CSF proteome in multiple sclerosis subtypes related to brain lesion transcriptomes

Maria L Elkjaer,Lene Wermuth,Tobias Sejbaek,Anja Hviid Simonsen,Richard Reynolds,Jan Baumbach,Zsolt Illes,Peter Høgh,Tim Kacprowski,Romain Marignier,Martin R Larsen,Pernille Lassen,Finn Sellebjerg,Alyaa Yakut Rashid,Helle H Nielsen,Arkadiusz Nawrocki

doi:10.1038/s41598-021-83591-5

Abstract

To identify markers in the CSF of multiple sclerosis (MS) subtypes, we used a two-step proteomic approach: (i) Discovery proteomics compared 169 pooled CSF from MS subtypes and inflammatory/degenerative CNS diseases (NMO spectrum and Alzheimer disease) and healthy controls. (ii) Next, 299 proteins selected by comprehensive statistics were quantified in 170 individual CSF samples. (iii) Genes of the identified proteins were also screened among transcripts in 73 MS brain lesions compared to 25 control brains. F-test based feature selection resulted in 8 proteins differentiating the MS subtypes, and secondary progressive (SP)MS was the most different also from controls. Genes of 7 out these 8 proteins were present in MS brain lesions: GOLM was significantly differentially expressed in active, chronic active, inactive and remyelinating lesions, FRZB in active and chronic active lesions, and SELENBP1 in inactive lesions. Volcano maps of normalized proteins in the different disease groups also indicated the highest amount of altered proteins in SPMS. Apolipoprotein C-I, apolipoprotein A-II, augurin, receptor-type tyrosine-protein phosphatase gamma, and trypsin-1 were upregulated in the CSF of MS subtypes compared to controls. This CSF profile and associated brain lesion spectrum highlight non-inflammatory mechanisms in differentiating CNS diseases and MS subtypes and the uniqueness of SPMS.

Highlights

Mechanisms that culminate in the progressive phase may provide a more individualized treatment approach and postpone the secondary phase[6]
By using F-test based feature selection, 11 proteins were able to distinguish the diseasegroups (Fig. 2a). These 11 proteins were used to conduct a linear discriminant analysis (LDA) that focuses on maximizing the separability among the known disease groups and healthy controls: there was no overlap between the different disease groups, and no influence of the technical batch effect (Fig. 2b)
Almost half of them were involved in axon-related processes (RTN4R, CNTNAP4, ADAM22, PCSK1N, NRXN1), which could indicate that the neurodegenerative mechanisms may be different between these brain diseases

Summary

Introduction

Mechanisms that culminate in the progressive phase may provide a more individualized treatment approach and postpone the secondary phase[6]. Hypothesis-generating exploratory omics are effective tools for revealing novel molecular pathways and quantifying differentially expressed molecules to identify multiple markers that may predict disease outcomes. We used a comprehensive two-stage approach, with an untargeted and a quantitative targeted method to characterize the molecular landscape of the CSF in different phases of MS. Based on the different protein abundances in 169 CSF samples, we: (i) clustered the diseases and MS subtypes to create the CSF proteomic landscapes across MS subtypes and an array of controls; (ii) selected 299 proteins that were quantified in 170 individual CSF of the MS subgroups and controls to identify novel unique protein markers across diseases; and (iii) linked the unique CSF proteins with MS brain lesion transcriptome signatures using multi-omics comparison across compartments and information levels[10] (Fig. 1)

Methods

Results

Conclusion