Functional Genomics and Proteomics in Personalized Medicine

---

By Eef Harmsen, PhD, Rob Sladek, MD and Dr. Andrew Orr, MD

21st Century Approaches to Complex Diseases

What is personalized medicine?

Personalized medicine is not new. When your doctor tells you that your cholesterol is too high, that you need to lose weight and that you should stop smoking, he or she is giving you personalized advice, tailored to your specific situation, that will help you reduce your chances of having a heart attack or stroke. Their advice, which is similar to the advice that is given to tens of thousands of Canadians, is based on solid clinical studies that have been performed over many decades. It's sound advice and if your personal medical situation fits, you should follow it - particularly if heart disease runs in your family.

If personalized medicine has been around for decades, why is it only recently attracting so much attention? Part of the reason is that we are starting to learn a lot more about how to predict disease risk for individual patients and in particular, how to identify people whose genes make them more likely to get diseases such as diabetes, stroke and cancer. Our knowledge of gene variants (person to person variation in the DNA sequence) has increased rapidly since the finalization of the sequencing of the Human Genome Project and the determination of the frequency of the DNA variation (polymorphisms) in people of different backgrounds (The HapMap project).

In addition, scientists from Canada and the USA have found, that not only people vary by single nucleotide polymorphisms (SNPs), but that some people differ in large blocks of DNA, which are deleted or inserted. Until recently, the major focus was to determine how genetic polymorphisms affected protein structure and function (coding SNPs). Functional genomics approaches, which use expression microarrays, have increasingly demonstrated that small differences in an individual's DNA may affect disease risk by altering genetic control gene activity and modify the amount of protein that is produced in cells of the body (regulatory SNPs) (FIGURE 1). Right now, these disease-associated polymorphisms provide a roadmap to possible molecular damage that causes disease. As we learn more about how these polymorphisms change the function of genes, proteins, cells and organs, we may be able to predict how small changes in the DNA sequence between different people cause illness, how to better predict how serious the illness may become and how to treat it most effectively.

Personalized medicine is based on this new genomics and proteomics knowledge and undoubtedly will result in important changes in how we diagnose and treat many common and chronic diseases. This will have significant impact for the health care of Canadians, since common diseases of the 21st century - such as cardiovascular diseases (stroke, atherosclerosis), metabolic syndrome (obesity and type II diabetes), arthritis and osteoarthritis, neurodegenerative disease like Alzheimer's (including those related to aging) and cancer - account for more than half of all health care expenditures in Canada. These chronic diseases occur as a result of environmental influences acting on genetically susceptible individuals and are called complex diseases, since they are frequently caused by the interacting effects of genetic variation in more than one gene. Furthermore, while a group of patients with cardiovascular disease may have similar symptoms, their individual illness may be caused by different groups of genetic variants affecting the function of different genes leading to different types of molecular and cellular damage. As a consequence, the traditional 'one size fits all' approach to drug therapy could result in incomplete or even harmful treatments (FIGURE 2). Personalized medicine, by taking into account the specific genetic or protein abnormalities in a specific individual, has the potential to substantially reduce the personal and socio-economic burden of many common illnesses.

The scientific and technologic basis of
personalized medicine

The cornerstone of personalized medicine is the identification of specific molecular mechanisms that cause disease in individual patients. Today's physician would be unlikely to tell a patient that they had "dropsy", since contemporary understanding of pathology, biochemistry and physiology would permit a much more specific diagnosis of heart, liver or kidney disease. Personalized medicine would permit patients to receive even more specific diagnoses: for example that their diabetes is caused by dysfunction of a particular protein that regulates insulin secretion in the pancreas. Such a precise diagnosis is scientifically appealing; however, medical science is still far from translating precise disease mechanisms into precise, safe and effective treatments.

In this context, it is important to realize that the link between genetic, protein or metabolic defects and disease involves many factors: while a patient may have a genetic defect that has been statistically associated with disease, environmental, dietary and lifestyle factors may prevent the disease from ever developing.

Key enabling technologies in the development of personalized medicine strategies include proteomic techniques and expression microarrays - the latter of which can measure the levels of tens of thousands mRNA species in a single tissue sample. These technologies can be used to identify molecular phenotypes in diseased tissues, and in turn can be used to discover biomarkers that can be used to diagnose classify externally identical forms of a disease, to follow its progress and to monitor its treatment (FIGURE 3).

In addition, these disease-associated molecules may be useful targets for treating disease. To do this, the molecular impairment needs to be linked to a pathway whose activity can be changed by a drug. This requires an extensive knowledge of 'functional genomics' or 'metabolomics' - sciences that seek to explain how changes in mRNA and cellular metabolites result from genetic differences among individuals.

These new technologies will provide more accurate diagnosis of an illness, the ability to treat a patient more precisely and the possibility of less frequent adverse drug reactions. For instance, people diagnosed with essential hypertension (hypertension without a known underlying cause), are frequently treated with a beta-blocker or a diuretic. If this doesn't reduce the blood pressure sufficiently or induces side effects, then their treatment will be expanded to include additional drugs - such as an ACE inhibitor or angiotensin II receptor blocker. If this also fails then another appropriate drug and dose is tried until one is found with acceptable efficacy and minimal side effects. In the foreseeable future, a personalized medical approach of a patient with hypertension will mean that the genetic and proteomic profile of the patient will improve the diagnosis of the underlying cause of the hypertension and allow the selection of a specific drug treatment, which is more effective and produces fewer side effects. In addition, the genomics and proteomic profiles of patients with hypertension will further enhance our insight into the cause of hypertension, which in its turn will form a solid basis for a rationale drug discovery strategy.

An alternative example is the application of a new drug Herceptin for the treatment of patients with metastatic breast cancer. Herceptin is a monoclonal antibody directed at the human epidermal growth factor receptor 2 (HER2): patients that overexpress HER2 in their tumors can be effectively treated using this antibody, while breast cancer patients who do not overexpress HER2 will not benefit from this medication.

The challenges of personalized medicine

Personalized medicine is a simple and attractive term that describes medical science's goal of understanding the molecular causes of disease: while this goal is rational and appealing, the effective development of personalized strategies for care faces a number of real and significant challenges. One major challenge ahead of us is to find which genetic variations cause common chronic diseases. This is a daunting task, considering that millions of common genetic polymorphisms have been discovered in humans in recent years.

However, large-scale genotyping studies in clinical cohorts are underway, which will indicate which gene variants are associated with a susceptibility to certain diseases. For instance, the Framingham study (heart disease) in the USA recently published hundreds of genes variants, which will increase the risk for heart disease. Similarly, a large Canadian study on Type 2 diabetes reported several gene variants increased patient's risk of type 2 diabetes. These and future discoveries, combined with biomarkers identified by clinical and histo-proteomic techniques will create a more accurate personal risk profiles for common diseases. They will also help provide tests to monitor the progress of disease and to identify the causes of variable drug response among individuals. Frequently, the initial optimism surrounding the use of a new marker to diagnose disease or predict disease risk is tempered by its reduced effectiveness in larger patient populations.

Consequently, the success of personalized medicine will critically depend on large prospective clinical trials to evaluate the suitability of genomic, proteomic and metabolomic markers for diagnosis and assessment of therapeutic responses.
A second large challenge in the next 10 years will be to develop effective approaches linking functional genetic polymorphisms to intracellular biochemical processes.

Many Canadian research groups are proposing to address these issues by developing a framework for personalized medicine that systematically links physiological processes in living cells to genetic changes that are associated with diseases or therapeutic responses. These research projects will take advantage of high-throughput genomics and proteomics technologies to discover new diagnostic and prognostic markers that can be used to explain patient-to-patient variation. The combination of complex disease genetics with genomics and proteomics in human cell models will close an important gap between disease gene discovery and its translation to molecular mechanisms and the use of these mechanisms to identify targets for clinical treatments.

Additional challenges play a crucial role in personalized medicine. For example, will the identification of 'high risk' patients result in the targeting of prevention strategies to a small group of individuals and hinder health-promoting behaviours for all? How will patients respond to learning that they were at significant risk of developing a serious illness? Will they enthusiastically follow mildly effective prevention strategies or will they simply accept the disease outcome as inevitable? What about patients who are told that they are at low risk?

Will they continue to follow health-damaging behaviours? To date, the outcome of studies looking at patient responses to genetic risk testing have been mixed. For example, patients with a high genetic risk of developing lung cancer frequently do not modify their cigarette consumption. This preliminary data suggest that genetic risk profiling will not necessarily influence behavior and additional policies will remain needed. In addition, there are concerns that extensive genotyping might have social implications in relation to increased premiums of health or life insurance premiums for people with specific medical risk or access to these data by prospective employers.

What role can Canada play in the development and application of personalized medicine?

We anticipate that personalized medicine will become part of the Canadian healthcare system within the foreseeable future. Canadian scientists are financially supported through Genome Canada, a federal funding agency promoting genomics research and the affiliated Genome Centres such as Genome BC, Genome Alberta, Genome Prairie, The Ontario Genomics Institute, Genome Québec and Genome Atlantic. These funding agencies have made significant investments into gene and biomarker discovery for complex diseases and their treatment. Similar studies have been supported by the Canadian Institute of Health Research (CIHR) and other national funding agencies.

With this support, Canadian scientists are conducting leading-edge research in genomics, proteomics and the study of functional regulatory variation. These studies need to be extended to involve clinical collaborations in fields such as arthritis, cardiovascular disease, metabolic disorders, cancer, pathology and surgical specialties. This will provide unique provincial and national opportunities for coordinated studies of large clinical cohorts, as well as biobanking of tissue samples and cell lines that can be used to dissect the physiologic consequences of specific genetic or protein abnormalities in individual patients. This work will be supported through Genome Canada's network of regional centres specializing in some of the most advanced genomics and proteomics technologies required to establish personalized approaches to clinical care.

We are at the dawn of a new medical era, which started with the sequencing of the human genome. An expanded Canadian genomics and proteomics infrastructure can with the many Canadian disease-specific clinical research networks, establish Canada as a leader in studying the pathogenesis and treatment of chronic diseases.

Eef Harmsen, PhD, is the project manager of 'The GRID (Gene Regulators in Disease) project' at McGill University and Genome Quebec Innovation Center in Montreal; Rob Sladek, MD, is the scientific director of the Diabetes Gene Discovery Group and an assistant Professor, Human Genetics at McGill University and Genome Quebec Innovation Centre in Montreal and Dr. Andrew Orr, MD, is a co-investigator with the Atlantic Medical Genetics and Genomics Initiative and assistant Professor Ophthalmology & Visual Sciences at Dalhousie University in Halifax.