Targeting of -3 analogues to the brain for pharmacological modulation of neurosignalling Han Siean Lee Bachelor of Biotechnology Drug Design and Development (Hons)   

 https://orcid.org/0000-0001-6058-6689     A thesis submitted for the degree of Doctor of Philosophy at The University of Queensland in 2019 Faculty of Medicine

  Abstract Relaxin-3, a member of the /relaxin superfamily, has been shown to be involved in modulating several neurophysiological process including food intake, stress, addiction, arousal, memory and learning in animal models. Its sequence is highly evolutionary conserved, which strongly suggests an important physiological function. Elucidation of the relaxin-3 three-dimensional structure showed that relaxin-3 adopts a structure similar to insulin, in which the two chains are linked by two inter-chain and one intra-chain disulfides. Residues vital for binding to its cognate receptor, relaxin family peptide 3 receptor (RXFP3), are found in the helical region of the B-chain whereas the C-terminal tail residues of the B-chain are important for activation of the receptor. Truncation of five residues from the C-terminus results in an antagonist, and antagonists and have been important for studying relaxin-3 function. However, the animal studies to date have almost exclusively been conducted via intracerebroventricular injection as these do not efficiently cross the blood-brain barrier (BBB). The BBB is a highly regulated barrier consisting of tight junctions between endothelial cells that prevents paracellular diffusion of most molecules into the brain, and it is decorated with transporters that pump out unwanted products. Many strategies for increasing brain delivery have been utilised over the years, including increasing lipophilicity and positive charge of the drug lead to facilitate membrane permeability and the use of molecular ‘Trojan horses’. These ‘Trojan horses’ can be nonspecific in nature, like cell penetrating peptides, or specifically target receptors expressed on the BBB to enter the brain through receptor-mediated transcytosis.

The main aim of this thesis was to design a series of relaxin-3 analogues based solely on the B-chain with increased potential for crossing the BBB when introduced through the systemic circulation, and with an ability to modulate food intake in mice models. In order to achieve this aim, structure-activity relationship studies were first conducted on the single-chain relaxin-3 antagonist (R3 B1-22R) to further understand the binding determinants of this variant. Using Fmoc-peptide synthesis modifications, including an alanine scan, were introduced and there effect on binding evaluated through competition binding assays. There were similarities between the antagonist residues involved in binding to RXFP3 and the relaxin-3 residues, but additional residues are involved in the antagonist and the binding conformations were distinct between the agonist and antagonist. Since linear relaxin-3 B-chain variants are unstructured in solution and easily degraded in serum, we employed several strategies to reintroduce structure into the single-chain analogues. These included substitution with helical promoting residues (Aib), global cyclisation, molecular grafting, and crosslinking of side chains through lactam bonds or hydrocarbon stapling, in order to improve stability and interactions with RXFP3. Aib and salt-bridges were well-tolerated in the single-chain

  antagonist, although it did not improve helicity or affinity for RXFP3 above R3 B1-22R. NMR spectroscopy studies conducted on peptide analogues with receptor interacting relaxin-3 residues grafted onto disulfide-rich scaffold, showed an increase in helical content and improved stability in serum from ~5 min to > 6 h. For agonist variants, RXFP3 potency was also improved over the unstructured B-chain, but intriguingly antagonist variants showed slight reductions in binding to RXFP3 relative to R3 B1-22R. Alternative stapling strategies and Aib substitutions into the single- chain relaxin-3 agonist were also explored but did not improve structure, affinity and potency of the agonist beyond what has been previously achieved in Ac-R3B10-27 [13/17 HC], a hydrocarbon stapled relaxin-3 agonist. Ac-R3B10-27 [13/17 HC] is more stable than the native relaxin-3 B-chain, but not as stable as the grafted analogues. The grafting scaffold used, the bee venom peptide apamin, is known to be BBB permeable, thus the apamin grafted relaxin-3 antagonist was chosen as one of the designs for evaluation of BBB penetration. Five other analogues containing BBB shuttles targeting different active transport mechanisms and five analogues with lipidation/cationisation motifs were also designed to compare the efficiency of each targeting approach. Pharmacological characterisation showed that the novel analogues retain binding to RXFP3, with lipidated analogues being the most stable in serum. Cytotoxicity can also be observed in some analogues, although it can be reversed by substituting Lys for Glu residues. In vitro permeability assays indicated that the modifications improved the permeability compared to R3 B1-22R, in both passive and active transport mechanism, however, their permeability is still poor.

Identification of receptor interacting residues provides essential information for efficient drug lead design and development. The data suggest that relaxin-3 analogues can be modified to be stable and incorporate BBB shuttle sequences without causing significant cytotoxicity or losing affinity for RXFP3. However, further optimisation of the sequences will be required for efficient targeting of RXFP3 using systemic delivery. The analysis of permeability itself is severely hampered by a difficulty in retrieving the peptide analogues from cells or tissue for quantification. Nonetheless, the lack of efficient permeability for the current analogue designs will assist in the design of new generation BBB permeable relaxin-3 analogues.

  Declaration by author This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance, survey design, data analysis, significant technical procedures, professional editorial advice, financial support and any other original research work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my higher degree by research candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the policy and procedures of The University of Queensland, the thesis be made available for research and study in accordance with the Copyright Act 1968 unless a period of embargo has been approved by the Dean of the Graduate School.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis and have sought permission from co-authors for any jointly authored works included in the thesis. 

  Publications included in this thesis Haugaard-Kedström, L.M., Lee, H.S., Jones, M.V., Song, A., Rathod, V., Hossain, M.A., Bathgate, R.a.D., and Rosengren, K.J. (2018). Binding conformation and determinants of a single-chain peptide antagonist at the relaxin-3 receptor RXFP3. J Biol Chem 293, 15765-15776.  Lee, H.S., Postan, M., Song, A., Clark, R.J., Bathgate, R.a.D., Haugaard-Kedstrom, L.M., and Rosengren, K.J. (2020). Development of Relaxin-3 Agonists and Antagonists Based on Grafted Disulfide-Stabilized Scaffolds. Front Chem 8, 87. 

  Submitted manuscripts included in this thesis No manuscripts submitted for publication.

Other publications during candidature Peer-reviewed articles:

Miles, J.J., Tan, M.P., Dolton, G., Edwards, E.S., Galloway, S.A., Laugel, B., Clement, M., Makinde, J., Ladell, K., Matthews, K.K., Watkins, T.S., Tungatt, K., Wong, Y., Lee, H.S., Clark, R.J., Pentier, J.M., Attaf, M., Lissina, A., Ager, A., Gallimore, A., Rizkallah, P.J., Gras, S., Rossjohn, J., Burrows, S.R., Cole, D.K., Price, D.A., and Sewell, A.K. (2018). Peptide mimic for influenza vaccination using nonnatural combinatorial chemistry. J Clin Invest 128, 1569- 1580.

Carstens, B.B., Berecki, G., Daniel, J.T., Lee, H.S., Jackson, K.A., Tae, H.S., Sadeghi, M., Castro, J., O'donnell, T., Deiteren, A., Brierley, S.M., Craik, D.J., Adams, D.J., and Clark, R.J. (2016). Structure-Activity Studies of Cysteine-Rich alpha-Conotoxins that Inhibit High-Voltage- Activated Calcium Channels via GABA(B) Receptor Activation Reveal a Minimal Functional Motif. Angew Chem Int Ed Engl 55, 4692-4696.

Conference abstracts: Han Siean Lee, Ross A.D. Bathgate, Angela Song, Joseph A. Nicolazzo, K. Johan Rosengren (2019) Development of novel relaxin-3 analogues for blood-brain barrier penetration. In 13th Australian Peptide Conference, Port Douglas, Australia (Poster presentation)

Han Siean Lee, Ross A.D. Bathgate, Joseph A. Nicolazzo, K. Johan Rosengren (2018) Development of novel relaxin-3 analogues for blood-brain barrier penetration. In 10th International Peptide Symposium, Kyoto, Japan (Oral presentation)

Han Siean Lee, Ross A.D. Bathgate, Joseph A. Nicolazzo, K. Johan Rosengren (2018) Development of novel relaxin-3 analogues for blood-brain barrier penetration. In 9th International Postgraduate Symposium, University of Queensland, Brisbane, Australia (Oral presentation)

Han Siean Lee, Ross A.D. Bathgate, Joseph A. Nicolazzo, K. Johan Rosengren (2017) Development of novel relaxin-3 analogues for blood-brain barrier penetration. In 12th International Conference on Cerebral Vascular Biology, Melbourne, Australia (Poster presentation)

  Han Siean Lee, Ross A.D. Bathgate, Joseph A. Nicolazzo, K. Johan Rosengren (2017) Development of novel relaxin-3 analogues for blood-brain barrier penetration. In 8th International Postgraduate Symposium, University of Queensland, Brisbane, Australia (Poster presentation)

Han Siean Lee, Ross A.D. Bathgate, Joseph A. Nicolazzo, K. Johan Rosengren (2017) Development of novel relaxin-3 analogues for blood-brain barrier penetration. In 12th Australian Peptide Conference, Noosa, Australia (Poster presentation)

Han Siean Lee, Ross A.D. Bathgate, Joseph A. Nicolazzo, K. Johan Rosengren (2016) Novel relaxin- 3 antagonist analogues for delivery across blood-brain barrier. In 7th International Postgraduate Symposium, University of Queensland, Brisbane, Australia (Poster presentation)

Contributions by others to the thesis My primary supervisor, Dr. Johan Rosengren was essential in the conception and direction of the project. He recorded and processed the NMR data and performed the majority of the analysis, and 3D structure determination. He also helped review my manuscripts and thesis chapters.

Shu Wang and Michael Postans assisted with some NMR analysis and assignments.

Dr. Ross Bathgate undertook the cell based assays and provided all the competition binding and activity data.

Dr. Joseph A. Nicolazzo conducted the cellular uptake of my novel analogues into the immortalised human brain endothelial cells.

Linda Haugaard-Kedström and Maryon Jones provided majority of the synthetic peptide analogues used in Chapter 3.

Shu Wang, James Daniel and Angela Song provided some of the synthetic peptide analogues used in Chapters 3 and 4.

  Statement of parts of the thesis submitted to qualify for the award of another degree No works submitted towards another degree have been included in this thesis.

Research Involving Human or Animal Subjects No animal or human subjects were involved in this research.

  Acknowledgements Firstly, I would like to thank my supervisors, especially Dr K. Johan Rosengren for his ongoing support and advise throughout my four years of PhD program. He has guided me throughout the project and his open-door policy was comforting as it allowed me to discuss various issues with him. I would also like to thank Richard Clark for showing me the ropes in peptide synthesis and for providing feedback in the manuscripts I have written. Both of them have been my supervisors even before I started my PhD journey and they are the best mentors anyone could ask for.

I also highly appreciate Vinod Kumar for his guidance in mass spectrometry and patience in answering all my questions. He has also provided one of the columns that was used frequently in my MS peptide quantification, and for that, I am very grateful. Angela has also been important for her technical assistance and it was a pleasure to work with her. Colton, thank you for being patient and kind to have a look over my NMR assignments to make sure it is correct.

It is also my pleasure to thank our collaborators, Ross Bathgate and Joe Nicolazzo for providing the data and samples for me to analyse. Without them, I would not have been able to complete my thesis.

Dr Lachlan Rash and Dr Trent Woodruff, my thesis committee, have been essential in providing feedback and suggestions to strengthen my thesis.

Thank you to The University of Queensland for the financial support through the scholarships provided, as I would not have been able to conduct my research.

I would also like to acknowledge my friend, Shreya, and colleagues/friends in the laboratory, especially Taylor, Elena and Bruno, for their words of encouragement and providing the very much needed distractions and shoulder to lean on at hard times. My laboratory mates have been very kind and understanding especially when I’m under more stress, and that made it all easier for me to get things done.

Last but not least, I would like to thank my family for their love and encouragement. Without their support and care throughout my life, I would not have been able to come this far. 

  Financial support This research was supported by an University of Queensland International Scholarship - Living Allowance and UQ Tuition fee waiver.

Faculty of Medicine Travel scholarship provided funding for my domestic and international conferences.

Shimadzu and Biotage provided the Travel Award for the Australian Peptide Conference 2017.

AB Sciex provided the Travel Award for the Australian Peptide Conference 2019.

Keywords relaxin-3, structure-activity relationship, grafting, peptide stapling, peptide shuttle, blood-brain barrier, permeability , peptide synthesis, pampa

  Australian and New Zealand Standard Research Classifications (ANZSRC) ANZSRC code: 030406 and Peptides, 40%

ANZSRC code: 110106 Medical Biochemistry Proteins and Peptides (incl. Medical Proteomics), 40%

ANZSRC code: 060112 Structural Biology (incl. Macromolecular Modelling), 20%

Fields of Research (FoR) Classification FoR code: 0601, Biochemistry and Cell Biology, 30%

FoR code: 0304, Medicinal and Biomolecular Chemistry, 40%

FoR code: 1115, Pharmacology and Pharmaceutical Sciences, 30%

  Table of Contents List of Figures and Tables ...... xiv List of Abbreviations ...... xviii Chapter 1 ...... 1 Introduction ...... 1 1.1 Bioactive peptides ...... 2 1.2 Human Insulin superfamily ...... 2 1.3 Discovery of relaxin-3 ...... 4 1.3.1 Structure-activity relationships ...... 5  1.3.2 Relaxin-3 modifications ...... 7 1.3.3 Biological functions of relaxin-3 ...... 10 1.3.3.1 Stress response ...... 12 1.3.3.2 Appetite regulation ...... 13 1.3.3.3 Alcohol dependence ...... 14 1.3.3.4 Arousal, memory and learning ...... 15 1.3.3.5 Non-neuronal function ...... 16 1.4 Blood-brain barrier ...... 17 1.4.1 Peptide drug delivery strategies across the BBB ...... 19 1.4.1.1 Non-specific targeting ...... 19 1.4.1.2 Ligand targeting towards BBB receptors ...... 22 1.4.1.3 Combinatorial shuttle and colloidal carrier system ...... 27 1.4.1.4 Alternative approaches ...... 29 1.5 Aims ...... 30 Chapter 2 ...... 31 Materials and methods ...... 31 2.0 Materials ...... 32 2.1 Peptide synthesis ...... 33 2.2 Oxidation of grafted peptides ...... 36 2.3 Peptide conjugation ...... 37 2.4 Hydrocarbon and halogen stapling ...... 38 2.5 Cell culture ...... 39 2.6 Competition binding assay ...... 40 2.7 cAMP functional assay ...... 41 2.8 Serum stability assay ...... 41 2.9 MTT assay ...... 42 2.10 In vitro BBB penetration assays ...... 43 2.10.1 Parallel artificial membrane permeability assay (PAMPA) ...... 43 2.10.2 Cellular uptake of peptides in hCMEC/D3 cells ...... 44 2.10.3 LC-MS analysis ...... 45 2.11 NMR spectroscopy ...... 49 2.12 Structure determination ...... 50

  Chapter 3 ...... 51 Binding conformation and determinants of a single-chain peptide antagonist at the relaxin-3 receptor RXFP3...... 51 3.1 Introduction ...... 53 3.2 Results ...... 56 3.2.1 Alanine scan of R3 B1-22R identifies amino acids involved in the RXFP3 interaction .. 56 3.2.2 Only a small subset of amino acid residues contribute specific interactions when binding to RXFP3 ...... 59 3.2.3 Arg23 at the R3 B1-22R C-terminus cannot be altered ...... 59 3.2.4 Can an Arg at position 23 in a full length B-chain improve affinity of an agonist? ...... 60 3.2.5 Modifications supporting a helical conformation of R3 B1-22R are well tolerated, but ones that disrupt it are not ...... 60 3.3 Discussion ...... 64 3.4 Conclusion...... 72 Chapter 4 ...... 73 Effect of peptide grafting and stapling on affinity and potency of single-chain relaxin-3 analogues at RXFP3 ...... 73 4.1 Introduction ...... 75 4.2 Results ...... 77 4.2.1 Peptide design rationale and synthesis ...... 77 4.2.2 α-Helical structure were restored in grafted relaxin-3 analogues ...... 82 4.2.3 Helicogenic amino acid but not thiol-based stapling methodology can improve secondary structure of single-chain relaxin-3 agonists ...... 84 4.2.4 Affinity and potency of some modified peptides for RXFP3 are retained ...... 86 4.2.5 Novel analogues showed improved stability in serum ...... 90 4.3 Discussion ...... 91 4.4 Conclusion...... 98 Chapter 5 ...... 99 Relaxin-3 antagonist analogues to improve BBB penetration ...... 99 5.1 Introduction ...... 100 5.2 Results ...... 103 5.1.1 Peptide analogue design rationale ...... 103 5.2.2 Modified single-chain relaxin-3 antagonist analogues retain binding affinity for RXFP3 ...... 105 5.2.3 Peptide analogues showed improvement in serum stability...... 107 5.2.4 Cytotoxicity can be observed in some novel analogues ...... 109 5.2.5 BBB penetration differ between active and passive mechanism targeting analogues .... 112 5.3 Discussion ...... 115 5.4 Conclusion...... 123 Chapter 6 ...... 124 Conclusion and future directions ...... 124 References ...... 132 Appendices...... 163

  List of Figures and Tables

List of Figures

Chapter 1 Figure 1-1 Sequences of the members in the insulin/relaxin superfamily...... 3 Figure 1-2 Relaxin-like peptide family members and their respective receptors...... 4 Figure 1-3 Amino acid comparison of human, porcine, mouse and rat relaxin-3...... 5 Figure 1-4 Relaxin-3/RXFP3 interaction model...... 7 Figure 1-5 Cross reactivity between relaxin peptides and relaxin receptors...... 10 Figure 1-6 Simplified schematic representation of relaxin-3 innervations and RXFP3 distribution by function...... 12 Figure 1-7 Effects of relaxin-3 agonist and antagonist on different rodent behaviours...... 16 Figure 1-8 Comparison between peripheral and brain capillaries...... 18 Figure 1-9 Overview of mechanism of transports for endogenous molecules across the BBB...... 19

Chapter 2 Figure 2-1 Standard curves for peptide quantitation by LC-MS/MS...... 48

Scheme 2-1 General steps of solid phase peptide synthesis...... 34 Scheme 2-2 On-resin lactam stapling...... 35 Scheme 2-3 Peptide oxidation protocols...... 37 Scheme 2-4 Peptide conjugation to plug-and-play shuttle analogues...... 38 Scheme 2-5 Stabilisation strategies for single-chain relaxin-3 agonist...... 39

Chapter 3 Figure 3-1 Structure and receptor interacting amino acid residues of relaxin-3...... 54 Figure 3-2 Competition binding of R3 B1-22R variants at RXFP3...... 58 Figure 3-3 Chemical structure of arginine and its variants...... 60

Figure 3-4 Secondary Hα shifts (experimental shifts – random coil shifts) for R3 B1-22R in H2O and 30% TFE solvents...... 63 Figure 3-5 Summary of the SAR data from point substitutions in R3 B1-22R...... 65 Figure 3-6 Differences in binding mode of R3 B1-22R and relaxin-3...... 70 Figure 3-7 Comparison between (A) relaxin-3 binding to active state homology model of RXFP3 and (B) R3 B1-22R binding to inactive state homology model of RXFP3...... 71

  Chapter 4 Figure 4-1 Structural comparison of (A) apamin (red) and (B) VhTI (blue) with (C) the relaxin-3 B- chain (green)...... 77 Figure 4-2 Comparison of secondary Hα chemical shift for apamin and VhTI based relaxin-3 analogues and native relaxin-3...... 82 Figure 4-3 Backbone superposition of the 20 best structures calculated based on NMR spectroscopy data for (A) analogue 3, (B) analogue 9 and (C) analogue 5...... 83 Figure 4-4 Comparison of secondary Hα chemical shift of Aib containing and stapled relaxin-3 analogues...... 85 Figure 4-5 Affinity and activity of relaxin-3 grafted agonists and antagonist on RXFP3...... 87 Figure 4-6 Binding affinity and potency of relaxin-3 agonist analogues at RXFP3...... 90 Figure 4-7 Stability of relaxin-3 analogues in serum...... 91 Figure 4-8 Comparison of the solution NMR structures of (A) analogue 3, (B) analogue 9, (C) analogue 5 and (D) analogue 1 (B-chain only)...... 95

Chapter 5 Figure 5-1 Binding affinity of relaxin-3 antagonist analogues conjugated to BBB shuttles...... 107 Figure 5-2 Stability of novel analogues in serum...... 109 Figure 5-3 Toxicity of BBB analogues in SH-SY5Y cells...... 111 Figure 5-4 Comparison of binding affinity of analogue 41 (EEKpE modified variant) to analogues 8 and 33...... 112 Figure 5-5 Concentration of novel analogues in different compartments of the PAMPA assay. ... 113

Chapter 6 Figure 6-1 Proposed workflow diagram of peptide-based drug discovery targeting the BBB……131

List of Tables

Chapter 1 Table 1-1 Summary of currently used relaxin-3 analogues...... 8 Table 1-2 Examples of combinatorial CNS drug delivery...... 28

Chapter 2 Table 2-1 LC-MS columns and IS concentrations...... 46 Table 2-2 MRMs used in the detection and identification of novel analogues...... 47

  Chapter 3 Table 3-1 Effects of point modifications on binding affinity for RXFP3...... 57 Table 3-2 Effects of helical supportive or disruptive modifications on binding affinity for RXFP3...... 62

Chapter 4 Table 4-1 Amino acid sequences of grafted RXFP3 agonists (3-7) and antagonists (9-10)...... 79 Table 4-2 Peptide sequence of novel relaxin-3 agonist analogues containing Aib and stapling...... 81 Table 4-3 Binding affinities and activities of grafted peptides...... 87 Table 4-4 Effect of helical promoting residue and stapling introduction on affinity and potency at RXFP3...... 89

Chapter 5 Table 5-1 Relaxin-3 antagonist peptide-shuttle candidates...... 104 Table 5-2 Binding affinity of potential BBB permeable relaxin-3 antagonist for RXFP3...... 106 Table 5-3 Half-life of novel BBB analogues...... 108 Table 5-4 Permeability of novel analogues in PAMPA...... 113 Table 5-5 Cellular uptake of novel analogues in hCMEC/D3 over time...... 114 Table 5-6 Peptide analogue net charge at pH 7...... 119 Table 5-7 Radiolabeled peptide uptake into hCMEC/D3 normalised to the total cell ...... 122

Supplementary figures and tables

Figure S-1. Competition binding of Arg23 agonist variants at RXFP3...... 164 Figure S-2. MS spectra and corresponding analytical HPLC traces of oxidised and purified analogues 3 – 6...... 168 Figure S-3. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 7, 9 and 10...... 169 Figure S-4. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 11 - 16...... 170 Figure S-5. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 16 - 22...... 171 Figure S-6. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 23 - 27...... 172 Figure S-7. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 3. .... 173 Figure S-8. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 4. .... 173

  Figure S-9. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 5. .... 174 Figure S-10. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 6. .. 174 Figure S-11. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 7. .. 175 Figure S-12. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 9. .. 175 Figure S-13. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 10.. 176 Figure S-14. Comparison of secondary Hα chemical shifts for analogue 3 and analogue 4...... 176 Figure S-15. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 28 – 32...... 178 Figure S-16. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 33 – 37...... 179 Figure S-17. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 38 – 41...... 180

Table S-1. Eu-H3/I5 competition binding of R3 B1-22R C-terminal extension variants...... 165 Table S-2. Physiochemical properties of point mutated R3 B1-22R analogues...... 166 Table S-3. Physiochemical properties of analogues of R3 B1-22R with helix supportive or disruptive modifications...... 167 Table S-4. Structural statistics from NMR based structure calculations...... 177

  List of Abbreviations

Amino acids

A Alanine R Arginine N Asparagine

D Aspartic acid C Cysteine E Glutamic acid

Q Glutamine G Glycine H Histidine

I Isoleucine L K Lysine

M F Phenylalanine P Proline

S Serine T Threonine W Tryptophan

Y Tyrosine V Valine

5-Ava 5-aminovaleric acid 6-MHA 6-maleimidohexanoic acid AC ACN Acetonitrile Aib Aminoisobutyric AMT Adsorptive-mediated transcytosis AMP Antimicrobial peptide AD Alzheimer’s disease At Atenolol BACE1 β-secretase BBB Blood-brain barrier BBCEC Bovine brain capillary endothelial cells BNST Bed nucleus of the stria terminalis BOC tert-butoxycarbonyl BSA Bovine serum albumin cAMP Cyclic adenosine monophosphate CHO Chinese hamster ovary CNS Central nervous system CRF Corticotropin-releasing factor CPP Cell penetrating-peptide CTX Chlorotoxin

  D2O Deuterium oxide DBx α,α’-dibromo-m-xylene DCA 1,3-dichloroacetone DELFIA Dissociation-enhanced lanthanide fluorescence immunoassay DIPEA N,N-Diisopropylethylamine DCM Dichloromethane DMEM Dulbecco’s Modified Eagle Medium DMF N,N-Dimethylformamide DMSO Dimethyl sulfoxide DODT 3,6-Dioxa-1,8-octane-dithiol DPDS 2,2'-Dipyridyldisulfide DQF-COSY Double-quantum filtered correlation spectroscopy DSS 4,4-dimethyl-4-silapentane-1-sulfonic acid ESI-MS Electrospray ionisation- mass spectrophotometry FA Formic acid FBS Fetal bovine serum FMOC Fluoren-9-ylmethyloxycarbonyl FUS Focused ultrasound GLUT1 Glucose transporter GPCR G-protein coupled receptor HATU 1-[Bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3- oxide hexafluorophosphate HBTU O-(6-Benzotriazol-1-yl)-N,N,N′,N-Tetramethyluroniumhexafluorophosphate HSA Human serum albumin HSQC Heteronuclear single quantum coherence ICV Intracerebroventricular IGF Insulin-like INSL Insulin-like IR IS Internal standard KO Knock-out LAT1 Large neutral amino acid transporter LDLR Low-density lipoprotein receptor LRP Low-density lipoprotein receptor-related proteins MRM Multiple reaction monitoring

  MTT 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide MV Methoxyverapamil NI Nucleus incertus NMR Nuclear magnetic resonance NOESY Nuclear Overhauser effect spectroscopy NP Nanoparticle NT p-97 Melanotransferrin Pa Pentenyl-alanine PAMPA Parallel artificial membrane permeability assay PBS Phosphate buffered saline PD Parkinson’s disease Pg Pentenyl-glycine P-gp P-glycoprotein PVN Paraventricular nucleus PYBOP Benzotriazol-1-yl-oxytripyrrolidinophosphonium hexafluorophosphate R3/I5 Relaxin-3 B-chain/INSL5 A-chain RCM Ring-closing metathesis RMT Receptor-mediated transcytosis RP-HPLC Reverse phase high-pressure liquid chromatography RPMI Roswell Park Memorial Institute RXFP Relaxin family peptide SAR Structure-activity relationship SPPS Solid phase peptide synthesis TCEP Tris(2-carboxyethyl)phosphine TIPS Triisopropylsilane TFA Trifluoroacetic acid TFE Trifluoroethanol TfR Transferrin receptor TOCSY 1H homonuclear total correlation spectroscopy VhTI Veronica hederifolia trypsin inhibitor



 

Chapter 1 Introduction



  1.1 Bioactive peptides

Ribosomally produced peptides consist of a combination of two or more of 20 proteinogenic amino acids, and their diversity can be further enhanced by posttranslational modifications such as the formation of disulfide bonds. Bioactive peptides have evolved to a specific biological function, and are found throughout all living organisms (Kastin, 2013). These peptides can have antimicrobial properties that are vital for the immunity of the organism, act as toxins for defense or predation (Marmiroli and Maestri, 2014) or be peptide hormones functioning peripherally or in the central nervous system (CNS) (Burbach, 2011;Leng and Sabatier, 2017). Many peptides have multiple functions, for example, insulin not only plays an important role in regulating blood glucose but is also important for the brain to regulate energy expenditure (Blazquez et al., 2014).

Peptides are attractive drug leads as they can be altered chemically; allowing manipulation of peptide sequence and structures towards high target specificity and affinity, lower toxicity and less accumulation in tissue (Henninot et al., 2018). However, peptides are rapidly cleared from the body, have poor stability due to protease degradation, low oral bioavailability, and can be potentially immunogenic (Bellmann-Sickert and Beck-Sickinger, 2010;Henninot et al., 2018). Recent developments in the delivery and stabilisation strategies of peptides have focussed on overcoming these limitations, and they have been used by pharmaceutical companies to develop treatments for diseases ranging from cancer, Alzheimer’ disease (AD) to diabetes (Bellmann-Sickert and Beck-Sickinger, 2010;Henninot et al., 2018). With over 150 peptides currently in clinical trials and 60 peptide drugs that have been approved, it is a clear sign that peptides are a class of drugs that is rapidly growing (Lau and Dunn, 2018). 

1.2 Human Insulin superfamily

The peptide hormones of the human insulin superfamily are one group of peptides considered attractive targets for pharmaceutical drug development. The human insulin superfamily comprises insulin, insulin-like growth factors (IGF-1 and IGF-2), and seven relaxin peptides, including relaxin-1-3 and insulin-like peptide (INSL) 3-6 (Figure 1-1) (Adham et al., 1993;Chassin et al., 1995;Conklin et al., 1999;Hsu, 1999). Although through phylogenetic analysis relaxin-3 is most closely related to INSL5 (Wilkinson et al., 2005), the

  presence of the binding motif RXXXRXXI in the B-chain, is a defining functional motif that is the signature of relaxin (Bullesbach and Schwabe, 2000;Bathgate et al., 2002).

A-chain B-chain

Insulin SLQKRGIVEQCCTSICSLYQLENYCN FVNQHLCGSHLVEALYLVCGERGFFYTPKT IGF-1 APQTGIVDECCFRSCDLRRLEMYCA GPETLCGAELVDALQFVCGDRGFYFNKPT IGF-2 RRSRGIVEECCFRSCDLALLETLCA YRPSETLCGGELVDTLQFVCGDRGFYFSRPA Relaxin-1 RPYVALFEKCCLIGCTKRSLAKYC KWKDDVIKLCGRELVRAQIAICGMSTWS Relaxin-2 QLYSALANKCCHVGCTKRSLARFC DSWMEEVIKLCGRELVRAQIAICGMSTWS Relaxin-3 DVLAGLSSSCCKWGCSKSEISSLC RAAPYGVRLCGREFIRAVIFTCGGSRW INSL3 AAATNPARYCCLSGCTQQDLLTLCPY PTPEMREKLCGHHFVRALVRVCGGPRWSTEA INSL4 RSGRHRFDPFCCEVICDDGTSVKLC ESLAAELRGCGPRFGKHLLSYCPMPEKTFTTTP INSL5 MSRQDLQTLCCTDGCSMTDLSALC VRSKESVRLCGLEYIRTVIYICASSRWRRHLEG INSL6 RKRRGYSEKCCLTGCTKEELSIAC SDISSARKLCGRYLVKEIEKLCGHANWSQFRFE

Figure 1-1 Sequences of the members in the insulin/relaxin superfamily. Conserved cysteine residues are highlighted in green and connecting lines show the disulfide connectivity.

All family members except IGF-1 and -2 comprise two peptide chains connected by two inter-chain disulfide bonds with the A chain containing an additional intra-chain disulfide (Schwabe and McDonald, 1977). These peptides are expressed as pre-prohormones that consist of an N-terminal for secretion, the B-chain, a non-conserved C-chain and the C- terminal A-chain (Shabanpoor et al., 2009). Subsequent protein processing results in the formation of the mature and functional A-B heterodimer peptide (Shabanpoor et al., 2009)

Despite the members of this superfamily being structurally similar, they bind and activate two different receptor types. Insulin, IGF-1 and IGF-2 bind to tyrosine kinase receptors, whereas relaxin-like peptides bind to G-protein coupled receptors (GPCR). Four GPCRs have been identified to date and are known as relaxin family peptide (RXFP) receptors (Figure 1-2).

  INSL3 Relaxin 2 Relaxin 3 INSL5

N N

N N

C C C C RXFP2 RXFP1 RXFP3 RXFP4  Figure 1-2 Relaxin-like peptide family members and their respective receptors. INSL3 and INSL5 bind to RXFP2 and RXFP4 respectively. Relaxin-2 binds to both RXFP1 and RXFP2. Relaxin-3 can bind to its cognate receptor, RXFP3 and also to RXFP1 and RXFP4. Bold arrow indicates high affinity binding while thin arrow indicates low affinity binding.

Despite being evolutionary related, these peptides have a wide array of physiological functions. Insulin’s main function is to lower blood glucose levels, but is also found in the brain where it modulates energy expenditure (Blazquez et al., 2014). In contrast, both IGFs are important growth hormones (Shabanpoor et al., 2009). Relaxin-1 is an orthologue of relaxin-2 with currently unknown function, and is only found in higher primates (Shabanpoor et al., 2009). Human relaxin-2 is the functional equivalent of the “relaxin” found in other mammals. Relaxin-2 is widely circulated in the human body; playing various roles ranging from being a pregnancy hormone, anti-fibrotic agent, cardio protectant to an anti-inflammatory agent (reviewed in (Samuel et al., 2007)). Relaxin-3 has been shown to play a vital role as a (Olucha-Bordonau et al., 2018). INSL3 is important in modulating fertility in both sexes, whereas for INSL4-6, no function has been determined thus far (reviewed in (Shabanpoor et al., 2009;Bathgate et al., 2013a)).

1.3 Discovery of relaxin-3

Relaxin-3 is the most recently identified member of the insulin superfamily. Bathgate and colleagues identified the novel relaxin-3 in both humans and mice through searches of the Celera Genomics data base (Bathgate et al., 2002), and found it to be highly expressed in the mouse and rat brain (Bathgate et al., 2002;Burazin et al., 2002). Interestingly, the relaxin- 3 sequence is highly conserved among different species (Figure 1-3), strongly indicating that relaxin-3 plays an important physiological role. Evolutionary analyses have identified relaxin- 3 as the ancestral relaxin (Wilkinson et al., 2005)

 

Human -MARYMLLLLLAVWVLTGELWPGAEARAAPYGVRLCGREFIRAVIFTCGGSRWRRSDILA 59 Porcine MAKRPLLLLLLAVWVLAGELWLRTEARASPYGVKLCGREFIRAVIFTCGGSRWRRSDMLA 60 Mouse --MAMLGLLLLASWALLGALGLQAEARPAPYGVKLCGREFIRAVIFTCGGSRWRRADILA 58 Rat --MAT-RGLLLASWALLGALVLQAEARPAPYGVKLCGREFIRAVIFTCGGSRWRRADILA 57 B-chain

Human HEAMGDTFPDADADEDSLAGELDEAMGSSEWLALTKSPQAFYRGRPSWQGTPGVLRGSRD 119 Porcine HEALGDVFSDTDSNAD---SELDEAMASSEWLALTKSPETFYGVQPGWQRTPGALRGSRD 117 Mouse HESLGDFFADGEANTDHLASELDEAVGSSEWLALTKSPQAFYGGRASWQGSPGVVRGSRD 118 Rat HDPLGEFFADGEANTDHLASELDEAVGSSEWLALTKSPQVFYGGRSSWQGSPGVVRGSRD 117 C-peptide

Human VLAGLSSSCCKWGCSKSEISSLC 142 Porcine VLAGLSSNCCKWGCSKSEISSLC 140 Mouse VLAGLSSSCCEWGCSKSQISSLC 141 Rat VLAGLSSSCCEWGCSKSQISSLC 140 A-chain Figure 1-3 Amino acid comparison of human, porcine, mouse and rat relaxin-3. Conserved amino acids are shaded. The C-peptide is cleaved off to form the mature peptide containing the A- and B-chains.

Initial research showed that relaxin-3 is able to interact with and activate three different GPCR receptors, namely RXFP1 (Sudo et al., 2003), RXFP3 (Liu et al., 2003b) and RXFP4 (Liu et al., 2003b) (Figure 1-2) with different binding affinity. Relaxin-3 activates RXFP1

(Sudo et al., 2003) and signal through Gαs, causing cyclic adenosine monophosphate (cAMP) accumulation (Liu et al., 2003b;Sudo et al., 2003). In contrast, relaxin-3 binds to RXFP3 and

RXFP4 at low nanomolar range and signal primarily via Gαi resulting in cAMP inhibition (Liu et al., 2003a;Liu et al., 2003b). Initially relaxin-3 was reported as a specific ligand for RXFP1, but subsequently it was concluded that RXFP3 is the endogenous receptor for relaxin-3. This is because RXFP3 is highly expressed in brain regions that overlap with relaxin-3 expression (Figure 1-6) and relaxin-3 is the only ligand known to bind to and activate RXFP3 (Liu et al., 2003b;Sutton et al., 2004).

1.3.1 Structure-activity relationships Using NMR spectroscopy techniques, the three-dimensional solution structure of relaxin-3 was determined and it was confirmed to have an insulin-like fold, which is stabilised by its three disulfide bonds, a motif conserved throughout the family (Figure 1-5C) (Rosengren et al., 2006). The relaxin-3 A-chain folds into two anti-parallel α-helices separated by a short β-strand. The B-chain contains a β-strand at the N-terminus and a longer α-helix is present through the middle segment. The B-chain helix lies perpendicular to the axes of A-chain helices (Rosengren et al., 2006) (Figure 1-5C).

  Mutational studies on the relaxin-3 B-chain have yielded interesting results. Six residues at the N-terminus of the B-chain are not required for binding and activation of RXFP3, RXFP4 or RXFP1 (Kuei et al., 2007). This is in line with sequence diversity as relaxin-2, relaxin-3 and INSL5 do not have a conserved N-terminus of the B-chain (Figure 1-1). In contrast, mutations of Arg12, Ile15, Arg16 and Phe20, which are highly conserved, resulted in peptides with reduced affinity and potency at RXFP3 (Kuei et al., 2007). Arg26 and Trp27 are critical for activation of RXFP3 as alanine substitution of both residues to R26A and W27R showed a total loss of activation, while retaining high affinity at RXFP3 (Kuei et al., 2007). Based on the relaxin-3 structure, B-chain residues Arg12, Ile15, Arg16 and Phe20 are mostly exposed on the B-chain helix while the C-terminus of the B-chain folds back to allow Trp27 to interact with the hydrophobic residues Val18 and Ile19 (Rosengren et al., 2006). Whether this conformation is the bioactive one has been debated though, as the tail linking the helix and activation domain Arg26,Trp27 is more flexible. The C-terminus of the B-chain is required to be a free acid as C-terminal amidation lowers binding and activity at RXFP3 (Shabanpoor et al., 2013).

Key residues on RXFP3 transmembrane helix 2 and exoloop 1 that are involved in the ligand/receptor interaction have also been identified (Figure 1-4). Receptor mutagenesis of negatively charged residues on RXFP3 showed that Arg12 and Arg16 of relaxin-3 B-chain potentially form electrostatic interactions with Glu244 and Asp145 respectively, bringing Arg26 in close contact with Glu141 that sits on top of transmembrane helix 2 of RXFP3 (Bathgate et al., 2013b). Zhang et al. used charge reversal mutations in both relaxin-3 and RXFP3 and concurs with the finding although they suggest that Arg12 is more likely to interact with Asp145 and not Glu244 as previously suggested. The difference however, may be due to structural disturbance from receptor mutagenesis, or Arg12 is chelated between the two negative charges (Zhang et al., 2014c). The C-terminus of the relaxin-3 B-chain is suggested based on modelling to adopt a more extended conformation instead of folding back against the helix when relaxin-3 is in complex with RXFP3, which allows hydrophobic interaction between Trp27 and Trp138 in RXFP3 (Rosengren et al., 2006;Hu et al., 2016). Recent research has suggested Asp128 is a control switch for RXFP3 activation and it is suggested that Asp128 keeps RXFP3 in a metastable state that can readily change conformation and relay binding events to intracellular signalling, as well as allowing internalization of the receptor once an agonist binds (Liu et al., 2016).

 



 Transmembrane2 Exoloop 1 Exoloop 2 LFVTNLALTDFQFVLTLPFWAV ENALDFKWPFGKAMCK TVKVMGEELCLVRFPDKLLGRDRQFWLGLYHSQ 128 138 141 145 244

Figure 1-4 Relaxin-3/RXFP3 interaction model. A) Model of the relaxin-3 RXFP3 complex. Relaxin-3 A- and B-chains are shown in cyan and green respectively, while RXFP3 is in pink. The key residues in RXFP3 that form electrostatic interactions with relaxin-3 binding residues Arg12, Arg16 and Arg26 are Glu141, Asp145 and Glu244. Trp27 inserts deep into the RXFP3 binding pocket. B) Sequences of human RXFP3 transmembrane 2 (highlighted in green), exoloop1 and exoloop2 segments with residues of interests identified.

1.3.2 Relaxin-3 modifications The pharmacological characterisation of relaxin-3/RXFP3 physiological functions has been difficult due to the relaxin-3 cross reactivity with RXFP1 and RXFP4 (Figure 1-2) (Liu et al., 2003a;de Avila et al., 2018). For instance, RXFP1 activation causes a significant increase in water intake when rats are injected intracerebroventricularly (icv) with relaxin-3; a response that is not associated with RXFP3 activation (de Avila et al., 2018). Therefore, the development of relaxin-3 analogues that are highly selective for RXFP3 as pharmacological tools for investigating relaxin-3/RXFP3 function has been a high priority in the field.

  In order to overcome the overlapping interactions that occur when relaxin-3 binds to both RXFP1 and RXFP3, R3/I5, a chimeric variant consisting of relaxin-3 B-chain and INSL5 A-chain was initially developed (Table 1-1) (Liu et al., 2005a). The relaxin-3 B-chain in R3/I5 retains a similar backbone conformation as native relaxin-3, even though the disordered N- terminus of the analogue is differently situated compared to the native peptide (Figure 1-5B) (Haugaard-Jonsson et al., 2009). Notably, the chimeric peptide retains its ability to bind and activate RXFP3 and RXFP4, but loses affinity for RXFP1 (Liu et al., 2005a). In addition to increased receptor specificity, R3/I5 has lower nonspecific binding signal when used as radioligand for binding studies and improved solubility since the INSL5 A-chain is more hydrophilic than relaxin-3 A-chain (Liu et al., 2005a). Truncation of R3/I5 at the C-terminus of the B-chain resulted in a relaxin-3 antagonist, R3B(Δ23-27)R/I5 (Table 1-1) (Kuei et al., 2007). The presence of an extra non-native C-terminal Arg at position 23 in this analogue, due to incomplete removal of a cleavage tag during recombinant protein processing (Kuei et al., 2007), was found to result in higher affinity binding to RXFP3 (Hossain et al., 2009).

Table 1-1 Summary of currently used relaxin-3 analogues.

Agonist modifications B-chain A-chain Relaxin-3 RAAPYGVRLCGREFIRAVIFTCGGSRW DVLAGLSSSCCKWGCSKSEISSLC R3/I5 RAAPYGVRLCGREFIRAVIFTCGGSRW

Antagonist modifications R3(B23-27)R/I5 RAAPYGVRLCGREFIRAVIFTCR

In order to reduce synthesis cost and effort, a minimised two-chain native relaxin-3 was developed. This simplified analogue, known as RXFP3-A2 (Table 1-1) consists of an N- terminally truncated relaxin-3 A-chain and a full relaxin-3 B-chain (Shabanpoor et al., 2012). The deletion of ten residues and removal of the intra-chain disulfide bond in RXFP3-A2 did not significantly affect its binding affinity and potency towards RXFP3 whilst retaining its selectivity for RXFP3 over RXFP1 (Shabanpoor et al., 2012). Furthermore, C-terminal truncation of the B-chain of this analogue in combination with the addition of an Arg at the C-

  terminus generated RXFP3-A3, which is a high affinity RXFP3-selective antagonist (Table 1- 1) (Shabanpoor et al., 2012).

More significantly, as the chain combination through regioselective disulfide bond formation is the major limitation to effective synthesis of , single-chain agonist and antagonist peptides based on relaxin-3 have also been developed. In these analogues, the native cysteine residues are substituted with serines (Haugaard-Kedström et al., 2011;Hojo et al., 2016;Praveen et al., 2019). Recently, Hojo et al. reported a truncated single B-chain relaxin-3 agonist with a “hydrocarbon staple” connecting the side chains of position 13 and 17 (Ac-R3B 10-27 [13/17 HC]) (Table 1-1) that showed only slightly decreased activation at RXFP3 and no activity at RXFP1, but retained activity at RXFP4 (Figure 1-5A) (Hojo et al., 2016). NMR spectroscopy analysis confirmed the staple was sufficient for restoring the helical structure of the B-chain, which is critical for agonist activity (Hojo et al., 2016). Interestingly, truncation and addition of Arg at the C-terminus of the B-chain as previously employed for R3B(Δ23- 27)R/I5 and RXFP3-A3 in a single chain linear variant, resulted in a highly competitive antagonist R3 B1-22R (Table 1-1), which is fully selective for RXFP3 and does not interact with either RXFP1 or RXFP4 (Figure 1-5A) (Haugaard-Kedström et al., 2011). Despite its high affinity, NMR studies have shown that R3 B1-22R is an unstructured peptide. Since without proper structure the single chain antagonist can still bind to RXFP3, it is indicative that the correct folding only needs to occur during ligand/receptor interaction and that for antagonists, conformational freedom may actually be advantageous (Haugaard-Kedström et al., 2011).

 

Figure 1-5 Cross reactivity between relaxin peptides and relaxin receptors. (A) Cartoon representation of relaxin-like family members binding to different RXFPs in comparison to modified agonist and antagonist analogues. Relaxin-3 analogues only bind to RXFP3 and RXFP4. (B) Comparison of structures between R3/I5 chimera (red) and (C) relaxin-3 (purple). Residues involved in receptor interaction remains spatially coordinated in R3/I5 similar to R3. Disulfide bonds are shown in yellow.

1.3.3 Biological functions of relaxin-3 Using immunocytochemistry and electron microscopy, relaxin-3 has been shown to be expressed in the neurons of the nucleus incertus (NI) of the rat (Tanaka et al., 2005) and mouse brain (Smith et al., 2010). From NI, relaxin-3 immunoreactive fibres project to the forebrain regions including the hypothalamus, septum, and hippocampus (Tanaka et al., 2005;Ma et al.,

  2007;Smith et al., 2010), which overlaps with regions expressing RXFP3 (Figure 1-6) (Sutton et al., 2004). These regions include the septohippocampal pathway, limbic system, sensory circuits, and regions that control arousal and stress responses (Smith et al., 2010). In the septohippocampal pathway, regions such as hippocampus and supramammillary nucleus (SuM) are known to regulate sleep/wake cycle and memory (Smith et al., 2010). On the other hand, paraventricular hypothalamic nucleus (PVN) and hypothalamus, a part of the limbic system, regulate feeding, metabolism and even stress. Other stress responsive regions include amygdala (part of limbic system) and bed nucleus of stria terminalis (BST) are also populated with RXFP3 (Smith et al., 2010). Additionally, the median raphe and periaqueductal gray, are associated with arousal control such as locomotion (Smith et al., 2010). These broad range innervations allow the relaxin-3/RXFP3 system to modulate different neurophysiological functions. An overview of relaxin-3 innervation and function is shown in Figure 1-6. Although the full complexity of the function of relaxin-3 in humans is not yet known, the potential of relaxin-3 as a key player in several physiological and pathophysiological functions ((Ma et al., 2017b;Olucha-Bordonau et al., 2018) makes it a very attractive neurosignalling system to therapeutically target.

 

Figure 1-6 Simplified schematic representation of relaxin-3 innervations and RXFP3 distribution by function. Amyg, amygdala; Arc, arcuate nucleus; BST, bed nucleus of stria terminalis; Cb, cerebellum; CgC, cingulate cortex; Cx, cerebral cortex; DBB, diagonal band of Broca; DG, dentate gyrus; DMH, dorsomedial nucleus of hypothalamus; DR, dorsal raphé nucleus; dSN, region dorsal to the substantia nigra; Hi, hippocampus; Hypo, hypothalamus; IC, inferior colliculus; IGL, intergeniculate leaflet; IPN, interpeduncular nucleus; LH, lateral hypothalamus; LPO, lateral preoptic area; MR, median raphé; NI, nucleus incertus; OB, olfactory bulb; PAG, periaqueductal gray; PnR, pontine raphé; PVA, paraventricular thalamic area; PVN, paraventricular hypothalamic nucleus; RSC, retrosplenial cortex; S, septum; SC, super colliculus; SuM, supramammillary nucleus; Thal, thalamus. Adapted from (Smith et al., 2014b).

1.3.3.1 Stress response Relaxin-3 has been shown to play a role in the stress responses through interactions with corticotropin-releasing factor (CRF). CRF type 1 receptor (CRF-R1), a mediator of CRF actions in response to stress, is co-expressed in relaxin-3 containing neurons at NI (Bittencourt and Sawchenko, 2000;Tanaka et al., 2005). Moreover, the endogenous receptor of relaxin-3, RXFP3 receptor, is highly expressed in the paraventricular nucleus (PVN) of the hypothalamus (Figure 1-6), a region where CRF is produced and commonly associated with stress and anxiety (Herman et al., 2002;Sutton et al., 2004;Smith et al., 2010). Anatomically, these evidence

  strongly indicate that relaxin-3 modulate stress responses by interacting with CRF. When rats were subjected to water immersion-restraint and repeated forced swim stress tests, an increase in relaxin-3 mRNA expression was observed (Tanaka et al., 2005;Banerjee et al., 2010), but its expression was reduced when a CRF-R1 antagonist was administered (Banerjee et al., 2010).

Since CRF also influence anxiety and depression (Lawther et al., 2015), a relaxin-3 agonist was investigated as a potential anxiolytic. Intracerebroventricular injection of the relaxin-3 agonist (RXFP3-A2) resulted in reduced stress behaviour in both rat and mice that had been pre-exposed to anxiety tests and drug respectively (Ryan et al., 2013a;Zhang et al., 2015a). The relaxin-3/RXFP3 system may not strongly modulate basal levels of anxiety, however, since in experimentally naïve (non-stressed) rats, RXFP3-A2 treatment did not alter anxiety behaviour of the rats compared to non-treated rats (Ryan et al., 2013a;Zhang et al., 2015a).

1.3.3.2 Appetite regulation Relaxin-3/RXFP3 signalling has also been implicated in metabolic control. Relaxin-3 agonists stimulate feeding behaviour without increased energy expenditure. Within an hour of icv injections of relaxin-3 or R3/I5, satiated rats showed a significant increase in food intake (McGowan et al., 2005;Kuei et al., 2007). In addition, pre-administration of a relaxin-3 antagonist analogue prior to icv injection of the relaxin-3 agonist analogues blocked this increased food intake (Kuei et al., 2007;Haugaard-Kedström et al., 2011;Shabanpoor et al., 2012). Chronic relaxin-3 injections, either by icv or iPVN, also showed increased feeding (Hida et al., 2006;McGowan et al., 2006) and additional cumulative weight gain in satiated rats (Hida et al., 2006;Sutton et al., 2009). R3/I5 expressed chronically in rats using a viral vector, rAAV1/2, also showed increased food consumption and an increase in body weight by 23%, with the effect evident from day 27 until the end of the 8 weeks of observation period (Ganella et al., 2013). Recently it has been revealed that the changes in feeding behaviour seem to differ between sexes (Calvez et al., 2017). When treated icv with 800 pmol relaxin-3, female rats gained weight but not in male rats (Lenglos et al., 2015). Chronically stressed and food restricted female rats also showed an increase in weight gain and higher expression of relaxin- 3 mRNA compared the male rats (Lenglos et al., 2013). The higher sensitivity of female rats to relaxin-3-induced food intake may be due to different expression of anorexigenic and

  orexigenic in male and female rats (Calvez et al., 2015;Lenglos et al., 2015). In a stress-induced binge eating model, female rats with higher expression of relaxin-3 were prone to develop a binge eating habit, but strikingly this effect could be prevented by administration of an RXFP3 antagonist, while no effect was seen in binge-eating resistant rats (Calvez et al., 2016).

Interestingly, in mice, the feeding behaviour is significantly different to rats. Agonist injections did not increase food intake, but antagonist peptide alone caused a reduction in food intake in mice models where the feeding drive is high (Smith et al., 2013b;Smith et al., 2014a). Similar effects were found in salt appetite studies when wild-type mice were salt depleted but when treated with the relaxin-3 antagonist R3 B1-22R showed a significant reduction in salt intake. Importantly these feeding effects are RXFP3 specific as the effects are absent in Rxfp3 gene knock-out (KO) mice (Smith et al., 2015). These results are important as they highlight RXFP3 as a potential therapeutic target to combat eating disorders, either treatment with agonist (to treat anorexia) or antagonist (to treat obesity) (Ganella et al., 2012).

1.3.3.3 Alcohol dependence Relaxin-3/RXFP3 system has also been indicated in regulating stress-induced alcohol seeking in rats (Ryan et al., 2013b). Using a rat model of alcohol use and alcohol-seeking, it was observed that rats that consumed more alcohol showed higher relaxin-3 mRNA expression (Ryan et al., 2014), but treatment with relaxin-3 antagonists led to a significant reduction in alcohol intake and stress-induced reinstatement of alcohol-seeking behaviour (Ryan et al., 2013b). The effect is likely to be regulated by RXFP3 as Rxfp3 KO mice showed a reduction in stress-induced alcohol preference compared to wild type mice (Walker et al., 2015), whereas Rxfp3 mRNA expression in the NI was markedly increased following chronic alcohol intake (Walker et al., 2017a). The bed nucleus of the stria terminalis (BNST), which is densely populated with RXFP3 (Sutton et al., 2004), appears to play an important role in regulating alcohol-seeking in rats, as similar effects (attenuation of alcohol seeking) were observed whether relaxin-3 antagonist analogues were injected into BNST or centrally administered (Ryan et al., 2013b). It is also likely that the BNST and central amygdala may be working in tandem in attenuating alcohol reinstatement since both brain regions are responsive to stress (Walker et al., 2017b), have high expression of RXFP3 and are innervated by relaxin-3 neurons from the NI (Sutton et al., 2004;Ma et al., 2007) (Figure 1-6). Notably, this is regulated by

  CRF/CRF1 signalling from NI as a CRF1 antagonist injected into NI resulted in reduced alcohol intake of yohimbine-induced stressed rats (Walker et al., 2017a).

1.3.3.4 Arousal, memory and learning Relaxin-3 fibres and RXFP3 binding sites are present at the septohippocampal pathway (Ma et al., 2007;Smith et al., 2010;Haidar et al., 2017) that produces the hippocampal theta rhythm, which is involved in behavioural activation and arousal (Ma et al., 2009;Ma et al., 2018). RXFP3 activation has been shown to enhance hippocampal theta activity in both anesthetized and conscious rats (Ma et al., 2009). It was also observed that rats given the antagonist R3(BΔ23-27)R/I5 has impaired theta-dependent spatial working memory performance, which can be reversed with co-administration of the agonist R3/I5 (Ma et al., 2009). Furthermore, acute injection and chronic expression of R3/I5 in rodents has been shown to lead to higher locomotor activity (Sutton et al., 2009;Smith et al., 2013a), while in contrast relaxin-3 KO mice exhibit circadian hypoactivity on running wheels (Smith et al., 2012). Hypoactivity during the active phase was also observed in both male and female Rxfp3 gene KO mice (Hosken et al., 2015), consistent with the behavioural changes observed previously. Impaired spatial memory was also observed when the dentate gyrus, a region vital for learning and memory, was depleted of RXFP3 (Haidar et al., 2017). Given that the major source of relaxin-3 is the NI, it is likely NI is also involved in regulating behavioural activation. An earlier study showed that stimulation of the NI region induces hippocampal theta rhythm in anaesthetised rats (Nunez et al., 2006), whereas NI inactivation using lidocaine caused spatial memory loss in rats when tested in Morris water maze (Nategh et al., 2015). Associated with this behavioural change, was a reduction in pCREB expression, a component vital to long term memory, in the hippocampus (Nategh et al., 2015). A more direct relationship between behavioural activation and NI was shown recently in awake rats using a chemogenetic approach (DREADDs). In this approach, a viral vector containing modified Gq-coupled human muscarinic receptor (hM3Dq) was injected into NI of rats to induce hM3Dq receptor expression that can only be activated by the pharmacologically inert clozapine-N-oxide (CNO) (Ma et al., 2017a). Rats stimulated with CNO in the NI region showed increased locomotor activity with an increased cortical electroencephalograph (EEG) desynchronization (Ma et al., 2017a). Figure 1-7 depicts some of the different effects that resulted from relaxin-3 analogues injections in rodent models.

 

Figure 1-7 Effects of relaxin-3 agonist and antagonist on different rodent behaviours. Excerpts of results taken from their respective articles. (A) Mice with induced-anxiety (pre- treated with FG-7142) spent less time in light zone, but the behaviour was reversed following relaxin-3 agonist treatment (Zhang et al., 2015a). (B) Wild-type mice treated with 3.8 nmol R3 B1-22R; following food depravation, showed reduced palatable food consumption which was not observed in relaxin-3 KO mice (Smith et al., 2014a). (C) Self administration of alcohol in rats treated with R3 B1-22R was also significantly reduced compared to baseline and vehicle treated (black bars) rats, with no change in water consumption (white bars) (Ryan et al., 2013b). (D) Rats given the antagonist R3(BΔ23-27)R/I5 has impaired theta-dependent spatial working memory performance, which can be reversed with co-administration of the agonist R3/I5 (Ma et al., 2009).

1.3.3.5 Non-neuronal function When administered peripherally, relaxin-3 promote antifibrotic actions via physiological off-target effects consistent with its observed in vitro pharmacology (Zhang et al., 2005). Relaxin-3-induced activation of RXFP1 increased expression of matrix metalloproteinase 2 levels when administered to rat ventricular fibroblasts in vitro (Bathgate et al., 2006) and also had an add-on effect on relaxin-2 in inhibiting transforming growth factor β1-stimulated removal (Hossain et al., 2011).

  In summary, structure-activity studies have provided the necessary knowledge to develop relaxin-3 agonists and antagonist, which in turn have been proven useful as pharmacological tools for the study of modulatory functions of relaxin-3/RXFP3. The involvement of relaxin-3/RXFP3 in several key brain functions makes it an attractive target to control feeding, as an anti-anxiety drug and for the treatment of other mental disorders. However, these results are all from peptides being delivered through central infusion into animal models (shown in Table 1 of (Kumar et al., 2017)). Unwanted RXFP1 related systemic effects can be eliminated by using selective analogues, however, extensive studies of the desired central effect are impeded by the inability of current peptides to be delivered via peripheral injection or even orally, as they are required to cross the blood-brain barrier (BBB). The development of new analogues that are able to be delivered peripherally is the most critical next step for advancement of the relaxin-3 field.

1.4 Blood-brain barrier

The importance of the protection of the brain is evident from the unique barrier that prevent harmful compounds from entering it from the systemic circulation. The main constituents of the blood-brain barrier (BBB) are the endothelial cells surrounding the blood capillaries. Basal membrane and other cells such as pericytes and astrocytes communicate with the endothelial cells, helping to maintain the physical and enzymatic barrier of the BBB (Figure 1-8). Together, they form the neurovascular unit (Deeken and Loscher, 2007;McConnell et al., 2017). In between the endothelial cells, tight junctions prevent paracellular diffusion of most molecules (McConnell et al., 2017). Endothelial cells are also decorated with multiple efflux transporters, the most active being P-glycoprotein (P-gp), which can actively pump out toxic materials and other unwanted compounds (Figure 1-8) (Deeken and Loscher, 2007).

 

Figure 1-8 Comparison between peripheral and brain capillaries. A) Molecules can diffuse through the gaps between peripheral capillary. B) Brain capillary consist of endothelial cells held closely together by tight junctions (TJs) that restricts penetration of water-soluble compounds into the brain. The basal membrane, astrocytes and pericytes surround the brain endothelial cells. Luminal surfaces of endothelial cells are decorated with multiple transporters. Organic anion and cation transporters (Oatp) 2 and ABC transporters such as multidrug resistance-associated proteins (MRP) 1, 2, 4, and 5, p- glycoprotein (P-gp) and breast cancer resistance protein (BCRP) lines the luminal membrane. MRP4, OAT3 and Oatp 2 are present on the abluminal side (adapted from (Deeken and Loscher, 2007)).

The BBB selectively allow nutrients required for brain functions to pass into the brain via multiple mechanisms. Small hydrophobic nutrients (molecular weight <500 Da) that are not efflux transporter substrates get into brain parenchyma by passive diffusion (Georgieva et al., 2014;Oller-Salvia et al., 2016b). Polar molecules such as glucose and amino acids use specific transporters - glucose transporter (GLUT1) and large neutral amino acid transporter (LAT1), respectively, to transverse the BBB (Georgieva et al., 2014). Larger molecules and peptides require either specific receptor-mediated transcytosis (RMT) or adsorptive-mediated transcytosis (AMT) through electrostatic interactions with endothelial membranes (Malakoutikhah et al., 2011;Patel and Patel, 2017) (Figure 1-9). Both of these transport pathways are saturable and involve vesicle formation to facilitate movement through the

  endothelial cell, bypassing of the degradation pathway and finally release of the contents into the parenchyma via exocytosis (Figure 1-9 C – D) (Oller-Salvia et al., 2016b).

Figure 1-9 Overview of mechanism of transports for endogenous molecules across the BBB. (A) Small, lipid soluble substrate passes through the membrane by passive diffusion. (B) Amino acids, glucose and nucleosides uses carrier-mediated transport. (C) Large endogenous molecules such as insulin and transferrin are recognise by receptors on the luminal side of the BBB and transcytosed into the brain parenchyma. (D) Endogenous plasma proteins such as cationised albumin is transported across using non-specific adsorptive-mediated transcytosis.

Most peptide drug leads and hydrophilic compounds are not BBB permeable. Therefore, many strategies have been investigated in order to allow therapeutic agents to cross the barrier to treat CNS diseases, such as Alzheimer’s disease (Wong et al., 2019), cancer and psychological disorders. These strategies have been extensively reviewed in (Malakoutikhah et al., 2011;Georgieva et al., 2014;Oller-Salvia et al., 2016b;Patel and Patel, 2017).

1.4.1 Peptide drug delivery strategies across the BBB

1.4.1.1 Non-specific targeting Several strategies involving unspecific uptake of drugs across the BBB has been considered. These potential therapeutic agents are designed to cross the BBB either via passive diffusion or using AMT. The former strategy involves increasing hydrophobicity of the peptide whereas the latter involves electrostatic interactions with cell membranes, which is achieved through the use of cell-penetrating peptides.

  1.4.1.1.1 Lipidation A variety of lipidation strategies have been reviewed to convert peptides to drug leads by altering the physiochemical properties of the peptides (Zhang and Bulaj, 2012). Lipidated peptides have increased plasma stability and prolonged elimination time due to serum albumin binding (Zhang et al., 2017). As part of a potential gene delivery strategy, it was observed that a fluorescently-labelled DNA/fatty acid complex could still be seen in circulation 48 h post injection, indicating a long half-life of the lipidated complex cargo (Shen et al., 2013). Myristoylation has also been shown to prolong peptide-siRNA complex stability in serum, allowing more time for cellular uptake of the siRNA complex (Youn et al., 2014). For peptides, palmitoylation of the 13-residue P3V8, a peptidylic inhibitor, has been shown to improve peptide stability in rat plasma and brain homogenate without diminishing P3V8’s ability to inhibit retinal degeneration in a Drosophila model (Zhang et al., 2017). Cationisation using polyamines have also been shown to improve CNS bioavailability of proteins and peptides (Poduslo and Curran, 1996;Zhang et al., 2009a) and a combination of lipidation and cationisation of neuropeptides (, neurotensin and neuropeptide W) was found to increase plasma stability and increase in vivo efficacy (Bulaj et al., 2008;Green et al., 2010;Green et al., 2011).

The position and length of the lipids added onto peptides can however play an important role as these factors determine receptor-binding capabilities of the lipidated peptides (Zhang and Bulaj, 2012). For example, the N-terminally lipidated -releasing peptide showed better metabolic stability while retaining its function as an anorexigenic neuropeptide (Maletinska et al., 2015), but the binding affinity of the Leu-enkephalin opioid peptide was reduced upon lipidation due to the steric blocking of the N-terminus (Wang et al., 2006). The effects of lipidation was also clearly shown in the development of , a GLP-1 agonist (Knudsen et al., 2000). Mono-lipids up to C16 attached through γGlu spacer at position 26 of GLP-1 retained activity whereas usage of fatty di-acids longer than C14 was detrimental to receptor activity. Even though positioning the lipid at a different GLP-1 site can further improve the peptide’s activity at GLP-1 receptor, palmitoylation was introduced at Lys26 of GLP-1 as the best compromise in minimal change in peptide sequence with improved plasma stability and without significantly compromising potency for the receptor (Knudsen et al., 2000). The slightly reduced potency did not hinder the ability of liraglutide to cause weight lost in mice that were given subcutaneous injection of the peptide (Sisley et al., 2014). Further

  development has led to perhaps the most striking example of clinical importance, , in which a different lipidation and minor sequence modifications are sufficient to extend the half-life in vivo to seven days, allowing a once weekly dosing regimen (Lau et al., 2015).

1.4.1.1.2 Cell penetrating peptides (CPP) Instead of modifying the chemistry of the drug lead itself, a molecular “Trojan horse”, a ligand that can cross the BBB, can sometimes be coupled to the drug lead to promote permeability without affecting the affinity of the therapeutic compound (Malakoutikhah et al., 2011).

CPPs are a family of peptides that are able to transverse cell membranes and can be generally be divided into three groups based on their chemical characteristics – hydrophobic, amphipathic and cationic CPPs (Milletti, 2012). Typically, CPPs range in size from 5-30 amino acids, and cross the cell membrane either through endocytosis or direct translocation (Milletti, 2012). CPPs can be optimised to specifically target different receptor expression profiles or tumour microenvironments to improve the delivery of drug across different biomembranes (Zhang et al., 2016). When linked to large cargoes, CPPs typically undergo endocytosis and is assumed to cross the BBB using adsorptive-mediated transcytosis (Herve et al., 2008).

The most commonly studied CPPs are cationic in nature, such as TAT, penetratin and polyarginine. TAT(47-57), derived from the HIV transcription factor (Vives et al., 1997), has been shown to successfully translocate neuroprotectants through the BBB to protect the brain from ischemic injury (Aarts et al., 2002;Cao et al., 2002;Kilic et al., 2005), and treat bacterial and yeast meningitis in animal model brains (Liu et al., 2009b;Wang et al., 2010). Penetratin is another cationic peptide sequence derived from the DNA binding domain of Antennapedia, a Drosophila transcription factor (Derossi et al., 1998). Penetratin has been shown to be able to transport the anticancer drug, doxorubicin, into the brain (Rousselle et al., 2000), and has also been used as a linker for a single chain antibody fragment targeting prion proteins in mice model brain (Skrlj et al., 2013). Polyarginine is a synthetic CPP (Milletti, 2012) that has been used, among other things, to deliver an antineoplastic agent into the brain (Liu et al., 2014) and for gene delivery into cells (Yoo et al., 2017). Amphipathic CPPs such as SynB1 have also been described to have BBB transport capabilities (Milletti, 2012). Brain uptake of a chemotherapeutic drug and an antibiotic have been successfully enhanced by conjugation with the SynB vector by 6-fold and 7-fold respectively. (Rousselle et al., 2000;Rousselle et al.,

  2002). SynB vector conjugation to dalargin, a hexapeptide analgesic, also improved its analgesic properties as observed in mice tested using the hot-plate assay (Rousselle et al., 2003).

Although CPPs can improve efficacy of CNS drug delivery, they can be susceptible to proteolytic degradation (Palm et al., 2007) unless modifications were introduced to prevent proteolysis (Gregori et al., 2017). Besides, CPPs generally have low specificity since they penetrate into most cells and organs and therefore, potentially need to be used in conjunction with targeting groups to specifically target cargo to the BBB (Sarko et al., 2010).

1.4.1.2 Ligand targeting towards BBB receptors The most promising approach to drug delivery into the CNS is perhaps by receptor- mediated transcytosis (RMT), which involves binding of ligands to specific receptors that have higher expression on the luminal endothelial cells of the BBB compared to other tissues (Oller- Salvia et al., 2016b). This specific targeting can reduce side effects in other parts of the body (Vlieghe and Khrestchatisky, 2010) and ideally, these receptors are capable of recognising a broad range of substrates and transcytosis can be mediated without affecting the physiological role of the receptors themselves (Oller-Salvia et al., 2016b).

Several receptors, including the transferrin receptor (TfR), insulin receptor (IR), low- density lipoprotein receptor (LDLR) and the low-density lipoprotein receptor-related proteins (LRP) 1 and LRP2 are expressed on the brain endothelial cells (Vlieghe and Khrestchatisky, 2010) and have been extensively studied. How these receptors have been exploited to date is discussed in the following sections.

1.4.1.2.1 Transferrin receptor (TfR) TfR is expressed on a variety of cells and have high level expression on the brain endothelial cells (Jefferies et al., 1984;Ponka and Lok, 1999). It binds to transferrin and mediates transport of iron bound to transferrin into the brain (Rouault and Cooperman, 2006).

Rodent specific antibodies targeting TfR have been used to explore the therapeutic capabilities of antibody-mediated receptor targeting for delivery of drugs for Alzheimer’s disease (AD) and ischemia to the brain (Jones and Shusta, 2007). Although these studies

  provide a proof-of-concept and are indicative of pre-clinical safety, the results were not directly translatable to human clinical trials, as these antibodies did not recognize the human TfR (Jones and Shusta, 2007). Besides, antibodies may not be ideal for delivering drug cargo across the BBB as a study have indicated the possibility of the transferrin receptor-targeting antibody associating too strongly to the brain microvessels and not being released into the brain parenchyma (Johnsen et al., 2017). Recently, a new antibody capable of binding to both human and nonhuman primate TfR, that heterodimerises with β-secretase (BACE1) antibody to form anti-TfR/BACE1, has been developed. BACE1 is an attractive target for AD as it is one of the proteases that cleaves amyloid precursor protein, in the formation of Aβ aggregates. The anti- TfR/BACE1 antibody was shown to have an increased brain uptake and reduce Aβ aggregates in a primate model (Yu et al., 2014). However, although antibody use may have some success in CNS therapies, its production is expensive and there may be immunogenicity issues (Oller- Salvia et al., 2016b).

As an alternative to antibodies, specific targeting of receptors using peptides have been investigated. For instance, a nine-mer cyclic peptide CRTIGPSVC (CRT), was used to deliver a gene therapy that when used together with ganciclovir, significantly supress brain tumour growth in mice model (Staquicini et al., 2011). Two peptide sequences that interact with the human TfR, HAIYPRH (HAI) and THRPPMWSPVWP (THR), were identified through phage display have also been studied as potential shuttles for BBB drug delivery (Lee et al., 2001). Preclinical studies have indicated that HAI can be used to target brain glioma (Kuang et al., 2016;Wei et al., 2016). Furthermore, a 12-mer THR peptide was also shown to assist in delivery of a β-sheet breaker peptide for AD treatment in a rat model (Prades et al., 2012). THR also has a potential for gene delivery as the peptide is able to interact with a recombinant adeno- associated virus vector and is BBB permeable in the immortalised brain endothelial cells hCMEC/D3, and in vivo in mouse when injected systemically (Zhang et al., 2018). A study has also indicated that introducing branched THR shuttles can improve cargo cell uptake by 2.6- fold and permeability by 1.4-fold in in vitro BBB models (Diaz-Perlas et al., 2018a).

1.4.1.2.2 Insulin receptor (IR) IR is another attractive target as it is expressed on the BBB (Uchida et al., 2011) and highly expressed in the brain parenchyma (Rhea and Banks, 2019). It is a tyrosine kinase receptor that binds to insulin and transports it into brain for regulation of food intake, energy

  metabolism and reproduction (Gerozissis, 2003). Insulin is not explored as a carrier protein as interruption of physiological insulin transport across the brain may result in lethal insulin overdosing (de Boer and Gaillard, 2007). Therefore, antibodies that target the IR have instead been evaluated. A humanized version of the antibody, known as HIRMAb, was developed (Boado et al., 2007b), and has been shown to ferry neuroprotectants into the brain parenchyma as potential treatment for ischemia and chronic neurodegenerative diseases such as Parkinson’s disease (PD) (Boado et al., 2007a;Boado et al., 2010). One of the most successful IR targeting approaches currently in clinical trials uses a fusion of α-L-iduronidase and HIRMAb to treat Hurley’s syndrome, a lysosome storage disorder, as a form of enzyme replacement therapy (Giugliani et al., 2018).

HIRMAb may be excellent at IR targeting; however a lack of recognition for the murine IR preclude initial preclinical studies in mice, necessitating the use of nonhuman primates for safety studies (Bell and Ehlers, 2014) which are more costly. Besides, an immune reaction can still occur when antibodies are used (Ohshima-Hosoyama et al., 2012). Although HIRMAb binds to a different receptor epitope than the endogenous insulin, competition between the antibody-modified drug conjugates and insulin for IR may also cause insufficient receptors to be available for one or the other interaction partner, leading to an inefficiency of the drug formulation or negative consequences for glucose metabolism (Georgieva et al., 2014).

1.4.1.2.3 Low-density lipoprotein receptor (LDLR) The main LDLR family members are LDLR, very low-density lipoprotein receptor, apolipoprotein E receptor, LRP1 and LRP2, and they all bind a variety of ligands (Hussain et al., 1999). LDLR functions as a plasma cholesterol regulator, by mediating the uptake and removal of lipoproteins containing apolipoprotein E and B (Hussain et al., 1999). These apoliproteins were covalently attached onto the surface of human serum albumin (HSA) nanoparticles (NP) and was successful in delivering loperamide loaded HSA-NP into the brain of mice via systemic injection. Following this treatment, an antinociceptive effect lasting an hour was observed in the tail-flick test which was not observed in mice given loperamide in solution (Kreuter et al., 2007). Another peptide variant, Peptide-22, also binds to LDLR but at a different binding site than LDL. It was decorated on NPs loaded with paclitaxel and showed to have 1.6-fold higher BBB permeability compared to undecorated NP and an ability to increase the survival rate of glioma-bearing mice (Zhang et al., 2013). Further optimisation of the peptide shuttle was carried out to improve stability and binding affinity to LDLR (Jacquot

  et al., 2016). The optimised ligand was then covalently conjugated to a human antibody and was successfully delivered into the brain when injected intravenously (iv) (Molino et al., 2017).

Targeting LDLR however may not be practical as LDLR expression is controlled by the brain lipid demand. Astrocytes will up-regulate luminal LDLR expression only when cholesterol level is limited in brain (Dehouck et al., 1994). This results in a saturable transport capacity that can be affected by food intake (Georgieva et al., 2014). Besides, LDLR is expressed in other tissues such as the liver and adrenal glands (Hussain et al., 1999;Jacquot et al., 2016), which may result in lack of homing specificity to the brain LDLR.  1.4.1.2.4 Low-density lipoprotein receptor-related protein (LRP) 1 LRP1 is highly expressed on the BBB and is responsible for transporting ligands such as lactoferrin, tissue plasminogen activator and secreted amyloid precursor protein (APP) (Hussain et al., 1999). Through sequence alignment of APP and other human proteins containing a Kunitz domain, a series of peptides known as angiopeps was identified, with Angiopep2 showing the highest transcytotic potential (Georgieva et al., 2014). Angiopep2 was shown to be able to translocate small molecules across the BBB to treat cancers (Regina et al., 2008;Thomas et al., 2009;Che et al., 2010;Bertrand et al., 2011) and fungal infection in the CNS (Shao et al., 2010). Its conjugate with paclitaxel, ANG1005, showed good tolerance in Phase I studies and is currently undergoing phase II clinical trials for high-grade glioma and breast cancer (Oller-Salvia et al., 2016b;Li and Tang, 2017). Although Angiopep2 has been used mainly to shuttle small molecules, the neuropeptide neurotensin (NT) has also been successfully conjugated to Angiopep2, resulting in 10 times more efficient BBB penetration than for native NT alone, while retaining NT’s analgesic properties (Demeule et al., 2014). An anti-HER2 antibody was also conjugated to Angiopep2 (ANG4043) for a treatment study of HER2-positive brain metastases (Regina et al., 2015). When mice were treated with ANG4043, they had a longer survival period than vehicle-treated mice (Regina et al., 2015).

Melanotransferrin (p97), like Tf, is an iron binding protein and has been identified as another potential ligand of LRP1 (Demeule et al., 2002). It is an attractive shuttle as it has low levels in blood, preventing competition between p97—drug conjugate and the endogenous material for binding to the receptor at the BBB (Karkan et al., 2008). In a proof of concept study, doxorubicin conjugated to p97 crossed the BBB into the brain parenchyma and was able

  to reduce tumour mass and increase survival rate in a brain glioma bearing mice model (Karkan et al., 2008). Currently, a dodecapeptide derived from p97 is ready to enter clinical trials (Oller- Salvia et al., 2016b). Two new peptide sequences have recently been discovered to bind to LRP1. Radiolabelled L57 (TWPKHFDKHTFYSILKLGKH) showed high BBB permeability in mice brain (Sakamoto et al., 2017), but further studies will be required to investigate its suitability for ferrying a cargo across the BBB. The other new peptide shuttle discovered, RAP12, facilitated delivery of paclitaxel loaded micelles across the BBB and improved survival of intracranial glioma suffering mice (Ruan et al., 2018).  1.4.1.2.5 Unknown active uptake mechanisms Apamin is bee venom peptide that consists of 18 amino acids and targets calcium- dependent K+ channels in the CNS (Hallworth et al., 2003). It is regarded as a neuroprotective agent, however, at high doses, undesirable neurotoxic effects have been observed. The residues responsible for its toxicity lie in the C-terminal region. Apamin is a very versatile natural peptide as it can be used as a scaffold (Volkman and Wemmer, 1997) to help maintain secondary structure of bioactive epitopes, and it is highly stable at various pH and temperatures due to its two disulfide bonds (Brazil et al., 1997;Dempsey et al., 2000). It was recently shown that apamin can be used to target the CNS to treat spinal cord injury (Wu et al., 2014), and it can also be modified to obtain a non-toxic apamin analogue that is resistant to proteases and crosses the BBB when tested in in vitro assay (Oller-Salvia et al., 2013). Apamin has also been successfully minimized into a nine-residue peptide, MiniAp4, that can shuttle a variety of fluorescent tags and nanoparticles into the brain (Oller-Salvia et al., 2016a). The exact mechanism of the transcellular transport of apamin is not yet known although it has been proposed to be receptor-mediated (Oller-Salvia et al., 2013). Another naturally occurring peptide toxin, chlorotoxin (CTX), has been isolated from scorpion venom. A similar optimisation process as for MiniAp4 was carried out for CTX to create MiniCTX3, which has the capacity to carry gold NP through a human BBB cellular model, although like for apamin its mechanism of action remains unclear (Diaz-Perlas et al., 2018b).

Through phage display screening techniques, several sequences have also been identified to promote BBB penetration. Using different strains of viruses, random gene sequences were displayed onto the phage coat and injected into systemic circulation of animal models or incubated with cells in in vitro BBB models (Diaz-Perlas et al., 2017). Phages that

  crossed the BBB were then collected, amplified by two to three more rounds of selection before the most common variants found in the brain were sequenced and identified (Diaz-Perlas et al., 2017). Using mice models, several brain homing peptides, such as cyclic PepC7 (CTSTSAPYC), GLA (GLAHSFSDFARDFVA), GYR (GYRPVHNIRGHWAPG) and TGN (TGNYKALHPHNG) were identified, (van Rooy et al., 2010;Li et al., 2011;Li et al., 2012) but these peptides’ ability to translocate therapeutic treatments and still be able to elicit effect at the site of interests remain to be determined. Only a handful of sequences identified from in vivo phage biopanning are able to transport drugs into brain with therapeutic benefits (Staquicini et al., 2011;Urich et al., 2015). In contrast, in vitro model phage selection is less technically challenging, less skill demanding and more cost efficient (Diaz-Perlas et al., 2017). Recently, using a co-culture of human brain capillary endothelial cells and bovine pericytes, SGVYKVAYDWQH (SGV), a 12-mer peptide, was identified to be BBB permeable. GFP conjugated to SGV showed 2-fold improvement in cellular uptake in an in vitro BBB model compared to the GFP-only control (Diaz-Perlas et al., 2017). Chemically synthesized peptide libraries have also been generated and tested against in vitro cellular BBB models to search for soluble and stable shuttles (Guixer et al., 2016).

1.4.1.3 Combinatorial shuttle and colloidal carrier system Different shuttles have also been used in conjunction with different colloidal formulations such as liposomes, nanoparticles (NP), micelles, emulsions and dendrimers for drug and gene delivery across the BBB (Spicer et al., 2018). Examples are shown in Table 1- 2. These nanocarriers may be preferred as they improve metabolic stability, and increase local concentration at the BBB for more efficient delivery into brain (Spicer et al., 2018). However, transport efficiency may depend on a combination of the cargo and colloidal carrier as Angiopep2 grafted onto liposomes showed poorer recovery of targeted particles in the brain than when an anti-Tf antibody was grafted onto liposomes (van Rooy et al., 2011).

  Table 1-2 Examples of combinatorial CNS drug delivery.

Peptide Colloidal Main cargo Observation Disease Ref shuttle carrier Angiopep2 Polymer-some Doxorubicin - Five-fold increase in Glioblastoma (Figueiredo et glioma cell uptake. al., 2016) - Toxic to U87MG glioblastoma cells NP Paclitaxel - Increased uptake into Brain glioma (Xin et al., glioma cells. 2011) - Inhibit proliferation and cell viability of glioma cells Dendrimer Tumour necrosis - Accumulation of Brain tumour (Huang et al., factor related complex in mice brain 2011) apoptosis-inducing - Increased survival in ligand plasmid mice bearing glioma DNA tumours Micelle Amphotericin B 1.6-fold increase in drug Fungal infection (Shao et al., accumulation in brain 2010) compared to non-decorated micelles TGN NP QSH (Aβ targeting QSH accumulate at Aβ region Alzheimer’s (Zhang et al., peptide), in mice brain disease 2014a) fluorescence probes Apamin Polymeric Enhanced recovery of mice Spinal cord injury (Wu et al., micelles from spinal cord injury 2014) TAT NP -gene Increased CGRP expression Cerebral (Tian et al., related peptide attenuates vasospasm in rat vasospasm 2013) (gene delivery) Liposome Doxorubicin 24-fold stronger cell Brain glioma (Qin et al., proliferation inhibition and 2011) increased survival period of C6 glioma rats in comparison to free doxorubicin NP Ritonavir NP delivered free drug is 800- HIV (Rao et al., fold higher in brain at 14 days 2008) compared to drug in solution. Penetratin Liposome Doxorubicin 1.4-fold enhanced Brain tumour (Sharma et al., and Tf translocation of doxorubicin 2014) across BBB with dual targeting liposome. 7-mer HAI NP EGFR siRNA - siRNA loaded HAI–NP Brain tumour (Wei et al., BBB penetration is TfR 2016) mediated - EGFR expression down regulated in glioma tumour cell of mice model Tf Dendrimer GFP Increased GFP expression by - (Huang et al., approximately 2-fold, after 2007) treatment with Tf conjugated dendrimer than without in both in vitro BCEC cells and mice model. THR NP CLPFFD Conjugate crosses the BBB AD (Prades et al., (β sheet breaker via TfR and recognizes Aβ 2012) peptide) aggregates in BBB co-culture model and rat model compared to unconjugated drug loaded NP

  1.4.1.4 Alternative approaches Highly potent and toxic drugs such as anticancer drugs can be delivered directly into the brain via intraventricular injection or intrathecally in order to achieve sufficient drug concentration in the brain and to prevent systemic toxicity, but it is an invasive approach (Barnabas, 2019). Alternatively, the BBB can be temporarily disrupted by using focused ultrasound (FUS) and combined with intranasal or iv injections in order to improve the drug delivery efficiency into CNS (Liu et al., 2010;Chen et al., 2014). A combination of FUS and doxorubicin loaded in cationic liposomes inhibited glioma without causing systemic toxicity in a rat model (Lin et al., 2016). More recently, it has been shown that an antibody uptake into the brain was significantly increased when used in conjunction with FUS to reduce tau phosphorylation in an Alzheimer’s disease mouse model (Nisbet et al., 2017). However, FUS is also invasive as the BBB is left open for somewhere between several hours and days depending on the acoustic parameters used (Chen et al., 2014) and the technique also requires highly skilled personnel, making it generally impractical (Malakoutikhah et al., 2011).

A less invasive approach involves intranasal delivery which can bypass the BBB and also allow drugs that are easily degraded in the systemic circulation to reach their target (Meredith et al., 2015;Barnabas, 2019). In the intranasal route, drugs enter CNS mainly via trigeminal and olfactory nerves located at the posterior region of nasal cavity (Agrawal et al., 2018). The approach has been shown to improve drug onset time with fewer side effects due to reduced systemic exposure (Quintana et al., 2016). Currently potential intranasal delivered CNS drugs are being studied for the treatment of neurodegenerative and psychiatric diseases (Meredith et al., 2015;Quintana et al., 2016;Agrawal et al., 2018). For instance, galantamine in chitosan nanoparticles was successfully delivered into different brain regions in a rat model (Hanafy et al., 2015), without any significant toxicity (Hanafy et al., 2016). Moreover, intranasal delivery of zolpidem encapsulated in polymers alleviates anxiety in mice (Borodina et al., 2018). A handful of drugs that use nasal delivery devices have been marketed (Mittal et al., 2014). However, several challenges with intranasal delivery are present such as nasal cavity irritation and a small surface area for absorption. In addition, the nasal region required for efficient intranasal delivery is inconvenient to target, resulting in compliance issue (Gizurarson, 2012;Quintana et al., 2016). Furthermore, pharmacodynamic responses are difficult to investigate due to differences in nasal cavity physiology of different species and the drugs absorption rate is different depending on whether a direct or indirect drug absorption pathway is favoured, leading to variation in dosing (Quintana et al., 2016).

  1.5 Aims

As discussed in the above sections, relaxin-3 is a neuropeptide with multiple functions in regulating appetite, stress, arousal and possibly psychological disorders. However, relaxin- 3 function can currently only be elucidated by icv injection, which is invasive and costly, and not ideal for models dealing with stress responses. In addition, the relaxin-3 analogues used are either unstructured and unstable, or does not bind to RXFP3 with sufficient affinity and selectivity. CNS disorders are very difficult to treat, as many drugs that could pharmacologically combat the diseases do not cross the blood-brain barrier. In order to validate RXFP3 as a novel bona fide therapeutic target, it is vital to be able to design and synthesize analogues with high selectivity and potency for RXFP3, that can efficiently cross the BBB under physiological conditions and have pharmacological effect at the targeted brain sites. This thesis addresses this key challenge through the following aims:

1. The interactions between the single chain antagonist R3 B1-22R and RXFP3 are unclear since SAR studies have mostly been focussed on relaxin-3. Therefore, the first aim is to determine the residues on the single-chain relaxin-3 antagonist that are interacting with RXFP3 and the conformation of the bound peptide.

2. Using the SAR results for R3 B1-22R and for relaxin-3, I will aim to design new generation antagonists and agonists that are based on single peptide chains but include disulfides or other cross-links for structural stability. Conformational restraints should improve in vivo half-life and ideally also improve the affinity and potency of the analogues for RXFP3.

3. Using optimised relaxin-3 analogues, I will aim to design variants with modifications that specifically target the BBB for delivery to RXFP3 in the brain. Analogues will be synthesised using Fmoc-based chemistry and their pharmacology evaluated.

4. In aim four, setting the scene for further preclinical validation in animal models, these novel analogues will be characterised in terms of toxicology and structural stability in vitro. The ability of these analogues to cross the BBB will also be investigated using in vitro models.

 

Chapter 2 Materials and methods

  2.0 Materials

All standard amino acids were purchased from GL Biochem (Shanghai, China) and Mimotopes (Victoria, Australia). Fmoc-Gly(Pentenyl)-OH (Fmoc-Pg), Benzotriazol-1-yl- oxytripyrrolidinophosphonium hexafluorophosphate (PyBOP), and 1- [Bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide hexafluorophosphate (HATU) were purchased from ChemPep (Florida, United States). Fmoc- Asp(Opis)-OH, Fmoc-Glu(Opis)-OH and Fmoc-Lys(Mtt)-OH were purchased from CS Bio (Shanghai, China). Fmoc-Dap(Mtt)-OH, and O-(6-Benzotriazol-1-yl)-N,N,N′,N Tetramethyluroniumhexafluorophosphate (HBTU) were purchased from Mimotopes (Victoria, Australia). Dimethyl sulfoxide (DMSO) was purchased Chem-Supply (South Australia, Australia). Fmoc-Trp(Boc) Tentagel S-PHB resin (0.23 mmol/g) and Tentagel-XV-RAM (0.21 mmol/g) were purchased from Rapp Polymere (Tuebingen, Germany). Fmoc-Trp(Boc)-Peg- PS (0.19 mmol/g) and PAL-PEG-PS resins (0.20 mmol/g) were purchased from Applied Biosystems (Victoria, Australia). LC-MS grade formic acid (FA) (OPTIMA) was purchased from Thermo Fisher Scientific (Queensland, Australia). Fmoc-(S)-2-(4-pentenyl)Ala-OH (Fmoc-Pa), Tris(2-carboxyethyl)phosphine (TCEP), N,N-Diisopropylethylamine (DIPEA), 3,6-Dioxa-1,8-octane-dithiol (DODT), triisopropylsilane (TIPS), Grubb’s catalyst (2nd generation), 1,3-dichloroacetone (DCA), dibromo-m-xylene (DBx), 2,2'-Dipyridyldisulfide (DPDS), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide(MTT), 4,4-dimethyl-

4-silapentane-1-sulfonic acid (DSS) and deuterium oxide (D2O)were purchased from Sigma Aldrich (NSW, Australia). All other solvents N,N-Dimethylformamide (DMF), dichloromethane (DCM), acetonitrile (ACN), trifluoroacetic acid (TFA), piperidine and methanol were purchased from Merck (Victoria, Australia) and were of peptide synthesis grade. All media for cell maintenance were obtained from Gibco, Thermo Fisher Scinetific (Australia) unless stated otherwise.

  2.1 Peptide synthesis

Peptide synthesis involves the formation of amide bonds between two amino acids and can be done in solution or using a solid support. Unlike solution phase synthesis that can be tedious, peptide synthesis using a solid support, as introduced by Merrifield, is more time efficient and easier to execute (Merrifield, 1963). In solid phase peptide synthesis (SPPS), the C-terminal amino acid is anchored onto a resin, allowing step-wise addition of amino acids to build the peptide chain in the direction of C- to N-terminus. Removal of the Nα-protecting group after each amino acid coupling allows the free amino group to interact with the activated carboxylic group of the incoming amino acid. Excess reagents used to drive SPPS reactions to completion can be removed via filtration and washing (Jensen et al., 2013). The repeated cycles of deprotection and coupling are conducted until the desired peptide sequence is obtained, and can be then released from its solid support through peptide cleavage. The process is illustrated in Scheme 2-1.

Depending on the orthogonal protecting groups chosen, resin chemistry and cleavage conditions, side-chains can remain protected during the release of assembled peptide from the resin. Thus, peptide fragments with different functional groups can be synthesised and conjugated together in solution to form the final product. The two common SPPS strategies for Nα-protection are based on tert-butoxycarbonyl (Boc)-chemistry or fluoren-9- ylmethyloxycarbonyl (Fmoc)-chemistry. The Fmoc-SPPS is generally preferred, as the peptide cleavage from the resin involves usage of TFA which is less hazardous than the hydrofluoric acid required for Boc-chemistry (Jensen et al., 2013).

 

Scheme 2-1 General steps of solid phase peptide synthesis.

All relaxin-3 antagonist variants were synthesised using Fmoc-based SPPS. The peptides were made on a 0.125 mmol scale on Pal-PEG-PS or Tentagel-XV-RAM resins, resulting in a favoured C-terminal amide (Haugaard-Kedström et al., 2011). Excess Fmoc- protected amino acids (4 eq), 0.5 MHBTU (4 eq), and 1 MDIPEA (4 eq) were used to couple each amino acid, and Fmoc-deprotection was carried out using 20% piperidine in DMF. Relaxin-3 agonist analogues were synthesized on Fmoc-Trp(Boc)-Tentagel S PHB resin or Fmoc-Trp(Boc)-Peg-PS, which gives rise to agonist variants with preferred acid at the C- terminus. Some variations to the standard approach were used when needed, e.g. analogue 7 was assembled on Fmoc-Trp(Boc)-Tentagel S PHB resin using 5 eq amino acid whereas analogues21 - 27 required 8 eq amino acids. Analogues 3 – 6 assembled on Fmoc-Trp(Boc)- Peg-PS used 4 eq amino acid. Unnatural amino acids coupling for the hydrocarbon-containing

  agonist analogues were done using 2 eq of Fmoc-Pa or Fmoc-Pg, 1.8 eq of HBTU, and 4 eq of DIPEA for a duration of 3 h at room temperature. Coupling of residues following these amino acid required 8 eq amino acid, 8 eq HATU and 12 eq DIPEA, shaken for 4 h at room temperature. In analogues 13, 14, 19 and 20, which contain consecutive Aib residues, the two residues following the first Aib required 8 eq of PyBOP and 16 eq of DIPEA shaken for 20 h at room temperature for complete coupling. Acetylation of the N-termini where required was carried out on resin using 9.2 mmol acetic anhydride and 2.7 mmol DIPEA in 15 ml DMF (2 x 5 min).  For lactam-bridge containing peptides, Fmoc-Asp(Opis)-OH, Fmoc-Glu(Opis)-OH, Fmoc-Lys(Mtt)-OH and Fmoc-Dap(Mtt)-OH were incorporated and side-chain deprotected on-resin using 1% TFA in DCM. Side-chain lactam bond cyclization was done on resin through treatment with HBTU (4 eq) and DIPEA (4 eq) under microwave heating, when required. PyBOP (1 eq) and DIPEA (4 eq) were used to form the lactam bond in MiniAp4 for analogue 28. General protocol for on-resin lactam-bridge formation is found in Scheme 2-2.

Fmoc

HN Opis Mtt O O

1) 1% TFA/DCM

2) HBTU/DIPEA (4:4) or PyBOP/DIPEA (1:4)

Fmoc

HN O Scheme 2-2 On-resin lactam stapling.

Upon completion of synthesis, final Fmoc deprotection of the N-terminus were carried out using 20% piperidine/DMF. Peptide containing resin were then washed with DMF and dried with DCM before the peptides were cleaved from resin using a mixture of

TFA/TIPS/DODT/H2O (92.5:2.5:2.5:2.5) for 2 h. TFA was evaporated off under vacuum and the peptides were precipitated using ice-cold diethyl ether. Precipitated peptides were redissolved in 50/50 buffer A (0.05% TFA in H2O) and buffer B (90% ACN and 0.045% TFA

  in H2O), before lyophilisation. The linear peptides were purified using C18 reversed phase columns on a Prominence HPLC system (Shimadzu) with a gradient of buffer A and buffer B. Characterisation of all analogues were conducted using electro-spray ionisation mass spectrometry (ESI-MS) on an API2000 (AB Sciex). Analogues were analysed for purity using an analytical RP-HPLC C18 column with a flow rate of 1 ml/min or 0.3 ml/min at a 1% gradient with the exception of analogue 13 which was analysed at a 2% gradient, and confirmed as > 95% pure. Theoretical and experimental masses and pI values for analogues in chapter 3 are presented in Table S-2 and Table S-3.

2.2 Oxidation of grafted peptides

The apamin grafted peptides were oxidised using random oxidation. The linear peptides were dissolved in 20 mM Tris, pH 8 at 0.25 mg/ml and stirred for 72 h at room temperature (Scheme 2-3A), according to previous reported conditions (Volkman and Wemmer, 1997). The reaction was stopped by lowering the pH to 3-4 with buffer A, before purification using RP- HPLC.

The linear VhTI grafted peptides were either oxidised using a random oxidation procedure or by regioselective disulfide bond formation (Scheme 2-3B-C). For random oxidation 0.1 mg/ml linear peptide was dissolved in 50 mM Tris, pH 8.6 and stirred at room temperature overnight. For regioselective disulfide bond formation, the first bond was formed by dissolving the Acm-protected linear peptide in 50% acetonitrile/H2O at a concentration of 0.33 mg/ml followed by addition of 0.1 ml/mg DPDS dissolved in methanol. The reaction was carried out overnight at room temperature before purification by RP-HPLC. In order to form the second disulfide bond, 0.5 mg/ml peptide was dissolved in 50% acetic acid in H2O and degassed with nitrogen for 5 min. 0.1 M HCl (0.1 ml/mg) was added to the peptide and I2 dissolved in 50% acetic acid was quickly added to the solution until it turned light brown. The peptide solution was degassed briefly and left stirring at room temperature for 2 h in the dark. The reaction was stopped using 1 M ascorbic acid, which was added until the solution changed to colourless. Buffer A was added to dilute the peptide solution before loading onto the RP- HPLC column for purification. The final peptides were characterised using analytical RP- HPLC, mass spectrometry and NMR spectroscopy.

 

Scheme 2-3 Peptide oxidation protocols. Panels A and B represents the random oxidation protocols for apamin and VhTI grafted peptides respectively. Panel C shows the scheme for regioselective oxidation. X – denote -OH or -NH2 functional groups for agonist and antagonist analogue synthesis respectively.

2.3 Peptide conjugation

Conjugation between peptide shuttles and relaxin-3 antagonist was conducted using the thiol-maleimide reaction. Additional Cys residue was added onto the peptide shuttle whereas the maleimide with a spacer, 6-maleimidohexanoic acid (6-MHA), was conjugated to relaxin- 3 antagonist, R3B6-22R. These modified shuttle and relaxin-3 antagonist were synthesised using the general peptide synthesis described in the previous section. The reaction scheme is shown in Scheme 2-4. Peptide shuttles containing a free cysteine residue were dissolved in degassed 4X PBS (pH 7.0) at a concentration of 0.5 mg/ml. 6MHA-R3B6-22R dissolved in DMF at 1.9 mg/ml was added to the solution of peptide shuttle in the presence of TCEP prior to being degassed again. The ratio of peptide shuttle/R36-22R/TCEP was 1.3:1:1. The peptide mixture was stirred for 2 h at room temperature under nitrogen conditions. The conjugation reaction was quenched using 20% acetic acid and diluted with water before purification using RP-HPLC.

  O O O H A N CH C NH + N 2 (CH2)5 O NH CH2 HS GVRLSGREFIRAVIFTSR-NH2

Peptide shuttle 6-MHA-R3B6-22R

pH 7.0, 2 h, TCEP

O H A N CH C NH2

CH2 O O S N (CH2)5 O NH

GVRLSGREFIRAVIFTSR-NH2 Scheme 2-4 Peptide conjugation to plug-and-play shuttle analogues.

2.4 Hydrocarbon and halogen stapling

Analogue 22 containing Pa residues was stapled using ring-closing metathesis (RCM) on resin, similar to a previously described protocol (Hossain et al., 2015). The reaction was initiated after the second Pa residue was coupled onto position 13 of relaxin-3 B-chain, and is conducted as shown in Scheme 2-5A. 5% of 0.4 M LiCl in DMF and 0.2 eq 2nd generation Grubb’s catalyst in DCM were added into 0.02 mmol Fmoc-protected on-resin peptide. All mixtures used were degassed before and after addition to the resin. The reaction mixture was shaken in the dark at room temperature for 48 h. Once the hydrocarbon stapling was completed, the protected on-resin peptide was washed with DCM and DMF before continuation of the synthesis.

Analogues 24 – 27 were synthesized and purified using RP-HPLC before thioether stapling was conducted in solution as shown in Scheme 2-5B. 1.5 eq TCEP was added to analogues containing minimal amount of AB buffer dissolved in 100 mM NH4HCO3 (3 mg/ml), and the reaction was stirred for 1 h at room temperature. 6 eq of DCA or DBx dissolved in minimal DMF was then added to the reaction and further stirred for 2 h at room temperature.

  The thioether stapling reaction was quenched using 4% TFA and samples diluted with buffer A before purification using RP-HPLC.

A

Fmoc Relaxin-3 13-27 Relaxin-3 13-27 Fmoc 5% 0.4 M LiCl 0.2 eq Grubb’s catalyst (2nd generation) 48 h stirring at room temperature

O

B 6 eq. dichloroacetone S S

Cl Cl Relaxin-3 10-27 SH SH O

Relaxin-3 10-27 100 mM NH4CO3 1.5 eq TCEP

S S

Relaxin-3 10-27 Br Br

6 eq. dibromo-m-xylene Scheme 2-5 Stabilisation strategies for single-chain relaxin-3 agonist.

2.5 Cell culture

Chinese hamster ovary CHO-K1 cells stably transfected with RXFP3 were maintained in DMEM/Hams F12 media supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin, 2 mM L-glutamine and 400 ng/ml G418 for RXFP3-expressing cell selection (Van der Westhuizen et al., 2005). Human immortalised brain endothelial cells hCMEC/D3 cells were maintained in EBM-2-MV media (Lonza, Walkersville, MD) supplemented with 2.5% fetal bovine serum, penicillin/streptomycin and various growth factors from EGM-2 SingleQuots Kit (Weksler et al., 2005). Human neuroblastoma SH-SY5Y were maintained in RPMI 1640 with 10% FBS, and 1% penicillin/streptomycin.

  2.6 Competition binding assay

Receptor binding assays traditionally use competition with radiolabelled ligands to measure affinity. However, due to the radioactivity safety issues, lanthanides have been developed as an alternative to radiolabelled binding assays. Two of the special features of chelated lanthanide is their large Stokes’ shift and long-lasting luminescence, allowing time- resolve fluorescence that improves signal-to-noise (Josan et al., 2011). The competition binding assay used in these studies employs Eu-DTPA as the lanthanide-chelating complex. Eu3+ is dissociated from the complex at low pH using dissociation enhanced lanthanide fluoroimmunoassay (DELFIA) solution. This assay was developed recently and have shown to be an efficient and safe cell-based assay for measuring binding to RXFP3 (Haugaard-Kedström et al., 2015).

Determination of binding affinity in this study was carried out as previously described (Haugaard-Kedström et al., 2015). CHO-K1 cells stably expressing RXFP3 were seeded into 96-well plates at a density of 5 x 104 cells/well. Cell media was aspirated off and washed with 200 μl PBS prior to cell incubation with 5 nM Eu-DTPA-R3 B1-22R and increasing concentrations of unlabelled relaxin-3 analogues (0.001 nM - 10 μM) in binding buffer (20 mM HEPES (pH 7.5), 1.5 mM CaCl2, 50 mM NaCl) and 1% BSA. Unlabelled R3 B1-22R was used as reference compound. Cells were incubated with the peptide for 2 h. Peptide containing media was then removed and cells were washed with 200 μl PBS before addition of the DELFIA europium enhancement solution. Fluorescence measurements of the europium labelled tracer were recorded with a 340 nm excitation and 614 nm emission wavelengths. The assays were conducted with a minimum of three independent experiments. Data are throughout presented as mean ± SEM. pKi were determined using one-site fit Ki and a Kd value of 26 nM. Statistical analyses were conducted using the one-way ANOVA with uncorrected Fisher’s LSD in GraphPad Prism 8.

  2.7 cAMP functional assay

In the canonical signalling pathway, RXFP3 is coupled to the Gαi/o protein. Activation of the receptor thus causes the inactivation of adenylyl cyclase (AC) which in turn reduces cAMP accumulation (Kocan et al., 2015). In the cAMP functional assays used for these studies, forskolin is used to raise cAMP levels by stimulation of AC. The changes in cAMP can be assessed using cAMP reporter gene assay. In this assay, cAMP induces β-galactosidase expression, whereby addition of its substrate induces a colour change which can be detected using absorbance measurement (Shabanpoor et al., 2012).

The ability of novel relaxin-3 agonist analogues to inhibit cAMP accumulation were tested in CHO-K1 cells stably expressing RXFP3 transfected with a pCRE-β-galactosidase reporter plasmid, as described previously (Shabanpoor et al., 2012). Cells were stimulated with 5 μM forskolin and incubated with increasing concentrations of relaxin-3 analogues for 6 h. Media was then aspirated and cells frozen at -80°C. The cells thawed to room temperature were lysed using 25 μL buffer A (100 mM Na2HPO4, pH 8.0, 2 mM MgSO4, and 0.1 mM MnCl2), shaken at room temperature for 10 min. 100 μL buffer B (100 mM Na2HPO4, pH 8.0, 2 mM

MgSO4, 0.1 mM MnCl2, 0.5% Triton X-100 and 40 mM β-mercaptoethanol) was then added to each well and shaken for further 10 min at room temperature before addition of 25 μL β- galactosidase substrate (chlorophenol red β-D galactopyranoside). Absorbance reading were taken at 570 nm to detect colour change. Each concentration point was conducted in triplicates and each experiment was repeated three times independently. Data are expressed as mean ± SEM and analysed using GraphPad prism8.

2.8 Serum stability assay

Peptides as drug leads have advantages including high target specificity. However, unstructured peptides are known to be highly susceptible to hydrolysis by proteases that are present in the serum, resulting in short half-life. Therefore, as part of the initial screening, determination of peptide stability early on in the drug development is essential. This would allow redesign or elimination of peptides with low stability, before proceeding further in the development, thereby saving cost in the long run.

  The in vitro serum stability of peptides was determined by peptide spiking as previously described (Hossain et al., 2015). 855 μL of human male pooled serum was spiked with 45 μl peptide (1 mg/ml) to a final concentration of 0.05 mg/ml. 100 μl samples were taken out at different time points and the protease activity was quenched by addition of 900 μl ammonium acetate (0.1 M, pH 3). Samples were left on ice for 30 min before remaining peptide was extracted using solid phase extraction cartridges (Oasis HLB 3cc, Waters). For the extraction, cartridges were activated with 6 ml MeOH and then preconditioned with 70% ACN/1% FA solution (3 ml). 3 ml 1% FA solution was used to equilibrate the cartridge prior to sample loading. After sample loading, a washing step using 3 ml of 1% FA solution was undertaken before peptide analogues were eluted with 30% ACN/1% FA solution (3 ml). Eluted samples were lyophilised and reconstituted in 200 μl 1% FA. The amount of intact peptide is reported as relative to the amount of peptide recovered from timepoint 0 h. Quantification was carried out using multiple reaction monitoring (MRM) on a triple-quad LC-MS (API2000, AB Sciex) without addition of internal standard. Peptides were eluted using Kinetex C18 column (Phenomenex) at a flow rate of 0.3 ml/min for analogues 8, 9, 32, 38 and 39. The columns used for other peptides are listed in Table 2-1. Data analysed using non-linear fit equation in Prism 8.

2.9 MTT assay

The MTT assay is a high-throughput cell-based assay used to screen potential peptide cytotoxicity on cells (Mosmann, 1983). Viable cells will be able to convert the water-soluble MTT into the insoluble purple formazan in mitochondria through cleavage of tetrazolium ring by succinate dehydrogenase (Fotakis and Timbrell, 2006). The exact mechanism of MTT reduction in cells is unknown but it has been suggested that the process is mediated by NADH (Riss et al., 2004). The formazan can be dissolved in DMSO, resulting in a visible colour change that can be quantified using a spectrophotometer (Riss et al., 2004).

In this cell viability assay, SH-SY5Y neuroblastoma cells were used. Cell assays were conducted using peptide quantified with Direct Detect (Merck Millipore). SH-SY5Y neuronal cells were seeded into 96-well plates at a cell density of 5 × 103 cells/well. The cells were then differentiated with 10 μM retinoic acid and further grown in the incubator for 48 h. Cells were then treated with increasing concentrations of peptide (in serum-free media) for 2.5 h and incubated at 37°C and 5% CO2. Water and Triton X-100 were used as negative and positive

  controls of cell death, respectively. 20 μM mellitin was also used as a peptide-based positive control. Media-containing peptide conjugates were then removed and replaced with 100 μl/ well of fresh media containing MTT (0.5 mg/ml). A stock solution of 5 mg/ml of MTT in PBS was sterile-filtered prior to dilution in cell media. Cells were then further incubated for 4 h. MTT-containing media was removed prior to DMSO (100 μl) addition into each well. Cell viability were measured in triplicates. Absorbance was measured at 570 nm using a POLARstar Omega plate reader (BMG Labtech), and the percentage of viable cell calculated by the equation below. Data were presented as mean ± SEM.    !        ! where  = absorbance reading and  = Triton-X.

2.10 In vitro BBB penetration assays 2.10.1 Parallel artificial membrane permeability assay (PAMPA) PAMPA was developed to assess passive permeability across biological membrane in the absence of efflux transporters(Kansy et al., 1998). This assay has been adapted to screen for permeability across intestinal barrier (Chen et al., 2008), skin (Ottaviani et al., 2006) and the BBB (Prades et al., 2012), depending on the lipids utilised. As the assay is conducted in a 96-well plate format, this allows high-throughput screening. In our study, the Gentest Pre- coated PAMPA plate system (Corning, Australia ) was used to investigate passive permeability (Chen et al., 2008). The assay was conducted as per manufacturer’s protocol.

The PAMPA 96-well plate was pre-warmed to room temperature before 300 μl of 30 μM, 50 μM, 70 μM or 100 μM of peptide analogues in PBS were added to the donor wells. 50 μM metverapamil was used as positive control and 100 μM atenolol was used as the negative control. 200 μl PBS was added to each of the acceptor wells. Analogue 41 was dissolved in 2% DMSO/ 15% ACN/ PBS and the same buffer was added to the acceptor wells. The acceptor plate is then lowered onto the donor plate to form a sandwich and incubated at room temperature for 5 h. After the incubation period, the donor and acceptor plates were separated. Samples from both donor and acceptor wells were kept at -20 °C until the analysis using LC-

  MS. Permeability of the compound was calculated as per manufacturer’s protocol using the formula:

                !  2 where is permeability in cm/s.A = filter area = 0.3 cm , = donor well volume = 0.3 ml

 = acceptor well volume = 0.2 ml,  = incubation time = 18000 s,    = compound concentration in acceptor well at time , = compound concentration in donor well at time

, and             

2.10.2 Cellular uptake of peptides in hCMEC/D3 cells Cell-based assays using primary cells or immortalised cell lines have been developed to study the transport capabilities of drug leads across the BBB. Although primary cells from different sources such as rodents, bovine and porcine, when co-cultured with astrocytes, are able to form strong tight junctions, cultivating the primary cells can be expensive and labour intensive (Helms et al., 2016). An immortalised human brain endothelial cell lines is on the other hand easier to culture, and can be obtained relatively easily. These cells are cultured as a monolayer and also express characteristics that closely mimics the human BBB (Weksler et al., 2013). One of the most well-characterised immortalised cell line, the human brain endothelial cells, hCMEC/D3, has been a favourite among scientists for studying drug uptake, and by extension transport across the BBB.

Investigation into the ability of the novel analogues in crossing the BBB was conducted in an in vitro assay using hCMEC/D3 cells similar to the protocol described by (Kooijmans et al., 2012). hCMEC/D3 cells were seeded at 27,000 cells/cm2 in a 24-well plate and incubated overnight. Peptides at 166 μg/ml were added to each well in the 24-well plate and incubated for 5 min, 30 min, or 90 min. At the respective timepoint, the media containing peptide was removed and cells were washed with ice-cold PBS prior to cell lysis using 200 μl 1 M NaOH for 2 h before neutralisation with 100 μl 2 M HCl. Cell lysates were then centrifuged and the 200 μl supernatant was removed and stored at – 20 °C until further analysis could be carried out using LC-MS. The degree of peptide uptake into the cells was calculated using the formula below:

            where= compound concentration in cell lysate,= volume of cell lysate = 0.3 ml,= compound concentration added to well = 166 μg/ml, and= volume of media containing analogue = 0.2 ml

2.10.3 LC-MS analysis Mass spectrometry (MS) is a powerful analytical technique used to identify and quantify peptides of interest in solution. In ESI-MS, sample droplets are ionised by an electrical field, which allows peptides to be detected in the form of mass-to-charge ratio (m/z) (Zhang et al., 2014b). HPLC coupled with MS (LC-MS) allows peak fractionation and identification of components in complex peptides mixtures, which saves time and prevents potential sample loss from manual peptide purification (Zhang et al., 2014b). The selectivity of peptides detected can be further improved by using multiple reaction monitoring (MRM) (LC-MS/MS). In LC- MS/MS, a precursor ion of a peptide is chosen (Q1) together with a corresponding fragmented ion (Q3). This pair of ions act as a fingerprint of the peptide of interest which would enable specific peptide identification even for different peptides that co-elute. Due to this technique’s high selectivity and sensitivity, LC-MS/MS is the leading analytical technique in pharmacokinetic studies of therapeutic drug leads in different biological matrices (Magotra et al., 2017;Kumar et al., 2018).

For each peptide analogue tested, a standard curve of peak area ratio (analyte/IS) versus peptide concentration ranging from 0.5 μg/ml – 32 μg/ml was plotted. For the control methoxyverapamil a concentration range of 31 ng/ml – 2 μg/ml was used. Each analyte sample was spiked with an internal standard (IS) at a final concentration as shown in Table 2-1 . The standard curve for each peptide has a R2 value of ≥ 0.99 (Figure 2-1). For the analysis of the PAMPA assays 1 μl or 10 μl from the donor and 50 or 100 μl from the acceptor well samples were used, whereas for the analysis of the cellular uptake assay 150 - 200 μl samples, depending on the analogue tested, were analysed. R3B6-22R (IS) was added to these samples for a final concentration of either 2.5 μg/ml and 5 μg/ml. For analogue 33, analogue 40 was used as IS at a final concentration of 5 μg/ml. These samples were concentrated using

SpeedVac before redissolving in 50% buffer A-FA (0.1%FA/H2O)/50% buffer B-FA (90%

ACN/0.1% FA/10% H2O) for LC-MS analysis. Peptides were eluted using a gradient of buffer A-FA and buffer B-FA at a flow rate of 0.3 ml/min. The list of analytical columns used for LC-

  MS quantification are shown in Table 2-1. Peak detection and identification were carried out using MRM on a triple-quad LC-MS (API2000, AB Sciex). MRMs used for each peptide quantification are shown in Table 2-2.The ESI-MS was run in positive ion mode with the source set at 400°C, ion spray voltage of 4500 V, and ion source Gas 1 and Gas 2 at 60 psi and 40 psi, respectively. Samples were injected into the MS within the seven min window of expected peak elution time and the rest were directed to waste to avoid signal suppression. Peak areas for analytes and IS were integrated and quantified using MultiQuant 2.1.1 software.

Table 2-1 LC-MS columns and IS concentrations.

Peptides Analogue Column description IS final concentration R3 B1-22R 8 Aquapore, OD300, 7 µm, 5.0 μg/ml KKKpK-Ava-R3B6-22R 33 300 Å, 30 × 2.1 (i. d.) mm 5.0 μg/ml KKKpK-R3B6-22R 34 (Brownlee) 5.0 μg/ml Palm-KKKK-Ava-R3B6-22R 35 5.0 μg/ml Palm-R3B6-22R 36 5.0 μg/ml R3B6-22R S10Kp 37 5.0 μg/ml CTHR-R3B6-22R 31 2.5 μg/ml CSGV-R3B6-22R 30 Aeris peptide, 3.6 µm XB- 2.5 μg/ml Methoxyverapamil C18, 100 x 2.1 (i. d.) mm 2.5 μg/ml Atenolol (Phenomenex) 2.5 μg/ml EEKpE-Ava-R3B6-22R 41 5.0 μg/ml RLS-R3B6-22R 29 Polar c18, 2.6 µm 100 Å, 50 2.5 μg/ml x 2.1 (i. d.) mm (Kinetex) Angiopep2-R3B6-22R 32 XTerra® MS C18 5 µm, 150 2.5 μg/ml ApaR3B 1-22R 9 x 2.1 (i. d.) mm (Waters) 2.5 μg/ml CMiniAp4-R3B6-22R 28 2.5 μg/ml

  Table 2-2 MRMs used in the detection and identification of novel analogues.

Peptide Analogue Q1 Q3 Methoxyverapamil 485.0 165.1 150.2 91.2 Atenolol 267.0 145.0 74.0 56.0 R3 B1-22R 8 656.5 70.0 120.0 86.2 R3B6-22R IS 516.4 70.0 120.0 86.0 ApaR3 B12-22R 9 584.9 69.7 83.9 120.1 CMiniAp4-R3B6-22R 28 654.6 70.2 84.0 120.1 RLS-R3B6-22R 29 699.2 69.9 86.2 112.0 CSGV-R3B6-22R 30 762.8 84.0 72.0 110.3 CTHR-R3B6-22R 31 963.0 70.2 110.0 159.1 Angiopep2-R3B6-22R 32 932.3 120.1 83.9 69.9 KKKpK-Ava-R3B6-22R 33 583.2 84.0 70.0 56.3 KKKpK-R3B6-22R 34 703.9 84.2 70.0 128.9 Palm-KKKK-Ava-R3B6-22R 35 583.3 84.0 70.0 128.8 Palm-R3B6-22R 36 768.1 70.1 86.1 119.9 R3B6-22R S10Kp 37 586.2 70.0 84.2 119.9 KKKpK-Ava-R3B6-22R F20Aib 40 570.7 84.1 56.0 69.9 EEKpE-Ava-R3B6-22R 41 729.5 84.0 70.0 120.0

 

Figure 2-1 Standard curves for peptide quantitation by LC-MS/MS. Weighting factor of 1/x is used for all standard curves with the exception of methoxyverapamil, R3 B1-22R (8), KKKpK-Ava-R3B6-22R (33) and R3B6-22R S10Kp (37) for which a weighting factor of 1/x2.

  2.11 NMR spectroscopy

Three dimensional (3D) atomic level structures of peptides and proteins can be determined using nuclear magnetic resonance (NMR) spectroscopy or x-ray crystallography. However, for small peptides NMR is more convenient as it can be used to determine peptide structure in solution without the need for crystallisation (Schroeder and Rosengren, 2020). In NMR spectroscopy the resonance frequencies, referred to as chemical shifts, of individual hydrogen, carbon and nitrogen atoms are determined. The chemical shifts are dependent on the environment surrounding the nuclei, primarily the local chemical structure, and together with the inter-nuclei interactions referred to as couplings, they result in signature patterns or “spin system” for each amino acid (Edwards and Reid, 2001). These can be identified using two dimensional (2D) double-quantum filtered correlation spectroscopy (DQF-COSY) and 1H homonuclear total correlation spectroscopy (TOCSY) (Edwards and Reid, 2001). Interactions between nuclei from amino acids that are close together in space can be detected in nuclear Overhauser effect spectroscopy (NOESY) (Edwards and Reid, 2001). Combination of these spectra are used for resonance assignments using sequential assignment strategies (Wüthrich, 1986). Additional shift assignments of interactions between hydrogen and 13C can be obtained from heteronuclear quantum coherence (HSQC) experiments. Secondary structure affects chemical shifts of the backbone in a predictive manner (Wishart et al., 1995). Therefore, it can be identified using e.g. deviations of peptide determined Hα shifts from the equivalent shifts in conformationally unrestrained amino acids in short random coil peptides (Wishart et al., 1995). The complete 3D structure of a peptide can be computationally calculated using structural constraints determined from NMR data using structure calculation programs such CYANA (Guntert, 2004). These restraints include inter-proton data based on the NOESY spectrum. Dihedral angle restraints based on coupling constants or generated based on the chemical shifts and database searches using e.g. TALOS-N (Shen and Bax, 2013). Hydrogen bonds can also be used for further refinement of peptide structure based on identification of hydrogen bond donors from analysis of solvent exchange or temperature dependence of amide protons (Cierpicki and Otlewski, 2001).

Samples for NMR spectroscopy were generally prepared containing 0.5 mg peptide in

0.5 ml of 90% H2O and 10% D2O and subjected to solution NMR spectroscopy studies. TOCSY with a mixing time of 80 ms, DQF-COSY and NOESY with a mixing time of 200 ms were recorded at 298K and 600 or 700 MHz using Avance III spectrometers equipped with a

  cryoprobe (Bruker). For the apamin variants, additional 13C HSQC data were recorded at natural abundance, allowing determination of 13C chemical shifts. Amide proton temperature coefficients were determined from TOCSY data sets recorded at temperatures 288-308K. The data were processed using Topsin and analysed in CARA (Keller, 2004). The data were referenced to 4,4-dimethyl-4-silapentane-1-sulfonic acid (DSS) at 0.0 ppm and secondary shifts determined using random coil shifts reported by Wishart et al. (Wishart et al., 1995).

2.12 Structure determination

Spectra of grafted relaxin-3 analogues were analysed using CARA (Keller, 2004). Resonance assignments were achieved using sequential assignment strategy (Wüthrich, 1986) and peaks in the NOESY spectra integrated and automatically assigned and translated into distance constraints using the software package CYANA (Guntert, 2004). Backbone dihedral angels (φ and ψ) where derived from chemical shifts using TALOS-N (Shen and Bax, 2013). Hydrogen bonds were identified from temperature coefficients and included as constraints for amides with a temperature dependence ΔδHN/ΔT > -4.6 ppm/K if a suitable hydrogen bond acceptor could be identified in preliminary structures (Cierpicki and Otlewski, 2001). The structure calculations were performed using torsion angle dynamics within CYANA (Guntert, 2004). The 20 lowest energy structure models from a total of 50 structures calculated were chosen to represent the solution structure and analysed using MOLPROBITY (Chen et al., 2010).



 

Chapter 3 Binding conformation and determinants of a single-chain peptide antagonist at the relaxin-3 receptor RXFP3 

  Citation: Haugaard-Kedström, L.M., Lee, H.S., Jones, M.V., Song, A., Rathod, V., Hossain, M.A., Bathgate, R.a.D., and Rosengren, K.J. (2018). Binding conformation and determinants of a single-chain peptide antagonist at the relaxin-3 receptor RXFP3. J Biol Chem 293, 15765-15776.  Author Contribution Han Siean Lee • Peptide synthesis • Statistical analysis • Figures 3-1 - 3-6 preparation • Table 3-1 – 3-2 preparation and finalisation • Supplementary Table 2-3 • Data analysis • NMR analysis • Manuscript drafting, review, editing and finalisation K. Johan Rosengren • NMR analysis • Data analysis • Statistical analysis • Manuscript drafting, review, editing and finalisation • Project funding • Project design • Project supervision and administration Linda Haugaard-Kedstrom • Project design • Peptide synthesis • Manuscript drafting Ross Bathgate • Competition binding assays • Statistical analysis • Data analysis • Supplementary Figure 1 and Table 1 • Manuscript review and editing • Project funding Maryon Jones • Peptide synthesis Angela Song • Peptide synthesis • Table 3-1 – 3-2 preparation Vishaal Rathod • Peptide synthesis Mohammed Akhter Hossain • Peptide synthesis of relaxin-3







  3.1 Introduction

Relaxin-3 is a highly conserved novel neuropeptide (Bathgate et al., 2002) that has been shown to modulate food intake (McGowan et al., 2005;Smith et al., 2014a), stress (Tanaka et al., 2005;Banerjee et al., 2010), arousal, memory and learning, and addiction in animal models (Ma et al., 2009;Ryan et al., 2013b). These functions are consistent with relaxin-3 innervation pathways and sites of expression of its endogenous receptor relaxin family peptide receptor-3 (RXFP3). Relaxin-3 has been shown to be highly expressed in the nucleus incertus in the rodent brain (Bathgate et al., 2002;Burazin et al., 2002). The relaxin-3 expressing neurons co-localize with the corticotropin-releasing factor type 1 receptor, and relaxin-3 expression has been shown to be up-regulated in stressed animals (Tanaka et al., 2005). Projection of relaxin-3 fibers to the paraventricular nucleus of hypothalamus and the bed nucleus strial terminalis, both regions with high RXFP3 expression, are thought to drive increased food intake and modulate pathways involved in addiction/ reward seeking, respectively (McGowan et al., 2005;Ryan et al., 2013b). Furthermore, projections of relaxin-3 neurons to the septohippocampus pathway modulate arousal and spatial memory (Ma et al., 2009).

Relaxin-3 is a member of the insulin/relaxin superfamily of peptide hormones. Although insulin and the insulin-like growth factors signal through tyrosine kinase receptors, the relaxin peptides, including relaxins 1-3, and insulin-like peptides 3-6 (INSL3-6) signal through RXFPs, which are G protein-coupled receptors (Shabanpoor et al., 2009). Currently four RXFPs have been identified. Relaxin-2 acts through RXFP1 (Hsu et al., 2002), INSL3 through RXFP2 (Kumagai et al., 2002), and INSL5 through RXFP4 (Liu et al., 2005b). Relaxin-3 is, in addition to its cognate receptor, RXFP3, also able to activate both RXFP1 and RXFP4 (Liu et al., 2003a;Liu et al., 2003b;Sudo et al., 2003).

 

Figure 3-1 Structure and receptor interacting amino acid residues of relaxin-3. The A-chain, shown in grey, consists of two antiparallel helices connected by a β-strand. The B-chain, shown in green, comprises a helical segment spanning from Gly11 to Cys22. Disulfides are shown in yellow. Side-chains important for binding to RXFP3 and for activating RXFP3 are shown in blue and red, respectively. The R3 B1-22R antagonist retains only the B- chain with the two cysteines at positions 10 and 22 replaced with Ser and the C-terminal five amino acid residues 23-27 replaced with a non-native Arg23.

Relaxin-3, like the other members of the insulin/relaxin superfamily, consists of two peptide chains cross-braced by one intra-chain and two inter-chain disulfide bonds (Figure 3- 1) (Rosengren et al., 2006). Multiple studies have investigated structure-activity relationships of relaxins in order to develop better analogues (Patil et al., 2017). For relaxin-3, amino acid residues Arg8, Arg12, Ile15, Arg16, Ile19 and Phe20 in the helical region of the relaxin-3 B- chain have been identified as important for binding to RXFP3. The same amino acid residues, except Arg12, are also essential for binding to RXFP4 (Kuei et al., 2007). For activation of RXFP3 and RXFP4, Arg26 and Trp27 of the B-chain tail are critical (Kuei et al., 2007). Interestingly, removal of N-terminal amino acid residues in the A-chain significantly reduces the binding affinity of relaxin-3 to RXFP1, whereas the affinity and potency for RXFP3 are unaffected (Hossain et al., 2008;Shabanpoor et al., 2012). Replacement of the relaxin-3 A- chain with the INSL5 A-chain results in increased specificity for RXFP3, by reducing the interaction with RXFP1 (Liu et al., 2005a;Haugaard-Jonsson et al., 2008). The ability of relaxin-3 to retain the interaction with RXFP3 even after changes to the A-chain strongly suggests that the A-chain functions only as a structural support, ensuring the correct conformation of the relaxin-3 B-chain (Shabanpoor et al., 2012). Indeed, it has now been shown that introducing a helical “staple” that supports the native conformation is sufficient to create

  a high affinity agonist (Hojo et al., 2016;Jayakody et al., 2016). Receptor mutagenesis studies have shed some light on the amino acid residues in the receptor that are involved in the relaxin- 3/RXFP3 interaction. Arg12, Arg16 and Arg26 of the relaxin-3 B-chain form electrostatic interactions with Glu244, Asp145 and Glu141 of RXFP3, respectively (Bathgate et al., 2013b;Zhang et al., 2014c). The flexible C-terminal of relaxin-3 has been suggested to undergo a conformational change to allow Trp27 to interact with Trp138 buried deep in the transmembrane region of RXFP3 (Hu et al., 2016).

Removal of the Arg26-Trp27 activation domain of relaxin-3 results in an antagonist peptide (Kuei et al., 2007). Intriguingly as a result of the recombinant production strategy a non-native Arg was left as an artifact in place of the C-terminal five amino acid residues during these studies, and this addition appeared to benefit affinity (Kuei et al., 2007). Given the interaction with RXFP3 is only dependent on the B-chain, this modification was further explored in single-chain variants, allowing the development of the antagonist R3 B1-22R (Haugaard-Kedström et al., 2011). The R3 B1-22R variant has binding affinity comparable with that of native relaxin-3 for RXFP3, despite R3 B1-22R being unstructured in solution (Haugaard-Kedström et al., 2011). R3 B1-22R may adopt a defined conformation upon binding to RXFP3, however its smaller size and flexibility will allow it to optimize interactions in a substantially different way than the larger and constrained relaxin-3. R3 B1-22R has been an important tool for studying the relaxin-3 system and shown that antagonizing RXFP3 reduces food intake in mice and alcohol seeking in rat addiction models (Ryan et al., 2013b;Smith et al., 2014a;Calvez et al., 2016). Therefore, an investigation of the structure-activity relationship of this peptide is a critical step in order to further develop it as a drug lead.

In this study, we have explored the structure activity relationships of R3 B1-22R through alanine scanning and followed up changes in pharmacology with additional substitutional approaches. The binding conformation was also explored by introducing helix- breaking amino acid residues, such as prolines, or helix-promoting features, such as side chain ‘staples’ to further understand the binding conformation of R3 B1-22R. We show that some features that are important in relaxin-3 are also important in the antagonist, but that additional amino acid residues also contribute to receptor binding. These findings provide new mechanistic insights into the activity of R3 B1-22R and highlight regions that can be explored for further improving the activity of this peptide.

  3.2 Results 3.2.1 Alanine scan of R3 B1-22R identifies amino acids involved in the RXFP3 interaction Using an alanine scan strategy a series of peptide variants, each carrying a single Ala substitution, was synthesized using standard Fmoc solid phase peptide synthesis. The influence of the amino acid side-chain substitutions in R3 B1-22R were assessed through a competition binding assays using an Eu-labelled R3 B1-22R tracer (Haugaard-Kedström et al., 2015). Truncation of the N-terminal part has previously been shown to be relatively well tolerated; thus, we focused on residue five onwards (Haugaard-Kedström et al., 2011). Strikingly removal of the majority of side-chains resulted in a reduction of the binding affinity of the mutant peptides to RXFP3, compared to R3 B1-22R (Table 3-1, Figure 3-2A). Substitution of amino acid residues before the helical region to Ala (Y5A, G6A, V7A, R8A, L9A, S10A) showed a modest but significant reduction in binding affinity by three to six-fold. A slightly larger reduction in affinity, up to 15-fold, was observed in the mutants R12A, F14A, I15A R16A, V18A, I19A, T21A, and S22A, in what constitutes the helical region in the native relaxin-3 structure. Intriguingly, although these included Arg8, Arg12, Ile15, Arg16 and Ile19, which all have been shown to be important for relaxin-3 binding (Kuei et al., 2007), amino acid residues that are not surfaced exposed in relaxin-3 due to the interaction with the A-chain, such as Phe14 and Val18 also appear to contribute to binding of the antagonist. In contrast, Phe20 which is a key residue for affinity in relaxin-3, can be replaced with an Ala without loss of affinity in R3 B1-22R. Position 17 is an Ala in the native sequence. Changing this to a polar Asn showed a significant 50-fold reduction in binding affinity compared to native R3 B1-22R. Ala17 is also buried in relaxin-3, and its replacement should be well tolerated if a similar binding position was adopted. A >300-fold reduction in binding affinity, essentially abolishing the interaction, was observed when the non-native Arg23 was mutated to Ala; this residue by far is the biggest contributor to affinity.

  Table 3-1 Effects of point modifications on binding affinity for RXFP3.

Peptides Sequence† pKI (± SEM) Alanine scan R3 B1-22R RAAPYGVRLSGREFIRAVIFTSR 7.69 ± 0.18 R3 B1-22R Y5A RAAPAGVRLSGREFIRAVIFTSR 7.12 ± 0.26b R3 B1-22R G6A RAAPYAVRLSGREFIRAVIFTSR 7.23 ± 0.20b R3 B1-22R V7A RAAPYGARLSGREFIRAVIFTSR 6.91 ± 0.22a R3 B1-22R R8A RAAPYGVALSGREFIRAVIFTSR 6.98 ± 0.08a R3 B1-22R L9A RAAPYGVRASGREFIRAVIFTSR 6.88 ± 0.13a R3 B1-22R S10A RAAPYGVRLAGREFIRAVIFTSR 7.05 ± 0.02a R3 B1-22R G11A RAAPYGVRLSAREFIRAVIFTSR 7.75 ± 0.08 R3 B1-22R R12A RAAPYGVRLSGAEFIRAVIFTSR 6.92 ± 0.07a R3 B1-22R E13A RAAPYGVRLSGRAFIRAVIFTSR 7.65 ± 0.01 R3 B1-22R F14A RAAPYGVRLSGREAIRAVIFTSR 6.78 ± 0.17a R3 B1-22R I15A RAAPYGVRLSGREFARAVIFTSR 6.49 ± 0.17a R3 B1-22R R16A RAAPYGVRLSGREFIAAVIFTSR 6.76 ± 0.16a R3 B1-22R A17N RAAPYGVRLSGREFIRNVIFTSR 5.98 ± 0.05a R3 B1-22R V18A RAAPYGVRLSGREFIRAAIFTSR 6.75 ± 0.02a R3 B1-22R I19A RAAPYGVRLSGREFIRAVAFTSR 6.51 ± 0.09a R3 B1-22R F20A RAAPYGVRLSGREFIRAVIATSR 7.51 ± 0.11 R3 B1-22R T21A RAAPYGVRLSGREFIRAVIFASR 6.75 ± 0.41a R3 B1-22R S22A RAAPYGVRLSGREFIRAVIFTAR 6.99 ± 0.05a R3 B1-22R R23A RAAPYGVRLSGREFIRAVIFTSA < 5 Abu scan R3 B1-22R V7Abu RAAPYG(Abu)RLSGREFIRAVIFTSR 6.96 ± 0.08a R3 B1-22R R8Abu RAAPYGV(Abu)LSGREFIRAVIFTSR 7.44 ± 0.08 R3 B1-22R R12Abu RAAPYGVRLSG(Abu)EFIRAVIFTSR 6.25 ± 0.03a R3 B1-22R F14Abu RAAPYGVRLSGRE(Abu)IRAVIFTSR 7.32 ± 0.17c R3 B1-22R I15Abu RAAPYGVRLSGREF(Abu)RAVIFTSR 6.12 ± 0.05a R3 B1-22R R16Abu RAAPYGVRLSGREFI(Abu)AVIFTSR 7.44 ± 0.07 R3 B1-22R A17Abu RAAPYGVRLSGREFIR(Abu)VIFTSR 7.59 ± 0.19 R3 B1-22R V18Abu RAAPYGVRLSGREFIRA(Abu)IFTSR 7.35 ± 0.15 R3 B1-22R I19Abu RAAPYGVRLSGREFIRAV(Abu)FTSR 6.37 ± 0.05a R3 B1-22R T21Abu RAAPYGVRLSGREFIRAVIF(Abu)SR 7.09 ± 0.04a Arginine R3 B1-22R R23Orn RAAPYGVRLSGREFIRAVIFTS(Orn) 5.40 ± 0.12a variants R3 B1-22R R23Har RAAPYGVRLSGREFIRAVIFTS(Har) 6.43 ± 0.15a R3 B1-22R R23K RAAPYGVRLSGREFIRAVIFTSK 5.85 ± 0.13a R3 B1-22R R23Cit RAAPYGVRLSGREFIRAVIFTS(Cit) < 5 R3 B1-22R R23Agb RAAPYGVRLSGREFIRAVIFTS(Agb) 6.96 ± 0.07a † All peptides were synthesized with an amidated C-terminus. a p < 0.001 versus R3 B1-22R b p < 0.01 versus R3 B1-22R c p < 0.05 versus R3 B1-22R

 

Figure 3-2 Competition binding of R3 B1-22R variants at RXFP3. The effect of representative R3 B1-22R substitutions including A) Ala substitutions B) Arg variants, C) helix breakers (Pro) or helix promotors (Aib), and D) side chain and backbone cyclization on the ability to compete for binding with europium labelled R3 B1-22R is shown. Data are presented as mean ± SEM from a minimum of three independent experiments.

  3.2.2 Only a small subset of amino acid residues contribute specific interactions when binding to RXFP3 Alanine is the smallest chiral amino acid residue. To investigate how specific the interactions contributed by the native side-chains were, we tested whether simply providing some additional ‘bulk’ was sufficient to restore some of the loss seen in the Ala variants. In this series, the non-protein encoded amino acid aminobutyric acid (Abu) was introduced, and again the affinity for RXFP3 was tested. Strikingly, improvements of binding affinity to a level not significantly different to the native peptide were seen for position Arg8, Phe14, Arg16, Val18 and Thr21 (Table 3-1). Substitution of the native Ala17 for Abu was also well tolerated; however, for Arg12, Ile15 and Ile19, no improvements were observed. At these three positions, the particular side-chain features thus appear required for the interaction, while for others, including the positive charges of Arg8 and Arg16, they are dispensable. These amino acid residues contact the receptor but can be replaced by other types, allowing different types of interactions.

3.2.3 Arg23 at the R3 B1-22R C-terminus cannot be altered The alanine scan highlighted the extraordinary importance of Arg23 for ensuring high affinity binding to RXFP3. Therefore, we set out to investigate whether any type of subtle change to this Arg could be tolerated, or even favoured, by introducing non-protein encoded Arg variants (Figure 3-3). Citrulline lacks a positive charge, while lysine and ornithine retain a positive charge but in the form of a smaller amine rather than the large guanidinium group. These modifications were all found to be detrimental for binding, reducing affinity by at least 70-fold. Homoarginine (Har) and norarginine (Agb) both retain the native guanidinium group, but its position is altered as the side-chain is extended or shortened by one carbon, respectively. For these the effect was less dramatic, yet the binding affinity was still reduced >20-fold. (Table 3-1, Figure 3-2B). Thus, both the nature and relative position of the Arg guanidinium group are optimal and critical for affinity. All analogues in this study were produced as C- terminal amides, as a free acid at Arg23 has been previously shown to be unfavorable and result in a ~10-fold drop in affinity (Haugaard-Kedström et al., 2011).

  NH2

NH NH 2 NH 2 NH2 HN NH2

HN NH HN O NH2 HN NH

H N COOH COOH 2 H2N H2N COOH H2N COOH H2N COOH H2N COOH Arginine (Arg, R) Norarginine (Agb) Homoarginine (Har) Citrulline (Cit) Lysine (Lys, K) Ornithine (Orn) Figure 3-3 Chemical structure of arginine and its variants. Non-proteinogenic amino acids were used to substitute arginine at position Arg23 to investigate the optimal features for binding to RXFP3.

3.2.4 Can an Arg at position 23 in a full length B-chain improve affinity of an agonist? Given the significance Arg23 in R3 B1-22R we wanted to explore whether introducing the same modification in a single chain variant retaining the activation domain would be beneficial for creating an agonist. The variant R3 B1-22RGSRW was designed to retain the relative spacing between the helical domain and the C-terminal Arg26-Trp27 activation domain, while in R3 B1-22RGGSRW the full five-residue C-terminal tail was added after Arg23. Notably, both variants showed poor binding with a pKi of ~5.5; thus the inclusion of Arg23 made no improvement over the linear B-chain (Figure S-1 and Table S-1), confirming that the interaction of Arg23 is not compatible with the endogenous receptor binding mode of native relaxin-3. A third variant that included a single Ala residue extension to the R3 B1-22R antagonist (R3 B1-22RA) was also produced. Again this variant showed significantly lower affinity than R3 B1-22R, highlighting that the chain cannot be C-terminally extended for Arg23 to be able to optimally engage RXFP3.

3.2.5 Modifications supporting a helical conformation of R3 B1- 22R are well tolerated, but ones that disrupt it are not Next, we turned to investigating the binding conformation of the flexible R3 B1-22R. Retro-inverso variants, in which the sequence order and chirality of each amino acid has been inverted, has been shown to be able to recreate the positioning of side chains in peptide loops, resulting in native like binding surfaces (Fischer, 2003). This strategy is however not compatible with recreating elements of secondary structure such as helices. Consistent with the need to refold into a helix the retro-inverso variant of R3 B6-22R showed no binding to RXFP3

  (Table 3-2). To further investigate whether the ability to form a helix upon binding to RXFP3 is required for R3 B1-22R, we incorporated Pro residues in the sequence. Pro residues are helix breakers due to the constrained conformation and lack of hydrogen bonding potential. Pro substitution at Leu9 and Glu13, before the helix and at the N-terminal part of the helix, resulted in a modest drop in affinity (Table 3-2, Figure 3-2C). In contrast, disrupting the C-terminal portion of the helix by the modifications A17P or F20P caused a total loss of binding.

Instead, substitutions promoting a helical structure were introduced. Aminoisobutyric acid (Aib) has, as a result of the substitution of its α proton with a methyl group, a conformationally restricted backbone that favours the dihedral angles adopted in helices (Mahalakshmi and Balaram, 2006). Single residue substitutions with Aib at positions Ala17, Val18 and Thr21 in the C-terminal part of the helix were all well tolerated, with neither variant showing impaired binding to RXFP3 relative to R3 B1-22R. Using a shorter version of the antagonist template R3 B6-22R, which has affinity comparable with R3 B1-22R (Table 3-2), we investigated whether incorporation of multiple Aib residues could drive helix formation and improve binding beyond the affinity of the native sequence. However, incorporating three Aib residues resulted in a reduction rather than improvement in binding affinity by approximately five-fold (Aib20,21,22). Binding affinity dropped a further 2.5 times when the positions Phe14, Val18 and Thr21 were replaced with Aib. This drop is likely due to the cumulative effect of removing side chains, each of which makes a small contribution to binding. Other helix promoting strategies were also explored, including introducing side chain ‘staples’ in the form of lactam bonds between helical positions i, and i+4. The sequence modifications to install Glu-Lys pairs at positions 16-20 and 18-22 were both found to completely disrupt binding, but notably some affinity was restored when the side chains were linked to stabilise a helical conformation. The substitution A17K, as previously noted, resulted in a significant drop in affinity; however, linking it to Glu13 via a lactam negated some of this effect. Finally, introducing a Lys-Asp pair at positions 13-17 also reduced affinity, but linking them to support a helical structure restored this drop to a closer to native affinity (Table 3-2, Figure 3-2D).

In addition to the short-range helical staples, we explored the effect of more global cyclization restraints. In the NMR structure of native relaxin-3, the N-terminal tail loops around and lies parallel to the helical segment, an arrangement that could be supported by a covalent

  link between these regions in the antagonist. Cyclizing R3 B1-22R via a lactam bond from the N-terminal amino group to the side chain carboxyl group of Glu13 caused a reduction in pKi from 7.69 to 6.78. Introducing the same linkage in the shorter antagonist version R3 B6-22R also reduced affinity, but to a lesser extent. To introduce a different anchor point, the native Glu13 was mutated to Gln13 and Thr21 to Glu21. Cyclizing the full-length peptide from the N-terminus to Glu21 in this variant was surprisingly well tolerated, despite this position being close to the key binding site of Arg23. However, cyclization of the shorter R3 B6-22R variant from the N-terminus to Glu21 resulted in complete loss of binding. In this variant, the N- terminal tail is too short to be able to wrap around to the C-terminal end of the helix without disrupting it.

Table 3-2 Effects of helical supportive or disruptive modifications on binding affinity for RXFP3.

Peptides Sequence† pKI (± SEM) Retro inverso R3 B6-22R retro inverso rstfivarifergslrvg < 5 Pro mutants R3 B1-22R L9P RAAPYGVRPSGREFIRAVIFTSR 6.82 ± 0.07a R3 B1-22R E13P RAAPYGVRLSGRPFIRAVIFTSR 6.47 ± 0.09a R3 B1-22R A17P RAAPYGVRLSGREFIRPVIFTSR < 5 R3 B1-22R F20P RAAPYGVRLSGREFIRAVIPTSR < 5 Aib variants R3 B1-22R A17Aib RAAPYGVRLSGREFIR(Aib)VIFTSR 7.49 ± 0.09 R3 B1-22R V18Aib RAAPYGVRLSGREFIRA(Aib)IFTSR 7.47 ± 0.13 R3 B1-22R T21Aib RAAPYGVRLSGREFIRAVIF(Aib)SR 7.22 ± 0.11c R3 B1-22R F14/V18/T21Aib RAAPYGVRLSGRE(Aib)IRA(Aib)IF(Aib)SR 6.47 ± 0.04a,b R3 B6-22R GVRLSGREFIRAVIFTSR 7.56 ± 0.17 R3 B6-22R F20/T21/S22Aib GVRLSGREFIRAVI(Aib,Aib,Aib)R 6.87 ± 0.12a,b Lactam bonds R3 B1-22R V18E S22K RAAPYGVRLSGREFIRAEIFTKR < 5

R3 B1-22R V18E S22K Lactam RAAPYGVRLSGREFIRAEIFTKR 6.34 ± 0.07a R3 B1-22R R16E F20K RAAPYGVRLSGREFIEAVIKTSR < 5

R3 B1-22R R16E F20K Lactam RAAPYGVRLSGREFIEAVIKTSR 5.32± 0.18a R3 B1-22R E13E A17K RAAPYGVRLSGREFIRKVIFTSR 5.82 ± 0.34a

R3 B1-22R E13E A17K Lactam RAAPYGVRLSGREFIRKVIFTSR 6.54 ± 0.17a R3 B1-22R E13K A17D RAAPYGVRLSGRKFIRDVIFTSR 5.83 ± 0.14a

R3 B1-22R E13K A17D Lactam RAAPYGVRLSGRKFIRDVIFTSR 7.09 ± 0.13a Cyclic R3 B1-22R RAAPYGVRLSGREFIRAVIFTSR 7.69 ± 0.18 variants R3 B1-22R cyclic RAAPYGVRLSGREFIRAVIFTSR 6.78 ± 0.06a R3 B6-22R GVRLSGREFIRAVIFTSR 7.56 ± 0.17

R3 B6-22R cyclic GVRLSGREFIRAVIFTSR 7.16 ± 0.08c,d R3 B1-22R S10A E13Q T21E RAAPYGVRLAGRQFIRAVIFESR 6.54 ± 0.07a

R3 B1-22R S10A E13Q T21E cyclic RAAPYGVRLAGRQFIRAVIFESR 6.82 ± 0.06a R3 B6-22R E13Q T21E GVRLSGRQFIRAVIFESR 7.44 ± 0.06

R3 B6-22R E13Q T21E cyclic GVRLSGRQFIRAVIFESR < 5 † All peptides were synthesized with an amidated C-terminus. a p < 0.001 versus R3 B1-22R b p < 0.001 versus R3 B6-22R c p < 0.01 versus R3 B1-22R

  Finally, we explored the inherent ability of R3 B1-22R to adopt a helical conformation in solution by studying the effect of addition of trifluoroethanol (TFE). TFE is well known to support the formation of helical structure; thus, we prepared a sample containing 70% water / 30% TFE and recorded 2D solution NMR spectroscopy data. The data were assigned using sequential assignment methods, and the secondary shifts, which are highly sensitive indicators of secondary structure, were determined by subtracting random coil chemical shifts from the observed chemical shifts. The comparison of Hα secondary shifts of R3 B1-22R in water and TFE, as well as the B-chain in relaxin-3 is presented in Figure 3-4. Remarkably, not only was the helical conformation of relaxin-3 restored, as the negative secondary shifts in the region 13-21 closely match the ones observed in the native peptide, but the positive secondary shifts resulting from the extended conformation in the region 8-11 were also observed in the linear peptide in TFE. Thus, R3 B1-22R readily adopts a native like conformation even in the absence of the A-chain in a more hydrophobic environment.

Figure 3-4 Secondary Hα shifts (experimental shifts – random coil shifts) for R3 B1-22R in H2O and 30% TFE solvents. For comparison, the secondary Hα shifts of the B-chain in native relaxin-3 are included. The secondary shifts, and consequently the secondary structure, in TFE closely mimics the structure in native relaxin-3.

  3.3 Discussion

Studies on relaxin-3 and related peptides have long been hampered by the complex and expensive synthesis of their two chain structure. The development of the single-chain relaxin- 3 antagonist R3 B1-22R was the first example of a variant that retained native-like binding affinity in a single peptide chain (Haugaard-Kedström et al., 2011). It has since been shown that linear variants of relaxin-2 can also be achieved, possessing potent anti-fibrotic activity through targeting RXFP1 (Hossain et al., 2016). These peptides have been a game changer as they allow large scale production and extensive in vivo studies into rodent physiology and behavior. Data showing that modulation of the relaxin-3/RXFP3 system controls important behaviors highlight R3 B1-22R as a potential therapeutic lead. However, further improvements of this peptide are required as unstructured amino acid sequences are readily degraded by proteases and current studies have relied on intracerebroventricular administration. For the peptide to be a viable candidate for further preclinical studies, it must be modified to be able to pass the blood brain barrier to engage its neuronal target after systemic administration. Although much is known about how relaxin-3 engages RXFP3, there is little information about how R3 B1-22R interacts with the receptor, both in terms of contributions from individual amino acid residues and in terms of its binding conformation. In this study, we investigated the changes to R3 B1-22R binding interactions that arise from amino acid substitutions to further understand the binding mode of the antagonist at RXFP3. These new insights set the scene for the development of next generation analogues.

The results from all point modifications introduced are summarized in Figure 3-5. As expected, and consistent with relaxin-3 structure activity data (Kuei et al., 2007), the Ala scan showed that Arg8, Arg12, Ile15, Arg16, and Ile19 contribute to RXFP3 binding. Notably, these contributions are significantly less pronounced than in relaxin-3, where many modifications at these sites resulted in a drop of affinity of 100-fold or more. Instead, a larger number of amino acid residues appear to make smaller contributions to binding in R3 B1-22R, as a reduction of side chain functionality through replacement with a small Ala residue at the majority of positions leading to a decrease in binding. These changes include positions that are involved in the interaction with the A-chain in native relaxin-3. Interestingly, Ala substitution at Phe20 did not show any change in binding affinity in the single-chain antagonist (Kuei et al., 2007). Taken together, these findings highlight significant differences in the binding of R3 B1-22R compared to relaxin-3 at RXFP3.

 

Figure 3-5 Summary of the SAR data from point substitutions in R3 B1-22R. Green circles denote position of R3 B1-22R that can be substituted with the mutant residue without causing a statistically significant change in binding affinity for RXFP3. Substitutions resulting in a drop of binding p<0.01 vs. R3 B1-22R are shown in yellow circles and substitutions resulting in a drop in affinity p<0.001 vs. R3 B1-22R in red circles. No changes to Arg23 were tolerated.

Given that many amino acid residues made small contributions, including ones that have evolved to maintain structural integrity in the two chain relaxin-3, as opposed to maintaining an optimal receptor interaction, we wanted to further investigate the type of interactions present (i.e. are they specific in terms of requiring a particular residue type, or can different types of side-chains provide similar level of binding contribution?). To analyze this, we made a series of mutants incorporating an Abu residue rather than Ala, increasing the side- chain ‘bulk’ by one carbon to allow additional receptor contacts. Indeed, we found that for both Phe14 and Val18, amino acid residues that would not be expected to be optimized for binding given their structural role in relaxin-3, incorporation of an Abu instead of Ala was sufficient to restore a native-like affinity. In contrast, for Arg12, Ile15 and Ile19, which are part of the native binding site, introducing an Abu made no improvement over an Ala. Intriguingly, at positions Arg8 and Arg16, which are critical for relaxin-3 binding, an improved affinity was observed when Abu was introduced instead of Ala. This suggest the positive charge is no longer essential for the interaction and the added peptide flexibility allows local adaptation of interactions in the antagonist. An Abu residue was tolerated at position Ala17, in contrast to the larger polar residue Asn, which resulted in a large drop in affinity. The latter likely induces clashes in the binding position of the antagonist. Overall, it appears the main contributions of interactions from the helical region are hydrophobic in nature. However, although the main contributions are hydrophobic in nature, there is no correlation between increased in hydrophobicity and improvement in affinity for the receptor. For example, the Phe20Ala mutant is significantly

  less hydrophobic but does not lose binding to RXFP3. Glu13Ala on the other hand has an increase in hydrophobicity but the mutant peptide does not show an improved affinity for RXFP3 compared to R3 B1-22R.

In contrast to the modest contributions from the helical amino acid residues, the C- terminal Arg23 is vital for the interaction with RXPF3, and an Ala substitution resulted in a >300-fold loss in binding affinity. Although Arg23 is not part of the original relaxin-3 sequence, it was serendipitously realized in a two-chain C-terminally truncated antagonist that an extra Arg residue at this position increased affinity (Kuei et al., 2007). Complete removal of the Arg in the two chain antagonist resulted in significantly reduced binding to RXFP3 (Hossain et al., 2009), but the effect is even more dramatic in the single-chain antagonist resulting in a complete loss of binding (Haugaard-Kedström et al., 2011). Here, we investigated whether the affinity can be further increased in the single-chain antagonist by using subtle variants of Arg. From Table 3-1, it was clear that the guanidium group is essential for binding. because replacement with an amine or amide functional group showed at least a 70-fold reduction in binding to RXFP3. Citrulline substitution resulted in the loss of measurable affinity. Surprisingly, even a change in side-chain length by one carbon (using Agb or Har) is sufficient to significantly disrupt the binding to the receptor. This is a strong indication that the spatial relationship between the key Arg binding site and the binding site of other groups are optimally targeted by R3 B1-22R, and no alteration to Arg at position 23 is tolerated. Given the significance of the Arg23 interactions, we tried to incorporate this residue in single chain variants that also included the activation domain of relaxin-3. However, in neither variant was the Arg able to increase affinity above the one already observed for the relaxin-3 B-chain. These data highlight that the binding mode of the antagonist is not compatible with the interactions required for receptor activation.

NMR studies have shown that the antagonist is disordered in solution (Haugaard- Kedström et al., 2011). It was suggested that a native conformation of R3 B1-22R would form when the peptide binds to RXFP3 (Haugaard-Kedström et al., 2011). Here, we used NMR spectroscopy to analyze the influence of changing the solvent conditions on the structure of R3 B1-22R. The addition of TFE induced a structure with striking resemblance to the one seen in native relaxin-3, confirming that it is a preferred conformation even in the absence of covalent fixation to the A-chain via disulfide bonds. In relaxin-3, the helix ends at Cys22; it is however possible that in the active conformation of R3 B1-22R, it extends all the way to include Arg23.

  If so, Arg23 may further contribute to activity by stabilizing this conformation through an electrostatic interaction between its positive charge and the negative end of the helical dipole. Such a conformation might also explain why a C-terminal amide is favored as the negative charge of a free C-terminus would interact unfavorably with the dipole and destabilize the helix.

Given the clear difference in binding contributions from individual amino acids between relaxin-3 and R3 B1-22R, we still wanted to further investigate the binding conformation of R3 B1-22R using sequence modifications. Retro-inverso strategy has been successful in a number of cases in mimicking the bioactivity of the parent L-peptide with an added advantage of remaining fully resistant to proteolytic activity (Wei et al., 2014). However, this strategy did not work in recapitulating affinity of the mutant relaxin-3 antagonist for RXFP3. It has been shown previously that this methodology is largely ineffective in peptides that are helical in nature when bound to their target protein (Li et al., 2010). This was demonstrated in tumour suppressing-negative inhibitor interaction, p53/MDM2, HIV-1 capsid protein-CAI peptide inhibitor and Src domain of Abl tyrosine kinase (Li et al., 2010). The affinity of these retro-inverso ligands for their respected proteins diminished significantly. Proline residues are known to break helical structures and favoring turns because of their constrained structure and lack of hydrogen bonding potential due to the missing amide proton. We therefore introduced Pro residues at different positions to see what effect conformational restraints would have on binding. Substitution at Leu9 or Glu13 was relatively well tolerated, however, binding was completely abolished with a Pro substitution at Ala17 or Phe20. Leu9 is located N-terminally to the helical segment and Glu13 in the N-terminal first helical turn, thus these changes may not majorly influence the overall conformation. In contrast, Ala17 and Phe20 are located towards the C-terminal end of the helix, closer to the key binding residue Arg23. At position 20, the native Phe and an Ala substitution are equally well tolerated, thus the Pro is unlikely to remove an interaction or introduce a clash. The effect is instead highly likely to be due its structural effect, preventing the formation of the correct helical conformation for the interaction with RXFP3.

Rather than disrupting the conformation, we investigated whether instead introducing modifications promoting a helical structure would be beneficial for binding. The achiral Aib residue contains an extra methyl group at the Cα carbon. The increased steric hindrance

  restricts the number of backbone Phi and Psi angle combinations that are favoured and has been shown to promote helicity in peptides (Mahalakshmi and Balaram, 2006). Introduction of a single Aib residue at positions Ala17, Val18 and Thr21 at the critical C-terminal end of the helix was well tolerated but did not improve binding. Multiple Aib residues were thus introduced to further strengthen the helical character. Rather than improving binding, this led to a modest reduction in binding. We speculate that this is likely a result of removal of multiple side chains that make some contribution to the binding, but the possibility that removal of too much of the freedom of the backbone is negatively influencing the favored backbone conformation cannot be ruled out.

Because Aib did not improve the binding interaction between R3 B1-22R and RXFP3, lactam bridge ‘staples’ were introduced. The constraints were introduced to mimic one turn of an α-helix (i, i+4). Similar stapling motif has been used for helix induction of a single-chain relaxin-3 agonist using lactam and hydrocarbon staples (Hojo et al., 2016) and also in the related peptide INSL3 (Shabanpoor et al., 2010). Although the affinity and activity of single- chain relaxin-3 agonists were successfully improved with helical stapling compared with linear B-chain analogs, no improvement was seen for the relaxin-3 antagonist. Notably, it was the installment of the point modifications that would allow the stapling that resulted in the drop in affinity. The formation of the lactam bridges did improve affinity for RXFP3 compared with the linear peptide variants, but not enough to rescue binding to a level comparable with the native single-chain antagonist. The positions of the lactams were chosen to face away from the key binding motif, thus if the antagonist was bound in a similar position to native relaxin-3, it would not interfere with the interaction. Indeed, this was the case for the stapled B-chain agonist, which would be required to bind in a native-like fashion to induce the structural changes required for receptor activation (Hojo et al., 2016). However, consistent with the point modification data, the unfavorable results from the stapling of the antagonist highlight a difference in binding position that brings both sides of the helix into contact with the receptor. The Lys-Asp lactam at position 13-17 was the one most tolerated for RXFP3 binding. The K- D lactam has been found to be one of the most promising strategies to induce α-helicity using a pentapeptide screen (de Araujo et al., 2014). The 13-17 position is furthest away from the C- terminal binding site and it was also found to be the one most suited for modification in the agonist variants (Hojo et al., 2016).

  Rather than trying to induce α-helicity, we tried alternative cyclizations. Cyclization has been widely used to stabilize peptides to improve protection against protease degradation, including in bioactive drug leads such as conotoxins (Clark et al., 2005). The relaxin-3 N- terminal region turns back along the helix in native relaxin-3, thus we envisaged that linking the N-terminus to a side-chain could provide some conformational support. However, none of the head-to-side-chain cyclizations improved binding affinity for RXFP3. For the shorter R3 B6-22R, cyclization from the N-terminus to position 21 via an introduced Glu residue resulted in complete loss of binding. This linkage likely compromised the ability to form the required helical conformation.

RXFP3 mutational studies primarily targeting acidic amino acid residues that are likely binding partners to the several important Arg residues in relaxin-3 have provided some insights into how native relaxin-3 binds to RXFP3 with high affinity (Liu et al., 2003b;Liu et al., 2005a). The peptide uses Arg12 and Arg16 in the helical domain (Kuei et al., 2007) to form electrostatic interactions with Asp145 and Glu244 in RXFP3, located in extracellular loops 1 and 2, respectively (Bathgate et al., 2013b). The neighboring Ile15, Ile19 and Phe20 in relaxin-3 likely provide complimentary hydrophobic contacts, creating an extensive continuous binding surface (Figure 3-6). The terminal Arg26-Trp27 residues are known to be critical for activation and target a second binding site deeper in the binding pocket of RXFP3. Glu141 at the top of TM2 has been identified as the binding partner for Arg26 and Trp138 as involved in the binding of Trp27 (Liu et al., 2003b;Liu et al., 2005a). We have shown here that the antagonist R3 B1- 22R likely adopts a helical conformation similar to relaxin-3 upon binding to RXFP3. It also utilizes amino acid residues in this helical region for interactions with RXFP3. The roles of features in RXFP3 in ligand recognition have also been explored (Wong et al., 2018a). The data from this study confirm that the binding sites of Arg12, Ile15, Arg16, Ile19 are likely the same as for native relaxin-3. Based on the hydrophobic residues mutated in the receptor, the refined model of agonist binding (Figure 3-7A) are in agreement with previous research with additional interactions identified. Ile15 and Ile19 interacts with L246 and L248 respectively, and these interactions are also present in the antagonist/receptor homology model (Figure 3- 7B). Similarly, for the antagonist, electrostatic interactions between Arg12 and acidic Glu244, and Arg16 and Asp145 of RXFP3 were identified through the novel acidic mutational studies conducted on the receptor. The lack of affinity loss of Phe20Ala in R3 B1-22R binding to RXFP3 could be explained by the different positioning of Phe20 that is further away from

  Phe364, preventing hydrophobic interactions that are present in the agonist (Figure 3-7). Notably, the antagonist relies far less on individual contributions from these positions, but instead, it is also able to target a second binding site in RXFP3, using features on the other face of the helix. We thus propose the smaller size in the absence of an A-chain allows a different positioning of the helix, probably deeper in the binding pocket, which promotes this interaction. This arrangement will allow Arg23 to insert deep into the pocket forming a perfect arrangement of interactions that are the key drivers of high affinity (Figure 3-6, Figure 3-7B). The interaction is shown to involve both a cation-π interaction with Trp138 and a salt bridge with Glu141 in RXFP3 (Wong et al., 2018a).

Figure 3-6 Differences in binding mode of R3 B1-22R and relaxin-3. Structural modifications in the form of staples, Aib substitutions and Pro substitutions strongly suggest that R3 B1-22R adopts a native like helical conformation when interacting with RXFP3. Important amino acid residues that are buried and involved in intra-molecular interactions in relaxin-3 do contribute to the RXFP3 interactions in R3 B1-22R. We propose the positioning of the helix differs in R3 B1-22R, creating an additional interaction surface. Positions that showed no tolerance to substitution are shown in red, positions with some tolerance are shown in yellow and positions where substitutions were widely accepted are shown in green. Positions in relaxin-3 that are not exposed were not probed for binding.

   

Figure 3-7 Comparison between (A) relaxin-3 binding to active state homology model of RXFP3 and (B) R3 B1-22R binding to inactive state homology model of RXFP3. (A) Refined model of relaxin-3/RXFP3 shows Ile15, Ile19 and Phe20 interacting with Leu246, Leu248 and Phe364 respectively. A salt bridge also occur between the Trp27 carboxy terminal and Lys271. (B) In the antagonist/RXFP3 model, only Ile15 and Ile19 have the same hydrophobic interaction as the agonist. Electrostatic interactions occur between Arg12/Glu244, and Arg16/Asp145. Amino acids on relaxin-3 agonist and antagonist that interacts with RXFP3 are shown in red. Black labeled residues of RXFP3 are interacting with the ligands. Figure adapted from (Wong et al., 2018a).

  3.4 Conclusion

In this study, we have reported extensive structure-activity data for the R3 B1-22R RXFP3 antagonist. This peptide has been shown to effectively reduce food intake and alcohol seeking in rodents, thus representing an important lead for treatment of obesity and addiction. Antagonist peptides are frequently developed from naturally occurring agonist ligands, and it was initially envisaged that the binding mode of R3 B1-22R would mimic the one of the native peptide. All evidence suggest that the antagonist adopts a helical conformation, but that it binds in a different configuration. These new insights suggest new avenues for the development of next generation antagonists. It is unlikely that stapling strategies will be beneficial for this peptide, given that its positioning in RXFP3 means that they are likely to introduce clashes with the receptor. Instead, efforts could be directed at modifying side chains, in particular for amino acid residues on the face of the helix that is buried in relaxin-3, as these represent a new non-native interaction that to date has not been optimized by nature or by medicinal chemistry approaches. This study also highlights the need for the development of individual SAR data for agonists and antagonists when targeting peptide GPCRs.

  

Chapter 4 Effect of peptide grafting and stapling on affinity and potency of single-chain relaxin-3 analogues at RXFP3

  Citation: Lee, H.S., Postan, M., Song, A., Clark, R.J., Bathgate, R.a.D., Haugaard-Kedstrom, L.M., and Rosengren, K.J. (2020). Development of Relaxin-3 Agonists and Antagonists Based on Grafted Disulfide-Stabilized Scaffolds. Front Chem 8, 87.

Author Contribution Han Siean Lee • Peptide synthesis • Statistical analysis • Figures 1, 3-7 preparation • Supplementary information figure preparation • Table 1-2 preparation and finalisation • Data analysis • 3D structure analysis • Serum assay • Manuscript drafting, review, editing and finalisation K. Johan Rosengren • NMR analysis • 3D structure analysis • Data analysis • Supplementary information figure preparation • Manuscript drafting, review, editing and finalisation • Project funding • Project design • Project supervision and administration Linda Haugaard-Kedstrom • Peptide synthesis • Table 1 and 2 preparation and finalisation • Figure 2 preparations • Serum assay • Manuscript drafting, editing, and finalisation • Project design Ross Bathgate • Competition binding assays • Activity assays • Statistical analysis • Manuscript review and editing • Project funding Michael Postan • Peptide synthesis • NMR analysis • 3D structure analysis Angela Song • Serum assay • Manuscript editing Richard Clark • Peptide synthesis and supervision • Manuscript editing and finalisation 

  4.1 Introduction

Following the determination of the solution NMR structure of the two-chain relaxin-3 peptide (Figure 1-5C) (Rosengren et al., 2006), detailed structure-activity relationship studies have been undertaken (reviewed in (Patil et al., 2017)). Studies on chimeric variants using combinations of A- and B-chains from different members of the insulin-like relaxin family identified the relaxin-3 B-chain as the key region for interacting with RXFP3 (Liu et al., 2005a). The first nine residues in the α-helix in the A-chain can be truncated (Hossain et al., 2008) and the A-chain disulfide bond removed (Shabanpoor et al., 2012) without severely affecting binding and potency at RXFP3. However, the overall structure of such analogues is severely affected, hence the function of the A-chain is likely to provide a structural support to the B-chain (Hossain et al., 2008). Point mutations throughout the relaxin-3 B-chain have identified residues that contribute to binding and activation of RXFP3 (Kuei et al., 2007). Arg8, Arg12, Ile15, Arg16, Ile19 and Phe20, which largely form a continuous binding surface on the B-chain helix, are important for binding. In addition, the two very last C-terminal residues (Arg26 and Trp27) are crucial for the activation of RXFP3 and their removal results in an antagonist (Kuei et al., 2007) (Figure 1-5C).  As the key residues are located in the B-chain, the B-chain alone can act as an agonist, albeit with significantly lower affinity (Liu et al., 2005a). A high affinity single chain antagonist based on a truncated B-chain in which the five C-terminal residues were exchanged for an Arg (R3 B1-22R) has been developed (Haugaard-Kedström et al., 2011). In this analogue the non-native Arg forms, as shown in chapter 3, significant interactions with RXFP3, compensating for the loss of affinity due to the compromised structure (Haugaard-Kedström et al., 2018;Wong et al., 2018a). However, linear peptides such as the R3 B1-22R antagonist and the relaxin-3 B-chain are generally readily broken down in serum and require redesign to be feasible drug leads.

For single chain agonists, the affinity can instead be improved by reintroducing helical structure in the absence of the A-chain through helical stapling using dicarba bonds (Shabanpoor et al., 2012;Hojo et al., 2016;Jayakody et al., 2016;Marwari et al., 2019). In addition, removal of protease-susceptible amide bond by hydrocarbon stapling has been shown to be quite effective in improving peptide stability (Green et al., 2013). Following on from the success of dicarba stapling in relaxin-3 agonists, it was interesting to investigate whether

  alternative helical support approaches could be introduced into single chain agonists and further improve binding and potency. Previous attempts using lactamisation and disulfide bond stabilising strategies were not fruitful (Shabanpoor et al., 2012). The failure might relate to the suitability of the staples utilised for the relaxin-3 B-chain. Thus, to address the side-chain linkage requirements, alternative stapling strategies were employed in this study including cysteine alkylation using halogen staples through reactions with dichlorocetone (DCA) and dibromo-m-xylene (DBx). These staples have been utilised previously to great success (Jo et al., 2012;Assem et al., 2015). Another strategy investigated in this study was to incorporate helicogenic Aib residues in single-chain relaxin-3 agonists.

The single-chain R3 B1-22R antagonist however, did not show improvement beyond the starting peptide affinity for RXFP3 when Aib residues were introduced, as shown in chapter 3. Global cyclisation and lactam bond stapling also did not improve the interaction between the antagonist and RXFP3 (Haugaard-Kedström et al., 2018). Therefore, strategies other than stapling needs to be considered when designing novel relaxin-3 antagonist analogues for improved affinity and stability. One such redesign concept is “molecular grafting” where essential residues for bioactivity are introduced at topologically equivalent positions of a scaffold with appropriate structure, as a way to improve enzymatic stability and/or activity of the binding motif (Wang and Craik, 2018). The key binding features of relaxin-3 are centred around the B-chain helix, thus an appropriate grafting scaffold for targeting RXFP3 would have to include a helix of similar length to native relaxin-3. Several α-helical scaffolds have been evaluated for peptide based drug development, including apamin (Li et al., 2009), the zinc finger Zif268 (McColl et al., 1999) and MCoTI-I (Ji et al., 2013). Apamin is an 18 amino acid long natural product derived from bee venom and consists of an α-helical region, which is stabilised by two disulfide bonds, and a flexible C-terminal (Figure 4-1A). Apamin has previously been used as grafting scaffold for the development of p53 inhibitors, and ligands for both the receptor and p32 (Li et al., 2009;Phan et al., 2010;Zhang et al., 2015b). Another scaffold identified as interesting for relaxin-3 grafting is a 34 amino acid residue trypsin inhibitor from the seeds of Veronica hederifolia (VhTI), which has a helix-loop-helix fold (Figure 4-1B). The structure is stabilised by two disulfide bonds cross-linking the helices at adjacent turns (Conners et al., 2007).

  In this study seven grafted relaxin-3 agonists and antagonists were designed by exploiting the two disulfide-stabilized α-helical peptide scaffolds apamin and VhTI (Figure 4- 1). The helical promoting residue, Aib, thioether stapling and pentenyl-glycine (Pg) were also introduced into single-chain relaxin-3 agonists as part of the investigation into alternative helical promoting strategies, synthesised as shown in Scheme 2-5. Their solution NMR structures, affinity for RXFP3 and their potential to induce RXFP3 activation were determined. 

Figure 4-1 Structural comparison of (A) apamin (red) and (B) VhTI (blue) with (C) the relaxin-3 B-chain (green). The apamin and VhTI scaffolds are stabilised by two disulfide bonds and include α-helices between residues 9-18 and 3-25, respectively.

4.2 Results 4.2.1 Peptide design rationale and synthesis The RXFP3 binding motif (Arg8, Arg12, Ile15, Arg16, Ile19 and Phe20) is predominantly situated on the solvent exposed side of theαhelix located in the native relaxin- 3 B-chain (Rosengren et al., 2006). The C-terminal Arg26 and Trp27, which are located in the highly flexible C-terminal part of the B-chain, extend from the helical segment and interact with the central pore of the transmembrane receptor helix bundle (Bathgate et al., 2013b;Hu et al., 2016;Wong et al., 2018a). These residues could thus potentially be grafted onto anαhelical scaffold and retain the native bioactive secondary structure. Table 4-1 lists seven novel analogues that were designed using the apamin and VhTI scaffolds and compares them to reference compounds. Analogues 1 (native relaxin-3), 2 (relaxin-3 B-chain) and 8(single-chain relaxin-3 antagonist) were used as reference compounds for comparison with the novel grafted analogues’ affinity and potency at RXFP3. 3 was designed using the apamin scaffold with

  relaxin-3 B-chain binding and activating motifs. In order to further improve the helical propensity of 3, 4 was designed to include Aib amino acids, known to induce helicity due to stereochemical constraint of an additional methyl group at the Cα (Mahalakshmi and Balaram, 2006). Val18 and Thr21 of relaxin-3 B-chain were substituted with Aib in 4, as it was shown in the previous chapter that at those positions, Aib substitution did not adversely affect affinity of R3 B1-22R for RXFP3 (Haugaard-Kedström et al., 2018). The trypsin inhibitor scaffold VhTI was also investigated by grafting the corresponding relaxin-3 residues Arg8, Ile15, Arg16, Phe20, and the flexible C-terminal residues Gly23, Gly24, Ser25, Arg26 and Trp27 to form analogue 5. The equivalent position of Arg12 is a Pro in VhTI, and may be structurally important, thus was retained in this variant. In addition, Arg20 of VhTI was mutated to an Ala to prevent charge and/or steric clashes. In analogue 6, the Pro of VhTI was replaced to include relaxin-3 Arg12 and the conservative change of Leu to Ile was also included to match Ile19 in relaxin-3. In analogue 7, the entire binding cassette was shifted one residue closer to the C- terminus to change the positioning of the motif on the helical surface to better match the topology of the interacting residues of relaxin-3. Finally, it has previously been shown that by truncating the last five native C-terminal residues in the B-chain and replacing them with an arginine, a single chain RXFP3 antagonist can be generated (Haugaard-Kedström et al., 2011). This forms the basis for constructing 9 and 10 that was grafted onto apamin and VhTI respectively.

  Table 4-1 Amino acid sequences of grafted RXFP3 agonists (3-7) and antagonists (9-10).

Ligand Sequence Analogue R3 DVLAGLSSSCCKWGCSKSEISSLC-OH 1

RAAPYGVRLCGREFIRAVIFTCGGSRW-OH R3 B1-27 RAAPYGVRLCGREFIRAVIFTCGGSRW-OH 2

Apamin (Apa) CNCKAPETALCARRCQQH-OH

Apa+R3B CNCKAPETARCAIRCVIFTSGGSRW-OH 3

Apa+R3B[V18Aib,T21Aib] CNCKAPETARCAIRCUIFUSGGSRW-OH 4

VhTI EQCKVMCYAQRHSSPELLRRCLDNCEK-OH

VhTI+R3B EQCKVMCYAQRHSSPELIRACLFNCGGSRW-OH 5

VhTI+R3B[G11,R12] EQCKVMCYAQRHSGRELIRACIFNCGGSRW-OH 6

VhTI+R3B[R12] EQCKVMCYAQRHSSPRLLIRCLIFCSGGSRW-OH 7

R3 B1-22R RAAPYGVRLSGREFIRAVIFTSR-NH2 8

Apa+R3 B12-22R CNCKAPETARCAIRCVIFTSR-NH2 9

VhTI+R3 B1-22R EQCKVMCYAQRHSSPELIRACIFTCR-NH2 10 U;aminoisobutyric acid. Bold letters indicate corresponding relaxin-3 B-chain peptide grafted onto the scaffold. Red letters; residues important for RXFP3 binding. Green letters; residues important to activate RXFP3.

The design of the second set of peptides was based on the success of hydrocarbon stapling positions of Ac-R3B10-27 [13/17 HC]; Glu13 and Ala17 (Table 4-2) (Hojo et al., 2016)). Aib residues were introduced at the same positions of the relaxin-3 B-chain (analogue 11), to observe if the simpler helical promoting strategy is sufficient to induce α−helical structure, improve affinity and activity of single-chain relaxin-3 agonist. Additional Aib residues were also introduced at positions Phe14, Val18, Thr21 or Ser22, which are not important for relaxin-3/RXFP3 interactions and have been used as stapling positions without adversely affecting ligand-receptor interaction (Jayakody et al., 2016;Marwari et al., 2019). These Aib residues were added into analogues containing Aibs at Glu13 and Ala17 to form a series of triple substituted agonist peptides, analogues 12 – 18. It was also of interest to investigate whether more than three Aib can be introduced, hence the design of 19 - 20. 21 was

  synthesised since affinity and potency of the linear Ac-R3B10-27 [13/17 HC] have not been assessed and it should reveal if the addition of hydrophobic side chains contribute to binding. Ac-R3B10-27 [13/17 HC] also has solubility issues due to its hydrophobicity, hence four Lys residues were conjugated to the N-terminus of the peptide, forming 22, to increase its hydrophilicity. For the alternative hydrocarbon stapling strategies, analogue containing hydrocarbon staple using pentenyl-glycine (Pg) staple was designed. This variant would contain the same side-chain staple as analogue Ac-R3B10-27 [13/17 HC], but lack the extra steric restriction of the backbone. Despite multiple attempts using different stapling conditions, the analogue could not be stapled and only its linear version (23) could be assessed. Another stapling strategy involving cysteine-based alkylation was also investigated. In this strategy, peptides containing Cys substitutions at Glu13 and Ala17 in the single-chain agonist were stapled using DCA (24 and 25) or DBx (26 and 27). Acetylated and non-acetylated variants were synthesised to investigate the role of the N-terminus. Table 4-2 lists the analogues containing Aib and side-chain staples. All peptides were synthesised using solid phase peptide synthesis and purified by RP-HPLC to obtain peptides with high degree of purity (>95%). HPLC traces and MS data of the synthesised analogues are included in the appendices (Figure S-2 - Figure S-6).

  Table 4-2 Peptide sequence of novel relaxin-3 agonist analogues containing Aib and stapling.

Ligand Sequence Analogue Ac-R3 B10-27 13,17 Pab Ac-SGR(Pa)FIR(Pa)VIFTSGGSRW Ac-R3B10-27 [13/17 HC]

Aib variants Ac-R3 B10-27 13,17 Aib Ac-SGR(Aib)FIR(Aib)VIFTSGGSRW 11 Ac-R3 B10-27 13,14,17 Aib Ac-SGR(Aib)(Aib)IR(Aib)VIFTSGGSRW 12 R3 B10-27 13,17,18 Aib SGR(Aib)FIR(Aib)(Aib)IFTSGGSRW 13 Ac-R3 B10-27 13,17,18 Aib Ac-SGR(Aib)FIR(Aib)(Aib)IFTSGGSRW 14 R3 B10-27 13,17,21 Aib SGR(Aib)FIR(Aib)VIF(Aib)SGGSRW 15 Ac-R3 B10-27 13,17,21 Aib Ac-SGR(Aib)FIR(Aib)VIF(Aib)SGGSRW 16 R3 B10-27 13,17,22 Aib SGR(Aib)FIR(Aib)VIFT(Aib)GGSRW 17 Ac-R3 B10-27 13,17,22 Aib Ac-SGR(Aib)FIR(Aib)VIFT(Aib)GGSRW 18 R3 B10-27 13,17,18,21 Aib SGR(Aib)FIR(Aib)(Aib)IF(Aib)SGGSRW 19 Ac-R3 B10-27 13,17,18,21 Aib Ac-SGR(Aib)FIR(Aib)(Aib)IF(Aib)SGGSRW 20

Hydrocarbon variants Ac-R3 B10-27 13,17 Paa Ac-SGR(Pa)FIR(Pa)VIFTSGGSRW 21 4K-R3 B10-27 13,17 Pab KKKKSGR(Pa)FIR(Pa)VIFTSGGSRW 22 Ac-R3 B10-27 13,17 Pga Ac-SGR(Pg)FIR(Pg)VIFTSGGSRW 23 R3 B10-27 13,17 DCAb SGRCFIRCVIFTSGGSRW 24 Ac-R3 B10-27 13,17 DCAb Ac-SGRCFIRCVIFTSGGSRW 25 R3 B10-27 13,17 DBxb SGRCFIRCVIFTSGGSRW 26 Ac-R3 B10-27 13,17 DBxb Ac-SGRCFIRCVIFTSGGSRW 27 Pa; (S)-2-(4-pentenyl) alanine). Pg; (S)-2-(4-pentenyl) glycine). DCA; dichloroacetone. DBx; α,α’- dibromo- xylene. a; linear analogue. b; stapled analogue. Ac; acetylated N-terminus. All analogues were synthesised with an acidic C-terminus.

  4.2.2 α-Helical structure were restored in grafted relaxin-3 analogues The apamin and VhTI scaffolds were chosen to mimic the desired structure of relaxin-3. Whether the peptide analogues had in fact been able to fold into their expected helical conformation was assessed by solution NMR spectroscopy. The NMR data for analogues 3, 4, 5, 9 and 10 were of good quality with sharp lines and excellent signal dispersion, consistent with a well-ordered structure. In contrast, the NMR data for the VhTI grafted analogues, 6 and 7, were of poor quality with limited dispersion, and in the case of analogue 7 broad lines, consistent with unstructured or misfolded peptides. NOESY data for all analogues are included as appendices, highlighting the difference in spectral quality (Figure S-7 - Figure S-13). Resonance assignment of the NMR data were completed for 3, 4, 5 and 9 using homonuclear sequential assignments strategies. Secondary Hα chemical shifts are excellent indicators of secondary structure, with stretches of negative values being indicative of helical structure while stretches of positive values being consistent with β-sheet (Wishart et al., 1995). A comparison of the secondary Hα shifts observed for the B-chain in native relaxin-3 (1) and analogues 3, 5 and 9, showed that important relaxin-3 residues can be grafted onto apamin and VhTI scaffolds to restore helicity in the secondary structure of single-chain analogues, as evident from the negative secondary shifts (Figure 4-2). Analogue 5 most closely mimic the negative values of relaxin- 3 throughout the entire helical region.

Figure 4-2 Comparison of secondary Hα chemical shift for apamin and VhTI based relaxin-3 analogues and native relaxin-3. The stretch of negative values from 15-24 correspond to the relaxin-3 B-chain helix. Residues numbers on the X-axis relate to the longest peptide sequence, analogue 5, while the sequences are aligned so that the key grafted residues are at the same position in each analogue. Green bar (1) represents the native relaxin- 3 B-chain. Red (3) and stripe bars (9) represent the relaxin-3 agonist and antagonist grafted onto the apamin scaffold respectively. Blue bar (5) represents the first generation analogue of relaxin-3 agonist grafted onto VhTI scaffold.

To further characterize the structures of the grafted analogues, the full 3D structures were calculated based on NMR data for analogues 3, 5 and 9. Interproton distance restraints were derived from cross-peak intensities in the NOESY data, hydrogen bond donors were identified from temperature coefficients and for the apamin variants where 13C chemical shifts could be determined from natural abundance HSQC data, TALOS-N was used to derive backbone dihedral angle restraints. These data were used as input for torsion angle dynamics calculations using CYANA. From a total of 50 structures generated, 20 structures with good CYANA target functions and stereochemistry based on MOLPROBITY (Chen et al., 2010) analyses were chosen to represent the solution structures of the variants. Structural superpositions of these are shown in Figure 4-3. From here it is clear that the disulfide bonded cores are well defined, while the C-terminal tail is disordered, consistent with the secondary shifts. The helical regions observed in the analogues match the helices seen in the native structures of the apamin and VhTI scaffolds. The structural statistics are given in Table S-4.

Figure 4-3 Backbone superposition of the 20 best structures calculated based on NMR spectroscopy data for (A) analogue 3, (B) analogue 9 and (C) analogue 5. Cystine side chains are shown in yellow. Cystines positions in the sequences are indicated by roman numerals.

  4.2.3 Helicogenic amino acid but not thiol-based stapling methodology can improve secondary structure of single-chain relaxin-3 agonists Helical properties of relaxin-3 agonist analogues containing strategic substitution of native relaxin-3 residues for Aib and alternative cysteine-based stapling were also investigated using NMR spectroscopy. Analogues 11, 12, 13, 15, 17, 19, 24 and 26 were studied by NMR spectroscopy and the secondary Hα shifts were determined. The NMR data were of good quality with sharp peaks, but generally had poor signal dispersion. The secondary shifts of these analogues were compared to native relaxin-3 B-chain (1), Ac-R3B10-27 [13/17 HC], and the single-chain relaxin-3 antagonist (8). Overall, the Aib containing relaxin-3 analogues showed a slight increase in helical content compared to 8, although the degree of helicity is significantly less than that observed in the native relaxin-3 B- chain (Figure 4-4 A-B). In both analogues 15 and 19, residue Phe20, which is N-terminally to the introduced Aib at position 21, resulted in a significant upfield shift compared to 8 or 13, which may suggest increased helicity in this region. As illustrated in Figure 4-4C, the secondary Hα shifts of DCA (24) and DBx (26) stapled relaxin-3 agonists closely resemble each other. The secondary Hα shifts were more negative at the stapled sites of Cys13 and Cys17 and the residue immediately before them, but chemical shifts of surrounding residues did not show any helical character, making it more similar to the unstructured R3 B1-22R (8). Overall none of the modifications were able to recapitulate the helical effect of the hydrocarbon staple of Ac-R3B10-27 [13/17 HC], but the C-terminal tail showed closely matching chemical shifts to both this stapled variant and native relaxin-3.

  Figure 4-4 Comparison of secondary Hα chemical shift of Aib containing and stapled relaxin-3 analogues. Secondary Hα chemical shift is presented as the recorded chemical shift subtracted by the random coil chemical shift for each amino acid. An extended region of negative values, indicate an α-helical structure. Analogues were compared to relaxin-3 B-chain of the native peptide (1) or Ac-R3B10-27 [13/17 HC] (HC) and R3 B1-22R (8). Panel A consist of comparison of analogues 11 (Ac-R3 B10-27 13, 17 Aib), 12 (Ac-R3 B10-27 13, 14, 17 Aib) and 17 (R3 B10-27 13, 17, 22 Aib). Panel B consist of comparison of analogues 13 (R3 B10-27 13, 17, 18 Aib), 15 (R3 B10-27 13, 17, 21 Aib) and 19 (R3 B10-27 13, 17, 18, 21 Aib). Panel C consist of comparison of cysteine-based stapled analogues 24 (R3 B10-27 13, 17 DCA) and 26 (R3 B10-27 13, 17 DBx).* indicates thiol stapling positions. Random coil shifts for Pa is not available. Oxidised Cys random coil value was used for calculation of secondary Hα chemical shift at stapled sites of analogues 24 and 26.

  4.2.4 Affinity and potency of some modified peptides for RXFP3 are retained Structural changes can affect ligand/receptor interaction (Hojo et al., 2016), hence it was interesting to investigate whether improved overall structure translated into improvements in binding and activation of RXFP3. The binding affinity of the grafted peptides were determined using CHO cells stably expressing RXFP3 by measuring increasing concentration of the grafted peptides’ ability to compete for the binding site with europium labelled R3 B1-22R (Haugaard-Kedström et al., 2015). Activation was measured by assessing the ability of the analogues to inhibit cAMP accumulation induced by forskolin in a reporter assay, as RXFP3 couples to an inhibitory G-protein. The binding and activation data are presented in Table 4-3 and Table 4-4 and illustrated in Figure 4-5 and Figure 4-6.

The apamin grafted agonist analogue 3, which included Arg12, Ile15, Arg16, Ile19 and Phe20, showed both significantly increased potency (pEC50 = 6.83) and affinity (pKi = 6.65) compared to relaxin-3 B-chain analogue 2 (pEC50 = 5.93 and pKi=5.91). Substitution of Aib at positions Val18 and Thr21 in analogue 4 however, led to a loss in affinity ( pKi = 5.70) and potency (pEC50 = 5.88) for RXFP3, relative to analogue 3. The VhTI based relaxin-3 agonists, 5 and 6, showed a dramatic loss in binding. Due to lack of binding, cAMP activity of analogues 5 and 6 were not determined. Repositioning of the relaxin-3 binding motif closer to the C-terminus in analogue 7 restored affinity of the analogue for RXFP3 to pKi of 6.91 but the peptide was not able to activate RXFP3 (pEC50 < 5). A similar result was obtained for the relaxin-3 antagonists grafted onto apamin and VhTI scaffolds. Analogue 9, showed a slightly weaker binding affinity (< 10-fold) for RXFP3 compared to R3 B1-22R (analogue 8). In comparison, relaxin-3 antagonist grafted onto VhTI scaffold (10) showed almost undetectable affinity for RXFP3 (Table 4-3 and Figure 4-5). Table 4-3 Binding affinities and activities of grafted peptides.

Agonists Analogue Binding affinity cAMP activity a pKi ± SEM [logM] pEC50 ± SEM [logM] R3 acid 1 7.69 ± 0.12 (3) 9.08 ± 0.07 (5) R3 B1-27 2 5.91 ± 0.21 (3)b 5.93 ± 0.02 (3)b,i Apa+R3B 3 6.65 ± 0.18 (3)b,g 6.83 ± 0.07 (3)b,e Apa+R3B [V18Aib,T21Aib] 4 5.70 ± 0.10 (3)b 5.88 ± 0.06 (3)b VhTI+R3B 5 <5 ND VhTI+R3B [G11, R12] 6 <5 ND VhTI+R3B [R12] 7 6.91 ± 0.01 (2)d,f <5 Antagonists R3 B1-22R 8 7.69 ± 0.18 (4)e No activity Apa+R3 B12-22R 9 6.76 ± 0.03 (3)c,f,h No activity VhTI+R3 B1-22R 10 <5 ND a pKi-values are calculated based on the Kd 26 nM for Eu-R3 B1-22R. Numbers in parentheses indicate the times each experiment has been repeated as independent triplicates. bp < 0.001, cp < 0.01 and dp <0.05 compared to R3 acid. ep < 0.001, fp < 0.01 and gp < 0.05 compared to R3 B1-27. hp < 0.05 compared to R3 B1-22R. i, from (Hojo et al., 2016). ND, not determined.

Figure 4-5 Affinity and activity of relaxin-3 grafted agonists and antagonist on RXFP3. (A) Competitive binding curves for grafted peptides using CHO cells stably expressing RXFP3 and the fluorescent tracer Eu-DTPA-R3 B1-22R. (B) Activity of grafted agonists using a reporter assay in CHO cells stably expressing RXFP3. Activation is measured as a reduction in forskolin induced cAMP activity. Data are shown as mean ± SEM of triplicate determinations from a minimum of three independent experiments. 

The hydrocarbon stapled relaxin-3 agonist, Ac-R3B10-27 [13/17 HC], was shown to have high affinity (pKi = 7.38) and potency (pEC50 = 8.48) in a previous study (Hojo et al., 2016). It was of interest to observe if substitution with smaller Aib residue containing additional methyl group at Cα without hydrocarbon stapling is sufficient to improve ligand-receptor interaction. As shown in Table 4-4, none of the Aib agonist variants show comparable affinity or potency for RXFP3 to Ac- R3B10-27 [13/17 HC]. However, analogue 11, containing Aib residues at the same position as the stapled Ac-R3B10-27 [13/17 HC], did show significant improvement in binding (pKi = 6.25) and activity (pEC50 = 7.17) at RXFP3 compared to analogue 2 (pKi = 5.91, pEC50 = 5.93), highlighting that the conformational restriction of the backbone is favourable even in the absence of a side chain staple. Additional Aib residues did not further improve RXFP3 binding or activation. Aib substitution of Val18 was tolerated, however Aib substitutions at positions Phe14 (analogue 12), Thr21 (analogue 16) or Ser22 (analogue 18) were found to be unfavoured. Binding data for 13 and 15 were not acquired as the acetylated analogues were found to have similar activity as the non-acetylated version. Increasing the number of Aib to four residues at Glu13, Ala17, Val18 and Thr21 of relaxin-3 B-chain in 19 and 20 also did not improve affinity of these analogues for RXFP3 in comparison to analogue 2. Variants with Aib at position 22 (analogues 17 and 18) may show signs of partial agonism, as these analogues only show a maximum cAMP accumulation inhibition of 50% at the highest concentration tested (Figure 4-6).

The lack of gain in affinity and activity at RXFP3 motivated the reconsideration of alternative stapling strategies. The linear Ac-R3B10-27 [13/17 HC] (21) showed similar affinity and activity for RXFP3 as 11 with a pKi of 6.03 and pEC50 of 7.38. Thus, the presence of pentenyl-alanine side chains does not further improve activity in a linear variant. The stapling of the hydrocarbon is vital to generate affinity and activity to the level of Ac-R3B10-27 [13/17 HC]. Additional Lys residues at the N-terminus of R3B10-27 [13/17 HC] did not significantly affect binding and activity of the hydrocarbon stapled analogue 22. Absence of the methyl group at Cα in linear agonist containing hydrocarbon Pg in analogue 23 resulted in a substantial loss in affinity for RXFP3, with pKi < 5, consistent with the results from the Aib variants and the linear Pa variant. Alternative stapling utilising Cys at Glu13 and Ala17 of the single-chain relaxin-3 agonist with DCA and DBx staples was not favourable and did not improve affinity of analogues 24 - 27 over 2. 

  Table 4-4 Effect of helical promoting residue and stapling introduction on affinity and potency at RXFP3.

c Peptide Analogue pKi [logM] pEC50 [logM] R3B1-27 2 5.91 ± 0.21d 5.93 ± 0.02d,i R3 B10-27 13,17 Pab Ac-R3B10-27 7.38 ± 0.03 8.48 + 0.06 [13/17 HC]

Aib variants Ac-R3 B10-27 13,17 Aib 11 6.25 ± 0.12e 7.17 + 0.11 Ac-R3 B10-27 13,14,17 Aib 12 5.83 ± 0.18d 6.26 + 0.06d R3 B10-27 13,17,18 Aib 13 ND 7.04 + 0.04f Ac-R3 B10-27 13,17,18 Aib 14 6.12 ± 0.11e 7.48 + 0.20h R3 B10-27 13,17,21 Aib 15 ND 5.99e Ac-R3 B10-27 13,17,21 Aib 16 5.77 ± 0.28d 5.73 + 0.04d R3 B10-27 13,17,22 Aib 17 5.95 ± 0.05d <5 Ac-R3 B10-27 13,17,22 Aib 18 6.02 ± 0.08e <5 R3 B10-27 13,17,18,21 Aib 19 5.70 ± 0.19d 5.75 + 0.27d Ac-R3 B10-27 13,17,18,21 Aib 20 <5 5.49 + 0.09d

Stapling variants Ac-R3 B10-27 13,17 Paa 21 6.03 ± 0.45e 7.38 + 0.16 4K-R3 B10-27 13,17 Pab 22 6.74 ± 0.16 8.19 + 0.14g Ac-R3 B10-27 13,17 Pga 23 <5 6.51 + 0.27d R3 B10-27 13,17 DCAb 24 <5 5.21 + 0.75d Ac-R3 B10-27 13,17 DCAb 25 5.32 ± 0.21d 5.32 + 0.76d R3 B10-27 13,17 DBxb 26 6.00 ± 0.44d 5.98 + 0.07d Ac-R3 B10-27 13,17 DBxb 27 5.98 ± 0.04d 6.07 + 0.03d Pa; (S)-2-(4-pentenyl) alanine). Pg; (S)-2-(4-pentenyl) glycine). DCA; dichloroacetone. DBx; α,α’- dibromo-xylene. a; linear analogue. b; stapled analogue. Ac; acetylated N-terminus. All analogues were c synthesised with an acidic C-terminus. pKi-values are calculated based on the Kd 26 nM for Eu-R3 B1- 22R. dp < 0.001, ep < 0.01 and fp<0.05 compared to Ac-R3B10-27 [13/17 HC]. gp < 0.001 and hp < 0.05 compared to R3 B1-27. i, from (Hojo et al., 2016)

 

Figure 4-6 Binding affinity and potency of relaxin-3 agonist analogues at RXFP3. (A) and (C) illustrate the binding of the Aib and stapled analogues respectively. Activity of Aib containing agonists (B) and stapled analogues (D) are also shown. Data shown as mean ± SEM of triplicate points from a minimum of three independent experiments. HC, Ac-R3B10- 27 [13/17 HC].

4.2.5 Novel analogues showed improved stability in serum To further evaluate the most promising grafted peptides, the half-life in serum was determined for analogues Ac-R3B10-27 [13/17 HC], 3 and 9. Stability of these analogues were referenced to 2 and 8 (Figure 4-7). The serum stability of 3 showed a remarkable increase (T1/2

= 12.8 h) compared to the linear R3 B1-27 (2) (T1/2 = 3.1 min). Similarly, stability of the grafted antagonist 9 (T1/2 = 6.6 h) was significantly improved compared to the linear R3 B1-22R (8)

(T1/2 = 4 min). This equates to a 250-fold and 75-fold improvement in serum stability for the agonist and antagonist, respectively. Hydrocarbon-stapled Ac-R3B10-27 [13/17 HC] was also more stable than the unstructured linear peptides, with a half-life of 58 min.

  Figure 4-7 Stability of relaxin-3 analogues in serum. Serum stability of the single-chain R3 B1-27 (2), Apa+R3B (3), R3 B1-22R (8), Apa + R3 B12-22R (9) and Ac-R3B10-27 [13/17 HC] as measured in % peptide remaining in serum over time is relative to the amount of respective peptide at time point 0 h. Data presented as mean ± SEM from at least three replicates.

4.3 Discussion

Relaxin-3 plays an important role in modulating several neurophysiological functions although its main role has yet to be determined. Since it has a great potential as a drug lead, it is more cost-effective to synthesise the peptide as a single-chain peptide rather than in its native form. The success of the design of single-chain relaxin-3 agonists and antagonists that has high affinity, selectivity and potency for RXFP3 (Haugaard-Kedström et al., 2011;Hojo et al., 2016) were major breakthroughs in drug design and development. However, peptides generally require substantial modification to address susceptibility to serum proteases (Figure 4-7).

Given the recent success with employing molecular grafting as a drug design approach (Li et al., 2009;Phan et al., 2010;Zhang et al., 2015b;Wang and Craik, 2018), we envisioned that single chain RXFP3 agonists and antagonists with improved stability, and possibly also affinity and efficacy, could be developed by transferring key residues from the two-chain relaxin-3 onto a stable single chain scaffold with appropriate structure. The scaffolds investigated here, apamin and VhTI, contains α-helix secondary structure and are stabilised by two disulfide bonds. Important relaxin-3 residues were positioned on the scaffold so that these residues remain on the solvent expose side. It is also evident from the previous study that native relaxin-3 B-chain, which is a weak agonist (Liu et al., 2005a), can be significantly improved using helical staples (Hojo et al., 2016;Jayakody et al., 2016;Marwari et al., 2019). In order to determine whether the enhanced helicity and activity is due to the presence of the hydrocarbon

  stapling or the constraint in the backbone, a series of analogues containing Aib residues incorporated at non-essential positions of the relaxin-3 B-chain, were designed and compared to linear Ac-R3B10-27 [13/17 HC] (21) and a Pg-stapled analogue. However, the Pg analogue could not be stapled, indicating that the lack of backbone constraint from the additional methyl group can lead to a more more flexible peptide which does not fold up into the correct conformation that allow the stapling of the side chain. The versatility of cysteine-based alkylation for ease of stapling with different linker sizes and functionality was also explored.  The grafted analogues adopt helical conformation as indicated by presence of long stretches of negative secondary Hα shifts unlike single chain R3 B1-22R (Haugaard-Kedström et al., 2011) that have shifts closer to zero. Hence, molecular grafting using apamin or VhTI scaffold can clearly restore helicity in a chimeric peptide although the C-terminal part of the helix appears to be less ordered for the apamin grafted analogues than in relaxin-3. Analogues 3, 4, 5 and 9 were all synthesized using random oxidation, but their correct disulfide arrangement could be confirmed by the NMR analysis. Similar oxidation protocols have been used in previous studies and resulted in the native disulfide conformation for apamin (Le- Nguyen et al., 2007;Li et al., 2009). On the other hand, poor quality of the NMR data of analogues 6 and 7 were observed even though these peptides were synthesized using regioselective disulfide bond formation by first forming the Cys7-Cys21 disulfide bond followed by the Cys3-Cys25 disulfide for analogue 6 and vice versa for analogue 7. This approach ensured the fold was not compromised because of the formation of an incorrect disulfide isomer. Given that analogues 6 and 7 are unable to adopt a native fold, the VhTI scaffold is clearly more sensitive to amino acid changes than the apamin structure.  A previous study have shown that addition of hydrocarbon stapling at Glu13 and Ala17 of the single-chain relaxin-3 agonist reintroduce helicity in an otherwise unstructured single- chain agonist (Shabanpoor et al., 2012;Hojo et al., 2016). The agonist variants containing Aib substitutions at the same positions also have helical tendencies as seen in 11, albeit they are not as helical as Ac-R3B10-27 [13/17 HC]. Substitution of Aib at Thr21 in analogues 15 and 19 may extend helicity towards the C-terminus, unlike the analogues with Aib substitution at Val18 (13) or Ser22 (17). This indicates that positioning of Aib is crucial in dictating secondary structure of a peptide. Nonetheless, increasing the number Aib residues did not markedly improve the helical nature of agonists variants, suggesting that double Aib substitution is

  sufficient for a weak helix structure for the relaxin-3 agonist. The comparison to the stapled agonist Ac-R3B10-27 [13/17 HC] indicated that backbone constraints from Aib alone are insufficient and require the staple for significantly increasing the helicity of the agonist. The alternative staples at Glu13 and Ala17 in analogues 24 and 26 were not successful in gaining α−helix structure as shown in the NMR studies that showed the secondary structure of analogues 24 and 26 are still compromised. This may be due to the size of the staple used as staple length can affect the degree of helicity it reinforced. Pa staple contains eight-atom staple, whereas usage of DCA and DBx resulted in seven-atom and nine-atom staples, respectively. Previous studies have demonstrated that a nine-atom staple using DBx conjugated to Cys or DCA stapling with homocysteine conveys an optimal linker length to stabilised a one turn of α-helix at i,i+4 position (Jo et al., 2012;Assem et al., 2015). The current results appear to contradict previous findings. Nevertheless, the lack of secondary structure in 24 and 26 compared to Ac-R3B10-27 [13/17 HC] are more likely to be affected by the Cα methyl group, as an eight-atom lactam bridge stapling did not elicit a shift to helix secondary structure in CD spectroscopy unlike Ac-R3B10-27 [13/17 HC] (Hojo et al., 2016). Thus, to conclusively determine whether an eight- or nine-atom staple is optimal to induce helicity in relaxin-3 agonist, further studies need to be conducted. It is possible that with the addition of Cα methyl group, a nine-atom staple will evoke a similar improvements in helicity seen in Ac-R3B10-27 [13/17 HC].

 Presence of a structured single-chain relaxin-3 agonist is vital as improvement in secondary structure can improve affinity and activity (Hojo et al., 2016;Jayakody et al., 2016;Marwari et al., 2019). The importance of a helical secondary structure is confirmed with the findings of the current study where apamin grafted agonist analogue 3 showed improvement in both affinity and potency compared to the relaxin-3 B-chain analogue 2. Surprisingly, substitution of Val18 and Thr21 with Aib in 4, designed to extend the helix closer to the C-terminus, led to a decrease in affinity and activity at RXFP3 to comparable levels of the native relaxin-3 B-chain (2). The drop in binding affinity is unexpected, given the secondary Hα shifts confirmed the addition of Aib indeed increased the degree of helicity of analogue 4 compared to analogue 3 (Figure S-14). This suggest that either a degree of flexibility is favourable to allow adaptation of the binding conformation, or that the Val and Thr side chains do contribute to binding in the context of this analogue. The apamin grafted antagonist (9) retained binding to RXFP3, but is weaker compared to the linear antagonist R3

  B1-22R. The NMR structural analysis confirmed that the apamin scaffold does confer helical structure, but that the C-terminal part of the peptide after the last cysteine, which includes the key antagonist binding residue Arg23, remains highly flexible. Again this is consistent with detailed structure-activity analysis of R3 B1-22R, which suggest a degree of flexibility is a prerequisite for optimal binding (Haugaard-Kedström et al., 2018;Wong et al., 2018b). The VhTI scaffold however, may not be suitable in the design of potent relaxin-3 analogues. Even though α-helix was successfully regained in the grafted analogue 5, the affinity of 5 for RXFP3 was poor (pKi < 5). The design in the subsequent analogues to ensure the poor binding to RXFP3 is not due to missing relaxin-3 binding residues, also did not lead to improvement in affinity for RXFP3. In analogue 6, NMR structural analysis showed that this version could not adopt the correct fold. Therefore, it can be concluded that indeed the Pro is likely a prerequisite for correct folding of the helix-turn-helix motif in VhTI. Comparing the NMR structures of our grafted analogues with the relaxin-3 B-chain, we noted that in both apamin based variants the binding motif was positioned on top of the helical structure, extending away from the stabilising core, in a very similar fashion to the arrangement in relaxin-3. In contrast, in the folded VhTI variant (analogue 5), the motif is rotated slightly around, which may prevent full access of key grafted residues, including Ile15 (Figure 4-8). Thus, based on the determined structure of analogue 5, a third VhTI grafted agonist variant, analogue 7, was designed. The positions of Arg12, Ile15, Ile19 and Phe20 were moved one residue towards the C-terminus to fit the positioning of the native relaxin-3 relative to the helical surface, which also allowed the structural Pro in native VhTI to be retained. This modification did improve affinity but not potency. Structural analysis using NMR showed the peptide was largely unstructured rather than being able to retain the VhTI fold, which would explain the lack of potency. We have previously shown that the single chain antagonist R3 B1-22R is able to bind efficiently to RXFP3 without being structured in solution (Haugaard-Kedström et al., 2011), and in fact the flexibility is favoured for being able to adapt to bind efficiently to RXFP3 in a slightly different conformation relative to native relaxin-3 (Haugaard-Kedström et al., 2018;Wong et al., 2018b). The VhTI grafted antagonist 10 was based on the first version of the VhTI agonist and retained the native Pro residue from VhTI instead of Arg12. The relative importance of Arg12 for binding to RXFP3 in antagonists is significantly less than in native relaxin-3 (Haugaard- Kedström et al., 2018;Wong et al., 2018b), thus we did not anticipate this being detrimental to affinity. However, analogue 10 showed very poor binding to RXFP3 (Table 4-3, Figure 4-8), which again is likely related to the positioning of the binding residues on the helical surface.

  Given the inability of our next generation VhTI agonist variants (analogues 6 and 7) to fold correctly, we did not pursue this scaffold further in terms of redesign of additional antagonists. Taken together, the apamin framework appears more adaptable and able to accommodate the non-native amino acids while remaining structurally stable. VhTI scaffold on the other hand, is more sensitive to sequence changes and may be too large to accommodate the interaction between the important residues of relaxin-3 and RXFP3 binding site. These results also imply that although molecular grafting in general leads to improved secondary structure and peptide activity, it is not always a direct correlation (Swedberg et al., 2016). The change in affinity for RXFP3 could also be attributed to increased hydrophobicity of the grafted analogues. Analogues 5, 6 and 10 with similar retention times, 28.2 min, 30.8 min and 28.3 min (Figure S-2 and S-3) respectively, were less hydrophobic than analogue 7 with retention time of 37.1 min (Figure S-3). As shown in Table 4-3, only analogue 7 showed affinity for RXFP3. All apamin grafted analogues have retention between 38 min and 39 min (Figure S-2 and S-3), which seems to be sufficiently hydrophobic as well to retain binding to RXFP3. However, hydrophobicity is not a main determinant in affinity for RXFP3. Analogue 4, the relaxin-3 agonist containing two Aibs grafted onto the apamin scaffold although has a retention of 38.4 min, its affinity for RXFP3 is compromised, unlike the other two apamin grafted peptides. 



 Figure 4-8 Comparison of the solution NMR structures of (A) analogue 3, (B) analogue 9, (C) analogue 5 and (D) analogue 1 (B-chain only). The grafted apamin analogues (A and B) and grafted VhTI (C) analogue show overall secondary structure similar to the native H3 relaxin B-chain (D). The grafted RXFP3 binding and activating residues are shown, highlighting differences in the positioning of the receptor binding motif in the various analogues.  

  The changes in affinity and potency of alternative stapling and Aib containing agonists reflects more closely with the degree of helicity obtained. The reinstatement of some helical secondary structure in analogue 11 showed slight improvement in both affinity and activity compared to 2, but is weaker in comparison to Ac-R3B10-27 [13/17 HC], that has a higher degree of helicity. Although analogue 12 also showed a small gain in helicity compared to R3 B1-22R (8), both its affinity and potency at RXFP3 are comparable to 2, which suggests that Phe14 side chain may be involved in RXFP3 interaction. On the other hand, the pair of analogues 13 and 14 that also contain adjacent Aib but at Val18, were shown to gain potency at RXFP3 compared to 2. Increasing the number of Aib substitutions with the addition of Aib at Thr21 however, caused a significant reduction in binding and potency as seen in analogues 19 and 20. Intriguingly, a similar drop in affinity and potency can be observed in analogues 15 and 16 that only contain additional Aib at Thr21. These analogues (13 – 16, 19 -20) have similar propensity towards helical structure, hence the difference in the binding affinity and activity would be more likely due to loss in side-chain functional groups. Therefore, based on comparison between these three pairs of analogues, the drastic reduction in potency at RXFP3 can be narrowed down to Thr21 that may be participating in binding and activating RXFP3. This is surprising as Thr21Aib in the antagonist was well-tolerated, as shown in chapter 3. Thus, this could provide more evidence that the binding mode of relaxin-3 agonist and antagonist are different. Using alternative staples also did not improve structure, binding or activity, and this is most likely due to the lack of Cα methyl group supporting helical structure formation (Figure 4-4C). Taken together, these results strongly indicate the importance of Cα backbone constraints at position Glu13 and Ala17. Having said that, the significant loss of binding of analogues containing less hydrophobic staple, m-xylene and dichloroacetone, could indicate a role played by the linker properties. Studies have shown that cellular properties could correlate better with the degree of hydrophobicity rather than helicity of a peptide (Sakagami et al., 2018;Yuen et al., 2020). As shown in the previous chapter, lactam stapling on the relaxin- 3 antagonist did not improve binding affinity for RXFP3. Similarly, lactam stapling on the single-chain agonist was unfavourable for binding (Shabanpoor et al., 2012) unlike the hydrocarbon staple (Hojo et al., 2016), which is more hydrophobic than a lactam staple (Tian et al., 2017). Analogues 24-27 containing the less hydrophobic dichloroacetone and m-xylene linker (Figure S-6) also showed a highly compromised binding to RXFP3, lending credence to the importance of hydrophobicity in this case. However, binding affinity comparison between Aib and hydrocarbon-containing analogues indicate that hydrophobicity may not be playing a

  major role since 11 and 21 have similar affinity for RXFP3 even though the latter is significantly more hydrophobic than the former (retention time of 11 (42.2 min) is shorter than 21 (50.5 min) in the analytical HPLC-Figure S-4 and S-5). Hence, hydrophobicity may only be playing a minor role in determining affinity of a relaxin-3 analogue for RXFP3, as similarly observed for the grafted analogues.

Unstructured peptides tend to have higher susceptibility to degradation in serum as this allows easy protease access to cleave the peptide. Presence of secondary structure, however, can provide some protection against proteases (Wang et al., 2014). Apamin, which in itself is a stable peptide (Dempsey et al., 2000;Oller-Salvia et al., 2013), have been used as a scaffold to improve stability (Weston et al., 2004), bioactivity (Weston et al., 2004;Zhang et al., 2015b), and as a template to develop a stable BBB penetrable shuttle (Oller-Salvia et al., 2013). Our results as depicted in Figure 4-7 is in agreement with previous findings. Linear peptides were fully degraded within 5 min, unlike the helically-structured grafted analogues that are stable for hours in 100% pooled human male serum. The stability of Ac-R3B10-27 [13/17 HC] which also contain a helix secondary structure was also improved compared to unstructured relaxin- 3 analogues, but, it is at least 6-fold less stable than the grafted analogues. This suggests that degradation can still occur at the more flexible C-terminus of the analogue.

  4.4 Conclusion

In conclusion, the relaxin-3 residues involved in binding and activity at RXFP3 were successfully grafted onto two different scaffolds, apamin and VhTI, to form potential agonist and antagonist analogues of relaxin-3. The relaxin-3 agonist based on the apamin scaffold showed improved structure, binding affinity and potency for RXFP3 compared to the linear relaxin-3 B-chain. Although one variant of the VhTI scaffold was also able to restore the α- helix structure of the native B-chain of relaxin-3, none had the desired pharmacology. Similarly, for the antagonist design, the use of the apamin scaffold improved helical structure, but relative to the single chain antagonist R3 B1-22R, a reduction in binding was observed. Alternative stapling strategies and substitution of helical-promoting residues proved to be inferior than the current hydrocarbon stapling in the design of potent relaxin-3 agonist. Although these agonist variants have the propensity to form helical structure, the binding and potency of these single-chain agonists was not improved compared to Ac-R3B10-27 [13/17 HC]. However, looking at the stability of the analogues, Ac-R3B10-27 [13/17 HC] is less stable in serum compared to apamin grafted analogues. The moderately weaker interacting profile of apamin based relaxin-3 analogues compared to Ac-R3B10-27 [13/17 HC], is compensated by their stability profile. Taking all aspects into account, the apamin chimeric peptides could potentially be therapeutic drug leads but will require further studies to investigate its in vivo activity. 

  





 Chapter 5 Relaxin-3 antagonist analogues to improve BBB penetration

  5.1 Introduction

It is clear that relaxin-3 antagonist analogues plays an important role as modulator of several physiological functions including food intake, stress and learning (Olucha-Bordonau et al., 2018); making relaxin-3 antagonist an attractive therapeutic agent. These biological functions were established using animal model where peptides were delivered directly into the brain via icv or viral vector injections (Kumar et al., 2017). In the previous chapters, the structure-activity relationship of the single-chain relaxin-3 antagonist were investigated and different strategies to re-introduce helicity to the B-chain that would consequently improve stability and/or the interaction with RXFP3 explored. Since RXFP3 is found predominantly in the brain (Matsumoto et al., 2000), it is imperative to also address the aspect of delivering relaxin-3 antagonist analogues into the brain through the BBB when introduced systemically. As the BBB prevents approximately 98% of substances from getting into the brain, delivering relaxin-3 antagonists at a sufficient quantity to have an effect in the brain after crossing the BBB is no easy task. Throughout the years, many strategies have been studied and these methodology have been briefly reviewed in chapter 1. These strategies take advantage of the four available transport mechanisms that allow important nutrients into the brain (Figure 1-9). In this study, two different strategies are being investigated – receptor mediated transcytosis (RMT) and lipidation for passive diffusion.

 Apamin, a bee venom component, is a known neurotoxic agent that is able to block calcium-activated potassium channels in the brain (Habermann, 1984). In the early years, apamin, known for its helical secondary structure (Pease and Wemmer, 1988) and stability (Dempsey et al., 2000;Oller-Salvia et al., 2013), was used as a framework to form chimeric peptides in a chaperon recognition study (Brazil et al., 1997), as a miniature enzyme (Weston et al., 2004), a DNA binding protein (Turner et al., 2004), and as a tumour suppressor activator (Li et al., 2009). Even though apamin used in conjunction with curcumin-loaded micelles has been successful in delivering curcumin through the BBB to treat spinal cord injury in a mouse model (Wu et al., 2014), there are potential issues of immunogenicity and toxicity of apamin (Cosand and Merrifield, 1977;Oller-Salvia et al., 2016a). By removing Arg13 and Arg14 of apamin, residues that are responsible for its toxicity, Oller-Salvia et al. were the first to show that this non-toxic analogue can cross the BBB using an in vitro model (Oller-Salvia et al., 2013). The sequence was minimised to form the highly stable MiniAp-4 BBB shuttle, which was shown to be able to deliver gold NPs and fluorophores across the BBB in in vitro and in

  vivo models, respectively (Oller-Salvia et al., 2016a). However, the exact mechanism of transcytosis for these non-toxic shuttles remain to be elucidated. Similarly, two other peptide shuttles that can cross BBB were discovered recently, but their transcytosis mechanism is still unknown. These peptide sequences were discovered using phage display high throughput screening. Using an in vitro human BBB cellular model, the 12-mer peptide sequence, SGVYKVAYDWQH (SGV), was identified (Diaz-Perlas et al., 2017), while phage screening using iv injection into a rat model resulted in the discovery of RLSSVDSDLSGC (RLS) (Urich et al., 2015). The potential of RLS as a brain penetrating shuttle was further established with its ability to ferry a BACE1 inhibitor into the rat brain to reduce amyloid plague in the cerebrospinal fluid (Urich et al., 2015). Although not much has progressed for these promising shuttles since then, they remain of interest for this study.

 Another selective peptide sequence that crosses the BBB, THRPPMWSPVWP (THR), was identified using phage display libraries incubated with a chicken embryo fibroblast cell line expressing the human TfR in vitro (Lee et al., 2001). Intriguingly, THR was shown to selectively bind to the human transferrin receptor (TfR), and does not bind to chicken TfR (Lee et al., 2001). Since TfR is highly expressed on the luminal BBB side (Jefferies et al., 1984), the receptor has long been seen as an attractive target for brain delivery. Antibodies have been developed for TfR in order to deliver BACE1 for the treatment of Alzheimer’s disease in rodent and nonhuman primate models (Yu et al., 2011;Niewoehner et al., 2014;Yu et al., 2014), a tumour necrosis factor inhibitor for PD treatment (Zhou et al., 2011), and an anticancer drug into the brain (Johnsen et al., 2017). Unlike the larger antibody-based shuttles, the THR peptide is unlikely to cause immunogenicity issues and would be less costly to develop. In addition, THR binds to TfR at a location different from transferrin (Lee et al., 2001), thus do not disrupt the iron transport into the brain (Rouault and Cooperman, 2006). The ability of THR to translocate cargo of different sizes, such as DNA material, NPs and fluorescent dyes across the BBB in both in vitro and in vivo models to treat AD has been demonstrated (Prades et al., 2012;Diaz-Perlas et al., 2018a;Zhang et al., 2018), making it an ideal peptide shuttle. Angiopep2 is another peptide shuttle that has generated high interest among researchers. This BBB shuttle was designed through sequence alignments of the Kunitz protease inhibitor domain of aprotinin, and LRP related ligands containing the same domain (Demeule et al., 2008b). Angiopep2 was shown to be able to cross the BBB in an in vitro BBB model and this is mediated in part by LRP1 (Demeule et al., 2008a), a receptor which is also highly expressed on the BBB (Hussain et al., 1999). Since its discovery, Angiopep2 have been used to deliver

  small molecules for brain cancer treatment (ANG1005, ANG1007 and ANG1009) (Regina et al., 2008;Che et al., 2010), an analgesic neuropeptide (Demeule et al., 2014) and a monoclonal antibody to treat HER2-positive intracranial cancer in animal models (Regina et al., 2015). ANG1005 is currently in phase II clinical trials (NCT02048059) (Tang et al., 2016) after its safety have been substantiated in phase I trials (Kurzrock et al., 2012;Drappatz et al., 2013). Angiopep2 has also been conjugated with NPs and liposomes as a potential treatment pathway for fungal infections and brain glioma (Xin et al., 2011;Shao et al., 2012;Gao et al., 2014). Even though Angiopep2 has been particularly utilised in studies where there is compromised BBB, the ability of the peptide shuttle in penetrating the BBB under normal physiological conditions has also been established using in vitro and in vivo models (Bertrand et al., 2010;Velasco-Aguirre et al., 2017).

Unlike RMT, passive diffusion using lipidation strategies lack selectivity in targeting the BBB. However, there have been many cases where lipidated peptides were successful in delivering therapeutic agents into the brain and other targeted organs (reviewed in (Zhang and Bulaj, 2012;Knudsen and Lau, 2019). Several lipidated neuropeptides, including liraglutide (Hunter and Holscher, 2012), galanin (Bulaj et al., 2008) and prolactin-releasing peptide (PrRP31) (Maletinska et al., 2015) have been successfully shown to cross the BBB to potentially treat neurodegenerative diseases, act as an anticonvulsant and regulate food intake respectively. Although different acyl length have been tested (Zhang et al., 2009b), C18 (palmitate) is the most commonly used acyl group (Bulaj et al., 2008;Zhang et al., 2009b;Maletinska et al., 2015). This strategy can be used in conjunction with cationisation, where the ‘KKKpK’ lysine-palmitoyl motif allows galanin and neuropeptide W analogues to reach the brain and elicit anticonvulsant activity (Bulaj et al., 2008;Green et al., 2011). Presence of cationisation would enable interaction with the negatively charged endothelial cells of the BBB to promote AMT-type translocation into the brain (Herve et al., 2008). In addition, the presence of an acyl group especially palmitate, significantly improve peptide stability (Zorzi et al., 2017). This in turn prolongs peptide circulation in the periphery, bioavailability and consequently increase the probability of the peptides to reach the BBB to transcytose into the brain parenchyma.

In this study, 11 different analogues of the single-chain relaxin-3 antagonist were designed and studied. These analogues contain BBB peptide shuttles conjugated to the N- terminus of the relaxin-3 antagonist, targeting passive transport or RMT. The peptides’ ability

  to bind to RXFP3, their stability and toxicity, and ability to permeate across the BBB in in vitro models were analysed.

5.2 Results 5.1.1 Peptide analogue design rationale In chapter 4 we described analogue 9, in which important relaxin-3 antagonist binding residues (Arg12, Ile15, Arg16 , Ile19 and non-native Arg23) were grafted onto the apamin scaffold. Given apamins ability to pass the BBB, this was the first analogue to be chosen as a candidate for further studies into BBB permeability. In chapter 3, we also showed that the binding affinity of an N-terminally truncated single-chain relaxin-3 antagonist, RB6-22R, is comparable to the full-length single-chain antagonist, R3 B1-22R (Haugaard-Kedström et al., 2018). Therefore, the remaining potential BBB penetrating analogues were designed based on the idea of conjugating R3B6-22R to BBB-penetrable peptide shuttles. Initially a plug-and- play concept, in which the antagonist could be combined with different shuttles via selective post-synthesis conjugation, was envisaged. Using this concept, R3B6-22R was designed to have an additional maleimide functional group with a spacer, known as 6-maleimidohexanoic acid, at the N-terminus since the C-terminal region is highly sensitive to modifications (Haugaard-Kedström et al., 2018). The maleimide allows the formation of a thiol-ether bond when reacted with the thiol group at the side-chain of a Cys in the peptide shuttles, as shown in the conjugation reaction in chapter 2. All variants are presented in Table 5-1. The first shuttle chosen was apamin’s minimised peptide, MiniAp4, to create analogue 28. Analogues 29 and 30 contains the peptide shuttles discovered using phage display. Analogues 31 and 32 contain the peptide shuttles targeting the transferrin receptor and LRP1 receptor, respectively. The shuttles in analogues 28, 30, 31 and 32 do not contain Cys residues, thus for the conjugation, a Cys was added to the peptide shuttles at either the N- or C-terminus.

We also wanted to investigate the effectiveness of lipidation and in particular the ‘KKKpK’ lysine-palmitoyl motif, for improving BBB penetration. Initially, the motif was added to the N-terminal region of R3B6-22R after a spacer, 5-aminovaleric acid (Ava), resulting in analogue 33. The importance of palmitoyl positioning in the analogues was then investigated. Analogue 34 lacks the spacer Ava, whereas in 35 and 36, effect of lipidation at the N-terminus was explored. Unlike 35, 36 does not have a spacer and lysine residues. In 37, the palmitoylated lysine residue was placed at the position of the original Cys10 of relaxin-3

  B chain. All analogues were synthesized as described in chapter 2, and purified to high purity (95%). MS and HPLC traces of the analogues are included in the appendices (Figure S-15 to Figure S-17).  Table 5-1 Relaxin-3 antagonist peptide-shuttle candidates.

Peptide Analogue Sequence Proposed delivery approach

ApaR3 B12-22R 9 H-CNCKAPETARCAIRCVIFTSR Active transport

CMiniAp4-R3B6-22R 28 C[Dap]KAPETALD XGVRLSGREFIRAVIFTSR Active transport

RLS-R3B6-22R 29 RLSSVDSDLSGC-XGVRLSGREFIRAVIFTSR Active transport

CSGV-R3B6-22R 30 CSGVYKVAYDWQH XGVRLSGREFIRAVIFTSR Active transport

CTHR- R3B6-22R 31 CTHRPPMWSPVWP XGVRLSGREFIRAVIFTSR TfR targeting

Angiopep2-R3B6-22R 32 TFFYGGSRGKRNNFKTEEYC-XGVRLSGREFIRAVIFTSR LRP1 targeting KKKpK-Ava-R3B6-22R 33 KKKPK(Ava)GVRLSGREFIRAVIFTSR Passive diffusion KKKpK-R3B6-22R 34 KKKPKGVRLSGREFIRAVIFTSR Passive diffusion Palm-KKKK-Ava-R3B6-22R 35 (Palm)KKKK(Ava)GVRLSGREFIRAVIFTSR Passive diffusion Palm-R3B6-22R 36 (Palm)GVRLSGREFIRAVIFTSR Passive diffusion R3B6-22R S10Kp 37 GVRL(KP)GREFIRAVIFTSR Passive diffusion X, 6-maleimidohexanoic acid; Ava, 5-aminovaleric acid. All peptides were synthesised with an amidated C-terminus.  

  5.2.2 Modified single-chain relaxin-3 antagonist analogues retain binding affinity for RXFP3 The synthesized and purified analogues were then tested for affinity for RXFP3, to investigate whether the modification introduced into the N-terminus of R3B6-22R affected the antagonist’s interaction with RXFP3. The novel analogues were compared to the single-chain antagonist R3 B1-22R (8) using competition binding assays. The results are summarised in Table 5-2 and Figure 5-1.

Amongst the active transport targeting analogues, 31 with a pKi of 7.22 has the highest affinity for RXFP3 and is comparable to R3 B1-22R (8). Analogues 28 and 30 are 6-fold weaker than 8, whereas 32 showed a 9-fold loss in binding to RXFP3 in comparison to 8. The lowest binding affinity for RXFP3 occurs in analogues 9 and 29, with a pKi of 6.36 and 6.41 respectively. For lipidated peptides containing the ‘KKKpK’ motif, the presence of a spacer made a significant difference in affinity for RXFP3. Analogue 33, containing both the motif and spacer, have comparable affinity for RXFP3 as 8. However, removal of spacer in 34 resulted in a drop in pKi to 5.29, which is at least a 180-fold loss of affinity for RXFP3 when compared to 8. Shifting the palmitoyl group from the Lys side-chain to the N-terminus of the motif in analogue 35, did not further improve binding to RXFP3 but the analogue remains comparable to the reference peptide. Removal of the Lys residues and spacer however, resulted in a moderate loss of affinity, as seen in 36 with a pKi of 6.56. Replacing the native Cys10 in the relaxin-3 B-chain with a palmitoylated Lys residue in 37 caused a similar drop in affinity (pKi = 6.41). Overall, although most of the novel analogues have some loss of binding to RXFP3, the effect was moderate and most analogues have comparable affinity. Therefore, it can conclude that the modifications introduced in this series of analogues are well-tolerated and can thus be used for further studies.

  Table 5-2 Binding affinity of potential BBB permeable relaxin-3 antagonist for RXFP3.

Peptide Analogue pKi ± SEM, [logM]g R3 B1-22R 8 7.69 ± 0.18 ApaR3 B12-22R 9 6.36 ± 0.03a CMiniAp4-R3B6-22R 28 6.92 ± 0.22c RLS-R3B6-22R 29 6.41 ± 0.23a CSGV-R3B6-22R 30 6.90 ± 0.07c CTHR-R3B6-22R 31 7.22 ± 0.31 Angiopep2-R3B6-22R 32 6.73 ± 0.13b KKKpK-Ava-R3B6-22R 33 7.38 ± 0.11 KKKpK-R3B6-22R 34 5.29 ± 0.16a,d Palm-KKKK-Ava-R3B6-22R 35 7.34 ± 0.12 Palm-R3B6-22R 36 6.56 ± 0.11b,f R3B6-22R S10Kp 37 6.41 ± 0.27a,e ap< 0.001, bp<0.01 and cp<0.05 compared to R3 B1-22R. dp< 0.001, ep<0.01 and fp<0.05 g compared to KKKpK-Ava-R3B6-22R (for comparison to analogues 34-37). pKi-values are calculated based on the Kd 26 nM for Eu-DTPA-R3 B1-22R. 

  Figure 5-1 Binding affinity of relaxin-3 antagonist analogues conjugated to BBB shuttles. (A) Data comparison between analogues targeting active transport mechanisms and (B) passive diffusion mechanisms. Data presented as mean ± SEM from at least three independent experiments.

5.2.3 Peptide analogues showed improvement in serum stability Next, it was of interest to find out whether these analogues would be stable enough in circulation to have an opportunity to pass through the BBB in vivo. The stability of these novel analogues were therefore tested in 100% human male pooled serum in vitro. Novel analogues were added to serum and incubated at 37°C. At different timepoints, an aliquot was removed and added into ammonium acetate (pH 3) to inactivate the protease activity. Serum half-life of novel analogues tested are summarised in Table 5-3 and illustrated in Figure 5-2. As shown in the previous chapter, the linear single-chain relaxin-3 antagonist (8) is fully degraded within ~5 min. The plug-and-play analogues, 30 and 32, containing the linear B-chain antagonist, also showed low stability in serum, as expected. The grafted analogue based on the stable apamin scaffold (9) has, as shown in chapter 4, an increased half-life of >6 h.

  Lipidation also strikingly improves stability, despite the peptides themselves remaining linear and unstructured. The lipidated peptides 34 – 36, are highly stable in serum with half- lives of each of these analogues exceeding 24 h. Analogue 33 is also significantly more stable than the parent single-chain relaxin-3 antagonist, with a half-life of 2.8 h, but is considerably less stable compared to the other lipidated analogues, suggesting the position of the lipidation is important. During the analysis, a major breakdown product was identified as resulting from cleavage at Phe20. In order to further improve the stability of analogues in serum, Phe20 of the R3B6-22R was therefore substituted with a non-native Aib amino acid in three new variants. This improved the half-life of analogue 38 9-fold compared to 8. Introducing the same modification in lipidated variants lead to further significant gain in stability as evident from 40, where Phe20Aib substitution relative to 33 resulted in a drastic improvement of half-life from ~3 h to more than 24 h. The same amino acid substitution in linear active transport variants such as 39 however, did not result in a significant change in the stability of the analogue when compared to the non-Aib analogue 32. The stability analysis of analogue 37 was hampered by poor solubility in serum leading to unreliable quantification and poor data. Nonetheless, most of the peptide was remaining at the end time point, thus it is concluded that 37 have a half-life of >6 h. Overall, lipidation can be seen as an effective strategy to increase half-life of peptide analogues in serum. Utilisation of non-natural amino acid Aib is also beneficial to improve peptide stability when introduced at the correct position of the single-chain relaxin-3 antagonist.  Table 5-3 Half-life of novel BBB analogues.

Peptide Analogue Serum half-life R3 B1-22R 8 4.8 min ApaR3 B12-22R 9 6.6 h CSGV-R3B6-22R 30 12.6 min Angiopep2-R3B6-22R 32 5.0 min KKKpK-Ava-R3B6-22R 33 2.8 h KKKpK-R3B6-22R 34 > 24 h Palm-KKKK-Ava-R3B6-22R 35 > 24 h Palm-R3B6-22R 36 > 24 h R3B6-22R S10Kp 37 > 6 h R3B6-22R F20Aib 38 42.9 min Angiopep2-R3B6-22R F20Aib 39 12.5 min KKKpK-Ava-R3B6-22R F20Aib 40 > 24 h

 

Figure 5-2 Stability of novel analogues in serum. Analogues were incubated at 37°C in 100% pooled male human serum. Each data point has at least three replicates and data are presented as mean ± SEM.

5.2.4 Cytotoxicity can be observed in some novel analogues It is also vital to assess any toxicity of the novel analogues to cells to evaluate what concentrations can be introduced systemically to achieve sufficient quantity for BBB targeting. To assess this, a MTT cytotoxicity assay with differentiated neuroblastoma cells, SH-SY5Y, was used. Prior to cell/peptide incubation, serum-containing media was removed and fresh serum-free media with peptide was added to each well. Viability of cells were determined using the absorbance of the formazan crystals formed in viable cells.

Comparison between novel analogues targeting active mechanism (A) and passive diffusion (B) clearly showed that the presence of the ‘KKKpK’ motif are more toxic to neuroblastoma cells (Figure 5-3). Analogue 9 did not show sign of cytotoxicity when incubated to 80 μM. On the contrary, analogues 28 and 30 only showed signs of cytotoxicity in SH-SY5Y when the concentration of these analogues were more than 50 μM, similar to the parent peptide, R3B6-22R. Cell viability was slightly decreased to 70% after 30 μM of 29 was incubated with cells. On the other hand, analogue 31 was well-tolerated up to 10 μM, before a 25% cell death was observed at approximately 20 μM. The cell viability drastically decreased at 39 μM of analogue 31, where more than 80% cell death occurred. Analogue 32 fare worse as the cell

  viability dropped by 50% when peptide concentration was increased from 7 μM to 14 μM. 90% cell death was observed at 28 μM.

The ‘KKKpK’ motif containing analogues are more toxic with higher cytotoxicity occurring at a lower peptide concentration. Cells incubated with almost 100 μM of 36 did not result in complete cell death, but appears to plateau close to 50% cell viability at 21.9 μM. In 37, 5.6 μM peptide was well-tolerated, but 50% cell death occurred when peptide concentration was doubled, and further increase in peptide concentration led to complete loss of cell viability. Analogues 33 – 35 were more similar in their cytotoxicity profile and are the most toxic analogues. Neuronal cells well-tolerated analogues 33 – 35 at low micromolar range, where 20% loss in cell viability occurred at approximately 1.6 μM and 2 μM of 33 and 35 respectively. Analogue 34 caused a 30% loss in cell viability at 4.8 μM. All three analogues caused a significant cell death in the subsequent increase in peptide concentrations. Presence of Phe20Aib in analogue 40 did not significantly alter its toxicity compared to analogue 33. To investigate the role of the positive charges in cell toxicity, a new analogue 41 was designed. In this variant, the lysines of KKKpK were replaced with negatively charged glutamic acids. This change in analogue 41 (EEKpE-Ava-R3B6-22R) successfully eliminated the cytotoxicity observed in 33. Since 41 does not show toxicity, we wanted to confirm that the modification didn’t affect binding to RXFP3. As shown in Figure 5-4, analogue 41 with a pKi of 7.95 has comparable, if not slightly better affinity than analogues 8 and 33 for RXFP3. Further studies on the stability was also done. Like analogue 37, the stability analysis of 41 was hampered due to solubility issues, but peptide quantitation showed most of 41 remained after 24 h. Overall, most of the active transport targeting analogues are non-toxic to neuroblastoma cells whereas the toxic analogues contain the ‘KKKpK’ motif. The toxicity however, can be reversed by removing the positive charges and substituting with negatively charged residues. 

 

Figure 5-3 Toxicity of BBB analogues in SH-SY5Y cells. Differentiated neuroblastoma cells were incubated with novel analogues that targets A) active transport mechanism and B) passive diffusion, for 2.5 h before addition of MTT (0.5 mg/ml). Toxicity of novel analogues were compared to parent relaxin-3 antagonist, R3B6-22R. Fluorescence readings were taken at 570 nm. Data are presented as mean ± SEM of three replicates. Each analogue were tested at least three times independently with the exception of 29 with a n=2.

 

Figure 5-4 Comparison of binding affinity of analogue 41 (EEKpE modified variant) to analogues 8 and 33. Data are presented as mean ± SEM from at least three independent experiments. Analogue 41 contain a pKi of 7.95 ± 0.22.

5.2.5 BBB penetration differ between active and passive mechanism targeting analogues Finally, the permeability of these novel analogues across the BBB was investigated, as one of the main goals of this study was to introduce relaxin-3 antagonist into systemic circulation and still can pharmacologically target RXFP3 in the CNS. Peptide permeability via passive diffusion was assessed using precoated parallel artificial membrane permeability assays (PAMPA). As per manufacturer’s protocol, novel analogues were added to the donor compartment at the bottom with the acceptor (filter plate) at the top. After 5 h incubation in PBS, samples from both donor and acceptor wells were removed and frozen at -20°C until analysis using LC-MS. Using standard curves prepared for each peptide analogue, the peptide concentration in the donor and acceptor wells were determined and the diffusion across the lipid membrane was calculated. Analogues 8, 9, 28 – 37, and 41 were tested in PAMPA, but only a few of those analogues could be detected in the acceptor compartment of the assay. The detected analogues in the acceptor compartment were analysed and the results of the permeability, Pe, are shown in Table 5-4. Figure 5-5 illustrates the concentration of analogues incubated at the start of the assay and the resulting concentration in each compartment after 5 h incubation. High permeability analogues are considered to surpass the diffusion threshold of

  1.5 x 10-6 cm/s (Chen et al., 2008). As shown in Table 5-4, only the positive control, methoxyverapamil, with a Pe value of 2.55 x10-5 cm/s, is a high permeable compound through passive diffusion. A comparison between analogues 33 – 35 showed that 33 has the lowest permeability of 0.81 x10-8 cm/s. The permeability increased by 20-fold in 34, with a Pe value of 16.18 x10-8 cm/s. 35 has the highest Pe (32.72 x10-8 cm/s) amongst the three analogues that could be detected. Analogues 33 - 35 all carry variants on the KKKpK motif, which is clearly beneficial for passive diffusion. Analogue 41 with the reverse charged motif could not diffuse through the membrane while analogue 36, which carries the lipid but no charges, was retained in the membrane, which only small amounts detected in the donor compartment.  Table 5-4 Permeability of novel analogues in PAMPA.

Peptides Analogues Pe(.10-8) cm/s Atenolol 1.54 ± 0.25 Methoxyverapamil 2.55x103 ± 2.56x102 KKKpK-Ava-R3B6-22R 33 0.81 ± 0.52 KKKpK-R3B6-22R 34 16.18 ± 4.31 Palm-KKKK-Ava-R3B6-22R 35 32.72 ± 6.39

Figure 5-5 Concentration of novel analogues in different compartments of the PAMPA assay. The initial peptide concentration is shown in black bars. After 5 h incubation, analogue concentration in donor and acceptor compartments are shown in white and grey bars, respectively. MV; methoxyverapamil and At; atenolol.

  The passive permeability of the novel analogues was also compared to the cellular uptake of analogues into the immortalised human brain endothelial cells, hCMEC/D3. In these assays, cells were incubated with 166 μg/ml peptide in serum-free media before cells were washed and lysed at 5 min, 30 min and 90 min respectively. Quantitative analyses were also carried out using LC-MS. The percentage of peptide uptake into the cells was calculated by comparing the peptide concentration in the cell lysates to the initial concentration. Six representative analogues were tested; 9, 28 – 31 and 6. Of the six, half could not be detected at all in the cell lysate using LC-MS. The percentage of peptide analogues found in the hCMEC/D3 cells is shown in Table 5-5. Analogue 28 showed relatively similar uptake at 5 min, 30 min and 90 min incubation without significant increase in peptide uptake in longer incubation period. Analogue 30 on the other hand, showed a slight increase in cell uptake over time, where the analogues taken up into the cells increased from 0.25% to almost 0.5% in 90 min. For 36, the change in cell uptake increased drastically from 0.04% at 5 min to 0.24% at 90 min

Table 5-5 Cellular uptake of novel analogues in hCMEC/D3 over time.

Peptides tested Analogue % cellular uptake 5 min 30 min 90 min CMiniAp4-R3B6-22R 28 0.046 ± 0.002 0.047 ± 0.002 0.054 ± 0.001 CSGV-R3B6-22R 30 0.252 ± 0.025 0.339 ± 0.064 0.478 ± 0.055 Palm-R3B6-22R 36 0.041 ± 0.011 0.055 ± 0.005 0.238 ± 0.014 ApaR3 B12-22R 9 ND ND ND RLS- R3B6-22R 29 ND ND ND CTHR- R3B6-22R 31 ND ND ND ND; not detected in LC-MS

Taken together, none of the designed analogues showed good permeability in either assay, but the trend of adding lipidation and positive charges did positively influence passive diffusion and modifications targeting active transport mechanisms did show a small effect in the cell uptake study.

  5.3 Discussion

Several relaxin-3 antagonist analogues are available as pharmacological tools to study the neuromodulatory effects of the relaxin-3/RXFP3 interaction, however, these studies require invasive icv injections of peptides (Haugaard-Kedström et al., 2011;Smith et al., 2014a). In order to be considered a potential pharmaceutical drug lead, these analogues require further modification to be able to cross the BBB to engage RXFP3 when introduced via systemic administration. In this study, 11 relaxin-3 antagonist analogues containing potential BBB peptide shuttles were designed to observe if the introduced modifications were effective in increasing permeability of relaxin-3 antagonist across the BBB into the brain.

The analogues designed were chosen to compare BBB penetrating capabilities between passive transport and active transport. Since apamin is a known BBB penetrating bee venom component, it has been utilised in a previous study to delivery curcumin for spinal cord injury in an animal model (Wu et al., 2014). Although it is a neurotoxin, analogue 9, which is a grafted “chimera” between apamin and the relaxin-3 antagonist, lacks residues Arg13 and Gln17 of native apamin, which are known to trigger toxicity (Labbe-Jullie et al., 1991). The smaller analogue MiniAp4 also lacks toxicity. For conjugation of this to R3B6-22R, a Cys was added to the N-terminus, as a previous study has demonstrated its ability to shuttle a variety of cargo (fluorescent labels and gold NPs) with an additional Cys at N-terminus (Oller-Salvia et al., 2016a). Similarly, fluorescent labelling has been performed at the N-terminus of SGV (Diaz- Perlas et al., 2017) and THR (Prades et al., 2015). Since modification at the N-terminus did not impair microscopy analysis in those studies, a Cys was positioned here in the current investigation. In contrast, Cys was added to C-terminus of Angiopep2, based on the previous work on neurotensin delivery in rats (Demeule et al., 2014). For RLS, the existing Cys in the sequence was utilised. Thiol-maleimide chemistry were used to conjugate peptide shuttle and R3B6-22R as it has also been used successfully for peptide conjugation and peptide delivery into the brain (Demeule et al., 2014). Considering the single-chain relaxin-3 antagonist, R3 B1- 22R have a short half-life of ∼ 5 min in serum (Figure 5-2), it was of interest to study the effects of lipidation, not only on improving the BBB penetrating capability of relaxin-3 antagonist, but also to improve the stability of the relaxin-3 antagonist. Positioning of the lipid can also affect affinity of the analogues for the receptor (Bulaj et al., 2008). Thus, a series of peptides were designed to address this issue. In 36, palmitoylated lysine residue were placed in the

  original Cys10 of relaxin-3 B-chain as palmitoylation usually occurs at a Cys residue of proteins (Aicart-Ramos et al., 2011).  The new variants were first assayed to investigate affinity for RXFP3. It is essential that the modifications introduced into relaxin-3 antagonist did not interfere with the ligand/receptor interaction. As shown in Table 5-2 and Figure 5-1A, analogues targeting active transport mechanisms, 28 – 32, generally retained good affinity for RXFP3 with 19-fold as the largest loss in affinity for RXFP3 observed in analogue 29 compared to 8. This is as expected as peptide shuttle conjugation in these analogues were carried out at the N-terminus of R3B6- 22R, which is not close to the vicinity of the RXFP3 binding residues of the relaxin-3 antagonist. Apamin grafted analogue 9 as discussed in chapter 4, consists of RXFP3 binding residues positioned in the same topological positions as the native relaxin-3 B-chain α-helix. The flexibility of the C-terminus of the grafted antagonist (shown in NMR structural studies) allowed efficient interaction with RXFP3. Lipidated analogues also retain binding to RXFP3, but the result strongly indicated that spacer and the position of palmitoyl group can influence affinity of the analogue for RXFP3. As illustrated in Figure 5-1B, drastic change in affinity was observed when 5-Ava spacer was removed in analogue 34. Removing the lysineresidues and placing the palmitoyl group at the N-terminus of relaxin-3 antagonist (36) resulted in a moderate loss in affinity for RXFP3 compared to analogues 33, 35 and 41 that contain spacer. Placing the palmitoyl group at Ser10 of relaxin-3 B-chain (37) also showed a similar drop in binding to RXFP3 as 36, which could be a result of steric hindrance, where palmitoyl group in close proximity to Arg12 might prevent efficient interaction of Arg12 with RXFP3. It is evident that both spacer and lipid positions can impact on receptor binding similar to the observations made in the development of systemically stable liraglutide, semaglutide (Madsen et al., 2007;Lau et al., 2015) and galanin (Bulaj et al., 2008;Robertson et al., 2012). It should be noted however, that the competition binding assays were conducted in the presence of 1% BSA, as lack of BSA led to too high background reading from non-specific binding in the assays. BSA is likely to bind to the palmitoyl group of the lipidated peptides, leading to a lower free peptide concentration available to bind to the receptor (Lau et al., 2015). Hence, there is a strong likelihood that these lipidated analogues actually have higher affinity for RXFP3 than reported; without serum albumin interference. Considering the different modifications designed resulted in only moderate loss with the exception of 34, all of these analogues were tested for their stability in serum as it is part of the main goal to administer the peptides systemically.

  Unstructured peptides are known to be more susceptible to protease degradation. Since analogue 28 - 32 like 8, contain the linear B-chain antagonist, it was predicted that the stability of these analogues in serum will be low. The short half-lives of two of those plug-and-play analogues tested, 30 and 32, indicated that regardless of presence of peptide shuttles that may be stable (Demeule et al., 2014;Urich et al., 2015), having the linear relaxin-3 antagonist unprotected is a liability for stability in serum. Thiosuccinimide formed from the thiol- maleimide chemistry is known to hydrolyse in presence of gluthathione, but it is not considered an issue since it was shown that at the gluthathione blood concentration, the spacer is stable (Diaz-Perlas et al., 2018a). As discussed in chapter 4, grafting the vital relaxin-3 binding residues onto a stable disulfide-rich scaffold (9) can significantly improve stability in serum due to increased structural constraints. On the other hand, presence of lipidation such as palmitoylation, allows peptide analogue to bind to serum albumin (Curry, 2009), which would reduce the amount of peptides exposed to proteolytic activity. As expected, even though 33 – 37 also contain the linear R3B6-22R, these palmitoylated analogues showed a significant improvement in serum stability than the non-palmitoylated analogues. The same strategy has been used successfully in the development of liraglutide and semaglutide (Knudsen et al., 2000;Lau et al., 2015). Positioning of palmitoyl group also influenced stability of the lipidated analogues but the effect is opposite to analogue affinity for RXFP3. Without 5-Ava spacer as seen in 34, 36 and 37, these analogues are more stable than 33, although 35 and 41 does not conform to the observed pattern. The Glu residues in analogue 41 may also able to interact with serum albumin (Zorzi et al., 2017), which may further contribute to its stability in serum. It was noted that analogue 33, our first-generation lysine-palmitoyl analogue, has a 35-fold improvement in serum stability compared to 8, but, this can be further improved. Substitution of Phe20, a major cleavage site in 33, with Aib in 38 and 40 resulted in a significant improvement in stability (Figure 5-2). This is supported by previous study that showed Aib substitution could prevent peptide recognition by serum proteases (Werner et al., 2016). The substitution position is also ideal as it was shown that Phe20 is non-critical in the relaxin-3 antagonist interaction with RXFP3 (Haugaard-Kedström et al., 2018). Nevertheless, as evident from the breakdown of analogue 39 containing the Angiopep2 peptide shuttle, other cleavage sites are present and may also have to be modified for further improvement. Intriguingly, there are contradictory reports in regards to Angiopep2 stability in plasma. Angiopep2 conjugated to neurotensin (Ang2002) only resulted in cleavage in rat plasma after a 7 h incubation (Demeule et al., 2014), however, in another study, the unconjugated shuttle showed a 30-min half-life in rat plasma (Wei et al., 2014) and was fully degraded within 10 min in 1% pancreatin

  (Fuster et al., 2019). It is likely that depending on the peptide conjugated and location, different cleavage sites may be exposed. THR in analogue 31 is also known to have a short serum half- life and have been replaced with its retro-inverso version in recent years as a protease-resistant shuttle for cargo delivery into the CNS (Prades et al., 2015), and it has been suggested that the more complex branched THR shuttle is more efficient in crossing the BBB (Diaz-Perlas et al., 2018a). As a preliminary study, 31 and 32 were be evaluated for its ability to cross the BBB (in a serum-free environment) and could be further modified using retro-inverso shuttles if the permeability looked promising.

Apart from assessing the stability of these novel analogues, cytotoxicity characterisation is essential as they provide an insight into the therapeutic window available for these novel analogues. It should be noted that the cytotoxicity assays were carried out without presence of serum during peptide incubation. Therefore, even the low stability analogues should not degrade within the duration of the incubation. The apamin-based analogues, 9 and 28, were well-tolerated in the neuronal cells. The lack of toxicity in these analogues is consistent with the toxicity-causing apamin residues, Arg13 and Gln17, being omitted in these analogues and Arg14 was also removed in 28 (Labbe-Jullie et al., 1991;Oller- Salvia et al., 2016a). Reduction of toxicity has also been shown to reduce immunogenicity of MiniAp4 by three orders of magnitude compared to apamin (Oller-Salvia et al., 2016a). Although not much is known about the toxicity of 29, it has been used effectively in vivo without detrimental effects (Urich et al., 2015). SGV in 30 was tested for cytotoxicity during its discovery where it was noted the cell viability remained above 90% at 100μM (Diaz-Perlas et al., 2017). Since R3B6-22R is also non-cytotoxic to the neuronal cells at high concentration of 50μM, the lack of toxicity observed in 29 and 30 was expected. There have been no reported cases of Angiopep2 or THR being toxic to cells to date, hence, the level of cytotoxicity observed in 31 and 32, were not anticipated. Previous studies using the same peptide shuttle in conjunction with NPs, an antibody or a small molecule did not result in unwanted cell death in in vitro or in vivo models (Kurzrock et al., 2012;Prades et al., 2012;Youn et al., 2014;Regina et al., 2015). However, this may be a result of the overall physicochemical characteristics of the peptide conjugates. R3B6-22R is highly basic with a net charge of +4 at pH 7 (Table 5-6).

  Table 5-6 Peptide analogue net charge at pH 7.

Peptide Analogue + charge at pH 7 R3B6-22R 4 ApaR3 B12-22R 9 4 CMiniAp4-R3B6-22R 28 4 RLS-R3B6-22R 29 3 CSGV-R3B6-22R 30 4.1 CTHR-R3B6-22R 31 5.1 Angiopep2-R3B6-22R 32 6 KKKpK-Ava-R3B6-22R 33 7 KKKpK-R3B6-22R 34 7 Palm-KKKK-Ava-R3B6-22R 35 7 Palm-R3B6-22R 36 3 R3B6-22R S10Kp 37 4 KKKpK-Ava-R3B6-22R F20Aib 40 7 EEKpE-Ava-R3B6-22R 41 1

Analogue 31 with a positive charge of five, caused a near complete cell death at 40 μM. Moreover, 32 containing Angiopep2 and a higher positive charge of six, showed a similar cell death profile but occurring at a lower concentration of 30 μM. Similar pattern can be observed in the lysine-palmitoyl analogues that were designed. Analogues 36 and 37 which contains less positive charges at physiological pH compared to 33 – 35, exhibit better cell viability profile (Figure 5-3). However, 36 and 37 despite containing less or the same charge number as R3B6- 22R respectively, still caused more cell death than R3B6-22R. Hence, it is very likely that the presence of palmitoylation in these analogues is sufficient to contribute to toxicity when a certain number of positive charges are present. In order to prove that cationisation is the major factor in the observed cytotoxicity, an analogue was designed to substitute Lys residues in 33 with Glu, forming 41. The MTT assays clearly demonstrated that cytotoxicity observed with 33 can be reversed when the positive charges were reduced from seven to one. Based on the cytotoxicity seen in amphipathic analogues containing positive charges and hydrophobicity (via palmitoylation), these analogues may behave in a similar fashion to the antimicrobial peptides’ (AMP) mode of action. The mammalian cell membranes are zwitterionic in nature and AMPs are known to favour disruption of bacterial cell membranes that are highly anionic (Silhavy et al., 2010;van Meer and de Kroon, 2011). Nevertheless, these novel analogues are still similar to AMPs such as the amphipathic mellitin, which does not discern between bacterial or mammalian cell membranes (Paterson et al., 2017). Even though several of the novel analogues have a degree of cytotoxicity, the results from the competition binding assays would not be compromised. As the highest concentration used is around 10 μM, the slightly

  toxic 31 and 32 would not have caused cell death. Moreover, the competition assays were conducted in presence of serum. As the lipidated analogues can bind to serum albumin, the effective peptide concentration in the assays would very likely be less than the peptide concentration threshold that caused cytotoxicity.

Even though some of the novel analogues exhibited cytotoxicity, it was of interest to investigate whether the modifications that were introduced into the relaxin-3 antagonist improved the efficiency of peptide permeability across biological barriers. PAMPA is routinely used to predict potential drug lead permeability via passive diffusion. Therefore, the lack of detection for the active transport targeting analogues was as predicted. Previous studies, at least for SGV and MiniAp4, showed that the shuttle efficiency for permeability across an in vitro BBB model is temperature dependent and affected by sodium azide, a clear sign of active transport mechanisms (Oller-Salvia et al., 2016a;Diaz-Perlas et al., 2017). Apamin is also almost undetected in the acceptor compartments of a PAMPA-BBB (Oller-Salvia et al., 2013). Thus, the relaxin-3 apamin grafted would not be expected to be different. Analogues 31 and 32, containing THR and Angiopep2 shuttle, targets TfR and LRP1 respectively. Thus, lack of receptors in the PAMPA model would have resulted in no translocation across the artificial lipids also for these peptides.

Amongst the lipidated analogues designed, analogue 41 was the most promising as it does not cause cytotoxicity, have high affinity for RXFP3 and is stable in serum. However, amongst the lipidated analogues designed to cross the BBB using passive transport, 41, was the one most retained in the donor compartment. Other lipidated analogues were detected in the acceptor compartment but still have permeability well below the threshold of 1.5 x 10-6 cm/s to be considered a good drug lead that crosses the BBB via passive diffusion (Chen et al., 2008). This is unexpected, as previous studies were quite successful in utilising palmitoylation alone to achieve significant BBB penetration when introduced peripherally (Knudsen et al., 2000;Maletinska et al., 2015). The lack of analogue 36 in the acceptor component could be explained by its hydrophobicity. From Figure 5-5, it can be observed that after 5 h incubation, only a small amount of peptide was left in the donor compartment but none was detected in the acceptor. This was most likely due to non-specific retention of the peptide analogue in the artificial lipid layer (Chen et al., 2008). There is however, an overall pattern that suggests that an increase in number of positive charges lead to an increase in permeability across the artificial lipids. Presence of positively charged residues can promote additional membrane interaction

  via adsorptive-mediated transcytosis (AMT) with anionic phospholipids in the PAMPA. Although, Green and colleagues noted that having too many Lys residues can be detrimental and thus opted for combinatorial strategy of lipidation-cationisation instead (Bulaj et al., 2008), this can as shown here lead to increased toxicity. A delicate balance is required as the BBB is highly anionic in nature (Ribeiro et al., 2012), which is likely the reason 41 containing negatively charged Glu residues cannot cross the BBB due to repelling charges.

Taking all aspects of the novel analogue characterisation that has been carried out thus far, six of the initial 11 designed analogues were chosen to be tested for their ability to be taken up into the immortalised human cerebral microvascular cells, hCMEC/D3. As peptides were incubated with the cells at a high concentration of 166 μg/ml, only 36, that is the least cytotoxic among the lipidated analogues, was chosen to be tested. Analogue 32 containing Angiopep2 was not tested since the hCMEC/D3 cell line express LRP1, the target of Angiopep2, at very low levels (Urich et al., 2012). Overall, the permeability results in the cell assay were disappointing. The below limit of detection of relaxin-3 antagonist-apamin chimeric peptide (9) was unexpected as apamin is a well-known peptide that enters the brain (Habermann and Cheng-Raude, 1975). The exact BBB penetration mechanism of apamin is yet to be elucidated, but it has been suggested that interactions with calcium-activated potassium channel may be involved during the BBB penetration (Wu et al., 2014). Therefore, grafting of non-natural residues onto apamin could potentially obstruct interactions between apamin and the ion channel. Although 28 could be detected in the cells, the uptake remained the same over 90 min, which suggests that longer incubation is required to see a more significant uptake. Lack of detection of 29 in the hCMEC/D3 cells could be a result of incorrect conjugation position. This is because it was shown previously that the BACE1 inhibitor was ferried into the CNS with the aid of RLS using biotinylation on the RLS N-terminus (Urich et al., 2015). As the current study utilised the existing C-terminus Cys residue in the RLS for conjugation to R3B6-22R, this modification may have prevented interactions of the RLS shuttle with the BBB. LC-MS peak detection of 31 indicated a peak broadening that does not correspond to the intended MRM, and thus could not be quantified. Since some of these analogues target unknown active mechanisms, it is unclear if these shuttles would be affected by efflux transporters. However, in this immortalised brain endothelial cell line, the most active efflux transport, p-gp, is expressed at a very low level (Urich et al., 2012), and it is likely these peptides are too large for p-gp anyway. Therefore, it is unlikely the lack of uptake into the cell over 90 min is due to

  efflux. There is possibility however, that the unknown receptors involved may be expressed at a low level in the immortalised hCMEC/D3 cells. Thus, even though 30 showed increased cellular uptake over 90 min, the low concentration observed could be a result of the level of expression of the responsible receptor. Rapid uptake of 36 from 30 min to 90 min was intriguing as in the PAMPA model, the analogue was mainly retained in the membrane. Therefore, the uptake seen in the cells could be a result of 36 being membrane bound and released when cells are lysed. Comparing it to a similar assay previously conducted by our collaborators using sensitive radiolabelled peptides applied at lower concentration, 33 have an increased uptake at 30 min and is significantly better than the parent antagonist, R3 B1-22R (Table 5-7). Thus, it is likely that both lipidation and cationisation is required for a better uptake into cells, which is consistent with the results obtained from the PAMPA.

Table 5-7 Radiolabeled peptide uptake into hCMEC/D3 normalised to the total cell protein.

Analogues Analogues 5 min* 30 min* Docosahexaenoic acid (high 2.447 ± 0.689 3.151 ± 0.237 brain uptake; positive control) R3 B1-22R 8 0.280 ± 0.083 0.257 ± 0.073 KKKpK-Ava-R3B6-22R 33 0.733 ± 0.072 0.622 ± 0.174 *Data are reported as cell uptake (ml/mg i.e. DPM/mg of total protein relative to initial spiking solution DPM/ml). Data are presented as mean ± SD.

Overall the peptide permeability was poor, but this may in part be related to technical challenges. Samples from both PAMPA and cellular uptake studies were not desalted prior to sample concentration and LC-MS analysis as it was expected that the LC methodology would be sufficient to alleviate potential ion suppression due to salts. It would likely be more accurate to repeat the analysis of the current analogues also using radiolabelling given that the detection sensitivity, and indeed the degree of uptake, was more significant using radiolabelling than LC-MS.

  5.4 Conclusion

Physiological effects stemming from the relaxin-3/RXFP3 interaction in the CNS has been studied extensively using animal models. However, these studies were conducted using the invasive icv injections. Hence, there is a real need to address the lack of pharmacological tools that would allow observation of behavioural changes when relaxin-3 antagonist is introduced systemically. Here, 10 new analogues were designed to include modifications targeting active and passive transport, at the N-terminus of relaxin-3 antagonist and one analogue was grafted onto apamin scaffold. The design of these potential BBB analogues were successful in generally retaining affinity for RXFP3. In terms of the stability and toxicity however, the results are more varied. A lysine-palmitoyl strategy using the ‘KKKpK’ motif improved the analogues stability, however, the increased in overall cationisation resulted in increased cytotoxicity in the neuronal cells. The neurotoxic effects nevertheless, can be reversed with substitution of Lys for Glu. The plug-and-play BBB analogues targeting active mechanism on the other hand, were less stable due to the presence of cleavable site on the single-chain relaxin-3 antagonist but were relatively less toxic compared to the lysine- palmitoyl peptide analogues. Using in vitro BBB models, lysine-palmitoyl analogues improved its uptake across the artificial lipids compared to only palmitoylated analogue, but the permeability is still poor. Uptake of analogues targeting active mechanisms into hCMEC/D3 cells were less than 1%, as was the case for the palmitoylated analogue that was tested. Although the active targeting analogues did not show promising uptake into the immortalised human BBB cells, it is very likely this is partly due to the lack of sensitivity on the LC-MS methodology. Therefore, we could not conclusively state that these designed analogues do not cross the BBB. Further studies will be required to improve quantitation of peptide analogues’ BBB permeability to conclusively determine the permeability of the designed analogues.

 

Chapter 6 Conclusion and future directions

  Since its discovery in 2002, relaxin-3 has been shown to play a vital role in modulating several neurophysiological processes. The most notable studies were monitoring food intake using rodent models. Relaxin-3 agonist treatment can induce increased food intake in satiated rats (Kuei et al., 2007) whereas treatment with the antagonist R3 B1-22R resulted in a loss of feeding drive in mice (Smith et al., 2014a). There are other neurological functions that have been indicated to be modulated by relaxin-3 signalling, including stress, addiction and anxiety (Olucha-Bordonau et al., 2018). As relaxin-3 is important in the various behavioural responses, the development of high affinity selective antagonist and agonists is critical.

Extensive SAR studies on relaxin-3 have been undertaken previously, and important residues for binding and activating RXFP3 determined (Kuei et al., 2007;Liu et al., 2009a;Shabanpoor et al., 2012;Bathgate et al., 2013b;Hu et al., 2016). SAR for the single-chain antagonist, R3 B1-22R, on the other hand, have not been systematically evaluated, but instead generally assumed to be similar to relaxin-3. As shown in chapter 3, alanine scanning provided important information in regards to the residues that engages RXFP3 and binding mode of the single-chain antagonist (Haugaard-Kedström et al., 2018;Wong et al., 2018a). The R3 B1-22R can be further reduced in size by removing five residues from the N-terminus without compromising affinity for RXFP3 but no changes can be made to the non-native Arg at the C- terminus to retain high binding affinity for RXFP3. Although agonist and antagonist share some binding residues; mainly Arg12, Ile5, Arg16, and Ile19, there are additional residues on the antagonist that form receptor interactions, as the removal of individual side chain functional groups, which are not involved in the binding of relaxin-3, results in impaired binding of R3 B1-22R. The R3 B1-22R is thought to form a helical conformation upon binding to RXFP3, hence, different strategies have been used to try and reintroduce helicity into the unstructured single-chain antagonist. Single substitution of helical promoting residue Aib at positions 17, 18 or 21 of the relaxin-3 B-chain are well tolerated but presence of multiple Aib in the same peptide led to a reduction in affinity for RXFP3. Cyclisation was also not successful for the single-chain antagonist, leading to a loss rather than gain of binding to the receptor. Point modifications using Glu-Lys or Lys-Asp pairs at positions that are non-essential for the relaxin- 3 binding motif resulted in a significant drop in R3 B1-22R binding to RXFP3, but binding can be rescued to a certain extent with the formation of a side-chain lactam staple. The recovery was however not to a comparable level to R3 B1-22R, which further proves that the relaxin-3 antagonist/RXFP3 interaction is distinct from the agonist.

  These distinctions were complemented with receptor mutagenesis study conducted by our collaborator (Wong et al., 2018a). Mutagenesis conducted on acidic residues of RXFP3 showed that Arg12 and Arg16 of the R3 B1-22R antagonist, just like the native relaxin-3, interacts with Glu244 and Asp145 respectively. Although from the peptide SAR, the positive charge of Arg16 of R3 B1-22R was deemed less essential since affinity for RXFP3 can be rescued with mutation to Abu. Receptor mutagenesis and modelling suggested that it is more likely Arg16 interacts with Asp145 (Wong et al., 2018a). Moreover, based on the receptor/R3 B1-22R homology model, the smaller R3 B1-22R could insert deeper into the RXFP3 binding cavity. Therefore, Arg23 of the antagonist can form ionic and cation-π interactions with both Glu141 and Trp138 as evidenced by the loss of R3 B1-22R binding to RXFP3 with Glu141A/Trp138A double mutation (Wong et al., 2018a). The receptor mutagenesis studies were also vital as it showed the potential secondary site of RXFP3 that are interacting with the opposite helix of R3 B1-22R, which face away from the binding motif of Arg12, Ile15, Arg16 and Ile19. His268A of transmembrane 5 (TM5), Thr346A and Ser349A of TM6 were shown to have affected R3 B1-22R binding with a 2-3-fold loss in binding (Wong et al., 2018a). Phe14A and Val18A of relaxin-3 antagonist caused an 8-fold loss of affinity to RXFP3 but was rescued with Abu substitution at those positions. In the agonist, both of these residues are prevented by the A-chain from contributing to binding. Using the updated R3B1-22R/RXFP3 interaction model, both Phe14 and Val18 could be well positioned to form interactions with the residues on TM5 and TM6. These results strongly indicated the importance of receptor mutagenesis studies in providing a more comprehensive view of how the peptide interacts with the receptor and will assist in development of future relaxin-3 antagonist analogues.

As Aib, lactam staples and cyclisation could not improve the pharmacological affinity of the single-chain relaxin-3 antagonist analogues, it was postulated that grafting may have a better prospect. Important relaxin-3 binding residues were grafted onto two disulfide-rich scaffolds, apamin and VhTI, and yielded agonist and antagonist analogues that showed improved secondary structure. These analogues showed good helical structure in the same region as the B-chain of native relaxin-3. However, only the apamin chimeric peptides exhibited acceptable affinity and activity for RXFP3. In the VhTI scaffold the topological positions of the relaxin-3 B-chain could not be replicated efficiently, which dropped the affinity for RXFP3 significantly. Redesign to adjust the positions did not allow the chimeric peptide to fold correctly, thus VhTI is more sensitive to changes. As a previous study have shown that

  the affinity and potency of single-chain relaxin-3 agonist can be improved by addition of a hydrocarbon staple to form Ac-R3B10-27 [13/17 HC] (Hojo et al., 2016), further work was conducted to investigate the potential of alternative stapling strategies for an agonist. None of the variants successfully restored the structure or activity to the level of the previous analogue. The NMR analysis have shown that analogues lacking Cα substitution have compromised secondary structure with secondary Hα shifts that are closer to 0 ppm (Figure 4-4). Thus, lack of backbone constraint on the Cα using halogen staples was postulated to be the main reason behind the weak affinity of these novel agonists for RXFP3 compared to Ac-R3B10-27 [13/17 HC]. Using the simpler helix-inducing strategy, Aib introduced at position Glu13 and Ala17 showed a moderate improvement in activity at RXFP3, consistent with this observation. However, incorporation of three to four Aib residues in the single-chain agonist did not further improve structure or activity. Comparison between the Aib-containing analogues strongly suggests that replacement of Thr21 is unfavourable and that a loss of potential side-chain interactions between Thr21 and RXFP3 affects binding and activation. Although Ac-R3B10- 27 [13/17 HC] exhibits higher binding affinity and potency for RXFP3 compared to the apamin chimeric analogues, the better stability profile of the chimeric analogues still makes them attractive leads. Further modifications may be able to further restore activity to the level of the stapled variant.

With decent progress made in SAR studies and identification of structured relaxin-3 analogues containing good affinity and activity, it was of high interest to investigate whether additional modifications could improve the BBB penetration of relaxin-3 antagonist variants. The apamin grafted antagonist was one of the analogues considered for BBB permeability since apamin is known to be able to cross the BBB. Five other analogues containing peptide shuttles targeting active transport across the BBB conjugated to the minimised relaxin-3 antagonist (R3B6-22R) and five other antagonist analogues with lipidation modifications for passive transport were also designed and synthesised for this study. The modifications were well tolerated in terms of affinity to the receptor, although active transport targeting analogues have poorer stability compared to most of the lipidated analogues, among which some have stability exceeding 24 h. Removal of a protease recognition site, Phe20, by Aib incorporation, was able to further improve the stability of the unstructured R3B6-22R and the first generation lipidated analogue, KKKpK-Ava-R3B6-22R, and the use of LC-MS for identifying breakdown products followed by targeted modifications is likely to lead to further improvement of most variants.

  Cytotoxicity also varied significantly between analogues targeting active and passive transport mechanism. The former analogues are generally well-tolerated by the neuronal cells at concentrations up to 50 μM, with Angiopep2 and THR peptide shuttles being tolerated up to 10 μM before starting to show signs of cytotoxicity. The lipidated and cationised analogues are more toxic to the neuronal cells in comparison, as low micromolar concentration can only be incubated with the cells without causing cell death. In the studies of BBB permeability, lack of permeability of the active targeting analogues across PAMPA was as predicted, but the lipidated analogues did show some permeability. The fact that overall the passive diffusion was limited is likely a result of the peptides despite lipidation being rather large and overall polar molecules. For this type of molecules an active transport mechanism may have more promise, yet disappointingly the active mechanism targeting analogues did not show good uptake into the immortalised hCMEC/D3 cells, despite the peptide shuttles chosen for this project having been shown to be effective molecular ‘Trojan horses’ in previous studies.

The results obtained in this thesis have made major inroads into our understanding of the differences between the pharmacology of agonists and antagonists of RXFP3, and while perhaps not delivering on the promise of generating highly effective systemically active compounds, have identified several limitations and considerations that need to be taken into account for future design. Firstly, It is clear that for agonist variants, the key to good pharmacology is the ability to replicate the native structure of relaxin-3, while the best antagonist to date R3 B1-22R is unstructured and for optimal binding of an antagonist, some degree of flexibility appears favourable. Thus different design principles will apply depending on whether it is an agonist or antagonist that is being optimised. Secondly, when utilising shuttles, the conjugation sites for the BBB penetrating analogues must be considered. The ones used here may not be ideal for some of the analogues, as the actual interacting residues on the peptide shuttles are generally not known. Usage of existing or addition of non-native Cys for conjugation could potentially interfere with and alter the shuttle - BBB interaction. From a technical point of view, the conjugation chemistry may also cause some issue in cellular uptake peptide quantitation. Thiol-maleimide conjugation reaction was used as they are practical with fast conjugation kinetic and have been shown to be a successful bioconjugate in the delivery of therapeutic drug leads across the BBB (Oller-Salvia et al., 2016a). However, the thiosuccinimide from the conjugation of maleimide and thiol group, can under circumstances hydrolyse in the presence of GSH and at basic pH depending on whether electron-withdrawing

  groups are in close proximity to the maleimide (Lyon et al., 2014). The N-alkyl maleimide use in this study was shown to be stable at high GSH concentration of 10 mM (Diaz-Perlas et al., 2018a), but the conjugate may be susceptible to pH changes. It has been reported that at basic pH, the N-alkyl maleimide removal from antibody conjugates does occur, although the rate of removal is slow, taking up to 3 d (Lyon et al., 2014). Due to the accessibility of our analogues compared to the larger antibody-drug conjugates, it is plausible the removal rate may be higher in our analogues. As the cellular uptake study required cell lysis under basic conditions, this issue could cause decreased detection of intact analogue in LC-MS. Thus, based on the current design of BBB analogues, the conjugation location and in extension, the spacer type and length, should be studied further as these factors can affect stability and BBB permeability of these novel analogues. Thirdly, lipidation is a clearly a powerful way to increase serum stability, even for unstructured analogues, thus is particularly useful for antagonists. This does however come at the risk of increased toxicity and it will be important to consider interplay with overall charge and positioning of charged residues when incorporating lipids. Technically, lipidation is challenging for the competition binding assays, the assay could not be conducted without presence of serum due to high background reading with serum removed. Since the lipidated analogues can interact with serum albumin, this could potentially lead to inaccurate pKi values and may not reflect the true binding affinity of the novel analogues for RXFP3. A different assay setup may be required for more accurate determination of pKi. Fourthly, hCMEC/D3 cells used in in vitro cellular uptake is a well-known model that mimics most of the human BBB (Weksler et al., 2013). However, being an immortalised cell line, some bias can occur due to different expression of some transporters and receptor, such as p-gp, LRP1 and GLUT- 1 compared to the primary human brain endothelial cells (Urich et al., 2012). To gain a better insight into the effectiveness of the active transport targeting analogues transcytosis, rather than uptake into monolayer cells, transcytosis across primary BBB endothelial cells would be preferred. For this, a co-culture of primary brain endothelial cells with astrocytes would be ideal, although assays use primary cells that are costly and more difficult to maintain. In addition, the lack of a positive control more closely related to a peptide in the cellular uptake assays using immortalized hCMEC/D3 cell is likely to have reduced the robustness of the assay. Finally, quantitation of peptide analogues permeability in in vitro assays was more technically challenging than expected. Without desalting the samples prior to LC-MS analysis, ionisation of the fragment used in MRM could have occurred and led to a loss in signal intensity since the altered fragmentation would not be detected by the MS. Hence, samples from PAMPA and in vitro cellular uptake may require processing using suitable C18 cartridges to

  prevent salt ionisation of the peptide of interest. If sample cleanup still does not translate to good signal intensity in the LC-MS, radiolabelling detection may need to be considered. This has the added benefit of requiring substantially less peptide for detection thus can eliminate any concerns about cell toxicity.

Several additional experiments and designs can be envisaged to strengthen our findings. Further investigation on binding and activity at the related receptors RXFP1 and RXFP4 should be conducted to investigate the selectivity of these analogues. Our data hint at contributions from Val18 or Thr21 in the B-chain agonist to receptor interactions. Based on the comparisons between Aib containing agonists, it seems likely that at least Thr21 is involved in interacting with RXFP3. Thus, relaxin-3 agonists grafted onto the apamin scaffold containing single Aib substitution at Val18 and Thr21 should also be synthesised, to investigate the difference in activity of these two variants. Although the novel analogues designed for BBB penetration retained binding to RXFP3, these analogues also showed some neuronal cytotoxicity. The reason for this is likely membrane disruption given the overall positive charge and hydrophobicity. In order to elucidate the mechanism behind the cell death, further experiments could be conducted using lactate dehydrogenase (LDH) assay which detects cytoplasmic enzymes in the cell culture supernatant due to membrane leakage. This would confirm the observed toxicity is due to membrane disruption. Given that separate targeting of active and passive transport mechanism seems inefficient, and that the stability benefits of cationisation and lipidation were generally offset by toxicity, a complete redesign of the BBB permeable analogues may be required. A dual-targeting strategy, incorporating a shuttle for active transport and a lipid but no cationisation, could provide a long-term solution. With dual- targeting, it is plausible that simultaneous increase in stability, bioavailability and selectivity for BBB endothelial cells can be achieved. In order to improve on the cellular uptake experiments using the immortalized human cell line, a more appropriate control will be included to test permeability of the new generation of BBB analogues. This would allow a more appropriate comparison and to observe any significant improvement in permeability of the new generation of BBB analogues. If permeability of the new generation of relaxin-3 antagonist BBB analogues in the in vitro BBB models are promising, these analogues should be investigated for their ability to reduce food intake in rodent models. Moreover, once an optimal conjugation to both lipidation and active targeting shuttles has been identified, more assays can be conducted to study its pharmacokinetics and biodistribution. The amenability of the R3B6-22R N-terminus would enable labelling for imaging analysis of biodistribution in

  rodent models. Real-time tracking of these labelled analogues can be conducted using PET- MRI to study the details of relaxin-3 antagonist biodistribution in the brain and whether the peptide reach the RXFP3-rich regions. Overall, the insights gained in this investigation is a significant step forward in terms of designing new generation of relaxin-3 antagonist analogues that can be delivered systemically and still have an effect in the brain. This investigation has also highlighted the need for an ideal peptide-based drug discovery as depicted in Figure 6-1.

Figure 6-1 Proposed workflow diagram of peptide-based drug discovery targeting the BBB. Design of peptide analogues will be based on a combination of peptide/target identification and delivery system based on a literature search. Potential BBB permeable analogues will be characterised before selection of lead analogues to be tested for their permeability across the BBB using in vitro assays. Further narrowing down of lead compounds with BBB permeability will progress into in vivo studies. Analogues with ideal in vivo preclinical results can be considered for clinical trials. Analogues that fail to exhibit acceptable receptor interaction, stability, is toxic or lack BBB permeability will be redesigned and reoptimized with a more suitable delivery system.

 













 References

  Aarts, M., Liu, Y., Liu, L., Besshoh, S., Arundine, M., Gurd, J.W., Wang, Y.T., Salter, M.W., and Tymianski, M. (2002). Treatment of ischemic brain damage by perturbing NMDA receptor- PSD-95 protein interactions. Science 298, 846-850.

Adham, I.M., Burkhardt, E., Benahmed, M., and Engel, W. (1993). Cloning of a cDNA for a novel insulin-like peptide of the testicular Leydig cells. J Biol Chem 268, 26668-26672.

Agrawal, M., Saraf, S., Saraf, S., Antimisiaris, S.G., Chougule, M.B., Shoyele, S.A., and Alexander, A. (2018). Nose-to-brain drug delivery: An update on clinical challenges and progress towards approval of anti-Alzheimer drugs. Journal of Controlled Release 281, 139-177.

Aicart-Ramos, C., Valero, R.A., and Rodriguez-Crespo, I. (2011). Protein palmitoylation and subcellular trafficking. Biochim Biophys Acta 1808, 2981-2994.

Assem, N., Ferreira, D.J., Wolan, D.W., and Dawson, P.E. (2015). Acetone-Linked Peptides: A Convergent Approach for Peptide Macrocyclization and Labeling. Angew Chem Int Ed Engl 54, 8665-8668.

Banerjee, A., Shen, P.J., Ma, S., Bathgate, R.A., and Gundlach, A.L. (2010). Swim stress excitation of nucleus incertus and rapid induction of relaxin-3 expression via CRF1 activation. Neuropharmacology 58, 145-155.

Barnabas, W. (2019). Drug targeting strategies into the brain for treating neurological diseases. J Neurosci Methods 311, 133-146.

Bathgate, R.A., Halls, M.L., Van Der Westhuizen, E.T., Callander, G.E., Kocan, M., and Summers, R.J. (2013a). Relaxin family peptides and their receptors. Physiol Rev 93, 405-480.

Bathgate, R.A., Lin, F., Hanson, N.F., Otvos, L., Jr., Guidolin, A., Giannakis, C., Bastiras, S., Layfield, S.L., Ferraro, T., Ma, S., Zhao, C., Gundlach, A.L., Samuel, C.S., Tregear, G.W., and Wade, J.D. (2006). Relaxin-3: improved synthesis strategy and demonstration of its high-affinity interaction with the LGR7 both in vitro and in vivo. Biochemistry 45, 1043-1053.

Bathgate, R.A., Oh, M.H., Ling, W.J., Kaas, Q., Hossain, M.A., Gooley, P.R., and Rosengren, K.J. (2013b). Elucidation of relaxin-3 binding interactions in the extracellular loops of RXFP3. Front Endocrinol (Lausanne) 4, 1-10.

Bathgate, R.A., Samuel, C.S., Burazin, T.C., Layfield, S., Claasz, A.A., Reytomas, I.G., Dawson, N.F., Zhao, C., Bond, C., Summers, R.J., Parry, L.J., Wade, J.D., and Tregear,

  G.W. (2002). Human relaxin gene 3 (H3) and the equivalent mouse relaxin (M3) gene. Novel members of the relaxin peptide family. J Biol Chem 277, 1148-1157.

Bell, R.D., and Ehlers, M.D. (2014). Breaching the blood-brain barrier for drug delivery. Neuron 81, 1-3.

Bellmann-Sickert, K., and Beck-Sickinger, A.G. (2010). Peptide drugs to target G protein- coupled receptors. Trends Pharmacol Sci 31, 434-441.

Bertrand, Y., Currie, J.C., Demeule, M., Regina, A., Che, C., Abulrob, A., Fatehi, D., Sartelet, H., Gabathuler, R., Castaigne, J.P., Stanimirovic, D., and Beliveau, R. (2010). Transport characteristics of a novel peptide platform for CNS therapeutics. J Cell Mol Med 14, 2827-2839.

Bertrand, Y., Currie, J.C., Poirier, J., Demeule, M., Abulrob, A., Fatehi, D., Stanimirovic, D., Sartelet, H., Castaigne, J.P., and Beliveau, R. (2011). Influence of glioma tumour microenvironment on the transport of ANG1005 via low-density lipoprotein receptor- related protein 1. Br J Cancer 105, 1697-1707.

Bittencourt, J.C., and Sawchenko, P.E. (2000). Do centrally administered neuropeptides access cognate receptors?: an analysis in the central corticotropin-releasing factor system. J Neurosci 20, 1142-1156.

Blazquez, E., Velazquez, E., Hurtado-Carneiro, V., and Ruiz-Albusac, J.M. (2014). Insulin in the brain: its pathophysiological implications for States related with central insulin resistance, type 2 diabetes and Alzheimer's disease. Front Endocrinol (Lausanne) 5, 161.

Boado, R.J., Hui, E.K., Lu, J.Z., and Pardridge, W.M. (2010). Drug targeting of erythropoietin across the primate blood-brain barrier with an IgG molecular Trojan horse. J Pharmacol Exp Ther 333, 961-969.

Boado, R.J., Zhang, Y., Zhang, Y., and Pardridge, W.M. (2007a). Genetic engineering, expression, and activity of a fusion protein of a human neurotrophin and a molecular Trojan horse for delivery across the human blood-brain barrier. Biotechnol Bioeng 97, 1376-1386.

Boado, R.J., Zhang, Y., Zhang, Y., and Pardridge, W.M. (2007b). Humanization of anti-human insulin receptor antibody for drug targeting across the human blood-brain barrier. Biotechnol Bioeng 96, 381-391.

Borodina, T., Marchenko, I., Trushina, D., Volkova, Y., Shirinian, V., Zavarzin, I., Kondrakhin, E., Kovalev, G., Kovalchuk, M., and Bukreeva, T. (2018). A novel

  formulation of zolpidem for direct nose-to-brain delivery: synthesis, encapsulation and intranasal administration to mice. J Pharm Pharmacol 70, 1164-1173.

Brazil, B.T., Cleland, J.L., Mcdowell, R.S., Skelton, N.J., Paris, K., and Horowitz, P.M. (1997). Model peptide studies demonstrate that amphipathic secondary structures can be recognized by the chaperonin GroEL (cpn60). J Biol Chem 272, 5105-5111.

Bulaj, G., Green, B.R., Lee, H.K., Robertson, C.R., White, K., Zhang, L., Sochanska, M., Flynn, S.P., Scholl, E.A., Pruess, T.H., Smith, M.D., and White, H.S. (2008). Design, synthesis, and characterization of high-affinity, systemically-active galanin analogues with potent anticonvulsant activities. J Med Chem 51, 8038-8047.

Bullesbach, E.E., and Schwabe, C. (2000). The relaxin receptor-binding site geometry suggests a novel gripping mode of interaction. J Biol Chem 275, 35276-35280.

Burazin, T.C., Bathgate, R.A., Macris, M., Layfield, S., Gundlach, A.L., and Tregear, G.W. (2002). Restricted, but abundant, expression of the novel rat gene-3 (R3) relaxin in the dorsal tegmental region of brain. J Neurochem 82, 1553-1557.

Burbach, J.P. (2011). What are neuropeptides? Methods Mol Biol 789, 1-36.

Calvez, J., De Avila, C., Matte, L.O., Guevremont, G., Gundlach, A.L., and Timofeeva, E. (2016). Role of relaxin-3/RXFP3 system in stress-induced binge-like eating in female rats. Neuropharmacology 102, 207-215.

Calvez, J., De Avila, C., and Timofeeva, E. (2017). Sex-specific effects of relaxin-3 on food intake and body weight gain. Br J Pharmacol 174, 1049-1060.

Calvez, J., Lenglos, C., De Avila, C., Guevremont, G., and Timofeeva, E. (2015). Differential effects of central administration of relaxin-3 on food intake and hypothalamic neuropeptides in male and female rats. Brain Behav 14, 550-563.

Cao, G., Pei, W., Ge, H., Liang, Q., Luo, Y., Sharp, F.R., Lu, A., Ran, R., Graham, S.H., and Chen, J. (2002). In Vivo Delivery of a Bcl-xL Fusion Protein Containing the TAT Protein Transduction Domain Protects against Ischemic Brain Injury and Neuronal Apoptosis. J Neurosci 22, 5423-5431.

Chassin, D., Laurent, A., Janneau, J.L., Berger, R., and Bellet, D. (1995). Cloning of a new member of the insulin gene superfamily (INSL4) expressed in human . Genomics 29, 465-470.

  Che, C., Yang, G., Thiot, C., Lacoste, M.C., Currie, J.C., Demeule, M., Regina, A., Beliveau, R., and Castaigne, J.P. (2010). New Angiopep-modified doxorubicin (ANG1007) and etoposide (ANG1009) chemotherapeutics with increased brain penetration. J Med Chem 53, 2814-2824.

Chen, H., Chen, C.C., Acosta, C., Wu, S.Y., Sun, T., and Konofagou, E.E. (2014). A new brain drug delivery strategy: focused ultrasound-enhanced intranasal drug delivery. PLoS One 9, e108880.

Chen, V.B., Arendall, W.B., 3rd, Headd, J.J., Keedy, D.A., Immormino, R.M., Kapral, G.J., Murray, L.W., Richardson, J.S., and Richardson, D.C. (2010). MolProbity: all-atom structure validation for macromolecular crystallography. Acta Crystallogr D Biol Crystallogr 66, 12-21.

Chen, X., Murawski, A., Patel, K., Crespi, C.L., and Balimane, P.V. (2008). A novel design of artificial membrane for improving the PAMPA model. Pharm Res 25, 1511-1520.

Cierpicki, T., and Otlewski, J. (2001). Amide proton temperature coefficients as hydrogen bond indicators in proteins. Journal of Biomolecular Nmr 21, 249-261.

Clark, R.J., Fischer, H., Dempster, L., Daly, N.L., Rosengren, K.J., Nevin, S.T., Meunier, F.A., Adams, D.J., and Craik, D.J. (2005). Engineering stable peptide toxins by means of backbone cyclization: stabilization of the alpha-conotoxin MII. Proc Natl Acad Sci U S A 102, 13767-13772.

Conklin, D., Lofton-Day, C.E., Haldeman, B.A., Ching, A., Whitmore, T.E., Lok, S., and Jaspers, S. (1999). Identification of INSL5, a new member of the insulin superfamily. Genomics 60, 50-56.

Conners, R., Konarev, A.V., Forsyth, J., Lovegrove, A., Marsh, J., Joseph-Horne, T., Shewry, P., and Brady, R.L. (2007). An unusual helix-turn-helix protease inhibitory motif in a novel trypsin inhibitor from seeds of Veronica (Veronica hederifolia L.). J Biol Chem 282, 27760-27768.

Cosand, W.L., and Merrifield, R.B. (1977). Concept of internal structural controls for evaluation of inactive synthetic peptide analogs: synthesis of [Orn13,14]apamin and its guanidination to an apamin derivative with full neurotoxic activity. Proc Natl Acad Sci U S A 74, 2771-2775.

Curry, S. (2009). Lessons from the crystallographic analysis of small molecule binding to human serum albumin. Drug Metab Pharmacokinet 24, 342-357.

  De Araujo, A.D., Hoang, H.N., Kok, W.M., Diness, F., Gupta, P., Hill, T.A., Driver, R.W., Price, D.A., Liras, S., and Fairlie, D.P. (2014). Comparative alpha-helicity of cyclic pentapeptides in water. Angew Chem Int Ed Engl 53, 6965-6969.

De Avila, C., Chometton, S., Lenglos, C., Calvez, J., Gundlach, A.L., and Timofeeva, E. (2018). Differential effects of relaxin-3 and a selective relaxin-3 receptor agonist on food and water intake and hypothalamic neuronal activity in rats. Behav Brain Res 336, 135-144.

De Boer, A.G., and Gaillard, P.J. (2007). Drug targeting to the brain. Annu Rev Pharmacol Toxicol 47, 323-355.

Deeken, J.F., and Loscher, W. (2007). The blood-brain barrier and cancer: transporters, treatment, and Trojan horses. Clin Cancer Res 13, 1663-1674.

Dehouck, B., Dehouck, M.P., Fruchart, J.C., and Cecchelli, R. (1994). Upregulation of the low density lipoprotein receptor at the blood-brain barrier: intercommunications between brain capillary endothelial cells and astrocytes. J Cell Biol 126, 465-473.

Demeule, M., Beaudet, N., Regina, A., Besserer-Offroy, E., Murza, A., Tetreault, P., Belleville, K., Che, C., Larocque, A., Thiot, C., Beliveau, R., Longpre, J.M., Marsault, E., Leduc, R., Lachowicz, J.E., Gonias, S.L., Castaigne, J.P., and Sarret, P. (2014). Conjugation of a brain-penetrant peptide with neurotensin provides antinociceptive properties. J Clin Invest 124, 1199-1213.

Demeule, M., Currie, J.C., Bertrand, Y., Che, C., Nguyen, T., Regina, A., Gabathuler, R., Castaigne, J.P., and Beliveau, R. (2008a). Involvement of the low-density lipoprotein receptor-related protein in the transcytosis of the brain delivery vector angiopep-2. J Neurochem 106, 1534-1544.

Demeule, M., Poirier, J., Jodoin, J., Bertrand, Y., Desrosiers, R.R., Dagenais, C., Nguyen, T., Lanthier, J., Gabathuler, R., Kennard, M., Jefferies, W.A., Karkan, D., Tsai, S., Fenart, L., Cecchelli, R., and Beliveau, R. (2002). High transcytosis of melanotransferrin (P97) across the blood-brain barrier. J Neurochem 83, 924-933.

Demeule, M., Regina, A., Che, C., Poirier, J., Nguyen, T., Gabathuler, R., Castaigne, J.P., and Beliveau, R. (2008b). Identification and design of peptides as a new drug delivery system for the brain. J Pharmacol Exp Ther 324, 1064-1072.

Dempsey, C.E., Sessions, R.B., Lamble, N.V., and Campbell, S.J. (2000). The asparagine- stabilized beta-turn of apamin: contribution to structural stability from dynamics simulation and amide hydrogen exchange analysis. Biochemistry 39, 15944-15952.

  Derossi, D., Chassaing, G., and Prochiantz, A. (1998). Trojan peptides: the penetratin system for intracellular delivery. Trends Cell Biol 8, 84-87.

Diaz-Perlas, C., Oller-Salvia, B., Sanchez-Navarro, M., Teixido, M., and Giralt, E. (2018a). Branched BBB-shuttle peptides: chemoselective modification of proteins to enhance blood-brain barrier transport. Chem Sci 9, 8409-8415.

Diaz-Perlas, C., Sanchez-Navarro, M., Oller-Salvia, B., Moreno, M., Teixido, M., and Giralt, E. (2017). Phage display as a tool to discover blood-brain barrier (BBB)-shuttle peptides: panning against a human BBB cellular model. Biopolymers 108.

Diaz-Perlas, C., Varese, M., Guardiola, S., Garcia, J., Sanchez-Navarro, M., Giralt, E., and Teixido, M. (2018b). From venoms to BBB-shuttles. MiniCTX3: a molecular vector derived from scorpion venom. Chem Commun (Camb) 54, 12738-12741.

Drappatz, J., Brenner, A., Wong, E.T., Eichler, A., Schiff, D., Groves, M.D., Mikkelsen, T., Rosenfeld, S., Sarantopoulos, J., Meyers, C.A., Fielding, R.M., Elian, K., Wang, X., Lawrence, B., Shing, M., Kelsey, S., Castaigne, J.P., and Wen, P.Y. (2013). Phase I study of GRN1005 in recurrent malignant glioma. Clin Cancer Res 19, 1567-1576.

Edwards, A.J., and Reid, D. (2001). Introduction to NMR of proteins. Curr Protoc Protein Sci Chapter 17, Unit 17 15.

Figueiredo, P., Balasubramanian, V., Shahbazi, M.A., Correia, A., Wu, D., Palivan, C.G., Hirvonen, J.T., and Santos, H.A. (2016). Angiopep2-functionalized polymersomes for targeted doxorubicin delivery to glioblastoma cells. Int J Pharm 511, 794-803.

Fischer, P.M. (2003). The design, synthesis and application of stereochemical and directional peptide isomers: a critical review. Curr Protein Pept Sci 4, 339-356.

Fotakis, G., and Timbrell, J.A. (2006). In vitro cytotoxicity assays: comparison of LDH, neutral red, MTT and protein assay in hepatoma cell lines following exposure to cadmium chloride. Toxicol Lett 160, 171-177.

Fuster, C., Varese, M., Garcia, J., Giralt, E., Sanchez-Navarro, M., and Teixido, M. (2019). Expanding the MiniAp-4 BBB-shuttle family: Evaluation of proline cis-trans ratio as tool to fine-tune transport. J Pept Sci 25, e3172.

Ganella, D.E., Callander, G.E., Ma, S., Bye, C.R., Gundlach, A.L., and Bathgate, R.A. (2013). Modulation of feeding by chronic rAAV expression of a relaxin-3 peptide agonist in rat hypothalamus. Gene Ther 20, 703-716.

  Ganella, D.E., Ryan, P.J., Bathgate, R.A., and Gundlach, A.L. (2012). Increased feeding and body weight gain in rats after acute and chronic activation of RXFP3 by relaxin-3 and receptor-selective peptides: functional and therapeutic implications. Behav Pharmacol 23, 516-525.

Gao, H., Zhang, S., Cao, S., Yang, Z., Pang, Z., and Jiang, X. (2014). Angiopep-2 and activatable cell-penetrating peptide dual-functionalized nanoparticles for systemic glioma-targeting delivery. Mol Pharm 11, 2755-2763.

Georgieva, J.V., Hoekstra, D., and Zuhorn, I.S. (2014). Smuggling Drugs into the Brain: An Overview of Ligands Targeting Transcytosis for Drug Delivery across the Blood-Brain Barrier. Pharmaceutics 6, 557-583.

Gerozissis, K. (2003). Brain insulin: regulation, mechanisms of action and functions. Cell Mol Neurobiol 23, 1-25.

Giugliani, R., Giugliani, L., De Oliveira Poswar, F., Donis, K.C., Corte, A.D., Schmidt, M., Boado, R.J., Nestrasil, I., Nguyen, C., Chen, S., and Pardridge, W.M. (2018). Neurocognitive and somatic stabilization in pediatric patients with severe Mucopolysaccharidosis Type I after 52 weeks of intravenous brain-penetrating insulin receptor antibody-iduronidase fusion protein (valanafusp alpha): an open label phase 1-2 trial. Orphanet J Rare Dis 13, 110.

Gizurarson, S. (2012). Anatomical and histological factors affecting intranasal drug and vaccine delivery. Curr Drug Deliv 9, 566-582.

Green, B.R., Klein, B.D., Lee, H.K., Smith, M.D., Steve White, H., and Bulaj, G. (2013). Cyclic analogs of galanin and by hydrocarbon stapling. Bioorg Med Chem 21, 303-310.

Green, B.R., Smith, M., White, K.L., White, H.S., and Bulaj, G. (2011). Analgesic neuropeptide W suppresses seizures in the brain revealed by rational repositioning and peptide engineering. ACS Chem Neurosci 2, 51-56.

Green, B.R., White, K.L., Mcdougle, D.R., Zhang, L.Y., Klein, B., Scholl, E.A., Pruess, T.H., White, H.S., and Bulaj, G. (2010). Introduction of lipidization-cationization motifs affords systemically bioavailable neuropeptide Y and neurotensin analogs with anticonvulsant activities. Journal of Peptide Science 16, 486-495.

Gregori, M., Taylor, M., Salvati, E., Re, F., Mancini, S., Balducci, C., Forloni, G., Zambelli, V., Sesana, S., Michael, M., Michail, C., Tinker-Mill, C., Kolosov, O., Sherer, M., Harris, S., Fullwood, N.J., Masserini, M., and Allsop, D. (2017). Retro-inverso peptide

  inhibitor nanoparticles as potent inhibitors of aggregation of the Alzheimer's Abeta peptide. Nanomedicine 13, 723-732.

Guixer, B., Arroyo, X., Belda, I., Sabido, E., Teixido, M., and Giralt, E. (2016). Chemically synthesized peptide libraries as a new source of BBB shuttles. Use of mass spectrometry for peptide identification. J Pept Sci 22, 577-591.

Guntert, P. (2004). Automated NMR structure calculation with CYANA. Methods Mol Biol 278, 353-378.

Habermann, E. (1984). Apamin. Pharmacol Ther 25, 255-270.

Habermann, E., and Cheng-Raude, D. (1975). Central neurotoxicity of apamin, crotamin, phospholipase A and alpha-amanitin. Toxicon 13, 465-473.

Haidar, M., Guevremont, G., Zhang, C., Bathgate, R., Timofeeva, E., Smith, C.M., and Gundlach, A.L. (2017). Relaxin-3 Inputs Target Hippocampal Interneurons and Deletion of Hilar Relaxin-3 Receptors in 'Floxed-RXFP3' Mice Impairs Spatial Memory. Hippocampus.

Hallworth, N.E., Wilson, C.J., and Bevan, M.D. (2003). Apamin-sensitive small conductance calcium-activated potassium channels, through their selective coupling to voltage-gated calcium channels, are critical determinants of the precision, pace, and pattern of action potential generation in rat subthalamic nucleus neurons in vitro. Journal of Neuroscience 23, 7525-7542.

Hanafy, A.S., Farid, R.M., and Elgamal, S.S. (2015). Complexation as an approach to entrap cationic drugs into cationic nanoparticles administered intranasally for Alzheimer's disease management: preparation and detection in rat brain. Drug Development and Industrial Pharmacy 41, 2055-2068.

Hanafy, A.S., Farid, R.M., Helmy, M.W., and Elgamal, S.S. (2016). Pharmacological, toxicological and neuronal localization assessment of galantamine/chitosan complex nanoparticles in rats: future potential contribution in Alzheimer's disease management. Drug Deliv 23, 3111-3122.

Haugaard-Jonsson, L.M., Hossain, M.A., Daly, N.L., Bathgate, R.A., Wade, J.D., Craik, D.J., and Rosengren, K.J. (2008). Structure of the R3/I5 chimeric relaxin peptide, a selective GPCR135 and GPCR142 agonist. J Biol Chem 283, 23811-23818.

Haugaard-Jonsson, L.M., Hossain, M.A., Daly, N.L., Bathgate, R.A., Wade, J.D., Craik, D.J., and Rosengren, K.J. (2009). Structural properties of relaxin chimeras. Ann N Y Acad Sci 1160, 27-30.

  Haugaard-Kedström, L.M., Lee, H.S., Jones, M.V., Song, A., Rathod, V., Hossain, M.A., Bathgate, R.a.D., and Rosengren, K.J. (2018). Binding conformation and determinants of a single-chain peptide antagonist at the relaxin-3 receptor RXFP3. J Biol Chem 293, 15765-15776.

Haugaard-Kedström, L.M., Shabanpoor, F., Hossain, M.A., Clark, R.J., Ryan, P.J., Craik, D.J., Gundlach, A.L., Wade, J.D., Bathgate, R.a.D., and Rosengren, K.J. (2011). Design, synthesis, and characterization of a single-chain peptide antagonist for the relaxin-3 receptor RXFP3. Journal of the American Chemical Society 133, 4965-4974.

Haugaard-Kedström, L.M., Wong, L.L., Bathgate, R.A., and Rosengren, K.J. (2015). Synthesis and pharmacological characterization of a europium-labelled single-chain antagonist for binding studies of the relaxin-3 receptor RXFP3. Amino Acids 47, 1267-1271.

Helms, H.C., Abbott, N.J., Burek, M., Cecchelli, R., Couraud, P.O., Deli, M.A., Forster, C., Galla, H.J., Romero, I.A., Shusta, E.V., Stebbins, M.J., Vandenhaute, E., Weksler, B., and Brodin, B. (2016). In vitro models of the blood-brain barrier: An overview of commonly used brain endothelial cell culture models and guidelines for their use. Journal of Cerebral Blood Flow and Metabolism 36, 862-890.

Henninot, A., Collins, J.C., and Nuss, J.M. (2018). The Current State of Peptide Drug Discovery: Back to the Future? J Med Chem 61, 1382-1414.

Herman, J.P., Cullinan, W.E., Ziegler, D.R., and Tasker, J.G. (2002). Role of the paraventricular nucleus microenvironment in stress integration. Eur J Neurosci 16, 381-385.

Herve, F., Ghinea, N., and Scherrmann, J.M. (2008). CNS delivery via adsorptive transcytosis. AAPS J 10, 455-472.

Hida, T., Takahashi, E., Shikata, K., Hirohashi, T., Sawai, T., Seiki, T., Tanaka, H., Kawai, T., Ito, O., Arai, T., Yokoi, A., Hirakawa, T., Ogura, H., Nagasu, T., Miyamoto, N., and Kuromitsu, J. (2006). Chronic intracerebroventricular administration of relaxin-3 increases body weight in rats. J Recept Signal Transduct Res 26, 147-158.

Hojo, K., Hossain, M.A., Tailhades, J., Shabanpoor, F., Wong, L.L., Ong-Palsson, E.E., Kastman, H.E., Ma, S.K., Gundlach, A.L., Rosengren, K.J., Wade, J.D., and Bathgate, R.A. (2016). Development of a single-chain peptide agonist of the relaxin-3 receptor using hydrocarbon stapling. J Med Chem.

Hosken, I.T., Sutton, S.W., Smith, C.M., and Gundlach, A.L. (2015). Relaxin-3 receptor (Rxfp3) gene knockout mice display reduced running wheel activity: implications for role of relaxin-3/RXFP3 signalling in sustained arousal. Behav Brain Res 278, 167-175.

  Hossain, M.A., Bathgate, R.A., Rosengren, K.J., Shabanpoor, F., Zhang, S., Lin, F., Tregear, G.W., and Wade, J.D. (2009). The structural and functional role of the B-chain C- terminal arginine in the relaxin-3 peptide antagonist, R3(BDelta23-27)R/I5. Chem Biol Drug Des 73, 46-52.

Hossain, M.A., Haugaard-Kedström, L.M., Rosengren, K.J., Bathgate, R.A., and Wade, J.D. (2015). Chemically synthesized dicarba H2 relaxin analogues retain strong RXFP1 receptor activity but show an unexpected loss of in vitro serum stability. Org Biomol Chem 13, 10895-10903.

Hossain, M.A., Kocan, M., Yao, S.T., Royce, S.G., Nair, V.B., Siwek, C., Patil, N.A., Harrison, I.P., Rosengren, K.J., Selemidis, S., Summers, R.J., Wade, J.D., Bathgate, R.a.D., and Samuel, C.S. (2016). A single-chain derivative of the relaxin hormone is a functionally selective agonist of the G protein-coupled receptor, RXFP1. Chemical Science 7, 3805- 3819.

Hossain, M.A., Man, B.C., Zhao, C., Xu, Q., Du, X.J., Wade, J.D., and Samuel, C.S. (2011). H3 relaxin demonstrates antifibrotic properties via the RXFP1 receptor. Biochemistry 50, 1368-1375.

Hossain, M.A., Rosengren, K.J., Haugaard-Jonsson, L.M., Zhang, S., Layfield, S., Ferraro, T., Daly, N.L., Tregear, G.W., Wade, J.D., and Bathgate, R.A. (2008). The A-chain of human relaxin family peptides has distinct roles in the binding and activation of the different relaxin family peptide receptors. J Biol Chem 283, 17287-17297.

Hsu, S.Y. (1999). Cloning of two novel mammalian paralogs of relaxin/insulin family proteins and their expression in testis and kidney. Molecular Endocrinology 13, 2163-2174.

Hsu, S.Y., Nakabayashi, K., Nishi, S., Kumagai, J., Kudo, M., Sherwood, O.D., and Hsueh, A.J. (2002). Activation of orphan receptors by the hormone relaxin. Science 295, 671- 674.

Hu, M.J., Shao, X.X., Wang, J.H., Wei, D., Liu, Y.L., Xu, Z.G., and Guo, Z.Y. (2016). Identification of hydrophobic interactions between relaxin-3 and its receptor RXFP3: implication for a conformational change in the B-chain C-terminus during receptor binding. Amino Acids 48, 2227-2236.

Huang, R.Q., Qu, Y.H., Ke, W.L., Zhu, J.H., Pei, Y.Y., and Jiang, C. (2007). Efficient gene delivery targeted to the brain using a transferrin-conjugated polyethyleneglycol- modified polyamidoamine dendrimer. FASEB J 21, 1117-1125.

  Huang, S., Li, J., Han, L., Liu, S., Ma, H., Huang, R., and Jiang, C. (2011). Dual targeting effect of Angiopep-2-modified, DNA-loaded nanoparticles for glioma. Biomaterials 32, 6832-6838.

Hunter, K., and Holscher, C. (2012). Drugs developed to treat diabetes, liraglutide and , cross the blood brain barrier and enhance neurogenesis. BMC Neurosci 13, 33.

Hussain, M.M., Strickland, D.K., and Bakillah, A. (1999). The mammalian low-density lipoprotein receptor family. Annu Rev Nutr 19, 141-172.

Jacquot, G., Lecorche, P., Malcor, J.D., Laurencin, M., Smirnova, M., Varini, K., Malicet, C., Gassiot, F., Abouzid, K., Faucon, A., David, M., Gaudin, N., Masse, M., Ferracci, G., Dive, V., Cisternino, S., and Khrestchatisky, M. (2016). Optimization and in Vivo Validation of Peptide Vectors Targeting the LDL Receptor. Mol Pharm 13, 4094-4105.

Jayakody, T., Marwari, S., Lakshminarayanan, R., Tan, F.C., Johannes, C.W., Dymock, B.W., Poulsen, A., Herr, D.R., and Dawe, G.S. (2016). Hydrocarbon stapled B chain analogues of relaxin-3 retain biological activity. Peptides 84, 44-57.

Jefferies, W.A., Brandon, M.R., Hunt, S.V., Williams, A.F., Gatter, K.C., and Mason, D.Y. (1984). Transferrin receptor on endothelium of brain capillaries. Nature 312, 162-163.

Jensen, K.J., Shelton, P.T., and Pedersen, S.L. (2013). Peptide Synthesis and Applications Second Edition Preface. Peptide Synthesis and Applications, 2nd Edition 1047, V-V.

Ji, Y., Majumder, S., Millard, M., Borra, R., Bi, T., Elnagar, A.Y., Neamati, N., Shekhtman, A., and Camarero, J.A. (2013). In vivo activation of the p53 tumor suppressor pathway by an engineered cyclotide. J Am Chem Soc 135, 11623-11633.

Jo, H., Meinhardt, N., Wu, Y., Kulkarni, S., Hu, X., Low, K.E., Davies, P.L., Degrado, W.F., and Greenbaum, D.C. (2012). Development of alpha-helical calpain probes by mimicking a natural protein-protein interaction. J Am Chem Soc 134, 17704-17713.

Johnsen, K.B., Burkhart, A., Melander, F., Kempen, P.J., Vejlebo, J.B., Siupka, P., Nielsen, M.S., Andresen, T.L., and Moos, T. (2017). Targeting transferrin receptors at the blood- brain barrier improves the uptake of immunoliposomes and subsequent cargo transport into the brain parenchyma. Sci Rep 7, 10396.

Jones, A.R., and Shusta, E.V. (2007). Blood-brain barrier transport of therapeutics via receptor- mediation. Pharm Res 24, 1759-1771.

  Josan, J.S., De Silva, C.R., Yoo, B., Lynch, R.M., Pagel, M.D., Vagner, J., and Hruby, V.J. (2011). Fluorescent and Lanthanide Labeling for Ligand Screens, Assays, and Imaging. Drug Design and Discovery: Methods and Protocols 716, 89-126.

Kansy, M., Senner, F., and Gubernator, K. (1998). Physicochemical high throughput screening: parallel artificial membrane permeation assay in the description of passive absorption processes. J Med Chem 41, 1007-1010.

Karkan, D., Pfeifer, C., Vitalis, T.Z., Arthur, G., Ujiie, M., Chen, Q., Tsai, S., Koliatis, G., Gabathuler, R., and Jefferies, W.A. (2008). A unique carrier for delivery of therapeutic compounds beyond the blood-brain barrier. PLoS One 3, e2469.

Kastin, A.J. (2013). Handbook of Biologically Active Peptides. San Diego, California: Elsevier.

Keller, R.L.J. (2004). The Computer Aided Resonance Assignment Tutorial. Switzerland: Cantina Verlag.

Kilic, E., Kilic, U., and Hermann, D.M. (2005). TAT-GDNF in neurodegeneration and ischemic stroke. CNS Drug Rev 11, 369-378.

Knudsen, L.B., and Lau, J. (2019). The Discovery and Development of Liraglutide and Semaglutide. Front Endocrinol (Lausanne) 10, 155.

Knudsen, L.B., Nielsen, P.F., Huusfeldt, P.O., Johansen, N.L., Madsen, K., Pedersen, F.Z., Thogersen, H., Wilken, M., and Agerso, H. (2000). Potent derivatives of -like peptide-1 with pharmacokinetic properties suitable for once daily administration. J Med Chem 43, 1664-1669.

Kocan, M., Ang, S.Y., and Summers, R.J. (2015). Orthosteric, Allosteric and Biased Signalling at the Relaxin-3 Receptor RXFP3. Neurochem Res.

Kooijmans, S.A., Senyschyn, D., Mezhiselvam, M.M., Morizzi, J., Charman, S.A., Weksler, B., Romero, I.A., Couraud, P.O., and Nicolazzo, J.A. (2012). The involvement of a Na(+)- and Cl(-)-dependent transporter in the brain uptake of amantadine and rimantadine. Mol Pharm 9, 883-893.

Kreuter, J., Hekmatara, T., Dreis, S., Vogel, T., Gelperina, S., and Langer, K. (2007). Covalent attachment of apolipoprotein A-I and apolipoprotein B-100 to albumin nanoparticles enables drug transport into the brain. Journal of Controlled Release 118, 54-58.

Kuang, Y., Jiang, X., Zhang, Y., Lu, Y., Ma, H., Guo, Y., Zhang, Y., An, S., Li, J., Liu, L., Wu, Y., Liang, J., and Jiang, C. (2016). Dual Functional Peptide-Driven Nanoparticles

  for Highly Efficient Glioma-Targeting and Drug Codelivery. Mol Pharm 13, 1599- 1607.

Kuei, C., Sutton, S., Bonaventure, P., Pudiak, C., Shelton, J., Zhu, J., Nepomuceno, D., Wu, J., Chen, J., Kamme, F., Seierstad, M., Hack, M.D., Bathgate, R.A., Hossain, M.A., Wade, J.D., Atack, J., Lovenberg, T.W., and Liu, C. (2007). R3(BDelta23 27)R/I5 chimeric peptide, a selective antagonist for GPCR135 and GPCR142 over relaxin receptor LGR7: in vitro and in vivo characterization. J Biol Chem 282, 25425-25435.

Kumagai, J., Hsu, S.Y., Matsumi, H., Roh, J.S., Fu, P., Wade, J.D., Bathgate, R.A., and Hsueh, A.J. (2002). INSL3/Leydig insulin-like peptide activates the LGR8 receptor important in testis descent. J Biol Chem 277, 31283-31286.

Kumar, J.R., Rajkumar, R., Jayakody, T., Marwari, S., Hong, J.M., Ma, S., Gundlach, A.L., Lai, M.K.P., and Dawe, G.S. (2017). Relaxin' the brain: a case for targeting the nucleus incertus network and relaxin-3/RXFP3 system in neuropsychiatric disorders. Br J Pharmacol 174, 1061-1076.

Kumar, V., Lee, J.D., Clark, R.J., and Woodruff, T.M. (2018). Development and validation of a LC-MS/MS assay for pharmacokinetic studies of complement antagonists PMX53 and PMX205 in mice. Sci Rep 8, 8101.

Kurzrock, R., Gabrail, N., Chandhasin, C., Moulder, S., Smith, C., Brenner, A., Sankhala, K., Mita, A., Elian, K., Bouchard, D., and Sarantopoulos, J. (2012). Safety, pharmacokinetics, and activity of GRN1005, a novel conjugate of angiopep-2, a peptide facilitating brain penetration, and paclitaxel, in patients with advanced solid tumors. Mol Cancer Ther 11, 308-316.

Labbe-Jullie, C., Granier, C., Albericio, F., Defendini, M.L., Ceard, B., Rochat, H., and Van Rietschoten, J. (1991). Binding and toxicity of apamin. Characterization of the active site. Eur J Biochem 196, 639-645.

Lau, J., Bloch, P., Schaffer, L., Pettersson, I., Spetzler, J., Kofoed, J., Madsen, K., Knudsen, L.B., Mcguire, J., Steensgaard, D.B., Strauss, H.M., Gram, D.X., Knudsen, S.M., Nielsen, F.S., Thygesen, P., Reedtz-Runge, S., and Kruse, T. (2015). Discovery of the Once-Weekly Glucagon-Like Peptide-1 (GLP-1) Analogue Semaglutide. J Med Chem 58, 7370-7380.

Lau, J.L., and Dunn, M.K. (2018). Therapeutic peptides: Historical perspectives, current development trends, and future directions. Bioorg Med Chem 26, 2700-2707.

Lawther, A.J., Clissold, M.L., Ma, S., Kent, S., Lowry, C.A., Gundlach, A.L., and Hale, M.W. (2015). Anxiogenic drug administration and elevated plus-maze exposure in rats

  activate populations of relaxin-3 neurons in the nucleus incertus and serotonergic neurons in the dorsal raphe nucleus. Neuroscience 303, 270-284.

Le-Nguyen, D., Chiche, L., Hoh, F., Martin-Eauclaire, M.F., Dumas, C., Nishi, Y., Kobayashi, Y., and Aumelas, A. (2007). Role of Asn(2) and Glu(7) residues in the oxidative folding and on the conformation of the N-terminal loop of apamin. Biopolymers 86, 447-462.

Lee, J.H., Engler, J.A., Collawn, J.F., and Moore, B.A. (2001). Receptor mediated uptake of peptides that bind the human transferrin receptor. Eur J Biochem 268, 2004-2012.

Leng, G., and Sabatier, N. (2017). Oxytocin - The Sweet Hormone? Trends Endocrinol Metab 28, 365-376.

Lenglos, C., Calvez, J., and Timofeeva, E. (2015). Sex-specific effects of relaxin-3 on food intake and brain expression of corticotropin-releasing factor in rats. Endocrinology 156, 523-533.

Lenglos, C., Mitra, A., Guevremont, G., and Timofeeva, E. (2013). Sex differences in the effects of chronic stress and food restriction on body weight gain and brain expression of CRF and relaxin-3 in rats. Genes Brain Behav 12, 370-387.

Li, C., Pazgier, M., Li, J., Li, C., Liu, M., Zou, G., Li, Z., Chen, J., Tarasov, S.G., Lu, W.Y., and Lu, W. (2010). Limitations of peptide retro-inverso isomerization in molecular mimicry. J Biol Chem 285, 19572-19581.

Li, C., Pazgier, M., Liu, M., Lu, W.Y., and Lu, W. (2009). Apamin as a template for structure- based rational design of potent peptide activators of p53. Angew Chem Int Ed Engl 48, 8712-8715.

Li, F., and Tang, S.C. (2017). Targeting metastatic breast cancer with ANG1005, a novel peptide-paclitaxel conjugate that crosses the blood-brain-barrier (BBB). Genes Dis 4, 1-3.

Li, J., Feng, L., Fan, L., Zha, Y., Guo, L., Zhang, Q., Chen, J., Pang, Z., Wang, Y., Jiang, X., Yang, V.C., and Wen, L. (2011). Targeting the brain with PEG-PLGA nanoparticles modified with phage-displayed peptides. Biomaterials 32, 4943-4950.

Li, J.W., Zhang, Q.Z., Pang, Z.Q., Wang, Y.C., Liu, Q.F., Guo, L.R., and Jiang, X.G. (2012). Identification of peptide sequences that target to the brain using in vivo phage display. Amino Acids 42, 2373-2381.

  Lin, Q., Mao, K.L., Tian, F.R., Yang, J.J., Chen, P.P., Xu, J., Fan, Z.L., Zhao, Y.P., Li, W.F., Zheng, L., Zhao, Y.Z., and Lu, C.T. (2016). Brain tumor-targeted delivery and therapy by focused ultrasound introduced doxorubicin-loaded cationic liposomes. Cancer Chemotherapy and Pharmacology 77, 269-280.

Liu, C., Chen, J., Kuei, C., Sutton, S., Nepomuceno, D., Bonaventure, P., and Lovenberg, T.W. (2005a). Relaxin-3/insulin-like peptide 5 chimeric peptide, a selective ligand for G protein-coupled receptor (GPCR)135 and GPCR142 over leucine-rich repeat- containing G protein-coupled receptor 7. Mol Pharmacol 67, 231-240.

Liu, C., Chen, J., Sutton, S., Roland, B., Kuei, C., Farmer, N., Sillard, R., and Lovenberg, T.W. (2003a). Identification of relaxin-3/INSL7 as a ligand for GPCR142. J Biol Chem 278, 50765-50770.

Liu, C., Eriste, E., Sutton, S., Chen, J., Roland, B., Kuei, C., Farmer, N., Jornvall, H., Sillard, R., and Lovenberg, T.W. (2003b). Identification of relaxin-3/INSL7 as an endogenous ligand for the orphan G-protein-coupled receptor GPCR135. J Biol Chem 278, 50754- 50764.

Liu, C., Kuei, C., Sutton, S., Chen, J., Bonaventure, P., Wu, J., Nepomuceno, D., Kamme, F., Tran, D.T., Zhu, J., Wilkinson, T., Bathgate, R., Eriste, E., Sillard, R., and Lovenberg, T.W. (2005b). INSL5 is a high affinity specific agonist for GPCR142 (GPR100). J Biol Chem 280, 292-300.

Liu, C., Kuei, C., Sutton, S., Shelton, J., Zhu, J., Nepomuceno, D., Hossain, M.A., Wade, J.D., Bathgate, R.A., Bonaventure, P., and Lovenberg, T. (2009a). Probing the functional domains of relaxin-3 and the creation of a selective antagonist for RXFP3/GPCR135 over relaxin receptor RXFP1/LGR7. Ann N Y Acad Sci 1160, 31-37.

Liu, H.L., Hua, M.Y., Chen, P.Y., Chu, P.C., Pan, C.H., Yang, H.W., Huang, C.Y., Wang, J.J., Yen, T.C., and Wei, K.C. (2010). Blood-brain barrier disruption with focused ultrasound enhances delivery of chemotherapeutic drugs for glioblastoma treatment. Radiology 255, 415-425.

Liu, L., Xu, K., Wang, H., Tan, P.K., Fan, W., Venkatraman, S.S., Li, L., and Yang, Y.Y. (2009b). Self-assembled cationic peptide nanoparticles as an efficient antimicrobial agent. Nat Nanotechnol 4, 457-463.

Liu, Y., Ran, R., Chen, J., Kuang, Q., Tang, J., Mei, L., Zhang, Q., Gao, H., Zhang, Z., and He, Q. (2014). Paclitaxel loaded liposomes decorated with a multifunctional tandem peptide for glioma targeting. Biomaterials 35, 4835-4847.

  Liu, Y., Zhang, L., Shao, X.X., Hu, M.J., Liu, Y.L., Xu, Z.G., and Guo, Z.Y. (2016). A negatively charged transmembrane aspartate residue controls activation of the relaxin- 3 receptor RXFP3. Arch Biochem Biophys.

Lyon, R.P., Setter, J.R., Bovee, T.D., Doronina, S.O., Hunter, J.H., Anderson, M.E., Balasubramanian, C.L., Duniho, S.M., Leiske, C.I., Li, F., and Senter, P.D. (2014). Self-hydrolyzing maleimides improve the stability and pharmacological properties of antibody-drug conjugates. Nat Biotechnol 32, 1059-1062.

Ma, S., Allocca, G., Ong-Palsson, E.K., Singleton, C.E., Hawkes, D., Mcdougall, S.J., Williams, S.J., Bathgate, R.A., and Gundlach, A.L. (2017a). Nucleus incertus promotes cortical desynchronization and behavioral arousal. Brain Struct Funct 222, 515-537.

Ma, S., Bonaventure, P., Ferraro, T., Shen, P.J., Burazin, T.C., Bathgate, R.A., Liu, C., Tregear, G.W., Sutton, S.W., and Gundlach, A.L. (2007). Relaxin-3 in GABA projection neurons of nucleus incertus suggests widespread influence on forebrain circuits via G- protein-coupled receptor-135 in the rat. Neuroscience 144, 165-190.

Ma, S., Hangya, B., Leonard, C.S., Wisden, W., and Gundlach, A.L. (2018). Dual-transmitter systems regulating arousal, attention, learning and memory. Neurosci Biobehav Rev 85, 21-33.

Ma, S., Olucha-Bordonau, F.E., Hossain, M.A., Lin, F., Kuei, C., Liu, C., Wade, J.D., Sutton, S.W., Nunez, A., and Gundlach, A.L. (2009). Modulation of hippocampal theta oscillations and spatial memory by relaxin-3 neurons of the nucleus incertus. Learn Mem 16, 730-742.

Ma, S., Smith, C.M., Blasiak, A., and Gundlach, A.L. (2017b). Distribution, physiology and pharmacology of relaxin-3/RXFP3 systems in brain. Br J Pharmacol 174, 1034-1048.

Madsen, K., Knudsen, L.B., Agersoe, H., Nielsen, P.F., Thogersen, H., Wilken, M., and Johansen, N.L. (2007). Structure-activity and protraction relationship of long-acting glucagon-like peptide-1 derivatives: importance of fatty acid length, polarity, and bulkiness. J Med Chem 50, 6126-6132.

Magotra, A., Sharma, A., Gupta, A.P., Wazir, P., Sharma, S., Singh, P.P., Tikoo, M.K., Vishwakarma, R.A., Singh, G., and Nandi, U. (2017). Development and validation of a highly sensitive LC-ESI-MS/MS method for estimation of IIIM-MCD-211, a novel nitrofuranyl methyl piperazine derivative with potential activity against tuberculosis: Application to drug development. J Chromatogr B Analyt Technol Biomed Life Sci 1060, 200-206.

  Mahalakshmi, R., and Balaram, P. (2006). Non-protein amino acids in the design of secondary structure scaffolds. Methods Mol Biol 340, 71-94.

Malakoutikhah, M., Teixido, M., and Giralt, E. (2011). Shuttle-Mediated Drug Delivery to the Brain. Angewandte Chemie-International Edition 50, 7998-8014.

Maletinska, L., Nagelova, V., Ticha, A., Zemenova, J., Pirnik, Z., Holubova, M., Spolcova, A., Mikulaskova, B., Blechova, M., Sykora, D., Lacinova, Z., Haluzik, M., Zelezna, B., and Kunes, J. (2015). Novel lipidized analogs of prolactin-releasing peptide have prolonged half-lives and exert anti-obesity effects after peripheral administration. Int J Obes (Lond) 39, 986-993.

Marmiroli, N., and Maestri, E. (2014). Plant peptides in defense and signaling. Peptides 56, 30-44.

Marwari, S., Poulsen, A., Shih, N., Lakshminarayanan, R., Kini, R.M., Johannes, C.W., Dymock, B.W., and Dawe, G.S. (2019). Intranasal administration of a stapled relaxin- 3 mimetic has anxiolytic- and antidepressant-like activity in rats. Br J Pharmacol.

Matsumoto, M., Kamohara, M., Sugimoto, T., Hidaka, K., Takasaki, J., Saito, T., Okada, M., Yamaguchi, T., and Furuichi, K. (2000). The novel G-protein coupled receptor SALPR shares sequence similarity with somatostatin and angiotensin receptors. Gene 248, 183- 189.

Mccoll, D.J., Honchell, C.D., and Frankel, A.D. (1999). Structure-based design of an RNA- binding zinc finger. Proc Natl Acad Sci U S A 96, 9521-9526.

Mcconnell, H.L., Kersch, C.N., Woltjer, R.L., and Neuwelt, E.A. (2017). The Translational Significance of the Neurovascular Unit. J Biol Chem 292, 762-770.

Mcgowan, B.M., Stanley, S.A., Smith, K.L., Minnion, J.S., Donovan, J., Thompson, E.L., Patterson, M., Connolly, M.M., Abbott, C.R., Small, C.J., Gardiner, J.V., Ghatei, M.A., and Bloom, S.R. (2006). Effects of acute and chronic relaxin-3 on food intake and energy expenditure in rats. Regul Pept 136, 72-77.

Mcgowan, B.M., Stanley, S.A., Smith, K.L., White, N.E., Connolly, M.M., Thompson, E.L., Gardiner, J.V., Murphy, K.G., Ghatei, M.A., and Bloom, S.R. (2005). Central relaxin- 3 administration causes hyperphagia in male Wistar rats. Endocrinology 146, 3295- 3300.

Meredith, M.E., Salameh, T.S., and Banks, W.A. (2015). Intranasal Delivery of Proteins and Peptides in the Treatment of Neurodegenerative Diseases. AAPS J 17, 780-787.

  Merrifield, R.B. (1963). Solid Phase Peptide Synthesis .1. Synthesis of a Tetrapeptide. Journal of the American Chemical Society 85, 2149-&.

Milletti, F. (2012). Cell-penetrating peptides: classes, origin, and current landscape. Drug Discov Today 17, 850-860.

Mittal, D., Ali, A., Md, S., Baboota, S., Sahni, J.K., and Ali, J. (2014). Insights into direct nose to brain delivery: current status and future perspective. Drug Deliv 21, 75-86.

Molino, Y., David, M., Varini, K., Jabes, F., Gaudin, N., Fortoul, A., Bakloul, K., Masse, M., Bernard, A., Drobecq, L., Lecorche, P., Temsamani, J., Jacquot, G., and Khrestchatisky, M. (2017). Use of LDL receptor-targeting peptide vectors for in vitro and in vivo cargo transport across the blood-brain barrier. FASEB J.

Mosmann, T. (1983). Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J Immunol Methods 65, 55-63.

Nategh, M., Nikseresht, S., Khodagholi, F., and Motamedi, F. (2015). Nucleus incertus inactivation impairs spatial learning and memory in rats. Physiol Behav 139, 112-120.

Niewoehner, J., Bohrmann, B., Collin, L., Urich, E., Sade, H., Maier, P., Rueger, P., Stracke, J.O., Lau, W., Tissot, A.C., Loetscher, H., Ghosh, A., and Freskgard, P.O. (2014). Increased brain penetration and potency of a therapeutic antibody using a monovalent molecular shuttle. Neuron 81, 49-60.

Nisbet, R.M., Van Der Jeugd, A., Leinenga, G., Evans, H.T., Janowicz, P.W., and Gotz, J. (2017). Combined effects of scanning ultrasound and a tau-specific single chain antibody in a tau transgenic mouse model. Brain 140, 1220-1230.

Nunez, A., Cervera-Ferri, A., Olucha-Bordonau, F., Ruiz-Torner, A., and Teruel, V. (2006). Nucleus incertus contribution to hippocampal theta rhythm generation. Eur J Neurosci 23, 2731-2738.

Ohshima-Hosoyama, S., Simmons, H.A., Goecks, N., Joers, V., Swanson, C.R., Bondarenko, V., Velotta, R., Brunner, K., Wood, L.D., Hruban, R.H., and Emborg, M.E. (2012). A monoclonal antibody-GDNF fusion protein is not neuroprotective and is associated with proliferative pancreatic lesions in parkinsonian monkeys. PLoS One 7, e39036.

Oller-Salvia, B., Sanchez-Navarro, M., Ciudad, S., Guiu, M., Arranz-Gibert, P., Garcia, C., Gomis, R.R., Cecchelli, R., Garcia, J., Giralt, E., and Teixido, M. (2016a). MiniAp-4: A Venom-Inspired Peptidomimetic for Brain Delivery. Angew Chem Int Ed Engl 55, 572-575.

  Oller-Salvia, B., Sanchez-Navarro, M., Giralt, E., and Teixido, M. (2016b). Blood-brain barrier shuttle peptides: an emerging paradigm for brain delivery. Chem Soc Rev.

Oller-Salvia, B., Teixido, M., and Giralt, E. (2013). From venoms to BBB shuttles: Synthesis and blood-brain barrier transport assessment of apamin and a nontoxic analog. Biopolymers 100, 675-686.

Olucha-Bordonau, F.E., Albert-Gasco, H., Ros-Bernal, F., Rytova, V., Ong-Palsson, E.K.E., Ma, S., Sanchez-Perez, A.M., and Gundlach, A.L. (2018). Modulation of forebrain function by nucleus incertus and relaxin-3/RXFP3 signaling. CNS Neurosci Ther 24, 694-702.

Ottaviani, G., Martel, S., and Carrupt, P.A. (2006). Parallel artificial membrane permeability assay: a new membrane for the fast prediction of passive human skin permeability. J Med Chem 49, 3948-3954.

Palm, C., Jayamanne, M., Kjellander, M., and Hallbrink, M. (2007). Peptide degradation is a critical determinant for cell-penetrating peptide uptake. Biochim Biophys Acta 1768, 1769-1776.

Patel, M.M., and Patel, B.M. (2017). Crossing the Blood-Brain Barrier: Recent Advances in Drug Delivery to the Brain. CNS Drugs.

Paterson, D.J., Tassieri, M., Reboud, J., Wilson, R., and Cooper, J.M. (2017). Lipid topology and electrostatic interactions underpin lytic activity of linear cationic antimicrobial peptides in membranes. Proc Natl Acad Sci U S A 114, E8324-E8332.

Patil, N.A., Rosengren, K.J., Separovic, F., Wade, J.D., Bathgate, R.a.D., and Hossain, M.A. (2017). Relaxin family peptides: structure-activity relationship studies. Br J Pharmacol 174, 950-961.

Pease, J.H., and Wemmer, D.E. (1988). Solution structure of apamin determined by nuclear magnetic resonance and distance geometry. Biochemistry 27, 8491-8498.

Phan, T., Nguyen, H.D., Goksel, H., Mocklinghoff, S., and Brunsveld, L. (2010). Phage display selection of miniprotein binders of the Estrogen Receptor. Chem Commun (Camb) 46, 8207-8209.

Poduslo, J.F., and Curran, G.L. (1996). Polyamine modification increases the permeability of proteins at the blood-nerve and blood-brain barriers. J Neurochem 66, 1599-1609.

  Ponka, P., and Lok, C.N. (1999). The transferrin receptor: role in health and disease. Int J Biochem Cell Biol 31, 1111-1137.

Prades, R., Guerrero, S., Araya, E., Molina, C., Salas, E., Zurita, E., Selva, J., Egea, G., Lopez- Iglesias, C., Teixido, M., Kogan, M.J., and Giralt, E. (2012). Delivery of gold nanoparticles to the brain by conjugation with a peptide that recognizes the transferrin receptor. Biomaterials 33, 7194-7205.

Prades, R., Oller-Salvia, B., Schwarzmaier, S.M., Selva, J., Moros, M., Balbi, M., Grazu, V., De La Fuente, J.M., Egea, G., Plesnila, N., Teixido, M., and Giralt, E. (2015). Applying the retro-enantio approach to obtain a peptide capable of overcoming the blood-brain barrier. Angew Chem Int Ed Engl 54, 3967-3972.

Praveen, P., Kocan, M., Valkovic, A., Bathgate, R., and Hossain, M.A. (2019). Single chain peptide agonists of relaxin receptors. Mol Cell Endocrinol 487, 34-39.

Qin, Y., Chen, H., Zhang, Q., Wang, X., Yuan, W., Kuai, R., Tang, J., Zhang, L., Zhang, Z., Zhang, Q., Liu, J., and He, Q. (2011). Liposome formulated with TAT-modified cholesterol for improving brain delivery and therapeutic efficacy on brain glioma in animals. Int J Pharm 420, 304-312.

Quintana, D.S., Guastella, A.J., Westlye, L.T., and Andreassen, O.A. (2016). The promise and pitfalls of intranasally administering psychopharmacological agents for the treatment of psychiatric disorders. Mol Psychiatry 21, 29-38.

Rao, K.S., Reddy, M.K., Horning, J.L., and Labhasetwar, V. (2008). TAT-conjugated nanoparticles for the CNS delivery of anti-HIV drugs. Biomaterials 29, 4429-4438.

Regina, A., Demeule, M., Che, C., Lavallee, I., Poirier, J., Gabathuler, R., Beliveau, R., and Castaigne, J.P. (2008). Antitumour activity of ANG1005, a conjugate between paclitaxel and the new brain delivery vector Angiopep-2. Br J Pharmacol 155, 185- 197.

Regina, A., Demeule, M., Tripathy, S., Lord-Dufour, S., Currie, J.C., Iddir, M., Annabi, B., Castaigne, J.P., and Lachowicz, J.E. (2015). ANG4043, a Novel Brain-Penetrant Peptide-mAb Conjugate, Is Efficacious against HER2-Positive Intracranial Tumors in Mice. Molecular Cancer Therapeutics 14, 129-140.

Rhea, E.M., and Banks, W.A. (2019). Role of the Blood-Brain Barrier in Central Nervous System Insulin Resistance. Front Neurosci 13, 521.

Ribeiro, M.M., Domingues, M.M., Freire, J.M., Santos, N.C., and Castanho, M.A. (2012). Translocating the blood-brain barrier using electrostatics. Front Cell Neurosci 6, 44.

  Riss, T.L., Moravec, R.A., Niles, A.L., Duellman, S., Benink, H.A., Worzella, T.J., and Minor, L. (2004). "Cell Viability Assays," in Assay Guidance Manual, eds. G.S. Sittampalam, N.P. Coussens, H. Nelson, M. Arkin, D. Auld, C. Austin, B. Bejcek, M. Glicksman, J. Inglese, P.W. Iversen, Z. Li, J. Mcgee, O. Mcmanus, L. Minor, A. Napper, J.M. Peltier, T. Riss, O.J. Trask, Jr. & J. Weidner. (Bethesda (MD)).

Robertson, C.R., Pruess, T.H., Grussendorf, E., White, H.S., and Bulaj, G. (2012). Generating orally active galanin analogues with analgesic activities. ChemMedChem 7, 903-909.

Rosengren, K.J., Lin, F., Bathgate, R.A., Tregear, G.W., Daly, N.L., Wade, J.D., and Craik, D.J. (2006). Solution structure and novel insights into the determinants of the receptor specificity of human relaxin-3. J Biol Chem 281, 5845-5851.

Rouault, T.A., and Cooperman, S. (2006). Brain iron metabolism. Semin Pediatr Neurol 13, 142-148.

Rousselle, C., Clair, P., Lefauconnier, J.M., Kaczorek, M., Scherrmann, J.M., and Temsamani, J. (2000). New advances in the transport of doxorubicin through the blood-brain barrier by a peptide vector-mediated strategy. Molecular Pharmacology 57, 679-686.

Rousselle, C., Clair, P., Smirnova, M., Kolesnikov, Y., Pasternak, G.W., Gac-Breton, S., Rees, A.R., Scherrmann, J.M., and Temsamani, J. (2003). Improved brain uptake and pharmacological activity of dalargin using a peptide-vector-mediated strategy. J Pharmacol Exp Ther 306, 371-376.

Rousselle, C., Clair, P., Temsamani, J., and Scherrmann, J.M. (2002). Improved brain delivery of benzylpenicillin with a peptide-vector-mediated strategy. J Drug Target 10, 309- 315.

Ruan, H., Chai, Z., Shen, Q., Chen, X., Su, B., Xie, C., Zhan, C., Yao, S., Wang, H., Zhang, M., Ying, M., and Lu, W. (2018). A novel peptide ligand RAP12 of LRP1 for glioma targeted drug delivery. J Control Release 279, 306-315.

Ryan, P.J., Buchler, E., Shabanpoor, F., Hossain, M.A., Wade, J.D., Lawrence, A.J., and Gundlach, A.L. (2013a). Central relaxin-3 receptor (RXFP3) activation decreases anxiety- and depressive-like behaviours in the rat. Behav Brain Res 244, 142-151.

Ryan, P.J., Kastman, H.E., Krstew, E.V., Rosengren, K.J., Hossain, M.A., Churilov, L., Wade, J.D., Gundlach, A.L., and Lawrence, A.J. (2013b). Relaxin-3/RXFP3 system regulates alcohol-seeking. Proceedings of the National Academy of Sciences of the United States of America 110, 20789-20794.

  Ryan, P.J., Krstew, E.V., Sarwar, M., Gundlach, A.L., and Lawrence, A.J. (2014). Relaxin-3 mRNA levels in nucleus incertus correlate with alcohol and sucrose intake in rats. Drug Alcohol Depend 140, 8-16.

Sakagami, K., Masuda, T., Kawano, K., and Futaki, S. (2018). Importance of Net Hydrophobicity in the Cellular Uptake of All-Hydrocarbon Stapled Peptides. Mol Pharm 15, 1332-1340.

Sakamoto, K., Shinohara, T., Adachi, Y., Asami, T., and Ohtaki, T. (2017). A novel LRP1- binding peptide L57 that crosses the blood brain barrier. Biochem Biophys Rep 12, 135- 139.

Samuel, C.S., Hewitson, T.D., Unemori, E.N., and Tang, M.L. (2007). Drugs of the future: the hormone relaxin. Cell Mol Life Sci 64, 1539-1557.

Sarko, D., Beijer, B., Garcia Boy, R., Nothelfer, E.M., Leotta, K., Eisenhut, M., Altmann, A., Haberkorn, U., and Mier, W. (2010). The pharmacokinetics of cell-penetrating peptides. Mol Pharm 7, 2224-2231.

Schroeder, C.I., and Rosengren, K.J. (2020). Three-Dimensional Structure Determination of Peptides Using Solution Nuclear Magnetic Resonance Spectroscopy. Methods Mol Biol 2068, 129-162.

Schwabe, C., and Mcdonald, J.K. (1977). Relaxin: a disulfide homolog of insulin. Science 197, 914-915.

Shabanpoor, F., Akhter Hossain, M., Ryan, P.J., Belgi, A., Layfield, S., Kocan, M., Zhang, S., Samuel, C.S., Gundlach, A.L., Bathgate, R.A., Separovic, F., and Wade, J.D. (2012). Minimization of human relaxin-3 leading to high-affinity analogues with increased selectivity for relaxin-family peptide 3 receptor (RXFP3) over RXFP1. J Med Chem 55, 1671-1681.

Shabanpoor, F., Bathgate, R.A., Wade, J.D., and Hossain, M.A. (2013). C-terminus of the B- chain of relaxin-3 is important for receptor activity. PLoS One 8, e82567.

Shabanpoor, F., Hughes, R.A., Zhang, S., Bathgate, R.A., Layfield, S., Hossain, M.A., Tregear, G.W., Separovic, F., and Wade, J.D. (2010). Effect of helix-promoting strategies on the biological activity of novel analogues of the B-chain of INSL3. Amino Acids 38, 121- 131.

Shabanpoor, F., Separovic, F., and Wade, J.D. (2009). The Human Insulin Superfamily of Polypeptide Hormones. Vitamins and Hormones Insulin and Igfs 80, 1-31.

  Shao, K., Huang, R., Li, J., Han, L., Ye, L., Lou, J., and Jiang, C. (2010). Angiopep-2 modified PE-PEG based polymeric micelles for amphotericin B delivery targeted to the brain. J Control Release 147, 118-126.

Shao, K., Wu, J., Chen, Z., Huang, S., Li, J., Ye, L., Lou, J., Zhu, L., and Jiang, C. (2012). A brain-vectored angiopep-2 based polymeric micelles for the treatment of intracranial fungal infection. Biomaterials 33, 6898-6907.

Sharma, G., Modgil, A., Zhong, T., Sun, C., and Singh, J. (2014). Influence of short-chain cell- penetrating peptides on transport of doxorubicin encapsulating receptor-targeted liposomes across brain endothelial barrier. Pharm Res 31, 1194-1209.

Shen, J., Yu, M., Meng, Q., Li, J., Lv, Y., and Lu, W. (2013). Fatty acid-based strategy for efficient brain targeted gene delivery. Pharm Res 30, 2573-2583.

Shen, Y., and Bax, A. (2013). Protein backbone and sidechain torsion angles predicted from NMR chemical shifts using artificial neural networks. Journal of Biomolecular Nmr 56, 227-241.

Silhavy, T.J., Kahne, D., and Walker, S. (2010). The bacterial cell envelope. Cold Spring Harb Perspect Biol 2, a000414.

Sisley, S., Gutierrez-Aguilar, R., Scott, M., D'alessio, D.A., Sandoval, D.A., and Seeley, R.J. (2014). Neuronal GLP1R mediates liraglutide's anorectic but not glucose-lowering effect. J Clin Invest 124, 2456-2463.

Skrlj, N., Drevensek, G., Hudoklin, S., Romih, R., Curin Serbec, V., and Dolinar, M. (2013). Recombinant single-chain antibody with the Trojan peptide penetratin positioned in the linker region enables cargo transfer across the blood-brain barrier. Appl Biochem Biotechnol 169, 159-169.

Smith, C.M., Blasiak, A., Ganella, D.E., Chua, B.E., Layfield, S.L., Bathgate, R.A., and Gundlach, A.L. (2013a). Viral-mediated delivery of an RXFP3 agonist into brain promotes arousal in mice. Ital J Anat Embryol 118, 42-46.

Smith, C.M., Chua, B.E., Zhang, C., Walker, A.W., Haidar, M., Hawkes, D., Shabanpoor, F., Hossain, M.A., Wade, J.D., Rosengren, K.J., and Gundlach, A.L. (2014a). Central injection of relaxin-3 receptor (RXFP3) antagonist peptides reduces motivated food seeking and consumption in C57BL/6J mice. Behavioural Brain Research 268, 117- 126.

  Smith, C.M., Hosken, I.T., Downer, N.L., Chua, B.E., Hossain, M.A., Wade, J.D., and Gundlach, A.L. (2013b). Pharmacological activation of RXFP3 is not orexigenic in C57BL/6J mice. Ital J Anat Embryol 118, 52-55.

Smith, C.M., Hosken, I.T., Sutton, S.W., Lawrence, A.J., and Gundlach, A.L. (2012). Relaxin- 3 null mutation mice display a circadian hypoactivity phenotype. Genes Brain Behav 11, 94-104.

Smith, C.M., Shen, P.J., Banerjee, A., Bonaventure, P., Ma, S., Bathgate, R.A., Sutton, S.W., and Gundlach, A.L. (2010). Distribution of relaxin-3 and RXFP3 within arousal, stress, affective, and cognitive circuits of mouse brain. J Comp Neurol 518, 4016-4045.

Smith, C.M., Walker, A.W., Hosken, I.T., Chua, B.E., Zhang, C., Haidar, M., and Gundlach, A.L. (2014b). Relaxin-3/RXFP3 networks: an emerging target for the treatment of depression and other neuro psychiatric diseases? Frontiers in Pharmacology 5.

Smith, C.M., Walker, L.L., Chua, B.E., Mckinley, M.J., Gundlach, A.L., Denton, D.A., and Lawrence, A.J. (2015). Involvement of central relaxin-3 signalling in sodium (salt) appetite. Exp Physiol 100, 1064-1072.

Spicer, C.D., Jumeaux, C., Gupta, B., and Stevens, M.M. (2018). Peptide and protein nanoparticle conjugates: versatile platforms for biomedical applications. Chem Soc Rev 47, 3574-3620.

Staquicini, F.I., Ozawa, M.G., Moya, C.A., Driessen, W.H., Barbu, E.M., Nishimori, H., Soghomonyan, S., Flores, L.G., 2nd, Liang, X., Paolillo, V., Alauddin, M.M., Basilion, J.P., Furnari, F.B., Bogler, O., Lang, F.F., Aldape, K.D., Fuller, G.N., Hook, M., Gelovani, J.G., Sidman, R.L., Cavenee, W.K., Pasqualini, R., and Arap, W. (2011). Systemic combinatorial peptide selection yields a non-canonical iron-mimicry mechanism for targeting tumors in a mouse model of human glioblastoma. J Clin Invest 121, 161-173.

Sudo, S., Kumagai, J., Nishi, S., Layfield, S., Ferraro, T., Bathgate, R.A., and Hsueh, A.J. (2003). H3 relaxin is a specific ligand for LGR7 and activates the receptor by interacting with both the ectodomain and the exoloop 2. J Biol Chem 278, 7855-7862.

Sutton, S.W., Bonaventure, P., Kuei, C., Roland, B., Chen, J., Nepomuceno, D., Lovenberg, T.W., and Liu, C. (2004). Distribution of G-protein-coupled receptor (GPCR)135 binding sites and receptor mRNA in the rat brain suggests a role for relaxin-3 in neuroendocrine and sensory processing. Neuroendocrinology 80, 298-307.

Sutton, S.W., Shelton, J., Smith, C., Williams, J., Yun, S., Motley, T., Kuei, C., Bonaventure, P., Gundlach, A., Liu, C., and Lovenberg, T. (2009). Metabolic and neuroendocrine

  responses to RXFP3 modulation in the central nervous system. Ann N Y Acad Sci 1160, 242-249.

Swedberg, J.E., Schroeder, C.I., Mitchell, J.M., Fairlie, D.P., Edmonds, D.J., Griffith, D.A., Ruggeri, R.B., Derksen, D.R., Loria, P.M., Price, D.A., Liras, S., and Craik, D.J. (2016). Truncated glucagon-like peptide-1 and exendin-4 alpha-conotoxin pl14a peptide chimeras maintain potency and alpha-helicity and reveal interactions vital for cAMP signaling in vitro. J Biol Chem 291, 15778-15787.

Tanaka, M., Iijima, N., Miyamoto, Y., Fukusumi, S., Itoh, Y., Ozawa, H., and Ibata, Y. (2005). Neurons expressing relaxin 3/INSL 7 in the nucleus incertus respond to stress. Eur J Neurosci 21, 1659-1670.

Tang, S.C., Kumthekar, P., Brenner, A.J., Kesari, S., Piccioni, D., Anders, C.K., Carillo, J.A., Chalasani, P., Kabos, P., Puhalla, S.L., Garcia, A., Tkaczuk, K., Ahluwalia, M.S., Lakhani, N., and Ibrahim, N. (2016). ANG1005, a novel peptide-paclitaxel conjugate crosses the BBB and shows activity in patients with recurrent CNS metastasis from breast cancer, results from a phase II clinical study. Annals of Oncology 27.

Thomas, F.C., Taskar, K., Rudraraju, V., Goda, S., Thorsheim, H.R., Gaasch, J.A., Mittapalli, R.K., Palmieri, D., Steeg, P.S., Lockman, P.R., and Smith, Q.R. (2009). Uptake of ANG1005, A Novel Paclitaxel Derivative, Through the Blood-Brain Barrier into Brain and Experimental Brain Metastases of Breast Cancer. Pharmaceutical Research 26, 2486-2494.

Tian, X.H., Wang, Z.G., Meng, H., Wang, Y.H., Feng, W., Wei, F., Huang, Z.C., Lin, X.N., and Ren, L. (2013). Tat peptide-decorated gelatin-siloxane nanoparticles for delivery of CGRP transgene in treatment of cerebral vasospasm. Int J Nanomedicine 8, 865-876.

Tian, Y., Jiang, Y., Li, J., Wang, D., Zhao, H., and Li, Z. (2017). Effect of Stapling Architecture on Physiochemical Properties and Cell Permeability of Stapled alpha-Helical Peptides: A Comparative Study. Chembiochem 18, 2087-2093.

Turner, E.C., Cureton, C.H., Weston, C.J., Smart, O.S., and Allemann, R.K. (2004). Controlling the DNA binding specificity of bHLH proteins through intramolecular interactions. Chem Biol 11, 69-77.

Uchida, Y., Ohtsuki, S., Katsukura, Y., Ikeda, C., Suzuki, T., Kamiie, J., and Terasaki, T. (2011). Quantitative targeted absolute proteomics of human blood-brain barrier transporters and receptors. J Neurochem 117, 333-345.

  Urich, E., Lazic, S.E., Molnos, J., Wells, I., and Freskgard, P.O. (2012). Transcriptional profiling of human brain endothelial cells reveals key properties crucial for predictive in vitro blood-brain barrier models. PLoS One 7, e38149.

Urich, E., Schmucki, R., Ruderisch, N., Kitas, E., Certa, U., Jacobsen, H., Schweitzer, C., Bergadano, A., Ebeling, M., Loetscher, H., and Freskgard, P.O. (2015). Cargo Delivery into the Brain by in vivo identified Transport Peptides. Sci Rep 5, 14104.

Van Der Westhuizen, E.T., Sexton, P.M., Bathgate, R.A., and Summers, R.J. (2005). Responses of GPCR135 to human gene 3 (H3) relaxin in CHO-K1 cells determined by microphysiometry. Ann N Y Acad Sci 1041, 332-337.

Van Meer, G., and De Kroon, A.I. (2011). Lipid map of the mammalian cell. J Cell Sci 124, 5- 8.

Van Rooy, I., Cakir-Tascioglu, S., Couraud, P.O., Romero, I.A., Weksler, B., Storm, G., Hennink, W.E., Schiffelers, R.M., and Mastrobattista, E. (2010). Identification of peptide ligands for targeting to the blood-brain barrier. Pharm Res 27, 673-682.

Van Rooy, I., Mastrobattista, E., Storm, G., Hennink, W.E., and Schiffelers, R.M. (2011). Comparison of five different targeting ligands to enhance accumulation of liposomes into the brain. J Control Release 150, 30-36.

Velasco-Aguirre, C., Morales-Zavala, F., Salas-Huenuleo, E., Gallardo-Toledo, E., Andonie, O., Munoz, L., Rojas, X., Acosta, G., Sanchez-Navarro, M., Giralt, E., Araya, E., Albericio, F., and Kogan, M.J. (2017). Improving gold nanorod delivery to the central nervous system by conjugation to the shuttle Angiopep-2. Nanomedicine (Lond).

Vives, E., Brodin, P., and Lebleu, B. (1997). A truncated HIV-1 Tat protein basic domain rapidly translocates through the plasma membrane and accumulates in the cell nucleus. J Biol Chem 272, 16010-16017.

Vlieghe, P., and Khrestchatisky, M. (2010). Peptide-based vectors for blood-brain barrier targeting and delivery of drugs to the central nervous system. Ther Deliv 1, 489-494.

Volkman, B.F., and Wemmer, D.E. (1997). Deletion of a single amino acid changes the folding of an apamin hybrid sequence peptide to that of . Biopolymers 41, 451-460.

Walker, A.W., Smith, C.M., Chua, B.E., Krstew, E.V., Zhang, C., Gundlach, A.L., and Lawrence, A.J. (2015). Relaxin-3 receptor (RXFP3) signalling mediates stress-related alcohol preference in mice. PLoS One 10, e0122504.

  Walker, L.C., Kastman, H.E., Koeleman, J.A., Smith, C.M., Perry, C.J., Krstew, E.V., Gundlach, A.L., and Lawrence, A.J. (2017a). Nucleus incertus corticotrophin-releasing factor 1 receptor signalling regulates alcohol seeking in rats. Addict Biol 22, 1641-1654.

Walker, L.C., Kastman, H.E., Krstew, E.V., Gundlach, A.L., and Lawrence, A.J. (2017b). Central amygdala relaxin-3/relaxin family peptide receptor 3 signalling modulates alcohol seeking in rats. Br J Pharmacol 174, 3359-3369.

Wang, C.K., and Craik, D.J. (2018). Designing macrocyclic disulfide-rich peptides for biotechnological applications. Nat Chem Biol 14, 417-427.

Wang, C.K., Gruber, C.W., Cemazar, M., Siatskas, C., Tagore, P., Payne, N., Sun, G., Wang, S., Bernard, C.C., and Craik, D.J. (2014). Molecular grafting onto a stable framework yields novel cyclic peptides for the treatment of multiple sclerosis. ACS Chem Biol 9, 156-163.

Wang, H., Xu, K., Liu, L., Tan, J.P., Chen, Y., Li, Y., Fan, W., Wei, Z., Sheng, J., Yang, Y.Y., and Li, L. (2010). The efficacy of self-assembled cationic antimicrobial peptide nanoparticles against Cryptococcus neoformans for the treatment of meningitis. Biomaterials 31, 2874-2881.

Wang, J., Hogenkamp, D.J., Tran, M., Li, W.Y., Yoshimura, R.F., Johnstone, T.B., Shen, W.C., and Gee, K.W. (2006). Reversible lipidization for the oral delivery of leu- enkephalin. J Drug Target 14, 127-136.

Wei, L., Guo, X.Y., Yang, T., Yu, M.Z., Chen, D.W., and Wang, J.C. (2016). Brain tumor- targeted therapy by systemic delivery of siRNA with Transferrin receptor-mediated core-shell nanoparticles. Int J Pharm 510, 394-405.

Wei, X., Zhan, C., Chen, X., Hou, J., Xie, C., and Lu, W. (2014). Retro-inverso isomer of Angiopep-2: a stable d-peptide ligand inspires brain-targeted drug delivery. Mol Pharm 11, 3261-3268.

Weksler, B., Romero, I.A., and Couraud, P.O. (2013). The hCMEC/D3 cell line as a model of the human blood brain barrier. Fluids Barriers CNS 10, 16.

Weksler, B.B., Subileau, E.A., Perriere, N., Charneau, P., Holloway, K., Leveque, M., Tricoire- Leignel, H., Nicotra, A., Bourdoulous, S., Turowski, P., Male, D.K., Roux, F., Greenwood, J., Romero, I.A., and Couraud, P.O. (2005). Blood-brain barrier-specific properties of a human adult brain endothelial cell line. FASEB J 19, 1872-1874.

  Werner, H.M., Cabalteja, C.C., and Horne, W.S. (2016). Peptide Backbone Composition and Protease Susceptibility: Impact of Modification Type, Position, and Tandem Substitution. Chembiochem 17, 712-718.

Weston, C.J., Cureton, C.H., Calvert, M.J., Smart, O.S., and Allemann, R.K. (2004). A stable miniature protein with oxaloacetate decarboxylase activity. Chembiochem 5, 1075- 1080.

Wilkinson, T.N., Speed, T.P., Tregear, G.W., and Bathgate, R.A. (2005). Evolution of the relaxin-like peptide family. BMC Evol Biol 5, 14.

Wishart, D.S., Bigam, C.G., Holm, A., Hodges, R.S., and Sykes, B.D. (1995). 1H, 13C and 15N random coil NMR chemical shifts of the common amino acids. I. Investigations of nearest-neighbor effects. J Biomol NMR 5, 67-81.

Wong, K.H., Riaz, M.K., Xie, Y., Zhang, X., Liu, Q., Chen, H., Bian, Z., Chen, X., Lu, A., and Yang, Z. (2019). Review of Current Strategies for Delivering Alzheimer's Disease Drugs across the Blood-Brain Barrier. Int J Mol Sci 20.

Wong, L.L.L., Scott, D.J., Hossain, M.A., Kaas, Q., Rosengren, K.J., and Bathgate, R.a.D. (2018a). Distinct but overlapping binding sites of agonist and antagonist at the relaxin family peptide 3 (RXFP3) receptor. Journal of Biological Chemistry 293, 15777- 15789.

Wong, L.L.L., Scott, D.J., Hossain, M.A., Kaas, Q., Rosengren, K.J., and Bathgate, R.a.D. (2018b). Distinct but overlapping binding sites of agonist and antagonist at the relaxin family peptide 3 (RXFP3) receptor. J Biol Chem 293, 15777-15789.

Wu, J., Jiang, H., Bi, Q.Y., Luo, Q.S., Li, J.J., Zhang, Y., Chen, Z.B., and Li, C. (2014). Apamin-mediated actively targeted drug delivery for treatment of spinal cord injury: more than just a concept. Molecular Pharmaceutics 11, 3210-3222.

WüThrich, K. (1986). NMR of proteins and nucleic acids. New York: John Wiley & Sons.

Xin, H., Jiang, X., Gu, J., Sha, X., Chen, L., Law, K., Chen, Y., Wang, X., Jiang, Y., and Fang, X. (2011). Angiopep-conjugated poly(ethylene glycol)-co-poly(epsilon-caprolactone) nanoparticles as dual-targeting drug delivery system for brain glioma. Biomaterials 32, 4293-4305.

Yoo, J., Lee, D., Gujrati, V., Rejinold, N.S., Lekshmi, K.M., Uthaman, S., Jeong, C., Park, I.K., Jon, S., and Kim, Y.C. (2017). Bioreducible branched poly(modified nona- arginine) cell-penetrating peptide as a novel gene delivery platform. J Control Release 246, 142-154.

  Youn, P., Chen, Y., and Furgeson, D.Y. (2014). A myristoylated cell-penetrating peptide bearing a transferrin receptor-targeting sequence for neuro-targeted siRNA delivery. Mol Pharm 11, 486-495.

Yu, Y.J., Atwal, J.K., Zhang, Y., Tong, R.K., Wildsmith, K.R., Tan, C., Bien-Ly, N., Hersom, M., Maloney, J.A., Meilandt, W.J., Bumbaca, D., Gadkar, K., Hoyte, K., Luk, W., Lu, Y., Ernst, J.A., Scearce-Levie, K., Couch, J.A., Dennis, M.S., and Watts, R.J. (2014). Therapeutic bispecific antibodies cross the blood-brain barrier in nonhuman primates. Sci Transl Med 6, 261ra154.

Yu, Y.J., Zhang, Y., Kenrick, M., Hoyte, K., Luk, W., Lu, Y., Atwal, J., Elliott, J.M., Prabhu, S., Watts, R.J., and Dennis, M.S. (2011). Boosting brain uptake of a therapeutic antibody by reducing its affinity for a transcytosis target. Sci Transl Med 3, 84ra44.

Yuen, T.Y., Brown, C.J., Tan, Y.S., and Johannes, C.W. (2020). Synthesis of Chiral Alkenyl Cyclopropane Amino Acids for Incorporation into Stapled Peptides. J Org Chem 85, 1556-1566.

Zhang, B., Sun, X., Mei, H., Wang, Y., Liao, Z., Chen, J., Zhang, Q., Hu, Y., Pang, Z., and Jiang, X. (2013). LDLR-mediated peptide-22-conjugated nanoparticles for dual- targeting therapy of brain glioma. Biomaterials 34, 9171-9182.

Zhang, C., Chua, B.E., Yang, A., Shabanpoor, F., Hossain, M.A., Wade, J.D., Rosengren, K.J., Smith, C.M., and Gundlach, A.L. (2015a). Central relaxin-3 receptor (RXFP3) activation reduces elevated, but not basal, anxiety-like behaviour in C57BL/6J mice. Behav Brain Res 292, 125-132.

Zhang, C., Wan, X., Zheng, X., Shao, X., Liu, Q., Zhang, Q., and Qian, Y. (2014a). Dual- functional nanoparticles targeting amyloid plaques in the brains of Alzheimer's disease mice. Biomaterials 35, 456-465.

Zhang, D.D., Wang, J.X., and Xu, D.G. (2016). Cell-penetrating peptides as noninvasive transmembrane vectors for the development of novel multifunctional drug-delivery systems. Journal of Controlled Release 229, 130-139.

Zhang, G., Annan, R.S., Carr, S.A., and Neubert, T.A. (2014b). Overview of peptide and protein analysis by mass spectrometry. Curr Protoc Mol Biol 108, 10 21 11-30.

Zhang, J., Qi, Y.F., Geng, B., Pan, C.S., Zhao, J., Chen, L., Yang, J., Chang, J.K., and Tang, C.S. (2005). Effect of relaxin on myocardial ischemia injury induced by isoproterenol. Peptides 26, 1632-1639.

  Zhang, L., and Bulaj, G. (2012). Converting Peptides into Drug Leads by Lipidation. Current Medicinal Chemistry 19, 1602-1618.

Zhang, L., Lee, H.K., Pruess, T.H., White, H.S., and Bulaj, G. (2009a). Synthesis and applications of polyamine amino acid residues: improving the bioactivity of an analgesic neuropeptide, neurotensin. J Med Chem 52, 1514-1517.

Zhang, L., Robertson, C.R., Green, B.R., Pruess, T.H., White, H.S., and Bulaj, G. (2009b). Structural requirements for a lipoamino acid in modulating the anticonvulsant activities of systemically active galanin analogues. J Med Chem 52, 1310-1316.

Zhang, Q., Yang, M., Sorensen, K.K., Madsen, C.S., Boesen, J.T., An, Y., Peng, S.H., Wei, Y., Wang, Q., Jensen, K.J., Zuo, Z., Chan, H.Y.E., and Ngo, J.C.K. (2017). A brain- targeting lipidated peptide for neutralizing RNA-mediated toxicity in Polyglutamine Diseases. Sci Rep 7, 12077.

Zhang, W.J., Wang, X.Y., Guo, Y.Q., Luo, X., Gao, X.J., Shao, X.X., Liu, Y.L., Xu, Z.G., and Guo, Z.Y. (2014c). The highly conserved negatively charged Glu141 and Asp145 of the G-protein-coupled receptor RXFP3 interact with the highly conserved positively charged arginine residues of relaxin-3. Amino Acids 46, 1393-1402.

Zhang, X., He, T., Chai, Z., Samulski, R.J., and Li, C. (2018). Blood-brain barrier shuttle peptides enhance AAV transduction in the brain after systemic administration. Biomaterials 176, 71-83.

Zhang, Y., Deng, C., Liu, S., Wu, J., Chen, Z., Li, C., and Lu, W. (2015b). Active targeting of tumors through conformational epitope imprinting. Angew Chem Int Ed Engl 54, 5157- 5160.

Zhou, Q.H., Sumbria, R., Hui, E.K., Lu, J.Z., Boado, R.J., and Pardridge, W.M. (2011). Neuroprotection with a brain-penetrating biologic tumor necrosis factor inhibitor. J Pharmacol Exp Ther 339, 618-623.

Zorzi, A., Middendorp, S.J., Wilbs, J., Deyle, K., and Heinis, C. (2017). Acylated heptapeptide binds albumin with high affinity and application as tag furnishes long-acting peptides. Nat Commun 8, 16092.



 

Appendices

 

Figure S-1. Competition binding of Arg23 agonist variants at RXFP3. The effect of reintroducing the C-terminal activation domain into R3 B1-22R was tested to see if the Arg23 modification could also improve affinity of an agonist. Variants were tested for their ability to compete for binding with europium labelled R3/I5. Data are presented as mean ± SEM from a minimum of three independent experiments. Both variants showed binding equivalent to the unmodified B-chain. Addition of a single Ala residue to R3 B1-22R also reduced affinity to closer to the unmodified B-chain.

  Table S-1. Eu-H3/I5 competition binding of R3 B1-22R C-terminal extension variants.   $#  !%  +  relaxin-3 RAAPYGVRLCGREFIRAVIFTCGGSRW 7.73 ± 0.04 (3) relaxin-3 B-chain RAAPYGVRLSGREFIRAVIFTSGGSRW 5.60 ± 0.09a,b (3) R3 B1-22R RAAPYGVRLSGREFIRAVIFTSR* 7.58 ± 0.11 (3) R3 B1-22RA RAAPYGVRLSGREFIRAVIFTSRA* 6.13 ± 0.04a (3) R3 B1-22RGSRW RAAPYGVRLSGREFIRAVIFTSRGSRW* 5.68 ± 0.12a,b (3) R3 B1-22RGGSRW RAAPYGVRLSGREFIRAVIFTSRGGSRW* 5.59 ± 0.08a,b (4) *( $#)'$ $$" %#   &"#%#    &"#%#  

   Table S-2. Physiochemical properties of point mutated R3 B1-22R analogues.     % "  "  !      !   R3 B1-22R RAAPYGVRLSGREFIRAVIFTSR 2622.04 2620.20 12.28  R3 B1-22R Y5A RAAPAGVRLSGREFIRAVIFTSR 2529.94 2529.40 12.70  R3 B1-22R G6A RAAPYAVRLSGREFIRAVIFTSR 2636.07 2635.40 12.28  R3 B1-22R V7A RAAPYGARLSGREFIRAVIFTSR 2593.99 2593.40 12.28  R3 B1-22R R8A RAAPYGVALSGREFIRAVIFTSR 2536.93 2536.60 12.10  R3 B1-22R L9A RAAPYGVRASGREFIRAVIFTSR 2579.96 2579.00 12.28  R3 B1-22R S10A RAAPYGVRLAGREFIRAVIFTSR 2606.04 2605.40 12.28  R3 B1-22R G11A RAAPYGVRLSAREFIRAVIFTSR 2636.07 2635.20 12.28  R3 B1-22R R12A RAAPYGVRLSGAEFIRAVIFTSR 2536.93 2538.32 12.10  R3 B1-22R E13A RAAPYGVRLSGRAFIRAVIFTSR 2564.00 2564.00 12.70  R3 B1-22R F14A RAAPYGVRLSGREAIRAVIFTSR 2545.94 2545.40 12.28  R3 B1-22R I15A RAAPYGVRLSGREFARAVIFTSR 2579.96 2580.84 12.28  R3 B1-22R R16A RAAPYGVRLSGREFIAAVIFTSR 2536.93 2537.92 12.10  R3 B1-22R A17K RAAPYGVRLSGREFIRKVIFTSR 2679.13 2678.40 12.29  R3 B1-22R A17N RAAPYGVRLSGREFIRNVIFTSR 2665.06 2664.80 12.28  R3 B1-22R V18A RAAPYGVRLSGREFIRAAIFTSR 2593.99 2593.40 12.28  R3 B1-22R I19A RAAPYGVRLSGREFIRAVAFTSR 2579.96 2581.16 12.28  R3 B1-22R F20A RAAPYGVRLSGREFIRAVIATSR 2545.94 2545.00 12.28  R3 B1-22R T21A RAAPYGVRLSGREFIRAVIFASR 2592.01 2591.40 12.28  R3 B1-22R S22A RAAPYGVRLSGREFIRAVIFTAR 2606.04 2605.60 12.28  R3 B1-22R R23A RAAPYGVRLSGREFIRAVIFTSA 2536.93 2539.75 12.10   R3 B1-22R V7Abu RAAPYG(Abu)RLSGREFIRAVIFTSR 2608.01 2607.80 12.28  R3 B1-22R R8Abu RAAPYGV(Abu)LSGREFIRAVIFTSR 2550.96 2550.20 12.10  R3 B1-22R R12Abu RAAPYGVRLSG(Abu)EFIRAVIFTSR 2550.96 2550.80 12.10  R3 B1-22R F14Abu RAAPYGVRLSGRE(Abu)IRAVIFTSR 2559.97 2558.80 12.28  R3 B1-22R I15Abu RAAPYGVRLSGREF(Abu)RAVIFTSR 2593.99 2593.40 12.28  R3 B1-22R R16Abu RAAPYGVRLSGREFI(Abu)AVIFTSR 2550.96 2550.60 12.10  R3 B1-22R A17Abu RAAPYGVRLSGREFIR(Abu)VIFTSR 2636.07 2636.00 12.28  R3 B1-22R V18Abu RAAPYGVRLSGREFIRA(Abu)IFTSR 2608.01 2607.20 12.28  R3 B1-22R I19Abu RAAPYGVRLSGREFIRAV(Abu)FTSR 2593.99 2593.40 12.28  R3 B1-22R T21Abu RAAPYGVRLSGREFIRAVIF(Abu)SR 2606.04 2606.00 12.28  R3 B1-22R R23Orn RAAPYGVRLSGREFIRAVIFTS(Orn) 2580.00 2580.00 12.10  R3 B1-22R R23Har RAAPYGVRLSGREFIRAVIFTS(Har) 2636.08 2635.20 12.28  R3 B1-22R R23K RAAPYGVRLSGREFIRAVIFTSK 2594.03 2592.80 12.12  R3 B1-22R R23Cit RAAPYGVRLSGREFIRAVIFTS(Cit) 2623.02 2622.80 12.10  R3 B1-22R R23Agb RAAPYGVRLSGREFIRAVIFTS(Agb) 2608.03 2607.00 12.28 %!#$! 

  Table S-3. Physiochemical properties of analogues of R3 B1-22R with helix supportive or disruptive modifications.    !  "( %!"  % ! !  $!  " $! !   ! #  R3 B6-22R retro inverso rstfivarifergslrvg 2063.41 2063.00 12.58  "!!  R3 B1-22R L9P RAAPYGVRPSGREFIRAVIFTSR 2606.00 2605.80 12.28  R3 B1-22R E13P RAAPYGVRLSGRPFIRAVIFTSR 2590.04 2589.40 12.70  R3 B1-22R A17P RAAPYGVRLSGREFIRPVIFTSR 2648.08 2647.80 12.28  R3 B1-22R F20P RAAPYGVRLSGREFIRAVIPTSR 2571.98 2570.20 12.28 # !  R3 B1-22R A17Aib RAAPYGVRLSGREFIR(Aib)VIFTSR 2636.07 2634.40 12.28  R3 B1-22R V18Aib RAAPYGVRLSGREFIRA(Aib)IFTSR 2608.01 2608.00 12.28  R3 B1-22R T21Aib RAAPYGVRLSGREFIRAVIF(Aib)SR 2606.04 2605.80 12.28  R3 B1-22R F14/V18/T21Aib RAAPYGVRLSGRE(Aib)IRA(Aib)IF(Aib)SR 2529.94 2529.60 12.28  R3 B6-22R GVRLSGREFIRAVIFTSR 2063.41 2063.20 12.58  R3 B6-22R F20/T21/S22Aib GVRLSGREFIRAVI(Aib,Aib,Aib)R 1983.37 1983.00 12.58 !  R3 B1-22R V18E S22K RAAPYGVRLSGREFIRAEIFTKR 2693.12 2692.60 11.95  R3 B1-22R V18E S22K Lactam RAAPYGVRLSGREFIRAEIFTKR 2675.12 2675.00 12.28  R3 B1-22R R16E F20K RAAPYGVRLSGREFIEAVIKTSR 2575.97 2576.00 11.69  R3 B1-22R R16E F20K Lactam RAAPYGVRLSGREFIEAVIKTSR 2557.97 2559.00 12.10  R3 B1-22R E13E A17K RAAPYGVRLSGREFIRKVIFTSR 2679.13 2678.80 12.29  R3 B1-22R E13E A17K Lactam RAAPYGVRLSGREFIRKVIFTSR 2661.13 2660.40 12.70  R3 B1-22R E13K A17D RAAPYGVRLSGRKFIRDVIFTSR 2665.11 2664.00 12.29  R3 B1-22R E13K A17D Lactam RAAPYGVRLSGRKFIRDVIFTSR 2647.11 2646.20 12.70 & R3 B1-22R RAAPYGVRLSGREFIRAVIFTSR 2622.04 2620.20 12.28 # !  R3 B1-22R cyclic RAAPYGVRLSGREFIRAVIFTSR 2604.04 2604.00 12.70  R3 B6-22R GVRLSGREFIRAVIFTSR 2063.41 2063.20 12.58  R3 B6-22R cyclic GVRLSGREFIRAVIFTSR 2045.41 2044.80 14.00  R3 B1-22R S10A E13Q T21E RAAPYGVRLAGRQFIRAVIFESR 2633.07 2633.20 12.28  R3 B1-22R S10A E13Q T21E cyclic RAAPYGVRLAGRQFIRAVIFESR 2615.07 2614.00 12.70  R3 B6-22R E13Q T21E GVRLSGRQFIRAVIFESR 2090.44 2090.60 12.58

 R3 B6-22R E13Q T21E cyclic GVRLSGRQFIRAVIFESR 2072.44 2071.80 14.00 (! $  &! '$!! ! " 

   Figure S-2. MS spectra and corresponding analytical HPLC traces of oxidised and purified analogues 3 – 6.

   Figure S-3. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 7, 9 and 10.

   Figure S-4. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 11 - 16. 

   Figure S-5. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 16 - 22. 

   Figure S-6. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 23 - 27.  

    Figure S-7. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 3. The sequential walk highlighting the NMR spin system assignments are shown by connecting lines. Hα-HN cross peaks are labelled with residue numbers. 

  Figure S-8. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 4. The sequential walk highlighting the NMR spin system assignments are shown by connecting lines. Hα-HN cross peaks are labelled with residue numbers.

    Figure S-9. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 5. The sequential walk highlighting the NMR spin system assignments are shown by connecting lines. Hα-HN cross peaks are labelled with residue numbers. 

  Figure S-10. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 6. Poor dispersion and lack of NOEs are consistent with a lack or ordered structure.

    Figure S-11. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 7. Poor dispersion, lack of NOEs and severe line broadening are consistent with a lack of ordered structure and potential aggregation. 

  Figure S-12. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 9. The sequential walk highlighting the NMR spin system assignments are shown by connecting lines. Hα-HN cross peaks are labelled with residue numbers. 

    Figure S-13. NOESY spectrum recorded at 298K with a mixing time of 200 ms of analogue 10. The spectrum is similar to that of analogue 5, which highlights that both the agonist and antagonist variants grafted on VhTI adopt similar well-structured folds 

  Figure S-14. Comparison of secondary Hα chemical shifts for analogue 3 and analogue 4. The stretch of negative values from 15-24 correspond to the relaxin-3 B-chain helix. The more negative values in the later part of this helix show that the helical nature is increased by the incorporation of Aib residues. Residues numbers on the X-axis relate to the longest peptide sequence in the main manuscript, analogue 5.  

  Table S-4. Structural statistics from NMR based structure calculations.  Analogue 3 Analogue 9 Analogue 5 MolProbity statistics Clashes (> 0.4 Å/1000 atoms) 6.73 ± 3.61 5.40 ± 4.13 3.81 ± 2.54 Poor rotamers 0.25 ± 0.64 0.25 ± 0.55 1.45 ± 1.19 Ramachandran outliers (%) 1.43 ± 2.93 0.50 ± 2.24 0.31 ± 1.40 Ramachandran favoured (%) 82.14 ± 8.82 94.00 ± 6.81 90.00 ± 5.88 MolProbity score 2.10 ± 0.20 1.58 ± 0.22 2.08 ± 0.28 MolProbity score percentile 69.80 ± 10.53 91.60 ± 5.93 69.80 ± 14.05 Distance restraints Intraresidue (i-j = 0) 111 69 117 Sequential (/i-j/ = 1) 85 54 106 Medium range (/i-j/ ≤ 5) 45 13 88 Long range (/i-j/ > 5) 11 6 27 Hydrogen bonds 6 7 NA Total 258 149 338 Dihedral angle restraints φ (phi) 11 12 NA ψ (psi) 12 13 NA Total 23 25 NA Average pairwise r.m.s.d.a (Å) Backbone atoms 0.14 ± 0.07 0.80 ± 0.34 1.23 ± 0.40b Heavy atoms 1.18 ± 0.25 1.86 ± 0.40 2.37 ± 0.53 a Calculated from 20 refined structures over residues 1-15 b Calculated from 20 refined structures over residues 3-25 

  

Figure S-15. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 28 – 32.

  

Figure S-16. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 33 – 37.

  

Figure S-17. MS spectra and the corresponding analytical HPLC trace of purified final product of analogues 38 – 41.