Proteasome stress in skeletal muscle mounts a long-range protective response that delays retinal and brain aging

Abstract

Neurodegeneration in the central nervous system (CNS) is a defining feature of organismal aging that is influenced by peripheral tissues. Clinical observations indicate that skeletal muscle influences CNS aging, but the underlying muscle-to-brain signaling remains unexplored. In Drosophila, we find that moderate perturbation of the proteasome in skeletal muscle induces compensatory preservation of CNS proteostasis during aging. Such long-range stress signaling depends on muscle-secreted Amyrel amylase. Mimicking stress-induced Amyrel upregulation in muscle reduces age-related accumulation of poly-ubiquitinated proteins in the brain and retina via chaperones. Preservation of proteostasis stems from the disaccharide maltose, which is produced via Amyrel amylase activity. Correspondingly, RNAi for SLC45 maltose transporters reduces expression of Amyrel-induced chaperones and worsens brain proteostasis during aging. Moreover, maltose preserves proteostasis and neuronal activity in human brain organoids challenged by thermal stress. Thus, proteasome stress in skeletal muscle hinders retinal and brain aging by mounting an adaptive response via amylase/maltose.

Keywords: aging; amylase; brain organoids; maltose; muscle-to-brain signaling; muscle-to-retina signaling; myokine; proteasome; proteostasis; stress response.

Fig. 1.. Proteasome Stress in Skeletal Muscle Preserves Protein Quality Control in Distant Non-muscle Tissues.

(A) RNA-seq data indicates that RNAi for proteasome components and associated factors (Prosβ1 and Ter94/VCP) induces a local compensatory transcriptional response in muscle characterized by the increased expression of chaperones and proteases/peptidases. The most significant categories of upregulated genes are shown (logratio>1 and p<0.05; compared to white^RNAi). (B-C) Western blot analyses of thoraces (consisting primarily of skeletal muscle) and heads (consisting primarily of brains and retinas) from flies with muscle-specific Prosβ1^RNAi and control white^RNAi driven by Mhc-Gal4. Western blot analyses of detergent-insoluble fractions indicate that, although there are no substantial effects in muscle (B), muscle-specific Prosβ1^RNAi improves protein quality control in heads, as indicated by the lower age-related accumulation of poly-ubiquitinated proteins in detergent-insoluble fractions (C). (D-E) Similar improvement of proteostasis is also found in head tissues of flies with muscle-specific Ter94^RNAi, compared to control white^RNAi and vermillion^RNAi. However, protein quality control is worsened in muscle (D), presumably due to insufficient compensation by transcriptional adaptive responses (A). (F-G) Drug-induced Prosβ1^RNAi expression in thoracic flight skeletal muscle reduces age-related accumulation of poly-ubiquitinated proteins in detergent-insoluble fractions, compared to uninduced controls and to control white^RNAi (G). The levels of ubiquitin, Ref(2)P/p62, α-tubulin, and/or or β-actin are shown in (B-G), together with the color-coded quantitation of ubiquitin levels normalized to α-tubulin or β-actin. The ages analyzed are 10, 30, and 60 days. (H) Immunostaining of retinas from 30-day-old flies for ubiquitin (red), Ref(2)P/p62 (green), and F-actin (blue). Drug-induced (+RU486) expression of Prosβ1^RNAi in thoracic flight skeletal muscle reduces the age-related accumulation of poly-ubiquitin protein aggregates in retinas during aging, compared to uninduced controls and to no transgene (+). The scale bar is 20μm. (I) Proteasome stress induced in skeletal muscle via RNAi for proteasome subunits (indicated in the scheme) reduces the accumulation of proteasome substrates (poly-ubiquitinated proteins) in distant tissues (retina and brain) during aging. Supplemental Fig. S1 and S2 report additional data related to Fig. 1.

Fig. 2.. Amyrel is a key Myokine that Preserves Protein Quality Control in Distant Tissues in Response to Proteasome Stress in Muscle.

(A) RNA-seq data indicates muscle-secreted factors (myokines) that are commonly induced by RNAi for subunits of the 20S proteasome (Prosβ1 and Prosβ5) and Ter94/VCP. The logratio of Prosβ1^RNAi, Prosβ5^RNAi, and Ter94^RNAi versus white^RNAi is shown. Non-significantly regulated myokines are shaded in gray, whereas upregulated and downregulated myokines with p<0.05 are highlighted in red and blue, respectively. Some of the most-significantly regulated myokines (yellow) were selected for follow up biochemical analyses (B). (B) Summary of the analysis of muscle-specific RNAi and overexpression for selected stress-induced myokines, based on the primary data shown in Supplemental Fig. S3 and S4. The levels of poly-ubiquitinated proteins found in the detergent-insoluble fractions of heads is indicated. (C) Drug-induced Amyrel overexpression limited to the thoracic flight muscle reduces the age-related increase in poly-ubiquitinated proteins found in detergent-insoluble fractions of heads. No effect is seen with mock drug treatment (i.e. ethanol alone) and in the absence of the Amyrel transgene. (D) GFP-tagged Amyrel expressed specifically in thoracic skeletal muscle is detected in muscle, in the circulation (hemolymph), and to a lower extent in the head, indicating that Amyrel is a muscle-secreted factor. (E) Muscle-specific Prosβ1^RNAi impedes the age-related accumulation of poly-ubiquitinated proteins in head tissues during aging. However, concomitant Amyrel^RNAi blunts the effects of Prosβ1^RNAi, compared to control yellow^RNAi. (F) Muscle-specific Amyrel overexpression improves protein quality control in heads of flies with pathogenic huntingtin (GMR-Htt-Q120) as indicated by the lower age-dependent increase in poly-ubiquitinated proteins found in detergent-insoluble fractions compared to controls. The ages analyzed are 10, 30, and 60 days. (G) Drosophila retinas with transgenic expression of GFP-tagged pathogenic huntingtin (Htt-Q72-GFP) driven by GMR-Gal4. Amyrel overexpression decreases the overall amount of Htt-Q72-GFP protein aggregates in 30-day-old flies, compared to control mCherry and no transgene (+). SD, n≥18. (H) Mutant tau^V337M induces retinal degeneration during aging, as exemplified by the appearance of a rough eye phenotype due to the loss and derangement of photoreceptor neurons in 30-day-old females. Amyrel largely prevents such age-related neurodegeneration. Supplemental Fig. S3 and S4 report additional data related to Fig. 2.

Fig. 3.. C/EBP Transcription Factors Mimic the Response to Proteasome Stress and Induce Amyrel Expression.

(A) Similar gene expression changes (R²=0.808) are induced by muscle-specific Prosβ1^RNAi and Prosβ5^RNAi. (B) No correlation (R²=0.029) is found when Prosβ1^RNAi and Prosβ5^RNAi are each compared to GFP^RNAi (all normalized to white^RNAi; n = 3). (C) Similar gene expression changes (R²=0.447; and R²=0.771 for p<0.05) are induced by muscle-specific Prosβ5^RNAi and CG6272^RNAi. (D) No correlation is found when compared control GFP^RNAi (R²=0.053; and R²=0.104 for p<0.05). (E-F) Comparison of muscle-specific Te94^RNAi versus Prosβ5^RNAi indicates some overlap in the gene expression changes induced (R²=0.248; and R²=0.305 for p<0.05), similar to the comparison of Te94^RNAi versus CG6272^RNAi (R²=0.202; and R²=0.257 for p<0.05). (G) No correlation (R²=0.068, and R²=0.097 for p<0.05) is found when Te94^RNAi is compared with control GFP^RNAi. In (A-G), the identity of x and y axes is specified in the figure panels. All RNA-seq datasets are normalized to white^RNAi; GFP^RNAi serves as additional control; n = 3. (H) Gene categories upregulated in muscle by RNAi for CG6272 (C/EBPγ) include peptidases/proteases and many other gene categories (outlined) that are similarly modulated by Prosβ1^RNAi and Prosβ5^RNAi (Fig. 1 and Supplemental Fig. 1). Genes with p<0.05 and logratio>1 were used for these analyses. (I) Myokines that are upregulated (red) and downregulated (blue) by proteasome stress (Fig. 2A) are similarly modulated by CG6272^RNAi, including Amyrel. (J-K) CG6272^RNAi increases Amyrel expression in skeletal muscle, similar to overexpression of slbo, homologous to C/EBPβ and C/EBPδ compared to controls (L-M); SD, n = 3. (N) RNAi for slbo reduces Amyrel expression in Prosβ1^RNAi muscle, compared to control white^RNAi. SD, n = 3. (O) Proteasome stress in skeletal muscle is sensed via antagonizing functions of C/EBPβ/δ (slbo) and C/EBPγ (CG6272) transcription factors, which induce a local adaptive response based on the transcriptional induction of proteases, and a systemic adaptive response via modulation of muscle-secreted factors (myokines). Supplemental Fig. S3 reports additional data related to Fig. 3.

Fig. 4.. Muscle-derived Amyrel Preserves Protein Quality Control in the Brain and Retina during Aging.

Immunostaining of retinas and brains from 60-day-old flies for poly-ubiquitinated proteins (red), Ref(2)P/p62 (green), and F-actin (blue). (A) Skeletal muscle-specific overexpression of Amyrel (Mhc>Amyrel) reduces the age-related accumulation of poly-ubiquitin protein aggregates in the brain and retinas of flies, compared to isogenic controls (Mhc>+). (B-C) Similar results are found in response to drug-induced (+RU486) expression of Amyrel in thoracic flight skeletal muscle with Act88F-GS-Gal4, compared to uninduced controls and no transgene (+), in both the retina (B) and brain (C). (D) Muscle-specific Prosβ1^RNAi systemically preserves protein quality control in the retina and brain during aging via Amyrel, as indicated by the higher levels of poly-ubiquitin protein aggregates found in the brains and retinas of flies with Prosβ1^RNAi+Amyrel^RNAi versus control Prosβ1^RNAi+yellow^RNAi. In (A-D), the scale bar is 20μm. (A-D) Quantitation of the area of poly-ubiquitin protein aggregates in the brains and retinas of flies with muscle-specific modulation of Amyrel, Prosβ1^RNAi+Amyrel^RNAi, and controls. The n and SD is indicated. In (A-D), higher (4x) magnification representative images are shown for each intervention. (E) Proteasome stress in skeletal muscle improves protein quality control in the brain and retina during aging via the stress-induced myokine Amyrel. Supplemental Fig. S3 reports additional data related to Fig. 4.

Fig. 5.. Muscle-derived Amyrel Promotes the Expression of Chaperones and Proteases in Head Tissues and Preserves Neuronal Function during Aging.

(A-B) Muscle-specific Amyrel overexpression induces transcriptional changes in head tissues (n = 3), including upregulation of chaperones (orange) and proteases (blue) (p<0.05 and LogFC>0.5). The score corresponds to −Log10(p-value). (C) Hsp23 overexpression preserves protein quality control, as indicated by the lower amount of Htt-Q72-GFP protein aggregates compared to controls (SD, n≥9). Conversely, RNAi for CG9733 and CG7142 proteases (upregulated by Amyrel) increases Htt-Q72-GFP protein aggregates compared to control RNAi (SD, n=11). (D) Drug-induced overexpression of Amyrel and Hsp23 in the CNS reduces the age-related increase in poly-ubiquitinated proteins (proteasome substrates) in detergent-insoluble fractions of heads (which consist primarily of brains and retinas). No effect is seen with mock drug treatments in the absence of transgenes. The ages analyzed are 10, 30, and 60 days. (E) The capacity for startle-induced negative geotaxis declines during aging but is preserved by muscle-specific Amyrel overexpression, compared to isogenic controls (n[batches of 25 flies] = 11; SEM). Supplemental Fig. S5 and S6 report additional data related to Fig. 5.

Fig. 6.. Amyrel-Produced Maltose and Maltose Transporters Promote Protein Quality Control in the Brain and Retina via the Transcriptional Induction of Chaperones and Proteases.

(A) Body levels of maltose increase in response to muscle-specific Prosβ1^RNAi, whereas glucose is inconsistently regulated. Similarly, muscle-specific Amyrel overexpression increases body maltose levels. SD, n = 5. (B) Head maltose levels increase in response to muscle-specific Prosβ1^RNAi and Prosβ5^RNAi compared to control mCherry^RNAi; SD, n≥4. (C) Western blot analysis of detergent-insoluble fractions from Drosophila S2R+ cells treated with porcine recombinant amylase or maltose and heat shocked for 6h at 37°C. Amylase and maltose decrease the heat-induced accumulation of poly-ubiquitinated proteins in detergent-insoluble fractions. (D) RNAi for the maltose transporter Slc45-2 driven by GMR-Gal4 increases the overall amount of Htt-Q72-GFP protein aggregates in retinas, compared to control mCherry^RNAi (n≥7, SD). Psa^RNAi is a positive control. (E) Slc45-2^RNAi driven in the CNS by elav-Gal4 reduces the expression of chaperones that are upregulated by Amyrel. SD, n = 3. (F) Slc45-2^RNAi increases age-related accumulation of poly-ubiquitinated proteins in detergent-insoluble fractions of heads, compared to control RNAi. The ages analyzed are 10 and 30 days (elav>Slc45-2^RNAi flies do not survive to 60 days). (G) Immunostaining for poly-ubiquitin (red), Ref(2)P/p62 (green), and F-actin (blue) of brains from 30-day-old flies. Slc45-2^RNAi leads to an increase in the age-related accumulation of poly-ubiquitinated protein aggregates in the brain compared to white^RNAi. (H) Head tissues from Slc45-2 ^null/null flies display age-dependent increase in poly-ubiquitinated proteins in detergent-insoluble fractions from heads, compared to controls. (I) Muscle-specific Amyrel overexpression does not improve protein quality control of head tissues during aging in the absence of Slc45-2. (J-L) Western blot analysis of detergent-insoluble fractions from human HEK293 cells treated with increasing maltose concentrations (0, 3, 5, 10, and 33 mg/mL) and iso-osmolar and iso-energetic controls (NaCl and glucose). Maltose preserves protein quality control in HEK293 cells that were heat shocked for 7h at 41.5°C, whereas treatment with the disaccharide cellobiose (10 mg/mL) does not. (M) Western blot analysis of detergent-insoluble fractions from human HEK293 cells treated with 10 mg/mL maltose and either control NT siRNAs or combined SLC45A3+A4 siRNAs. SLC45 RNAi partially prevents the protective action of maltose. (N) The stress-induced amylase Amyrel produces maltose, which improves proteostasis via SLC45 maltose transporters and transcriptional induction of chaperones. Supplemental Fig. S5 and S6 report additional data related to Fig. 6.

Fig. 7.. Maltose Preserves Protein Quality Control in Human Cortical Brain Organoids Challenged by Thermal Stress.

(A) Western blots of detergent-insoluble fractions from human cortical brain organoids treated with maltose (mg/mL) and heat shocked for 7h at 41.5°C. “-HS” denotes non heat-shocked control organoids. (B) Immunostaining for ubiquitin and p62 indicates that maltose (40 mg/mL) prevents the accumulation of poly-ubiquitin protein aggregates after heat shock. The scale bar is 20 μm. The n and SD are indicated. (C) Cluster analysis of RNA-seq data (n = 3; 2095 genes). Maltose partially prevents the gene expression changes induced by thermal stress and maintains the expression of genes involved in proteostasis. (D) Multielectrode array recording of neuronal activity from cortical organoids treated with maltose and heat shocked (SEM with n=18 for each condition; each n represents a well with an organoid slice). Maltose treatment preserves neuronal activity, which is compromised by thermal stress. The p-value represents the row factor from two-way ANOVA, which indicates the effect of treatment at each time point (p<0.01). Supplemental Fig. S7 reports additional data related to Fig. 7.