Clinical and neuroimaging correlates of antiphospholipid antibodies in multiple sclerosis: a preliminary study

Multiple sclerosis (MS) is an immune-mediated neurodegenerative disorder of human central nervous system, which is initially characterized by loss of myelin/oligodendrocyte complex followed by progressive neuronal loss and axonal degeneration [1‐3]. Clinically, the majority of MS patients present with a relapsing-remitting course and within a few years, a large number of these patients with or without treatment with immunomodulatory agents enter another phase of disease known as secondary progressive MS. Neuropathologically, MS lesions within MS are characterized by perivenular infiltration of myelin basic protein-activated CD4 T lymphocytes as well as reactive macrophages which orchestrate the massive inflammatory cascade within the CNS [2]. Another arm of the immune system, the humoral immune system-autoantibodies as well as activated complement system-also play a significant role in the pathogenesis of MS [2, 3]. The abnormal activation of both cellular and humoral immune arms combined with disruption of the blood brain barrier (BBB), activation of the cerebral endothelial cells, and loss of adjacent tight and adherent endothelial junctions [4‐6] precede formation of perivenular demyelinating lesions

The first antiphospholipid antibody (APLA) identified was anti-cardiolipin (aCL) in 1941, seen in false-positive syphilis tests. The lupus anticoagulant (LAC), which is believed to be a manifestation of APLA, was originally associated with a hemorrhagic diathesis [7, 8] but in the 1980s, a stronger association with thrombosis was found, first called aCL syndrome or Hughes syndrome, now known as antiphospholipid syndrome (APS) [9‐12]. A major advance was the realization that nearly all APLA are in fact directed not against phospholipids (PL) per se, but against PL-binding proteins [13]. The first such cofactor identified was β2GPI, said to be associated with thrombosis, but many others were subsequently identified, now numbering in the dozens. This discovery has greatly broadened the definition of APLA, and makes clear that APLA detected by standard ELISA methods are in reality very heterogeneous [14, 15].

High frequencies of APLA are seen in autoimmune disorders other than systemic lupus erythematosus (SLE), not necessarily associated with thrombosis, such as in the bleeding disorder, immune thrombocytopenic purpura (ITP) [16] and in MS.

The neuropsychiatric manifestations of APLA (with or without APS) are well known [17, 18] and in some instances resemble those of MS [19‐23]. The reported frequencies of positive APLA in MS have ranged from 10% or less [24‐26] to 44% [27] and to 88% [20]. Such wide discrepancies are common in the APLA literature and are attributable to variations in methodological details (which are inadequately specified in some of the above cited reports) as well as criteria of patient selection such as distinguishing clinical state. To our knowledge, no report to date has established a clear association between APLA and the clinical state or radiologic imaging data in MS patients. The present study was motivated to clarify these uncertainties using the same standardized methods we have applied in several other studies [16, 28, 29] and a well-defined patient population.

Demographic features of MS subjects are presented in Table 1. The study subjects with RRMS consisted of 20 females and 4 males with a mean duration of neurological symptoms prior to diagnosis of 5.3 months. None of the MS subjects had ever experienced any thromboembolic or thrombo-occlusive syndromes.

The principal finding of this study was a much higher frequency of APLA positive test results for MS patients in exacerbation compared to remission. Figure 1 shows graphically the results for subjects with IgM antibodies against the pure phospholipids (PL). It is observed that the increased frequency of seropositivity ranged from 2-fold to 4-fold in exacerbation compared to remission. This was statistically significant in every case at p < 0.004. Interestingly, no positive IgG was encountered.

It is apparent that if all patients were combined rather than separated by exacerbation vs. remission, the frequency of positive tests would be reduced. For example, on the basis of all 24 patients, the prevalence of aPS was 54% and that of aFVII was 41.6% (not shown).

Figure 2 shows positive results for the two protein antigens, β2GPI and FVII/VIIa. Both of these autoantibodies were significantly and more frequently present in exacerbation compared to remission, p = 0.004. Remarkably, none of the patients in remission exhibited seropositivity for anti-FVII but 10 of 17 (59%) in exacerbation were positive.

A summary of frequency of seropositivity of APLA in MS patients and controls is shown in Table 2. Plasma of 19 of our MS subjects tested positive in at least one of the 12 tests (6 antigens, IgG and IgM). All these antibodies were of IgM family. No positive IgG was detected. The right-most bar of Figure 2 shows frequency of at least one positive of the 12 tests between exacerbation and remission. This difference was also significant (p < 0.01).

Next, we analyzed possible correlations between the presence of T1-weighted post contrast enhancing lesions on brain MRI and the presence of various sub-classes of APLA. Results are shown on Table 3. A significant correlation between contrast-enhancing lesions and 3 of the 6 target antigens (FVII, PC, PS) was found (p < 0.0002, 0.0065, 0.0014, respectively). There was a trend toward correlation for anti-PE and anti-CL (p = 0.056 for both). Interestingly, there was no trend of correlation for anti-β2GPI.

The central finding of this paper is a close association between APLA and certain clinical and neuroradiologic features of MS. We surmise that previous studies failed to detect this association because they did not distinguish clinical states of the patients, or used methods unable to make this discrimination.

The frequency of MS plasma positive for APLA in this study was higher than in most but not all previous reports, since as noted in the introduction, Ijdo and colleagues found 88% positivity [20]. The number of tests performed on each specimen in our study was larger than most and would help to increase the likelihood of finding at least one positive test in a given subject. Furthermore, few if any prior studies clearly separated exacerbation from remission, resulting in a reduced apparent prevalence, since we clearly show a much greater frequency of positivity in exacerbation.

Our finding of IgM exclusively (no IgG) was also unexpected but is not entirely inconsistent with prior reports since most find a preponderance of IgM over IgG. For example, Sugiyama et al found 14 of 32 positive for IgM but only 3 of 32 positive for IgG [27]. It is unlikely that our finding of exclusively IgM reflects a systematic methodological error since other studies in progress at the same time (such as of ITP) revealed nearly equal numbers of IgG and IgM seropositivity. Additionally, another recent study on clinical and neuroimaging correlates of autoreactive antibodies in MS patients revealed similar finding [33]. Through a retrospective study, the investigators determined that APLA mainly of IgM type was present in 55% of their study subjects with MS. It remains unknown now why, in our MS subjects, exclusively IgM APLA was found; however, we hypothesize that since the subjects in this study were all newly diagnosed they had not yet class-switched to IgG. Alternatively, one may suggest that a class-switching occurs in a small fraction of APLA+ MS patients, but majority of these APLA+ MS patients are defective in class-switching [34].

The clinical significance of the findings of this study is unknown; however, there are two broad possibilities: first, that APLA in MS may contribute causally to the disease pathogenesis, or that APLA are secondary (epiphenomenal) with no role in promotion of the inflammatory cascade of MS. Regarding the first concept, it is important to point out that our assay of aPC, aPS, aPE and aCL does not identify the specific plasma antigen responsible for a positive test. As remarked in the introduction, many APLA antigens are now known, and certainly many more exist, therefore it is possible that APLA in MS are directed against some specific antigen that could be involved, for example, in compromise of the BBB. A number of available reports document cross-reaction of some APLA with endothelial cells and platelets [35‐39], raising the possibility that the rise of specific types of APLA in the MS subjects could trigger exacerbations by further compromise of the BBB. We have shown complex effects of plasma from MS patients on brain microvascular endothelial cells in tissue culture [40, 41].

It should be stressed that our assay against the four pure phospholipids will detect any PL-binding protein, not necessarily those thus far identified as APLA. It is therefore possible that we are detecting an antigen specifically involved in compromise of the BBB, possibly responsible for the endothelial activation that we reported in exacerbations of MS [42].

It is of interest that only anti-β2GPI, which is believed to be a risk factor for thrombosis, was not associated with MS radiologic imaging, while anti-FVII was most closely associated (Table 2). The pathophysiological significance of these findings remain unclear but it may be relevant to note that FVIIa is closely associated with tissue factor (TF), FXa, and TF pathway inhibitor (TFPI) in normal hemostasis, and that TF has numerous actions apart from thrombosis and hemostasis [43].

In light of these findings, further investigation of APLA in a larger cohort of MS subjects is warranted to clarify the significance of these autoantibodies in MS. Further understanding the role of these autoantibodies in MS subjects may eventually translate into more effective treatments.

The findings of this preliminary study demonstrate presence of IgM APLA during exacerbations of MS. Currently, the significance of these autoantibodies in pathogenesis of MS remains unknown. Longitudinal studies to measure both IgM and IgG classes of APLA in larger cohorts of patients with relapsing-remitting MS are necessary to establish the significance of these autoantibodies and assess their pattern of expression before and after treatment with disease modifying agents.