GDF-15 as a biomarker of aging

Background: The aging process is accompanied by the gradual development of chronic systemic inammation (inamm-aging). Growth differentiation factor-15 (GDF-15) is associated with inammation and known to be a stress-induced factor. The present study aimed to explore the association of GDF-15 with aging. Methods: In this cross-sectional study, serum GDF-15, hematological parameters, and biomedical parameters were determined in 120 healthy individuals (23-83 years old, males). Three telomere related parameters, including telomere length, telomerase activity, and the expression of human telomerase reverse transcriptase (hTERT) mRNA were also quantied. Results: The older group has a higher levels of GDF-15 and lower expression of hTERT mRNA, and PBMC telomerase activity (p<0.001). In individuals with high GDF-15 levels, they were older, and presented with the lower level of hTERT mRNA and T/S ratio (p<0.01). Spearman correlation analysis shows that GDF-15 positively correlated with age (r=0.664, p<0.001), and negatively correlated with telomere length (r=-0.434, p<0.001), telomerase activity (r=-0.231, p=0.012), and hTERT mRNA (r=-0.206, p=0.024). Furthermore, in multivariate regression analysis, GDF-15 levels showed a statistically signicant linear and negative relationship with PBMC telomerase activity (β-coecient=-0.583, 95% CI -1.044 to -0.122, p=0.014), telomere length (β-coecient=-0.200, 95% CI -0.305 to -0.094, p<0.001), and hTERT mRNA (β-coecient=-0.207, 95% CI -0.312 to -0.102, p<0.001) after adjusting for confounders. Conclusions: In conclusion, our results show that circulating GDF-15 is the potential biomarkers


Background
The growth in human life expectancy in the last century was substantial, resulting in adults aged 65 years or older represent 8.5% of the world population, and it is expected that the number of individuals aged 80 or over will represent 17% of the world population by 2050, which brings a great economic burden to the society [1]. Therefore, nding biomarkers of senescence that can be translated to the clinical setting are of particular interest.
Several aging biomarkers such as telomere length, telomerase activity, and the expression of human telomerase reverse transcriptase (hTERT) mRNA have been reported may in uence the risk and progression of multiple aging conditions [2]. Uncontrolled telomere shortening may trigger immune senescence, chromosomal degradation, and cellular dysfunction, which have been related to the occurrence of age related diseases, including cardiovascular disease (CVD) and neurodegeneration [3]. Telomere length is believed to be a vitally important biomarker of aging, and many studies show that reductions in telomere length might be the independent risk factor for CVD [4,5]. Telomerase is important to maintain the relative stability of telomere length. It consists of TERT, the protein component of telomerase, and RNA component. The expression of hTERT mRNA and telomerase activity decreased with age [6]. Recently, growing evidence shows that TERT is protective in the microcirculation against prolonged vascular stress [7,8]. Among the different biomarkers proposed, including cell-free DNA and circulating extracellular RNAs [9,10], growth differentiation factor (GDF-15) appears particularly promising due to the ease of collecting specimens and the low costs. However, little is known about how serum GDF-15 changes with age.
High levels of GDF-15 was rst reported in 1997 in macrophages and acted as an autocrine regulatory molecule and belonged to the transforming growth factor β (TGF-β) superfamily [11]. In the last two decades, GDF-15 has received lots of attention because of its multiple phenotypic functions, with roles in the regulation of obesity, cancer, nervous system disease, metabolism and CVD. Furthermore, serum GDF-15 levels are potential diagnostic markers for aging-related diseases, such as cognitive impairment, frailty, and CVD [12][13][14].
Here, to gain a comprehensive assessment of the association between GDF-15 and aging. We measured GDF-15 levels, and three telomere related parameters, including telomere length, telomerase activity, and the expression of hTERT mRNA in individuals composed of 53 young (23 to 39 years old), 39 middleaged (50-66 years old), 28 old (67-83 years old).

Study population
The research was approved by the Medical Ethics Review Committee of Renmin Hospital, Wuhan University, China. All the participants were asked to provide written informed consent in accordance with the Renmin Hospital of Wuhan University Ethics Committee. The study conforms to the principles outlined in the declaration of Helsinki.
We performed a cross-sectional study with a total of 120 individuals aged 23-83 years old were recruited from the physical examination center of Renmin Hospital of Wuhan University. Participants who are free of major chronic conditions and functional impairments are enrolled. In detail, 53 young (20-39 years old), 39 middle-aged (50-66 years old), 28 old (67-83 years old) individuals, and the age of the participants was de ned by birth certi cates stated at the time of recruitment [2]. Participants affected by infectious diseases, pulmonary edema, chronic renal function, or acute kidney injury, or those receiving thrombolysis treatment, or on immunosuppressive treatment were excluded from the study. Venipuncture was performed in the morning after the participants had fasted for at least 8 h.

Laboratory methods
Levels of the secretory form of human GDF-15 in crude serum samples were measured in duplicate using a commercial enzyme-linked immunosorbent assay (ELISA) kit (Quantikine, R&D Systems, USA) with intra-and inter-assay coe cient of variation < 6% and 2.8%, respectively. The assay detection range was 7.8-500 pg/mL, and samples were diluted 10 times for detection.

PBMCs isolation
Peripheral blood mononuclear cells were isolated from peripheral blood by centrifugation with Ficoll-Paque Plus (GE Healthcare). PBMCs were used for the measurement of telomerase activity and telomere length. All cells were then stored at −80°C until analysis Telomerase enzymatic activity assay Telomerase activity were assayed using a photometric ELISA based on the telomeric repeat ampli cation protocol (TRAP, Roche) as previously described [6]. Brie y, 2 × 10 5 PBMCs were isolated from each sample and stored at −80°C pending analysis. Samples containing telomerase are added to the 3 'end of the biotinylated synthetic primer P1-TS, then P1-TS were used to amplify these products by PCR. The hybrid product was xed to the microporous plate coated by biotin-labeled of streptomyces a nities primers, then peroxidase-linked digoxin antibodies were used to detect the PCR products. Finally, TMB is metabolized by peroxidase to form the colored products. Added stop reagent per well to stop color development and measured the absorbance within 30 min.
PBMCs telomere length assay Total DNA was isolated from PBMCs using the MiniBEST Universal Genomic DNA Extraction Kit 5.0 (Takara) according to the manufacturer's instructions. DNA samples were measured for concentration and purity using the NanoDrop (Thermo Fisher Scienti c). The quantitative PCR method was used to measure telomere length relative to standard reference DNA (T/S ratio), as described previously [15]. Primer sequences were as follows [6] and synthesized by Invitrogen: telg ACACTAAGGTTTGGGTTTGGGTTTGGGTTTGGGTTAGTGT, telc TGTTAGGTATCCCTATCCCTATCCCTATCCCTATCCCTAACA; albu CGGCGGCGGGCGGCGCGGGCTGGGCGGAAATGCTGCACAGAATCCTTG, albd GCCCGGCCCGCCGCGCCCGTCCCGCCGGAAAAGCATGGTCGCCTGTT.
Reverse transcriptase-quantitative PCR RNA was extracted by Trizol reagent (Takara) according to the manufacturer's instructions. RNA was quanti ed by Nanodrop SD-2000 spectrophotometer (PLCO) to measure the absorbance (A260 nm) and samples were then stored at -80°C until use. The cDNA was reverse-transcribed using a PrimeScript™ RT reagent kit with gDNA Eraser (Takara) in an ABI 9902 thermal cycler. The reactions were incubated in the thermal cycler for 2 min at 42 °C, 15 min at 37°C, and 5 sec at 85°C and then held at 4°C.
The cDNA was ampli ed by PCR using hTERT-speci c primer and GAPDH primer pairs. PCR was performed for 40 cycles 95℃ 30s (each cycle consisting of 95°C for 30s, 64°C for 90s, and 72°C for 30s). The PCR products were analyzed by electrophoresis on a 1% agarose. Primer sequences were as follows [6] and synthesized by Invitrogen: hTERT-For CGGAAGAGTGTCTGGAGCAA, Rev GGATGAAGCGGAGTCTGGA; GAPDH-For AGAAGGCTGGGGCTCATTTG, Rev AGGGGCCATCCACAGTCTTC.

Statistical analysis
Statistical analysis was performed using IBM SPSS software, version 20.0 (IBM, Armonl, NY, USA), GraphPad Prism 7.0 (GraphPad Software, La Jolla, CA, USA). Continuous variables were expressed as mean value ± SD or median, percent 25-percent 75, according to normality of distribution. Differences among groups were analyzed by the one-way ANOVA for Gaussian distributed data and the Kruskal-Wallis H test where at least one column was not normally distributed. The correlation analysis was made by Spearman coe cient. The associations among GDF-15 with T/S ratio, leukocyte hTERT mRNA levels, and PBMC telomerase activity were performed by univariate and multivariable linear regression, respectively. And we transformed GDF-15 levels and leukocyte hTERT mRNA to the log10 scale for linear regression analysis. A two-sided p value of < 0.05 was considered statistically signi cant.

Baseline characteristics
Demographic data for the cohort as related to chronological age are presented in Table 1. The difference in the red series is observed, presented in the lower level of hemoglobin and RBC in the older group (p<0.01). Also signi cant differences are observed in the number of the platelets and lymphocytes, the older group has fewer platelets and lymphocytes (p<0.05). And the levels of AST, TC, TG, TC/HDL-c, and glucose increased with age. On the contrary, levels of Alb decreased with age. Furthermore, the older group has a higher levels of GDF-15 and lower expression of hTERT mRNA, and PBMC telomerase activity. No signi cant differences in the distribution of the number of WBC, Neu, Mono, neutrophil to lymphocyte ratio, and levels of ALT, BUN, creatinine and LDL-c are observed. Table 2 shows the baseline characteristics as strati ed by GDF-15 tertiles. The medians of GDF-15 levels varied from 338.52 pg/mL in the bottom tertile up to 640.71 pg/mL in the top tertile. Patients with high GDF-15 levels are older, had higher number of blood monocytes, when compared to patients with low GDF-15 levels (p<0.05). Kim et al, showed that the frailty and frailty criterion were signi cantly associated with lower erythrocyte levels of long-chain n-3 PUFA [16]. Consistent with the previous study, in our study, the older group has less RBC. And the levels of AST and TC/HDL-c increased with GDF-15 levels. With regard to the numbers of WBC, Neu, LYM, neutrophil to lymphocyte ratio, platelets and levels of Hb, ALT, BUN, creatinine, TC, TG, HDL-c, LDL-c and glucose, no signi cant statistical difference is observed among different GDF-15 tertiles.

Telomere-related parameters across different GDF-15 tertiles.
Telomere length in PBMCs, relative levels of hTERT mRNA in leukocytes, and the activity of PBMC telomerase strati ed by GDF-15 tertiles are presented in Table 3. Signi cant difference in the telomererelated parameters were observed, presented in the lower level of hTERT mRNA and T/S ratio in individuals with high GDF-15 levels (p<0.01). Moreover, the measurements of PBMC telomerase activity showed a consistent direction of effect over GDF-15 groups. Patients with high GDF-15 levels had a lower PBMC telomerase activity.

Univariate and multivariate linear regression
To further explore the association of GDF-15 with T/S ratio, leukocyte hTERT mRNA levels, and PBMC telomerase activity, we performed multivariable linear regression, respectively. Table 4 displays the results of the univariate and multivariate regression analysis of PBMC telomerase activity and GDF-15. In the univariate linear regression analysis, PBMC telomerase activity positively associated with numbers of LYM, RBC, PLT, and levels of Alb, glucose, TC. In addition, PBMC telomerase activity negatively associated with GDF-15 levels. In the multivariate linear regression analysis, PBMC telomerase activity showed a statistically signi cant linear and negative relationship with GDF-15 levels (β-coe cient=-0.583, 95% CI -1.044 to -0.122, p=0.014) after adjusting for Neu, RBC, Alb and TC. Table 5 shows that telomere length is negatively associated with GDF-15 levels (β-coe cient=-0.200, 95% CI -0.305 to -0.094, p<0.001) after adjusting for glucose, TG, and HDL-c.
The association between hTERT mRNA and GDF-15 is also further evaluated by multivariate regression analysis. As summarized in Table 6, the expression of hTERT mRNA showed a statistically signi cant linear and negative relationship with GDF-15 levels (β-coe cient=-0.207, 95% CI -0.312 to -0.102, p<0.001) after adjusting for glucose.

Discussion
The current study explored the associations between serum GDF-15 and aging. Our result shows the older group has a higher levels of GDF-15, lower hTERT mRNA expression, and PBMC telomerase activity. The comparision of baseline characteristics among different GDF-15 tertiles shows a difference in the number of monocyte and erythrocyte, where the high GDF-15 group presents a lower number of monocyte and erythrocyte. These results coincide with those previously observed, metabolic alterations that occur in monocyte and erythrocyte during aging [17,18]. In individuals with high GDF-15 tertile, they displayed high levels of AST, and TC/HDL-c. This is in accordance with Tang et al, which demonstrates that blood lipid pro les have been ambiguously reported to be associated with aging-related disease. According to the Spearman correlation method, GDF-15 levels presented positively correlated with age, and negatively correlated with PBMC telomerase activity, telomere length, and WBC hTERT mRNA expression. Moreover, in the multivariate linear regression, statistically signi cant linear and negative relationship of GDF-15 with PBMC telomerase activity, telomere length, and hTERT mRNA expression were observed.
GDF-15 is stress-induced factor, the biological roles of GDF-15 are context-dependent and may vary with the stage of the disease [19][20][21]. In a systematic research, aimed to identify genes regulated in aging, and age-related diseases, GDF-15 is one of the high priority candidates [13]. In human endothelial cells, GDF-15 promotes radiation-induced senescence through the ROS-mediated p16 pathway, and contribute to the development of atherosclerosis [22]. GDF-15 levels are upregulated under the stress of ischemiareperfusion, and it is important for the generation of reactive oxygen species and the development of senescence [23]. Furthermore, Talia et al. showed GDF-15 was associated with aging related impairment and changes, and predicted to be the potential biomarker of cognitive decline [24]. Consistent with previous studies, the olders displayed higher levels of GDF-15. And in high GDF-15 tertile group, individuals displayed high levels of AST, and TC/HDL-c. These biomedical parameters are correlated with aging-related disease, such as CVD.
The progression of aging is accompanied by the changes of hematological parameters. RBCs are the most abundant cell in our body, constitute approximately 83% of the total host cells [25]. During their 120day lifespan, RBCs transport molecules of oxygen to all parts of the body [26]. When RBCs are used as the model to explore system metabolism in the context of cellular oxidant stress, presented that aged individuals are associated with a stressed erythropoiesis phenotype and contributed to the so-called "anemia in the elderly" [27]. Furthermore, aging promotes erythrocyte phagocytosis and contribute to the development of atherothrombosis [28]. Also, Pluta, et al demonstrated the effects of erythrocytes as the biomarker of Alzheimer's Disease [29]. Monocytes probably account for about 10% of all circulating leukocytes [30]. Monocytes play vital role in antigen presentation, tissue repair, in ammatory processes, and affect many age-related health situations, including atherosclerosis, and Alzheimer's disease [31,32]. In patients with frailty, a higher number of monocytes is associated with frailty [33]. And monocytes present age-related changes [34,35]. In the current study, coincide with previous studies, our results showed the olders displayed lower number of erythrocytes. And in individuals with high GDF-15 tertile, they presented lower RBCs, and higher number of monocytes. According to the Spearman correlation method, GDF-15 levels showed negatively correlated with RBC, positively correlated with monocytes.
A meta-analysis of a case-control study revealed increased TC levels and decreased HDL-c levels are observed to be associated with an elevated risk of aging-related, such as Alzheimer's disease [36]. In germ-free mice, Albouery, et al proposed brain lipid composition changes with age, and microbiota alterations may be the regulator [37]. Our results are in line with these studies showing in individuals with high GDF-15 tertile, they displayed higher TC/HDL-c, and a statistically signi cant positive association was found between GDF-15 and TC/HDL-c.
Telomere length, and telomerase activity decline with age and are now considered to be the potential markers re ecting physical age that correlate with various age-related pathological changes including Alzheimer's disease, and cancers [38,39]. Enzo, et al identi ed that telomere length and telomerase activity in T cells were the biomarkers of high-performing centenarians [2]. Consistent with previous studies, declined PBMC telomerase activity, and hTERT mRNA were observed in the older group. In the high GDF-15 tertile group, telomere length, telomerase activity, and hTERT mRNA were all signi cantly declined (p<0.05). And elevated GDF-15 levels presented negatively correlated with decreased PBMC telomerase activity, telomere length, and WBC hTERT mRNA expression. Furthermore, the results of multivariate linear regression analysis showed GDF-15 negatively associated with telomerase length, leukocyte hTERT mRNA levels, and PBMC telomerase activity.
In the study of GDF-15 as a biomarker of aging, we also take the regulatory effect of GDF-15 on in ammation into consideration. Our study suggests circulating GDF-15 not only as a biomarker of age also distinguish person from health state.

Strengths And Limitations
Circulating GDF-15 levels displayed positively correlated with age, and statistically signi cant linear and negative relationship with PBMC telomerase activity, telomere length, and hTERT mRNA expression were observed. In addition, to evaluate aging by GDF-15 is more convenient, the detection of GDF-15 levels is more easier than the assays of telomere related parameters. And GDF-15 is helpful to evaluate risk factors of developing aging related diseases. This study suffers from some limitations, due to the small number of subjects in the subgroups, making it limited for us to try our power to transform our ndings into a real predictor. Larger cohorts are needed to minish inter-individual difference, as well as any potential association with nutritional differences. Moreover, further work is needed to elucidate the mechanism of GDF-15 on aging, both in vitro and in vivo.

Conclusion
In male individuals with different ages, the olders has higher levels of GDF-15. And GDF-15 levels were negatively associated with telomere related parameters, after adjustment for confounders. Moreover, as an important regulator in in ammation, GDF-15 showed positive correlation with monocytes, and "bad lipids", including LDL-c and TC/HDL-c. Hence, our study suggests circulating GDF-15 levels could be used as a biomarker of age and helpful to evaluate risk fators for the development of aging related diseases.
Abbreviations WBC white blood cell, Neu Neutrophil, LYM lymphocyte, Mono mononuclear cell, N/L ratio neutrophil to lymphocyte ratio, RBC red blood cell, Hb hemoglobin, ALT alanine aminotransferase, AST aspartate aminotransferase, Alb albumin, BUN blood urine nitrogen, TC total cholesterol, TG triglyceride, HDL-c highdensity lipoprotein cholesterol, LDL-c low-density lipoprotein cholesterol, T/S ratio telomere length relative to standard reference DNA, GDF-15 growth differentiation factor-15. Data are presented as the mean value ± SD or median, percent 25-percent 75. Differences among groups were analyzed by the one-way ANOVA and Kruskal-Wallis H test.