Mechanical overload-induced release of extracellular mitochondrial particles from tendon cells leads to inflammation in tendinopathy

Tendinopathy is prevalent and debilitating, with substantial personal and economic burdens. Mechanical overload is a leading etiological factor, yet the mechanistic link between overload and tendinopathy remains unclear. Prior work implicates extracellular matrix degeneration and dysregulation as well as inflammation, with interactions between resident tendon cells and immune cells driving matrix breakdown. Subcellular stress responses, including lysosomal quality control, ER stress, and mitochondrial dynamics, influence inflammation and regeneration, but their roles in tendinopathy are not well defined. This study asks whether mechanical loading alters mitochondrial organization in tendon cells and whether mitochondria-related extracellular particles mediate crosstalk with immune cells to promote inflammation in tendinopathy. The authors hypothesize that overload damages the mitochondrial network, promotes release of extracellular mitochondrial particles (ExtraMito), and that these particles drive macrophage chemotaxis and inflammatory cytokine production.

The paper situates tendinopathy as involving ECM degeneration/dysregulation and inflammation, with stromal tendon cells interacting with infiltrating immune cells to orchestrate inflammatory cascades and matrix degradation. It highlights subcellular responses—lysosomal quality control, ER stress, mitochondrial dynamics—as modulators of inflammation and regeneration in other contexts, suggesting analogous roles in tendinopathy. Prior evidence shows macrophage infiltration and proinflammatory cytokines (e.g., IL-1β, TNF-α, CXCL1, IL-6) in tendinopathy, particularly early in disease. The evolving nomenclature and biology of extracellular particles/vesicles are noted, including mitochondrial-origin vesicles described as mitovesicles or extracellular mitochondria in other systems, where they can influence inflammatory responses. However, the specific contribution of the mitochondrial component versus nonmitochondrial components within heterogeneous extracellular particle populations has been unclear.

• Human tendon transcriptomics: Unhealthy rotator cuff tendons (n=2) and healthy hamstring tendons (n=3) collected with ethics approval. Total RNA extracted (TRIzol), RIN>7.0. TruSeq Stranded mRNA libraries prepared and sequenced (Illumina NovaSeq 6000, 150 bp paired-end). Reads cleaned, aligned to GRCh38 with HISAT2; QC with FAST-QC; counts via HTSeq; expression by RPKM; DEGs via DESeq2 (P<0.05). Hallmark gene set enrichment (MSigDB h.all.v7.4) and GO analyses (org.Hs.eg.db, clusterProfiler). Microarray validation: GEO GSE26051 analyzed with limma (P<0.05), GSEA (hallmark, GO BP) with 1000 permutations. Macrophage infiltration inferred by xCell using TPM-normalized RNA-seq (RNA-seq=True; significance P<0.2 per xCell guidance).
• Mouse tendon cell culture and 3D constructs: Tendon cells isolated from 6–8-week-old C57BL/6 mice (Achilles and patellar tendons; collagenase II digestion). Cells expanded to P3. 3D constructs formed by generating ECM-rich monolayer sheets using CTGF (25 ng/mL) and ascorbic acid (4.4 µg/mL) for 6 days, then mounted on tissue hooks in a programmable uniaxial bioreactor and cultured in EV-depleted FBS medium.
• Mechanical loading regimens: Cyclic uniaxial strain at 0%, 3%, 6%, or 9% (0.25 Hz; 8 h/day; 16 h rest; 6 days; 37°C, 5% CO2). 6% defined as normal load, 9% as overload based on gene expression responses; 0–3% considered underload.
• Mitochondrial labeling and protection: Live-cell mitochondrial imaging with MitoTracker Red (MTR). For genetic mitochondrial labeling, CellLight Mitochondria-RFP used. Antioxidant protection with N-acetylcysteine (NAC) 2.5 mM during 9% or 0% strain.
• Imaging and quantification: CLSM (Nikon A1 series) for live 3D imaging; Z-stacks (0.1–0.15 µm). Image processing (denoising, deconvolution, 3D rendering) with NIS-Elements, Imaris, ImageJ. Mitochondrial morphology quantified as mitochondrial footprint, aspect ratio, and branch junctions/mitochondrion. TEM used to visualize mEP ultrastructure.
• RT-qPCR: RNA from constructs; cDNA synthesis (M-MLV). qPCR with SYBR Green. 36B4 as control. 2^-ΔΔCt analysis; triplicate technical repeats and three biological replicates.
• Immunoblotting: Lysates from constructs or mEP pellets; RIPA buffer with inhibitors; SDS-PAGE and nitrocellulose transfer. Primary antibodies: β-actin, VDAC, TOM20, CD63, Annexin A2, Lamin A/C, Albumin; HRP secondaries; chemiluminescence detection. Quantified with ImageJ.
• Finite element analysis: 2D plane stress model (ANSYS v2021R1) to illustrate uneven strain distribution; isotropic, linear elastic, incompressible; E=10 GPa; mesh size 0.2 mm; applied 3% strain.
• mEP isolation and fractionation: Conditioned media clarified (400 g 5 min; 1500 g 10 min) and pelleted mEPs at 18,000 g for 30 min. Density gradient separation in iodixanol (40% load; overlay 20–5% steps; 200,000 g 16 h, 4°C); fractions collected and pelleted (18,000 g 30 min). mEP size assessed by flow cytometry calibration beads (SSC) and by dynamic light scattering (Zetasizer). Membrane content assessed with PKH26 staining by flow cytometry. Marker analysis for compartmental proteins via immunoblot.
• ExtraMito detection and depletion: Constructs pre-stained with MTR (mitochondria) and CFSE (cytoplasm) before loading; post-loading CLSM for ExtraMito forms (free mitochondria and mitochondria-encapsulated mEPs). Mitochondria depleted from mEPs using MACS with anti-TOM20 microbeads (Miltenyi); flow-through designated TOM20− mEPs; retained TOM20+ eluted. Density gradient fractions probed for TOM20/VDAC to define ExtraMito-enriched fractions.
• Macrophage assays: RAW 264.7 cells starved (serum-free DMEM) for 24 h. Boyden chamber chemotaxis assay (8 µm pores) with mEPs as chemoattractants; staining and OD560 readout. mEP dosing normalized to tendon construct mass/length. Cytokine production measured after 8 h incubation with mEPs (or PBS/LPS control) using LEGENDplex Mouse Macrophage 13-plex bead-based assay by flow cytometry; dose–response with 10-fold concentration gradients. Cytokine content within mEPs also measured. Morphological effects assessed by CLSM after RAW 264.7 exposure to MTR-labeled ExtraMito.
• Statistics: Student’s t-test for two-group comparisons; one-way or two-way ANOVA with Tukey’s post hoc for multiple groups. n≥3 biological replicates; imaging from ≥5 areas per replicate; data as mean ± SEM; significance at P<0.05.

• Human tendon transcriptomics: Identified 1083 upregulated and 1261 downregulated genes (P<0.05, |logFC|>2) in tendinopathy vs healthy. Hallmark enrichment highlighted “hypoxia” and “inflammatory response” as top events. GO cellular components enriched for extracellular terms, including “extracellular vesicle” and “extracellular organelle.” GO biological processes enriched for inflammatory functions such as “leukocyte migration” and “cell chemotaxis.” Independent microarray dataset (GSE26051) GSEA corroborated activation of inflammatory response pathways and extracellular components.
• 3D mechanical model validates load regimes: Compared to 0% strain, 6% strain increased tenogenic gene expression in constructs—Scx (2.51-fold), Mkx (3.39-fold), Tnmd (1.58-fold), Col1a1 (2.45-fold). Relative to 6% strain, 9% strain decreased Scx (0.52), Mkx (0.60), Tnmd (0.76), Col1a1 (0.66), indicating overload impairs cell function; 0–3% underload had minimal effect on tenogenesis.
• Mitochondrial network is load-sensitive: Live-cell CLSM showed elongated, interconnected networks at 6% strain, with largest mitochondrial footprint, highest aspect ratio, and most branch junctions/mitochondrion among groups. Overload (9%) led to fragmented, punctate mitochondria. VDAC and TOM20 protein levels were highest at 6% strain, confirming increased mitochondrial content with normal loading.
• Tendon cells release ExtraMito particles: mEPs (100–1000 nm) contained membranes (>75% PKH26+), displayed characteristic ultrastructure (TEM), and were enriched for CD63 and Annexin A2 with low Lamin A/C and Albumin. Two ExtraMito forms were observed: free extracellular mitochondria and mitochondria-encapsulated mEPs. Immunoblotting of mEPs showed ExtraMito markers (VDAC, TOM20) were highest in the 9% overload condition (mEP9%).
• mEPs drive macrophage responses: RAW 264.7 chemotaxis was significantly greater toward mEP9% than toward mEP0%, mEP3%, or mEP6%. xCell analysis suggested increased macrophage infiltration in human tendinopathy tissues. mEP exposure increased multiple cytokines vs PBS (IL-6, CCL22, G-CSF, CCL17, CXCL1, IL-23, IL-18, TNF-α, IL-12p40). Among these, IL-6, CXCL1, and IL-18 levels varied with loading; mEP9% induced the highest IL-6 and CXCL1; dose–response analyses showed mEPs mediated cytokine release in a concentration-dependent manner, with mEP9% eliciting particularly strong IL-6 and IL-18 responses across doses and higher CXCL1 at higher doses. Cytokines were below detection in mEPs themselves, indicating macrophage origin.
• ExtraMito are essential for chemotaxis but not sufficient alone: Depletion of mitochondria from mEP9% via anti-TOM20 MACS reduced VDAC/TOM20 and significantly decreased RAW 264.7 chemotaxis versus sham-depleted controls. Purified ExtraMito9% (density fractions 4–8) retained chemotactic activity comparable to unfractionated mEP9%, whereas mEP9% fractions with fewer ExtraMito (fractions 1–3) showed negligible chemotaxis, underscoring ExtraMito’s necessity. However, TOM20+ particles alone did not significantly increase chemotaxis over PBS, indicating ExtraMito are necessary but not sufficient by themselves. Depleting ExtraMito did not significantly change mEP9%-induced IL-6, CXCL1, or IL-18 secretion, suggesting non-mitochondrial mEP components primarily drive cytokine induction.
• Mitochondrial protection mitigates chemotaxis: NAC (2.5 mM) preserved mitochondrial network integrity under 9% overload (increased footprint, aspect ratio, branch junctions) and significantly reduced mEP9%-induced macrophage chemotaxis.

The study connects mechanical overload to inflammation in tendinopathy via mitochondrial dynamics and extracellular signaling. Overload disrupted the intracellular mitochondrial network in tendon cells and increased the release of extracellular mitochondrial particles (ExtraMito) within mEPs. Functionally, mEPs from overloaded cells enhanced macrophage chemotaxis and proinflammatory cytokine secretion, aligning with transcriptomic evidence of leukocyte migration and inflammatory activation in human tendinopathy. Mechanistically, ExtraMito were indispensable for chemotaxis, while cytokine induction (IL-6, CXCL1, IL-18) was largely mediated by other mEP components. This delineation suggests division of labor within mEP cargo: mitochondrial elements bias macrophage migratory behavior (potentially through effects on polarization), whereas non-mitochondrial constituents trigger cytokine production. Protecting mitochondrial networks with NAC reduced chemotactic signaling from overloaded tendon cells, highlighting mitochondrial homeostasis as an upstream modulator of immune microenvironment remodeling. These insights provide a mechanistic framework for how mechanical signals are transduced into inflammatory cues through extracellular mitochondrial communication in tendinopathy and suggest broader relevance to other overload-associated inflammatory diseases.

This work identifies tendon cell-derived extracellular mitochondrial particles (ExtraMito) as key mediators linking mechanical overload to inflammation in tendinopathy. Normal loading promotes interconnected mitochondrial networks, whereas overload fragments mitochondria and elevates ExtraMito release within mEPs. mEPs from overloaded constructs drive macrophage chemotaxis and proinflammatory cytokine secretion; ExtraMito are essential for chemotaxis but not for cytokine induction, which appears to rely on non-mitochondrial mEP components. Preserving mitochondrial integrity with NAC reduces overload-induced chemotactic signaling. These findings reveal a novel mechanobiological mechanism in tendinopathy and propose mitochondria-targeted strategies to modulate immune responses. Future studies should: (1) develop methods to isolate and manipulate specific mEP subcomponents (free vs encapsulated mitochondria) to dissect synergistic effects; (2) validate findings in vivo and in human primary cells; (3) assess mitochondria-targeted therapeutics for tendinopathy; and (4) explore relevance to other mechanical overload pathologies.

• The human RNA-seq cohort was small (n=2 diseased, n=3 healthy), which may limit generalizability despite validation with external microarray data.
• The macrophage studies used the RAW 264.7 cell line; responses may differ in primary macrophages in vivo.
• Tools to isolate the mitochondrial portion encapsulated within mEPs were not available; thus, potential synergy between free mitochondria and mitochondria-encapsulated mEPs could not be tested. The mitochondrial fraction alone did not suffice to induce chemotaxis, indicating incomplete understanding of cooperative components.
• The study relies on an in vitro 3D uniaxial loading model; in vivo mechanical and tissue complexities were not fully replicated.