Results of segregation analyses support the possibility that risk of Early Onset Periodontitis (EOP) may be due to a single major gene. Alternatively, several genes of moderate effect may contribute significantly to increasing risk of the disease. Clinical subject recruitment has continued to be through long-standing collaborations with periodontists in Virginia. We conducted linkage analyses spanning the entire human genome to evaluate these hypotheses. Families with two or more close relatives affected by EOP were ascertained in the state of Virginia, and the country of Chile. DNA was extracted from blood and highly polymorphic markers distributed throughout the human genome were typed using the polymerase chain reaction. Linkage analyses were performed using a dominant model of disease transmission which is most strongly supported by the segregation analysis studies, as well as additional analyses based on assumptions of alternative modes of disease gene transmission. We also applied nonparametric methods of inferring linkage and/or linkage disequilibrium which do not require assumptions about the disease's genetic architecture. We evaluated African-American families and Caucasian families, both separately and combined, and also evaluated LJP and G-EOP subtypes of EOP both separately and combined for all analyses. Dr. David Duffy of Queensland Institute of Medical Research in Australia contributed substantially to these statistical analyses. For our genome scan for EOP susceptibility we used the non-parametric SIBPAL program to assess evidence of linkage. We performed LOD score analyses using dominant, recessive and intermediate modes of transmission using the FASTLINK program. We obtained six regions with LOD scores above 2.0. We recognize that as a consequence of performing so many multiple comparisons in a genome search we expect to obtain several such LOD scores by chance alone. In fact, current analyses suggest that one would expect to obtain a LOD score of 3.3 about 5% of the time in a genome scan of a complex disease even in the absence of any true susceptibility loci. Thus this threshold constitutes the standard type I error criterion. In our study, we also performed multiple analyses based on different disease definitions, modes of inheritance and racial groups, so additional adjustments need to be made. However, we often found results to be highly correlated among these alternative analyses, so a simple Bonferroni method would be overly conservative. Our strongest finding is on chromosome 19 with a multipoint LOD score of 3.72 (Figure 1). Our second highest multipoint LOD score 2.90 is for a region of chromosome 8 implicated for both LJP and G-EOP susceptibility. We also conducted follow-up studies of our finding of a strong association of IL-1 polymorphisms with generalized EOP and obtained supportive results in an independent case-control design data set. To follow up on these findings, we sought to develop alternative quantitative measures of risk of this disease. Using multivariate statistical procedures such as factor analysis and discriminant function analysis, we identified independent, continuous alternative measures which classify individual subjects according to levels of severity of EOP disease. We simultaneously developed automated computer software for performing genome-wide analyses of such quantitative data and will complete these statistical analyses in the coming year. It is well-established that quantitative measures can be much more powerful for gene mapping analyses. Our goal is to re-evaluate our mapping results using these alternative measures to see if greater support can be obtained for some of the regions identified above using the discrete, binary classification of subjects as affected or unaffected. During the past year we developed a new initiative aimed at identifying genetic and environmental factors underlying risk of adult periodontitis. We completed an analysis of a data set of monozygotic and dizygotic adult twin pairs. These results indicated that risk of adult periodontal disease has a substantial hereditary basis (50%). Importantly, this estimate was not diminished by adjustment for behavioral variables such as smoking or frequency of dental visits, further pointing to a biological basis for the genetic component of differences in risk to this common disease.
| Michalowicz, B S; Diehl, S R; Gunsolley, J C et al. (2000) Evidence of a substantial genetic basis for risk of adult periodontitis. J Periodontol 71:1699-707 |
