• Log in with Facebook Log in with Twitter Log In with Google      Sign In    
  • Create Account
  LongeCity
              Advocacy & Research for Unlimited Lifespans

Photo

Growth differentiation factor 15 protects against the aging‐mediated systemic inflammatory response in humans and mice

aging inflammation mitochondria senescence t cell

  • Please log in to reply
No replies to this topic

#1 Engadin

  • Guest
  • 198 posts
  • 580
  • Location:Madrid
  • NO

Posted 31 July 2020 - 04:15 PM


.

 

 

 

 

 

 

O P E N   A C C E S S   S O U R C E :   Aging Cell

 

 

 

 

 

 

 

Abstract
 
Mitochondrial dysfunction is associated with aging‐mediated inflammatory responses, leading to metabolic deterioration, development of insulin resistance, and type 2 diabetes. Growth differentiation factor 15 (GDF15) is an important mitokine generated in response to mitochondrial stress and dysfunction; however, the implications of GDF15 to the aging process are poorly understood in mammals. In this study, we identified a link between mitochondrial stress‐induced GDF15 production and protection from tissue inflammation on aging in humans and mice. We observed an increase in serum levels and hepatic expression of GDF15 as well as pro‐inflammatory cytokines in elderly subjects. Circulating levels of cell‐free mitochondrial DNA were significantly higher in elderly subjects with elevated serum levels of GDF15. In the BXD mouse reference population, mice with metabolic impairments and shorter survival were found to exhibit higher hepatic Gdf15 expression. Mendelian randomization links reduced GDF15 expression in human blood to increased body weight and inflammation. GDF15 deficiency promotes tissue inflammation by increasing the activation of resident immune cells in metabolic organs, such as in the liver and adipose tissues of 20‐month‐old mice. Aging also results in more severe liver injury and hepatic fat deposition in Gdf15‐deficient mice. Although GDF15 is not required for Th17 cell differentiation and IL‐17 production in Th17 cells, GDF15 contributes to regulatory T‐cell‐mediated suppression of conventional T‐cell activation and inflammatory cytokines. Taken together, these data reveal that GDF15 is indispensable for attenuating aging‐mediated local and systemic inflammation, thereby maintaining glucose homeostasis and insulin sensitivity in humans and mice.
 
 
 
1 INTRODUCTION
 
Aging is a major risk factor for various chronic diseases, including type 2 diabetes, neurodegenerative diseases, and malignancies, which are closely related to systemic subclinical inflammation in the absence of overt infections in the elderly (Ortega Martinez de Victoria et al., 2009; Meda et al., 1995; Multhoff, Molls, & Radons, 2011; Yi et al., 2019). Such systemic inflammatory responses are also termed “metaflammation” or “inflammaging” in humans (Franceschi et al., 2007; Hotamisligil, 2017; Sanada et al., 2018). Given that mitochondrial damage contributes to various senescent processes with a distinct pro‐inflammatory secretory phenotype (Wiley et al., 2016), the fact that the disruption of mitochondrial function is linked to age‐related pathologies is not surprising (Lane, Hilsabeck, & Rea, 2015).
 
Progressive mitochondrial dysfunction occurs across species during the aging process (Yi, Chang, & Shong, 2018). Oxidative damage to cellular macromolecules, or stress arising from mitochondrial DNA (mtDNA) mutation and increased reactive oxygen species (ROS), is a key hallmark of aging physiology (Yi et al., 2018). Although higher levels of ROS induced by mitochondrial stress are involved in cellular damage and the inflammatory response, they also provide the first line of host defense (Pellegrino et al., 2014). Paradoxically, elevated ROS levels increase the lifespan of worms, flies, and mice through an adaptive response (Yi et al., 2018), termed mitohormesis. The secretion of mitokines during cellular stress is a critical response that may reflect disease severity acting as disease markers. They may also regulate disease progression, which makes them a potential therapeutic target for chronic diseases caused by mitochondrial dysfunction (Yi et al., 2018).
 
Growth differentiation factor 15 (GDF15) is a well‐known mitokine that is induced by defects in mitochondrial oxidative phosphorylation or by the unfolded protein response (UPRmt) pathway in mammals (Chung, Ryu, et al., 2017; Khan et al., 2017). GDF15 production is regulated by the mTORC1 kinase through an integrated mitochondrial stress response in patients with mitochondrial myopathy (Khan et al., 2017). Plasma GDF15 levels increase during metabolic stress‐mediated tissue inflammation, including in insulin resistance and type 2 diabetes (Kempf et al., 2012; Yi, 2019). Skeletal muscle‐specific UPRmt is also related to the promotion of lipolysis and fatty acid oxidation in adipose tissues by GDF15 production, thereby protecting the organism against high fat, diet‐induced obesity, and insulin resistance (Chung, Ryu, et al., 2017). Additionally, GDF15 improves alcohol‐ or chemically induced chronic liver injury by suppressing the infiltration of neutrophils, monocytes, and activated T cells in the liver (Chung, Kim, et al., 2017). Although the fibroblast growth factor 21 (FGF21) maintains peripheral T‐cell homeostasis by attenuating thymic immune senescence with age (Youm, Horvath, Mangelsdorf, Kliewer, & Dixit, 2016), the immunometabolic role of GDF15 in the aging process is incompletely understood.
 
In this study, we provide evidence that GDF15 exerts a protective effect on tissue inflammation in metabolic organs, such as the liver and adipose tissues, in humans and mice. Through complementary human and animal experiments supported by the reanalysis of large‐scale human transcript datasets, we demonstrate that GDF15 is required for the prevention of aging‐induced development of metabolic diseases by regulating tissue and systemic inflammation. Taken together, the immune regulatory role of GDF15 reveals the dynamic interplay between the metabolic and immune systems and contributes to delays in aging‐induced systemic inflammation.
 
 
 
2 RESULTS
 
2.1 Elderly subjects exhibit higher levels of serum GDF15 and hepatic GDF15 expression
 
In an experiment on 12 male C57BL/6 WT mice, we noticed that serum GDF15 levels were elevated in old (20‐month‐old) mice compared to young (8‐week‐old) mice (Figure S1a). Subsequently, we recruited 70 participants of which the demographics and baseline characteristics are summarized in Table S1. Again, we observed a significant positive association between the age of the study subjects and serum levels of GDF15 (Figure 1a). Elderly subjects (≥60 years) also exhibited significantly higher levels of serum GDF15, compared with younger subjects (≤40 years) (Figure 1b). We further noticed that hepatic Gdf15 expression was higher in old mice (20‐month‐old) compared to young mice (8‐week‐old) (Figure S1b). Likewise, hepatic GDF15 expression was remarkably increased in elderly subjects compared with young people (Figure 1c). We confirmed this age‐related increase in hepatic GDF15 expression in two independent large human transcript datasets: (1) a liver microarray dataset (Innocenti et al., 2011) (Figure 1d) and (2) the RNA‐Seq data of the human Genotype‐Tissue Expression (GTEx) project (Consortium, 2015) (Figure 1e). In both datasets, GDF15 expression decreases in very young subjects (up to 30 years old), remains constant between 30 and 50 years of age, and then increases again after 50 years old. These non‐linear age effects are significant in both the microarray dataset (limma analysis, p  = 4 × 10−5) and the GTEx RNA‐Seq dataset (edgeR‐zingeR analysis, p  = 0.04). The former would even remain significant in a genome‐wide screen (q‐value =0.001). Average GDF15 expression is 65% higher in 60‐ to 81‐year‐old subjects as compared to 20‐ to 40‐year‐old subjects (corrected for gender and ancestry, p  = 0.06) (Innocenti et al., 2011). Gdf15 is also highly expressed in murine livers compared to other tissues (Figure S1c). If we equate 6‐months‐old mice to 30‐year‐old humans and 14‐month‐old mice to 50‐year‐old humans (Fox, 2007), this trend can also be observed in C57BL/6 JN mice (Figure S1d) (Tabula Muris et al., 2018). The lower Gdf15 expression in very old mice (27 months old) might be due to survival bias as only ~50% of mice reach this age.
 
 
acel13195-fig-0001-m.jpg
 
 
Figure 1. GDF15 correlates positively with aging‐induced systemic inflammation in humans.
(a) Correlation analysis of serum GDF15 levels in human subjects. (b) Serum levels of GDF15 in young (≤ 40; n = 14) and elderly (≥60; n = 24) subjects. © Hepatic expression of GDF15 in young (≤ 40; n = 8) and elderly (≥60; n = 8) subjects. (d,e) The effect of age on hepatic GDF15 expression in (d) a microarray dataset showing patient‐averaged hepatic log2‐transformed GDF15 intensities for 202 patients (Innocenti et al., 2011), and (e) a GTEx RNA‐Seq dataset with log2‐transformed GDF15 expression in transcripts per million (TPM) for 226 liver biopsies. Men are denoted as black circles, women as red triangles. The blue trend lines are obtained by fitting regression models with linear and quadratic age effects to the data. The transparent blue bands denote the 95% confidence intervals corresponding to these models. (f) Serum levels of TNF in young (≤ 40; n = 14) and elderly (≥60; n = 24) subjects. (g) Quantitation of mtDNA levels in ccf‐DNA from plasma in study participants. (h) Serum levels of GDF15 in subjects with the 20% lowest (bottom; n = 14; mean age, 46.4 years old) or 20% highest (top; n = 14; mean age, 65.5 years old) plasma levels of ccf‐mtDNA copy numbers. Data are expressed as mean ± SEM. *p  < 0.05, **p  < 0.01 ((a): simple linear regression, (b,c), (f–h): two‐tailed t‐tests)
 
 
 
2.2 GDF15 is linked to inflammation and mitochondrial stress
 
We also found that elderly subjects exhibited higher levels of serum TNF and increased hepatic TNF expression (Figure 1f and Figure S1e). The subjects with a higher level (top 25%) of serum GDF15 exhibited significantly elevated hepatic TNF expression compared with those with lower levels (bottom 25%) of serum GDF15 (Figure S1f). Old subjects also showed a reduction in the frequency of naïve (CD45RA+CD45RO−) CD4+ and CD8+ T cells and an increase in the frequency of memory (CD45RA−CD45RO+) CD4+ and CD8+ T cells in the peripheral blood (Figure S1g–i). Moreover, the absolute numbers of naïve CD4+ and CD8+ T cells were reduced in older subjects, but the number of memory CD8+ T cells was increased (Figure S1j). In addition, the population of memory CD8+ T cells showed a positive correlation with serum levels of GDF15 (Figure S1k). The production of granzyme B in senescent CD4+ and CD8+ T cells was also higher in elderly subjects, compared with young subjects (Figure S1l–o).
 
Mitochondrial damage is closely linked with the systemic inflammatory response in human diseases, and circulating mitochondrial DNA is also associated with inflammation during aging (Picca et al., 2018). GDF15 is a well‐known mitokine that is secreted during mitochondrial stress and damage. Thus, we also measured mitochondria DNA (mtDNA) levels by quantitating the copy number of mitochondria DNA in circulating‐cell‐free (ccf) DNA in the plasma of young and elderly subjects. Although there is a high variability in circulating mtDNA in the elderly population which may be explained by individual differences in mitochondrial dysfunction, the mtDNA levels were significantly higher in elderly subjects compared with the younger controls (Figure 1g). Subjects with higher levels (top 20%) of ccf‐mtDNA copy number in plasma exhibited significantly higher levels of serum GDF15 compared with those with lower levels (bottom 20%) (Figure 1h). In addition, GDF15 levels were positively correlated with ccf‐mtDNA copy number in plasma (Figure S2a). We also found that tissue‐specific deficiency of CR6‐interacting factor‐1 (Crif1)‐induced dysfunction of mitochondrial oxidative phosphorylation exhibited an increase in serum levels of GDF15 in mice at 8 weeks of age compared with controls (Figure S2b) (Choi et al., 2020; Chung, Ryu, et al., 2017). Collectively, our data show that increases in both hepatic GDF15 expression and serum levels of GDF15 are associated with aging‐related inflammation and mitochondrial damage.
 
 
 
2.3 Analysis of transcriptome datasets from the Genotype‐Tissue Expression (GTEx) project
 
To further investigate the relationship between GDF15 and inflammatory response at the transcriptome level, we utilized GTEx RNA‐Seq data from the liver, adipose tissue, and skeletal muscle to observe whether GDF15 expression is associated with systemic inflammation in humans. Differential expression gene analysis (DEA) was performed by dividing the data into two groups (top 25% and bottom 25% group) based on GDF15 expression levels. First, DEA was performed in the liver (Figure 2a). The –log10(q‐value) for GDF15 was equal to 191.7, confirming that each group was well‐differentiated by the expression of GDF15 (Figure S3a). The DEA results indicated that 6,314 up‐regulated and 6307 down‐regulated genes differed between the GDF15 top 25% group and the bottom 25% group (Figure 2b). Next, pathway analysis using the Kyoto Encyclopedia of Genes and Genomes (KEGG) was performed to define pathways that differ between groups in terms of GDF15 expression. Of the total 299 available pathways, the top 25% group with the highest GDF15 levels exhibited 240 up‐regulated pathways and 4 down‐regulated pathways compared with the bottom 25% group. We found that mTOR signaling pathway and mitochondria‐related pathways were up‐regulated in the GDF15 top 25% group compared to the GDF15 bottom 25% group. We also observed an increase in inflammation‐related pathways, including the TNF signaling and IL‐17 signaling pathways. On the other hand, the pathways related valine, leucine and isoleucine degradation, and fatty acid degradation were significantly down‐regulated (Figure 2c). A correlation analysis on the GTEx liver expression data revealed that the expression levels of AP1 , TNFAIP3 , NOD2 , and CD44 , which play important roles in inflammation including TNF signaling, showed significant positive correlation with GDF15 expression (Figure 2d).
 
 
acel13195-fig-0002-m.jpg
 
 
Figure 2. Analysis of related pathways regulated by GDF15 expression in liver and adipose tissue in the GTEx database.
(a) Distribution of 226 hepatic GDF15 expression levels (log2(TPM + 0.001)) for human subjects in GTEx. The red and blue boxes represent the top 25% (n = 57) and the bottom 25% (n = 57) of the group according to GDF15 expression levels, respectively. (b) Number of DEGs between the top 25% and the bottom 25% according to GDF15 levels. The red, blue, and gray represent the number of up‐regulated, down‐regulated, and un‐regulated genes, respectively. © KEGG pathway analysis of the DEA results. The red, blue, and gray boxes indicate up‐regulated, down‐regulated, and un‐regulated pathways, respectively. Bar plots representing the up‐regulated (red) and down‐regulated (blue) pathways for significantly enriched pathways. The pathways shown in these bar plots were selected from the significant pathways (FDR <0.1) in the KEGG analysis. (d) Hepatic GDF15 expression correlates positively with AP1 , TNFAIP3 , NOD2 , and CD44 gene expression. The correlation analysis was conducted by GEPIA2 in the GTEx liver dataset (R: Pearson's correlation coefficient). (e) Distribution of 633 subcutaneous adipose tissue GDF15 expression levels (log2(TPM + 0.001)) for human subjects in GTEx. The red and blue boxes represent the top 25% (n = 158) and bottom 25% (n = 158) groups according to GDF15 levels, respectively. (f) Number of DEGs between the top 25% and bottom 25% groups according to GDF15 levels. The red, blue, and gray boxes represent the number of up‐regulated, down‐regulated, and un‐regulated genes, respectively. (g) KEGG pathway analysis of the DEA results. The red, blue, and gray boxes indicate up‐regulated, down‐regulated, and un‐regulated pathways, respectively. Bar plots represent the up‐regulated (red) and down‐regulated (blue) pathways for significantly enriched pathways. The pathways shown in these bar plots were selected from the significant pathways (FDR <0.1) in the KEGG analysis. (h) Adipose tissue GDF15 expression correlates positively with AP1 , TNFAIP3 , NOD2 , IFNG , CD44 , CD11B , and CCL2 gene expression. The correlation analysis was conducted by GEPIA2 in the GTEx subcutaneous adipose tissue dataset (R: Pearson's correlation coefficient)
 
 
 
The role of GDF15 in subcutaneous adipose tissue was analyzed next. As in the liver, groups were divided into top 25% and bottom 25% groups, according to GDF15 expression levels (Figure 2e). The q‐value for GDF15 was approximately 0, indicating that each group was well‐divided in terms of GDF15 expression (Figure S3b). For the DEA results, the top 25% group contained 9841 up‐regulated genes and 7588 down‐regulated genes, compared with the bottom 25% group (Figure 2f). KEGG analysis showed that inflammation‐related pathways were up‐regulated and mitochondria‐related pathways were down‐regulated, similar to what we observed in the liver (Figure 2g). Correlation analyses showed that, in addition to the inflammation‐related genes (AP1 , TNFAIP3 , NOD2 , and CD44 ) observed in the liver, genes such as IFNG , CD11B , and CCL2 also correlated positively with GDF15 mRNA expression levels (Figure 2h). Similar data were observed in skeletal muscle (Figure S4a–d).
 
 
2.4 Role of GDF15 on longevity and metabolic phenotypes in mouse populations
 
To further explore the link between GDF15 and the metabolic features of aging, we analyzed whether GDF15 impacts lifespan and metabolic phenotypes in the BXD mouse genetic reference population, which is composed of ~160 genetically different mouse strains (Andreux et al., 2012). First, we analyzed the overall impact of hepatic Gdf15 expression on murine lifespan without considering their strain (Figure 3a, b). BXD mice with high expression (top 25%) of Gdf15 transcripts in the liver lived significantly shorter than mice with low expression levels (bottom 25%) (Figure 3a,b). Higher Gdf15 expression in the liver is tightly associated with gene sets involved in mitochondrial stress and quality control (Figure S5a,b). To identify the Gdf15‐expression‐associated metabolic phenotypes, we also analyzed the major phenotypes of the individual BXD mice with higher (top 25%; Gdf15‐Hi) or lower expression (bottom 25%; Gdf15‐Lo) of hepatic Gdf15 transcripts (Figure 3c). The Gdf15‐Hi BXD strains placed on a normal chow diet exhibited glucose intolerance (Figure 3d) and a lower respiratory exchange ratio (RER) compared with the Gdf15‐Lo strains (Figure 3e,f). Although we found no differences in food intake between chow‐fed Gdf15‐Lo and Gdf15‐Hi animals, body weight (Figure S6a,b) and fat mass were significantly higher in the latter, whereas the lean mass was not altered (Figure S6c,d). The chow‐fed Gdf15‐Hi group also had heavier liver masses (Figure S5e) and higher serum levels of liver injury markers (Figure 3g,h) compared to the chow‐fed Gdf15‐Lo group.
 
 
acel13195-fig-0003-m.jpg
 
 
 
Figure 3. Impact of Gdf15 on metabolic phenotypes and survival in BXD mouse reference populations.
(a) Violin plot visualizing the distribution of 241 BXD mice by hepatic Gdf15 expression. The red and blue boxes represent the top and bottom 25% of BXD populations, respectively. (b) Kaplan–Meier plot showing the survival curves for the top (n = 37) and bottom (n = 36) 25% of mice corresponding to the red and blue squares, regardless of their BXD line (log‐rank (Mantel–Cox) test p  = 0.039; Gehan–Breslow–Wilcoxon test p  < 0.0001, hazard ratio = 0.5644). © The mean expression of hepatic Gdf15 in each BXD strain at 29 weeks of age. The Gdf15‐low (blue) or Gdf15‐Hi (red) group consisted of 10 BXD lines each with, respectively, the lowest and highest hepatic Gdf15 expression levels. (d–o) Metabolism and inflammation‐related phenotypes of Gdf15‐low (blue circle) and Gdf15‐Hi (red square) groups (n = 3 to 5 mice per BXD line). The area under curve for the oral glucose tolerance test (OGTT AUC) (d), the respiratory exchange ratio (RER) during day and night (e,f), alanine aminotransferase (ALT), and aspartate aminotransferase (AST) (g, h) were obtained from the Gdf15‐low (blue) or Gdf15‐Hi (red) groups fed a normal chow diet at 29 weeks of age. The OGTT AUC (i), RER_day (j), RER_night (k), AST (l), ALT (m), plasma TNF (n), and plasma IL‐10 (o) were from the Gdf15‐low (blue) or Gdf15‐Hi (red) groups under a high‐fat diet at 29 weeks of age. Values (d–o) are mean ± SEM. *p  < 0.05, **p  < 0.01, ***p  < 0.001 ((d–o): two‐tailed t tests).
 
 
 
 
 
 
 
.../...
 
 
 
 
 
 
 
 
 
 
.

 







Also tagged with one or more of these keywords: aging, inflammation, mitochondria, senescence, t cell

3 user(s) are reading this topic

0 members, 3 guests, 0 anonymous users