Back to the Top
Deal Colleagues,
Say we are dealing with a situation that total plasma
drug concentrations are not reflective of target
tissue levels (equilibrium between plasma and tissues
has not been achieved or will be achieved at very late
time points e.g., 72 h post-dosing). How do you
handle this situation? What is the value of plasma
concentration-time data?
Rostam
Back to the Top
Check the blood to plasma ratios. Blood is what perfuses tissues and
organs.
Stanley Cotler
Non-Clinical Drug Safety
Nutley, NJ 07110
Back to the Top
Dear Rostam,
You raise a very important point - what is the value of plasma
concentration-time data? If we are expecting therapeutic or assessing the
risk of adverse effects in a tissue site, the PK we are interested in is
that of the tissue, not the plasma. However, we use plasma as a biomarker
of tissue levels with the assumption that they are highly correlated.
If the tissue and plasma (or more correctly blood, as has been pointed out)
levels do not follow each other strictly in time, then a probable
explanation is that there may be a permeability barrier between the
vasculature and the tissue. This can be at the level of the endothelium or
the cellular plasma membranes. In other words the organs are no longer
behaving as a simple, single, well stirred compartments. An alternative is
non-linear tissue binding, perhaps to specific proteins.
The way to understand this is to consider using physiology based PK
modelling (PBPK). There are several examples of situations where models
containing two (or more) compartments per organ have been used to explain
then predict the types of kinetics you observe. They are parameterised by
the actual volumes of the different parts of the tissue (e.g. vascular,
interstitial, cellular), permeability x surface area coefficients (PS) and
either unbound drug partition coefficients (Kpu) or fractions unbound (e.g.
fu, fe, fc for blood, extracellular fluid and cellular fluid repsectively).
A good starting reference is Kawai et al (1998) J. Pharmacol. Exp. Therap
287, 457-468.
PBPK models can be developed for the whole body (ideally), or using
arterial blood levels as a forcing function to drive tissue levels via the
above parameters. Venous blood levels could be used as a surrogate if you
are content with assuming negligable arteriovenous differences. Whole body
models have the advantage of being scalable between species by substituting
parameter values of flows, volumes, unbound fractions, intrinsic CL, etc.
for the alternative species.
Best regards, Phil
Philip Lowe
Preclinical Safety Modelling Group
Novartis Pharma AG
CH-4002 Basel
Back to the Top
Dear Philip Lowe,
In your letter to Rostam, you are making an important point in stating
"the organs are no longer behaving as a simple, single, well stirred
compartments. An alternative is non-linear tissue binding, perhaps to
specific proteins." Further, "the PK we are interested in is that of the
tissue, not the plasma. However, we use plasma as a biomarker of tissue
levels with the assumption that they are highly correlated."
This is a major problem in most of the current PK studies in drug
development.
The assumption that plasma can be used as biomarker of tissue levels and
that plasma levels and specific(!) tissue levels are highly correlated
cannot be made. No PK modeling will help. Information from in vitro
studies neither.
What is needed is pharmacokinetics of receptor-drug binding. Current
methods do not provide such information because of lack of sensitivity and
resolution.
As an example, for steroids, plasma half-life may be in the range of 15 to
30 min, while receptor-steroid half-life may range from one to several
hours and vary among different target tissues.
This problem has been discussed during January of 2001.
Walter E. Stumpf
Chapel Hill, NC
Back to the Top
Dear Colleagues,
As Stanley suggested I looked at the blood data and it
seems to correlate better with tissue data but even
that does not indicate equilibrium. Philip suggested
using PBPK modeling (and I totally agree with him)
which is an establish approach in Environmental
Toxicology and Risk Assessment but not a common
practice in pharmaceutical industries yet. Walter
raises an interesting and important point with regard
to PK of receptor-drug binding and also mentioned ėno
PK modeling will helpî. My question is without
knowing the concentration at the target site or
tissue(s) of interest how PK of receptor-drug binding
can help?
Walter, you mentioned that this problem was discussed
during January of 2001, where? Would you please
provide more detail/references?
Thanks
Rostam
Back to the Top
Dear Rostam
I appreciate your response.
The records of the PharmPK letters are kept at Boomer Home Page:
PharmPK.-a-.boomer.org
Discussions about Receptor-Drug Pharmacokinetics took place in January 0f
2001. Regards, Walter
[The url is http://www.boomer.org/pkin/. The archive are listed about
2/3rds of the way down the page. They are searchable (again) using
the URL http://www.boomer.org/pkin/simple.html and are available on
CD, http://www.boomer.org/CDinfo.html. I wonder if this is the
discussion http://www.boomer.org/pkin/PK01/PK2001010.html or
http://www.boomer.org/pkin/PK01/PK2001132.html - db]
Back to the Top
Dear Rostam:
It seems to me that when serum concentrations are not reflective
of tissue (or other) concentrations, or of effect, that is what
multicompartment modeling is all about. If you have data elsewhere, use it.
If not, see what you can to in fitting a multicompartment model to your
serum data. Reuning et al, for example, made a really nice model of digoxin
in which the effect correlates much better with tissue concentrations than
with serum levels. Even though serum data was all there was, one could fit
a multicompartment model to that data, and then show its empirical
correlation with data of effect. We have happily used this model in our
clinicial USC*PACK software for years, and it has been most useful. The
patient's clinical status correlates much better with the concentrations in
the peripheral nonnserum compartment than with the serum levels. We use the
data of the levels to make an individual Bayesian posterior model, and we
compare the patient's clinical response with the behavior of the peripheral
compartment that we see in the plots of the behavior of the drug. See
Reuninng, Sams, and Notari, Role of Pharmacokinetics in Drug Dosage
Adjustment. Pharmacologic Effects, Kinetics, and Apparent Volume of
Distribution of Digoxin. J. Clin. Pharmacol. 13: 128-141, 1973. There must
be a ton of similar references. You might also look at our web
site www.lapk.org and click on the teaching topics and the software
there.
Very best regards,
Roger Jelliffe
Roger W. Jelliffe, M.D. Professor of Medicine,
Laboratory of Applied Pharmacokinetics,
USC Keck School of Medicine
2250 Alcazar St, Los Angeles CA 90033, USA
email= jelliffe.at.hsc.usc.edu
Our web site= http://www.lapk.org
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)