In this issue of the journal, Leonard Levin, MD, PhD has proposed a potentially unifying hypothesis regarding the pathogenesis of 3 apparently disparate optic neuropathies that are characterized in part by cecocentral scotomas: Leber hereditary optic neuropathy (LHON), vitamin B12-deficiency (nutritional) optic neuropathy, and ethambutol-related (toxic) optic neuropathy (1).
This hypothesis, based on theoretical grounds as well as both in vitro and murine experiments, is that these diseases have, as a common final pathway, the generation of superoxide free radicals within retinal ganglion cells (RGCs), leading to RGC death. Dr Levin further hypothesizes that because of their small size compared with fibers from RGCs elsewhere in the retina, the axons comprising the papillomacular (PM) bundle are more likely to be damaged from superoxide free radicals than their larger counterparts. This occurs because, according to Sadun (2), the energy expenditure of an axon is related to its surface area, whereas its content of mitochondria is constrained by the cell volume. Thus, the small P-cell axons of the PM bundle, which have the smallest ratio of volume to surface area, have the least margin for error in the setting of energy depletion and, hence, may be at the greatest disadvantage in energy dependence for maintaining efficient axon transport. Dr Levin also implicates the anatomy of the crossing fibers of the chiasm, emphasizing that these fibers may undergo a small deformation as they cross under and over each other, rendering them more susceptible to metabolic compromise than their noncrossed counterparts. He concludes his manuscript by emphasizing the speculative nature of his hypothesis (it is, after all, a hypothesis) and the need to validate it.
Dr Levin's hypothesis is complex, implicating at least 3 different factors for the development of these 3 apparently disparate optic neuropathies: the development of superoxide free radicals, and the susceptibility of the small axons of the PM bundle as well as the crossing fibers in the chiasm to metabolic and other forms of stress. The experiments that he and his colleagues have performed provide circumstantial evidence supporting this hypothesis. So how do we validate it?
The first decision to be made is whether data from an animal model more closely related to humans are needed. Data obtained from many animals (e.g., rat, mouse) may or may not be pertinent to human disease as these animals often have different responses from humans to inflammation and other types of insults (3). For example, substances that showed significant experiment promise in providing neuroprotection in rats after cerebrovascular insults failed to provide the same protection in humans. The same issues may apply to the use of oxygen free radical scavengers. Therefore, it may be more appropriate to test the hypothesis in nonhuman primates, the optic nerves of which are more closely related to those of humans, and the responses to various insults are more similar to those of humans than those of rats or mice.
The second consideration is how to plan a clinical trial (4,5). Clinical trials first require an assessment of drug safety. Only when safety has been established, trials with different concentrations of drug can be evaluated for efficacy. Thus, the first issue is whether a drug already approved for human use is available. If so, that drug already has a safety profile, and it may be used in a trial without a major safety arm, assuming that it is to be given in doses and routes for which it has already been approved. If the drug has not been approved for human use, it will require an assessment of its safety (Phase 1 trial). This, in turn, requires a statistical analysis to determine the number of subjects required to determine with a high probability that the drug is safe. One of the major errors of some clinical trials is that a statistician is consulted after the trial has been performed. In such cases, the trial results may be spurious because the drug has been evaluated in an inadequate number of subjects. Similar considerations apply when assessing the efficacy of a drug. One must consider the number of subjects required to provide adequate information regarding efficacy, whether assessed (with or without dose escalation) against a known treatment or a placebo (Phase 2 and 3 trials). This number is dependent on many factors, particularly the natural history of the disorder. For example, patients with LHON who harbor the 11778 mutation in their mitochondria have, at most, a 4% likelihood of spontaneous visual recovery. Thus, fewer subjects would be needed for a clinical trial of a superoxide radical scavenging drug than would be needed for, for example, a clinical trial of a drug for the treatment of nonarteritic anterior ischemic optic neuropathy (NAION), which has about a 40% spontaneous rate of visual recovery of 3 lines or more. However, just because a drug works in a group of subjects with a specific disorder does not mean that it will work in another. Dr Levin postulates that LHON, vitamin B12-deficiency optic neuropathy, and ethambutol optic neuropathy have a common pathogenesis. To prove this, not only does a particular drug or target need to provide benefit in patients with LHON but also in patients with the other 2 optic neuropathies. Proving his hypothesis will require more than a trial in patients with LHON.
This brings up a final consideration: a clinical trial must be economically feasible. One always hopes that the results will save society money in the long run, either by identifying optimum treatment for a previously untreatable or poorly treatable disorder or by showing that a current treatment either is ineffective (think systemic corticosteroids for NAION) or even potentially harmful (think optic nerve sheath fenestration for NAION). Several years ago, at the request of a pharmaceutical company, several members of the Neuro-Ophthalmology Research Disease Investigator (NORDIC) put together a proposal for a clinical trial of a substance to treat patients with LHON. When the company was shown the proposed budget necessary to perform the study in what the NORDIC committee thought was an appropriate manner, the company's executives decided that the study was too costly for what it would accomplish.
Ultimately, evidence-based medicine clearly is the best way to provide optimum medical care for our patients, and in most cases, the evidence is best derived from controlled clinical trials. Although such trials are not always straightforward and may be extremely costly, this is what translational research is all about.
1. Levin LA. Superoxide generation explains common features of optic neuropathies associated with cecocentral scotomas. J Neuroophthalmol. 2015;35:xx.
2. Sadun AA. Acquired mitochondrial impairment as a cause of optic nerve disease. Trans Am Ophthalmol Soc. 1998;46:881–923.
3. Gladstone DJ, Black SE, Hakim AM; for the Heart and Stroke Foundation of Ontario Centre of Excellence in Stroke Recovery. Toward wisdom from failure. Lessons from neuroprotective stroke trials and new therapeutic directions. Stroke. 2002;33:2123–2136.
4. Friedman LM, Furberg CD, DeMets DL. Fundamentals of Clinical Trials, 4th edition. New York, NY: Springer, 2010.
5. Piantadosi S. Clinical Trials: A Methodologic Perspective, 2nd edition. Hoboken, NJ: John Wiley and Sons, 2005.