Actions of the Small Molecule Ligands SW106 and AH-3960 on the Type-1 Parathyroid Hormone Receptor

Abstract

The parathyroid hormone receptor-1 (PTHR1) plays critical roles in regulating blood calcium levels and bone metabolism and is thus of interest for small-molecule ligand development. Of the few small-molecule ligands reported for the PTHR1, most are of low affinity, and none has a well-defined mechanism of action. Here, we show that SW106 and AH-3960, compounds previously identified to act as an antagonist and agonist, respectively, on the PTHR1, each bind to PTHR1-delNT, a PTHR1 construct that lacks the large amino-terminal extracellular domain used for binding endogenous PTH peptide ligands, with the same micromolar affinity with which it binds to the intact PTHR1. SW106 antagonized PTHR1-mediated cAMP signaling induced by the peptide analog, M-PTH(1–11), as well as by the native PTH(1–9) sequence, as tethered to the extracellular end of transmembrane domain (TMD) helix-1 of the receptor. SW106, however, did not function as an inverse agonist on either PTHR1-H223R or PTHR1-T410P, which have activating mutations at the cytoplasmic ends of TMD helices 2 and 6, respectively. The overall data indicate that SW106 and AH-3960 each bind to the PTHR1 TMD region and likely to within an extracellularly exposed area that is occupied by the N-terminal residues of PTH peptides. Additionally, they suggest that the inhibitory effects of SW106 are limited to the extracellular portions of the TMD region that mediate interactions with agonist ligands but do not extend to receptor-activation determinants situated more deeply in the helical bundle. The study helps to elucidate potential mechanisms of small-molecule binding at the PTHR1.

The parathyroid hormone receptor-1 (PTHR1) is a family B G-protein-coupled receptor (GPCR) that mediates the actions of 2 endogenous peptide ligands: PTH, in the endocrine control of blood calcium levels, and PTH-related protein (PTHrP), in the paracrine control of cell differentiation programs in the skeleton and other developing tissues (1). As for each family B GPCR, the PTHR1 binds its peptide ligand, as represented by the PTH(1–34) fragment, via a 2-site mechanism, by which the (15–34) portion of the ligand first interacts with the receptor's amino-terminal extracellular domain (ECD) to provide binding energy to the complex, and the N-terminal(1–14) portion of the ligand interacts with the receptor's transmembrane domain (TMD) region containing the 7 membrane-spanning helices and interconnecting loops to induce the conformational changes involved in receptor activation (2, 3).

The specifics of the ECD component of the interaction for the PTHR1 have been revealed by high resolution crystal structures of the isolated ECD in complex with the (15–34) portion of either PTH (4) or PTHrP (5). Such a crystallographic analysis has not yet been reported for the TMD region of the PTHR1, and so information on ligand interactions occurring in this portion of the receptor is still limited to that which can be inferred from the available mutagenesis and cross-linking data (6–8). The recently reported crystal structures of the TMD regions of the glucagon receptor (GCGR) (9) and the corticotropin-releasing factor receptor type 1 (CRFR1) (10) provide the first 3-dimensional views of this region of any family B GPCR, and although these structures were obtained in complex with a small-molecule antagonist, rather than a cognate peptide ligand, they nevertheless reveal the general topographical features of the likely peptide-binding surface. Thus, each of the TMD structures showed a wide V-shaped cavity formed on the extracellularly exposed surface of the hepta-helical bundle, and this cavity would presumably accommodate the N-terminal, pharmacophoric portion of the peptide ligand (2, 11).

Previous studies on short N-terminal PTH fragment analogs demonstrate that ligand interactions to the TMD region of the PTHR1 can be sufficient to induce potent receptor activation and signal transduction responses. Analogs such as M-PTH(1–14) and M-PTH(1–11) thus exhibit the same potencies on PTHR1-delNT, a PTHR1-construct that lacks the ECD, as they do on the intact PTHR1 (12, 13). The importance of the N-terminal residues of PTH in mediating such receptor activation responses is further underscored by the lack of signaling potency in N-terminally truncated peptides, such as PTH(7–34) analogs (14), or in N-terminally intact peptides that have conserved valine-2 replaced by a bulky tryptophan or benzoyl-phenylalanine (Bpa) (15, 16). Such ligands can function not only as antagonists on the PTHR1 (14) but also, at least in some cases, as inverse agonist, and thus can suppress the high basal signaling of constitutively active PTHR1 mutants (15–19).

The development of small-molecule ligands that mimic the actions of the agonist peptides has been a challenging goal for the PTHR1, as it has been for each family B GPCR (3). For the PTHR1, a potent, orally active agonist compound would be of considerable interest, because it could lead to better treatments for diseases such as osteoporosis (20) and hypoparathyroidism (21). So far, however, no such ligand has been reported. Of the few nonpeptide PTHR1 ligands that have been described, only one, AH-3960, was found to exhibit agonist activity, albeit with a potency for stimulating cAMP responses in cells that is several orders of magnitude weaker than that of PTH(1–34) (EC_50s = ∼1 × 10⁻⁴M vs 1 × 10⁻⁷M) (22). The other small-molecule ligands reported for the PTHR1 were found to act as antagonists and mostly with affinities in the micromolar range (23, 24).

The mechanisms by which any of the identified small-molecule PTHR1 ligand binds to the receptor remain largely undefined. One antagonist compound, however, called SW106, was identified by its capacity to inhibit the binding of a modified PTH(1–14) radioligand to the intact PTHR1, and so this compound can at least be hypothesized to interact with the receptor's TMD region (24). Elucidating the mechanisms by which any nonpeptide ligand interacts with the PTHR1 could lead not only to a better understanding of the structure-activity relationships that operate at this receptor but also to the development of more potent mimetics and ultimately to new therapeutic approaches. We therefore sought to define further the mechanisms by which SW106 and AH-3960 interact with the PTHR1. We used a series of pharmacological assays employing various PTHR1 constructs and PTH peptide analog probes to assess modes of binding and signal modulation. We thus could specifically test the hypothesis that the compounds bind to the TMD region of the receptor and also assess whether or not SW106 functions as an inverse agonist on constitutively active PTHR1 mutants.

Materials and Methods

Peptides

All PTH and PTHrP peptide used were based on the human amino acid peptide sequence except where the species is indicated, and each contained a C-terminal amide. The analogs used were: PTH(1–34); [Ala^1,12,Aib³,Nle⁸,Gln¹⁰,Har¹¹,Trp¹⁴,Tyr¹⁵]-PTH(1–15) (M-PTH(1–15)); [ACPC¹,Aib³,Gln¹⁰,Har¹¹]-PTH(1–11) (M-PTH-11)); [Deg^1,3,Trp²,Gln¹⁰,Har¹¹,Ala¹²,Trp¹⁴]-PTH(1–14) (Trp²-M-PTH(1–14)); [Deg^1,3,Bpa²,Gln¹⁰,Har¹¹,Ala¹²,Trp¹⁴]-PTH(1–14) (Bpa²-M-PTH(1–14)); [DTrp¹²,Tyr³⁴]-bovine PTH(7–34), (DTrp¹²-PTH(7–34)) and [Leu¹¹,DTrp¹²,Trp²³,Tyr³⁶]-PTHrP(7–36) (DTrp¹²-PTHrP(7–36)), in which ACPC is aminocyclopentane carboxylic acid, Deg is diethylglycine, and Aib is α-aminoisobutyric acid. Radioligands used were ¹²⁵I-PTH(1–34) (¹²⁵I-[Nle^8,21,Tyr³⁴]-rat PTH(1–34)) and ¹²⁵I-M-PTH(1–15) and were prepared by chloramine-T-based radioiodination followed by reversed-phase HPLC purification.

Compound synthesis and handling

SW106 and AH-3960 were synthesized based on published procedures (25, 26). Both compounds were verified for purity (HPLC, ¹H-NMR) and identity (Mass-spec, ¹H-NMR). Compounds were maintained as 100mM stock solutions in dimethyl sulfoxide (DMSO); mild turbidity appearing in initial 300-fold dilutions in aqueous assay buffer dissipated with brief mixing and not apparent in subsequent serial dilutions.

Cell transfection and PTHR1 plasmids

Cells were transfected with plasmid DNA encoding either the wild-type human PTHR1 (PTHR1 or PTHR1-wild-type [WT]); a hPTHR1 derivative, including PTHR1-delNT, in which Leu²⁵ just after the predicted signal sequence cleavage site at Tyr²³/Ala²⁴ is joined via a -Gly-Ser-Gly- linker to Glu¹⁸² of helix-1 of the TMD1 region; PTHR-Tether(1–9), in which the prepro leader sequence of human PTH, followed by the rat PTH(1–9) sequence (Ala-Val-Ser-Glu-Ile-Gln-Leu-Met-His), is joined at His⁹ via a Gly₍₄₎ linker to Glu¹⁸² of the hPTHR1 (17); the constitutively active mutants: PTHR1-H223R, PTHR1-T410P, or PTHR1-H223R/T410P (PTHR1-HR/TP), the latter containing both the H223R and T410P activating mutations, or the human PTHR2. The plasmid vector used for expressing PTHR1-WT and PTHR1-delNT was pCDNA3.1(+), and that for all other PTHR1 constructs was pCDNA1. For binding assays, COS-7 cells were transiently transfected to express either PTHR1-WT or PTHR1-delNT. For signaling assays, an HEK-293 cell line (GS-22A) was prepared that stably expresses the luciferase-based pGlosensor-22F (Glosensor) cAMP reporter plasmid (Promega Corp) (27). The GS-22A cells were further transfected to stably express either PTHR1-WT (GP-2.3 cells), PTHR1-delNT (GD-5 cells), or PTHR1-H223R (GHR-10 cells). In some experiments, GS-22A cells were transiently transfected to express either the PTHR1-WT or a PTHR1 mutant.

PTHR1 binding

Binding to the PTHR1 was assessed using ¹²⁵I-PTH(1–34) or ¹²⁵I-M-PTH(1–15) tracer radioligands and membranes prepared from PTHR1-transfected cells, as described (28). For competition assays, membranes were prepared from COS-7 cells transiently transfected to express PTHR1 or PTHR1-delNT, along with a Gαs mutant (Gα_S^ND), which enriches for high affinity, G protein-coupled receptor states, and thus promotes binding of the ¹²⁵I-M-PTH(1–15) radioligand to the PTHR1 in COS-7 cell membranes (28). Binding reactions in 96-well plates were incubated at room temperature for 90 minutes, then processed by vacuum filtration, and after washing, the filters were removed and counted for γ irradiation. Nonspecific binding was determined in reactions containing an excess (5 × 10⁻⁷ M) of unlabeled M-PTH(1–15) or PTH(1–34). Curves were fit to the data using a 4-parameter sigmoidal dose-response equation.

For dissociation time-course assays, ¹²⁵I-M-PTH(1–15) or ¹²⁵I-PTH(1–34) radioligands were prebound to membranes prepared from GP-2.3 cells for 90 minutes; then, an excess (5 × 10⁻⁷M) of unlabeled M-PTH(1–15) or rPTH(1–34) was added, either alone (control) or in the presence of a candidate modulating agent. At times thereafter, and immediately before ligand addition (T = 0), aliquots were withdrawn, vacuum filtered, and, after washing, the filters were removed and counted for bound radioactivity. Nonspecific binding was determined in reactions containing an excess (5 × 10⁻⁷M) of unlabeled M-PTH(1–15) or PTH(1–34) during the prebinding phase. Curves were fit to the data using a biphasic decay equation, in which the rapid phase component (Span-1) was constrained to a t_1/2 of less than 5 minutes; the slow phase component (Span-2) was constrained to a t_1/2 of more than 5 minutes, and the plateau was constrained to 0 (28).

cAMP signaling

Signaling via the cAMP/protein kinase-A pathway was assessed using HEK-293-derived cells stably expressing the Glosensor cAMP reporter plasmid, along with a PTHR derivative, expressed either stably or transiently. For assays, confluent cells in 96-well plates were pretreated with luciferin for 20 minutes at room temperature, and then a test ligand or medium (vehicle) was added and incubations continued for an additional period of up to 90 minutes. Luminescence arising from intracellular cAMP binding to the Glosensor reporter protein was measured in real-time at 2-minute intervals during both the pretreatment and ligand-addition phases using a PerkinElmer Envision plate reader. Ligand dose-response curves were generated using the data obtained at the peak response time, typically by 20 minutes after ligand addition. For antagonism experiments, the agonist and antagonist were added simultaneously, unless, as indicated in the figure legend, the antagonist was added 15 minutes before the agonist.

PLC/IP₃ signaling

GP-2.3 cells were treated with ligand reagents for 30 minutes at 37°C, and then intracellular levels of inositol monophosphate (IP₁), a stable metabolite of IP₃, were quantified using an antibody-based, homogeneous time-resolved fluorescence assay (IP-One Tb kit, CisBio).

Data calculations

Data were processed using Excel 2008 (Microsoft Corp) and Prism 5.0 (GraphPad Software, Inc). Aggregate data are expressed as mean ± SEM. Curves were fit to the data using a 4-parameter, nonlinear regression function. Statistical analyses were performed using the Student's t test (two-tailed, unequal variances), significance was inferred from P values of less than 0.05.

Results

Mechanisms of compound binding to the PTHR1

We first sought to explore the mechanisms of binding used by the 2 compounds, SW106 and AH-3960, at the PTHR1, and in particular, the relative roles of the TMD vs the ECD regions of the receptor in mediating such binding. We thus performed radioligand competition assays in membranes prepared from COS-7 cells transiently transfected to express either the intact PTHR1 or PTHR1-delNt, a receptor construct that lacks nearly the entire ECD region. Two PTH-based tracer radioligands were used: ¹²⁵I-PTH(1–34), which requires interactions to both the ECD and TMD regions for high affinity binding, and thus binds detectably only to the intact PTHR1, and ¹²⁵I-M-PTH(1–15), which interacts only with the receptor's TMD, as enhanced by the incorporated “M” substitutions, and thus binds detectably to both the PTHR1 and to PTHR1-delNT (12, 13). On the intact PTHR1, SW106 and AH-3960, at the highest concentration tested (100μM), each only partially (∼40%) inhibited the binding of ¹²⁵I-PTH(1–34), whereas they more completely (∼90%) inhibited binding of ¹²⁵I-M-PTH(1–15), and with apparent affinities (IC₅₀s) in the low- to mid-μM range (Figure 1, A and B, and Table 1). On PTHR1-delNT, each compound inhibited binding of ¹²⁵I-M-PTH(1–15) with a maximum efficacy and apparent affinity that was similar to that observed for each compound on the intact PTHR1 as assessed with the ¹²⁵I-M-PTH (1–15) radioligand (Figure 1C). The control peptide M-PTH(1–15) also inhibited binding of ¹²⁵I-M-PTH(1–15) to PTHR1-delNT and to the intact PTHR1 with comparable affinities, whereas PTH(1–34) exhibited an affinity on PTHR1-delNT that was approximately 700-fold weaker than that on the intact PTHR1 (Figure 1, B and C, and Table 1).

Figure 1.

Ligand binding to PTHR1-WT and PTHR1-delNT. Radioligand competition assays were performed in membranes prepared from COS-7 cells transiently transfected to express either the intact wild-type (WT) PTHR1 (A and B) or PTHR1-delNt (C); tracer radioligands were either ¹²⁵I-PTH(1–34) (A) or ¹²⁵I-M-PTH(1–15) (B and C). Reactions contained varying concentrations of the unlabeled ligands: PTH(1–34), M-PTH(1–15), SW106, and AH3960. Schematics at the top show the structures of SW106 and AH-3960, and those to the right of each graph illustrate the PTHR1 construct used in the assay as well as the radioligand and its mode of binding. Note that the binding of ¹²⁵I-PTH(1–34) to PTHR1-delNT was too weak to be detected. Data are mean ± SEM of 3 or more experiments, each performed in duplicate. Binding curve-fit parameters are reported in Table 1.

Table 1.

Ligand-Binding to PTHR1-WT and PTHR1-delNt

	WT/125I-PTH(1–34)		WT/125I-M-PTH(1–15)		delNT/125I-M-PTH(1–15)		n
	pIC50/nM	Min. (%)	pIC50/nM	Min. (%)	pIC50/nM	Min. (%)
M-PTH(1–15)	8.13 ± 0.08	11 ± 3	9.78 ± 0.11	1 ± 1	9.63 ± 0.08	1 ± 0	12
	7.33 nM		0.17 nM		0.23 nM
SW106	5.70 ± 0.12	52 ± 6	5.32 ± 0.14	16 ± 3	5.47 ± 0.04	5 ± 3	10
	2008 nM		4733 nM		3363 nM
AH-3960	4.03 ± 0.29	10 ± 7	4.78 ± 0.22	10 ± 7	5.45 ± 0.13	23 ± 7	7
	93 513 nM		16 509 nM		3555 nM
Trp2-M-PTH(1–14)	5.32 ± 0.32	21 ± 13	5.81 ± 0.18	4 ± 2	5.25 ± 0.13	9 ± 3	7
	4775 nM		1565 nM		5679 nM
Bpa2-M-PTH(1–14)	5.60 ± 0.16	6 ± 3	5.96 ± 0.1	0 ± 0	5.57 ± 0.11	1 ± 1	5
	2515 nM		1094 nM		2714 nM
PTH(1–34)	9.38 ± 0.07	0 ± 0	10.07 ± 0.06	0 ± 0	7.23 ± 0.05	3 ± 2	6
	0.421 nM		0.086 nM		59 nM

Assays were performed in membranes prepared from COS-7 cells transiently cotransfected to express either hPTHR1-WT or hPTHR1-delNt, each with a high-affinity Gs mutant. Binding inhibition constants are expressed as -log M (pIC₅₀) and corresponding nM value below in italic. Min. (%), fit-derived minimum of radioligand binding at maximum ligand concentration. Data are mean ± SEM of the number experiments, indicated (n).

Although such competition data cannot rule out the possibility that the compounds modulate PTH radioligand binding via indirect, allosteric effects, they are at least consistent with a more direct, competitive mode of inhibition. That SW106 and AH-3960 each bound to PTHR1-delNT with an affinity that was similar to the affinity with which it bound to the intact PTHR1, as assessed with ¹²⁵I-M-PTH(1–15), suggests that each compound, like M-PTH(1–15), interacts mainly, if not solely, with the receptor's TMD region. The failure of each compound to completely inhibit the binding of ¹²⁵I-PTH(1–34) to the intact receptor is consistent with this hypothesis, because it likely reflects the absence of direct inhibitory effects of the compounds on binding interactions that occur between the ECD portion of the receptor and the (15–34) portion of PTH(1–34) (5, 29).

Effects of compounds on cAMP signaling at the PTHR1

We next assessed the capacities of AH-3960 and SW106 to modulate cAMP signaling at the PTHR1. We used for these studies the luciferase-based “Glosensor” cAMP reporter (27), which we stably expressed in HEK-293 cells to obtain the cell line GS-22A. The GS-22A cells were further transfected with either vector-plasmid DNA or with a PTHR plasmid of interest. GS-22A cells transiently transfected with plasmid vector DNA mediated a robust cAMP response to isoproterenol, which activates β₂-adrenergic receptors endogenous to HEK-293 cells, but they exhibited almost no response to PTH(1–34), AH-3960, or SW106 (Figure 2A). In contrast, GS-22A cells transiently transfected to express the PTHR1 exhibited a robust and potent cAMP response to PTH(1–34) (maximal luminescence increase vs basal of ∼60-fold at 1 × 10⁻⁸M ligand; EC₅₀ = ∼0.2nM), and a considerably weaker, yet significant response to AH-3960 (maximal increase vs basal of ∼22-fold at 3 × 10⁻⁵M ligand; P = .0007) (Figure 2B). A small, yet significant increase in cAMP was also detected with SW106 (maximal increase vs basal of ∼5-fold at 3 × 10⁻⁵M ligand; P = .02) (Figure 2B); thus, SW106, previously characterized as a PTHR1 antagonist (24), has some, albeit very weak, intrinsic agonist activity on the PTHR1.

Figure 2.

Ligand effects on cAMP signaling and receptor specificity. Ligand-induced cAMP responses were assessed in GS-22A cells, which are HEK-293-derived cells that stably express the luciferase-based Glosensor cAMP reporter; the cells were further transiently transfected with either pcDNA3.1 vector DNA (A) or with a plasmid encoding the PTHR1 (B). Responses were also assessed in GP-2.3 cells, which are GS-22A-derived cells transfected to stably express the PTHR1 (C), and in G2R-5 cells, which are GS-22A cells transfected to stably express the PTHR2 (D). Cells were treated with varying concentrations of ligand for 15–20 minutes, and cAMP-dependent luminescence was recorded. Ligands used were AH-3960, SW106, PTH(1–34), M-PTH(1–11), TIP39, which is a PTHR2-specific agonist peptide, and isoproterenol, which activates endogenous β₂-adrenergic receptors. Data are mean ± SEM of 3 experiments, each performed in duplicate.

The specificity of the cAMP signaling actions of AH-3960 was further explored using GS-22A-derived cells that were stably transfected to express either the PTHR1 (GP-2.3 cells) or the PTHR2 (G2R-5 cells). PTH(1–34) induced potent, dose-dependent increases in cAMP in both GP-2.3 cells and G2R-5 cells (EC₅₀s = ∼0.3nM and 2.3nM, respectively), whereas TIP39, the cognate peptide ligand for the PTHR2 (30), induced a robust and potent (EC₅₀ = ∼0.2nM) agonist response in G2R-5 cells but exhibited little or no activity in GP-2.3 cells (Figure 2, C and D). These cAMP responses observed for PTH(1–34) and TIP39 in the Glosensor-expressing cells are fully consistent with the selectivity profiles previously reported for these peptide ligands acting on the 2 PTHR subtypes (30). AH-3960, over the dose range of 1μM to 300μM, induced clear and dose-dependent increases in cAMP in GP-2.3 cells (maximum increase vs basal = 73-fold, P = .002) (Figure 2 C) but was much less active in G2R-5 cells (maximum increase vs basal = 12-fold, P = .02) (Figure 2D). M-PTH(1–11) was a potent agonist in GP-2.3 cells but, like AH-3960, was considerably less active in G2R-5 cells. These findings confirm that AH-3960 functions as a low-affinity agonist at the PTHR1 (22) and further extend the specificity profile of the compound by showing that it is considerably less potent on the PTHR2 than on the PTHR1.

Antagonist actions of SW106 on agonist-induced cAMP signaling at the PTHR1

We next evaluated the capacity of SW106 to function as an antagonist of ligand-induced cAMP signaling at the PTHR1 and, more specifically, the relative roles that the receptor's ECD vs TMD regions play in mediating any such antagonist action. Thus, we assessed the capacity of the compound to shift the cAMP dose-response curves generated by several different agonist ligands in GP-2.3 cells (stable expression of PTHR1) as compared with in GD-5 cells, which are GS-22A-derived cells that stably express PTHR1-delNT. In GP-2.3 cells, SW106 (10μM) had little or no effect on the potency curve obtained for PTH(1–34) (pEC₅₀ values in the absence and presence of SW106 were 9.32 ± 0.26 and 9.46 ± 0.26, respectively, P = .71; corresponding E_Max. values were 303 ± 38 counts per second [cps] × 10⁻³ and 271 ± 39 cps × 10⁻³, respectively) (Figure 3A). In contrast, the compound caused a marked rightward shift in the PTH(1–34) potency curve obtained in GD-5 cells (Figure 3B) (note that the ∼700-fold weaker potency that PTH(1–34) exhibited in GD-5 cells, as compared with in GP-2.3 cells, is again explained by the absence of interactions between the (15–34) portion of the ligand and the ECD). Analyses of the time course of cAMP (luminescence) development in GP-2.3 cells and GD-5 cells treated with PTH(1–34) at a submaximal concentration (1 × 10⁻⁹M and 3 × 10⁻⁷M, respectively) in the absence vs presence of SW106 further highlighted the difference in the capacity of the compound to inhibit signaling induced by PTH(1–34) in the 2 cell lines (Figure 3, A and B, insets).

Figure 3.

Antagonist effects of SW106 on cAMP signaling responses in cells expressing the intact wild-type (WT) PTHR1 or PTHR1-delNT. The effects of SW106 (1 × 10⁻⁵M) on cAMP dose-response curves obtained for different agonist ligands were assessed in either GP-2.3 cells (stable Glosensor and PTHR1) (A and C), or in GD-5 cells (stable Glosensor and PTHR1-delNT) (B and D). The cells were preloaded with luciferin and then treated with varying concentrations of an agonist ligand with or without SW106 (1 × 10⁻⁵M) and peak luminescence signals (cps × 10⁻³; observed within 20 min of ligand addition) were measured. SW106 caused rightward shifts in the agonist-dose response curves obtained for PTH(1–34) on PTHR1-delNT (B) and for M-PTH(1–11) and AH-3960 on the PTHR1 (C) and PTHR1-delNT (D), but it had little or no effect on the isoproterenol curves. Insets in A and B show the time course of luminescence development in the corresponding cells after treatment with vehicle or PTH(1–34) (1 × 10⁻⁹M, A; 3 × 10⁻⁷M, B), with or without SW106. In A, the maximum luminescence signals in vehicle-treated cells in the absence and presence of SW106 were 3606 ± 447 and 8738 ± 778 cps, respectively (P < .0001); those in B were 1348 ± 104 and 1456 ± 105 cps (P = .5); those in C were 1602 ± 119 and 2090 ± 197 cps (P = .04); and those in D, determined only in the absence of SW106, were 495 ± 65 cps. The inhibitory actions of SW106 were further assessed by comparing the effects of 3 different concentrations of the compound, vs vehicle control, on the cAMP dose-response curve obtained for M-PTH(1–11) in GP-2.3 cells (E); replotting of the responses observed in these experiments at the near-maximal concentration (1 × 10⁻⁹M) of M-PTH(1–11) confirmed the dose dependency of SW106 inhibition (F). The effects of adding SW106 (1 × 10⁻⁵) either concomitantly with an agonist, as done in A–F, or 15 minutes before the agonist were assessed in GP-2.3 cells for PTH(1–34) (G) and M-PTH(1–11) (H); the pEC₅₀ values obtained for PTH(1–34) without SW106, with previous SW106, and with concomitant SW106 were: 9.77 ± 0.23; 9.55 ± 0.25 (P = .6) and 9.58 ± 0.26 (P vs no SW106 = 0.6; P vs previous SW106 = 0.9); and the corresponding pEC₅₀ values for M-PTH(1–11) were: 9.98 ± 0.13, 8.36 ± 0.02 (P = .006), and 8.77 ± 0.03 (P vs no SW106 = 0.009; P vs previous SW106 = 0.007). Data are mean ± SEM of 5 (A and B), 4 (C and D), or 3 (E–H) experiments, each performed in duplicate.

SW106 caused a marked reduction in the cAMP signaling potency of M-PTH(1–11) in both the GP-2.3 and GD-5 cells (pEC₅₀s in GP-2.3 cells were 9.86 ± 0.06 vs 8.87 ± 0.06 in the absence vs presence of SW106, respectively, P < .0001, and corresponding E_Max. values were 305 ± 20 × 10⁻³ vs 268 ± 20 cps × 10⁻³, P = .7; pEC₅₀s in GD-5 cells were 9.83 ± 0.12 and 8.93 ± 0.15 in the absence and presence of SW106, respectively, P = .004, and corresponding E_Max. values were 134 ± 10 × 10⁻³ and 128 ± 9 cps × 10⁻³, P = .7) (Figure 3, C and D). SW106 had little or no effect on the cAMP signaling responses obtained for isoproterenol in either GP-2.3 cells or GD-5 cells, whereas it caused a moderate rightward shift in the dose-response curves obtained in these cells for AH-3960. The antagonist effect of SW106 on cAMP signaling induced by M-PTH(1–11) in GP-2.3 cells was dose dependent, and thus characterized by an IC₅₀ of about 3 × 10⁻⁵M (Figure 3, E and F). Separate studies performed in GP-2.3 cells revealed that adding SW106 to the cells 15 minutes before the agonist, PTH(1–34) or M-PTH(1–11), modestly enhanced the antagonist effect of the compound, as compared with adding SW106 concomitantly with agonist, as done in the experiments of Figure 3, A–F. However, enhancement was only significant for M-PTH(1–11) (Figure 3, G and H).

Although these studies cannot rule out the possibility of inhibition by indirect or allosteric mechanisms, the finding that SW106 caused approximately parallel, rightward shifts in the cAMP dose-responses curve obtained for M-PTH(1–11) in both GP-2.3 and GD-5 cells but had no effect on the cAMP signaling actions of isoproterenol supports the conclusion that the compound acts as a specific, competitive inhibitor of agonist-induced cAMP signaling at the PTHR1 (24). They furthermore are consistent with the interpretation that the compound mediates these antagonist actions by binding to the receptor's TMD region. The above results also confirm that AH-3960 acts as a weak agonist at the PTHR1 (22) and that it mediates these agonist actions by interacting with the PTHR1 TMD region (Figure 3, C and D).

Antagonist actions of SW106 on agonist-induced PLC/IP₃ signaling at the PTHR1

We then assessed the capacity of SW106 to inhibit PTHR1-mediated signaling via the PLC/IP₃ pathway. Signaling via this pathway was assessed by measuring intracellular levels of IP₁, a stable metabolite of inositol triphosphate. GP-2.3 cells were thus treated for 30 minutes with either vehicle or a test agonist ligand, each in the absence or presence of SW106 (1 × 10⁻⁵M). Figure 4A shows that SW106 caused a modest reduction in the IP₁ response induced by PTH(1–34) (1 × 10⁻⁷M) and more significant reductions in the IP₁ responses induced by M-PTH(1–14) (1 × 10⁻⁷M) and M-PTH(1–11) (1 × 10⁻⁶M). The compound also caused a modest (27%) reduction in IP₁ levels in cells treated with ATP (1 × 10⁻⁵M), which activates PLC signaling via endogenous P2Y purinergic receptors, but this effect did not reach significance. Thus, as was found for its effects on PTH ligand-induced cAMP signaling, SW106 was more effective in inhibiting PLC signaling responses induced by shorter-length N-terminal PTH peptide analogs than in it was in inhibiting PLC responses induced by PTH(1–34). The inhibitory effects of SW106 on the M-PTH(1-11)-induced PLC signaling were dose dependent and thus characterized by an IC₅₀ value of about 1 × 10⁻⁶M (Figure 4B).

Figure 4.

Antagonist effects of SW106 on PLC/IP₃ signaling. The capacity of SW106 to inhibit agonist-induced signaling via the PLC/IP₃ pathway was assessed in GP-2.3 cells. The cells were stimulated with vehicle (veh), PTH(1–34) (1 × 10⁻⁷M), M-PTH(1–14) (1 × 10⁻⁷M), M-PTH(1–11) (1 × 10⁻⁶M), or ATP, which activates endogenous PY2 receptors (1 × 10⁻⁵M), each in the absence or presence of SW106 (1 × 10⁻⁵M) for 30 minutes at 37°C, and cellular concentrations of IP₁, a stable IP₃ metabolite, were measured (A). Dose dependency of SW106 inhibition of IP₁ formation was assessed by treating cells with M-PTH(1–11) (1 × 10⁻⁷M) or vehicle each in the absence or presence of SW106 at varying concentrations (B). Data are mean ± SEM of 3 experiments, each performed in duplicate; P vs the same agonist ligand without SW106: *, P < .05; **, P < .01.

Ligand effects on constitutively active PTHR1 mutants

We then explored the capacity of SW106 to function as an inverse agonist at constitutively active mutant PTHRs. We prepared for these studies a line of GS-22A-derived cells, called GHR-10, that stably express the constitutively active PTHR1-H223R mutant along with the Glosensor cAMP reporter. Real-time analysis of basal intracellular cAMP formation, as revealed by the development of luciferase-derived luminescence initiated by the addition of the luciferin substrate, revealed that ligand-independent cAMP formation rates were higher in GHR-10 cells than in either GP-2.3 cells or the parental GS-22A cells (Figure 5A). Upon the addition of PTH(1–34) (100nM), the maximum cAMP level attained (∼20 min after ligand addition) was lower in GHR-10 cells than in GP-2.3 cells (Figure 5B). The agonist actions of PTH(1–34) in GHR-10 cells were dose dependent (Figure 5C). Both DTrp¹²-PTHrP(7–36) and Bpa²,M-PTH(1–14) induced dose-dependent reductions in cAMP levels, consistent with the known inverse agonist actions of these ligands (31), whereas SW106 resulted in no consistent change in the cAMP signaling activity of PTHR1-H223R, as assessed in these GHR-10 cells (Figure 5D).

Figure 5.

Effects on cAMP signaling in cells stably expressing PTHR1-H223R. Basal rates of cAMP formation were assessed in GHR-10 cells, which stably express the Glosensor cAMP reporter along with the constitutively active mutant, PTHR1–H223R, in GP-2.3 cells, which stably express the Glosensor reporter and PTHR1-WT, and in parental GS-22A cells (control), which stably express only the Glosensor reporter; cAMP-dependent luminescence (as cps) was measured for 24 minutes after the addition of luciferin (A). The cells were then treated with PTH(1–34) (1 × 10⁻⁷M), and luminescence was monitored for an additional 2 hours (B). Ligand dose-response effects on cAMP signaling via PTHR1–H223R were assessed in GHR-10 cells for PTH(1–34) (C) and several test inverse agonist ligands: SW106, DTrp¹²-PTHrP(7–36), or Bpa²-M-PTH(1–14) (D); the cells were preloaded with luciferin and then treated for 10 minutes with vehicle or varying concentrations of ligand. Asterisks in D indicate statistical comparisons with vehicle-treated cells: *, P < .05; **, P < .001. Data are mean ± SEM of 5 experiments, each performed in duplicate.

We further assessed the inverse agonist potential of SW106 using 2 additional constitutively active PTHR1 mutants, PTHR1-T410P, and the double-mutant, PTHR1-HR/TP, which contains both the H223R and T410P mutations and has been shown to exhibit considerably higher constitutive cAMP signaling activity than either single mutant receptor alone (18). These receptors, in parallel with the wild-type PTHR1 and PTHR1-H223R, were transiently transfected into GS-22A, cells and their basal and ligand-stimulated cAMP signaling actions were assessed. As expected, the basal cAMP signaling levels in cells transfected with PTHR1-HR/TP were considerably (∼8-fold) higher than those in cells expressing either single mutant receptor, whereas the basal cAMP signals in cells expressing PTHR1-H223R or PTHR1-T410P were about 2-fold higher than those in cells transfected with wild-type PTHR1 (Figure 6A).

Figure 6.

Time-course analyses of basal and ligand-modulated cAMP production in cells transiently transfected to express wild-type or a constitutively active mutant PTHR1. GS-22A cells transiently transfected to express either PTHR1-WT, PTHR1–H223R, PTHR1–T410P, or PTHR1–H223R/T410 were treated with luciferin, and ligand-independent (basal) cAMP-dependent luminescence was recorded for 24 minutes (A; insets show the graph with rescaled y-axis, and a schematic indicates the locations of the mutations). The same cells were then (t = 0) treated with vehicle, PTH(1–34) (1 × 10⁻⁷M), or a test inverse agonist: SW106 (1 × 10⁻⁵M), DTrp¹²-PTHrP(7–36) (1 × 10⁻⁶M), Trp²-M-PTH(1–14) (1 × 10⁻⁵M), and cAMP levels were assessed for an additional 52 minutes (B–F). Data are mean ± SEM of 9 experiments; in each experiment, the preligand (basal) measurements were obtained from 24 replicate wells (A) and the postligand measurements from duplicate wells (B–F).

After the addition of PTH(1–34) (1 × 10⁻⁷M), the cAMP signal increased maximally, at about 15 minutes and relative to the corresponding signal in vehicle-treated cells, by 51-fold in cells expressing the wild-type PTHR1, by 12-fold in cells expressing PTHR1-H223R, by 24-fold in cells expressing PTHR1-T410P, and, after a small initial increase, it decreased by as much as 23% in cells expressing PTHR1-HR/TP (Figure 6B; note that the corresponding responses obtained in these experiments in vehicle-treated cells for each receptor are displayed in Figure 6, C–F). The cAMP signaling profiles revealed by these Glosensor-based time-course analyses for the 3 constitutively active receptors in both the basal- and PTH(1–34)-stimulated states, including the paradoxical lowering of cAMP by PTH(1–34) with PTHR1-HR/TP, are consistent with previous comparative studies on these receptors, in which intracellular cAMP signaling was measured directly by RIA (18, 31).

Panels C–F of Figure 6 show time-course responses observed in the same experiment in the cells treated with a test antagonist or inverse agonist ligand. Of note, SW106 (1 × 10⁻⁵M) induced partial agonist responses (relative to vehicle and PTH(1–34) treatments) in cells expressing PTHR1, PTHR1-H223R, or PTHR1-T410P and little or no change in cAMP levels in cells expressing PTHR1-HR/TP (Figure 6, C–F). Thus, consistent with data shown in Figures 5D and 2B, SW106 did not function as an inverse agonist on any of these mutant receptors but rather tended to exhibit partial agonist behavior. In contrast, DTrp¹²-PTHrP(7–36) (1 × 10⁻⁶M), which induced a transient partial agonist response in cells expressing the PTHR1 (Figure 6C), induced the expected inverse agonist response on each the 3 constitutively active receptors (Figure 6, D–F). The analog Trp²-M-PTH(1–14) exhibited partial agonism on the wild-type PTHR1, inverse agonism on PTHR1-H223R and PTHR1-HR/TP, and had little or no effect on signaling in cells expressing PTHR1-T410P. These findings thus confirm the H223R-selective inverse agonist properties of N-terminally intact PTH analogs containing a Trp or Bpa substitution at position-2 as well as the nonselective inverse agonist actions of DTrp¹²-containing PTHrP(7–36) analogs (17, 18).

A separate series of experiments were performed to evaluate the effects of AH-3960, with several other control ligands, on cAMP signaling by these constitutively active PTHR mutants. Consistent with the data shown in Figure 6, time-course analyses revealed that DTrp¹²-PTH(7–34) functioned as an inverse agonist on each constitutively active receptor, Trp²-M-PTH(1–14) and Bpa²-M-PTH(1–14) functioned as inverse agonists on PTHR1-H223R and PTHR1-HR/TP, and SW106 functioned as a weak partial agonist on each receptor (Figure 7, A–D). Interestingly, AH-3960 exhibited agonist activity not only on PTHR1 and on each single mutant receptor, but also on PTHR1-HR/TP, and indeed was the only ligand to induce an appreciable increase in cAMP at the double-mutant receptor. Dose-response analyses confirmed that AH-3960 acted as an agonist at the double-mutant receptor and indeed resulted in maximum cAMP levels that were, if anything, greater than those attained with PTH(1–34) acting on the wild-type PTHR1 (Figure 7, E and F). This finding is notable in that AH-3960 is the only ligand found to date to function as a clear cAMP agonist on PTHR1-HR/TP. Although the mechanistic basis for the effect of any ligand at these constitutively active receptors remains largely unclear, the robust activity that AH-3960 exhibited on PTHR1-HR/TP seems at least consistent with the notion that constitutively active GPCRs, in general, can potentially offer a sensitive means for detecting agonist activities in otherwise weakly active ligands (32).

Figure 7.

Ligand effects AH-3960 on cAMP signaling by wild-type and constitutively active PTHR1 mutants. Time-course analyses were performed to determine the effects of AH-3960 and additional control ligands on cAMP production in GS-22A cells transiently transfected to express either PTHR1-WT (A), PTHR1–H223R (B), PTHR1–T410P (C), or PTHR1–H223R/T410 (D). The cells were preloaded with luciferin for 20 minutes and then (t = 0) treated with vehicle, PTH(1–34) (1 × 10⁻⁷M), AH-3960 (1 × 10⁻⁵M), SW106 (1 × 10⁻⁵M), DTrp¹²-PTH(7–34) (1 × 10⁻⁶M), Trp²-M-PTH(1–14) (1 × 10⁻⁵M), or Bpa²-M-PTH(1–14) (1 × 10⁻⁵M), and cAMP-dependent luminescence was assessed for 90 minutes (insets of A–C show the graphs with rescaled y-axes). Dose-response experiments were performed for PTH(1–34) (E) and AH-3960 (F) in GS-22A cells transiently transfected to express either the wild-type PTHR1 or the indicated constitutively active mutant receptor; plotted are the peak cAMP-dependent luminescence signals observed over the time course for each ligand concentration. Data are mean ± SEM of 3 experiments, each performed in triplicate.

Effects of SW106 on cAMP signaling by PTHR-Tether(1–9)

We then tested the capacity of SW106 to inhibit the signaling activity of a PTHR1-delNT derivative in which the PTH(1–9) sequence is recombinantly tethered to the extracellular end of TMD helix-1. In this construct, His⁹ of the PTH sequence is joined via a tetra-glycine linker to Glu¹⁸² at the extracellular end of helix-1, and Ala¹ of the PTH sequence is the free N terminus (33, 34). When transiently transfected into GS-22A cells, and in the absence of exogenous ligand, this PTHR-Tether(1–9) construct produced a cAMP-dependent luminescence signal that was several orders of magnitude greater than the basal level seen in cells transfected with PTHR1 or with control vector (Figure 8A). Consistent with previous studies (33), the addition of Trp²-M-PTH(1–14) caused a marked reduction in signaling by PTHR-Tether(1–9), whereas D-Trp¹²-PTH(7–34) was inert (Figure 8B). SW106 rapidly and dose dependently reduced cAMP signaling by PTHR-Tether(1–9), and the maximum inhibitory effect was greater than that seen for Trp²-M-PTH(1–14) analog. Thus, SW106 effectively inhibited PTHR1 signaling actions mediated by the first 9 amino acids of PTH interacting with the TMD region of the receptor.

Figure 8.

Ligand effects on cAMP signaling by a tethered PTH(1–9)-PTHR1 construct. Luciferase-based cAMP assays were performed in GS-22A cells transiently transfected with either vector (pCDNA3.1), PTHR1-WT, or PTHR-Tethered(1–9), in which the PTH(1–9) sequence is joined via a tetra-glycine linker to the extracellular end of TMD helix-1 (inset). Ligand-independent (basal) cAMP-dependent luminescence was assessed for 18 minutes after the addition of luciferin (A); the cells were then treated with vehicle, SW106 (1 × 10⁻⁵, 3 × 10⁻⁶, or 1 × 10⁻⁶M), DTrp¹²-PTH(7–34) (3 × 10⁻⁶M), or Trp²-M-PTH(1–14) (1 × 10⁻⁵M) (B), and luminescence was recorded for an additional 30 minutes. Data are mean ± SEM of 4 experiments, in each of which preligand data (A) were obtained from 24 replicate wells, and postligand addition data (B) were obtained from triplicate wells.

Ligand effects on dissociation of PTH-PTHR1 complexes

To further explore the mechanisms by which SW106 and AH-3960 bind to the PTHR1, we assessed the capacities of the compounds to modulate the rate at which bound ¹²⁵I-PTH(1–34) and ¹²⁵I-M-PTH(1–15) radioligands dissociated from the PTHR1. By this approach, if the added compound ligand increases the rate of dissociation of a prebound peptide radioligand, then the compound can be inferred to bind to a site on the receptor distinct from that used by the radioligand; if, on the other hand, the compound does not affect radioligand dissociation, then the compound can be inferred to bind to a receptor site that overlaps with that used by the peptide ligand (3, 35). As shown in Figure 9, the addition of SW106 caused a definite increase in the rate at which ¹²⁵I-PTH(1–34) dissociated from the PTHR1, whereas it had no effect on that rate at which ¹²⁵I-M-PTH(1–15) dissociated. Analysis of the effect of the compound on the 2-phase decay curve fit to the ¹²⁵I-PTH(1–34) dissociation data revealed an approximate 10% increase in the fraction of complexes that comprised the rapid component of the decay curve, together with a significant decrease in the t_1/2 values derived for both the rapid and slow components (Table 2). AH-3960 did not cause a significant change in the dissociation rate of either radioligand, although it tended to slightly slow the rate of dissociation of ¹²⁵I-PTH(1–34). The addition of guanosine 5′-O-[gamma-thio]triphosphate (GTPγS), well known to allosterically lower the affinity states of many GPCRs by promoting the dissociation of receptor•G protein complexes (36), had little effect on the dissociation rate of ¹²⁵I-PTH(1–34), whereas it markedly accelerated the dissociation rate of the shorter-length radioligand; these effects of GTPγS are fully consistent with those observed in previous studies on the dissociation of these 2 radioligands from the PTHR1 (28).

Figure 9.

Ligand effects on dissociation of PTH radioligands from PTHR1-WT. The effects of SW106, AH-3960, and GTPγS on the rates of dissociation of ¹²⁵I-PTH(1–34) (A) and ¹²⁵I-M-PTH(1–15) (B) from the PTHR1-WT were analyzed in membranes prepared from GP-2.3 cells. The radioligands were prebound to the receptor for 90 minutes; then an excess (5 × 10⁻⁷M) of unlabeled PTH(1–34) (A) or M-PTH(1–15) (B) was added, either alone (control) or in the presence of either SW106 (3 × 10⁻⁵M), AH-3960 (1 × 10⁻⁵M), or GTPγS (5 × 10⁻⁵M), and at times thereafter, aliquots were withdrawn and vacuum filtered; the filters were then rinsed, detached, and counted for bound γ radioactivity. Data are expressed as a percent of the radioactivity specifically bound immediately before (t = 0) addition of cold PTH ± compound ligand or GTPγS. Data are mean ± SEM of 4 experiments. Curves were fit to the data using a biphasic decay equation and constraining the plateau to 0%, t_1/2-1 to less than 5 minutes and t_1/2-2 to more than 5 minutes; fit parameters are reported in Table 2. The schematic (C) illustrates a plausible mechanism by which SW106 could destabilize complexes formed by PTH(1–34) (via simultaneous binding of the two ligands) but have no effect on complexes formed by M-PTH(1–15), as discussed further in the text.

Table 2.

Effects on Dissocation of PTH Radioligand Analogs From PTHR1-WT

125I-PTH Radioligand	Control	SW106		AH-3960		GTPγS	P
125I-rPTH(1–34)
Span-1 (%)	23 ± 2	33 ± 3	0.022	13 ± 4	0.048	20 ± 5	0.576
t1/2-1 (min)	3.1 ± 0.5	1.0 ± 0.4	0.018	1.7 ± 0.8	0.18	2.2 ± 0.8	0.407
Span-2 (%)	77 ± 3	67 ± 3	0.04	88 ± 4	0.067	80 ± 5	0.642
t1/2-2 (min)	134 ± 10	56 ± 3	0.00074	105 ± 6	0.037	93 ± 11	0.022
125I-M-PTH(1–15)
Span-1 (%)	35 ± 2	39 ± 2	0.14	32 ± 1	0.29	63 ± 4	0.001
t1/2-1 (min)	0.91 ± 0.24	1.3 ± 0.4	0.53	30.9 ± 0.3	0.96	0.5 ± 0.1	0.16
Span-2 (%)	65 ± 2	61 ± 2	0.13	68 ± 1	0.26	37 ± 4	0.0044
t1/2-2 (min)	87 ± 10	106 ± 15	0.32	85 ± 8	0.89	35 ± 3	0.0047

Dissocation of ¹²⁵I-M-PTH(1–15) or ¹²⁵I-PTH(1–34) from PTHR1-WT in membranes prepared from GP-2.3 cells assessed either in the absence (control) or presence of the indicated compound or peptide ligand. Data were fit to a biphasic decay equation constrained such that T1/2–1 and T1/2–2 were less than and greater than 5 minutes respectively, and the plateau equal to zero %. P indicates significance of difference vs control. Data are mean ± SEM of 5 experiments.

These results, showing that neither SW106 nor AH-3960 altered the rate of dissociation of ¹²⁵I-M-PTH(1–15) from the PTHR1, are consistent with the hypothesis that the compounds each bind to a site in the receptor that at least overlaps with that used by the N-terminal residues of PTH. That SW106 did increase the rate of dissociation of ¹²⁵I-PTH(1–34) is not inconsistent with this interpretation, because it could potentially by explained by the binding of the compound to the unoccupied TMD site of some fraction of the ¹²⁵I-PTH(1–34)-bound receptors, in which the radioligand is anchored to the receptor only by interactions between the PTH(15–34) segment and the ECD (Figure 9C). Binding of SW106 to the available TMD site would thus prevent the formation of the more complete PTH(1–34)-PTHR1 complex, in which the peptide ligand interacts with both the ECD and TMD regions, such that the compound would cause a net destabilization of the overall population of PTH(1–34)-PTHR1 complexes and, hence, an apparent increase in the rate of ¹²⁵I-PTH(1–34) dissociation. Such a model of simultaneous binding to the PTHR1 has been discussed previously for N-terminal and C-terminal fragments of PTH ligands interacting, respectively, with the receptor's TMD and ECD regions (37). The reason why AH-3960 did not similarly accelerate the rate of ¹²⁵I-PTH(1–34) dissociation is not clear but may in part be due to the moderately weaker affinity with which this ligand binds to the receptor (Figure 1A).

Discussion

The current absence of a potent mimetic ligand for the PTHR1, despite considerable discovery effort, likely reflects an incomplete understanding of the basic mechanisms by which any such nonpeptide ligand might interact with this receptor. To probe such mechanisms, we explored the PTHR1-binding and signaling actions of SW106 (24) and AH-3960 (22), 2 compounds that had been previously identified to act as a weak antagonist and a weak agonist, respectively, on the PTHR1, but which were otherwise not well characterized in terms of modes of action. Our main findings establish that each ligand interacts mainly, if not exclusively, with the receptor's TMD region and thus does not require interactions to the receptor's ECD to mediate its action on the receptor. This is shown by the nearly equivalent binding and signaling properties that each compound exhibited on the PTHR1-delNT construct as compared with on the intact PTHR1. Nonpeptide ligands for several other family B GPCRs have also been shown by similar pharmacological approaches using ECD-truncation constructs to bind to the TMD region of the target receptor (3, 38, 39). That the TMD region of at least some of the family B GPCRs, which, in general, have proven to be difficult targets for small-molecule ligand development, can accommodate such nonpeptide ligands is more recently highlighted by the crystal structures of the TMD regions of the CRFR1 (9) and GCGR (10), as each was obtained as a complex with a bound small-molecule, albeit an antagonist in each case and resolved only in the CRFR1 (11).

Our studies also revealed that the SW106 does not function as an inverse agonist on the constitutively active mutants, PTHR1-H223R, PTHR1-T410P, or PTHR1-HR/TP. This result was somewhat unexpected, given that many of the peptide ligands originally developed as PTHR1 antagonists were subsequently found to behave as inverse agonists when tested on these constitutively active receptors (17, 18, 31). Inverse agonists are of interest for the PTHR1, because such a ligand could, in principal, be used to treat Jansen's chondrodysplasia, a rare, yet dominant disease of skeletal development and mineral ion homeostasis that is caused by PTHR1 activating mutations, including H223R and T410P (19). Although the underlying mechanisms of inverse agonism at the PTHR1 are not fully understood, the current data at least suggest that SW106 does not mimic the effects of the known peptide determinants of such activity, which are the Gly¹²→DTrp substitution in the PTH(7–34) scaffold, and the Val²→BPA/Trp substitution in N-terminally intact PTH analogs (17, 18, 31).

Underscoring the absence of intrinsic inverse agonist activity in SW106, we found that the compound, in fact, exhibited weak, yet measurable agonist actions on the wild-type PTHR1, as well as on the constitutively active mutants. Detection of this agonist activity, not observed previously (18), was likely facilitated by the use of the Glosensor cAMP reporter, which enabled real-time, luminescence-based quantification of intracellular cAMP production. The lack of inverse agonist activity observed for SW106 on PTHR1-HR/TP is particularly noteworthy, because all peptide ligands studied to date, including PTH(1–34), behave as inverse agonists on this double-mutant receptor (18). Interestingly, AH-3960 also did not function as an inverse agonist on PTHR1-HR/TP but rather induced a positive agonist response. These findings overall suggest that the specific interactions by which SW106 and AH-3960 mediate their actions on the receptor are distinct from those by which any of the PTH peptide ligands, agonists or antagonists, mediate their actions, at least as considered in the context of the constitutively active mutant PTHR1s.

Although the current functional studies do not provide direct information on the specific binding sites used by the compound ligands within the PTHR1 TMD region, they nevertheless may provide some clues as to potential binding mechanisms. Thus, that SW106 dose dependently inhibited the agonist activity of the M-PTH(1–11) analog peptide on PTHR1-delNT, as well as the signaling activity of the native PTH(1–9) sequence, as tethered to the end of helix-1 of the TMD region, suggests that the compound binds to a site that at least partially overlaps with that used by the N-terminal(1–9) portion of PTH and thus presumably within a pocket formed between the extracellular ends of the TMD helices (8, 34). That neither SW106 nor AH-3960 altered the rate at which ¹²⁵I-M-PTH(1–15) dissociated from the receptor is consistent with this interpretation and, moreover, suggests that neither compound alters the peptide-binding pocket via an allosteric mechanism, although such a mechanism cannot be strictly ruled out. In this regard, the finding that GTPγS strongly destabilized the complexes formed between ¹²⁵I-M-PTH(1–15) and the PTHR1 confirms that the orthosteric, peptide ligand-binding pocket in the PTHR1 TMD region can be impacted by at least this type of G protein-based allosteric modulation (28, 37).

In the crystal structure of the CRFR1 TMD region, the compound ligand, CP-376395, was observed to be situated in an unexpectedly deep pocket formed in the lower portions of the helical bundle and well below the extracellularly exposed V-shaped cavity that likely serves as the orthosteric binding site for the peptide ligand (2, 10). Among the receptor residues in proximity to this deep drug-binding pocket are those corresponding to H223 and T410 in the PTHR1, which reside in the lower portions of TM helices 2 and 6, respectively. Given that CP-376395 can function as an inverse agonist on the CRFR1, it is conceivable that the compound, as bound in this deep location, can directly interfere with activation-related movements that normally would occur locally around this binding site and involve residues at or near the H223 and T410 positions (10). It is unknown whether such a deep drug-binding site, which was not seen in the GCGR TMD structure, is present in the PTHR1. In any case, the finding that SW106 is not an inverse agonist on the PTHR1-H223R and PTHR1-T410P mutants seems to suggest that the compound does not bind to any such site that is proximal to H223 and T410. Rather, the results seem more consistent with the binding of the compound to a site closer to or within the extracellularly exposed surface of the orthosteric binding pocket used by the N-terminal portion of the PTH peptide ligand (2, 11).

Binding of SW106 to a site within the more extracellularly exposed surface of the PTHR1 would also be generally consistent with a previous molecular modeling analysis, which suggested that the compound could fit into a predicted pocket formed between the extracellular ends of several TMD helices (24). The binding of SW106 to a site within the exposed peptide-binding cavity might also more easily account for the capacity of the compound to competitively block binding and hence signaling actions induced by the N-terminal portion of PTH agonist peptides but have no influence on activation events induced by mutations at the H223 and T410 sites and, thus, originating closer to the receptor's cytoplasmic surface. In this regard, the functional properties of SW106 would seem more similar to those of a neutral antagonist, such as the PTHrP(5–36) analogs that contain the native glycine, rather than DTrp residue at position 12, and thus do not suppress signaling by either the H223R or T410P mutant receptors (18).

Further studies employing more direct biophysical approaches are needed to define more precisely the specific molecular mechanisms by which SW106 and AH-3960 bind to the PTHR1 and modulate its signaling activity. The present results, establishing that these compounds mediate their distinct pharmacological actions by binding specifically to the TMD region of the PTHR1, provide new information that should prove useful in future efforts aimed at discovering and characterizing new small-molecule ligands for this receptor.