Background
The field of function of CXCL13 has been constantly growing since its discovery in 1998 [
1]. Initially, the essential role of CXCL13 was seen in the establishment and maintenance of lymphoid tissue microarchitecture [
2]. Accordingly, CXCL13 deficient mice fail to develop lymph nodes [
3], and B-cell homing to lymph node follicles requires CXCL13 and its exclusive receptor CXCR5 [
4]. Some years later, evidence for a role in the formation of ectopic lymphoid tissue in chronic inflammation such as multiple sclerosis or rheumatoid arthritis was also found [
5,
6]. Finally, the detection of CXCL13 expression in
Helicobacter pylori gastritis [
7], pulmonary tuberculosis [
8] or
Bartonella henselae infection [
9] suggested a role of this chemokine in chronic bacterial infections as well. Its influence on leucocyte migration to the site of infection, however, has not been evaluated so far.
Recently, we and others observed a strong up-regulation of CXCL13 expression in an acute bacterial infection, in Lyme neuroborreliosis (LNB) [
10,
11]. In LNB, the spirochete
Borrelia burgdorferi (
B.b.) invades the cerebrospinal fluid (CSF) [
12]. The host immune system reacts to the invading spirochetes with a local inflammation, leading to an intrathecal accumulation of leucocytes. A hallmark of this CSF-pleocytosis in LNB is the accumulation of activated B cells and plasma cells. The percentage of B cells in the CSF of LNB patients reaches up to 80%, clearly exceeding other CNS infections [
13]. B cells show a substantial migration only to very few chemokines, namely, CCL19, CCL21, CXCL12, and CXCL13 [
14].
In previous studies, we measured high concentrations of CXCL13 in the CSF of patients with LNB, even before the intrathecal production of
B.b.-specific antibodies has started [
10,
15]. Cell culture experiments have shown, that PBMC produce CXCL13 in response to an incubation with
B.b. through the interaction of the TLR2 receptor of the innate immune system with spirochete outer surface proteins [
16]. This
in vitro study is further supported by findings in the rhesus monkey model of LNB, where the CXCL13 expression at the spinal nerve roots correlated with the spirochete load and resident immune cells have been identified as source of this chemokine [
11,
17]. Based on these data, we speculated that
in vivo, the high intrathecal concentration of CXCL13 in LNB patients directs the B cells to the CSF, leading to the observed B cell enriched CSF pleocytosis [
12].
To further assess the role of CXCL13 for the B cell rich infiltrate in the CSF of LNB patients, we (1) examined the chemotactic activity of CSF samples from patients with LNB on human B cells, compared with that of patients with a non-inflammatory CNS diseases (NIND), and neurosyphilis (NS) as another spirochete CNS disease in a chemotaxis assay, (2) determined the concentrations of CXCL13 and other B cell attracting chemokines in CSF/serum pairs of the three patient groups, and (3) tried to elucidate the contribution of these chemokines to the chemotactic activity of LNB by using specific neutralizing antibodies.
Methods
B-cell isolation and stimulation
Human peripheral blood mononuclear cells were extracted from venous blood samples from a healthy male donor using ficoll gradient centrifugation. Subsequent untouched B cells were isolated by Midi Macs System
® (Miltenyi Biotec, Bergisch-Gladbach, Germany). The purity of the (CD19
+) B cells, examined repetitively before and during the experiments by a FACS analysis, was always higher than 98.4%. More than 95% of isolated cells were shown by the Trypan blue exclusion test to be viable. In order to stimulate the B cells to increase their migratory ability, they were incubated for 16 hours at 37°C and 5% CO
2 in RPMI1640 medium supplemented with 10%FCS, 100 units penicillin and 100 μg streptomycin/ml, 20 ng/ml IL4 and 1 μg/ml CD40L according to [
14]. Thereafter, the cells were pelleted again and diluted to a working concentration of 2.6 × 10^6 cells/ml for the migration assays and 2.1 × 10^6 cells/ml for the inhibition assays.
CSF/blood samples
Blood was drawn and lumbar puncture was performed for diagnostic purposes after the patient's informed consent was obtained. All samples were frozen at -30°C. Paired CSF and blood samples were obtained from the following groups:
(i) 14 patients (9 males) with non-inflammatory CNS disease (NIND). The mean cell count was 1.8 [0.3-4.7] cells/μl, and the mean CSF/serum albumin ratio was 8.7 [3.3-31.0]. The mean age was 64.7 [48-74] years. Three of these patients suffered from chronic pain, two others had an epileptic seizure, and the further patients were diagnosed amyotrophic lateral sclerosis, vascular compression of the brain stem, Parkinson's disease, transverse myelitis, Guillain-Barré syndrome, multiple sclerosis, herniated vertebral disc, somatoform disorder or progressive ataxia.
(ii) 18 patients (11 males) with acute (duration of symptoms < 6 months) Lyme neuroborreliosis (LNB) before initiation of therapy. The mean cell count was 146 [1-600] cells/μl (for details, please refer to table
1), the mean CSF/serum albumin ratio was 26.0 [3.9-43.9]. The mean age was 57.8 [15-72] years. The diagnosis of LNB was based on the following criteria: 1) typical clinical picture (e.g. meningoradiculitis, cranial neuritis or meningitis) 2) lymphocytic CSF pleocytosis, and 3) increased intrathecal
B.b.-specific antibody production. Detailed data concerning basic CSF findings as well as the
B.b.-specific antibody index and the percentage of plasma cells are given in table
1.
Table 1
Clinical data and basic CSF parameters of the LNB patients
1
| f | 69 | BS | 53 | 15 | 43.9 | 1.7 |
2
| m | 15 | CNP, M | 600 | 0 | 5.4 | 4 |
3
| f | 70 | BS | 90 | 0 | 4.83 | n.d.+ |
4
| m | 62 | BS, CNP | 246 | 16 | 39.7 | 7.3 |
5
| f | 69 | BS | 109 | 5 | 15.8 | 13.3 |
6
| f | 61 | BS | 217 | 1 | 37.9 | 40.7 |
7
| m | 66 | BS, CNP | 16 | 0 | 8.9 | <1.5# |
8
| m | 67 | BS, M | 52 | 20 | 10.3 | 34 |
9
| m | 61 | BS | 1* | 8 | 30.3 | 10.9 |
10
| m | 72 | CNP | 166 | 3 | 21.6 | 62.7 |
11
| m | 59 | BS | 100 | 0 | 23.4 | 2.8 |
12
| m | 38 | M | 117 | 11 | 26 | 168 |
13
| m | 55 | BS, CNP | 368 | n.d. | 19.1 | 38.6 |
14
| m | 50 | CNP | 12 | 0 | 4.5 | 7.7 |
15
| f | 46 | BS | 14 | 0 | 3.9 | 1.8 |
16
| m | 59 | BS | 30 | 0 | 13.2 | 2.5 |
17
| f | 59 | BS | 255 | 5 | 33.2 | 17.8 |
18
| f | 62 | CNP, M | 180 | 1 | 10.5 | 4.3 |
(iii) 14 patients (13 males) with neurosyphilis (NS). The mean cell count was 17.4 [5-45] cells/μl and the mean CSF/serum albumin ratio was 5.9 [3.2-12.5]. The mean age was 31.8 [19-48] years. The diagnosis was based on serological proof of a syphilitic infection (positive
Treponema particle agglutination and fluorescent treponemal antibody absorbed tests) and an elevated CSF white blood cell count according to the criteria described by Marra et al. [
18]. All patients with syphilis were in the second stage of disease (secondary syphilis) and negative for the human immunodeficiency virus (HIV).
Migration assay
The migration assay was performed in a disposable 96-well system (ChemoTx, Neuroprobe) with polycarbonate membrane filters (5 μm pore). 30 μl of the CSF samples were added to the lower chamber and 50 μl B cell solution was added onto the membrane. The chamber was incubated for 1 h at 37° in humidified air with 5% CO2. After incubation, the migrated cells in the lower chamber were counted in a Fuchs-Rosenthal-chamber.
For better comparison of the results from different experiments, a migration index was calculated. We used medium as negative control value and 500 ng/ml rhCXCL13 (R&D Systems, Minneapolis, MN, USA) as a positive control.
The results are presented in per cent. A MI of 100% implicates, that the sample had the same chemotactic effect as 500 ng rhCXCL13.
FACS analysis
Anti-human monoclonal antibodies against CD19 (PC5, J3-119, IgG1) and CD27 (PE, 1A4, IgG1), obtained by Immunotech (Marseille Cedex, France) were used. For flow cytometry 50 μl cell suspension (2 × 106/ml) of isolated B cells and either B cells, which did migrate to rhCXCL13 or LNB CSF, and those that have not migrated were incubated with 5 μl of CD19 and 10 μl of CD27 each for 15 minutes at room temperature in the dark. 1 ml PeliLyse lysing reagent (Hiss Diagnostics GmbH, Freiburg, Germany) was added to each of the tubes and incubated for another 10 minutes in the dark at room temperature. The cells were washed and resuspended in 200 μl PBS (Merck KGaA, Darmstadt, Germany) and afterwards immediately analysed.
Analysis was performed on a FACS Canto II flow cytometer (Becton Dickinson, San Jose, USA) using FACSDiva software. 10,000 events were acquired. Data were displayed as two colour plots (CD19/CD27). The data of migrated B cells were compared with those of non migrated B cells.
ELISA
From the CSF/serum samples used for the migration assays, sufficient volume for the quantitative measurement of the chemokines CXCL12, CCL19 and CCL21 was available from 12 NIND (mean cell count 1.89/μl, mean total CSF protein 553 mg/l, mean age 65.2 years, and 58% male), 14 NB (mean cell count 375.6/μl, CSF protein 1,186 mg/l, mean age 56.4 years, and 64% male) and all NS patients. CXCL13 levels in CSF and serum samples were measured in all samples used in the migration assays. A part of the samples have already been analysed for CXCL13 before [
10,
16]. The measurements were done by ELISA (R&D) according to the recommendations of the manufacturer.
Inhibition assay
To investigate the chemotactic impact of CXCL13, CXCL12 and CCL19 on the human B cells, the respective chemokines in the CSF samples from LNB patients were blocked with neutralizing antibodies. For the isolated blockade of CXCL13 activity, 10 μg/ml of the neutralizing polyclonal CXCL13 antibody or an isotype control antibody (10 μg/ml) (all R&D) was used. For triple inhibition assays, we added additional 10 μg/ml polyclonal CCL19 and 50 μg/ml polyclonal CXCL12 antibody or analogous amounts of isotype control antibodies. In pre-tests, we could show that the CXCL13 and CCL19 antibodies inhibit >99% of the chemotactic effect of 50 ng/ml rhCXCL13 and rhCCL19, respectively. In contrast, though testing several available neutralizing antibodies in different concentrations, the blockade of CXCL12 activity only reached 42%. Cross reactivity between all applied antibodies was tested and not observed. All samples were preincubated on a shaker with the respective antibody for 30 min at RT. As negative control, we used a mixture from 12 CSF samples from NIND patients as they would represent the baseline of migratory activity in vivo, 500 ng/ml rhCXCL13 served as positive control. To perform the assay, sufficient CSF volume from 12 patients was available, but only 9 samples were tested as their MI was higher than all NIND samples. In addition, 3 LNB CSF samples were either heated at 60°C for 10 minutes to denature protein structures, filtrated for 90 min at 14 g using a membrane with a 3 kDa Nominal Molecular Weight Limit (Millipore), or the B cells were preincubated for 90 min with 5 μg/ml PTx. All experiments were performed in duplicate.
Statistics
In order to assess statistical significant differences of the chemotactic activity of CSF samples obtained from patients with NIND, LNB, or NS on isolated B cells we used one way analysis of variance (ANOVA) and Bonferroni adjustment. This statistical test was also used to measure the significant differences between the chemokine concentrations of these groups in the ELISA assays. The paired t-test was used to measure the differences in the migration behaviour of B cells between antibody- and isobody-control groups in the inhibition assays. The data were expressed as mean ± SD.
Discussion
The mechanisms underlying the B cell rich CSF pleocytosis in LNB are still undefined. Here we demonstrate that CSF of LNB patients is chemotactic to peripheral blood human B cells, and that CXCL13 is a major regulator of B cell recruitment in acute LNB.
The cells that migrated towards the CSF of LNB patients in our
in vitro setting were predominantly CD27
+ B cells. The surface marker CD27
+ on B cells indicates, that these cells are activated and can produce 5- to 100-fold more immunoglobulins than CD27
- cells [
22]. Earlier studies on the immune response during the course of
B.b. meningoradiculitis revealed, that large numbers of B cells in the CSF and a prominent intrathecal production of antigens are characteristic for LNB [
13], suggesting that these B-cells are from the (CD27
+) mature type. The exact rate of CD27
+ cells in the CSF in LNB, however, has not been determined so far. Results from patients with other neuroinflammatory disorders like multiple sclerosis, viral infections, or neurosyphilis showed that 85% of the total B cell population in the CSF are CD27
+, in contrast to 31% in the peripheral blood [
23]. Taken these studies together, CD27
+ B cells appear to be the main migrating B cell population in neuroinflammation, both
in vivo and in our
in vitro experimental setup. Therefore, our
in vitro model seems to be a suitable tool to search for the responsible chemotactic agent for B cell immigration in LNB.
This chemotactic agent should fulfil four criteria: a) it has to attract B cells, b) it has to be elevated in the CSF of LNB patients and there must be a gradient over the blood-CSF barrier, c) there should be an association between the gradient and the intensity of B-cell-migration, and d) its neutralization should abrogate the observed B cell migration to LNB CSF. The four known major B cell attracting chemokines are CCL19, CCL21, CXCL12, and CXCL13, as shown by Brandes et al. [
14]. From recent studies, it is known that CXCL13 is elevated in the CSF, but not the blood of LNB patients [
10,
16]. In the current study, we further showed that rhCXCL13, when used in concentrations as found in LNB CSF samples, is chemotactic to B cells. Data concerning the concentration of other B-cell attracting chemokines in LNB CSF are rarely found in the literature. Only CXCL12 was explicitly measured in a small collective of 6 LNB patients with a mean of 24.1 ng/ml in the CSF of LNB patients and 5.5 ng/ml in NIND patients [
24,
25]. Elevated levels of CCL19 in the CSF of patients with various infectious diseases (including LNB patients) were determined by two study groups [
26,
27] with a mean value between 173 - 250 pg/ml compared to 26 - 62 pg/ml in NIND patients. CCL21 in contrast was not found at measurable concentrations in any of these studies. Exact data for LNB patients or serum concentrations to calculate the gradient of CCL19 or CCL21 were not available. Therefore, we measured the concentrations of CCL19, CCL21, CXCL12 and CXCL13 in CSF/serum pairs of patients with LNB. The most prominent results were found as expected for CXCL13 [
10,
16] with a more than 300 fold difference in the CSF between LNB and NIND patients and a mean CSR of approximately 150.
All of the LNB patients had not been treated with antibiotics before the lumbar puncture. One of the NS patients received an antibiotic therapy for two weeks before CSF sampling and showed the lowest CXCL13 CSF value of all NS patients. This observation fits perfectly to recent studies demonstrating that (I) viable spirochetes are needed for the production of CXCL13 in monocytic cells [
16] and (II) CSF CXCL13 concentrations rapidly decreases under antibiotic therapy in LNB patients [
10].
The difference between the patient groups is less impressive for CCL19 but still compatible with a functional chemotactic role: The mean CSF concentration is 8 fold higher than in NIND patients and the mean CSR equals 5.
The results obtained for CXCL12 are lower than those from Pashenkov et al. [
24], but the high constitutive expression of CXCL12 in the CNS and the lack of an elevated CSR, arguing against a functional role for B cell immigration, was confirmed in our study. The concentration of CCL21 in the CSF, as expected from previous studies, is hardly measurable in any patient group and a functional role therefore very unlikely. In view of these results, CXCL13, but also CCL19 and less likely CXCL12 were identified as putative chemotactic agents in LNB. This was further underlined by the association analysis. High CXCL13 and CCL19, but not CXCL12 concentrations in LNB CSF were associated with a high MI. Finally, the neutralization experiments underlined the key role of CXCL13, since preincubation with anti-CXCL13 antibodies significantly reduced the migration activity of B cells. Therefore CXCL13 fulfils all criteria as stated above.
CCL19 instead, which signals via the CCR7 receptor, appears to play a functional role for B-cells only in single cases (as seen in patient Nr. 3, Fig.
4). However, this chemokine could be important for the immigration of T-lymphocytes instead, as suggested in the literature. Activated T-lymphocytes strongly upregulate CCR7 and efficiently migrate to CCL19 [
28], and T-lymphocytes found in the CSF of patients with (not further specified) inflammatory CNS-diseases are CCR7 positive [
27]. As CCL19 is much stronger chemotactic for T than for B cells, its role in LNB could be the attraction of T cells, but this would have to be determined in further studies.
For the association analysis as shown in Figure
3, the cut-off for an increased migration index was set at 80, as the CSF of all NIND-patients had a migration below this value. As seen in Figure
3A and
3B and also in Figure
1A, there are four CSF samples from LNB patient with a migration index below 80. The one of them with the highest MI also has a high CSR for CXCL13 and resembles in its basic CSF findings the other LNB patients (cell count 368/μl, albumin CSF/serum quotient 19.1). The three others, however, appear to be a distinct group. They are combining a low CSR for both CXCL13 (7-18.5) and CCL19 (0.7-1.7) with a low migration index. Compared to the remaining LNB patients, they also have a lower cell count (19 ± 10 vs. 146 ± 152 cells/μl, p < 0.01), albumin CSF/serum quotient (7.2 ± 5.2 vs. 19.6 ± 13, p = 0.02), and in none of them, plasma cells are found in the CSF (in contrast to a mean of 5% in the other LNB patients, p < 0.01.). Although these data should be interpreted with caution, given the small sample size of this subgroup, the observation that in all three patients low-grade inflammatory CSF changes were paralleled by very low CSRs for CXCL13 and CCL19 (as compared to other LNB patients) further underlines the functional role of both chemokines. It also has to be mentioned that the less pronounced CSF alterations in this LNB subgroup are not reflected by less (or different) clinical symptoms or by clear differences in age or gender. However, we are not aware of the duration between the onset of symptoms and lumbar puncture. Both the quantity and quality of CSF abnormalities (e.g. CSF pleocytosis and CSF chemokine levels) may change over time after infection, which may be an interesting topic to be addressed in future studies.
As the inhibition of CXCL13, CCL19 and CXCL12 does not completely abrogate the chemotactic activity of LNB CSF samples, additional factors have to be involved. In a study by Dubois et al., the supernatant of cultured dendritic cells was found to be chemotactic for B cells [
29]. As the supernatant was insensitive to PTx treatment, the authors stated that the relevant factors have to be different from classical chemoattractants. In our study, however, the chemotactic activity of both, rhCXCL13 and LNB CSF was reduced by around 80% by interfering with the PTx pathway, and also heat treatment. In addition, the chemotactic activity of LNB CSF samples and rhCXCL13 was completely blocked by eliminating all substances larger than 3 kDa. Chemokines weigh between 6 and 14 kDa [
30], signal predominantly via the PTx-sensitive pathway [
31], and are as proteins heat sensitive. This indicates that, besides CXCL13, another "similar" chemokine could be responsible for the remaining chemotactic activity after blockade of CXCL13. Another possible candidate could be the complement, as the size of complement factors ranges from 24 to 410 kDa, they are heat sensitive and - as shown for C5a - signal via PTx-sensitive G proteins [
32]. In addition, it has been demonstrated, that
B.b. can induce the production of complement in vitro and in vivo [
33,
34] and that B cells respond chemotactic to C5a [
35]. However, there are even more substances to be considered: components of the Borrelia themselves could also contribute to the chemotaxis. In a study by Benach et al., large proteins of the borrelial membrane like the Outer surface protein A or flagellin had a chemotactic effect on leucocytes [
36]. Taken together, further studies are warranted to identify all factors involved in B cell attraction.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
TAR and UK have conceived of the study, TAR has established the experimental setting and written the manuscript, AP has carried out most of the experiments, MA has done the FACS analysis, MW and SK have collected the CSF and serum samples, CS and MK have participated in the design of the study and the statistical analysis and helped drafting the manuscript. HWP and UK have participated in the design and coordination of the study and helped to draft the manuscript. All authors read and approved the final manuscript.