Microdialysis of dopamine interpreted with quantitative model incorporating probe implantation trauma

Abstract

Although microdialysis is widely used to sample endogenous and exogenous substances in vivo, interpretation of the results obtained by this technique remains controversial. The goal of the present study was to examine recent criticism of microdialysis in the specific case of dopamine (DA) measurements in the brain extracellular microenvironment. The apparent steady-state basal extracellular concentration and extraction fraction of DA were determined in anesthetized rat striatum by the concentration difference (no-net-flux) microdialysis technique. A rate constant for extracellular clearance of DA calculated from the extraction fraction was smaller than the previously determined estimate by fast-scan cyclic voltammetry for cellular uptake of DA. Because the relatively small size of the voltammetric microsensor produces little tissue damage, the discrepancy between the uptake rate constants may be a consequence of trauma from microdialysis probe implantation. The trauma layer has previously been identified by histology and proposed to distort measurements of extracellular DA levels by the no-net-flux method. To address this issue, an existing quantitative mathematical model for microdialysis was modified to incorporate a traumatized tissue layer interposed between the probe and surrounding normal tissue. The tissue layers are hypothesized to differ in their rates of neurotransmitter release and uptake. A post-implantation traumatized layer with reduced uptake and no release can reconcile the discrepancy between DA uptake measured by microdialysis and voltammetry. The model predicts that this trauma layer would cause the DA extraction fraction obtained from microdialysis in vivo calibration techniques, such as no-net-flux, to differ from the DA relative recovery and lead to an underestimation of the DA extracellular concentration in the surrounding normal tissue.

Fig. 1

Schematic diagram for the characterization of microdialysis probes in vitro. The dopamine concentrations in the perfusate and dialysate are indicated by $C_{\text{d}}^{\text{in}}$ and $C_{\text{d}}^{\text{out}}$ and the volumetric flow rate of these solutions by Q_d. The probe is held concentrically in a stream of external medium pumped axially along the probe membrane at a mean velocity of v_ext. The probe and external medium are maintained at 37°C by immersion in a temperature controlled water bath.

Fig. 2

Determination of well-stirred in vitro dopamine extraction fraction, $E_{\text{vitro}}^{\text{ws}}$, by extrapolation. The points indicate extraction fraction values, E_vitro, calculated according to the definition, equation 6, from measured values of the inflow and steady-state outflow concentrations, $C_{\text{d}}^{\text{in}}$ and $C_{\text{d}}^{\text{out}}$, for various external solution mean velocities, v_ext. The intercept at 1/v_ext = 0 obtained by linear regression is $E_{\text{vitro}}^{\text{ws}}$. The nominal membrane lengths and $E_{\text{vitro}}^{\text{ws}}$ (mean ± SE) values are: ■, 1 mm, 0.244 ± 0.002, n = 4; ●, 2 mm, 0.405 ± 0.009, n = 4; ▲, 3 mm, 0.529 ± 0.017, n = 4; ▼, 4 mm, 0.604 ± 0.007, n = 4.

Fig. 3

Determination of the length of membrane available for dop-amine exchange by regression of the well-stirred extraction fraction intercepts, $E_{\text{vitro}}^{\text{ws}}$, from Fig. 2 plotted as a function of the nominal membrane length, l_n.

Fig. 4

In vivo dopamine extraction fraction, E_vivo, and apparent extracellular concentration, $C_{\text{e}}^{\text{app}}$, obtained by the concentration difference (no-net-flux) technique. Probes implanted in anesthetized rat striatum were alternately and randomly perfused with artificial cerebrospinal fluid containing 0, 100, 200 or 400 nM dopamine. According to equation A20, E_vivo is the coefficient of proportionality between the perfusate concentration difference for inflow and outflow, $C_{\text{d}}^{\text{in}}-C_{\text{d}}^{\text{out}}$, and the inflow value, $C_{\text{d}}^{\text{in}}$. Each point is a mean ± SE from five rats and the slope and intercept at $C_{\text{d}}^{\text{in}}=C_{\text{d}}^{\text{out}}$ by linear regression are E_vivo = 0.52 ± 0.01 and $C_{\text{e}}^{\text{app}}=13\pm 13$ nM, respectively, and E_vivo = 0.54 ± 0.10 and $C_{\text{e}}^{\text{app}}=16\pm 7$ nM if the 400 nM data are excluded.

Fig. 5

Geometry for proposed model of microdialysis with a traumatized layer of tissue immediately adjacent to the probe membrane. The radial distance from the probe axis is denoted by r and the axial distance from the inlet end of the membrane is z. The inner and outer radii of the membrane (m) are r_i and r_o, respectively, and the thickness of the trauma layer (tr) is δ. The surrounding tissue layer (n) is of effectively semi-infinite extent.

Fig. 6

Because of the dependence on the unknown microdialysis trauma layer thickness, δ, the rate constant for dopamine clearance in the trauma layer, $k_{\text{e}}^{\text{tr}}$, can be any value between zero and $k_{\text{e}}^{\text{app}}$, the apparent clearance rate constant obtained from the no-net-flux analysis. For δ greater than 13 μm, $k_{\text{e}}^{\text{tr}}$ is within 5% of the apparent clearance rate constant, $k_{\text{e}}^{\text{app}}$.

Fig. 7

The discrepancy between the apparent extracellular dopamine concentration from the no-net-flux intercept, $C_{\text{e}}^{\text{app}}$, and the true value far from the probe, $C_{\text{e}}^{\infty }$, increases as the trauma layer thickness increases. The same dependence on δ applies to the discrepancy between the true relative recovery, R, and the in vivo extraction fraction, E_vivo. For this illustrative calculation from equation 12, release is assumed to be abolished in the trauma layer $(C_{\text{e}}^{\ast }=0)$.

Fig. 8

If dopamine release is abolished in the layer of traumatized tissue while uptake continues to occur throughout both this layer and the adjacent normal tissue, the extraction fraction, E_vivo, exhibits a monotonic reduction as uptake is progressively inhibited. However, the true relative recovery, R, based on the dopamine extracellular concentration in the normal tissue increases as the degree of uptake inhibition increases (except in range of nearly complete inhibition). Uptake inhibition is simulated by proportionate reduction in both the normal and trauma layer uptake rate constants, $k_{\text{e}}^{\text{n}}$ and $k_{\text{e}}^{\text{tr}}$, respectively. The abscissa is the fraction reduction in these rate constants. In the absence of trauma, the curve for R would become the same as that for E_vivo and likewise exhibit a monotonic decrease with uptake inhibition. For these illustrative curves, an arbitrary trauma layer thickness of δ = 20 μm was chosen.

Fig. 9

Predicted radial profiles in endogenous dopamine extracellular concentration produced by microdialysis sampling $(C_{\text{d}}^{\text{in}}=0)$ are strongly dependent upon the unknown thickness of the trauma layer. Profiles in the dialysate, probe membrane fluid and ECS of striatal tissue have been calculated from equations A27–A31 for δ = 10, 20 and 30 μm and the values, $C_{\text{e}}^{\text{app}}=16$ nM and E_vivo = 0.54. Concentrations are normalized with respect to the distant undisturbed extracellular concentration, $C_{\text{e}}^{\infty }$. Axial-averaging over the membrane length is indicated by the angle brackets. The extracellular concentration at the probe-tissue and trauma layer–normal tissue interfaces are denoted by $C_{\text{e}}^{\text{o}}$ and $C_{\text{e}}^{\delta }$, respectively, and the radially averaged concentration in the dialysate by C_d.