Endotoxins are toxic molecules embedded in the outer membrane of gram-negative bacteria, released when the bacteria die or break apart. Chemically, they are lipopolysaccharides (LPS), large molecules built from fats and sugars that serve as both structural scaffolding for the bacterium and a potent trigger of the human immune system. Unlike toxins that bacteria actively secrete, endotoxins are part of the bacterial body itself, which is how they got their name in the late nineteenth century when Richard Pfeiffer first described a heat-stable poison that could not be separated from the bacterial cell during the organism’s life.
Three Layers, Three Jobs
An LPS molecule has three distinct regions, each with a different role. At the base sits lipid A, a fatty structure that anchors the whole molecule into the bacterium’s outer membrane. Above lipid A is a short core sugar chain, and extending outward from that is the O-antigen, a long, repeating polysaccharide chain that projects into the surrounding environment.1PubMed Central. Structural characterization of a model gram-negative bacterial surface using lipopolysaccharides from rough strains of Escherichia coli LPS sits exclusively on the outer leaflet of the outer membrane, meaning it faces outward like molecular armor.2PubMed Central. On the essentiality of lipopolysaccharide to Gram-negative bacteria
Each layer does something different for the bacterium. Lipid A forms a tight hydrophobic barrier that makes the outer membrane difficult for antibiotics and detergents to penetrate. The core sugars connect the fatty anchor to the outermost decorations. And the O-antigen is where gram-negative species show off their individuality: its structure varies enormously between bacterial strains, and it helps bacteria resist being killed by the complement system and other host defenses.3PubMed Central. Lipopolysaccharide O-antigens-bacterial glycans made to measure This layered architecture means that an LPS molecule simultaneously works as structural glue, a permeability shield, and an identity badge for the bacterium.
Lipid A Is Where the Danger Lives
From the human body’s perspective, the most important part of LPS is lipid A. This is the piece the immune system reacts to, and it is responsible for nearly all the toxic effects associated with gram-negative infections. Lipid A consists of a sugar backbone decorated with multiple fatty acid chains and phosphate groups. Variations in those fatty acid chains, specifically their length, number, and arrangement, determine how strongly a given lipid A molecule activates the immune system.4PubMed Central. Variation, Modification and Engineering of Lipid A in Endotoxin of Gram-Negative Bacteria
The classic highly stimulatory form of lipid A, found in bacteria like Escherichia coli and Salmonella, carries six fatty acid chains. Remove one or two of those chains, or alter the phosphate groups, and the molecule’s ability to trigger inflammation drops dramatically. This structural sensitivity is not just a curiosity. It is the basis for one of the most successful vaccine technologies of recent decades, as we will see later.
How Your Body Detects LPS
The immune system has evolved a specialized alarm system for detecting lipid A. When gram-negative bacteria die in the body, whether killed by antibiotics, the immune system itself, or simply running through their life cycle, fragments of LPS scatter into surrounding tissues and blood. These fragments are first grabbed by a circulating protein called LPS-binding protein (LBP), which shuttles them to another protein on the surface of immune cells called CD14. CD14 then hands the LPS off to a receptor complex made of Toll-like receptor 4 (TLR4) and a small helper protein called MD-2.5PubMed. Initial responses to endotoxins and Gram-negative bacteria
This relay system is remarkably sensitive. Humans can mount an immune response to picogram quantities of LPS, amounts so small they are essentially invisible by most chemical tests. Once MD-2 binds to lipid A, it causes two TLR4 molecules to pair up, which is the signal that something dangerous is present. Research has shown that this pairing happens entirely through contacts on the outside of the cell. The interior portion of TLR4 is not even required for the receptor to be pulled into the cell during the inflammatory response.6Immunity. CD14 and MD-2 Transport Lipopolysaccharide and Toll-like Receptor 4 into Cells to Promote Signaling
Two Signaling Roads Lead to Inflammation
Once TLR4 is activated, the signal branches into two distinct downstream pathways. One pathway, called the MyD88 pathway, rapidly turns on genes that produce classic inflammatory molecules like TNF-α and interleukin-6. The other, called the TRIF pathway, takes a slower route and produces a different set of signals, including type I interferons, which are more commonly associated with antiviral defenses. Both pathways work together to generate the full inflammatory response to LPS.7PubMed Central. Dual signaling of MyD88 and TRIF is critical for maximal TLR4-induced dendritic cell maturation
This dual-pathway design makes TLR4 unique among immune receptors and helps explain why LPS triggers such a broad and powerful inflammatory response. It also explains why certain substances can selectively dampen parts of the response without eliminating all of it. Alcohol is one example: binge drinking suppresses both TLR4 pathways in human immune cells, reducing the production of TNF-α, interleukin-6, and interferon-β, which partly explains why heavy drinkers are more vulnerable to gram-negative infections.8PubMed Central. Human Binge Alcohol Intake Inhibits TLR4-MyD88 and TLR4-TRIF Responses but Not the TLR3-TRIF Pathway
A Second Alarm Inside the Cell
TLR4 is not the only way the body senses LPS. If gram-negative bacteria invade a cell or if LPS somehow reaches the cytoplasm (the interior compartment of a cell), an entirely separate detection system kicks in. In humans, two enzymes called caspase-4 and caspase-5 can directly bind to lipid A inside the cell.9PubMed. NLRP3 inflammasome activation downstream of cytoplasmic LPS recognition by both caspase-4 and caspase-5 When they do, the cell essentially self-destructs through a process called pyroptosis, a violent form of cell death that spills inflammatory contents into the surrounding tissue and alerts neighboring immune cells.
This intracellular sensing system acts as a backup. Even if bacteria manage to slip past the TLR4 alarm on the cell surface, the caspase system catches them once they are inside. Pyroptosis also triggers the NLRP3 inflammasome, a large protein complex that amplifies inflammation further by processing the potent cytokine interleukin-1β.10Trends in Pharmacological Sciences. Intracellular Lipopolysaccharide Sensing as a Potential Therapeutic Target for Sepsis The caspase-4 and caspase-5 systems appear to have some overlap in function, suggesting that the body maintains redundant safeguards against gram-negative intruders.
When the Response Goes Too Far
At controlled levels, the inflammatory response to LPS is protective. It recruits immune cells, opens blood vessels to allow white blood cells into infected tissue, and triggers fever. The problems begin when LPS enters the bloodstream in large amounts, as happens during severe gram-negative infections. This can spark a systemic inflammatory cascade that damages the host far more than the bacteria themselves. Inflammatory mediators like cytokines, nitric oxide, and eicosanoids flood the circulation, causing blood pressure to plummet, blood to clot inappropriately in small vessels, and organs to start failing one by one.11Trends in Molecular Medicine. Endotoxins in Gram-Negative Bacteria: Structure and Impact
This is the pathology behind gram-negative sepsis, which remains one of the leading causes of death in intensive care units worldwide. The cruel irony is that antibiotic treatment can temporarily worsen the problem: killing large numbers of bacteria releases a burst of LPS from their disintegrating membranes, sometimes provoking a more intense inflammatory flare than the living infection did. This phenomenon has driven decades of research into drugs that might neutralize LPS directly. Polymyxin B, an old antibiotic, binds tightly to lipid A and can neutralize its toxicity, but its own side effects limit its usefulness in this role.12PubMed. Mapping residue-specific contacts of polymyxin B with lipopolysaccharide by saturation transfer difference NMR
Endotoxin Tolerance and Immune Memory
Repeated or prolonged exposure to low doses of LPS does something unexpected: instead of ramping up inflammation further, the immune system dials it down. This phenomenon, called endotoxin tolerance, involves a reprogramming of immune cells so that they produce fewer inflammatory cytokines upon subsequent LPS exposure.13PubMed. Molecular mechanisms of innate memory and tolerance to LPS The effect is not a blanket shutdown; tolerance selectively dampens the most dangerous inflammatory outputs while keeping, or even boosting, certain antimicrobial defenses. Studies in mouse macrophages have shown that low-dose LPS pretreatment inhibits nitric oxide production while simultaneously increasing production of interleukin-6, suggesting the immune system reshuffles its priorities rather than simply going quiet.14PubMed Central. Low-dose lipopolysaccharide (LPS) pretreatment of mouse macrophages modulates LPS-dependent interleukin-6 production in vitro
Endotoxin tolerance has real clinical relevance. It may explain why patients with chronic infections or prolonged ICU stays sometimes develop an “immune paralysis” that leaves them vulnerable to secondary infections. Their immune cells have seen so much LPS that they have downregulated their ability to respond to new threats. Understanding the molecular switches that control tolerance is an active area of sepsis research.
Metabolic Endotoxemia and the Gut
You do not need a raging infection to have LPS circulating in your blood. Your gut is home to enormous numbers of gram-negative bacteria, and small amounts of LPS routinely leak through the intestinal lining into the bloodstream. Normally these trace amounts are cleared without incident, but when the gut barrier becomes more permeable, as happens with high-fat diets, obesity, and heavy alcohol use, LPS levels in the blood can rise chronically. This low-grade, persistent elevation is called metabolic endotoxemia.15PubMed Central. Regulation of Gut Microbiota and Metabolic Endotoxemia with Dietary Factors
Metabolic endotoxemia has become a focus in research on obesity, type 2 diabetes, non-alcoholic fatty liver disease, and even neurodegenerative conditions. The hypothesis is that a chronic trickle of LPS from the gut maintains a low but constant level of inflammation that contributes to insulin resistance, fat accumulation in the liver, and damage to tissues over years.16PubMed Central. The metabolic endotoxemia and gut microbiota: research trajectories and hot trends across the centuries (1999-2024) The precise mechanisms linking gut-derived LPS to each of these conditions are still being worked out, and the field has more promising leads than hard answers at this point. But the basic connection between a leaky gut, circulating endotoxin, and chronic inflammation is well enough established that it has reshaped how researchers think about metabolic disease.
How Bacteria Remodel Lipid A to Hide
Bacteria are not passive targets in the arms race with the immune system. Many gram-negative pathogens actively modify their lipid A structure to reduce detection by TLR4. These modifications include adding extra chemical groups that mask the phosphate residues TLR4 recognizes, changing the number of fatty acid chains, or swapping in different sugar backbones. The net effect is a lipid A that triggers a weaker immune response, buying the bacterium time to establish an infection.17PubMed Central. Fortifying the barrier: the impact of lipid A remodelling on bacterial pathogenesis
Some pathogens go further. Helicobacter pylori, the bacterium that causes stomach ulcers, produces a lipid A so heavily modified that it barely registers on the TLR4 system at all. Commensal bacteria in the gut tend to have mildly stimulatory lipid A structures, enough to keep the immune system primed but not enough to cause damaging inflammation, a balance that appears to have been shaped by millions of years of coexistence.18PubMed. Structure and function: Lipid A modifications in commensals and pathogens The enzymes responsible for these modifications can act during or after lipid A synthesis, giving bacteria flexibility to adjust their outer coat in response to environmental signals like temperature, pH, or the presence of antimicrobial peptides.4PubMed Central. Variation, Modification and Engineering of Lipid A in Endotoxin of Gram-Negative Bacteria
Outer Membrane Vesicles as LPS Delivery Vehicles
Bacteria do not only release LPS when they die. Living gram-negative bacteria routinely pinch off small bubbles from their outer membrane called outer membrane vesicles (OMVs). These vesicles carry a full complement of outer membrane components, including large amounts of LPS, and can travel far from the parent bacterium. OMVs act as natural delivery vehicles, ferrying virulence factors and immune-stimulating molecules directly to host cells without the bacterium needing to be in direct contact.19PubMed Central. Virulence and immunomodulatory roles of bacterial outer membrane vesicles
This means the immune system encounters LPS not just from dead bacteria floating in body fluids, but also from these purposefully shed packages produced by living, growing bacteria. OMVs are now thought to play a role in seeding inflammation at sites distant from the actual infection and may help explain how gram-negative bacteria can cause tissue damage in organs they never physically reach.
Turning Endotoxin Into a Vaccine Ingredient
The same inflammatory power that makes LPS dangerous also makes it useful, if the structure is carefully modified. Monophosphoryl lipid A (MPL) is a chemically detoxified version of lipid A, first derived from Salmonella LPS, that retains the ability to activate TLR4 and boost adaptive immune responses but produces far less of the dangerous inflammatory cascade. MPL was among the first of a new generation of TLR-based vaccine adjuvants used at large scale in human populations.20PubMed Central. Putting endotoxin to work for us: monophosphoryl lipid A as a safe and effective vaccine adjuvant It appears in licensed vaccines against hepatitis B, human papillomavirus, and malaria, among others.
The trick lies in what was removed. Full lipid A has two phosphate groups and six fatty acid chains; MPL has one phosphate group, which dramatically reduces its ability to drive the MyD88-dependent inflammatory arm while preserving the TRIF-dependent arm that helps shape long-lasting immune memory. Safety evaluations have confirmed that MPL produces far fewer side effects than native LPS, with broad activity across different branches of the immune system.21PubMed. Safety evaluation of monophosphoryl lipid A (MPL): an immunostimulatory adjuvant The development of MPL is one of the cleaner examples of taking a dangerous biological molecule and engineering it, through structural understanding, into something medically beneficial.
Testing for Endotoxin in Pharmaceuticals
Endotoxins are extraordinarily heat-stable. Standard sterilization procedures that kill bacteria do not destroy LPS. In fact, sterilizing a contaminated product can make the endotoxin problem worse by lysing bacteria and releasing their LPS into solution.22PubMed Central. Atlantic Horseshoe Crabs and Endotoxin Testing: Perspectives on Alternatives, sustainable Methods, and the 3Rs This makes endotoxin testing a regulatory requirement for injectable drugs and medical devices. Even a tiny amount of LPS contamination in an IV drug can cause fever, chills, and in severe cases circulatory collapse.
For decades, the standard test has relied on a clotting enzyme extracted from the blood of Atlantic horseshoe crabs. This lysate-based test is exquisitely sensitive to LPS but raises conservation concerns, since hundreds of thousands of crabs are bled annually. Recombinant factor C (rFC) assays, which use a lab-produced version of the horseshoe crab enzyme, have emerged as a synthetic alternative. Studies across a range of pharmaceutical products have found rFC to be accurate and reliable for endotoxin detection.23PubMed Central. A Study on the Application of Recombinant Factor C (rFC) Assay Using Biopharmaceuticals Independent evaluations across many product types have reached the same conclusion, supporting rFC as a suitable replacement for the animal-derived reagent.24PDA Journal of Pharmaceutical Science and Technology. Application of Recombinant Factor C Reagent for the Detection of Bacterial Endotoxins in Pharmaceutical Products
Removing endotoxin from a product once it has been detected is its own challenge. Most removal methods rely on non-specific adsorption, where charged or hydrophobic surfaces grab LPS out of solution. The problem is that these surfaces also grab useful proteins, reducing the yield of the drug being manufactured. Newer approaches use antibodies that target lipid A specifically, pulling out endotoxin while leaving the therapeutic protein intact. One such system using anti-lipid A antibody-coated microparticles demonstrated high endotoxin removal capacity while preserving more protein than non-specific methods across formulations of increasing complexity.25PubMed Central. Recovering What Matters: High Protein Recovery after Endotoxin Removal from LPS-Contaminated Formulations Using Novel Anti-Lipid A Antibody Microparticle Conjugates
An Ancient Molecule Still Being Understood
Evolutionary analysis of the enzymes that build the lipid A portion of LPS has revealed that the pathway did not appear all at once. Some gram-negative bacteria carry only a subset of the nine enzymes typically involved in lipid A synthesis, and others have none at all, suggesting the pathway was assembled in stages over evolutionary time. Within the Proteobacteria, which include many medically important gram-negative species, gene duplication events followed by specialization appear to have shaped the modern biosynthetic machinery.26PubMed Central. Evolution of the Kdo2-lipid A biosynthesis in bacteria The implication is that endotoxin, far from being a fixed feature of gram-negative life, is a product of ongoing evolutionary tinkering, still being modified by selective pressures today. That helps explain why lipid A structures vary so widely across species and why some gram-negative bacteria have drifted away from the canonical LPS model altogether.