Role of Viruses in the Causation of Human Cancer

To prove a causal relationship between a putative cancer-causing virus and human cancer is not a

simple task. Such proof relies on evidence that is to a fair extent circumstantial. This evidence includes

(1) epidemiological data showing a correlation between living in an area of endemic viral infection and a type of cancer; (2) serological evidence of antibody titers to viral antigens in patients with a given cancer type; (3) evidence for insertion of viral DNA into a cancer-bearing host’s cell genome; (4) evidence for a consistent chromosomal translocation, particularly those involving an oncogene, in virally infected patients; (5) data showing that viral infection of cells in culture or transfection of viral genes into cells

causes cell transformation and the ability of such cells to produce tumors in nude mice; and (6)

development of cancers of the suspected target organ in transgenic mice produced by embryonic

gene transfer of viral genes.

On the basis of this sort of evidence, some human cancers are considered to be caused by viral infection either directly or indirectly. By ‘‘directly,’’ I mean that the viral gene(s) can themselves cause cells to become malignant (sometimes also requiring the loss of a tumor suppressor gene). By ‘‘indirectly,’’ I mean that viral infection may simply cause the progression of malignant cell growth by producing an

immunodeficiency state (e.g., the occurrence of non-Hodgkin’s lymphoma in HIV-infected patients)

or by stimulating the proliferation of already transformed cells. Sometimes viral infection acts in concert with other infectious agents or chemical carcinogens. Such is the case for malarial infection of Epstein-Barr virus (EBV)– infected patients and for aflatoxin exposure of individuals bearing the hepatitis B viral genome in their liver cells (see below). The types of human cancer thought to be caused by viral

infection and the strength of epidemiological associations Epstein-Barr virus has been linked to four different types of human cancer: Burkitt’s lymphoma (BL), nasopharyngeal carcinoma (NPC),

B-cell lymphomas in immunosuppressed individuals such as HIV-infected patients, and some

cases of Hodgkin’s lymphoma.149 The evidence is strongest for an association with BL and NPC.

Infection with EBV does not by itself cause cancer. On average, across the world, about 90% of the population may be infected by the time they reach adulthood.

 In some endemic areas, the incidence rate approaches 100%. In developing countries, EBV infection often occurs in young childhood. In more affluent societies, EBV infection tends to occur as the ‘‘kissing age’’ of adolescence or young adulthood is reached, and manifests itself as infectious mononucleosis.

In developing countries, particularly in equatorial Africa, concomitant or subsequent infection with the malarial parasite induces B-cell proliferation and an immunodeficiency state that leads to malignant transformation and progression. There is a consistent chromosomal translocation involving immunoglobulin genes, usually on chromosome 14, and sequences within or adjacent to the c-myc gene locus on chromosome. The role of EBV in NPC is less well characterized, but the evidence for an association includes high serum antibody titers against EBV antigens and the presence of EBV DNA in NPC cells. Similar evidence suggests an association between EBV infection and induction of some

B-cell lymphomas and some Hodgkin’s disease cases in immunosuppressed individuals, although

the exact role of EBV remains to be elucidated.

Read more...

VIRAL CARCINOGENESIS

It has long been suspected that various forms of cancer, particularly certain lymphomas and

leukemias, are caused or at least ‘‘co-caused’’ by transmissible viruses. This theory has had its ups

and downs during the first half of this century, and it was not generally accepted until the 1950s

that viruses can cause malignant tumors in animals. The known carcinogenic effects of certain

chemicals, irradiation, chronic irritation, and hormones did not fit with the idea of an infectious

origin of cancer. In early experiments, the basic assay to determine whether cancer could

be induced by a transmissible agent involved transmititng malignant disease by inoculation of

filtered extracts prepared from diseased tissues. If the disease occurred in animals inoculated

with such filtrates, it was assumed to be caused by a virus. In 1908, Ellermann and Bang137

transmitted chicken leukemia by cell-free, filtered extracts and thus were among the first to

demonstrate the viral etiology of this disease. In 1911, Rous induced sarcomas in chickens by

filtrates obtained by passing tumor extracts through filters that were impermeable to cells

and bacteria. These findings remained dormant for two decades until Shope showed, in 1933,

that the common cutaneous papillomas of wild rabbits in Kansas and Iowa were caused by a

filterable agent.139 It was later found that when these tumors were transplanted subcutaneously

they became invasive squamous cell carcinomas.  

In 1934, Lucke´ observed that kidney carcinomas commonly found in frogs in New

England lakes could be transmitted by lyophilized cell-free extracts.141Twoyears later, Bittner

demonstrated the transmission of mouse mammary carcinoma through the milk of mothers to

offspring.142 This was the first documented example of transmission of a tumor-inducing virus

from one generation to another. Drawing on the experiments of Bittner, Gross postulated that

mouse leukemia was also caused by a virus and that occurrence of the disease in successive generations of mice was due to transmission of virus from parents to offspring. The proof of this hypothesis eluded Gross for a number of years until he was prompted, by evidence based on transmission of Coxsackie viruses to newborn mice, to attempt inoculation of mice less than

48 hours old. Using this approach, he successfully transmitted mouse leukemia by injecting

filtered extracts preparedfrom borgans of inbred AK or C58 mice, which have a high incidence

of ‘‘spontaneous’’ leukemia, or from embryos of these mice, into newborn C3H mice, which have

a very low incidence of leukemia. These experiments demonstrated for the first time that mouse

 leukemia is caused by a virus and that the virus is transmitted in its latent form through embryos.

This led to the isolation of a mouse leukemia virus.The isolated virus was also found to induce

Leukemias and lymphomas in inbred strains of mice. Electron-microscopic studies145 showed that

the mouse leukemia virus is spheroid, has a diameter of about 100 nm, and contains a dense,

centrally located ‘‘nucleus’’ separated from the external envelope by a clear circular zone. The

Gross mouse leukemia virus was classified as a type C virus, a term now used to describe a wide

variety of RNA-containing oncogenic viruses of similar morphology.

The RNA oncoviruses have been classified by morphological criteria. Intracytoplasmic type A

particles were initially observed in early embryos of mice and in certain murine tumors. These A

particles are noninfectious, bud into intracellular membranes rather than through the plasma

membrane, and thus stay within the cell. They have an active reverse transcriptase and exist as

a proviral form in chromosomal DNA. Type B viruses have spikes on their outer envelope, bud

from cells, and have been identified primarily in murine species, mouse mammary tumor virus

(MMTV) being an example. Type C viruses have been found widely distributed among birds and

mammals, can induce leukemias, sarcomas, and other tumors in various species, and have certain

gene sequences that are homologous to ‘‘transforming’’ sequences isolated from various

human tumors (see below). Another subgroup, type D RNA oncoviruses, has been isolated

from primate species but their oncogenic potential is not well established. The subtypes of

RNA tumor viruses, known as Retroviridiae, share a genetically related genome containing a

gag-pol-env gene sequence coding for virus internal structural proteins, the special type of

RNA-directed DNA polymerase called reverse transcriptase and viral envelope proteins, respectively.

Thus, they most likely share a common evolutionary heritage.146 However, distinct subclasses of retrovirus evolution, based on pol gene sequence homologies, have been found; one major pathway gives rise to mammalian type C viruses and a second to A, B, D, and avian type C oncoviruses.146 A more recent addition to the retrovirus classification is the human Tcell leukemia virus (HTLV), isolated from patients with certain forms of adult T-cell leukemias (discussed later). The pol gene of HTLV

appears to have evolved from a progenitor common to the types A, B, D, and avian C oncoviruses

rather than from the mammalian C type.146 If true, this would be unusual because most mammalian type C viruses share antigenic determinants among several gag, pol, and env gene products, suggesting a common progenitor for this subclass of retroviruses.

Read more...

DNA REPAIR MECHANISMS

Not all interactions of chemicals and irradiation with DNA produce mutations. In fact, all cells

have efficient repair mechanisms that repair such lesions. DNA repair mechanisms include sets of

enzymes that survey DNA for specific kinds of damage, remove the altered portion ofDNA, and

then restore the correct nucleotide sequence. The important role of DNA repair in human

cancer has been established by the finding that a number of inherited defects in DNA repair systems

predispose individuals to getting cancer. These diseases include xeroderma pigmentosum,

ataxia telangiectasia, Fanconi’s anemia, Bloom’s syndrome, Cokayne’s syndrome, and hereditary

retinoblastoma. There are several types of DNA repair systems, a number of which have been preserved

from bacteria to humans. These include (1) abnormal precursor degradation, e.g., the hydrolysis of the oxidized nucleotide triphosphate 8-hydroxy-dGTP to its nucleotide 8-OHdGMP, preventing incorporation into DNA; (2) a visible light-activated photoreactivation repair mechanism for removal of UV-induced cyclobutane pyrimidine dimmers; (3) strand break repair via an action of DNA ligase, exonuclease,

and polymerase activities; (4) base excision repair that recognizes simple base alterations such

as cytosine deamination to uracil and requires the action of (a) a purine or pyrimidine glycosylase

that breaks the deoxyribose-base bond, (b) an endonuclease to cleave at the abasic site, (c) a phosphodiesterase to clip away the ‘‘naked’’ abasic site, (d) DNA polymerase, and (e) DNA

ligase to refill and reclose the site; (5) nucleotide excision repair that recognizes bulky DNA base

adducts, pyrimidine dimers, and base crosslinks and requires the concerted action of enzymes and recognition factors (see below); and (6) 06-alkyguanine-DNA alkyltransferase that recognizes and removes small alkyl adducts from DNA. In mammalian cells, key repair mechanisms are base excision repair, nucleotide excision repair, transcription-coupled repair, homologous recombination and end joining, and mismatch repair.

 Excision repair is the most generalDNArepair mechanism in higher organisms. Base excision repair removes damage such as deaminated bases, oxidized or ring-opened bases generated by hydroxyl or superoxide radicals, and abnormally methylated bases such as 3-methyladenine.126 Nucleotide excision repair requires sequential steps of (1) preincision recognition of damage; (2) incision of the damaged DNA strand at or near the damaged site; (3) excision of the damaged site and local removal of nucleotides in both directions from the defect in the affected DNA strand; (4) repair replication to replace

the excised region, using the undamaged strand as a template; and (5) ligation to join the repaired

sequence of nucleotides at its 30 end to the contiguous DNA strand.125 DNA repair is usually very accurate, but if repair cannot occur prior to or during DNA replication it may be error prone. This errorprone, post-replication repair seems to be brought into play by certain types of agents or

when a cell is overwhelmed by damage that it cannot handle by excision repair before the cell

enters S phase during the next round of cell division. In this case, the new DNA is synthesized

on templates that still contain damaged bases, leading to mispairing or recombinational events that transfer damaged bases to daughter strands. For example, in mammalian cells, 5% to 30% of UV-induced thymidine dimers are transferred from parental to daughter strands during postreplication repair.129

Nucleotide excision repair (NER) of DNA in eukaryotic cells requires several gene products. Some of these gene products appear to be identical or highly homologous in yeast, rodents, and humans.130,131 A number of defects in the NER system have been found by studying mutations in cells from patients with xeroderma pigmentosum, in whom at least nine different kinds of mutations (i.e., nine different  complementation groups) have been found.125 Some of these XP genes have been cloned and found to

be highly homologous to yeast RAD genes that are required for excision repair in Saccharomyces

cerevisiae.130–133 Some of the cloned human genes also correct repair defects in mutant rodent cells and are called excision repair crosscomplementing (ERCC) genes.

Read more...

MULTIPLE MUTATIONS IN CANCER

In most cases, it takes years for a full-blown invasive, metastatic cancer to develop from a

small clone of initiated cells. This process might take 20 years or more, during which time an

initiated clone of cells undergoes clonal expansion via multiple cell doublings. As these clones

expand, various cells in the population accumulate multiple genetic alterations, some of which

facilitate dysregulated cell proliferation and some of which lead to cell death. These genetic

alterations can include point mutations, chromosomal translocations, gene deletions, gene

amplifications, loss of genetic heterozygosity (LOH), and loss of genetic imprinting (LOI).

This accumulation of genetic defects that occurs during clonal expansion of transformed cells is

due to ‘‘genetic instability.’’ The cause of this genetic instability is not clearly understood, but

it includes defects in cell replication checkpoint controls and decreased ability to repair DNA

damage.

There is evidence for the accumulation of thousands of mutations in cancer cells derived

from human tumors. For example, examination of the colon tumor–derivedDNA from patients with

hereditary non-polyposis colon cancer (HNPCC) reveals that as many as 100,000 repetitive DNA

sequences are altered from the mismatch DNA repair defects that these patients’ cells harbor

(reviewed in Reference 122). Mismatch repair defects have also been noted in ‘‘sporadic’’ (not

known to be hereditary) cancers. As noted earlier, one hypothesis explaining the genetic instability of transformed cells is the mutator phenotype hypothesis, championed by Loeb and colleagues.122 This hypothesis states that an ‘‘initial mutator [gene] mutation generates further mutations including mutations

in additional genetic stability genes, resulting in a cascade of mutations throughout the genome.’’ The molecular defect that could provide this phenotype could be a mutation in DNA polymerases that leads to error-prone DNA replication. The mutator phenotype would have to be generated early in tumorigenesis for this hypothesis to be valid. There are a number of arguments against this idea, such as observations

that there is not necessarily an increased mutation rate in cancer cells over that of normal cells123 and that a similar ‘‘evolution’’ of genetically altered cancer cells could arise by clonal selection followed by clonal expansion of cells with a genetic alteration that provides a proliferative advantage.

Read more...

GENETIC SUSCEPTIBILITY AND CANCER

As was noted above, there are a number of inherited cancer susceptibility gene mutations,

such as xeroderma pigmentosum, Fanconi’s anemia, and ataxia telangiectasia. These types of

inherited defects that lead to cancer are generally caused by a deficiency in DNA repair pathways.

Almost certainly we have only scratched the surface of inherited cancer susceptibility

genes that make an individual more prone to developing cancer. Other susceptibility genes

may include alterations in the metabolic enzymes that metabolize drugs and environmental

toxins, polymorphisms in genes that regulate utilization of certain essential nutrients such as

folic acid, or inherited mutations in tumor suppressor genes.

The completion of the Human Genome Project allows a systematic approach to discovering

the genetic alterations thatmakeindividuals prone to developing various diseases. The Environmental

Genome Project is producing a catalogue of variation in genes involved in catabolizing toxins,

nutrient metabolism, and DNA repair.121 These data, which will be largely generated by detection

of single nucleotide polymorphisms (SNPs), will enable toxicologists and cancer biologists to predict

individual susceptibility to diseases triggered or promoted by environmental pollutants, diet,

and other lifestyle factors. Some examples of this SNP analysis approach are the increased susceptibility

of individuals with altered folate metabolism genes to develop leukemia after benzene

exposure and the ethnic variation in the BRCA1 gene SNPs that affect susceptibility to

breast cancer.

Read more...

Ultraviolet Radiation

Ultraviolet radiation–induced lesions, generated by UV-B (280–320nm wavelength) or UV-A

(320–400nm wavelength), result from DNA damage, which is converted to mutations during

cellular repair processes. UB-B and UV-A generate different types of DNA damage and DNA

repair mechanisms (reviewed in Reference 113). Irradiation with UV-B produces cyclobutane

pyrimidine dimers that are repaired by nucleotide excision repair. If left unrepaired,

C?T and CC?TT base transitions occur. UVA- induced DNA damage produces mostly oxidative

lesions via photosensitization mechanisms and is repaired by base excision repair.

UV-B and UV-A also produce different effects on the immune system and elicit different transcriptional

and inflammatory responses. While the specific mechanisms by which UV radiation

induces basal cell or squamous cell carcinomas or melanoma are not clear, a number of signal

transduction pathways are affected that can either lead to apoptosis or to increased cell proliferation.

UV irradiation activates receptor tyrosine kinases and other cell surface receptors. It also enhances phosphorylation by ligand-independent mechanisms via inhibition of protein tyrosine phosphatase activity. Liganddependent cell surface receptor activation can also occur by activation of autocrine or paracrine

release of growth factors from keratinocytes, melanocytes, or neighboring fibroblasts. It is clear, however, that better animal models are needed to clearly define the mechanisms by which UV light causes human cancer. OXYGEN FREE RADICALS, AGING, AND CANCER The diseases of aging include  cardiovascular disease, decline in function of the immune system, brain dysfunction,

Read more...

Ionizing Radiation

The history of radiation carcinogenesis goesback a long way.The harmful effects of X-rays were observed

soon after their discovery in 1895 by W. K. Ro¨ ntgen. The first observed effects were acute, such as reddening and blistering of the skin within hours or days after exposure. By 1902, it became apparent that cancer was one of the possible delayed effects of X-ray exposure. These cancers, which included  leukemia, skin cancers, lymphomas, and brain tumors, were usually seen in radiologists only after long-term exposure before adequate safety measures were adopted, thus it was thought that there was a

safe threshold for radiation exposure. The hypothesis that small doses of radiation might also cause cancer was not adopted until the 1950s, when data from atomic bomb survivors in Japan and certain groups of patients treated with Xrays for noncancerous conditions, such as enlarged thyroids, were analyzed. These and other data led to the concept that the incidence of radiation-induced cancers might increase as a linear, nonthreshold function of dose. Thus the debate about whether there is a safe threshold pertains to radiation carcinogenesis, just as it does to chemical carcinogenesis.

In radiation carcinogenesis, the damage to DNA, and hence its mutagenic and carcinogenic effect, is due to the generation of free radicals as the radiation passes through tissues. The amount of radical formation and ensuing DNA damage depend on the energy of the radiation. In general, X-rays and gamma rays have a low rate of linear energy transfer, generate ions sparsely along their tracks, and penetrate deeply into tissue. This profile contrasts with that of charged particles, such as protons and a particles, which

have a high linear energy transfer, generate many more radical ions locally, and have low penetration through tissues. Thedamage toDNA can include single- and double-strand breaks, point mutations due to misrepair deletions, and chromosomal translocations.107–109 The molecular genetic events that follow radiation damage to cells include (1) induction of early-response genes such as c-jun and Egr-1; (2) induction of later-response genes such as tumor necrosis factor-a (TNF-a), fibroblast growth factor

(FGF), and platelet-derived growth factor-a (PDGF-a); (3) activation of interleukin-1 (IL-1) PKC110; and (4) activation of oncogenes such as c-myc and K-ras.111 Induction of these genes may be involved in the cellular responses to irradiation and in the longer-range effects that lead to carcinogenesis. At any rate, the production of clinically detectable cancers in humans after known exposures generally occurs after long latent periods. Estimates of these latent periods are 7 to 10 years for leukemia, 10– 15 years for bone, 27 years for brain, 20 years for thyroid, 22 years for breast, 25 years for lung, 26 years for intestinal, and 24 years for skin cancers.

A more recent example of nuclear fallout leading to environmental exposure to radiation is the Chernobyl accident, which happened onApril 26, 1986. A steam explosion blew the lid off the reactor. The graphite core caught fire and over 1019 becquerels (Bqs) of radioisotopes were released, producing a fallout that covered much of Belarus, Northern Ukraine, and part of the Russian Federation. Estimates are that

10–20 million people were exposed to significant fallout. There were some deaths due to acute

radiation sickness from high levels of exposure. However, the long-term effects are still being

recorded. So far, the reliable reports of increases in cancer incidences are mostly limited to thyroid

cancer.112 This finding is in contrast to cancer incidence among atomic bomb survivors in Japan, some of whom developed cancers of various types, including cancers of the thyroid, breast, lung, stomach, esophagus, bladder, leukemia, and lymphoma (although the incidence of cancers in Japanese atomic bomb survivors was less than would have been predicted by radiation exposure). The reason for this discrepancy is most likely that those exposed to the Chernobyl fallout received primarily dosage from

b-emitters, mostly isotopes of iodine, which concentrates in the thyroid. Atomic bomb survivors,

by contrast, received whole-body irradiation from neutrons and gamma rays.

Another interesting point about the Chernobyl survivors is that the type of thyroid cancer they developed, mostly among those under 2 years of age if they were exposed, were 98% papillary, many with an unusual morphology, whereas in non-exposed populations, only 67% of childhood thyroid cancers are papillary.112 Expression of two families of oncogenes, the c-ret and ras families, has been shown to be

involved in papillary thyroid cancers. The oncogene c-ret is a receptor tyrosine kinase activated

by gene rearrangement, and two of these, ret-ptc 1 and ret-ptc 3, are activated in papillary carcinomas. Since c-ret is activated by rearrangement, the high proportion of doublestrand DNA breaks seen in radiation-induced papillary carcinomas of the thyroid may explain its activation. Since the thyroid is not the only tissue that concentrates iodine, malignancies of other tissues that also concentrate iodine, such as the breast, salivary gland, and stomach, may appear in higher incidence as time goes on. Moreover,

other isotopes including cesium were present inthe fallout, and inhabitants of parts of the Ukraine and Belarus are still exposed to low levels of radioactive cesium. The long-term effects, if any, of such exposure is not yet clear.

Read more...

IRRADIATION CARCINOGENESIS

A number of the points made about chemical carcinogenesis are also true for radiation-induced

carcinogenesis. Both X-rays and ultraviolet (UV) radiation, for example, produce damage to DNA.

As with chemical carcinogens, this damage induces DNA repair processes, some of which are

error prone and may lead to mutations. The developmentofmalignanttransformationincultured

cells after irradiation requires cell proliferation to ‘‘fix’’ the initial damage into a heritable change and

then to allow clonal proliferation and expression of the typical transformed phenotype.105 Fixation

appears to be complete after the first postirrad  transformation requires an additional 12 rounds

Measured cancer risk at high dose Increasing incidence of tumors Increasing. Linear curve. Curves with this appearance

are not usually found experimentally in dose– response assays, and the idea that a dose–response

curve could take such a form is now considered obsolete. (From America’s War on ‘‘Carcinogens’’: Reassessing

the Use of Animal Tests to Predict Hum  Increasing dose . Nonlinear threshold. (From America’s War on ‘‘Carcinogens’’: Reassessing the Use of Animal Tests to Predict Human Cancer Risk, p. 53, with permission.) of cell division. Thus, as in the case of chemical carcinogenesis, a promotion phase is required for

full expression of the initiated malignant alteration. Moreover, when low doses of chemical

carcinogens and X-rays are used together, these two types of agents act synergistically to produce

malignant transformation.105

When cells are exposed to UV light in the 240 to 300 nm range, the bases acquire excited energy

states, producing photochemical reactions between DNA bases (reviewed in Reference 106). The principal products in DNA at biologically relevant doses of UV light are cyclobutane dimers formed between two adjacent pyrimidine bases in the DNA chain. Both thymine–thymine and thymine–cytosine dimmers are formed. That formation of these dimers is linked to mutagenic events .

Heavy exposure to sunlight induces similar changes in human skin, and the degree of exposure

to sunlight is closely related to the incidence of skin cancer. Whether continuing exposure to UV rays in sunlight is the promoting agent in skin cancer or additional promoting events are required is not clear, but it seems that UV irradiation is a complete carcinogen, just as some chemicals are—that is, it has both initiating and promoting activities. Patients who cannot efficiently repair UV-induced damage,

such as those with xeroderma pigmentosum, have a much higher risk of developing malignant skin tumors.

Read more...

Experimental Models for the Study of Carcinogenesis

A number of models for the study of carcinogenesis have been developed over the years. Historically, two of the most useful ones have been the initiation-promotion model of mouse skin carcinogenesis (the ‘‘skin-painting’’ model) and the induction of liver cancers in rats. The classic model of carcinogenesis is the single application of an initiating agent such as a polycyclic aromatic hydrocarbon followed by the

continuous application of a promoting agent like TPA to the backs of shaved mice. Much of what we know about tumor initiation, promotion, and progression has come from this model system. Initiation and promotion during mouse skin carcinogenesis produce multiple benign squamous papillomas. A few squamous cell carcinomas eventually arise from the papillomas over many months. However, malignant conversion can be speeded up by exposure of papillomabearing mice to mutagens, which activates oncogenes such as H-ras and causes loss of tumor suppressor genes such as p53, as noted above.

The mouse skin carcinogenesis model is also a useful one in which to study the role of diet and

chemopreventive agents in carcinogenesis. For example, calorie-restricted diets have been shown to reduce the number and size of papillomas during and following promotion with TPA in DMBA-initiated SENCAR mice.85 Furthermore, the latency period for occurrence of carcinomas was increased and

the total number of carcinomas was decreased. Applicationofapigenin,aplantalkaloid,86retinoic

acid,87 and prostratin, a nonpromoting phorbol ester88 have been shown to inhibit the promotion

phase (appearance of papillomas) of mouse skin carcinogenesis.

Multistage carcinogenesis has also been observed for liver tissue. For example, Peraino et al.89 observed that a 3-week exposure of rats to AAF in the diet produced only a small number of hepatomas after several months, but if the animals were subsequently treated with Phenobarbital for several months after carcinogen feeding was discontinued, a high incidence of hepatomas was noted. Similar results have been obtained by Kitagawa et al.,90 who fed rats a nonhepatocarcinogenic dose of 2-methyl-N,Ndimethyl-

4-aminoazobenzene for 2 to 6 weeks, and then a dietary administration of Phenobarbital for 70 weeks. By 72 weeks, many large hepatocellular carcinomas had developed in the phenobarbital-treated animals, whereas only a few small tumor nodules were observed in the rats not given phenobarbital. Thus, the action of phenobarbital appears to be analogous to that of TPA in the mouse skin system—that is, it ‘‘fixes’’ the damage to cells induced by an initiating agent and causes a clone of cells arising from a

damaged cell to proliferate. However, whereas TPA stimulates DNA synthesis and hyperplasia in skin, phenobarbital produces only a transient and relatively small increase in DNA synthesis in liver. Perhaps that is all that is needed to fix the carcinogenic damage and to allow for the initial proliferation of a damaged clone of cells.

Once the damaged clone is present, it could undergo alteration due to its genetic instability and gradually progress to a detectable malignant tumor. This idea is supported by the experiments of Pitot et al.,91 who treated rats with a single dose of diethylnitrosamine by intubation 24 hours after partial hepatectomy (partial removal of the liver), which stimulates DNA synthesis and cell proliferation in the remaining

tissue. If the animals were then treated, starting 8 weeks later, with phenobarbital in the diet for 6 months, many small, phenotypically heterogeneous foci characterized by glucose-6- phosphatase–deficient areas, ATPase-deficient areas, and g-glutamyltranspeptidase-containing areas developed in the liver. Many of these animals also had hepatomas, for which the enzyme-altered foci appear to represent the early stage of neoplastic development. Thus in this case, phenobarbital appears to have stimulated the replication of dormant initiated cells, which, in the absence of the promoter, would not have proliferated. If each enzyme-altered focus observed in these experiments were a clone derived from a single cell, about 104 to 105 cells in the liver were ‘‘initiated’’ by diethylnitrosamine, and a very small number of these subsequently underwent clonal proliferation during phenobarbital feeding.91 Thus the conversion of these abnormal foci, or early nodules, as they have been called, to a malignant neoplasm is a rare event.

Newermodels of carcinogenicity have involved the use of knock-out or knock-in rodent models, in which various oncogenes, tumor-suppressor genes, or susceptibility genes have been engineered into or out of rodent embryos (usually mice). This process has enabled the definition of some of the genes that are key to various steps in the tumor-initiation promotion and progression steps. These tumor models are now being superceded by conditional genetic knock-out models in mice that allow for the controlled expression of oncogenes or tumor suppressor genes in a way that more closely mimics ‘‘spontaneously’’ arising

human cancers.

Read more...

Mechanisms of Tumor-Promoting Agents

The terms tumor promotion, tumor progression, and multistage carcinogenesis are overlapping and somewhat redundant. Some people use these terms interchangeably and some use them to define discreet steps in the carcinogenesis process. Mechanistically, tumor promotion and progression are a continuum, even though they appear to be ‘‘multistage.’’ Promotion involves a clonal expansion (proliferative phase), and progression usually refers to the genetic alteration phase. But as was noted above, the genes involved in these steps are overlapping or similar. Nevertheless, studies of chemical carcinogenesis models have been used to define and discriminate initiation events and promotion or progression events, and these studies have been useful in determining the genetic and biochemical steps

involved in these steps, as well as providing targets for drug therapy and chemoprevention.

The isolation and characterization of tumorpromoting agents have provided the tools to study the mechanisms of tumor promotion in vitro and in vivo. The reader is reminded that these agents are primarily defined by their ability to promote skin carcinogenesis in the mouse skin-painting assay, and the mechanisms by which they do this may or may not be relevant to the mechanism of tumor promotion and

progression during carcinogenesis in other organs in experimental animals or in humans. Nevertheless, the study of these compounds has been extremely useful in determining the biochemical actions of tumor promoters. Of the promoting agents examined, the phorbol esters have been the most widely studied. Still, one must ask: what the ‘‘phorbol esters’’ are in human carcinogenesis. Most likely they are factors

to which we are continually exposed through our diet, cigarette smoke, and other kinds of environmental

agents. This answer leads to a second question: Do all these agents act through the same receptor or, if not, through the same biochemical steps? The answer is not known, but the list of potential promoters in the human environment is so large that it seems unlikely that they would all act by means of the same

proximal (‘‘receptor’’) mechanism. More likely, they act through different steps in a cascade leading

to the same end point—namely, clonal expansion of initiated cells and progressive selection of genetically variant populations of tumor  cells.

Tumor-promoting phorbol esters produce a wide variety of biochemical changes in cells. A number of these changes may be related to the ability of these agents to promote the growth of initiated tumor cells in vivo. Many of the cellular changes induced by phorbol esters are reminiscent of characteristics of the transformed phenotype (see Chapter 4). The effects of phorbol esters on cultured cells include (1) induction of ornithine decarboxylase, 50-nucleotidase, ATPase, and plasminogen activator activities; (2)

stimulation of sugar transport, DNA synthesis, and cell proliferation; and (3) alteration of cell morphology with a loss of cell surface fibronectin and the appearance of a diffuse pattern of actincontaining cytoskeletal elements (reviewed in Reference 60). In addition, phorbol esters stimulate anchorage-independent growth of adenovirus- transformed cells61 and inhibit the terminal differentiation of chicken myoblasts62 and chondroblasts,63 murine lipocytes,64 erythroleukemia cells,65 and neuroblastoma cells.66

Tumor-promoting phorbol esters also transform mouse embryo fibroblasts treated with ultraviolet

light67 and enhance the transformation of human lymphocytes by Epstein-Barr virus.68 These cell culture effects are exerted by low concentrations (nonomolar range) of phorbol esters, and there is generally a correlation between the potencies of phorbol esters for the cell culture effects and their potencies as promoters in mouse skin carcinogenesis. Phorbol esters share a number of biological properties

with epidermal growth factor (EGF) and may act by mechanisms similar to EGF. An interesting observation suggests that TPA can induce neoplastic transformation of fibroblasts from humans genetically predisposed to cancer.69 In these experiments, fibroblasts derivedfrom individuals with familial  adenomatosis of the colon and rectum were treated with TPA in culture and then injected into athymic mice. Cultures treated with TPA produced tumors in  These results indicate that the fibroblasts from

adenomatosis patients exist in an ‘‘initiated’’ state due to the dominant mutation that produces

the disease, and that this dominantly inherited trait can be induced to undergo malignant progression by treatment with promoting agents alone. This observation supports the idea that initiation of cancer is a mutagenic event and has profound implications for human cancer. For example, if the promoting agents present in our environment could be identified and exposure to them eliminated or significantly diminished,  could human cancer be prevented? This approach could conceivably be more effective

than eliminating exposure to initiating agents, since exposure to them need be only very short

and is irreversible. Completely preventing exposure to initiating agents over a lifetime is not

practical; however, if the promotion phase takes 15 to 20 years, expanding it to 30 to 40 years would mean that most individuals could have a life expectancy approaching normal before they developed a fatal cancer.

Read more...

Central Dogma of Tumor Progression

The standard concept of how cancer starts is that malignant tumors arise from a single cell transformed by a chemical carcinogen, oncogenic virus, radiation damage, endogenous genetic damage caused by oxidative insult to DNA, or any of a host of other potential ways (e.g., chronic infections with a bacteria such as H. pylori or with a parasite such as schistosomiasis, or hormonal imbalance). Once the initiated cell starts to undergo clonal expansion, it undergoes multiple genetic changes, due to genetic instability,

leading to an invasive metastatic cancer. This progression is thought to occur sequentially,

as exemplified by the work of Vogelstein and colleagues on colon cancer.46 The idea here

is that colon cancer goes through a series of ‘‘evolutionary’’ changes from hyperplasia, to earlystage

adenoma, to late-stage adenoma, to carcinoma, and finally to metastatic cancer.

There is, however, another point of view proposed by Weinberg and colleagues.56,57 This hypothesis, for which there are supportive clinical data, states that the genes involved in driving invasiveness and metastasis may be expressed early in the progression pathway and actually be the same genes involved in a selective growth advantage for these cells. These cells may be lurking even in early-stage cancers. That is, some cancers are predestined almost from the beginning to evolve into invasive, metastatic

tumors and some are not. This possibility has huge implications for cancer screening, diagnosis,

and choice of therapy. Numerous women receive a diagnosis of ductal carcinoma in situ of the breast based on mammography screening, and many men receive a diagnosis of prostate cancer based on a prostate-specific antigen (PSA) test and subsequent biopsy. And yet many of these patients have indolent tumors that would not affect their overall life expectancy, and they still often undergo significant surgical

and drug treatments. The problem is that we are only beginning to be able to tell (e.g., by gene

expression arrays) which of these so-called early-stage cancers will be lethal and which ones won’t.

Another point of the Weinberg theory is that the genetic alterations that occur during tumor progression do not necessarily occur in a given sequence and are probably different for different cancers.56 One might even suggest that they may be different in different patients who have the same histological tumor type. Ultimately, however, these genetic and phenotypic changes lead to a similar loss of cell proliferation control and expression of a panoply of genes (maybe not the identical ones) that make some tumors invasive and metastatic. There are clinical data supporting some of these concepts. In a study by van de Vijver et al.,58 it was determined that the gene expression profile of breast cancers was a much better predictor of disease outcome in patientswith breast cancer than standard clinical and  histopathological staging. Indeed, they could restratify patients listed as low risk or high risk by clinical staging into a more accurate prognostic outcome category (based on actual metastasisfree survival) through gene expression arrays. In addition, Al-Hajj et al.59 were able to identify and isolate the more tumorigenic cells from a

heterogeneous population of breast tumors in eight of nine patients. These more aggressive cell types were identified by their cell surface markers and by repeated passage in nude mice. Each time the more aggressive cells were injected  into nude mice they produced tumors, whereas the marker-negative cells did not grow. These data suggest that the aggressive tumorigenic cells can be prospectively identified in initial tumor biopsies containing mixed populations of cells and can be used to discriminate patients

with potentially more aggressive tumors.

Read more...

Mechanisms of Tumor Promotion and Progression

Tumor-initiating agents most likely act by interacting with DNA to induce mutations, gene rearrangements, or gene amplification events that produce a genotypically altered cell. What happens next is that the initiated cells undergo a clonal expansion under the influence of promoting agents that act as mitogens for the transformed cell type. As will be discussed later, these promoting actions appear to be mediated

by cell membrane events, although a direct action of promoters on DNA has also been proposed.

It is important to note that multiple clones of cells are likely to be initiated by a DNAdamaging agent in vivo and that, through a rare second event, one or a small number of these clones progresses to malignant cancer. It may be useful to think of the promotion phase as the stage of cell proliferation and clonal expansion induced by mitogenic stimuli and of the progression phase as the gradual evolution

of genotypically and phenotypically altered cells that occurs due to genetic instability of the

progressing cells. This process leads to the development of cell heterogeneity within a tumor, an idea first described by Foulds53 and later expanded by Nowell.54 During the progression phase, which can take many years in humans, individual tumors develop heterogeneity with respect to their invasive and metastatic characteristics, antigentic specificity, state of cellular differentiation, and responsiveness to hormones, drugs, and immune-modulating agents.

Presumably, some powerful selection process goes on to favor the growth of one progressing cell type

over another. This preferential selection may be due to a certain cell type developing a growth

advantage in the host’s tissues over its peers, as proposed by Nowell, or to the host’s immunologic

defense system being able to recognize and destroy some cell types better than others, thus

providing the selection pressure for expansion of one clone over another, or to a combination of

these factors. Experimental evidence supports such a selection of tumor cells growing in vivo.

For example, Trainer and Wheelock55 have shown that during the growth of L5178Y lymphoma

cells in mice, a continual selection of cells with a decreasing ability to be killed by cytolytic T lymphocytes (CTL) ‘‘armed’’ against the tumor occurs, until an ‘‘emergent phenotype’’ appears that is highly resistant to the CTL cells.

Read more...

Mechanisms of Tumor Initiation

Initiation of malignant transformation of normal cells by a carcinogenic agent involves a permanent,

heritable change in the gene expression of the transformed cell. This could come about by either direct genotoxic or mutational events, in which a carcinogenic agent reacts directly with DNA, or by indirect or ‘‘epigenetic’’ events that modulate gene expression without directly reacting with the base sequence of DNA. Most investigators favor the mutational theory of carcinogenesis—that is, that the initiating events involve a direct action on the genome. The mutational theory depends on three kinds of evidence:

1. Agents that damage DNA are frequently carcinogenic. As discussed previously, chemical carcinogens are usually activated to form electrophilic agents that form specific reaction products with DNA. The extent of formation of some of these reaction products, for example, alkyl-O6-guanine, has been shown to correlate with mutagenicity and carcinogenicity of certain chemical agents. Ultraviolet and ionizing

radiation also interact with DNA at doses that are carcinogenic.

2. Most carcinogenic agents are mutagens. A number of in vitro test systems using mutational events in microorganisms have been developed to rapidly screen themutagenic potential of various chemical agents. One of the best known of these, the Ames test, is based on certain characteristics of specially developed strains of the bacterium Salmonella typhimurium. The tester strain, amutant line that requires exogenous histidine for its growth (hisauxotroph), has a poor excision repair mechanism and an increased permeability to exogenously added chemicals. Using this system, together with a liver microsomal fraction that has the capacity to activate most chemical carcinogens metabolically, Ames and colleagues have shown that about 90% of all carcinogens tested are also mutagenic.36 Moreover,

few noncarcinogens show significant mutagenicity in this test system. Malignant transformation can be induced in a variety of cultured mammalian cells by agents that are mutagenic for the same cells. For example, carcinogenic polycyclic hydrocarbons cause mutations, as measured by induction of resistance to 8-azaguanine, ouabain, or elevated temperature, in Chinese hamster V79 cells if the cells are

cocultured with lethally irradiated rodent cells that can metabolize the hydrocarbons to their electrophilic, activemetabolite.37,38 In these studies,mutagenicity was obtained with the carcinogenic hydrocarbons 7,12- dimethylbenz(a)anthracene, benzo(a)pyrene, and 3-methylcholanthrene. There was no mutagenicity with a noncarcinogenic hydrocarbon, and the degree of mutagenicity was related to the degree of carcinogenicity of the chemicals in vivo.

3. Incidence of cancer in patients with DNArepair deficiencies is increased. In individuals with certain recessively inherited disorders, the prevalence of cancer is significantly higher than in the general population. 39 The connecting link between these disorders is the inability to repair certain kinds of physical or chemical damage to DNA. The high incidence of cancer in these diseases constitutes the best available evidence for a casual relationship between mutagenicity and carcinogenicity in humans.

One example of xeroderma pigmentosum (XP) is characterized by extreme sensitivity of the skin to sunlight and is the most widely studied of the repair-deficient human diseases. Virtually 100% of affected

individuals will eventually develop some form of skin cancer. In addition, heterozygotes who carry the XP gene but do not have the disease appear to have a higher incidence of nonmelanoma skin cancer.40

All individuals with XP are defective in  repair of ultraviolet damage to DNA, and most of them have a defect in the excision repair pathway. The repair defect ranges from 50% to 90% repair efficiency in cells from different patients, and there is good correlation between the severity of the molecular defect and the

extent of the disease. The defect in most patients appears to be at the nicking or incision step of excision repair, although patients in one complementation group have normal excision repair and are defective

in postreplication repair.39 The XP cells are also less efficient at repairing chemically induced damage to their DNA.

Read more...

Tumor Initiation, Promotion, and Progression

 The idea that development of cancer is a multistage process arose from early studies of virusinduced

tumors and from the discovery of the cocarcinogenic effects of croton oil. Rous and colleagues found that certain virus-induced skin papillomas in rabbits regressed after a period of time and that papillomas could be made to reappear if the skin was stressed by punching holes in it or by applying such irritant substances as turpentine or chloroform. These findings led Rous and his associates to conclude that tumor cells could exist in a latent or dormant state and that the tumor induction process and subsequent

growth of the tumor involved different mechanisms, which they called ‘‘initiation and promotion.’’27 Theterm cocarcinogen was coined by Shear, who discovered that a basic fraction of creosote oil enhanced the production of mouse skin tumors by benzo(a)pyrene.28 In 1941, Berenblum 29 reported that among mice receiving a single skin painting of a carcinogen, such as methylcholanthrene, only a small number of animals developed papillomas, but if the same area of skin was later painted repeatedly with

croton oil, which by itself is not carcinogenic, almost all the animals developed skin carcinomas.

Taken together, the data of these investigators suggested a multistage mechanism for carcinogenesis.

Studies of the events involved in the initiation and promotion phases of carcinogenesis were

greatly aided by the identification of agents that have primarily an initiating activity, such as urethane

or a low dose of a ‘‘complete’’ carcinogen (see below), and by the purification of the components

of croton oil that have only a promoting activity. Diesters of the diterpene alcohol phorbol

were isolated from croton oil and found to be the tumor-promoting substances.30,31 Of these,

12-O-tetradecanoylphorbol-13-acetate (TPA) is the most potent promoter.32 A scheme used to study the initiation–promotion phases of mouse skin carcinogenesis  is depicted in Figure 2–8. Typically, tumor initiation is brought about by the single application of an initiator, such as urethane, or a subcarcinogenic

dose of an agent with both initiating and promoting activity, such as the polycyclic hydrocarbon

benzo(a)pyrene; promotion is carried out by repeated application of a phorbol ester, such as TPA (e.g., three times a week).31,33 Benign papillomas begin to appear at 12 to 20 weeks and by about 1 year, 40% to 60% of the animals develop squamous cell carcinomas. If the promoting agent is given alone, or before the initiating agent, usually no malignant tumors occur.

The progression stage of carcinogenesis is an extension of the tumor promotion stage and results from it in the sense that the cell proliferation caused by promoting agents allows the cellular damage inflicted by initiation to be propagated, and the initiated cells are clonally expanded. This propagation of damaged cells in which genetic alterations have been produced leads to the production of more genetic alterations.

This genetic instability is the hallmark of the progression phase of carcinogenesis and leads to the chromosomal translocations and aneuploidy that are frequently seen in cancer cells.34 Such alterations in the genome of the neoplastic cell during the progression phase lead to the increased growth rate, invasiveness, and metastatic capability of advanced neoplasms. Some of the gene expression alterations

that occur during tumor initiation and promotion are shown in Figure 2–9 (see color insert). Evidence for multistage induction of malignant tumors has also been observed for mammary gland, thyroid, lung, and urinary bladder and in cell culture systems (reviewed in Reference 9), thus it seems to be a general phenomenon. This experimental evidence is consistent with the observed clinical history of tumor development in humans after exposure to known carcinogens—that is, initial exposure to a known

chemical or physical carcinogen, a long lag period during which exposure to promoting agents

probably occurs, and finally the appearance ofa malignant tumor.

Several characteristics of tumor initiation, promotion, and progression provide some insight into the mechanisms involved in these processes. Initiation can occur after a single, brief exposure to a potent initiating agent. The actual initiation events leading to transformation into a dormant tumor cell appear to occur within one mitotic cycle, or about 1 day for the mouse skin system.32 Furthermore, initiation appears to be irreversible; the promoting agent can be given for up to a year later and a high percentage of

tumors will still be obtained. Thus, the initiation phase only requires a small amount of time, it is irreversible, and it must be heritable because the initiated cell conveys the malignant alteration to

its daughter cells. All these properties are consistent with the idea that the initiation event involves a genetic mutation, although other ‘‘epigenetic’’ explanations are possible. The promotion phase, by contrast, is a slow, gradual process and requires a more prolonged exposure to the promoting agent. Promotion occupies the greater part of the latent period of carcinogenesis, is at least partially reversible,

and can be arrested by certain anticarcinogenic agents .Tumor promotion is a cell proliferation phase that propagates the initiated damage and leads to the emergence of an altered clone of cells. Most promoting agents are mitogens for the tissue in which promotion occurs. Tumor progression requires continued

clonal proliferation of altered cells, during which a loss of growth control and an escape from host defense mechanisms become predominant phenotypic traits. This process allows growth to progress to a clinically detectable tumor. The later events in the tumor progression phase are also thought to be irreversible because of the pronounced changes in the genome that have occurred leading into this phase. Agents thatare ‘‘pure’’ progression-causing agents are hard  to identify, but the free radical–generating agent

benzoylperoxide appears to be a progressioninducing agent during experimental epidermal  carcinogenesis.35 It should be noted that some potent carcinogens are ‘‘complete carcinogens’’ in that at certain doses they can by themselves induce a cancer. Such agents include polycyclic aromatic

hydrocarbons, nitrosamines, certain aromatic amines, and aflatoxin B1. When these agents are given in sufficient dose to animals during cancercausing protocols, they can cause DNA damage and produce tissue necrosis, which is itself enough to stimulate several rounds of cell proliferation in response to the tissue damage. In this situation, the promotion–progression phases are often collapsed in time, resulting in the production of aneuploid malignant cells.

Read more...