Proinflammatory Responses to Lipo-oligosaccharide of Neisseria meningitidis Immunotype Strains in Relation to Virulence and Disease

Abstract

Inflammatory responses to lipo-oligosaccharide (LOS) contribute to the severity of meningococcal disease. Strains that express the L(3,7,9) LOS immunotypes are isolated from the majority of patients, but other immunotypes are isolated predominantly from carriers. Inflammatory responses elicited from a human monocytic cell line (THP-1) that had been pretreated with vitamin D3 (VD3) were compared after stimulation with purified LOSs from standard immunotype strains. The neutralizing effects of normal human serum and serum from mice immunized with strain B:2a:P1.5,2:L3 were compared. LOSs of immunotypes L3, L7, L8, and L9 induced significantly higher levels of tumor necrosis factor-α and interleukin-6, compared with other immunotypes. Normal human serum neutralized the proinflammatory responses to LOSs of all immunotypes tested. Immune mouse serum neutralized inflammatory responses against LOSs from immunotypes with epitopes cross-reactive with L(3,7,9) moieties. Antibodies found in normal human serum and immune mouse serum to the oligosaccharide, core, and lipid A moieties of meningococcal endotoxin contribute to neutralizing activity.

Inflammatory responses to lipo-oligosaccharide (LOS) contribute to the severity of disease caused by Neisseriameningitidis [1–4]. Several studies have shown a close relationship between severity and fatality of meningococcal sepsis with plasma levels of LOS [5] and markers for inflammation such as interleukin (IL)-1 [6], IL-6 [7], IL-8 [8], tumor necrosis factor (TNF)-α [9], and acute phase protein C3 [10]. Although there have been 12 immunotypes of LOS (L1-L12) described, L3, L7, and L9 (hereafter “L[3,7,9]”) are associated with rapid progressive meningitis and septicemia and are isolated from >90% of patients with disease due to serogroup B or C infection. Immunotypes L10–L12 are found exclusively on group A meningococci. Other types are obtained primarily from asymptomatic carriers [11, 12].

Immunotypes L3, L7, and L9 are thought to be similar in their immunochemical structures [13, 14], with immunotype L3 being sialylated by endogenous sialyl transferases. Immunotypes L3 and L7 are found in serogroupsB andC meningococci, and they have similar core components, phospho-ethanolamine (PEA) (1→3) heptose (HepII). Immunotype L9 is expressed on group A strains.

The sialylated phenotype is associated with resistance to complement-mediated killing bymasking the terminal galactose with N-acetylneuraminic acid (sialyl). This mechanism is thought to reduce the recognition of the epitope by anti-LOS antibodies directed against the nonsialylated epitopes [15, 16]. Free or membrane- bound sialyl-L(3,7,9) also up-regulates neutrophil activation markers and results in increased injury of epithelial cell lines [17]. Sialyl-L(3,7,9) phenotypes can evade the complement-mediated bacteriolysis cascade [18]. This phenotype also reduces complement- and anti-LOS antibody-mediated phagocytosis by professional phagocytes [19, 20].

The expression of multiple immunotypes within ameningococcal population is thought to allow the organism to diversify its antigenic structure. Selective pressures due to the presence of antibodies in the host to 1 LOS immunotype allow the strain to express other immunotypes, thus increasing their chance of survival. Sialylation [21] and the expression of the paragloboside gene cluster (Igt ABE) [22] are the main phase-variable phenotypes known.

Meningococcal meningitis and septicemia are exclusively human diseases, and a suitable animal model does not exist that reflects the genetic and environmental factors that contribute to susceptibility to or severity of disease. Inflammatory responses need to be assessed in relation to the genetically controlled variability of the inflammatory responses of individuals to bacterial antigens and toxins [2, 3, 23–25]. Environmental factors such as virus infection [26] and exposure to cigarette smoke [27] can also influence inflammatory responses to bacteria or their antigens. To attempt to minimize the effects of these potential confounding factors, an in vitro model with a human monocytic cell line was developed for the initial screening of cytokine release, induced by LOS preparations from meningococci. The objective was to use the model system to test the hypothesis that the L(3,7,9) LOSs induced higher levels of inflammatory cytokines than other immunotypes.

Materials and Methods

Cell Lines and Culture Conditions

THP-1. The human monocytic cell line THP-1 was obtained from the European Collection of Animal Cell Cultures (ECACC). Cells were grown to 10⁴–10⁶ cells/mL in RPMI 1640 cell culture medium (Sigma) supplemented with fetal calf serum (FCS; 5% vol/ vol; Gibco), L-glutamine (1%, wt/vol; Gibco), penicillin (100 IU/ mL), and streptomycin (200 mg/mL; Gibco) for ⩽18 weeks after establishment of the cell line. The FCS did not contain antibodies against any of the tested bacterial strains as determined by wholecell ELISA (WCE) [28].

L929. The mouse fibroblast cell line L929 was obtained from the ECACC. Cells were grown in 75 cm³ tissue culture flasks (Becton Dickinson) to 70% confluence in growth medium that contained Dulbecco's modified Eagle medium (Sigma) supplemented with FCS (5%, vol/vol; Gibco), L-glutamine (1%, wt/vol; Gibco), penicillin (100 IU/mL), and streptomycin (200 mg/mL; Gibco) at 37°C with 5% CO₂ [23].

Analysis of Cell Surface Antigens

THP-1 cells were induced to express the lipopolysaccharide (LPS) receptor CD14 by incubation with 10−⁷ M vitamin D3 (1.25-dihydroxyvitamin D3 [VD3]; Calbiochem) for 72 h [29, 30]. The primary and secondary antibodies used to detect cell surface antigens in the flow cytometry assay were fluorochrome-conjugated primary antibodies against the major histocompatibility II receptor (CD3/ CD4; TCS Biologicals), complement receptor a chain (CD11b; DAKO), matrix-bound LPS receptor (CD14;DAKO), immunoglobulin (IgG3) high-affinity receptor (CD64; Serotec), fibronectin receptor β chain (CD29; DAKO), and vitronectin receptor β chain (CD51; DAKO). Antibodies bound to the surface of the cells incubated in the presence or absence of VD3 were detected by flow cytometry, as described elsewhere [31].

Extraction of LOS from Meningococcal Strains

The characteristics of the meningococcal immunotype reference are summarized in table 1. All strains were grown for 18 h in 5%(vol/ vol) CO₂ on human blood agar (HBA). The medium contained lysed whole blood (100 mL) obtained from the Scottish National Blood Transfusion Service, special peptone (23 g; Difco), corn starch (1 g; Sigma), NaCl (4.5 g; Sigma), D-glucose (1 g; Sigma), technicalgrade agar (10 g; Oxoid), K₂HPO₄ (4 g; BDH), and KH₂PO₄ (1 g; BDH) in 900 mL of distilled water. Cells were harvested fromplates, washed in sterile pyrogen-free PBS, centrifuged at 1000 g, and resuspended in pyrogen-free distilled water.

Table 1

Lipo-oligosaccharide (LOS) immunotypes of meningococcal reference strains.

The hot phenol-water method described by Hancock and Poxton [32] was used to extract the LOS. The purified LOS contained protein contaminants of <1% (wt/wt), as assessed against a standard of bovine serum albumin (Sigma) [33]. LOS was lyophilized, and the amount of LOS was determined by weight. LOS was resuspended (1μg/mL;wt/vol) in RPMI 1640 medium (Sigma) and filtered through a 0.22-μm membrane filter. Aliquots were stored at −27°C, and 2 samples from each batch were tested for sterility. Escherichia coli (O26:B6) endotoxin was purchased from Sigma.

Induction of Proinflammatory Cytokines

THP-1 cells were incubated for 72 h with 10−⁷M VD3 to induce expression of the CD14 cell surface antigen [29, 30]. Triplicate samples of the differentiated cells were challenged for 6 h in RPMI 1640 medium that contained 5% (vol/vol) FCS as a source of endotoxinbinding protein [34], supplemented with 10-fold dilutions of LOS from the individual meningococcal immunotypes (L2–L11) or E. coli endotoxin (strain 026:B6). A range of concentrations, from 1 pg/mL to 100 ng/mL, was examined in initial studies, and a concentration of 100 pg/mL was used in subsequent experiments. To determine the concentration to be used in the neutralizing assays, a pool of serum from 8 adult donors with no history of meningococcal disease or immune mouse serum was serially diluted and tested by use of WCE for binding to meningococcal immunotype L3. A final dilution of 1:1000 was used in the neutralization experiments. The viability of cells was assessed by use of trypan blue exclusion.

Detection of Cytokines

An ELISA was used to detect IL-6 [23], and a bioassay was used to detect TNF-α [35].

Confocal Microscopy

After flow cytometric analysis, immature and VD3-differentiated THP-1 cells (50 μL), stainedwith fluorochrome-conjugated idiotypic antibodies (control) and conjugated anti-CD14, were placed on a microscope slide (Greiner) and sealed under a cover slip with clear nail polish. Specific staining was confirmed by scanning confocal microscopy (3-μm depth) and phase-contrast microscopy (magnification, ×670; J. Bard, University of Edinburgh).

Immune Mouse Serum

Themeningococcal immunotype strain B:2a:P1.5,2:L3 was grown on HBA and killed by heating for 60 min at 100°C. Bacteria (10⁹ in 100 μL) in adjuvant- and pyrogen-free saline (Sigma) were injected into the tail vein of 36-week-old male BALB/c mice on 3 consecutive days. This was followed by repeated intravenous inoculations with the same dose and batch of antigen at weeks 4, 8, 12, and 16. At week 20, LOSs (100 μL, 100 ng/mL) obtained by hot phenolwater extraction of the immunotype L3 strain were injected. Three days after the final injection, blood was collected aseptically by cardiac puncture, allowed to clot, and centrifuged at 500 g for 15 min at 4°C. The supernatant was collected and diluted in pyrogen-free saline (1:100). The complement was inactivated by heat treatment (56°C for 30 min), and the serum samples were stored in aliquots (1 mL) at −27°C. The production of antibodies was covered by an animal license obtained from the British Home Office. Antibodies to L3 in samples taken from the mice before immunization and at the end of the immunization schedule were assessed by WCE [28].

Statistical Analysis

The mean, SD, and Student's t test were calculated by use of Minitab for the Apple Macintosh (Minitab). To determine whether the data were normally distributed, normal probability plots were used [36]. Regression and analysis of variance showed that cytokine levels were distributed normally. Probability values were calculated with a confidence interval of 5% against the negative control treatedwith PBS only or the E. coli O26:B6 LPS. Two-sided analysis (5% confidence level) was carried out by use of a paired t test for different endotoxin samples.

To assess the inflammatory response of known LOS antigens, samples were grouped, and a 2-sided analysis was performed. First, immunotype strains that expressed the major L3, L7, and L9 immunotypes were grouped and assessed to determine whether the inflammatory response was greater than the response elicited by non-L(3,7,9) immunotypes (L4, L6, L10, and L11) or strains that expressed L(3,7,9) as a minor antigen (L2, L5, and L8) (table 1). Next, immunotypes with PEA (1→3) HepII in their LOS core structure, but that did not express L(3,7,9) as a major antigen, were grouped (L8, L10, and L11) and compared with meningococcal LOS that did not have PEA moieties in the core of the major immunotype (35) [37].

Results

Expression of cell surface antigens. A high proportion of undifferentiated THP-1 cells (70%–99%) bound antibodies to the complement receptor (CD11b), the high-affinity IgG receptor (CD64), and the receptors for fibronectin (CD41) and vitronectin (CD51), both of which are associated with meningococcal adherence and endocytosis. Binding of antibodies to the MHC class II antigen (CD4) or membrane-bound LPS receptor (CD14) was detected on <5% of cells. Incubation of THP-1 cells with VD3 for 72 h significantly enhanced expression of CD14 (4.3%–89.9%; P <05).

The negative control contained differentiated and undifferentiated cells incubated with fluorescein isothiocyanate (FITC)-labeled anti-idiotypic mouse antibodies and did not show any detectable staining. Confocal microscopy confirmed that CD14 was expressed in VD3-differentiated cells, but undifferentiated THP1 did not bind FITC-labeled anti-CD14 antibodies.

>Time course for induction of TNF-α and IL-6 from differentiated THP-1 cells. In 3 independent experiments, TNF-α was detected from 3 h after exposure to LOSs from the L2 and L3 strains. Concentrations ranging from 1 pg/mL to 100 ng/mL were studied (tables 2 and 3). The peak level for all LOSs tested was observed at 6 h at an optimal concentration of 100 pg/mL. IL-6 was detected from 3 h, and the peak level was reached by 6 h. Viability of THP-1 cells (trypan blue exclusion) was <95% for LOSs at concentrations > 1 ng LOS/mL. In subsequent experiments, cells were incubated with 100 pg/mL purified LOS, and samples were taken at 6 h for assessment of TNF-α and IL-6.

Table 2

Time and dose response of tumor necrosis factor (TNF)-α (IU/mL) release from vitamin D3-differentiated THP-1 cells challenged with endotoxin (lipo-oliogosaccharide [LOS]) from meningococcal immunotype L2 or equal volume of PBS (control) (n = 3).

Table 3

Time and dose response of tumor necrosis factor (TNF)-α (IU/mL) release from vitamin D3-differentiated THP-1 cells challenged with endotoxin (lipo-oligosaccharide [LOS]) from meningococcal immunotype L3 or an equal volume of PBS (control) (n = 3).

Effect of LOS immunotype on cytokine levels. In 6 independent experiments, equivalent amounts of endotoxin (100 pg/mL) from E. coli and the meningococcal immunotypes were assessed for their induction of proinflammatory cytokines. E. coli endotoxin induced significantly lower levels of TNF-α and IL-6 than LOSs from each meningococcal immunotype tested (P<003; figures 1A and 2A). Immunotypes L3, L7, L8, and L9 induced significantly higher levels of TNF-α and IL-6, compared with strains that expressed non-L(3,7,9) immunotypes (P<.01). L7 LOSs elicited the highest levels of TNF-α, compared with all other immunotypes tested, including those that expressed L(3,7,9) (P<.01; figure 1A). L3 LOSs elicited TNF-α levels that were significantly higher than those of non-L(3,7,9) strains and L8(3,7,9) (P<.01), but there were no significant differences observed between results for L3 and L9. L9 LOSs elicited TNF-α levels significantly higher than those of non-L(3,7,9) strains and L8(3,7,9) (P<.006), but there were no significant differences observed between results for L3 and L9 (P>.19).

Figure 1

Release of tumor necrosis factor (TNF; IU/mL) from vitamin D3-differentiated THP-1 cells challenged with endotoxin from meningococcal immunotypes (L2-L11) or from Escherichia coli (lipopolysaccharide [LPS]; 100 pg/mL) (A), endotoxins coincubated with pooled human serum (final dilution, 1:1000) (B), and endotoxins coincubated with mouse serum produced by immunization with the meningococcal immunotype L3 (final dilution, 1:1000) (C); n = 6. Error bars, SD. PBS was used as a negative control.

Figure 2

Release of interleukin (IL)-6 (pg/mL) from vitamin D3-differentiated THP-1 cells challenged with endotoxin from meningococcal immunotypes (L2–L11) or from Eschericheria coli (lipopolysaccharide [LPS]; 100 pg/mL) (A), endotoxins coincubated with pooled human serum (final dilution, 1:1000) (B), and endotoxins coincubated with mouse serum produced by immunization with the meningococcal immunotype L3 (final dilution, 1:1000) (C); n = 6. Error bars, SD. PBS was used as a negative control.

LOSs of L7 elicited the highest levels of IL-6, compared with all other immunotypes tested, including those that expressed L(3,7,9) (P <.01; figure 1A). L3 LOSs elicited IL-6 levels significantly higher than those of all non-L (3,7,9) strains tested (P<.006), except L8 (P=.980) or L9 (P=.864). Similar patterns were observed for L9.

Cytokine levels in relation to the presence or absence of the major L(3,7,9) structure. To assess the inflammatory response of known LOS antigens, samples were grouped, and a 2-sided paired analysis was performed, where immunotypes L3, L7, and L9 were grouped to assess the inflammatory responses in comparison with immunotypes that did not express L(3,7,9) as their major LOS antigen (table 1) (L2, L4, L5, L6, L8, L10, and L11). The results with immunotypes that contained L(3,7,9) were significantly higher for TNF-α (P<.01) and IL-6 (P <.01).

Cytokine levels induced by LOS immunotypes in relation to core structure. To assess the role of different core structures, cytokine responses to LOSs from immunotypes known to have PEA (1→3) HepII structure in their LOS were grouped (L3, L7, L8, L9, L10, and L11) and assessed in relation to responses elicited by LOS from strains that do not have the PEA (1→3) HepII core structure (L2, L4, L5, and L6) ("table 1) [37]. Levels of TNF-α (P<.01) and IL-6 (P<.006), induced by LOS structures that contained PEA(1→3) HepII, were significantly higher, compared with those without PEA (1→3) HepII. The analysis was repeated excluding immunotypes that expressed L(3,7,9) as their major LOS antigen. Cytokine levels induced by LOSs of immunotypes with the PEA (1→3) core structure (L8, L10, and L11) had significantly higher responses than LOSs from immunotypes with other core structures (L2, L4, L5, and L6), although L2 and L5 strains expressed L(3,7,9) as minor antigens, TNF-α (P < .01) or IL-6 (P < .01).

Effect on cytokine levels after treatment of LOS with pooled human serum. In 6 experiments, cytokine levels for LOSs from the different meningococcal immunotypes incubated with the pooled human serum were significantly reduced, compared with the results obtainedwith the LOSs alone. In the presence of serum, the levels of both TNF-α and IL-6 were reduced to the levels obtained with the PBS control (P < .01) (figures 1B and 2B).

Effect on cytokine levels after treatment of LOSs with immune mouse serum induced by the L3 strain. In 6 experiments, cytokine levels for LOSs from the different immunotypes incubated with the L3-induced immune mouse serum strain were significantly lower for LOSs of immunotype strains L3, L5, L7, and L9, compared with results obtained with the LOSs in the absence of serum (P<03). Treatment of the E. coli LPSwith the immune mouse serum did not significantly reduce the release of either TNF-α or IL-6 (figures 1C and 2C). The pooled serumsample obtained from mice prior to immunization did not significantly reduce cytokine levels for any of the LOS or E. coli LPS samples (P > .65).

Compared with results obtained with LOSs alone, TNF-α levels, induced by LOSs coincubated with immune mouse serum, were reduced 6-fold for immunotypes L3 and L7 (P < .01) and 4-fold for immunotype L9 (P < .01). Immunotype L8 that coexpressed the L(3,7,9) immunotype showed a reduction of TNF-α levels by ∼40% (P < .01). Release of TNF-a by immunotype L5 was also reduced significantly (40%) by treatment with the mouse serum (P < .01; figure 1C).

Compared with results obtained with the LOSs alone, IL-6 release was reduced by treatment of the LOS preparations with immune mouse serum by 85%–90% for immunotypes L3, L7, and L9 (P < .01). Immunotypes L5 and L8(3,7,9) were reduced by 50% and 35%, respectively (P < .01), L2 was reduced by 12% (P < .02), and L1 was reduced by 5% (P < .01; figure 2C).

Discussion

Previous investigations of the proinflammatory responses to meningococcal LOSs used a variety of methods: in vivo observations in patients with meningococcal disease [1, 38], ex vivo experiments with peripheral blood mononuclear cells (PBMC) [39–42] or granulocytes (PMNL) [4, 43], in vivo mouse studies [44–46], or whole-blood samples from patients or first-degree relatives of patients who died of meningococcal disease [2, 3, 27]. Although these models for the assessment of inflammation reflect the diversity of the inflammatory responses to LOS challenge, some factors might limit their use as a standardized model system for the assessment of the inflammatory responses induced by bacterial antigens or screening of potential vaccine candidates.

Findings from in vivo studies that have used mouse models are only partially transferable to the inflammatory responses in humans. Meningococci are exclusively human pathogens. Patients who develop fulminant septicemia showed circulating LOS plasma levels of 210–170,000 ng/L, and patients with meningitis showed plasma LOS levels of 25–260 ng/L [1]. This indicates that relatively small amounts of LOSs are associated with the pathology of meningococcal disease in humans. The lethal dose of purified LOS in mice is ∼10–100 ng LOS/ mouse [40, 47]. The typical weight of a mouse is 20 g, which reflects a level of LOS in mice that is up to 50 times greater than that in humans.

Because of the genetically controlled variability of the inflammatory response of individuals to bacterial antigens and toxins [2, 3, 24, 25, 27], the use of PBMC and PMNL from individual donors will result in a wide range of inflammatory responses to LOS challenge. The effect of cigarette smoking and symptomatic or asymptomatic viral infections needs to be considered in relation to the results obtained with blood samples fromindividual donors [27, 31].

The use of the monocyte cell line was considered to have several advantages. THP-1 cells and other monocytic-like cell lines (U937, MonoMac6, and HL-60) are readily available through national cell culture collections. Large numbers of cells can be produced within a relatively short period of time for use in studies of inflammatory responses or protective effects of monoclonal and polyclonal antibodies produced against candidate vaccine antigens or convalescent or normal human serum. The use of cell lines provides a greater reproducibility among different laboratories over time. The method reduces the effects of genetically controlled individual variation in the inflammatory responses [2, 3, 24, 25] and the effects of environmental factors such as exposure to cigarette smoke [27] or effects of recent or current virus infections [31].

Limitations of this model include several considerations. First, the use of cell lines does not take into account the genetic and environmental variability found in human populations [3, 24, 27]. Second, tumor cell lines are immature cells and do not express all the cell surface-antigen phenotypes found in normal human monocytes and macrophages. Third, the involvement of PMNL, other leukocytes, and antigen-presenting cells on the inflammatory process cannot be assessed.

Comparison of responses elicited by LOS immunotypes. In the present model system, immunotypes L3, L7, L8, and L9 induced significantly higher levels of TNF-α and IL-6, compared with other immunotypes. These findings could partly explainwhy the immunotype L(3,7,9) ismost frequently isolated frompatients with serogroup B or C meningococcal disease and other immunotypes are associated with carriage [11, 12]. The LOS moieties responsible for different proinflammatory responses are not clear. Differences in LOS core, oligosaccharide, or lipid A structure might affect the inflammatory responses in the absence of LOSspecific antibodies.

The oligosaccharide chain length was shown to affect the bioactivity of meningococcal LOSs, with mutant cells that expressed short LOS moieties being less active than their wild-type forms [44, 45]. Immunotypes L(3,7,9), L2, and L5 express identical oligosaccharidemoieties, but their ability to induce inflammatory cytokines varied greatly. Sialylation of the terminal galactose residue might play a role in the ability to induce inflammation. Immunotype reference strains L1–L8 are thought to be fully or partially sialylated. Immunotype L7 induced the highest levels of cytokines, but these were not significantly higher, compared with the structurally related immunotypes L3 and L9.

Core antigens might also contribute to induction of these responses. With the exception of immunotypes L10 and L11, LOSs from strains that express PEA on the second core heptose induced significantly higher cytokine levels (P < .01), compared with immunotypes that expressed glucose (1→3) or that lacked a functional group. LOSs from immunotype L(3,7,9), which was reported to bind the monoclonal antibody that detects the epitope expressed in the PEA (1→3) HepII core antigen [28, 37], induced the highest levels of cytokines, which suggests that both core structure and α-chain moieties might alter the bioactivity of meningococcal LOSs.

The lipid A moiety of meningococcal LOSs is thought to be heterogeneous [48–51]. Monophosphorylation of meningococcal lipidAis thought to be less toxic, compared with diphosphorylated LOSs [14].With limited information on the lipid A moieties and little published evidence for a correlation between lipid A structure and bioactivity, it is not possible to assess the findings of the present study in relation to the structure of lipid A.

Assessment of results in relation to vaccine development. The THP-1 cell line was useful for screening for the induction of proinflammatory cytokines by LOS preparations and eliminated the variability of genetic or environmental factors. The LOSs from strains that expressed the L(3,7,9) moiety as the major immunotype induced significantly higher levels of proinflammatory cytokines, compared with LOSs obtained from other meningococcal immunotypes. The results obtained with the in vitro model complement the clinical and epidemiological findings and could partly account for the predominance of L(3,7,9) strains among isolates from patients with meningococcal disease. The results confirm the toxicity of the meningococcal endotoxin.

Acknowledgements

We are grateful to W. D. Zollinger (Walter Reed Army Hospital for Medical Research, Washington, DC) for supplying the lipooligosaccharide immunotype strains.

Presented in part: 12th International Pathogenic Neisseria Conference, Galveston, Texas, 12-17 November 2000 (abstract 063).

Animal experimentation guidelines were followed in accordance with an animal licence issued by the British Home Office.

Financial support: Meningitis Association of Scotland, the Lloyds TSB Foundation, Institute for Scientific Evaluation of Naturopathy, University of Cologne.

a

Present affiliations: Institute for Scientific Evaluation of Naturopathy, University of Cologne Medical School, Cologne, Germany (J.M.B. and D.M.W.); Discipline of Immunology and Microbiology, University of Newcastle, Newcastle, Australia (C.C.B.); Faculty of Medical Technology, Great Al Fateh University for Medical Sciences, Tripoli, Libya (O.E.A.); Department of Molecular and Cellular Pathology, Ninewells Hospital and Medical School, University of Dundee, Dundee, Scotland (A.E.G.).