Autoradiographic studies with 125 I-[Tyr 10 ]N/OFQ(1-11) and

Autoradiographic studies, using ¹²⁵I-[Tyr¹⁰]N/OFQ(1-11) and ¹²⁵I-[Tyr¹⁴]N/OFQ as radioligands, showed that their distributions in the brain of either mice (Mathis et al., 1999) or rats (Letchworth et al., 2000) are different. Binding parameters of

125I-[Tyr¹⁰]N/OFQ(1-11) revealed an affinity (KD) of 235 pM, which is over 100-fold

lower than its Ki in inhibition of against ¹²⁵I-[Tyr¹⁴]N/OFQ binding in mouse brain. The binding site of ¹²⁵I-[Tyr¹⁰]N/OFQ(1-11) was, therefore, proposed to be the high affinity site for ¹²⁵I-[Tyr¹⁴]N/OFQ, based from their similar maximal binding densities (Bmax) (Table 5). The binding density of ¹²⁵I-[Tyr¹⁰]N/OFQ(1-11) is one sixth of that of

125I-[Tyr¹⁴]N/OFQ (1.3+0.2 and 7.8+0.3 fmol/mg, respectively) in the rat midbrain PAG

(Letchworth et al., 2000), suggesting that [Tyr¹⁰]N/OFQ(1-11) might affect a portion of N/OFQ-sensitive NOP receptors. Recently, Dr. Pasternak’s group showed that, in both wild type and NOP-knockout mice, N/OFQ displayed similar Ki values in competing the binding of ¹²⁵I-[Tyr¹⁰]N/OFQ(1-11) at the high affinity site (Majumdar et al., 2009).

This suggests that the high affinity binding site of N/OFQ still exists in NOP-knockout mice.

6.4. Functional heterogeneity of NOP receptors revealed from the results of Ro

64-6198

As reported previously, we have proven that Ro 64-6198 activates only a subset of the N/OFQ-sensitive NOP receptors in rat PAG (Chiou et al., 2004). Several studies also indicate that Ro 64-6198 mimicked some, but not all, of the effects of N/OFQ (Table 7).

N/OFQ (i.c.v.) produced anxiolytic effect in both rats and mice but Ro 64-6198 (i.p.) was less effective in producing anxiolytic effect in mice at the doses unaffecting motor activity (Jenck et al., 2000). N/OFQ affects locomotor activity in a biphasic manner, increasing locomotion at lower doses but decreasing it at higher doses. Ro 64-6198 failed to fully reproduce the effect of N/OFQ, instead, it inhibited locomotor activity at all doses (Kuzmin et al., 2004). Ro 64-6198 mimicked the effects of N/OFQ to inhibit

contraction of mouse vas deferens, but the effects of Ro 64-6198 were not blocked by neither NOP nor MOP receptor antagonists (Rizzi et al., 2001). Ro 64-6198 regulated urination, but not hypotension or bradycardia, centrally as did N/OFQ (Dayan et al., 2001). In a neuropathic pain model, Ro 64-6198 was less effective in reducing allodynic responses than N/OFQ if given by i.t. administration, but was equopotent as N/OFQ if given by intraplantar injection.(Obara et al., 2005). In addition, Ro 64-6198 attenuated the expression of morphine sensitization at higher doses via a mechanism not blocked by a NOP receptor antagonist (Kotlinska et al., 2005). Although Ro 64-6198 increased food intake as N/OFQ, it did not decrease alcohol consumption as N/OFQ (Economidou et al., 2006a). However, it reduced alcohol self-administrtion and prevented relapse-like alcohol drinking (Kuzmin et al., 2007). Taken together, these effects of Ro 64-6198 suggest that there are functional heterogeneity of NOP receptors.

Aims of Study Hypotheses:

1. NOP receptors are functionally heterogeneous in vlPAG neurons. This was

proven by our previous study using Ro 64-6198 and can also be revealed by

(+)-5a Compound, [Tyr¹⁰]N/OFQ(1-11) or new NOP receptor antagonists.

2. [Tyr¹⁰]N/OFQ(1-11) is an agonist for the high, but not low, affinity binding site

of N/OFQ in rodent brain. The Ro 64-6198-sensitive NOP receptors might be

the NOP receptors exhibiting high binding affinity for N/OFQ.

Aim 1: To investigate if (+)-5a Compound can also affect a portion of NOP

receptors in vlPAG neurons.

(+)-5a Compound has a structure backbone analogous to Ro 64-6198 and Ro 64-6198 acts differently from N/OFQ at NOP receptors in rat PAG slices. If our hypothesis 1 is true, (+)-5a Compound should, like Ro 64-6198, also affect a portion of NOP receptors in vlPAG neurons.

Aim 2: To investigate if (+)-5a Compound affects the same subset of NOP

receptors that Ro 64-6198 effects.

If both (+)-5a Compound and Ro 64-6198 affect a portion of NOP receptor in

vlPAG neurons, if they affect the same subset of NOP recpeotrs will be examined.

Aim 3: To characterize the neurochemical and morphological properties of

the neurons sensitive to Ro 64-6198 and/or (+)-5a Compound.

If (+)-5a Compound affects the same subset of NOP receptors that Ro 64-6198 effects. The neurochemical and morphological properties of the neurons sensitive and insensitive to Ro 64-6198 and (+)-5a Compound will be examined.

Aim 4: To investigate if [Tyr¹⁰]N/OFQ(1-11), an agonist for the high affinity

binding site of N/OFQ in rodent brain, also affect a portion of NOP receptors in

vlPAG neurons.

Binding studies showed that ¹²⁵I[Tyr¹⁴]N/OFQ (low-affinity site) and

125I[Tyr¹⁰]N/OFQ(1-11) (high-affinity site) displayed distinct distribution sites in rodent

brains. The binding densities of high-affinity site is one sixth of that of low-affinity sites in the rat PAG, suggesting that [Tyr¹⁰]N/OFQ(1-11), which had higher selectivity for the high-affinity site than N/OFQ(1-11), might affect a portion of N/OFQ-sensitive NOP receptors. [Tyr¹⁰]N/OFQ(1-11) had higher selectivity for the specific binding site of

125I-[Tyr¹⁰]N/OFQ(1-11) than its parent peptide, N/OFQ(1-11) (Table 6). Therefore, we

used [Tyr¹⁰]N/OFQ(1-11) as a tool to investigate if it can reveal the heterogeneity of

NOP receptors in vlPAG neurons.

Aim 5: To investigate if [Tyr¹⁰]N/OFQ(1-11) affects the same subset of NOP

receptors that (+)-5a Compound does.

If the aim 4 is true, then the pharmacological files of [Tyr¹⁰]N/OFQ(1-11) will be compared with that of (+)-5a Compound to validate if they affect the same subset of NOP receptors.

Aim 6: To characterize the pharmacological properties of two novel NOP receptor

antagonists, Compound 24 and SB-612111, and examine if they can differentiate

the NOP receptors which are all sensitive to N/OFQ in vlPAG neurons.

We will further examine if the heterogeneity of NOP receptors can be revealed by NOP receptor antagonists, Compound 24 and SB-612111. In addition, their pharmacological properties in vlPAG neurons will be quantitatively characterized.

Materials and methods：

All experiments were conducted with Wistar rats of 9-18 days old and conformed to the guidelines of the Institutional Animal Care and Use Committee of the College of Medicine, National Taiwan University. All efforts were made to minimize the number of animals used.

1. Brain slice preparations

The midbrain blocks containing the PAG were rapidly dissected from postnatal Wistar rats. Coronal slices (300 or 400 μm) were then sectioned with a vibrotome (microslicer DTK-100, Dosaka) and equilibrated at room temperature in the artificial cerebral spinal fluid (aCSF). The aCSF consisted of (in mM) 117 NaCl, 4.5 KCl, 2.5 CaCl2, 1.2 MgCl2, 1.2 NaH2PO4, 25 NaHCO3 and 11.4 dextrose (pH=7.4), and was oxygenated with 95% O2/5% CO2. After equilibration for at least one hour, one slices was mounted on a submerged recording chamber and continuously perfused with aCSF at a rate of 2-3 ml/min.

2. Electrophysiological recordings

Blind patch-clamp whole cell recording was performed with 4-8 MΩ glass microelectrodes filled with the internal solution consisting of (in mM): 125 K⁺

gluconate, 5 KCl, 0.5 CaCl2, 5 BAPTA, 10 HEPES, 5 MgATP and 0.33 GTPtris (pH=7.3). To elucidate if (+)-5a Compound, like N/OFQ, also induced the NOP receptor-mediated GIRK currents, a hyperpolarization voltage ramp protocol was applied. After whole cell configuration was formed, the potential of the recorded neuron was held at -70 mV, stepped to -60 mV for 100 ms, ramped from -60 mV to -140 mV for 400 ms, and then stepped back to -70 mV (Fig. 3, inset). The membrane currents elicited by voltage ramps were acquired through an Axopatch 200B amplifier (Molecular Devices/Axon Instruments, Union City, CA) with a pClamp 7 software (Molecular Devices/Axon Instruments, Union City, CA) and simultaneously recorded with a chart recorder (Gould RS3200) at a low frequency response of 10 Hz to monitor the time course of drug effects. The access resistance (10-15 MΩ) was monitored during the recording period. Only those neurons with unchanged access resistance before and after drug treatments were accepted to ensure that the clamp efficiency was not deteriorated during the recording period.

3. Quantitative analysis of NOP receptor ligands

3.1. NOP receptor agonists

The effect of the tested NOP receptor agonist was quantified by the percent increment of the membrane current at -140 mV (I-140), taking its own I-140 before

treatment as 100%. An increment of I-140 greater than 5% and the induced-current having a reversal potential at around -90 mV (the equilibrium potential of K⁺ ions) was considered to be effective (defined as agonist-sensitive). For establishing the concentration-response curves of (+)-5a Compound or [Tyr¹⁰]N/OFQ(1-11), the percent increment of I-140 in each neuron was normalized to the maximal effect (Emax) produced by 1 μM N/OFQ, which was 39.4%+4% increment (n=26) (Chiou et al., 2002a). The EC50 values of (+)-5a Compound and [Tyr¹⁰]N/OFQ(1-11) were determined by the refection point of its concentration-response curve produced by logistic fitting:

E=Emax/[1+(D/EC50)ⁿ], where E represents the percentage of increment, Emax the maximal increment, D the concentration of agonist and n the Hill coefficient. To quantitatively evaluate the antagonistic effect of various receptor antagonists, their interactions with (+)-5a Compound or [Tyr¹⁰]N/OFQ(1-11) were examined in the same neuron. Given that not all neurons were sensitive to (+)-5a Compound or [Tyr¹⁰]N/OFQ(1-11) (see Results), it was not practical to pre-apply the intended antagonist. Therefore, the tested receptor antagonistwas applied to (+)-5a Compound or [Tyr¹⁰]N/OFQ(1-11)-sensitive neurons after the response of (+)-5a Compound or [Tyr¹⁰]N/OFQ(1-11) had reached a steady state, which usually took 20-25 min. The response of the tested antagonist was continuously monitored thereafter. In the study verifying whether NOP receptor-mediated GIRK channels were functional in those

(+)-5a Compound or [Tyr¹⁰]N/OFQ(1-11)-unresponsive neurons,N/OFQ was examined after (+)-5a Compound or [Tyr¹⁰]N/OFQ(1-11) had been applied for atleast 20 min. In this set of experiments, Ro 64-6198 was tested in some (+)-5a Compound-unresponsive neurons before N/OFQ was added in order to verify if the neurons unresponsive to (+)-5a Compound were also insensitive to Ro 64-6198. [Tyr¹⁰]N/OFQ(1-11) was tested in the (+)-5a Compound-treated neurons in order to verify the correlation with (+)-5a Compound.

3.2. NOP receptor antagonists

To quantitatively estimate the antagonistic effect of Compound 24 or SB-612111 against N/OFQ- or DAMGO-induced GIRK current, Compound 24 or SB-612111 was applied after the response to N/OFQ or DAMGO had reached a steady state. The GIRK current induced by N/OFQ or DAMGO was quantified as the percent increment of the membrane current recorded at -140 mV (I-140) in each neuron, taking its own I-140 before treatment as 100%. For establishing the concentration-response curve of Compound 24 or SB-612111 against N/OFQ- or DAMGO-induced GIRK current, the inhibitory effect of Compound 24 or SB-612111 in each neuron was calculated, taking the effect of N/OFQ or DAMGO in the same neuron as 100 %. The IC50 values of Compound 24 and SB-612111 were determined by the refection point of its concentration-response curve

in its inhibition of 0.1 μM N/OFQ-induced GIRK current produced by logistic fitting:

I=Imax/[1+(D/IC50)ⁿ], where I represents the percentage of inhibition, Imax the maximal inhibition, D the concentration of antagonist and n the Hill coefficient.

4. Immunofluorescence staining

For immunofluorescence studies, 0.2% Lucifer yellow (LY) was added in the internal solution. After recording, the slice containing the recorded neuron which had been filled with Lucifer yellow was fixed, re-sectioned and subjected to an immunofluorescent staining of glutamic acid decarboxylase-67 (GAD67), a synthesizing enzyme of GABA (Erlander et al., 1991). Briefly, after recording, the slices were fixed with 4% paraformaldehyde at 4 °C for one day, and then dehydrated in 30%

sucrose. Dehydrated slices were embedded and re-sectioned into 50 μm sections with a cryostat microtome (Leica CM3050S, Leica Microsystems, Nussloch, Germany). Slice sections were rinsed and washed with phosphate buffered saline (PBS) 3 times, followed by 0.3% Triton X-100 containing PBS (PBST) plus 0.5% bovine serum albumin (BSA) and then blocked in PBST containing 1% BSA and 10% normal goat serum (NGS) for 1 h. Then, slice sections were incubated with the mouse monoclonal antibody against GAD67 (diluted 1:1000) (Chemicon, Temecula, CA) in PBST containing 1% BSA overnight at 4 °C. Slice sections were then washed with PBST 3

times, followed by secondary antibody, Alexa 594 (diluted 1:100) (Invitrogen, Carlsbad, CA) for 1 hr. Fluorescent images were acquired with Zeiss LSM 510 Meta confocal microscope (Zeiss, Thornwood, NY) and edited with LSM 510 software (Zeiss, Thornwood, NY).

5. Morphometric analysis

To compare the morphology of (+)-5a Compound-sensitive and -insensitive neurons, the recorded neurons were reconstructed and analyzed. In brief, the recorded neuron was filled with biocytin (1%), which had been added in the internal solution during recording. After recording, slices were fixed with 4% paraformaldehyde at 4 °C for one day. After several washes with PBS, sections were treated with 1% H2O2 diluted in PBS for 3 hr, then incubated with 1% Triton-X 100 containing PBS plus 10% NGS and 2% BSA overnight at room temperature. Biocytin-filled neurons were then revealed using a avidin-biotin-peroxidase complex (ABC) kit (PK-4000, Vector Laboratories, Burlingham, CA), followed by a 3,3’-diaminobenzidine (DAB) kit (SK-4100, Vector Laboratories, Burlingham, CA). The DAB staining was monitored under a microscope (ECLIPSE TE 2000-U, Nikon, Japan) and stopped when appropriate by rinsing the sections with PBS. The sections were then mounted with Histokitt mounting medium (Assistent, Germany). The image of the identified neuron was captured under the

Olympus BX51 microscope (Olympus, Tokyo, Japan) equipped with a CX9000 CCD camera (MicroBrightField, Williston, VT), and reconstructed 3-dimensionally using Neurolucida (MicroBrightField, Williston, VT). Morphometric analysis of the recorded neurons, including the soma area and dendritic arborization, was performed by Neurolucida Explorer. The dendritic complexity was analyzed by the Sholl concentric ring analysis.

6. Chemicals

(+)-5a Compound and Ro 64-6198 were synthesized as reported previously (Jenck et al., 2000; Kolczewski et al., 2003). UFP-101 was a generous gift from Drs. Calo’ and Guerrini (University of Ferrara, Ferrara, Italy). Compound 24 was synthesized at University of Ferrara, Italy. We appreciate the generous gift of SB-612111 from Dr Toll (SRI international, Menlo Park, California, USA) (Khroyan et al., 2009). N/OFQ and AF-DX-116 were purchased from Tocris (Bristol, UK) and naloxone was from Sigma (St. Louis, MO). All drugs were dissolved in de-ionized water except (+)-5a Compound ,Ro 64-9198, Compound 24 and SB-612111 which were dissolved in dimethyl sulfoxide (DMSO). The final concentration of DMSO was kept below 0.1%, which did not affect the membrane currents elicited by voltage ramps (Chiou et al., 2004).

7. Statistics

Data are expressed as the mean+S.E.M. with the number of neurons tested.

Usually one neuron was recorded in each slice and 3-4 slices were taken from one rat.

Student t test, Student pair t test, one sample t test, one way ANOVA and chi-square test were used for statistical analysis and a p value < 0.05 was considered to be significant.

Results

1. Functional heterogeneity of NOP receptors reveal by (+)-5a Compound

1.1. (+)-5a Compound activated GIRK channels in vlPAG neurons.

(+)-5a Compound (10 μM) shifted the holding current (Ihold in Fig. 3A) outwardly and increased the membrane current elicited by a hyperpolarization ramp command from -60 to -140 mV in vlPAG neurons (Fig. 3A). Fig. 3B shows the current-voltage (I-V) curves of the membrane currents in a vlPAG neuron before and after treatment with 10 μM (+)-5a Compound, and after washout. The current induced by (+)-5a Compound, obtained by subtracting the current in the control from that in the presence of (+)-5a Compound, displayed inward rectification (Fig. 3C). The mean reversal potential of (+)-5a Compound-elicited current was -90.5+1.3 mV (n=49), resembling the equilibrium potential of K⁺ ions estimated by the Nernst equation. Therefore, the K⁺ channel in vlPAG neurons activated by (+)-5a Compound, as by N/OFQ (Chiou et al., 2002a), is the GIRK channel, an inwardly rectifying K⁺ channel that is coupled to G proteins (Ikeda et al., 1997).

1.2. (+)-5a Compound was less potent and less efficacious than N/OFQ.

The effect of (+)-5a Compound on GIRK channels was concentration-dependent (0.1-30 μM). To establish its concentration-response curve (Fig. 4A), the magnitude of

GIRK channel activation induced by (+)-5a Compound was quantified from the percent increment of I-140 obtained in each neuron as described in the Methods and then, after normalized to the maximal increment (39%+4% increment, n=26) produced by 1 μM N/OFQ (Chiou et al., 2002a), expressed as the percentage of the maximal effect of N/OFQ. (+)-5a Compound, at 10 μM, produced a maximal effect of 47%, as compared to that generated by N/OFQ (Fig. 4A). For a comparison, the concentration-response curves of N/OFQ and Ro 64-6198, taken from our previous studies (Chiou et al., 2004;

Chiou et al., 2002a), are also shown in Fig. 4A. In this study, 0.3 μM N/OFQ was also tested and reproduced a similar increment (40.2%+5.4%, n=32) of I-140 (filled square in Fig. 4A) as it did before (37.2%+5.3%, n=17) (Chiou et al., 2002a). The estimated EC50

of (+)-5a Compound was 605+2 nM, which is about 12 times higher than that of N/OFQ, 52+6.8 nM (Chiou et al., 2002a). The concentration-response curves of (+)-5a Compound and Ro 64-6198 are similar (Fig. 4A) with Emaxs of 46.9%+4.6% and 61.5%+5.6% (Chiou et al., 2004), respectively.

1.3. (+)-5a Compound activated GIRK channels in about half of the recorded

neurons.

(+)-5a Compound (0.1-30 μM) activated GIRK channels in 49 out of 92 (53%) recorded neurons and had no effect in the remaining 43 neurons. The responsiveness (>

5% increment of I-140 and the induced current having a reversal potential around -90 mV) of a neuron was independent of its location; either at the superficial or deeper layer of the slice. The numbers of the neurons sensitive to (+)-5a Compound or Ro 64-6198 over the total numbers of tested neurons under various tested concentrations are depicted in Fig. 4A. A scatter plot for the responses of all neurons tested with (+)-5a Compound showed that there was a sharp cutoff between (+)-5a Compound-sensitive and -insensitive neurons (Fig. 4B). Even at the highest tested concentration (30 μM), (+)-5a Compound was still ineffective in some neurons (Fig. 4), as observed before with Ro 64-6198 (Chiou et al., 2004).

1.4. The effect of (+)-5a Compound was antagonized by NOP, but not opioid or M2

muscarinic, receptor antagonists.

In (+)-5a Compound-sensitive neurons, the effect of (+)-5a Compound (0.1-30 μM) in 23 out of 35 (66%) neurons recorded was antagonized by UFP-101 (Fig. 3A), which competitively antagonized the effect of N/OFQ in the same preparation (Chiou et al., 2005). However, in the remaining 12 neurons, the effects of (+)-5a Compound were not reduced by UFP-101. The effect of (+)-5a Compound was not affected by naloxone, a non-selective μ-opioid receptor antagonist. The I^-140 increments after treatment with 10 μM (+)-5a Compound in the absence or presence of 1 μM naloxone were not

significantly different (119.8%+2.5% vs. 120.3%+2.4%, n=3).

Previous binding studies (Kolczewski et al., 2003) showed that (+)-5a Compound, at a higher concentration (10 μM), also had affinity at histamine H3, muscarinic and σ receptors, and sodium channels. It is unlikely that (+)-5a Compound activated GIRK channels through histamine H3 receptors since there are few histamine H3 receptors in the PAG (Pillot et al., 2002), and activation of σ receptors will block, not activate, K⁺ channels (Zhang and Cuevas, 2005). Among the five subtypes (M1-M5) of muscarinic receptors, the M2 subtype exists in the PAG and activation of M2 muscarinic receptors also results in GIRK channel activation (Sanada et al., 2007). Therefore, we further examined if AF-DX-116, a selective antagonist of M2 muscarinic receptors, would antagonize the effect of (+)-5a Compound. The result showed that AF-DX-116 did not alter the effect of (+)-5a Compound. After treatment with AF-DX-116 (3 μM), the I^-140 increased by 10 μM of (+)-5a Compound was 99.9%+2.5% (n=6) of that before treatment.

1.5. N/OFQ activated GIRK channels via NOP receptors in (+)-5a

Compound-insensitive vlPAG neurons.

In those (+)-5a Compound-insensitive neurons, we further tested the effect of N/OFQ to verify if their NOP receptors and GIRK channels were functional. In 41 out

of 43 neurons which were insensitive to the pretreatment of (+)-5a Compound, N/OFQ (0.3 μM) activated GIRK channels (Fig. 5).in a comparable magnitude as in naïve neurons (135.1% + 3.2%, n=41 vs. 137.2%+5.3%, n=17). The effects of N/OFQ in these neurons were antagonized by UFP-101 (Fig. 5). This result suggests that (+)-5a Compound is ineffective in a subset of NOP receptors of vlPAG neurons, at which N/OFQ displays similar efficacy as in all neurons, This characteristic of (+)-5a Compound is similar to that of Ro 64-6198 (Chiou et al., 2004). Therefore, we further investigated the interactions between (+)-5a Compound and Ro 64-6198 in the same neurons.

1.6. (+)-5a Compound-insensitive neurons were also insensitive to Ro 64-6198, and

vice versa.

In a neuron that was unaffected by (+)-5a Compound, Ro 64-6198 also failed to induce any membrane current change (Fig. 6A). The same result was found in other 8 tested neurons. The I-140 values after treatment with 10 μM (+)-5a Compound were 99.7%+0.5% (n=9) of the controls and were 98.2%+1.0% (n=9) after further treatment with 10 μM Ro 64-6198. In these neurons, N/OFQ activated GIRK channels in a manner antagonized by UFP-101 (Fig. 6A-C). Conversely, in 5 neurons which were unresponsive to Ro 64-6198, further addition of (+)-5a Compound also had no effect

(Fig. 6D). The I-140 values after treatment with 10 μM Ro 64-6198 were 100.6%+0.6%

(n=5) of the controls, and were 99.6%+0.6% (n=5) after further treatment with 10 μM (+)-5a Compound. In these neurons, N/OFQ did activate GIRK channels (Fig. 6D) in a manner blocked by UFP-101 (data not shown).

1.7. (+)-5a Compound precluded the effect of Ro 64-6198.

In (+)-5a Compound-sensitive neurons, which have been treated with the maximal effective concentration of (+)-5a Compound (10 μM), further addition of Ro 64-6198 (10 μM) did not cause any additional change in membrane currents elicited by voltage ramps (Fig. 7A) in all of the 11 neurons tested. The I-140 values after treatment with 10 μM (+)-5a Compound were 118.8%+3% of the controls (n=11), and were 116.8%+4%

(n=11) after further treatment with 10 μM of Ro 64-6198. This result suggests that (+)-5a Compound precluded the effect of Ro 64-6198.

1.8. N/OFQ further enhanced GIRK current in (+)-5a Compound-sensitive

neurons.

In contrast to Ro 64-6198, N/OFQ (0.3 μM) further enhanced GIRK current in the neurons which were responsive to (+)-5a Compound (Fig. 7B). After further treatment with 0.3 μM N/OFQ, the I^-140 values in the presence of the maximal effective

concentration (10 μM) of (+)-5a Compound were increased from 119.6%+3.6% to 132.6%+5.9% of the controls (n=8). The latter response is comparable to that produced by 0.3 μM N/OFQ alone (137.2%+5.3%, n=17). When N/OFQ was reduced to 50 nM, which is equi-effective to 10 μM of (+)-5a Compound, its increment on I^-140 in 10 μM (+)-5a Compound-pretreated neurons was also the same as in non-pretreated neurons (129.2%+2.8%, n=4 vs. 129.6%+3.1%, n=4). This suggests that even at the maximal effective concentration, (+)-5a Compound did not interact with equi-effective concentration of N/OFQ.

1.9. Most of the (+)-5a Compound-sensitive neurons were GABAergic.

To characterize the differences between (+)-5a Compound-sensitive and -insensitive neurons, their electrophysiological properties, morphometric features and neurotransmitter identity were compared. These results are presented on Table 8 and Figure 8.

As shown on Table 8, the mean resting membrane potential, input resistance and capacitance of (+)-5a Compound-sensitive (n=78) and -insensitive (n=70) neurons were not significantly different. Morphometric analysis data showed that (+)-5a Compound-sensitive (n=8 from 6 rats) and (+)-5a Compound-insensitive (n=6 from 5 rats) neurons had comparable soma size and total dendritic length. (+)-5a

Compound-sensitive neurons had more primary dendrites (Fig. 8A), resembling the triangular and multipolar cells described in the Golgi study (Beitz and Shepard, 1985).

Besides, compared with (+)-5a Compound-insensitive neurons (Fig. 8B), the dendrites of (+)-5a Compound-sensitive neurons had more branching nodes leading to greater dendritic orders and more terminal tips. Using Sholl analysis, these two types of neurons could also be differentiated from their dendritic complexity (Fig 8C). Neurons sensitive to (+)-5a Compound had more intersections in the Sholl concentric rings

Besides, compared with (+)-5a Compound-insensitive neurons (Fig. 8B), the dendrites of (+)-5a Compound-sensitive neurons had more branching nodes leading to greater dendritic orders and more terminal tips. Using Sholl analysis, these two types of neurons could also be differentiated from their dendritic complexity (Fig 8C). Neurons sensitive to (+)-5a Compound had more intersections in the Sholl concentric rings