Configuration, optimization and evaluation of a novel instrumental platform for automated SPE-LC-MS/MS analysis of drugs in whole blood

The thesis describes the configuration, optimization and evaluation
of a novel instrumental platform for fully automated SPE-LC-MS/MS
analysis of small molecules, such as drugs, in whole blood. The
immunosuppressant Cyclosporine A was chosen as a model analyte, as
this drug is predominantly bound to erythrocytes. First,
anticoagulated blood is converted into so-called Cell-Disintegrated
Blood (CDB) by heat-shock or cryogenic treatment. CDB represents a
homogenous blood sample and consists of subcellular particles which
do not sediment on standing and do not clog capillaries, sieves or
HPLC column packings. For in-line treatment of anticoagulated whole
blood, i.e. generation of CDB, a sample mixing unit, two special
liquid handling units and two home-made sample processing modules
were embedded into a XYZ-autosampler. The module for heat-shock
treatment consists of a stainless-steel capillary jacketed with a
heating sleeve. Under optimal conditions for sampling and in-line
processing of 20 µL of whole blood, it takes 13 seconds at 75 °C to
generate CDB. The latter is stored in a holding loop before further
treatment. For cryogenic treatment of a blood sample, a
stainless-steel processing needle with a large inner diameter was
installed in one of the liquid handling units. The autosampler was
programmed to introduce the processing needle containing the blood
sample (40 µL) into a stand-pipe, which is located in a
thermo-flask filled with liquid nitrogen. The processing needle
therefore contacts liquid nitrogen and the blood sample is
snap-frozen. Optimal conditions were found to be 10 seconds for
snap-freezing at -196 °C and 60 seconds for thawing at room
temperature. A CDB sample obtained either by heat-shock or
cryogenic treatment is further processed by being pumped via a
switching-valve through an in-line filter to retain cell nuclei and
“cell debris”. It was found that a depth filter packed with
spherical hydrophilic silica is optimal. This filter allows at
least 200 analysis cycles before it has to be replaced. Next, the
CDB sample is pumped on-line via another switching-valve through a
SPE column (50 x 0.5 mm ID) at a high flow rate. Due to the special
packing material and the very small inner diameter, a high linear
flow velocity is achieved and turbulent flow is generated. By this,
high-molecular matrix components such as proteins are eluted in the
void volume to waste. The low-molecular weight target analyte
Cyclosporine A and the Internal Standard Cyclosporine D are
retained and extracted by reversed phase partitioning
chromatography (RPC). After fractionation of CDB on the SPE column,
the analyte and the IS are transferred to a series-connected
analytical column and separated from residual matrix components by
RPC. Finally, the analyte is detected by a tandem mass spectrometer
applying electrospray ionization (ESI) and multiple reaction
monitoring (MRM). The optimized method has a total analysis time of
less than 11 minutes. The analytical procedure and the instrumental
platform were validated for heat-shock treated blood samples with
respect to linearity, range (10 - 1000 ng/mL), lower limit of
quantitation (10 ng/mL), intra-day and inter-day accuracy and
precision, as well as matrix-independent and matrix-dependent
recovery (around 100 %). It was shown that the electrospray induced
ionization is suppressed by approximately 25 %. These matrix
effects, however, can be totally compensated for by the addition of
an Internal Standard, i.e. Cyclosporine D. A comparison with a
semi-automated SPE-LC-MS/MS method, established in the Institute,
revealed a very good agreement. This was shown by Passing and
Bablok plots. The robustness of the fully automated SPE-LC-MS/MS
analysis platform was monitored during 500 consecutive analysis
cycles with heat-shock treated blood samples. The relative standard
deviation for the signal response was 15.6 % for Cyclosporine A and
15.2 % for Cyclosporine D. The back pressure of the total system
rose only by 52 bar. These findings show that, despite its
instrumental and chromatographic complexity, the described analysis
platform fulfills the prerequisites to be used in routine
clinical-chemical analysis.