To investigate the kinetic Th1/Th2 immunopathogenic mechanisms of Haemophilus influenzae meningitis, we established a murine experimental model of meningitis and elucidated the Th1/Th2 immune responses in T1/T2 doubly transgenic mice based on a BALB/c background under the control of the IFN-γ (interferon-γ)/IL-4 (interleukin-4) promoters respectively. NTHi (non-typeable Haemophilus influenzae) meningitis was induced in these mice by inoculation with either a colonized (CNTHi) or invasive (INTHi) strain of NTHi. Mice inoculated with CNTHi displayed a less severe degree of disease in terms of clinical symptoms, mortality rate and brain histopathology. Conversely, INTHi-inoculated mice had more severe clinical symptoms. CNTHi-inoculated mice had a more significant Th1 response in terms of a higher percentage and longer maintenance of Th1 cells, and more production of IFN-γ from strain-specific antigen-stimulated splenocytes than INTHi-inoculated mice. In contrast, INTHi-inoculated mice had a more significant Th2 response. This was due to a significant increase in IL-4-producing CD4+ T-cells (Th2 cells) and more production of IL-4 from strain-specific antigen-stimulated splenocytes accompanied by a rapid decline of Th1 cells in INTHi-inoculated mice. In conclusion, the preferential Th1/Th2 trend in this murine model of NTHi meningitis is correlated with clinical severity as well as isolated characteristics of the pathogens themselves.

INTRODUCTION

Bacterial meningitis is recognized as one of the leading causes of infection-related death, and this disease still continues to be a significant health concern [1]. In the past, NTHi (non-typeable Haemophilus influenzae) was rarely the cause of meningitis in children [2,3] but, since the Hib (H. influenzae type b) vaccine era, it has become more prominent in fully Hib-vaccinated children [4,5]. An extended study [6] demonstrated that the incidence of non-Hib diseases in all children under 5 years of age was elevated, compared with a decreased incidence of Hib disease, and the majority of non-Hib strains isolated were NTHi strains. NTHi has thus become a more central cause of H. influenzae disease, and these micro-organisms must be considered as potential pathogens of bacterial meningitis in children. NTHi is a common commensal organism in the human upper respiratory tract and also an important cause of infectious disease. The pathogenesis of NTHi infection involves multiple steps and the interplay of both bacterial and host factors. The mechanism of NTHi infection begins with bacterial colonization of the nasopharynx, involving establishment on the mucosal surface and evasion of local immune mechanisms, eventually leading to disease [7]. However, the potential diversity of Th1/Th2 immunity in pathogenic NTHi strains isolated from infectious cases or colonized strains isolated from nasopharynx of healthy children is rarely reported and needed to be clarified further.

Various microbes may induce diverse Th1/Th2 immune responses under different conditions. The potential involvement of Th1/Th2 cytokines in bacterial meningitis has been studied, and some reports indicate the early presence of Th1-related cytokines in CSF (cerebrospinal fluid) of infants and children with bacterial meningitis [8,9]. Increased levels of Th2-related cytokines have also been identified in CSF and serum of infants and children with bacterial meningitis [10,11]; however, these results may not entirely dissect the diversity of Th1/Th2 responses in the host with bacterial meningitis. NTHi with both an extracellular and intracellular niche in the human respiratory tract has been suggested [12].

To detect Th1/Th2 cells directly in vivo, we have generated T1/T2 doubly transgenic mice. These T1/T2 doubly transgenic mice bear two transgenes, which express two distinct cell-surface markers: a human THY1 transgene (hTHY1) under the control of the murine Igng [gene encoding IFN-γ (interferon-γ)] promoter, and a murine Thy1.1 transgene (mThy1.1) under control of the murine Il4 [gene encoding IL-4 (interleukin-4)] promoter, designated as T1 and T2 respectively [13]. These transgenic mice have been used previously [13] as a murine model to monitor the in vivo development of Th1 or Th2 cells during Listeria monocytogenes or Schistosoma mansoni infections. The detection of both intracellular IFN-γ production and T1 transgene expression demonstrates that the Th1-dominant response was induced in mice infected with L. monocytogenes. A similar correlation between IL-4 and T2 transgene expression was observed with S. mansoni infection in vivo [13]. These T1/T2 doubly transgenic mice have been used as a monitor system in autoimmune diabetes [14,15]. Thus these transgenic mice provide a valuable system for tracing Th1 or Th2 cells in vivo.

Murine models of bacterial meningitis have been induced by direct intracranial injection of Escherichia coli, Streptococcus pneumoniae or L. monocytogenes [16]. Nevertheless, murine models of H. influenzae meningitis are still limited [17,18]. In the present study, we have established a murine model of experimental meningitis by intracranial inoculation with variant strains of NTHi either isolated from bacterial colonization of the nasopharynx of healthy children, defined as CNTHi (colonized NTHi), or isolated from CSF of patients with NTHi meningitis before prescription of antibiotics, defined as INTHi (invasive NTHi), to investigate the differences in clinical symptoms and kinetic Th1/Th2 immune responses in mice with NTHi meningitis.

METHODS AND MATERIALS

Animals

T1/T2 doubly transgenic mice were generated on a BALB/c background [13] and, subsequently, bred and raised at the Animal Center of the National Defense Medical Center following the Guidelines of the National Science Council of the Republic of China. BALB/c mice were purchased from the National Laboratory Animal Center (the proved code of certification for animal experiment in this study was NDMC&TSGH-IACUC-04-048). Female BALB/c mice were crossed with T1/T2 doubly transgenic male mice. Mice were housed in a monitored light/dark-cycle-barred environment, and were given standard laboratory chow and water ad libitum. Transgenic mice were screened by PCR as reported previously [13]. In brief, primers Tg1 and Tg2 were used to amplify the mIFN-γ:hThy1 transgene (T1), and primers Tg3 and Tg4 were used to amplify the mIL-4:mThy1.1 transgene (T2). The DNA sequences of the PCR primers were as follows: Tg1, 5′-GCTGTCTCATCGTCAGAGAGC-3′; Tg2, 5′-TCAAGGACAGGAGATCTTAGGG-3′; Tg3, 5′-CCAAGATATCAGAGTTTCCAAGG-3′; and Tg4, 5′-AGAGGCTACTTCCCGGGATG-3′.

A 600 bp DNA fragment was amplified by primers Tg1/Tg2 (see Figure 1A), and another 600 bp DNA fragment was also amplified by primers Tg3/Tg4 (see Figure 1B).

Screening of T1/T2 doubly transgenic mice

Figure 1
Screening of T1/T2 doubly transgenic mice

The screening of T1/T2 doubly transgenic mice was analysed by PCR using 0.2 μg of tail DNA as the PCR template. Both T1- (A) and T2- (B) amplified DNA fragments were screened from five littermate mice. Lane M, molecular size marker; lanes 1, 3 and 4, positive results of screening; lanes 2 and 5, negative results of screening; lane P, positive control; lane N, negative control.

Figure 1
Screening of T1/T2 doubly transgenic mice

The screening of T1/T2 doubly transgenic mice was analysed by PCR using 0.2 μg of tail DNA as the PCR template. Both T1- (A) and T2- (B) amplified DNA fragments were screened from five littermate mice. Lane M, molecular size marker; lanes 1, 3 and 4, positive results of screening; lanes 2 and 5, negative results of screening; lane P, positive control; lane N, negative control.

Bacterial strains

After receiving informed consent from the healthy subjects and patients, clinical isolates of NTHi were obtained from hospitals throughout Taiwan. CNTHi strains (numbers 255 and 63) were isolated from the nasopharynx of healthy children, who were well and without significant allergic diseases and immune-deficient disorders, as well as without infectious diseases for at least 2 weeks prior to recruitment. Alternatively, INTHi strains (numbers 450 and 645) were isolated in pure culture from CSF of patients with meningitis before antibiotic treatment [19].

Murine model for NTHi-induced meningitis

Mice (6–8-weeks old) were anaesthetized via intramuscular injection of 100 mg of ketamine/kg of body weight (Imalgene 1000; Merial Laboratoire) and 20 mg of xylazinum/kg of body weight (Rompum; Bayer), and subsequently inoculated intracranially with 40 μl of PBS (controls) or 40 μl of PBS containing 6×106 colony-forming units of either CNTHi or INTHi. For clinical evaluation, mice were inoculated with PBS (n=10), CNTHi (n=11) or INTHi (n=13) and were monitored daily. The clinical severity of the disease was scored by a method modified from previous studies [20,21]. Disease severity in clinical scores was graded from 0–5 by the following criteria: 0 (normal), normal activity and ambulation; 1 (minimal disease), decreased spontaneous activity, but still turned upright within 5 s; 2 (moderate disease), unable to turn upright within 5 s; 3 (severe disease), profound lethargy without ambulation; 4, demonstrated seizure or coma; and 5, death. We observed the clinical symptoms and graded their severity for 4 weeks after bacterial inoculation. If mice died, their clinical scores were assessed; however, these mice were subsequently excluded from the statistical analysis of disease severity.

Histological examination

At the indicated time points, mice were killed and their brains were quickly removed and fixed overnight in 10% (w/v) formalin (pH 7.4). Coronal sections were embedded in paraffin, cut into 6–10-μm thick sections and stained with haematoxylin/eosin for histological evaluation.

Isolation of splenocytes

Spleens were collected aseptically from mice at the indicated time points after inoculation, and lymphocyte subpopulations and cytokine production were analysed. Spleens were harvested in RPMI 1640 medium and minced, and RBCs (red blood cells) were depleted with Tris-buffered ammonium chloride. The remaining cell pellets, representing the total splenic mononuclear cell population, were resuspended in RPMI 1640 medium. The non-adherent lymphocyte population was harvested, washed and resuspended in PBS containing 1% (v/v) FBS (fetal bovine serum) for flow cytometric analysis [22].

Flow cytometry

Prepared splenocytes (106 cells in 0.1 ml of PBS) were incubated on ice and stained with the following marker-specific antibodies (0.5 μg of antibody/106 cells): FITC-conjugated anti-(mouse CD90.1) (mThy1.1; clone OX-7), anti-(mouse CD19) (clone 1D3) and anti-(mouse DX5) antibodies; PE (phycoerythrin)-conjugated anti-(human CD90) (hThy1; clone 5E10) and anti-(mouse CD8α) (Ly2; clone 53-6.7) antibodies; and an APC (allophycocyanin)-conjugated anti-(mouse CD4) antibody (L3T4, clone RM4-5). Flow cytometric analysis was performed with a FACSCalibur cell sorter (Becton Dickinson), and data were analysed with CellQuest software.

Antigen preparation

OMPs (outer-membrane proteins) were extracted from the variant strains of NTHi by using a method published previously [23]. In brief, selected bacterial isolates were cultured in a 250 ml Erlenmeyer flask containing 50 ml of medium and incubated at 37 °C in a rotary shaker/incubator. Bacteria were harvested by centrifugation at 40000 g for 20 min at 4 °C. The pelleted bacteria were resuspended in 10 ml of 10 mmol/l Hepes buffer (pH 7.4) and sonicated four times in a sonicator (MSE Soniprep 150; Sanyo) for 15 s in an ice bath. Intact cells and large debris were removed by centrifugation at 1700 g for 20 min, and the total membrane preparation was harvested from the supernatant by centrifugation at 100000 g for 60 min at 4 °C. The clear gel-like pellet was resuspended in 1 ml of 10 mmol/l Hepes buffer (pH 7.4) and was extracted with an equal volume of 2% (w/v) sodium lauryl sarcosinate in 10 mmol/l Hepes buffer (pH 7.4) for 30 min at room temperature. The detergent-insoluble fraction was harvested by centrifugation at 100000 g for 60 min at 4 °C. OMPs were resuspended in distilled water at a protein concentration of 0.5–3.0 mg/ml. OMPs were solubilized in electrophoresis sample buffer and examined by SDS/PAGE on 12% (w/v) polyacrylamide gels. Fractions were stored at −30 °C for up to 3 weeks.

Silver staining

After electrophoresis was completed, the gel was incubated in a fixative solution containing 40% (v/v) ethanol and 5% (v/v) acetic acid and rocked overnight. According to the method of Tsai and Frasch [24], the fixative solution was replaced with 0.7% (v/v) periodic acid/40% (v/v) ethanol/5% (v/v) acetic acid oxidizer solution for 5 min, followed by 15 rinses in 500 ml of distilled water. The gel was then rocked for 15 min in staining solution [90% (w/v) NH4OH, 0.1 mol/l NaOH and 20% (w/v) AgNO3], washed three times in distilled water and developed by adding a mixture of 50 mg/l citric acid and 37% (w/v) formaldehyde. Development was terminated by adding 5% (v/v) acetic acid. The LPS (lipopolysaccharide) content, including a commercial standard of LPS (E. coli K-235; Sigma) or probable LPS contaminant of extracted OMPs from variable NTHi, was visualized as typical dark-brown-stained bands on a clear gel background by silver staining.

Cell cultures and lymphocyte proliferation assays

Mice were killed 7 days after inoculation. RBC-depleted splenocytes were plated in RPMI 1640 medium supplemented with 2 mmol/l L-glutamine (Sigma), 10% (v/v) heat-inactivated FBS, 50 μmol/l 2-mercaptoethanol (Sigma), 50 mg/ml gentamicin sulphate and 0.1 mol/l Hepes (pH 7.4) (BioFluids) at 1.25×106 RBC-depleted splenocytes/well. These were seeded in triplicate in a 96-well flat-bottomed plate (BD Biosciences). Cells were stimulated with Con A (concanavalin A; 5 μg/ml; Sigma) or with specific antigens (OMPs; 15 μg/ml) for 72 h, or with PBS as a negative control. Samples were incubated at 37 °C in 5% CO2. After 3 days, cultured splenocytes were pulsed with 1 μCi/well [3H]methyl thymidine (Amersham Biosciences) overnight for 16 h. The plates were then harvested on to glass fibre, and the incorporated [3H]methyl thymidine was detected with a TopCount Liquid Scintillation Counter (Packard).

Cytokine determination in cultured splenocytes re-stimulated with strain-specific antigen

RBC-depleted splenocytes were collected as described above and were plated in RPMI 1640 medium supplemented with 2 mmol/l L-glutamine, 10% (v/v) heat inactivated FBS, 50 μmol/l 2-mercaptoethanol, 50 mg/ml gentamicin sulphate and 0.1 mol/l Hepes at 1.25×106 splenocytes/ml, and incubated with or without strain-specific antigen (OMPs, 15 μg/ml) for 60 h at 37 °C in 5% CO2. Supernatants were harvested for detection of IFN-γ and IL-4 by ELISA, according to the manufacturer's protocol (R&D Systems). Briefly, standards or samples were added to microtitre plates coated with a monoclonal antibody specific to the cytokine of interest, and plates were incubated for 2 h. After washing, an HRP (horse-radish peroxidase)-conjugated cytokine-specific antibody was added to each well, and the plates were incubated for 2 h and washed. Substrate solution was then added, plates were incubated for 30 min and the reaction was terminated by the addition of stop solution. Cytokine concentrations were measured with a MRX microplate reader (Dynex Technologies) at 450 nm. The minimum detectable dose of IL-4 is typically <2 pg/ml.

Statistical analysis

Results of disease severity were tested by ANOVA, followed by post-hoc analysis (Scheffe's and Bonferroni's test). Significant differences in mortality between control and experimental groups were examined with the Kaplan–Meier log-rank test. To compare the variation in splenic lymphocytes and the data from ELISA for different groups, one-way ANOVA was performed, followed by Student's t test. Results are means±S.E.M. A P value <0.05 was deemed to be statistically significant.

RESULTS

Screening of the transgenic mice

T1/T2 doubly transgenic mice were used from doubly positive results screened routinely by PCR (Figure 1). The clinical symptoms of these doubly transgenic mice were then determined after various inoculations.

Clinical severity of murine experimental meningitis

Mice inoculated with sterile PBS did not have an altered health status; however, all mice inoculated with the variant strains of NTHi developed various degrees of clinical symptoms at day 1 after injection. The disease severity score gradually reduced at 3 days after infection in CNTHi-infected mice (Figure 2A). In contrast, INTHi-infected mice recovered progressively 7 days following inoculation (Figure 2A). The mortality rates of mice after inoculation with variant strains of NTHi were determined, and no deaths occurred in any of the mice inoculated with either PBS or CNTHi. However, the mortality rate was 23.1% (three out of 13) in mice inoculated with INTHi (Figure 2B), although some surviving mice still displayed residual clinical sequelae, e.g. limb paralysis, weakness, prominent weight loss or irritation. These results indicate that mice inoculated with CNTHi had a lower degree of clinical severity, whereas mice with INTHi meningitis had significant clinical symptoms.

Disease severity and mortality rate after inoculation

Figure 2
Disease severity and mortality rate after inoculation

Mice were injected intracranially with PBS, CNTHi or INTHi and were evaluated at the indicated time points after challenge. (A) The disease severity of the clinical scores was graded as: 0, normal; 1, minimal disease; 2, moderate disease; 3, severe disease; 4, seizure or coma; or 5, died. (B) The mortality of experimental mice (PBS, n=10; CNTHi, n=11; INTHi, n=13) was recorded daily for at least 28 days after inoculation. Results of disease severity were tested by ANOVA, followed by post-hoc analysis. Significant differences in mortality between control and experimental groups were examined by the Kaplan–Meier analysis log-rank test. *P<0.05 compared with the control group; and #P<0.05 when the INTHi group was compared with the CNTHi group.

Figure 2
Disease severity and mortality rate after inoculation

Mice were injected intracranially with PBS, CNTHi or INTHi and were evaluated at the indicated time points after challenge. (A) The disease severity of the clinical scores was graded as: 0, normal; 1, minimal disease; 2, moderate disease; 3, severe disease; 4, seizure or coma; or 5, died. (B) The mortality of experimental mice (PBS, n=10; CNTHi, n=11; INTHi, n=13) was recorded daily for at least 28 days after inoculation. Results of disease severity were tested by ANOVA, followed by post-hoc analysis. Significant differences in mortality between control and experimental groups were examined by the Kaplan–Meier analysis log-rank test. *P<0.05 compared with the control group; and #P<0.05 when the INTHi group was compared with the CNTHi group.

Histopathology of brains during acute and convalescent experimental meningitis

To verify the clinical symptoms in mice inoculated with CNTHi and INTHi, we examined brain slices during the acute (Figures 3A–3C) and convalescent (Figures 3D–3F) stages of disease. Brain slices of mice inoculated with PBS had normal histology without inflammation of meninges at both acute (Figure 3A) and convalescent (Figure 3D) stages. However, meningeal inflammation was observed in all of the brains from mice inoculated with either CNTHi or INTHI during the acute stage, although the degree of meningeal inflammation was less severe in CNTHi-inoculated mice (Figure 3B) compared with INTHi-inoculated mice (Figure 3C).

Histological evaluation in brain during acute and convalescent stage

Figure 3
Histological evaluation in brain during acute and convalescent stage

Histological studies were performed on day 2 and day 21, representing acute (A–C) and convalescent (D–F) stages after inoculation with PBS, CNTHi or INTHi. Brain sections were stained with haematoxylin/eosin. Normal histology was observed in the controls (A and D). During the acute stage, leptomeningeal inflammation (arrowhead) was observed in brains of CNTHi-inoculated mice (B). Apparent leptomeningeal inflammation (arrowhead) was observed in brains of INTHi-inoculated mice (C). During the convalescent stage, mild leptomeningeal enhancement (arrowhead) was observed in brains of mice inoculated with CNTHi (E). Obvious leptomeningeal inflammation (arrowhead) with a wedge-shaped focal infiltration in the parenchyma (arrow) was seen in mice inoculated with INTHi (F). Magnification, ×100.

Figure 3
Histological evaluation in brain during acute and convalescent stage

Histological studies were performed on day 2 and day 21, representing acute (A–C) and convalescent (D–F) stages after inoculation with PBS, CNTHi or INTHi. Brain sections were stained with haematoxylin/eosin. Normal histology was observed in the controls (A and D). During the acute stage, leptomeningeal inflammation (arrowhead) was observed in brains of CNTHi-inoculated mice (B). Apparent leptomeningeal inflammation (arrowhead) was observed in brains of INTHi-inoculated mice (C). During the convalescent stage, mild leptomeningeal enhancement (arrowhead) was observed in brains of mice inoculated with CNTHi (E). Obvious leptomeningeal inflammation (arrowhead) with a wedge-shaped focal infiltration in the parenchyma (arrow) was seen in mice inoculated with INTHi (F). Magnification, ×100.

Histopathology was investigated further during the convalescent stage. Only a slight enhancement of meningeal inflammation without parenchyma infiltration was seen in mice inoculated with CNTHi (Figure 3E); however, marked meningeal inflammation was present in the brains of mice with INTHi meningitis at this stage (Figure 3F). In addition, a wedge-shaped focal inflammatory infiltration of the brain was observed in mice inoculated with INTHi (arrow in Figure 3F). Overall, pathological changes in the brain were more severe in mice inoculated with INTHi compared with CNTHi at either the acute or convalescent stage.

The same experiments were also performed in either wild-type or BALB/c mice and similar results were observed with respect to disease severity compared with NTHi-inoculated T1/T2 doubly transgenic mice in terms of mortality rate, clinical score and histopathology.

Kinetic changes in lymphocyte subsets in spleens of mice after infection

In previous studies [25,26], mice or rats with meningitis had rapid and systemic immunological responses in their peripheral lymphoid organs, including cytokine production and lymphocyte activation. To investigate systemic immunity during infection further, the number and subpopulation of splenic lymphocytes at selected time points (days 3, 7 and 14) after inoculation were analysed (Figure 4). Interestingly, the number of splenic lymphocytes decreased significantly in both INTHi- and CNTHi-inoculated mice at day 3, and this decrease was statistically significant in INTHi- compared with CNTHi-inoculated mice. Nevertheless, there was no statistical difference among the groups at either day 7 or day 14 (Figure 4A).

Distribution of splenic lymphocyte subsets after challenge

Figure 4
Distribution of splenic lymphocyte subsets after challenge

The number (A) and subpopulation [CD4+ T-cells (B), CD8+ T-cells (C) and CD19+ B-cells (D)] of splenic lymphocytes at days 3, 7 and 14 after inoculation were analysed. *P<0.05 compared with the control group; #P<0.05 when the INTHi group was compared with the CNTHi group. Results shown are representative of three mice per time point in each group.

Figure 4
Distribution of splenic lymphocyte subsets after challenge

The number (A) and subpopulation [CD4+ T-cells (B), CD8+ T-cells (C) and CD19+ B-cells (D)] of splenic lymphocytes at days 3, 7 and 14 after inoculation were analysed. *P<0.05 compared with the control group; #P<0.05 when the INTHi group was compared with the CNTHi group. Results shown are representative of three mice per time point in each group.

The percentage of major lymphocyte subsets, including CD4+ T-cells, CD8+ T-cells and CD19+ B-cells, was evaluated further. All RBC-depleted splenocytes from CNTHi- and INTHI-inoculated mice had increased percentages of CD4+ T-cells throughout the entire evaluation period. As expected, an elevated percentage of CD4+ T-cells in spleens of CNTHi-inoculated mice was observed compared with INTHi-inoculated mice throughout the course of the study (Figure 4B). The percentage of CD8+ T-cells increased in INTHi- and CNTHI-inoculated mice at days 3 and 7 after infection (Figure 4C). The role of B-cells in mice with INTHi and CNTHi meningitis was also evaluated. Mice inoculated with CNTHi had a significantly decreased percentage of CD19+ B-cells in spleens during the evaluation period (Figure 4D), whereas INTHi-inoculated mice only had a low percentage of CD19+ B-cells at day 1 after infection (results not shown), but not at days 3, 7 and 14. (Figure 4D).

In addition, a similar trend was observed in the distribution of splenic lymphocyte subsets and antigen-specific lymphocyte proliferative assay and cytokine production after OMP stimulation in the T1/T2 doubly transgenic, wild-type and BALB/c mice (results not shown).

Kinetic changes of Th1 cells (CD4+ hThy1+) and Th2 cells (CD4+ mThy1.1+) in spleens of T1/T2 doubly transgenic mice inoculated with CNTHi and INTHi

At day 3 after infection, CNTHi- and INTHi-inoculated mice had a significantly higher percentage of Th1 cells in the spleen, which was particularly evident in mice with CNTHi meningitis. Interestingly, CNTHi-inoculated mice had a persistent and high percentage of Th1 cells during the first week after infection, suggesting that CNTHi-inoculated mice underwent a dominant Th1 response during this period (Figure 5A). In comparison, a proportional increase in Th2 cells in mice with INTHi meningitis was observed (Figure 5B). At day 14 after infection, Th1 or Th2 cells were barely detectable in spleens from all infected groups. Furthermore, splenocytes were doubly stained with antibodies against hThy1 for T1 cells and DX5 for pan-natural killer cells to rule out the possibility that T1 cells contained other IFN-γ-producing natural killer cells. However, very few cells were observed expressing both hThy1 and DX5 markers simultaneously (results not shown). Moreover, the percentage of DX5-positive cells from infected mice was the same as that in control mice (results not shown).

Development of Th1 and Th2 cells in T1/T2 doubly transgenic mice

Figure 5
Development of Th1 and Th2 cells in T1/T2 doubly transgenic mice

Splenic lymphocytes from transgenic mice were stained with marker-specific antibodies [FITC-conjugated anti-(mouse CD90.1) (mThy1.1), PE-conjugated anti-(human CD90) (hThy1); and APC-conjugated anti-(mouse CD4)] prior to flow cytometric analysis. The numbers in the right upper quadrant of the dot plots represent the percentage of Th1 cells (CD4+hThy1+ T-cells) (A) or Th2 cells (CD4+mThy1.1+ T-cells) (B) at different time points after infection. Similar results are repeated at least three times at the indicated time point in each group.

Figure 5
Development of Th1 and Th2 cells in T1/T2 doubly transgenic mice

Splenic lymphocytes from transgenic mice were stained with marker-specific antibodies [FITC-conjugated anti-(mouse CD90.1) (mThy1.1), PE-conjugated anti-(human CD90) (hThy1); and APC-conjugated anti-(mouse CD4)] prior to flow cytometric analysis. The numbers in the right upper quadrant of the dot plots represent the percentage of Th1 cells (CD4+hThy1+ T-cells) (A) or Th2 cells (CD4+mThy1.1+ T-cells) (B) at different time points after infection. Similar results are repeated at least three times at the indicated time point in each group.

Strain-specific antigen preparation and silver staining of NTHi OMP extracts

As the endotoxic–lipoprotein complex is difficult to dissociate from proteins extracted from Gram-negative bacteria, we were concerned about LPS contamination in the NTHi OMP extracts potentially interfering with our results. Therefore silver staining was used to check for LPS contamination. Various preparations were examined by SDS/PAGE (Figure 6A) and silver staining (Figure 6B), including standard LPS (E. coli K-235) as a positive control, BSA as a negative control and OMPs extracted from either CNTHi or INTHi. The LPS standard was not detectable by SDS/PAGE, but a band was present upon silver staining with a molecular mass of approx. 10–15 kDa. In contrast, the BSA standard was detected by SDS/PAGE with a molecular mass of 67 kDa, but no silver staining was observed in the area corresponding to the LPS standard. Silver staining of OMPs extracted from CNTHi and INTHi revealed some indistinct bands in a similar area as the LPS standard, implying some LPS contamination in the OMP extracts.

SDS/PAGE and silver staining of OMP extracts

Figure 6
SDS/PAGE and silver staining of OMP extracts

OMPs extracted from either CNTHi (8 μg) or INTHi (8 μg) were examined by SDS/PAGE (A) and silver staining (B), along with an LPS standard (8 μg; molecular mass, 10–15 kDa) and BSA (8 μg; molecular mass, 67 kDa). The LPS standard was not visible by SDS/PAGE, but was visible by silver staining (arrow). In contrast, BSA standard was visible by SDS/PAGE with a molecular mass of approx. 67 kDa, but was not detected from silver staining corresponding to the LPS standard. Relatively indistinct staining at a similar area as the LPS standard was observed in OMPs extracted from CNTHi and INTHi. Molecular-mass markers in kDa (Kd) are shown on the right.

Figure 6
SDS/PAGE and silver staining of OMP extracts

OMPs extracted from either CNTHi (8 μg) or INTHi (8 μg) were examined by SDS/PAGE (A) and silver staining (B), along with an LPS standard (8 μg; molecular mass, 10–15 kDa) and BSA (8 μg; molecular mass, 67 kDa). The LPS standard was not visible by SDS/PAGE, but was visible by silver staining (arrow). In contrast, BSA standard was visible by SDS/PAGE with a molecular mass of approx. 67 kDa, but was not detected from silver staining corresponding to the LPS standard. Relatively indistinct staining at a similar area as the LPS standard was observed in OMPs extracted from CNTHi and INTHi. Molecular-mass markers in kDa (Kd) are shown on the right.

Lymphocyte proliferation and cytokine production from cultured splenocytes following strain-specific antigen stimulation

OMPs were extracted from the CNTHi and INTHi strains of NTHi to produce strain-specific antigens to re-stimulate splenocytes from individual mice at day 7 after inoculation. Con A was used as a positive control to stimulate cultured splenocytes and to assess lymphocyte proliferation in all groups, whereas PBS was used as a negative control. There was no significant proliferation of cultured splenocytes with PBS in any of the three groups (Figure 7A). Significant proliferation was detected in cultured splenocytes from control mice stimulated with Con A (Figure 7A). In addition, significant splenic lymphocyte proliferation was observed after stimulation with either Con A or strain-specific antigens in mice inoculated with CNTHi and INTHi (Figure 7A). However, even though silver staining suggested the presence of a potentially small amount of LPS in the OMP extracts from INTHi and CNTHi, there was no significantly proliferative response in the control group after stimulation with these OMPs, suggesting that the presence of any contaminating LPS had a minor effect (Figure 7A). This confirmed that the OMPs were suitable for consideration as strain-specific antigens. Therefore the secretion of IFN-γ and IL-4 from splenocytes stimulated with these strain-specific antigens was investigated. IFN-γ secretion from splenocytes stimulated with strain-specific antigens was significantly elevated in both the INTHi and CNTHi groups (Figure 7B). However, the production of IFN-γ was significantly higher in the CNTHi group compared with the INTHi group (Figure 7B). The production of IL-4 from splenocytes stimulated with strain-specific antigens was only detectable in the CNTHi and INTHi groups (Figure 7C). However, the production of IL-4 was significantly higher in the INTHi group compared with CNTHi group (Figure 7C).

Effect of strain-specific antigens on lymphocyte proliferation and cytokine production in splenocytes

Figure 7
Effect of strain-specific antigens on lymphocyte proliferation and cytokine production in splenocytes

(A) In the splenic lymphocyte proliferation assays, RBC-depleted splenocytes (1.25×106 splenocytes/ml) were isolated from mice on day 7 after PBS, CNTHi or INTHi inoculation. Strain-specific antigens (extracted OMPs at a concentration of 15 μg/ml), Con A (5 μg/ml) and PBS were used to stimulate the cultured splenocytes. (B and C) RBC-depleted splenocytes (1.25×106 splenocytes/ml) were cultured without or with strain-specific antigens (individually extracted OMPs at a concentration of 15 μg/ml) for 60 h. Supernatants were collected and IFN-γ (B) and IL-4 (C) were detected by ELISA. *P<0.05 compared with control mice; #P<0.05 when the INTHi group and CNTHi group were compared. Results shown are representative of triplicate determinations. The minimum detectable dose of IL-4 is typically <2 pg/ml.

Figure 7
Effect of strain-specific antigens on lymphocyte proliferation and cytokine production in splenocytes

(A) In the splenic lymphocyte proliferation assays, RBC-depleted splenocytes (1.25×106 splenocytes/ml) were isolated from mice on day 7 after PBS, CNTHi or INTHi inoculation. Strain-specific antigens (extracted OMPs at a concentration of 15 μg/ml), Con A (5 μg/ml) and PBS were used to stimulate the cultured splenocytes. (B and C) RBC-depleted splenocytes (1.25×106 splenocytes/ml) were cultured without or with strain-specific antigens (individually extracted OMPs at a concentration of 15 μg/ml) for 60 h. Supernatants were collected and IFN-γ (B) and IL-4 (C) were detected by ELISA. *P<0.05 compared with control mice; #P<0.05 when the INTHi group and CNTHi group were compared. Results shown are representative of triplicate determinations. The minimum detectable dose of IL-4 is typically <2 pg/ml.

DISCUSSION

Bacterial meningitis, with its long-term neurological sequelae and high mortality rate, is still a serious issue in public health. In recent years, studies using animal-based models have intensified our understanding of the mechanisms underlying neuronal injury and innate immune responses in bacterial meningitis [1,27]. However, the study of Th1/Th2 immunity in bacterial meningitis in murine models is still limited. In the present study, we have established a murine meningitis model by direct intracranial injection of various NTHi strains (CNTHi and INTHi) and have confirmed that this disease was induced in inoculated mice in terms of clinical symptoms and brain meningeal inflammation.

NTHi is a common commensal organism in the human upper respiratory tract, but can also be an important cause of infectious disease [7,28]. A comprehensive Th1/Th2 paradigm of NTHi infection remains unclear, which is attributed to the complicated adaptive immune response against this micro-organism [12,29,30]. Although many factors are involved in this complex immune response, including host immunity, mechanisms of colonization and invasion, virulence of the bacteria and antigenic variations [3133], these responses have not been investigated fully in experimental NTHi meningitis. In the present study, we report various clinical symptoms and diverse kinetic Th1/Th2 immune responses in mice inoculated with various NTHi strains isolated from either healthy subjects (strain 255) or CSF of patients with NTHi meningitis (strain 450). We have also demonstrated a correlation of pathological changes in brain histology with clinical symptoms. In addition, other strains of CNTHi (strain number 63) and INTHi (strain number 645) were also used in the present study, and similar results were observed with regard to clinical outcomes and kinetic immune responses (results not shown). However, the potential mechanisms associated with the lower virulence of CNTHi and the higher pathogenicity of INTHi still remain to be clarified and require further investigation. In addition, we repeated this experiment with either BALB/c mice or wild-type mice and both sets of mice displayed similar results to those seen in the T1/T2 doubly transgenic mice in terms of disease severity, histopathology and the kinetic trend of splenic lymphocyte subsets (results not shown), suggesting that the expression of these transgenes in the T1/T2 doubly transgenic mice does not interfere with adaptive immune response.

The characteristics of variant bacterial strains contributing to different immunoregulatory effects have been evaluated. A previous study [34] suggested that the difference in expression of Th1/Th2 cytokine profiles induced by two different bacterial strains (S. pneumoniae and Hib) may have an effect on the diverse immunomodulation depending on the aetiology. Our present results suggest that CNTHi mediates a more dominant and long-lasting Th1 response, whereas, in contrast, INTHi triggers a stronger Th2 reaction. Therefore we propose that these variant strains of NTHi isolated from different sources with diverse Th1/Th2 propensity lead to different clinical outcomes. Nevertheless, the potential roles of different lymphocyte subsets are not well known in bacterial meningitis. Interestingly, further investigation of the kinetic changes in lymphocyte subsets in spleens of mice during infection in the present study revealed that the number of lymphocytes in INTHi-inoculated mice was less than those in CNTHi-inoculated mice, and this difference may provide a link between the diversity of the Th1/Th2 paradigm during infection in variant NTHi strains. In addition, we observed that the lowest splenocyte count in INTHi-inoculated mice occurred at day 3 after infection, a time period when these mice had the most severe clinical symptoms. We consider that this kinetic change in splenocyte count may correlate with disease severity. Furthermore, these immunopathogenic responses truly reflect characteristics of the pathogens themselves, since the host factor is restricted by using inbred mice in this experimental model of NTHi meningitis.

Raziuddin et al. [35] have reported increases in CD4+ T-cells, CD8+ T-cells and regulatory T-cells, but less significant increases in the population of B-cells, from patients with bacterial meningitis. A comparable finding was observed in the present study, with the increase in CD4+ T-cells and decrease in CD19+ B-cells being observed in CNTHi-inoculated mice in particular. Foxwell et al. [36] reported that CD8+ T-cells play an essential role in pulmonary clearance in rats after NTHi infection. Our present study supports the proposal that the potential role of CD8+ T-cells in the acute infectious stage may correlate with clearance of pathogens. In brief, the varied distribution of lymphocyte subsets implies that the potential effects of the Th1/Th2 paradigm in NTHi meningitis are multifaceted. The use of T1/T2 doubly transgenic mice can facilitate the direct detection of Th1-type and Th2-type lymphocytes by flow cytometry with surface immunofluorescent staining, even if there is a temporal lag in peak expression of IFN-γ, IL-4 and transgenic surface markers. Nevertheless, these transgenic mice allow us to practically trace the Th1/Th2 paradigm faithfully and are easy to handle [1315]. We also observed that Th1 or Th2 cells in spleens from both infected groups were barely detected at day 14, which implies a kinetic process of CD4 effector T-cells after infection. Recently, the divergent production of Th1/Th2 cytokines in peripheral lymphocytes from healthy subjects or patients with chronic lung disease stimulated by antigens has been reported [37], and an essential comment from this report is that strain-specific immune responses should be evaluated [12].

To dissect further the complicated adaptive immunity occurring during NTHi infection, OMPs were extracted in order to assess antigen-specific splenic lymphocyte re-stimulation. The specific antigenicity of OMPs was confirmed initially in lymphocyte proliferation assays. However, we were concerned about the possible presence of endotoxin contaminants persisting in the extracted OMPs and, thus, silver staining was used to examine the presence of contaminating LPS. A shown by silver staining, a slight residual amount of LPS was present in these extracts; however, there was no significant response to OMPs in the splenic lymphocyte proliferation assay in control mice, implying that the potential effects of contaminating LPS are rather slight due to the high purity of the extracted OMPs from both CNTHi and INTHi. Nevertheless, it is hard to exclude any endotoxin effects in the present study. Therefore ascertaining the exact dose of LPS by quantitative methods, such as the Limulus assay, should be considered. Furthermore, the use of various doses of LPS in proliferative assays may clarify the LPS response, and LPS-inhibition experiments should be done in the future using LPS, LPS+polymyxinB, OMP+polymyxinB and BSA.

Finally, we characterized further the Th1/Th2 polarity in mice infected with variant NTHi and the potential effects of residual LPS in the OMP extracts by analysing the secretion of IFN-γ and IL-4 from cultured splenocytes with strain-specific antigen re-stimulation. We previously inoculated LPS directly as the method to induce meningitis (S. J. Chen, J. T. Huang, H. K. Sytwu and C. C Wang, unpublished work). No death was observed in these mice after LPS intracranial inoculation, but disease symptoms were evident from day 1, became exacerbated up to a maximum on day 2, and subsequently declined rapidly such that, by day 5, mice had recovered totally. Moreover, we found an increased trend in total splenocyte counts in LPS-inoculated mice until day 7 after inoculation. Therefore we selected day 7 after NTHi infection to reflect adaptive immunity after infection. The antigenic diversity and ambiguous immunopathogenic mechanism of NTHi make developing effective vaccines challenging [38], and it is still unclear what potential mechanisms CNTHi, with low pathogenicity, uses to ‘transform’ its initial characteristics into an invasive phenotype (INTHi) with higher virulence. To recognize NTHi as invasive or colonized strains is controversial and, therefore, further studies are required to identify the genomic polymorphisms and antigenic diversities between CNTHi and INTHi.

In conclusion, our results have shown that mice infected with different sources of NTHi have a diverse Th1/Th2 paradigm, which is linked to variation in disease severity. CNTHi mediates a more dominant and long-lasting Th1 response, which might correlate with its characteristically lower virulence, leading to a favourable prognosis. In contrast, INTHi triggers a stronger Th2 reaction, which might correlate with its higher virulence, leading to a worse outcome. These findings from a murine model of NTHi meningitis may provide a potential strategy in developing a novel vaccine against the threat of NTHi infection.

Abbreviations

     
  • APC

    allophycocyanin

  •  
  • Con A

    concanavalin A

  •  
  • CSF

    cerebrospinal fluid

  •  
  • FBS

    fetal bovine serum

  •  
  • Hib

    Haemophilus influenzae type b

  •  
  • IFN-γ

    interferon-γ

  •  
  • IL-4

    interleukin-4

  •  
  • LPS

    lipopolysaccharide

  •  
  • NTHi

    non-typeable Haemophilus influenzae

  •  
  • CNTHi

    colonized NTHi

  •  
  • INTHi

    invasive NTHi

  •  
  • OMP

    outer-membrane protein

  •  
  • PE

    phycoerythrin

  •  
  • RBC

    red blood cell

We thank Min-Chuan Huang, Yen-Lin Wang and Shu-Ying Tsai for their technical assistance, Ming-Hsien Chiang for statistical assistance, Wan-Ling Tzou and Li-Tzu Yeh for their assistance with the silver staining, and Dr Shu-Fen Wu and Dr Jung-Tung Hung for helpful discussions. This work was supported by grants from the National Science Council, Taiwan, Republic of China (NSC-93-2320-B-016-020 to C.-C.W., and NSC-93-3112-B-016-001 to H.-K. S.), and partly by the C. Y. Foundation for Advancement of Education, Sciences and Medicine.

References

References
1
Nathan
B. R.
Scheld
W. M.
New advances in the pathogenesis and pathophysiology of bacterial meningitis
Curr. Infect. Dis. Rep.
2000
, vol. 
2
 (pg. 
332
-
336
)
2
Faden
H.
Meningitis caused by nontypable Haemophilus influenzae in a four-month-old infant
Pediatr. Infect. Dis. J.
1991
, vol. 
10
 (pg. 
254
-
255
)
3
Gratten
M.
Barker
J.
Shann
F.
, et al. 
Non-type b Haemophilus influenzae meningitis
Lancet
1985
, vol. 
i
 (pg. 
1343
-
1344
)
4
Cuthill
S. L.
Farley
M. M.
Donowitz
L. G.
Nontypable Haemophilus influenzae meningitis
Pediatr. Infect. Dis. J.
1999
, vol. 
18
 (pg. 
660
-
662
)
5
O'Neill
J. M.
St Geme
J. W.
III
Cutter
D.
, et al. 
Invasive disease due to nontypeable Haemophilus influenzae among children in Arkansas
J. Clin. Microbiol.
2003
, vol. 
41
 (pg. 
3064
-
3069
)
6
Heath
P. T.
Booy
R.
Azzopardi
H. J.
, et al. 
Non-type b Haemophilus influenzae disease: clinical and epidemiologic characteristics in the Haemophilus influenzae type b vaccine era
Pediatr. Infect. Dis. J.
2001
, vol. 
20
 (pg. 
300
-
305
)
7
Rao
V. K.
Krasan
G. P.
Hendrixson
D. R.
Dawid
S.
St Geme
J. W.
III
Molecular determinants of the pathogenesis of disease due to non-typable Haemophilus influenzae
FEMS Microbiol. Rev.
1999
, vol. 
23
 (pg. 
99
-
129
)
8
Kornelisse
R. F.
Hack
C. E.
Savelkoul
H. F.
, et al. 
Intrathecal production of interleukin-12 and γ-interferon in patients with bacterial meningitis
Infect. Immun.
1997
, vol. 
65
 (pg. 
877
-
881
)
9
Chonmaitree
T.
Baron
S.
Bacteria and viruses induce production of interferon in the cerebrospinal fluid of children with acute meningitis: a study of 57 cases and review
Rev. Infect. Dis.
1991
, vol. 
13
 (pg. 
1061
-
1065
)
10
Torre
D.
Zeroli
C.
Ferraro
G.
, et al. 
Cerebrospinal fluid levels of IL-6 in patients with acute infections of the central nervous system
Scand. J. Infect. Dis.
1992
, vol. 
24
 (pg. 
787
-
791
)
11
Matsuzono
Y.
Narita
M.
Akutsu
Y.
Togashi
T.
Interleukin-6 in cerebrospinal fluid of patients with central nervous system infections
Acta Paediatr.
1995
, vol. 
84
 (pg. 
879
-
883
)
12
Murphy
T. F.
Immunity to nontypeable Haemophilus influenzae: elucidating protective responses
Am. J. Respir. Crit. Care Med.
2003
, vol. 
167
 (pg. 
486
-
487
)
13
Hsieh
S.
Chen
N.
Tarbell
K.
, et al. 
Transgenic mice expressing surface markers for IFN-γ and IL-4 producing cells
Mol. Immunol.
2000
, vol. 
37
 (pg. 
281
-
293
)
14
Sung
H. H.
Juang
J. H.
Lin
Y. C.
, et al. 
Transgenic expression of decoy receptor 3 protects islets from spontaneous and chemical-induced autoimmune destruction in nonobese diabetic mice
J. Exp. Med.
2004
, vol. 
199
 (pg. 
1143
-
1151
)
15
Hung
J. T.
Liao
J. H.
Lin
Y. C.
, et al. 
Immunopathogenic role of TH1 cells in autoimmune diabetes: Evidence from a T1 and T2 doubly transgenic non-obese diabetic mouse model
J. Autoimmun.
2005
, vol. 
25
 (pg. 
181
-
192
)
16
Cauwels
A.
Frei
K.
Sansano
S.
, et al. 
The origin and function of soluble CD14 in experimental bacterial meningitis
J. Immunol.
1999
, vol. 
162
 (pg. 
4762
-
4772
)
17
Marks
M. I.
Ziegler
E. J.
Douglas
H.
Corbeil
L. B.
Braude
A. I.
Induction of immunity against lethal Haemophilus influenzae type b infection by Escherichia coli core lipopolysaccharide
J. Clin. Invest.
1982
, vol. 
69
 (pg. 
742
-
749
)
18
Brodeur
B. R.
Tsang
P. S.
Hamel
J.
Larose
Y.
Montplaisir
S.
Mouse models of infection for Neisseria meningitidis B,2b and Haemophilus influenzae type b diseases
Can. J. Microbiol.
1986
, vol. 
32
 (pg. 
33
-
37
)
19
Wang
C. C.
Siu
L. K.
Chen
M. K.
, et al. 
Use of automated riboprinter and pulsed-field gel electrophoresis for epidemiological studies of invasive Haemophilus influenzae in Taiwan
J. Med. Microbiol.
2001
, vol. 
50
 (pg. 
277
-
283
)
20
Echchannaoui
H.
Frei
K.
Schnell
C.
Leib
S. L.
Zimmerli
W.
Landmann
R.
Toll-like receptor 2-deficient mice are highly susceptible to Streptococcus pneumoniae meningitis because of reduced bacterial clearing and enhanced inflammation
J. Infect. Dis.
2002
, vol. 
186
 (pg. 
798
-
806
)
21
Irazuzta
J. E.
de Courten-Myers
G.
Zemlan
F. P.
Bekkedal
M. Y.
Rossi
J.
III
Serum cleaved Tau protein and neurobehavioral battery of tests as markers of brain injury in experimental bacterial meningitis
Brain Res.
2001
, vol. 
913
 (pg. 
95
-
105
)
22
Wu
S. F.
Liu
T. M.
Lin
Y. C.
, et al. 
Immunomodulatory effect of decoy receptor 3 on the differentiation and function of bone marrow-derived dendritic cells in nonobese diabetic mice: from regulatory mechanism to clinical implication
J. Leukocyte Biol.
2004
, vol. 
75
 (pg. 
293
-
306
)
23
Barenkamp
S. J.
Munson
R. S.
Jr
Granoff
D. M.
, et al. 
Subtyping isolates of Haemophilus influenzae type b by outer-membrane protein profiles
J. Infect. Dis.
1981
, vol. 
143
 (pg. 
668
-
676
)
24
Tsai
C.
Frasch
C. E.
A sensitive silver stain for detecting lipopolysaccharides in polyacrylamide gels
Anal. Biochem.
1982
, vol. 
119
 (pg. 
115
-
119
)
25
Andersson
A.
Dai
W. J.
Di Santo
J. P.
Brombacher
F.
Early IFN-γ production and innate immunity during Listeria monocytogenes infection in the absence of NK cells
J. Immunol.
1998
, vol. 
161
 (pg. 
5600
-
5606
)
26
Diab
A.
Zhu
J.
Lindquist
L.
Wretlind
B.
Link
H.
Bakhiet
M.
Cytokine mRNA profiles during the course of experimental Haemophilus influenzae bacterial meningitis
Clin. Immunol. Immunopathol.
1997
, vol. 
85
 (pg. 
236
-
245
)
27
Scheld
W. M.
Koedel
U.
Nathan
B.
Pfister
H. W.
Pathophysiology of bacterial meningitis: mechanism(s) of neuronal injury
J. Infect. Dis.
2002
, vol. 
186
 (pg. 
S225
-
S233
)
28
Turk
D. C.
The pathogenicity of Haemophilus influenzae
J. Med. Microbiol.
1984
, vol. 
18
 (pg. 
1
-
16
)
29
Weiser
J. N.
Pan
N.
Adaptation of Haemophilus influenzae to acquired and innate humoral immunity based on phase variation of lipopolysaccharide
Mol. Microbiol.
1998
, vol. 
30
 (pg. 
767
-
775
)
30
Kodama
H.
Faden
H.
Harabuchi
Y.
Kataura
A.
Bernstein
J. M.
Brodsky
L.
Cellular immune response of adenoidal and tonsillar lymphocytes to the P6 outer membrane protein of non-typeable Haemophilus influenzae and its relation to otitis media
Acta Otolaryngol.
1999
, vol. 
119
 (pg. 
377
-
383
)
31
Kyd
J.
Cripps
A.
Nontypeable Haemophilus influenzae: challenges in developing a vaccine
J. Biotechnol.
1999
, vol. 
73
 (pg. 
103
-
108
)
32
Foxwell
A. R.
Kyd
J. M.
Cripps
A. W.
Nontypeable Haemophilus influenzae: pathogenesis and prevention
Microbiol. Mol. Biol. Rev.
1998
, vol. 
62
 (pg. 
294
-
308
)
33
Gilsdorf
J. R.
Antigenic diversity and gene polymorphisms in Haemophilus influenzae
Infect. Immun.
1998
, vol. 
66
 (pg. 
5053
-
5059
)
34
Diab
A.
Zhu
J.
Lindquist
L.
Wretlind
B.
Bakhiet
M.
Link
H.
Haemophilus influenzae and Streptococcus pneumoniae induce different intracerebral mRNA cytokine patterns during the course of experimental bacterial meningitis
Clin. Exp. Immunol.
1997
, vol. 
109
 (pg. 
233
-
241
)
35
Raziuddin
S.
el-Awad
M. E.
Mir
N. A.
Bacterial meningitis: T cell activation and immunoregulatory CD4+ T cell subset alteration
J. Allergy Clin. Immunol.
1991
, vol. 
87
 (pg. 
1115
-
1120
)
36
Foxwell
A. R.
Kyd
J. M.
Karupiah
G.
Cripps
A. W.
CD8+ T cells have an essential role in pulmonary clearance of nontypeable Haemophilus influenzae following mucosal immunization
Infect. Immun.
2001
, vol. 
69
 (pg. 
2636
-
2642
)
37
King
P. T.
Hutchinson
P. E.
Johnson
P. D.
Holmes
P. W.
Freezer
N. J.
Holdsworth
S. R.
Adaptive immunity to nontypeable Haemophilus influenzae
Am. J. Respir. Crit. Care Med.
2003
, vol. 
167
 (pg. 
587
-
592
)
38
Poolman
J. T.
Bakaletz
L.
Cripps
A.
, et al. 
Developing a nontypeable Haemophilus influenzae (NTHi) vaccine
Vaccine
2000
, vol. 
19
 (pg. 
S108
-
S115
)