Phycocyanobilin as a Functional Food-Derived Nutraceutical Candidate for Modulating the RAGE/NOX4 Axis in Neurodegenerative Disorders

Abstract

Background/Objectives: Neurodegeneration associated with diabetes and metabolic dysfunction involves interconnected processes, including advanced glycation end product (AGE)-related signaling, RAGE/NOX4-dependent oxidative stress, dysregulated endoplasmic reticulum (ER) stress, and mitochondrial apoptosis. Phycocyanobilin (PCB), a tetrapyrrolic chromophore of C-phycocyanin, has been proposed to exert pleiotropic cytoprotective effects; however, its actions within glycation-associated neuronal stress pathways remain incompletely defined. Methods: Differentiated SH-SY5Y neurons were exposed to AGEs (300 μg/mL) for a 24 h period to examine whether PCB modulates neuronal injury along the RAGE–NOX4–oxidative-stress–ER-stress–mitochondrial axis. The selective RAGE antagonist TTP488 (100 μmol/L) was included as a pharmacological reference. Neuronal viability, neurite integrity, intracellular and mitochondrial reactive oxygen species, ER stress signaling, and apoptotic markers were assessed using complementary biochemical, molecular, and functional assays. Results: PCB pretreatment (10–50 μmol/L) significantly improved neuronal viability, preserved neurite structure, and reduced oxidative stress under the AGE challenge. These effects were accompanied by attenuation of AGEs-induced upregulation of RAGE and NOX4 expression, suppression of PERK–eIF2α–ATF4–CHOP signaling, restoration of mitochondrial apoptotic balance, inhibition of caspase activation, and reduced DNA fragmentation. The overall protective profile of PCB was comparable to that observed with TTP488 at the level of downstream pathway modulation. Conclusions: These findings suggest that PCB mitigates glycation-associated neuronal injury through coordinated regulation of oxidative, ER stress, and mitochondrial apoptotic pathways linked to RAGE/NOX4 signaling, supporting further investigation of PCB as a functional food-derived bioactive in metabolic stress-related neurodegeneration.

1. Introduction

Neurodegenerative disorders such as Alzheimer’s disease (AD), Parkinson’s disease, and vascular dementia share common pathological features, including chronic oxidative stress, metabolic dysregulation, neuroinflammation, and progressive loss of neuronal resilience [1]. These disturbances are increasingly recognized as being closely linked to impaired glucose metabolism and accelerated cellular aging, particularly under conditions such as diabetes and insulin resistance [1]. A key biochemical process connecting metabolic stress to neuronal dysfunction is glycation, a non-enzymatic reaction in which reducing sugars covalently modify proteins, lipids, or nucleic acids, ultimately leading to the formation of advanced glycation end products (AGEs) [2]. Glycation is accelerated by chronic hyperglycemia, oxidative stress, aging, and metabolic imbalance, resulting in the accumulation of structurally altered and functionally compromised macromolecules [2]. Excessive glycation burden has been implicated in diabetic complications, vascular pathology, and neurodegenerative diseases, where AGEs accumulation contributes to synaptic dysfunction, oxidative injury, and inflammatory signaling [3]. Clinically and experimentally, glycation is commonly assessed using indicators such as AGE levels, glycated hemoglobin (HbA1c), and activation of the receptor for advanced glycation end products (RAGE) [4]. Among these, AGEs–RAGE engagement represents a central signaling axis linking glycation stress to downstream oxidative and proteostatic disturbances [3]. RAGE functions as a multi-ligand pattern recognition receptor that binds AGEs, amyloid-β (Aβ), and other stress-associated molecules, thereby amplifying intracellular oxidative stress and inflammatory cascades [5]. Under hyperglycemic or AGEs-enriched conditions, persistent RAGE activation disrupts redox homeostasis, promotes neuroinflammation, facilitates Aβ transport across the blood–brain barrier, and accelerates neuronal degeneration [6].

Beyond its established role in redox and inflammatory signaling, RAGE activation has been shown to induce endoplasmic reticulum (ER) stress, a cellular response characterized by the accumulation of misfolded or unfolded proteins within the ER lumen [7]. This stress response is mediated in part through NADPH oxidase (NOX)-derived reactive oxygen species (ROS) production [8]. Among NOX isoforms, NOX4 is highly expressed in neurons and constitutes a major contributor to ROS generation in the central nervous system, playing a pathogenic role in diabetes-related cognitive decline and neurodegenerative disorders [8]. Excessive ROS not only disturbs cellular redox equilibrium but also activates ER stress pathways, particularly the protein kinase RNA-like endoplasmic reticulum kinase (PERK) branch of the unfolded protein response (UPR), a major pro-apoptotic route [7]. PERK-driven eukaryotic initiation factor 2α (eIF2α) phosphorylation is associated with translational repression and preferential transcription factor 4 (ATF4) synthesis, culminating in C/EBP homologous protein (CHOP) upregulation [9]. Elevated CHOP disrupts B cell lymphoma 2 (BCL2) protein family homeostasis, destabilizes mitochondrial membranes, and promotes cytochrome c release and subsequent caspase activation, thereby mechanistically linking sustained ER stress to mitochondrial dysfunction and neuronal apoptosis [10]. Collectively, these events establish a self-amplifying pathological loop in which glycation-driven RAGE activation intensifies NOX4-dependent oxidative stress, exacerbates ER stress, and accelerates mitochondrial injury, ultimately contributing to progressive neuronal loss [11].

Given the central role of the AGEs–RAGE–NOX4–ROS axis in linking metabolic and glycation stress to neurodegeneration, pharmacological modulation of this pathway represents a rational target for therapeutic investigation [12]. Low-molecular-weight RAGE antagonists, including TTP488 (Azeliragon), have progressed into clinical trials [13]; however, large-scale Phase III trials failed to achieve primary efficacy endpoints, highlighting the limitations of single-target RAGE inhibition and the need for multitarget modulators capable of simultaneously attenuating oxidative stress, ER stress, and mitochondrial dysfunction [14].

C-phycocyanin (C-PC), a water-soluble biliprotein abundantly derived from cyanobacteria such as Arthrospira platensis (Spirulina), has attracted increasing attention for its antioxidant, anti-inflammatory, and neuroprotective properties [15]. Previous work from our group identified C-PC as an inhibitor of AGE–RAGE–related oxidative stress, accompanied by downregulation of PERK–CHOP–mediated ER stress pathways, preserves mitochondrial integrity, and attenuates apoptosis in AGEs-challenged SH-SY5Y neurons [16]. However, the large pigment-protein architecture of C-PC raises concerns regarding its bioavailability, stability, and capacity for blood–brain barrier penetration, suggesting that its chromophore may represent the principal bioactive component [17,18]. Phycocyanobilin (PCB), a linear tetrapyrrolic chromophore structurally related to biliverdin and bilirubin, exhibits intrinsic antioxidant, anti-inflammatory, and NOX-inhibitory properties that align closely with the biological effects historically attributed to C-PC [19]. Unlike the holoprotein, PCB is a small, lipophilic molecule with superior pharmacokinetic potential, including improved systemic absorption and potential blood–brain barrier permeability [20]. Mechanistically, PCB has been reported to suppress NOX activity and limit ROS generation, suggesting that it may intervene at a proximal regulatory node within the glycation–RAGE–NOX–ROS cascade, thereby interrupting downstream ER stress and mitochondrial apoptotic signaling. Despite these promising attributes, whether PCB can directly modulate RAGE-associated signaling under glycation stress remains insufficiently defined. To address this gap, the present study employed an AGEs-challenged SH-SY5Y neuronal model that recapitulates key features of glycation-driven neurotoxicity [21]. Using the selective RAGE antagonist TTP488 as a pharmacological comparator, we aimed to determine whether PCB can suppress AGEs-driven RAGE/NOX4 signaling and oxidative stress, reduce ER stress signaling, and preserve mitochondrial function. Through this approach, we sought to establish PCB as a mechanistically precise, bioavailable, and functionally relevant nutraceutical candidate for counteracting glycation-associated neurodegeneration.

2. Materials and Methods

2.1. Establishment and Differentiation of SH-SY5Y Neuronal Cell Model

A human neuroblastoma-derived neuronal cell system, SH-SY5Y (CRL-2266; ATCC, Manassas, VA, USA), was utilized in this study. Cells were cultured at 37 °C in a humidified incubator with 5% CO₂ using a complete growth medium consisting of a DMEM/Ham’s F-12 formulation (Cat# SLM-243-B, Sigma-Aldrich, St. Louis, MO, USA) supplemented with fetal bovine serum (10%; Cat# F2379, Sigma-Aldrich), non-essential amino acids (1%; Cat# 11140050, Thermo Fisher Scientific, Waltham, MA, USA), and penicillin–streptomycin (100 U/mL and 100 μg/mL; Cat# 15140148, Thermo Fisher Scientific). Routine culture and expansion were carried out under these standard conditions.

Neuronal differentiation was initiated by plating SH-SY5Y cells onto 100 mm dishes at a density of 1 × 10⁶ cells/cm², followed by attachment in complete culture medium. When cultures reached approximately 40–50% confluence, all-trans retinoic acid (RA; 10 μmol/L; Cat# R2625, Sigma-Aldrich, St. Louis, MO, USA) was added to promote neuronal differentiation for five days, consistent with previously described methodologies [22]. Neurite outgrowth was assessed by direct phase-contrast microscopic observation (Zeiss Axiovert 135; Zeiss, Oberkochen, Germany). Primary neurite projections extending from the cell soma were manually counted and followed by digital image analysis conducted with ImageJ, version 1.6.0 (NIH, Bethesda, MD, USA). For each experimental condition, neurite analysis was conducted on randomly selected microscopic fields, with at least 100 cells evaluated per condition across five independent experiments. After completion of the differentiation period, cells were harvested and replated at a density of 1 × 10⁵ cells per well in six-well plates. Cultures were allowed to stabilize until reaching experimental confluence and were then detached using 0.05% trypsin prepared in phosphate-buffered saline (PBS, pH 7.4). All experimental treatments were initiated only after five days of RA-induced neuronal differentiation.

2.2. Establishment of AGE-Induced Metabolic Stress and Pharmacological Agents Exposure Conditions

Differentiated SH-SY5Y cells were seeded onto 6-well plates at a density of 2 × 10⁶ cells per well to obtain adequate material for downstream molecular assays. After reaching optimal confluence, cells were harvested with 0.05% trypsin prepared in PBS (pH 7.4; Cat# P4474, Sigma-Aldrich, St. Louis, MO, USA). Stock preparations of bovine serum albumin (BSA; Cat# A8806, Sigma-Aldrich, St. Louis, MO, USA) and BSA-derived advanced glycation end products (AGEs; Cat# 121800-M, Sigma-Aldrich, St. Louis, MO, USA) were made at 1 mg/mL in PBS, passed through 0.22 μm sterile filters, aliquoted, and kept at −20 °C until required. Phycocyanobilin (PCB; Cat# P2172, Sigma-Aldrich, St. Louis, MO, USA) was dissolved in PBS to obtain a 1 mmol/L stock solution, sterilized via 0.2 μm syringe filtration (Minisart^®, Sartorius, Germany), and stored protected from light at 4 °C. TTP488 (Azeliragon; 3-[4-[2-butyl-1-[4-(4-chlorophenoxy)phenyl]imidazol-4-yl]phenoxy]-N,N-diethylpropan-1-amine; Cat# HY-50682, MedChemExpress, Monmouth Junction, NJ, USA) was dissolved in dimethyl sulfoxide (DMSO; Cat# 2650, Sigma-Aldrich, St. Louis, MO, USA) to prepare a 10 mmol/L stock solution. All reagents were diluted to their working concentrations immediately before experiments, ensuring that the final DMSO percentage in culture medium did not exceed 0.1% (v/v), a concentration shown to be non-toxic to neuronal cells [23]. Control wells received medium containing an equivalent amount of DMSO.

For cytotoxicity screening, cells were exposed to increasing concentrations of BSA or AGEs (100–400 μg/mL) for 24 h, either alone or combined with PCB (10, 30, or 50 μmol/L), a concentration range previously reported to exert neuroprotection in ischemic injury models [24]. From these assays, 300 μg/mL AGEs were identified as the concentration that consistently reduced viability to approximately 50% of untreated control levels and were therefore selected for subsequent mechanistic experiments. In pharmacological studies, differentiated SH-SY5Y cells were pre-incubated with PCB (10–50 μmol/L) or TTP488 (100 μmol/L) for 1 h, followed by incubation with AGEs (300 μg/mL) for 24 h at 37 °C. The selected concentration of TTP488 was based on earlier findings demonstrating its ability to downregulate NLR family pyrin domain-containing 3 activation, attenuate pro-apoptotic cascades, and suppress ROS formation in AD-related cellular systems [25]. In the present study, TTP488 was used as a functional comparator to benchmark pathway-level modulation of AGEs–RAGE–NOX4 signaling, rather than to establish definitive receptor-level specificity.

2.3. Cell Viability Analysis

Cellular metabolic activity, used as an index of cell viability, was assessed using the Cell Counting Kit-8 assay (CCK-8; Cat# 96992, Sigma-Aldrich, St. Louis, MO, USA) [26]. SH-SY5Y cells were distributed into 96-well plates at a density of 5 × 10³ cells per well and cultured for 24 h to allow stabilization prior to experimental manipulation. Cells were then exposed for 24 h to BSA or AGEs (100–400 μg/mL), either alone or in combination with PCB (10–50 μmol/L) or TTP488 (100 μmol/L). Following treatment, 10 μL of CCK-8 reagent was added to each well, and plates were incubated at 37 °C for 2 h. Absorbance corresponding to the enzymatic conversion of the tetrazolium-based substrate to a water-soluble formazan product was measured at 450 nm using a SpectraMax M5 microplate reader (Molecular Devices, Sunnyvale, CA, USA). Viability values were normalized to vehicle-treated controls, which were defined as 100%, and all results were expressed as relative percentages.

2.4. Assessment of Intracellular and Mitochondrial ROS Production

A fluorescence-based assay employing the redox-sensitive probe 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA; Cat# 35845, Sigma-Aldrich, St. Louis, MO, USA) was used to assess intracellular oxidative status [27]. Although DCFH-DA is a sensitive but non-specific probe and may be influenced by probe loading, esterase activity, and peroxidase-mediated oxidation, it is widely used as a general indicator of intracellular ROS and is commonly used to assess relative changes in cellular oxidative status under controlled conditions [27]. Cells were incubated with DCFH-DA (5 μg/mL) for 30 min at 37 °C following experimental treatments. Fluorescence was recorded at 488/525 nm with a SpectraMax M5 reader (Molecular Devices, Sunnyvale, CA, USA), with interval measurements used to confirm signal stabilization and endpoint values used for normalized quantification.

Mitochondrial superoxide production was evaluated using the mitochondria-targeted probe MitoSOX™ Red (Cat# M36008, Invitrogen, Carlsbad, CA, USA) [28]. Following treatment, cells were incubated with 5 μmol/L MitoSOX™ Red for 10 min at 37 °C in the dark and washed twice with phosphate-buffered saline (PBS, pH 7.4). Fluorescence was visualized using a Leica fluorescence microscope (Leica Microsystems, Wetzlar, Germany) and quantified using the microplate reader. Detection was performed at excitation/emission wavelengths of 396/610 nm, reflecting the instrument-specific filter configuration rather than the canonical spectral maxima of the probe.

Following completion of fluorescence analysis, cell samples were collected and disrupted, and overall protein content was quantified using a bicinchoninic acid-based colorimetric assay (BCA kit; Cat# ab102536, Abcam, Cambridge, UK). Fluorescence signals were adjusted based on total protein levels to account for variations in cell number and dye handling, and results were reported relative to the vehicle-treated control group. Representative images were obtained from five randomly selected fields per condition across five independent experiments.

2.5. Analysis of ER Stress Markers and Apoptosis-Associated Proteins

To assess the effects of the experimental treatments on ER stress and apoptotic signaling, cellular lysates were analyzed using a panel of commercially available ELISA kits targeting RAGE/NOX4 pathways, unfolded protein response markers, and apoptosis regulators. RAGE levels were measured using an ELISA kit from R&D Systems (Cat# DRG00, Minneapolis, MN, USA), and NOX4 levels were assessed using an assay from Antibodies.com (Cat# A78534, St. Louis, MO, USA). Total PERK and phosphorylated PERK were quantified using ELISA kits (Cat# CBP2027 and Cat# CB5548, respectively), while total eIF2α and phosphorylated eIF2α were measured using ELISA kits (Cat# CB5226 and Cat# CBP1538, respectively), all obtained from Assay Biotechnology (San Jose, CA, USA). ATF4 and CHOP concentrations were determined using kits from Signosis (Cat# TE-0039, Santa Clara, CA, USA) and Fine Biotech (Cat# EM1933, Wuhan, China), respectively. Apoptosis-associated proteins Bcl-2 (Cat# CBCAB00158) and Bax (Cat# CBCAB00157) were quantified using ELISA kits obtained from Assay Genie (Dublin, Ireland). For caspase activation analysis, caspase-9 (Cat# APT139) and caspase-3 (Cat# APT131) activities were measured using substrate cleavage-based assays from Sigma-Aldrich (St. Louis, MO, USA), employing Ac-LEHD-pNA and Ac-DEVD-pNA as the respective chromogenic substrate, respectively. Absorbance was recorded at 450 nm for ELISA measurements and at 405 nm for caspase assays using a SpectraMax M5 microplate reader (Molecular Devices, Sunnyvale, CA, USA). Protein expression and enzymatic activities were normalized to the total protein content, which was determined using the BCA protein assay kits (Cat# ab102536; Abcam, Cambridge, UK). All values were expressed relative to the vehicle-treated control group.

2.6. Quantitative Real-Time PCR Analysis

RNA extraction from SH-SY5Y cells was performed using RNAiso Plus (Cat# 9108; Takara Bio Inc., Shiga, Japan). RNA quantity and purity were verified by spectrophotometric measurement at 260/280 nm with a NanoDrop™ 2000 instrument (Thermo Fisher Scientific, MA, USA), and samples meeting a quality threshold of A260/A280 = 1.8–2.0 were selected for downstream procedures. RNA quality was further evaluated by agarose gel electrophoresis to verify integrity. For complementary DNA synthesis, 1 μg of purified RNA from each sample was converted to cDNA using the PrimeScript^® RT Reagent Kit (Cat# RR037A; Takara Bio Inc., Shiga, Japan). Reverse transcription was carried out in a total reaction volume of 20 μL containing a mixture of oligo(dT) and random hexamer primers. The synthesized cDNA was subsequently diluted 1:10 with nuclease-free water and stored at −20 °C until further use.

Quantitative real-time PCR analysis was performed using a SYBR Green-based detection chemistry (SYBR^® Premix Ex Taq™ II; Cat# RR820A, Takara Bio Inc., Shiga, Japan) on a CFX96 real-time PCR platform (Bio-Rad Laboratories, Hercules, CA, USA). Reactions were carried out in a final volume of 20 μL containing SYBR master mix, gene-specific forward and reverse primers (0.4 μmol/L each), diluted cDNA template (2 μL), and nuclease-free water. Thermal cycling conditions included an initial enzyme activation step at 95 °C for 30 s, followed by 40 amplification cycles consisting of denaturation at 95 °C for 5 s and annealing/extension at 60 °C for 30 s. Product specificity was verified by post-amplification melt curve analysis over a temperature range of 65–95 °C. Primer sequences for all analyzed genes are provided in

Table 1. The primer sequences used for real-time PCR assay.

Genes	Forward Primers (5′-3′)	Reverse Primers (5′-3′)
RAGE	CACCTTCTCCTGTAGCTTCAGC	AGGAGCTACTGCTCCACCTTCT
NOX4	GCCAGAGTATCACTACTCCAC	CTCGGAGGTAAGCCAAGAGTGT
PERK	GTCCCAAGGCTTGGAATCTGTC	CCTACCAAGACAGGAGTTCTGG
eIF2α	CTGGACCTCATGCAGCTTTAGC	CTCCATAGTAGGAGCTCCTGTC
ATF4	TTCTCCAGCGACAAGGCTAAG	CTCCAACATCCAATCTGTCCCG
CHOP	GGTATGAGGACCTGCAGAGGT	CTTGTGACCTCTGCTGGTTCTG
Bcl-2	ATCGCCCTGTGATGACTGAGT	GCCAGGGAAATCAAACAGAGGC
Bax	TCAGGATGCGTCCACCAGAAG	TGTGTCCACGGCGGCAATCATC
Caspase-9	GTTTGAGGACCTTCGACCAGCT	CAACGTACCAGGAGCCACTCTT
Caspase-3	GGAGCGAATCAATGGACTCTGG	GCATCGACATCTGTACCAGACC
GAPDH	GTCTCCTCTGACTTCAACAGCG	ACCACCCTGTTGCTGTAGCCAA

, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as the internal normalization control. All reactions were conducted in technical triplicate, with no-template controls included to exclude reagent contamination. Cycle threshold values were determined using CFX Manager software (version 3.0; Bio-Rad), and relative mRNA expression levels were calculated using the comparative Ct approach [29]. Gene expression data are presented as fold changes relative to the vehicle-treated control group.

2.7. Assessment of Cytochrome c Translocation from Mitochondria to Cytosol

The redistribution of cytochrome c was analyzed as an indicator of mitochondrial outer membrane permeabilization [30]. Following experimental treatments, cells were gently lysed in ice-cold mitochondrial isolation buffer, and subcellular fractions were separated through sequential centrifugation. The homogenates were first cleared of nuclei and cellular debris (800× g, 20 min), after which mitochondrial pellets were obtained (10,000× g, 15 min). The resulting supernatant was further centrifuged (16,000× g, 25 min) to yield the cytosolic fraction. Cytochrome c levels in mitochondrial and cytosolic extracts were quantified using a commercial ELISA kit (Cat# ab210575; Abcam, Cambridge, UK) based on antibody capture and horseradish peroxidase-mediated color development, with absorbance recorded at 450 nm using a SpectraMax M5 microplate reader (Molecular Devices, USA). Protein content in each fraction was determined by the BCA protein assay kits (Cat# ab102536; Abcam, Cambridge, UK), and cytochrome c abundance was normalized to total protein and expressed relative to vehicle-treated controls.

2.8. Quantification of Apoptosis-Associated DNA Fragmentation

The extent of apoptosis-related DNA breakdown was determined using a nucleosome-based ELISA system (Cell Death Detection kit, Cat# 11544675001; Roche Molecular Biochemicals, Mannheim, Germany) [31]. After treatment, the cytosolic portion of each sample was isolated and introduced into wells pre-coated with an anti-histone antibody to capture nucleosomal particles released during apoptotic chromatin condensation. Subsequently, a peroxidase-linked anti-DNA antibody was applied to detect the bound nucleosomes, generating immune complexes in proportion to the level of DNA fragmentation. Chromogenic development was initiated by adding the ABTS substrate and allowing the reaction to proceed for 10 min at 20 °C under light-protected conditions. Absorbance at 405 nm was recorded using a SpectraMax M5 microplate reader (Molecular Devices, Sunnyvale, CA, USA). The resulting optical density values were adjusted based on protein concentrations obtained using the BCA assay (Cat# ab102536; Abcam, Cambridge, UK).

2.9. Statistical Analysis

Data are presented as mean ± standard deviation (SD). Each dataset was derived from five independent biological experiments (n = 5), with three technical measurements obtained for each treatment condition. For each biological experiment, technical triplicates were averaged to generate a single value per condition prior to statistical analysis, and biological replicates were used as the unit of analysis for all inferential statistics. Statistical evaluation of differences among groups was conducted by one-way analysis of variance (ANOVA). For experimental designs involving multiple factors, one-way ANOVA was performed separately within predefined strata. Upon detection of a significant overall effect, pairwise comparisons between groups were conducted using Tukey’s multiple-comparison procedure. Data distribution and homogeneity of variance were assessed prior to ANOVA, and no major violations of test assumptions were observed. Statistical analyses were performed using SigmaPlot 14 (Systat Software, Inc., San Jose, CA, USA), and statistical significance was defined as a probability value (p) below 0.05.

3. Results

3.1. Establishing an AGEs-Induced SH-SY5Y Cell Injury Model

Exposure of SH-SY5Y cells to bovine serum albumin at concentrations between 100 and 400 μg/mL for either 24 or 48 h had no measurable effect on cell viability (Figure 1, left panel). After 24 h of treatment, AGEs at 100 μg/mL already reduced survival to approximately 80% of control levels, while 200 μg/mL further decreased viability to about 70% (Figure 1, right panel). A more pronounced reduction was observed after 24 h of incubation with higher concentrations of AGEs, with cell viability decreasing to approximately 50% at 300 μg/mL and approaching 40% at 400 μg/mL (Figure 1, right panel) Prolonging the incubation period to 48 h markedly intensified AGEs-induced cytotoxicity, decreasing cell viability to under 60% at 100 μg/mL AGEs and to nearly 35% at 400 μg/mL (Figure 1, right panel). Based on these viability data, the condition of 300 μg/mL AGEs for 24 h, which reduced SH-SY5Y survival to approximately 50% of vehicle-treated controls, was selected as the optimal challenge for reliable evaluation of protective interventions in downstream experiments. (Figure 1, right panel).

3.2. PCB Protects SH-SY5Y Neurons Against AGEs-Induced Cytotoxicity and Neurite Degeneration

Microscopic observation revealed that a 24 h exposure to 300 μg/mL AGEs markedly impaired neuronal morphology, characterized by shrunken cell bodies and a pronounced loss of neurite extensions compared with the control group (Figure 2A). In contrast, pretreatment with PCB or TTP488 substantially preserved neuronal architecture, maintaining both cell density and neurite networks in a concentration-dependent manner (Figure 2A).

Quantitative analysis of cell viability reveals that 300 μg/mL AGEs exposure for 24 h reduced SH-SY5Y viability to nearly 50% of control values, confirming its cytotoxic effects (Figure 2B, upper panel). PCB supplementation at 10 μmol/L or 30 μmol/L significantly ameliorated this decline, restoring viability to ~65% and ~75% of controls, respectively, while 50 μmol/L PCB produced a near-complete recovery comparable to the control (Figure 2B, upper panel). TTP488 (100 μmol/L) pretreatment also conferred marked protection, elevating viability to levels indistinguishable from the control group (Figure 2B, upper panel).

Assessment of neurite outgrowth provided additional support for the protective actions of PCB and TTP488 (Figure 2B, lower panel). Exposure to AGEs (300 μg/mL) markedly reduced the neurite number per cell to ~50% of control levels, indicating substantial impairment of neuronal differentiation and connectivity. Pretreatment with PCB restored neurite formation in a concentration-dependent manner, with 30 and 50 μmol/L significantly reversing the AGEs-induced deficit and producing neurite densities comparable to those of controls. TTP488 (100 μmol/L) exerted a similarly robust protective effect, fully preserving neurite outgrowth and confirming its potent antagonism of AGEs-mediated neurotoxicity. Neither PCB nor TTP488 altered baseline cell viability or neurite morphology in the absence of AGEs challenge (Figure 2B).

3.3. PCB Protects SH-SY5Y Cells from AGEs-Induced Intracellular and Mitochondrial Oxidative Stress

Intracellular ROS generation was assessed using DCF-DA fluorescence staining (Figure 3A). Control cells exhibited minimal fluorescence, consistent with low basal ROS production. In contrast, stimulation with AGEs (300 μg/mL) induced a marked elevation in fluorescence intensity, reaching nearly threefold above control levels. PCB pretreatment attenuated this increase in a concentration-dependent manner, reducing intracellular ROS to approximately 2.4-, 1.9-, and 1.5-fold of control at 10, 30, and 50 μmol/L, respectively. The reduction achieved with 50 μmol/L PCB was comparable to that observed with 100 μmol/L TTP488 (~1.7-fold of control).

Mitochondrial ROS accumulation was quantified using MitoSOX Red staining (Figure 3B). Control cells displayed faint red fluorescence, whereas AGEs (300 μg/mL) exposure led to a substantial increase, elevating mitochondrial ROS to ~3.3-fold of control. PCB pretreatment effectively suppressed this rise, lowering MitoSOX fluorescence to ~2.5-fold at 10 μmol/L, ~2.1-fold at 30 μmol/L, and ~1.5-fold at 50 μmol/L. The suppression observed at the highest PCB concentration closely matched the inhibitory effect of 100 μmol/L TTP488 (~1.6-fold of control).

3.4. PCB Suppresses AGEs-Induced Activation of the RAGE–NOX4 Signaling Pathway in SH-SY5Y Cells

PCB mitigates AGEs-triggered activation of the RAGE–NOX4 axis; we examined protein abundance and mRNA expression in SH-SY5Y cells exposed to AGEs, with or without PCB pretreatment (Figure 4). AGEs stimulation markedly increased the protein levels of RAGE and NOX4 compared with control cells (p < 0.01). Pretreatment with PCB at 10, 30, or 50 μmol/L significantly and concentration-dependently suppressed the AGEs-induced elevations of RAGE and NOX4 proteins (all p < 0.05 vs. AGEs-treated vehicle group), with the greatest reduction observed at 50 μmol/L.

A similar trend was observed at the transcriptional level. AGEs exposure robustly enhanced RAGE and NOX4 mRNA expression (both p < 0.01), increasing each to ~3.2-fold and ~3.0-fold of control. PCB pretreatment effectively reversed these transcriptional elevations in a concentration-dependent manner, lowering RAGE and NOX4 mRNA expression to ~2.5-fold at 10 μmol/L, ~2.0-fold at 30 μmol/L, and ~1.4-fold at 50 μmol/L. TTP488 (100 μmol/L) showed a comparable inhibitory effect, reducing both transcripts to ~1.3-fold of control (all p < 0.05 vs. AGEs-treated vehicle group).

3.5. PCB Inhibits AGEs-Associated PERK–eIF2α–ATF4–CHOP Signaling in SH-SY5Y Cells

As shown in Figure 5A, exposure to AGEs (300 μg/mL) markedly increased PERK and eIF2α phosphorylation, elevating their phosphorylation ratios to approximately 3.3–4.0-fold relative to controls (both p < 0.01). A concentration-dependent mitigation of these increases was observed following PCB pretreatment. Partial suppression of PERK- and eIF2α-associated phosphorylation was observed with PCB at a concentration of 10 μmol/L (p < 0.05 vs. AGEs-treated vehicle group), whereas 30 μmol/L produced a more pronounced suppression. The greatest inhibition was observed with 50 μmol/L PCB, which restored phosphorylation ratios to near-baseline levels. Notably, the inhibitory effects of PCB were comparable to those of TTP488. AGEs (300 μg/mL) stimulation also significantly elevated ATF4 and CHOP protein expression to ~2.5-fold and ~3-fold of control, respectively (both p < 0.01). A concentration-dependent reduction in ATF4 and CHOP abundance was observed following PCB pretreatment, and at 50 μmol/L, expression levels approximated those seen with TTP488 (100 μmol/L).

Parallel trends were observed at the transcriptional level (Figure 5B). AGEs (300 μg/mL) exposure significantly upregulated PERK, eIF2α, ATF4, and CHOP mRNA expression to approximately three times that of control cells (all p < 0.01). PCB pretreatment produced a graded reduction in transcript abundance, with the most pronounced inhibition at 50 μmol/L PCB (p < 0.01 vs. AGEs-treated vehicle group). TTP488 (100 μmol/L) elicited a similar suppressive effect across all examined genes.

3.6. PCB Inhibits AGEs-Triggered Mitochondrial Dysfunction and Intrinsic Apoptosis in SH-SY5Y Cells

As shown in Figure 6A, stimulation with AGEs (300 μg/mL) induced a marked shift in the Bcl-2 family balance, characterized by decreased Bcl-2 abundance together with increased Bax expression at the mRNA and protein levels. PCB pretreatment progressively counteracted this pro-apoptotic shift in a concentration-dependent manner, restoring Bcl-2 expression and suppressing Bax upregulation. The response observed with 50 μmol/L PCB closely approximated that induced by TTP488 (100 μmol/L).

AGEs exposure (300 μg/mL) markedly increased the activity and transcription of caspase-9 and caspase-3 (Figure 6B). PCB pretreatment significantly attenuated these AGEs-evoked increases in a concentration-dependent fashion, with the most pronounced inhibition at 50 μmol/L PCB. The extent of suppression achieved at this concentration was nearly indistinguishable from that observed with 100 μmol/L TTP488 (p < 0.05 vs. AGE-treated vehicle group).

AGE (300 μg/mL)-induced mitochondrial dysfunction was further evidenced by substantial cytochrome c release from mitochondria into the cytosol (Figure 6C). PCB pretreatment markedly prevented this pathological redistribution, retaining cytochrome c within the mitochondrial fraction while reducing its cytosolic accumulation in a concentration-dependent manner. The protective capacity of 50 μmol/L PCB was comparable to that of 100 μmol/L TTP488 (p < 0.05 vs. AGE-treated vehicle group).

The DNA fragmentation assay (Figure 6D) showed that AGE (300 μg/mL) exposure markedly increased nuclear fragmentation (p < 0.01 vs. control). PCB treatment reduced DNA fragmentation in a graded fashion, and at a concentration of 50 μmol/L provided a level of protection comparable to 100 μmol/L TTP488 (p < 0.01 vs. AGE-treated vehicle group).

4. Discussion

The present study provides compelling mechanistic evidence that PCB, the tetrapyrrolic chromophore of C-PC, is a potent multitarget modulator of the AGEs–RAGE–NOX–ROS–ER stress axis in neurons. Using an AGEs-challenged SH-SY5Y model that recapitulates key features of diabetes- and glycation-associated neurotoxicity [32], we demonstrate that PCB not only maintains neuronal viability and morphology but also attenuates the interconnected cascade of oxidative stress, ER stress activation, mitochondrial destabilization, and intrinsic apoptosis. Notably, the protective profile of PCB closely paralleled that of the selective RAGE antagonist TTP488, yet its ability to influence multiple nodes within the pathway suggests a broader regulatory scope. Collectively, these findings reinforce the emerging view that PCB rather than the C-PC holoprotein serves as the principal bioactive contributor to the neuroprotective effects historically attributed to Spirulina-derived phycobiliproteins. Moreover, this work aligns with the expanding trend in functional foods and nutritional supplements, where naturally derived bioactive compounds are increasingly recognized as modulators of disease-relevant molecular pathways [33].

These results align with the framework positioning RAGE as a central integrator of metabolic, oxidative, and inflammatory stress in neurodegeneration [2]. In our model, AGEs exposure reproduced this paradigm by upregulating RAGE and NOX4, elevating cytosolic and mitochondrial ROS, and activating the PERK–eIF2α–ATF4–CHOP pathway. A central finding is that PCB restores redox balance in both oxidative compartments, which are frequently dysregulated in neurodegeneration, and limits NOX4-derived ROS in a manner consistent with its structural similarity to the endogenous antioxidants biliverdin and bilirubin [34]. By curbing ROS amplification, PCB interrupts the feed-forward cycle that intensifies ER stress and prevents maladaptive activation of the PERK–eIF2α–ATF4–CHOP pathway at an early stage of the AGEs–RAGE axis [8]. This combined reduction of oxidative and ER stress highlights PCB as a multitarget modulator capable of intervening where metabolic and proteostatic disturbances converge, supporting its potential as a bioactive compound for next-generation functional foods and nutraceuticals aimed at neuroprotection.

Beyond ER stress regulation, our findings further highlight the capacity of PCB to preserve mitochondrial integrity, a critical determinant that often distinguishes reversible cellular stress from irreversible apoptotic progression [35]. In the present model, AGEs exposure shifted neurons toward a pro-apoptotic Bcl-2 family profile, characterized by reduced Bcl-2 expression, increased Bax levels, mitochondrial cytochrome c efflux, and subsequent caspase-9/3 activation, consistent with the canonical sequence of mitochondria-mediated neuronal injury [36]. PCB pretreatment effectively counteracted these changes by sustaining Bcl-2 expression, limiting Bax induction, preventing cytochrome c efflux, and attenuating downstream caspase activation. These findings indicate that the protective influence of PCB extends across both early regulatory checkpoints and late execution phases of the intrinsic apoptotic pathway.

At a broader signaling level, PCB exerted inhibitory effects across multiple AGEs-associated stress responses, as reflected by reduced RAGE and NOX4 expression, attenuation of intracellular and mitochondrial ROS accumulation, and suppression of PERK–eIF2α–ATF4–CHOP pathway activation. This coordinated pattern suggests that PCB modulates interconnected cellular stress networks rather than acting through a single defined molecular target. While the overall profile of protection observed with PCB paralleled that of the reference compound TTP488, such similarity should be interpreted cautiously and at the level of functional outcomes rather than receptor-specific mechanisms [37]. Consistent with previous reports, TTP488 attenuated oxidative stress, ER stress signaling, and apoptotic activation in this experimental context [13]. Importantly, PCB does not inhibit the chemical formation of AGEs. Instead, our data support the interpretation that PCB mitigates glycation-driven neuronal injury by modulating downstream cellular responses following AGEs exposure. In this model, AGEs stimulation amplified oxidative stress, activated unfolded protein response pathways, and engaged mitochondrial apoptotic signaling. PCB pretreatment attenuated each of these pathological responses, consistent with disruption of a feed-forward stress amplification loop linking oxidative stress, proteostatic imbalance, and mitochondrial dysfunction. By limiting NOX4-associated ROS accumulation, PCB reduced ER stress signaling, preserved mitochondrial integrity, restrained cytochrome c release, and attenuated caspase-dependent apoptosis. It should be noted that small-molecule modulators applied at relatively high concentrations in in vitro systems may exert off-target effects [12,13]. Accordingly, the functional similarity observed between PCB and TTP488 should not be interpreted as evidence of shared receptor-level activity. Rather, these findings support a model in which PCB acts as a downstream modulator of glycation-related stress signaling, thereby enhancing neuronal resilience under metabolic and oxidative stress conditions.

An additional conceptual advance of this study lies in revisiting the relative contributions of PCB and C-PC to the biological activities historically attributed to the C-PC holoprotein [38]. Although C-PC has long been described as an antioxidant, anti-inflammatory, and neuroprotective agent, its large pigment protein architecture raises uncertainties regarding bioavailability and central nervous system penetration [20]. Our previous work identified C-PC as a modulator of PERK–CHOP signaling and mitochondrial apoptosis under AGEs stress [16], yet the present data show that PCB alone can reproduce these effects in the same neuronal model while offering clearer mechanistic resolution. When considered together with the favorable physicochemical characteristics of PCB, including its small molecular size and acidic stability that may facilitate more efficient absorption following oral intake [39], the findings raise the possibility that C-PC may function in part as a carrier complex in which PCB represents a principal bioactive contributor. Further studies across diverse experimental systems and dosing strategies will be needed to more clearly define the relative therapeutic roles of PCB and C-PC. These considerations also point to the potential of PCB as a naturally derived bioactive compound suitable for future development in functional foods and nutraceutical formulations targeting neuroprotection.

Despite the strength of these findings, several limitations warrant consideration and help define important directions for future research. First, this study relied on a single in vitro neuronal model and a single predominant insult (AGEs). Although PCB showed protective activity in vitro at concentrations of 10–50 μmol/L, these conditions primarily serve to define mechanistic feasibility and should not be taken to reflect attainable plasma or brain levels through dietary supplementation. Pharmacokinetic bridging, including assessment of systemic bioavailability and central nervous system distribution, has not yet been established and will need to be addressed in future in vivo studies to more accurately define the translational and nutraceutical relevance of PCB. Future investigations should also assess whether PCB mitigates neurotoxicity induced by other RAGE ligands, such as Aβ, to clarify its broader applicability across RAGE- and NOX4-mediated pathological pathways. Importantly, the observed reduction in RAGE mRNA and protein expression should be distinguished from direct inhibition of RAGE activation. The present data indicate modulation of RAGE abundance and downstream signaling responses under AGEs challenge, but do not directly demonstrate interference with ligand–receptor binding or receptor activation. Future studies employing ligand–RAGE binding assays or RAGE knockdown or knockout models will be required to determine the directness of PCB–RAGE interactions. Moreover, although PCB pre-treatment was applied to retinoic acid-differentiated SH-SY5Y neurons exhibiting stable baseline viability, we did not perform dedicated long-term neuronal stability or structural integrity assessments, such as detailed neurite morphometry or cytoskeletal marker analyses. Incorporation of direct neuronal stability measurements in future studies will further strengthen the mechanistic interpretation of PCBs’ neuroprotective effects. Collectively, addressing these issues will be essential to determine whether PCB can be developed as a safe and effective, naturally derived neuroprotective compound.

Although the present study is mechanistic and cell-based, its nutraceutical framing is broadly consistent with emerging clinical observations suggesting that multi-ingredient food supplementation may be associated with concurrent changes in inflammatory markers and neuroendocrine mediators such as orexin-A, which are implicated in stress and neuroimmune regulation [40]. These observations provide contextual support for the translational relevance of our findings, without implying direct mechanistic equivalence, and indicate that multi-target modulation observed at the cellular level could potentially be reflected in systemic and neuroendocrine changes in human settings.

In conclusion, this study provides mechanistic support for the neuroprotective actions of PCB in the context of chronic metabolic and glycation-associated stress. Using an in vitro neuronal model, we demonstrate that PCB modulates multiple interconnected stress pathways, including attenuation of RAGE–NOX4–associated oxidative signaling, restoration of redox homeostasis, suppression of PERK–eIF2α–ATF4–CHOP pathway activation, and preservation of mitochondrial integrity. Through these coordinated effects, PCB interferes with feed-forward stress cascades that contribute to neuronal vulnerability under conditions of metabolic imbalance. Importantly, these findings should be interpreted as cell-based proof-of-concept. PCB should therefore be regarded as a promising naturally derived candidate that warrants further investigation in animal models to validate its neuroprotective potential and translational relevance in glycation-associated neurodegenerative and metabolic stress contexts.