STX4 Is Essential for Maintaining Healthy Energy Factories in Skeletal Muscle

STX4 floxed mice (STX4^fl/fl) were obtained from Dr. Sidney Whiteheart (University of Kentucky) [S1] and crossed with heterozygous human skeletal actin (HSA)‐rtTA/TRE‐Cre recombinase positive mice (Cre^+/−) (JAX, #012433) to produce doxycycline (Dox)‐inducible STX4^fl/fl;Cre^+/− mice (C57BL6/J background). Male and female skmSTX4‐iKO mice were generated by providing 2‐mg/mL Dox in drinking water for 14 days for Stx4 gene excision [8]. Non‐Dox inducible control (CTRL) mice for skmSTX4‐iKO experiments were on the STX4^fl/fl;Cre^−/− background and provided 2‐mg/mL Dox in the drinking water. All mice were housed on Pure‐O'Cel bedding in groups of three to five and fed a standard maintenance chow diet (13% of kCal from fat; Picolab #5053, Fort Worth, TX) starting at 4 weeks of age to maintain a non‐obese phenotype. All experimental procedures were approved by the Institutional Animal Care and Use Committee of City of Hope (Duarte, CA, USA; protocol #23041).

Whole protein lysates of cell cultures and primary mouse tissue were lysed as previously described [8]. Per standard procedures, protein lysates were resolved using in‐house generated 10%–12% SDS‐PAGE gels and transferred onto polyvinylidene difluoride (PVDF) membranes for immunoblotting. Membranes were incubated with primary antibodies (Table S1) overnight at 4°C. Secondary antibodies Goat Anti‐Rabbit IgG (H L)‐HRP Conjugate (Bio‐Rad, #172‐1019) or Goat Anti‐Mouse IgG (H L)‐HRP Conjugate (Bio‐Rad, #172‐1011) were incubated for 1 h at room temperature. Protein bands were detected using a Chemic‐Doc Touch gel documentation system (Bio‐Rad) with Amersham ECL (Cytiva, #RPN2106), ECL prime (Cytiva, #RPN2232) or Super Signal (Fisher Scientific, #34095). Protein levels were quantified using Image Lab 6.1 (Bio‐Rad) and normalized against tubulin housekeeping protein levels or ponceau staining for total protein lysate.

20‐week‐old CTRL and skmSTX4‐iKO mice underwent intraperitoneal insulin tolerance testing (IPITT) as previously described [8].

20‐week‐old CTRL and skmSTX4‐iKO mice underwent intraperitoneal glucose tolerance testing (IPGTT) as previously described [8].

Following an initial overnight acclimatization, 26‐week‐old CTRL and skmSTX4‐iKO male mice were metabolically phenotyped with data collected every hour starting from 6 AM across 72 h in an indirect calorimetry cage system (Promethion, Sable Systems, Las Vegas, NV, USA) that measured oxygen consumption (VO₂), CO₂ production (VCO₂), to determine respiratory exchange rate (RER), energy expenditure, food and water consumption.

OCR measurements of flexor digitorum brevis (FDB) muscle harvested from both legs of 20‐week‐old CTRL and skmSTX4‐iKO male mice were performed as previously described [8].

Forelimb grip strength of the CTRL and skmSTX4‐iKO male mice was performed using a grip strength meter (Harvard Apparatus) with three repeated measurements for each mouse tested, as previously described [S2].

Preparation, fixation, embedment and imaging of soleus and Tibialis anterior (TA) muscle from skmSTX4‐iKO mice using FEI Tecnai 12 transmission electron microscope equipped with a Gatan OneView CMOS camera was performed as previously described [8]. Mitochondrial area in four to six 6500× TEM images per mouse was calculated using Image J 1.54p software set at 600 distance in pixels: 1‐μm scale bar.

Citrate synthase activity was measured in 2 μg of gastrocnemius (GAS) homogenate as per the manufacturer's instructions using the MitoCheck Citrate Synthase Activity Assay Kit (Cayman Chemicals, #701040) as previously described [S3].

Rat immortalized L6.GLUT4myc skeletal muscle cells (Kerafast #ESK202‐FP) myoblast growth and myotube differentiation were performed as previously described [8, 9]. For insulin signalling, myotubes were serum‐ and glucose‐starved in oxygenated FCB buffer (125‐mM NaCl, 5‐mM KCl, 1.8‐mM CaCl₂, 2.6‐mM MgSO₄, 25‐mM HEPES, 2‐mM pyruvate and 2% [wt/vol] BSA [bovine serum albumin]) (Fisher Scientific, #BP9704100) for 2 h prior to 10‐nM insulin (Sigma‐Aldrich, #I5500) stimulation for 5 min.

For STX4 knockdown studies, differentiation day (D)4 L6.GLUT4myc myotubes were transfected with 100‐nM STX4‐targeting siRNA (siSTX4) (ThermoFisher, Assay ID: s135769) sequence or non‐specific targeting siRNA control (siCON) (ThermoFisher, #4390843) in a dropwise fashion using RNAiMax (ThermoFisher, #13778075) and Opti‐MEM (Gibco, #31985088) overnight, and fresh media was added the next day as previously described [8]. These myotubes underwent 96‐h total siSTX4 knockdown to ensure optimal reduction of STX4 protein abundance.

L6.GLUT4myc myoblasts were cultured in 60‐mm dishes and underwent myotube differentiation and 96‐h siCON or siSTX4 transfection at the D4 stage as mentioned above (see 2.11). For the final 6 h, the siCON and siSTX4 L6.GLUT4myc myotubes were incubated in 5‐mM [U‐¹³C]glucose (Sigma, #389374) labelled glucose‐free DMEM (ThermoFisher, #11966025) alongside a non‐radioactive labelled 5‐mM glucose (Sigma, #G6152) DMEM control for the analysis of glycolytic and tricarboxylic acid (TCA) cycle metabolites as previously described [8].

Total DNA isolation from rat L6.GLUT4myc myotubes and mouse GAS muscle was performed per DNeasy Blood and Tissue Kit instructions (Qiagen, #69504). Mitochondrial DNA (mtDNA) copy number was calculated as the ratio of mtDNA represented by respective primers (Integrated DNA Technologies [IDT]) for mitochondrial genome‐encoded genes compared to nuclear genome‐encoded reference genes using the QuantiTect SYBR Green PCR Kit (Qiagen, #204143) for real‐time qPCR. Rat and mouse primers are described in (Table S2) and were previously used [8, 9].

Total RNA isolation from L6.GLUT4myc cell line was performed per RNeasy Plus Mini Kit instructions (Qiagen, #74136) and from mouse primary tissue with TriReagent (Sigma, #93289) as previously described [8]. Complementary DNA (cDNA) was generated per High‐Capacity cDNA Reverse Transcription Kit instructions (ThermoFisher, #4368814). The QuantiTect SYBR Green PCR Kit and primers (IDT, Table S2) were used for qPCR detection. Relative gene expression was quantified against the validated respective housekeeping gene controls for each model (IDT, Table S2).

L6.GLUT4myc myoblasts at 80% confluence underwent a 6‐h transfection with mKeima‐Red‐Mito‐7 (mt‐Keima) plasmid DNA, a gift from Michael Davidson (Addgene, #56018) via dropwise fashion of L6 Avalanche (EZ Biosystems, #EZT‐L600‐1) and Opti‐MEM according to the manufacturer's instructions. This was subsequently followed by a 42‐h transfection with siCON or siSTX4 via RNAiMax and Opti‐MEM. The cells were then exposed to ±10‐μM carbonyl cyanide 3‐chlorophenylhydrazone (CCCP) (ThermoFisher, #L06932‐MC), a known mitochondrial uncoupler [S4–S5] for 4 h to promote mitophagy. Cells were imaged using a Leica LSM 900 confocal fluorescence microscope (Zeiss) at 40× objective for quantification of mean fluorescence intensity using Zen Blue (Zeiss) on lysosomal‐associated mitochondria (pH 4.5, ex: 561 nm) [S5]; nuclei were stained by NucBlue (ThermoFisher #R37605).

The siCON‐ or siSTX4‐treated L6.GLUT4myc myotubes were exposed to ±10‐μM CCCP for 4 h and then treated with 100‐nM MitoTracker Red (ThermoScientific, #M7510) and 50‐nM LysoTracker Green (ThermoScientific, #L7526) for 30 min at 37°C in a 5% CO₂ incubator. The myotubes were imaged using an LSM 900 fluorescence confocal microscope using a 63X oil‐immersed objective; mitochondria (red)–lysosome (green) colocalization (yellow) was quantified using Pearson's coefficient in Zen Blue (Zeiss); nuclei were visualized with NucBlue.

Statistical analyses were performed using GraphPad PRISM (San Diego, CA, Version 10.1.0). Prior to statistical analysis, datasets underwent normalization testing using the Shapiro–Wilk, with normalized data established at p > 0.05 and non‐normalized at p < 0.05. Unpaired two‐tailed t‐test (normalized) or Mann–Whitney U test (non‐normalized) was used for single comparisons between genotype (CTRL vs. skmSTX4‐iKO) and cells (siCON vs. siSTX4). One‐way ANOVA with Tukey's multiple comparison was used for the quantification of mitophagy in cells, and two‐way ANOVA with uncorrected Fisher's LSD test, multiple comparison t‐test was used for OXPHOS protein and mRNA expression quantification (except for isotope tracing, which used Šídák's multiple comparison test). A multiple comparison t‐test was used for line graphs. The area over the curve (AOC) for IPITT and the area under the curve (AUC) for IPGTT, RER and energy expenditure were analysed using an unpaired two‐tailed Student's t‐test. Graphs illustrate individual values, means and standard deviation (SD) unless otherwise indicated, with statistical significance of p < 0.05 of minimum n = 3 animals or independent passages per group.

We generated skmSTX4‐iKO mice using a skeletal muscle‐specific gene knockout model as previously used [8, 19] by crossing Dox‐inducible HSA‐rtTA/TRE‐Cre^+/− and Stx4^fl/fl construct lines to excise the mouse Stx4 gene. Compared to Dox‐treated noninducible single Stx4^fl/fl;Cre^−/− (CTRL) mice, we observed significant depletion in STX4 protein abundance in soleus (oxidative fibre–rich), TA (glycolytic fibre–rich) and GAS (mixed fibres) skeletal muscles (Figure 1a) with no significant change in nonskeletal muscle tissues (heart and liver) (Figure 1b). Stx4 mRNA levels across these muscle groups were also decreased (Figure S1a–c). Significantly reduced gains in body weight (Figure 1c) and weights of large muscle groups such as quadriceps and GAS for the skmSTX4‐iKO male mice were also detected (Table S3). Furthermore, IPITT (Figure 1d) and IPGTT (Figure S1d) revealed significantly decreased insulin sensitivity (Figure 1d) and impaired glucose tolerance (Figure S1d) in non‐obese skmSTX4‐iKO male mice compared to age‐matched CTRL. Female skmSTX4‐iKO mice showed similar changes (Figure S2 and Table S4). These data indicate that inducible STX4 ablation selectively in skeletal muscle leads to the development of insulin resistance.

To assess the impact of skeletal muscle–specific STX4 ablation on whole body metabolism in skmSTX4‐iKO mice, we initially acclimatized CTRL and skmSTX4‐iKO male mice in metabolic caging units overnight and then captured data every hour for up to 72 h. The RER was significantly lower in skmSTX4‐iKO male mice (Figure 2a), suggesting a shift towards fat as a predominant metabolic fuel [8]. Interestingly, we observed a significant decrease in energy expenditure in the skmSTX4‐iKO male mice compared to their CTRL counterparts (Figure 2b). However, our metabolic cage measurements showed no significant differences in total food and water consumption (Figure S3). Based on previous studies in adipose tissue [S6], the decreased RER (Figure 2a) and energy expenditure (Figure 2b) could be a consequence of decreased mitochondrial activity. Thus, we next assessed mitochondrial function via OCR in isolated primary FDB myofibers from CTRL and skmSTX4‐iKO mice [8]. Notably, we found a significant reduction of OCR in the skmSTX4‐iKO myofibers that flatlines across the basal and maximal respiration OCR phases compared to the CTRL (Figure 2c). Next, we assessed whether the skeletal muscle mitochondrial dysfunction presented in skmSTX4‐iKO male mice was associated with weakened physical performance, that is, muscle strength, as observed in conditions like aging [S7] and diabetes [S8]. Thus, we measured forelimb grip strength and found that skmSTX4‐iKO male mice presented ~25% reduced grip strength as compared to their CTRL counterparts (Figure 2d). These findings suggest that STX4 ablation in skeletal muscle impedes whole body energy efficiency and physical performance through impaired mitochondrial function.

As STX4 is located at both the sarcolemma and OMM in skeletal muscle [8] and its loss negatively impacts mitochondrial function, we next tested whether STX4 ablation impacts changes in skeletal muscle mitochondrial structure. To assess mitochondrial structure, we analysed mitochondria in CTRL and skmSTX4‐iKO mouse soleus (oxidative slow twitch) and TA (glycolytic fast twitch) by TEM, as previously described [8]. In the soleus of skmSTX4‐iKO male mice, smaller and more fragmented mitochondria were observed, with reduced mitochondrial area, as compared to CTRL (Figure 3a). In TA of skmSTX4‐iKO male mice, a larger mitochondrial area was attributed to a bloated, swollen, damaged mitochondrial structure with abnormal cristae, as compared to CTRL (Figure 3b). Despite being non‐obese, these skmSTX4‐iKO mice presented mitochondrial damage reminiscent of those observed in HFD‐exposed insulin‐resistant obese mice [8]. Given the differences observed in mitochondrial damage between the oxidative‐rich (soleus) (Figure 3a) and glycolytic‐rich (TA) (Figure 3b) skeletal muscle in skmSTX4‐iKO male mice, we next evaluated changes in electron transport chain (ETC) complexes between these two different muscle types. We observed a significant decrease in ETC complexes I, II and IV in oxidative fibre–rich skmSTX4‐iKO soleus muscle compared to CTRL mice (Figure 3c). In contrast, ETC complexes in the glycolytic fibre–rich skmSTX4‐iKO TA muscle were unchanged (Figure 3d), suggesting that STX4 ablation has a greater impact on ETC protein abundance in oxidative fibre types. We found a similar mitochondrial phenotype and ETC heterogeneity between soleus and TA muscle in female skmSTX4‐iKO mice (Figure S4). These data demonstrate that STX4 depletion impairs skeletal muscle mitochondrial structure, supporting that not only is STX4 necessary for the mitochondria in skeletal muscle but also displays differential impacts on oxidative versus glycolytic muscles.

With the damaged mitochondrial phenotype in STX4‐depleted muscle, we next assessed citrate synthase activity, a marker of the TCA cycle in glucose metabolism [8]. However, we observed no significant difference in the GAS muscle between CTRL and skmSTX4‐iKO male mice (Figure 4a). With the TCA cycle being a dynamic process, we next utilized a live culture system of the rat L6.GLUT4myc myotube cell line as successfully used in previous studies on mitochondrial activity [8, 9]. Following a paradigm of 96‐h siRNA transfection, we found that siSTX4 L6.GLUT4myc myotubes showed significant decreases in ETC I, II and IV complexes compared to siCON (Figure 4b,c). With decreases in these ETC complexes, we next assessed the impact of STX4 ablation on the TCA cycle for glucose metabolism via incubation with [U‐¹³C]glucose labelled media, as previously performed [8]. However, we observed no significant difference in the levels of glycolytic (Figure 4d) or TCA (Figure 4e) metabolites between siSTX4 and siCON L6.GLUT4myc myotubes, suggesting that STX4 ablation has no impact on the TCA cycle for mitochondrial glucose metabolism in skeletal muscle. Similarly, we found STX4 ablation does not alter markers of lactate conversion (glycolysis), fatty acid oxidation or canonical insulin signalling in skeletal muscle (Figure S5). Overall, our data suggest that the effect of STX4 ablation on inducing insulin resistance and mitochondrial dysfunction does not involve glycolytic and mitochondrial glucose metabolism in skeletal muscle.

We previously established that STX4 plays an important role in mitochondrial dynamics via its interaction with mitochondrial fission protein Drp1 [8]. Given the coordinated nature of MQC [5, 6], STX4 may function in other mitochondrial processes, such as mitochondrial biogenesis. Thus, we assessed mtDNA copy number via qPCR of mtDNA levels and found significantly depleted mtDNA copy number in skmSTX4‐iKO mouse muscle (Figure 5a). To determine whether this decrease in mtDNA copy number was attributed to impaired mitochondrial biogenesis, we next evaluated PGC1‐α, a well‐established master regulator of mitochondrial biogenesis and key mitochondrial biogenesis regulators NRF1 and TFAM [5, 6, 7]. Given the short half‐life of PGC1‐α [9] and NRF1 [S9–10] proteins, we assessed mitochondrial biogenesis genes' mRNA levels and found a significant decrease of Ppargc1a (PGC1‐α) and Tfam (Figure 5b) in oxidative soleus muscle from skmSTX4‐iKO mice only, with no significant difference in Nrf1 expression (Figure 5b). However, we found no significant difference in mitochondrial biogenesis genes in glycolytic TA muscle (Figure S6). Similarly, our siSTX4 L6.GLUT4myc myotubes presented significant decreases in mtDNA copy number (Figure 5c). Interestingly, the mRNA levels across all biogenesis genes (Figure 5d) and protein abundances for PGC1‐α and NRF1 (Figure 5e,f) were reduced in siSTX4 L6.GLUT4myc myotubes, which could be attributed to the key role mitochondrial biogenesis plays in establishing the mitochondrial network in myotube differentiation [20, 21]. Taken together, these findings suggest that STX4 ablation reduces mtDNA copy number and leads to mitochondrial biogenesis dysregulation.

Mitochondrial biogenesis is interlinked with mitochondrial turnover by mitophagy [5, 6, 7]. Thus, we next evaluated the requirement for STX4 in mitophagy. Being a dynamic process, we initially assessed mitophagy using live cell imaging by dual transfection of siCON or siSTX4 L6.GLUT4myc myoblasts with mt‐Keima plasmid, a mitochondrial targeting dual excitation protein to distinguish non‐lysosomal‐associated mitochondria from lysosomal‐associated mitochondria undergoing mitophagy (Figure 6a) [S5]. Although we observed no significant difference in mt‐Keima red signal under basal conditions, STX4 knockdown significantly reduced mt‐Keima red signal under CCCP exposure (a mitophagy inducer) [S4–S5], consistent with a tempered mitophagic response (Figure 6b,c). To confirm that the reduced mt‐Keima red was due to impaired mitochondria–lysosomal colocalization, we used MitoTracker (red) and Lysotracker (green) reagents to mark both organelles and evaluated their colocalization (yellow) (Figure 6d). We observed reduced colocalization in siSTX4 10‐μM CCCP‐exposed L6.GLUT4myc myotubes (Figure 6e,f) versus control, suggesting that STX4 ablation in skeletal muscle impairs mitochondrial turnover by mitophagy. With one recent report suggesting a predominance of canonical PINK1‐PARKIN mitophagy in oxidative‐rich skeletal muscle [11], we assessed these proteins in both the soleus and TA muscle from skmSTX4‐iKO male mice. Interestingly, we showed reduced PARKIN (Figure 7a) and PINK1 (Figure 7b) protein abundance in both skmSTX4‐iKO soleus and TA muscle, suggesting that canonical mitophagy is impacted by STX4 ablation in both oxidative‐ and glycolytic‐rich muscle. Similarly, we observed reduced PARKIN (Figure 7c) and PINK1 (Figure 7d) protein abundance in siSTX4 L6.GLUT4myc myotubes. Altogether, our data suggest that STX4 could be required for mitochondrial turnover by mitophagy in skeletal muscle and that this is associated with the canonical PINK1‐PARKIN pathway.