LPS Cell Culture: Inflammatory Signaling and Cytokine Response

Lipopolysaccharide, or LPS, is one of the most widely used tools in immunology research for a simple reason: adding it to cultured cells reliably triggers a cascade of inflammatory signaling that mirrors much of what happens during a bacterial infection. When LPS lands on the surface of a macrophage or monocyte in a dish, it activates Toll-like receptor 4 (TLR4) and sets off two distinct signaling branches that together drive production of pro-inflammatory cytokines like TNF-α, IL-6, and IL-1β, along with anti-inflammatory mediators like IL-10. But the details of how that happens, how quickly, and how consistently depend on choices that researchers make about cell type, LPS source, dose, and culture conditions.

How LPS Reaches Its Receptor

LPS does not simply bump into TLR4 and flip a switch. The process involves a multi-step handoff between accessory proteins that shuttle individual LPS molecules out of their natural aggregates and deliver them to the receptor complex. LPS in solution forms micelles, and a protein called LPS-binding protein (LBP) attaches lengthwise along the surface of these micelles. A single LBP molecule can then catalyze multiple rounds of LPS transfer to CD14, a co-receptor that grabs one LPS molecule at a time and quickly dissociates from the LBP-LPS complex through charge-based interactions.¹ CD14 then hands the single LPS molecule off to the TLR4-MD2 complex, and this final transfer step depends on TLR4 itself being present.²

This relay system matters in cell culture because CD14 can exist in two forms: a membrane-anchored version on the cell surface and a soluble version (sCD14) found in serum. Cells that express little or no membrane CD14, like endothelial and epithelial cells, depend heavily on sCD14 supplied by the culture medium to respond to LPS at all. That is why serum concentration in the medium can dramatically change how cells behave when LPS is added, a point covered in more detail below.

Two Signaling Arms Downstream of TLR4

Once LPS engages TLR4-MD2, the receptor activates two parallel intracellular signaling pathways, each organized around a different adaptor protein. The first and faster route runs through the adaptor MyD88. Within minutes of LPS exposure, MyD88-dependent signaling triggers degradation of an inhibitor protein called IκB, which frees the transcription factor NF-κB to move into the nucleus and switch on genes for early pro-inflammatory cytokines, particularly TNF-α.³ At the same time, mitogen-activated protein kinases (MAPKs) are phosphorylated within five to twenty-five minutes after LPS treatment, contributing to activation of additional transcription factors.⁴ Work in astrocytes has shown that while NF-κB activation depends on MyD88, some MAPK activation proceeds independently of it, revealing that these two kinase cascades are not simply redundant branches of the same pathway.⁵

The second route runs through two different adaptor molecules, TRIF and TRAM, and kicks in somewhat later. This MyD88-independent arm is responsible for inducing type I interferons (IFN-α/β) and a set of chemokines including RANTES and IP-10.⁶ TRAM’s role is specific to TLR4; it does not participate in signaling from other Toll-like receptors. TRIF, by contrast, also handles signaling from TLR3, which detects double-stranded RNA.⁷ The interferon output from this arm has downstream consequences: in macrophages, LPS-induced type I interferons trigger production of IL-27, which in turn drives expression of IL-10, a major anti-inflammatory cytokine. So the same LPS stimulus that launches the inflammatory response also sets in motion the machinery to rein it back in, though the anti-inflammatory wave arrives later.⁸

Intracellular LPS and the Pyroptosis Pathway

TLR4 sits on the cell surface (or in endosomes), but LPS can also reach the cytoplasm during a genuine bacterial infection or through experimental delivery methods. Intracellular LPS activates a completely separate detection system. In human cells, cytoplasmic LPS directly binds and activates caspase-4 and caspase-5 (caspase-11 in mice). These caspases cleave a protein called gasdermin D (GSDMD), and the freed fragment of GSDMD punches pores in the cell membrane, causing a form of inflammatory cell death known as pyroptosis.⁹ The gasdermin D pores also allow potassium to leak out, which triggers assembly of the NLRP3 inflammasome and subsequent processing of IL-1β, amplifying the inflammatory signal further.¹⁰

This pathway is distinct from the classical TLR4 route and operates in multiple cell types, not just professional immune cells. For researchers, it means that the method of LPS delivery matters. Simply adding LPS to the culture medium primarily engages TLR4 at the cell surface, while transfecting LPS into the cytoplasm engages the caspase-4/5/11 pathway instead. The two routes produce overlapping but different downstream profiles, and conflating them can lead to confusing results.

Metabolic Rewiring During LPS Stimulation

LPS does not just activate transcription factors; it fundamentally reshapes how the stimulated cell generates energy. Within hours of LPS exposure, macrophages shift away from oxidative phosphorylation and toward glycolysis, a metabolic switch sometimes called the Warburg effect in immune cells. This shift is not merely a byproduct. Succinate, a metabolic intermediate in the citric acid cycle, accumulates during this transition, and its oxidation by succinate dehydrogenase drives production of mitochondrial reactive oxygen species (ROS) that amplify pro-inflammatory gene expression.¹¹ Blocking succinate oxidation with the inhibitor dimethyl malonate shifts the macrophage toward an anti-inflammatory gene expression profile, showing that the metabolic state itself is a control point for inflammation.

Accumulated succinate also stabilizes a transcription factor called HIF-1α under normal oxygen conditions, which would not usually happen. Stabilized HIF-1α then drives transcription of IL-1β, linking metabolic reprogramming directly to a specific cytokine output.¹² Alongside this, LPS activates NADPH oxidase (NOX2) in macrophages, generating additional reactive oxygen species outside the mitochondria.¹³ These oxidative signals feed back into NF-κB and MAPK pathways, reinforcing the inflammatory program already underway.

Cell Type Shapes the Response

Not all cells in a culture dish respond to LPS in the same way or even through the same receptor. Macrophages and monocytes are the canonical responders, but endothelial cells, epithelial cells, fibroblasts, and even astrocytes all express TLR4 and can mount responses. The magnitude and character of those responses differ substantially. Comparing commonly used murine macrophage lines, RAW 264.7 cells and bone marrow-derived macrophages produce the highest TNF-α levels after LPS stimulation, outperforming splenic macrophages and bone marrow-derived dendritic cells. RAW cells also express much higher levels of surface CD14 (roughly 89% positive) compared with bone marrow-derived macrophages (about 53% positive), which likely contributes to their robust responsiveness.¹⁴

When the question involves a barrier tissue rather than a free-floating immune cell, the picture gets more complex. In a model of the alveolar-capillary barrier, epithelial cells grown alone showed no loss of barrier integrity when exposed to LPS at 10 or 20 μg/ml from either side. Only when epithelial cells were co-cultured with endothelial cells and LPS was applied from the endothelial (basolateral) side did barrier function drop, suggesting that soluble factors released by the endothelial cells mediated the damage.¹⁵ This is a reminder that single-cell-type cultures can miss interactions that dominate the biology in a real tissue.

Dose matters too, and not always in a simple more-is-more fashion. In human pulmonary microvascular endothelial cells, low doses of LPS actually tightened the cell monolayer by upregulating junction proteins and promoting cell migration, while high doses broke the barrier down by suppressing those same proteins.¹⁶ Biphasic responses like this are easy to miss if experiments only test one concentration.

Species Differences and Cross-Species Translation

One of the most persistent headaches in LPS research is that murine and human cells do not respond identically. TLR4 has evolved differently across species, leading to variation in its affinity for LPS, its expression patterns across tissues, and the downstream signals it generates.¹⁷ In practical terms, the human monocyte line THP-1 responds to lower concentrations of LPS than the murine RAW 264.7 line, and lower doses produce a greater fold increase in NF-κB nuclear translocation in THP-1 cells.¹⁸ This means that dose-response curves generated in one species cannot simply be applied to the other, and findings from murine macrophage cultures should be validated in human cells before drawing translational conclusions.

Why Serum in the Medium Matters So Much

Fetal bovine serum (FBS) is a standard supplement in cell culture, and its presence has an outsized effect on LPS experiments because serum provides both LBP and sCD14. Cells that lack membrane-bound CD14, such as endothelial and epithelial cells, rely on serum-derived sCD14 to shuttle LPS to TLR4. Experiments have shown that even high concentrations of LPS (1 μg/ml) fail to increase endothelial permeability under serum-free conditions, but adding as little as 3% normal serum restores the response. Blocking sCD14 with a neutralizing antibody eliminates the effect, and adding recombinant sCD14 can fully replace serum.¹⁹

LBP’s role in serum is to accelerate LPS transfer to CD14 but is not always strictly required. In some systems, LBP greatly enhanced the cytotoxic effect of LPS when combined with sCD14, but sCD14 alone was sufficient to enable LPS activity.²⁰ Work on human endothelial and epithelial cells confirmed that sCD14 is the essential serum component for LPS activation and that LPS-LBP complexes transfer LPS to sCD14, which then engages a cellular receptor.²¹

For anyone designing LPS experiments, the practical takeaway is that switching serum lots, reducing serum concentration, or moving to serum-free medium can fundamentally alter LPS responsiveness. Controlling for this variable, or at minimum reporting it, is essential for reproducible results.

LPS Purity and Bacterial Source

Not all LPS preparations are created equal, and the differences go beyond potency. The lipid A portion of LPS is the moiety that actually sits inside the TLR4-MD2 pocket and triggers signaling, and its structure varies between bacterial species. E. coli LPS carries predominantly six acyl chains on its lipid A, making it a strong TLR4 agonist. LPS from other bacteria can have four or five acyl chains instead, and these underacylated forms activate TLR4 much less efficiently. Testing with Shigella flexneri LPS, which has mostly tetraacylated lipid A, produced significantly lower NF-κB reporter activity than E. coli LPS in cells expressing TLR4.²² Similarly, different lipid A structures from Porphyromonas gingivalis LPS differentially modulated expression of TLR4 and TLR2 in human gingival fibroblasts, with the hexaacylated form behaving more like E. coli LPS and a tetra/pentaacylated form preferentially engaging TLR2.²³

Beyond the bacterial source, the extraction method matters. Standard commercial LPS preparations frequently contain trace amounts of contaminating lipoproteins, sometimes called “endotoxin protein.” These contaminants activate TLR2, which recognizes bacterial lipoproteins, not LPS. When commercial LPS was subjected to a phenol re-extraction to remove these contaminants, TLR2-dependent signaling vanished entirely, while TLR4-dependent signaling remained intact.²⁴ This is a significant practical concern. If an experiment aims to study TLR4-specific signaling, using standard-grade LPS without re-extraction can produce confounding TLR2 activation that muddies the results. Ultrapure or re-extracted LPS preparations eliminate this problem but cost more and are not always the default choice in published work.

Endotoxin Tolerance

Cells exposed to LPS once do not always respond the same way the second time. Repeated or prolonged LPS exposure can induce a state called endotoxin tolerance, where cells become hyporesponsive to subsequent stimulation. In bone marrow-derived macrophages from standard C57BL/6 mice, repeated LPS treatment induced robust tolerance, with previously activated genes failing to turn back on at the same level. Strikingly, macrophages from a different mouse strain (BTBR) failed to develop normal tolerance and instead showed hyperresponsive gene expression on re-stimulation. The difference was mirrored at the level of chromatin accessibility, suggesting that genetic background shapes how epigenetic marks are laid down during the first LPS encounter.²⁵

For anyone running multi-day LPS experiments or priming protocols, tolerance is not a nuisance variable to ignore. It is a genuine biological phenomenon that can flatten cytokine readings and mislead interpretation if the experimental design assumes each LPS dose hits a naive cell population. Washing cells between treatments and including untreated time-matched controls helps, but understanding that tolerance can set in within hours of the first exposure is key.

From Dish to Organism

Cell culture provides a controlled, reductible system, but it strips away much of the biology that governs LPS responses in a living organism. A direct comparison of human in vivo endotoxemia (injecting a small amount of LPS into healthy volunteers) versus ex vivo whole-blood stimulation found that some immune markers moved in opposite directions between the two models. Markers tied to neutrophil activity, such as CD177 and CD64, were among the most discrepant, likely because in vivo LPS exposure activates circulating neutrophils and recruits fresh ones from bone marrow, something that cannot happen in a blood sample sitting in a well. The ex vivo system also lacks tissue compartments and the feedback loops that connect immune cells to endothelium, stroma, and nervous system signals.²⁶

None of this means cell culture data are unreliable, but it does mean they answer a narrower question than researchers sometimes assume. A macrophage in a well can tell you whether a compound blocks NF-κB translocation. It cannot tell you whether that blockade matters when the macrophage is embedded in splenic tissue receiving signals from dendritic cells and T cells at the same time.

LPS-Stimulated Cells as a Drug Screening Platform

Despite these limitations, LPS-stimulated cell cultures remain one of the most practical tools for early-stage screening of anti-inflammatory compounds. THP-1 human monocytes stimulated with LPS and assessed for TNF-α release have been validated as a rapid screening system for both dietary supplements (such as N-acetylcysteine and green tea catechins) and pharmacologic agents (such as statins and PPAR-γ agonists).²⁷ The readout is straightforward: does the test compound reduce TNF-α production after LPS challenge? Hits from this kind of assay still need confirmation in more complex systems, but the throughput and cost make LPS-based cell culture assays a natural first filter.

LPS has also been used to build inflammation models in whole organisms for drug discovery purposes, including zebrafish, where LPS exposure induces quantifiable inflammatory responses that can be screened against compound libraries.²⁸ These in vivo models capture some of the tissue-level complexity lost in cell culture while retaining enough throughput for systematic screening.

Exosomes and Paracrine Communication After LPS Stimulation

An aspect of LPS-stimulated cultures that is easy to overlook is what happens outside the cell. Macrophages activated by LPS release exosomes, small membrane-bound vesicles, that carry a distinct cargo of cytokines and microRNAs. Out of 16 cytokines secreted by RAW 264.7 cells after LPS stimulation, 10 were detected inside the exosomes those cells released. The exosomal cargo included TNF-α, two anti-inflammatory mediators (G-CSF and IL-1Ra), and several chemokines including CCL3 and CCL4.²⁹ These exosomes can prime recipient cells for an immune challenge before those recipient cells have encountered LPS themselves. MicroRNAs packaged in the same vesicles add another layer of post-transcriptional regulation, fine-tuning the inflammatory response in cells that receive them.³⁰

This means that conditioned medium from LPS-stimulated macrophages is not just a soup of free-floating cytokines. It contains vesicle-packaged signals with their own regulatory logic. Researchers using conditioned medium transfer experiments or co-culture systems should consider whether exosomal cargo, rather than soluble cytokines alone, accounts for the effects they observe in recipient cells.