Structure–Activity Relationship of Plesiomonas shigelloides Lipid A to the Production of TNF-α, IL-1β, and IL-6 by Human and Murine Macrophages

Plesiomonas shigelloides is a Gram-negative bacterium that is associated with diarrheal disease in humans. Lipopolysaccharide (LPS) is the main surface antigen and virulence factor of this bacterium. The lipid A (LA) moiety of LPS is the main region recognized by target cells of immune system. Here, we evaluated the biological activities of P. shigelloides LA for their abilities to induce the productions of proinflammatory cytokines (TNF-α, IL-1β, and IL-6) by human and murine macrophages [THP-1 macrophages and immortalized murine bone marrow-derived macrophages (iBMDM)]. Four native P. shigelloides LA preparations differing in their phosphoethanolamine (PEtn) substitution, length, number, and saturation of fatty acids were compared with Escherichia coli O55 LA. The bisphosphorylated, hexaacylated, and asymmetric forms of the P. shigelloides and E. coli LA molecules had similar activities in human and murine macrophages, indicating that shortening of the acyl chains in P. shigelloides LA had no effect on its in vitro activities. The PEtn decoration also had no impact on the interaction with the toll-like receptor 4/MD-2 receptor complex. The heptaacylated form of P. shigelloides LA decorated with 16:0 exhibited strong effect on proinflammatory activity, significantly decreasing the levels of all tested cytokines in both murine and human macrophages. Our results revealed that despite the presence of shorter acyl chains and an unsaturated acyl residue (16:1), the bisphosphorylated, hexaacylated, and asymmetric forms of P. shigelloides LA represent highly immunostimulatory structures.

by mediating the interaction of LPS with pattern recognition receptors, such as toll-like receptor 4 (TLR4) (1), on monocytes/ macrophages. As a result, signaling pathways are triggered followed by transcription factors activation (e.g., NF-κB) and production of proinflammatory cytokines (e.g., TNF-α, IL-1β, IL-6) that initiate and shape the host's immune response against the pathogen. In the case of bacteremia caused by Gram-negative bacteria, an excessive response to LPS may lead to sepsis and septic shock (2)(3)(4).
Bacteria produce LA molecules that are generally conserved, but show some structural variability in terms of the following: the number, length, and saturation of fatty acids; asymmetry; the carbohydrate backbone composition; and the presence of additional substituents, such as phosphate groups (P), ethanolamine (Etn), phosphoethanolamine (PEtn), 4-amino-4-deoxyl-arabinose (l-Ara4N), and d-galacturonic acid (d-GalpA) (1,2). Structurally different LA molecules show different potencies in activating the mammalian macrophages (5). E. coli LA, which displays strong immunostimulatory or endotoxic activity in the mammalian host (6), consists of a bisphosphorylated carbohydrate backbone substituted with six asymmetrically distributed fatty acids ( Figure 1A) (1,7).
Several important pathogenic bacteria (i.e., E. coli, K. pneumoniae, Shigella flexneri, Yersinia pestis, Neisseria meningitidis) have been shown to modify their LA in ways that significantly alter TLR4-dependent signaling (1,11). Such modifications include the expression of enzymes responsible for palmitoylation (leading to heptaacylated LA) (12) and the additions of P, PEtn, and/or l-Ara4N ( Figure 1A) (1). Thus, it has been hypothesized that the expression of different LA types may be a mechanism through which pathogens modulate or evade the host immune response (6).
The bacterium, P. shigelloides, is associated with diarrheal disease in humans. It has been implicated in gastroenteritis outbreaks among tropical travelers and in patients who have ingested contaminated food or water, and reportedly stands as the third most common cause of diarrhea among travelers in Japan and China (13). P. shigelloides causes acute bacterial gastroenteritis and extra-intestinal infections, such as sepsis, meningitis, cellulitis, and septic arthritis (14). The extra-intestinal diseases caused by P. shigelloides are associated with a high mortality rate, even when patients are given appropriate antibiotic therapy and intensive care (15, 16).
The structures of P. shigelloides LPS, including that of the LA region, have been extensively studied. Natural P. shigelloides LA molecules comprise a mixture of structures that differ in the numbers, lengths, and saturations of their acyl chains, as well as in PEtn substitutions. Three strains (O-serotypes) have been characterized to date with reference to LA among 85 O-serotypes of P. shigelloides (17)(18)(19), and detailed structural analyses suggest that there is significant conservation of this region among all studied O-serotypes. In general, the LA of P. shigelloides comprises a β-GlcpN4P-(1 → 6)-α-GlcpN1P disaccharide backbone substituted with the following primary fatty acids: 14:0[3-(R)-OH] at N-2 and N-2' and (R)-3-hydroxydodecanoic acid [12:0(3-OH)] at O-3 at O-3' . The heterogeneity of P. shigelloides LA reflects that the acyl residues at N-2' and O-3' are substituted with the following secondary acyls: (1) cis-9-hexadecenoic acid (9c-16:1) or hexadodecanoic acid (16:0) at N-2' and 12:0 at O-3' (strain dependent); (2) 14:0 at N-2' and 12:0 at O-3'; and (3) 12:0 at N-2' and O-3' . Other types have additional substitutions by PEtn and 16:0 as a seventh acyl chain (17)(18)(19) (Figure 1B). The LA of P. shigelloides differs from those of E. coli mainly in the length and saturation of their fatty acids. Our previously published report discussed the biological in vitro and in vivo activity of P. shigelloides (serotype O54) LPS, which was found to have a stronger effect on cytokine production by murine macrophages and 10-fold higher in vivo toxicity in the actinomycin D-sensitized mouse model, compared to E. coli O55 LPS (19). We concluded that the non-typical structures of the core oligosaccharide, which lacks phosphate residues, and the LA, which has shorter fatty acid residues than E. coli LA, might be responsible for the higher biological activity. This work provided some interesting insight into the structure-in vitro activity relationships of P. shigelloides LA in comparison with E. coli LA.
To avoid the potential for the heterogeneity of LA molecules to influence their biological activity in an experimental setting, structure-activity relationships LA-based studies are usually performed using chemically synthesized LA analogs, such as tetra-, penta-, hexa-, and heptaacylated E. coli LA (8,9,(20)(21)(22)(23)(24), or LA isolated from genetic mutants (10,25,26). Since synthetic LA analogs and mutated strains are not commercially available for P. shigelloides, we screened 85 P. shigelloides O-serotypes (unpublished results) and identified four strains that showed relatively low LA heterogeneity, and whose typical LA structures showed between-strain differences with respect to their palmitoylation and PEtn substitution. In the present study, we used the welldefined LA molecules from these four P. shigelloides strains to examine the effect of structural heterogeneity on the biological response triggered by LA in human (THP-1) and immortalized murine (iBMDM) macrophages. Human and murine models have been chosen, since species-specific recognition was demonstrated for TLR4/MD-2 complexes (4,27). This is the first report to assess the structure-in vitro activity relationships of P. shigelloides LA with respect to inducing the production of proinflammatory cytokines (TNF-α, IL-1β, and IL-6). We demonstrate how variations in the structure of P. shigelloides LA (the length of the acyl groups, the presence of PEtn, and the seventh acyl chain) affect its recognition by human and murine TLR4/MD-2 receptor complexes.

Bacterial strains and growth conditions
Bacteria of P. shigelloides serotypes O14, O30, O61, and O75 and E. coli O55:B5 [Polish Collection of Microorganism (PCM)-224] were obtained from the PCM of the Hirszfeld Institute of Immunology and Experimental Therapy PAS (Wroclaw, Poland). Lyophilized bacteria were reconstituted, and strains were cultured on agar plates. Bacteria were washed from the solid medium with buffered phosphate saline (PBS) and used to inoculate liquid medium (Davis, 1.5 l). After 48 h of culture at 37°C, the bacteria were killed with 0.5% phenol, centrifuged (4000 × g/30 min/4°C), resuspended in water, and lyophilized.

cell lines and reagents
The human monocytic cell line THP-1 was purchased from the European Collection of Authenticated Cell Cultures (ECACC). immortalized murine bone marrow-derived macrophages (iBMDM) from wild-type mice (iBMDM WT), and TLR4 knockout mice (iBMDM TLR4 −/− ) were obtained from BEI Resources.

Microextraction and analysis of la
The rapid, small-scale procedure used to directly isolate the LA molecules from freeze-dried P. shigelloides and E. coli O55 cells was performed as described previously with slight modifications (28). Briefly, bacterial cells (10 mg) were suspended in 400 µl of 1 M isobutyric acid-ammonium hydroxide (5:3, v/v), vortexed (20 min), sonicated (20 min), and incubated for 2 h at 100°C. The mixture was cooled in ice water and centrifuged (2000 × g/15 min/4°C). The supernatant was diluted with water (300 µl) and lyophilized. The samples were washed twice with 400 µl of methanol and centrifuged (2000 × g/15 min/4°C). Finally, the LA-containing sediment was suspended in water and lyophilized. To obtain E. coli LA, the LPS of E. coli O55 (Sigma-Aldrich) was subjected to mild hydrolysis with 1.5% acetic acid (15 min/100°C) as described previously (29). The LA-containing sediment was washed, resuspended in water, and freeze-dried. All LA samples were analyzed by matrix-assisted laser-desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry (MS) using an UltrafleXtreme (Bruker, Germany) instrument. The MALDI-TOF MS spectra of LA were obtained in the negative ion mode. 2′,4′,6′-Trihydroxyacetophenone [10 mg/ml in 1:1 AcN/mQ (v/v)] was used as a matrix. Each LA sample was tested for the presence of potential immunostimulatory impurities, such as other lipids, proteins, peptides, and nucleic acids, using MALDI-TOF MS and the appropriate matrix (2′,4′,6′-trihydroxyacetophenone for lipids and nucleic acids, sinapinic acid for proteins, and α-cyano-4-hydroxy cinnamic acid for peptides).

la solubilization for In Vitro assays
Lipid A was dissolved in water (1 mg/ml) by the addition of triethalamine (TEA) (5 µl/ml) and vortexed to obtain a stock solution (1 mg/ml, pH 8.1). For stimulation experiments, the LA stock solution was diluted with culture medium. Prior to cell stimulations, all LA samples were analyzed by MALDI-TOF MS within the range of m/z 500-3000 to validate the LA structures.

cell Viability assay
To examine the viability of LA-stimulated cells, we performed a colorimetric assay using a modification (30,31) of the tetrazolium salt [MTT; (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide)] method (32). Briefly, cells (THP-1 macrophages and iBMDM) in the medium dedicated for the cell line were plated (10 4 cells/well; 100 µl) to triplicate wells of a 96-well flat-bottomed plate, and incubated in a humidified atmosphere at 37°C with 5% CO2 for 24 h. Cells were stimulated with 10, 1, 0.1 µg LA (1 mg/ml stock solution of LA in the medium supplemented with 5 µl of TEA) for 24 h. To determine viability, cells were incubated with 25 µl MTT (5 mg/ml stock solution) for 4 h at 37°C, and the medium was replaced with DMSO (100 µl). Absorbance was read at after 30 min at 570 nm. Cells incubated with medium were used as a reference samples.

cytokines analysis by elisa
The concentrations of cytokines in supernatants collected from LA-stimulated human and murine macrophages were measured using human and murine TNF-α, IL-1β, and IL-6 ELISA kits according to the manufacturer's instructions (BioLegend). In all in vitro experiments, the same preparations of structurally welldefined ligands (LA O55 and LA I-IV) were used.

Data analyses
Each experiment was repeated at least three times (from the stimulation stage) in duplicate. Data are presented as the median ± SD. The results were compared by one-way ANOVA with the Tukey-Kramer multiple comparison test, and differences were considered significant at p < 0.05 versus the control (TEA-supplemented cell cultures) or the appropriate LA type (in Figures 3 and 4).

resUlTs
isolation and structural analysis of la  Lipopolysaccharide (LA) O55 was isolated by both microextraction and mild acid hydrolysis of commercially available LPS. Both LA O55 preparations revealed the same structure and heterogeneity. LA O55 (Figure 2A, inset structure) comprised effect of the Various la Types on TnF-α, il-1β, and il-6 Production in Murine Macrophages To compare the effects of in vitro stimulation of murine iBMDM with P. shigelloides LA I-IV, we used ELISA to examine the releases of the proinflammatory cytokines, TNF-α, IL-1β, and IL-6 ( Figure 3). Two preparations of E. coli O55 LA, one isolated by microextraction of the bacterial mass and one isolated by mild acid hydrolysis of commercially available O55 LPS, were used as positive controls and revealed similar stimulatory effects (data not shown for O55 LA isolated by mild acid hydrolysis of LPS). When TEA-supplemented medium was used as additional control, no biological effect was observed. Our MTT viability assay showed that the LA preparations (dose range, 1-0.01 µg of LA) had no toxic effect on cells. Experiments were performed using WT and TLR4 −/− murine macrophages. iBMDM TLR4 −/− did not demonstrate any measurable cytokine release upon stimulation with LA, which supports the involvement of the TLR4/MD-2 receptor complex in the LA-induced production of cytokines, and indicates that our preparations were free from other immunostimulants ( Figure S1 in Supplementary Material). We observed significant dose-dependent differences in the production levels of TNF-α and IL-6 by LA (I-IV)-stimulated murine macrophages compared with control (unstimulated) cells. No production of IL-1β was observed (data not shown), which is in agreement with the previous reports (33)(34)(35).
The bisphosphorylated, hexaacylated, and asymmetric forms of E. coli O55 and P. shigelloides LA I demonstrated similar abilities to induce murine macrophages to produce proinflammatory cytokines, although the production levels were slightly lower in cells treated with LA I. There was no significant difference in the amount of TNF-α produced by cells stimulated with LA O55 versus LA I for most of the tested LA concentrations, with the exception of the 1 µg dose (p < 0.05). However, LA I induced significantly less production of murine IL-6 than LA O55 at doses of 1.0, 0.1, and 0.01 µg (p < 0.05-0.001) (Figure 3). The observed trend can be explained only by the difference in the average length and saturation of the acyl groups. Moreover, a significant portion of the LA I preparation comprised the hexaacylated form with of 16:1 acyl chain instead of 14:0 or 12:0 ( Figure 2B).
Phosphoethanolamine substitution (relative abundance 42%) did not influence the activity of LA II. No statistically significant differences were observed between LA I and LA II for both cytokines (Figure 3).
In comparison with bisphosphorylated and hexaacylated LA I, the presence of the seventh acyl chain of 16:0 (relative abundance 51%) in LA III had a profound effect on the in vitro productions of murine TNF-α and IL-6. LA III was at least approximately two-and approximately fivefold less active (p < 0.001) than hexaacylated LA I in triggering the productions of TNF-α ( Figure 3A) and IL-6 ( Figure 3B), respectively.
The presence of PEtn (4%) and the decreased contribution of the 16:0 acyl chain (7%) were sufficient to attenuate the activity LA IV in a manner similar to that seen for LA III; there was no significant difference in the abilities of LA III and LA IV to trigger the production of TNF-α and IL-6 ( Figure 3).

effect of the Various la Types on TnF-α, il-1β, and il-6 Production in human Macrophages
We next compared the abilities of P. shigelloides LA (I-IV) to stimulate the productions of TNF-α, IL-1β, and IL-6 by human-THP-1 macrophages in vitro (Figure 4). Our MTT viability assay showed that the LA preparations (dose range, 1-0.01 µg of LA) had no toxic effect on cells. The tested ligands (LA I-IV, LA O55) dose-dependently stimulated human macrophages in most cases, although some deviations were seen for production of IL-1β by cells treated with 0.01 µg LA I or 1.0 µg LA III (Figure 4B).
Escherichia coli O55 LA and P. shigelloides LA I (Figure 2A,B) demonstrated similar abilities to induce proinflammatory cytokine production by human macrophages. No significant difference was observed in the abilities of LA O55 and LA I to induce the productions of TNF-α, IL-1β, and IL-6 at any tested dose (Figure 4), with the exception of the TNF-α production triggered by 0.1 µg LA (p < 0.05) (Figure 4A).
Phosphoethanolamine substitution (relative abundance 42%) did not influence the ability of LA II to stimulate the productions of TNF-α or IL-1β (Figures 4A,B). For IL-6, doses of 0.01 and 0.001 µg yielded higher cytokine production for cells treated with LA II versus LA I (p < 0.05-0.001) (Figure 4C).
The palmitoylation that was characteristic of LA III (relative abundance 51%) significantly decreased the in vitro effects (p < 0.05-0.001) of most of LA doses on the productions of TNF-α, IL-1β, and IL-6 in comparison to those triggered by the same doses of LA I (Figure 4), with the exception of 0.001 µg for TNF-α ( Figure 4A) and 0.01 and 0.001 µg for IL-6 ( Figure 4C).
The presence of PEtn (4%) and the decreased contribution of the 16:0 component (7%) were sufficient to attenuate the activity of LA IV in a manner similar to that seen for LA III; there was no significant difference in the cytokine productions triggered by the tested doses of LA III and LA IV (Figure 4).

DiscUssiOn
Recognition of LPS, a major component of the outer membrane of Gram-negative bacteria, by the TLR4/MD-2 complex is   Comparing the activity of the bisphosphorylated, hexaacylated, and asymmetric LA I with that of E. coli O55 LA demonstrated that decreasing the chain length to two to four acyl groups of 12 carbon atoms or two acyl groups of 12 carbon atoms and the presence of the one 16:1 (9c-16:1) (Figures 2A,B) (19) did not influence the in vitro activity of LA I in murine or human macrophages. LA I seemed to be a slightly weaker activator of murine macrophages, although not to a statistically significant degree. As mentioned above, the general structural differences between LA O55 and LA I (both of which are bisphosphorylated and hexaacylated) include a difference in acylation, with the latter having a higher contribution of shorter [12:0/12:(3-OH) instead of 14:0/14:(3-OH)] and unsaturated (9c-16:1 instead of 14:0 or 12:0) acyl residues on an identical bisphosphorylated disaccharide backbone. Little published data have focused on the structure-activity relationships of LA with direct reference to the acyl chain lengths. Generally, shorter acyl chains improve LA solubility and decrease aggregate formation improving cell activation. However, too long or too short acyl chains could impair effective LA/TLR4/MD-2 interactions (27). A previous comparison of TNF-α production in murine BMDM stimulated with LPS from E. coli, Y. pestis, and Psychrobacter cryohalolentis showed that shortening the average acyl chain length to 10-12 carbon atoms (as seen for P. cryohalolentis) slightly attenuated the in vitro activity of whole LPS. Moreover similar TNF-α productions were stimulated by the LPS of E. coli and that of Y. pestis, which has bisphosphorylated hexaacylated LA molecules substituted with two l-Ara4N residues and 9c-16:1 instead of 12:0 (37). Consistent with this, we herein found that the presence of 9c-16:1 in P. shigelloides LA I did not influence the tested biological activity in human and murine macrophages. By contrast, Zhang et al. investigated a synthetic analog of bisphosphorylated, hexaacylated, and asymmetric E. coli LA (5 acyl chains of 14 carbon atoms and 1 of 12 carbon atoms) and a counterpart built of shorter acyl chains (2 acyl chains of 14 carbon atoms and 4 of 12 carbon atoms), and found that the latter showed higher potency in stimulating the productions of TNF-α, IL-1β, and IL-6 by murine macrophages (38). This discrepancy among the published data for natural LA preparations may be explained by small differences in the length between acyl chains built of 10/12 and 14 carbon atoms that depending on host cell and co-existing forms of LA with additional substituents in native preparations may decrease or increase in vitro activity of LA. It was suggested that fatty acid chain length within 12-14 carbon atoms ensures biological activity maintenance. Despite the small differences in the average acyl chain lengths, all of the discussed examples have the bisphosphorylated, hexaacylated, and asymmetric structure that is the most important prerequisite for the interaction Effect of P. Shigelloides LA on Cytokines Level Frontiers in Immunology | www.frontiersin.org December 2017 | Volume 8 | Article 1741 between LA and cell receptor. Conversely, little or no endotoxic activity was found for LA carrying more extended acyl residues (at least 18-carbon long) and branched and unsaturated acyl residues (e.g., Legionella pneumophila LA with the very-long-chain fatty acid, 27-hydroxyoctacosanoic acid) (1). The PEtn decoration of P. shigelloides LA II did not impact its biological activity on murine and human macrophages. In general, no significant difference was observed in the activities of LA I and LA II, except that the production of IL-6 by human macrophages was significantly increased by 0.01 or 0.001 µg of LA II. The two phosphate groups of LA are known to be critical for complete activation of the macrophages, and an additional PEtn substitution was shown to affect the biological activity of Cronobacter sakazakii LPS (39,40). The presence of PEtn and l-Ara4N moieties in LA decreases its net negative charge, which has been shown to increase resistance to polymixin B and other cationic antimicrobial peptides (CAMPs), as well as to complement-mediated killing (10,25,26). Thus, for classical LA structures (E. coli-like), PEtn substitution seems to be an important factor for other activities than TNF-α, IL-1β, and IL-6 production by macrophages. The degree of LA phosphorylation and phosphoethanolaminylation can be correlated to inflammatory potential of bacteria and ability to induce immune tolerance in vitro (41). For example, N. meningitidis LA mixture of hexaacylated forms with 2P and PEtn, 3P, or 2P, along with apparently less abundant pentaacyl LA molecules with 2P, was more active to trigger proinflammatory cytokine response of THP-1 cells than the hexaacylated and bisphosphorylated LA. This suggested that increasing the number of phosphoryl substituents increases the potency of TNF-α induction. Moreover, the decreases in P and PEtn substitution observed for invasive isolates of N. meningitidis were correlated with increased occurrence of septicemia. The most highly phosphorylated bacterial isolates had relatively diminished ability to survive systemically and cause systemic disease (42). Among the tested structures, only the bisphosphorylated and heptaacylated LA III structure decorated with 16:0 showed a significant alteration in the proinflammatory activity. LA III triggered significantly less production of all tested cytokines in both murine and human macrophages: it was at least ~2-5and 2-4-fold less active than hexaacylated LAI and O55 LA in murine and human macrophages, respectively. In the LA molecules of E. coli and P. shigelloides, palmitoylation is related to the addition of a palmitate (16:0) to the 14:0(3-OH) at N-2 of the α-D-GlcpN residue within the carbohydrate backbone structure (12,29,(43)(44)(45).
Palmitoylation and the addition of PEtn are dependent on environmental conditions and can directly protect the bacterium against host immune defenses. Such modification alters the ability of LA to activate defense mechanisms through TLR4/MD-2 receptor complex-mediated signal transduction (4,(46)(47)(48). Heptaacylated native LA of S. enterica sv. Minnesota Re deep rough mutant strain (R595) and the synthetic analog, compound 516, were shown to moderately induce the inductions of TNFα, IL-1β, and IL-6 in human monocytes (THP-1 cell line) (48). In some studies, compound 516-treated human mononuclear cells from healthy donors or isolated monocytes/macrophages exhibited 10-to 100-fold decreases in IL-1 production compared to those triggered by E. coli O55 and its synthetic analog, compound 506 (8,49). For murine macrophages, the ability of LA to trigger TNF-α and IL-1β production was also significantly decreased for synthetic analogs of the bisphosphorylated and asymmetric heptaacylated LA of S. enterica sv. Typhimurium compared to synthetic bisphosphorylated and asymmetric hexaacylated E. coli LA (38). Palmitoylation has been well described for S. enterica sv. Typhimurium, where the PhoP/PhoQ regulatory system palmitoylates LA in response to environmental conditions (50). In response to specific environmental signals, PhoQ phosphorylates PhoP, leading to the activation or repression of over 40 different genes, including pagP, which encodes a palmitoyltransferase responsible for transferring palmitate to LA. The same study demonstrated that the addition of palmitate provides resistance to certain CAMPs by increasing the integrity of the outer membrane and preventing the translocation of CAMPs across the bilayer (46,50). Palmitoylation has also been described in E. coli and K. pneumoniae, and similar modifications at various positions in the LA structure have been described for Legionella pneumophila, Pseudomonas aeruginosa, Bordetella bronchiseptica, Y. enterocolitica, and Y. pseudotuberculosis LPS (19). This is the first report concerning the biological activity of P. shigelloides LA with respect to structure-activity relationships. Even though they have relatively short acyl chains and unsaturated acyl residue (16:1), the bisphosphorylated, hexaacylated, and asymmetric forms of P. shigelloides LA represent a highly immunostimulatory structure. Moreover, similar trends in in vitro activity of LA I-IV were observed for human and murine macrophages, what was generally in agreement with previously published data. An interaction between LA/LPS and TLR4/ MD-2 complex is species-specific, where depending on species (e.g., human or murine MD-2) the same LA modification may elicit opposed host response. Moreover, several key differences in the amino acid sequences of human and murine TLR4/MD-2 receptor complexes have been shown to alter their ability to recognize different types of LA (27). According to the literature, most important structural features of LA for recognition by the TLR4 complex are acyl chains number and pattern (symmetric vs. asymmetric), and the presence of very short or long acyl chains (27). Five of the acyl chains are buried in a hydrophobic cavity of the human TLR4 adaptor molecule, MD-2, while the sixth chain mediates TLR4 binding and heterodimerization, and subsequent intracellular signaling (27,51,52). In general, different behavior of human and murine cells was observed for tetraacylated forms. We have shown that in laboratory conditions P. shigelloides does not synthesize penta-or tetraacylated forms. By searching for the diversity of P. shigelloides LA, we have shown that this species is able to utilize some LA modification systems that might be useful to evade the immune response, for example, palmitoylation and PEtn substitution. Even though PEtn substitution does not influence in vitro cytokine induction of isolated LA, it may increase resistance of the bacteria against CAMPs. Heptaacylated forms may promote bacterial survival by decreasing cytokine production. Thus, natural P. shigelloides LA heterogeneity may be prominent for the immune system modulation and pathogenesis, since this species was shown to Effect of P. Shigelloides LA on Cytokines Level Frontiers in Immunology | www.frontiersin.org December 2017 | Volume 8 | Article 1741 reFerences be able to adhere, internalize and multiply within mammalian cells in vitro (16,53