Abstract
dc:description.abstractThere is a pressing need to facilitate and accelerate new treatments for Alzheimer’s disease. An improved mechanistic understanding of the human disease will facilitate new treatments. Human pathophysiological biomarkers sensitive to treatment and disease mechanisms will help accelerate such new treatments. To this end, biophysically-informed dynamic causal models can support inferences around laminar and cell-specific disease effects from human non-invasive imaging. This thesis aims to use dynamic causal models to elucidate cellular-level effects of Alzheimer’s disease on cognition to (i) improve our understanding of disease mechanisms and (ii) identify mechanistically-informative biomarkers for interventional studies. The objectives of my thesis are to: (i) assess the reliability and sensitivity of magnetoencephalography evoked responses from a mismatch negativity paradigm to the severity and progression of Alzheimer’s disease and to employ dynamic causal modelling to (ii) reproduce these neurophysiological responses, (iii) test hypotheses of cellular-level disease effects in people living with Alzheimer’s disease based on preclinical- and post mortem- derived findings and (iv) to investigate how Alzheimer’s disease, its severity and progression affect cells, neurotransmitter systems and cortical networks underlying cognition. This thesis first details the study protocol used for the data collection of the primary study used in this thesis (the New Therapeutics in Alzheimer’s disease study, NTAD, PMCID: PMC9756184). NTAD is a multi-site project which aims to assess magnetoencephalography as a potential biomarker platform for experimental medicine studies of novel compounds. To this end, NTAD includes clinical, cognitive and neuroimaging measures from participants with mild cognitive impairment or early dementia from Alzheimer’s disease and neurologically-healthy control participants. The analyses herein use baseline, test-retest and 16-month follow up data from the Cambridge NTAD cohort. Group comparisons of the NTAD Cambridge cohort confirmed clinical, cognitive and anatomical changes typical of mild cognitive impairment and early Alzheimer’s dementia. For magnetoencephalography, sensor-level analyses demonstrated attenuated mismatch negativity responses from 140ms to 160ms for people with Alzheimer’s disease. Mismatch negativity responses were more attenuated when disease was more severe (lower Mini-Mental State Examination scores) and after annual follow up (follow up versus baseline). An absolute, intraclass correlation model of the test-retest mismatch negativity amplitude illustrated excellent reliability. To be suitable as a biomarker platform, magnetoencephalography must be reliable and sensitive to disease. Magnetoencephalography biomarkers should ideally also be informative about disease mechanisms at the level of cortical microcircuits and their cells or receptors. The cortical generators underlying the neurophysiological deficit are assessed in chapters 4, 5 and 6. Chapter 4 tests the hypothesis that Alzheimer’s disease affects the gain and connectivity of pyramidal cells. The analysis used in this chapter inverts a dynamic causal model with pyramidal cell gain and pyramidal cell extrinsic connectivity parameters to evoked magnetoencephalographic responses to the mismatch negativity task. I confirm that superficial pyramidal cell gain modulation and extrinsic connectivity between pyramidal cells are reduced in Alzheimer’s disease and reduce further over time. Chapter 5 reports how Alzheimer’s disease affects channel time constants and channel-mediated signalling. I inverted dynamic causal models with ion channel parameters to magnetoencephalographic responses to the mismatch negativity task. This analysis confirms that abnormal NMDA-channel function explains the neurophysiological deficit, more than AMPA or GABA channel changes. Alzheimer’s disease progressively changes NMDA channel time constants in primary auditory and inferior frontal regions. Chapter 6 examines the ability of dynamic causal models to identify the mechanism of action of a currently licenced medication for Alzheimer’s disease, memantine. To do this, I use magnetoencephalography data recorded during mismatch negativity paradigms from (1) neurologically-healthy people from a double-blind placebo-controlled crossover study of the NMDA antagonist memantine and (2) people with Alzheimer’s disease and controls from the Cambridge NTAD study. I first tested where memantine acts in the canonical microcircuit by inverting dynamic causal models to evoked responses from healthy controls on placebo versus memantine. Next, I assessed how the strength of this blockade affects cognitive function and whether it is affected by Alzheimer’s disease progression. I confirm in humans that memantine increases and Alzheimer’s disease decreases NMDA-receptor inhibition. The biomarkers identified in this thesis are particularly important to the field as they quantify cellular-level effects of disease mechanisms on cognition. These biomarkers could consequently be used by experimental medicine studies to evaluate drug effects on pyramidal cell and NMDA channel function that underlie neurophysiological deficits. Future studies could similarly tailor biologically-informed models to identify novel targets-of-interest and assess their relationship with disease severity and progression. Therapeutic target-engagement and effect on cognition could then be measured to determine the potential of novel therapeutics. The techniques used and biomarkers identified in this thesis, that quantify cellular properties of an individual, could inform personalised treatment plans in clinic. Future studies could determine whether these techniques can accurately determine which therapeutics, or therapeutic combination, are most beneficial to a patient. These techniques could similarly be applied to identify cellular-level disease effects underlying performance in other cognitive tasks. For example, in the NTAD study, magnetoencephalography was recorded in tasks assessing episodic memory and paired associates learning. These cognitive domains are affected early in the course of Alzheimer’s disease. As such, disease mechanisms underlying poorer performance in these tasks may be early markers of cognitive involvement in Alzheimer’s disease. Future studies could therefore apply the techniques illustrated in this thesis to identify biomarkers that provide insight into disease mechanisms affecting cognitive performance in other domains. In conclusion, this thesis uses mechanistic modelling to identify disease effects on the cortical microcircuit that can explain the observed neurophysiological deficit. In line with preclinical and pathological findings, the mechanistic modelling confirmed hypotheses that in Alzheimer’s disease there is reduced pyramidal cell connectivity, impaired NMDA-channel kinetics and opposing effects on the cortical microcircuit of disease and its therapeutics.
Degree
thesis:*- Name dc:type.qualificationname
- Doctor of Philosophy (PhD)
- Level dc:type.qualificationlevel
- Doctoral
- Grantor dc:publisher.institution
- University of Cambridge
- Year dc:date.issued
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Lanskey, Juliette
- Advisors dc:contributor.advisor
-
- Rowe, James
- Henson, Rik
Subjects
dc:subject × 10Rights
dc:rightsIdentifiers
dc:identifier.*- DOI dc:identifier.doi
- https://doi.org/10.17863/CAM.114777
- OAI identifier oai:identifier
- oai:www.repository.cam.ac.uk:1810/378299