Clinicopathological Correlation: Dopamine and Amyloid PET Imaging with Neuropathology in Three Subjects Clinically Diagnosed with Alzheimer’s Disease or Dementia with Lewy Bodies

INTRODUCTION

As the population of the United States ages, there is an increasing burden of neurodegenerative diseases in the elderly. Dementia with Lewy bodies (DLB) is second only to Alzheimer’s disease (AD) as a leading cause of dementia in elderly individuals, and Parkinson’s disease (PD) is overall the second most common neurodegenerative disease [1]. Even though clinical criteria have been established for antemortem diagnosis and for distinguishing these conditions from each other, there remains significant misdiagnosis when compared to gold-standard pathologic confirmation [2, 3], and clinical recognition of DLB is achieved in only a minority of subjects [4]. Imaging biomarkers based on the underlying pathology have been employed to improve diagnostic accuracy for clinical and research purposes. 2-[¹⁸F] fluoro-2-Deoxy-D-glucose (FDG) positron emission tomography (PET) imaging has been in use for more than 25 years to help in the diagnosis of AD by assessing the cerebral metabolic rate of glucose [5], and there are now three PET tracers that are FDA-approved for the demonstration of neuritic plaques [6–8], allowing greatly improved AD diagnostic accuracy [9]. The ability to enrich clinical trials with subjects that have the targeted molecular pathology, and to identify neurodegenerative disorders in the prodromal or preclinical phase, is critical for increasing clinical trial efficiency of potential disease-modifying therapies [10].

Vesicular monoamine transporter 2 (VMAT2) is a nonselective transporter responsible for packaging of dopamine, norepinephrine, and serotonin into monoamine vesicles [11]. There is interest in PET imaging of VMAT2 to identify the dopamine-deficient state. The imaging of VMAT2 has potential advantages over dopamine transporter imaging (DAT) as it is not affected by dopaminergic drugs [11]. AV-133 (florbenazine) is a fluorinated VMAT2 PET ligand designed to measure dopaminergic degeneration by in vivo assessment of striatal VMAT2 density in patients with PD [12]. Besides differentiating PD from essential tremor and drug-induced parkinsonism, PET imaging of VMAT2 may help in differentiating DLB from AD [13]. AV-45 (florbetapir, Amyvid™) is the first PET ligand to receive US FDA approval for the assessment of neuritic plaque amyloid [6].

The current study reports three patients with an antemortem clinical diagnosis of AD or DLB who underwent both AV-45 and AV-133 PET scans to assess the extent of cortical amyloid burden as well as striatal dopaminergic degeneration, respectively (ClinicalTrials.gov identifier: NCT01503944). Recruited subjects included those with clinical diagnoses of AD, DLB, and PD as well as clinically normal subjects. Clinical and imaging results were compared to the neuropathological diagnosis at autopsy.

PATIENTS AND METHODS

All patients provided written informed consent to participate in the AV133-B03 study (ClinicalTrials.gov identifier: NCT00857506), sponsored by Avid Radiopharmaceuticals. Results and methodology from the trial have previously been reported [13]. The study was approved by the Institutional Review Board at Banner Sun Health Research Institute (BSHRI) in Sun City, AZ.

Briefly, subjects were > 50 years and had to be able to tolerate two PET imaging sessions. DLB subjects were enrolled if they met the diagnostic criteria for probable DLB [14] and did not meet the National Institute of Neurological and Communicative Disorders and Stroke and Alzheimer’s disease and Related Disorders Association (NINCDS-ADRDA) criteria for probable AD [15]. AD subjects were enrolled if they met the NINCDS criteria for probable AD, had a Mini-Mental State Examination (MMSE) score at screening between 10 and 24 inclusive and had no symptoms of parkinsonism.

Subjects were examined by neurologists, given a clinical diagnosis, and subsequently had PET imaging with AV-133 and AV-45. Three study sub-jects who had also been enrolled in the Arizona Study of Aging and Neurodegenerative Disorders (AZSAND) [16] were autopsied after death by the Brain and Body Donation Program (http://www.brainandbodydonationprogram.org). The AV-133 and AV-45 PET scans were independently read while blinded to clinical diagnosis by 5 experienced nuclear medicine physicians. The interpretation given by the majority of the five readers for each image was used for analysis. For AV-133, the readers conducted a binary assessment of the presence or absence of dopaminergic degeneration based on the intensity and distribution of uptake in the striatum. Quantitative imaging measures were not used to make the binary reads. The neuropathological findings were recorded blinded to clinical diagnosis and imaging results by a single neuropathologist (T.G.B.). Research, clinical, and neuropathological records were available for review. Final clinical and clinicopathological diagnoses were compared.

Immunohistochemical staining for VMAT2 was performed and the staining intensity quantified on the left anterior caudate nucleus of the imaged subjects as well as in three clinically normal autopsied subjects without parkinsonism, dementia, or postmortem Lewy body disease. Cryostat sections (10μm) of fresh-frozen left anterior caudate nuclei were incubated overnight with 1 : 3,000 dilutions of rabbit VMAT2 antibody (Invitrogen, OTI10C11), followed by biotinylated anti-rabbit IgG and an avidin-biotin-peroxidase complex (Vector), with intervening wash steps, then developed in 3,3’-diaminobenzidine with saturated nickel ammonium sulfate. Three photomicrographs were taken at 400x magnification of each section and then analyzed with ImageJ for area occupied by stained tissue elements exceeding a standard optical density threshold. The means for each case were compared graphically and with an unpaired, 2-tailed t-test.

RESULTS

Case results are summarized in Table 1.

Case 1

A 68-year-old male had symptoms of cognitive impairment and right upper extremity tremor beginning in October 2006, when he scored 27/30 on the MMSE. By March 2007, he declined to 22/30 on the MMSE, and comprehensive neuropsychological testing showed impairment in verbal and visual memory, naming, executive function, semantic fluency, and visuospatial function. In the fall of 2007, he had been diagnosed with parkinsonism and dementia and was being treated with carbidopa-levodopa, with a beneficial response. Acetylcholinesterase inhibitors were started, but he was intolerant of these. His cognitive function then rapidly declined so that by September 2010, his MMSE score was 2/30. He had been having visual hallucinations when seen in September 2009 and a diagnosis of DLB was made. His UPDRS motor score was initially 4 in September 2008 and worsened to 33 in October 2009. He had no history of REM sleep behavior disorder (RBD), and no fluctuations of alertness, cognition, or motor function. The final clinical diagnostic impression was DLB at the time of entry into the AV133-B03 trial and his AV-133 PET scan, done in July 2010, was consistent with bilateral dopaminergic depletion, in both putamen and caudate nucleus, according to 5/5 study readers (Fig. 1a). The AV-45 PET scan in May 2010 was read as positive by all 5 readers. He died of cardiac arrest in May 2012, two years after PET imaging.

OPEN IN VIEWER

Fig. 1

AV-133 PET images for the three studied cases. Case 1 (top) was read as having bilateral dopaminergic depletion in putamen and caudate by all 5 study readers. Case 2 (middle) was read as normal, without dopaminergic depletion in caudate or putamen, by all 5 readers. Case 3 (bottom) was read by all 5 readers as having a decreased signal in the left putamen, while 3/5 noted a decrease in right putamen and one indicated the decreases were asymmetrical, left > right; all 5 readers reported no decrease in left or right caudate nuclei.

General autopsy findings were notable for coronary atherosclerosis and clinically unsuspected prostatic adenocarcinoma, confined to the prostate. The brain weight was 1,302 grams. Gross neuropathological findings included moderate depigmentation of the substantia nigra and mild atrophy of the hippocampus and parahippocampal gyrus. Microscopic examination with hematoxylin and eosin (H&E) showed moderate to severe depletion of pigmented neurons from the substantia nigra and locus coeruleus, with Lewy bodies within remaining neurons in both regions. Immunohistochemical staining for phosphorylated α-synuclein showed frequent immunoreactive neuronal perikaryal inclusions and fibers in the olfactory bulb, brainstem, amygdala, and throughout the cerebral cortex, meeting Unified Staging System for Lewy Body Disorders (USSLB) stage IV (neocortical) [17]. Gallyas, Campbell-Switzer, and thioflavin-S staining showed frequent diffuse and neuritic neocortical amyloid plaques, Braak neurofibrillary stage V neurofibrillary change and Thal amyloid phase V, A3B3C3, “high” Alzheimer’s disease neuropathological change (ADNC) [18]. The final clinicopathological diagnoses were DLB and AD.

VMAT2 immunohistochemical staining

Sections of left anterior caudate nucleus from the three AV-133-imaged cases and three control subjects all showed dense punctate staining, consistent with dopaminergic presynaptic terminals originating from the substantia nigra (Fig. 2). Quantification of the area occupied by stained tissue elements showed decreased areas for the AV-133-imaged subjects (Fig. 3A, C, E) with diagnoses of AD/DLB (Cases 1 and 2; 38.8%and 13.8%area occupied, respectively) and MSA (Case 3; 60.6%area occupied), as compared to the three control subjects (Fig. 3B, D, F; mean 70.9%SD 6.4%.

OPEN IN VIEWER

Fig. 2

Sections stained with an immunohistochemical method for VMAT2, showing the left anterior caudate nucleus from the three AV-133-imaged subjects (A, C, E) and three control subjects (B, D, F) without clinical dementia or parkinsonism. Scale bar (F) is 50μm and serves for A–F.

OPEN IN VIEWER

Fig. 3

Quantification of staining intensity in caudate nucleus sections stained with an immunohistochemical method for VMAT2, for the three AV-133-imaged subjects and three control subjects (B, D, F) without clinical dementia or parkinsonism.

DISCUSSION

These three cases, although insufficient as a statistical sample, demonstrate that functional imaging of dopaminergic neuronal systems with AV-133 holds promise for the clinical identification of DLB cases, although borderline cases may present difficulties. As in prior reports on DLB subjects with combined amyloid PET and dopaminergic imaging [19], all three subjects had increased cortical signal for AV-45, consistent with the presence of amyloid neuritic plaques, and neuropathological examination confirmed this, with all three subjects meeting intermediate or high ADNC criteria [16]. Clinicopathological concordance for AV-133 was not unanimous, however.

The first subject met clinical and neuropathological diagnostic criteria for DLB [20] and AV-133 PET imaging was read as having decreased striatal uptake consistent with dopaminergic degeneration. As the final clinicopathological diagnosis was also DLB, PET imaging confirmed the clinical diagnosis in this case.

At the time of PET imaging, the second subject was given a clinical diagnosis of AD without DLB. PET AV-45 imaging was read as positive for neocortical amyloid but AV-133 PET imaging was read by all 5 readers as being negative for dopaminergic degeneration. Subsequently, this patient developed rest tremor, visual hallucinations, and movement difficulties, and was diagnosed with DLB. The neuropathological diagnosis was AD and DLB. The AV-133 PET scan was therefore not consistent with the final clinical or neuropathological diagnoses.

The failure of some DLB subjects to show reduced striatal dopaminergic functional imaging has been previously reported by others and the underlying reason for the discrepancy may be due to lesser degrees of loss of nigrostriatal dopaminergic neuronal and nerve terminals in DLB as compared to PD [21–25]. The complete AV-133 clinical trial results, as previously reported, found that 3 of 11 subjects clinically diagnosed with DLB had a binary normal qualitative read and normal levels of striatal AV-133 uptake [13]. For our Case 2, 30 months had elapsed between AV-133 imaging and death, during which time dopaminergic degeneration would presumably have advanced, so that repeat AV-133 imaging more proximal to death may have been positive. Another possibility for these unexpected results may be due to persisting noradrenergic or serotonergic striatal innervation in some patients, as VMAT2 serves these neurotransmitter systems as well. The clinical failure to detect DLB is a well-known deficiency of current DLB diagnostic criteria, as up to 80%of autopsy-diagnosed DLB subjects are never diagnosed during life [4, 26]. It is possible that AV-133 could increase diagnostic sensitivity for DLB but the potential for this would require a larger autopsy-confirmed study.

The third subject met clinical criteria for probable DLB [18] and was found to have evidence of dopaminergic neuronal loss and amyloid deposition on PET imaging. On neuropathological examination, she was found to have pathognomonic features of MSA and AD. The concomitant pathology of MSA and AD is rare [27]. The cognitive deficits and parkinsonism seen in this case were likely due to a combination of these two separate pathologies, and PET imaging showed two different pathologies (AD and dopaminergic depletion) as well. Aside from PD and DLB, other disorders with striatal dopaminergic depletion include MSA, corticobasal degeneration (CBD), and progressive supranuclear palsy (PSP). Dopaminergic functional imaging identifies nigrostriatal degeneration but not the specific cause [28]. This subject had markedly asymmetric AV-133 striatal uptake that was matched by autopsy findings of markedly asymmetric striatal atrophy. A prior autopsy-confirmed study has reported increased striatal asymmetry of dopamine transporter binding in MSA as compared to PD, but this was not sufficiently distinct to provide a reliable differentiation of individual subjects [29].

This small clinicopathological series, although insufficient to be statistically evaluated, and subject to the selection bias that is unavoidable with autopsy studies, shows that AV-133 imaging may be useful for detecting striatal dopaminergic degeneration but may not be sensitive in subjects with early or mild loss and may not be able to distinguish between DLB and MSA. Prior published autopsy-confirmed work indicates that dopaminergic functional imaging ligands of other types cannot distinguish PD from DLB, PSP, CBD, and MSA [26]. The presence of amyloid pathology on PET imaging does not necessarily reflect clinical disease as positivity can be seen in cognitively normal individuals as well [10]. However, in the setting of dementia, a positive AV-45 scan is likely to be a highly sensitive and specific indicator that AD is at least one cause [9]. Since most autopsied cases of DLB concurrently meet neuropathological criteria for intermediate or high ADNC, but most subjects with DLB are clinically unrecognized, amyloid imaging in conjunction with a dopaminergic functional scan would likely increase both sensitivity and specificity for the dual diagnosis, since other causes of striatal dopaminergic depletion are either reliably separated by clinical criteria (e.g., PD versus DLB) or are much less prevalent, in neuropathological series, as concurrent diagnoses with AD (i.e., AD with DLB or PD is much more common than AD with PSP, CBD, or MSA).

The immunohistochemical comparison of VMAT2 staining in the AV-133-imaged subjects and three controls all showed decreased staining density in the diseased subjects as compared to the three controls, supporting 2 of 3 of the visual reads made by the clinical trial readers. The one discrepant read, as compared to the postmortem staining, may have been due to this subject having had the longest time interval between imaging and death, allowing progression of striatal dopaminergic terminal loss from non-apparent on PET imaging to obvious on immunohistochemical stains.

In summary, the findings in these three subjects show that PET imaging with AV-133 and AV-45 may be helpful in identifying underlying amyloid plaques and striatal dopaminergic degeneration co-existing in the same individual.