Polyhedron 80 (2014) 134–141

Contents lists available at ScienceDirect

Polyhedron

journal homepage: www.elsevier .com/locate /poly

Synthesis, spectral characterization, cytotoxicity and enzyme-inhibitingactivity of new ferrocene–indole hybrids

http://dx.doi.org/10.1016/j.poly.2014.03.0060277-5387/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +381 18533015; fax: +381 18533014.E-mail address: nikoradulovic@yahoo.com (N.S. Radulovic).

Niko S. Radulovic a,⇑, Dragan B. Zlatkovic a, Katarina V. Mitic a, Pavle J. Randjelovic b, Nikola M. Stojanovic c

a Department of Chemistry, Faculty of Science and Mathematics, University of Niš, Višegradska 33, RS-18000 Niš, Serbiab Department of Physiology, Faculty of Medicine, University of Niš, Bulevar dr Zorana Ðin -dica 81, RS-18000 Niš, Serbiac Faculty of Medicine, University of Niš, Bulevar dr Zorana Ðin -dica 81, RS-18000 Niš, Serbia

a r t i c l e i n f o

Article history:Received 11 January 2014Accepted 4 March 2014Available online 14 March 2014

Dedicated to Professor Vukadin Leovac onthe occasion of his 70th birthday.

Keywords:IndoleFerroceneArtemia salinaAcetylcholinesterase inhibitionMyeloperoxidase activity

a b s t r a c t

Incorporation of a ferrocene moiety into the 2-phenylindole scaffold has been recently reported to dras-tically improve the cytotoxic activity of the parent compounds. In our search for new promising cytotoxicagents we designed and prepared two new ferrocene–indole hybrids, 2-(3-ferrocenylphenyl)-1H-indoleand 2-(4-ferrocenylphenyl)-1H-indole, utilizing the Fischer indole synthesis as the key step. Detailedspectral analyses, including 1D and 2D NMR in various solvents, have been carried out to corroboratethe structures of the synthesized compounds. In this work, we also present the first results of biologicalstudies of these two compounds. Both compounds showed weak anticholinesterase activity but highcytotoxicity against rat peritoneal macrophages and the crustacean Artemia salina. Also, both compoundsshowed significant myeloperoxidase inhibiting activity, thus suggesting a potential use in inflammatorydisorders. The results of these tests are very encouraging as they also suggest possible cytotoxic activitiesof these compounds against human cancer cell lines.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Today a growing number of medicinal and synthetic chemistsdevote their time and effort in the search for new and/or more po-tent pharmacologically active compounds that might eventuallyrepresent therapeutic agents in the combat against human ill-nesses of the modern world. It is not enough to arrive at a com-pound possessing the desired activity but we need a compoundthat exerts this beneficial effect without being toxic in the applieddose [1]. Hence, it is paramount to know the pharmacological win-dow of the potential drug and this starts by determining its toxicityversus targeted activity. Among a vast number of possible biologi-cal assays, one can argue that a ‘‘screening’’ that includes cytotox-icity tests against (in vivo) an animal species and (in vitro)specialized cell culture of a higher organism (murine), as well asan in vitro enzyme inhibiting assays should convey a wide spec-trum of relevant data. For example, the brine shrimp cytotoxicityassay, cytotoxicity against rat peritoneal macrophages and acetyl-cholinesterase inhibition test would nicely suit this screeningpurpose.

Artemia salina is one of the standard organisms for testing thecytotoxicity of chemicals [2,3]. The Artemia bioassay is attractive

to researchers due to the commercial availability and possible longstorage of the cysts, quickness of the assay and since it complieswith animal ethics guidelines in many countries [3].

Macrophages are a heterogeneous population of mononuclearphagocytes present in all tissues of the body. They play a crucialrole in regulating and executing most homeostatic, immunologicaland inflammatory processes [4–6]. Tissue macrophages are thefirst line of defense against infection by pathogens prior to themigration of polymorphonuclear neutrophils and monocytes. Gen-erally, macrophages’ major functions include engulfing, degrada-tion of self or foreign materials and the regulated production ofinflammatory mediators such as pro-inflammatory cytokines,prostaglandins and reactive oxygen and nitrogen species. The pro-duction of oxygen species, often referred to as respiratory burst, isa tightly regulated mechanism, involving NADPH-oxidase andmyeloperoxidase (MPO) as key enzymes. Previous studies haveshown that MPO, like enzyme found in all phagocytes (neutrophilsand macrophages) [7], exerts both antimicrobial and cytotoxicproperties [8,9]. MPO reacts with hydrogen peroxide (H2O2) con-verted from the extra oxygen consumed in the respiratory burstto form a complex that can oxidize a large variety of substances.Among the latter are the chloride ions, which are oxidized to hypo-chlorous acid, with the subsequent formation of chlorine andchloramines. These products are powerful oxidants that haveimportant roles in host defense by destroying a variety of targets

N.S. Radulovic et al. / Polyhedron 80 (2014) 134–141 135

including bacteria, fungi and viruses. However, the oxidant activityof MPO might contribute to tumor initiation as it is known to causeDNA damage [10]. Namely, MPO is released by neutrophils andmacrophages, which are known to invade and infiltrate tumors,and this enzyme can be a useful specific marker of phagocytes infil-tration in premalignant and neoplastic lesions [11]. On the otherhand, macrophages are, also, professional antigen-presenting cellsand play an important role in inducing of T cell activation and reg-ulating the adaptive immune response [12].

Acetylcholinesterase (AChE) is an enzyme that hydrolyzes ace-tylcholine at neuromuscular junctions and cholinergic synapsesin the brain [13]. AChE activity controls the transmission of nerveimpulses and prevents continuous activation of postsynaptic cho-linergic receptors. It is vital in maintaining normal function ofthe entire nervous system [14]. Interestingly, the expression ofAChE is not restricted to cholinergic nervous tissues and can evenbe found in different types of tumors [15]. Moreover, AChE mayhave some basic functions such as cell differentiation and celladhesion [16].

Modern organometallic chemistry started with the recognitionof the structure of ferrocene [17,18]. Ferrocene or its derivativesare used, among other things, as non-toxic anti-knocking fuel addi-tives [19], as ligand scaffolds [20] and in the synthesis of carbonnanotubes [21]. It has gained a lot of interest in medicinal chemis-try and is frequently used as a substituent for the phenyl or alkylgroups [22]. There are several traits that make ferrocene an attrac-tive building block: lipophilicity (it is more lipophilic than ben-zene, thus it is expected that ferrocene derivatives have greaterbioavailability than the parent phenyl compounds), stability (dueto an 18 p electron configuration of the iron(II) center) and the rep-utation of being a ‘‘safe’’ metallocene (unsubstituted ferrocene is aparticularly non-toxic compound–abnormally high doses of ferro-cene were tested on dogs for 3 months in the study by Yeary[23]. There was no observed acute toxicity even at the doses of1 g kg�1).

It is impossible to predict the effect of incorporation of ferro-cene in molecules that possess biological activity. For example,Loev and Flores [24] reported the first synthesis of ferrocenylatedpharmacologically active compounds – yet the prepared ferroceneanalogues of amphetamine and phenytoin showed no activity atall. Fortunately, a number of studies exist where ferrocene intro-duction improved the activity of the parent compound, or changedthe activity profile [22]. Ferroquine is 35 times more active thanchloroquine against drug-resistant strains of Plasmodium falcipari-um [25]. In addition to the primary mechanism of quinoline action,fluorescent probe studies in infected red blood cells showed an-other mechanism based on the production of HO� (by the metall-ocenic moiety) at work [26,27]. Inclusion of ferrocene in theselective estrogen receptor modulator tamoxifen (anti-tumoragent) is reported to have fixed some of the shortcomings of theparent compound. For other examples of successful ferrocene-arylsubstitutions, the reader is referred to a review by Fouda et al. [28].

The indole nucleus is present in an array of compounds possess-ing interesting biological activities [29]. There are a very limitednumber of reported compounds that contain both the indole andferrocene moieties [30,31], which is surprising due to an expectedcontribution of both structural fragments to the overall possibleactivity of such hybrids. Prompted by the lack of such chemical/biological investigations and motivated by the work of Quiranteet al. [31], we recently engaged in a study of the synthesis and bio-logical/toxicological evaluation of ferrocene–indole hybrids. Theauthors of the aforementioned paper noted that the 2-phenylin-dole scaffold represented a potent antimitotic agent and prepareda series of 3-ferrocenylmethyl derivatives – two of the synthesizedcompounds showed 25-fold increase in the cytotoxic activity com-pared to 2-phenylindole. While Quirante et al. [31] introduced a

ferrocene moiety in the position 3 of the indole ring, we decidedto prepare a ferrocene–indole hybrid tethered by an aromatic ring.We arrived at the first amounts of these new compounds using thetraditional Fischer indole synthesis. Detailed spectral analyses,including 1D and 2D NMR in various solvents, have been carriedout to corroborate the structures of the synthesized compounds.

Hence, in this work we report the first results on the biological/pharmacological activity of two new ferrocene–indole hybrids,2-(3-ferrocenylphenyl)-1H-indole and 2-(4-ferrocenylphenyl)-1H-indole. We evaluated the biological activity of the two compoundsthrough a cytotoxicity assay using A. salina and AChE inhibitoryactivity. Also, the direct influences of ferrocene–indole hybrids oncells involved in the inflammatory reactions were studied bymeans of viability and MPO enzymatic activity in rat peritonealmacrophages.

2. Results and discussion

2.1. Synthesis

Ferrocenylacetophenones 2a and 2b were prepared by couplingof ferrocene with appropriate diazonium salts (Scheme 1). A reac-tion of these ketones with phenylhydrazine yielded hydrazoneswhich were directly subjected to a Fischer indolization to givecompounds 3a and 3b. Fisher indolization has not been performedon a ferrocene-containing hydrazone thus far. Hence, this is thefirst report of a de novo synthesis of indoles in the presence of a fer-rocene moiety. While Fischer indole synthesis is known to be cat-alyzed be various acids [29], only polyphosphoric acid gavereasonable yields. The use of conc. H2SO4, p-toluenesulfonic acid,glacial acetic acid and BF3� Et2O gave no desired product at all. Per-haps this is the reason why previous reports are lacking. Also,please note that we tried to synthesize the third regioisomer (2-(2-ferrocenylphenyl)-1H-indole) but failed – only traces of theproduct were detected. In spite of the general notion that the fer-rocene moiety is stable under acidic conditions (an array of ferro-cene derivatives are obtained in acidic media), the combination ofhigh temperature and high acidity needed to drive the Fischer in-dole synthesis turned out to be detrimental to the metallocene(we observed a high degree of decomposition to inorganic ironcompounds during these reactions). Although the reaction yieldswere poor (ca. 20%), this synthetic methodology provides a meansof acquiring such new structures of ferrocene–indole hybrids-theferrocene containing substituent is bonded through carbon-2 ofthe indole core and tethered by an aromatic ring, whereas only3-ferrocenyl substituted indoles and non-tethered derivatives havebeen synthesized so far [30,31]. Currently, new higher yieldingsynthetic approaches are being tried out in our laboratory.

2.2. Spectral characterization

Mass spectrometry of the two synthesized compounds con-firmed their molecular formula (C24H19FeN). The mass spectra(fully shown in Supplemental material) were dominated by themolecular ion and its isotopic ions (expected for compounds withhigh degree of aromaticity). The monoisotopic mass of the molec-ular ion (377) corresponded to the masses of ions containing onlythe most abundant isotopes (12C, 1H, 56Fe and 14N). The spectrawere also characterized by high intensities of [M�2] ions, a resultof the presence of 54Fe atoms. Experimentally determined intensi-ties of ions (from [M+3] to [M�2]) correlated well with the theo-retically calculated values for a compound of the molecularformula C24H19FeN. Fragment ions were of very low intensities,but the ion at m/z 56 (Fe+) found in almost all ferrocene-containingcompounds was present in both of the compounds.

Scheme 1.

136 N.S. Radulovic et al. / Polyhedron 80 (2014) 134–141

The indoles 3a and 3b were fully characterized by 1D and 2DNMR experiments. The NMR spectra were recorded in various sol-vents and the chemical shifts are listed in Table 1 (the carbon num-bering is shown in Scheme 1). Assignment of 13C and 1H chemicalshifts was accomplished through the use of 2D spectra (HSQC,HMBC, 1H–1H-COSY and NOESY).

2.3. NMR analysis of 3a dissolved in DMSO-d6

1H NMR spectra of both compounds showed the usual patternof monosubstituted ferrocene (Fc) systems–two magneticallyinequivalent ortho (H-200 and H-500) protons at d 4.91 downfieldfrom the signals of the two meta (H-300 and H-400) protons at d4.40, and a strong signal (5H, d 4.07) due to the five equivalenthydrogens of the unsubstituted cyclopentadienylidene (Cp) ring.None of the Fc protons had any HMBC (Fig. 1) correlations withthe carbon atoms from the rest of the molecule, but the NOESYspectrum (Fig. 1) revealed a correlation (and spatial proximity) be-tween the signals at d 4.91 (H-200) and aromatic signals at d 7.50and 8.02. No such cross-peaks were observable for the proton atd 4.40 (H-300), thus, confirming that the downfield protons were in-deed in the ortho position.

Also, the presence of 1,3-disubstituted benzene ring could beeasily observed in the spectra. H-20 signal appeared as a pseudo-triplet, coupled with H-40 and H-60 with almost equal small cou-pling constants (1.6 Hz), indicating that the positions ortho to thisproton were substituted. A (pseudo-triplet) signal at d 7.39 (H-50)was split with larger J values (7.6 and 7.8 Hz), indicative of twoneighboring ortho protons, and finally protons at d 7.50 (H-40)and 7.70 (H-60) were both doublets of doublets of doublets (withone large and two small J values). These two protons could be dif-ferentiated by NOESY – H-40 correlated with the protons of Fcgroup, whereas H-60 correlated with the signals of the indole moi-ety (Fig. 1).

The assignment of the five protons of the indole core was alsostraightforward. H-3 proton was shielded by the conjugated nitro-gen electron pair, in agreement with the low dC (99.3 ppm) value ofthe carbon directly attached. C-3 carbon correlated with the signalat d 7.57 which must belong to H-9. All other CH protons in themolecule were located more than three bonds from C-3. H-9 is cou-pled to H-8 (ddd). The remaining two protons were at d 7.45 (H-6,dd, correlates with the NH proton in NOESY, Fig. 1), and 7.13 (H-7,ddd).

The protonated C-atoms were easily assigned with the help ofthe HSQC spectrum. Quaternary carbons had to be assigned fromtheir HMBC correlations (Fig. 1). Carbons at 129.2 ppm (C-4) and137.6 ppm (C-5) correlated with H-8 and H-7, respectively. Fromthe remaining three unassigned carbons, only the signal at d138.2 ppm did not correlate with H-50, so this carbon should be

in the indole ring (C-2). The last two carbons (C-10 and C-30) unas-signed could not be readily distinguished since they only displayedone strong cross peak with H-50 (through three bonds) in the HMBCspectrum (Fig. 1). This is why we decided to measure the NMRspectra of this compound in other solvents to be able to assign sig-nals to all of the carbons in an unambiguous manner. In an analo-gous way, we analyzed 1D and 2D spectra of 3a recorded in CDCl3,CD3CN and CD3OD, where weaker two-bond couplings of the prob-lematic carbons (C-10 with H-60 and C-30 with H-40) were observa-ble. Thus, we succeeded in the task and the full assignations aregiven in Table 1.

2.4. NMR analysis of 3b in DMSO-d6

The para isomer had a much less complicated 1H NMR spectrumdue to the symmetry of a AA0BB0-coupling pattern of the centralbenzene core. H-20 and H-30 could again be differentiated fromthe NOESY (H-30 correlated with Fc protons) or HMBC (H-30 corre-lated with the only quaternary Fc carbon), as summarized in Fig. 1.Most of the carbon (and all proton) resonances were readily assign-able through a coupling pattern from the 2D spectra (especiallyHMBC) as in the case of 3a. The assignation of three carbon signals(130.1, 138.3 and 138.7 ppm) turned out to be less straightforwarddue to a limited number of observed HMBC cross-peaks. We recog-nized the noted HMBC interactions (130.1 ppm with 7.60 ppm, and138.7 ppm with 7.77 ppm) as couplings between carbon and pro-tons across three bonds (as opposed to smaller two-bond cou-plings) and allocated 130.1 ppm to C-10 and 138.7 ppm to C-40.That meant that the only remaining carbon signal at 138.3 ppm be-longs to C-2 of the indole ring. The value of this chemical shift is inaccordance with that of the corresponding carbon in the meta iso-mer 3a. We also tried to record the NMR spectra of this isomer inother solvents but failed owing to the poor solubility of 3b in theavailable to us deuterated solvents other than DMSO.

2.5. Biological activity

2.5.1. Evaluation of acute toxicity against A. salinaBoth compounds exhibited strong cytotoxic activity against A.

salina. Lower LD50 were observed for the meta-isomer 3a(20.7 lM and <13.3 lM after 24 and 48 h exposures, respectively)than the para-isomer 3b (24 h exposure: LD50 = 120.4 lM; 48 hexposure: LD50 = 66.5 lM). Some ferrocene-containing compoundsare already known to exhibit strong cytotoxic activity towards A.salina [32,33], and the observed cytotoxicity of 1 and 2 can beprobably attributed to the presence of both 2-phenylindole andferrocene moieties. Activity of the ferrocene compounds are usu-ally explained by the formation of cytotoxic ferrocenium-ionswhich is distinguished by both great solubility in water and the

Table 11H and 13C NMR chemical shifts of 2-(3-ferrocenylphenyl)-1H-indole (1) and 2-(4-ferrocenylphenyl)-1H-indole (2).

Assignment 2-(3-Ferrocenylphenyl)-1H-indole 2-(4-Ferrocenylphenyl)-1H-indole

Type of carbon DMSO-d6 Chloroform-d Acetonitrile-d3 Methanol-d4 Type of carbon DMSO-d6

13C 1H 13C 1H 13C 1H 13C 1H 13C 1H

C-2 Cq 138.2 138.1 140.3 138 Cq 138.3C-3 CH 99.3 6.98

J3,6 = 0.7 HzJ3,1 = 2.1 Hz

100.0 6.91J3,6 = 0.9 HzJ3,1 = 2.1 Hz

99.1 6.92J3,6 = 0.8 HzJ3,1 = 2.2 Hz

98.3 6.86J3,6 = 0.7 Hz

CH 98.6 6.86J3,6 = 0.7 HzJ3,1 = 2.1 Hz

C-4 Cq 129.2 129.4 129.1 129.2 Cq 129.3C-5 Cq 137.6 136.9 137.2 137.5 Cq 137.6C-6 CH 111.7 7.45

J6,7 = 8.1 HzJ6,3 = 0.7 HzJ6,8 = 1.9 Hz

110.9 7.46J6,7 = 7.9 HzJ6,3 = 0.9 HzJ6,8 = 1.9 Hz

111.1 7.50J6,7 = 8.1 HzJ6,3 = 0.9 HzJ6,8 = 1.9 Hz

110.7 7.44J6,7 = 8.1 HzJ6,3 = 0.7 HzJ6,8 = 1.7 Hz

CH 111.6 7.41J6,7 = 8.0 HzJ6,3 = 0.7 HzJ6,8 = 1.7 Hz

C-7 CH 122.0 7.13J7,9 = 1.2 HzJ7,8 = 7.1 HzJ7,6 = 8.1 Hz

122.4 7.25J7,9 = 1.2 HzJ7,8 = 7.2 HzJ7,6 = 7.9 Hz

122.2 7.20J7,9 = 1.2 HzJ7,8 = 7.1 HzJ7,6 = 8.1 Hz

121.3 7.12J7,9 = 1.2 HzJ7,8 = 7.1 HzJ7,6 = 8.1 Hz

CH 121.8 7.09J7,8 = 7.1 HzJ7,6 = 8.0 HzJ7,9 = 1.2 Hz

C-8 CH 119.8 7.03J8,6 = 1.0 HzJ8,7 = 7.1 HzJ8,9 = 7.9 Hz

120.3 7.20J8,6 = 1.9 HzJ8,7 = 7.2 HzJ8,9 = 7.9 Hz

119.8 7.09J8,6 = 1.9 HzJ8,7 = 7.1 HzJ8,9 = 8.0 Hz

119.1 7.02J8,6 = 1.7 HzJ8,7 = 7.1 HzJ8,9 = 7.9 Hz

CH 119.7 7.00J8,7 = 7.1 HzJ8,9 = 7.9 HzJ8,6 = 1.7 Hz

C-9 CH 120.5 7.57J9,8 = 7.9 HzJ9,7 = 1.2 Hz

120.7 7.71J9,8 = 7.9 HzJ9,7 = 1.2 Hz

120.2 7.62J9,8 = 8.0 HzJ9,7 = 1.2 Hz

119.8 7.57J9,8 = 7.9 HzJ9,7 = 1.2 Hz

CH 120.3 7.53J9,8 = 7.9 HzJ9,7 = 1.2 Hz

C-10 Cq 132.7 or 140.2 132.5 132.4 132.8 Cq 130.1C-20 CH 122.6 8.02

J20 ,40 = 1.6 HzJ20 ,60 = 1.6 Hz

122.9 7.79J20 ,40 = 1.6 HzJ20 ,60 = 1.6 Hz

122.4 7.93J20 ,40 = 1.7 HzJ20 ,60 = 1.7 Hz

122.0 7.96J20 ,40 = 1.6 HzJ20 ,60 = 1.6 Hz

CH 125.5 7.78J20 ,30 = 8.3 Hz

C-30 Cq 132.7 or 140.2 140.4 140.3 140.1 CH 126.6 7.60J30 ,20 = 8.3 Hz

C-40 CH 125.4 7.50J40 ,50 = 7.8 HzJ40 ,60 = 1.2 HzJ40 ,20 = 1.6 Hz

125.7 7.50J40 ,50 = 7.9 HzJ40 ,60 = 1.2 HzJ40 ,2 = 1.6 Hz

125.3 7.54J40 ,50 = 7.8 HzJ40 ,60 = 1.2 HzJ40 ,2 = 1.7 Hz

124.7 7.48J40 ,50 = 7.7 HzJ40 ,60 = 1.1 HzJ40 ,2 = 1.6 Hz

Cq 138.7

C-50 CH 129.4 7.39J50 ,40 = 7.8 HzJ50 ,60 = 7.6 Hz

129.0 7.40J50 ,40 = 7.9 HzJ50 ,60 = 7.4 Hz

129.0 7.42J50 ,40 = 7.8 HzJ50 ,60 = 7.4 Hz

128.5 7.37J50 ,40 = 7.7 HzJ50 ,60 = 7.7 Hz

C-60 CH 123.1 7.70J60 ,50 = 7.6 HzJ60 ,40 = 1.2 HzJ60 ,20 = 1.6 Hz

122.8 7.51J60 ,50 = 7.4 HzJ60 ,40 = 1.2 HzJ60 ,20 = 1.6 Hz

122.8 7.64J60 ,50 = 7.4 HzJ60 ,40 = 1.2 HzJ60 ,20 = 1.7 Hz

122.4 7.63J60 ,50 = 7.7 HzJ60 ,40 = 1.1 HzJ60 ,20 = 1.6 Hz

C-100 Cq 85.0 85.0 84.8 84.9 Cq 84.7C-200 CH 66.9 4.91 66.7 4.76 66.5 4.83 66.3 4.81 CH 66.6 4.82C-300 CH 69.4 4.40 69.1 4.41 69.1 4.39 68.7 4.38 CH 69.5 4.37C-1000 CH 69.9 4.07 69.8 4.13 69.6 4.07 69.2 4.07 CH 69.8 4.04NH Het-H 11.50 (broad s) 8.35 (broad s) 9.85 (broad s) / Het-H 11.42

Fig. 1. Key HMBC and NOESY correlations in 3a and 3b.

N.S. Radulovic et al. / Polyhedron 80 (2014) 134–141 137

presence of two cyclopentadienyl rings which ensures penetra-tions of membranes [34]. The antitumor effect of 2-phenylindolescaffold was reported many times in the literature (for more infor-mation, refer to Quirante et al. [31] and the references cited there-in). It is also possible that the presence of both the indole andferrocene cores results in a synergistic cytotoxic interaction. Theobservation of Neuse [35] that water solubility should play oneof the key roles when developing new Fc-based anti-cancer agents

is in agreement with this synergism – the polar indole ring mightcompensate for the less hydrophilicity of the Fc group.

2.5.2. AChE (acetylcholinesterase) inhibitory activityInhibition of AChE leads to accumulation of the neurotransmit-

ter acetylcholine (ACh) in cholinergic synapses and neuromuscularjunctions, which causes over-stimulation of cholinergic receptorsand results in cholinergic toxicity [36]. This cholinergic toxicity in-cludes amplified sweating and salivation, diarrhea, increased bron-chial secretion, miosis, muscular cramps and other central nervoussystem effects [37]. Also it is interesting to note that cytotoxicitytests on immortalized and primary cell lines derived from lung tu-mor showed an AChE inhibition-dependent selective reduction ofcell viability, statistically significant [38].

Compound 3a exhibited a weak anti-AChE activity with an IC50

value of 255 lM as compared to the reference drug physostigmine(0.30 lM). Solubility issues limit the available dose of the para-iso-mer (3b). The saturated solution (ca. 100 lg/ml in the medium) ofthis compound did not have any appreciative effect on the activityof AChE, thus the IC50 value for the para-isomer (3b) could not bedetermined. These results suggest that the tested substances donot have a potential toxic anti-AChE activity which could lead toundesirable side effects.

138 N.S. Radulovic et al. / Polyhedron 80 (2014) 134–141

It is interesting to notice that physostigmine, compound used asa positive control in our tests belongs to the class of indoles, butshowed 1000-fold higher activity than 3a. As a matter of fact, in-dole alkaloids are well-known for their cholinesterase inhibitingactivities [39]. One possible explanation for the observed lack ofactivities is that the inclusion of the bulky ferrocene group intothe indole scaffold prevented the formation of strong bindinginteractions between the enzyme and the substrate in the activecenter, but this hypothesis needs to be confirmed by additionalin vitro and/or in silico studies.

Fig. 2. The influence of various concentrations of ferrocene–indole hybrids (3a and3b) on rat peritoneal macrophage viability, expressed as % of cell viability (a) and asODs (595 nm) � 1000 (b). The number of cell samples per treatment was eight.Statistically significant differences (t-test): #p < 0.0001, 0 (PBS) vs. 3a or 3bconcentrations; ap < 0.0001, 0 (PBS) vs. 3a concentrations; bp < 0.0001, 0 (PBS) vs.3b concentrations.

Fig. 3. The in vitro effects of ferrocene–indole hybrids (3a and 3b) on rat peritonealmacrophage MPO activity. The number of cell samples per treatment was eight.Statistically significant differences (t-test): ⁄p < 0.05 vs. 0 (PBS).

2.5.3. Evaluation of cytotoxity against macrophages of 3a and 3b andthe corresponding macrophage MPO activities

To explore the in vitro cytotoxic activity of new ferrocene–in-dole hybrids, rat peritoneal macrophages were treated with thesecompounds for 1 h, and cell viability was determined in a crystalviolet assay. Such a short period was sufficient for 3a and 3b to ex-ert significant cytotoxic action. Namely, macrophage viability wasreduced by more than 50% in the presence of the lowest tested con-centrations of these compounds (IC50 was in the range 0.5–1.0 lM)(Fig. 2a), that was in correlation with the measured ODs (Fig. 2b).Additionally, the observed cytotoxic effect was dose-dependentand higher concentrations of 3a and 3b markedly reduced the via-bility of peritoneal macrophages in comparison with the control(PBS).

Both 3a and 3b, at concentration of 5 and 1 lM, respectively,decreased the MPO enzymatic activity in rat peritoneal macro-phages in a statistically significant manner (Fig. 3). MPO inhibitingactivity of indole compounds is well covered in literature [40,41],and the accepted mechanism of action is that the indoles act as1e� substrates for the reactive intermediate MPO–Fe(IV) species.However, a possible contribution of the Fc moiety to the MPO-inhibiting activities of 1 and 2 should not be neglected – ferroceneis easily oxidized to the ferrocenium (Fe3+) state, so it can reactwith MPO–Fe(IV) and prevent the formation of HOCl. This is, how-ever, purely speculative since, to our knowledge, no studies of thepossible use of ferrocene in MPO inhibition exist.

In our experiments peritoneal macrophages were used as amodel for studying direct effects of ferrocene–indole hybrids oninflammatory cell viability and their activity. Physiological signifi-cance of ferrocene–indole hybrids-induced modulation of macro-phage functions could be in regulation of an inflammatoryreaction. This study has provided evidence that 3a and 3b dimin-ished both peritoneal macrophage viability and the MPO enzy-matic activity. In fact, if we consider the correlation with in vivoconditions, these synthetic hybrids could lead macrophages intocell death significantly reducing the number of mature immunecells and contribute to the inefficient inflammatory response. Onthe other hand, the diminished MPO activity also abrogates a suc-cessful immune reaction against microbe infections and pathogenelimination. These findings strongly support the assumption thatferrocene–indole hybrids could modulate macrophage activities,implying their immunosuppressive properties. The results reachedby several research groups confirmed that a higher MPO expres-sion level can be correlated with an increased risk of various typesof cancer, such as hepatoblastoma [42], urinary bladder cancer [43]and ovarian cancer [44]. These observations additionally corrobo-rate our hypothesis that these new ferrocene–indol hybrids canbe regarded as potential anti-cancer drugs.

In conclusion, our current in vitro data suggest that these com-pounds exert the influence on tissue macrophage functions proba-bly via a non-receptor/receptor interplay, which is dependent onthe local microenvironment, xenobiotic metabolism andmacrophage activation when correlated to the in vivo conditions.In the future our further experiments will be focused on in vivo

investigations of ferrocene–indole hybrids immunomodulatorycapacity(ies).

3. Conclusions

Two new ferrocene–indole hybrids were tested in four assays(AChE inhibitory activity, brine shrimp lethality assay and macro-phage viability and MPO assays) in order to screen for their biolog-ical activity. Although the compounds exerted weak inhibition ofacetylcholinesterase (IC50 values: 255 lM for the meta and>265 lM for the para isomer), they demonstrated significant cyto-toxic activities against A. salina (24 h LD50 values were 20.7 lM for

N.S. Radulovic et al. / Polyhedron 80 (2014) 134–141 139

the meta and 120.4 lM for the para substituted compounds) andrat macrophages (IC50 around 0.5 lM for both compounds). Sucha pharmacological profile gives justification for further studies oftheir chemistries and biological properties (especially cytotoxicactivity on human cancer cell lines).

4. Materials and methods

4.1. Synthesis and spectral characterization

4.1.1. GeneralAll chemicals were commercially available and used as

received, except that the solvents were purified by distillation.Preparative medium-pressure liquid chromatography (MPLC)separations were performed with pre-packed silica gel (40–63 lm) polypropylene cartridges under isocratic conditions usinga Büchi system (C 610, Büchi, Switzerland). TLC was performedon Merck silica gel plates, layer thickness 0.2 mm with 60 F254indicator. Visualization was accomplished with UV light and phos-phomolybdic acid (10% w/v in ethanol). Mass spectra were re-corded on a Hewlett Packard 5975B mass selective detectorcoupled with a Hewlett Packard 6890N gas chromatographequipped with a fused silica capillary column DB-5MS (5% phen-ylmethylsiloxane, 30 m � 0.25 mm, film thickness 0.25 lm, AgilentTechnologies, USA). 1H and 13C NMR spectra were recorded on aBruker Avance II spectrometer (Bruker, Germany) operating at400 and 100.6 MHz, respectively. 2D experiments (1H–1H COSY,HSQC, HMBC and NOESY) were run on the same instrument withthe usual pulse sequences. All NMR spectra were measured at25 �C in CDCl3, CD3OD, CD3CN and CD3SOCD3 with TMS as internalstandard.

4.1.2. Syntheses of 3- and 4-ferrocenylacetophenones (2a and 2b)The diazonium salts of 3- and 4-aminoacetophenones were pre-

pared by the slow addition of ice cold sodium nitrite solution(1.30 g (18.8 mmol) of NaNO2 in 10 ml of water) to the stirred solu-tions of the corresponding amine (2.0 g, 14.8 mmol) in aqueousH2SO4 (1 M, 37.5 ml). The resulting solution was added to a cooledand stirred suspension of ferrocene (3.50 g, 18.8 mmol) in glacialacetic acid (200 ml) under inert atmosphere. The solution was stir-red at room temperature for 24 h, and then poured to 400 ml ofwater. The mixture was extracted with 3 � 50 ml of dichlorometh-ane, and the combined extracts were neutralized with saturatedsodium carbonate solution. After the separation, CH2Cl2 phasewas dried (with anh. MgSO4) and concentrated in vacuo. The prod-ucts (2a and 2b) were purified by MPLC using dichloromethane asan eluent. Yields for 2a and 2b were 1.51 g (33.4%) and 1.31 g(26.8%), respectively. NMR spectra of the obtained pure substancescorresponded to literature values [45].

EI-MS m/z (rel. intensity, %) of 2a: 304 (100, M), 305 (22.3, M+1),139 (9.2), 261 (7.8, M�Ac), 302 (7.5, M�2), 121 (7.1), 144 (6.7), 260(6.0), 56 (4.9, Fe), 203 (4.3).

EI-MS m/z (rel. intensity, %) of 2b: 304 (100, M), 261 (22.1,M�Ac), 305 (16.4, M+1), 144 (8.9), 139 (6.9), 302 (6.5, M�2), 260(5.9), 121 (5. 2), 262 (4.4), 56 (4.1, Fe).

4.1.3. Indolization of 2a and 2bTwo drops of acetic acid, 0.72 g (6.66 mmol) of phenylhydr-

azine, 1.00 g (3.29 mmol) of 2a or 2b and 10 ml of ethanol werecombined and refluxed for 2 h and the resulting slurry was evapo-rated to dryness. The raw hydrazone intermediate was dried onhigh vacuum for 30 min and then 10 ml of polyphosphoric acidwas added. The mixture was heated at 100 �C for an hour, cooledto room temperature and quenched by a careful addition of50 ml of ice cold water. After the extraction with dichloromethane

(3 � 20 ml), the combined organic phases were washed with H2O,dried over anh. MgSO4 and the solvent evaporated. The products(3a and 3b) were purified by MPLC using hexane/CH2Cl2 (3:1,v/v) as an eluent. Yields for 3a and 3b were 270 mg (21.8%) and188 mg (15.2%), respectively.

EI-MS m/z (rel. intensity, %) of 3a: 377 (100, M), 378 (28.6, M+1),188 (11.9), 254 (10.1), 375 (8.7, M�2), 121 (5.6), 376 (4.3, M�1),379 (4.2, M+2).

EI-MS m/z (rel. intensity, %) of 3b: 377 (100, M), 378 (29.5,M+1), 256 (10.0), 121 (8.5), 375 (8.3, M�2), 254 (7.1), 379 (6.2,M+2), 376 (4.1, M�1).

For 1H and 13C NMR see Table 1.

4.2. Biological studies

4.2.1. GeneralThioglycolate medium was acquired from the Institute of Virol-

ogy, Vaccines and Sera ‘‘Torlak’’, Belgrade, Serbia. Male Wistar ratsaged 3 months were obtained from the Vivarium of the Institute ofBiomedical Research, Medical Faculty, University of Niš, Niš, Serbia.Animals were kept under standard laboratory conditions at22 ± 2 �C and 60% humidity, with food and water availablead libitum. All experimental procedures with the animals were con-ducted in compliance with The European Council Directive ofNovember 24th, 1986 (86/609/EEC) and were also approved bythe local Ethics Committee.

4.2.2. Evaluation of acute toxicity against A. salinaA teaspoon of lyophilized cysts of the brine shrimp A. salina was

added to 1 l of artificial sea water. The suspension, termostated at25 �C, was aerated with a mild air flow and kept under constantillumination. Most of the cysts hatched after approximately 48 h.Freshly hatched nauplii were transferred from the hatching beakerinto Petri dishes containing the tested compounds dissolved inDMSO and diluted with artificial seawater. Final concentrationsof the tested samples were in the range 5–50 lg/ml, whereas thefinal concentration of DMSO was much less than 1% (v/v). DMSOwas inactive under the stated conditions as demonstrated by anegative control. Brine shrimps were not fed during the test. Thetested samples were not aerated and the test dishes were left atroom temperature, under constant illumination. Dead nauplii werecounted after 24 and 48 h. LC50 (concentration lethal to 50% ofnauplii) were determined after statistical analysis. All tests wereperformed in triplicate. Solutions of varying concentrations ofSDS were used as the positive control.

4.2.3. AChE (acetylcholinesterase) inhibitory activityAChE inhibitory activities of compounds 3a and 3b were mea-

sured by a quantitative colorimetric assay based on Ellman’s meth-od [46]. In a 96-well plate, 25 ll of 15 mM acetylthiocholine iodide,125 ll of 3 mM 5,50-dithiobis-(2-nitrobenzoic acid) in buffer A(50 mM Tris–HCl, pH 7.9 containing 0.1 M NaCl and 0.02 M MgCl2),50 ll of buffer B (50 mM Tris–HCl, pH 7.9 containing 0.1% bovineserum albumin) and 25 ll of test compound solutions (0.156–40.0 lg of the test substance per 1 ml of 10% DMSO (v/v); geomet-rically increasing concentrations) were mixed and the absorbance(405 nm) was recorded on every 15 s during 1 min (5 measure-ments) on a microplate reader (Multiscan Ascent, Labsystems,Finland). After that, 25 ll of AChE (0.22 U/ml in buffer B) wasadded to each well and the plate was incubated at 25 �C for10 min. The absorbance was again measured in 15 s intervals (8points in total). A solution of 10% DMSO (v/v) was used as thenegative control. The absorbance resulting from the spontaneoushydrolysis of the substrate (the absorbance measured before theaddition of the enzyme (blind probe)) was subtracted from thatrecorded after the addition of AChE. For validation, different

140 N.S. Radulovic et al. / Polyhedron 80 (2014) 134–141

concentrations of physostigmine (several concentrations in therange 0.12–15 lmol/ml) served as a positive control. Each experi-ment was carried out in triplicate and repeated three times.

4.2.4. Evaluation of cytotoxicity against macrophages and thecorresponding macrophage MPO activities4.2.4.1. Isolation of peritoneal exudate cells. Rats were intraperitone-ally (i.p.) injected with thioglycolate medium in order to inducerecruitment of circulatory monocytes into the peritoneum and pro-liferation of resident peritoneal macrophages [47,48]. The cellswere harvested by peritoneal lavage with phosphate buffered sal-ine (PBS) 7 days following the i.p. injection when the majority ofthe cells in peritoneal exudates were macrophages, as previouslydescribed [49]. The peritoneal lavages were centrifuged(1200 rpm, 10 min) and resulting cell pellets were washed. Afterthis procedure, the cell viability was determined to be >95% bythe trypan blue dye exclusion method and the concentration ofindividual cell suspensions was adjusted to 2.5 � 106 cell/ml andused as such in the assays.

4.2.4.2. Crystal violet assay. Staining with crystal violet was used todetermine the viability of adherent cells (Flick and Gifford) [50].After the 60 min incubation of cells together with test drugs (3aor 3b in the final concentrations of 0.5, 1, 5, 10 and 50 lM), cellsamples were washed twice with PBS to remove non-adherentdead cells, and adherent viable cells were then stained with 1%(w/v) crystal violet solution at room temperature. Plates were thor-oughly washed with PBS, and the crystal violet dye was dissolvedin absolute ethanol. The absorbance of the dissolved dye, corre-sponding to the number of adherent (viable) cells, was measuredin a microplate reader (Multiscan Ascent, Labsystems, Finland) at595 nm. The results are expressed as% viability relative to un-treated control samples and as ODs (595 nm) � 1000. Each exper-iment was repeated three times.

4.2.4.3. Determination of MPO enzymatic activity. To measure theMPO activity in cells adhered to tissue culture plates, a slightlymodified method by Mitic et al. [51] was used. Briefly, individualcell suspensions of peritoneal macrophages (eight individual cellsamples in duplicate) were plated in 96-well flat-bottomed tissueculture plates (NUNC, Roskilde, Denmark) previously filled withtest substances (3a or 3b in the final concentrations of 0.5, 1, 5,10 and 50 lM). The plates were incubated for 60 min at 37 �C in95% air–5% CO2. Non-adherent cells were removed by washingthe plates twice with warmed (37 �C) PBS, resulting in the forma-tion of a uniform macrophage monolayer at the bottom of thewells. MPO activity was determined using 1,2-diaminobenzene(o-phenylenediamine) in citric buffer activated with H2O2 and Tri-ton X-100. The reaction was stopped with H2SO4 solution (1 M).Optical densities (ODs) were determined on a Multiscan Ascent(Labsystems, Finland) at 540 nm. The results are expressed asODs (540 nm) � 1000. Each experiment was repeated three times.

4.3. Statistical analysis

Data were analyzed by one-way ANOVA, followed by Fisher’sPLSD test and Student’s t-test. Differences are regarded as statisti-cally significant if p < 0.05. Results are presented as mean ± SE. Thestatistics were performed using statistical package SPSS 20.0 forWINDOWS.

Acknowledgments

This work was supported by the Ministry of Education, Scienceand Technological Development of Serbia [Project No. 172061].

This work is part of a PhD thesis of Dragan B. Zlatkovic supervisedby Niko S. Radulovic.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.poly.2014.03.006.

References

[1] P.D. Josephy, B. Mannervik, Molecular Toxicology, Oxford University Press,New York, 2006.

[2] M.W. Toussaint, T.R. Shedd, W.H. van der Schalie, G.R. Leather, Environ.Toxicol. Chem. 14 (1995) 907.

[3] D.R. Ruebhart, I.E. Cock, G.R. Shaw, Environ. Toxicol. 23 (2008) 555.[4] D.M. Mosser, J. Leukoc. Biol. 73 (2003) 209.[5] R.D. Stout, J. Suttles, J. Leukoc. Biol. 76 (2004) 509.[6] H.S. Kim, Y.J. Kim, H.K. Lee, H.S. Ryu, J.S. Kim, M.J. Yoon, J.S. Kang, J.T. Hong, Y.

Kim, S.B. Han, Food Chem. Toxicol. 50 (2012) 3190.[7] J.E. Krawisz, P. Sharon, W.F. Stenson, Gastroenterology 87 (1984) 1344.[8] S.J. Klebanoff, J. Bacteriol. 95 (1968) 2131.[9] P.J. Edelson, Z.A. Cohn, J. Exp. Med. 138 (1973) 318.

[10] C.E. Van Rensburg, A.M. Van Staden, R. Anderson, E.J. Van Rensburg, Mutat.Res. 265 (1992) 255.

[11] L. Roncucci, E. Mora, F. Mariani, S. Bursi, A. Pezzi, G. Rossi, M. Pedroni, D. Luppi,L. Santoro, S. Monni, A. Manenti, A. Bertani, A. Merighi, P. Benatti, C. Gregorio,M.P. de Leon, Cancer Epidemiol. Biomarkers 17 (2008) 2291.

[12] L.A. Pozzi, J.W. Maciaszek, K.L. Rock, J. Immunol. 175 (2005) 2071.[13] M.G. Lionetto, R. Caricato, A. Calisi, M.E. Giordano, T. Schettino, BioMed Res.

Int. (2013), http://dx.doi.org/10.1155/2013/321213.[14] P. Taylor, Z. Radic, Annu. Rev. Pharmacol. 4 (1994) 281.[15] H. Soreq, Y. Lapidot-Lifson, H. Zakut, Cancer Cells 3 (1991) 511.[16] D.H. Small, S. Michaelson, G. Sberna, Neurochem. Int. 28 (1996) 453.[17] G. Wilkinson, M. Rosenblum, M.C. Whiting, R.B. Woodward, J. Am. Chem. Soc.

74 (1952) 2125.[18] E.O. Fischer, W. Pfab, Z. Naturforsch. B 7 (1952) 377.[19] V.E. Emelyanov, L.S. Simonenko, V.N. Skvortsov, Chem. Technol. Fuel Oil 37

(2001) 224.[20] P. Stepnicka, Ferrocenes: Ligands, Materials and Biomolecules, John Wiley &

Sons Ltd, Chichester, England, 2008.[21] D. Conroy, A. Moisala, S. Cardoso, A. Windle, J. Davidson, Chem. Eng. Sci. 65

(2010) 2965.[22] M. Gielen, E. Tiekink, Metallotherapeutic Drugs and Metal-Based Diagnostic

Agents: The Use of Metals in Medicine, John Wiley & Sons Ltd., Chichester,England, 2005.

[23] R.A. Yeary, Toxicol. Appl. Pharmacol. 15 (1969) 666.[24] B. Loev, M. Flores, J. Org. Chem. 26 (1961). 3595-3595.[25] B. Pradines, T. Fusai, W. Daries, V. Laloge, C. Rogier, P. Millet, E. Panconi, M.

Kombila, D. Parzy, J. Antimicrob. Chemother. 48 (2001) 179.[26] F. Dubar, C. Slomianny, J. Khalife, D. Dive, H. Kalamou, Y. Guérardel, P. Grellier,

C. Biot, Angew. Chem. Int. Ed. Engl. 52 (2013) 7690.[27] F. Dubar, S. Bohic, C. Slomianny, J.C. Morin, P. Thomas, H. Kalamou, Y.

Guérardel, P. Cloetens, J. Khalife, C. Biot, Chem. Commun. (Camb.) 48 (2012)910.

[28] M.F.R. Fouda, M.M. Abd-Elzaher, R.A. Abdelsamaia, A.A. Labib, Appl.Organomet. Chem. 21 (2007) 613.

[29] J.J. Li, Name Reactions in Heterocyclic Chemistry, John Wiley & Sons, Hoboken,New Jersey, 2004.

[30] K.K. Lo, J.S. Lau, N. Zhu, New J. Chem. 30 (2006) 1567.[31] J. Quirante, F. Dubar, A. González, C. Lopez, M. Cascante, R. Cortés, I. Forfar, B.

Pradines, C. Biot, J. Organomet. Chem. 696 (2011) 1011.[32] Z.H. Chohan, M.M. Naseer, Appl. Organomet. Chem. 21 (2007) 1005.[33] Z.H. Chohan, J. Enzyme Inhib. Med. Chem. 24 (2009) 169.[34] P. Kopf-Maier, H. Kopf, E.W. Neuse, J. Cancer Res. Clin. Oncol. 108 (1984) 336.[35] E.W. Neuse, Macromol. Symp. 172 (2001) 127.[36] M. Lotti, in: P.S. Spencer, H.H. Schaumburg (Eds.), Experimental and Clinical

Neurotoxicology, second ed., Oxford University Press, New York, 2000, p. 898.[37] L.G. Costa, Clin. Chim. Acta 366 (2006) 1.[38] A. Zovko, K. Sepcic, T. Turk, M. Faimali, F. Garaventa, E. Chelossi, L. Paleari, C.

Falugi, M.G. Aluigi, C. Angelini, S. Trombino, L. Gallus, S. Ferrando, WSEASTrans. Biol. Biomed. 3 (2009) 58.

[39] E.L. Konrath, C. dos Santos Passos, L.C. Klein-Júnior, A.T. Henriques, J. Pharm.Pharmacol. 65 (2013) 1701.

[40] I. Sliskovic, I. Abdulhamid, M. Sharma, H.M. Abu-Soud, Free Radical Biol. Med.47 (2009) 1005.

[41] E. Malle, P.G. Furtmuller, W. Sattler, C. Obinger, Br. J. Pharmacol. 152 (2007)838.

[42] S. Pakakasama, T.T. Chen, W. Frawley, Int. J. Cancer 106 (2003) 205.[43] R.J. Hung, P. Boffetta, P. Brennan, C. Malaveille, U. Gelatti, D. Placidi, A. Carta, A.

Hautefeuille, S. Porru, Carcinogenesis 25 (2004) 973.[44] S.H. Olson, M.D. Carlson, H. Ostrer, S. Harlap, A. Stone, M. Winters, C.B.

Ambrosone, Gynecol. Oncol. 93 (2004) 615.[45] T.G. Traylor, J.C. Ware, J. Am. Chem. Soc. 89 (1967) 2304.

N.S. Radulovic et al. / Polyhedron 80 (2014) 134–141 141

[46] N.S. Radulovic, M.Z. Mladenovic, P.D. Blagojevic, Z.Z. Stojanovic-Radic, T. Ilic-Tomic, L. Senerovic, J. Nikodinovic-Runic, Food Chem. Toxicol. 62 (2013) 554.

[47] R.D. Eichner, T.C. Smeaton, Scand. J. Immunol. 18 (1983) 259.[48] M.J. Melnicoff, P.K. Horan, P.S. Morahan, Cell. Immunol. 118 (1989) 178.

[49] Y. Oghiso, Y. Yamada, Y. Shibata, J. Leukoc. Biol. 52 (1992) 421.[50] D.A. Flick, G.E. Gifford, J. Immunol. Methods 68 (1984) 167.[51] K. Mitic, S. Stanojevic, N. Kuštrimovic, V. Vujic, M. Dimitrijevic, Peptides 32

(2011) 1626.