Reflections on the use of 10 IARC carcinogenic characteristics for an objective approach to identifying and organizing results from certain mechanistic studies

Abstract

To find a scientifically based method for evaluating mechanistic data related to risks to human beings, a new protocol for identifying, organizing, and summarizing mechanistic data for decision-making on cancer hazard identification was proposed by the International Agency for Research on Cancer and by an international working group of multidisciplinary experts. This Commentary examined the 10 key carcinogens’ characteristics proposed in the context of several paradigms assumed in the using of these 10 characteristics. These characteristics were assumed to represent a “carcinogen’s” mechanism of action but what was ignored were characteristics of the mechanisms of the “initiation,” “promotion,” and “progression” carcinogenic process. Challenges were made to the interpretation of genotoxicity data as well as from concepts and findings related to the promotion phase and the role of adult human stem cells. Reliance of interpretation of “genotoxicity” data (molecular-DNA lesions in DNA; induction of free radicals/oxidative stress markers; phenotypic surrogates of gene mutations), as well as from lesions in genomic versus mitochondrial DNA, or in the target cells for the carcinogenic process in either in vitro cultures or in vivo tissues, makes this “objective” use of the data questionable. A challenge to the “dedifferentiation” hypothesis of cancer was made. Because of an agent being misclassified as “genotoxic”—rather than an “epigenetic”—agent (which works by threshold levels; can be blocked; and must be present at critical times during development and at regular, sustained chronic exposures) could lead to unwise policy decisions.

Introduction

In the recent article, “Key Characteristics of carcinogens as a basis for organizing data on mechanisms of carcinogenesis”, ¹ an attempt was made to review, to explicate the 10 carcinogen characteristics, and to use these 10 characteristics, named by IARC, ² on two carcinogens, namely, benzene and polychlorinated biphenyls. While acknowledging the need to find a scientifically sound basis for identifying agents as contributors to human carcinogenesis, given the implicit and explicit assumptions behind this new approach, this “Commentary” reviews the proposed approach and finds uncertainty in some of the rationale and application of the data. In particular, the statement in this article is as follows,

The more recent description by Hanahan and Weinberg of hallmarks of cancer is not predicated on morphology or the impact of carcinogens, but on changes in gene expression and cell signaling (Hanahan and Weinberg 2011). These hallmarks are the properties of cancer cells and neoplasms, and are not characteristic of the agents that cause cancer.

seemed to contradict one known property of chemicals that are known to modulate gap junctional intercellular communication (GJIC).

Starting with a shared and positive note, the authors of this article correctly state that carcinogenesis is a multi-stage, multi-mechanism process, consisting of the “initiation,” “promotion,” and “progression” concepts. ^3,4 Therein, the first conceptual problem emerges. Referring to a chemical that produces a cancer as a “carcinogen” implies that it produces all the three steps of carcinogenesis, which have different operational characteristics, such as being, either irreversible or potentially interruptible or reversible, creates the first problem. A given chemical does not have all three properties “initiator,” “promoter,” or “progressor”, that is, an epigenetic promoter cannot be an initiator. Whatever that agent is, for example, ultraviolet light, benzene, poly brominated biphenyl, asbestos, human papilloma virus, poly aromatic hydrocarbons, alcohol, none of them, by themselves; produce all these different operational steps or their underlying mechanisms of action. It would seem that a more rigorous characterization of that agent should be a “carcinogenic initiator” or a “carcinogenic promoter” or a “carcinogenic progressor.” The reason for this is that the mechanisms responsible for these distinct steps are very different.

After reading the article, especially the description of each of the 10 characteristics, it seemed that two of the prevailing paradigms in the field of human carcinogenesis were both implicit and explicit in this new attempt to find order out of the current chaos in evaluating the “carcinogenic” potential (and other toxicities) of agents to which we are exposed. The first paradigm is the one that strongly derived from the paper by Dr Bruce Ames, ⁵ namely, “Carcinogens as mutagens.” The second paradigm was the assumption that the first step of carcinogenesis involved is the conversion of a normal cell to a malignant cell and the induction of “immortality” in that normal “mortal” cell. ⁶

This first important paradigm, namely, that mutations are the drivers of cancer cells, has some validity; however, carcinogenesis is more than mutagenesis! It can be argued that epigenetic mechanisms might be the driver and the rate-limiting phase of the carcinogenic process through the promotion phase.

The operational fact of the “initiating” event is that it is “irreversible.” If an organism is exposed to an agent that induces a cell that, after exposure to the other “operational” process/agent clonally amplies this single initiated cell (the “promoter”), one assumes DNA damage has occurred which can lead to a mutation. Herein, more complicating factors are found. First, the “initiated cell” might have previously existed prior to the exposure to the alleged agent, and the agent, such as chronic smoking, could have been the “promoter” of this preexisting “initiated” cell. ⁷ This possibly explains why lung cancers can occur in nonsmokers, even those not exposed to downstream smoke, ⁸ because these preexisting “initiated” lung cells were promoted by other agents.

The second major problem is to determine mutations in cells using in vitro surrogates of direct changes in the genome, such as drug resistance, for example, thymidine kinase mutation, ouabain resistance, or 6-thioguaine resistance. Without the means to detecting the oncogenic mutation in the single initiated cell that converts this normal cell to one that could, but does not necessarily, lead to a cancer, one must recognize the limitation and artifacts built into these various in vitro assays. ^{9 –11} Currently, the prevailing assumption of presumptive positive results in these in vitro assays to measure altered phenotypic changes is that the altered phenotype is the result of a true mutation. Clearly, this is not always the case. In other words, after exposure of normal drug sensitive cells to a presumptive genotoxic agent, one recovers a drug-resistant cell. It is often assumed that this drug-resistant cell’s phenotype was due to a mutation, when in point of fact, the same phenotype could be the result of the gene conferring that drug sensitivity could have been transcriptionally turned off, leading to a drug resistant phenotype. This would lead one to call the agent producing an altered phenotype a mutagen, when, in point of fact, it is an epigenetic agent. This is very serious because mutagens are very dangerous and produce irreversible changes in the genome, whereas an epigenetic agent, while also could be dangerous, especially during development, has to achieve threshold levels to be effective and could, as a tumor promoter, be interrupted or blocked by anti-tumor promoters and anti-oxidants. Lastly, to measure actual lesions in DNA, as a measure of genotoxicity of an agent, is, itself, subject to alterative explanations. Extracting DNA from exposed tissues, which contain three types of cells, for example, the few organ-specific adult stem cells, the many progenitors, and terminally differentiated cells, will be from a mixture of all these cell types. If the bulk of the DNA measured demonstrates the induction of DNA lesions, how does one know that those lesions occurred equally in all cell types, including the one that became initiated? These different cell types are epigenetically different and have different metabolism to protect or repair the damage in its DNA. In addition, when extracting the DNA from this population of different cell types, any attempt to separate the genomic DNA from the mitochondrial DNA is rarely carried out. This could be a very misleading factor if this separation of the two types of DNA is not done.

The second prevailing paradigm relates to the idea that to start the carcinogenic process of “initiation”, one must convert this single normal cell, which is assumed to be a differentiated, normal cell, that is destined to be mortal, such as proliferating human fibroblast cells. ¹² With the very convincing study, ⁶ a population of primary fibroblasts, when exposed to the oncogene, myc, leads to the recovery of a few “immortalized” cells, which after exposure to another oncogene, Ha-Ras, ultimately gave rise to a few neoplastic cells. This study strongly supported the assumption that the first step was to “reprogram” a mortal cell to the “immortal”, but not yet neoplastic state. This was assumed to be the initiation event. However, given the recent advances in the biological characteristics of stem cells, we now know that the definition of stem cells is that, under different signals, the stem cell could divide, symmetrically, to produce two stem cells that can self-renew or it could divide, asymmetrically, to produce one self-renewing stem cell daughter and one progenitor daughter that can terminally differentiate. By definition, a stem cell is naturally “immortal” until it is induced to become terminally differentiated or “mortal.” This prevailing hypothesis of the initiating event is the “dedifferentiation” or “reprogramming” of a “mortal” cell to become an “immortal” cell.

With these two important prevailing paradigms in mind, new discoveries of the biology of cancer and “cancer stem cell” ^13,14 should be examined in the context of this new 10 characteristics of “carcinogens.” A delineation of the fundamental challenge to this new protocol follows. It is based on what seems to be a real problem in distinguishing chemicals that might be a “genotoxicant,” a non-genotoxic cytotoxicant, or a non-cytotoxic “epigenetic” chemical. A systematic analysis of many of the elements of the multi-stage, multi-mechanism of carcinogenesis, must include stem cells as “targets” for the carcinogenic “initiating” event, and distinguishing characteristics of mutagens and epigenetic agents as mechanisms of the multi-stage, multi-mechanism process of carcinogenesis.

A. All cancers have a single cell origin.

It is generally accepted that all cancers are derived from a single normal cell. Even though the ultimate tumor, derived from that single cell, contains many cells with very different genotypes and phenotypes, they all are progenitors of that single initiated cell. ^15,16

B. Initiation step is irreversible and occurs in a normal cell that gives rise to the “cancer stem cell.”

Most, if not all, agree that cancers, including the so-called “cancer stem cell,” originated from a single “normal” cell. Therefore, the initiation event had to occur in this normal cell. The “initiation” event is also an IRREVERSIBLE step.

C. Mechanism for the initiating event is assumed to be a mutagenic event.

Since the “initiation” step is irreversible, it is assumed that mutagenesis by “errors in DNA Repair” or “errors in DNA Replication” is the underlying mechanism. As was correctly noted in the study by Smith et al., ¹ gene mutations can be either due to “errors in DNA repair”, as in the genetically predisposed human skin cancer prone, namely xeroderma pigmentosum, ^{17 –19} or due to “errors in DNA replication”, as in the Blooms syndrome. ²⁰

D. Not all positive clones in “mutation assays” are true mutations.

However, another major challenge to the idea that all agents that induce free radicals and induce oxidative stress are mutagenic to the genome in the targeted “initiated” cell. To reiterate, all the assays that have been designed to detect “mutagens” as phenotypic surrogates for actual mutations in cells (e.g. drug resistant phenotypes, 6-thioguine resistance, thymidine kinase minus cells, ouabain resistance, etc.). All of these assays have major flaws and false positive limitations. ^{9 –11} Most importantly, the use of different species and cell types (primary cells, non-normal or immortalized cancer cells in sparse or confluent cultures, 2-D or 3-D in vitro cultures, etc.) can dramatically influence both the results and interpretation of data. A comparison of one chemical, namely 2,4 dinitrofluorobenzene, caused genotoxicity in bacteria but acted as an epigenetic agent in mammalian cells. ²¹ Today, with very sophisticated analytical detection, DNA lesions can be measured in targeted tissues. However, since these lesions are not measured in the target single cell that leads to the initiated event, that, then, leads to another problem, especially if the lesions are measured in differentiated cells that have large numbers of mitochondria and have metabolic systems to convert chemicals to electrophiles and to generate free radicals. Another problem with the detection of minute quantities of DNA lesions will be discussed later in the context of the need for threshold levels during the promotion phase.

E. Any agent associated with a cancer can interact with an adult organ-specific stem cell, a finite-life progenitor, or differentiated cell.

If a genotoxic or “initiating” agent ultimately leads to a cancer, then the cancer that might result from that exposure is derived from that single targeted cell. The reason that this event is a rare event, which is probably due to the fact that there are very few target or stem cells in vivo or in vitro. When a radiation photon, a free radical-inducing chemical, or an “oncogenic” or “immortalizing” virus interacts with the whole organism, it ultimately interacts with a cell. In the whole organism or a population of primary cells in vitro, that cell could be an organ-specific adult stem cell; a progenitor and finite-living cell, or a terminally differentiated cell in any organ (liver, skin, breast, pancreas, immune tissue, etc.). Even in primary in vitro cultures, the early cultures can have all three cell types. ²² On further subculturing, especially in current media and 20% oxygen, the cultures ultimately senesce. ¹²

F. There are two opposing hypotheses on the origin of cancers: The stem cell hypothesis and the “de-differentiation” or “re-programming” hypothesis.

There are two hypotheses on the origin of the cancer or “initiated” cell, namely, the “stem cell” hypothesis ^{23 –27} and the “de-differentiation” ²⁸ or the up-to-date term, “re-programming” hypothesis. This is the fundamental point in our quest to find an accurate understanding of how a cancer is formed. To examine this important point, the “reprogramming” hypothesis, in the view of Dr S. Yamanaka’s Nobel prize-observation of creating “induced pluripotent stem” (“iPS”) cells from differentiated fibroblast primary cultures (Yamanaka, 2006), seems to support the Weinberg hypothesis that the first step of carcinogenesis is to “immortalize” a “mortal” cell by “reprogramming.” However, again, an alternative explanation to these valid observations can be made. Organ-specific adult stem cells are the target for initiated, cancer stem cells. ^{29 –34} Basically, the argument against the reprogramming of differentiated cells that will be responsible for starting the carcinogenic progress is from the operational definition of an iPS cell, that is, when the iPS cell is put back into an adult organism, it must form teratomas. If reprogramming existed in adult human beings, why is it that sarcomas and carcinomas are formed, but not teratomas?

G. Initiation as the irreversible blockage of asymmetric cell division of organ-specific adult stem cells.

The late Dr Van R. Potter, in his “oncogeny as partially blocked ontogeny” hypothesis, ²⁵ assumed that a stem cell, after exposure to an initiator, blocked the normal epigenetic mechanisms needed for terminal differentiation. That, then, leads one to define what is a stem cell, as opposed to a progenitor and terminally differentiated cell. A stem cell is defined as a cell that can divide indefinitely and has the ability, on receiving the right external signal, either to divide symmetrically to form two stem cell daughters or asymmetrically to form one stem cell daughter and one progenitor cell that can terminally differentiate, apoptose, or senesce. If the organ-specific stem cell is initiated by either an error of DNA repair, as in ultraviolet-irradiated skin, or starting out as a normal “immortal” cell, it can no longer divide asymmetrically to terminally differentiate. It can now live long enough to accrue other mutations and epigenetic changes if it is stimulated to proliferate in a sustained, chronic fashion.

H. First step of carcinogenesis is the blockage of “mortalization” rather than the induction of “immortalization” of the normal “mortal” cell.

This now brings in the powerful paradigm of Land et al., ⁶ which showed that a primary culture of cells in vitro, exposed to an oncogene, such as myc, could lead to the recovery of some “immortalized” cells. The “immortalization” of these primary fibroblasts was further explained. ³⁵ It has been assumed that all the cells in that primary culture were “mortal” cells, as was done when Takahashi and Yamanaka ³⁶ exposed their primary cells to embryonic genes, such as Oct4, Sox2, and so on. Weinberg, then, hypothesized that the first step of carcinogenesis was to “immortalize” a “mortal” differentiated cell. That immortalized, but not yet neoplastic, cell could then be transformed, neoplastically, with another type of oncogene, such as Ha-Ras. However, the alternative hypothesis is that both of these solid scientifically important observations (“immortalization” of a few cells in a primary culture by the Weinberg group or by the induction of a few rare iPS cells by Takahashi and Yamanaka) have an alternative explanation. The myc gene actually blocked the “mortalization” in the few adult stem cells in the primary cultures rather than to induce “immortalization” of the bulk of differentiated cells in the primary culture. ³⁴ The same explanation could easily explain the origin of iPS cells.

I. Do “immortalizing viruses” reprogram “mortal” normal differentiated cells or do they block the differentiation of normal “immortal” organ-specific adult stem cells?

With the Nobel prize given to Dr Zur Hausen for his observation that “oncogenic viruses” are associated with several human cancers, ³⁷ the same explanation can be applied to tissues (in vivo and in primary cultures in vitro). The SRC and HPV E-6, E-7 genes, when expressed in organ-specific adult stem cells, that are naturally “immortal,” will block asymmetric cell division of these stem cells. For example, expressed large T antigen in a cell can render nonfunctional p53 and RB proteins that would block the differentiation of a stem cell. ³⁸ These would be “initiated” but not neoplastically transformed cells.

J. The epigenetic clonal expansion of initiated cells by epigenetic agents.

By the operational definition of promotion, this process must take place after a single cell is “initiated.” ³⁹ Animal experiments verify this phenomenon. Promotion must take place after initiation in a sustained, regular, and chronic exposure to agents or conditions, such as wound healing and massive cell death, ⁴⁰ which are, themselves, non-genotoxic and in the absence of “anti-tumor promoters.” ⁴¹ These promoting agents must be given at “threshold” levels. ^42,43 Phenobarbital, TCDD, DDT, phthalates, estrogen, EGF, and scores of others can induce oxidative stress and DNA lesions in differentiated cells. However, they are not mutagenic to the nuclear genome of cells but can reversibly inhibit GJIC. ^44,45

The promotion process is potentially interruptible or even reversed. ⁴⁶ Some of the successes of epidemiology included the detection of the association of smoking with lung cancer; alcohol was associated with liver cancer, and sun light was associated with skin cancer. This is because individuals, who are heavy smokers, chronic alcoholics, and light skinned individuals exposed to sun, are at higher risk to cancers. Smoking one cigarette a day, having a beer or two a week, or being heavily pigmented or having skin not exposed to high levels of ultraviolet light does not cause non-genetically predisposed individuals to these cancers. These promoting agents, for examples, chemicals in cigarette smoke, alcohol, or ultraviolet light, while very different, seem to work by a shared underlying mechanism. Are they mutagens, cytotoxicants, or epigenetic agents?

K. Cancer cells, which lack growth control and lack the ability to terminally differentiate or apoptose, are “immortal” and do not have functional GJIC.

One of the generally accepted characteristics of cancer cells is that (a) they lack growth control; (b) they do not terminally differentiate or apoptose under normal conditions; and (c) they appear “immortal.” It is here that one of the well validated cell characteristics of cancer cells (original observation of Dr Werner Loewenstein ⁴⁷ ) is they lack functional GJIC.

Moreover, not cited as a “cancer characteristic” in this new proposed protocol to find the carcinogenic potential of agents, the fact that one of the fundamental cellular processes, needed for metazoan growth control, differentiation, and apoptosis, is GJIC. GJIC can be reversibly inhibited by tumor promoters, including specific congers of poly brominated biphenyls and poly chlorinated biphenyls, ^48,49 in a reversible, threshold fashion in the absence of anti-tumor promoters. Most, if not all, of these “promoting” chemicals reversibly inhibit GJIC at threshold levels. These promoting chemicals can be both natural compounds, such as hormones, ⁵⁰ growth factors, ⁵¹ or cytokines, ⁵² as well as synthetic chemicals. ^44,45 However, it should not be concluded that all chemicals can inhibit GJIC at non-genotoxic and non-cytotoxic levels. Very structurally similar chemical, such as 2-,3-,4-,2’-,3-’,4’-poly brominated biphenyl, blocks GJIC and is a rodent liver tumor promoter, while 3-,4-,5-,3’-,4’-,5’-polybrominated biphenyl is not a rodent tumor promoter at non-cytotoxic concentrations. ^48,49,53,54

L. Chemicals can under different conditions be either oxidants or antioxidants.

While it is beyond the scope of this Commentary, it has to be noted that these “promoting” chemicals can, depending on specific circumstances, be both oxidants (promoters) or anti-oxidants (chemo-preventive agents). ⁵⁵ One of the first hypothesis that oxidative stress was the mechanism behind tumor promotion. ⁵⁶ Retinoids can be anti-promoters under one set of conditions ⁵⁷ or promoters in another. ⁵⁸ Several other environmental toxicants, such as DDT and TCDD, have been shown, which, under most circumstances, work as tumor promoters and, under other experimental circumstances, work as anti-cancer agents. ^59,60

M. Cytotoxicants and natural cytokines released by cell death can be tumor promoters.

Promotion can occur after cell death caused by ultraviolet light (a powerful DNA damaging agent and point mutagen as well as a cytotoxicant at high doses) that would cause inflammatory cytokines to stimulate compensatory hyperplasia after sun burn to skin. Alcohol or chloroform, non-mutagenic cytotoxic agents, can cause cell death and have been classified as “carcinogens.” Most likely, they should be classified as “carcinogenic-indirect endogenous promoters” ⁶¹ and are probably working as inflammatory agents. ^{62 –64}

N. Natural endogenous chemicals, which work by different biochemical mechanisms, can be tumor promoters.

Promotion can be brought about by normal endogenous factors, such as growth factors, ⁵¹ hormones, ⁵⁰ and cytokines/chemokines, ⁵² which can all modulate GJIC. Promoting endogenous and exogenous chemicals is carried out by many different intracellular signaling mechanisms. TPA and DDT are two classic tumor promoters. TPA uses the protein kinase C pathway, while DDT appears to work via alteration in intracellular calcium. ⁶⁵ In addition, Smith et al. ¹ correctly noted two different receptor-dependent mechanisms that can be characteristic to “carcinogens”, estrogen, while it works via an estrogenic receptor as a promoter and mitogen, and, at certain stages of cancer development or supersaturated concentrations, can act via receptor-independent mechanisms to be an anti-mitogen. ^30,66 .

O. Promoters can inhibit GJIC in normal progenitor, differentiated cells, and initiated cells.

While promoters can reversibly inhibit gap junctions in both normal cell and initiated cells, the normal non-stem or progenitor cells will eventually senesce, whereas the initiated cells will proliferate and not apoptose, as both are cell-regulated functions of gap junctions. ⁶⁷

P. Some “carcinogens” can act epigenetically to be both tumor promoters for initiated cells and inducers of differentiation of cancer cells.

Some classic non-mutagenic tumor promoters can block gap junctions, act as epigenetic agents to initiated cells, and yet can induce terminal differentiation of cancer cells. Benzene can induce epigenetically terminal differentiation of human leukemia cells. ^68,69 Because it can effectively induce differentiation, induce oxidative stress, produce free radicals, and even block TPA action, it cannot be characterized as a “mutagen” as has been done before in the literature. It clearly works as an epigenetic agent.

Q. Adult organ stem cells with few mitochondria metabolize via glycolysis.

If one accepts that the organ-specific stem cell is the target cell for the initiation phase, and one accepts the existence of DNA lesions in cells of target tissues after exposures to chemicals that can induce oxidative stress and free radicals, stem cells are undifferentiated and metabolized via glycolysis rather than by oxidative phosphorylation because they have very few mitochondria just opposite to their terminally differentiated derivatives. ^{70 –75} This observation alone should have some utility in any assessment of mechanisms in the multi-stage, multi-mechanism process of carcinogenesis, in that it should put adult organ-specific stem cells as the target cell to be characterized and monitored.

R. Restoration of gap junctional communication can occur by epigenetic regulation of connexin genes.

While cancer chemo-preventive agents either block the inhibition of GJIC by non-genetic, but epigenetically acting tumor promoters, several anti-cancer therapeutic agents, such as SAHA, ⁷⁶ an inhibitor of histone deacetylase, can induce GJIC in non-gap junction expressing cancer cells.

S. The promotion phase is the rate-limiting step of human carcinogenesis.

While preventing the initiation or mutagenic event from occurring is admirable (e.g. don’t sit in the sun too long; do not get exposed to too much ionizing radiation), one can never reduce to zero levels, the mutations from occurring via errors in DNA replication of stimulated cell division in our organ-specific stem cells. On the other hand, aside from childhood cancers, the promotion phase must involve in the exposure of decades of sustained, regular, and threshold concentrations, in the absence of anti-promoters. Consequently, to prevent human cancers, intervention of the tumor promotion phase must be our objective to reduce cancers. Therefore, the major objective of this Commentary is to characterize and identify potential human, organ-specific, gender, and developmental stage agents that can amplify initiated cells by epigenetic mechanisms, which include agents that could induce several intra-signals that can modulate cell–cell communication and transcriptional alter gene expression. While the reference to the Hanahan and Weinberg statement, “…changes in gene expression and cell signaling” is partially correct, it fails to note that these intra-cellular signaling pathways are important, but insufficient, in which these signals can modulate GJIC. In other words, extra-, intra-, and gap junctional inter-cellular communication was evolutionary coordinated to regulate cell proliferation, cell differentiation, and apoptosis in metazoans. ^77,78

T. Understanding cellular, biochemical, and molecular mechanisms is important to epidemiology.

While some epidemiologists might argue that the understanding mechanisms of the three phases of carcinogenesis, especially the tumor promotion phase, it is not necessary to know that the mechanisms by which smoking, drinking too much alcohol, or sitting in the sun too long “cause” cancer, since by stopping smoking, stopping excessive drinking, or wearing sun screen will reduce risks to these cancers. However, to understand that the initiation step occurs in stem cells and that promotion is a rate limiting and preventable or interruptible process of the initiated stem cell, which can occur via many different intracellular biochemical mechanisms (which could involve thresholds, synergism, additivity, or antagonistic interactions to bring about modulations in cell-cell communication), would help to design future epidemiological studies as well as interpret all too often conflicting results of these kinds of studies.

Epidemiology has the potential to identify that there might be an association between an agent and a disease. However, it cannot identify the underlying mechanism or a cause–effect relationship. ⁷⁹ Determining the “cause–effect” link requires taking into consideration the dose–response, the temporal relationship, and biological factors (e.g. species, gender, developmental stage, target cell type, etc.], as well as considering possible confounding factors (e.g. exposures to other potential additive, antagonistic or synergistic agents, consistency, etc.). ⁸⁰ Understanding mechanisms of action can, in addition, add to the power of epidemiological studies in the design and interpretation of the study. ⁸¹ To know a chemical might act as a tumor promoter by acting epigenetically could explain why the same chemical could be a teratogen or a reproductive endocrine disruptor, an atherosclerotic agent, or a fetal neural toxicant as well as being a pharmacologically effective sedative. It could help regulators to understand why a chemical, such as thalidomide, can have both toxic and beneficial potential simply by understanding the mechanism in which it acts under different conditions. ^{82 –85}

Acknowledging that the new 10 carcinogen characteristics approach is a new and empirical strategy, it is argued here that understanding some basic biology of the carcinogenic process and the incorporation of new facts and mechanisms of the three phases of carcinogenesis should bring new light to this approach. While all 10 “characteristics” of “carcinogens” have some validity, review of the literature for these characteristics associated with a chemical will find thousands of papers, in which, because (a) different assays, techniques, and living model systems were used and (b) invalid paradigms of mutagenesis, cell killing, and epigenetic mechanisms in the multi-stage carcinogenesis process influenced interpretation of results, will lead to continuous confusion about how these chemicals work. Several concepts must be incorporated in any system to characterize how agents (radiation, chemicals, and microbes) affect adult organ-specific stem cells by altering cell–cell communication. These comments have been made to be seriously considered in the search of a new paradigm of predicting the potential carcinogenicity and potential toxicity and pathology of the tens of thousands natural and synthetic chemicals to which we are exposed. Therefore, while our goal is still to prevent and treat cancers better than we have in the past (e.g. the new “Precision Medicine” approach and the “Cancer Moon Shot” project), there is sufficient scientific mechanistic knowledge of the multi-stage process, in particular the promotion process, to use existing knowledge in a new manner.

It has been argued that the promotion phase and characteristics and mechanisms of action of “promoters” should be the primary focus to identify those agents which could lead to lowering the risk to human cancer. Of course, knowing if an agent might be mutagenic to the germ or somatic cells is extremely important. However, in the case of many diseases, such as birth defects and cancer, knowing the epigenetic potential is of utmost importance. Understanding the mechanisms of action of these promoting or epigenetic agents, especially during the in utero development of the embryo and fetus, might lead to a major moral objective. ⁸⁶ Since there will be an estimated three billion more babies born by the end of this century, trying to prevent the uterine environment from modulating the number of adult organ-specific stem cells, which might (a) increase the risk of the number of target cells for the initiation of the carcinogenic process later in life (the Barker hypothesis ^87,88 ) and (b) modulate cell–cell communication by epigenetic mechanisms controlling the proliferation, differentiation, and apoptosis of cells during critical periods of development, ⁸⁹ should be our moral imperative. This potential altered biological process during of disrupting adult organ-specific stem cells during development, if modulated by these epigenetic-acting chemicals, might be related to many later in life diseases, such as autism.