Regulatory and ethical challenges of personalized medicine

10 years with the genome

Forward progress in the 10 years since the first draft of the human genome has been slow due to the challenges of the genome’s complexity, and the confounding influence of environmental factors [1]. Translation to the clinic is likely to be accelerated by an increased focus on better information management for genomic data and a clearer regulatory path to approval for drugs and companion diagnostics [101]. As the technological hurdles fall, ethical and regulatory issues come more to the fore. This short overview will examine some of them.

Possibilities

In a scene from the science fiction drama, Gattaca (1997), a genetics expert counsels expectant parents on the genetic options for their future child. ‘Wild-type’ individuals, or those with their natural genetic state left intact, are considered inferior to those who are intelligent, beautiful and talented by design. In the British newspaper The Times [102] Francis Collins predicted that within 10 years, genotyping could be part of every individual’s health screening due to lower testing costs and increased access. He cautioned that over-regulation of commercial DNA testing services would be a mistake suggesting that it would be wrong to withhold genetic information from individuals, most of whom he believes, could interpret the data without medical oversight.

While the Gattaca and the Francis Collins views differ dramatically, the underlying premise is the same; that the knowledge in our genes can be used to create healthier and more functional human beings. For this to prevail there is much to be done in improving our ability to understand and communicate the relevance of genetic information.

Finding meaning in data

Biobanks have been essential to our understanding of how genes correlate with phenotypes. However, there are several issues regarding samples and data disposition. First, does the patient have the right to the genetic information derived during the use of their samples for study? Second, how much additional information regarding implications should be included? Third, in the testing of samples contained in a biobank, how broad does the informed consent have to be to allow tangential testing of data from samples acquired for other purposes? Consensus appears to be emerging that specific permission is not required for previously collected samples, provided the samples are appropriately anonymized, coded and not prohibited by the consent process [2]. While this solution might seem straightforward, an altercation between Arizona State University and the Havasupai Indian Tribe of Arizona illustrates the complexity of what constitutes consent for genetic research, particularly for native communities [3,103]. Between 1990 and 1994, the University had acquired samples of blood from the tribe, with their consent, to understand their susceptibility to Type II diabetes. However, the consent statement allowed researchers to ‘study the causes of behavioral/medical disorders,’ which the researchers extended to include the genetics of schizophrenia, inbreeding and the geographic origins of the Havasupai. The published research offended the tribe, threatening the long oral history of their origins in the Grand Canyon. The University eventually paid US$700,000 to 41 individuals for misuse of their DNA. This case illustrates the cultural divide between scientific researchers who are conducting good science, and the public they intend to serve.

Another contentious debate continues around whether genetic data gathered from individuals for research purposes should be shared back with the subject [4,5]. Beslow and Burke recently concluded that the question should be considered more fully in the clinical design since it is dependant on experimental context and genetic findings [4]. Kohane and Taylor have further proposed that the participant’s desires and the communicability of the message be taken into account in determining study design and reporting governance [5].

Personal genetic information: to know or not to know?

From a scientific perspective, the more data that can be collected and analyzed, the more likely the knowledge can be translated to medical benefit. For individual DNA, issues arise when we consider an unarguable fact about genetic data. Once the data is known, it can never be unknown. In medicine, patients acquire labels throughout their life that can be advantageous or detrimental. Our ability to obtain information about our genetic signature provides the option for many more labels that, given their permanence, warrants careful consideration.

In May of this year, Walgreens announced that they would sell gene-testing services in their stores but later reversed their decision after their direct-to-consumer (DTC) partner, Pathway Genomics, was challenged by the US FDA citing concerns about the analytical and clinical accuracy of the tests [104]. While supportive of personalized medicine [6], the FDA is stepping up regulatory focus on personal genotyping. Five commercial genotyping companies received letters from the FDA in June [104] advising the companies that their tests are considered medical devices and are therefore subject to regulatory oversight.

There is debate on how far regulation should go; to genes for disease risk, drug response or even to less contentious SNP patterns such as those recently associated with longevity [105]? The University of California at Berkley recently revealed plans to offer voluntary genetic testing for genes that metabolize lactose, alcohol and folates to 5500 incoming freshmen [106]. The Center for Genetics and Society and others have denounced the initiative saying it underestimates the potential harmful use of the information [107,108].

Several issues arise when individuals test positive for genes associated with potentially fatal diseases, some of which are illustrated with the development of the genetic test for Huntingdon’s disease, a hereditary brain disease that appears in middle age. The genetic association was discovered by researchers using samples from inhabitants of a small village in Venezuela [7]. The inhabitants of this village do not have access to the resulting genetic test and many are not even aware that it exists, which has raised ethical concerns. For genes that have been associated with serious diseases for which there is no known prevention or treatments, it has been assumed that individuals may suffer psychological harm by knowing their genetic status. However, a recent study conducted by Green et al. did not find any short-term detrimental effect in patients who had a parent with Alzheimer’s disease and also an ApoE variant associated with disease risk [8]. If the context and the ambiguity of the results are made clear, patients appear able to accept the result with little anxiety.

It is assumed that for tests ordered through a physician, the doctor will provide the counseling regarding risks and benefits. However, Freedman et al. revealed uncertainty amongst USA physicians regarding the risks and benefit of testing for cancer-related genes, citing multiple areas of concern [9]. Physician education needs to be improved and it is an open question whether the availability of genetic counseling should be mandated with DTC genetic tests.

Masha Gessen has eloquently described the ethical challenges of testing positive for a potentially fatal genetic mutation in her book Blood Matters [10]. Gessen is part of a family lineage that carries a deleterious mutation in the BRCA1 tumor-suppressor gene making them more likely to develop ovarian and breasts cancer [11,12,109]. Gessen’s book vividly describes the difficult issues raised by the genetic label itself, and the decisions she faced regarding prophylactic surgery.

Regulation: how much is enough?

While over-regulation of genetic testing may delay the translation of genetic findings to meaningful application, it is likely that a degree of regulatory oversight will eventually be warranted to minimize test inaccuracies and misinterpretation of data. A recent example of testing error occurred when 23andme reported on their blog site [110] that a number of consumer’s samples were incorrectly processed due to an error in placement of 96-well plates used in the test. Risks of misinterpretation are high given that genetic variants are rarely predictive of disease. A recent paper that examined some of the ethical issues arising from whole-genome association studies for multigenic diseases concluded that the identification of one or a few genes can only be useful if the tests meet three criteria [111]. These are:

• ▪ That the data are reproducible and applicable to the target population;

• ▪ That that data has significant impact to the patient;

• ▪ That the resulting knowledge would lead to some beneficial outcome for the patient.

Regulatory standards should at least extend to data accuracy, marketing of genetic tests and possibly analysis.

Better understanding of the genetic signatures of disease and of the variants that govern drug metabolism promise improved efficacy and minimal safety risk [13,14]. Regulatory considerations apply particularly to safety, and especially to the impact of the drug event, the likelihood in relation to genotype, and the number of patients that may be affected [15,16]. If the genetic determinants of adverse events in a particular population can be discovered, then genetic screening may be employed to avoid prescribing potentially harmful drugs to patients at risk. Screening for the HLA-B*5701 allele, before treatment with the HIV drug abacavir (GlaxoSmithKline, London, UK) significantly reduces the incidence of hypersensitivity to the drug [17]. Should such screens be compulsory when shown to be effective, and should genetic data guide the risk management approach of regulatory agencies? This allele is not carried by 94% of the white population tested, implicating race as a potential proxy for risk management purposes. Ethics and regulation overlap in such cases [18]. Duster has suggested that this amounts to the medicalization of race, and underestimates nonmedical aspects of susceptibility [19]. Wadman has asked if race is sufficient to equate with genetic risk and found the answer to be no, with more discussion needed [20].

It takes a village

As the reductionism of the past two centuries has attempted to create a simpler, more scaleable brand of medicine, the ethical challenges have become amplified as decisions about the health of the many have come to outweigh the needs of the few. In personalized medicine, the patient is at the center of a large genomic village where the ethical and regulatory challenges coalesce around the difficult task of treating the individual fairly and at manageable cost. Technology generally outpaces ethics, and personalized medicine is no exception. However, the scene is clearly set for open, integrative collaboration, so the challenges can be met head on.