The focal point of the investigation was PSD-95, or postsynaptic density protein-95, a scaffold protein that anchors NMDA and AMPA glutamate receptors and organizes the signaling cascades responsible for synaptic plasticity. Because PSD-95 physically links receptor localization with calcium/calmodulin-dependent protein kinase II (CaMKII), a central regulator of activity-dependent plasticity, changes in its expression and clustering translate into changes in how excitatory synapses are built and maintained. Disruption of PSD-95 has been implicated in a striking range of conditions, including DLG4-related synaptopathy, a rare genetic disorder caused by pathogenic variants in the gene encoding PSD-95, as well as autism spectrum disorder, schizophrenia, Alzheimer’s disease, and Fragile X syndrome. This makes PSD-95 a high-value target for any intervention claiming to support synaptic health.

The experimental design was deliberately symmetric and composition-controlled. The team used differentiated human SH-SY5Y neuroblastoma cells, a widely employed neuronal-like model, matured over 14 days with retinoic acid and BrainPhys neuronal medium. After differentiation, cultures were treated for 24 hours with DHA alone (50 micromolar), EPA alone (50 micromolar), or defined DHA:EPA molar mixtures of 2:1, 1:1, and 1:2, with the total fatty acid concentration held constant at 50 micromolar across all conditions. This five-point series spanned the full compositional range from pure DHA to pure EPA, allowing the researchers to isolate the effect of the DHA:EPA ratio rather than simply total omega-3 exposure. Vehicle-treated and untreated controls confirmed that the ethanol carrier, kept at or below 0.04 percent, had no detectable influence.

Quantification combined high-content immunofluorescence imaging with Western blot densitometry. Using an Opera Phenix High Content Screening System and a standardized segmentation pipeline in Columbus software, the team measured PSD-95 fluorescence intensity, PSD-95 puncta density within MAP2-positive neurites, and the pixel-wise overlap of PSD-95 with synaptophysin, a presynaptic marker, to approximate Manders’ colocalization coefficients. Protein abundance was assessed by Western blot with stain-free total-protein normalization, a method chosen over single housekeeping proteins because it accounts for lane-to-lane loading variability. An MTT assay confirmed that none of the treatments reduced cellular metabolic activity, meaning the molecular changes could not be attributed to toxicity.

The results were unambiguous. DHA alone produced the largest increase in PSD-95 protein abundance by immunoblot, the strongest increase in PSD-95 puncta density (adjusted p = 0.0005), and the greatest total colocalization with synaptophysin (adjusted p < 0.0001). The proportion of PSD-95 apposed to synaptophysin-positive regions also rose with DHA (adjusted p = 0.0214), whereas EPA actually reduced this proportion relative to control (adjusted p = 0.0079). Among the mixed formulations, the DHA-dominant 2:1 ratio most closely resembled pure DHA, while the equimolar and EPA-dominant compositions showed progressively attenuated responses. Critically, the PSD-95 puncta increases extended into neuronal processes: puncta density normalized to neurite length was significantly elevated by DHA, 2:1, and 1:1 DHA:EPA, indicating that the effect is not confined to the cell body.

The glutamatergic signaling proteins told a consistent story. DHA produced the strongest increase in NMDA receptor expression (adjusted p < 0.0001), with the 2:1 mixture mirroring that response, the 1:1 mixture showing a smaller rise, and the 1:2 mixture failing to differ from control. AMPA receptor expression followed a similar but less pronounced pattern, with DHA again the strongest inducer (adjusted p = 0.0008). CaMKII levels increased across every treatment, though DHA remained the most potent. EPA alone, however, stood out for the wrong reason: it significantly reduced NMDA receptor levels relative to control (adjusted p = 0.0056), demonstrating that EPA is not biologically inert at the synapse but can actively work against the postsynaptic receptor profile that DHA supports.

Why would two structurally similar fatty acids, differing only in two carbons and one double bond, behave so differently? The answer lies in their divergent fates in the brain. DHA is the most abundant long-chain omega-3 in the central nervous system, comprising 10 to 20 percent of neuronal membrane phospholipids and exceeding 30 percent in some synaptic and retinal membranes. It is preferentially esterified into phosphatidylserine and phosphatidylethanolamine, where its six double bonds promote membrane curvature and acyl chain disorder, shaping the lateral organization of membrane microdomains that host glutamate receptors. EPA, by contrast, normally accounts for less than 0.5 percent of brain fatty acids and is rapidly turned over, consistent with a signaling role rather than a structural one. Enzymes such as long-chain acyl-CoA synthetase 6, lysophosphatidylethanolamine acyltransferase 2, and phosphatidylserine synthase 2 preferentially channel DHA into synaptic phospholipids, providing a plausible mechanistic basis for the differential retention the study observed, though these pathways were not directly measured.

The findings carry immediate implications for how omega-3 research and supplementation should be interpreted. Clinical omega-3 formulations vary enormously in DHA:EPA ratio, and prior clinical literature hints that this matters: EPA-predominant preparations have shown greater efficacy in depression, while DHA-containing formulations dominate cognitive and neurodegenerative studies. The new data suggest that when the goal is postsynaptic scaffold stability and glutamatergic signaling, diluting DHA with EPA may attenuate the benefit. The authors are careful to note a key design limitation: because total fatty acid concentration was fixed, increasing the EPA share necessarily decreased the absolute DHA dose, so the study cannot distinguish a true DHA-EPA interaction from a simple reduction in DHA exposure. Other caveats include the acute 24-hour exposure, the use of an immortalized cell line that cannot capture the cellular diversity of the human cortex, and the absence of electrophysiological measurements, meaning the protein changes should be read as evidence of molecular reorganization rather than proven functional enhancement.

Even with those limitations, the study delivers a clear and potentially consequential message: omega-3 effects on synaptic biology depend on fatty acid identity and relative composition, not merely total omega-3 intake. DHA and EPA are not interchangeable, and the DHA:EPA ratio is a modifiable variable that shapes postsynaptic scaffold organization. The research team suggests that future work in human iPSC-derived neurons, lipidomic profiling of membrane remodeling, pathway-inhibition experiments targeting candidates such as PI3K/Akt and MAPK/ERK, and electrophysiological recordings will be needed to determine whether DHA-dominant formulations can translate their molecular signature into genuine synaptic and network benefits. For conditions marked by disrupted PSD-95 organization, from rare genetic synaptopathies to Alzheimer’s disease, the possibility that a carefully tuned lipid composition could stabilize the postsynaptic apparatus offers a compelling, testable direction for lipid-based therapeutics.

Subject of Research: Effects of DHA and EPA omega-3 fatty acid ratios on PSD-95 scaffold organization and glutamatergic signaling in human neuronal cells

Article Title: DHA-dominant Omega-3 compositions enhance PSD-95 organization and glutamatergic protein expression

Article References: Medina, G., Baird, S. D., Rodriguez-Alcocer, L. A., & MacKenzie, A. E. (2026). DHA-dominant Omega-3 compositions enhance PSD-95 organization and glutamatergic protein expression. Frontiers in Synaptic Neuroscience, 18, Article 1947075. https://doi.org/10.3389/fnsyn.2026.1947075

Image Credits: AI Generated

DOI: 10.3389/fnsyn.2026.1947075

Keywords: DHA, EPA, omega-3 fatty acids, PSD-95, synaptic plasticity, glutamate receptors, NMDA receptor, AMPA receptor, CaMKII, SH-SY5Y neurons, postsynaptic density, synaptophysin