Are senolytics damaging for long-term health?
A conversation with Dr. Michael Fossel about the potential downsides of senolytics and the promise of telomerase therapy
(I had some assistance from LEAF’s Steve Hill in the first part of this interview)
Dr. Michael Fossel has been focused single-mindedly on telomeres and telomerase therapy for decades now. He worries that the field of longevity science’s new infatuation with senolytics may be damaging if researchers don’t fully take into account not only positive short-term results from senolytics in various model organisms, but also the long-term results — which he argues can be quite damaging.
He makes this case in a 2019 review paper and also touched on it in my interview with him last year here. Fossel’s 2004 textbook, Cells, Aging and Human Disease (Oxford University Press) also covers some of these issues and the relevant biological context of his concerns.
Fossel’s new review presents a useful schematic showing the causal pathways of cellular senescence and potential interventions. Telomerase therapy, and other possible means for re-lengthening telomeres, is, he argues, the most upstream and potentially most effective means for reversing cellular senescence and thus aging more generally.
I asked to interview Fossel about his concerns because so many researchers in this field are now working with senolytics in various ways and few seem to be considering the potential for long-term damaging effects from elimination of a significant number of cells from various tissues (which is the direct and desired impact of senolytics).
I’m not entirely convinced by the available evidence or Fossel’s arguments here that we know the long-term effects of senolytics. However, the cautionary argument certainly makes sense and it is reasonable and logical to prefer rejuvenation of senescent cells, if possible, rather than killing them (with senolytics). The question then becomes: do we have or will we soon have reliable rejuvenation therapies?
We conducted the below interview via email.
Your new review on senolytics suggests that senolytics may cause more harm than good. Can you summarize your objections and concerns?
Here is the argument: 1) theoretically, senolytics should make things worse and 2) the available data support this theoretical concern. To use an analogy, imagine that you have a factory in which 10 of the 100 factory workers are feeling overworked and tired. Furthermore, their complaints are disrupting the other workers. You have two possible interventions. You can:
1. Fire the 10 workers, thereby removing the complainers. The result is that the remaining 90 workers are now overworked and they too begin to complain. You end up with 30 workers who are now complaining and disrupting your factory. This is the senolytic approach.
2. You could improve the health and conditions of the 10 workers who are overworked and complaining. You now have 100 workers who are doing an excellent job. This is the telomerase therapy approach.
In the first case, your factory has a problem and you make it worse. In the second case, your factory has a problem and you solve the problem.
This figure from my new paper illustrates the same point in terms of nine cells subjected to senolytics, with the result being temporary short-term improvement followed by decline and a worse situation than we started with.
Ok, but this does not take into account replacing that pool of “workers” by bringing in fresh stem cells, as proposed in the Hallmarks of Aging approach, so if you keep on draining the pool and not topping it up with new cells, it seems that your stated concerns could pose a risk. However, use of senolytics in isolation is not the long-term solution proposed either in the Hallmarks of Aging or SENS theoretical frameworks. Rather, the idea is to periodically remove the senescent cells and replace them with healthy working cells. Does this two-step “remove and replace” approach reduce your concerns?
You have to keep a few points in mind. 1) Will the stem cells populate as desired? 2) If you do get stem cell population, that requires cell division, which shortens telomeres, which accelerates cell senescence, and once again you have accelerated pathology. 3) Why would you bother recruiting stem cells when you can much more easily reset cell senescence in the resident cells of the tissue? 4) The long-term data (what there is of it) support the failure of senolytics. I would not rely on the flaws of the Hallmarks of Aging paper, let alone SENS, as reliable sources of clinical prediction.
Again: remember where those “new cells” come from: you are accelerating senescence in the stem cell pool. The only way to “replace them with healthy working cells” is to simply and effectively reset gene expression, taking senescing cells and turning them into functionally young cells. This was first demonstrated 21 years ago in human cells in vitro, then repeatedly in human tissues in vitro.
What gives you such confidence that telomerase therapy will work when there are almost no clinical trials or even pre-clinical trials completed with telomerase therapy on humans at this point (unfortunately), and nothing that has shown significant telomere lengthening?
My new editorial presents the relevant data. There have been repeated studies on human cells, human tissues, and animals (in vivo). In addition, the telomerase activator papers support this position. Telomere re-lengthening has been demonstrated repeatedly and with predictably positive in vivo results in cells, tissues, and animals.
Also, we know from the work of Irina Conboy that elements in the inflammatory factors such as excessive TGF-beta (present in SASP) inhibit stem cell mobility, including blocking TERT and thus telomerase. The argument could be made that this chronic inflammation is worse versus leaving the senescent cells in situ because it impairs tissue regeneration, and removing those pro-inflammatory factors should improve tissue regeneration. We also see this reflected in animal studies where increased median lifespan and healthspan have been demonstrated following senolytic treatment. So if we have to choose senolytics or simply leaving the senescent cells in situ, is not removing them the lesser of two evils?
While telomerase can rescue senescent cells from cell cycle arrest, isn’t it the case that many senescent cells are that way for a reason (DNA damage, mutations, and so on), which cause the cells to shut down? The concern here is if you bring damaged cells back online using telomerase you are potentially allowing broken cells with aberrant behavior to re-enter the cell cycle. Is this not an issue?
You are confusing cause and effect. This is a common mistake, unfortunately, and continues to result in the muddled thinking that results in poor clinical results when people take small-molecular approaches to age-related disease. I refer you to Maria Blasco’s recent paper on this issue (Muñoz-Lorente, et al. 2019), as well as the first five chapters of my 2004 textbook.
Is there a role in long-term human healthcare, as longevity science progresses, for one or two rounds of senolytic treatment to stop the damage from stem cell SASP, even if we are irreversibly eliminating some cells that would otherwise perform positive roles if they were normal and healthy? In other words, even if senolytics shouldn’t be considered long-term parts of our rejuvenation program, could one or two uses provide long-term beneficial care combined with other programs?
No. In all cases, you simply accelerate the pathology. It makes no sense because the alternative — telomerase therapy — is effective, beneficial, and conclusive. To use another analogy, asking if you should use senolytic therapy “for one or two rounds” is like asking how to treat an infection: should you use one or two rounds of blood-letting or should you use an antibiotic. The use of senolytics is counter-productive and naïve.
What telomerase therapy are you referring to? Are there any commercially available telomerase therapies available yet?
Telomerase therapy (generically) refers to the ability to add, activate, or express telomerase in living cells. See my A&D paper for newer details, but the topic was reviewed in my textbook 15 years ago. Telomerase therapy (specifically) refers to the addition of an hTERT gene via viral vector resulting in active telomerase function. While there are commercially available telomerase activators, there are no commercially available telomerase therapies yet.
Would senolytics make sense for progerics or normal elderly who need help now and can’t afford to wait for what might be better solutions?
No. Again: both a knowledge of human pathology and attention to the data suggest that senolytic use will prove to accelerate disease. This will be particularly true in progeria.
Which studies in particular support the assertion that senolytics accelerate pathology, given the well-documented role of senescent cells in multiple age-related diseases?
The role of senescent cells is not only well-documented, but (as you well know) I am the only one in the world who has published a textbook with that as its primary thesis. Read it. As for the first half of your statement, I again refer you to my published paper on senolytics.
Justice et al 2019, the first pilot study results with senolytics in humans, describes a DQ pilot study with 14 idiopathic pulmonary fibrosis (IPF) patients, a prelude to phase 1 clinical trials. Their team is already testing the same treatment on 15 lung cancer patients and planning a similar study on 20 kidney disease patients. What are your thoughts on these results? How do they fit with your suggestions that senolytics may be harmful in the long run even if they provide some short-term improvements in various ways?
If they follow their patients over a period of months to years, they will find that they have accelerated aging and age-related disease in the affected tissues. To use another analogy, using senolytics in age-related disease is like using steroids in acute infections. The initial result may look impressive, but (whether using senolytics in age-related disease or steroids in acute infections) the morbidity and mortality will prove to be tragic.
It seems that we can only speculate on these issues, as these long-term follow-ups have not yet been done, right? However, we can look at mouse studies to give us some idea and senolytics have been shown to increase median lifespan and healthspan in murine models. Given that senescent cells are a common mechanism in both mice and humans and are a proposed secondary cause of aging, the consequence of primary hallmarks (DNA damage, epigenetic alterations, telomere attrition, and loss of proteostasis), it doesn’t seem unreasonable to think the results in humans would be similar. Given all that, why would removing senescent cells be significantly different in humans than it is in mice?
I don’t see any credible data that supports the contention that “senolytics have been shown to increase median lifespan and healthspan in murine models”.
What about the Baker paper you cited above, which argues that “senescent cell clearance extends lifespan”?
I don’t see that the Baker paper data is sufficient support for their claims. Their approach may increase median (but not maximum) lifespan in murine models, but the acceleration of the curve shows that pathology increases after an initial “honeymoon period”
Are you suggesting that while median lifespan may be increased that healthspan isn’t?
Yes, but not the maximum, i.e., it has no effect on the fundamental pathology involved. More importantly, the curve shows accelerated pathology. The most they have done is to “rectangularize” the curve, rather than affect aging per se. Beyond that, I suspect that additional animal and particularly human data will not support even the median lifespan increase.
Doesn’t their data show in each scenario that median lifespan increased? Isn’t that all we need to know if we’re asking the question “did this intervention increase median lifespan?”
We can do better than that. There are two problems. The first is that the pathology will recur with a vengeance as they accelerate cell aging. The second is that almost anything (e.g., good diet, seatbelts, immunizations) extends median lifespan, but none of them extend maximum lifespan. Maximum lifespan CAN be extended via telomerase therapy.
There seem to be various ways to revert SCs back to youthful type, including various NAD+ pathways, as described in this recent paper (Neelakantan et al. 2019) on mouse SCs that revert in the presence of NNMT inhibitors. Are all of these pathways ultimately pointing to telomere re-lengthening as the mechanism of action for reversion or are there various different ways to achieve reversion?
No. What you’re seeing are partial effects based upon interventions that address one of several dozen “downstream” biomarkers rather than an “upstream” intervention that modulates all of those downstream biomarkers. In short, you are treating effects, not causes. To use a musical analogy, if an orchestra was playing Mozart and is now playing a John Cage piece, you could go around to each and every musician and one-by-one instruct them to ignore the conductor (the telomere) and play their own part of the Mozart composition, but the most effective and efficient way would be to simply ask the conductor to change scores and use the previous Mozart score.
Moreover, the use of small molecular approaches (such as NAD+) incurs increased risks of adverse effects with the result that (once you use dozens of small molecules to change each separate issue) your risks of side effects and interactive side effects is severely detrimental. The use of a single, large-molecular approach — which is inherently specific — avoids such pharmaceutical side effects, avoids interactive side effects, and addresses the entire range of pathology at the root cause.
Many researchers have recently become excited about the potential for partial cellular reprogramming with enhanced Yamanaka factors (Ocampo et al. 2016, e.g.). What do you think of this therapeutic option? Does it overlap in any way with telomerase therapy in terms of the underlying mechanisms? Or is this another “downstream” mechanism?
The Yamanaka factors are downstream and only a partial invocation of what occurs when we reverse cell senescence.