BackgroundUnderstanding the relationship between Alzheimer's disease (AD) biomarkers and cognitive performance in diverse populations is essential to uncover how social and demographic factors influence brain pathology and clinical manifestations. Studying these associations in underrepresented groups, can enhance the global applicability of diagnosis using AD biomarkers and therapeutic strategies.MethodThis study included 72 participants with an average of 8.9 [SD ±5.1] years of education, ranging from 0 to 21 years, from a tertiary memory clinic in Brazil. All of them underwent CSF biomarker analysis and a neuropsychological evaluation as part of their diagnostic workup. Among the participants, 30 were diagnosed with AD (39 men and 33 women; mean age 65.1 years [SD ±8.2], range 43–83) with CSF biomarker levels of p‐tau 87.0 (±41.7), total tau 765 (±611), and AB‐42 579 (±195); 16 had behavioral variant frontotemporal dementia (bvFTD) (13 men and 3 women; mean age 64.3 years [SD ±7.9], range 43–76) with p‐tau 49.1 (±21.2), total tau 405 (±346), and AB‐42 875 (±334); seven had mild cognitive impairment (MCI) (1 men and 6 women; mean age 62.3 years [SD ±9.3], range 50–78) with p‐tau 63.4 (±36.7), total tau 315 (±322), and AB‐42 1050 (±330); and 19 had dementia without a specific classification (11 men and 8 women; mean age 63.8 years [SD ±9.7], range 43–82) with p‐tau 51.8 (±24.7), total tau 389 (±262), and AB‐42 941 (±379).ResultOur analysis revealed significant correlations between AD CSF biomarkers and cognitive performance. Total tau was negatively correlated with verbal memory performance (z_scoreRAVLT, Spearman's rho = ‐0.265, p = 0.025), while beta‐amyloid 42 (AB‐42) was positively correlated with verbal memory (z_scoreRAVLT, Spearman's rho = 0.266, p = 0.024). However, AB‐42 did not show significant correlation with global cognitive function (z_scoreMattis, Spearman's rho = 0.210, p = 0.076). No significant correlations between phosphorylated tau and cognitive measures emerged, as well.ConclusionIn a Brazilian population with a diverse educational level, total tau biomarkers showed significant correlations with verbal memory tests. However, phosphorylated tau and AB‐42 did not demonstrate such correlations. These findings suggest that the cognitive impairment does not follow amyloidosis, but seems to reflect neurodegeneration.

BackgroundNon‐human primates (NHPs) are becoming a critical component of brain aging research, since many aspects of aging cannot be properly modeled in rodents. Our earlier published study utilizing a NHP model of sporadic AD‐related pathology that develops abundant CAA, the squirrel monkey (SQM), indicated that treatment with TLR9 agonist, CpG‐ODN 2006, safely ameliorates CAA while providing cognitive benefits. We advanced our research to evaluate the efficacy and safety of CpG‐ODN 1018, which has shown excellent safety profiles in clinical trials, including COVID vaccine trials and as an adjuvant in the FDA approved hepatitis B vaccine (HEPLISAV‐B/Dynavax). The present study was designed to assess CpG‐ODN immunostimulatory patterns and to monitor disease progression by a combination of behavioral measures and fluid/imaging biomarker signatures of pathology development.MethodGeriatric SQMs have received either CpG‐ODN or saline injections every 5 weeks. The efficacy of CpG‐ODN in triggering an immune response was analyzed using the Nanostring nCounter System. SIMOA and Luminex were used to measure plasma and CSF biomarkers of AD pathogenesis, neurodegeneration, and neuroinflammation. A touchscreen‐based Automated Cognitive Testing System (ACTS) was implemented within SQM cohorts to assess cognitive changes.ResultLongitudinal Nanostring analyses demonstrated significant upregulation of IFN‐inducible genes and specific cytokine‐chemokine genes (IFIT2, Mx2/MxB, GBP1, MIG, IP10, MCP1) following CpG‐ODN administration. Our preliminary analyses of plasma/CSF Ab40, Ab42 and Ab42/Ab40 ratio, NfL, GFAP, Neurogranin, and TREM2 levels did not reveal differences over time (post‐6th injection compared to baseline levels) or between our treatment groups. Subsequent characterization of biomarkers of inflammation and cerebrovascular dysfunction is underway. The potential of using ACTS to longitudinally monitor cognitive function in socially living SQMs was also demonstrated. Furthermore, no evidence of complications such as amyloid‐related imaging abnormalities (ARIA) was observed in our CpG‐ODN treated monkeys, confirming the safety of our immunomodulatory approach. We are expanding on these findings with longitudinal monitoring of biofluid/imaging biomarker profiles, as well as cognitive performance at selected intervals during the treatment period.ConclusionOur comprehensive NHP study will further validate this concept of immunomodulation as a safer approach to effectively ameliorate dementia‐related pathology, particularly CAA, and support the potential clinical application of CpG‐ODN 1018.

BackgroundToxic Aβ oligomers (AβO) are implicated in the progression of Alzheimer’s disease (AD). PMN310 is a humanized IgG1 monoclonal antibody that binds to a computationally derived three‐dimensional epitope specific to misfolded Aβ in AβO. Because PMN310 inhibits toxicity of AβO and does not bind to plaque, thereby potentially limiting risk of ARIA, it is being developed as a therapy for early AD. Results from a single ascending dose study of PMN310 (A Phase 1a Study of PMN310 In Healthy Volunteers NCT06105528) indicate that PK parameters and CSF concentrations of PMN310 were linearly dose‐dependent and CSF concentrations were 100‐600 times the estimated AβO molar concentration. The plasma half‐life (t ½) was approximately 17.5 days and the CSF t ½ was approximately 27 days.MethodsPRECISE‐AD, NCT06750432, is a placebo‐controlled, multiple‐ascending dose study of PMN310 to evaluate safety, tolerability, PK, PD, and preliminary efficacy of multiple intravenous infusions of PMN310 in patients with early Alzheimer's disease. The study consists of three staggered dosing arms of 350 mg, 700 mg, 1400 mg. Patients will be randomized 3:1, PMN310: placebo and will receive either PMN310 or placebo once every 28 days for a total of 12 infusions. The study will enroll 128 patients with either Stage 3 or 4 AD. Diagnosis will be determined by clinical criteria, CSF and plasma biomarkers, and Aβ PET. MRI scans will be done at months 2, 4, 6, 9, 12 to detect potential ARIA. Plasma biomarkers (pTau217, pTau 243, GFAP, SNAP25, neurogranin, Aβ42/Aβ40, NfL) will be measured at baseline and at 3‐month intervals. Biomarkers in CSF will be measured at baseline, 6, 12 months. Cognitive outcomes (CDR‐SB, ADAS‐cog, ADAS‐ADL IADRS Clinical Impression of Change) will be assessed at baseline, month 6 and month 12.ResultsThe proposed study has sufficient power to detect at least one ARIA event. The proposed sample size has sufficient power to provide statistically meaningful insight into effects of PMN310 on biomarkers and clinical outcomes.ConclusionsPRECISE‐AD will be the first study to examine the effects of a monoclonal antibody directed solely against AβO on biomarkers associated with AD pathology and clinical outcomes.

BackgroundSporadic frontotemporal dementia (FTD) can affect both younger (<65 years old, Early‐Onset, eoFTD) and older individuals (≥65 years old, Late‐Onset, loFTD). The mechanisms behind varying onset ages are intriguing yet underexplored. This study aimed to investigate whether differences in neurodegeneration intensity, microglial activation, and vascular comorbidities distinguish eoFTD from loFTD or differently impact longitudinal cognitive decline, potentially elucidating mechanisms of different vulnerability in these subgroups.Method52 consecutive patients with sporadic FTD, stratified according to age at onset (eoFTD, n = 28; loFTD, n = 24), underwent lumbar puncture with CSF dosage of amyloid‐β42, phosphorylated‐tau (p‐tau), total‐tau (t‐tau), albumin quotient (Qalb), neurofilament light chains (NfL), and soluble TREM‐2 (sTREM‐2). Eight vascular risk factors were computed in a cumulative vascular score (VS). First, we compared CSF levels of each biomarker and VS between eoFTD and loFTD. Then, we used multivariate regression analyses to verify the association of CSF NfL, sTREM‐2, VS and Qalb, with cognitive decline at 18±3 months follow‐up (ΔMMSE) in each group.ResultThe groups were comparable in age, sex and in the prevalence of clinical FTD phenotypes (eoFTD: 21 bvFTD/8 PPA; loFTD: 18 bvFTD/6 PPA). We found no differences in CSF biomarkers and sTREM‐2, and a trend toward higher CSF NfL levels in loFTD compared to eoFTD (p = 0.061). Qalb and VS values were similar. Longitudinal evaluation further indicated that ΔMMSE were comparable between eoFTD and loFTD (p = 0.853). Multivariate regressions showed a negative association between NfL and ΔMMSE in the whole cohort (β=‐0.528, p = 0.005), with higher effect size in eoFTD (β=‐0.809, p = 0.011) but no significant association in loFTD. Age, Qalb values, sTREM‐2 levels and VS values were not associated with ΔMMSE in any subgroup.ConclusionOur results suggest a stronger association between NfL levels and cognitive decline in eoFTD than in loFTD, despite generally higher CSF NfL levels in the latter. This may reflect greater neurodegeneration susceptibility or a more direct link with cognitive decline in younger patients. Notably, no associations were found between ΔMMSE and vascular risk factors, microglial activation, or blood‐brain barrier permeability—factors linked to cognitive decline in Alzheimer's disease. These findings highlight potential FTD‐specific mechanisms requiring further investigation to tailor therapeutic strategies

BackgroundCognitive reserve is an increasingly studied concept. According to this framework, individuals with higher reserve can better cope with pathology, delaying symptom onset. The Cognitive Reserve Questionnaire (CRQ) is a brief (<2 min) eight‐item scale validated in the Spanish population. Recently, it has been used to create normative data for a screening cognitive test (TMA‐93), showing greater variability in scores than the isolated education level. This study aimed to explore sex differences in cognitive reserve, as measured by the CRQ, in a biologically‐confirmed Alzheimer's disease (AD) cohort enriched with illiterate individuals from southern Spain.MethodCross‐sectional study, including patients from the Outpatient Memory Clinic at Torrecárdenas Hospital in Almería (2023–2024), diagnosed with AD based on CSF biomarkers (p‐tau 181/Aβ42 < 0.023; Elecsys automated platform). Baseline sociodemographic data (age, sex, education, Fototest score) and biomarker levels (Aβ42, p‐tau 181, t‐tau) were collected, along with CRQ scores, reported by a reliable informant. We analyzed sex differences in total CRQ scores (Mann‐Whitney U test) and examined the distribution of each CRQ item score across sexes.Result53 AD patients (median age: 70 years, IQR: 66–75) were included (31 females, 22 males). No significant differences were found between sexes in age, education level, Fototest score, or biomarker levels (Table). Notably, only 37.7% of participants had education beyond primary school. Both sexes exhibited low cognitive reserve (median CRQ score: 7/25, IQR: 4–11), with significantly lower scores in females (Figure 1; W = 212; p = 0.01977). At the item level, males scored higher across all items (Figure 2A, 2B), with notable differences in occupation, reading, skills development courses, and languages spoken.ConclusionCognitive reserve, which moderates the relationship between brain pathology and cognitive outcomes, may be influenced by sex. Developing feasible cognitive reserve assessments applicable to illiterate individuals, which also account for sex differences, is essential. Women may have lower cognitive reserve, particularly in populations with a high prevalence of illiteracy. Future studies are needed to explore the sex‐specific moderating effects of AD pathology on cognition and the potential generational changes in cognitive reserve across sexes.

BackgroundEnrollment in clinical trials for early Alzheimer's disease is typically slow due to low rates of amyloid positivity but essential confirmatory testing using CSF biomarkers or amyloid PET imaging increases patient burden and operational challenges. By prescreening participants with blood biomarkers to enrich for amyloid positivity, enrollment can be accelerated, making trials more accessible and reducing overall costs.MethodsWe incorporated PrecivityAD (C2N Diagnostics) and Elecsys pTau181 (Roche Diagnostics) assays into the ongoing screening for PROGRESS‐AD (EU CT: 2023‐505083‐11‐00; NCT06079190; GSK Study ID 219867), a phase 2 double‐blind study enrolling approximately 282 patients with early Alzheimer's disease to receive anti‐sortilin mAb, GSK4527226/AL101, or placebo. Plasma samples from participants were tested and were considered amyloid positive for PrecivityAD APS score ≥ 36 (includes both intermediate and high range) or pTau181 ≥1 pg/ml. These results were compared to amyloid PET and CSF (Lumipulse β‐amyloid (1‐42/1‐40) ratio). Historical results were acceptable utilizing FDA approved tracers or CSF ratios with tests that meet regional clinical and regulatory requirements. Confirmatory PET or CSF results were available for PrecivityAD (n = 172) and pTau181 (n = 190) for ROC analysis with and without Beggs and Greene correction. Human biological samples were sourced ethically, and their research use was in accord with terms of the informed consents under an IRB/REC approved protocol.ResultsFor pTau181, the positive percent agreement (PPA) and negative percent agreement (NPA) with comparators were 85.0% [78.4 – 89.8] and 59.5% [42. – 75.] respectively. PrecivityAD had a PPA and NPA of 64.6% [57.1‐71.5] and 42.9% [15.8 – 74.9] respectively. The positive predictive value (PPV) for participants using PrecivityAD was 96.4% compared to the PPV of 90.0% from participants using pTau181. Overall, this reduced the screen failure rate to confirm amyloid positivity to 9.9%.ConclusionsUtilizing blood‐based biomarkers significantly decreased the number of participants required to have unnecessary CSF or PET testing. Limitations of this study are incomplete confirmatory results for amyloid negative individuals, regional PET/CSF confirmatory test bias, and the delay between historical result and blood biomarker testing. Including blood‐based biomarkers in screening can reduce participant burden, enhance access and enrollment and improve clinical trial design.

BackgroundMany patients with bipolar disorder (BD) report cognitive problems. Pathophysiology of cognitive impairment in BD is unclear, although prevalence of dementia in BD is high. In Older-Age Bipolar Disorder (OABD) patients with cognitive complaints, neurodegeneration may play a role. This could occur in at least 2 ways: (1) BD with 'comorbid' diagnosis of dementia; (2) specific neurobiological processes can underlie a cognitive impairment that is intrinsic of BD (ie, BD-related cognitive impairment).Methods102 OABD patients were selected from the Amsterdam Dementia Cohort study. Diagnostic workup included clinical and neuropsychological assessment, CSF biomarkers and MRI visual rating scales. About half of patients had depressive symptoms. We (1) examined which neurological diagnoses were identified by the memory clinic as the main cause of cognitive complaints. Subsequently, (2) in the remaining OABD patients with an unknown cause, we performed linear regression between biomarkers of neurodegeneration and composite cognitive score.Results29 OABD patients (28.4%) received a neurological diagnosis, 6 of which Alzheimer's Disease. In the remaining 73 (71.6%) OABD patients, a lower Aβ42 CSF concentration was related to lower composite cognitive scores (B = -0.143, P = 0.034), whereas CSF T-Tau, P-Tau, and MRI markers were not.ConclusionIn most OABD patients visiting a memory clinic, a neurological cause of cognitive complaints was not identified despite extensive diagnostic work-up. Altered amyloid metabolism may be an extra biological factor in the multifactorial puzzle that is BD-related cognitive impairment. Future studies should investigate a large range of biomarkers in relation to cognition in BD, including amyloid.

Recently, Alzheimers disease (AD), as one of most prevalent neurodegenerative disorders, has emerged as the main diseases affecting the elderly. It induces progressive memory loss, cognitive decline, and hippocampal atrophy [1], leading to the decline of the quality of individuals life and casing the burden of society. This problem of widespread dementia is not a temporary one either, but one that would have growing impacts as societies around the world begin to age. Although amyloid PET imaging and CSF biomarkers (e.g., A42, p-tau, total tau) [2] can be the diagnosis markers for the detection of AD pathology, non-invasive methods for the detection of AD at the early stage are still unclear.The breakdown of the blood-brain barrier (BBB) has been proved as an early event in AD pathogenesis [3]. Neutrophils, as key cells in the process of inflammation, infiltrate brain tissue following BBB breakdown, exacerbating neuroinflammation and neuronal apoptosis [4]. Several studies have identified the neutrophil-related biomarkers (e.g., NGAL, MPO, and NLR) [5] for AD diagnosis, indicating that the blood-based biomarkers related to neutrophil dysfunction offer a promising approach for early AD detection. However, these current biomarkers can not cover all the diagnosis of early AD development. Therefore, herein this study, by integrating multiple human and mouse single-cell RNA sequencing (scRNA-seq) datasets about blood neutrophils for BBB functions, we selected the 17 high potential biomarkers from more than 2,000 genes for blood neutrophil dysfunctions and BBB breakdown, we identified genes as potential non-invasive biomarkers for the early diagnosis of AD. By conducting such a study and identifying neutrophil-related transcripts across species, we aim to contribute to the development of a non-invasive, scalable approach to AD detection that could be implemented on large-scale screenings.

BackgroundConcentration‐based fluid biomarkers represent an informative and cost‐effective way to detect and monitor Alzheimer's disease (AD) pathology. However, non‐AD‐related inter‐individual variation in biofluids can also affect biomarker concentrations. We previously identified several reference proteins that, in the AT(N) classification framework, improved concordance between CSF Aβ42 and Aβ‐positron emission tomography (PET), as well as between CSF p‐tau181 and tau‐PET.1 However, it is still unclear what effect reference proteins have on the relationship between CSF AD biomarkers and the load of AD pathology. It is also unclear if plasma AD biomarkers can be improved by accounting for reference proteins in a similar manner.MethodsUsing the Swedish BioFINDER‐2 cohort (n = 1702, 50.7% male, mean [SD] age 68.4 [12.2] years), we compared the associations between tau/Aβ‐PET load and CSF biomarkers (MTBR‐tau243, p‐tau217, p‐tau181, p‐tau205, Aβ42, SNAP‐25, neurogranin) alone versus in a ratio with a reference protein (e.g. CSF Aβ40 or non‐phosphorylated tau [np‐tau]) in univariate linear regression models. We repeated this analysis for plasma biomarkers.ResultsCSF Aβ40 normalization significantly strengthened the associations of several core CSF AD biomarkers, including CSF MTBR‐tau243, p‐tau isoforms and synaptic biomarkers, with tau‐PET (ΔR2=0.064‐0.24) and Aβ‐PET (ΔR2=0.016‐0.28), Figure 1. CSF np‐tau normalization mainly improved concordance between CSF biomarkers and Aβ‐PET (ΔR2=‐0.0059‐0.19). The strongest association with tau‐PET was observed for MTBR‐tau243/Aβ40 (R‐squared=0.78, compared to 0.65 for non‐normalized MTBR‐tau243), and with Aβ‐PET for p‐tau217/np‐tau (R‐squared= 0.65, compared to 0.46 for non‐normalized p‐tau217). For core plasma AD biomarkers, including MTBR‐tau243 and p‐tau isoforms, associations with tau‐PET were enhanced by using plasma Aβ40 or np‐tau as references (ΔR2=0.0019‐0.14), while associations with Aβ‐PET mainly improved with np‐tau (ΔR2=0.018‐0.16), Figure 2. The findings were successfully replicated in Knight ADRC and TRIAD for improved biomarker associations with both tau‐PET (Table 1) and Aβ‐PET.ConclusionsNormalization to reference proteins (i.e., Aβ40 or np‐tau) enhances the associations between CSF and plasma biomarkers with the load of tau and Aβ pathology in the brain, making already high‐performing AD and synaptic fluid biomarkers even more precise.Reference1. Karlsson, L. et al. Cerebrospinal fluid reference proteins increase accuracy and interpretability of biomarkers for brain diseases. Nat Commun15, (2024).

BackgroundThe enzyme indoleamine‐2,3‐dioxygenase (IDO1) is highly expressed in astrocytes in response to inflammatory stimuli associated with Alzheimer's disease (AD). IDO1 plays a critical role in the conversion of tryptophan to kynurenine, a process that facilitates immune suppression under inflammatory conditions. Recent studies have revealed a novel role for IDO1 in brain metabolism, showing that its activity impairs astrocytic glycolysis and reduces metabolic support to neurons in preclinical models of AD, thereby exacerbating amyloid and tau pathologies. However, the functional implications of gain‐ or loss‐of‐function single nucleotide polymorphisms (SNPs) in the IDO1 gene among individuals with AD remain poorly understood. This study investigates whether IDO1 SNP carriership contributes to metabolic disruptions across the AD continuum.MethodsWe assessed 619 cognitively unimpaired (CU) and impaired (CI) individuals from ADNI with available baseline FDG‐PET, IDO1 genotyping, CSF biomarkers, and cognitive assessments. SNP haplotype assignment was conducted for the combination of five identified IDO1 SNPs (rs7820268, rs62512635, rs77622610, rs3739319, rs10108662). Generalized linear mixed‐models were performed to assess the effect of haplotype carriership on FDG‐PET for the five most frequent haplotypes. The haplotype with a significant impact on brain metabolism was subsequently tested for its interaction with APOE4 and association with cognitive decline or CSF biomarkers using linear models and voxel‐wise regression analysis (p <0.05).ResultsThe most frequent IDO1 SNP haplotype (15.5%) corresponds to carriers of one minor allele for rs7820268, rs3739319, and rs10108662, referred to here as the 10011 haplotype (Table 1A). Carriership of this haplotype was associated with distinct alterations in brain glucose metabolism, characterized by hypermetabolism in CU and hypometabolism in CI individuals, independent of Aβ and tau status (Figure 1B). CI carriers of the 10011 haplotype exhibited prominent hypometabolism in the temporal region (Figure 2A), an effect potentialized by APOE4 (Figure 2B). In addition, CI carriers presented severe cognitive impairment, as confirmed by four different assessments (Figure 2C). Finally, glucose hypometabolism was associated with higher levels of CSF pTau181 and GAP43 (Figure 3A‐B).ConclusionTo the best of our knowledge, this is the first in vivo clinical evidence demonstrating that an IDO1 SNP exacerbates brain hypometabolism and worsens cognitive decline in AD.

BackgroundThe enzyme indoleamine‐2,3‐dioxygenase (IDO1) is highly expressed in astrocytes in response to inflammatory stimuli associated with Alzheimer’s disease (AD). IDO1 plays a critical role in the conversion of tryptophan to kynurenine, a process that facilitates immune suppression under inflammatory conditions. Recent studies have revealed a novel role for IDO1 in brain metabolism, showing that its activity impairs astrocytic glycolysis and reduces metabolic support to neurons in preclinical models of AD, thereby exacerbating amyloid and tau pathologies. However, the functional implications of gain‐ or loss‐of‐function single nucleotide polymorphisms (SNPs) in the IDO1 gene among individuals with AD remain poorly understood. This study investigates whether IDO1 SNP carriership contributes to metabolic disruptions across the AD continuum.MethodsWe assessed 619 cognitively unimpaired (CU) and impaired (CI) individuals from ADNI with available baseline FDG‐PET, IDO1 genotyping, CSF biomarkers, and cognitive assessments. SNP haplotype assignment was conducted for the combination of five identified IDO1 SNPs (rs7820268, rs62512635, rs77622610, rs3739319, rs10108662). Generalized linear mixed‐models were performed to assess the effect of haplotype carriership on FDG‐PET for the five most frequent haplotypes. The haplotype with a significant impact on brain metabolism was subsequently tested for its interaction with APOE4 and association with cognitive decline or CSF biomarkers using linear models and voxel‐wise regression analysis (p <0.05).ResultsThe most frequent IDO1 SNP haplotype (15.5%) corresponds to carriers of one minor allele for rs7820268, rs3739319, and rs10108662, referred to here as the 10011 haplotype (Table 1A). Carriership of this haplotype was associated with distinct alterations in brain glucose metabolism, characterized by hypermetabolism in CU and hypometabolism in CI individuals, independent of Aβ and tau status (Figure 1B). CI carriers of the 10011 haplotype exhibited prominent hypometabolism in the temporal region (Figure 2A), an effect potentialized by APOE4 (Figure 2B). In addition, CI carriers presented severe cognitive impairment, as confirmed by four different assessments (Figure 2C). Finally, glucose hypometabolism was associated with higher levels of CSF pTau181 and GAP43 (Figure 3A‐B).ConclusionTo the best of our knowledge, this is the first in vivo clinical evidence demonstrating that an IDO1 SNP exacerbates brain hypometabolism and worsens cognitive decline in AD.

BackgroundOlfactory deficits predict future dementia and Alzheimer’s Disease (AD). It is crucial to unravel the mechanisms of olfactory dysfunction in the pre‐dementia stages to understand their potential as early markers. We aimed to examine the associations of AD‐related CSF biomarkers and central olfactory system volumes with odor identification (OID) in subjective cognitive impairment (SCI), mild cognitive impairment (MCI), and AD.MethodIndividuals with SCI (n = 154; Mage=58.5, SDage = 5.7; %females=64.9), MCI (n = 51; Mage=61.6, SDage = 5.0; %females=49.0), and AD (n = 31; Mage=58.7, SDage = 5.0; %females=54.8) were recruited from the Karolinska University Hospital Memory Clinic, Solna, Sweden. We examined within‐group associations of OID, assessed with the Sniffin' Sticks test (number of correct identifications, range 0 to 16), with CSF biomarkers (42 amino acid form of amyloid‐β [Aβ42], tau phosphorylated at threonine 181 [p‐tau181], and neurofilament light chain [NfL]), and central olfactory system volumes extracted using cNeuro cMRI software (Combinostics Oy; left hippocampus, amygdala, left parahippocampal gyrus, entorhinal cortex, orbitofrontal cortex, insula, and left caudate). Data were analyzed using one‐tailed partial correlations for all groups (controlled for age, sex, and education) and multiple linear regressions applying stepwise model selection for SCI and MCI. The alpha level was set at 0.05.ResultSCI (M = 14.1, SD = 2.2) outperformed MCI (M = 12.7, SD = 3.5) and AD groups (M = 12.8, SD = 2.8) in OID. In SCI, Aβ42 (r = 0.136) and orbitofrontal volume (r = 0.194) were associated with OID. In MCI, hippocampal (r = 0.347), parahippocampal (r = 0.274), and caudate volumes (r = 0.343) were associated with OID. In AD, hippocampal volume (r = 0.557) was associated with OID. Aβ42 and orbitofrontal volume were significant predictors of OID in the linear regression model in SCI (adjusted‐R2 = 0.102); whereas NfL, entorhinal, and insula volumes were significant predictors of OID in MCI (adjusted‐R2 = 0.497).ConclusionOlfactory deficits were associated with amyloid CSF levels and lower orbitofrontal volume in individuals with SCI, suggesting that amyloid deposition and orbitofrontal degeneration may play a role early in the disease spectrum. In MCI and AD, degeneration of medial temporal lobe and subcortical structures may contribute to these deficits.

BackgroundAlzheimer's disease (AD) is characterized by pathological amyloid‐β accumulation. Positron emission tomography (PET) imaging remains a widely accepted gold standard for quantifying amyloid burden via standardized uptake value ratios (SUVRs); however, its prohibitive cost and limited availability restrict broader clinical adoption. Although recent deep learning models show promise in predicting amyloid burden, they often require complete clinical data or actual PET images—conditions rarely satisfied in routine clinical practice. To overcome these barriers, we propose a masked multimodal–multitask deep learning framework that integrates synthetic PET scans to improve amyloid burden prediction under real‐world data constraints.MethodWe analyzed 2,043 longitudinal observations from 968 ADNI‐2 and ADNI‐3 participants (mean age: 72 years; male/female: 491/477). Each observation included T1‐weighted, T2*‐weighted, and FLAIR MRI, AV45‐PET, demographic data, APOE4 status, cognitive assessments, and CSF biomarkers. Our approach involved two main steps:1. Synthetic PET generation: A latent diffusion model (LDM) was trained to generate synthetic AV45‐PET scans from MRI sequences.2. Amyloid burden prediction: A deep learning network integrated these synthetic PET images with available clinical data through a masked embedding attention mechanism to explicitly handle missing inputs and predict both continuous amyloid SUVRs and masked clinical features.We compared our method against three baselines: (i) MRI‐only, (ii) non‐imaging–only, and (iii) a multimodal model without synthetic PET. Performance on a held‐out test set was evaluated using mean absolute error (MAE) for SUVR prediction and area under the curve (AUC) for amyloid positivity (SUVR > 1.11).ResultFor SUVR prediction, the baseline models achieved MAEs of 0.20, 0.13, and 0.13, whereas our approach reached 0.11. For amyloid positivity classification, the baselines yielded AUCs of 0.48, 0.89, and 0.90, while our model attained 0.93.ConclusionBy reducing reliance on cost‐intensive PET scans and handling missing data more effectively, our framework could substantially broaden access to early AD screening in diverse clinical settings. Future work will involve testing in larger, more varied cohorts, applying different tracers, and further evaluate its real‐world impact on AD management.

BackgroundAlzheimer's disease (AD) is commonly diagnosed when neuronal damage is already established and irreversible. A differential diagnosis in the mild cognitive impairment (MCI) stage is one of the greatest challenges nowadays. Blood biomarkers, and specifically plasma extracellular vesicles (pEVs), are gaining much interest as a new biomarkers’ source for the early stages of AD. This work aims to evaluate the proteomic profile of pEVs from patients with MCI and AD to explore their potential as AD screening tools.MethodspEVs were isolated by ultracentrifugation from patients with MCI Aβ(+) (n = 50), MCI Aβ(‐) (n = 50), and with AD dementia (n = 43). Nanoparticle tracking analysis (NTA) and cryo‐TEM were used to characterize the pEVs. CSF, serum and pEVs proteomics were carried out by using the multiplex PEA technology of Olink© proteomics, Inflammation and Neurology Explore 364 panels (728 proteins).ResultsCharacterization results showed that isolated plasma fraction corresponded in shape, size and concentration to EVs. pEVs’ biomarkers correlated with common AD signatures (CSF Aβ42 and pTau181, plasma pTau181, MMSE, Age and Qalb) with the same pattern than CSF biomarkers as previously described. Several neurology proteins of the pEVs (e.g. MMP‐8, MMP‐9, IGF2R or NDRG1) didn’t exhibit differences between the MCI Aβ(+) vs AD dementia groups, whilst MCI Aβ(‐) vs AD dementia did. Likewise, many pEVs neurology proteins significantly correlated (rho>0.30 and p <0.05) with their CSF homonyms, not instead with their serum homonyms (e.g. APP, NOS1, CXCL11, SOD2, BCAN …).ConclusionPreliminary results suggest that the biomarkers signature of pEVs could inform about the AD pathology in the prodromal stages of AD continuum. However, further experiments are still needed for a better understanding of the EVs’ role in the AD development and pathology dissemination.

Typical Alzheimer's disease (AD) and Limbic-predominant Age-related TDP-43 Encephalopathy (LATE) are neurodegenerative diseases that present with a similar amnestic syndrome but different proteinopathies. While these diseases are usually described as cortical diseases, our previous cross-sectional study reported white matter (WM) fiber bundle alterations in both diseases with distinct patterns, suggesting a possible role of WM damage in these diseases. To further test this hypothesis, in this work we investigated the evolution of these alterations after two years in early AD and presumed LATE using high-quality diffusion MRI and fixel-based analysis. We included 16 AD patients, 12 presumed LATE patients, and 15 healthy controls based on strict clinical and pathophysiological criteria (CSF biomarkers and amyloid and tau PET imaging) (Table 1). All participants underwent two 3-tesla brain MRIs two years apart. Using multi-shell diffusion MRI, we first performed whole-brain fixel-based analyses to investigate the progression of fixel metrics over two years within each group of patients (no covariates) and between patients and controls (covariates: age and sex). We then performed tract-based analyses, using age and sex as covariates, to compare the rate of change in fiber density and cross-section (FDC) between patients and controls at the level of tracts of interest that were previously identified in the whole-brain analyses and in our previous cross-sectional study. We found that damage to WM fiber bundle progressed in both patient groups after two years (Figure 1), and that this decline was more marked in patients than in controls in tracts connecting the temporal lobe to the parietal and frontal lobes (arcuate fasciculus and middle longitudinal fasciculus) (Figure 2). In addition, AD patients exhibited a more pronounced decline than controls in the cingulum, the inferior longitudinal fasciculus and the temporopulvinar bundle of Arnold, while presumed LATE exhibited a more pronounced decline in the ventral section of the superior longitudinal fasciculus. These results show that WM fiber bundle alterations are not only detectable in patients with AD and presumed LATE but that they worsen over time, more markedly than in controls, underscoring the importance of WM fiber bundle alterations in these diseases.

BackgroundThis study aims to explore plasma inflammation biomarkers in Alzheimer's disease dementia (ADD) and dementia of other neurodegenerative subtypes (NonADD), and associations with CSF AD biomarkers.Method245 participants including controls and ADD, and 328 participants including controls, ADD and NonADD, were selected from cohort 1 and 2, respectively (Figure 1). CSF biomarkers (Aβ42, ptau181, ttau) and plasma inflammatory biomarkers including counts of neutrophil, lymphocyte, platelet, CRP, Albumin were measured. Statistics were performed to assess prediction performances and correlations.ResultCombination of two CSF ratios, four inflammation ratios and ApoE4 showed optimal performance in distinguishing ADD from controls (Figure 2A). Combination of two CSF biomarker ratios, two inflammation ratios (neutrophil‐to‐lymphocyte, NLR and platelet‐to‐lymphocyte, PLR) and ApoE4 showed optimal performance in distinguishing ADD from NonADD (Figure 2B). Plasma inflammation biomarkers especially NLR showed positive correlation with CSF AD biomarkers in ADD (Figure 3).ConclusionPlasma inflammation biomarkers showed added value in the diagnosis and differential diagnosis of ADD.

BackgroundAlphaherpesviruses, including Herpes Simplex Virus Type 1 (HSV‐1) and Varicella‐Zoster Virus (VZV), are human neurotropic viruses with a high prevalence worldwide. The relationships between neurotropic Herpesviruses and Alzheimer's disease (AD) have been repeatedly emphasized. Our project is based on the hypothesis that neuro‐invasion of alphaherpesviruses, especially of HSV‐1, could trigger or contribute to the formation of early lesions in AD.MethodTo test our hypothesis, we use a multidisciplinary approach based first on two different animal models for studying the neuropathological consequences of HSV‐1 infection: the cotton‐rat model and the mouse model. Animals were infected in the upper lip and sacrificed at different time points after infection. Histological analyses were conducted to detect viral proteins, microglia, Aß and pTau deposits by immunostaining. Various stains were used to highlight tissue damage. Viral RNAs (lytic or latent phase of HSV‐1) were detected by in situ hybridization (RNAscope). Biochemical assays of Aß/ Tau proteins were performed using the Meso Scale Discovery technology. Another aspect of the project analyses the Shatau cohort, a cohort of human individuals (AD and non demented cases) for which we have serological data for HSV‐1 and VZV, in vivo measurements of Locus Coeruleus integrity, neuroimaging markers of brain accumulation of Aß and pTau proteins, and AD biomarkers in cerebrospinal fluid.ResultIn the different animal models of peripheral HSV‐1 infection we were able to evidence a neuro‐invasion characterized by the presence, in the brainstem, of viral proteins, lytic/latent viral genomes and neuroinflammatory status. In addition, we detected, by immunohistochemistry and biochemical dosages, evidence of abnormal deposition of Aß and hyperphosphorylated tau proteins evocative of AD. Our results in the analysis of the Shatau cohort indicate that, as hypothesized, anti‐viral antibody titrations are increased in AD vs non‐demented control patients. In addition, VZV titers appear surprisingly to be associated with AD biomarkers for the reduction of the Locus Coeruleus integrity and increased in CSF biomarkers, while association with HSV‐1 titers are weaker or absent.ConclusionOur results strengthen the hypothesis of a causal connection between infection and neuro‐invasion of alphaherpesviruses and AD neuropathologies.

Biofluid‐based biomarkers detect amyloid, tau and neurodegeneration (ATN) pathophysiological changes prior to onset of symptoms. But are there biomarkers that can predict the onset of these pathologies, i.e., predict AD before the brain changes have developed? Genetics, proteomics and ultrasensitive immunoassays were performed in 193 subjects with AD across the cognitive spectrum. In APOE ε4 carriers, i.e., in individuals at increased risk of AD, with normal amyloid β and cognition, CSF biomarkers for neuronal injury were normal. Compared to the control group, ten proteins (apoE4 fragment, tau, BACE1, β‐nerve growth factor [NGF], macrophage inflammatory protein‐1β [CCL4], osteopontin [SPP1], AXL receptor tyrosine kinase [AXL], heparin‐binding EGF‐like growth factor [HBGF], carbonic anhydrase‐1 [CA1], interferon gamma‐induced protein‐10 [CXCL10]) showed altered levels. The top KEGG pathway enriched was the Toll‐like receptor signaling pathway (pFDR = 0.00033; including SPP1, CXCL10, and CCL4). Upregulated amyloidogenic APP‐processing and increased neuronal and astroglial activity may precede amyloid positivity according to CSF biomarkers. Data replicating these findings in larger cohorts will be presented.

BackgroundDementia is a growing concern throughout the developing world and is severely underdiagnosed among the Brazilian population. Despite remarkable progress in the biomarker field in recent years, local testing and validation of plasma biomarkers of AD and dementia is still lacking in Brazil and Latin America.MethodsIn this longitudinal cohort study of 145 participants, the diagnostic performance of plasma biomarkers was assessed based on clinical diagnosis and CSF biomarker positivity. Follow‐up data of up to 4.7 years were used to determine biomarker performance in predicting diagnostic conversions. The study was conducted at the Memory Clinic at the D’Or Institute for Research and Education (IDOR) in Rio de Janeiro. Participants were volunteers referred to the service. All were native Brazilians, had Portuguese as their first language and 60+ years of age. They were clinically diagnosed and underwent psychiatric and laboratory assessments. Diagnoses outside the scope of the study were excluded. CSF biomarker data was available for 36% of the sample. Participants were categorized as cognitively unimpaired (CU; n = 49) or cognitively impaired (CI; n = 96), with CI participants further classified as aMCI, Alzheimer's disease (AD), Lewy body dementia, or vascular dementia. Plasma Tau, Aβ40, Aβ42, NfL, GFAP, pTau231, pTau181 and pTau217 were measured at initial and follow‐up visits using a SIMOA HD‐X instrument. To assess the diagnostic performance of each biomarker, receiver operating characteristic (ROC) curve analysis was performed.ResultsThe pTau217/Aβ42 ratio was the top‐performing biomarker for discrimination between CSF‐biomarker‐negative and CSF‐biomarker‐positive subjects (AUC=0.98, 95% CI: [0.94 – 1.00]), followed by pTau217 alone (AUC=0.94, 95% CI: [0.88 – 1.00]. When discriminating participants based on their cognitive status (CU x CI), pTau217 and pTau217/Aβ42 ratio reached identical AUCs of 0.82 (95% CI: [0.75 – 0.89]). Similar results were seen when discriminating CU from all‐cause dementia, with pTau217 and pTau217/Aβ42 ratio tied at the highest AUCs (AUC=0.87, 95% CI: [0.80 – 0.94]).ConclusionsIn this first biomarker profiling of a Brazilian dementia cohort, plasma pTau217 confirmed its potential as a clinically useful diagnostic tool. This study comprises an initial step towards local validation and adoption of dementia biomarkers in Brazil and Latin America.

Alzheimer's disease (AD) is characterized by tau and amyloid-β (Aβ) pathologies, influenced by factors such as sex, aging, APOE genotype, and sleep. Sleep is essential for memory consolidation, but its interaction with these factors and AD biomarkers remains unclear. This study examined how sex, APOEε4 status, and CSF AD biomarkers (Aβ42/40, p-tau) affect the relationship between sleep and overnight memory retention in older adults. Sixty-six Wisconsin Registry for Alzheimer's Prevention participants (41 women, 25 ε4-carriers, mean age 61.8±6.1 years) encoded word pairs in the evening, underwent overnight polysomnography, and completed a morning memory test. Memory retention was calculated as the morning-evening difference, with sleep stages quantified. CSF biomarkers (p-tau, Aβ42/40) were assessed using the exploratory NeuroToolKit panel on Cobas® analyzers (Roche Diagnostics International Ltd, Rotkreuz, Switzerland) under strict quality control. Moderated mediation analysis (PROCESS model 11) examined relationships between sleep stages (predictor), memory retention (outcome), biomarkers (mediators), and sex/APOEε4 status (moderators). A sex×APOEε4 interaction predicted memory (p=0.02) where female ε4-non-carriers exhibited more forgetting than ε4-carriers (p=0.02). Male ε4-carriers had worse memory retention than female ε4-carriers (p=0.01). Female ε4-carriers had more N3-sleep stage (p=0.01), correlating with better memory retention (r=0.72, p=0.02), while male ε4-carriers had more N2-sleep (p=0.04), also linked to better memory retention (r=0.68, p=0.03). Moderated mediation analysis demonstrated that N2-sleep negatively predicted p-tau levels across participants (p=0.03), suggesting a potential protective effect, while higher p-tau was linked to lower overnight memory retention (p=0.04). A significant interaction between sex and APOEε4 (p=0.05), N2 and APOEε4 (p=0.07-trend), and a main effect of APOEε4 (p=0.04) on Aβ42/40 levels suggest that APOEε4 status is a determinant of Aβ42/40. The findings indicate that the relationship between N2-sleep and Aβ42/40 is not uniform but are moderated by whether an individual carries the APOEε4 allele. No significant associations were found with N3-sleep (p>0.05). Our findings show that NREM sleep stages influence overnight memory retention in older men and women, depending on APOE genotype and CSF AD biomarkers, suggesting that sex differences in long-term memory retention are shaped by interactions between sleep and AD pathology.