Viral and Nonviral Oncogene Theories of Cancer

Rather than writing a review with a limited scope, we attempted to survey the span of medical and molecular biology research in the field of oncology during the last six decades: from the early breakthroughs in viral oncology and the novel notions of virus and viral cancer (onc) genes; to the role of DNA cancer viruses and C retroviruses in experimental carcinogenesis; through the discovery of cellular oncogenes and the biology of growth factors; to the expansion of the concept of oncogene and its types; and, finally, to the realization that cancer is a multiplicity of different disorders that appear to arise through nonviral auto-oncogenic processes involving adaptive changes and epigenetic responses to cancer-promoting pressures in the external and internal environment of the organism. Throughout, we have searched for the integration of an oncogenic vector with different degrees of transformation, seeking the commonality of proliferative disorders, somatic cancer and leukemia. We re-examine what separates transformation, benign and malignant, from differentiation, and how their reversible switch deploys graded responses related to states of hypersensitivity to, or independence from, key physiological growth factors. “In a virus-induced cancer, a normal cell is altered by an oncogenic virus. The malignant cell continues to grow and divide, and considered by itself is healthy. A cell, however, is not an independent unit but a dependent part of an organism. An organism controls the growth and multiplication of the normal cell but not of the malignant one, which behaves as an independent unit. Its multiplication causes the death of the organism. The oncogenic virus, although it only modifies a cell, kills the organism and is therefore pathogenic.” A. Lwoff, 1958 [1] 1. Emergence of virology and research into the viral etiology of cancer: viruses vs living systems The term “cancer” means ‘crab’ in Latin, and comes from the description of a tumor as a central mass that invades surrounding tissue with crab-like claws. Cancer has been known as a form of disease for at least 4,000 years, since the time of known papyri which describe cervical cancer and surgical trepanation for the removal of brain tumors. But the biology of cancer, ie oncology, is only barely a century old. It began in the form of a viral theory of cancer. Paradoxically, by the time that its development spurned by the study of bacteriophage that infect bacteria had led to a full molecular and biochemical elucidation of DNA and RNA viruses, and their involvement in infection and malignancy, most cancers could no longer be regarded as being viral in origin. The notion that infection with viruses (from virus, Latin for ‘poison’ or ‘venom’) may cause cancer dates back to Peyton Rous discovery of the viral etiology of a malignant chicken tumor in 1911 [2], when he reproduced the sarcoma in chicken by injection of a cell-free filtrate [3]. R. Shoppe repeated the demonstration with a cell-free filtrate from a rabbit fibroma in 1932 [4]. These experiments suggested that at least the first three postulates of Koch could be confirmed: the agent was isolated from the pathological lesion, could be transmitted to a like host and, once transmitted, caused the same disease. Ludwik Gross went a step further in the 1950’s, and showed that virus-induced fibrosarcomas in the newborn mice could be prevented by immunization of the mother, fulfilling Koch’s 4th postulate. He also discovered that murine leukemia could be transmitted with a filtrate [5]. As detailed by Gross in a review of the findings from research in the 1950’s, leukemic viruses could be isolated from spontaneous leukemias in high-leukemic inbred mice strains, from virus-induced leukemias in rodents, from radiation-induced leukemias in low-leukemic inbred mice strains and from transplanted mouse sarcomas and carcinomas [6]. Social resistance in the form of peer-review rejection, media derision and unavailability of funds to the modest beginnings of virology and oncology was nearly as intense as that met, nearly Journal of Biophysics, Hematology and Oncology, Vol. 1, 4:1-79 April 2010 2 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 in parallel, by Reich’s medical and ‘bion’ research in what he called “the cancer biopathy” [7]. Lwoff, Luria, Gross, and many others, suffered ignominious attacks. Gross, in particular, was the subject of a concerted ad hominem campaign [8], which eventually ceased when a new generation of researchers trained with molecular techniques began corroborating his findings. A substantial component of the irrationality of the epoch was a consequence of the then ongoing revolution in the biological understanding of what is a virus and how it is distinguishable from an organism, be it a micro-organism. The definition of virus which has since remained was that magistrally legated by André Lwoff in the 1950’s [9]. It is based on the notion that viruses are neither organisms, nor cells or mere molecules. Lwoff ’s accent lies in the definition of the living as an “organism”, as “an independent unit of integrated and interdependent structures and functions” [9]. An independent existence requires: (1) an autonomous energy metabolism (“the presence of a Lipmann system”) that permits growth; (2) an ability to autonomously replicate in toto, ie replicate not only their genomic DNA or genetic material but reproduce the entirety of their organellar, cellular, tissue and organic structures; and (3) the ability to multiply, whether by mitotic (self-)division and proliferation or by a combination of the latter with a form of propagation. Now note that Lwoff ’s requirements for an “independent existence” do not define the major trait of Lwoff ’s concept of an “organism”. Instead, this trait emphasizes the dependence of the parts on the whole and the interdependence of structures and functions in the organism. This conceptualization is nearly parallel to Reich’s definition of the organism as a system having a unity of function and structure between very different organs. Both Lwoff and Reich explicitly reject the notion that the organism is the sum of its parts. But it must be said that the concept of ‘dependence’ is too wide and therefore somewhat vague as is the concept of independence or autonomy. One may grow cells from a metazoon outside of its system, in in vitro conditions where that dependence is abrogated even if it is still residually manifested by the specificity of factor requirements. One may grow organelles or endosymbionts (eg dinoflagellates, etc) independently from the host cells, and so on. Moreover, no living system is ultimately independent of other living systems (as the concept of ecology well illustrates) or from the physico-chemical conditions of its environment (fluid and energy media). In this respect, Reich provided a better formulation of the concept of biological unity when he described the major trait of a living system as being provided by “the function of the whole in each individual part” [10]. It is not so much “the dependence of the part on the whole”, as it is the fact that the whole functionally resides in each of the parts, in the form of a function of the part which is, in that part, already the function of the whole. The whole is a part apart from the parts, but it is composed of the parts of the whole in each of the parts. The unity of structure is merely a consequence of the unity of function that results from the consistent function of the whole in the functioning of each of the parts. But we should rather prefer to call Correa & Correa Oncogene Theory of Cancer 3 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 ‘system’ what Lwoff and Reich call an ‘organism’ a ‘discrete living system’, it being understood that such a system is always open through energy and fluid (molecular) continuity with an outside. Similarly, the entirety of the requirements for autonomy in energy metabolism, in morphogenesis as well as in replication and division is a corollary of the self-ordering property of a living system, which is the expression of the capacity of the system to increase its internal energy by energy capture, conversion and accumulation [11]. Lwoff argued that metazoa, protozoa and prokaryotes all form organismic entities (living systems) a cell in a metazoic system or a colonial lifeform being as much a dependent part as is an organelle in a protozoon [9]. An organelle Lwoff suggested does not have an independent existence, even if it once did (such as mitochondria). Nor do viruses, he stated. In fact, in the wake of Beijerinck and Bawden, Lwoff claimed that “it is clear that viruses have more in common with cellular organelles than with micro-organisms” [9]. As proof, he mentioned that both organelles and viruses present nucleic acid or genetic continuity (this observation only applies to DNA-carrying organelles) and depend upon the cellular metabolism of the host for their replication and multiplication. This is all the more provocative as Lwoff placed one more accent in the definition of a virus that he was providing: the ability to be infectious which, in his strict view, had to distinguish a virus from “all the normal cell structures which can penetrate into another cell” [9]. The distinguishing trait was “the introduction into an organism of a foreign entity able to multiply, to produce a disease, and to reproduce infectious entities” [9], with the term ‘foreign’ being the connection to an outside that ruled out the virus as being an organelle. Thus, we immediately remark the weakness of this distinction of a virus from any “dependent part” of a living system, and the petition of principle it incurs. If viruses are parallel to cellular organelles and do not exist or multiply independently from host cells, then one should conclude that their original creation must, in each instance, have been a biological production a cellular emission of a “normal cell structure able to penetrate another cell”. Before any original virus ever had the chance of being multiplied, it was a truly ‘endogenous’ virus. The concept of ‘infection’ as a linkage to an outside is a secondary characteristic, no matter how important, since it is effectively inseparable from the notion of ‘production of a viral particle from an inside’, by a living system. This weakness, however, remained the ‘unspoken’ of medical virology which gives one the measure of contrast with Reich’s “orgonomic” approach to the etiology of cancer, which claimed that cancer was not infectious, and instead was an acquired, endogenous disorder, where even the production of ‘Tbacilli’ (or mycoplasma [7]) was an endogenous (and ‘heterogenic’) production. Curiously, Lwoff actually addressed these problems in that magistral 1957 lecture. He proposed, in fact, what was a radically new concept of viruses, that they were truly endogenous in origin: “viruses (...) originated from some pathological constituents of their host cell” [9]. Using the bacteriophage as example, “the Journal of Biophysics, Hematology and Oncology, Vol. 1, 4:1-79 April 2010 4 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 endogenous theory” (Lwoff ) held precisely as we just suggested above that viruses originated in bacterial genes as dependent parts of the bacterial system: “The prophage is obviously not independent. It behaves as a dependent part of an organism” [9]. But in thus employing the term ‘endogenous’ to designate the origin of bacteriophage and any other viruses by extension as effectively heterogenic, Lwoff explicitly refused assimilation of ‘endogenesis’ to ‘heterogenesis’: “The inquisitors of faith have tried, and are still trying, to ridicule the endogenous theory by brandishing the threadbare scarecrow of heterogenesis” [9]. Yet, the only avenue left for Lwoff to take that could dissimilate the two concepts was not the one he chose for ‘endogenesis’ of viruses could only be different from ‘heterogenesis’ to the extent that the latter could only be said of living systems and viruses clearly do not constitute living systems. However, when it comes to the biological ontogeny of viruses qua viruses, to say that they are endogenous is still to say that they are heterogenic, that they (viruses) arise from different and irreducible elements, from cells, from their host cells. Moreover, for all effects and purposes, this made any concept of infection secondary to the processes that, endogenously to a cell, heterogenically create a virus. And while it is apparent that viruses have the ability to infect in a manner analogous to bacteria and protozoa, this ability is a passive one, since as Lwoff was well aware host cellular susceptibility (permissiveness) and immunity are the ultimate determinant factors of whether an infection succeeds or not. Irrespective of the shortcomings in Lwoff ’s definition of a virus, it is evident today that viruses constitute a unique category of biological entities that cannot be regarded as forming biological systems. They should, rather, be thought as dependent parts of biological systems. Viruses are not considered to be living systems, to constitute organisms or even cells, since they cannot autonomously replicate and multiply, and they lack metabolism. Viruses neither grow nor divide. Thus the existence of a virus does not require a source of energy, only its production and reproduction do. But for these ‘purposes’, every virus depends upon the protein and genetic machinery of a host cell that it must infect. Further, viruses can only be multiplied in the form of their nucleic acid. There are no known living systems whose genome is composed of RNA, but all living systems contain both DNA and RNA. Conversely, the genome of viruses may be composed of either DNA or RNA, and viruses only contain one type of nucleic acid from which the virus may be multiplied. Viruses are therefore strictly parasitic molecular assemblages of proteins and genetic code fragments. They are not living in the cellular or systematic senses, but alive in the molecular sense that all viruses are biological productions, or nucleoproteinic biomolecular constructs. The key component of each virus is the genome, their genetic sequence, which may come in single or double stranded varieties of either DNA or RNA. By the late 1950’s, viruses and virus-like particles (see Fig. 1) were being isolated from a variety of somatic tumors and leukemias in animals, and from some human leukemias (even though lackCorrea & Correa Oncogene Theory of Cancer 5 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 Journal of Biophysics, Hematology and Oncology, Vol. 1, 4:1-79 April 2010 6 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 Fig. 1 Representation of electron micrographs of typical DNA and RNA tumor viruses, and virus-like mycoplasma particles that are bacterial, budding from a host cell. The different particles can be readily distinguished under the transmitted electron microscope, but are ordinarily indistinguishable under light microscopy. In live mounts, however, mycoplasma have characteristic forms of motion (gliding, jumping). In fixed mounts, mycoplasma also stain Gram negative. Note that mycoplasma may be as small as 100 nm in diameter or slightly less. Fig. 1A Herpes simplex 1 about to detach from a human host cell. Fig. 1BFriend erythroleukemia virus budding from murine host cell surface (osmium tetroxide stain). Fig. 1C Mycoplasma hominis, a virus-like particle, in the process of budding from a human host cell (after Barile 1967, Figure 2). ing an etiological relationship, as it turned out), and it was now possible to determine whether viral genomes were composed of DNA or RNA. The infectivity of purified RNA from Tobacco Mosaic Virus was demonstrated in 1956 [12]. In 1952, Renato Dulbecco devised the first in vitro assay for tumor viruses, employing tissueculture of fibroblasts grown in monolayers [13]. Cells that became malignantly transformed by virus infection lost contact-inhibition, ceased producing a monolayer and adhering to the bottom of the culture vessel, to form distinct foci or plaques that, if harvested and inoculated into experimental animals, consistently produced the same fibrosarcomas from which the viruses had been isolated. It was now possible to produce and quantitate viral particles in vitro. We should note that by the 1950’s, it had also become well understood that chemical mutagens and ionizing radiation (from cosmic, solar, geological and man-made sources, including far ultraviolet light) could induce nonhereditary cancer, but the suspicion then arose that they might do so indirectly, via the production of viruses. During the 1920’s and 1930’s, most oncology research had gone into establishing that physical (X-rays) and chemical carcinogens acted as mutagens. The first two classes of identified chemical carcinogens were the nitrogen mustards and the nitrosoamines. By the 1950’s, it became apparent that chemical mutagens (aromatic amines, polycyclic hydrocarbons, steroid hormones, asbestos) induced mostly epithelioid tumors, called carcinomas, whereas ionizing radiation seemed to induce melanomas, but most frequently leukemias. Most generally, when viruses infect permissive cells two main responses occur. The acute infectious response is associated with massive viral replication and the lysis of the host cell, and is thus referred to as the lytic or cytopathic response. But persistent infection may instead take place, with the virus either (1) replicating independently of host replication (whether inserted into the host genome or leading an extra-chromosomal existence, eg an episome), and most often so slowly that it does not lyse the host cell; or, if unable to replicate, (2) with the virus being inserted into the host genome at or near critical gene sites, and leading a latent existence there (in DNA viruses and bacteriophages this is referred to as lysogeny proper; note that upon induction, lysogenic phages leave the chromosomal integration site and begin a lytic reproductive cycle). In the first case, one speaks of chronic infection (for example, infection of the trigeminal ganglion cells with the most common type of human herpes virus, HSV-I), and in the second, of a latent infection with no virus production. 2. DNA tumor viruses It was early research with bacteriophages in the 1940’s and early 1950’s on their ability to ‘transform’ bacteria that led to interest in eukaryotic viruses and investigation of the possibility that viral infection might cause malignant transformation of the eukaryotic cells of metazoa. Correa & Correa Oncogene Theory of Cancer 7 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 Early work with the murine polyoma and the African green monkey SV40 DNA viruses showed that, though not oncogenic in their hosts of origin, they could cause a variety of tumors when injected into newborn animals of susceptible hosts [14]. Infection of living animals or tissue culture cells exhibited alternative courses a lytic and a transforming responses. Whereas the lytic response produced large numbers of viral particles in permissive cells, in a small number of nonpermissive cells that were transformation-susceptible the virus integrated in the host genome and yet no trace of it could be found (cell lines may be permissive to one DNA virus and nonpermissive to another; eg NIH3T3 cells are permissive for polyoma and nonpermissive for SV40). Infection with a single viral particle sufficed for transformation. Transformed cells presented no production of viral particles at all, these viruses behaving in these cells like the lysogenic phage of bacteria. However, it was remarkable that while integration of the viruses was observed in their hosts of origin, no transformation or tumor formation occurred. By itself, integration was not sufficient to initiate transformation [15], the latter requiring expression of virus-specific transforming antigens. Human papilloma viruses are responsible for a variety of warts benign epidermal tumors but, according to zur Hausen’s 1977 suggestion, these warts may convert into squamous cell carcinomas if exposed to X-rays or if they persist longer than 5 years [16]. Since then, DNA of cervical, uterine, vulvar and penile carcinomas was shown to contain human papilloma virus (HPV) homologous sequences in 61% of German patients but only in 35% of Brazilian and Kenyan patients [17], which suggests the HPV presence could just as well be a coincidence. The viral homologous sequences are required for carcinogenesis and tumorigenicity [18], yet no intact HPV particles have been isolated from these carcinomas. Based on epidemiological considerations alone, these carcinomas appear to have a viral etiology of an infectious nature [18]. However, it is far from clear whether infection with HPV causes these various carcinomas, whether it requires super-infection with other viruses (such as HSV-2), or follows in the footsteps, or still whether it interacts with other systemic risks of neoplasia. As in infection with the Epstein-Barr virus (see below), these tumors cannot be the direct result of primary infection without some other factor(s) intervening, likely a multiplicity of factors. This is evident from epidemiological studies, as they show that only a very small fraction of those infected with HPV develop carcinomas. Transforming abilities have been found for many DNA viruses, whether they integrated in host chromosomes (like polyoma virus, adenovirus, papilloma) or not (like fibrosarcoma-inducing pox virus). However, unlike oncogenic retroviruses (see below), DNA tumor viruses do not transduce cellular oncogenes; they transform the host cells mostly by the production of virus-specific proteins (eg the nuclear “T”-antigen characteristic of early infection with polyoma), or by targeted mutagenesis (insertions, deletions, inversions) that activates DNA transcription. The main trait of most DNA Journal of Biophysics, Hematology and Oncology, Vol. 1, 4:1-79 April 2010 8 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 tumor viruses is that they appear to integrate in the host genome only randomly. There are two cautions to this: (1) that these studies have all been carried out with either laboratory animals or cell lines that were experimentally infected (and typically with high multiplicities of infection); and (2) that the translocations induced by the Epstein-Barr virus (EBV) appear to be selective. Though the existence of animal models demonstrating causation of sarcomatous tumors by DNA viruses emboldened the belief that cancer had a viral etiology, the evidence for DNA viruses causing human cancers is, to this day, practically nil. In defiance of Koch’s postulates, the epidemiology of cancer does not match the spread of any infectious DNA viruses, with the notable exception of the EBV. Even when DNA viruses are able to transform cells in tissue culture, like the herpes simplex viruses HSV-1 and HSV-2 that have high homology with EBV, none of the resulting transformants contain HSV [19], and there are no human tumors that can be causatively linked to them. The herpesvirus family (Herpesviridae) was the main DNA virus candidate to play a central role in viral oncogenesis in humans. Some of the herpesviruses are present in virtually all human populations, and they are expressed in many human ailments and diseases cold sores, shingles, otitis (and likely in the etiology of Ménière’s Syndrome), venereal disease, infectious mononucleosis (once thought to be an infectious form of leukemia), birth defects, in the induction of Burkitt’s B-cell lymphoma and, possibly, of nasopharyngeal carcinoma. They also account for many animal diseases. Herpesvirus B simiae, which only causes cold sores in monkeys, is the cause of a fulminating fatal encephalomyelitis in humans. Herpesviruses are replicated inside the host cell, and their particles bud from the cell surface in a manner analogous to retroviruses (see Fig1A and compare with Fig 1B). As a consequence, they can be mistaken for the latter under high-power darkfield light microscopy and very low power electron microscopy, as well as for cell-adsorbed mycoplasma, all the more easily as these particles (viral and virus-like) have the same size range. The herpes virion (or nucleocapsid) is composed of a core containing an histone-packed, double-strand of two 105 kD linear DNA molecules, and a protein capsid. Replication of viral DNA takes place inside the nucleus of infected cells, and it typically (or most frequently) involves episomal circularization of the linear genome (analogous to phage DNA processing). The DNA polymerase produces multiple concatemeric DNA molecules to be packed into an equal number of virions. Inside the host cell cytoplasm, only nucleocapsid forms exist. The complete viral particle is present only outside of cells, whether in the process of being ‘emitted’ (released) from infected cells, in free state, or adsorbing to the surface of cells that it will subsequently infect. Outside the cell, the virion is frequently surrounded by an amorphous gel the tegument that it took from the host cytoplasm, and by a lipid bilayer envelope that it took from the host cell membrane during budding (see Fig 1A). Whereas the complete virus the nucleocapsid typically Correa & Correa Oncogene Theory of Cancer 9 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 measures 180 nm in diameter (as is the case with the EBV), the complete virion typically measures 0.2 to 0.3 μm, so it is a ‘large virus’. Dennis Burkitt first observed that, given association with the spread of malaria, the geographical distribution of a B-cell lymphoma affecting East African children also appeared to be related to climactic factors [21]. In 1964, M. Epstein identified a herpesvirus in tissue cultures of lymphoblasts from a Burkitt’s lymphoma cell line [22]. Only some of the lymphoblasts produced virus (which was cytopathic), and virus-producing cells could be increased to 20% with L-arginine depletion [23]. The acute form of infection with EBV is the now prevalent infectious mononucleosis (IM), a glandular fever which was once disseminated mostly among lower socioeconomic strata [24]. IM is a polyclonal lymphoid leukocytosis easily diagnosed by the vacuolated (“moth-eaten” appearance) cytoplasm of a significant portion (up to 35%) of the lymphocytes in peripheral blood (see Fig. 2). IM is an insidious disease that persists lifelong in peripheral blood B-lymphocytes, and may have episodic manifestations. The absence of replicating EBV inside most Burkitt cells suggested the virus existed in some proviral form, and this was subsequently confirmed when it was found that all Burkitt cells carried multiple copies of the viral genome [25], most in the form of unintegrated circular DNA and a few, in some cases, integrated in the linear virion form [26]. Journal of Biophysics, Hematology and Oncology, Vol. 1, 4:1-79 April 2010 10 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 Fig. 2 Starry-sky pattern of vacuolated CD19+ cells is characteristic of pre-pre B-cells (“large lymphocytes”) in the bone marrow of patients with Burkitt’s lymphoma, leukemia and other B-cell ALL. These abnormal lymphocytes are also present in peripheral blood (on the left, 1150x mag.), where they are easily mistaken for monocytes, and in cell lines derived from these leukemic patients (on the center and right, 1750x mag.). Luminera Infinity-1 camera, Wright’s stain. Originally it was thought that EBV infection and induction of transformation was restricted to B-lymphocytes, but its role in inducing nasopharyngeal carcinoma [27] in epithelial cells has now been demonstrated as involving linear virion integration in the genome of these cells [28-29]. EBV infection is thought to be predominantly carried out via the nasopharyngeal route [30]. The most distinctive in vitro trait of EBV infection of B-cells is the establishment of continuous cell cultures, ie immortalization. The EBV encodes a family of six nuclear antigens (EBNAs), one of which, EBNA-2, is the transcriptional activator [31] responsible for the induction of the B-cell activation antigen, CD23 [32], and for the immortalization of cells infected in vitro by EBV [33]. Thus, current thought is still that IM predisposes an infected individual to later on express a B-cell type of malignancy. Detection of anti-EBV antibodies indicates previous IM of varying severity, most primary infections being so mild that they went by unnoticed [34]. B-cells from patients with Burkitt’s lymphoma present two main translocations that they share in common with other B-cell lymphomas (and thus are not specific to EBV-induced lymphomas): reciprocal 8 to 14 chromosome translocation [35] and a translocation of chromosome 8 to either chromosome 2 (t(2;8)) or 22 (t(8;22)) [36] (see below and Table 2). Both translocations involve activation of the immunoglobulin heavy chain gene IgH and the c-myc oncogene [37-38]. 3. Retroviruses and their cellular origin: viral and cellular oncogenes All RNA viruses capable of inducing tumors in animals happen to be retroviruses. The name ‘retrovirus’ comes from the enzyme REverse TRAnscriptase (RT), to give “retra”. Retroviruses carry the pol gene that encodes for this enzyme, and they have the particularity of being the only RNA viruses with a diploid genome. In infection, retroviral particles in the range of 0.25 to 0.35 μm are adsorbed to host cell surfaces where they bind stereospecific glycoproteins of both cellular and viral origin (thus, while encounters with host cells may be treated as a chance occurrence, infectivity is not random but relatively specific), and subsequently penetrate through the plasma membrane (mimicking endocytosis). All retroviruses replicate by insertion of a double-stranded DNA copy (a provirus) of their RNA genomes into the genome of a host cell. A single-stranded DNA sequence (rDNA) is first produced from one of the RNA strands by the RT enzyme, and then a complementary DNA (cDNA) strand is polymerized. A major step in the understanding of the function and origins of retroviruses was the independent discovery of reverse transcriptase in 1970 by Harold Temin’s group [39] and David Baltimore [40], which confirmed Temin’s 1962 hypothesis that RNA oncogenic viruses reproduced in the form of a DNA provirus copy that integrated in the host genome. Identified in oncogenic retroviruses, RT synthesized rDNA, ie complementary DNA from an RNA template breaking the first rung of the so-called DNA dogma: genetic information also flowed in reverse from Correa & Correa Oncogene Theory of Cancer 11 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 RNA to DNA. Temin went as far as suggesting that retroviruses functioned as intercellular messengers with a role in differentiation and “could provide part of a mechanism for inheritance of some acquired characters”. Integration into the host genome is part of the lifecycle of retroviruses, though proviruses may exist in single, circular episomal form, typically prior to integration. Dynamic states of equilibrium between integrated and episomal proviruses may be a strategy of the virus to retain a potential for multivariant adaptive change [41]. Similarly, it may also be a strategy of the host cell itself. Once integrated, replication-competent retroviruses can generate either ordinary m-RNA transcripts or genomic RNA transcripts, each 6 to 10 kbp long. The latter are then joined to form double stranded RNA and packed within a virion (a protein capsule), with virion proteins having been translated from the viral m-RNA transcripts. No other known viruses are packed in double-stranded form inside virions (see Fig.s 1B & 3). Replication-competent retroviruses are also the only known viruses that can transform host cells while simultaneously replicating inside of them, and do so independently of synchronism with cell replication (note that they would thus differ in this from the classic scheme of organelle replication). The replicated retroviral virion buds from the host cell surface by taking with it a bit of its cortical or gelated cytoplasm (see Fig. 1B), along with a portion of the plasma membrane (note that this parasitic graft is a residual marker of the truly heterogenic ontogeny of retroviruses). At high magnification with special techniques of light microscopy, budding of retroviral particles from a host cell is indistinguishable from the budding replication of mycoplasma from filaments of a parent mycoplasma that adsorbed to the surface of host cells (see Fig. 1C). Both would appear as if heterogenically created from the host, yet neither one likely is the retrovirus being a replica of an infecting particle absorbed by the host cell sometime in the past, and the mycoplasma dividing by budding from another mycoplasma or its common stalk frequently embedded in the plasma-membrane of a previously infected host cell. Today there is, of course, a greater acceptance of the notion that at least each type of retrovirus must have once arisen heterogenically from cells (see below), from cellular genes, but this notion only applies to singular events of speciation (whether successful or not), and not to the ordinary host-dependent but non-synchronous replication and packaging of retroviruses in infected cells, which is the routine case when one observes the budding of a retroviral particle from a host cell surface with light or electron microscopy. In a manner of speaking, the speciation of every viral type could be said to be ‘endogenous’, and precisely to the extent that viral ontogeny must be heterogenic. ‘Endogenous’ would here be opposed to what is ‘infectious’ or gained by infection, but would be so by subtending the notion of heterogenic ontogeny [42]. But modern molecular biology of retroviruses conceptualizes ‘endogenous Journal of Biophysics, Hematology and Oncology, Vol. 1, 4:1-79 April 2010 12 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 viruses’ rather differently even though also in opposition to the concept of ‘viruses acquired by infection’. Molecular genetics defines as ‘endogenous’ all viruses that are transmitted genetically to offspring and maintained as normal Mendelian genes. Vertical transmission (as opposed to horizontal transmission by infection) depends upon integration of viral DNA into the genome of host germ cells. Some endogenous viruses are silent (latent, ie under repressor control), others partially expressed and still others fully expressed and replicated. In evolutionary terms, the consensus is that retroviruses are recent “acquisitions”. But this does not explain what their source is, whether they first infected animal cells as if coming out of nowhere, or were first created and emitted by animal cells themselves. Gross held that integration and vertical transmission were second to horizontal acquisition [43]. This is likely the case in laboratory investigations that use high multiplicities of infection. However, after the discovery that information exchange flows from DNA to RNA but also in reverse, back to DNA, Temin’s “protovirus hypothesis” [44] explicitly considered that mutations and unusual recombinational events might create tumor viruses de novo, ie heterogenically [45]. This view was practically a neoLamarckian perspective on the creation and evolution of retroviruses that challenged the more traditional viral oncogene view of retroviruses which pins these down to either vertical transmission or horizontal infection. By admitting to the de novo creation of viruses from reshuffled genomic sequences carrying the requisite genetic elements, or from rescued viral footprints for that matter, the protovirus hypothesis questioned whether retroviruses and their involvement in specific oncogenic processes Correa & Correa Oncogene Theory of Cancer 13 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 really functioned in the traditional sense of Koch’s postulates as agents or causes of disease, and invited speculation as to whether viruses were mere symptoms of the disease messages emitted by cells subject to degenerative or pathogenic processes, messages that targeted and altered hormonal (growth factor) networks that controlled and modulated cell metabolism, growth and differentiation. There were sets of cellular genes that had a capacity to escape the genome by utilizing viral packaging, and tumor formation might simply reflect errors in the back and forth flux of genetic information that affected the ‘escapist genes’ and might prompt their escape. The central notion of the protoviral theory of oncogenesis is that retroviruses transduced cancer-causing genes (oncogenes) that were originally derived from host cell DNA (from “protooncogenes”), but which were no longer tissue dependent for their regulation and expression. This evolutionary relationship was evident early on when high homologies were found between the viral oncogenes and the host genomic sequences of the proto-oncogenes, indicating that these oncogene sequences were highly conserved between cells and retroviruses. Whereas proto-oncogenes encoded proteins or enzymes typically phosphorylating kinases (tyrosine kinases, but also threonine and serine kinases) that regulated normal cell growth and development, and whereas their expression was both tissue-specific and developmental stage-specific, viral oncogenes encoded mutant variations of these gene-products. In particular in retroviruses, these oncogenes had lost intron and regulatory sequences present in proto-oncogenes, this fact pointing to their heterogenic origin from cellular mRNA rather than cellular DNA. Expression of retroviral oncogenes is solely regulated by long, repeating nucleic acid sequences present at the start and end of their genome, and called long-tandem repeats (LTRs). A typical retroviral genome structure is shown in the upper part of Fig. 4, for the Moloney Murine Leukemia Virus (Mo-MuLV). The facts that there are hotspots for the genomic integration of retroviruses, and that these are preferentially located near or in the sequences of cellular proto-oncogenes also pointed to the cellular origin of retroviral oncogenes. The marked cellular tropism of a retrovirus is a function of the oncogene it transduces. So it might not be surprising if the targets of retroviruses turn out to be growth factor receptors that control cellular growth, differentiation and metabolism. In normal cells, there are no integrated defective retroviruses with their viral oncogenes (voncs), only active cellular oncogenes (c-oncs) under normal regulation and the odd normal retrovirus without a v-onc. And, indeed, all viruses found in normal or wildtype germ cells do not seem to carry a v-onc. This, as Arthur Axelrad taught these authors, led MacFarlane Burnet once to state that since the syringe of the investigator was the only vector of viral cancer, viruses could not be taken seriously as causative agents in human cancer and the successful induction and transmission of cancer by viruses in animals were mere laboratory curiosities without clinical value. Journal of Biophysics, Hematology and Oncology, Vol. 1, 4:1-79 April 2010 14 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 In evolutionary terms, the biopoiesis of retroviruses is indistinguishable from their heterogenesis from host cells: retroviruses and their oncogenes evolutionarily arose from cells, and not the other way around that cells acquired oncogenes because they were infected with retroviruses. Since, ontogenetically speaking, infection must have come second with respect to true endogenous virus emission or production, viruses should be viewed as cellular signals. The fact that, whether by the adaptive pressure of the syringe or by nature, they can be induced to pick up cellular oncogenes and transduce them in a mutagenized form, or affect expression of critical cellular genes by nonrandom hotspot insertion near them, demonstrates that they are cellular signals designed to alter the normal growth, metabolism and differentiation of tissue. Their vertical transmission identifies simply an hereditary predisposition to possible neoplasia, just as their lateral transmission defines infection, Correa & Correa Oncogene Theory of Cancer 15 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 Fig. 4 Typical genomic structures of replication-competent (Moloney Murine Leukemia Virus, Mo-MuLV) and replication-deficient (Friend Spleen Focus Forming Virus, F-SFFV) retroviruses. The latter have large deletions in the pol and env genes. The Friend SFFV comes in two varieties, anemia-inducing and polycythemia-inducing. The latter produces a surface protein (gp55) from the env gene that renders host cells (erythroid progenitors) independent of EPO, the hormone that normally regulates RBC production. which depends not just upon their infectivity potential, but also on encounters with hosts that are susceptible, genetically and dynamically (see below). 4. The broad classes of retroviruses Two broad classes of retroviruses initially appeared to exist: leukemia-inducing and sarcomainducing. Retroviruses that transform fibroblasts in culture only induce formation of solid tumors (sarcomas) in connective tissue, whereas most retroviruses only induce hyperplasia (not neoplasia) of blood cells, and are thus known as leukemia (leukocytosis-inducing, to be more exact) viruses. The transforming sarcoma-inducing retroviruses are typically replication-defective (mostly due to lack of a complete viral envelope-coding env gene, see Fig. 4); once inserted into the host genome, they replicate only when the host genome replicates. They become therefore latent viruses. Whereas the leukemia-inducing nontransforming retroviruses were considered to be replication-competent but not cytopathic, ie they did not lyse the host where they independently replicated (see Fig.s 4 & 5). Sarcoma-inducing retroviruses can only replicate independently of host cell replication when rescued by a leukemia-inducing retrovirus that is referred to as a “helper” virus. For a long time it was thought that there were no carcinoma-inducing retroviruses, but that retroviruses or their genes would mediate, as co-carcinogens, the effect of chemical mutagens in carcinogenesis. Nontransformed fibroblast cell lines that were “spontaneously immortalized” in culture (it is unknown how) could be transformed into tumor cells by chemical mutagenesis; moreover, whereas primary rodent fibroblasts failed to transform when exposed to certain chemical mutagens, exposure after infection with murine leukemia viruses resulted in transformation [46]. A two-step model of induced carcinogenesis was proposed, where viral infection worked as an ‘initiator’ that altered a normal cell into a pre-malignant cell, and the chemical or physical mutagen functioned as a ‘promoter’ inducing neoplastic transformation. The infecting virus might be defective and remain silently integrated until the mutagen intervened to induce its expression or activity and result in neoplasia. We should remark that this was the beginning of a new conceptualization of cancer as a process that crossed defined stages with distinct characteristics, initiation events being required before fully developed neoplasia manifests itself. Oncogenesis could therefore be thought as a vector of transformation, with pre-neoplastic lesions launching the vector. Thus the concept of an oncogenic vector emerged directly from the study of viral oncogenesis. The broad distinction of two classes of retroviruses was subsequently corrected with the discovery of a carcinoma-inducing retrovirus with strong oncogenicity [47], that contained two unrelated and independently expressed oncogenes, v-myc and v-mil(raf ). Localization of the corresponding proto-oncogenes which the virus transduced also showed these were unlinked [48]. Transformation by Journal of Biophysics, Hematology and Oncology, Vol. 1, 4:1-79 April 2010 16 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 Myc/Mil(Raf ) proteins was a cooperative event that by increasing the expression of c-Fos and c-Jun proteins (more on this below) activated DNA transcription of genes implicated in the control of cell migration, invasion and metastasis [49]. 5. Retrovirus-induced immortalization, transformation and the differentiation switch 5.1. Immortality and transformation Central to the determination of whether leukemia-inducing retroviruses were able to induce neoplastic transformation was the distinction between induction of immortality and malignant transformation, which required determination of what exactly were the obligatory and universal traits of the latter and the conditions under which preneoplastic cells progressed to full neoplasia. The relation between immortalization and transformation was from the beginning inseparable from the conceptualization of a ‘tumorigenic process’ ( the ‘oncogenic vector’) that crossed distinct stages and, through successive epigenetic and then genomic and phenotypic alterations, finally produced fully tumorigenic, malignantly transformed cells. Thus oncology would come to realize that immortalized cells were not transformed neoplastically, and not all malignantly transformed cells were tumorigenic or had the same tumorigenic potential. Correa & Correa Oncogene Theory of Cancer 17 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 Fig. 5 Diagram of the reproductive cycle of replication-competent retrovirus, from infection (at the top) to extrusion of a complete replicated viral particle (at the left) budding from the host cell membrane. Hundreds of copies may be produced by the same cell. Normal diploid mammalian cells have a limited lifespan and proliferative potential, clocked by the number of their cellular divisions, with a final phase of senescence before they die (see Fig. 6). This sequence of events is observed regularly in tissue culture of primary cells. In contrast, immortalized cells can give rise to cell lines (even though the precise origin of these adaptations to tissue culture has not, for the most part, been to this day established) because their proliferative potential is unlimited (see Fig. 6). Chemical mutagens can induce immortalization [50] (even though these tests were performed with sera that, back when these studies were conducted, were still commonly contaminated with mycoplasmas, not to mention chemical contaminants that are part of serum, such as undefined toxins, including bacterial endotoxin, and various growth and regulatory factors). So can infection with certain viruses. Most importantly, all malignantly transformed cells have been immortalized, but not all immortalized cells are malignantly transformed ie capable of tumor formation or tissue invasion or present gross genotypic alterations characteristic of neoplastic cells. Immortality is an essential step in the process of the malignant transformation of cells [50-51]. Defined DNA fragments from Herpes simplex virus type 2 (HSV-2) were able to either immortalize and transform, or just immortalize, embryonic rodent cells grown in vitro, again indicating that neoplastic transformation involved at least two steps [50]. Experiments like these led to the hope that there might be universal pathways to oncogenesis, that only a few oncogenes might be involved, and that, at least with respect to the transformation of the immortalized stage, single-hit event kinetics might suffice [52]. Yet, HSV-2 is not tumorigenic in vivo. Nevertheless, the in vitro findings appeared to tally with the oldest phenotypic model of oncogenesis, which held that cells first reversed or regressed to an embryonal stage (anaplasia), recovering some pluripotential state and engaging in some degree of excessive proliferation (hyperplasia), and only afterward would they become malignantly transformed into some other tissue (metaplasia), thought of either as an amoeboid ‘tissue’ (a tumor can be compared to a colony or colonies of protozoa, including parallel tissue-invading abilities) or a malignant and abortive variant of embryonic differentiation. The malignant transformation of the cell was ‘metaplastic’ and the tumorseeding ability was ‘neoplastic’ (growth of ‘new tissue’). In this model, immortalization or non-senescence would likely be connected to pre-neoplastic, hyperplastic stages of proliferation. However, hopes that either immortalization or transformation could be achieved by simple event kinetics would evaporate by the late 1990’s, as it became clear that cell lines were established by virtue of mostly undefined changes in large arrays of very different types of genes, and not just changes in oncogenes and tumor-suppressor genes. Even induction of immortality is not an all-or-none process, presenting “different degrees of reduction in the commitment of cells to a non-proliferative state” [50]. Likewise, there are different degrees of transformation. Journal of Biophysics, Hematology and Oncology, Vol. 1, 4:1-79 April 2010 18 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 Correa & Correa Oncogene Theory of Cancer 19 © Akronos Publishing, Canada, 2010 ISSN 1920-3799 Fig. 6 In Vitro Cell Growth Normal vs Transformed No. of cell divisions