The low-density lipoprotein receptor-related protein-1 (LRP1) in Schwann cells controls mitochondria homeostasis in peripheral nerves

Abstract

Following peripheral nerve injury, Schwann cell (SC) survival is imperative for successful nerve regeneration. The low-density lipoprotein receptor-related protein-1 (LRP1) has been identified as a pro-survival SC plasma membrane signaling receptor, however, the responsible mechanisms underlying SC homeostasis remain incompletely understood. Herein, we establish that LRP1 largely manages mitochondrial dynamics and bioenergetics in SCs by limiting mitochondria fission, maintaining healthy mitochondria membrane potentials, and reducing lactate production associated with peripheral sensitization. When SC LRP1 is suppressed, inner-mitochondria-linked pathways in peripheral nerve proteome are dramatically altered, and cristae integrity in unmyelinated C-fibers is compromised. SC LRP1 protected sensory neurons from mitochondrial dysfunction and modulated mitochondria-related biological pathways in the DRG transcriptome. Conditional deletion of LRP1 in SCs induces pain-related behaviors in mice without nerve injury. Results point to a significant role for LRP1 in SC mitochondrial homeostasis and advance our understanding of the sensory neuron response to alterations in SC bioenergetics.

Graphical Abstract

LRP1 regulation of mitochondrial dynamics and bioenergetics in SCs and effect on sensory neurons. LRP1 in SCs balances mitochondria fission and fusion, but when LRP1 is absent, DRP1 is increased and fission is more prevalent. Without SC LRP1, mitochondria are less efficient, larger and appear stressed. In addition to ATP production by oxidative phosphorylation, alternative energy pathways and the production of lactate increases in the absence of SC LRP1. These mitochondrial and metabolic changes in SCs are associated with ultrastructural aberrations in C-fiber mitochondria and alterations in the neuroenergetic DRG transcriptome that likely contribute to painful peripheral neuropathy.

1. Introduction

Painful peripheral neuropathy is highly prevalent in adults and is a leading cause of disability worldwide. The burden associated with painful peripheral neuropathy increases with age and poor health. Often painful peripheral neuropathy is neuropathic, a particularly severe type of chronic pain. Over the past decades, peripheral glia, namely Schwann cells (SCs), have been identified as potential novel cellular targets for painful peripheral neuropathy treatment. However, there are inconsistencies in our understanding of the molecular mechanisms underlying SC function in pain outcomes. When SCs are dysfunctional due to a loss of Erb2 receptor (Chen et al., 2006), loss of GluN1 subunit of the NMDAR receptor (Brifault et al., 2020) or loss of GABA-B receptor (Faroni et al., 2014), aberrations in Remak bundles and pain processing are observed. Yet recently, functional SCs have been shown to play a direct role in pain transmission, forming a nociceptive glial-neural complex in nociceptive nerves (Abdo et al., 2019). Thus, a comprehensive molecular understanding of SC function with sensory axons is critical for developing effective, non-addictive therapies and treatments for painful peripheral neuropathy.

The low-density lipoprotein receptor-related protein-1 (LRP1) is an endocytic and cell-signaling receptor for structurally and functionally diverse ligands (Strickland et al., 2002, Gonias and Campana, 2014). The effects of LRP1-initiated cell-signaling are distinct in different cell types and include roles in microglia neuroinflammation (Brifault et al., 2019), innate immunity (Mantuano et al., 2017), neuronal homeostasis (Herz and Bock, 2002, Liu et al., 2015), and recently in astrocytes (Zhou et al., 2024). In peripheral nerve injury, LRP1 is substantially upregulated in SCs (Campana et al. 2006) and facilitates many well-described changes in SC physiology that are associated with activation of the SC Repair Program (Arthur-Farraj et al., 2012). These include promoting SC survival and migration via PI3K/Akt and ERK1/2 cell signaling (Campana et al., 2006, Mantuano et al., 2008, Mantuano et al., 2011). These LRP1-induced activities during injury likely contribute to the SC repair phenotype (Flütsch et al., 2016). Accordingly, conditional deletion of LRP1 in SCs results in failed sensory recovery after traumatic peripheral nerve injury (Orita et al., 2013). Activation of LRP1 may improve painful peripheral neuropathy (Garcia-Fernandez et al., 2021, Wang et al., 2022), while reduced expression of Lrp1 may serve as a biomarker for diabetic peripheral neuropathy diagnosis and progression (El-Horany et al., 2019).

Mitochondria are dynamic organelles that have significant roles in the metabolic regulation of cells based on their ability to generate adenosine triphosphate (ATP), the main cellular energy source (Miettinen and Björklund, 2017, Murphy et al., 2016). Mitochondria morphology is compartmentalized, interactive, dynamic, and mobile. Structurally, they harbor two distinct membranes: the outer and inner membranes that define functionally diverse compartments. Imbalances in metabolic homeostasis are greatly associated with human diseases, and greater than 70% of humans with inherited mitochondrial deficits develop chronic pain (van den Ameele et al., 2020). However, most metabolic studies linked to neuropathic pain states have only focused on mitochondria aberrations in the cell bodies and axons of sensory neurons, while changes in SC mitochondria have not been revealed. Thus, the role of SC mitochondria is an underestimated contributor to the abnormalities observed in painful peripheral neuropathies, despite some studies indicating that SC-specific mitochondrial dysfunctions could be sufficient to cause or contribute to neuropathic and chronic pain conditions (Viader et al., 2011, Ino and Iino, 2017). While mitochondrial proteins such as Prohibitin 1 (Phb1) have been conditionally deleted in SCs (Della-Flores Nunes et al., 2021) and lead to aberrant developmental myelination and demyelinating neuropathies in adult mice, very little is known about SC mitochondrial homeostasis regulated by cell surface signaling receptors that are linked to neuropathic pain states.

In this study, we discovered that the endocytic and cell signaling receptor, LRP1, is a potent regulator of mitochondrial dynamics and bioenergetics in SCs. LRP1 maintains healthy mitochondria membrane potentials and reduces the production of extracellular lactate, which has the potential to damage and sensitize peripheral nerves. Indeed, nerves without SC LRP1 have elevated superoxide dismutase 1 (SOD1), suggesting alterations in nerve redox states (Wang et al., 2018). We demonstrate that SC LRP1 modulates the size and density of mitochondria in neighboring small unmyelinated C-fibers, and when LRP1 is ablated in SCs, mitochondria in the cell bodies of DRG neurons are dysfunctional. These results are consistent with alterations in the mitochondrial peripheral nerve proteome and DRG transcriptome. Our novel findings enhance understanding of SC mitochondria homeostasis and how LRP1 regulates SC bioenergetics, the importance of SC homeostasis in maintaining sensory neuron mitochondrial functions, and strengthen the role of SC mitochondria in developing neuropathic pain states.

2. Materials and Methods

2.1. Animals

We crossed mice bearing a floxed Lrp1 gene (originally described by Rohlmann et al., 1996), with mice expressing Cre recombinase underthe control of the Mpz promoter (Feltri et al., 1999). This allowed us to generate mice in which Lrp1 was deleted specifically in SCs (Lrp1fl/fl referred to as scLRP1^+/+; Lrp1fl/fl:Mpz-Cre—and scLRP1^−/−) (Orita et al., 2013, Poplawski et al., 2018). Recombination in sciatic nerves of scLRP1^−/− animals was confirmed by mass spectroscopy (Fig. 3D) and resulted in a significant reduction of Lrp1 protein (Fig. S2B and S2C). The P0 promoter used in our studies is reported to be inactive in CNS glia and in DRG neurons (Feltri et al., 1999, Feltri et al., 2002) and confirmed by us (Brifault et al., 2020). Breeding procedures and animal experiments were performed according to protocols approved by the Veterans Administration (VA) and conform to VA Guidelines for Animal Use. All animal experiments performed were also approved by the Institutional Animal Care and Use Committee (IACUC) of the University of California San Diego (UCSD) and New York University (NYU) and were conducted strictly under the guidelines for animal experimentation of UCSD and NYU. Mice were housed using a 12 h:12 h light:dark cycle with ad libitum access to food and water.

Fig. 3. The sciatic nerve proteome reveals alterations in bioenergetic pathways when SC LRP1 is deleted.

A Principal Component (PC) analysis of 8 female sciatic nerve samples divided into 2 groups (4 scLRP1^+/+ reported in blue, and 4 scLRP1^−/− reported in red) by LC-MS/MS using an LTQ Orbitrap and available SEAQUEST software. B Proteomic analysis of the results defined 3 groups of proteins. The largest group of proteins was identified in both groups. C Differential abundance analysis using Limma quantified the differences between proteins found in both groups, and the results were summarized using a volcano plot. Proteins with log fold change (LogFC) > 0 are up in scLRP1^+/+ (compared to scLRP1^−/−) with a negative value representing down in scLRP1^+/+. The proteins with an adjusted p < .05 (LRP1) are colored in red, while those with a nominal p-value < .05 are colored in orange. Black points are not significant. D Abundance of LRP1 in the two genotypes. Individual points represent the intensity in each sample on a log scale. E A functional enrichment analysis using the combination of scLRP1^+/+ only, scLRP1^−/− only, and differentially abundant (p < .01) proteins was performed on all metabolic pathways in 3 pathway databases: KEGG, Reactome (REAC), and WikiPathways (WP). The dot plot shows the significantly enriched pathways with a negative log of the false discovery rate (-Log(FDR)) on the x-axis. The size of the dots represents the overlap between the pathway proteins and the merged input list. F A network representation of the enriched metabolic pathways (triangles) and proteins (circles) from those pathways in the merged list of scLRP1^+/+ only, scLRP1^−/− only, and differentially abundant proteins. These edges are supplemented with edges between proteins from the medium confidence STRING interactome. Protein nodes are colored by log fold change (for the differentially abundant proteins) or only identified in the scLRP1^+/+ (red) or scLRP1^−/− (blue) group. G Immunoblotting of uninjured sciatic nerves isolated from scLRP1^+/+ and scLRP1^−/− mice (N=3 per genotype) for MT-CO1 (37 kDa), COX7C (10 kDa), and ATP5MK (7 kDa). Calnexin (90 kDa) served as a loading control for each blot.

2.2. Pain behavioral assays

In all pain-related behavior assays, male and female mice (4–5 months) were utilized. Mechanical sensitivity. Mice were placed in a plexiglass box (22×22×14 cm) on a metal grid floor and acclimated to the behavior testing facility for at least 60 min. Mechanical sensitivity (tactile allodynia) was tested by applying 0.04 to 4 g von Frey filaments (Stoelting) to the plantar surface of the left hind paw. Filaments were presented consecutively ascending or descending using the up-down method previously described and modified for the mouse (Poplawski et al., 2018, Chaplan et al., 1994). The filament that caused paw withdrawal 50% of the time (the 50% PWT) was determined. All experiments were performed by an investigator blinded to the mouse identity. Thermal hyperalgesia. Mice were placed in a plastic chamber on a 25 °C glass surface. Thermal hyperalgesia was measured by the paw thermal stimulator (IITC Life Sciences) with a radiant heat source delivering a thermal stimulus to the left hind paw of each mouse with a cutoff of 20 seconds (Cheah et al., 2017). Paw withdrawal latency was measured as a mean of 3 trials at 5-minute intervals. Cold allodynia. The tests were performed as previously described (Toma et al., 2017) with modifications. Mice were placed in a plexiglass box on a metal grid floor and acclimated for at least 30 minutes before testing. 20μL of acetone was projected onto the plantar surface of the hind paw from the pipette tip. Time spent licking, shaking, and holding the hind paw was recorded by a stopwatch for 2 minutes. Cold response. Cold hyperalgesia was measured using a well-established protocol (Brenner et al., 2012). Mice were kept in a restrainer box with a transparent glass sheet. A dry ice pellet was pressed against the glass underneath the hind paw with consistent pressure using a syringe. The withdrawal latency (paw withdrawal away from the glass) was recorded with a stopwatch. The maximum time allowed for withdrawal was 20 seconds to avoid potential tissue damage. Nesting behavior. The tests were performed as previously described (Deacon, 2006). Briefly, mice were single-housed one hour before the dark phase without enrichment. Each cage was given 3.0 g of Nestlet (#NES3600, Ancare). The nests were assessed after 14 hours on a rating scale of 1–5 as follows: 1 = more than 90% Nestlet intact; 2 = 50%-90% Nestlet intact; 3 = Nestlet mostly shredded, but not identifiable nest site; 4 = an identifiable but flat nest (the walls of the nest that are higher than the curled-up mouse body height for less than 50% of the circumference); 5= a near perfect nest, over 90% (nestlets are torn and the nest has walls higher than the mouse body height for more than 50% circumference). Inner Zone Duration. Anxiety-like behaviors were evaluated by a Behavioral Spectrometer that is capable of recording different types of behaviors, including time spent in the inner zone (Brodkin et al., 2014). The Behavioral Spectrometer comprises a 40 cm² arena with a CCD camera mounted in the center of the ceiling and a door aperture in the front area of the arena. Mouse movement and behaviors are assessed by a floor-mounted vibration sensor and 32 wall-mounted infrared transmitter and receiver pairs. Mice were individually placed in the center of the behavioral spectrometer, and their behavior was recorded, tracked, evaluated, and analyzed using a computerized video tracking system (Viewer3, BiObserve, DE) for 15 min. The arena area was divided into an inner zone and outer zone, and time spent in the inner zone was determined.

2.3. Transmission electron microscopy (TEM) and morphological assessments

Male and female mice (4 months) were deeply anesthetized with isoflurane and subjected to intracardiac perfusion with fresh 20 mM sodium phosphate, 150 nM NaCl, pH 7.4 (PBS), followed by glutaraldehyde (2.5%) in PBS. Sciatic nerves (5 mm) were collected 2 mm distal from the DRG. Collected nerves were prepared for TEM, as we described previously (Brifault et al., 2020). Briefly, nerves were immersed in modified Karnovsky’s fixative (2% glutaraldehyde in 0.10 M sodium cacodylate buffer) and further post-fixed in 1% osmium in 0.15 M cacodylate buffer for 1 hour on ice. Tissues were stained with 2% of uranyl acetate (UA) for 1 hour at 4 °C and then were dehydrated in different concentrations of ethanol (EtOH) (50%, 70%, 90%, and 2 times in 100%) for 10 minutes per concentration on ice. Tissues were then washed with acetone and embedded in Durcupan epoxy resin. Ultrathin sections (60 nm) were cut on a Leica UCT microtome with a diamond knife and were post-stained with 2% uranyl acetate for 5 min and Sato’s lead stain for 1 min. Nerve sections were placed on 300 mesh copper grids for imaging; these grids were viewed using a JEOL 1400 plus Transmission Electron Microscope at 80kV with a Gatan 4kx4k camera. Grids were viewed with the nerves being imaged at 2000x, 5000x, 15000x, and 30000x magnification.

For each genotype and sex, at least three animals were processed and analyzed. For each animal, over 825 axons in the sciatic nerve were imaged and quantified. For each sciatic nerve, ~10–12 random images at 2000x were used for assessment of all morphological parameters. To quantify mitochondria area and density, ~6 images at 15,000x were randomly selected and at least 50 mitochondria per animal were measured using OneNote and Image J Fiji v2.0 as described previously (Dell-Flora Nunes et al., 2021). Only individual fibers and Remak bundles with distinct structures were included in the quantification.

2.4. LC-MS/MS and analysis by mass spectrometry interaction statistics

Sciatic nerves were isolated from female mice and ophthalmic forceps were used to remove the epineurium. Nerves were rinsed in PBS to remove the blood and snap frozen in cryogenic storage tubes. Fractured tissue was transferred to a 1 mL milliTUBE containing an AFA fiber (#520135, Covaris) in 200 μL of 50 mM HEPES pH 8.5, 150 mM NaCl, and 2% Triton X-114 and sonicated with a M220 Focused-Ultrasonicator (#500295, Covaris) for 600 seconds. Extracted proteins were clarified of insoluble material by centrifugation at 15,000 x g for 20 minutes at 4 °C. Protein concentrations were determined with the Micro BCA colorimetric assay (#23235, ThermoFisher Scientific). Aliquots containing approximately 5 μg of protein were processed using the SP3 protocol as previously described (Hughes et al., 2019, Zhang et al., 2023). Briefly, the sample aliquots were brought to 50 μL volume, and disulfide bonds were reduced and alkylated simultaneously with 10 mM TCEP, 40 mM 2-chloroacetamide in 50 mM HEPES pH 8.5, and 1% sodium deoxycholate at 70 °C for 10 minutes, then cooled on ice. Proteins were precipitated and captured following the addition of 10 μL of a washed 10 μg/μL suspension of SpeedBeads (#44152105050350, Cytiva) and 400 μL of ethanol. After shaking for 10 minutes at room temperature, the beads were magnetically captured and washed 3 times with 200 μL of 80% ethanol in water. Proteins were digested on the beads in 50 μL of 50 mM HEPES pH 8.5, 1% sodium deoxycholate, and 10 ng/μL trypsin (#V5280, Promega) overnight at room temperature with shaking sufficient to maintain the beads in suspension. The digest was diluted 10-fold with 80% acetonitrile and 1% formic acid, then separated from the beads magnetically, and the resulting peptides were captured on 2 mm discs of Empore Cation (CDS Analytical) fitted into 1,000 μL pipette tips (#791000, Sartorius). Detergents and other contaminants were removed by washing the tips serially with 1) ethyl acetate; 2) 80% acetonitrile, 1% formic acid; and 3) 10% acetonitrile, 0.2% formic acid. Peptides were eluted directly into injection vials with freshly prepared 80% acetonitrile and 5% ammonium hydroxide and immediately dried down in a centrifugal vacuum evaporator. One-fifth of the recovered peptides from each sample were subsequently analyzed by liquid chromatography-tandem mass spectrometry. In-house capillary columns were constructed from 360 μm OD and 100 μm internal diameter × 30 cm fused silica tubing (#987651–5582, Molex) with laser-pulled tips (#P-2000, Sutter Instruments) and were packed with Reprosil-PUR 3 μm C18-AQ (#rs13.9tde, Dr. Maisch GmBH). Solvents A and B consisted of 0.1% formic acid in water and 80% acetonitrile with 0.1% formic acid, respectively. A 180-minute linear gradient from 2% to 35% solvent B was used for chromatographic separation. Peptides were analyzed with an Orbitrap Elite (ThermoFisher Scientific) mass spectrometer using nano-electrospray ionization with an applied voltage of 1,800 V. MS1 spectra were acquired at a resolution of 120,000, and the 15 most abundant precursor ions were selected for fragmentation by higher energy collision dissociation. MS2 spectra were acquired at a resolution of 15,000. Dynamic exclusion parameters were a list size of 500, a mass window of ±7 ppm, and a duration of 1 minute. Automatic gain control settings were MS1 target 1 × 106, maximum inject time 100 ms; MS2 target 4 × 104, maximum inject time 100 ms.

Spectrum matching and protein identification, validation, and quantification were performed as described (Zhang et al., 2023, Kong et al., 2017, Yu et al., 2021). The protein database used for the search was the Mus musculus reviewed sequence database downloaded from UniProt. The results were analyzed and processed as previously described (Hughes et al., 2019). False discovery analysis was performed using the Benjamini, Krieger, and Yekutieli method in GraphPad Prism 10.0. Causal analysis of proteomic data was performed in Ingenuity Pathway Analyses (IPA) upstream analysis software (QIAGEN). For IPA, the binary values were imputed using local minimum intensities. Enrichment analyses for gene ontology (KEGG, Reactome, Wiki pathways) were performed using clusterProfiler 4.2.2 R package on R 4.1.0.

2.5. Immunoblotting

Protein concentration of tissue and cell extracts were determined by BCA assay. Protein extracts (15 or 20 μg/lane) were subjected to SDS-PAGE on 4–20% gradient gels and then electrotransferred to nitrocellulose or PVDF membranes using a BioRad Trans-Blot Turbo Transfer System (#1704150, BioRad). Membranes were blocked with blotting grade blocker (5% nonfat dry milk) (#170–6404, Bio-Rad) in 10 mM Tris–HCl, 150 mM NaCl, pH 7.5, 0.1% Tween 20 (TBS-T) for 1 h at room temperature. The primary antibodies ATP5MK (#17716–1-AP, Proteintech, 1:1000), MT-CO1 (#ab203912, Abcam, 1:1000), COX7C (#PA5–68018, ThermoFisher Scientific, 1:1000), Calnexin (ADI-SPA-860-D, Enzo, 1:1000), VDAC-1 (#ab14734, Abcam, 1:1000), COX4 (#38563, Cell Signaling, 1:1000), β-Actin (#3700, Cell Signaling, 1:1000), MFN2 (#12186–1-AP, Proteintech, 1:1000), and DRP1 (#12957–1-AP, Proteintech, 1:1000), LRP1 (#64099, Cell Signaling, 1:1000), SOD1 (#ab51254, Abcam, 1:1,000), and p-PDHA1-S293 (#ab92696, Abcam, 1:1000, used for nerve lysates), PDH (#ab110333, Abcam, 1:500, used for immunoblotting) and PDH (#3205 Cell Signaling, 1:1000; used for IF) and were diluted in 5% BSA/TBS-T or MILK/TBS-T and incubated with the membranes overnight at 4 °C. Membranes were washed with TBS-T and incubated with horseradish peroxidase (HRP)-conjugated secondary anti-mouse IgG (#7076, Cell Signaling, 1:2500) or anti-rabbit IgG (#7074, Cell Signaling, 1:2500) in TBS-T with nonfat dry milk for 1 h at room temperature followed by enhanced chemiluminescence (#32109, ThermoFisher Scientific). Blots were imaged using a BioRad ChemiDoc Imaging System.

2.6. Isolation and purification of adult mouse primary SCs

Adult mice (3–5 months) were anesthetized with 3% isoflurane (#502017, VETone), maintained with 2% isoflurane, and subjected to sciatic nerve crush injury, as previously described (Poplawski et al., 2018, Azzouz et al., 1996). Seven days later, nerves at the crush site and distal to the injury were collected and placed in HEPES. The epineurium was removed, and nerves were subsequently moved into a 35 mm dish containing Dulbecco’s Modified Eagle Medium (DMEM) (#11885, Gibco), 100 U/mL penicillin and 100 μg/mL streptomycin (P/S; #15140122, Gibco), and 200 U/mL collagenase D1 (#SCR139, Millipore Sigma), and incubated for 45 min at 37 °C in a humidified atmosphere with 5% CO₂. After centrifugation, pre-warmed 0.25 % trypsin (#15140122, Gibco) diluted in PBS (#10010023, Gibco) was added to the pellet and incubated for 10 min at 37 °C in a humidified atmosphere with 5% CO₂. DMEM supplemented with 10% heat-inactivated FBS (#10-082-147, Gibco) was added, and nerves were triturated gently 10–15 times using a 10 mL syringe equipped with an 18-G and then, a 21-G needle. After centrifugation, the pellet was resuspended in preheated DMEM, supplemented with 10% FBS, 100 U/ml penicillin, 100 μg/mL streptomycin, 1 nM Neuregulin-1 (NRG1) (#9875-NR-050, R&D System), 2 μM forskolin (#3828S, Cell signaling), and cultured in 60 mm petri dish pre-coated with 100 μg/mL poly-D-lysine (#P6407, Millipore Sigma) and 10 μg/mL laminin (#23017015, Gibco). When cellular confluence was about 90%, the medium was replaced with DMEM supplemented with 10% FBS and 1% P/S, and cells were detached with a trypsin solution (0.05% trypsin/0.02% EDTA). After centrifugation, the pellet was treated with DMEM supplemented with 10% FBS, 1% P/S, and 2μg/mL anti-Thy1.2 antibody (#14-0903-82, ThermoFisher Scientific) for 20 min at 37 °C in a humidified atmosphere with 5% CO₂. Subsequently, 10 ug/mL rabbit complement (#S7764, Millipore Sigma) was added to the cell suspension for 1 hr at 37 °C in a humidified atmosphere with 5% CO₂. The cells were moved into a T25 flask and were passaged no more than three times for the studies. There were no significant metabolic or morphological changes in passage numbers 1–2.

2.7. RT-qPCR

RNA was reverse-transcribed using the iScript cDNA synthesis kit (#1708891, Bio-Rad). RT-qPCR was performed using TaqManTM gene expression products and the CFX Connect Real-Time system (Bio-Rad). The relative change in gene expression was calculated using the 2ΔΔCt method. 18S mRNA was measured as a standard. The primer-probe sets used included: 16S (#Mm04260181_s1), ND1 (#Mm04225274_s1), and 18S (#Hs99999901_s1). All probes were purchased from ThermoFisher Scientific. The relative change in mRNA expression was calculated using the 2ΔΔCT method with 18S mRNA as an internal normalizer. All results were presented as the fold-change in mRNA expression relative to controls.

2.8. Isolation of mitochondrial subcellular fraction

Mitochondria Isolation Kit for Tissue with Dounce Homogenizer (#ab110169, Abcam) was used to isolate mitochondrial fractions from cultured mouse SCs following the manufacturer’s guidance. Briefly, trypsinized SCs were collected and pelleted by centrifugation at 300 x g and subsequently were frozen and then thawed for three cycles to weaken the cell membranes. Subsequently, cells were resuspended in Buffer A and incubated for 10 min on ice, followed by Dounce homogenization. After centrifugation, the supernatant was saved, and the pellet was resuspended in Buffer B to the same volume as Buffer A, following the same rupturing steps. Combined supernatants were centrifuged at 8,000g for 15 min at 4 °C to obtain the mitochondrial fraction as a pellet. The pellet was resuspended in Buffer C. The supernatant was collected as a cytoplasmic fraction. Protein concentration was determined by BCA (#23225, ThermoFisher Scientific). Immunoblotting was used to validate the purity of enriched mitochondrial fractions with COX4 and cytoplasmic factions with β-Actin.

2.9. Primary cultured human and rat SCs and siRNA transient transfections

Human SCs (#10HU-188, iXCells Biotechnologies) were cultured in a medium (#MD-0055B, iXCells Biotechnologies) supplemented with 10% heat-inactivated FBS (#MD-0094, iXCells Biotechnologies), 1% of antibiotic/antimycotic (#MD-0095, iXCells Biotechnologies), and 0.2% of Schwann cell growth supplement (#MD-0095, iXCells Biotechnologies) at 37 °C in a humidified atmosphere with 5% CO₂.

Rat SCs were isolated from postnatal (day 1) rat pups and maintained as previously described (Sadri et al., 2022, Martellucci et al., 2024). Schwann cells were cultured in a complete medium consisting of Dulbecco’s Modified Eagle Medium (DMEM) (#11885, Gibco), supplemented with 10% heat-inactivated FBS (#10-082-147, Gibco), 100 U/ml penicillin, 100 μg/ml streptomycin, 21 μg/ml bovine pituitary extract (#P1476, Millipore Sigma) and 4 μM forskolin (#3828S, Cell signaling) at 37 °C in humidified atmosphere with 5% CO₂. To confirm the purity of the primary cultures, isolated SCs were immunostained with anti-S100 β and Dapi. Schwann cells were greater than 90% pure. Schwann cells used for all studies were passaged no more than six times.

Gene silencing in primary SCs with siRNAs was performed as previously described (Mantuano et al., 2008). Human and rat-specific ON-TARGETplus SMARTpool siRNA, targeting Lrp1, and pooled NTC siRNA, were from Horizon Discovery. SCs were transfected with siRNA by electroporation using the Cell Line Nucleofector Kit V (#VCA-1003, Lonza). Briefly, cell suspensions were combined with Lrp1-specific (200nM) or NTC siRNA (200 nM) and electroporated with an SCs-specific program in a Nucleofector 2b device (Lonza). Experiments were performed 48/56 hours after transfection when LRP1 levels were significantly reduced, as described by Martellucci et al. (Martellucci et al., 2024).

2.10. Dual-label immunofluorescence and confocal microscopy

Mouse SCs were plated in a 24-well chamber (#3526, Corning) on PDL-coated 12-mm coverslips (#1203J81, Thomas Scientific) at the density of 6.0 × 10⁴ cells/well. The day after, the cells were fixed in 4% paraformaldehyde for 10 min at 4 °C, washed with PBS, and subsequently permeabilized with 0.01% Triton X-100 for 10 min. Nonspecific binding sites were blocked for 1hr with blotting grade blocker (5% nonfat dry milk) (#170–6404, Bio-Rad). Immunostaining was performed using mouse anti-PDH (#ab110333, Abcam) conjugated with anti-mouse IgG Alexa Fluor-488 (#A-121202, ThermoFisher Scientific) and rabbit anti-Sox10 (#MA5–32335, ThermoFisher Scientific) conjugated with anti-rabbit IgG Alexa Fluor-594 (#A-21442, ThermoFisher Scientific). Coverslips were mounted on slides using the mounting medium Aqueous Fluoroshield (#ab104139, Abcam) with DAPI for nuclear labeling. Immunofluorescence was imaged using an Axio imager Z2 with calibri7 LED equipped with Plan-Apochromat 20x/0.8 M27 objective and Plan-Apochromat 63x/1.40 Oil DIC M27 objective. The cells were stimulated with the 385nm and 475nm LED at 10% intensity. 3008 × 4096 pixels images were acquired and analyzed with ZEN 3.6 (ZEN pro) software.

2.11. Respirometry measurements by Seahorse analyses

Cellular bioenergetics analysis was performed using an Agilent Seahorse XFe96 Analyzer with Seahorse XF Mito Stress Test, following the manufacturer’s instructions. Briefly, 6 × 10⁴/mm² primary cultured mouse SCs were seeded into poly-L-lysine and laminin pre-coated XF96 plates (#103010–100, Agilent) 18 h before the assay. NucBlue^™ Live ReadyProbes^™ Reagent (Hoechst 33342) (#62249, ThermoFisher Scientific) live cell stain was performed for 20 minutes at 37°C in a CO₂-free incubator for outgassing to stain nuclei for cell counting. During the assay, cells were kept in a non-buffered XF DMEM base medium (pH 7.4, no phenol red) supplemented with 1 mM pyruvate, 2 mM glutamine, and 10 mM glucose. For oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) analysis, 2 μM oligomycin was used to inhibit mitochondrial ATP synthesis, 2 μM carbonyl cyanide-p-trifluoromethoxyphenyl-hydrazon (FCCP) was used to stimulate maximal mitochondrial respiration, and 0.5 μM rotenone and 0.5 μM antimycin A were used to block mitochondrial respiration completely. Parameters of mitochondrial respiration were calculated as follows: basal respiration = (last measurement before oligomycin injection) - (minimum rate measurement after rotenone/antimycin A injection); ATP-linked respiration = (last measurement before oligomycin injection) - (minimum rate measurement after oligomycin injection); proton leak = (minimum rate measurement after oligomycin injection) - (minimum rate measurement after rotenone/antimycin A injection); maximal respiration = (maximum rate measurement after FCCP injection) - (non-mitochondrial oxygen consumption); spare respiratory capacity = (maximum respiration) - (basal respiration). Basal and maximal respiration (OCR), as well as glycolytic capacity (ECAR) values, were plotted after cell count normalization with FluxNorm⁴⁵. Primary mSC cultures (biological replicates) using five independent cultures with 4–6 technical replicates per culture were used per genotype.

2.12. Measurement of lactate levels

Lactate was measured in the conditioned medium (CM) and cellular lysates from primary mouse SC cultures. Briefly, mouse SCs were seeded in pre-coated T25 flasks with 100 μg/mL poly-D-lysine (#P6407, Millipore Sigma) and 10 μg/mL laminin (#23017015, Gibco). At 75–80% confluency, culture media was replaced with DMEM base medium (#11885, Gibco) and subsequently collected within 12 hours. Each sample was centrifuged at 10000 x g for 10 minutes and deproteinized using 10 kDa spin columns (#ab93349). To avoid Lactate degradation, cells were homogenized and treated with a Deproteinizing Sample Preparation Kit – TCA (#ab204708, Abcam). Lactate levels were measured using the Fluorometric lactate assay kit (#ab65330, Abcam). The level of lactate in the CM and cells was normalized with cellular protein concentration measured by BCA.

2.13. Mitochondrial membrane potential (MMP) assay

Mitochondrial membrane potential dissipation was monitored by measuring tetramethylrhodamine methyl ester (TMRM) fluorescence as previously described (Pavlov et al., 2010, Yu et al. 2024). The day before the experiment, 6.0 × 10⁴ SCs per well were plated in a 24-well chamber (#25–105, Genesee Scientific) with 12-mm coverslips (#1203J81, Thomas Scientific) pre-coated with poly-D-lysine and laminin as described above. On the day of the experiment, the cells were washed twice with PBS and DMEM with Mitotracker 100 nM (#M7514, ThermoFisher Scientific) and Tetramethylrhodamine (TMRM) 15 nM (#T668, ThermoFisher Scientific) were added for 30 mins at 37 °C. Subsequently, the cells were washed twice with Hibernate E 1x (#HELF, TransnetYX), and a solution of Hibernate E 1x + Glucose 1g/L + TMRM 1 nM was added to maintain the cells until the live cell imaging. 2 μM FCCP (#C2920, Millipore Sigma) was used as positive control. Live cells were imaged using an Axio imager Z2 with calibri7 LED equipped with Plan-Apochromat 20x/0.8 M27 objective and Plan-Apochromat 63x/1.40 Oil DIC M27 objective. Independent mouse SC cultures were quantified for each genotype. Quantification analyses were performed based on integrated density (24 cells/culture) and regions of interest (ROIs) (Wolken and Arriaga, 2014). For each image, 3 regions of interest (ROI) were selected and the TMRM and Mitotracker Green fluorescence intensity of each ROI was measured using ImageJ Fiji 1.54f. The Fire Lookup Table (LUT) was applied using ImageJ Fiji 1.54f for TMRM immunofluorescence. The TMRM and Mitotracker Green ratio was calculated and plotted using GraphPad Prism 10.

2.14. Primary dorsal root ganglion neuron (DRG) cultures

Adult mice (3–5 months) were anesthetized with inhalational isoflurane. DRGs were collected and purified as described before (Malin et al., 2007, Ye et al., 2014) in an artificial cerebrospinal fluid (aCSF) medium. Isolated DRG neurons were seeded onto the 25-mm-diameter coverslip pre-coated with laminin (15 μg/ml) and poly-D-lysine (20 μg/ml) and collected for RNA-Seq analyses after 24 hours.

2.15. RNA-sequencing and data analysis

L4 and L5 DRGs from the left and right sides of each mouse were isolated, pooled, and snap-frozen in liquid nitrogen (N = 3 per genotype). RNA was extracted with the AllPrep DNA/RNA Micro Kit (#80284, QIAGEN). RNA quality was assessed on an Agilent 2100 Bioanalyzer. Samples with RNA integrity numbers ≥8 were used for RNA sequencing (RNA-seq). RNA-seq was conducted at the NYU Genome Technology Center. All samples were processed in the same time period, following the same protocol to limit batch effects and other confounders using the Illumina RNA-seq platform. The cDNA library was prepared via the standard Illumina protocol. Three samples were pooled into each lane using standard protocols and sequenced by 75-bp paired-end sequencing on an Illumina HiSeq 2500. A Phi-X positive control provided by Illumina was spiked into all lanes at a concentration of 0.3% to monitor sequencing quality. The sample error rate was < 2%, and the distribution of reads per sample in a lane was within a reasonable tolerance. Data generated during sequencing runs were transferred to the high-performance computing (HPC) cluster, with individual base calls transferred for downstream analysis. The target for average reads per sample was approximately 25 million. The QC pipeline included: 1) quality check of the raw sequencing data using FastQC (v 0.11.9) and MultiQC (v 1.9); 2) mapping the sequencing reads to the human genome (build 102) using HISAT2 (v 2.2.1), followed by SAMtools (v 1.12) to convert BAM (Binary Alignment Map) into SAM (Sequence Alignment Map) files; 3) assembly of RNA-seq reads into transcripts using StringTie (v 2.1.4); and 4) calculation of expression levels from read counts, producing a gene count matrix. To improve the rigor of gene expression analysis, a minimum read depth of 10 gene counts in at least 90% of samples was required for retention in the matrix. The Ensembl identifiers (ID) of the gene counts were annotated to Entrez IDs using the EnrichmentBrowser (v.2.18.2) package in R. The Entrez IDs were annotated to gene symbols using Homo sapiens (v.1.3.1). Gene expression was compared using DESeq2 (v.1.32.0), which assumes a negative binomial distribution using the mean and variance estimated from gene counts and p-values calculated using the Wald test. GSEA was performed using GSEA 3.0 (http://www.broadinstitute.org/gsea/). Adjusted p<0.05 and a false discovery rate (FDR) q < 0.25 were considered significant. DEGs were selected when presented with an adjusted p-value < 0.05 and a log2FC (fold change) > 1 for upregulated genes and < −1 for downregulated genes. The Database for Annotation, Visualization, and Integrated Discovery (DAVID, http://david.abcc.ncifcrf.gov/, version 6.8) was used to perform GO and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses. Set enrichment analysis was used for the pathway by selecting non-significant differentially expressed genes specified as the “background universe” and accounting for multiple testing using a false discovery rate of q<0.1. Differentially altered pathways were evaluated using the enrich plot package in R to visualize functional enrichment (i.e., dot plot).

2.16. Statistics

Data collection and analyses were performed blinded to genotype. Statistical analysis was performed using GraphPad Prism 10.0 (GraphPad Software). Results are presented as the mean ± SEM. Data were analyzed using a Mann-Whitney U test for non-parametric data. A Student’s t-test was used for parametric data. For each analysis, the statistical test used is reported in the figure legend.

3. Results

3.1. Conditional deletion of LRP1 in SCs increases pain-related behaviors in uninjured mice

Previously, we demonstrated that mice with conditional deletion of LRP1 in Schwann cells (scLRP1^−/−) manifest mechanical hypersensitivity (Orita et al. 2013). To better understand changes in pain processing, we tested several distinct pain modalities in both male and female scLRP1^+/+ and scLRP1^−/− transgenic mice. Mechanical hypersensitivity, evaluated by von Frey filaments, revealed a significant reduction in paw withdrawal thresholds (PWT) (p < 0.05) in both male and female scLRP1^−/− mice (Fig. 1A). Using the Hargreaves test, we determined that scLRP1^−/− mice have reduced paw withdrawal latencies (PWL) (p < 0.05), indicating thermal hyperalgesia (Fig. 1B), which appeared to be more significant in male mice. We also measured responses to cold using an acetone test and exposure to dry ice. Female, but not male, scLRP1^−/− mice exposed to acetone showed increased flinching, licking, and holding their affected paw (Fig. 1C). However, both female and male scLRP1^−/− mice showed a greater cold response when exposed to dry ice when compared to scLRP1^+/+ mice (Fig. 1D). The cold severity of dry ice likely explains the difference between the cold allodynia tests. Next, we utilized nesting behavior as a model to evaluate the mice’s well-being and pain. Only female scLRP1^−/− mice revealed a significant reduction in nesting (Fig. 1E). The behavioral spectrometer tests also demonstrated a difference in response by sex; female scLRP1^−/− mice, but not male mice, spent less time exploring the inner zone, the novel zone, and preferred the outer zone, which is consistent with anxiety-like emotion (Fig. 1F). Overall, female scLRP1^−/− mice consistently demonstrated greater changes in evoked and experiential pain-related behaviors, indicating a slightly more significant neuropathic phenotype than scLRP1^−/− male mice.

Fig. 1. Adult mice with genetic deletion of SC LRP1 demonstrate pain-related behaviors in spontaneous, affective, and evoked pain modalities.

Uninjured female (blue square) and male (blue circle) scLRP1^+/+ mice, and female (orange square) and male (orange circle) scLRP1^−/− mice were tested behaviorally. A Mechanical hypersensitivity was evaluated by von Frey filaments on the plantar surface of the left hind paw (N=18 per genotype: 9 male, 9 female). B Thermal hyperalgesia was measured by the Hargreaves test (N=20 per genotype: 10 male, 10 female). C Cold allodynia evaluated by acetone test (N=18 in scLRP1^+/+ mice: 9 male, 9 female; N=17 in scLRP1^−/− mice: 8 male, 9 female). D Cold response measured by pressing a dry ice pellet against the glass underneath the hind paw. The withdrawal latency was recorded (N=20 per genotype; 10 male, 10 female). E Nesting behavior tests. Determined after 14 hours of nesting (N=17 in scLRP1^+/+ mice: 8 male, 9 female; N=18 in scLRP1^−/− mice: 9 male, 9 female). F Anxiety-like behavior as shown by time spent in the inner zones and measured by a behavior spectrometer. Time was measured in seconds (N=18 per genotype: 9 male, 9 female). Mann-Whitney U tests were used to analyze differences in male, female, and mixed mice between the two genotypes (*p < .05, **p < .01, ***p < .001, and ****p < .0001).

3.2. Mitochondria number and size increase in myelinated SCs when LRP1 is absent

To determine whether alterations in pain processing were associated with nerve ultrastructure, we performed transmission electron microscopy (TEM) studies in sciatic nerves from scLRP1^+/+ and scLRP1^−/− male and female mice. Representative images showing myelinated fibers in scLRP1^+/+ and scLRP1^−/− female sciatic nerves are shown in Fig. 2A. The number of myelinated fibers was not different between genotypes, indicating that myelinated SCs remained viable even when LRP1 was conditionally deleted in SCs (Fig. 2B). This result was consistent for both male and female mice when analyzed by sex and genotype (Fig. S1A). We counted mitochondria in myelinated axoplasm and observed no changes in mitochondria number between the genotypes (Fig. 2C). This finding was also observed in both female and male mice analyzed by sex and genotype (Fig. S1B). However, when we counted mitochondria in myelinated SC cytoplasm, the number of mitochondria in scLRP1^−/− nerves was significantly increased (Fig. 2D). Both male and female scLRP1^−/− mice demonstrated increased mitochondria in myelinated SC cytoplasm, but only scLRP1^−/− females were significantly increased (Fig. S1C). To determine whether the SC cytoplasmic area differed between the genotypes, we calculated the area of SC cytoplasm per myelinated fiber. SC cytoplasmic area in myelinated scLRP1^−/− fibers was significantly increased compared to wild type (Fig. 2E). Although SC cytoplasmic area was increased in female scLRP1^−/− nerve, only male mice were significantly increased (Fig.S1D). High- powered TEM images of myelinated SC cytoplasm (Fig. 2F) in scLRP1^−/− nerves revealed a significant increase in size compared to controls (Fig. 2G) however, changes in mitochondria content density between the genotypes were not statistically significant (Fig. 2H). In some cases, myelinated fibers had similar SC cytoplasmic areas in both scLRP1^−/− and scLRP1^+/+ mice, but the number of mitochondria in the scLRP1^−/− SC cytoplasm was greater than double (Fig. S1E).

Fig. 2. Deletion of LRP1 increases the mitochondria number in myelinated SCs.

A Representative TEM images of myelinated fibers from scLRP1^+/+ (upper panels) and scLRP1^−/− (lower panels) sciatic nerves. Red asterisks show increased mitochondria in the SC cytoplasm of myelinated fibers. Scale bar 2 μm. B Quantification of the number of myelinated fibers in scLRP1^+/+(blue circles) and scLRP1^−/− (orange circles) nerves. C Quantification of the number of mitochondria in myelinated axons. D Quantification of the number of mitochondria in myelinated fibers. E Quantification of myelinated SC cytoplasmic area per 25 μm² image. Data are expressed as mean ± SEM. N=6 (scLRP1^+/+ mice) and N=7 (scLRP1^−/− mice). Student’s t-test (*p < .05, **p < .01). F Representative TEM images of SC cytoplasm of myelinated fibers from scLRP1^+/+ (upper panels) and scLRP1^−/− (lower panels) sciatic nerves. Red asterisks show increased mitochondria number and size. Scale bar 500 nm. G Quantification of mitochondria area per 25 μm² image. H Quantification of mitochondria content density in myelinated SCs. Data are expressed as mean ± SEM. N=3 per genotype. Student’s t-test (***p < .001).

3.3. The sciatic nerve proteome reveals significant alterations in bioenergetic pathways when SC LRP1 is absent

Next, we applied a discovery approach to identify changes in protein expression and biological pathways in sciatic nerves of female scLRP1^−/− compared to scLRP1^+/+ mice. We focussed on female mice since our pain and TEM data indicated more sustained aberrations in scLRP1^−/− female mice. Tryptic peptides and their proteins of origin were identified by LC-MS/MS. Principal components (PC) of 8 samples divided into 2 mouse groups (4 scLRP1^+/+, 4 scLRP1^−/−) were analyzed. The first two principal components accounted for 46% of the variance in the dataset, with PC2 most highly correlated with the groups (Fig. 3A). Proteomic analysis of the results defined 3 groups of proteins (Fig. 3B). These were defined as only identified in scLRP1^+/+ samples, only identified in scLRP1^−/− samples, and identified in both groups (scLRP1^+/+ and scLRP1^−/−); the latter represents the largest group of proteins identified. Differential abundance analysis using Limma quantified the differences between proteins found in both groups (Fig. 3C), confirming the decreased expression of LRP1 in nerves with conditional deletion for LRP1 in SCs (Fig. 3D). Comparing three different databases, KEGG, Reactome, and Wikipathway, we found altered metabolic and biologic pathways in scLRP1^−/− mice, as reported in the bubble plot (Fig. 3E). As shown in the metabolic pathway and protein interaction network, abundant proteins were perturbed in cholesterol and sugar metabolism pathways that are consistent with LRP1 function (Fig. 3F). In addition, we noted significant alterations in pathways with inner mitochondrial membrane proteins and thus, validated some by immunoblotting (Fig. 3G). We detected increased levels of Mitochondrially Encoded Cytochrome C Oxidase I (MT-CO1) and COX7C, both subunits of Complex IV, and decreased level of Adenosine Triphosphate (ATP) Synthase Subunit K (ATP5MK), a subunit of Complex V essential for proper christae structure (Siegmund et al., 2018), in scLRP1^−/− nerves compared to scLRP1^+/+ nerves, that are consistent with the LC-MS/MS data (https://figshare.com/s/77479a30a1c56fad3b23). Because changes in mitochondrial status are associated with sensory processing and pain (Doyle and Salvemini, 2021), behavior responses associated with distinct pain modalities were measured in male and female scLRP1^+/+ and scLRP1^−/− transgenic mice.

3.4. Deletion of LRP1 increases mitochondrial content and promotes mitochondrial fission in SCs, and metabolic stress in sciatic nerves

We investigated whether the increased number of mitochondria observed by TEM would correspond to increased mitochondrial DNA abundance. Sciatic nerves from scLRP1^+/+ and scLRP1^−/− mice were subjected to RT-qPCR to quantify the expression levels of mitochondrially encoded genes 16S (the mitochondrial large subunit ribosomal RNA) and ND1 (the mitochondrial-encoded NADH dehydrogenase 1). Data in Fig. 4A demonstrate that in the absence of SC LRP1, the expression of mitochondrial genes 16S and ND1 is significantly increased compared to control nerves. Increased levels of mitochondrial DNA are consistent with the increased number of mitochondria in myelinated SCs, as observed in TEM image analyses.

Fig. 4. Mitochondrial content, dynamics and nerve health are altered when LRP1 is absent in SCs.

Mitochondria DNA copy number (mtDNA/gDNA) of 16S and ND1 from A sciatic nerves, and B cultured mouse SCs. 18S mRNA was measured as a standard. Data are expressed as fold change compared to wild type (mean ± SEM) of N=3–6 independent studies and analyzed by Student’s t-test (*p < .05; **p < .01). C Immunoblot of mitochondrial (mito) and cytosolic (cyto) subcellular fractions from primary cultured mouse SCs. Protein extracts (20 μg) per well were loaded and probed for the mitochondrial-encoded MT-CO1 and VDAC-1. COX4 served as a nuclear-encoded mitochondrial protein marker, and β-Actin served as a cytosolic marker. D Immunoblots of primary cultured mouse SC lysates and E cell extracts from human SCs electroporated with non-targeting control (NTC) or siRNA targeting human LRP1 (siLRP1) probed for MFN2, DRP1, and LRP1. Calnexin and β-Actin served as loading controls. F,G Densitometric analyses of immunoblots for MFN2 and DRP1. The ratio of MFN2 to β-Actin and DRP1 to β-Actin were calculated for each sample. Data are expressed as mean ± SEM of 3 independent cultures and analyzed by Student’s t-test (*p < .05). H Representative dual immunofluorescence and confocal images of PDH+ mitochondria (green), Sox10 (red), and Dapi (blue) in primary mouse SCs (far left and middle panels, respectively). Magnified images are from boxes to show details. scLRP1^+/+ in box 1 (far right, upper panel), and scLRP1^−/− in box 2 (far right, lower panel). Scale bar 10 μm. I Representative immunoblot from sciatic nerves for phospho-PDH (43 kDa). Total-PDH served as a loading control. J Densitometric analyses of immunoblots showing the ratio of phospho-PDH to PDH calculated for each sample. Data are expressed as mean ± SEM, N=6 nerves per genotype, and analyzed by Student’s t-test (*p < .05). K Representative immunoblot from sciatic nerves for SOD1 (16 kDa). Calnexin (90 kDa) served as a loading control. L Densitometric analyses of immunoblots showing the ratio of SOD1 to Calnexin was calculated for each sample. Data are expressed as mean ± SEM, N=4 nerves per genotype, and analyzed by Student’s t-test (*p < .05).

Next, we isolated SCs from sciatic nerves and established primary SC cultures. Phase contrast images and immunofluorescence were performed to characterize the viability of primary cultured mouse SCs between the genotypes and can confirm LRP1 levels. As shown in Fig. S2, primary mouse SCs displayed similar morphology, although LRP1-deficient SCs were slightly less spindle-shaped compared to controls (Fig. S2A), consistent with previously published studies in rat SCs (Mantuano et al., 2010). Schwann cells isolated from scLRP1^−/− nerves were not immunopositive for LRP1 (Fig. S2B,C) as anticipated, but remained immunopositive for S100 β (Fig. S2D,E). To confirm that LRP1-deficient SCs contributed to increased mitochondrial DNA in the nerve, we measured 16S and ND1. Both mitochondrial DNA markers were elevated in SCs from the scLRP1^−/− nerve, further supporting the increase in mitochondria number (Fig. 2A,D, Fig. 4B). We then isolated mitochondria from primary cultured SCs. Immunoblot analyses of mitochondrial lysates revealed that mitochondrial proteins MT-CO1 and voltage-dependent anion channel 1 (VDAC-1) were increased in scLRP1^−/− mitochondria as compared to control scLRP1^+/+ mitochondria (Fig. 4C). The nuclear-encoded subunit 4 of cytochrome c oxidase (COX4) was immunoblotted and remained unchanged between the genotypes. β-Actin was used as a cytosolic marker (Fig. 4C) to confirm the purity of the mitochondrial isolation.

In healthy conditions, mitochondria divide and fuse to specific cytoplasmic regions in the cell, making them dynamic organelles. Defects in mitochondrial dynamics could lead to changes in mitochondria number and location in SCs. Given that LRP1 regulates mitochondrial dynamics in hepatocytes (Chinnarasu et al., 2021), we sought to determine if mitochondrial dynamics were altered by LRP1 ablation in SCs. We performed immunoblot analyses of SC extracts using mitofusin-2 (MFN2), a marker of mitochondrial fusion, and dynamin-related protein 1 (DRP1), a marker of mitochondrial fission. Primary mouse SCs deficient in LRP1 demonstrated reduced levels of MFN2 and increased levels of DRP1 compared to control SCs (Fig. 4D). These studies were repeated in primary cultured human (Fig. 4E) and rat (Fig. S3) SCs using siRNA silencing of LRP1, and showed identical results for MFN2 and DRP1 expression (Fig. 4E–G; Fig. S3). Collectively, the data reveal that mitochondria in LRP1 deficient SCs are more prone to splitting, rather than merging together, and supports a compelling mechanism for increased mitochondria number in myelinated SCs. Furthermore, both long-term LRP1 deficiency (mSCs) and in SCs with acute LRP1 deficiency (hSCs and rSCs) appear to have alterations in mitochondrial dynamics.

We also directly measured whether mitochondria distribution within the SC cytoplasm was consistent across genotypes. Using dual-label immunofluorescence with pyruvate dehydrogenase (PDH) and Sox10 antibodies, we discovered that the mitochondria were not uniformly distributed within the cells when comparing genotypes (Fig. 4H). Next, we examined the effect of SC LRP1 ablation on metabolic and stress-related regulators in whole sciatic nerve lysates. Protein extracts from scLRP1^+/+ and scLRP1^−/− nerves were immunoblotted for phosphorylated PDH subunit A1 (p-PDH), an inactive form of PDH. PDH catalyzes the conversion of pyruvate to Acetyl-CoA and, by phosphorylation, tightly and reversibly regulates cellular metabolism linking glycolysis to the citric acid cycle and lipogenesis (Rardin et al., 2009). LRP1 ablation significantly increased the ratio of phospho-PDH (Fig. 4I,J) to total PDH, suggesting that these nerves may have a glycolytic shift in their metabolism. We also measured superoxide dismutase-1 (SOD-1) (Fig. 4K,L), an enzyme that is upregulated in response to an imbalance of reactive oxidation species production, metabolism, and detoxification (Godfellow et al., 2020), and observed increased levels of this enzyme in SC LRP1 ablated nerves. These findings suggest that peripheral nerves lacking LRP1 in SCs have altered metabolic and oxidative homeostasis.

3.5. LRP1 deficiency alters mitochondrial bioenergetics in cultured mouse SCs

Given that LRP1 has been implicated in neuronal insulin signaling and bioenergetics (Herz and Bock, 2002, Liu et al., 2015), we next focused on mitochondrial respiration in real-time using a Seahorse XF96 Metabolic Flux Analyzer. Primary mouse SCs isolated from transgenic mice nerves were cultured in 96-well plates, and oxygen consumption rates (OCR) were measured. SCs from scLRP1^−/− nerves exhibited elevated basal and maximal respiration rates compared to scLRP1^+/+ SCs, suggesting that LRP1 deficiency promotes metabolic adaptations of SCs (Fig. 5A–C). After injecting the ATP synthase inhibitor, oligomycin, we observed a significantly increased level of ATP-linked respiration in scLRP1^−/− SCs (Fig. S4A) and proton leak (Fig. S4B), suggesting that ATP production may be less efficient. To estimate the maximal potential respiration that the cells can sustain, we used FCCP, a proton ionophore (uncoupler) to drive maximal electron transport. No changes were detected in spare respiratory capacity in scLRP1^−/− SCs (Fig. S4C). Additionally, we assessed the extracellular acidification rate (ECAR) as a measure of glycolytic flux. LRP1-deficient SCs exhibited higher ECAR levels compared to controls, which suggests a glycolytic shift in these cells (Fig. 5D). To better understand the source of ATP, Proton Efflux Rates were examined for mitoATP production to glycoATP production from our Seahorse studies. We observed a significant increase in mitoATP and only a slight increase (trend only P<0.1) in glycoATP in SCs from scLRP1^−/− mice compared to scLRP1^+/+ mice (Fig. S5A–C). Next, we measured lactate levels. Conditioned medium (CM) (Fig. 5E) and lysates from cultured mouse SCs (Fig. S4D) deficient in LRP1 demonstrated greater than 2.5- and 3.4-fold increase, respectively, in lactate levels over 12 hours, further supporting the possibility that glycolysis or alternative pathways for fuel were operational in these SCs. Next, to evaluate activation levels of PDH specifically in mSCs, we measured phospho-PDH and observed a significant upregulation of the phospho-PDH/PDH ratio (Fig. 5F) in LRP1 deficient SCs, indicating that the conversion of pyruvate to Acetyl-CoA required for oxidative phosphorylation may be limited.

Fig. 5. Mitochondrial respiration increases in LRP1 deficient SCs and is dysfunctional.

Cellular bioenergetics analysis by Agilent Seahorse XFe96 Analyzer. A The real-time oxygen consumption rate (OCR) of cultured mouse SCs from (+/+) scLRP1^+/+ and (−/−) scLRP1^−/− nerves. B Basal (OCR_basal), C maximal respiration (OCR_FCCP), D the real-time extracellular acidification rate (ECAR) in cultured SCs as indicated by the Seahorse assay. Samples isolated from scLRP1^+/+ mice are reported in blue (+/+), while samples from scLRP1^−/− mice are reported in orange (−/−). Data are expressed as mean ± SEM of five independent cultures (biological replicates), and analyzed by Student’s t-test (*p < .05, and **p < .01). E Lactate levels in conditioned media (CM) from mouse SCs after 12 h. Data are expressed as mean ± SEM of 6 independent cultures and analyzed by Student’s t-test (****p < .0001). F Representative image of phospho-PDH (43 kDa) in mouse SC lysates. Immunoblots were analyzed by densitometry. The ratio of phospho-PDH to total-PDH was calculated for each sample. Data are expressed as mean ± SEM of 3 independent cultures and analyzed by Student’s t-test (*p < .05). G Representative images of the mitochondrial membrane potential (MMP) (ΔΨm)-TMRM (upper panels, fire LUT), and ΔΨm-insensitive Mitotracker (lower panels, green) observed by confocal microscopy in cultured SCs. The left panels show control SCs and the right panels show LRP1 deficient SCs. Scale bar 10 μm. The pseudocolor scale indicates low to high intensity. FCCP served as positive control. H Quantification of MMP as a ratio of TMRM and MT Green integrated density. Data are presented as mean ± SEM and analyzed by Student’s t-test (*p < .05). I ROI quantification of MMP changes by calculating the ratio of TMRM and MT Green for each mitochondrion from SC cytoplasm, N=60 mitochondria per genotype, four independent experiments, and analyzed by Student’s t-test (*p < .05).

3.6. Mitochondria in LRP1 deficient SCs are dysfunctional

We evaluated the mitochondrial membrane potential (MMP). We measured MMP using the TMRM and Mitotracker green (MT Green). TMRM is a cell-permeant dye that accumulates in active mitochondria with intact membrane potentials. MT Green probe stains all mitochondria in live cells and serves as a control. As shown in Fig. 5G, both scLRP1^−/− and scLRP1^+/+ SCs exhibit MT Green staining (green), while only control SCs demonstrate substantial TMRM staining (red). LRP1-deficient SCs show TMRM staining comparable to FCCP-treated SCs. Quantification of TMRM integrated density across the entire field of view, normalized to MT Green, revealed a significant decrease in membrane potential in LRP1-deficient SCs (Fig. 5H). To refine this analysis, MT Green was used to identify specific regions of interest (ROIs) corresponding to mitochondria. When TMRM fluorescence within these ROIs was normalized to the MT Green signal, the decrease in membrane potential in LRP1-deficient SCs was more pronounced (Fig. 5I). While a decrease in MMP often points to mitochondrial dysfunction, it can also indicate a response to increased metabolic activity or adaptation (Zorova et al., 2018).

3.7. Mitochondria size is increased in C-fibers and is associated with a decrease in mitochondria content density when LRP1 is conditionally deleted in SCs

We hypothesized that alterations in SC mitochondrial status may impact C-fibers, which are involved in evoked and spontaneous pain-related behaviors in mice. To investigate this, we examined the ultrastructure of scLRP1^+/+ and scLRP1^−/− nerves using TEM. Representative images in Fig. 6A show Remak bundles with normal mitochondria in non-myelinating SCs and C-fibers from scLRP1^+/+ mice (top panels). In contrast, Remak bundles from scLRP1^−/− nerves showed abnormal SC processes and cytoplasm, indicative of dysfunctional SCs, as previously described (Orita et al., 2013). The number of Remak bundles was unchanged between genotypes, confirming the viability of non-myelinating SCs remained consistent (Fig. 6B). Quantification of mitochondria in Remak SCs revealed no changes between genotypes (Fig. 6C), while mitochondria size and density remained unchanged (Fig. 6D). In neighboring C-fibers, the number of mitochondria was equivalent between genotypes (Fig. 6E), however, the mitochondria size (area) was increased (Fig. 6F), accompanied by a reduction in mitochondrial content density (Fig. 6G). These findings suggest that changes in SC bioenergetics influence the cellular ultrastructure of neighboring C-fiber mitochondria. Alternatively, Remak SCs of LRP1-deficient mice may simply be dysfunctional, causing aberrations in Remak SC structure and axonal swelling,

Fig. 6. Loss of LRP1 in SCs induces damage to small unmyelinated C-fiber mitochondrial ultrastructure.

A Representative TEM images of Remak bundles and C-fibers in female scLRP1^+/+ (upper panels) and scLRP1^−/− (lower panels) mice by transmission electron microscopy (TEM). Mitochondria in scLRP1^−/− mice appeared bigger and lighter in color than controls (red arrowheads). Scale bar 2 μm (left panels), 500 nm (middle panels), 200 nm (right panels). B Quantification of Remak bundle numbers. C Quantification of the mitochondria number. and D the mitochondria content density in Remak bundle SCs. E Quantification of the mitochondria number. F Quantification of mitochondria area per 25 μm² image, and G mitochondria content density in Remak bundle axons. Samples isolated from scLRP1^+/+ mice are reported in blue (+/+), while samples from scLRP1^−/− mice are reported in orange (−/−). Data are expressed as mean ± SEM. N=3 scLRP1^+/+ mice and N=4 scLRP1^−/− mice. Data are analyzed by Student’s t-test (*p < .01).

3.8. Transcriptome analyses of DRG neurons from scLRP1^−/− mice show robust alterations in neuroenergetic pathways

Next, we tested whether LRP1 deficiency in SCs altered the neuronal bioenergetic transcriptome in acutely cultured primary sensory neurons. We confirmed that LRP1 mRNA levels are unchanged between the DRG neuron genotypes by RNA-seq, as anticipated (Fig. 7A). We performed pathway analyses of differentially expressed DRG neuron genes from scLRP1^+/+ and scLRP1^−/− mice and used the GSEA database Gene Ontologies (GO) to analyze changes in the Cellular Components (CC) and Biological Processes (BP). The relevant and significant pathways in the GO:CC show a large enrichment and abundant changes in plasma membrane protein complexes were observed in DRG neurons when LRP1 was deleted from SCs (p < .01; Fig. 7B). We also noted that mitochondrial protein-containing complexes and, specifically, the inner mitochondrial membrane protein complex pathways are less enriched and significantly altered in scLRP1^−/− DRGs (p < .001). When we analyzed pathways in the GO:BP, we noted that all pathways associated with oxidative phosphorylation and ATP biosynthetic processes were significantly less enriched in DRG neurons from scLRP1^−/− mice as compared to scLRP1^+/+ mice (Fig. 7C,D). Specifically, alterations in pathways associated with oxidative phosphorylation indicated that a small number of genes in the mitochondrial electron transport chain, NADH dehydrogenase complex assembly, and ATP synthesis are less enriched in DRG neurons from scLRP1^−/− mice. The bubble plot in Fig. 7D reveals a significant number of genes that are less enriched in the ATP biosynthetic process and specifically in ribonucleotide assembly and metabolism. Collectively, these pathway analyses show that loss of LRP1 in SCs affects DRG neuron metabolism.

Fig. 7. Alterations in neuroenergetic pathways of the DRG neuron transcriptome from scLRP1^−/− mice.

A LRP1 raw gene count in DRG neurons from female scLRP1^−/− (orange) and scLRP1^+/+ (blue) mice. B Bubble plots of perturbed metabolic pathways in DRG neurons from scLRP1^−/− mice, determined by Gene Set Enrichment Analysis (GSEA) for Gene Ontology (GO) Cellular Components (CC) and C,D Biological Process (BP). Adjusted p < .05 and a false discovery rate (FDR) q < .25 were considered significant. The size of the dots indicates the number of genes. E Representative confocal images of the mitochondrial membrane potential (MMP) (ΔΨm)-TMRM (upper panels, fire LUT) and MT Green (lower panels, green) in cultured DRG neurons from scLRP1^+/+ mice (left panels), and scLRP1^−/− mice (right panels). Scale bar 10 μm. The pseudocolor scale indicates low to high intensity. F ROI quantification of MMP changes by calculating the ratio of TMRM and MT Green for each mitochondrion, N=58–66 mitochondria per genotype, three independent experiments, and analyzed by Student’s t-test (*p < .05). ROIs from scLRP1^+/+ mice are reported in blue (+/+), while DRG neurons from scLRP1^−/− mice are reported in orange (−/−).

Last, we evaluated the MMPs in scLRP1^−/− and scLRP1^+/+ DRG neurons acutely cultured for less than 24 hours. MT Green (green) confirmed viable mitochondria in both scLRP1^−/− and scLRP1^+/+ neurons, however only the cell bodies of DRG neurons from scLRP1^+/+ mice, and not scLRP1^−/− mice, demonstrated substantial TMRM staining (red) (Fig. 7E). MT Green was used to identify specific ROIs corresponding to mitochondria. When TMRM fluorescence within these ROIs was normalized to MT Green signal, the decrease in membrane potential in scLRP1^−/− SCs was significantly reduced (Fig. 7F). Collectively, the data show that LRP1 deletion in SCs not only changes SC bioenergetics but also alters mitochondria function in neighboring primary sensory neurons as well.

4. Discussion

The importance of LRP1 cell signaling and its role in endocytosis after injury is linked to the survival of SCs under stress (Campana et al., 2006, Mantuano et al., 2008, Mantuano et al., 2011, Flütsch et al., 2016). However, in the absence of injury, when levels of LRP1 in SCs are lower, its role in peripheral nerve physiology is completely unknown. Findings in this study reveal entirely novel evidence that LRP1 plays a pivotal role in SC mitochondria homeostasis in uninjured peripheral nerves. Specifically, we report that SCs are greatly affected by the loss of LRP1, with LRP1 appearing critical in regulating mitochondrial dynamics, maintaining mitochondrial structure, and managing cellular respiration. These results were unexpected since LRP1 is not known to be concentrated in the mitochondria and is mainly localized to the cell surface plasma membrane and in vesicular transport machinery (Stickland et al., 2002, Gonias and Campana, 2014). Thus, the direct regulation of mitochondrial dynamics and function by an endocytic and cell surface protein is compelling.

The mitochondrial alterations we identified, while not necessarily causative, appear to be associated with painful peripheral neuropathy. Our findings are consistent with earlier work in human patients, demonstrating that altered mitochondrial responses and morphology in SCs are associated with peripheral neuropathies that are painful (Kalichman et al., 1998). Accordingly, we observed changes in mitochondria dynamics in human SCs, when LRP1 was transiently silenced, that are consistent with mitochondrial changes observed in rodent SCs when LRP1 was absent. These findings support clinical relevance and suggest that SC LRP1 may serve as a therapeutic target. Nonetheless, most work on peripheral nerve mitochondria in the pain field has focussed on sensory neurons, not SCs (Doyle and Salvemini, 2021). For example, previous studies in sensory neurons indicate that female, but not male mice, are more resistant to mitochondrial stress, have greater hyperexcitability and unique TNFα signaling in nociceptors (Maguire et al., 2022). In female mice, loss of LRP1 in SCs affected not only evoked pain measurements but spontaneous and affective pain-related behaviors, which were not observed in male mice. These findings are consistent with females generally showing a lower pain threshold and pain tolerance, and a greater sensitivity to pain (Keogh and Boerer, 2024) as well as hormonal changes that affect the pain experience (Gulati et al. 2023). However, since we observed a greater number of mitochondria in myelinated SCs of female scLRP1^−/−mice, mechanistically, LRP1 may be a more effective therapeutic target in females.

Previously, a role for LRP1 in glucose and lipid metabolism in neurons (May et al., 2004), adipose tissue (Hofmann et al., 2007), and liver (Ding et al., 2016) has been shown. Data herein significantly expands this knowledge to peripheral glia and provides novel insights into LRP1 as an orchestrator of peripheral nerve bioenergetics via SCs that was not previously appreciated. Indeed, the proteome from SC LRP1-deficient nerves is significantly altered, and bioenergetic pathways, including cholesterol metabolism, are vastly perturbed. Our findings are consistent with those found in LRP1-deficient oligodendrocytes, which shows that cholesterol homeostasis is highly dysregulated (Lin et al., 2017). Specifically, LRP1-deficient oligodendrocytes have reduced cholesterol levels and the ability to differentiate. In scLRP1^−/− nerves, pathway analyses showed that cholesterol metabolism with Bloch pathways is highly altered. Bloch pathways synthesize cholesterol, and although the exact contributions to cholesterol levels are unknown, they are highly operational in glia (Genaro-Mattos et al, 2019). Thus, it is likely that these key LRP1-regulated cholesterol synthesizing pathways play a major role in maintaining cellular cholesterol levels and differentiation, not only in SCs but in oligodendrocytes as well. We also observed significant regulation of amino and nucleotide sugar metabolism pathways that may be related to lactate production, and peptide hormone metabolism pathways in scLRP1^−/− nerves. Regulation of these pathways by LRP1 is novel and currently under investigation in our laboratory.

After careful analysis of respiration in primary cultures of mouse SCs by Seahorse analyses, we learned that LRP1 regulates ATP-linked respiration. ATP is the primary cellular source and energy repository and is mainly synthesized in mitochondria by oxidative phosphorylation (Kadenbach, 2021). Initially, we explained the enhanced respiration observed in LRP1-deficient SCs with greater mitochondria numbers and mass. However, we noted that increased ATP-linked respiration can also reflect an escalation in ATP demand related to either a physiological adaptation or a stress response (Hill et al., 2012, Underwood et al., 2020, Zhou et al., 2022) since mitochondrial mass and mtDNA content are often early molecular events in human cells responding to endogenous or exogenous oxidative stress (Lee et al., 2000). We tested whether mitochondria function was altered and discovered that SCs with LRP1 deficiency also had reduced mitochondrial membrane potential (MMP). Collectively, we interpret these results to more likely be a consequence of high ATP turnover and metabolic adaptation, which is further supported by increases in proton leak and SOD1. Similar results have been observed in murine embryonic fibroblasts with a BCL-2 family member, BIM, deficiency (Wali et al., 2018). The authors suggest that these findings could link metabolism and cell death; however, in our studies, we found no evidence of greater SC death in LRP1-deficient mouse SCs in vitro or in uninjured nerves in vivo. Our hypothesis for greater ATP turnover is further validated by the increased levels of cytochrome c oxidase subunit 1 (MT-C01) in isolated mitochondria from LRP1- deficient SCs. MT-CO1 is a subunit of Complex IV that is directly involved in the ATP feedback inhibition system and, under stressed conditions, can lose its ability to sense energy in the cell (Arnold, 2012). Collectively, our data support a compensatory-like bioenergetic phenotype and mitochondrial stress response that ultimately manifests in mitochondrial dysfunction.

The metabolic coupling of SCs and peripheral axons is still poorly understood, and few molecules in SCs are known to regulate axon stability. Recently, studies in mice with conditional deletion of Rheb, a small GTPase with sequence homology to Ras that binds mTOR and activates mTORC1 complex activity, in SCs showed changes in myelin status and loss of peripheral axon stability due to chronic SC-produced lactate exposure to neurons (Jia et al., 2021). These findings are controversial as previous reports suggest that increased lactate levels by SCs were beneficial to neurons after peripheral nerve injury (Babetto et al., 2020) and that a SC glycolytic shift may serve as an adaptive mechanism that supports the bioenergetics of injured axons. Data herein suggest that LRP1-deficient SCs release abundant lactate into the conditioned medium that may result from a slight shift toward glycolysis and/or activation of alternative metabolic pathways, such as glutaminolysis (DeBerardinis et al., 2007). Activation of glutaminolysis is consistent with our proteomics results showing changes in amino sugar metabolism pathways when SC LRP1 is absent. While extracellular lactate may be a fuel source for neighboring cells or even promote axonal regeneration (Babetto and Beirowski, 2022), the studies by Jia et al. (Jia et al., 2021) indicate that increased lactate produced by SCs increases chronic reactive oxygen species and damages peripheral nerves. The latter are consistent with findings herein, showing increased pain outcomes, aberrations in C-fiber mitochondria and increased SOD1 levels, in mice that produce greater SC-derived lactate. Nonetheless, the importance and complexity of SC mitochondrial homeostasis in peripheral axon stability and bioenergetics is highlighted when LRP1 is conditionally deleted in SCs.

Our current knowledge on the underlying mechanisms of SC mitochondria function in peripheral neuropathies is rather limited and mostly includes data from conditional knockouts of mitochondrial proteins such as prohibitin (Della-Flora Nunes et al., 2021) or mitochondrial transcription factor A gene (Tfam) (Viader et al., 2011). When SC prohibitin 1 is deleted from SCs, a fast progressing demyelinating neuropathy occurs that is related to the activation of the c-Jun and mTORC1 pathways (Della-Flora Nunes et al., 2021). When Tfam is deleted in the mitochondria of SCs, an integrated stress response is initiated, resulting in a dramatic depletion of lipid components that are essential to myelin maintenance. Thus, it appears that mitochondrial dysfunction significantly alters the ability of SCs to maintain myelin appropriately, and accordingly myelinated fibers in nerves of scLRP1^−/− mice have thinner myelin (Orita et al. 2013). More recently, leptin receptors were identified on SCs and shown to regulate metabolic adaptation during peripheral nerve regeneration (Sudaram et al., 2023). These findings are interesting given that LRP1 regulates leptin signaling and leptin receptor complexes in neurons (Liu et al., 2011), and are consistent with alterations in protein hormone metabolism pathways in scLRP1^−/− mice identified by our proteomic studies. However, in contrast to LRP1, leptin receptors alone do not appear to play a role in mitochondrial homeostasis without injury (Sudaram et al., 2023).

The present study adds knowledge to the regulation of SC bioenergetics, however, how loss of LRP1 reduces mitochondrial integrity and function has yet to be identified. One possible explanation includes the observed shift in mitochondrial dynamics towards greater mitochondrial fission (increased DRP1) and reduced fusion (decreased MFN2). This shift is required to make new mitochondria and is often characterized by different levels of membrane potential as an attempt to keep sustained mitochondrial and cellular homeostasis (Quintana-Cabrera and Scorrano, 2023). The reductions in SC mitochondrial membrane potential from LRP1-deficient SCs may reflect this delicate balance. Previous work in hepatocytes shows that LRP1 deficiency alters and compromises mitochondrial fusion and enhances mitochondrial fission, increasing functional defects and mitochondrial fragmentation (Chinnarasu et al., 2021). Because the lack of LRP1 causes alterations in SC physiology (Orita et al., 2013), the increased level of fission in LRP1-deficient SCs may be a response to that stress. Moreover, SCs without LRP1 appear to have a more aged phenotype which displays increased respiration and substantial changes in mitochondrial dynamics (Son et al., 2017).

Interestingly, while the number of axonal mitochondria in myelinated and non-myelinated fibers was unchanged between genotypes, mitochondria in small diameter axons (C-fibers) of scLRP1^−/− nerves appeared sick. Most mitochondria were large, swollen, light-colored, and associated with decreased mitochondria content densities, suggesting aberrations in their cristae structure. We report that ATPMK5, a constituent of Complex V that is required to shape mitochondrial cristae (Siegmund et al., 2018), is significantly reduced in LRP1-deficient nerves and likely contributes to these C-fiber ultrastructural abnormalities. Accordingly, perturbations in inner mitochondria pathways were observed in the DRG neuron transcriptome. We hypothesize that SC LRP1 is an essential regulator of sensory nerve mitochondrial homeostasis.

5. Conclusion

In summary, we demonstrate a novel role for LRP1 in mitochondrial homeostasis in SCs. Our results point to a previously underappreciated role of healthy SC metabolic function in preventing the development and maintenance of painful peripheral neuropathy. When the metabolic homeostasis is shifted in SCs, neighboring vulnerable sensory axons become unstable. The modulation of bioenergetics by SC LRP1 appears to be glia- and neuro-protective in the PNS and could be targeted to improve outcomes in painful peripheral neuropathy.

Highlights.

• LRP1 is critical in SC mitochondrial dynamics by balancing mitochondria fission and fusion.

• LRP1 manages bioenergetics in SCs by maintaining mitochondria membrane potential and reducing the production of lactate.

• SC LRP1 protects sensory neurons from mitochondrial dysfunction and modulates DRG neuroenergetics.

• SC LRP1 controls SC metabolic homeostasis that is associated with C-fiber ultrastructure and the prevention of painful peripheral neuropathy.

Appendix A. Supplementary material

Figures S1, S2, and S3.

Data Availability

Agilent Seahorse XF96e Metabolic Flux Analyzer data normalization is performed via the custom-written macro “FluxNorm”, available at https://github.com/pekkurnazlab/FluxNormalyzer. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. The raw LC-MS/MS and RNA-seq files for the characterization of the samples are available and browsable in the public repository in Figshare (https://figshare.com/s/77479a30a1c56fad3b23).