Quantitative microdialysis using modified ultraslow microdialysis: Direct rapid and reliable determination of free brain concentrations with the MetaQuant technique

https://doi.org/10.1016/j.jneumeth.2008.12.010Get rights and content

Abstract

The only method to quantify free extracellular levels of drugs in the brain of living animals is microdialysis. However, quantitative microdialysis has been hampered by methodological issues for decades. The problems arise from the need to establish the in vivo recovery for appropriate quantitation. In dealing with these issues the “dynamic no-net-flux” (DNNF) method seemed to be the experimental method of choice. Major disadvantages were, however, the need for a very high degree of bioanalytical precision and accuracy and the need for a large number of animals. Moreover, today we know that the experimental data are not always straightforward.

To improve robustness and practicality of quantitative microdialysis sampling we modified the ultraslow microdialysis approach. Ultraslow microdialysis uses very low microdialysis flow rates (<200 nl/min) which increase recovery (both in vivo and in vitro) to over 90%. However, new practical issues arise when attempting to work with these flow rates. The resulting very low volumes and long lag times make this method very impractical for general application. In the modified version, addition of a carrier flow after the dialysis process has been completed, which negates the problems of long lag times and low volumes. The resulting dilution of the dialysis sample concentration can simply be mathematically corrected.

In the current study we measured the free brain levels of two CNS compounds using the classic DNNF and the new modified ultraslow dialysis method.

Modified ultraslow microdialysis was shown to generate robust data with the use of only small numbers of rats. The method is a promising tool for common straightforward screening of blood–brain barrier penetration of compounds into the brain.

Introduction

Measuring free brain levels of experimental CNS-active compounds in an early stage of their development is crucial to acquire vital information about their blood–brain barrier (BBB) penetrating properties. To date it is common practice to use total brain and/or CSF samples to ascertain this information. However, determination of CSF or total tissue levels yields data that may not necessarily reflect the actual free brain levels (De Lange and Danhof, 2002).

The only way to monitor free brain concentrations directly is by in vivo intracerebral microdialysis. This method provides access to the extracellular fluid (ECF) of the brain and enables repeated sampling in a living subject.

The basic mechanism of the microdialysis method, briefly, is to insert a small microdialysis probe into the brain and perfuse its lumen with a physiological solution. This probe contains a semi-permeable membrane that allows small molecules to penetrate the membrane down their concentration gradient. The effluent, the dialysate, is continuously collected into vials and analysis of the compound(s) of interest can, thus, be performed off-line. The concentration of a compound in the dialysate is therefore indicative of the free concentration in the extracellular fluid surrounding the semi-permeable membrane of the microdialysis probe. The rate and extent of concentration equilibration between the lumen of the probe and the fluid surrounding the semi-permeable membrane of the probe (concentration recovery) can be estimated in vitro, but never really reflects the probe's performance in vivo. Instead, it merely reveals the semi-permeable membrane transport characteristics as such. This is because in vivo concentration recovery is to a large extent affected by physiological processes in the tissue (De Lange et al., 1997a, De Lange et al., 1997b). These processes, and hence in vivo recovery, may be time- and concentration-dependent (Bungay et al., 1990, Morrison et al., 1991). Therefore the in vivo concentration recovery cannot be predicted simply from in vitro tests alone (Stahle, 1991).

For many years the determination of the in vivo recovery, vital to obtain quantitative data, has proven to be a most eluding challenge. Quite a few different methods have been developed, including both theoretical (Lindefors et al., 1989, Bungay et al., 1990, Morrison et al., 1991) and experimental approaches (Lonnroth et al., 1987, Larsson, 1991, Scheller and Kolb, 1991, Yokel et al., 1992, Olson and Justice, 1993, Bouw and Hammarlund-Udenaes, 1998, Wang et al., 1993). These methods also range from simplified approaches with many incorporated assumptions to complex procedures with almost no assumptions (De Lange et al., 2001a). Important experimental approaches are the retrodialysis by drug (Scheller and Kolb, 1991), the no-net-flux (NNF) method, retrodialysis by calibrator (Yokel et al., 1992), or by drug and calibrator (Bouw and Hammarlund-Udenaes, 1998) and the dynamic no-net-flux (DNNF, Olson and Justice, 1993) method.

While the retrodialysis approach is relatively easy, it has to be assumed that in vivo concentration recovery is independent of concentration in the probes’ surrounding, while also no time-dependency is assumed, or at least a time-dependency being similar for the drug and calibrator. The DNNF almost has no assumptions. To date is was the only experimental method that takes into account time- and concentration-dependency of in vivo recovery. Instead of serial perfusion of individual animals with different concentrations via the probe like in the NNF, a group of animals are continuously perfused with one of the perfusion concentrations selected around the expected ECF concentrations. Different groups are perfused with different concentrations and the results are combined at each time point. Regression of the mean data points of the different groups at a particular point in time will give the actual concentration in the ECF with the associated in vivo concentration recovery value at that time (Xie et al., 1999, De Lange et al., 2001b). However, this is a very complex and time-consuming protocol, with another important disadvantage from an ethical point of view: the requirement of a significantly larger number of experimental animals. Moreover, long standing experience with the use of the DNNF has made clear that often highly variable outcomes are generated (data not published).

In the current study we developed a new microdialysis sampling system. We adopted the principle that microdialysis extraction becomes quantitative (i.e. concentration recovery is close to 100%) when the dialysis flow rate is decreased from the more common rates of 1.5–2 μl/min down to 0.1 μl/min (Menacherry et al., 1992; Cremers et al., 2001).

In practice it is notoriously difficult in in vivo microdialysis experiments to work with flow rates as low as 0.1 μl/min, because of the long lag times (between probe and collection vial) and small sample volumes.

Therefore we modified the conventional microdialysis probe setup by introducing an additional flow (carrier flow) to merge with the ultra-slow dialysate immediately downstream from the microdialysis membrane. The carrier fluid is delivered at a higher flow rate (typical 0.9 μl/min) that shortens the lag time and increases final sample volume for easier handling. The obvious dilution of dialysate by the carrier fluid requires employment of sensitive techniques such as mass spectrometry. As the microdialysis exchange has already accomplished a close to 100% concentration recovery after leaving the dialysis membrane, the dilution of the sample by the carrier flow and collection systems can be mathematically corrected after analysis. This design greatly facilitates the ease of use of the sampling system in vivo (Cremers and Ebert, 2007).

In the current study, the new microdialysis sampling system, named MetaQuant, was validated for two known CNS compounds by evaluating rate and extent (dynamics) of their in vitro concentration recovery, and by comparing in vivo MetaQuant data to those obtained by the DNNF.

Section snippets

Specifications of a MetaQuant probe

Fig. 1 shows a diagram of the MetaQuant microdialysis probe (Brainlink BV, Groningen, The Netherlands). An artificial cerebrospinal fluid (147 mM NaCl, 3.0 mM KCl, 1.2 mM CaCl2, and 1.2 mM MgCl2 containing 0.2% albumin) is typically delivered at 0.1 μl/min through the middle (yellow) inlet. The carrier flow (typically ultra-purified water) is delivered through the left (blue) inlet at 0.9 μl/min. The right outlet thus yields a combined outflow of 1 μl/min, consisting of 0.10 μl/min dialysate with 100%

In vitro recovery

Decreasing the flow rate through the dialysis probe typically increased the relative recovery of gaboxadol from 40% at 1 μl/min to over 90% at flow rates below 0.2 μl/min (Fig. 2A). At the same time, absolute amounts recovered decreased as the increase in recovery was outweighed by the decrease in volume.

The dynamic properties of changes in recovery to an instantaneously applied increase in external bulk fluid concentration were characterized by a delay of 30% in the first 10 min. This means the

Discussions and conclusions

The research presented in this paper was designed to develop an advanced quantitative, highly practical and direct method for assessment of free brain concentrations of (new) CNS-targeted drugs. Quantitative microdialysis data are extremely useful in many research applications including, but not limited to, investigation of pharmacokinetic–pharmacodynamic relationships in drug development. The purpose of the current study was to improve the practical use of quantitative microdialysis, and

Conclusions

The present study shows that modified ultraslow microdialysis (“MetaQuant”) technique can be employed to quantitate free brain levels without the need to perform complex approaches with an aggregate statistic. In addition, it is important to note that this approach requires considerably fewer animals and even created a more robust measurement.

We therefore conclude that the modified ultraslow microdialysis technique is an advanced quantitative method for assessment of free brain concentrations

References (20)

There are more references available in the full text version of this article.

Cited by (0)

View full text