- Email: [email protected]
In vitro Characterization of MK-3328: A Novel Fluorinated Positron Emission Tomography Tracer for Plaques
S40
Alzheimer’s Imaging Consortium IC-P: Imaging Posters
decline of brain reserve or cognitive reserve and increase of vulnerability to Ab toxicity, and may explain gradual increase of the prevalence or incidence of AD after early sixties. IC-P-098
IN VITRO CHARACTERIZATION OF MK-3328: A NOVEL FLUORINATED POSITRON EMISSION TOMOGRAPHY TRACER FOR BETA-AMYLOID PLAQUES
Cyrille Sur1, Zhizhen Zeng1, Eric Hostetler1, Brett M. Connolly1, Patricia J. Miller1, Stacey O’Malley1, Tsin-Bau Chen1, Christopher Culberson2, Scott Harrison2, Jim Mulhearn2, Scott Wolkenberg2, James Barrow2, Sandra Sanabria1, Jacquelynn J. Cook1, Richard Hargreaves1,3, David L. Williams1, 1Imaging Merck and Co Inc., West Point, PA, USA; 2Medicinal Chemistry Merck and Co Inc., West Point, PA, USA; 3Neuroscience Franchise Merck and Co Inc, West Point, PA, USA. Contact e-mail: [email protected] Background: Amyloid-b (Ab) plaques have been successfully visualized in the brain of Alzheimer’s disease (AD) patients with positron emission tomography (PET) tracers such as [11C]PIB. Yet, the short half-life of carbon-11 (20min) limits its use to clinical facilities close to a cyclotron. In order to support multicenter clinical trials for novel AD therapies, the present work aimed at developing a fluorine-18 (half-life 110min) Ab PET tracer. Methods: In vitro binding studies were performed on cortical homogenates from AD and aged-matched controls with [3H]MK-3328, [3H]PIB and [3H]lazabemide. Autoradiographic mapping of [3H]MK-3328 binding site expression was performed on AD brain sections and compared to the distribution of Ab plaques and astrocytes detected by immunocytochemistry with 6E10 and GFAP antibodies, respectively. Results: Structural similarity searches and extensive medicinal chemistry optimization identified azabenzoxazole compounds that exhibited promising pharmacological properties. The azaindole [3H]MK-3328 binds to cortical Ab plaques with a Kd of 17 6 4 nM (n ¼ 5) and a Bmax of 1600 6 419 nM (n ¼ 5) yielding a Bmax/Kd ratio of 95 6 15 (n ¼ 5). Screening against a diverse set of enzymes and receptors unexpectedly revealed an interaction between MK-3328 and monoamine oxidase B (MAO-B). Studies with purified enzyme indicated that [3H]MK-3328 binds MAO-B with a Kd of 6 6 3 nM (n ¼ 3). Specific [3H]MK-3328 binding was also observed in non-AD human cortical membrane preparations and fully displaced by MAO-B inhibitor lazabemide. Saturation studies in AD cortex demonstrated that MAO-B levels accounted for only 17 6 3 % (n ¼ 5) of Ab levels measured by [3H]PIB. Despite the contribution of MAO-B binding to background signal, autoradiographic experiments with [3H]MK3328 revealed a punctated expression pattern in AD cortical areas comparable to [3H]PIB that was not blocked by lazabemide. Immunocytochemistry on adjacent sections supported a close association between [3H]MK-3328 positive areas, Ab plaques labeled with 6E10 antibody and astrocytes stained with GFAP antibody. Conclusions: Despite unanticipated MAO-B binding, these results indicate that MK-3328 has an in vitro pharmacological profile supporting its development as a fluorine-18 PET tracer for the detection of Ab deposits in AD patients. On going clinical studies will establish the value of this novel imaging agent for the detection of Ab plaques in AD patients. IC-P-099
METRICS FOR RESTING STATE NETWORKS IN MCI AND ALZHEIMER’S DISEASE
Kelvin O. Lim1,2, Laura Hemmy1,2, Bryon A. Mueller1, Chris Bell1, Sue J. Rottunda2, Michael A. Kuskowski1,2, John R. McCarten1,2, 1University of Minnesota, Minneapolis, MN, USA; 2Veterans Affairs Medical Center, Minneapolis, MN, USA. Contact e-mail: [email protected] Background: Resting state fMRI has begun to gain acceptance as a valid approach for assessing brain networks without a task. Graph Theory provides a powerful framework for quantifying complex networks. Recent work used graph theory to analyze resting fMRI in Alzheimer’s patients and healthy controls and found evidence of group differences (Supekar et al., PLOS Computational Biology, 2008). Methods: Because there is no standard method for establishing the correlation threshold to binarize the correlation matrix, we
examined metrics of the correlations between nodes represented in the correlation matrix. We sought to determine if differences between AD and MCI subjects could be detected by the mean of the mean correlation at each node (meannode) and the mean of the variance of the correlation at each node (withnode). Clinical data were obtained from the work-up of 78 subjects with AD (possible or probable) or MCI presenting to a VA Medical Center Memory Loss Clinic. Diagnoses were determined through systematic consensus diagnosis based upon standardized history, neurological exam, neuropsychological evaluation, occupational therapy evaluation, labs, and MRI protocol. Consensus was obtained from a neurologist, geropsychiatrist, internist, and neuropsychologist. Resting fMRI data were collected on a 1.5T system with TR ¼ 2sec for 6 min. Preprocessing included registration and slice-time adjustment. Time courses were extracted from ninety regions using an anatomical atlas and a wavelet analysis was applied to compute frequency dependent correlation matrices for the frequency interval .125-.25 Hz. Results: Subjects were 93% male and 77.83 (6.70 SD) years old. Imaging was performed an average of 59.84 (58.1 SD) days difference from the cognitive outcomes. No diagnostic group differences were found for either correlation matrix metric. However, both the meannode and withnode metric significantly correlated with cognitive performance measures of attention and executive ability (e.g. Stroop, Trail Making). Conclusions: Although network metrics did not differentiate between AD and MCI, they were significantly related to cognition. These metrics may be better indices of processing ability as a general cognitive resource (both basic and more demanding) than either memory or disease-specific severity. Limitations of this preliminary study include the small number of subjects and mostly male sample. IC-P-100
CROSS-SECTIONAL AND LONGITUDINAL STUDY OF AGE-RELATED SHRINKAGE IN THE SUPERIOR, MIDDLE AND INFERIOR FRONTAL GYRUS IN OLD AND VERY OLD AGE
Olof E. Lindberg1, Carl-Henrik Ehrenkrona1, Linnea Engstro¨m1, ¨ hrndahl2, Yi Zhang1, Laura Fratiglioni3, Leif A. Svensson2, Eva O Lars Ba¨ckman3, Sari Karlsson3, Lars-Olof Wahlund1, 1Division of Geriatric Medicine and Department of Neurobiology, Health Care Science and Society, Karolinska University Hospital, Huddinge, Sweden; 2Department of Medical Physics, Karolinska University Hospital, Huddinge, Sweden; 3 Aging Research Centre (ARC), Karolinska Institutet, Stockholm, Sweden. Contact e-mail: [email protected] Background: Cross-sectional and longitudinal structural MRI studies have shown that different parts of the brain are differentially affected by age-related shrinkage (ARS). However, relatively few studies have examined morphological trajectories in old and very old age. We aimed to describe the development of ARS in the superior (SFG), middle (MFG) and inferior frontal gyrus (IFG) in healthy 60 - 96 year old people. Methods: In the cross-sectional study a random sample of 499 persons were examined with structural MRI. Follow-up data were available for 300 subjects. The total volume (white + grey matter) of frontal gyri was obtained by manual segmentation. Cross-sectional ARS was estimated by comparing mean volume of region/intra cranial volume (V/I) for subjects at the age of 60 with the mean V/I for subjects above 86, divided by difference in number of years between young and old subjects. Longitudinal ARS was calculated by dividing the total decrease between time point 1 and time point 2 with the interval between scans for each single subject. Thus longitudinal ARS was the mean atrophy rate for the whole sample. Results: In agreement with previous reports, we found significant ARS in all brain regions in the cross-sectional MRI examinations. The mean atrophy rate in these dorsolateral prefrontal gyri was estimated to 0,78% per year in cross sectional data, whereas the analyses of longitudinal data led to a mean atrophy rate of frontal gyri of 0,66% per year. Conclusions: Cross-sectional data slightly overestimate annual rate of atrophy in the frontal gyri in comparison to the atrophy rate detected with follow-up analysis. The discrepancy between results from cross-sectional and longitudinal analyses may be due to different health status in different birth cohorts of the studied population. If valid, this interpretation might imply that the future generation of old people may have a better preservation of the brain.
Alzheimer’s Imaging Consortium IC-P: Imaging Posters
decline of brain reserve or cognitive reserve and increase of vulnerability to Ab toxicity, and may explain gradual increase of the prevalence or incidence of AD after early sixties. IC-P-098
IN VITRO CHARACTERIZATION OF MK-3328: A NOVEL FLUORINATED POSITRON EMISSION TOMOGRAPHY TRACER FOR BETA-AMYLOID PLAQUES
Cyrille Sur1, Zhizhen Zeng1, Eric Hostetler1, Brett M. Connolly1, Patricia J. Miller1, Stacey O’Malley1, Tsin-Bau Chen1, Christopher Culberson2, Scott Harrison2, Jim Mulhearn2, Scott Wolkenberg2, James Barrow2, Sandra Sanabria1, Jacquelynn J. Cook1, Richard Hargreaves1,3, David L. Williams1, 1Imaging Merck and Co Inc., West Point, PA, USA; 2Medicinal Chemistry Merck and Co Inc., West Point, PA, USA; 3Neuroscience Franchise Merck and Co Inc, West Point, PA, USA. Contact e-mail: [email protected] Background: Amyloid-b (Ab) plaques have been successfully visualized in the brain of Alzheimer’s disease (AD) patients with positron emission tomography (PET) tracers such as [11C]PIB. Yet, the short half-life of carbon-11 (20min) limits its use to clinical facilities close to a cyclotron. In order to support multicenter clinical trials for novel AD therapies, the present work aimed at developing a fluorine-18 (half-life 110min) Ab PET tracer. Methods: In vitro binding studies were performed on cortical homogenates from AD and aged-matched controls with [3H]MK-3328, [3H]PIB and [3H]lazabemide. Autoradiographic mapping of [3H]MK-3328 binding site expression was performed on AD brain sections and compared to the distribution of Ab plaques and astrocytes detected by immunocytochemistry with 6E10 and GFAP antibodies, respectively. Results: Structural similarity searches and extensive medicinal chemistry optimization identified azabenzoxazole compounds that exhibited promising pharmacological properties. The azaindole [3H]MK-3328 binds to cortical Ab plaques with a Kd of 17 6 4 nM (n ¼ 5) and a Bmax of 1600 6 419 nM (n ¼ 5) yielding a Bmax/Kd ratio of 95 6 15 (n ¼ 5). Screening against a diverse set of enzymes and receptors unexpectedly revealed an interaction between MK-3328 and monoamine oxidase B (MAO-B). Studies with purified enzyme indicated that [3H]MK-3328 binds MAO-B with a Kd of 6 6 3 nM (n ¼ 3). Specific [3H]MK-3328 binding was also observed in non-AD human cortical membrane preparations and fully displaced by MAO-B inhibitor lazabemide. Saturation studies in AD cortex demonstrated that MAO-B levels accounted for only 17 6 3 % (n ¼ 5) of Ab levels measured by [3H]PIB. Despite the contribution of MAO-B binding to background signal, autoradiographic experiments with [3H]MK3328 revealed a punctated expression pattern in AD cortical areas comparable to [3H]PIB that was not blocked by lazabemide. Immunocytochemistry on adjacent sections supported a close association between [3H]MK-3328 positive areas, Ab plaques labeled with 6E10 antibody and astrocytes stained with GFAP antibody. Conclusions: Despite unanticipated MAO-B binding, these results indicate that MK-3328 has an in vitro pharmacological profile supporting its development as a fluorine-18 PET tracer for the detection of Ab deposits in AD patients. On going clinical studies will establish the value of this novel imaging agent for the detection of Ab plaques in AD patients. IC-P-099
METRICS FOR RESTING STATE NETWORKS IN MCI AND ALZHEIMER’S DISEASE
Kelvin O. Lim1,2, Laura Hemmy1,2, Bryon A. Mueller1, Chris Bell1, Sue J. Rottunda2, Michael A. Kuskowski1,2, John R. McCarten1,2, 1University of Minnesota, Minneapolis, MN, USA; 2Veterans Affairs Medical Center, Minneapolis, MN, USA. Contact e-mail: [email protected] Background: Resting state fMRI has begun to gain acceptance as a valid approach for assessing brain networks without a task. Graph Theory provides a powerful framework for quantifying complex networks. Recent work used graph theory to analyze resting fMRI in Alzheimer’s patients and healthy controls and found evidence of group differences (Supekar et al., PLOS Computational Biology, 2008). Methods: Because there is no standard method for establishing the correlation threshold to binarize the correlation matrix, we
examined metrics of the correlations between nodes represented in the correlation matrix. We sought to determine if differences between AD and MCI subjects could be detected by the mean of the mean correlation at each node (meannode) and the mean of the variance of the correlation at each node (withnode). Clinical data were obtained from the work-up of 78 subjects with AD (possible or probable) or MCI presenting to a VA Medical Center Memory Loss Clinic. Diagnoses were determined through systematic consensus diagnosis based upon standardized history, neurological exam, neuropsychological evaluation, occupational therapy evaluation, labs, and MRI protocol. Consensus was obtained from a neurologist, geropsychiatrist, internist, and neuropsychologist. Resting fMRI data were collected on a 1.5T system with TR ¼ 2sec for 6 min. Preprocessing included registration and slice-time adjustment. Time courses were extracted from ninety regions using an anatomical atlas and a wavelet analysis was applied to compute frequency dependent correlation matrices for the frequency interval .125-.25 Hz. Results: Subjects were 93% male and 77.83 (6.70 SD) years old. Imaging was performed an average of 59.84 (58.1 SD) days difference from the cognitive outcomes. No diagnostic group differences were found for either correlation matrix metric. However, both the meannode and withnode metric significantly correlated with cognitive performance measures of attention and executive ability (e.g. Stroop, Trail Making). Conclusions: Although network metrics did not differentiate between AD and MCI, they were significantly related to cognition. These metrics may be better indices of processing ability as a general cognitive resource (both basic and more demanding) than either memory or disease-specific severity. Limitations of this preliminary study include the small number of subjects and mostly male sample. IC-P-100
CROSS-SECTIONAL AND LONGITUDINAL STUDY OF AGE-RELATED SHRINKAGE IN THE SUPERIOR, MIDDLE AND INFERIOR FRONTAL GYRUS IN OLD AND VERY OLD AGE
Olof E. Lindberg1, Carl-Henrik Ehrenkrona1, Linnea Engstro¨m1, ¨ hrndahl2, Yi Zhang1, Laura Fratiglioni3, Leif A. Svensson2, Eva O Lars Ba¨ckman3, Sari Karlsson3, Lars-Olof Wahlund1, 1Division of Geriatric Medicine and Department of Neurobiology, Health Care Science and Society, Karolinska University Hospital, Huddinge, Sweden; 2Department of Medical Physics, Karolinska University Hospital, Huddinge, Sweden; 3 Aging Research Centre (ARC), Karolinska Institutet, Stockholm, Sweden. Contact e-mail: [email protected] Background: Cross-sectional and longitudinal structural MRI studies have shown that different parts of the brain are differentially affected by age-related shrinkage (ARS). However, relatively few studies have examined morphological trajectories in old and very old age. We aimed to describe the development of ARS in the superior (SFG), middle (MFG) and inferior frontal gyrus (IFG) in healthy 60 - 96 year old people. Methods: In the cross-sectional study a random sample of 499 persons were examined with structural MRI. Follow-up data were available for 300 subjects. The total volume (white + grey matter) of frontal gyri was obtained by manual segmentation. Cross-sectional ARS was estimated by comparing mean volume of region/intra cranial volume (V/I) for subjects at the age of 60 with the mean V/I for subjects above 86, divided by difference in number of years between young and old subjects. Longitudinal ARS was calculated by dividing the total decrease between time point 1 and time point 2 with the interval between scans for each single subject. Thus longitudinal ARS was the mean atrophy rate for the whole sample. Results: In agreement with previous reports, we found significant ARS in all brain regions in the cross-sectional MRI examinations. The mean atrophy rate in these dorsolateral prefrontal gyri was estimated to 0,78% per year in cross sectional data, whereas the analyses of longitudinal data led to a mean atrophy rate of frontal gyri of 0,66% per year. Conclusions: Cross-sectional data slightly overestimate annual rate of atrophy in the frontal gyri in comparison to the atrophy rate detected with follow-up analysis. The discrepancy between results from cross-sectional and longitudinal analyses may be due to different health status in different birth cohorts of the studied population. If valid, this interpretation might imply that the future generation of old people may have a better preservation of the brain.