Glucocorticoid programming of dopamine neurons. Virdee et al.

Materials and Methods(full details)

All procedures were conducted under license in accordance with the United Kingdom Animals (Scientific Procedures) Act of 1986.

Animals

Sprague-Dawley rats were purchased for timed matings (Harlan Olac, Blackthorn, Beicester, Oxfordshire, UK) as described previously (McArthur et al., 2005; McArthur et al., 2006; McArthur et al., 2007a). On arrival, males and females were caged separately and allowed to acclimatize and recover from transport stress for 1 week before being housed together overnight. The presence of vaginal plugs the following morning indicated mating (designated gestational day 1, GD1). Pregnancy was confirmed approximately 6 days later by palpation. Timed-pregnant rats were housed 5 per cage until GD 15, when they were individually housed in solid-bottomed cages in preparation for the day of birth, designated postnatal day one (P1). Offspring were weaned at P21, at which time males and females were housed separately in standard cages and allowed to mature to young adulthood (3-4 months) undisturbed, apart from standard husbandry, before commencement of the experimental procedures. All animals were maintained under a 12/12 hour reversed light/dark cycle (lights off at 7 am) and controlled temperature (21-230C) and humidity (63%) with standard rat chow and drinking water available ad libitum.

Antenatal glucocorticoid treatment (AGT)

The synthetic GC, dexamethasone (dexamethasone sodium phosphate, Fauliding Pharmaceuticals Plc., Royal Leamington Spa, UK), was added to the drinking water of pregnant dams at a final concentration of 0.5µg/ml onGD 16 - 19; control dams continued to receive normal drinking water. This non-invasive, stress-free regimen delivers a dose of dexamethasone to the mother (~ 0.075 mg/kg/day, based on an approximate water intake of 50 ml /day (McArthur et al., 2005)), which is approximately 3-fold lower than that used in perinatal medicine (Ballard et al., 1995; Jobe and Soll, 2004), 4-fold less than that calculated to mimic the physiological levels of GCs required to mature the fetal rat lung at GD 18-21 (Samtani et al., 2006a, b), insufficient to alter body weight at term (Kreider et al., 2006) and has no significant effects on the outcome of the pregnancies or maternal behavior(McArthur et al., 2006; McArthur et al., 2007a). Thus, offspring were reared by their biological mothers in order to avoid potential effects of cross-fostering, which canalter both maternal behaviors (Maccari et al., 1995) and DA transmission in adult offspring (Kane et al., 2004). Potential effects of litter-of-origin were minimized by selecting no more than 2 offspring per litter to contribute to the male and female AGT (dexamethasone exposure) and control (normal drinking water) groups.Female subjects were not selected for specific stages of the estrous cycle, which involves stressful processes of handling and probing to sample vaginal cytology; daily handling of males would not mimic this invasive procedure and therefore has uncertain equivalence as a control (Greenspan et al., 2007). Moreover, such stressors could potentially interfere with the results of this study.

Immunohistochemistry

Immunohistochemical identification of tyrosine hydroxylase immunoreactivity (TH-IR) was used to identify DA cell bodies in the VTA and SNc and their respective nerve terminals in the ventral striatum (NAc, core and shell) and dorsal striatum (caudate-putamen, CPu). Adult offspring were decapitated at 12 weeks of age between 09:00 and 10:00 (McArthur et al., 2007b). Brains were rapidly removed and immersed in 4% formaldehyde in phosphate-buffered saline (PBS), (0.1M NaH2PO4.2H2O, 0.1M Na2HPO4.12H2O, 0.15M NaCl (all VWR International, UK), pH 7.4) for one week, cryoprotected by immersion in 20% sucrose in PBS for 48h, then frozen and stored at -80°C. Serial coronal cryosections (20µm) were colleceted and stored in an antifreeze solution (0.1M NaH2PO4.H2O, 0.05M Na2HPO4, 0.15mM NaCl, 50% v/v ethanediol (all VWR International), 1% w/v polyvinylpyrrolidone, 0.1% w/v NaN3 (both Sigma-Aldrich, UK)) at -20°C until required. Immunostaining was carried out on free-floating sections (McArthur et al., 2007a). Briefly, sections were permeabilized with 0.05% Triton X-100 in PBS; non-specific binding sites were saturated by incubation for 1h with 10% normal goat serum (NGS) in PBS, and sections were incubated at room temperature overnight under constant agitation with the rabbit polyclonal anti-tyrosine hydroxylase (TH) (Millipore Ltd., UK) diluted 1:3000 in 1% NGS, as the primary antibody for detection of the rate-limiting DA synthetic enzyme. Sections were washed in 1% NGS in PBS and TH immunoreactivity TH-(IR) was visualized by incubation with AF594-conjugated goat anti-rabbit IgG (Invitrogen Ltd., UK), diluted 1:500in 1% NGS, for 1h at room temperature. Sections were finally washed (3 x 10 min in PBS), mounted on glass microscope slides and allowed to air-dry before being placed under cover-slipsusing an aqueous mountant (Vectashield, Vector Laboratories, Peterborough, UK).

Estimates of DA neuron numbers were based on protocols for measuring the effects of sex steroid hormones on cell numbers in specific brain nuclei (Balthazart et al., 2008; McArthur et al., 2011; Yamamura et al., 2011). Images of the SNc and VTA were captured and digitized using a CoolSNAP-Procf camera (Roper Scientific, Marlow, UK) attached to a NikonEclipse E800 microscope (Media Cybernetics, UK), with an image analysis software package (Image ProPlus 4.5, Media Cybernetics, Finchampstead, UK). At low magnification(x35), which included the whole of the coronal plane of the SNc and VTA, regional boundaries were outlined in digitized images, as defined by the presence of TH-IR neuronal groups and neuroanatomical landmarks, which were the thalamus dorsally and the substantia nigra pars reticulata ventrally for the SNc, and the parabrachial pigmented nucleus dorsally and the interfascicular and interpeduncular nuclei medially for the VTA. The adjacent SNc and VTA are also distinguished by the third cranial nerve tract running between the two nuclei. As illustrated previously (McArthur et al., 2007a), this is most obvious at level B, defined as -5.1 to -5.4 mm relative to bregma according to the parcellation scheme of Carman et al (Carman et al., 1991)where the SNc/VTA are divided into four levels (A-D), each spanning 300 µm from -4.8 mm to -6.0 mm with respect to bregma. Therefore, the present study examined alternate sections within this region for estimation of TH-IR cell counts, which we have shown consistently to be increased in AGT-exposed animals relative to controls(McArthur et al., 2005; McArthur et al., 2007a). TH-IR cell counts and densities were estimated by overlaying a counting grid (50 µm x 50 µm) on the regions of interest and immunoreactive cells were maually tagged (x100) in systematically, randomly selected visual fields within the grid; the number of TH-IR cells per field, grid-square area and number of grid squares sampledwere used to calculate cell densities. The level B volumes of the SNc and VTA were calculated according to the method of Cavalieri, as detailed elsewhere (Brodski et al., 2003; McArthur et al., 2007a). Briefly, the average cross sectional area for both regions (defined by the presence of TH-IR neuronal groups, as above), measured in every second section, was multiplied by the total number of sections and the section thickness, with allowance for tissue shrinkage using an electronic microcator (McArthur et al., 2007a). The total number of level B SNc and VTA TH-IR neurons was estimated by relating their numerical cell density to the volume, and used to calculate the group average (n = 6 rats per group).These methods provide cell estimates which are in good agreement with others using sterological or non-sterological methods (Brodski et al., 2003; Dewing et al., 2006; McArthur et al., 2007a; Balthazart et al., 2008; Yamamura et al., 2011).

For analysis of TH-IR fiber density in the CPu(innervated primarily by SNc DA neurons) and NAc, (innervated primarily by VTA DA neurons) core and shell, imagesof a 1µm optical section were captured using a Leica SP5 confocal microscopefitted with a 63x oil immersion objective.Threshold masks were applied using ImageJ software, at a standardised intensity for all samples. Percentage area cover was then measured by systematic random sampling of matrix regions of the CPu and NAc, allowing for background correction, in at least 6 different sections per animal.The average fiber density for each region per animal was used to calculate the group average (n = 6 rats per group).

Autoradiography

The DA transporter (DAT), D1-type (D1) and D2-type (D2) receptors were analyzed in the CPu, NAc (core and shell) regions in order to assess the effects of AGT on pre- and post-synaptic markers of DA function. Male and female adult offspring of control and AGT-exposed dams (n=6 per group) were decapitated at 12 weeks of age between 9-10am. Brains were quickly removed (n=6 per group), frozen at -40 ºC, and stored at -80 ºC until required. Serial coronal sections (20µm) were cut using a cryostat (Bright Instruments Ltd) maintained at -22°C. Sections were collected between +1.7 mm and +2.2 mm relative to bregma for the NAc core and shell, whereas for the CPu serial sections were collected between -0.2 and +1.6mm relative to bregma.Levels ofDAT, D1 and D2 were quantified using binding of the highly specific ligands [125I]-RTI 121, [3H]-SCH23390 and [3H]-raclopride, respectively,following our protocol used previously for DAT autoradiography(McArthur et al., 2007b). Briefly, after a pre-incubation period (30 min) adjacent sections were incubatedfor 1h in 0.1M PBS(pH 7.4) containing either radioligand alone (total binding) or radioligand plus excess non-radioactive ligand (to assess non-specific binding). These were, respectively,[125I]-RTI 121 (15pM; 2200Ci/mmol; PerkinElmer, UK) and GBR 12935 (10µM; Tocris, UK) for DAT; [3H]-SCH23390 (200pM; 85Ci/mmol; PerkinElmer) and SCH23390 (20µM; Tocris, UK)for D1-type receptors; [3H]-raclopride (3nM; 82.8Ci/mmol; PerkinElmer) and sulpiride (300µM; Tocris, UK)for D2-type receptors. Binding buffer for D1-type receptor assays also included 1µM CP809101 (Tocris, UK) to block potential 5-HT1C and 5-HT2C receptor binding. Sections were then washed in 0.1M PBS at 4°C, allowed to air dry, and exposed for 2 weeks ([125I]) or 6 weeks ([3H]) at -80°C to Kodak Biomax MS photographic film (PerkinElmer, UK)in a lead-lined cassette, alongside microscale standards (ARC Inc., St Louis, USA) impregnated with [125I] (0.18–93nCi/mg) or [3H] (0.14–489.1nCi/mg); signals from [3H]-SCH23390 and [3H]-raclopride binding assays were amplified using phosphor intensifying screens (PerkinElmer, UK). Films were developed and digitized using an MCID Core system attached to a CoolSNAP Procf camera (Interfocus Imaging Ltd., Cambridge, UK). Analysis within the defined brain regions (Paxinos and Watson, 1998)was performed by systematic random sampling using NIH ImageJ, and densities were quantified as femtomoles/mg equivalent. Average values from six different sections per region per animal were used to calculate the group means (n=6 per group). Specific radioligand binding sites were determined by subtraction of non-specific binding from total binding.

In-vivo microdialysis

Extracellular levels of DAand its metabolites were assessed byin vivo microdialysis coupled with electrochemical detection using commercially-available microdialysis probes (MAB 6 15kDa cut-off PES,Microbiotech AB, Sweden). Probes were implanted stereotactically into the NAc core through a burr hole in the skull (anterior-posterior: 1.2 mm forward of bregma; lateral: ± 2 mm from the midline; ventral: -7.5 mm from the cortical surface) in rats anaesthetized with urethane (1.2 g/kg i.p.). The probes were perfused continuously with artificial CSF (in mM: 147 NaCl, 3 KCl, 1.3 CaCl2, 1 MgCl2, 0.2 NaH2PO4 and 1.3 Na2HPO4) for 3h at a flow rate of 1 L/min prior to the collection of samples. After this equilibration period, three 20min baseline dialysate samples were collected (20 L) into polypropylene vials. After three baseline collections, rats were injected with amphetamine (0.8 mg/kg i.p) and sample collection (20 L/20 min) continued for an additional 2 h. All samples were promptly frozen on dry ice before storage at -70oC pending analysis by HPLC coupled with electrochemical detection. On completion of sample collection, rats were sacrificed by cervical dislocation and their brains rapidly removed into formalin for subsequent verification of probe placement.DA was determined in dialysate samples by HPLC and electrochemical detection. Separation was obtained using a Hypersil analytical column (100 x 4.6 mm inner diameter, 3m; Phenomenex, UK) and a mobile phase consisting of 8.82 g/L trisodium citrate, 2.03 gm/L NaH2PO4, 500 mg/L Na-1-octane sulfonic acid, 22.5% methanol, 25 mg/L EDTA and 1 ml/L triethylamine, pH 2.7 adjusted using orthophosphoric acid. DA was detected by oxidation using a Coulochem II detector (ESA 5014) equipped with a guard cell (+300mV) and a dual electrode analytical cell (E1, -150 mV; E2 +150mV). Chromatographic data were acquired and processed using Gyncosoft V4.4 (Dionex, UK).

Behavioral analysis

Adult progeny of control and AGT dams were tested during the dark phase of the light-dark cycle using the following range of established behavioral paradigms that have been shown to depend on the integrity of midDA transmission.

Amphetamine-induced locomotor activity

Following habituation to the environment, each animal was subjected to a further five test sessions of 120 min in the photocell testing chambers.On the day of testing, animals were weighed and injected intraperitoneallywith either vehicle (0.9% saline) or d-amphetamine at varying doses (0.2, 0.4, 0.8 or 1.2 mg/kg free base weight in 0.9% saline), according to a Latin square design. Immediately after injection, animals were placedin the centre of the activity chambers and locomotor activity was assessed by recording the number of photocell interruptions during the 2 h period.

Cocaine self-administration

Under isoflurane anaesthesia,chronic in-dwelling catheters were inserted into the right jugular vein and mounted in the mid-scapular region, as detailed elsewhere (Caine et al., 1992). Following surgery, subjects were housed individually with food and water made freely available. Catheters were flushed daily with heparinized saline (0.15-0.2 ml) and after 7 days’ recovery, drug self-administration training commenced.Each operant chamber was equipped with an active and inactive lever. Pressing the active lever activated an infusion pump to deliver cocaine intravenously in a volume of 100 L 0.9% sterile saline over 3 seconds, and simultaneously activated a cue light located above the active lever that remained illuminated for 20 seconds during which time both levers were retracted (time out period).Responses on the inactive lever were recorded but had no consequences.Training sessions were carried out for 7 consecutive days at each dose on a fixed-ratio schedule of reinforcement with ascending doses of cocaine (MacFarlan Smith, Edinburgh; 0.25, 0.5 and 1 mg/kg/100 L), and lever presses were recorded. Sessions were terminated after a maximum of 50 cocaine infusions had been self-administered or a period of 2h, whichever came first.

Spontaneous open field locomotor activity

Spontaneous locomotor activity in test-naïve animals was monitored inactivity chambers (259 mm x 476 mm x 209 mm; Allentown Inc)fitted with infrared photocell beams that were enclosed in ventilated sound-proofboxes. Testing was conducted during the dark phase of the light-dark cycle and beam breaks were recorded in bins of 5 min over a 90 min period using a computer that was interfaced with the photocells and situated in an adjacent room.

Prepulse inhibition (PPI) of acoustic startle

The PPI test was conducted in adult rats (male and female control and AGT-exposed groups, n=8)that had been drug- and test-free since birth in order to investigate the effects of AGT and sex on baseline PPI values. Three Plexiglas startle chambers designed for rats (Kinder Scientific, UK) were used and each chamber was housed in a ventilated sound-attenuating (-35dB) cabinet. Acoustic stimuli were delivered through a speaker mounted on the ceiling of the cabinet. A piezoelectric accelerometer mounted below the Plexiglas frame detected and transduced motion from within the chamber. Acoustic stimulus intensities and response sensitivities were calibrated to be near-identical in each startle chamber using the manufacturer’s startle calibration system.For testing, rats were placed into the startle chamber in a manner that prevented rearing but allowed free movement. Rats were subjected to a 5 min habituation period of background broadband white noise of 65dB before being presented with a series of trials composed of mixture of three trial types (startle pulse-alone trials, prepulse-pulse trials and no-stimulus trials). Each session started with six pulse-alone trials, each involving exposure of the animal to a 40ms burst of 120dB noise. These pulse-alone trials served to habituate and stabilize the startle response but were not used in the subsequent analysis. These trials were followed by prepulse-pulse trials in which four levels of prepulse (+3, +6, +12 and +16 dB above background for 20ms) were presented 100ms (onset to onset) before the startle stimulus (120 dB, 40ms). Each of the prepulse-pulse trials were presented 5 times in a trial block that also included 10 ten presentations of the pulse-alone trials, all delivered in a pseudorandom sequence. The average value of the startle response obtained from the ten pulse-alone signals was used in the calculation of startle reactivity and its inhibition by a prepulse. The discrete trial types were separated by a variable inter-trial interval with a mean of 15 s (range8-22 s). On no-stimulus trials, the startle response was measured in the absence of an acoustic signal. No-stimulus trials were intermixed between other trial types but were not used in the calculation of inter-trial intervals. All PPI tests were performed in the dark phase of the light-dark cycle and prepulse inhibition (PPI) was calculated as the % inhibition of the startle response obtained in the pulse-alone trials at each of the four prepulse intensities, for each animal by the expression: PPI = 100% x (1 – [mean reactivity in prepulse-pulse trials/mean reactivity in pulse-alone trials]).

Pavlovian appetitive learning(autoshaping)

In the autoshaping experimental paradigm, learning results from the association of a conditioned stimulus (CS+) that predicts food delivery and is measured relative to a second identical stimulus (CS-) that is explicitly unpaired with food reward. With repeated exposures animals acquire discriminated approach behavior to the CS+. Animals were trained using automated chambers (Med Associates, USA; 29.5 x 32.5 x 23.5 cm) and the ‘lever autoshaping’ software client (see as detailed elsewhere. Briefly, rats that had been placed on a food-restricted regime sufficient to maintain body weight at no less than 85% of free feeding weight were first habituated to the testing chambers for two consecutive days for 20 min each day with sugar pellets (45 mg; Noyes dustless pellets; Sandown Scientific, Hampton, UK) being placed in the central food magazine and the house light illuminated. Autoshaping commenced on day three. In each chamber, one lever was designated as the CS+ and the other as the CS-, with levers counterbalanced across animals and across experimental groups. For each trial, the CS+ or CS- lever was inserted into the chamber for a period of 15 sec and then retracted. A sugar pellet was delivered in the central food magazine following the withdrawal of only the CS+ lever. The CS+ and CS- trials were separated by an inter-trial interval that ranged between 45 and 75 s (average of 60 s) and the order of presentation in each chamber was random, with no more than two trials of either type occurring in sequence. Each autoshaping session consisted of 50 trials (25 CS+, 25 CS-) and sessions were conducted for 7 consecutive days. For each trial only the first contact with the CS, and not subsequent responses, was recorded as a lever press response, and contact with either lever was interpreted as approach behaviour. Data were analyzed in blocks of 50 trials as mean CS+ and CS- approach scores. The mean difference between CS+ and CS- scores was taken as a measure of discriminative learning.