Review Article

Molecular Epidemiologic Markers:
A New Concept in the Preventive Medicine with Special Attention to the Prevention of Cancer

István Ember1, Árpád Németh1, Csaba Varga1, Pál Perjési2, István Arany3, Katalin Fehér1, Katalin Németh1, Zsuzsanna Dombi1, and István Kiss1

1 Department of Public Health, Medical School, University of Pécs, Pécs, Hungary
2 Department of Biochemistry, Medical School, University of Pécs, Pécs, Hungary
3 Department of Internal Medicine, Division of Nephrology, University of Arkansas for Medical Sciences, Little Rock, AR, USA

Corresponding author: Prof. István Ember
    Department of Public Health
    Medical School
    University of Pécs
    Szigeti u. 12
    H-7643 Pécs, Hungary
    Telephone: Phone: +36-72-536395
    E-mail: istvan.ember@aok.pte.hu

CEJOEM 2005, Vol.11. No.1.: 3–15


Key words:
Early phase of carcinogenesis, gene alterations, biomarker, chemopreventive substances, risk assessment, primary prevention


Abstract:
In early diagnosis of cancer (secondary prevention), there was no real breakthrough, although amplification and/or overexpression of onco-suppressor genes had been often observed in routine pathological studies. Numerous genetic alterations could be associated with the early phase of carcinogenesis. Examination of these alterations, immediately after the exposure, could give us the opportunity to use them as biomarkers. A relationship was proved between the early genetic alterations caused by the carcinogens and tumor development. The effects were measured in “short-term” animal model in thymus, spleen, lymph nodes, bone marrow, liver, kidneys and lungs at 24, 48 and 72 hours after the treatment. Considering individual susceptibility and polymorphism of carcinogen metabolizing enzymes, we can apply these biomarkers to identify the “high risk” individuals and groups of the population and this method could be suitable to decrease cancer morbidity, follow up the effectiveness of therapy and chemopreventive agents.



INTRODUCTION

Despite the revolutionary discoveries in cancer research of the last decade, as new information about onco-suppressor gene cascades, communication between cells, detection of apoptosis (Wild et al., 2002), there has been little success to find specific, diagnostic, pathological and morphological alterations that would be characteristic of each tumor type (Balmain et al., 2003). The advances in molecular genetics have led to the hypothesis that tumors are basically “genetic diseases”, mainly induced by environmental factors (Minamoto et al., 1999). This is true primarily for sporadic tumors, constituting 90% of all tumors. Environmental chemical factors are responsible for 90% of these tumors (Knudson, 2002).
     It is known that ectopic activation of onco- and suppressor genes by environmental chemical factors leads to development of tumors (Todd and Wong, 1999). However, this activation is complicated and interrelated that requires multi-component systems. The detectable alterations in onco/suppressor genes and related pathways are not tumor specific and they are often caused by various environmental carcinogens, i.e., chemical, physical, and biological factors. On the other hand, they can serve as a basis to develop more effective therapy and achieve greater ability to differentiate tumor types in the diagnosis. In early diagnosis (secondary prevention) as yet no real breakthrough occurred, even though amplification and/or overexpression of onco/suppressor genes had been often observed in routine pathological studies. However, these findings have never been employed at the early diagnostic period, they have been applied in advanced stages of tumorigenesis rather. From the point of view of prevention, a real achievement would be to introduce the application of these early alterations to the premorbid diagnostics of tumorigenesis. Unfortunately, at this time it seems impossible due to the lack of knowledge of tumor-specific changes. Consequently, the application of these genetic changes in tumor pathology and early diagnostics faces limitations, not only in the secondary prevention of sporadic tumors but also in their early diagnosis and screening. Nevertheless, this body of knowledge is effectively used in the field of tertiary prevention, i.e., in the diagnosis of tumor relapses and follow-up of their possible remission, refinement of individually designed therapeutic regimens, and in molecular and predictive clinical epidemiology (Hawk et al., 2000). On the other hand, in tertiary prevention the genetic approach has seemed less effective to the early detection of metastases, despite the numerous promising signs. With familial tumors it is not the case, because the expression of certain oncogenes is predictable, e.g., that of BRCA1 in breast cancer. The molecular method (assessment of onco/suppressor genes), therefore, could be quite effective both in primary and secondary prevention of tumors (Owen, 2001; Garcea et al., 2003), however, with methodological and ethical limitations.
     In the multi-step model of carcinogenesis, the chance to apply these markers is confined due to their nonspecific nature (Perera, 1995), even though the general sequential steps of the tumorigenetic process are known (Balmain et al., 2003). Thus, by applying step-by-step models we have the opportunity to identify high-risk groups but the early diagnosis of environmental cancers is far away. A good example is the sequential predictive model (Fearon and Vogelstein, 1990; Fig. 1), or other related models (Bellacosa, 2003).


Fig. 1. Molecular sequential model of cancer formation (Fearon and Vogelstein, 1990)


EVOLUTION OF MOLECULAR-EPIDEMIOLOGOCAL EVIDENCE
AND OUR RESEARCH DIRECTIONS

Numerous genetic alterations could be associated with the early phase of carcinogenesis. However, alterations in the incubation period are describable only by means of molecular biological methods (Brambilla et al., 2003). It means opportunity to assess cancer risk both at the individual and group levels (Ember et al. 2002). Examining these alterations immediately after exposure could give us the opportunity to use them for risk assessment, as early biomarkers. It is worthwhile to emphasize that they are unrelated to the classical tumor markers (Wild and Turner, 2001). Moreover, during the last 20 years, it has become clear that the models of multistep carcinogenesis (today molecular carcinogenesis) might vary. Earlier they emphasized the importance of the sequential steps. Today it is believed that certain somatic (oncogenic) alterations are interrelated and appear at a stage that depends on individual variability (Knudson, 2000). However, sometimes the morphological counterparts of a given state could provide a basis to find the adequate biomarkers (Ember et al., 2002).
     In this context, we cannot speak about diagnostic methods, but rather risk estimation and analysis. Fig. 2 presents the pathogenesis taken from the first exposure to the actual disease. This scheme can also be useful in evaluating other diseases beside tumors. It has become clear that before the morphological/clinical manifestations of tumors there are gene alterations and the analysis and evaluation of these alterations would be highly informative on the actual condition or status of tumor risk among patients or various groups. The validity of preventive markers implies the need of parameters that differ from those used for diagnosis (Brambilla et al., 2003). The term prevention is defined in Fig. 3, which also includes the early stages of carcinogenesis and molecular biological alterations (markers). At this point of the tumorigenetic process, i.e., before morphological changes, the fields of prevention, epidemiology, and pathology (using similar molecular biological methods) have historically pursued separate approaches.


Fig. 2. Cancer formation


     
Tumor research, the big expectation of the 80’s, prior to breakthroughs by molecular biology, employed many nonspecific data. Today the use of DNA chip techniques also yields nonspecific data. The simultaneous measurement of thousands of genes can be useful for clinicians to assess better the condition/status of the patient, its response to therapy, but it is less informative for early and more specific diagnosis. It can be said that DNA chip microarray technology offers the possibility of broad quantitative risk assessment, a concept to which less investment is required than to the present methods. The survey of somatic status is not curative, but as a part of preventive/predictive medicine, the disease manifestation could be prevented by risk group survey. Still, it is clear that preventive measures are now based on more solid scientific data than previously. Fig. 4 presents the role of biological methods (i.e., onco/suppressor gene examination) that give the opportunity of prevention and quantitative risk estimation as a part of molecular and predictive epidemiology. These measures are not diagnostic elements and naturally conform to different criteria (they are nonspecific and approximate), therefore we interpret them as molecular epidemiological biomarkers.


Fig. 3. Preventive molecular epidemiology


Fig. 4. Expression of onco/suppressor genes (H-ras, p53, c-myc), as early biomarkers of chemical carcinogenesis


     
Before the discovery of onco- and suppressor genes, biotransformation by various adducts, carcinogens, and events at the chromosomal level represented the markers that were used (Knudson, 2000; Wild and Turner, 2001; Ember et al., 2002). Now when individual susceptibility, e.g., polymorphism of enzymes metabolizing various carcinogens is considered (Groopman et al., 1996; Vlastos et al., 2003), we could apply molecular epidemiological biomarkers to identify individuals or groups at risk of carcinogenesis, to decrease cancer morbidity and mortality, and to follow-up the effectiveness of the therapy. By means of these tools we have the possibility to monitor early molecular events of carcinogenesis at the risk assessment level. At the same time we have a chance to assign the way of intervention into these early events (primary prevention before the manifestation of tumors) and measure its effects with biomarkers.
     Chemopreventive materials are synthetic or naturally occurring substances – mainly of plant and rarely of animal origin – which exert their effect before the morphological detection of the tumor is possible (Bartsch, 1999; Fenech, 2002; Visakorpi, 2003). Based on their chemical, physical, or biological characteristics, their application is collectively termed chemoprophylaxis in analogy with the successful prevention of pox and other infective diseases. Their mechanism of action is different and covers a broad range, including inhibition of chemical binding to macromolecules, modulation of biotransformation, and altering the somatic mutation and repair system, etc. Their main action is inhibition of carcinogenesis (anti-initiators, antioxidants, anti-promoters, etc). We presented an animal model that not only meets the criteria of biomarkers, detection of chemical carcinogen exposure, and early biological effects (Fig. 5), but also other preconditions. This system would satisfy many other functional aspects as well. Essentially, this approach applies a mouse strain, which is especially sensitive to the effect of chemical carcinogens and which responds to their effects by overexpressing certain key onco- and suppressor genes (Ha-ras, c-myc, and p53) soon after exposure (within 24–72 hours) (Hong and Lippman, 1995; Warren and Shields, 1997; Howell at al., 2002). The model signals carcinogenic effects at an early phase, which otherwise could only be detected in long-term studies (Shureiqi et al., 2000; Turini and Dubois, 2002). On the one hand, it is capable of signaling exposure, the early effect of the examined chemical, physical, or biological carcinogens as a part of primary prevention. On the other hand, it gives an opportunity to quantitatively assess the risk of carcinogenesis in human populations. By this approach, we can develop a complex molecular epidemiological marker system (Fig. 6), which could provide information on tumorigenic factors endangering humans with potential biological effects which could give the opportunity of primary prevention (chemoprophylaxis, hygienic actions, immunological and genetic interventions) and thereby to formulate an integral part of primary preventive systems in molecular and preventive epidemiology (Ember et al., 2002). In addition, this method is capable to measure and follow-up the effectiveness of applied chemoprophylactic interventions at the level of genes (Gyöngyi et al., 2001b). The model is a “short-term” sensitive animal strain (CbA/Ca/H-2k) model for determining carcinogenicíty by monitoring gene alterations (Ember et al., 1999; Ember et al., 2000). This model proved a relationship between the early genetic alterations (gene amplification, gene expression and mutation) induced by carcinogens and tumor development in the thymus, spleen, lymph nodes, bone marrow, liver, kidneys, and lungs at 24, 48, and 72 hours after the treatment with the carcinogens. From these organs and the white blood cells of the peripheral blood RNAs were isolated. The expression of three genes (H-ras, c-myc, and p53) was investigated. H-ras, with its key role in chemical carcinogenesis, is known to play a significant role in signal transduction, while p53 is a tumor suppressor gene involved in the regulation of apoptosis, DNA repair, and cell-cycle. The gene c-myc exerts its effect in the proliferation and immortalization processes. The early alterations of these three genes detected at three time points predict the biological effects of carcinogens (Gyöngyi et al., 2001a). The model signalizes if any effect of a chemopreventive agent applied and the stage of tumorigenesis in which the intervention has occurred. The investigation of various chemical carcinogens in this model is not only an important primary preventive tool, but due to consistent early biological effects it can also be used to monitor exposure. It is necessary to note that as a biomarker it can be used for risk estimation but not to establish diagnosis.


Fig. 5. A biomarker test system in an animal model


Fig. 6. Complex risk assessment model


     
The above mentioned method has predictive value, since it indicates the fate of affected animals or individuals by examining gene expression and amplification. By means of the Kaplan-Mayer survival curves, the model offers the opportunity to measure the effectiveness of individually designed chemotherapy protocols (chemoprevention) together with the follow-up stages of the disease and metastasis. The significance of the animal model experiment allows one to test chemopreventive agents. However, early tests of carcinogenesis before the manifestation of tumors are not able to test chemotherapeutical effectiveness. For the latter, long-term experiments are necessary along with in vitro models. To our knowledge, this is the first in vivo short-term model particularly utilizing predictive molecular biomarkers.
     With the application of chemopreventive agents, we realized that a significant decrease in oncogene expression could be reached by treatment with these chemicals. Examining many potential chemopreventive agents, we concluded that the model is capable of testing the effects of potential chemopreventive agents, too, even those of different mechanisms of action. Our investigations showed that the model was able to detect chemopreventive effects as efficiently as they had been detected in long-term models. In other words, this model can be considered a chemopreventive testing model that enables us to investigate the initiation and all steps of the carcinogenic effect as shown in the case of the carcinogen (7,12-dimethylbenz[a]anthracene) (DMBA) (Ember et al., 1998b).
     First we wanted to rely on the early diagnosis and the screening pathway. However, as tumor-specific genetic alterations were lacking, the animal experiments suggested us to direct our efforts to identify susceptible and exposed individuals by applying early biological markers and quantifying risk assessment by means of their molecular-predictive character. Thus, a way was found from secondary to primary prevention, with due attention to the ethical and technical limitations and the fact that the orthodox therapeutic medicine cannot give proper answers to molecular medicine in terms of pharmacological development or in effectiveness of therapeutic interventions.
     For this reason, in the Western European countries ethical limitations to genetically based (somatic) diagnostic measures have been raised (Ember et al., 2000; Vahakangas, 2004). For quantitative risk assessment, molecular epidemiological developments on animal-based experiments and further human epidemiological studies are equally required. In the new animal model, we found gene alterations (biomarkers) that can also be used as early approximative biomarkers not only in epidemiology but also in chemopreventive models as a part of the primary prevention complex extrapolated to humans (Ember et al, 1998a). This so far has been achieved in relation to risk assessment. Its capability to analyze the effectiveness of interventions is, however, a task yet to be achieved. In spite all of these considerations, a high number of statistical studies are required, mainly due to the individual variability of onco-suppressor genes, to enable us to use the model at a population level. At the same time, we should not forget to apply individual gene polymorphism (enzyme polymorphism) to risk assessment.


CONCLUSIONS

In summary, our suggested model aptly applies to epidemiology, prevention, and prediction and can be used as an integral part of primary tumor prevention. Our mouse model can also be used as a complex model (exposure, early risk assessment) and it can play a role in testing the effectiveness of chemoprophylaxis. However, in the absence of better therapies, chemoprevention is a natural, effective tool. In accordance with the preventive (chemopreventive) methods, this model provides information in the “silent” period of chemical carcinogenesis, i.e., when no morphological and biochemical signs are present. At this stage, chemopreventives cannot be investigated. Therefore, the ability to test chemopreventives would help in formulating interventional and preventive strategies with an estimated effectivity of 30–35%, which has not been met by other known means to decrease morbidity and mortality. For this purpose, we developed our molecular biological, genetic, and mathematical model of high scientific and experimental background. In order to measure their effects, we investigated numerous potential chemopreventive substances, e.g., we studied the changes in oncogene expression caused by the known carcinogenic DMBA, as a positive control. We did not expect to measure the therapeutic effects in this early induction model, since we were far from the point of tumor development, but we examined substances that inhibited overexpression of certain oncogenes. We concluded that for measuring the effects of substances in this period, we do not have any other possibility but using molecular epidemiological markers in in vivo “short-term” or classical “long-term” induction models.
     Our model seems to be suitable for testing preventive substances because it utilizes biomarkers, which are more potent and useful, at this early stage, than the other late markers. Therefore this animal model is applicable as a molecular epidemiological predictive test for pre-signaling the effects of a chemical carcinogen and protection against it as well. It is fast, cheap, and interpretable. According to the literature, no model like this – capable of monitoring both the individual steps of tumorigenesis and the effects of interventions undertaken at those stages – is available. Therefore, our model is suitable to test the strategy of potential preventive substances, which exert their effect in the early stage of tumorigenesis while controlling the effectiveness of interventions.
     In summary, by establishing this animal model we have a molecular, epidemiological and preventive marker system capable of indicating the risk on the one hand, and serving as a model for testing substances with preventive characters, on the other. The next step would be extrapolation of the model to humans.


ACKNOWLEDGEMENT

The present work was supported by grants of the National Fund of Scientific Research (OTKA No.: T037279), and the Council of Health Sciences (ETT No.: 323/2003) of Hungary.
     We are grateful to Ms. Mardelle Susman (Department of Microbiology and Immunology, UTMB, Galveston, TX) for reviewing this manuscript.


REFERENCES

BALMAIN, A., GRAY, J., and PONDER, B. (2003). “The genetics and genomics of cancer.” Nature Genet. 33/Suppl: 238–244.

BARTSCH, H. (1999). “Keynote address: exocyclic adducts as new risk markers for DNA damage in man.” IARC Sci. Publ. 150:1–16.

BELLACOSA, A. (2003). “Genetic hits and mutation rate in colorectal tumorigenesis: versatility of Knudson’s theory and implications for cancer prevention.” Genes Chromosomes Cancer 38: 382–388.

BRAMBILLA, C., FIEVET, F., JEANMART, M., DE FRAIPONT, F., LANTUEJOUL, S., FRAPPAT, V., FERRETTI, G., BRICHON, P. Y., and MORO-SIBILOT, D. (2003). “Early detection of lung cancer: role of biomarkers.” Eur. Respir. J. 39/Suppl.: 36s–44s.

EMBER, I., GYÖNGYI, Z., KISS, I., GHODRATOLLAH, N., and ARANY, I. (2002). “The possible relationship between onco/suppressor gene expression and carcinogen exposure in vivo: evaluation of a potential biomarker in preventive and predictive medicine.” Anticancer Res. 22: 2109–2116.

EMBER, I., KISS, I., and FALUHELYI, ZS. (1998a). “Gene expression changes as potential biomarkers of tumour bearing status in humans.” Eur. J. Cancer Prev. 7:347–348.

EMBER, I., KISS, I., and NOWRASTEH, G. (1999). “Different H2 haplotypes have a strong influence on oncogene action.” Anticancer Res. 19:1181–1185.

EMBER, I., KISS, I., and PUSZTAI, ZS. (1998b). “Effect of 7,12-Dimethylbenz(a)-antracene on onco/suppressor gene action in vivo, short term experiment.” Anticancer Res. 18:445–448.

EMBER, I., PUSZTAI, Z., GYÖNGYI, Z., and KISS, I. (2000). “1-Nitropyrene induces elevated expression of oncogenes and tumor suppressor genes 24 hours after treatment in CBA/Ca mice.” Anticancer Res. 20:1563–1566.

FEARON, E. R. and VOGELSTEIN, B. (1990). “A genetic model for colorectal tumorigenesis.” Cell 61:759–767.

FENECH, M. (2002). “Biomarkers of genetic damage for cancer epidemiology.” Toxicology 27: 411–416.

FRAPPAT, V., FERRETTI, G., BRICHON, P. Y., and MORO-SIBILOT, D. (2003). “Early detection of lung cancer: role of biomarkers.” Eur. Respir. J. 39/Suppl.: 36s-44s.

GARCEA, G., SHARMA, R. A., DENNISON, A., STEWARD, W. P., GESCHER, A., and BERRY, D. P. (2003). “Molecular biomarkers of colorectal carcinogenesis and their role in surveillance and early intervention.” Eur. J. Cancer 39:1041–1052.

GROOPMAN, J. D., WANG, J. S., and SCHOLL, P. (1996). “Molecular biomarkers for aflatoxins: from adducts to gene mutations to human liver cancer.” Can. J. Physiol. Pharmacol. 74:203–209.

GYÖNGYI, Z., EMBER, I., KISS, I., and VARGA, CS. (2001a). “Changes in expression of onco- and suppressor genes in peripheral leukocytes – as potential biomarkers of chemical carcinogenesis.” Anticancer Res. 21:3377–3380.

GYÖNGYI, Z., NÁDASI, E., VARGA, C., KISS, I., and EMBER, I. (2001b). “Long-term effects of 1-nitropyrene on oncogene and tumor suppressor gene expression.” Anticancer Res. 21:3937–3940.

HAWK, E., VINER, J. L., and LAWRENCE, J. A. (2000). “Biomarkers as surrogates for cancer development.” Curr. Oncol. Rep. 2:242–250.

HONG, W. K. and LIPPMAN, S. M. (1995). “Cancer chemoprevention.” J. Natl. Cancer Inst. Monogr. 17:49–53.

HOWELL, W. M., CALDER, P. C., and GRIMBLE, R. F. (2002). “Gene polymorphisms, inflammatory diseases and cancer.” Proc. Nutr. Soc. 61:447–456.

KNUDSON, A. G. (2000). “Chasing the cancer demon.” Ann. Rev. Genet. 34:1–19.

KNUDSON, A. G. (2002). “Cancer genetics.” Am. J. Med. Genet. 22:96–102.

MINAMOTO, T., MAI, M., and RONAI, Z. (1999). “Environmental factors as regulators and effectors of multistep carcinogenesis.” Carcinogenesis 20:519–527.

OWEN, R. W. (2001). “Biomarkers in colorectal cancer.” IARC Sci. Publ. 154:101–111.

PERERA, F. P. (1995). “Molecular epidemiology and prevention of cancer.” Environ. Health Perspect. 103/Suppl: 233–236.

SHUREIQI, I., REDDY, P., and BRENNER, D. E. (2000). “Chemoprevention: general perspective.” Crit. Rev. Oncol. Hematol. 33:157–167.

TODD, R. and WONG, D. T. (1999). “Oncogenes.” Anticancer Res. 19:4729–4746.

TURINI, M. E. and DUBOIS, R. N. (2002). “Primary prevention: phytoprevention and chemoprevention of colorectal cancer.” Hematol. Oncol. Clin. North Am. 16:811–840.

VAHAKANGAS, K. (2004). “Ethical aspects of molecular epidemiology of cancer.” Carcinogenesis 25:465–471.

VISAKORPI, T. (2003). “The molecular genetics of prostate cancer.” Urology 62/Suppl 1: 3–10.

VLASTOS, A. T., SCHOTTENFELD, D., and FOLLEN, M. (2003). “Biomarkers and their use in cervical cancer chemoprevention.” Crit. Rev. Oncol. Hematol. 46:261–273.

WARREN, A. J. and SHIELDS, P. G. (1997). “Molecular epidemiology: carcinogen-DNA adducts and genetic susceptibility.” Proc. Soc. Exp. Biol. Med. 216:172–180.

WILD, C. P., LAW, G. R., and ROMAN, E. (2002). “Molecular epidemiology and cancer: promising areas for future research in the post-genomic era.” Mutat. Res. 29:3–12.

WILD, C. P. and TURNER, P. C. (2001). “Exposure biomarkers in chemoprevention studies of liver cancer.” IARC Sci. Publ. 154:215–222.


Received: 2 February 2005
Accepted: 20 May 2005

| Back |