Traditionally, medicine has been based on the disease and not the patient. This means that for the same disease, e.g. childhood leukemia, two very different patients — let’s say a 5-year-old Spanish girl of Arab origin and a 7-year-old boy in Sweden — will very likely be treated with the same chemotherapy protocol. This is one of the reasons why therapy outcomes are indeed diverse.
There are patients that respond extraordinarily well to therapy, those who also show a good response but with adverse effects, those who do not respond to treatment at all, and there are even people for whom the drug is not effective while also having to cope with toxicity. What has come to be known as personalized or precision medicine aims to confront this scenario with a one-size-does-not-fit-all strategy, using data from the patient, most importantly their genes, to prevent, diagnose and treat diseases.
As state-of-the-art as it may sound, this is not a new concept. I first learned of this idea in the early nineties while attending a seminar on pharmacogenetics, which is the study of the influence of genes on the pharmacological response. I remember the speaker saying that in a not so distant future we would all have health cards with microchips (actually it was described more like barcodes back then) containing all our genetic information. The idea was that at the time of prescribing, the doctor would get some kind of warning in the event that the prescribed drug constituted a risk to the patient’s health. The system could also be used to choose the right dose of a medication looking for the highest possible efficiency. All this based on the stored genetic data. Over the years I have attended similar talks on several occasions, always raising the same question: why treat everyone in the same way if we are all different? The solution to this question was then called personalized medicine, and its implementation in the mid-long term was almost a certainty.
However, more than twenty years later, the clinical routine does not include this therapeutic approach and it does not appear it will any time soon. This is somewhat surprising because the technology needed to implement the system as it was hypothesized is pretty much available. In Spain and other countries of the EU, there are personal health cards that are as advanced as to include tax data to calculate how much the patient has to pay for medications, store electronic prescriptions, or plan chronic treatments in advance to avoid additional visits to the doctor. In addition, the evolution of genetic analysis techniques in the last two decades has been enormous, not only regarding capacity (thousands of mutations of a DNA sample can now be analyzed in a relatively short time) but also from an economic point of view. The competition generated by the increasing number of companies performing genetic testing has brought down prices significantly. Sequencing a person’s whole genome cost $100,000 in 2001, while it can be done for roughly $1000 in 2018.
So, why is personalized medicine not a day-to-day reality yet? The answer is simple: the expectations were too high. We oversimplified the situation, thinking that knowing an individual’s genome would give us the key to designing the perfect therapy. Over the years, however, we have understood that genetics, being important, is in most cases just another factor that can affect the outcome of a treatment. The age of the patient, their environment, diet, interactions with other drugs, or the presence of other pathologies, to give some examples, influence the therapeutic response of the patient with equal or greater weight. In addition, the development of concepts such as epigenetics (changes in gene expression not involving DNA alterations) tells us that the personalized medicine approach is far more complex than expected.
Furthermore, the generalized implementation of these individualized therapies requires that all stakeholders: patients, clinicians, researchers, industry, regulatory agencies, and governments must closely collaborate to make it a reality (Wafi et al. 2018). Each of these players perceive obstacles that are hampering the routine use of personalized medicine: lack of access for many patients; lack of universal protocols and guidelines; problems often generated by the use of confidential information; lack of training necessary for the clinician to interpret test results; a perception by policymakers that the cost of generalized genetic analyses is unaffordable, especially for public health systems; the long and burdensome review periods to get approval by regulatory agencies; or concerns about performing universal genetic screening when many of the clinically important mutations are rare.
On the bright side, some policies are being developed to overcome these difficulties. In 2015, President Obama launched the Precision Medicine Initiative in the US, a long-term research endeavor involving multiple research centers that aimed to bring precision medicine to all areas of health (Collins et al. 2015). In the EU, a significant investment on this therapeutic strategy was started in 2010 within the frame of the research programs FP7 and Horizon 2020 (Nimmesgern et al. 2017), and agencies such as the European Alliance for Personalised Medicine have been created.
Also, an effort is being made to make relevant pharmacogenetic information available to both patients and clinicians. The FDA issues a regularly updated list of pharmacogenetic biomarkers in drug labeling. These are mutations (around 200 in the last list) that are known to affect drug exposure, clinical response variability, adverse events, dosing, or mechanisms of drug action. In Europe, according to data from the European Medicines Agency (EMA), the percentage of medications with pharmacogenomic markers in their product label is less than 15%. In theory, the determination of these genetic variants would provide the clinician with valuable information to individualize the therapy. In reality, pharmacogenetic testing is only routinely performed in a handful of the drugs affected by these mutations.
Now, the fact that genetic tests are not regularly being done at the doctor’s office does not by any means imply that screening for genetic variants is not useful, quite the opposite. It is simply a matter of lowering expectations, of understanding that we will probably never have a primary care doctor whose prescriptions are determined by the genetics of a patient; while at the same time recognizing that, sometimes, knowing whether the patient has a mutation or not can be the difference between life and death, and this is not a rhetorical figure.
The case of abacavir, for instance, is the paradigm of personalized medicine. Abacavir is a highly efficient drug prescribed for HIV infection and has in fact been one of the agents responsible for the spectacular improvement in AIDS management. On the other hand, this drug may induce hypersensitivity reactions in about 4-8 % of patients. The symptoms are initially non-specific and they are easily overlooked or even mistaken by a common cold. If the reaction goes indeed unnoticed and abacavir is administered again, the already alerted immune system may cause a potentially deadly reaction, which at the very least will lead to termination of the treatment. Here is where personalized medicine comes into play. Ten years ago it was shown that this reaction was only present in patients with a certain HLA allele (a key component of the immune system). A simple, cost-effective test before starting treatment is able to identify those patients carrying the hazardous HLA variant, who should in no circumstances be administered abacavir.
There are other therapies that may also benefit from genetic screening, although these tests are not so commonly carried out. Clopidogrel is an antiplatelet medication administered after the placement of a coronary stent and also used to reduce the risk of heart disease and stroke. This is actually a prodrug that needs to be activated in the patient’s body before it can efficiently reduce platelet aggregation. As the activation is carried out by a single gene, those patients with important mutations in it will not have an efficient activation and hence will be at risk of therapeutic failure. The FDA has included a warning in the drug label describing this information and the convenience of performing a genetic test prior to administration. These tests can now easily identify those individuals who are at risk (approximately 2 in 100), whose doctors should consider other antiplatelet alternatives.
A similar case is that of Tamoxifen, the cornerstone of therapy in certain breast tumors, and a drug that also needs to be activated. Women with alterations in the responsible gene, defects that are easily detectable, will not be able to activate the compound adequately and consequently will be at higher risk of cancer relapse. Genetic testing can also be helpful in situations when dosing can be challenging, such as it occurs with warfarin. Too much dose of this widely-used anticoagulant may cause bleeding, and too little may have no effect. There are available algorithms that by combining genetic data with other physiological variables of the patient allow for a better adjustment of the dose. There are also cases when even a single gene may be important for more than one drug. Mutations in the TPMT gene increase the risk of serious toxicity in patients treated with 6-mercaptopurine for leukemia and also decrease the correct activation of azathioprine, which is used to treat autoimmune diseases. Screening for relevant mutations in this gene has proved to be useful in both pathologies.
Finally, oncology is a field where personalized medicine is more settled, although in this case the tests are usually designed to identify particularities of the cancer cell so the industry can create new agents designed to attack these specific targets. Unfortunately, a great deal of these targeted therapies, as they are called, are only offered through clinical trials and are not yet standard treatment options. There are some exceptions, like the recently FDA-approved agent rucaparic, which is utilized in advanced ovarian cancer according to the BRCA gene status, or the existing tests aimed to detect EGFR mutations, which confer sensitivity to tyrosine kinase inhibitors in lung cancer patients.
I have no doubts that personalized medicine is the way to go, but I am equally certain of the existence of numerous limitations. We need to be aware that the level of required cooperation between all stakeholders involved might prove a tough nut to crack. There is also a necessity for a greater investment in basic research for the identification of new biomarkers. In this regard, the advent of new techniques such as next-generation sequencing may be decisive (Inaba et al. 2017). In my opinion, the secret to establishing a scenario where genetic screening may become something we can use on a daily basis, is to take the information provided by the tests for what it is: one more tool available to the clinician which can be negligible sometimes, be able to improve therapy in many occasions, or play a really crucial role in specific cases.
References:
- Collins SF, Varmus H. A new initiative on precision medicine. N Engl J Med 2015;372(9):793-5. doi: 10.1056/NEJMp1500523.
- Inaba H, Azzato EM, Mullighan CG. Integration of Next-Generation Sequencing to Treat Acute Lymphoblastic Leukemia with Targetable Lesions: The St. Jude Children’s Research Hospital Approach. Front Pediatr 2017;5:258. doi: 10.3389/fped.2017.00258.
- Nimmesgern E, Norstedt I, Draghia-Akli R. Enabling personalized medicine in Europe by the European Commission’s funding activities. Per Med 2017;14(4):355–365. doi: 10.2217/pme-2017-0003.
- Wafi A, Mirnezami R. Translational –omics: Future potential and current challenges in precision medicine. Methods 2018 Article in press.