Integrative Metabolomics and Systems Pharmacology Reveal PPARγ-Centered Antidiabetic Mechanisms of Caulerpa racemosa and Its Bioactive Compounds

Abstract

Type 2 diabetes mellitus (T2DM) is a complex metabolic disorder requiring safe, multitarget therapeutic strategies. Marine macroalgae represent an underexplored source of bioactives with pleiotropic metabolic effects. This study investigated the antidiabetic potential of an ultrasound-assisted ethanolic extract of Caulerpa racemosa (UAECr) and its key phytosterol, campesterol, through an integrative framework combining metabolomics, network pharmacology, molecular docking, molecular dynamics simulation, and in vitro validation. Untargeted ultra-high-performance liquid chromatography–high-resolution mass spectrometry (UHPLC–HRMS) metabolomics characterized UAECr constituents, followed by in silico bioactivity prediction, target-network analysis, molecular docking, and 100 ns molecular dynamics simulation of the peroxisome proliferator-activated receptor gamma (PPARγ)–campesterol complex. Functional validation was performed in differentiated 3T3-L1 adipocytes assessing glucose uptake, PPARγ expression, dipeptidyl peptidase 4 (DPP-4) inhibition, and cytotoxicity. Metabolomics identified campesterol as a prominent bioactive. Network pharmacology highlighted PPARγ as a central hub, supported by strong docking affinity of campesterol toward PPARγ (−11.4 kcal/mol) and DPP-4 (−8.3 kcal/mol). Molecular dynamics simulations demonstrated stable PPARγ–campesterol interactions, with preserved protein compactness and low residue fluctuation. In vitro, UAECr and campesterol significantly enhanced glucose uptake (up to 134% vs. control, p < 0.001), upregulated PPARγ expression (4-fold, p < 0.0001), and moderately inhibited DPP-4 activity (p < 0.01) without cytotoxicity. C. racemosa-derived extracts and campesterol exert antidiabetic effects primarily via stable PPARγ-mediated insulin sensitization with complementary DPP-4 modulation, supporting its potential as a marine-derived functional food candidate.

1. Introduction

Diabetes mellitus (DM) is a chronic metabolic disorder that now affects over half a billion people globally, and its prevalence continues to rise [1]. Type 2 diabetes (T2D) is the most common form (roughly 80–90% of cases) and is characterized by insulin resistance, hyperglycemia, and progressive β-cell dysfunction [1]. The growing “diabesity” epidemic is a major public health challenge, contributing significantly to cardiovascular disease, kidney failure, neuropathy, and mortality. Despite advances in treatment, managing T2D remains difficult because its pathophysiology is complex and multi-factorial [1]. Standard antidiabetic drugs—including insulin sensitizers, secretagogues, and incretin-based agents—often have limited long-term efficacy and can cause serious side effects [1,2]. For instance, metformin frequently induces gastrointestinal intolerance, sulfonylureas can provoke hypoglycemia and weight gain, and thiazolidinediones (TZDs) are linked to cardiovascular risks and edema [1]. In addition, the cost of lifelong pharmacotherapy is high, especially in low-income settings [2]. These limitations underscore the urgent need for safer, more affordable, and multitarget therapeutic alternatives [1,2].

One promising solution is the development of natural-product-based therapeutics. Nutraceuticals and phytochemicals often exert pleiotropic effects—for example, by reducing oxidative stress and inflammation while simultaneously modulating metabolic pathways [1]. Indeed, a growing body of literature highlights marine organisms, particularly macroalgae (seaweeds), as underexplored sources of novel antidiabetic agents [3,4]. Marine macroalgae live in extreme environments and produce unique metabolites (e.g., sulfated polysaccharides, terpenes, polyphenols, carotenoids, and sterols) to adapt to stress, and many of these compounds have demonstrated antioxidant and anti-inflammatory activities [3,5]. Remarkably, seaweed extracts have shown strong potential to alleviate metabolic syndrome features in preclinical studies [1,2], yet they have been less studied than terrestrial plants. Given their richness in bioactive molecules and generally favorable safety profiles, marine algae are gaining attention for drug discovery and functional food development [2,3].

The green seaweed Caulerpa racemosa (commonly called “sea grapes”) is a widespread edible alga with a rich phytochemical repertoire. It contains proteins and soluble fibers (sulfated polysaccharides) as well as large amounts of polyphenols, flavonoids, and pigment molecules [3,5]. Caulerpa spp. also produce various long-chain fatty acids and phytosterols; for example, green algae are known to contain sterols such as clionasterol and poriferasterol, implying that related sterols like campesterol may be present [3]. These constituents confer potent antioxidant capacity. In fact, one study found C. racemosa extracts to be especially rich in phenolics and radical-scavenging activity compared to many other seaweeds [3]. In addition to antioxidant effects, Caulerpa extracts have shown promising metabolic benefits. Dietary or in vitro administration of C. racemosa preparations has been reported to improve lipid and glucose profiles in animal models [5]. In vitro assays demonstrate that carotenoid-rich extracts of C. racemosa can inhibit key carbohydrate-digesting enzymes: for example, fractions from C. racemosa strongly inhibit α-glucosidase and α-amylase [6]. Such enzyme inhibition is a validated antidiabetic mechanism. Moreover, the holistic mixture of Caulerpa metabolites may target critical regulators of glucose homeostasis. For instance, the nuclear receptor peroxisome proliferator-activated receptor gamma (PPARγ) is a master regulator of insulin sensitivity and lipid metabolism, and many natural products can act as partial agonists of PPARγ [7]. Separately, the enzyme dipeptidyl peptidase 4 (DPP-4) degrades incretin hormones and is a proven antidiabetic drug target (inhibitors like sitagliptin/linagliptin improve insulin secretion). Recent molecular docking studies have even predicted that Caulerpa-derived compounds (e.g., the alkaloid caulerpin) can bind and inhibit DPP-4 [8]. Together, these findings suggest that C. racemosa harbors diverse bioactives (polyphenols, carotenoids, fatty acids, sterols) that may modulate multiple diabetes-relevant pathways, including PPARγ and DPP-4.

Because T2D involves many interacting pathways, modern anti-diabetic research increasingly uses multi-target and systems-level approaches. Network pharmacology and metabolomics tools can capture the complexity of botanical extracts by linking their chemical constituents to multiple protein targets [4,9]. In this study, we employ an integrated strategy: untargeted ultra-high-performance liquid chromatography–high-resolution mass spectrometry (UHPLC–HRMS) is used to catalog the major metabolites in the ultrasonic-assisted aqueous extract of C. racemosa (UAECr). These identified compounds are then subjected to in silico screening (bioactivity prediction, network pharmacology and molecular docking) against T2D-related targets (with a focus on PPARγ and DPP-4). Finally, we perform in vitro validation by testing UAECr (and pure campesterol) in cell-based assays for PPARγ activation and DPP-4 inhibition. Such an integrative pipeline—combining metabolomics, computational modeling, and biological assays—has proven powerful for elucidating multi-component drug mechanisms [4,9].

This study aims to evaluate the antidiabetic potential of UAECr and its principal phytosterol, campesterol, through a multi-target investigative framework integrating untargeted metabolomic profiling, in silico network pharmacology, molecular docking, molecular dynamics, and cell-based experimental validation. To the best of our knowledge, this is the first study to comprehensively combine these complementary approaches to elucidate the molecular mechanisms by which C. racemosa bioactives, particularly campesterol, modulate key antidiabetic targets, including PPARγ and DPP-4. By bridging natural product chemistry with systems pharmacology, this integrative strategy provides novel mechanistic insights into the antidiabetic activity of C. racemosa and advances its scientific rationale as a functional food–derived candidate for metabolic disease management.

2. Results

2.1. Untargeted Metabolomic Profiling of UAECr Reveals Diverse Bioactive Constituents

Untargeted UHPLC–HRMS analysis successfully characterized the UAECr. As summarized in

Table 1. Untargeted metabolomic profiling of UAECr identified by UHPLC–HRMS.

Name	Formula	Annot. Delta Mass [ppm]	Calc. MW	m/z	RT [min]	Reference Ion	Peak Area
4-Methoxycinnamic acid	C10H10O3	0.91	178.06316	179.07043	13.433	[M + H] + 1	12,621,041.3
Campesterol	C28H48O	0.61	400.37076	383.36747	16.762	[M + H − H2O] + 1	10,411,215.1
Alpha-Linolenic acid	C18H30O2	0.49	278.22472	279.23199	11.677	[M + H] + 1	393,577,014.0
Ferulic acid	C10H10O4	0.97	194.0581	177.05481	8.723	[M + H − H2O] + 1	79,514,797.3
Oleic acid	C18H34O2	0.59	282.25605	283.26331	11.709	[M + H] + 1	28,504,328.8
Palmitoleic acid	C16H30O2	0.71	254.22476	255.23206	10.312	[M + H] + 1	27,168,919.7
Linoleic acid	C18H32O2	0.97	280.2405	263.23721	12.515	[M + H − H2O] + 1	26,824,332.9
5-Nitrovanillic acid	C8H7NO6	0.84	213.02751	212.02022	2.289	[M − H] − 1	8,048,046.28
Palmitoleic acid	C16H30O2	0.84	254.22479	255.23209	13.57	[M + H] + 1	5,428,443.73
Gallicynoic acid H	C16H28O4	0.13	284.1988	283.19168	7.066	[M − H] − 1	4,628,172.78
N-Cyclohexyl-4-(N-heptanoylisoleucyl)-2-methyl-1-piperazinecarboxamide	C25H46N4O3	−2.49	450.35587	468.38971	11.19	[M + NH4] + 1	103,416,066.0
DG(12:0/22:6(4Z,7Z,10Z,13Z,16Z,19Z)/0:0)[iso2]	C37H60O5	0.45	584.44434	585.4516	13.866	[M + H] + 1	75,471,393.6
2E,4E-undecadienoic acid	C11H18O2	0.83	182.13083	183.13811	11.749	[M + H] + 1	50,136,169.2

, a diverse range of metabolites was identified, encompassing phenolic acids, sterols, fatty acids, and lipid-derived compounds. Major constituents included α-linolenic acid, oleic acid, linoleic acid, palmitoleic acid, ferulic acid, and campesterol, with high mass accuracy (Δppm < ±1) and distinct retention times, supporting confident putative annotation. Among the detected compounds, campesterol emerged as a notable sterol component, exhibiting a substantial peak area and a characteristic dehydration ion ([M + H − H₂O]⁺), consistent with sterol fragmentation behavior. The presence of multiple unsaturated fatty acids and phenolic compounds suggests a metabolomic profile associated with metabolic regulation, lipid homeostasis, and insulin sensitivity.

2.2. Quantitative Determination of Campesterol

HPLC-ELSD analysis confirmed the presence of campesterol in UAECr, with a retention time consistent with that of the authentic reference standard. The campesterol peak exhibited satisfactory chromatographic resolution and reproducibility. Quantitative analysis revealed that UAECr contained 3.42 ± 0.18 mg campesterol per g extract. The analytical method demonstrated excellent linearity across the calibration range (R² = 0.998). Precision assessment showed good repeatability with a relative standard deviation (RSD) of 2.1%. The method sensitivity was supported by a limit of detection (LOD) of 0.74 µg/mL and a limit of quantification (LOQ) of 2.25 µg/mL. These validation parameters indicate that the HPLC-ELSD method is reliable, precise, and sufficiently sensitive for campesterol quantification in UAECr.

2.3. In Silico Bioactivity, Toxicity, and Drug-Likeness Screening Identifies Campesterol as a Key Candidate

To prioritize bioactive constituents for downstream analysis, in silico bioactivity prediction, toxicity assessment, and drug-likeness evaluation were conducted for major UAECr metabolites (

Table 2. In silico bioactivity prediction, toxicity assessment, and drug-likeness evaluation of major metabolites identified in UAECr.

Compounds/Peptides	Pa Score			Toxicity Model Computation Analysis		Drug-Likeness
	Insulin Promoter	Antidiabetic	Cholesterol Antagonist	Predicted LD50 (mg/kg)	Toxicity Class	Lipinski Rule	Pfizer Rule	GSK
4-Methoxycinnamic acid	0.459	0.376	0.602	2980	5	Accepted	Accepted	Accepted
Campesterol	0.332	NA	0.955	890	4	Accepted	Rejected	Rejected
Alpha-Linolenic acid	0.597	0.412	0.852	10,000	6	Accepted	Rejected	Rejected
Ferulic acid	0.346	0.274	0.604	1772	4	Accepted	Accepted	Accepted
Oleic acid	0.648	0.243	0.798	1190	4	Accepted	Rejected	Rejected
Palmitoleic acid	0.648	0.243	0.798	48	2	Accepted	Rejected	Rejected
Linoleic acid	0.597	0.284	0.804	10,000	6	Accepted	Rejected	Rejected
5-Nitrovanillic acid	NA	NA	0.521	270	3	Accepted	Accepted	Accepted
Gallicynoic acid H	0.590	NA	0.698	1000	4	Accepted	Accepted	Accepted
N-Cyclohexyl-4-(N-heptanoylisoleucyl)-2-methyl-1-piperazinecarboxamide	0.380	NA	NA	1500	4	Accepted	Accepted	Rejected
DG(12:0/22:6(4Z,7Z,10Z,13Z,16Z,19Z)/0:0)[iso2]	NA	0.349	0.921	37,000	6	Rejected	Rejected	Rejected
2E,4E-undecadienoic acid	0.430	0.290	0.654	3389	4	Accepted	Rejected	Accepted

NA, not applicable.

). Several compounds demonstrated predicted insulin-promoting and antidiabetic activities, with Pa scores exceeding the preliminary screening threshold. Notably, campesterol exhibited a high predicted cholesterol antagonist activity (Pa = 0.955) and acceptable acute toxicity parameters (predicted LD₅₀ = 890 mg/kg; toxicity class 4). While campesterol violated certain drug-likeness rules related to lipophilicity, its safety profile and bioactivity prediction supported its selection as a representative lead compound. Other fatty acid derivatives, including α-linolenic and linoleic acids, also showed favorable antidiabetic-related Pa scores but were deprioritized due to broader target promiscuity and physicochemical limitations.

2.4. Network Pharmacology and Pathway Enrichment Highlight PPARγ-Centered Antidiabetic Mechanisms

Network pharmacology analysis revealed that UAECr-derived metabolites interact with a tightly connected protein network implicated in glucose and lipid metabolism (Figure 1A). Key hub targets included PPARγ, DPP4, AKT serine (AKT2), protein tyrosine phosphatase non-receptor type 1 (PTPN1), and glucokinase (GCK), indicating a multi-target regulatory profile rather than a single-target mechanism. Gene Ontology enrichment analysis demonstrated that the predicted targets were significantly associated with biological processes related to chemical and organic substance response, metabolic regulation, and cellular homeostasis(Figure 1B). Molecular function analysis further highlighted enrichment in binding activity, catalytic activity, and nuclear receptor-related functions, supporting transcriptional regulation of metabolic pathways (Figure 1C). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment revealed significant associations with general metabolic pathways, insulin signaling-related pathways, advanced glycation end-products (AGEs)—receptor for advanced glycation end-products (RAGE) signaling in diabetic complications, and apoptosis-related signaling (Figure 1D). Although canonical insulin signaling pathways were not the top statistically enriched KEGG categories, enrichment analysis is inherently data-driven and reflects the distribution of overlapping predicted targets. Many enriched pathways identified in this study (e.g., metabolic pathways, AGE–RAGE signaling, apoptosis, and nuclear receptor-related processes) are functionally integrated with insulin sensitivity, adipocyte differentiation, and PPARγ-mediated transcriptional regulation. Because PPARγ itself is a master regulator of adipogenesis and insulin responsiveness, modulation of this nuclear receptor may influence insulin signaling indirectly rather than through direct enrichment of classical insulin pathway components. Therefore, the enrichment profile observed here supports a broader metabolic regulatory mechanism rather than isolated pathway activation.

Collectively, these findings suggest that UAECr bioactives may exert antidiabetic effects through coordinated modulation of insulin sensitivity, glucose uptake, and inflammatory-metabolic crosstalk, with PPARγ as a central regulatory node.

2.5. Molecular Docking Supports Strong Binding of Campesterol to PPARγ and DPP-4

Molecular docking analysis was conducted to evaluate the binding affinity of UAECr-derived metabolites toward key antidiabetic targets (

Table 3. Molecular docking scores (kcal/mol) of UAECr-derived metabolites against key antidiabetic protein targets.

Compounds/Peptides	DPP4 (PDB 3G0B)	PPARγ (PDB 2HFP)	GCK (PDB 1V4S)	PTPN1 (PDB 2HNP)	AKT2 (PDB 1MRV)
Sitagliptin (Control)	−8.2	NA	NA	NA	NA
Pioglitazone (Control)	NA	−8.5	NA	NA	NA
Metformin (Control)	NA	NA	−5.0	−5.7	−5.1
4-Methoxycinnamic acid	−6.7	−6.1	−6.5	−5.7	−6.8
Campesterol	−8.3	−11.4	−7.6	−8.1	−6.9
Alpha-Linolenic acid	−6.1	−6.4	−7.1	−5.3	−6.2
Ferulic acid	−6.7	−6.6	−6.8	−5.8	−6.8
Oleic acid	−5.1	−6.1	−6.5	−4.9	−5.8
Palmitoleic acid	−5.4	−6.0	−6.4	−5.2	−6.0
Linoleic acid	−5.7	−6.5	−6.2	−5.4	−5.9
5-Nitrovanillic acid	−7.1	−6.0	−6.6	−5.9	−6.1
Gallicynoic acid H	−6.0	−5.9	−7.0	−5.5	−6.0
N-Cyclohexyl-4-(N-heptanoylisoleucyl)-2-methyl-1-piperazinecarboxamide	−7.2	−8.5	−7.7	−7.1	−6.4
DG(12:0/22:6(4Z,7Z,10Z,13Z,16Z,19Z)/0:0)[iso2]	The ligand was not subjected to docking analysis because it exceeded the acceptable limit of rotatable bonds (n = 30), which may significantly reduce docking accuracy.
2E,4E-undecadienoic acid	−5.5	−5.8	−6.2	−5.0	−6.0

NA, not applicable.

). Among all tested compounds, campesterol demonstrated the strongest binding affinity to PPARγ (−11.4 kcal/mol), exceeding that of the reference agonist pioglitazone (−8.5 kcal/mol). Campesterol also exhibited favorable docking scores toward DPP-4 (−8.3 kcal/mol) and PTPN1 (−8.1 kcal/mol), supporting its multi-target interaction profile. In contrast, smaller phenolic and fatty acid compounds displayed moderate docking affinities, while highly flexible lipid species were excluded due to excessive rotatable bonds. These docking results corroborate the network pharmacology findings and support the prioritization of campesterol for experimental validation. Based on its consistently strong binding affinity toward PPARγ and DPP-4, favorable safety profile, and commercial availability as a well-characterized phytosterol, campesterol was selected as a representative lead compound for subsequent in vitro experimental validation alongside the whole extract (UAECr). The inclusion of campesterol in the in vitro assays was intended to enable a direct comparison between the biological effects of the complex extract and those of a single, commercially accessible bioactive constituent.

Molecular docking analysis revealed that campesterol derived from UAECr exhibited favorable binding affinities toward both DPP-4 (PDB ID: 3G0B) and PPARγ (PDB ID: 2HFP) (

Table 4. Molecular Docking Interactions of Campesterol Derived from UAECr with DPP-4 and PPARγ.

Campesterol with DPP4 (PDB 3G0B)
Campesterol with PPARG (PDB 2HFP)

). The docking scores indicated energetically stable ligand–receptor complexes, supporting the potential bioactivity of campesterol against glucose metabolism-related targets. In the DPP-4 complex, campesterol was predicted to occupy the catalytic binding pocket, where stabilization was primarily mediated through hydrophobic interactions with residues commonly associated with substrate recognition. Additional van der Waals contacts contributed to ligand anchoring within the active site environment. Although campesterol lacks strong polar functional groups, weak hydrogen bond interactions were observed, suggesting supplementary electrostatic stabilization.

For PPARγ, campesterol demonstrated a binding orientation consistent with accommodation inside the ligand-binding domain (LBD). The interaction pattern was dominated by hydrophobic contacts with key residues lining the LBD, indicating compatibility with the receptor’s lipophilic cavity. This binding mode resembles the characteristic interactions of natural lipid-like ligands, implying that campesterol may act as a modulatory ligand rather than a full agonist.

Overall, the docking profiles support the hypothesis that campesterol may contribute to the antidiabetic potential of UAECr through dual targeting mechanisms involving DPP-4 inhibition and PPARγ modulation. These computational findings align with the experimental in vitro observations, where campesterol-containing extracts demonstrated metabolic regulatory effects.

2.6. Cytotoxicity Assessment Demonstrates Low Toxicity of UAECr and Campesterol

The cytotoxic effects of the UAECr, campesterol, and the reference drug pioglitazone were evaluated in 3T3-L1 cells using the MTT assay. As shown in Figure 2A, treatment with UAECr and campesterol at concentrations ranging from 6.25 to 50 µg/mL for 24 h did not induce any significant reduction in cell viability compared with the vehicle control. Cell viability remained consistently above 95% across all tested concentrations, indicating a favorable safety profile. A comparative analysis among treatment groups (Figure 2B) further confirmed that neither UAECr nor campesterol exerted cytotoxic effects within the biologically active concentration range. Pioglitazone also exhibited minimal cytotoxicity, consistent with its known pharmacological profile. These results demonstrate that subsequent functional assays were conducted at non-cytotoxic concentrations.

2.7. UAECr and Campesterol Enhance Glucose Uptake in Differentiated Adipocytes

The effects of UAECr and campesterol on insulin-stimulated glucose uptake were assessed in fully differentiated 3T3-L1 adipocytes using the 2-deoxy-2-[(7-nitro-2,1,3-benzoxadiazol-4-yl)-amino]-D-glucose (2-NBDG) assay. As illustrated in Figure 3A, both UAECr and campesterol enhanced glucose uptake in a concentration-dependent manner. Significant increases were observed starting from 12.5 µg/mL, with maximal stimulation achieved at 50 µg/mL. Campesterol induced a more pronounced effect than the crude extract, reaching approximately 134% of control levels at the highest concentration. Comparative analysis across treatment groups (Figure 3B) revealed that the glucose uptake–enhancing effect of campesterol was comparable to that of the positive control pioglitazone, while UAECr exhibited a moderate but significant enhancement. These findings indicate that both the extract and its major sterol constituent improve cellular glucose utilization, consistent with an insulin-sensitizing mechanism.

2.8. Upregulation of PPARγ Expression by UAECr and Campesterol

To further elucidate the molecular basis underlying the enhanced glucose uptake, PPARγ mRNA expression was quantified by reverse transcription quantitative polymerase chain reaction (RT-qPCR). As shown in Figure 4A, treatment with UAECr and campesterol resulted in a dose-dependent upregulation of PPARγ expression in differentiated adipocytes. Campesterol markedly increased PPARγ transcript levels, reaching an approximately four-fold induction at 50 µg/mL relative to the vehicle control. UAECr also significantly elevated PPARγ expression, although to a lesser extent than the purified compound. Comparative visualization (Figure 4B) suggested that campesterol consistently outperformed the crude extract across all tested concentrations, with effects comparable to or exceeding those of pioglitazone. These results support the involvement of PPARγ activation in mediating the observed metabolic effects.

2.9. UAECr and Campesterol Exhibit Moderate DPP-4 Inhibitory Activity

The inhibitory effects of UAECr and campesterol on DPP-4 enzymatic activity were evaluated using a fluorometric assay. As depicted in Figure 5A, both treatments inhibited DPP-4 activity in a concentration-dependent manner, although the magnitude of inhibition was lower than that of the reference inhibitor sitagliptin. Campesterol exhibited stronger inhibitory activity than UAECr, achieving approximately 56% inhibition at 50 µg/mL, whereas the extract reached around 46% inhibition at the same concentration. The comparative analysis shown in Figure 5B highlights the distinct mechanistic profiles of the tested compounds, as pioglitazone displayed minimal DPP-4 inhibition, consistent with its role as a PPARγ agonist rather than an incretin-based modulator. These findings indicate that UAECr and campesterol possess moderate DPP-4 inhibitory activity that may complement their insulin-sensitizing effects.

SwissADME Analysis Highlights Bioactive Rather than Drug-like Properties of Campesterol

In silico absorption, distribution, metabolism, and excretion (ADME) and drug-likeness analysis using SwissADME (Figure 6) revealed that campesterol possesses high lipophilicity and low aqueous solubility, resulting in limited predicted gastrointestinal absorption. While campesterol violated certain drug-likeness rules, its bioavailability score and pharmacokinetic profile support its classification as a bioactive functional compound rather than a conventional small-molecule drug. These properties are consistent with the observed in vitro efficacy and support the potential of campesterol as a nutraceutical or functional food-derived bioactive targeting metabolic regulation.

2.10. Molecular Dynamics Simulation Analysis of the PPARγ–Campesterol Complex

The molecular dynamics simulation over 100 ns demonstrated that the PPARγ–campesterol complex remained structurally stable throughout the trajectory (Figure 7). The radius of gyration (Rg) profile showed minimal variation, indicating preserved protein compactness despite minor anisotropic fluctuations along individual spatial axes. Solvent-accessible surface area (SASA) analysis revealed moderate oscillations around a stable baseline, suggesting consistent solvent exposure without evidence of major conformational opening or collapse. Residue-level flexibility assessed by root mean square fluctuation (RMSF) indicated generally low to moderate fluctuations across most residues, with higher mobility confined to terminal regions and flexible loops, while residues within the core and ligand-binding domain remained relatively rigid. In addition, salt bridge analysis displayed a nearly constant profile, reflecting stable electrostatic interactions within the protein, and the Lennard–Jones short-range interaction energy fluctuated within a narrow range, supporting the stability of non-bonded van der Waals interactions. Collectively, these results indicate that campesterol binding does not disrupt the global fold of PPARγ and instead supports a dynamically stable complex under physiological-like conditions.

3. Discussion

This integrative analysis revealed that UAECr and its principal sterol, campesterol, exert antidiabetic effects through a dual-target mechanism involving both PPARγ activation and DPP-4 inhibition (Figure 8). Untargeted metabolomics identified campesterol among several bioactives in UAECr, and network pharmacology highlighted PPARγ as a central hub of the predicted target network (Figure 1). This PPARγ-centric profile aligns with molecular docking: campesterol docked favorably into the PPARγ ligand-binding domain (Table 3), suggesting it can act as a partial agonist, while also fitting into the DPP-4 active site. PPARγ is a key regulator of glucose and lipid metabolism, and its activation enhances insulin sensitivity and glucose uptake [10,11]. DPP-4 is a validated T2DM target that degrades incretins (glucagon-like peptide-1 or GLP-1; gastric inhibitory polypeptide or GIP); inhibiting DPP-4 prolongs incretin action and lowers blood glucose [12]. The concurrent engagement of both targets could synergistically improve glycemic control, as PPARγ agonists enhance glucose utilization while DPP-4 inhibitors sustain insulinotropic signals.

Campesterol emerged as the major bioactive constituent within UAECr. In vitro, campesterol was consistently more potent than the crude extract at 50 μg/mL; it achieved 56% DPP-4 inhibition (versus 46% for UAECr) and increased insulin-stimulated glucose uptake by 134% of control (versus 118% for UAECr). Correspondingly, campesterol upregulated PPARγ gene expression to a 3–4-fold extent at high dose, substantially higher than the 2-fold increase seen with UAECr alone. These findings indicate that campesterol likely drives much of the extract’s activity. Phytosterols like campesterol and β-sitosterol have been reported to activate PPAR pathways [13]; our results extend this by showing campesterol also directly inhibits DPP-4. Notably, previous work predicted dockings of campesterol to DPP-4 but found little in vitro inhibition by crude phytosterols [14]. In contrast, we demonstrate significant DPP-4 inhibition by purified campesterol, supporting our docking and network predictions of its dual targeting. Thus, campesterol is a promising lead molecule for antidiabetic development.

Our results are in line with prior studies on seaweeds and plants that modulate similar pathways. Marine carotenoids and polyphenols have demonstrated PPARγ activation; for example, fucoxanthin from brown algae activates PPARγ, promoting fatty acid oxidation and glucose uptake [10]. Brown-algal phlorotannins have also been shown to inhibit DPP-4 and other metabolic enzymes [15]. Terrestrial plants offer parallels; extracts of Thymelaea hirsuta and Terminalia chebula were recently reported to act as dual PPARα/γ agonists and to enhance glucose uptake in adipocytes [16]. Like UAECr, these natural extracts raised cellular glucose uptake without undesirable adipogenesis, underscoring the therapeutic advantage of balanced PPAR modulation. The convergence of our findings with these reports suggests that multi-component extracts from algae or plants can work through complementary mechanisms to improve glucose homeostasis.

Importantly, in silico analyses supported the experimental data. Network pharmacology predicted that UAECr metabolites target PPARγ-centered pathways involved in lipid or glucose metabolism. This is consistent with reports that network-based analyses of herbal antidiabetics often identify PPAR-related pathways as key hubs [16]. We also used molecular docking to validate these targets. Campesterol showed strong binding poses and favorable binding energies with both PPARγ and DPP-4 (comparable to other known agonists/inhibitors). As noted in the literature, combining network pharmacology with docking is a robust strategy to link compounds to targets [16]. For example, Lin et al. found that over 90% of predicted active plant compounds docked well to antidiabetic targets in silico, confirming their network predictions [16]. The docking results thus lend credence to the proposed mechanism, campesterol can physically interact with PPARγ and the DPP-4 active site, explaining the dual-target effects.

Molecular dynamics simulations further strengthened the docking-based hypothesis by demonstrating that the PPARγ–campesterol complex remained dynamically stable throughout the 10 ns trajectory. The relatively constant radius of gyration (Rg) indicates preserved global compactness of the PPARγ structure upon ligand binding, suggesting that campesterol does not induce destabilizing conformational strain. Consistently, SASA fluctuations were moderate and centered around a stable baseline, implying that ligand engagement does not promote abnormal exposure or collapse of the ligand-binding domain. RMSF revealed that structural fluctuations were largely confined to terminal and loop regions, whereas residues within the PPARγ ligand-binding pocket exhibited low mobility, supporting sustained ligand anchoring. In parallel, the stable salt bridge profile and narrowly fluctuating Lennard–Jones short-range interaction energy indicate persistent electrostatic and van der Waals interactions that contribute to complex stabilization. Collectively, these MD findings suggest that campesterol behaves as a structurally compatible PPARγ ligand, maintaining receptor integrity while supporting a stable binding mode, a dynamic profile consistent with partial agonist behavior. This dynamic stability provides mechanistic support for the observed upregulation of PPARγ expression and enhanced glucose uptake in adipocytes, reinforcing PPARγ-mediated insulin sensitization as a central molecular mechanism underlying the antidiabetic activity of C. racemosa bioactives.

The in vitro assays corroborated the safety and efficacy of UAECr and campesterol at active doses. MTT viability tests showed that neither UAECr (≤50 µg/mL) nor the comparator pioglitazone impaired adipocyte survival (cell viability remained ≈95–100%). This indicates that the pro-glucose effects occur without cytotoxicity, an essential criterion for therapeutic development. In glucose uptake assays, UAECr and especially campesterol significantly enhanced insulin-stimulated uptake in 3T3-L1 adipocytes. Campesterol at 25–50 µg/mL elevated uptake by 30–35% above control, on par with the effect of pioglitazone, confirming potent insulin-sensitizing activity. Consistent with this, quantitative PCR showed dose-dependent upregulation of PPARγ expression by both treatments. Campesterol again outperformed UAECr at 50 µg/mL, which induced a 4-fold increase in PPARγ mRNA versus a 2-fold for the extract (Figure 3). These increases in PPARγ levels likely underlie the enhanced GLUT4-mediated uptake. On the enzyme side, the DPP-4 inhibition assay revealed that campesterol dose-dependently inhibited DPP-4, whereas the crude extract was less active. This DPP-4 inhibition further supports an incretin-mediated mechanism complementing PPARγ effects. In summary, the biological assays validate that UAECr and campesterol are safe and exert multiple insulin-sensitizing actions in vitro.

Although campesterol produced measurable DPP-4 inhibition (≈56% at 50 µg/mL), this effect is moderate and substantially weaker than that of synthetic DPP-4 inhibitors such as sitagliptin. An IC₅₀ could not be reliably determined within our tested concentration range, and therefore the enzymatic data should be interpreted as supportive rather than definitive evidence of potent DPP-4 blockade. Published work on phytosterols reports similarly weak or absent DPP-4 effects. Gupta et al. (2020) found that major phytosterols (stigmasterol, β-sitosterol and campesterol by docking) did not translate to meaningful in vitro inhibition (IC₅₀ values for stigmasterol and β-sitosterol were >50 mg/mL under their assay conditions), concluding phytosterol supplements are unlikely to provide clinically relevant DPP-4 inhibition [14]. Reviews of plant-derived DPP-4 inhibitors emphasize that potent DPP-4 activity is more commonly associated with polyphenols, peptides, or glycosylated terpenoids rather than simple sterols, which more often show partial or weak inhibition at high concentrations [12,17]. Taken together, our data and the literature support a mechanistic interpretation in which campesterol acts predominantly via PPARγ-mediated insulin sensitization (strong docking to PPARγ, stable MD behavior, ~4-fold PPARγ mRNA up-regulation, and robust glucose uptake enhancement), with only moderate and likely ancillary DPP-4 inhibition. Thus, the term “PPARγ-centered with complementary DPP-4 modulation” better reflects the weight of the evidence than describing campesterol as an equally balanced dual-target agent.

Although the theoretical concentration of campesterol delivered by UAECr at 50 µg/mL (~0.17 µg/mL, based on 3.42 mg/g extract) is markedly lower than the effective range observed for pure campesterol (12.5–50 µg/mL), UAECr represents a complex phytochemical matrix rather than a single-compound preparation. The extract contains multiple bioactive classes, including unsaturated fatty acids and phenolic compounds, which are known to influence insulin sensitivity, inflammatory signaling, and PPAR-related pathways. Consequently, the biological effects of UAECr are unlikely to be attributable solely to campesterol concentration, but may instead reflect additive or synergistic interactions among co-existing metabolites. The significant enhancement of glucose uptake and upregulation of PPARγ expression observed with UAECr supports the contribution of this multi-component profile. Such behavior aligns with the established concept of phytochemical synergy in botanical and marine-derived extracts. Future studies involving fractionation, reconstitution, and combination analyses would help clarify individual versus cooperative metabolite contributions.

Strengths of this study include the multi-pronged approach linking chemical profiling, computational prediction, and biological validation. Our untargeted metabolomics guided the focus on campesterol, and network pharmacology placed its activity in context of a PPARγ-centered antidiabetic network. Docking then confirmed plausible direct interactions, and in vitro tests closed the loop by demonstrating functional efficacy on glucose uptake and enzyme inhibition. This integrative pipeline is aligned with best practices for natural product drug discovery [16].

Limitations should be noted. All functional assays were in vitro, using cell lines and recombinant enzyme with a limited number of independent biological experiments; in vivo efficacy and bioavailability of campesterol or UAECr remain to be tested. Campesterol is lipophilic and may have limited solubility or metabolism issues in vivo. Also, the whole extract is complex and batch-variable. Future work should include animal diabetic models to confirm efficacy and safety, and pharmacokinetic studies to assess campesterol exposure. Nevertheless, the consistency of our multi-level data strengthens confidence in the mechanism. Despite these limitations, the convergence of computational, metabolomic, and experimental observations provides a coherent, hypothesis-generating framework for the proposed mechanism.

Finally, this study findings highlight campesterol as a key bioactive from C. racemosa with dual PPARγ-agonist and DPP-4-inhibitor properties. This dual-target profile is attractive for diabetes management because it combines improved insulin sensitivity with enhanced incretin action. The study supports further investigation, one path is a nutraceutical formulation enriched in campesterol or standardized UAECr for glycemic support. Another is medicinal chemistry to optimize campesterol analogs with greater potency or better drug-like properties. Our results provide a clear rationale and starting point for such lead optimization. By elucidating the molecular underpinnings, this work advances C. racemosa and campesterol toward potential therapeutic use in T2D.

4. Materials and Methods

4.1. Sequential Maceration and Ultrasound-Assisted Extraction of Caulerpa racemosa

Fresh Caulerpa racemosa was sourced from a certified commercial aquaculture farm operated by PT Sarana Mukti Sustainable Nutrient (Jepara, Central Java, Indonesia). The seaweed was cultivated under controlled pond conditions and harvested at commercial maturity. Upon arrival at the laboratory, samples were rinsed with distilled water, air-dried, and ground into powder for subsequent extraction procedures.

A sequential extraction approach combining maceration followed by ultrasound-assisted extraction was employed to obtain the UAECr. Briefly, 1 g of dried and powdered C. racemosa simplicia was initially subjected to maceration with 10 mL of an ethanol–water mixture (80:20, v/v; ethanol ≥ 70%, Sigma-Aldrich, St. Louis, MO, USA) at room temperature under continuous gentle stirring for 24 h to facilitate preliminary solvent penetration and extraction of readily soluble compounds. The macerate was then filtered, and the solid residue was retained for further extraction.

Subsequently, the residue obtained from the maceration step was re-extracted using ultrasound-assisted extraction. The residue was resuspended in 10 mL of fresh ethanol–water solution (80:20, v/v) and subjected to sonication using a JOANLAB UC-60 ultrasonic cleaner (JoanLab Indonesia, Jakarta, Indonesia) operating at a frequency of 40 kHz and an effective power of 60 W. Sonication was performed for 10 min, during which the extraction temperature was maintained at 40 °C using an external water bath to enhance cell wall disruption while minimizing thermal degradation of heat-sensitive pigments and bioactive compounds.

Following ultrasound treatment, the extract was clarified by filtration through Whatman Grade 1 filter paper (Whatman/Cytiva, Marlborough, MA, USA), and the combined filtrates from the maceration and ultrasound steps were pooled. The resulting extract was concentrated under reduced pressure at 40 °C using a Büchi R-210 rotary evaporator (Büchi Labortechnik AG, Flawil, Switzerland) equipped with a V-850 vacuum controller and B-491 heating bath (Büchi Labortechnik AG, Flawil, Switzerland) until approximately 5 mL of concentrated extract remained. The concentrated extract was then stored at −20 °C until further analysis.

4.2. Metabolomic Profiling of UAECr via UHPLC-ESI-MS/MS

Untargeted metabolomic profiling of the UAECr was performed at Corpora Science Research Laboratory, Yogyakarta, Indonesia, using a UHPLC–HRMS platform (Thermo Scientific, Waltham, MA, USA) [18,19]. Chromatographic separation was carried out using a Thermo Scientific™ Vanquish™ Horizon UHPLC system equipped with a binary pump. An Accucore™ Phenyl-Hexyl analytical column (100 × 2.1 mm, 2.6 μm particle size) was employed and maintained at 40 °C throughout the analysis. The mobile phase consisted of (A) MS-grade water containing 0.1% formic acid and (B) MS-grade acetonitrile containing 0.1% formic acid (Sigma-Aldrich, St. Louis, MO, USA), delivered at a flow rate of 0.3 mL/min. The gradient elution program started at 5% B, increased linearly to 90% B over 16 min, held at 90% B for 4 min, and then returned to 5% B until a total run time of 25 min. The injection volume was 5 μL. High-resolution mass spectrometric detection was conducted using a Thermo Scientific™ Orbitrap™ Exploris 240 HRMS operated in Full MS/data-dependent MS² (dd-MS²) acquisition mode with polarity switching (ESI+ and ESI−). Full MS scans were acquired at a resolution of 60,000 FWHM over an m/z range of 70–800, with a maximum injection time of 100 ms and an intensity threshold of 5000. Data-dependent MS² spectra were collected at a resolution of 30,000 FWHM, using normalized collision energies of 30, 50, and 70, with nitrogen as the collision gas. Ionization was achieved using an Optamax™ NG heated electrospray ionization (H-ESI) source, with spray voltages set at 3500 V (positive mode) and 2500 V (negative mode). The sheath, auxiliary, and sweep gas flows were set to 35, 7, and 1 arbitrary units, respectively. The ion transfer tube and vaporizer temperatures were maintained at 300 °C and 320 °C, respectively.

Raw UHPLC–HRMS data were processed using Thermo Scientific™ Compound Discoverer™ software (version 3.5) for peak detection, spectral deconvolution, alignment, and annotation. Retention time alignment tolerance was set to 0.2 min, and mass tolerance was set to 5 ppm. Features detected in blank samples were excluded to minimize background interference. Peak intensities were normalized using total ion current (TIC) normalization to correct for injection variability and instrument signal drift. Relative abundance values were calculated based on normalized peak areas and used for comparative profiling of metabolite intensity. Compound annotation was performed through spectral matching against the mzCloud™ and related high-resolution spectral libraries. Only features meeting the following confidence criteria were retained: best match mzCloud spectral similarity score > 99%; mass accuracy with delta mass (Δppm) approaching 0 (≤±1 ppm); and consistent isotopic pattern and characteristic MS/MS fragmentation behavior. Putative identification was assigned according to Level 2 confidence (MSI guidelines), based on accurate mass and high-quality MS/MS spectral matching.

4.3. Quantification of Campesterol by HPLC-ELSD

Quantitative determination of campesterol in UAECr was performed using high-performance liquid chromatography coupled with an evaporative light scattering detector (HPLC-ELSD), a method suitable for non-UV-absorbing phytosterols. Chromatographic separation was achieved on a reversed-phase C18 column (250 × 4.6 mm, 5 µm) maintained at 35 °C. The mobile phase consisted of methanol–water (95:5, v/v) delivered at a flow rate of 1.0 mL/min under isocratic conditions. The ELSD was operated under the following conditions: nebulizer gas (nitrogen) pressure, 3.5 bar; drift tube temperature, 40 °C; gain, 8. Campesterol reference standard (≥98% purity; Sigma-Aldrich, St. Louis, MO, USA) was used to construct an external calibration curve over the concentration range of 5–200 µg/mL. UAECr samples were dissolved in methanol, filtered through a 0.22 µm membrane filter, and injected at a volume of 20 µL. Quantification was based on peak area integration relative to the calibration curve. Method validation parameters including linearity, precision, limit of detection (LOD), and limit of quantification (LOQ) were evaluated following standard analytical guidelines. Campesterol content was expressed as mg per g extract (mg/g UAECr).

4.4. Molecular Docking and In Silico Analysis of UAECr

4.4.1. Bioactivity Prediction and Pharmacokinetic Profiling

The therapeutic potential of phytochemicals identified from UAECr against T2D was evaluated using the PASS (Prediction of Activity Spectra for Substances) platform provided by WAY2DRUG (https://way2drug.com/PassOnline/, accessed 20 December 2025) [20]. This web based tool employs a structure activity relationship (SAR) algorithm that compares submitted molecular structures with a curated library of biologically characterized compounds to estimate probable pharmacological activities. Compounds exhibiting a probability of activity (Pa) greater than 0.7 were classified as strong candidates, while a Pa threshold of 0.4 was applied for preliminary screening [21]. Higher Pa values indicate greater predictive reliability. To further assess pharmaceutical suitability, pharmacokinetic behavior, toxicity risk, and drug-likeness were analyzed using ADMET descriptors and Lipinski’s Rule of Five [22,23]. Molecular features were derived from canonical SMILES notations and processed using ADMETLab 3.0 and ProTox-III platforms, with compound information obtained from the PubChem database [22,24,25].

4.4.2. Protein Target Identification and Analysis

Putative protein targets of UAEcr-derived compounds were predicted using the SuperPred server version December 2025 (https://prediction.charite.de/; accessed on 20 December 2025) [26]. Canonical SMILES representations of each compound were submitted to the SuperPred server. Predicted targets were ranked according to probability scores generated by the SuperPred algorithm. To focus on high-confidence interactions and minimize spurious associations inherent to similarity-based prediction tools, only top-ranked targets (probability score ≥ 80%) were retained for subsequent network analysis. This filtering strategy was applied to enhance specificity and ensure a manageable and biologically interpretable target network. Genes and proteins implicated in hepatocellular carcinoma were retrieved from the GeneCards database (https://www.genecards.org/; accessed on 20 December 2025) [27]. Overlapping targets between UAEcr-derived compounds and T2D-related genes were identified using Venn diagram analysis generated via the Bioinformatics web tool (https://bioinformatics.psb.ugent.be/webtools/Venn/; accessed on 20 December 2025). These shared targets were subsequently selected for protein–protein interaction network construction.

4.4.3. Network Pharmacology Analysis

Protein–protein interactions among the overlapping targets were investigated using the STRING database (version 12.0; https://string-db.org/, accessed on 20 December 2025), which integrates experimentally validated and computationally predicted associations based on physical binding and functional linkage [28]. The analysis was limited to Homo sapiens, and all receptor-related interactions were included. A stringent confidence score threshold of 0.9 was applied to ensure high-reliability interactions [29].

4.4.4. Gene Ontology and Pathway Enrichment Analysis

Biological interpretation of the predicted target genes was carried out using pathway and functional annotation based on the KEGG and Gene Ontology (GO) resources. KEGG pathway reconstruction was performed using Python in combination with the Bioservices package (version 1.11.2), which enables automated querying of the KEGG database (https://www.genome.jp/kegg/pathway.html, accessed on 20 December 2025) [30,31]. Compound names and molecular formulas were first matched to KEGG compound identifiers and subsequently linked to their annotated metabolic and signaling pathways.

Gene Ontology analysis was conducted to categorize UAECr associated targets into the three principal domains of biological process and molecular function using curated annotation datasets [32]. Statistical enrichment was evaluated using a hypergeometric distribution to identify GO terms and KEGG pathways that were significantly over-represented compared with the background gene universe [33]. To control for multiple hypothesis testing, p-values were adjusted using the Benjamini–Hochberg false discovery rate (FDR) method, and terms with FDR-corrected values below 0.05 were considered statistically significant. Enrichment strength was expressed as the ratio between observed and expected gene counts [34].

4.4.5. Molecular Docking and Binding Interaction Analysis

Protein ligand docking simulations were conducted using CB-Dock2, a blind docking platform guided by automated cavity detection (https://cadd.labshare.cn/cb-dock2/index.php; accessed on 21 December 2025) [35]. This system integrates binding-site recognition, docking, and homologous structure fitting to improve both docking precision and computational efficiency. The software identifies potential ligand binding cavities, automatically defines docking boxes, and ranks binding poses based on AutoDock Vina affinity scores [36]. Three dimensional interaction patterns were visualized to examine ligand receptor binding modes. To further refine pocket prediction, the CurPocket curvature-based cavity detection algorithm was applied. Three dimensional structures of UAECr-derived ligands were retrieved from PubChem, whereas target protein structures including DPP4 (PDB 3G0B), PPARγ (PDB 2HFP), GCK (PDB 1V4S), PTPN1 (PDB 2HNP), and AKT2 (PDB 1MRV) were downloaded from the RCSB Protein Data Bank (https://www.rcsb.org; accessed on 21 December 2025). All crystallographic water molecules were automatically removed by the CB-Dock2 server prior to docking to avoid interference with ligand binding.

4.4.6. Molecular Dynamics Simulation

Molecular dynamicssimulations were conducted using the MDsim360 web-based platform to evaluate the dynamic behavior and stability of the protein–ligand complex under near-physiological conditions. The docked protein–ligand complex was used as the starting structure for all simulations. The system was parameterized using the CHARMM36 force field [37], and explicit solvation was performed with the TIP3P water model. Protonation states of titratable residues were assigned at pH 7.0 to reflect physiological conditions. Protein and ligand chains were defined according to their respective chain identifiers. The complex was embedded in a cubic simulation box with a minimum distance of 1.0 nm between the solute and the box boundaries to prevent artificial self-interactions under periodic boundary conditions (PBC). To achieve electroneutrality and physiological ionic strength, KCl ions were added at a concentration of 0.15 M. Following system preparation, energy minimization was performed to relieve unfavorable steric interactions. The minimized system was subsequently subjected to a 100 ns production MD simulation using a classical MD integrator. The system temperature was maintained at 310 K using the V-rescale thermostat, while pressure was controlled at 1 bar using the Parrinello–Rahman barostat. Short-range van der Waals and Coulombic interactions were treated with a cutoff distance of 1.2 nm. All covalent bonds involving hydrogen atoms were constrained using the H-bonds constraint algorithm, allowing for stable numerical integration. The MD trajectories were analyzed to assess the conformational stability and compactness of the protein–ligand complex. Structural and energetic parameters, including RMSF, radius of gyration (Rg), SASA, interaction energy components, and salt bridge dynamics, were evaluated throughout the simulation period.

4.5. In Vitro Study

UAECr was prepared as described previously. Campesterol (≥98% purity; SigmaAldrich) (Sigma-Aldrich, St. Louis, MO, USA) was used as a representative single bioactive compound, while pioglitazone (Sigma-Aldrich, St. Louis, MO, USA) served as the positive control. Stock solutions of UAECr, campesterol, and pioglitazone were prepared in dimethyl sulfoxide (DMSO) (Sigma-Aldrich, St. Louis, MO, USA) and diluted with culture medium prior to treatment. The final concentration of DMSO was maintained at 0.1% (v/v) in all experimental groups, including the vehicle control.

4.5.1. Cell Culture and Adipocyte Differentiation

Mouse 3T3-L1 preadipocytes (ATCC, Manassas, VA, USA) were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) (Sigma-Aldrich, St. Louis, MO, USA) and 1% penicillin–streptomycin (Sigma-Aldrich, St. Louis, MO, USA) at 37 °C in a humidified atmosphere containing 5% CO₂. Cells were seeded at a density of 1 × 10⁴ cells per well in 96-well plates (100 µL/well) and maintained until reaching full confluence. Cells were maintained until confluence [38,39].

Adipocyte differentiation was induced by seeding cells at a density of 2.0 × 10⁴ cells/cm² and allowing them to reach full confluence, followed by treatment with a standard hormonal cocktail (Sigma-Aldrich, St. Louis, MO, USA) containing 0.5 mM 3-isobutyl-1-methylxanthine (IBMX), 0.25 µM dexamethasone (Sigma-Aldrich, St. Louis, MO, USA), and 10 µg/mL insulin (Sigma-Aldrich, St. Louis, MO, USA) for 48–72 h. Subsequently, cells were maintained in DMEM supplemented with insulin alone, with medium replacement every two days. Fully differentiated adipocytes (day 8–10) were used for glucose uptake and gene expression analyses.

4.5.2. Cell Viability Assay (MTT Assay)

Cell viability was assessed using the MTT assay (Sigma-Aldrich, St. Louis, MO, USA) to determine the cytotoxic profile of the test samples [40]. 3T3-L1 cells were seeded into 96-well plates and treated with UAECr, campesterol, or pioglitazone at final concentrations of 0, 6.25, 12.5, 25, 50, 100, 200, and 400 µg/mL for 24 h. Vehicle-treated cells (0.1% DMSO) served as the control.

Following treatment, MTT solution was added to each well and incubated for 4 h at 37 °C, after which the resulting formazan crystals were dissolved in DMSO and the absorbance was measured at 570 nm using a microplate reader. Cell viability was calculated relative to the vehicle control according to the following equation:

$$\mathrm{Cell} \mathrm{Viability} (\%)={\displaystyle \frac{{\mathrm{Abs}}_{\mathrm{treatment}}}{{\mathrm{Mean} \mathrm{Abs}}_{\mathrm{vehicle}}}}\times 100$$

${\mathrm{Abs}}_{\mathrm{treatment}}$ = absorbance value of cells treated with the test sample. Mean ${\mathrm{Mean} \mathrm{Abs}}_{\mathrm{vehicle}}$ = mean absorbance value of the vehicle-treated control group.

4.5.3. Glucose Uptake Assay

Glucose uptake was evaluated in fully differentiated 3T3-L1 adipocytes using 2-NBDG fluorescent assay (Sigma-Aldrich, St. Louis, MO, USA). Prior to treatment, cells were serum-starved for six hours to enhance insulin sensitivity [40]. Adipocytes were treated with UAECr, campesterol, or pioglitazone at 0, 6.25, 12.5, 25, and 50 µg/mL in the presence of insulin. Vehicle-treated cells were used as the baseline control. After treatment, 2-NBDG was added and incubated under dark conditions. Cells were then washed with phosphate-buffered saline (PBS) (Sigma-Aldrich, St. Louis, MO, USA), and fluorescence intensity was measured using a microplate reader. Glucose uptake was expressed as a percentage relative to the vehicle control using the following formula:

$$\mathrm{Glucose} \mathrm{Uptake} (\%)={\displaystyle \frac{{\mathrm{RFU}}_{\mathrm{treatment}}}{{\mathrm{Mean} \mathrm{RFU}}_{\mathrm{vehicle}}}}\times 100$$

${\mathrm{RFU}}_{\mathrm{treatment}}$ = relative fluorescence unit measured in cells treated with the test sample. Mean ${\mathrm{Mean} \mathrm{RFU}}_{\mathrm{vehicle}}$ = mean fluorescence intensity of the vehicle-treated control group. An increase above 100% indicated enhanced glucose uptake activity compared to basal conditions.

4.5.4. Quantitative Real-Time PCR Analysis of PPARγ Expression

To elucidate the molecular mechanism underlying the observed metabolic effects, PPARγ gene expression was analyzed by quantitative real-time PCR (RT-qPCR) (Sigma-Aldrich, St. Louis, MO, USA) [41,42]. Differentiated 3T3-L1 adipocytes were seeded at a density of 2 × 10⁵ cells per well in 6-well plates and treated with UAECr, campesterol, or pioglitazone at concentrations of 0, 6.25, 12.5, 25, and 50 µg/mL for 24 h. Vehicle-treated cells served as the control group. After treatment, total RNA was isolated using a silica membrane-based RNA extraction kit (Sigma-Aldrich, St. Louis, MO, USA) according to the manufacturer’s instructions. RNA concentration and purity were assessed spectrophotometrically by measuring absorbance at 260 and 280 nm. Complementary DNA (cDNA) was synthesized from 1 µg of total RNA using a commercial reverse transcription kit (Sigma-Aldrich, St. Louis, MO, USA) following the manufacturer’s protocol. Quantitative real-time PCR was performed using gene-specific primers for peroxisome PPARγ and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as the internal reference gene. Primer sequences were as follows, PPARγ forward (5′–3′): TGGAGCCTAAGTTTGAGTTTGC; PPARγ reverse (5′–3′): TGACGATCTGCCTGAGGTCT; GAPDH forward (5′–3′): AGGTCGGTGTGAACGGATTTG; GAPDH reverse (5′–3′): TGTAGACCATGTAGTTGAGGTCA. Relative gene expression levels were calculated using the comparative Ct (2^−ΔΔCt) method. Fold-change values (2^−ΔΔCt) were used for both statistical analysis and graphical presentation.

Relative gene expression was calculated using the comparative Ct (2^−ΔΔCt) method as follows:

$$\mathsf{\Delta }Ct=C{t}_{\mathrm{P}\mathrm{P}\mathrm{A}\mathrm{R}\mathsf{\gamma }}-C{t}_{\mathrm{GAPDH}}$$

$$\mathsf{\Delta }\mathsf{\Delta }Ct=\mathsf{\Delta }C{t}_{\mathrm{treatment}}-\mathrm{Mean} \mathsf{\Delta }C{t}_{\mathrm{vehicle}}$$

$$\mathrm{Relative} \mathrm{Expression} (\mathrm{Fold} \mathrm{Change})=2^{-\mathsf{\Delta }\mathsf{\Delta }Ct}$$

$C{t}_{\mathrm{P}\mathrm{P}\mathrm{A}\mathrm{R}\mathsf{\gamma }}$= threshold cycle (Ct) value of the target gene (PPARγ). $C{t}_{\mathrm{GAPDH}}$ = Ct value of the housekeeping gene (GAPDH). $\mathsf{\Delta }Ct$ = normalized Ct value of the target gene relative to GAPDH. $\mathsf{\Delta }\mathsf{\Delta }Ct$ = difference between $\mathsf{\Delta }Ct$ of the treated sample and the mean $\mathsf{\Delta }Ct$ of the vehicle control. Fold change represents relative gene expression normalized to the vehicle control, which was set to 1.0. For statistical analysis, $2^{-\mathsf{\Delta }\mathsf{\Delta }Ct}$ values were used, while fold-change values were used for data visualization.

4.5.5. DPP-4 Inhibition Assay

DPP-4 inhibitory activity of UAECr and campesterol was evaluated using a fluorometric enzymatic assay based on the cleavage of a fluorogenic substrate (Sigma-Aldrich, St. Louis, MO, USA) [43]. Sitagliptin (Sigma-Aldrich, St. Louis, MO, USA) was used as the positive control inhibitor. Stock solutions of UAECr, campesterol, and sitagliptin were prepared in DMSO and diluted in assay buffer (Sigma-Aldrich, St. Louis, MO, USA) to obtain final test concentrations of 0, 6.25, 12.5, 25, and 50 µg/mL, while maintaining the final DMSO concentration at 0.1% (v/v) for all wells, including the vehicle control. Briefly, the reaction mixture was prepared in black 96-well microplates to minimize background fluorescence. Each well contained DPP-4 enzyme solution (Sigma-Aldrich, St. Louis, MO, USA), fluorogenic substrate, assay buffer, and the corresponding test sample or control. The plate layout included (I) blank wells (substrate + buffer without enzyme) to determine background fluorescence, (II) vehicle wells (enzyme + substrate + 0.1% DMSO) representing 100% enzyme activity, and (III) inhibitor wells (enzyme + substrate + test samples or sitagliptin at specified concentrations). After gentle mixing, the plate was incubated under light-protected conditions at 37 °C for 30 min. Fluorescence signals were recorded using a microplate reader at excitation/emission (Ex/Em) = 360/460 nm, and results were expressed as relative fluorescence units (RFU). DPP-4 activity and inhibition were calculated after blank correction as follows:

$$\mathrm{Activity} (\%)={\displaystyle \frac{RF{U}_{\mathrm{sample}}-{\overline{RFU}}_{\mathrm{blank}}}{{\overline{RFU}}_{\mathrm{vehicle}}-{\overline{RFU}}_{\mathrm{blank}}}}\times 100$$

$$\mathrm{Inhibition} (\%)=100-\mathrm{Activity} (\%)$$

$RF{U}_{\mathrm{sample}}$ = fluorescence signal from wells containing enzyme, substrate, and test sample. ${\overline{RFU}}_{\mathrm{vehicle}}$ = fluorescence signal from vehicle control wells representing 100% enzyme activity. ${\overline{RFU}}_{\mathrm{blank}}$ = fluorescence signal from blank wells without enzyme. Activity (%) represents residual DPP-4 enzymatic activity relative to the vehicle control. Inhibition (%) represents the percentage reduction in DPP-4 activity induced by the test sample. All measurements were performed in at least triplicate, and results were expressed as mean ± standard deviation (SD).

4.6. Statistical Analysis

Statistical analysis was conducted using GraphPad Prism Premium 10 (GraphPad Software, Inc., San Diego, CA, USA) on an Apple MacBook computer. All experiments were performed with at least two independent experiment with three biological replicates. Data were expressed as mean ± standard deviation (SD). Statistical differences among groups were analyzed using two-way analysis of variance (ANOVA), followed by Tukey’s post hoc test. A p-value < 0.05 was considered statistically significant.

5. Conclusions

This study provides integrative mechanistic evidence that C. racemosa exhibits antidiabetic activity through a coordinated, multi-target mechanism centered on PPARγ activation. By combining untargeted metabolomics, network pharmacology, molecular docking, molecular dynamics simulation, and cell-based validation, we demonstrate that UAECr contains bioactive metabolites capable of modulating key pathways involved in insulin sensitivity and glucose homeostasis.

Campesterol emerged as a principal contributor, showing strong and dynamically stable interactions with PPARγ, as confirmed by molecular dynamics analyses, alongside the moderate inhibition of DPP-4. In vitro, UAECr and campesterol significantly enhanced glucose uptake and PPARγ expression in adipocytes without inducing cytotoxicity, indicating an insulin-sensitizing profile rather than nonspecific enzyme inhibition. Although campesterol displays limited classical drug-likeness, its biological efficacy and safety profile support its positioning as a functional food-derived bioactive.

Overall, these findings establish a robust molecular rationale for the antidiabetic potential of C. racemosa, highlight the value of integrating molecular dynamics into natural product research, and warrant further in vivo and translational investigations to advance marine algae-based strategies for metabolic health.