The Influence of Aging on the Unfolded Protein Response in Human Skeletal Muscle at Rest and Following Acute Exercise

Abstract

Background:

The unfolded protein response (UPR) is a proteostatic process that is activated in response to endoplasmic reticulum stress. It is currently unclear how aging influences the chronic and adaptive UPR in human skeletal muscle. Here we determined the effect of aging on UPR activation at rest, in response to exercise, and the associations with muscle function.

Methods:

Thirty young (20–35 yrs) and 50 older (65–85 yrs) individuals were enrolled. Vastus lateralis biopsies were performed at rest and 3hrs and 48hrs after a single bout of resistance exercise. The abundance of UPR-related transcripts and proteins were measured by RNA sequencing and Western blotting, respectively. Fractional synthetic rates (FSR) of muscle protein were determined by mass spectrometry following intravenous infusion of ¹³C₆ phenylalanine.

Results:

Older adults demonstrated elevated transcriptional and proteomic markers of UPR activation in resting muscle. Resting UPR gene expression was negatively associated with muscle strength and power in older adults. The UPR is similarly activated by acute resistance exercise in young and older adults and positively associated with muscle function but not the anabolic response to exercise.

Conclusions:

Skeletal muscle from older adults exhibits chronically activated UPR, which accompanies functional decline. The adaptive UPR is a proteostatic mechanism that is upregulated in response to exercise in young and older adults and positively associated with muscle function.

INTRODUCTION

Skeletal muscle is a critical determinant of health span in older adults because of its influence on metabolic health, physical function, mobility, and quality of life. With advancing age, skeletal muscle mass and strength progressively decline (1), leading to frailty and disability (2). A wide range of factors have been mechanistically linked with detrimental muscle phenotypes of aging, including dysregulated protein metabolism (3). The preventive and restorative potential of physical activity on skeletal muscle function is evident across all age groups, including older adults (4), but continued efforts are needed to understand why aging attenuates the adaptive responses to exercise (5) and the recovery from disuse (6). Several potential mechanisms have been proposed, including dysregulated skeletal muscle protein synthesis and anabolic signaling in older adults (i.e., anabolic resistance) (5, 7) and impaired protein breakdown (8). However, recent evidence suggests that muscle anabolic processes are not limited with aging, but compromised proteostatic mechanisms such as protein folding and protein degradation may underly impaired recovery from disuse in aging (9). The extent to which muscle atrophy is caused by impaired anabolic vs. proteolytic processes is debated (8, 10), but it is clear that derangements in proteostatic mechanisms are highly relevant to age-related muscle loss and attenuated adaptive responses to exercise.

The unfolded protein response (UPR) is an important quality control process in skeletal muscle that may play a role in age-related skeletal muscle phenotypes (11). The UPR is governed by three interconnected arms: inositol-requiring protein 1 (IRE1α), activating transcription factor 6 (ATF6), and protein kinase RNA (PKR)-like ER kinase (PERK). When unfolded or misfolded proteins accumulate in the endoplasmic reticulum (ER) lumen, these UPR sensors become activated and induce transcription of genes involved in protein folding, (12), ER membrane expansion (13), and phosphorylation of eukaryotic translation initiation factor 2 (eIF2α) (14), which inhibits protein translation. Prolonged ER stress can trigger apoptosis through expression of ATF4 and the pro-apoptotic C/EBP homologous protein (CHOP, also known as DDIT3). Early evidence suggests that the UPR may be involved in the development of age-related skeletal muscle phenotypes (11). Cellular damage accumulates with aging, and the efficiency of repair mechanisms is reduced (15, 16). Aging is also accompanied by chronic inflammation, oxidative stress, and the accumulation of unfolded or misfolded proteins (17); conditions that may trigger the UPR in aged tissues. In rodents, the UPR is activated or dysregulated with aging in a variety of tissues (18, 19), but much less is known about the role of the UPR in skeletal muscle (16). Skeletal muscle from aged rats exhibits transcriptional and proteomic evidence of altered UPR (20). In mice, aging alters the expression of UPR and ER stress markers, but in a manner that is highly dependent on muscle group (21). In aged mice, disruption of ER homeostasis contributed to lower muscle protein synthesis and loss of muscle mass (22). Limited data is available on aging human skeletal muscle, but it is conceivable that the UPR may act as a protective response against age-related accumulation of unfolded or misfolded proteins with aging.

The UPR is a molecular response pathway that is consistently activated following exercise in rodents (23, 24) and humans (25, 26). Older rats exhibited elevated UPR markers and attenuated recovery following disuse compared to young animals (27), but how aging influences this molecular exercise response pathway in humans is not well understood. Skeletal muscle UPR activation after acute resistance exercise was similar in young and older men (25, 26). We recently found that the expression of UPR genes was modestly elevated in skeletal muscle from a small cohort of older adults compared to young at rest, and that the transcriptional markers of UPR activation were attenuated following a single bout of exercise in older adults (28). The need for continued research into this topic is highlighted by current ambiguity regarding the protective vs. detrimental role of UPR activation in aging skeletal muscle. The current study was conducted to determine the influence of age on UPR transcripts and proteins in human skeletal muscle and their associations with muscle phenotypes. An additional objective was to determine how aging influences UPR activation after acute resistance exercise and whether chronic or adaptive UPR activation is associated with the anabolic response to exercise.

METHODS

Study design and participant characteristics.

This work is a secondary analysis of biospecimens and phenotype data from a previous study (29) where data on body composition, muscle function, and muscle protein synthesis have been reported. The new data contained in this manuscript include muscle gene expression and protein abundance on a subset of 80 individuals for which there was sufficient muscle tissue for RNA sequencing and immunoblotting. The study (NCT03350906) was approved by the Mayo Foundation Institutional Review Board (IRB #17-004403) and followed the principles outlined in the Declaration of Helsinki. Thirty younger (20–35 years) and 50 older (65–85 years) men and women were included in this analysis. All participants provided written informed consent. Potential participants were excluded for history of cigarette, tobacco, alcohol, or substance abuse; untreated or uncontrolled thyroid disease; diabetes or fasting plasma glucose > 126 mg/dL; blood clotting disorders; anemia (hemoglobin < 11 g/dL for females, <12 g/dL for males); liver disease (aspartate aminotransferase [AST] > 144 IU/L or alanine aminotransferase [ALT] > 165 IU/L); renal insufficiency (serum creatinine > 1.5 mg/dL); or coronary artery disease. Participants were also excluded if they were taking any medication that could potentially influence the outcomes or increase the risks of the study, such as insulin, metformin, anticoagulants, opiates, tricyclic antidepressants, barbiturates, or benzodiazepines.

Outpatient phenotyping

Body composition was measured using dual-energy X-ray absorptiometry (DEXA; GE Lunar iDXA, GE Healthcare, Chicago, IL, USA). Peak whole-body oxygen consumption (VO₂peak) was assessed using indirect calorimetry during a graded treadmill test by using the standard Bruce protocol. Unilateral knee extensor strength and power were measured using a pneumatic leg extension machine (Keiser Air300, Keiser Corporation). Participants were instructed to wear a 3-axis accelerometer (wGT3X-BT, Actigraph) during waking hours for a period of 2 weeks to assess daily physical activity levels.

Inpatient study procedures

At least 1 week after the outpatient visit, participants reported to the Mayo Clinic Clinical Research and Trials Unit (CRTU) for an overnight study visit. Participants were provided with 3 days of weight-maintaining meals prior to the inpatient study. Participants were admitted to the CRTU at approximately 1730hrs. Following an evening meal at 1800hrs, participants remained fasting until the end of the visit. At 0400hrs the next morning, a peripheral intravenous catheter was placed for the infusion of isotopic labeled phenylalanine (details below), and a retrograde intravenous catheter was placed in the opposite hand for blood sampling. At 0830hrs, a pre-exercise percutaneous muscle biopsy of the vastus lateralis muscle was performed using a modified Bergstrom needle and local anesthetic (2% lidocaine). Muscle tissue was immediately frozen in liquid nitrogen before storage at −80°C. At approximately 0900hrs, participants performed exercise with the contralateral leg, completing 8 sets of 10 reps of single-leg extensions at 70% of 1RM, determined at the outpatient visit. Post-exercise muscle biopsies were performed at 3hrs and 48 hours following exercise.

RNA extraction and sequencing

Frozen muscle samples were pulverized, suspended in QIAzol^® lysis reagent, and homogenized using a Bead Ruptor 24 bead mill homogenizer (Omni International). Total RNA was isolated and purified using the miRNeasy^® Mini Kit (Qiagen). RNA quality and concentration of the RNA were assessed using the Agilent 2100 Bioanalyzer and the NanoDrop ND-2000 Spectrophotometer, respectively. Subsequently, RNA samples were submitted to the Mayo Clinic Genome Analysis Core Laboratory for RNA sequencing. For RNA library preparation, 200 ng of total RNA was used following the manufacturer’s instructions for the TruSeq Stranded mRNA Sample Prep Kit (Illumina, San Diego, CA). The concentration and size distribution of the completed libraries was determined using an Agilent Bioanalyzer DNA 1000 chip (Santa Clara, CA) and Qubit fluorometry (Invitrogen, Carlsbad, CA). The libraries were then sequenced at 37 samples per lane following the standard protocol for the Illumina NovaSeq^™ 6000. Individual lane loading was performed using the NovaSeq XP 4-Lane kit. The flow cells were sequenced as 100 X 2 paired end reads using the NovaSeq S4 sequencing kit and NovaSeq Control Software v1.8.0. Base-calling was performed using Illumina’s RTA version 3.4.4.

Protein extraction

The lysis buffer for protein extraction contained 1x Cell Lysis Buffer (Cell Signaling, #9803), cOmplete Mini EDTA-free Protease Inhibitor Cocktail (Roche, #04693159001) and Halt Phosphatase Inhibitor Cocktail (Thermo Fisher, #78420). Immediately prior to tissue homogenization, fresh lysis buffer was added at a ratio of 1:20 to approximately 40mg of pulverized, frozen muscle tissue powder inside 1.5ml tubes pre-filled with 1.4mm ceramic beads (Omni International, #19-617). Tissue homogenization was performed using a Bead Ruptor 24 bead mill homogenizer (Omni International) at 4°C, using a speed setting of 5 for 3 cycles of 15s on and 10s off. After homogenization, the samples were maintained at 4°C with constant agitation on a rocking platform for 30 mins. Subsequently, the homogenates were transferred to new 1.5ml tubes and centrifuged at 12,000 RPM for 20 mins at 4°C. The resulting supernatant was transferred to new 1.5ml tubes and stored in multiple aliquots at −80°C for downstream assays. Protein concentrations were estimated using the Pierce 660nm Protein Assay Kit (Thermo Fisher, #22662) following the manufacturer’s instructions.

Western blotting

Samples were prepared for denaturing gel electrophoresis with NuPAGE LDS sample buffer (Thermo Fisher #NP0007) and reducing agent (Thermo Fisher, #NP0004) according to the manufacturer’s instructions. Denatured samples, along with the Chameleon Duo protein ladder (LI-COR, #928-60000), were loaded onto 17-well NuPAGE 4 to 12%, Bis-Tris, 1.0 mm, Mini Protein Gels (Thermo Fisher, #NP0329BOX). Electrophoresis was performed in NuPAGE MOPS SDS running buffer (Thermo Fisher, #NP0001) using the XCell SureLock Mini-Cell (Thermo Fisher, #EI0001) following the manufacturer’s instructions. Proteins were then transferred from the gels to Immobilon-FL PVDF membranes (Millipore, #IPFL00010) using NuPAGE Transfer Buffer (Thermo Fisher, #NP0006) and the XCell II Blot Module (Thermo Fisher, #EI9051) following the manufacturer’s instructions. The membranes were allowed to dry overnight at 4°C, followed by total protein staining using the Revert 700 Total Protein Stain kit (LI-COR, #926-11010) and imaged using the LI-COR Odyssey FC imaging system and LI-COR Image Studio software (v5.2). The membranes were then de-stained and blocked for 1 hr at room temperature (RT) with shaking in Intercept Blocking Buffer (LI-COR, #927-60001). Subsequently, the Blocked membranes were incubated overnight at 4°C with shaking in primary antibody solutions containing 1x Intercept Blocking Buffer and 0.2% Tween 20 (Sigma-Aldrich, #P1379). Afterward, the membranes were washed four times for 5 mins each with shaking at RT in TBS-T (1X Tris-Buffered Saline [Biorad, #1706435] and 0.1% Tween 20). Next, the membranes were incubated with secondary antibody solutions consisting of 1x Intercept Blocking Buffer, 0.2% Tween 20, and .015% SDS (Promega, # V6551) for 1 hr covered and shaking at RT. Following the incubation, membranes were washed four times with TBS-T and given a final wash in 1X TBS. Membranes were imaged as described earlier. The signal intensity of the target protein bands was determined and normalized to the total protein stain using the LI-COR Empira Studio (v2.3) qualitative Western blot analysis workflow. All full blots for each target protein, total protein stains, and linear range tests for each antibody are available in Supplemental Digital Content 1.

Antibodies

The appropriate antibody concentration and protein loading amount for each gel were determined empirically by performing a linear range analysis (30) in Empira Studio (v2.3, LI-COR). All gels were loaded with 5µl of the Chameleon Duo protein ladder and 50µg of tissue homogenate per sample, except for CALR detection, where 30µg of protein per sample was loaded. The primary antibodies and their concentrations used are: BiP 1:1000 (Cell Signaling, #3183); CHOP 1:1000 (Cell Signaling, #2895); eIF2α 1:10,000 (Santa Cruz, #133227); P-eIF2α (Ser51) 1:500 (Cell Signaling, #9721), CALR 1:2000 (Invitrogen, MA1-91034); IRE1α 1:1000 (Cell Signaling, 14C10); GADD45α 1:500 (Cell Signaling, #4632); GADD34 1:1000 (Invitrogen, #PA1-139). All secondary antibodies were used at a concentration of 1:20,000, including IRDye 800CW Goat anti-Mouse IgG (LI-COR, #926-32210); IRDye 800CW Goat anti-Rabbit IgG (LI-COR, #926-32211); IRDye 680RD Goat anti-Mouse IgG (LI-COR, #926-68070); IRDye 680RD Goat anti-Rabbit IgG (LI-COR, #926-68071).

Skeletal muscle protein synthesis rates

Skeletal muscle protein fractional synthesis rates (FSR) were determined from the incorporation of isotopically labeled phenylalanine into muscle proteins. At 0500hrs on the inpatient study visit, a priming bolus of [¹³C₆]phenylalanine (1mg/kg FFM) was administered intravenously, followed by a continuous infusion at 1mg/kg FFM/hr. The isotope enrichment of [¹³C₆]phenylalanine in plasma, protein-free muscle tissue fluid (TF), and mixed muscle protein (MMP) was measured using liquid chromatography tandem mass spectrometry (LC-MS-MS), as previously described (7). FSR was calculated as previously described (31) using the standard precursor-product equation: FSR (%/hr) = Ep₂ – Ep₁/[E_precursor x t] x 100, where Ep₁ is the protein-bound ¹³C₆ phenylalanine enrichment in mixed plasma proteins before isotope infusion, Ep₂ is the protein-bound enrichment of phenylalanine in the muscle biopsy, E_precursor is the muscle tissue fluid free amino acid enrichment, and t is the elapsed time from tracer administration to muscle collection.

Statistical analyses

Unpaired ttests were used to compare body composition, physical characteristics, muscle performance, physical activity levels, and muscle protein synthesis rates in young and older groups. Univariate linear regression was used to evaluate associations between UPR-related transcripts or proteins and physical performance outcomes (VO2 peak, muscle strength, muscle power) or muscle protein synthesis rates. Statistical analyses were performed using GraphPad Prism version 10.0.0 (GraphPad Software, Boston, Massachusetts USA). For muscle RNA sequencing data, differential expression analysis was carried out using a negative binomial generalized log-linear model in the edgeR R package (R Foundation). A list of 278 UPR genes was curated from the Molecular Signatures Database (MSigDB) Hallmark unfolded protein response M5922, Reactome unfolded protein response M10294, and the Kyoto Encyclopedia of Genes and Genomes (KEGG) protein processing in endoplasmic reticulum - Homo sapiens hsa04141 gene lists, plus a few additional UPR-related genes of interest based on literature reviews, which include BID, CASP2, NFKB1, GADD45A, and TP53 (32). The gene set enrichment analysis (GSEA) used only the MSigDB Hallmark UPR gene list. RNA sequencing data is available through the Gene Expression Omnibus (GEO accession GSE249921).

RESULTS

Participant characteristics

Young and older adults were similar in height, weight, body mass index, and body composition (Table 1). The steps per day were also similar, with a trend (P = 0.06) towards less moderate-to-vigorous intensity physical activity (MVPA) in the older group (Table 1). Total lean mass and appendicular skeletal muscle index were similar in young and older adults (Figure 1A), consistent with general good health in this cohort of participants and the absence of evidence of frailty or obvious sarcopenia. Nonetheless, whole-body cardiorespiratory fitness (VO₂peak), leg muscle strength (1RM), and leg muscle power were significantly lower in older compared to young adults (Figure 1A).

Table 1. Participant Characteristics and Body Composition.

Values are presented as the mean ± standard deviation. MVPA = daily minutes in moderate to vigorous physical activity. F= female, M= male. P values were calculated using a two-tailed unpaired t test.

Characteristic	Younger	Older	P value
Sex	15F/15M	24F/26M
Age (years)	27.1 ± 4.1	71.5 ± 4.4
Height (cm)	171.3 ± 8.6	169.5 ± 10.0	0.39
Weight (kg)	73.8 ± 11.3	75.1 ± 14.8	0.66
Body mass index	24.9 ± 2.7	26.0 ± 3.6	0.17
Lean mass (kg)	48.2 ± 8.5	46.6 ± 10.0	0.47
Total body fat (%)	31.7 ± 8.0	34.7 ± 7.5	0.09
Daily step counts	7540.0 ± 1896.6	7768.7 ± 2997.9	0.71
MVPA (minutes)	35.6 ± 19.9	26.6 ± 20.4	0.06

Figure 1. Comparisons of Pre-exercise Unfolded Protein Response-Related Gene Expression Between Young and Older Participants.

(A) Physiological fitness measurements for lean mass, skeletal muscle index, VO₂peak (VO2), 1 repetition maximum (1RM), and peak power (PP), n= 30 young, 50 older. Bar plots are colored light gray for younger participants and dark gray for older participants. Individual points representing female participants are colored red and males are colored blue. P values were determined by a two-tailed unpaired t test. (B) Volcano plot of pre-exercise (pre-ex) skeletal muscle UPR-related gene expression between older and young participants. Transcripts with a false discovery rate adjusted P value (FDR) of > 0.05 (below the horizontal dotted line) are represented by gray points. Transcripts with an FDR < 0.05 (above the horizontal line) are colored light red or light blue if they have a positive or negative log₂ fold change (log₂FC) in the older grouper compared to the young group, respectively. Genes were considered significantly differentially expressed (DE) if they had an FDR < 0.05 and a log₂FC (vertical dotted lines) of > 0.5 (dark red) or < −0.5 (dark blue). (C) Heatmap of individual participant gene expression pre-ex. Each column represents an individual participant, with young on the left and older on the right. The displayed UPR-related genes had the greatest difference in gene expression between the older and young groups. Values represent the log₂FC-transformed ratio of gene expression relative to the average expression in the young group. This value is represented by red if positive, white if 0, and blue if negative. (D) Heatmap of significant correlations between UPR-related gene expression pre-ex and muscle phenotypes. Pearson’s correlation coefficients for the younger or older groups were calculated by correlating gene expression to VO2, 1RM, and PP. Correlations are represented by red if positive, white if 0, and blue if negative. Correlations were considered statistically significant if the P value was ≤ 0.05 and are indicated on the heatmap by a *.

Pre-exercise UPR-related gene expression in young and older skeletal muscle

We curated a UPR-related gene list from the Molecular Signatures Database (MSigDB) Hallmark, MSigDB Reactome, and the Kyoto Encyclopedia of Genes and Genomes (KEGG) UPR gene lists. In resting skeletal muscle, a total of 273 UPR-related genes were detected in the RNA sequencing dataset, which contained over 18,000 transcripts (Figure 1B). A total of 113 genes exhibited significantly altered expression levels between young and older adults, as determined by a false discovery rate corrected P value (FDR) < 0.05. Among these, 53 genes showed higher expression levels in older adults, while 60 genes exhibited lower expression levels. Of these genes, 6 were classified as significantly differentially expressed (DE) with FDR < 0.05 and log₂ fold change (log₂FC) > 0.5 or < −0.5. Specifically, DNAJC5B, DDIT4, and GADD45A were upregulated in older skeletal muscle, whereas ATF3, SLC7A5 and SLC14A were downregulated. GADD45A (GADD45α) is a nuclear stress response protein that promotes cell cycle arrest in response to ER stress in a P-eIF2α dependent manner (33) and has been associated muscle age-related muscle loss (34). The heatmap in Figure 1C shows the log₂FC in gene expression for each individual young and older participant compared to the average of the young group. The top UPR-related genes with the greatest differences in gene expression between the young and older group are shown. Notably, several of these genes (DDIT4, DDX11, EXOSC10, GADD45A, NABP1, TSPYL2, YWHAZ) are involved in DNA damage repair or regulation of cell cycle progression. We next investigated whether the pre-exercise expression of UPR-related genes were associated with skeletal muscle strength (1RM), peak power (PP) or whole-body cardiorespiratory fitness (VO₂) in young and older adults (Figure 1D). The majority of upregulated UPR-related transcripts in older adults were significantly negatively associated with 1RM and PP, but not VO₂peak. These associations were evident in older adults, but largely absent in young adults. These results demonstrate that healthy older adults exhibit transcriptional evidence of an upregulated UPR in skeletal muscle in the absence of obvious sarcopenia. The negative associations between UPR-related gene expression and muscle function suggests that chronically upregulated UPR in skeletal muscle may be linked with an age-related decline in muscle function.

UPR-related gene expression in young and older skeletal muscle after resistance exercise

To determine if the adaptive UPR is dysregulated with aging in skeletal muscle, we compared the UPR-related transcriptional response to acute resistance exercise in young and older adults (Figure 2A–D). In young individuals, a total of 153 UPR-related transcripts were significantly altered 3 hours post-exercise compared to pre-exercise (FDR < 0.05), with 106 increased and 47 decreased (Figure 2A). Of these genes, 27 were categorized as differentially expressed (FDR < 0.05 and log₂FC > 0.5 or < −0.5); 26 were upregulated and 1 was downregulated. In older adults, a total of 151 genes had expression significantly altered 3 hours post-exercise (93 increased, 58 decreased). Of these, 18 were considered differentially expressed (Figure 2B); 17 were upregulated and 1 was downregulated. A total of 18 differentially expressed genes were in common between young and older adults 3 hours post-exercise (Figure 2E), including core UPR effectors EIF2AK3 (eIF2α), DDIT3 (CHOP), and PPP1R15A (GADD34) and several protein chaperones (CRYAB, HSPA1A, HSPA1B, BAG3, DNAJA1, DNAJA4, DNAJB4). Nine genes were differentially expressed in young but not older adults 3 hours after resistance exercise. These genes included two chaperones (HSP90AA1 and HSPH1), a transcription factor (NFYA), a translation initiation factor (EIF4A1) and a gene involved in ribosome biogenesis (RRBP1). Together, the results indicate that UPR-related gene expression is induced 3 hours following a single bout of resistance exercise in both young and older adults, but the induction is modestly greater in young adults.

Figure 2. Comparisons of UPR-related Gene Expression Between Young and Older Participants After Acute Resistance Exercise.

(A-D). Volcano plots of the fold changes in skeletal muscle UPR gene expression at 3 hours post-exercise vs pre-exercise (3hr post/pre) in young (A) and older participants (B) or 48hr post/pre in young (C) and older (D) participants. Gray points represent genes with a false discovery rate adjusted P value (FDR) of > 0.05 (below the horizontal dotted line). Light red or light blue points represent genes with an FDR < 0.05 (above the horizontal line) and a positive or negative log₂ fold change (log₂FC) post/pre, respectively. Dark red or dark blue points represent genes that are considered significantly differentially expressed (DE), as defined by an FDR of < 0.05 and a log₂FC (vertical dotted lines) of > 0.5 or < −0.5, respectively. (E) Venn diagram illustrating the comparison of DE UPR genes between young 3hrs post/pre (red), older 3hrs post/pre (purple), young 48hrs post/pre (yellow) and older 48hrs post/pre (pink). Areas of non-overlap represent DE genes unique to a specific condition. Overlapping areas between two or more circles represent shared DE genes between the corresponding two or more conditions. (F) Gene set enrichment analysis (GSEA) plot depicting significant enrichment of UPR-related genes 3hr or 48hr post/pre in young and older participants. Colored tick marks below the x-axis represent individual genes and their rank in the ordered dataset. Left-side enrichment indicates a positive enrichment score (ES), reflecting increased UPR expression post/pre. Colors: Red - young 3hrs post/pre; blue - older 3hrs post/pre; yellow - young 48hrs post/pre; pink - older 48hrs post/pre (G) Bar plot depicting the fold change in muscle fractional synthesis rates (FSR) 3hrs post/pre. Bar plots are colored light gray for younger participants and dark gray for older participants. Individual points representing female participants are colored red, and males are colored blue. P value was determined by a two-tailed unpaired t test. n = 30 young, 48 older. (H) Heatmap of individual participant gene expression 3hr and 48hr post/pre. Each column represents an individual participant. The 37 DE UPR genes from Figure 3A–E are displayed. Values represent the log₂ fold change (log₂FC) in gene expression post/pre. This value is represented by red if positive, white if 0, and blue if negative. (I) Heatmap of significant correlations between UPR gene fold change (log₂FC) post/pre and muscle phenotypes. Pearson’s correlation coefficients for the younger or older groups were calculated by correlating the fold change in gene expression to VO₂peak (VO2), 1 repetition maximum (1RM), peak power (PP), or FSR. Correlations are represented by red if positive, white if 0, and blue if negative. Correlations were considered statistically significant if the P value was ≤ 0.05 and are indicated on the heatmap by a *.

A subset of muscle samples from 12 young and 13 older participants was used to examine UPR-related gene expression at 48 hours post-exercise. Body composition and physical function of individuals in this subset were representative of the larger cohort (Supplemental Digital Content 2). In young adults at 48 hours post-exercise, the expression levels of a majority of the UPR-related genes approached pre-exercise values (Figure 2C). Of the 64 significantly altered UPR-related genes, 47 had increased and 17 had decreased expression compared to pre-exercise values. Ten genes exceeded thresholds for differential expression (6 upregulated, 4 downregulated). In older adults, of the 133 significantly altered UPR-related genes, 102 were significantly increased and 31 were significantly decreased at 48 hours post-exercise compared to pre-exercise values (Figure 2D). Of these, 16 genes met the criteria for differential expression, with 13 upregulated and 3 downregulated. A comparison of differentially expressed genes in young and older adults at 48 hours reveals 9 common genes involved in recovery and repair processes after cellular stress (CCL2, HYOU, NABP1, TUBB2A, CALR, and PDIA4). Seven differentially expressed genes specific to the older group include genes involved in protein folding and quality control (KDELR3, TXNDC5, DNAJA4). Taken together, the gene expression data at 48 hours post-exercise show that while most UPR-related genes are close to their pre-exercise values, several remain differentially expressed, particularly in older adults.

Gene set enrichment analysis (GSEA) was performed using the MSigDB Hallmark UPR gene list at both post-exercise muscle biopsy time points (Figure 2F). The significant positive enrichment scores (ES) in young and older adults reinforces the notion that the UPR is transcriptionally activated in the hours and days following a single bout of exercise. Furthermore, the ES values were similar between the young and older groups, indicating that, on the pathway level, UPR gene expression is similarly upregulated after exercise in both young and older individuals. The inter-individual heterogeneity in the post-exercise UPR transcriptional response is evident in the heatmap in Figure 2H, which shows the change in expression of UPR-related genes from pre-exercise to 3 hours or 48 hours post-exercise in individual young and older participants.

Next, we determined if the induction of UPR-related genes following resistance exercise was associated with the anabolic response to acute resistance exercise, measured from the post-exercise fold change in fractional synthesis rate (FSR) of muscle proteins (Figure 2G). The FSR fold change was not statistically different between young and older adults (Figure 2G), but there was notable heterogeneity in the anabolic response in both age groups, with some individuals exhibiting robust increases in FSR after exercise, while others exhibited minimal or no increase. The FSR values for the subset of 25 participants with data at 48 hours post-exercise are given in Supplemental Digital Content 2. In Figure 2I, we determined the extent to which the post-exercise fold change in UPR-related gene expression was predictive of muscle strength, power, cardiorespiratory fitness, or muscle anabolic response to exercise. A notable major pattern was observed in young individuals where the 3-hours post exercise fold change in the gene expression of 21 UPR-related genes was significantly positively associated with muscle strength. This group of genes includes heat shock proteins (HSPA1A, HSPA1B, DNAJA4, DNAJA1, HSPH1, and HSP90AA1), EIF2AK3 (PERK), translation initiation factors (EIF4E and EIF4A1), genes involved in cellular stress response (GADD45A, BAG3, CRYAB) and protein trafficking and quality control (SEC24D and CALR). This pattern was attenuated in older adults who demonstrated positive associations between muscle strength and the fold induction of only 8 of the same UPR-related genes: SLC7A5, EIF2AK3, EIF4E, GADD45A, TUBB2A, EIF4A1, YWHAZ, and SEC24D. The post-exercise fold change in FSR was not significantly associated with the fold induction of any UPR-related genes following exercise except in older adults where CALR, SHC1, and PDIA4 showed positive and UBE2D1 showed negative associations with FSR. In sum, the fold change in UPR-related gene expression following acute resistance exercise was positively associated with muscle function in young and older adults, albeit less pronounced in older adults. Importantly, the variability in anabolic response to exercise could not be explained on the basis of transcriptional UPR activation.

UPR-related protein expression in young and older adults at rest and following acute resistance exercise

Since the UPR is regulated by the abundance and phosphorylation of key proteins, we performed Western blotting for several UPR targets in muscle biopsies from the previously identified subset of 12 young and 13 older participants at baseline, 3 hours post-exercise, and 48 hours post-exercise. Representative blots are shown in Figure 3G. For each target protein, normalized gene counts for the corresponding gene are provided from the RNA sequencing dataset on the same muscle samples on which Western blotting was performed (Figures 3A–C). Resting, pre-exercise mRNA abundance of HSPA5, CALR, and GADD45A were significantly higher in older skeletal muscle compared to young (Figure 3A). BiP, the protein product of HSPA5, was also significantly increased in older adults, as was CHOP (DDIT3; Figure 3D). At rest, older adults exhibited modestly higher GADD45α protein abundance, although this was less than the difference in GADD45A gene expression (Figure 3D). At 3 hours post-exercise, HSPA5, DDIT3, GADD45A, and PPP1R15A mRNAs were upregulated compared to pre-exercise. Furthermore, the change in HSPA3 and GADD45A gene expression was significantly higher in young compared to older adults (Figure 3B). However, the corresponding fold changes in the protein products of these genes were modest and similar in young and older adults (Figure 3E). At 48 hours post-exercise, mRNA abundance approached baseline values with the exception of CALR, which remained elevated (Figure 3C). Protein abundance was similar in young and older adults at the 48 hour post-exercise timepoint (Figure 3F).

Figure 3: Comparisons of UPR-related protein expression between young and older participants at rest and after acute resistance exercise with correlations to gene expression.

Normalized RNA-seq mRNA expression data in reads per kilobase million (RPKMs) (A-C) and corresponding Western blot protein abundance data in arbitrary units (AU) (D-F) for young and older participants pre-exercise (pre-ex), 3hrs or 48hrs post-exercise vs pre-exercise (referred to as 3hr or 48hr post/pre). The 7 targets, with their common protein names listed first followed by the official corresponding gene name in parentheses, are: BiP (HSPA5), CHOP (DDIT3), P-eIF2α/eIF2α (EIF2S1), CALR, IRE1α (ERN1), GADD45α (GADD45A), and GADD34 (PPP1R15A). P-eIF2α/eIF2α represents the abundance of Ser51 phosphorylated eIF2α over the abundance of total eIF2α. All bands were normalized to the total protein stain internal loading control. Bar plots are colored light gray for younger participants and dark gray for older participants. Individual points representing female participants are colored red, and males are colored blue. P values were determined by a two-tailed unpaired t test. (G) Representative Western blots. (H) Heatmap of significant correlations between UPR protein expression and muscle phenotypes. Pearson’s correlation coefficients were calculated by correlating protein abundance with VO₂peak (VO2), 1 repetition maximum (1RM), peak power (PP), or muscle tissue fluid fractional synthesis rate (FSR). Correlations are represented by red if positive, white if 0, and blue if negative. Correlations were considered statistically significant if the P value was ≤ 0.05 and are indicated on the heatmap by a *. (I-J) XY plots illustrating the log2 fold change (Log2FC) in UPR-related mRNA and protein abundance 3hr post/pre (I) or 3hr post/pre mRNA abundance and 48hrs post/pre protein abundance (J) for individual participants. Light gray and dark gray points represent young and older participants, respectively.

These discordant patterns prompted us to more closely examine the relationship between gene expression and protein abundance. At the 3 hours post-exercise timepoint, the 3 most upregulated genes (DDIT3, GADD45A, PPP1R15A) were not accompanied by corresponding increases in protein abundance (Figure 3I). To address the possibility that changes in protein abundance are temporally mismatched from acute changes in gene expression, we also examined the relationships between the change in gene expression 3 hours post-exercise to the change in protein abundance 48 hours post-exercise, but still found weak agreement (Figure 3J). Together, the data suggest that older adults have upregulated mRNA and protein UPR markers at rest. Furthermore, the induction of the UPR following exercise is modestly attenuated in older adults based on gene expression, but it is similar in young and older adults at the level of protein abundance and phosphorylation. The relationship between UPR-related mRNA and protein expression is not straightforward and likely involves additional layers of post-transcriptional, translational, and post-translational regulatory control. Indeed, we do not observe consistent predictive patterns between UPR protein markers and muscle phenotypes or the anabolic response to exercise (Figure 3H). However, a notable preliminary observation is a significant positive association between the fold-induction of IRE1alpha 3 hours post exercise and the anabolic response to exercise (FSR) in older adults. Furthermore, we also observe a general pattern in older adults where muscle function is positively associated with increases in UPR protein markers in muscle 48 hours post-exercise. Supplemental Digital Content 3 provides all of the summary statistics for correlation analyses performed in this manuscript.

DISCUSSION

We evaluated the influence of aging on the chronic and adaptive unfolded protein response in human skeletal muscle. Our key findings are: 1) older adults exhibit elevated transcriptional and proteomic markers of UPR activation in resting muscle, 2) basal UPR gene expression is negatively associated with muscle strength and power in older adults, 3) the UPR is similarly activated by acute resistance exercise in young and older adults, and 4) post-exercise UPR signaling is positively associated with muscle function but not the anabolic response to exercise. Chronically activated UPR is a feature of aged skeletal muscle that accompanies functional decline, and the adaptive UPR is a proteostatic mechanism that is upregulated in response to exercise in young and older adults and positively associated with muscle function.

Our observation that older adults exhibit elevated abundance of UPR-related transcripts and proteins in skeletal muscle aligns with previous studies demonstrating heightened UPR activity at rest in older rodents (11). Specifically, the key molecules BiP and CHOP were found to be elevated with aging in a variety of rodent tissues (18, 19), including skeletal muscle (20, 21, 35). In humans, heightened levels of BiP, a major ER chaperone and UPR activator, suggest increased ER stress and UPR activation in skeletal muscle with aging (12, 17). The pro-apoptotic factor CHOP becomes activated during persistent ER stress, and our observations of elevated CHOP protein in aged human skeletal muscle suggests chronic ER stress and a transition from pro-survival to pro-apoptotic UPR signaling. In conjunction with this new evidence for increased ER stress and UPR activation in skeletal muscle in the context of human aging, we find that this transcriptional pattern of UPR activation in older adults is associated with reduced muscle strength and power production in the same muscle group. Notably, the abundance of UPR-related transcripts was largely dissociated from cardiorespiratory fitness (VO₂ peak) despite being reduced in older compared to young to a similar degree as muscle strength and power. This observation supports the notion that upregulated UPR and ER stress in aging skeletal muscle may be more strongly linked with muscle-specific than whole-body functional phenotypes.

The UPR is transiently activated in response to acute exercise and is believed to be an important exercise-responsive proteostatic pathway in skeletal muscle (25, 36, 37). Consistent with prior literature, our GSEA results demonstrate that UPR gene expression is enriched 3hours and 48 hours following a single bout of resistance exercise in both young and older age groups (Figure 2F), supporting the notion that the UPR is an exercise-responsive pathway in skeletal muscle (11, 23–26, 36, 37). A goal of the current study was to determine how aging influences UPR induction following acute exercise. Slightly more UPR-related genes were differentially expressed following exercise in young compared to older adults (Figure 2A, B, E), although the post-exercise UPR transcriptional patterns were largely similar in both age groups. Likewise, young and older adults did not demonstrate any remarkable differences in the post-exercise fold change in protein abundance of key UPR-related molecules (Figure 3E, F). Although some published data suggests that aging may blunt induction of the UPR or other stress-responsive pathways in older skeletal muscle following exercise (16, 37), the current study suggests that post-exercise UPR activation is maintained in older adults, in agreement with other reports (25, 26). It is important to highlight that the current study was performed in independent, community-dwelling older adults without evidence of frailty or major disease, and the results should not be generalized across the entire spectrum of aging, including pre-frail or frail individuals. Although the UPR-related transcriptional response was similar in young and older adults 3 hours post exercise, we observed residual expression of UPR-related genes in older adults 48 hours after exercise (Figure 2C–D). Notably, most of these differentially expressed genes at 48 hours were not differentially expressed at 3 hours post-exercise. We speculate that this temporal pattern of UPR-related transcripts is related to distinct molecular chaperones required to manage the ER stress in the early post-exercise phase versus the molecules needed for ongoing repair, protein refolding, and cellular remodeling in the later phase of recovery. The possibility that that UPR activation after acute resistance exercise may be prolonged in older adults requires further interrogation.

An additional objective of this work was to investigate if heterogeneity in the anabolic response to exercise could be explained on the basis of UPR activation, which acts as a molecular brake on protein translation. We anticipated a negative association between UPR activation and post-exercise muscle protein synthesis, which could link heightened UPR activation to anabolic resistance. We demonstrate that the anabolic response to exercise was similar in young and older adults, with notable heterogeneity in the fold induction of muscle protein synthesis (Figure 2G). Although age-related anabolic resistance has been documented (38–40) this is not a universal finding (41) and highlights that evidence of this phenomenon likely depends on subject characteristics, exercise modality, nutritional support of protein synthesis, and other factors. Nevertheless, we can clearly identify individuals within each age group who exhibit robust or modest anabolic responses to the exercise stimulus, providing an opportunity to investigate the relationship between anabolic response and UPR activation. Unexpectedly, the anabolic response to exercise was not strongly associated with any clear transcriptional or protein markers of UPR activation in skeletal muscle. This finding is consistent with a previous report in aged rodents (21) and suggests that while the UPR is an exercise responsive molecular pathway, it does not explain why some older adults exhibit a muted anabolic response to exercise. Alternatively, post-exercise UPR activation may be temporally mismatched from the acute anabolic response in muscle, providing negative feedback to protein translation in response to ER stress. We cannot ignore the possibility that sampling muscle at 3 hours and 48 hours post exercise may have missed a window of time when UPR signaling events may have influenced the initial anabolic response in the first 3 hours following exercise.

Although our data do not support the hypothesis that age-related anabolic resistance can be explained by interindividual heterogeneity in UPR activation, a number of limitations of this work must be considered. The invasive nature of the human muscle biopsy technique limits the temporal resolution of post-exercise measurements of gene expression, protein abundance, and fractional synthesis rates. It is possible that we missed the ideal window to capture the peak post-exercise response in these measurements. Indeed, we observed a general mismatch between transcriptional signals and abundance of protein products for UPR-related molecules, with the exception of BiP which was elevated in skeletal muscle from older adults at the transcript and protein levels at rest (Figure 3A–F). Further interrogation of this transcript-protein mismatch is shown in Figure 3I which demonstrates dissimilar patterns in fold changes in mRNA and protein of key UPR-related molecules at 3hrs post-exercise. While a temporal lag between mRNA expression and the protein abundance is likely, we did not observe any significant associations between the mRNA fold change at 3 hours and the protein abundance at 48 hours (Figure 3J). The 3-hour and 48-hour post-exercise time points provide two discrete windows into UPR activation dynamics, but they cannot capture the full time course of UPR activation. Another limitation is that we were restricted in the number of UPR-related protein targets that we could interrogate by Western blotting and the sample size of the cohort used for protein abundance measurements. Future studies could leverage advanced mass spectrometry techniques to provide better coverage of the abundance and activation of a larger number of UPR-related proteins in all 3 arms of the pathway. Lastly, care should be taken when generalizing the results from this study of relatively healthy older adults to more clinical situations of frailty, sarcopenia, and the spectrum of age-related disease. These less healthy subsets of the aging population require further investigation to understand the role of adaptive and maladaptive UPR in skeletal muscle phenotypes and exercise response.

Dysregulated proteostatic pathways, including the UPR, are increasingly recognized as contributors to diminished muscle quality and function with aging and chronic disease. There is ongoing research into the potential utility of UPR-modulating strategies in conditions of chronic ER stress, such as obesity, diabetes, neurodegeneration, cancer, and skeletal muscle myopathies (42, 43). Overexpression of BiP in obese mice improved insulin sensitivity and reduced liver triglyceride and cholesterol levels (44), while overexpression of protein disulfide isomerase (PDI) improved cardiomyocyte survival in a mouse model of myocardial infarction (45). Another approach is the use of chemical chaperones, which are small molecules that facilitate proper protein folding. Treatment of myopathic mice with tauroursodeoxycholic acid (TUDCA) lowered BiP and ATF6 levels, decreased triglyceride and lipid levels, and improved muscle strength (46). Similarly, administration of sodium 4-phenylbutyrate (4PBA) to myopathic mice resulted in reduced ER stress and UPR activation, stimulated protein synthesis, and improved muscle function (47). Other approaches include inhibiting specific UPR molecules, enhancing the degradation of abnormally folded proteins, improving the protein folding environment of the ER, and enhancing regulators of protein trafficking or secretion (42, 43). Whether these strategies are safe and effective in humans remains to be determined.

CONCLUSIONS

In conclusion, we provide evidence supporting that healthy older adults exhibit heightened UPR activation in the skeletal muscle, which is negatively associated with muscle strength and power in the absence of frailty or obvious sarcopenia. Furthermore, we demonstrate that the adaptive UPR is an exercise-responsive molecular pathway that is similarly upregulated in skeletal muscle of healthy young and older adults, but unrelated to the post-exercise anabolic response. Altogether, this work suggests that chronically activated UPR accompanies the functional decline in aging skeletal muscle, and the adaptive UPR is a proteostatic mechanism that is upregulated in response to exercise and positively associated with muscle function.