Intrathecal synthesis of oligoclonal IgM against myelin lipids predicts an aggressive disease course in MS (original) (raw)
Immunophenotypic characteristics of B cell subpopulations from CSF of MS patients and controls. The percentage of total B lymphocytes (CD19+ cells) present in CSF was analyzed in 29 MS patients with OCMBs (M+ MS patients), 52 MS patients without OCMBs (M– MS patients), and 21 controls (Figure 1A). We found that both M+ (6.33% ± 0.73%, mean ± SE) and M– (3.46% ± 0.30%) groups had increased percentages of B lymphocytes compared with the control group (0.37% ± 0.10%) (P < 0.0001). In addition, differences were also found in B cell percentages between M+ and M– patients (P = 0.0005). Representative dot plots showing the CSF B lymphocytes from an M+ MS patient, an M– MS patient, and a control are shown in Figure 1B.
Study of B cells in CSF of MS patients with and without OCMBs. (A) Both M+ MS and M– MS OCMBs showed a higher percentage of CD19+ cells in CSF when compared with the control group (P < 0.0001). In addition, M+ MS patients show a higher percentage of CD19+ cells when compared with M– MS patients (P = 0.0005). (B) Representative dot plots showing CD19+ cells present in CSF of an M+ MS patient (left), an M– MS patient (center), and a patient from the control group (right). CD19-PerCP-Cy5.5, CD19–peridinin-chlorophyll-protein–Cy-Chrome 5.5.
To study whether the differences in B cell percentages found between M+ and M– patients could be ascribed to a specific B cell subpopulation, we analyzed CSF CD5-positive and -negative B cell subsets (Figure 2). Considerable differences were found in the CD5+ subset between M+ (3.94% ± 0.73%) and M– MS (1.31% ± 0.17%) patients (P < 0.0001). No differences were found in the CD5– subset between M+ (2.41% ± 0.35%) and M– (2.17% ± 0.21%) MS patients, although they were increased when compared with the control group, where practically no B cells were found in CSF. Thus, the CSF CD19+CD5– subpopulation is increased in both M+ and M– patients when compared with the control group, and the CSF CD19+CD5+ subpopulation is mostly increased in M+ patients and is responsible for the increased percentage of B cells observed when MS M+ patients are compared with the M– group.
B cell subsets in MS. The percentage of CD19+CD5+ cells is higher in CSF of M+ MS patients when compared with M– MS patients (P < 0.0001). There are no significant differences in CSF CD19+CD5– cells between the 2 groups of patients.
To study intrathecal B cell activation in MS, we performed additional labeling with anti-CD20 in CSF of 9 M+ and 19 M– patients, since it has been described that disappearance of this marker is associated in mature B lymphocytes with activation and differentiation toward high-rate Ig-secreting cells (26). We found that nearly all CD5– B cells expressed CD20 in both the M+ (99.61% ± 1.17%) and the M– (99.29% ± 0.41%) groups. Conversely, a fraction of CD5+ B cells had lost CD20 expression. This loss of CD20 was higher in the M+ group, where only 31.45% ± 5.60% of CD5+ B cells expressed CD20, than in the M– group, where 70.05% ± 7.18% of CD5+ B cells expressed this marker (P = 0.0009; data not shown).
Study of the specificity of the OCMBs present in a subset of MS patients. CD19+CD5+ cells are the predominant B cell subpopulation present in CSF of MS M+ patients. Since these cells are characterized by the secretion of IgM antibodies that recognize nonproteic antigens, we investigated whether OCMBs were directed against myelin lipids. The presence of lipid-specific OCMBs (LS-OCMBs) was investigated in paired CSF and serum samples from 53 MS patients with OCMBs. Results are shown in Table 1.
Presence of LS-OCMBs in CSF of MS patients
We observed that 46 of the 53 patients (86.8%) had LS-OCMBs. The major antigen was phosphatidylcholine (PC). Oligoclonal IgM recognizing this phospholipid was found in 37 patients. In 30 of them, PC was the only antigen recognized by the oligoclonal IgM. A representative example of this pattern is shown in Figure 3A. In the other 7 patients, an additional reactivity against other myelin lipids was also detected. A representative example (Figure 3B) shows the presence of OCMBs against PC, phosphatidylethanolamine, phosphatidylinositol, and sphingomyelin. Finally, 9 patients showed oligoclonal IgM reacting only against myelin glycolipids, sphingomyelin being the antigen most frequently recognized. The recognition of sphingomyelin by the oligoclonal IgM present in CSF is shown in a representative example (Figure 3C). We could not detect OCMBs against any of the myelin lipids in the 7 remaining patients. A representative example is shown in Figure 3D.
Study of OCMB specificity. The presence of total IgM bands (T) and of oligoclonal IgM reacting against PC, phosphatidylethanolamine (PEA), phosphatidylinositol (PI), phosphatidylserine (PS), sphingomyelin (SP), gangliosides (GA), sulphatides (SU), and membranes coated with Polypep (NC, negative control) were studied in paired serum (S) and CSF (C) samples from 53 MS patients with OCMBs restricted to CSF as determined by IEF and immunodetection. Shown are representative examples of the 4 patterns that were found. (A) OCMBs against PC were detected. (B) OCMBs against PC, PEA, PI, and SP can be observed. (C) OCMBs recognizing SP were found. (D) The OCMBs did not recognize any of the lipids studied.
To investigate the specificity of the binding of OCMBs to the lipids, we inhibited IgM binding to the lipid-coated membranes with relevant and irrelevant antigenic preparations. A representative example (Figure 4) shows that binding of PC-specific IgM to PC-coated membranes was inhibited with this lipid but not with gangliosides, thus confirming the specificity of these antibodies.
Inhibition of IgM binding to lipid-coated membranes. Four aliquots of CSF from a patient presenting OCMB against PC were analyzed by IEF. Proteins were transferred to: an uncoated nitrocellulose membrane (1), a PC-coated nitrocellulose membrane (2), a PC-coated nitrocellulose membrane embedded with a solution of PC that inhibits binding of specific IgM to the antigen-coated membrane (3), and a PC-coated nitrocellulose membrane embedded with a solution of gangliosides that does not produce any inhibition (4).
Next, IgM specificity of M+ patients that underwent lumbar puncture in the first stages of the disease (laboratory-supported MS) or after suffering at least 2 separate attacks (clinically definite MS; see Methods) were analyzed separately (Table 2). It was found that all patients with a clinically definite MS at lumbar puncture had LS-OCMBs. However, only 15 of the 22 patients studied in the first stages of the disease had them.
Distribution of LS-OCMBs in patients classified according to disease duration
Ten of these M+ patients studied in the first stages of the disease had also been included in the study of CSF B cell subpopulations. The percentage of CD19+CD5+ cells was 4.22% ± 1.21% in the 7 patients with LS-OCMBs and 1.35% ± 0.34% in the 3 with OCMBs that did not recognize myelin lipids. Although differences between the 2 groups were not significant (P = 0.1), probably due to the small number of patients studied, the percentage of CD5+ B cells present in patients without LS-OCMBs resembles that of M– patients.
Relationship between the presence of LS-OCMBs and disease course. It has been demonstrated that the presence of OCMBs restricted to CSF is an unfavorable prognostic factor in MS (12, 14). We wished to evaluate whether LS-OCMBs were a more accurate prognostic marker than total OCMBs (regardless of lipid specificity). To study this hypothesis, we analyzed the evolution of 48 MS patients that underwent lumbar puncture after the first attack of the disease. Fifteen of them showed LS-OCMBs in CSF, the other 7 showed CSF OCMBs that did not recognize myelin lipids, and the remaining 26 did not have OCMBs restricted to CSF. Clinical and immunological characteristics of these patients are summarized in Table 3.
Clinical and immunological data from the patients included in the study
The time elapsed between the first and the second relapse has been reported to be a prognostic marker in MS (27). We evaluated it and compared the probability of remaining free of a second relapse in this group of patients divided according to the presence or absence of OCMBs, regardless of lipid specificity (Figure 5A), or LS-OCMBs (Figure 5B). Comparisons were made by means of Kaplan-Meier tests.
Probability of remaining free of a second relapse, which was evaluated by a Kaplan-Meier test in 48 MS patients that had undergone lumbar puncture after the first attack of the disease. First, the disease course in the 22 patients with OCMBs was compared with that of the 26 patients without them (A). Secondly, the disease course in the 15 patients with LS-OCMBs (LS-OCMB+) was compared with that of the group formed by the 26 patients without OCMBs plus the 7 patients with OCMBs that did not recognize myelin lipids (LS-OCMB–) (B).
Patients with OCMBs, regardless of lipid specificity, developed a second relapse earlier than patients lacking them, as we have previously described (12). Five years after disease onset, the probability of remaining free of relapses was 17.9% for patients with OCMBs and 32.5% for patients without OCMBs (P = 0.0004). However, differences increased dramatically when the presence or absence of LS-OCMBs was considered. All patients with LS-OCMBs suffered a second relapse 11 months after disease onset. Conversely, 36.4% of patients lacking LS-OCMBs remained free of relapses after 60 months of follow-up (P < 0.0001). Thus, the presence of IgM reacting against myelin lipids strongly associates with early appearance of a second relapse in MS.
The impact of the presence of OCMBs and LS-OCMBs in the early onset of new relapses was further evaluated by multivariate Cox regression. The presence of total OCMBs (regardless of lipid specificity) resulted in an increased risk of suffering a second relapse (hazard ratio, 3.54; confidence interval, 1.65–7.55). However, the risk was considerably higher in patients with LS-OCMBs (hazard ratio, 11.41; confidence interval, 4.26–30.58). These results confirm that LS-OCMBs are more accurate than total OCMBs in predicting the early onset of a second relapse in MS.
The relationship of the presence of LS-OCMBs with the clinical course of the disease was further explored by monitoring the number of relapses and the progression of disability during follow-up. Although IFN-β treatment is known to be effective in decreasing the number of relapses in MS patients (28–30) and could introduce bias in our study, for ethical reasons treatment was offered to all patients who had suffered at least 2 relapses 6 months after disease onset. The probability of remaining without treatment during follow-up was evaluated by means of a Kaplan-Meier test in patients with and without LS-OCMBs. Results are shown in Figure 6. The probability of remaining without treatment was 0% after 24 months of follow-up for patients with LS-OCMBs. Conversely, at the end of the study, it was still 55.7% for patients without LS-OCMBs (P < 0.0001).
Probability of remaining free of IFN-β treatment during follow-up. It was lower in the 15 patients with CSF-restricted LS-OCMBs (LS-OCMB+) than in the 33 patients without LS-OCMBs (LS-OCMB–) (P < 0.0001).
The number of relapses during follow-up was evaluated in both groups. Although most patients with LS-OCMBs began IFN-β treatment earlier than those without LS-OCMBs, the former suffered more relapses 2.7 (2.7 ± 0.7) than the latter (0.8 ± 0.2) (P = 0.0004). Results are shown Figure 7A.
Evaluation of disease course. The number of relapses during follow-up (A) and the neurological disability at the end of the study (B) were monitored in patients with and without LS-OCMBs. Comparisons were made by a Mann-Whitney U test. Patients with LS-OCMBs had more relapses (P = 0.0004) and became more disabled (P = 0.02) than those without LS-OCMBs.
The neurological disability experienced by the patients during the study was also evaluated. Although it has been reported that increase in the expanded standard disability status scale (EDSS; see Methods) score is generally moderate in the first few following the disease onset, after follow-up, the score was 1.68 ± 0.36 for patients with LS-OCMBs and 0.85 ± 0.16 for those lacking LS-OCMBs (P = 0.02) (Figure 7B).
As we have demonstrated, the presence of LS-OCMBs not only associates with a higher number of relapses but also with greater disability in MS, even at the first stages of the disease.