Personalized Medicine is a Science

Students who complete a new pre-med course at Kettering University in Michigan will receive a minor in biochemistry. The university’s Biochemistry Department and its Bio-Engineering Laboratory support the program. The idea is to prepare the students for med school, the biotech industry, or graduate school. A Kettering faculty member said: “As medicine becomes increasingly technical, having a strong foundation in science with an understanding of engineering will make the difference.”

A George Mason University-based developer of nanotechnologies for medicine (which we discuss below) seems to imply that med schools may not be capable of taking advantage of the sort of science-oriented students Kettering aims to turn out. He claimed that not having a medical school at George Mason was more of an advantage to clinical and translational research. His program partners included a local hospital system, private-practice physicians, and overseas cancer centers.

That position may be a little extreme, but regardless of whether scientific education and research in pathology, radiology, and molecular medicine takes place in med schools or engineering departments, the results will mean a radical shift in diagnostic practice in coming years as individualized genomic diagnosis and more targeted and integrated pathological and radiological diagnostic methods become the norm.

At the hospital, integrated diagnostics will seamlessly incorporate pathological, radiological, and molecular tools in a “department of diagnostic medicine.”[i] One of the rare examples of such a facility, in Brazil, provides a full suite of diagnostic services across many specialties and combines pathology and radiology results in one report. No such center exists in the United States, which is hampered in part by physician resistance and IT limitations to integrating radiology and pathology legacy systems and data.

Obstacles to Personalized Medicines

Traditional trial-and-error diagnosis is being replaced by precision-targeted diagnosis that reveals an individual’s disease or predisposition to disease, enabling precise therapies to be tailored to the individual. However, major obstacles remain before such “personalized medicine” reaches its full potential. For example:

• Pharmaceutical companies are slow to embrace targeted drug discovery and companion diagnostic tests to assess patient eligibility.

• Many physicians will be reluctant to learn and incorporate personalized medicine practices.

• Identifying genetic and proteomic pathways in individual diseases is difficult (but not impossible, as projects such as the Cancer Genome Atlas have already begun to show.)

• FDA approval of personalized medicine technologies will not guarantee Medicare or other insurer reimbursement for them. (In any case, genetic tests are virtually free of regulatory oversight.)

• The technologies of personalized medicine, such as genetic diagnostic tests, are currently expensive, though they could save money in the long run through earlier and more accurate diagnosis and treatment.

Genetic Diagnostic Tests

What is evidently not an obstacle to personalized medicine is the questionable validity of many current genetic diagnostic tests. More than 20 companies in the US today sell such tests directly to the consumer, claiming they help customers discover their predispositions to disease, personality traits, and drug addiction. The tests can today detect hundreds of thousands of genetic anomalies in a sample of saliva or blood. So far, about 100 such anomalies have been confirmed to influence specific diseases or behavioral tendencies, but the number is growing rapidly.

The probabilistic nature even of valid tests is problematic: What does an n percent chance of succumbing to a disease really mean to the patient? And what if—as happens often—the validity of a genetic link is modified or overturned by later research? Patients may base life-changing decisions on products that genome superstar J. Craig Venter says “are going on the marketplace as though they were underarm deodorant.”

Genetic tests will become valid and reliable, eventually, but to do so will take massive amounts of data from a wide array of people. Some direct-to-consumer genomics companies not only accept that but also have begun to collect it. Customers of 23andMe (which has just slashed the price of its test from $899 to $399) share information about their genes, their health, and their personalities on the company’s website.

It’s Happening

Obstacles notwithstanding, personalized medicine will become the standard of care. The question is when, not if. The first steps have already been taken, as, for example, with personalized asthma medicine: “Protein profiles”­­­—active combinations of proteins—that indicate different subtypes of asthma have been identified. The resulting ability to diagnose a patient’s asthma subtype will enable treatments to be tailored to that patient and result in better outcomes. However, more protein profiles remain to be discovered and a simple blood or breath test needs to be developed to detect the profile present in a patient.

For another example: A small clinical trial has demonstrated that cancer treatment can be tailored to the genetic profile of a patient’s tumor and result in better outcomes. The ultimate goal is to take a molecular fingerprint of someone’s tumor and assign treatment based on molecular defects. Already, it has been found that certain non-small-cell lung cancer patients susceptible to EGFR (epidermal growth factor receptor) inhibitors have a specific molecule in their cancer cells that can be targeted with an EGFR inhibitor drug such as Iressa. In a small trial, the median time that such patients survived without their cancer progressing was nine months when treated with Iressa, versus four months for those who received chemotherapy.

The drug Abraxane is one of several FDA-approved nanotherapies for late-stage breast cancer in women who have failed other therapy. It uses a protein-bound nanoparticle system to deliver Taxol to the tumor cells.

Cancer Proteomics

These examples illustrate that disease proteins, not genes, are the proper and effective target for drugs. That is because proteins do the actual work in the cell, for good or ill. Genes and gene expression merely indicate whether or not proteins may be present, not whether they are active. For instance: About half of multiple sclerosis patients respond to interferon beta therapy, but half do not. The difference appears to be genetic, but it is not enough to know that. We need to understand the proteomic mechanisms and pathways regulated by the genes.

Therefore, “The road to personalized medicine will be paved with proteins,” says the researcher mentioned at the beginning of this issue who is less than enamoured of med schools and who has co-developed nanotechnologies for clinical proteomics. One of his technologies—an engineered nanoparticle—detects biomarkers in blood, urine, or other body fluid, and amplifies the sparse concentration of these markers to aid in identification. Another of his technologies, a protein chip, then detects which proteins in a large sample are actively contributing to a patient’s given disease condition. Those proteins become the target of new classes of therapies designed to turn them off.

So a genetic diagnosis may take us close to a patient’s disease, and with luck, to the actual proteins causing it; but a proteomic diagnosis hones right in. In cancer, every tumor is different at the molecular level, yet today patients who have the same cancer generally receive the same therapy. Fewer than 20 percent of them respond to the “generic” therapy, which means 80 percent receive treatment that is ineffective and toxic. In fact, the active protein in one patient’s breast cancer could be the same protein causing lung cancer in another patient, which suggests that both patients could be effectively treated by the same drug.

The protein chip measures the active pathways (pre-identified as such by the nanoparticle detector) in a patient’s diseased cells. It is currently being used in studies to examine macular degeneration in the eye, fat cells in obese patients, and blood vessel inflammation in heart disease, among others. By telling physicians the level of activation of the protein drug targets in each patient sample, it should enable them to precision-tailor their therapies.

As of this writing, the nanoparticle detector had identified an active signaling pathway in patients who failed to lose much weight after bariatric surgery. Subsequent protein chip analysis of fat cells taken from bariatric surgery candidates appeared to predict how well each patient would respond to the surgery in terms of weight loss and diabetes resolution.

The FDA currently treats nanotechnology no differently than any other new technology that would be incorporated into FDA-regulated products drugs—biologics, devices, food, food additives, nutritional supplements, and cosmetics. It does, however, recognize there are many unknowns and is working with other agencies, including the National Cancer Institute, to determine what additional information is needed to evaluate biosafety and other factors.

It is time for all of us to get used to hearing the terms “personalized medicine” and “nanomedicine.” It is time for clinicians to start preparing to practice them. And it is time for medical schools and engineering schools to start integrating their curricula to produce the next generation of clinicians and clinical researchers.