Catecholamine release and uptake in the mouse prefrontal cortex

Authors


Address correspondence and reprint requests to R. Mark Wightman, Department of Chemistry and Neuroscience Center, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599–3290, USA. E-mail: rmw@unc.edu

Abstract

Monitoring the release and uptake of catecholamines from terminals in weakly innervated brain regions is an important step in understanding their importance in normal brain function. To that end, we have labeled brain slices from transgenic mice that synthesize placental alkaline phosphatase (PLAP) on neurons containing tyrosine hydroxylase with antibody–fluorochrome conjugate, PLAP-Cy5. Excitation of the fluorochrome enables catecholamine neurons to be visualized in living tissue. Immunohistochemical fluorescence with antibodies to tyrosine hydroxylase and dopamine β-hydroxylase revealed that the PLAP labeling was specific to catecholamine neurons. In the prefrontal cortex (PFC), immunohistochemical fluorescence of the PLAP along with staining for dopamine transporter (DAT) and norepinephrine transporter (NET) revealed that all three exhibit remarkable spatial overlap. Fluorescence from the PLAP antibody was used to position carbon-fiber microelectrodes adjacent to catecholamine neurons in the PFC. Following incubation with l-DOPA, catecholamine release and subsequent uptake was measured and the effect of uptake inhibitors examined. Release and uptake in NET and DAT knockout mice were also monitored. Uptake rates in the cingulate and prelimbic cortex are so slow that catecholamines can exist in the extracellular fluid for sufficient time to travel ∼100 µm. The results support heterologous uptake of catecholamines and volume transmission in the PFC of mice.

Abbreviations used
aCSF

artificial cerebral spinal fluid

Cg1

cingulate cortex

CP

caudate putamen

DAT

dopamine transporter

DβH

dopamine β-hydroxylase

KO

knockout

LC

locus coeruleus

NET

norepinephrine transporter

PLAP

placental alkaline phosphatase

PFC

prefrontal cortex

PLC

prelimbic cortex

SNc

substantia nigra pars compacta

TH

tyrosine hydroxylase

VTA

ventral tegmental area.

The prefrontal cortex (PFC) is a region of the brain that is involved in spatial working memory and decision making (Damasio 1995; Williams and Goldman-Rakic 1995; Murphy et al. 1996; Jentsch et al. 1997; Zahart et al. 1997; Seamans et al. 1998). Improper function of neurotransmitters in this region has been associated with a series of disorders such as schizophrenia, depression, drug abuse, and attention deficit disorders. The catecholamines have received particular attention since both norepinephrine (NE) and dopamine (DA) terminals are found in this region (Thierry et al. 1973; Berger et al. 1974; Berger et al. 1976). DA terminals in the PFC primarily project from the ventral tegmental area (VTA) (Emson and Koob 1978; Lindvall et al. 1978). NE terminals arise from the locus coeruleus (LC) (Loughlin et al. 1982). However, the density of catecholaminergic terminals in the PFC is low when compared with the caudate-putamen (CP), a region that is densely innervated with dopaminergic neurons (Palkovits 1979). In addition, the PFC has been shown to contain fewer DA transporters (DAT) than the CP, as evaluated by immunohistochemistry (Sesack et al. 1998).

Regulation of catecholaminergic neurotransmission in the PFC differs from that in other brain regions due to structural and functional differences. For example, the dopaminergic terminals in the PFC possess secretion-regulating and uptake-modulating autoreceptors (Talmaciu et al. 1986; Starke et al. 1989; Meiergerd et al. 1993; Cass and Gerhardt 1994), but lack synthesis-regulating autoreceptors (Bannon et al. 1981; Galloway et al. 1986) that are found in other DA terminal regions. The rate of DA uptake in the PFC has been found to be slower than in the CP or nucleus accumbens (Garris et al. 1993; Cass and Gerhardt 1995; Wayment et al. 2001). The slow clearance of DA in the PFC, which correlates with the low amount of DAT in this region (Sesack et al. 1998), suggests that DA can diffuse farther from its release site than in the CP. This diffusion would increase the likelihood of volume transmission, extrasynaptic transmission that occurs over long distances, in the PFC. Similarly, NE uptake in the cortex has been shown to be quite slow (Mitchell et al. 1994).

Due to the low immunoreactivity for DAT in the PFC, it has been proposed that NE terminals may be responsible for DA clearance from the extracellular fluid (Carboni et al. 1990; Pozzi et al. 1994; Cass and Gerhardt 1995). This uptake of DA by NE terminals is termed heterologous uptake. This type of uptake is feasible because DA actually has a higher affinity for the NE transporter (NET) than NE (Horn 1973; Raiteri et al. 1977). Indeed, pharmacological inhibition of NET has been found to increase the extracellular concentrations of DA in the PFC (Carboni et al. 1990), consistent with heterologous uptake (Di Chiara et al. 1992; Tanda et al. 1997; Yamamoto and Novotney 1998).

The relatively low levels of releasable DA and NE in the PFC has hindered extensive investigation of catecholamine neurotransmission in this region by voltammetric microelectrodes. Multiple placements of the electrode are required to find the release sites. In this work, we have investigated catecholaminergic neurotransmission in the brain slices prepared from transgenic mice that enable the catecholamine-containing neurons to be labeled with an antibody–fluorochrome conjugate (De Waele et al. 1982; Gustincich et al. 1997; Hochstetler et al. 2000). These animals have been genetically altered so that any neuron containing tyrosine hydroxylase (TH), the rate-limiting enzyme in catecholamine synthesis, will express a human placental alkaline phosphatase (PLAP) on its surface. Binding of the PLAP antibody conjugated to a fluorochrome enables visualization of the catecholamine neurons in brain slices, facilitating positioning of the electrode adjacent to release sites. Using this approach, catecholamine release and uptake has been investigated with cyclic voltammetry in the cingulate cortex (Cg1) and the prelimbic cortex (PLC) of mice. Consistent with the close proximity of DAT and NET in these regions, the results are entirely consistent with the concept of heterologous uptake. Recordings in transporter knockout (KO) animals also support this view.

Materials and methods

Animals

A transgenic mouse line in which catecholamine containing cells express the gene for the human placental alkaline phosphatase was used (Gustincich et al. 1997). DAT knockout (Giros et al. 1996) and NET knockout (Xu et al. 2000) mice were utilized as well. Mice were 2–6-months-old at the time of the experiments. Food and water was provided ad libitum. Animal care was provided in accordance with institutional guidelines.

Slice preparation

Brain slices were prepared as previously described (Jones et al. 1995). Briefly, mice were killed by decapitation, the brain was rapidly removed, and coronal slices (200 µm thick) were obtained using a vibrating microtome (Model NVSL, World Precision Instruments, Inc., Sarasota, FL, USA). During this step, the brain and brain slices were maintained in an artificial cerebral spinal fluid (aCSF) that had been chilled on ice. The aCSF contained 126 mm NaCl, 2.5 mm KCl, 1.2 mm NaH2PO4, 2.4 mm CaCl2, 1.2 mm MgCl2, 25 mm NaHCO3, 11 mm glucose, and 20 mm HEPES. The pH was 7.4 and the buffer was continually saturated with 95% O2/5% CO2.

Labeling with antibodies

The monoclonal PLAP antibody, E6 (De Waele et al. 1982) was conjugated to Cy5 fluorochrome (Amersham Life Sciences Inc., Pittsburgh, PA, USA) as previously described (Gustincich et al. 1997). The antibody–fluorochrome conjugate (5 µL per 1 mL of aCSF) was incubated with the slices in a culture plate at room temperature for 5 min followed by 15 min at 37°C. Then, the culture plate was removed from the incubator and the incubation buffer was saturated with 95% O2/5% CO2 for another 30 min.

To examine PLAP/TH colocalization, some slices were also stained with a secondary immunoreaction to a TH antibody. After the initial PLAP antibody–fluorochrome labeling, slices were fixed in 2% formaldehyde in Sorenson's buffer (0.15 m Na2HPO4/KH2PO4 at pH 7.4) for 4 h. The slices were rinsed three times with 1 × PBS (Tissue Culture Facility, Chapel Hill, NC, USA). Then, normal goat serum (NGS, Life Technologies, Rockville, MD, USA) was combined with 0.2% BSA/PBS solution (1 : 10, NGS : BSA/PBS). The slices were incubated in this blocking solution for 1 h. An antibody to TH from rat PC12 cells (Chemicon International, Inc., Temecula, CA, USA) was diluted 10-fold with PBS and then further diluted 10-fold with NGS/BSA/PBS solution. Slices were incubated with this solution for 15 h and rinsed with PBS. A secondary antibody, Alexafluor®-488 anti-rat (Molecular Probes, Eugene, OR, USA), was diluted 1 : 2000 and incubated for 3–4 h with the slices. Then, the slices were rinsed with PBS, placed on slides, and examined by confocal microscopy.

The colocalization of PLAP and dopamine β-hydroxylase (DβH) was also examined. The slices were labeled with the PLAP antibody and placed in the blocking solution. The DβH polyclonal antibody (Diasorin Inc., Stillwater, MN, USA) was diluted 10-fold with PBS and further diluted 100-fold with NGS/BSA/PBS solution. Slices were incubated with this solution for 15 h and then rinsed with PBS. The secondary antibody from the Alexafluor®-488 anti-rabbit series (Molecular Probes) was diluted 1 : 2000 and incubated with the slices for 4 h. Then, the slices were rinsed with PBS and examined.

For the triple labeling of DAT, NET, and PLAP, the slices were first incubated with PLAP and then the blocking solution. The rat anti-DAT monoclonal antibody (Chemicon International, Inc., Temecula, CA, USA) was diluted 1 : 2000 with NGS/BSA/PBS. Slices were incubated with this solution for 16 h and rinsed with PBS. The secondary antibody AlexaFluor® 488 anti-rat (Molecular Probes) was diluted 1 : 2000 and incubated with the slices for 4 h. The slices were rinsed twice with PBS and incubated with the blocking solution for 1 h. Following the incubation with the blocking solution, the anti-rat NET antibody (Chemicon International, Inc.) was diluted 1 : 2000 with NGS/BSA/PBS. Slices were incubated with this solution for 16 h and rinsed with PBS. The secondary antibody AlexaFluor® 568 anti-rabbit (Molecular Probes) was diluted 1 : 2000 and incubated with the slices for 4 h.

Confocal microscopy

Throughout this work, anatomical locations were determined by comparison with a mouse brain atlas (Franklin and Paxinos 1997). The brain slices containing the desired regions were placed onto microscope slides with glass wells to maintain their structural integrity. Into these wells, VectaShield (Vector Laboratories Inc., Burlingame, CA, USA) was added. After the slices were placed in mounting medium, they were coverslipped, and sealed with clear nail varnish.

The confocal images were collected on a Leica TCS-NT laser scanning confocal microscope mounted on a Leica DMRI inverted microscope (Leica Microsystems, Confocal Microscopy, Germany). Photomultiplier tubes were used to detect the fluorescence from the fluorochromes in the multiple-labeling experiments. Images were initially imported into Adobe Photoshop 5.5 (Adobe Systems Incorporated, San Jose, CA, USA) where they were rendered in false color using the CMYK color scheme. Images of the same region stained for two different antibodies were combined using the difference mode. Estimation of the degree of overlap of the regions labeled with different antibodies employed Scion Imaging software (Scion Corporation, Frederick, MD, USA). In cell body regions, the labeled area was selected with the ‘magic wand’ tool of the Scion software. When labeling was more diffuse, such as in regions intensely innervated with terminals like the CP, the unstained regions were selected and subtracted from the total area. The remainder was considered the labeled area. For weakly innervated regions such as the PFC and Cg1, regions of fluorescence were selected manually. Overlap was defined as the ratio of the labeled areas for different antibodies.

Electrochemical experiments

Fast scan cyclic voltammetry was performed with an amplifier (Axopatch 200B, Axon Instruments, Foster City, CA, USA). The computer-generated, triangular waveform was scanned from −0.4 V to 1.0 V and back at 300 V/s with a 10-Hz repetition rate (Michael et al. 1998). Background-subtracted cyclic voltammograms were obtained and the peak oxidative current for catecholamine was integrated over a 100-mV window in successive voltammograms. This current was converted to concentration based upon a calibration in aCSF after the experiment. The cyclic voltammograms do not allow distinction between DA and NE, but do provide signals for these two substances that are different from ascorbate, serotonin, dihydroxyphenylacetic acid, and other interferences. The cyclic voltammogram for l-DOPA has peaks at similar locations as the catecholamines. However, its amplitude is one-tenth that of DA and its reverse wave is poorly defined giving it a unique appearance. For every reported result in slices, the substance detected was a catecholamine as confirmed by its voltammogram.

Electrodes were constructed by a modified procedure for preparing microcylinder electrodes (Cahill et al. 1996). Approximately 75–100 µm of the carbon fiber (5 µm radius, Thornel T650, Amoco Corp., Greenville, SC, USA) was left protruding beyond the cut glass prior to sealing the electrodes with epoxy (Epon 828 with 14% w/w m-phenylenediamine, Miller-Stephenson Chemical Co., Danbury, CT, USA). The electrode tips were then coated with an insulating layer of electrodeposition paint (BASF, Münster, Germany) (Hochstetler et al. 2000).

One day prior to use, the electropainted electrodes were beveled on a diamond dust-embedded polishing wheel (Sutter Instrument Co., Novato, CA, USA), cleaned in isopropyl alcohol for 10 min, coated with Nafion for 10 min, and heat treated for 15 min. The exposed area is defined by the cross section of the fiber. Electrodes were back filled with a solution containing 4 m potassium acetate/150 mm KCl, and electrical connection to the head stage (CV201A/Axopatch 200 A, Axon Instruments, Foster City, CA, USA) was made by a chlorided silver wire. All electrochemical potentials were applied versus a quasi-Ag/AgCl reference electrode.

Stimulated release

The slices were placed in the perfusion/recording chamber (Model RC-22C, Warner Instrument Co., Hamden, CT, USA) and, consistent with prior work (Bull et al. 1990), they were perfused with preheated (32°C), oxygenated aCSF. Slices containing the PFC were incubated for 30 min in 100 µm L-DOPA and subsequently rinsed for 1 h with aCSF. A stimulating electrode was fabricated from two insulated tungsten microelectrodes (FHC Inc., Bowdoinham, ME, USA). They were placed on the slice surface and lowered approximately 20 µm. A carbon-fiber microelectrode was lowered 75 µm into the tissue and placed approximately 100–150 µm from the stimulating electrodes (Jones et al. 1995). Catecholamine release was evoked with biphasic (2 ms for each phase), constant current (48 µA) pulses. Typically, 30 pulses at 30 Hz stimulation was used. Each site was stimulated to obtain a reproducible response (at least three measurements taken every 5 min). Effects of DAT and NET inhibition were examined with GBR-12909 (Sigma, St Louis, MO, USA) and protriptyline (Sigma), respectively.

Data analysis

The concentration of evoked catecholamine, [CA]p, was determined from the maximum amplitude. Uptake rates following stimulated release were fit to the Michaelis–Menten equation:

inline image

Diffusion through the Nafion film was accounted for by convolution (Kawagoe et al. 1992). Fits were adjusted until a maximum correlation coefficient between the data and model were obtained. Before uptake inhibition, Vmax values were evaluated using a previously published value for Km (Ross 1991). After exposure to an uptake inhibitor, competitive inhibition was assumed and curve fits were obtained by adjusting the Km value while keeping the originally obtained value for Vmax. Results are shown as the mean of at least four replications. Statistically significant changes were determined by anova (p < 0.05).

High pressure liquid chromatography

Sections were punched from slices with a 2-mm diameter cork borer. The tissue was weighed wet and then homogenized in 40 µL of 0.1 N HClO4 using a sonic dismembrator (Fisher, Model 60, Pittsburgh, PA, USA). The homogenized tissue was centrifuged at 250 × g for 6 min in a microcentrifuge. The supernatant was removed and filtered using a 0.2-µm syringe microfilter. Injections (500 nL) were made onto a reverse phase column (C-18, 3 µm, 150 × 0.3 mm, BetaBasic). The mobile phase contained 100 mm sodium acetate, 0.7 mm sodium dodecyl sulfate, 0.1 mm EDTA, 10 mm NaCl and 10% acetonitrile (pH 3.5) at a flow rate of 9 µL/min. Catecholamines were detected with a glassy carbon electrode (BAS, West Lafayette, IN, USA) at a potential of 0.75 V versus a Ag/AgCl reference electrode.

Results

Imaging catecholamine neurons with antibodies for PLAP and TH

Initial experiments showed that the PLAP antibody bound to regions of known dopaminergic innervation in brain slices from the transgenic mice. To ensure that labeling was specific to catecholamine neurons, confocal images were obtained in slices that were labeled with both the PLAP and TH antibodies. For catecholamine cell body regions, there was an excellent correspondence of the TH and PLAP labeling (Figs 1a–c). Taking the difference of the two CMYK images in Adobe Photoshop (right panel) revealed that there was a significant amount of overlap (black–navy blue coloring). In the cell body regions, 95 ± 4% of the area stained by PLAP was also labeled by TH. The high degree of colabeling in these regions confirms the specificity of this approach to locate catecholamine containing neurons. In the CP (Fig. 1d), labeling with both antibodies was widespread with coincident unlabeled areas (black circular regions) that appear as circular white spots in the difference images. Presumably, these circular regions arise from noncatechoaminergic fibers of passage. TH labeled 92 ± 3% of the region labeled by PLAP. Diffuse labeling was observed in two regions of the prefrontal cortex: the cingulate cortex (Cg1, Fig. 1e) and the prelimbic cortex (PLC, Fig. 1f). In both regions, 81 ± 5% of the area labeled by PLAP was also labeled by TH. The extent of colabeling is summarized in Table 1.

Figure 1.

Confocal images of mouse brain slices labeled with TH and PLAP. Slices from mice genetically altered to express PLAP on TH-expressing cells were labeled with fluorescently tagged antibodies to both PLAP and TH as described in the Materials and methods section. The images shown employ the CMYK color scheme and encode TH labeling as yellow and PLAP labeling in the same section as blue. To the right of each pair of images are the difference images that reveal labeling colocalization. Black shows results from colocalization and white indicates the absence of labeling. The drawings on the left side of the images show the approximate location of the images shown. (a) SNc, (b) LC, (c) VTA, (d) CP, (e) Cg1, and (f) PLC. Scale bar = 50 µm.

Table 1.  Spatial overlap of the fluorescence from antibodies in the CP, Cg1, and PLC
Brain regionTH/PLAPDAT/PLAPNET/PLAP
  1. Areas labeled by the fluorescent antibodies were determined as described in the Materials and methods section. In all cases, the area of brain tissue labeled by the PLAP antibody exceeded the area labeled by the other antibodies used. Slices from at least seven mice (minimum of four slices per animal) were examined for each region.

CP92 ± 3%90 ± 4%2 ± 1%
Cg181 ± 5%39 ± 5%90 ± 8%
PLC81 ± 5%68 ± 3%89 ± 3%

Imaging catecholamine neurons with antibodies for PLAP and DβH

NE neurons contain DβH (the enzyme required for NE synthesis from DA) as well as TH. Therefore, confocal images were obtained in slices that were labeled with antibodies to PLAP and DβH (Fig. 2). Very little labeling for DβH was found in the CP (data not shown) indicating the specificity of the binding. In areas of sparse innervation such as the Cg1 and PLC of the prefrontal cortex, PLAP and DβH were found to be colocalized (Fig. 2a and b). In the difference images on the left, the PLAP labeling appeared as larger, circular areas of fluorescence and the DβH labeling was speckled. The cell bodies of the LC were easily seen with both antibodies (Fig. 2c). In fact, the DβH labeling was punctate and enabled clusters of vesicles to be seen at the cellular membrane (Fig. 2d).

Figure 2.

Confocal images of mouse brain slices labeled with DβH and PLAP antibodies. Slices from mice as in Fig. 1 were labeled with fluorescently tagged antibodies to PLAP and DβH as described in the Materials and methods section. Images are presented in false color using the procedures described in Fig. 1. (a) Cg1, (b) PLC, and (c) LC. The region identified by the box in (c) is shown in expanded form in (d). Scale bars are given for each set of images.

Imaging catecholamine neurons with antibodies for NET, DAT and PLAP

In the CP, there was almost no labeling with the NET antibody (Fig. 3a, Table 1), but extensive labeling was found with the DAT antibody (Fig. 3a) with high overlap with PLAP (Table 1). These results again confirm the specificity of the binding of the antibodies in the mouse brain. In the Cg1 and PLC, staining of the NET and DAT was quite weak (Figs 3b and c). In the Cg1, 90 ± 8% of the area labeled by the PLAP antibody was labeled by the NET antibody (Table 1), but only 39 ± 5% of this area was labeled by the DAT antibody with the greatest overlap occurring toward the midline. At the most medial locations of the PLC, PLAP staining is absent as is NET binding. Indeed, the overlap of area labeled by the NET antibody with that labeled by the PLAP antibody was similar to the Cg1 (89 ± 3%, Table 1). However, there was a greater overlap of the DAT to PLAP (68 ± 3%) in the PLC. Thus, the two transporters are found in close proximity to one another in the PFC at the level of magnification shown in Fig. 3 (overlapped areas indicated by shading on the cartoon of the brain slice).

Figure 3.

Confocal images of mouse brain slices labeled with antibodies to NET, DAT, and PLAP antibodies. Slices from mice as in Fig. 1 were labeled with antibodies as described in the Materials and methods section. Each antibody was encoded with a different fluorescent dye enabling all to be examined in the same slice. (a) CP, (b) Cg1, and (c) PLC; the arrows indicate the midline. Drawings of the approximate location of the images shown are given in the left panels. For (b) and (c), the stippled regions in the drawing indicates the regions stained by the NET and DAT antibodies. The circle in the Cg1 and the oval in the PLC represent the areas of greatest overlap of the DAT and NET antibody labeling. Scale bar = 100 µm. Because of the low density of staining in the cortical regions, the PMT gain was adjusted to a much higher value for (b) and (c) than for (a).

Stimulated release in the PFC

For the investigation of catecholamine release and uptake in the Cg1 and PLC, slices were incubated with the PLAP antibody to enable release sites to be visualized. In most instances, excitation of neurons with electrical stimulation delivered with a stimulating electrode adjacent to the carbon-fiber electrode was unable to evoke detectable release, even when placed in areas of PLAP localization. To replenish catecholamine stores that may have been depleted during slice preparation, slices containing the PLC and Cg1 were incubated with 100 µm L-DOPA for 30 min followed by perfusion with normal aCSF for 1 h prior to experiments. Release was still not observed when the electrode was lowered into nonfluorescent regions (Fig. 4a). However, release was observed at sites of localized fluorescence following l-DOPA loading (an example from the PLC is shown in Fig. 4b). HPLC analysis of punches taken from slices incubated with 100 µml-DOPA and perfused for the same period of time revealed an increase in DA tissue content following incubation with l-DOPA in both regions, but no increase in NE (Table 2). The amount of NE found in this work is quite similar to that reported for mouse cortex (283 ± 16 ng/g tissue, Gibson 1988). Levels of l-DOPA were undetectable, indicating that the perfusion after exposure to l-DOPA removed that which was not converted to dopamine.

Figure 4.

Stimulated release in the Cg1 of a brain slice from a mouse genetically altered to express PLAP on TH-expressing cells. A brain slice was incubated with the fluorescently labeled PLAP antibody and l-DOPA (100 µm) as described in the Methods and materials section. The carbon fiber electrode was placed adjacent to the stimulating electrode. The electrical stimulus was a 30 pulse, 30 Hz stimulation delivered in the time indicated by the bar under the concentration versus time trace. (a) Release monitored from an unlabeled region. (b) Release detected with the electrodes positioned in an area of localized fluorescence. The cyclic voltammogram in (b) was recorded at the maximum response. The solid line superimposed on the data in (b) is the simulated response (see Methods and materials section).

Table 2.  Amounts of NE and DA in the Cg1 and PLC
Brain regionNE
control
DA
control
NE/DA
control
NE
l-DOPA
DA
l-DOPA
NE/DA
l-DOPA
  1. Tissue content before (control) and after incubation with L-DOPA (100 µm for 30 min followed for 1 h by incubation in aCSF). Values reported as ng/g tissue.

Cg1100 ± 15111 ± 190.9 ± 0.270 ± 6240 ± 250.33 ± 0.1
(n = 5)(n = 5) (n = 6)(n = 6) 
PLC204 ± 19225 ± 220.9 ± 0.1200 ± 30336 ± 440.6 ± 0.1
(n = 5)(n = 5) (n = 6)(n = 6) 

At sites in the cortex where release was observed, the species detected was a catecholamine as revealed by the cyclic voltammogram (insert, Fig. 4b). Following release, the catecholamine disappeared from the electrode. The disappearance was fit to Michaelis–Menten kinetics assuming a Km of 0.210 µm in all locations, and the Vmax was determined (solid line superimposed on the data in Fig. 4b). Mean values for Vmax for the two cortical regions and the CP are summarized in Table 3. The value of Vmax in the PLC is approximately twice as large as in the Cg1, but it is almost one-thirtieth of the maximal uptake rate measured at locations of high PLAP density in the CP. In accord with the tissue content, maximal release amplitude with 30 pulse, 30 Hz stimulations in the Cg1 were approximately two-thirds of that in the PLC (0.32 ± 0.09 and 0.50 ± 0.05 µm, respectively). These levels are quite similar to those evoked in vivo with electrical stimulations of similar frequency (Garris and Wightman 1994). In contrast, a single pulse evoked 1.7 ± 0.1 µm DA in the CP. In this region, l-DOPA loading was not required to obtain measurable release.

Table 3.  Maximal rate for uptake (Vmax) in different brain regions of transgenic and knockout mice
Brain regionAnimalVmax (nm/s)n
  1. Uptake rates were measured following electrically evoked release of catecholamines with a locally applied stimulus. The stimulus consisted of a single pulse in the CP and 30 pulses at 30 Hz in the Cg1 and PLC. The measurements in the Cg1 and PLC were obtained from slices preloaded with l-DOPA. Results are given as the mean ± SEM. aanova single factor statistical analysis ( d.f. = 13, p < 0.001) reveals value is statistically significant from the wild type. bNo statistical difference found between knockout and wild type. Only a limited supply of NET KO animals were available for this study, so significance was not evaluated for these measurements.

CPPLAP3700 ± 40039
CPDAT KO12 ± 1a7
Cg1PLAP60 ± 926
Cg1DAT KO66 ± 17b4
Cg1NET KO35 2
PLCPLAP121 ± 2217
PLCDAT KO118 ± 46b4
PLCNET KO58 2

Effects of DAT, NET and MAO inhibition in Cg1 and PLC

To inhibit uptake by catecholamine transporters, competitive inhibitors were applied to the slice for an hour, and uptake following evoked release was compared with that prior to the drug exposure. An example is shown in Fig. 5(a) for 1 µm GBR-12909, a DAT inhibitor. Incubation with 10 nm GBR-12909 did not affect uptake in the Cg1 (data not shown, Km was 110 ± 10% of the predrug value), but uptake was slowed with 1 µm GBR-12909 (Table 4). In separate experiments, NET was inhibited with protriptyline. Neither 10 nm nor 1 µm (data not shown) protriptyline significantly altered uptake in the Cg1, but uptake was slowed with 10 µm protriptyline (a representative example is shown in Fig. 5b). The changes in Km for the highest doses of GBR-12909 and protriptyline obtained by kinetic modeling (as indicated by the solid lines in Fig. 5) are summarized in Table 4.

Figure 5.

Effects of pharmacological inhibition of DAT and NET in the Cg1 and PLC. Slices as in Fig. 4 were locally stimulated with the electrodes located in fluorescently labeled regions. The bar underneath the concentration traces indicates the beginning and end of the stimulation. Stimuli were repeated at five minute intervals and at least three measurements were averaged to improve signal to noise. Superimposed measurements were all made at the same location and are from a single slice. The measured responses are given by the points and the solid lines are simulations of the measured responses. (a,b) Measurements in the Cg1. (a) A predrug response superimposed on the response obtained following 1 h incubation with 1 µm GBR-12909. (b) Responses before and following one hour incubation with protriptyline (10 µm). (c,d) Responses in the PLC. (c) Responses before and after incubation with GBR-12909 (1 µm). (d) Responses before and after incubation with protriptyline (10 µm).

Table 4.  Change in apparent Km in the Cg1 and PLC of PLAP, DAT KO, and NET KO mice treated with uptake or metabolism inhibitors
Brain regionKm (% of predrug value
following 10 µm protriptyline)
Km (% of predrug value
following 1 µm GBR-12909)
Km (% of predrug value
following 10 µm pargyline)
  1. anova single factor statistical analysis (minimum of 7 d.f.). aSignificantly different from predrug value with p < 0.002. Only a limited supply of NET KO animals were available for this study so significance was not evaluated for these results.

Cg1 (PLAP)700 ± 10 (n = 12)a380 ± 70 (n = 10)a116 ± 11 (n = 4)
Cg1 (DAT KO)2500 ± 600 (n = 4)a  
Cg1 (NET KO) 1800 (n = 2) 
PLC (PLAP)4800 ± 160 (n = 6)a1300 ± 300 (n = 7)a112 ± 10 (n = 4)
PLC (DAT KO)1800 ± 500 (n = 4)a  
PLC (NET KO) 5000 (n = 2) 

In the PLC, 10 nm GBR-12909 significantly altered uptake (Km increase of 400 ± 130%, p < 0.05), and, at 1 µm, uptake was dramatically slowed (examples are shown in Fig. 5c). Incubation with 10 nm or 1 µm protriptyline (data not shown) resulted in only slight changes in uptake in the PLC. However, 10 µm protriptyline significantly altered the uptake rate (Table 4, an example is shown in Fig. 5d).

Another possible mechanism of removal of released catecholamine is metabolism by MAO. However, inhibition of MAO with 10 µm pargyline did not significantly affect clearance rates in either region (Table 4).

Inhibition of NET in the DAT knockout mouse

To further understand the contribution of the NET in the clearance described above, release and uptake were examined in the PFC in brain slices from DAT KO mice. As with the slices from PLAP mice, measurements were made after 30 minute incubation with l-DOPA (100 µm) and an hour perfusion period. Although catecholamine release sites could not be visualized, approximately 6% of the electrode placements yielded measurable release. In the sites where the release amplitudes were similar to those seen in sites of high PLAP localization, the Vmax values in the Cg1 and PLC were not significantly different from those measured in the PLAP animals (Table 3). In contrast, large differences in uptake rates were seen in the CP (Table 3), consistent with previous reports (Giros et al. 1996). In slices from DAT KO animals, incubation with 10 µm protriptyline in both the Cg1 and PLC resulted in a dramatic decrease in uptake (Table 4, representative examples in Figs 6a and b).

Figure 6.

Effects of inhibition of the NET in the DAT knockout mouse and DAT in the NET knockout mouse. Catecholamine release was evoked by 30 pulse, 30 Hz stimulations delivered at the time indicated by the horizontal bars. (a,b) Results obtained in slices from DAT KO mice. (a) Electrodes in Cg1 before and after 10 µm protriptyline. (b) Electrodes in the PLC before and after 10 µm protriptyline. (c,d) Results obtained in slices from NET KO mice. (c) Electrodes in Cg1 before and after 1 µm GBR-12909. (d) Electrodes in the PLC before and after 1 µm GBR-12909.

Inhibition of DAT in the NET knockout mouse

Release and uptake in brain slices from the NET KO mice were also monitored to examine uptake in animals with only DAT. Finding sites of release was similarly difficult as in slices from the DAT KO. The maximal rate for uptake (Vmax) for each region was about half of that obtained in the PLAP mice (Table 3). In slices from two mice, it was found that inhibition of DAT with 1 µm GBR-12909 in both the Cg1 and PLC virtually eliminated the rate of disappearance of released catecholamine (Table 4, representative examples in Figs 6c and d).

Discussion

The results in this paper are the first characterization of catecholamine uptake and release in the mouse PFC. In many respects, the anatomical findings reported here are quite similar to those in the rat PFC that has been quite thoroughly characterized (Lindvall et al. 1978; Lindvall and Bjorklund 1987; Sesack et al. 1998). Both NE and DA innervate the Cg1 and PLC of the mouse brain, as revealed by the antibodies to TH and DβH. Transporters for both catecholamines are also found in both regions of the PFC, but NET is distributed over a broader range. An advantage, however, of the mouse brain is the availability of genetically altered animals that allow specific aspects of brain function to be examined. In this study, we utilized a mouse that expresses PLAP on cells containing TH. Catecholamine-containing neurons can be visualized in respiring tissue through fluorescent antibody labeling. This approach allowed placement of voltammetric electrodes at sites in sparsely innervated regions with 10 µm resolution, and enabled real-time observation of local release and uptake in these regions. Uptake was found to be remarkably slow and revealed for the first time the apparently heterologous character of uptake on the spatial scale of a few microns. Indeed, in mice with a genetic KO of either one of the transporters, uptake was still operant throughout the PFC and governed the lifetime of the catecholamines in the extracellular space. Thus, extracellular catecholamines in the PFC are not restricted to their release sites, but their diffusion over long distances is restricted by uptake by both NET and DAT.

Central to the success of this work is the suitability of PLAP to serve as a marker for catecholamine neurons in the genetically altered mouse. The fluorescent antibody to PLAP enables the clear visualization of catecholaminergic cell bodies and also allows dendrites, terminals, and axons to be visualized. The labeling by the PLAP antibody of various anatomical regions corresponds well with the densities of these neurons previously reported using more traditional techniques in the rat brain (Thierry et al. 1973; Berger et al. 1974; Berger et al. 1976; Palkovits 1979; Lindvall and Bjorklund 1987; Sesack et al. 1998). We find that the catecholamine-containing neurons labeled with the PLAP fluorochrome conjugate are unaltered with respect to the kinetics of release and uptake (data not shown). Thus, the uptake values reported here are anticipated to be quite similar to those in slices prepared from wild-type animals.

The images of catecholaminergic cell bodies obtained with PLAP and TH antibodies clearly reveal their spatial colocalization. In the CP, a region of dense dopaminergic terminal density, close overlap of both TH and DAT with PLAP is found. As anticipated, colocalization of DβH with PLAP was found in regions of noradrenergic innervation since neurons that synthesize norepinephrine also synthesize dopamine.

In much of the PFC, regions of DAT and NET antibody binding were colocalized. This result is in contrast to the CP that was virtually devoid of NET antibody binding. Since saturation of binding sites was not attempted, we cannot use these images to quantitatively compare the density of NET and DAT in the PFC. However, qualitative comparisons of the same antibody–fluorochrome conjugate in different regions is possible, since the same PMT settings were utilized in each region. When labeled with DAT, fluorescence from the CP was much brighter than in the Cg1 and PLC, consistent with results in rat brain that show that the density of DAT is more dense in the CP than in the PFC (Sesack et al. 1998). Our ratios of areas of overlap of DAT with PLAP and NET with PLAP give a relative measure of the spatial distribution of DAT and NET sites in the PFC. These ratios indicate that NET is distributed quite similarly in the Cg1 and PLC, and in both regions the area of overlap of DAT with PLAP is less than the overlap of NET with PLAP. However, there is twice as much overlap area of DAT with PLAP the PLC compared with the Cg1. Thus, we conclude that DA and NE axons coexist in broad regions of both the PLC and Cg1 in the mouse brain, but that NET is more broadly distributed than DAT.

The bright PLAP antibody fluorescence, visible via epifluorescence with a CCD, enables voltammetric electrodes to be placed in close proximity to release sites. Since broad regions of the PFC (at least on the spatial scale of the voltammetric electrode) are devoid of catecholamine containing terminals, considerable searching is required in unlabeled tissue to find these sites. Despite the ability to visualize catecholaminergic processes, we were unable to measure release from these sites in freshly prepared slices of the PFC. It is well known that, as an animal is killed, catecholamine neurons release vast amounts of their neurotransmitter (Phebus et al. 1986). We reasoned that in a region with weak uptake such as the PFC, these stores might not be refilled by the transporter upon reoxygenation of the living tissue prepared from the mouse brain. Thus, l-DOPA was used to replenish the stores of catecholamines in the respiring brain tissue (Jones et al. 1994).

l-DOPA is transported into rat cerebral cortical tissue by an amino acid transport system (Garcia-Sancho and Herreros 1975). The system is distinct from the catecholamine transporters since it has a Km in the 10−5 m range and is Na+ independent. l-DOPA may also interact with the DAT since it decreases DA clearance rates in the basolateral amygdaloid nucleus of the rat brain (Jones et al. 1994). This effect is reversible, however, and was not apparent with the incubation times used in this work. l-DOPA increases DA releasable stores in vivo (May et al. 1988), in slices (Gibson 1988), and in cultured neurons (Puopolo et al. 2001). In this work, incubation with l-DOPA caused a selective increase in DA content in both the Cg1 and PLC (Table 2) without altering NE. This result is surprising since NE and DA synthesis rates are quite similar (Neff and Costa 1966). However, since it is likely that DA can be released from NE terminals, we believe that this is the catecholamine whose release was detected in both regions of the PFC (the electrode cannot distinguish between NE and DA).

The maximal catecholamine uptake rates in the PLC are twice as fast as in the Cg1, although both are remarkably slow in comparison to the CP (Table 3). The measured rates are quite similar to those measured in these regions in the rat brain for NE (Mitchell et al. 1994) and DA (Garris and Wightman 1994) uptake. Uptake by both DAT and NET is certainly playing a role in the clearance since the disappearance rates are 5–10 times greater in the PFC regions than in the CP from animals that are devoid of DAT. However, pharmacological inhibition of NET and DAT in the PFC regions was remarkably ineffective. For example, in the Cg1, the use of 1 µm GBR-12909, at a concentration 1000 times larger than Ki for DA uptake inhibition (Anderson 1989), caused little change in the uptake rate. This same dose was four times more effective in the PLC. In rats, a similar lack of sensitivity to inhibition of DAT in the PFC has been reported (Cass and Gerhardt 1995). Inhibition of clearance by the NET inhibitor, protriptyline, was also more effective in the PLC than the Cg1, but, again, to obtain measurable effects, concentrations much greater than its Ki (1 nm for NE uptake inhibition, Richelson and Pfenning 1984) were used. The low rates and poor response to pharmacological inhibition is likely a result of the low density of both NET and DAT in the cortical regions. Due to the temperature of our recordings (32°C), the Vmax values reported here for all regions may be lower than in vivo by perhaps more than a factor of two (Xie et al. 2000). Nevertheless, the data clearly show that the relative rates for catecholamine uptake in the cortical regions are much less than in the CP. Unlike results from rat brains (Wayment et al. 2001), we found no effect of MAO inhibition on catecholamine clearance rates.

In both the Cg1 and PLC, there is a greater spatial overlap of NET with PLAP than of DAT with PLAP, suggesting a greater role for NET in controlling catecholamine concentrations in the PFC. It has been previously shown that selective NE uptake inhibitors can increase extracellular DA levels in the rat PFC (Carboni et al. 1990; Di Chiara et al. 1992; Yamamoto and Novotney 1998). Since DA and NE are both good substrates for each others transporters, clearance by heterologous uptake of DA into NE terminals has been proposed as a controlling mechanism in rats (Carboni et al. 1990; Izenwasser et al. 1990; Pozzi et al. 1994; Cass and Gerhardt 1995). The data reported here support such a mechanism in the mouse brain. Furthermore, the slow rates of uptake also support the view that volume transmission can occur for catecholamines in the PFC (Garris and Wightman 1995; Sesack et al. 1998). For example, the slow rate of uptake in the Cg1 allows catecholamines to diffuse with a half life of more than 5 s before being intercepted by a binding site. As estimated previously, in this time interval, catecholamines could diffuse more than 100 µm (Mitchell et al. 1994). Although this is quite far when compared with the size of the synapse, it is still localized when compared with the dimensions of the PFC (i.e. twice the size of the dimension mark in Fig. 1), and it is sufficient to allow a released catecholamine to reach either DAT or NET.

The use of the NET and DAT KO mice enabled these ideas to be further explored. Release was not observed in the PFC of either animal unless the tissue was first incubated with l-DOPA. The DAT KO mice had similar Vmax values in the Cg1 and PLC as the PLAP mice and clearance was protriptylene sensitive. Thus, even in the absence of DAT, NET was capable of maintaining ‘normal’ rates of uptake in both PFC regions. Consistent with this, it has recently been proposed that clearance of DA by the NET plays a role in cocaine self administration in DAT KO animals because it might control extracellular DA in these animals (Carboni et al. 2001). In the limited number of NET KO mice evaluated, Vmax values were nearly half that seen in the PLAP mouse in both the Cg1 and PLC. This decrease in uptake rates supports the greater importance of the NET in both regions.

In conclusion, this study provides a local view of the extracellular lifetime of catecholamines in cortical regions of the mouse brain. The use of antibodies to DAT and NET enabled the proximal location of transporters in the Cg1 and PLC to be identified. The ability to label catecholaminergic neurons with the PLAP antibody–fluorochrome conjugate allowed sites of catecholamine release and uptake to be found in the Cg1 and PLC demonstrating the spatial heterogeneity of catecholamine release and uptake. The slow uptake in the Cg1 and PLC, apparently due to the low density of DA and NE transporters when compared with the CP, enables volume transmission over dimensions approaching 100 µm. Since uptake sites for both catecholamines are found within these dimensions, heterologous uptake can readily occur, at least in the mouse brain.

Acknowledgement

This research was supported by NIH (NS 38879).

Ancillary