Several molecular genetic studies have been conducted by our laboratory staff over the past year. A collaboration with epidemiologists at the NCI continues to study the interaction between molecular and classical risk factors of oral cancer. We are using biological samples already collected in a case-control study of oral and pharyngeal cancer conducted in Puerto Rico. The study focused on Puerto Rico because of the especially high incidence of oral cancer known to occur there. Our analyses thus far have focused on applying highly sensitive techniques for detection of human papilloma virus using polymerase chain reaction methods. We extracted DNA from 150 paraffin embedded sections of tumor tissue. Preliminary results indicate that this virus is infrequently present in these oral cancers from Puerto Rico. We have also characterized a polymorphism at the aryl hydrocarbon receptor, a key mediator of the detoxification pathways involving P450 cytochrome. Statistical analyses suggest that this molecular variant may interact with exposure to risk factors such as smoking and drinking in influencing risk of oral cancer. We have been developing plans and protocols for performing molecular analyses of several additional DNA polymorphisms that have been shown to be associated with risk of other forms of cancer. Statistical analyses will be performed for candidate genes using this collection of study subjects. Molecular analyses of p53, HPV and other appropriate markers such as those in regions known to be associated with Loss of Heterozygosity (LOH) will also be conducted on available tumor tissues.A second molecular epidemiological study of oral cancer is being conducted in Athens, Greece, in collaboration with Dr. Anthanasios Zavras, a dentist and Ph.D. student in the Department of Oral Health Policy and Epidemiology at the School of Dental Medicine, Harvard University. We are using a case-control design, based on recruitment in hospitals in Athens. Our laboratory?s role is to perform the molecular and statistical analyses for this study, as well as provide guidance about the study?s design and performance. Risk factors, family histories of cancer and biological samples (oral rinses and brushings, tumor biopsy or surgical specimens (if available) and blood spots on filter paper) are being delivered to our laboratory for patients and controls. We have begun to evaluate genes potentially responsible for oral cancer, such as those of the cytochrome P450 pathway and in a preliminary analysis found a lack of association with the CYP1A1 gene in this population. In addition to studies of DNA polymorphisms at many other genes involved in detoxification of carcinogens, we plan to also evaluate the role of human papilloma virus, including interaction with variation at HLA genes. Molecular analyses will be conducted to assess mutations in oncogenes and tumor suppresser genes in tumor tissues, and these results will be correlated with the subjects? exposure to risk factors such as smoking and drinking as well as protective factors including the consumption of fresh fruits and vegetables. Statistical analyses will be performed for candidate genes using this collection of study subjects. Molecular analyses of p53, HPV and other appropriate markers such as those in regions known to be associated with Loss of Heterozygosity (LOH) is being conducted on available tumor tissues. In Taiwan, we are collecting multiplex families, and also over 100 simplex families with both parents and at least one unaffected sibling of the OC case available for sampling. Simplex families such as these usually involve OC cases with relatively early age of disease onset (since parents are still alive) and might, therefore, been riched for high genetic susceptibility as has been the case for early onset for ms of other cancers. In addition, such sampling designs are suitable for family-based transmission disequilibrium gene mapping approaches which may provide some advantages over traditional case-control designs. As described above for NPC, we are recruiting OC families in Taiwan using both the Cancer Registry (4,238 letters sent, responses received from 762 or 18%) and hospital clinics. In the course of this family ascertainment effort, we have also distributed 753 segregation analysis questionnaires for OC, and have received 429 (57%) completed. Analyses of these data will compliment our gene mapping studies. Biospecimens from 448 subjects are currently being processed in Taiwan, and risk factor and family history questionnaires are being coded for key entry. Sections from paraffin-embedded tumor tissue are being obtained from approximately half of these OC cases for molecular analyses. A new collaboration with Dr. Rolando Herrero of the International Agency for Research on Cancer (IARC) will provide our laboratory with access to the biospecimens and risk factor data from the ongoing IARC Multinational Oral Cancer Study. With over 1,200 cases and 1,200 controls already collected from countries around the world, and with recruitment ongoing, this resource will provide unprecedented power for genetic epidemiology research for this complex disease. Our laboratory will be responsible for DNA extraction from buffy coat samples, analyses of SNPs in cancer susceptibility and nicotine/alcohol addiction candidate genes, and statistical analyses in partnership with IARC investigators. Analyses of a SNP in the ADH3 gene suggest an effect in our Greek sample that is in the opposite direction of that reported previously in the Puerto Rico OC Study. This example illustrates how unstable findings of SNP associations have so often been, with different studies either failing to replicate or even suggesting associations in opposite directions. We also recently found that a reported association of smoking behavior with SNPs in the CYP2A6 gene may be an artifact of a faulty laboratory assay. Our manuscript raises concerns about the future credibility of gene mapping approaches for complex diseases unless more caution is applied to their interpretation and reporting. Since we believe these approaches truly do have great potential, our goal is to ensure that our own work is carried out with high standards of quality for laboratory and risk factor data, statistical analysis and cautious interpretation. We believe it will be especially important that adequate numbers of subjects are employed in future studies, to minimize the pursuit of too many false positive findings that turn out to be artifacts of small samples. This is why we have devoted so much effort to building adequate research resources. This program also incorporates the former research project number DE00654-05, Genetic Epidemilogical Studies of Nasopharyngeal Cancer in China. Summary follows. Nasopharyngeal carcinoma (NPC) occurs much more commonly in southern China and Taiwan than in other parts of the world. Epstein-Barr virus is suspected to play a role in disease etiology, but supporting data are quite limited. The disease exhibits clear familial aggregation, and our previously reported segregation analyses suggest that a recessively transmitted major gene may be involved. Our current studies are aimed at determining the basic causes of this disease using molecular epidemiological approaches. We have obtained questionnaire data regarding the family history of this disease and correlation with risk factors from a large number of subjects. As of February 1999, we have obtained DNA, serum, oral rinses, oral brushes and saliva for 1,003 subjects from 179 families with two or more members affected by NPC. We also obtained 1,576 risk factor questionnaires from our multiplex family members. Epstein Barr Virus (EBV) anitbody titres are being assessed on serum samples from all study subjects. A whole genome scan has been completed for approximately 767 DNA samples using highly informative Short Tandem Repeat (STR) markers of the Applied Biosystems Version 2.0 set. Laboratory assays of candidate gene SNPs are also ongoing. We used fluorescently labeled oligonucleotide primers incorporated during PCR. Products were sized by denaturing PAGE on Applied Biosystems (Perkin Elmer) Model 373 and 377 automated DNA analyzers. GENESCAN and GENOTYPER computer programs (Applied Biosystems) were used to identify microsatellite alleles from the gel images. Statistical analyses are currently being performed on these data using methods thought to be most powerful and robust for complex diseases. Some analyses are general and are being applied to all complex diseases under study in the Branch. Other analyses involve novel approaches specific for NPC such as using EBV titres as covariates for predicting NPC risk. Transmission disequilibrium tests (TDTs) are being performed on the candidate genes being assayed including SNPs at HLA loci and CYP2E1 that have been previously suggested to be associated with risk of NPC. Extensive questionnaire data collected for all participating subjects are being evaluated to identify the most important environmental risk factors for this disease. Statistical methods which account for relationships among family members will be used for these analyses. EBV antibody titres are being analyzed as a dependent variable of direct interest for understanding the genetic basis of host defense. Gene mapping studies of these immune response phenotypes have required minimal additional cost, since biospecimens and substantial numbers of genotypes are already available for these subjects. Sections from paraffin-embedded tumor tissue are being obtained from approximately half of these NPC cases for molecular analyses.
| Zavras, A I; Douglass, C W; Joshipura, K et al. (2001) Smoking and alcohol in the etiology of oral cancer: gender-specific risk profiles in the south of Greece. Oral Oncol 37:28-35 |
| Winn, D M; Diehl, S R; Brown, L M et al. (2001) Mouthwash in the etiology of oral cancer in Puerto Rico. Cancer Causes Control 12:419-29 |
