Toward Angiogenesis Inhibitors Based on the Conjugation of Organometallic Platinum(II) Complexes to RGD Peptides

A novel conjugate between a cyclometalated platinum(II) complex with dual antiangiogenic and antitumor activity and a cyclic peptide containing the RGD sequence (‐Arg‐Gly‐Asp‐) has been synthesized by combining solid‐ and solution‐phase methodologies. Although peptide conjugation rendered a non‐cytotoxic compound in all tested tumor cell lines (± αVβ3 and αVβ5 integrin receptors), the antiangiogenic activity of the Pt‐c(RGDfK) conjugate in human umbilical vein endothelial cells at sub‐cytotoxic concentrations opens the way to the design of a novel class of angiogenesis inhibitors through conjugation of metallodrugs with high antiangiogenic activity to cyclic RGD‐containing peptides or peptidomimetic analogues.


Introduction
Angiogenesis, the process by which new blood vessels are formed from pre-existing vessels,p lays ac rucial role in several events, such as embryogenesis, growth, andw oundh ealing. [1,2] Likewise, it is widely accepted that aberrant angiogenesis occurs in other pathological conditions, including cancer growth and metastasis, due to the necessity of supplyingn utrients and oxygen to malignant tumor tissues.O wing to the potentialo fa ngiogenesis as at herapeutic target for blocking cancer growth, in recent decades, many efforts have been devoted to the development of novel antiangiogenic agents. [3] United States Food and Drug Administration (FDA)-approved angiogenesis inhibitors based on antibodies (bevacizumab) or small molecules (sorafenib and sunitinib)a re currently being used in clinics for the treatment of various cancers, either alone or in combination with other chemotherapeutic agents. [4] However,t hey are not exemptf rom drawbacks, including severely toxic side effects. [5] Bioactivep eptides have also demonstrated great potential in antiangiogenic therapies, as exemplified by cilengitide (1; Scheme 1). This cyclic RGD pentapeptide developed by Kessler and co-workersi nt he early 1990s is currently in late-stage clinical trials for the treatment of glioblastoma, as wella si np reclinicals tudies in combination with classical therapies, such as radio and cytotoxic agents, due to its lacko ft oxicitya nd high drug tolerance [6][7][8] (e.g., 1 with cetuximab, cisplatin, and5 -fluorouracili nm etastatic squamous cell cancer of the head and neck; [9] 1 with cetuximab and platinum-based chemotherapy as first-line treatment in advanced non-small-cell lung cancer). [10] Peptide-based drugs, such as 1,o ffer several advantages over small-molecule-based drugs or largea ntibodies because key factors, such as enzymatic stability,t arget specificity, and toxicity, can be fine-tuned throughc hemical modifications (e.g.,c yclization, N-methylation, and incorporation of d-amino acid residues) to obtain privileged bioactive structures. [11,12] Compound 1 is the result of constraining the RGD motif in its optimum conformation for binding to the a V b 3 integrin recep-An ovel conjugate between ac yclometalated platinum(II) complex with dual antiangiogenic anda ntitumor activity and a cyclic peptide containing the RGD sequence (-Arg-Gly-Asp-) has been synthesized by combining solid-and solution-phase methodologies. Although peptidec onjugation rendered a non-cytotoxic compound in all tested tumor cell lines (AE a V b 3 and a V b 5 integrin receptors), the antiangiogenic activity of the Pt-c(RGDfK) conjugate in human umbilical vein endothelial cells at sub-cytotoxic concentrationso pens the way to the designo fan ovel class of angiogenesis inhibitors through conjugation of metallodrugs with high antiangiogenic activity to cyclic RGD-containing peptides or peptidomimetic analogues.
[a] Dr.A .Zamora, A. Gandioso tor (and, to al esser extent,toa V b 5 and a 5 b 1 ). [6][7][8] Integrins are a class of heterodimeric transmembrane receptors that mediate cell-cellc ommunication and cell-extracellular matrix interactions, and many of them,i ncluding a V b 3 and a V b 5 ,a re involved in tumora ngiogenesis. [13] For this reason, al arge number of RGD-containing peptides (including 1 and other analogues)o r peptidomimetics have been developed to target integrin receptors with the aim of searching for novel anticancer agents. In addition, the particularo verexpression of a V b 3 , a V b 5 ,a nd a 5 b 1 integrins in tumor cells offer the possibilityo fu sing RGD peptides as delivery vectors in both therapy and diagnosis. [14][15][16] On the other hand, the development of nonconventional metal complexes with antiangiogenic properties offers great potentiali nc hemotherapy as an alternative to classical platinum drugs based on cisplatin and derivatives. In spite of the relativelyl arge number of nontoxic angiogenesis inhibitors based on Ru II ,R h II ,a nd Ir III complexes reported in recent years, [17][18][19] bioactivem etallodrugs with dual cytotoxicity and antiangiogenic properties are scarce, [20][21][22][23] particularly in the platinum field. Recently,w ed escribed af amilyo fc yclometalated Pt II complexes of the type [Pt(dmba-R)(dmso)Cl] (dmba = C,N-dimethylbenzylamine;R= MeO, Me, H, Br,F ,C F 3 ,a nd NO 2 substituents at positions 4a nd 5o ft he phenyl ring;S cheme1) exhibitingI C 50 values in the sub-micromolar range in the A2780 cell line and inhibition of angiogenesis at ac oncentration close to their IC 50 values. [24] Because the introduction of substituents into the C,N-chelatingl igand did not influence their dual antiangiogenic and cytotoxic activities in vitro, attachment to ac arriera gent was envisaged as ap romising strategyt oi mprove the pharmacological properties of such novel metallodrugs.
Herein,w ef ocus on the conjugation of ap romising organoplatinum(II) complex with dual antitumor and antiangiogenic activity,[ PtCl(dmba)(dmso)] (2;S cheme 1), [24] to c(RGDfK), a cyclic conjugatable version of the RGD motif, with the aim of exploring the antiangiogenic activity of the resulting conjugate. Solid-phase peptide synthesis (SPPS) in combination with solution-phase approaches have been used for the preparation of the bioconjugate (3;S cheme 1), and the effect of peptide conjugation on the biological activity of the parent Pt II complex has been investigated in ap anel of human cancerc ell lines that overexpress or not integrin receptors relatedt oa ngiogenesis.

Results and Discussion
Synthesis and characterization of the Pt-c(RGDfK) conjugate (3) The preparation of 3 relied on the synthesis of the peptide and subsequent conjugation to the metal complex in solution (Scheme 2). This approach is especially useful for the preparation of bioconjugates that incorporate sensitive metallodrugs based on Pt II and Pt IV complexes. [25][26][27][28][29] The conjugation of platinum complex 2 with the peptidet hrough an amide bond required the incorporation of ac arboxylic function in the C,N-chelating ligand.T herefore, 4-[(dimethylamino)methyl]benzoic acid was prepared through the reactiono f4 -(chloromethyl)benzoica cid with dimethylamine, and subsequently cyclometalationo fcis-[PtCl 2 (DMSO) 2 ]a fforded complex 4 in 55 %y ield (Scheme 2A), which was characterized by NMR spectroscopy (Figures S1-S3 in the Supporting Information). The 1 HNMR spectrum shows all diagnostic resonances, including the CH 6 of the dmba ligand flankedb y 195 Pt satellitesa nd ab road signal at high field corresponding to the carboxylic acidf unction required for conjugation. Complex 4 wasa lso characterized by positive-ion high-resolution (HR) ESI-MS.
On the other hand, it is worth noting that the well-established angiogenesis inhibitor 1 cannot be conjugated to a cargo drug as it is because three amino acids are essential for binding to the integrin receptors (RGD triad), d-Phe is involved in hydrophobic interactions and the N-methylated Valr esidue has no derivatizable functional group. Although the incorporation of N-methyl aminoa cids turned out to be one of the most important structural modifications of 1, [30] replacement of NMeVal by Lyso rG lu provides new functional groups that can be further functionalized, while retaining the integrin binding affinity. [31] Thus, ac onjugatable version of 1,c (RGDfK), has been employed and further modified through the e-amino functionoft he lysine residue. As shown in Scheme2B, peptide 5,w hich incorporates ap olyethylene glycol spacer at the side chain of the lysine residueo fc (RGDfK), was first synthesized by combining SPPS for the assembly of the peptides equence and solution-phaser eactions for the cyclization and removal of the side-chain protecting groups.B riefly,t he linear cyclopentapeptide was assembled on 2-chlorotrityl chloride resin by using standard Fmoc/tBu chemistry,c leaved under mildly acidic conditions (1 %A cOH for3 0min), and cyclizedw ithP yBOP in the presenceo fD IPEA to afford the protected peptide c[-Arg(Pbf)-Gly-Asp(tBu)-d-Phe-Lys(linker)-]. Finally,t he remaining sidechain protecting groups were eliminated by acidic treatment, and desired peptide 5 was purified by reversed-phaseH PLC and characterized by ESI-MS.
Conjugate 3 was synthesized through the reaction of peptide 5 with complex 4 (Scheme 2), which was previously activated with HATU and DIPEA in anhydrous DMF (0.8 m LiCl) for 10 min. These conjugation conditions proved to be better than those with ac arbodiimide (e.g.,1 -ethyl-3-(3-dimethylaminopropyl)carbodiimide) and 1-hydroxybenzotriazoleo rH ATUa nd DIPEA in anhydrous DMF without LiCl as ac oupling agent. The expected conjugate was obtained as aw hite solid (overall yield 22 %) after purification by reversed-phase HPLC and lyophilization ( Figure S4 With 3 in hand, we first studied the expression levels of a V b 3 and a V b 5 integrinsi napanel of cancerc ell lines to choose the best model for investigating how the cytotoxic activity of the parentc omplex( 2)w as affectedd ue to peptide conjugation. In good agreement with previous studies, [25,26] we found that SK-MEL-28 cells expressed high levels of a V b 3 integrin, whereas av ery low expression of both integrins was found in CAPAN-1 cells (Figure 2). These results indicate that the melanoma cell line is ag ood modelo fc ells that overexpress a V b 3 ,w hereas pancreatic cancer cells can be used as the negative control. In contrast, MDA-MB-231 cells exhibit similare xpression levelso f a V b 5 to that of SK-MEL-28, but the negligible expression of a V b 3 makes breast cancer cells as uitable model of cells that overexpress a V b 5 .

Cytotoxicity studies
The antiproliferativea ctivities of 1-3 were determined in selected cell lines by using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction assay,a sw ell as that of cisplatin, which was included for comparison purposes. First, it is worth noting that the cytotoxic activity of the parent Pt complex (2)w as lower in all cell lines tested (Table 1a nd Figure3)c ompared with A2780 cells, which were previously found to be highly sensitive to this compound (IC 50 = 0.98 mm). [24] Second, the cytotoxicity of 2 was drastically reduced (IC 50 > 100 mm)a fter conjugation to the peptide. This result was not completely unexpected upon taking into account the low cytotoxic activity of 2 in a V b 3 /a V b 5 + cells, and the fact that conjugation was usually accompanied by ar eduction of the antitumor activity of the free metallodrug. [32][33][34][35][36] Furthermore, it was found that 1 was not cytotoxic in any of the cell lines tested. Interestingly,w eo bserved am arked effect of 3 on the morphologyo fa V b 3 -overexpressing SK-MEL-28 cells and a V b 5 -overexpressing MDA-MB-231 cells. Indeed, the exposure of the cells to 1 or 3 induced morphological changes followed by cell detachment from the plate surface and as light decreasei nt he number of adherent cells. In contrast, no disruption of the morphology of the a V b 3 /a V b 5 À CAPAN-1 cells was observed at equimolar concentrations of the compounds (Figure 4). Thus, the similar behavior of 1 and 3 indicates that the conjugate is able to retain parto ft he activity of the a V b 3 antagonist peptide, which was previously found to cause detachmento f HMEC-1e ndothelial and glioma cells, leading to the induction of apoptosis by directa ctivation of caspase-8. [37] However,t he cell viability of SK-MEL-28, MDA-MB-231, and CAPAN-1 was not affected after treatment with either 1 or 3;t his might be attributed to the distinct origin of the cells tested. In fact, inhibition of cell adhesion has been found to be dependent on the matrix used in glioma cells, [38] whereas the effect on endothelial and tumor cells still remains unclear. [37] On the other hand, disruption of the integrin-mediated cell adhesion by 1 has been also linked to its antiangiogenic properties, [38] which prompted us to investigate the antiangiogenic activity of the conjugate (see below). The HUVEC cell line was chosen for the experiments because it constitutively expressed integrins involved in angiogenesis (a V b 3 , a V b 5 ,a nd a 5 b 1 ). [39] Prior to the experiments, the antiproliferativea ctivities of 1-3 in HUVECs were studied to subsequently work under sub-cytotoxicc oncentrations. As previously found for a V b 3 -a nd a V b 5 -overexpressing cell lines (SK-MEL-28 and MDA-MB-231, respectively), both 1 and 3 were found to be nontoxic (IC 50 > 100 mm;s ee Ta ble 1).

Angiogenesis inhibition studies
Considering that endothelial cell proliferation is essential in the multistep process of angiogenesis, in vitro vascular tube formation assay was measured2 4hafter incubating HUVECsw ith conjugate 3 at concentrationso f1 ,1 0, and 100 mm.D rug 1 and complex 2 were used for comparativep urposes and the negative control (untreated cells) wasc onsidered as 100 %o f capillaryformation.The resultsofthe two-dimensional Matrigel assay (Figure 5a nd Figures S6 and S7) were quantified as percentage capillaryf ormationb yc alculating the followingp arameters:n umber of master segments, number of meshes (network or polygonal structures), and number of master junctions. The formationo fc apillary-like structures was inhibited by complex 2 in ad ose-dependent manner( Figures S6 and  S7), although it should be noted that at high concentrations its cytotoxicity plays the central role (IC 50 = 4.61 mm;T able 1). In contrast, as light inhibitory effect on HUVECsw as observed for 1 and 3 under all concentrationst ested. In both cases,h igher activity was found at 10 mm,w ith no further enhancement at 100 mm.I nterestingly,t he effect of 3 ((18.47 AE 7.41) %; p > 0.05)   was higher with respect to 1 (0 %; p > 0.05) at the lower concentration( 1mm; Figure 5). However, it should be noted that the use of 1 at low doses can stimulate angiogenesis, whereas potent activity to block angiogenesis has been demonstrated in vivo. [40] In addition, the lack of activity of 1 in our in vitro model may be attributed to its inability to mimic crucial aspects of capillary morphogenesis in vivo. [41] Overall, the results are in agreement with the fact that peptidec onjugation alters the global mechanism of action of the free Pt II drug. In fact, peptidec onjugation supposes, as previously stated, the loss of anticancer activity, whereas the antiangiogenic properties seem to be closer to those of the a V b 3 antagonist peptides. The overall resultss uggest that the antiangiogenic properties of RGD-containing peptidesc an be modified through conjugation to metal complexesw ith high antiangiogenic activity.

Conclusions
We have described the synthesis and characterization of a novel conjugate between ac yclometalated platinum(II) complex with dual antiangiogenica nd antitumor activity and a cyclic RGD-containing peptideb yc ombining SPPS and solution-phase methodologies. Peptide conjugation resulted in a non-cytotoxic compound (IC 50 > 100 mm)i na ll tested tumor cell lines (AE a V b 3 and a V b 5 integrin receptors), which suggested that the mechanism of action of the parentP tc omplex had been modified. On the other hand, as imilar angiogenesis pattern of activity to that of reference drug 1 at sub-cytotoxic concentrations was found in HUVECs for 3,w hich opened the way to exploring the development of novel potentiala ngiogenesisi nhibitors through conjugation of compounds with high antiangiogenic activity,s uch as metal complexes or small organic molecules, [42] to cyclic RGD-containingp eptides or peptidomimetic analogues.

Experimental Section
General:C ommon chemicals and solvents (HPLC-grade or reagentgrade quality) were purchased from commercial sources and used  without further purification. Milli-Q water was directly obtained from aM illi-Q system equipped with a5 000 Da ultrafiltration cartridge. Peptide-grade DMF was from Scharlau. Fmoc-protected amino acids, resins, and coupling reagents for solid-phase synthesis were obtained from Novabiochem, Bachem, or Iris Biotech. Fmocl-Lys(Boc-AEEA)-OH was purchased from Iris. Solid-phase syntheses were performed manually in ap olypropylene syringe fitted with a polyethylene disk. Compounds 2 [24] and 5 [25,26] were synthesized and characterized as previously reported. All assayed compounds displayed ap urity of ! 95 %, as determined by HPLC analysis. Compound 1 was purchased from Bachem.
Integrin expression analysis:T he expression of a V b 3 and a V b 5 integrins on the cell surface was analyzed by means of double immunofluorescence. The cells were incubated for 30 min at 4 8C with monoclonal antibodies against human a V b 3 (clone LM609; Millipore, Te mecula, CA, USA) and a V b 5 integrin (sc-81632;S anta Cruz Biotechnology,S anta Cruz, CA, USA) or medium alone as a negative control. After washing with PBS (Gibco-BRL), cells were incubated for an additional 30 min at 4 8Cw ith the Alexa-Fluor 488conjugated goat anti-mouse IgG antibody (Invitrogen, Carlsbad, CA, USA). Next, the cell fluorescence was analyzed by using a FACSCalibur flow cytometer (Becton Dickinson Immunocytometry Systems, San Jose, CA, USA) equipped with CellQuest TM software (Becton Dickinson). 10 000 cells were analyzed in each experiment. Fluorescence intensity was represented on af our orders of magnitude log scale (1-10 000).
Cell viability studies:T he cytotoxicity of the compounds in SK-MEL-28, CAPAN-1, MDA-MB-231, and HUVECs was determined by the MTT tetrazolium reduction assay.C ompounds 1 and 3 were dissolved in Milli-Q water to obtain 1mm stock solutions. Cisplatin (Sigma-Aldrich, St. Louis, MO) and 2 were dissolved in DMSO to provide a1 0mm solution, which was then diluted in Milli-Q to obtain a1m m stock solution. Appropriate aliquots of these solu- ChemMedChem 2018ChemMedChem , 13,1755ChemMedChem -1762 www.chemmedchem.org 2018 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim tions were diluted in the cell culture medium to obtain the final working solutions. Aliquots of 5000 CAPAN-1, 4000 SK-MEL-28, 5000 MDA-MB-231, and 2500 HUVECs were seeded on 96-well plates 24 hp rior to the treatments. Then, cells were treated for 48 hw ith the corresponding compound at concentrations ranging from 0t o1 00 mm.A fter removal of the treatment, cells were washed with PBS. Before the MTT colorimetric assay,r epresentative images (10 )o ft he cells exposed to the each treatment were obtained with aC KX41 inverted microscope (Olympus). Then, cells were incubated for 2h in the dark with fresh culture medium (100 mL) containing MTT (10 mL). The medium was discarded and DMSO (100 mL) was added to each well to dissolve the purple formazan crystals. Plates were agitated at RT for 2min and the absorbance of each well was determined with an absorbance microplate reader (EL 800, BioTek, Winooski, VT,U SA) at aw avelength of 570 nm. Three replicates for each compound were used, and all treatments were tested at least in three independent experiments. For each treatment, the cell viability was determined as ap ercentage of the control untreated cells, by dividing the mean absorbance of each treatment by the mean absorbance of the untreated cells. The IC 50 was established for each compound by using af ourparameter curve fit (Gen5 Data Analysis Software, BioTek).
Tube-like formation assay:T oa ddress the role played by the conjugation of the Pt complex to 1 in the angiogenesis process, the tube-like formation assay was used. Separate experiments were performed by using Matrigel Basement Matrix Growth Factor Reduced (Corning 356231) as as ubstrate for HUVECs. Cells (2 10 4 in complete medium) were seeded onto a9 6-well culture plate. Different concentrations of 1-3 (1, 10, or 100 mm)w ere added to the wells and incubated at 37 8Ci nahumidified 5% CO 2 atmosphere for 24 h. Then, images of each well were acquired at am agnification of 10 and analyzed by using the Angiogenesis Analyzer for ImageJ from Gilles Carpentier.D ata analysis was performed by normalizing the results with the negative control (untreated cells), which was considered as 100 %o fc apillary formation.
Statistical analysis:M ean values, expressed as mean AE standard error of the mean, were compared by using the two-tailed nonparametric Mann-Whitney U-test (GraphPad Prism software, version 5.04 for Windows) and as tatistically significant value was taken as p < 0.05.