Unravelling the plasma proteome: Pioneering biomarkers for differential dementia diagnosis

Abstract

Introduction: Diagnosing Alzheimer's disease (AD) is challenging due to overlapping symptoms with other dementias and the invasiveness of current biomarkers. This study introduces the NULISA platform, a novel proteomics technology, to evaluate diagnostic accuracy of known biomarkers and uncover novel biomarkers underlying different dementias.

Methods: We analyzed plasma and cerebrospinal fluid (CSF) samples from 248 participants diagnosed with Alzheimer's disease (AD), dementia with Lewy bodies (DLB), frontotemporal dementia (FTD), and mild cognitive impairment (MCI). Plasma biomarkers were evaluated using regression models, receiver operating characteristics curve (ROC) analysis, and pathway enrichment.

Results: Plasma phosphorylated Tau217 (pTau217) demonstrated the highest diagnostic accuracy for AD, DLB, and FTD (area under the curve [AUCs]: 0.9, 0.84, and 0.79, respectively). CXCL1 (fractalkine), synaptosomal-associated protein 25 (SNAP25), triggering receptor expressed on myeloid cells 1 (TREM1), β-synuclein, and tyrosine kinase (TEK) are expressed differently in DLB and FTD than AD. Ingenuity pathway analyses revealed astrocytic, synaptic, and inflammatory pathways as shared and distinct mechanisms across these dementia types.

Conclusion: Our findings establish plasma pTau217 as a robust diagnostic marker. This study provides new plasma biomarkers for differential diagnosis of dementias with a noninvasive method.

Highlights: Plasma pTau217 showed high diagnostic accuracy for AD, DLB, and FTD. CXCL1, SNAP25, TREM1, β-synuclein, and TEK are novel markers distinguishing other dementias from AD. Noninvasive plasma biomarkers enable diagnosis and differentiation of dementias.

Keywords: AD; DLB; FTD; dementia; pTau; plasma biomarkers.

FIGURE 1

Violin plots illustrating plasma biomarker levels of pTau217, Aβ42, NfL, and GFAP across patients with AD, DLB, FTD, and MCI. Each violin plot shows the distribution of normalized NPQ values for each biomarker within the diagnostic groups, with mean or median values highlighted by red squares. Statistically significant differences between diagnostic groups were assessed using adjusted p‐values (p < 0.05, FDR‐adj indicated) from group comparisons. Aβ42, amyloid‐beta 42; AD, Alzheimer's disease; DLB, dementia with Lewy bodies; FDR‐adj, post hoc pairwise Dunn's test with false discovery rate adjustment; FTD, frontotemporal dementia; GFAP, glial fibrillary acidic protein; MCI, mild cognitive impairment; NfL, neurofilament light chain; NPQ, normalized protein quantification; pTau‐217, phosphorylated tau‐217.

FIGURE 2

ROC curves for CSF amyloid status, AD, DLB, and FTD diagnosis. The ROC curves display the diagnostic performance of a model combining plasma pTau217 levels, age, and sex for predicting AD, DLB, FTD diagnoses, and amyloid status in CSF, separately. The AUC for AD, DLB, FTD, and amyloid positivity are 0.9, 0.82, 0.79, and 0.87, respectively. AD, Alzheimer's disease; AUC, area under the curve; CSF, cerebrospinal fluid; DLB, dementia with Lewy bodies; FTD, frontotemporal dementia; ROC, receiver operating characteristic.

FIGURE 3

Correlation heatmap between CSF and plasma biomarkers for all cohort. The heatmap displays the correlation between CSF biomarkers and MMSE on the y‐axis and plasma biomarkers on the x‐axis. The numbers represent Spearman correlation coefficients, indicating the strength and direction of the relationships between the variables. Blue indicates positive correlations, and red indicates negative correlations. Only FDR corrected p‐value < 0.05 are showed. Aβ42, amyloid‐beta 42; Aβ40, plasma amyloid‐beta 40; CSF, cerebrospinal fluid; FDR, false discovery rate; MMSE, Mini‐Mental State Examination; NfL, neurofilament light chain; p181, phosphorylated tau‐181; p217, phosphorylated tau‐217; p231, phosphorylated tau‐231; pTau181, phosphorylated tau 181; pTau/Aβ42, ratio of phosphorylated tau 181 to Aβ42; Aβ42/Aβ40, ratio of Aβ42 to Aβ40; p181/Aβ42, ratio of phosphorylated tau 181 to Aβ42; p217/Aβ42, ratio of phosphorylated tau 217 to Aβ42; p231/Aβ42, ratio of phosphorylated tau 231 to Aβ42; tTau, total tau; tTau/Aβ42, ratio of total tau to Aβ42.

FIGURE 4

Multinomial logistic regression coefficients for biomarkers differentiating DLB and FTD from AD. Biomarkers that above zero are increased in DLB compared to AD. Biomarkers that below zero are increased in AD compared to DLB or FTD. AD, Alzheimer's disease; BACE1, beta‐site amyloid precursor protein cleaving enzyme 1; CST3, cystatin C; CX3CL1, chemokine (C‐X3‐C motif) ligand 1 (fractalkine); DLB, dementia with Lewy bodies; FTD, frontotemporal dementia; GDI1, GDP dissociation inhibitor 1; GFAP, glial fibrillary acidic protein; IGFBP7, insulin‐like growth factor binding protein 7; IL15, interleukin‐15; MAPT, microtubule‐associated protein tau; pTau181, phosphorylated Tau‐181; pTau217, phosphorylated Tau‐217; pTau231, phosphorylated Tau‐231; PTN, pleiotrophin; SNAP25, synaptosomal‐associated protein 25; SNCB, beta synuclein; TEK, tyrosine kinase; TREM1, triggering receptor expressed on myeloid cells 1.

FIGURE 5

Volcano plot of biomarkers across the AD continuum based on ordinal regression. Red biomarkers are significantly increasing with disease stage. AD, Alzheimer's disease; GFAP, glial fibrillary acidic protein; MAPT, microtubule‐associated protein tau; NEFL, neurofilament light chain; PGF, placental growth factor; pTau181, phosphorylated tau‐181; pTau217, phosphorylated tau‐217; pTau231, phosphorylated tau‐231; SNAP25, synaptosomal‐associated protein 25.

FIGURE 6

RF ROC curve for dementia diagnosis, differentiated biomarkers. The ROC curves display the diagnostic performance of a regression model combining only differently expressed plasma biomarkers, age, and sex for predicting AD, DLB, FTD diagnoses. The AUC for AD, FTD, and DLB are 0.88, 0.87, and 0.8, respectively. AD, Alzheimer's disease; AUC, area under the curve; DLB, dementia with Lewy bodies; FTD, frontotemporal dementia; RF, random forest; ROC, receiver operating characteristic.