Genes related to maintenance of autophagy and successful aging

ABSTRACT Considering aging as a phenomenon in which there is a decline in essential processes for cell survival, we investigated the autophagic and proteasome pathways in three different groups: young, older and oldest old male adults. The expression profile of autophagic pathway-related genes was carried out in peripheral blood, and the proteasome quantification was performed in plasma. No significant changes were found in plasma proteasome concentrations or in correlations between proteasome concentrations and ages. However, some autophagy- and/or apoptosis-related genes were differentially expressed. In addition, the network and enrichment analysis showed an interaction between four of the five differentially expressed genes and an association of these genes with the transcriptional process. Considering that the oldest old individuals maintained both the expression of genes linked to the autophagic machinery, and the proteasome levels, when compared with the older group, we concluded that these factors could be considered crucial for successful aging.

such as cancer 13,14 , and suggested a correlation between proteasome concentrations and health status 10,15,16 .
Several diseases related to aging show accumulation of oxidized proteins and the failure of autophagic pathways is suggested as a possible cause 17 . The autophagy-lysosome pathway is a cytoplasmic-restricted degradation system, related to the degradation of organelles, proteins and protein aggregates 18 . Under adequate levels of nutrients, growth factors and reactive oxygen species, autophagy is at basal levels (constitutive) with normal protein biosynthesis. However, autophagy can be induced by a stressor and the production of proteins interrupted 19 .
As the molecular characterization of the autophagic machinery may allow the development of tools for a better physiological and molecular evaluation of successful aging, our objective was to quantify the expression of genes involved in autophagic machinery regulation in young, older and oldest old adult male individuals.

Volunteers
The individuals selected for this study were previously recruited by the Department of Preventive Medicine and Discipline of Geriatrics and Gerontology of the Universidade Federal de São Paulo for a different study but the samples were not fully used. All volunteers signed a free and informed consent form. For the current study, they signed an authorization for sample use, following norms determined by the Research Ethics Committee of the Universidade Federal de São Paulo, which approved the study (# 451631/2013). The sample consisted of male volunteers, distributed into three groups: individuals aged between 20 and 30 years (young group, n = 15), individuals aged between 60 and 70 years (older group, n = 13) and individuals between 85 and 105 years old (oldest old group, n = 10). Individuals with neoplasias or severe unmanaged diseases, such as heart diseases, gastrointestinal diseases, type 2 diabetes, or with neurological and psychiatric antecedents were excluded.

Samples previously collected
Peripheral blood was collected in EDTA tubes, centrifuged at 3,000 rpm for 10 minutes and the separated plasma stored at -20°C. In addition, 5 mL of blood was collected in specific tubes (PaxGene RNA collection tubes -PreAnalytiX, Switzerland) for total RNA extraction using the PaxGene kit (PaxGene blood RNA isolation kit -PreAnalytiX, Switzerland). After verification of integrity and purity, total RNA was stored at -80°C.

Proteasome
To perform proteasome quantification in plasma, we used enzyme-linked immunosorbent assay -Proteasome ELISA Kit (Enzo Life Sciences, BML-PW0575, EUA), which employs specific antibodies for the 20S proteasome subunit. The product absorbance was detected using the SpectraMax M2 apparatus (Molecular Devices, USA).

Gene expression
Total RNA was quantified using the NanoDrop 8000 (Thermo Scientific, USA). For complementary DNA synthesis, we used the RT2 First Strand Kit (QIAGEN, Germany) plus 625 ng of RNA. Cycling parameters comprised a holding stage at 42ºC for 15 minutes, followed by inactivation at 95°C for 15 minutes. The expression profile of 84 autophagic pathway-related genes was analyzed in peripheral blood RNA samples using the Superarray-RT2 Profiler™ PCR Array System (QIAGEN, Germany -PAHS-084ZD-24) in the 7500 PCR Real-Time System (Applied Biosystems, USA). In addition, ACTB, B2M, GAPDH, HPRT1 and RPLP0 genes were evaluated as an endogenous control. Thermal cycling conditions comprised an initial denaturation at 95°C for 15 seconds and annealing and extension at 60°C for one minute (Table 1).
Gene expression quantification was obtained using ΔCT calculation, and the endogenous control CT was obtained from the arithmetic mean of two endogenous controls that showed a lower variation between groups (standard deviation > 0.1). Then, the relative gene expression was calculated by CT comparative method (ΔΔCT) using the following formula: FC = 2-ΔΔCT = 2-(ΔCT interest group -ΔCT reference group). For better visualization of variation, data were presented by fold regulation (FR) (if FC was greater than 1, FR = FC, if the HR was less than 1, the FR = -(1/FC)), which represents the number of times a gene is expressed in one group in relation to the other. Both the older and oldest old groups were compared with the young group (reference group).

Network and enrichment analysis
To investigate interactions and pathways shared by differentially expressed genes, two online software applications were used: GeneMANIA 20 (www.genemania.org]) and Enrichr 21 (http://amp.pharm.mssm.edu/Enrichr). GeneMANIA allows genes with shared or functionally similar properties to be identified. These analyses may be performed with genes of interest or with 20, 50 or 100 other genes in the interaction. In the present study, we used 20 genes. Enrichr enables enrichment analysis, where genes of interest are searched and compared in databases to verify possible pathways and over-represented cellular processes in which they may participate.

Statistical analysis
The normality of data was analyzed with Shapiro-Wilk' s test and, when necessary, normalized using the Z-score. Data was compared using one-way analysis of variance (ANOVA). We used the Pearson' s correlation test to compare age and proteasome levels in each age group. Data are presented as mean ± standard error. The level of significance was set at p ≤ 0.05. However, the p value was corrected by the Benjamini-Hochberg or less than -1.50 (genes with decreased expression) were used to select differentially expressed genes, and to exclude those potentially subject to methodological noise. Thus, the differentially expressed genes were those included in one of the following conditions: 1) pBH ≤ 0.05, independent of the FR value or 2) p ≤ 0.05 and FR ≥ 1.50 or FR ≤ -1.50.

RESULTS
Autophagic pathway gene expression and proteasome levels were evaluated in the individuals from three different age groups (mean ± standard deviation): young, 24.3 ± 2.2 years: older, 65.5 ± 3.0 years; and oldest old, 91.9 ± 6.1 years (F = 0.619 and p = 0.545; Figure 1). Additionally, plasma proteasome levels were not related to the individuals' ages in each group ( Figure 2). However, when the oldest individual (105 years) was excluded from the oldest old group analysis, a statistically significant correlation was observed ( Figure 2D).   Regarding gene expression, from the 84 genes linked to autophagic machinery, only five were differentially expressed according to the adopted criteria: ATG4C, BCL2L1, EIF2AK3, EIF4G1 and TP53 ( Table 2). The ATG4C gene was significantly less expressed in the oldest old group when compared with the young group (1.91-fold decrease); in addition, there was also a difference in the older group when compared with the oldest old (1.47-fold increase; p = 0.031). The BCL2L1 gene was significantly more expressed in the oldest old when compared with the young group (increase of 1.91 times). The EIF2AK3 gene was significantly less expressed in the older group (1.46fold decrease), as well as in oldest old individuals when compared with the young group (1.44-fold decrease). The EIF4G1 gene was significantly less expressed in the older and oldest old when compared with the young group (decrease of 1.47 and 1.32 times, respectively). The TP53 gene was significantly less expressed in the older and oldest old when compared with the young group (decrease of 1.57 and 1.66-fold, respectively).
In the network analysis, we observed that from the five differentially expressed genes, only two showed evidence of some interaction -TP53 and BCL2L1 (Figure 3). When the other 20 genes were added, we observed that four of the five genes showed some type of interaction, the exception being ATG4C (Figure 4). The enrichment analysis was divided     into two stages: the first was done with the five differentially expressed genes, and the second with the differentially expressed genes plus the genes that showed the most frequent pathways in the network analysis (HSPA5, SIN3A and EIF2S1). In the first step, the following databases were used: TRANSFAC and JASPAR PWMs, ENCODE TF ChIP-seq 2015, ESCAPE, ENCODE TF ChIP-seq and GO Biological Process 2013. The databases used in the second stage were: ChEA, TRANSFAC and JASPAR PWMs, ENCODE TF ChIP-seq 2015, transcription factor PPIs, ESCAPE, ENCODE TF ChIP-seq. All databases used in the first and second stages showed direct or indirect linkage of the genes analyzed with the transcription process.

DISCUSSION
The accumulation of macromolecules and damaged organelles is one of the most predominant alterations found in aged cells, and the main cause is related to a deficient autophagic process 22 . Studies in C. elegans and D. melanogaster have shown that the loss of function of autophagy genes is related to an accumulation of damaged organelles and proteins, accelerated aging and shortened life span 23,24,25 . To evaluate the contribution of the autophagic machinery in successful aging, we quantified the expression of 84 genes related to the autophagic pathway in young, older and oldest old individuals; five presented with differential expression between the studied groups: ATG4C, BCL2L1, TP53, EIF2AK3 and EIF4G1.
The ATG4C encodes a protein with protease activity involved in autophagic vacuole formation. However, studies suggest that this protein is not essential to generate the basal level of autophagy required, since knockout mice for the ATG4C gene exhibit normal development 24 . In contrast, knockouts for this gene are more likely to develop fibrosarcoma when exposed to carcinogenic chemicals compared with wild-type animals 26 . The lower expression of ATG4C observed in the oldest old people group does not suggest lower autophagic activity per se, but may contribute to a higher risk of these individuals developing tumors, a condition that could be related to aging. On the other hand, the increased expression of BCL2L1 observed in the older and oldest old groups indicates that autophagy levels decrease during aging 27 . The BCL2L1 is a co-regulator of autophagy and apoptosis and proteins from the BCL-2 family may also interact with p53 in the induction of autophagy. P53 exhibits tumor suppressor activity and the ability to control autophagic processes and cellular senescence 28,29 . In the current study, there was decreased TP53 expression in both the oldest old and older groups in relation to the young group, suggesting that the autophagic process decreases with increased age. In addition, decreased expression of EIF2AK3 and EIF4G1 in both the oldest old and older individuals reflects the body' s declining ability to maintain reticulum homeostasis and cellular processes with increasing age 19,30 .
The EIF2AK3 and EIF4G1 proteins, respectively, are associated with endoplasmic reticulum homeostasis and the initiation of translation of mRNAs related to mitochondrial activity and cellular bioenergetics 19,30 .
Studies have suggested that proteasome activity declines during cellular senescence and aging in both animal models and humans 31 . However, a study performed by Chondrogianni and colleagues showed similar functional proteasomes in human fibroblasts cultures from centenarian and young donors 8 . In the current study, we evaluated, for the first time, the plasmatic proteasome levels in the young, older and oldest old groups and we did not observe a significant difference between them. Although there is no evidence that plasmatic proteasome concentrations reflect the intracellular proteasome activity, we hypothesized that the similarity of plasma proteasome concentrations between the groups found in our samples could be one of the factors contributing to the longevity in the oldest old group. In fact, we previously observed that these same oldest old individuals had a more favorable lipid profile compared with the other groups 32 . An increase in SIRT2 expression in the oldest old people was also observed when compared with the young group (unpublished data). The increase in SIRT2 seems to contribute to the promotion of longevity by increasing levels of autophagy 33 . More recently, several studies have shown the impact of caloric restriction on sirtuin levels, which in turn act on autophagic pathways and contribute to increased life expectancy 34 . In the network analysis of differentially expressed genes, we identified interactions between the TP53 and BCL2L1 genes, which was expected, as several studies have shown the promotion of autophagy by the interaction of TP53 with the Bcl-2 family proteins [35][36][37] . However, when we added 20 other genes to this network, four of the five differentially expressed genes had some type of interaction, with ATG4C being the exception (Figure 4). The interaction between the four genes is related to the regulation of transcription, an extremely important process for cell functioning 38 . During the aging process, some genes have increased expression, such as those related to cell adhesion and immune response 39 , while others have decreased expression, such as genes that participate in lipid metabolism 39 and those involved in the electron transport chain 40,41 .
In conclusion, the ATG4C, BCL2L1, TP53, EIF2AK3 and EIF4G1 genes differed preferentially when comparing the oldest old and older with the young group, suggesting that autophagy and some processes like maintenance of metabolism and control of gene expression are impaired when the individual ages. On the other hand, the similarity in the expression pattern observed between the older and oldest old suggests that the maintenance of these pathways related to homeostasis plays an important role in increasing life expectancy. In general, these findings, together with the maintenance of proteasome levels observed in the oldest old individuals, point to the maintenance of autophagy as a crucial factor for longevity.