ABSTRACT
Cytokine responses of chronic Q fever patients to the intracellular bacterium Coxiella burnetii have mostly been studied using ex vivo stimulation of immune cells with heat-killed C. burnetii due to the extensive measures needed to work with viable biosafety level 3 agents. Whether research with heat-killed C. burnetii can be translated to immune responses to viable C. burnetii is imperative for the interpretation of previous and future studies with heat-killed C. burnetii. Peripheral blood mononuclear cells (PBMCs) of chronic Q fever patients (n = 10) and healthy controls (n = 10) were stimulated with heat-killed or viable C. burnetii of two strains, Nine Mile and the Dutch outbreak strain 3262, for 24 h, 48 h, and 7 days in the absence or presence of serum containing anti-C. burnetii antibodies. When stimulated with viable C. burnetii, PBMCs of chronic Q fever patients and controls produced fewer proinflammatory cytokines (interleukin-6 [IL-6], tumor necrosis factor alpha, and IL-1β) after 24 h than after stimulation with heat-killed C. burnetii. In the presence of Q fever seronegative serum, IL-10 production was higher after stimulation with viable rather than heat-killed C. burnetii; however, when incubating with anti-C. burnetii antibody serum, the effect on IL-10 production was reduced. Levels of adaptive, merely T-cell-derived cytokine (gamma interferon, IL-17, and IL-22) and CXCL9 production were not different between heat-killed and viable C. burnetii stimulatory conditions. Results from previous and future research with heat-killed C. burnetii should be interpreted with caution for innate cytokines, but heat-killed C. burnetii-induced adaptive cytokine production is representative of stimulation with viable bacteria.
INTRODUCTION
Coxiella burnetii is the causative microorganism of Q fever, a highly infectious zoonosis that is prevalent worldwide. Infection is contracted by inhalation of contaminated aerosols. After inhalation and infection, the obligate intracellular growing C. burnetii invades the alveolar macrophages. Replication and outgrowth of Coxiella take place in a peculiar compartment called the Coxiella-containing vacuole (CCV) within macrophages or monocytes (1). Unlike other intracellular bacteria, C. burnetii does not try to escape phagolysosomes but instead uses the acidic compartment to activate an important virulence factor, the type IV secretion system (T4SS), through which it modulates host cell metabolism and gene expression to enhance its CCV survival (2).
In humans, acute Q fever is most often associated with a relatively mild flu-like disease or, in some cases, pneumonia or hepatitis. Initial infection occurs asymptomatically in 60% of infections. In a minority of infected individuals, however, initial symptomatic or asymptomatic infection persists, often manifesting as an endocarditis or infection of an aneurysm or vascular prosthesis (1). During such persistent infection, also referred to as chronic Q fever or persistent focalized Q fever infection (3), C. burnetii is not eradicated by the host immune response. Individuals with heart valve defects, aneurysms, or vascular prostheses are susceptible to develop a persistent infection (4). During the major Q fever epidemic in the Netherlands, approximately 30% of individuals with known aorto-iliac disease and serological evidence of Q fever infection developed chronic Q fever (5). Therefore, an additional risk factor resulting in a defective systemic or local immune response is likely to be required for the development of chronic Q fever.
To evaluate the systemic immune response in chronic Q fever patients, various studies examined the cytokine and chemokine response to heat-killed C. burnetii. Nearly all of these studies showed a robust innate cytokine response in chronic Q fever patients, with a strikingly more profound production of interferon gamma (IFN-γ) than in Q fever seronegative controls who lacked IFN-γ response (6–9). Various studies also showed a higher production of anti-inflammatory interleukin-10 (IL-10) in chronic Q fever patients (10–13).
Due to its highly infectious nature, C. burnetii is classified as a biosafety level 3 (BSL3) agent (14). Therefore, nearly all stimulation studies were performed with heat-killed C. burnetii or with the avirulent phase II organisms to prevent researchers from exposure and to circumvent complicated and expensive biosafety procedures. Whether results obtained in experiments with heat-killed bacteria can be translated to the immune response to viable C. burnetii, armed with tools to influence host cell metabolism, has not been evaluated.
Therefore, in the present study, we compare the cytokine-inducing capacities of strains of heat-killed and viable C. burnetii in peripheral blood mononuclear cell (PBMC) cultures of chronic Q fever patients and healthy controls.
RESULTS
Viable C. burnetii induces few cytokines in PBMCs of Q fever naive controls in the presence of Q fever naive serum.In a small study (n = 4) aimed to justify the final experimental setup, the effect of heat-killed and viable bacteria on cytokine production of PBMCs of Q fever naive controls in the presence of Q fever naive serum was examined. To this end, cells were stimulated with two C. burnetii strains, either viable or heat-killed Nine Mile or 3262 strain. It appeared that the production of tumor necrosis factor alpha (TNF-α), IL-1β, and IL-6, mostly secreted by monocytes, was lower after stimulation with viable C. burnetii than with heat-killed C. burnetii (see Fig. S1 in the supplemental material). Generally, the Nine Mile and 3262 strains induced similar effects, although not for every cytokine. Levels of IL-10 production induced by the 3262 strain were similar for viable and heat-killed bacteria, but the live Nine Mile strain showed a higher cytokine response than the heat-killed bacteria. Due to the small sample size, results are not statistically significant.
The levels of cytokines IFN-γ, IL-17, and IL-22, mostly secreted by T cells and, therefore, belonging to the adaptive immune response, were determined in the supernatants after 48 h and 7 days. Stimulation with viable and heat-killed bacteria from either of the strains led to similar cytokine production (Fig. S2). Cytokine concentration in the unstimulated cells remained below the detection limit. However, IL-17 production seemed to be higher after stimulation with viable than heat-killed bacteria. Due to the small sample size, results are not statistically significant.
The effect of anti-C. burnetii IgG phase I antibodies in the C. burnetii stimulation experiments was determined with PBMCs from four chronic Q fever patients and four healthy controls. The controls and patients responded similarly to the serum of a chronic Q fever patient with high anti-C. burnetii antibody titers (1:4,096) (Fig. 1).
(A to H) TNF-α and IL-1β production of peripheral blood mononuclear cells after 24 h of stimulation with heat-killed C. burnetii NM (106 bacteria/ml) (A and E), C. burnetii 3262 (106 bacteria/ml) (B and F), Candida albicans conidiae (105/ml) (C and G), and Escherichia coli LPS (10 ng/ml) (D and H). Cells (5 × 105) of chronic Q fever patients or healthy controls were incubated without serum, with Q fever naive serum, or with serum from a chronic Q fever patient with high IgG phase I titers. Abbreviations: HK, heat-killed; NM, Nine Mile strain; ns, nonsignificant; pts, patients. Q fever naive and Q fever serum conditions were compared with the Mann-Whitney test.
Viable C. burnetii induces less innate cytokine production than heat-killed C. burnetii in PBMCs of chronic Q fever patients and Q fever naive controls in the presence of serum containing Q fever antibodies.PBMCs of 10 chronic Q fever patients and 10 healthy Q fever naive controls were stimulated with viable or heat-killed C. burnetii. Cells of chronic Q fever patients were used in the experiments, because it is unknown whether these patients, who are assumed to have a defective immune response to C. burnetii, respond to viable bacteria in a manner similar to that of healthy individuals. To mimic the in vivo situation, the cells were incubated with serum from a patient with chronic Q fever containing high titers of anti-C. burnetii IgG phase I. Two strains of C. burnetii were used, the laboratory strain Nine Mile and the predominant strain during the Q fever outbreak in the Netherlands (C. burnetii 3262) that infected approximately 40,000 individuals.
The innate cytokines TNF-α, IL-1β, and IL-6 were measured after 24 h of stimulation (Fig. 2). Cytokine production in unstimulated cells remained below the detection limit of the tests.
(A to D) Innate cytokine production after heat-killed (HK) and viable (L) Coxiella burnetii stimulation in peripheral blood mononuclear cells of chronic Q fever patients (n = 10) and healthy Q fever naive controls (n = 10). The Dutch outbreak strain 3262 (Cb 3262) and the Nine Mile strain (Cb NM) were used to stimulate the cells with 2 × 106 bacteria per condition. Cells (5 × 105) were incubated in the presence of serum containing anti-C. burnetii antibodies for 24 h, after which IL-6, IL-1β, and TNF-α were measured. IL-10 production was measured after 48 h. *, P < 0.05; **, P < 0.01. Nonsignificant results are not indicated by a symbol, and results were analyzed with the Wilcoxon signed-rank test.
In chronic Q fever patients, TNF-α production of viable C. burnetii 3262-stimulated cells was not different from the TNF-α production induced by heat-killed C. burnetii. However, viable C. burnetii Nine Mile induced less TNF-α production than heat-killed C. burnetii Nine Mile (P < 0.01). In controls, both C. burnetii strains showed the same effect, viable C. burnetii induced less TNF-α production than heat-killed (P < 0.01). Chronic Q fever patients showed high variability in cytokine production, and the cytokine responses were not significantly different from those of healthy controls.
The IL-1β production induced by viable C. burnetii 3262 in chronic Q fever patients was similar to that of heat-killed C. burnetii 3262, in contrast to the lower cytokine production induced by viable C. burnetii NM than heat-killed C. burnetii NM (P < 0.01). In controls, for both strains the viable C. burnetii condition led to less cytokine production (P < 0.01). As for TNF-α, IL-1β production by PBMCs of patients was highly variable, and there was no difference with IL-1β production by PBMCs of healthy controls.
Chronic Q fever patients and controls readily produced the proinflammatory cytokine IL-6. In chronic Q fever patients, for both C. burnetii strains, viable C. burnetii induced less IL-6 than heat-killed C. burnetii (C. burnetii 3262, P < 0.05; C. burnetii Nine Mile, P < 0.01). For the C. burnetii 3262 strain, healthy controls showed similar levels of IL-6 production under the viable and heat-killed conditions. With regard to the Nine Mile strain, viable C. burnetii induced less IL-6 than heat-killed C. burnetii (P < 0.01). IL-6 production was not different between chronic Q fever patients or controls in any of the stimulations.
The anti-inflammatory cytokine IL-10 was measured after 48 h. In chronic Q fever patients, viable C. burnetii 3262 induced less IL-10 than heat-killed C. burnetii (P < 0.01), although this was not the case for the Nine Mile strain. PBMCs of controls produced less IL-10 after viable C. burnetii 3262 and Nine Mile strain stimulation than after heat-killed C. burnetii stimulation (P < 0.01 and P < 0.05, respectively).
Bacterial replication, assessed with quantitative PCR (qPCR), did not take place within the first 48 h of incubation (Fig. 3). Cell lysis, measured by lactate dehydrogenase (LDH) release, did not differ between unstimulated and stimulated conditions after 24 h (Fig. 3).
(A and B) Lactate dehydrogenase (LDH) assay to assess necrotic cell death after 24 h (A) or 7 days (B) of stimulation with heat-killed (HK) and live C. burnetii Nine Mile (NM) strain or 3262 strain. There was no difference in cell death compared to that for the medium control. (C and D) C. burnetii genomic copies between start of experiment and 7 days. Cb NM copy numbers significantly decreased during cell culture, while C. burnetii 3262 did not significantly decrease. *, P < 0.05; **, P < 0.01; ns, nonsignificant. Results were analyzed with the Wilcoxon signed-rank test (A and B) and one-way ANOVA with Bonferroni's multiple-comparison test (C and D).
Lower innate cytokine production induced by viable bacteria does not lead to reduced adaptive cytokine production.After 7 days, IFN-γ, IL-17, and IL-22 production was determined in the supernatants, and IFN-γ production was also measured after 48 h (Fig. 4). IL-10 production was also measured after 7 days but was undetectable under all conditions.
(A to D) Adaptive cytokine production after heat-killed (HK) and viable (L) Coxiella burnetii stimulation in peripheral blood mononuclear cells of chronic Q fever patients (n = 10) and healthy Q fever naive controls (n = 10). The Dutch outbreak strain 3262 (Cb 3262) and the Nine Mile strain (Cb NM) were used to stimulate the cells with 2 × 106 bacteria per condition. Cells (5 × 105) were incubated in the presence of serum containing anti-C. burnetii antibodies for 48 h and 7 days, after which IFN-γ production was measured. IL-17 and IL-22 production was measured after 7 days. Nonsignificant results are not indicated by a symbol, and results were analyzed with the Wilcoxon signed-rank test.
IFN-γ is crucial in the defense against intracellular pathogens and known to be highly produced upon stimulation with heat-killed C. burnetii by chronic Q fever patients, in contrast to healthy controls (6, 7). In contrast to the mostly innate cytokines, IFN-γ production was not influenced by C. burnetii viability in chronic Q fever patients. Controls also showed no differences in cytokine production between viable and heat-killed C. burnetii conditions. At both time points and regardless of the viability of the bacteria, IFN-γ production was higher in chronic Q fever patients after stimulation with C. burnetii 3262 (viable; median, 638 pg/ml; range, 78 to 46,733) than in controls (median, 77 pg/ml; range, 8 to 890 pg/ml; P < 0.01) or with C. burnetii Nine Mile (viable patients, median, 728 pg/ml; range, 78 to 41,295 pg/ml; controls, median, 59 pg/ml; range, 8 to 900 pg/ml). In chronic Q fever patients and controls, IL-17 and IL-22 production also did not differ between heat-killed and viable C. burnetii, which was apparent in both strains. There was no difference between chronic Q fever patients and controls regarding IL-17 and IL-22 production.
After 7 days, bacterial numbers did not exceed the number of bacteria used to initially stimulate the cells (Fig. 3). There was no difference in cell death, assessed by LDH release and trypan blue staining, between the unstimulated and stimulated conditions after 7 days of culture (Fig. 3).
Chemokine CXCL9 is not differentially produced after incubation with viable or heat-killed Coxiella.CXCL9, also known as monokine induced by interferon gamma (MIG), is a chemokine that is primarily induced by IFN-γ. It predominantly attracts T cells and thereby plays a role in orchestrating the T-helper 1 cell response (15). We previously reported that CXCL9 is scarcely produced after heat-killed C. burnetii stimulation of PBMCs of healthy controls, but evidently it is produced by cells taken from chronic Q fever patients (16). Apart from altered cytokine production, dysfunctional chemokine production may prevent adequate influx of T effector cells, affecting the adaptive immune response. Therefore, we determined the CXCL9 concentration after 24 h of incubation (Fig. 5). There was no significant difference in CXCL9 concentration after live or heat-killed C. burnetii incubation.
CXCL9 production after heat-killed (HK) and viable (L) Coxiella burnetii stimulation in peripheral blood mononuclear cells of chronic Q fever patients (n = 10) and healthy Q fever naive controls (n = 10). The Dutch outbreak strain 3262 (Cb 3262) and the Nine Mile strain (Cb NM) were used to stimulate the cells with 2 × 106 bacteria per condition. Nonsignificant results are not indicated by a symbol, and results were analyzed with the Wilcoxon signed-rank test.
DISCUSSION
In the present study, we showed that viable C. burnetii induces fewer of the innate cytokines TNF-α, IL-1β, IL-6, and IL-10 than heat-killed bacteria but equal amounts of T-cell-derived (adaptive) cytokines IFN-γ, IL-17, and IL-22 and the IFN-γ-dependent chemokine CXCL9. Hence, data obtained in studies evaluating innate cytokine production induced by heat-killed C. burnetii cannot be extrapolated to the real-life situation where infection is induced by viable C. burnetii. Regarding the interpretation of previous research with heat-killed C. burnetii, the in vivo immune activation due to viable bacteria has likely been overestimated. On the other hand, measurements of the adaptive cytokine production, such as that with an IFN-γ assay (17), reflect the response to viable bacteria appropriately. Consistent with previous studies performed by our research group (6, 7, 16), heat-killed C. burnetii induced higher IFN-γ and CXCL9 production in chronic Q fever patients than in controls, which could now also be confirmed for viable C. burnetii.
Multiple theories may explain the attenuated cytokine production after incubation with viable bacteria compared to that of heat-killed bacteria. It is conceivable that heat killing increases the release and exposure of bacterial pathogen-associated molecular patterns (PAMPs) and thereby affects signaling by PAMP receptors. The uptake, internalization, and phagosomal pathway of viable and dead bacteria are less likely to influence cytokine production, because viable and dead bacteria are both taken up passively via actin-dependent phagocytosis, and the maturation of the CCV occurs irrespective of C. burnetii replication or protein synthesis (18, 19). However, once they arrive in the CCV, viable bacteria are likely to influence the host cell due to the type IV secretion system (T4SS), which can deliver bacterial effector proteins into the cytosol of the host. For example, C. burnetii is able to inhibit apoptosis, promote fusion of the CCV with autophagosomes, and modulate host gene expression, including genes of the immune response. Reportedly the genes IL-8, CCL2, CXCL1, and SPP1, which control cell trafficking and IL-12 production, are actively downregulated by C. burnetii (20). Thus, several probable mechanisms may explain the lower proinflammatory innate cytokine induction after stimulation with viable rather than nonviable C. burnetii.
In addition to the observed differences in cytokine induction by viable and heat-killed C. burnetii, we noticed a trend toward higher anti-inflammatory IL-10 production after stimulation with viable bacteria in the presence of Q fever naive serum. If this observation can be validated, it supports the assumption that the induction of IL-10 early in the course of the disease is one of the immune evasion strategies of (viable) C. burnetii. Similar to our observation, Mycobacterium tuberculosis has been reported to enhance IL-10 production and downregulate proinflammatory cytokines (21). Differences between strains in the induction of IL-10 may also be one of the virulence factors of C. burnetii. In a mouse model of Q fever, the Nine Mile strain appeared to be more virulent than the 3262 strain (22). Notably, in the presence of C. burnetii IgG phase I antibodies, this shift toward higher IL-10 production was absent. In Q fever patients, phase I antibodies appear during the transition to chronic infection. The role of anti-C. burnetii IgG antibodies has remained controversial. It was shown that opsonization of C. burnetii favors replication in human-derived monocytes, while unopsonized bacteria survive poorly (23). However, Fcγ receptor (FcγR)-mediated phagocytosis by dendritic cells induces cytokine production, whereas non-antibody-containing serum with C. burnetii does not induce changes in dendritic cells (24). In mice, FcγR- or complement-mediated phagocytosis does not play a role in vivo (24). If the observed trend of higher IL-10 production after stimulation with viable bacteria can be validated, our results suggest that in the chronic phase of the disease, C. burnetii IgG phase I antibodies prevent a possible deleterious overproduction of IL-10 and thus may be considered in part protective.
A note of caution is due here, because all patients used antimicrobial drugs, which are known to influence cytokine production through inhibited transcription (25) and may also affect bacterial elimination. It should also be noted that patients were recruited multiple months after they started antibiotic treatment. Furthermore, the viable bacterial inoculum inevitably contained a small percentage of dead bacteria, which could have affected the cytokine production. The percentage of dead bacteria under the viable stimulatory condition is unknown; however, C. burnetii is very resilient and can resist many chemical compounds, and so large percentages of dead C. burnetii bacteria were unlikely.
Conclusions.Heat-killed C. burnetii is frequently used as an alternative for viable C. burnetii due to the requirement of dedicated facilities. However, innate proinflammatory cytokine production induced by viable bacteria are lower than that by heat-killed C. burnetii. Levels of adaptive cytokine production are comparable between induction by viable and heat-killed C. burnetii in humans. The results of this study can be used to interpret previous and future research with heat-killed C. burnetii regarding C. burnetii-specific cytokine production in healthy individuals and chronic Q fever patients.
MATERIALS AND METHODS
Subjects.Healthy controls (n = 10) were hospital personnel, all seronegative for Q fever, as tested by immunofluorescence assay (IFA). Chronic Q fever patients (n = 10) were recruited from the outpatient clinic of the Radboud University Medical Center Nijmegen. Patients were diagnosed according to the Dutch Consensus Guidelines on chronic Q fever diagnosis (26); this guideline categorizes diagnosis as proven, probable, or possible chronic infection and is based on serology, PCR, clinical symptoms, risk factors, and diagnostic imaging. All patients fulfilled the criteria of proven chronic Q fever and were being treated with antibiotics, mostly doxycycline, and hydroxychloroquine at the time of blood drawing. Patient characteristics are shown in Table S1 in the supplemental material. Serum with anti-C. burnetii IgG phase I antibodies was drawn from a chronic Q fever patient with high IgG phase I titers (1:4,096) determined by IFA. Q fever naive serum was drawn from a healthy control who participated in the study.
Bacteria.C. burnetii strains were cultured on Buffalo green monkey cells at Wageningen Bioveterinary Research (former Central Veterinary Institute), the Netherlands, as described previously (27). Two strains were used, the frequently used Nine Mile (RSA 693) strain and the most prominent Dutch outbreak strain, 3262, isolated in 2009 (28). The Nine Mile strain is a laboratory strain isolated from a tick and is generally used for research purposes (29). The C. burnetii 3262 strain had been isolated from a goat during the Dutch epidemic and is pathogenic to humans.
Bacterial DNA, measured to quantify the amount of bacteria, was isolated using a DNA isolation kit (DNeasy blood and tissue kit; Qiagen, the Netherlands). Quantification by qPCR was done as described by Roest et al. (28). In brief, a single-copy gene encoding a C. burnetii specific hypothetical protein (GenBank accession number AY502846) was targeted using the forward primer 5′-ATAGCGCCAATCGAAATGGT-3′, the reverse primer 5′-CTTGAATACCCATCCCGAAGTC-3′, and the NED-labeled probe 5′-CCCAGTAGGGCAGAAGACGTTCCCC-3′. To quantify the bacterial numbers, the commercially available C. burnetii standard positive-control dilution series was used (Adiavet Cox, Adiagene, Marcy l' Etoile, France). qPCR was performed on a 7500 Fast real-time PCR system (Applied Biosystems, USA). After determination of the amount of both Coxiella batches, they were brought to a concentration of 4 × 107 bacteria/ml. The first half was heat killed at 99°C for 30 min (30) while the second half remained viable. Both were aliquoted and frozen at −80°C until the experiment, which was performed within 3 months. Determination of C. burnetii gene copies at specific time points (4 h, 24 h, 48 h, and 7 days) after stimulation experiments in frozen cell pellets was executed by following the same DNA isolation and qPCR protocol as that described above.
PBMC isolation and stimulation.PBMCs from all individuals were isolated according to a standardized protocol, described in detail by Oosting et al. (31). In short, cells were separated by density centrifugation with Ficoll Paque (GE Healthycare, Zeist, the Netherlands), taken from the interphase layer, and washed twice in cold phosphate-buffered saline. The cells were counted with a Z1 Coulter particle counter (Beckman Coulter, Woerden, the Netherlands) and brought to a concentration of 5 × 106 cells/ml in RPMI 1640 (Life Technologies/Invitrogen, Breda, the Netherlands) supplemented with 2 mM GlutaMAX and 1 mM pyruvate without antibiotics. Cells were subsequently transported on ice to the BSL3 facility. In a round-bottom 96-well plate (Corning, New York, USA), 5 × 105 cells were incubated with 10% human serum and either Q fever naive serum or serum containing IgG phase I and II antibodies (original IFA titers of 1:4,096). Subsequently, the cells were stimulated with either 2 × 106 heat-killed or viable C. burnetii Nine Mile bacteria and with 2 × 106 heat-killed or viable C. burnetii 3262 bacteria. The infection index, i.e., the ratio of bacteria to PBMCs, was 4:1 and the estimated number of bacteria per monocyte was 20:1, identical to that used in earlier studies on the measurements of cytokine production in chronic Q fever patients and healthy controls (6, 32). Positive and negative controls were heat-killed Candida albicans conidia (105/ml) and culture medium, respectively. After 24 h, 48 h, and 7 days, the supernatants were harvested after centrifugation and stored at −80°C. Cell death at 24 h and 7 days was assessed by measuring lactate dehydrogenase (LDH) in the supernatants using Cytotox96 nonradioactive cytotoxicity assay (Promega Corporation, Madison, WI, USA) and additionally at 7 days by staining with trypan blue and subsequent counting by microscopy. The remaining cell pellets were frozen at −80°C to assess C. burnetii replication during the experiments.
Cytokine measurements.Cytokine concentration was measured at different time points in the PBMC culture supernatants using enzyme-linked immunosorbent assay (ELISA) according to the manufacturer's instructions. After 24 h, the concentration of TNF-α, IL-6, IL-1β, and CXCL9 (R&D Systems, Minneapolis, MN, USA) was measured. After 48 h, the concentrations of IL-10 and IFN-γ (Sanquin, Amsterdam, the Netherlands) were determined. After 7 days, the concentrations of IL-10, IFN-γ, IL-17, and IL-22 (R&D Systems) were measured.
Statistical analysis.All data were analyzed using GraphPad Prism 5.0. Medians and interquartile ranges were calculated for cytokine production. Differences between cytokine production induced by heat-killed and viable bacteria were analyzed with a Wilcoxon signed-rank test. Differences in number of bacteria were analyzed by means of one-way analysis of variance (ANOVA) with Bonferroni post hoc tests. A P value below 0.05 was considered statistically significant.
Ethical consideration.This study was approved by the local institutional review board (NL35784.091.11; CMO Arnhem-Nijmegen). All subjects provided written informed consent.
ACKNOWLEDGMENTS
We thank Michiel Kroese, Betty van Gelderen, and Marianne Vahl for their assistance with the experiments and Dimitrios Frangoulidis, Bundeswehr Institute of Microbiology, for providing the C. burnetii Nine Mile strain.
A.J., T.S., and A.D. performed the experiments; P.C.W. performed additional experiments; M.V.D., A.K., H.J.K., and L.A.B.J. designed the study; C.B. collected patient material; A.F.M.J. drafted the manuscript; all authors reviewed and approved the final manuscript.
This study was funded by the Q-Support Foundation. The funders had no role in study design, experiments, or drafting of the manuscript.
We have no conflicts of interest to declare.
FOOTNOTES
- Received 4 May 2018.
- Returned for modification 30 May 2018.
- Accepted 17 July 2018.
- Accepted manuscript posted online 23 July 2018.
Supplemental material for this article may be found at https://doi.org/10.1128/IAI.00333-18.
REFERENCES
- Copyright © 2018 American Society for Microbiology.
