Periodontal Research: Molecular Determinants of Risk With a Commentary on Post-Genomic Periodontal Research Musings of an Experimentalist Steven Offenbacher Center for Oral and Systemic Diseases, University of North Carolina, Chapel Hill. i CONFERENCE OVERVIEW The 20th ICPR was one of the first periodontal research conferences of the post-genomic era. It was organized with the purpose of providing periodontal scientists with an overview of what the future may hold using microbial and human genomics as tools to explore pathogenesis and redene diagnosis, preven- tion, and treatment. It brought experts in molecular genomics and analytical modeling methods from dis- parate areas to provide discussions ranging from inte- grating genomic data into epidemiological risk mod- els to designing laboratory experiments to explore changes in transcriptome expression in cells, tissues, and animal models. Elegant techniques were shown that begin to fully characterize the biodiversity of the oral biolm and to characterize the virulence genes of organisms with genomic approaches. The goal of the conference was to begin to bring these molecular meth- ods of analyses to bear on risk assessment of peri- odontal infection on oral and systemic disease. The intent was to begin the process of identifying specic genetic markers that confer risk or resistance, and cer- tain microbial virulence genes that are critical to patho- genesis to reconstruct denitions of diseases that are not necessarily generalizable to the population at large, but provide a detailed, specic characterization at a patient level or within a smaller subset of individuals. Papers were presented that provided new insights into the thought processes that will be needed to establish new clusters of clinical and laboratory ndings to dene periodontal disease syndromes and to provide new denitions of disease that better characterize the oral condition as it serves as an exposure for systemic complications. Commonalities of pathogenesis were raised when periodontal disease was considered as a triad of clinical signs, oral infection, and inammatory response rather than a single cluster of clinical nd- ings. These mechanisms extend from local tissue destruction to pathways that involve the vasculature and pregnancy. Papers on the inuence of lipid and carbohydrate metabolism, as well as obesity and dia- betes as metabolic conditions, were discussed as to how they modify periodontal disease expression or serve as systemic stressors. Discussions ranged from lipid modulation of inammation to discussions of the differences between anthropomorphic definitions of central obesity, body mass index, and waist-to-hip ratio on disease models. These papers and discussions helped to clarify how obesity and diabetes serve as effect modiers as one examines the role of oral infec- tion on periodontal status and systemic complications such as cardiovascular disease. The research trend manifested by this ICPR meet- ing is clear. We are rapidly moving into an era when molecular laboratory findings, especially protein or DNA-based, using host or pathogen biological samples will be tightly integrated into our traditional clinical denitions of disease. It is the promise of the future that we will eventually be able to better dene the dis- ease at a patient level to provide personalized med- icine. That is, based upon our new complex diag- nostic profiling that includes an individuals genetic background, behaviors, and exposures, we will then be able to offer a specic preventive or therapeutic strategy that is customized to the individual. This Con- ference was a successful step towards describing these approaches to explore this formidable challenge of personalized medicine. And, as a consequence of reection and debate on this transitional thinking and in the spirit of the ICPR, which traditionally fosters an open-minded exchange of controversial ideas, the fol- lowing personal commentary on this post-genomic future is offered. COMMENTARY Many of us in science can remember the excitement generated by the original scientific breakthrough of Watson and Crick that started it all 50 years ago and can now marvel at the realization of milestone achieve- ments with the completion of the human genome pro- ject and the recent sequencing of several genomes of relevant oral pathogens. It is a phenomenal time to be in science, as these shared national scientific land- marks of achievement rival the collective effort of putting a manned spacecraft on the moon. These new genetic roadmaps should help us navigate our exper- imental pathways as we try to decipher the complex intricacies of microbial virulence and pathogenesis, 2009 3/26/03 1:12 PM Page i Reections on the 20th International Conference on Periodontal Research Volume 7 Number 1 December 2002 gene-environment interactions, periodontal disease sus- ceptibility and resistance traits, and the role of peri- odontal disease as a systemic exposure. The magni- tude of the breadth and complexity of the information has birthed its own discipline (bioinformatics) and is purported to eventually enable us to achieve the com- mon goal of personalized medicine that includes den- tistry. That is, the prevention and treatment of human ailments will be repackaged not to deliver what gen- erally works best and is globally safe for everyone, but rather, in the context of optimizing the prevention and intervention specically for the unique patient using information about an individuals genetic characteris- tics, exposures, biomarker patterns, and behaviors. This is the promise of medicine for the future that is her- alded by the omics. But exactly how the new post- genomic era will create new directions for dental research, the context in which dental scientists will con- tribute to the future of molecular dentistry and the main- stream of biomedical research, is just beginning to unfold. Inevitably, changing trends in research are precipi- tated by new discovery and the creation of new tools. Landmark technological advances that create new methods can sometimes create self-validating scien- tic lines of investigation rather than simply provide better tools for scientic discovery. Thus, as we move into this new area of genomics, one is reminded of the afterglow of post-Apollo landing times when we glee- fully discussed living on Mars, just as today with the sequencing of the human genome some now forecast the imminent end of human suffering from disease. At times, this optimism is fueled by the ecumenical media hype for one trendy research method or another. But the history of biomedical research is spotted with exam- ples of ineffective lines of investigation, suggesting that sometimes the tools we use drive the scientic ques- tions we pose rather than the other way around. For example, with the perfection of electron microscopy, structural biology became the science of the day, and then there was the era of monoclonal antibodies. How many research facilities still harbor antigen-specific monoclonal cell lines that may have been of interest but were never characterized? With the explosion in mol- ecular biology over the last 2 decades, it seems that virtually every scientific discipline has now been renamed to include the word molecular. And in recent years, how many laboratories have studied a unique gene that has been cloned, sequenced, and, via trans- genic models, been over- or underexpressed, and still remains without a clear function? With the sequencing of the human genome and a growing list of microbial genomes providing the criti- cal blueprints for potential experiments, some specic thoughts came to mind. I would like to briey raise some points of discussion and, perhaps, controversy. These comments are not intended to criticize or dimin- ish the tremendous advances that have been realized by the milestone achievements of the past several decades. Instead, these comments are made to pre- sent a personal perspective that may, for some, pos- sibly minimize well-intended experimental meander- ings as we embark into the new post-genomics era. 1. My rst point is that the human genome sequence was done on a composite individual to produce a lim- ited sequence map for those genes that code for pro- teins. Those open-reading frames are regions that are transcribed into messenger RNA and then translated into proteins. That represents only about 25% to 35% of the entire DNA sequence. There are huge areas of DNA in between these open-reading frames that are thought to play some regulatory and structural role for maintaining the genome. The data in the human genome map contain, for example, the entire sequence for type I collagen alpha chain. It is often popular to cite in the literature that the differences between the genes for man and a gorilla differ by less than 5%. It is no surprise that the sequence of type I collagen for man is almost identical to that of a gorilla. It has been argued that the differences between gorillas and men can be explained by that infinitesimally small, but somehow magically transforming, 5% that represents the tiny difference in the coding regions, as compar- ing our 2 closely aligned species. From my perspec- tive that concept is analogous to saying that tigers, zebras, and convicts are almost identical that is, from the viewpoint of black stripes. After all, collagen is a basic building block of connective tissue and it would be inefcient for us not to have utilized this pro- tein as an evolutionarily conserved molecule. Clearly, genes that code for proteins representing the common building blocks of life that are needed for a biologi- cally viable mammal should logically be quite similar. These sequences are conserved because mother nature found them to be useful so they need not be reinvented. But what about those unknown sequences contained within the non-coding DNA? Many years ago these genes were termed nonsense genes, as the gene products did not result in protein sequences and there is tremendous diversity in these sequences. It is the unique combination of sequences in this region of the total DNA that is used to incriminate a suspected murderer, for example. There are regions of DNA that contain so-called random repeat areas of DNA that do not code for protein but can be used to identify a person uniquely and establish parentage and familial linkages. However, it might be argued that it would also seem totally evolutionarily inefcient that any DNA sequence would be conserved, such as these non-cod- ing regions, if they are indeed nonsense or non-func- tional. It would seem more plausible that these sequences are not, in fact, random and contain infor- mation as yet to be discovered. These are the sequences that we inherited from our parents that give us our special personalities and physical features and they are probably no more random than those regions ii 2009 3/26/03 1:12 PM Page ii Ann Periodontol Offenbacher that encode for protein that we also considered to be random gibberish just 50 years ago. One might recall that prior to Watson and Crick, the entire 4-base sequence of the entire chromosomal DNA was also considered a random chain of bases. We now know that it is clearly not random and that it has a critical importance with regard to lineage and heritance. This would suggest that there is much more of the genetic code yet to be unraveled. The limited nature of the DNA sequence decoding data also raises questions regarding the general util- ity of the human genome data to ascertain risk for dis- ease and assess response to therapy. For example, when forensic scientists seek to identify a person uniquely, it is not done by comparing DNA from an individual to the human genome sequence data. After all, using just these data, we could not really distinguish very efciently if the DNA sequence were from a mur- der suspect or a gorilla. One might then easily ques- tion how we are to use these sequence data to iden- tify genes which confer risk for diseases especially most of them that have familial tendencies like peri- odontal disease, cardiovascular disease, and diabetes. And what about the other 65% to 75% of the DNA that has not been sequenced? These genes are clearly inherited from our parents. It would not appear likely that those genes would have been retained evolution- arily along our ancestry if they performed no function. Would not these regions be more likely to harbor the differences that account for both survival genes and familial patterns of disease inheritance? It is the non- coding region of the DNA that is used to identify ones parentage unequivocally. If one wants to prove that a DNA sample is from a specific person, then the sequences are determined for the highly variable non- coding regions of the DNA. Those sequences not only identify us as a specic person, but arguably it is those sequences that uniquely dene us as a biological entity. Logically, the fact that I am more like my father than a gorilla is probably due to the fact that I inherited those non-coding regions from my father and not a gorilla (more-or-less...). Indeed, most of the DNA that uniquely denes us is not within the human genome sequence base. It would appear likely that the dis- covery of relatively rare, single gene mutations (i.e., genetic defects) that result in altered proteins not crit- ical for survival will be elucidated rst by the use of the human gene bank data, and that complex polygenic diseases will be more difficult to discover with this dataset. Admittedly, there will also be important link- ages which may co-segregate with those genes of inter- est, so it will provide useful information to dene hot spots of the genome that confer risk. Furthermore, the gene sequences that are known do serve as impor- tant islands of information and points of reference within the vast and expansive sea of unknown DNA code. The logical consequences that follow from this line of thinking should be raised from a scientific view- point, even if controversial. 2. We probably have not completely broken the genetic code. We understand the function of some of the DNA sequences that code for proteins, but we are only beginning to understand how promoter regions define the transcriptome function. Messenger RNA (mRNA) synthesis, mRNA splicing, and mRNA turnover are ultimately regulated by DNA and RNA sequences and by key interactions of the DNA and RNA with pro- teins that often require metabolic alteration such as phosphorylation, acetylation, or conformational shifts via ligand binding with other molecules, including lipids. For example, the open-reading frame gene sequences for prostaglandin synthase or tumor necrosis factor that translate into the primary amino acid sequences for these proteins are virtually identical from person to person, but the sequences upstream and downstream of the coding region can differ signicantly from per- son to person. Often these are single nucleotide poly- morphisms or SNPs, which are inherited genes that usually are not in encoding regions of the genome. There are DNA sequences that provide docking bays that allow proteins to bind and which function by enhancing the transcription of the specic mRNA mol- ecules that code for these proteins. These DNA regions are called response elements. Among other mecha- nisms, different SNPs can alter the binding of nuclear transcriptional factors to these elements and modu- late transcription levels. Thus, differences in these SNPs may ultimately account for genetic differences in the amount of mRNA of a critically important biological molecule such as a cytokine or growth factor that reg- ulates the host response to challenges and may result in a phenotypic difference in host susceptibility or resis- tance. Many genes have similar response element sequences that enable them to be expressed or mod- ulated in a concerted manner. In the future, the bio- logical responsiveness and function of a molecule may be predicted based upon the sequences of these reg- ulatory elements that ank the open-reading frames. The study of these regions will undoubtedly be made much easier with the sequencing of the surrounding coding regions, but this is a region likely to be where true differences in hosts can be genetically identied to enable personalized medicine approaches. Further- more, it would appear likely that there are several other kinds of genetic codes that determine how genes are regulated that we have yet to discover. 3. The genes that make us different are also likely to include those genes that explain the differences among us in disease resistance and susceptibility. This may seem to be an obvious point, but it is not a pop- ular point of view. We tend to prefer the concept that we are all genetically the same, or almost the same, and that diseases are thrust upon us as a consequence of environmental exposures, behaviors, or, even bet- ter, external inuences that we cannot often control iii 2009 3/26/03 1:12 PM Page iii Reections on the 20th International Conference on Periodontal Research Volume 7 Number 1 December 2002 such as socioeconomic status, stress, or educational attainment. However, the data are consistently telling a different story. Studies of genetically identical twins reared together or apart continue to provide compelling scientic evidence that many major human illnesses that affect the majority of the population ranging from cardiovascular disease, to diabetes, to periodontal dis- ease have a substantive genetic component. 1-4 Over and over, studies tend to reach the same conclusion: that about half of the variance in the expression of health or disease can be explained by genes. This means that half of the disease or the trait can, on aver- age, be explained by the genes that we inherit from our lineage and suggest that the remaining part is due to environmental exposures or behaviors. Thus, this line of thinking would suggest that in order to understand the contribution of genes to risk, we will by necessity need to better understand the role of those DNA sequences that are beyond the coding regions. We will need detailed SNP maps or haplotype analyses to establish patterns of inheritance of multiple genes that confer risk. Furthermore, we will need to better under- stand the complexities of transcriptional biology as well as translational and post-translational regulation before we can appreciate the functional impact of these genetic determinants on the disease phenotype. In concluding this commentary, I would like to re- emphasize that these points are raised as a caution- ary and certainly not a negative note. For example, genetic models of diseases have often used transgenic or knock-out experimental approaches that over- express or underexpress (or express dominant nega- tive proteins) to determine the role of that specic mol- ecule in disease expression or to create new models of disease. Even if these models do not faithfully repro- duce the human condition, they often lead to impor- tant insights into mechanisms of pathogenesis. Stud- ies that examine the role of regulatory genes using these experimental approaches are in a minority, but hold promise. Thus, important strides are being made. However, admitting that there are large genetic differ- ences between and among individuals (and differences that we dont understand) is sometimes a more con- troversial and difcult issue to grasp. However, using differences in DNA sequences that are unique to the individual or the individuals lineage to confer suscep- tibility or resistance should not be viewed similarly to using DNA evidence to identify who someones true parents are (and that they are not a gorilla). Suggest- ing that there are substantive, not minor, genetic dif- ferences between individuals is not to suggest that these differences (desirable or undesirable genes) fall along ethno-racial, cultural, or socioeconomic group- ings. However, it does suggest that reluctance to explore these genetic differences will hamper a full understanding of how gene-environment interactions result in disease. In other words, the sooner we come to grips with the realities and limitations of what we know, and acknowledge that we are just at the very beginning of this emerging eld of genomics, the more likely we will not at some point in the future reminisce with embarrassment at our forecasts of soon living in Martian colonies with personalized full dental plans. ACKNOWLEDGMENTS At this 20th session of the ICPR, the organizers gave a special award of recognition to Robert Singer, PhD., Senior Scientist at Procter & Gamble, in acknowledg- ment of his scientic leadership and the impact he has made in periodontal research and the biomedical sci- entic community. This occasion was in honor of his retirement from Procter & Gamble. His insight and leadership within the industry have had a tremendous impact over the past few decades in growing oral health care from its historical roots in soap chemistry to a growing age of biopharmaceuticals and biomedical device applications. His work has extended beyond basic research and corporate product development, as he has been a strong and unfaltering champion for basic and clinical periodontal research. His advocacy and the sole support of Procter & Gamble have made this ICPR conference a thriving tradition and an entity that promotes new research, young investigators, and the open exchange of ideas. Our committee and the conference attendees would like to thank P&G for its gracious support of the 20th ICPR and especially want to acknowledge the outstanding periodontal research contributions of Bob Singer, who is a true colleague and friend. I would also like to give my special thanks to my co-chair, Dr. Ray Williams, Chair of Periodontology, whose outstanding efforts were critical to the success of the Conference and for personally assuring that it was a well-organized, informative, pleasurable, and memorable experience for all of us. Thanks also to the UNC School of Dentistry (special thanks to Sharon Grayden and Amy Williams, UNC Continuing Educa- tion), the Department of Periodontology, the Compre- hensive Center for Inflammatory Disorders, and the UNC Center for Oral and Systemic Diseases. I want to thank Jim Beck, Ray Williams, and Gary Armitage for their thoughtful suggestions on this commentary. REFERENCES 1. Michalowicz BS, Aeppli D, Virag JG, et al. Periodontal ndings in adult twins. J Periodontol 1991;62:293-299. 2. Michalowicz BS, Aeppli DP, Kuba RK, et al. A twin study of genetic variation in proportional radiographic alveolar bone height. J Dent Res 1991;70:1431-1435. 3. Michalowicz BS, Wolff LF, Klump D, et al. Periodontal bacteria in adult twins. J Periodontol 1999;70:263-273. 4. Michalowicz BS, Diehl SR, Gunsolley JC, et al. Evidence of a substantial genetic basis for risk of adult periodonti- tis. J Periodontol 2000;71:1699-1707. iv 2009 3/26/03 1:12 PM Page iv