Birth size and the serum level of biological age markers in men

Previous studies showed that intrauterine growth restrictions, resulting in smaller body size at birth, are associated with altered development and the risk of age-related diseases in adult life. Thus, prenatal development may predict aging trajectories in humans. The study aimed to verify if body size at birth is related to biological age in adult men. The study sample consisted of 159 healthy, non-smoking men with a mean age of 35.24 (SD 3.44) years. Birth weight and length were taken from medical records. The ponderal index at birth was calculated. Biological age was evaluated based on serum levels of s-Klotho, hsCRP, DHEA/S, and oxidative stress markers. Pregnancy age at birth, lifestyle, weight, cortisol, and testosterone levels were controlled. The results showed no relationship between birth size and s-Klotho, DHEA/S level, inflammation, or oxidative stress. Also, men born as small-for-gestational-age (N = 49) and men born as appropriate-for-gestational-age (N = 110) did not differ in terms of biological age markers levels. The results were similar when controlled for pregnancy week at birth, chronological age, BMI, testosterone, or cortisol level. The results suggest that there is no relationship between intrauterine growth and biomarkers of aging in men aged 30–45 years from the affluent population.

higher testosterone level (M = 518.98,SD = 171.61)compared to men who did not practice sport (M = 462.81,SD = 186.81)(t(157) = 1.97, p = 0.05, Cohen's d = − 0.31).The two groups did not differ in terms of the other markers of biological age or chronological age (in each case p > 0.23).Men who smoked in the past and men who have never smoked did not differ in terms of markers of biological age (in each case p > 0.24).S-Klotho level was lower in men who drank alcohol more often (F(2,156) = 5.06, p = 0.007, η 2 = 0.06).These three groups did not differ in terms of other markers of biological age (p > 0.11).DHEA, DHEAS-S levels, and the composite index of biological age were negatively correlated with age.hsCRP was positively related to adiposity and the composite index of biological age was negatively related to adiposity.DHEA/S and composite index of biological age were also positively related to cortisol levels.hsCRP correlated negatively with testosterone levels (Table 2).
Birthweight was positively related to pregnancy week at birth and TAC levels and negatively with cortisol level.The ponderal index was not related to any controlled variable (Table 3).
The relationship between birth size and markers of biological age.The results of zero-order correlations showed that the s-Klotho level showed no relationship between birth weight or ponderal index and markers of biological age (Table 4).Also, there was no difference between men classified as SGA (N = 49) and classified as AGA (N = 119) comparing the mean values of the levels of markers of biological age (in each case p > 0.28).The relationship between birth size and biological age parameters controlled for cofounders.S-Klotho level was not related to birth weight or ponderal index when controlled for pregnancy week at birth and frequency of the alcohol drinking.S-Klotho level was only related to the frequency of alcohol drinking (Table 5).There was no difference in the mean s-Klotho levels between men from the three terciles of birthweight (F(2,156) = 0.04, p = 0.96, η 2 < 0.001) or ponderal index (F(2,156) = 0.23, p = 0.80, η 2 = 0.003).DHEA level was not related to birth weight or ponderal index when controlled for pregnancy week at birth, age, adiposity, and cortisol level (factors that may impact DHEA levels-Table 2).DHEA was positively related to cortisol levels (Model 1 and 2-Table 6).Similarly, there was no relationship between DHEA-S and birth weight or ponderal index when controlled for pregnancy week at birth, chronological age, and cortisol level.DHEA-S was negatively related to age (Model 3 and 4-Table 6).There was no difference in the mean DHEA levels between men from the three terciles of birthweight (F(2,156) = 0.71, p = 0.49, η 2 = 0.01) or ponderal index (F(2,156) = 0.12, p = 0.88, η 2 = 0.001).There was also no difference in the mean DHEA-S levels between men from the three terciles of birthweight (F(2,156) = 0.48, p = 0.62, η 2 = 0.006) or ponderal index (F(2,156) = 0.30, p = 0.74, η 2 = 0.004).www.nature.com/scientificreports/hsCRP level was not related to birth weight or ponderal index when controlled for pregnancy week at birth, adiposity and regular physical activity.Inflammation was positively related only to adiposity level (Model 1 and 2-Table 7).We have also run similar model including also testosterone and cortisol and we also found no relationship between birth size and hsCRP level and the model with higher number of predictors had worse fit (BIC = 273.181vs BIC = 265.72).Additionally, we verified if adiposity moderated the relationship between birth weight/ponderal index and hsCRP by introducing the interaction between birthweight and body adiposity in the model.The model showed no relationship between birth weight and adiposity (p = 0.13 for birth weight; p = 0.94 for ponderal index).There was no difference in the mean inflammation levels between men from the three terciles of birthweight (F(2,156) = 0.47, p = 0.63, η 2 = 0.006) or ponderal index (F(2,156) = 0.31, p = 0.73, η 2 = 0.006).
OS level was not related to birth weight or ponderal index when controlled for pregnancy week at birth and TAC (in order to account for a possible impact of antioxidant capacity on the relationship between birth size and oxidative stress level (Table 8).There was no difference in the mean OS levels between men from the three terciles of birthweight (F(2,156) = 0.17, p = 0.85, η 2 = 0.002) or ponderal index (F(2,156) = 0.91, p = 0.41, η 2 = 0.01).
The composite index of biological age was also not related to birth weight or ponderal index when controlled for chronological age, adiposity, and cortisol level.Biological age was negatively related to adiposity level and positively to cortisol level (Table 9).There was no difference in the mean biological age index between men from the three terciles of birthweight (F(2,156) = 0.76, p = 0.47, η 2 = 0.01) or ponderal index (F(2,156) = 0.29, p = 0.40, η 2 = 0.004).

Discussion
The results of this study showed no relationship between birth size and physiological markers of biological age in men between 30 and 45 years.Birth weight but not the ponderal index was only positively related to TAC level and negatively to cortisol level.Furthermore, the levels of biological age markers were mainly related to lifestyle factors, adiposity, and testosterone level.hsCRP was negatively related to physical activity and testosterone level and positively to adiposity.The S-Klotho level was negatively related to the frequency of alcohol drinking.DHEA and DHEA-s were negatively related to age and positively to cortisol levels.
The results of the previous research on the relationship between birth size and senescence showed contradictory results.Animal studies have shown that growth restriction during pregnancy and subsequent catch-up Table 7.The results of multiple regression analyses for the relationship between hsCRP level and birth weight (Model 1) or ponderal index (Model 2) controlled for cofounders (N = 159).Significant values are in bold.a Coded as 0 = "no" and 1 = "yes".www.nature.com/scientificreports/growth diminishes telomere length in rats that ultimately have also a reduced life span [72][73][74][75] .In human adults with a mean age of 43.04 years, birth weight was inversely associated with p21 gene expression but not with p16 expression, markers of cellular senescence.What is interesting, the expression of p21 but not p16, was positively correlated with BMI gain at 29 years and current BMI 76 .In men, lower birth weight predicted advanced biological aging based on epigenetic clocks 21 .Furthermore, no association was found between birth weight and telomere length in humans 24,77,78 .Masterson et al. 79 showed that heavier birth weight was associated with longer telomere length but the relationship was attenuated by maternal age at birth and probably rather related to childhood growth.Other studies show that BMI in childhood may be more important for senescence than birth size 80,81 .In this study, no relationship between birthweight and biological age was found as measured as a composite index of physiological markers of biological age.We also found no relation between birth weight and inflammation level, also when body adiposity was controlled.This result is in line with sparse studies showing no relationship between birth size and inflammation 82 and at variance with previous studies which showed that birth weight was negatively associated with adolescent or adult CRP levels [83][84][85] .However, previous research has also shown that this relationship may be mediated by an elevated body fat mass among individuals with lower birthweight [85][86][87] .Birth weight-adult inflammation risk is mediated by later BMI and BMI change and inflammatory risk related to birthweight can be offset by weight loss 88 .However, in our study, although hsCRP level was positively related to body adiposity, physical activity, and testosterone level, these factors did not moderate the relationship between birth weight and inflammation level.Other studies suggest that breastfeeding 89 , life-course socioeconomic status 90 , and diet 91 are important factors in moderating the relationship between birth weight and adult inflammation.Also, the role of such modifiers as sex or age is still unclear 92 .We cannot exclude the possibility that some lifestyle factors that were not controlled in our study might impact the relationship between inflammation and birth size.
We found no relationship between birth size and s-Klotho level in adult men.Previous research has shown that newborns with SGA have lower levels of klotho 93 .No research, so far investigated if the relationship persists to adulthood.What is interesting, the s-Klotho level was also not related to physiological markers of cardiometabolic risk in this group 94 .This is in line with other research showing no relationship between s-Klotho level and health outcomes [95][96][97] .It is possible that the role of s-Klotho as a marker of biological age may depend on age and may be more valid in older individuals 94,98 Also, other research suggested sex differences in the relationship between the s-Klotho level and health components 99 that might be explained by different profiles of sex hormones and the interaction between s-Klotho and testosterone or estradiol levels 100,101 .However, the physiological effects of sex steroid hormones on klotho levels have not been completely elucidated and require further studies.Furthermore, the only factor related to klotho level was the frequency of alcohol drinking which is in line with the previous research [102][103][104] .Quintero-Platt et al. 105 observed higher s-Klotho levels among alcoholics than in controls, but these differences were dependent on the presence of cirrhosis.This higher s-Klotho plasma levels in cirrhotic patients could be due to a pathologic status, where the liver function is impaired, and thus not applicable to healthy individuals.Several studies showed an increase in s-Klotho plasma levels in diabetics 106 , patients with acromegaly 107 , and patients with autosomal dominant polycystic kidney disease 108 .However, it is also important to note that many studies show a positive correlation between s-Klotho and markers of health and biological age 109,110 , also in patients with various disorders 111 .
We also found no relationship between size at birth and DHEA/S level in adult men.Most of the research on the relationship between birth size and adrenal androgens level was conducted in children or young adults and the results are equivocal.For instance, previous research showed a relationship between birth size and DHEA and DHEAS levels in women but not in men of age 18-22 years 112 On the other hand, higher DHEAS levels at 7 years old were also associated with lower birth weight and higher adiposity 113 .Although, experimental data in animals and recent human observations have suggested that an alteration in the set point of www.nature.com/scientificreports/ the hypothalamic-pituitary-adrenal axis entails important long-term change that occurs in association with reduced fetal growth 33 we did not observe the relationship between DHEA/S level and birth weight in men of 30-45 years.We only confirmed that DHEA/S level was only inversely related to chronological age and positively to cortisol level.Many known or suspected causes or conditions associated with adverse (poor or excessive) fetal growth or preterm birth have been associated with oxidative stress (OS).Although OS may be one of the mechanisms linking impaired fetal growth with certain chronic diseases in adulthood 114,115 we found no relationship between birth size and OS level in men of 30-45 years old.Previous research showed increases in oxidative stress markers in SGA children [116][117][118] and male rats exposed to maternal protein restrictions 119 or placental insufficiency 116 .The mechanisms of oxidative stress programming may be through directly modulating gene expression or indirectly through the effects of certain oxidized molecules, which could be compensated by increased antioxidant levels 114 .Interestingly, we found a positive relationship between birth weight, but not ponderal index, and serum total antioxidant capacity.Animal studies suggest that the antioxidant expression and activity are up-regulated in the IUGR (intrauterine growth restriction) rats that are normotensive in young adulthood, implicating a compensatory mechanism that may be protective against the generation of reactive oxygen species in adult IUGR rats 119 .Also, human studies have shown that SGA children have elevated OS markers and also some antioxidants 120 .Possibly, oxidative stress in adults born as small for gestational age can be attenuated by elevated antioxidant levels.
This research study has several methodological strengths.Study subjects were homogeneous in terms of age and ethnic background.Information on birth weight and gestational age at birth was drawn from data recorded by hospital or clinic staff in the maternal and child health handbook.Several methodological limitations of the current study should also be clarified.First, the study had a cross-sectional design, thus we cannot exclude the influence of various factors that could impact the relationship during ontogeny.Second, biological age markers were assayed only once.Despite the rigorous study protocol, we cannot exclude the possibility of some intraindividual variability in the measured markers.Finally, the relatively small sample size may limit the statistical power of the analysis, and small differences in biological age as a function of birth weight may not have been detected.

Materials and methods
This study was a part of a broader project on factors related to men's health.All methods were carried out by relevant guidelines and regulations and were reviewed and approved by the Commission of Bioethics at Wroclaw Medical University (nr 222/2019).All men signed written informed consent for participation in the study.

Participants.
Information on the ongoing study was posted in local newspapers, on social media, and broadcasted on the local radio.209 men (M age = 36.14± 3.53) were recruited from an urban Polish population after the initial screening to exclude chronic and acute health problems.Information on birth parameters was available for 183 men.From this sample, 24 men were excluded due to the following reasons: (1) chronic diseases-e.g.diabetes (N = 3); (2) regular smoking (N = 8); (3) CRP > 10 mg/l, indicating ongoing inflammation (N = 1); (4) incomplete measurements of physiological markers of biological age (N = 12).Health status was evaluated based on inflammation, blood morphology, and self-reported health and none of the participants exhibited symptoms of infection.Thus, the final analyses were conducted on 159 healthy men of mean age 35.24 ± 3.42 years (29.83-44.29 years).An a priori power analysis was conducted using G*Power 126 for minimum sample size estimation.Results indicated the required sample size to achieve 80% power for detecting a medium effect at a significance criterion of α = 0.05 and five predictors is 92 for the regression analysis.Thus, the obtained sample size of N = 159 is adequate to test the study hypothesis.

Procedures.
Participants were asked to refrain from physical activity, heavy meals, and alcohol for 24 h before the study visit.A fasting blood sample was taken between 7:30 a.m. and 9:00 a.m. for blood biochemical and hormone analyses.In addition to serum markers of biological age, also testosterone and cortisol levels were measured.Age-related decline in testosterone levels is linked with increased vulnerability to multiple chronic diseases 121 .We controlled also cortisol level as its level contributes to the pathophysiological mechanisms of aging 122 .Adiposity may impact the relationship between birth weight and measures of the adult biological condition 123 thus participants' body adiposity was measured with bioimpedance analysis (SECA mBCA 515).
The participants completed a questionnaire about demographic data, current and past health problems, medications, smoking, and alcohol drinking patterns.Among the participants, 28 individuals declared that they had regularly smoked in the past (quit at least one year before the visit) and 131 stated that they had never smoked.Participants were also asked about the level of physical activity: the number and length (in minutes) of training per week and the type of sport practiced.The types of sports activities practiced by the participants were comparable in terms of intensity (running, biking, swimming, football, basketball, calisthenics, tennis, squash, CrossFit, and strength workout).The participants were divided into two categories: (1) physically active (N = 86), which included individuals who declared regular physical activity at least 60 min/week; and (2) inactive (N = 73) i.e. with no regular sport activity.There were no professional sportsmen in the study sample.Participants were also asked how often they drink alcohol and based on their answers they were divided into three groups: (1) rarely-i.e.once per month or less often (N = 40), (2) sometimes-i.e.2-4 times per month (N = 72), (3) ofteni.e.2-3 times per week (N = 47).
Birth parameters.Information on participants' birth weight and length, and health condition at birth was obtained from the medical records (personal "health books").The ponderal index at birth was calculated as birth 0.1 × weight/(birth length) 3 and expressed in kg/m 3  www.nature.com/scientificreports/intrauterine growth restrictions.We controlled for pregnancy week at birth, as it may impact the birth size also of children born in term 124 .
Biological age markers measurement.Blood samples were centrifuged and serum was collected and stored at − 80 °C until the analyses.Sample preparation and all test procedures followed the manual supplied with each ELISA kit.Serum soluble circulating Klotho (s-Klotho) level was measured with enzyme-linked immunosorbent assay (ELISA) using commercial kits (IBL ® Code no 27998; with inter-assay and intra-assay coefficient of variation less than 11.4% and less than 3.5% respectively with assay sensitivity of 6.15 pg/ml), following the manufacturer's protocol.Calibrators (standards supplied with the kit) and serum samples were assayed in duplicate and the average absorbance value was used to calculate hormone concentration.The standard curve was created by plotting mean absorbance values for each standard (Y axis) against its concentration (X axis).Total s-Klotho concentration was calculated in relation to the standard curve, multiplied by the dilution ratio, and expressed in pg/ml.Serum 8-epi-PGF2α level was assayed with an Elabscience competitive-ELISA kit (cat.No E-EL-0041) with intra-and inter-CV less than 7%.Protein carbonyls were assayed with a MyBioSource ELISA kit (cat.No MBS3802635).DNA/RNA oxidative damage was assayed with a Cayman ELISA kit (cat.No 589320) with interand intra-CV less than 12%.The kit allows to measure of DNA/RNA oxidative damage, asses 8-OH-dG from DNA, 8-OHG from RNA, and 8-hydroxyguanine from either DNA or RNA.
Serum hsCRP level was measured using a DEMEDITEC ELISA kit (cat no 740011) with inter-and intra-CV less than 6.3% and 6.9% respectively.
Serum DHEA and DHEA-S levels were assayed using a DEMEDITEC competitive ELISA kit.DHEA (cat no DEH3344) level was measured with inter-and intra-CV less than 6.9% and less than 6.9% respectively.DHEA-S (cat no DEH3366) level was measured with inter-and intra-CV less than 12.2% and less than 6.8% respectively.
Serum total antioxidant capacity (TAC) was assayed with a Cayman ELISA kit (cat.No 709001), with an inter-assay CV is 3% and an intra-assay CV is 3.4%.The kit allows for assessing of aqueous-and lipid-soluble antioxidants, thus the combined antioxidant capacities of all its constituents including vitamins, proteins, lipids, glutathione, uric acid, etc. were assessed.
Serum cortisol level was measured using a DEMEDITEC ELISA kit (cat no DEH3388) with inter-and intra-CV less than 9.2% and less than 7.2% respectively.
Serum samples and calibrators (and controls if supplied with a kit) were assayed in duplicate in all ELISA tests.Absorbance (OD-optical density) was measured using a microplate reader (ASUS UVR340) with at λ recommended by each manual.The calculation of results was based on the calibration curve (for absorbance value for each sample the corresponding concentration from the calibration curve was determined).The average value of sample concentration (from duplicate measures) was used in analyses.
Quantitative measurement of total testosterone was evaluated by a certified analytical laboratory (DIAGNOSTYKA ® ) using a Cobas analyzer and expressed in ng/ml.

Statistical analyses.
Normality was assessed based on Kolmogorov-Smirnov's tests, kurtosis, skewness, and plot visual inspection.The results of Kolmogorov-Smirnov's test showed that values of age, birth weight, ponderal index, DHEAS-S, TAC, body adiposity, cortisol, and testosterone levels had a normal distribution.Distribution of pregnancy week at birth was assessed as normal based on the z-scores values for kurtosis and skewness that were below 3.29.S-klotho, DHEA, hsCRP, 8-isoepiprostaglandine, and RNA/DNA levels were log-transformed due to positive results of the Kolmogorov-Smirnov's test, high kurtosis and/or skewness levels.
An aggregate measure incorporating oxidative stress markers should more accurately assess the levels of antioxidant potential and OS than a single biomarker 125 .Thus, an aggregate measure of oxidative stress was calculated by Z scoring each OS biomarker (each assessing different outcomes of oxidative stress) and averaging for each individual the three Z scores.
Similarly, an aggregate measure of biological age was calculated by Z scoring each biomarker (s-Klotho, hsCRP, DHEA, DHEA-S, 8-isoepiprostaglandine, RNA/DNA, PCs levels) and averaging for each individual the seven Z scores.Z scores of hsCRP and oxidative stress markers were multiplied by − 1 to account for the inverse relationship with biological age.The higher values of this composite measure indicated younger biological age.
Similarly, as in our previous study 94 , the t-test was used to verify if participants who smoked in the past (N = 28) and participants who have never smoked (N = 131) differed in terms of markers of biological age.T-test was also used to verify if participants classified as physically active (N = 86) and participants classified as inactive (N = 73) differed in terms of markers of biological age.ANOVA was used to verify if participants differed in terms of levels of s-Klotho and cardiometabolic risk markers levels depending on how often they drink alcohol: First, the relationship between birth parameters and biological age markers levels was verified with Pearson correlation analysis.The mean levels of markers of biological age were compared between men classified as small-for-gestational-age (SGA; N = 49) and men classified as appropriate for gestational age (AGA; N = 110) using a T-test.Participants were divided into three groups according to the values of terciles of birth weight: www.nature.com/scientificreports/used to test the association between levels of biological age markers.In each regression analysis, the pregnancy week at birth was controlled as it may impact an individual's birth size.We controlled also other factors related to the markers of biological age such as chronological age, body adiposity, alcohol use, testosterone and cortisol level, and physical activity.The predictors for each analysis were selected based on the results of the t-test and Pearson correlation analysis for the relationship between biological age markers and potential confounders.Analyses were performed with Statistica 12.0 software.The results were interpreted as statistically significant if p < 0.05.

Table 2 .
The relationship between markers of biological age and controlled variables (N = 159).Significant values are in bold.a Composite measure of oxidative stress.b Composite measure of biological age.c Serum total antioxidant capacity.

Table 3 .
The relationship between measures of birth size and controlled variables (N = 159).Significant values are in bold.a Serum total antioxidant capacity.

Table 5 .
The results of regression analyses for the relationship between birth weight (Model 1) or ponderal index (Model 2) and s-Klotho level (log) controlled for cofounders (N = 159).Significant values are in bold.

Table 9 .
The results of multiple regression analyses for the relationship between biological age index and birth weight (Model 1) or ponderal index (Model 2) controlled for cofounders (N = 159).Significant values are in bold.