Structure- Activity Relationship Study and Function-Based Petidomimetic Design of Human Opiorphin with Improved Bioavailability Property and Unaltered Analgesic Activity

Opiorphin QRFSR-peptide is an endogenous human regulator that was discovered using a functional biochemical approach [1,2]. Its characterization demonstrated that it is an authentic physiological dual inhibitor of Zn-dependent metallo-ectopeptidases, neutral endopeptidase (NEP EC3.4.21.11) and aminopeptidase N (AP-N EC3.4.11.2). These enzymes are implicated in the rapid inactivation of endogenous circulating opioid agonists, namely the enkephalins. As a consequence, opiorphin improves the specific binding and affinity of enkephalin-related peptides to membrane opioid receptors [3]. The enkephalin neuropeptides play key roles in the control of nociceptive transmission and in the modulation of the activity of cerebral structures governing the motivation and the adaptive balance of emotional states [4-8]. By increasing the half-life of circulating enkephalins, opiorphin, at systemically or centrally active doses (1-2 mg/kg I.V. or 5-10 μg/kg I.C.V.), produces analgesia in various murine models of pain [1,9,10]. At equivalent doses, opiorphin also exerts antidepressant-like effects in the standard model of depression, the forced swim test [11,12]. All opiorphin-induced effects are specifically mediated via endogenous enkephalin-related activation of μ and/or δ opioid pathways.


Introduction
Opiorphin QRFSR-peptide is an endogenous human regulator that was discovered using a functional biochemical approach [1,2]. Its characterization demonstrated that it is an authentic physiological dual inhibitor of Zn-dependent metallo-ectopeptidases, neutral endopeptidase (NEP EC3. 4.21.11) and aminopeptidase N (AP-N EC3. 4.11.2). These enzymes are implicated in the rapid inactivation of endogenous circulating opioid agonists, namely the enkephalins. As a consequence, opiorphin improves the specific binding and affinity of enkephalin-related peptides to membrane opioid receptors [3]. The enkephalin neuropeptides play key roles in the control of nociceptive transmission and in the modulation of the activity of cerebral structures governing the motivation and the adaptive balance of emotional states [4][5][6][7][8]. By increasing the half-life of circulating enkephalins, opiorphin, at systemically or centrally active doses (1-2 mg/kg I.V. or 5-10 µg/kg I.C.V.), produces analgesia in various murine models of pain [1,9,10]. At equivalent doses, opiorphin also exerts antidepressant-like effects in the standard model of depression, the forced swim test [11,12]. All opiorphin-induced effects are specifically mediated via endogenous enkephalin-related activation of µ and/or δ opioid pathways.
The discovery of opiorphin is the first demonstration of the existence of a physiological regulator of enkephalin bioavailability in humans. As an upstream modulator of opioid pathways in humans, it is thus of major interest from a therapeutic points of view. Indeed, endogenous human opiorphin appears to intervene in the process of adaptation mediated by enkephalins that is associated with nociception. As a consequence, opiorphin is a promising template for the design of a new class of drug-candidates able to efficiently alleviate a number of severe and chronic pain syndromes, without morphine side effects. The actions of opiorphin could be induced at a specific opioid receptor restricted pathway dynamically stimulated by natural effectors, such as enkephalins, that are recruited according to the nature, duration and intensity of the stimulus. This mechanism of action avoids excessive stimulation of ubiquitously distributed opioid receptors and prevents serious side effects such as respiratory depression, sedation, constipation, physical and psychic dependence and tolerance that have been reported in the case of µ-opioid agonists. We previously demonstrated that opiorphin subchronic intake does not develop significant abuse liability or antinociceptive drug tolerance. In addition, anti-peristalsis is not observed [10].
Measurement of Ectopeptidase Activities using 96-well fluorimetric assays: Under conditions of initial velocity measurement (steady state), hydrolysis of substrates was measured by real-time monitoring of their metabolism rate by the respective recombinant and membrane-bound peptidases, in the presence and absence of tested inhibitory compound (concentrations ranging from 0.01 to 100 µM).
Measurement of NEP-endopeptidase activity using FRET specific peptide-substrate, Abz-dR-G-L-EDDnp: Using the black half-area 96 well micro-plate, the standard reaction consisted of enzyme (12 ng) in 100 mM Tris-HCl pH 7 containing 200 mM NaCl and 0.05% Brij 35 (100 µl final volume). The substrate (15 µM final concentration) was added after preincubation for 10 min at 28°C and the kinetics of appearance of the fluorescent signal (RFU) was directly analyzed for 20-40 min at 28°C (2 to 3 min interval successive measures) by using a fluorimeter micro-plate reader (monochromator Infinite 200-Tecan) at 320 nm and 420 nm excitation and emission wavelengths, respectively.
Measurement of NEP-Carboxy DiPept idase activity using FRET specific peptide-substrate Abz-R-G-F-K-DnpOH: Using the black halfarea 96 well microplate, the standard reaction consisted of enzyme (2.5 ng) in 100 mM Tris-HCl pH 6.5 containing 50 mM NaCl and 0.05% Brij 35 (100 µl final volume). The substrate (4 µM final concentration) was added after pre incubation for 10 min and the kinetics of appearance of the fluorescent signal (RFU) was directly analyzed for 20-40 min at 28°C (2 to 3 min-interval successive measures) using the fluorimeter reader at 320 nm excitation and 420 nm emission wavelengths.
In addition, the intra-molecularly quenched fluorogenic peptide, Mca-BK2 (2.5 µM final concentration), was submitted to hydrolysis by 2 ng rhNEP under the same experimental conditions as those described above. Under these conditions the hNEP-enzyme acted upon Mca-R-P-P-G-F-S-A-F-K-(Dnp)-OH as a CarboxyDiPeptidase preferentially cleaving the A-F bond but also as an EndoPeptidase cleaving the G-F bond.
To measure ECE1-ectopeptidase activity, the same protocol described previously was applied except that Mca-BK2 substrate was used at 7.5 µM final concentration and rhECE1 at 5 ng final concentration.
Measurement of DPPIV activity using FRET specific peptidesubstrate G-P-7amido-4-Mca: Using the black half-area 96 well microplate, the standard reaction consisted of enzyme (7 ng) in 100 mM Tris-HCl pH 8 (100 µl final volume). The substrate (5 µM final concentration) was added after preincubation for 10 min and the kinetics of appearance of the fluorescent signal (RFU) was directly analyzed for 20-40 min at 28°C (2.3 min-interval successive measures) using the fluorimeter reader at 380 nm excitation and 460 nm emission wavelengths.
Measurement of AP-N-ectopeptidase activity using Ala-AMC substrate: Using the black half-area 96 well microplate the standard is their rapid degradation by circulating peptidases and the limited permeation of peptides across biological barriers. In order to search for functional derivatives of opiorphin endowed with improved half-life stability and bioavailability properties a Structure-Activity Relationship (SAR) study on opiorphin was first carried out. Then opiorphin analogs were tested for their inhibitory potency against the two membrane-anchored human ectoenzymes, NEP that has both endopeptidase and carboxydipeptidase activities and AP-N, by using selective fluorescence-based enzyme in vitro assays [13][14][15]. Comparative degradation kinetics were done using experimental in vitro systems to evaluate metabolic half-life in human plasma, which is a reliable prediction model for in vivo stability. Metabolic stability parameters in human liver microsomes were also determined.
The final aim of the research described here was to design and analyze functional analogs of opiorphin that display in vivo bioavailability properties superior to the native peptide, in particular, an increase in circulating peptidase resistance and in permeation across epithelial and endo-epithelial membrane barriers, without affecting in vitro and in vivo biological properties, namely, selective inhibition of human NEP and AP-N ectoenkephalinases and potent inhibition of pain behavioral responses in rat model.

Chemicals
All peptides, human opiorphin and opiorphin derivatives, were synthesized by Genosphere Biotechnologies (Paris-France). Analytical RP-HPLC and electrospray MS confirmed the purity (≥ 95%) and molecular mass of the synthesized peptides.

FRET-based Enzyme In Vitro assays
Formal kinetic analysis was performed for each assay using realtime fluorescence monitoring of specific substrate hydrolysis.

Sources of the human ectopeptidases:
Human recombinant NEP and human recombinant AP-N (devoid of their respective N-terminal cytosol and transmembrane segment) were purchased from R&D Systems (France) and used as a pure source of peptidases. Membraneanchored NEP and AP-N expressed by a human cell line in culture (serum-free medium), namely, LNCaP epithelial prostate cells, were also used as a source of native ectopeptidases. Human recombinant DPPIV (dipeptidyl amino peptidase IV) and human recombinant ECE-1 (endothelin converting enzyme), purchased from R&D Systems, were used to assess compound specificity.
FRET is the distance-dependent transfer of energy from a donor fluorophore (Abz=ortho-aminobenzoyl or Mca=7-methoxycoumarin-reaction consisted of enzyme (3.5 ng) in 100mM Tris-HCl pH 7.0 (100 µl final volume). The Ala-AMC substrate (20 µM final concentrations) was added after preincubation for 10 min at 28°C and the kinetics of appearance of the signal was monitored for 20-40 min at 28°C using the fluorimeter reader at 380 nm excitation and 460 nm emission wavelengths.
Measurement of membrane human NEP-endopeptidase activity using tritiated substance P substrate: the method used was previously described in Wisner et al. and Rougeot et al. [1,16].
The background rate of substrate autolysis, representing the fluorescent signal obtained in the absence of enzyme, was subtracted to calculate the initial velocities in RFU (Relative Fluorescent Unit)/ min. Data were analyzed using Magellan 6.0 software to evaluate initial velocities and with Excel Microsoft software. IC 50 estimates were obtained from a sigmoidal curve fit to a plot of % inhibitory activity versus log inhibitor concentration, using Prism software. For each curve, inhibitors were tested across a range of concentrations differing in half log unit increments.
In vitro pharmacokinetic and metabolic studies: Human blood was collected in pre-chilled tubes containing 1% sodium citrate (buffered at pH 7) and kept at 4°C. The plasma was collected after centrifugation at 400 × g for 30 min at 4°C, then aliquoted and stored at -80°C.
Peptide solutions were extemporaneously prepared in order to add the appropriate concentration of peptidein a volume of 10 µl, thus avoiding dilution of the plasma. The plasma peptide solutions were then mixed and incubated in a shaking water bath at 37°C with a continuous and slight shaking for the preset kinetic time period. The reaction was stopped by cooling the tubes simultaneously in ice and by the addition of 0.1N final concentration of HCl. For opiorphin PK experiments, a mixture of 1 µg or 40 µg QRFSR-peptide, containing 100 or 500.10 3 cpm QR [ 3 H-F] SR (3.6 Ci/mmole, CEA-Saclay), was used. Controls, in which protease free human plasma (Methanol/TFA extract) is substituted for fresh plasma, were included. For certain experiments, different inhibitors of plasma peptidases were added immediately prior to the addition of opiorphin-peptide, bestatin, an inhibitor of aminopeptidases or GEMSA and an inhibitor of carboxypeptidase B.
All samples were stored at -80°C until subjected to Sep-Pak extraction and RP-HPLC chromatography.
C18 Solid-phase extraction: Acidified (HCl 0.1N final concentration) and clarified biological samples were applied to C18-SepPak cartridges (Waters, France) preconditioned with three successive cycles of methanol (Lichrosolv, Merck) and pure water and ultimately maintained in 0.1% TFA-water. After applying the samples to the top of the cartridge and washing with 0.1%TFA-water (5 ml), the analytes were eluted with 100% methanol containing 0.1% TFA (5 ml). The fractions were collected at 4°C, frozen at -80°C and then lyophilized at -110°C for 48 h. Under these conditions, recovery of the marker QR [ 3 H-F] SR-opiorphin, added to plasma samples, was 76 ± 5 % (mean ± SD for n=20).
Finally, dried extracts were re-suspended in 250 µl pyrolyzed water at 4°C then centrifuged 30 min at 4000 rpm and +4°C to quantify opiorphin-related components by radioactivity measurement (radiometer Wallac, PerkinElmer) and RP-HPLC in conjunction with PDA and radiometer analyses and/or ELISA-Opiorphin immunoassays.
Reverse phase C18-HPLC Chromatography: RP-HPLC, coupled with online PDA (224 nm) and radiometric (150-TR PerkinElmer) detection, was used to separate, identify and semi-quantify the different opiorphin-related molecular forms contained in human plasma extracts from in vitro PK experiments. Reversed Phase-High Performance Liquid Chromatography (RP-HPLC) used a C18-bonded stationary phase and an acetonitrile mobile phase in the presence of 0.1% trifluoroacetic acid (TFA, Sigma-France).
The re-suspended extracts (equivalent to 100 µl initial plasma volume), obtained during the above-described procedures, and were applied to the top of the C18/RP-HPLC analytical column (150×4.5 mm Luna 5 µ Phenomenex-France) under TFA 0.1%-water solvent equilibrium conditions. The various components were eluted and isolated according to their hydrophobic characteristics, in a 25min linear gradient from 0% to 50% acetonitrile (Lichrosol, Merck), containing 0.1% TFA at a 1 ml/min flow rate (Surveyor HPLC system, Thermo Scientific-France). The entire HPLC system was thermoregulated at 12°C. Each fraction (1 ml) was collected and lyophilized at -110°C for 48 h. Each chromatographic profile was driven, integrated and analyzed by the ChromQuest software. The peak height values of each peak of interest as well as those for a defined inner standard peak were calculated. Eluted fractions were collected at a 1 min timeinterval. Each fraction was lyophilized at -110°C for 48 h. In opiorphin PK experiments, the content of radioactivity of each sample, i.e., crude plasma, plasma extracts, HPLC fractions, was determined to evaluate the recovery of each processing step.
The opiorphin-like content of samples (SepPak extracts and/or HPLC fractions) was also measured using a quantitative and specific immunoassay (competitive-ELISA) developed in the laboratory [17].

Immunoassay for opiorphin:
The recently published protocol was used to assess the opiorphin-like content of samples [17]. Optimized assay conditions are summarized as follows: For the coating, 40 ng of the Y-[(CH 2 ) 12 ]-QRFSR peptide per 200 µl coating buffer (100 mM potassium phosphate, P H 7.1) were added to individual wells on a 96well micro-titration plate and incubated overnight at +4°C with light shaking. In parallel, 100 µl of standard or samples, that were serially diluted 2-fold with incubation buffer (200 mM Tris-HCl, pH 7.5+150 mM NaCl+0.1% Tween 20+0.1% bovine serum albumin), were preincubated in Screen Mates tubes (Matrix, Thermo Scientific-France) overnight at 10°C, in the presence of 100 µl anti-opiorphin antibody diluted at 1/80 000. The following day, after washing 5 times with washing buffer (1 tablet PBS-Sigma in 200 ml pure pyrolyzed water + 0.1% tween 20), 250 µl of saturation buffer (20 mM Tris-HCl, pH 7.5+150 mM NaCl+0.1% Tween 20+0.5% gelatin) were added to the individual coated-wells and incubated for at least 1 h at 20°C. Then, after washing, 100 µl of the pre-incubated immunological reaction were transferred onto the coated and saturated micro-titration plates and incubated 1.30 h at 10°C in a humid atmosphere. After washing, 100 µl of the anti-rabbit IgG conjugated to HRP (Pierce, ThermoScientific-France), diluted in Tris buffer (20 mM Tris-HCl, pH 7.5+150 mM NaCl+0.1% Tween 20+0.1% BSA) at 1/3 000, were added to each well and incubated for 1 h at 20°C. After incubation an ultimate wash was performed and 100 µl of the HRP chromogenic substrate (StepUltraTMB-ELISA, ThermoScientific-France) were added and incubated for 30-45 min at 20°C. Finally, the reaction was stopped by adding 100 µl 4N H 2 SO 4 . Plates were read at 450 mm with a microplate spectrophotometer (Infinite M200, Tecan-France) and the results In vivo studies using a rat pain model Animals: Male Wistar rats (Harlan, France) weighing 250-280 g were used in this study. After a 7-day acclimatization period, they were weighed and randomly housed according to the treatment groups in a room with a 12 h alternating light/dark cycle (9:00 pm/9:00 am) and controlled temperature (21 ± 1°C) and hygrometry (50 ± 5%). Food and water were available ad libidum. They were experimentally only tested once. Chemicals: Opiorphin analog (Genosphere Biotechnologies, France) was dissolved in vehicle solution (55% of PBS 100 mM-45% of Acetic acid 0.01N) and systemically (I.V.) injected, 10 to 15 min prior to the behavioral tests, at doses ranging from 0.5 to 2 mg/kg body weight. Morphine HCl (Francopia, France) was dissolved in saline (0.9% sodium chloride in distilled water) and injected I.V. 15 min before the behavioral test, at 2 mg/kg dose. All drugs were administered in a volume of 1 ml/kg body weight.

The Formalin Test:
The previously prescribed protocol [1,10,16] was used to assess the analgesic potency of opiorphin analog in a chemical-induced inflammatory pain model. Groups of 8 rats were used for each experiment. 50 µl of a 2.5% formalin solution was injected under the surface of the left hind paw 10-15 min after I.V. injection of opiorphin analogs, morphine or vehicle. The duration of formalin-injected paw licking and the number of inflamed paw flinches and body tremors were recorded for a period of 60 min after formalin administration. The behavioral scores were expressed as means ± standard error of the mean (SEM) for n=8 rats.

Statistical Evaluation:
The significance of differences between groups was evaluated using the Kruskal-Wallis one-way analysis of variance (KWT, a non-parametric method) for comparison between several independent variables across the experimental conditions. When a significant difference among the treatments was obtained, the Mann-Whitney post hoc test (MWT) was applied to compare each treated group to the control one. For all statistical evaluations, the level of significance was set at P < 0.05. All statistical analyses were carried out using the software StatView®5 statistical package (SAS, Institute, Inc., USA).

Structure-activity relationship study
In order to identify the amino acid residues or functional groups required for opiorphin inhibitory potency toward both AP-N and NEP human ectopeptidases, the molecular relationship of structure to activity, namely Structure-Activity Relationship (SAR), of opiorphin native peptide was first investigated. The inhibiting activity of each modified compound was evaluated toward human recombinant NEP (rh-NEP) and AP-N (rh-AP-N), the residual enzyme activity was measured by continuous fluorimetric assays in the presence of specific fluorescent substrate.
The importance of the free C-terminal carboxyl group of the QRFSR-COOH peptide in inhibitory potency toward rhNEP, in particular, rhNEP CarboxyDiPeptidase activity. Indeed, the amidation of the C terminal (QRFSR-CONH 2 ) gives rise to a compound displaying diminished inhibitory potency toward rhNEP.
The key role played by the aromatic side chain ofPhe 3 residue (QRFSR) in the inhibitory potency of opiorphin toward rhNEP and rhAP-N activities. Indeed, substitution with a Tyr residue (QRYSR) led to a compound displaying up to an 8-fold decrease in rhAP-N inhibition potency and a slight decrease in rhNEP inhibition potency. Substitution by an Ala residue led to a compound with completely diminished inhibitory potency toward both rhNEP and rhAP-N [18].
The importance of the RFS central residues of the QRFSR peptide in the inhibitory potency of opiorphin toward rhNEP. The compounds QRGPR -QHNPR -QRFPR displayed equivalent inhibitory potency toward rhAP-N but a low or totally diminished inhibitory potential for rhNEP.
The importance of the guanidium side chains of the Arg 2 (R 2 ) and Arg 5 (R 5 ) residues in the inhibitory potency of opiorphin toward rhAPN. Indeed, their respective substitution by the εamine side chain of Lys residue (QKFSR and QRFSK) led to compounds displaying more than a 10-fold decrease in rhAP-N inhibitory potency while showing equivalent rhNEP inhibitory potency. Their respective substitution by an Ala residue confirmed these results [18].
In summary, there is a clear structural selectivity in the functional interaction of opiorphin with both human NEP and AP-N ectoenkephalinases. The aromatic residue of Phe 3 plays a critical role in the interactions of opiorphin with both targets. In addition, the C-terminal FSR tri-amino acids constitute the minimal active sequence for NEP inhibition; moreover, FSR-peptide is 10 times more active than the natural QRFSR-peptide in its inhibition potency toward rhNEP. Conversely, it seems that the entire amino acid sequence of opiorphin is required for full rhAP-N inhibition. In general, our results demonstrate that any change in the intra-peptide sequence inhibits or even abolishes at least one of the two inhibitory activities. In contrast, addition of an amide link with a Tyr residue at the N-terminal position of the peptide ([Y]-QRFSR) does not reduce the inhibitory potency of the peptide toward either human target and does not affect its antinociceptive potency in a pain rat model [1].

Metabolism of the native opiorphin peptide
In order to evaluate the half-life of circulating opiorphin in the human bloodstream, the fate of the natural peptide was analyzed using in vitro kinetic models. The metabolic profile of opiorphin native peptide in human plasma as a function of incubation time at 37°C is shown in Figure 1A. The major metabolism products, generated following a 60-min incubation period of 1 µg QRFSR/QR [ 3 H-F] SR per500 µl  Figure 1A), to the parent Gln 1 -RFSR-peptide. This result does not concur with a previous report indicating that pGlu formation (in enzymatic or non-enzymatic processes) minimizes susceptibility to degradation by aminopeptidase [19]. It is also interesting to point out that pGlu 1 -RFSR peptide is an efficient NEP inhibitor. To a lesser extent, a more hydrophilic molecular population was also observed on the HPLC profile, reaching a maximum from the 2 min time-point and remaining stable at about 12% over the 30 min incubation period. The chromatographic and kinetic behaviors of this population lead us to suggest that it could result from an opiorphin-related product binding to a human plasma component.

Selection of potent bioactive opiorphin peptidomimetics:
The peptidomimetic strategy consists of altering the physical characteristics of a peptide without changing its biological activity. Here we wished to design and select functional derivatives of opiorphin that would display in vivo bioavailability properties superior to the native peptide, in particular increased resistance against proteolytic degradation. Several modifications are known to improve the metabolic stability of peptides. Conventional modifications consist of protecting the NH 2 -and COOH-terminal ends by N-acetylation and C-amidation, respectively. However, SAR studies (see above) reveal that these modifications inhibit or even abolish opiorphin inhibitory potencies. Alternatively, amino acids can be selectively substituted with non-natural amino acids, most notably by a D-enantiomer or β-amino acid [20]. However, as previously reported, ⇓ changes on the structural conformation of N-and C-terminal amino acids (N-and C-terminally homologated opiorphin, ß 2 hGln-Arg-Phe-Ser-ß 3 hArg),while increasing by about 7-fold the metabolic half-life of the modified opiorphin in human plasma, reduced by up to 10-fold its inhibitory potency toward both targets. This indicated that the relocation of the terminal carboxy and/ or amino groups has an impact on opiorphin interaction with the enkephalin-inactivating NEP and AP-N [21].
A third possibility to increase the enzymatic stability of peptides is to reduce their peptide character (pseudo peptides), substituting peptide bonds with isosteric surrogates. The isosters most frequently used are the reduced peptide bond (methyl-amino, CH 2 -NH), the retro-inverso link, the aza group or polyethylene chain spacers such as (CH 2 ) 6 or (CH 2 ) 12 . Depending on the chemical residue in corporated, the most direct consequences are increased resistance to the lytic action of circulating peptidases and an increase in lipophilicity that serves to facilitate transport across biological barriers [20,22]. However, such chemically stabilized peptides can lose some, if not all, of their biological activity, such as the retro-inverso D-amino acid opiorphin analog that lost its ability to inhibit NEP (unpublished observations by Rougeot C).
Consistent with the above, most of the QRFSR-peptide changes failed to reproduce the biological activity of the natural peptide. However, a series of opiorphin derivatives were screened and selected step by step on the basis of their dual inhibitory potency for hNEP and/ or hAP-N. To test for specificity, hit compounds were further tested with respect to other members of the metallo-ectopeptidase family, such as DPPIV and ECE. Here we present only functional opiorphin derivatives displaying significant in vitro inhibitory activity toward human NEP and AP-N. human fresh plasma, were isolated by RP-HPLC in conjunction with PDA (224 nm) and radiometric detections and semi-quantified using Chromquest software. The data are expressed by relative peak height. The addition of tracer quantity of tritiated opiorphin established the drug plasma concentration with high precision even for small amounts of compound not usually detected using standard PDA detection. Finally, analyses with Kinetica software were used to predict from the concentration-time course the metabolic half-life of the native compound, either from plasma-induced hydrolysis and/or chemical changes.
Native QRFSR-peptide disappears from human plasma with a metabolic half-life evaluated at 5 min (R 2 =0.88, n=5 time points over the 8 min time course of incubation). One metabolite appeared as early as 1 min after incubation, reaching maximum relative levels after 30 min incubation. Its appearance inversely correlated with the disappearance of Gln 1 -RFSR native peptide ( Figure 1A). The maximal appearance after 30 min incubation was blocked in the presence of150 µM bestatin, a selective inhibitor of amino peptidases ( Figure 1B). In contrast, its appearance was not affected by 150 µM GEMSA, a selective carboxy peptidase B inhibitor ( Figure 1B). This result suggests that opiorphinis primarily hydrolyzed to an RFSR-peptide metabolite resulting from the activity of a plasma exo-aminopeptidase, potentially a glutamyl peptidase. Interestingly, the RFSR-peptide is about 3 fold less inhibitory than the native QRFSR-peptide toward both rhNEP and rhAP-N. To increase the 5 minutes half-life of native opiorphin, changes were designed at the level of this sensitive site.
Two additional radioactive molecular populations were distinguished on the RP-HPLC chromatograms during the time course

[C]-QRFSR peptide
We previously showed that addition of a Tyr residue at the N-terminal position of the opiorphin-peptide does not affect its in vitro inhibitory potency or its in vivo antinociceptive properties [1]. Potent NEP and AP-N inhibitors were designed on the basis that the molecules contain a strong metal-coordinating group [23]. These observations were also used to design an opiorphin peptidomimetic carrying at the N-terminal moiety a Cys-thiol functional group that is a strong Zn atom-coordinating group.

[C]-[amino-hexanoic-acid spacer]-QRFSR peptide
In an attempt to protect the opiorphin derivative against degradation by circulating amino peptidases and thus increase its metabolic stability, a [CH 2 ] 6 polyethylene bridge [amino-hexanoic-acid spacer] was substituted for the peptide bond joining the Zn-chelating Cys 0 and the Glu 1 amino acids.
Surprisingly, the additional polyethylene bridge between the Cys 0 and Glu 1 residues of the [C]-QRFSR peptide was caused a decrease in inhibitory potency, of more than one order of magnitude relative to [C]-QRFSR peptide, toward NEP enzyme and particularly toward NEP-carboxypeptidase activity, whereas no difference in affinity towards AP-N was detected relative to [C]-QRFSR peptide.

QRF-[S-O-Octanoyl]-R peptide
Comparative conformational analyses of the opiorphin peptide revealed that the hydroxyl group of the Ser 4 residue does not seem to play a critical role in its bioactive conformation for hNEP [18]. Therefore, initially we tested the product resulting from esterification by octanoic acid, [CH 2 )8], of the serine hydroxyl group of the QRFSR peptide.
As shown in Figure 2  In the biologically relevant in vitro assay, using substance P, the physiological NEP substrate and human cell membranes as sources of native human NEP, the [C]-[(CH 2 ) 6 ]-QRF-[S-O-(CH 2 ) 8 ]-R peptide prevented, in a concentration dependent manner, substance P cleavage mediated by membrane-bound hNEP-Endopeptidase (mhNEP-Endo) activity with an IC 50 at 1.6 ± 0.4 µM (r 2 =0.95, n=13 determination points) ( Figure 5). Under the same assay conditions, it appears to be at least five times more potent than opiorphin natural peptide toward hNEP [1]. In addition using fluorescent substrates with human cell membranes as sources of native hNEP, the [C]-[(CH 2 ) 6 ]-QRF-[S-O-(CH 2 ) 8 ]-R peptide inhibited in a concentration dependent manner the mhNEP-Endo activity with an IC 50 at 1.6 ± 0.4 µM, and mhAP-N activity with an IC 50 at 0.9 ± 0.1 µM ( Figure 5). Thus, the designed analog presents similar affinity towards human NEP and AP-N, whether they are in a native membrane-anchored or recombinant soluble conformation.
In vitro assays using human recombinant DPP4 or ECE-1 revealed that the [C]-[(CH 2 ) 6 ]-QRF-[S-O-(CH 2 ) 8 ]-R compound did not inhibit rhECE1 and rhDPPIV-ectopeptidase activities even at 100 µM final concentration. These results indicate that, similarly to opiorphin, the opiorphin analog shows excellent selectivity with respect to related zinc-metallo peptidases, such as ECE1 (closely structurally related to NEP with 40% sequence identity) and DPPIV that is involved among other endopeptidases, including NEP, in the inactivation of the substance Pandbradykinin.
Di-peptide analogs can be metabolically more resistant to peptidase degradation. We tested the cystine-dipeptide (single disulfide bond connecting the Cys 1

[dCys]-QRF-[Ser-O-octanoyl]-[dArg]
Another strategy to protect peptide compounds against degradation by circulating peptidases is the replacement of the N-term and C-term amino acid residues, which are major targets for degradation by circulating exopeptidases, by their respective D-enantiomer.
As shown in Figure 6, [dC]-QRF-[S-O-(CH 2 ) 8 ]-[dR] derivative peptide inhibited, in a concentration dependent manner, rhNEP-Endo activity with an IC 50 at 4 ± 1 µM (r 2 =0.97, n=30 determination points) and rhNEP-CDP activity with an IC 50 at 21 ± 1 µM (r 2 =0.99, n=30 determination points). Strikingly, this derivative was at least 200 times more potent against rhAP-N activity than against rhNEP with an IC 50 at 0.022 ± 0.002 µM (r 2 =0.98, n=43 determination points). We also used human cell membranes as a source of native membrane-bound hNEP and hAP-N and confirmed that the [dC]-QRF-[S-O-(CH 2 ) 8 ]-[dR] peptide displays an unbalanced inhibitory profile. Indeed, it showed a dose-dependent inhibition of mhNEP-Endo activity with an IC 50 at 9 ± 1 µM (r 2 =0.98, n=21 determination points) and of mhNEP-CDP activity with an IC 50 at 37 ± 5 µM (r 2 =0.95, n=21 determination points). In addition, it appeared to be 30-100 times more potent toward mhAP-N activity than toward mhNEP(IC 50 at 0.3 ± 0.1 µM, r 2 =0.95,  Furthermore, substitution of the L-Arg 5 by its respective D-enantiomer clearly affected the inhibitory potency of the compound toward hNEP-carboxydipeptidase. The related [dC]-QRF-[S-O-(CH 2 ) 8 ]-R peptide inhibited mhNEP-CDP activity with an IC 50 at 2.6 ± 0.3 µM (r 2 =0.98, n=30 determination points), about ten times more potent than the D-Arg 5 counterpart. Such a difference leads us to propose the existence of a stereo-chemical requirement for optimal interaction of the peptide with the catalytic site of NEP. Conversely, the substitution of the L-Cys 0 by its respective D-enantiomer clearly enhanced the inhibitory potency of the compound toward hAP-N (about 50 times more potent than the L-Cys 0 counterpart) and may be due to the fact that its spatial conformation provides tight binding to the AP-N target.
derivative probably displays some superior in vivo bioavailability properties compared to native opiorphin peptide, such as a possible gain in circulating amino-and carboxy -peptidase resistance. However, it's very modest gain in hNEP inhibitory potency, combined with a distinctly unbalanced bioactive profile eliminated it as a suitable candidate molecule. Therefore, only the C-[(CH 2 ) 6 ]-QRF-[S-O-(CH 2 ) 8 ]-R derivative was retained for further exploration.

Metabolism and Toxicity profile of the best performing opiorphin functional derivative
Metabolism in fresh human plasma: We established overall in vitro pharmacokinetic and metabolic parameters, based on an in vitro time-dependent system, using opiorphin or its derivative incubated in human plasma. Kinetica software, which is used to predict the metabolic half-life (T½) of the parent peptide from the concentrationtime course, was used in this study.
As shown above, in vitro kinetic analyses in human plasma revealed that the native QRFSR-peptide disappears with a half-life evaluated at 5 min. Its disappearance results in part from the cyclization of Gln 1 (16% maximum) but mainly from the hydrolytic removal of both Gln 1 -and pGlu 1 -peptides by plasma amino peptidases (reaching a maximum of 84% of the parent peptide at 60 min incubation) and also, to a small extent (12%), from potential complex formation.  8 ]-R derivative is more metabolically stable in human plasma than opiorphin native peptide. In addition, it is important to point out that the major biotransformation product of the parent derivative, the cystine-dipeptide, is as active as the parent peptide in fluorescence-based NEP and AP-N assays.
All together, the data showed that, as expected, opiorphin derivative   Drug Absorption and in vitro Cytotoxicity: A range of in vitro ADME-Tox assays provided by Cerep Laboratories (Celle L'Evescault-France) allowed us to evaluate a number of factors including drug absorption and membrane permeability with the A-B permeability and P-glycoprotein ATPase efflux system [24]. The Caco-2/TC7 (pH 6.5/7.4) human cell line gives an indication of the intestinal epithelial transport potential of compounds [24]. Metabolic stability, using human liver microsomes and in vitro cytotoxicity in cell-based assays that measure cellular parameters such as cell viability, nuclear size and mitochondrial membrane potential using the HepG2 human cell line can also be evaluated.
Our data demonstrate that, compared to the reference positive and negative controls, no apparent in vitro human cell toxicityis observed for either QRFSR native peptide or [C]-[(CH 2 ) 6 ]-QRF-[S-O-(CH 2 ) 8 ]-R derivative peptide, incubated at 10, 30 and 100 µM final concentrations for 72 h at 37°C. For example, relative to controls at 100 µM, the peptides increased cell proliferation by 1and 12%, respectively and reduced nuclear size and mitochondrial membrane potential only by 1 and 5%, respectively. However, there is a clear decrease in the metabolic stability of the opiorphin derivative in the presence of human liver microsomes compared with the native opiorphin peptide: at 10 µM final concentration and after 60 min incubation, 2.5% of the parent [C]-[(CH 2 ) 6 ]-QRF-[S-O-(CH 2 ) 8 ]-R compound remains versus 47% remaining in the case of opiorphin. Surprisingly, the opiorphin derivative, although endowed with higher lipophilicity than opiorphin native peptide, did not display significantly increased trans-membrane cell permeability over the 60 min incubation-period at 37°C, as the apparent permeability coefficient of both tested compounds was <0.2×10-6 cm/s (10 µM test concentration and HPLC-MS/MS detection method). However, this result is probably due to the cellular model used, namely, TC7 human epithelial intestinal cells derived from the CaCO2 cell line, and known to express membrane-bound NEP and AP-N ectoenzymes. The cell line, therefore, is not an appropriate model for permeability studies of NEP and/or AP-N-inhibitor-ligands. Indeed, the mean recovery of the compounds in donor samples was dramatically low (0% for QRFSR and 14% for the derivative) due mainly to binding to TC7 cell membranes.
We then tested in vivo acute toxicity using a rat model provided by CERB (Centre de Recherches Biologiques, Baugy-France). CERB experimental conditions are based on a stepwise procedure, each step uses 3 male rats for each compound. No mortality occurred among the animals treated with QRFSR natural peptide at 100 mg/kg maximum dose, administered as a bolus in the caudal vein. This dose is a 100-fold the effective I.V. dose in the rat pain model. In contrast, the rat treated with a 100 mg/kg dose of [C]-[(CH 2 ) 6 ]-QRF-[S-O-(CH 2 ) 8 ]-R analog died 3 minutes after treatment; however, no mortality occurred among the 3 animals treated at 30 mg/kg I.V. These animals were further observed for general clinical and neurobehavioral signs, based on the Irwin method, for 14 days [25]. No clinical signs were observed during the course of the study of both peptides. Body weight gain was normal and no gross organ or tissue changes were detected by necropsy.
In conclusion, under the experimental conditions adopted by CERB, opiorphin natural QRFSR peptide administered intravenously at 100 mg/kg and    and body tremors were recorded over the 60 min-test period. The formalin test measures the behavioral response to a chemical-induced inflammatory nociception, which induces two distinct nociceptive phases separated a stationary interphase: a early acute phase (first 10 min after formalin injection) followed by a late phase in which a more tonic pain is elicited.
Here, we demonstrate that the [C]-[(CH 2 )6]-QRF-[S-O-(CH 2 ) 8 ]-R functional opiorphin analog inhibits, in a dose-dependent manner, the pain behavior induced by long-acting chemical stimuli with significant antinociceptive effect at 0.5, 1 and 2 mg/kg I.V. doses over early and later phases of the test (Figure 8). Thus, compared to the control vehicle rats, the opiorphin analog-treated rats at 1 and 2 mg/kg dose spent significantly less time in paw licking over the first 10 min-test period, from 161 ± 19 sec (vehicle) to 89 ± 15 sec (1 mg/kg) and 77 ± 6 sec (2 mg/kg) (P<0.05 and 0.01 vs vehicle by Mann-Whitney U-test, MWT, n=8 rats/group) as morphine-treated rats at 2 mg/kg I.V. dose (43 ± 13 sec, P<0.01 by MWT). The 1 and 2 mg-treated rats also spent significant less time in paw licking over the second 10-30 min period, from 468 ± 48 sec (vehicle) to 287 ± 28 sec (1 mg/kg) and 224 ± 27 sec (2 mg/kg) (P<0.01 and 0.001 vs vehicle by MWT, n=8 rats/group). The 0.5 mg/kg-treated rats also spent at least 30% less time in inflamed paw licking over pain periods: 98 ± 10 sec (early phase) and 334 ± 54 sec (late phase) compared to vehicle-treated rats 161 ± 19 sec and 468 ± 48 sec, respectively (P<0.05 vs vehicle by MWT, n=8 rats/group). From 30 min post-formalin injection, the duration of paw licking decreased in a parallel manner in both vehicle-and opiorphin analog-treated rats and their behavioral responses to the test compound as well as to morphine were not significant. Conversely, during this 30-60 min period, the control vehicle rats exhibited an important increase in the total number of formalin-injected paw flinches and body tremors. And systemic administration of opiorphin analog at 2 mg/kg significantly reduced this pain behavioral score throughout the 30 to 60 min timeperiod, from 300 ± 31 (vehicle) to 218 ± 15 (P ≤ 0.05 vs vehicle by MWT, n=8 rats/group).
This model was previously used for testing native opiorphin activity and we demonstrated that opiorphin, at 1 and 2 mg/kg I.V. doses inhibits nociception in both acute early and tonic late phases of the test by primarily activating µ-opioid pathways [1,10].
Thus, our data clearly indicate that the [C]-[(CH 2 ) 6 ]-QRF-[S-O-(CH 2 ) 8 ]-R opiorphin analog inhibits nociception induced by acute and long-acting chemical stimuli in the rat model. Strikingly, although metabolically more resistant and more potent in its ability to inhibit enkephalin-degrading ectopeptidases, the opiorphin analog-induced pain reduction in the formalin test is similar to the opiorphin natural peptide, in terms of dose effect, delay and duration of action. This could be due to the loss of a significant proportion of active derivative by dimerization and/or by hepatic metabolism in vivo in rats.

Conclusion
The goal of the study described here was to design and characterize functional analogs of opiorphin that display in vivo bioavailability properties superior to the native peptide. The inhibitory potency of the main functional derivatives toward human NEP and AP-N is summarized in Table 1. A close structural selectivity in the functional interaction of opiorphin with both human NEP and AP-N targets was first demonstrated by SAR studies, thus limiting the possibilities of chemical changes. Nevertheless, results of the study clearly demonstrate that addition of a N-terminal Zn-chelating group, a Cys-thiol group and replacement of the first labile peptide bond by a polyethylene surrogate, a [CH 2 ] 6 linker, and, finally, substitution of Ser 4 by a octanoyl-Ser, Ser-O-[CH 2 ] 8 , to the native opiorphin amino acid sequence produced a high performing C-[(CH 2 ) 6 ]-QRF[S-O-[CH 2 ] 8 ]-R derivative. This designed analog displays reinforced inhibitory potency toward hAP-N activity (more than 10-fold increase) and toward hNEP-Endopeptidase and CarboxyDiPeptidase activities (more than 40-fold increase) relative to the QRFSR natural peptide. Moreover, the analog shows increased stability in human plasma compared to unmodified opiorphin. Finally, we demonstrate that it retains the full analgesic activity characteristic of the opiorphin native peptide, in terms of delay of action and effective doses, in the behavioral formalin-induced pain rat model. If we consider that the maximum effective analgesic dose for the two compounds is 1 mg/kg I.V., the safety-effectiveness ratio is estimated at 30 for the designed analog and at 100 for the native peptide.