Systems-Level Epigenetic Changes of Aging and Disease
April 30, 2021ID6552
To CiteDCA Citation Guide
- 00:00It's now my great pleasure to
- 00:02introduce Doctor Morgan Levine,
- 00:04who is a ladder rank assistant
- 00:06professor in the Department of
- 00:08Pathology and a member of both the
- 00:11Yale Combined Programs in Computational
- 00:13Biology and Bioinformatics,
- 00:15as well as the Yale Center
- 00:17for Research on Aging.
- 00:19She has extensive experience using
- 00:21systems level and machine learning
- 00:23approaches to track epigenetic,
- 00:24transcriptomic and proteomic changes
- 00:26with aging and with developing
- 00:28measures of risk stratification
- 00:30for major chronic diseases.
- 00:32The floor is yours up to living.
- 00:34Thank you so much, Nicole.
- 00:36So I'm going to change course a little bit
- 00:39and a lot of the talks they had been on.
- 00:42Genetics and genomics an I'm actually in
- 00:43talk a little bit more about epigenetics.
- 00:46So before I begin I do have conflict of
- 00:48interest so I have IP associated with some
- 00:51of the markers that I'm just guessing
- 00:53today and I'm also consultant with a
- 00:55biotech company called Elise Team Health.
- 00:58So my love is very interested in
- 01:00studying the biology of aging,
- 01:02and I think it's really important to kind
- 01:04of address 2 important questions there.
- 01:06So number one.
- 01:07What do we mean when we're saying
- 01:09we study aging and #2?
- 01:10Why is it actually important to
- 01:12study aging and I'm going to answer
- 01:14the second question first.
- 01:15'cause it's actually an easier
- 01:17question to address.
- 01:18So some people may not be aware,
- 01:21but aging is actually the biggest
- 01:23risk factor for most of the chronic
- 01:26diseases that people suffer and
- 01:28die from in the US and worldwide.
- 01:30So this is actually a chart.
- 01:33the Y axis here is actually in
- 01:35log incident rate.
- 01:36So actually what this shows is
- 01:38that the risk for all of these
- 01:41chronic diseases actually rises
- 01:43exponentially as a function of age,
- 01:45and it's actually been estimated
- 01:47using demographic data.
- 01:48That if we were to slow the aging process
- 01:51down by something equivalent to 7 years,
- 01:54so this idea of 60s,
- 01:55the new 50 maybe 60 is the new 53 we
- 01:58would actually cut the disease burden
- 02:00in half in the US and worldwide.
- 02:08So to address kind of that first
- 02:10question of what do I mean when I say
- 02:13that we study aging and also annoy
- 02:15me when I say if we were actually to
- 02:18slow the rate of aging by 7 years,
- 02:21I think it's important to really
- 02:22contrast when we're talking about
- 02:24something like chronological aging,
- 02:26which as everyone knows,
- 02:27is basically the time in years,
- 02:29months or days since birth.
- 02:30And actually this should
- 02:32have a positive connotation.
- 02:33It's filled with experiences and
- 02:35contributions back to society.
- 02:36But the reason that chronological
- 02:38age actually get such negative
- 02:40connotation is that it's actually
- 02:41tide to something that in we in the
- 02:44field talk about as biological aging.
- 02:46And there isn't a single kind of
- 02:48definition for what we mean when
- 02:50we talk about but biological aging.
- 02:52But at least in my lab,
- 02:54we like to think of this as kind
- 02:56of the robustness or kind of
- 02:58specificity of a living system.
- 03:00So all living systems are basically
- 03:02set up to function in a certain way,
- 03:05but overtime.
- 03:06That specificity tends to actually decline,
- 03:08and this can lead to things like
- 03:11manifestations of disease or
- 03:12basically risk of system failure,
- 03:14which we might describe this death.
- 03:17So actually what my life is very
- 03:19interested in doing is can we quantify
- 03:21biological aging and contrast that
- 03:23to something like chronological age?
- 03:25And this is important for a number
- 03:28of reasons,
- 03:29so number one as we know
- 03:30chronological age is not modifiable,
- 03:32but we think biological age is,
- 03:35and this is actually conserved.
- 03:36Therefore, as a clinical endpoint.
- 03:38So if there are interventions aimed at
- 03:40actually slowing the aging process,
- 03:42we wouldn't have to wait 10 or
- 03:4520 years to use something like.
- 03:47Mortality or disease incidents.
- 03:49As an endpoint,
- 03:50we can actually measure biological
- 03:52age in real time.
- 03:53It's also just important for a
- 03:56basic understanding of drivers
- 03:57of the aging process.
- 03:59An identification of potential therapeutics,
- 04:01and finally we actually think that
- 04:03knowing biological agent help with
- 04:05primary and secondary prevention
- 04:06through risk stratification.
- 04:08So understanding who in the population is
- 04:10most at risk of various chronic diseases.
- 04:15So my lab and others have actually
- 04:18developed things called epigenetic clocks,
- 04:20which is mostly what I'm going to
- 04:22talk about today to actually try and
- 04:24get at biological aging and measure
- 04:27it in different tissues and organs.
- 04:29So I think she meant clocks
- 04:31are based on DNA methylations,
- 04:33so this is the addition of a methyl group
- 04:36to CPG dinucleotide and DNA metalation,
- 04:38and epigenetics in general are
- 04:40really important in kind of
- 04:42defining cellular states.
- 04:43So as most of you know.
- 04:45Essentially, most of the cells in your.
- 04:49You know,
- 04:49in the body of the same genetics
- 04:51but the average genetics are going
- 04:53to dictate cell state and South
- 04:56phenotype an actually decades ago is
- 04:58shown that actually your epigenetic
- 05:00pattern can also differentiate a
- 05:02young cell from an old cell and
- 05:04using this information we and others
- 05:06have actually developed these.
- 05:08What are called epigenetic locks,
- 05:09where we're looking at hundreds or
- 05:12thousands of CPG is across the genome
- 05:14measuring metalation and actually from
- 05:16there we can get out of predicted age.
- 05:19So this is showing an example of
- 05:21one of these measures were here.
- 05:23We have chronological age plotted on the
- 05:25X axis and epigenetic age on the Y axis,
- 05:28and as you can see,
- 05:30we can get actually a very good
- 05:32indicator of aging based solely.
- 05:34This is from 513 CPG's across the genome.
- 05:37But far more important than just
- 05:39kind of this parlor trick that we
- 05:42can actually predict someone's age.
- 05:43We're really interested in kind
- 05:45of this variance within age,
- 05:47so asking the question of people
- 05:48who are predicted older than they
- 05:50actually are chronologically,
- 05:51or these individuals more at
- 05:53risk for different diseases and
- 05:56conditions associated with aging.
- 05:58So we've looked at this in a
- 06:00number of different issues,
- 06:02so first this is data from the
- 06:04Framingham Heart Study on that was done.
- 06:07The study was done by my postdoc.
- 06:09Albert Higgins Chen were.
- 06:10Basically we can look at a one V
- 06:13Association for one standard deviation
- 06:15increase in your epigenetic age based
- 06:17on in contrast to your chronological
- 06:20age and what we find is that when
- 06:22looking at all 'cause mortality,
- 06:24this one standard deviation increases
- 06:26associated with almost a two fold increase.
- 06:29Risk in all 'cause mortality.
- 06:31We also can look at associations with
- 06:34a number of different conditions,
- 06:36so again,
- 06:37we find that older epigenetic age
- 06:39relative to chronological age is
- 06:41associated with different anthropometric
- 06:43factors like BMI and waist circumference.
- 06:45Functioning factors like heart rate,
- 06:47systolic blood pressure associated
- 06:49with lipid levels, and blood glucose,
- 06:51different health behaviors that we
- 06:53think are dealing serious to health,
- 06:56like cigarette smoking, heavy alcohol use,
- 06:58and a number of diseases.
- 07:00Cardiovascular disease,
- 07:01coronary heart disease, coronary heart.
- 07:03Failure, depression, sleep apnea,
- 07:04and this is actually not an exhaustive list.
- 07:07We've actually associated this with a
- 07:09number of different diseases as well.
- 07:11So on the left here,
- 07:13this is all epigenetic gauge
- 07:15measured in blood.
- 07:16But we can also measure
- 07:17opportunity in specific tissues,
- 07:19so my student care.
- 07:20Thresh is a student in the
- 07:22CMB program here at Yale,
- 07:24is interested in epigenetic aging and brain.
- 07:26So this is data from what's called
- 07:28the Ross Mapco part where we have
- 07:31postmortem brain samples and we've measured.
- 07:33Gauging in dorsal lateral prefrontal
- 07:35cortex and what we show here is
- 07:38that higher epigenetic aging is
- 07:41associated with postmortem diagnosis
- 07:42of Alzheimer's disease and this is
- 07:45based on neural pathology diagnosis.
- 07:47But we've also shown that even
- 07:50pre mortem this also tracks with
- 07:53things like cognitive decline.
- 07:55I'm so that just goes to show that
- 07:58Abby Janicke agent can be measured
- 08:00again across different issues,
- 08:02so we also have a collaborations with
- 08:05doctors whose Diane Hofstadter and
- 08:07actually what we've looked at here,
- 08:09is normal breast tissue from women
- 08:11with and without breast cancer.
- 08:14And again we show an accelerated
- 08:16epigenetic aging phenotype in women who
- 08:18do have breast cancer versus those who don't.
- 08:21With the hypothesis that being
- 08:23being that may be accelerated.
- 08:25Aging in a given tissue may actually
- 08:28predispose that tissue to tumorigenesis
- 08:30and this is prior to any treatment.
- 08:33And similarly,
- 08:34we've looked at data in liver and
- 08:36we find again higher epigenetic
- 08:39aging is associated with obesity,
- 08:41NAFLD and Nash in in the liver samples.
- 08:46So again,
- 08:47we can measure epigenetic aging
- 08:48and essentially any different
- 08:50tissue and cell type.
- 08:51And actually we use the exact same
- 08:53algorithm when we're doing this.
- 08:55These aren't tissue specific algorithms,
- 08:57so we can actually also use this to look
- 08:59at kind of between tissue differences,
- 09:02so there's this longstanding hypothesis
- 09:03in aging research that even though all
- 09:06your tissues are essentially aging,
- 09:07they do this in an asynchronous manner,
- 09:10so they're not all aging at the same rate.
- 09:13So what we've done here is
- 09:15we've looked at epigenetic.
- 09:16In a variety of different
- 09:18tissue and cell types,
- 09:19and basically what we find again
- 09:22is that your tissues and tissues do
- 09:24not seem to age at the same rate,
- 09:27so all these on the bottom
- 09:29are actually from brain.
- 09:30So it seems that brain not only
- 09:32seems epigenetically younger,
- 09:33but the rate of increase in
- 09:35epigenetic aging also seems slower,
- 09:37and then some of these more
- 09:39proliferative tissues like colon,
- 09:40epidermis and some of the blood cell types.
- 09:44We can also just compare within
- 09:47a tissue tumor.
- 09:48Epigenetic measures from tumor
- 09:50samples versus normal an across
- 09:53different cancer subtypes or
- 09:55types we find in red here.
- 09:57The tumor samples seem to have
- 10:00accelerated epigenetic aging
- 10:01compared to the normal controls.
- 10:06So one kind of outstanding question
- 10:08in this field is really, you know,
- 10:10mechanistically what's driving up.
- 10:11The genetic aging, and are there
- 10:14actually ways to intervene and modify
- 10:16the rate of the Fiji night change?
- 10:18So one of my sons Christmas here,
- 10:20who's in the M2P2 program here at Yale
- 10:23is actually working on developing
- 10:24in vitro models of epigenetic aging,
- 10:27so we can show that.
- 10:29Simply by passaging different cells.
- 10:31We've done this in mouse
- 10:33embryonic fibroblasts.
- 10:33We've done this in human astrocytes
- 10:35that we can actually recapitulate
- 10:37the exact same epigenetic patterns
- 10:39that we're seeing in vivo.
- 10:41So this is the method that Chris
- 10:43generated where he can show
- 10:45this increase in Abidjan cage
- 10:47as a function of cell passage,
- 10:49and then it uses the exact same
- 10:52equation and show that we can also see
- 10:55these exact same changes with age in
- 10:57vivo in a variety of different issues,
- 11:00so liver long.
- 11:01Kidney blood adipose however
- 11:03lesser extent and sculpt muscle,
- 11:05which is perhaps not surprising 'cause
- 11:07it's tends to be mostly post mitotic.
- 11:13But then again, probably the
- 11:15most important question here
- 11:16is whether this is modifiable,
- 11:18so this is actually on the left here.
- 11:22A collaboration that we did with Doctor
- 11:25David Sinclair at Harvard and his
- 11:27then student at the time using Luann.
- 11:30Basically, this is.
- 11:32Fishing nature last year where
- 11:34we had a test on Ted off system.
- 11:38For those of you familiar with Yamanaka
- 11:41factors were actually doing what
- 11:43we call epigenetic reprogramming.
- 11:45So this was a three factor system with OSK.
- 11:49And basically what they did is they
- 11:52crushed the optical nerve and they show
- 11:55that when you turn on OSK you actually
- 11:58get Rigo growth of these ganglia,
- 12:00neural, neural neural ganglion cells.
- 12:03But we also see change in the
- 12:06epigenetic aging pattern as well,
- 12:08so these are to be intact.
- 12:10So with without the crushing but
- 12:12with when you get injury you see
- 12:15an increase in epigenetic age.
- 12:17But then when you.
- 12:19Express to SK.
- 12:21You get a reduction in the
- 12:23epigenetic age again.
- 12:24And we actually show this
- 12:26as well in an aged model,
- 12:28not just an injury model,
- 12:30where with aging we show an increase
- 12:32in the epigenetic age in these cells.
- 12:34But with OSK we we can actually show
- 12:37in deceleration in upper GI gauge.
- 12:39And we've shown in a number
- 12:41of different cell types too.
- 12:42When you look at when you take a
- 12:44somatic cell and convert it back
- 12:46to induce player problem stencil,
- 12:48you also get a reversion or or
- 12:51basically complete erase erasing
- 12:52of this epigenetic aging pattern.
- 12:54And similarly,
- 12:55we've also looked in studies in
- 12:57mice on dietary intervention,
- 12:59so here this is a epigenetic
- 13:01gauge measure and mice in blood.
- 13:04And here in pink we have my
- 13:06sister on caloric restriction,
- 13:08so we showed a clinically restricted
- 13:10mice seem to have lower epigenetic age.
- 13:13They were all started at the same
- 13:16time on chloric restriction,
- 13:18so we actually show that it
- 13:21actually just also decreases the
- 13:23rate of epigenetic aging over time.
- 13:26So as takeaways,
- 13:28biological aging is actually
- 13:29distinct from chronological aging,
- 13:31and we can actually use things
- 13:34like epigenetic clocks that are
- 13:36based on DNA metalation to actually
- 13:39track biological aging,
- 13:40quantify it in various cells,
- 13:42and tissue types between person
- 13:44differences in the discordance between
- 13:46chronological inopportune engages
- 13:48biologically meaningful and actually
- 13:50has implications for things like morbidity,
- 13:52mortality,
- 13:53risk.
- 13:53An epigenetic clocks can be used
- 13:56to study asynchronous aging across
- 13:59different issue install States and
- 14:02also show distinctions in tissue
- 14:05state or such as acceleration
- 14:07at in tumor Genesis.
- 14:09They can also be used to
- 14:11facilitate in vitro aging models
- 14:14for discovery of therapeutics,
- 14:16and potentially the most important
- 14:18thing is that epigenetic aging
- 14:20phenomena is actually modifiable,
- 14:22but it's yet to show whether a
- 14:25deceleration epigenetic aging
- 14:27will actually manifest as as a
- 14:29reduced risks and things like
- 14:32morbidity and mortality.
- 14:33And with that I just want to
- 14:35acknowledge the people in my lab.
- 14:37My collaborators at Yellen
- 14:38elsewhere and also my funding from
- 14:40the National Institute on Aging,
- 14:42and I'm happy to take questions
- 14:43either via email or in the chat.
- 14:45Thank you so much.