Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Disease-causing receptor mutations are widely believed to lose function as a result of inability to engage in receptor-ligand or receptor-effector binding interactions. Recent observations challenge this view and suggest that receptor misfolding and subsequent misrouting is a mechanism that results in loss of receptor function (Conn et al., 2002; Janovick et al., 2002; Leaños-Miranda et al., 2002). We recently reported (Janovick et al., 2002) a pharmacologic stratagem relying on an antagonistic peptidomimetic to correct the routing of a naturally occurring mutant GnRHR and correctly route it to the plasma membrane. This "pharmacoperone" could then be removed and the rescued GnRHR was shown to bind ligand with similar specificity and affinity as the WT receptor and couple to its effector protein. Subsequently, the arresting observation was made (Leaños-Miranda et al., 2002) that 11 of the 13 reported human mutants of the GnRHR that cause hypogonadotropic hypogonadism (Janovick et al., 2002; Leaños-Miranda et al., 2002) could be rescued by this approach, as could many "manufactured" mutants (truncations, deletions, and Cys substitutions), suggesting the generality of this approach. Based on this and a similar observation for the V2 receptor (Morello et al., 2000), we proposed (Conn et al., 2002) that this approach might be generally applicable to correcting diseases for which the etiology is misrouted or misfolded proteins, such as cystic fibrosis, nephrogenic diabetes insipidus, hypercholesterolemia, cataracts, Alzheimer's, retinitis pigmentosa, and others (Bellotti et al., 1999; Brooks, 1999; Radford and Dobson, 1999; Deen et al., 2000; Kopito and Ron, 2000; Morello et al., 2000; Sanders and Nagy, 2000; Dobson, 2001; Ellgaard and Helenius, 2001; Gregersen et al., 2001; Hammarstrom et al., 2001; Helenius, 2001; Sherman and Goldberg, 2001; Couzin, 2002; Hartl and Hayer-Hartl, 2002; Luque et al., 2002). Since rescue molecules need not bind mutants at ligand binding sites to serve as stabilizing scaffolding, it is likely that constituents of pharmaceutical archives that have failed agonist and antagonist screens may include unrecognized rescue molecules. We suggested (Conn et al., 2002) a set of useful characteristics for pharmacoperones, among these: 1) specificity for the molecule being rescued, 2) ability to arrive at the correct cellular locus (enter the cell, get to the endoplasmic reticulum, and remain stable long enough to bind the nascent molecule), and 3) ability to dissociate from the molecule (or, in the alternative, not compete with the physiological ligand) after it arrives at the appropriate target locus. In the present study, we examine structural homologs of three different chemical classes of potential pharmacoperones to determine whether: 1) rescue is a general characteristic of different chemical families that interact with the ligand binding site of a particular receptor; 2) efficacy is related primarily to receptor binding affinity; 3) some mutants cannot be rescued by any compound; and 4) an agent that rescues one mutant is equally effective at rescuing others.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Natural sequence GnRH (NIDDK National Hormone and Peptide Program, Bethesda, MD) and the GnRH agonist buserelin [D-tert-butyl-Ser⁶,des-Gly¹⁰,Pro⁹,ethylamide-GnRH (Hoeschst-Roussel Pharmaceuticals, Somerville, NJ)] were obtained as specified. The following chemical structures (collectively referenced as "the drugs") were used; those of the quinolone class are prefaced by the letter "Q" and those of the indole class by the letters "In" and were produced by Merck and Company (Rahway, ,NJ) (Ashton et al., 2001a,b,c): Q89 [7-chloro-2-oxo-4-{2-[(2S)-piperidin-2-yl]ethoxy}-N-pyrimidin-4-yl-3-(3,4,5-trimethylphenyl)-1,2-dihydroquinoline-6-carboxamide], Q76 [N-(7-chloro-3-(3,5-dimethylphenyl)-2-oxo-4-{2-[(2S)-piperidin-2-yl]ethoxy}-1,2-dihydroquinolin-6-yl)-N'-cyclopropylurea], Q08 [(2S)-2-(2-{[7-chloro-6-[(6,7-dimethoxy-3,4-dihydroisoquinolin-2(1H)-yl)carbonyl]-3-(3,5-dimethylphenyl)-2-oxo-1,2-dihydroquinolin-4-yl]oxy}ethyl)piperidinium trifluoroacetate], In30 [(2S)-2-[5-[2-(2-azabicyclo[2.2.2]oct-2-yl)-1,1-dimethyl-2-oxoethyl]-2-(3,5-dimethylphenyl)-1H-indol-3-yl]-N-{2-[4-(methylsulfinyl)phenyl]ethyl}propan-1-amine], In31b [(2S)-N-[2-(4-carboxyphenyl)ethyl]-2-[5-[1,1-dimethyl-2-(4-methylpiperazin-1-yl)-2-oxoethyl]-2-(3,5-dimethylphenyl)-1H-indol-3-yl]propan-1-aminium trifluoroacetate], and In3 [(2S)-2-[5-[2-(2-azabicyclo[2.2.2]oct-2-yl)-1,1-dimethyl-2-oxoethyl]-2-(3,5-dimethylphenyl)-1H-indol-3-yl]-N-(2-pyridin-4-ylethyl)propan-1-amine]. Erythromycin-derived macrolides were prepared by Abbott Laboratories (Bush et al., 1999; Diaz et al., 1999) and are prefaced by the letter "A". They are as follows: A-7662.0 (erythromycin A), A-64755.0 [11-deoxy-11-[carboxy-phenylethylamino]-6-O-methyl-erythromycin A 11,12-(cyclic carbamate)]; A-177775.0 [3'-N-desmethyl-3'-N-cyclopentyl-11-deoxy-11-[carboxy-(3,4-dichlorophenylethylamino)]-6-O-methyl-erythromycin A 11,12-(cyclic carbamate)], A-222509.0 [3',3'-N-desmethyl-3',3'-N-cyclopropylmethyl-11-deoxy-11-[carboxy-(3-chloro,4-fluoro-phenylethylamino)]-6-O-methyl-erythromycin A 11,12-(cyclic carbamate)]. Dulbecco's modified Eagle's medium, OPTI-minimal essential medium, lipofectamine, and polymerase chain reaction reagents (Invitrogen, Carlsbad, CA) were obtained as indicated. Restriction enzymes, modified enzymes, and competent cells for subcloning were purchased from Promega (Madison, WI). Other reagents were of the highest degree of purity available from commercial sources.

Vector Construction. WT hGnRHR cDNA in pcDNA3 was subcloned into pcDNA3.1 at KpnI and XbaI restriction enzymes sites. All the GnRHR mutants were constructed, sequenced, and prepared as previously described (Janovick et al., 2002). The WT human and rat GnRH receptor (Janovick et al., 2002; Leaños-Miranda et al., 2002) and naturally occurring mutants of the hGnRHR, associated with human hypogonadotropic hypogonadism (Achermann and Jameson, 2001), were N¹⁰K, T³²I, E⁹⁰K, Q¹⁰⁶R, A¹²⁹D, R¹³⁹H, S¹⁶⁸R, C²⁰⁰Y, S²¹⁷R, R²⁶²Q, L²⁶⁶R, C²⁷⁹Y, and Y²⁸⁴C. Manufactured mutants of the rGnRHR include the shortest rat GnRHR c-terminal truncation mutant that results in receptor loss of function (des^325-327-rGnRHR) (Brothers et al., 2002), two deletion mutants [des^237-241-rGnRHR, which has been rescuable previously (Janovick et al., 2002), and des^260-265-rGnRHR, which has not been rescuable previously (Janovick et al., 2002)], and two Cys mutants [C²²⁹A-GnRHR, which was not rescuable previously with the indole IN3, and C²⁷⁸A-GnRHR, which was rescued with this agent (Janovick et al., 2002].

Transient transfection of COS-7 cells. Wild-type (WT) hGnRHR and altered receptors were transiently expressed in COS-7 cells as described (Leaños-Miranda et al., 2002). Cells were transfected with 0.05 µg of DNA/well [for inositol phosphates (IP)] or 0.1 µg of DNA/well (for saturation binding studies) containing 1% DMSO (vehicle) or 1 µg/ml of each indole, quinoline or erythromycin macrolide, prepared in the vehicle. The quantification of IP production and saturation binding were made 27 h after the start of transfection. From transfection until 18 h before assay, drug was present; during the last 18 h, drug was not present. The quantification of IP production and saturation binding were made as described (Huckle and Conn, 1987).

Statistics. The data shown are the means ± S.E.M. from triplicate (IP) determinations. For clarity in three-dimensional graphics, the S.E.M. bars were omitted. In all experiments, the standard deviation was typically less than 10% of the corresponding mean. Each experiment was repeated three or more times with similar results. The results shown are from a single experiment.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Indoles show the efficacy (assessed by IP production) of each of the drugs tested on rescue of a single mutant, E90K. Transfection of cells and measurement of IP production were measured as described under Materials and Methods.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Figures 1 (indoles), 2 (quinolones), and 3 (erythromycin macrolides) show the efficacy (assessed by IP production) of each of the drugs tested as pharmacoperones of a single mutant, E⁹⁰K. This mutant was selected because of its low basal expression in the absence of the first drug examined, IN3, and its pronounced response (both IP production and radioligand binding) to rescue by IN3 (Janovick et al., 2002). For each chemical class, the data are presented with the lowest IC₅₀ value (for the hGnRHR, shown in figures) first. The data indicate that a concentration of 1 µg/ml is, in most cases, the dose that elicits an optimum rescue response.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Quinolones show the efficacy (assessed by IP production) of each of the drugs tested on rescue of a single mutant, E90K. Transfection of cells and measurement of IP production were measured as described under Materials and Methods.

View larger version (28K): [in this window] [in a new window]   Fig. 3.   Erythromycin macrolides show the efficacy (assessed by IP production) of each of the drugs tested on rescue of a single mutant, E90K. Transfection of cells and measurement of IP production were measured as described under Materials and Methods.

View larger version (40K): [in this window] [in a new window]   Fig. 4.   Indoles show the effect of the 1 µg/ml concentration of each member of the specified class of compounds in rescuing each component of the receptor palette. This figure shows IP production in the presence of 107 M buserelin. The transfection of cells and measurement of IP production occurred as described under Materials and Methods.

View larger version (31K): [in this window] [in a new window]   Fig. 5.   Quinolones show the effect of the 1 µg/ml concentration of each member of the specified class of compounds in rescuing each component of the receptor palette. This figure shows IP production in the presence of 107 M buserelin. The transfection of cells and measurement of IP production occurred as described under Materials and Methods.

Figures 4 (indoles), 5 (quinolones), and 6 (erythromycin macrolides) show the effect of the 1 µg/ml concentration of each of the drugs in rescuing each component of the receptor palette. In each figure, the data shows IP production in the presence of drug and 10⁷ M buserelin. For reference, Fig. 7 shows the unrescued coupling of the receptor in the absence (Fig. 7, upper image) or presence (Fig. 7, lower image) of 10⁷ M buserelin. In the absence of buserelin none of the rescue agents produced a response greater than basal levels for any vector. The member of each drug class with the lowest affinity for the human GnRHR is oriented closest to the viewer. This data allows the following conclusions to be made: 1) The efficacy of these drugs (measured by the ability of a fixed dose to rescue receptor) is consistent with the binding affinity of each class of molecule for the WT receptor; 2) peptidomimetics that were successful in rescuing one mutant usually rescued all mutants (that could be rescued by any of the peptidomimetics); and 3) particular mutants (human: S¹⁶⁸R and S²¹⁷R; rat des^260-265 -GnRHR and C²²⁹A-GnRHR) could not be rescued by any drug tried, suggesting that these were either grossly deformed or that loss-of-function was due to inability to bind ligand.

View larger version (35K): [in this window] [in a new window]   Fig. 6.   Erythromycin macrolides show the effect of the 1 µg/ml concentration of each member of the specified class of compounds in rescuing each component of the receptor palette. This figure shows IP production in the presence of 107 M buserelin. The transfection of cells and measurement of IP production occurred as described under Materials and Methods.

View larger version (19K): [in this window] [in a new window]   Fig. 7.   This figure shows the unrescued (no drug) coupling of receptor in the absence (upper image) or presence (lower image) of 107 M buserelin.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, we used a palette of WT and loss-of-function mutants of the GnRHR. These were selected as representatives of receptor truncations, sequence deletions, and point mutations at Cys residues, as well as all point mutants that have been reported to cause hypogonadotropic hypogonadism in humans (Achermann and Jameson, 2001; Janovick et al., 2002; Leaños-Miranda et al., 2002). This palette was used to assess the efficacy of chemically distinct drugs as pharmacoperones, structures that serve as molecular scaffolding, cause mutants to -fold correctly, and thereby avoid detection by the cellular quality control apparatuses. Our data indicate that structures from all three different chemical classes can rescue defective mutant folding and allow defective mutants to route to the plasma membrane, bind ligand, and couple to effectors. Accordingly, exploitation of this approach is not a unique feature of the indole class, a member of which class provided the first proof of principle for rescue of hypogonadotropic hypogonadism-causing mutations (Janovick et al., 2002). Efficacy of these drugs (measured by the ability of a fixed dose to rescue receptor) is proportional to the binding affinity of the molecules for the WT receptor. Particular mutants (human: S¹⁶⁸R and S²¹⁷R; rat des^260-265-GnRHR, and C²²⁹A-GnRHR) could not be rescued by any drug tried, suggesting that these were either grossly deformed or that loss-of-function resulted from the inability to bind ligand. Because addition of green fluorescent protein or HA-tags in themselves result in rescue, we have not been able to determine whether these particular loss-of-function mutants reside at the plasma membrane.

We were initially surprised by the lack of rescue specificity for different drugs (that is, all effective agents rescued virtually the same mutants). Because they were designed as GnRHR peptidomimetics, however, all of the drugs examined would be expected to compete with the natural ligand for these receptors, even though the precise locus of binding may differ slightly. Accordingly, it is conceivable that they serve as similarly located nuclei, stabilizing the ligand binding site of the mutants. This observation also leads to the consideration that the ability of the mutants to escape degradation by the quality control apparatus of processing may be, as an evolutionary matter, related to stability of the natural ligand binding site.

Alternatively, pharmacoperones that do not compete for the natural ligand binding site can certainly be envisioned since the existence of molecule-stabilizing allosteric interactions suggests that such compounds would exist (Conn et al., 2002). The ability to bind outside the natural ligand binding site may make such molecules advantageous as therapeutic agents because it might not be necessary to remove these before activating the receptor with an agonist. It is conceivable that such agents might show a more heterogeneous pattern of mutant rescue by interacting at diverse sites than the heterogeneous pattern measured for the peptidomimetics in the present study. Likewise, shipwrecking (Conn et al., 2002) compounds that interfere with the structure of wild type receptors may be highly heterogeneous for the same reason.

It is notable that the normal expression of the human GnRHR, but not the rat GnRHR, can be increased by the agents that were successful in rescuing receptors. It has become clear that the presence of Lys¹⁹¹ in the primate sequence (Arora et al., 1999; Maya-Núñez et al., 2000, 2002) limits the percentage of the synthesized receptor that reaches the plasma membrane to about 50% (Conn et al., 2002). The pharmacoperones function to override this inhibition, presumably due to structural stabilization, but are unable to do so in the rodent sequence, which lacks this "extra" amino acid and is already more efficiently routed to the membrane. The ability of all the rescue agents with high receptor binding affinity to increase the expression of the hWT sequence at the plasma membrane suggests that in vivo use of these agents as antagonists may first increase the expression of the receptor. This observation merits consideration since the therapeutic goal of antagonists is usually to decrease endocrine output in the cases of prostatic and breast cancer, for example. If the peptidomimetic, unlike their peptidic counterpart actually increases receptor expression, the results may be initial worsening of the disease if persistent occupancy of the receptor is not maintained since endogenous occupancy may activate the newly expressed receptors. A comparable event occurs in the cases of agonist therapy designed to desensitize the receptor, since there is an initial phase of clinical flare, apparently due to initial occupancy but before the development of desensitization.

Interestingly, the opioid receptor, also subject to desensitization, has also been reported to only process about half of the synthesized receptor to the plasma membrane (Petäjä-Repo et al., 2001), and concerns expressed about the significance of this event in the treatment of addiction with cell permeant receptor antagonists. The optimal drug dose needed for rescue appears, on one hand, to be governed by reaching a sufficient concentration but, on the other hand, is inhibited by excessive concentrations, as these cannot be completely washed out of the receptor to allow agonist occupancy.

The large number of disorders that are recognized as being caused by receptor misrouting (in contrast to inability to bind ligand or couple to effectors) has grown significantly in recent years (Conn et al., 2002). This indicates that misfolded receptors are a much more common disease etiology than was previously suspected and presents a new approach for pharmacological intervention. Accordingly, therapeutic approaches based on restoration of the receptor to its correct cellular locus now present a novel means of disease treatment, potentially applicable to hypogonadotropic hypogonadism, cystic fibrosis, nephrogenic diabetes insipidus, hypercholesterolemia, retinitis pigmentosa, cataracts, Alzheimer's, and other diseases (Conn et al., 2002). The data in this study provide the basis for structural design of pharmacoperones, as well as for "shipwrecking" agents (Conn et al., 2002), that cause the misrouting of proteins that would, in their absence, route correctly.