SENS is hard , yes , but not too hard to try : a reply to Warner

Elsewhere in this issue, Huber Warner writes a critique of my “Strategies for Engineered Negligible Senescence” (SENS) proposal for the dramatic postponement of age-related functional decline. His commentary advances the debate on SENS to a proper level of measured discourse, something that has been regrettably lacking hitherto, and for this I applaud him. However, his conclusion that SENS is currently so unlikely to succeed that we should not pursue it is, I argue here, incorrect, being based on a range of factual and logical errors. Warner's contribution to the SIAAR 2005 forum in this issue is most welcome, not least for its departure from the intemperate tone that has recently characterised public criticisms of SENS. The fact that it is Warner who has raised this debate to the appropriate level of decorum is especially significant given that he was the corresponding author of a previous, unedifyingly strongly-worded critique. In my opinion, however, his concerns about the efficacy, safety and feasibility of SENS are misplaced, since they are based on several important errors of logic, fact and interpretation. Efficacy. Warner begins by saying that a repair (as opposed to prevention) strategy such as SENS requires us to know how much damage is tolerable. I can see no basis whatever for this assertion, which is made without any supporting argument and which Warner thus presumably feels is selfevident. I contend that in fact the opposite is self-evident: specifically, that the only substantial influence that this threshold will have on what strategy works is with regard to the frequency and/or thoroughness with which the strategy in question is implemented. Warner is also wrong when he suggests that we do not know how much damage is tolerable: we indeed know this to an adequate approximation, simply because we can measure the levels of each category of damage in people of various ages and hence of various implied remaining life expectancies. This is imprecise, certainly, but since SENS's goal is to maintain damage at youthful levels (i.e., much lower than life-threatening levels), it is accurate enough. The comprehensiveness of the seven SENS categories is a central plank of my argument that SENS can succeed in the timeframe I have predicted, which is why I have presented highly detailed arguments showing that, in contrast to how they are generally interpreted, available data strongly support the conclusion that non-oncogenic nuclear DNA damage is harmless in a currently-normal lifetime. However, these arguments are only now in press, so Warner is justified in this criticism of SENS as hitherto published. In regard to types of protein damage other than those which SENS enumerates, I claim that the null hypothesis is that these phenomena do not contribute to age-related decline, so long as no evidence has come to light that implicates them. But my key point here is this: if any of these other phenomena do in fact contribute appreciably to aging, the fact that it has for so 2 long been impossible to generate strong evidence for that contribution suggests strongly that the most efficient way to obtain such evidence will be to unmask it by removing the types of damage for which such evidence already exists: in other words, to implement SENS. Warner also repeats a remarkable inference that has been made and refuted previously: that the lack of life extension effects of any SENS strand in isolation constitutes evidence that they will also fail in combination. Since the only interventions that have to date reproducibly increased the maximum lifespan of any mammal (excluding short-lived strains) are merely elicitations of the organism’s evolved response to nutrient deprivation or genetic emulations of that response, it seems quite reasonable to conclude that the time has come to explore more complex, composite interventions if we wish to postpone aging more than such elicitation can achieve. Finally in this section, Warner highlights the overoptimism of my prediction in October 2000 that all 13 mitochondrially-encoded proteins would be successfully expressed from nuclear transgenes (“allotopic expression,” AE) in cell culture within five years. As I have explained elsewhere, my overoptimism on that score was purely with regard to whether serious efforts to engineer such transgenes would begin soon. I anticipated at that time that a recently-conducted demonstration of successful allotopic expression of one of these genes would imminently be published in Science, where it was in review, but in fact, and for reasons unrelated to the quality of the study, it was only published four years later (in, I am proud to say, this journal). AE has long been considered realistic by many mitochondriologists but overambitious by funding agencies, a disconnect that a high-profile report of success might have abruptly altered. From a scientific standpoint, however, no new reasons have emerged to suggest that AE is infeasible, so I still claim that it can be achieved within five years of the initiation of adequate funding. I therefore disagree with Warner that my erroneous prediction implies anything about the futility or otherwise of trying to predict the time course of scientific progress. Safety. SENS does not discuss germ-line gene therapy, so Warner’s point that such therapy is ethically problematic at this time is not relevant to SENS. Somatic gene therapy indeed has a long way to go, but, in view of the continued and accelerating effort to advance it, and especially since gene therapy (even with non-randomly-inserting vectors) already works rather well in mice, its current limitations are hardly reasons not to develop the appropriate cargo in anticipation of such advances. Rather, we should consider how foolish and guilty we will feel if breakthroughs in gene delivery are made and we have not begun to develop the constructs that would be worth delivering. Warner then discusses WILT, the SENS strand proposed to address cancer. However, in rejecting it he simply lists aspects of it that have been thoroughly addressed in my publications and offers no challenges to my analysis given there. Thus, he gives no specific reasons for doubting my claim that it can be implemented safely and effectively within a few decades. In my view, to call confidence in WILT a “leap of faith” on the basis of such little analysis is itself a leap of faith. Feasibility. The fact that SENS is a repair and maintenance strategy gives it an advantage for testing in long-lived mammals that Warner apparently overlooks: namely, that effective interventions of that type can yield positive results in a small fraction of the organism’s lifespan. SENS, if it works, should greatly reduce the mortality rates of organisms to which it is applied, even when those organisms have had no treatment before that time. It is thus realistic to expect that statistically significant results in primates with a life expectancy of, say, 30 years could be obtained after only a year or two, using animals that were already approaching their life expectancy at birth when treatment began. Biomarkers of remaining life expectancy (whose relevance for assessing the efficacy of treatments is in any case questionable) are therefore not required. Warner’s remarks concerning human trials are also short-sighted, in my view, since they presume that such trials would be conducted in a regulatory environment resembling today’s. I have argued that this is in fact unlikely: that public attitudes to the risk of experimental treatments will probably shift dramatically when the prospect of defeating aging becomes widely appreciated. Moreover, the 3 “escape velocity” of subsequent SENS development (the rate at which the SENS components would need to be improved in efficacy in order to deliver the same degree of life extension as if they were perfect) has recently been shown to be quite modest: roughly a doubling of efficacy every 40 years. Warner concludes that, since (in his opinion) SENS is so unlikely to work, we should not allocate resources to it but instead continue to focus our resources on funding additional exploratory research. There is indeed a threshold probability of success below which it would be inappropriate to pursue a venture so ambitious and expensive as the SENS agenda; I have never disputed this. However, one can only assess accurately whether SENS’s probability of success exceeds that threshold by examining its details closely. The shortcomings of Warner’s analysis reinforce my belief that the overwhelming majority of those who have expressed pessimism concerning SENS’s prospects have yet to study my publications (and, most importantly, the relevant experimental work that I cite) thoroughly enough to make such an assessment.