Brain, Vol. 124, No. 3, 480-492,
March 2001
© 2001 Oxford University Press
Two subsets of dendritic cells are present in human cerebrospinal fluid
1 Divisions of Neurology, 2 Infectious Diseases and 3 Ophthalmology, Karolinska Institutet, Huddinge University Hospital, Stockholm, Sweden
Correspondence to:
Dr Mikhail Pashenkov, Division of Neurology, Huddinge University Hospital, R54, SE-14186 Huddinge, Sweden E-mail: Mikhail.Pashenkov{at}neurotec.ki.se
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and methodsResultsDiscussionReferences |
|---|
Little is known about the presence of dendritic cells in the human CNS. To investigate the occurrence of dendritic cells in the CSF, paired blood/CSF samples from patients with multiple sclerosis, acute optic neuritis, Lyme neuroborreliosis, other inflammatory neurological diseases and non-inflammatory neurological diseases were examined using flow cytometry. Almost all CSF samples contained myeloid (lin–CD11c+HLA-DR++CD123dim) and plasmacytoid (lin–CD11c–HLA-DR+CD123high) dendritic cells. In non-inflammatory neurological diseases, dendritic cells of either subset only constituted up to 1% of CSF mononuclear cells. Myeloid CSF dendritic cells were elevated in optic neuritis, neuroborreliosis and other inflammatory neurological disorders, while plasmacytoid dendritic cells were elevated in all neuroinflammatory conditions studied, with especially high numbers in neuroborreliosis. Numbers of CSF dendritic cells correlated with the common parameters of CNS inflammation. The myeloid dendritic cells in CSF expressed higher levels of HLA-DR, CD86, CD80 and CD40 than those in blood, whereas expression of these molecules by plasmacytoid dendritic cells was equal in blood and CSF. Both CSF and blood dendritic cells expressed the chemokine receptor CCR5. This is the first demonstration that dendritic cells are present in human CSF and that plasmacytoid dendritic cells are present in a non-lymphoid compartment. Myeloid and plasmacytoid dendritic cells in CSF may contribute to orchestration of the local immune responses.
dendritic cells; cerebrospinal fluid; Lyme disease; multiple sclerosis; optic neuritis
APC = antigen-presenting cell; FACS = fluorescence activated cell sorting; FITC = fluorescein isothiocyanate; IFN = interferon; mAb = monoclonal antibodies; MFI = mean fluorescence intensity; PE = phycoerythrin; PerCP = periclinin chlorophyll protein; Th1(2) = T-helper 1(2)
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Dendritic cells are potent antigen presenting cells able to activate naïve antigen-specific T cells, and are assumed to consist of several pools of cells (Steinman et al., 1997
; Banchereau and Steinman, 1998). The `sentinel' pool of dendritic cells is represented by CD1a+ Langerhans cells of epidermis and mucosal surfaces. The so-called myeloid dendritic cells (CD11c+) are found in many tissues, at sites of antigen delivery in secondary lymphoid organs, and in blood (Kohrgruber et al., 1999; Robinson et al., 1999). The so-called plasmacytoid dendritic cells, which are CD11c–, have been identified in blood and in T-cell-rich areas of secondary lymphoid organs (Grouard et al., 1997; Olweus et al., 1997; Kohrgruber et al., 1999; Robinson et al., 1999). A brief comparison of myeloid and plasmacytoid dendritic cells is presented in Table 1.
|
Different dendritic cell subsets may play different roles in an immune response. Myeloid dendritic cells take up locally delivered antigens, then mature and migrate into secondary lymphoid organs, in order to present the antigens to T cells. In vitro, monocyte-derived myeloid dendritic cells have been shown to induce T-helper 1 (Th1) or Th2 responses, depending on the maturation conditions (Rissoan et al., 1999
; Vieira et al., 2000). The main function of plasmacytoid dendritic cells appears to be production of large amounts of type I interferons (IFN) in response to infections with enveloped viruses or bacteria (Feldman et al., 1994; Svensson et al., 1996; Cella et al., 1999; Siegal et al., 1999). Upon in vitro culture with IL-3 (interleukin-3), these cells mature into potent antigen-presenting cells (APC) and induce Th2 responses (Grouard et al., 1997; Rissoan et al., 1999). The ability of plasmacytoid dendritic cells to produce type I IFNs makes them especially important with regard to multiple sclerosis, as IFN-ß is one of two drugs proven to be efficient in combating this disease.
The CNS is considered an immunoprivileged site. One of the mechanisms of such immune privilege might be the lack of potent APC, including dendritic cells, within the CNS. Recent findings have challenged this view. OX-62+MHC (MHC = major histocompatibility complex) class II+ dendritic cells were found in the meninges and choroid plexus of normal rats (McMenamin et al., 1999). Massive infiltration of dendritic cells into the brain is observed in DTH (delayed type hypersensitivity) reactions against Mycobacterium tuberculosis sequestered within rat brain tissue (Matyszak and Perry, 1996). Dendritic-like (dendriform) cells emerging on top of murine astroglial monolayers cultured with the growth factor GM-CSF (granulocyte–macrophage colony-stimulating factor) have been described. These cells are CD11c+ and act as potent APC (Fischer and Bielinsky, 1999). Human microglia express CD4, CD11c and HLA-DR (HLA = human leucocyte antigen) and can serve as APC, thus resembling myeloid dendritic cells (Ulvestad et al., 1994). Thus, there is evidence that dendritic cells may develop from the cells resident in the CNS and/or migrate to the CNS from the periphery.
To our knowledge, data on the presence of `classical' dendritic cells in human CNS are lacking. However, human blood contains two well-characterized and easily identifiable subsets of dendritic cells, namely lin–CD11c+CD123dim myeloid dendritic cells and lin–CD11c–CD123high plasmacytoid dendritic cells (O'Doherty et al., 1994; Thomas and Lipsky, 1994; Olweus et al., 1997; Kohrgruber et al., 1999; Robinson et al., 1999). We have reported recently preliminary data which shows that plasmacytoid dendritic cells are present in the CSF and expanded during CNS infections (Link et al., 1999). Here, we report on the presence of dendritic cells in the CSF of patients with multiple sclerosis, optic neuritis, Lyme neuroborreliosis, other inflammatory neurological diseases and non-inflammatory neurological diseases, and compare the phenotype of blood and CSF dendritic cells.
|
Patients and methods |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Patients
Paired blood (20 ml) and CSF (10–20 ml) samples were collected from 22 patients (14 females, eight males) with clinically definite multiple sclerosis (Poser et al., 1983
), eight patients (six females, two males) with optic neuritis, four patients (three females, one male) with Lyme neuroborreliosis, six patients (three females, three males) with other inflammatory neurological diseases and 13 patients (six females, seven males) with non-inflammatory neurological diseases. All these patients were undergoing a standard diagnostic procedure for their neurological disease. In all patients, total CSF cell counts, CSF : plasma albumin ratio and IgG index were determined. All patients included in the study and mentioned above had to have >10 000 cells in their CSF samples, due to the technical limitations of flow cytometry. The albumin ratio was calculated as albuminCSF (mg/l) : albuminplasma (g/l) and used as a measure of blood–brain barrier permeability (Link and Tibbling, 1977). The IgG index was calculated as
 |
).
The median age of patients with multiple sclerosis was 39 years (range 24–67 years), the median duration of multiple sclerosis was 8 years (range 0.5–41 years), and the median EDSS (Expanded Disability Status Score) (Kurtzke, 1983) was 2.5 (range 0–6). All patients with multiple sclerosis had oligoclonal IgG bands in CSF determined by agarose isoelectric focusing (Kostulas and Link, 1982) and more than four multiple sclerosis-like lesions on MRI of the brain. None of the multiple sclerosis patients had ever received any immunomodulatory therapy, including steroids or IFN-ß. CSF cell counts ranged between 1 and 27 mononuclear cells per microlitre, and nine patients had a mononuclear pleocytosis (more than 5 mononuclear cells per microlitre).
Eight patients had acute monosymptomatic optic neuritis. Their median age was 35 years (range 24–46 years). All patients had oligoclonal IgG bands in the CSF, which were absent in corresponding sera. Cell counts in CSF ranged from 2 to 25 mononuclear cells per microlitre; six patients had mononuclear pleocytosis.
Four patients had acute meningoencephalitis due to Borrelia burgdorferi infection. The median age of these patients was 59 years (range 21–64 years), CSF cell counts varied between 24 and 930 mononuclear cells per microlitre. All patients had oligoclonal bands in CSF, which were absent in serum.
Among six patients with other inflammatory neurological diseases, four had aseptic meningitis or encephalitis, one had chronic demyelinating polyneuropathy and one had Behciet syndrome with CNS involvement, which was manifest by focal neurological deficit and mononuclear CSF pleocytosis. The median age of the other inflammatory neurological diseases patients was 44 years (range 27–67 years), CSF cell counts varied between 2 and 288 mononuclear cells per microlitre, and five patients had mononuclear pleocytosis. One patient with other inflammatory neurological diseases (aseptic meningoencephalitis) had oligoclonal bands in CSF, which were absent in serum.
The group of non-inflammatory neurological diseases consisted of 13 patients with no clinical evidence of any inflammatory process within the CNS. Five of these patients had ischaemic stroke, two had tension headache, and one patient each had transient ischaemic attack, vertigo, depression, isolated cerebellopathy, pseudotumour cerebri and Alzheimer's disease. CSF cell counts varied between 1 and 5 mononuclear cells per microlitre. The median age of this group was 56 years (range 21–81 years). Age differences between the groups were not significant.
Preparation of mononuclear cells from blood and CSF samples
Mononuclear cells were isolated from heparinized blood samples by centrifugation over Lymphoprep (Nycomed, Oslo, Norway) density gradient (
= 1.078), washed three times with Dulbecco's modified Eagle medium (Life Technology, Paisley, UK), counted and resuspended in PBS (phosphate-buffered saline) supplemented with 1% BSA (bovine serum albumin) for subsequent immunostaining.
Ten to twenty millilitres of CSF was obtained from each patient. Within 20–30 min of the lumbar puncture, CSF was subjected to cell counting. CSF was centrifuged at 100 g, supernatants were discarded and cells gently resuspended in PBS containing 1% BSA.
Antibodies and cytokines
Monoclonal antibodies (mAbs) used in the study are listed in Table 2. Fluorescein isothiocyanate (FITC)-labelled and phycoerythrin (PE)-labelled irrelevant IgG1 were from Dakopatts (Copenhagen, Denmark). Unlabelled IgG1 was from Becton Dickinson (Mountain View, Calif., USA). Secondary PE-labelled goat anti-mouse Ab was from Serotec (Oxford, UK), biotinylated and unlabelled goat anti-mouse Abs were from Dakopatts, and avidin–biotin–peroxidase complex was from Vector Laboratories (Burlingame, Calif., USA). Purified human IFN-
(Interferon Alfanative) was from BioNative AB (Umeå, Sweden).
|
Identification, enumeration and immunophenotyping of dendritic cell subsets in blood and CSF by flow cytometry
Flow cytometric identification of dendritic cells in blood and CSF was based on their lack of markers for T cells, B cells, natural killer cells or monocytes, positivity for HLA-DR and expression of CD123 (for plasmacytoid dendritic cells) and CD11c (for myeloid dendritic cells) (Olweus et al., 1997
; Kohrgruber et al., 1999). Blood mononuclear cells (2 x 105) or 10–100 x 103 CSF mononuclear cells were resuspended in 100 µl of 1% BSA/PBS and stained with a cocktail of FITC-labelled anti-CD3, CD14, CD16, CD19 and CD56 (lineage cocktail), PE-labelled anti-CD123 and periclinin chlorophyll protein (PerCP)-labelled anti-HLA-DR (20 min, 4°C). Cells were washed with 1% BSA/PBS, resuspended in PBS and analysed by a three-colour FACScan flow cytometer and CellQuest software (both from Becton Dickinson). Mononuclear cells were distinguished from debris by light scattering, and 50 000 events (at least 5000 in case of CSF) within the mononuclear cell gate were acquired. Cocktaildim/– cells were gated (Fig. 1), and plasmacytoid dendritic cells were identified as cocktail–HLA-DR+CD123high cells, while myeloid dendritic cells were cocktaildimHLA-DR++CD123dim (Fig. 1). The dim staining of myeloid dendritic cells with the cocktail of lineage-specific mAbs was apparently due to dim expression of CD14 by this cell subset (Thomas and Lipsky, 1994). As an alternative, anti-CD123 mAb was replaced with PE-labelled anti-CD11c. Myeloid dendritic cells were then cocktaildimHLA-DR++CD11c+, and plasmacytoid dendritic cells were cocktail–HLA-DR+CD11c–. Both approaches gave similar percentages of dendritic cells. Control samples were stained with irrelevant PE-IgG1.
|
Since CSF samples from patients with non-inflammatory neurological diseases contained few cells, it was not always possible in these patients to perform all three stainings (anti-CD123, anti-CD11c and isotype control) on one sample. The following `priority list' for the staining of CSF was then applied: (i) cocktail/HLA-DR/CD123; (ii) cocktail/HLA-DR/CD11c; and (iii) cocktail/HLA-DR/isotype control. Corresponding blood samples were always stained with cocktail/HLA-DR/isotype control, cocktail/HLA-DR/CD123 and cocktail/HLA-DR/CD11c. Blood and CSF samples stained with isotype-matched control IgG never displayed a signal above autofluorescence.
Absolute numbers of myeloid and plasmacytoid dendritic cells in CSF (per millilitre) were calculated by multiplying their percentages among all CSF mononuclear cells (determined by flow cytometry) by the total cell count.
To detect expression of co-stimulatory molecules on dendritic cells, blood or CSF mononuclear cells were stained with the lineage cocktail (FITC-labelled), PerCP-labelled anti-HLA-DR and PE-labelled mAb against a particular costimulatory molecule (CD86, CD80 or CD40). Control samples were stained with irrelevant isotype-matched PE-IgG. To examine expression of costimulatory molecules differentially on plasmacytoid and myeloid dendritic cells by a three-colour flow cytometer, the following approach was used. Since CD123high plasmacytoid dendritic cells were cocktail–HLA-DR+ and CD123dim myeloid dendritic cells were cocktaildimHLA-DR++, it was possible to differentiate between myeloid and plasmacytoid dendritic cells by using only two channels (FITC and PerCP) of the FACScan flow cytometer, while the third one (PE) was used for detection of costimulatory molecules (see detailed description in the legend to Fig. 1
).
To detect expression of chemokine receptors on dendritic cells, cells were stained with unlabelled mAbs against chemokine receptors, secondary PE-labelled goat anti-mouse Ab, and, finally, with FITC-lineage cocktail and PerCP-labelled anti-HLA-DR. The same approach as above was used to differentially estimate expression of chemokine receptors on plasmacytoid and myeloid dendritic cells.
Expression of HLA-DR, costimulatory molecules and chemokine receptors by plasmacytoid or myeloid dendritic cells was presented as mean fluorescence intensity (MFI) of the PerCP-signal (for HLA-DR) or PE-signal (for all other molecules) displayed by entire subsets of myeloid or plasmacytoid dendritic cells.
FACS
To assess morphology of plasmacytoid CSF dendritic cells, they were isolated from one CSF sample by means of FACS. For this purpose, CSF mononuclear cells were stained with FITC-lineage cocktail and PerCP-labelled anti-HLA-DR, and cocktail–HLA-DR+ cells (plasmacytoid dendritic cells) were sorted by a FACSVantage cell sorter (Becton Dickinson).
Immunodetection of IFN-
Intracellular IFN- in blood or CSF mononuclear cells was detected by flow cytometry. We did not use Golgi-blocking agents in these experiments, since (i) we wanted to study immediate ex vivo expression of IFN- and (ii) previous work (Greenway et al., 1995; Svensson et al., 1996; Siegal et al., 1999; Cella et al., 2000) showed that intracellular staining for IFN- does not require blocking of the Golgi complex. Blood or CSF mononuclear cells were first stained with FITC-lineage cocktail and PerCP-labelled anti-HLA-DR, and antigenic sites of these antibodies were blocked by excess of unlabelled goat anti-mouse Ab. Cells were then fixed with 4% ice-cold paraformaldehyde, permeabilized with 0.1% saponin and stained with anti-IFN- mAb (clone MMHA-2, 5 µg/ml) and secondary PE-labelled goat anti-mouse Ab (1 : 30). Dendritic cells were gated as described above. Control samples were stained with either irrelevant IgG1 or anti-IFN- mAb pre-absorbed overnight with purified human IFN-, which totally abolished the specific signal. The IFN- staining usually represented a curved shift, rather than positivity of a proportion of the cells. Therefore, the expression of IFN- by plasmacytoid dendritic cells, myeloid dendritic cells, and non-dendritic cells in blood and CSF was judged by MFI displayed by all cells of the respective subset. To correct for non-specific staining, these MFI values were subtracted by the background MFI for respective cell subsets, determined by staining with isotype control mAb.
To confirm the flow cytometry data, IFN- was also detected immunocytochemically in sorted plasmacytoid CSF dendritic cells and non-dendritic cells from one patient with neuroborreliosis. A paraformaldehyde–saponin–peroxidase procedure described elsewhere was used (Litton et al., 1997). Slides were examined by light microscopy at x400. The same controls as for flow cytometry were used (see above).
Statistics
Paired comparisons (blood versus CSF) were done by Wilcoxon signed rank test. Three or more groups were compared by non-parametric Kruskal–Wallis test: if the P value in this test was <0.05, differences between pairs of groups were tested further by non-parametric Mann–Whitney U-test. Correlations were tested by non-parametric Spearman rank test. Throughout the text, data are expressed as median (min–max).
|
Results |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Detection and immunophenotyping of dendritic cells in CSF and blood
When CSF mononuclear cells were double-stained with FITC-lineage cocktail and PerCP-labelled anti-HLA-DR, a cocktaildim/–HLA-DR+ cell population was found in the majority of CSF samples (Fig. 1
). This population expressed CD4 and was usually minor, but was expanded in some cases up to 17% of CSF mononuclear cells. Triple-staining with the addition of anti-CD123 (IL-3R) or anti-CD11c revealed that the cocktail–HLA-DR+CD4+ cells represented two subpopulations, namely CD123highCD11c–HLA-DR+ and CD123dimCD11c+HLA-DR++. Among them, the latter but not the former co-expressed CD1a. Both subpopulations expressed CD86 and CD40, with the higher levels those of CD123dimCD11c+HLA-DR++ cells. The CD123dim subpopulation expressed low levels of CD80. Both subpopulations were negative for CD25 (IL-2R) (Fig. 1). The cocktaildimHLA-DR++CD123dim but not cocktail–HLA-DR+CD123high subpopulation expressed CD45RO, and both subpopulations were negative for CD83 (data not shown).
Cells with these surface phenotypes are well known in blood as myeloid dendritic cells [cocktaildim(CD14dim)CD123dimCD11c+HLA-DR++] and plasmacytoid dendritic cells [cocktail–CD123dimCD11c+HLA-DR+] (O'Doherty et al., 1994; Thomas and Lipsky, 1994; Olweus et al., 1997; Kohrgruber et al., 1999; Robinson et al., 1999). Both dendritic cell subpopulations were readily detectable in all blood mononuclear cell samples from our patients. We thus came to the conclusion that CSF contains myeloid and plasmacytoid dendritic cells.
Morphology and surface phenotype of sorted plasmacytoid CSF dendritic cells
The numbers of cells present in the great majority of the CSF samples are not enough to isolate dendritic cells by means of fluorescence- or magnetic-activated cell sorting. We were able to sort plasmacytoid dendritic cells from one CSF sample obtained from a patient with acute Lyme neuroborreliosis. The initial number of CSF mononuclear cells available was 1.5 x 106, of which 14.5% expressed a surface phenotype of plasmacytoid dendritic cells (Fig. 2). The yield of these cells after sorting was 1.5 x 105, with a purity of 95%. When stained by routine Giemsa staining, they displayed a typical plasmacytoid morphology (Fig. 3). Additional phenotyping showed that these cells were CD1a–CD83–, but expressed CD86 and high level of CD54 (intracellular adhesion molecule, ICAM-1) (Fig. 2). CD11c was detected on 5% of CD123high plasmacytoid dendritic cells, CD80 was not detectable. Additionally, plasmacytoid dendritic cells expressed the chemokine receptor CCR5. Thus, morphology and surface marker profile of these cells corresponded exactly to that of immature plasmacytoid blood dendritic cells (Olweus et al., 1997; Cella et al., 1999; Kohrgruber et al., 1999). These cells were poor stimulators of an allogenic mixed leucocyte reaction (not shown), which is in agreement with previous findings (O'Doherty et al., 1994; Kohrgruber et al., 1999).
|
|
Due to their low numbers, we have so far been unable to sort CD123dimCD11c+ myeloid dendritic cells from the CSF.
Enumeration of dendritic cells in blood and CSF
Percentages of dendritic cells in blood are similar in all patient groups
When analysed in all patients, the proportions of myeloid and plasmacytoid dendritic cells among blood mononuclear cells were 0.3% (0.1–0.8%) and 0.2% (0–0.9%), respectively. There were no differences between patient groups (data not shown).
Low numbers of dendritic cells in CSF from patients with non-inflammatory neurological diseases
CSF samples from 13 patients with non-inflammatory neurological diseases, without CSF pleocytosis, were considered as reference. In these samples, myeloid dendritic cells constituted a median 0.4% (range 0–1.2%) and plasmacytoid dendritic cells 0.3% (0–1.1%) of CSF mononuclear cells. These values were not different from those in blood: 0.3% (0.1–0.5%) for myeloid dendritic cells and 0.2% (0.1–0.9%) for plasmacytoid dendritic cells (Fig. 4). Absolute numbers of myeloid and plasmacytoid dendritic cells in patients with non-inflammatory neurological diseases were, respectively, 9 (0–60) and 9 (0–40) per millilitre of CSF (Table 3).
|
|
Increased numbers of CSF dendritic cells in early multiple sclerosis
Percentages of plasmacytoid dendritic cells in CSF of multiple sclerosis patients were higher than in corresponding blood samples [0.6% (0–1.9%) versus 0.2% (0–0.4%), P = 0.001]. Percentages of myeloid dendritic cells in CSF were not different from those in blood [0.3% (0–3.1%) versus 0.3% (0.2–0.6%)], although some multiple sclerosis patients had rather high frequencies of myeloid dendritic cells in CSF, up to 3.1% of CSF mononuclear cells (Fig. 4
).
Absolute numbers of plasmacytoid CSF dendritic cells were higher in multiple sclerosis than in non-inflammatory neurological diseases (P < 0.01), while numbers of myeloid CSF dendritic cells were similar in the two groups (Table 3). Numbers of plasmacytoid and myeloid CSF dendritic cells correlated negatively with the duration of multiple sclerosis (r = –0.47, P = 0.03 and r = –0.41, P = 0.05). Thus, the shorter the duration of multiple sclerosis, the higher the absolute numbers of dendritic cells in CSF. There was no correlation between numbers of CSF dendritic cells and multiple sclerosis patients' disability scores (EDSS).
Three and the two consecutive CSF samples were obtained from two patients with short duration of multiple sclerosis, at 3-month intervals. CSF dendritic cells in these two patients remained relatively stably elevated. Numbers of myeloid dendritic cells were 400, 280 and 210 per millilitre CSF in one patient, and 100 and 70 per millilitre in the other, while numbers of plasmacytoid dendritic cells were 190, 120 and 200, and 120 and 50 per millilitre CSF, respectively.
All patients with optic neuritis included in this study had oligoclonal bands in their CSF. Seventy-five per cent of such patients develop clinically definite multiple sclerosis within 5 years (Söderström et al., 1998) and thus can be considered as having early multiple sclerosis. In a manner similar to patients with multiple sclerosis, patients with optic neuritis had elevated percentages of plasmacytoid dendritic cells in CSF compared with blood [0.6% (0.3–1.1%) versus 0.2% (0–0.5%), P < 0.01] (Fig. 4). Percentages of myeloid dendritic cells in these patients were also higher in CSF than in blood [0.8% (0.4–2.2%) versus 0.2% (0.1–0.5%), P < 0.01]. Furthermore, absolute numbers of myeloid and plasmacytoid CSF dendritic cells in optic neuritis were higher than in non-inflammatory neurological diseases and in multiple sclerosis (Table 3). These findings in optic neuritis are consistent with the above observation of higher numbers of CSF dendritic cells in patients with a shorter duration of multiple sclerosis.
Accumulation of plasmacytoid dendritic cells in CSF in neuroborreliosis
Percentages of plasmacytoid dendritic cells in CSF from the four patients with neuroborreliosis were 4.4% (2.1–14.5%), i.e. much higher than in corresponding blood samples [0.3% (0.2–0.4%)] and in CSF from patients with non-inflammatory neurological diseases, multiple sclerosis and optic neuritis (P < 0.001 for all comparisons) (Fig. 4). Absolute numbers of plasmacytoid CSF dendritic cells in patients with neuroborreliosis were 17 850 (500–133 400) per millilitre CSF, higher than in all other patient groups (P < 0.001 for all comparisons) (Table 3). From one of the patients, three consecutive samples were obtained, one on admission to hospital and the rest at 1-week intervals. The percentages of plasmacytoid dendritic cells in these samples were 14.5, 17.6 and 7.1%, with corresponding absolute numbers being 133 400, 75 700 and 34 100 per millilitre of CSF.
Percentages of myeloid CSF dendritic cells in patients with neuroborreliosis were the same as in the other groups and did not differ from the figures in corresponding blood samples. However, absolute numbers of these cells in CSF, due to high pleocytosis, were still higher than in non-inflammatory neurological diseases or multiple sclerosis (Table 3).
In patients with other inflammatory neurological diseases (aseptic meningitis and/or encephalitis, neuro-Behciet syndrome, chronic demyelinating polyneuropathy), percentages of plasmacytoid dendritic cells in CSF were also elevated compared with blood [0.8% (0.2–3.6%) versus 0.1% (0.1–0.3%), P < 0.05], but to a lesser extent than in neuroborreliosis. Particularly, increased percentages and absolute numbers of plasmacytoid CSF dendritic cells were observed in a patient with neuro-Behciet syndrome (3.6% of CSF mononuclear cells, 10 200 cells per millilitre). Absolute numbers of plasmacytoid and myeloid CSF dendritic cells were higher in this group than in non-inflammatory neurological diseases (P < 0.01 and <0.05, respectively) (Table 3).
Correlations between numbers of CSF dendritic cells and common markers of CNS inflammation
Correlations between the numbers of CSF dendritic cells and common indicators of CSF inflammation, such as total CSF cell counts and CSF IgG index, are presented in Table 4. Total CSF cell counts reflect inflammatory infiltration of the CNS/CSF compartment, and CSF IgG index intrathecal IgG synthesis. As can be seen from Table 4, absolute numbers of myeloid and plasmacytoid CSF dendritic cells (analysed in all patients) correlate significantly with total CSF cell counts (r = 0.67 and r = 0.75, respectively) and CSF IgG indices (r = 0.33 and r = 0.54, respectively). Additionally, percentages of plasmacytoid CSF dendritic cells also correlated with total CSF cell counts and IgG index; in other words, the higher the CSF pleocytosis or IgG index, the more frequent are plasmacytoid dendritic cells among CSF mononuclear cells.
|
Neither percentages nor absolute numbers of plasmacytoid and myeloid CSF dendritic cells correlated with CSF : plasma albumin ratio, the parameter that reflects the permeability of the blood–brain barrier (data not shown).
Expression of HLA-DR and costimulatory molecules by myeloid and plasmacytoid dendritic cells in CSF and blood
Expression of HLA-DR on myeloid and plasmacytoid dendritic cells was studied in all CSF and blood samples, and expression of CD86, CD80 and CD40 was studied in 10 patients (seven with multiple sclerosis and three with neuroborreliosis). Myeloid blood dendritic cells expressed higher levels of HLA-DR and CD86 than plasmacytoid blood dendritic cells, while expression of CD80 and CD40 was the same in both dendritic cell subsets in blood. Expression of HLA-DR, CD86, CD80 and CD40 on myeloid dendritic cells was significantly higher in CSF than in corresponding blood samples, indicating that myeloid dendritic cells in CSF are more mature than in blood (Fig. 5). In contrast, expression of these four molecules on plasmacytoid dendritic cells was equal in CSF and in blood. Thus, plasmacytoid CSF dendritic cells retained the same maturity as in blood. There were no differences in expression of HLA-DR, CD86, CD80 and CD40 by blood or CSF dendritic cells between patients with multiple sclerosis and neuroborreliosis.
|
Expression of chemokine receptors by dendritic cells in blood and CSF
Since an accumulation of dendritic cells in CSF from patients with neuroinflammation was observed, we assessed expression of CCR5, a receptor for RANTES (regulated on activation, normal T-cell expressed and secreted), MIP-1
(macrophage inflammatory protein-1) and MIP-1ß, by CSF and blood dendritic cells, as these chemokines could play a role in the attraction of dendritic cells into CSF. Plasmacytoid and myeloid dendritic cells in both CSF and blood did express CCR5, and myeloid dendritic cells in CSF displayed even higher expression of CCR5 than in blood (P < 0.05) (Fig. 5).
Expression of IFN- by blood and CSF dendritic cells
Using flow cytometry, we examined directly ex vivo blood and CSF dendritic cells from six multiple sclerosis patients for the presence of intracellular IFN-, assuming that plasmacytoid CSF dendritic cells might have been triggered in vivo to produce IFN-. Both dendritic and non-dendritic cells in blood and CSF contained detectable intracellular IFN- (Table 5). This is consistent with a previous report that mononuclear cells constitutively express certain isoforms of IFN- intracellularly (Greenway et al., 1995). The specificity of IFN- staining was confirmed by overnight pre-absorption of anti-IFN- mAb with purified human IFN-, which resulted in complete abrogation of the fluorescent signal. In all samples, MFI of IFN- was higher in CSF mononuclear cells than in blood mononuclear cells, but did not differ between between plasmacytoid dendritic cells, myeloid dendritic cells and non-dendritic cells in the same compartment (Table 5). A similar tendency was observed in a patient with Lyme neuroborreliosis (data not shown). Accordingly, immunocytochemistry of plasmacytoid dendritic and non-dendritic cells sorted from the CSF of another patient with neuroborreliosis showed that cells brightly stained for IFN- were present in both cell subpopulations (Fig. 6). The presence of intracellular IFN- does not mean that it is secreted and plays a biological role. However, given the low total numbers of CSF dendritic cells, we have so far been unable to examine secretion of IFN- by sorted plasmacytoid or myeloid CSF dendritic cells.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
The CNS is considered an immunoprivileged site, i.e. immune reactions within the CNS are strictly controlled and foreign antigens sequestered within the CNS do not normally induce immune responses. However, the CNS is still affected by inflammatory diseases, such as infections, as well as multiple sclerosis, where the neuroinflammation may have an autoimmune origin (Brosnan and Raine, 1996
). Dendritic cells are crucial in inducing and regulating immune responses against pathogens and autoantigens (Steinman et al., 1997; Banchereau and Steinman, 1998; Dittel et al., 1999). Evidence is now accumulating that dendritic cells are present in the CNS and may participate in immune reactions in this compartment (Ulvestad et al., 1994; Matyszak and Perry, 1996; Fischer and Bielinsky, 1999). The present study aimed to investigate whether dendritic cells are present in human CSF, which is routinely examined in clinical conditions for evidence of CNS inflammation. A methodological problem always exists when dealing with a rare cell population, such as dendritic cells; i.e. how precise is enumeration of these rare cells. We tried to overcome this dilemma by (i) obtaining large CSF samples (10–20 ml or more), which represent a substantial part of patient's total CSF; (ii) acquiring at least 5000 events per test in case of CSF and 50 000 events in case of blood when running flow cytometry; and (iii) making several stainings of one CSF sample with the same antibody combinations, which showed that percentages of cocktaildim/–HLA-DR+ CSF mononuclear cells (total CSF dendritic cells) did not change significantly from test to test.
The main findings in this study are: (i) both myeloid (CD11c+CD123dim) and plasmacytoid (CD11c–CD123high) dendritic cells are present in human CSF; (ii) only very few dendritic cells are present in CSF from patients with non-inflammatory neurological diseases who lack pleocytosis; (iii) absolute numbers of myeloid and plasmacytoid CSF dendritic cells are elevated in different neuroinflammatory conditions, such as multiple sclerosis, optic neuritis, neuroborreliosis or aseptic meningoencephalitis; (iv) in acute Lyme neuroborreliosis, CSF is strongly enriched on plasmacytoid dendritic cells; and (v) levels of myeloid and plasmacytoid CSF dendritic cells correlate with total CSF cell counts and CSF IgG index, but not with CSF : plasma albumin ratio. These findings demonstrate that the accumulation of dendritic cells of either subset in the CSF is directly related to inflammation in the CNS. Furthermore, in acute neuroinflammation, like in the case of neuroborreliosis, initially elevated numbers of CSF dendritic cells can drop dramatically over 2 weeks, while in multiple sclerosis, a chronic disease, numbers of CSF dendritic cells can remain elevated over several months.
An interesting finding is elevation of myeloid and plasmacytoid CSF dendritic cells in our patients with shorter duration of multiple sclerosis and with multiple sclerosis-type monosymptomatic optic neuritis. Previous data have shown that ~75% of optic neuritis patients with oligoclonal bands in their CSF develop multiple sclerosis (Söderström et al., 1998), so they can be viewed as having early multiple sclerosis. Our data thus suggest a role for dendritic cells in initiation of the inflammatory process in multiple sclerosis.
Myeloid blood dendritic cells are potent APC (O'Doherty et al., 1994; Thomas and Lipsky, 1994; Kohrgruber et al., 1999) and can activate naïve T cells at dendritic : T-cell ratios of <1 : 50–1 : 100. Myeloid CSF dendritic cells may play a similar role by presenting antigens delivered within the CNS and inducing antigen-specific T cell responses, even though their percentages among CSF mononuclear cells do not usually exceed 1–2%. Myeloid CSF dendritic cells express higher levels of HLA-DR, CD40, CD80 and CD86 than myeloid dendritic cells in blood or plasmacytoid dendritic cells in blood and CSF. This suggests a higher antigen-presenting and T cell-stimulatory capacity of myeloid CSF dendritic cells. Indeed, we found a correlation between percentages of myeloid dendritic cells in CSF and the ability of CSF leucocytes to stimulate allogeneic T cells; this correlation was absent in the case of blood (data not shown). However, it remains to be established whether antigen presentation by CSF dendritic cells does indeed occur in situ or in secondary lymphoid organs, where dendritic cells migrate from the CNS.
Plasmacytoid blood dendritic cells are considered immature, or even named dendritic cell precursors (Rissoan et al., 1999). They are poor APC when freshly isolated from peripheral blood (Kohrgruber et al., 1999), but produce large amounts of type I IFNs in response to infection with bacteria and enveloped viruses, and thus have been implicated in antiviral and antibacterial defence (Feldman et al., 1994; Svensson et al., 1996; Cella et al., 1999; Siegal et al., 1999). Viruses or the cytokine IL-3 cause maturation of plasmacytoid dendritic cells, increasing expression of HLA-DR and co-stimulatory molecules, but diminishing the ability to produce type I IFNs (Grouard et al., 1997; Siegal et al., 1999). In our experiments, plasmacytoid dendritic cells in CSF expressed the same levels of HLA-DR, CD86, CD80 and CD40 as in blood. Thus, it would be logical to think that the function of plasmacytoid CSF dendritic cells is the production of type I IFNs and that these cells in CSF may have been primed to produce IFN-. Staining of blood and CSF mononuclear cells revealed, however, that levels of IFN- are generally elevated in these cells compared with their blood counterparts, but there is no difference between plasmacytoid dendritic cells, myeloid dendritic cells and non-dendritic cells in the same compartment. These data are reminiscent of those published by Cella and colleagues, who found that basically all leucocytes in inflamed lymph nodes are positive for IFN-, although in their experiments the plasmacytoid dendritic cells expressed more IFN- than other cells (Cella et al., 2000). The presence of IFN- in the cells does not mean that it is biologically active, and further studies focusing on secretion of IFN- by isolated CSF dendritic cells and non-dendritic cells are necessary. It is also possible that expression of IFN- by plasmacytoid dendritic cells will increase further upon their direct contact with high concentrations of pathogens in meninges or brain parenchyma. The increased numbers of plasmacytoid dendritic cells in CSF from multiple sclerosis and optic neuritis patients may indirectly indicate possible involvement of viruses in the pathogenesis of these diseases.
Recent studies have shown that, depending on the mode of activation and specific cytokine milieu, myeloid and plasmacytoid dendritic cells can induce Th1- or Th2-type immune responses (Rissoan et al., 1999; Vieira et al., 2000). We observed that only myeloid CSF dendritic cells express CD80, a costimulatory molecule previously shown to be associated with Th1 responses (Kuchroo et al., 1995). This finding is in favour of Th1-deviating potential of myeloid CSF dendritic cells. However, direct experiments are required to demonstrate whether subsets of CSF dendritic cells can contribute to Th1/Th2 deviation of immune responses within the CNS.
Another question arising from this study regards the sources of dendritic cells in CSF. Dendritic cells can emerge in CSF from CNS-resident cells or from blood. It is likely that the source of plasmacytoid CSF dendritic cells is plasmacytoid dendritic cells from blood but not CNS-resident cells, since no cells with a plasmacytoid morphology have ever been described in normal meninges, CNS parenchyma, or brain cell cultures. The origin of myeloid dendritic cells in the CSF is less clear. They might arise from blood myeloid dendritic cells or from blood monocytes upon their transendothelial migration (Randolph et al., 1998), or represent dendriform cells, possibly arising from microglia (Fischer and Bielinsky, 1999). Plasmacytoid and myeloid dendritic cells in blood and CSF express CCR5, which is a receptor for RANTES and MIP-1. These two chemokines are upregulated in the brain and CSF of multiple sclerosis patients (Sorensen et al., 1999), and B. burgdorferi induces expression of RANTES and MIP-1 in vitro (Sprenger et al., 1997). Our preliminary results show that myeloid, and to a lesser extent plasmacytoid, blood dendritic cells respond chemotactically to RANTES. Thus, the system CCR5/RANTES may be involved in the chemoattraction of dendritic cells to CSF. Further studies are needed to investigate migration of dendritic cells through the blood–brain barrier.
In conclusion, plasmacytoid and myeloid dendritic cells are present in human CSF and are elevated in different neuroinflammatory conditions. The levels of CSF dendritic cells parallel the degree of CNS inflammation. The probable role of CSF dendritic cells in immunoregulation within the CNS needs to be investigated further.
|
Acknowledgments |
|---|
We wish to thank M. C. M. Kouwenhoven, L. Rinaldi, N. Teleshova and V. Özenci for valuable discussions and comments regarding this paper. This study was supported by the Swedish Medical Research Council, the Swedish Association of the Neurologically Disabled (NHR), and the Karolinska Institute.
|
References |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Banchereau J, Steinman RM. Dendritic cells and the control of immunity. [Review]. Nature 1998; 392: 245–52.[Medline]
Brosnan CF, Raine CS. Mechanisms of immune injury in multiple sclerosis. [Review]. Brain Pathol 1996; 6: 243–57.[ISI][Medline]
Cella M, Jarrossay D, Facchetti F, Alebardi O, Nakajima H, Lanzavecchia A, et al. Plasmacytoid monocytes migrate to inflamed lymph nodes and produce large amounts of type I interferon. Nat Med 1999; 5: 919–23.[ISI][Medline]
Cella M, Facchetti F, Lanzavecchia A, Colonna M. Plasmacytoid dendritic cells activated by influenza virus and CD40L drive a potent Th1 polarization. Nat Immunol 2000; 1: 305–10.[ISI][Medline]
Dittel BN, Visintin I, Merchant RM, Janeway CA Jr. Presentation of the self antigen myelin basic protein by dendritic cells leads to experimental autoimmune encephalomyelitis. J Immunol 1999; 163: 32–9.
Feldman SB, Ferraro M, Zheng HM, Patel N, Gould-Fogerite S, Fitzgerald-Bocarsly P. Viral induction of low frequency interferon-alpha producing cells. Virology 1994; 204: 1–7.[ISI][Medline]
Fischer HG, Bielinsky AK. Antigen presentation function of brain-derived dendriform cells depends on astrocyte help. Int Immunol 1999; 11: 1265–74.
Greenway AL, Hertzog PJ, Devenish RJ, Linnane AW. Constitutive and virus-induced interferon production by peripheral blood leukocytes. Exp Hematol 1995; 23: 229–35.[ISI][Medline]
Grouard G, Rissoan MC, Filgueira L, Durand I, Banchereau J, Liu YJ. The enigmatic plasmacytoid T cells develop into dendritic cells with interleukin (IL)-3 and CD40-ligand. J Exp Med 1997; 185: 1101–11.
Kohrgruber N, Halanek N, Groger M, Winter D, Rappersberger K, Schmitt-Egenolf M, et al. Survival, maturation, and function of CD11c– and CD11c+ peripheral blood dendritic cells are differentially regulated by cytokines. J Immunol 1999; 163: 3250–9.
Kostulas VK, Link H. Agarose isoelectric focusing of unconcentrated CSF and radioimmunofixation for detection of oligoclonal bands in patients with multiple sclerosis and other neurological diseases. J Neurol Sci 1982; 54: 117–27.[ISI][Medline]
Kuchroo VK, Das MP, Brown JA, Ranger AM, Zamvil SS, Sobel RA, et al. B7–1 and B7–2 costimulatory molecules activate differentially the Th1/Th2 developmental pathways: application to autoimmune disease therapy. Cell 1995; 80: 707–18.[ISI][Medline]
Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983; 33: 1444–52.
Link H, Tibbling G. Principles of albumin and IgG analyses in neurological disorders. III. Evaluation of IgG synthesis within the central nervous system in multiple sclerosis. Scand J Clin Lab Invest 1977; 37: 397–401.[ISI][Medline]
Link H, Huang YM, Xiao BG. Dendritic cells in experimental allergic encephalomyelitis and multiple sclerosis. [Review]. J Neuroimmunol 1999; 100: 102–10.[ISI][Medline]
Litton MJ, Andersson J, Björk L, Fehniger T, Ulfgren AK, Andersson U. Cytoplasmic cytokine staining in individual cells. In: Debets R, Savelkoul HF, editors. Human cytokine protocols. Clifton (NJ): Humana Press; 1997. p. 34–44.
Matyszak MK, Perry VH. The potential role of dendritic cells in immune-mediated inflammatory diseases in the central nervous system. Neuroscience 1996; 74: 599–608.[ISI][Medline]
McMenamin PG. Distribution and phenotype of dendritic cells and resident tissue macrophages in the dura mater, leptomeninges, and choroid plexus of the rat brain as demonstrated in wholemount preparations. J Comp Neurol 1999; 405: 553–62.[ISI][Medline]
O'Doherty U, Peng M, Gezelter S, Swiggard WJ, Betjes M, Bhardwaj N, et al. Human blood contains two subsets of dendritic cells, one immunologically mature and the other immature. Immunology 1994; 82: 487–93.[ISI][Medline]
Olweus J, BitMansour A, Warnke R, Thompson PA, Carballido J, Picker LJ et al. Dendritic cell ontogeny: a human dendritic cell lineage of myeloid origin. Proc Natl Acad Sci USA 1997; 94: 12551–6.
Poser CM, Paty DW, Scheinberg L, McDonald WI, Davis FA, Ebers GC, et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 1983; 13: 227–31.[ISI][Medline]
Randolph GJ, Beaulieu S, Lebecque S, Steinman RM, Muller WA. Differentiation of monocytes into dendritic cells in a model of transendothelial trafficking. Science 1998; 282: 480–3.
Rissoan MC, Soumelis V, Kadowaki N, Grouard G, Briere F, de Waal Malefyt R, et al. Reciprocal control of T helper cell and dendritic cell differentiation. Science 1999; 283: 1183–6.
Robinson SP, Patterson S, English N, Davies D, Knight SC, Reid CD. Human peripheral blood contains two distinct lineages of dendritic cells. Eur J Immunol 1999; 29: 2769–78.[ISI][Medline]
Siegal FP, Kadowaki N, Shodell M, Fitzgerald-Bocarsly PA, Shah K, Ho S, et al. The nature of the principal type 1 interferon-producing cells in human blood. Science 1999; 284: 1835–7.
Söderström M, Ya-Ping J, Hillert J, Link H. Optic neuritis: prognosis for multiple sclerosis from MRI, CSF, and HLA findings. Neurology 1998; 50: 708–14.
Sorensen TL, Tani M, Jensen J, Pierce V, Lucchinetti C, Folcik VA, et al. Expression of specific chemokines and chemokine receptors in the central nervous system of multiple sclerosis patients. J Clin Invest 1999; 103: 807–15.[ISI][Medline]
Sprenger H, Krause A, Kaufmann A, Priem S, Fabian D, Burmester GR, et al. Borrelia burgdorferi induces chemokines in human monocytes. Infect Immun 1997; 65: 4384–8.[Abstract]
Steinman RM, Pack M, Inaba K. Dendritic cell development and maturation. In: Ricciardi-Castagnoli, P, editor. Dendritic cells in fundamental and clinical immunology, Vol. 3. New York: Plenum Press; 1997. p. 1–6.
Svensson H, Johannisson A, Nikkila T, Alm GV, Cederblad B. The cell surface phenotype of human natural interferon-alpha producing cells as determined by flow cytometry. Scand J Immunol 1996; 44: 164–72.[ISI][Medline]
Thomas R, Lipsky PE. Human peripheral blood dendritic cell subsets. Isolation and characterization of precursor and mature antigen-presenting cells. J Immunol 1994; 153: 4016–28.[Abstract]
Ulvestad E, Williams K, Bjerkvig R, Tiekotter K, Antel J, Matre R. Human microglial cells have phenotypic and functional characteristics in common with both macrophages and dendritic antigen-presenting cells. J Leukoc Biol 1994; 56: 732–40.[Abstract]
Vieira PL, de Jong EC, Wierenga EA, Kapsenberg ML, Kalinski P. Development of Th1-inducing capacity in myeloid dendritic cells requires environmental instruction. J Immunol 2000; 164: 4507–12.
Received July 10, 2000. Revised October 20, 2000. Accepted October 30, 2000.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. M. Santer, T. Yoshio, S. Minota, T. Moller, and K. B. Elkon Potent Induction of IFN-{alpha} and Chemokines by Autoantibodies in the Cerebrospinal Fluid of Patients with Neuropsychiatric Lupus J. Immunol., January 15, 2009; 182(2): 1192 - 1201. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. L. Zozulya, S. Ortler, J. Lee, C. Weidenfeller, M. Sandor, H. Wiendl, and Z. Fabry Intracerebral Dendritic Cells Critically Modulate Encephalitogenic versus Regulatory Immune Responses in the CNS J. Neurosci., January 7, 2009; 29(1): 140 - 152. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|