Characterization of Cerebrospinal Fluid via Data-Independent Acquisition Mass Spectrometry

Abstract

Cerebrospinal fluid (CSF) is in direct contact with the brain and serves as a valuable specimen to examine diseases of the central nervous system through analyzing its components. These include the analysis of metabolites, cells as well as proteins. For identifying new suitable diagnostic protein biomarkers bottom-up data-dependent acquisition (DDA) mass spectrometry-based approaches are most popular. Drawbacks of this method are stochastic and irreproducible precursor ion selection. Recently, data-independent acquisition (DIA) emerged as an alternative method. It overcomes several limitations of DDA, since it combines the benefits of DDA and targeted methods like selected reaction monitoring (SRM). We established a DIA method for in-depth proteome analysis of CSF. For this, four spectral libraries were generated with samples from native CSF ( n = 5), CSF fractionation (15 in total) and substantia nigra fractionation (54 in total) and applied to three CSF DIA replicates. The DDA and DIA methods for CSF were conducted with the same nanoLC parameters using a 180 min gradient. Compared to a conventional DDA method, our DIA approach increased the number of identified protein groups from 648 identifications in DDA to 1574 in DIA using a comprehensive spectral library generated with DDA measurements from five native CSF and 54 substantia nigra fractions. We also could show that a sample specific spectral library generated from native CSF only increased the identification reproducibility from three DIA replicates to 90% (77% with a DDA method). Moreover, by utilizing a substantia nigra specific spectral library for CSF DIA, over 60 brain-originated proteins could be identified compared to only 11 with DDA. In conclusion, the here presented optimized DIA method substantially outperforms DDA and could develop into a powerful tool for biomarker discovery in CSF. Data are available via ProteomeXchange with the identifiers PXD010698, PXD010708, PXD010690, PXD010705, and PXD009624.

Reproducibility, Specificity and Accuracy of Relative Quantification Using Spectral Library-based Data-independent Acquisition

“…Most of them demonstrated an improved peptide and protein identification rate using spectral library-based DIA (e.g. (24,25)) even in single shot analyses (16). In addition, a few studies showed an improved quantification reproducibility based on the coefficient of variation (CV) between technical replicates (1,26).…”

Blood Contamination in CSF and Its Impact on Quantitative Analysis of Alpha-Synuclein

“…Digestion was performed as reported in Reference [6] with some modifications. Measures of 20 µL CSF were mixed with 50 µL 0.1% Rapigest™ (Waters, Eschborn, Germany) in 50 mM NH 4 HCO 3 and dithiothreitol (DTT) (final concentration 5 mM) and heated at 60 • C for 30 min, followed by cooling down at room temperature for 10 min.…”

“…MS analysis enables the simultaneous detection and quantification of several hundred up to thousands of proteins in complex samples [22,23]. For CSF proteome analysis, we used an established workflow, which results in the identification of over 700 proteins in CSF samples [6]. We examined the identification pattern of the three blood proteins hemoglobin (HB) (subunit alpha and beta), carbonic anhydrase 1 (CAH1), and catalase (CATA), which have previously been used as blood contamination markers in CSF proteomics [21].…”

“…Consequently, the presence of only minimal blood portions in generated CSF samples will lead to an increased overall aSyn level, and the actual level of aSyn released from brain due to neurodegenerative processes cannot be assessed. Moreover, blood contamination has a substantial effect on the CSF protein composition as well as detection of low-abundance CSF-and/or brain-specific proteins [3,6]. These can be masked by highly abundant blood proteins, which makes it difficult to identify disease-relevant molecules with presumably low quantities.…”

In Vivo Large Scale Mapping Of Protein Turnover In The Human Cerebrospinal Fluid

“…We used as reference plasma proteome the union of the plasma proteome database (Nanjappa et al, 2014) (PPD) and proteins annotated as plasma in UniProtKB/Swissprot. Similarly, we defined a reference CSF proteome by taking the union of CSF proteins reported in two comprehensive lists (Barkovits et al, 2018;Fernandez-Irigoyen et al, 2015). Potentially novel CSF proteins were (HUGO symbols for short) CEP290, FBXW10, KSR2, LOX, SH2D3A, SIK2, SPEN, TMEM212 and immunoglobulins (IGHV3-43, IGHV3-74, IGKV3D-15, IGKV4-1).…”