Putative Kappa Opioid Heteromers As Targets for Developing Analgesics Free of Adverse Effects

Abstract

It is now generally recognized that upon activation by an agonist, β-arrestin associates with G protein-coupled receptors and acts as a scaffold in creating a diverse signaling network that could lead to adverse effects. As an approach to reducing side effects associated with κ opioid agonists, a series of β-naltrexamides 3–10 was synthesized in an effort to selectively target putative κ opioid heteromers without recruiting β-arrestin upon activation. The most potent derivative 3 (INTA) strongly activated KOR-DOR and KOR-MOR heteromers in HEK293 cells. In vivo studies revealed 3 to produce potent antinociception, which, when taken together with antagonism data, was consistent with the activation of both heteromers. 3 was devoid of tolerance, dependence, and showed no aversive effect in the conditioned place preference assay. As immunofluorescence studies indicated no recruitment of β-arrestin2 to membranes in coexpressed KOR-DOR cells, this study suggests that targeting of specific putative heteromers has the potential to identify leads for analgesics devoid of adverse effects.

Introduction

Although homomeric G protein-coupled receptors (GPCRs) have been reported to be functional, more recent studies have suggested that many members of the GPCRs may exist as oligomeric heteromers.¹⁻⁶ The existence of opioid heteromers has been demonstrated using a wide range of experimental techniques including coimmunoprecipitation, immunocytochemistry, bioluminescence, and Foerster resonance energy transfer (BRET and FRET, respectively).^6,7 It is now generally recognized that following G protein activation, the β-arrestin that associates with its target can act as scaffold for binding a host of intracellular mediators that initiate other signaling pathways, thereby leading to a diverse signaling network.⁸⁻¹⁵ As this network may be a source of some adverse effects associated with opioid analgesics, we have considered the possibility that receptor biasing could be mediated via activation of a specific heteromer whose interaction with β-arrestin is reduced or modified.

The present study describes an approach to developing analgesics that target putative κ opioid receptor (KOR)-containing heteromers without producing the adverse effects known to be associated with this receptor. On the basis of the structural requirements of ligands (1,¹⁶2¹⁷) that activate κ/μ opioid receptor (KOR-MOR) and κ/δ opioid receptor (KOR-DOR) heteromers, a focused library of compounds was evaluated for activating putative KOR heteromers without recruitment of β-arrestin-2. This led to the identification N-2′-Indolylnaltrexamine 3 (INTA), which produces potent antinociception in mice without aversion, tolerance, or dependence.

Chemistry

Given that 2(17) was reported to possess δ−κ agonist activity, the influence of an indole group on agonist selectivity was explored through a series of β-naltrexamine amides 3–10 that contain an indoyl moiety or closely related isosteric groups. These ligands were synthesized in a fashion similar to that described previously.¹⁸

Biological Results

Intracellular Ca²⁺ Release Studies in HEK293 Cells

Initial studies involved testing target compounds 3–10 for agonist activity using an intracellular calcium release assay in human embryonic kidney 293 (HEK293) cells as described previously.¹⁸ The stably tranfected HEK293 cell lines contained chimeric G protein (Δ6-G_αqi-myr)¹⁶⁻¹⁸ with a transiently transfected single opioid receptor (MOR, DOR, or KOR) or a pair of coexpressed opioid receptors (MOR-DOR, MOR-KOR, or DOR-KOR). Calcium release was measured on a FlexStation3 apparatus using a FLIPR Ca²⁺ kit (Molecular Devices). Concentration–response curves were established by measuring fluorescence for 90 s after the addition. Appropriate control studies, DAMGO for MOR, U69,593 for KOR, and DPDPE for DOR, expressing cells at 10 μM, confirmed the functional expression of the receptors in the assay (Supporting Information Figure S1). Concentration–response curves were plotted as a change in relative fluorescence units (ΔRFU). The data are displayed graphically (Figure 1, Supporting Information Table S1).

Figure 1.

Intracellular Ca²⁺ release profiles at multiple opioid receptors HEK293 cells. ^aData are mean ± SEM (n = 3–5). RFU, relative fluorescence unit.

Among the eight congeners, indole derivative 3 showed significantly greater stimulation of calcium release via activation of putative KOR-MOR and KOR-DOR heteromers with EC₅₀s in the 10^–12 M range. Although not superimposable, the curves were similar except at high concentration (≥10^–7 M). Interestingly, only weak activation of singly expressed δ, κ, or μ receptors was observed. It is noteworthy that methyl substitution of the indolic nitrogen afforded derivative 4 with little if any activation of singly or coexpressed receptors, suggesting the involvement of the hydrogen bonding as a factor contributing to the activity of 3.

Regioisomers 5 and 6 of 3 exhibited selectivity profiles that differed substantially from that of 3 in that KOR-DOR heteromer activation was reduced, especially for 5, and activation of MOR-KOR activation was reduced or lost. Conformational differences of the indoyl moiety due to the loss of the intramolecular hydrogen bonding between the carboxamide carbonyl group and indole NH may play a role in the lower efficacy of 5 and 6, as suggested by the finding that the N-Me derivative (4) of compound 3 possesses low potency. The 5′-chloro analogue 7 equally activated MOR-KOR and DOR-KOR expressing cells, but it was somewhat less efficacious than 3. The activity of the 5′-fluoro analogue 8 was different from that of 7 in that activation of KOR-DOR was apparently lower and an increase in activation at MOR-DOR was observed. Significantly, isosteric replacement of indole nitrogen with a sulfur or oxygen to afford benzothiophene 9 and benzofurane 10 afforded low activity without selectivity, again suggesting the importance of the indolic nitrogen for confering potency via hydrogen bonding to the carboxamide carbonyl group.

Antinociception and Tolerance Studies

Table 1 profiles the antinociception data of compounds 3–10 using the mouse tail-flick assay after intracerebroventricular (icv) or intrathecal (i.t.) administration¹⁹⁻²¹ (Table 1). Acute tolerance was measured only for compounds displaying a full agonist profile by comparing the ED_80–90 dose measured on day 1 to the same dose measured 24 h later on the same mice (Table 1).¹⁸

Table 1. Antinociception of 3–10 in Mice.

compounda	mode of administration	ED50 pmol/mouse (95% CI)	24 h toleranceb	ratio icv/i. t
INTA	i.t.	21.27 (13.90–32.54)	no	59
3	icv	1252.2 (948–1652)	no
4	i.t.	23.06 (16–33)	no	24
	icv	550.4 (325–933)	yes
5	i.t.	29.39 (21.91–39.43)	no	NAd
	icv	partial agonistc
6	i.t.	partial agonist		NA
	icv	partial agonist
7	i.t.	partial agonist		NA
	icv	partial agonist
8	i.t.	295.0 (140.04–620.0)	yes	NA
	icv	partial agonist
9	i.t.	partial agonist		NA
	icv	partial agonist
10	i.t.	191.9 (133–276)	no	3
	icv	580.8 (417–808)	no

^aPeak times for the dose response curves were as follows for i.t.: 3, 4, 8, and 10, all 10 min; 5, 6, 7, 9, all 20 min. For icv: 3, 4, and 10, 10 min.

^bAcute tolerance was calculated using the highest dose of the dose–response curve on day 1 and repeated on day 2. If there was no significant difference between the 2 days, the animals were said to be not tolerant.

^cPartial agonist is defined as when the maximum %MPE was ≤60%

^dNA: not applicable.

Compound 3 was found to be 59-fold more potent when given i.t. vs icv (ED₅₀ = 21.27 pmol per mouse i.t. versus 1252 pmol per mouse icv). These values were in the same range as 1.¹⁶ No acute tolerance was observed for 3 by either of these routes of administration. Compound 3 was also found to be highly potent when administered by the subcutaneous (sc) route with an ED₅₀ of 0.97 mg/kg (0.74–1.28), which is 9-fold more potent than morphine (7.8 mg/kg). When administered orally, 3 was 4-fold less potent [ED₅₀ of 9.08 mg/kg (7.51–10.80)] than morphine [2.51 mg/kg (1.8–3.4)].

The N-methyl analogue 4 had identical potency to 3 via the i.t. route, with an icv/i.t. ED₅₀ ratio approximately half that of 3. Acute tolerance was observed when 4 was administered icv. A possible reason for the discrepancy between in vivo and the cell-based data could be due to the greater lipophilicity of 4 or to targeting a different KOR heteromer. The 3′-regioisomer 5 displayed i.t. antinociception which was similar to that of 3 and was not accompanied by 24 h tolerance. In contrast to 3, it was found to be a partial agonist when administered i.t., thereby precluding use of the naloxone-precipitated jumping assay. As the 4′-regioisomer 6 exhibited mixed agonist–antagonist activity after i.t. or icv administration, no ED₅₀ values were obtained. The 5′-chloro analogue 7 possessed a similar mixed agonist–antagonist profile, while the 5′-fluoro derivative 8 was found to be a weak full agonist with tolerance by the i.t. route and a partial agonist when administered icv. Partial agonism by both routes of administration was also observed with benzothiaphene analogue 9. The benzofuran congener 10 behaved as a weak full agonist without tolerance by both routes of administration.

In Vivo Studies of 3 with Opioid Antagonists

Initial studies involved use of the nonselective antagonist, naloxone, via the i.t. route (Supporting Information Figure S2). Naloxone (500 nmol/mouse) produced a potent antagonism, as indicated by a 15-fold rightward shift, which is consistent with an opioid receptor-mediated mechanism. The pharmacologic selectivity of compound 3 was evaluated with selective antagonists norBNI^22,23 (κ), NTB²⁴ (δ), and β-FNA^25,26 (μ) used alone or in combination. Each antagonist when measured individually using the standard antagonist dose did not shift the dose–response curve of 3. To measure for synergism, a theoretical AD₅₀ was calculated based on the ratio of each antagonist to the other antagonist. The theoretical AD₅₀ was compared to the actual AD₅₀ measured. Synergism was considered significant if the 95% of the confidence intervals (CI) did not overlap the theoretical AD₅₀.^27,28 NTB (1) + norBNI (25) was 6-fold more potent in the presence of the other antagonist, while the combination norBNI (1) + β-FNA (3) was 9 times more potent when compared to being tested individually. (Supporting Information Figure S3 (parts A and B)). These data are consistent with the cell-based targeting by 3 to both KOR-DOR and KOR-MOR putative heteromers that were observed in the calcium mobilization studies.

Compound 3 Does Not Produce Tolerance or Physical Dependence

Chronic tolerance was evaluated using a modified methodology of Fairbanks et al.²⁹3 was administered i.t. (ED_80–90 dose) twice on day 1 and day 2. On the third day, the mice were injected with the same dose and retested to compare the ED_80–90 value. Tolerance is indicated by a significantly higher ED₅₀ value. On day 3, compound 3 did not induce significant tolerance, as its ED₅₀ was 29.47 pmol (16.97–51.23) compared to the control ED₅₀ [22.05 pmol (14.71–33.05)] (Supporting Information Figure S4).

Evaluation of physical dependence was conducted in mice that were injected sc with 3, 3 times daily for 4 days (2.5–10 mg/kg) and 10 mg/kg on day 5, followed 3 h later by a single 10 or 50 mg/kg dose of sc naloxone.³⁰ The degree of physical dependence was estimated by counting the number of jumps over a 10 min period. When compared to a morphine control (71 jumps), 3 (14 jumps) failed to reveal significant physical dependence with either 10 or 50 mg/kg of naloxone (Supporting Information Figure S5).

Place Conditioning Studies of 3

Standard κ opioid agonists are known to be aversive in the dose range for antinociception.³¹⁻³⁴ Because no drug-induced dependence of 3 was observed in the naloxone-induced jumping test, the effect of sc-administered 3 on place preference was measured in mice after four or eight days of conditioning. The most noteworthy feature in these experiments, was a robust dose-dependent place preference in the 0.3–10 mg/kg dose range (Figure 2A), which overlapped with the sc ED₅₀ dose for antinociception. A comparative study with the κ-selective ligand, salvinorin A,³⁵⁻³⁸ induced aversion at all doses tested (0.1, 0.3, and 1.0 mg/kg (Figure 2B). As naloxone treatment of mice blocked the place preference induced by 3 (Figure 2C), it suggests that the rewarding effect of derivative 3 was mediated via opioid receptors (see also Supporting Information Table S3). This dose of naloxone (1.0 mg/kg) produced no aversion on its own.

Figure 2.

(A) 3 (INTA) did not produce aversion in mice. A strong dose-dependent rewarding effect was observed in a wide range of doses (0.3–10 mg/kg, sc). (B) On another hand, salvinorin A (0.3–1.0 mg/kg, sc) in the same protocol produced a strong aversion. (C) 3 condition preference was reversed by naloxone. *p < 0.05, **p < 0.01, ***p < 0.001; Bonferroni t test comparison versus vehicle control (same test). Test 1 and test 2 were performed following four and eight days of conditioning, respectively.

Activation of Putative KOR-DOR Heteromers by 3 Does Not Lead to Recruitment of β-Arrestin2

Compound 3 possessed the most favorable in vivo profile from the standpoint of parenteral and oral antinociception, lack of tolerance and dependence, and absence of aversion in conditioned place aversion. As adverse effects have been associated with β-arrestin2 recruitment,³⁹⁻⁵⁰ we investigated how 3 would affect such recruitment upon activation in the presence of singly expressing or coexpressing opioid receptors in HEK293 cells. Similar experiments also were carried out with standard μ, δ, and κ agonist ligands (DAMGO,⁵¹ DPDPE,⁵² U69593,⁵³ respectively) as controls. Cells were treated with immunofluorescent β-arrestin2 primary antibodies (goat) and secondary antigoat antibodies to image the receptors. A brighter cell surface membrane was the criterion for positive β-arrestin2 recruitment. Cells with singly expressing receptors showed recruitment when exposed to 3 or the three standard opioid ligands (Figure 3). However, with KOR-DOR coexpressing cells, 3 did not recruit β-arrestin2 to cell membranes and MOR-KOR appeared to exhibit a reduced recruitment. Thus, the effect of 3 on β-arrestin2 clearly appears to be different for cells that coexpress KOR-DOR when compared to those expressing KOR-MOR, MOR, DOR, and KOR.

Figure 3.

Representative high power fluorescent micrographs of control HEK293 cells and those singly and doubly expressing opioid receptors that were stained for β-arrestin2. INTA (3)- and NNTA-treated MOR-KOR HEK293 cells showed some recruitment of β-arrestin2, while INTA-treated KOR-DOR HEK293 cells showed no recruitment of β-arrestin2. 6′-GNTI-treated KOR-DOR HEK 293 cells exhibited little recruitment. Treatment of nontransfected HEK293 cells with 3, NNTA (1), 6′-GNTI (2), or any of the standard opioid control ligands did not induce β-arrestin2 recruitment (not shown).

It has recently been reported that 2 inhibits recruitment of β-arrestin to KOR and possibly KOR-DOR in cultured cells.⁵⁴⁻⁵⁶ There are, however, notable selectivity differences as 3 inhibits recruitment only of KOR-DOR, leaving KOR unaffected. This difference can simply be explained by the involvement of different mechanisms for 3 and 2, particularly because the ligands differ architecturally.

Discussion

It is now generally recognized that following G protein activation, β-arrestin associates with the phosphorylated receptor and can function as a scaffold for the binding of a host of intracellular mediators that initiate other signaling pathways, thereby leading to a diverse signaling network.⁸⁻¹⁵ As this network has been considered to be a source of some adverse effects, it has been proposed that a viable approach to the development of drugs with fewer side effects can be accomplished by reducing, modifying, or preventing interaction of β-arrestin with the GPCR.^{39,40,42,43,45,46,56−58}

There is burgeoning evidence for the possible involvement of opioid receptor heteromers in promoting desired and/or adverse pharmacologic effects.^16,59 For examples, early studies have suggested that interaction between μ and δ opioid receptors are obligatory for the development of tolerance and physical dependence.^59a−59c More recent studies have provided additional support for this concept with bivalent ligands that target putative MOR-DOR heteromers.^59d−59h Given evidence involving G protein switching from G_i/0 (MOR) to G_z (MOR-DOR) upon activation,⁶⁰ this could account for lack of side effects in some cases. Very recent studies with a μ agonist/mGluR5 antagonist bivalent ligand⁶¹ and with 1,¹⁶ a ligand that selectively targets MOR-KOR heteromers are consistent with this concept. Thus, it occurred to us that the constitution of different opioid receptor heteromers could also affect recruitment of β-arrestin.⁶²⁻⁶⁶ On the basis of this concept, we have synthesized and screened ligands that selectively activate κ opioid receptors coexpressed with MOR and DOR in an effort to reduce or eliminate the well-known aversive effects associated with κ agonists. If some opioid receptor heteromers do not readily recruit β-arrestin upon activation, perhaps their adverse effects might be reduced or eliminated. In this regard, 1, which selectively activates putative MOR-KOR heteromers, produces potent antinociception without tolerance or dependence. However, not all adverse effects were eliminated, as some aversion was noted at 10 times its ED₅₀ dose.¹⁶ The study of 1 raised the possibility that its adverse effects might be due to a switch in G protein that differs from that involved in the activation of MOR-DOR, which in turn could reduce the recruitment of β-arrestin.

With the above concept in mind, we have synthesized and tested a focused library based on the structural requirements of the μ–κ agonist, 1,¹⁶ and 2,¹⁷ which was reported to activate KOR-DOR heteromer. After initial screening in HEK-293 cells singly expressing MOR-, KOR-, and DOR- or coexpressing MOR-DOR, KOR-MOR-, and KOR-DOR using the calcium mobilization assay, we have found that the β-naltrexamine derivative 3 is highly potent in activating both KOR-MOR and KOR-DOR heteromers in the pM range (Figure 1). In this regard, it is noteworthy that small structural changes of 3 produced substantial changes in potency or selectivity.

In mice, the i.t. potency of 3 was equivalent to that of morphine. However, by the icv route, 3 was ∼60-fold less potent. Significantly, compound 3 produced neither tolerance nor dependence. Other members of the series were either less potent and/or produced tolerance. In view of the favorable in vivo pharmacological profile of 3, i.t. antagonism studies were performed using standard opioid antagonists. As the ED₅₀ of 3 was strongly right-shifted by a single dose of naloxone, it appears that its action is mediated via opioid receptors. In an effort to evaluate the in vivo selectivity of 3 at opioid receptors, we investigated the effect of κ (norBNI), δ (NTB), and μ (β-FNA) antagonists on antinociception when administered separately or in combination. The observation that norBNI/NTB or norBNI/β-FNA combinations exhibited synergism is consistent with interaction of 3 with KOR-DOR and KOR-MOR heteromers. These data support the selectivity profile of 3 shown in the calcium mobilization assay (Figure 1).

Immunofluorescent imaging of β-arrestin2 recruitment in HEK293 cells singly or coexpressing opioid receptors revealed that receptor activation by 3 led to recruitment of β-arrestin in all these cell lines except KOR-DOR (Figure 3). The observation that the KOR-DOR/KOR-MOR agonist (3) induced recruitment of β-arrestin in KOR-MOR cells but did not produce aversion in mice (Figure 2A) was intriguing because the κ–μ selective agonist, 1, has been reported to be aversive at 10× its ED₅₀ dose.¹⁶

The fact that 3 exhibited conditioned place preference that was reversed by naloxone suggested that the reversal was due to interaction of naloxone with a putative opioid receptor heteromer. However, in light of the absence of naloxone-induced jumping, the place preference does not appear to be coincidental with dependence. Becaise compound 3 targets both KOR-MOR and KOR-DOR in HEK293 cells, some aversion would be expected in the conditioned place preference assay if these putative heteromers behave independently. This raises the possibility that perhaps the KOR-MOR and KOR-DOR that are targeted by 3 do not behave independently but rather function as an higher-order oligomer, as there are reports for the existence of tetramers among class A GPCRs.^2,67−69 Such a higher-order organization might lead to a pharmacologic profile that would differ from individual heterodimers that are not intimately associated. In any case, the present study suggests that targeting of KOR-DOR heteromer may be a viable approach to the design of analgesics devoid of the side effects associated with the presently employed analgesics. These data also reinvigorate the interest for the development of κ or κ-associated agonist therapeutics that are not peripherically restricted.

These and prior studies suggest that targeting putative κ-opioid receptor heteromers is a viable approach to the development of analgesics that are free of side effects through modulation or inhibition of β-arrestin2-dependent signaling pathways that may promote adverse effects. Also, the combinatorial possibilities that exists for KOR-containing heteromers affords more targets than homomers. In this regard, nearly two dozen heteromers containing opioid protomers or opioid and nonopioid protomers have been reported in cultured cells.^{1−6,61,70−77} From this perspective, the targeting of opioid heteromers has the potential of identifying new leads for development of ligands devoid of adverse effects for managemen of pain.

Experimental Section

Chemistry

For material and method, see ref (18).

Analytical data confirmed the purity of the products was ≥95%.

General Protocol for Amide Bond Formation

As reported prevously,¹⁸ 6-β-naltrexamine (1 equiv) was dissolved in DCM and benzotriazole-1-yl-oxy-tris(dimethylamino)-phosphonium hexafluorophosphate (BOP) (2 equiv), and the appropriate aryl carboxylic acid (2.4 equiv) was added. N,N-Diisopropylethylamine (DIPEA) (2.8 equiv) was then added to the mixture. The solution was stirred for 2–16 h at room temperature and concentrated to dryness. The resulting solid was dissolved in anhydrous methanol (3 mL) and K₂CO₃ (7.3 equiv) was added. The mixture was stirred at room temperature for 1 h, concentrated, and purified by SiO₂ chromatography. The targeted compounds were then converted into hydrochloride salt for pharmacological evaluation.

17-Cyclopropylmethyl-3,14β-dihydroxy-4,5α-epoxy-6β-[(2′-indolyl)acetamido]morphinan (3, INTA)

Purification by flash chromatography [DCM/MeOH 99/1] and then precipitation from an acetone/hexanes mixture provided 3 as a white solid. Yield 78%. ¹H NMR (DMSO-d₆) δ: 0.11 (m, 2H), 0.46 (m, 2H), 0.84 (m, 1H), 1.25–1.60 (m, 4H), 1.82–1.99 (m, 2H), 2.17 (m, 1H), 2.33 (m, 2H), 2.58 (m, 2H), 3.02 (m, 2H), 3.69 (m, 1H), 4.68 (d, 1H, J = 7.8 Hz), 4.90 (bs, 1H, OH-14), 6.52–6.57 (m, 2H, H₁, H₂), 7.01 (m, 1H), 7.15 (m, 2H), 7.40 (d, 1H, J = 7.6 Hz, H), 7.61 (d, 1H, J = 8.0 Hz), 8.63 (d, 1H amide, J_NH = 8.3 Hz), 9.03 (bs, 1H, OH-3), 11.51 (s, 1H, NH indole). ¹³C NMR (DMSO-d₆) δ: 3.52, 3.64, 9.20, 22.13, 24.80, 30.04, 30.26, 43.65, 47.01, 51.10, 58.36, 61.72, 69.56, 90.65, 102.20, 112.22, 116.98, 118.38, 119.59, 121.46, 123.15, 123.45, 127.04, 131.32, 131.75, 136.33, 140.43, 142.03, 160.63; mp 232–234 °C. Anal. Calcd for C₂₉H₃₁N₃O₄: C, 71.73; H, 6.43; N, 8.65. Found: C, 71.65; H, 6.39; N, 8.32. ESI-TOF MS calculated for C₃₁H₃₂N₂O₄, m/z 485.574; found, 486.123 (MH)⁺.

17-Cyclopropylmethyl-3,14β-dihydroxy-4,5α-epoxy-6β-[1′-N-methyl-(2′-indolyl)acetamido]morphinan (4)

Purification by flash chromatography [hexanes/AcOEt 25/75] and then precipitation from an acetone/hexanes mixture provided 4 as a white solid. Yield 91%. ¹H NMR (DMSO-d₆) δ: 0.01 (m, 2H), 0.40 (m, 2H), 0.71 (m, 1H), 1.11 (m, 1H), 1.36 (m, 2H), 1.51 (m, 2H), 1.76 (m, 1H), 2.05 (m, 2H), 2.22 (m, 2H), 2.88 (d, 1H, J = 8.4 Hz), 2.98 (d, 1H, J = 5.7 Hz), 3.92 (s, 3H), 4.01 (m, 1H), 4.42 (d, 1H, J = 7.9 Hz), 6.39 (d, 1H, J = 8.2 Hz), 6.63 (d, 1H, J = 8.1 Hz), 6.81 (s, 1H), 7.01 (t, 1H, J = 7.1 Hz), 7.13–7.17 (m, 2H), 7.22 (d, 1H, J = 8.3 Hz), 7.31 (d, 1H amide, J_NH = 8.9 Hz), 7.48 (d, 1H, J = 7.9 Hz). ¹³C NMR (DMSO-d₆) δ: 3.78, 3.99, 9.40, 22.61, 22.33, 29.08, 31.59, 36.81, 43.93, 47.26, 50.02, 59.36, 62.33, 70.15, 92.53, 104.12, 110.06, 117.70, 118.97, 120.30, 121.82, 123.89, 124.24, 126.08, 130.68, 132.15, 139.05, 139.92, 143.28, 162.35; mp 196–197 °C. Anal. Calcd for C₃₀H₃₃N₃O₄: C, 72.13; H, 6.66; N, 8.41. Found: C, 72.23; H, 6.64; N, 8.39. ESI-TOF MS calculated for C₃₀H₃₃N₃O₄, m/z 499.601; found m/z, 500.123 (MH⁺).

17-Cyclopropylmethyl-3,14β-dihydroxy-4,5α-epoxy-6β-[(3′-indolyl)acetamido]morphinan (5)

Purification by flash chromatography [DCM/MeOH 99/1] and then precipitation from an acetone/hexanes mixture provided 5 as a white solid. Yield 82%. ¹H NMR (CD₃OD) δ: 0.22 (m, 2H), 0.57 (m, 2H), 0.91 (m, 1H), 1.47–1.73 (m, 4H), 1.98 (m, 1H), 2.28 (m, 2H), 2.74 (m, 2H), 3.12 (d, 1H, J = 8.1 Hz), 3.16 (d, 1H, J = 5.8 Hz), 3.93 (m, 1H), 4.64 (d, 1H, J = 7.8 Hz), 6.60 (d, 1H, J = 8.2 Hz), 6.64 (d, 1H, J = 8.2 Hz), 7.11–7.40 (m, 2H), 7.42 (d, 1H, J = 7.6 Hz), 7.91 (s, 1H), 7.61 (dd, 1H, J = 0.9 Hz, J = 6.9 Hz). ¹³C NMR (CD₃OD) δ: 3.99, 4.12, 14.44, 23.74, 25.76, 30.49, 31.29, 35.76, 49.45, 52.43, 58.75, 60.92, 63.89, 71.73, 93.41, 111.78, 112.73, 119.00, 120.21, 121.89, 121.96, 123.48, 125.44, 127.26, 128.65, 128.94, 138.08, 142.43, 143.87, 168.34; mp 170–174 °C. Anal. Calcd for C₂₉H₃₀N₃O₄: C, 71.73; H, 6.43; N, 8.65. Found: C, 71.65; H, 6.39; N, 8.32. ESI-TOF MS calculated for C₃₁H₃₂N₂O₄, m/z 485.574; found 486.124, (MH)⁺.

17-Cyclopropylmethyl-3,14β-dihydroxy-4,5α-epoxy-6β-[(5′-indolyl)acetamido]morphinan (6)

Purification by flash chromatography [DCM/MeOH 99/1] and then precipitation from an acetone/hexanes mixture provided 6 as a white solid. Yield 75%. ¹H NMR (CD₃OD) δ: 0.19 (m, 2H), 0.57 (m, 2H), 0.89 (m, 1H), 1.47–1.73 (m, 2H), 1.82–1.99 (m, 2H), 1.96 (m, 1H), 2.29 (m, 2H), 2.46 (m, 2H), 2.73 (m, 2H), 3.12 (d, 1H, J = 8.3 Hz), 3.22 (d, 1H, J = 5.9 Hz), 3.97 (m, 1H), 4.68 (d, 1H, J = 7.8 Hz), 6.55 (dd, 1H, J = 0.7 Hz, J = 3.1 Hz), 6.60 (d, 1H, J = 8.2 Hz), 6.65 (d, 1H, J = 8.1 Hz), 7.31 (d, 1H, J = 3.2 Hz), 7.42 (d, 1H, J = 8.5 Hz), 7.63 (dd, 1H, J = 1.7 Hz, J = 8.6 Hz), 8.14 (d, 1H, J = 1.2 Hz). ¹³C NMR (CDCl₃) δ: 4.12, 4.19, 12.31, 23.62, 25.65, 31.31, 34.29, 48.34, 49.61, 53.21, 60.15, 63.82, 71.85, 93.26, 103.56, 111.95, 118.75, 120.07, 120.42, 121.44, 121.63 (2×), 122.57, 126.41, 127.21, 129.02, 135.61, 139.56, 143.82, 171.56; mp 170–172 °C. Anal. Calcd for C₂₉H₃₁N₃O₄: C, 71.73; H, 6.43; N, 8.65. Found: C, 72.65; H, 6.19; N, 8.12. ESI-TOF MS calculated for C₃₁H₃₂N₂O₄, m/z 485.574; found, 486.157 (MH)⁺.

17-Cyclopropylmethyl-3,14β-dihydroxy-4,5α-epoxy-6β-[(5′-chloro-2′-indolyl)acetamido]morphinan (7)

Purification by flash chromatography [Hexanes/AcOEt 15/85] and then precipitation from an acetone/hexanes mixture provided 7 as a white solid. Yield 75%. ¹H NMR (CDCl₃) δ: 0.14 (m, 2H), 0.55 (m, 2H), 0.86 (m, 1H), 1.08 (m, 2H), 1.36 (m, 2H), 1.82–1.99 (m, 2H), 2.17 (m, 1H),1.94 (m, 2H), 2.39 (m, 2H), 3.02 (d, 1H, J = 8.3 Hz), 3.14 (d, 1H, J = 5.9 Hz), 3.47 (m, 1H), 4.55 (d, 1H, J = 7.8 Hz), 6.44 (s, 1H), 6.58 (d, 1H, J = 8.2 Hz), 6.74 (d, 1H, J = 8.1 Hz), 6.97 (d, 1H amide, J = 9.3 Hz), 7.20 (dd, 1H, J = 2.0 Hz, J = 8.7 Hz), 7.27 (d, 1H, J = 8.8 Hz), 7.49 (s, 1H), 9.92 (d, 1H, NH indole). ¹³C NMR (CDCl₃) δ: 3.80, 4.01, 9.41, 24.93, 25.61, 33.94, 36.49, 47.46, 49.20, 50.27, 59.30, 62.24, 70.18, 93.21, 101.62, 112.87, 118.11, 119.61, 121.13, 124.53, 124.89, 126.08, 128.58, 130.63, 131.83, 134.53, 139.81, 142.61, 160.75; mp 178–180 °C. Anal. Calcd for C₂₉H₃₀ClN₃O₄: C, 66.98; H, 5.81; N, 8.08. Found: C, 66.72; H, 5.84; N, 7.89. ESI-TOF MS calculated for C₂₉H₃₀ClN₃O₄, m/z 518.182; found, 520.096 (MH)⁺.

17-Cyclopropylmethyl-3,14β-dihydroxy-4,5α-epoxy-6β-[(5′-fluoro-2′-indolyl)acetamido]morphinan (8)

Purification by flash chromatography [hexanes/AcOEt 25/75] and then precipitation from an acetone/hexanes mixture provided 8 as a white solid. Yield 89%. ¹H NMR (CD₃OD) δ: 0.17 (m, 2H), 0.55 (m, 2H), 0.89 (m, 1H), 1.54–1.72 (m, 4H), 1.99 (m, 1H), 2.17–2.30 (m, 2H), 2.42 (m, 2H), 2.69 (m, 2H), 3.07 (d, 1H, J = 8.2 Hz), 3.14 (d, 1H, J = 5.9 Hz), 3.93 (m, 1H), 4.64 (d, 1H, J = 7.8 Hz), 6.23 (d, 1H, J = 8.2 Hz), 6.28 (d, 1H, J = 8.1 Hz), 6.63 (td, 1H, J = 2.5 Hz, J = 9.2 Hz), 6.73 (s, 1H), 6.92 (dd, 1H, J = 2.5 Hz, J = 9.6 Hz), 7.05 (m, 1H). ¹³C NMR (CDCl₃) δ: 4.17, 4.51, 10.21, 23.56, 25.76, 31.33, 31.88, 45.37, 49.21, 52.94, 60.23, 63.75, 71.79, 93.12, 104.21, 106.56, 106.80, 114.09, 114.19, 118.66, 120.09, 125.43, 129.19, 132.53, 134.07, 134.92, 141.84, 143.72, 142.61, 163.54. ¹⁹F NMR (CD₃OD) δ −125.99; mp 178–180 °C. Anal. Calcd for C₂₉FH₃₀N₃O₄: C, 69.17; H, 6.00; N; 8.34. Found C, 69.59; H, 5.97; N, 8.45. ESI-TOF MS calculated for C₂₉FH₃₀N₃O₄, m/z 503.565; found, 504.036 (MH⁺)

17-Cyclopropylmethyl-3,14β-dihydroxy-4,5α-epoxy-6β-[(2′-benzothiophene)acetamido]morphinan (9)

Purification by flash chromatography [hexanes/AcOEt 25/75] and then precipitation from an acetone/hexanes mixture provided 9 as a white solid. Yield 77%. ¹H NMR (CD₃OD) δ: 0.17 (m, 2H), 0.53 (m, 2H), 0.87 (m, 1H), 1.25–1.62 (m, 4H), 1.99 (m, 1H), 2.12–2.29 (m, 2H), 2.47 (m, 2H), 2.73 (m, 2H), 3.06 (d, 1H, J = 8.3 Hz), 3.12 (d, 1H, J = 5.9 Hz), 3.89 (m, 1H), 4.63 (d, 1H, J = 7.8 Hz), 6.57 (d, 1H, J = 8.2 Hz), 6.69 (d, 1H, J = 8.2 Hz), 7.28–7.42 (m, 2H), 7.86–7.89 (m, 2H), 7.96 (s, 1H). ¹³C NMR (CD₃OD) δ: 4.13, 4.54, 11.62, 23.61, 25.87, 31.47, 32.78, 49.61, 50.83, 51.78, 53.93, 61.18, 64.74, 72.32, 93.83, 113.15, 114.52, 118.52, 120.89, 121.12, 122.94, 123.51, 123.64, 124.21, 124.83, 128.67, 134.57, 139.95, 143.80, 171.51; mp 192–194 °C. Anal. Calcd for C₂₉H₃₀N₂O₄S: C, 69.30; H, 6.02; N; 5.57. Found: C, 68.67; H, 5.74; N, 5.88. ESI-TOF MS calculated for C₂₉H₃₀N₂O₄S m/z, 502.627; found, 503.869 (MH⁺).

17-Cyclopropylmethyl-3,14β-dihydroxy-4,5α-epoxy-6β-[(2′-benzofuryl)acetamido]morphinan (10)

Purification by flash chromatography [hexanes/AcOEt 25/75] and then recrystallized from MeOH provided 10 as a white solid. Yield 92%. ¹H NMR (CD₃OD) δ: 0.19 (m, 2H), 0.56 (m, 2H), 0.93 (m, 1H), 1.44–1.70 (m, 4H), 2.04 (m, 1H), 2.19–2.32 (m, 2H), 2.45 (m, 2H), 2.71 (m, 2H), 3.09 (d, 1H, J = 8.3 Hz), 3.16 (d, 1H, J = 5.9 Hz), 3.96 (m, 1H), 4.69 (d, 1H, J = 7.8 Hz), 6.61 (d, 1H, J = 8.2 Hz), 6.66 (d, 1H, J = 8.1 Hz), 7.32 (t, 1H, J = 7.6 Hz), 7.46–7.49 (m, 2H), 7.62 (d, 1H, J = 8.2 Hz), 7.73 (d, 1H, J = 7.7 Hz). ¹³C NMR (CD₃OD) δ: 4.12, 4.57, 11.71, 23.59, 25.56, 31.36, 31.61, 48.78, 49.82, 51.48, 52.88, 60.16, 63.76, 71.76, 92.82, 111.38, 112.84, 118.68, 119.87, 119.99, 120.15, 121.47, 122.78, 123.74, 124.89, 128.19, 133.47, 139.74, 142.81, 168.42; mp 162–164 °C. Anal. Calcd for C₂₉H₃₀N₂O₅: C, 71.59; H, 6.21; N, 5.76. Found: C, 71.47; H, 6.15; N, 5.71. ESI-TOF MS calculated for C₂₉H₃₀N₂O₅m/z, 486.564; found, 487.04 (MH⁺).

Intracellular Calcium Release

The assay was performed as described previously.¹⁸ Briefly, HEK-293 cells were cultured at 37 °C in Dulbecco’s Modified Eagle’s Medium supplemented with 10% bovine calf serum and Pen/Strep antibiotics. These cells were transiently transfected with 200 ng/20000 cells of the opioid receptor cDNA using OptiMEM medium (Invitrogen) and Lipofectamine 2000 (Invitrogen, Carlsbad CA) reagent according to manufacturer’s protocol (1:2 wt/vol ratio for DNA:Lipofectamine; 16 μg DNA, 32 μL Lipofectamine for singly expressed receptors, 12 μg of each receptor DNA (24 μg total), 48 μL Lipofectamine for doubly expressed receptors). The cells were seeded into 96-well plates (half area; Corning) at 20000 cells/well after 24 h and assayed 48 h after transfection using the FLIPR calcium kit (Molecular devices) in a Flexstation-III apparatus (Molecular devices).

Animals

Male ICR-CD1 mice (17–25 g; Harlan, Madison, WI) employed in the testing were housed in groups of eight in a temperature- and humidity-controlled environment with unlimited access to food and water. They were maintained on a 12 h light/dark cycle. All experiments were approved by the Institutional Animal Care and Use Committee of the University of Minnesota (Minneapolis, MN).

Antinociceptive Testing

For the measurement of antinociception, the tail-flick latency assay^19,20 was employed (Tail Flick Analgesia Meter, Columbus Instruments, Columbus, Ohio). The tail-flick response was elicited by applying radiant heat to the dorsal side of the tail. The intensity of the heat was set so that the mouse flicks its tail within 2–3 s. The test latency was measured before drug treatment (control), and again after the drug treatment (test) at the peak time of the compound, a 10 s maximum cutoff time was used to prevent damage to the tail. Antinociception was quantified as the percent maximal possible effect (%MPE): %MPE = (test – control/10 – control) × 100²¹. Three groups of 8–10 mice were used for each dose–response curve, and each mouse was used only once. Peak times are established by measuring the antinociception 5, 10, 20, 30, and 60 min after administration of the compound being tested. ED₅₀ values with 95% confidence intervals (CI) were computed with GraphPad Prism 4 by using nonlinear regression methods.

Antagonist Synergy

The AD₅₀ for each antagonist was measured by changing the dose of the antagonist and keeping the agonist dose (ED₉₀) constant.^27,28 The theoretical AD₅₀ was based on the ratio of the two antagonsits being tested together. The theoretical was then compared to the actual AD₅₀ measured if there was no overlapping within the confident limits the compound was said to be synergistic.

Conditioned Place Preference

Place conditioning procedures are used to assess the rewarding and aversive effects of psychoactive drugs.³⁴ An unconditioned stimulus (US; drug) was paired with a conditioned stimulus (CS) and rewarding or aversive effects of the US cause a preference for or against the CS (context) to develop. Similar protocols have recently been used in rats to demonstrate both rewarding³⁶ and aversive properties^37,38 of the κ opioid receptor-selective agonist salvinorin A.

Apparatus

A two-chambered place preference apparatus (22 × 45 × 20 cm³) was used. The test chamber was divided into two compartments separated by an opaque wall with a guillotine-style door. The floors of the two experimental chambers differed in the floor construction. One floor was made of a series of 2.3 mm stainless steel rods, centered every 6.4 mm, and was termed the “rod” floor. The other floor, called the “mesh” floor, was made up of thin wire mesh with openings of 2.5 mm × 2.5 mm. This apparatus allowed for an unbiased (rod vs mesh) experimental design. All data was collected by video camera, and movements were analyzed by ANY-maze (Stoelting Co., Wood Dale, IL).

Preconditioning

Prior to the beginning of the experiment, mice were habituated to the experimenter through 3 days of handling and saline injections. After habituation to handling and injections, initial chamber bias was determined for each mouse by placing them in the test chamber and allowed free exploration of the apparatus for 20 min in a “pre-test”. Total time spent in rod- and mesh-floored sides was recorded and analyzed.

Conditioning

Three experiments were conducted: (1) INTA dose–response, (2) salvinorin A dose–response, and (3) antagonism of INTA with naloxone. In experiment 1, mice were assigned to have INTA (0.3, 1.0, 3.0, or 10.0 mg/kg, sc), vehicle (saline w/1% DMSO) associated with either the rod- or mesh-floored chamber (CS+), and saline paired to the other chamber (CS-chamber). In experiment 2, mice were assigned to have salvinorin A (0.3, 1.0, or 3.0 mg/kg, sc) or vehicle associated with either the rod- or mesh-floored chamber (CS+ chamber) and saline paired to the other chamber (CS-chamber). In experiment 3, mice were given an injection of naloxone (1.0 sc) or vehicle and a second injection of INTA (3.0 mg/kg) or vehicle prior to CS+ sessions and saline prior to CS-sessions. All drug group assignments were done pseudorandomly and counterbalanced for pretest performance. In the conditioning trials, each subject was injected with the drug or saline 20 min prior to being confined to the corresponding paired chamber for 30 min. A total of four drug and four saline trials were performed with one conditioning trial performed per day alternating between drug and saline (8 total days).

Test Sessions

Two tests were performed. On the day following the fourth conditioning session, mice were tested as before in the preconditioning test (free exploration for 20 min). A second test was to be performed after the eighth conditioning session. Place preference induced by the US was determined by the difference in time spent in the CS+ chamber during the test session as compared to pretest. Experiments 1 and 2 were analyzed across test sessions, and two-way ANOVA (with Bonferroni-corrected t test comparisons) was used to determine the main effects of conditioning and drug (INTA or salvinorin A) and the interaction of the two. Because of the multiple drug treatments in experiment 3, test 2 was separately analyzed with two-way ANOVA for the main effects of naloxone and INTA treatments and the interaction of the two.

Immunofluorescence Study for β-arrestin2 Recruitment

HEK293 cells (100 mm2 tissue culture plate, 80–100% confluent) were transiently transfected with MOP, KOP, or DOP (16 μg DNA in 32 μL Lipofectamine 2000 (Invitrogen)), or MOP + KOP, MOP + DOP, or KOP + DOP (12 μg DNA for each receptor in 48 μL of Lipofectamine 2000). After 24 h, cells were diluted 1:100 and replated in 8-well tissue culture slides (B-D Falcon) and incubated overnight. Ligands were diluted initially in 10% DMSO and 90% water. Serial dilutions were prepared in Dulbecco’s medium without serum and added to the wells to achieve the concentrations listed. Ligands were incubated with the cells for 60 min at 37 °C. The media was then gently removed, and cells were fixed using 4% paraformaldehyde in 0.2 M Sorenson’s buffer for 20 min. The fixative was gently removed, and the cells were washed three times with 1× phosphate buffered saline (PBS). Primary and secondary antibodies were diluted in in 0.3% Triton X-100, 1% normal donkey serum, 1% bovine serum albumin, and 0.01% sodium azide. Anti-β-arrestin2 (goat) antibodies (R&D Systems) were added to each well at a concentration of 1 μg/mL and incubated at 4 °C overnight on an oscillating shaker plate. Cells were washed as above and incubated with antigoat Northern Lights 493 (R&D Systems) for 30 min at room temperature. The cells were washed as above, except that during the last wash, DAPI (1 ug/mL) was added and incubated for 15 min at room temperature. The cells were mounted using i-Brite Plus (Neuromics) mounting medium and covered with a glass coverslip. High power images were recorded using a Leica DMI4000 fluorescent microscope and processed using ImageJ software (NIH).

Glossary

Abbreviations Used

GPCR

G-protein-coupled receptor

KOR

κ-opioid receptor

MOR

μ-opioid receptor

DOR

δ-opioid receptor

INTA

N-2′-indolyl-β-naltrexamine

NNTA

N-2′-naphthoyl-β-naltrexamine

6′-GNTI

6′-guanidinonaltrindole

FRET

fluorescence resonance energy transfer

BRET

bioluminescence resonance energy transfer

HEK293

human embryonic kidney 293

FLIPR

fluorescence imaging plate reader

RFU

relative fluorescence units

icv

intracerebroventricular

i.t.

intrathecal

MPE

percentage of maximum effect

DAMGO

[d-Ala²,N-MePhe⁴,Gly⁵]enkephalin

β-FNA

β-funaltrexamine

NTB

naltriben

DPDPE

[d-Pen²,d-Pen⁵]enkephalin

NorBNI

norbinaltorphimine

AD

antagonist dose

sc

subcutaneous

CPP

condition place preference

Supporting Information Available

EC₅₀s for the calcium mobilization assay, competitive binding for 3, and antagonist, tolerance, and dependence study data. This material is available free of charge via the Internet at http://pubs.acs.org.

The authors declare no competing financial interest.

Funding Statement

National Institutes of Health, United States