Main

During the initial stage of bacterial and aseptic meningitis, the CSF is characterized by a marked increase in neutrophils. A number of cytokines increase in the CSF, and play critical roles in the local inflammatory responses in the meningitis. Transient elevations of the levels of IL-1, IL-8(1, 2), IL-6(3), TNF-α, and GM-csf(4) have been observed in patients with meningitis. Intracisternal injections of IL-1 and TNF-α in the CSF increase the permeability of the blood-brain barrier(5). The transient elevations of the levels of G-csf and IL-8 in the CSF could lead to the transient accumulation of neutrophils in the CSF by recruiting neutrophils from the peripheral blood at the initial stage of meningitis(6, 7). To explore the coordination of another cytokine with these cytokines, we investigated MIP-1α, which can induce an infiltration of predominantly neutrophils into the spinal cavity.

MIP-1α is a member of a family of β-chemokines, and the infiltration of neutrophils MIP-1α induces is usually followed by an influx of monocytes(8–10). MIP-1α also induces fever(11), a loss of appetite(11) and wound healing(12)in vivo; it causes a significant oxidative burst(8) and inhibits the proliferation of hematopoietic stem cells(9) in vitro. This cytokine is mainly produced by activated monocytes/macrophages. The MIP-1α levesl in sera, synovial fluid and the bronchoalveolar lavage fluid are increased in some inflammatory diseases(13,14). This chemokine in the CSF has been reported to be increased in experimental bacterial meningitis using animal models(5, 10, 15). To our knowledge, however, little is known about the kinetics of the MIP-1α levels in the CSF of human meningitis and its cellular sources. To analyze the presence and role of MIP-1α in the CNS in children with meningitis, we examined the MIP-1α levels in the CSF of such patients. Our data demonstrated that MIP-1α levels are markedly elevated in the CSF only at the initial symptomatic stage of meningitis. We also studied which cells produce MIP-1α in the CSF of meningitis patients using RT-PCR-assisted amplification and cellular cultures.

METHODS

Study subjects. We studied 10 patients with bacterial meningitis, 22 with aseptic meningitis and 19 control subjects without meningitis (Table 1). They ranged in age from 14 d to 14 y and were admitted to our hospital during the period from June 1991 to June 1996. The following criteria were used for the diagnosis of meningitis: 1) all of these symptoms-fever, headache, vomiting, and neck stiffness: 2) cell counts greater than 35/μL in the CSF; 3) in bacteriologic studies, bacteria detected in the CSF in bacterial meningitis and sterile CSF found in aseptic meningitis. In the bacterial meningitis group, five different bacteria were detected by the bacterial cultures of the CSF: Streptococcus pneumoniae in four patients, Escherichia coli in two,Hemophilus influenzae in two, and Streptococcus agalactiae and Listeria monocytogenes in one patient each. The pathogens for aseptic meningitis were determined using neutralization tests and viral cultures as described previously(6); mumps virus in six patients, echovirus 30 in four echovirus 9 in two, and uncertain causes in 10 individuals. The first day of illness was determined as the day when all of the symptoms of meningitis had first occurred. The symptomatic stage was designated as the period when any of the meningeal symptoms or signs existed, and the recovery stage as the period after all of the symptoms and signs disappeared. The control subjects fit the following criteria: 1) cell counts less than 5/μL in the CSF; 2) sterile CSF found in bacteriologic studies. The non-meningitis patients included three with epilepsy, four with febrile convulsions, two with leukemia, and 10 with both fever and the meningeal signs.

Table 1 Clinical characteristics of study subjects
Full size table

Sample collection. We obtained institutional approval from the responsible committee and a full informed consent from each patients' parents. To reduce the patients' risk throughout this study, CSF samples were obtained only when lumbar punctures were clinically indicated as previously described(16). The CSF samples were immediately spun down at 150× g for 10 min, and the supernatant was stored at -80°C; the cells were used for the subsequent procedures. Serum samples were collected at the same time and stored at -80°C.

Cell culture. MNCs (1 × 106/mL) separated from the CSF cells were incubated in RPMI 1640 medium supplemented with 10% FCS for 24 h as previously described(16). The culture supernatant was harvested by centrifugation at 400 × g for 10 min and stored at -80°C.

Assays for MIP-1α and other cytokines. The concentrations of MIP-1α were measured in duplicate by an ELISA (Amersham, Little Chalfont, UK). The assay was based on the double ligand method(17). Briefly, standard recombinant human MIP-1α or test samples were dispensed into a plate precoated with a murine anti-MIP-1α MAb. The MIP-1α was then sandwiched with a polyclonal antibody against MIP-1α conjugated with horseradish peroxidase. The lower limits of detection were 6 ng/L for CSF or cell culture supernatant, and 10 ng/L for serum.

The concentrations of IL-6 and IL-8 were assayed by an ELISA (Toray-Fuji Bionics, Inc., Tokyo, Japan). The G-csf levels were quantified by a chemiluminescent enzyme immunoassay (Chugai Pharmaceutical Co., Tokyo) that was more sensitive(18) than the method used in our previous report(6). The minimal detectable concentrations were 25 ng/L (IL-6), 12 ng/L (IL-8), and 1 ng/L (G-csf). These assay systems did not react with other cytokines.

Detection of mRNA by RT-PCR. The RT-PCR was done as previously described(16, 19). Briefly, first-strand cDNA was synthesized from total RNA collected from 1 × 106 of fresh cells in the CSF and amplified by PCR. The oligonucleotide primers for human MIP-1α were designed as 5′-CTG/CTG/CTT/CAG/CTA/CAC/CT-3′(sense) and 5′-ACA/GTG/TGG/CTG/TTT/GGC/AA-3′ (anti-sense)(20). The β-actin primers were made as previously described(21). The reaction consisted of 29 cycles(94°C for 1 min, 55°C for 1 min, 72°C for 1 min) for MIP-1α or 21 cycles for β-actin. The PCR products were electrophoresed and transferred onto nylon filters, and then hybridized with the following probes. An MIP-1α oligonucleotide probe, 5′-TTC/CTA/ACC/AAG/CGA/AGC/CGG/CAG/GTC/TGT-3′(20), was labeled with 32P by the 5′-end-labeling method(21). A β-actin probe, 0.7 kb of the EcoRI-BamHI cDNA fragment, was labeled with32 P by the random priming method, as previously described(22).

Statistical analysis. The results are expressed as mean± SD unless otherwise noted. Variables were transformed to common logarithms before some statistical analyses. The probability of a significant difference was determined using the Mann-Whitney U test. The relationships between the MIP-1α levels and other indices were estimated by Pearson's correlation coefficients. Differences were considered significant when the two-tailed p value was less than 0.05.

RESULTS

The gender and age of the patients were not significantly different among the groups. The counts of total leukocytes, MNCs, and neutrophils in the CSF samples of the children with meningitis were significantly higher than those of the controls without meningitis (Table 1).

MIP-1α levels in the CSF. We examined the MIP-1α levels in the CSF of the children with and without meningitis. At the symptomatic stage, MIP-1α was detected in the CSF of all 10 patients with bacterial meningitis (17 of 18 samples) and in 17 of the 22 patients with aseptic meningitis (21 of 30 samples). The mean MIP-1α levels in the CSF at the symptomatic stage were 289.9 ± 270.7 ng/L (the maximum value, 1499.0 ng/L) in the bacterial meningitis group and 16.1 ± 12.5 ng/L in the aseptic meningitis group. These values were significantly different(p = 0.0008) from each other, and significantly higher (p< 0.0001) than those of the control patients without meningitis. None of the samples from the control children contained detectable levels of MIP-1α (<6 ng/L). At the recovery stage, MIP-1α was detected in the CSF of only 3 of 11 samples from the bacterial meningitis group and 5 of 17 from the aseptic meningitis group. These levels were significantly lower than those at the symptomatic stage (bacterial meningitis, p = 0.0013; aseptic meningitis, p = 0.0004).

Figure 1 illustrates time-dependent changes in the MIP-1α levels in the CSF. The MIP-1α levels decreased with the improvement of clinical symptoms. These levels dropped rapidly in the aseptic meningitis group and decreased more slowly in the bacterial meningitis group.

Figure 1
figure 1

The time course of MIP-1α levels in the CSF of patients with meningitis. Squares, bacterial meningitis (n= 29); circles, aseptic meningitis (n = 47); closed symbols, symptomatic stage; open symbols, recovery stage of meningitis.

Full size image

MIP-1α levels in sera. MIP-1α was detected in 13 of 28 serum samples. Figure 2 shows the MIP-1α levels in the CSF and sera that were collected at the same time. Because the lower detectable limits of MIP-1α were different between the CSF samples and the serum samples, we analyzed only 12 paired samples with over 10 ng/L MIP-1α in both CSF and sera. The MIP-1α levels were significantly higher in the CSF than in the sera (p = 0.0464).

Figure 2
figure 2

MIP-1α levels in the CSF samples and serum samples collected at the same time (p = 0.0464, n = 12).Squares, bacterial meningitis; circles, aseptic meningitis(n = 6 each).

Full size image

Correlations between MIP-1α levels and clinical indices. High levels of neutrophil counts in the CSF were found within 5 d of the onset of meningitis. The neutrophil counts decreased in parallel with the improvement of clinical symptoms, as previously shown(6). At the initial stage of meningitis, the predominant cells in the CSF were neutrophils, but within a few days of onset, the majority of cells shifted to MNCs, which consisted of monocytes and lymphocytes (data not shown). We analyzed the relationships between the MIP-1α levels in the CSF and some clinical data in both the bacterial and aseptic meningitis groups. Figure 3 shows the significant correlation between the MIP-1α levels and neutrophil counts in the CSF (r = 0.750, p < 0.0001, n = 76); an especially strong correlation was observed in the group with bacterial meningitis (r = 0.868, n = 29). The mean value of the MIP-1α (log ng/L)/neutrophil counts (log/L) ratio was 1.003 ± 0.576, and there were no significant differences in this ratio between the two types of meningitis or between symptomatic and recovery stages.

Figure 3
figure 3

Correlations between MIP-1α levels and neutrophil counts in the CSF of patients with meningitis (r = 0.750,p < 0.0001). Squares, bacterial meningitis (n= 29); circles, aseptic meningitis (n = 47); closed symbols, symptomatic stage; open symbols, recovery stage of meningitis; solid line, linear regression line. The mean value of the MIP-1α (log ng/L)/neutrophil counts (log/L) ratio was 1.003 ± 0.576, and there were no significant differences in this ratio between the two types of meningitis or between the symptomatic and recovery stages.

Full size image

There were also significant correlations between the CSF MIP-1α levels and either the total cell counts (r = 0.682, p < 0.0001) or the MNC counts (r = 0.478, p < 0.0001) in the CSF. However, the MIP-1α concentrations in the CSF did not correlate with the grade or duration of fever, headache, or vomiting.

Correlations between MIP-1α levels and other cytokines in the CSF. We analyzed the relationships between the levels of MIP-1α and several cytokines in the CSF of the patients with bacterial and aseptic meningitis. All of the cytokines measured were increased at the symptomatic stage, and decreased with the improvement of meningeal symptoms.Figure 4 shows the significant relationships between the MIP-1α levels and both the G-csf (r = 0.666, p < 0.0001, n = 42) and IL-8 (r = 0.645, p < 0.0001, n = 43) levels in the CSF of the meningitis patients. There were especially strong correlations in the patients with bacterial meningitis(r = 0.856 in G-csf. r = 0.769 in IL-8; n = 9). The MIP-1α levels were also significantly correlated with the IL-6 levels in the CSF of the meningitis patients (r = 0.774, p< 0.0001, n = 30).

Figure 4
figure 4

Correlations between the levels of MIP-1α and G-csf (A; r = 0.666, p < 0.0001, n= 42) or IL-8 (B; r = 0.645, p < 0.0001,n = 43) in the CSF of the meningitis patients. Squares, bacterial meningitis; circles, aseptic meningitis; closed symbols, symptomatic stage; open symbols, recovery stage of meningitis; solid lines, linear regression lines.

Full size image

Production of MIP-1α by CSF cells. The expression of MIP-1α mRNA in fresh CSF cells in meningitis was examined by an RT-PCR technique. The CSF cells were collected on the 1st to 4th d from the onset of the illness. Figure 5A shows that the human MIP-1α-specific hybridization signals were found in seven of nine samples. Two samples (lanes 3 and 7) were from children with bacterial meningitis and the other seven were from children with aseptic meningitis. High levels of expression of the MIP-1α mRNA were shown in the two patients with bacterial meningitis. Figure 5B demonstrates markedly high levels of MIP-1α in the culture-supernatant from patients with both types of meningitis (1917 ± 2887 ng/L).

Figure 5
figure 5

Expression of MIP-1α mRNA and its protein by cells in the CSF of patients with meningitis. (A) The signals for human MIP-1α and β-actin mRNA in fresh CSF cells detected by RT-PCR amplification and hybridization. Lanes 3 and 7, bacterial meningitis; lanes 1, 2, 4-6, 8, and 9, aseptic meningitis.(B) The MIP-1α protein levels produced by cultures of CSF MNCs from meningitis patients.

Full size image

DISCUSSION

Our study revealed that the MIP-1α levels in the CSF were significantly increased at the stage in which any meningeal symptoms were manifested in both bacterial and aseptic meningitis, and decreased with the improvement of meningeal signs or symptoms. The neutrophil counts were also increased in the CSF at the symptomatic stage, and then decreased with the drop in the MIP-1α levels. We observed significant relationships between the CSF MIP-1α levels and CSF neutrophil counts. Elevated levels of G-csf(6, 7) and IL-8(1, 2) in the CSF have been reported to lead to the accumulation of neutrophils in the foci of inflammation. In the present patient series, there were significant relationships between the levels of MIP-1α and the levels of both G-csf and IL-8. MIP-1α has biologic activities that can activate the functions of neutrophils and induce neutrophil chemotaxis(8). Saukkonen et al.(10) have shown that intracisternal injections of MIP-1α induce the elevation of blood-brain barrier permeability and leukocytosis in the CSF in a rabbit model. Mice with a disruption of the gene encoding MIP-1α were resistant to inflammatory reactions in experimental myocarditis and pneumonitis models(23). We therefore suspect that elevated levels of MIP-1α in the CSF can induce the accumulation of neutrophils from the peripheral blood into the spinal cavity, in coordination with the functions of G-csf and IL-8. The presence of MIP-1α might modify the sequelae of meningitis.

The predominant cell type in the CSF was neutrophils at the initial stage of meningitis, but it then shifted to MNC within a few days of the onset. MIP-1α has attractant activities for monocytes and lymphocytes(24), in addition to neutrophils, which have MIP-1α receptors(25, 26). MIP-1α can induce the migration of monocytes across the blood-brain barrier in experimental bacterial meningitis(10). We observed that an elevation in the number of MNCs consisting of monocytes and lymphocytes followed the neutrophil accumulation. Increases in IL-8 and M-csf in patients with meningitis have been found by us (our unpublished data) and other investigators(1, 2). We thus speculate that MIP-1α in the CSF also plays a role in the accumulation of monocytes in coordination with M-csf and in the accumulation of lymphocytes in coordination with IL-8 at the local inflammatory foci in human meningitis. Furthermore, because MIP-1α can induce the proliferation of Schwann cells in vitro(27), it may act, in part, to heal the injury in meningitis.

We next studied how MIP-1α increased in the CSF in the patients with bacterial and aseptic meningitis. The main cellular source of MIP-1α is monocytes/macrophages in human beings and animals(8, 13, 22, 28). MIP-1α can also be produced by B and T lymphocytes(28–30), neutrophils(31), eosinophils(32), and endothelial cells(33) in humans, and by fibroblasts(34) and astrocytes(15) in mice. Although the expression of the mRNA of MIP-1α has been reported in synovial fluid MNCs of patients with rheumatoid arthritis(13, 24) and in the brains of animals with experimental meningoencephalitis(35), little is known about where MIP-1α is produced and what cells release MIP-1α in human meningitis. Our data showed that the MIP-1α levels in the CSF are remarkably high at the initial stage, and that its levels in sera are significantly lower than those in the CSF. The serum MIP-1α levels did not fluctuate during the progression of the illness. These results suggest that the elevated levels of MIP-1α in the CSF are not due to a switch-over from the blood, but rather to its production in the foci of inflammation i.e., in the spinal cavity. We detected MIP-1α mRNA in the CSF cells using PCR amplification, and high levels of MIP-1α protein in the culture media from MNCs in the CSF in vitro. These results provide direct evidence that MIP-1α in the CSF is produced, at least in part, by MNCs in the CSF of both bacterial and aseptic meningitis.

The production of MIP-1α is stimulated by lipopolysaccharide(5, 31, 32), TNF-α(13) and GM-csf(31). Lipopolysaccharide is one of the bacterial components; GM-csf is increased in the CSF in both types of meningitis, and TNF-α is increased in bacterial meningitis. In our study, the MIP-1α levels in the CSF were significantly higher and its mRNA was expressed more strongly in the bacterial meningitis group than in the aseptic meningitis group. We suspect that bacterial lipopolysaccharide may directly stimulate the production of MIP-1α in bacterial meningitis, and that a cytokine such as GM-csf may lead to the induction of MIP-1α production in not only bacterial but also aseptic meningitis. Because the MIP-1α levels in the CSF at the symptomatic stage were significantly different between both forms of meningitis, its levels in the early days are thought to be a useful index of whether a case is of bacterial or aseptic meningitis.

In summary, MIP-1α is produced in the CSF during the initial stage of meningitis. In the CSF of patients with meningitis, MIP-1α may act as a major cytokine that induces the accumulation of neutrophils in coordination with G-csf and IL-8, and may contribute to the infiltration of monocytes and lymphocytes.