Influence of Cognitive Status, Age, and APOE-4 Genetic Risk on Brain FDDNP Positron-Emission Tomography Imaging in Persons Without Dementia FREE

Neurodegeneration associated with aging progresses along a continuum,¹ but it has been categorized according to the degree of cognitive impairment. In normal aging, mild memory concerns with minimal objective cognitive deficits have been observed in nearly half of people by 50 years of age.² Such awareness of memory changes is usually stable and not a risk factor for future cognitive decline.³ Mild cognitive impairment (MCI) is a more advanced form of age-related cognitive decline in which people notice memory changes and neuropsychological tests often confirm problems with delayed recall, although non–memory-related cognitive domains may also be impaired.⁴ People experiencing this transitional state between normal aging and dementia are still able to live independently, but they have an increased risk for developing dementia. A recent study that followed patients with MCI for 30 months reported conversion rates to Alzheimer disease (AD) ranging from 27% to 49%, depending on the subtype of MCI.⁵ The prevalence of MCI may be as high as 19% in people older than 65 years and 29% in those older than 85 years.⁶ When cognitive decline interferes with daily functioning and impairs not just memory but other mental abilities, dementia is often diagnosed.⁷ Alzheimer disease, which accounts for most cases, is insidious in its onset and progressive in its course.⁸ The prevalence of AD in individuals aged 71 years and older approaches 10%,⁹ and by 85 years has been reported to be as high as 50%.¹⁰

Age is the strongest known risk factor for AD. The estimated annual incidence of AD in a community-based sample ranged from 0.6% for people aged 65 to 69 years to 8.4% for those 85 years and older.¹¹ In a meta-analysis of 23 studies,¹² the incidence of AD increased exponentially with age until 90 years.

In addition to cognitive status and age, many genetic and nongenetic¹³ factors contribute to the risk for developing AD. For the common forms of late-onset AD, the major genetic risk is associated with the apolipoprotein E (APOE) gene on chromosome 19, which has 3 allelic variants (2, 3, and 4) and 5 common genotypes (2/3, 3/3, 2/4, 3/4, and 4/4). The APOE-4 allele increases risk and decreases the average age of dementia onset in a dose-related fashion (ie, AD risk is lowest for the 3/3 genotype, higher for the 3/4 genotype, and highest for the 4/4 genotype),¹⁴ while APOE-2 lowers the risk.¹⁵

Because APOE-4 accounts for only part of the genetic risk for AD, family history of dementia, regardless of whether an individual is an APOE-4 carrier, may increase the risk for developing AD. People with a first-degree relative with dementia have a 10% to 30% increased risk of developing the disorder,¹⁶ although a recent investigation reported that family history of dementia was associated with increased risk of dementia and AD only in APOE-4 carriers.¹⁷

Less education also appears to increase the risk for AD.¹⁸ Although this association suggests that the cognitive stimulation resulting from higher education protects the brain,¹⁹ other factors may explain the higher risk of dementia in people with less education including unhealthy lifestyles, lower cognitive reserve,²⁰ and higher prevalence of small vascular lesions.²¹

In patients with AD, 2 proteins, β-amyloid (in senile plaques) and tau (in neurofibrillary tangles), accumulate abnormally in the brain in a predictable spatial pattern; however, these proteins also appear to accumulate before people develop dementia and increase gradually with age.^{22 - 23} For a definitive diagnosis of AD, high cerebral concentrations of amyloid senile plaques and tau neurofibrillary tangles must be present in the brain at autopsy.⁸

Newly developed small molecule tracers used in conjunction with positron emission tomography (PET) have made it possible to obtain measures of these abnormal protein deposits in living people.^{24 - 26} For example, the amyloid-binding radiotracer Pittsburgh Compound B has been used with PET imaging to demonstrate significantly greater cortical Pittsburgh Compound B retention in patients with AD vs controls.²⁴

Our group has developed a small molecule, 2-(1-{6-[(2-[F-18]fluoroethyl)(methyl)amino]-2-naphthyl} ethylidene)malononitrile (FDDNP), for use as an in vivo chemical marker of cerebral amyloid and tau proteins.²⁶ This molecule and its parent compound, DDNP, are both fluorescent and may be used with confocal fluorescence microscopy to clearly visualize plaques and tangles in vitro in brain specimens of patients with AD.²⁷ Initial studies have found that global FDDNP-PET binding (average of temporal, parietal, posterior cingulate, and frontal regions) in MCI is intermediate between controls with normal cognition and patients with AD, and that subjects who progress clinically over time show corresponding increases in FDDNP signal.²⁸ Moreover, 3-dimensional cortical surface projection images of FDDNP-PET show patterns remarkably similar to those expected from autopsy studies demonstrating regional brain accumulation patterns of plaques and tangles.^{1 ,22 ,28}

In this study, we determined whether previously reported risk factors for developing a clinical diagnosis of AD were also associated with plaque and tangle accumulation as measured with FDDNP-PET binding in volunteers without dementia. We hypothesized that several established risk factors, including impaired cognitive status, older age, APOE-4 genetic risk for AD, family history of dementia, and lower educational achievement would be associated with increased regional cerebral FDDNP-PET binding.

We performed neuropsychiatric evaluations, cognitive assessments, and PET scanning on 63 volunteers without dementia from a larger longitudinal study.²⁸ Subjects were recruited through study advertisements regarding mild memory concerns, media coverage, and referrals from physicians and families. Although all subjects had mild memory concerns, patients with any form of dementia were excluded. Also excluded were individuals taking medications that might affect cognition, such as sedatives, or those taking nonsteroidal antiinflammatory drugs, which bind to amyloid plaques and may affect FDDNP binding values.²⁷

Subjects had screening laboratory tests and structural imaging scans (3-dimensional magnetic resonance imaging [MRI] or computed tomography [CT]) to rule out other causes of cognitive impairment (eg, stroke, tumor)⁸ and for coregistration with PET scans for region-of-interest image analyses. Computed tomography scans instead of MRI were performed on 4 subjects because they could not tolerate MRI (eg, owing to claustrophobia, metal in body). Subjects with vascular lesions on the MRI or CT scan were excluded from the study. In addition to the Mini-Mental State Examination²⁹ and Hamilton Rating Scale for Depression,³⁰ a neuropsychological test battery³¹ was administered to assess 5 cognitive domains: (1) memory, including logical memory, selective reminding, and complex figure recall; (2) language, including Boston naming and letter and category fluency; (3) attention and information-processing speed, including Trail Making A, Stroop Color (Kaplan version), and Digit Symbol; (4) executive functioning, including Trail Making B, Stroop Interference (Kaplan version), Wisconsin card sort, and perseverative errors; and (5) visuospatial, including block design, complex figure copy, and visual retention.

To diagnose MCI, we used standard diagnostic criteria for amnestic MCI (ie, memory impairment without other cognitive impairments), which include (1) patient awareness of a memory problem, preferably confirmed by another person; (2) memory impairment detected with standard assessment tests; and (3) ability to perform normal daily activities.⁴ For a broad definition of MCI, we also used guidelines to identify subjects with other MCI subtypes, including those with memory impairment and additional cognitive deficits.³² The diagnosis was corroborated by clinical judgment⁴ and included subjects with MCI who scored 1 SD or more less than age-corrected norms, as this threshold for impairment yields high sensitivity for predicting dementia.³³ To balance increased sensitivity with specificity, we required impairment on at least 2 neuropsychological tests within 1 of the 5 cognitive domains.³⁴ Subjects in the MCI group did not meet diagnostic criteria for AD,^{7 - 8} and the presence of memory concerns was documented using a standardized subjective memory instrument (Memory Functioning Questionnaire)³⁵ and clinical interview.

Volunteers with 1 or more first-degree relatives (ie, sibling or parent) with AD or dementia were classified as having a positive family history of dementia. Prior educational achievement was quantified according to years and months completed, beginning with elementary school (ie, first grade).

All clinical assessments were performed within 4 weeks of scanning procedures, and clinicians were blinded to the results of FDDNP-PET scans. Written informed consent was obtained in accordance with the University of California, Los Angeles Human Subjects Protection Committee procedures. Cumulative radiation dosimetry for all scans was below the mandated maximum annual dose and in compliance with state and federal regulations. Two minor adverse events occurred during PET scanning: one subject developed minor bruises at venipuncture sites, and another subject experienced a transient headache.

All DNA was obtained from blood samples. The APOE genotypes were determined using standard techniques as previously described.¹⁴ Genetic data were available for 72 subjects.

As previously described, FDDNP was prepared at very high specific activities (>37 gigabecquerel [GBq]/μmol).^{26 ,28} All scans were performed with the ECAT HR or EXACT HR+ tomograph (Siemens-CTI, Knoxville, Tennessee) with subjects supine and with the imaging plane parallel to the orbitomeatal line. A bolus of FDDNP (320-550 megabecquerel [MBq]) was injected via an indwelling venous catheter, and consecutive dynamic PET scans were performed for 2 hours. Scans were decay corrected and reconstructed using filtered back-projection (Hann filter, 5.5 mm full-width at half-maximum) with scatter and measured attenuation correction. The resulting images contained 47 contiguous slices with plane separation of 3.37 mm (ECAT HR) or 63 contiguous slices with plane separation of 2.42 mm (EXACT HR+). Results did not differ significantly according to the scanner used.

The FDDNP binding data were quantified using Logan graphical analysis with the cerebellum as the reference region for time points between 60 and 125 minutes.^{28 ,36} The slope of the linear portion of the Logan plot is the relative distribution volume, which is equal to the distribution volume of the tracer in a region of interest divided by that in the reference region. The relative distribution volume parametric images were generated and analyzed using regions of interest traced on the coregistered MRI or CT scans for left and right parietal, medial temporal (limbic regions including hippocampus, parahippocampal areas, and entorhinal cortex), lateral temporal, posterior cingulate, and frontal regions, as previously described.²⁸ Each regional relative distribution volume or binding value was expressed as an average of left and right regions. Rules for region-of-interest drawing were based on the identification of gyral and sulcal landmarks with respect to the atlas of Talairach and Tournoux.³⁷ All PET scans were read and regions of interest drawn by individuals who were blinded to clinical assessments and genotype. Repeat scans performed on the same 2 subjects within several weeks indicated stability of these measures (≤3% SD of regional values).

Data were screened for outliers and normality assumptions. Descriptive statistics were computed for the entire sample and for the MCI and control subjects separately. The t test was used to compare the continuous variables of cognitive groups and χ² tests were used for categorical variables. General linear models were used to determine which variables—age, APOE-4 status, family history, education, and cognitive status (MCI vs normal)—were associated with regional FDDNP binding in the entire sample. We first included all of these risk factors as predictors in the general linear model. We then computed the final model by deleting the risk factors that did not contribute significantly. All tests were 2-tailed, and a significance level of P = .05 was used for all inferences.

Subjects were middle-aged or older (range, 47-87 years; mean [SD] age, 66.8 [10.7] years) and educated (mean [SD] education, 17.0 [2.9] years). They showed minimal impairment on cognitive testing (mean [SD] Mini-Mental State Exam scores, 28.7 [1.5]), and 44 (59%) had a family history of dementia in at least 1 first-degree relative. Of the 36 patients with MCI, 17 showed memory impairment consistent with amnestic MCI and 19 had amnestic MCI plus deficits in other cognitive areas. The memory symptoms of all other subjects were normal for their age. Of the 72 subjects with genetic data, 34 (47%) were APOE-4 carriers (Table). Of these, 4 subjects were homozygotes (4/4 genotype).

For all regions of interest studied, cognitive status (ie, diagnosis of MCI vs normal aging) was associated with increased FDDNP binding (medial temporal F_1,67 = 6.04, P < .02; lateral temporal F_1,68 = 8.34, P < .005; parietal F_1,70 = 7.46, P < .008; posterior cingulate F_1,70 = 3.83, P < .05; and frontal F_1,68 = 9.27, P < .003). Older age was associated with increased lateral temporal FDDNP binding (F_1,68 = 5.41, P < .02). Subjects' APOE-4 status was associated with higher frontal FDDNP binding; APOE-4 carriers showed more binding than noncarriers (F_1,68 = 3.93, P < .05).

In the MCI group, older age was associated with increased medial temporal (F_1,29 = 4.08, P < .05) (Figure) and lateral temporal (F_1,34 = 6.73, P < .01) FDDNP binding, and APOE-4 carriers had more medial temporal FDDNP binding than noncarriers (F_1,29 = 5.22, P < .03). In the group without MCI, APOE-4 carriers had more frontal FDDNP binding than noncarriers (F_1,37 = 4.40, P < .04).

Family history of dementia and years of educational achievement were not associated with increased FDDNP binding values. No associations between Hamilton Depression Scale scores and FDDNP binding were found.

These findings indicate that impaired cognitive performance, older age, and APOE-4 genetic risk for AD are associated with increased brain FDDNP-PET binding in persons without dementia. Moreover, the degree of cognitive impairment (ie, normal aging vs MCI) appears to influence the interactions among risk factors. For example, in the MCI group, APOE-4 carriers show higher medial temporal FDDNP binding, whereas in normal aging, APOE-4 carriers demonstrate higher frontal binding. Overall, the results are consistent with our hypotheses and with previous clinical and postmortem studies demonstrating a relationship between such risk factors and amyloid plaque and tau tangle formation in the brain. By contrast, the other risk factors we tested, family history of dementia and prior years of education, were not found to be associated with higher FDDNP binding values.

Brain deposition of plaques and tangles follows a pattern in which tau tangles accumulate initially in the entorhinal cortex in normal aging and then spread to medial temporal regions as MCI develops; concentrations of medial temporal tangles become intermediate between those of normal aging and AD.^{22 - 23 ,38} The finding in our study that APOE-4 status was associated with FDDNP binding in the medial temporal region of patients with MCI is interesting in light of autopsy studies showing that this region is among the earliest to demonstrate increased plaque and tangle accumulation.^{22 - 23 ,38} Also, neuritic and diffuse plaques and tangles in patients with MCI are widely distributed throughout the neocortex and limbic structures.³⁸ This spatial pattern and progression of abnormal protein accumulation may be consistent with an interaction between plaque and tangle accumulation. At some critical point in neurodegeneration, β-amyloid peptides may accelerate age-related tangle accumulation, which would otherwise progress relatively slowly with age.²³ Tangle load has been associated with cognitive decline in older individuals, but plaque load has not consistently demonstrated such an association.³⁸ The findings that FDDNP binds both plaques and tangles, particularly in the medial temporal lobe, may explain, in part, the association between higher FDDNP binding values and impaired cognitive function. Moreover, the regional pattern of FDDNP binding appears consistent with plaque and tangle accumulation patterns observed in autopsy studies.^{22 - 23 ,28 ,38}

This is the first study to explore and demonstrate that a genetic risk for AD is associated with increased FDDNP-PET binding in persons without dementia. These results are consistent with previous neuropathological studies demonstrating increased plaque and tangle formation in middle-aged and older APOE-4 carriers without dementia.^{39 - 40} For example, in a study of persons without dementia who died between the ages of 50 and 93 years, APOE-4 carriers showed a premature appearance of β-amyloid and neurofibrillary tangles.⁴⁰ By contrast, autopsy studies of patients with AD find that APOE-4 heterozygotes do not show increased plaque and tangle accumulation, whereas APOE-4 homozygotes do show increased accumulation.⁴¹ Thus, the effect of the APOE-4 allele on cerebral plaque and tangle formation may only occur early in the course of neurodegeneration.

Clearly APOE-4 lowers the age of clinical dementia onset, but surprisingly, several studies do not demonstrate acceleration of clinical progression of the disease in APOE-4 carriers.^{42 - 46} Consistent with such findings, APOE-4 has been reported to accelerate transitions from normal aging to MCI, but not from MCI to dementia.^{46 - 49} While controversial, these results suggest that APOE-4 may have a larger effect on a central precipitating event like amyloid plaque deposition, arguably a poor correlate of clinical progression or initial pathology in medial temporal regions, as observed in this study with FDDNP.

Previous autopsy studies of individuals without dementia ranging from young adults to elderly persons also have demonstrated that plaque and tangle formation is age-related.^{22 ,50} Other research has demonstrated interactions among these various risk factors. For example, APOE-4 carrier status may lead to increased tangle accumulation in relatively young age groups. In an autopsy study of asymptomatic younger adults (mean age, 38 years), tangle formation was significantly greater in APOE-4 carriers compared with controls.⁵¹ Sex may also modify the effect of APOE-4 on the deposition of AD brain pathology; in a study of 729 brains examined by routine autopsy, an association between the APOE-4 allele and plaques was found only for women aged between 60 and 79 years, whereas the association was found for men in all age groups.⁵² In the present study, we did not find sex to be associated with greater FDDNP binding.

Subjective memory concerns and minimal decline in memory ability compared with young adults are expected with normal aging.² Although cognitive impairment is a risk factor for dementia, it is also a consequence of the brain lesions causing AD. The results of this study suggest that in vivo measures of plaques and tangles are associated with increased cognitive impairment, but other factors besides plaques and tangles can contribute to cognitive impairment including cerebrovascular disease and head trauma.^{53 - 54}

Revised research criteria for the diagnosis of AD have been proposed.⁵⁵ These criteria include the presence of early episodic memory impairment along with 1 or more abnormal biomarker such as molecular neuroimaging with PET or cerebrospinal fluid analysis of β-amyloid or tau proteins. Our findings that FDDNP binding patterns differ according to the degree of cognitive impairment (ie, normal aging vs MCI) suggest that FDDNP-PET might be a useful tool in applying such revised research diagnostic criteria. Additional studies clarifying the patterns of FDDNP binding and other molecular imaging techniques in AD, MCI, and normal aging will likely have an effect on the use of such diagnostic criteria.

Family history of dementia was not associated with higher FDDNP binding values. Previous studies have found that family history of dementia increases the risk for neurodegeneration⁵⁶ and is associated with subsequent cognitive decline⁵⁷ and lower scores on neuropsychological testing.⁵⁸ Family history of dementia is an established risk factor for AD,⁵⁹ but not all studies have confirmed such a risk.^{60 - 61} Moreover, the effect of family history on risk for dementia may be age-dependent—some studies have found the effect in persons older than 75 years,⁶² while other reports suggest that the effect of familial or genetic factors on dementia risk diminishes with increasing age.¹⁷ Misclassification in the assessment of dementia history and cohort effects (ie, relatives may be more likely to report dementia in siblings than in parents) may also diminish the accuracy of family history estimates. The relatively small sample size also may explain why family history was not associated with increased FDDNP binding values.

This small sample also may explain why we did not find prior educational achievement to influence our results. In addition, the lack of variance in years of education in these subjects may have minimized any effect of education in the present analysis. Other methodological issues could have influenced these results as well, including partial volume effects⁶³ and use of a relatively educated sample who may not be representative of the general population.

Despite such limitations, these results, that FDDNP-PET may be an informative biological marker for people at risk for dementia, are encouraging. An important potential application of emerging technologies such as FDDNP-PET is in early detection of neurodegeneration. Our finding that greater FDDNP binding is associated with increased cognitive impairment in individuals without dementia suggests that this approach might be useful in detecting people at risk for dementia, which would also be useful for identifying candidates for clinical trials of prevention treatments. These results suggest that in future clinical trials using FDDNP-PET, stratifying subject groups according to APOE-4 carrier status, age, and cognitive status may be an informative strategy.

Correspondence: Gary W. Small, MD, Semel Institute, 760 Westwood Plaza, Ste 88-201, Los Angeles, CA 90024 (gsmall@mednet.ucla.edu).

Submitted for Publication: January 16, 2008; final revision received June 21, 2008; accepted June 23, 2008.

Financial Disclosure: The University of California, Los Angeles, owns a US patent (6,274,119) titled “Methods for Labeling β-Amyloid Plaques and Neurofibrillary Tangles” that uses the approach outlined in this article and has been licensed to Siemens. Drs Small, Huang, Cole, and Barrio reported that they are among the inventors, have received royalties, and will receive royalties on future sales. Dr Small reports having served as a consultant and/or having received lecture fees from Abbott, Brainstorming Co, Dakim, Eisai, Forest, Myriad Genetics, Novartis, Ortho-McNeil, Pfizer, Radica, Siemens, and VerusMed, as well as having received stock options from Dakim. Dr Lavretsky reports having received lecture fees from Eisai, Janssen, and Pfizer and having received a grant from Forest. Dr Huang reports having received lecture fees from GlaxoSmithKline. Dr Barrio reports having served as a consultant and having received lecture fees from Nihon Medi-Physics Co, Bristol-Meyer Squibb, PETNet Pharmaceuticals, and Siemens. Drs Ercoli, Siddarth, Burggren, Kepe, Miller, Thompson, Phelps, and Bookheimer have no financial conflicts of interest.

Funding/Support: This study was supported by grants P01-AG024831, AG13308, P50 AG 16570, MH/AG58156, MH52453, AG10123, and M01-RR00865 (General Clinical Research Centers Program) from the National Institutes of Health; contract DE-FC03-87-ER60615 from the Department of Energy; the Rotary CART Fund; the Fran and Ray Stark Foundation Fund for Alzheimer's Disease Research; the Ahmanson Foundation; the Larry L. Hillblom Foundation; the Lovelace Foundation; the Judith Olenick Elgart Fund for Research on Brain Aging; the John D. French Foundation for Alzheimer's Research; and the Tamkin Foundation. No company provided support of any kind for this study.

Previous Presentations: Presented at the Annual Scientific Meeting of the American College of Neuropsychopharmacology; December, 2007; Boca Raton, Florida.

Additional Contributions: The authors also thank Andrea Kaplan, MD, Deborah Dorsey, RN, Gwendolyn Byrd, MD, and Teresann Crowe-Lear, MD, for help in subject recruitment, data management, and study coordination, and Gerald Timbol and Anasheh Halabi for help in image processing.