Introduction
Susac syndrome (SuS) is a rare disorder that is thought to be caused by autoimmune-mediated occlusions of microvessels in the brain, retina and inner ear [
1] which lead to a characteristic clinical triad of central nervous system (CNS) dysfunction, visual disturbances and hearing deficits [
1‐
5]. Typical findings in patients with SuS include branch retinal artery occlusions (BRAO) detectable on retinal fluorescein angiography, characteristic callosal lesions on cranial MRI, and evidence of sensorineural hearing loss [
1,
6]. The three index events defining SuS may occur simultaneously or, more often, successively. The disease may be monophasic or may follow a relapsing or a chronic-progressive course.
The exact prevalence of SuS is unknown but is considered to be low; since the syndrome was first described in 1973 only about 300 patients have been reported worldwide [
1]. Accordingly, most of our current knowledge on SuS is based on case reports or small case series, only very few of which have included more than four patients [
1].
Although spontaneous recovery and long-term remission have been described, many patients respond to immunosuppressive agents, suggesting a possible autoimmune pathogenesis [
1,
5,
7]. A role for anti-endothelial cell antibodies (AECA) in the pathogenesis of SuS was proposed by Susac and co-workers in two review articles in 2007 with reference to unpublished data [
7,
8]. However, no original studies had been published by then to substantiate this claim. When examining the only serum samples from a patient with SuS regularly seen at our center at that time, we indeed found evidence of an anti-endothelial humoral immune response [
9], in keeping with those anecdotal reports. However, it remained unknown whether AECA are a common phenomenon in SuS or whether that was an isolated finding. A more recent study seems to confirm the presence of AECA in SuS, but its results are challenged by a number of methodological issues [
10].
Here, we provide a summary of the most important clinical and paraclinical findings associated with SuS from one of the largest single cohorts of patients with SuS studied to date. In addition, we present data on the frequency, titers and clinical relevance of AECA in SuS.
Patients and methods
Twenty-five patients with SuS and 70 controls, comprising 25 patients with multiple sclerosis (MS) (relapsing remitting MS in 18, secondary progressive MS in 5, and primary progressive MS in 2), 25 patients with connective tissue disorders (CTD) with CNS involvement (systemic lupus erythematosus (SLE) in 17, Sjögren syndrome (SS) in 1, SLE and secondary SS in 1, Sharp syndrome in 5, and scleroderma in 1) and 20 healthy controls (HC), were analyzed for serum AECAs using a tissue-based indirect immunofluorescence assay (IFA) [
11]. Patients with SuS were further stratified according to clinical, fluorescence angiographic and MRI findings: While all patients in group I (n = 20) had encephalopathy, hearing loss and visual disturbances (‘complete SuS’), group II consisted of patients (n = 5) in whom only two of the three typical sites (brain, retina, inner ear) were affected but in whom a diagnosis of SuS was likely due to the presence of BRAO (‘limited SuS’). In addition, 12 follow-up samples from 7 patients with definite SuS were available, adding up to a total number of 107 samples tested, including 37 from patients with SuS.
All patients and controls were of Caucasian origin (countries of origin are given in Table
1). The sex ratio (m:f) was 1:2.6 in the SuS group (1:3 in the definite SuS subgroup) and 1:3.4 in the control group (
P = n.s.). The median age at blood sampling was 32 years in the SuS group and 35 years in the control group (
P = n.s.) (Table
1).
Table 1
Epidemiological data from the total cohort and disease and control subgroups
All | 90 | 90 | 33 (16 to 79) | | 1:3.1 | |
SuS
| 25 | 37 | 32 (19 to 62) |
P = n.s.b,d
| 1:2.6 |
P = n.s.c,d
|
Definitive SuS | 20 | 32 | 32 (20 to 62) | | 1:3 | |
Limited SuS | 5 | 5 | 30 (19 to 44) | | 1:1.5 | |
Controls
| 70 | 70 | 35 (16 to 79) | | 1:3.4 | |
HC | 20 | 20 | 28 (24 to 50) | | 1:3 | |
MS | 25 | 25 | 35 (21 to 67) | | 1:2.6 | |
CTD | 25 | 25 | 44 (16 to 79) | | 1:5.3 | |
All serum samples were stored at −80°C until testing. The median time between SuS onset and blood sampling was 5 years (range 0 to 18), and the median time between the most recent SuS attack and blood sampling was 39.5 months (range 0 to 165). While 20 samples were taken from untreated patients, 15 were taken during periods of active immunomodulatory or immunosuppressive treatment; in 2 cases the exact treatment status at the time of blood sampling was unknown. A total of 35.1% (13/37) of samples were obtained during phases of active disease. All patients with SuS had been treated at some point in time, most commonly with platelet aggregation inhibitors or anticoagulants (acetylsalicylic acid in 17, clopidogrel in 1, dipyridamole in 1, and fondaparinux in 1). Five patients were treated with nimodipine and 1 with losartan. Immunomodulatory (IM) and immunosuppressive (IS) treatments used at some point in time included steroids in 18 patients, cyclophosphamide in 10, intravenous immunoglobulins in 9, mycophenolate mofetil in 6, azathioprine in 4, and methotrexate in 2; overall, 23/25 patients (92%) had received IM and/or IS treatments at least once.
Antibodies to CNS tissue were analyzed by indirect immunofluorescence on adult primate (
Macaca mulatta), rat, and mouse cerebellum cryosections (Euroimmun, Luebeck, Germany) as described [
9,
11,
12]. Briefly, unfixed or fixed cryosections (10% formalin for 4 min and 3-[(3-cholamidopropyl) dimethylammonio]-1-propanesulfonate (CHAPS) 1% in PBS for another 4 min), respectively, were blocked with 10% goat serum and subsequently incubated with patient serum for 1 h. Human IgG, IgM and IgA binding to CNS tissue were detected by the use of prediluted polyclonal goat anti-human IgG, anti-human IgM, or anti-human IgA, respectively, conjugated to fluorescein isothiocyanate (FITC) (Euroimmun). Sections were mounted using ProLongGold mounting medium (Invitrogen, Darmstadt, Germany) containing 4′,6-diamidino-2-phenylindole (DAPI; 1:1000). Each incubation step was followed by three washes in PBS. For evaluation of IgG subclasses, serum samples were tested by immunohistochemistry (IHC) on mouse cerebellum sections as described above, with the following modifications: unconjugated sheep anti-human IgG antibodies specific for IgG subclasses (Binding Site, Germany) were substituted for the FITC-labeled goat anti-human IgG antibody, and Alexa Fluor (AF) 568-labelled donkey anti-sheep IgG (Invitrogen; absorbed against primate IgG) was used to detect the subclass-specific antibodies. In addition, all AECA-positive samples were tested for neuromyelitis optica (NMO)-IgG/aquaporin4 (AQP4) antibodies, the staining pattern of which shows similarities with that of AECA when tested by IHC, using a highly sensitive and specific commercial cell-based assay employing recombinant human AQP4 as previously described [
13‐
15]. In addition, double staining of monkey cerebellum sections with a commercial antibody to AQP4, detected using AF568-labelled anti-rabbit IgG, and sera of AECA positive patients, detected using FITC-labeled, anti-human IgG was performed as previously described [
12].
All data were analyzed in an anonymized fashion as required by the institutional review board of the University of Heidelberg. The Mann–Whitney U-test (2-tailed) was used to test for significant differences between continuous variables, and Fisher’s exact test (2-tailed) to compare proportions. All tests should be understood as constituting exploratory data analysis, such that no adjustments for multiple testing have been made. Microsoft Excel 2003 and GraphPad Prism 4 were used for statistical analyses.
Discussion
This is one of the largest series of patients with SuS so far. It is a potential advantage that our data are derived from a single series with uniform inclusion criteria. By contrast, almost all of our previous knowledge on SuS was based on reports of single cases and small case series, only a few of which included more than four patients. The completeness of data provided in those reports, some of which were published only in abstract form, varied widely, as did the inclusion criteria for patients with limited SuS. We therefore applied stricter criteria for the inclusion of patients with abortive SuS, all of whom had to present with BRAO, and aimed at collecting a more comprehensive and complete set of data than in some of the previous reports.
Our study reveals a broad spectrum of neurological signs and symptoms associated with SuS, including, in addition to headache, motor, sensory, brainstem and cerebellar symptoms, aphasia, and cognitive and memory decline. In addition, all patients with SuS had periventricular and/or callosal brain lesions on MRI. These findings, the mostly relapsing course, and the occasional presence of CSF-restricted OCB underlines the relevance of SuS as a rare differential diagnosis of MS, especially if disease starts with isolated encephalopathic symptoms, which was the case in 60% of our patients. The 7-T MRI and optical coherence tomography have recently been shown to be of help in the differential diagnosis of MS and SuS [
6,
18].
The findings that visual and hearing impairments were bilateral in 92% and 96% of cases, respectively, that 64% of the patients had neurological symptoms at last follow-up and, importantly, that more than half of them exhibited brain atrophy, despite both immunosuppressive and antithrombotic treatments having been tried in most cases, characterize SuS as a severe disease.
Importantly, from both the pathogenetic and the diagnostic point of view, around 30% of patients with definite SuS were positive for serum AECA. The fact that AECA were present at high titers (up to 1:17,500) in some of these patients and were found to belong to the complement-activating IgG1 and Ig3 subclass in all of them, suggests that AECA might possibly play a role in a subset of patients with SuS. IgM antibodies, which are considered even stronger complement activators, were present in around half of the cases. Of note, AECA were present in all five samples obtained from a single patient with SuS taken over a period of 29 months, indicating that AECA seropositivity was not an accidental or transitory finding.
A possible role for AECA in some patients with SuS is further supported by the finding that median AECA titers were significantly higher in the SuS group than in the control group. AECA titers of >1:320 were exclusively found in patients with SuS; by contrast, they were demonstrated in none of the 70 controls, including 25 patients with CTD. Of special note, frequency and titers of AECA were higher in the SuS group than in the control group despite a higher rate of immunosuppressive treatment at the time of blood sampling in the SuS group (neither the HC nor the MS patients had ever received long-term IS treatment).
As AECA seronegativity was not strictly associated with IS/IM treatment or remission, the lack of AECA seropositivity in some patients could indicate that
1.
SuS is an etiologically heterogeneous syndrome, which is not caused by AECA in all cases; and/or
2.
AECA are an optional secondary phenomenon following endothelial damage of other origin (for example, T cell mediated) not occurring in all patients.
In fact, a broad range of conditions are capable of causing microinfarctions and/or BRAO, and etiological heterogeneity has been demonstrated over the last few years in a number of autoimmune CNS syndromes, such as limbic encephalitis, subacute cerebellar degeneration, neuromyelitis optica (NMO) and myasthenia gravis [
19‐
22]. Importantly, AECA were absent also in a number of patients with active disease, demonstrating that the presence of AECA, at least in some cases, is not a
conditio sine qua non for SuS flares.
Interestingly, almost 80% of all patients with available data showed an elevated CSF/serum albumin ratio, which in the absence of substantial CSF flow alterations (as seen, for example, in spinal canal stenosis) is thought to mainly reflect blood–brain barrier dysfunction. This is well in line with results from histopathological and ophthalmoscopic (fluorescein dye leakage, arterial wall plaques) studies showing vascular damage in SuS.
The absence of CSF-restricted OCB in most patients does not
per se argue against SuS being an autoimmune CNS disorder and against AECA being involved in the pathogenesis of CNS damage, since the antibodies’ antigen might well be on the luminal side of the endothelium and thus directly accessible to AECA in the peripheral blood. Similarly, autoantibodies to AQP4 in NMO are mainly produced in the periphery [
23‐
25]. OCBs were found to be frequently missing also in other CNS disorders of putative autoimmune etiology [
26‐
28].
Since publication of our index case [
9], a single study on AECA in SuS has been published, the authors of which strongly advocated a role of such antibodies in the pathogenesis of the syndrome [
10]. However, it is unknown whether the antibodies detected in that study were indeed AECA, whether they bound to brain endothelial antigens, and whether they were specific for SuS, since only endothelial cells (EC) but no control cells (as naturally included in the tissue sections used in the present study) were used in some of the experiments, cutaneous EC instead of brain EC were employed, no control sera were used in the IHC experiments, statistical significance levels were provided only for a single assay, and the histopathological data demonstrating endothelial pathology were derived from the only of all patients that had not been tested for AECA. Moreover, patient numbers were relatively low. Finally, only 5 of the 11 patients tested for AECA had the typical triad of SuS, and MRI data, which could have supported the diagnosis especially in the ‘limited SuS’ cases, were not supplied.
Unless the exact role of AECA in the pathogenesis of SuS has been better defined, we believe that the detection of AECA alone, especially if present only at low titer, does per se not justify B cell or antibody targeted treatments. However, an effect of IS/IM treatment at least in a subset of patients with SuS is suggested by a number of anecdotal reports and small retrospective case series. It is recommended that IS/IM treatment of patients with SuS should be performed in the context of controlled clinical trials or, at least, treatment registers in the future. Such trials or registers should include standardized serum testing for AECA as a prerequisite for investigating the role of AECA as well as the effect of IS/IM treatment on AECA serostatus and titers in SuS in a more definite way.
For routine autoantibody testing, many clinical laboratories use mouse or rat tissue instead of primate tissue. While rodent tissue sections are more easily available, the use of non-primate tissue may be problematic due to interspecies differences in antigen structure. To evaluate whether primate tissue can be replaced by rodent tissue, all SuS samples seropositive on primate tissue were tested in addition on mouse and rat cerebellum tissue sections. Based on that direct comparison, we recommend that future studies on AECA in SuS using IIF should use primate tissue, which yielded higher sensitivity, titers and specificity than rodent tissues in the present study.
It is a potential limitation of our study that some patients were treated with IS/IM drugs at the time of blood sampling and that some samples were taken during periods of clinical remission. However, as mentioned above, AECA were negative also in a relevant number of samples taken from untreated patients or taken during active disease. Moreover, several of our AECA-positive patients had very high titers also during remission (similarly, autoantibodies remain detectable during remission in many neurological and non-neurological autoimmune disorders such as AQP4-Ab-positive NMO [
15,
29], myasthenia gravis [
30] or SLE [
31]. These findings suggest that treatment and disease activity are not the only factor determining AECA serostatus; instead, AECA may be truly absent in some cases of SuS as discussed above.
In summary, our study provides systematic clinical and paraclinical information derived from a single large series of patients with SuS. In addition, we demonstrate that complement-activating IgG1 and IgM AECA are present in a subset of patients with definite SuS, in some cases at high level. Future studies are now warranted to evaluate the exact pathogenic impact of AECA in SuS and to identify their antigenic target.
Competing interests
KF and WS are employees of Euroimmun AG, Lübeck. Euroimmun kindy provided the tissue sections used in this study. The other authors have no competing interests.
Authors’ contributions
SJ, IK, JD, FP and BW conceived the study. SJ designed the study, performed the experiments, analyzed the data, and wrote the manuscript. SJ, IK, JD, JS-G, ZI, EE, CC, MR, OA, XM, EBR, FP and BW were involved in patient care and retrospective data analysis. KF and WS were involved in antibody testing. All authors were involved in critically revising the manuscript for important intellectual content. All authors read and approved the final manuscript.