This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
Neisseria meningitidis is a gram negative bacterium, which can cause sepsis if it gains access to the bloodstream. This systemic inflammatory state can be rapidly fatal and thus the search for the cells (and mechanisms) involved and for therapeutic strategies has gained much interest. Lipooligosaccharide (LOS), a constituent of the outer membrane of the bacteria, is a key mediator of sepsis. By engaging toll like receptor 4 on the surface of cells of the immune system, it ultimately leads to production of proinflammatory cytokines such as Tumour Necrosis Factor α. Exessive levels of these cytokines contribute to the development of sepsis. Monocytes are the main producers of TNFα in the blood and have been divided into two subsets: proinflammatory and classical. We investigate here the TNFα response of monocytes to LOS treatment. Our results showed that the proinflammatory monocyte subset produced over two fold higher levels of TNFα than the classical monocytes. This may indicate that this subset plays an important role in the excessive production of TNFα seen in meningococcal sepsis, a possiblity that in vivo work may confirm.
We explore a method of detoxification of LOS which involves conversion of the lipid A component from a diphosphoryl- to a monophosphoryl- form by alkaline phosphatase. This was able to substantially reduce (by about 60%) the TNFα production by the proinflammatory monocyte subset in two experiments out of eight. It may be that this effect is donor dependent, a possibility that further work may uncover. Perhaps certain individuals do not develop sepsis as endogenous alklaine phosphatase is able to sufficiently dephosphorylate the LOS.
Neisseria meningitidis is the leading cause of rapidly fatal sepsis (1). Sepsis is a systemic inflammatory state that occurs due to the presence of high levels of bacteria in the bloodstream. This gram negative bacterium exists as a nasopharyngeal commensal in around 10% of the population during periods of endemic disease (2). If a pathogenic strain gains access to the submucosa and subsequently the bloodstream then sepsis may occur. Factors which may contribute to this include bacteria virulence (such as the presence of a capsule: a polysaccharide structure attached to the outer membrane which provides protection from the immune system) and the immune status of the individual (such as whether the individual has or has recently had a viral infection) (3).
Figure 1: Schematic of a section of the cell wall of Neisseria menigitidis. From Tzeng and Stephens (1)
The endotoxin of gram negative bacteria is central to the events which can lead to sepsis. Endotoxin is present in the outer membrane of the bacteria (figure 1) and during growth and lysis, is released as outer membrane vesicles. The endotoxin of N. meningitidis, lipooligosaccharide (LOS), consists of a lipid A moiety and a core polysaccharide. It lacks the O polysaccharide which is present in lipopolysaccharide (LPS), another form of endotoxin. The lipid A moiety (figure 2b) is the component which confers toxicity. It interacts with the Toll-Like Receptor 4 (TLR4)/Myeloid Differentiation factor-2 (MD-2)/Cluster of Differentiation 14 (CD14) complex on the surface of monocytes, macrophages, neutrophils and endothelial cells (4). This ultimately leads to the production of pro-inflammatory cytokines such as Tumour Necrosis Factor α (TNFα), Interleukin 1 β (IL-1β) and Interleukin 6 (IL-6) by the cell.
TNFα has been described as a 'master cytokine' (5). Among other roles it is able to lead to the proliferation and migration of cells as well as regulating the production of other cytokines by those cells. Although initially, through these effects, TNFα production is beneficial in protecting the body against infection by N. meningitidis, excessive TNFα levels in the blood have been found to correlate with the outcome of meningococcal disease (6). TNFα can excessively activate the blood coagulation cascade which can lead to a life threatening phenomenon called Disseminated Intravascular Coagulation (DIC). This involves formation of microthrombi, and can lead to organ failure (7). DIC also contributes to shock, the other hallmark of sepsis.
Monocytes are the principal cellular source of TNFα in the blood (7) and have been previously shown to produce TNFα in response to LOS (8). Monocytes have been divided into two main subsets in humans: Classical (CD14++CD16-HLADR+) and Pro-inflammatory (CD14+CD16++HLADR++) (5). The pro-inflammatory monocytes are more mature and make up around 10% of the monocyte population. They were characterized by the production of large quantities of TNFα, little Interleukin 10 (IL-10) (9) and by the fact that their numbers increase in infection. Belge et al. (5) showed that pro-inflammatory monocytes produced 3 fold higher TNFα levels in response to LPS as compared with classical monocytes. For pam3csk4, a ligand for Toll-Like Receptor 1 (TLR1) and Toll-Like Receptor 2 (TLR2), they showed that the pro-inflammatory monocytes produce 10 fold higher TNFα levels than classical monocytes. However, the results of Frankenberger et al. (9) indicated that the two monocyte subsets produce comparable levels of TNFα in response to LPS.
Figure 2: (a) Diagram of LPS: composed of lipid A, core polysaccharide and O-antigen. Adapted from Miller et al. (10). (B): The lipid A component of Neisseria meningitidis LOS. Circled is the phosphate which is predicted to be cleaved by AP, if the mechanism is the same as with the lipid A of LPS. Adapted from Rietschel et al. (11).
Endotoxin has been proposed as a natural substrate of a phosphate transferase called Alkaline Phosphatase (AP), as a host defence strategy (12). This enzyme is present in many tissues of the body (13). AP has been shown to remove a phosphate from the lipid A component of Escherichia coli and Salmonella minnesota endotoxin (12). It has been demonstrated that up to 30% of the LPS molecules were dephosphorylated. The two phosphate groups on the lipid A molecule have been shown to be important for the toxicity of lipid A (11). The two phosphate groups are thought to contribute, along with the acyl chains of lipid A, to binding to MD-2. When the lipid A acyl chains (shown in figure 2) intercalate into the hydrophobic core of the β sandwich of MD-2, this interaction is stabilized by ionic interactions between the negatively charged phosphate groups of lipid A and basic residues of MD-2 (14).
Figure 3: The homodimerization of two TLR4 molecules and the resulting MyD88-MAL and TRIF-TRAM signalling pathways (15).
Upon engagement of the TLR4 complex the MyD88-MAL and the TRIF-TRAM adapter signalling pathways and activated (figure 3). Monophosphoryl lipid A (MPLA) from S. minnesota has only one phosphate in its lipid A moiety and has been shown as a result to confer a bias towards TRIF-TRAM signalling and away from MyD88-MAL signalling (16). The MyD88-MAL pathway is much more pro-inflammatory than the TRIF-TRAM as this signalling pathway leads to the rapid activation of Nuclear Factor κB (NFκb) and hence expression of pro-inflammatory cytokines. Therefore AP has been proposed as a therapy for sepsis.
Among other work, a study has shown that AP is able to reduce E. coli mediated sepsis in a mouse model. (17). Therefore AP may be able to dephosphorylate LOS and reduce LOS induced sepsis. To date no results have been published in terms of treatment of N. meningitidis LOS with AP. Previous therapies for sepsis have involved neutralization of the pro-inflammatory mediators or endotoxin using antibodies. However, most of these have been ineffective in patients (17). Dephosphorylation of LOS as a therapeutic strategy is attractive as the result is a low toxicity form of LOS, but still with useful immunostimulative properties.
We hypothesise that higher TNFα levels will be produced by the pro-inflammatory monocyte subset than the classical monocyte subset in response to LOS from N. meningitidis. We also hypothesise that AP treatment of LOS will reduce this TNFα production. To test this hypothesis we stimulated whole peripheral blood with LOS and measured the TNFα levels produced by the two subsets. We also looked at the effect of pretreating the LOS with AP on the level of TNFα produced.
Materials and Methods
The hot phenol method, as described by Prendergast et al. (18) was used to extract LOS from Neisseria meningitidis strain MC58. N. meningitidis was grown to confluence on six chocolated horse blood agar plates (Oxoid). Bacteria were harvested into 12ml PBS (2ml per plate). They were washed 3 times with 9 ml PBS before finally being resuspended in 9ml dH20. 0.75ml aliquots were mixed with 0.75ml liquid phenol (90%; Sigma, preheated to 65°C). Samples were vortexed for 1 min and then incubated at 65°C for 10 min, with intermittent vortexing. The samples were then cooled on ice and then centrifuged using a benchtop centrifuge for 3 min at 14,000rpm. The supernatant was transferred to a fresh tube and mixed with 0.75ml of Di-ethyl-Ether to remove residual phenol. Samples were mixed, centrifuged for 3 min at 14,000rpm and the top layer was discarded. This was repeated three times. Samples were left with the lid open overnight to allow evaporation of residual ether. Benzonase nuclease (Novagen 70746-3) was used at a concentration of 25 Units/µl. The samples were then incubated for 1 hr at 60°C with Proteinase K (20µg/ml Company) to remove the benzonase nuclease. Samples were heated at 100°C for 3 min and stored at -20°C.
Purpald assay for quantification of LOS
This method is based on that of Lee and Tsai (19) and compares the unknown sample against known concentrations of E. coli LPS (Sigma; L2654). 50µl of LOS and 50 µl of LPS at various concentrations were placed in the wells of a 96-well microtitre plate.
50µl of 32mM sodium metaperiodate (NaIO4; 0.2M NaOH) was added to each sample and the plate was incubated in the dark for 25 min. 50µl of 136mM purpald reagent (in 2M NaOH) was added to each sample and the plate was then incubated in the dark for 20 min. 50µl of 64mM NaIO4 (in 0.2M NaOH) was added to each sample and the plate was incubated in the dark for 20 min. Finally, 20µl of isopropanol was added to each sample before reading at 540nm.
Extracted LOS was run out alongside appropriate control samples on 14% acrylamide gels. 1x SDS page sample buffer (0.625M Tris-HCl; pH 6.8, 5% SDS, 25% glycerol, 12.5% β-mercaptoethanol, 0.002% bromophenol blue, made up to 25ml with dH2O) was used. The running buffer used was 25mM Tris, 250mM glycine, 0.1 % SDS. Gels were run at 60 volts for 40min.
Gel was stained with Simply Blue SafeStain (Invitrogen LC6060) for 1h. The gel was washed 3 times for 5 min in dH2O before being photographed.
The silver staining method was based on that of Tsai and Frasch (20). The gel was fixed by incubating overnight in a solution of 40% (v/v) ethanol and 5% (v/v) glacial acetic acid. The gel was then transferred to an oxidative solution (0.7% w/v periodic acid, 40% (v/v) ethanol and 5% (v/v) glacial acetic acid) for 5 min before being washed 3 times for 10 min in distilled water. This wash stage was repeated three more times. The gel was then transferred into a pre-staining solution (28ml 0.1M NaOH, 2ml NH4OH and 115ml distilled water), 5 ml of freshly made 20% (w/v) AgNO3 was gradually added and the gel was incubated for 10 min to allow staining. The gel was then washed twice for 5 min in distilled water to remove excess stain before being transferred to developing solution (10mg citric acid and 100µl formaldehyde in 200ml distilled water). The gel was incubated in this solution in the dark until bands could be visualized. The developing reaction was halted by rinsing the gel several times in dH2O.
1% agarose gel: agarose (Invitrogen 15510-019) in Tris Acetate EDTA (TAE) buffer (Tris base 100mM, acetic acid 50mM and EDTA 2.5mM) with SYBR safe DNA gel stain (Invitrogen S33102). The TAE buffer was used as the running buffer. The loading buffer used was 0.005% bromophenol blue in 50% glycerol. Direct Load PCR 100bp Low Ladder (Sigma D3687) was used. The samples were run on the gel at 100 volts for 40 min.
Whole peripheral blood was used as done by Belge et al. (5). Firstly, 600µl blood per condition was washed with 600µl RPMI-1640 medium (Sigma, R0883) with 10% Fetal Calf Serum (FCS) (Autogen Bioclear) (R10 medium) in sterile FACS tubes. The tubes were centrifuged for 5 min at 300g. The supernatant was then discarded and 600µl R10 medium was added to each pellet.
Quantofix® phosphate test kit (Sigma Fluka 37210) or an assay from Katewa et al. (22) were used to determine LPS LOS dephosphorylation. Alkaline phosphatase was used at a concentration of 1 µg/ml and LPS and LOS were used at a concentration of 10 µg/ml. LPS and LOS were diluted into 240µl AP buffer or Tris-HCl. Then 80µl of 3N H2SO4, 40µl of 2.5% (w/v) ammonium molybdate tetrahydrate (prepared in 3 N H2SO4) (Acros Organics 12054-85-2) and 40µl of reducing reagent were sequentially added. The reducing reagent was composed of Hydrazine Sulfate (20mg/ml, Acros Organics 10034-93-2) and Ascorbic acid (20mg/ml, Acros Organics 134-03-2).
Determination of alkaline phosphatase enzymatic activity
Alkaline phosphatase from calf intestine and human placenta were used. p-nitrophenyl phosphate (pNPP) (1mg/ml, Sigma N2640) was used as a substrate for alkaline phosphatase. The optical density was measured at 405nm.
Alkaline phosphatase treatment
LPS/LOS dephosphorylation was done in alkaline phosphatase buffer (50ml Tris-HCl; pH 9.5, 10ml 5M NaCl, 0.5ml 1M MgCl2 x 6H2O) or Tris-HCl buffer (0.2mmol/L; pH 7.5). Calf intestinal mucosa alkaline phosphatase (1 µg/ml Fluka 79390) or human placental alkaline phosphatase (can you use AP as an abbreviation) (1.5 Units/ml, Sigma P3895) and LPS/LOS were incubated for 1 hr at RT and alkaline phosphatase was then heat-inactivated at 90°C for 15 min.
Cells were treated with Neisseria meningitidis lipooligosaccharide (LOS) (100ng/ml) or one of the two controls: Salmonella minnesota R595 lipopolysaccharide (LPS) (100ng/ml, Sigma L6261 and Pam3Cys-Ser-(Lys)4âˆ™3HCl (10µg/ml, Alexis L23978/f). Salmonella minnesota R595 ultra pure lipopolysaccharide (LPS) (110ng/ml, Calbiochem 437628) was also used in one experiment. Calf intestinal mucosa alkaline phosphatase (1 µg/ml Fluka 79390) or human placental alkaline phosphatase (1.5 Units/ml, Sigma P3895) and LPS/LOS were incubated together for 1 hr at RT and after this time the alkaline phosphatase was inactivated. Alkaline phosphatase was inactivated by heating at 90°C for 15 min. Alkaline phosphatase buffer (50ml Tris-HCl; pH 9.5, 10ml 5M NaCl, 0.5ml 1M MgCl2 x 6H2O. QSF 500ml) or Tris-HCl buffer (0.2mmol/L; pH 7.5) was used to dilute the alkaline phosphatase. Alkaline phosphatase concentrations given are the concentration at which LPS and LOS are treated at. Cells were incubated at 37°C for 5 hrs after treatment. For intracellular staining, Brefeldin A (BFA) readymade solution (10µg/ml, Sigma B5936-200UL) was added to cells before incubation. In some experiments Levamisole/ (-)-Tetramisole hydrochloride (1mM, Sigma L9756) was added to cells 30 min before BFA treatment.
1ml Dulbecco's phosphate buffered saline (PBS) (Sigma D8537) was added to the cells after incubation and they were centrifuged for 5 min at 300g. The supernatant was then discarded and 2ml red blood cell lysing buffer (Sigma R7757, 8.3g/L ammonium chloride in 0.01M Tris-HCl buffer; pH 7.5±0.2) was added to the pellets. This was mixed for 1 min. Cells were washed once with 1ml PBS before fixation.
Membrane staining was done using anti-CD14-FITC (Beckman Coulter, Clone: RMO52), anti-CD16-PE (Beckman Coulter, clone: 3G8) and anti-HLA-DR-PC5 (Beckman Coulter, Clone: Immu-357). The antibodies were diluted at a ratio of 1 in 10 in PBA (Bovine Serum Albumin 0.5%, Azide 0.1% in PBS). The cells were incubated at 4 °C for 45 min and then washed in 1ml PBS. The cells were fixed overnight in 500µl fixing solution (0.5% formaldehyde) and kept at 4 °C.
For intracellular staining, 500µl permeabilizing solution (1ml PBS, 0.1% BSA, 0.1% Saponin) was added to each tube and they were incubated for 5 min at RT. The cells were centrifuged and stained with anti-TNFα-PE (Beckman Coulter, clone: IPM2) (1 in 10) in 100µl of permeabilizing solution. Cells were incubated at 4°C for 45 min after staining. The cells were washed twice with permeabilization solution and 500µl fixing solution (0.5% formaldehyde in PBS) was added to each tube before flow cytometric analysis. Initially isotype controls were used to set up the protocol. Flow cytometry was performed using a Coulter FC 500 flow cytometer (Beckman Coulter) and 300000 events were collected for each sample. Dead cells were excluded by forward and side scatter characteristics.
Measurement of cytokine levels in cell supernatants
A cytometric bead array (CBA) assay from BenderMed Systems was performed according to the manufacturer's instructions. A Human basic kit (BMS8420FF) was used along with an IL-8 simplex kit (BMS8204FF), an IL-6 simplex kit (BMS8213FF), an IL-1β simplex kit (BMS8224FF) and a TNFα simplex kit (BMS8224FF).
For the LPS, pam3csk4 and LOS results (figure 4c) parametric one way Anova was performed. The error bars show the standard error of the mean.
Upon extraction, quantification of LOS referring to a E. coli LPS standard curve was carried out. From the purpald assay, the LOS concentration of the preparation was calculated to be 1.645 mg/ml (see figure 4). (How do you getthis value?)
Figure 4: E. coli LPS standard curve for quantification of LOS. A purpald assay, for quantification of endotoxin, was performed on increasing concentrations of LPS and the wavelengths recorded. The purpald assay involves oxidation of unsubstituted terminal vicinal glycol groups of sugar residues such as heptose in LPS. This then yields quantitative formaldehyde measurable by purpald reagent. This curve was used to determine the concentration of the LOS which had been extracted from the bacteria via the hot phenol method.
We then confirmed the presence of LOS in our sample. In Figure 5.1 a silver stained band can be clearly seen in the LOS sample lane. This shows that the preparation does indeed contain LOS as silver staining is dependent on the presence of periodate-sensitive cis-glycols in the core and side chain polysaccharide of the LPS molecules (Kropinski et al, 1986). The LOS sample was prepared so as to contain 8µg, via the purpald assay calculations. However it is close in intensity to the E. coli LPS band. This suggests that there is a ~4 fold overestimation of the concentration of LOS by the Purpald assay.
We then assessed the purity of the LOS extract. Initially we looked for the presence of protein contamination e.g. lipopeptides. This was done by running the sample on a gel and staining with SimplyBlue Safestain. In figure 5.2 we can see no protein bands in the LOS sample lane. This indicates that there is no protein contamination in the sample. However, there may be trace amounts of protein that the stain is not sensitive enough to show.
Figure 5: 14% acrylamide gels stained for protein and endotoxin. Lane 1: LOS 8µg (based on purpald assay results). Lane 2: N. meningitidis whole cell lysate 10µl. Lane 3: E. coli LPS 1 µg. Lane 4: ColourPlus Protein Marker 10 µl. The LPS and the whole cell lysate were used as a control for the presence of LOS. The whole cell lysate was used also as a protein control.
We then sought to determine possible contaminations by nucleic acid. Initial nanodrop spectrophotometer readings suggested that there was nucleic acid contamination in the LOS sample. The initial nucleic acid reading was 420.00ng/μl and the reading after benzonase nuclease treatment was 430.48 ng/μl. This suggests that there is nucleic acid contamination that benzonase is not able to remove. We then decided to visualize on an agarose electrophoresis the possible contamination. Figure 6 suggests also that there is nucleic acid contamination of the LOS samples. We observed the presence of 3 bands with the LOS sample, only one with the LPS. Upon benzonase nuclease treatment, the bands appear weaker, suggesting that benzonase nuclease is able to partially remove this contamination. This contradicts the nanodrop spectrophotometer results.
Figure 6: 1% agarose gel stained for nucleic acids. This gel contains both 15 µg and 40 µg of LOS, benzonase treated LOS and LPS. Ladder used was Direct Load PCR 100bp Low Ladder. Gel was stained with SYBR Stain and visualised using a UV transilluminator.
TNFα production by human monocyte subsets in response to LOS and controls
Now that we had obtained a LOS preparation, we could look at the TNFα response by the two monocyte subsets. We also looked at the response to LPS and Pam3csk4, as the TNFα response by the subsets to these bacterial components has been previously characterized.
figure 7a shows that, the CD14+HLADR++ monocytes had around 7 fold higher CD16 levels than the CD14++HLADR+ monocytes. This was calculated from the median fluorescence intensity (MFI) of PE, to which the anti-CD16 antibody is coupled. Also 66.2% of these cells were CD16 positive whereas 2.3% of CD14++HLADR+ cells were. This indicates that our gating, which was done via levels of CD14 and HLA-DR, is reliable. Then, Figure 7b and 7c show the TNF production after LPS, LOS or Pam3csk4 treatment.
We observed, as expected, an upregulation of TNF production by both monocyte populations in response to LOS, LPS and pam3csk4. The results for LPS and LOS for the proinflammatory monocytes were found to be significant (p<0.001) in comparison to the results from the unstimulated cells. We can see from this graph that the proinflammatory monocytes consistently produced around 2 fold more TNFα than the classical monocytes in response to LPS and LOS (give some values). In both cases this was found to be statistically significant with a p value of under 0.001. The average MFI for the pro-inflammatory monocytes in response to pam3csk4 was around 9 fold higher than for the classical monocytes. (p ≤0.05). Overall, a slightly higher MFI was observed upon stimulation with LPS than LOS, but this was not found to be statistically significant.
Figure 7: Flow cytometric data showing gating of the two monocyte subsets and TNFα production in response to LOS and controls. (a) results from one experiment to provide an example of how the two monocyte subsets were gated. The cells were stained for HLA-DR and CD14 and the two monocyte subsets were gated using levels of these two markers as a guide. (b) levels of TNFα produced when the cells were treated with LPS, pam3csk4 and LOS. (c) average TNFα levels from 6 experiments. Unstimulated monocytes were used as a negative control. ***p<0.001, *p<0.05, parametric one-way Anova. MFI: Median Fluorescence Intensity
Determination of LOS and LPS dephosphorylation
Poelstra et al. (13) showed that cAP can dephosphorylate LPS (as did Verweij et al. (17) for plAP). The lipid A moiety is similar between LPS and LOS but we did want to confirm that AP can dephosphorylate LOS. Therefore we used two phosphate determination asssays to detect released phosphate. Using the Quantofix kit (figure 12a), both the LPS alone and cAP treated LPS were shown to contain 50mg/l phosphate. Both the LOS alone and AP treated LOS were shown to contain 3mg/l phosphate. The nanodrop spectrophotometer results (figure 12b) show that this assay has detected similar levels of phosphate in the LPS and LOS samples as in the AP treated LPS and LOS samples. As well for the inactivated AP treated LPS and LOS samples. The LOS was found to contain just over 5 fold less phosphate than the LPS. This experiment was repeated in AP buffer rather than Tris-HCl, with similar results (not shown)
Figure 12: The results of two assays for the detection of LOS and LPS dephosphorylation by calf intestinal AP. The image in (a) shows the scale and dipsticks from the Quantofix kit for phosphate determination. The graph in (b) contains the results of the phosphate determination assay adapted from Katewa et al. (21). The free phosphate reacts with ammonium molybdate to produce phosphomolybdate. This is then reduced by ascorbic acid and hydrazine sulfate to form a soluble blue product. The optical density of this blue colour was measured using a nanodrop spectrophotometer at 750nm. In this experiment the cAP was diluted in Tris-HCl buffer.
The effect of alkaline phosphatase treatment of LOS and LPS on TNFα production by the two monocyte subsets
Since the phosphate determination assays didn't work, we used the pNPP substrate as a method of confirming the activity of the alkaline phosphatase. However this can only give us an indication of whether the LOS is being dephosphorylated. Previous work (13, 17) has shown that alkaline phosphatase can dephosphorylate LPS. From the pNPP experiments with AP (figure 8a, 9a and 10a) we could see that the AP enzymes had high activity at under 0.5µg/ml for calf intestinal AP (cAP) and under 0.75 Units/ml for human placental AP (plAP). Upon these findings, we decided to use double these concentrations, to be sure of sufficient enzyme activity. These experiments also indicated that 15 minutes incubation at 90°C was enough to heat inactivate the two kinds of AP.
We then treated the LOS and LPS with alkaline phosphatase and measured the TNFα levels produced by the two subsets. We have concentrated on the pro-inflammatory monocytes, as this is the subset which appears to be principle in the excessive TNFα production seen in meningococcal sepsis. The median fluorescence intensity (MFI) gives us an indication of the TNFα production. When cAP was used at a concentration of 1µg/ml in AP buffer (figure 8b and c), there was a slight reduction in MFI by the pro-inflammatory monocytes in response to both LPS and LOS. For donor 1 (figure 8b) this was by 28% for LPS and 25% for LOS. In the case of donor 2 (figure 8c) there was a decrease in MFI of 16% in response to both LPS and LOS. In all cases the TNFα produced in response to the heat inactivated LPS/LOS is used as 100%. cAP was then used at a concentration of 10µg/ml (figure 8d and e). In one donor (figure 8d), there was a 42% reduction in TNFα produced by the pro-inflammatory monocytes upon stimulation with AP treated LPS. Similarly, for LOS there was a 60% reduction in TNFα levels upon AP treatment. However, when this experiment was repeated using another donor (figure 8e), no reduction in MFI upon AP treatment of LPS and LOS was observed.
Experiments were then done using cAP in Tris-HCl (pH 7.5) buffer rather than AP buffer (pH 9.5) ast Poelstra et al. (12) state that dephosphorylation of endotoxin occurs at pH 7.5. When cAP was used at a concentration of 1µg/ml and 10µg/ml (figures 9b and 9c), there was little difference in TNFα production by the pro-inflammatory monocytes upon AP treatment of LPS and LOS. When plAP in Tris-HCl buffer was tried, a reduction (40%) was seen for LOS but no difference was seen for LPS for one donor (figure 10b). In the other donor (figure 10c), a reduction was seen for LOS (16%) and a slight reduction was seen for LPS (10%).
(a)In figure 10a, levamisole was able to reduce the alkaline phosphatase activity by 83%, if we consider 100% to be the optical density when 0.75 Units/ml untreated human placental alkaline phosphatase was used. Levamisole was added to the cells before stimulation to attenuate any endogenous leukocyte AP activity in the cells as this could be masking an effect of the exogenous AP. This isoenzyme of alklaine phosphatase is present on the surface of polymorphonuclear cells such as neutrophils (12), but has not been found to be present on the surface of monocytes (22). In figure 10b it seems that when cells are pre-treated with levamisole, the MFI in response to stimulation was in the same range as the levels seen from experiments where the cells were not pre-treated. However, in the repeat of this experiment using blood from another donor (10c), the MFI upon stimulation were generally higher than has been seen for the other experiments.
Figure 8: TNFα production by human monocyte subsets in response to calf alkaline phosphatase treated LOS and LPS (in alkaline phosphatase buffer). (a) Measure of AP activity. Optical density at 405 nm of the soluble yellow product produced when pNPP is dephosphorylated by calf alkaline phosphatase (cAP). (b) TNFα levels produced by the two monocyte populations in response to 1 µg/ml cAP treated LPS and LOS are shown. TNFα levels produced by the monocytes in response to inactivated cAP treated LPS and LOS were used as a control. TNFα readings from alkaline phosphatase and inactivated alkaline phosphatase treated monocytes were used as negative controls. (c) A repeat of this experiment, using another donor. (d) calf intestinal alkaline phosphatase was used at 10 µg/ml in alkaline phosphatase buffer. (e) a repeat of the experiment shown in (d), using another donor.
Figure 9: TNFα production by human monocyte subsets in response to calf alkaline phosphatase treated LOS and LPS (in Tris-HCl buffer). (a) Optical density at 405 nm of the soluble yellow product produced when pNPP is dephosphorylated by cAP as well as when pNPP was treated with heat inactivated alkaline phosphatase. (b) the TNFα levels produced by the two monocyte populations in response to 1 µg/ml calf intestinal alkaline phosphatase treated LPS and LOS are shown. (c) the experiment was repeated using another donor but the calf alkaline phosphatase was used at a concentration of 10 µg/ml.
Figure 10: TNFα production by levamisole treated human monocyte subsets in response to human alkaline phosphatase treated LOS and LPS (in Tris-HCl buffer). (a) Optical density of the soluble yellow product produced when pNPP is dephosphorylated by human placental alkaline phosphatase (p1AP). This was read at 405nm using a plate reader. It also shows the optical density when pNPP was treated with heat inactivated alkaline phosphatase. In this experiment the alkaline phosphatase was diluted in Tris-HCl buffer. In (b) the TNFα levels produced by the two monocyte populations in response to 1 µg/ml human placental alkaline phosphatase are shown. The human placental alkaline phosphatase was diluted in Tris-HCl buffer. In (c) the experiment was repeated using another donor.
Use of a CBA assay to determine the effect of AP on TNFα, IL-1β, Il-6 and IL-8 production of leukocytes in response to LOS and LPS.
(a)We then wanted to complete the data by looking at TNFα levels in the supernatant of the whole blood. The results of figure 11 indicate again that LOS stimulation leads to high levels of TNFα production and that neither cAP nor plAP were able to reduce the levels of TNFα, IL-1β, IL-6 or IL-8 in the supernatant from whole peripheral blood from two donors.
(d)Can you add IL-6, IL-8… on the dot blot
Figure 11: Cytometric bead array (CBA) assay results. For each cytokine, a fluorescent bead population coated with specific antibodies was used to bind that cytokine. A biotin labelled secondary antibody, also specific for that cytokine then binds the complex. PE labelled streptavidin is added and binds to the biotin. (a) the two bead populations from the CBA assay. This example is the supernatant of the cells treated with LOS from donor 1. The two sets of bead populations can be differentiated by size. The bead populations within region 0 are gated upon in (b) and the populations in region 1 are gated upon in (c). The two bead populations in each region can be differentiated by their levels of fluorescence, measured by FL4. The levels of each cytokine are measured by FL2, which detects the levels of streptavidin-PE attached to the beads. In (b) we can see high levels of TNF (top) and IL-1beta (bottom). In (c) we can see even higher levels of IL-6 (top) and IL-8 (bottom). The assay was done with two donors and for each both calf intestinal (cAP) and human placental alkaline phosphatase (p1AP) were used. Average cytokine concentrations from the supernatants of the cells from the two donors are shown when human placental (d) or calf (e) alkaline phosphatase was used. This was done in Tris-HCl. For (e) we have only one repeat for inactivated cAP treated LPS.
Stimulation of cells with benzonase treated LOS and ultra pure LPS
To investigate the effect of nucleic acid contamination of the LOS preparation we stimulated cells with benzonase treated LOS and ultrapure LPS. Upon stimulation of cells with benzonase nuclease treated LOS, the MFI was 25% lower than when cells were stimulated with untreated LOS (data not shown). Stimulation with ultrapure LPS showed a 10% reduction in MFI in comparison to LPS. The ultra pure LPS used was from the same strain of S. Minnesota as the LPS used. This was done investigate the effect of any potential contamination in the LPS. In the one experiment where the benzonase treated LOS was treated with plAP, a 44% reduction was seen in the MFI, similar to the 40% reduction upon p1AP treatment of LOS (figure 10b) seen in the same experiment. Upon plAP treatment of ultrapure LPS, a reduction in MFI of 10% was seen.
This study involved stimulation of human monocyte subsets with LOS from Neisseria meningitidis. The purpald assay used E. coli LPS as the standard from which the cocnentration of the extracted LOS was estimated. As LPS and LOS are different in chemical structure, it is likely that there is some degree of inaccuracy. If indeed the concentration of LOS was overestimated it is not a big problem as we have an experiment with different doses of LOS with the same secretion of TNF. ?
The pro-inflammatory monocytes were characterized upon their propensity to produce high levels of TNFα. Indeed in figure 4 we observed that the pro-inflammatory monocytes consistently produce over 2 fold more TNFα than the classical monocytes, in response to LOS, LPS and pam3csk4. This was found to be statistically significant. Belge et al. (5) described a 3 fold difference so our results here were similar to theirs. It would indeed appear that the pro-inflammatory subset of monocytes plays a role in the excessive production of TNFα in meningococcal sepsis. However in vivo work would need to be done to confirm this. Experiments could be done for example with a mouse model of sepsis, similar to that used by Verweij et al. (17). The reason why pro-inflammatory monocytes produce more TNFα than classical monocytes is not clearly understood. The pro-inflammatory monocytes express less CD14 on their surface and Belge et al (5) showed that the two monocyte subsets contained similar levels of TLR4 messenger Ribonucleic Acid (mRNA). Therefore it may be post receptor mediated events that cause pro-inflammatory monocytes to produce more TNFα (and less IL-10). However, Skinner et al. (23) found that TLR4 was found to be expressed 2.5x more on CD14dimCD16+ monocytes than on CD14+CD16- monocytes, and therefore this may have an influence.
Stimulation of the pro-inflammatory monocytes with LPS and LOS led them to produce 3 fold higher TNFα levels than pam3csk4, a synthetic ligand of TLR1 and TLR2. Indeed Belge et al. (5) showed 5 fold higher TNFα levels in response to LPS than pam3csk4. The average MFI for the pro-inflammatory monocytes in response to pam3csk4 was around 9 fold higher than for the classical monocytes, which is in accordance with Belge et al. (5) who reported a 10 fold difference. Belge et al. (5) suggest that part of the reason for this is that the pro-inflammatory monocytes express two fold higher levels of TLR2 on their surface than the classical monocytes do.
These findings contrast with those of Frankenberger et al. (9) who observed that comparable levels of mRNA encoding TNFα could be found in the two subsets upon LPS stimulation. Belge et al. (5) suggest that a possible reason that they had different results than Frankenberger et al. (9) is that when whole peripheral blood is used there is an absence of interfering signals. We used whole blood for this reason. In preliminary experiments we used PBMCs (data not shown) and this did give results similar to those found by Frankenberger et al. (9). Another reason we did not use PBMCs is that monocytes can develop tolerance to endotoxin. Wilson et al. (24) have shown that pre-treatment of monocytes with a low dose of LPS impairs the TNFα response when subsequently stimulated with a higher dose of LPS. In the experiments using PBMCs, the classical monocytes produced as much or more TNFα than the pro-inflammatory monocytes. It may be that the pro-inflammatory monocytes develop tolerance to endotoxin more efficiently or more readily than the classical monocytes. The TNFα production by the pro-inflammatory monocytes in response to LOS was slightly lower than for LPS. However this was not found to be statistically significant.
Calf intestinal AP was used as per Poelstra et al. (12) and human placental AP was used as per Verweij et al. (17). Initially AP buffer was used (pH 9.5) but Poelstra et al. (12) state that dephosphorylation of endotoxin occurs at pH 7.5. Therefore Tris-HCl at pH 7.5 was tried but no dramatic difference was seen in TNFα levels upon cAP treatment of LPS and LOS. Indeed Verweij et al. (17) showed that AP can dephosphorylate endotoxin to the same degree at both pH 7.5 and 9.8. The percentage decrease in MFI produced by the pro-inflammatory monocytes in response to AP treated LOS and LPS in comparison to inactivated AP treated LOS and LPS varied from 13% to 60%. However, in a few cases, the MFI in response to the inactivated AP treated LOS and LPS was higher than in response to the AP treated LOS and LPS. This percentage increase varied from 7% to 26%. Therefore, we decided that we cannot trust the significance of a decrease below around 30%. This leaves the results of figure 8d, when for LPS there was a 42% decrease and for LOS there was a 60% decrease and the results of figure 10b where a decrease of 40% was observed for LOS.
The CBA assay results (figure 8) indicate that AP (both human and calf) treatment of LPS and LOS is not able to reduce the production of TNFα by monocytes. However, the CBA results can only be used as an indication of the cytokine levels produced by monocytes because we are not able to differentiate between cell types in this assay. This assay also suggested that the levels of other pro-inflammatory cytokines: Il-1β, IL-6 and IL8, are not reduced when LPS and LOS are treated with AP.
It may be that the effect is donor dependent. Perhaps exogenous alkaline phosphatase can only reduce TNFα production in individuals who have low endogenous alkaline phosphatase activity/levels. This would be reflected in high TNFα levels being produced by the cells of these individuals. We were not able to observe this trend with just two donors. It may be that these inviduals are more at risk from sepsis for this reason and would benefit from alkaline phosphatase treatment as a therapeutic strategy if infection occurred. In two donors, cells were pretreated with levamisole. One donor (figure 10c) produced higher levels of TNFα than the others, suggesting high endogenous AP activity. As no effect of pIAP was seen this fits in with the idea postulated. More repeats would be needed to check that this was not due to other effects such as donor variation. To investigate this idea of donor variation further, perhaps the experiment should be repeated using blood from the two donors where an effect was seen to see whether similar results would be obtained. This should be repeated with and without pre-treatment of the cells with levamisole. If this theory is correct then the levamisole should have little effect.
The results for AP treatment of LPS are perhaps not as expected due to previous findings. Bentala et al. (25) showed that MPLA does not induce production of TNFα in vitro or in vivo, in comparison to LPS which does. Verweij et al. (17) showed that administration of AP to mice which had been infected with E. coli increased survival. However in this experiment no significant elevation in TNFα levels was detected when E. coli alone or both E. coli and AP were administered. A possible explanation is that it takes time for endotoxin to be released from the surface of the bacteria and TNFα levels were measured only up until 4 hours.
The results of figure 5a, 6a and 7a show that we can rule out the possibility that both batches of APs were defective. We were not able to determine whether LOS was being dephosphorylated by AP as it seems that the Quantofix kit and the Katewa et al. (21) method, despite what is stated in this paper, are able to detect bound phosphate as well as released phosphate. In the future perhaps another assay should be tried, such as that in Chandrarajan et al (26), a paper that was not available to us.
It is important to address the LOS contamination issue. Bacterial DNA contains unmethylated CpG islands, which may be recognized in the endosome of the monocytes by Toll-Like Receptor 9 (TLR9). Signalling through TLR9 would lead to the production of pro-inflammatory cytokines such as TNFα via the MyD88 adapter signalling pathway. Low levels of TLR9 have been found to be present on the cell surface of freshly isolated human monocytes (27), so it is possible that the nucleic acid contaminants are causing some of the TNFα production. Sjolinder et al. (28) have demonstrated that TLR9 plays an important role in defence against N. meningitidis and Sprong et al. (29) have shown that non-LOS components contribute to roughly half of TNFα production by PBMCs, therefore indicating that a control for such components is important. We observed a 25% reduction in MFI when using the benzonase treated LOS, suggesting that 25% of the TNFα levels produced are as a result of the nucleic acid contaminants. This would have to be repeated to draw any solid conclusions. A 44% reduction in MFI was observed upon AP treatment of benzonase nuclease treated LOS. As a similar effect of plAP treatment was observed for the non benzonase treated LOS, we cannot draw conclusions as to the effect of the nucleic acid contamination. However if in other donors the nucleic acid contaminants are stimulating production of high levels of TNFα, this could potentially mask any decrease in TNFα production in response to AP treatment of LOS.
Polymyxin B treatment of LOS was initially trialled as a control for the nucleic acid contamination. This antibiotic has been shown to be able to form a complex with LPS, disrupting its morphology (30). Indeed when the S. minnesota LPS was treated with polymyxin B, there was an average reduction MFI of 90% in comparison to LPS alone (data not shown). However for LOS polymyxin B had little effect. Indeed Baldwin et al. (31) showed that polymyxin B was able to reduce experimental shock in rabbits mediated by E. coli LPS but not N. meningitidis LOS. In the future a LOS deletion mutant strain of N. meningitidis, such as generated by Steeghs et al. (32), could potentially be used as a control for the DNA contamination, and indeed any other non-LOS components.
In conclusion, our results indicate that the proinflammatory monocyte subset may play an important role in the excessive production of TNFα seen in meningococcal sepsis, a possiblity that in vivo work may confirm. Therefore these cells may be important cells to target therapeutically. We cannot confirm whether AP could be used to detoxify LOS in sepsis patients. Its effect may be donor dependent, a possibility that further work may reveal.