Succinimide-Based Conjugates Improve IsoDGR Cyclopeptide Affinity

Subscriber access provided by Kaohsiung Medical University

Article

Succinimide-based conjugates improve isoDGR cyclopeptide affinity to ## without promoting integrin allosteric activation v

3

Francesca Nardelli, Cristina Paissoni, Giacomo Quilici, Alessandro Gori, Catia Traversari, Barbara Valentinis, Angelina Sacchi, Angelo Corti, Flavio Curnis, Michela Ghitti, and Giovanna Musco J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00745 • Publication Date (Web): 08 Jun 2018 Downloaded from http://pubs.acs.org on June 8, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Succinimide-based conjugates improve isoDGR cyclopeptide affinity to αvβ3 without promoting integrin allosteric activation. Francesca Nardelli,a‡ Cristina Paissoni,a,b‡ǂ Giacomo Quilici,a Alessandro Gori,c Catia Traversari,d Barbara Valentinis,d Angelina Sacchi,a Angelo Corti,a Flavio Curnis,a* Michela Ghitti,a§* and Giovanna Muscoa§* AUTHOR ADDRESS. a

IRCCS Ospedale San Raffaele, Via Olgettina 60, 20132 Milan (Italy).

b

Dipartimento di Chimica, Università degli Studi di Milano, Via Golgi 19, 20133 Milan (Italy).

c

Istituto di Chimica del Riconoscimento Molecolare, CNR. Via Mario Bianco 9, 20131, Milan

(Italy). d

Molmed, SpA, Via Olgettina Milano, 58, 20132 Milan (Italy).

ACS Paragon Plus Environment

1

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 38

ABSTRACT

The isoDGR sequence is an integrin-binding motif that has been successfully employed as tumor vasculature-homing molecule or for the targeted delivery of drugs and diagnostic agents to tumors. In this context we have previously demonstrated that the product of the conjugation of c(CGisoDGRG) (1) to 4-[N-maleimidomethyl] cyclohexane-1-carboxamide, (cyclopeptide 2) can be successfully used as a tumor-homing ligand for nanodrug delivery to neoplastic tissues. Here, combining NMR, computational and biochemical methods we show that the succinimide ring contained in 2 contributes to stabilizing interactions with αvβ3, an integrin overexpressed in the tumor vasculature. Furthermore, we demonstrate that various cyclopeptides containing the isoDGR sequence embedded in different molecular scaffolds, do not induce αvβ3 allosteric activation and work as pure integrin antagonists. These results could be profitably exploited for the rational design of novel isoDGR-based ligands and tumor targeting molecules with improved αvβ3-binding properties and devoid of adverse integrin activating effects.


2

Page 3 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


INTRODUCTION One of the main causes of systemic toxicity and adverse side effects of current anticancer drugs is their inability to differentiate between neoplastic and normal tissues. A promising strategy to tackle these shortcomings relies on a drug-delivery approach based on the conjugation of therapeutic compounds to molecules capable of recognizing receptors over-expressed in neoplastic tissues, such as tumor-homing antibodies or peptides.1,2 An important class of target receptors suitable for this purpose is represented by cell adhesion proteins abnormally expressed by tumor vessels, including integrins αvβ3, αvβ5 and α5β1, all involved in the regulation of angiogenesis and tumor growth.3–5 These integrins can be targeted by peptides containing the arginine-glycine-aspartate (RGD) sequence,6 a tripeptide motif also present in various cell adhesion molecules of the extracellular matrix.7 As a result, RGD peptides have been exploited for delivering a variety of different compounds to tumors, such as liposomes, nanoparticles, DNA complexes, viral particles, anti-angiogenic compounds, chemotherapeutic drugs, imaging compounds, cytokines and many others.6,8 A growing body of evidences suggests that also peptides containing the isoaspartate-glycine-arginine (isoDGR) sequence can work as efficient integrin ligands and as tumor-homing molecules.9–12 This motif was discovered in aged fibronectin, an extracellular-matrix protein, following spontaneous deamidation of asparagine at its asparagine-glycine-arginine (NGR) site,9 via cyclic imide formation, eventually resulting in DGR and isoDGR, in a typical ratio of 1:3.11 By combining NMR and computational methods we have previously shown that at variance to CDGRC, CisoDGRC, a peptide cyclized through a disulphide bridge, can accommodate into the canonical RGD binding site of αvβ3.10,13,14 Herein, the arginine and isoaspartate side chains anchor to αvβ3 headpiece recapitulating the classical electrostatic interactions at the basis of ligand/integrin recognition.10,13,14 However, despite their


3


Page 4 of 38

similarity, RGD and isoDGR peptides have different effects on integrin conformation upon binding. Indeed, RGD containing ligands can induce marked integrin conformational changes upon binding,15–18 causing receptor activation and potential detrimental agonist-like pharmacological effects.18–20 Conversely, the CisoDGRC peptide does not induce αvβ3 allosteric activation, thereby working as a pure integrin antagonist.13 This property may represent an important advantage for CisoDGRC application as a pure drug vehicle, in light of the numerous pharmacological and toxicological implications related to integrin activation.18 Whether this feature is a specific CisoDGRC peculiarity or is a general intrinsic property of the isoDGR sequence, independent from the molecular scaffolds and linkers used for coupling to effector molecules, is unknown. Conversely, previous studies have shown that the molecular environment in which the isoDGR motif is embedded can influence its affinity and selectivity towards different integrins.12,14,21–25 In this context, we have recently shown, that the chemical conjugation of the head-to-tail-cyclized hexapeptide c(CGisoDGRG) (Figure 1, cyclopeptide 1) to 4-[N-maleimidomethyl] cyclohexane-1-carboxamide (MCCA), the amidated derivative of the bifunctional crosslinking agent sulfosuccinimidyl 4-[N-maleimidomethyl] cyclohexane-1carboxylate (sulfo-SMCC), resulting in conjugates 2 (Figure 1, Figure S1), increases the affinity and selectivity of the peptide for αvβ3, conceivably through additional interactions via the chemical linker.12 Notably, conjugate 2 has been shown to work as an efficient ligand for delivering drugs, fluorescent nanoparticles, radioactive compounds and tumor necrosis factor-α (TNF)-loaded gold nanoparticles to tumors in animal models.12 These observations prompted us to investigate the structural determinants of the improved affinity of cyclopeptide 1 for αvβ3 after conjugation to MCCA and to assess whether this modification preserves the integrin antagonist activity of the isoDGR sequence or not.


4

Page 5 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Combining ligand based NMR experiments with computational techniques and biochemical methods, here we provide evidence that the succinimide ring of cyclopeptide conjugate 2 contributes to αvβ3 binding through additional polar interactions, without promoting integrin allosteric activation. Furthermore, using various cyclopeptides having the isoDGR sequence embedded in different molecular scaffolds, we support the notion that the lack of allosteric activation is a general property of the isoDGR motif, regardless of the molecular scaffolds and linkers used for compounds conjugation.

Figure 1. 2D representation of cyclopeptide 1 and of its conjugates 2, 3 and 4. Structures are shown with carbon, nitrogen, oxygen and sulfur atoms in black, blue, red and yellow, respectively.


5


Page 6 of 38

RESULTS AND DISCUSSION The chemical linker does not affect cyclopeptide 1 conformation We have previously shown in competitive binding assays that MCCA conjugation to the head-totail cyclopeptide 1 (i.e. conjugate 2) improves by five times the Ki towards αvβ3 as compared to the non-derivatized cyclopeptide, constituting a new promising scaffold for selective αvβ3 targeting (Table 1, Table S1).12 To understand the structural basis underlying the improved affinity, we first asked whether the chemical linker contained in conjugate 2 interacts with the peptide macrocycle and influences its structure, possibly promoting a favorable pre-organized binding conformation. To this end we acquired classical homonuclear 2D (NOESY, ROESY, TOCSY) and heteronuclear experiments (1H-13C HSQC, HMBC) on both 1 and 2 (Tables S2, S3). In solution, the conjugate 2 was always present as a diastereomer mixture, because of the tautomeric equilibrium of the succinimide ring, eventually resulting in the rapid carbon C1 racemization. Even though the two forms could be separated by HPLC (Figure S2A-C), each single diastereomer rapidly racemized at physiological pH to equimolar ratio of R and S forms in position C1, as assessed by time course 1H-1D NMR spectra (Figure S2D). Because of this rapid racemization, a pooled product of the two diastereomers was used for subsequent studies. The chemical shifts of the two diastereomers were identical, except for those assigned to Cysteine 1 (1C) and to the succinimide ring (Table S3). Cyclopeptide 1 and its conjugate 2 shared similar 13

C and 1H chemical shifts within the macrocycle, with chemical shift differences smaller than

0.30 ppm and 0.06 ppm, respectively, with the exception of residues 1C (protons and carbons) and of 6G (amide proton) (Figure 1), whose chemical shifts were influenced by the direct proximity to the linker (Figure S3, Table S2 and S3). The two molecules shared also similar 3

JHN-Hα and 3JHα-Hβ coupling constants (Table S2 and S3) and identical ROE patterns within the


6

Page 7 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


macrocycle (Table S4), implying similar conformations. Moreover, the linker, which did not show any ROE effect with the cyclopeptide, tumbled independently from the rest of the molecule, as assessed by NOESY experiments showing positive and negative NOEs (Figure S4) for the linker and the macrocycle, respectively. Thus, we conclude that the highly flexible linker does not interact with the cyclopeptide and does not influence the macrocycle overall conformation.

Table 1. Inhibition constant of cyclopeptide 1, its conjugates 1-4 and cilengitide for αvβ3 as determined by competitive binding assay.a Molecule

Replicates b

Ki [nM]

p-value c

1

13

103.0 ± 18.0 d

-

2

10

19.9 ± 3.5

< 0.0001

3

7

26.5 ± 5.1

0.0004

4

10

21.6 ± 4.4

< 0.0001

d

4 1.0 ± 0.1 0.0001 cilengitide Binding was measured by a competitive binding assay, as described in the Experimental Section, using a complex made by an N-terminal acetylated isoDGR peptide biotinylated at the ε-amino group of the lysine, (acetyl-CisoDGRCGVRSSSRTPSDKY-bio), and a streptavidinperoxidase conjugate. The Kd of the probe for αvβ3 is 1.3 nM. For each competitor, the inhibition constant, Ki, and the associated standard error are reported. a

b

Number of independent experiments (each in duplicate).

c

One-way ANOVA with Dunnett's multiple comparisons post-test using as control cyclopeptide 1. The Ki of cyclopeptide 1 and of cilengitide, previously reported12 have been determined using the same assay used for conjugates 2, 3, and 4. d

Conformational sampling of the cyclopeptide 1


7


Page 8 of 38

The presence of the β-bond in the backbone of isoDGR-containing cyclopeptides usually augments the range of accessible interconverting conformations,14 whose identification is challenging for standard spectroscopic techniques. As a matter of fact, both in cyclopeptide 1 and in its conjugate 2 the macrocycle was characterized by the paucity of NOE and ROE long range effects, supporting the notion that in aqueous solution it is in rapid equilibrium between different conformations on the NMR time scale. Prompted by our previous experience on RGD and isoDGR containing cyclopeptides,14,26 we reasoned that computational enhanced sampling techniques could represent a valid strategy to exhaustively explore the conformational space of such a highly dynamic system. In particular, Metadynamics in its Bias Exchange variant (BEMetaD)27 represents a valuable method for the acceleration of rare events and provides a reliable estimate of the populations of the accessible conformers in cyclopeptides.28 Therefore, based on the NMR observations, suggesting that the chemical linker of conjugate 2 does not influence the macrocycle conformational space, we performed BE-MetaD simulations on cyclopeptide 1 to characterize the macrocycle conformational equilibrium. Simulations and analysis details are provided in Supplementary Information. In agreement with the expected high conformational variability, the reconstructed free energy of 1 was characterized by four main minima (all with a population between 14% and 30%) (Figure 2). We scrutinized the four families of conformers for turns, as defined by Chou et al.,29 and hydrogen bonds patterns with VMD H-bonds plugin,30 using a donor-acceptor distance and angle cut off of 3.5 Å and 35°, respectively. Only H-bonds present in at least 40% of structures were considered. Of note, we identified only single β and α turns (Figure 2). Double β-turns, which are a hallmark of cyclic hexapeptides,31 were absent, conceivably because of the presence of the isoD Cβ carbon within the cycle backbone, that induced high conformational heterogeneity.14


8

Page 9 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2. BE-MetaD conformational sampling of cyclopeptide 1. A representative backbone conformation of each minimum populated more than 10% is reported. Structures are shown in sticks with nitrogen, oxygen, sulfur and polar hydrogen atoms in blue, red, yellow and white, respectively. The presence of α and β turns is highlighted coloring the carbon atoms of the involved residues in dark green and orange, respectively. H-bonds were represented by dashed lines.

The chemical linker of peptide conjugate 2 contributes to the interaction with αvβ3 To gain a structural understanding of the binding mode of conjugate 2 to αvβ3 in a native-like context and to investigate whether the chemical linker was directly involved in receptor interactions, we performed Saturation Transfer Difference (STD) experiments on conjugate 2 (1 mM) in the presence of native membranes derived from melanoma (MSR3) cell lines (75-125 x106 cells), displaying high levels of αvβ3 expression, as determined by FACS analysis (Figure S5). STD is a popular ligand-based NMR method, that is widely used to probe ligand-protein interactions of medium-weak affinity (high nM to mM) and to offer ligand-specific binding information at atomic resolution.32–34 The method is based on the transfer of saturation from the


9


Page 10 of 38

receptor to bound ligands that are in fast exchange with the binding site and allows for the identification of those ligand hydrogens which are closest to the receptor. Herein, the use of membrane-bound receptors is gaining increasing interest,35,36 as compared to whole cells systems or to reductionist approaches focusing only on the extracellular part of the isolated receptor, whose availability often represents a limiting factor. On the one hand the method preserves native-like conditions while avoiding cell internalization of the ligand,36,37 and on the other hand it requires reduced amounts of receptor as compared to experiments performed with isolated receptors, while allowing for a more efficient transfer of saturation, increased tumbling time and increased sensitiveness. Unambiguous STD effects with relative STD percentage (STD%) bigger than 50% were observed both for protons of the peptidic macrocycle (3isoD-HN, Hβa/b; 5R-Hε, 6G-HN) and of the linker (H5a, H6, H8ax, H9ax, H10, H12a/b) (Figure 1, Figure 3A,B) herewith indicating a spatial proximity of the chemical linker to membrane bound integrins. The interaction was further supported by tr-NOE experiments (Figure 3C), which highlighted i) an increase of the NOE effect of the protons within the macrocycle upon membrane addition and ii) a change in sign of the linker NOE effects, ascribable to an increase of the rotational correlation time of the bound ligand fraction. Moreover, in the presence of cell membranes the observed ROEs pattern was similar to that of the free molecule (Figure S6), suggesting that in the bound state

the

backbone

macrocycle

did

not

experience

major


10

Page 11 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. STD and tr-NOESY experiments on conjugate 2 in the presence of MSR3 membranes. (A) Amide (left) and aliphatic (right) regions of the STD experiment, on- and offresonance spectra are shown in the lower and top panel, respectively. (B) Relative STD percentage. Low (25%

Succinimide-Based Conjugates Improve IsoDGR Cyclopeptide Affinity

Recommend Documents