Characterization of a Potential KOR/DOR Dual Agonist with No Apparent Abuse Liability via a Complementary Structure-Activity Relationship Study on Nalfurafine Analogs

Abstract

Discovery of analgesics void of abuse liability is critical to battle the opioid crisis in the US. Among many strategies to achieve this goal, targeting more than one opioid receptor seems promising to minimize this unwanted side-effect while achieving a reasonable therapeutic profile. In the process of understanding the structure-activity relationship of nalfurafine, we identified a potential analgesic agent, NMF, as a dual KOR/DOR agonist with minimum abuse liability. Further characterizations, including primary in vitro ADMET studies (hERG toxicity, plasma protein binding, permeability and hepatic metabolism), and in vivo pharmacodynamic and toxicity profiling (time course, abuse liability, tolerance, withdrawal, respiratory depression, body weight, and locomotor activity) further confirmed NMF as a promising drug candidate for future development.

Graphical Abstract

INTRODUCTION

Over 20% of people in the United States experience daily pain¹ and approximately 10% of the patients who receive standard opioid treatment may develop addiction to the medications.² Despite vast efforts devoted to non-addictive analgesics from both academia and industry, there is no drugs that can replace the mainstay opioid analgesics for moderate-to-severe pain management, highlighting the urgent demand for more effective and safer pain medications.

It has been widely accepted that mu opioid receptor (MOR) activation produces analgesia as well as notorious abuse liability. So considerable attempts targeting different proteins or different pain mechanisms have been made and led to many interesting drug candidates.^3-8 One example is the development of G protein-biased MOR agonist, which is based on the hypothesis that the abuse-related and respiration-depressing side effects result from β-arrestin rather than G-protein-mediated signaling pathways. However, contrasting to the original hypothesis, these “G-protein biased” MOR agonists, including the newly marketed TRV-130 (Oliceridine, Figure 1), still possess undesired abuse liability and cause respiratory depression.^3,9,10

Figure 1.

Compounds targeting one or more opioid receptors.

On the other hand, there are three other members in the opioid receptor family: kappa opioid receptor (KOR), delta opioid receptor (DOR), and nociception/orphanin FQ peptide receptor (NOP), and activation on any of them may mediate analgesic effects.¹¹ Although agonists of the KOR, DOR, and NOP also carry side effects, such as dysphoria, seizures, and altered renal function, respectively, none of them cause euphoria or severe dependence like MOR agonists.^11-13 However, selective agonists solely targeting the KOR, DOR, or NOP have not achieved satisfactory clinical outputs either.¹¹

Developing lead compounds to modulate multiple targets simultaneously is an emerging strategy in drug discovery.^13-15 Shifting from “one-drug-one-target-one-disease” to polypharmacology has provided more options for treating complex diseases.^14,15 More importantly, some multitarget drugs have been shown to carry enhanced effectiveness and reduce unwanted side effects. In addition, compared with certain co-formulated drugs, some multitarget-directed single-drug molecules have shown much less complex PK/PD relationships or potential drug-drug interactions (DDIs).^14-16

Accordingly, an imminent strategy to develop analgesics is to bind with high affinity to multiple opioid receptors. This new generation of multi-functioning analgesics is expected with enhanced effectiveness, reduced side effects, and uncomplicated PK/PD relationships.^14-16 The most recent efforts have been focusing on MOR/KOR, MOR/DOR and MOR/NOP mixed ligands, which have led to many new chemical entities and novel lead structures.

Evidence has shown that both MOR and KOR are expressed in supraspinal regions such as periaqueductal grey and rostral ventral medulla, spinal cord, and dorsal root ganglia. Before dual MOR/KOR agonists were developed, the KOR agonist spiradoline had already been observed to potentiate fentanyl-induced antinociception when co-administered, implying the possibility of dual acting agents.¹⁷ Nalfurafine (NFU, Figure 1) is a potent MOR/KOR dual agonist which is used in Japan as an antipruritic for hemodialysis patients, and also showed potent analgesia in a variety of pain models in different animals.^18-22 More importantly, NFU at doses producing sufficient antinociception caused no reinforcing effects in rats and monkeys.^4,23 Furthermore, NFU was the first KOR agonist on the market, which was reportedly devoid of concerning side effects from KOR activation, such as dysphoria and psychomimetisis preclinically and clinically. Other examples of clinically efficacious dual MOR/KOR agonists such as pentazocine, nalbuphine, and dihydroetorphine (Figure 1), however, still retain the limitations of MOR agonists including abuse liability and respiratory depression.¹¹

Similar to MOR/KOR mixed ligands, the development of dual MOR/DOR agonists has aimed to attenuate tolerance and dependence in pain management. Another reason for conjointly targeting MOR/DOR is the discovered crosstalk and co-localization between these two receptors.²⁴ The co-administration of MOR agonist and DOR agonist, such as fentanyl and SNC80, has first shown the potentiated overall analgesic actions.²⁵ Then a proof-of-concept MOR/DOR dual-acting enkephalin-like peptide has been observed with antiallodynic and antihyperalgesic effects.²⁶ Recently reported small molecules including fentanyl-like and acylaminomorphinan compounds have also exhibited antinociception comparable to morphine, but the abuse potential of these MOR/DOR dual agonists still remains to be investigated.^27,28

In addition, there are many mixed ligands that target three or even all four opioid receptors. MP1104 is a successful example of compounds which possess high affinity and efficacy to the MOR, KOR and DOR. Such “low selectivity” did not hamper its efficacy or safety, instead, MP1104 was observed with potent analgesia and less side effects including dependence and respiratory depression.^29,30 Moreover, a novel “pan-agonist”, Cebranopadol, with high efficacy at all four opioid receptors, shows a broad analgesic activity and is under phase III clinical trials for the treatment of severe neuropathic and nociceptive pain.³¹

Altogether, ligands with a multi-opioid receptor pharmacology have been approved as potential clinically suitable pain treatments. But more extensive research is warranted in order to provide more comprehensive perspectives in all kinds of multi-target opioid agonists. For instance, limited attention, compared to other mixed opioid receptor ligands, has been paid to dual KOR/DOR agonists as analgesics,^12,32 while both target proteins naturally lack abuse liability.

Besides the likely lower abuse liability to be developed, another rationale to study KOR/DOR dual agonists as analgesics is the observations that co-administration of a KOR agonist and a DOR agonist can elicit a synergistic response of antinociception.^33,34 Although the existence of KOR/DOR heterodimer is still debatable, growing in vivo evidence indicating KOR/DOR co-localization and cross-talking both centrally and peripherally.^35-39

To our notice, NFU elicits its analgesic effects via dual agonistic function at MOR/KOR while it is also a DOR agonist with a moderate potency. Although some NFU derivatives have been synthesized and reported previously, many of these compounds have not been investigated or reported extensively in biological or pharmacological assays.^40-44 Taking its clinical success into consideration, NFU is an apparently preferred starting point to explore potential KOR/DOR agonists. Therefore, in the present work, we conducted a complementary structure-activity relationship (SAR) study to systematically understand the multi-opioid receptor pharmacology of NFU analogs in order to reveal potentially valuable information for developing dual KOR/DOR agonists.

RESULTS AND DISCUSSION

Molecular Design

The structural modifications on the NFU skeleton were focused on the 3-hydroxy (3-OH) group, C6 configuration, and amide nitrogen methylation. The 3-hydroxy group has been believed as critical for opioid receptor binding. But recent structural biology and pharmacology studies have revealed that 3-OH may be necessary for MOR binding, but not for the KOR.⁴⁵ Hence, it was hypothesized that 3-dehydroxylation may diminish MOR affinity, reducing the risk of abusive effects, but maintain KOR affinity. Secondly, even though methylation on amide nitrogen is not commonly seen in epoxymorphinan compounds with side chains, NFU has shown not only potent activities in many disease models but also a desired PK profile, including a 56% oral bioavailability.⁴⁶ In such, it was hypothesized that methylation on the amide nitrogen may improve the physiochemical properties of newly designed ligands. The design of different configurations of C6 are to give the side chain different conformational flexibility in order to achieve a preferred binding mode in the target proteins. Herein the SAR studies of 3-OH, C6 configuration, and amide nitrogen methylation are presented (Figure 2).

Figure 2.

Molecular design.

Chemical Syntheses

In the designed synthetic routes, two starting molecules, naltrexone and 3-dehydroxy naltrexone, were first needed to prepare the target compounds. Naltrexone was provided by NIDA, and 3-dehydroxy naltrexone was synthesized from naltrexone via activation and hydrogenation following the reported methods (Scheme 1).⁴⁷

Scheme 1.

Stereoselective synthetic methods for intermediate amines (1–8). (a). DMF, K₂CO₃, 1-phenyl-5-chlorotetrazole, r.t., 12 h. Yield: 85%. (b). H₂, 45-50 psi, w.t. 30% Pd/C, AcOH, 50-55 °C, 10 h. Yield: 75%. (c). BnNH₂, PhH, PTSA, reflux, 10 h. (d). NaBH₄, EtOH, 4Å MS, r.t. overnight. Yield of c-d: 26-28%. (e). H₂, 60-70 psi, MeOH, HCl, w.t. 28% Pd/C, r.t. 23 h. Yield: 43-83%. (f). TEA, MeCN, BnNH₂, HCOOH, 1% w.t. [Ru(p-cymene)Cl₂]₂, r.t. 40 or 70 h. Yield: 48-50%. (g). Bn₂NH, PhCOOH, Tol, PTSA, reflux, 18-30 h. (h). NaCNBH₃, EtOH, 4Å MS, r.t., overnight. Yield of g-h: 30-57%. (i). H₂ 60-70 psi, MeOH, HCl, w.t. 50% Pd/C, 70-72 h. Yield: 63-66%. (j). TEA, MeCN, CH₃NH₃•HCl, HCOOH, 2% wt. [Ru(p-cymene)Cl₂]_2, r.t. 24 h. Yield: 96-98% (k). BnNHCH₃, PhCOOH, PhH, PTSA, reflux, 24 h. (l). NaCNBH₃, EtOH, 4Å MS, r.t., overnight. (m). H₂ 60 psi, MeOH, HCl, 35% wt. Pd/C, r.t. 48 h. Yield of k-m: 19-20%.

Eight enantiopure intermediates were required to prepare the target compounds. Part of the protocols to get these essential intermediate amines (1-8) were made available from previous studies from others and our lab,^41,48,49 and a systematic description for all synthetic routes are briefly discussed in the following. For preparing 1-8, at least one stereoselective synthetic route was adopted and established with reasonable yields. Taking advantage of the unique “T-shape” epoxymorphinan skeleton, the stereoselectivity was introduced by either Noyori catalysts (f, j, Scheme 1), which required no chiral ligands in this case, or a conformation-induced-steric-hindrance method (c-d, g-h, k-l, Scheme 1). For comparison, the method using asymmetric catalysts has advantages including simple one-step setup, higher yields, and carcinogen-free. On the other hand, the conformation-induced-steric-hindrance method (c-e) is more time-efficient for obtaining compounds 1 or 5.

Additionally, “one-pot” methods were also explored and established as alternative synthetic routes for the four N-methylated intermediate amines (Scheme 2).⁴⁸ In such approaches, all starting materials were added together and allowed to reflux for 24 h. In both reactions, starting with naltrexone or 3-deydroxynaltrexone, both C6-epimers were generated while the 6β-epimers (4, 8) were the major products, respectively.

Scheme 2.

One-pot method for methylated amines (3-4, 7-8) and proposed mechanisms. (n). NaBH₃CN, MeOH, CH₃NH₃•HCl, reflux, 24 h. Yields: 9% for 3 and 62% for 4; 35% for 7 and 50% for 8.

The reaction mechanism of “one-pot” reaction was proposed here (Scheme 2) for the first time. The coexistence of both chair (A) and boat (B) conformations of ring C is believed to account for the formation of both epimers. Once all starting materials were heated in methanol, amination by methylamine took place on the 6-position carbonyl group. The resulted iminiums could present in the form of a pair of stereoisomers, A and B’ (Scheme 2). In the case of A, the reducing agent cyanoborohydride favored the less hindered β-face (red arrow), yielding compound 3 or 7, the 6α-epimers. On the other hand, a substantial A^1,3 strain between iminium hydrogen and epoxy ether oxygen in B’ led to a conformational change of ring C, from chair to boat. Subsequently, the iminium (B) with a more stable ring C was generated, to which the hydride mainly approached from the less hindered α-face (blue arrow), giving the 6β-epimer, 4 or 8.

The “one-pot” approach (Scheme 2) not only demands less effort in reaction setup and monitoring than the conformation-induced steric hindrance methods (Scheme 1), but also requires shorter reaction time. Hence, it is preferred to use a one-pot method to synthesize intermediates 4 and 8.

After all intermediate amines were acquired, the amide formation using (E)-3-(furan-3-yl)acryloyl chloride was conducted to prepare the final target compounds (compounds 9 to 16, Scheme 3). One extra step of 3-hydrolyzation (Scheme 3) was performed for compounds with the 3-hydroxy group. Then all target compounds were converted into their hydrochloride salts. Compound 12 (nalfurafine, or NFU) was prepared and tested along through all the following studies.

Scheme 3.

Preparation of target compounds 9-16. (o). DCM, TEA, (E)-3-(furan-3-yl)acryloyl chloride, 0°C to r.t., 3-5 h. (p). K₂CO₃, MeOH, 0°C to r.t., 16-48 h. (q). MeOH, methanolic HCl, 0°C to r.t., overnight. Reaction (p) was only conducted to 1, 3, 5, and 7. Overall yields: 10-50%.

In Vitro Binding and Functional Studies

Radioligand binding assays

Radioligand competitive binding assays are frequently performed to determine the binding affinity and selectivity of potential GPCR ligands. In this membrane-based assay, the binding affinity of the target compounds were determined based on their ability to compete and replace the corresponding radioligands at each receptor.^50,51

As shown in Table 1, all compounds except 13 acted as high-affinity KOR ligands with sub- or single-digit nanomolar K_i values. Among them, compound 10, 11 and 12 possessed the highest affinity (K_i < 0.2 nM). While 9-12 showed relatively higher binding affinities compared with their 3-dehydroxy counterparts 13-16, respectively, 14-16 still carried sub- or single-digit nanomolar K_i. Hence, 3-dehydroxygenation seemed not absolutely necessary for KOR binding, which supported our hypothesis. Comparing each C6-epimer pair, though C6 configuration seemed to have no consistent impact on KOR affinity, the dramatic affinity difference between 13 and 14 was interesting and might be useful in designing 3-dehydroxy KOR ligands. On the contrary, comparing the KOR binding affinity of 11, 12, and 16 with their des-methyl counterparts 9, 10, and 14 respectively, the role of amide methyl group seemed not as obvious in recognizing the KOR except for the case of compound 15 vs 13.

Table 1.

Binding Affinity Results for compound 9 to 16 at the KOR, DOR, and MOR.

Cmpd	R	C6*	R’	Binding Affinity Ki (nM) ± SEM			Selectivity
				KOR	DOR	MOR	M/K	D/K	D/M
9	-OH	α	-H	1.96 ± 0.17	20.0 ± 3.3	1.23 ± 0.12	0.6	10	16
10	-OH	β	-H	0.18 ± 0.03	4.74 ± 0.85	0.32 ± 0.05	1.8	26	15
11 (NMF)	-OH	α	-CH3	0.11 ± 0.01	0.64 ± 0.11	0.47 ± 0.06	4.1	5.6	1.4
12 (NFU)	-OH	β	-CH3	0.19 ± 0.01	116 ± 20	0.72 ± 0.09	3.8	603	161
13	-H	α	-H	45.7 ± 6.6	77.6 ± 5.9	87.3 ± 8.4	1.9	1.7	0.9
14	-H	β	-H	0.98 ± 0.27	340 ± 96	25.1 ± 2.6	26	345	14
15	-H	α	-CH3	0.46 ± 0.08	116 ± 24	9.59 ± 0.49	21	249	12
16	-H	β	-CH3	2.91 ± 0.22	1000 ± 190	123 ± 12	42	343	8.1

Note: SEM stands for standard error of mean, calculated from at least three individual experiments.

In contrast to KOR, the DOR binding affinity was dramatically reduced when the 3-OH group was removed (Table 1), i.e. compound 9 to 13 (Ki: 20.0 ± 3.3 to 77.6 ± 5.9 nM), 10 to 14 (K_i: 4.74 ± 0.85 to 340 ± 96 nM), 11 to 15 (K_i: 0.64 ± 0.11 to 116 ± 24 nM), and 12 to 16 (K_i: 116 ± 20 to 1000 ± 190 nM). Therefore, the 3-OH seemed essential for DOR binding. On the other hand, the impact of C6 configuration on DOR binding seemed not conclusive from these eight compounds. Comparing DOR binding affinity of the N-methylated compounds with their des-methyl counterparts, 10 with 12, 13 with 15, and 14 with 16, the compounds without N-methylation showed much lower K_i values, i.e. non-methylated amide seemed to be preferred for the DOR, except for the case of compounds 9 and 11.

The 3-hydroxy group played a critical role in MOR binding, that is, compounds 9 to 12 are potent MOR ligands with 0.3-1.2 nM K_i values (Table 1), whereas the MOR binding affinity of their 3-dehydroxy counterparts, 13 to 16, were lowered by 20-170 folds, respectively. It seemed that 3-dehydroxylation could diminish MOR affinity. The other two chemical features, however, C6 configuration and N-methylation seemed to have no evident impact on MOR binding.

In summary, 3-dehydroxylation on the epoxymorphinan skeleton may diminish MOR and DOR affinity, but not that of KOR, which could help enhance the selectivity for the KOR over both the MOR and DOR (Table 1). N-Methylation might be beneficial to KOR affinity but not to DOR binding.

[³⁵S]-GTPγS binding assays

Functional assays are usually performed to determine the efficacy and potency of GPCR ligands. [³⁵S]-GTPγS binding assay and calcium mobilization (flux) assay are two common and established functional assays used for evaluating opioid receptor ligands. The [³⁵S]-GTPγS assay is applied to test the direct activation on a receptor level, the activation of which by the GPCR can be quantified by measuring the amount of the radiolabeled GTP analog bound to the cell membrane.⁵¹

As shown in Table 2, compound 9 to 16 all acted as high-efficacy KOR agonists and seven out of eight were observed with > 90% efficacy. Among them, compound 10, 11 and 12 (NFU) are the most potent agonists with subnanomolar EC₅₀ values. Impressively, compound 11 possessed picomolar level potency in the [³⁵S]-GTPγS assay, almost four times more potent than NFU (12). Compounds 9, 14, 15, and 16 possessed single-to-double-digit nanomolar high potency, while the EC₅₀ of 13 is at three-digit nanomolar level. Comparing the EC₅₀ values of each pair of compounds (with or without 3-hydroxy group), 9 (15.3 ± 1.9 nM) vs 13 (135 ± 25 nM), 10 (0.36 ± 0.02 nM) vs 14 (46.0 ± 6.8 nM), 11 (0.075 ± 0.003 nM) vs 15 (2.52 ± 0.38 nM), and 12 (0.26 ± 0.02 nM) vs 16 (59.0 ± 7.1 nM), the ones with 3-hydroxy group always presented a higher potency. It seemed that, although not required for KOR binding affinity, 3-OH might reinforce KOR-activation. Then comparing the compounds with and without N-methylation, the methylated ones all possessed similar or lower EC₅₀ values, hence, the methyl group on the amide nitrogen atom might be beneficial to KOR potency.

Table 2.

[³⁵S]-GTPγS Functional Assay Results of Compounds 9 to 16.^a

Cmpd	R	*	R’	KOR		DOR		MOR
				EC50, nM	%Emaxb	EC50, nM	%Emaxc	EC50, nM	%Emaxd
9	-OH	α	-H	15.3 ± 1.9	73.7 ± 2.5	83.6 ± 28.8	108 ± 6	8.41 ± 0.98	25.9 ± 1.0
10	-OH	β	-H	0.36 ± 0.02	95.2 ± 4.9	24.8 ± 4.7	103 ± 11	1.36 ± 0.14	40.8 ± 2.2
11 (NMF)	-OH	α	-CH3	0.075 ± 0.003	92.5 ± 3.0	2.68 ± 0.74	113 ± 11	0.59 ± 0.03	34.3 ± 0.5
12 (NFU)	-OH	β	-CH3	0.26 ± 0.02	95.9 ± 4.1	141 ± 15	65.7 ± 6.4	0.51 ± 0.12	30.3 ± 3.7
13	-H	α	-H	135 ± 25	103 ± 2	714 ± 87	102 ± 8	283 ± 49	22.8 ± 2.1
14	-H	β	-H	46.0 ± 6.8	97.9 ± 2.4	1500 ± 140	107 ± 13	167 ± 23	27.7 ± 1.1
15	-H	α	-CH3	2.52 ± 0.38	98.4 ± 2.4	242 ± 28	108 ± 7	132 ± 10	55.8 ± 2.9
16	-H	β	-CH3	59.0 ± 7.1	94.6 ± 2.9	2100 ± 400	107 ± 8	1800 ± 1100	17.5 ± 4.4

^a.All values are presented as mean ± SEM.

^b.%E_max = 100 was defined using 5 μM U50,488H.

^c.%E_max = 100 was defined using 5 μM SNC80.

^d.%E_max = 100 was defined using E_max of 3 μM DAMGO.

As shown in Table 2, all eight compounds acted as DOR full agonists except for NFU with a moderate-to-high efficacy activation. Among them, compound 11 is the most potent agonist with single-digit nanomolar EC₅₀, followed by compounds 9 and 10 which showed double-digit nanomolar EC₅₀ values. Similar to the binding affinity pattern, compound 9 to 12 (EC₅₀: 83.6 ± 28.8, 24.8 ± 4.7, 2.68 ± 0.74, 141 ± 15 nM) showed much higher potency than their 3-dehydroxy counterparts 13 to 16 (EC₅₀: 714 ± 87, 1500 ± 140, 242 ± 28, 2100 ± 400 nM), indicating that 3-hydroxy group may be important for the potency of compounds as DOR agonists.

From the results of [³⁵S]-GTPγS assay for MOR (Table 2), all compounds achieved low to moderate efficacies at the MOR, and the agonism potency (EC₅₀) of compound 9 to 12 are 50-3500 times higher than those of their 3-dehydroxy counterparts, 13 to 16, respectively. It might be concluded that 3-dehydroxylation, the MOR affinity-deteriorating factor, may also diminish agonist potency on the MOR. And as expected, a reasonably linear correlation between binding affinity and potency was found for each compound on all three receptors (Figure S1). Such correlation may be useful to predict compounds’ potency from their affinity in the future study.

It was noticed that some previously reported compounds, e.g. compound 11 and 12, showed somewhat varied binding affinity and functional potency compared with literature,⁴⁴ which is probably due to the difference between membrane-based assays vs tissues-based ones.

Primary In Vivo Characterization

Warm-water tail immersion assays

Because the in vitro studies showed that all compounds acted as efficacious dual KOR/DOR agonists, all eight compounds were selected to be tested for their antinociceptive effects. Herein, this is the first report to evaluate this set of compounds as high-efficacy dual KOR/DOR agonists for their antinociceptive studies. The first in vivo approach we chose was a warm water immersion study.^52,53 A single-dose (10 mg/kg) screening was conducted first in mice to see whether the compounds could exhibit antinociceptive effects in vivo at this initial dose (Figure 3).

Figure 3.

Warm Water Tail Immersion Assays of compounds 9 to 16 at 10 mg/kg. (Mean values of %MPE for each group were presented and the error bar stands for SE (Standard Error). Compared with vehicle, ns: no significant difference at p < 0.05. #p < 0.1, **p < 0.01, ***p < 0.001, ****p < 0.0001)

In this assay, each compound was injected systemically (s.c.) to a group of six mice. After 20 minutes, the tail flick response times were determined. The longer duration their tails stayed in water before flicking, the higher antinociceptive effects the compound possessed. The antinociceptive effect was quantified using %MPE.

As shown in Figure 3, six out of eight compounds presented an antinociception effect in mice in the single-dose test. Surprisingly, no significant increase of %MPE was observed from compound 9 considering its even higher KOR and DOR agonist potency than compound 14 and 16 in vitro (Table 2). The absence of antinociception for compound 13 might mainly be due to its intrinsically low potency or unfavorable pharmacokinetics profile.

Afterwards, dose-response studies were carried out to evaluate the antinociception potency of six compounds. As shown in Table 3, all six compounds were more potent than the KOR agonist U50,488H. Moreover, four of them, compounds 11, 12, 15, and 16 have exhibited higher antinociception potency than the MOR agonist morphine. And the antinociceptive effects were believed to be mediated by both KOR and DOR activation while KOR may play a major role (Table 2 and 3).

Table 3.

Antinociception Potency ED₅₀ in Tail-flick Assays.

Compound	ED50 (95% CL), mg/kg
Morphine	2.34 (1.57 - 3.50)54
Fentanyl	0.015 (0.013 – 0.018)55
U50,488H	8.84 (5.51 - 14.17)54
10	2.414 (1.720 - 3.388)
11 (NMF)	0.037 (0.014 - 0.096)
12 (NFU)	0.046 (0.016 - 0.126)
14	3.827 (1.224 -11.97)
15	0.341 (0.196 - 0.604)
16	1.610 (1.057 - 2.452)

In addition, one interesting observation from Table 3 was that the methylated compounds presented higher antinociception potency than their des-methyl counterparts. For example, compound 12 was 50 times more potent than compound 10, though they possessed a similar KOR and DOR potency in vitro. Meanwhile, compound 16 was more potent than compound 14 in the tail immersion assays even though it was the opposite case in the [³⁵S]-GTPγS assay. Therefore, the amide N-methylation might play a critical role in increasing lipophilicity, which helped compounds 12 and 16 distribute through the body. As an electron-donating group, the methyl group was also likely to enhance the metabolic stability by making the amide and the acryloyl less susceptible to hydrolysis and Michael addition, respectively.

More importantly, compound 11 (NMF) possessed a comparable ED₅₀ to NFU (12). Considering its more potent DOR agonism compared to NFU, a KOR/DOR dual activating mechanism may play a major role in its antinociception potency.

Antinociception mechanism studies

Observed as the most potent agonist in vitro and in vivo, NMF was selected for further pharmacology and pharmacokinetics characterizations. In vivo selectivity study was firstly carried out to verify the proposed KOR/DOR dual activation mechanism of the antinociceptive effects of NMF.

Therefore, KOR-selective antagonist nor-BNI, DOR-selective antagonist NTI, and MOR-selective irreversible antagonist β-FNA were co-administered with NMF in warm-water tail immersion assays, respectively.³⁰

As shown in Figure 4, both nor-BNI and NTI were able to reduce the antinociception produced by 0.1 mg/kg NMF, the calculated ED₉₀, profoundly as expected. Moreover, their blocking effects seemed addable as the combination of two antagonists completely abolished the antinociception effect of NMF in mice. In addition, the irreversible MOR antagonist β-FNA was not able to block the antinociception of NMF antinociception effect at the dose tested, which indicated the possibility of minimum involvement of the MOR, which was likely to be caused by its low efficacy at the MOR (%Emax = 34.3). In comparison, NFU has been characterized as a selective agent to the MOR and KOR^20,56,57, but not the DOR⁵⁷. This is very intriguing since the only structural difference between these two compounds is their C6 chirality.

Figure 4.

Blocking the antinociceptive effects of NMF by selective KOR antagonist and DOR antagonist in warm water tail immersion assay, respectively. (nor-BNI and β-FNA were given s.c. at a dose of 10 mg/kg 24 h prior to 0.1 mg/kg NMF administration. NTI was injected s.c. at a dose of 15 mg/kg 30 minutes prior to 0.1 mg/kg NMF injection. Compared to 0.1 mg/kg NMF group: ****p < 0.0001. Compared to vehicle: #p < 0.05; ^: n.s. Error bars represent SE)

Time-course studies

In the warm water tail immersion time-course study, both morphine and NMF showed peak antinociceptive effects around 0.5 h. The onset of action for 10 mg/kg morphine and 0.1 mg/kg NMF seemed to require several minutes, while the antinociceptive effects of 0.5 mg/kg NMF was observed immediately after administration. The duration of antinociception was approximately 3 hours for 10 mg/kg morphine and 0.1 mg/kg of NMF, and the antinociceptive effects produced by 0.5 mg/kg NMF lasted up 8 h and were significantly greater than morphine from 3 h (Figure 5). As a comparison, NFU showed 3-4 hours action time at 0.003 mg/kg in AA-induced abdominal constriction⁵⁷, less than 2 hours at 0.03 mg/kg in a hot plate studies⁵⁷, and about 1.5 hours at 0.1 nmol (i.t.) in warm water tail withdrawal studies.⁵⁸

Figure 5.

Time course study of NMF and morphine. (All administration routes were s.c. compared to vehicle: *p<0.05, **p<0.01, ***p<0.001, *p<0.0001; Compared with the morphine group: ^^p<0.01, ^^^p<0.001, ^^^^p<0.0001. Error bars represent SD)

In Vivo Pharmacodynamic and Toxicity Studies

Abuse liability studies via self-administration method

As discussed above, abuse liability is the main concern for all opioid analgesics. As shown in Figure 6, fentanyl functioned as a reinforcer and 3.2 μg/kg/infusion fentanyl maintained significantly higher rates of responding than saline. The present fentanyl results are consistent with previous results demonstrating intravenous fentanyl self-administration in rats. ^59-62 In contrast to fentanyl, no NMF dose, across a 100-fold range, reinforced the responding rates in rats. Though the effective analgesic doses of NMF in rats remain to be examined, the current results have further supported the low efficacy of NMF at the MOR in vivo, as concluded from the in vitro and in vivo studies. More importantly, as a non-reinforcer in rats, NMF would be predicted to have low-to-no abuse liability, and further evaluation and development are warranted.

Figure 6.

Self-administration of NMF and fentanyl in rats. (Rats were trained to self-administer (i.v.) 32 μg/kg/infusion heroin. All points represent the mean (± SEM) number of infusions for NMF doses (n=10; 5 males and 5 females) and fentanyl (n=5; 3 males and 2 females). Filled symbols: significantly different from saline at p <0.05.

Tolerance potential

Chronic administration of analgesics may lead to tolerance, which has been widely observed for clinically used MOR agonists. In our study, both NMF and morphine groups developed antinociceptive tolerance on day 3 (Figure 7B and 7C), whereas NMF (0.1 mg/kg) retained a higher antinociceptive effect than morphine (10 mg/kg) in morphine-tolerated mice (Figure 7B and 7C) though the potency of NMF seems reduced under morphine-tolerated conditions.^30,63 Similarly, NFU also retained its antinociceptive effect in the morphine-tolerated mice.⁶⁴

Figure 7.

Drug tolerance and cross tolerance. (A) Experimental design. Mice were administered a vehicle, 10 mg/kg morphine, 0.1 mg/kg NMF or 0.5 mg/kg NMF twice daily for consecutive four days. (s.c.) (B) Antinociceptive effects of compounds were evaluated daily. Compared with D1 (day 1), *p<0.05, ***p<0.001. (C) On day 5, vehicle, 10 mg/kg morphine or 0.1 mg/kg NMF was given to the morphine-treated group, after which the warm-water tail immersion assay was again performed. *p<0.05, **p<0.01. (Error bars represent SD.)

Weight loss

Alterations in weight have been observed in subjects chronically exposed to morphine. Recently, the underlying mechanism of metabolic influence from opioids is also being studied.^65-67 Therefore, along with tolerance studies, weight changes were also recorded for a preliminary prediction of the effects of NMF on metabolism, and for informing animal care and potential patient care. Habituating in the same environment, the weight of mice in the vehicle group did not fluctuate significantly but the morphine group lost weight significantly after only one day. The chronic administration of 0.1 mg/kg NMF seemed to have a negligible influence in weight, whereas when the dose was increased to 0.5 mg/kg, weight loss was observed on Day 4. Therefore, compared to morphine, NMF seemed less likely to cause weight loss (Figure 8).

Figure 8.

Weight change in mice received chronic administration of vehicle, NMF or morphine. (compared to Day 1, *p<0.05, ****p<0.0001. Error bars represent SD.)

Withdrawal

Physical dependence usually occurs in subjects who repeatedly receive traditional opioid analgesics. Such dependence can always be visualized in morphine-dependent mice by injecting naloxone (NLX). NLX-precipitated withdrawal symptoms include wet dog shakes, paw tremors, jumps, and diarrhea.^68,69 The mice who received 0.1 mg/kg, a dose resulting in maximum antinociception in the same mice, of NMF twice daily for consecutive four days showed no withdrawal symptoms at all after challenged with 1 mg/kg NLX. Also, the mice in the 0.5 mg/kg NMF group exhibited no jumps or signs of diarrhea after NLX challenge. Although 0.5 mg/kg NMF group showed a few wet dog shakes and paw tremors after NLX injection, all of the observed withdrawal symptoms were significantly fewer than those of morphine-dependent mice. Therefore, NMF seems to result in less dependence, even at a very high dose, which is an advantage as a potential drug candidate to treat chronic pain. Moreover, if NMF (or future NMF analogs) is to be used for its antinociceptive effects in patients who suffer from opioid use disorders and receive treatments which are MOR antagonists/partial agonists, the occurrence of withdrawal symptoms is unlikely to be a concern while the pain in these patients can be alleviated (Figures 7C and 9).

Figure 9.

Withdrawal and diarrhea symptoms precipitated by naloxone (NLX) in chronic analgesic-administration mice. (In comparison to morphine pellet group: *p<0.05, **p<0.01, ****p<0.0001. Percentage of diarrhea was calculated as the number of mice who had diarrhea divided by the total mouse number in each group. The severity of diarrhea was not quantified. Error bars represent SEM.)

Respiratory depression

Agonism at the MOR can cause depressed respiration. As NMF exhibited partial MOR agonism in vitro, whole-body plethysmography was utilized to assess the possibility of NMF to induce respiratory depression. Overall, NMF (0.1 mg/kg) showed no significant effects on the respiration of tested mice except for insignificant decreases in frequency and minute volume which were observed in the first 10-15 mins after the administration. Therefore, the limited respiratory effects of NMF from this study agreed with the KOR-DOR dual agonist feature of the compound, but it should be also noted that higher doses of NMF might lead to MOR-mediated respiratory depression due to its partial agonism on the MOR (Figure 10). To compare, NFU did not induced respiratory depression at 0.001 mg/kg in monkeys either.⁷⁰

Figure 10.

The respiratory effects of NMF and morphine in mice. A) Respiratory frequency (breaths per min) change. B) Minute volume (mL/min) change. C) Tidal volume (mL) changes. (Error bars represent SD.)

Locomotor activity

Many KOR agonists have sedative side effects, which might be interpreted as “positive” results in pain-stimulated models, so evaluating the existence and degree of sedation NMF could elicit is important. On the other hand, hyperactivity has been observed in mice that received morphine injections.³⁰ Therefore, locomotor activity tests were employed to evaluate locomotor effects of NMF in mice quantitively, serving both purposes abovementioned.

As shown (Figure 11), no significant activity-reducing effects were observed after injection of 0.1 mg/kg of NMF or 0.5 mg/kg of NMF. This indicated the antinociceptive effects, rather than sedative effects, were resulted in the higher %MPE value in the warm-water tail immersion studies. Additionally, although mice in 0.1 mg/kg NMF group traveled more and made more jump and ambulatory attempts, no significant hyperactivity elicited from NMF. Overall, the absence of significant hypo- or hyper-locomotor activity suggest a great potential of NMF as a drug candidate for analgesia.

Figure 11.

Spontaneous locomotor activity change after drug administration. (Activity data was recorded for 30 min with the injection time as time zero. Compared to vehicle: ns, no significant difference at p<0.05. Error bars represent SD.)

Primary In Vitro ADMET Studies on NMF

Cardiac toxicity

Inhibition of K_v11.1 (simply denoted as hERG), a subunit of a potassium channel, can potentially prolong QT-interval and lead to a fatal tachyarrhythmia. This has made hERG a critical anti-target in drug development.⁷¹ Therefore, NMF was evaluated in an automated patch-clamp hERG inhibition assay. In this assay, the inhibition of tail current from a series of concentrations of NMF and a reference compound, E-4031, were tested concurrently using CHO-K1 cell line. The IC₅₀ of E-4031 was 18 nM and NMF 6.7 μM, respectively. Taking the extraordinary analgesic potency into consideration, NMF is not likely to cause cardiac toxicity by blocking hERG under potential therapeutic doses.

Plasma protein binding

Plasma protein binding (PPB) plays a significant role not only in chemical-induced toxicity but also in modulating drug concentration at the target sites, thereby influencing the in vivo efficacy. The t_1/2-determining parameters, Cl and V_d, are also partially dependent on PPB. Therefore, PPB assessment is a necessary routine study in drug discovery and development.^72,73 Additionally, interspecies difference in PPB, although not common, is sometimes encountered.^74,75 In order to better predict human PK/PD from preclinical species, the PPB of NMF was tested using both human and rat plasma. NMF exhibited 68% and 76% PPB in human plasma and rat plasma, respectively (Table 4). Both PPBs were higher than those of morphine, but very much lower than some alkaloids such as mitragynine and speciociliatine, which also functioned as opioid agonists and showed 90-99% PPB.⁷⁶

Table 4.

Plasma Protein Binding of NMF, NFU, and Morphine.

	NMF	NFU77	Morphine
plasma, human	68%	74%	35%78
plasma, rat	76%	68%	85%79

^a.NMF was tested at a concentration of 10 μM.

^b.Nalfurafine was tested at 2 nM.⁶²

Permeability studies

In vitro caco-2 permeability assay is a commonly used and accepted surrogate for predicting human intestinal absorption.^80-82 A compound with a greater-than-10x10⁻⁶ cm/s P_app has been conventionally considered as highly permeable, but the criteria and experiment condition are inconsistent through the literature.^80-84 So we evaluated NMF along with internal standards including propranolol, labetalol, and ranitidine, to more accurately predict the absorption feature of NMF and facilitate comparisons (Table 5).

Table 5.

Permeability of NMF and model drugs.

	A-B permeability (x10−6 cm/s)	B-A Permeability (x10−6 cm/s)	%fa82
NMF	10.1 ± 0.4	31.8 ± 0.8	-
Propranolol	23.8 ± 0.4	25.6 ± 0.7	100%
Labetalol	18.5 ± 0.3	34.2 ± 0.4	90%
Ranitidine	0.50 ± 0.02	1.3 ± 0.2	50%

^a.Test concentration of all compounds was 10 μM.

^b.%f_a: extent of absorption in humans. (Error bars represent SEM).

The absorptive permeability (A→B) of NMF was 10.1 x10⁻⁶ cm/s, which was slightly lower than the highly permeable drugs, propranolol (%f_a = 100%) and labetalol (%f_a = 90%) (Table 5). Also, P_app (A-B) of NMF was 20 times higher than Ranitidine, which showed a 50% absorption rate. Therefore, NMF was believed to possess a moderate-to-high permeability and an absorption rate ranging from 50% to 90%.

On the other hand, NMF showed a secretive transport permeability of 31.8 x10⁻⁶ cm/s, yielding an efflux ratio of 3.1. The efflux transport may pose a barrier for intestinal permeability, however according to the empirical rules summarized by Wang, J., et al.⁸⁵, this efflux ratio should not be a significant risk for reduced permeability.

Metabolism profiling

Satisfactory oral bioavailability has become critically important in drug development, in which not only permeability but also hepatic metabolic stability plays an important role. Hence, in vitro or ex vivo metabolism studies have become routine for preliminary examinations at an early discovery stage.⁸⁶ For this reason, NMF was incubated in human and rat liver S9 fractions for rapid evaluation of overall liver metabolism in these two species. Assuming first-order kinetics (Figure 5), the half-life of NMF was calculated to be over 60 minutes in the human liver S9 fraction, but less than 15 minutes in the rat. Such relatively high stability in the human liver and low in the rat was also observed with other drugs including imipramine, propranolol, and clozapine, which were tested in parallel in this study (Table 6). In short, NMF has shown reasonable stability in human liver metabolism and relatively lower stability in rat liver. This should be considered for all future in vivo preclinical and translational studies.

Table 6.

Half-lives of compounds in human and rat liver S9 fractions.

	NMF	imipramine	propranolol	clozapine	terfenadine	verapamil
T1/2 (human liver S9, min)	> 60	> 60	> 60	> 60	10	18
T1/2 (rat liver S9, min)	10	11	13	10	14	10

CONCLUSIONS

Opioid analgesics have been on the frontline of pain management for many years while their abuse liability has been recognized since their introduction to the field. Development of analgesics by targeting two or more opioid receptors seems a valid strategy based on recent effort in this regard. Through a complementary structure-activity relationship study on nalfurafine, the only KOR agonists approved clinically, we identified one of its analogs as a potent analgesic with minimum abuse liability by acting as a KOR/DOR dual agonist. Follow-up in vitro PK and in vivo PD studies demonstrated its drug-like properties as well as potential as a lead for future development. Meanwhile, more extensive examination on several principal side effects of KOR and DOR agonists should be considered in the future. For KOR agonists, one principal side effect is “dysphoria” that can be examined using either a drug discrimination assay or a place conditioning assay. For DOR agonists, the principal side effect is convulsions. Besides the above, clinically favorable administration routes should be explored in due course as well. In all, we believe this chemical entity as a KOR/DOR dual agonist provides us a viable avenue to develop novel analgesics with minimum abuse liability.

EXPERIMENTAL SECTION

Chemistry.

General methods.

Naltrexone (NTX) was obtained as a free base through the NIDA Drug Supply Program. Other reagents were purchased from commercial vendors (such as Sigma-Aldrich and Aldlab Chemicals) and used without further purification. Flash column chromatography was performed with silica gel columns (230-400 mesh, Merck). ¹H (400 MHz) and ¹³C (100 MHz) nuclear magnetic resonance (NMR) spectra were recorded with tetramethylsilane as the internal standard on a Bruker Ultrashield 400 Plus spectrometer. High resolution mass spectroscopy (HRMS) was performed on an Applied Bio Systems 3200 Q trap with a turbo V source for TurbolonSpray. HPLC analysis was done with a Varian ProStar 210 system on Microsorb-MV 100-5 C8/C18 column (250 mm × 4.6 mm) at 254 nm, eluting with acetonitrile/water (0.1% TFA) (85/15) at 1 mL/min over 30 min. Melting points were determined using OptiMelt automated melting point system (Fisher Scientific).

17-Cyclopropylmethyl-4,5α-epoxy-14-hydroxy-morphinan-6-one (S2)

See reference ⁴⁸

6α-Amino-17-cyclopropylmethyl-3, 14-dihydroxy-4, 5α-epoxy-morphinan dihydrochloride (1)

Conformation-induced Steric Hinderance Stereoselective Method: Naltrexone (1 g, 2.9 mmol), p-toluene sulfonic acid (PTSA, 60 mg, 0.3 mmol), and 170 mL anhydrous benzene were added to an oven-dried pear-shape round-bottom flask. After allowing the mixture to stir for 5 min, benzylamine (1.0 mL, 9.2 mmol) was added dropwise. Then a Dean-Stark apparatus was installed and the reaction was heated to reflux under N₂ for 10 h. After solvent was evaporated to almost dryness and reaction mixture was cooled down to r.t., 20 mL anhydrous EtOH and 2 g 4 Å MS were added and stirred for 5 min. Then NaBH₄ (117 mg, 3.1 mmol) was added and the reaction mixture was allowed to stir at r.t. overnight. The reaction mixture was filtered through celite and the filtrate was concentrated. Then the residue was washed with H₂O and extracted with DCM (3 x 100 mL). The combined organic layers were dried over sodium sulfate and concentrated. Purification by flash column chromatography using DCM/MeOH/0.2% NH₃·H₂O (DCM : MeOH from 100:1 to 50:1) gave light yellow oil-like intermediate 6α-benzylamino-17-cyclopropylmethyl-3, 14-dihydroxy-4, 5α-epoxy-morphinan (370 mg, 0.8 mmol, 28%). Noyori catalyst Method: Naltrexone (1 g, 2.9 mmol) and 4.0 mL anhydrous acetonitrile were added to a 50 mL round-bottom flask. A muddy suspension was observed and immediately turned into clear yellow solution after anhydrous triethylamine (TEA, 2.2 mL, 15.6 mmol) was added. Then benzylamine (1.0 mL, 9.2 mmol) was added dropwise followed by dropwise addition of formic acid (1.5 mL, 39.0 mmol). After 10 min, dichloro(p-cymene)Ru(II)dimer (11mg) was dissolved in 1.0 mL anhydrous acetonitrile and added dropwise. After reaction was allowed to stir for 70 h at r.t. the solvent was removed, pH was adjusted to 9 using NH₃·H₂O at the end of which no further salt formation was observed. Then, DCM (7 x 30 mL) was used to extract the product. Normal phase purification was performed on an NH₃·H₂O-basified column using DCM/MeOH/1% NH₃·H₂O (DCM : MeOH from 200:1 to 170:1) as the eluent. Pure oil-like intermediate 6α-benzylamino-17-cyclopropylmethyl-3, 14-dihydroxy-4, 5α-epoxy-morphinan was obtained (604 mg, 1.4 mmol, yield 48%). Then 6α-benzylamino-17-cyclopropylmethyl-3, 14-dihydroxy-4, 5α-epoxy-morphinan (220 mg, 0.5 mmol) was added to a hydrogenator bottle and 15 mL anhydrous methanol was used as the solvent. The pH of the reaction mixture was adjusted to approximately 2 using 0.4 mL concentrated hydrochloride. Then 10% palladium on carbon (62 mg) was transferred to the bottle and shaken well. Then the bottle was set on hydrogenator at r.t. and ran for 23 h. Then the catalyst was filtered off through celite and the filtrate was concentrated as the crude product. Cold mixture of methanol and isopropanol (1:9) was used to crystallize and light yellow powder was obtained (90 mg, 0.2 mmol, yield 43%). ¹H NMR (400 MHz, MeOD) δ 6.78 (d, J = 8 Hz, 1H), 6.69 (d, J = 8 Hz, 1H), 3.81 (dt, J₁ = 16 Hz, J₂ = 4 Hz, 1H, H_7axial), 3.31-3.27 (m, 2H), 2.95-2.68 (m, 5 H), 2.52 (m, 1H), 1.88-1.82 (m, 1H), 1.75-1.72 (m, 1H), 1.59 (dd, J₁ = 16 Hz, J₂ = 8 Hz, 1H), 1.21-1.09 (m, 1H), 1.02 (m, 1H), 0.73- 0.68 (m, 1H), 0.36 (m, 1H). HRMS C₂₀H₂₆N₂O₃ m/z calc. 342.1943, found [M+H]⁺ 343.2022.

6β-Amino-17-cyclopropylmethyl-3, 14-dihydroxy-4, 5α-epoxy-morphinan dihydrochloride (2)

Naltrexone (500 mg, 1.4 mmol), benzoic acid (400 mg, 1.8 mmol), p-toluene sulfonic acid (PTSA, 15 mg, 0.05 mmol), and 75 mL anhydrous toluene were added to a 250 mL oven-dried pear-shape round-bottom flask. After allowing the mixture to stir for 5 min, 20 mL anhydrous EtOH was added and the reaction mixture turned into a transparent solution. Then dibenzylamine (360 mg, 1.8 mmol) was added dropwise, followed by additional 25 mL anhydrous toluene. Next, a Dean-Stark apparatus was attached to the flask and reaction mixture was heated to reflux under N₂ overnight, during which 20 mL liquid was carefully removed from Dean-Stark apparatus. On the following day, the solvent was evaporated to almost dryness and reaction mixture was cooled down to r.t.. Then 20 mL fresh anhydrous EtOH and 2 g 4Å MS were added. After 5 min, NaCNBH₃ (100 mg, 1.5 mmol) was added and the reaction was allowed to stir at r.t. overnight. Then the reaction mixture was filtered through celite and concentrated in vacuo. After water-washing and DCM-extraction, methanol recrystallization was employed to give the desired intermediate, 17-cyclopropylmethyl-6β-dibenzylamino-3, 14-dihydroxy-4, 5a-epoxy-morphinan as fine white powder (390 mg, 0.8 mmol, yield 57%). Then 17-cyclopropylmethyl-6β-dibenzylamino-3, 14-dihydroxy-4, 5a-epoxy-morphinan (400 mg, 0.7 mmol) was added to a hydrogenator bottle and dissolved in 40 mL anhydrous methanol, resulting in a clear yellow solution. The pH of the reaction mixture was adjusted to approximately 2 using 0.4 mL concentrated hydrochloride. Then, the catalyst, 10% palladium on carbon (200 mg), was transferred to the bottle and shaken well. After the bottle was set on Parr hydrogenator under 60 psi, the hydrogenation reaction was allowed to run at r.t. for three days. Then the catalysts were filtered off and the filtrate was concentrated to a yellow oil. A mixture of methanol and isopropanol (1:9 or 1:10) was used to recrystallize and yield a pure product (2, 210 mg, 0.5 mmol, yield 66%). ¹H NMR (400 MHz, MeOD) δ 6.65 (d, J = 8 Hz, 1H), 6.61 (d, J = 8 Hz, 1H), 4.43 (d, J = 8 Hz, 1H), 3.58 (d, J = 4 Hz, 1H), 2.95-2.80 (m, 4H), 2.66-2.60 (m, 1H), 2.40 (d, J = 8 Hz, 1H), 1.97-1.90 (m, 1H), 1.87-1.85 (m, 3H)1.71-1.65 (m, 1H), 1.70-1.61 (m, 2H), 1.47-1.39 (m, 2H), 0.96-0.90 (m, 1H), 0.65-0.56 (m, 2H), 0.31-0.27 (m, 2H). HRMS C₂₀H₂₆N₂O₃ m/z calc 342.1943, found [M-H₂O+H]⁺ 325.1908, [M+H]⁺ 343.2014.

17-Cyclopropylmethyl-3, 14-dihydroxy-4,5α-epoxy-6α-methylaminomorphinan (3)

Noyori catalyst method: Naltrexone (168 mg, 0.5 mmol) and methylamine hydrochloride (139 mg, 1.8 mmol) were added to a round-bottom flask with 1.3 mL anhydrous acetonitrile and stirred for 15 min. Anhydrous triethylamine (0.7 mL) was added dropwise followed by the addition of 0.4 mL formic acid. Dichloro(p-cymene)Ru(II)dimer (3 mg) was then dissolved in 0.5 mL anhydrous acetonitrile and added. After reaction was allowed to stir for 24 h at r.t., a large amount of white solid was precipitated out. Filtration was performed to give off-white solid and dark green filtrate. The solids were dried and obtained as the salt form of the desired product (3), which was neutralized by NH₃·H₂O later. The filtrate was concentrated and adjusted to pH = 8-9 using NH₃·H₂O. Ethyl acetate was used for extraction and the organic layers were collected and concentrated. The products from the filtrate and the solids were combined (135 mg, 0.5 mmol, yield 98%). Note: If reaction mixture turned into a green suspension with little white precipitation, the solvent was removed directly and pH was adjusted to 8-9 using NH₃·H₂O. After extraction and concentration, the resulting residue was purified by flash silica column chromatography using DCM/MeOH/0.5%NH₃·H₂O (DCM : MeOH from 50:1 to 30:1). ¹H NMR (400 MHz, MeOD) δ 6.45 (d, J = 8 Hz, 1H), 6.34 (d, J = 8 Hz, 1H), 4.53 (d, J = 4 Hz, 1H, H₅), 2.97 (d, J = 8 Hz, 1H), 2.91 (d, J = 16 Hz, 1H), 2.87 (dt, J₁ = 16 Hz, J₂ = 4 Hz, 1H, H₇), 2.54-2.50 (m, 1H), 2.45 (dd, J₁ = 20 Hz, J₂ = 8 Hz, 1H), 2.32 (s, 1H), 2.28-2.16 (m, 2H), 2.15- 2.06 (m, 2H), 1.56-1.35 (m, 3H), 1.31 (dd, J₁ = 12 Hz, J₂ = 10 Hz, 1H), 0.77-0.65 (m, 1H), 0.64, -0.57 (m, 1H), 0.41-0.33 (m, 2H), 0.03-0.03 (m, 2H). HRMS C₂₁H₂₈N₂O₃ m/z calc 356.2100, found [M-H₂O-C₄H₆+H]⁺ 285.1613, [M-C₄H₈+H]⁺ 303.1722, [M-H₂O+H]⁺ 339.2078, [M+H]⁺ 357.2190, [M+Na⁺]⁺ 379.2016.

17-Cyclopropylmethyl-3, 14-dihydroxy-4,5a-epoxy-6β-methylaminomorphinan (4)

Naltrexone (450 mg, 1.3 mmol), methylamine hydrochloride (900 mg, 13.3 mmol), and 15 mL methanol was added in a round-bottom flask and stirred at r.t. After 5 min, NaCNBH₃ (130 mg, 2.1 mmol) was added and reaction mixture was heated to reflux for 20 h. After the reaction was completed, light yellowish solution was observed with some white precipitates at the flask bottom. After removal of the solvent, chloroform was added to dissolve the mixture and NaHCO₃ solution was added to adjust pH to approximately 9. Chloroform (3 x 70 mL) was used for extraction, after which organic layers were combined, dried over sodium sulfate, and evaporated to dryness. The purification was performed using flash column chromatography with EtOAc/MeOH/0.5% NH₃·H₂O (EtOAc : MeOH from 20:1 to 12:1) to give the 6β-compound 4 (280 mg, 0.8 mmol, yield 62%) and its 6a-epimer (3, 40 mg, yield 9%) (total yield 71%). ¹H NMR (400 MHz, CDCl₃) δ 6.66 (d, J = 8 Hz, 1H), 6.54 (d, J = 8 Hz, 1H), 5.45 (broad s, 1H), 4.48 (d, J = 8 Hz, 1H), 3.05 (d, J = 8 Hz, 1H, H₅), 3.00 (d, J = 16 Hz, 1H), 2.62 (dd, J₁ = 12 Hz, J₂ = 4 Hz, 1H), 2.59-2.52 (m, 2H), 2.47 (s, 3H), 2.36 (d, J = 8 Hz, 1H), 2.22 (td, J₁ = 12 Hz, J₂ = 4 Hz, 1H), 2.13 (td, J₁ = 12 Hz, J₂ = 4 Hz, 1H), 1.92-1.82 (m, 1H), 1.69-1.60 (m, 2H), 1.46-1.35 (m, 2H), 0.88-0.80 (m, 1H), 0.54-0.50 (m, 2H), 0.14-0.10 (m, 2H). HRMS C₂₁H₂₈N₂O₃ m/z calc 356.2100, found [M-H₂O-NHCH₃-C₃H₅-C₃H₅N+H]⁺ 213.0765, [M-H₂O-NHCH₃-C₃H₅+H]⁺ 267.1230, [M-H₂O-C₄H₆+H]⁺ 285.1573, [M-H₂O-NHCH₃+H]⁺ 308.1620, [M-H₂O+H]⁺ 339.2047, [M+H]⁺ 357.2159, [M+Na⁺]⁺ 379.1967.

6α-Amino-17-cyclopropylmethyl-4,5α-epoxy-14-hydroxy-morphinan dihydrochloride (5)

Conformation-induced steric hinderance stereoselective method: 3-Dehydroxy-naltrexone (1.2 g, 3.8 mmol), PTSA (74 mg, 0.4 mmol) and 200 mL anhydrous benzene were added to a 500 mL oven-dried pear-shape round-bottom flask. After allowing the mixture to stir for 5 min, benzylamine (0.6 mL, 5.7 mmol) was added dropwise. Then a Dean-Stark apparatus was attached to the flask and the reaction was heated to reflux under N₂ for 24 h, during which 20 mL liquid was cautiously removed from the bottom of Dean-Stark apparatus. Before reduction, solvent was evaporated to almost dryness and reaction mixture was cooled down to r.t. 40 mL anhydrous EtOH was added followed by 3 g 4 Å MS and stirred for 5 min. NaBH₄ (214 mg, 5.7 mmol) was then added and the reaction mixture was allowed to stir at r.t. overnight. After filtration through celite, the reaction mixture was concentrated, washed with H₂O and extracted with DCM. Then the combined organic layers were dried over sodium sulfate and concentrated. Purification by flash column chromatography using DCM/NH₃·H₂O (NH₃·H₂O : DCM from 0 to 1%) gave intermediate 6α-benzylamino-17-cyclopropylmethyl-4,5a-epoxy-14-hydroxy-morphinan (390 mg, 1.0 mmol, yield 26%) and starting material S2 (250 mg, 0.8 mmol). Noyori catalyst method: 3-Dehydroxy-naltrexone (1 g, 3.1 mmol) and 5.0 mL anhydrous acetonitrile were added to a 50 mL round-bottom flask, followed by addition of anhydrous TEA (2.2 mL, 15.6 mmol). Then, benzylamine (1.0 mL, 9.3 mmol) was added dropwise followed by dropwise addition of formic acid (1.5 mL, 40 mmol). After 10 min, dichloro(p-cymene)Ru(II)dimer (20 mg) was dissolved in 1.0 mL anhydrous acetonitrile and added dropwise. After reaction was allowed to stir for 40 h at r.t., the solvent was removed, pH was adjusted to 8-9 using NH₃·H₂O. Then, DCM (3 x 50 mL) was used to extract the product. After concentration, normal phase purification was performed on NH₃·H₂O-basified column using DCM/MeOH/1% NH₃·H₂O (DCM : MeOH from 200:1 to 170:1). Pure oil-like intermediate 6α-benzylamino-17-cyclopropylmethyl-4,5a-epoxy-14-hydroxy-morphinan was obtained (730 mg, 1.5 mmol, yield 50%). Then 6α-Benzylamino-17-cyclopropylmethyl-4,5a-epoxy-14-hydroxy-morphinan (280 mg, 0.6 mmol) was added to a hydrogenator bottle and 23 mL anhydrous methanol was used as the solvent. The pH of the reaction mixture was adjusted to approximately 2 using 0.3 mL concentrated hydrochloride. Then, catalyst, 10% palladium on carbon (112 mg), was transferred to the bottle and shaken well. The bottle was set on a Parr hydrogenator at r.t. and ran for 24 h under 60 psi. When the reaction was completed, the catalyst was filtered off and the filtrate was concentrated to an oil-like crude product. Cold methanol (or methanol : isopropanol = 1:9) was used to crystallize and yellow powder was obtained (206 mg, 0.5 mmol, yield 83%). ¹H NMR (400 MHz, DMSO-d₆) δ 9.02 (s, 1H), 8.48 (s, 1H), 7.20 (t, J = 8 Hz), 6.78 (d, J = 8 Hz), 6.70 (d, J = 8 Hz), 4.82 (d, J = 4 Hz, 1H), 4.04 (d, J = 8 Hz, 1H), 3.73 (s, 1H), 3.28-3.21 (m, 3H), 3.17-3.05 (m, 2H), 2.75-2.69 (m, 2H), 1.98-1.94 (m, 1H), 1.68 (d, J = 12 Hz, 2H), 1.44-1.40 (m, 1H), 1.11-1.09 (m, 1H), 0.94-0.85 (m, 1H), 0.70-0.60 (m, 2H), 0.51-0.40 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆) δ 155.13, 131.44, 129.71, 127.69, 119.33, 109.01, 88.22, 69.30, 61.10, 56.66, 52.34, 46.03, 44.96, 28.67, 27.07, 23.62, 21.31, 5.68, 5.09, 2.63. HRMS C₂₀H₂₆N₂O₂ m/z calc 326.1994, found [M-H₂O-C₄H₇+H⁺]⁺ 255.1493, [M-H₂O+H⁺]⁺ 309.1959, [M+H⁺]⁺ 327.2069.

6β-Amino-17-cyclopropylmethyl-4,5α-epoxy-14-hydroxy-morphinan dihydrochloride (6)

3-Dehydroxy-naltrexone (600 mg, 1.8 mmol), benzoic acid (248 mg, 2.0 mmol), PTSA (30 mg, 0.1 mmol), 75 mL anhydrous toluene, and 20 mL anhydrous EtOH were added to a 250 mL round-bottom flask. After allowing the mixture to stir for 15 min, dibenzylamine (0.4 mL, 1.9 mmol) was added dropwise. Then, a Dean-Stark apparatus was applied to the flask and the reaction mixture was heated to reflux under N₂ overnight. In the following morning, solvent was evaporated to almost dryness and reaction mixture was cooled down to r.t.. After 20 mL fresh anhydrous EtOH and 2.2 g 4Å MS were added and stirred for 5 min, NaCNBH₃ (140 mg, 2.2 mmol) was added and the reaction was allowed to stir at r.t. overnight. In the work-up, reaction mixture was first filtered through celite and the filtrate was concentrated in vacuo. The residue was washed with H₂O and extracted with DCM (6 x 50 mL). Then, the combined organic layers were dried over sodium sulfate and concentrated to a dark yellow oil. Purification by flash column chromatography using DCM/MeOH/0.5% NH₃·H₂O (MeOH : DCM from 0 to 1:100) gave the N-dibenzyl intermediate, 6β-dibenzylamino-17-cyclopropylmethyl-4,5a-epoxy-14-hydroxy-morphinan (280 mg, 0.55 mmol, yield 30%). Next, the N-dibenzyl intermediate was added to a hydrogenator bottle and methanol was used as the solvent. The pH of the reaction mixture was adjusted to approximately 2 using concentrated hydrochloride. Then, the catalyst, 10% palladium on carbon (90 mg), was transferred to the bottle and shaken well. The bottle was set on a hydrogenator under 60 psi at r.t. for three days. After filtration, filtrate was concentrated and washed by a small amount of methanol, resulting in the desired product 6 (yield 63%). ¹H NMR (400 MHz, DMSO-d₆) δ 9.01 (s, 1H), 8.58 (s, 3H), 7.22 (t, J = 8 Hz, 1H), 6.87 (d, J = 8 Hz, 1H), 6.80 (d, J = 8Hz, 1H), 6.56 (s, 1H), 4.77 (d, J = 8 Hz, 1H), 4.13 (s, 1H), 4.00 (s, 1H), 3.39-3.34 (m, 2H), 3.19-3.12 (m, 4H), 3.05 (d, J = 8Hz, 1H), 2.93-2.88 (m, 1H), 2.79-2.66 (m, 1H), 2.48-2.44 (m, 1H), 2.02 (q, J = 12 Hz, 1H), 1.88 (d, J = 12 Hz, 1H), 1.76 (d, J = 12 Hz, 1H), 1.43 (d, J = 12Hz, 1H), 1.29 (t, J = 12Hz, 1H), 1.10-1.07 (m, 1H), 0.69-0.65 (m, 1H), 0.63-0.58 (m, 1H), 0.54-0.50 (m, 1H), 0.43-0.39 (m, 1H). ¹³C NMR (101 MHz, DMSO) δ 155.67, 132.08, 130.19, 128.34, 119.85, 109.51, 88.76, 69.84, 61.53, 57.20, 52.88, 49.04, 45.51, 29.24, 27.65, 24.19, 21.80, 6.33, 5.66, 3.21. HRMS C₂₀H₂₆N₂O₂ m/z calc 326.1994, found [M+H⁺]⁺ 327.2059.

17-Cyclopropylmethyl-4,5α-epoxy-14-hydroxy-6α-mehthylaminomorphinan (7)

3-Dehydroxy-naltrexone (300 mg, 0.9 mmol) and methylamine hydrochloride (124 mg, 1.8 mmol) were added to a round-bottom flask with 1.0 mL anhydrous acetonitrile. Then, 0.8 mL anhydrous TEA was added followed by the dropwise addition of 0.4 mL formic acid and stirred for 30 min. Dichloro(p-cymene)Ru(II)dimer (8 mg) was then dissolved in 0.8 mL anhydrous acetonitrile and added dropwise to the reaction mixture. Reaction was allowed to stir for 24 h at r.t.. Next, white precipitate (free base of 7) was filtered off and filtrate was concentrated. The residual oil then was dissolved in ethyl acetate and adjusted to pH = 9 using NH₃·H₂O. Then, a small amount water was used to wash the organic layer and ethyl acetate (3 x 20 mL) was used for extraction. Next, the organic layers were concentrated to give light green oil-like residue. The combined product was purified by flash column chromatography (DCM/MeOH/1%NH₃·H₂O) to yield the pure product 7 (300 mg, 0.9 mmol, yield 96%). ¹H NMR (400 MHz, CDCl₃) δ 7.04 (t, J = 8 Hz, 1H), 6.61 (d, J = 8 Hz, 1H), 6.57 (d, J = 8 Hz, 1 H), 5.02 (s, 1H), 4.71 (d, J = 4 Hz, 1H), 3.11-3.05 (m, 3H), 2.71-2.61 (m, 2H), 2.53 (s, 3H), 2.41-2.32 (m, 2H), 2.27-2.21 (m, 2H), 1.77-1.54 (m, 3.5H), 1.44-1.37 (m, 1.5 H), 0.94-0.84 (m, 1H), 0.83-0.69 (m, 1H), 0.58-0.52 (m, 2H), 0.18-0.12 (m, 2H). ¹³C NMR (101 MHz, CDCl₃) δ 159.73, 134.62, 130.03, 128.81, 117.79, 105.72, 88.80, 69.99, 62.16, 59.69, 55.35, 46.12, 43.09, 34.15, 33.51, 29.95, 23.53, 21.00, 9.40, 3.97, 3.80. HRMS m/z C₂₁H₂₈N₂O₂ m/z calc 340.2151, found [M-H₂O+H]⁺ 323.2100, [M+H]⁺ 341.2211.

17-Cyclopropylmethyl-4,5α-epoxy-14-hydroxy-6β-methylaminomorphinan (8)

Stereoselective method: 3-Dehydroxy-naltrexone (484 mg, 1.5 mmol), benzoic acid (454 mg, 3.7 mmol), PTSA (28 mg, 0.2 mmol), and 50 mL anhydrous benzene were added to an oven-dried pear-shape round-bottom flask. After allowing the mixture to stir for 15 min, 0.2 mL benzylmethylamine was added dropwise. Next, a Dean-Stark apparatus was attached to the flask and reaction mixture was heated to reflux under N₂ for 24 h. On the next day, the solvent was evaporated to almost dryness and reaction mixture was cooled down to r.t. Then, 30 mL anhydrous EtOH was added followed by 2 g 4Å MS and stirred for 5 min. NaCNBH₃ (140 mg, 2.2 mmol) was then added and the reaction was allowed to stir at r.t. overnight. After the reduction was completed, the reaction mixture was filtered through Celite. The filtrate was concentrated, washed with H₂O and extracted with DCM (3 x 70 mL). The combined organic layers were dried over sodium sulfate and concentrated, after which the resulted intermediate 6β-N,N-benzylmethyl-3-dehydroxynaltrexamine (250 mg, 0.58 mmol) was added to a hydrogenator bottle and 20 mL methanol was used as the solvent. The pH of the solution was adjusted to approximately 2 using concentrated hydrochloride, and 10% palladium on carbon (78 mg) was transferred to the bottle and shaken well. After the hydrogenation was allowed to run under 60 psi at r.t. for two days, the catalyst was filtered off and the filtrate was concentrated. Cold methanol was used to crystalize and gave the final salt (97 mg, 0.3 mmol, overall yield 19%). One-pot method: 3-Dehydroxy-naltrexone (300 mg, 0.9 mmol), methylamine hydrochloride (622 mg, 9.2 mmol), NaCNBH₃ (75 mg, 1.2 mmol), and 10 mL anhydrous methanol were added to a round-bottom flask and stirred at r.t for 5 min. Then, reaction mixture was heated to reflux for 24 h. After methanol was removed, chloroform was added to dissolve the mixture and sodium bicarbonate solution was added to adjust pH to approximately 8. Chloroform was used for extraction and the resulting organic layers were combined, dried over sodium sulfate, and evaporated to dryness. The purification was conducted on normal phase silica column with EtOAc/MeOH/1% NH₃·H₂O (EtOAc : MeOH from 50:1 to 10:1) eluent system to give two desired compounds in approximately 50% yield for the beta-epimer (8, 157 mg) and 35% (7, 110 mg) yield for the alpha-epimer (total yield 85%). ¹H NMR (400 MHz, CDCl₃) δ 7.07 (t, J =7.8 Hz, 1H), 6.67 (d, J = 8 Hz, 1H), 6.65 (d, J = 8 Hz, 1H ), 4.36 (d, J = 8 Hz, 1 H), 3.49 (s, 1H), 3.10-3.09 (d, J = 8 Hz, 1H), 3.07 (d, J = 16 Hz, 1H), 2.68-2.62 (m, 2H), 2.46 (s, 3H), 2.41-2.33 (td, J₁ =12 Hz, J₂ = 8 Hz, 1H) 2.37 (d, J = 8 Hz, 2H), 2.27-2.18 (td, J₁ =12 Hz, J₂ = 8 Hz, 1 H), 2.12-2.06 (td, J₁ = 12 Hz, J₂ = 4Hz, 1H), 1.75-1.69 (m, 2 H) 1.64-1.59 (m, 5H), 1.44-1.31 (m, 3H), 0.90-0.81 (m, 2H), 0.55-0.49 (m, 2H), 0.17-0.09 (m, 2H). HRMS C₂₁H₂₈N₂O₂ m/z calc 340.2151, found [M-H₂O-CH₃NH-C₃H₅+H]⁺ 251.1309, [M-H₂O-CH₃NH+H]⁺ 292.1698, [M-H₂O+H]⁺ 323.2121, [M+H]⁺ 341.2230, [M+Na⁺]⁺ 363.2049, [2M+Na⁺]⁺ 703.4216.

General Procedures for the Final Compounds

Compounds 9 to 12

The corresponding intermediate amine (1-4) was dissolved in anhydrous DCM (0.1 M) and cooled down using an ice bath. Then, anhydrous TEA (4.0 equiv) was added dropwise. The reaction was allowed to stir for half an hour, followed by addition of (E)-3-(3-furanyl)-2-propenoyl chloride (3.5 equiv). After the starting material was consumed, the reaction was quenched using methanol and concentrated to dryness. After the crude product was redissolved in methanol, addition of K₂CO₃ was followed. The hydrolysis was completed after 14-50 h, and the purification was conducted using flash column chromatography (DCM/MeOH/1%NH₃·H₂O) gave pure products in form of free base. The free base was dissolved in anhydrous methanol, followed by addition of methanolic hydrochloride (1.3 equiv.) at 0 °C. The reaction was allowed to stir overnight and diethyl ether was added to precipitate the final hydrochloride salt (overall yield 10-30%).

Compounds 13 to 16

The corresponding intermediate amine (5-8) was dissolved in anhydrous DCM (0.1 M) and cooled down using an ice bath. Anhydrous TEA (4.0 equiv) was added dropwise. Reaction was allowed to stir for half an hour, followed by addition of (E)-3-(3-furanyl)-2-propenoyl chloride (2.0 equiv). After the starting material was consumed, the reaction was quenched using methanol and concentrated to dryness. Purification was done using flash column chromatography (DCM/MeOH/1%NH₃·H₂O) gave product in the form of free base. Then, the free base was dissolved in anhydrous methanol, followed by addition of methanolic hydrochloride (1.3 equiv.) at 0 °C. The reaction was allowed to stir overnight and then diethyl ether was added to precipitate the final hydrochloride salt (overall yield 20-50%).

(2E)-N-[6α-17-(Cyclopropylmethyl)-3,14-dihydroxy-4,5α-epoxy-morphinan-6-yl]-3-(3-furanyl)-2-propenamide hydrochloride (9)

The title compound was obtained following the general procedure for final compounds as a white solid. ¹H NMR (400 MHz, DMSO-d₆) δ 9.25 (s, 1H), 8.86 (broad s, 1H), 8.02 (s, 1H), 7.86 (d, J = 8 Hz, 1H), 7.74 (d, J = 4 Hz, 1H), 7.35 (d, J = 16 Hz, 1H), 6.73 (d, J = 8 Hz, 1H), 6.70 (d, J = 4 Hz, 1H), 6.57 (d, J = 8 Hz, 1H), 6.55(d, J = 16 Hz, 1H), 6.29 (s, 1H), 4.67(d, J = 4 Hz, 1H, H₅), 4.54-4.47 (m, 1H), 3.91(d, J = 8 Hz, 1H, H_7axial), 3.29-3.24 (m, 1H), 3.10-3.03 (m, 1H), 3.00-2.94 (m, 1H), 2.77-2.67 (m, 1H), 2.48-2.44 (m, 1H), 1.94-1.85 (m, 1H), 1.64 (d, J = 12 Hz, 1H), 1.11-1.03 (m, 1H), 0.97-0.92 (m, 1H), 0.73-0.69 (m, 1H), 0.65-0.59 (m, 1H), 0.52-0.46 (m, 1H), 0.43-0.37 (m, 1H). ¹³C NMR (101 MHz, DMSO-d₆) δ 165.00, 146.48, 145.30, 144.89, 139.31, 129.62, 129.23, 123.14, 122.57, 122.33, 119.60, 118.71, 107.98, 87.93, 69.82, 61.54, 57.51, 45.69, 45.65, 45.62, 30.69, 29.75, 23.95, 20.19, 6.16, 5.63, 3.04. HRMS m/z C₂₇H₃₀N₂O₅ calc 462.2155, found [M+H]⁺ 463.2245. Melting point: decomposed at 214.3 °C. HPLC: R_t = 7.106 min; purity: 98.2%.

(2E)-N-[6β-17-(Cyclopropylmethyl)-3,14-dihydroxy-4,5α-epoxy-morphinan-6-yl]-3-(3-furanyl)-2-propenamide hydrochloride (10)

The title compound was obtained following the general procedure for final compounds as a white solid. ¹H NMR (400 MHz, DMSO-d₆) δ 9.29 (s, 1H), 8.82 (s, 1H), 8.33 (d, J = 8 Hz, 1H), 8.00 (s, 1H), 7.72 (s, 1H), 7.31 (d, J = 12 Hz, 1H), 6.73 (s, 1H), 6.72 (d, J = 8 Hz, 1H), 6.65 (d, J = 8 Hz, 1H), 6.31 (d, J = 12 Hz, 1H), 6.17 (s, 1H), 4.60 (d, J = 8 Hz, 1H, H₅), 3.85 (d, J = 4 Hz, 1H), 3.56-3.47 (m, 1H), 3.35-3.30 (m, 2H), 3.09 (d, J = 8 Hz, 1H), 3.04 (d, J = 8 Hz, 1H), 2.86 (t, J = 12 Hz, 1H), 2.45-2.39 (m, 1H), 1.82-1.71 (m, 2H), 1.59-1.55 (m, 1H), 1.45 (d, J = 8 Hz, 1H), 1.37 (t, J = 12 Hz, 1H), 1.27-1.21 (m, 1H), 1.11-1.06 (m, 1H), 0.69-0.65 (m, 1H), 0.61-0.58 (m, 1H), 0.53-0.49 (m, 1H), 0.44-0.39 (m, 1H). ¹³C NMR (101 MHz, DMSO-d₆) δ 165.24, 145.28, 144.87, 142.62, 141.81, 130.13, 129.76, 123.03, 122.19, 121.05, 119.75, 118.44, 108.08, 90.47, 70.22, 62.22, 57.22, 51.25, 46.99, 46.10, 29.90, 27.82, 24.25, 23.50, 6.20, 5.57, 3.12. HRMS m/z C₂₇H₃₀N₂O₅ calc 462.2155, found [M+H]⁺ 463.2222, [M+Na]⁺ 485.2022. Melting point: decomposed at 231.5 °C. HPLC: R_t = 7.201 min; purity: 99.0%.

(2E)-N-[6β-17-(Cyclopropylmethyl)-3,14-dihydroxy-4,5α-epoxy-morphinan-6-yl]-3-(3-furanyl)-N-methyl-2-propenamide hydrochloride (11)

The title compound was obtained following the general procedure for final compounds as a white solid. ¹H NMR (400 MHz, DMSO-d₆) δ 9.32 (s, 1H), 8.83 (s, 1H), 8.06 (s, 1H), 7.73 (s, 1H), 7.45 (d, J = 16 Hz, 1H), 7.03 (s, 1H), 6.97 (d, J = 16 Hz, 1H), 6.73 (d, J = 8 Hz, 1H), 6.60 (d, J = 8 Hz, 1H), 6.27 (s, 1H), 5.04 (d, J = 12 Hz, 1H), 4.71 (d, J = 4 Hz, 1H, H₅), 3.92 (d, J = 8 Hz, 1H), 3.37-3.27 (m, 2H), 3.14-3.09 (m, 1H), 3.05 (s, 1H), 2.94-2.88 (m, 2H), 2.74-2.68 (m, 1H), 2.46-2.43 (m, 1H), 1.99-1.91 (m, 1H), 1.64-1.56 (m, 2H), 1.44-1.41 (m, 1H), 1.24-1.06 (m, 2H), 0.71- 0.60 (m, 2H), 0.49-0.40 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆) δ 166.42, 163.47, 146.17, 145.21, 145.11, 139.53, 132.45, 129.39, 123.61, 122.55, 119.81, 118.87, 118.58, 108.65, 89.47, 69.42, 61.59, 57.50, 49.44, 46.16, 45.71, 32.06, 30.58, 30.29, 23.89, 18.40, 6.17, 5.66, 3.05. HRMS C₂₈H₃₂N₂O₅ m/z calc 476.2311, found [M+H]⁺ 477.2201. Melting point: decomposed at 231.5 °C. HPLC: R_t = 7.417 min; purity: 98.4%.

(2E)-N-[6β-17-(Cyclopropylmethyl)-3,14-dihydroxy-4,5α-epoxy-morphinan-6-yl]-3-(3-furanyl)-N-methyl-2-propenamide hydrochloride (12)

The title compound was obtained following the general procedure for final compounds as a light yellow solid. ¹H NMR (400 MHz, DMSO-d₆) Abundance ratio of two tautomers is 2 : 3 = T1 : T2. δ 9.69 (s, 0.6 H, T2), 9.29 (s, 0.4H, T1), 8.85 (s, 1H), 8.04 (s, 0.4 H, T1), 7.92 (s, 0.6 H, T2), 7.73 (s, 0.4 H, T1), 7.67 (s, 0.6 H, T2), 7.37 (d, J = 16 Hz, 0.4H, T1), 7.23 (d, J = 16 Hz, 0.6H, T2), 7.01 (s, 0.4 H, T1), 6.91 (d, J = 16 Hz, 0.4H, T1), 6.85 (d, J = 8 Hz, 0.6H, T2), 6.71 (d, J = 8 Hz, 1H), 6.65 (d, J = 8 Hz, 0.4H, T1), 6.63 (s, 0.6 H, T2), 6.50 (s, 0.6 H, T2), 6.40 (s, 0.4 H, T1), 6.36 (d, J = 16 Hz, 0.6H, T2), 4.93 (d, J = 8 Hz, 0.4H, T1, H₅), 4.86 (d, J = 8 Hz, 0.6H, T2, H₅), 4.20 (m, 0.4 H, T1), 3.86 (m, 1H), 3.59 (m, 0.6 H, T2), 3.42- 3.38 (m, 2H), 3.16 (s, 1.2 H, T1), 3.09-3.05 (m, 2H), 2.93 (s, 1.9 H, T2), 2.89-2.87 (m, 2H), 2.60-2.52 (m, 2H), 2.22-2.06 (m, 1H), 1.48-1.25 (m, 3H), 0.70-0.66 (m, 1H), 0.61-0.59 (m, 1H), 0.53-0.50 (m, 1H), 0.44-0.41 (m, 1H). ¹³C NMR (101 MHz, DMSO-d₆) δ 165.90, 144.46, 144.17, 142.09, 141.34, 132.01, 130.88, 123.07, 122.97, 118.97, 118.52, 117.80, 107.64, 69.69, 69.60, 61.34, 61.15, 56.61, 56.57, 46.38, 45.90, 45.84, 30.36, 30.21, 22.93, 22.48, 5.71, 5.11, 2.58. HRMS C₂₈H₃₂N₂O₅ m/z calc 476.2311, found [M+H]⁺ 477.2394. Melting point: decomposed at 221.2 °C. HPLC: R_t = 7.563 min; purity: 98.9%.

(2E)-N-[6α-17-(Cyclopropylmethyl)-4,5α-epoxy-14-hydroxy-morphinan-6-yl]-3-(3-furanyl)-2-propenamide hydrochloride (13)

The title compound was obtained following the general procedure for final compounds as a white solid. ¹H NMR (400 MHz, DMSO-d6) δ 8.91 (s, 1H), 8.02 (s, 1H), 7.94 (d, J = 8 Hz, 1H), 7.74 (s, 1H), 7.35 (d, J = 16 Hz, 1H), 7.18 (t, J = 8 Hz, 1H), 6.75 (d, J = 8 Hz, 1H), 6.68 (d, J = 8 Hz, 1H), 6.67 (s, 1H), 6.52 (d, J = 16 Hz, 1H), 6.34 (s, 1H), 4.68 (d, J = 4 Hz, 1H), 4.57-4.48 (m, 1H), 3.95 (d, J = 8 Hz, 1H), 3.47 (d, J = 20 Hz, 1H), 3.29-3.26 (m, 1H), 3.17 (dd, J₁ = 20 Hz, J₂ = 8 Hz, 1H), 3.14-2.94 (m, 1H), 2.76-2.70 (m, 1H), 2.46-2.41 (m, 1H), 1.95-1.86 (m, 1H), 1.65-1.53 (m, 1H), 1.47-1.38 (m, 1H), 1.08 (m, 1H), 0.99-0.84 (m, 1H), 0.72-0.59 (m, 2H), 0.52-0.38 (m, 1H). ¹³C NMR (101 MHz, DMSO-d₆) δ 165.03, 159.77, 159.12, 145.32, 144.90, 133.15, 130.27, 129.65, 123.13, 122.25, 118.96, 107.95, 106.69, 69.74, 69.74, 61.39, 57.54, 45.49, 45.48, 44.93, 30.60, 29.63, 24.65, 20.05, 6.14, 5.64, 3.04. HRMS C₂₇H₃₀N₂O₄ m/z calc 446.2206, found [M+H]⁺ 447.2285. HPLC: R_t = 8.514 min; purity: 98.8%.

(2E)-N-[6β-17-(Cyclopropylmethyl)-4,5α-epoxy-14-hydroxy-morphinan-6-yl]-3-(3-furanyl)-2-propenamide hydrochloride (14)

The title compound was obtained following the general procedure for final compounds as a white solid. ¹H NMR (400 MHz, DMSO-d₆) δ 8.90 (s, 1H), 8.36 (d, J = 8 Hz, 1H), 8.01 (s, 1H), 7.73 (s, 1H), 7.30 (d, J = 16 Hz, 1H), 7.19 (t, J = 8 Hz, 1H), 6.84 (d, J = 8 Hz, 1H), 6.75 (d, J = 8 Hz, 1H), 6.74 (s, 1H), 6.30 (d, J = 16 Hz, 1H), 6.26 (s, 1H), 4.61 (d, J = 8 Hz, 1H, H₅), 3.91(d, J = 4 Hz, 1H), 3.56-3.42 (m, 2H), 3.19 (dd, J₁ = 20 Hz, J₂ = 8 Hz, 1H), 3.10-3.01 (m, 2H), 2.91-2.84 (m, 1H), 2.46-2.41 (m, 2H), 1.84-1.74 (m, 2H), 1.57-1.52 (m, 1H), 1.44 (d, J = 8 Hz, 1H), 1.40-1.33 (m, 1H), 1.11-1.06 (m, 1H), 0.73-0.66 (m, 1H), 0.63-0.57(m, 1H), 0.55-0.49 (m, 1H), 0.45-0.39 (m, 1H). ¹³C NMR (101 MHz, DMSO-d₆) δ 165.17, 156.31, 145.30, 144.96, 131.69, 129.97, 129.85, 128.84, 123.01, 122.04, 119.24, 109.46, 108.07, 90.85, 70.13, 61.98, 57.16, 51.00, 46.72, 45.92, 45.41, 39.39, 29.93, 27.67, 24.17, 6.19, 5.59, 3.11. HRMS C₂₇H₃₀N₂O₄ m/z calc 446.2206, found [M+H]⁺ 447.2296. Melting point: decomposed at 232.9 °C. HPLC: R_t = 7.657 min; purity: 97.9%.

(2E)-N-[6a-17-(Cyclopropylmethyl)-4,5a-epoxy-14-hydroxy-morphinan-6-yl]-3-(3-furanyl)-N-methyl-2-propenamide hydrochloride (15)

The title compound was obtained following the general procedure for final compounds as a white solid. ¹H NMR (400 MHz, DMSO-d₆) δ 8.89 (s, 1H), 8.14 (s, 0.2H), 8.06 (s, 0.5H), 7.73 (s, 0.5H), 7.66 (s, 0.2H), 7.46 (d, J = 16 Hz, 1H), 7.20 (t, J = 8 Hz, 1H), 7.22-7.18 (m, 0.2H), 7.01 (s, 0.5H), 6.96 (d, J = 16 Hz, 1H), 6.78 (d, J = 8 Hz, 1H), 6.70 (d, J = 8 Hz, 1H), 6.71-6.69 (m, 0.5H), 6.34 (s, 1H), 5.10-4.98 (m, 1H), 4.72 (d, J = 4 Hz, 1H), 3.97-3.93 (m, 1H), 3.50-3.45 (m, 1H), 3.30-3.18 (m, 2H), 3.07-2.79 (m, 1H), 2.72-2.6 (m, 1H), 2.03-1.91 (m, 1H), 1.64-1.53 (m, 2H), 1.45-1.42 (m, 1H), 1.24-1.05 (m, 3H), 0.73-0.61 (m, 2H), 0.52-0.40 (m, 1H). ¹³C NMR (101 MHz, DMSO-d₆) δ 163.48, 145.14, 137.15, 133.22, 130.47, 128.03, 127.96, 127.37, 123.59, 122.58, 119.28, 118.79, 108.64, 96.41, 75.60, 69.38, 61.47, 61.39, 57.53, 45.45, 45.42, 39.59, 39.38, 30.40, 30.33, 30.28, 24.59, 6.15, 5.65, 3.06, 0.58. HRMS C₂₈H₃₂N₂O₄ m/z calc 460.2362, found [M+H]⁺ 461.2417. Melting point: decomposed at 248.3 °C. HPLC: R_t = 8.705 min; purity: 99.0%.

(2E)-N-[6β-17-(Cyclopropylmethyl)-4,5α-epoxy-14-hydroxy-morphinan-6-yl]-3-(3-furanyl)-N-methyl-2-propenamide hydrochloride (16)

The title compound was obtained following the general procedure for final compounds as a white solid. ¹H NMR (400 MHz, CDCl₃) δ 7.57 (s, 1H), 7.50 (d, J = 16 Hz, 1H), 7.39 (s, 1H), 7.16 (t, J = 8 Hz, 1H ), 6.76 (d, J = 8 Hz, 1H), 6.70 (d, J = 8 Hz, 1H), 6.42 (s, 1H), 6.38 (d, J = 16 Hz, 1H), 4.58 (d, J = 8 Hz, 1H, H₅), 3.72-3.66 (m, 1H), 3.14 (s, 1H), 3.12 (d, J = 12 Hz, 1H), 3.03 (s, 3H), 2.71 (d, J = 8 Hz, 1H), 2.67 (d, J = 4 Hz, 1H), 2.40 (d, J = 4 Hz, 2H), 2.28 (td, J₁ = 8 Hz, J₂ = 4 Hz, 2H), 2.11 (td, J₁ = 12 Hz, J₂ = 4 Hz, 1H), 1.71 (dt, J₁ = 12 Hz, J₂ = 4 Hz, 1H), 1.49-1.45 (m, 3H), 0.90-0.81 (m, 1H), 0.58-0.53 (m, 2H), 0.18-0.11 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆) δ 163.48, 155.83, 145.21, 145.09, 144.99, 144.86, 131.19, 129.96, 129.84, 129.15, 123.42, 119.01, 109.51, 87.61, 70.14, 61.77, 61.65, 57.14, 46.69, 45.68, 45.65, 29.41, 27.40, 24.11, 21.11, 6.19, 5.60, 3.10. HRMS C₂₈H₃₂N₂O₄ m/z calc 460.2362, found [M+H]⁺ 461.2414. Melting point: decomposed at 251.2 °C. HPLC: R_t = 8.251 min; purity: 97.3%.

In Vitro Biological Evaluations.

Competitive radioligand binding assay.

[³H]Naloxone was used to label the MOR and [³H]diprenorphine was used to label the KOR and DOR. The K_d and B_max values of the MOR, KOR, and DOR were determined using 5 μM NTX, U50,488H and SNC80, respectively. The cells were centrifuged at 1000 g for 10 min and at 50,000 g for 10 min in membrane buffer. After homogenization in 4 mL TME buffer (50 mM Tris, 3 mM MgCl₂, and 0.2 mM EGTA, pH 7.4.), the Bradford assay was conducted to determine the concentration of the membrane protein. 300 μL TME buffer and 50 μL radioligand solutions was added to all test-tubes. 50 μL of unlabeled ligands and testing compounds solutions were added to their respective test-tubes. 50 μL Radioligand solutions were also to 3 scintillation vials containing 4 mL scintillation fluid as standard. Finally, 100 μl of membrane protein (30 μg/tube) was added to all the test-tubes to afford a total volume of 500 μl. All the test-tubes were vortexed, and incubated at 30 °C for 90 min. After incubation, the samples containing bound radioligands were filtered using a Brandel harvester. The filter papers containing filtered samples were then transferred into the scintillation vials filled with 4 mL of scintillation fluid. After 9 h, the samples were quantified using the liquid scintillation counter. Competition for bound radioligand was calculated using nonlinear regression analysis to determine the IC₅₀ values with GraphPad 6.0 software. The Ki values were determined from the IC₅₀ values using the Cheng-Prusoff equation. The assay was performed in duplicates and repeated at least three times.

[³⁵S]-GTPγS binding assay.

10 μg of mMOR-CHO, mKOR-CHO or mDOR-CHO membrane protein was incubated with 20 μM GDP, 0.1 nM [³⁵S]-GTPγS, assay buffer (TME + 100 mM NaCl), and varying concentrations of the testing compounds for 90 min in a 30 °C water bath. Nonspecific binding was determined with 20 μM unlabeled GTPγS. 3 μM DAMGO, 5 μM U50,488H or 5 μM SNC80 was included as maximally effective concentration of a full agonist for the MOR, KOR or DOR, respectively. Assay buffer was used for all the dilutions. 250 μL Assay buffer, 50 μL of GDP, 50 μL of cold GTPγS, testing compounds, and 50 μL [³⁵S]-GTPγS were added to the test-tubes accordingly. 50 μL [³⁵S]-GTPγS was also added to 2 scintillation vials as standards. Finally, 100 μL of membrane protein were added and all test tubes were incubated at 30 °C for 90 min. Similar work-up and quantification procedures were performed to the competitive radioligand binding assay. Percent DAMGO/U50,488H/SNC80-stimulated [³⁵S]-GTPγS binding was defined as (net-stimulated binding by ligand/net-stimulated binding by 3 μM DAMGO/5 μM U50,488H/5 μM SNC80) × 100. The normalized data were subjected to nonlinear regression analysis to determine EC₅₀ and E_max values using GraphPad 6.0 software.

hERG (automated patch-lamp) toxicity.

hERG toxicity study was performed in CHO-K1 cell line. The degree of inhibition (%) was obtained by measuring the tail current amplitude, which is induced by a one second test pulse to -40 mV after a two second pulse to + 20 mV, before and after drug incubation (the difference current was normalized to the control). Concentration (log) response curves were fitted to a logistic equation (three parameters assuming complete block of the current at very high test compound concentrations) to generate estimates of the 50 % inhibitory concentration (IC₅₀). The concentration response relationship of the test compound was constructed from the percentage reductions of current amplitude by sequential concentrations.

Caco-2 permeability.

The apparent permeability coefficient (P_app) of the test compound was calculated as follows:

$${\mathrm{P}}_{\mathrm{app}}(\mathrm{cm}∕\mathrm{s})=\frac{{\mathrm{V}}_{\mathrm{R}}\ast {\mathrm{C}}_{\mathrm{R},\text{end}}}{\mathrm{Delta}\mathrm{t}}\ast \frac{1}{\mathrm{A}\ast ({\mathrm{C}}_{\mathrm{D},\text{mid}}-{\mathrm{C}}_{\mathrm{R},\text{mid}})}$$

where V_R is the volume of the receiver chamber. C_R,end is the concentration of the test compound in the receiver chamber at the end time point, Δt is the incubation time and A is the surface area of the cell monolayer. C_D,mid is the calculated mid-point concentration, which is the mean value of the donor concentration at time 0 and the concentration at the end time point. C_R,mid is the mid-point concentration in the receiver side. Concentrations of the test compound were expressed as peak areas of the test compound. Fluorescein was used as the cell monolayer integrity marker. Fluorescein permeability assessment (in the A-B direction at pH 7.4 on both sides) was performed after the permeability assay for the test compound. The cell monolayer that had a fluorescein permeability of less than 1.5 x 10⁻⁶ cm/s.

Protein binding.⁸⁷

Dialysis membranes were soaked in DI water, 30% ethanol and isotonic sodium phosphate buffer subsequently. Plasma was obtained by centrifuge from fresh blood. The dialysate side of the 96-well dialysis apparatus was loaded with 0.15 mL of phosphate buffer (0.05 M sodium phosphate in 0.07 M NaCl, pH 7.5). The same volume of plasma spiked with 10 mM test compound was pipetted into the sample side. After 8 h incubation at 37 °C, post-dialysis plasma and buffer volumes were recorded and 90 mL of phosphate buffer was added to every 10 mL of plasma, and then precipitated with two volumes of acetonitrile. The quantification data was collected using LC-MS and calculated as following:

$$\text{Protein binding}(\%)=\frac{{\text{Area}}_{\mathrm{p}}-{\text{Area}}_{\mathrm{b}}}{{\text{Area}}_{\mathrm{p}}}\ast 100$$

$$\text{Recovery}(\%)=\frac{{\text{Area}}_{\mathrm{p}}+{\text{Area}}_{\mathrm{b}}}{{\text{Area}}_{\mathrm{c}}}\ast 100$$

Where Area_p = Peak area of analyte in protein matrix; Area_b = Peak area of analyte in buffer; Area_c = Peak area of analyte in control sample

Hepatic metabolism S9 fraction incubation⁸⁸

0.1 μM of NMF or reference compounds was tested in human liver S9 plus 1 mM UDPGA or rat liver S9 plus 1 mM UDPGA, respectively. At time 0, 15, 30, 45 and 60 minutes of incubation, the concentration of each compound was determined using LC-MS. After the experiment, metabolic stability, expressed as percent of the parent compound remaining, was calculated by comparing the peak area of the compound at the time point relative to that at time-0. The half-life (T_1/2) was estimated from the slope of the initial linear range of the logarithmic curve of compound remaining (%) vs. time, assuming the first-order kinetics.

In Vivo Studies.

Animals.

Mice: 5-8 Week 25-35 g male Swiss Webster mice (Envigo Laboratories, Frederick, MD, USA) were housed in cages (5 maximal per cage) in animal care quarters and maintained at 22 ± 2 °C on a 12 h light-dark cycle, except for the mice used for respiration measurement who were maintained in the reversed light-dark cycle. Food (standard chow) and water were available ad libitum. The mice were brought to the lab (22 ± 2 °C, 12 h light/dark cycle) and allowed 18 h to recover from transport. All studies used at least six mice for each group, and withdrawal studies were performed in the respective mice that were used in tolerance studies. Rats: A total of 10 rats (5 males and 5 females) Sprague-Dawley rats were acquired at approximately 8-10 weeks of age (Envigo Laboratories, Frederick, MD, USA) and surgically implanted with custom-made jugular catheters and vascular access ports (Instech, Plymouth Meeting, PA, USA) as described previously.⁸⁹ Rats were singly housed in a temperature and humidity-controlled vivarium that was maintained on a 12 h light/dark cycle. Water and food (Teklad Rat Diet, Envigo) were provided ad libitum in the home cage. Protocols and procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at Virginia Commonwealth University Medical Center and comply with the recommendations of the IASP (International Association for the Study of Pain).

Warm-water immersion assay.

The tail-flick test was performed using a water bath with the temperature maintained at 56 ± 0.1°C. The distal one-third of the tail was immersed perpendicularly in water, and the mouse rapidly flicked the tail from the bath at the first sign of discomfort. The duration of time the tail remained in the water bath was counted as the baseline latency. Untreated mice with baseline latency ranging from 2 to 4 seconds were used. Test latency was obtained 20 min later after injection (s.c.). A 10-second maximum cutoff latency was used to prevent any tissue damage. Antinociception was quantified as the percentage of maximal possible effect (%MPE), which was calculated as %MPE= [(test latency – control latency)/(10 – control latency)] × 100. Each compound was tested in a group of six mice. In dose-response studies, each dose of each compound was tested in a group of six mice. The doses were designed to give %MPE from 80%-20% for an accurate ED₅₀. ED₅₀ values were calculated using the least squares linear regression analysis followed by calculation of 95% confidence interval by the Bliss method. One-way ANOVA followed by the post-hoc Dunnett test was performed to assess the significance using Prism 8.0 software (GraphPad Software, San Diego, CA).

Drug self-administration.

Heroin hydrochloride and fentanyl hydrochloride were provided by the National Institute on Drug Abuse Drug Supply Program (Bethesda, MD). Modular operant chambers located in sound-attenuating cubicles (Med Associates, St. Albans, VT) were equipped with two retractable levers and LED lights. Intravenous (i.v.) compound solutions were delivered as described previously.⁸⁹ After each behavioral session, catheters were flushed with gentamicin (0.4 mg) and catheter patency was verified at the end of each experiment by instantaneous muscle tone loss following methohexital (0.5 mg) administration. Rats were initially trained to respond for i.v. heroin (32 μg/kg/infusion) under a fixed ratio (FR) 5 / time out 20 s schedule of reinforcement during daily 2 h sessions. Each session began with a non-contingent infusion of the available heroin dose followed by a 60 s time out. The response period was signaled by the extension of only the right lever and illumination of the right green stimulus light. Following each response requirement completion, the lever was retracted, the green light was extinguished, and a heroin dose was infused. This schedule was in effect until the number of heroin infusions earned per session was within 20% of the running mean for three consecutive sessions. Subsequently, saline was substituted for heroin every other session (i.e., SDSDS; S, saline; D, drug) until the number of saline infusions earned was at least 75% lower than the number of heroin infusions earned during the preceding heroin session for two consecutive alternations. The same experimental program was utilized during the saline substitution sessions, using the same infusion duration as a 32 μg/kg/infusion of heroin of 5 s per 300g of rat weight. Once training criteria were met, test sessions were inserted into the sequence (i.e., DTSTD or STDTS; T, test) to evaluate responding maintained by a range of NMF doses (1-100 μg/kg/infusion). Fentanyl doses (0.32-10 μg/kg/infusion) were also determined as a positive control. Saline and each unit NMF and fentanyl dose was tested once in each rat using a counterbalanced dosing order. The primary dependent measure for the NMF and fentanyl self-administration studies was the number of infusions earned per session and these data were plotted as a function of drug dose. Data were analyzed using a one-way repeated-measures ANOVA followed with Dunnet’s post-hoc tests using Prism 9.0 software.

Toxicity, tolerance and cross tolerance.

Vehicle, 0.1 mg/kg NMF, 0.5 mg/kg NMF, and 10 mg/kg morphine was given (s.c.) to four groups of mice twice a day. Weight was measured before drug administration every day and warm water tail immersion was performed to each mouse 20 min after the first injection in the day. On day 5, vehicle, 10 mg/kg morphine, and 0.1 mg/kg NMF, were given to the mice who received morphine injections for 4 days continuously, respectively. Then cross tolerance was evaluated using warm water tail immersion experiment. Weight change and tolerance data were analyzed using t-test, i.e., data collected from day 2-4 was compared to day 1, and cross tolerance data was analyzed using one-way ANOVA followed with Dunnet’s post-hoc tests using Prism 8.0 software.

Withdrawal studies.

morphine-dependent mice: 75 mg morphine pellets were implanted as described previously.⁹⁰ In brief, mice were anesthetized with 2.5% isoflurane, neck was shaved and cleaned with povidone-iodine, and then a 1-cm horizontal incision was made at the base of the neck. A 75-mg morphine pellet was inserted in the space before closing the site with Clay Adams Brand, MikRon AutoClip 9-mm wound clips (BD Diagnostics, Sparks, MD). The animals were allowed to recover in their home cages where they remained throughout the experiment. NMF-dependent mice: 0.1 mg/kg NMF or 0.5 mg/kg NMF was given subcutaneously to mice every 12 hours for four consecutive days. On day 5, 1 mg/kg of naloxone was administered to NMF-treated and morphine-pelleted mice, 3 minutes after which the numbers of paw tremors, wet dog shake, and jumping were counted for a period of 20 minutes for each mouse. And the occurrence of diarrhea was recorded as 0 or 1 for each mouse at the end of each withdrawal-observation experiment. Data was analyzed using one-way ANOVA followed with Dunnet’s post-hoc tests using Prism 8.0 software.

Measurement of respiration.

Respiration was measured in freely moving mice using plethysmography chambers (EMKA Technologies, France) supplied with a 5% CO₂ in air mixture (BOC Gas Supplies, UK) as described previously.⁹¹ Mice were habituated to plethysmograph chambers for 15 min before the experimentation. A 5-min baseline respiration period was recorded prior to challenge with any compound. Rate and depth of respiration were recorded and averaged over 1 min periods and converted to minute volume (rate × tidal volume). Tidal volume was calculated from the raw inspiration and expiration data⁹¹. Data were normalized as percentage of baseline and analyzed using a one-way repeated-measures ANOVA followed with Dunnet’s post-hoc tests using Prism 9.0 software.

Locomotor activities.

Activity chambers (Med Associates, St. Albans, VT) were used for locomotor activity study. Each individual chamber has closeable doors and a ventilation system. The interior of the chamber consists of a 27 x 27 cm Plexiglas enclosure that is wired with photo-beam cells connected to a computer console that counts the activity of the animal. Mice were habituated to the chamber for 20 minutes 24 h before the experiment. On the day of experiment, mice were injected with a desired dose of the compound subcutaneously and placed in the chambers immediately. Ambulatory counts, jumps, distance traveled, and average speed were monitored and recorded for 30 minutes. Data were analyzed using a one-way repeated-measures ANOVA followed with Dunnet’s post-hoc tests using Prism 9.0 software.

Figure 12.

Metabolism study of NMF in human and rat liver S9 fractions. Error bars represent SD.

ABBREVIATIONS AND ACROYNMS

CHO

Chinese hamster ovary

CL

clearance

CNS

central nervous system

DDI

drug-drug interaction

DOR

delta opioid receptor

FDA

Food and Drug Administration

KOR

kappa opioid receptor

MOR

mu opioid receptor

MPE

maximum possible effects

P_app

apparent permeability coefficient

PD

pharmacodynamics

PK

pharmacokinetics

PPB

plasma protein binding

V_d

volume of distribution