RELAXIN AND RELATED PEPTIDES: FIFTH INTERNATIONAL CONFERENCE

Activation of Relaxin-Related Receptorsby Short, Linear Peptides Derived from

a Collagen-Containing PrecursorRonen Shemesh,a Chen Hermesh,a Amir Toporik,a

Zurit Levine,a Amit Novik,a Assaf Wool,a Yossef Kliger,a

Avi Rosenberg,a Ross A. D. Bathgate,b and Yossi Cohena

aCompugen Ltd., Tel-Aviv, IsraelbHoward Florey Institute, University of Melbourne, Victoria 3010, Australia

In a screening effort based on algorithmic predictions for novel G-protein-coupled re-ceptor (GPCR) peptide activators, we were able to identify and examine two novel pep-tides (P59 and P74) which are short, linear, and derived from a natural, previouslyunidentified precursor protein containing a collagen-like repeat. Both peptides seemedto show an apparent cAMP-related effect on CHO-K1 cells transiently transfected witheither LGR7 or LGR8, usually after treatment with cAMP-generating forskolin, com-pared to the same cells treated with forskolin plus relaxin. This activation was notfound for the relaxin-3 receptor (GPR135). In a set of follow-up experiments, both pep-tides were found to stimulate cAMP production, mostly upon initial stimulation ofcAMP production by 5 μM forskolin in cells transfected with either LGR7 or LGR8.In a dye-free cell impedance GPCR activation assay, we were able to show that thesepeptides were also able to activate a cellular response mediated by these receptors.Although untransfected CHO-K1 cells showed some cellular activation by both relaxinand at least one of our newly discovered peptides, both LGR7- and LGR8-transfectedcells showed a stronger response, indicating stimulation of a cellular pathway throughactivation of these receptors. In conclusion, we were able to show that these newly dis-covered peptides, which have no similarity to any member of the relaxin–insulin-likepeptide family, are potential ligands for the relaxin-related family of receptors and assuch might serve as novel candidates for relaxin-related therapeutic indications. Bothpeptides are linear and were found to be active after being chemically synthesized.

Key words: relaxin; GPCR; collagen; receptor; ligand; peptide; agonist

Introduction

Known and uncharacterized G-protein-coupled receptors (GPCRs) currently consti-tute major targets for drug action and develop-ment, and >30% of all marketed therapeuticsact on them.1,2 GPCRs usually have seventransmembrane domains. Upon binding of aligand to an extracellular portion or fragmentof a GPCR, a signal is transduced within the

Address for correspondence: Ronen Shemesh. ronens@compugen.co.il

cell that results in a change in a biological orphysiological property or behavior of the cell.GPCRs, along with G-proteins and effectors(intracellular enzymes and channels modulatedby G-proteins), are the components of a mod-ular signaling system that connects the state ofintracellular second messengers to extracellu-lar inputs.3,4 The GPCRs seem to be of criticalimportance to both the central nervous systemand peripheral physiological processes. TheGPCR superfamily is diverse, and sequencingof the human genome has revealed >850 genesthat encode these receptors.5 There is great di-versity within the GPCRs which is matched

Relaxin and Related Peptides: Fifth International Conference: Ann. N.Y. Acad. Sci. 1160: 78–86 (2009).doi: 10.1111/j.1749-6632.2009.03827.x C© 2009 New York Academy of Sciences.

78

Shemesh et al.: Relaxin-Related Receptor-Activating Peptides 79

by the great variety of ligands that activatethem. Known drugs target only ∼30 membersof the GPCR family, mainly biogenic aminereceptors. Thus, there is an enormous poten-tial for the pharmaceutical industry to exploitthe remaining family members, including the>100 orphan receptors for which no existingligands have been identified.6 There are ongo-ing efforts to identify new GPCRs and to de-orphanize known GPCRs which can be usedthen to screen for new agonists and antagonistswith potential prophylactic and therapeuticproperties.6,7

The hormone relaxin and its family mem-bers (insulin-like peptide 3 [INSL3] and otherinsulin-like peptides) are GPCR peptide lig-ands that are emerging as multifunctional fac-tors in a broad range of target tissues,8,9

including several nonreproductive organs, inaddition to relaxin’s historical role as the “preg-nancy hormone” and in male and female fertil-ity. The many diverse and vital roles of relaxininclude homeostatic roles in mammalian preg-nancy and aging as well as therapeutic effectsin the treatment of fibrosis, inflammation, car-dioprotection, vasodilation, and wound healing(angiogenesis), among other pathophysiologi-cal conditions.

The P74 (CGEN-25009), P59 (CGEN-25010), and P59-s (CGEN-25011) peptideswere discovered by Compugen (Tel Aviv, Israel)through a previously disclosed GPCR peptideagonist’s discovery platform.10 Based on thismethod, Compugen was able to identify mul-tiple novel peptide ligands for GPCRs, includ-ing ligands for the formyl peptide receptor-likepeptide 1 (FPRL1) and MAS proto-oncogene1 (MAS1) GPCRs, for which significantin vivo efficacy has already been demonstrated.Compugen’s P74, P59, and P59-s are shortlinear peptides derived from a natural, previ-ously unidentified precursor protein contain-ing a collagen-like repeat. These peptides werefound to activate GPCRs LGR7 and LGR8,receptors for relaxin and INSL3, respectively,and therefore may have therapeutic activitiesfor various clinical indications, including labor

complications, infertility, and congestive heartfailure.

Methods

GPCR Screening for cAMP Accumulationwith Candidate Peptides

Applied Cell Sciences transiently transfectedcloned cDNA of different GPCRs into CHO-K1 cells using their proprietary protocol.Transfected cells were plated into 24 wellsof a 96-well plate. Cells were pretreatedwith 0.5 mM isobutylmethylxanthine (IBMX[Sigma, St. Louis, MO, USA]) (stimulationbuffer) for 10 min at 37◦C and then stimu-lated for 20 min with either of the tested pep-tides (for stimulatory [Gs] functional exami-nation) or preincubated with 10 μM forskolin(for inhibitory [Gi] functional examination) fol-lowed by a 20-min stimulation with either of thetested peptides. Either with or without forskolinpretreatment, intracellular cAMP was assayedusing the Hit Hunter cAMP kit (DiscoveRxCorporation), according to the manufacturer’srecommended protocol. Data were convertedto nanomoles of cAMP by running a standardcAMP curve.

RXFP3 Activation by P59

The abilities of P59 and H3 relaxin (Euro-screen) to activate CHO-K1 cells transfectedwith GPR135, a non-LGR, relaxin-related re-ceptor, were tested. For cAMP concentrationanalysis, cells (GPR135; Euroscreen, catalogno. ES-656-A) grown to mid-log phase inculture medium without antibiotics, were de-tached with phosphate-buffered saline (PBS)–EDTA, centrifuged, and resuspended in assaybuffer at a concentration of 7.5 × 105 cells/mL.The test was performed in 96-well plates. Foragonist testing, 12 μl of cells (5 × 103 cells/well)was mixed with 12 μl of agonist, either H3 re-laxin or P59, at increasing concentrations. Theplates were then incubated for 30 min at roomtemperature. After addition of the lysis buffer,

80 Annals of the New York Academy of Sciences

Figure 1. Compugen’s GPCR peptide prediction platform was used successfully todiscover peptides that activate the LGR7 receptor. (A) Compugen’s GPCR peptide agonist dis-covery platform as described in the text. (B) P59 stimulates a dose-dependent cAMP responsein LGR7-transfected cells. CHO-K1 cells transiently transfected with LGR7 (and treated withIBMX) were pretreated with forskolin and then challenged with either H2 relaxin (as a positivecontrol) or P59 or P74 at increasing concentrations. As can be seen, P59 results formed abell-shaped curve, as at lower concentrations there was an increase in cAMP (Gs effect)with the 30 and 100 nM doses compared to H2 relaxin, whereas relaxin showed a dose-dependent decrease (Gi) effect at higher concentrations (>100 nM). P74 showed no effectin this assay. (The work presented was done by Applied Cell Sciences.) (C) P59 stimulates a

Shemesh et al.: Relaxin-Related Receptor-Activating Peptides 81

cAMP concentrations were estimated, accord-ing to the manufacturer’s specification, withthe homogeneous time-resolved fluorescence(HTRF) kit for cAMP measurement from Cis-Bio International (catalog no. 62AM2PEB).

CRE Response of LGR7- andLGR8-Transfected CHO-K1 Cellsto Stimulation by P59 and P74

The pCRE–β-galactosidase (pCRE–β-Gal)assay is a standard and well-established assayat the Howard Florey Institute (Melbourne,Australia) for measurement of cAMP responseelement (CRE) activity.11 The experimentationwas done on transiently transfected CHO-K1cells pretreated with 5 μM forskolin.

Cell Impedance Real-Time Monitoringof GPCR Activation

The ACEA Biosciences (ROCHE) RT-CESscreen is a noninvasive and label-free assayfor GPCRs that can be used with both engi-neered and nonengineered cell lines. The as-say is based on cell–electrode impedance tomeasure minute changes in cellular morphol-ogy as a result of ligand-dependent GPCRactivation.12

Compugen’s GPCR Peptide AgonistDiscovery Platform

A peptide database was created using acleavage-predicting algorithm13 on all theNCBI nonredundant (nr) (proteome in Fig. 1A)potentially secreted proteins (human proteinswith a signal peptide prediction) and Swiss-Prot/Uniprot signal peptide annotated pro-teins. Out of this data set, designated thesecretome (Fig. 1A), hundreds of thousandspeptides were generated by the classifier (pep-

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−dose-dependent cAMP response in LGR7-transfected cells. Results shown are the cAMP stimu-lation (Gs) dose response at 3–100 nM by H2 relaxin and P59 on LGR7-transfected CHO-K1cells that were treated with IBMX. Both P59 and H2 showed a dose-dependent increase incAMP. (The work presented was done by Applied Cell Sciences.)

tidome in Fig. 1A) as a bulk data set on whichwe used a machine learning approach aimingat discriminating between GPCR- and non-GPCR-activating peptides based on distinctparameters (GPCR peptide ligands in Fig. 1A).The thousands of predicted GPCR-activatingpeptides underwent a manual review for dis-tinct GPCR-related peptide features, and agroup of hundreds of “high-reliability ligands”(Fig. 1A) was selected and inspected for nov-elty. The most reliable 33 novel peptides weresynthesized and screened for activation of 152selected GPCRs by using both Ca2+ and cAMPaccumulation assays.

Results

Screening Experiment

Thirty-three novel peptides were selectedand synthesized (by Pepscan Inc., The Nether-lands). All peptides were tested for activation on152 GPCRs selected according to two majorcriteria:

(1) Clinical implication for agonist activity orestimated activity (based on knockout ex-periments)

(2) Probability that the GPCR would be ac-tivated by a peptide (based on family andsequence resemblance)

Eight hits (25%) were found by screening allpeptides against all receptors at a high concen-tration (1 μM). Six peptides activated a cal-cium flux (Gq)-related pathway on four recep-tors (MRGX1, MRGX2, MAS1, and FPRL1),and all of them showed a clear dose–responseactivation pattern on these receptors. At leasttwo of the peptides were further examined forin vivo function and showed significant efficacyin rodent disease models.10,14

82 Annals of the New York Academy of Sciences

Figure 2. P59 and P74 stimulate dose-dependent cAMP responses mediated by LGR7 and LGR8 butnot by GPR135. (A–D) LGR7- and LGR8-transfected CHO-K1 cells were treated with 5 μM forskolin toprestimulate cAMP (it was previously reported in the literature that CHO-K1 cells need a cAMP “boost” witha low concentration of forskolin in order to be examined for cAMP activation). (A) LGR7-transfected cellswere challenged with increasing doses of H2 relaxin (as a positive control), P74, or P59-s. The bell-shapedactivation pattern for P74 is shown. Each point represents three unrelated repeats for each experiment. (B)

Shemesh et al.: Relaxin-Related Receptor-Activating Peptides 83

In a cAMP assay done across a small sub-set of the above 152 GPCRs, two novel pep-tides were discovered: P74 (CGEN-25009) andP59 (CGEN-25010). Both peptides (togetherwith a short subset termed P59-S, or CGEN-25011) were derived from the same precur-sor, which is a hypothetical protein containingboth a collagen-like domain and a C1Q do-main. Both the above peptides were able toshow a distinct cAMP inhibition (Gi) responseat 1 μM on two different receptors, LGR7 andLGR8.10

Functional and Dose–Response Analysis

A cAMP accumulation assay demonstrateda dose-dependent activation for one of the pep-tides (P59) on one of the receptors (LGR7), ineither the presence (Fig. 1B, partial) or absence(Fig. 1C) of forskolin. The activation in the pres-ence of forskolin (Fig. 1B) was only shown at twopeptide concentrations (i.e., 30 and 100 nM).

We went further to test the abilities of P74and P59 to influence cAMP activity inducedby 5 μM forskolin in CHO-K1 cells tran-siently transfected with LGR7/pCRE-β-Gal orLGR8/pCRE-β-Gal in parallel to H2 relaxinand INSL3, respectively. This work was doneexternally by the Howard Florey Institute. Aspresented in Figure 2 (A and B for LGR7 andC and D for LGR8), we could see a clear dose-dependent effect on LGR7/8-transfected cells.The bell shape activation curve of P59 and P74(Fig. 2A–D) seemed to show the same pattern

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−LGR7-transfected cells were challenged with increasing doses of H2 relaxin (as a positive control), P59,or P59–amide. A dual-effect activation pattern was found for P59, whereas a slight activation pattern wasalso seen with P59–amide. Each point represents three unrelated replicates for each experiment. (C) LGR8-transfected cells were challenged with increasing doses of INSL3 (as a positive control), P74, or P59-s.The bell-shaped activation pattern for P74 is shown. Each point represents three unrelated repeats for eachexperiment. (D) LGR7-transfected cells were challenged with increasing doses of H2 relaxin (as a positivecontrol), P59, or P59–amide. A dual-effect activation pattern is shown for P59, and an activation pattern isalso shown for P59–amide. Each point represents three unrelated repeats for each experiment. (The workpresented here was performed by The Howard Florey Institute.) (E) cAMP dose–response stimulation. GPR135-transfected cells were challenged with increasing concentrations of relaxin 3 (H3 relaxin, as a positive control)and P59. Compared with H3 relaxin, no stimulation by P59 was shown on GPR135. Relaxin is representedby the circles and the P59 is represented by the triangles. (The work presented was performed by EuroscreenLtd.)

as was shown for P59 on LGR7 (Fig. 1B–C),where the cAMP increase was found to bedose dependent only for lower peptide concen-trations; however, we could see distinct differ-ences between the examined peptides (to thisexperiment we added an N-terminal amidatedversion of P59 in addition to P59, P74, andP59-S). This experiment was repeated threetimes (Fig. 2A–D show the overall averages ofall three experiments). We can therefore con-clude that all of the above peptides show acAMP dose-dependent activity on both recep-tors LGR7 (Fig. 2A and B) and LGR8 (Fig. 2Cand D). In contrast with LGR7 and LGR8, P59showed no ability to activate a cAMP responsein GPCR135-transfected cells (in comparisonwith the natural ligand, H3 relaxin). Figure 2Eshows the results of this assay (done as an out-source service by Euroscreen Ltd.).

Real-Time Activation Assay Tested withthe ACEA Cell Impedance System

ACEA Biosciences (ROCHE) has designedelectronic cell sensor arrays embedded in thebottom of the wells of microtiter plates thatare capable of measuring minute changes incell morphology. The electronic sensors mea-sure changes in cell–substrate impedance asa result of the disruption of the ionic en-vironment due to cell and cell morphologydynamics. The main advantage offered by mea-suring cell–substrate impedance and cell mor-phology is that both exogenously expressed and

84 Annals of the New York Academy of Sciences

Figure 3. Compugen’s P74 and H2 relaxin activate a dose-dependent downstream cellular responseas determined with the ACEA cell impedance measurement device. (A) LGR7-transfected CHO-K1 cells.The cell index was normalized to T1 (after peptide challenge). There are clear strong effects of both H2relaxin and P74 on LGR7-transfected CHO-K1 cells. The effect of relaxin is moderate in comparison toBSA. However, the effect of P74 was significant. (B) LGR8-transfected CHO-K1 cells. The cell index wasnormalized to T1 (after peptide challenge). There are clear strong effects of both H2 relaxin and P74 onLGR7-transfected CHO-K1 cells. The effects of both H2 relaxin and P74 were apparent (not tested for statistical

Shemesh et al.: Relaxin-Related Receptor-Activating Peptides 85

endogenous receptors can be assayed withoutthe need for engineering the cell with promis-cuous G-proteins and reporters or labeling thecells with dyes. In addition, since the readoutis noninvasive, multiple stimulations with thesame ligand or different ligands can be per-formed to assess events such as desensitizationand receptor cross talk. Finally, another majoraspect of using cell–substrate impedance andcell morphology as readouts is that potentiallyall GPCRs, regardless of signaling pathways,can be functionally monitored.

The screening was performed on transientlytransfected CHO-K1 cells and was conductedin two stages:

1. Screening phase: challenging the trans-fected CHO-K1 cells with 10 μM of eachscreened peptide, using calcitonin as anassay internal positive control. (The calci-tonin receptor is endogenously expressedby CHO-K1 cells.15) The negative controlwas 0.1% bovine serum albumin (BSA).H2 relaxin at 1 μM was a positive controlfor the transfected receptor.

2. Dose–response phase: challenging thetransfected CHO-K1 cells with increasingconcentrations of each screened peptide,using H2 relaxin as a positive control forthe transfected receptor.

For the screening phase (Fig. 3A and B), themain outcome of this phase was that P74 wasable to activate a cellular response in LGR7-transfected (Fig. 3B) and LGR8-transfected(Fig. 3B) cells at high concentrations (10 μM).The positive control (H2 relaxin at 1 μM)showed a similar cellular activation on LGR8-transfected cells (Fig. 3B) and a lower but

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−significance). P59 showed no effect compared to BSA. (C) Cell index for LGR7-transfected CHO-K1 cellschallenged with increasing concentrations of H2 relaxin. There is a clear dose response for the higherconcentrations, with an estimated 50% effective concentration of ∼100 nM. The arrow indicates the additionof relaxin (H2). (D) A dose–response curve for the data shown in panel C. (E) Cell index for LGR7-transfectedCHO-K1 cells challenged with increasing concentrations of P74. There is a clear dose response for the higherconcentrations, with an estimated 50% effective concentration of ∼300 nM. The arrow indicates the additionof the peptide P74. (F) A dose–response curve for the data in panel E.

distinct activation on LGR7-transfected cells(Fig. 3A).

In the dose–response phase, activation ofboth H2 relaxin and P74 showed dose-dependent activation on LGR7-transfectedcells (Fig. 3C and F). P59 also showed a weakerdose-dependent activation (data not shown),and P59-S showed a moderate activation at thehighest concentration (10 μM).

Conclusions

Novel GPCR peptide agonists were discov-ered and validated as activators of severalGPCRs. Among these peptides, two were foundto specifically activate receptors of the relaxin-related family (LGRs LGR7 and LGR8). Boththe P59 and P74 peptides were derived froma collagen/C1Q-containing hypothetical pre-cursor. The peptides were not able to activatethe non-LGR relaxin 3 receptor (GPR135).Some of the peptides were able to activate someof the orphan LGRs (LGR4 and LGR6; datanot shown). The novel peptides are simple andlinear and were able to activate the receptorswhen chemically synthesized.

Conflicts of Interest

The authors declare no conflicts of interest.

References

1. Jacoby, E., R. Bouhelal, M. Gerspacher, et al. 2006.The 7 TM G-protein-coupled receptor target family.ChemMedChem 1: 761–782.

86 Annals of the New York Academy of Sciences

2. Lundstrom, K. 2005. The future of G protein-coupled receptors as targets in drug discovery. IDrugs

8: 909–913.3. Pierce, K.L., R.T. Premont & R.J. Lefkowitz. 2002.

Seven-transmembrane receptors. Nat. Rev. Mol. Cell

Biol. 3: 639–650.4. Eglen, R.M., R. Bosse & T. Reisine. 2007. Emerg-

ing concepts of guanine nucleotide-binding protein-coupled receptor (GPCR) function and implicationsfor high throughput screening. Assay Drug Dev. Technol.

5: 425–451.5. Hopkins, A.L. & C.R. Groom. 2002. The druggable

genome. Nat. Rev. Drug Discov. 1: 727–730.6. Gilchrist, A. 2004. Second annual GPCRs: From or-

phan to blockbuster. Expert Opin. Ther. Targets 8: 495–498.

7. Schlyer, S. & R. Horuk. 2006. I want a new drug: G-protein-coupled receptors in drug development. Drug

Discov. Today 11: 481–493.8. Samuel, C.S., T.D. Hewitson, E.N. Unemori, et al.

2007. Drugs of the future: The hormone relaxin.Cell. Mol. Life Sci. 64: 1539–1557.

9. Van Der Westhuizen, E.T., R.J. Summers, M.L.Halls, et al. 2007. Relaxin receptors—new drug tar-gets for multiple disease states. Curr. Drug Targets 8:91–104.

10. Shemesh, R., A. Toporik, Z. Levine, et al. 2008.Discovery and validation of novel peptide agonistsfor G-protein coupled receptors. J. Biol. Chem. 283:34643–34649.

11. Halls, M.L., R.A. Bathgate & R.J. Summers. 2007.Comparison of signalling pathways activated bythe relaxin family peptide receptors, RXFP1 andRXFP2, using reporter genes. J. Pharmacol. Exp. Ther.

320: 281–290.12. Yu, N., J.M. Atienza, J. Bernard, et al. 2006. Real-time

monitoring of morphological changes in living cellsby electronic cell sensor arrays: An approach to studyG protein-coupled receptors. Anal. Chem. 78: 35–43.

13. Kliger, Y., E. Gofer, A. Wool, et al. 2008. Predict-ing proteolytic sites in extracellular proteins: Onlyhalfway there. Bioinformatics 24: 1049–1055.

14. Hecht, I., R. Jiang, AL. Sampaio, et al. 2009. Anovel peptide agonist of FPRL1 (ALX) displays anti-inflammatory and cardioprotective effects. J Pharma-

col. Exp. Ther. 2008 Nov 20. [Epub ahead of print].15. George, S.E., P.J. Bungay & L.H. Naylor. 1997. Func-

tional coupling of endogenous serotonin (5-HT1B)and calcitonin (C1a) receptors in CHO cells to acyclic AMP-responsive luciferase reporter gene. J.

Neurochem. 69: 1278–1285.