Cancer is widely considered a genetic disease involving nuclear mutations in oncogenes and tumor suppressor genes. This view persists despite the numerous inconsistencies associated with the somatic mutation theory. In contrast to the somatic mutation theory, emerging evidence suggests that cancer is a mitochondrial metabolic disease, according to the original theory of Otto Warburg. The findings are reviewed from nuclear cytoplasm transfer experiments that relate to the origin of cancer. The evidence from these experiments is difficult to reconcile with the somatic mutation theory, but is consistent with the notion that cancer is primarily a mitochondrial metabolic disease.
Introduction
The prevailing view today is that cancer is a genetic disease involving nuclear mutations in oncogenes and tumor suppressor genes (Hanahan and Weinberg, 2011). A typical tumor is thought to contain two to eight so-called driver gene mutations that regulate the tumorigenic phenotype (Vogelstein et al., 2013; Hou and Ma, 2014). The nuclear genomic instability seen in nearly all types of tumor cells is considered the primary cause of the cancer’s hallmarks that include sustained proliferative signaling, evasion of growth suppressors, resistance to cell death, replicative immortality, enhanced angiogenesis, and activation of invasion and metastasis (Hanahan and Weinberg, 2011). The somatic mutations thought to be the origin of cancer arise randomly during DNA replication in normal noncancerous stem cells (Tomasetti and Vogelstein, 2015). The somatic mutation theory reigns as the most widely accepted view of the origin of cancer and is the justification for developing personalized genetic therapies for managing the various forms of the disease (Vaux, 2011; McLeod, 2013; Hou and Ma, 2014). Despite numerous inconsistencies associated with the somatic mutation theory (Rous, 1959; Sonnenschein and Soto, 2000; Soto and Sonnenschein, 2004; Baker and Kramer, 2007; Burgio and Migliore, 2015), the theory is presented as if it were dogma in most current college textbooks of genetics, biochemistry, and cell biology, and is the mainstay of the National Cancer Institute in stating that, Cancer is a genetic disease?that is, it is caused by changes to genes that control the way our cells function, especially how they grow and divide (http://www.cancer.gov/cancertopics/what-is-cancer).
As an alternative to the somatic mutation theory, emerging evidence suggests that cancer is primarily a mitochondrial metabolic disease (Seyfried and Shelton, 2010; Hu et al., 2012; Verschoor et al., 2013; Seyfried et al., 2014). The view of cancer as a metabolic disease originated with the experiments of Otto Warburg (Warburg, 1956a,b; Burk et al., 1967). Respiratory insufficiency is the origin of cancer according to Warburg’s theory. All other phenotypes of the disease, including the somatic mutations, arise either directly or indirectly from insufficient respiration (Warburg, 1956a; Seyfried, 2012a; Seyfried et al., 2014). Warburg’s metabolic theory was also in line with the concepts of C. D. Darlington and others showing that cancer is largely a cytoplasmic mitochondrial disease (Woods and Du Buy, 1945; Darlington, 1948). Proponents of the somatic mutation theory, however, consider the abnormal energy metabolism of tumor cells as simply another phenotype that is programmed by proliferation-inducing oncogenes and defective tumor suppressor genes (Hanahan and Weinberg, 2011). In light of the overwhelming acceptance of the somatic mutation theory, it would be good to reconsider data from the nuclear transfer experiments that are inconsistent with the somatic mutation theory.
The rationale for the nuclear transfer experiments is to determine whether the genome of somatic cells can direct normal development (Gurdon and Wilmut, 2011). These same types of experiments can also be used to test the somatic mutation theory of cancer. If nuclear somatic mutations are the origin of cancer cells, then the hallmark cancer phenotype, dysregulated cell proliferation, should occur following the transfer of a tumor nucleus into a normal cell cytoplasm. In other words, the somatic mutations in the tumor cell nucleus should determine the tumorigenic phenotype of abnormal cell growth. On the other hand, if mitochondrial dysfunction is the origin of cancer cells, then the tumorigenic phenotype should follow the type of mitochondria in the cell. In other words, mitochondria from non-cancerous cells should suppress tumorigenesis whereas mitochondria of tumor cells should enhance tumorigenesis regardless of whether the nucleus present is from a normal cell or from a tumor cell. It would therefore be important to consider the findings from the nuclear-cytoplasm transfer studies, as I previously described (Seyfried, 2012d).
Normal Cytoplasm Suppresses Tumorigenesis in Cell Cybrids
Suppression of tumorigenicity was observed when the cytoplasm of enucleated normal cells was fused with nucleated tumor cells to form cybrids (Seyfried, 2012d). Cybrids contain a single nucleus with a mixture of cytoplasm from two different cells. To examine the influence of cytoplasm on the expression of tumorigenicity in cybrids, Koura fused intact B16 mouse melanoma cancer cells with cytoplasts (absent nucleus) from non-tumorigenic rat myoblasts (Koura et al., 1982). The reconstituted cybrids exhibited a unique morphology and cellular arrangements different from that of the parental cells. Tumorigenicity was reduced in all the reconstituted clones and cybrids soon after their isolation, but tumorigenicity re-appeared in some clones after extended cultivation of the cells in vitro (Koura et al., 1982; Seyfried, 2012d). The effects of the unnatural cell culture environment on mitochondrial respiration could account in part for the reversion to tumorigenicity seen in some clones (Warburg, 1956a; Kiebish et al., 2009). The Koura et al findings showed that normal cytoplasm, containing mitochondria from non-tumorigenic cells, could suppress the malignant phenotype of tumor cells (Seyfried, 2012d). Although these observations were not linked to Warburg’s theory, the findings question the dominant role of the nucleus in the origin of tumorigenesis.
In a more comprehensive series of experiments, Israel, and Schaeffer demonstrated that suppression of malignancy could reach 100% in cybrids containing tumorigenic nuclei and normal cytoplasm (Israel and Schaeffer, 1987). On the other hand, tumors formed in 97% of mice implanted with cybrid cells derived by fusion of cytoplasts from malignant cells (nucleus absent) with karyoplasts from normal cells (nucleus present). The important feature of their study was that the non-transformed and the transformed cells were all derived from an original cloned progenitor cell with a common nuclear and cytoplasmic background (Israel and Schaeffer, 1988; Seyfried, 2012d). These findings showed that normal cell nuclei could not suppress tumorigenesis when placed in tumor cell cytoplasm. In other words, normal nuclear gene expression, which would presumably include tumor suppressor genes, was unable to suppress malignancy. An alternative view is that the cytoplasm of the tumor cell could reprogram the nucleus to become tumorigenic. These findings are consistent with the view of Darlington who showed that it was the cytoplasm, rather than the nucleus, that determined the tumorigenic state of the cells (Darlington, 1948). Israel and Schaeffer did not identify the molecular basis for the cytoplasmic control of tumorigenesis, but they did suggest that epigenetic changes in nuclear gene expression might be responsible for the phenomenon (Seyfried, 2012d).
It is obvious that the findings of Israel and Schaeffer are inconsistent with the somatic mutation theory, but their observations would support the concepts of Warburg’s theory. These investigators, however, did not connect their observations to Warburg’s theory. Instead they linked their observations to a potential epigenetic phenomenon (Israel and Schaeffer, 1988). It is important to recognize that mitochondria are a powerful extra nuclear epigenetic system that can control nuclear gene expression through the retrograde signaling system (Minocherhomji et al., 2012; Seyfried, 2012e). A personal account of the Israel and Schaeffer studies in light of the competing theories of cancer has appeared (Christofferson, 2014).
The findings of Israel and Schaeffer that normal cytoplasm could suppress tumorigenicity were also consistent with the observations of Shay and Werbin (Shay and Werbin, 1988; Shay et al., 1988; Seyfried, 2012d). Shay and Werbin identified several factors that could influence the outcome of cybrid experiments designed to uncover cytoplasmic suppressors of tumorigenicity. These influencing factors included, (1) the relative amounts of non-tumorigenic and tumorigenic cytoplasm in cybrids; (2) the time interval that cybrids are passaged in culture prior to testing their tumorigenicity; (3) whether or not mutagenesis with carcinogens were used to introduce genetic markers in the cells; and (4) the type of cell combinations that were used. It would not therefore be surprising that varied results could occur with the cybrid experiments if the confounding variables were not controlled. In general, however, the observations of Shay and Werbin were consistent with the conclusion of the Israel and Schaeffer experiments. Although Shay and Werbin mentioned a possible role for mitochondria in the suppressive effects of the cytoplasm on tumorigenesis, they also did not consider their results in light of Warburg’s metabolic theory (Seyfried, 2012d).
Howell and Sager, however, were aware of the relationship of Warburg’s theory and the findings from the various cybrid studies (Howell and Sager, 1978; Seyfried, 2012d). These investigators speculated that the results from the cybrid experiments could help distinguish whether it was the cytoplasm or the nucleus or that determined tumorigenicity. They showed that cytoplasm of non-tumorigenic normal cells suppressed the rate and extent of tumor formation in nude mice when fused with nucleated tumorigenic counterparts (Seyfried, 2012d). Howell and Sager stated; if tumor cell mitochondria are defective, as Warburg postulated, then suppression could result from the introduction of mitochondria from normal cells into cybrids (Howell and Sager, 1978). These findings like those of Koura, Israel and Schaeffer, and Shay and Werbin supported Warburg’s theory and are difficult to explain with the somatic mutation theory (Seyfried, 2012d).
To further evaluate the role of the cytoplasm and the nucleus in the control of malignancy, Jonasson and Harris conducted several interesting studies in human/mouse hybrids. These investigators evaluated in vivo tumor malignancy in a range of hybrid clones derived from fusions of a malignant mouse melanoma with diploid human fibroblasts and lymphocytes (Jonasson and Harris, 1977). They observed that the human diploid cells were as effective as the mouse diploid cells in suppressing the malignancy of the mouse melanoma cells, despite the preferential elimination of the human chromosomes in the hybrid clones. Malignancy was also suppressed in a hybrid clone where only a single human X chromosome was present. Jonasson and Harris showed that this clone continued to produce few tumors, even after they used back selection to remove this remaining X chromosome. These findings suggested that no human nuclear genetic material was responsible for suppression of malignancy. These findings would rule out a nuclear epigenetic explanation for suppression of tumorigenesis, but would not exclude an extra-nuclear (mitochondrial) epigenetic explanation.
Jonasson and Harris also constructed hybrids between the melanoma cells and human fibroblasts that were irradiated before cell fusion (Jonasson and Harris, 1977). They showed that the incidence of tumor take in nude mice was greater in crosses between the mouse melanoma cells and the irradiated human fibroblasts than in crosses between the melanoma cells and the un-irradiated human fibroblasts (Jonasson and Harris, 1977). These investigators concluded that the suppression of malignancy involved the participation of a radio-sensitive extra-chromosomal element. The findings from the Jonasson and Harris studies were interesting for several reasons (Seyfried, 2012d). First, their observations were consistent with those of several other cybrid studies suggesting that factors in normal cytoplasm could suppress tumorigenicity. Second, no human nuclear genetic material was responsible for the suppressive effect. Lastly, high-dose gamma radiation could destroy the cytoplasmic factor that was responsible for tumor suppression. This last observation was consistent with the findings of both Warburg and Darlington in showing that high-dose radiation destroys mitochondrial respiration and the cytoplasmic plasmagene, which has multiple characteristics of mitochondria (Darlington, 1948; Warburg, 1956a). Low dose radiation can cause nuclear mutations but not cancer, whereas high dose radiation damages both the nucleus and mitochondria and can cause cancer.
It is interesting that Jonasson and Harris excluded the mitochondria in preference to a centrosome origin for the suppression effect of cytoplasm on tumorigenesis (Jonasson and Harris, 1977). Their opinion was based largely on the findings of other investigators showing that no human mitochondrial DNA or proteins were detected in the human-mouse cybrids. More recent studies in transmissible cancers, however, show that tumor mitochondria can integrate with normal mitochondria in some tumors (Rebbeck et al., 2011). I suggested that this integration might reduce or partially correct the respiratory insufficiency in the tumor cell mitochondria thus suppressing tumorigenicity (Seyfried, 2012d). The work of King and Attardi also support this possibility in showing that exogenous mtDNA could enhance respiration in cells lacking functional mtDNA (King and Attardi, 1988, 1989). The more recent findings of Tan et al. also support the possibility in showing that mtDNA can be transferred horizontally from host cells to tumor cells in the microenvironment (Tan et al., 2015). Viewed collectively, these observations are in general agreement with Warburg’s original theory.
It is not possible, however, to exclude all influence of the nuclear genome in the suppression of tumorigenicity. Saxon and co-workers showed that the microcell transfer of Chromosome 11 suppressed tumorigenicity in HeLa cells (Saxon et al., 1986). They suggested that a tumor suppressor gene could be present on Chromosome 11. These findings also suggest an interaction between Chromosome 11 and mitochondria (Seyfried, 2012d). Is it possible that a gene on Chromosome 11 facilitates mitochondrial respiration thus suppressing tumorigenicity in the HeLa cells? (Seyfried, 2012d). It is also interesting that defects on Chromosome 11 have been associated with the Wilms kidney tumor and with childhood neuroblastoma. Further studies will be needed to determine if tumorigenic suppression involves interactions between mitochondrial respiration and genes on Chromosome 11.
Evidence from rho0 Cells Supporting a Mitochondrial Origin of Tumorigenesis
Singh and co-workers showed that the exogenous transfer of wild type mitochondria to cells with depleted mitochondria DNA (rho0 cells) could reverse the altered expression of the APE1 DNA repair protein and the tumorigenic phenotype, thus providing evidence for the role of mitochondria in the suppression of tumorigenicity (Singh et al., 2005). Mitochondrial respiration appears responsible for the efficiency of APE1-mediated DNA repair. The rho0 cells have impaired respiration due to the lack mtDNA that is essential for normal cellular respiration. It is my view that transfer of normal mtDNA to the rho0 cells will restore respiration, turn off the mitochondria/nuclear retrograde response, and prevent nuclear genomic instability (Seyfried, 2012d). These findings suggest that it is efficient mitochondrial respiration that prevents cancer. The more recent studies of Cruz-Bermudez support these observations (Cruz-Bermdez et al., 2015). I also described how mitochondrial enhancement therapies could prevent cancer (Seyfried, 2012b).
Wallace, and colleagues also provided support for the importance of respiration in the origin of prostate cancer (Petros et al., 2005). These investigators introduced the T8993G pathogenic mtDNA mutation into PC3 prostate cancer cells through cybrid transfer to determine whether the mutant pancreatic tumors expressed increased ROS levels and growth rate. The engineered PC3 prostate cancer cells were then tested for tumor growth in nude mice. The resulting mutant T8993G cybrids produced tumors that were seven-times larger than those produced from the wild-type cybrids. In contrast to the rapid growth of the mutant cybrids, the wild-type cybrids grew very slowly in the mice. Significantly more ROS were also produced in the tumors derived from T8993G mutant cybrids than tumors without this mutation. The carcinogenic and mutagenic action of ROS will damage respiration and produce nuclear genomic instability (Waris and Ahsan, 2006; Klaunig et al., 2010; Seoane et al., 2011). Additional experiments from the Wallace group and more recently from Cruz-Bermudez and co-workers showed that introduction of mtDNA mutations could reverse the anti-tumorigenic effect of normal mitochondria in cybrids (Petros et al., 2005; Cruz-Bermdez et al., 2015). These findings indicate that some mtDNA mutations can play an important role in the etiology of cancer and that cancer can be best defined as a type of mitochondrial metabolic disease. These findings are also more in line with Warburg’s theory than with the somatic mutation theory.
Normal Cytoplasm Suppresses Tumorigenesis In Vivo: The Lucke Frog Renal Tumor
Substantial information exists showing that the nuclei of tumor cells can be reprogrammed to form normal tissues when they are transplanted into normal cytoplasm, despite the continued presence of the tumor-associated genomic defects in the cells of the derived tissues (Seyfried, 2012d). McKinnell, Deggins, and Labat showed that cell nuclei from frog renal tumors could direct normal frog development following transplantation of the renal tumor cell nucleus into an enucleated normal egg cell (McKinnell et al., 1969). The experiments involved implantation of nuclei, isolated from Lucke frog renal cell tumors, into fertilized enucleated eggs from normal diploid frogs. Importantly, the cells of the renal tumor were triploid in containing three copies of all chromosomes. Triploid tadpoles developed normally from the triploid tumor cell nuclei, and revealed functional tissues of many types. This experimental strategy made it possible to distinguish development initiated by the transplanted nucleus from development influenced by an inadvertently retained maternal diploid nucleus (McKinnell et al., 1969). The investigators showed that ciliated epithelium propelled the tadpoles in the culture dishes. The tadpoles swam when stimulated. The tadpoles had functional receptors, nerve tissue, and striated muscle necessary for swimming. Cardiac muscle pumped blood cells through the gills. Suckers secreted abundant mucus. Clearly seen were a pronephric ridge, eye anlage, nasal pit, and open mouth, as was the differentiation of the head, body, and the tail. The tail fin regenerated after being clipped for chromosome study. Moreover, sections of embryos developed from transplanted triploid tumor nuclei revealed apparent normal development of the brain, spinal cord, optic cup with lens, auditory vesicle, somites, pronephric tubules, pharynx, midgut, and notochord. No evidence of abnormal cell growth was seen in any of the organs or tissues examined (Seyfried, 2012d). These findings showed that nuclei derived from tumor cells could direct normal developmental and did not induce dysregulated cell growth, the signature phenotype of tumorigenesis. It is interesting that the tadpoles containing tumor nuclei could not complete development to normal adult frogs. It remains unclear if the tumor-associated nuclear defects were responsible for preventing late stage development of the frogs.
The findings from the Lucke frog experiments are consistent with the mitochondrial metabolic theory, but are difficult to reconcile with the somatic mutation theory of cancer (Seyfried, 2012d). The enucleated egg would contain the mitochondria from the normal egg cytoplasm. These mitochondria would direct normal energy homeostasis during development. It is my view that normal energy homeostasis suppresses tumorigenesis despite the presence of the tumor nucleus and somatic mutations (Seyfried, 2012d). Later studies suggested that loss of the Lucke tumor herpes virus was linked to the loss of tumorigenicity (Carlson et al., 1994). This virus was considered the etiological agent responsible for the origin of the renal tumors. It is now known, however, that the herpes virus can alter mitochondrial function to induce tumorigenesis (D’agostino et al., 2005; Seyfried, 2012c). Indeed, Ackerman and Kurtz showed that herpes viruses have an intimate attachment to mitochondria that causes dysfunctional respiration (Ackermann and Kurtz, 1952). Hence, the replacement of virus-damaged mitochondria with normal mitochondria from the host could the suppress tumorigenesis despite the presence of the renal tumor nucleus (Seyfried, 2012d). The findings from the frog renal tumor are similar to those described above from the cell cybrid experiments, and cast doubt on the somatic mutation theory as an explanation for this type of cancer.
Normal Cytoplasm can Suppresses Tumorigenic Phenotypes in Mice
Findings similar to those obtained with the Lucke frog renal tumor were also obtained following nuclear transfer in mouse tumors. Morgan and colleagues showed that nuclei from a mouse brain tumor, arising from cerebellar granule cells (medulloblastoma), could direct normal development when the tumor nuclei were transplanted into enucleated somatic cells (Li et al., 2003). Figure 1 from their study shows that normal embryonic tissues and germ cell layers can be formed from cells containing the tumor nuclei. These investigators showed that the transfer of the tumor cell nucleus into normal cytoplasm suppressed the tumorigenic phenotype despite the continued presence of the mutant nuclear gene (Patched) that was thought responsible for the original tumorigenic phenotype (Li et al., 2003; Seyfried, 2012d). The transplanted medulloblastoma nuclei produced post-implantation embryos that underwent normal tissue differentiation and early stage organogenesis. Importantly, no malignancies or abnormal cell growth were seen in any of the recipient mice. Normal proliferation control was observed in cultured blastocysts indicating that nuclear somatic mutations alone were not likely responsible for the original tumorigenic phenotype (Li et al., 2003).