Small Particles Released by Injured Muscle May Worsen Kidney Damage

Animal experiments were approved by the Institutional Animal Care and Use Committee of Ewha Womans University (No. 20‐003) and Wonkwang University (WKU23‐61). Male C57BL/6J mice were purchased from Central Lab Animal Inc. (Seoul, Korea) and were acclimated for 1 week. Grip strength was measured using a calibrated grip strength apparatus (BIOSEB, Vitrolles, France). Plasma, kidney and gastrocnemius muscle were collected upon euthanasia. Plasma blood urea nitrogen (BUN) levels were measured using an ELISA kit (Arbor Assays, Ann Arbor, MI, USA).

1Tibial nerve denervation modelTibial nerve denervation was performed in 5‐week‐old mice as previously described [12]. Briefly, the right tibial nerve was cut with microdissecting scissors, and the end of the transected tibial nerve was sutured to the adjacent biceps femoris muscle. UUO surgery was performed 7 days after denervation. In a separate study, mice were kept for 4 weeks after denervation, to evaluate the long‐term effect of muscle denervation on kidney. 2Kidney disease modelUUO surgery was performed as previously described [13], and the mice were euthanized after 5 days. To establish the CKD model, mice were fed a 0.25% adenine‐supplemented diet for 7 days [14]. 3GW4869 administrationGW4869 (Sigma‐Aldrich, St. Louis, MO, USA), an inhibitor of EV biogenesis and release [15], was administered to 5‐week‐old mice at 5 mg/kg/day for 12 days via intraperitoneal or intramuscular injections, starting on the day of denervation. Control mice received equivalent volumes of 0.005% dimethyl sulfoxide (DMSO). 4Exercise study in miceSeven‐week‐old mice underwent moderate‐intensity exercise on a RotaRod (Ugo Basile SRL, Gemonio, Italy) at a constant speed of 5.8 m/min for 1 h. For acclimation, mice were placed on the rotating rod at approximately 3 m/min for 5 min 1 day prior. Mice were sacrificed 15 min after completing the exercise [16].

C2C12 cells (ATCC, Manassas, VA, USA) were maintained at 37°C in a humidified incubator with 5% CO₂ and cultured in high‐glucose DMEM (Gibco, Waltham, MA, USA) supplemented with 10% fetal bovine serum (Gibco). Myotube differentiation was induced using 2% horse serum for 5 days. For conditioned medium (CM) production, differentiated myotubes were stimulated with dexamethasone or transforming growth factor beta 1 (TGFβ1), washed with filtered phosphate‐buffered saline (PBS) and incubated with EV‐depleted medium for 40 h [17]. EVs were depleted from horse serum by ultracentrifugation at 100 000× g for 18 h at 4°C (Type 100 Ti rotor; Beckman Coulter, Indianapolis, IN, USA) [18]. mProx24 mouse renal tubular epithelial cells were supplied by Dr. Takeshi Sugaya (St. Marianna University School of Medicine, Kanagawa, Japan) and cultured in high‐glucose DMEM supplemented with 10% fetal bovine serum [13]. miR‐21a inhibitor (Thermo Fisher Scientific) was treated to C2C12 myotubes or mProx24 cells for 24 h. All in vitro experiments were conducted using mProx cells at passages 23–27 and C2C12 cells at passages 8–10. Each replicate was treated with EVs isolated from individual mouse or human plasma samples.

The human study adhered to the Declaration of Helsinki and was approved by the Institutional Review Boards of Ewha Womans University Medical Center and registered on the International Clinical Trials Registry Platform (KCT0001137) [19]. Participant characteristics are detailed in Table S1. All subjects provided informed consent. The exercise group performed knee‐strengthening resistance exercises (1 set at 10RM, 3 times/week) and received Korean mistletoe extract (2 g/day) for 12 weeks. The control group received no intervention. Plasma samples were collected for small EV isolation. Also, human muscle data (GSE165632↗) from the Gene Expression Omnibus were used to analyse miRNA expression among older adult men (aged 65–80 years) with different exercise habits. Specifically, the dataset included five sedentary subjects and nine senior sportsmen, including four predominantly resistance trained and five predominantly endurance trained. Data were analysed using muscle tissues obtained from vastus lateralis.

Plasma or CM was centrifuged at low speed to remove large debris, filtered through a 0.22 μm filter and purified using commercial kits (EX‐03, EX‐01; Rosetta Exosome Inc., Seongnam, South Korea) or a qEV size exclusion column (qEVsingle/35 nm; IZON, Medford, MA, USA). The column separates particles based on size, retaining those larger than 35 nm. As EVs typically elute early, the first 0.8 mL of eluate was collected. Purified EVs were then resuspended and diluted in filtered PBS. The number of EVs was measured by dynamic light scattering (Zetasizer Nano‐ZS90; Malvern Panalytical Ltd., Malvern, UK) and EV concentration was analysed using a Nanoparticle Tracking Analyser (NS300; Malvern Panalytical Ltd.). To isolate EVs from muscle, freshly harvested gastrocnemius muscle was harvested from mice 2 weeks after denervation and incubated in RPMI 1640 medium (Gibco) containing 2 mg/mL type 2 collagenase (Worthington Biochemical Corporation, Worthington, NJ, USA) and 40 U/mL DNase I (Roche, Basel, Switzerland) at 37°C for 30 min [20]. After differential centrifugation, the supernatants were filtered through a 0.45‐μm membrane and centrifuged at 100 000× g for 70 min. Isolated EVs were either used to treat mProx24 cells for 24 h (50 μg) or injected into mice via the tail vein for 5 days (50 μg).

Isolated EVs were diluted to 1:100 in PBS, and 5 μL diluted EVs were dropped on Formvar–carbon‐coated EM grids. The grids were stained with 2% uranyl acetate for 2 min at RT and removed using filter paper. EV morphology was evaluated by transmission electron microscopy (H‐7650; Hitachi‐Science & Technology, Tokyo, Japan).

mRNA levels were measured by qPCR using a SYBR Green PCR Master Mix kit (Applied Biosystems, Waltham, MA, USA) in a StepOne real‐time PCR system (Applied Biosystems). mRNA expression was normalized to 18S rRNA levels. We confirmed that 18S rRNA levels did not differ among groups (data not shown). miRNA was extracted from EVs, and cDNA synthesis was conducted using TaqMan Advanced miRNA Assays (Thermo Fisher Scientific, Waltham, MA, USA). miR‐21a‐3p expression was measured by qPCR using a TaqMan Advanced Master Mix 2X (Thermo Fisher Scientific). The primer sequences are listed in Table. S2

Cell lysate was mixed with 5× sample buffers and heated at 95°C for 6 min. Total protein concentrations were measured using the Bradford method (Bio‐Rad, Hercules, CA, USA), and whole lysates were subjected to SDS‐PAGE. After membrane blocking with 5% skim milk, the target proteins were probed using various antibodies (Table). The protein bands were detected using ChemiDoc MP Imaging System (Bio‐Rad). Kidney protein levels were normalized to that of glyceraldehyde 3‐phosphate dehydrogenase (GAPDH). S3

For IHC staining, anti‐collagen 1 (COL1; SouthernBiotech, Birmingham, AL, USA), anti‐neutrophil gelatinase‐associated lipocalin (NGAL; Abcam, Cambridge, UK) and anti‐alpha‐smooth muscle actin (αSMA; Abcam) antibodies were used. For fluorescence images, anti‐CD63 (Abcam) antibody was used. Sample sections were incubated with primary antibodies overnight and then with Alexa 568‐conjugated anti‐rabbit (Invitrogen, Carlsbad, CA, USA) antibody at room temperature for 1 h. Images were captured using a Zeiss microscope (Carl Zeiss, Oberkochen, Germany). Randomly selected digital images (20–25) from each mouse were quantified using Image‐Pro Plus 4.5 software (Media Cybernetics, Rockville, MD, USA).

Public RNA‐Seq datasets from the Gene Expression Omnibus database were analysed to identify differentially expressed genes (DEGs) in denervated muscle (GSE62812↗) and obstructed kidneys (GSE38117↗). In GSE62812↗, gastrocnemius muscles were collected 14 days after sciatic nerve denervation or sham surgery (n = 3/group), and DEGs were identified by comparing the two groups. Similarly, in GSE38117↗, gene expression was compared between kidneys with and without UUO (n = 3/group). miRNA expression changes were analysed in denervated muscle (GSE162794↗, n = 4/group) and obstructed kidneys (GSE42719↗, n = 4/group). Differential expression was considered significant with a fold change ≥ 1.5 or ≤ 0.8 and adjusted p‐value < 0.05. Predicted miRNAs targeting Ppargc1a were validated using miRbase (https://www.mirbase.org/↗). Gene ontology enrichment analysis of 155 candidate genes was performed using g: Profiler (https://biit.cs.ut.ee/gprofiler/gost↗) [21].

All results are expressed as the mean ± standard error (SE). Differences among multiple groups were assessed using one‐way or two‐way analysis of variance (ANOVA) in GraphPad Prism (version 8.0.2), followed by Tukey's post hoc test for pairwise comparisons. A two‐tailed Pearson correlation analysis was used to assess the relationship between biomarkers. Statistical significance was set at p < 0.05.

To determine the effect of muscle wasting on kidney injury progression, we performed UUO in tibial nerve‐denervated mice (Figure 1a). Muscle atrophy was confirmed by lower muscle mass and strength in both DEN and DEN + UUO groups compared with control (Figure 1b,c). Atrogin‐1 and Myog expressions were higher (Figure 1d,e), whereas mitochondria‐related gene expressions were lower in denervated mice (Figure 1f). Two‐way ANOVA revealed significant interactions between denervation and UUO for Ppara (F = 10.55, p = 0.005), Mcad (F = 5.349, p = 0.0344) and Cytc (F = 5.467, p = 0.0327). UUO alone caused mild but significant muscle injury, evidenced by lower muscle weight and higher Atrogin‐1 expression (Figure 1b,d), along with lower expression of Ppara, Mcad and Cytc (Figure 1f). However, muscle strength in UUO mice without denervation did not differ from control (Figure 1c). Notably, the extent of muscle injury was comparable between DEN and DEN + UUO groups.

Next, we examined the effect of muscle denervation on the kidney. Elevated plasma BUN levels in UUO mice were aggravated in DEN + UUO mice (Figure 2a). UUO‐induced kidney inflammation, as indicated by upregulated Mcp‐1, Il‐1, Il‐6, F4/80, Tnfa, Vcam‐1 and Icam‐1, and fibrosis markers, including Acta2, Col1, Col4, Fn and Tgfb1, were higher in DEN + UUO mice compared with UUO alone (Figure 2b). Comparable results were observed for αSMA, fibronectin (FN) and COL1 protein levels (Figure 2c,d), with significant interactions between denervation and UUO for αSMA (F = 47.19, p < 0.0001) and FN (F = 19.06, p = 0.0001). Similar findings were observed in the adenine diet‐induced CKD model, where denervation further elevated BUN levels and enhanced expression of inflammation and fibrosis markers in the kidneys of adenine‐fed mice (Figure S1).

Notably, 12 days of denervation alone was sufficient to induce IL‐1, IL‐6, TNFα and Vcam‐1 expression in the kidney, along with higher protein levels of αSMA, FN and COL1 (Figure 2b–d). To determine whether prolonged denervation could further exacerbate renal injury, a separate cohort of mice was observed for 4 weeks post‐denervation without UUO surgery or adenine‐diet supplement (Figure S2). Compared with control, denervated mice exhibited significantly lower body weight and muscle strength (Figure S2a,b), along with increased albuminuria (Figure S2c). Consistently, renal expression of inflammatory and fibrotic genes was elevated in these mice (Figure S2d). These findings support a causative role of denervated muscle in the development of kidney injury.

We hypothesized that small EVs from denervated muscle could mediate crosstalk with the kidney. To test this hypothesis, we first isolated EVs from mouse plasma. The expression of small EV markers (CD9, CD63, CD81 and Alix), as well as EV particle number, was higher in plasma EVs isolated from denervated mice compared with sham controls (Figure 3a,b), while average vesicle size was unaltered (Figure 3c). Notably, Alix and Atrogin‐1 expression levels were positively correlated in both gastrocnemius muscle (r = 0.7921, p < 0.0001) and differentiated myotubes (r = 0.6382, p < 0.0001) (Figure 3d). Consistently, muscle protein EV markers were higher in denervated mice compared with sham controls (Figure 3e). Gene expression also confirmed that muscle EV production is induced by denervation and unaffected by UUO (Figure S3a). In contrast, kidney EV production was significantly induced by UUO (Figure S4). Next, we treated cultured myotubes with dexamethasone, a synthetic glucocorticoid known to induce muscle injury [22], or TGFβ1, which cause both muscle wasting [23] and kidney fibrosis [13]. As expected, these treatments induced the expression of EV markers (Figure S3b,c). Notably, genes related to EV secretion (Rab11a, Rab11b and Rab35) were also increased.

We further examined whether small EVs could be targeted to the injured kidney. PKH67‐labelled plasma EVs from control mice were intravenously administrated to UUO mice (Figure 3f). EVs accumulated in the obstructed kidney but not in the control kidney, as indicated by higher fluorescence intensity (1.57 × 10⁸ vs. 1.03 × 10⁸). mProx24 proximal tubular epithelial cells were also treated with plasma EVs for 1 h, confirming EV uptake into the cells (Figure 3f). Collectively, these results imply that EVs could be effectively delivered to the damaged kidney.

Although muscle injury promotes EV production and release, it remains unclear whether these EVs contribute to the progression of renal injury. To address this, we utilized GW4869 to block EV generation and release [15] (Figure 4a). Circulating EVs were effectively removed by intraperitoneal GW4869 administration, confirmed by diminished EV marker expression in plasma (Figure 4b). Plasma BUN levels were lower in GW4869‐treated denervated UUO mice compared with DMSO‐treated controls (Figure 4c). Genetic markers of inflammation and fibrosis were significantly lower in GW4869‐treated mice (Figure 4d). Partial inhibition of induced genes implied the involvement of non‐EV factors in muscle–kidney crosstalk. The anti‐fibrotic effect of GW4869 was more evident in protein levels, where the elevated expression of FN, αSMA and COL1 in denervated UUO mice was attenuated by GW4869 treatment (Figure 4e,f). As plasma EVs could be derived from tissues other than muscle, we repeated the experiments by intramuscularly injecting GW4869. Diminished CD63 and TSG101 expression in muscle and plasma, respectively, confirmed that the EVs were effectively removed (Figure S5a,b). Similar to the intraperitoneal injection results, upregulated inflammatory and fibrotic genes in denervated UUO mice kidneys were significantly reduced by EV depletion in muscle (Figure S5c). Moreover, the Masson's trichrome and αSMA stained areas induced by denervation in UUO mice were completely ameliorated (Figure S5d,e). These results indicate that plasma EVs in denervated mice are mostly derived from muscle, and their depletion can attenuate the effect of muscle denervation on renal injury.

To provide direct evidence for the role of EVs in muscle–kidney crosstalk, mProx24 cells were used (Figure 5a). The TGFβ1‐induced downregulation of E‐cadherin and upregulation of Acta2, Col1 and Col4 in mProx24 cells were aggravated following pretreatment with plasma EVs from denervated mice compared with EVs from sham controls (Figure 5b). Next, we isolated EVs from the CM of C2C12 myotubes (Figure 5c). EVs isolated from the CM of dexamethasone‐ and TGFβ1‐treated myotubes aggravated E‐cadherin downregulation and Acta2 upregulation in TGFβ1‐treated mProx24 cells (Figure 5d). The pathogenic role of muscle‐derived EVs was evaluated in vivo by injecting the EVs isolated from denervated muscle (Figure 6a) and TGFβ‐treated C2C12 myotubes (Figure 6c). Five‐day injection of damaged muscle‐derived EVs resulted in aggravated UUO‐induced expression, as shown by Masson's trichrome, NGAL and αSMA staining, compared with EVs from sham mice (Figure 6b) or control myotubes (Figure 6d).

We hypothesized that the increased miRNAs within EVs could mediate the muscle–kidney crosstalk. As the increased miRNAs would decrease target mRNAs in both muscle and kidney, we first identified the genes commonly downregulated in denervated muscle and fibrotic kidney (Figure 7a and Table S4). Functional enrichment analysis showed that the 155 commonly downregulated genes were enriched in mitochondrial function (Figure S6). Thus, we narrowed down the gene list to those related to mitochondria and fibrosis. Among the four genes related to fibrosis (Mapk14, Ndufs6, Ppargc1a and Ldha), Ppargc1a was the only one associated with kidney fibrosis [24, 25]. Among the miRNAs commonly upregulated in denervated muscle and kidney fibrosis (Table S5), we screened for those targeting Ppargc1a and identified miR‐21a‐3p. To confirm, we measured Ppargc1a expression in denervated muscle and obstructed kidney, both of which exhibited significantly lower levels compared with sham controls (Figure 7b). miR‐21a‐3p content was higher in EVs isolated from the CM of dexamethasone‐ and TGFβ1‐treated C2C12 myotubes, as well as in EVs from denervated muscle, compared with controls or sham mice (Figure 7c). A negative correlation was observed between Ppargc1a and miR‐21a‐3p expression in both muscle (r = −0.6558, p = 0.0017) and kidney (r = −0.8611, p < 0.0001) (Figure 7d). Treatment with miR‐21a inhibitor resulted in increased PGC1α protein in mProx24 cells and Ppargc1a gene expression in myotubes (Figure 7e). These results show that miR‐21a‐3p, as cargo in small EVs, is upregulated in damaged muscle and targets Ppargc1a to downregulate its expression.

Next, two strategies were used to examine the role of miR‐21a‐3p‐containing EVs in mediating kidney injury progression. First, C2C12 myotubes were treated with miR‐21a inhibitor before isolating EVs from CM (Figure 7f). Second, mProx24 cells were pretreated with miR‐21a inhibitor before stimulation with EVs isolated from either denervated mice or the CM of C2C12 myotubes (Figure S7a). In both cases, TGFβ1‐induced mProx24 injury was ameliorated by depleting miR‐21a (Figure 7f and Figure S7b–d), implying that increased miR‐21a‐3p is the major component of muscle‐derived EVs responsible for kidney fibrosis progression.

While muscle atrophy can aggravate kidney disease, exercise has beneficial effects in ameliorating kidney disease [26]. Therefore, we sought to determine the effect of exercise‐induced EVs on kidney fibrosis (Figure 8a). In mice, a single bout of RotaRod exercise effectively upregulated Ppargc1a and reduced miR‐21a‐3p levels in muscle (Figure 8b). Next, mProx24 cells were treated with plasma EVs isolated from these mice. TGFβ1‐induced Acta2, Fn, Col1 and Tgfb1 expression was inhibited by EVs from exercised mice compared with those from control mice (Figure 8c). Consistent with the mouse data, data mining revealed that muscle miR‐21a‐3p expression was lower in senior sportsmen trained primarily in either resistance or endurance exercise compared with sedentary subjects (Figure 8d). We previously reported that resistance exercise combined with Korean mistletoe extract (KME) supplementation mitigates age‐related muscle loss and strength decline [19]. We utilized the plasma from these individuals to isolated EVs (Figure 8e). Compared with controls, higher muscle mass and lower fat mass was observed in the exercised group (Table S1) Treatment of mProx24 cells with isolated EVs resulted in lower expression of TGFβ1‐induced genes when using EVs from the exercise plus KME group compared with those from the controls (Figure 8f). These results partly show that miR‐21a‐3p inhibition and small EV content regulation in muscle could account for the anti‐fibrotic effect of exercise on the kidney.