(Re)Setting Epigenetic Clocks: An Important Avenue Whereby Social Conditions Become Biologically Embedded across the Life Course

Abstract

Research on biological embedding of the social environment has been expedited by increased availability of biomarkers. Recently, this arsenal of measures has been expanded to include epigenetic clocks that indicate in years the extent to which an individual is older or younger than their chronological age. These measures of biological aging, especially GrimAge, are robust predictors of both illness and time to death. Importantly for sociologists, several studies have linked social conditions to these indices of aging. The present study extends this research using longitudinal data from a sample of 223 black women participating in the Family and Community Health Study. We find that changes in income and living arrangements over an 11-year period predict changes in speed of biological aging. These results provide further support for the idea that epigenetic aging is a mechanism whereby social conditions become biologically embedded. The utility of epigenetic clocks for sociological studies of health are discussed.

Keywords

biological aging biological embedding epigenetic clocks health disparities

For decades, self-ratings of health have been the most popular method for measuring health status. Unfortunately, a large body of research now indicates that there are significant discrepancies between self-ratings of health and other, more biologically based assessments. This discrepancy appears to increase with age so that onset of chronic illness in old age often has little impact on the elderly’s self-reported health (Idler and Cartwright 2018). And studies often find weak income gradients in self-reported health (Au and Johnson 2014), contrary to research using biological assessments of illness (Marmot 2015). Such findings raise questions regarding what self-reported health actually assesses. Furthermore, they suggest that research relying on self-ratings of health may either overestimate or fail to identify significant social determinants of physical illness (Au and Johnston 2014). For instance, Dowd and Todd (2011) found that failure to take into account gender, age, education, and racial-ethnic differences in health self-appraisals lead to significant underestimation of health inequalities. Furthermore, recent longitudinal studies have reported that childhood trauma is only related to adult self-rated health when the respondent recalls such childhood events, whereas objective indicators of childhood trauma predict biomarkers of health regardless of whether adult respondents recall these childhood events (Berg et al. 2020; Osborn and Widom 2019; Reuben et al. 2016).

In addition to these validity and reliability concerns, self-ratings of health are of little utility when investigating the mechanisms whereby social experiences get under the skin and make people sick. In the past 15 years, biological embedding of the social environment has emerged as an important research focus in medical sociology as well as in other health-related disciplines. Investigation of this issue has required that objective biomarkers of health be substituted for previously utilized self-report measures. Sociologists have begun to utilize, for example, biological indicators of cardiometabolic processes, allostastic load, systemic inflammation, and telomere length (Harris and Schorpp 2018).

In recent years, this arsenal of biomarkers has been expanded to include epigenetic clocks that indicate in years the extent to which an individual is older or younger than their chronological age (Bell et al. 2019; Ryan 2020). These epigenetic indices of speed of biological aging have been shown to be powerful predictors of a variety of chronic illnesses as well as time to death (Hillary et al. 2020; Levine et al. 2018; Lu et al. 2019; McCrory et al. 2020). And importantly for social and behavioral scientists, research using these clocks has provided evidence that social conditions and relationships may influence an individual’s speed of biological aging (Brody et al. 2016; Lei et al. 2019; Simons et al. 2016, 2020; Zannas et al. 2015). Sociologists have been concerned with the way that the stress and adversity experienced by members of low-status groups (e.g., persons of color, low social class) may foster physiological weathering, which, in turn, gives rise to early illness and death (Geronimus 2001). Epigenetic clocks might be viewed as robust indicators of such weathering (Simons et al. 2016, 2020). The fact that these epigenetic clocks predict morbidity and mortality far in advance makes them powerful instruments for investigating illness and aging without long-term follow-up studies (Belsky et al. 2020; Ryan 2020).

Regrettably, most medical sociologists are not familiar with these epigenetic measures. The present article attempts to remedy this situation. We begin by providing a description of the various epigenetic clocks that have been developed, placing special emphasis on the recently developed Grim Index of accelerated aging (Lu et al. 2019). Having described the Grim Index, we use longitudinal data from a sample of 223 middle-aged African American women to examine the extent to which indicators of reduced income and loss of independence predict accelerated Grim aging. Our analyses are unique because they go beyond cross-sectional explorations to investigate the extent to which changes in social conditions over an 11-year period are associated with change in speed of epigenetic aging. To our knowledge, this is the first article to consider such effects.

Backgound

Epigenetic Indices of Biological Aging (Epigenetic Clocks)

As a consequence of recent gains in average life span, the proportion of the world’s population over 65 years of age has gone from 5% (roughly 130 million) in 1950 to a projected 17% (1.6 billion) by 2050 (He, Goodkind, and Kowal 2016). Unfortunately, this decline in mortality has not been paired with a corresponding decrease in chronic illness. Thus, individuals often spend a prolonged period of their lives in ill health. In an attempt to increase healthy aging, medical research has focused on identifying causes and treatments for the various diseases of old age. Increasingly, however, it has become apparent that these illnesses share the same root cause—biological aging—and that increasing the health span will require a better understanding of the fundamental causes of aging (Kennedy et al. 2014). This fact argues for more research on factors, including social events and circumstances, that accelerate biological aging (Moffitt 2020). Research on accelerated aging requires, of course, the availability of robust measures of the speed of biological aging (Belsky et al. 2015). Biological aging consists of a gradual decline with age in system integrity, resulting in morbidity, disability, and ultimately death. In the past decade, it has become evident that broad-based epigenetic changes can be used to assess the rate at which this progression is taking place.

Epigenetic regulation involves biochemical mechanisms that influence genome expression by either up-regulating or down-regulating particular genes. One of the most pervasive and well studied of these mechanisms is methylation. This process occurs when a methyl group attaches to a segment of deoxyribonucleic acid (DNA) at a CpG site (i.e., a DNA region where a cytosine nucleotide is positioned next to a guanine nucleotide separated by one phosphate), which causes the inhibition of gene expression. Since the 1960s, researchers have been aware of the strong association between age and DNA methylation (Koch and Wagner 2011). This finding spurred interest in the possibility of formulating epigenetic (methylomic) measures of aging.

Hannum et al. (2013) and Horvath (2013) developed the first epigenetic indices of aging. Hannum et al. (2013) identified 71 sites, and Horvath (2013) identified 353 sites on the human genome where methylation changed with age. For some sites, methylation increased with age, whereas it decreased with age at others. For both the Hannum et al. (2013) and Horvath (2013) indices, the correlation between age and the weighted sum of methylation scores was roughly .85 to .90. Nearly all of the markers in these indices lay within or near genes with known functions associated with age-related conditions, including Alzheimer’s disease, cancer, tissue degradation, DNA damage, and oxidative stress (Hannum et al. 2013; Horvath 2013).

Importantly, these epigenetic indices can be used as “biological clocks” to assess, at any point during adulthood, the extent to which an individual is experiencing accelerated or decelerated biological aging (Hannum et al. 2013; Horvath 2013). This can be done by calculating the discrepancy between a person’s chronological age and the age predicted using the epigenetic index. The resulting difference indicates in years the extent to which an individual is biologically older or younger than his or her chronological age (i.e., whether the person is experiencing accelerated or decelerated aging).

The epigenetic measures of Horvath (2013) and Hannum et al. (2013) have become widely accepted measures of biological aging. Accelerated aging using these measures has been linked to frailty, some types of cancer, cardiovascular disease, both Alzheimer’s and Parkinson’s disease, and all-cause mortality (Chen et al. 2016; Marioni et al. 2015). And importantly for social scientists, the two indices have been linked to social factors such as economic hardship (Simons et al. 2016), neighborhood disadvantage (Lei et al. 2019), racial discrimination (Brody et al. 2016), and cumulative stress (Zannas et al. 2015). Although these epigenetic clocks demonstrate many advantages, they have also been shown to suffer from various limitations. For example, they are not related to many types of cancer, their relation with cardiovascular disease is weak, and they show inconsistent associations with health risk behaviors such as smoking and biomarkers such as elevated blood sugar, hypertension, and lipoproteins (Horvath and Raj 2019).

Recently, however, two new epigenetic measure have been developed that overcome many of these weaknesses The first, called DNAm PhenoAge (Levine et al. 2018), was developed using composite clinical measures of phenotypic age as well as chronological age to capture differences in life span and health span. The index is based on 513 CpG sites related to several known aging pathways, including those linked to proinflammatory and antiviral processes, growth hormone signaling, multicellular growth, regulation of DNA methylation, transcriptional and translational machinery, and DNA damage recognition and repair (Horvath and Raj 2019; Levine et al. 2018). The index is strongly related to age, but also, in contrast to the first-generation measures of Horvath (2013) and Hannum et al. (2013), it is related to biomarkers of cardiometabolic dysregulation and various chronic illnesses as well as to sociodemographic attributes (income and education) and lifestyle risk factors such as diet, smoking, and exercise (Horvath and Raj 2019; Levine et al. 2018).

The most recent epigenetic measure of aging, the GrimAge Index, was developed using methylation at 1,030 sites (Lu et al. 2019). The GrimAge Index was formulated by identifying methylation signatures associated with seven plasma proteins known to be involved in age-related diseases plus pack years of smoking. The index is called GrimAge after the Grim Reaper (Lu et al. 2019). Like its predecessors, the residuals obtained by regressing GrimAge on chronological age indicate the extent to which an individual is experiencing accelerated or decelerated aging. Three recent studies have reported that GrimAge performed better than other biological clocks, including DNAm PhenoAge, in terms of predicting various states of morbidity and mortality (Hillary et al. 2020; Li et al. 2020; McCrory et al. 2020). GrimAge was a robust predictor, for example, of shorter life span, coronary heart disease, diabetes, frailty, visceral adiposity, fatty liver, and cognitive decline. Indeed, analysis found that those in the top 5% of accelerated GrimAge had a mortality risk over 4.5 times greater than those in the bottom 5% (Lu et al. 2019). Although these findings are impressive, the world of epigenetic clocks is a rapidly changing one. There is every reason to believe that ever more powerful biological clocks will continue to be developed.

A limitation of the GrimAge clock is that most of the validation studies have been on Europeans and non-Hispanic white Americans. An exception, however, is a study by Lu et al. (2019) in which GrimAge predicted mortality and various physiological pathological states across a variety of samples, including 1,700 black individuals in the Jackson Heart Study. And Roshandel et al. (2020) found that GrimAge predicted neuropathy among people with diabetes in a sample of several hundred African Americans. Although these studies are encouraging, there is clearly a need for more validation studies with African Americans and other persons of color.

Social Determinants of Epigenetic Aging

Evidence suggests that roughly 40% of variation between individuals in the “ticking” of their epigenetic clocks is genetically determined (Bell et al. 2019; Ryan 2020). Consequently, trajectories of epigenetic aging tend to emerge early in life and are somewhat stable across the life course (Li et al. 2020). Still, this leaves roughly 60% of the variance in speed of biological aging to be explained by developmental, behavioral, and environmental factors. In keeping with this idea, studies have linked lifestyle factors, such as poor diet and lack of exercise (Levine et al. 2018; Quach et al. 2019), and developmental events, such as puberty, pregnancy, and menopause, to accelerated aging (Ryan 2020).

Earlier it was argued that epigenetic clocks might be viewed as indicators of physiological weathering. If this is the case, chronic exposure to adverse social conditions should also predict accelerated epigenetic aging. Consonant with this view, recent research using African American samples has reported that chronic economic hardship (Simons et al. 2016), long-term residence in a disadvantaged neighborhood (Lei et al. 2019), persistent exposure to racial discrimination (Brody et al. 2016), and cumulative stressful events (Zannas et al. 2015) are related to accelerated aging. All of these studies used first-generation measures of accelerated aging (i.e., the Horvath [2013] and Hannum et al. [2013] indices). Given how recently GrimAge was developed, there has been only limited investigation of the extent to which it is associated with social circumstances. However, using a sample of roughly 500 African Americans, Simons et al. (2020) recently reported that persistent financial problems, exposure to discrimination, low education, and residing in a disadvantaged neighborhood each predicted accelerated aging and that the sum of these chronic conditions exerted a robust effect. Similarly, using a large sample of older African Americans, Zhao et al. (2019) reported that GrimAge was more strongly related to education, alcohol consumption, and BMI than were the other epigenetic clocks.

None of these various studies, regardless of the clock that was utilized, found a link between acute stressors and speed of aging. Rather, consonant with the weathering hypothesis, it appears to be exposure to chronic, persistent adversity that accelerates biological aging. Although such findings suggest that social conditions may impact speed of aging, the evidence would be more compelling if it could be shown that changes in social conditions are associated with changes in the speed of aging. Such a finding would also be important for policymakers.

Resetting the Clock: Predicting Changes in Accelerated Aging

Given the challenge of obtaining repeated blood draws from a panel of individuals over an extended period of time, there has been little consideration of the extent to which an individual’s speed of epigenetic aging might change across the life course. The few studies that have investigated this issue report stabilities ranging from .22 to .82 (Li et al. 2020; Marioni et al. 2019), with the stabilities being especially strong for the second-generation clocks (Li et al. 2020). Using a sample of middle-aged African American women, the present study focuses on change in GrimAge over 11 years. The stability coefficient for this time period was .76. Although this is a rather high stability, it indicates that change in GrimAge still occurs for some individuals in late middle age. Indeed, GrimAge assessed when the respondents were in their late 40s explained only a little over half of the variation in GrimAge when they were in their late 50s.

The finding that some individuals continue to demonstrate alterations in their speed of aging well into late middle age raises questions regarding the determinants of such variation. One possibility is that changes in their social environment foster changes in speed of epigenetic aging. The present study uses longitudinal data from a sample of middle-aged, African American women to investigate this idea.

Detecting an association between changing social circumstances and changes in speed of aging presents a significant challenge. As noted, it is exposure to chronic or persistent conditions such as economic hardship, low education, interpersonal discrimination, and residing in a disadvantaged neighborhood (i.e., cumulative adversity) that predicts speed of aging (Lei et al. 2019; Simons et al. 2020; Zannas et al. 2015), with acute events having little effect. Chronic conditions, by definition, show minimal change over time. Hence, they are of limited utility when the goal is to examine the association between changing social conditions and shifts in speed of aging. It is the case, however, that many individuals, especially those of color, experience considerable change, for better or worse, in financial resources and their living situation during their 50s and 60s (Carr 2019). These changes, along with the rash of adverse circumstances that they often produce, provide a window for investigating the potential link between alterations in social conditions and fluctuations in speed of aging. The present article presents analyses investigating this possibility.

The Present Study

The current study investigates speed of aging using 223 aging African American women participating in the Family and Community Health Study (FACHS). Such a sample is fitting given the vast literature showing that African Americans suffer a greater prevalence and earlier onset of chronic illness and disability than other ethnic groups (Lim et al. 2018). Furthermore, the most prominent race differences in health emerge in middle age, and the mortality rates for middle-aged black women have actually worsened since 1990 (Geronimus et al. 2010). FACHS respondents provided blood draws in 2008 (Wave 5) and again in 2019 (Wave 8). This enabled us to obtain assessments of GrimAge at two time points separated by 11 years. Our goal was to identify shifting social circumstances that might predict changes in GrimAge over this 11-year period.

There is rather strong evidence that cohorts show increases in income inequality over the life course (Crystal, Shea, and Reyes 2017; Deaton and Paxon 1994). Such findings are consistent with the assertion by cumulative advantage/disadvantage (CAD) theory that advantage increases exposure to opportunity while disadvantage increases exposure to risk and adversity (Dannefer 2003; Ferraro and Shippee 2009). The result is cumulative inequality within a cohort as income and wealth continue to fan out as individuals age (Crystal, Shea, and Reyes 2017; Deaton and Paxon 1994). Childhood poverty, for example, influences the probability of living in a disadvantaged neighborhood, attending a poor school, and early termination of education, which in turn leads to a low-paying job, little or no accumulation of wealth, and lack of retirement income during old age (Carr 2019). This chain is amplified for low-income black Americans who also face the challenge of institutionalized racism at every step in this process. These arguments suggest that income inequality will accrue during the 11-year period covered by the current study, with some individuals increasing and others decreasing their economic resources. Research has established that in general, Americans view financial insecurity as the most stressful of stressors (American Psychological Association 2015), and such circumstances are likely even more upsetting for aging individuals striving to establish the overall meaning and significance of their lives. Based on this idea, we expect that the changes in income experienced by our late-middle-aged respondents will be associated with acceleration of speed of biological aging.

Furthermore, we expect that much of the impact of alterations in income on speed of aging will be explained by the cascade of stressful circumstances and the unraveling of everyday life often associated with such financial changes. Financially strapped aging individuals must often accept monetary assistance from their adult children. In addition, fiscal problems combined with institutional racism regarding housing (redlining, foreclosures, rental discrimination) may require sharing a residence with an adult child. Based on these realities, we predict that over the 11-year period covered by the present study, changes regarding housing, receipt of financial assistance from family members, and coresidence with family members will predict increases in speed of biological aging. Furthermore, we expect that these fluctuating circumstances will account for much of the effect of alterations in income on change in speed of aging. Specifically, we make the following predictions:

Hypothesis 1: Based on CAD, we expect that income will show a fanning effect and that decreased household income will be associated with an increase in speed of biological aging.

Hypothesis 2: Financial difficulties often require turning to family members for assistance. Given the stress often associated with receiving such aid, we expect that increases in financial assistance from adult children will be associated with an increase in speed of biological aging.

Hypothesis 3: It is well established that housing discrimination, redlining, foreclosure, gentrification, and so on make it difficult for low-income African Americans to find adequate housing (Massey 2007). This is especially true for elderly individuals. We expect that increased difficulties relating to housing (denied rental applications or home loans) will be associated with an increase in speed of biological aging.

Hypothesis 4: Financial problems and difficulties finding housing often necessitate sharing residence with a family member such as an adult child. There is necessarily a certain amount of stress involved with the transition from living independently to sharing a household with others. This is especially the case when the arrangement assumes that the grandparent will provide day care for grandchildren. Although grandparents are often happy to be involved in the lives of their grandchildren, day care is often a physically and psychologically taxing experience for aging individuals, many of whom suffer from declining health (Carr 2019). Given these realities, we predict that beginning to share a residence with an adult child or children will be associated with an increase in speed of biological aging.

Hypothesis 5: Finally, it is unlikely that any one of these changes in isolation exerts a major effect on biological aging. Rather, these stressful events tend to occur together, and it is the cumulative effect that is likely most consequential. We test this idea by summing changes in these four circumstances to form a cumulative index of social disadvantage. We expect that high scores on this index of adverse changes will display a rather robust association with increases in speed of biological aging.

In testing these hypotheses, it was important that physical disability be included as a control variable. Given the documented impact of epigenetic aging on health, a reduction in physical functioning is likely related to accelerated GrimAge. And a relationship would also be expected between diminished physical functioning and our predictor variables: reduction in income, taking money from adult children, having housing difficulties, and moving in with adult children. Thus, any association found between changes in the predictors and change in GrimAge might be, at least in part, spurious due to the association of both with change in physical functioning. We take this possibility into account by including change in physical disability in our models.

Data and Methods

Participants and Procedures

We tested our hypotheses using data collected at Wave 5 (2008) and Wave 8 (2019) from the primary caregivers in FACHS. FACHS is an ongoing study of several hundred African American families that was initiated in 1997. All of the families had a fifth grader at study inception. Using a stratified random sampling procedure, the sampling strategy was intentionally designed to generate families representing a range of socioeconomic statuses and neighborhood settings. Details regarding recruitment were described by Gibbons and colleagues (2004) and Simons and colleagues (2011). The protocol and all study procedures were approved by the Institutional Review Board at the University of Georgia (Title: FACHS IV; Protocol No. Study00000172). Computer-assisted interviews were conducted at each wave and took on average about 2 hours to complete.

At Wave 1, about half of the sample resided in Georgia (n = 422) and the other half in Iowa (n = 467). The vast majority of primary caregivers were women. Given this fact, only women were included in the present study. Mean caregiver age at Wave 5 was 48.5 years (SD = 9.23), 17.8% had less than a 12th-grade education, and 24.8% were married. The majority (68.5%) lived in large urban areas, 12.2% lived in the suburbs, and 19.3% lived in rural areas. Average per capita income was $11,069. Eleven years later, at Wave 8, average per capita income had risen to $18,001, whereas the other demographics remained roughly similar to Wave 5.

Within two weeks of the Wave 5 and Wave 8 psychosocial interviews, a certified phlebotomist visited the home and collected four tubes of blood (30 ml) from each consenting participant. Given the logistics of scheduling home visits by phlebotomists, only members of the sample still residing in Georgia or Iowa at Waves 5 and 8 were identified as eligible for the blood draws. Blood was obtained from 72% (n = 366) of the women at Wave 5 who met this criteria. Of these individuals, 239 were still living in the study area and agreed to provide blood at Wave 8. Unfortunately, 16 of these individuals had missing data and had to be dropped from the analysis. This left 223 women who served as the sample for the present study.

Comparisons of these individuals with those who did not provide blood did not reveal any significant differences with regard to either demographic characteristics or the independent variables (e.g., income: t = 1.042, ns; financial pressure: t = 1.231, ns; chronological age: t = 1.133, ns) at the initial wave of the FACHS study. The only exception was diet. Respondents who gave blood had a slightly poorer diet than those who did not (about 3/20 of a standard deviation, p < .05).

The tubes of blood (30 ml) drawn from each participant were shipped on the same day to a laboratory for preparation. Upon arrival, the tubes were inspected to ensure anticoagulation and aliquots of blood were diluted 1:1 with phosphate buffered saline (pH 8.0). Mononuclear cell pellets were separated from the diluted blood specimen using a centrifuge with ficoll (400 g, 30 minutes). The mononuclear cell layer was removed from the tube using a transfer pipette, resuspended in a phosphate buffered saline solution, and briefly centrifuged again. The resulting cell pellet was resuspended in a 10% DMSO/RPMI solution and frozen at 8.0° C until use.

The Illumina EPIC 850 BeadChip was used to assay genome-wide DNA methylation. Participants were randomly assigned to 16 sample “slides/chips” with groups of eight slides being bisulfite converted in a single plate, resulting in two “batches/plates.” A replicated sample of DNA was included in each plate to aid in assessment of batch variation and to ensure correct handling of specimens. Prior to normalization, methylation data were filtered based on these criteria: (a) samples containing 1% of CpG sites with detection p < .05 were removed, (b) sites were removed if a bead count of less than three was present in 5% of samples, and (c) sites with a detection p < .05 in 1% of samples were removed. The methylation data were then used to calculate a GrimAge score for each participant through the public online DNA Methylation Age Calculator (https://dnamage.genetics.ucla.edu/).

Measures

Accelerated epigenetic aging

At Waves 5 (2008) and 8 (2019), epigenetic aging was assessed using the GrimAge Index (Lu et al. 2019). This index estimated an individual’s biological age based on methylation assessments at 1,030 sites scattered across the human genome. In the study sample, the correlation between GrimAge and chronological age was .751 at Wave 5 and .769 at Wave 8. To transform GrimAge into an accelerated aging score, we regressed GrimAge on chronological age. The resulting values had a zero correlation with chronological age. A positive value on this variable indicated in years accelerated epigenetic aging, whereas a negative value indicated in years decelerated aging. As expected, accelerated GrimAge was significantly correlated with biomarkers of health such as C-reactive protein, hemoglobin A1C, various inflammatory cytokines, and chronic illness.

Family per capita income

Respondents reported their annual household income from all sources (e.g., wages, interest, business profit, etc.) at Waves 5 and 8. This variable was measured as an ordinal variable with 16 categories, ranging from 0 (<$10,000) to 15 (≥$200,000); total household income was based on the midpoint of these categories. Family per capita income was calculated by dividing the total household income by the number of family members.

Living with adult children

At Waves 5 and 8, respondents were asked, “During the past 12 months, did your children live with you?” Response format for this item ranged from 0 (“never”) to 4 (“all year”). Those whose children were deceased were coded as 0.

Financial support from children

The item asked respondents to indicate whether their children contribute financially to the support of parents. Response format for this item ranged from 0 (“no”) to 3 (“yes, regularly”). Those whose children were deceased were coded as 0.

Difficulty in housing

Respondents answered two questions regarding how often (1 = “never,” 4 = “frequently”) they had been refused rental housing and had difficulty getting a home loan because of race or ethnic background. These two items were not available at Wave 5 (2008). Thus, we utilized data from Wave 6 (2010) and Wave 8 (2019).

Education

At Wave 8, respondents reported on their highest level of education, with the range from 1 = “first grade” to 20 = “doctoral degree.”

Health promotion behaviors

We used a six-item scale developed by the authors for this study. Respondents’ involvement in exercise over the previous week was assessed using the item, “How often do you exercise?” Respondents were asked to report on their regular eating behavior, including, “How often do you watch what you eat (i.e., nutrition—such as eating fruits and vegetable regularly)?” “How often do you diet to lose weight?” “How often do you eat fatty foods like potato chips or ice cream?” and “How often do you eat fast food?” We reverse-coded the last two items. Finally, respondents were asked, “How often do you have checkups with a doctor?” Responses for these items ranged from 1 (“never”) to 4 (“regularly”). We created a measure of health promotion behaviors by averaging these six items.

Physical disabilities

At Waves 5 and 8, respondents were asked whether their health limited their ability to engage in activities (1 = “not limited at all,” 3 = “limited a lot”), including (a) moderate activities, such as moving a table, pushing a vacuum cleaner, bowling, or playing golf, and (b) climbing several flights of stairs. We generated a measure of physical disability by averaging these two items.

Statistical Analysis

We used STATA 16 for all analyses. Given that we only had two waves of accelerated Grim aging, change scores (Δ) for study variables were calculated using the residuals from the regression of Wave 8 scores on Wave 5 scores. A residual change score of 0 indicated no change between Waves 5 and 8. A positive value for the residual change score suggested an increase over time, whereas negative values indicated a decrease over time. Hierarchical regression models with robust standard errors were used to examine our hypotheses. Independent and dependent variables were entered into the models in the following steps. Model 1 tested the impact of Δ in family per capita income on Δ in accelerated aging after controlling for education and health promotion behaviors. Model 2 added the three stressful transitions (living with children, financial support from children, and difficulty in housing) to the regression to evaluate the extent to which Δ in each of them is associated with Δ in accelerated aging. Finally, Model 3 examined the impact of Δ in cumulative disadvantage—income (reversed-coded) and the three stressful transition variables—on Δ in accelerated aging.

Having established the impact of change in cumulative adversity on change in aging, we reran this regression controlling for cell type. This allowed us to determine the extent to which the association of Δ cumulative disadvantage with Δ GrimAge was due to the methylation sites used to assess the latter being confounded with age-related changes in the distribution of various immune cell types. Furthermore, to establish the stability of our findings, we reran our regression using fixed effect analysis, thereby controlling for unmeasured time-invariant variables. Lastly, to better interpret the association between Δ cumulative disadvantage and Δ GrimAge, we graphed the estimated value of accelerated aging as a function of cumulative social disadvantage.

Results

Table 1 provides the descriptive statistics for the study variables. To begin, the top of the table shows that over an 11-year period, some study participants demonstrated an increase in speed of biological aging while others showed a decrease in speed of aging. More specifically, 73 individuals decelerated their speed of aging by at least a year, whereas 68 individuals accelerated their speed of aging by at least a year.

Table 1.

Descriptive Statistics for Study Variables with Data from Family and Community Health Study (FACHS), Waves 2008–2019.

	Full Sample(N = 223)	More Than 1 Year Younger (n = 73)	More Than 1 Year Older (n = 68)
Variables	% or Mean (SD)	% or Mean (SD)	% or Mean (SD)
Per capita income (2008)	11,069.80 (10,384.99)	11,784.70 (10,510.33)	11,135.15 (10,791.39)
Per capita income (2019)	18,001.04 (16,183.00)	21,517.37 (17,878.54)	15,052.17 (14,736.43)
Living with children (2008)	2.34 (1.58)	2.49 (1.56)	2.19 (1.64)
Never	22.42%	19.18%	26.47%
Rarely	7.17%	6.85%	8.82%
Some	22.87%	21.92%	20.59%
Often	9.42%	9.59%	7.35%
All year	38.12%	42.47%	36.76%
Living with children (2019)	.98 (1.52)	.59 (1.23)	1.54 (1.76)
Never	66.82%	78.08%	51.47%
Rarely	2.24%	2.74%	4.41%
Some	11.66%	9.59%	8.82%
Often	4.48%	1.37%	8.82%
All year	14.80%	8.22%	26.47%
Financial support from children (2008)	.46 (.71)	.55 (.75)	.47 (.74)
No	66.37%	60.27%	67.65%
Occasionally	21.08%	24.66%	17.65%
Regularly	12.56%	15.07%	14.71%
Financial support from children (2019)	.38 (.66)	.23 (.51)	.65 (.77)
No	71.75%	80.82%	52.94%
Occasionally	18.39%	15.07%	29.41%
Regularly	9.87%	4.11%	17.65%
Difficulty in housing (2010)	1.25 (.53)	1.16 (.44)	1.30 (.57)
Difficulty in housing (2019)	1.31 (.54)	1.24 (.50)	1.46 (.66)
Cumulative disadvantage (2010)	.00 (1.00)	–.01 (.98)	.01 (1.10)
Cumulative disadvantage (2019)	.00 (1.00)	–.35 (.87)	.51 (1.10)
Education (2019)	12.94 (2.15)	13.36 (2.14)	12.60 (2.34)
< 12th Grade	15.25%	10.96%	22.06%
High school graduate or GED	38.57%	38.36%	38.24%
Some college or associate’s degree	31.84%	28.77%	27.94%
Bachelor’s degree	4.04%	8.22%	2.94%
Postgraduate	10.31%	13.70%	8.82%
Health promotion behaviors (2008)	15.09 (3.13)	14.74 (3.01)	15.10 (3.23)
Health promotion behaviors (2019)	15.95 (2.97)	15.96 (2.94)	15.82 (2.94)
Physical disabilities (2008)	3.88 (.48)	3.89 (.43)	3.78 (.51)
Physical disabilities (2019)	3.78 (.63)	3.70 (.54)	3.91 (.59)

For each of the study predictor variables, Table 1 presents the mean and standard deviation for the full sample, for those whose speed of aging decelerated by at least a year, and for those whose speed of aging accelerated by at least a year. Overall, these variables show a cohort fanning effect. That is, in most cases, the means for the accelerated versus decelerated groups are more similar in 2008 than in 2019. Per capita income, for example, shows a difference between the two groups of $649 in 2008, but the gap widened to $6,465 by 2019. Similarly, the percentage of respondents receiving regular financial support from children was 15.07% compared to 14.71% in 2008 but expanded to 4.11% versus 17.65% by 2019. And although the mean for cumulative disadvantage was virtually the same for the two groups in 2008/2010, a difference in means of –.35 versus .51 emerged by 2019.

The 2008 percentages for living with adult children requires some explanation. In 2008, the children in the study were roughly 21 to 22 years of age, and hence many of them (38%) still lived at home. In subsequent years, however, most all of them moved out and began living independently. Albeit, by 2019, when they were roughly 32 to 33 years of age, many were again sharing residence with their parents. As expected, the proportion sharing a residence was much higher for those who experienced an acceleration in speed of aging compared to those whose speed of aging decelerated. As shown in Table 1, the difference is 8.22% versus 26.47%, respectively. The takeaway message from Table 1 would seem to be that individuals who have increased their speed of aging over the 11-year period from 2008 to 2019 were more likely to report having experienced greater increases in various adverse circumstances than those who showed a decrease in their speed of aging.

Table 2 presents the bivariate correlations for changes in the study variables. The table shows that changes in each of the four predictor variables (income, financial assistance, sharing a residence, and housing difficulties) are significantly related to Δ in GrimAge. The correlations are particularly strong for financial assistance from children (r = .296, p ≤ .01) and living with children (r = .314, p ≤ .01). There is also a robust correlation of .389 between Δ cumulative disadvantage and Δ GrimAge. Change in physical disability is correlated with Δ GrimAge (r = .136, p ≤ .05), but it is not significantly related to any of the other study variables. This suggests that changes in our predictor variables were not in response to parents’ increased physical limitations. Finally, the bivariate correlation between change in health risk behaviors and Δ in GrimAge fails to achieve significance.

Table 2.

Correlations, Means, and Standard Deviations among Study Variables with Data from Family and Community Health Study (FACHS), Waves 2008–2019.

	1	2	3	4	5	6	7	8	9
1. Δ Accelerated GrimAge	—
2. Δ Per capita income	−.160*	—
3. Δ Living with children	.314**	−.138*	—
4. Δ Financial support from children	.296**	−.159*	.536**	—
5. Δ Difficulty in housing	.162*	−.064	−.006	−.016	—
6. Δ Cumulative disadvantage	.389**	−.568**	.696**	.700**	.435**	—
7. Education (2019)	−.014^†	.223**	−.070	−.037	−.197**	−.219**	—
8. Δ Health promotion behaviors	−.075	.088	.065	.060	−.055	−.007	.095	—
9. Δ Physical disabilities	.136*	−.053	.069	.007	.011	.058	.063	.088	—
Mean	.000	.000	.000	.000	.000	.000	12.937	.000	.000
SD	2.266	14023	1.497	.655	.469	1.000	2.149	2.619	.630

Note: N = 223; Δ = change in a variable from 2008 to 2019.

†

p ≤ .10, *p ≤ .05, **p ≤ .01 (two-tailed tests).

As an aside, it is interesting to note that the magnitude of the cross-sectional correlations at Wave 8 between GrimAge and the adversity variables was very comparable to those reported in Table 2 for the change variables. For example, the correlations of GrimAge with per capita income, living with adult children, receiving financial support, housing difficulties, and cumulative disadvantage were –.213, .278, .318, .101, and .380, respectively. The biggest contrast is for per capita income, where the change score correlation is –.160, whereas the cross-sectional correlation is –.213.

Table 3 shows the results of hierarchical regression to examine the unique effects of the various adverse events on biological aging. Model 1 shows that Δ in per capita income continues to have a significant effect on Δ in GrimAge (β = –.322, p ≤ .05) after controlling for another component of socioeconomic status—education. Model 2 adds changes in various adverse circumstances to the regression model. Consonant with the study hypotheses, Δ in living with adult children, Δ in financial support from adult children, and Δ in housing difficulties are all significant predictors of Δ in GrimAge. Furthermore, consistent with the mediation argument, the effect of Δ in per capita income is no longer significant after introducing these adverse conditions into the model. Finally, Model 3 adds Δ physical disability and Δ health behaviors to the regression. The effect of Δ physical disability is significant, whereas the effect of Δ health behaviors is not. Importantly, however, Δ in financial support from children, Δ living with adult children, and Δ in housing difficulties all continue to show significant associations with Δ GrimAge after the effects of physical disability and health behaviors are taken into account.

Table 3.

Regression Models Examining the Effects of Social Factors and Health Behaviors on Change in Accelerated GrimAge between 2008 and 2019 with Data from Family and Community Health Study (FACHS), Waves 2008–2019.

	Δ Accelerated GrimAge
	Model 1		Model 2		Model 3
	b (SE)	β	b (SE)	β	b (SE)	β
Δ Per capita income	−.322* (.155)	−.142	−.192(.150)	−.085	−.155(.152)	−.069
Δ Living with adult children			.313* (.120)	.207	.304* (.120)	.201
Δ Financial support from children			.598* (.275)	.173	.634* (.271)	.183
Δ Housing difficulties			.732* (.329)	.151	.703* (.316)	.145
Education (2019)	−.087(.068)	−.082	−.047(.069)	−.044	−.051(.068)	−.049
Δ Health promotion behaviors					−.079(.050)	−.091
Δ Physical disabilities					.457* (.204)	.127
Constant	1.122(.905)		.605(.912)		.662(.905)
R ²	.032		.159		.182

Note: Unstandardized (b) shown with robust standard errors in parentheses. Δ = change in a variable from 2008 to 2019. Per capita income and cumulative disadvantage are standardized by z transformation. N = 223.

p ≤ .05 (two-tailed tests).

As noted previously, we expected that the cumulative effect of these adverse changes would be much more consequential for biological aging than any of the adverse conditions by themselves. Model 1 in Table 4 shows that this is the case. Net the controls for education, health promotion behaviors, and disability, Δ cumulative disadvantage shows what might be considered a rather robust association with Δ GrimAge (β = .374, p ≤ .01). Next, we introduced controls for cell type (Bell et al. 2019). This allowed us to determine the extent to which the association of Δ cumulative disadvantage with Δ GrimAge was due to the methylation sites used to assess the latter being confounded with age-related changes in the distribution of various immune cell types. Model 2 shows that most of the effect of Δ cumulative disadvantage remains after controlling for distributions of cell types. This suggests that cumulative adversity is related to biological changes beyond age-related shifts in the immune system. Finally, we reran our regression using fixed effect analysis. This allowed us to control for various unmeasured time-invariant variables. As shown in Model 3, the effect of Δ cumulative disadvantage on Δ GrimAge remains significant (p ≤ .01) in this analysis. This finding provides further, although certainly not definitive, support for a causal relationship between these variables.

Table 4.

Regression Models Examining the Effects of Cumulative Disadvantage, Health Behaviors, and Cell-Type Variations on Change in Accelerated GrimAge between 2008 and 2019 with Data from Family and Community Health Study (FACHS), Waves 2008–2019.

	Δ Accelerated GrimAge
	Model 1		Model 2		Fixed Effect Model
	b (SE)	β	b (SE)	β		b (SE)
Δ Cumulative disadvantage	.847** (.150)	.374	.671** (.134)	.296	Cumulative disadvantage	.321** (.115)
Education (2019)	−.034(.069)	−.032	−.065(.060)	−.062	Time	−.367(.236)
Δ HPB	−.069(.051)	−.080	−.021(.046)	−.024	HPB	−.033(.044)
Δ Physical disabilities	.445* (.200)	.124	.311(.200)	.086	Physical disabilities	.141(.183)
Δ CD8+ T cells			−16.877** (2.953)	−.286	CD8+ T cells	−16.735** (3.539)
Δ CD4+ T cells			−9.184** (3.000)	−.212	CD4+ T cells	−10.679** (2.829)
Δ Natural killer cells			2.260(5.117)	.021	Natural killer cells	−4.594(5.211)
Δ B cells			−7.265(5.227)	−.094	B cells	−9.200†(4.603)
Δ Monocytes			5.039(8.683)	.054	Monocytes	2.775(6.510)
Constant	.437(.908)		.846(.802)		Constant	3.389* (1.423)
R ²	.172		.348		Rho	.859

Note: Unstandardized (b) shown with robust standard errors in parentheses. Δ = change in a variable from 2008 to 2019. HPB = Health promotion behaviors. Per capita income and cumulative disadvantage are standardized by z transformation. Education is omitted from the fixed effect model because it is subsumed among the unmeasured time-invariant variables controlled by this procedure. N = 223.

p ≤ .05, **p ≤ .01 (two-tailed tests).

Figure 1 presents a graph depicting the association between Δ cumulative index and Δ GrimAge. The distribution of Δ cumulative disadvantage is presented below the graph and indicates that the distribution of changes in GrimAge are roughly normally distributed. The graph shows that a two standard deviation increase in cumulative disadvantage is associated with roughly a one year and nine month acceleration of biological aging, whereas a two standard deviation decrease in cumulative disadvantage is associated with roughly a one year and nine month deceleration in biological aging. The latter finding is important because it suggests that the combination of increased income, avoidance of housing problems, and maintenance of financial and residential independence during late middle age may serve to decelerate biological aging.

Figure 1.

Graph Depicting the Association between Change in the Index of Cumulative Disadvantage and Change in Accelerated GrimAge with Data from Family and Community Health Study (FACHS), Waves 2008–2019.

Finally, we performed sensitivity analyses by examining the association of our changes in disadvantage with two other recently developed clocks. The first was PhenoAge (Levine et al. 2018), a second-generation clock described earlier. In addition, we used DunedinPoAm (Belsky et al. 2020), a recently published clock that has shown early promise as a measure of pace of biological aging. The results, which are presented in the Supplemental Material available in the online version of the article, replicate many of the findings reported in Tables 3 and 4 for GrimAge: Δ living with children and Δ financial support from children are associated with Δ PhenoAge, whereas Δ in financial support from children and Δ in housing difficulties are related to Δ DunedinPoAm. Furthermore, Δ cumulative disadvantage continues to predict changes in both PhenoAge and DunedinPoAm after introducing controls for Δ health promotion and Δ physical disabilities. These analyses provide additional support for the view that changes in chronic stressors during late middle age tend to affect speed of biological aging.

Discussion

For decades, sociologists have used self-ratings to assess health status. Importantly, however, such measures tell us nothing about the proximal, intervening mechanisms whereby social factors become biologically embedded and influence morbidity and mortality. Self-ratings of health are of limited value when the issue is how social experiences get under the skin. Recognizing that this is the case, sociologists have begun to incorporate an array of biomarkers into their research on health and illness. The result has been that sociology now has a seat at the table with biomedical researchers who can no longer dismiss findings regarding the impact of social conditions on health.

Albeit, staying abreast of recent advances in biomarker research represents a significant challenge to medical sociologists who are often not well trained in human biology and lack access to systematized procedures for learning the ins and outs of such measures. The present article was an attempt to demonstrate the usefulness of one relatively new and quite popular set of biomarkers: the epigenetic clocks. These measures of speed of biological aging are noteworthy because they have been shown to be robust predictors of onset of chronic illness and time to death (Hillary et al. 2020; Levine et al. 2018; Li et al. 2020; Lu et al. 2019). Indeed, these indices might be considered indicators of biological “weathering” as discussed by Geronimus and colleagues (2006, 2010). They define weathering as the physiological deterioration experienced by black Americans in response to cumulative social and economic adversity that gives rise to early illness and death. Epigenetic clocks might be viewed as robust markers of such weathering (Simons et al. 2016, 2020).

Importantly, there is now good evidence that speed of aging, as assessed by these clocks, is associated with various adverse social conditions (Brody et al. 2016; Lei et al. 2019; Simons et al. 2016, 2020). To this point, however, research on the social determinants of epigenetic aging, regardless of the index used, has been cross-sectional. Support for a causal relation would be more compelling if it could be shown that changes in social conditions are associated with changes in speed of aging. Such research represents a considerable challenge, however, because it is chronic conditions that have been linked to speed of aging, and chronic conditions, by definition, show minimal change over time. To circumvent this problem, the present study focused on life changes faced by many black Americans later in life. Findings indicated that changes in income, aid from adult children, sharing residence with an adult child, and difficulties relating to housing predicted change in GrimAge. A cumulative index formed by summing these various adverse events was an especially robust predictor of changes in GrimAge (β = .374) because a two standard deviation increase in this measure was associated with roughly a one year and nine month acceleration in in speed of aging. These findings remained stable following controls for cell type and reanalysis using fixed effect modeling. Furthermore, sensitivity analyses using two other recently developed biological clocks produced a roughly similar pattern of results. This might be considered the strongest evidence yet of a causal relationship between social factors and speed of biological aging.

These findings might be seen as adding to our understanding of the mechanisms whereby socioeconomic status is a fundamental cause of illness (Phelan, Link, and Tehranifar 2010). In addition to avenues such as access to health care and a healthy lifestyle, the fundamental cause argument posits that low socioeconomic status individuals experience the health consequences of greater exposure to stressful, adverse conditions. This assertion raises questions regarding the biological pathways whereby class-related stressors get “under the skin” or “biologically embedded” and foster illness and ultimately death. The present study suggests that at least in part, adverse conditions associated with low socioeconomic status increase risk for early morbidity and mortality because they are associated with accelerated biological aging. Past research reports that disparities in income and wealth tend to expand over the life course (Crystal, Shea, and Reyes 2017; Deaton and Paxon 1994), and this fanning effect was evident in the present study. Prior studies also indicate that health disparities tend to accrue or fan out over the life course (Pavalko and Caputo 2013; Willson, Shuey, and Elder 2007), although selective mortality may mute this effect among the very old. Findings from the present study suggest that these two phenomena may be linked given that the cascade of stressful events fostered by life course changes in financial circumstances may augment disparities in biological aging and associated morbidities.

Like all research, the present study suffered from various limitations. To begin, the sample consisted of only 223 respondents. There is clearly a need to replicate the findings with a larger sample. Second, the study participants all resided in either the Southeastern or Midwestern regions of the United States. Although there is little reason to believe that the associations found between changes in adverse conditions and changes in speed of aging are specific to these regions, it is important that our findings be replicated with a more nationally representative sample. Finally, the study sample consisted only of middle-age, African American women. On the one hand, this was appropriate based on the heavy burdens placed on this group (Carr 2019). Given the high unemployment and incarceration rates experienced by black men, economic survival of the family must often be shouldered by black women who must engage in sustained high-effort coping (Geronimus et al. 2006). As a result, they are at high risk for biological wear and tear or weathering (Geronimus et al. 2006). Still, although it is important to investigate adversity and biological aging among impoverished black women, our results need to be replicated with other stigmatized and disadvantaged groups including black men, Latinos, Native Americans, and more generally, the working poor.

Another limitation was the omission of variables involving perceived stress and coping resources often posited to be mediators/moderators of the impact of the social environment on biomarkers of health (Pearlin et al. 2005). Such variables were not included in the present study given the complex set of findings reported in past research. One study found, for example, that self-control is associated with slower aging among persons of high class origin but with faster aging among those of lower class origin (Miller et al. 2015). Research with veterans reports that those who evince no posttraumatic stress disorder even though exposed to traumatic events tend to display faster epigenetic aging (Boks et al. 2015). Similarly, in another study, veterans suffering from posttraumatic stress disorder exhibited greater accelerated aging than those without posttraumatic stress disorder, but only if they were also high on self-efficacy (Mehta et al. 2018). Such findings would seem to suggest that stress and adversity accelerate biological aging but that those who seem to be handling it best may be at greater risk for accelerated aging than those who appear distressed (Ryan 2020). Clearly, more research is needed regarding the complicated effect of perceived stress and coping resources on speed of aging. Space limitations, however, prevented consideration of these intricacies in the present article.

In conclusion, recent research indicates that epigenetic clocks are important indicators, if not drivers, of biological aging and that they can serve as surrogates for a variety of biomarkers of pathology as well as time to death (Ryan 2020). This has given rise to a flurry of research regarding pharmaceutical and molecular interventions that might slow or reset these epigenetic clocks. Findings from in vivo animal experiments (Kane and Sinclair 2019), as well as human experiments involving hormone supplementation (Fahy et al. 2019) and calorie restriction (Belsky et al. 2020), suggest that epigenetic aging may, at least to some degree, be slowed and perhaps even reversed. Evidence is mounting, however, that this issue might also be investigated through a sociological lens. As noted, past research has identified a link between social factors and speed of epigenetic aging. Importantly, this effect has not been limited to the impact of adversity on accelerated aging. Rather, it is also the case that positive social conditions have been linked to decelerated aging. Indeed, in the present study 73 out of 223 women reduced their speed of aging by over a year. This deceleration was predicted by sizable increases in income and maintaining financial and household independence. This finding, coupled with the results of prior cross-sectional research, suggests that prevention, and even reversal, of accelerated biological aging among low-status and disadvantaged groups might best be addressed through social policies and programs that target structural and cultural conditions relating to social injustice, racism, and economic inequality (Bailey et al. 2017; Massey 2017).

Supplemental Material

sj-pdf-1-hsb-10.1177_00221465211009309 – Supplemental material for Re(Setting) Epigenetic Clocks: An Important Avenue Whereby Social Conditions Become Biologically Embedded across the Life Course

Supplemental material, sj-pdf-1-hsb-10.1177_00221465211009309 for Re(Setting) Epigenetic Clocks: An Important Avenue Whereby Social Conditions Become Biologically Embedded across the Life Course by Ronald L. Simons, Man-Kit Lei, Eric Klopach, Mark Berg, Yue Zhang and Steven S. R. Beach in Journal of Health and Social Behavior

Footnotes

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institute on Aging (R01 AG055393); the National Heart, Lung, Blood Institute (R01 HL118045); the National Institute on Child Health and Human Development (R01 HD080749); National Institute on Drug Abuse (R21 DA034457), and the Centers for Disease Control (R49 CE003095-01). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Supplemental Material

The Supplemental Material is available in the online version of the article.

ORCID iDs

Ronald L. Simons

Eric Klopach

Mark Berg

Author Biographies

Ronald L. Simons is regents professor in the Department of Sociology and codirector of the Center on the Biological Embedding of Social Events and Relationship at the University of Georgia. He has been funded continuously by NIH for over 25 years. His research utilizes longitudinal data to investigate avenues whereby social events and circumstances such as racial discrimination, economic hardship, neighborhood disorder, and relationship conflict become biologically embedded and influence speed of aging, health, and mortality. He is especially interested in how these processes produce class and race health inequalities.

Man-Kit Lei is assistant professor of sociology and affiliated with the Center for Family Research as the University of Georgia. His research utilizes innovative quantitative methods to examine health-related outcomes such as chronic inflammation and rate of biological aging. Recent articles have investigated the impact, for better or worse, of the social environment on biomarkers of health, especially among minority populations.

Eric Klopack is currently a postdoctoral scholar at the University of Southern California Leonard Davis School of Gerontology. He is investigating the methylation signatures associated with adverse conditions such as low socioeconomic status or chronic discrimination. Furthermore, he is concerned with how later life circumstances may affect, either countering or amplifying, DNAm patterns associated with childhood adversity.

Mark Berg is associate professor of sociology and director of the Crime and Justice Policy Program at the University of Iowa. His research is funded by the Centers for Disease Control and examines interpersonal disputes, antisocial behavior, and the life course determinants of various health outcomes. His work utilizes multilevel and longitudinal quantitative data to examine situational, social ecological, and life course contexts of social behavior and biomarkers of health.

Yue Zhang is a doctoral student in the Department of Sociology at the University of Georgia. Her research utilizes longitudinal models to investigate the extent to which factors such as religiosity, social networks, and romantic partner support serve to buffer the deleterious impact of stressful conditions on biomarkers of health and aging.

Steven R. H. Beach is regents professor of psychology and codirector of both the Center for Family Research and the Center on the Biological Embedding of Social Events and Relationship at the University of Georgia. His scholarship focuses on the way that community, family, and marital process affect mental and physical health. His pioneering research has shown that strengthening close relationships, including marriage and parenting relationships, can play a critical role in reducing depression and physical health problems and can buffer the impact of stress from economic and social factors.

References

American Psychological Association. 2015. “Stress in America Press Room.” http://www.apa.org/news/press/relaeases/stress/index.aspx?tab=10.

Nicole

Johnston

David W

. 2014. “Self-assessed Health: What Does It Mean and What Does It Hide?” Social Science and Medicine 121:21–28.

Bailey

Zinzi A.

Kriger

Nancy

Agenor

Madina

Graves

Jamine

Linos

Natalia

Bassett

Mary T

. 2017. “Structural Racism and Health Inequalities in the USA: Evidence and Interventions.” The Lancet 386:1453–63.

Bell

Christopher G.

Lowe

Robert

Adams

Peter D.

Baccarelli

Andrea A.

Beck

Stephan

Bell

Jordana T.

Christensen

Brock C.

, et al 2019. “DNA Methylation Aging Clocks: Challenges and Recommendations.” Genome Biology 20(1):249. doi:10.1186/s13059-019-1824-y.

Belsky

Daniel W.

Caspi

Avshalom

Arseneault

Louise

Baccarelli

Andrea

Corcoran

David L.

Gao

Hannon

Eiliss

, et al 2020. “Quantification of the Pace of Biological Aging in Humans through Blood Test, the Dunedinpoam DNA Methylation Algorithm.” eLife 9:e54870. doi:10.7554/eLife.54870.

Belsky

Daniel W.

Caspi

Avshalom

Houts

Renate

Cohen

Harvey J.

Corcoran

David L.

Danese

Andrea

Harrington

HonaLee

, et al 2015. “Quantification of Biological Aging in Young Adults.” Proceedings of the National Academy of Sciences 112(30):E4104–10.

Berg

Mark T.

Lei

Man-Kit

Beach

Steven

Simons

Ronald L.

Simons

Leslie G

. 2020. “Childhood Adversities as Determinants of Cardiovascular Disease Risk and Illness Burden in Adulthood.” Journal of Youth and Adolescence 49(6):1292–308.

Boks

Marco P.

Mierlo

Hans C.

Rutten

Bart P. F.

Radstake

Timothy R. D. J.

Witte

Lot De

Geuze

Elbert

Horvath

Steve

, et al 2015. “Longitudinal Changes of Telomere Length and Epigenetic Age Related to Traumatic Stress and Post-traumatic Stress Disorder. Psychoneuroendocrinology 51: 506–512.

Brody

Gene H.

Miller

Gregory E.

Tianyi

Beach

Stephen S. R.

Chen

Edith

. 2016. “Supportive Family Environments Ameliorate the Link between Racial Discrimination and Epigenetic Aging.” Psychological Science 27(4):530–41.

10.

Carr

Deborah

. 2019. Golden Years? Social Inequality in Later-Life. New York, NY: Russell Sage Foundation.

11.

Chen

Brian H.

Marioni

Riccardo E.

Colicino

Elena

Peters

Marjolein J.

Ward-Caviness

Cavin

Tsai

Pei-Chien

Roetker

Nicholas S.

, et al 2016. “DNA Methylation-Based Measures of Biological Age: Meta-analysis Predicting Time to Death.” Aging 8(9):1844–59.

12.

Crystal

Stephen

Shea

Dennis G.

Reyes

Adriana M

. 2017. “Cumulative Advantage, Cumulative Disadvantage, and Evolving Patterns of Late-Life Inequality.” Gerontologist 57(5):910–20.

13.

Dannefer

Dale

. 2003. “Cumulative Advantage/Disadvantage and the Life Course: Cross-fertilizing Age and Social Science Theory.” The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences 58(5):S327–37.

14.

Deaton

Angus

Paxon

Christina

. 1994. “Intertemporal Choice and Inequality.” Journal of Political Economy 102(3):437–67.

15.

Dowd

Jennifer Beam

Todd

Megan

. 2011. “Does Self-reported Health Bias the Measurement of Health Inequalities in U.S. Adults? Evidence Using Anchoring Vignettes from the Health and Retirement Study.” The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences 66(4):478–89.

16.

Fahy

Gregory M.

Brooke

Robert T.

Watson

James P.

Good

Zinaida

Vasanawala

Shreyas S.

Maecker

Holden

Leipold

Michael D.

Lin

David T. S.

Kobor

Michael S.

Horvath

Steve

. 2019. “Reversal of Epigenetic Aging and Immunoscenescent Trends in Humans.” Aging Cell 18(6):e13028. doi:10.1111/acel.13028.

17.

Ferraro

Kenneth

Shippee

Tetyana P

. 2009. “Aging and Cumulative Inequality: How Does Inequality Get under the Skin?” Gerontologist 49(3):333–43.

18.

Geronimus

Arline T

. 2001. “Understanding and Eliminating Racial Inequalities in Women’s Health in the United States: The Role of the Weathering Conceptual Framework.” Journal of the American Medical Women’s Association 56(4):133–36.

19.

Geronimus

Arline T.

Hicken

Margaret

Keene

Danya

Bound

John

. 2006. “Weathering and Age Patterns of Allostatic Load Scores among Blacks and Whites in the United States.” American Journal of Public Health 96(5):826–33.

20.

Geronimus

Arline T.

Hicken

Margaret

Pearson

Jay A.

Seashols

Sarah J.

Brown

Kelly L.

Cruz

Tracey D

. 2010. “Do U.S. Black Women Experience Stress-Related Accelerated Biological Aging?” Human Nature 21(1):19–38.

21.

Gibbons

Frederick X.

Gerrard

Meg

Cleveland

Michael

Wills

Thomas

Brody

Gene H

. 2004. “Perceived Discrimination and Substance Use in African American Parents and their Children: A Panel Study.” Journal of Personality and Social Psychology 86(4):517–29. doi:10.1037/0022-3514.86.4.517.

22.

Hannum

Gregory

Guinney

Justin

Zhao

Ling

Zhang

Hughes

Guy

Sadda

SriniVas

Klotzle

Brandy

, et al 2013. “Genome-Wide Methylation Profiles Reveal Quantitative Views of Human Aging Rates.” Molecular Cell 49(2):359–67.

23.

Harris

Kathleen M.

Schorpp

Kristen M

. 2018. “Integrating Biomarkers in Social Stratification and Health Research.” Annual Review of Sociology 44:361–86.

24.

Wan

Goodkind

Daniel

Kowal

Paul

. 2016. An Aging World. U.S. Census Bureau, Report P95-16-1. Washington, DC: U.S. Census Bureau. https://www.census.gov/content/dam/Census/library/publications/2016/demo/p95-16-1.pdf.

25.

Hillary

Robert

Stevenson

Anna

McCartney

Daniel

Campbell

Archie

Walker

Rosie M.

Howard

David M.

Ritchie

Craig W.

, et al 2020. “Epigenetic Measures of Aging Predict the Prevalence and Incidence of Leading Causes of Death and Disease Burden.” Clinical Epigenetics 12(1). doi:10.1186/s13148-020-00905-6.

26.

Horvath

Steve

. 2013. “DNA Methylation Age of Human Tissues and Cell Types.” Genome Biology 14:R115. https://genomebiology.com//14/10/R115.

27.

Horvath

Steve

Raj

. 2019. “DNA Methylation Based Biomarkers and the Epigenetic Clock Theory of Aging.” Nature Reviews Genetics 19(6):371–84.

28.

Idler

Ellen

Cartwright

Kate

. 2018. “What Do We Rate When We Rate Our Health? Decomposing Age-Related Contributions to Self-rated Health. Journal of Health and Social Behavior 59(1):74–93.

29.

Kane

Alice E.

Sinclair

David A

. 2019. “Epigenetic Changes during Aging and Their Reprogramming Potential.” Critical Reviews in Biochemistry and Molecular Biology 54(1):61–83.

30.

Kennedy

Brian K.

Berger

Shelley L.

Brunet

Anne

Campisi

Judith

Cuervo

Ana Maria

Epel

Elissa S.

Franceschi

Claudio

, et al 2014. “Geroscience: Linking Aging to Chronic Disease.” Cell 159(4):709–13.

31.

Koch

Carmen M.

Wagner

Wolfgang

. 2011. “Epigenetic-Aging-Signature to Determine Age in Different Tissues.” Aging 3(10):1018–27.

32.

Lei

Man-Kit

Simons

Ronald L.

Beach

Steven S. R. H.

Philibert

Robert A

. 2019. “Neighborhood Disadvantage and Biological Aging.” Journals of Gerontology, Series B: Psychological Sciences and Social Sciences 74(7):50–59.

33.

Levine

Morgan E.

Ake

Quach

Austin

Chen

Brian H.

Assimes

Themistocles L.

Bandinelli

Stefania

Hou

Lifang

, et al 2018. “An Epigenetic Biomarker of Aging for Lifespan and Healthspan.” Aging 10(4):573–91.

34.

Xia

Ploner

Alexander

Wang

Yunzhang

Magnusson

Patrik Ke

Reynolds

Chandra

Finkel

Deborah

Pedersen

Nancy L.

Jylhävä

Juulia

Hägg

Sara

. 2020. “Longitudinal Trajectories, Correlations and Mortality Associations of Nine Biological Ages across 20 Years.” eLife 9:e51507. doi:10.7554/eLife.51507.

35.

Lim

Eujung

Gandhi

Krupa

Davis

James

Chen

John J

. 2018. “Prevalence of Chronic Conditions and Multimorbidities in a Geographically Defined Geriatric Population with Diverse Races and Ethnicities.” Journal of Aging and Health 30(3):421–44.

36.

Ake T.

Quach

Austin

Wilson

James G.

Reiner

Alex P.

Aviv

Abraham

Raj

Kenneth

Hou

Lifang

, et al 2019. “DNA Methylation GrimAge Strongly Predicts Lifespan and Healthspan.” Aging 11(2):303–27.

37.

Marmot

Michael

. 2015. The Health Gap: The Challenges of an Unequal World. New York, NY: Bloomsbury.

38.

Massey

Douglas S

. 2007. Categorically Unequal: The American Stratification System. New York, NY: Russell Sage Foundation.

39.

Massey

Douglas S

. 2017. “Why Death Haunts Black Lives.” Proceedings of the National Academy of Science 114:800–802.

40.

Mehta

Divya

Bruenig

Dagmar

Lawford

Bruce

Harvey

Wendy

Carrillo-Roa

Tania

Morris

Charles P.

Jovanovic

Tanja

Young

Ross McD

Binder

Elisabeth B.

Voisey

Joanne

. 2018. “Accelerated DNA Methylaton Aging and Increased Relilience in Veterans: The Biological Cost of Soldiering on.” Neurobiology of Stress 8:112–19.

41.

Marioni

Riccardo E.

Shah

Sonia

McRae

Alan F.

Chen

Brian H.

Colicino

Elena

Harris

Sarah E.

Gibson

Jude

, et al 2015. “DNA Methylation Age of Blood Predicts All-Cause Mortality in Later Life.” Genome Biology 16:25 doi:10.1186/s13059-015-0584-6.

42.

Marioni

Riccardo

Suderman

Matthew

Chen

Brian H.

Horvath

Steve

Bandinelli

Stefania

Morris

Tiffany

Beck

Stephan

, et al 2019. “Tracking the Epigenetic Clock across the Human Life Course: A Meta-analysis of Longitudinal Cohort Data.” The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences 74(1):57–61.

43.

McCrory

Cathal

Fiorito

Giovanni

Hernandez

Belinda

Polidoro

Silvia

O'Halloran

Aisling M.

Hever

Ann

Cheallaigh

Cliona Ni

, et al 2020. “GrimAge Outperforms Other Epigenetic Clocks in the Prediction of Age-Related Clinical Phenotypes and All-Cause Mortality.” The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences. doi:10.1093/gerona/glaa286.

44.

Miller

Gregory

Tianyi

Chen

Edith

Brody

Gene

(2015). “Self-control Forecasts Better Psychosocial Outcomes but Faster Epigentic Aging in Low-SES Youth.” Proceedings of the National Academy of Sciences 12(33):10325–30.

45.

Moffitt

Terrie E

. 2020. “Behavioral and Social Research to Accelerate the Geroscience Translation Agenda.” Ageing Research Reviews 63:101146. doi:10.1016/j.arr.2020.101146.

46.

Osborn

Meg

Widom

Cathy Spatz

. 2019. “Do Documented Records and Retrospective Reports of Childhood Maltreatment Similarly Predict Chronic Inflammation?” Psychological Medicine 50(14):2406–15.

47.

Pavalko

Eliza K.

Caputo

Jennifer

. 2013. “Social Inequality and Health across the Life Course.” American Behavioral Scientist 57(8):1040–56.

48.

Pearlin

Leonard

Schieman

Scott

Fazio

Elena M.

Meersman

Stephen C

. 2005. “Stress, Health, and the Life Course: Some Conceptual Perspectives.” Journal of Health and Social Behavior 46(2):205–19.

49.

Phelan

Jo C.

Link

Bruce G.

Tehranifar

Parisa

. 2010. “Social Conditions as Fundamental Causes of Health Inequalities: Theory, Evidence, and Policy Implications.” Journal of Health and Social Behavior 51(S28):S28–40.

50.

Quach

Austin

Levine

Morgan

Tanaka

Toshiko

Ake T.

Chen

Brian H.

Ferrucci

Luigi

Ritz

Beate

, et al 2019. “Epigenetic Analysis of Diet, Exercise, Education, and Lifestyle Factors.” Aging 9(2):419–46.

51.

Reuben

Aaron

Moffitt

Terrie

Caspi

Avshalom

Belsky

Daniel W.

Harrington

Honalee

Schroeder

Felix

Hogan

Sean

Ramrakha

Sandhya

Poulto

Richie

Danese

Andrea

. 2016. “Lest We Forget: Comparing Retrospective and Prospective Assessments of Adverse Adulthood Experiences in the Prediction of Adult Health.” Journal of Child Psychology and Psychiatry 57(10):1103–12.

52.

Roshandel

Delnaz

Chen

Zhuo

Canty

Angelo J.

Bull

Shelley B.

Natarajan

Rama

Paterson

Andrew D.

DCCT/EDIC Research Group. 2020. “DNA Methylation Age Calculators Reveal Association with Diabetic Neuropathy in Type 1 Diabetes.” Clinical Epigenetics 12(1):52 doi:10.1186/s13148-020-00840-6.

53.

Ryan

Calen P

. 2020. “‘Epigenetic clocks’: Theory and Applications in Human Biology.” Human Biology Toolkit. doi:10.1002/ajhb.23488.

54.

Simons

Ronald L.

Lei

Man-Kit

Beach

Steven S. R. H.

Brody

Gene H.

Philibert

Robert

Gibbons

Frederick

. 2011. “Social Environment, Genes, and Aggression: Evidence Supporting the Differential Susceptibility Perspective.” American Sociological Review 76(6):883–912.

55.

Simons

Ronald L.

Lei

Man-Kit

Beach

Steven S. R.

Philibert

Robert

Cutrona

Carolyn

Gibbons

Frederick

Barr

Ashley

. 2016. “Economic Hardship and Biological Weathering: The Epigenetics of Aging in a U.S. Sample of Black Women.” Social Science and Medicine 150:192–200.

56.

Simons

Ronald L.

Lei

Man-Kit

Klopack

Eric

Beach

Steven R. H.

Gibbons

Frederick X.

Philibert

Robert A

. 2020. “The Effects of Social Adversity, Discrimination, and Health Risk Behaviors on the Accelerated Aging of African Americans: Further Support for the Weathering Hypothesis.” Social Science and Medicine doi:10.1016/j.socscimed.2020.113169.

57.

Willson

Andrea E.

Shuey

Kim M.

Elder

Glen H

. 2007. “Cumulative Advantage Processes as Mechanisms of Inequality in Life Course Health.” American Journal of Sociology 112(6):1886–924.

58.

Zannas

Anthony S.

Arloth

Janine

Carrillo-Roa

Tania

Lurato

Stella

Roh

Simone

Ressler

Kerry J.

Nemeroff

Charles B.

, et al 2015. “Lifetime Stress Accelerates Epigenetic Aging in an Urban, African American Cohort: Relevance of Glucocorticoid Signaling.” Genome Biology 16:266. doi:10.1186/s13059-015-0828-5.

59.

Zhao

Wei

Ammous

Farah

Ratliff

Scott

Liu

Jiaxuan

Miao

Mosley

Thomas H.

Kardia

Sharon L. R.

Smith

Jennifer A

. 2019. “Education and Lifestyle Factors Are Associated with DNA Methylation Clocks in Older African Americans.” International Journal of Environmental Research and Public Health 16(17):3141 doi:10.3390/ijerph16173141.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.21 MB