one could introduce the embryonic genes (i.e., OCT4, Sox 2, etc.) into a large population of the rare adult organ-specific stem cells than from a primary population of somatic cells, in which there are a few adult stem cells. The recent demonstration of a rare population of cells (MUSE cells) in skin could preferentially be converted into iPS cells than in the non-MUSE cells of the skin (Wakao et al., 2011). Clearly, the frequency of adult stem cells in primary cultures would be higher in early primary cultures than late passage primary cultures or even in primary cultures of young donors as opposed to very old individuals. Given that growing progenitor and stem cells in vitro under low oxygen and high antioxidants (Linning et al. 2004; Csete, 2005; Pervaiz et al., 2009; Mohyeldin et al., 2010; Brigelius-Flohe and Flohe, 2011) should increase the frequency of recovering iPS cells, than under normal in vitro culturing conditions of high oxygen tension and few antioxidants in the medium. What has yet to be performed is the use of these embryonic genes to see if the frequency of iPS cells would be higher in a pure population of other organ-specific adult stem cells (e.g., liver, breast, and pancreas).

Since the original hypothesis was to examine the question as to the origin of iPS cells from both the developmental biology field and from the cancer field, we must now ask: “Are there similar observations from the cancer field that would support the idea that ‘reprogramming’ of somatic differentiated cells do not give rise to the cancer cell ?” As noted before, no one has neoplastically transformed a normal fibroblast or epithelial cell with a single oncogene or with any known physical or chemical carcinogen (DiPaolo, 1983; Rhim, 1993; Kuroki and Huh, 1993). Only by, immortalizing a primary culture of human cells with an oncogene or “immortalizing virus, could one subsequently obtain neoplastically transformed cells with other activated oncogenes or carcinogens' (Land et al., 1983). However, looking at the problem differently, since this process of obtaining an immortalized cell from a primary culture, it might suggest that the immortalized nontumorigenic cell, rather than being reprogrammed from a mortal state to an immortalized state, the process actually selected out of the primary culture the few rare adult organ-specific stem cells, which are naturally “immortal.” The “immortalizing viruses” actually did nothing to reprogram the differentiated fibroblasts or epithelial cells (which died when the culture went through crisis). However, the immortalizing genes of these viruses actually rendered nonfunctional the critical gene products (e.g., p53 and RB) needed for differentiation of the few stem cells (Trosko, 2003, 2006, 2009; Trosko and Tai, 2006). As a result, only a few immortalized cells were isolated (about the same frequencies seen in the generation of iPS cells from primary cultures).

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 were derived from normal adult organ-specific stem cells that persisted to metabolize via the Warburg pathway and retained normal elements of the tissues, from which they were derived during the evolution of the multistep, multimechanism process of carcinogenesis. Besides the observation that only the adult organ-specific human breast stem cell was able to be neoplastically transformed (Chang et al., 2001), it has been shown that genomic mutations found in human colorectal cancer could generate intestinal carcinomas in mice only when forced to occur in the stem cells (Barker et al., 2009).

To get at another set of experimental observations that supports the hypothesis that adult organ-specific stem cells, which metabolize via the Warburg process, are the target cell for cancer, another set of genes seem to correlate with the OCT4A gene of normal ES -, iPS-, normal adult organ specific adult-, and “cancer” stem cells and the Warburg metabolism (Trosko and Kang, 2012).

A major observation was reported that showed lactate dehydrogenase A (LDHA) appears to be a direct marker to trace the evolution of a Warburg metabolizing cancer cell (Le et al., 2010; Hirschhaeuser et al., 2011). Lactate dehydrogenase A is needed for the conversion of pyruvate to lactate, coupled with the recycling of NAD+ (Semenza et al., 1996). In this study, they showed that reduction of the LDHA gene by siRNA or a chemical inhibitor could induce oxidative stress and inhibit tumor progression. The previous studies, showing that reduction of LDHA reduced neoplastic transformation and delayed tumor formation, gave rise to the idea that LDHA was critical for tumor initiation (Fantin, 2006; Xie et al., 2009).

Therefore, the Le et al. (2010) studies demonstrated that induction of oxidative stress, due to increased ROS production via the reduction of LDHA activity, seems to run counter intuitively. However, when one reviews some of the literature of how classic tumor promoters, such as phorbol ester, which induces oxidative stress (Cerruti, 1985), cause initiated cells (presumably, adult stem cells) to proliferate and blocks apoptosis of these initiated cells, yet, can induce differentiation in cancer cells (Huberman and Callahan, 1979), one can surmise oxidative stress is an epigenetic inducer of signaling mechanisms to alter gene expression (Finkel, 2003; Upham and Trosko, 2009). Persistent exposures to tumor promoting agents or to stable downregulation of p53 and RB cause oxidative stress and oxidation-stress induced signaling (Finkel, 2003; Lebedeva et al., 2009; Upham and Trosko, 2009), for mitogenesis, not mutagenesis. On the other hand, exposures to antioxidants can ameliorate the oxidative stress induced by tumor promoting agents (Leone et al., 2012) (Fig. 3).

Figure 3.

Scheme of postulated actions of phytochemicals in cancer chemoprevention. Transformation of the normal adult stem cell into an initiated cell starts the carcinogenesis process. Initiated cells might undergo the tumor promotion into preneoplastic cells by a number of means, such as chronic inflammation, surgery, cell killing, growth factors, and exogenous chemicals that can cause initiated cells to proliferate and to inhibit apoptosis. During this promotion phase, other changes that allow a single initiated cell to invade and to metastasize, constitutes the progression phase. The malignant tumor consists of a mixture of cancer stem cells and cancer nonstem cells. Phytochemicals might interfere with different steps of this process. Some chemopreventive phytochemicals inhibit metabolic activation of the procarcinogens or stimulate the detoxification systems, avoiding their subsequent interaction with DNA or other cellular target and therefore blocking tumor initiation (blocking agents). Other phytochemicals might act by suppressing the later steps (mostly promotion) of multistage carcinogenesis (suppressing agents). Some phytochemicals can act as both blocking and suppressing agents. Permission granted by Springer Publishing, Phytochemistry Reviews Journal.

Again, this calls for careful interpretation of what various chemicals do to the different cell types in the tissue (normal adult stem cells, initiated adult stem cells, normal progenitor cells, normal progenitor and normal differentiated cells, cancer stem cells and cancer nonstem cells). With newer technologies to measure both the expressed gene proteins for stem cells and for these metabolic enzymes and metabolic activity on a single cell basis, it might be possible to resolve the problem, namely, “Did the original target cell for the initiation of the carcinogenic process get a differentiated cell to revert back from the oxidative phosphorylation state to the glycolytic state?” or “Did the original adult stem cell, which metabolizes via glycolysis and has high LDHA activity, get frozen in this state through out the whole promotion/progression phase of carcinogenesis?”

Related to the normal reduced number of mitochondria and of ROS-induced oxidative stress in normal stem cells because of glycolysis and high activity of LDHA, the fact that the loss of p53 has been shown to be associated with the mitochondrial DNA and altered mitochondrial ROS homeostasis (Lebedeva et al., 2009), suggests another link, connecting the adult stem cell to the cancer stem cell. The fact the SV40 can “immortalize” normal human breast adult stem cells to the initiated state, where the stem cell has its p53 and RB gene products rendered nonfunctional by the large T antigen, the stem cell would not be shutting off its LDHA gene and the number of functioning mitochondria to shift the cell's metabolism from glycolysis to oxidative phosphorylation. This keeps the immortalized or initiated adult stem cells in a low oxidative stress state and to continue metabolism via the Warburg metabolism. Since the mammalian mitochondrial proteome comprises about 1130–1500 genes, of which mitochondria mtRNA encodes only 13 genes, the majority of the mitochondrial genes are nuclear and are, in large part, related to p53 function (Lopez et al., 2000; Ma et al., 2007; Pagliarini, 2008; Scarpulla, 2008). Therefore, measuring the expression of LDHA in normal adult stem cells, in phorbol ester-treated normal differentiated cells, in immortalized (initiated), but not tumor tumorigenic cells, in tumorigenic cancer stem cells and in cancer nonstem cells, could help to resolve this major question of the origin of cancer stem cells. It should also be performed on the embryonic stem cells, the normal differentiated cells and on the iPS cells derived from those cells. Both approaches should be used to understand the “stoichiometric” versus “elite” models of the origin of iPS cells and the “stem cell” hypothesis versus the de-differentiation or reprogramming hypothesis.

It has been previously shown that many tumors have higher LDHA levels than normal tissues (Goldman et al., 1964). The same issue, related to generalizing that a given marker, such as LDHA, exists for normal tissue, has to take into account that the normal tissue is a mixture of a few adult stem cells, progenitor and differentiated cells (normal tissues) and cancer tissues are variable mixtures of cancer stem cells (Webster et al., 2007) and cancer nonstem cells. The normal adult stem cell, which metabolizes via glycolysis, would be masked by the metabolism of the progenitor and differentiated cells of normal tissue. This is the reason why many investigators interpret the obvious differences seen between normal tissues and cancer tissues as meaning a reversion of the oxidative phosphorylation metabolism back to the glycolytic mechanism has occurred. The number of adult stem cells, which metabolize via glycolysis, is very rare in normal tissues, where most cells in normal tissue metabolize via oxidative phosphorylation, whereas the numbers of cancer cells metabolizing by aerobic glycolysis is higher than in the normal tissues.

Therefore, in summary, by examining the iPS and the cancer stem cell phenomena, the question of the origin of each type of cell reviewed, using as a common phenotype, namely, the nature of energy metabolism, seems to be consistent with the hypothesis that (a) the cell of origin for both types of cells was the adult organ-specific stem cell; and (b) the two types of cells were not derived from reprogramming somatic differentiated cells, but, rather, by either over-expressing stemness genes in the adult origin-specific adult stem cell (i.e., the iPS cells) or by inhibiting the repression of stemness genes in the adult stem cells by the “initiation process” of carcinogenesis (i.e., cancer stem cells). The fact that iPS cells form teratomas and continue to retain some expressed epigenetic “finger prints” of the tissues of origin suggest that they are relatively normal, but could not reprogram these stable adult organ-specific markers. On the other hand, the fact that this speculated reprogramming or de-differentiation of a normal, differentiated organ-specific somatic cell does not form teratomas or teratocarcinomas during the carcinogenesis process, but rather carcinomas or sarcomas, seems not to be the process by which cancer stem cells are derived.

In the context of expanding isolated stem cells, in vitro, for regenerative therapy, it has been noted that exposure to hyperoxia, not only shortens their life span, but increases genetic instability (Estrada et al., 2012). This raises another possibility that a false function was been attributed to telomeres and telomere shortening in “aging” or senescence of cells. Since most of the studies of telomeres during life span of cells in vitro or during aging have been performed when nonstem cells, such as fibroblasts, were grown at 20% oxygen or when cells were biopsied from individuals at various ages. The telomeres in these studies seemed to suggest that their erosion correlated with cell passage level or with chronological age. Quite possibly, a false interpretation of the telomere shortening has been made as to their possible role in the aging process. If the role of stem cells is (a) to maintain the genome stability for both the species and individual, (b) to metabolize via glycolysis as a means to reduce the generation of ROS's for energy production, and (c) to have a low turnover of stem cells in their niches as a means to reduce errors of DNA replication, then this was an evolutionary means to reduce genomic instability. So, therefore, the notion that senescence of the nonstem cell progenitors was a biological means to minimize cancer of those cells that needed oxidative phosphorylation for differentiation, might be misleading. Consequently, even though some higher risk of genomic instability occurred in a cell, not destined to serve in a stemness capacity, that cell, not being immortal, would never give rise to an “immortal cancer stem cell.” Resolving this possibility might be to examine early passage organ-specific adult stem cells from individuals at different ages.

In an effort to put the origin of the iPS cells and of the cancer stem cells into an integrated framework to explain the fact that stem cells (a) do not have many mitochondria, (b) do metabolize by glycolysis in an hypoxic niche, and (c) that cancer stem cells metabolize via the Warburg glycolytic process, the process of carcinogenesis, consisting of the initiation of a single normal adult stem cell, followed by the clonal expansion by the promotion process, must be considered. The promotion process is brought about by agents that induce oxidative stress, such as phorbol esters. It has been observed that, in premalignant skin epidermal cells (the presumptive expanded initiated stem cells), enhancing mitochondrial respiration suppresses the tumor promoter, TPA, -induced M2 isoform of pyruvate kinase (PKM2) expression (which is highly expressed in cancer cells) and cell transformation. PKM2 is responsible catalyzing the final step of aerobic glycolysis and an indicator of the Warburg effect (Wittwer, 2011).

Lastly, linking the later observation to a very recent experimental finding, when the tumor suppressor gene (Pten) was over expressed in mice, the mice were smaller because they had few cells, burned energy at a higher rate, the cells consumed less glucose and produced more ATP by cellular respiration and developed tumors later after exposed to carcinogens (Garcia-Cao et al., 2012). This might suggest the over expression of Pten forced adult stem cells to have less self-renewal and more asymmetric cell division to differentiate early. In terms of the multistage carcinogenic process, while the numbers of organ-specific stem cells might be lower, thus lowering the number of initiated cells, clearly the promotion phase, needing the suppression of mitochondria numbers, of differentiation or oxidative phosphorylation, and of an increase of glucose consumption, would be restricted, thus protracting the appearance of the tumors.


Stem Cells, the Warburg Hypothesis and Cancer

To bring the role of biological evolution of stem cells and of the multicellular organisms in a oxygenated physical world and to understand the origin of role of stem cells, as the origin of cancer and cancer stem cells, the critical observation of the Warburg phenomenon needs to be examined. Warburg noticed that cancer tissues metabolized via aerobic glycolysis, whereas normal tissues metabolized glucose by oxidative phosphorylation. Recently, it had been noticed that normal stem cells have few mitochondria and metabolized via glycolysis, whereas their differentiate offspring metabolized glucose via oxidative phosphorylation. Attempts to examine the number and quality of mitochondria and mt-DNA of iPS cells, embryonic (ES) cells, and normal differentiated fibroblasts were made. The observations that were made suggested that the numbers of mitochondria of ES and iPS cells were nearly identical (low), whereas the numbers of mitochondria in the differentiated fibroblasts were high. That suggested, if reprogramming of the genome fibroblast did occur, it had to reprogram the numbers of mitochondria, also. The assumption that, since oxidative metabolism was occurring in the fibroblast, ROS-induced mitochondrial DNA damage would be extensive. When the quality of mt-DNA damage was examined in the iPS compared to the ES cells' mt-DNA, the two were similar, namely, little DNA damage compared to the differentiated fibroblast. If the iPS were “reprogrammed” at the genomic DNA level, there had to be a selection to eliminate mitochondria with extensive mt-DNA damage, when there was a reduction in the number of mitochondria during this reprogramming process. However, using “Occam's razor,” a more plausible explanation for these results would be that the stem cell, with few mitochondria, would perform glycolysis metabolism and produce little ROS-induced mt-DNA damage, and gave rise to the iPS cells, which also had few mitochondria, metabolized via glycolysis, and had little mt-DNA damage. To get an iPS cell by reprogramming the genomic DNA via epigenetic mechanisms, could, in principle, occur. However, to obtain an iPS cell with few mitochondria with little mt-DNA damage or mutations cannot occur via epigenetic mechanisms. That could only occur by “back mutations.” That would be impossible. These observations actually support the adult stem cells being the origin of the iPS cells.