Objectives: Evidence that chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) is an autoimmune disease was sought, by studying cellular and humoral immune responses to peripheral nerve myelin proteins.
Methods: 40 CIDP, 36 healthy control subjects (HC) and subjects with non-immune mediated neuropathies (other neuropathies, ON) for antibodies were studied by ELISA and cellular responses by cytokine ELISPOT (INFγ, IL10) and ELISA (IL17) to synthetic peptides representing P0, P2 and PMP22.
Results: Antibodies to P0, P2 or PMP22 peptides were detected in only a minority of CIDP, both not treated (nT-CIDP) and treated (T-CIDP). IgG antibodies to P280–105 were significantly more frequent in CIDP than in HC (4/30 vs 0/32; p<0.05) but the difference from ON (1/25) was not significant. In ELISPOT assays, IFNγ was detected at a low frequency in CIDP and did not differ from HC or ON. In contrast, IL10 responses against P21–85 were more frequent in nT and T-CIDP (7/24 and 3/16) than HC (0/36; p<0.001 and p<0.05, respectively). The production of IL17 in cell-culture supernatants was not increased.
Conclusions: Antibodies to non-conformational antigenic epitopes of myelin proteins rarely occur in CIDP. None of the myelin protein peptides elicited IFNγ responses, but P2 elicited IL10 responses significantly more often in CIDP patients than in controls. This reactivity may be part of an antigen-specific Th2 type pathogenetic or regulatory mechanism or represent a transitory epiphenomenon due to nerve damage. In our study, P2 was the protein antigen most likely to be involved in the aberrant immune responses in CIDP.
Statistics from Altmetric.com
Chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) is an acquired disorder of the peripheral nervous system with a probable autoimmune pathogenesis. The disease can present with the classic form (typical CIDP) characterised by distinctive clinical, electrophysiological and pathological features or with several variants (atypical CIDP).1 The target antigens and immunological mechanisms underlying the disease have not yet been identified, although the involvement of both humoral and cell-mediated immune responses has been documented.2 2–5
Gangliosides have been much studied in inflammatory neuropathy, but antibodies to them are uncommon in CIDP.2 6 Peripheral nerve myelin proteins are important, potentially antigenic components of myelin.7 Myelin protein zero (P0), myelin protein 2 (P2) and peripheral myelin protein 22 (PMP22) are the principal candidates as target antigens because they can each induce experimental autoimmune neuritis (EAN).2 8 Antibodies to myelin proteins have been reported with variable frequencies in CIDP in previous studies but not consistently.4 9–16 Cellular reactivity to myelin proteins has only been studied infrequently.5 17–20 Neither the frequency nor the Th1-Th2 polarity of the cell mediated responses to these antigens has been clearly demonstrated.
We sought to clarify this conflicting evidence by measuring in the blood of CIDP and controls antibody and T cell responses directed against peptides of P0, P2 and PMP22. The synthetic overlapping peptides used in this study represented the almost complete sequence of these proteins, therefore investigating a more comprehensive range of epitopes than had been done before.
Forty CIDP patients were recruited at Guy’s Hospital in London (UK) between July 2005 and July 2007 (supplementary tables 1, 2). The study was approved by the Guy’s Hospital ethics committee, and participants gave informed consent. Patients met the Possible, Probable or Definite criteria for Typical or Atypical CIDP of the EFNS/PNS Joint Task Force.1 Twenty-four patients were not being treated (nT-CIDP), and 16 were on immunosuppressant or immunomodulatory treatment (T-CIDP).
The 36 healthy controls (HCs) were relatives of patients or volunteers without symptoms or history of peripheral nerve disease. There were 18 men and 18 women, mean age 53.0 (SD 11.6) years.
The other neuropathy (ON) control group for the antibody study were 25 patients with chronic idiopathic axonal polyneuropathy (CIAP).21 For the cytokine study, ON control samples were freshly collected from 22 patients with non-immune-mediated peripheral neuropathies. This group was heterogeneous: eight CIAP, seven hereditary neuropathy (four Charcot-Marie-Tooth type 1A and three type 2), two small fibre neuropathy, one diabetic, one paraneoplastic and one amyloid neuropathy. There were 11 men and 11 women, mean age 58.5 (12.9) years.
For the antibody study, serum was separated from clotted blood and stored at −70°C until use. For the cytokine study, peripheral blood mononuclear cells (PBMC) were freshly isolated on a density gradient (Lymphoprep, Norway) from heparinised samples.
We used 23 synthetic overlapping peptides representing almost the entire sequence of P0, P2 and PMP22 as reported in our study of GBS,22 except for the addition of the cytoplasmic domain peptide P0210–234.
Detection of anti-peptide antibodies by ELISA
Nunc Maxisorp 96-well plates (NUNC) were coated with 5 μg/ml of each peptide in carbonate coating buffer for 4 h at 37°C. Plates were washed and blocked for 1 h at 37°C or overnight at 4°C with 1% bovine serum albumin (Sigma Aldrich, St Louis, Missouri) in phosphate buffer saline. After washing steps, human serum was incubated at 37°C for 1 h at 1:50 or increasing dilutions up to 1:1600. After washing, bound antibody was detected with mouse anti-human IgG or IgM monoclonal secondary antibody conjugated to alkaline phosphatase (Sigma) and developed with p-nitrophenylphosphate.
All samples were tested in triplicate. The optical density (OD) obtained with serum alone was subtracted from the OD of the serum with the peptide of interest. A sample was deemed positive if the OD was more than the mean plus three SD of the sera of healthy controls.
Interferon γ (IFNγ) and interleukin 10 (IL10) production by PBMC was measured by cytokine ELISPOT as described in our previous study22 after a 48 h culture of freshly isolated PBMC in the presence of test peptide. The 23 peptides were pooled in two or three groups for each protein and used at the final concentration of 10 μg/ml for each peptide. P0 peptides were divided into three panels: P0 I (P030–54, P050–74, P070–94) corresponding to the first part of the extracellular domain of P0; P0 II (P090–114, P0110–134, P0130–154) corresponding to the second part of the extracellular domain of P0; P0 III (P0170–194, P0190–214, P0210–234, P0230–248) corresponding to part of the transmembrane region and the cytoplasmic tail. P2 peptides were divided into two panels: P2 I (P21–26, P220–45, P240–65, P260–85) corresponding to the first 85 amino acids of the protein and P2 II (P280–105, P2100–125, P2120–132) corresponding to the remainder. PMP22 peptides were divided into two panels: P22 I (PMP2220–45, PMP2240–65, PMP2280–105) corresponding to the first extracellular domain and part of the transmembrane region of the protein; P22 II (PMP22100–125, PMP22120–145, PMP22140–160) corresponding to the rest of the transmembrane region and the second extracellular domain.
Control wells contained medium with an equivalent concentration of peptide diluent (dimethyl sulfoxide, DMSO) alone. Tetanus toxoid (Aventis Pasteur, Maidenhead, UK) at a final concentration of 200 ng/ml was used as a positive control. Assays were performed in triplicate, and the mean spots per 3×105 cells were calculated. The values of DMSO alone were subtracted from the values in test wells for each set of peptides.
We considered as positive values higher than three SD above the mean of the healthy controls.
Detection of IL17 in cell culture supernatants by ELISA
The supernatants of the cell cultures of the ELISPOT study were collected after 48 h of incubation. Approximately 400 μl of supernatant was collected for each condition and stored at −80°C. The concentration of interleukin 17 (IL17) was determined using a commercial kit (R&D Systems, Amersham, UK) according to the manufacturer’s instructions. IL17 levels were expressed in picograms per ml (pg/ml), and the detection threshold was 7 pg/ml. The samples were tested in duplicate, and the mean values of wells with DMSO alone were subtracted from the mean values in test wells for each set of peptides.
All statistical analyses were performed using GraphPad Prism V.2.01 for Windows. For the antibody and cytokine study, the proportions of patients (nT-CIDP and T-CIDP) with positive responses above cut-off values were compared with the control groups (HC and CIAP) using χ2 or Fisher exact tests. The total number of spots produced in the ELISPOT study and the ELISA for IL17 were analysed using the Kruskal–Wallis test followed by the Dunn multiple comparison test. Tests of significance were two-tailed.
Thirty patients with CIDP, 14 men and 16 women, were enrolled in the antibody study. Their clinical features are summarised in supplementary table 1 (online supplementary material). The other neuropathy group consisted of 25 CIAP patients, 12 men and 13 women, mean age 65.1 (7.5) years. There were no significant differences in gender distribution between groups, but the ON group was significantly older than the HC (p<0.01).
Single peptides were tested for both IgG and IgM antibodies, except PMP2280–105 and PMP22140–160, due to difficulties in dissolving these peptides in coating buffer. We deemed as high titre IgG and IgM responses 200 or above and as low titre IgG and IgM responses below this value.
IgM to P0, P2 and PMP22
Three out of 30 CIDP sera (one nT-CIDP, titre 1:200; two T-CIDP, titres 1:200 and 1:800) and two out of 32 HC (titre 1:50 and 1:100) reacted with P220–45 (p = 0.66), but titres above 200 were detected only in CIDP sera (p = 0.1). One out of 30 CIDP sera (titre 1:400) and one out of 25 CIAP (titre 1:200) reacted with P2120–132, both at high titre, and two out of 32 HC reacted with PMP2240–65 at low titre (titre 1:50 and 1:100). No other reactivity to P2, PMP22 or P0 peptides was detected above cut-off values.
Considering the total number of patients and controls reacting to any of the peptides of the three proteins, there was no significant difference between patients and controls (4/30 CIDP, 4/32 HC, 1/25 CIAP, p = 0.46). Nevertheless, a titre above 200 was present only in the four CIDP and the CIAP patients. This proportion was significantly increased in CIDP compared with HC (4/30 CIDP vs 0/32 HC, p = 0.05), but not with CIAP (4/30 CIDP vs 1/25 CIAP; p = 0.36).
IgG to P0, P2 and PMP22
IgG reactivity to the single peptides was detected more frequently than IgM. The results are reported in table 1. IgG to P0 peptides were detected in seven out of 30 CIDP, three out of 32 HC and two out of 25 CIAP (p = 0.17). IgG responses at titres above 200 did not differ between patients and controls. Three CIDP patients (CIDP20, 27, 39), one HC (HC10) and one CIAP (CIAP64) showed multiple reactivities to two peptides of P0 which mostly represented adjacent sequences.
We found a trend in an increased proportion of patients with CIDP and IgG antibodies to P280–105 compared with the other groups (4/30 CIDP, 0/32 HC,1/25 CIAP, p = 0.07). When only untreated patients were considered, there was a significant difference between nT-CIDP and HC (4/18 CIDP vs 0/32 HC, p = 0.008), but not CIAP (4/18 CIDP vs 1/25 CIAP, p = 0.14). There were no significant differences in the proportion of IgG responses at high titre, and none of the subjects reacted to more than one peptide of P2.
IgG to PMP22 peptides were detected in three out of 30 CIDP (all untreated), one out of 32 HC and two out of 25 CIAP (p = 0.54). IgG antibodies at a titre above 200 were detected in only one CIAP patient. One CIDP patient (CIDP39) had low titre antibodies to two peptides of PMP22.
Two CIDP patients and none of the controls responded to more than one protein (CIDP39 to all, CIDP52 to P2 and PMP22). Considering the total number of patients and controls reacting to any of the peptides of the three proteins, there was no statistical difference between patients and controls (14/30 CIDP, 8/32 HC, 7/25 CIAP, p = 0.15). Similarly, there was no significant difference among groups in the total number of responses at high titres (5/30 CIDP, 4/32 HC, 5/25 CIAP, p = 0.74).
There was no significant difference in age and gender distribution between IgG responders and non-responders in either CIDP or controls, although CIDP responders tended to be more frequently males (CIDP responders: 9/14 males, CIDP non-responders: 5/16 males, p = 0.14) and HC responders tended to be older (mean age 60.4 (10.8) vs 53.5 (11.3), p = 0.11).
Cytokines responses to myelin proteins
All 40 patients with CIDP were enrolled in the study of cytokine responses. The clinical features are summarised in supplementary table 2 (online supplementary material). The controls consisted of 36 HC and 22 ON. There were no significant differences in age or gender distribution between groups. We compared the results either as the total number of spots or as the proportion of responders in each group.
Most samples showed absent or very low IFNγ responses to the tested peptides (data not shown). There were no significant differences either in the total number of spots or in the proportion of responders in response to the three groups of peptides of P0, P2 and PMP22. In the unstimulated cells, there was a trend in an increased number of IFNγ spots in T-CIDP compared with the other groups (T-CIDP 4.7 (7.0), nT-CIDP 1.5 (1.9), HC 1.2 (1.6), ON 1.2. (1.3), p = 0.38), and the difference reached statistical significance considering the proportion of responders (3/16 T-CIDP, 1/24 nT-CIDP, 1/36 HC, 0/22 ON, p = 0.05), suggesting underlying immune activation in this group.
Most of the peptides elicited low IL10 responses except P22 II and P0 III (fig 1). After stimulation with these two groups of peptides, both patients and controls showed a non-significant increase in responses. There was a trend towards an increase in total IL10 spots in T-CIDP (T-CIDP 3.9 (11.1), nT-CIDP 2.0 (2.9), HC 1.1 (1.8), ON 1.0 (1.5), p = 0.09) as well as in the proportion of positive responses (2/16 T-CIDP, 1/24 nT-CIDP, 0/36 HC, 0/22 ON, p = 0.08) in response to P22 I.
Although there was no significant difference in the total number of IL10 spots in response to P2 I, the proportion of positive tests in CIDP was significantly higher than in the other groups (7/24 nT-CIDP, 3/16 T-CIDP, 0/36 HC, 2/22 ON, p = 0.003). The difference was significant in both untreated and treated CIDP compared with HC (p = 0.0009 and p = 0.02, respectively) but not with ON (fig 2A). The response to P2 I was tested on a second occasion in seven out of 10 CIDP responders, five nT-CIDP and two T-CIDP (fig 3A). The second sampling was usually performed after a long interval from the first test, sometimes as much as 1 year. On this second occasion, the IL10 response to the single peptides belonging to the P2 I group was also tested (fig 3B). The response remained above cut-off in only two out of seven CIDP, both nT-CIDP (CIDP15 and CIDP29). In these two patients the reactivity was to P220–45 in CIDP15 and to P220–45 and P240–65 in CIDP29.
There was a significant increase in the total number of spots in nT-CIDP in response to P2 II (nT-CIDP 0.8 (1.3), T-CIDP 0.1 (0.2), HC 0.4 (1.0), ON 0.6 (0.8), p = 0.01); this increase was not significant, comparing the proportion of responders (1/24 nT-CIDP, 0/16 T-CIDP, 2/36 HC, 0/22 ON, p = 0.56) (fig 2B).
IL17 production in cell-culture supernatants was detectable above threshold in most patients and controls, although at low levels (data not shown). There were no significant differences between patients and controls in the concentration of IL17 after stimulation with the sets of peptides. Interestingly, one untreated CIDP patient (CIDP15) displayed the highest concentrations of IL17 in response to both P2 and P0 peptides.
Antibodies to single myelin protein peptides were detected at a low frequency in both patients and controls with no significant differences. There was an increased proportion of IgG antibodies to P280–105 in untreated CIDP compared with HC but not with ON. To other peptides, IgG or IgM responses were detectable only in a few patients and controls without any significant differences between groups. The overall reactivity to all peptides did not differ between groups. Thus, B cell responses to these peptides can be found in only a minority of patients and may not be specific for CIDP.
Antibody reactivity to bovine P2 has been shown previously in small proportions of CIDP patients.12 14 In a recent study, Inglis et al11 reported no significant differences in IgG antibodies to human peptides P214–25 and P261–70, but found a significant increase to P214–25 in GBS at the peak of the disease. This contrasted with our finding in CIDP of an increased frequency of responses to the sequence P280–105. There was no increase in frequency of antibodies to P261–70, the minimum component of P2 capable of inducing EAN in the rat, in either Inglis et al’s study or our own. Neither P214–25 nor P280–105 has been recognised as neuritogenic in animal models. The significance of these responses is therefore not clear. These antibodies may represent an epiphenomenon due to exposure of intracellular nerve antigens, such as P2, during the demyelinating process.
In our study, the proportion of subjects with IgG anti-P0 peptides was not significantly different in patients compared with controls, although individual subjects with CIDP had high titres. In previous studies of CIDP, performed mainly by western blot, there was evidence of antibodies to a P0-like band in 12–28% of patients.4 9 The differences between studies might be explained by the different sensitivity of the methodology. The use of peptides as antigens may have underestimated the frequency of these responses due to the lack of conformational and glycosylated epitopes.
Although we could detect significantly increased numbers of responders to P280–105 and occasional responses of individuals to other peptides in our study, we cannot however conclude that any of these proteins are common targets of a causative humoral response in CIDP.
In our study, we found low IFNγ responses to our myelin protein peptides and no significant differences between patients and controls. However, there was an increased proportion of patients with IL10 reactivity to four P2 peptides representing the first 85 residues of this intracellular protein. There was also an increase in the total number of PBMC secreting IL10 after stimulation with residues 80–132 of P2, although the intensity of the responses was low. Both responses were significantly increased compared with HC but not with ON. It is unclear whether the IL10 responses were directed to specific immunodominant regions of P2 or part of non-specific reactivity secondary to epitope spreading.
Seven out of 10 CIDP patients responding to the residues 1–85 of P2 were tested on a second occasion. The IL10 response persisted above cut-off in only two out of seven, thus indicating that reactivity fluctuates over time. The individual peptides responsible for the reactivity within these residues could be identified in only two subjects, and in both reactivity to P220–45 was present (fig 3B).
A few studies have investigated antigen-specific cellular responses to myelin protein in CIDP. Increased proliferative lymphocyte responses to bovine P2 have been reported in some studies17 18 but not confirmed in others.19 20 In a recent study with synthetic human peptides, Csurhes et al5 reported an increased number of PBMC spontaneously secreting IFNγ as well as IL5 in CIDP. In contrast with our study, they also reported a significantly higher proportion of patients with CIDP with increased numbers of PBMC secreting IFNγ in response to PMP-2251–64, but not IFNγ or IL5, to P2.
Previous studies have mostly shown a Th1 polarisation of the immune response in the blood and cerebrospinal fluid of CIDP patients.23 24 However, in one study, there was an increased proportion of IL4 secreting CD4+ T cells in the blood, suggesting a predominant Th2 response.25 Importantly, the cytokine profile in CIDP might differ according to the disease stage. Inoue et al26 reported an increased number of IL4 secreting cells in the remission stage. Twenty-one of our 40 CIDP samples (18/24 nT-CIDP and 3/16 T-CIDP) were collected in the remission stage. Thus, the IL10 response detected in our series might represent an activation of the Th2 pathway towards myelin proteins during remission. It is noteworthy that our ELISPOT methodology detected IL10 rather than IL4 or IL5, as IL10 is also associated with a tolerogenic/regulatory type of response.27 CIDP is a chronic condition, and these data might indicate the activation of regulatory mechanisms targeting myelin proteins to control the inflammatory process.
Our myelin peptides did not elicit the production of IL17 above cut-off except in occasional patients. Interestingly, one patient (CIDP15) who displayed persistent IL10 reactivity to P21–85 peptides also displayed high concentrations of IL17 to all P2 and P0 peptides. There have been no studies so far on the IL17 responses to myelin proteins in CIDP. IL17-secreting T cells (Th17) are a subset of CD4+ T cells that have been recently recognised as highly pro-inflammatory and responsible for severe autoimmune reactions.28 Mei et al23 reported an increase in the concentration of IL17, along with IL6 and IL8, in the cerebrospinal fluid of untreated CIDP patients compared with other neurological diseases. Recent studies have suggested that IL17, rather than IFNγ, plays a crucial role in the inflammatory responses in multiple sclerosis.29 Although there is still debate about this finding,30 it would be worth investigating more extensively the role of this cell subset in the pathogenesis of CIDP.
In our study, P2 was the most likely candidate target antigen of both antibody and cell-mediated responses. Given the low frequency of these reactivities, the antibody reactivity to one peptide and IL10 response to a pool of different peptides in the same molecule, it is unlikely that P2 or the other myelin proteins tested are the principal targets of a causative immune response in CIDP. The responses which we have found may only be an epiphenomenon, and the search for antibodies and cellular responses in CIDP should now be directed to the large number of other possible Schwann cell and myelin protein or glycolipid antigens.
This work was possible thanks to the financial support of the European Federation of Neurological Societies (fellowship to LS) and J Maas. We thank I Gray for technical support, and our CIDP patients and healthy controls for participating in the study.
Competing interests: None.
Funding: European Federation of Neurological Societies (fellowship to LS).
Ethics approval: Ethics approval was provided by the Guy’s Hospital ethics committee.
Patient consent: Obtained.
▸ Additional tables are published online only at http://jnnp.bmj.com/content/vol80/issue3
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.