Editor’s Note: The U.S. Transhumanist Party features this article by our guest Steve Hill, originally published by our allies at the Life Extension Advocacy Foundation (LEAF) on May 9th, 2018. In this article Mr. Hill interviews Dr. Sarah Constantin, a researcher with a focus on machine learning at The Longevity Research Institute. This is an excellent article, especially if you want to learn more of the hard science behind longevity research. The topics of the interview range from deep learning being applied to pharmacology, to optimal mouse strains, and ideal areas of research to target age-related diseases.
~Bobby Ridge, Assistant Editor, June 30, 2019
Today, we have an interview with the Longevity Research Institute, a new group that launched in April 2018. The goal of the Institute is to identify therapies that can demonstrably extend healthy human lifespan by 2030 at the latest.
Searching for longevity
There are dozens of compounds and therapies that have been demonstrated to increase the lifespan of mammals. Recently, there have been some impressive examples of rejuvenation in animals using a variety of approaches, including partial cellular reprogramming, stem cell therapy, and senescent cell removal. More importantly, in many of these studies, age-related diseases have been delayed or even reversed.
Unfortunately, very few of these studies have had independent follow-ups or replication, and that is slowing down progress. The Longevity Research Institute is aiming to bridge the gap between basic science and commercial drug development.
It has chosen the field of aging research as its area of focus for one simple reason: age-related diseases are the leading cause of death globally. Heart disease, stroke, cancer, diabetes, Parkinson’s, Alzheimer’s and many more diseases are all caused by the various processes of aging.
The data from hundreds of animal studies tell us that aging is not a one-way process and that the rate of aging is something we can slow down or even reverse. Experimental results show that we can increase the healthy lifespan of animals significantly and delay the onset of age-related diseases in doing so. If we could translate those findings to humans, we could potentially increase the healthy period of life, known as health span, or even increase our lifespan beyond current norms while remaining healthy.
The majority of aging research consists of basic science that focuses on the mechanisms of aging, studies involving invertebrates like worms or fruit flies, and experiments that examine the effect of therapies on biomarkers of aging. However, the Longevity Research Institute believes that the way to find effective treatments that could translate to humans is by testing interventions on mammals to see if they increase lifespan or if they delay or reverse symptoms of aging, such as frailty, cognitive decline, and the prevalence of age-related diseases. Robust mammalian lifespan studies are quite rare in aging research due to their long duration and thus cost; the Institute believes they are worth doing despite this challenge.
Its philosophy is to be skeptical of results that depend on too many uncertain assumptions, such as particular mechanisms of aging or analogies between invertebrate and human biology. It believes that the closest way to measure the health and lifespan of a human is to measure the same things in mammals.
Replicating and Extending Lifespan Results
The majority of studies that have been shown to increase lifespan are rarely independently replicated to confirm the findings. There are therapies that, decades later, still have had no follow-up, and the Longevity Research Institute would like to change this situation.
To that end, it will be engaged in grant writing to obtain funds so that researchers studying aging will be able to conduct lifespan studies in mice and rats. The Longevity Research Institute also plans to commission its own studies and contract research organizations to carry them out.
It has a long list of promising interventions and is considering becoming involved with carboxyfullerenes, epithalamin, and stem cell transplants, for example. It is also interested in testing combinations of therapies to see if they have synergistic effects.
As translational research on aging is really a new, uncharted territory, the Institute is working with the Interventions Testing Program and METRICS to design reproducible animal studies. As part of that process, it will be testing genetically heterogeneous animals and using blind, randomized studies to reduce bias. A blind experiment is an experiment in which information about the test is hidden from participants, to reduce or eliminate bias, until after a trial outcome is known.
Best practices and transparency
Establishing best practices and protocol for translational aging research is a top priority here, and its work could help set the stage for future translational efforts. If superbly designed research protocols can be designed and made accessible to everyone, then they could be a real help in standardizing aging research and ensuring that the quality of results is the best it can be.
As part of its commitment to transparency and knowledge sharing, a condition of funding projects is that all experimental data will be made freely available to the public, as will pre-registration of experimental designs. The Institute will further protect this open science initiative by using blockchain technology to make immutable, publicly accessible records of everything it does.
We had the opportunity to talk with Sarah Constantin, Ph.D. and one of the key figures at the Longevity Research Institute, about their work. Sarah is a data scientist specializing in machine learning.
Your group believes that we need to conduct lifespan studies in mice in order to confirm that something might translate. However, some researchers believe that using multiple biomarkers of aging allows them to project, within a reasonable margin of error, changes to potential lifespan. This is becoming more relevant as the accuracy of biomarkers, and the use of comprehensive biomarker panels, becomes more commonplace. How do you respond to this?
There’s some very interesting stuff going on with biomarkers of aging. We’re able to predict mortality with AUCs of 0.8-0.9, which is quite good, with aging biomarkers, including things like blood panels of inflammatory and metabolic markers, DNA methylation, and phenotypic markers such as BMI and frailty. Some of these biomarkers are things we’re planning to measure in our animal studies, and they should give us interim results on whether the interventions we’re testing affect the predictors of aging. I still believe that we can be most confident in whether a treatment promotes longevity when we’ve tracked its effects throughout an organism’s lifespan. We do know of examples (such as calorie restriction in primates) in which it’s equivocal whether the treatment extends lifespan but it clearly improves age-related biomarkers, and you have to do a lifespan study to distinguish those cases.
Advances in deep learning and systems pharmacology are allowing us to project interactions and potential therapies far more efficiently than ever before. What are your thoughts on these approaches, and will you be looking to use them in your work?
The deep learning and systems pharmacology approaches are actually where I started in biotech; I did machine learning at Recursion Pharmaceuticals, which is taking those approaches for doing phenotypic screens for genetic disease treatments. I think they’re really useful for drug discovery, at the beginning of the pipeline, where they can enable you to search a wider space of drug candidates. At LRI, we’re starting all the way at the other end of the pipeline, with drugs that have already been tested and shown promise in vivo. However, once we make some progress on those, then yes, it could make sense to start doing some of these machine learning-enabled approaches.
What is the ideal mouse strain for aging research, particularly lifespan studies, in your view?
Well, the Interventions Testing Program at the National Institute of Aging is using three-way heterozygous mouse crosses, which I think is the ideal. A single inbred strain of mouse doesn’t have much genetic diversity, so often what you’re testing is the effect of a treatment on that particular strain of mouse, and the results won’t transfer to another strain.
The use of progeria mice is common in aging research due to the shorter study time, but these models are often criticized as not being representative of true aging; what are your thoughts on the prevalence of progeria mice in aging research, and are they a relevant model for what we are trying to achieve?
I think progeria mice are an imperfect proxy. There are a lot of different kinds of progeria, and they exhibit some but not all of the typical symptoms of natural aging. I’d have more confidence in studies done on aged mice than progeric mice.
We see that you have a strong commitment to ensuring public access to scientific knowledge. What inspired you to make such a wonderful and strong commitment to open science?
Well, coming from a data science background, I’m hyper-aware of how easy it is to fool yourself with data. You can massage anything into a spurious result if you test enough hypotheses and pick your subgroups artfully. Really, the best way to guard against that is to share the raw data so that people can run their own analyses. Making science more open is how you make it more trustworthy.
Is there a publically viewable list of the targets that you are interested in testing?
The list is still evolving, but some of the first things we’re looking into testing are carboxyfullerenes, which seem to have neuroprotective and life-extending effects, and epithalamin, which is a pineal gland-derived peptide that’s been reported to extend lifespan and even reduce human mortality. Both of these are sort of in the sweet spot of not being the subject of that much research to date, but what there is is very promising, so the value of information is high.
What is likely to be your first target for studies, and what is the rationale behind your choice?
I think people should know that there’s a lot of low-hanging fruit in aging research — treatments that we have reason to believe might work but that we’d still have to test. The misperceptions are either that life extension is so speculative that we’ll never get there or that we already know how to do it and you just have to take the right supplements to live forever. I think the reality is that we’ll have to do a lot of experimental work, but it’s highly possible that, in time, we might find something that extends healthy lifespan in humans.
We would like to thank Sarah for taking the time to do this interview with us, and we look forward to seeing her team’s progress in the near future.
Steve Hill serves on the LEAF Board of Directors and is the Editor in Chief, coordinating the daily news articles and social media content of the organization. He is an active journalist in the aging research and biotechnology field and has to date written over 500 articles on the topic as well as attending various medical industry conferences. In 2019 he was listed in the top 100 journalists covering biomedicine and longevity research in the industry report – Top-100 Journalists covering advanced biomedicine and longevity created by the Aging Analytics Agency. His work has been featured in H+ Magazine, Psychology Today, Singularity Weblog, Standpoint Magazine, and, Keep Me Prime, and New Economy Magazine. Steve has a background in project management and administration which has helped him to build a united team for effective fundraising and content creation, while his additional knowledge of biology and statistical data analysis allows him to carefully assess and coordinate the scientific groups involved in the project. In 2015 he led the Major Mouse Testing Program (MMTP) for the International Longevity Alliance and in 2016 helped the team of the SENS Research Foundation to reach their goal for the OncoSENS campaign for cancer research.