|Home | About | Journals | Submit | Contact Us | Français|
Activation of the hypothalamic-pituitary-adrenal (HPA) axis is associated with increased extracellular serotonin (5-HT) in limbic brain regions. The mechanism through which this occurs remains unclear. One way could be via HPA axis-dependent impairment of serotonin transporter (SERT) function, the high-affinity uptake mechanism for 5-HT. Consistent with this idea, we found that 5-HT clearance rate in hippocampus was dramatically reduced in mice exposed to repeated swim, a stimulus known to activate the HPA axis. However, this phenomenon also occurred in mice lacking SERT, ruling out SERT as a mechanism. The organic cation transporter 3 (OCT3) is emerging as an important regulator of brain 5-HT. Moreover, corticosterone, which is released upon HPA axis activation, blocks 5-HT uptake by OCT3. Repeated swim produced a persistent elevation in plasma corticosterone, and consistent with prolonged blockade by corticosterone, we found that OCT3 expression and function was reduced in these mice. Importantly, this effect of repeated swim to reduce 5-HT clearance rate was corticosterone-dependent, as evidenced by its absence in adrenalectomized mice, in which plasma corticosterone levels were essentially undetectable. Behaviorally, mice subjected to repeated swim spent less time immobile in the tail suspension test than control mice, but responded similarly to SERT and norepinephrine transporter selective antidepressants. Together, these results show that reduced 5-HT clearance following HPA axis activation is likely mediated, at least in part, by the corticosterone-sensitive OCT3, and that drugs developed to selectively target OCT3 (unlike corticosterone) may be candidates for the development of novel antidepressant medications.
A consistent observation from preclinical studies is that activation of the hypothalamic-pituitary-adrenal (HPA) axis raises levels of extracellular fluid (ECF) serotonin (5-HT) in various forebrain regions, including hippocampus - a critical neural locus for antidepressant medications (Campbell and Macqueen, 2004; Leonardo and Hen, 2006). This effect appears to generalize across different types of HPA axis activators, including exposure to aversive stimuli (Wilkinson et al., 1996; Amat et al., 1998; Linthorst et al., 2002; Hajos-Korcsok et al., 2003; Beekman et al., 2005), environmental enrichment (Brenes et al., 2008; Xu et al., 2009) and exercise (Meeusen and De Meirleir, 1995; Gomez-Merino et al., 2001). These increases are not always associated with activation of serotonergic dorsal raphe neurons innervating the forebrain or by increased 5-HT turnover and release (Lanfumey et al., 2008), suggesting involvement of mechanisms regulating ECF levels of 5-HT after release.
The serotonin transporter (SERT) is the principle high-affinity, low-capacity uptake mechanism responsible for clearing released 5-HT from the extracellular space (Blakely et al., 1998). Whether HPA-axis activation negatively affects SERT function to produce increased ECF 5-HT has not been extensively investigated and the literature that does exist is mixed, with decreases (Maines et al., 1999; McKittrick et al., 2000; Jahng et al., 2007) or no change (Watanabe et al., 1993; Fernandez et al., 2001; Dias Soares et al., 2003; Williams et al., 2005; Chen et al., 2008) in SERT expression in ex vivo rodent forebrain preparations being reported. These discrepancies could be due to differences in the magnitude and/or chronicity of HPA-axis activation or the specific brain region examined. Methodological differences and inaccuracy of ex vivo measurements of SERT status could also be contributing factors, given none of the studies have directly examined SERT function in vivo. However, a highly feasible and intriguing possibility is that HPA-axis activation does not increase ECF 5-HT by impairing SERT-mediated clearance, but that other 5-HT clearance mechanisms are affected.
Of particular note is the organic cation transporter 3 (OCT3), which acts as a low-affinity, high-capacity transporter of 5-HT (Koepsell et al., 2007). We recently unmasked the contribution of OCT3 to 5-HT clearance in knockout mice lacking SERT, where OCT3 is upregulated (Baganz et al., 2008). Furthermore, and consistent with a contribution to HPA-axis-related impairment of 5-HT clearance, OCT3 is widely expressed in forebrain targets of glucocorticoids, including hippocampus (Gasser et al., 2009), and is functionally inhibited by corticosterone (CORT) (Gründemann et al., 1998; Gasser et al., 2006). Beyond this, however, possible functional interactions between the HPA-axis and OCT3 have yet to be determined.
Here we determine the effects of repeated swim on 5-HT clearance in mouse hippocampus in vivo, using high-speed chronoamperometry. We then employed a combination of mutant, pharmacological and behavioral approaches to elucidate the transporter mechanisms underlying these effects. Our data indicate that repeated HPA-axis activation induced by swim compromises 5-HT clearance in vivo. Moreover, we found this effect to be independent of SERT, but dependent upon corticosterone, and associated with loss of OCT3 expression and function.
In the majority of experiments male C57BL/6J mice were used and purchased from Jackson Lab (Bar Harbor, Maine, USA). In a subset of experiments, male SERT knockout (KO) mice (backcrossed to C57BL/6J for >10 generations) were used. We maintain a breeding colony of SERT wild-type (+/+), heterozygote (+/−) and homozygote KO (−/−) (originally provided by Dr. Dennis Murphy (NIMH)). SERT KO mice were generated and maintained as described previously (Bengel et al., 1998). Male mice weighing 25 to 30 g, were used for all experiments. All animal procedures were in strict accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All measures were taken to limit the number of animals used and to minimize animal discomfort.
Mice were housed on a standard light cycle with lights on at 7:00 a.m. and off at 7:00 p.m. In mice, circulating corticosterone levels typically peak around the onset of the active period (Moore and Eichler, 1972; Buijs et al., 1993; La Fleur et al., 1999; Li et al., 2006), which would equate to ~6:00 – 7:00 p.m. in our studies. All experimental manipulations and data acquisition took place prior to this peak. Mice were swum between 9:00 a.m. and 10:30 a.m. Tail suspension tests (TST) were carried out between 9:00 a.m. and 12:00 noon, chronoamperometry recordings were made over the period from 10:00 a.m. and 6:00 p.m. and trunk blood was collected at approximately 6:00 p.m.
We selected repeated daily exposure to 10 min of swim as it is a known activator of the HPA axis and produces transient increases in extracellular 5-HT. For example, Boyce-Rustay and co-workers (2007) showed that plasma corticosterone levels measured 30 min after either one 10 min swim or 14 days of 10 min/day swim were significantly elevated compared to non-swum control mice. Microdialysis studies revealed that exposure to one or two consecutive swims (2 hours apart) causes a rapid increase in ECF 5-HT in mouse hippocampus, peaking within a few min following swim and declining to baseline levels approximately 40 min later (Yoshitake et al., 2004).
Animals were housed in groups of two or four. Half of the mice in each group received a 10-min swim, for 14 days, while the other half remained in the home cage, as previously described (Boyce-Rustay et al., 2007). The swim tank consisted of a transparent Plexiglas cylinder (20 cm diameter) filled half way with water (24 ± 1 °C). Immediately following the swim, mice were towel dried before being returned to their home cage. In most instances separate cohorts of mice were used for the in vivo electrochemistry, neurochemistry and behavioral studies. Quantitative autoradioradiography was carried out using brains from mice that were harvested immediately following chronoamperometric recordings. All experiments commenced 24 hr after the final swim.
In vivo chronoamperometry was carried out according to the methods described in Daws and Toney (2007). We construct our own carbon fiber electrodes and a detailed description can be found in Williams et al., (2007). Recording electrode/micropipette assemblies were constructed using a single carbon-fiber (30 µm diameter; Specialty Materials Inc., Lowell, Massachusetts, USA), which was sealed inside fused silica tubing (Schott North America, Inc., Elmsford, New York, USA). We based our procedure for electrode construction on modifications of published methods (Gerhardt, 1995; Perez et al., 2006). Carbon fiber electrodes (30 µm tip diameter) were coated with Nafion® (5% solution, Aldrich Chemical Co, Milwaukee, WI), to prevent interference from anionic substances in extracellular fluid as previously described (Daws and Toney, 2007). Electrodes were tested for sensitivity to the 5-HT metabolite, 5-hydroxyindoleacetic acid (5-HIAA, 250 µM) and calibrated with accumulating concentrations of 5-HT (0–4 µM). Only electrodes displaying a selectivity ratio for 5-HT over 5-HIAA greater than 500:1 and a linear response (r2 ≥ 0.9) to 5-HT were used.
The electrochemical recording assembly consisted of a Nafion-coated, single carbon fiber electrode attached to a 4-barreled micropipette such that their tips were separated by ~200 µm. Depending on the experimental design (see details below), barrels were filled with combinations of either 5-HT (200 µM), histamine (200 µM), fluvoxamine (400 or 800 µM), corticosterone (2 mM) or vehicle phosphate buffered saline (PBS) +/− ethanol (EtOH) (6 mM). Serotonin and histamine were prepared in 0.1M PBS with 100 µM ascorbic acid added as an antioxidant and the pH adjusted to 7.4. The electrode-micropipette recording assembly was lowered into the CA3 region of the dorsal hippocampus (AP, −1.94 from bregma; ML, +2.0 from midline; DV, −2.0 from dura) (Franklin and Paxinos, 1997) of anesthetized mice. We chose to study the CA3 region because this is a neural target for both stress (Watanabe et al., 1992; Holmes & Wellman 2009) and antidepressant drugs (de Montigny et al., 1990), and because we have previously shown that 5-HT clearance in this region is mediated primarily by SERT (and not norepinephrine, dopamine or organic cation transporters) when exogenously applied 5-HT concentrations are 1 micromolar or less (Daws et al., 2005; Daws and Toney, 2007; Baganz et al., 2008; Daws, 2009).
For all experiments, mice were anesthetized by intraperitoneal (i.p.) injection (2 ml/kg body weight) of a mixture of chloralose (35 mg/kg) and urethane (350 mg/kg). A tube was inserted into the trachea to facilitate breathing and mice were then placed into a stereotaxic frame. Body temperature was maintained at 36–37 °C by a water circulated heating pad.
High-speed chronoamperometric recordings were made using the FAST-12 and FAST-16 systems (Quanteon, Nicholasville, Kentucky, USA) (Montañez et al., 2003). Oxidation potentials consisted of 100 msec pulses of +0.55 volts. Each pulse was separated by a 900 msec interval during which the resting potential was maintained at 0.0 V. Voltage at the active electrode was applied with respect to a Ag/AgCl reference electrode positioned in the extracellular fluid of the ipsilateral superficial cortex. The oxidation and reduction currents were digitally integrated during the last 80 msec of each 100 msec voltage pulse.
At the conclusion of the experiment, an electrolytic lesion was made to mark the placement of the electrode tip. The brain was removed, rapidly frozen on dry ice and stored at −80°C until use. At this time brains were thawed to −15°C and sliced into 20 µm thick sections for histological verification of electrode localization. Only data from mice in which the electrode was confirmed to be in the CA3 region of the hippocampus were included in data analyses. An insufficient number of placements falling outside this region prevented assessment of drug effects on 5-HT clearance in other hippocampal regions.
The electrodes were not post-calibrated in these studies. We have carried out post-calibrations in the past and do find that electrodes decline in sensitivity to 5-HT over time; depending on the duration of the experiment this decline can range from 5 to 50%. Importantly, however, this decline occurs at a very similar rate, therefore the inter-electrode variability between experiments is very constant. Because of this we ensure that the timing of protocols and drug challenges is kept as similar as possible between experiments (but ensuring a randomized design to avoid order effects of drug administration). That is, when comparing 5-HT clearance rates between control and swum mice, in wild-type or SERT KO mice, this is done in a time matched manner. In addition, after establishing a stable baseline, (i.e. 3 consecutive 5-HT signals of the same amplitude, time course and clearance rate), we routinely switch out electrodes during the experiment if we find that signal amplitude (the best indicator of lost sensitivity) for a given amount of 5-HT declines by more than 20%. We find this occurs in <5% of experiments and with equivalent frequency across treatment groups. Combined, these strategies minimize potential confounds that may be introduced by loss of electrode sensitivity.
It is known that some compounds can interfere with the sensitivity of carbon fiber electrodes to detect neurotransmitter (Davidson et al., 2000). Because of this we tested the effect of fluvoxamine, corticosterone and EtOH (vehicle for corticosterone) on electrode sensitivity for 5-HT. To do this, Nafion coated carbon fiber electrodes were calibrated to 5-HT (0.2 to 1.0 µM in 0.2 µM increments) in a beaker containing PBS according to standard protocol (Daws and Toney, 2007). Each electrode was calibrated twice, the first time in the absence of drug or vehicle and the second time after the addition of drug or vehicle. Drug concentrations were based on those estimated to reach the recording electrode following pressure-ejection into brain. Based on our own estimates and that of others (Gerhardt and Palmer, 1987; Daws et al., 2006) the concentration of drug pressure-ejected a distance of ~200 µm from the recording electrode is diluted in the order of 10- to 200-fold by the time it reaches the recording site. Thus, in these experiments we tested drugs at both the low and high end of this range. The concentrations tested were 2 µM and 80 µM for fluvoxamine, 10 µM and 200 µM for corticosterone and 30 µM and 600 µM for EtOH. An equivalent volume of PBS was added as the control.
Sensitivity was determined as the slope of the calibration curve. As expected under control conditions, the electrode’s sensitivity for 5-HT declined after the first calibration by ~30% (69 ± 15%, n=3, slope of second calibration as a percent of first). Neither corticosterone nor EtOH, at either concentration, changed the sensitivity of the Nafion-coated carbon fiber electrode for 5-HT compared to the PBS control. Expressed as a percent of the control condition the sensitivity of the electrode following corticosterone was 100 ± 26% (n=5, 10 µM and 200 µM pooled) and following EtOH, 115 ± 14% (n=6, 30 µM and 600 µM pooled). There was a non-significant trend for fluvoxamine (2.0 µM) to reduce the sensitivity of the electrode for 5-HT (PBS 100 ± 23% n=3, vs fluvoxamine 68 ± 15%, n=6). The higher fluvoxamine concentration (80 µM) reduced the sensitivity of the electrode for 5-HT (15 ± 3%, n=3). It is important to note however, that we do not find any evidence for fluvoxamine interfering with electrode sensitivity for 5-HT under our in vivo recording scenario. For example, 5-HT signals remain reproducible following intrahippocampal application of fluvoxamine (200 µM barrel concentration) to SERT knockout out mice (Montañez et al., 2003). Together, these results support the view that drugs are diluted much in excess of 10-fold, and perhaps even more than 200-fold, by the time they diffuse the 200 µm from the micropipette to the recording electrode. None of the conditions significantly changed the selectivity of the electrodes for 5-HT over 5-HIAA. The implication is that data presented in the current study are not confounded by drugs interfering with the sensitivity or selectivity of the carbon fiber electrode for 5-HT.
We first assessed basal 5-HT clearance in the CA3 region of the hippocampus of mice exposed to repeated swim (10 min/day for 14 days) or not exposed to swim. Serotonin (2.5, 5.0, 10.0, or 20.0 pmol) was locally pressure ejected into the hippocampus to obtain 5-HT signals with increasing amplitude (µM). Current induced by oxidation of 5-HT at the carbon fiber recording electrode was measured as a function of time and converted to micromolar units based on a calibration factor determined prior to the experiment in vivo (for details see Daws and Toney, 2007).
Exogenous 5-HT was applied into the CA3 region of the hippocampus by pressure-ejection (5–25 psi for 0.25–3.0 sec). Advantages of this approach are that clearance can be measured without an associated ‘release’ component and that measurements can be made in vivo with excellent temporal (msec) resolution. The amount of 5-HT pressure ejected was adjusted so that baseline peak signal amplitudes did not exceed 1.0 µM. By keeping signal amplitudes in this range we can maintain the sensitivity of the electrode for 5-HT for several hours, and importantly, provide conditions that favor SERT mediated 5-HT uptake.
Once reproducible 5-HT electrochemical signals were obtained, fluvoxamine or PBS vehicle was applied locally into the CA3 region of hippocampus 2 min before the next application of 5-HT. Different pmol amounts of fluvoxamine were delivered by varying the barrel concentration of drug (400 or 800 µM) as well as the volume ejected. This drug application protocol was chosen to cause minimal disturbance to the baseline electrochemical signal and to allow sufficient time for drugs to diffuse to the recording site. Three signal parameters were analyzed; the peak signal amplitude, the T80 time course parameter, defined as the time for the signal to decline by 80% of the peak signal amplitude and the Tc, a measure of clearance rate, defined as the slope of the most linear portion of the descending limb of the signal, between T20 and T60 (the time for the signal to decline by 20% to 60% of the peak signal amplitude).
It is important to note that the concentration of drug reaching the recording electrode is approximately 200 fold less than the concentration of drug within the barrel (Daws et al., 2006). Thus, barrel concentrations were determined to achieve concentrations of drug reaching the recording site in a range that would yield pharmacologically and physiologically relevant concentrations.
These experiments were carried out as described for fluvoxamine, with two notable exceptions. To favor detection of OCT3 mediated 5-HT clearance, 5-HT signals in a range of 2–4 µM were obtained. Second, the barrel concentration of corticosterone was 2 mM. We have previously shown that a corticosterone barrel concentration of 100 µM does not affect 5-HT clearance in C57BL/6J mice, but does impede 5-HT clearance in SERT deficient mice where OCT3 expression is elevated (Baganz et al., 2008). Therefore, we used this relatively high concentration of corticosterone to increase the likelihood of detecting an ability of corticosterone (an OCT3 blocker) to inhibit 5-HT clearance in C57BL/6J mice exposed to repeated swim, compared to control mice. The vehicle for corticosterone was PBS containing 6 mM EtOH. This would yield a concentration of approximately 0.03 mM EtOH at the recording site, a concentration that does not influence 5-HT clearance when signal amplitudes are kept below 1 µM (Daws et al., 2006).
To directly assess the function of OCT3, we measured clearance of the high-affinity substrate, histamine (200 µM barrel concentration and delivering 140 pmol), in the CA3 region of the hippocampus of mice exposed to 0 or 14 days of swim using chronoamperometry. These experiments were carried out in the same way as described above for 5-HT clearance with the following exceptions. The sensitivity for carbon fiber electrodes is much lower for histamine than 5-HT. Therefore, to increase the surface area of the electrode we created micro-fractures in the carbon fiber by applying a potential of +1.50 V for 100 msec and stepping back to a resting potential of −1.50 V for 100 msec for a period of 10 min prior to coating the electrode with Nafion, calibrating in vitro and recording in vivo. Histamine is oxidized at a higher potential than 5-HT, and so for in vivo recordings 100 msec pulses of +1.0V were applied (compared to 0.55V for 5-HT recordings). Each pulse was separated by a 900 msec interval during which the resting potential was maintained at 0.0 V.
A separate cohort of animals was exposed to repeated swim, or not, as described above and 24 hr following the final swim, mice were sacrificed by rapid decapitation; brains were rapidly removed, frozen on dry ice, and stored at −80°C until use. Coronal, 20 µm brain sections were cut in a cryostat and mounted onto gelatin-coated slides.
To determine SERT expression level, we used quantitative autoradiography to measure binding of the SERT-specific radioligand, [3H]-cyanoimipramine (CN-IMI) (80–85 Ci/mmol: American Radiolabeled Chemicals, St. Louis, Missouri, USA) in hippocampus of whole brain sections, as described by Hensler et al. (1994). Slide-mounted sections were incubated with CN-IMI (1 nM) in a buffer of 50 mM Tris, pH 7.4, and 120 nM NaCl at 4°C for 24 hr. Nonspecific binding was defined in the presence of 1 µM fluvoxamine. Sections were then washed in cold buffer (4°C) and dipped in cold distilled water. At the end of the autoradiographic assay, sections were dried on a slide warmer at 60°C and apposed to Kodak Biomax MR film (Amersham, Piscataway, New Jersey, USA) for 4 weeks to generate autoradiograms. Analysis of the digitized autoradiograms was performed using the image analysis program NIH Image, version 1.47 (NIH, Bethesda, Maryland, USA), and optical densities of brain images were converted to femtomoles/mg protein (fmol/mgpt). The concentration of [3H]-CN-IMI used is approximately 8 times its KD value (Kovachich et al., 1988) so the values obtained approximate Bmax values.
Brains were removed from mice sacrificed by rapid decapitation and hippocampi were dissected out on ice and stored at −80°C until use. Samples were suspended in 500 µL of ice cold PBS (0.1 M, pH 7.4) and pulsed in centrifuge. The supernatant was removed, and the pellets were then homogenized in 500 µL of homogenizing buffer (25mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 25mM sucrose, 1.5mM MgCl2, 50mM NaCl, pH=7.2, aprotinin, leupeptin, pepstatin, phenylmethanesulfonylfluoride (PMSF)) using glass Potter-Elvehjem tissue grinders. Total protein concentrations of each sample were measured by Bradford Protein Assay (Bradford, 1976). Samples were solubilized in a buffer containing bromophenyl blue, heated for 5 minutes at 95°C, and separated by electrophoresis using SDS-polyacrylamide gel electrophoresis (SDS-PAGE). Protein was then transferred to a polyvinylidine difluoride (PVDF) membrane (Immobilon-P, Millipore, Bedford, MA, USA). The membrane was blocked in 5% nonfat dry milk in 1X TBS-Tween at room temperature for 1 hour, washed twice with 1X TBS-Tween, and incubated overnight at 4°C with polyclonal OCT3 antibody (Alpha Diagnostics International, San Antonio, Texas, USA; Catalogue # OCT31-A). The primary antibody was then removed, and the membrane was washed 3 times and incubated for 1 hour at 4°C with the secondary ECL donkey anti-rabbit IgG horseradish peroxidase-linked antibody (1:2500) (GE Healthcare, Little Chalfont, Buckinghamshire, UK). Bound antibody was detected on x-ray film using enhanced chemiluminescence reagents (GE Healthcare, Little Chalfont, Buckinghamsire, UK). Note that affinity purified rabbit polyclonal OCT3 primary antibodies used here were generated against a 19 amino acid C-terminal cytoplasmic domain of mouse OCT3. This peptide has no significant sequence homology with other mouse OCTs. Together with our previously published immunohistochemical data using the same antibody and controlling for non-specific staining (Baganz et al., 2008), the antibody used in these studies appears to be highly specific for OCT3.
Mice were anesthetized with isoflurane (3%) and bilateral adrenalectomy performed according to Castro (1974). Sham-operated animals were surgically incised to visually identify intact adrenal glands. Incisions were sutured and animals were allowed to recover from surgery for 1 week prior to exposure to repeated forced swim. Drinking water was replaced with 0.45% saline for ADX mice.
Immediately following chronoamperometry recordings in ADX or sham operated mice (i.e. ~30 hr after the last swim), the animals were sacrificed, and trunk blood was collected for analysis of corticosterone content. Blood was centrifuged (6000 min−1 for 15 min) with 300 µL ethylenediaminetetraacetic acid (EDTA, anticoagulant) and plasma was separated according to methods described in Li et al., 1999 (and Dr. Li at Univ Kansas, Lawrence, Kansas – personal communication). Plasma corticosterone concentration was measured using a radioimmunoassay kit purchased from MP Biomedicals, LLC (Diagnostic Division, 29525 Fountain Pkwy. Solon, Ohio 44139, Cat No. 07120102) as described by the manufacturer.
A separate cohort of mice was used to assess the effect of the SSRIs fluvoxamine and fluoxetine, as well as the selective norepinephrine uptake inhibitor, desipramine (DMI) on behavior. Mice were exposed to 0 or 14 days of swim as described. Twenty-four hr after the final swim, mice were administered fluvoxamine (30 mg/kg), fluoxetine (15 mg/kg), desipramine (DMI) (15 mg/kg), dissolved in 0.9% v/v saline, injected intraperitoneally at 10 ml/kg body weight), or saline control. Thirty min following injection, the TST was carried out as previously described (Steru et al., 1985). Mice were secured and suspended by the distal end of the tail to a flat metallic surface for 6 min. The presence or absence of immobility, defined as the absence of movement other than passive swaying, was sampled every 5 s over 6-min by a trained observer who remained blind to treatment (Boyce-Rustay et al., 2006).
Data are presented as the mean and standard error of the mean and analyzed with ANOVA followed by Bonferroni or Newman-Keuls post-hoc comparisons or t-tests for independent samples. The threshold for statistical significance was set at p < 0.05
Serotonin, 5HIAA, histamine, corticosterone, desipramine, fluvoxamine, urethane and chloralose were purchased from Sigma (Houston, Texas, USA). Fluoxetine was purchased from LKT Laboratories (St. Paul, Minnesota, USA).
To determine if impaired SERT activity could account for increased ECF 5-HT in hippocampus following repeated swim (Yoshitake et al., 2004) we first measured clearance of exogenously-applied 5-HT from ECF in the CA3 region of the hippocampus in vivo, using high-speed chronoamperometry. Male C57BL/6J mice were subjected to 14 consecutive days of 10 min/day swim and, 1 day later, 5-HT clearance was measured, under anesthesia, in comparison to age/weight-matched mice not exposed to swim. We found that 5-HT clearance kinetics in controls were similar to those previously reported for mice on this genetic background (Montañez et al., 2003). However, clearance rate was markedly reduced in mice with a history of repeated swim, as compared to controls (Figure 1A,B). Specifically, while clearance rate increased with increasing concentrations of 5-HT (analysis of variance (ANOVA): F3,54=10.98; P<0.01), clearance rate was significantly slower in mice with a history of repeated swim than controls (F1,54=23.96; P<0.01) (Fig 1A,B). There was also a significant interaction between repeated swim and 5-HT concentration (F3,54=4.31, P<0.01) due to slower 5-HT clearance rate in mice exposed to repeated swim compared to controls at the highest (10 and 20 pmol) 5-HT concentrations (Bonferroni post hoc tests: P<0.05).
We found that reduced 5-HT clearance rate was associated with a decreased time course for clearance; quantified as the time taken for the 5-HT signal to decay to 80% of its peak amplitude (T80). T80 values in mice exposed to repeated swim were significantly longer than in controls at the highest (20 pmol) 5-HT concentration (effect of swim: F1,54=9.42; P<0.01; Bonferroni post hoc: P<0.05) (Fig 1C). Importantly, while 5-HT signal amplitude showed an expected increase with increasing 5-HT concentration (effect of concentration F3,54=14.92, P<0.01), this measure did not differ between swum and control groups (Fig 1D). This indicates that slower 5-HT clearance in mice exposed to repeated swim was not an artifact of factors that can affect 5-HT diffusion from the micropipette through the extracellular milieu to the recording electrode, such as changes in volume fraction or tortuosity. Thus, the major novel result of this experiment is that repeated swim produced a profound impairment in hippocampal 5-HT clearance in vivo. The finding that this deficit was selectively manifested at high 5-HT concentrations (>5 micromolar) was notable, because this is a concentration range where non-SERT clearance mechanisms are known to begin to exert an influence over clearance (for review see Daws, 2009; Daws et al., 2005; Baganz et al., 2008). Thus, although SERT would be the most likely a priori mechanism explaining the 5-HT clearance-impairing effects of repeated swim, other transporter mechanisms could also be involved. To address this issue, we first conducted a set of experiments to establish whether SERT function was altered by repeated swim.
To determine the consequences of repeated swim for SERT function, we began by testing whether the SERT-inhibiting effects of the SSRI, fluvoxamine (FVX), were intact after repeated swim. Assessing the capacity of FVX to inhibit 5-HT clearance provides a sensitive measure of changes in SERT function (Montañez et al., 2003). One day after the repeated swim regime, we locally-applied increasing concentrations of FVX into the CA3 hippocampus and then examined clearance of exogenously applied 5-HT using chronoamperometry, as above. Here, we limited the concentration of 5-HT reaching the recording electrode to <1 micromolar in order to optimize conditions for detecting SERT, as opposed to non-SERT, mediated 5-HT uptake (Daws and Toney, 2007; Baganz et al., 2008).
Consistent with our results above (Fig 1B,C) baseline, pre-FVX, clearance of this low 5-HT concentration was not different between repeated swim and control groups (Tc: 6.41±0.30 nM/s and 6.98±0.47 nM/sec; T80: 81±5 s, and 84±5 s, for the same signal amplitude: 0.53±0.01µM and 0.56±0.02 µM; n=21 and n=14, respectively), providing further evidence that repeated swim does not affect clearance when SERT is the dominant transport mechanism. In line with our previous findings (Montañez et al., 2003), FVX dose-dependently inhibited 5-HT clearance (F4,63=5.79; P<0.01). However, 5-HT clearance after FVX did not differ between mice subjected to repeated swim and control mice, as measured by T80 time course (Fig 2A) as well as clearance rate (Fig 2B). Furthermore, in CA3 hippocampal tissue taken at the completion of the chronoamperometry analysis and processed for quantitative autoradiography with the SERT-selective radioligand [3H]-cyanoimipramine (CN-IMI), we found no difference in maximal SERT-binding between groups (1192 ± 83 fmol/mgpr vs 1261 ± 69 fmol/mgpr for control and mice exposed to repeated swim respectively). Thus, the retention of the ability of FVX to inhibit 5-HT clearance, together with the lack of changes in SERT expression, demonstrates there was no clear loss of SERT function following repeated swim. One caveat here is that FVX and 5-HT might target different binding pockets within SERT (Anderson et al., 2009), and therefore our FVX data cannot rule out the possibility that repeated swim impaired SERT function by interfering with the translocation of 5-HT via SERT.
As another means to assay SERT function we turned to a behavioral paradigm. Wild-type C57BL/6J mice were subjected to repeated swim and tested on the tail suspension test (TST) one day later. The TST is a simple measure of ‘behavioral despair’ that is reliably sensitive to acute treatment with antidepressant drugs; which reduce the time mice spend immobile (e.g. Cryan and Holmes, 2005). We tested groups of swum and non-swum mice 30 min after i.p. injection of either saline, the SERT-targeting antidepressants FVX and fluoxetine (FLX), or the norepinephrine transporter (NET)-targeting antidepressant desipramine (DMI). This provided another means to evaluate possible changes in SERT (and NET) function after repeated swim, this time at the behavioral level.
We found that FVX, FLX and DMI treatments all significantly and equivalently reduced immobility time, regardless of swim history (effect of drug: F3,40=59.98; P<0.01) (Fig 2C). Time spent immobile (scored every 5 s over a 6 min test as the absence of movement other than passive swaying), was 38 ± 2 counts and 29 ± 2 counts for non-swum mice and mice exposed to repeated swim respectively (t14=3.54; P<0.01); (i.e. 53 ± 2% and 41 ± 3% of a total of 72 possible counts were spent immobile in non-swum and swum mice respectively). These data provide further evidence that SERT (and NET) function is unaltered by repeated swim, and that functional alteration in other transporter mechanisms likely drive the reduced rate of 5-HT clearance observed in our in vivo electrochemistry studies.
To definitively rule out a role for SERT, we turned to a gene mutant approach. Mutant mice with constitutive gene KO of SERT (Murphy et al., 2008) and maintained on a congenic C57BL/6J background were exposed to repeated swim and then, 1 day later, clearance of 5-HT from ECF of hippocampus measured, using the same procedures as above. As expected (Montañez et al., 2003), clearance was significantly slower in control SERT KO mice, relative to the values reported above in control C57BL/6J mice tested under the same experimental conditions. This was apparent for clearance rate at the 10 and 20 pmol 5-HT concentrations (effect of genotype: F3,89=22.06; P<0.01, Bonferroni post hocs: P<0.01) (Fig 3A). Strikingly, however, not only did SERT KO fail to abolish the effects of repeated swim on 5-HT clearance, but these effects were actually augmented in the KO mice. That is, repeated swim significantly reduced 5-HT clearance at 10 and 20 pmol, below the already compromised basal 5-HT clearance rate in SERT KO (i.e., no swim control) (Bonferroni post hocs: P<0.01) (Fig 3A). The effects of repeated swim and SERT genotype are summarized for the 20 pmol concentration in Fig 3B. Importantly, as we showed in Figure 1D, while 5-HT signal amplitude showed an expected increase with increasing 5-HT concentration (effect of concentration F3,89 = 17.13, P<0.01), this measure did not differ between swum and control groups (Fig 3C).
Taken together, this series of experiments makes two important points. First, while repeated swim clearly impairs 5-HT clearance, these effects are not associated with any observable changes in SERT function or expression. The second major conclusion is that the 5-HT clearance-impairing effect of repeated swim is retained in the absence of SERT. Therefore, loss of SERT function is unlikely to account for the impairment in 5-HT clearance, and other transporter mechanisms appear to be involved. We next explored one such possible mechanism.
OCT3 is a low-affinity, high-capacity 5-HT transporter (Koepsell et al., 2007) that we and others have shown to regulate 5-HT clearance predominantly under conditions of high 5-HT, and to be upregulated in SERT KO mice (Chen et al., 2001; Feng et al., 2005; Baganz et al., 2008). To test whether OCT3 downregulation might underlie the sluggish rate of 5-HT clearance after repeated swim, we first performed western blot analysis of OCT3 protein levels in whole hippocampus of C57BL/6J mice with a history of repeated swim. OCT3 density values (normalized to β-actin) were significantly lower (t18=2.40; P<0.05) after repeated swim, relative to non-repeated swim counterparts (Fig 4A). To determine OCT3 function in vivo, we turned to measurement of histamine clearance. Histamine is a substrate for OCT3 but not other biogenic amine transporters (Schneider et al., 2005; Gasser et al., 2006). Supporting the utility of histamine clearance as a probe for OCT3 function, we previously found that histamine clearance is faster, and inhibited more robustly by the OCT3 antagonist D-22, in SERT KO mice, where OCT3 is upregulated (Baganz et al., 2008). Therefore, we would predict an impairment in histamine clearance if repeated swim impairs OCT3 function. Consistent with this prediction, when examined one day after the last swim, the T80 time course of locally-applied histamine clearance in CA3 hippocampus was significantly longer in C57BL/6J mice exposed to repeated swim than in controls (t6=2.99; P<0.05) (Fig 4B). Clearance rate of histamine was also significantly slower in the swum mice than in controls (t6=9.31, P<0.01) (Fig 4C). Thus, both hippocampal OCT3 expression and an OCT3-mediated clearance function are reduced by repeated swim.
We recently demonstrated that when the concentration of applied 5-HT was low (signal amplitude <1.0 µM), FVX efficiently inhibited hippocampal 5-HT clearance, while the OCT3 antagonist, D-22, was ineffective (Baganz et al., 2008). Conversely, at higher 5-HT concentrations (> 2.0 µM), the pattern of drug effects was reversed – with D-22, but not FVX, inhibiting 5-HT clearance (Baganz et al., 2008). This pattern of drug effect is consistent with OCT3 operating as a low-affinity, high-capacity transporter for 5-HT. It also predicts that antagonists of OCT3, such as corticosterone, should exert demonstrable effects on 5-HT clearance only when 5-HT concentrations are high, as measured by chronoamperometry. Indeed, we previously found that hippocampally-applied corticosterone (55 pmol) did not affect clearance of exogenously applied 5-HT when the concentration of 5-HT was low and OCT3 is putatively quiescent (Baganz et al., 2008).
Here, we tested whether corticosterone would reduce 5-HT clearance at a high, OCT3-recruiting, 5-HT concentration (> 2.5 µM), and also whether this effect would be blunted by repeated swim due to OCT3 downregulation. One day after subjecting C57BL/6J mice to 14 days of 10 min/day repeated swim, we locally-applied corticosterone (PBS or EtOH vehicle) and measured 5-HT clearance in CA3 hippocampus. We used a corticosterone concentration of 2 mM in the micropipette in order to achieve corticosterone concentrations at the recording electrode in the range of 10 micromolar after accounting for dilution through the extracellular matrix (Daws et al., 2006).
The first notable finding was that corticosterone significantly reduced 5-HT clearance rate as compared to PBS (effect of drug: F2,29=21.41; P<0.01) (Fig 5A). This demonstrates that exogenous corticosterone is sufficient to mimic the 5-HT clearance-disrupting effects of repeated swim under conditions where OCT3 is a dominant transport mechanism. One potential caveat to note here is that corticosterone was diluted with a small amount of EtOH (< 0.8 nmol) and it is possible that this contributed to the clearance-inhibiting effect of 5-HT given evidence that, at albeit much higher concentrations (10 nmol), EtOH inhibits 5-HT clearance under high endogenous 5-HT concentrations (Daws et al., 2006). Notwithstanding, the major finding was that while corticosterone reduced 5-HT clearance rate both in mice subjected to repeated swim and control mice, the magnitude of the effect was significantly less in repeatedly swum mice (effect of swim: F1,29=25.65; P<0.01) (Fig 5A). This finding is consistent with a model in which repeated swim causes functional downregulation of OCT3 to produce loss of hippocampal 5-HT clearance. In our next set of experiments, we interrogated this model further by asking whether disrupting a step upstream of OCT3 downregulation would block the effects of repeated swim.
The glucocorticoid, corticosterone, is released following HPA-axis activation by a wide range of environmental factors, including stressors and exercise, and targets various brain regions especially the CA3 subregion of the hippocampus (McEwen, 1999). As noted, we have previously shown that repeated swim causes significant corticosterone release in C57BL/6J mice even after 14 days (Boyce-Rustay et al., 2007). In much the same way that biogenic amine transporters, including SERT (Hirano et al., 2005) and NET (Weinshenker et al., 2002), are downregulated by repeated administration of transporter blocking drugs, repeated swim induction of corticosterone release could cause OCT3 downregulation, in turn producing the accompanying deficit in 5-HT clearance we observe.
To test this hypothesis, we eliminated corticosterone by adrenalectomizing (ADX) C57BL/6J mice prior to repeated swim. We examined 5-HT clearance 1 day after the final swim; limiting our analysis to one, high concentration of 5-HT (20 pmol), at which repeated swim effects were most clearly manifest in our earlier experiments. Replicating our earlier experiment, 5-HT clearance rate was significantly slower in sham mice subjected to repeated swim than in sham non-swum controls (effect of swim: F1,20=4.47, P<0.05) (Fig 6A). By contrast, while ADX had no effect on clearance rate in non-swum control mice, the reduction in clearance rate produced by repeated swim was completely prevented by ADX (Fig 6A). Radioimmunoassay (RIA) analysis of plasma samples taken immediately after the chronoamperometry recordings (i.e. approximately 30 hr after the last swim), confirmed that corticosterone was elevated in non-ADX swum mice relative to non-ADX, non-swum controls (effect of swim: F1,20=5.89, P<0.05), and that corticosterone was essentially absent in ADX mice (irrespective of swim history) (effect of ADX: F1,20=262.50, P<0.01) (Fig 6B).
Taken together, these results suggest that corticosterone is necessary for the repeated swim-induced deficit in hippocampal 5-HT clearance. Furthermore, they lend further support to our hypothesis that this deficit is driven by loss of function of OCT3 and not other transporters, such as the corticosterone-insensitive plasma membrane monoamine transporter (PMAT).
A consistent finding in the literature is that activation of the HPA axis is followed by an elevation in ECF 5-HT in limbic brain regions, including hippocampus (Wilkinson et al., 1996; Amat et al., 1998; Linthorst et al., 2002; Hajos-Korcsok et al., 2003; Beekman et al., 2005). Using repeated swim to activate the HPA axis, the major finding of the present study is that inhibition of 5-HT clearance from ECF is one mechanism through which this occurs. However, we found that the decrease in 5-HT clearance rate following repeated swim was not a consequence of impaired function or expression of the high-affinity uptake mechanism for 5-HT, the SERT, a finding consistent with reports that other HPA-axis activators, including repeated restraint (Watanabe et al., 1993), fox odor exposure (Dias Sores et al., 2003), corticosterone ingestion (Fernandez et al., 2001), 3,4-methylenedioxymethamphetamine administration (Williams et al., 2005), and exercise (Chen et al., 2008), fail to alter SERT expression or [3H]5-HT uptake in in vitro preparations from various brain regions, including hippocampus.
An important observation is that 5-HT clearance was most robustly impaired at concentrations where we have previously shown in hippocampus (using in vivo chronoamperometry) that SERT is less likely to dominate 5-HT uptake, but where low-affinity – high-capacity transporters for 5-HT, such as OCT3, are known to come into play (Daws, 2009). Indeed, at SERT-predominating concentrations, our in vivo studies revealed no effect of repeated swim on hippocampal 5-HT clearance rate and no effect on the ability of the SERT blocker, FVX to inhibit 5-HT clearance. Moreover, the effect of repeated swim to reduce 5-HT clearance (at non-SERT predominating concentrations) was maintained in SERT KO mice, ruling out SERT as a mechanism and strongly implicating the involvement of another transporter capable of 5-HT uptake. A particularly salient transporter in this regard is the corticosterone-sensitive OCT3. Our subsequent experiments interrogated its potential role in repeated swim induced impairment of 5-HT clearance.
A role for OCT3 in regulating 5-HT neurotransmission has only recently been recognized. Some of the first studies by Feng and co-workers (2005) showed that infusion of the OCT3 blocker, D-22, into the hypothalamus of rats increased ECF levels more so than an equivalent dose of SSRI. These findings suggest that ECF 5-HT levels in this brain region under baseline conditions are sufficiently high to recruit OCT3-mediated 5-HT uptake. Feng and co-workers went on to show that infusion of corticosterone, an OCT3 antagonist, potentiated the elevation in ECF 5-HT after administration of d-fenfluramine, a 5-HT releasing agent (Feng et al., 2009), and more recently that mild restraint potentiated the increase in 5-HT following infusion of D-22 into the hypothalamus (Feng et al., 2010). Together these studies suggest that corticosterone released in response to HPA-activation may indeed produce increases in ECF 5-HT by blocking 5-HT clearance via OCT3.
Consistent with this idea, we found that repeated HPA-activation, by means of repeated swim, led to decreased expression and function of OCT3 in hippocampus. It is known that repeated blockade of SERT or NET with selective antagonists leads to decreased expression and function of these transporters (Weinshenker et al., 2002; Hirano et al., 2005). Our data are in keeping with repeated blockade of OCT3 by corticosterone producing analogous effects on OCT3 expression and function. Reduced OCT3 protein was associated with a decrease in the clearance rate of histamine, a substrate for OCT3 but not SERT. Moreover, there was a blunted ability of intrahippocampally applied corticosterone to inhibit 5-HT clearance in mice exposed to repeated swim. Perhaps the most compelling evidence for a role of the corticosterone-sensitive OCT3 in mediating the effect of repeated swim to slow 5-HT clearance was the finding that this effect was lost in ADX mice, devoid of endogenous corticosterone. Certainly other biological signaling molecules, such as mineralocorticoids and androgens, that are likely reduced or eliminated after ADX cannot be ruled out. However, given the high affinity of OCT3 for corticosterone and our results showing inhibition of 5-HT clearance by exogenous application of corticosterone (Fig 5), makes corticosterone a likely candidate.
Of course there are many other transporters capable of 5-HT uptake. For example, OCT1, OCT2, PMAT, NET and dopamine transporter (DAT) are capable of clearing 5-HT with low-affinity/high-capacity (Engel et al., 2004; Zhou et al. 2007; for reviews see Koepsell et al., 2007 and Daws, 2009). However, we believe that their contributions to the results of the present study are minimal, if any, for several reasons. Primary among them are our findings measuring histamine clearance in hippocampus. Histamine is a substrate for OCT3 and not other biogenic amine transporters including SERT, NET and DAT (Gasser et al., 2006; Schneider et al., 2005). OCT1 and OCT2 transport histamine with very low efficiency, if at all, (Gründemann et al., 1999; Koepsell et al., 2007). In addition, OCT2 and DAT are sparsely expressed in hippocampus (Javitch et al., 1985; Gorboulev et al., 1997; Alnouti et al., 2006; Koepsell et al., 2007). Likewise, although PMAT is highly expressed in hippocampus it has relatively low histamine transport efficiency (Engel et al., 2004; Gasser et al., 2006). A key difference between OCT3 and PMAT, that is particularly salient in the context of the current study, is that corticosterone acts as an antagonist of OCT3 but not PMAT (Engel et al., 2004; Gasser et al., 2006; Koepsell et al., 2007). Corticosterone is also about 100-fold more selective for OCT3 than OCT1 or OCT2 (Hayer-Zillgen et al., 2002). Taken together, our data point to two likely mechanisms by which repeated swim impairs 5-HT clearance. The first, via the acute actions of corticosterone to block OCT3 mediated 5-HT clearance and second, via the effect of chronic OCT3 blockade by corticosterone (produced by repeated swim) to decrease OCT3 expression and function. One other mechanism that cannot be discounted is 5-HT competing with itself for uptake by OCT3. For example, as a result of OCT3 down-regulation and persistently higher levels of corticosterone in mice exposed to repeated swim, basal levels of ECF 5-HT might be predicted to be higher than those in control mice. A higher ambient level of ECF 5-HT could potentially contribute to the slower clearance rate of exogenously applied 5-HT in mice exposed to repeated swim.
The ability of the SERT selective blockers, FLX and FVX, and the NET selective blocker, DMI, to decrease immobility time in the TST was not diminished. This provides a behavioral readout in support of our contention that changes in neither SERT nor NET underlie the effect of repeated swim to slow 5-HT clearance from extracellular fluid in brain. We also observed that immobility time in saline injected mice exposed to repeated swim was modestly, but significantly less than in saline injected unswum mice. Clearly more comprehensive behavioral analyses are required before the significance of this observation can be clarified. However, it is a finding supported by literature showing that, contingent upon dose and duration of treatment, exogenously applied corticosterone can reduced immobility time (Stone and Lin, 2008; Zhao et al., 2009). Similarly, environment enrichment (another activator of the HPA-axis) that leads to similar increases in plasma corticosterone in mice as reported here, also led to decreased immobility time in the TST (Xu et al., 2009). The effect of environmental enrichment to reduce immobility time was prevented by ADX (Xu et al., 2009). There are also reports that elevated endogenous corticosterone is sometimes associated with antidepressant- and anxiolytic-like effects in rats and (C57BL/6J) mice (Swiergiel et al., 2008; Mozhui et al., 2010).
While further studies are needed, particularly in light of the limitations of the TST and FST as assays for ‘depression’ (Cryan & Holmes, 2005), these data provide important behavioral parallels to our neurochemical results, supporting lack of involvement of SERT and NET in mediating reduced 5-HT clearance after repeated swim.
Our data show that repeated swim leads to impaired 5-HT clearance via a corticosterone-sensitive transporter, most likely OCT3. Repeated swim caused a modest, but persistent elevation in plasma corticosterone and corticosterone is a potent blocker of OCT3. Thus, consistent with our findings that the OCT3 blocker, D-22, produces antidepressant-like effects in mice with elevated extracellular 5-HT (Baganz et al., 2008) and those of others (Kitaichi et al., 2005), the present studies support the idea that modest elevations in plasma corticosterone may act in the same way to produce antidepressant-like effects. Clearly corticosterone itself is not a viable candidate for the treatment of depression due to its many and varied actions at other receptors and clear pro-depressant effects under many conditions (for review see McEwen, 2008). However, our data build on growing evidence that OCT3 is an important modulator of brain 5-HT and a promising candidate for the development of novel, OCT3 selective drugs for the treatment of depression and related disorders.
We thank Drs. Randy Blakely, Ana Carneiro, Georgianna Gould, Julie Hensler, Wouter Koek, and Glenn Toney for helpful discussions. We also thank Teri Frosto-Burke, Jaclyn Munn, David Aguilar and Steven Alvarado for excellent technical assistance. This work was supported by National Institutes of Health Grant R01-MH6448 (L.C.D.), a NARSAD Independent Investigator Award (L.C.D.) and NIAAA intramural research program (A.H.).