Quantification of Biological Aging

Daniel W. Belsky, PhD Assistant Professor of Population Health ScienceDuke University School of Medicine

Implications for Clinical Trials of Therapies to Extend Healthy Lifespan

dbelsky@duke.eduhttp:/sites.duke.edu/danbelsky

U24AG047121 (Kraus, Peiper)P30 AG028716 (Schmader, Morey)P30AG034424 (O’Rand)R01 AG032282 (Moffitt)R21AG054846 (Belsky)

Outline

IntroductionBiological aging is a treatment target for healthspan extension

Part 1. Human trials of geroprotectorsChallenges & opportunities

Part 2. Quantification of biological aging in young adults

Part 3. Testing biological aging in a geroprotector trial

Conclusion

4

The global population is aging

Strategies are needed to extend healthy lifespan

5https://www.nimh.nih.gov/health/statistics/disability/us-leading-disease-disorder-categories-by-age.shtml

Aging itself is a leading risk factor for many diseases

6

Kennedyetal.2014Cell

Agingisabiologicalprocessagradualandprogressivedeclineinsystemintegrity

Lopez-Otin etal.2013Cell

7

GeroprotectiveIntervention?

8

Evolving theoretical models of aging

AcceleratedAging

Diseaseà Disability/Frailtyà Death

Early-lifeAdversity

Belsky etal.2015PNASMoffittetal.2016JGeron AMedSci

9

Evolving theoretical models of aging

AcceleratedAging

Diseaseà Disability/Frailtyà Death

Early-lifeAdversity

Belsky etal.2015PNASMoffittetal.2016JGeron AMedSci

10

Evolving theoretical models of aging

AcceleratedAging

Diseaseà Disability/Frailtyà Death

Early-lifeAdversity

Belsky etal.2015PNASMoffittetal.2016JGeron AMedSci

20 30 40 50 60 70 80 90 100DecadeofLife

Morbidityà

RisingMorbidity withAdvancingAge

11

• Exposures accumulate from early life • Changes to physiology precede disease onset• Preventive intervention must begin early

Belsky 2017

20 30 40 50 60 70 80 90 100DecadeofLife

Morbidityà

RisingMorbidity withAdvancingAge

GeroprotectiveIntervention

12

• Exposures accumulate from early life • Changes to physiology precede disease onset• Preventive intervention must begin early

Belsky 2017

20 30 40 50 60 70 80 90 100DecadeofLife

Morbidityà

RisingMorbidity withAdvancingAge

GeroprotectiveIntervention

13

• Exposures accumulate from early life • Changes to physiology precede disease onset• Preventive intervention must begin early

Belsky 2017

Outline

IntroductionBiological aging is a treatment target for healthspan extension

Part 1. Human trials of geroprotectorsChallenges & opportunities

Part 2. Quantification of biological aging in young adults

Part 3. Testing biological aging in a geroprotector trial

Conclusion

15

Ideal Geroprotector Trial Design

Rx

Control

HealthspanExtension

Belsky 2017

16

Ideal Geroprotector Trial Design

Rx

Control

HealthspanExtension

Rx

Control

HealthspanExtension

Belsky 2017

17

Ideal Geroprotector Trial Design

Rx

Control

HealthspanExtension

Rx

Control

HealthspanExtension

Months

Belsky 2017

18

Ideal Geroprotector Trial Design

Rx

Control

HealthspanExtension

Rx

Control

HealthspanExtension

Months Decades

Belsky 2017

19

Problem: Healthspan follow-up from midlife prevention takes too long

20 30 40 50 60 70 80 90 100DecadeofLife

Morbidityà

RisingMorbidity withAdvancingAge

GeroprotectiveIntervention

Belsky 2017

20

Solution: Measure aging processes as healthspansurrogate endpoint for trial

20 30 40 50 60 70 80 90 100DecadeofLife

Morbidityà

RisingMorbidity withAdvancingAge

GeroprotectiveIntervention

Belsky 2017

21

Alternative Geroprotector Trial Design

Rx

Control

SlowerPaceofAging

NormalPaceofAging

Years

Belsky 2017

22

Alternative Geroprotector Trial Design

Biologica

lAge

Y1 Y2 Y3

Control

Treatment

Rx

Control

SlowerPaceofAging

NormalPaceofAging

Years

23

Alternative Geroprotector Trial Design

Biologica

lAge

Y1 Y2 Y3

Control

Treatment

Rx

Control

SlowerPaceofAging

NormalPaceofAging

Years

Belsky 2017

24

Part 1 Summary

• Biological aging may be a modifiable risk factor for age-related disease and disability

• New (and old) therapies to slow biological aging now have proof-of-concept in animal models, and more are on the way

• Translation of midlife geroprotective interventions to humans faces a barrier – follow-up takes too long

• Translation to humans can be accelerated using study designs that measure biological aging as a surrogate endpoint for healthspanextension

Outline

IntroductionBiological aging is a treatment target for healthspan extension

Part 1. Human trials of geroprotectorsChallenges & opportunities

Part 2. Quantification of biological aging in young adults

Part 3. Testing biological aging in a geroprotector trial

Conclusion

26

27

www.moffittcaspi.com

28

The Dunedin Longitudinal Study

*Percentassessed,ofthosewhowerealiveateachage.

Age Year Number Percent*

Birth 1972-733 1975-76 1037 100%5 1977-78 991 967 1979-80 954 929 1981-82 955 9211 1983-84 925 9013 1985-86 850 8215 1987-88 976 9518 1990-91 993 9721 1993-94 992 9726 1998-99 980 9632 2004-05 972 96

38 2010-12 961 95%45 2017-18 ?? ??

29

The Dunedin Longitudinal Study

*Percentassessed,ofthosewhowerealiveateachage.

Age Year Number Percent*

Birth 1972-733 1975-76 1037 100%5 1977-78 991 967 1979-80 954 929 1981-82 955 9211 1983-84 925 9013 1985-86 850 8215 1987-88 976 9518 1990-91 993 9721 1993-94 992 9726 1998-99 980 9632 2004-05 972 96

38 2010-12 961 95%45 2017-18 ?? ??

30

Can we measure the rate of aging in young, healthy humans?

31

The Pace of Aging

Quantification of biological aging in young adultsDaniel W. Belskya,b,1, Avshalom Caspic,d,e,f, Renate Houtsc, Harvey J. Cohena, David L. Corcorane, Andrea Danesef,g,HonaLee Harringtonc, Salomon Israelh, Morgan E. Levinei, Jonathan D. Schaeferc, Karen Sugdenc, Ben Williamsc,Anatoli I. Yashinb, Richie Poultonj, and Terrie E. Moffittc,d,e,f

aDepartment of Medicine, Duke University School of Medicine, Durham, NC 27710; bSocial Science Research Institute, Duke University, Durham, NC 27708;cDepartment of Psychology & Neuroscience, Duke University, Durham, NC 27708; dDepartment of Psychiatry & Behavioral Sciences, Duke UniversitySchool of Medicine, Durham, NC 27708; eCenter for Genomic and Computational Biology, Duke University, Durham, NC 27708; fSocial, Genetic, &Developmental Psychiatry Research Centre, Institute of Psychiatry, Kings College London, London SE5 8AF, United Kingdom; gDepartment of Child &Adolescent Psychiatry, Institute of Psychiatry, King’s College London, London SE5 8AF, United Kingdom; hDepartment of Psychology, The HebrewUniversity of Jerusalem, Jerusalem 91905, Israel; iDepartment of Human Genetics, Gonda Research Center, David Geffen School of Medicine, University ofCalifornia, Los Angeles, CA 90095; and jDepartment of Psychology, University of Otago, Dunedin 9016, New Zealand

Edited by Bruce S. McEwen, The Rockefeller University, New York, NY, and approved June 1, 2015 (received for review March 30, 2015)

Antiaging therapies show promise in model organism research.Translation to humans is needed to address the challenges of anaging global population. Interventions to slow human aging willneed to be applied to still-young individuals. However, most humanaging research examines older adults, manywith chronic disease. Asa result, little is known about aging in young humans. We studiedaging in 954 young humans, the Dunedin Study birth cohort,tracking multiple biomarkers across three time points spanningtheir third and fourth decades of life. We developed and validatedtwo methods by which aging can be measured in young adults,one cross-sectional and one longitudinal. Our longitudinal mea-sure allows quantification of the pace of coordinated physiologicaldeterioration across multiple organ systems (e.g., pulmonary,periodontal, cardiovascular, renal, hepatic, and immune function).We applied these methods to assess biological aging in younghumans who had not yet developed age-related diseases. Youngindividuals of the same chronological age varied in their “biologicalaging” (declining integrity of multiple organ systems). Already,before midlife, individuals who were aging more rapidly were lessphysically able, showed cognitive decline and brain aging, self-reported worse health, and looked older. Measured biologicalaging in young adults can be used to identify causes of agingand evaluate rejuvenation therapies.

biological aging | cognitive aging | aging | healthspan | geroscience

By 2050, the world population aged 80 y and above will morethan triple, approaching 400 million individuals (1, 2). As the

population ages, the global burden of disease and disability isrising (3). From the fifth decade of life, advancing age is asso-ciated with an exponential increase in burden from many differentchronic conditions (Fig. 1). The most effective means to reducedisease burden and control costs is to delay this progression byextending healthspan, years of life lived free of disease and dis-ability (4). A key to extending healthspan is addressing the prob-lem of aging itself (5–8).At present, much research on aging is being carried out with

animals and older humans. Paradoxically, these seemingly sen-sible strategies pose translational difficulties. The difficulty withstudying aging in old humans is that many of them already haveage-related diseases (9–11). Age-related changes to physiologyaccumulate from early life, affecting organ systems years beforedisease diagnosis (12–15). Thus, intervention to reverse or delaythe march toward age-related diseases must be scheduled whilepeople are still young (16). Early interventions to slow aging canbe tested in model organisms (17, 18). The difficulty with thesenonhuman models is that they do not typically capture thecomplex multifactorial risks and exposures that shape humanaging. Moreover, whereas animals’ brief lives make it feasible tostudy animal aging in the laboratory, humans’ lives span manyyears. A solution is to study human aging in the first half of thelife course, when individuals are starting to diverge in their aging

trajectories, before most diseases (and regimens to managethem) become established. The main obstacle to studying agingbefore old age—and before the onset of age-related diseases—is the absence of methods to quantify the Pace of Aging inyoung humans.We studied aging in a population-representative 1972–1973 birth

cohort of 1,037 young adults followed from birth to age 38 y with95% retention: the Dunedin Study (SI Appendix). When they were38 y old, we examined their physiologies to test whether this youngpopulation would show evidence of individual variation in agingdespite remaining free of age-related disease. We next tested thehypothesis that cohort members with “older” physiologies at age 38had actually been aging faster than their same chronologically agedpeers who retained “younger” physiologies; specifically, we testedwhether indicators of the integrity of their cardiovascular, meta-bolic, and immune systems, their kidneys, livers, gums, and lungs,and their DNA had deteriorated more rapidly according to mea-surements taken repeatedly since a baseline 12 y earlier at age 26.We further tested whether, by midlife, young adults who wereaging more rapidly already exhibited deficits in their physicalfunctioning, showed signs of early cognitive decline, and lookedolder to independent observers.

ResultsAre Young Adults Aging at Different Rates? Measuring the aging pro-cess is controversial. Candidate biomarkers of aging are numerous,

Significance

The global population is aging, driving up age-related diseasemorbidity. Antiaging interventions are needed to reduce theburden of disease and protect population productivity. Youngpeople are the most attractive targets for therapies to extendhealthspan (because it is still possible to prevent disease in theyoung). However, there is skepticism about whether agingprocesses can be detected in young adults who do not yet havechronic diseases. Our findings indicate that aging processes canbe quantified in people still young enough for prevention ofage-related disease, opening a new door for antiaging thera-pies. The science of healthspan extension may be focused onthe wrong end of the lifespan; rather than only studying oldhumans, geroscience should also study the young.

Author contributions: D.W.B., A.C., R.P., and T.E.M. designed research; D.W.B., A.C., R.H.,H.J.C., D.L.C., A.D., H.H., S.I., M.E.L., J.D.S., K.S., B.W., A.I.Y., R.P., and T.E.M. performedresearch; M.E.L. contributed new reagents/analytic tools; D.W.B., A.C., R.H., H.H., andT.E.M. analyzed data; and D.W.B., A.C., and T.E.M. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: dbelsky@duke.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1506264112/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1506264112 PNAS Early Edition | 1 of 7

MED

ICALSC

IENCE

SPS

YCHO

LOGICALAND

COGNITIVESC

IENCE

SPN

ASPL

US

Belsky et al. 2015 PNAS

Agingischaracterizedbyagradualandprogressivedeclineinsystemintegrity

Therateofagingcanbeinferredfromtherateofdeclineinintegrityacrossmultipleorgansystems

32

The Pace of Aging

Quantification of biological aging in young adultsDaniel W. Belskya,b,1, Avshalom Caspic,d,e,f, Renate Houtsc, Harvey J. Cohena, David L. Corcorane, Andrea Danesef,g,HonaLee Harringtonc, Salomon Israelh, Morgan E. Levinei, Jonathan D. Schaeferc, Karen Sugdenc, Ben Williamsc,Anatoli I. Yashinb, Richie Poultonj, and Terrie E. Moffittc,d,e,f

aDepartment of Medicine, Duke University School of Medicine, Durham, NC 27710; bSocial Science Research Institute, Duke University, Durham, NC 27708;cDepartment of Psychology & Neuroscience, Duke University, Durham, NC 27708; dDepartment of Psychiatry & Behavioral Sciences, Duke UniversitySchool of Medicine, Durham, NC 27708; eCenter for Genomic and Computational Biology, Duke University, Durham, NC 27708; fSocial, Genetic, &Developmental Psychiatry Research Centre, Institute of Psychiatry, Kings College London, London SE5 8AF, United Kingdom; gDepartment of Child &Adolescent Psychiatry, Institute of Psychiatry, King’s College London, London SE5 8AF, United Kingdom; hDepartment of Psychology, The HebrewUniversity of Jerusalem, Jerusalem 91905, Israel; iDepartment of Human Genetics, Gonda Research Center, David Geffen School of Medicine, University ofCalifornia, Los Angeles, CA 90095; and jDepartment of Psychology, University of Otago, Dunedin 9016, New Zealand

Edited by Bruce S. McEwen, The Rockefeller University, New York, NY, and approved June 1, 2015 (received for review March 30, 2015)

Antiaging therapies show promise in model organism research.Translation to humans is needed to address the challenges of anaging global population. Interventions to slow human aging willneed to be applied to still-young individuals. However, most humanaging research examines older adults, manywith chronic disease. Asa result, little is known about aging in young humans. We studiedaging in 954 young humans, the Dunedin Study birth cohort,tracking multiple biomarkers across three time points spanningtheir third and fourth decades of life. We developed and validatedtwo methods by which aging can be measured in young adults,one cross-sectional and one longitudinal. Our longitudinal mea-sure allows quantification of the pace of coordinated physiologicaldeterioration across multiple organ systems (e.g., pulmonary,periodontal, cardiovascular, renal, hepatic, and immune function).We applied these methods to assess biological aging in younghumans who had not yet developed age-related diseases. Youngindividuals of the same chronological age varied in their “biologicalaging” (declining integrity of multiple organ systems). Already,before midlife, individuals who were aging more rapidly were lessphysically able, showed cognitive decline and brain aging, self-reported worse health, and looked older. Measured biologicalaging in young adults can be used to identify causes of agingand evaluate rejuvenation therapies.

biological aging | cognitive aging | aging | healthspan | geroscience

By 2050, the world population aged 80 y and above will morethan triple, approaching 400 million individuals (1, 2). As the

population ages, the global burden of disease and disability isrising (3). From the fifth decade of life, advancing age is asso-ciated with an exponential increase in burden from many differentchronic conditions (Fig. 1). The most effective means to reducedisease burden and control costs is to delay this progression byextending healthspan, years of life lived free of disease and dis-ability (4). A key to extending healthspan is addressing the prob-lem of aging itself (5–8).At present, much research on aging is being carried out with

animals and older humans. Paradoxically, these seemingly sen-sible strategies pose translational difficulties. The difficulty withstudying aging in old humans is that many of them already haveage-related diseases (9–11). Age-related changes to physiologyaccumulate from early life, affecting organ systems years beforedisease diagnosis (12–15). Thus, intervention to reverse or delaythe march toward age-related diseases must be scheduled whilepeople are still young (16). Early interventions to slow aging canbe tested in model organisms (17, 18). The difficulty with thesenonhuman models is that they do not typically capture thecomplex multifactorial risks and exposures that shape humanaging. Moreover, whereas animals’ brief lives make it feasible tostudy animal aging in the laboratory, humans’ lives span manyyears. A solution is to study human aging in the first half of thelife course, when individuals are starting to diverge in their aging

trajectories, before most diseases (and regimens to managethem) become established. The main obstacle to studying agingbefore old age—and before the onset of age-related diseases—is the absence of methods to quantify the Pace of Aging inyoung humans.We studied aging in a population-representative 1972–1973 birth

cohort of 1,037 young adults followed from birth to age 38 y with95% retention: the Dunedin Study (SI Appendix). When they were38 y old, we examined their physiologies to test whether this youngpopulation would show evidence of individual variation in agingdespite remaining free of age-related disease. We next tested thehypothesis that cohort members with “older” physiologies at age 38had actually been aging faster than their same chronologically agedpeers who retained “younger” physiologies; specifically, we testedwhether indicators of the integrity of their cardiovascular, meta-bolic, and immune systems, their kidneys, livers, gums, and lungs,and their DNA had deteriorated more rapidly according to mea-surements taken repeatedly since a baseline 12 y earlier at age 26.We further tested whether, by midlife, young adults who wereaging more rapidly already exhibited deficits in their physicalfunctioning, showed signs of early cognitive decline, and lookedolder to independent observers.

ResultsAre Young Adults Aging at Different Rates? Measuring the aging pro-cess is controversial. Candidate biomarkers of aging are numerous,

Significance

The global population is aging, driving up age-related diseasemorbidity. Antiaging interventions are needed to reduce theburden of disease and protect population productivity. Youngpeople are the most attractive targets for therapies to extendhealthspan (because it is still possible to prevent disease in theyoung). However, there is skepticism about whether agingprocesses can be detected in young adults who do not yet havechronic diseases. Our findings indicate that aging processes canbe quantified in people still young enough for prevention ofage-related disease, opening a new door for antiaging thera-pies. The science of healthspan extension may be focused onthe wrong end of the lifespan; rather than only studying oldhumans, geroscience should also study the young.

Author contributions: D.W.B., A.C., R.P., and T.E.M. designed research; D.W.B., A.C., R.H.,H.J.C., D.L.C., A.D., H.H., S.I., M.E.L., J.D.S., K.S., B.W., A.I.Y., R.P., and T.E.M. performedresearch; M.E.L. contributed new reagents/analytic tools; D.W.B., A.C., R.H., H.H., andT.E.M. analyzed data; and D.W.B., A.C., and T.E.M. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: dbelsky@duke.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1506264112/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1506264112 PNAS Early Edition | 1 of 7

MED

ICALSC

IENCE

SPS

YCHO

LOGICALAND

COGNITIVESC

IENCE

SPN

ASPL

US

Belsky et al. 2015 PNAS

GingivalRecession

LTLWBCCRP

VO2max

CreatinineBUN

BMI,WHR

FEV1FEV1/FVC

TotCholesterolHDL,TriglyceridesLipoprotein(a)ApoB100/ApoA1

HbA1C

MeanArterialPressure

33

The Pace of Aging

Quantification of biological aging in young adultsDaniel W. Belskya,b,1, Avshalom Caspic,d,e,f, Renate Houtsc, Harvey J. Cohena, David L. Corcorane, Andrea Danesef,g,HonaLee Harringtonc, Salomon Israelh, Morgan E. Levinei, Jonathan D. Schaeferc, Karen Sugdenc, Ben Williamsc,Anatoli I. Yashinb, Richie Poultonj, and Terrie E. Moffittc,d,e,f

aDepartment of Medicine, Duke University School of Medicine, Durham, NC 27710; bSocial Science Research Institute, Duke University, Durham, NC 27708;cDepartment of Psychology & Neuroscience, Duke University, Durham, NC 27708; dDepartment of Psychiatry & Behavioral Sciences, Duke UniversitySchool of Medicine, Durham, NC 27708; eCenter for Genomic and Computational Biology, Duke University, Durham, NC 27708; fSocial, Genetic, &Developmental Psychiatry Research Centre, Institute of Psychiatry, Kings College London, London SE5 8AF, United Kingdom; gDepartment of Child &Adolescent Psychiatry, Institute of Psychiatry, King’s College London, London SE5 8AF, United Kingdom; hDepartment of Psychology, The HebrewUniversity of Jerusalem, Jerusalem 91905, Israel; iDepartment of Human Genetics, Gonda Research Center, David Geffen School of Medicine, University ofCalifornia, Los Angeles, CA 90095; and jDepartment of Psychology, University of Otago, Dunedin 9016, New Zealand

Edited by Bruce S. McEwen, The Rockefeller University, New York, NY, and approved June 1, 2015 (received for review March 30, 2015)

Antiaging therapies show promise in model organism research.Translation to humans is needed to address the challenges of anaging global population. Interventions to slow human aging willneed to be applied to still-young individuals. However, most humanaging research examines older adults, manywith chronic disease. Asa result, little is known about aging in young humans. We studiedaging in 954 young humans, the Dunedin Study birth cohort,tracking multiple biomarkers across three time points spanningtheir third and fourth decades of life. We developed and validatedtwo methods by which aging can be measured in young adults,one cross-sectional and one longitudinal. Our longitudinal mea-sure allows quantification of the pace of coordinated physiologicaldeterioration across multiple organ systems (e.g., pulmonary,periodontal, cardiovascular, renal, hepatic, and immune function).We applied these methods to assess biological aging in younghumans who had not yet developed age-related diseases. Youngindividuals of the same chronological age varied in their “biologicalaging” (declining integrity of multiple organ systems). Already,before midlife, individuals who were aging more rapidly were lessphysically able, showed cognitive decline and brain aging, self-reported worse health, and looked older. Measured biologicalaging in young adults can be used to identify causes of agingand evaluate rejuvenation therapies.

biological aging | cognitive aging | aging | healthspan | geroscience

By 2050, the world population aged 80 y and above will morethan triple, approaching 400 million individuals (1, 2). As the

population ages, the global burden of disease and disability isrising (3). From the fifth decade of life, advancing age is asso-ciated with an exponential increase in burden from many differentchronic conditions (Fig. 1). The most effective means to reducedisease burden and control costs is to delay this progression byextending healthspan, years of life lived free of disease and dis-ability (4). A key to extending healthspan is addressing the prob-lem of aging itself (5–8).At present, much research on aging is being carried out with

animals and older humans. Paradoxically, these seemingly sen-sible strategies pose translational difficulties. The difficulty withstudying aging in old humans is that many of them already haveage-related diseases (9–11). Age-related changes to physiologyaccumulate from early life, affecting organ systems years beforedisease diagnosis (12–15). Thus, intervention to reverse or delaythe march toward age-related diseases must be scheduled whilepeople are still young (16). Early interventions to slow aging canbe tested in model organisms (17, 18). The difficulty with thesenonhuman models is that they do not typically capture thecomplex multifactorial risks and exposures that shape humanaging. Moreover, whereas animals’ brief lives make it feasible tostudy animal aging in the laboratory, humans’ lives span manyyears. A solution is to study human aging in the first half of thelife course, when individuals are starting to diverge in their aging

trajectories, before most diseases (and regimens to managethem) become established. The main obstacle to studying agingbefore old age—and before the onset of age-related diseases—is the absence of methods to quantify the Pace of Aging inyoung humans.We studied aging in a population-representative 1972–1973 birth

cohort of 1,037 young adults followed from birth to age 38 y with95% retention: the Dunedin Study (SI Appendix). When they were38 y old, we examined their physiologies to test whether this youngpopulation would show evidence of individual variation in agingdespite remaining free of age-related disease. We next tested thehypothesis that cohort members with “older” physiologies at age 38had actually been aging faster than their same chronologically agedpeers who retained “younger” physiologies; specifically, we testedwhether indicators of the integrity of their cardiovascular, meta-bolic, and immune systems, their kidneys, livers, gums, and lungs,and their DNA had deteriorated more rapidly according to mea-surements taken repeatedly since a baseline 12 y earlier at age 26.We further tested whether, by midlife, young adults who wereaging more rapidly already exhibited deficits in their physicalfunctioning, showed signs of early cognitive decline, and lookedolder to independent observers.

ResultsAre Young Adults Aging at Different Rates? Measuring the aging pro-cess is controversial. Candidate biomarkers of aging are numerous,

Significance

The global population is aging, driving up age-related diseasemorbidity. Antiaging interventions are needed to reduce theburden of disease and protect population productivity. Youngpeople are the most attractive targets for therapies to extendhealthspan (because it is still possible to prevent disease in theyoung). However, there is skepticism about whether agingprocesses can be detected in young adults who do not yet havechronic diseases. Our findings indicate that aging processes canbe quantified in people still young enough for prevention ofage-related disease, opening a new door for antiaging thera-pies. The science of healthspan extension may be focused onthe wrong end of the lifespan; rather than only studying oldhumans, geroscience should also study the young.

Author contributions: D.W.B., A.C., R.P., and T.E.M. designed research; D.W.B., A.C., R.H.,H.J.C., D.L.C., A.D., H.H., S.I., M.E.L., J.D.S., K.S., B.W., A.I.Y., R.P., and T.E.M. performedresearch; M.E.L. contributed new reagents/analytic tools; D.W.B., A.C., R.H., H.H., andT.E.M. analyzed data; and D.W.B., A.C., and T.E.M. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: dbelsky@duke.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1506264112/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1506264112 PNAS Early Edition | 1 of 7

MED

ICALSC

IENCE

SPS

YCHO

LOGICALAND

COGNITIVESC

IENCE

SPN

ASPL

US

Belsky et al. 2015 PNAS

GingivalRecession

LTLWBCCRP

VO2max

CreatinineBUN

BMI,WHR

FEV1FEV1/FVC

TotCholesterolHDL,TriglyceridesLipoprotein(a)ApoB100/ApoA1

HbA1C

MeanArterialPressure

34

The Pace of Aging

Quantification of biological aging in young adultsDaniel W. Belskya,b,1, Avshalom Caspic,d,e,f, Renate Houtsc, Harvey J. Cohena, David L. Corcorane, Andrea Danesef,g,HonaLee Harringtonc, Salomon Israelh, Morgan E. Levinei, Jonathan D. Schaeferc, Karen Sugdenc, Ben Williamsc,Anatoli I. Yashinb, Richie Poultonj, and Terrie E. Moffittc,d,e,f

aDepartment of Medicine, Duke University School of Medicine, Durham, NC 27710; bSocial Science Research Institute, Duke University, Durham, NC 27708;cDepartment of Psychology & Neuroscience, Duke University, Durham, NC 27708; dDepartment of Psychiatry & Behavioral Sciences, Duke UniversitySchool of Medicine, Durham, NC 27708; eCenter for Genomic and Computational Biology, Duke University, Durham, NC 27708; fSocial, Genetic, &Developmental Psychiatry Research Centre, Institute of Psychiatry, Kings College London, London SE5 8AF, United Kingdom; gDepartment of Child &Adolescent Psychiatry, Institute of Psychiatry, King’s College London, London SE5 8AF, United Kingdom; hDepartment of Psychology, The HebrewUniversity of Jerusalem, Jerusalem 91905, Israel; iDepartment of Human Genetics, Gonda Research Center, David Geffen School of Medicine, University ofCalifornia, Los Angeles, CA 90095; and jDepartment of Psychology, University of Otago, Dunedin 9016, New Zealand

Edited by Bruce S. McEwen, The Rockefeller University, New York, NY, and approved June 1, 2015 (received for review March 30, 2015)

Antiaging therapies show promise in model organism research.Translation to humans is needed to address the challenges of anaging global population. Interventions to slow human aging willneed to be applied to still-young individuals. However, most humanaging research examines older adults, manywith chronic disease. Asa result, little is known about aging in young humans. We studiedaging in 954 young humans, the Dunedin Study birth cohort,tracking multiple biomarkers across three time points spanningtheir third and fourth decades of life. We developed and validatedtwo methods by which aging can be measured in young adults,one cross-sectional and one longitudinal. Our longitudinal mea-sure allows quantification of the pace of coordinated physiologicaldeterioration across multiple organ systems (e.g., pulmonary,periodontal, cardiovascular, renal, hepatic, and immune function).We applied these methods to assess biological aging in younghumans who had not yet developed age-related diseases. Youngindividuals of the same chronological age varied in their “biologicalaging” (declining integrity of multiple organ systems). Already,before midlife, individuals who were aging more rapidly were lessphysically able, showed cognitive decline and brain aging, self-reported worse health, and looked older. Measured biologicalaging in young adults can be used to identify causes of agingand evaluate rejuvenation therapies.

biological aging | cognitive aging | aging | healthspan | geroscience

By 2050, the world population aged 80 y and above will morethan triple, approaching 400 million individuals (1, 2). As the

population ages, the global burden of disease and disability isrising (3). From the fifth decade of life, advancing age is asso-ciated with an exponential increase in burden from many differentchronic conditions (Fig. 1). The most effective means to reducedisease burden and control costs is to delay this progression byextending healthspan, years of life lived free of disease and dis-ability (4). A key to extending healthspan is addressing the prob-lem of aging itself (5–8).At present, much research on aging is being carried out with

animals and older humans. Paradoxically, these seemingly sen-sible strategies pose translational difficulties. The difficulty withstudying aging in old humans is that many of them already haveage-related diseases (9–11). Age-related changes to physiologyaccumulate from early life, affecting organ systems years beforedisease diagnosis (12–15). Thus, intervention to reverse or delaythe march toward age-related diseases must be scheduled whilepeople are still young (16). Early interventions to slow aging canbe tested in model organisms (17, 18). The difficulty with thesenonhuman models is that they do not typically capture thecomplex multifactorial risks and exposures that shape humanaging. Moreover, whereas animals’ brief lives make it feasible tostudy animal aging in the laboratory, humans’ lives span manyyears. A solution is to study human aging in the first half of thelife course, when individuals are starting to diverge in their aging

trajectories, before most diseases (and regimens to managethem) become established. The main obstacle to studying agingbefore old age—and before the onset of age-related diseases—is the absence of methods to quantify the Pace of Aging inyoung humans.We studied aging in a population-representative 1972–1973 birth

cohort of 1,037 young adults followed from birth to age 38 y with95% retention: the Dunedin Study (SI Appendix). When they were38 y old, we examined their physiologies to test whether this youngpopulation would show evidence of individual variation in agingdespite remaining free of age-related disease. We next tested thehypothesis that cohort members with “older” physiologies at age 38had actually been aging faster than their same chronologically agedpeers who retained “younger” physiologies; specifically, we testedwhether indicators of the integrity of their cardiovascular, meta-bolic, and immune systems, their kidneys, livers, gums, and lungs,and their DNA had deteriorated more rapidly according to mea-surements taken repeatedly since a baseline 12 y earlier at age 26.We further tested whether, by midlife, young adults who wereaging more rapidly already exhibited deficits in their physicalfunctioning, showed signs of early cognitive decline, and lookedolder to independent observers.

ResultsAre Young Adults Aging at Different Rates? Measuring the aging pro-cess is controversial. Candidate biomarkers of aging are numerous,

Significance

The global population is aging, driving up age-related diseasemorbidity. Antiaging interventions are needed to reduce theburden of disease and protect population productivity. Youngpeople are the most attractive targets for therapies to extendhealthspan (because it is still possible to prevent disease in theyoung). However, there is skepticism about whether agingprocesses can be detected in young adults who do not yet havechronic diseases. Our findings indicate that aging processes canbe quantified in people still young enough for prevention ofage-related disease, opening a new door for antiaging thera-pies. The science of healthspan extension may be focused onthe wrong end of the lifespan; rather than only studying oldhumans, geroscience should also study the young.

Author contributions: D.W.B., A.C., R.P., and T.E.M. designed research; D.W.B., A.C., R.H.,H.J.C., D.L.C., A.D., H.H., S.I., M.E.L., J.D.S., K.S., B.W., A.I.Y., R.P., and T.E.M. performedresearch; M.E.L. contributed new reagents/analytic tools; D.W.B., A.C., R.H., H.H., andT.E.M. analyzed data; and D.W.B., A.C., and T.E.M. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: dbelsky@duke.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1506264112/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1506264112 PNAS Early Edition | 1 of 7

MED

ICALSC

IENCE

SPS

YCHO

LOGICALAND

COGNITIVESC

IENCE

SPN

ASPL

US

Belsky et al. 2015 PNAS

GingivalRecession

LTLWBCCRP

VO2max

CreatinineBUN

BMI,WHR

FEV1FEV1/FVC

TotCholesterolHDL,TriglyceridesLipoprotein(a)ApoB100/ApoA1

HbA1C

MeanArterialPressure

35

The Pace of Aging

Quantification of biological aging in young adultsDaniel W. Belskya,b,1, Avshalom Caspic,d,e,f, Renate Houtsc, Harvey J. Cohena, David L. Corcorane, Andrea Danesef,g,HonaLee Harringtonc, Salomon Israelh, Morgan E. Levinei, Jonathan D. Schaeferc, Karen Sugdenc, Ben Williamsc,Anatoli I. Yashinb, Richie Poultonj, and Terrie E. Moffittc,d,e,f

aDepartment of Medicine, Duke University School of Medicine, Durham, NC 27710; bSocial Science Research Institute, Duke University, Durham, NC 27708;cDepartment of Psychology & Neuroscience, Duke University, Durham, NC 27708; dDepartment of Psychiatry & Behavioral Sciences, Duke UniversitySchool of Medicine, Durham, NC 27708; eCenter for Genomic and Computational Biology, Duke University, Durham, NC 27708; fSocial, Genetic, &Developmental Psychiatry Research Centre, Institute of Psychiatry, Kings College London, London SE5 8AF, United Kingdom; gDepartment of Child &Adolescent Psychiatry, Institute of Psychiatry, King’s College London, London SE5 8AF, United Kingdom; hDepartment of Psychology, The HebrewUniversity of Jerusalem, Jerusalem 91905, Israel; iDepartment of Human Genetics, Gonda Research Center, David Geffen School of Medicine, University ofCalifornia, Los Angeles, CA 90095; and jDepartment of Psychology, University of Otago, Dunedin 9016, New Zealand

Edited by Bruce S. McEwen, The Rockefeller University, New York, NY, and approved June 1, 2015 (received for review March 30, 2015)

Antiaging therapies show promise in model organism research.Translation to humans is needed to address the challenges of anaging global population. Interventions to slow human aging willneed to be applied to still-young individuals. However, most humanaging research examines older adults, manywith chronic disease. Asa result, little is known about aging in young humans. We studiedaging in 954 young humans, the Dunedin Study birth cohort,tracking multiple biomarkers across three time points spanningtheir third and fourth decades of life. We developed and validatedtwo methods by which aging can be measured in young adults,one cross-sectional and one longitudinal. Our longitudinal mea-sure allows quantification of the pace of coordinated physiologicaldeterioration across multiple organ systems (e.g., pulmonary,periodontal, cardiovascular, renal, hepatic, and immune function).We applied these methods to assess biological aging in younghumans who had not yet developed age-related diseases. Youngindividuals of the same chronological age varied in their “biologicalaging” (declining integrity of multiple organ systems). Already,before midlife, individuals who were aging more rapidly were lessphysically able, showed cognitive decline and brain aging, self-reported worse health, and looked older. Measured biologicalaging in young adults can be used to identify causes of agingand evaluate rejuvenation therapies.

biological aging | cognitive aging | aging | healthspan | geroscience

By 2050, the world population aged 80 y and above will morethan triple, approaching 400 million individuals (1, 2). As the

population ages, the global burden of disease and disability isrising (3). From the fifth decade of life, advancing age is asso-ciated with an exponential increase in burden from many differentchronic conditions (Fig. 1). The most effective means to reducedisease burden and control costs is to delay this progression byextending healthspan, years of life lived free of disease and dis-ability (4). A key to extending healthspan is addressing the prob-lem of aging itself (5–8).At present, much research on aging is being carried out with

animals and older humans. Paradoxically, these seemingly sen-sible strategies pose translational difficulties. The difficulty withstudying aging in old humans is that many of them already haveage-related diseases (9–11). Age-related changes to physiologyaccumulate from early life, affecting organ systems years beforedisease diagnosis (12–15). Thus, intervention to reverse or delaythe march toward age-related diseases must be scheduled whilepeople are still young (16). Early interventions to slow aging canbe tested in model organisms (17, 18). The difficulty with thesenonhuman models is that they do not typically capture thecomplex multifactorial risks and exposures that shape humanaging. Moreover, whereas animals’ brief lives make it feasible tostudy animal aging in the laboratory, humans’ lives span manyyears. A solution is to study human aging in the first half of thelife course, when individuals are starting to diverge in their aging

trajectories, before most diseases (and regimens to managethem) become established. The main obstacle to studying agingbefore old age—and before the onset of age-related diseases—is the absence of methods to quantify the Pace of Aging inyoung humans.We studied aging in a population-representative 1972–1973 birth

cohort of 1,037 young adults followed from birth to age 38 y with95% retention: the Dunedin Study (SI Appendix). When they were38 y old, we examined their physiologies to test whether this youngpopulation would show evidence of individual variation in agingdespite remaining free of age-related disease. We next tested thehypothesis that cohort members with “older” physiologies at age 38had actually been aging faster than their same chronologically agedpeers who retained “younger” physiologies; specifically, we testedwhether indicators of the integrity of their cardiovascular, meta-bolic, and immune systems, their kidneys, livers, gums, and lungs,and their DNA had deteriorated more rapidly according to mea-surements taken repeatedly since a baseline 12 y earlier at age 26.We further tested whether, by midlife, young adults who wereaging more rapidly already exhibited deficits in their physicalfunctioning, showed signs of early cognitive decline, and lookedolder to independent observers.

ResultsAre Young Adults Aging at Different Rates? Measuring the aging pro-cess is controversial. Candidate biomarkers of aging are numerous,

Significance

The global population is aging, driving up age-related diseasemorbidity. Antiaging interventions are needed to reduce theburden of disease and protect population productivity. Youngpeople are the most attractive targets for therapies to extendhealthspan (because it is still possible to prevent disease in theyoung). However, there is skepticism about whether agingprocesses can be detected in young adults who do not yet havechronic diseases. Our findings indicate that aging processes canbe quantified in people still young enough for prevention ofage-related disease, opening a new door for antiaging thera-pies. The science of healthspan extension may be focused onthe wrong end of the lifespan; rather than only studying oldhumans, geroscience should also study the young.

Author contributions: D.W.B., A.C., R.P., and T.E.M. designed research; D.W.B., A.C., R.H.,H.J.C., D.L.C., A.D., H.H., S.I., M.E.L., J.D.S., K.S., B.W., A.I.Y., R.P., and T.E.M. performedresearch; M.E.L. contributed new reagents/analytic tools; D.W.B., A.C., R.H., H.H., andT.E.M. analyzed data; and D.W.B., A.C., and T.E.M. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: dbelsky@duke.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1506264112/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1506264112 PNAS Early Edition | 1 of 7

MED

ICALSC

IENCE

SPS

YCHO

LOGICALAND

COGNITIVESC

IENCE

SPN

ASPL

US

Belsky et al. 2015 PNAS

GingivalRecession

LTLWBCCRP

VO2max

CreatinineBUN

BMI,WHR

FEV1FEV1/FVC

TotCholesterolHDL,TriglyceridesLipoprotein(a)ApoB100/ApoA1

HbA1C

MeanArterialPressure

0

5

10

15

Perc

ent

0.0 0.5 1.0 1.5 2.0Years of Physiological Change Per One Chronological Year

Pace of Aging 26-38

36

The Pace of Aging

Quantification of biological aging in young adultsDaniel W. Belskya,b,1, Avshalom Caspic,d,e,f, Renate Houtsc, Harvey J. Cohena, David L. Corcorane, Andrea Danesef,g,HonaLee Harringtonc, Salomon Israelh, Morgan E. Levinei, Jonathan D. Schaeferc, Karen Sugdenc, Ben Williamsc,Anatoli I. Yashinb, Richie Poultonj, and Terrie E. Moffittc,d,e,f

aDepartment of Medicine, Duke University School of Medicine, Durham, NC 27710; bSocial Science Research Institute, Duke University, Durham, NC 27708;cDepartment of Psychology & Neuroscience, Duke University, Durham, NC 27708; dDepartment of Psychiatry & Behavioral Sciences, Duke UniversitySchool of Medicine, Durham, NC 27708; eCenter for Genomic and Computational Biology, Duke University, Durham, NC 27708; fSocial, Genetic, &Developmental Psychiatry Research Centre, Institute of Psychiatry, Kings College London, London SE5 8AF, United Kingdom; gDepartment of Child &Adolescent Psychiatry, Institute of Psychiatry, King’s College London, London SE5 8AF, United Kingdom; hDepartment of Psychology, The HebrewUniversity of Jerusalem, Jerusalem 91905, Israel; iDepartment of Human Genetics, Gonda Research Center, David Geffen School of Medicine, University ofCalifornia, Los Angeles, CA 90095; and jDepartment of Psychology, University of Otago, Dunedin 9016, New Zealand

Edited by Bruce S. McEwen, The Rockefeller University, New York, NY, and approved June 1, 2015 (received for review March 30, 2015)

Antiaging therapies show promise in model organism research.Translation to humans is needed to address the challenges of anaging global population. Interventions to slow human aging willneed to be applied to still-young individuals. However, most humanaging research examines older adults, manywith chronic disease. Asa result, little is known about aging in young humans. We studiedaging in 954 young humans, the Dunedin Study birth cohort,tracking multiple biomarkers across three time points spanningtheir third and fourth decades of life. We developed and validatedtwo methods by which aging can be measured in young adults,one cross-sectional and one longitudinal. Our longitudinal mea-sure allows quantification of the pace of coordinated physiologicaldeterioration across multiple organ systems (e.g., pulmonary,periodontal, cardiovascular, renal, hepatic, and immune function).We applied these methods to assess biological aging in younghumans who had not yet developed age-related diseases. Youngindividuals of the same chronological age varied in their “biologicalaging” (declining integrity of multiple organ systems). Already,before midlife, individuals who were aging more rapidly were lessphysically able, showed cognitive decline and brain aging, self-reported worse health, and looked older. Measured biologicalaging in young adults can be used to identify causes of agingand evaluate rejuvenation therapies.

biological aging | cognitive aging | aging | healthspan | geroscience

By 2050, the world population aged 80 y and above will morethan triple, approaching 400 million individuals (1, 2). As the

population ages, the global burden of disease and disability isrising (3). From the fifth decade of life, advancing age is asso-ciated with an exponential increase in burden from many differentchronic conditions (Fig. 1). The most effective means to reducedisease burden and control costs is to delay this progression byextending healthspan, years of life lived free of disease and dis-ability (4). A key to extending healthspan is addressing the prob-lem of aging itself (5–8).At present, much research on aging is being carried out with

animals and older humans. Paradoxically, these seemingly sen-sible strategies pose translational difficulties. The difficulty withstudying aging in old humans is that many of them already haveage-related diseases (9–11). Age-related changes to physiologyaccumulate from early life, affecting organ systems years beforedisease diagnosis (12–15). Thus, intervention to reverse or delaythe march toward age-related diseases must be scheduled whilepeople are still young (16). Early interventions to slow aging canbe tested in model organisms (17, 18). The difficulty with thesenonhuman models is that they do not typically capture thecomplex multifactorial risks and exposures that shape humanaging. Moreover, whereas animals’ brief lives make it feasible tostudy animal aging in the laboratory, humans’ lives span manyyears. A solution is to study human aging in the first half of thelife course, when individuals are starting to diverge in their aging

trajectories, before most diseases (and regimens to managethem) become established. The main obstacle to studying agingbefore old age—and before the onset of age-related diseases—is the absence of methods to quantify the Pace of Aging inyoung humans.We studied aging in a population-representative 1972–1973 birth

cohort of 1,037 young adults followed from birth to age 38 y with95% retention: the Dunedin Study (SI Appendix). When they were38 y old, we examined their physiologies to test whether this youngpopulation would show evidence of individual variation in agingdespite remaining free of age-related disease. We next tested thehypothesis that cohort members with “older” physiologies at age 38had actually been aging faster than their same chronologically agedpeers who retained “younger” physiologies; specifically, we testedwhether indicators of the integrity of their cardiovascular, meta-bolic, and immune systems, their kidneys, livers, gums, and lungs,and their DNA had deteriorated more rapidly according to mea-surements taken repeatedly since a baseline 12 y earlier at age 26.We further tested whether, by midlife, young adults who wereaging more rapidly already exhibited deficits in their physicalfunctioning, showed signs of early cognitive decline, and lookedolder to independent observers.

ResultsAre Young Adults Aging at Different Rates? Measuring the aging pro-cess is controversial. Candidate biomarkers of aging are numerous,

Significance

The global population is aging, driving up age-related diseasemorbidity. Antiaging interventions are needed to reduce theburden of disease and protect population productivity. Youngpeople are the most attractive targets for therapies to extendhealthspan (because it is still possible to prevent disease in theyoung). However, there is skepticism about whether agingprocesses can be detected in young adults who do not yet havechronic diseases. Our findings indicate that aging processes canbe quantified in people still young enough for prevention ofage-related disease, opening a new door for antiaging thera-pies. The science of healthspan extension may be focused onthe wrong end of the lifespan; rather than only studying oldhumans, geroscience should also study the young.

Author contributions: D.W.B., A.C., R.P., and T.E.M. designed research; D.W.B., A.C., R.H.,H.J.C., D.L.C., A.D., H.H., S.I., M.E.L., J.D.S., K.S., B.W., A.I.Y., R.P., and T.E.M. performedresearch; M.E.L. contributed new reagents/analytic tools; D.W.B., A.C., R.H., H.H., andT.E.M. analyzed data; and D.W.B., A.C., and T.E.M. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: dbelsky@duke.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1506264112/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1506264112 PNAS Early Edition | 1 of 7

MED

ICALSC

IENCE

SPS

YCHO

LOGICALAND

COGNITIVESC

IENCE

SPN

ASPL

US

Belsky et al. 2015 PNAS

GingivalRecession

LTLWBCCRP

VO2max

CreatinineBUN

BMI,WHR

FEV1FEV1/FVC

TotCholesterolHDL,TriglyceridesLipoprotein(a)ApoB100/ApoA1

HbA1C

MeanArterialPressure

20

30

40

50

60

Estim

ated

Bio

logi

cal A

ge

26 32 38Chronological Age

67

AgingSlow

AgingFast

37

Cross-sectional measurement of biological age

Quantification of biological aging in young adultsDaniel W. Belskya,b,1, Avshalom Caspic,d,e,f, Renate Houtsc, Harvey J. Cohena, David L. Corcorane, Andrea Danesef,g,HonaLee Harringtonc, Salomon Israelh, Morgan E. Levinei, Jonathan D. Schaeferc, Karen Sugdenc, Ben Williamsc,Anatoli I. Yashinb, Richie Poultonj, and Terrie E. Moffittc,d,e,f

aDepartment of Medicine, Duke University School of Medicine, Durham, NC 27710; bSocial Science Research Institute, Duke University, Durham, NC 27708;cDepartment of Psychology & Neuroscience, Duke University, Durham, NC 27708; dDepartment of Psychiatry & Behavioral Sciences, Duke UniversitySchool of Medicine, Durham, NC 27708; eCenter for Genomic and Computational Biology, Duke University, Durham, NC 27708; fSocial, Genetic, &Developmental Psychiatry Research Centre, Institute of Psychiatry, Kings College London, London SE5 8AF, United Kingdom; gDepartment of Child &Adolescent Psychiatry, Institute of Psychiatry, King’s College London, London SE5 8AF, United Kingdom; hDepartment of Psychology, The HebrewUniversity of Jerusalem, Jerusalem 91905, Israel; iDepartment of Human Genetics, Gonda Research Center, David Geffen School of Medicine, University ofCalifornia, Los Angeles, CA 90095; and jDepartment of Psychology, University of Otago, Dunedin 9016, New Zealand

Edited by Bruce S. McEwen, The Rockefeller University, New York, NY, and approved June 1, 2015 (received for review March 30, 2015)

Antiaging therapies show promise in model organism research.Translation to humans is needed to address the challenges of anaging global population. Interventions to slow human aging willneed to be applied to still-young individuals. However, most humanaging research examines older adults, manywith chronic disease. Asa result, little is known about aging in young humans. We studiedaging in 954 young humans, the Dunedin Study birth cohort,tracking multiple biomarkers across three time points spanningtheir third and fourth decades of life. We developed and validatedtwo methods by which aging can be measured in young adults,one cross-sectional and one longitudinal. Our longitudinal mea-sure allows quantification of the pace of coordinated physiologicaldeterioration across multiple organ systems (e.g., pulmonary,periodontal, cardiovascular, renal, hepatic, and immune function).We applied these methods to assess biological aging in younghumans who had not yet developed age-related diseases. Youngindividuals of the same chronological age varied in their “biologicalaging” (declining integrity of multiple organ systems). Already,before midlife, individuals who were aging more rapidly were lessphysically able, showed cognitive decline and brain aging, self-reported worse health, and looked older. Measured biologicalaging in young adults can be used to identify causes of agingand evaluate rejuvenation therapies.

biological aging | cognitive aging | aging | healthspan | geroscience

By 2050, the world population aged 80 y and above will morethan triple, approaching 400 million individuals (1, 2). As the

population ages, the global burden of disease and disability isrising (3). From the fifth decade of life, advancing age is asso-ciated with an exponential increase in burden from many differentchronic conditions (Fig. 1). The most effective means to reducedisease burden and control costs is to delay this progression byextending healthspan, years of life lived free of disease and dis-ability (4). A key to extending healthspan is addressing the prob-lem of aging itself (5–8).At present, much research on aging is being carried out with

animals and older humans. Paradoxically, these seemingly sen-sible strategies pose translational difficulties. The difficulty withstudying aging in old humans is that many of them already haveage-related diseases (9–11). Age-related changes to physiologyaccumulate from early life, affecting organ systems years beforedisease diagnosis (12–15). Thus, intervention to reverse or delaythe march toward age-related diseases must be scheduled whilepeople are still young (16). Early interventions to slow aging canbe tested in model organisms (17, 18). The difficulty with thesenonhuman models is that they do not typically capture thecomplex multifactorial risks and exposures that shape humanaging. Moreover, whereas animals’ brief lives make it feasible tostudy animal aging in the laboratory, humans’ lives span manyyears. A solution is to study human aging in the first half of thelife course, when individuals are starting to diverge in their aging

trajectories, before most diseases (and regimens to managethem) become established. The main obstacle to studying agingbefore old age—and before the onset of age-related diseases—is the absence of methods to quantify the Pace of Aging inyoung humans.We studied aging in a population-representative 1972–1973 birth

cohort of 1,037 young adults followed from birth to age 38 y with95% retention: the Dunedin Study (SI Appendix). When they were38 y old, we examined their physiologies to test whether this youngpopulation would show evidence of individual variation in agingdespite remaining free of age-related disease. We next tested thehypothesis that cohort members with “older” physiologies at age 38had actually been aging faster than their same chronologically agedpeers who retained “younger” physiologies; specifically, we testedwhether indicators of the integrity of their cardiovascular, meta-bolic, and immune systems, their kidneys, livers, gums, and lungs,and their DNA had deteriorated more rapidly according to mea-surements taken repeatedly since a baseline 12 y earlier at age 26.We further tested whether, by midlife, young adults who wereaging more rapidly already exhibited deficits in their physicalfunctioning, showed signs of early cognitive decline, and lookedolder to independent observers.

ResultsAre Young Adults Aging at Different Rates? Measuring the aging pro-cess is controversial. Candidate biomarkers of aging are numerous,

Significance

The global population is aging, driving up age-related diseasemorbidity. Antiaging interventions are needed to reduce theburden of disease and protect population productivity. Youngpeople are the most attractive targets for therapies to extendhealthspan (because it is still possible to prevent disease in theyoung). However, there is skepticism about whether agingprocesses can be detected in young adults who do not yet havechronic diseases. Our findings indicate that aging processes canbe quantified in people still young enough for prevention ofage-related disease, opening a new door for antiaging thera-pies. The science of healthspan extension may be focused onthe wrong end of the lifespan; rather than only studying oldhumans, geroscience should also study the young.

Author contributions: D.W.B., A.C., R.P., and T.E.M. designed research; D.W.B., A.C., R.H.,H.J.C., D.L.C., A.D., H.H., S.I., M.E.L., J.D.S., K.S., B.W., A.I.Y., R.P., and T.E.M. performedresearch; M.E.L. contributed new reagents/analytic tools; D.W.B., A.C., R.H., H.H., andT.E.M. analyzed data; and D.W.B., A.C., and T.E.M. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: dbelsky@duke.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1506264112/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1506264112 PNAS Early Edition | 1 of 7

MED

ICALSC

IENCE

SPS

YCHO

LOGICALAND

COGNITIVESC

IENCE

SPN

ASPL

US

Belsky et al. 2015 PNAS

AlbuminAlkalinePhosphataseBUNCreatinineCRPHbA1cSBPTotalcholesterolCMVFEV1

Klemera &Doubal 2006Mech AgDev

Levine2013JGeron A

Klemera-Doubal MethodBiologicalAge

38

Cross-sectional measurement reflects recent rate of change

Quantification of biological aging in young adultsDaniel W. Belskya,b,1, Avshalom Caspic,d,e,f, Renate Houtsc, Harvey J. Cohena, David L. Corcorane, Andrea Danesef,g,HonaLee Harringtonc, Salomon Israelh, Morgan E. Levinei, Jonathan D. Schaeferc, Karen Sugdenc, Ben Williamsc,Anatoli I. Yashinb, Richie Poultonj, and Terrie E. Moffittc,d,e,f

aDepartment of Medicine, Duke University School of Medicine, Durham, NC 27710; bSocial Science Research Institute, Duke University, Durham, NC 27708;cDepartment of Psychology & Neuroscience, Duke University, Durham, NC 27708; dDepartment of Psychiatry & Behavioral Sciences, Duke UniversitySchool of Medicine, Durham, NC 27708; eCenter for Genomic and Computational Biology, Duke University, Durham, NC 27708; fSocial, Genetic, &Developmental Psychiatry Research Centre, Institute of Psychiatry, Kings College London, London SE5 8AF, United Kingdom; gDepartment of Child &Adolescent Psychiatry, Institute of Psychiatry, King’s College London, London SE5 8AF, United Kingdom; hDepartment of Psychology, The HebrewUniversity of Jerusalem, Jerusalem 91905, Israel; iDepartment of Human Genetics, Gonda Research Center, David Geffen School of Medicine, University ofCalifornia, Los Angeles, CA 90095; and jDepartment of Psychology, University of Otago, Dunedin 9016, New Zealand

Edited by Bruce S. McEwen, The Rockefeller University, New York, NY, and approved June 1, 2015 (received for review March 30, 2015)

Antiaging therapies show promise in model organism research.Translation to humans is needed to address the challenges of anaging global population. Interventions to slow human aging willneed to be applied to still-young individuals. However, most humanaging research examines older adults, manywith chronic disease. Asa result, little is known about aging in young humans. We studiedaging in 954 young humans, the Dunedin Study birth cohort,tracking multiple biomarkers across three time points spanningtheir third and fourth decades of life. We developed and validatedtwo methods by which aging can be measured in young adults,one cross-sectional and one longitudinal. Our longitudinal mea-sure allows quantification of the pace of coordinated physiologicaldeterioration across multiple organ systems (e.g., pulmonary,periodontal, cardiovascular, renal, hepatic, and immune function).We applied these methods to assess biological aging in younghumans who had not yet developed age-related diseases. Youngindividuals of the same chronological age varied in their “biologicalaging” (declining integrity of multiple organ systems). Already,before midlife, individuals who were aging more rapidly were lessphysically able, showed cognitive decline and brain aging, self-reported worse health, and looked older. Measured biologicalaging in young adults can be used to identify causes of agingand evaluate rejuvenation therapies.

biological aging | cognitive aging | aging | healthspan | geroscience

By 2050, the world population aged 80 y and above will morethan triple, approaching 400 million individuals (1, 2). As the

population ages, the global burden of disease and disability isrising (3). From the fifth decade of life, advancing age is asso-ciated with an exponential increase in burden from many differentchronic conditions (Fig. 1). The most effective means to reducedisease burden and control costs is to delay this progression byextending healthspan, years of life lived free of disease and dis-ability (4). A key to extending healthspan is addressing the prob-lem of aging itself (5–8).At present, much research on aging is being carried out with

animals and older humans. Paradoxically, these seemingly sen-sible strategies pose translational difficulties. The difficulty withstudying aging in old humans is that many of them already haveage-related diseases (9–11). Age-related changes to physiologyaccumulate from early life, affecting organ systems years beforedisease diagnosis (12–15). Thus, intervention to reverse or delaythe march toward age-related diseases must be scheduled whilepeople are still young (16). Early interventions to slow aging canbe tested in model organisms (17, 18). The difficulty with thesenonhuman models is that they do not typically capture thecomplex multifactorial risks and exposures that shape humanaging. Moreover, whereas animals’ brief lives make it feasible tostudy animal aging in the laboratory, humans’ lives span manyyears. A solution is to study human aging in the first half of thelife course, when individuals are starting to diverge in their aging

trajectories, before most diseases (and regimens to managethem) become established. The main obstacle to studying agingbefore old age—and before the onset of age-related diseases—is the absence of methods to quantify the Pace of Aging inyoung humans.We studied aging in a population-representative 1972–1973 birth

cohort of 1,037 young adults followed from birth to age 38 y with95% retention: the Dunedin Study (SI Appendix). When they were38 y old, we examined their physiologies to test whether this youngpopulation would show evidence of individual variation in agingdespite remaining free of age-related disease. We next tested thehypothesis that cohort members with “older” physiologies at age 38had actually been aging faster than their same chronologically agedpeers who retained “younger” physiologies; specifically, we testedwhether indicators of the integrity of their cardiovascular, meta-bolic, and immune systems, their kidneys, livers, gums, and lungs,and their DNA had deteriorated more rapidly according to mea-surements taken repeatedly since a baseline 12 y earlier at age 26.We further tested whether, by midlife, young adults who wereaging more rapidly already exhibited deficits in their physicalfunctioning, showed signs of early cognitive decline, and lookedolder to independent observers.

ResultsAre Young Adults Aging at Different Rates? Measuring the aging pro-cess is controversial. Candidate biomarkers of aging are numerous,

Significance

The global population is aging, driving up age-related diseasemorbidity. Antiaging interventions are needed to reduce theburden of disease and protect population productivity. Youngpeople are the most attractive targets for therapies to extendhealthspan (because it is still possible to prevent disease in theyoung). However, there is skepticism about whether agingprocesses can be detected in young adults who do not yet havechronic diseases. Our findings indicate that aging processes canbe quantified in people still young enough for prevention ofage-related disease, opening a new door for antiaging thera-pies. The science of healthspan extension may be focused onthe wrong end of the lifespan; rather than only studying oldhumans, geroscience should also study the young.

Author contributions: D.W.B., A.C., R.P., and T.E.M. designed research; D.W.B., A.C., R.H.,H.J.C., D.L.C., A.D., H.H., S.I., M.E.L., J.D.S., K.S., B.W., A.I.Y., R.P., and T.E.M. performedresearch; M.E.L. contributed new reagents/analytic tools; D.W.B., A.C., R.H., H.H., andT.E.M. analyzed data; and D.W.B., A.C., and T.E.M. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: dbelsky@duke.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1506264112/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1506264112 PNAS Early Edition | 1 of 7

MED

ICALSC

IENCE

SPS

YCHO

LOGICALAND

COGNITIVESC

IENCE

SPN

ASPL

US

Belsky et al. 2015 PNAS

AlbuminAlkalinePhosphataseBUNCreatinineCRPHbA1cSBPTotalcholesterolCMVFEV1

Klemera &Doubal 2006Mech AgDev

Levine2013JGeron A

Klemera-Doubal MethodBiologicalAge

39

Does variation in rate of aging predict signs of aging?

40

Faster aging predicts poor physical function

Quantification of biological aging in young adultsDaniel W. Belskya,b,1, Avshalom Caspic,d,e,f, Renate Houtsc, Harvey J. Cohena, David L. Corcorane, Andrea Danesef,g,HonaLee Harringtonc, Salomon Israelh, Morgan E. Levinei, Jonathan D. Schaeferc, Karen Sugdenc, Ben Williamsc,Anatoli I. Yashinb, Richie Poultonj, and Terrie E. Moffittc,d,e,f

aDepartment of Medicine, Duke University School of Medicine, Durham, NC 27710; bSocial Science Research Institute, Duke University, Durham, NC 27708;cDepartment of Psychology & Neuroscience, Duke University, Durham, NC 27708; dDepartment of Psychiatry & Behavioral Sciences, Duke UniversitySchool of Medicine, Durham, NC 27708; eCenter for Genomic and Computational Biology, Duke University, Durham, NC 27708; fSocial, Genetic, &Developmental Psychiatry Research Centre, Institute of Psychiatry, Kings College London, London SE5 8AF, United Kingdom; gDepartment of Child &Adolescent Psychiatry, Institute of Psychiatry, King’s College London, London SE5 8AF, United Kingdom; hDepartment of Psychology, The HebrewUniversity of Jerusalem, Jerusalem 91905, Israel; iDepartment of Human Genetics, Gonda Research Center, David Geffen School of Medicine, University ofCalifornia, Los Angeles, CA 90095; and jDepartment of Psychology, University of Otago, Dunedin 9016, New Zealand

Edited by Bruce S. McEwen, The Rockefeller University, New York, NY, and approved June 1, 2015 (received for review March 30, 2015)

Antiaging therapies show promise in model organism research.Translation to humans is needed to address the challenges of anaging global population. Interventions to slow human aging willneed to be applied to still-young individuals. However, most humanaging research examines older adults, manywith chronic disease. Asa result, little is known about aging in young humans. We studiedaging in 954 young humans, the Dunedin Study birth cohort,tracking multiple biomarkers across three time points spanningtheir third and fourth decades of life. We developed and validatedtwo methods by which aging can be measured in young adults,one cross-sectional and one longitudinal. Our longitudinal mea-sure allows quantification of the pace of coordinated physiologicaldeterioration across multiple organ systems (e.g., pulmonary,periodontal, cardiovascular, renal, hepatic, and immune function).We applied these methods to assess biological aging in younghumans who had not yet developed age-related diseases. Youngindividuals of the same chronological age varied in their “biologicalaging” (declining integrity of multiple organ systems). Already,before midlife, individuals who were aging more rapidly were lessphysically able, showed cognitive decline and brain aging, self-reported worse health, and looked older. Measured biologicalaging in young adults can be used to identify causes of agingand evaluate rejuvenation therapies.

biological aging | cognitive aging | aging | healthspan | geroscience

By 2050, the world population aged 80 y and above will morethan triple, approaching 400 million individuals (1, 2). As the

population ages, the global burden of disease and disability isrising (3). From the fifth decade of life, advancing age is asso-ciated with an exponential increase in burden from many differentchronic conditions (Fig. 1). The most effective means to reducedisease burden and control costs is to delay this progression byextending healthspan, years of life lived free of disease and dis-ability (4). A key to extending healthspan is addressing the prob-lem of aging itself (5–8).At present, much research on aging is being carried out with

animals and older humans. Paradoxically, these seemingly sen-sible strategies pose translational difficulties. The difficulty withstudying aging in old humans is that many of them already haveage-related diseases (9–11). Age-related changes to physiologyaccumulate from early life, affecting organ systems years beforedisease diagnosis (12–15). Thus, intervention to reverse or delaythe march toward age-related diseases must be scheduled whilepeople are still young (16). Early interventions to slow aging canbe tested in model organisms (17, 18). The difficulty with thesenonhuman models is that they do not typically capture thecomplex multifactorial risks and exposures that shape humanaging. Moreover, whereas animals’ brief lives make it feasible tostudy animal aging in the laboratory, humans’ lives span manyyears. A solution is to study human aging in the first half of thelife course, when individuals are starting to diverge in their aging

trajectories, before most diseases (and regimens to managethem) become established. The main obstacle to studying agingbefore old age—and before the onset of age-related diseases—is the absence of methods to quantify the Pace of Aging inyoung humans.We studied aging in a population-representative 1972–1973 birth

cohort of 1,037 young adults followed from birth to age 38 y with95% retention: the Dunedin Study (SI Appendix). When they were38 y old, we examined their physiologies to test whether this youngpopulation would show evidence of individual variation in agingdespite remaining free of age-related disease. We next tested thehypothesis that cohort members with “older” physiologies at age 38had actually been aging faster than their same chronologically agedpeers who retained “younger” physiologies; specifically, we testedwhether indicators of the integrity of their cardiovascular, meta-bolic, and immune systems, their kidneys, livers, gums, and lungs,and their DNA had deteriorated more rapidly according to mea-surements taken repeatedly since a baseline 12 y earlier at age 26.We further tested whether, by midlife, young adults who wereaging more rapidly already exhibited deficits in their physicalfunctioning, showed signs of early cognitive decline, and lookedolder to independent observers.

ResultsAre Young Adults Aging at Different Rates? Measuring the aging pro-cess is controversial. Candidate biomarkers of aging are numerous,

Significance

The global population is aging, driving up age-related diseasemorbidity. Antiaging interventions are needed to reduce theburden of disease and protect population productivity. Youngpeople are the most attractive targets for therapies to extendhealthspan (because it is still possible to prevent disease in theyoung). However, there is skepticism about whether agingprocesses can be detected in young adults who do not yet havechronic diseases. Our findings indicate that aging processes canbe quantified in people still young enough for prevention ofage-related disease, opening a new door for antiaging thera-pies. The science of healthspan extension may be focused onthe wrong end of the lifespan; rather than only studying oldhumans, geroscience should also study the young.

Author contributions: D.W.B., A.C., R.P., and T.E.M. designed research; D.W.B., A.C., R.H.,H.J.C., D.L.C., A.D., H.H., S.I., M.E.L., J.D.S., K.S., B.W., A.I.Y., R.P., and T.E.M. performedresearch; M.E.L. contributed new reagents/analytic tools; D.W.B., A.C., R.H., H.H., andT.E.M. analyzed data; and D.W.B., A.C., and T.E.M. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: dbelsky@duke.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1506264112/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1506264112 PNAS Early Edition | 1 of 7

MED

ICALSC

IENCE

SPS

YCHO

LOGICALAND

COGNITIVESC

IENCE

SPN

ASPL

US

Belsky et al. 2015 PNAS

41

Faster aging predicts declining brain health

Quantification of biological aging in young adultsDaniel W. Belskya,b,1, Avshalom Caspic,d,e,f, Renate Houtsc, Harvey J. Cohena, David L. Corcorane, Andrea Danesef,g,HonaLee Harringtonc, Salomon Israelh, Morgan E. Levinei, Jonathan D. Schaeferc, Karen Sugdenc, Ben Williamsc,Anatoli I. Yashinb, Richie Poultonj, and Terrie E. Moffittc,d,e,f

aDepartment of Medicine, Duke University School of Medicine, Durham, NC 27710; bSocial Science Research Institute, Duke University, Durham, NC 27708;cDepartment of Psychology & Neuroscience, Duke University, Durham, NC 27708; dDepartment of Psychiatry & Behavioral Sciences, Duke UniversitySchool of Medicine, Durham, NC 27708; eCenter for Genomic and Computational Biology, Duke University, Durham, NC 27708; fSocial, Genetic, &Developmental Psychiatry Research Centre, Institute of Psychiatry, Kings College London, London SE5 8AF, United Kingdom; gDepartment of Child &Adolescent Psychiatry, Institute of Psychiatry, King’s College London, London SE5 8AF, United Kingdom; hDepartment of Psychology, The HebrewUniversity of Jerusalem, Jerusalem 91905, Israel; iDepartment of Human Genetics, Gonda Research Center, David Geffen School of Medicine, University ofCalifornia, Los Angeles, CA 90095; and jDepartment of Psychology, University of Otago, Dunedin 9016, New Zealand

Edited by Bruce S. McEwen, The Rockefeller University, New York, NY, and approved June 1, 2015 (received for review March 30, 2015)

Antiaging therapies show promise in model organism research.Translation to humans is needed to address the challenges of anaging global population. Interventions to slow human aging willneed to be applied to still-young individuals. However, most humanaging research examines older adults, manywith chronic disease. Asa result, little is known about aging in young humans. We studiedaging in 954 young humans, the Dunedin Study birth cohort,tracking multiple biomarkers across three time points spanningtheir third and fourth decades of life. We developed and validatedtwo methods by which aging can be measured in young adults,one cross-sectional and one longitudinal. Our longitudinal mea-sure allows quantification of the pace of coordinated physiologicaldeterioration across multiple organ systems (e.g., pulmonary,periodontal, cardiovascular, renal, hepatic, and immune function).We applied these methods to assess biological aging in younghumans who had not yet developed age-related diseases. Youngindividuals of the same chronological age varied in their “biologicalaging” (declining integrity of multiple organ systems). Already,before midlife, individuals who were aging more rapidly were lessphysically able, showed cognitive decline and brain aging, self-reported worse health, and looked older. Measured biologicalaging in young adults can be used to identify causes of agingand evaluate rejuvenation therapies.

biological aging | cognitive aging | aging | healthspan | geroscience

By 2050, the world population aged 80 y and above will morethan triple, approaching 400 million individuals (1, 2). As the

population ages, the global burden of disease and disability isrising (3). From the fifth decade of life, advancing age is asso-ciated with an exponential increase in burden from many differentchronic conditions (Fig. 1). The most effective means to reducedisease burden and control costs is to delay this progression byextending healthspan, years of life lived free of disease and dis-ability (4). A key to extending healthspan is addressing the prob-lem of aging itself (5–8).At present, much research on aging is being carried out with

animals and older humans. Paradoxically, these seemingly sen-sible strategies pose translational difficulties. The difficulty withstudying aging in old humans is that many of them already haveage-related diseases (9–11). Age-related changes to physiologyaccumulate from early life, affecting organ systems years beforedisease diagnosis (12–15). Thus, intervention to reverse or delaythe march toward age-related diseases must be scheduled whilepeople are still young (16). Early interventions to slow aging canbe tested in model organisms (17, 18). The difficulty with thesenonhuman models is that they do not typically capture thecomplex multifactorial risks and exposures that shape humanaging. Moreover, whereas animals’ brief lives make it feasible tostudy animal aging in the laboratory, humans’ lives span manyyears. A solution is to study human aging in the first half of thelife course, when individuals are starting to diverge in their aging

trajectories, before most diseases (and regimens to managethem) become established. The main obstacle to studying agingbefore old age—and before the onset of age-related diseases—is the absence of methods to quantify the Pace of Aging inyoung humans.We studied aging in a population-representative 1972–1973 birth

cohort of 1,037 young adults followed from birth to age 38 y with95% retention: the Dunedin Study (SI Appendix). When they were38 y old, we examined their physiologies to test whether this youngpopulation would show evidence of individual variation in agingdespite remaining free of age-related disease. We next tested thehypothesis that cohort members with “older” physiologies at age 38had actually been aging faster than their same chronologically agedpeers who retained “younger” physiologies; specifically, we testedwhether indicators of the integrity of their cardiovascular, meta-bolic, and immune systems, their kidneys, livers, gums, and lungs,and their DNA had deteriorated more rapidly according to mea-surements taken repeatedly since a baseline 12 y earlier at age 26.We further tested whether, by midlife, young adults who wereaging more rapidly already exhibited deficits in their physicalfunctioning, showed signs of early cognitive decline, and lookedolder to independent observers.

ResultsAre Young Adults Aging at Different Rates? Measuring the aging pro-cess is controversial. Candidate biomarkers of aging are numerous,

Significance

The global population is aging, driving up age-related diseasemorbidity. Antiaging interventions are needed to reduce theburden of disease and protect population productivity. Youngpeople are the most attractive targets for therapies to extendhealthspan (because it is still possible to prevent disease in theyoung). However, there is skepticism about whether agingprocesses can be detected in young adults who do not yet havechronic diseases. Our findings indicate that aging processes canbe quantified in people still young enough for prevention ofage-related disease, opening a new door for antiaging thera-pies. The science of healthspan extension may be focused onthe wrong end of the lifespan; rather than only studying oldhumans, geroscience should also study the young.

Author contributions: D.W.B., A.C., R.P., and T.E.M. designed research; D.W.B., A.C., R.H.,H.J.C., D.L.C., A.D., H.H., S.I., M.E.L., J.D.S., K.S., B.W., A.I.Y., R.P., and T.E.M. performedresearch; M.E.L. contributed new reagents/analytic tools; D.W.B., A.C., R.H., H.H., andT.E.M. analyzed data; and D.W.B., A.C., and T.E.M. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: dbelsky@duke.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1506264112/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1506264112 PNAS Early Edition | 1 of 7

MED

ICALSC

IENCE

SPS

YCHO

LOGICALAND

COGNITIVESC

IENCE

SPN

ASPL

US

Belsky et al. 2015 PNAS

42

Faster aging predicts subjective signs

Quantification of biological aging in young adultsDaniel W. Belskya,b,1, Avshalom Caspic,d,e,f, Renate Houtsc, Harvey J. Cohena, David L. Corcorane, Andrea Danesef,g,HonaLee Harringtonc, Salomon Israelh, Morgan E. Levinei, Jonathan D. Schaeferc, Karen Sugdenc, Ben Williamsc,Anatoli I. Yashinb, Richie Poultonj, and Terrie E. Moffittc,d,e,f

aDepartment of Medicine, Duke University School of Medicine, Durham, NC 27710; bSocial Science Research Institute, Duke University, Durham, NC 27708;cDepartment of Psychology & Neuroscience, Duke University, Durham, NC 27708; dDepartment of Psychiatry & Behavioral Sciences, Duke UniversitySchool of Medicine, Durham, NC 27708; eCenter for Genomic and Computational Biology, Duke University, Durham, NC 27708; fSocial, Genetic, &Developmental Psychiatry Research Centre, Institute of Psychiatry, Kings College London, London SE5 8AF, United Kingdom; gDepartment of Child &Adolescent Psychiatry, Institute of Psychiatry, King’s College London, London SE5 8AF, United Kingdom; hDepartment of Psychology, The HebrewUniversity of Jerusalem, Jerusalem 91905, Israel; iDepartment of Human Genetics, Gonda Research Center, David Geffen School of Medicine, University ofCalifornia, Los Angeles, CA 90095; and jDepartment of Psychology, University of Otago, Dunedin 9016, New Zealand

Edited by Bruce S. McEwen, The Rockefeller University, New York, NY, and approved June 1, 2015 (received for review March 30, 2015)

Antiaging therapies show promise in model organism research.Translation to humans is needed to address the challenges of anaging global population. Interventions to slow human aging willneed to be applied to still-young individuals. However, most humanaging research examines older adults, manywith chronic disease. Asa result, little is known about aging in young humans. We studiedaging in 954 young humans, the Dunedin Study birth cohort,tracking multiple biomarkers across three time points spanningtheir third and fourth decades of life. We developed and validatedtwo methods by which aging can be measured in young adults,one cross-sectional and one longitudinal. Our longitudinal mea-sure allows quantification of the pace of coordinated physiologicaldeterioration across multiple organ systems (e.g., pulmonary,periodontal, cardiovascular, renal, hepatic, and immune function).We applied these methods to assess biological aging in younghumans who had not yet developed age-related diseases. Youngindividuals of the same chronological age varied in their “biologicalaging” (declining integrity of multiple organ systems). Already,before midlife, individuals who were aging more rapidly were lessphysically able, showed cognitive decline and brain aging, self-reported worse health, and looked older. Measured biologicalaging in young adults can be used to identify causes of agingand evaluate rejuvenation therapies.

biological aging | cognitive aging | aging | healthspan | geroscience

By 2050, the world population aged 80 y and above will morethan triple, approaching 400 million individuals (1, 2). As the

population ages, the global burden of disease and disability isrising (3). From the fifth decade of life, advancing age is asso-ciated with an exponential increase in burden from many differentchronic conditions (Fig. 1). The most effective means to reducedisease burden and control costs is to delay this progression byextending healthspan, years of life lived free of disease and dis-ability (4). A key to extending healthspan is addressing the prob-lem of aging itself (5–8).At present, much research on aging is being carried out with

animals and older humans. Paradoxically, these seemingly sen-sible strategies pose translational difficulties. The difficulty withstudying aging in old humans is that many of them already haveage-related diseases (9–11). Age-related changes to physiologyaccumulate from early life, affecting organ systems years beforedisease diagnosis (12–15). Thus, intervention to reverse or delaythe march toward age-related diseases must be scheduled whilepeople are still young (16). Early interventions to slow aging canbe tested in model organisms (17, 18). The difficulty with thesenonhuman models is that they do not typically capture thecomplex multifactorial risks and exposures that shape humanaging. Moreover, whereas animals’ brief lives make it feasible tostudy animal aging in the laboratory, humans’ lives span manyyears. A solution is to study human aging in the first half of thelife course, when individuals are starting to diverge in their aging

trajectories, before most diseases (and regimens to managethem) become established. The main obstacle to studying agingbefore old age—and before the onset of age-related diseases—is the absence of methods to quantify the Pace of Aging inyoung humans.We studied aging in a population-representative 1972–1973 birth

cohort of 1,037 young adults followed from birth to age 38 y with95% retention: the Dunedin Study (SI Appendix). When they were38 y old, we examined their physiologies to test whether this youngpopulation would show evidence of individual variation in agingdespite remaining free of age-related disease. We next tested thehypothesis that cohort members with “older” physiologies at age 38had actually been aging faster than their same chronologically agedpeers who retained “younger” physiologies; specifically, we testedwhether indicators of the integrity of their cardiovascular, meta-bolic, and immune systems, their kidneys, livers, gums, and lungs,and their DNA had deteriorated more rapidly according to mea-surements taken repeatedly since a baseline 12 y earlier at age 26.We further tested whether, by midlife, young adults who wereaging more rapidly already exhibited deficits in their physicalfunctioning, showed signs of early cognitive decline, and lookedolder to independent observers.

ResultsAre Young Adults Aging at Different Rates? Measuring the aging pro-cess is controversial. Candidate biomarkers of aging are numerous,

Significance

The global population is aging, driving up age-related diseasemorbidity. Antiaging interventions are needed to reduce theburden of disease and protect population productivity. Youngpeople are the most attractive targets for therapies to extendhealthspan (because it is still possible to prevent disease in theyoung). However, there is skepticism about whether agingprocesses can be detected in young adults who do not yet havechronic diseases. Our findings indicate that aging processes canbe quantified in people still young enough for prevention ofage-related disease, opening a new door for antiaging thera-pies. The science of healthspan extension may be focused onthe wrong end of the lifespan; rather than only studying oldhumans, geroscience should also study the young.

Author contributions: D.W.B., A.C., R.P., and T.E.M. designed research; D.W.B., A.C., R.H.,H.J.C., D.L.C., A.D., H.H., S.I., M.E.L., J.D.S., K.S., B.W., A.I.Y., R.P., and T.E.M. performedresearch; M.E.L. contributed new reagents/analytic tools; D.W.B., A.C., R.H., H.H., andT.E.M. analyzed data; and D.W.B., A.C., and T.E.M. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: dbelsky@duke.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1506264112/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1506264112 PNAS Early Edition | 1 of 7

MED

ICALSC

IENCE

SPS

YCHO

LOGICALAND

COGNITIVESC

IENCE

SPN

ASPL

US

Belsky et al. 2015 PNAS

43

Part 2 Interim Summary

• The rate of aging can be measured in healthy young adults

Repeated-measures clinical-exam and blood-test data can track aging-related changes decades in advance of disease onset

• The aging rate in healthy young adults is already variable

• A faster rate of aging correlates with deficits in physical and cognitive function and subjective signs of aging

44

Do early-life risks accelerate the page of aging?

Belsky etal.2017AgingCell

Age Year Number Percent*

Birth 1972-733 1975-76 1037 100%5 1977-78 991 967 1979-80 954 929 1981-82 955 9211 1983-84 925 9013 1985-86 850 8215 1987-88 976 9518 1990-91 993 9721 1993-94 992 9726 1998-99 980 9632 2004-05 972 96

38 2010-12 961 95%45 2017-18 ?? ??

45

Life-Course Design

Belsky etal.2017AgingCell

Age Year Number Percent*

Birth 1972-733 1975-76 1037 100%5 1977-78 991 967 1979-80 954 929 1981-82 955 9211 1983-84 925 9013 1985-86 850 8215 1987-88 976 9518 1990-91 993 9721 1993-94 992 9726 1998-99 980 9632 2004-05 972 96

38 2010-12 961 95%45 2017-18 ?? ??

46

Life-Course Design

Belsky etal.2017AgingCell

0

5

10

15

Perc

ent

-6 -4 -2 0 2Childhood Self Control

Childhood Self-Control

0

5

10

15

Perc

ent

60 80 100 120 140Childhood IQ

Childhood IQ

0

5

10

15

Perc

ent

-3 -2 -1 0 1 2Childhood Health

Childhood Health

0

10

20

30

40

Perc

ent

-1 0 1 2 3 4Adverse Childhood Experiences

Adverse Childhood Experiences

0

5

10

15

Perc

ent

50 60 70 80 90 100Age of Longest Lived Grandparent

Age of Longest-Lived Grandparent

0

5

10

15

20

Perc

ent

1 2 3 4 5 6Childhood Social Class

Childhood SES

47

Observationalratings,parent,teacher,&self-reportsofhyperactivity,lackofpersistence,inattention,impulsiveaggression,impulsivity

Ages3-11years

WechslerIntelligenceScalesforChildren(WISC)Ages7,9,11,13

LungfunctionBloodpressureAnthropometryBalance&MotorClinicalinterview

Ages3,5,7,9,11

5typesofchildharmPhysicalabuseEmotionalabuseSexualabusePhysicalneglectEmotionalneglect

5typesofhouseholddysfunctionSubstanceabuseMentalillnessIncarcerationPartnerviolenceParentallossAges3-15

BasedonparentaloccupationBirth-age15

FamilialLongevity

Ageoflongest-livedgrandparent

ChildhoodSES ACEs ChildhoodHealth ChildhoodIQ ChildhoodSelf-Control

Belsky etal.2017AgingCell

48

Midlife pace of aging has origins in childhood

Belsky etal.2017AgingCell

49

<15minbattery5-pointscale

Belsky etal.2017AgingCell

Construct Measurement

FamilialLongevity Grandparent >80y(lifeexpectancy)

ChildhoodSocialClass Low (e.g.low-skilloccupation)

ACEs CDCACEInventory

ChildhoodIQ Loweducation

ChildhoodSelf-Control 5-item Nurserating

Retrospective Personal-History Risk Assessment for Screening Geroprotector Trial Participants

50

5332

15

54

1927

4322

35

4733

20

45

32

23

4031

29

31 28 41 30 26 43 28 14 58

0%

20%

40%

60%

80%

100%

NoRisks 1Risk

2+Risks

NoRisks 1Risk

2+Risks

NoRisks 1Risk

2+Risks

CohortPrescriptionRegister

(N=686)HospitalAdmission

(N=183)

Fast

Normal

Slow

PaceofAging

Belsky etal.2017AgingCell

Personal-history risk assessment identifies fast-aging group

51

Part 2 Summary

• The rate of aging can be measured in healthy young adults

Repeated-measures clinical-exam and blood-test data can track aging-related changes decades in advance of disease onset

• The aging rate in healthy young adults is already variable

• A faster rate of aging correlates with deficits in physical and cognitive function and subjective signs of aging

• Midlife aging rate has origins in childhood

Possible to screen participants to balance enrollment in geroprotector trials

Outline

IntroductionBiological aging is a treatment target for healthspan extension

Part 1. Human trials of geroprotectorsChallenges & opportunities

Part 2. Quantification of biological aging in young adults

Part 3. Testing biological aging in a geroprotector trial

Conclusion

53

Alternative Geroprotector Trial Design

Biologica

lAge

Y1 Y2 Y3

Control

Treatment

Rx

Control

SlowerPaceofAging

NormalPaceofAging

Years

54

https://calerie.duke.edu/

Weindruch &Walford 1982ScienceColmanetal.2009ScienceMattison etal.2014NatureColmanetal.2014NatComm

55

N=145Randomizedto25%CR(12%achieved)N=75RandomizedtoAL(2%CRonaverage)

Ravussin etal.2015JGMS

56

Belsky etal.2017JGBS

T1 T2 T3

Baseline 12mo 24mo

Biologica

lAge

Y1 Y2 Y3

Control

Treatment

Hypothesis

CR

AdLibitum

57

Belsky etal.2017JGBS

T1 T2 T3

Baseline 12mo 24mo

Biologica

lAge

Y1 Y2 Y3

Control

TreatmentCR

AdLibitumAlbuminAlkalinePhosphataseBUNCreatinineCRPHbA1c(glucose)SBPTotalcholesterolUricAcidWBC

Klemera &Doubal 2006Mech AgDevLevine2013JGBSCohenetal.2013Mech AgDevLietal.2015AgingCell

Repeatedmeasuresofbiologicalage• Klemera-Doubal methodBiologicalAge• HomeostaticDysregulation

58

Belsky etal.2017JGBS

T1 T2 T3

Baseline 12mo 24mo

Klemera-Doubal methodBiologicalAgeatbaseline

59

Belsky etal.2017JGBS

T1 T2 T3

Baseline 12mo 24mo

-1

0

1

2

3

Cha

nge

in K

DM

Bio

logi

cal A

ge (y

ears

)

Baseline 12-months 24-months CALERIE Assessment

60

Belsky etal.2017JGBS

T1 T2 T3

Baseline 12mo 24mo

-1

0

1

2

3

Cha

nge

in K

DM

Bio

logi

cal A

ge (y

ears

)

Baseline 12-months 24-months CALERIE Assessment

Ad LibitumCaloric Restriction

Randomized to

61

Part 3 Summary

• Moderate (~10%) CR slows the rate of biological aging as measured from physiology

• Implementation of biological aging measures within already-collected data from RCTs can advance validation efforts

• Long-term follow-up will be needed (and should be planned for)

Outline

IntroductionBiological aging is a treatment target for healthspan extension

Part 1. Human trials of geroprotectorsChallenges & opportunities

Part 2. Quantification of biological aging in young adults

Part 3. Testing biological aging in a geroprotector trial

Conclusion

63

Conclusions

• Aging is a lifelong process with effects already manifest by midlife

• The midlife rate of aging is variable in apparently healthy adults

Interventions to slow aging should begin early

• The midlife rate of aging can be measured

• The midlife rate of aging can be modified (e.g. by CR)

Geroprotector trials should consider the rate of change in biological age as a surrogate endpoint for healthspan extension

64

Not all measures of aging measure the same thing

Belsky et al. AJE 2017

65

Not all measures of aging measure the same thing

Belsky et al. AJE 2017

66

Next Steps

• Refine measures to quantify biological aging

Critical validation is that rate of change predicts healthspan

• Expand scope of intervention testing

Exercise trials and other lifestyle interventions are low hanging fruit

But Rx is the frontier

67

ThankYou!