Variation of the net charge, lipophilicity and side chain flexibility in Dmt¹-DALDA: effect on opioid activity and biodistribution

Abstract

The influence of the side chain charges of the second and fourth amino acid residues in the peptidic μ opioid lead agonist Dmt-D-Arg-Phe-Lys-NH₂ ([Dmt¹]-DALDA) was examined. Additionally, to increase the overall lipophilicity of [Dmt¹]-DALDA and to investigate the Phe³ side chain flexibility, the final amide bond was N-methylated and Phe³ was replaced by a constrained aminobenzazepine analogue. The in vitro receptor binding and activity of the peptides, as well as their in vivo transport (brain in- and efflux and tissue biodistribution) and antinociceptive properties after peripheral administration (i.p. and s.c.) in mice were determined. The structural modifications result in significant shifts of receptor binding, activity and transport properties. Strikingly, while [Dmt¹]-DALDA and its N-methyl analogue, Dmt-D-Arg-Phe-NMeLys-NH₂, showed a long-lasting antinociceptive effect (>7h), the peptides with D-Cit² generate potent antinociception more rapidly (maximal effect at 1h post-injection) but also lose their analgesic activity faster, when compared to [Dmt¹]-DALDA and [Dmt¹,NMeLys⁴]-DALDA.

Introduction

Due to the present limitations of the pharmacotherapy of pain, the development of novel medications for the treatment of severe and chronic pain represents a high priority goal. To date, these pathologies rely mostly on opioid analgesics, but the current available opioid drugs produce a number of side-effects that limit their clinical use. A considerable amount of work has been done and is still needed to target more efficacious and non-addicting analgesics in order to solve this major health care problem. One research avenue consists of the discovery and development of a potent peptide-based analgesic, with reduced undesired effects, requiring a compound with a high affinity for μ and/or δ opioid receptors (MOR and DOR, resp.), a potent in vitro agonist activity, a strong antinociceptive effect in vivo, an improved resistance to enzymatic degradation and the ability to cross the blood-brain barrier (BBB), allowing systemic administration. More specifically, it was evidenced that compounds with a dual MOR/DOR activity present beneficial pharmacological effects in comparison to highly selective MOR agonists. There have been reports on synergistic effects between MOR agonists and DOR agonists.^1,2 Therefore, compounds with a mixed μ agonist/δ agonist profile have the potential to be effective for the treatment of pain with reduced side effects, as lower doses would be required.³ In parallel to the development of such dual agonists, mixed MOR agonists/DOR antagonists were also designed in search of a better pain management.^4-6 Among other beneficial effects, the observation that the selective blockade of DOR suppressed morphine tolerance and dependence motivated the development of compounds with this mixed profile.⁷

Earlier studies, performed by Schiller and coworkers,⁸ on the dermorphin (H-Tyr-D-Ala-Phe-Gly-Tyr-Pro-Ser-NH₂)-derived tetrapeptide H-Dmt-D-Arg-Phe-Lys-NH₂ 1 ([Dmt¹]DALDA, Figure 1), identified this peptide as a highly potent MOR agonist with a high binding affinity at the MOR (K_i^μ = 0.143 nM, 7-fold higher than morphine) and a high metabolic stability, when incubated in sheep blood (t_1/2 = 1.8 h).⁹ The C-terminal amidation and insertion of a D-Arg residue in position 2 are likely to assist in the high stability of this peptide.⁹ Moreover, [Dmt¹]DALDA showed extraordinary potency in the rat tail-flick assay after intrathecal administration (3000 times more potent than morphine),¹⁰ and a subcutaneous (s.c.) administration also induced a potent analgesic response,^11,12 which in turn indicated that this molecule is able to cross the BBB. Nevertheless, the structural features of this tri-charged peptide responsible for its ability to cross the BBB have not clearly been determined.

Figure 1.

Structures of Dmt-DALDA 1, Phe and the conformationally constrained 4-amino-1,2,4,5-tetrahydro-2-benzazepin-3-one (Aba)-containing dipeptide 2 (with Xxx: an α-amino acid residue).

The introduction of a conformational constraint and the N-methylation of peptidic amide bonds are well-known procedures to improve the stability and the activity of (opioid) peptides.^13,14 Previously, we have used the 4-amino-1,2,4,5-tetrahydro-2-benzazepin-3-one (Aba) scaffold 2 (Figure 1) as a tool for introducing a local conformational constraint in opioid peptides.^15-19 The synthetic Aba residue can be considered as a constrained phenylalanine where the phenyl group is fixed to the amine of the next amino acid via a methylene bridge.¹⁷ Previous work on opioid peptides indicated that replacement of Phe by Aba in dermorphin-based sequences improves the DOR affinity considerably, while preserving the binding to the MOR.^17,18 Additionally, peptides containing this constraint represent highly active antinociceptive agents when tested in vivo, as recently reported.¹⁹ Similarly, amide N-methylation in opioid peptide sequences have shown to result in retained opioid receptor binding and activation,²⁰ as well as improved BBB passage²¹ when compared to the non-methylated analogues. Therefore, both conformational side chain restriction and N-methylation prove to be efficient techniques in improving the pharmacokinetic and pharmacodynamic profile of dual MOR/DOR opioid peptides.²²

The present work describes the chemical modification of the MOR lead tetrapeptide [Dmt¹]DALDA (Figure 1). The different modifications that were carried out comprise: i) changing the overall net charge in the peptide by replacing the D-Arg² residue by a D-Cit² and/or the Lys⁴ residue by a Nle⁴, ii) the introduction of a N-methyl group in the amide bond connecting residues in position 3 and 4, in order to protect the peptide against enzymatic degradation and to increase the overall lipophilicity of the peptide, and iii) introducing a conformational constraint by means of an Aba³ residue that replaces Phe³ in the Dmt¹[DALDA] sequence. The synthesized peptides have been investigated for their opioid receptor binding affinity and agonist potency, BBB permeability, plasma stability and antinociceptive activities. Structure-activity relationship (SAR) studies relating to the substitution pattern in targeted positions within this series were pursued.

Results and discussion

Synthesis

All standard and N-methylated amino acids were obtained from commercial sources, whereas the conformationally constrained aminobenzazepinones of type 2 were prepared as described below, and used in standard solid phase peptide synthesis.

1. Synthesis of the 4-amino-2-benzazepin-3-one scaffolds

The preparation of the required Aba-Lys/Nle dipeptidomimetics 8a,b was performed using a previously reported²³ and optimized procedure²⁴ (Scheme 1). The phthaloyl-protected (S)-o-cyano-phenylalanine 3, was obtained by reaction of methyl 2-[(succinimidooxy)-carbonyl]benzoate²⁵ with the commercially available o-(S)-cyanophenylalanine. The corresponding o-formyl intermediate 4 was formed by hydrogenation with Raney Nickel in an acetic acid/water/pyridine mixture. Reductive amination of 4 with lysine and norleucine amino acid methyl esters resulted in secondary amines 5a/b, which were subsequently used for the dicyclohexylcarbodiimide-induced ring closure to present the desired aminobenzazepinones 6a/b. Ester hydrolysis, followed by phthaloyl-removal with hydrazine hydrate and final Fmoc-protection gave the required Fmoc-Aba-Lys(Z)/Nle-OH building blocks 8a/b which were used in the peptide synthesis.

Scheme 1.

Synthesis of Fmoc-Aba-Lys(Z)-OH 8a and Fmoc-Aba-Nle-OH 8b dipeptide mimetics.

Reaction conditions: a) H₂ (50 psi), RaNi, H₂O/AcOH/Py 1:1:1, 50°C, 38h, 78%; b) 1.1 eq HCl.H-Lys(Z)-OMe or HCl.H-Nle-OMe, 2.5 eq. NaCNBH₃, 20wt% MgSO₄, CH₂Cl₂, NMM or AcOH, r.t., 14h; c) 1.1. eq. DCC, Py, CH₃CN/ H₂O 1:1, r.t., 14h; d) acetone/HCl 1N 1:1, r.t., reflux, 16h; e) i. 4 eq. NH₂NH₂.H₂O, EtOH, reflux, 1.5h and ii. 1.2 eq. Fmoc-OSu, 2.2 eq. NaHCO₃, acetone/water 1:2, r.t, 2h.

2. Peptide synthesis

The peptides in this study were synthesized manually using N^α-Fmoc- or N^α-Boc-protected amino acids on Rink amide resin (0.31 mmol scale) and MBHA resin (0.29 mmol scale) using DIC and 1-hydroxybenzotriazole (HOBt) or O-(benzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium tetrafluoroborate (TBTU)/N,N-diisopropylethylamine (DIPEA) as the coupling reagents/mixtures. A threefold excess of the building blocks (Fmoc-Aba-Lys(Z)-OH, Fmoc-Aba-Nle-OH, Fmoc-Lys(Boc)-OH, Boc-Lys(Z)-OH, Boc-NMeLys(Z)-OH, Fmoc-Nle-OH, Boc-NMeNle-OH, Fmoc- or Boc-Phe-OH, Fmoc-D-Arg(Pbf)-OH, Boc-D-Arg(Tos)-OH, Boc-D-Cit-OH and Boc-Dmt-OH) and activating agents was applied and dry DMF was used as a solvent. Fmoc-deprotections were carried out by treating the resin twice (5 min and 15 min) with 20% piperidine in DMF. Boc-Dmt-OH was used in a final coupling, to directly present the fully unprotected peptide after acidic cleavage from the resin. Boc-Dmt-OH was coupled with 1,3-diisopropylcarbodiimide (DIC) and HOBt in order to avoid side reactions when uronium-based coupling reagents were used. Final cleavage of the peptide as well as the 2,2,4,6,7-pentamethyldihydrobenzofuran-5-sulfonyl (Pbf) side chain protection group removal was accomplished by treatment with TFA/CH₂Cl₂/anisol 90:5:5 for 90 min (Rink resin) or standard HF/anisol cleavage conditions for 60 min (MBHA resin). The crude peptides were obtained after evaporation of the cleavage mixture, precipitation in cold Et₂O, dissolving in glacial acetic acid and subsequent lyophilization. Final purification was performed by RP-HPLC on a SUPELCO DiscoveryBIO wide pore preparative C18 column in 45% overall yield and was >95% pure as determined by analytical RP-HPLC. The structures of the compounds were confirmed by high-resolution electrospray (ESP) ionization mass-spectrometry.

Biological evaluation and structure-activity relationships

In vitro opioid receptor binding and activity (Table 1)

Table 1.

Structures, receptor MOR/DOR binding affinities and in vitro potency data

Structure	Comp.	MOR Ki (nM) c	DOR Ki (nM) c	KOR Ki (nM)	Ki ratio μ/δ/κ	GPI (μ) IC50 (nM) c	MVD (δ) IC50 (nM) c
H-Dmt-D-Arg-Phe-Lys-NH2d	1	0.143	2100	22.3	1/11400/156	1.41	23.1
H-Dmt-D-Arg-Phe-Nle-NH2	9	0.23	15.2	16.2	1/66/70	2.2	6.2
H-Dmt-D-Cit-Phe-Lys-NH2	10	0.57	50.0	20.9	1/88/37	0.31	1.4
H-Dmt-D-Cit-Phe-Nle-NH2	11	0.79	3.98	98.5	1/5/125	0.89	4.7
H-Dmt-D-Arg-Phe-NMeLys-NH2	12	0.26	309	56.1	1/1190/216	8.3	7.0
H-Dmt-D-Arg-Phe-NMeNle-NH2	13	0.26	16.0	14.4	1/62/55	4.35	1.5
H-Dmt-D-Cit-Phe-NMeLys-NH2	14	0.94	39.2	11.5	1/42/12	1.01	4.4
H-Dmt-D-Cit-Phe-NMeNle-NH2	15	2.92	5.03	>10.000	1/2/>3420	0.65	1.5
H-Dmt-D-Arg-Aba-Lys-NH2	16	0.61	38.6	ND	1/64/-	76.0	64 (IC25)
H-Dmt-D-Arg-Aba-Nle-NH2	17	0.43	7.46	1.12	1/17/3	1.45	3.1
H-Dmt-D-Cit-Aba-Lys-NH2	18	2.62	56.9	ND	1/22/-	49.0	1000
H-Dmt-D-Cit-Aba-Nle-NH2	19	0.93	3.18	11.1	1/3/12	0.78 (IC35)	1.16
H-Dmt-D-Arg-Aba-Gly-NH2e	20	0.15	0.60	118	1/4/787	0.32	0.42

^cValues represent means of 3-6 experiments. The GPI functional assay is representative of MOR activation, whereas the MVD is a DOR receptor-representative assay. Binding affinities of compounds for MOR and DOR opioid receptors were determined by displacement of [³H]DAMGO ([D-Ala²,NMePhe⁴,Gly-ol⁵]enkephalin) and [³H]DSLET ([D-Ser²,Leu⁵]enkephalin-Thr⁶) from rat brain membrane binding sites, respectively. KOR receptor binding affinities were determined by displacement of [³H]U69,593 from guina pig ileum membrane binding sites.

^dData taken from ref.8,

^eData taken from ref.15

The investigated peptides can be subdivided in three series: i) peptides 1 and 9 to 11 in which the side chain charges were altered, ii) the analogous series where the α-amine of the C-terminal residue is methylated (peptides 12 to 15) and iii) the tetrapeptides with restricted side chain flexibility after a Phe³→Aba³ substitution (compounds 16 to 19). Peptide 20 (Dmt-D-Arg-Aba-Gly-NH₂) was reported recently,¹⁹ but is included in this work as a reference compound.

In general, the replacement of D-Arg² in the H-Dmt-D-Arg-Phe-Lys-NH₂ analogues by D-Cit² preserves the low- to subnanomolar MOR binding affinity, produces shifts in κ opioid receptor (KOR) binding, and enhances the DOR affinity in most cases (except for 18 vs. 16). This MOR/DOR selectivity shift is in agreement with the in vitro functional activities at both receptor types, as indicated by the results of the guinea pig ileum (GPI) and mouse vas deferens (MVD) bioassays. Structurally, D-Cit resembles D-Arg but it is not charged at physiological pH. The binding and activation of MOR and DOR confirms that the charge of the guanidinium moiety in D-Arg² is not crucial for receptor recognition, being in line with earlier findings of other opioid peptides such as, for example, the dermorphin, endomorphin and enkephalin peptides that bear respectively a D-Ala, Pro or Gly residue in position 2. Similarly, the side chain charge of the Lys residue in position 4 was removed via replacement with Nle, a residue that lacks the terminal amino group in the side chain. As shown in Table 1, this modification did not markedly affect the interaction with the MOR receptor, but produced a significant 55- (1 vs. 9), 19- (12 vs. 13) and 5- fold (16 vs. 17) improvement of both DOR binding and in vitro functional activity [4- (1 vs. 9), 5- (12 vs. 13) and 21- fold (16 (IC₂₅) vs. 17)]. In this case, the favorable interaction with DOR may be a result of not only the absence of a side chain charge, but also other factors like hydrogen bonding or steric effects (amine function vs. hydrogen) have to be considered. In respect to the KOR affinity, the Lys⁴→Nle⁴ substitution slightly improves binding when D-Arg² is preserved (1 vs. 9 and 12 vs. 13). However, when combined with the D-Arg²→D-Cit² modification, a significant loss of KOR affinity is observed.

Previously, we have shown that constraining the conformation of Phe³ in dermorphin-derived tetrapeptides into a 4-amino-tetrahydro-2-benzazepin-3-one ring 2 leads to highly potent peptides.^18,19 To distinguish the influence of the ring constraint from that of the N-alkylation of the Gly⁴ residue, the “ring opened” analogs containing an amide N-methylation were prepared, and were shown to be of equal potency.²¹ Therefore, the N-methylated Lys⁴ analogues of Dmt¹-DALDA were now also investigated. The N-methylation of the Xxx³-Yyy⁴ amide bond, on the other hand, influences the cis-trans isomer equilibrium of the peptide bond between these amino acids and stabilizes this bond against enzymatic degradation, but also increases the overall lipophilicity of the peptide analogues. Clearly, when comparing the peptides of the first series (peptides 1 and 9 to 11) with the respective peptide analogues of the second series (12 to 15), the receptor binding data were very similar, whereas the functional GPI and MVD assays gave values slightly in disfavor of MOR activation, but in favor of DOR activity.

In contrast to the first eight analogues in Table 1, for which high MOR and DOR receptor binding and potent opioid activity are obtained, the next analogues including the conformationally constrained Aba-Lys dipeptide (16 and 17), presented in vitro activity data with more pronounced shifts. When Aba³-Lys⁴ was inserted in the peptide (16), the MOR affinity was retained, but the DOR affinity increased by a factor 23, in agreement with previous findings.¹⁸ However, the ability to activate MOR and DOR decreased substantially. Because of this drop in activity, we did not perform any further KOR binding studies for compounds 16 and 18. The ability of DOR activation was restored in the Nle⁴ analogue 17 and concomitantly improved KOR binding considerably (IC₅₀ κ 1.12 nM). A combination of D-Cit² and Aba³-Nle⁴ substitutions in 19 resulted in the conversion of the potent MOR selective Dmt¹-DALDA 1 into a potent non-selective MOR partial agonist/DOR full agonist 19.

Previously, the opioid sequence Dmt¹-D-Arg²-Aba³-Gly⁴-NH₂ 20 was identified as a potent antinociceptive opioid peptide.¹⁹ The present data showed that the constrained amino-azepinone scaffold should not be used in opioid tetrapeptides in combination with a Lys⁴ amino acid, but preferably in combination with a Gly⁴ residue, when aiming at combined MOR/DOR agonists.

Brain influx and efflux

Determination of the blood-brain barrier transport properties (initial influx K_in, V_br, brain capillary and parenchyma fraction and efflux k_out) form important parameters to determine the “druggability” of central nervous system (CNS)-compounds. Therefore, the transport properties of the peptides 1, 9, 10, 12, 14 and 20 were determined in CD-1 mice. After labeling with ¹²⁵I, the radiolabeled peptides were injected into the jugular vein and at specified time points after injection, the mice were decapitated and the brain was isolated for analysis (Table 2). These data clearly indicate an influx into the mouse brain for all peptides: peptides 12, 10, 20 and 14 show the highest initial K_in values, indicating a rapid and high influx of the radiolabelled peptides into the brain, while peptides 9 and 1 demonstrate slower influx properties. Moreover, the apparent brain distribution volumes, obtained from 15 min time interval and calculated according to Gjedde,²⁶ indicate relative high values of 27 μl/g, compared to 15 μl/g for the negative control Bovine serum albumin (BSA). Based on the capillary depletion results, a higher parenchyma versus capillary distribution for all peptides is observed: the amount of peptide entering the brain is mostly transferred from the capillaries to the parenchyma. This is in accordance with previous distribution results of other opioid peptides.²⁷ However, despite the good brain influx properties of peptide 20, the net effect is expected to be reduced due to the significant brain to blood efflux transport. Also peptide 1 shows a non-negligible efflux, while this is less observed for peptides 10 and 12, and not at all for peptides 9 and 14.

Table 2.

Blood-brain barrier transport data of the [¹²⁵I]-peptides (mean ± SEM).

Peptide	Kin initial (μl/(g×min))	Vbr (μl/g)	Capillary fraction (%)	Parenchyma fraction (%)	kout (min-1)
1	9.44 ± 0.41	19.25 ± 1.02	23.12 ± 10.26	76.88 ± 10.26	0.058 ± 0.11
9	8.48 ± 1.61	27.43 ± 4.80	32.97 ± 9.99	67.03 ± 9.99	−0.039 ± 0.11
10	14.91 ± 0.10	27.12 ± 3.76	19.39 ± 15.81	80.61 ± 15.81	0.027 ± 0.022
12	18.11 ± 0.78	21.99 ± 2.81	25.95 ± 0.95	74.05 ± 0.95	0.028 ± 0.030
14	11.35 ± 0.61	17.13 ± 2.36	21.84 ± 1.32	78.16 ± 1.32	0.010 ± 0.038
20	14.89 ± 1.34	24.01 ± 3.41	30.95 ± 19.75	69.05 ± 19.75	0.096 ± 0.046

Additionally, and in an effort to predict and potentially rank the ability of the peptides to permeate the BBB via use of an in vitro test, all compounds were subjected to a Caco-2 permeability assay (Supporting information). Very low peptide transport values through the Caco-2 cell monolayers (in both directions, i.e. apical→basal (A-B) and basal→apical (B-A)) were observed. However, a positive observation that was made from this study consisted of a low B-A transport, which is indicative of not hitting p-glycoprotein, an important efflux transporter that is present at high concentration in the luminal membrane of the brain endothelia.

In vivo tissue biodistribution

Immediately after brain dissection, the spleen, liver, lungs, heart and kidneys were collected, the tissues were weighed, and the amount of radioactivity was measured. For each tissue, the percentage of mass-corrected injected dose (ID) was then calculated as described in the experimental section. The tissue distribution plots (Figure 2) show a strikingly high liver accumulation for peptide 9 as a consequence of the Lys⁴ replacement by Nle⁴. Also for peptide 10, a high liver, as well as kidney extraction is observed. This higher kidney uptake is also seen for peptides 1 and 20, despite the structural differences between both peptides. The two peptide analogues containing the amide N-methyl groups (peptides 12 and 14) present a more balanced distribution between the different organs. The serum levels observed 15 min after administration were the highest for peptide 1, the lowest for peptide 9 and intermediate for peptides 10, 12, 14 and 20.

Figure 2.

Mean distribution of the [¹²⁵I]-peptides into the different tissues.

In vitro metabolic stability

The in vitro metabolic stability of unlabeled peptides 1 and 9 was determined in mouse plasma and brain homogenate using UPLC RP C18-chromatography, while the other peptides were investigated using the Prevail Organic Acid column (HPLC). All peptides could undeniably be distinguished from tissue-components, resulting in a reliable data evaluation. The peptides were observed to be very stable in mouse plasma and brain homogenate for at least two hours (Table 4). There is also no adsorption nor chemical degradation observed for the peptides under investigation. The BBB characteristics can thus be assigned to the native peptides. Comparatively, on the contrary, endomorphin-1 and endomorphin-2 for example show very poor half-lives in mouse plasma: a t_1/2 of 4-7 min indicates a rapid enzymatic degradation of these peptides.²⁸

Table 4.

Peptide recovery after 2 hours of in vitro incubation

Peptide	Mean recovery in plasma (%)	Mean recovery in brain (%)
1	94.7	102.6
9	90.2	103.7
10	98.3	96.2
12	112.4	94.0
14	103.9	95.0
20	107.2	99.6

In vivo antinociceptive properties

Given the superior potency of peptides 1 and 10 and the absence of promising in vitro affinity and potency data for the [Aba-Lys]- and [Aba-Nle]-containing peptides (compounds 16 to 19), only the selected [Dmt¹]DALDA analogues 1, 9, 10, 12, 14 and 20 were investigated in vivo for their antinociceptive activity using the tail-flick test in mice, after intraperitoneal (i.p.) and subcutaneous (s.c.) administration. The antinociception, resulting from i.p. injection of the compounds, is shown in Figure 3, with morphine as a positive control. All peptides generated an antinociceptive effect significantly different from control mice receiving i.p. saline. As a reference peptide, the naturally occurring opioid peptide dermorphin (Tyr-D-Ala-Phe-Gly-Tyr-Pro-Ser-NH₂) was included in the study.²⁹ Figure 3 illustrates the inability of dermorphin to induce a significant antinociceptive effect at the applied dose after i.p. injection.

Figure 3.

Antinociceptive characteristics of the [Dmt¹]-DALDA analogues after i.p. administration. Left: Time response curves of the peptides (dose of 1 μmol/kg, n = 10). Morphine and dermorphin at a dose of 1 μmol/kg (n = 10) were used as positive controls. The error bars represent the standard error of the mean (SEM). All data were saline-corrected. Right: Area under the curve normalized to a dose of 1 μmol/kg, based on antinociception measured during 120 min post-injection.

Peptide 9 had a relatively low antinociceptive activity, similar to morphine and dermorphin, at 60 min post i.p. injection. In contrast, the parent peptide 1 as well as the peptide bearing a D-Cit residue in position two (10) produced potent responses. Peptide 1 showed a maximum antinociceptive activity at 60 min post i.p. injection, while the maximum antinociceptive activity of peptide 10 was reached at an earlier time point, i.e. at 45 min. This can be explained by the lower initial K_in value together with the high serum concentration observed at 15 minutes post injection obtained for peptide 1. Peptide 10 showed good initial brain influx properties, resulting in a relatively rapid onset of antinociception. The D-Arg²→D-Cit² substitution changes the polarity of 1 in such a way that brain influx is accelerated (see K_in Table 2). The rapid relapse of antinociception, on the other hand, can be explained by the kidney and liver extraction, seen at 15 min post injection. Both peptides gave responses that were superior to the response of peptide 20. Despite the high initial K_in value of peptide 20, the efflux and kidney uptake characteristics highly influence its analgesic profile, as a slow and shallow antinociceptive response is produced by this peptide. Peptides 12 and 14 showed a rapid increase of antinociception over time, reaching the highest plateau values at 60-90 min post i.p. injection. During the first 45 min after injection, peptide 12 reached potent antinociception at a slower rate than peptide 14. After the maximal effect, i.e. at the plateau decline phase, both peptides still showed a high antinociceptive activity (13 - 15 times higher than the antinociception of morphine). These analgesic profiles can be explained by the high initial brain influx properties (K_in) and insignificant outflux rate (k_out) of both peptides. Taken together, the relatively high serum concentrations after 15 min i.p. and the absence of massive tissue extraction, provide an explanation for the plateauing characteristics. Finally, the disappointing antinociceptive activity of peptide 9 can be understood by the low initial K_in, indicating a slow onset of antinociception, together with the excessive liver extraction results.

When analyzing the dose-normalized area under the curve values (AUCs, 1 μmol/kg) (Figure 3) of the initial 120 min post injection, we could conclude that peptide 14 (1412 %.min) had the highest antinociceptive effect, followed by peptide 12 (1187 %.min), peptide 1 (750 %.min), peptide 10 (521 %.min) and peptide 20 (253 %.min). The positive control morphine had a much lower AUC, i.e. 66 %.min, similar to peptide 9 (65 %.min) and dermorphin (46 %.min).

Looking at the structural characteristics of peptides 1-14, it is clear that the C-terminal amino acid is the most determining factor for the antinociceptive activity: the presence of NMeLys (12 and 14) is preferred, followed by Lys (1 and 10), while other amino acids like Nle (9) decreased its antinociceptive activity significantly. The effect of varying Arg and Cit at position 2 is less pronounced, but cannot be neglected. It shows a confounding effect with the amino acid on position 4: if this is NMeLys, then Cit² gives a higher activity (14 versus 12), while if position 4 is Lys, the reverse is true and Arg² gives the highest activity (1 versus 10).

In addition to the primary evaluation of antinociceptive properties of peptides 1, 9, 10, 12, 14 and 20 after i.p. administration, a further in vivo study assessing the dose- and time-dependent antinociceptive effects after s.c. administration was performed using the mouse tail-flick test. Antinociceptive properties of the six peptides, given s.c., were compared to those of morphine and are illustrated in Figure 4. The calculated ED₅₀ values with 95% confidence intervals are shown in Table 5. The overall profile of antinociceptive effects after s.c. administration closely match that determined after i.p. administration.

Figure 4.

Antinociceptive activities of peptides 1, 9, 10, 12, 14 and 20 and morphine after s.c. administration in the mouse tail-flick test. Dose- and time-dependent antinociceptive response presented as %MPE. Data are shown as mean ± SEM.

Table 5.

Antinociceptive potencies of peptides 1, 9, 10, 12, 14 and 20 in comparison to morphine after s.c. administration in the mouse tail-flick test

Compound	ED50 μmol/kg (95% confidence limits)a	Relative potency to Morphine
1	0.15 (0.07-0.32)	44
9	>6	n.d.
10	0.23 (0.10-0.55)	29
12	0.13 (0.06-0.30)	51
14	0.16 (0.07-0.36)	41
20	0.47 (0.18-1.2)	14
Morphine	6.6 (3.5-12.6)	1

^aED₅₀ was determined at the time of peak effect (i.e. 2 h for peptide 1 and 12, 1h for peptides 10, 14, and 20 and 30 min for morphine). n.d.: not determined.

Dose-dependent and long duration of action lasting up to 6-7 h was shown by peptides 1 (Figure 4A), 12 (Figure 4D) and 20 (Figure 4F), while a shorter antinociceptive effect was produced by peptides 10 (Figure C) and 14 (Figure E) lasting about 2-3 h, still being longer than the effect of morphine. Compound 9 showed no dose-related antinociception, and failed to reach a 50% effect after s.c. administration of 6 μmol/kg (Figure 4B). All peptides, except 9, were considerably more potent in eliciting an analgesic action than morphine (Figure 4 and Table 5).

In line with earlier reports,^11,12,30 [Dmt¹]DALDA 1 is a potent opioid antinociceptive agent when administered to mice s.c. in the tail-flick test, having a 44-fold increase in potency relative to morphine and a prolonged duration of action (Figure 4A). The high potency that characterizes [Dmt¹]DALDA 1 is, similarly to the i.p. administration, completely lost upon deletion of the Lys⁴ side chain amino group resulting in analogue 9 (Figure 4B). On the contrary, compound 10 (Figure 4C), as well as its NMeLys⁴ analogue 14 (Figure 4E), induce a potent effect with a peak at 1h. This observation clearly shows that only the removal of the Lys⁴ side chain functionality is detrimental for antinociception. Following s.c. administration, peptide 12, the N-methylated analogue of [Dmt]¹DALDA 1, produced a prolonged antinociception action with a peak effect around 1-4 h and lasting up to 7 h (Figure 4D), a profile similar to that of the parent peptide. While it was equipotent to [Dmt]¹DALDA, compound 12 was about 51-fold more potent than morphine (Table 5). This finding was also true when comparing peptide 10 and its N-methyl analogue 14, which show comparable potencies and similar time course of antinociceptive action in mice after s.c. administration, and an augmented potency relative to morphine. On this basis, N-methylation of the Lys residue in position 4 did not result in any major alterations in antinociceptive activities when injected s.c., while its replacement with Nle has a detrimental effect. Compound 20 also produced a significant antinociceptive action after s.c. administration with a peak effect at 1h (Figure 4F), but with a somewhat shorter duration of action than peptides 1 and 12. It was 3- and 4-fold less potent in eliciting an antinociceptive effect, respectively. However, compared to morphine, peptide 20 still shows 14-fold increased analgesic potency after s.c. administration. Exchanging the D-Arg² residue in peptides 1 and 12 with a D-Cit residue, resulted in analogues 10 and 14, respectively, which retained high antinociceptive potency, while having a shorter duration of action (Figure 4C and 4E).

Altogether, the new peptidic [Dmt¹]DALDA analogues 10, 12 and 14 emerged as potent antinociceptive agents after i.p. and s.c. administration, with several folds increased potency as compared to morphine. They differ in kinetic behavior, like BBB-characteristics and tissue distribution, resulting in different kinetic profiles of the antinociceptive activity.

Conclusions

From the substitution of the residues of [Dmt¹]-DALDA at position 2 and 4, by structurally related, but uncharged amino acids (i.e. D-Arg² by D-Cit² and Lys⁴ by Nle⁴), we conclude that side chain groups bearing the positive charges in [Dmt¹]DALDA are not needed for efficient in vitro receptor binding and activation. However, the removal of either side chain charges did diminish the well-established MOR subtype receptor selectivity of [Dmt¹]DALDA. Enhanced DOR binding was observed when one charge and even more when the two charges were eliminated. This trend was conserved throughout the ‘parent’ (1 and 9 to 11), the N-methylated (12 to 15) and the constrained “Aba” series (16 to 19). In regard to KOR affinity, the introduced modifications clearly had an impact on the binding affinity of the investigated peptides to this opioid receptor subtype. Low nanomolar KOR binding affinities were determined for peptide analogues 13, 14, 17 and 19 (with K^κ_i values between 1.12 and 14.4 nM). Because of the liabilities related to mu opioid analgesics such as morphine, ligands of KOR have also received considerable attention as potential analgesics.³¹ In particular, KOR agonists have therapeutic potential in pain conditions associated with inflammation.

When looking at the transport properties of the peptide analogues in this study, all synthesized [Dmt¹]-DALDA peptides show influx into the brain parenchyma. The tissue distribution, on the other hand, indicated that the charged aminoalkyl chain of Lys⁴ and the guanidine moiety of D-Arg², are of significant importance for the biodistribution of the peptides. Deletion of these charges, via D-Arg²→D-Cit² and Lys⁴→Nle⁴ substitutions, decreases the 3+ net charge of the peptide, resulting in enhanced liver uptake, and potentially influences brain accessibility. As shown in this work, subtle structural modifications can influence the biodistribution of the peptide analogues and could lead to a modulated peptide delivery to the CNS, thus affecting the ratio of centrally vs. peripherally induced analgesia.

The ability of mice to sense a pain stimulus was verified after peripheral administration of a selection of opioid peptides. In contrast to the reference peptide [Dmt¹]DALDA 1, which showed a very potent analgesic effect, the low antinociceptive effect of peptide 9 could be rationalized by a limited brain intake after massive liver extraction. Peptide 20 showed a rapid brain influx, but also a significant efflux, resulting in lower antinociception compared to the other peptides such as 1, 12 and 14.

Peptides 12 and 14 had the highest antinociceptive effect during the first 2h of an i.p. injection, measured as AUC relative to morphine (normalized at 1 μmol/kg), followed by peptides 1 and 10. The presence of NMeLys⁴ at the carboxyl terminus of the peptide sequence is thus preferred for antinociceptive activity, followed by Lys⁴, while other amino acids like Gly⁴ and Nle⁴ significantly decreased the antinociception.

When the peptide analogues were investigated over a longer s.c. post-injection period (up to 7h) in a dose-dependency study, it became apparent that two distinct antinociception profiles were obtained. Comparable ED₅₀ values were measured for peptides 1 and 12 (Dmt-D-Arg-Phe-(NMe)Lys-NH₂) in comparison with analogues 10 (H*Dmt-D-Cit-Phe-Lys-NH₂) and 14 (Dmt-D-Cit-Phe-(NMe)Lys-NH₂), but the latter peptides produced their peak effect more rapidly (1h vs. 2h post-injection for 1/12). The steep increase in potency was also accompanied by a rapid decrease after reaching the maximum in antinociceptive effect. These observations are interesting for the development of peptide-based pharmaceuticals aiming at different pain types. Peptide 10 could serve as a lead molecule for the treatment of acute (e.g. post-operative pain), whereas the long-lasting Dmt-D-Arg-Phe-NMeLys-NH₂ 12 could be useful for addressing chronic pain conditions (e.g. neuropathic pain). As the compounds investigated in this study are potent dual MOR/DOR (and even KOR, see 17) agonists, they could potentially be useful for the treatment of pain with reduced side effects and be effective at low administered doses.

Experimental section

General

Thin layer chromatography (TLC) was performed on plastic sheet pre-coated with silica gel 60F₂₅₄ (Merck, Darmstadt, Germany) using specified solvent systems. Mass Spectrometry (MS) was recorded on a Micromass Q-Tof Micro spectrometer using electrospray (ESP) ionization (positive or negative ion mode). Data collection was done with Masslynx software. Analytical Reversed Phase (RP)-HPLC was performed using an Agilent 1100 Series system (Waldbronn, Germany) with a Supelco Discovery BIO Wide Pore® (Bellefonte, PA, USA) RP C-18 column (25 cm × 4.6 mm, 5 μm) using UV detection at 215 nm. The mobile phase (water/acetonitrile) contained 0.1% TFA. The gradient consisted of a 20 min run from 3 to 97% acetonitrile at a flow rate of 1mL/min. Preparative HPLC was performed on a Gilson apparatus and controlled with the software package Unipoint. The RP C18-column (Discovery BIO Wide Pore 25 cm × 21.2 mm, 10 μm) was used under the same conditions as the analytical RP-HPLC, but with a flow rate of 20 mL min^-1. A purity of more than 95% was determined for all compounds by analytical RP-HPLC using the conditions described above. ¹H-NMR and ¹³C-NMR spectra were recorded at 250 MHz and 63 MHz, respectively, on a Bruker Avance 250 spectrometer or at 500 MHz and 125 MHz on a Bruker Avance II 500. Calibration was done with TMS (tetramethylsilane) or residual solvent signals as an internal standard. The solvent used is mentioned in all cases and the abbreviations used are: s (singlet), d (doublet), dd (double doublet), t (triplet), br s (broad singlet) and m (multiplet).

Opioid receptor binding assays

Opioid receptor binding studies were performed as described in detail elsewhere.³⁵ Binding affinities for MOR and DOR were determined by displacing, respectively, [³H]DAMGO (Multiple Peptide Systems, San Diego, CA) and [³H]DSLET (Multiple Peptide Systems) from rat brain membrane binding sites, and KOR receptor affinities were measured by displacement of [³H]U69,593 (Amersham) from guinea pig brain membrane binding sites. Incubations were performed for 2 h at 0 °C with [³H]DAMGO, [³H]DSLET and [³H]U69,593 at respective concentrations of 0.72, 0.78 and 0.80 nM. IC₅₀ values were determined form log-dose displacement curves, and K_i values were calculated from the IC₅₀ values by means of the equation of Cheng and Prusoff,³² using values of 1.3, 2.6 and 2.9 nM for the dissociation constants of [³H]DAMGO, [³H]DSLET and [³H]U69,593, respectively.

Functional GPI and MVD assays

The guinea pig ileum (GPI)³³ and mouse vas deferens (MVD)³⁴ bioassays were carried out as described in detail elsewhere.^35,36 A dose-response curve was determined with [Leu⁵]enkephalin as standard for each ileum and vas preparation and IC₅₀ values of the compounds being tested were normalized according to a published procedure.³⁷

Blood-brain barrier transport

Animals

Male Institute for Cancer Research, Caesarean Derived-1 (ICR-CD-1) mice (Harlan Laboratories, Venray, The Netherlands), weighing 25-30 g, were used according to the Ethical Committee principles of laboratory animal welfare as approved by our institute (Ghent University, Faculty of Veterinary Medicine, 2009-052). Animals were housed under standard laboratory conditions in a temperature and humidity regulated environment with food and water ad libitum.

Reagents and peptides

Bovine serum albumin (BSA), calcium chloride dehydrate, D-glucose, HEPES, magnesium sulphate heptahydrate, potassium chloride, disodium hydrogen phosphate dihydrate, sodium dihydrogen phosphate monohydrate, sodium chloride, sodium lactate, urethane, and Krebs-Henseleit buffer were all purchased from Sigma (St. Louis, MO, USA) or Merck KGaA (Darmstadt, Germany). Dextran was obtained from AppliChem GmbH (Darmstadt, Germany), while dermorphin was purchased from Peptide Synthetics (PPR, Hampshire, United Kingdom). Morphine hydrochloride was acquired from Belgopia (Louvain-La-Neuve, Belgium).

Peptides ¹²⁵I radiolabeling

The Dmt¹-DALDA peptides and controls (dermorphin and BSA) were labeled by the Iodogen method.^37,39 Briefly, a 1 g/l peptide solution was prepared by dissolving about 1 mg in phosphate buffer (pH 7.4, 130 mM). A Iodo-Gen® coated tube was previously rinsed with 1 ml phosphate buffer. Subsequently, 50 μl NaI (20 nmol/10 μl) and 10 μl of Na¹²⁵I solution (3700 MBq/ml) (Perkin Elmer, Boston, MA, USA) were transferred into this Iodo-Gen® coated tube. After 6 min of incubation at room temperature, the oxidation reaction of iodide to iodonium was stopped by removing the solution from the tube and adding the iodonium solution to a tube containing 50 μl of peptide solution (1 μg/Kl). The iodination reaction of the peptide was allowed to proceed another 6 min at room temperature, after which the reaction mixture was purified with an Ag-filter. After Ag-purification, the radioactivity of the filtrate was measured using a dose calibrator (Veenstra, The Netherlands) to dilute the ¹²⁵I-peptide appropriately.

In addition, a quality control was performed on the filtrated peptide using a Vydac Everest C18 (250 mm × 4.6 mm I.D., 300 Å, 5 μm particle size) HPLC column (Grace Vydac, Hesperia, CA, USA) in an oven set at 30°C, with a mobile phase consisting of A) 0.1 % w/v formic acid in water, and B) 0.1 % w/v formic acid in acetonitrile. A linear gradient was employed as follows:

Peptide 1: 0-50 min going from 98 % to 50 % A.

Peptides 9, 12, 14 and 20: 0-40 min going from 98 % to 80 % A.

Peptide 10: 0-40 min going from 98 % to 60 % A.

The flow rate was set at 1.0 ml/min with an injection volume of 100 μl.

The radio-HPLC apparatus consisted of a LaChrom Elite L-2130 pump with degasser, a LaChrom Elite L-2300 column oven, a LaChrom Elite L-2400 UV detector set at 215 nm (all Hitachi, Tokyo, Japan), a Rheodyne 7725i manual injector with 100 μl sample loop (Rheodyne, Rohnert Park, CA, USA) and a Berthold LB500 HERM radioactivity detector (Berthold Technologies, Bad Wildbad, Germany) equipped with EZChrom Elite version 3.1.7 software for data acquisition (Scientific Software, Pleasanton, CA, USA).

Multiple time regression (MTR)

In order to determine whether the peptides could enter the brain, multiple time regression analysis was performed. The CD-1 mice were anesthetized with intraperitoneal (i.p.) urethane (3 g/kg) and the jugular vein and carotid artery were isolated. 200 μl of the radiolabeled peptides (30 000 cpm/μl) in Lactated Ringer’s (LR) solution containing 1 % BSA were each injected separately into the jugular vein. At specified time points after injection, blood was obtained from the carotid artery and the mouse was decapitated. The brain was removed and weighed, after which radioactivity was determined in a gamma counter (Wallac Wizard automatic gamma counter, PerkinElmer, Shelton, CT, USA). The blood was centrifuged at 10 000 × g for 15 min at 21 °C and the radioactivity of the obtained serum was measured.

The theoretical background of the multiple time regression analysis is based on the Gjedde equation:²⁶

$$\frac{A_m\left(t\right)}{C_p\left(t\right)}=K_{\mathit{in}}\frac{{\int }_0^t{C}_p\left(t\right)\cdot \mathit{dt}}{C_p\left(t\right)}+V_i$$

where A_m(t) is the amount of radioactivity in brain at time t, C_p(t) the amount of radioactivity in serum at time t, K_in the brain influx rate constant and V_i the initial brain distribution volume.

The exposure time (exp.t) is defined as the integral of the serum radioactivity over time divided by the radioactivity at time t: $\mathit{exp}.t=\frac{{\int }_o^t{C}_p\left(t\right)\cdot \mathit{dt}}{C_p\left(t\right)}$ The integral of the radioactivity over time is represented by the area under the curve (AUC).

Finally, the brain/serum ratios (μl/g) were plotted versus exposure time. The initial slope of the linear portion of this relationship over the first minutes measures the unidirectional initial influx rate (K_in initial) from blood to brain. The intercept represents the brain volume of distribution (V_br).

As a system suitability controls of this method, BSA, the impermeable vascular marker, as well as dermorphin, a weak positive influx control peptide,²⁸ were labeled with I-125 and their influx investigated as described above. In compliance with literature data,⁴⁰ a mean influx transfer constant of 0.0849 ± 0.0847 μl/(g × min) was obtained for BSA, confirming negligible influx. The mean K_in over 15 minutes for dermorphin was 0.173 ± 0.137 μl/(g × min), while the initial K_in was 5.52 ± 1.41 μl/(g × min) if the first minutes are considered, justifying its choice as a weak positive brain influx control.

Capillary depletion

This method was performed to determine to what extent the peptides, taken up by the brain, crossed the capillary wall instead of being trapped in the capillary cells. The method of Triguero et al.⁴¹ as modified by Gutierrez et al.⁴² was used. Summarizing, CD-1 mice were anesthetized with urethane (3 g/kg i.p.) and the jugular vein was isolated. 200 μl of the diluted peptide solution (10 000 cpm/μl) was injected in the jugular vein. At specified time points after injection, blood was collected from the abdominal aorta and the brain was perfused manually with 20 ml of LR buffer after severing the jugular veins. Subsequently, the brain was collected, homogenized with 0.7 ml of ice-cold capillary buffer (10 mM HEPES, 141 mM NaCl, 4 mM KCl, 2.8 mM CaCl₂, 1 mM MgSO₄, 1 mM NaH₂PO₄ and 10 mM D-glucose adjusted to pH 7.4) in a pyrex glass tube and mixed with 1.7 ml of 26 % ice-cold dextran solution in capillary buffer. The resulting solution was weighed and centrifuged in a swinging bucket rotor (Eppendorf, Hamburg, Germany) at 5400 × g for 30 min at 4 C. Pellet (capillaries) and supernatant (parenchyma and fat tissues) were collected, weighed and measured in a gamma counter. After centrifuging the obtained blood, 50 μl of serum was collected and measured in a gamma counter. Compartmental BBB-distribution was evaluated from plotting capillaries/serum and parenchyma/serum activity ratios versus time.

Brain to blood transport

This method was performed to quantify the amount of peptide pumped out of the brain by efflux transport.⁴³ CD-1 mice were anesthetized with urethane (3 g/kg i.p.) and the skin of the skull was removed. A hole was then made into the lateral ventricle, 1 mm lateral and 0.34 mm posterior to the bregma, with a 22 G needle carrying a tubing guard at a constant depth of 2 mm. The anesthetized mice received an intracerebroventricular (i.c.v.) injection of 1 μl of LR solution containing the labeled peptide (25 000 cpm/μl) and 1 % BSA at a speed of 360 μl/h for 10 sec using a syringe pump (KDS100, KR analytical, Cheshire, UK). At specified time points after injection, blood was collected from the abdominal aorta and the mouse was decapitated. Subsequently, the whole brain was collected, weighed and measured in a gamma counter. The efflux rate constant k_out (min⁻¹) was calculated from the linear regression of the natural logarithm of the residual radioactivity in brain versus time, applying first order kinetics.

Tissue distribution

Immediately after brain collection, spleen, liver, lungs, heart and kidneys were collected.⁴⁴ Again, the tissues were weighed and measured in the gamma counter. For each tissue, the percentage of time-corrected injected dose (ID) was then calculated:

$$\text{Percentage}\mathit{ID}(\%)=\frac{\frac{A_{\text{tissue}}}{W_{\text{tissue}}}}{\frac{A_{\text{does injected}}}{W_{\text{animal}}}}\times 100$$

where A_tissue is the radioactivity of the tissue, w_tissue the weight of the tissue, A_{dose injected} the amount of radioactivity injected into the vein and w_animal the body weight of the mouse.

In vitro metabolic stability

In order to obtain information regarding the metabolic fate of peptides, the non-labeled peptides were tested in vitro by incubating in mouse plasma and mouse brain homogenate.⁴⁵ In brief, 100 μg of the non-labeled peptide dissolved in Krebs-Henseleit buffer (pH 7.4) was incubated in plasma or brain homogenate (250 μl of mouse plasma or 150 μg of tissue protein, respectively) at 37°C while shaking. At suitable time points, aliquots were taken and immediately transferred into microtubes containing 10 % v/v trifluoric acetic acid (TFA) solution in water to obtain a final concentration of 1:10 v/v solvent/aliquot. The samples were then heated for 5 min at 95°C, followed by cooling in an ice-bath for 30 min. After centrifugation of the samples at 15000 × g for 15 min at 5 °C, the supernatants were analyzed using UPLC-PDA and HPLC-UV. Placebos (mouse plasma or tissue homogenate), controls (initially inactivated tissue proteases) and reference solutions (peptide dissolved in buffer) were similarly processed. The time-dependent decrease of the native peptide was quantified and from the linear regression between the quantified decrease and time, the half-life (t_1/2) was determined as:

$$t_{\frac{1}{2}}=\frac{\ln \left(2\right)}{k}$$

where k is the first order rate constant.

If the calculated half-life was more than two times the experimental time, i.e. more than 4 hours, no quantitative half-life was given and the results expressed as “stable”.

UPLC-PDA analyses were performed using Waters Acquity H-Class Bio-samples FTN, a Waters Acquity H-Class BioQuaternary Solvent Manager, a Waters Acquity H-Class column module and a Waters Acquity H-Class Photodiode Array Detector (UV spectrum was recorded between 190 and 400 nm) equipped with Waters Empower Pro software version 2 (all Waters, Milford, MA, USA). Analysis was performed using a Waters Acquity UPLC BEH300 C18 column (100 × 2.1 mm I.D., 300 Å, 1.7 μm particle size) in an oven set at 35 °C. The following solvent system was used: A) 10 mM ammonium carbonate and B) acetonitrile. The gradient method consisted of a linear increase from 5 % B to 60 % B in 20 min, followed by 10 min reconditioning with initial composition 95/5 v/v A/B. The flow rate was set at 0.58 ml/min and the injection volume was 5 μl.

The degradation of the peptides was also studied by use of an HPLC system consisting of a Waters Alliance 2695 Separations module, a Waters column heater and a Waters 2487 Dual wavelength absorbance detector, fitted with Empower Pro software version 2 (all Waters, Milford, MA, USA). Quantification of the degradation was performed at 210 nm. Analysis was achieved using a Grace Alltech Prevail Organic Acid column (250 × 4.6 mm I.D., 110 Å, 5 μm particle size) in an oven set at 30 °C, using A) 25 mM phosphate buffer pH 8.0 and B) acetonitrile. The following linear gradient method was used: from 10% B to 70% B in 45 min, followed by 15 min with the initial composition (90/10 v/v A/B) to recondition the system. The flow rate was set at 1.0 ml/min with an injection volume of 20 μl.

Antinociceptive properties

A. After i.p. administration

The same type of mice as used for the BBB transport studies were used in this assay. Antinociception was assessed using the 55 °C tail withdrawal method in CD-1 mice. Each mouse (10 mice per group) was first tested for consistency in baseline latency by immersing its tail in the water and recording the time to response. Mice were then administered the controls as well as the test compounds at a dose of 5 mg/kg body weight and tested for antinociception at various time points afterwards. Antinociception was calculated as % antinociception or % MPE (percent maximum possible effect) using the following formula: 100 × (test latency – baseline latency) / (10 - baseline latency). To avoid tissue damage, a maximum score was assigned (100%) to animals that failed to respond within 10 s (=cut-off time).

A negative control consisted of a physiological 0.9 % saline solution. Two positive controls were also included, i.e. morphine and dermorphin, at a dose of 5 mg/kg. All controls were i.p. administered.

The amount of antinociception was determined using the AUC:

$$\mathit{AUC}=\sum _{i=1}^n\left(\frac{(y_i+y_{i-1})}{2}\times (x_i-x_{i-1})\right)$$

within the 120 min time points, y the percentage of antinociception and x the time after injection (min). Assuming first order kinetics, the AUC of the peptides and of morphine is converted to a theoretical AUC_p, respectively AUC_m, equivalent to a dose of 1 μmol active moiety/kg. The morphine normalized AUC for the peptides (AUC_normalized) is then calculated by:

$${\mathit{AUC}}_{\text{normalized}}=\left(\frac{{\mathit{AUC}}_p}{{\mathit{AUC}}_m}\right)\times 100$$

B. After s.c. administration

Male CD-1 mice (25-35 g) were purchased from Charles River Laboratories (Sulzfeld, Germany). Animals were housed under standard laboratory conditions in a temperature-regulated environment under a controlled 12 h light/dark cycle with food and water ad libitum. All animal procedures were approved by the Austrian Ethical Committee on Animal Care and Use in line with international laws and policies.Morphine hydrochloride was obtained from Gatt-Koller GmbH (Innsbruck, Austria). All drugs were dissolved in sterile physiological saline (0.9%) and were administered s.c. in a volume of 10 μl per 1 g body weight.

The radiant heat tail-flick test was used to assess antinociceptive effects of tested drugs after s.c. administration in mice using an UB 37360 Ugo Basile analgesiometer (Ugo Basile s.r.l., Varese, Italy) as described.⁴⁶ The reaction time required by the mouse to remove its tail due to the radiant heat was measured and defined as the tail-flick latency. A cut-off time of 10 s was also used in order to minimize tissue damage, and the % antinociception or MPE for each dose administered was calculated as above. At least three doses were tested and five to six animals per dose were used. Antinociceptive ED₅₀ values and 95% confidence limits were calculated by the method of Litchfield and Wilcoxon (1949).⁴⁷

Compound Synthesis and characterization

Both Fmoc-Aba-Lys-OH 8a and Fmoc-Aba-Nle-OH 8b were prepared according to reported procedures (see supporting information).^23,24 Subsequently, all peptides were assembled as mentioned before by standard SPPS protocols employing Fmoc- or Boc-protected amino acid residues or dipeptides (8a and 8b).

Peptide characterization

H-Dmt-D-Arg-Phe-Lys-NH₂ 1

HPLC (standard gradient) : t_ret = 8.46 min. TLC R_f (EBAW) : 0.50 ; TLC R_f (BAW) : 0.24 ; ESI-HRMS [M+H⁺] : m/z = 640.3950 (calculated for C₃₂H₄₈H⁺N₉O₅ : 640.3929)

H-Dmt-D-Arg-Phe-Nle-NH₂ 9

HPLC (standard gradient) : t_ret = 10.57 min. TLC R_f (EBAW) : 0.62 ; TLC R_f (BAW) : 0.70 ; ESI-HRMS [M+H⁺] : m/z = 625.3793 (calculated for C₃₂H₄₇H⁺N₈O₅ : 625.3820)

H-Dmt-D-Cit-Phe-Lys-NH₂ 10

HPLC (standard gradient) : t_ret = 8.68 min. TLC R_f (EBAW) : 0.52 ; TLC R_f (BAW) : 0.35 ; ESI-HRMS [M+H⁺] : m/z = 641.3759 (calculated for C₃₂H₄₇H⁺N₈O₆ : 641.3770)

H-Dmt-D-Cit-Phe-Nle-NH₂ 11

HPLC (standard gradient) : t_ret = 11.06 min. TLC R_f (EBAW) : 0.72 ; TLC R_f (BAW) : 0.74 ; ESI-HRMS [M+H⁺] : m/z = 626.3651 (calculated for C₃₂H₄₆H⁺N₇O₆ : 626.3661)

H-Dmt-D-Arg-Phe-NMeLys-NH₂ 12

HPLC (standard gradient) : t_ret = 8.43 min. TLC R_f (EBAW) : 0.52 ; TLC R_f (BAW) : 0.24 ; ESI-HRMS [M+H⁺] : m/z = 654.4064 (calculated for C₃₃H₅₀H⁺N₉O₅ : 654.4086)

H-Dmt-D-Arg-Phe-NMeNle-NH₂ 13

HPLC (standard gradient) : t_ret = 10.67 min. TLC R_f (EBAW) : 0.62 ; TLC R_f (BAW) : 0.70 ; ESI-HRMS [M+H⁺] : m/z = 639.3983 (calculated for C₃₃H₄₉H⁺N₈O₅ : 639.3977)

H-Dmt-D-Cit-Phe-NMeLys-NH₂ 14

HPLC (standard gradient) : t_ret = 8.67 min. TLC R_f (EBAW) : 0.55 ; TLC R_f (BAW) : 0.37 ; ESI-HRMS [M+H⁺] : m/z = 655.3915 (calculated for C₃₃H₄₉H⁺N₈O₆ : 655.3926)

H-Dmt-D-Cit-Phe-NMeNle-NH₂ 15

HPLC (standard gradient) : t_ret = 11.81 min. TLC R_f (EBAW) : 0.72 ; TLC R_f (BAW) : 0.74 ; ESI-HRMS [M+H⁺] : m/z = 640.3823 (calculated for C₃₃H₄₈H⁺N₇O₆ : 640.3817)

H-Dmt-D-Arg-Aba-Lys-NH₂ 16

HPLC (standard gradient) : t_ret = 8.21 min. TLC R_f (EBAW) : 0.52 ; TLC R_f (BAW) : 0.24 ; ESI-HRMS [M+H⁺] : m/z = 653.3913 (calculated for C₃₃H₄₈H⁺N₉O₅ : 652.3929)

H-Dmt-D-Arg-Aba-Nle-NH₂ 17

HPLC (standard gradient) : t_ret = 11.25 min. TLC R_f (EBAW) : 0.75 ; TLC R_f (BAW) : 0.74 ; ESI-HRMS [M+H⁺] : m/z = 637.3826 (calculated for C₃₃H₄₇H⁺N₈O₅ : 637.3820)

H-Dmt-D-Cit-Aba-Lys-NH₂ 18

HPLC (standard gradient) : t_ret = 8.43 min. TLC R_f (EBAW) : 0.55 ; TLC R_f (BAW) : 0.32 ; ESI-HRMS [M+H⁺] : m/z = 653.3794 (calculated for C₃₃H₄₇H⁺N₈O₆ : 653.3770)

H-Dmt-D-Cit-Aba-Nle-NH₂ 19

HPLC (standard gradient) : t_ret = 11.57 min. TLC R_f (EBAW) : 0.77 ; TLC R_f (BAW) : 0.74 ; ESI-HRMS [M+H⁺] : m/z = 638.3659 (calculated for C₃₃H₄₆H⁺N₇O₆ : 638.3661)

H-Dmt-D-Arg-Aba-Gly-NH₂ 20

HPLC (standard gradient) : t_ret = 8.63 min. TLC R_f (EBAW) : 0.55 ; TLC R_f (BAW) : 0.45 ; ESI-HRMS [M+H⁺] : m/z = 581.3220 (calculated for C₂₉H₃₉H⁺N₈O₅ : 581.3195)

Abbreviations used

Aba

4-amino-1,2,4,5-tetrahydro-2-benzazepin-3-one

AUC

area under the curve value

BBB

blood-brain barrier

BSA

Bovine Serum Albumin

Cit

citrulline

CNS

central nervous system

DAMGO

[D-Ala²,NMePhe⁴,Gly-ol⁵]enkephalin

Dmt

2′,6′-dimethyl-(S)-tyrosine

DOR

δ–opioid receptor

DSLET

[D-Ser²,Leu⁵]enkephalin-Thr⁶

ESP

electrospray ionization mass-spectrometry

GPI

Guinea pig ileum

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

HOBt

1-hydroxybenzotriazole

ID

injected dose

i.p.

intraperitoneal

i.v.

intravenous

KOR

κ–opioid receptor

LR

Lactated Ringer’s (solution)

k_out

brain efflux rate coefficient

MOR

μ–opioid receptor

MPE

maximal possible effect

MTR

Multiple time regression

MVD

mouse vas deferens

Nle

norleucine

NMM

N-methyl morpholine

K_in

unidirectional brain influx or uptake transfer constant

Pbf

2,2,4,6,7-pentamethyldihydrobenzofuran-5-sulfonyl

Phth

phthaloyl

RaNi

Raney Nickel

SAR

structure-activity relationship

s.c.

subcutaneous

SEM

standard error of the mean

TBTU

O-(Benzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium tetrafluoroborate

V_br

brain distribution volume

Z

benzyloxycarbonyl.