Assessment of β-Amyloid in a Frontal Cortical Brain Biopsy Specimen and by Positron Emission Tomography With Carbon 11–Labeled Pittsburgh Compound B FREE

Aggregates of β-amyloid (Aβ) in the neuropil together with hyperphosphorylated tau (HPτ) seen in neurons and neuronal processes are considered diagnostic hallmark lesions of Alzheimer disease (AD).¹ Based on current knowledge, at the first phase presumably Aβ deposits in neocortical regions; at the second phase, in the allocortex; at the third phase, additionally in the diencephalic nuclei and striatum; at the fourth phase, additionally in distinct brainstem nuclei; and, finally, at the fifth phase, also in the cerebellum and additional brainstem nuclei.² Variation in this pattern of deposition has been reported in subjects carrying the presenilin 1 mutation.³ So far, the only confident method to assess Aβ aggregates and HPτ in the brain is the histological analysis of tissue samples obtained either during life or at autopsy,^1- 2 a major methodological obstacle considering clinical drug trials of early AD.

Imaging of Aβ aggregates by Pittsburgh Compound B (PiB) positron emission tomography (PET) seems a promising method for noninvasive evaluation of patients with suspected AD.^3- 8 A case report described that a patient with dementia with Lewy bodies showed a positive correlation between carbon 11–labeled (¹¹C) PiB PET findings during life and postmortem assessment of Aβ aggregates 3 months later.⁹

Cognitive impairment, the leading symptom of AD, is also included in the clinical triad of normal-pressure hydrocephalus (NPH).¹⁰In NPH, the diagnostic accuracy is increased by intracranial pressure (ICP) monitoring,¹¹ and a tiny cortical biopsy specimen can be obtained through the bur hole for differential diagnosis. The risk of complications associated with this invasive procedure has been low. In fact, 22% to 42% of the patients with symptoms suggestive of NPH showed AD pathological lesions in frontal cortical samples.^12- 14

In this study, our objectives were to assess Aβ aggregates both applying noninvasive [¹¹C]PiB PET and invasive surgery. We compared [¹¹C]PiB PET results in 10 patients with known histopathological features (ie, the presence or absence of Aβ aggregates and HPτ in frontal cortical samples obtained during ICP monitoring).

Altogether, 125 patients underwent ICP monitoring with frontal cortical biopsy for suspected NPH at the Department of Neurosurgery, Kuopio University Hospital, between January 2004 and February 2007.

Medical records of the patients, who were 75 years or younger, were first screened according to the medical history by an independent neurologist. Exclusion criteria included poor general health, severe dementia, or severe concomitant diseases affecting the ability to cooperate in the PET examination or contraindication for magnetic resonance imaging. Approximately 50 patients fulfilled the study criteria. These patients were contacted by telephone and only 10 patients agreed to participate in the [¹¹C]PiB PET study (Table 1). Based on the neuropathological findings in the frontal cortical biopsy specimen, patients were divided into 2 groups. Six patients had Aβ (3 of them had also HPτ) pathological lesions in the biopsy specimen, whereas 4 patients had no AD-related pathological lesions in the biopsy specimen (Table 1). Close to the PET investigation, the patients were evaluated for cognitive impairment using the Clinical Dementia Rating Scale (CDR),¹⁶ CDR sum of boxes, Mini-Mental State Examination,¹⁷ Consortium to Establish a Registry for Alzheimer's Disease (CERAD) neuropsychological test battery,¹⁸ and total score of CERAD constructed as suggested by Chandler and colleagues.¹⁵

A right frontal 12-mm bur hole was made under local anesthesia. The standard site was 2 cm right from the midline ahead of the coronary suture. Prior to insertion of the intraventricular monitoring catheter, cylindrical cortical brain biopsy specimens of 2 to 5 mm in diameter and 3 to 7 mm in length were obtained through the bur hole. The samples were placed in buffered formalin and, after overnight fixation, were embedded in paraffin.

Consecutive 7-μm-thick sections were stained with hematoxylin-eosin, Bielschowsky silver impregnation technique, and immunohistochemical (IHC) methods. Briefly, deparaffinized sections were manually immunostained with antibodies directed to HPτ and Aβ (Table 2). β-amyloid antibodies included both 6F/3D (reactive to amino acid residue 10-15), labeling both parenchymal aggregates (ie, plaques) as well as cerebral amyloid angiopathy, and 4G8 (residue 18-22), labeling primarily parenchymal Aβ aggregates and especially fleecy and diffuse aggregates seen at early stages. The labeled streptavidin-biotin method (Histostain-Plus Kit; Zymed, San Francisco, California) was used with romulin 3-amino-9-ethylcarbazole chromogen (Biocare Medical, Walnut Creek, California). The sections were counterstained with Harris hematoxylin (Merck, Darmstadt, Germany), dehydrated, and mounted in DePex (BDH Laboratory Supplies, Poole, England). Omission of primary antibodies revealed no detectable staining.

The assessment of stained sections was carried out under light microscopy at magnifications ×100 to ×200. Cellular or neuritic HPτ structures were sought and rated as negative or positive. In Bielschowsky silver–stained and in Aβ-IHC–stained sections, fleecy, diffuse, and dense plaques were counted in magnification ×100 within the whole visual field (3.14 mm²) composed of gray matter.

[¹¹C]PiB was produced by the reaction of 6-OH-BTA-0 and [¹¹C]methyl triflate, as reported earlier.⁷ The radiochemical purity of the tracer was more than 98% in all [¹¹C]PiB studies. A mean (SD) 442.8 million (90.1 million) Bq (range, 255-530 million Bq) (to convert to curies, multiply by 2.7×10⁻¹¹) of [¹¹C]PiB was injected intravenously as a bolus and all patients underwent a 90-minute dynamic PET scan with a GE Advance PET scanner (General Electric Medical Systems, Milwaukee, Wisconsin) in the 3-dimensional scanning mode, as described earlier.⁷ Positron emission tomography imaging was performed without the knowledge of the neuropathological data of the patient.

Before the voxel-based statistical analysis and automated region-of-interest (ROI) analysis, dynamic images were first computed into quantitative parametric images. Parametric images representing [¹¹C]PiB uptake in each pixel were calculated as a region-to-cerebellum ratio of the radioactivity concentration over 60 to 90 minutes, as described earlier.⁷

Voxel-based statistical analyses of [¹¹C]PiB data were performed using Statistical Parametric Mapping version 99 (SPM99) and MATLAB 6.5 for Windows (MathWorks, Natick, Massachusetts), using procedures described in detail earlier.⁷ Briefly, spatial normalization of parametric images was performed using a ligand-specific [¹¹C]PiB template.⁷ The between-group comparison equaling 2-sample t tests and testing the difference in ratio values was performed as an explorative analysis covering the whole brain. Multiple comparison–corrected P values <.01 were considered significant.

To obtain quantitative regional values of [¹¹C]PiB uptake, automated ROI analysis was performed, as described earlier.⁷ Briefly, the standardized ROIs were defined using Imadeus software (version 1.50; Forima Inc, Turku, Finland) on the magnetic resonance imaging template image representing brain anatomy in accordance with MNI space (Montreal Neurological Institute database). Because this method is based on a common stereotactic space (ie, spatial normalization of the images), the operator-induced error in defining ROIs individually for each subject can be avoided. The ROIs were positioned bilaterally on the frontal cortex, lateral temporal cortex, medial temporal lobe, inferior parietal lobe, occipital cortex, cerebellar cortex, and subcortical white matter.⁸ The average regional ratio values of [¹¹C]PiB uptake were calculated using these ROIs from spatially normalized parametric ratio images (see “Statistical Parametric Mapping Analysis” subsection) and subjected to statistical analysis conducted using SPSS for Windows (release 12.0.1; SPSS Inc, Chicago, Illinois).

The study was approved by the Kuopio University Hospital Research Ethics Board. All patients provided a written informed consent prior to their participation.

The cognitive status of the 10 patients (Mini-Mental State Examination score, CDR score, CDR sum of boxes, and CERAD total score) at the time of the [¹¹C]PiB PET scan and the histological and IHC findings in the frontal cortical biopsy specimens are presented in Table 1 and Figure 1. Occasional cytoplasmic HPτ was seen in only 3 of the 10 patients. No neuropil or plaque-associated neurites (ie, no HPτ-labeled neuritic plaques) were seen. No cerebral amyloid angiopathy was seen in any of our subjects either.

The between-group SPM analysis (Figure 2) showed significantly higher [¹¹C]PiB uptake in the frontal, parietal, and lateral temporal cortices (phase 1 of regional Aβ deposition) and striatum (phase 3) in patients with Aβ aggregates in the frontal cortex compared with those without notable Aβ aggregates in the brain biopsy specimen.

The automated ROI analysis showed that the patients with Aβ aggregates had higher [¹¹C]PiB uptake in the lateral frontal and lateral temporal cortices (phase 1), anterior and posterior cingulate gyri (phases 2-3), and caudate nucleus (phase 3) than the patients without Aβ aggregates (Table 3). The difference did not reach significance in the medial temporal lobe, inferior parietal and occipital cortices (phases 1-2), or putamen and thalamus (phase 3). Figure 3 indicates representative transaxial slices of parametric [¹¹C]PiB images.

[¹¹ C]PiB uptake in the right frontal cortex (ROI) correlated (Pearson r = 0.85; P = .002) with the amount of Aβ aggregates in the right frontal cortical biopsy specimen (Figure 4).

This is, to our knowledge, the first study where living patients were assessed regarding their Aβ deposition in the brain both by means of surgery and imaging. Our results indicate that Aβ deposition in a frontal cortical biopsy specimen obtained during life is in line with the [¹¹C]PiB uptake found in PET imaging, suggesting that [¹¹C]PiB PET reflects brain Aβ deposition.

In suspected NPH, we routinely monitor intraventricular ICP and obtain a small right frontal cortical biopsy specimen to exclude or verify a specific neurodegenerative process.^11- 12,14Alzheimer disease–related lesions (ie, HPτ) are primarily seen in the temporoparietal regions,¹ whereas the deposition of Aβ starts from the neocortex, proceeding via central structures to the subtentorial structures.² Thus, a frontal cortical biopsy specimen with Aβ aggregates without HPτ would suggest early AD.² Six of our patients displayed AD-related pathological lesions in the cortical biopsy specimen, but none of them had severe dementia.

In our study, the most significant differences in [¹¹C]PiB uptake between patients with Aβ immunoreactivity in the frontal cortical biopsy specimen and those without were seen in the frontal cortex (phase 1), the lateral temporal cortex (phase 1), the anterior and posterior cingulate gyri (phases 2-3), and the caudate nucleus (phase 3). The patients with the highest Aβ load in the biopsy specimen had also the highest [¹¹C]PiB uptake in PET imaging. The correlation and SPM analysis showed that the [¹¹C]PiB uptake increased with increasing Aβ load in the biopsy specimen. Our findings are congruent with the previous PET data from patients with AD^4,7 or mild cognitive impairment^8,19 and from healthy controls.

The study groups with or without frontal cortical Aβ aggregates were similar in age, sex, and time from the cortical biopsy to PET imaging. Thus, the patients were suitable for the main objective of the study: methodological comparison of the [¹¹C]PiB PET imaging findings and cortical biopsy specimen to indicate Aβ deposition in the brain independently from the cognitive status of the patients. Case 5 was excluded from the SPM analysis because there was only 1 fleecy plaque in the biopsy specimen, and the presence or absence of that case did not change the result. The interval between the surgery and imaging (Table 1) might to some extent skew the interpretation of our results. In parallel with the increase of IHC/Aβ labeling, an increase in the plaque count was noted while using silver stain. The surgically obtained cortical biopsy sample was small and thus false-negative results are possible, and the absence of Aβ aggregates in the frontal cortex does not securely reflect the Aβ deposition in other isocortical brain regions. Also, the false-positive result of sampling of a very small area of high plaque load is possible. To validate the clinical significance of the surgically obtained frontal cortical biopsy specimen in diagnosing brain amyloidosis, further systematic assessment, preferably by postmortem verification of brain pathological lesions in subjects in whom a biopsy specimen has been taken during life, needs to be carried out. However, all the patients with normal biopsy results also had negative PiB results despite the interval between biopsy and PET imaging. On the other hand, along with the potential technical errors of PET imaging and labeling of the histological samples, case 6 (IHC positive but PiB negative) emphasizes the potential that there might be types of Aβ deposits that PiB does not detect. Diagnosis of NPH was based on clinical symptoms and intraventricular ICP monitoring. Four of the 7 patients with NPH also had concomitant AD-related lesions (Aβ) in the cortical biopsy specimen. This is in line with the previous findings of notable comorbidity of NPH and AD-related pathological lesions.^12- 14

This study supports the use of [¹¹C]PiB PET in the evaluation of Aβ deposition in, for example, mild cognitive impairment, AD, or NPH. Large and prospective studies are required to verify whether [¹¹C]PiB PET will become a tool in diagnosing AD. Another potential use of [¹¹C]PiB would be the quantitative monitoring of Aβ deposits in the brain in subjects under treatment in pharmaceutical trials of early AD targeting amyloid accumulation.

Correspondence: Ville Leinonen, MD, PhD, Department of Neurosurgery, Kuopio University Hospital, PO Box 1777, 70211 Kuopio, Finland (ville.leinonen@kuh.fi).

Accepted for Publication: February 5, 2008.

Published online: August 11, 2008 .

Author Contributions:Study concept and design: Alafuzoff, Savolainen, Tapiola, Pirttilä, J. Rinne, Jääskeläinen, Soininen, and J. O. Rinne. Acquisition of data: Leinonen, Alafuzoff, Suotunen, Savolainen, Någren, and Soininen. Analysis and interpretation of data: Leinonen, Alafuzoff, Aalto, Pirttilä, and J. O. Rinne. Drafting of the manuscript: Leinonen and Aalto. Critical revision of the manuscript for important intellectual content: Alafuzoff, Aalto, Suotunen, Savolainen, Någren, Tapiola, Pirttilä, J. Rinne, Jääskeläinen, Soininen, and J. O. Rinne. Statistical analysis: Aalto. Obtained funding: Soininen and J. O. Rinne. Administrative, technical, and material support: Alafuzoff, Aalto, J. Rinne, Jääskeläinen, and Soininen. Study supervision: Savolainen, Pirttilä, J. Rinne, Jääskeläinen, Soininen, and J. O. Rinne.

Financial Disclosure: None reported.

Funding/Support: The study was supported by grant 5772720 from the Kuopio University Hospital, the Academy of Finland (project 205954), and the Sigrid Juselius Foundation.