CAR Accidents: Unexpected and Serious Toxicity in Gene Transfer Immunotherapy

This month's issue of Molecular Therapy- the premium journal covering developments in gene transfer- reports two deaths in recent cancer gene transfer studies. Both studies involved a similar anti-cancer strategy, in which a patient's T cells are genetically modified to mount a strong and sudden immune attack against the patient's cancer (the particular genetic modification is known as "CAR," for chimeric antigen receptors). Both were phase 1 studies. Both patients died from what looks like "cytokine storm"- the same phenomenon that caused life threatening toxicity in the Tegenero TGN1412 study in 2005. In one case, the authors attribute death to the gene transfer; in the other, the authors categorize the death as possibly related to the gene transfer (the latter was previously described at ASGT in 2009).

In all likelihood, these patients (or at least, one of the patients) will be the third or fourth death in gene transfer that is clearly attributable to gene transfer. Don't expect too much public hand-wringing or media coverage, however: in both cases, the patients were adults and had terminal cancer. I have not made a careful study of these particular trials or the strategies they employ. So the following thoughts about these deaths should be read with caution:

1- Unpredictability: These deaths point, once again, to the unpredictability of strategies aimed at training the immune system to respond against tumors. Immune systems are notoriously difficult to model in animals, and as a result, every human study is essentially a shot in the dark. The authors of one of the reports sagely urge that phase 1 studies using similar strategies begin at low doses.

2-Where's the Toxicology? Neither report mentions anything about observing similar toxicities in preclinical studies. Indeed, neither report even mentions preclinical toxicology studies. One wonders why: were they done? how were they done? what was observed? For example, both studies involved immunosuppression co-interventions aimed at enhancing the effects of the T-cells (in one case, administration of the drug cyclophosphamide; in the other, use of nonmyeloablative conditioning). Were toxicology studies performed in animals receiving these immunosuppression treatments?

3-Did Investigators Give Preclinical Studies Their Best Shot at Producing Similar Toxicity? Both phase 1 studies were supported by preclinical studies using mice that lacked functional immune systems. One has to wonder how useful it is to test immunotherapies in mice that lack properly functioning immune systems. From what I can tell, in neither the first nor the second case did investigators perform preclinical studies that simultaneously delivered modified T-cells and immunosuppressive drugs.

4- Please: No More Gratuitous Appeals to the Integrity of the Investigators. An editorial by a leading expert on CARs accompanies the reports in Molecular Therapy and provides a very helpful summary and context of the events. It ends, however, with the statement "it is a great credit to all investigators involved that they have been so forthcoming in providing detailed reports of serious adverse events." I heard similar sentiments expressed when one of the deaths was presented at a scientific meeting last year. True- the research team did provide an unusually extensive report and investigation, including autopsy. However, careful and public reporting of serious adverse events is exactly what researchers are supposed to do in phase 1 studies involving highly innovative approaches; praising them for coming forward with this kind of information is a bit like congratulating Canada every time it holds a democratic election. One has to wonder whether there is a reserve of trial deaths that are never investigated or reported. (photo credit: Steve Kay 2008)

Testing Testing...: Personal Medicine, Breast Cancer, and Policy

Personalized medicine is supposed to usher an era in which treatments are tailored to individuals. And HER2 testing has long been seen as heralding the promise of personalized medicine: tumors that test positive for an amplified HER2 gene are more likely to be responsive to drugs, like trastuzumab, that block the HER2 receptor.

Some may see HER2 testing as foreshadowing a perfect future in which treatment decisions are coupled to molecular diagnostics. But Gina Kolata in the New York Times instead tells a story of shadows. Like all medical tests, HER2 testing is error prone: some tumors test positive when they are in fact negative, and others test negative when they are in fact positive. And some results are just plain ambiguous, with parts of the tumor being positive and other parts being negative. Kolata describes the challenges that women and their physicians confront when interpreting test results.

Most troubling in this story is the role, or lack thereof, played by regulatory agencies like the FDA. Quoting Kolata: "there is a proliferation of laboratories offering tests without F.D.A. oversight. But, for now, the agency has no specific plan to regulate the tests, in part because of lack of money." If FDA is not prepared to regulate tests because of resource constraints, and prescription decisions are likely to be increasingly coupled to diagnostic tests, it logically follows that FDA is not prepared to regulate the approval and use of newer generation, test-based drugs. In other words, FDA seems unable to establish the validity of labeling indications for drugs that rely on diagnostic tests. This can't be a good thing for patients, physicians, or third party payers (but is great for the makers of drugs and diagnostics, who thrive in this kind of clinical and regulatory uncertainty!) (photo credit: crafty dame, breast cancer cells, 2009)

Teaching Kills Blogging: Somewhat Recent Developments...

Dear Faithful Readers: Teaching has cut my blogging to a trickle, though the teaching has now begun to taper off. My silence is not for want of major developments in the last two months. Among a few highlights:

• Obama picks members for his Bioethics advisory panel: White house recently announced membership of its "Presidential Commission for the Study of Bioethical Issues." The group is smaller than past Presidential panels. Its membership is lean on working bioethicists (3 or 4 who clearly fit the classic definition-- all others scientists, clinicians, federal employees, university administrators, or a disease advocate).

• Health care reform (+ Translational Research) passes in the U.S.: Among the intriguing elements here is the relationship between reform and biomedical research. When Clinton proposed healthcare reform in the 1990s, there was much consternation in the research community that this would spell a retreat from investment in basic research. Indeed, failure to enact reform propelled a massive expansion of the NIH budget through the 1990s. This time around, healthcare reform has specifically integrated basic research. The law includes language creating a "Cures Acceleration Network" that would fund up to $15M/year in translational research (though the budget will depend on direct appropriation from Congress, and there is no certainty that it will be funded).

• Gene Patents Voided: Following an ACLU challenge, a U.S. District Court Judge threw out Myriad Genetics' patent on BRCA1 and BRCA2 (genes associated with hereditary breast cancer; the company markets a $3K per pop test for mutations in the genes) by ruling that the genes are "products of nature." Products of nature are not patentable, though products purified from nature (e.g. enzymes, wood chemicals, etc.) are. The logic behind the decisions is that genes are better thought of as information rather than as chemicals, and that information extracted from the natural entities does not have distinct properties in the way that chemicals do. If ever there were a demonstration of the power of metaphors; suffice it to say, biotechnology companies will appeal. (photo credit: aurelian s 2008)

Ark, Troubled Waters, and Rainbows for Gene Transfer

This morning I awoke to a news report by National Public Radio's Joe Palca on promising developments in gene transfer. In it, Palca provided a good account of the field's travails, as well as some encouraging developments in the last few years. The story ended with the prediction that the coming "months and years" would bring landings for more common disorders like AIDS and cancer.

Coincidentally, the just released March issue of Nature Biotechnology ran a report on a front-runner for gene transfer commercialization: biotechnology company Ark Therapeutics gene transfer gliobastoma product Cerepro. The application for licensure of this product in Europe was unsuccessful (press release here). Recall that, last June, I described what seemed like unimpressive results from a phase 3 trial that were reported at an annual meeting of the American Society of Gene Therapy. Apparently, European drug regulators weren't impressed either (they cited flaws in trial design, including a small sample size and unconcealed allocation; Ark has asked the agency to re-examine their application).

But for those awaiting the first commercialization of a gene transfer product in a country with a robust drug regulatory system, there is still some indication that the rains may be subsiding: according to the report in Nature Biotechnology, Amsterdam Molecular Therapeutics has filed with EMEA for marketing authorization of their AAV product for a rare hereditary disorder, LPL deficiency; the company will soon file in Canada as well (the disorder is more prevalent in Quebec) (photo credit: Occhiovivo 2007)

Canada Human Research Ethics Policies: Take 2

In the Vancouver Olympics, Canada is tied for the most gold medals as of this writing. Will Canada also "own the podium" when it comes to providing a clear and effective voice for ethical human research?

Recall that, in a previous post, I mentioned that Canada was presently undertaking a major overhaul of its main research ethics policy- the Tricouncil Policy Statement. After issuing an intital draft, the panel charged with revising the document (PRE) released a second draft. In a few days, the comment period closes for this second draft. The policy will then be revised again and finalized.

I will be submitting comments on the new version, along with my research collaborators. Here are a few important problems that carry over from the previous version:

1) The draft, like the old version, does not quite get the foundations for research ethics right. As a consequence, principles like respect for persons or justice are made subservient to beneficence. This is not an obscure philosophical point: it sends a message to investigators and research ethics committees that these other principles matter less.

2) The application of justice in studies involving economically or socially disadvantaged populations is somewhat muddled. The draft defines justice in terms of what it is not, or what researchers should not do, rather than providing an affirmative description of what justice entails. Consider what it would mean if the application of "respect for persons" merely meant that researchers should avoid enrolling patients who did not consent, or that researchers should refrain from studies if risks are unacceptable.

3) The language on undue inducement has problems: The TCPS language implies that concerns about undue inducement arise out of a concern for risk (see point 1 above). It doesn't. It is unethical to pay individuals to override certain moral commitments (in government, we call this bribery), even in the absence of risk to the individual.

4) The chapter on clinical trials sends conflicting messages, and seems to imply that it is ok for doctors and institutions to put patients at medical disadvantage by enrolling them in research. Hard to imagine, if this is accepted, how conscientious doctors could ever participate in research, much less refer their patients to studies. Of particular concern, the chapter on clinical trials seems like it is trying to accommodate the unethical standards established by the International Council on Harmonization's Good Clinical Practice.

I hope the panel can correct these (and other) flaws, while preserving the many qualities contained in the proposed revisions. (photo credit: pmorgan 2004)

Transplanting Autoimmune Research

What's the difference between testing a typical small molecule drug, and testing a novel cell therapy strategy? And where might the latter raise ethical challenges that the former doesn't? These questions are extensively discussed in my book, and given human drama in a recent story by Jennifer Couzin-Frankel in the Feb 12, 2010 issue of Science ("Replacing an Immune System Gone Haywire").

Couzin-Frankel describes the numerous difficulties that researchers have faced in attempting to validate autologous bone marrow transplantation for the treatment of (often nonlethal but highly debilitating) autoimmune disorders like type 1 diabetes, Crohn's disease, and multiple sclerosis. The idea of this procedure is to "reset" the immune system by purging patients of their bone marrow cells, and then returning healthy bone marrow to them. The approach has shown some promise for certain autoimmune disorders. However, response is highly variable and unpredictable, and validating and applying bone marrow transplantation for autoimmune disorders is beset by numerous ethical and logistical difficulties.

A major one is the risk-benefit balance: bone marrow transplantation requires exposing patients to the dangers of the transplantation procedure (6.6% mortality in one report of lupus patients). And yet, the procedures appear to work better in patients whose disease is not yet advanced. Testing the procedure therefore requires recruiting more or less healthy, at risk patients (sometimes children) into studies that expose them to serious risk of mortality. Clinicians understandably balk at referring their patients to such studies, making recruitment very difficult.

A second challenge is funding: many of these approaches involve using the patient's own bone marrow cells. There is nothing to patent-- and hence, little commercial interest in bone marrow transplantation for autoimmune disorders. This deprives this promising line of research needed resources.

And all this creates the perfect storm for a series of ethical challenges not directly addressed in this article (but covered in my book and articles): the siting of such studies in low and middle-income settings. Prohibitive costs, plus extreme difficulty recruiting patients who are otherwise eligible for somewhat effective and extremely expensive monoclonal antibody therapies, makes the siting of such trials in economically disadvantaged settings very attractive. This gives rise to what I have elsewhere called "expedient" justification for recruitment. Not surprisingly, then, one of the first trials of the procedure was performed in Brazil, and the article closes by mentioning that ongoing trials involving high-income country researchers are recruiting from São Paulo, Prague, China, and Argentina. This is good news if people in those settings have a reasonable prospect of having widespread and affordable access to bone marrow transplantation once it becomes validated. But it is troubling indeed if people in these countries will be bearing considerable burdens for the sake of knowledge benefits that will primarily (or most expeditiously) accrue to patients in high-income settings. (photo credit: Wellcome Images, Compact Bone, 2009)

Cooperation and Medical Research

Why do patients cooperate with medical researchers? So asks sociologists Mary Dixon-Woods and Carolyn Grant in a study analysis appearing in the June 2009 issue of Social Science and Medicine. You might think the answer is simple: they think they will benefit; they want to contribute to medical knowledge; or, they trust researchers who invite them. These are the simple and pat answers that have dominated the research ethics literature until now. However, Dixon-Woods and Tarrant probe deeper by asking why it is that patients feel safe and justified in joining research studies.

Using interviews from three different clinical studies, Dixon-Woods and Tarrant identify five recurring themes:

1- Research as Moral Act: Volunteers perceive participation as a moral act: they are willing to participate and cooperate only insofar as they are "able to defend the moral character" of their actions

2- Research as Risk: There was an awareness of risk and research scandals, and volunteers perceived a need to defend their choice to enter a study (rather than feeling like the choice was an obvious one that needed no explanation)

3- Trust in Regulation: Volunteers perceived that the research enterprise was regulated- that transgression of errant researchers would be "subject to punishment," though they had no familiarity at all with specific oversight structures

4- Signals of Trustworthiness: Participants sought "cues" and "signals" as to the values and trustworthiness of researchers. Informed consent, for example, was taken less as a substantive process than as a signal of the researcher's openness and integrity. Participants had strong expectations that researchers and affiliated institutions shared values of cooperation. And they saw the healthcare setting and affiliation of research personnel as "signaling" a kind of trustworthiness."

5- Trust of Professions: Building on item 4, participants made "swift" judgments about the trustworthiness of research personnel- not so much on the basis of their personal characteristics as their affiliation with trusted professions and institutions

The report helpfully contextualizes these findings within a broad sociological literature on giving, cooperation, and trust. Though the authors shy from offering specific prescription, two key themes emerge: oversight systems should focus on meeting these expectations; and regulation (whether external or internal to the profession), far from impeding research, creates social conditions in which patients can feel comfortable bearing risks imposed by strangers for the sake of strangers. (photocredit: Lucas 2008)