Sample Preparation in Global Metabolomics of Biological Fluids and Tissues - Proteomic and Metabolomic Approaches to Biomarker Discovery - page 55

Low speed pre-
centrifugation
1000-3000 rcf 5 min

High-speed
centrifugation
14000 rcf 5 min

0.20 µm filtration or
sodium azide addition
(optional)

Dilution with water
1:1 to 1:10 ratio

Urine

Storage at or below
-25°C

Solvent precipitation with cold
methanol or methanol/ethanol
1:4 ratio

Evaporation/reconstitution
(optional)

Anticoagulant addition and
centrifugation

Plasma

Storage at -80°C

Check for blood
contamination by
measuring hemoglobin
or apolipoprotein B100

Immediate low-speed
centrifugation
3000 rpm 10 min

Solvent precipitation
with methanol
1:4 or 1:10 ratio

Evaporation/reconstitution

CSF

Storage at -80°C

Uniphasic or
biphasic solvent
extraction with cold
solvent

Rinse to remove
blood

Liquid nitrogen
quenching

Homogenization at
low temperature

Evaporation/
reconstitution

Tissue

Storage at -80°C

FIGURE 3 Common recommended protocols for the analysis of urine, plasma/serum, CSF, and biological tissues in
combination with LC-MS. Similar protocols may be employed in combination with GC-MS analysis after the inclusion of an
appropriate derivatization step.

provided by solvent precipitation are (1) ioniza-
tion suppression, which can occur due to the
competitive nature of the electrospray ionization
(ESI) process if using LC-MS detection; (2) signif-
icantly

appropriate anticoagulant (heparin, citrate, eth-
ylenediaminetetraacetic acid [EDTA]) to prevent

fibrinogen clotting cascade followed by centrifu-
gation to isolate the supernatant. Serum is
obtained after the natural process of clotting takes
place without the addition of anticoagulants, fol-
lowed by centrifugation to isolate the superna-
tant. The clotting process needed to obtain
serum is carried out for 30 to 60 min at room
temperature, which (1) increases the likelihood
of enzymatic conversion and other degradation
processes to occur at least for some of the more
labile metabolites and (2) facilitates possible loss
of some metabolites during clot formation/
precipitation. 44,45 Clotting on ice and strictly stan-
dardizing serum e clot contact time can eliminate
bias introduced by residual enzymatic activity
and reduce the transport of metabolites between
different compartments to address the

400
samples 41 ); and (3) contamination of the ion
source, which necessitates frequent MS source
cleaning. In fact, several publications to date
advocate metabolomic sample runs of no more
than 100 samples to ensure good data quality,
with source cleaning between each data set. 41,42
Furthermore, the issues of metabolite losses due
to coprecipitation with proteins and/or poor
solubility in selected extraction solvent have not
yet been adequately investigated. 43

reduced column lifetime (

w

Plasma versus Serum in Global
Metabolomics
The first step in development of a blood meta-

bolomic sample preparation method is the selec-
tion of plasma versus serum, and both matrices
have been successfully used to date. The use of
whole blood is generally avoided due to its
poor stability, heterogeneous nature, and poor
compatibility with the majority of common
sample preparation methods. Plasma is obtained
from whole blood by the addition of

rst

concern. 46
Several recent studies comparing metabolite
pro les obtained from plasma and serum found

similar metabolite coverage (correlation r ¼ 0.81

cant
differences detected for only a small number of
the metabolites. 45,47,48 The reported differences
between serum and plasma include (1) high levels

0.10 for plasma versus serum), with signi

the