Characterization of a selective antagonist of neuropeptide Y at the Y2 receptor. Synthesis and pharmacological evaluation of a Y2 antagonist.

Neuropeptide Y (NPY) is a potent inhibitor of neurotransmitter release through the Y2 receptor subtype. Specific antagonists for the Y2 receptors have not yet been described. Based on the concept of template-assembled synthetic proteins we have used a cyclic template molecule containing two β-turn mimetics for covalent attachment of four COOH-terminal fragments RQRYNH2 (NPY 33-36), termed T4-[NPY(33-36)]4. This structurally defined template-assembled synthetic protein has been tested for binding using SK-N-MC and LN319 cell lines that express the Y1 and Y2 receptor, respectively. T4-[NPY(33-36)]4 binds to the Y2 receptor with high affinity (IC50 = 67.2 nM) and has poor binding to the Y1 receptor. This peptidomimetic tested on LN319 cells at concentrations up to 10 μM shows no inhibitory effect on forskolin-stimulated cAMP levels (IC50 for NPY = 2.5 nM). Furthermore, we used confocal microscopy to examine the NPY-induced increase in intracellular calcium in single LN319 cells. Preincubation of the cells with T4-[NPY(33-36)]4 shifted to the right the dose-response curves for intracellular mobilization of calcium induced by NPY at concentrations ranging from 0.1 nM to 10 μM. Finally, we assessed the competitive antagonistic properties of T4-[NPY(33-36)]4 at presynaptic peptidergic Y2 receptors modulating noradrenaline release. the compound T4-[NPY(33-36)]4 caused a marked shift to the right of the concentration-response curve of NPY 13-36, a Y2-selective fragment, yielding a pA2 value of 8.48. Thus, to our best knowledge, T4-[NPY(33-36)]4 represents the first potent and selective Y2 antagonist.

Neuropeptide Y (NPY) 1 is a 36-amino acid peptide amide distributed widely in the central and peripheral nervous system (1)(2)(3). NPY exerts many biological effects, especially on cardiovascular, metabolic, food intake, behavior, anxiety, and endocrine regulation (4). Several lines of evidence suggest potential roles for NPY in the pathophysiology of hypertension, obesity, diabetes, and psychiatric disorders (5). NPY acts through a number of G-protein-coupled receptors termed Y1, Y2, Y4/PP, Y3, and Y1-like receptors (4). Only Y1, Y2, and Y4/PP receptors have been cloned (6 -10). Y1 receptors are present in the sympathetic nervous system mainly postsynaptically and mediate vasoconstriction. Those of the Y2 subtype are present prejunctionally and inhibit the release of catecholamines (11). Furthermore, it has been demonstrated that NPY Y2 receptors are located on noradrenergic nerve terminals within the hypothalamus and other brain regions, exerting an inhibitory action on [ 3 H]norepinephrine release evoked by appropriate concentrations of potassium ions (12). Y2 receptors are also involved in endocrine control; NPY has been reported to inhibit through the Y2 receptor potassium-stimulated glutamate release (13), ␣-melanocyte-stimulating hormone release (14), release of luteinizing hormone in a steroid-free environment (15), prolactin release (16), and to potentiate the secretion of vasopressin from the neurointermediate lobe of the rat pituitary gland (17).
The assembly of bioactive peptides on topological template molecules according to the template-assembled synthetic proteins (TASP) concept has been shown to induce or stabilize specific conformations of various peptides and consequently, to modify their biological and pharmacokinetic properties (25). We have, for example, synthesized a TASP molecule able to bind and to stimulate selectively the angiotensin II AT2 receptor (26). Here, we design a molecule composed of four truncated NPY peptide fragments (NPY 33-36) attached to a cyclic car-rier molecule via their NH 2 termini. This TASP molecule was investigated for binding by NPY Y1 and Y2 receptors, and its antagonistic activity was established by its ability to prevent the increase in intracellular calcium induced by NPY using a new methodology based on the analysis of Ca 2ϩ increase in single cells. The biological antagonistic properties of the compound were confirmed by measuring the blockade of the inhibitory action of a Y2 agonist on K ϩ -induced [ 3 H]norepinephrine release from perfused rat hypothalamic synaptosomes.

Synthesis of the TASP T 4 -[NPY(33-36)] 4
For the effective synthesis of the TASP (27) T 4 -[NPY(33-36)] 4 (III in Fig. 1) chemoselective ligation methods were applied. Oxime bond formation (28) was used to attach the functionally modified NH 2 -terminal (aminoxy group) peptide fragments to the cyclic peptide template, T 4 (I). T 4 contains four lysine residues acting as attachment sites and the ␤-turn mimic 8-aminomethyl-2-naphthoic acid (29). The ⑀-amino groups of lysine were transformed to aldehydic functions by reaction with glyoxylic acid 1,1-diethylacetal and subsequent hydrolysis (30) to yield I. Solid phase synthesis of the partial NPY sequence Arg-Gln-Arg-Tyr (II) was performed on Rink amide resin using N-(9-fluorenylmethoxycarbonyl) chemistry and 2,2,5,7,8-pentamethylchroman-6-sulfonyl, trityl, and tert-butyl protection for the functional groups of the Arg, Gln, and Tyr side chains (31). The ligation reaction proceeded as follows. The tetrakis aldehyde T 4 was dissolved in 1 M sodium acetate, and the pH was adjusted to 5 with acetic acid. A 1.2-fold excess of the tetrapeptide II (with respect to the aldehyde groups) in 1 M sodium acetate (pH 5) was added, and the mixture was stirred at room temperature for 15 h. The crude product was purified directly by semipreparative reverse phase HPLC, and the isolated TASP III was characterized by electrospray-mass spectrometry (Fig. 2) and amino acid analysis.

Cell Cultures
SK-N-MC cells were derived from a human neuroblastoma and were cultured according to the American Type Cell Culture recommendations. LN319 cells, obtained from a human glioblastoma, were grown in Dulbecco's modified Eagle's medium supplemented with 5% fetal calf serum, glutamine, 100 IU of penicillin, and 100 g/ml streptomycin in a 5% CO 2 , 95% air incubator at 37°C. Tissue culture media were purchased from Life Technologies, Inc. (Basel, Switzerland), and fetal calf serum was obtained from Seromed (Berlin, Germany). The cells were used for experiments from passage 190 to 210. To obtain reproducible data over 2 years of experiments, 70% confluent cells were washed with phosphate-buffered saline and harvested using 0.15% trypsin containing 0.4 mM EDTA. Cells were further diluted 1:3 and plated onto either 60-mm cell culture dishes (Nunc, Denmark) or 12-mm glass coverslips (Huber and Co, AG, Reinach, Switzerland). Media were changed every 3 days.
Binding Assays Y1 Binding Assay-Binding of iodinated NPY (Amersham, Buckingamshire, UK, 74 TBq/mmol) was performed by incubation at 37°C for 1 h of various peptide dilutions in Eagle's minimum essential medium containing 0.5% bovine serum albumin, 4 mM MgCl 2 , and 10 mM Hepes with SK-N-MC cells that exhibit exclusively Y1 receptors. Cells were then washed three times with buffer and lysed in 1% Nonidet P-40 (Fluka, Neu-Ulm, Germany), 8 M urea, 3 M acetic acid. Nonspecific binding was estimated by adding 1 M NPY to the incubation mixture. Displacement curves were obtained by incubation of various concentrations of competitive peptides together with a nonsaturating dose of iodinated NPY. At the end of the incubation period, cells were washed and lysed. Bound radioactivity was determined by ␥-counting. Halfmaximal inhibition of the binding, obtained with 125 I-NPY, is given as the IC 50 . Each point represents the mean Ϯ S.D. of at least four experiments.
Y2 Binding Assay-We used a human glioblastoma cell line, LN319, for Y2 binding studies (24). Prior to performing the binding experiments, adhering LN319 cells were harvested in 50 mM Tris (pH 7.5), which contained 100 mM NaCl, 4 mM MnCl 2 , 1 mM EGTA, 0.1% bovine serum albumin, 0.25 mg/ml bacitracin, and incubated at room temperature for 45 min. Bound radioactivity was determined after separating the unbound fraction by centrifugation.

Determination of cAMP
Six-well plates, containing confluent LN319 cell cultures, were washed and incubated at 37°C for 1 h in Eagle's minimum essential medium containing 0.5% bovine serum albumin, 4 mM MgCl 2 , 10 mM Hepes, 100 M papaverin, and 10 M forskolin and one of the peptides to be tested in varying dilutions. Cells were washed once in 100 mM sodium phosphate buffer (pH 7.5) and lysed with 0.75 ml of 0.1 M HCl. After centrifugation, the supernatant was recovered and lyophilized. cAMP concentration was measured by a radioimmunoassay using a commercially available kit (Amersham).

Antagonistic Properties of T 4 -]NPY(33-36)] 4 on the Free Cytosolic Calcium Response to NPY
LN319 cells were plated on glass coverslips 48 h before intracellular free calcium measurements. Intracellular free calcium concentration [Ca 2ϩ ] i was determined using the fluorescent probe fluo-3/AM. The dye was loaded into the cells by adding the acetoxymethyl ester fluo-3/AM (2.5 M) from a 1 mM stock in dry dimethyl sulfoxide to the culture medium (Dulbecco's modified Eagle's medium or Eagle's minimum essential medium without serum or complements) and incubating the cells for 30 min at room temperature in the dark. Pluronic acid (2 l/ml) in 25% dimethyl sulfoxide was added to fluo-3/AM to disperse the dye. After loading, the cells were washed three times with medium and placed in a chamber with 0.5 ml of physiological saline solution containing 140 mM NaCl, 2 mM CaCl 2 , 4.6 mM KCl, 1.0 mM MgCl 2 , 10 mM glucose, and 10 mM Hepes (pH 7.4). Tween 20 (Pierce) 0.0008% was present in the medium to prevent NPY from sticking to the walls of the exposed surfaces. Fluorescence images of the intracellular calcium localization were obtained with a laser-scanned confocal microscope (MRC 500 confocal imaging system, Bio-Rad) equipped with an argon ion laser and a fluorescein (488 nm) or rhodamine (514 nm) filter cartridge. The scanner and detectors were attached to an inverted microscope (Diaphot, Nikon).
The confocal microscopy technique with video recording provides serial readings, at 5-s intervals, of the individual fluorescence for a set of cells. The changes in Ca 2ϩ were evaluated in single cells on whole images containing 5-15 cells using the NIH image analyzer program (29 -385 cells were used to test each concentration of peptides). In each experiment it uses a fixed delineation of the cell borders, entered with a pointer device on the image screen. For each cell, five intensity readings were recorded. The base-line fluorescence was determined by averaging two consecutive images (F base ); the signal induced by adding the T 4 -[NPY(33-36)] 4 solution (F antag ); the peak response induced by adding the NPY solution (F NPY ); the maximal response observed after adding the nonfluorescent Ca 2ϩ ionophore A-23187 (10 M) to saturate the intracellular dye with calcium and thereby obtain maximal fluorescence (F max ); and the minimal fluorescence (F min ) was measured after addition of an excess of EDTA (5 mM). The F antag values were not used for further analysis, after it had been demonstrated that T 4 -[NPY(33-36)] 4 did not induce any significant response.
The intracytoplasmic calcium concentrations at the peak of the NPY effect were derived from fluorescence readings by using the formula (32) in nmol/liter: Attempts were made to correct Ca NPY value for the base-line level; however, as the peak response was poorly correlated with the base-line value (r 2 ϭ 0.13, r log/log 2 ϭ 0.24), this only added noise without modifying the results; so Ca NPY was retained as the response variable. It was transformed to logarithmic functions to normalize its distribution. Means and standard deviations of log 10 Table I for the corresponding number of cells and geometric means for each NPY and T 4 -[NPY(33-36)] 4 level.

cients of variation of the original Ca NPY values).
A dual response Hill model was used to relate the log 10 (Ca NPY ) values to the NPY concentrations, to account for the observed fading of the response at very high doses of the agonist (33). The T 4 -[NPY(33-36)] 4 was considered as a competitive antagonist, affecting the apparent EC 50 of NPY. Thus the mathematical expression of the pharmacodynamic model was as follows, where E min is the base-line (no effect) value; E max , the maximal response; EC 50 , the concentration of NPY associated with a half-maximal response; FC 50 , the concentrations associated with a one-half

Antagonistic Properties of T 4 -[NPY(33-36)] 4 at Presynaptic Peptidergic Y2 Receptors Modulating Norepinephrine Release
Preparation of Synaptosomes-Nerve endings were prepared from the hypothalamus of adult Wistar rats (200 -250 g) according to the method of Gray and Whittacker with minor modifications (34). Briefly, the rat hypothalamus was homogenized using a Teflon-glass tissue grinder (clearance 0.25 mm) in 40 volumes of 0.32 M sucrose buffered at pH 7.40 with phosphate. The homogenate was centrifuged (5 min, 1,000 ϫ g) to remove nuclei and debris, and the supernatant was again centrifuged (20 min, 12,000 ϫ g) to isolate crude synaptosomes. The synaptosomal pellet (at a protein concentration of 0.6 -0.8 mg/ml) was then resuspended in a physiological medium with the following composition (in mM): 125 NaCl, 3  for 15 min at 37°C. The labeled particles were then distributed in several parallel superfusion chambers and superfused at 37°C with continuously oxygenated medium, at a rate of 0.6 ml/min (36). After 10 min, to equilibrate the system and to reach a constant spontaneous efflux, T 4 -[NPY(33-36)] 4 was added and perfusion continued for a further 15 min. The synaptosomes were depolarized with 15 mM KCl for 90 s (substituting for an equimolar concentration of NaCl). The NPY fragment, NPY 13-36 (Peninsula Laboratories, Merseyside, U. K.) was added concomitantly with K ϩ . Fractions were collected every min, and the radioactivity (present as [ 3 H]norepinephrine) in each fraction and in the filters was determined after separation on Biorex 70 columns (37).
Evaluation of Results-The [ 3 H]norepinephrine found in each fraction collected was calculated as a percentage of the total [ 3 H]norepinephrine recovered (fractions plus filters). The concentration-dependent effects of NPY 13-36 were calculated as follows. The area of baseline efflux curve was subtracted from the area of the total release curve obtained in the absence and in the presence of the compound tested. The areas under the release curves of the time course were recorded for each experiment according to the Newton-Cotes integration formula. Data obtained according to this method were used to calculate the percentage inhibition of the K ϩ -evoked release of [ 3 H]norepinephrine in the presence of NPY 13-36 or in the presence of NPY 13-36 plus T 4 -[NPY(33-36)] 4 . The apparent pA2 value for the antagonist was calculated by means of the Schild regression analysis according to the following formula: pA2 ϭ log(EЈ/E Ϫ 1) Ϫ logB, where EЈ and E are those concentrations of the agonist which caused half-maximum effects in the presence and in the absence of the antagonist, respectively. B is is the concentration of the antagonist.

RESULTS
Binding Assays-As described above, the SK-N-MC and LN319 cells express Y1 and Y2 receptor subtypes, respectively. For competitive binding studies, in addition to the native NPY, we used peptides with differential selectivity for Y1 and Y2 binding. The Leu-31-and Pro-34-substituted NPY has been shown to be a Y1 agonist (38), whereas the NPY 13-36 has been reported to bind preferentially to the NPY Y2 receptor subtype. Fig. 3 depicts the results of binding experiments obtained with the two cell lines. SK-N-MC cells (Fig. 3A) bind NPY and Leu-31, Pro-34 NPY equally well as shown by the similar competition displacement curves (Fig. 3A). In contrast, NPY 13-36 binding was 2,000-fold less as this cell line does not express Y2 receptors.  4 was also added to angiotensin II type 1 and type 2 and muscarinic receptor preparations at concentrations up to 10 M, and there was no binding. A slight interaction was observed at a high concentration of the antagonist in binding experiments with the ␣ 1B -adrenergic receptor (42% inhibition at 20 M). cAMP Measurements-To assess whether this peptidomimetic had intrinsic agonist properties we tested its ability to inhibit cAMP accumulation in LN319 cells. Whereas NPY inhibits forskolin-stimulated cAMP accumulation in LN319 cells with an IC 50 of 2.5 nM, T 4 -[NPY (33)(34)(35)(36) (Table II). The fitting of the general pharmacodynamic model provided estimates of parameters, which are indicated in Table II. The half-fading concen-tration of NPY, FC 50 , was estimated with poor precision, in accordance with the few points on the descending part of the dose-response curve. This model explained most of the variance in the data (r 2 ϭ 0.91). The intercorrelations between the parameter estimates were acceptable (highest value r ϭ 0.77). The Hill coefficient ␥ departed significantly from unity, based on its 95% confidence interval; this was not the case for ␦.
NPY also increased in a dose-dependent manner the number of cells that get a rise in free calcium (18 and 71% at 0.1 and 100 nM concentrations, respectively), whereas T 4 -[NPY (33)(34)(35)(36)] 4 decreased the number of cells that responded to NPY (data not shown). To express the calcium response in an all-ornone fashion, cells were considered as responding if their peak Ca 2ϩ was higher than the mean Ϯ 2 S.D. observed after vehicle only (see Table I). The dose-response curve in the cells that responded to NPY provided similar results when compared with those observed for the mean of responding ϩ nonresponding cells (see Fig. 5).
Antagonistic Properties of T 4 -[NPY (33)(34)(35)(36)] 4 Fig. 6 shows the dose-response inhibition curve of NPY 13-36 on the release of [ 3 H]norepinephrine evoked by depolarization with 15 mM KCl from hypothalamic nerve endings. Assuming that inhibition of 40% represented the maximum response, then a 6.5 nM concentration of NPY 13-36, which caused 20% inhibition, roughly produced the half-maximum effect. Therefore these IC 20   pendent manner (IC 50 values in the presence of 1, 10, and 100 nM were 9, 22, and 109 nM, respectively). The concentrationresponse curve of the NPY 13-36 fragment was shifted to the right by the addition of 10 nM antagonist (Fig. 7). The antagonist appeared competitive for the Y2 receptor that modulated norepinephrine release because the effect of each antagonist concentration could be overcome by high doses of the agonist NPY 13-36. To assess competitive antagonism of T 4 -[NPY (33)(34)(35)(36)] 4 , the apparent pA2 value for this compound was deter-mined using three different concentrations (1, 10, and 100 nM  Fig. 7.

DISCUSSION
The present sudy shows that T 4 -[NPY (33)(34)(35)(36)]4 is a potent and selective ligand for the NPY Y2 receptor and exhibits in vitro antagonistic properties in two models of Y2-mediated effects. Several NPY Y1 receptor antagonists have been characterized and demonstrated to inhibit the pharmacological vasopressor effect of NPY. This compound displays a high affinity for the Y2 receptor (IC 50 ϭ 62 nM) but exhibits 2 orders of magnitude less interaction with the Y1 receptor. The antagonistic properties of the molecule have been confirmed by demonstrating inhibition of function.
For these functional studies, we used an original experimental approach involving single-cell recordings of cytosolic calcium responses to a biochemical stimulus. Other investigators have reported a similar technique on a limited number of cells (39 -41). The ability to perform hundreds or thousands of single-cell measurements required the development of appropriate statistical methods to cope with the large amount of data. The pharmacological characterization of an antagonist with this new methodology cannot rely on simple Shild plots drawn across average points. Thus, we used computer-based nonlinear regression methods, with special attention to the variability structure displayed by the data.
Preliminary explorations led us to use the intracytoplasmic calcium concentrations at the peak of the NPY effect as the response variable. Correction of this value for the base-line calcium level could be theoretically justified. However, this only added noise to the data, as the peak response was poorly correlated with the base-line value, and this correction was discarded. The response variable showed a strongly skewed distribution. A logarithmic transformation brought the results closer to a Gaussian distribution, allowing the application of a least square approach to estimate pharmacodynamic parameters. This observa- tion indicates a log-normal behavior of calcium concentrations, consistent with its contamination by multiplicative randomness and warrants the description of results by geometric rather than arithmetic means. Biologically, the NPY stimulus is indeed linked to the intracytoplasmic calcium response by a cascade of amplification steps. The high response variability must be emphasized: most coefficients of variation associated with the geometric means exceeded 100%. A pharmacodynamic study could have recorded the average response of a cluster of cells, a piece of tissue, or a whole organ. In that situation, however, only a global effect would have been measured, without consideration for the heterogeneity of individual cell responses. Moreover, averaging would have distorted the signal in many ways. Thus, the results shown here may not reflect what might be observed in multiple cell or tissue preparations. On the other hand, the individual cell behavior itself can be considered as the integration of many subcellular quantal responses (42,43). Interestingly, when the same procedure of calculation was used for only the responding cells, we obtained an almost identical mathematical model. This finding may indicate that the free calcium response to NPY does not obey an all-or-none rule. Thus, the shape of the dose-effect relationship at the cellular level not only reflects the dose-effect relationship at the subcellular scale but also its variable nature (42).
Despite the high number and variability of the single-cell responses, a pharmacodynamic model was fitted satisfactorily to the whole data set by nonlinear regression. We have chosen a dual Hill model, able to account for the fading of the response at high doses, which is frequent in the field of peptide pharmacology. This model included a competitive antagonism, which adequately describes the pharmacodynamics of T 4 -[NPY (33)(34)(35)(36)] 4 in this experimental setting. However, all terms of the model must be interpreted with caution. As this was not a binding study, the EC 50 , FC 50 , and AC 50 cannot simply be equated with receptor affinities. Neither can they be considered equivalent to values measured in a living organism. This may account for the apparent differences observed between the pharmacodynamic parameters of the two functional assays we have used. Nevertheless, we cannot exclude the possibility that LN319 cells and hypothalamic synaptosomes exhibit different Y2 receptors (44). The slope coefficients ␥ and ␦ have also no straightforward interpretation; they rely strongly on the variance structure of the underlying phenomena (42).
The NPY gene has recently been disrupted in mice (45). These NPY-deficient mice exhibit seizures, an effect that may be due to inhibition of glutamate release which is known to be mediated by Y2 receptors present on the endings of presynaptic excitatory neurons of NPY (13). Therefore, this compound may be a useful tool for the study of the role of NPY in various disorders such as seizures or abnormalities in the reproductive axis.