Corticosteroid therapy in a patient with cerebral amyloid angiopathy-related inflammation
© Kimura et al.; licensee BioMed Central Ltd. 2013
Received: 26 September 2012
Accepted: 27 February 2013
Published: 16 March 2013
Skip to main content
© Kimura et al.; licensee BioMed Central Ltd. 2013
Received: 26 September 2012
Accepted: 27 February 2013
Published: 16 March 2013
We studied longitudinal changes of the levels of anti-amyloid β (anti-Aβ) antibody, amyloid β (Aβ) protein, and interleukin 8 (IL-8) in cerebrospinal fluid (CSF) of a patient with cerebral amyloid angiopathy-related inflammation (CAA-ri) in whom steroid treatment resulted in clinical improvement. The diagnosis of CAA-ri was established with brain biopsy. Levels of anti-Aβ 42 antibody, Aβ 40, Aβ 42 and IL-8 in CSF were measured in the CAA-ri patient at 23 time points in the 8-month clinical course. These CSF samples were divided into 2 groups: those obtained before (n = 12) and those after (n = 11) oral corticosteroid therapy was started. We compared these levels between CSF samples obtained before and after therapy. The mean levels of anti-Aβ 42 antibody and IL-8 were significantly higher in CSF samples of the CAA-ri patient before oral corticosteroid therapy than those after therapy. A positive correlation was noted between levels of anti-Aβ 42 antibodies and IL-8 in CSF of this patient. There were no significant differences of mean levels of Aβ 40 and Aβ 42 between CSF samples obtained before and after oral corticosteroid therapy. It was possible that the autoinflammatory process with anti-Aβ 42 antibodies and IL-8 may have been involved in the pathogenesis of CAA-ri, and that corticosteroid therapy directly affected levels of anti-Aβ 42 antibody and IL-8. In summary, CAA-ri encephalopathy is a relapsing or progressive disorder and may be treatable by adequate immunosuppressive therapy. The anti-Aβ 42 antibody in CSF is a useful biological marker for therapeutic monitoring of CAA-ri.
Cerebral amyloid angiopathy-related inflammation (CAA-ri) is characterized by sub-acute confusion, progressive cognitive decline, seizure or headaches; reversible focal subcortical and/or cortical T2 hyperintensities on magnetic resonance imaging (MRI); and neuropathological evidence of cerebral amyloid angiopathy (CAA) and associated vascular or perivascular inflammation [1–3]. Although the apolipoprotein E ε4/ε4 genotype is strongly associated with CAA-ri , the pathogenesis of CAA-ri is unknown. Clinical reports have noted a response to immunosuppressive treatment, suggesting that this syndrome may be a treatable form of CAA [4, 5]. The syndrome may be diagnosed non-invasively, based on a characteristic combination of clinical and radiographic features , but the only way to confirm the diagnosis is through brain biopsy. Two previous studies have demonstrated the presence of anti-amyloid β (anti-Aβ) antibodies associated with CAA-ri [7, 8]. In the present study, we longitudinally analyzed anti-Aβ antibody, amyloid β (Aβ) protein and interleukin 8 (IL-8) in cerebrospinal fluid (CSF) samples of a patient who had the diagnosis of CAA-ri confirmed with brain biopsy samples.
On admission to our hospital the patient’s consciousness deteriorated further and she required mechanical respiratory treatment. She was treated with intravenous antibiotics, antifungal drugs and acyclovir with intravenous methylprednisolone (1,000 mg/day for 3 days without subsequent oral corticosteroids). Her symptoms transiently improved but her consciousness progressively deteriorated again, and she required further mechanical respiratory treatment. At 49 days after admission, a repeat MRI scan with T2-weighted and FLAIR images showed more extensive hyperintensity lesions in the leptomeninges and sulci, asymmetric subcortical white matter of the occipital lobes, and hyperintensity lesions that extended through the white matter (Figure 1).
We measured the levels of anti-Aβ 42 antibody, Aβ 40 and Aβ 42 in CSF at 23 time points in the patient with CAA-ri during the 8-month clinical course by Enzyme-Linked Immunosorbent Assay (ELISA). These CSF samples were divided into 2 groups: those obtained before (0 to 103 days after admission; n = 12) and those after (≥116 days after admission; n = 11) oral corticosteroid therapy was started.
Demographic and clinical features of control subjects*
Gender, female / male
7 / 7
5 / 2
Age at evaluation (y)
73 ± 9
42 ± 9
Age at onset (y)
72 ± 9
30 ± 10
Disease duration (y)
1.9 ± 0.7
12 ± 8
Prednisolone, 5 mg/day (1 patient); interferon β1b, 8 MIU/every other day (2 patients); mizoribine, 50 mg/day (1 patient)
We measured the levels of anti-Aβ 42 antibody, Aβ 40 and Aβ 42 in 2 serum samples obtained before oral corticosteroid therapy in the patient with CAA-ri by ELISA. We also measured these levels in serum samples of 7 patients with multiple sclerosis and 10 of 14 patients with Alzheimer’s disease. The CSF and serum samples of all subjects were collected and stored at −30°C. Commercially available kits were used for measuring levels of anti-Aβ 42 antibody (Human anti-Aβ42 ELISA kit, DRG International, Mountainside, NJ, USA), Aβ 40 and Aβ 42 (Human Aβ 40 and Aβ 42 brain ELISA kits, Millipore, Billerica, MA, USA). The patient with CAA-ri had polynuclear pleocytosis. Therefore, we also measured the levels of the neutrophil chemotactic factor IL-8 in CSF, which is a major mediator of the inflammatory response, in the CAA-ri patient and control subjects. The IL-8 levels were measured with a commercially available kit (Human IL-8 ELISA kit, eBioscience, San Diego, CA, USA). These commercial kits used solid phase ELISA that was based on the sandwich principle. The level of anti-Aβ 42 antibody was reported as units, which were defined commercially from a calibration curve as an indicator of concentration (not as units in enzymatic assays that describe turnover rate of substrate under defined conditions and time). The levels of Aβ40, Aβ42 and IL-8 were reported as concentrations (pg/mL). This study was approved by the institutional review board of Gifu University Graduate School of Medicine, Gifu City, Japan.
The Mann–Whitney test was used to compare the mean levels of anti-Aβ 42 antibody, Aβ 40, Aβ 42 and IL-8 between CSF samples obtained before and after oral corticosteroid therapy was started. The Spearman rank correlation was used to assess the correlation between the anti-Aβ 42 antibody level, IL-8 level, total cell count and total protein concentration in the CSF samples of the patient with CAA-ri. Statistical significance was defined by P ≤0.05.
Levels of anti-Aβ 42 antibody, Aβ 40, Aβ 42 and IL-8 in CSF or serum of the patient with CAA-ri and control subjects*
Anti-Aβ 42 Ab level (U)
Aβ 40 (pg/mL)
Aβ 42 (pg/mL)
CAA-ri (1 patient)
All values CSF
1.5 ± 0.9
1030 ± 640
72 ± 8
340 ± 330
38 ± 24
1490 ± 327
Before oral corticosteroids CSF
1.9 ± 0.9†
1130 ± 850
78 ± 101
570 ± 290‡
38 ± 24
1490 ± 327
After oral corticosteroids CSF
1.4 ± 0.4†
920 ± 290
65 ± 69
80 ± 52‡
Alzheimer’s disease (14 patients) CSF
0.8 ± 0.8
4990 ± 2080
540 ± 160
12 ± 46
29 ± 19
1330 ± 1040
52.7 ± 120
Multiple sclerosis (7 patients) CSF
1.2 ± 0.9
4180 ± 2080
770 ± 450
0 ± 0
54 ± 22
683 ± 785
The mean Aβ 40 and Aβ 42 levels of the CSF samples obtained before and after oral corticosteroid therapy, and all values combined, were lower in the CAA-ri patient than control subjects. The mean IL-8 level of the CSF samples obtained before and after oral corticosteroid therapy, and all values combined in the CAA-ri patient, were greater than those in control subjects.
This study showed that the mean level of anti-Aβ 42 antibody in the patient with CAA-ri was significantly greater in CSF samples obtained before starting oral corticosteroid therapy than those obtained after. After oral corticosteroid therapy was started in the present CAA-ri patient, the level of anti-Aβ 42 antibody significantly decreased and the clinical symptoms improved. This supports the concept that anti-Aβ 42 antibodies in CSF may mediate the autoimmune reaction occurring in CAA-ri, and that corticosteroids had a therapeutic effect on elevated anti-Aβ 42 antibody in the CSF of the patient with CAA-ri.
The mean level of anti-Aβ 42 antibody of CSF samples obtained before oral corticosteroid therapy was greater in the CAA-ri patient than control subjects. Using CSF/serum quotient diagrams (Reibergrams), we evaluated whether this increased level of anti-Aβ 42 antibody of CSF samples obtained before oral corticosteroid therapy in the CAA-ri patient was a result of blood–brain barrier damage or specific intrathecal synthesis . The increased level of anti-Aβ 42 antibody was not explained by blood–brain barrier damage, but was caused by specific intrathecal synthesis of anti-Aβ antibodies (quotient-anti-Aβ 42 antibody: 54.2 × 10-3; quotient-total IgG: 16.9 × 10-3). This confirms the findings of recent reports [7, 8]. However, the present study is limited because there was only 1 patient with CAA-ri. Future studies including a larger number of CAA-ri patients and control subjects are expected.
The radiographic abnormalities of the present patient with CAA-ri progressed and did not improve despite corticosteroid treatment. The beneficial effects of corticosteroid treatment may vary with the pathological subtype of CAA-ri. There have been 2 subtypes of CAA-ri described, including a non-vasculitic form (perivascular infiltration) and a vasculitic form (transmural granulomatous angiitis) . These pathological forms can occur together in the same patient . Neuropathological findings of the present study suggested that our patient had concurrent perivascular infiltration and transmural granulomatous angiitis. In general, patients with transmural granulomatous angiitis may not benefit as much as patients with perivascular infiltration by immunosuppressive therapy. A previous long-term follow-up study showed that CAA-ri can be a relapsing or progressive disorder . The present CAA-ri patient was treated twice with intravenous methylprednisolone pulse therapy. The CSF level of anti-Aβ 42 antibody was immediately and strongly reduced after intravenous methylprednisolone pulse therapy. However, this antibody level increased within a few weeks after the first intravenous methylprednisolone pulse therapy that was not followed with oral corticosteroids. The present study showed that this antibody level was progressively increasing, according to the reduction of oral corticosteroid. If the CSF level of anti-Aβ 42 antibody correlates with the ongoing autoimmune process, the patient’s history suggests that the inflammatory process resumed after the first methylprednisolone pulse, and it may be advisable to place some CAA-ri patients on immunosuppressive treatment. Repeat measurement of this antibody level in CSF may help monitor the course of the disease and efficacy of treatment.
The mean IL-8 level in the present CAA-ri patient was also significantly greater in CSF samples obtained before than those after oral corticosteroid therapy (Table 2), and were greater in the CAA-ri patient than in control subjects (Table 2). Moreover, a positive correlation was noted between levels of anti-Aβ 42 antibodies and IL-8 in CSF of the CAA-ri patient (Figure 7). Small heat shock proteins are associated with Aβ deposits in CAA and induce a cerebral inflammatory reaction . These small heat shock proteins induce the production of IL-8 by human leptomeningeal smooth muscle cells and human brain astrocytes in vitro. Therefore, an autoinflammatory process with anti-Aβ 42 antibodies and IL-8 may be involved in the pathogenesis of CAA-ri. However, caution is advised in the interpretation of these findings because the results of this study may have been influenced by an unusual presentation for this patient. The polynuclear pleocytosis in our patient was atypical compared with previous reports of CAA-ri with normal to mild mononuclear pleocytosis [1, 3]. A possible explanation for this finding is that the present patient may have had more severe leptomeningeal vessel inflammation than previously reported CAA-ri patients. Another possibility is that the initial presentation was caused by complications of viral meningoencephalitis. Inflammatory markers including IL-8 may have been high if this patient had pre-existing viral meningoencephalitis. More studies are needed to clarify the pathological role of IL-8 in CAA-ri.
The mean levels of Aβ 40 and Aβ 42 in CSF were lower in the CAA-ri patient than control subjects (Table 2). This confirms the findings of a previous case report in which a patient with CAA-ri had a low level of Aβ 42 level in CSF  and other reports that showed low levels of Aβ 40 and Aβ 42 in CSF of patients with CAA [14, 15]. Previous histological data support the deposition of Aβ 40 and Aβ 42 proteins in CAA . In CAA, both Aβ 40 and Aβ 42 may be trapped in the cerebral vasculature, and deplete these peptides from CSF. The diagnosis of CAA-ri requires recognition of CAA [1, 17]. Echo gradient imaging and presence of microbleeds could be used as a useful screening tool in CAA. We could not perform this imaging in the present study and could not make the diagnosis of CAA-ri until the brain biopsy was performed. Echo gradient imaging can help speed up the diagnostic process and prompt an earlier biopsy. It is possible that the low level of Aβ 40 and Aβ 42 in CSF could be an alternative screening marker of CAA and CAA-ri. Further studies are required to assay levels of Aβ 40 and Aβ 42 in CSF in more patients with CAA and CAA-ri.
The encephalopathy CAA-ri is treatable by adequate immunosuppressive therapy. We demonstrated the possible therapeutic effect of corticosteroids on elevated anti-Aβ 42 antibody and IL-8 in CSF of the patient with CAA-ri. It may be helpful to place some CAA-ri patients on immunosuppressive therapy, and anti-Aβ 42 antibody in CSF may be a potentially useful biological marker for the therapeutic monitoring of CAA-ri.
Written informed consent was obtained from the next of kin of the patient for publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.
AK, TS, NY, YH and TI are members of the Departments of Neurology and Geriatrics, Gifu University Graduate School of Medicine, and MT is a member of the Informative Clinical Medicine, Gifu University Graduate School of Medicine. HT is a member of the Department of Pathology, Brain Research Institute, Niigata University.
Cerebral amyloid angiopathy
Cerebral amyloid angiopathy-related inflammation
Enzyme-linked immunosorbent assay
Fluid-attenuated inversion recovery
Magnetic resonance imaging
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.