Aims

The majority of Alzheimer’s disease (AD) cases develop sporadically, although different genetic factors have been linked to the illness. The most studied is the APOE genotype, whose ε4 allele increases the likelihood of developing AD, reducing the age of onset. Many studies have analyzed AD biomarkers in APOE-stratified cerebrospinal fluid (CSF) collections, but few addressed whether the apolipoprotein E (apoE) protein is altered in AD.

Methods

ApoE protein levels from CSF samples of individuals with varying disease states (control and AD) and APOE genotype (ε3/3, ε3/4, ε4/4) from 2 independent cohorts were studied via SDS-PAGE (reducing and non-reducing conditions) and native-PAGE.

Results

We found that apoE is present in CSF as ~36 kDa species, but also at 34 kDa, probably representing immature glycoforms. Despite the absolute amount of CSF-apoE protein levels being significantly higher in AD patients, the ratio of mature/immature apoE glycoforms was lower compared to controls. Furthermore, a high molecular weight SDS-PAGE-resistant apoE band of ~100 kDa appeared almost exclusively in AD patients. ApoE3 and apoE2 form disulphide-linked dimers in CSF, and apoE homodimers have been reported at ~100 kDa (as opposed to the predicted ~70 kDa), whereas apoE4 cannot form these dimers as it lacks a cysteine residue. We corroborated the presence of these species under non-reducing conditions, although we also observed the presence of 100 kDa apoE β-mercaptoethanol-resistant species in AD samples, including apoE4 carriers. Results were corroborated in two independent cohorts.

Conclusions

In conclusion, the study shows that apoE protein levels and conformational state appear to be affected in AD CSF.

Aims

To assess the ability of plasma Aβ42/Aβ40 ratio as determined by a novel antibody-free HPLC-MS/MS method (ABtest-MS, Araclon Biotech) to detect early brain Aβ deposition in individuals at risk of Alzheimer’s disease (AD).

Methods

We have developed a unique method that quantifies Aβ40 and Aβ42 directly from plasma. Since no immunoprecipitation or enzymatic digestion steps are required, sample preparation time and cost are drastically reduced. A comprehensive analytical validation was performed.

Aβ40 and Aβ42 plasma levels were measured with ABtest-MS in 200 individuals with subjective cognitive decline from the Fundació ACE Healthy Brain Initiative (FACEHBI) cohort. Participants underwent comprehensive neurological evaluation and cognitive testing, APOE genotyping and ¹⁸F-florbetaben (FBB)-PET brain imaging. Cutoff for early Aβ-PET positivity was established at >13.5 centiloids (CL).

The ability of plasma Aβ42/Aβ40 to detect early Aβ-PET positivity was determined with logistic regression models and ROC curve analysis.

Results

Aβ42/Aβ40 plasma ratio showed a significant negative correlation with FBB-PET CL values (rho=-0.390; p=1.16e-08). Additionally, Aβ-PET(+) individuals presented lower Aβ42/Aβ40 levels than subjects in the Aβ-PET(-) group (p=5.93e-12).

Aβ42/Aβ40 plasma ratio detected abnormal brain Aβ status with AUC of 0.87 and accuracy of 82%. After adjusting for age, sex and APOE4 genotype, AUC further improved to 0.89.

Conclusions

This novel ABtest-MS method accurately identifies early amyloid deposition in individuals at risk of AD. On-going analysis of the follow-up visits of the FACEHBI cohort will assess the ability of baseline Aβ42/Aβ40 plasma ratio as measured with ABtest-MS to predict future conversion to mild cognitive impairment and/or dementia.

Aims

Background

Proteomics studies showed differential expression of numerous proteins in dementias, but have barely led to novel biomarker tests for clinical use. The Marie Curie MIRIADE-project is designed to experimentally evaluate development strategies to accelerate the validation and implementation of novel biomarkers. We hypothesize that a major hurdle is the use of different technologies for discovery (mass spectrometry) compared to the technology for large scale validation (immunoassays). This problem might be overcome by using novel immune-based platforms for both discovery and validation (=no technology-switch). In this study we investigated this alternative development pipeline with the aim to identify and validate specific cerebrospinal fluid (CSF) biomarkers for frontotemporal dementia (FTD).

Methods

Method

We analysed >600 proteins using antibody-based Olink proteomics in 394 CSF samples from controls (n=195) and FTD patients (n=199, including 34 FTLD-Tau and 54 FTLD-TDP-subtypes). Nested (generalized) linear modeling was applied to define differences and diagnostic classification models.

Results

We identified >50 CSF proteins dysregulated in FTD and one between FTLD subtypes. We detected biomarker signatures for specific diagnosis of FTD (8-9 proteins, AUC>0.85), but not to discriminate FTLD subtypes. We next developed and successfully verified custom multiplex panels for these proteins, which are currently being clinically validated in independent cohorts.

Conclusions

The discovery and verification of the novel biomarker panels to identify FTD using immunoassays shows the effectiveness of the MIRIADE strategy. The pipeline could be used for the development of other biomarkers to improve the biological diagnosis and patient care for dementias.

Aims

To study the diagnostic performance of plasma phospho-tau231 (p-tau231) to identify mild cognitive impairment (MCI) and dementia due to Alzheimer’s disease (AD) in a memory clinic setting and compare its utility with plasma neurofilament light chain (NfL).

Methods

We retrospectively enrolled consecutive patients investigated for a cognitive complaint at the Cognitive Neurology Centre, Hospital Lariboisière, France, including 30 neurological controls, 63 patients with MCI due to AD (AD-MCI), 103 with AD dementia, 72 with MCI due to other disorders (MCI-other) and 35 with non-AD dementia (other dementia). Plasma p-tau231 and plasma NfL were measured using the Simoa platform. We used receiver operating characteristic curves and DeLong test to study performance of plasma markers alone and in combination with clinical measures for diagnosis of AD.

Results

Plasma p-tau231 and plasma NfL were increased in AD patients from the MCI stage (both p<0.0001). Alone, plasma p-tau231 and plasma NfL discriminated patients with AD-MCI from controls with similar performance (respectively AUC=0.86 and AUC=0.80, p_DeLong=0.283). A clinical model (age, sex and MMSE) yielded an AUC of 0.87 but did not differ from plasma p-tau231 or plasma NfL (respectively, p_DeLong=0.283 and p_DeLong=0.063). Plasma p-tau231 distinguished AD-MCI from MCI-other (AUC=0.78), as well as AD dementia from other dementia individuals (AUC=0.85) both better than plasma NfL (respectively, AUC=0.68 and AUC=0.73). Combining plasma p-tau231, plasma NfL and clinical variables did not significantly outperform plasma p-tau231 alone.

Conclusions

Plasma p-tau231 is effective for the diagnosis of AD at MCI stage in a memory clinic setting, and displays higher diagnostic performance than plasma NfL.

Aims

Cerebral amyloid angiopathy (CAA) is characterized by the accumulation of amyloid-beta (Aβ) in the cerebral vasculature and has been associated with intracerebral haemorrhages and dementia. CAA diagnosis is based on the identification of lobar (micro-)bleeds and cortical superficial siderosis on MRI. However, these markers are indirect and reflect only late-stage manifestations. We evaluated the diagnostic potential of the peptides Aβ38, Aβ40, Aβ42 and Aβ43 in cerebrospinal fluid (CSF) to discriminate CAA from controls.

Methods

CSF Aβ peptides were quantified in a discovery cohort of sporadic CAA patients (n=27; sCAA) and age-matched controls (n=40), a validation cohort of sCAA patients (n=43) and age-matched controls (n=36), and in a cohort of Dutch-type hereditary CAA (D-CAA) (asymptomatic; n=10, symptomatic; n=12) and age-matched controls (n=22 and n=28, respectively). All comparisons were adjusted for age and sex by linear regression.

Results

In the discovery cohort, we found decreased levels of Aβ38 (p=0.021), Aβ40 (p=0.031), Aβ42 (p<0.0001) and Aβ43 (p<0.0001) in sCAA patients. In the validation cohort, we could confirm decreased levels of Aβ38 (p<0.0001), Aβ40 (p<0.0001), Aβ42 (p<0.0001) and Aβ43 (p<0.0001) in sCAA patients. We also found decreased levels of Aβ38 (p=0.002), Aβ40 (p=0.031), Aβ42 (p<0.0001) and Aβ43 (p=0.001) in asymptomatic D-CAA patients, and in symptomatic D-CAA patients (Aβ38: p<0.0001, Aβ40: p<0.0001, Aβ42: p<0.0001, Aβ43: p<0.0001).

Conclusions

Aβ peptides may have strong potential as biomarkers for the diagnosis of sCAA and D-CAA. In addition, decreased levels of Aβ peptides in asymptomatic D-CAA patients indicate that abnormal Aβ deposition is already present in the preclinical stage.

Aims

To quantify proteome-wide changes in CSF across normal controls and individuals diagnosed with AD.

Methods

Using quantitative liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) and multiplex tandem mass tags (TMT) we comparatively analyzed the CSF proteomes from 300 samples collected fom participants in the Emory Goizueta ADRC and affiliated cohorts, including 150 cognitively normal individuals (avg. MoCA=26.5) and 150 subjects with clinical diagnoses of AD (avg. MoCA=14.8). Samples across these two diagnostic groups were matched for age and sex. AD cases demonstrated low Aβ_1–42 (299.8 pg/mL) and high tau (114.8 pg/mL) levels compared to controls (546.8 pg/mL and 54.5 pg/mL, respectively) as measured by ELISA. Following regression for age and sex, both differential expression and co-expression analysis were used to identify CSF proteins related to Aβ_1–42 and tau concentration and clinical phenotypes.

Results

A total of approximately 1,800 proteins were quantified across all 300 samples. Synaptic proteins (GAP43, YWHAZ) and proteins involved in energy metabolism (PKM, PGAM1) were increased in AD and correlated strongly with Aβ_1–42 and tau levels by ELISA. Proteins with links to microglia and neuroinflammation (SPP1, MIF) were also increased in AD and correlated to total tau levels. Tau levels were also identified and quantified by mass spectrometry and correlated to tau ELISA measurements. Co-expression network analyses revealed modules of CSF proteins that strongly overlap with protein networks in AD postmortem brain tissue.

Conclusions

These results reveal robust CSF protein biomarkers highly reflective of underlying disease pathology in brain with potential to assess disease progression and therapeutic target engagement.

Aims

The AT(N) framework has been widely used to classify Alzheimer’s disease (AD). However, the biological network and pathway relating to the AT(N) framework remain incompletely understood. The objective of this study is to perform an in depth proteomic profiling of cerebrospinal fluid (CSF) and plasma from AD patients to gain a better understanding of its pathogenesis.

Methods

We used mass spectrometry to measure 2535 proteins in CSF samples in 371 subjects (125 controls, 152 mild cognitive impairment [MCI], and 94 AD) selected from the EMIF-AD study. SOMAscan platform was used to measure 4001 proteins in plasma in 972 subjects (372 controls, 409 MCI, and 191 AD) from the same study. We firstly performed both linear regression and protein co-expression network to identify differentially expressed proteins and co-expressed protein modules associated with the AT(N) framework in CSF. We then validated such associations in plasma samples.

Results

In CSF, we identified three modules, all of which were significantly associated with the AT(N) framework. Furthermore, two of these modules were preserved in plasma proteomics and the AT(N) associations were also maintained. In addition, we identified 245 proteins in CSF that were significantly associated with the AT(N) framework after multiple correction. Of these, 21 proteins were also significantly associated with the AT(N) framework in plasma.

Conclusions

We identified both protein modules and single proteins relating to the AT(N) framework in CSF and then validated such associations in plasma. These proteins provide tractable targets for further biomarkers and mechanistic studies of Alzheimer’s pathophysiology.

Aims

Currently, molecular diagnosis of Alzheimer`s disease (AD) is based on cerebrospinal fluid (CSF) biomarkers. It is performed once first symptoms appear and neuronal damage is already irreversible. Therefore, an accurate early diagnosis is one of the great challenges in the AD field. Growing interest has been focused in plasma-derived exosomes (pEXO), which have shown to reflect the molecular alterations occurring at a central level in the different stages of AD. Thus, the aim of this project is to evaluate the diagnostic potential of pEXOs through the analysis of their biomarkers´ cargo in an early-onset mild cognitive impairment (EOMCI) cohort.

Methods

Data from 80 participants with EOMCI were analyzed. pEXOs were isolated by using ultracentrifugation method. Concentration of 184 biomarkers in of CSF, plasma and pEXOs were analyzed by using ProSeek® multiplex immunoassay of Olink Proteomics. Neuroimaging were performed by brain MRI. CSF AD biomarkers were used to determine the Aβ status of the participants.

Results

pEXOS from EOMCI Aβ+ participants showed an increased total protein amount compared to EOMCI Aβ-. Several biomarkers showed different levels in CSF and pEXOs samples of EOMCI Aβ+ vs EOMCI Aβ-, but not in plasma. Several biomarker of EOMCI Aβ+ pEXOs samples were significantly correlated with some brain areas and p-Tau levels, but not in EOMCI Aβ-. Further studies are needed to validate these results.

Conclusions

pEXOs could have the potential to be a suitable tool for the early diagnosis of AD.Conclusion: pEXOs could have the potential to be a suitable tool for the early diagnosis of AD.