MS Condition The ion source is heated (e.g., 400 50 C) with gas flows. Ion
spray voltage is usually high (typically 5000 eV in +ESI or4500 eV inESI).
The parameters of multiple reaction monitor (MRM) for quantification are
varied upon the analytes, with normally the collision energy (CE) of 20–50 eV.
In brief, the short column (10–50 mm) and sharp mobile phase gradients or
isocratic elution with high organic solvent content (30%), facilitating a quick
HPLC run (2–10 min), are often used for the detection of single or a few
compounds, as is the case for microsomal stability studies. On the contrary, the
conventional column (50–150 mm) and the smooth gradients, resulting in a
slow HPLC run (15–30 min), are preferably applied for the detection of several
analytes such as both parent and metabolites. Moreover, because of the
specificity of LC/MS/MS (MRM) methods, it is possible to quantify several
analytes in the injection samples pooled from the different reaction mixtures
during a single LC run (Hakala et al., 2005).
13.5 PREDICTION OF HUMAN HEPATIC CLEARANCE
13.5.1 Aims
Forin vitrodata, PK properties, particularly for the liver, are potentially
predictable (Lin and Lu, 1997). Compared to other possible predictive
approaches, such as allometric scaling using animal models, predictions based
onin vitrohuman metabolic data have indeed been shown to be reasonably
accurate and cost effective (Zuegge et al., 2001). However, it must be borne in
mind that the intrinsic differences betweenin vitroandin vivosystems inevitably
challenge the consistency between pharmacokinetics and in vitro results
(Masimirembwa et al., 2003; Zhang and Wong, 2005). Human physiology is
complex and dynamic, and can by no means be completely simulatedin vitro.The
following should, therefore, be viewed as a simplification.
While several fundamental PK terms including hepatic clearance (CLh),
extraction (Eh) and availability (Fh), could be estimated fromin vitrointrinsic
clearance (CL
0
int), a pharmacokinetic model describing the concentration of the
drugs/xenobiotics in the liver, serving as a link, should be employed. Such
potential predictive models include the well-stirred (Rowland et al., 1973;
Wilkinson and Shand, 1975), the parallel-tube (Pang and Rowland, 1977a,
1977b; Winkler et al., 1973), the distributed (Roberts and Rowland 1986a; Bass
et al., 1978), and the dispersion models (Roberts and Rowland, 1986b, 1986c).
Generally speaking, the predictive powers of these models appear to be drug
clearance-dependent, with few differences when applied to low clearance drugs;
however, they show potential disparities for high clearance drugs (Iwatsubo
et al., 1997; Ito and Houston, 2004). The traditional well-stirred model,
although not always the most accurate, has been shown to be generally suitable
forin vitro–in vivocorrelations (Ito and Houston, 2004); thus, it is chosen as the
one for demonstration in the following content.
434 DETERMINATION OF METABOLIC RATES AND ENZYME KINETICS