The CSF proteome has been less studied than that of serum, plasma and many other human biological fluids and tissues. Nonetheless, CSF is a very informative sample with close relation to the brain and a precious source for biomarkers and the exploration of molecular mechanism in humans. Its reduced use links primarily to the invasiveness of its collection performed with a lumbar puncture.
The analytical challenges faced when performing proteomic analysis of CSF are numerous: i) the limited concentration of proteins (less than 1% of that in plasma), ii) the variable total protein concentration between individuals, iii) the important dynamic range in the concentrations of proteins (about 8 orders of magnitude), and iv) the high amount of salts present in CSF.
While several of these complications are alleviated today with dedicated sample preparation and improved liquid chromatography and mass spectrometry instrumentation, an important parameter to consider is the overall time needed for the proteomic analysis of CSF samples. As such, some trade-offs remain between obtaining deep CSF proteome coverages and analysing large numbers of CSF samples at the same time. We reasoned that two dedicated proteomic workflows are still needed to fully satisfy those two dimensions.
We then developed a general protocol for higher throughput proteomic analyses of CSF samples with optional steps to expand into deeper proteome coverage when needed. These optional steps rely on depletion of abundant proteins and sample pre-fractionation before liquid chromatography tandem mass spectrometry. These steps significantly increase the analysis time (about four times), without many possibilities for parallel sample treatment. However, they should allow the operator to identify more than 3000 proteins in a CSF sample.
When performing a clinical research study, a sufficient number of samples is required to empower the relevant identification of biomarkers for instance. In this case, inclusion of those above-mentioned optional steps for the proteomic analysis of CSF is limited because they are too time consuming. Our core protocol accommodates the preparation of 96 CSF samples in parallel to increase throughput. We use a liquid handling workstation to improve reproducibility and reduce manual repetitive tasks and the risk of errors; such an automated platform has required a careful validation for robust sample preparation over several months. Nonetheless, all steps are transposable at the bench when such robotic system is not available. This neat workflow should allow robust identification/quantification of hundreds of proteins in hundreds to thousands of CSF samples.
In summary, the proteomic protocol described in chapter 9 of the present Methods in Molecular Biology volume is versatile enough to let the user choose the best compromise between throughput and depth of analysis. Because, as of today, complete coverage of the CSF proteome is not attainable at a sufficient throughput required for clinical research proteomics.
Note: Image by John Hain from Pixabay.