Comparison of Signaling Pathways Activated by the Relaxin Family Peptide Receptors, RXFP1 and RXFP2, Using Reporter Genes

The discovery of relaxin was originally based on its effects on the pubic ligament in guinea pigs (Hisaw, 1926) and led to the identification of its roles in parturition in many mammals and in implantation in humans (Hayes et al., 2004; Shirota et al., 2005). Research since has revealed a broader spectrum of physiological effects that are not solely based on actions within the reproductive system and include roles for relaxin in the cardiovascular system, collagen remodeling, wound healing, fibrosis prevention, and as a neuropeptide (for detailed review, see Bathgate et al., 2006a).

Relaxin is a two-chain peptide hormone that belongs to the insulin/relaxin peptide family. This family encompasses the relaxins: human gene 1 (H1) relaxin, H2 relaxin, and H3 relaxin; insulin; insulin-like growth factors I and II; and the insulin-like (INSL) peptides INSL3, INSL4, INSL5, and INSL6. The relaxin family peptides are limited to the relaxins (H1, H2, and H3) and INSL (INSL3, INSL4, INSL5, and INSL6). Recently, four G protein-coupled receptors (GPCR) were identified as the receptors for relaxin family peptides: LGR7 and LGR8 were identified as the receptors for H2 relaxin (Hsu et al., 2002) and INSL3 (Kumagai et al., 2002), respectively, and GPCR135 and GPCR142 were identified as the receptors for H3 relaxin (Liu et al., 2003b) and INSL5, respectively (Liu et al., 2004). Based on recent International Union of Pharmacology Committee on Receptor Nomenclature and Drug Classification recommendations, the receptors are referred to as relaxin family peptide receptor (RXFP) 1 (LGR7; H2 relaxin), RXFP2 (LGR8; INSL3), RXFP3 (GPCR135; H3 relaxin), and RXFP4 (GPCR142; INSL5) (Bathgate et al., 2006a).

RXFP1 and RXFP2 are leucine-rich repeat-containing GPCRs (LGR) that share 60% amino acid sequence homology (Hsu et al., 2002). LGRs are unique GPCRs in that the ectodomain of the receptor is relatively large and distinctive. The ectodomain of RXFP1 and RXFP2 comprises 10 leucine-rich repeats and a low-density lipoprotein class A module at the amino terminal, which is necessary for signal transduction (Scott et al., 2006). RXFP1 and RXFP2 each contain two ligand binding sites: a high affinity leucine-rich repeat site and a lower affinity transmembrane domain site (Sudo et al., 2003; Halls et al., 2005). In addition, there is a degree of overlap in ligand specificity for these receptors in humans: H2 relaxin binds to both RXFP1 and RXFP2 (Hsu et al., 2002; Halls et al., 2005); INSL3 binds to RXFP2 (Kumagai et al., 2002; Halls et al., 2005); whereas H3 relaxin binds with relatively high affinity to RXFP1 (Sudo et al., 2003), RXFP3 (Liu et al., 2003b), and RXFP4 (Liu et al., 2003a) but not RXFP2 (Sudo et al., 2003).

The signaling pathways mediated by RXFP1 and RXFP2 are not yet well characterized. Thus far, most research has focused on relaxin-mediated signaling events. Increases in cAMP have been observed in some target tissues and cell lines (Fei et al., 1990; Parsell et al., 1996) but not in others (Palejwala et al., 1998; Kompa et al., 2002). In addition, there is evidence for mitogen-activated protein kinase [extracellular signal-regulated kinase (Erk) 1/2] activation in particular cell types, which may occur with (Bartsch et al., 2001; Zhang et al., 2002) or without (Palejwala et al., 1998, 2001) associated increases in cAMP. In terms of cAMP signaling, RXFP1, when stably expressed in human embryonic kidney (HEK) 293T cells, was recently shown to modulate cAMP accumulation by coupling to multiple G proteins, including a Gα_s-mediated stimulation, a Gα_oB-mediated inhibition, and a delayed Gα_i3-Gβγ-mediated stimulation (Halls et al., 2006) via phosphoinositide-3-kinase (PI3K) and protein kinase Cζ isoform (PKCζ) (Nguyen et al., 2003; Nguyen and Dessauer, 2005). This complex interplay between pathways and relative influence of pathway component expression may provide an explanation for the rich variety of responses observed following relaxin stimulation across multiple cell lines and target tissues. Evidence also exists for relaxin-mediated increases in nitric oxide (Nistri and Bani, 2003; Conrad and Novak, 2004) and separately for direct activation of the glucocorticoid receptor (GR) (Dschietzig et al., 2004), which additionally suggested that relaxin might be able to translocate across the cell membrane. In contrast, little is known regarding signaling through RXFP2, with only cAMP accumulation studied in detail. Despite this, there is evidence for both Gα_s-coupling (Hsu et al., 2002; Kumagai et al., 2002) and an inhibitory pertussis toxin (PTX)-sensitive pathway (Kawamura et al., 2004) mediated by Gα_oB (Halls et al., 2006), which suggests that RXFP2-mediated signaling is also complex.

Thus, to gain a more comprehensive overview of the signaling events initiated by RXFP1 and RXFP2, we have measured the effect of receptor stimulation using reporter gene technology as a screen for additional signaling pathways. Stimulation of RXFP1 showed activation of the cAMP-response element (CRE) reporter, inhibition of nuclear factor κB (NF-κB)-stimulated reporter gene activity in response to both H2 relaxin and H3 relaxin, and activation of the glucocorticoid-response element (GRE) reporter by H2 relaxin and INSL3. The only reporter gene activated following RXFP2 stimulation was CRE, in response to all of the peptides tested.

Materials and Methods

Hormones and Reagents. Recombinant H2 relaxin was provided by BAS Medical (San Mateo, CA). H3 relaxin and human INSL3 were both chemically synthesized by Dr. John Wade at the Howard Florey Institute (Victoria, Australia). For detailed descriptions of H2 relaxin, H3 relaxin, and INSL3, see the review by Bathgate et al. (2006). Wortmannin was purchased from Calbiochem (Victoria, Australia), and PTX was purchased from Sigma-Aldrich (Sydney, Australia).

Reporter Gene Constructs. Reporter gene constructs (Clontech, Mountain View, CA) contained the secreted alkaline phosphatase (SEAP) reporter gene under the control of a specific transcription factor consensus sequence (Table 1). SEAP is engineered to be heat-resistant. All of the reporter genes were tested for activity using serum as a control stimulus [10% fetal bovine serum (FBS)]. All of the reporter genes, when expressed in HEK293T cells, were activated.

Cell Culture. Parent HEK293T cells (ATCC number CRL-1573; American Type Culture Collection, Manassas, VA) and HEK293T cells stably expressing either the RXFP1 or RXFP2 receptor were maintained in RPMI 1640 medium supplemented with heat-inactivated FBS (10% v/v), penicillin (100 U/ml), streptomycin (100 μg/ml), and l-glutamine (2 nM) (all Trace Biosciences, Sydney, Australia). All of the tissue culture plates and flasks were precoated with poly-l-lysine (0.1 mg/ml) (Sigma-Aldrich) before use. Cells were maintained at 37°C in a CO₂ water-jacket incubator (Thermo Electron Corporation, Waltham, MA) at 5% CO₂ and 85% humidity.

Parent HEK293T cells (as stated) and HEK293T cells stably expressing RXFP1 or RXFP2 were seeded into 12-well plates at 3.5 × 10⁵ cells/well in 10% FBS RPMI 1640 medium. After 24 h, cells were transiently cotransfected with the reporter gene of interest and a β-galactosidase plasmid using Metafectene (Biontex, Munich, Germany) as per manufacturer's instructions (Halls et al., 2005). Cells transiently expressing both the reporter gene of interest and the β-galactosidase plasmid were partially serum-starved (0.5% FBS RPMI 1640 medium) at approximately 30 h following transfection and used 48 h following transfection.

Reporter Gene Time Course Assays. SEAP reporter genes encode a truncated form of the human placental alkaline phosphatase gene that lacks membrane anchoring domains, allowing efficient secretion of the protein from transfected cells. Levels of SEAP activity in culture medium are directly proportional to changes in intracellular concentrations of SEAP mRNA and protein (Berger et al., 1988) and, as such, can give an indication of activity at the particular cis-acting enhancer element incorporated into a reporter vector. Detection of any enhancer activation is based on the interaction of SEAP with the fluorescent substrate 4-methylumbelliferyl phosphatase (4-MUP).

HEK293T cells stably expressing the receptors of interest and transiently expressing one of eight reporter genes were exposed to concentrations of relaxin family peptides previously shown to generate maximal responses (Sudo et al., 2003; Halls et al., 2005, 2006). After preincubation with PTX (16 h, 100 ng/ml) or wortmannin (30 min, 100 nM) and/or stimulation of transfected cells with H2 relaxin (100 nM), H3 relaxin (100 nM), INSL3 (100 nM), or vehicle (0.1% TFA) as stated, media samples were taken at indicated time points (0, 4, 8, 12, and 24 h) and frozen at –20°C.

Detection of SEAP. Following collection of all of the samples, samples were thawed and added to 96-well white optiplates (PerkinElmer, Victoria, Australia) in duplicate. Samples were heat-treated (65°C for 30 min) to denature any endogenous alkaline phosphatase and then cooled on ice before addition of SEAP buffer (0.39 M diethanolamine, and 0.98 mM MgCl₂, pH 10.3). The fluorescent substrate 4-MUP (0.2 mM) (Sigma-Aldrich) was incubated with the samples for 1 h in the dark. Plates were then read on a Fusion-α microplate reader (PerkinElmer) with excitation at 360 nm and emission at 440 nm.

Detection of β-Galactosidase. For all of the reporter gene assays, a plasmid encoding β-galactosidase was cotransfected with the reporter gene of interest. The β-galactosidase was constitutively expressed and secreted by the cell and thus was used as a measure of transfection efficiency between and within experiments. Following collection of all of the samples and detection of SEAP, a further portion of the sample was added to an additional white optiplate (PerkinElmer) for detection of β-galactosidase. Detection of β-galactosidase using the fluorescent substrate 4-methylumbelliferyl β-d-galactopyranoside was performed as described by Gee et al. (1999). In brief, β-galactosidase buffer (200 mM Na₂PO₄, 2 nM MgCl₂, and 200 mM β-mercaptoethanol, pH 7.0) was added to the samples, followed by addition of the fluorescent substrate 4-methylumbelliferyl β-d-galactopyranoside (0.2 mM) (Sigma-Aldrich), and incubated for 1 h in the dark. Plates were read as for SEAP samples.

Experimental Analysis and Statistics. SEAP readings were standardized for the β-galactosidase-induced fluorescence values generated for each sample set to account for variations in transfection efficiency. Results are expressed as the fractional change in response compared with the effect of vehicle alone at each time point and plotted against time. Values are expressed as mean ± S.E.M. of n experiments as stated. Statistical analyses were performed on raw data using a GraphPad Prism two-way analysis of variance (ANOVA) with Bonferroni's post-tests. Statistical significance is defined by p < 0.05 (indicated by *), p < 0.01 (indicated by **), and p < 0.001 (indicated by ***).

Results

Reporter Gene Activation by Relaxin Family Peptides following Stimulation of RXFP1 Receptors Stably Expressed in HEK293T Cells. The activation of eight reporter genes—CRE, GRE, NF-κB, activator protein 1 (AP1), serum-response element (SRE), nuclear factor of activated T cells (NFAT), E-box DNA binding element (Myc), and heat shock element (HSE) (Table 1)—was assessed after stimulation of the RXFP1 receptor by its cognate ligand H2 relaxin and the related peptides H3 relaxin and INSL3. All of the peptides were used at a single final concentration of 100 nM, previously shown to induce maximal functional responses, to determine the absolute effect of each peptide (Sudo et al., 2003; Halls et al., 2005, 2006). Three of the eight reporter genes showed activity above or below baseline (vehicle treatment) after 24-h stimulation: CRE, NF-κB, and GRE. At the CRE reporter (Fig. 1A), only H2 relaxin and H3 relaxin stimulation of RXFP1 induced a response. These two peptides gave responses that were significantly increased compared with vehicle at all of the time points tested. INSL3 had no effect at any time point compared with vehicle treatment alone. At the NF-κB reporter (Fig. 1B), INSL3 produced no significant effect compared with vehicle treatment alone. However, H2 relaxin stimulation decreased the production of SEAP through NF-κB, with the inhibition being significant from 8 h onward. H3 relaxin stimulation of RXFP1 also decreased SEAP production; however, this was only significant at 24 h. At the GRE reporter (Fig. 1C), both H2 relaxin and INSL3 treatment caused responses that were significantly elevated compared with baseline (vehicle treatment) at all of the time points. H3 relaxin did not significantly affect SEAP production in this instance.

A recent study has suggested that relaxin binds to and activates the GR in a manner independent of the RXFP1 receptor (Dschietzig et al., 2004). Based on this, as well as the activation of GRE in HEK-RXFP1 cells (cells stably expressing the RXFP1 receptor), we also tested the effect of relaxin family peptides (H2 relaxin, H3 relaxin, and INSL3, all 100 nM) on activity of the GRE reporter gene transiently expressed in parent HEK293T cells (cells not expressing either RXFP1 or RXFP2 receptors). When assessed at all of the time points up to 24 h, there was no discernable effect of any of the relaxin family peptides tested on the GRE reporter gene expressed in HEK293T cells (Fig. 1D). To assess the possibility of GR interaction stemming from RXFP1-specific coupling to G_i/G_o proteins (Halls et al., 2006), we tested the effect of PTX on the GRE response. There was no effect of this inhibitor on the GRE response through RXFP1 (data not shown).

The response of five other reporter genes was also assessed after stimulation of RXFP1 with H2 relaxin, H3 relaxin, INSL3, or vehicle treatment alone over a 24-h period (Fig. 2). These included AP1, SRE, NFAT, Myc, and HSE. These five reporter genes showed no change from vehicle-treated cells when stimulated with any of the peptides tested at any of the time points. As a positive control, all of the reporter genes were stimulated using 10% FBS, and all showed activation (data not shown).

Reporter Gene Activation by Relaxin Family Peptides after Stimulation of RXFP2 Receptors Stably Expressed in HEK293T Cells. Reporter gene activation was also assessed after stimulation of RXFP2 by its cognate ligand INSL3 and the related peptides H2 relaxin and H3 relaxin (all 100 nM). Again, all of the peptides were used at a single final concentration of 100 nM, previously shown to induce maximal functional responses, to determine the absolute effect of each peptide (Sudo et al., 2003; Halls et al., 2005, 2006). Only the CRE reporter showed activation after RXFP2 stimulation over the 24-h period (Fig. 3A). INSL3 stimulation of RXFP2 caused significant elevation of reporter activity above vehicle-treated cells at all of the time points. H2 relaxin activation of the reporter gene was also significantly increased compared with vehicle at all time points. Interestingly, H3 relaxin also showed elevated reporter gene activation above baseline (vehicle-treated cells), although this was only significant after periods longer than 8 h.

The remaining seven reporter genes—AP1, SRE, NFAT, Myc, HSE, GRE, and NF-κB—were not altered compared with baseline after stimulation with INSL3, H2 relaxin, or H3 relaxin (all 100 nM) (Fig. 3). As a positive control, all of the reporter genes were stimulated using 10% FBS, and all showed activation (data not shown). Thus, activation of the RXFP1 receptor affected CRE, NF-κB, and GRE reporter genes, whereas activation of RXFP2 receptors stimulated only the CRE reporter.

Effect of Irreversible G_i/G_o Inhibition on the CRE Response after Stimulation of RXFP1 and RXFP2. To examine the CRE responses in greater detail, we assessed the effect of PTX preincubation (irreversible G_i/G_o inhibition; 100 ng/ml, 16 h preincubation) on the CRE response generated after stimulation of RXFP1 with H2 relaxin (100 nM) (Fig. 4A) and RXFP2 with INSL3 (100 nM) (Fig. 4B). In both instances, PTX significantly increased the response of the CRE SEAP reporter gene, compared with peptide alone.

Effect of Irreversible G_i/G_o and PI3K Inhibitors on the Response Generated through the NF-κB Reporter Gene after Stimulation of RXFP1. Only HEK293T cells stably expressing RXFP1, but not cells stably expressing RXFP2, showed inhibition of the NF-κB reporter gene after stimulation with either H2 relaxin or H3 relaxin. The major difference between these two receptors (in terms of cAMP signaling: the major signaling pathway identified for these receptors) is the ability of RXFP1 receptors to couple to Gα_i3, release G-βγ, and stimulate PI3K activation and translocation of PKCζ (Nguyen et al., 2003; Nguyen and Dessauer, 2005; Halls et al., 2006). Interestingly, PKCζ has been found to interact with the scaffolding protein p62 to locate PKCζ in the NF-κB signaling pathway (Moscat and Diaz-Meco, 2005). This suggests that the activation and translocation of PKCζ may additionally provide a possible pathway mechanism for the observed interaction with the NF-κB reporter gene. However, because of the long period over which these assays are conducted, only irreversible inhibitors may be effectively used. Thus, we examined the effect of irreversible inhibition of the signaling proteins upstream of PKCζ (in an RXFP1 context): PI3K and Gα_i3.

HEK-RXFP1 cells transiently expressing the NF-κB reporter gene were pretreated with the irreversible PI3K inhibitor wortmannin (100 nM, 30 min preincubation) or the irreversible G_i/G_o inhibitor PTX (100 ng/ml, 16 h preincubation) (Fig. 5). There was no effect of either of the two inhibitors on vehicle-treated (Fig. 5A) or INSL3-treated cells (data not shown). In the presence of either wortmannin or PTX, the inhibitory effect of H2 relaxin stimulation of RXFP1 on the NF-κB reporter gene was significantly reversed at each time point (Fig. 5B). The inhibition exerted by H3 relaxin stimulation of RXFP1 on the NF-κB reporter gene was also significantly inhibited by wortmannin or PTX at 24 h (Fig. 5C).

Discussion

Reporter genes enable a measurement of the transcriptional activity resulting from receptor stimulation, providing an indication of the signaling pathways activated by receptors of interest. Thus, the use of reporter gene assays can provide a general screen of potential signaling pathways. This is advantageous in this setting because very little is known regarding the range of signaling pathways activated after stimulation of the recently deorphanized RXFP1 or RXFP2 receptors. Most studies to date have focused on cAMP signaling, although the importance of this signaling pathway in some endogenous systems is still not certain. Thus, reporter genes can potentially generate avenues for more focused signaling studies.

In this study, RXFP1 stimulation affected the transcriptional activity of CRE, NF-κB, and GRE reporter genes, whereas RXFP2 stimulation only affected the CRE reporter. The wider range of reporters activated by RXFP1 is unsurprising. RXFP1 can activate mitogen-activated protein kinase in some cells (Palejwala et al., 1998, 2001; Zhang et al., 2002) and is involved in both nitric oxide (Nistri and Bani, 2003; Conrad and Novak, 2004) and GR signaling (Dschietzig et al., 2004). In addition, in cAMP accumulation studies, RXFP1 activates a biphasic response with time, initially involving integration of Gα_s stimulation and Gα_oB inhibition, and finally stimulation via Gα_i3-G-βγ-PI3K-PKCζ (Nguyen et al., 2003; Nguyen and Dessauer, 2005; Halls et al., 2006). In contrast, there is only evidence for modulation of cAMP signaling by RXFP2, and this receptor can only couple to Gα_s and Gα_oB in this context (Halls et al., 2006).

The CRE reporter can be activated by members of the CREB family, including CREB, ATF-1, and CREM. Therefore, this reporter integrates signals from the cAMP/protein kinase A, calcium/calmodulin, Erk1/2, p38 kinase, and PI3K/Akt signaling pathways (for detailed review, see Lonze and Ginty, 2002). Many studies have shown RXFP1 and RXFP2 can couple to Gα_s and increase cAMP (Hsu et al., 2000, 2002; Halls et al., 2005, 2006). However, there is no evidence for increased calcium (Bani et al., 2002), Akt (Zhang et al., 2002), or phosphorylated p38 kinase (Heeg et al., 2005) levels in response to relaxin. In contrast, Erk1/2 activation can occur in a number of cells (Zhang et al., 2002; Dschietzig et al., 2003) and may contribute to the overall CRE response. Taken together, the observed CRE response in HEK293T cells is likely mainly the result of cAMP production.

This CRE activation would immediately appear instinctive because of the well reported effects of RXFP1 and RXFP2 stimulation on cAMP accumulation. However, the use of reporter gene technology has revealed a number of interesting phenomena. Because of the previously reported coupling of these two receptors to multiple G proteins (Halls et al., 2006), we examined the effect of the irreversible G_i/G_o inhibitor PTX on the receptor-stimulated CRE response. Both receptors showed significantly increased CRE responses in the presence of PTX, indicating that coupling to an inhibitory G_i/G_o protein was involved in pathways leading to CRE activation. Thus, integration of both Gs and inhibitory G_i/G_o signaling produces the observed CRE response. This is surprising because a great proportion of the RXFP1-mediated cAMP accumulation response occurs through a stimulatory Gα_i3 pathway (Halls et al., 2006). However, the effect of PTX in this study (enhancement of the response) suggests that Gα_i3 signaling does not contribute to the CRE response observed.

This may infer that RXFP1-mediated signaling is compartmentalized, such that only Gα_s and inhibitory G_i/G_o (probably Gα_oB) pathways are colocalized to affect CRE transcription.

It was also interesting to note the CRE response to H3 relaxin stimulation of RXFP2. This seems to contradict previous reports that indicate little to no binding of H3 relaxin to this receptor at concentrations used in this study (Sudo et al., 2003). However, more recent work showed slight H3 relaxin binding to human RXFP2 (pKi 6.97) (Bathgate et al., 2006b) and binding to rat RXFP2 with relatively high affinity (pIC₅₀ 7.3) (Scott et al., 2005). In addition, weak cAMP accumulation was observed at human RXFP2 with H3 relaxin stimulation (1 μM) (Bathgate et al., 2006b). Because the reporter gene assay occurs over an extended time period, it is likely that a small level of binding and receptor activation could generate enough cAMP to stimulate the small SEAP response. Thus, although H3 relaxin may bind with very low affinity to human RXFP2, it does so sufficiently to induce a CRE response.

In addition to modulation of the CRE reporter gene, stimulation of RXFP1 with both relaxins also affected the NF-κB reporter. INSL3 had no effect, consistent with the relative inability of this peptide to bind RXFP1 at the concentrations used (Halls et al., 2005). NF-κB is reported to be activated by all of the G proteins except Gα_s (for review, see Vanvooren et al., 2001), and activation of RXFP2 did not affect NF-κB. Taken together, this suggests that NF-κB inhibition occurs via a RXFP1-specific pathway. Recent work has identified Gα_i3-βγ-PI3K-PKCζ (in the context of cAMP accumulation) as such a pathway (Halls et al., 2006). This theory was confirmed by use of the irreversible inhibitors wortmannin (PI3K inhibitor) and PTX. Both inhibitors reversed the inhibitory effects of the two relaxin peptides but had no effect on vehicle- or INSL3-treated cells. Despite this, it is unclear which element of this pathway leads to NF-κB inhibition because reporter gene studies are limited as a result of the time scale and subsequent requirement for irreversible signaling inhibitors. Nevertheless, this, together with the CRE data, suggests compartmentalization of RXFP1-mediated signaling such that only Gα_s and inhibitory G_i/G_o (potentially Gα_oB) signaling affects CRE transcription, whereas only stimulatory G_i/G_o (potentially Gα_i3) signaling negatively affects NF-κB transcription.

Previous reports have alternatively suggested that relaxin stimulation activates NF-κB in some cell lines. In THP-1 cells, relaxin stimulation resulted in an increased level of NF-κB/DNA binding activity after 30 min (Ho and Bagnell, 2005). Relaxin stimulation for 45, 60, or 90 min also increased NF-κB/DNA binding in HeLa, human umbilical vein endothelial, and human vascular smooth muscle cells (Dschietzig et al., 2003). In addition, luciferase reporter gene assays showed NF-κB activation after 45-min relaxin stimulation in bovine aortic endothelial and HeLa cells (Dschietzig et al., 2003). This suggests that the inhibition of NF-κB observed in this study may specifically occur over a longer time period, and thus has important ramifications for situations that involve acute versus chronic exposure to relaxin.

It was recently reported that relaxin binds to and activates the GR (Dschietzig et al., 2004). The effects of relaxin on cytokine production in macrophages were alleviated by a GR antagonist, and relaxin was finally shown to physically interact with and compete for GR binding. In addition, relaxin activated a GRE-luciferase reporter over a time period of 6 h in HeLa, THP-1, and HEK293T cells (Dschietzig et al., 2004). In this current study, only cells stably expressing RXFP1 caused an increase in transcriptional activity of the GRE reporter after stimulation by H2 relaxin and INSL3, but not H3 relaxin. Interestingly, the observed INSL3 stimulation of GRE supports recent work showing a low-affinity interaction of INSL3 with RXFP1 (pKi 5.97) (Bathgate et al., 2006b).

To determine receptor specificity for the GRE response, we examined the effect of relaxin family peptides in HEKRXFP2 and parent HEK293T cells. Unexpectedly, there was no effect on GRE compared with vehicle-treated cells. This suggests, contrary to Dschietzig et al. (2004), that GRE activation in HEK293T cells indeed requires the presence of RXFP1. There was no effect of PTX on the GRE response (data not shown); thus, a GR interaction is not mediated by RXFP1-specific coupling to G_i/G_o proteins. This work suggests that RXFP1 is in some way required for the interaction of H2 relaxin and INSL3 with the GR, which consequently implicates ligand-directed signaling by H2 relaxin and INSL3 (but not H3 relaxin) to induce the GRE response. Alternatively, H2 relaxin and INSL3 may recognize a binding site on RXFP1 subtly distinct from the H3 relaxin site, which can induce the conformation necessary for GR interaction. Clearly, we require greater understanding of the relaxin family peptide-GR relationship to accurately interpret these data.

Examination of the signaling pathways activated by RXFP1 and RXFP2 utilizing reporter genes has confirmed the importance of cAMP-mediated signaling in HEK293T cells. The CRE response to stimulation of each receptor was similar and is affected by integration of G_s and inhibitory G_i/G_o pathways. RXFP1 activation also caused NF-κB inhibition, which was in turn reversed by G_i/G_o and PI3K inhibition. In addition, only RXFP1-expressing cells increased GRE activation, perhaps indicating that the relaxin-GR interaction is receptor-specific. The difference in transcription factors influenced by these two receptors, and particularly the greater range influenced by RXFP1, potentially provides scope for the wide and varied range of relaxin family peptide effects in numerous systems.