The sample preparation procedure was adopted and modified from the method described by (Solihah et al., 2019). In this study, preparations began by removing any remaining meat from shark cartilage. The clean cartilage was then cut into small pieces to accelerate drying, washed, drained for approximately 30 min, and weighed. The pieces were dried in a drying machine at 50±2°C for 2 days. Drying was conducted in two phases to prevent color change. A total of 286 g of dry cartilage was obtained and milled into a flour. This resulted in 265 g of shark cartilage meal, which was stored in a plastic bag in a freezer at -20°C until extraction.

The extraction procedure was adapted from the method described by (Agustin et al., 2016). Shark cartilage meal was first washed with n-hexane at a 1:10 (w/v) ratio. A total of 15 g of cartilage meal was mixed with 150 mL of hexane and stirred on a magnetic stirrer for one hour. The meal was separated from hexane by decantation and dried overnight to evaporate any residual hexane. Subsequently, the samples were placed in an oven for 15 min, cooled, and reweighed. Hexane defatting was performed as an initial step to remove non-polar lipids and impurities that could interfere with extraction. Subsequent water extraction was specifically selected to efectively isolate and maximize the yield of the targeted polar bioactive compounds, such as glycosaminoglycans and water-soluble proteins. The washed shark cartilage meal was then extracted with 150 mL of distilled water using a hot plate magnetic stirrer for 8 h at 45±2°C. The mixture was centrifuged at 4,000 rpm for 10 min. The resulting filtrate was collected in an Erlenmeyer flask and stored at until chemical screening.

Bioactive compound screening was conducted using Liquid Chromatography-High Resolution Mass Spectrometry (LC-HRMS). The analysis was performed using LC-HRMS (Thermo Fisher Scientific, USA).

Samples were injected with a volume of 100 µL, and mass detection was carried out using a Q ExactiveTM Orbitrap Mass Spectrometer in fullscan mode at 70,000 resolution. The solution was vortexed at 2,000 rpm for approximately 2 min and centrifuged at 6,000 rpm for 2 min. The sample preparation procedures before LC-HRMS injection difered depending on the sample type. For the solid shark cartilage meal, 50 mg of the sample was diluted with distilled water to a final volume of 1,500 µL. For the liquid meal extract, 1,500 µL of the previously stored filtrate was used directly. Both prepared solutions were vortexed at 2,000 rpm for approximately 2 min and centrifuged at 6,000 rpm for 2 min. The supernatant was collected, passed through a 0.22 µm syringe filter, and transferred to a vial. The sample was placed in an autosampler and injected into the LC-HRMS system with an operation time of 30 min.

The pharmacokinetics of the bioactive compounds were evaluated using the SwissADME platform (Website link.swissadme.ch) based on key drug-likeness parameters, including Lipinski’s rule of 5, Veber, and Egan criteria. Following the methodology of (Rai et al., 2023), compounds with higher drug-likeness scores were considered to have greater potential as drug candidates. Subsequently, compounds that met the drug-likeness criteria were assessed for toxicity using the ProTox online tool (Linked resourcetox.charite.de/protox3/) to identify those that were safe and non-toxic for human use. As stated by (Banerjee et al., 2018), compounds in toxicity classes 4 and 5 are considered nontoxic when ingested.

The PerMM web server (Linked resourcepermm.phar.umich.edu/) was used to evaluate the ability of each compound to penetrate the lipid bilayer of the cell membranes. Membrane permeability is a crucial characteristic that regulates the transport of solutes and solvents across cellular boundaries (Frallicciardi et al., 2022).

The pharmacokinetic properties of the compounds, including absorption, distribution, metabolism, and excretion (ADME), were predicted using the pkCSM application (Website link.au/pkcsm/prediction). Predictions were performed using the SMILES representations of compounds obtained from the PubChem database.

This study focused on NF-kB p65 (PDB ID: 2RAM) and TNF-α convertase (PDB ID: 3EWJ) as potential anti-inflammatory targets. Molecular docking procedures were performed using AutoDock Tools in PyRx Software to investigate the interactions between the identified ligands and target proteins.

Molecular docking was performed within the specific active sites of each protein, such as Lys221, Gln241, Asp243, Val244, Arg246, and Gln247 for NF-kB p65 protein with grid setting Center X:19.1730 Y: 19.9956 Z:61.2699 and Dimensions (A) X: 16.8862 Y: 17.6627 Z: 19.0387, while the active site of TNF-α were Glu406, Leu348, Asn389, Ala439, Gly346, Tyr390, Ala439, Val402, Ile438, His405 with grid setting Center X: 4.3066 Y:6.8428 Z: 27.990 and Dimension (A) X: 18.5942, Y: 24.5849 Z: 18.4775. The selection of these specific amino acid residues for the active sites was based on previously validated structural studies and active pocket characterizations of NF-kB p65 (Giridharan & Srinivasan, 2018) and TNF-α convertase (Zheng et al., 2010). The binding afinity was recorded in kcal/mol, and the binding poses of the ligands were saved and visualized using Discovery Studio software.

A total of 2,030 g of fresh cartilage was processed into 256 g meal. The rendement of the cartilage meal of blue shark was estimated to be 12.7 %. This indicates that the water content of fresh shark cartilage is very high. (Sulityowati et al., 2015) observed that the water content of fresh shark bones was 74%, and that of shark cartilage meal was 5.40%. According to (Putranto et al., 2015), the higher the rendement, the higher the economic value and efectiveness of the product.

Blue shark cartilage was extracted with 150 mL of distilled water, repeated six times, obtaining an average extract of 75.17±6.88 mL, then weighed to obtain an average weight of 74.21±7.00 g. (Wijaya & Satriawan, 2023) stated that the amount of rendement obtained is afected by the type of solvent, while according to (Damaiyanti, 2015), the rendement is inversely proportional to the extract quality.

LC-HRMS enables thorough protein analysis and identification of peptide markers in samples, ofering a strong, reliable, and standardized approach for authentication (Windarsih et al., 2024). The results of LC-HRMS analysis of the blue shark cartilage meal and its extract are presented as a whole chromatogram (graphic chromatogram) in Figure 1.

The chromatograms of shark cartilage meal (A) and shark cartilage extract (B) show that the compounds began to elute at approximately 0.84 min, with well-defined chromatographic peaks. LC-HRMS analysis detected 64 compounds in the shark cartilage meal and 141 compounds in the extract. Of these, only compounds with an mzCloud Best Match value >80% were subjected to in silico analysis. In the shark cartilage meal sample, there were 20 compounds with a match value >80% (Table 1), whereas 37 compounds were identified in the extract (Table 2). According to (Castaño-Ortiz et al., 2023), compounds with a spectral similarity of more than 80% to the database and that have been visually validated are considered positively identified and documented, along with their m/z values and retention times

LC-HRMS analysis revealed a diverse profile of low-molecular-weight metabolites in both the meal and extract. While several identified bioactive compounds are amino acids (such as L-norleucine, L-phenylalanine, and D-(+)-proline), others belong to distinct chemical classes. Hypoxanthine and guanine are naturally occurring purine derivatives, whereas uracil and isocytosine are pyrimidines, which are typical products of nucleic acid and energy metabolism in animal tissues. Furthermore, other nitrogenous metabolites and quaternary ammonium compounds, such as betaine, choline, and creatine were detected.

However, it is important to note the detection of synthetic substances such as MDPV, DEET, and α-Pyrrolidinopropiophenone (α-PVP) in the extract. These compounds are highly unlikely to be endogenous to shark cartilage and are presumed to be environmental contaminants, marine bioaccumulation products, or artifacts introduced during sample handling and extraction.

Drug-likeness evaluation was performed using the Lipinski, Ghose, Veber, and Egan rules to determine which compounds met all the criteria. High druglikeness scores indicate that molecules have a greater potential to serve as efective drug candidates (Rai et al., 2023). Tables 3 and 4 demonstrate the drug-likeness predictions for shark cartilage meal and shark cartilage extract, respectively.

Table 3 shows that two compounds (L-Phenylalanine and trans-3-Indoleacrylic acid) in shark cartilage meal passed four drug-likeness parameters. Table 4 shows that the shark cartilage extract has 8 compounds, namely 3,5-di-tert-butyl-4- hydroxybenzaldehyde, L-Phenylalanine, 13(S)-HOTrE, Olomoucine, Piperine, DEET, α-Pyrrolidinopropiophenone, and Piperanine, which passed 4 drug-likeness parameters (Lipinski, Veber, Ghose, and Egan). According to (Silva et al., 2023), a common approach to determine whether a molecule possesses suitable pharmacological traits (druglikeness) is to evaluate its adherence to the criteria outlined in Veber’s and Lipinski’s Rule of 5. The drug-likeness prediction in this study showed that all compounds in shark cartilage meal and its extract had no violations based on Lipinski, Veber, and Egan parameters (Fig 1A and Fig 2A). According to (Ursu et al., 2011), the concept of drug-likeness was developed to provide a useful direction during the initial stages of drug discovery, aiming to enhance the chances of a compound progressing into and through clinical trials.

Drug-likeness parameters, such as Lipinski’s Rule of Five, serve as preliminary physicochemical filters for oral bioavailability rather than definitive predictors of pharmacological eficacy. Furthermore, a clear distinction must be made between the identified compounds based on their primary biological roles. Many of the compounds that passed these filters, such as betaine, creatine, choline, and various amino acids (e.g., L-phenylalanine, L-norleucine, and D-(+)-proline), are fundamental nutritional metabolites and dietary components. Although biological functions were identified, no traditional drug candidates were observed. However, the more complex secondary metabolites and unique structural derivatives identified in this study, such as trans-3- indoleacrylic acid, piperine, piperanine, and olomoucine, represent actual lead compounds and potential drug candidates for further pharmacological development.

The membrane permeabilities of these compounds were predicted to evaluate their ability to cross the cell membrane and assess their therapeutic efectiveness. Figure IB shows that the 13 compounds in shark cartilage meal possess the potential for membrane penetration, and the energy transfer value of the 13 compounds penetrating the phospholipid membrane is the minimum transfer energy. According to (Frallicciardi et al., 2022), membrane permeability is a key characteristic of cell membranes that regulates the transport of solutes and solvents across cells and within their internal compartments.

As shown in Figure 2, the 23 compounds were evaluated to predict their membrane permeability, determine their ability to cross the cell membrane, and assess their therapeutic efectiveness. The identified compounds have the potential to penetrate the membrane barrier. Among the bioactive compounds, betaine requires relatively high energy to pass through the phospholipid bilayer, suggesting low lipophilicity. This reduced lipophilicity is attributed to the higher molecular weight of the compound compared to the other compounds. Membrane permeability is crucial in drug development (Bennion et al., 2017). A drug intended to act on an intracellular target is largely inefective when it cannot cross the cell membrane.

The Pharmacokinetic and Toxicity Community System (pkCSM) is a webbased tool designed to predict the pharmacokinetic properties and toxicity profiles of chemical compounds (Pires et al., 2015). The pharmacokinetic characteristics of a compound describe its absorption, distribution throughout the body, metabolism, and excretion (ADME). The pharmacokinetic and toxicity properties of shark cartilage meal and shark cartilage extract are shown in Tables 5 and 6, respectively.

According to (Chander et al., 2017), a compound quality is categorized as very good at an absorption rate of >80%, good at >30%- 80%, and poor at <30%. As presented in Table 5, there are nine compounds in shark cartilage meal with very good absorption (absorption >80%): betaine, hypoxanthine, choline, caprolactam, nicotinamide, creatinine, D-(+)- proline, trigonelline, and trans-3-Indoleacrylic acid. Table 6 demonstrates the presence of 17 compounds in the shark cartilage extract with very good absorption (absorption >80%), such as betaine, hypoxanthine, choline, caprolactam, 3,5-di-tert-Butyl-4- hydroxybenzaldehyde, 2,2,6,6-Tetramethyl-1-piperidinol (TEMPO), 13(S)-HOTrE, Creatinine, 7-Hydroxycoumarine, D-(+)- Proline, Trigonelline, Olomoucine, DL-Carnitine, Piperine, Uracil, DEET and Piperanine.

Skin permeability is frequently regarded as a key factor influencing the potential of a chemical to cause sensitization (Alves et al., 2015). This property reflects a molecule’s capacity to penetrate the skin, which is largely determined by both its physicochemical characteristics and the physicochemical and biological properties of the skin. A compound with relatively low skin permeability has a log Kp value of >-2.5. Table 5 reveals that there are 11 compounds with a log Kp value <-2.5 and two compounds with a log Kp value >-2.5, L-norleucine and L-phenylalanine, in shark cartilage meal. Table 6 shows that shark cartilage extract contains 19 compounds with log Kp value <-2.5 and five compounds with log Kp value >-2.5, namely L-norleucine, L-Phenylalanine, 7-Hydroxycoumarine, and valine. Therefore, both shark cartilage meal and its extract have relatively high skin permeabilities. Compounds with favorable skin permeability are potential candidates for the development of new drugs for transdermal delivery (Dwi et al., 2020).

According to Smith et al. (2015), the volume of distribution at steady state (VDss) refers to the theoretical volume that a drug would need to occupy to achieve the observed plasma or blood concentration relative to the administered dose, while an excessively high or low volume of distribution leads to a reevaluation of the compound’s suitability as a drug candidate (Yates & Arundal, 2007). A compound with a low volume of distribution has a log VD value of <-0.15 and >0.45 (Pires et al., 2015). Table 5 shows that eight compounds have a log VD value <-0.15, four compounds have a value of -0.15< log VD <0,45, and one >0.45, namely hypoxanthine in shark cartilage meal. Table 6 shows that the shark cartilage extract has 11 compounds with a log VD value of <-0.15, 10 compounds with 0.15< log VD <0.45, and 2 compounds with log VD >0.45, which are hypoxanthine and olomoucine. Mansoor and Mahabadi (2023) added that drugs with a high volume of distribution tend to exit the bloodstream and accumulate in tissues outside the vascular system, implying that a larger dose is necessary to reach the desired plasma concentration. However, drugs with a low Vd tend to remain in the plasma. A smaller dose is suficient to achieve the same plasma concentration.

The blood-brain barrier (BBB) functions as a selective and partially permeable restraint that helps regulate internal stability in the central nervous system (CNS). As drugs must reach the CNS to exert their intended efects, evaluating their ability to cross the BBB is essential for the early development of CNS-targeted therapies (Kumar et al., 2022). Based on PreADMET, a compound with high absorption to the CNS has a log BB value >0.3, moderate absorption to the CNS >-0.1 but <0.3, and low absorption to the CNS <-1.0 (Kang, 2015). Table 5 shows that there are 10 compounds in shark cartilage meal with a log BB value >-0.1 but <0.3, while Table 6 shows the presence of 19 compounds in shark cartilage meal with a log BB value >-0.1 but <0.3. This indicates that a middle absorption to the CNS was exhibited. However, one compound (DEET) had a high absorption to the CNS with a log BB value >0.3, indicating a high absorption to the CNS. According to (Huang et al., 2024), assessing BBB permeability is important to avoid unwanted CNS side efects, including drowsiness, respiratory suppression, nausea, vomiting, dizziness, confusion, and anxiety in patients. In silico prediction of a compound’s ability to cross the BBB plays a major role in the discovery of new drugs for brain and neurological disorders.

The toxicity profile of a compound plays a significant role in determining its therapeutic potential. Favorable toxicity characteristics can facilitate the eficient and rapid advancement of lead compounds into viable drug candidates (Pires et al., 2015). The toxicity evaluations of the 13 bioactive compounds in shark cartilage meal and the 23 bioactive compounds in shark cartilage extract are presented in Figures 3 and 4, respectively. Figure 3 shows that the toxicity evaluation of the 13 bioactive compounds in shark cartilage meal indicated that they were non-toxic in terms of immunotoxicity, mutagenicity, and cytotoxicity. Moreover, the screening results suggested that nicotinamide, guanine, and trans-3-indoleacrylic acid may cause liver damage. In addition, toxicity screening showed that trans-3-indoleacrylic acid and D-(+)-proline potentially caused nerve damage. Despite this, these compounds are included in toxicity classes 4 and 5, except for hypoxanthine, L-norleucine, and guanine, which are in class 3. According to (Banerjee et al., 2018), compounds that fall under toxicity classes 4 and 5 are considered non-toxic when ingested.

Toxicity evaluation revealed 13 bioactive compounds in shark cartilage meal and 23 in shark cartilage extract (Figures 3 and 4).

Figure 4 shows that the toxicity evaluation of the 23 bioactive compounds indicated that shark cartilage extract was non-toxic and safe in terms of hepatotoxicity, mutagenicity, and cytotoxicity. The screening results suggested that piperine causes immune damage. Toxicity screening showed that 7-Hydroxycoumarine and D(+)-proline potentially caused nerve damage. However, the toxicity class also indicates that the compounds are still in classes 4 and 5, except for hypoxanthine, L-norleucine, and 2,2,6,6-Tetramethyl-1-piperidinol (TEMPO), which are included in class 3. Based on (Banerjee et al., 2018), the compounds were considered non-toxic when ingested by fish.

A total of 13 bioactive compounds from shark cartilage meal and its extract that met the criteria for drug-likeness, membrane permeability, and pharmacokinetics, with nontoxic properties, were subsequently analyzed for molecular docking. Molecular docking analysis showed that the most favorable binding interaction against the NF-kB p65 protein was exhibited by emetine, which was used as a control drug (-4.9 kcal/mol). Among the identified constituents, Trans-3-indoleacrylic acid (-4.6 kcal/mol) and L-phenylalanine (-4.4 kcal/mol) demonstrated the lowest binding energies to NF-kB p65. These compounds interacted with the protein through hydrogen and hydrophobic bonds at residues Arg246, Gln247 and Lys221 (Table 7). However, it is crucial to clarify that L-phenylalanine is a fundamental proteinogenic amino acid. Its moderate docking score merely reflects the structural binding plausibility within the active site and should not be overinterpreted as evidence of pharmacological inhibition.

A similar binding trend was observed for the TNF-α convertase protein. Trans-3-indoleacrylic acid and L-phenylalanine displayed the most favorable binding afinities among the extracted compounds (-7.0 and -6.5 kcal/mol, respectively), although their afinities remained lower than that of the control ligand (-7.3 kcal/mol) (Table 8). The control ligand interacted with TNF-α through Glu406 (electrostatic interaction), Leu348, Asn389, Ala439, Gly346, and Tyr390 (hydrogen bonds), and Ala439, Val402, and His405 (hydrophobic bonds). The same residues were also included in the interaction between phenylalanine and TNF-α, such as Ala439, His405, and Val402. While Ala439, His405, and Val402 are also involved in Trans-3-indoleacrylic acid (Table 8).

The binding afinities of L-phenylalanine (-4.4 kcal/mol) and Trans-3- indoleacrylic acid (-4.6 kcal/mol) against NFkB p65 showed moderate binding potential comparable to that of the control drug (-4.9 kcal/mol). Furthermore, both compounds exhibited moderate binding plausibility with TNF-α protein, with afinities of -6.5 kcal/mol and -7.0 kcal/mol, compared to the control ligand (-7.3 kcal/mol). This strong interaction is facilitated by hydrogen bonds formed between the ligand and the protein (Maarif et al., 2025). Docking scores in the range of -4 to -7 kcal/mol denote moderate binding afinity. Furthermore, molecular docking merely suggests binding plausibility and does not provide definitive evidence of biological inhibition. Subsequent in vitro and in vivo