Lack of association between cortical amyloid deposition and glucose metabolism in early stage Alzheimer´s disease patients.

BACKGROUND: Beta amyloid (Aβ) causes synaptic dysfunction leading to neuronal death. It is still controversial if the magnitude of Aβ deposition correlates with the degree of cognitive impairment. Diagnostic imaging may lead to a better understanding the role of Aβ in development of cognitive deficits. The aim of the present study was to investigate if Aβ deposition in the corresponding brain region of early stage Alzheimer´s disease (AD) patients, directly correlates to neuronal dysfunction and cognitive impairment indicated by reduced glucose metabolism. PATIENTS AND METHODS: In 30 patients with a clinical phenotype of AD and amyloid positive brain imaging, 2-[18F] fluoro-2-deoxy-d-glucose (FDG) PET/CT was performed. We extracted the average [18F] flutemetamol (Vizamyl) uptake for each of the 16 regions of interest in both hemispheres and computed the standardized uptake value ratio (SUVR) by dividing the Vimazyl intensities by the mean signal of positive and negative control regions. Data were analysed using the R environment for statistical computing and graphics.
RESULTS: Any negative correlation between Aβ deposition and glucose metabolism in 32 dementia related and corresponding brain regions in AD patients was not found. None of the correlation coefficient values were statistically significant different from zero based on two-sided p- value.
CONCLUSIONS: Regional Aβ deposition did not correlate negatively with local glucose metabolism in early stage AD patients. Our findings support the role of Aβ as a valid biomarker, but does not permit to conclude that Aβ is a direct cause for an aberrant brain glucose metabolism and neuronal dysfunction.

Introduction

Alzheimer’s disease (AD) is the most common cause of dementia in the elderly people. The hallmark pathologies include beta-amyloid (Aβ) depositions in plaques and brain vessels, tau pathology, microglia activation, and inflammation. The causes of AD are unknown; however, Aβ may play an important role in development of AD. According to the amyloid cascade hypothesis, Aβ causes synaptic dysfunction and neuronal death leading to cognitive impairment.[1,2]

However, several studies suggested only a modest correlation between Aβ pathology and cognition. There is little correlation between amyloid plaques location at autopsy and affected brain regions according to patient’s clinical symptoms.[3] A large burden of amyloid plaques is found in subjects without cognitive deficits.[4] Amyloid plaque formation does occur late in hippocampus, although this structure is the first to fail clinically.[5] Indeed, the reduction of plaque load in the brain in therapeutic trials has not yielded to cognitive benefit in AD patients.[6]

In contrast, the burden of neurofibrillary tangles, consisting of hyperphosphorylated tau, correlates with the degree of cognitive impairment in AD.[7] In vivo studies using PET tracers for tau replicated these findings showing that more advanced Braak stages are associated with decreased global cognitive status.[8,9] Tau pathology leads to a reduced glucose metabolism correlating with cognitive decline.[10] Thus, tau pathology rather than amyloid accumulation, may contribute to cognitive dysfunction in AD patients.

Imaging may lead to a better understanding the role of Aβ in the development of cognitive deficits. Techniques such as 2-[18F]fluoro-2-deoxy-d-glucose (FDG) Positron Emission Tomography (PET) or amyloid PET are widely used for supporting the diagnosis of dementia. To exclude cerebral pathologies, such as tumors, subdural hematoma or normal pressure hydrocephalus, magnetic resonance imaging (MRI) is used. Morphologically, MRI may detect a hippocampal volume reduction in AD [11]. Amyloid PET allows in vivo detection of amyloid plaques indicating the pathophysiologic state of dementia. Several tracers, such as 18[F] florbetapir, 18[F] flutemetamol and 18[F] florteban are approved by the US Food and Drug Administration (FDA). Timing of amyloid accumulation is at a preclinical stage of AD.[12]

However, healthy individuals without cognitive symptoms can have a positive amyloid PET scan. Thus, amyloid imaging may be helpful for differential diagnosis in early onset dementia, particularly to rule out AD dementia. FDG PET is an important tool to detect early neurodegenerative dementia, differentiate neurodegenerative dementia or comorbidity of other neurodegenerative disease. FDG PET is described as a neuronal injury biomarker in AD and neuronal dysfunctionis indicated by reduced glucose metabolism.[13] In AD patients, FDG PET demonstrates a glucose metabolic reduction in the parietotemporal association cortices, posterior cingulate and precuneus regions.[14] In the later stages of AD, hypometabolic regions spread to the frontal association cortices.[15] In patients with mild cognitive impairment (MCI), the transitional stage between aging and AD, FDG PET appears to add the greatest prognostic information.[16]

According to the amyloid cascade hypothesis, in brain regions with amyloid deposition, neuronal injury and an altered FDG metabolism may be expected. However, several studies showed controversial findings regarding the correlation of amyloid deposition and glucose metabolism.[17, 18, 19] Thus, it is still not clear if amyloid deposition comparably to tau pathology is a leading cause of cognitive impairment.

In this study, we included patients with a clinical phenotype of AD with cognitive dysfunction and amyloid deposition in cortical brain areas as detected by Aβ binding PET tracer.

The aim was to investigate if Aβ deposition in the corresponding brain region directly correlates to neuronal dysfunction indicated by reduced glucose metabolism. Such a correlation should be easy and convenient to recognize when clinically AD is at a relatively early stage and impaired glucose metabolism is still restricted to certain areas and not globalized. Therefore, our cohort consists mostly of patients at the early stage of disease.

Patients and methods

Patients

Ninety patients underwent amyloid PET analysis at the Institute of Nuclear Medicine, Kepler University Hospital, Neuromed Campus, Linz, between January 2016 and November 2017. These patients were assigned from Departments of Gerontology, Neurology or Psychiatry of various hospital institutions located in upper Austria. At least all amyloid positive patients (n = 30) underwent a comprehensive clinical and neuropsychological evaluation. Probable AD was diagnosed clinically and according to the S3 guidelines for dementia by experienced clinicians.[20] For screening, the mini mental state examination (MMSE) according to Folstein was performed.[21] Each subject underwent computed tomography or MRI to rule out any structural abnormalities, such as brain tumours, hematomas, hydrocephalus or ischemia, as the cause for dementia. Vitamin deficiencies and thyroid abnormalities were excluded by blood analysis. FDG PET, amyloid PET and neuropsychological testing were acquired within 60 days. Brain Aβ deposition was quantified by performing PET scans using the tracer 18[F] flutemetamolm (Vizamyl). In a standardized procedure patients were rated as amyloid positive (n = 30, age 65.0 ± 14.3 years) or amyloid negative (n = 60, age 64.6 ± 8.7 years).

FDG PET was used to evaluate brain glucose metabolism. It was rated in clinical routine by experienced clinicians blinded to clinical symptoms using the Neuro Q, Version 3.5, 2007-analysis system, a schematic summary, comparing the patients scan to the scan of an asymptomatic control group. For our investigation, we selected 32 dementia related brain regions as regions of interest (Table 1). The regions of interest were based on brain areas, which are suggested by the Neuro Q program and include all dementia related brain regions. For each region NeuroQ compares the FDG metabolism to that of a cohort of normal persons, numbers indicate extent of standard deviation in comparison to a normal situation. Negative numbers represent relative hypometabolism.

Table 1

Regions of interest

superior frontal cortex

middle frontal cortex

Inferior frontal cortex

anterior cingulate cortex

posterior cingulate cortex

sensorimotoric cortex

superior lateral temporal cortex

medial anterior temporal cortex

medial posterior temporal cortex

inferior lateral anterior temporal cortex

inferior lateral posterior temporal cortex

superior parietal cortex

inferior parietal cortex

parietotemporal cortex

primary visual cortex

associative visual cortex

For correlation, we manually defined cortical areas of representative gyri in dementia related brain regions of amyloid positive PET scans and excluded non cortical areas and sulci to avoid bias. Pons and the cerebellum represent the reference regions for positive and negative control within each brain. Further, we evaluated the Vizamyl uptake value ratio in corresponding brain regions. We extracted the average Vizamyl uptake for each of the 16 regions of interest in both hemispheres and computed the standardized uptake value ratio (SUVR) by dividing the Vimazyl intensities by the mean signal of the individual positive and negative control regions as mentioned above.

Positron emission tomography

All PET scans were obtained with a Philips Gemini GXL PET/CT. For amyloid imaging patients received 185 MBq of 18F-Flutemetamol i.v. The tracer was distributed by GE healthcare Austria. PET/CT images were obtained 60 min after tracer injection.

FDG PET had been scheduled on a different day. Patients fasted for a minimum of 6 h before FDG injection to ensure standardized metabolic conditions. Blood glucose level was measured and had to be < 160 mg % in all patients. 185 MBq of FDG was injected i.v.. PET images were acquired 30 min post injection (3D acquisition). The scanner acquires transaxial planes, simultaneously covering an 18 cm axial field of view. Eliminating the sub-sampling required in conventional techniques, line-of-response (LOR) removes averaging and consequent image degradation. Detector material is gadolinium oxyorthosilicate (GSO) with a crystal size of 4 x 6 x 30 mm. A 6-slice helical CT – Philips brilliance air 6 – was used for attenuation correction. For further evaluation, data were transferred to a Hermes Medical Solutions, Sweden work station (HERMES).

PET interpretation was done visually by two experienced nuclear medicine specialists in knowledge of clinical data of the patient and consensually. The procedure and technical data are described in detail by Pichler et al.[22,23]

Statistics

Data were analysed using the R environment for statistical computing and graphics.[24] Scatter plots of the amyloid mean score versus the FDG mean score visualize separately for each region and side the association. To test for association between these two scores, the Pearson correlation coefficient was determined for each region and side. Two-sided p-values were calculated assuming normally distributed data. P-values were for each side corrected for multiple testing using Holm’s method.[25]

The study was done in accord with ethical standards and in accord with the Helsinki Declaration of 1975.

Results

In clinical routine, 90 amyloid PET scans for diagnosis of dementia were performed. In detail, AD was diagnosed in 30 patients (65.0 ± 14.3 years), 16 male and 14 female. All patients with amyloid positive PET scan underwent also PET scanning with FDG and computed tomography (CT) or MRI, as well as neuropsychological and clinical examination.

In the AD group mean score MMSE was 23 ± 5 (n = 30).

Figures 1 and 2 demonstrate FDG PET and amyloid positive PET images of patients with the clinical suspected diagnosis of AD.

Figure 1

2-[18F] fluoro-2-deoxy-d-glucose (FDG) and amyloid brain PET/CT of 59-year-old woman. (A) FDG brain PET/CT of a 59-year-old woman with a history of fluctuating cognitive impairment (mini mental state examination [MMSE] = 14/30). Glucose hypometabolism was demonstrated in the parietal dorsolateral and temporolateral, and occipatal cortical areas. The glucose metabolism in the left temporomesial area is weak. The other cortical structures show a slight attenuation of FDG metabolism. Basal ganglia show more intense uptake compared to the cortical areas. This FDG brain PET study shows the typical picture of abnormal glucose metabolism that occurs in Alzheimer´s disease (AD) and is additionally compatible with pronounced microvascular changes. (B) On the amyloid PET a non-specific tracer accumulation from the pons to the basal ganglia is evident. PET images of the white matter demonstrate individual non-specific enrichments. In the frontal and temporal cortices as well as sporadically in the parietal cortical areas, a pathological tracer accumulation occurs. This global cortical tracer uptake is consistent with the neuropathology of AD.

Figure 2

FDG and amyloid brain PET/CT of a 70-year-old woman. (A) 2-[18F] fluoro2-deoxy-d-glucose (FDG) brain PET/CT of a 70-year-old woman, who presented with a history of cognitive decline (mini mental state examination [MMSE] = 15/30). The glucose metabolism in the cerebral cortex is inhomogenous and moderately attenuated. In the cerebellum, normal glucose metabolism was demonstrated. This FDG brain PET study does not show the typical picture of abnormal glucose metabolism that occurs in Alzheimer´s disease (AD), but temporomesial and temporolateral some decreased tracer uptake can be observed. Additionally the images are compatible with pronounced microvascular changes. (B) Amyloid PET images demonstrate pathologically increased tracer accumulation in the entire brain, more pronounced in the frontal and temporal cortical areas. This is compatible with the diagnosis of AD.

In 43 subjects with amyloid negative PET scans a MMSE was performed. In the non AD group the mean score MMSE was 26 ± 3 (n = 43). Subjects with negative amyloid PET (n = 60, 24 female, 36 male, age 64.6 ± 8.7) received the following diagnosis according to ICD-10: affective disorders (n = 33), Parkinson’s disease (n = 3), psychoorganic syndrom (POS, n = 3), Hashimoto’s encephalopathy (n = 1), hepatic encephalopathy (n = 1), frontotemporal dementia (FTD, n = 8), vascular dementia (vaD, n = 8) and β-amyloid associated angiopathy (n = 3).

As shown in Figure (3) and Figure (4) we did not find any significant negative correlation between amyloid deposition and glucose metabolism in 32 dementia related and corresponding brain regions in AD patients. The estimated correlation coefficient values differed between 0.48 and -0.32. None of the correlation coefficient values were statistically significant different from zero based on two-sided p- value at significance level 0.05 after correcting for multiple testing for each side using Holm´s method. No statistical evidence was found to confirm a negative correlation between amyloid deposition and glucose metabolism in general or specifically for some brain regions.

Discussion

There is an ongoing debate to which extent amyloid is related to AD pathology. The concept of a direct mechanism leading to clinical manifestation lead to various trials of vaccination therapies. As therapeutic success was disappointing the pathophysiological role of amyloid in AD had to be rediscussed.

In AD pathology, the amyloid cascade hypothesis may play a fundamental role.[2] Plaques in AD brains consist of insoluble Aβ peptides cleaved by different secretases from the amyloid precursor protein (APP).[26] The cleavage results in Aβ-40 with a length of 40 amino acids and Aβ-42 with a length of 42 amino acids, which is the plaque proning form. Distinct plaque subtypes with low (diffuse plaques) and high (cored or neurotic plaques) proportion of fibrillar components have been identified.[27]

In fact, insoluble Aβ exceeds soluble forms of Aβ by a factor of about 100-fold in AD brain.[28] However, Aβ does not correlate well with clinical symptoms and anti-amyloid pharmaceuticals have failed to improve significantly patient’s symptoms[3,6], even when amyloid deposits are efficiently reduced.[29] As a possible exception, the updated analysis of the EMERGE trial showed a significant reduction in decline of global functions for the patients treated with a high dose of aducanumab.[30]

Thus, it is still not clear if Aβ directly leads to neuronal dysfunction. High levels of Aβ may subsequently lead to a downstream of pathological events, including tau pathology, inflammation, oxidative stress, excitotoxicity, loss of synaptic connections, and cell death, causing the clinical symptoms of AD. Levels of prefibrillar Aβ forms, such as soluble oligomers and protofibrils, correlate better than plaques with disease severity.[31] This may indicate that soluble species are the neurotoxic form of Aβ leading to neurodegeneration.[31]

Aβ deposition is a valid biomarker to support AD diagnostic.[32] Available PET radioligands visualizing Aβ bind to insoluble fibrils, such as Aβ plaques. Recently, several 18F-labeled tracers were designed including flobetapir ([18F]AV-45), flutemetamol ([18F]GE067), florbetaben ([18F] BAY94–9172) [28, 33-36]. We used [18F]flutemetamol PET as a surrogate marker for brain amyloid deposition. Several studies suggested a high correlation between [18F]flutemetamol retention and neuropathologic findings.[37, 38, 39, 40] However, amyloid specific tracers may not be able to provide an accurate measurement of Aβ. In fact, there is a lack of data of in vivo Aβ specificity.[41] Amyloid PET scan is able to rule out an AD diagnosis. Amyloid PET may have an additive, but primarily confirmatory role as a diagnostic marker in patients suspected of early-onset AD. An amyloid-positive PET scan often supports or changes diagnosis into AD.[42] Amyloid pathology is also present in other forms of dementia and interpreted as mixed or copathology and not always the primary cause of the clinical manifestation of dementia.

FDG PET, a neuroimaging tool in AD, plays an important role in discriminating different forms of dementia.[12,34] Decreased glucose metabolism in temporal and parietal cortex indicates synaptic dysfunction and in contrast to amyloid deposition, this occurs mainly in the symptomatic phase of AD.[28] In the present study, we investigated a possible correlation between local amyloid deposition and glucose metabolism in dementia related corresponding brain regions in AD patients. In an early stage of AD impaired glucose metabolism is still restricted to certain areas and not globalized.[43] Our cohort consists mostly of patients at the early stage of disease, which is a result of reasonable referrals. Clinical impact for different diagnosis of dementia in an advanced state of disease is questionable, because all therapeutical strategies are more helpful at an early phase of the disease. The availability of both PET modalities in 30 patients diagnosed for AD in the present study allowed investigating the correlation of amyloid deposition and glucose metabolism, retrospectively.

Scatter plots of the amyloid standardized uptake value (SUV) (on the x- axis) and the 2-[18F] fluoro-2-deoxy-d-glucose (FDG) values represented by NeuroQ (on the y-axis) for the 30 amyloid positive patient for the left side. Each panel represents a brain regions with the name indicated in the strip. Least-squares regression lines with slope proportional to the Pearson correlation coefficient indicate the association. The Pearson correlation (r) and the associated p-value (P) are shown in the label on the top of each panel.

To avoid bias due to computer assisted measurement we manually designated the cortical areas of dementia related brain regions and evaluated the SUVR of [18F]flutemetamol. Compared to previous investigations[17, 18, 19] we included a higher number of AD patients. Engler et al. found a negative correlation with metabolism in parietal cortex in 16 AD patients.[17] Another author showed that a higher amyloid tracer uptake correlated with lower regional glucose metabolism in 19 AD patients.[18] We did not find any negative correlation of amyloid tracer uptake and glucose metabolism in corresponding brain areas in 30 AD patients. Our data supports the concept that amyloid depositions may not be the direct cause of dysfunctional metabolism.

One limitation of this study is that measurement of Aβ40, Aβ 42 and phosphorylated tau in the cerebrospinal fluid (CSF) was not performed[44] and thus, the correlation between CSF amyloid levels and amyloid tracer uptake could not be shown. If Aβ deposition plays a role in the development of cognitive deficits in AD, the lack of direct correlations requires other involved mechanisms such as tau pathology.[2,45,46]

Scatter plots of the amyloid standardized uptake value (SUV) (on the x-axis) and the 2-[18F] fluoro-2-deoxy-d-glucose (FDG) values represented by NeuroQ (on the y-axis) for the 30 amyloid positive patient for the right side. Each panel represents a brain regions with the name indicated in the strip. Least-squares regression lines with slope proportional to the Pearson correlation coefficient indicate the association. The Pearson correlation (r) and the associated p-value (P) are shown in the label on the top of each panel.

The hyperphosphorylation and abnormal aggregation of tau, a microtubule-associated protein essential to neuronal stability and functioning, is a hallmark in AD pathology. Tau imaging revealed that neurofibrillary tangels are mainly located in the hippocampus and associative cortical regions.[47] Tau PET imaging may serve as a valuable and early biomarker for the localization of neuronal injury. In contrast to Aβ accumulation, tau may cause cognitive decline mediated by glucose hypometabolism. Indeed, tau pathology was observed in brain regions related to clinical symptoms and overlapped with areas of hypometabolism.48.49 Exactly what we were not able to show for the relation of amyloid and glucose metabolism. Aβ may play an indirect role in AD pathology in the development of NFTs. It is likely, that Aβ indirectly promotes tau phosphorylation through upregulation of kinases such as GSK-3β and CDK5, which phosphorylate tau.[50]

Conclusios

The focus of the study was to investigate the correlation between local amyloid deposition and glucose metabolism in vivo at corresponding brain areas. Therefore, we manually designated the areas of interest in an early stage of disease, which in that manner has not been performed before. We showed that regional amyloid deposition did not correlate negatively with local glucose metabolism. Our findings support the role of Aβ as a valid biomarker, but does not permit to conclude that Aβ is a direct cause for an aberrant glucose metabolism and neuronal dysfunction. On the contrary our data added a piece of puzzle to the concept that amyloid does not directly cause AD pathology but has to be considered as a prerequisite only.