In the clinic of the future, a person’s genomic information will be a routine part of their medical record, no more exotic than height, weight and family history. The unique genetics of each patient will be employed by the doctor and his computer to custom-tailor treatments for every condition imaginable, from HIV and cancer to allergies and weight gain. Most of the guesswork will be removed, as physicians can prescribe drug doses that maximize effectiveness and minimize side effects for each individual, rather than playing the averages of research results.
Most physicians and scientists agree that this vision of genetic medicine will become reality…someday. But will gene-based decision-making become a significant part of the clinical world in the coming years, or the coming decades? That was the question that hovered over Translating Genomics into Personalized Therapy, an afternoon symposium at the University of Chicago last Friday. With four speakers from different corners of medicine speaking to an audience spanning several medical and scientific departments, the event was equal parts optimism and caution – a glimpse at both the promise and limitations of genetic medicine.
If the question is “what can genetic information do for me today?,” the answer was “Depends.” Each speaker at Friday’s symposium reported on the current state of genetic research for a different medical problem: HIV/AIDS, breast cancer, psychiatric disease, and blood disorders. And while some fields are on the verge of real, practical use of genetics in the clinic – for diagnosis or determining treatments, or even reprogramming DNA – others are only just beginning to study the potential of genomics.
On one side of the spectrum were blood disorders such as hemophilia and SCID (X-linked severed combined immunodeficiency, aka “bubble boy syndrome”), for which gene therapies are already under investigation. Arthur Nienhuis of St. Jude Children’s Research Hospital talked about efforts to cure these diseases by inserting a new gene into a patient’s DNA, using a lab-designed virus. The first experiments with this gene therapy led to an unexpected severe side effect of leukemia in some patients, but efforts at St. Jude have worked to develop safer means of inserting new genes, Nienhuis said. Researchers are also looking at other virus delivery systems, including one based on a defanged HIV virus, to improve the efficiency of treatments for blood disorders. But Nienhuis admitted that the road to gene therapy for these diseases was much straighter than it would be for other illnessess, since blood disorders are often caused by only a single “broken” gene.
On the other hand, psychiatric disorders have consistently proven themselves difficult to impossible to be traced back to their genetic roots. Margit Burmeister of the University of Michigan described that frustration in her talk, recapping how genetic association studies involving thousands of subjects have failed to reveal likely candidates genes for diseases such as schizophrenia or bipolar depression. Those are illnesses that run in families, but Burmeister reminded us that a familial trait is not necessarily heritable and genetic.
“Going to medical school, speaking Mandarin and Catholicism also run in families,” Burmeister said.
Whether due to environmental factors, diagnostic variability, or multiple gene interactions, few promising psychiatric illness genes have been discovered. But there have been some smaller successes, Burmeister said, including the discovery of a variant of the receptor for the neurotransmitter serotonin that appears to increase a person’s chance of depression after stressful life events. An upcoming publication from Burmeister’s group will detail the impact of that gene variant on new doctors going through a very predictable stressful period – the first year of residency.
Burmeister also profiled some small successes in using genetic information to tailor types and doses of drugs to schizophrenics, an pharmacogenomic advance that was expanded upon in the talk by David Haas from Vanderbilt University. Haas’ specialty is HIV/AIDS, where treatment advances have made what was once a death sentence into a livable chronic disease. Now that dozens of drugs to treat HIV exist, doctors are seeking new ways of calculating the best combination of those drugs for each patient using genetic information.
Once again, the sci-fi future of feeding a patient’s genotype into a computer that spits out a treatment plan remains far off, but several small findings are starting to add up to a useful whole, Haas said. Gene variants associated with side effects such as jaundice, renal failure and even “brain disconnected from body” psychotropic experiences have been located, informing physicians about how to prescribe medications. Such pharmacogenomics is often touted as the vanguard of first-world medicine, but Haas emphasized that such information would also be useful in developing countries where HIV rates are skyrocketing.
“I really do think this is going to have the biggest impact in resource-limited countries, where you really don’t have the resources to do very intensive monitoring long term,” Haas said. “So what we can learn from a genetic test could have a huge impact if we can get treating a patient right the first time – preventing toxicity and treatment failure.”
That kind of genetic test has been developed in the laboratory of Charles Perou, from the University of North Carolina, to quickly categorize a patient’s breast tumor into different groups that require different treatment approaches. Studying more than 2,000 genes, Perou’s group found that breast tumors can be linked into six categories with different prognoses and sensitivities to treatment; for example, Luminal-A tumors are easily treated with surgery and have high survival rates, while Basal-like tumors are more difficult to treat and often recur.
“Breast cancer is not one disease, it’s at least five different diseases,” Perou said, adding that each patient’s tumor is also singular within those groups. “The gene expression pattern – the portrait – of a tumor is as unique and reproducible as a picture.”
Sequencing 2,000 genes for each breast cancer patient is not practical for a clinic, so Perou and his colleagues developed a simpler, 50-gene test that is 93% as accurate as the full comparison. The results of that test can direct physicians down the correct treatment path earlier and more accurately – in some cases, even recommending against costly, unpleasant chemotherapy in tumors that are likely to regress fully with only hormonal treatment.
“We can spare many of the patients who don’t need chemo, who aren’t benefiting, and we can target it to those who need it,” Perou said. “It’s undoing the current medical practice of giving a therapy to everyone to be safe; now we can figure out some people who don’t need it and they’ll still have a good outcome.”