Assistant Professor of Neuroscience and Neurology, Feil Family Brain and Mind Research Institute,
Department of Neurology, Weill Cornell Medical College

Dr. Ishii reports no financial relationships relevant to this field of study.

SYNOPSIS: In a study cohort from the Alzheimer’s Disease Neuroimaging Initiative, cerebrospinal fluid levels of cholecystokinin were associated with better outcomes that may reflect compensatory protection as Alzheimer’s disease pathology progresses. However, because of significant study limitations, these findings need to be validated in additional studies.

SOURCE: Plagman A, Hoscheidt S, McLimans KE, et al. Cholecystokinin and Alzheimer’s disease: A biomarker of metabolic function, neural integrity, and cognitive performance. Neurobiol Aging 2019;76:201-207.

Weight loss is a common clinical manifestation of Alzheimer’s disease (AD), but the underlying mechanisms have not been entirely elucidated. Cholecystokinin (CCK) is a gut hormone that can suppress appetite. CCK also is an abundant neuropeptide with putative roles in regulating hippocampal and memory function. Thus, CCK is well positioned to be involved in AD pathogenesis; however, little is known about the role of CCK in AD. Therefore, Plagman et al set out to determine whether cerebrospinal fluid (CSF) levels of CCK were associated with the onset and severity of AD and if CSF CCK levels were related to changes in cognition, neuroimaging, and AD biomarkers such as amyloid-beta and tau.

This study was a retrospective analysis of cross-sectional data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Baseline CSF data were available from 287 subjects: 86 cognitively normal, 135 with mild cognitive impairment (MCI), and 66 with AD. MCI and AD were classified based on clinical criteria. The levels of various proteins were measured in the CSF by a proteomics approach using Multiple Reaction Monitoring Mass Spectrometry, which targeted quantitation of 567 peptides representing 221 proteins in a single run. For this study, the CSF proteomics panel referring to the peptide AHLGALLAR was chosen to represent CCK levels.

In the study population, there was no significant difference among the three groups in years of education, APOE E4 carrier percentage, or age. As expected, cognitive function was significantly worse for the MCI and AD groups compared to the cognitively normal group. CSF CCK levels were significantly lower in the AD group compared to the MCI or cognitively normal groups, but there was substantial overlap among all groups. A logistic regression analysis found that higher CSF CCK expression levels predicted a decreased likelihood of MCI or AD with a per ng/mL increase in CSF CCK corresponding to roughly 65% less likelihood of being diagnosed with AD compared to cognitively normal or MCI. However, there was no relation between CSF CCK levels and increased risk when comparing cognitively normal vs. MCI, cognitively normal vs. AD, or MCI vs. AD individually. Among the MCI subjects, a per ng/mL increase in CCK was related to a 61.7% less likelihood of progressing to AD.

Regarding AD CSF biomarkers, CSF CCK levels were not associated with CSF amyloid-beta42 levels, but higher CSF CCK levels were associated with higher CSF total tau (β ± SE = 37.857 ± 4.799; F = 62.237; P < 0.001) and CSF p-tau-181 levels (β ± SE = 10.046 ± 1.630; F = 37.992; P < 0.001). Higher CSF CCK levels were also related to better global cognition score, memory function factor scores, and executive function factor scores. Furthermore, CSF tau and p-tau-181 mediated associations with CSF CCK for several of these cognitive scores. Finally, using a voxelwise analysis of T1-weighted MRI data, higher CSF CCK was significantly associated with greater grey matter volume in several large clusters, including the cingulate cortex and parahippocampal gyrus. However, there was no significant association between CSF CCK levels and FDG-PET glucose uptake.

COMMENTARY

Based on these findings, the authors proposed CSF CCK as a biomarker that is associated with metabolic function, neural integrity, and cognitive performance in AD. However, there are significant limitations that need to be addressed. First, the methods used to measure CCK levels needs to be rigorously validated. Plagman et al relied on the measurement of a small peptide fragment by mass spectrometry and assumed this is equivalent to measuring the full-length CCK protein. Based on the data provided, there is no way to distinguish if degraded protein fragments or biologically active full-length proteins are measured. Even if the measurement of CCK in the CSF was accurate, it is not known if CSF levels reflect physiological levels of CCK in the brain parenchyma. Second, this study provides no direct evidence to support the role of CCK in regulating metabolic function in AD. If CSF CCK reflected CCK’s role as a satiety hormone, there should be a significant association of CSF levels of CCK with body weight, eating behavior, or other measures of metabolic function. Third, the study presents intriguing data that higher CSF levels of CCK are associated with better outcomes, suggesting a protective effect for CCK. However, there is a discrepancy between this possible protective role of CCK and the association of higher CSF CCK levels with higher CSF tau and p-tau181 levels, which may reflect an association between higher CSF CCK levels and increasing neuronal injury. The authors postulated that as AD progresses, CCK actually is increased to protect the brain from any further injury, which is mitigated by progressive neurodegeneration. Although this is plausible, additional evidence is needed to support this claim, such as by conducting longitudinal studies measuring CCK levels as AD progresses or mechanistic studies in model systems. Regardless of the exact role of CCK in AD, this study does provide evidence to support the use of CSF CCK as a potential biomarker of cognitive impairment and volumetric loss. However, because of the significant limitations and concerns from this study, additional well-designed studies are clearly needed to validate these findings.