Advanced Parental Age at Conception and Sex Affects Mitochondrial DNA Copy Number in Human and Fruit Flies

Abstract

Aging is a multifactorial trait caused by early as well as late-life circumstances. A society trend that parents deliberately delay having children is of concern to health professionals, for example as advanced parental age at conception increases disease risk profiles in offspring. We here aim to study if advanced parental age at conception affects mitochondrial DNA content, a cross-species biomarker of general health, in adult human twin offspring and in a model organism. We find no deteriorated mitochondrial DNA content at advanced parental age at conception, but human mitochondrial DNA content was higher in females than males, and the difference was twofold higher at advanced maternal age at conception. Similar parental age effects and sex-specific differences in mitochondrial DNA content were found in Drosophila melanogaster. In addition, parental longevity in humans associates with both mitochondrial DNA content and parental age at conception; thus, we carefully propose that a poorer disease risk profile from advanced parental age at conception might be surpassed by superior effects of parental successful late-life reproduction that associate with parental longevity.

In westerns countries, an increasing proportion of parents deliberately decide to delay having children and advanced parental age at conception increases the risk of miscarriages as well as the disease risk profile in offspring (1). This has been of concern to health professions and parental age at conception may be an important life circumstance influencing biological aging.

A central biological mechanism in all cells is the mitochondrion, which is a multifunctional cellular compartment that produces main energy resources such as adenosine triphosphate through oxidative phosphorylation. The oxidative phosphorylation processes induce toxic by-products such as reactive oxygen species, which can cause biological damage at high contents, but also act as a signaling molecule and have crucial roles in normal physiological processes (2) The mitochondrion has its own 16,569 base pair double-stranded circular mitochondrial DNA (mtDNA) (3). Mitochondrion function is regulated by biogenesis, disposal (autophagy or mitophagy), and morphology (fusion or fission) of the mitochondrion. Approximately 1,200 nuclear genes influence the mitochondria function in humans (4). At present, Tower and colleagues (5,6) highlighted that the mitochondrion acts differently in males and females as it is regulated by the sex-specific hormones and even the different gene dosages for genes located on the sex chromosomes may contribute to the mitochondrial sex difference.

The mtDNA content, that is the mitochondrial DNA copy number (mtDNA CN), is considered a proxy for the function of the mitochondrion, but in accordance with the estimated heritability of mtDNA CN, ranging from 33% to 65% (7–9), environmental factors such as toxin exposures and oxidative stress (10) also have a strong interindividual influence. When measured in blood, mtDNA CN correlates positively with good health and declines with age in elderly individuals (11,12). A decrease in mtDNA CN has been linked to multiple diseases, including cancer, cardiovascular diseases, and diabetes (3,7,8,13–16), and an elevated mtDNA CN associates with increased longevity, as demonstrated in two independent cohort-based populations (11,12). These findings were also confirmed in a longevity family study by Yong-Han and co-workers (17). Contradicting this, van Leeuwen and co-workers (18) found in the Leiden longevity study that mtDNA CN was lower among offspring of nonagenarians than their age-matched spouses. The authors suggested that mitochondrial biogenesis is not increased in long-lived families, but rather the mitochondrial function is preserved. In a Long Life Family Study, Sun and co-workers (19) demonstrated that mothers surviving to unusually old age were more likely to have given birth to their last child at an older maternal age. It could, therefore, be hypothesized that other factors such as parental age at conception may influence the link between mitochondrial biogenesis and longevity for offspring in family-based studies. Findings from participants from the Vitality 90+ study were supportive of the latter, suggesting that paternal age at conception was associated with transcriptomic changes involved in mitochondrial function (20).

Recent literature demonstrates that the amount of, possibly harmful, de novo mutations in offspring increases with paternal age at conception and that paternal age at conception attributes several fold more de novo mutations than maternal age at conception (21,22). Opposite to this harmful effect of paternal age at conception, favorable transgenerational impact was seen for leukocyte telomere length (LTL), as these are longer in offspring of older fathers, a paradox that is supported by telomere length found to be longer in sperm from older men (23). There are also indications of transgenerational influence on mtDNA CN (24–26). Studying maternal influence, for example maternal age at conception, on offspring mtDNA CN may be biologically meaningful as the mitochondrion is transmitted from the mother to the offspring, although mtDNA replicates independent of the cell cycle. For example, mtDNA CN was observed to be higher in aneuploidy embryos from older women than younger women (24) and maternal influence of mtDNA CN was demonstrated in domestic animals, but here mtDNA content in oocytes decreases with advanced maternal age at conception in horses (25) and cattle (26).

In this study, we make use of two species, Homo sapiens and Drosophila melanogaster (D melanogaster) primarily to investigate if advanced maternal age at conception influences mitochondrial functions in aging individuals, and if it is influenced differently in males and females. We used mtDNA CN as a surrogate measure of mitochondrial function in a cohort of middle-aged and elderly human twins. In parallel mtDNA, CN was measured in a cohort of young, middle-aged, and old D melanogaster produced by adults from each of the three age-classes. In addition, we use the twin design to investigate if maternal age at conception influences the family effect estimate of the mtDNA CN in biometric twin models. As secondary investigations, we study associations to parental longevity.

Methods

Human Study Population

The twins included were from two twin cohorts: The Middle Aged Danish Twins Studies and the Longitudinal Study of Ageing Danish Twins. Included in the study were 1,475 twins from middle-aged twins of which participants were distributed in 591 complete twin pairs. Middle-aged twins were 46–67 years of age when the study was initiated in 1998 (27). Participants were revisited from 2008 to 2011 and blood samples for this study were donated and used in the present work (28). Data sampling has been described elsewhere (11). The study included 671 twins from elderly twins of which participants were distributed in 291 complete twin pairs. Elderly twins were aged 70 years and older (29,30). Characteristics of the two sample populations are listed in Supplementary Table 1.

Parental Data—Human Data

Parental age at death and age when their offspring were born were defined by a questionnaire as described previously (31). Parental age at conception was computed by the average of the mean of two values by each twin in a pair, except if they differed by more than 5 years. Then they were treated as missing. For that reason, in middle-aged twins, paternal age at conception was treated as missing for four pairs, and maternal age at conception was treated as missing for five pairs. Parental age at death was the average of the mean of reported ages and computed from dates of birth and death, except if they differed by more than 5 years in which case they were set to missing (mothers: N_pairs = 7; fathers: N_pairs = 6). Twins of nonagenarians were called so if the father or mother was reported dead after 90 years of age or was alive and at least 90 years of age at the time of assessment. In total, 10 mothers and 10 fathers were alive at the time of assessment. Three pairs were omitted as the zygosity status was not known.

The same data-cleaning procedure was used for the elderly twins. For maternal age at conception, 17 pairs were treated as missing and for paternal age at conception 25 pairs were treated as missing. Parental age at death was exclusively self-reported from one question.

mtDNA CN Measurements—Human Samples

Blood samples were donated in 2008–2011 for the middle-aged twins and in 1996–1997 for the elderly twins. DNA was extracted from peripheral blood that was centrifuged at 1,000g for 15 minute followed by removal of plasma (buffy coat), using a standard manual DNA extraction kit (Gentrapure; Qiagen-LGC, Hertfordshire, UK), whereas DNA was extracted from blood samples from the elderly twins by a standard manual salting out protocol including a protease digestion step. MtDNA CN was measured as described in a previous article (11).

Drosophila melanogaster Study Design

A mass bred population of D melanogaster was established from the offspring of 30 naturally inseminated females caught in Karensminde orchard at the Danish peninsula of Jutland (55°56042.46ʺN, 10°12045.31ʺE) in October 2012. The results presented in this article are based on experiments performed in March and April 2017, that is 104 generations after flies were brought to the laboratory. Before performing the current experiment, the population was held at a size of minimum 1,000 individuals per generation on a standard Drosophila diet composed of yeast (60 g/L water), sucrose (40 g/L water), oatmeal (30 g/L water), and agar (16 g/L water) mixed with tap water. Following autoclaving, nipagin (12 mL/L water) and acetic acid (1.2 mL/L water) were added to the diet. Before performing the experiments and during the experiments, flies were kept at 25°C and 12-hour light/12-hour dark cycles.

Flies were density controlled during the development from eggs to adults one generation before initiating the experiment and all flies analyzed for mtDNA CN were also density controlled during development. This was done by collecting 30 fertilized eggs into vials with 7 mL of standard Drosophila diet. Newly emerged virgin flies (less than 8 hours old) were collected from the vials they developed in, separated into males and females, and distributed into new vials with 7 mL medium. At respectively 3, 19, and 35 days of age males and females were mixed and allowed to reproduce. Eggs were then produced by females from each of these age-classes and distributed into vials with 30 eggs in each. Emerging flies were separated in males and females and transferred to new vials with fresh food every second day. Three replicates with five males and three replicates with five females were sampled from each of the three parental age-classes when flies reached the ages 3, 19, and 35 days. Thus, in total we collected three replicates each with five flies per sex when flies reached the ages 3, 19, and 35 days for each of the three parental age-classes totaling to 27 samples per sex. These flies were used for mtDNA CN measurements.

mtDNA CN Measurements in D melanogaster

mtDNA CN analyses were performed as described previously using triplicate samples and SYBR Green technology (Applied Biosystems) (11). For D melanogaster one 100 bp polymerase chain reaction was targeted to the mitochondrial NADH dehydrogenase 4 L gene (ND4L) using the primer sequences 5′-TAAGAAAATTCCGAGGGATTCA-3′ for the forward primer and 5′-GGTCGAGCTCCAATTCAAGTTA-3′ for the reverse primer. Primers were originally published elsewhere (32). To quantify the amount of mtDNA, another 126 bp polymerase chain reaction targeted to the nuclear rosy gene the primer sequences were for the forward primer 5′-GGTGGTGAGCCTGTTCTTCAAG-3′ and the reverse primer 5′-ACTGGTGTGTGGAATGTCTCGG-3′. Primers were originally published by Wu and colleagues (33). The amplification was preheated at 95°C for 20 seconds followed by a 40 cycle program of 0.3 seconds at 95°C, 15 seconds at 52°C, and 30 seconds at 72°C.

Statistical Analysis Human Data

Within-pair dependence of monozygotic and dizygotic twin pairs was assessed by estimating the within-pair correlation adjusting for sex, age, and cohort with standard assumptions on the correlation structure. Classical biometric variance component models, ACE and ADE, were also fitted to data adjusting the mean structure for sex, age, and cohort effects. These were compared to AE, CE, and E models using the AIC (34,35).

For analyzing the influence of paternal age at conception and maternal age at conception on mtDNA CN correlation, we divided fathers and mothers into two age groups at the median age (33 and 30 years, respectively). Again, monozygotic and dizygotic within-pair correlations stratified by paternal or maternal age group and correcting for sex, age, and cohort were estimated. Analysis was performed using the R-package mets (34,35).

We modeled the association between parental age at birth and mtDNA CN by a generalized additive mixed model with sex, age, zygosity, and paternal age at birth as fixed effects. The effect of paternal age at conception and maternal age at conception was modeled using a cubic spline with four basis functions. The model was fitted using the R-package gamm4 (R package mets).

We used a linear mixed model with twin pair as random effects and sex, age, and paternal or maternal age group. Owing to possible nonindependence within twin pairs, the analysis was performed using the robust estimator of variance. Sex-specific analyses were performed using interaction terms. The statistical calculations were performed using Stata, version 14.2 (StataCorp).

Statistical Analysis D melanogaster Data

To test whether the medians of the mtDNA CN were significantly different across age-classes (3, 19, and 35 days of age), Kruskal–Wallis tests were conducted keeping sex (males and females) and parental age at conception (3, 19, and 35 days) separated to control for the effect of sex and parental age at conception. The pairwise comparisons testing if mtDNA CN was significantly different across age of the flies (3 vs 19 vs 35 and 3 vs 35 days of age) were made using Mann–Whitney U-tests again keeping sex and parental age at conception separated.

To test whether the age of the parents affected the medians of the mtDNA CN in offspring flies, we performed a Kruskal–Wallis test keeping the sex and the age of the offspring separately.

The pairwise comparisons for testing if mtDNA CN was significantly different across offspring flies with different parental age at conception (parental age of 3 vs 19 vs 35 and 3 vs 35 days of age) were made using Mann–Whitney U-tests, again keeping sex and offspring age separated.

Further, within each parental age-class, we tested whether mtDNA CN differed between similarly aged males and females using Mann–Whitney U-tests.

Results

Data Description

Characteristics of the human twin cohorts are listed in Supplementary Table 1. Data were available for both paternal ages at conception. Parents’ survival was divided into those surviving to be nonagenarians (90+) and those who died younger. As expected, there was a strong correlation between the paternal ages at conception (r = .76). Fathers who lived to be nonagenarians had an older mean paternal age at conception (Δ: 5.5 years, p < .001), and also an older mean maternal age at conception was observed for their spouses (Δ: 2.9 years, p < .001). For mothers who lived to be nonagenarians only a slightly older mean maternal age at conception was observed (Δ: 1.3 years, p = .09), and similarly a slightly older mean paternal age at conception was found for their spouses (Δ: 1.1. years p = .29).

mtDNA CN in Twin Pairs and Parental Age at Conception

Analyses were carried out to deduce if parental age of conception influences familial effects. In groups of equal sizes, the correlation of mtDNA CN in twin pairs was estimated separately for monozygotic and dizygotic twins. In the two twin cohorts, a notable high monozygotic twin pair similarity in mtDNA CN (r = .48) was observed for offspring of mothers with a maternal age at conception below the mean (<30 years of age), whereas for monozygotic offspring of mothers with a maternal age at conception above the mean the correlation was low. For dizygotic offspring at any maternal age at conception, the correlation in mtDNA CN was low. Similar evidence was observed for paternal age at conception (mean: 33 years of age; Figure 1). This suggests the presence of familial effects that may be modified by parental age at conception. The broad sense heritability (H²) of mtDNA CN in twin offspring adjusted for effects of gender, cohort, and sex was estimated to 0.18 (95% CI = –0.08, 0.43; ACE model), and there were only minor common environmental effects (Table 1). Results from the same model showed the H² of mtDNA CN was higher for twin offspring of mothers below the mean maternal age at conception (H²: 0.34, 95% CI = 0.19, 0.49).

Figure 1.

Plots of twin mitochondria DNA copy number (mtDNA CN) correlation within parental age groups of equal sizes. (A) Plots of maternal age at conception and (B) plots of paternal age at conception from the twin cohorts. Correlations included an adjustment for sex and cohort.

Table 1.

Estimates of Genetic and Environmental Effects on Mitochondria DNA Copy Number in the Twins ACE, ADE, AE, and CE Model were Adjusted for Sex and Cohort

	H 2 (95% CI)	A (95% CI)	C (95% CI)	D (95% CI)	E (95% CI)	AIC
ACE	0.18 (–0.08, 0.43)	0.18 (0.03, 0.79)	0.03 (–0.16, 0.23)	—	0.79 (0.69, 0.88)	16,822.21
ADE	0.22 (0.14, 0.31)	0.22 (0.14, 0.31)	—	0 (0,0)	0.78 (0.69, 0.86)	16,822.34
AE	0.22 (0.14, 0.31)	0.22 (0.14, 0.31)	—	—	0.78 (0.69, 0.86)	16,820.34
CE	—	—	0.16 (0.10, 0.23)	—	0.83 (0.77, 0.90)	16,822.08

mtDNA CN, Sex, and Advanced Parental Age at Conception

As displayed in Figure 2, there were no evidence that mtDNA CN deteriorates with parental age rather mtDNA CN was slightly elevated in offspring of advanced ages for both parental ages at conception. For example, twins born to mothers of advanced maternal age at conception (≥40 years old) had approximately six more mtDNA copies pr. nuclear DNA than twins born to younger mothers (β = 6.6, SE = 3.8, p = .08), adjusted for age and cohort.

Figure 2.

Deviation of mitochondria DNA copy number (mtDNA CN) from expected mean value, depending on parental age at conception in the twin cohorts. (A) Plots of normalized mtDNA CN with maternal age at conception and (B) plots of normalized mtDNA CN with paternal age at conception. Data are adjusted for age, sex, and cohort.

mtDNA CN was also found to be higher in women than in men, so the difference between sexes and the effect of advanced maternal age at conception was investigated further. A significant elevation in mtDNA CN was observed in female offspring of parents with increasing maternal (β = 0.29, SE = 0.12, p = .02) and paternal (β = 0.31, SE = 0.13, p = 0.02) age at conception, adjusted for age and cohort. For instance, female offspring of mothers with advanced maternal age at conception (40+ years at birth) had higher mtDNA copies pr. nuclear DNA than male offspring (β = 19.6, SE = 5.9, p = .001), a difference twice the magnitude observed for twin offspring of younger mothers (β = 8.3, SE = 1.4, p < .001), adjusted for age and cohort. The sex difference is most pronounced at middle ages and diminishes with age as demonstrated in Figure 3.

Figure 3.

Plots of mitochondria DNA copy number (mtDNA CN) with age within maternal age at conception groups. Males (blue line) and females (red line) mtDNA ratio (medians) at the maternal age at conception below 35, 35–40 years, and above 40 years. The model was adjusted for cohort differences.

mtDNA CN and Parental Longevity

Offspring mtDNA CN association with parental longevity was explored further for two reasons; first, to see if we could replicate findings from previous studies of mtDNA CN in relation to life span in long-lived families, and second, because twin parents who lived to be nonagenarians in the current work were also older when the twins were born. Here, offspring to nonagenarian fathers had on average six more mtDNA copies pr. nuclear DNA (β = 5.7, SE = 2.5, p = .02) thus indicating that paternal longevity also associates with mtDNA CN, whereas no significant increase was observed for offspring of nonagenarian mothers (β = –1.5, SE = 1.8, p = .42).

Advanced Parental Age at Conception and mtDNA CN in D melanogaster

To investigate the effects of sex and advanced parental age at conception on the mtDNA CN level in another species, we conducted experiments using the model species D melanogaster. First, we investigated if the mtDNA CN varies with age in male and female D melanogaster and observed a general tendency for a reduction in mtDNA CN with age (Figure 4). The age reduction in mtDNA CN was, however, only significant for a fraction of the tests performed, that is males of parents 3 and 19 days old and females of parents being 19 days old (Figure 4 and Supplementary Table 2). Therefore, to robustly confirm conclusion about the effect of age on the mtDNA CN more extensive studies are needed. Second, we explored if parental age at conception influenced the mtDNA CN in offspring investigated at different ages. We found that parental age at conception had a strong influence on mtDNA CN in their offspring in both females and males of all ages (3, 19, and 35 days old) with only one exception (females with 35-day-old parents). Results for male offspring showed that the older the parents were the lower the mtDNA CN. In female offspring, individuals with a parental age at conception of 19 days had the highest mtDNA CN (Figure 4 and Supplementary Table 3). Interestingly, the paternal age effect was larger than the age-related decline (Figure 4). Third, the mtDNA CN was investigated for differences between male and female offspring at the same age and same parental age at conception. Female offspring tended to have a higher mtDNA CN than male offspring at the same age and parental age at conception in aging flies (19 and 35 days old), and with the largest sex differences at advanced parental age at conception (Figure 4 and Supplementary Table 4). However, for 3-day-old flies with 3-day-old parents, we observe an unexpected significantly higher mtDNA CN in males compared to females (Figure 4 and Supplementary Table 4).

Figure 4.

Plots of mean mitochondria DNA copy number (mtDNA CN) within parental age group in Drosophila melanogaster males (red line) and females (black line). Males mtDNA ratio (medians) at the age of 3, 19, and 35 days of age from parents that are 3, 19, and 35 days of age. The whiskers present the upper and lower limits of the 95% confidence intervals.

Discussion

More and more individuals deliberately decide to delay to have children (1). This is a concern for health professionals as it is argued that advanced parental age at conception has undesirable effects on the offspring (1). We find evidence that both parental age at conception associate with elevated mtDNA CN, although mainly in female offspring. Elevated mtDNA CN is considered an indication of good health and therefore stands in contrast to the possible undesirable effects of advanced parental age at conception. As expected, female offspring have a higher mtDNA CN level than male offspring, a difference that was most explicit for offspring born to mothers with advanced maternal age at conception. Supportive evidence was found in the D melanogaster studies where the mtDNA CN level was higher in females than males, which was particularly pronounced at advanced parental age at conception (19+ days old). Also, noteworthy parental age effect in fruit flies was in the present work larger than the effect of age per se.

Advanced paternal age at conception may have adverse effects for offspring prenatally or early in life, for example, causing greater risk of stillbirth, cleft palate, and early-life mortality, as well as later in life where offspring have increased risks of neurodevelopmental disorders, such as Autism spectrum disorders and Schizophrenia (36–38). Recent studies have also shown that advanced paternal age at conception associates with higher amounts of de novo mutations, of which some are likely harmful (21,22). Likewise, advanced maternal age at conception is associated with increased risk of undesirable effects for offspring, for example, trisomy 21 (39). Only a few previous studies have included investigations of the influence of parental age on mtDNA CN in aging individuals (17,18,40). Initially, we anticipated that advanced parental age at conception would associate inversely with mtDNA CN, as higher mtDNA CN level is associated with better health and lower mortality (11,12). Unexpectedly, an opposing positive association was observed in the present work. The results were, however, in accordance with studies showing that advanced paternal age at conception associates with longer LTL in twin offspring (23). MtDNA CN and LTL do seem to correlate and are likely related in a pathway dubbed the telomere–mitochondria axis (41–43). However, we do not believe the paternal age at conception association with mtDNA CN from the present work is simply a consequence of mtDNA CN and LTL correlation, as the mtDNA CN correlation in twin pairs declines with parental age at conception in the present work (Figure 2) rather than increases with paternal age at conception as it was observed for LTL (23).

Our finding that parental age at conception influences mtDNA CN differently in females and males supports the hypothesis that the mitochondrion is actually differently regulated in the two sexes, as proposed by Tower and colleagues (5,6). The authors also argue for a biological cause for this, for example, by the action of the sex-specific hormones estrogen and testosterone, and/or because females have a different gene dosage than males. As the parental age effect is mainly observed in females, an excess maternal estrogenic or related hormonal influence from the hypothalamic–pituitary–ovarian axis, that is progesterone, Luteinizing hormone and Follicle-stimulating hormone, is a likely mechanistic explanation for the maternal influence on mtDNA CN (44). We can only speculate if this could be a direct influence, influence through epigenetic mechanism, or caused by genetic variations, but there are needs of supporting evidence of the results in the present work. Yet, at least one study has demonstrated leukocyte mtDNA CN is altered in women who use sex-specific hormones, that is by hormone replacement therapy (45). A key finding in this work is the parental longevity association with both mtDNA content and parental age at conception in humans. This cautiously proposes that disease risk profile from advanced parental age at conception in adults may be surpassed by superior effects from successful late-life reproduction that associates with longevity. This supports late reproduction being linked to parental longevity, as previously shown in the Long Life Family Study where maternal age at their last childbirth associated with longevity (19). In addition, nonagenarian fathers became fathers of twins later in life than father who did not live to be 90 years old and the offspring mtDNA CN also associated with paternal longevity, which was supported in the longevity family study by Yong-Han and co-workers (17) and the fact that mtDNA CN associate with survival in cohort studies (11,12).

The present twin study included middle-aged to elderly twins and the positive association found between advanced maternal age at conception and offspring mtDNA CN may thus be subjected to age-related changes, in the sense that the association may diverge at older ages. Also, our study is limited to only including twins, thus further studies are needed to investigate if associations also apply to singletons. However, the findings from the fruit fly study corroborate that the parental age at conception is of importance, even across evolutionary very distinct species, and may be more important than the effect of individuals' age per se.

By making use of the twin design, we were able to illustrate that parental age at conception is not only associated with mtDNA CN but may also modify the presence of familial effects (Figure 1). Although highly speculative this might indicate that maternal age at conception influences heterogeneity among the many mitochondria in each cell during early-life formation. The overall heritability of mtDNA CN in the present work was estimated to be lower than what has been reported previously in American twins (65%) and in a family-based study (33%) (7,9). These previous reports included on average younger persons than the twins in this study, and this may thus indicate that age influences the mtDNA CN heritability, although other factors, such as technical procedures, randomness, sampling, and genetic variability between populations, may contribute to the diversity in heritability estimates.

Conclusion

We found no evidence that advanced paternal age deteriorates the health measure mtDNA CN, yet females had higher mtDNA CN than males in both human and fruit flies, and the difference was higher at advanced maternal age at conception. Also, advanced maternal age influences the presence of familial effects in biometric twin models. Association between parental longevity and both mtDNA content and longevity suggests poorer disease risk profile from advanced parental age might be surpassed by superior effects of parental successful late-life reproduction that associates with parental longevity.

Funding

This work was supported by the European Union’s Seventh Framework Programme (FP7/2007-2011) under grant agreement n° 259679, The Danish National Program for Research Infrastructure 2007 (09-063256), from the Danish Agency for Science, Technology and Innovation, the US National Institutes of Health (P01 AG08761) and Odense University Hospital Free Research Fund. C.P. was supported by the Aalborg Zoo Conservation Foundation (grant number 2017-3). The Danish Aging Research Center is supported by a grant from the VELUX Foundation.

Conflict of Interest

The authors have no conflicts of interest.