Receptor-Based Pharmacophore Model for

Subscriber access provided by UNIV OF PITTSBURGH

Article

A minimalist hybrid ligand/receptor-based pharmacophore model for CXCR4 applied to a small-library of marine natural products led to the identification of Phidianidine A as a new CXCR4 ligand exhibiting antagonist activity Rosa Maria Vitale, Monica Gatti, Marianna Carbone, Federica Barbieri, Vera Felicità, Margherita Gavagnin, Tullio Florio, and Pietro Amodeo ACS Chem. Biol., Just Accepted Manuscript • DOI: 10.1021/cb400521b • Publication Date (Web): 08 Oct 2013 Downloaded from http://pubs.acs.org on October 9, 2013

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 free 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 accessible to all readers and 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.

ACS Chemical Biology 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 27

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

ACS Chemical Biology

A minimalist hybrid ligand/receptor-based pharmacophore model for CXCR4 applied to a smalllibrary of marine natural products led to the identification of Phidianidine A as a new CXCR4 ligand exhibiting antagonist activity. Rosa Maria Vitale†,#, Monica Gatti§,#, Marianna Carbone†, Federica Barbieri§, , Vera Felicità†,‡, Margherita Gavagnin†, Tullio Florio*,§, and Pietro Amodeo*,† †

Institute of Biomolecular Chemistry (ICB) of the National Research Council (CNR), Via Campi Flegrei 34, 80078 Pozzuoli, Napoli, Italy ‡ Department of Health Science, “Magna Græcia” University, 88100, Catanzaro, Italy § Section of Pharmacology, Department of Internal Medicine, University of Genova, 16132 Genova, Italy Center of Excellence for Biomedical Research (CEBR), University of Genova, 16132 Genova, Italy

Abstract Here we present a minimal hybrid ligand/receptor-based pharmacophore model (PM) for CXCR4, a chemokine receptor deeply involved in several pathologies, such as HIV infection, rheumatoid arthritis, cancer development/progression and metastasization. This model, considerably simpler than those thus far proposed for this receptor, has been used to search for new CXCR4 inhibitors in a small marine natural product library available at ICB-CNR Institute (Pozzuoli, NA, Italy), since natural products, with their naturally-selected chemical and functional diversity, represent a rich source of bioactive scaffolds; computational approaches allow searching for new scaffolds with a minimal waste of possibly precious natural product samples; and our “stripped-down” model substantially increases the probabilities of identifying potential hits even in small-sized libraries. This search, also validated by a systematic virtual screening of the same library, has led to the identification of a new CXCR4 ligand, phidianidine A (PHIA). Docking studies supported PHIA activity and suggested its possible binding modes to CXCR4. Using the CXCR4-expressing/CXCR7negative GH4C1 cell line we show that PHIA inhibits CXCL12-induced DNA synthesis, cell migration, and ERK1/2 activation. The specificity of these effects was confirmed by the lack of PHIA activity in GH4C1 cells in which siRNA highly reduces CXCR4 expression and the lack of cytoxicity of PHIA was also verified. Thus PHIA represents a promising lead for a new family of CXCR4 modulators with 1 ACS Paragon Plus Environment


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 27

wide margins of improvement in potency and specificity offered by the small and very simple underlying PM. Introduction CXCR4, along with its cognate ligand CXCL12, has emerged as an important pharmacological target, to

treat

several

pathologies

such

as

HIV

infection,

rheumatoid

arthritis,

and

cancer

development/progression and metastasization. In particular, the CXCL12/CXCR4 system is involved in the tumorigenesis of several cancer histotypes and might determine the metastatic destination of tumor cells (1–7). CXCR4 shares its ligand, CXCL12, with CXCR7 a receptor whose biological function and significance are less characterized (8). Ligand binding in CXCR4 involves both the N-terminus and the transmembrane (TM) ligand binding site, which, as for other GPCRs, can be subdivided into two interconnected subpockets: the minor pocket, embracing TM helices 1,2,3,7 (TMS1), and the major pocket, comprising TM helices 3,4,5,6,7 (TMS2). Apart from peptides derived from CXCR4-binding proteins (9, 10), the most-extensively studied CXCR4 ligands originated by experimental library-screening or by combinatorial approaches, aimed at identifying HIV-entry blockers or, more recently, anticancer agents. The two main classes of CXCR4 modulators are either peptide/peptidomimetics derivatives of polyphemusin II, a 18-residue peptide isolated from the American horseshoe crabs (Limulus polyphemus) (11–13), or organic molecules derived from/related to the bicyclam compound AMD3100 (14). In both classes, in addition to optimization of previous hits, development of new molecules occurred by a progressive reduction of the original scaffolds and the associated pharmacophoric models. They shrank from six residues forming two triples of residues (11) to four or three functional groups (15) for polyphemusin II-derived molecules; and from the complex polyamino-macrocycles of the original bicylam compounds to the simpler three-vicinal-plus-one-distal functional group model present in isoquinoline-based AMD070

2 ACS Paragon Plus Environment

Page 3 of 27

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


and derivatives (16). A different approach, based on a virtual screening of the National Cancer Institute's (NCI) Open Chemical Repository Collection on a homology model of CXCR4 (17) has recently led to the identification of a rather simple lead structure, then used as a scaffold to search for other candidate CXCR4 ligands, but no PM was derived from this series of compounds. On the other hand, the recently-determined crystallographic structures of CXCR4 (18) complexed with two antagonists, the polyphemusin II-derived peptide CVX15, mainly spanning TMS2, and the organic inhibitor IT1t, binding into TMS1, unveiled the ligand-receptor interaction patterns in the two TM subpockets of the two complexes, thus paving the way to receptor-based drug-discovery on this target, since TM site is critical for both CXCL12-induced activation and antagonist inhibition of CXCR4. A closer look at the superposition of the two complexes (Fig. 1a) shows that they share a rather large region in the middle of the overall TM site, interacting with a common subset of CXCR4 residues: Tyr116, Asp187, Arg188 and Glu288, forming a saddle between the two subsites. Outside of this region, CVX15 forms polar interactions with Asp171, Asp262 in TMS2, while IT1t interacts with Asp97 in TMS1. Furthermore, both subpockets exhibit hydrophobic cavities featuring clusters of aromatic side-chains where ligands form several important hydrophobic interactions (Tyr190, Val196, Phe199, Gln200, His203, Tyr255 for CVX15 and Trp94, Trp102, Val112, Ile185, Arg183 for IT1t). By combining information on structural features of both receptor and ligands from the CXCR4 complexes with CVX15 and IT1t, we developed a loose, minimal hybrid-ligand/receptor-based PM, featuring only two functional groups, which in principle is capable of matching different combinations of HAPs within CXCR4 intrahelical site. This model was tested by using it to search for new potential modulators of the CXCR4 chemokine receptor in a small library of natural products (NPs) previously characterized at ICB-CNR for which suitable quantities to confirm predicted activities were available (see “Supporting information”). NPs are invaluable sources of molecular scaffolds endowed with highly diversified bioactivities 3 ACS Paragon Plus Environment


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 4 of 27

optimized by natural selection. In many cases the fraction of molecules exhibiting interesting activities is significantly higher in NP- than in synthetic-organic-libraries (19), thus making worthwhile even the screening of small NP-libraries to identify novel scaffolds active on biologically- or pharmacologicallyinteresting targets. The use of computational approaches to screen NP-libraries has the distinct advantage over totally-experimental approaches of sparing precious amounts of available NP samples, since the latter are only needed to confirm the activity of candidate hits identified in silico. The pharmacophore-based search led to the identification of a new CXCR4 candidate ligand, the indole alkaloid phidianidine A (PHIA), whose CXCR4-binding ability was assessed by docking, which provided possible CXCR4-PHIA binding modes. A much more time-consuming, yet accurate virtual screening of the same NP library on CXCR4 TM site confirmed PHIA as a candidate CXCR4 ligand, thus supporting the predictive power and efficiency of our PM. This prediction was then validated by pharmacological assays. PHIA exhibited antagonism to CXCL12-induced proliferation, migration and ERK1/2 activation, using the rat pituitary adenoma cell line GH4C1, which we previously showed to be responsive to this chemokine (20, 21). Importantly, GH4C1 expresses CXCR4 but, unlike most of the commonly used cell lines, does not release CXCL12, thus avoiding autocrine/paracrine CXCR4 activation loops, a significant confounding factor when characterizing CXCR4 ligands. Results and Discussion Pharmacophore Model. The search for a minimalist PM significantly simpler than those previously developed for CXCR4 started from the crystallographic structures of CXCR4 in complex with the polyphemusin II-derived peptide CVX15 and the organic inhibitor IT1t, and subsequently integrated features from these two compounds, but also from other known CXCR4 ligands. In spite of its large size and complex structure articulated into two subsites, CXCR4 TM site exhibits as emerging features the multiple occurrences both of negatively-charged residues (Asp97, Asp171, Asp187, Asp262 and Glu288),


Page 5 of 27

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


and of hydrophobic cavities, larger in TMS2, lined with clusters of aromatic side-chains and flanked and/or bottomed by polar groups, eventually provided by residues also contributing hydrophobic groups, like Arg188 and different Tyr residues. Therefore a large number of possible pairs of hydrophobic cavity-negatively-charged groups, which, by representing a distinctive signature of this site, will be henceforth named “hydrophobic-acidic pairs” (HAP), can be identified within the CXCR4 TM region. Thus, simultaneous targeting of the negatively-charged residues with positively-charged groups and the hydrophobic cavities with bulky, preferentially aromatic moieties may represent a successful strategy to identify/obtain good CXCR4 ligands. In this view, it is no surprise that available potent CXCR4 ligands, regardless of their chemical nature, ranging from proteins to relatively-small organic molecules, all share multiple occurrences of these two elements, and that the derived PMs were based on different combinations of basic groups and bulky (generally two- or three-ring) aromatic moieties. The experimental structures of CXCR4-IT1t and CXCR4-CVX15 complexes showed that IT1t and CVX15 interact with different receptor HAPs (see Introduction and Fig. 1a), thus suggesting that, while the blockade of residues critical for receptor activation is obviously an important general requisite for CXCR4-antagonists, the exact residues and, consequently, HAPs to be targeted exhibit some degree of redundancy, since different choices of subsites and HAPs can all lead to efficient CXCR4 antagonists. Altogether, these observations led us to elaborate a “loose-and-stripped-down” minimalistic PM for CXCR4, based on targeting only one, unspecified, HAP by single complementary pairs of bulky hydrophobic/aromatic and positively-charged groups, henceforth named “hydrophobic-basic pairs” (HBPs). This model is loose in its definition of spatial relationships between its functional groups, since it only includes an upper limit of 18 Å for the distance between their centers of mass, corresponding to the largest separation among all possible CXCR4 TM-site HAPs (i.e. the distance between the TMS2 5 ACS Paragon Plus Environment


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 6 of 27

hydrophobic pocket hosting the large aromatic naphthalene side-chain of CVX15 in 3OE0 and the TMS1 Asp97 residue interacting with IT1t in 3OE8). We did not adopt more stringent criteria because we wanted to develop a general model, not oriented toward any class or library of potential ligands and thus we wanted to avoid the rejection of new classes of CXCR4 ligands and to ensure the selection of mostly-diversified molecular scaffolds, even at the expense of possible higher rates of false positives and sub-optimal binding and/or selectivity of the identified leads. In this view, this PM can be considered both as a possible standalone signature for relatively small/simple CXCR4 ligands, and as a distinctive part of more complex ones, eventually featuring further HBPs. Additional requirements on size and/or occurrence of functional groups can be used to enforce either of the two interpretations of the model within a search protocol. Pharmacophore-based library search of CXCR4 ligands. Since we are currently developing a molecular database of molecules characterized at ICB-CNR, we used its already-available query tools to search for potential CXCR4 ligands compatible with this PM in a 250-molecule subset for which stored amounts of samples for subsequent functional tests existed (Table S1 in “Supporting Information”). This library, in spite of its relatively reduced size, featured a considerable diversity of chemical scaffolds in terms of size, polarity, number, nature of functional groups, and overall bulkiness and conformational freedom. Molecules endowed with the required functional groups were selected by a functional-group search of bulky-hydrophobic or aromatic moieties and of basic groups, while the proper maximum spacing was enforced by applying an upper-limit cut-off on the distance between the centers of mass of the PM functional groups to the fully-extended conformations of the selected compounds. This search produced only one candidate: phidianidine A (PHIA), an indole alkaloid isolated from the marine opisthobranch mollusc Phidiana militaris (22) (Fig. 2), featuring a methyleneindole hydrophobic/aromatic group, and a pentylguanidine basic group. They are linked together by a 36 ACS Paragon Plus Environment

Page 7 of 27

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


amino-1,2,4-oxadiazole moiety, which by its potential extra H-bond donor amino group and acceptors N and O ring atoms and by the rigidity associated to the five-membered ring, represents the only departure from a totally-ideal PM template. However, in spite of this rigid group, the molecule exhibits a good overall flexibility ensured by nine rotatable bonds and by the lack of bulky groups between the indole and guanidine end groups. Molecular Docking. To confirm CXCR4-binding properties of PHIA, it was docked inside the CXCR4 TM site. Although crystal structures of CXCR4 complexes represent convenient receptor templates for docking, they can however exhibit ligand-induced deformations that, in turn, might bias the pose sampling. To attenuate this problem by including the effects of protein dynamics, a molecular dynamics (MD) approach, described along with the overall docking protocol in “Supporting Information”, provided an ensemble of 29 representative CXCR4-monomer structures exhibiting conformational diversity and lacking site-occlusion from global or local protein motions. They were used together with one protein monomer from each 3OE0 and 3OE8 PDB entry as the target in docking calculations, which resulted in twelve CXCR4-PHIA poses selected for analysis and comparisons (Fig. 1b-h). However, these poses can be generated by combining a smaller number of local ligand-receptor interaction patterns of the three main ligand fragments: three for the indole ring, four for the 3-amino1,2,4-oxadiazole group, and three for the guanidinium moiety (Table 1). The indole group fits, with different tilts and depths, the same TMS2 aromatic site occupied by CVX15 naphthalene moiety in eight poses (Fig. 1b,d,e); substantially overlaps the guanidinium of CVX15 Arg2 in three poses (Fig. 1g,h); and spans the region occupied by IT1t in one pose (Fig. 1f). The guanidinium moiety forms Hbonds with either Asp97 (seven poses, in four of which also with Asp187, Fig. 1b,c,g,h), or Glu288 (four poses, in one of which also with Asp97, Fig. 1b,d), or Asp262 (two poses Fig. 1e,f). Although the 3amino-1,2,4-oxadiazole group forms H-bonds with Glu288 in five poses (Fig. 1b,c,h), with Gln200 in three poses (Fig. 1e), with Arg188 in three poses (Fig. 1d,f) and with Asp97 in one pose (Fig. 1g), a 7 ACS Paragon Plus Environment


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 8 of 27

CXCR4-interaction analysis on the different PHIA groups (Table S2 in “Supporting Information”) shows that its contribution amounts on average to less than 4% of the total CXCR4-PHIA interaction energy. However, this group, while rather marginal for binding energetics, is quite influential on the complex structure. Indeed, when analysing the overall arrangement of PHIA inside CXCR4 binding site, special attention was dedicated to the aforementioned balance between the number of rotatable bonds/absence of bulky groups, which, in principle, could favour compact ligand conformations, and the presence of a rigid straight spacer, which instead introduces a bias toward more extended structures. In this view, the overwhelming contribution of this latter feature appears quite evident, since eleven out of the twelve selected poses exhibit extended or semi-extended ligand conformations. Consequently, although PHIA (26 heavy-atoms) is slightly smaller than IT1t (27 heavy-atoms) and significantly smaller than CVX15(1-3) (40 heavy-atoms), in ten of the poses, characterized by extended conformations, the ligand spans both CXCR4 intrahelical subsites. It only occupies a single subsite in 3OE0_MD_fr2366, where it protrudes out TMS2 toward the region occupied by the CVX15(14-16) segment in 3OE0 structure, still exhibiting a fully-extended conformation, and in 3OE0_1, which features the only partly-folded ligand conformation of the set, where PHIA only interacts with TMS1, being fully overlapped to IT1t in 3OE8 structure. In summary, molecular docking shows the potential involvement in PHIA binding of CXCR4 residues that are known to be important for the modulation of the receptor activity such as Asp97, Asp187, Asp262, Glu288, along with the cluster of aromatic residues in TMS2 and TMS1, since in most poses PHIA spans over both receptor subpockets. This latter feature can be ascribed to the 3-amino-1,2,4-oxadiazole group, which, acting as a rigid spacer, pushes its indole and guanidinium group toward spatially-distant HAPs. Virtual Screening. To confirm the efficiency of a filtering based on our PM, the same NP library was used for a virtual screening of the CXCR4 TM site (see “Supporting Information” for details). This approach, while requiring substantially longer computing times than the pharmacophore-based 8 ACS Paragon Plus Environment

Page 9 of 27

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


screening (by a Nm/Nph factor, where Nm=number of molecules in the database, Nph=number of hits in the pharmacophore-based search, i.e. 250/1=250 in the present case), provided the same result by selecting PHIA as main CXCR4 ligand candidate. In vitro functional tests to validate computational prediction. The docking results prompted us to assay the biological effects of PHIA on CXCR4. However, we directly performed functional tests rather than competitive binding assays with [125I]-SDF1α (CXCL12) and/or anti-CXCR4 mAbs since previously characterized CXCR4 features such as different receptor oligomerization states, or the presence of distant binding subsites (23), can determine partial loss of correlation between measured receptor binding and different biological activities of a ligand. In particular, the CXCL12 multistep binding mechanism involves spatially distant regions of the receptor: chemokine N-loop interacts with receptor N-terminus and in turn CXCL12 N-terminus interacts with CXCR4 TM region to trigger receptor activation. This feature, by making possible the simultaneous binding of ligand (in TMS) and chemokine (to CXCR4 N-terminus), may determine a decoupling between real ligand binding and chemokine displacement from the receptor, thus producing in turn a net ligand binding underestimation. The non-linear responses observed in ligand-displacement assays on CXCR4 with vMIP-II protein and derived peptides when using [125I]-CXCL12-α vs. 12G5 antibody as reference ligands (24), being supportive of an only-partial overlap among receptor subsites responsible for binding of CXCL12, vMIP-II and 12G5 antibody, could represent a possible example of this behaviour, and they can be expected to occur even more easily for small ligands that fill only a fraction of the large CXCR4 TMS site. Moreover, several CXCR4 inhibitors exhibited low-correlation between functional and binding assays toward the wild-type (wt) receptor, which turned into high-correlation when binding was instead measured on a CXCR4 constitutionally-active mutant (25).



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 10 of 27

PHIA reduce CXCL12-induced proliferation in GH4C1. As previously reported (20), we show that CXCL12 (50 nM) induces a significant increase in DNA synthesis of 72h serum-starved for GH4C1 cells (+32%, after 24 hrs, p0.01, as reported in Fig. 3a). Similarly, also the treatment with AMD3100 was ineffective in GH4C1 siRNA/CXCR4 clone 17 and 29 cells (Fig. 3b). To obtain a more easily detectable cell growth arrest, we evaluated, in another series of experiments, the effects of a prolonged inhibition of CXCR4 by PHIA on CXCL12 induced proliferation after 48 hrs of treatment. In particular we observed that in GH4C1 wt, after 48 hrs of treatment, CXCL12 increased DNA synthesis (+51%, vs. untreated control cells), measured by BrdU incorporation assay, but it was ineffective in GH4 siRNA/CXCR4 clones 3, 17, and 29 (+7, -15, and +9%, respectively). Moreover, we confirmed the dose-dependent inhibition of PHIA on CXCL12-treated GH4C1 wt cells although a more pronounced effect was observed, since it was statistically significant already at 500 nM and maximal for 10 and 50 µM concentrations, and a higher statistical significance (p99% of the light transmission in a wavelength region between 490 and 700 nm. Thus in these experimental condition it is possible to discriminate the position of GFP-labeled cells in the two sides of the inserts. In these experiments, we used GH4C1 wt cells transfected with GFP in comparison with GH4C1 clones co-transfected with CXCR4/siRNA and GFP plasmids (GH4C1-GFP siRNA/CXCR4 clones G3 and G7). Cells (104) were placed on top of inserts in F10 medium/10% FCS as controls or in the presence of PHIA 50µM or AMD3100 10µM, and were placed in a 24-well with companion plate containing 600µl of F10/10% FCS with or without CXCL12 (50 nM) as chemoattractant molecule. Plates were then incubated at 37°C for 24 hrs and then read using a Tecan i-control, 1.7.1.12 infinite 200Pro plate reader at 488/520nm (excitation/emission filter). Interestingly, in the absence of FCS in the upper chamber no migration was observed even in the presence of high CXCL12 concentrations in the lower chamber, while a significant migration was observed by FCS alone (data not shown); thus this experimental condition was considered as control value. In the presence of FCS, CXCL12 confirmed itself to be a powerful chemoattractant factor for GH4C1 cells increasing by 300% cell migration, as compared to cells in which the chemokine was not added in the lower chamber. Both PHIA (10 and 50 µM) and AMD3100 (10 µM) added in the top insert completely suppressed CXCL12-induced migration. Importantly, the presence of PHIA also slightly


Page 13 of 27

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


reduced basal (FCS-induced) migration of these cells, an event that did not occur with AMD3100 (Fig. 5). In the GH4C1 clones in which CXCR4 expression was silenced, CXCL12 did not reduce cell migration and PHIA and AMD3100 had no effects in both basal and CXCL12-stimulated conditions. Intracellular mechanisms mediating the antiproliferative activity of PHIA. Next, we investigated the effects of PHIA on CXCR4 signal transduction evaluating ERK1/2 phosphorylation/activation (28). Our results show that PHIA significantly inhibited CXCL12-dependent ERK1/2 phosphorylation at the concentrations of 0.1, 1 and 10 µM. Moreover, treatment with PHIA (1 and 10 µM) also reduced basal ERK1/2 phosphorylation (Fig. 6a). The quantification of these effects by densitometric analysis, and the statistical evaluation, is reported in Figure 6b. Conclusions. Here we present a new minimalistic hybrid ligand/receptor-based PM for CXCR4, founded on just one, properly-spaced HBP. The use of this model to filter compounds in a small database of NPs led to the identification of a new potential CXCR4 ligand, PHIA, whose biological and pharmacological characterization fully confirmed the theoretical prediction, thus validating the use of this PM for the identification of new CXCR4 ligands. The proposed PM can be considered, in principle, somewhere in between a full-standalone motif and a distinctive signature, present in multiple copies and/or alongside other functional groups, for CXCR4 ligands. PHIA and the recently-discovered NSC56612 (plus its related compounds) (17) represent a closer approximation to the first interpretation of the model, since they add to the PM prescriptions only the 3-amino-1,2,4-oxadiazole group and a secondary nitrogen atom, respectively, both located within the linear linker joining their HBPs. On the contrary, PM signature occurs within all other known CXCR4 ligands, including the cognate agonist CXCL12 and the antagonist vMIP-II proteins, either in multiple copies, or as a part of more complex CXCR4-binding motifs. Both PHIA and NSC56612-related ligands most closely corresponding to our PM contain at least one additional functional group within the linker chain; thus minor stabilizing 13 ACS Paragon Plus Environment


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 14 of 27

effects from other groups cannot be totally ruled out, although their contributes to the overall stability of CXCR4-ligand complexes, as suggested by docking results on both PHIA and NSC56612, seem quite secondary. However, a quantitative characterization of the binding modes of this ligand class requires a direct (X-ray crystallography) and/or indirect (SAR studies on ligand analogs/derivatives) structural investigation, possibly involving both PHIA derivatives and new scaffolds from an extension of the search to other databases. The interesting pharmacological properties exhibited in in vitro assays make PHIA a promising lead for the development of a new family of CXCR4 modulators, also in view of the wide potential margin of improvement offered by its quite simple structure. While being an organic molecule, PHIA can also be viewed as a “bridge” between peptidic and organic CXCR4 ligands since its indole and guanidinic functional groups also occur in sidechains of His and Arg residues, respectively. The ability of PHIA, in contrast to AMD3100, to reduce basal CXCR4 activity on relevant parameters including cell migration, is reminiscent of a CXCR4 inverse agonist activity, also considering that no effects were observed in CXCR4-downregulated cells, although further studies are required to delve deeper into this issue. If confirmed, these data could open novel pharmacological uses, besides those already proposed for true CXCR4 antagonists such as AMD3100. Recently, PHIA compound has also been identified as partial agonist of µ-opioid receptor (MOR) (29). Because the colocalization of CXCR4 and MOR in multiple brain areas and their mutual interactions (30) had been previously demonstrated, this compound could also be evaluated as a multiligand agent for the treatment of pain in neuroimmune disease. Associated Content Supporting Information Methods for: Molecular Mechanics and Dynamics, Clustering; Binding Site Volume and Rotamer


Page 15 of 27

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


Analysis; Docking and Virtual Screening; Cell Cultures; Short-interfering RNA Preparation and Transfection; RNA Extraction; cDNA Synthesis and qRT-PCR; Western Blot; BrdU Incorporation, Cell Survival, and In Vitro Migration Assays. Details on: Molecular Dynamics of the empty CXCR dimer; Expression of CXCR4 and CXCR7 in GH4C1 cell lines, Generation of CXCR4/siRNA GH4C1 cell lines. This material is available free of charge via the Internet at http://pubs.acs.org. Author Information Corresponding Authors * E-mail: [email protected]; [email protected]. Equally-contributing Authors #

These authors contributed equally.

Notes The authors declare no competing financial interest. Acknowledgement Technical assistance. The authors acknowledge Mr. Salvatore Donadio for the ICB Molecular Database programming and management. Financial Support. This work was supported by Associazione Italiana per la Ricerca sul Cancro (AIRC) 2012 [Grant IG13563] to T.F. References 1. Bajetto, A., Barbieri, F., Dorcaratto, A., Barbero, S., Daga, A., Porcile, C., Ravetti, J. L., Zona, G., Spaziante, R., Corte, G., Schettini, G., and Florio, T. (2006) Expression of CXC chemokine receptors 1-5 and their ligands in human glioma tissues: role of CXCR4 and SDF1 in glioma cell proliferation and migration., Neurochem. Int. 49, 423–432. 2. Bajetto, A., Barbieri, F., Pattarozzi, A., Dorcaratto, A., Porcile, C., Ravetti, J. L., Zona, G., Spaziante, R., Schettini, G., and Florio, T. (2007) CXCR4 and SDF1 expression in human meningiomas: a proliferative role in tumoral meningothelial cells in vitro., Neuro. Oncol. 9, 3– 11. 15 ACS Paragon Plus Environment


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 16 of 27

3. Barbieri, F., Bajetto, A., Stumm, R., Pattarozzi, A., Porcile, C., Zona, G., Dorcaratto, A., Ravetti, J.-L., Minuto, F., Spaziante, R., Schettini, G., Ferone, D., and Florio, T. (2008) Overexpression of stromal cell-derived factor 1 and its receptor CXCR4 induces autocrine/paracrine cell proliferation in human pituitary adenomas., Clin. Cancer Res. 14, 5022– 5032. 4. Koshiba, T., Hosotani, R., Miyamoto, Y., Ida, J., Tsuji, S., Nakajima, S., Kawaguchi, M., Kobayashi, H., Doi, R., Hori, T., Fujii, N., and Imamura, M. (2000) Expression of stromal cellderived factor 1 and CXCR4 ligand receptor system in pancreatic cancer: a possible role for tumor progression., Clin. Cancer Res. 6, 3530–3535. 5. Muller, A., Homey, B., Soto, H., Ge, N., Catron, D., Buchanan, M. E., McClanahan, T., Murphy, E., Yuan, W., Wagner, S. N., Barrera, J. L., Mohar, A., Verastegui, E., and Zlotnik, A. (2001) Involvement of chemokine receptors in breast cancer metastasis., Nature 410, 50–56. 6. Robledo, M. M., Bartolome, R. A., Longo, N., Rodriguez-Frade, J. M., Mellado, M., Longo, I., van Muijen, G. N., Sanchez-Mateos, P., and Teixido, J. (2001) Expression of functional chemokine receptors CXCR3 and CXCR4 on human melanoma cells., J. Biol. Chem. 276, 45098–45105. 7. Teicher, B. A., and Fricker, S. P. (2010) CXCL12 (SDF-1)/CXCR4 pathway in cancer., Clin. Cancer Res. 16, 2927–2931. 8. Sun, X., Cheng, G., Hao, M., Zheng, J., Zhou, X., Zhang, J., Taichman, R. S., Pienta, K. J., and Wang, J. (2010) CXCL12 / CXCR4 / CXCR7 chemokine axis and cancer progression., Cancer Metastasis Rev. 29, 709–722. 9. Elisseeva, E. L., Slupsky, C. M., Crump, M. P., Clark-Lewis, I., and Sykes, B. D. (2000) NMR studies of active N-terminal peptides of stromal cell-derived factor-1. Structural basis for receptor binding., J. Biol. Chem. 275, 26799–805. 10. Crump, M. P., Elisseeva, E., Gong, J., Clark-Lewis, I., and Sykes, B. D. (2001) Structure/function of human herpesvirus-8 MIP-II (1-71) and the antagonist N-terminal segment (1-10)., FEBS Lett. 489, 171–5. 11. Tamamura, H., Imai, M., Ishihara, T., Masuda, M., Funakoshi, H., Oyake, H., Murakami, T., Arakaki, R., Nakashima, H., Otaka, A., Ibuka, T., Waki, M., Matsumoto, A., Yamamoto, N., and Fujii, N. (1998) Pharmacophore identification of a chemokine receptor (CXCR4) antagonist, T22 ([Tyr(5,12),Lys7]-polyphemusin II), which specifically blocks T cell-line-tropic HIV-1 infection., Bioorg. Med. Chem. 6, 1033–1041. 12. Tamamura, H., Hori, A., Kanzaki, N., Hiramatsu, K., Mizumoto, M., Nakashima, H., 16 ACS Paragon Plus Environment

Page 17 of 27

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


Yamamoto, N., Otaka, A., and Fujii, N. (2003) T140 analogs as CXCR4 antagonists identified as anti-metastatic agents in the treatment of breast cancer., FEBS Lett. 550, 79–83. 13. Arakaki, R., Tamamura, H., Premanathan, M., Kanbara, K., Ramanan, S., Mochizuki, K., Baba, M., Fujii, N., and Nakashima, H. (1999) T134, a small-molecule CXCR4 inhibitor, has no cross-drug resistance with AMD3100, a CXCR4 antagonist with a different structure., J. Virol. 73, 1719–1723. 14. De Clercq, E., Yamamoto, N., Pauwels, R., Baba, M., Schols, D., Nakashima, H., Balzarini, J., Debyser, Z., Murrer, B. A., Schwartz, D., Thornton, D., Bridger, G., Fricker, S., Henson, G., Abrams, M., and Picker, D. (1992) Potent and selective inhibition of human immunodeficiency virus (HIV)-1 and HIV-2 replication by a class of bicyclams interacting with a viral uncoating event., Proc. Natl. Acad. Sci. U. S. A. 89, 5286–5290. 15. Ueda, S., Kato, M., Inuki, S., Ohno, H., Evans, B., Wang, Z., Peiper, S. C., Izumi, K., Kodama, E., Matsuoka, M., Nagasawa, H., Oishi, S., and Fujii, N. (2008) Identification of novel non-peptide CXCR4 antagonists by ligand-based design approach., Bioorg. Med. Chem. Lett. 18, 4124–9. 16. Miller, J. F., Gudmundsson, K. S., Richardson, L. D., Jenkinson, S., Spaltenstein, A., Thomson, M., and Wheelan, P. (2010) Synthesis and SAR of novel isoquinoline CXCR4 antagonists with potent anti-HIV activity, Bioorg. Med. Chem. Lett. 20, 3026–3030. 17. Kim, J., Yip, M. L. R., Shen, X., Li, H., Hsin, L.-Y. C., Labarge, S., Heinrich, E. L., Lee, W., Lu, J., and Vaidehi, N. (2012) Identification of anti-malarial compounds as novel antagonists to chemokine receptor CXCR4 in pancreatic cancer cells., PLoS One 7, e31004. 18. Wu, B., Chien, E. Y. T., Mol, C. D., Fenalti, G., Liu, W., Katritch, V., Abagyan, R., Brooun, A., Wells, P., Bi, F. C., Hamel, D. J., Kuhn, P., Handel, T. M., Cherezov, V., and Stevens, R. C. (2010) Structures of the CXCR4 chemokine GPCR with small-molecule and cyclic peptide antagonists., Science 330, 1066–1071. 19. Paterson, I., and Anderson, E. A. (2005) The renaissance of natural products as drug candidates., Science 310, 451–3. 20. Florio, T., Casagrande, S., Diana, F., Bajetto, A., Porcile, C., Zona, G., Thellung, S., Arena, S., Pattarozzi, A., Corsaro, A., Spaziante, R., Robello, M., and Schettini, G. (2006) Chemokine stromal cell-derived factor 1alpha induces proliferation and growth hormone release in GH4C1 rat pituitary adenoma cell line through multiple intracellular signals., Mol. Pharmacol. 69, 539– 546. 21. Massa, A., Casagrande, S., Bajetto, A., Porcile, C., Barbieri, F., Thellung, S., Arena, S., Pattarozzi, A., Gatti, M., Corsaro, A., Robello, M., Schettini, G., and Florio, T. (2006) SDF-1 17 ACS Paragon Plus Environment


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 18 of 27

controls pituitary cell proliferation through the activation of ERK1/2 and the Ca2+-dependent, cytosolic tyrosine kinase Pyk2., Ann. N. Y. Acad. Sci. 1090, 385–398. 22. Carbone, M., Li, Y., Irace, C., Mollo, E., Castelluccio, F., Di Pascale, A., Cimino, G., Santamaria, R., Guo, Y.-W., and Gavagnin, M. (2011) Structure and Cytotoxicity of Phidianidines A and B: First Finding of 1,2,4-Oxadiazole System in a Marine Natural Product, Org. Lett. 13, 2516–2519. 23. Scholten, D. J., Canals, M., Maussang, D., Roumen, L., Smit, M. J., Wijtmans, M., de Graaf, C., Vischer, H. F., and Leurs, R. (2012) Pharmacological modulation of chemokine receptor function., Br. J. Pharmacol. 165, 1617–43. 24. Zhou, N., Luo, Z., Luo, J., Hall, J. W., and Huang, Z. (2000) A novel peptide antagonist of CXCR4 derived from the N-terminus of viral chemokine vMIP-II., Biochemistry 39, 3782– 3787. 25. Wang, Z., Tamamura, H., Frilot, N., Broach, J., Fujii, N., and Peiper, S. C. (2007) Screening and characterization of cyclic pentapeptide CXCR4 antagonists/inverse agonists using a pheromone responsive reporter gene in Saccharomyces cerevisiae: Utility of G protein coupled receptor constitutively active mutants. In Chemokine Biology - Basic Research and Clinical Application - Vol. II (Neote, K., Letts, L. G., and Moser, B., Eds.), pp 61–77, Birkhäuser, Basel. 26. Bajetto, A., Bonavia, R., Barbero, S., Piccioli, P., Costa, A., Florio, T., and Schettini, G. (1999) Glial and neuronal cells express functional chemokine receptor CXCR4 and its natural ligand stromal cell-derived factor 1., J. Neurochem. 73, 2348–2357. 27. De Clercq, E. (2009) The AMD3100 story: the path to the discovery of a stem cell mobilizer (Mozobil), Biochem. Pharmacol. 77, 1655–1664. 28. Pattarozzi, A., Gatti, M., Barbieri, F., Wurth, R., Porcile, C., Lunardi, G., Ratto, A., Favoni, R., Bajetto, A., Ferrari, A., and Florio, T. (2008) 17beta-estradiol promotes breast cancer cell proliferation-inducing stromal cell-derived factor-1-mediated epidermal growth factor receptor transactivation: reversal by gefitinib pretreatment., Mol. Pharmacol. 73, 191–202. 29. Brogan, J. T., Stoops, S. L., and Lindsley, C. W. (2012) Total synthesis and biological evaluation of phidianidines A and B uncovers unique pharmacological profiles at CNS targets., ACS Chem. Neurosci. 3, 658–64. 30. Heinisch, S., Palma, J., and Kirby, L. G. (2011) Interactions between chemokine and muopioid receptors: anatomical findings and electrophysiological studies in the rat periaqueductal grey., Brain. Behav. Immun. 25, 360–72.


Page 19 of 27

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


Table 1. Energetics and intermolecular contacts in representative models of the CXCR4-PHIA complex. Autodock 4.2 binding free energies (first column, values in kcal mol-1) and receptor residues exhibiting distances

Receptor-Based Pharmacophore Model for

Recommend Documents