- BIOETHICS FORUM ESSAY
Arthritis Gene Therapy Trials and Tribulations
Every now and then the field of gene therapy is thrust into turmoil when a clinical trial subject dies. First there was the 1999 death of Jesse Gelsinger, followed a few years later by outbreaks of leukemia in France. Then in July of this year, just as things had calmed down again, a young woman named Jolee Mohr died shortly after receiving a second dose of gene therapy for her rheumatoid arthritis.
Last week, the FDA removed the clinical hold that had been placed on the trial in which Mohr had been enrolled, effectively exonerating the gene transfer procedure as the cause of her death. But a few days later, the NIH’s Recombinant DNA Advisory Committee (RAC) refused definitively to rule out gene therapy as a factor. We will never know for sure whether the patient would be alive had she not enrolled for this protocol, but the whole episode raises several questions concerning the study design, patient selection, approval processes, and data collection and reporting.
Rheumatoid arthritis is a debilitating, systemic, autoimmune disorder that can affect most of the joints in the body. Treatment has dramatically improved because of new drugs that block the actions of tumor necrosis factor (TNF). These drugs need to be delivered by frequent injection or infusion. The two that concern us here, Enbrel and Humira, are normally self-administered by subcutaneous injection. Although many patients mount an impressive therapeutic response to these drugs, the effect can be incomplete, so that a few joints remain symptomatic.
The gene therapy agent in question (tgAAC94) consists of a recombinant adeno-associated virus containing an Enbrel gene. This vector is injected directly into symptomatic joints so that they produce the Enbrel protein within the affected tissues. Ideally, gene therapy would free patients from the need to administer the protein through frequent, repeated injections.
Mohr, who had suffered from rheumatoid arthritis for 15 years, received a second injection of tgAAC94 into her right knee on July 2, 2007. Twenty-two days later she was dead withhistoplasmosis (a fungal infection) and a puzzling leakage of blood into her abdomen associated with a massive hematoma that compressed and impaired several organs, particularly her kidneys and lungs.
As someone who has worked on arthritis gene therapy since its inception, I am obviously a supporter of this endeavor. However, there are several issues with the present trial that need to be aired.
One of the most controversial is that anti-TNF genes were introduced into the joints of subjects who were already receiving anti-TNF proteins. At the time she received tgAAC94, Jolee Mohr was also taking Humira and two other anti-inflammatory and immunosuppressive drugs. According to the sponsors of the trial, Targeted Genetics, individual joints that remain symptomatic in spite of circulating anti-TNF proteins may benefit from additional, gene-mediated TNF blockade within the joint. This may be true, but it is hard to find supporting evidence for this assertion in the refereed literature.
Another concern is that, even if the gene therapy were working, it is would be very difficult to confirm because of the trial’s design. When subjects in early phase trials are taking multiple drugs, it is hard to know which of the drugs is responsible for any changes that occur, especially in rheumatic conditions with spontaneous flares and remissions. On top of that, the most commonly used scoring systems for assessing rheumatoid arthritis, the American College of Rheumatology and Disease Activity Scores, measure global changes in large numbers of joints and cannot be applied to an individual joint. Other confounding circumstances in this trial are the options to include patients with diseases other than rheumatoid arthritis and to inject any of five different joints.
Muddying the waters even further is the fact that Jolee Mohr had been on Enbrel previously, but stopped taking it after two years when her arthritis flared up again. If the Enbrel protein had stopped working, as may have been the case, then administering the Enbrel gene would be pointless.
Our understanding of this case would be much improved if we knew how much Enbrel protein was expressed in patients receiving tgAAC94. Given that gene therapists agonize over the level and duration of transgene expression within the host, it is remarkable that we still don’t know how much Enbrel was present in the joints of patients in this study. According to Targeted Genetics, these values are difficult to obtain because the company uses a TNF binding assay to measure transgene expression, and this assay is unreliable in the presence of high background Humira which also binds TNF. In 2003, when the company sought RAC’s approval of an earlier, Phase I study, it was developing a specific ELISA assay that would permit such measurements; indeed, determining Enbrel protein levels in sera and joint fluids was one of the secondary goals of that study. But apparently the subsequent Phase I/II study, which was exempt from public RAC review, went ahead without these endpoints having been satisfied.
When this sort of gene therapy approach to treating arthritis was first suggested over 15 years ago, it was seen as a safer, longer-lasting, more effective and less expensive alternative to the repeated, systemic administration of recombinant proteins. Because arthritis is a nonlethal, nongenetic disease, using gene therapy to treat it has always been controversial. The recent tragedy will do nothing to alleviate the controversy, but it should not be allowed to doom the entire field. Rather, it emphasizes the importance of clinical trial design, execution, and oversight.
Chris Evans is the Robert Lovett Professor of Orthopaedic Surgery at Harvard Medical School.