Association of amyloid angiopathy with microbleeds in logopenic progressive aphasia: an imaging-pathology study

Abstract

Background:

Cerebral microbleeds (MB) and superficial siderosis (SS) are frequent neuroimaging findings in patients with logopenic progressive aphasia (LPA), often with frontal lobe predilection. Cerebral amyloid angiopathy (CAA) is hypothesized to be the major pathologic determinant of MB/SS in these patients; however, neuroimaging-pathologic data are limited.

Methods:

All patients who had been prospectively recruited by the Neurodegenerative Research Group at Mayo Clinic (Rochester, MN) between 2010 and 2015 that met the following inclusion criteria were included: 1) received an antemortem LPA diagnosis; 2) had a gradient-recalled echo (GRE) T2-weighted magnetic resonance imaging (MRI) performed; 3) died and completed a brain autopsy. Demographic, genetic, neuroimaging, clinical and pathologic characteristics were compared between patients with/without MB/SS. Two-tailed Fisher’s Exact and Wilcoxon rank sum tests were used for comparison of categorical and continuous variables, respectively.

Results:

Thirteen patients met inclusion criteria, six (46%) had MB/SS on MRI. Moderate/severe CAA was associated with the presence of MB/SS, p=0.029. As expected, MB/SS most frequently involved the frontal lobes, followed by the parietal lobes. No clear associations were found between regional MB/SS distribution and regional distribution of CAA or hypometabolism on [¹⁸F]-fluorodeoxyglucose PET. There was some evidence for a regional association between MB/SS and uptake on Pittsburgh compound B although not in all patients. No formal statistical analyses to assess topographic relationships were performed due to small sample size.

Conclusions:

The presence of MB/SS is a strong indicator of underlying moderate/severe CAA in LPA, although the biological mechanisms underlying the topographic distribution of MB/SS remain unclear.

Introduction

Cerebral microbleeds (MBs), small focal hypointense lesions, and superficial siderosis (SS) detectable on gradient-recalled echo (GRE) T2-weighted magnetic resonance imaging (MRI), occur in approximately one-third of patients with logopenic progressive aphasia (LPA) (1, 2), an atypical Alzheimer’s disease (AD) phenotype. The presence of MB/SS have been associated with cognitive decline and risk of dementia in the general population (3), although there is no evidence suggesting that MB/SS affect the rate of cognitive decline in AD patients (4, 5). Whereas multiple risk factors such as older age, hypertension and the presence of white matter hyperintensities are linked to the presence of MB/SS (6–8), there is mounting evidence of an association between MB/SS and cerebral amyloid angiopathy (CAA) (9), especially in AD patients(1, 2). Several studies have reported predominantly cortical topography of MB/SS in AD with the occipital lobe being the most frequently affected site supporting the hypothesis of CAA being the dominant risk factor (4). However, the spatial distribution of MB/SS in atypical AD, including LPA, seems to have a more frontal predilection (1, 2). Additionally, MB/SS are more numerous when compared to those with typical (amnestic) AD (2). Neuroimaging-neuropathologic data are scarce, and hence it remains unclear to what extent CAA contributes to the presence of MB/SS, specifically in LPA, and whether there is a topographic relationship between the severity of CAA and the presence of MB/SS.

In this autopsy study we sought to determine if the severity of autopsy-confirmed CAA is associated with presence of MB/SS in LPA patients and whether there is a region-specific association.

Materials and Methods

We identified all patients who had been prospectively recruited by the Neurodegenerative Research Group at Mayo Clinic (Rochester, MN) between 2010 and 2015 that met the following inclusion criteria: 1) received an antemortem LPA diagnosis by published criteria (10); 2) had a GRE T2*-weighted MRI; 3) died and completed a brain autopsy. This study was approved by the Mayo Clinic institutional review board. All participants and/or their proxies signed a written informed consent form before taking part in any research activities in accordance with the Declaration of Helsinki.

All patients underwent a standardized MRI imaging protocol at 3T that included a 2D GRE T2*-weighted sequence (TR/TE=200/20ms; flip angle=12°; FOV=20cm; in-plane matrix=256×224; phase FOV=1.00; slice thickness=3.3mm; bandwidth =15.63 kHz) to detect the presence of MBs. MBs were defined as homogeneous hypointense lesions up to 10 mm in diameter in the grey/white matter. The regional location of MBs was assigned using the Mayo Clinic adult lifespan template that was registered into the space of the T2*-weighted image (2, 11). MBs were assigned as either lobar, cerebellar, brainstem or deep/supratentorial grey/white matter. All patients also underwent amyloid-beta PET imaging with Pittsburgh compound B (PiB) and [¹⁸F]fluorodeoxyglucose (FDG) PET. Patients were injected with ~628MBq (range, 385–723MBq) of PiB or ⁓459 MBq (range, 367–576 MBq) of ¹⁸F-FDGand after a 40-minute (PiB) or 30-minute (FDG) uptake period, a 20-minute PiB or 8-minute FDG scan was obtained. A global PiB standardized uptake value ratio (SUVR) was generated by normalizing to the cerebellar crus grey matter and FDG SUVR was normalized to pons. Individual-level patterns of hypometabolism were assessed using CortexID (GE Healthcare) whereby activity at each voxel is z-scored to an age-segmented normative database.

All autopsied cases were evaluated according to standard neuropathologic examination following cortical sampling according to CERAD (12) with thioflavin S fluorescent microscopy used to assign CAA scores (13) as: none, mild, moderate and severe. Arteriolosclerosis was rated semi-quantitatively as: none, mild, moderate and severe. Presence of lacunar and large infarcts, microinfarcts and hemorrhages were recorded, and a vascular score was assigned based on published guidelines (14): 0=no vascular lesions; 1=mild arteriolosclerosis/CAA only; 2=moderate/severe arteriolosclerosis/CAA only; 3=presence of cortical microinfarcts without lacunar/large infarcts; 4=presence of lacunar/large infarcts. All patients underwent apolipoprotein E ɛ4 (APOE ɛ4) genotyping. All statistical analyses were performed in JMP Pro 14.1.0 (SAS Institute Inc.) software. Two-tailed Fisher’s Exact and Wilcoxon rank sum tests were used for categorical and continuous variables, respectively. Significance was set at p<0.05.

Results

A total of 13 LPA patients met our inclusion criteria. Their demographic, genetic, neuroimaging and pathologic characteristics are summarized in Table 1. Out of 13 patients, 46% (6/13) had MB/SS. The presence of moderate/severe CAA was more frequent in those with MB/SS compared to those without (83% versus 14%, p=0.029). There were no other differences between those with and without MB/SS. Patients with multiple MB/SS had those distributed bilaterally whereas single MB/SS did not always localize to dominantly-affected hemisphere. MB/SS localized most frequently in the frontal and parietal lobes; no clear associations were identified between the location of the MB/SS and the regional severity of CAA or the regional distribution of hypometabolism within patients (Figure 1). The region with the peak uptake on PiB-PET coincided with the region with the most MB/SS in four patients, but not in patients 2 and 6 (Figure 1). However, no formal statistical analyses to assess topographic relationships were performed due to small sample size.

Table 1.

Demographic, genetic, neuroimaging and pathologic characteristics of LPA patients with and without microbleeds/superficial siderosis on GRE T2*-weighted MRI

	All n=13	No MB/SS n=7	MB/SS n=6	p-value
Demographics
Female, n, %	4 (31%)	2 (29%)	2 (33%)	>0.99
APOE ε4 carrier, n, %†	9 (69%)	4 (57%)	5 (83%)	0.56
Education, years	16 (12 – 18)	16 (14 – 18)	16 (12 – 18)	0.28
Family history, n, %	5 (38%)	3 (43%)	2 (33%)	>0.99
Age at onset, years	63 (53 – 78)	62 (54 – 78)	65 (53 – 75)	0.39
Age at death, years	70 (61 – 86)	69 (64 – 86)	76 (61 – 84)	0.28
Neuroimaging
Time from MRI to death, years	4.4 (0.3 – 6.9)	4.1 (0.3 – 6.9)	5.4 (3.6 – 6.5)	0.13
Time from PiB-PET to death, years	4.3 (0.3 – 6.9)	4.0 (0.3 – 6.9)	5.1 (0.4 – 6.5)	0.52
Global PiB SUVR	2.1 (1.17 – 2.76)	2.1 (1.17 – 2.76)	2.1 (1.49 – 2.7)	0.57
Frontal PiB SUVR	2.48 (1.34 – 3.3)	2.48 (1.34 – 3.31)	2.35 (1.47 – 3.05)	0.72
Temporal PiB SUVR	2.36 (1.27 – 3.0)	2.38 (1.27 – 2.97)	2.18 (1.47 – 2.91)	0.52
Parietal PiB SUVR	2.38 (1.34 – 3.2)	2.42 (1.34 – 3.15)	2.23 (1.62 – 3.03)	0.43
Occipital PiB SUVR	1.87 (1.4 – 2.86)	1.87 (1.40 – 2.86)	1.80 (1.65 – 2.72)	0.72
Microbleeds	0 (0 – 23)	n/a	2 (0 – 23)	n/a
Superficial siderosis	0 (0 – 2)	n/a	0 (0 – 2)	n/a
Primary pathologic diagnosis, n, %
Alzheimer’s disease	10 (77%)	5 (71%)	5 (83%)	>0.99
Diffuse LBD‡	2 (15%)	1 (14%)	1 (17%)	>0.99
Frontotemporal lobar degeneration	1 (8%)	1 (14%)	0 (0%)	>0.99
Pathologic characteristic§
CAA, n, %	10 (77%)	5 (71%)	5 (83%)	>0.99
moderate/severe	6 (46%)	1 (14%)	5 (83%)	0.029
Arteriolosclerosis, n %	9 (69%)	6 (86%)	3 (50%)	0.27
moderate/severe	2 (15%)	1 (14%)	1 (17%)	>0.99
Microinfarcts, n %	0 (0%)	0 (0%)	0 (0%)	>0.99
Lacunar/large infarcts, n, %	0 (0%)	0 (0%)	0 (0%)	>0.99
Vascular score/4	2 (0 – 2)	1 (1 – 2)	2 (0 – 2)	0.17
Atherosclerosis, n, %	3 (23%)	1 (14%)	2 (33%)	0.56
Hemorrhages, n, %¶	2 (15%)	0 (0%)	2 (33%)	0.19

Data are presented as median (range) for continuous variables; n/a – not applicable

^†APOE ɛ4 carriers had the following haplotypes: 7/9 - APOE ɛ3/APOE ɛ4; 2/9 - APOE ɛ2/APOE ɛ4

^‡Both patients had a secondary pathologic diagnosis of Alzheimer’s disease: A2 B3 C2 – no MB/SS case; A1 B2 C1– MB/SS case

^§Neuropathologic examination of both hemispheres was performed in 4/13 (31%) cases 3 of which were those with MB/SS; in the remaining cases the dominant hemisphere was examined in 6/9 (67%) cases and non-dominant hemisphere was examined in 3/9 (33%) one of each was the patient with MB/SS (patient 2). In patients with MB/SS whose pathologic evaluation was limited to one hemisphere, at least one of MB/SS were present on the side which was neuropathologically examined

^¶One patient had subarachnoid hemorrhage in superior frontal regions; another patient had acute and chronic subdural hematomas

Figure 1. Distribution of microbleeds/superficial siderosis across the brain regions and their topographic correlation with regional severity of cerebral amyloid angiopathy, regional PiB-PET uptake and FDG-PET hypometabolism within each patient.

Non-lobar microbleeds were observed in supratentorial deep grey and white matter junction in patients 2 and 3 and in brainstem in patient 4

Discussion

In this study we found that the presence of MB/SS in LPA is associated with the presence of moderate/severe CAA at autopsy. The topographic distribution of MB/SS showed a tendency to match more closely with amyloid-beta uptake on PET rather than the regional distribution of CAA, although associations were not consistent across all patients.

MB/SS occurred in almost half of the LPA patients in this study which is even higher than we previously reported (1) underscoring that it is a very frequent neuroimaging finding in LPA. In keeping with previous studies, the majority of MB/SS were cortical and a predilection for the frontal lobe was observed in three (50%) of the patients (1, 2); however, two patients showed MBs only in the parietal lobe. Most importantly, in this study, we found pathologic evidence that moderate/severe CAA is associated with presence of MB/SS in LPA. In fact, only one patient who had moderate/severe CAA at autopsy did not have evidence for MB/SS on imaging. Hence, when MB/SS are observed on imaging, our data suggest that these patients likely have moderate/severe CAA. Four patients showed only mild CAA and none of them had evidence for MB/SS on imaging, suggesting a severity threshold for the development of MB/SS. Moreover, CAA was the only variable associated with MB/SS, with no other demographic, genetic or autopsy vascular relationships, emphasizing that perhaps CAA is the dominant risk factor of MB/SS development in LPA.

The underlying mechanism explaining the development of MBs in the one patient without CAA is unclear as no vascular abnormalities at autopsy were found. In fact, the patient had a primary pathologic diagnosis of diffuse Lewy body disease (LBD); LBD was associated with frontal MBs regardless of co-existing AD/CAA in one study (15).

Interestingly, while most MB/SS co-occurred with some degree of regional CAA we did not observe a higher frequency/number of MB/SS in the regions with the most severe CAA;nor did we observe a topographic relationship between MB/SS location and regional PiB uptake or regional hypometabolism. However, no formal statistical analysis was performed due to modest sample size. This is somewhat consistent with another neuroimaging-neuropathologic study of MB/SS and CAA where MB/SS occurred at the sites with reduced CAA (16). Therefore, it is possible that CAA, though undeniably involved in the mechanisms of MB/SS in AD, might indirectly contribute to fragility of blood vessels at specific sites through alteration of microvascular network and vascular autoregulation with changes in systemic blood pressure(17). Therefore, MB/SS might occur at the weak points of circulation or in vessels with anatomical abnormalities which cannot handle these changes in blood pressure potentially caused by moderate/severe CAA upstream or downstream of the area of MB/SS(16).

Whereas maximum Whereas PiB uptake matched the location of MB/SS in the majority of cases, this relationship was unclear since in two cases the location of MB/SS was observed at the point of minimal PiB uptake. In two of the cases where peak PiB uptake coincided with the location of MB/SS, PiB SUVR values had minimal variation across regions, therefore it is questionable whether the slightly higher PiB SUVR value in a specific region represented a true peak. Lastly, regional PiB uptake did not match the severity of CAA once again emphasizing that increased vascular amyloid-beta deposition while contributing to overall vessel fragility resulting in MB/SS might not cause vessel rupture at a specific site.

Ample neuroimaging data and detailed neuropathologic evaluation are the strengths of the present study. However, the modest sample size might have decreased our ability to detect additional associations, e.g. genetic and demographic. Furthermore, the neuropathologic evaluations in most cases examined only one hemisphere and so we cannot rule out the possibility that CAA was present in the other hemisphere in some cases.

Microbleeds/superficial siderosis are frequent in patients with LPA and when present are associated with moderate/severe CAA. However, the degree of CAA does not explain the fragility of cerebral vessels at specific sites; therefore, further research is warranted to understand the biological mechanisms governing the development of MB/SS.

Data Availability Statement

Anonymized data will be available from the corresponding author upon request from any qualified investigator for purposes of replicating procedures and results.