Extracellular Gelsolin Binds Lipoteichoic Acid and Modulates Cellular Response to Proinflammatory Bacterial Wall Components¹ Free

Gelsolin is an ∼84-kDa actin-binding protein first identified in the cytoplasm. Intracellular gelsolin is involved in the remodeling of actin filaments associated with cell shape changes and movement (1, 2). Cells from gelsolin-null mice exhibit a variety of motility and cytoskeletal defects. Gelsolin-null fibroblasts have pronounced actin stress fibers, a phenotype consistent with an inability to sever and remodel actin filaments (3). The functions of extracellular gelsolin are less well defined. Initially its involvement in scavenging F-actin polymers released into the circulation during cell death was the only role considered (4, 5). This preventative nature of gelsolin is an idea supported by observations that a reduction in plasma gelsolin as well as the detection of circulating gelsolin-actin complexes were reported in a variety of human and animal injury states (4, 6, 7).

One of the most important clinical examples of secondary injury in which the pathological implication of circulating F-actin was proposed is adult respiratory distress syndrome (ARDS)³ (8). Actin release in excess of the actin scavenger molecules’ capacities affects the lung because of its size and rich blood flow through narrow vessels, leading to the characteristic inflammatory changes of ARDS. In established ARDS cases, plasma gelsolin levels were found to be on average 30% of normal values (4). Recently, repletion with exogenous plasma gelsolin was found to be beneficial in mice subjected to hyperoxia (9), a condition that often results in ARDS development. Furthermore, the treatment of mice with exogenous plasma gelsolin significantly blunted neutrophil recruitment to the lungs (9), and gelsolin was able to attenuate vascular permeability associated with burn injury in rats (5). Reduction in plasma gelsolin levels was also observed in patients with sepsis, myocardial infarction, hepatitis, myonecrosis (10), and trauma (11, 12). However, an insufficiency in the blood F-actin buffering system has not been observed, and the involvement of plasma gelsolin in inflammatory mediator transport has been proposed (13).

Plasma gelsolin binds LPS from various bacteria, and some LPS-induced cellular functions are neutralized by gelsolin and by a peptide based on gelsolin residues 160–169 (14). The gelsolin P2 peptide (residues 150–169) binds LPS with greater affinity than it binds lysophosphatidic acid (LPA) (15), a potent extracellular agonist that influences endothelial cell migration and proliferation (16). Plasma gelsolin was also found to interfere with platelet-activating factor (PAF)-induced cellular activation in vitro, suggesting a protective mechanism for gelsolin in vivo (17). These findings suggest a new role for plasma gelsolin in inflammatory response and, in this context, actin may directly affect gelsolin binding to LPS, LPA, or PAF.

LPS and lipoteichoic acid (LTA) represent the major virulence factors of Gram-negative and Gram-positive bacteria, respectively. LTA concentrations can reach higher levels at infectious sites compared with LPS. Reported local tissue concentrations of LTA can be as high as 26 μg/ml (18), which may be associated with the fact that 10⁷ Gram-positive bacteria contain as much as 1 μg of LTA whereas 10⁷ Gram-negative bacteria contain only 20 ng of LPS (19). The primary transmembrane proteins that are activated by proinflammatory bacterial moieties such as LPS and LTA belong to the TLR family. Delivery of bacterial molecules from external fluids to the cell membrane and ultimately to TLR2 or TLR4 (dominant receptors for LTA and LPS, respectively) is complex and involves a number of other factors such as sCD14, LPS-binding protein (LBP), MD2, and moesin (20, 21).

One important factor that determines the toxicity of bacterial products to host cells is the geometry of their aggregation state. Low and high toxicity of LPS was associated with packing into lamellar and hexagonal phases, respectively (22). LTA and LPS obtained from different bacteria species induce the release of inflammatory cytokines such as TNF-α, IL-1β, IL-6, and IL-8 (19, 23). All TLR signaling pathways elicit MyD88- or TRIF (TIR domain-containing adaptor-inducing IFN-β)-dependent activation of the transcription factor NF-κB. Those signaling cascades involve recruitment of different proteins such as IL-1R-associated kinase (IRAK), TNFR-associated factor 6 (TRAF6), TGF-β-activated kinase-1 (TAK1), IKK complex, and MAPK (24). In RAW 264.7 macrophages, LTA was also found to activate the PI3K/AKT pathway and p38 MAPK-kinase, which in turn initiates NF-κB activation (25). TLR2 activation, contrary to earlier observation (25), was also proposed to function as a serum-independent factor, indicating significant differences between LTA- and LPS-mediated host cell activation (26, 27). LTA is also a potent pathogenicity factor that causes cardiac dysfunction in Gram-positive sepsis (28), may cause neuronal death (29), and determines clinical outcome in patients with pneumococcal meningitis (18).

In this study we report that in addition to the capability to bind endotoxin (14), extracellular gelsolin binds LTA from different Gram-positive bacteria strains. The result of this binding is the inhibition of gelsolin’s F-actin depolymerizing activity and compromised ability of LTA to activate endothelial cells, as measured by E-selectin expression, activation of the transcription factor NF-κB, and neutrophil adhesion. Gelsolin was also found to inhibit the release of IL-8 from human neutrophils subjected to LTA, LPS, and heat-inactivated bacteria treatment.

QRLFQVKGRR (gelsolin residues 160–169) and QRL peptides from gelsolin were prepared by solid phase peptide synthesis and fluorescently labeled at their N termini by reaction with succinimidyl esters of rhodamine B as previously described (30). The cathelicidin-derived antibacterial peptide LL37 peptide was purchased from Bachem. Ultrapure LPS from Prophyromonas gingivalis (tlrl-pglps), which is a ligand for TLR2, was from InvivoGen. Macrophage-activating lipopeptide-2 (Malp-2) corresponding to the isomer originally isolated from Mycoplasma fermentans, which signals via TLR2 and TLR6, was from Alexis Biochemicals. Human LBP peptide (86–99 aa) that binds lipid A and neutralizes LPS was from Hycult Biotechnology. LPS from Escherichia coli (serotype O26:B6), Pseudomonas aeruginosa 10 (L9143), and Klebsiella pneumoniae (L4268) and LTA from Staphylococcus aureus (L2551), Streptococcus faecalis (L4015), and Bacillus subtilis (L3265) were purchased from Sigma-Aldrich. According to the manufacturer’s quality control, the preparation of LTA contained <1 ng of LPS/1 mg of LTA, and therefore LPS would contribute <10 pg/ml culture medium in the highest LTA concentration used. Purified LTA from S. aureus was prepared as described previously (23, 31). To calculate molar concentrations of LPS and LTA, we have used the lowest range of their reported molecular mass in buffer without divalent cations (stock solution was made in H₂O). Recombinant human plasma gelsolin (rhGSN) was obtained from Biogen Idec. Solution of human albumin was from Baxter Healthcare. Human aortic endothelial cells (HAECs) were obtained from BioWhittaker. ELISA kit for IL-8 determination was from BioLegend. Heat-inactivated P. aeruginosa (PAO1) and B. subtilis (American Type Culture Collection (ATCC) 6051) were obtained by autoclaving their suspension (10⁸ CFU/ml) in PBS for 1 h at ∼120°C. Monomeric G-actin was prepared from an acetone powder of rabbit skeletal muscle as described previously (32).

To determine potential effects of LTA and other TLR agonists on the structure of gelsolin, OD at 280 nm was measured in solutions containing varying concentrations of LTA (from S. aureus) and Malp-2 added to 0.1 mg/ml of rhGSN in PBS. OD was also measured with PIP₂ and phosphatidylcholine as positive and negative controls, respectively. A decrease in tyrosine and tryptophan fluorescence, due to decreased absorbance, has been previously documented as an assay for PIP₂ binding to gelsolin (33). The fluorescence of rhodamine B-QRLFQVKGRR (PBP10) or rhodamine B (RhB)-QRL (λ_ex 565 nm, λ_em 590 nm) was measured 15 min after addition of various concentrations of PIP₂ or LTA from different bacterial strains to 2 μM peptide solutions (PBP10 or RhB-QRL) in buffer A (10 mM Tris, 10 mM MES (pH 7.0)). The expectation was that peptides bound to lipids would alter rhodamine B fluorescence similarly to previous observation with LPS (14) and PIP₂ (34).

Samples of cerebrospinal fluid, saliva, bile, and blood after collection were immediately centrifuged (2000 × g, 20 min), subjected to total protein analysis, and frozen. The study was approved by the Medical University of Bialystok Ethics Committee for Research on Humans and Animals, and written consent was obtained from all subjects. After being thawed, gel sample buffer was added to the samples, which were then boiled and subjected to electrophoresis on 10% polyacrylamide gels in the presence of SDS. After electrophoresis, proteins were transferred to polyvinylidene difluoride membrane (Amersham Biosciences), which were blocked by incubation in 5% (w/v) nonfat dry milk dissolved in TBS-T (150 mM NaCl, 50 mM Tris, 0.05% Tween 20 (pH = 7.4)). Following transfer, proteins were probed using a monoclonal anti-human gelsolin Ab (Sigma-Aldrich, G4896) used at a 1/10,000 dilution in TBS-T. HRP-conjugated secondary Abs were used at a 1/20,000 dilution in TBS-T. Immunoblots were developed with the FujiFilm LAS-300 system using an ECL Plus HRP-targeted chemiluminescent substrate (Amersham Biosciences). The detection limit for gelsolin in our experimental condition for Western blot analysis was 5 ng, and the highest volume of samples applied was 20 μl. Data revealed the presence of gelsolin in all plasma and cerebrospinal fluid (CSF) samples. Gelsolin was not detected in saliva and bile samples within the detection limit for this study (data not shown).

To prepare F-actin, 150 mM KCl and 2 mM MgCl₂ were added to G-actin solutions and allowed to incubate for 1 h at room temperature. The severing activity of rhGSN (50 nM), blood serum (2.5 μl), and cerebrospinal fluid (25 - 50 μl) was measured in 0.4 μM pyrene-labeled F-actin samples in the presence of LTA (S. aureus), LPS (E. coli), LPA, or PIP₂. The fluorescence intensity of pyrene F-actin was monitored for 5 min (λ_ex 365 nm, λ_em 386 nm) using an SL-5B spectrofluorometer (PerkinElmer). Calculation of severing activity based on the rate of fluorescence decrease was performed as previously described (35).

HAECs were grown in an incubator at 37°C and 5% CO₂ in endothelial cell basal medium-2 with supplements (Cambrex Bio Science). NF-κB translocation was measured after a 2-h incubation with 10 ng/ml TNF-α (positive control), 10 μg/ml Malp-2, 0.1 μg/ml LPS from P. gingivalis, 10 μg/ml purified LTA, or with these TLR agonists that had been preincubated with 1–10 μM of rhGSN. The intracellular location of NF-κB was observed using a mAb to the NF-κB/subunit p65 (Molecular Probes), and cell nuclei were detected by counterstaining with 4′,6-diamidino-2-phenylindole dihydrochloride (Sigma-Aldrich). Individual cells were counted to assess NF-κB localization as nuclear if the two stains colocalized or as cytoplasmic if they did not (14).

HAECs were placed in four chamber slides at 1 × 10⁵ cells/ml and allowed to adhere to the surface for 24 h. Cells were then incubated in serum-free media for 6 h and placed in either no serum, 10 μg/ml LTA, 10 μg/ml LTA + 4 μM gelsolin, or 10 ng/ml TNF-α for 24 h. Cells were then fixed in 4% paraformaldehyde for 20 min at room temperature, incubated in ammonium chloride for 20 min, and treated with a mouse anti-human E-selectin Ab (BD Pharmingen, 1/100 dilution in 0.5% BSA), followed by a rabbit anti-mouse secondary Ab (1/100 dilution in 0.5% BSA). Cells were then rinsed three times in PBS and viewed with a ×63 lens.

To determine whether gelsolin is able to affect the adhesivity of neutrophils to LTA-stimulated HAECs, we measured the adhesion of calcein-AM-labeled neutrophils to confluent cultures of HAECs treated with LTA with or without gelsolin using a previously employed method (36). Endothelial cells were placed in 24-well plates and treated at 37°C for 8 h with 1–10 μg/ml of LTA or LTA + 2 μM of gelsolin. Neutrophils were isolated from human blood using the endotoxin-free Lympholyte-poly kit (Cedarlane Laboratories) and resuspended in RPMI 1640 media (Invitrogen), labeled for 30-min with 2 μM calcein-AM, resuspended with RPMI 1640 media, and incubated with HAECs. After 30 min, unbound neutrophils were washed off and the number of bound cells was determined fluorometrically using a Fluoroskan Ascent FL multiple plate reader (Labsystems). Percentage adhesion was expressed as: [(fluorescence after washing the plates − background fluorescence)/(fluorescence before washing the plates − background fluroescence)] × 100. Bound neutrophils were viewed with a Leica microscope using a ×40 objective. Images were acquired using a CoolSNAP HQ camera.

Neutrophils treated with E. coli LPS (100 ng/ml) or S. aureus LTA (5 μg/ml) with or without 2 μM gelsolin after a 2 h incubation were viewed using a Leica microscope with a ×40 objective.

Neutrophils (5 × 10⁶ cells/ml) suspended in RPMI 1640 buffer containing 2% human albumin were activated with highly purified LTA from S. aureus (0.1–10 μg/ml), LPS from E. coli (10 ng/ml), conventionally purified LTA from S. aureus (5 μg/ml), or dilutions of autoclaved bacterial suspension (1 μl/ml) with or without rhGSN (0.5–4 μM) or LBP peptide (10 μM). Cell-free neutrophil supernatants were collected by centrifugation at 5000 × g for 5 min and stored at −80°C until cytokine determination. IL-8 was measured using a sandwich ELISA, according to the manufacturer’s instructions. The detection limit was 30 pg/ml.

To test the hypothesis that binding of gelsolin to the bacterial wall components LPS or LTA will prevent LL37 membrane insertion, we evaluated LL37 antibacterial activity in the presence of rhGSN. The bactericidal activities of the LL37 peptide against Gram-negative kanamycin-resistant P. aeruginosa (PAO1) and Gram-positive B. subtilis (ATCC 6051) was measured as previously described (37). Bacteria were grown to mid-log phase at 37°C (controlled by the evaluation of OD at 600 nm) and resuspended in PBS. The bacteria suspensions were then diluted in 100 μl of solutions containing antibacterial agents by themselves or with 2 μM rhGSN. After a 1-h incubation at 37°C, the suspensions were placed on ice and diluted 10- to 1000-fold. Then, 10 μl aliquots of each dilution were spotted on P. aeruginosa isolation agar or Luria-Bertani agar plates for overnight culture at 37°C. The number of colonies at each dilution was counted the following morning. The CFU per milliliter of the individual samples were determined using the dilution factor.

The MIC of the LL37 peptide was determined by a microbroth dilution method (38) with Mueller-Hinton broth (MH) or MH supplemented with 2 mM MgCl₂ with or without the addition of rhGSN. A series of 2-fold dilutions of LL37 in 0.25× MH broth were prepared from a stock solution and placed in 96-well plates to which dilutions of B. subtilis bacteria were then added. After incubation for 18 h at 37°C, the bacterial concentration was measured as the OD at 595 nm, and the MIC was read as the lowest concentration resulting in inhibition of detectable bacterial growth.

LPS and LTA molecules are amphipathic and form aggregates of varying sizes. These aggregates can be evaluated using DLS spectroscopy (39). In the absence of surface-active agents and divalent cations, LPS and LTA self-assemble into micellar structures. The LPS or LTA aggregate size (hydrodynamic diameter) was determined using a DynaPro 99 DLS instrument. The method measures the diffusion constant of the aggregates from the autocorrelation function of scattered light intensity. The diameter is calculated from the relation D = kT/6πηR_h, where D is the translational diffusion constant, η is the solvent viscosity, k is Boltzman’s constant, and R_h is the hydrodynamic radius. To determine whether gelsolin affects the aggregation state of LPS and LTA, solutions of bacterial wall products were evaluated before and after 20 min of incubation with either gelsolin or BSA (62 μM of each).

Data are reported as means ± SD from three to six experiments. Differences between means were evaluated using the unpaired Student’s t test, with p < 0.05 being taken as the level of significance.

The effect of LTA on the fluorescence of the gelsolin-derived PBP10 peptide is shown in Fig. 1,A. As previously reported, upon interaction of PBP10 with LPS (14), there is an initial decrease in fluorescence at low LPS/peptide ratios; as the amount of LPS increased, so did the level of peptide fluorescence, suggesting insertion of the peptide-bound rhodamine B into a more hydrophobic environment (30). At the molar ratios tested, only the first stage of decreased fluorescence was seen with LTA or PIP₂. LTA from different Gram-positive bacteria, including purified LTA from S. aureus, had similar effects on PBP10 fluorescence. There was no significant fluorescence change after adding LTA to a control peptide with the sequence RhB-QRL (data not shown). Binding of LTA to intact gelsolin was evident from a change in UV absorbance shown in Fig. 1 B. Purified LTA decreases the absorbance of gelsolin, with a maximal decrease of ∼20%, similar to that seen with LPS from P. gingivalis and PIP₂. Malp-2 and zwitterionic phosphatidylcholine had no effect on gelsolin absorbance. This result indicates that in addition to gelsolin’s ability to bind LPS (14, 40) from Gram-negative bacteria, gelsolin is also able to interact with LTA, a Gram-positive bacterial wall component.

When added to 0.05 μM rhGSN, 5 and 10 μM LTA inhibit 40% and 90% of gelsolin’s severing activity, respectively (Fig. 2,A). On a molar basis, LTA has potency similar to LPA, another known gelsolin inhibitor. When added to whole human serum (Fig. 2,B) or CSF (Fig. 2 C), LTA and PIP₂ were also able to inhibit actin-severing activity, although larger amounts of lipids were needed compared with inhibition of pure gelsolin in aqueous solution. The ability of LTA to inhibit the actin-depolymerizing activity in serum is specific to gelsolin, because the actin-sequestering activity of Gc-globulin (vitamin D-binding protein), the other component of the plasma actin scavenger system (35, 41), was not affected by LTA (data not shown).

Translocation of NF-κB to the nucleoplasm represents another consequence of bacterial wall product interaction with cell membrane TLRs. As shown in Fig. 3,A, TNF-α and Malp-2 induce NF-κB translocation from cyto- to nucleoplasm of HAECs, but the effect was not inhibited in the presence of rhGSN. On the other hand, NF-κB translocation induced with LPS from P. gingivalis or purified LTA (Fig. 3 B) was almost completely prevented by 2 and 10 μM rhGSN, respectively.

LPS and LTA have been reported to induce E-selectin expression on HUVECs, with a maximum increase between 4 and 8 h (42). Similarly, we detected an increase in E-selectin expression on HAECs after activation with LTA. As shown in Fig. 4 A, 2 μM gelsolin partially prevents LTA-mediated E-selectin expression on HAECs. Fluorescence quantification revealed that the average intensity of E-selectin markers was 180 ± 75, 250 ± 95, 355 ± 180, and 260 ± 120 for control, TNF-α, LTA, and LTA + gelsolin samples, respectively.

Similar to HUVECs (42) and human lung microvascular endothelial cells (36), treatment of HAECs with LTA increases neutrophil adhesion to the cell surface. The adhesion of neutrophils to LTA (1–10 μg/ml)-treated HAECs for 8 h is shown in Fig. 4, B and C. Recombinant plasma gelsolin effectively prevents LTA-induced activation of HAECs that translates to a decrease in neutrophil adhesion. Quantification of fluorescence from calcein-AM-labeled neutrophils documents a significant decrease in neutrophil adhesion to HAEC activated with 10 μg/ml LTA in the presence of 2 μM rhGSN.

Neutrophil activation is associated with marked changes in cellular morphology. Typically, shortly after activation by LPS or LTA, neutrophils adopt an elongated shape with a rough surface and form protrusions and aggregates (Fig. 5). After 2 h of incubation with LPS or LTA in the presence of rhGSN, the morphological predictors of neutrophil activation were less pronounced and were limited to a lower population of cells (Fig. 5). This result suggests that gelsolin may modulate neutrophil activation by sequestering bacterial wall products and buffering their interaction with TLRs.

Unstimulated human neutrophils produce very low, but detectable, amounts of cytokines, and after activation they produce and release several cytokines, including IL-8, TNF-α, and G-CSF at levels 10–50 times higher compared with the resting state. Although IL-8 is produced by a variety of cell types, neutrophils are the major source of this proinflammatory cytokine (19). In the presence of LTA, time-dependent induction of IL-8 release was observed, with a maximum reached at 24 h (data not shown). IL-8 secretion induced by addition of purified LTA was also concentration dependent. The amount of IL-8 released after exposure to 10 ng/ml of LPS was comparable to that observed after neutrophil activation with 5–10 μg/ml of LTA or 1 μl/ml of heat-inactivated bacteria (Fig. 6). Neutrophils coincubated with LPS, purified and nonpurified LTA, or lysed bacteria in the presence of rhGSN released significantly lower amounts of IL-8. In the case of activation with purified LTA, we observed a partial but progressive decrease of released IL-8 when an increased concentration of rhGSN was added. In agreement with previous observations, gelsolin may function as an inhibitor of LTA/LPS-induced IL-8 synthesis (19). The inhibition of IL-8 was also observed after LPS addition in the presence of LBP peptide (aa 86–99) that binds lipid A and neutralizes LPS. Gelsolin addition to neutrophil samples 30 min after LPS or LTA treatment did not prevent IL-8 release to the same extent as when gelsolin was added together with the stimuli (data not shown). This result supports the hypothesis that gelsolin acts as an LTA/LPS buffer preventing their agonistic effect on TLRs, although gelsolin may also interfere with other mediators involved in the regulation of neutrophil synthesis or release of IL-8.

rhGSN by itself had no effect on bacterial growth or the ability of the synthetic human cathelicidin-derived antibacterial peptide LL37 to kill B. subtilis or P. aeruginosa (Fig. 7,A). The MIC value for the LL37 peptide with or without gelsolin (Fig. 7 B) was unchanged. These data suggest that gelsolin is unable to interact with LTA molecules within the intact bacterial wall in the same way as takes place upon LTA release from dividing or dying bacteria.

Amphiphilic molecules such as LPS, lipid A, and LTA form aggregates in aqueous environments above a critical micellar concentration. The actual structure of these aggregates is not a constant, but depends on concentration, solvent conditions, and the effects of other solutes that can co-assemble in the micelles (20, 22). Accordingly, the average size of LPS and LTA molecules was observed to decrease as their solutions were diluted (Fig. 8,A). Previous analyses of aqueous LPS suspensions by negative staining and platinum shadowing revealed the presence of small globular aggregates (diameter 20–80 nm) and short filaments (43, 44). Upon the addition of serum proteins, mainly albumin, LPS aggregates become larger (>200 nm) and form structures several micrometers in length (44). In our study we observed a decrease in LPS and LTA aggregate size (Fig. 8) in the presence of purified gelsolin. This effect is consistent with a gelsolin-specific interaction with LPS and LTA.

Many of the biological activities of Gram-negative bacterial endotoxin are shared by LTA, a complex glycolipid from the outer wall of Gram-positive bacteria (45). Whereas most LPS- or LTA-mediated cellular effects are similar, LPS and LTA share few steps in their activation mechanisms and signal transduction pathways (27, 42, 46). The finding that gelsolin binds to LPS (14), LPS from P. gingivalis that acts as a specific agonist of TLR2, and purified LTA with approximately the same strength as LPA and PIP₂ suggests that gelsolin may be involved in both LTA and LPS presentation to TLR4 and TLR2, acting potentially as a scavenger or delivery promoter of immunogenic bacterial wall components. This proposed function is similar to gelsolin’s involvement in the presentation of LPA to cellular receptors (13). The specificity of gelsolin’s effects on acidic bioactive lipids is enforced by lack of gelsolin’s effect on TNF-α and Malp-2 activation of NF-κB translocation in endothelial cells. As LTA and LPS are potent mediators of the innate immune response, efficacy of the host recognition may determine survival during bacterial infection. It is also possible that the binding of gelsolin to LTA or LPS determines their interaction with other proteins such as LBP, CD14, or MD2 and factors such as lipoprotein involved in host cell detection and elimination of bacterial products. Gelsolin may have a buffering effect on the availability of LTA/LPS for these targets. Therefore, LPS (29) and LTA binding to gelsolin may determine the immune response at different steps of the signaling pathway. LTA, LPS, LPA, and PAF bind to at least one common site within the gelsolin molecule (14, 47, 48), indicating the possibility for competition among these lipids. However, the possible role of gelsolin in LPS and LTA recognition may also differ from the role of LBP or CD14 protein. In healthy subjects, blood gelsolin is present at much higher concentrations than the other high-affinity ligands, and, unlike LBP and CD14, gelsolin is not an acute-phase protein (49, 50).

Gelsolin levels are decreased in several pathological inflammatory states, and lowered gelsolin levels have potential for identifying patients at risk for adult respiratory distress syndrome or multiple organ failure in some settings. Repletion of plasma gelsolin was shown to be beneficial in murine hyperoxic lung injury (51) and endotoxemia (52), and gelsolin was able to attenuate vascular permeability associated with burn injury in rats (5). The hypothesis that plasma gelsolin may have utility in clinical settings such as ARDS, burn, and sepsis is supported by the results of this study. In this context, prevention of a decrease in blood gelsolin during sepsis (10) may protect the host from bacterial wall product-mediated increase of cytokines and systemic complications that can lead to septic shock and multiorgan failure. The first animal study in which gelsolin was used to prevent septic shock (52) indicated an increase in animal survival when an injection of LPS was followed by rhGSN administration. The mechanism of gelsolin’s role in this effect is poorly understood, and it is still an open question as to which of gelsolin’s activities—F-actin severing, inactivation of LPS/LTA, or buffering of other lipid mediators—translates to an increase in survival during endotoxemia or in other critical care stages.

The presence of gelsolin in CSF (53) combined with the observed decrease in chronic immune-inflammatory diseases such as multiple sclerosis (54) and the previously reported ability of LTA to induce neuronal cell death (29, 55) and microglial cell activation (56) suggest the potential for gelsolin’s involvement in modulating LTA/LPS interaction with its ligands in the CNS. This possibility is supported by data in Fig. 2 showing that, as in blood, gelsolin-dependent severing activity of CSF is decreased by LTA or LPS.

Bacterial wall molecules LPS and LTA are considered to be major targets for bactericidal activity of LL37. This activity occurs stepwise, beginning with attachment to the outer membrane, which induces a conformational change in the peptide resulting in insertion into the membrane (37, 57, 58, 59). In our study we found that gelsolin had no effect on either the growth of Gram-positive and Gram-negative bacterial strains or on the antibacterial activity of cathelicidin-derived LL37. This observation suggests that interaction of gelsolin with LPS and LTA is unlikely to occur when those molecules are packed in organized bacterial membrane structures or that gelsolin is trapped on bacterial surfaces before it can reach those molecules. Based on this observation, we suggest that gelsolin may interfere with bacterial LPS and LTA upon their release from dying bacteria. Additionally, data presented in Fig. 8 indicate that this interaction may affect LPS and LTA aggregation state.

In conclusion, a combination of biochemical and cellular studies show that the interaction of gelsolin with LPS and LTA results in both inhibition of gelsolin’s actin-severing activity and the ability of the bacterial toxins to induce an immune response in vitro. Gelsolin interaction with proinflammatory agents highlights its potential as a treatment for subjects with severe infection, or in urgent conditions when gelsolin blood concentration decreases.

R. Bucki and P. Janmey began a sponsored research agreement with Critical Biologics in May 2008 which is not related to the present study. This manuscript was submitted before the sponsored research agreement was initiated, and no support from Critical Biologics enabled any of the work presented in this manuscript.