Translational studies on biological signatures, risk profiles, and prevention of dementia

Neurodegenerative disorders that cause dementia, including Alzheimer's disease (AD), are multifaceted in their risk factor profiles and pathological changes. Today, we can use both fluid and imaging biomarkers to identify individuals with certain pathological changes, such as amyloid-beta (AB) aggregation, tau phosphorylation, and neurodegeneration, which can guide targeted interventions. However, these biomarkers account for a very limited biological signature. The mixed pathologies and risk factor profiles complicate the development of treatments and diagnostic tools for neurodegenerative disorders. There is an urgent need to expand on the biological signatures in neurodegenerative disorders and their relation to risk profiles to advance the understanding of both similarities and differences between patients, beyond and among different diagnostic and biomarker- classified groups. Knowledge about risk profiles and associated biological signatures could provide valuable information for targeted treatment and intervention options.

A great proportion of dementia cases are attributed to modifiable risk factors, providing a window of opportunity for lifestyle-related interventions to prevent dementia. A way to circumvent the need to identify what lifestyle factor(s) to change in which individual to gain optimal cognitive benefits is to apply a multimodal lifestyle intervention simultaneously targeting several risk factors. Optimally, a holistic lifestyle intervention could elicit synergistic effects between the lifestyle domains, further enhancing the effect of a multimodal lifestyle change compared to only single-domain interventions. While a multimodal lifestyle intervention is likely beneficial for many, irrespective of the risk profile, there is currently limited knowledge about risk profile-stratified effects. Furthermore, although multimodal lifestyle interventions have demonstrated clinical efficacy, it is currently not known which mechanistic changes such interventions evoke and whether the brain changes are dependent on risk and pathological profile. This doctoral thesis aims to unravel biological signatures, risk profiles, and prevention of dementia.

The first three constituent papers, Papers I, II, and III, investigate biological signatures and dementia-associated risk profiles. In Paper I, we measured 49 cerebrospinal fluid (CSF) proteins from a diverse memory clinic population, consisting of the full dementia continuum and different AD biomarker profiles, to assess biological variation within the population by applying principal component analysis. To our surprise, we found that most of the biological variance (up to 52%) between patients in our cohort was not attributed to pathology, but rather to physiological differences between patients. Nevertheless, we found a biological signature (accounting for 5% of biological variance), driven by gliosis and synaptic proteins, that was strongly associated with AD pathology, neurodegeneration, and cognitive impairment.

Paper II studied memory clinic patients as well, comprising patients at different stages in the dementia continuum without the common comorbidities of hypertension, hypercholesterolemia, and diabetes. We measured 11 CSF markers relevant to vascular function, cholesterol dysmetabolism, inflammation, oxidative stress, and glucose homeostasis. Despite not having diagnoses of comorbidities, we identified a cluster of patients with higher CSF levels of markers indicating higher oxidative stress and cholesterol dysmetabolism. The two identified patient clusters were independent of clinical diagnoses and detectable brain AD pathology.

Paper III studied a population of cognitively unimpaired older adults who had been dichotomized into those at greater and those at lower risk of developing AD, based on CSF measures of Aß42. We analyzed 67 CSF proteins to assess the biological signature related to the amyloid-associated dementia risk profile. We identified several proteins to be associated with a greater risk of developing AD, but these proteins did not predict cognitive decline during 3.5 years in the studied cohort.

In the last paper, Paper IV, we applied a multimodal lifestyle intervention to three different mouse models: healthy (WT), cholesterol dysmetabolism (CYP27Tg), and brain amyloidosis (AppNL-G-F). The lifestyle intervention elicited cognitive gains in WT and AppNL- G-F mice, but not in the CYP27Tg mice. WT mice who underwent the lifestyle intervention had enhanced synaptic functioning and plasticity in the hippocampus, as determined by proteomics analysis, while in AppNL-G-F mice we did not detect differences in the hippocampal proteome. In CYP27Tg mice, despite non-detectable cognitive improvement, the lifestyle intervention revived homeostasis in the hippocampus. The lifestyle intervention ameliorated model-specific pathology in both CYP27Tg and AppNL- G-F mice.

Collectively, the results from the constituent papers indicate that among patients in the dementia continuum, there are biological signatures that extend beyond the clinical diagnosis and AD-related biological stratification, and signatures that are associated with dementia-related risk profiles. We further established in animal models that the effects of lifestyle interventions, even if beneficial, are not universal but depend on dementia risk profiles. These findings endorse future efforts to deepen the knowledge regarding biological signatures for different dementia risk profiles. This information could be used to recognize how different people would benefit from lifestyle interventions.

List of scientific papers

I. Vilma Alanko, Sára Mravinacová, Anette Hall, Göran Hagman, Rosaleena Mohanty, Eric Westman, Peter Nilsson, Miia Kivipelto, Anna Månberg, Anna Matton. Biological signatures in the Alzheimer's continuum discriminate between diagnosis-related and -unrelated associations to ATN categories. Brain Communications. 2025 Feb 21;7(2):fcaf078. https://doi.org/10.1093/braincomms/fcaf078

II. Makrina Daniilidou, Francesca Eroli, Vilma Alanko, Julen Goikolea, Maria Latorre-Leal, Patricia Rodriguez-Rodriguez, William J Griffiths, Yuqin Wang, Manuela Pacciarini, Ann Brinkmalm, Henrik Zetterberg, Kaj Blennow, Anna Rosenberg, Nenad Bogdanovic, Bengt Winblad, Miia Kivipelto, Delphine Ibghi, Angel Cedazo-Minguez, Silvia Maioli, Anna Matton. Alzheimer's disease biomarker profiling in a memory clinic cohort without common comorbidities. Brain Communications. 2023 Aug 25;5(5):fcad228. https://doi.org/10.1093/braincomms/fcad228

III. Vilma Alanko, Oliver Robinson, Chinedu Udeh-Momoh, Sára Mravinacová, Jennie Olofsson, Janice Wong, Peter Nilsson, Anna Månberg, Miia Kivipelto, Lefkos Middleton, Anna Matton. Discrepancy between CSF protein profiles associated with amyloid positivity versus cognition in cognitively unimpaired older adults - Findings from the CHARIOT:PRO CSF substudy. [Manuscript]

IV. Vilma Alanko, Francesca Eroli, Ákos Végvári, Alina Solomon, Tobias Hartmann, Per Nilsson, Miia Kivipelto, Silvia Maioli, Anna Matton. Mouse PAW: Reverse-Translation of the FINGER Multimodal Lifestyle Trial Improves Memory and Dementia-Related Mechanisms in Female Mice. [Manuscript]