Cerebrospinal Fluid Anti-Amyloid-β Autoantibodies and Amyloid PET in Cerebral Amyloid Angiopathy-Related Inflammation

INTRODUCTION

Cerebral amyloid angiopathy-related inflammation (CAA-ri) is a rare type of meningoencephalitis that affects a subgroup of CAA patients who develop vascular inflammation. Clinically, CAA-ri usually presents with cognitive decline and focal neurological symptoms, white matter abnormalities on T2-weighted images, and multiple microhemorrhages on T2^*-GRE weighted images [1]. In some cases, reaching a diagnosis is complex, and invasive procedures, such as a cerebral biopsy, are required. This complexity, together with the variable clinical course and the usual responsiveness of CAA-ri to immunosuppressive treatment [2], highlights the need of biomarkers to allow early diagnosis.

The CAA-ri pathogenesis remains unknown. The genetic studies have found an association with the APOE ɛ4 allele [3]. The better characterization of CAA-ri through biomarkers have allowed the development of diagnostic criteria for probable CAA-ri based on typical clinico-radiological findings without requiring a biopsy [2, 3].

Cerebrospinal fluid (CSF) and amyloid positron emission tomography (PET) are increasingly recognized as markers of CAA, however, there are few studies evaluating biomarkers in CAA-ri [4–6]. It is worthy to mention that increased CSF anti-Aβ autoantibodies in the acute phase of CAA-ri were first reported by DiFrancesco et al. [7], while Piazza et al. recently suggested that they might be a specific marker of CAA-ri [8].

We describe a case series of four CAA-ri patients in whom CSF AD biomarkers, anti-Aβ autoantibodies, and amyloid PET biomarkers were analyzed.

METHODS

RESULTS

DISCUSSION

The present study analyses CSF anti-Aβ autoantibodies and amyloid PET in CAA-ri. This pilot study suggests that CSF analysis and amyloid PET might add diagnostic specificity to the proposed clinical criteria for CAA and that CSF anti-Aβ autoantibodies might help diagnose and monitor the response to treatment in CAA-ri [8].

The clinical presentation consisted of focal neurological symptoms and rapidly progressive cognitive decline. MRI examinations showed, besides CAA-related findings, asymmetrical and confluent WML, indicating vasogenic edema that remitted shortly after corticosteroids. However, clinical prognosis was poor. Two patients died during the follow-up and one developed dementia.

The CSF data showed a pattern compatible with an underlying CAA [10]. Low levels of Aβ₄₀ and Aβ₄₂ in CSF could be an alternative diagnostic marker of CAA [4, 12]. Some [4, 8], but not all [5], previous works have reported increased CSF Aβ₄₂ and Aβ₄₀ levels in the acute phase of CAA-ri. We did not find this elevation, but the antecedent MCI (prodromal AD) in three out of four patients, which might have affected CSF AD biomarker levels, should be noted.

Florbetapir-PET showed widespread cortical amyloid deposition. Regions presenting with inflammation in the acute phase seemed to present lower retention. This pilot study might suggest a reduction of Aβ uptake after remission that must be confirmed with further dedicated studies. There is only one recent case report with Pittsburg Compound B PET in CAA-ri [13]. This finding suggests a relationship between inflammation and amyloid clearance, a liaison that has been described in the Bapineuzumab and gantenerumab studies [14–16]. Beyond the mechanistic importance of amyloid PET, amyloid imaging might aid in the diagnosis of patients with focal syndromes and suspected CAA in the acute phase given the aforementioned conflicting results of CSF Aβ₄₀ and Aβ₄₂ levels at this stage [6, 13].

The genetic analyses showed that the APOE ɛ4 allele was overrepresented (when compared to Spanish population) [17]. This has been already reported in CAA-ri [2, 3]. Of note, beyond the clinical similarities between CAA-ri and the amyloid related imaging abnormalities observed in AD patients treated with anti-amyloid therapies [8, 14–16], were more frequent in the higher-dose groups of patients carrying the APOE ɛ4 allele and with a major vascular amyloid burden [18].

Autoantibody titer were high during the acute phase and decreased to normal values after corticosteroids [4, 8]. To our knowledge this is the first study replicating the findings of Piazza et al. on the utility of anti-Aβ autoantibodies titer to diagnose CAA-ri. We suggest that CSF anti-Aβ autoantibodies should be added to the diagnostic criteria for CAA-ri in order to reduce the number of cerebral biopsies. Moreover, CSF anti-Aβ autoantibodies might be used to monitor treatment response. This is usually done with MRI, but although clinical response to immunosuppressive therapy is associated with lesion volume reductions, CAA-ri is also associated with irreversible WML [2, 18].

The main limitation of the study is the absence of neuropathological confirmation in our patients. Nonetheless, our findings support the hypothesis that anti-Aβ autoantibodies play a role in the pathophysiological process of CAA-ri. Other limitations include the absence of a control group of sporadic CAA cases or AD that should be explored in future works. Finally, brain atrophy might have affected the quantification of Aβ deposition.

In conclusion, our results support the utility of CSF core AD biomarkers, CSF anti-Aβ autoantibodies, and amyloid PET in CAA-ri. However, further studies with larger samples and neuropathological confirmation are needed to confirm these findings.