PLoS ONE (Jan 2007)

CSF metabolic and proteomic profiles in patients prodromal for psychosis.

  • Jeffrey T-J Huang,
  • F Markus Leweke,
  • Tsz M Tsang,
  • Dagmar Koethe,
  • Laura Kranaster,
  • Christoph W Gerth,
  • Sonja Gross,
  • Daniela Schreiber,
  • Stephan Ruhrmann,
  • Frauke Schultze-Lutter,
  • Joachim Klosterkötter,
  • Elaine Holmes,
  • Sabine Bahn

DOI
https://doi.org/10.1371/journal.pone.0000756
Journal volume & issue
Vol. 2, no. 8
p. e756

Abstract

Read online

BACKGROUND: The initial prodromal state of psychosis (IPS) is defined as an early disease stage prior to the onset of overt psychosis characterized by sub-threshold or more unspecific psychiatric symptoms. Little is known regarding the biochemical changes during this period. METHODOLOGY/PRINCIPAL FINDINGS: We investigated the metabolic/proteomic profiles of cerebrospinal fluid (CSF) of first-onset drug naïve paranoid schizophrenia patients (n = 54) and individuals presenting with initial prodromal symptoms (n = 24), alongside healthy volunteers (n = 70) using proton nuclear magnetic resonance ((1)H-NMR) spectroscopy and surface enhanced laser desorption ionization (SELDI) mass spectrometry, respectively. Partial least square discriminant analysis (PLS-DA) showed that 36%/29% of IPS patients displayed proteomic/metabolic profiles characteristic of first-onset, drug naïve schizophrenia, i.e., changes in levels of glucose and lactate as well as changes in a VGF-derived peptide (VGF23-62) and transthyretin protein concentrations. However, only 29% (n = 7) of the investigated IPS patients (who to date have been followed up for up to three years) have so far received a diagnosis of schizophrenia. The presence of biochemical alterations in the IPS group did not correlate with the risk to develop schizophrenia. CONCLUSIONS/SIGNIFICANCE: Our results imply that schizophrenia-related biochemical disease processes can be traced in CSF of prodromal patients. However, the biochemical disturbances identified in IPS patients, at least when measured at a single time point, may not be sufficient to predict clinical outcome.