Secreted Phosphoprotein 1 (SPP1) Modulates Intracellular Metabolic Programs in Mature Adipocytes Revealed by Untargeted Metabolomics

TL;DR

This study shows how different amounts of SPP1 affect metabolism in fat cells, with low doses causing more metabolic changes than high doses.

Contribution

The study reveals SPP1's novel role in modulating intracellular metabolic programs in mature adipocytes using untargeted metabolomics.

Findings

Low-dose SPP1 (100 ng/mL) caused more pronounced intracellular metabolic remodeling than high-dose SPP1 (500 ng/mL).

Low-dose SPP1 altered carbohydrate- and amino acid-related metabolic pathways, including the pentose phosphate pathway and alanine metabolism.

High-dose SPP1 selectively induced matrix metalloproteinase-12 (MMP-12), indicating remodeling-associated responses.

Abstract

Secreted phosphoprotein 1 (SPP1), also known as osteopontin (OPN), is a multifunctional secreted factor implicated in inflammatory remodeling of adipose tissue. Here, we applied untargeted liquid chromatography–mass spectrometry (LC–MS)-based metabolomics to examine how different concentrations of SPP1 influence metabolism in differentiated 3T3-L1 adipocytes. We found that low-dose SPP1 (100 ng/mL) was associated with more pronounced intracellular metabolic remodeling than higher-dose SPP1 (500 ng/mL). Multivariate analyses, including principal component analysis (PCA), partial least squares–discriminant analysis (PLS-DA), and sparse PLS-DA, revealed a distinct metabolic profile in the low-dose group, whereas the high-dose group showed a comparatively attenuated metabolic response. Pathway enrichment analysis identified alterations in carbohydrate- and amino acid-related metabolic networks, including the pentose phosphate pathway; pentose and glucuronate interconversions; and alanine, aspartate, and glutamate metabolism. In contrast, higher-dose SPP1 was associated with selective induction of matrix metalloproteinase-12 (MMP-12), suggesting the engagement of remodeling-associated responses at elevated SPP1 levels. Together, these findings identify SPP1 as a modulator of metabolic programs in mature adipocytes and highlight its capacity to influence intracellular metabolic remodeling.

Figures

Full text

1. Introduction

Secreted phosphoprotein 1 (SPP1), also known as osteopontin (OPN), is a multifunctional glycoprotein that integrates inflammatory signaling with cellular functional adaptation across diverse tissues [1]. Unlike classical extracellular matrix (ECM) proteins, SPP1 belongs to the matricellular protein family and primarily functions as a soluble signaling mediator rather than a structural scaffold [2]. Through interactions with multiple receptors, including integrins and CD44, SPP1 activates context-dependent intracellular signaling pathways, such as the phosphatidylinositol 3 kinase (PI3K)–protein kinase B (AKT) pathway and focal adhesion-associated cascades, conferring pleiotropic and cell-type-specific biological effects [3].

In adipose tissue, elevated SPP1 expression is observed under metabolically stressful conditions, including aging and obesity, and is associated with immune cell infiltration, tissue remodeling, and impaired insulin sensitivity [4,5]. Although SPP1 is often enriched in senescent or inflamed adipose tissue, its functional relevance is not limited to senescence-associated inflammation, suggesting a broader role in regulating adipocyte function.

While prior studies have demonstrated that SPP1 influences adipocyte differentiation and lipid accumulation in developmental contexts [6,7], whether SPP1 directly reprograms metabolic pathways in fully differentiated adipocytes remains unclear. In particular, the effects of SPP1 on lipid species composition, central carbon metabolism, and energy-related metabolic networks in mature adipocytes have not been systematically examined.

Given the close interplay between inflammatory signaling and adipocyte metabolism, we hypothesized that SPP1 directly modulates metabolic programs in mature adipocytes. To test this hypothesis, we performed untargeted liquid chromatography–mass spectrometry (LC–MS)-based metabolomic profiling of differentiated 3T3-L1 adipocytes treated with recombinant human SPP1 at low (100 ng/mL) and high (500 ng/mL) concentrations—levels substantially lower than those commonly used in adipogenesis studies [6]. In parallel, we assessed inflammatory and remodeling-associated responses to determine whether SPP1-induced metabolic reprogramming occurs independently of broader inflammatory changes.

2. Results

2.1. SPP1 Treatment Does Not Overtly Alter Adipocyte Differentiation or Lipid Accumulation

Differentiated 3T3-L1 adipocytes were treated with vehicle (distilled water, DW), low-dose (100 ng/mL), or high-dose (500 ng/mL) SPP1 according to the experimental workflow shown in Figure 1a. Representative Oil Red O staining revealed no overt differences in lipid droplet accumulation among treatment groups (Figure 1b), indicating that SPP1 exposure did not markedly affect adipocyte differentiation status or gross lipid storage under the conditions examined. Following confirmation that SPP1 treatment did not overtly alter adipocyte differentiation or lipid accumulation, intracellular metabolites were extracted and subjected to untargeted LC–MS-based metabolomic profiling. Data processing and feature filtering were performed as summarized in Figure 1c, yielding a curated metabolite dataset for downstream analyses.

2.2. Unsupervised PCA Reveals Limited Global Separation of Metabolic Profiles

Using the curated metabolite dataset, principal component analysis (PCA) was first performed to assess global variance and intrinsic sample clustering. The first three principal components (PCs) accounted for 57.1% of the total variance, with PC1, PC2, and PC3 explaining 29.0%, 17.0%, and 11.1% of the variance, respectively (Figure 2a).

Despite this substantially explained variance, PCA score distributions showed considerable overlap among the vehicle (V), low-dose SPP1 (SL), and high-dose SPP1 (SH) groups, with no clear clustering by treatment. Pairwise comparisons between PCs did not reveal statistically significant group separation (PC1 vs. PC2, p = 0.877; PC1 vs. PC3, p = 0.195; PC2 vs. PC3, p = 0.635), indicating that SPP1-induced metabolic alterations are subtle and not sufficient to drive robust global separation in an unsupervised framework.

2.3. Supervised PLS-DA Reveals Dose-Dependent and Biphasic Metabolic Separation

To further interrogate treatment-associated metabolic differences, supervised partial least squares–discriminant analysis (PLS-DA) modeling was applied using a three-component model. In contrast to PCA, PLS-DA score plots demonstrated improved separation among the three groups (Figure 2b,c), indicating the presence of treatment-related metabolic signatures.

Notably, the low-dose SPP1 group (SL) exhibited the most pronounced separation, particularly along component 1, whereas the high-dose SPP1 group (SH) occupied an intermediate position between SL and vehicle (V) controls. This distribution is consistent with a non-linear, dose-dependent metabolic response to SPP1.

To identify metabolites contributing to group discrimination, variable importance in projection (VIP) scores were calculated for component 1. Given that over 130 metabolites exhibited VIP scores exceeding 2.0, a more stringent threshold (VIP > 3.0) was applied to reduce feature redundancy and facilitate clearer interpretation. Several metabolites displayed high VIP scores (>3.0), indicating strong contributions to treatment separation (Figure 2d). The accompanying abundance patterns revealed distinct group-specific trends. For example, petromyzonol showed the highest abundance in the vehicle (V) group and progressively lower levels in the SH and SL groups, whereas p-nitroacetophenone exhibited the highest abundance in the SL group. In contrast, metabolites such as 4-phorbol 12,13-dibutyrate and prednisolone hemisuccinate were the most abundant in SH, intermediate in vehicle, and the lowest in SL. Together, these opposing trends support the presence of biphasic metabolic effects across the SPP1 concentrations.

2.4. Sparse PLS-DA Further Refines Group-Specific Metabolic Features

To further refine the identification of metabolites driving group separation, sparse PLS-DA (sPLS-DA) was performed using the top 10 features and a three-component model. Compared with conventional PLS-DA, sPLS-DA yielded clearer separation among the vehicle (V), low-dose SPP1 (SL), and high-dose SPP1 (SH) groups across multiple component combinations (Figure 3a–c).

Consistent with previous analyses, the SL group occupied a distinct position relative to both the V and SH groups, while the SH group once again showed intermediate or divergent positioning depending on the component examined. In particular, separation along component 2 recapitulated the ordered pattern observed earlier, with samples distributed along a V > SH > SL axis (Figure 3c), further supporting a dose-dependent and non-linear metabolic response to SPP1.

To identify metabolites contributing to this refined separation, loading scores for components 1 and 2 were examined. For component 1, prenyl thioacetate and ethychlozate were among the highest-ranking contributors, whereas hopantenic acid and prenyl thioisobutyrate showed the strongest contributions to component 2 (Figure 3d,e). Notably, p-nitroacetophenone was once again identified among the top contributors in component 2, consistent with its prominence in both the VIP and univariate analyses. The recurrence of these metabolites across independent analytical approaches highlights their potential relevance to SPP1-induced metabolic differences.

2.5. Volcano Plot Analysis Identifies Recurrent Metabolite Changes Associated with Low-Dose SPP1

To further characterize group-wise metabolic differences, univariate volcano plot analysis was performed with a particular focus on comparisons involving the low-dose SPP1 group (SL). In the vehicle versus SL comparison, several metabolites met nominal significance criteria (log_2_|fold change| > 2, raw p < 0.05) (Figure 4a).

Among these metabolites, p-nitroacetophenone, dimethyl sulfate, and hopantenic acid were downregulated in the SL group relative to the vehicle control, whereas 4-phorbol 12,13-dibutyrate, petromyzonol, and prednisolone hemisuccinate were upregulated. These metabolites were among the most strongly shifted features in the comparison, as indicated by both fold-change magnitude and statistical significance. Similarly, the comparison between high-dose SPP1 (SH) and SL revealed a distinct set of differentially abundant metabolites (Figure 4b). Rebimastat, 4-phorbol 12,13-dibutyrate, and prednisolone hemisuccinate were increased in the SH group relative to the SL group, whereas droxidopa, hopantenic acid, and dexamethasone were decreased in the same comparison.

Although no metabolites remained significant following false discovery rate (FDR) correction—likely reflecting the limited sample size (n = 3 per group)—several metabolites, including p-nitroacetophenone, hopantenic acid, 4-phorbol 12,13-dibutyrate, and prednisolone hemisuccinate, were repeatedly identified across univariate and multivariate analyses. The recurrence of these features supports their association with SPP1-induced metabolic differences despite limited statistical power.

2.6. Enrichment and Pathway Analyses Highlight Perturbed Carbohydrate and Amino Acid Metabolism

To further visualize group-specific metabolic patterns, a heatmap was generated using the top 15 metabolites identified via statistical filtering (Figure 5a). Hierarchical clustering based on metabolite abundance revealed clear separation among the three treatment groups, with samples clustering according to SPP1 doses.

An examination of relative metabolite abundances revealed three distinct patterns.

Pattern 1: Metabolites such as hopantenic acid were present at lower levels in the vehicle control compared with both SPP1-treated groups (SL and SH);Pattern 2: Metabolites including prenyl thioacetate were elevated in the SH group relative to both the SL and vehicle groups;Pattern 3: In contrast, metabolites such as prednisolone hemisuccinate and 4-phorbol 12,13-dibutyrate were decreased in the SL group but remained higher in both the SH and vehicle groups.

These distinct abundance patterns further support a non-linear and dose-dependent metabolic response to SPP1 treatment.

The enrichment results were summarized as an overview of enriched metabolite sets, with up to the top 25 pathways displayed according to the default MetaboAnalyst over-representation analysis (ORA) workflow. Among these, five pathways met the statistical significance criteria, including pentose and glucuronate interconversions (KEGG: map00040); the pentose phosphate pathway (KEGG: map00030); alanine, aspartate and glutamate metabolism (KEGG: map00250); tryptophan metabolism (KEGG: map00380); and tyrosine metabolism (KEGG: map00350) (Figure 5b). Pathway impact analysis further highlighted carbohydrate- and amino acid-related pathways as the dominant contributors to the observed metabolic differences, with pentose-related processes showing the highest impact (Figure 5c).

Together, these results indicate that low-dose SPP1 induces a distinct metabolic state characterized by selective alterations in carbohydrate and amino acid metabolic pathways in mature adipocytes relative to high-dose exposure, consistent with a biphasic metabolic response.

2.7. SPP1-Induced Metabolic Alterations Are Not Accompanied by Pronounced Inflammatory Activation

Finally, to determine whether SPP1 treatment elicited inflammatory or tissue remodeling responses under the experimental conditions used, secreted cytokines and ECM-related factors were quantified (Figure 6). The expression levels of pro-inflammatory cytokines, including tumor necrosis factor-alpha (TNF-α), interleukin-6 (IL-6), and interleukin-1 beta (IL-1β), and resistin, were largely comparable across the vehicle, low-dose SPP1 (SL), and high-dose SPP1 (SH) groups (Figure 6a).

Similarly, most ECM remodeling markers, including tissue inhibitor of metalloproteinases-1 (TIMP-1), matrix metalloproteinase-2 (MMP-2), and matrix metalloproteinase-9 (MMP-9), showed no significant differences among treatment groups. A modest but statistically significant increase in matrix metalloproteinase-12 (MMP-12) expression was observed in the high-dose SPP1 (SH) group compared with the vehicle and SL conditions (Figure 6b).

Collectively, these results indicate that SPP1 treatment under the conditions examined does not induce pronounced inflammatory activation or broad ECM remodeling in mature adipocytes and that the metabolic alterations observed are not accompanied by marked inflammatory responses.

3. Discussion

This study identifies SPP1 as a modulator of intracellular metabolic programs in mature adipocytes, with distinct metabolic responses observed at different SPP1 concentrations. Using untargeted LC–MS-based metabolomics, we show that low-dose SPP1 (100 ng/mL, SL) elicits a metabolically distinct state characterized by altered intracellular metabolite composition and pathway engagement relative to the vehicle and higher-dose conditions (500 ng/mL, SH). Importantly, these metabolic changes occurred without detectable alterations in adipocyte differentiation status or broad pro-inflammatory cytokine secretion, indicating that SPP1 can directly influence adipocyte metabolism rather than acting indirectly through overt inflammatory activation.

To our knowledge, this is the first study to apply untargeted metabolomics to examine SPP1-induced metabolic remodeling in fully differentiated adipocytes. A central contribution of this work is the demonstration that SPP1 directly reprograms intracellular metabolic pathways in these cells. Although SPP1 is frequently discussed in the context of obesity, aging, and senescence-associated tissue dysfunction, the present study does not employ senescence models or assess canonical senescence markers. Instead, our findings establish a cell-autonomous metabolic effect of SPP1, independent of developmental differentiation or generalized inflammatory responses, in mature adipocytes. This distinction is important, as prior studies have largely focused on SPP1-driven adipogenesis, immune cell recruitment, or tissue-level remodeling, leaving unresolved whether SPP1 can directly modulate core metabolic networks in differentiated adipocytes [6,7].

At the level of individual metabolomic features, low-dose SPP1 was associated with pronounced shifts in several metabolites exhibiting large fold changes and strong statistical support. For example, p-nitroacetophenone, dimethyl sulfate, and hopantenic acid were reduced relative to the vehicle control, whereas 4-phorbol 12,13-dibutyrate, petromyzonol, and prednisolone hemisuccinate were increased. The comparison between high- and low-dose SPP1 further revealed a distinct set of differentially abundant features, including increased rebimastat expression and decreased droxidopa and dexamethasone concentrations (Figure 4). Because these features span diverse chemical classes and primarily serve as markers of global metabolomic separation, interpretation was primarily guided by pathway-level enrichment analyses. The enrichment heatmap highlights coordinated metabolite changes within pathways, providing a more robust and biologically interpretable overview of SPP1-induced metabolic remodeling in mature adipocytes (Figure 5).

At the pathway level, low-dose SPP1 preferentially perturbed carbohydrate- and amino acid-related metabolic networks, including the pentose phosphate pathway (PPP); pentose and glucuronate interconversions; and alanine, aspartate, and glutamate metabolism. The PPP plays a central role in adipocyte metabolism by supporting nicotinamide adenine dinucleotide phosphate (NADPH) generation for reductive biosynthesis and redox homeostasis and has been implicated in adipose tissue inflammation and insulin sensitivity [8,9]. Nicotinamide adenine dinucleotide phosphate (NADP) exists in reduced (NADPH) and oxidized (NADP^+^) forms [10]. However, the present study does not directly assess the NADPH/NADP^+^ balance, reactive oxygen species, glucose flux, or glucose-6-phosphate dehydrogenase (G6PD) activity. Therefore, while the data indicate PPP-associated metabolite remodeling in response to SPP1, functional consequences related to redox regulation or biosynthetic capacity remain to be determined.

Altered pentose and glucuronate interconversion further suggests modulation of carbohydrate-handling pathways that are suppressed in insulin-resistant states [11], and these pathways have been linked to aging-associated metabolic signatures, including circulating glucuronic acid levels predictive of biological aging and longevity [12]. While these associations raise the possibility that SPP1-responsive metabolic pathways intersect with aging-related metabolic programs, such interpretations remain speculative in the absence of targeted functional or in vivo validation.

Enrichment of alanine, aspartate, and glutamate metabolism implicates amino acid handling as an additional axis of SPP1-induced metabolic remodeling. Adipose tissue actively participates in systemic amino acid exchange and interconversion [13,14,15], and the dysregulation of these pathways has been linked to insulin resistance and metabolic diseases [16]. Our findings indicate that SPP1 alters intracellular amino acid-related metabolic networks in adipocytes; however, whether these changes reflect altered uptake, transamination, or mitochondrial utilization cannot be resolved without targeted flux analyses.

In contrast to the pronounced intracellular metabolic remodeling observed at lower SPP1 concentrations, higher SPP1 exposure was associated with selective induction of MMP-12 (Figure 6). While MMP-12 has been implicated in adipose tissue inflammation and remodeling [17,18], the present data do not address its downstream functional impact. The observed induction of MMP-12 is therefore interpreted as an adipocyte-derived remodeling-associated signal rather than evidence of overt ECM reorganization or macrophage-driven tissue remodeling.

Several limitations of this study should be acknowledged. The modest sample size and the absence of metabolites reaching FDR significance limit statistical inference. In addition, the lack of functional assays assessing glucose uptake, lipid turnover, mitochondrial respiration, or the redox state constrains mechanistic interpretation. Furthermore, this study was conducted using murine 3T3-L1 adipocytes, and species-specific differences in adipocyte physiology and metabolism may limit direct extrapolation to human adipose tissue. While established cell lines such as 3T3-L1 are experimentally tractable and cost-effective, freshly isolated adipocytes allow for the evaluation of diverse in vivo conditions that may not be fully recapitulated in cell line-based systems [19]. Nevertheless, the consistency of pathway-level trends across analytical approaches and the reproducibility of key enrichment patterns support the robustness of the observed dose-dependent metabolic responses.

Future studies incorporating targeted metabolic flux analyses and functional assays will be required to determine how SPP1-induced metabolite remodeling translates into physiological changes in adipocyte function. In addition, whether SPP1-induced metabolic remodeling in mature adipocytes is affected under senescent or aging-associated conditions remains an important question for future investigation. Extending these analyses to primary human adipocytes and in vivo models, which provide greater physiological relevance but are accompanied by donor-specific heterogeneity [20], will further clarify the translational relevance of SPP1-mediated metabolic remodeling. Together, such approaches will provide insights into the contextual factors that shape SPP1-driven metabolic responses.

In summary, this study expands the functional landscape of SPP1 by identifying it as a direct regulator of metabolic programs in mature adipocytes. Using untargeted LC–MS-based metabolomics, we demonstrate that SPP1 modulates intracellular metabolic states in a concentration-dependent manner, with distinct metabolic responses observed at different SPP1 concentrations. Importantly, these metabolic changes occurred without detectable alterations in adipocyte differentiation status or broad pro-inflammatory cytokine secretion, indicating that SPP1 can directly influence adipocyte metabolism rather than acting indirectly through overt inflammatory activation. To our knowledge, this is the first study to apply untargeted metabolomics to define SPP1-induced metabolic remodeling in fully differentiated adipocytes.

4. Materials and Methods

4.1. Cell Culture and Differentiation

In this study, we used 3T3-L1 adipocytes as a standardized and widely accepted in vitro model for studying adipocyte biology and metabolic regulation, owing to the availability of highly developed and reproducible differentiation protocols [21]. Murine preadipocyte 3T3-L1 cells were obtained from the Korean Cell Line Bank (KCLB, 10092.1, Lot No. 20235) and cultured for subsequent differentiation and treatment experiments as part of this study. Cells were differentiated for three weeks using a modified protocol based on the widely accepted method described in [22], as previously applied in our study [23]. Briefly, the 3T3-L1 preadipocytes were maintained in high-glucose DMEM (Gibco, Cat. No. 11965, Waltham, MA, USA) containing 10% newborn calf serum (Gibco, 26010) and 1% Penicillin–Streptomycin (10,000 U/mL; Gibco, 15140). The medium was changed every 48 h. Differentiation was initiated with a medium containing 10% FBS (Gibco, 10082), 1.7 μM bovine insulin (Sigma-Aldrich, Cat. No. I0516, St. Louis, MO, USA), 1 μM dexamethasone (Sigma D4902), and 0.5 mM 3-isobutyl-1-methyl xanthine (Sigma I7018). The medium was replaced every 48 h with DMEM containing 10% FBS and 1.7 μM bovine insulin.

4.2. SPP1 Preparation and Treatment

Human recombinant SPP1 (also known as osteopontin, OPN; PeproTech, Cat. No. 120-35-50UG, Lot No. 1108467 F0324, Cranbury, NJ, USA) was dissolved in distilled water to prepare a 50 μg/mL stock solution and stored at 4 °C. Three treatment groups were established as follows:

• (1)Vehicle control (V), treated with distilled water;
• (2)Low-dose SPP1 (SL), treated with 100 ng/mL SPP1;
• (3)High-dose SPP1 (SH), treated with 500 ng/mL SPP1.

Treatments were administered twice at 48 h intervals, and all treatments were performed in biological triplicates.

4.3. Treatment Timeline and Sample Collection

Following completion of the three-week differentiation period, the culture medium was replaced (designated as day 0), and cells were stabilized for an additional 2 days. The first SPP1 treatment was administered from day 2 to day 4 by supplementing the culture medium with SPP1. Afterward, the medium was replaced, followed by another 2-day stabilization period. The second SPP1 treatment was performed from day 6 to day 8 in the same manner. On day 8, the medium was replaced again, and the cells were stabilized for an additional 2 days.

On day 10, culture supernatants were first collected for secreted factors analysis, clarified via centrifugation to remove cellular debris, snap-frozen in liquid nitrogen, and stored at −80 °C until further use. Subsequently, the adherent cells were washed twice with phosphate-buffered saline (PBS), scraped, and collected using centrifugation. The resulting cell pellet (more than 10^7^ cells) was snap-frozen in liquid nitrogen and stored at −80 °C until further analysis. The overall experimental timeline is illustrated in Figure 1a. Intracellular lipid accumulation was assessed via Oil Red O staining using a commercial kit (Oil Red O Stain Kit, Abcam, Cat. No. ab150678, Cambridge, UK), which was performed according to the manufacturer’s instructions, and the stained lipid droplets were visualized under a light microscope (Figure 1b).

4.4. Untargeted Metabolomics via LC-MS/MS and Data Processing

Untargeted metabolomics analysis was outsourced to Ebiogen Inc. (Seoul, Republic of Korea). Cell pellets from nine biological samples were submitted for analysis. Metabolite profiling was performed using LC-MS/MS (Q-Exactive, Thermo Fisher Scientific, Waltham, MA, USA) equipped with a nano-UHPLC system operated in positive ion mode.

4.4.1. Metabolite Extraction

Cells were lysed in 500 μL of 50 mM ammonium bicarbonate using an ultrasonicator. Protein concentrations were measured using a BCA Protein Assay Kit (Thermo Fisher Scientific), and 200 μg of total protein from each sample was used for metabolite extraction. The volume was adjusted to 250 μL with 50 mM ammonium bicarbonate, followed by the addition of 1000 μL of 100% methanol. Samples were vortexed for 1 min and incubated at −20 °C for 1 h. After centrifugation at 14,000× g for 10 min, the supernatant was transferred to a new tube and dried using a speed vacuum concentrator. Dried metabolites were reconstituted in 100 μL of 0.1% formic acid in water and filtered through a 0.22 μm spin filter; then, 1 μL of the resulting solution was injected for LC-MS/MS analysis.

4.4.2. LC-MS/MS Analysis

Chromatographic separation was performed using a PepMap™ 100 trapping column (C18, 3 μm, 100 Å, 75 μm × 2 cm) in combination with a PepMap™ RSLC analytical column (C18, 2 μm, 100 Å, 75 μm × 25 cm). The mobile phases consisted of solvent A (water with 0.1% formic acid) and solvent B (80% acetonitrile with 0.1% formic acid).

The gradient elution was programmed as follows (time in minutes/% solvent B): 0/4%, 5/4%, 32/28%, 46/50%, 46.1/96%, 50.1/96%, 50.4/4%, and 60/4%.

The flow rate was set to 400 nL/min, and mass spectra were acquired over an m/z range of 100 to 1000.

4.4.3. Data Processing

Raw LC-MS data were processed using Compound Discoverer™ 3.3 (Thermo Fisher Scientific). The overall workflow is schematically summarized in Figure 1c.

Feature selection was based on the identification criteria of Metabolomics Standards Initiative (MSI) Levels 2 and 3, with background signals from the blank and solvent samples excluded. Redundant features were removed based on the average peak area intensity across samples.

4.5. Multiplex Profiling of Secreted Cytokines and Extracellular Matrix-Related Proteins

Secreted cytokines and extracellular matrix(ECM)-related proteins in culture supernatants were quantified using a bead-based multiplex immunoassay based on Luminex technology. The assay was performed as a contract research service by KOMA Biotech (Seoul, Republic of Korea) using the Mouse Premixed Multi-Analyte Kit (Luminex Assay; R&D Systems, Cat. No. RND-LXSAMSM-14; LXSAMSM panel) according to the manufacturer’s instructions. The multiplex panel included the following mouse analytes: adiponectin, MCP-1/CCL2, IL-6, IL-1β, IL-10, IL-16, MMP-2, MMP-9, MMP-12, TIMP-1, RANTES/CCL5, resistin, TNF, and GDF-15.

The culture supernatants collected on day 10 were thawed on ice, gently mixed, and clarified via centrifugation at 10,000× g for 4 min at 4 °C to remove cellular debris. Samples were diluted at a ratio of 1:2 using the assay buffer provided by the kit and analyzed in technical triplicates. Standards and blanks were prepared in duplicate using the supplied assay buffer. Sample incubation with the bead–antibody mixture was carried out for 2 h at room temperature with gentle agitation, followed by washing and detection steps performed in accordance with the manufacturer’s protocol.

Data acquisition was performed using a Luminex detection system (Luminex, Austin, TX, USA), and analyte concentrations were calculated from standard curves generated for each target using the Quantist software (Bio-Techne, Minneapolis, MN, USA), as provided by the service provider. All procedures were conducted in strict accordance with the provided protocol.

4.6. Data Analysis

The processed metabolomics data were subsequently analyzed using MetaboAnalyst 6.0 (https://www.metaboanalyst.ca) (accessed on 30 December 2025). Preprocessing steps included median normalization, log_10_ transformation, and Pareto scaling to reduce technical variation and enhance comparability across samples (Figure 1c). Statistical analyses included principal component analysis (PCA), partial least squares–discriminant analysis (PLS-DA), sparse PLS-DA (sPLS-DA), variable importance in projection (VIP) score calculation, unpaired t-tests (p < 0.05, |fold change| > 2), and hierarchical clustering. Volcano plots were generated to identify differentially expressed metabolites. Pairwise comparisons were performed among the following groups: V (vehicle-treated control), SL (low-dose SPP1, 100 ng/mL), and SH (high-dose SPP1, 500 ng/mL).

For secreted cytokine and extracellular matrix-related protein analyses, quantitative data were visualized and statistically analyzed using the GraphPad Prism software (GraphPad Software, San Diego, CA, USA; version 10). Data are presented as the mean ± SEM, and statistical significance was assessed using unpaired two-tailed Student’s t-tests, as indicated in the figure legends.