APOA4 Protects Chondrocytes and Modulates Wnt Signaling in Osteoarthritis

Introduction

Osteoarthritis (OA) is the most common degenerative joint disease, marked by the gradual breakdown of articular cartilage, changes in the subchondral bone, synovial inflammation, and subsequent pain and functional limitations in those affected.^1,2 Despite its widespread prevalence and significant socioeconomic impact, existing treatments mainly focus on alleviating symptoms, such as pain and limited joint mobility, without effectively halting or reversing the progression of the disease.³ Thus, gaining a deeper insight into the molecular mechanisms that regulate cartilage homeostasis and degeneration is crucial for identifying potential novel targets for disease-modifying therapies. OA is increasingly recognized as a whole-joint pathology involving coordinated abnormalities across articular cartilage, subchondral bone, synovium, meniscus, ligaments, and the infrapatellar fat pad (IFP), which together drive structural deterioration and symptoms in the knee.⁴

Recent studies have highlighted the infrapatellar fat pad (IFP) as an active participant in knee joint homeostasis and OA pathogenesis. The IFP functions as an active metabolic and immune organ within the joint; in OA it develops inflammation and fibrosis and can interact with altered biomechanics to exacerbate structural damage and symptoms.⁵ Far beyond passive cushioning, IFP secretes adipokines, cytokines and extracellular matrix–modulating proteins that influence periarticular tissues.^6,7 Proteomic profiling of IFP from OA patients versus non-OA controls has revealed differential expression of several factors, among which apolipoprotein A4 (APOA4) was found to be significantly up-regulated in the OA-derived tissue.⁸ In our preliminary work, we observed APOA4-related changes consistent with these observations, which motivated the present study. In this study, the IFP proteome was profiled by mass spectrometry across OA and non-OA cohorts; APOA4 emerged among differentially abundant proteins within pathway contexts related to extracellular-matrix organization and inflammation.⁸ However, the functional relevance of APOA4 in OA remains unexplored. To our knowledge, no study has directly examined APOA4 in human articular cartilage or chondrocytes, and its impact on matrix metabolism, inflammatory signaling, or Wnt-related pathways in the osteoarthritic joint remains largely unknown. This gap motivated us to investigate APOA4 in joint tissues and chondrocytes in the present work. In parallel, synovial fluid proteomics in osteoarthritis cohorts has established the detectability of lipid-binding apolipoproteins in joint fluids, and APOA4 has been detected in human OA synovial fluid by mass-spectrometry–based profiling.^9–12 Moreover, IFP inflammation has been linked to pain sensitization, and knee OA commonly features peripheral and central sensitization that may amplify nociceptive input from inflamed periarticular tissues.¹³ Emerging evidence links lipid metabolism and lipoprotein signaling to joint degeneration: chondrocytes sense and respond to exogenous lipids/apolipoproteins, modulating matrix turnover, oxidative stress, and inflammatory pathways; alterations in the intra-articular lipid/proteomic milieu associate with structural damage and symptoms in OA.^14,15 Within this framework, APOA4, an apolipoprotein with anti-inflammatory and metabolic regulatory properties, may influence chondrocyte homeostasis via lipid–receptor–signaling axes, thereby linking IFP secretory profiles to chondrocyte responses.¹⁶

APOA4, a secreted glycoprotein predominantly synthesized in the intestine and liver, is classically known for its role in lipid metabolism and anti-inflammatory properties in cardiovascular and metabolic disorders.^17,18 Beyond local joint factors, accumulating evidence indicates that systemic dyslipidaemia and altered lipid and lipoprotein metabolism are linked to osteoarthritis risk and progression, including effects on cartilage structure and chondrocyte phenotype. In particular, perturbations in HDL-related pathways and apolipoprotein profiles have been associated with OA incidence and severity in both experimental models and clinical cohorts. Within this context, APOA4, as a chylomicron- and HDL-associated apolipoprotein with anti-inflammatory and antioxidant properties, is well positioned to connect systemic metabolic status with the joint microenvironment and cartilage extracellular matrix homeostasis. In human studies, circulating apoA-IV associates with cardiometabolic outcomes, and mechanistic work in human endothelial cells shows NF-κB suppression via DHCR24, supporting a broader anti-inflammatory profile relevant to joint disease.¹⁹ Emerging evidence indicates that APOA4 can modulate cellular proliferation, antioxidant responses and cytokine secretion in various cell types.^20,21 These pleiotropic effects suggest that APOA4 may contribute to tissue protection and repair beyond its established systemic functions, warranting investigation in the context of joint degeneration. Although a dedicated APOA4 receptor has not been defined in joint tissues, studies in metabolic and vascular systems suggest that APOA4 can signal via interactions with lipoprotein particles, membrane binding partners and Gαs–cAMP–coupled pathways, thereby modulating oxidative stress and NF-κB–driven inflammation. These mechanisms provide plausible routes through which APOA4 might intersect with Wnt and inflammatory signaling in chondrocytes, a question that remains to be clarified.

We therefore employed RNA-seq to systematically evaluate the transcriptomic changes and implicate the Wnt pathway. Canonical Wnt/β-catenin signaling is a pivotal pathway in skeletal development and cartilage homeostasis.²² Sustained dysregulation of Wnt signaling has been implicated in the progression of osteoarthritis (OA).^23,24 Chondrocytes also integrate mechanical loading via integrin–FAK, YAP/TAZ, and Wnt/β-catenin nodes that govern matrix anabolism/catabolism and inflammatory crosstalk, providing a mechanistic interface between altered joint mechanics and Wnt-related readouts in OA.²⁵ The relationship between APOA4 and Wnt signaling in chondrocytes is unknown; however, proteomic and transcriptomic data hint at a potential link whereby APOA4 up-regulation coincides with alterations in Wnt-related readouts, and our RNA-seq/GSEA shows a negative enrichment trend (NES < 0; nominal P = 0.008; FDR q = 1.00), suggesting suppression rather than activation. Consistent with this rationale, our preliminary observations supported pursuing RNA-seq to interrogate Wnt-related signaling in the current study. Our mechanistic expectation primarily concerns canonical Wnt/β-catenin signaling: we used AXIN2/LEF1 as directional readouts and applied Wnt3a activation (“rescue”) to assess whether APOA4-associated effects align with the β-catenin axis.^26–28 We note that non-canonical Wnt branches (eg, Wnt/PCP, Wnt/Ca²⁺) and their downstream consequences were not systematically interrogated here and remain priorities for future work; accordingly, our Wnt-related conclusions should be interpreted as β-catenin-focused and hypothesis-generating.

In this study, we systematically characterize APOA4 expression in OA patient IFP tissue, synovial fluid and serum, and examine its direct effects on human chondrocytes in vitro. Using gain- and loss-of-function approaches, inflammatory challenge models, and RNA-sequencing, we investigate the potential role of APOA4 and its influence on Wnt/β-catenin signaling. Given that APOA4 is measurable in joint tissues and fluids, we also explore whether OA-associated increases in APOA4 might represent a compensatory, potentially chondroprotective response and provide preliminary support for APOA4 as a biomarker candidate, while fully acknowledging that dedicated diagnostic studies are needed. Building on prior literature and our preliminary transcriptomic signals, we formulated a testable, hypothesis-generating model in which APOA4 may exert chondroprotective effects, at least in part, by modulating—rather than simply activating—canonical Wnt/β-catenin signaling. Accordingly, we pre-specified primary readouts (AXIN2, LEF1, COL2, ACAN, MMP3, MMP13) and employed qPCR together with Wnt3a activation for functional interrogation.

Methods

Ethical Approval and Informed Consent

The study protocol was approved by the Institutional Review Board of Hefei First People’s Hospital. All procedures conformed to the Declaration of Helsinki. Written informed consent was obtained from all participants prior to sample collection.

Clinical Samples and Baseline Characteristics

Participants with radiographic knee osteoarthritis (OA) and non-OA controls were recruited at Hefei First People’s Hospital, Hefei, China. Inclusion and exclusion criteria are detailed below. OA severity on weight-bearing radiographs was graded using the Kellgren–Lawrence (KL) system by two blinded readers; discrepancies were adjudicated by consensus.²⁹ Demographic and clinical characteristics (age, sex, BMI, medication history including intra-articular corticosteroids, and per-donor KL grade) are summarized in Supplementary Table S1.

Inclusion/Exclusion Criteria

OA group (end-stage knee OA undergoing total knee arthroplasty): radiographic KL III–IV; no other inflammatory or metabolic joint diseases; no intra-articular corticosteroid injections within 3 months prior to surgery.

Control group (non-OA; patients undergoing diagnostic arthroscopy or trauma/orthopedic procedures): no clinical or radiographic evidence of OA (KL 0–I); no inflammatory arthritides; no intra-articular corticosteroid injections within 3 months.

Sample Collection and Processing

We collected infrapatellar fat pad (IFP) tissues, synovial fluid (SF), and peripheral serum from OA and control donors with the following numbers: OA—IFP (n = 3), SF (n = 11), serum (n = 11); Controls—IFP (n = 3), SF (n = 11), serum (n = 11).

IFP tissue: harvested intra-operatively under sterile conditions, trimmed, and partitioned. One portion was fixed in 4% paraformaldehyde for hematoxylin–eosin (HE) staining and immunohistochemistry; the other was snap-frozen in liquid nitrogen and stored at −80 °C for molecular assays.

Synovial fluid: aspirated aseptically, centrifuged at 3000 × g for 10 min to remove cells/debris, and supernatants stored at −80 °C.

Serum: peripheral blood collected into coagulation tubes, allowed to clot at room temperature for 30 min, then centrifuged at 2000 × g for 10 min; sera were aliquoted and stored at −80 °C.

All specimens were processed and biobanked within 3 hours of collection to ensure sample integrity.

Histology and Immunohistochemistry

Here, “validation” denotes verification of APOA4 detectability and group-wise differences in serum and synovial fluid by ELISA in an independent set of samples, complementing tissue-level observations. The abbreviation IFP has been defined previously and is used hereafter without re-expansion. IHC staining was quantified using the H-score method, calculated as ∑(Pi × Ii), where Pi is the percentage of cells at each staining intensity (0–3) and Ii is the corresponding intensity.³⁰

Paraffin Embedding and Sectioning of IFP Tissue

Infrapatellar fat pad (IFP) samples fixed in 4% paraformaldehyde were processed by dehydration through a series of graded ethanol concentrations, followed by clearing with xylene and embedding in paraffin. Tissue sections were cut to a thickness of 5 μm using a rotary microtome and then placed on glass slides.

Hematoxylin and Eosin (HE) Staining

The paraffin sections were dewaxed in xylene for two 5-minute intervals, rehydrated through a series of graded ethanol solutions (100%, 95%, and 70%, each for 3 minutes), and rinsed in running tap water. The sections were then stained with hematoxylin for 5 minutes, washed with water, differentiated using 1% acid alcohol, and blued with ammonia water. Eosin staining was applied for 2 minutes, followed by dehydration through graded ethanol solutions (70%, 95%, 100%, each for 3 minutes), clearing in xylene, and mounting under coverslips.

Immunohistochemistry (IHC) for APOA4

Antigen Retrieval: Sections were placed in 10 mM citrate buffer (pH 6.0) and heated at 95°C for 15 minutes, then allowed to cool to room temperature.

Endogenous Peroxidase Blocking: Sections were incubated with 3% hydrogen peroxide in methanol for 10 minutes at room temperature.

Blocking: The sections were incubated with 5% bovine serum albumin (BSA) in PBS for 30 minutes at room temperature.

Primary Antibody Incubation: The slides were incubated overnight at 4°C with a rabbit anti-APOA4 antibody at a 1:200 dilution.

Secondary Antibody Incubation: Sections were then incubated with HRP-conjugated anti-rabbit IgG (1:500) for 1 hour at room temperature.

Chromogenic Detection: DAB (3,3′-diaminobenzidine) was applied for 2–5 minutes until brownish-yellow signals appeared, followed by hematoxylin counterstaining for 1 minute.

Mounting: After dehydration and clearing, sections were mounted with neutral gum.

Semi-Quantitative Scoring

For each section, five randomly chosen fields at 200× magnification were captured. The staining intensity (ranging from 0 to 3) was multiplied by the percentage of positively stained cells (ranging from 0% to 100%) to determine the H-score, with a range of 0 to 300.

Chondrocyte Source and Characterization (C28/I2)

In-vitro assays used the human immortalized chondrocyte line C28/I2. Experiments were performed with passages P3 to minimize passage-related drift. C28/I2 was selected for its reproducible chondrocytic phenotype and responsiveness to matrix/inflammatory cues, with broad use in OA research. Baseline morphology and key markers (eg, COL2, ACAN, MMP13) were checked prior to experiments.^31,32

ELISA for APOA4 Detection

Sample Preparation

Synovial fluid and serum samples were thawed on ice and then centrifuged at 2000 × g for 10 minutes at 4°C to eliminate any precipitates. The resulting supernatants were collected for ELISA analysis.

ELISA Procedure

A commercially available human APOA4 ELISA kit was utilized for the analysis. All reagents and samples were allowed to reach room temperature before use. A standard curve was generated by serially diluting the standard (0–2000 pg/mL), with each concentration measured in duplicate. A total of 100 µL of either the standard or sample was added to the wells of a 96-well plate pre-coated with the capture antibody. The plates were incubated at 37°C for 2 hours.

After discarding the contents, 300 µL of wash buffer was added, and the wells were washed four times. Then, 100 µL of biotin-labeled detection antibody was added to each well and incubated for 1 hour at 37°C.

After four additional washes, 100 µL of streptavidin-HRP was added and incubated for 30 minutes at 37°C.

Following five washes, 100 µL of TMB substrate solution was added and incubated in the dark for 15 minutes. The reaction was stopped with 50 µL of stop solution, and absorbance was measured at 450 nm, with a reference wavelength of 570 nm.

Data Analysis

A four-parameter logistic regression model was employed to generate the standard curve and determine the APOA4 concentrations in the samples. All samples were analyzed in triplicate.

Serum and synovial fluid APOA4 were quantified using a commercial ELISA kit (Cloud-Clone, SEA835Hu, Wuhan, China), following the manufacturer’s instructions. Absorbance was read at 450 nm on SpectraMax iD3 (Molecular Devices, San Jose, CA, USA). Reagent sources and catalog numbers are listed in Supplementary Table S2.

Chondrocyte Culture and APOA4 Treatment

The human chondrocyte line C28/I2 was obtained from Kerafast (Boston, MA, USA; RRID: CVCL_0U51), authenticated by STR profiling, and routinely tested negative for mycoplasma. Recombinant human APOA4 (Sino Biological, 11617-H08H) was reconstituted in PBS/0.1% BSA; the vehicle control was PBS/0.1% BSA at matched concentration across groups.

The human articular chondrocyte cell line C28/I2 was cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin, and incubated in a humidified chamber at 37°C with 5% CO₂. Cells at passages 3 to 6 were used for all experiments.

Before treatment, cells were seeded at a density of 1×10⁵ cells per well in 6-well plates and allowed to adhere overnight. Recombinant human APOA4 was dissolved in double-distilled water (ddH₂O) and added to the culture medium to achieve a final concentration of 5 nM. The control group received an equal volume of PBS.

After 72 hours of treatment, the cells were washed twice with ice-cold PBS and collected for subsequent qRT-PCR analysis.

Overexpression and Silencing Experiments

Overexpression of APOA4 in Chondrocytes

To examine the effects of APOA4 overexpression, C28/I2 chondrocytes were transfected with a plasmid encoding human APOA4 using Lipofectamine 3000, following the manufacturer’s guidelines. Cells were plated in 6-well plates at a density of 1×10⁵ cells per well and incubated overnight. After 24 hours, the cells were collected for protein and RNA extraction, which were then used for Western blot and qRT-PCR analysis. The efficiency of overexpression was verified by qRT-PCR. Details of all siRNAs—including target gene, vendor and catalog/ID, sequences (sense/antisense; 5′→3′), modifications/purification, working concentration, and negative control—are provided in Supplementary Table S3.

Silencing of APOA4 in Chondrocytes

To knock down APOA4 expression, C28/I2 chondrocytes were transfected with siRNA targeting APOA4 using Lipofectamine 3000, following the manufacturer’s instructions. A non-targeting siRNA served as the control. After 48 hours of transfection, the cells were harvested for protein and RNA extraction. The silencing efficiency was confirmed through qRT-PCR.

Full-length human APOA4 cDNA was cloned into pcDNA3.1(+) (custom clone; GenScript, Piscataway, NJ, USA); empty vector served as control. Transfections used Lipofectamine 3000 (Thermo Fisher Scientific, Waltham, MA, USA) per the manufacturer’s protocol.

Knockdown employed chemically synthesized APOA4 siRNAs (Nanjing Jixi Biotechnology Co., Ltd., Nanjing, China; H337-siAPOA4-1 and H337-siAPOA4-2); sequences and ordering information are provided in Supplementary Table S3. Final siRNA concentration was 50 nM unless otherwise stated; a non-targeting siRNA was used as negative control.

RNA Extraction and Quantitative Real-Time PCR (qRT-PCR)

RNA isolation: Total RNA was extracted from treated chondrocytes using the RNeasy Mini Kit, following the manufacturer’s guidelines. The purity and concentration of the RNA were evaluated using a NanoDrop spectrophotometer, with acceptable A₂₆₀/A₂₈₀ ratios between 1.8 and 2.0.

Reverse transcription: A total of 1 µg of RNA was reverse-transcribed into cDNA using the iScript™ cDNA Synthesis Kit in a 20 µL reaction, with incubation at 25 °C for 5 minutes, 46 °C for 20 minutes, and 95 °C for 1 minute.

qRT-PCR: Quantitative PCR was carried out using a CFX96 Real-Time System with iTaq™ Universal SYBR^® Green Supermix. Each 20 µL reaction mixture contained 10 µL of SYBR Green Mix, 0.5 µM forward primer, 0.5 µM reverse primer, and 2 µL of cDNA template. The cycling conditions were as follows: 95 °C for 3 minutes; 40 cycles of 95 °C for 10 seconds and 60 °C for 30 seconds; followed by melt-curve analysis from 65 °C to 95 °C. All samples were tested in triplicate. Primer sequences are provided in Supplementary Table S4.

Data analysis: Relative gene expression was calculated using the 2^–ΔΔCt method, normalizing to GAPDH.

Total RNA was isolated with RNeasy Mini Kit (Qiagen, 74104, Hilden, Germany). RNA quantity/purity were assessed using NanoDrop 2000c (Thermo Fisher Scientific, Waltham, MA, USA). cDNA was synthesized with [RT kit, Company]; qPCR used PowerUp SYBR Green Master Mix (Thermo Fisher Scientific, A25742) on QuantStudio 5. Relative expression was calculated using 2^–ΔΔCt, normalized to GAPDH. Target genes: AXIN2, LEF1, COL2, ACAN, MMP3, MMP13.

Cell Proliferation Assay (CCK-8 and Colony Formation Assay)

CCK-8 Assay

To evaluate cell proliferation, C28/I2 chondrocytes were plated in 96-well plates at a density of 1 × 10⁴ cells per well and treated with either recombinant APOA4 (5 nM) or PBS (as a control) for 24, 48, or 72 hours. Following the treatment period, 10 µL of CCK-8 solution (Dojindo, Cat. No. CK04) was added to each well and incubated for 1 hour at 37 °C. Absorbance was measured at 450 nm using a microplate reader. Cell viability was determined by calculating the ratio of absorbance in the treated wells to that in the control wells.

Colony Formation Assay

C28/I2 chondrocytes were seeded at a density of 500 cells/well in 6-well plates and treated with recombinant APOA4 (5 nM) or PBS (control). After 7 days of incubation, cells were fixed with 4% paraformaldehyde for 15 min, stained with 0.5% crystal violet for 30 min, and washed thoroughly with PBS. Colonies were counted, and the number of colonies formed was recorded. The assay was performed in triplicate.

Inflammatory Model and Treatment with IL-1β and APOA4

Inflammatory Stimulation

To mimic osteoarthritic (OA) conditions, C28/I2 chondrocytes were seeded at 1×10⁵ cells/well in 6-well plates and allowed to adhere overnight. The following day, cells were treated with 10 ng/mL recombinant human IL-1β for 24 hours to induce an inflammatory response. Control cells received an equal volume of PBS.

Combined APOA4 Treatment

After 24 hours of IL-1β stimulation, cells were co-treated with 5 nM recombinant human APOA4 or PBS. Following an additional 72-hour incubation, cells were harvested for protein or RNA extraction. Expression levels of matrix anabolic and catabolic markers were analyzed by qRT-PCR.

Western Blot and qRT-PCR

Protein extraction and RNA isolation were performed as described above. Western blot was used to analyze the expression of COL2 and MMP13, while qRT-PCR was employed to quantify the gene expression of these markers.

RNA Sequencing and Pathway Analysis

RNA Sequencing (RNA-Seq)

C28/I2 chondrocytes were treated with 5 nM recombinant human APOA4 or PBS (control) for 72 hours, after which the cells were harvested for RNA sequencing (RNA-seq) analysis. The raw sequencing data were initially filtered using fastp (v0.23.2) to remove low-quality reads and trim adapter sequences. Clean reads were further processed with in-house scripts to eliminate duplication biases introduced during library preparation and sequencing. Briefly, clean reads were first clustered based on their unique molecular identifiers (UMIs), with reads sharing the same UMI sequence grouped into the same cluster. Pairwise alignment was then performed within each cluster, and reads with over 95% sequence identity were grouped into sub-clusters. Multiple sequence alignment was conducted on the sub-clusters to generate a consensus sequence for each one. These steps removed any errors and biases caused by PCR amplification or sequencing. The resulting sequences were used for standard mRNA-seq analysis. Reads were mapped to the reference genome using STAR software (version 2.7.6a), and the number of reads mapped to exonic regions of each gene and transcript was counted using featureCounts (Subread-1.5.1; Bioconductor). Gene expression was then quantified. Differentially expressed genes and transcripts between groups were identified using the edgeR package (version 3.40.2), with a p-value cutoff of 0.05 and a fold-change cutoff of 2. Gene ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis for differentially expressed genes were conducted using KOBAS software (version 2.1.1) with a p-value cutoff of 0.05. Alternative splicing events were detected using rMATS (v4.1.2) with a p-value cutoff of 0.05 and an absolute Δψ of 0.1. After alignment, reads were assembled de novo using StringTie (v2.1.4), and new transcripts were identified by discarding reads with the class code “=”.

Gene Set Enrichment Analysis (GSEA) was conducted using the GSEA software, with the GO and KEGG background datasets sourced from the MSigDB database. The protein-protein interaction network of differentially expressed genes was analyzed through the STRING database. Transcription factors were annotated using two different databases: the plantTFdb database for plants and the animalTFdb database for animals. Individual SNPs and InDels were identified with bcftools (v1.17) and filtered using the bcftools filter module. The filtered SNPs and InDels were subsequently annotated with Variant Effect Predictor (VEP, v106.1) software. Time-series transcriptome data were clustered using the R package Mfuzz (v2.58.0) to identify groups of genes exhibiting similar expression patterns.

qPCR Validation (Wnt Pathway and Cartilage Phenotype Genes)

To validate the RNA-seq–indicated involvement of the Wnt pathway and assess changes in cartilage-related phenotype genes, we performed qPCR for AXIN2, LEF1, COL2, ACAN, MMP3, and MMP13. Primer sequences are provided in Supplementary Table S4. Gene expression was calculated using the 2^–ΔΔCt method with GAPDH as the internal control. Each condition included n = 3 independent biological replicates with technical triplicates; melt-curve analysis confirmed a single amplification peak. Statistical tests are described in “Statistical analysis”.

Wnt Pathway Activation (Wnt3a Rescue Assay)

Recombinant human Wnt3a (100 ng/mL) was used to activate canonical Wnt/β-catenin signaling. Wnt3a was reconstituted in PBS/0.1% BSA, aliquoted, and stored at −80 °C to avoid repeated freeze–thaw cycles. Chondrocytes were pretreated with Wnt3a for 24 h, then co-incubated at the same concentration with or without APOA4 (5 nM) for a total of 72 h. qPCR readouts included AXIN2, LEF1, COL2, ACAN, MMP3, and MMP13. To maintain ligand activity, culture medium containing Wnt3a ± APOA4 was refreshed every 24 h. The vehicle control was PBS/0.1% BSA, matched across groups. Experimental groups were: vehicle control, APOA4, and APOA4 + Wnt3a. Unless otherwise specified, n = 3 independent biological replicates with technical triplicates were used; statistical procedures are described in “Statistical analysis”.

Statistical Analysis

Unless otherwise specified, data are presented as mean ± SD from independent biological replicates.

For two-group comparisons, we used two-tailed unpaired Student’s t-tests (Welch’s correction applied when variances were unequal). For ≥3 groups with a single factor, we used one-way ANOVA followed by Tukey’s (equal variance/n) or Šidák’s post hoc tests as appropriate. For time-course or two-factor designs, we used two-way ANOVA with Šidák’s post hoc tests. When distributional assumptions were not met, non-parametric alternatives were applied (Mann–Whitney U for two groups; Kruskal–Wallis with Dunn’s tests for ≥3 groups).

For omics-level analyses (eg, RNA-seq), P values were adjusted using the Benjamini–Hochberg false discovery rate (FDR); unless otherwise stated, FDR < 0.05 was considered significant. All tests were two-sided with α = 0.05. Significance in figures is annotated as: P<0.05 (*), P<0.01 (**), P<0.001 (***), P<0.0001 (****); “ns” denotes not significant.

Analyses were performed using GraphPad Prism (v9.1/v10.1) and R (v4.3) with relevant packages (eg, stats, car; for RNA-seq differential expression edgeR as specified in the RNA-seq subsection). Exact n and the statistical test used are reported in the corresponding Figure Legends.

Full details of instruments and software (model, manufacturer, location, and version) are provided in Supplementary Table S5.

Reagent sources and catalog numbers are listed in Supplementary Table S2; siRNA sequences are provided in Supplementary Table S3.

Results

Clinical Sample Analysis

Immunohistochemical (IHC) analysis revealed that APOA4 expression in the infrapatellar fat pad (IFP) tissues of OA patients (n = 3) was significantly greater than in the control group (n = 3) (p < 0.05) (Figure 1A and B). Hematoxylin and eosin (HE) staining of the IFP tissues showed a notably higher level of inflammation in the OA group compared to the controls. IHC staining further emphasized APOA4 expression in the tissue sections, which was substantially increased in the OA group (Figure 1C and D).

Figure 1 Immunohistochemical (IHC) and hematoxylin–eosin (HE) staining of infrapatellar fat pad (IFP) from osteoarthritis (OA) patients and non-OA controls. (A and B) IHC for APOA4 in IFP from control (A) and OA (B) samples. APOA4 staining is markedly increased in OA. Representative images; scale bars = 100 μm. The bar graph on the right shows quantitative analysis of APOA4-positive area (% positive area) corresponding to the IHC images, demonstrating significantly higher APOA4 expression in OA than in controls (n = 3 per group; data are mean ± SD; two-tailed unpaired t test; ****P < 0.0001). (C and D) HE staining of IFP from control (C) and OA (D), illustrating increased inflammatory cell infiltration and stromal alterations in OA. Scale bars = 100 μm.

ELISA Analysis of Synovial Fluid and Serum

ELISA analysis demonstrated that APOA4 levels were significantly higher in both synovial fluid (n = 11) and serum (n = 11) from the OA group compared to the control group (p < 0.05). The increased APOA4 expression in OA samples suggests it may play a role in the pathogenesis of OA (Figure 2).

Figure 2 APOA4 levels in serum and synovial fluid of osteoarthritis (OA) patients and healthy controls determined by ELISA. (A) Serum APOA4 concentrations were significantly higher in OA patients compared with normal controls (p < 0.01). (B) Synovial fluid APOA4 concentrations were markedly elevated in the OA group compared with controls (****p < 0.0001). Data are presented as mean ± SD. n = 11 per group; Statistics: two-tailed unpaired t-test (Welch correction when appropriate); data shown as mean ± SD.

Effect of APOA4 on Matrix Anabolic and Catabolic Markers

APOA4 treatment (5 nM, 72 h) significantly upregulated matrix-anabolic markers and downregulated catabolic mediators in chondrocytes—COL2 (P = 0.0006), ACAN (P = 0.0055), MMP3 (P = 0.0249), and MMP13 (P = 0.0214); two-tailed unpaired t-test, n = 3 (mean ± SD). These findings suggest that APOA4 may exert protective effects on chondrocytes (Figure 3).

Figure 3 Effects of APOA4 on anabolic and catabolic gene expression in chondrocytes. C28/I2 chondrocytes were treated with PBS or recombinant APOA4 (5 nM) for 72 h. Relative mRNA expression levels were quantified by qRT-PCR and normalized to GAPDH. (A) COL2, (B) ACAN, (C) MMP3, and (D) MMP13. APOA4 treatment significantly increased the expression of anabolic markers COL2 and ACAN, while significantly reducing the expression of catabolic enzymes MMP3 and MMP13 compared with PBS-treated controls. Data are presented as mean ± SD from three independent biological replicates (n = 3). Statistical significance was assessed using a two-tailed unpaired Student’s t-test. *P < 0.05, **P < 0.01, ***P < 0.001.

Overexpression and Silencing of APOA4

Chondrocytes were transfected with an APOA4 overexpression plasmid (OE-APOA4) or empty vector (EV). OE-APOA4 significantly increased matrix-anabolic markers (COL2, ACAN) and reduced catabolic mediators (MMP3, MMP13) relative to EV. Data are presented as mean ± SD; exact P values are reported in the panels (significance notation: * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001). See Figure 4. Conversely, siRNA-mediated knockdown of APOA4 (si-APOA4) produced a reciprocal pattern compared with the overexpression experiments. Relative to the non-targeting control (NC), si-APOA4 decreased COL2 and ACAN and increased MMP3 and MMP13. Group data are shown as mean ± SD with exact P values annotated in the panels; the same significance notation is used (* P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001). These mirror-image effects between OE-APOA4 and si-APOA4 support a causal role for APOA4 in maintaining an anabolic matrix program in chondrocytes (see Figure 5).

Figure 4 APOA4 overexpression (OE-APOA4) shifts matrix gene expression in chondrocytes. qPCR results showed that, compared with the empty vector (EV) control, APOA4 overexpression in chondrocytes upregulated the chondroprotective markers COL2 and ACAN while downregulating the catabolic markers MMP3 and MMP13. Bars represent mean ± SD; n = [3] biological replicates. Statistics: [two-tailed unpaired t-test/one-way ANOVA + Tukey, as appropriate]; exact P values are indicated in panels (significance notation: ** P < 0.01; *** P < 0.001; **** P < 0.0001). Abbreviations: OE-APOA4, APOA4 overexpression; EV, empty vector; qPCR, quantitative polymerase chain reaction; SD, standard deviation; ns, not significant.

Figure 5 APOA4 knockdown alters the expression of matrix-related genes in chondrocytes. APOA4 knockdown downregulated the chondroprotective markers COL2 and ACAN, while upregulating the matrix-degrading/catabolic markers MMP3 and MMP13 in chondrocytes. Bars represent mean ± SD; n = [3] biological replicates per group. Statistics: [one-way ANOVA + Tukey/two-tailed unpaired t-test vs NC, as appropriate]; exact P values are indicated in panels (significance notation: * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001). Abbreviations: si-APOA4, APOA4-targeting small interfering RNA; NC, non-targeting control; qPCR, quantitative polymerase chain reaction; SD, standard deviation; ns, not significant.

Cell Proliferation

Both CCK-8 and colony formation assays indicated that APOA4 treatment notably boosted the proliferative ability of chondrocytes. The APOA4-treated group showed a significantly greater number of colonies compared to the control group (p < 0.05) (Figure 6). In the CCK-8 assay, APOA4 treatment for 72 h significantly increased cell viability compared with vehicle (Figure 7), consistent with the transcriptional pro-anabolic shift. The same significance notation is used (* P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001).

Figure 6 Effect of APOA4 on chondrocyte colony formation. Representative images of colony formation assays in chondrocytes treated with PBS (A) or 5 nM APOA4 (B). The number of colonies was significantly higher in the APOA4-treated group compared with the control group (p < 0.05). Two-tailed unpaired Student’s t-test; mean ± SD; n = 3 independent experiments; P<0.05 considered significant.

Figure 7 Effect of APOA4 on chondrocyte proliferation assessed by CCK-8 assay. Chondrocytes were treated with PBS (control) or APOA4 (5 nM) for 24, 48, and 72 h. Cell viability was assessed using the CCK-8 assay, and absorbance was measured at 450 nm. APOA4 treatment significantly increased chondrocyte viability at 72 h compared with the PBS control. Data are presented as mean ± SD from n = 3 independent biological replicates. Statistical analysis was performed using two-way ANOVA (factors: treatment and time), followed by Šídák’s multiple comparisons test. Significance indicates comparison between PBS and APOA4 groups at the same time point (*P < 0.05, **P < 0.01, ***P < 0.001, ***P < 0.0001).

Effects of APOA4 on Inflammatory Response

IL-1β treatment significantly upregulated the expression of MMP3 and MMP13, while downregulating the levels of COL2 and ACAN in chondrocytes. However, co-treatment with APOA4 (5 nM) significantly inhibited the IL-1β-induced increase in MMP3 and MMP13, while maintaining the expression of COL2 and ACAN (p < 0.05) (Figure 8). These findings suggest that APOA4 may help mitigate IL-1β-induced inflammatory damage.

Figure 8 Effects of APOA4 on IL-1β-induced changes in anabolic and catabolic gene expression in chondrocytes. Chondrocytes were treated with PBS, IL-1β, or IL-1β plus APOA4 (5 nM) for 24 h. (A and B) qRT-PCR analysis showed that IL-1β stimulation significantly reduced the expression of anabolic markers COL2 and ACAN, whereas (C and D) it markedly increased the expression of catabolic enzymes MMP3 and MMP13 compared with PBS control. Co-treatment with APOA4 significantly attenuated IL-1β-induced MMP3 and MMP13 upregulation and partially restored COL2 and ACAN expression. Data are presented as mean ± SD from n = 3 independent biological replicates. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post hoc test. Significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001.

RNA-Seq Implicates Wnt Signaling in APOA4-Mediated Transcriptional Responses

Transcriptomic enrichment prioritized the Wnt signaling pathway with a negative GSEA trend (NES < 0; FDR q = 1.00), suggesting suppression rather than activation; this inference is suggestive and interpreted in light of the functional assays. GO/KEGG enrichment analyses identified the Wnt signaling pathway as one of the significantly prioritized pathways. GO/KEGG enrichment prioritized the Wnt signaling pathway with a negative enrichment trend in GSEA (NES<0; nominal P=0.008; FDR q=1.00), suggesting suppression rather than activation (Figures 9 and 10). RNA-seq profiling identified a set of differentially expressed genes (DEGs) in chondrocytes after APOA4 treatment. Using |log₂ fold change| ≥ 1 and FDR < 0.05 as thresholds, we detected 163 up-regulated and 95 down-regulated genes (Figure 9B), consistent with the volcano plot (Figure 9A). GO/KEGG enrichment of DEGs highlighted extracellular-matrix (ECM) organization, inflammatory/immune modulation, and transcriptional regulation (Figure 9C). Hierarchical clustering showed clear separation between APOA4 and control groups (Figure 9D). Full expression matrices, DEG lists, and enrichment outputs are provided in Supplementary Tables S6–S8. At the functional level, GO enrichment analysis of the RNA-seq differentially expressed genes highlighted multiple biological processes and molecular functions related to inflammation and cellular regulation (Figure 10). In parallel, KEGG GSEA suggested a negative enrichment trend for Wnt signaling (NES = −1.314; nominal P = 0.008; FDR q = 1.00), indicating a suppressive rather than activating tendency under the present conditions (Supplementary Table S9). Because the signal did not survive multiple-testing correction, we interpret this pattern as hypothesis-generating, which guided targeted qPCR readouts and a Wnt3a activation assay.

Figure 9 RNA-seq differential expression and functional annotation in APOA4-treated chondrocytes. (A) Volcano plot of DEGs; dashed lines indicate |log2FC| = 1 and FDR = 0.05 thresholds. (B) Counts of up-regulated (163) and down-regulated (95) genes meeting |log2FC| ≥ 1, FDR < 0.05. (C) Bubble plot of GO/KEGG terms enriched among DEGs (bubble size = input gene count; color = rich factor). (D) Heatmap of DEGs showing hierarchical clustering and group separation. Full outputs are provided in Supplementary Tables S6–S8.

Figure 10 Gene Ontology (GO) enrichment analysis of RNA-seq differentially expressed genes in APOA4-treated chondrocytes. Bar plot shows the number of genes mapped to representative enriched GO terms; colors indicate GO domains (Biological Process, Cellular Component, and Molecular Function).

qPCR Validation and Functional Interrogation of Wnt/β-Catenin Signaling

qPCR validation consistent with the RNA-seq trend. Following APOA4 (5 nM, 72 h) treatment, Wnt targets AXIN2 and LEF1 decreased, whereas the anabolic markers COL2 and ACAN increased, and the catabolic mediators MMP3 and MMP13 decreased (Figure 11). These findings support that APOA4 exerts chondroprotective effects at least in part by attenuating Wnt/β-catenin signaling. Statistical tests and sample sizes are reported in the figure legend and “Statistical analysis”. Wnt3a activation partially offset the effects of APOA4. Under Wnt3a (100 ng/mL) activation (co-incubated with APOA4 for 72 h), AXIN2 and LEF1 increased relative to the APOA4 group; meanwhile, COL2 and ACAN decreased compared with APOA4 (but remained overall higher than PBS), whereas MMP3 and MMP13 increased compared with APOA4 (yet remained overall lower than PBS) (Figure 11B). These findings indicate that Wnt3a activation partially reverses the chondroprotective effects of APOA4, supporting that APOA4 acts, at least in part, by attenuating Wnt/β-catenin signaling. Statistical tests and sample sizes are provided in the figure legend and “Statistical analysis”.

Figure 11 APOA4 modulates Wnt/β-catenin targets and matrix-related genes in chondrocytes, and its effects are partially reversed by Wnt3a. (A) qPCR analysis of Wnt target genes (AXIN2, LEF1) and matrix-related genes (ACAN, COL2, MMP3, MMP13) in C28/I2 chondrocytes treated with PBS (control) or 5 nM APOA4 for 72 h. Expression levels were normalized to GAPDH and are shown relative to the PBS group. APOA4 decreased AXIN2 and LEF1, increased ACAN and COL2, and reduced MMP3 and MMP13. (B) Wnt3a rescue experiment. Cells were pretreated with recombinant human Wnt3a (100 ng/mL) for 24 h and then co-treated with or without 5 nM APOA4 for a total of 72 h; qPCR readouts were obtained for the same gene panel as in (A). Wnt3a partially reversed APOA4-induced changes in AXIN2/LEF1 and in matrix-anabolic (ACAN, COL2) and catabolic (MMP3, MMP13) markers. Data are presented as mean ± SD from n = 3 independent biological replicates (each in technical triplicate). In (A), comparisons between PBS and APOA4 were performed using a two-tailed unpaired t test; in (B), one-way ANOVA followed by Tukey’s post hoc test was applied. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Discussion

Osteoarthritis (OA) is a progressive degenerative joint disease characterized by cartilage degradation, subchondral bone remodeling, and synovial inflammation.³³ Despite significant progress in understanding the pathogenesis of OA in recent years, effective therapeutic strategies remain limited.³⁴ In this study, we investigated the role of apolipoprotein A4 (APOA4) in OA, with a focus on its expression profile, protective effects on chondrocytes, and potential underlying mechanisms. This is an exploratory study. While our findings align with the transcriptomic trend, multiple-testing constraints limit causal inference. In conjunction with qPCR and Wnt3a rescue, the data suggest that APOA4 may confer chondroprotection by modulating Wnt/β-catenin, warranting validation in vivo and in larger cohorts.

Increased Expression of APOA4 in OA Samples

Our findings revealed that APOA4 expression was significantly elevated in the infrapatellar fat pad tissue, synovial fluid, and serum of OA patients compared with controls.³⁵ This is consistent with previous proteomic studies that identified increased APOA4 levels in OA tissues, suggesting that APOA4 may be involved in OA pathogenesis.⁸ The elevated APOA4 expression may represent a compensatory response to ongoing cartilage degradation and inflammation, as increased APOA4 has also been reported in various inflammation-related conditions. Notably, our earlier observations were aligned with these patterns and informed the hypotheses tested here. The observed elevation of APOA4 may reflect a compensatory response aimed at counteracting catabolic remodeling and inflammation; alternatively, it may serve as a marker of disease activity without sufficient functional impact in vivo. Our data support the former possibility at the cellular level, yet definitive causality will require in vivo perturbation of the APOA4–Wnt axis.

APOA4 Protects Chondrocytes by Regulating Matrix Synthesis and Degradation

In vitro experiments demonstrated that APOA4 treatment significantly upregulated cartilage matrix anabolic markers COL2 and ACAN, while downregulating catabolic enzymes MMP3 and MMP13.³⁶ This suggests that APOA4 may contribute to the balance between matrix synthesis and degradation, which is critical for maintaining cartilage homeostasis. Additionally, APOA4 enhanced chondrocyte proliferation, further supporting its potential as a therapeutic factor promoting cartilage repair.³⁷ These effects were validated by overexpression and knockdown experiments: APOA4 overexpression increased COL2 and ACAN levels and reduced MMP3 and MMP13 expression, whereas APOA4 knockdown had the opposite effects.

APOA4 Attenuates Inflammatory Damage in Chondrocytes

The inflammatory microenvironment in OA, particularly the presence of IL-1β, exacerbates cartilage destruction by upregulating catabolic enzymes such as MMP3 and MMP13.³⁸ In this study, we found that APOA4 significantly inhibited IL-1β-induced expression of MMP3 and MMP13, while partially preserving the expression of COL2 and ACAN. These results indicate that APOA4 may alleviate inflammatory injury in OA cartilage through its anti-inflammatory properties.

APOA4 and Wnt/β-Catenin Signaling in Chondrocytes

Contrary to a simple activation model, our data suggest that APOA4 is associated with modulation of Wnt/β-catenin signaling, consistent with a negative GSEA trend (NES < 0; FDR q = 1.00, hence suggestive) and partial reversal by Wnt3a.³⁹ Taken together, APOA4 may confer chondroprotection in part by modulating—rather than enhancing—Wnt/β-catenin activity; however, this interpretation remains preliminary pending orthogonal biochemical and in-vivo validation. Our mechanistic exploration was primarily focused on the transcriptional level through qPCR of target genes AXIN2 and LEF1. The absence of protein-level evidence—specifically the expression of total/active $\beta$-catenin and its nuclear translocation—is a significant limitation that warrants cautious interpretation of the APOA4-Wnt axis at this stage.

Potential Therapeutic Implications

This study highlights APOA4’s chondroprotective potential through effects on matrix homeostasis and inflammatory responses, together with suggestive modulation of Wnt/β-catenin signaling (RNA-seq negative GSEA trend, FDR q = 1.00; cellular perturbation by qPCR and partial reversal with Wnt3a). These observations are hypothesis-generating rather than confirmatory. At present, APOA4 should be viewed as a context-dependent candidate modulator of the Wnt pathway and a potential therapeutic avenue for OA.⁴⁰ Biochemical and in-vivo validation—including β-catenin (total/active) Western blotting, TOPFlash reporter assays, inhibitor epistasis (eg, XAV939/IWR-1), studies in primary/explant/3D organoid systems,⁴¹ and DMM/ACLT models—will be required to establish mechanism and translational feasibility. Given the modest human sample size and the in-vitro, single-cell-line nature of our mechanistic data, these translational implications should be interpreted with caution. At present, APOA4 is best regarded as a context-dependent candidate modulator of matrix remodeling and Wnt/β-catenin signaling, rather than as an established therapeutic agent. Confirmation in primary chondrocytes, 3D/ex vivo cartilage models, and in-vivo OA models, as well as in larger clinical cohorts, will be required before any clinical application can be realistically considered.

APOA4 in Lipid Metabolism and Translational Hurdles

Beyond the joint, APOA4 is a well-recognized regulator of lipid metabolism, satiety, and anti-inflammatory responses in cardiovascular and metabolic settings, acting within the gut–liver–adipose axis and circulating lipoprotein pools.⁴² In this context, our data support the notion that APOA4 may represent one node of a broader adipokine/lipid mediator network that influences cartilage homeostasis and OA progression. From a translational standpoint, however, several hurdles remain. First, APOA4 has systemic pleiotropic effects, and it is unclear whether local intra-articular delivery, systemic administration, or modulation of endogenous APOA4 would provide the optimal risk–benefit profile. Second, dose–response and safety have not been characterized in OA models, particularly in the setting of common comorbidities such as obesity and dyslipidemia. Third, inter-individual variability in metabolic status may shape the chondrocyte response to APOA4. These considerations underscore that APOA4 should presently be viewed as a metabolically-linked candidate target, and that rigorous in-vivo, pharmacologic, and safety studies will be essential before any clinical translation can be contemplated. These observations raise the possibility that circulating and synovial APOA4 could be explored as biomarker candidates; however, we did not assess associations with disease severity or longitudinal outcomes, and any diagnostic or prognostic application remains speculative and will require dedicated validation studies. While the specific receptor for APOA4 in chondrocytes remains to be identified, LRP1 is a promising candidate given its documented expression in human chondrocytes and its role in regulating the endocytosis of matrix metalloproteinases. Whether APOA4 directly binds to LRP1 to trigger chondroprotective signaling requires further ligand-binding assays.

Translational Implications

From a translational perspective, our findings raise several testable scenarios. First, in the context of phenotypically heterogeneous OA, APOA4 may help to capture a “metabolic–inflammatory” joint environment, complementing current concepts of metabolic and inflammatory OA phenotypes⁴³ and potentially informing future patient stratification or enrichment strategies in clinical trials.⁴⁴ Second, the observation that APOA4 is detectable in joint tissues and fluids, and is associated with more favorable matrix and inflammatory readouts in vitro, suggests that APOA4-based or APOA4-mimetic approaches might eventually be considered within the broader paradigm of joint-targeted, anti-inflammatory and matrix-supportive intra-articular interventions. In this regard, intra-articular agents such as carboxymethyl chitosan (CM-C), which has shown early pain relief and functional improvement in advanced knee OA patients who are non-responders to hyaluronic acid,⁴⁵ provide a useful framework for thinking about how novel biologically informed treatments could be integrated into clinical practice, including treatment intervals and patient-reported outcomes. Finally, elevated APOA4 in serum and synovial fluid should currently be regarded as a candidate biomarker signal rather than a validated diagnostic tool, as we did not assess correlations with radiographic severity or longitudinal outcomes; however, these preliminary data may guide the design of future studies that evaluate APOA4 as a diagnostic, prognostic or pharmacodynamic biomarker in OA. Altogether, the present work should be viewed as hypothesis-generating and exploratory, providing a clinical–translational framework for future mechanistic and interventional studies rather than definitive evidence of APOA4-targeted therapy.

Methodological Considerations and Limitations

Beyond RNA-seq, we combined qPCR validation with a Wnt3a rescue assay, providing suggestive evidence that APOA4 is associated with Wnt/β-catenin modulation. Nevertheless, we did not complete Western blotting for β-catenin (total/active) and downstream proteins, TOPFlash reporter assays, or Wnt-inhibitor epistasis in this revision; causal inference should therefore remain cautious. Future work will incorporate TOPFlash, protein/nuclear localization of β-catenin/TCF, and bidirectional tests using XAV939/IWR-1 to further validate the role of the Wnt/β-catenin axis in this context.⁴⁶

Limitations

This study has limitations. First, the clinical sample size is relatively small. Second, experiments were conducted in the C28/I2 cell line and have not yet been functionally verified in primary cells or in vivo. Third, although we now provide functional Wnt perturbation at the cellular level, in vivo validation of the APOA4–Wnt axis remains to be done. Finally, the current work should be regarded as exploratory and hypothesis-generating; larger cohorts and animal studies are warranted to generalize these findings. The sample size and in-vitro setting may limit generalizability; the transcriptomic GSEA signal did not survive FDR correction, rendering the inference suggestive. Mechanistic validation in animal models and larger cohorts is needed. Regarding the concentration of recombinant APOA4 used in this study (10ug/mL), this dosage was selected based on our preliminary dose-screening and previously reported anti-inflammatory effective ranges in other mesenchymal cell types. However, we acknowledge that the lack of a complete dose-response curve in the current study limits our ability to determine the absolute optimal concentration for OA treatment. The IL-1β stimulation used here represents a simplified inflammatory context and may not capture the complexity of OA pathophysiology. Follow-up studies will evaluate multi-cytokine conditions (eg, TNF-α/IL-6/IFN-γ), human cartilage explants/3D organoids, and in vivo validation in DMM/ACLT OA models. Our IL-1β-only stimulation represents a simplified inflammatory context and may not capture the multi-factorial nature of OA inflammation, particularly the intricate crosstalk between diverse joint and immune cell populations.⁴⁷ Future work will incorporate conditioned medium from differentiated macrophages to better recapitulate the cytokine and synovial immune microenvironment.⁴⁸ While 72 h is appropriate for transcriptional and proliferation endpoints, longer time points (≥7–14 days) and 3D micromass/hydrogel cultures, human cartilage explants, or engineered osteoarthritic cartilage organoids⁴⁹ will be incorporated to assess matrix accumulation and tissue-level phenotypes with higher physiologic relevance.

Taken together, our data are consistent with the notion that APOA4 may exert protective effects on chondrocytes; however, this interpretation remains preliminary and requires validation at the protein level and in vivo. These constraints mean that our findings should be viewed as preliminary and hypothesis-generating, and their generalizability and translational relevance await confirmation in independent, larger patient cohorts and in-vivo OA models.

Conclusion

APOA4 is upregulated in osteoarthritis and is associated with a chondroprotective phenotype in vitro, characterized by re-balancing of matrix metabolism (↑ COL2/ACAN, ↓ MMP3/MMP13), support of proliferation indices, and attenuation of IL-1β-induced injury. At the signaling level, transcriptomic enrichment, decreased AXIN2/LEF1, and partial reversal by Wnt3a collectively suggest that APOA4 may modulate—rather than activate—canonical Wnt/β-catenin signaling. However, these associations remain correlative and do not establish causality.

Given the small human cohorts, the in-vitro, single-cell-line design, and the incomplete functional interrogation of Wnt/β-catenin (no β-catenin protein readouts, TOP/FOP reporters, or pharmacologic/genetic Wnt blockade yet), our findings should be regarded as preliminary and hypothesis-generating. At present, APOA4 is best viewed as a context-dependent candidate modulator and potential therapeutic target in OA, rather than as an established therapy. Future work will test the APOA4–Wnt axis in vivo (eg, DMM/ACLT models), incorporate protein-level and reporter readouts (β-catenin total/active, TOP/FOP, β-catenin/TCF localization), apply pharmacologic and genetic Wnt perturbation (eg, XAV939/IWR-1, β-catenin knockdown), and evaluate dose–response, safety, and external validation in larger clinical cohorts and primary/3D/ex vivo cartilage systems.