All chemicals and reagents were analytical grade. S1PC™ (S-1-propenyl-l-cysteine) were obtained from Wakunaga Pharmaceutical Co., Ltd., Japan. SARS-COV-2 Spike recombinant glycoprotein (cat. no. ab49046) was purchased by Abcam. The purity was >90% as determined by SDS-PAGE.

S1PC was analyzed by GC-MS as TBDMS derivatives according to Jiménez-Martín et al (23).

S1PC was dissolved in 0.1 N HCl to a final concentration of 4 mg/ml. In total, 5 µl these solutions were spiked with 10 µl internal standard (3,4-dimethoxybenzoic acid, 0.1 mg/ml) and dried under N₂. Subsequently, 30 µl pure N-tert-butyldimethylsilyl-N-methyltrifluoroacetamide, followed by 30 µl pyridine, was added. The mixture was then heated at 80˚C for 1 h. The sample was neutralized afterwards with sodium bicarbonate and subjected to GC-MS analysis. The same derivatization protocol was used for the AGE powder (4 mg/ml 0.1 M HCl).

For all GC-MS analyses, an Agilent 7890B gas chromatograph coupled to a 5977B quadrupole mass selective detector (Agilent Technologies, Inc.) was employed. For chromatographic separations, an Agilent HP5ms fused-silica capillary column (30 m x 0.25 mm i.d.) (Agilent Technologies, Inc.) was used, coated with 5%-phenyl-95%-dimethylpolysiloxane (film thickness, 0.25 µm) as stationary phase. Splitless injection was performed at 280˚C. The column temperature program was set to 70˚C (1 min), then to 300˚C at a rate of 20˚C/min and held for 10 min. The carrier gas was helium at a constant flow of 1.0 ml/min. The spectra were obtained in the electron impact mode at 70 eV ionization energy, ion source of 280˚C and ion source vacuum of 10-5 Torr. MS analysis was performed simultaneously in total ion current (mass range scan in the range of m/z 50-600 at a rate of 0.42 scans per sec) and selected ion chromatogram mode. GC-SIM-MS analysis was performed by selecting the following ions: m/z 332 for S1PC and m/z 239 for 3,4-dimethoxybenzoic acid (internal standard).

The human bronchial epithelial IB3-1 cell line (Thermo Fischer Scientific, Inc.) (21) was cultured in LHC-8 medium (Gibco; Thermo Fischer Scientific, Inc.), supplemented with 5% FBS (Biowest) without antibiotics at a temperature of 37˚C and 5% CO₂ (21). The BNT162b2 vaccine (COMIRNATY™; lot. No. FP8191) was obtained from the Hospital Pharmacy of the University of Padova. The S1PC powder was freshly dissolved in culture medium, normally up to 50 mM, before each experiment. The solution was kept in the dark and used only once (15). For treatment with the BNT162b2 vaccine, IB3-1 cells were seeded at 200,000 cells/ml concentration. After 24 h at 37˚C, 0.5 µg/ml vaccine (22) was added just before the indicated concentrations of S1PC were added for an additional 48 h at 37˚C of treatment, for determining the effects on S1PC on BNT162b2-induced gene expression. Following incubation, cells were detached from the plate by trypsinization, counted using a Beckman Coulter^® Z2 cell counter (Beckman Coulter, Inc.) viability assay was performed using the Muse Annexin V & Dead Cell reagent (Merck Millipore), and RNA was extracted for RT-qPCR analysis using TRIzol reagent (Thermo Fischer Scientific, Inc.).

For quantification of the relative mRNA content, 500 ng of total cellular RNA was reverse transcribed to cDNA with the Taq-Man Reverse Transcription Kit (cat no. N8080234; Applied Biosystems; Thermo Fisher Scientific, Inc.), as described by Gasparello et al (24). qPCR experiments were performed using an assay consisting of a PCR primer pair and a fluorescently labeled 5' nuclease probe or SYBR Green. Assays IDs: i) Hs.PT.58.40226675 (HEX) for IL-6; ii) Hs.PT.58.38869678.g (Cy5) for IL-8; iii) Hs.PT.58.20610757 (FAM) for granulocyte-colony stimulation factor (G-CSF); iv) Hs.PT.58.38905484 (FAM) for NF-κB p50; and v) Hs.PT.58.22880470 (FAM) for NF-κB p65. The primers and probes for ribosomal protein L13a (RPL13A) and b-actin and were as follows: β-actin forward, 5'-ACGATGGAGGGGAAGACG-3' and reverse, 5'-ACAGAGCCTCGCCTTTG-3'; β-actin probe, 5'-/5Cy5/CCTTGCACATGCCGGAGCC/3IAbRQSp/-3'; RPL13A forward, 5'-GGCAATTTCTACAGAAACAAGTTG-3' and reverse, 5'-GTTTTGTGGGGCAGCATACC-3'; RPL13A probe, 5'-/5HEX/CGCACGGTC/ZEN/CGCCAGAAGAT/3IABkFQ/-3' (Applied Biosystems; Thermo Fisher Scientific, Inc.). The primers used for the amplification of BNT162b2 Spike sequences using a Master Mix with the SYBR Green intercalating Dye were forward 5'-CGAGGTGGCCAAGAATCTGA-3' and reverse, 5'-TAGGCTAAGCGTTTTGAGCTG-3', and β-actin forward 5'-CCTCGCCTTTGCCGATCC-3' and reverse, 5'-GGATCTTCATGAGGTAGTCAGTC-3' (Integrated DNA Technologies), according to Aldén et al (25). qPCR amplification of cDNA was performed at 95˚C for 1 min, then 50 cycles of 95˚C for 15 sec and 60˚C for 1 min, using the CFX96 Touch Real-Time PCR Detection System (Bio-Rad Laboratories, Inc.). Relative expression was calculated using the comparative quantification cycle (Cq) method (2^-ΔΔCq method) (26) and the endogenous controls human β-actin and RPL13A, used as normalizer. RT-qPCR reactions were performed in duplicate for both target and normalizer genes (24).

All the computational methodologies were performed on a 32 Core AMD Ryzen 93,905x, 3.5 GHz Linux Workstation (O.S. Ubuntu 20.04) equipped with GPU (Nvidia Quadro RTX 4000, 8 GB; Nvidia Corporation). The Toll-like receptor 4 (TLR4) dimer structure was derived from available information (27). The structure of S1PC was drawn and minimized using the Avogadro software (ver. 1.2.0; Avogadro Chemistry) (28). A blind docking simulation was performed on the entire TLR4 dimer surface using the AutoDock Vina software (ver. 1.2.3; Center for Computational Structural Biology) (29). The top scoring complex was submitted to all-atom unbiased molecular dynamics (MDs) simulation, as described by Zurlo et al (22), using the GROMACS ver. 2025.1 software, open source (30), patched with Plumed ver. 2.6.5(31) under the Charmm36 force field (32). The complex was included in a rectangular box of 8x10x7 nm length, solvated and neutralized using 0.15 M sodium chloride. The full system was submitted to energy minimization and equilibrated under constant temperature and volume and constant temperature and pressure (NPT) conditions. Long-range electrostatic interactions were modelled using the Particle Mesh Ewald algorithm (33). The LINCS (34), Nosé-Hoover (35) and Parrinello and Rahman (36) algorithms were used in the simulations for restraints and as thermostat and barostat, respectively. MDs were conducted under the NPT conditions for 50 nsec with 2 fsec time steps. Root-mean-squared deviation (RMSD), root-mean-squared fluctuation, number of hydrogen bonds and interaction energy were obtained through the ‘rms’, ‘rmsf’, ‘hbond’ and ‘energy’ routines implemented in GROMACS.

Effects on cellular viability and apoptosis Annexin V and Dead Cell assay were performed using the flow cytometry-based Muse Cell Analyzer (Merck Millipore) instrument, according to the protocols supplied by the manufacturer. Cells were washed with sterile DPBS 1X, trypsinized, and 150,000 cells were suspended in LHC-8 medium and diluted (1:2) with Muse Annexin V & Dead Cell reagent (Annexin V-PE and 7-AAD) (Merck Millipore) and analyzed. After an incubation of 15 min at room temperature in the dark, samples were acquired and data were analysed using the Muse 1.5 Analysis Software with the Annexin V and Dead Cell Software Module (Merck Millipore) (37).

For NF-κB (p105/p50 and p65) protein quantification, 20 µg total protein extract in RIPA Buffer (Thermo Fisher Scientific, Inc.) quantified with BCA kit (Pierce; Thermo Fisher Scientific, Inc.) were denatured for 5 min at 98˚C and loaded onto a SDS polyacrylamide (8%) gel in Tris-glycine buffer (25 mM Tris, 192 mM glycine and 0.1% SDS). The electrotransfer to 0.2-µm nitrocellulose membranes was performed overnight at 360 mA and 4˚C in electrotransfer with CAPS buffer (25 mM Tris, 192 mM glycine, CAPS 10 mM and 10% methanol). Obtained membranes were stained in Ponceau S solution (Sigma-Aldrich; Merck KGaA) to verify proteins transfer and incubated in 25 ml blocking buffer (TBS-T with 5% nonfat dry milk (Cell Signalling Technology, Inc.) for 1 h at room temperature. After three washes in TBS-T 1X (containing Tween-20 at 0.1%), membranes were incubated overnight at 4˚C in primary antibodies (NF-κB p105/p50 Ab; cat. no. GTX133711; 1:5,000 dilution; GeneTex, Inc.). The day after, membranes were washed in TBST 1X and incubated for 1 h at room temperature, with an appropriate HRP-conjugated secondary antibody (anti-rabbit IgG HRP-conjugated; cat. no. 7074P3; 1:2,000 dilution; Cell Signalling Technology, Inc.). β-actin (primary antibody: Cat. no. 4970S; 1:1,000; Cell Signalling Technology, Inc.) was used as a normalization control. After incubation with the ECL Solution (Claity™ ECL Substrate; Bio-Rad Laboratories, Inc.) the gel images were acquired with the ChemiDoc (Bio-Rad Laboratories, Inc.) with the software Image Lab version 6.1.0, used also for densitometric analysis (Bio-Rad Laboratories, Inc.).

Proteins released into culture supernatants were measured using Bio-Plex Human Cytokine 27-plex Assay (cat. no. M500KCAF0Y; Bio-Rad Laboratories, Inc.), as suggested by the manufacturer. The assay allows the multiplexed quantitative measurement of 27 cytokines/chemokines [including FGF basic, Eotaxin, G-CSF, granulocyte macrophage-colony stimulating factor, IFN-γ, IL-1β, IL-1ra, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8, IL-9, IL-10, IL-12 (p70), IL-13, IL-15, IL-17A, C-X-C motif chemokine ligand 10, chemokine ligand (CCL)2, CCL3, CCLe4, PDGFBB, CCL5, TNF-α and VEGF] in a single well (21).

Briefly, an amount of 50 µl cytokine standards or samples (diluted supernatants recovered from IB3-1 cells) was incubated with 50 µl anti-cytokine conjugated beads in a 96-well filter plate for 30 min at room temperature with shaking. The plate was washed by vacuum filtration three times with 100 µl Bio-Plex Wash Buffer, 25 µl diluted detection antibody was added to each well and the plate was incubated for 30 min at room temperature with shaking. After three filter washes, 50 µl streptavidin-phycoerythrin was added and the plate was incubated for 10 min at room temperature with shaking. Finally, the plate was washed by vacuum filtration three times, the beads were suspended in Bio-Plex Assay Buffer and the plate was read by a Bio-Rad 96-well plate reader. Collected data were analyzed by the Bio-Plex Manager Software (version 6.2; Bio-Rad Laboratories, Inc.) (21).

The data are presented as the mean ± standard deviation of at least three independent experiments. Statistical differences between/among groups were analyzed using one-way ANOVA. Prism (v. 9.02) by GraphPad software (Dotmatics) was used (followed by Bonferroni's test). P#x003C;0.05 was considered to indicate a significant difference.

The effects of exposure of IB3-1 cells to the COVID-19 BNT162b2 vaccine (22) were first analyzed. After the treatment, total cellular RNA was extracted for RT-qPCR analysis, whereas culture supernatants were isolated for the analysis of secreted proteins, to characterize BNT162b2-induced alteration of the secretome profile. IB3-1 cells treated with BNT162b2 vaccine were found to produce large amounts of SARS-CoV-2 Spike protein. To obtain this information, western blotting was performed using protein extracts from IB3-1 cells treated with the BNT162b2 vaccine (Fig. 1). A concentration of 0.5 µg/ml of BNT162b2 was found to be sufficient to induce S-protein (Figs. 1, S1 and S2) and cytokine and chemokine production (Fig. S3), with lower inhibitory effects on cell viability compared with 1 and 2 µg/ml of BNT162b2(22). The western blotting data shown in Fig. 1 confirm that S-protein is produced by BNT162b2 treated cells, which was reported elsewhere in other cellular model systems (22,38,39). In addition, IB3-1 cells treated with the BNT162b2 vaccine were found to accumulate large amounts of SARS-CoV-2 Spike mRNA (Fig. S2) as discussed elsewhere (40). This suggests that the Spike protein was expressed by IB3-1 cells treated with the BNT162b2 vaccine. When RT-qPCR and Bio-plex analyses were performed using RNA or secreted materials from BNT162b2-treated IB3-1 cells, it was found that they exhibited the increased expression and production of pro-inflammatory factors, including IL-6, IL-8, G-CSF, GM-CSF and IP-10 (Figs. S3 and S4).

Fig. 2 shows the GC-MS analysis of S1PC as a di-TBDMS derivative. Both the total ion current (TIC) chromatogram and the selected ion chromatogram (SIM) are shown. Fig. 2A and B revealed the presence of two partially co-eluting peaks at 10.55 and 10.57 min (Fig. 2B), exhibiting identical electron impact mass spectra (Fig. 2C) and the corresponding cis- and trans-isomers of S1PC. Both cis- and trans-S1PC have previously been identified in AGE, with the cis-form believed to be generated through isomerization of its trans form (41). Furthermore, the GC-MS analyses presented in Fig. 2 confirmed the purity levels of the S1PC preparation procedure, which was found to be >95%.

Considering the effects of the spike protein on the NF-κB pathway (42-44), RT-qPCR was next performed to measure p50 and p65 mRNA expression in BNT162b2-treated IB3-1 cells (Fig. 3), following on from a previous study on the same system using SARS-CoV-2 spike protein (45). Cells were therefore exposed for 24 h to BNT162b2 and then cultured for an additional 48 h in the presence of increasing concentrations of S1PC.

In total, two important conclusions could be gathered from the results shown in Fig. 3. BNT162b2 stimulated the accumulation of NF-κB p50 (Fig. 3A) and NF-κB p65 (Fig. 3B) mRNA expression, even though NF-κB was already present at high concentrations in the cytoplasm of IB3-1 cells before BNT162b2 treatment (21,45). In addition, after the BNT162b2-treated IB3-1 cells were cultured with S1PC, reversal of the accumulation of NF-κB p50 (Fig. 3A) and NF-κB p65 (Fig. 3B) mRNA expression was observed. The effect of S1PC resembled that observed using the NF-κB inhibitor sulforaphane (46) on the IB3-1 cellular model system (45). In agreement with the results shown in Fig. 3, the effects of 50 mM S1PC on NF-κB were also evident on protein level according to the western blot analysis using an antibody recognizing the p50/p105 NF-κB subunits (Fig. S5).

To verify if the BNT162b2-induced production of proinflammatory mRNA expression could be affected in IB3-1 cells treated with the BNT162b2 vaccine in the presence of S1PC, RT-qPCR analysis was performed (Fig. 4). The BNT162b2 vaccine was used at 0.5 µg/ml to minimize its anti-proliferative effects (22). Analysis of mRNA accumulation was performed 48 h after treatment. IL-1β, IL-6 and IL-8(47), G-CSF (48) and GM-CSF (49) were first considered, before experimentally focusing on IL-6, IL-8 and G-CSF, which are much more expressed in the IB3-1 experimental model system employed as reported by Gasparello et al (21,45).

The results shown in Fig. 4 demonstrate that the expression of IL-6 (Fig. 4A), IL-8 (Fig. 4B) and G-CSF (Fig. 4C) mRNA expression was significantly upregulated in IB3-1 cells after exposure to the BNT162b2 vaccine. The BNT162b2-mediated induction of expression was more efficient compared with that of SARS-CoV-2 Spike protein (Fig. S4 and data not shown). No major changes in the accumulation of mRNA expression of RPL13A were observed. The RPL13A mRNA was studied, due to its role in mitochondrial metabolism and that its expression is highly stable in several cellular systems (50,51).

The effects of different concentrations (range 1-100 µM) of S1PC on BNT162b2-stimulated IB3-1 cells were next studied by RT-qPCR, using β-actin as internal control (Fig. 4). A concentration-dependent inhibition of the accumulation of IL-6 (Fig. 4A), IL-8 (Fig. 4B) and G-CSF (Fig. 4C) mRNAs was detectable, suggesting that 10 µM S1PC is sufficient to inhibit to some extent the expression of IL-6, IL-8 and G-CSF induced in these cells by the BNT162b2 vaccine. However, higher S1PC concentrations were found to be more effective. By contrast, no inhibitory effects were observed in the RPL13A mRNA expression levels (Fig. 4D).

The results shown in Fig. 4 indicate that S1PC can be proposed as an anti-inflammatory component of AGE to be considered in pre-clinical studies (52,53). However, further studies on the biological effects of S1PC are required to assess possible anti-proliferative and/or cytotoxic effects associated with the inhibition of proinflammatory gene expression.

To analyze in depth the effects of S1PC on cell viability, the proportion of live/dead cells was analyzed using the MUSE^® Annexin V & Dead Cell Kit, which was used to discriminate among live cells, apoptotic cells and dead cells (24,37).

The results obtained are shown in Fig. 5. Fig. 5A shows that treatment of IB3-1 cells with the BNT162b2 vaccine is associated with the reduction of cell viability, whereas S1PC did not induce anti-proliferative effects in IB3-1 cells treated with 0.5 µg/ml of the BNT162b2 vaccine. In addition, Fig. 5B-D indicates that treatment of IB3-1 cells with the BNT162b2 vaccine is associated with a decrease of the % live cells (Fig. 5B) and an increase of the % of dead cells (Fig. 5C). No major effects of S1PC on % live cells or % dead cells could be observed (Fig. 5B and C).

To propose the possible mechanism of action of S1PC, a molecular docking analysis was performed. Preliminary analyses demonstrated a lack of binding of S1PC to NF-κB. The binding of low-molecular-weight drugs to this protein has however been previously reported, such as trimethylangelicin and analogues (54), corilagin (55) and sulforaphane (45,47). In these cases, efficient interactions with NF-κB were found. However, no evidence of molecular interaction between S1PC and NF-κB could be found in in the present docking analysis (data not shown). Since TLR are upstream regulators of the NF-κB signaling (56-58), the possible interaction between S1PC and TLR4 was assessed using the docking AutoDock Vina software (Fig. 6) (29). The results obtained indicate that S1PC is able to bind in silico that to the Toll-IL-1 receptor domain of TLR4. The amino acids involved in the S1PC-TLR4 interactions are His685, Val655, Arg722 and Tyr657.

To further sustain the reliability of the molecular interaction reported in Fig. 6, the computed model was submitted to 50 nsec of all-atom unbiased molecular dynamics simulation. The results obtained demonstrated that the complex remained stable, as can be seen from the Cα-RMSD values calculated over the simulation time (Fig. 7A). In particular, the hydrogen bonds reported in Fig. 7B were retained during the entire molecular dynamic's simulation, yielding an estimated interaction energy of -65.2±7.3 Kcal/mol (computed as the sum of short-range Lennard-Jones and short-range Coulomb contributions over the 50 nsec of simulation). In addition, binding with S1PC was found to reduce the intermolecular interaction between the TLR4 domains, as revealed by both the reduction in H-bond numbers (Fig. 7B) and the increased per-residue RMSF values in comparison to the apo complex (Fig. 7C-E).