Prof. William G Thilly

The work of the laboratory encompasses the study of genetic change in humans, both germinal and somatic changes causing or increasing the age specific probability of disease. We begin with a mathematical analysis of the age specific mortality rates for human disease which we have collected and computed for individual birth cohorts using data from 1900 forward. With these data and quantitative models of time dependent accumulation of somatic mutations we have been able to calculate the fraction of the American population at risk of death by diseases such as cancer at any site, cardiovascular or cerebrovascular disease, diabetes and several others. Our basic data set is available at http://cehs4.mit.edu.ezproxy.canberra.edu.au These studies have shown us that for the most prevalent human cancers unconnected with smoking, those of the colon, breast and prostate have shown no change in the population at risk from birth cohorts of the mid-nineteenth century to the present day. Other cancers such as leukemia, lymphoma and brain cancer have shown marked increases apparently unconnected to cigarette use. Stomach cancer has shown a precipitate decline in risk fraction. We interpret these data to demonstrate a clear environmental risk condition for cancers for which the risk has changed for the birthyear cohorts examined, roughly 1840 -1950. Thilly's approach also shows that the risk of lung cancer has risen as a simple linear function of cigarette usage with the same relationship observed for both males and females.

Further analyses of these data indicated that the parameters of somatic mutation rates and adenoma growth rates were historically constant or nearly constant even while the fraction of risk for certain cancers, including lung cancer significantly increased or decreased. This in turn suggests that environmental determinants of risk for these diseases are not effecting mutation rates or cell turnover or growth rates in normal tissues or adenomas.

These hypotheses derived from analyses of the public health record appear to be consistent with direct measurements made in the laboratory in experimental animals and humans. In studies of mitochondrial DNA mutations in cultured human cells and various human organs, an essentially identical pattern of point mutations was discovered. When these studies were extended to the mitochondrial DNA of bronchial epithelial cells of identical twins discordant for cigarette smoking, the number and patterns of mutation were identical. This near identity led to the conclusion that mitochondrial point mutations are spontaneous in origin and not induced by exogenous mutagenic chemicals despite their presence in the environment. [It has since been found that a subset of these mutations are created when human mitochondrial DNA polymerase is used to copy the mitochondrial DNA sequence studied.]

In nuclear DNA, the work is extended by direct comparison of smokers and nonsmokers using assays for point mutations at six separate positions in the H-RAS, TP53 and HPRT genes. First results indicate that the increase in mutant fractions in smokers is less than twofold, a value consistent with that derived from analysis of mortality data for lung cancer in birth cohorts which did or did not have access to cigarettes. Furthermore, it was discovered that a significant subset of human inherited and somatic point mutations [in the HPRT gene] are created by treating human cells with a concentration of a chemical which induces formation of 106 methyl guanine in DNA at levels found in normal human tissues. One interpretation is that spontaneous formation of this intermediate accounts for about a quarter of human base pair substitutions which account for about 3/4 of all human inherited disease.

The technology for performing the mutation assays in human cells and organs was developed in the laboratory in the past decade. Much of it is based on tour development of constant denaturing capillary electrophoresis (CDCE) which permits the isolation of measure and sequence mutants more than 95% of the human genome with a sensitivity of 2 x 106. CDCE technology itself is based on application of statistical mechanics to calculation of equilibrium melting temperatures for DNA duplex sequences; its application permits separation of mutant from nonmutant sequences.

More recently, Thilly's lab has extended the use of CDCE to the determination of the number and kind of point mutations in human populations. These inherited mutations are a mixture of a vast majority which have no effect on reproductive capacity (nondeleterious) or longevity (nonharmful) and a small faction which are deleterious, harmful or both. We have already demonstrated the discovery of inherited point mutations in mixtures of blood samples from 5000 juveniles at levels below 103 and have discovered some alleles at these low frequencies which are elevated in African American populations. Thilly has proposed a means to study human populations in such a way that deleterious alleles and harmful alleles may be readily identified within demographically distinct groups.

Current Challenges
Studies of organs and tissues to identify causes of cancer remain the focus of the research in Thilly's laboratory. He is currently trying to create a successor Center to the CEHS to house this research, and welcomes contact from companies interested in sponsoring research on environmental causes of cancer and the cellular processes leading to cancer.

His work on epidemiology and the assembly of the database has been his own project, not funded at MIT. He would like to seek some funding to support its expansion.