Learn more: PMC Disclaimer | PMC Copyright Notice
Natural killer cell subsets in cerebrospinal fluid of patients with multiple sclerosis
Associated Data
Abstract
Changes in blood natural killer (NK) cells, important players of the immune innate system, have been described in multiple sclerosis (MS). We studied percentages and total cell counts of different effector and regulatory NK cells in cerebrospinal fluid (CSF) of MS patients and other neurological diseases to gain clearer knowledge of the role of these cells in neuroinflammation. NK cell subsets were assessed by flow cytometry in CSF of 85 consecutive MS patients (33 with active disease and 52 with stable MS), 16 with other inflammatory diseases of the central nervous system (IND) and 17 with non-inflammatory neurological diseases (NIND). MS patients showed a decrease in percentages of different CSF NK subpopulations compared to the NIND group. However, absolute cell counts showed a significant increase of all NK subsets in MS and IND patients, revealing that the decrease in percentages does not reflect a real reduction of these immune cells. Remarkably, MS patients showed a significant increase of regulatory/effector (CD56bright/CD56dim) NK ratio compared to IND and NIND groups. In addition, MS activity associated with an expansion of NK T cells. These data show that NK cell subsets do not increase uniformly in all inflammatory neurological disease and suggest strongly that regulatory CD56bright and NK T cells may arise in CSF of MS patients as an attempt to counteract the CNS immune activation characteristic of the disease.
Introduction
Multiple sclerosis (MS) is considered to be a T cell-mediated autoimmune disease that results in demyelination and axonal loss inducing irreversible neurological deficits. Recent studies suggest that the innate immune system plays an important role in both the initiation and progression of MS by influencing the effector function of T and B cells 1. Natural killer (NK) cells contribute to both effector and regulatory functions of innate immunity via their cytotoxic activity and their ability to secrete pro- and anti-inflammatory cytokines and growth factors 1. Mechanisms by which NK cells could have an impact on autoimmune responses include a rapid cytokine release before autoreactive T cell differentiation and modulation of interactions between autoreactive T and B lymphocytes and antigen-presenting cells 2. Several NK subsets responsible for different functions have been identified 3. Based on the surface expression of CD56, NK cells have been classified as effector (CD56dim) or regulatory (CD56bright) cells. Effector NK cells are characterized by intracellular expression of perforin and granzymes, which are proteolytic enzymes involved in target killing 4,5. They may be subdivided further into CD56+CD16+ (cytotoxic effector cells) 6 and CD56+CD3+ (NK T cells), which may produce a variety of cytokines and control other immune cells 7. Conversely, CD56bright NK regulatory cells represent fewer than 10% of peripheral blood NK cells, express low levels of perforin and are able to secrete large amounts of cytokines 1,8. In addition, CD56–CD16+ NK cells are a defective NK subset with impaired cytolytic function that increases in viraemic HIV and hepatitis C virus (HCV)-infected individuals 9,10. A high level of these dysfunctional cells reveals a disturbance in innate cellular immunity that associates with an impaired ability to respond to anti-viral treatment with interferon (IFN)-α and ribavirin 10.
It has been suggested that NK cells play a key role at the interface of innate and adaptive responses in autoimmune diseases 11,12. Depending on the cell subtype and milieu, both pathogenic and protective roles in the central nervous system (CNS) have been attributed to different NK subpopulations in MS and its animal model, experimental autoimmune encephalomyelitis (EAE) 13–17. In keeping with this, NK cell lines may induce lysis of oligodendrocytes. Alternatively, circulating cytotoxic NK cells may lyse autoantigen-specific encephalitogenic T cells, dampening disease activity in EAE. In addition, CNS-resident NK cells may inhibit T helper type 17 (Th17) differentiation in EAE by interacting with microglia cells 2.
It is considered that immunophenotyping of cerebrospinal fluid (CSF) cells in MS reflects cellular events within brain parenchyma and may be useful to gain insight into MS pathophysiology 18. Different studies have described changes in the percentages of NK cells in the CSF of MS patients 19–22. Moreover, the beneficial effect of daclizumab, a humanized monoclonal antibody (mAb) that blocks the binding site on the IL-2Rα chain (CD25), is linked to the expansion of the CD56bright cells in CSF 23. We considered that it would be of interest to investigate whether changes in CSF NK cells occurring in MS are relative or absolute, and if they associate with a particular NK cell subset. In addition, we studied the relationship between CSF NK subpopulations and MS activity, which has not been explored so far.
Material and methods
Patients
This study was approved by the ethical committee of Hospital Ramón y Cajal (Madrid, Spain). Written informed consent was obtained from all patients before entry into the study. We included 85 consecutive patients diagnosed with relapsing–remitting MS (RRMS), according to modified McDonald criteria 24. We monitored disease duration, current disability measured by the Expanded Disability Status Scale (EDSS) score and disease status (relapse or remission). Relapses were defined as worsening of neurological impairment or the appearance of new symptoms attributable to MS, lasting at least 24 h and preceded by stability of at least 1 month. Remission was considered when patients were free of relapses or progression for at least 3 months before inclusion in the study. We also divided our patients according to clinical inflammatory disease activity (IDA) status. The active group included patients with the presence of either a relapse or new magnetic resonance imaging (MRI) gadolinium-enhanced lesions (n = 33). The stable group included MS individuals in remission who did not show any of these variables related to disease activity (n = 52).
Eighty-one MS patients (95·29%) had not received any disease-modifying treatment before lumbar puncture (LP). The remaining four cases were treated previously with Rebif (three cases) or Betaferon (one case). Another three patients received high-dose steroids before inclusion in this study. In all these cases a washout period of 3 months between treatment cessation and LP was established.
We also studied 16 patients who could be defined as inflammatory neurological disease controls (IND), according to the latest consensus definitions 25. Patients with the following diagnoses were included in this group: limbic encephalitis (n = 2), post-infectious cerebelitis (n = 1), bilateral optic neuritis (n = 4), herpes simplex encephalitis (n = 1), headache associated with neurological deficits and CSF lymphocytosis (HaNDL) syndrome (n = 3), transverse myelitis (n = 1), vasculitis (n = 2), pachymeningitis (n = 1) and neuromyelitis optica (n = 1). Seventeen patients with the following non-inflammatory neurological disease controls (NIND) were also studied: stroke (n = 1), normal pressure hydrocephalus (n = 3), dementia (n = 4), epilepsy (n = 1), pseudotumour cerebri (n = 5), amyotrophic lateral sclerosis (n = 1), pineal cyst (n = 1) and arachnoid cyst (n = 1). No significant differences were found in age and sex between MS [age: 38·86 ± 1·0 years (mean ± standard error); 75% females], IND (age: 42·09 ± 4·87 years; 66·7% females) and NIND (age: 46·26 ± 3·56; 58·8% females) groups. Basic CSF parameters of the patients included in the three groups are shown in the Supporting information, Fig. S1.
MRI scans of the brain were performed within a month of CSF sampling at 1·5 Tesla (Phillips Gyroscan NT, Eindhoven, the Netherlands). New brain-active gadolinium-enhanced lesions, T1 and T2 lesion load were studied following a standard protocol 26
Samples
Samples were always obtained for clinical purposes. Peripheral blood (PB) samples were collected at the time of the LB and were processed in parallel with CSF. Fresh CSF samples (4–6 ml) were centrifuged at 500 g for 15 min and the cellular pellet resuspended in 100 μl of phosphate-buffered saline (PBS) to be labelled as described below.
Flow cytometry analysis
CSF and PB cells were analysed for expression of surface markers using flow cytometry. The following monoclonal antibodies were used: control mouse isotypes and anti-human CD3, CD16, CD45 and CD56 (BD Biosciences, San Jose, CA, USA). Cells were labelled with optimal concentrations of these monoclonal antibodies. CSF cell staining was performed at 4°C in the dark and washes and incubations were carried out in PBS. Whole PB samples were labelled for 20 min at room temperature, and then lysed with 2 ml of lysis solution [fluorescence activated cell sorter (FACS) Lysis Solution; Becton Dickinson, San jose, CA, USA]. Cells were then washed twice. Data acquisition was performed with a FACSCanto II cytometer and analysed with FACSDiva software (BD Immunocytometry Systems, San Jose, CA, USA).
An initial region was set around cells expressing intermediate to high CD45 with low to intermediate side-scatter, and then a second region was set on the forward-/side-scatter dot-plot to exclude debris or apoptotic cells and include lymphocytes. Only cells that included both regions were accepted for analysis. A minimum of 500 events were collected for analysis of antigen staining in CSF. The cursor was set so that fewer than 1% of the cells in each sample stained positively with the isotype control antibodies. The percentage and total counts of cells that stained positively was recorded for each sample. The results were reported as percentages of total lymphocytes and as absolute cell counts. CD56dim and CD56bright NK cell subsets were identified according to the staining intensity with the specific mAb. Representative examples of cell gating are shown in Fig. 1 and the Supporting information, Fig. S2. Every sample was analysed by immunologists blind to clinical data.
Representative dot-plots showing gating strategy to select natural killer (NK) cells for analysis. (a) Cerebrospinal fluid (CSF) lymphocytes were identified on a dot-plot display with side-scatter (SSC) and CD45. A gate was set around CD45+ bright cells with low side-scatter. (b) A second region was then set on the forward (FSC)-/side-scatter dot-plot to select a lymphocyte population free of debris and apoptotic cells. (c) Finally, CSF NK cells were identified on a dot-plot display with SSC and CD56. A gate was then set around CD56dim cells and a second region was set around CD56bright cells.
Statistical analysis
Results were analysed with the Prism version 5·0 statistical package (GraphPad Software, San Diego, CA, USA). We used the Mann–Whitney U-test for comparisons between two groups and the Kruskal–Wallis test with Dunn's post-hoc test for comparisons between more than two groups. P-values below 0·05 were considered significant.
Results
The percentages of different NK cell subsets in CSF samples of our three patient groups (MS, IND and NIND) are depicted in Table 1. No differences were found in either the percentages of total NK CD56+ or regulatory NK CD56bright cells between the three patient groups. However, MS patients showed lower percentages of total effector NK (CD56dim) cells than the NIND group (P < 0·001). This decrease was due to the NK T (CD56dim CD3+) subset, which was diminished in MS patients compared to NIND patients (P < 0·001). The defective CD56–CD16+ NK subset was also diminished in MS patients compared to the NIND group (P < 0·01). These data demonstrate a relative decrease of effector NK cells in MS patients compared to non-inflammatory diseases of the CNS. Other inflammatory diseases of the CNS also showed a decreased NK T cell percentage compared to NIND (P < 0·05).
Table 1
Cerebrospinal fluid (CSF) percentages of different natural killer (NK) cell subpopulations
| CSF NK cells | MS | IND | NIND | P-values |
|---|---|---|---|---|
| n = 85 | n = 16 | n = 17 | ||
| NK CD56+ cells | 3·00 (2·20–4·15) | 2·85 (2·02–4·72) | 5·30 (2·95–8·25) | P = 0·04; a,b,c: n.s. |
| CD56bright cells | 1·40 (0·80–2·25) | 0·90 (0·32–2·62) | 1·60 (0·85–2·95) | P = 0·19; a,b,c: n.s. |
| CD 56dim cells | 1·40 (1·0–2·0) | 1·70 (1·07–2·57) | 2·90 (1·70–4·5) | P = 0·001; a,c: n.s., b*** |
| CD 56dimCD3+ | 1·10 (0·70–1·45) | 1·20 (0·72–1·50) | 2·0 (1·40–3·60) | P = 0·001; a: n.s., b***, c* |
| CD56dimCD16+ | 0·30 (0·15–0·50) | 0·35 (0·20–0·65) | 0·30 (0·20–0·90) | P = 0·52; a,b,c: n.s. |
| CD56−CD16+ cells | 1·30 (0·80–2·10) | 1·45 (0·92–3·3) | 3·2 (1·55–5·30) | P = 0·003; a,c: n.s., b** |
Values shown are calculated as percentage of lymphocytes in CSF and presented as median (25–75% interquartile range). MS = multiple sclerosis; IND = other inflammatory neurological diseases; NIND = non-inflammatory neurological diseases. P = P-value of overall comparison (Kruskal–Wallis test); a = Dunn's post-hoc test comparisons between a: MS versus IND, b: MS versus NIND, c: IND versus NIND.
We also studied the ratio between regulatory and effector (CD56bright/CD56dim) NK cells in CSF. MS patients showed a significantly higher ratio than the IND and NIND groups, which presented similar CD56bright/CD56dim ratio values (Fig. 2a). This was the best variable to discriminate MS and IND patients in our series. Representative dot-plots are shown in Fig. 2b.
(a) CD56 bright/dim ratio in cerebrospinal fluid (CSF) of patients with multiple sclerosis (MS), other inflammatory neurological diseases (IND) and non-inflammatory neurological diseases (NIND). P = P-value of overall comparison (Kruskal–Wallis test). *P < 0·05 in Dunn's post-hoc test comparison. (b) Representative dot-plots showing CD56 cells from a patient with MS, other inflammatory neurological disease (IND) and non-inflammatory neurological disease (NIND).
We next studied total CSF leucocyte counts in the three groups, and observed that MS and IND patients showed clearly higher cell numbers than the NIND group (P < 0·0001). Accordingly, they also had increased values of all NK subsets (Table 2).
Table 2
Absolute number of different natural killer (NK) cell subpopulations in cerebrospinal fluid (CSF)
| Absolute cell counts | MS | IND | NIND | P-values |
|---|---|---|---|---|
| n = 85 | n = 16 | n = 17 | ||
| Total cell number | 1219 (485·3–2738) | 1439 (698·6–6426) | 197·6 (150·1–327·3) | P < 0·0001 a: n.s., b***, c*** |
| NK CD56+ cells | 31·43 (14·21–86·11) | 49·98 (18·77–83·41) | 13·20 (6·71–14·75) | P < 0·0001 a: n.s., b***, c*** |
| CD56bright cells | 14·66 (5·37–49·96) | 10·30 (4·67–35·83) | 3·98 (2·40–5·81) | P < 0·0001 a: n.s., b***, c* |
| CD56dim cells | 17·61 (6·96–37·82) | 30·83 (10·99–53·60) | 5·63 (3·42–9·96) | P < 0·0001 a: n.s., b***, c: *** |
| CD56dimCD3+ | 11·48 (5·57–24·12) | 19·0 (7·33–42·09) | 3·53 (2·88–8·56) | P = 0·0001 a: n.s., b***, c*** |
| CD56dimCD16+ | 3·11 (0·96–9·39) | 6·55 (1·22–18·54) | 0·66 (0·34–2·15) | P = 0·0022 a: n.s., b**, c** |
| CD56−CD16+ cells | 18·85 (7·31–36·08) | 31·05 (10·03–105·0) | 6·59 (2·78–19·62) | P = 0·0066 a: n.s., b*, c** |
Values are calculated as number of cells/ml and presented as median (25–75% interquartile range). MS = multiple sclerosis; IND = other inflammatory neurological diseases; NIND = non-inflammatory neurological diseases; P = P-value of overall comparison (Kruskal–Wallis test); a: Dunn's post-hoc test comparison between MS versus IND, b: Dunn's post-hoc test comparison between MS versus NIND, c: Dunn's post-hoc test comparison between IND versus NIND.
Although no significant differences were found between the MS and IND groups, IND patients tended to show higher number of NK effector cells and, conversely, the MS group seem to have higher CSF CD56bright cell counts (Table 2).
The distribution of the NK cell subsets in peripheral blood is shown in the Supporting information, Table S1. No significant differences were observed between the three groups of patients.
In a further subanalysis, we assessed the influence of MS clinical inflammatory disease activity (IDA) on CSF NK cell subsets. We found no differences in CSF NK cell percentages associated with inflammatory status (Supporting information, Table S2). However, the study of absolute cell numbers provided significant results (Table 3). We found a moderate increase in total cell numbers in the active MS group (P = 0·033). Differences were higher when studying CD56dim cells. This was due mainly to an increase of the CD56dimCD3+ subpopulation in active patients (P = 0·018). The increase of CD56dimCD16+ cells was less relevant (P = 0·043). No significant differences were found in CD56bright or CD56–CD16+ subsets in relation to disease activity.
Table 3
Absolute cell numbers in cerebrospinal fluid (CSF)of active and stable multiple sclerosis (MS) patients
| MS active | MS stable | P-values | |
|---|---|---|---|
| n = 33 | n = 52 | ||
| Total cell number | 1854 (802·1–3885) | 849·9 (427·0–2567) | P = 0·033 |
| NK CD56+ cells | 55·81 (21·30–128·8) | 24·08 (12·43–77·39) | P = 0·022 |
| CD56bright cells | 19·31 (9·12–80·95) | 12·16 (4·65–42·83) | P = 0·064 |
| CD 56dim cells | 22·59 (10·66–54·80) | 13·29 (5·98–33·65) | P = 0·013 |
| CD 56dim CD3+ | 16·25 (8·73–33·90) | 9·77 (5·17–19·43) | P = 0·018 |
| CD56dim CD16+ | 5·80 (1·66–13·27) | 2·13 (0·75–8·39) | P = 0·043 |
| CD56−CD16+ cells | 23·86 (7·84–37·65) | 14·14 (5·99–34·71) | P = 0·21 |
Values are calculated as number of cells/ml and presented as median (25–75% interquartile range).
We found no differences in percentages or total numbers of any NK cell subset between active MS patients and the IND group (data not shown). However, the bright/dim ratio remained higher in active MS [1·0 (0·57–1·89), median (25–75% interquartile range)] when compared to IND patients [0·43 (0·25–1·28), P = 0·019].
Discussion
Several studies have implicated NK cells in the pathogenesis of MS. Using functional assays, NK cell dysfunction related to MS activity was found in the blood and CSF of patients 27–29. In addition, immunophenotyping studies revealed that blood NK cell subpopulations seem to play a role in disease activity and response to treatment. In keeping with this, the expansion of regulatory CD56bright NK cells associated with a good response to daclizumab 23,30 and IFN-β 31,32, while CD56dim cells decreased after immunomodulatory therapy 32.
The percentage of effector NK cells has been shown to be decreased in the CSF of MS compared to NIND patients 19. We have confirmed these differences for MS, and observe that patients with IND also have a lower proportion of these cells. This suggests strongly that the decrease of effector NK cells is associated not only with MS, but with CNS inflammation. In contrast, the CSF CD56bright/CD56dim ratio seemed to be increased specifically in MS patients compared to IND or NIND groups. We assessed this by comparing the values of this ratio in patients with active MS and IND. Active MS patients also showed significantly higher values of this ratio than those with other inflammatory diseases. This further supports the association of high levels of the CSF CD56bright/CD56dim ratio with MS and may reflect the existence of different immunoregulatory mechanisms in MS and IND. Further studies will demonstrate the role of CSF CD56bright NK cells in MS.
Conversely to the CSF data, we did not find differences in the percentages of any NK cell subset in peripheral blood between our three groups of patients, thus confirming previously published data on the poor correlation between systemic and CSF lymphoid cell subsets 23.
We next aimed to study absolute numbers of CSF NK cell subsets in our series, as percentages and absolute numbers of CSF cells can differ widely, depending on the clinical setting 33. All NK cell subsets were increased significantly in MS and IND compared to NIND, which suggests that these cells may play a role in CNS inflammation. However, we observed that the increase in the number of regulatory CD56bright cells was much clearer in MS patients than in IND. This confirms the CD56bright/CD56dim ratio data. Conversely, the expansion in CD56–CD16+ defective NK cells was more pronounced in IND patients, suggesting that the NK cell profile can differ between different inflammatory neurological diseases.
The relationship of CSF NK subpopulations and MS activity has not been studied extensively. We explored this in 33 patients with active disease and 52 with stable MS. The total CSF cell count was increased moderately in active patients. However, when exploring regulatory or defective NK cells, we did not find any changes between active and stable patients. Differences were found only in effector subsets, due mainly to an increase in NK T cell numbers. This subset is a heterogeneous lymphoid population that may have both immune-enhancing and immunosuppressive roles. It can be classified by studying the heterogeneity of T cell receptor (TCR) rearrangements into two categories, type I and type II 34. Type I, also denominated as invariant NK T (iNK T) cells, shows strong immune regulatory properties. iNK T cells exert a protective role in EAE 35, and expand in MS in response to IFN-β treatment 36.
The role of type II NK T cells has not yet been demonstrated so clearly, but these cells can contribute to enhance the inflammatory response in different autoimmune diseases 37. The low number of cells present in CSF prevented us from further investigating NK T isotypes in our patients. Future research will identify the nature of NK T cells present in the CSF of MS patients and demonstrate their relationship with disease activity.
In summary, our data show that absolute CSF cell numbers add important information to previous percentage data and could contribute to reveal the role of NK cells in inflammatory neurological diseases.
Acknowledgments
We acknowledge Asunción Fernández for her relevant work with patients and Belén Bonilla and Daniel Carpio for their technical work. This work was supported by grants from Plan Estatal de I+D+I 2013–2016, PI12-00239 from FIS, Instituto de Salud Carlos III and FEDER and SAF 2012-34670 from Ministerio de Economía y Competitividad. Raquel Alenda is recipient of a research contract of REEM from the Instituto de Salud Carlos III (Spain).
Disclosure
None declared.
Author contributions
E. R. M., L. M. V. and J. C. A. C. designed the study. C. P., R. A., E. R. and M. E. performed the experiments. L. C. F. and S. S. M. obtained patient samples and clinical data. E. R. M. and L. M.V. analysed the results. E. R. M., L. M. V. and J. C. A. C. drafted the manuscript. All the authors made a critical review of the manuscript.
Supporting Information
Additional Supporting information may be found in the online version of this article at the publisher's web-site:
Fig. S1. Basic cerebrospinal (CSF) parameters of the multiple sclerosis (MS), other inflammatory diseases of the central nervous system (IND) and non-inflammatory neurological diseases (NIND) patients included in the present study. (a) Albumin index, (b) immunoglobulin (Ig)G index, (c) number of leucocytes/mm3. P = P-value of overall comparison (Kruskal–Wallis test) and Dunn's post-hoc test comparison between groups. *P < 0·05; **P < 0·01; ***P < 0·001.
Fig. S2. Representative dot-plots showing gating strategy in cerebrospinal (CSF) to select natural killer (NK) cell subsets for analysis. (a) CSF NK cells were identified on a dot-plot display with side-scatter (SSC) and CD56. A gate was then set around CD56dim cells. (b) CD56dimCD3+ and CD56dimCD16+ cells were identified in a CD16, CD3 two-colour dot-plot. (c) CSF CD16+ cells were identified on a dot-plot display with SSC and CD16. (d) CD56–CD16+ cells were identified in a CD16, CD56 two-colour dot-plot.
Table S1. Peripheral blood (PB) percentages of different natural killer (NK) cell subpopulations.
Table S2. Percentages of natural killer (NK) cell subtypes in cerebrospinal (CSF) of active and stable multiple sclerosis (MS) patients.