(Karl-Ludwig Poggemann/Wikimedia Commons)

Recently, in the hospital cafeteria, I overheard a group of researchers discussing the upcoming availability of whole-genome sequencing to physicians. “We should devise a way to study how physicians will use this,” said one of them—underscoring the disruptive nature of the transformation that is currently happening in medicine.

The ability to immediately obtain whole-genome sequences from patients holds enormous potential for understanding and treating human disease. The list of studies reporting successful diagnosis of otherwise elusive orphan conditions is already too long to recount—more than 600 articles in PubMed as of the date of this posting—including poignant examples of advancing clinical care. Some are transformative achievements, such as a recently published study that identified mutations causing a complex childhood movement disorder. Whole-genome sequencing not only yielded a fundamentally novel biologic discovery, but also led to an effective therapy, resulting in clinical improvements in both patients in the study.

However, widespread use of whole-genome sequencing comes with novel responsibilities. It can yield unexpected results, which will need to be communicated to patients responsibly. The CLARITY contest, announced by Children’s yesterday, is an excellent start: Companies will compete to interpret DNA sequence information from three patients and report it in a form that a physician can easily use to guide patient counseling and care.

By virtue of producing an individual’s unique DNA sequence, clinical use of whole-genome sequencing will create other challenges: how to maintain individual privacy and confidentiality.

The recent decision by the National Collegiate Athletic Association to impose mandatory screening of athletes for sickle cell trait is just one example of how knowledge of the genetic basis of disease may be used outside of medicine. Similar information can be used by health insurance companies to decide insurance eligibility and prices, by employers to guide hiring practices (no matter how illegal), and by individuals to select mates or decide the fate of a pregnancy. Because whole-genome sequences correspond uniquely to each individual, the distribution of whole genome sequencing data for research purposes is tightly regulated, requiring users to be authorized by the National Institutes of Health and to abide by a strict Code of Conduct.

However frightening privacy abuses may be, the more daunting challenge is to integrate the sequencing results of many individuals. This integration is essential since only 1 percent of the 3 billion nucleotides of the human genome contain genes, and only a fraction of these have known functions in human physiology. A full understanding of their function and contribution to disease requires integrating genomic data across currently distinct medical specialties. An oncologist seeking to understand the genetic predisposition for the development of a particular tumor needs access to test results from other individuals with that tumor, but understanding what the sequence variants mean for other diseases requires access to still more individuals and their clinical records.

Add to that the fact that analysis of the whole-genome sequencing data is never completely finished, as illustrated by the recent example of a patient with acute myeloid leukemia (AML). The original analysis, reported in 2008, found 10 genes with mutations. A subsequent analysis reported in 2010, using improved methods, identified additional mutations, including one in a fundamental gene found not just in the original patient but in more than 20 percent of adult patients with AML. This discovery opened important areas of clinical and basic research, and potentially new therapeutic approaches.

Overzealous controls on the use of whole-genome sequencing have the potential to curtail such advances. Data integration and future testing require explicit consent from patients, often years and even decades before the questions and tools for performing such analyses are developed. Current practices by the Institutional Review Boards that regulate biomedical research limit investigators’ ability to contact study participants to discuss test results and solicit further information. Yet, this communication is crucial to avoid unapproved use of genetic samples, as emphasized by the recent legal settlement involving the Havasupai Indian tribe, and to return potentially useful knowledge to patients.

Currently, health providers from all fields of medicine, like my cafeteria neighbors, are asking the same question: How will we use whole-genome sequencing in our practice? Hospitals, physicians and scientists will need to educate themselves and their patients about the potential benefits and risks of this emerging technology.

Alex Kentsis, MD, PhD, is a pediatric hematologist-oncologist at the Dana-Farber/Children’s Hospital Cancer Center who is applying translational genomic and proteomic technologies to advance the treatment of children with tumors and inflammatory diseases.

“What is proteomics?” Answering this simple question was the motivation for the Proteomics 2011, an annual symposium hosted by Judith and Hanno Steen of the Steen & Steen lab and The Proteomics Center at Children’s, featuring global innovators and local advances in proteomics at Children’s Hospital Boston, held last week. As a video at the start of the symposium showed, it’s a question that elicits a wide range of answers:

Video courtesy of the Carino Agency.

Symposium host Hanno Steen gives his opening remarks. (Photo: Chewie Lin)

As noted in the symposium’s opening remarks, the word “proteome,” referring to whole sets of proteins, is only 17 years old, and already the field of proteomics has reached unprecedented scale. Mass spectrometry (mass spec for short), a key technology for proteomic research, can now be used to measure thousands of biomolecules simultaneously and do so in amounts that defy intuition (one part per quintillion, that’s 1 particle in 1,000,000,000,000,000,000!).

This combination of sensitivity and flexibility has enabled a wide range of discoveries in biology and medicine, and the symposium featured several of the latest advances. Ruedi Aebersold from the ETH Zurich in Switzerland and Michal Sharon from the Weizmann Institute in Israel presented new approaches to measuring the interactome, the whole set of molecular interactions in cells, both globally and specifically with respect to human diseases.

Ben Garcia from Princeton University talked about recent advances in epigenetics brought about by the development of functional proteomics, allowing the simultaneous profiling of hundreds of chemical modifications of proteins that affect gene expression.

Harvard Medical School's Dan Finley opens his talk on the proteasome. (Photo: Chewie Lin)

And Dan Finley from Harvard Medical School’s Department of Cell Biology described how enhanced understanding of the proteasome and mechanisms of protein degradation can be used to develop new drug treatments.

Members of The Proteomics Center and the Steen & Steen lab presented some their latest work as well. Sasha Singh discussed FLEXIQinase, a new assay based on the lab’s earlier FLEXIQuant assay that leaps beyond the conventional “on/off” methods used to study molecular switches like kinases. Marc Kirchner presented a transformative method for analyzing multiplex mass spec data, which should greatly enhance our ability to understand complex proteomes.

Sasha Singh listens to an audience question about FLEXIQinase. (Photo: Chewie Lin)

Finally, I presented latest results from a large collaborative project spanning several departments at Children’s Hospital on ways to use proteomics to improve the diagnosis and treatment of wide variety of human diseases. In particular, mass spec-based proteomics allowed my colleagues and me to identify new diagnostic markers of Kawasaki disease, a serious inflammatory and heart illness for which we have no definitive diagnostic test and which is mimicked by many childhood conditions.

In all, Proteomics 2011 suggested that “next generation proteomics” will enable unprecedented discovery of how proteins are regulated by functional modifications, how protein networks control biology, and how their study in patients could improve the diagnosis and treatment of a wide variety of human diseases.

The Proteomics 2011 speakers and hosts (L-R): Dan Finley, Reudi Aebersold, Michal Saron, Alex Kentsis, Judith Steen, Sasha Singh, Joao Paolo of the Steen & Steen lab, Ben Garcia, Marc Kirchner, Hanno Steen. (Photo: Chewie Lin)

[Ed. Note: Proteomics is still a growing field, and as greater analytic (read: computational and instrumental) power becomes available it should have immeasurable impact on our understanding of health and disease. Especially if in the future efforts to integrate molecular profiles with clinical and natural histories start to include proteomic data. So tell us: Where do you think the next generation of proteomic breakthroughs will come from?]

Pediatricians are accustomed to caring for patients with rare diseases. But as all physicians know, common diseases can also behave in rare ways, either presenting in unusual forms or responding only to particular treatments. Recent advances in molecular medicine have confirmed this intuition, particularly in cancer, whose varieties can be virtually as unique as we are ourselves.

Oncologists have formed cooperative groups to meet this challenge long ago. But they remain outmatched by cancer itself: its biology is complex, with different tumors growing as a result of hundreds to thousands of aberrations in cancer genomes and cells. And, until recently, they’ve been limited by the process of traditional clinical trials.

We are beginning to understand some of the genetic aberrations, such as loss of the BRCA1 gene, which predisposes women to breast and ovarian cancer early in life. In addition to causing cancer, loss of BRCA1 also makes cancer cells dependent on a functionally related protein PARP, allowing them to be specifically targeted with PARP inhibitors.

As a result, a number of ongoing clinical trials are investigating the safety and efficacy of these smart drugs. It is possible that incorporating PARP inhibitors into established cancer treatment regimens will lead to transformative improvements for patients with specific loss of the BRCA1 pathway.

What is assured, however, is that these traditional clinical trials will take many years to complete. Combining a smart drug with other personally targeted therapies and conventional drugs may take decades. And, smart as the drugs may be, they may not get a fair test: conventional clinical trials of new cancer therapies typically involve patients with advanced forms of disease, which is often different from disease at initial presentation, and may shortchange drugs that would otherwise be effective.

And paradoxically, as cure rates improve as a result of combination of intensive chemotherapy, radiotherapy and surgery, the number of patients eligible for traditional clinical trials decreases, delaying and sometimes making it impossible to test new targeted therapies that may be less toxic and more effective.

Now, a new clinical trial, termed I-SPY 2 TRIAL (Investigation of Serial Studies to Predict Your Therapeutic Response with Imaging and Molecular Analysis) aims to improve and expedite the drug-testing process for patients with breast cancer, bringing us closer to personalized cancer treatment in general.

Instead of testing new drugs in patients with advanced disease individually and sequentially, and waiting for differences in survival, patients with early forms of breast cancer will be monitored for the immediate effects of new drugs on their tumors in real time. The trial will investigate new drugs designed to target specific aberrations of the individual patient’s breast cancer cells, combining drugs based on their efficacy in shrinking or causing molecular response in tumors. In this way, the trial is intended to “learn” over time how to predict responses to each drug, allowing doctors to identify patients with tumors that are likely to respond to specific combinations of drugs.

This innovative approach would dramatically transform our ability to develop new therapies, certainly for patients with cancer, and perhaps for other diseases as well. The need is made more dire by the exploding knowledge about the molecular causes of disease that makes common diseases rare. Physicians need to take heed of the I-SPY 2 TRIAL, as lessons learned from its outcomes are likely to define the future of medicine for doctors and patients alike.

Not long ago I sat in a room with a young patient and her parents, struggling to devise a treatment that would slow down the growth of her aggressive tumor, which continued in spite of intensive chemotherapy. We knew that the tumor was distinct — it responded to certain combinations of chemotherapy but not others — but we knew little about what drove its growth, and less about how to target our treatment for cure.

While most patients with the common childhood lymphoblastic leukemia can today expect to be cured, progress for those with many rare types of tumors has been disappointing. Their rarity often hinders and sometimes prevents effective clinical trials. Patients are too few or too scattered for any one doctor or hospital, so testing new combinations of existing treatments usually takes decades. Discovery and development of new therapies for rare diseases can last a whole lifetime.

The same is true for many other rare childhood or adult diseases. Take Parkinson’s disease, a neurodegenerative disorder that causes loss of motor function and speech. A recently completed study to investigate a single genetic contribution required a tour-de-force collaboration among 16 hospitals, spread across five continents.

Since the genetic basis of Parkinson’s disease is multifactorial, many more studies like this will be needed. At this rate, they are likely to take many decades. But with advances in information technology — such as internet-based social networks and electronic medical records to connect distant patients and doctors — couldn’t we do this faster?

Spurred by personal interest, the founders of 23andMe, a privately held company providing genetic information directly to consumers, are rising to this challenge. They are uniting a 10,000+ person online social community in which they plan to carry out a genome-wide association study, seeking the many genes that contribute to the development of Parkinson’s disease.

Patients are joining briskly. If results follow suit, some of the genetic alterations discovered may prove to be targets for new drugs, and may even lead to ways to delay or prevent the onset of symptoms. The potential for good is great indeed.

There are also new challenges. Participants must be able to understand the study results, which will likely involve abstruse entities like single nucleotide polymorphisms. Some results may be unexpected, and will need to be communicated responsibly and meaningfully. For example, about 1 in 20 participants will learn that they have factor V Leiden trait, associated with increased risk of blood clots and strokes. Should they do anything differently?

Coriell Personalized Medicine Collaborative, Navigenics, deCODE, Knome and others are exploring similar approaches. Patients will increasingly be drawn to them, but will need to be able to translate this wealth of information into actionable plans — be they treatment for illnesses that ail them currently, or for potential conditions that may or may not materialize. The coming era of personalized medicine will require unprecedented patient advocacy and different ways of thinking about disease by doctors.

Investigators at Children’s Hospital Boston, drawing on the hospital’s diverse patient population, have begun an even more ambitious genetics cohort study called the Gene Partnership Project. The project plans to enroll 10,000 children and their families every year, who will be able to communicate securely with scientists, physicians, geneticists and ethicists who are part of the study. This aptly-termed “informed cohort” will receive recommendations about the study’s results interpreted by a multi-disciplinary team, including information on ways to treat current diseases or possibly prevent future ones.

The project’s even more ambitious goal is to understand the currently mysterious interaction between genetic predisposition towards disease, environmental exposure and lifestyle. Such a patient-centered approach is likely to lead to fundamental advances in our understanding of both rare conditions, like Kawasaki disease, and common ones, like autism.

What about my patients? Sometimes our dedication in searching for a treatment pays off; other times, we grieve with families. Nonetheless, the day when every patient and their newly diagnosed tumor can undergo complete genetic testing is very near. If we are able to connect patients, genes and their doctors, rare tumors will become far less mysterious, and treatments considerably more rational. And with this, the promise of personalized medicine — the ability to predict, prevent and definitively treat disease — could become a reality. Patients and doctors alike can barely wait.

Electrospray needle of a mass spectrometer

As a medical student at the last century’s end, I was taught to practice evidence-based medicine, to use the scientific method instead of the largely anecdotal, experiential practice of the physicians that came before. At this century’s beginning, medicine has begun yet another tectonic shift, termed personalized medicine.

Striving to use information about individual patients to their own benefit is probably as old as medicine itself. But I quickly learned, first as a resident and now as a Hematology and Oncology fellow at Children’s Hospital Boston, that accurate diagnosis remains challenging and perfect treatment, elusive. Frequently, we are not confident in knowing a patient’s exact disease, and treat using therapies that work better for some people than others without being able to predict the outcome.

Whereas Hippocrates described dozens of human conditions, today’s diseases number in the thousands, with the OMIM database alone listing more than 2,700. And the numbers keep growing: some diseases like cancer are so varied that they may be virtually as unique as we are ourselves.

Uncovering the crucial parts of a patient’s medical history and disease biology, performing just the right test to confirm the correct diagnosis and picking the best treatment can become overwhelming. But rapid improvements in technology can now help us make more rational choices.

The first is information technology, where the pace of progress is thrilling: just look at the thousands of patients sharing information at patientslikeme, reporting which treatments work and which don’t, and the hundreds of medical apps for the iPhone. These platforms, along with patient-controlled health records and sites like TuAnalyze (for patients with diabetes) are letting patients capture their own detailed medical histories. In the future, these histories will become essential to diagnosis and care.

At the same time, recent advances in molecular profiling are helping us classify disease and understand treatment in ever more specific ways. Mass spectrometry is an especially powerful tool, with its simultaneous abilities to identify biomolecules and measure them in amounts that defy intuition (one part per quintillion, that’s 1,000,000,000,000,000,000!). Recently, I worked with the Children’s Proteomics Center and Emergency Department, using “mass spec” to find biomarkers of appendicitis, a common surgical emergency that often escapes correct diagnosis, even by seasoned physicians. Our team found accurate markers in a simple urine sample, and a clinical trial is testing their performance as a diagnostic test. Several new collaborative projects are seeking diagnostic markers of Kawasaki disease, kidney tumors, neuroblastoma and leukemia, hoping to identify not only accurate diagnostic approaches, but also causes of diseases that can be targeted with new treatments. Early results are promising!

The ability to not simply define a generic disease or treatment, applying imperfectly to many people, but to define an individual’s disease and treatment, in just the right way, would allow us to deliver cures with assurance, not mere expectation. To conclude — with confidence — that a mysterious constellation of symptoms is a real disease with defined causes, specific diagnostic tests and effective treatment is to practice medicine that’s not only personal, but fundamentally rational.