Porphyromonas gingivalis lipopolysaccharide induces cognitive dysfunction, mediated by neuronal inflammation via activation of the TLR4 signaling pathway in C57BL/6 mice

You are sitting in a dentist's waiting room. Someone across from you has chronic periodontitis — swollen, bleeding gums — the kind of persistent infection that affects nearly half of all adults at some point in their lives. It feels local, mundane, a mouth problem. But researchers at Shanghai Jiao Tong University have now traced a molecule shed by the bacteria in that inflamed tissue all the way to the brain, where it switches on an inflammatory cascade that looks strikingly like the early stages of neurodegeneration. The paper by Jing Zhang and colleagues asks a pointed question: can a bacterial toxin from your gums, delivered into the bloodstream, impair learning and memory — and if so, how? The bacterium at the center of this story is Porphyromonas gingivalis, a Gram-negative anaerobe that dominates the microbial community in chronic periodontitis. Like all Gram-negative bacteria, it carries lipopolysaccharide, or LPS, in its outer membrane. LPS is essentially a molecular flag that the immune system reads as "danger," triggering the release of cytokines, reactive oxygen species, and other inflammatory mediators. However, P. gingivalis-LPS is not the same molecule as the Escherichia coli LPS that most immunology textbooks use as the standard example. The two differ structurally, and they engage the immune system differently — activating different receptor combinations and producing different cytokine profiles. That distinction matters because most of the prior experimental work linking bacterial LPS to cognitive dysfunction used E. coli-LPS. The P. gingivalis variety had barely been studied in this context, despite a striking detail: Poole and colleagues had previously detected P. gingivalis-LPS in brain tissue at autopsy in patients with Alzheimer's disease. Association without mechanism is not an explanation. Zhang and colleagues set out to build the mechanism. They worked with sixty eight-week-old male C57BL/6 mice, divided into five groups: saline control, P. gingivalis-LPS alone, E. coli-LPS alone, a TLR4 inhibitor called TAK-242 alone, and P. gingivalis-LPS plus TAK-242. TLR4, or Toll-like receptor 4, is the primary surface receptor that recognizes bacterial LPS and initiates downstream immune signaling. TAK-242 selectively blocks it. Mice received intraperitoneal injections of P. gingivalis-LPS at 5 milligrams per kilogram, with TAK-242 given one hour before LPS in the combination group. Behavioral testing started seven days later, once any acute motor effects had resolved. Three behavioral tests were used, and the choice of three is deliberate — each probes something different. The Open Field Test measures spontaneous locomotion and anxiety. The Morris Water Maze tests spatial learning and memory, requiring mice to navigate a circular pool to find a hidden platform using visual cues around the room. The Passive Avoidance Test measures fear-based memory: a mouse that remembers receiving a foot shock in the dark compartment of a two-chamber device will hesitate before re-entering it. The logic is that if you only ran one test and saw a deficit, you couldn't rule out that the mice were simply sick or lethargic. Running all three lets you isolate the cognitive effect. And that isolation is exactly what the results delivered. In the Open Field Test, there were no significant differences between any groups — not in total distance traveled, time in the center, rearing behavior, or grooming. The LPS-treated mice were not sluggish. They were not anxious. Physically, they appeared normal. Then came the Morris Water Maze, and the story changed completely. During five days of training, all mice gradually learned to find the hidden platform faster — but the P. gingivalis-LPS group fell further and further behind. By day five, control mice needed just 24 seconds on average to find the platform. LPS-treated mice needed nearly 63 seconds. That gap didn't close. The probe trial, conducted on day six with the platform removed, made the memory impairment unmistakable. Control mice spent about 36 percent of their swimming time in the quadrant where the platform had been. P. gingivalis-LPS mice spent only 18 percent there, essentially searching at random. TAK-242 brought that figure back up to 35 percent, nearly indistinguishable from controls. The Passive Avoidance Test told the same story: LPS-treated mice re-entered the dark shock compartment after only 54 seconds, compared to 166 seconds for controls. TAK-242 pushed that latency back up to 134 seconds. The inhibitor didn't just nudge the numbers — it largely restored normal behavior. So, what was happening inside the brain? Zhang and colleagues used immunohistochemistry to examine the cerebral cortex and hippocampus — the regions most associated with spatial learning and memory. Both microglia, the brain's resident immune cells, and astrocytes, which support and regulate neurons, were activated in LPS-treated animals. Activated microglia showed enlarged cell bodies and irregular protrusions; activated astrocytes appeared hypertrophied with more numerous GFAP-positive cells compared to controls. This morphological change is the physical signature of neuroinflammation taking hold. The molecular data confirmed it. In cortex tissue, TNF-alpha messenger RNA rose approximately fourfold in LPS-treated mice compared to controls. Interleukin-8 rose fourfold as well. Interleukin-1 beta climbed about threefold, and interleukin-6 about twofold. Protein measurements by enzyme-linked immunosorbent assay paralleled those messenger RNA changes — all four cytokines were significantly elevated at the protein level. And all of it was attenuated by TAK-242, which reduced both messenger RNA and protein levels of every cytokine measured and prevented the microglial and astrocyte activation seen in the hippocampus and cortex. That brings us to the signaling chain — the molecular mechanism that connects LPS arriving in the bloodstream to inflammation erupting in the brain. TLR4 is a pattern-recognition receptor that evolved to detect bacterial surface molecules. When P. gingivalis-LPS engaged it, TLR4 and its co-receptor CD14 were both upregulated — roughly twofold at the messenger RNA level, with elevated protein confirmed by western blot. Downstream, Zhang and colleagues measured increased IRAK1 protein, which is part of the canonical signaling module that TLR4 recruits after activation. IRAK1 then drives activation of NF-kappa-B — or nuclear factor kappa B — the transcription factor that controls inflammatory gene expression. The readout for NF-kappa-B activation here was the ratio of phosphorylated p65 to total p65, with higher phosphorylation indicating an active signaling state. That ratio was elevated in LPS-treated mice, tying receptor engagement to transcriptional output. The receptor specificity is worth pausing on. P. gingivalis-LPS increased TLR4 and CD14 significantly but did not significantly change TLR2 or TLR3 expression. E. coli-LPS, by contrast, produced an approximately 2.5-fold increase in TLR2. This differential pattern supports the conclusion that P. gingivalis-LPS was signaling primarily through TLR4 rather than other Toll-like receptors — and it's why blocking TLR4 specifically with TAK-242 was sufficient to prevent both the molecular inflammation and the cognitive deficits. TAK-242 reduced TLR4 and CD14 messenger RNA, lowered IRAK1 protein, and brought the phospho-p65 to p65 ratio back down. The whole cascade collapsed when TLR4 was blocked at the top. The picture that emerges is a linear chain: periodontal bacteria shed LPS, which enters systemic circulation, reaches the brain, binds TLR4 and CD14, recruits IRAK1, activates NF-kappa-B, and drives production of TNF-alpha, interleukin-1 beta, interleukin-6, and interleukin-8 in the cortex and hippocampus — precisely the regions that underwrite spatial learning and fear-based memory. The behavioral deficits follow directly. There are real limits to acknowledge here. This is a mouse model. LPS was injected intraperitoneally — a shortcut to systemic exposure that bypasses the slower, sustained process of natural periodontal infection. The behavioral assays are proxies for human cognition, not direct measures of it. A single injection one week before testing is not the same as the chronic, low-grade exposure that characterizes periodontitis over years. Within those limits, though, the TLR4 specificity is a meaningful signal. A single pathway inhibitor was sufficient to prevent both the neuroinflammation and the memory impairment. That points to a targetable mechanism, not just a general inflammatory mess. The question the study leaves open is the one that will drive the next decade of work: does controlling periodontal infection in humans measurably reduce neuroinflammatory signaling? The mouth-brain axis, as Zhang and colleagues demonstrate here, is a frontier worth watching closely. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.