Collagen II hydrogel-mediated sustained delivery of lacosamide attenuates cartilage degeneration and pain in osteoarthritis

Abstract

Osteoarthritis (OA) lacks effective disease-modifying therapies. Nav1.7 has recently been identified as a regulator of both chondrocyte metabolism and pain, establishing it as a dual-acting therapeutic target in OA. This study evaluated lacosamide (LCM) as a disease-modifying candidate and developed a cartilage-targeted delivery strategy to enhance its translational potential. Sodium channel inhibitors (carbamazepine, oxcarbazepine, and LCM) were evaluated in primary human OA chondrocytes, cartilage explants, and a murine DMM model with systemic or intra-articular administration. Pain behavior was assessed by von Frey and open-field testing. A collagen II–based thermoresponsive hydrogel was developed for sustained intra-articular delivery, and outcomes were assessed by histology, immunohistochemistry, ELISA, and gene expression analyses. LCM was the most potent inhibitor in suppressing IL-1β–induced catabolism and promoting anabolism in human chondrocytes, showing greater efficacy at lower concentrations than carbamazepine or oxcarbazepine. Systemic LCM reduced cartilage degeneration and pain in murine OA and outperformed carbamazepine at equivalent doses, while intra-articular administration achieved superior protection and analgesia at one-tenth the dose. Mechanistically, LCM increased HSP70 and midkine secretion to drive anabolic and anti-catabolic responses. Sustained delivery via a collagen II-based hydrogel prolonged joint retention and enhanced therapeutic durability. Together with its clinical efficacy in Nav1.7 mutation–related neuropathy, these findings underscore LCM's translational potential as a dual-acting disease-modifying therapy for OA. LCM is a dual-acting Nav1.7 inhibitor that alleviates pain and modifies OA progression. Collagen II–based intra-articular delivery enhances efficacy and supports clinical translation as a noninvasive, non-opioid disease-modifying therapy.

Highlights

• •Lacosamide modulates chondrocyte metabolism through Nav1.7 inhibition.
• •Lacosamide outperforms carbamazepine in protecting cartilage and reducing OA pain.
• •Collagen II–based hydrogel enables sustained intra-articular LCM delivery.
• •Hydrogel-released LCM provides long-lasting structural and analgesic benefits in OA.
• •Combining Nav1.7 inhibition with biomaterial delivery offers a disease-modifying OA therapy.

1. Introduction

Osteoarthritis (OA) is the most prevalent form of arthritis and a leading cause of chronic disability worldwide [[1], [2], [3], [4]]. It is characterized by progressive degeneration of articular cartilage, subchondral bone remodeling, osteophyte formation, and synovial inflammation, ultimately leading to joint pain, stiffness, and loss of function [[5], [6], [7], [8]]. The pathophysiology involves a complex interplay of mechanical, inflammatory, and metabolic factors that disrupt joint homeostasis [[9], [10], [11]]. Despite its significant burden on quality of life and healthcare systems, current pharmacological interventions for OA remain largely symptomatic, focusing on pain relief rather than disease modification [[12], [13], [14]]. Non-steroidal anti-inflammatory drugs (NSAIDs), acetaminophen, corticosteroids, and in some cases, opioids are routinely used, but these treatments fail to halt structural deterioration and are often associated with substantial side effects [[15], [16], [17], [18], [19], [20]]. There is, therefore, an urgent medical need to identify therapeutic agents that can simultaneously attenuate OA-associated pain and prevent or reverse joint degeneration.

Our recent work identified the voltage-gated sodium channel Nav1.7, previously thought to be specific to peripheral pain-sensing neurons, as a novel chondrocyte-expressed molecule associated with OA [21]. Nav1.7 is encoded by the SCN9A gene and plays a critical role in action potential generation in nociceptors [[22], [23], [24]]. Serial genetic ablation in mouse models highlighted the dual roles of Nav1.7, with neuron-expressed Nav1.7 affecting pain perception and chondrocyte-expressed Nav1.7 governing OA progression [25,26]. Pharmacological blockade of Nav1.7 with selective Nav1.7 inhibitor PF-04856264 demonstrated simultaneous attenuation of OA progression and pain relief [27]. These highlight the potential of Nav1.7 blockers as innovative disease-modifying drugs for OA, addressing both joint protection and non-opioid pain relief [21,[28], [29], [30], [31], [32]].

To explore the translational potential of Nav1.7 inhibition, we investigated whether the clinically-approved, non-selective sodium channel inhibitor has similar therapeutic effects on OA. Our previous study demonstrated that carbamazepine (CBZ), a first-generation anticonvulsant that is known to act on Nav1.7 [33,34], attenuated cartilage loss and reduced OA-associated pain in a dose-dependent manner [21,28]. However, CBZ's clinical limitations, including autoinduction of its own metabolism, extensive protein binding leading to variable serum levels, and severe drug-drug interactions [[35], [36], [37], [38]], pose major barriers to long term use. Moreover, systemic CBZ exposure is associates with serious adverse reactions, such as Steven‐Johnson Syndrome and Toxic Epidermal Necrolysis Syndrome, particularly in individuals carrying HLA-B1502, HLA-B1511, or HLA-A3101 alleles, which increases risk up to tenfold [35,[39], [40], [41]]. Chronic CBZ use can also cause metabolic derangements such as hypercholesterolemia, folate depletion, thyroid suppression, and hepatotoxicity reflected by elevated ALT, AST, and GGT levels in up to one-third of patients [38,[42], [43], [44]]. These safety concerns significantly limit CBZ's suitability for chronic diseases such as OA.

Second- and third-generation Nav inhibitors, oxcarbazepine (OXC) and lacosamide (LCM), were developed to overcome these challenges. OXC, a structural analog of CBZ, exhibits reduced hepatic enzyme induction and lower immunogenicity, though it carries a risk of hyponatremia at higher doses [35,45]. LCM, a third-generation functionalized amino acid compound, offers even greater clinical advantages: it enhances slow inactivation of sodium channels rather than fast inactivation, has minimal cytochrome P450 interactions, low protein binding, and is primarily renally metabolized [44,[46], [47], [48]]. LCM achieves consistent bioavailability, improved tolerability, and a superior safety profile compared to CBZ and OXC [44,49,50], making it an attractive candidate for repurposing as a disease-modifying OA therapy. Importantly, although LCM is a non-selective sodium channel inhibitor, its clinical efficacy in Nav1.7-related human pain disorders has been directly demonstrated. In a randomized, placebo-controlled, double-blind clinical trial, LCM significantly alleviated pain in patients with Nav1.7 mutation–related small-fiber neuropathy, providing strong human genetic validation for Nav1.7 as a clinically actionable analgesic target [51]. These findings support the translational premise that pharmacologic inhibition of Nav1.7 can yield meaningful analgesic benefit in patients and reinforce the rationale for repurposing LCM as a dual-acting therapeutic candidate targeting both pain and disease progression mechanisms in OA.

Given that long-term systemic delivery of sodium channel inhibitors can lead to systemic toxicity, localized intra-articular administration provides a compelling strategy to maximize local efficacy while minimizing off-target exposure [30,52]. Moreover, hydrogel-based delivery systems enable sustained intra-articular release, maintaining therapeutic concentrations in the joint microenvironment and enhancing treatment durability [53,54].

In this study, we systematically compared the efficacy and safety of three clinically approved Nav inhibitors, CBZ, OXC, and LCM, in regulating chondrocyte metabolism in vitro and mitigating OA progression and pain in vivo. We further developed a collagen II-based hydrogel for sustained intra-articular release of Nav inhibitors to enhance localized therapeutic outcomes. Our results identify LCM as the most potent Nav inhibitor, exhibiting superior chondroprotective and analgesic effects at lower concentrations compared to CBZ and OXC. These findings establish a strong foundation for repurposing LCM as a dual-action, biomaterial-deliverable therapy for OA that integrates disease modification with non-opioid pain relief.

2. Materials and methods

2.1. Mice

All animal protocols were reviewed and approved by the Yale University School of Medicine Institutional Animal Care and Use Committee (IACUC) and were performed following ethical standards for the use of animals in research (IACUC protocol number: 2023-20497). Mice were kept under standard 12-h light/dark cycles with unrestricted access to food and water. Experimental animals were C57BL/6 mice obtained from The Jackson Laboratory. All animals used were sex- and age-matched. OA was induced using the destabilization of the medial meniscus (DMM) model [55], in which the medial meniscotibial ligament of the right knee was transected under ketamine and xylazine anesthesia. Post-operatively, mice were randomized into different treatment groups within each housing cage. LCM (Sigma, SML3059) was delivered either by daily oral gavage at doses of 1, 10, or 50 mg/kg or by alternate-day intra-articular injections at doses of 0.1 or 1 mg/kg, starting 4 weeks after DMM surgery for 8 weeks. Similarly, CBZ (Sigma, C4024) was administered orally (10 mg/kg daily) or intra-articularly (1 mg/kg every other day) following the same treatment timeline. For sustained release, an LCM-loaded hydrogel was injected intra-articularly at 3.15 mg/kg once every 4 weeks for a total of 8 weeks, beginning 4 weeks post-surgery.

2.2. Human subjects research

Experiments involving human tissues were carried out in accordance with Yale IRB-approved protocol (IRB #2000035962). Cartilage tissues were harvested from patients undergoing total knee arthroplasty at Bridgeport Hospital Milford Campus. These samples were dissected into ∼1 mm fragments and digested overnight with 0.25% collagenase II. The following day, liberated chondrocytes were filtered, collected, and expanded in DMEM supplemented with 10% fetal bovine serum, 50 U/mL penicillin, and 0.05 mg/mL streptomycin. Treatments were conducted using indicated agents on these primary cultures.

For the full-thickness cartilage explant assay, human tibia plateaus were obtained from 3 patients with OA undergoing total knee arthroplasty. Cartilage explants were collected from the same regions of each tibial plateau using a 3-mm biopsy punch and randomly distributed into different groups and treated with 10 ng/ml IL-1β, 10 ng/ml IL-1β plus 10 nM LCM, or 10 ng/ml IL-1β plus 10 nM LCM plus HSP70 antibody or midkine antibody in DMEM medium for 5 days. The supernatant was collected and spun at 200g at 4 °C, followed by ELISA assay.

2.3. Cell culture and treatment

C28/I2 [56] cells and primary human chondrocytes were cultured in DMEM containing 10% fetal bovine serum and antibiotics (50 U/mL penicillin and 0.05 mg/mL streptomycin). To simulate an inflammatory state, 20 ng/mL IL-1β was added to the culture medium. For inhibitor studies, cells were pretreated with or without 10 nM LCM before further incubation, after which anabolic and catabolic marker expression was evaluated. Antibodies against HSP70 and midkine were applied to abolish LCM-induced alterations in chondrocyte metabolism, with anti-rabbit IgG serving as the control.

When cells reached 80% confluency, the medium was changed to DMEM supplemented with ITS Liquid Media Supplement (Sigma-Aldrich, I3146) containing 10 nM LCM for 2 days. The conditioned medium was then collected and centrifuged to remove cell debris. Residual LCM was removed by dialysis against fresh medium before use in subsequent experiments.

2.4. Quantitative real-time PCR (qRT-PCR)

Total RNA was isolated using the RNeasy Mini Kit (Qiagen, Cat. 74104), followed by cDNA synthesis and real-time PCR, as previously established. Gene-specific primers are provided in Supplementary Table S1. Expression levels were calculated using the 2^−ΔΔCT method and normalized to internal controls, with results expressed as fold change.

2.5. ELISA

Conditioned media from C28/I2 cells, primary human chondrocytes or cartilage explants treated with 10 nM LCM were collected, and HSP70 and midkine levels were quantified using commercial ELISA kits (Abcam, ab133060 and ab193761) in accordance with the manufacturer's protocols.

2.6. Immunofluorescence staining and quantification

C28/I2 cells were seeded onto glass coverslips and treated as indicated. Cells were fixed with ice-cold methanol for 5 min and then blocked with 5% bovine serum albumin for 1 h. Samples were subsequently incubated with primary antibodies against COL2 (1:200, ProteinTech, 28459-1-AP) or MMP13 (1:200, ProteinTech, 18165-1-AP) overnight at 4 °C, followed by incubation with fluorophore-conjugated secondary antibodies for 1 h at room temperature. Nuclei were counterstained with DAPI, and coverslips were mounted using antifade mounting medium. Images were acquired using a fluorescence microscope (Zeiss) under identical exposure settings across experimental groups. Fluorescence intensity was quantified using ImageJ. For each sample, regions of interest (ROIs) were manually drawn around individual cells or cell clusters, and background fluorescence was subtracted prior to measurement. The integrated density within each ROI was then normalized to the ROI area to obtain the average optical density (AOD), At least three representative fields were quantified for each condition. All values were normalized to the control group and expressed as fold change.

2.7. Histological and immunohistochemical staining

Twelve weeks post-surgery, mouse knee tissues were collected, decalcified with 10% EDTA for three weeks, embedded in paraffin, and sectioned at 6 μm. Serial sections were stained with Safranin-O and H&E. Cartilage degradation was evaluated on Safranin-O–stained sections using the OARSI scoring system (0–24 for human samples [57]–6 for mouse samples [58]) by blinded observers. Synovial inflammation was scored from H&E-stained slides (0–3 scale) [59], while osteophyte development and subchondral bone thickness were assessed [60], and osteophyte maturity graded (0–3) [61].

For immunohistochemical staining, paraffin sections were deparaffinized and subjected to sequential enzymatic digestion using 0.1% trypsin, 0.25 U/mL chondroitinase ABC, and 1 U/mL hyaluronidase. Primary antibodies used included Col II, aggrecan neoepitope, Mmp13, Adamts-5, CD31. Detection utilized the Vectastain Elite ABC kit, with DAB as the chromogen. Counterstaining was done with 1% methyl green, and imaging was performed on a Zeiss Axioscope A1. ImageJ software was used to quantify staining intensity via global threshold.

2.8. Safranin O staining

Paraffin-embedded sections of knees and ankles were deparaffinized with graded xylene and ethanol. Slides were stained with 2% hematoxylin (23412, MilliporeSigma) for 5 min, 1% Safranin O (S8884, Sigma-Aldrich) for 1 h, and counterstained with 0.02% Fast Green (F7258, Sigma-Aldrich) for 1 min. Samples were then dehydrated, coverslipped, and imaged using a Carl Zeiss Axio Scope A.1 microscope.

2.9. Behavioral assessment

Mechanical allodynia was assessed using von Frey filaments (North Coast Medical Inc., CA, USA) [62] ranging from 0.04 to 2.0 g, starting at 0.4 g. After 15-min acclimation on wire mesh, filaments were applied to the plantar surface, and 50% withdrawal thresholds were determined using the up-down method. Nociceptive responses included paw withdrawal, licking, or shaking. Examiners were blinded to group assignments.

Spontaneous locomotion was tested in a 40 × 40 × 40 cm PVC open field. Mice explored freely for 3 min while being video-recorded. Movement parameters were analyzed using behavioral tracking software.

2.10. Preparation of GO/PNIPAM/Collagen

Graphene oxide (GO) was diluted to 0.065 mg/mL from a 2 mg/mL stock. A mixture of 6.5 μL 1 M NaOH and 4 μL 7.5% NaHCO₃ was added, followed by type II collagen (∼3.9 mg/mL). Then, 20 μL PNIPAM solution (2 g PNIPAM in 15 mL PBS) was incorporated. Separately, LCM (650 mg) was dissolved in DMF and added carefully to prevent bubble formation. The mixture was incubated at 37 °C for 5–10 min to induce gelation. All preparation was conducted on ice.

2.11. Gelation and rheological analysis

The gelation time of hydrogels was determined using the test tube inversion method. Briefly, 1 mL of the hydrogel solution was transferred into 2 mL test tubes and incubated in a 37 °C water bath. Gelation process was monitored by gently tilting the tubes and assessing the fluidity of the samples. The point at which the hydrogel no longer flowed upon inversion was recorded as the gelation time.

Rheological characterization of hydrogels during gelation was performed using a DHR-3 rheometer (TA Instruments, USA) equipped with a 20 mm parallel plate geometry and Peltier temperature control. To prevent premature gelation, hydrogel precursor solutions were prepared on ice, thoroughly mixed, and 350 μL was loaded onto the pre-cooled lower plate. The gap was set to 450 μm, and the sample was allowed to equilibrate. A temperature sweep from 5 °C to 45 °C was then conducted at a heating rate of 1 °C/min using small-amplitude oscillatory shear (1% strain, 1 Hz) within the linear viscoelastic region (LVR). The gelation temperature was defined as the point where the storage modulus (G′) surpassed and remained above the loss modulus (G″). To determine gelation time, isothermal time sweeps were carried out at 20, 25, 30, and 37 °C under the same oscillatory conditions, with G′ and G″ continuously recorded until both moduli reached a plateau. All tests were conducted in triplicate to ensure reproducibility.

2.12. Water uptake measurement and hydrolytic degradation study

To evaluate water uptake, hydrogels were prepared as described earlier. Their initial weight (W₀), was recorded, and they were then immersed in 1x phosphate-buffered saline (PBS, Gibco) at a constant temperature of 37 °C. At specific intervals (0, 24, and 48 and 72 h), the hydrogels were carefully removed from the PBS, gently blotted with filter paper to remove excess surface moisture, and reweighed (W_t) The swelling ratio in percentage was determined using the following formula:

Swelling ratio (%) = (W_t-W₀)/W₀ × 100 Eqn. 1

To assess the degradation rate, hydrogels were individually incubated in either PBS or a 0.1% trypsin solution at 37 °C for a duration of six weeks. The incubation media was refreshed on a weekly basis to ensure consistent enzymatic and buffer conditions throughout the study. At predetermined weekly intervals (weeks 1, 2, 3, 4, 5, 6, 7 and 8), samples were retrieved, dried overnight at 37 °C, and their dry weight (W_dry) was measured. The degradation rate in percentage was calculated using the equation below:

Degradation Rate (%) = (W_initial-W_dry)/W_initial × 100 Eqn. 2

For both the swelling and degradation studies, all measurements were conducted in triplicate to ensure the reproducibility of the results.

2.13. In vitro drug release using mass spectroscopy

Mass spectrometric analysis was conducted using a Shimadzu QToF 9030 LC-MS system coupled with a Nexera LC-40D xs UHPLC. The system components included a CBM-40 Lite controller, DGU-405 degasser, two LC-40D XS pumps, SIL-40C XS autosampler, and CTO-40S column oven. UV data were collected with a Shimadzu Nexera Photodiode Array Detector (SPD M − 40) across a wavelength range of 190–800 nm. Samples were maintained at 4 °C in the autosampler compartment. A 3 μL aliquot of each sample was injected onto a Shim-pack Scepter C18-120 column (1.9 μm, 2.1 × 100 mm), equilibrated at 40 °C. Separation was achieved using a binary gradient consisting of solvent A (HPLC-grade water with 0.1% formic acid) and solvent B (HPLC-grade acetonitrile with 0.1% formic acid). Mass spectra were acquired on the QToF 9030 in positive electrospray ionization (ESI) mode, with the electrospray needle held at +4.5 kV. Nebulizer gas flow was set to 2 L/min, heating gas at 10 L/min, interface temperature at 300 °C, dry gas at 10 L/min, desolvation line at 250 °C, and heating block at 400 °C. Spectra were recorded over an m/z range of 50 to 2000. Data analysis and charge deconvolution were performed using LabSolutions Insight Version 3.8 SP1. Measurements and data post-processing were performed with LabSolutions 5.97 Realtime Analysis and PostRun.

Compound	Wavelength/nm	Mass M + H+/m/z
LCM	194 ± 4	251.1390

2.14. Biocompatibility test of hydrogels

To assess the biocompatibility of the hydrogels, chondrocytes were seeded into the lower compartments of 24-well plates at a density of 5 × 10⁴ cells per well and allowed to adhere for 3 h. Hydrogels were sterilized by immersion in 70% ethanol for 15-30 min with gentle agitation, followed by three rinses with sterile PBS. The sterilized hydrogels were then placed into transwell inserts (pore size: 0.4 μm), which allowed for indirect contact with the adhered cell layer. These transwells, each containing about 150 mg of hydrogel, were then positioned into the wells with the cells. A total of 700 μL of DMEM medium was added to the bottom well, and 500 μL was added to the transwell chamber to fully cover the hydrogel. The plates were incubated at 37 °C in a humidified 5% CO₂ atmosphere for durations of 1, 3, and 7 days (D1, D3, and D7). Following the incubation period, the transwell inserts were removed, and an MTT assay was performed to quantify cell viability. In brief, 100 μL of MTT reagent (5 mg/mL in PBS) was added to each well and incubated for 3 h. After incubation, the supernatant was removed, and the formazan crystals were dissolved in 100 μL of DMSO for 20 min. The resulting-colored solution was transferred to a 96-well plate, and absorbance was measured at 450 nm using a microplate reader (Stat Fax, USA). Relative cell viability was calculated by normalizing to untreated control wells using the following formula:

Viability (%) = (Mean OD of treatment / Mean OD of control) × 100 Eqn. 3

Each experimental condition was tested in a minimum of three biological replicates and repeated across three independent experiments (total n = 3).

2.15. Scanning electron microscopy (SEM)

SEM was performed according to a previously protocol from our laboratory. Briefly, hydrogels (n = 3) underwent a series of preparation steps. First, they were thoroughly rinsed three times with double-distilled water (DW). Subsequently, the samples were fixed in 2.5% aqueous glutaraldehyde (Electron Microscopy Sciences, EM Grade 8%) in 0.1 M cacodylate buffer. This fixation was carried out for 1 h at room temperature on a shaker. Following fixation, the hydrogels were rinsed three times with DW, with each rinse lasting 5 min. To reveal their internal morphology, hydrogels were carefully sectioned with a scalpel to expose the cross-sectional structure. Dehydration was then achieved using a graded ethanol series (30%, 50%, 70%, 95%, and 100%), with samples immersed for 10-15 min at each concentration. After ethanol dehydration, samples were subjected to two sequential incubations, each lasting 20-30 min, in 100% hexamethyldisilazane (HMDS; Electron Microscopy Sciences). The hydrogels were then transferred to fresh 100% HMDS and allowed to air dry overnight in a desiccator. For imaging, the dried samples were mounted onto SEM stubs using double-sided carbon tape (Uline Industrial Double-Sided Foam Tape, ULINE, USA). To improve conductivity, they were sputter-coated with a 10 nm layer of gold (Au). Imaging was performed using a Hitachi SU8230 field-emission scanning electron microscope (FE-SEM) operating at an accelerating voltage of 10 kV and a beam current of 86 pA.

2.16. In vivo hydrogel retention assessment via IVIS imaging

For the analysis of hydrogel retention in vivo, the collagen was labeled with IVISense 680 NHS Fluorescent Dye (1 mg; VivoTag) following the manufacturer's recommended protocol. To ensure the removal of any unconjugated dye, the mixture was purified using Zeba™ Dye and Biotin Removal Columns (Thermo Fisher Scientific) through centrifugation at 1000×g for 2 min. The fluorescently tagged collagen was then incorporated into the final hydrogel formulation as previously described. For the animal study, 8-week-old C57BL/6 mice were prepared for imaging. They were anesthetized using isoflurane, and their knee joint regions were shaved to facilitate both the injection procedure and subsequent imaging. Each mouse received a 10 μL intra-articular injection of the IVISense 680-labeled hydrogel. Unilateral injections were performed, with three mice assigned to each group. Fluorescence imaging was conducted at predetermined time points (days 1, 7, 14, 21, and 28) using the IVIS Lumina X5 system (PerkinElmer, MA, USA), with an excitation wavelength of 600 nm and emission at 670 nm.

2.17. Statistical analysis

Results are presented as mean ± standard deviation (s.d.). Student's t-test was used for two-group comparisons, and one-way ANOVA with Bonferroni post hoc test for multiple groups. A P-value <0.05 was considered statistically significant.

3. Results

3.1. LCM exhibits superior efficacy among tested Nav inhibitors in regulating chondrocyte metabolism

To assess the translational relevance of Nav1.7 inhibition in human chondrocyte biology, we examined the effects of three clinically used, non-selective Nav inhibitors which are known to act on Nav1.7, including CBZ, OXC and LCM [63,64], in primary human chondrocytes isolated from patients with advanced knee OA (Kellgren–Lawrence grade 3-4). Based on our previous findings that 10 μM CBZ effectively regulate chondrocyte metabolism [21], this concentration was used for initial comparison across all three compounds. All three inhibitors significantly increased the expression of anabolic markers COL2 and ACAN, and decreased IL-1β-induced expression of catabolic genes, including MMP13 and ADAMTS5. Among them, LCM showed the most potent effects in regulating chondrocyte metabolism, whereas OXC showed comparatively weaker efficacy at the same concentration (Fig. 1A and B). Similar trends were observed in the human chondrocyte cell line C28/I2, confirming the reproducibility of these effects (Fig. S1A and B). Given its superior effects in both primary and immortalized human chondrocytes, subsequent studies focused primarily on LCM.

Fig. 1.

Non-selective Nav1.7 inhibitors regulate chondrocyte metabolism. (A) mRNA levels of COL2 and ACAN in primary human OA chondrocytes treated with 10 μM LCM, CBZ, or OXC for 24h. (B) mRNA levels of MMP13 and ADAMTS5 in primary human OA chondrocytes stimulated with 20 ng/mL IL-1β and 10 μM LCM, CBZ or OXC for 24h. (C, D) COL2 and ACAN (C), MMP13 and ADAMTS5 (D) mRNA expression in human OA chondrocytes treated with various doses of LCM with/without IL-1β stimulated conditions. (E, F) mRNA levels of anabolic markers Col2 and Acan, as well as catabolic enzymes Mmp13 and Adamts5, in primary chondrocytes isolated from WT and chondrocyte specific Nav1.7 knockout mice under basal conditions or following stimulation with 20 ng/mL IL-1β, with or without 10 μM LCM co-treatment for 24 h. (G) Representative immunofluorescence staining showing COL2 (green) and MMP13 (green) expression in C28/I2 chondrocytes treated with 20 ng/ml IL-1β and 10 nM LCM for 48h. Nuclei were counterstained with DAPI (blue). Scale bar = 50 μm. (H) Quantification of average optical density (AOD) shown in G. (I) Safranin O/Fast Green and IHC staining of human OA cartilage explants cultured with 20 ng/ml IL-1β and 10 nM LCM for 5 days. Scale bar = 100 μm. (J) Quantification of GAG area and OARSI scores in I. (K) Quantification of IHC staining in I. Data are presented as mean ± SD. Statistical analysis: one-way ANOVA with post hoc comparison (A-F and H, J, K).

To determine the optimal concentration of LCM for modulating chondrocyte metabolism, we performed a wide-range dose-response screening across a range of concentrations (10 fM to 10 μM). We observed a distinct non-monotonic, biphasic response rather than a linear dose-dependency. Specifically, 10 nM LCM emerged as the functional optimum, inducing the most robust upregulation of anabolic markers and potent downregulation of catabolic markers in both primary human OA chondrocytes and C28/I2 cells. At concentrations both above (10 μM) and below (10 pM) this peak, the metabolic modulation was significantly attenuated, defining a clear therapeutic window centered at the nanomolar level. (Fig. 1C, D, Fig. S1C and D).

To determine whether these effects are mediated through Nav1.7, we isolated primary chondrocytes from Nav1.7 ^fl/fl (WT) and chondrocyte-specific Nav1.7 knockout mice. In line with our precious study [21], Nav1.7 deficiency did not alter basal anabolic marker expression, however, knockout of Nav1.7 significantly blocked IL-1β induced catabolism (Fig. 1E and F). Importantly, the metabolic actions of LCM were completely abolished in Nav1.7-deficient chondrocytes. Unlike WT cells, LCM did not enhance anabolic gene expression and further suppress IL-1β induced catabolic genes in Nav1.7 knockout chondrocytes (Fig. 1E and F). These findings demonstrate that the Nav channel blocker LCM regulates chondrocyte metabolism specifically by inhibiting Nav1.7, establishing Nav1.7 as the critical molecular target through which LCM exerts its chondroprotective effects.

Consistently, immunofluorescence staining in C28/I2 cells under IL-1β stimulation demonstrated that LCM restored COL2 expression and suppressed MMP13 induction (Fig. 1G and H).

To further validate these findings in physiologically relevant context that preserves the extracellular matrix architecture and cell-matrix interactions, we employed an ex vivo human OA cartilage explant model using full-thickness cartilage harvested from tibial plateaus of patients undergoing total knee arthroplasty. Under inflammatory stimulation, LCM treatment consistently downregulated MMP13 and upregulated COL2 expression measured by immunohistochemistry analysis of the explant sections (Fig. 1I–K), confirming that Nav inhibitors directly protects cartilage integrity in native human tissue.

3.2. LCM enhances HSP70 and midkine secretion to regulate chondrocyte metabolism

Mechanistically, our previous studies demonstrated that both non-selective Nav1.7 inhibitor CBZ and selective Nav1.7 inhibitor PF-04856264 regulate the chondrocyte secretome, particularly HSP70 and midkine, which in turn affects chondrocyte biology and OA progression [21,28]. To determine whether LCM acts through similar mechanisms, we investigated its effect on chondrocyte secretome and metabolism. As expected, LCM significantly increased the secretion of both HSP70 and midkine in primary human chondrocytes (Fig. 2A). Conditioned medium collected from LCM-treated primary human chondrocytes promoted the expression of anabolic markers and suppressed IL-1β-induced expression of catabolic molecules, recapitulating the effects of direct LCM treatment (Fig. 2B and C). Consistent with these functional changes, subsequent antibody neutralization experiments further revealed that blockade of HSP70 abolished the anabolic effects of LCM-conditioned medium, whereas blockade of midkine eliminated its anti-catabolic effects (Fig. 2D and E). Similar results were obtained in the human chondrocyte cells line C28/I2, confirming that LCM-induced HSP70 and midkine secretion to regulate chondrocyte anabolism and catabolism, respectively (Fig. S2A–E). Immunofluorescence staining in C28I2 cells revealed that IL-1β stimulation markedly increased MMP13 expression, while treatment with LCM-conditioned medium effectively suppressed this induction. Neutralization of midkine abolished the inhibitory effect of the LCM-conditioned medium on MMP13 expression. Conversely, LCM-conditioned medium enhanced COL2 expression, an effect that was abrogated by HSP70 neutralization (Fig. 2F–H). Furthermore, LCM treatment enhanced the secretion of HSP70 and midkine in human OA cartilage explant supernatants (Fig. S3A). When these supernatants collected from LCM-treated explant were applied to untreated cartilage explants, they upregulated anabolic markers while simultaneously downregulating IL-1β-induced mRNA expression of catabolic markers in human OA cartilage explants as assayed by qRT-PCR (Fig. S3B and C), and increased COL2 while reducing MMP13 protein expression as assessed by immunohistochemical staining (Fig. 2I-N). Importantly, neutralization of HSP70 in the LCM-conditioned supernatant abolished the anabolic effects, while neutralization of midkine reversed the anti-catabolic effects (Fig. 2I–N, Fig. S3D and E), indicating that both HSP70 and midkine are essential mediators of LCM's chondroprotective effects. Together, these findings indicate that LCM regulates chondrocyte metabolism through a secretome-mediated mechanism involving HSP70 and midkine, acting via the same pathway as CBZ and the selective Nav1.7 inhibitor PF-04856264 [21].

Fig. 2.

LCM regulated chondrocyte metabolism through enhancing HSP70 and midkine secretion. (A) ELISA quantification of midkine and HSP70 levels in conditioned medium (CM) collected from human OA chondrocytes treated with 10 nM LCM for 48 h. (B) mRNA levels of COL2 and ACAN in chondrocytes cultured with LCM- CM for 24 h. (C) mRNA levels of MMP13 and ADAMTS5 in human OA chondrocytes treated with 20 ng/mL IL-1β and LCM-CM for 24 h. (D) mRNA expression of COL2 and ACAN in chondrocytes treated with LCM-CM in the presence or absence of anti-HSP70 (αHSP70) antibody for 24 h. (E) mRNA expression of MMP13 and ADAMTS5 in human OA chondrocytes stimulated with 20 ng/ml IL-1β and LCM-CM in the presence or absence of anti-midkine (αMDK) antibody for 24 h. (F, G) Representative immunofluorescence staining of MMP13 (F) and COL2 (G) in C28/I2 chondrocytes with/without IL-1β, LCM-CM, anti-HSP70 antibody and anti-midkine antibody. Nuclei were counterstained with DAPI (blue). Scale bar = 50 μm. (H) Quantification of average optical density (AOD) of MMP13 and COL2 immunofluorescence shown in F and G. (I) Safranin O/Fast Green and IHC staining for MMP13 in human OA cartilage explants cultured with 20 ng/mL IL-1β in the presence or absence of LCM-CM or anti-midkine antibody for 5 days. Scale bar = 100 μm. (J) Safranin O/Fast Green and IHC staining for COL2 in cartilage explants cultured with LCM-CM with or without anti-HSP70 antibody for 5 days. Scale bar = 100 μm. (K) Quantification of GAG area and OARSI scores in I. (L) Quantification of IHC staining in I. (M) Quantification of GAG area and OARSI scores in J. (N) Quantification of IHC staining in J. Data are presented as mean ± SD. Statistical analysis: unpaired two-tailed Student's t-test (A, B) and one-way ANOVA with post hoc comparison (C-E and H, K-N).

3.3. Systemic LCM administration alleviates OA in a dose-dependent manner

We next sought to determine the impact of LCM on cartilage degeneration and OA-associated pain in the destabilization of the medial meniscus (DMM) mouse model, with CBZ serve as a positive control. DMM-operated mice received daily oral administration of CBZ (10 mg/kg body weight) or LCM (1, 10, or 50 mg/kg body weight) (Fig. 3A). In line with previously findings [21], 10 mg/kg CBZ significantly reduced cartilage loss and mitigated synovitis, subchondral sclerosis, and osteophyte formation (Fig. 3B, C, Fig. S4A). However, CBZ treatment had only a mild effect on OA-associated pain behavior compared with untreated DMM mice. (Fig. 3D and E). Remarkably, low dose LCM (1 mg/kg body weight) achieve a comparable structural protection and pain relief to that of 10 mg/kg CBZ (Fig. 3B–E). Medium (10 mg/kg body weight) and high dose (50 mg/kg body weight) LCM provided even stronger protection against cartilage destruction and significantly reduced OA-associated pain behavior, as demonstrated by improved locomotor activity in the open field test and reduced mechanical allodynia in the von Frey assay (Fig. 3B–E, Fig. S4A). To be noted, the protective and analgesic effects against OA of LCM plateaued between the medium and high doses, indicating a dose-dependent benefit that reached maximal efficacy at the medium dose. Immunohistochemical staining further showed that both CBZ and LCM treatment significantly increased Col2 levels while decreased Mmp13, Adamts5 and Aggrecan neoepitope levels in articular cartilage (Fig. 3F, Fig. S4B). Among these, LCM at 10 mg/kg provided the most robust protection, surpassing CBZ at the same dose.

Fig. 3.

Systemic administration of LCM attenuates OA progression and pain. (A) Schematic of the experimental outline. Twelve-week-old male WT mice underwent DMM surgery. Beginning 4 weeks post-surgery, mice were treated daily with either CBZ (10 mg/kg) or LCM (1, 10, or 50 mg/kg) by oral gavage for a total duration of 8 weeks. (n = 6). (B) Representative Safranin O/Fast Green stained images of knee joints at 12 weeks post-DMM surgery, systemically treated with CBZ (10 mg/kg/day) or LCM (1, 10, or 50 mg/kg/day) (n = 6). Scale bar = 100 μm. (C) Quantification of OARSI scores, subchondral bone plate (SBP) thickness, osteophyte formation, and synovitis. (D, E) Behavioral outcomes assessed by open-field travel distance (D) and von Frey test (E). (F) Representative IHC staining of Col2, Aggrecan neoepitope, Mmp13, and Adamts5 in joint sections (n = 6). Scale bar = 50 μm ∗Vehicle versus LCM (50 mg/kg/day), ^#Vehicle versus LCM (10 mg/kg/day),(∗^,#P < 0.05; ∗∗^,##P < 0.01). Data are presented as mean ± SD. Statistical analysis: one-way ANOVA with post hoc comparison (C) and unpaired two-tailed Student's t-test (D, E).

To assess vascular alterations associated with OA progression and pain, we examined an angiogenic marker, CD31-positive endothelial cells in joint tissues [65]. Compared with untreated OA controls, CBZ significantly reduced CD31-positive cells in subchondral bone; however, this reduction was modest. Similarly, the lowest dose of LCM produced a comparable decrease in CD31 expression relative to CBZ. In contrast, medium- and high-dose LCM treatment resulted in a more pronounced reduction in CD31-positive cells. These findings indicate that while both CBZ and low-dose LCM partially suppress OA-associated vascular changes, higher doses of LCM are more effective in attenuating pathological vascular responses (Fig. S5). These results suggest that LCM provides dose-dependent structural and symptomatic relief in OA, exhibiting superior efficacy to CBZ at the same dose and demonstrating its potential as a more effective disease-modifying OA therapy with combined chondroprotective and analgesic effects.

3.4. Local LCM delivery protects against OA with superior efficacy at lower doses

Given the strong systemic efficacy of LCM, we next tested whether local administration could achieve comparable chondroprotective and analgesic effects at lower doses while minimizing potential systemic side effects. DMM-operated mice received intra-articular injections of CBZ (1 mg/kg body weight) or LCM (0.1 mg/kg or 1 mg/kg body weight) three times per week for 8 weeks (Fig. 4A). Both CBZ and LCM markedly attenuated cartilage degeneration, synovitis, osteophyte formation and subchondral bone sclerosis compared with vehicle-treated DMM controls (Fig. 4B, C, Fig. S6A). Notably, 0.1 mg/kg LCM provided structural and symptomatic protection comparable to that achieved with 1 mg/kg CBZ, whereas 1 mg/kg LCM yielded the strongest therapeutic benefit. Mice treated with 1 mg/kg LCM exhibited significantly improved locomotor activity and reduced mechanical allodynia, indicating robust relief of OA-associated pain (Fig. 4D and E). Immunohistochemical analyses further confirmed that both CBZ and LCM treatments enhanced Col2 expression and suppressed MMP13 and Aggrecan neoepitope levels in articular cartilage, reflecting reduced catabolic activity and preserved extracellular matrix integrity (Fig. 4F, Fig. S6B). Among all treatment groups, 1 mg/kg LCM produced the most pronounced protective effects on cartilage structure and matrix homeostasis. Together, these results demonstrate that local delivery of LCM provides potent, dose-dependent protection against OA-induced structural damage and pain, achieving comparable or greater efficacy than CBZ at one-tenth the dose. This highlights LCM's potential as a superior and locally effective disease-modifying OA therapy.

Fig. 4.

Intra-articular delivery of LCM protects against OA. (A) Schematic of the experimental outline. Twelve-week-old male WT mice underwent DMM surgery. Beginning 4 weeks post-surgery, mice received intra-articular injections of CBZ (1 mg/kg, three times per week) or LCM (0.1 or 1 mg/kg, three times per week) for a total duration of 8 weeks (n = 6). (B) Representative Safranin O/Fast Green stained sections of knee joints at 12 weeks post-surgery, treated with or without CBZ (1 mg/kg, 3 × /week) or LCM (0.1 or 1 mg/kg, 3 × /week) intra-articularly (n = 6). Scale bar = 100 μm. (C) Quantification of OARSI scores, subchondral bone plate (SBP) thickness, osteophyte formation, and synovitis. (D, E) Behavioral outcomes assessed by open-field travel distance (D) and von Frey test (E). (F) Representative IHC staining of Col2, Aggrecan neoepitope, Mmp13, and Adamts5 in joint sections (n = 6). Scale bar = 50 μm ∗Vehicle versus LCM (1 mg/kg, 3 × /week), ^#Vehicle versus LCM (0.1 mg/kg, 3 × /week),(∗^,#P < 0.05). Data are presented as mean ± SD. Statistical analysis: one-way ANOVA with post hoc comparison (C) and unpaired two-tailed Student's t-test (D, E).

3.5. Generation of LCM-loaded collagen II-based hydrogel

Although repeated intra-articular injections of LCM effectively protected against OA progression, the need for frequent administration limits clinical feasibility. We next sought to improve the duration and consistency of its therapeutic effects. To achieve this, we encapsulated LCM within a biocompatible collagen II-based hydrogel designed for controlled intra-articular release. This strategy aimed to maintain effective local LCM concentrations over an extended period, thereby maximizing therapeutic efficacy while minimizing injection frequency. The hydrogel was formulated from a composite of poly (N-isopropylacrylamide) (PNIPAM), type II collagen, and trace graphene oxide (GO), forming a physically crosslinked, injectable network. PNIPAM served as the thermo-responsive backbone that undergoes a reversible sol–gel transition near physiological temperature (∼32 °C), allowing liquid injection at low temperature and rapid gelation within the joint cavity. Type II collagen provided natural biocompatibility and enzymatically degradable sites for tissue remodeling, while GO contributed to enhanced mechanical strength and homogeneous drug distribution through π–π and hydrogen-bonding interactions. This PNIPAM/Collagen/GO hybrid hydrogel thus combines temperature sensitivity, biodegradability, and sustained-release capacity, enabling it to serve as a long-acting depot for LCM delivery.

3.6. Rheological properties of hydrogel

The dynamic rheological properties of the hydrogel formulation were thoroughly investigated to define its gelation characteristics, including gelation temperature and isothermal gelation times. The hydrogel exhibited clear temperature-sensitive gelation behavior. At 5-25 °C, both storage modulus (G′) (red line) and loss modulus (G″) (green line) remained low (∼10⁻⁶ MPa), with G″ > G′, indicating a viscous liquid-like state. As temperature rose beyond ∼25 °C, G′ increased sharply and surpassed G″ at ∼30–31 °C, marking the sol–gel transition. Continued heating to 45 °C led to a two-log increase in G′, confirming formation of a stable, elastic network. Isothermal gelation kinetics revealed faster gelation at higher temperatures: no gelation at 20 °C, slow onset at ∼650 s at 25 °C, accelerated transition at ∼200 s at 30 °C, and rapid gelation within ∼100 s at 37 °C (Fig. 5A, Fig. S7A).

Fig. 5.

Thermo-responsive hydrogel properties and intra-articular delivery performance of LCM. (A) Rheological temperature sweep of the PNIPAM-based hydrogel showing sol–gel transition near 30–31 °C, with G′ (red) surpassing G″ (green). (B) Swelling behavior at 25 °C and 37 °C over 72 h, showing increased hydration below LCST and reduced swelling above LCST (n = 3). (C) Degradation profile over 8 weeks in PBS and 0.1% trypsin at 37 °C, with faster degradation under enzymatic conditions (n = 3). (D) In vitro release kinetics of LCM from hydrogel at 37 °C, demonstrating sustained, burst-free release. (E) SEM image of lyophilized hydrogel showing porous architecture suitable for nutrient diffusion (n = 5). Scale bar = 40 μm. (F) Chondrocyte viability and proliferation on days 1, 3, and 7, assessed by MTT assay in response to unloaded or LCM-loaded hydrogel (n = 5). (G) In vivo fluorescence imaging after intra-articular injection of IVISense 680-labeled hydrogel, with retention monitored up to day 56 (n = 5). (H) Quantification of mean fluorescence intensity (MFI) confirming superior retention in the hydrogel group. Data are presented as mean ± SD. Statistical analysis: one-way ANOVA with post hoc comparison (F).

3.7. Swelling behavior, degradation, and sustained drug release

The hydrogel demonstrated classic thermo-responsive swelling, with a significantly higher swelling ratio at 25 °C (below the LCST of ∼32 °C) due to hydrophilic polymer-water interactions. At 37 °C (above LCST), the hydrogel network collapsed into a denser, hydrophobic structure, restricting water absorption and reducing the swelling ratio (Fig. 5B). Degradation studies over 8 weeks revealed excellent stability in PBS (∼25.66% mass loss), while exposure to 0.1% trypsin resulted in accelerated degradation (∼57.2%) due to enzymatic cleavage of the collagen component (Fig. 5C), demonstrating environment-dependent biodegradability. Drug release assays showed a sustained and controlled release profile of LCM, with no initial burst and ∼99% cumulative release over time (Fig. 5D). At 37 °C, the PNIPAM collapse reduced mesh size and restricted drug diffusion, extending the release duration.

3.8. Evaluation of cytocompatibility and structural characterization of hydrogels

SEM imaging revealed an interconnected porous architecture favorable for nutrient diffusion and cellular compatibility (Fig. 5E). The biocompatibility of the PNIPAM-based hydrogels was assessed using primary human chondrocytes cultured indirectly with either unloaded or LCM-loaded hydrogels in a transwell system. High cell viability was observed on day 1 (D1) in all groups, with no evidence of cytotoxicity (Fig. 5F). By day 3 (D3), cell proliferation had increased across all conditions. Notably, by day 7 (D7), chondrocytes exposed to LCM-loaded hydrogels exhibited significantly higher proliferation compared to those cultured with unloaded hydrogels, indicating enhanced bioactivity of LCM. LCM released from hydrogels was collected and applied to primary chondrocytes to evaluate its effects on anabolic and catabolic gene expression. At equivalent concentrations, hydrogel-released LCM was as effective as free LCM in promoting chondrocyte anabolism and suppressing catabolic gene expression (Fig. S7B and C), suggesting that the hydrogel delivery system preserves the bioactivity of LCM.

3.9. Hydrogel retention analysis via IVIS imaging in mice

To assess in vivo retention and depot formation, the PNIPAM-based hydrogel incorporating IVISense 680-labeled collagen was injected intra-articularly into mouse knees, and fluorescence was monitored over 56 days. A free dye control group (IVISense 680 in PBS without hydrogel) was included to determine whether the hydrogel matrix enhanced local retention compared to unencapsulated dye. The hydrogel group exhibited strong localized fluorescence at the joint site on day 0 immediately after injection, which gradually declined over time but remained clearly detectable up to day 42, indicating prolonged residence in the intra-articular space (Fig. 5G). In contrast, the free dye group showed a rapid loss of fluorescence, with signal intensity dramatically reduced by day 7 and almost completely absent by day 14, suggesting fast clearance of the unencapsulated dye from the joint (Fig. S7D). Quantification of fluorescence intensity over time (Fig. 5H, Fig. S7E) confirmed this difference. The hydrogel group retained approximately 10% of its initial fluorescence at day 56, whereas the free dye group dropped below 3% by day 14. This striking contrast highlights the prolonged retention capability of the hydrogel, likely due to in situ gelation above the LCST of PNIPAM, forming a stable, thermally crosslinked depot that physically traps the dye and resists diffusion.

3.10. Sustained intra-articular release of LCM via hydrogel enhances and extends its therapeutic efficacy in OA

To investigate the therapeutic potential of hydrogel-mediated delivery of LCM for OA treatment, we intra-articularly injected 10 μL of LCM-loaded hydrogel (each 10uL of gel contain 63.75 μg LCM) into DMM-induced OA mice. Control groups received either blank hydrogel or PBS (Fig. 6A). Compared to control groups, LCM-hydrogel treatment provided prolonged chondroprotection, with significantly reduced OARSI scores, synovitis, osteophyte formation and subchondral bone sclerosis at 12 weeks post-surgery (Fig. 6B, C, Fig. S8). LCM-hydrogel treatment also significantly alleviated OA-associated pain, as evidenced by increased locomotor activity and elevated paw withdrawal threshold (Fig. 6D and E). Immunohistochemical analyses revealed sustained suppression of MMP13 and ADAMTS5 expression and enhanced COL2 deposition in the LCM-hydrogel group (Fig. 6F and G). No signs of local inflammation or systemic toxicity were observed, indicating good biocompatibility and safety of the hydrogel system (Fig. S9). Together, these results demonstrate that sustained intra-articular delivery of LCM via collagen II-based hydrogel provides prolonged structural and symptomatic protection against OA, offering a clinically translatable strategy that maximizes therapeutic efficacy while minimizing dosing frequency and systemic exposure.

Fig. 6.

Hydrogel-mediated delivery of LCM provides prolonged joint retention and therapeutic efficacy. (A) Schematic of the experimental outline. Twelve-week-old male WT mice underwent DMM surgery. Beginning at 4 weeks post-surgery, mice received intra-articular injections of either blank hydrogel or LCM-loaded hydrogel once monthly for 2 months (n = 6). (B) Safranin O/Fast Green–stained sections of knee joints 12 weeks post-DMM surgery, following intra-articular injection of LCM-loaded hydrogel (n = 6). Scale bar = 100 μm. (C) Quantification of OARSI scores, SBP thickness, osteophyte size, and synovitis in each group. (D, E) Behavior test outcomes assessed by open-field movement (D) and von Frey test (E). (F) Representative IHC staining of Col2, Aggrecan neoepitope, Mmp13, and Adamts5 in joint sections (n = 6). Scale bar = 50 μm. (G) Quantification of IHC-positive staining in (F). ∗Vehicle versus LCM-loaded hydrogel, (∗P < 0.05; ∗∗P < 0.01, ∗∗∗P < 0.001). Data are presented as mean ± SD. Statistical analysis: one-way ANOVA with post hoc comparison (C, G) and unpaired two-tailed Student's t-test (D, E).

4. Discussion

Although highly selective Nav1.7 inhibitors, such as PF-04856264, have provide valuable mechanistic tools and demonstrated analgesic efficacy in preclinical models, they have not yet been approved for clinical use and remain limited by developmental, regulatory, or pharmacokinetic challenges. As a result, their immediate clinical translation for chronic joint diseases such as OA remains uncertain. In contrast, LCM is an FDA-approved antiepileptic drug with a well-established safety profile in humans, predictable pharmacokinetics, and minimal drug-drug interactions. These features make LCM positioned for rapid clinical repurposing. Importantly, despite its non-selective sodium channel inhibition, our data demonstrate that LCM effectively modulates Nav1.7-dependent pathways in chondrocytes and cartilages at relatively low doses, resulting in robust chondroprotective and analgesic effects. Thus, the choice of LCM reflects a strategic balance between biological efficacy and real-world translational feasibility, enabling direct progression from preclinical validation to potential clinical application. Notably, LCM has already demonstrated clinical efficacy in patients carrying pathogenic SCN9A variants. A randomized, placebo-controlled clinical trial showed that LCM significantly alleviated pain in Nav1.7-associated small-fiber neuropathy, establishing that pharmacological modulation of Nav1.7 yields clinically meaningful analgesia in humans. This prior human validation substantially strengthens the translational impact of the present study and supports LCM as a clinically credible disease-modifying therapeutic candidate for OA [51].

Through integrated in vitro, ex vivo, and in vivo analyses, we demonstrate that LCM outperforms other clinically used Nav blockers, including CBZ and OXC, in modulating chondrocyte metabolism, attenuating OA progression, and reducing OA-associated pain. Among these agents, LCM exhibited the greatest potency in promoting anabolic markers such as COL2 and ACAN while suppressing catabolic mediators including MMP13 and ADAMTS5, achieving comparable or superior effects at markedly lower concentrations. Interestingly, we observed a distinct non-monotonic, biphasic response rather than a linear dose-dependency, suggesting that optimal modulation of Nav1.7 activity is required to maintain chondrocyte metabolic balance. This non-linear relationship may reflect the delicate interplay between sodium flux, calcium signaling, and transcriptional regulation in chondrocytes.

Consistent with our previous reports that Nav1.7 blockade by CBZ or PF-04856264 modulates the chondrocyte secretome [21], LCM significantly enhanced the secretion of HSP70 and midkine, two key chondroprotective factors. Neutralization studies confirmed that HSP70 mediates the anabolic response, whereas midkine primarily contributes to anti-catabolic activity. These findings highlight a paracrine mechanism through which LCM exerts long-range regulatory effects on cartilage homeostasis. While LCM shares this core mechanism with CBZ, the molecular basis for its superior efficacy remains incompletely understood. This advantage may arise from LCM's distinct pharmacological properties, including preferential enhancement of slow Nav1.7 inactivation, improved bioavailability, and lower effective dose requirements. It is also plausible that LCM engages additional signaling pathways beyond HSP70 and midkine regulation. Nevertheless, the precise upstream and downstream signaling events linking Nav1.7 inhibition to HSP70 and midkine induction remain incompletely defined. Future studies employing transcriptomic, proteomic, or signaling pathway analyses will be required to uncover these potential mechanisms and further optimize Nav1.7-targeted therapeutics for OA. In the DMM-induced OA mice model, systemic LCM administration alleviated cartilage destruction and OA-associated pain in a dose-dependent manner, with medium-dose treatment providing the most pronounced effect. At equivalent doses, LCM outperformed CBZ in both joint protection and pain reduction. These findings likely reflect LCM's pharmacokinetic advantages including low protein binding, minimal hepatic enzyme induction, and more stable bioavailability, compared to CBZ, whose clinical application is constrained by autoinduction, hepatotoxicity, and hypersensitivity risks. The observation that higher doses of LCM did not confer additional benefits suggests a therapeutic plateau, consistent with its non-monotonic, biphasic response observed in vitro. Beyond systemic administration, we also demonstrated that local intra-articular delivery of LCM yields potent therapeutic benefits. Intra-articular LCM outperformed CBZ in reducing joint degeneration and pain-related behaviors, highlighting the value of targeted delivery in minimizing systemic exposure while maximizing local efficacy.

Although repeated intra-articular injection of free LCM provided therapeutic benefit, it required relatively frequent dosing to maintain efficacy, which may limit clinical feasibility. To address this limitation, we developed a collagen II–based thermosensitive hydrogel that enables sustained intra-articular release of LCM, prolongs joint retention, and maintains therapeutic efficacy while significantly reducing injection frequency. This material-assisted delivery strategy has the potential to improve safety, patient compliance, and long-term treatment outcomes. Collagen II, the principal extracellular matrix component of articular cartilage, provides a biologically relevant and mechanically compatible scaffold that enhances joint residency and supports chondrocyte homeostasis. While direct intra-articular injection of free LCM did provide chondroprotective and analgesic benefits, these effects required frequent repeat dosing to maintain therapeutic levels in the joint. In contrast, incorporation of LCM into cartilage-targeting collagen II-based hydrogel enabled prolonged intra-articular retention, significantly reducing injection frequency while achieving equal or superior structural and symptomatic protection. Importantly, hydrogel-based localization optimized the pharmacokinetic profile of LCM while minimizing systemic exposure, an essential consideration for chronic joint diseases such as OA, where cumulative off-target effects, injection burden, and poor patient adherence limit treatment success. Although both CBZ and LCM are FDA-approved antiepileptic drugs, their pharmacological profiles diverge substantially. While CBZ's clinical utility is limited by hepatic and hematologic toxicities, LCM's superior tolerability and minimal drug-drug interactions make it an attractive repurposing candidate. Compared with other emerging OA therapies, such as direct intra-articular drug injections and stem cell treatments, LCM delivered via our collagen II–based hydrogel offers several distinct advantages. Direct injection of bioactive compounds typically suffers from rapid clearance from the joint, requiring frequent dosing and limiting sustained therapeutic effects [66,67], while stem cell therapy faces challenges of cell survival, engraftment, and high cost [68]. In contrast, our hydrogel provides localized, sustained release of LCM, prolonging retention in the joint, maintaining effective drug concentrations over time, and potentially reducing systemic exposure and associated side effects. Moreover, the collagen II matrix may enhance cartilage compatibility and integration, supporting cartilage preservation while delivering the drug. Together, these features suggest that hydrogel-mediated delivery of LCM could provide superior efficacy, convenience, and safety relative to conventional intra-articular therapies and highlight its potential as a disease-modifying OA treatment. Our findings further suggest that intra-articular hydrogel-based delivery could reduce systemic exposure and adverse events while maintaining therapeutic efficacy. This strategy exemplifies the synergistic integration of pharmacology and biomaterials to achieve localized, disease-modifying therapy for OA (Fig. 7). Nevertheless, optimization of dosing regimens, injection intervals, and comprehensive evaluation of long-term joint safety will be essential and should be addressed in future large-animal studies and clinical trials. In summary, this work identifies LCM as a potent, dual-action Nav1.7 inhibitor capable of modifying OA progression and alleviating pain. By coupling LCM with a hydrogel-based sustained-release system, this work provides a foundation for the development of non-invasive, non-opioid, and local disease-modifying OA therapies with clear translational potential.

Fig. 7.

Schematic illustration of the delivery routes and cartilage-protective mechanisms of LCM in osteoarthritis. Oral administration of LCM leads to systemic absorption through the gastrointestinal tract followed by blood circulation to the joint. Intra-articular delivery provides direct deposition of LCM or LCM-loaded hydrogel into the knee cavity, enabling localized and sustained exposure. Within chondrocytes, LCM acts via Nav1.7 to enhance anabolic processes while suppressing catabolic responses. LCM stimulates the secretion of HSP70, which promotes anabolism, and reduces midkine-mediated catabolic signaling. Together, these effects collectively contribute to chondroprotection in OA. Created with BioRender.com.

CRediT authorship contribution statement

Chaopeng He: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Guiwu Huang: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Lida Moradi: Validation, Methodology, Investigation, Data curation. Jingwei Bi: Methodology, Investigation, Formal analysis, Data curation. Xinyu Yang: Methodology, Investigation. Xin Liu: Methodology, Investigation. Xudong Cui: Methodology, Investigation. Arya Varthi: Formal analysis, Data curation. Daniel H. Wiznia: Resources. Stephen G. Waxman: Methodology, Conceptualization. Wenyu Fu: Writing – review & editing, Validation, Supervision, Project administration, Methodology, Formal analysis, Conceptualization. Chuan-Ju Liu: Writing – review & editing, Validation, Supervision, Resources, Project administration, Methodology, Funding acquisition, Conceptualization.

Data availability statement

All other relevant data from this study are available from the corresponding authors upon reasonable request.

Ethics approval and consent to participate

All animal procedures were carried out in accordance with institutional guidelines and approved by the Institutional Animal Care and Use Committee of Yale University. Human subjects research was performed according to the protocols approved by Yale University Institutional Review Board (IRB Study Number 2000035962). Patients and/or the public were not involved in the design, or conduct, or reporting or dissemination plans of this research.

Funding

This work was supported partly by NIH research grants R01AR078035, R01AR062207, R01AR076900 and R01NS070328.

Declaration of competing interest

The authors declare no competing interests.

Appendix A. Supplementary data

The following is the Supplementary data to this article: