Metabolites (Jan 2022)

Different Metabolomic and Proteomic Profiles of Cerebrospinal Fluid in Ventricular and Lumbar Compartments in Relation to Leptomeningeal Metastases

  • Ji-Woong Kwon,
  • Ji Hye Im,
  • Kyue-Yim Lee,
  • Byong Chul Yoo,
  • Jun Hwa Lee,
  • Kyung-Hee Kim,
  • Jong Heon Kim,
  • Sang Hoon Shin,
  • Heon Yoo,
  • Ho-Shin Gwak

DOI
https://doi.org/10.3390/metabo12010080
Journal volume & issue
Vol. 12, no. 1
p. 80

Abstract

Read online

The different molecular profiles of cerebrospinal fluid (CSF) between ventricular and lumbar compartments remain elusive, especially in the context of leptomeningeal metastasis (LM), which affects CSF flow. We evaluated CSF metabolomic and proteomic profiles based on the compartments and the diagnosis of spinal LM, proved by MRI from 20 paired ventricular and lumbar CSF samples of LM patients, including 12 spinal LM (+) samples. In metabolome analysis, 9512 low-mass ions (LMIs) were identified—7 LMIs were abundant in all lumbar versus paired ventricular CSF samples, and 3 LMIs were significantly abundant in all ventricular CSF. In comparisons between spinal LM (+) CSF and LM (−) CSF, 105 LMIs were discriminative for spinal LM (+) CSF. In proteome analysis, a total of 1536 proteins were measured. A total of 18 proteins, including complement C3, were more highly expressed in all lumbar CSF, compared with paired ventricular CSF, while 82 proteins, including coagulation factor V, were higher in the ventricular CSF. Of 37 discriminative proteins, including uteroglobin and complement component C8 gamma chain, 4 were higher in all spinal LM (+) CSF versus spinal LM (−) CSF. We further evaluated metabolic pathways associated with these discriminative proteins using the Gene Ontology database. We found that 16/17 spinal LM (+) pathways, including complement activation, were associated with lumbar discriminative proteins, whereas only 2 pathways were associated with ventricular-discriminative proteins. In conclusion, we determined that metabolite and protein profiles differed between paired lumbar and ventricular CSF samples. The protein profiles of spinal LM (+) CSF showed more similarity with the lumbar CSF than the ventricular CSF. Thus, we suggest that CSF LMIs and proteins could reflect LM disease activity and that LM-associated differences in CSF are more likely to be present in the lumbar compartment.

Keywords