Contents lists available at ScienceDirect

Journal of Inorganic Biochemistry

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

Review article

Synthesis, photophysical and cellular characterisation of folate andmethotrexate labelled luminescent lanthanide complexes

Zhangli Dub,1, Jing Sund,1, Christie A. Baderb, Doug A. Brooksb, Minqi Lid, Xun Lia,⁎,Sally E. Plushb,c,⁎⁎

a Department of Medicinal Chemistry, Key Laboratory of Chemistry and Chemical Biology (Ministry of Education), School of Pharmaceutical Sciences, ShandongUniversity, Ji'nan, Shandong, P.R. Chinab Mechanisms in Cell Biology and Diseases Research Group, School of Pharmacy and Medical Sciences, Sansom Institute for Health Research, University of South Australia,Adelaide, Australiac Future Industries Institute, University of South Australia, Mawson Lakes, SA 5095, Australiad Department of Bone Metabolism, School of Stomatology, Shandong University, Shandong Provincial Key Laboratory of Oral Tissue Regeneration, Ji'nan, P.R. China

A B S T R A C T

In this work we have developed a series of highly emissive europium(III) and terbium(III) complexes tethered toeither folic acid (FA) or methotrexate (MTX), with the aim of developing visual probes that enable the imaging offolate receptors in cancer cells. The synthesis, photophysical properties and cellular behaviour are reported forfour new lanthanide Ln(III) complexes, where either FA or MTX are tethered to 1,4,7-tris(carbonylmethyl)-10-(4′-quinolineacetic acid, (7′-acetamido)-1′,2′-dihydro-2′-oxo)-1,4,7,10-tetraazacyclododecane Ln(III) complex,and Ln(III) = Eu(III) or Tb(III); herein referred to as Eu-FA, Eu-MTX, Tb-FA or Tb-MTX. All four complexeswere found to be sensitive to the presence of the folate receptor in a range of cell lines. The MTX conjugatesshowed different cellular specificity in an oral adenosquamous carcinoma cell line (CAL-27) compared with theanalogous FA conjugates. This suggests that it is viable to explore differences in folate receptors using folate vs.anti-folate probes, with labels that have different emissive properties (e.g. Eu-FA vs. Tb-MTX). The MTXcomplexes were found to be the most cytotoxic, with Eu-MTX showing greater cytotoxicity than free MTX or theisostructural Tb-MTX. This suggested that there could be a synergistic effect on toxicity for the Eu(III) chelateand the MTX components of the complex.

1. Introduction

Folates (also called vitamin B9, folacin, pteroyl-L-glutamic acid, orpteroyl-L-glutamate) are essential for the maintenance of the humangenome and cell health, due to their central role in key metabolicfunctions, such as RNA and DNA biosynthesis [1,2]. The critical role offolates in DNA synthesis, repair and methylation [3] has been mainlyattributed to the reduced tetrahydrofolate form [4], which has an im-portant function in rapid cell division and growth. Reduced amounts offolate have been associated with the onset of cancer, including color-ectal [5], breast [6] and lung cancer, presumably due to increased DNAdamage and an enhanced mutation rate [7]. Folate can be internalisedinto cells via at least three transport mechanisms: the reduced folate

carrier (RFC) [8], which is an anion exchanger and mainly transportsreduced folate; the proton-coupled folate transporter (PCFT) [9], whichcan transport folate in an acidic environment; and the folate receptors(FRs) [10], which have high affinity and can transport folic acid (FA)into cells via endocytosis [11,12].

In contrast to the RFC and PCFT, which are ubiquitously expressedin both tumour and normal tissues [13], the FRs usually have minimalexpression in normal tissue, but increased expression in malignantcells/tissues, including, colorectal, ovarian, breast, lung, cervical, renal,kidney, brain and nasopharyngeal carcinomas; and this increased ex-pression has been linked with tumour progression [14–16]. This, makesthe FR a potential target for cancer diagnosis/therapy; [17,18] whereFA (small and stable over broad range of temperatures and pH) can be

http://dx.doi.org/10.1016/j.jinorgbio.2017.10.003Received 11 August 2017; Received in revised form 19 September 2017; Accepted 8 October 2017

⁎ Correspondence to: Dr. Xun Li, Department of Medicinal Chemistry, Key Laboratory of Chemistry and Chemical Biology (Ministry of Education), School of Pharmaceutical Sciences,Shandong University, Ji'nan, Shandong, P.R. China.

⁎⁎ Correspondence to: Dr. Sally E. Plush, Mechanisms in Cell Biology and Diseases Research Group, School of Pharmacy and Medical Sciences, Sansom Institute for Health Research,University of South Australia, Adelaide, Australia.

1 These two authors contributed equally to this work.E-mail addresses: tjulx2004@sdu.edu.cn (X. Li), sally.plush@unisa.edu.au (S.E. Plush).

Journal of Inorganic Biochemistry 178 (2018) 32–42

Available online 10 October 20170162-0134/ Crown Copyright © 2017 Published by Elsevier Inc. All rights reserved.

MARK

used as a biocompatible and non-immunogenic targeting motif tocovalently conjugate with an optical imaging agent or a therapeutic[19]. Anti-folates, such as methotrexate (MTX) [20], have been directlyexploited as therapeutic agents [21]. MTX was first synthesised in the1940's [22] and is still extensively used in the treatment of tumours,including acute lymphocytic leukaemia [23], breast [24], head andneck cancers [25]. It is also used in certain autoimmune diseases [26].MTX competitively inhibits dihydrofolate reductase, a key step in thereduction of folate to its tetrahydrofolate form, inhibiting DNA andRNA synthesis to cause cell apoptosis [27]. Folates and anti-folates canalso be used as targeting agents and this has been successfully employedin a wide variety of imaging applications, with clinical trials in humansunderway for magnetic resonance imaging (MRI), computed tomo-graphy (CT) and positron emission tomography (PET) imaging of solidtumours [28,29]. Fluorescent agents tethered to FA have been in-vestigated for intra-operative identification of malignant disease [30].In contrast to the conventional modalities of MRI, CT and PET, opticalimaging agents offer sub-micrometre spatial resolution, allowing forrapid results at the sub-cellular level [31]; where early stage biologicalchanges can be more easily detected using low cost and more accessibleinstrumentation. This opens opportunities for the non-invasive diag-nosis of cutaneous and subcutaneous cancers and for example endo-scopic diagnosis of oesophageal malignancies, where the issues of tissuetransparency are usually negated [28]. Furthermore, the use of imagingagents for microscopy will allow cell biologists working in the field ofcancer to investigate and monitor changes at a cellular level prior to,during and post intervention, in real time. To achieve these importantimaging outcomes, there is a need to be able to monitor the opticalsignal for long periods of time, without photo-bleaching and to be ableto differentiate the signal from endogenous fluorescence. Consequently,new imaging technology is required to address this unmet need.

Luminescent lanthanide (Ln) complexes are particularly attractiveas imaging agents, due to their distinctive photo-physical properties[32]. They offer extended resistance to photobleaching, have largeStokes shifts (large separation between absorption and emission wa-velengths) that avoid concentration-dependant self-absorption pro-blems, and exhibit long luminescence lifetimes (range 1 μs to 5 ms) thatenable time gated measurement of luminescence to avoid either non-specific background or specific auto-fluorescence. Furthermore, com-pared with fluorescent bands (> 200 nm bandwidths), lanthanideshave sharp emission bands (10–20 nm bandwidths) that do not overlapwith one another, resulting in optimal signal to noise ratios [33].Lanthanides have an advantage over other luminescent metal ions, asthey are considered to be isostructural, which means that the emissionrange can be tuned by ‘swapping’ the metal, using the same syntheticstrategy to deliver complexes with theoretically similar physiochemicaland biological targeting properties, but with different wavelengthemissions [34]. One key consideration with lanthanide ions istheir intrinsically low extinction coefficients (in the order of0.5–3 M−1 cm−1), which necessitates incorporating an antenna groupinto the lanthanide complex, to effectively populate the lanthanide'sexcited state [35]. Although a wide range of lanthanide complexes havebeen evaluated for the optical imaging of cells, most notably by Parkerand co-workers [36–39], further development is required to establish atrue structure activity relationship (SAR) and to determine the criticalparameters for the design and optimisation of these imaging agents.This design process may be further complicated in the case of folates astargeting groups, as there are differences in cellular uptake efficiencyand receptor binding affinities for substituted analogues of folic acid[40–43].

We and others have recently explored the relationship between thefolate moiety (FA consists of three moieties; pterin head group, p-aminobenzoate and glutamate residue), the length of the linker be-tween the folate and lanthanide complex, and the site of conjugation ofthe lanthanide complex; and the effects upon cellular uptake and lu-minescent emission for folate targeted Eu(III) conjugates [44,45]. This

has shown that each portion of FA is critical and that when a shortlinker is used to tether a luminescent Eu(III) complex using the γ-car-boxylic acid of FA, higher cellular uptake is achieved [44]. However, tobe optimal for live cell imaging, these complexes need to be highlyemissive and excitable within biologically compatible wavelengths.

In this study, we report the synthesis, photophysical and cellularproperties of four new imaging agents, Eu-FA, Eu-MTX, Tb-FA and Tb-MTX, which each consists of a highly luminescent antenna complex,covalently conjugated using a short linker, to one of the carboxylic acidgroups present in either FA or MTX (Fig. 1). Both Eu(III) and Tb(III)complexes were explored to evaluate any differences in emission and tocompare the uptake/targeting for these two isostructural ions. It isimportant to ascertain how changes in the core metal may influence thepotential to develop near infrared emitting complexes or even to use Gd(III) as an effective MRI contrast agent. We also aimed to explore howuptake, cytotoxicity and emission varied in a range of cells when folate(in the form of FA) was substituted for the anti-folate, MTX. Differencesin the uptake and cytotoxicity of FA and MTX have implications for thetargeting and potential treatment of cancer, as FA is often given as anantagonist drug to lessen and limit side effect toxicity (e.g. non-tumourtoxicity) during MTX treatment. Therefore, it is important and mean-ingful that cell biologists are equipped with tools that have the potentialto clearly visualise how these two analogues behave in cells, to developbetter strategies for cancer treatment.

2. Experimental

2.1. General materials and methods

1,4,7,10-tetraazacyclododecane (cyclen) was obtained commer-cially from Strem, USA. All other solvents and chemicals used wereobtained from Sigma-Aldrich or Merck, Australia. Sephadex G10 resinswere purchased from Sigma-Aldrich, Australia. Thin layer chromato-graphy (TLC) was performed on silica gel 60 F254 plates obtained fromMerck, Australia. The infrared spectroscopy (IR) was recorded on aShimadzu FTIR-8400S. Electrospray ionization mass spectrometry (ESI-MS) was recorded using a Perkin-Elmer® Scoex API 3000. 1H NMR and13C NMR were recorded on a Bruker 500 MHz NMR spectrometer. Allchemical shifts are given in ppm with coupling constants in Hz. All pHmeasurements were conducted on an Orion Ross pH meter. Opticalspectroscopy experiments were recorded in 100% water at constantionic strength (I= 0.01 (NaCl)) using a Varian CARY 50 UV–Visspectrophotometer or a Varian Cary Eclipse spectrophotometer at roomtemperature. High pressure liquid chromatography (HPLC) was per-formed on a Shimadzu LC-20AD with manual injection fitted with a60 Å, 4 μM Nova-Pak phenyl analytical column (3.9 × 150 mm) with aflow rate of 0.8 mL/min and was analysed using a Shimadzu SPD-20Adetector at 280 nm. Linear elution was used with mobile phases A (H2O+ 0.1% TFA) and B (CH3CN + 0.1% TFA); 15%–60% over 10 min.Elemental analysis was conducted in Campbell MicroanalyticalLaboratory, University of Otago.

2.2. Synthesis

2.2.1. 2,2′,2″-(10-(2-(4-(2-(2-(tert-butoxycarbonylamino)ethylamino)-2-oxoethyl)-2-oxo-1,2-dihydroquinolin-7-ylamino)-2-oxoethyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetic acid.Eu (Eu.3)

Compound Eu.1 [46] (131 mg, 0.17 mmol) was dissolved in DMSOusing sonication (170 W, 30 min), after which DIPEA (45 mg,0.35 mmol) was added. The reaction mixture was then stirred for10 min. Then HOBt (27 mg, 0.20 mmol) and BOP (159 mg, 0.36 mmol)were added. Following this, a solution of tert-butyl 2-ami-noethylcarbamate (60 mg, 0.37 mmol), DIPEA (23 mg, 0.18 mmol) inDMSO (2 mL) was added. The reaction mixture was then left to stir atroom temperature for 24 h. The crude product was then precipitatedout by the addition of diethyl ether/acetone (10 mL, 7/3 (v/v)). The

Z. Du et al. Journal of Inorganic Biochemistry 178 (2018) 32–42

33

precipitate was collected by centrifugation at 8300g for 5 min and usedin the next step without any further purification. Yield: 127 mg (83%).IR (KBr) v = 3433 (br), 2978, 2920, 2870, 1616 (br), 1389, 1319,1281, 1246, 1165, 1084, 1018 cm−1. 1H NMR (D2O, 500 MHz): δ33.92, 30.87, 29.66 (d), 11.87 (br), 10.74 (br), 9.02 (br), 7.81 to 0.32(m), −1.07, −3.08, −3.56, −3.89, −4.99, −6.68, 8.06 (d),−10.42, −12.12 (d), −13.52 (d), −14.63, −15.94 (d), −17.24. ESI-MS+: m/z = 897.4 (M + 1)+.

2.2.2. 2,2′,2″-(10-(2-(4-(2-(2-aminoethylamino)-2-oxoethyl)-2-oxo-1,2-dihydroquinolin-7-ylamino)-2-oxoethyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetic acid.Eu (Eu.5)

Compound Eu.3 (170 mg, 0.19 mmol) was dissolved in a DCM/TFA(5 mL, 1/1 (v/v)) solution and stirred at room temperature overnight.The solvent was then removed under reduced pressure to yield a brownresidue. The residue was then redissolved in water (8 mL) and the pH ofthe solution adjusted to 7 using 1 M NaOH. The solvent was removed byreduced pressure to yield the crude product. The crude product waspurified by a Sephadex G10 column in CH3OH/H2O (1/1) to yield ayellow crystal. Yield: 133 mg (88%). M.p. > 300 °C. IR (KBr):v = 3414 (br), 3063, 2986, 2916, 2870, 1597 (br), 1404, 1323, 1281,1242, 1204, 1177, 1130, 1084, 1018 cm−1. 1H NMR (D2O, 500 MHz): δ34.08 (d), 30.84 (br), 29.46 (br), 11.90 (br), 7.80–0 (m), −0.36,−1.02, −3.03 (br), −3.97, −4.93, −8.10 (m), −10.34, −11.73,−12.13, −12.95, −13.44, −14.33, −14.66, −15.91 (d), −17.26.ESI-MS+: m/z = 797.5 (M + 1)+.

2.2.3. Eu-FAFolic acid (48 mg, 0.11 mmol) was dissolved in DMSO (4 mL).

DIPEA (28 mg, 0.21 mmol) was added, followed by the sequentiallyaddition of HOBt (15 mg, 0.11 mmol) and BOP (53 mg, 0.13 mmol).The reaction mixture was stirred for 5 min, then was mixed with Eu.5(88 mg, 0.11 mmol) together with DIPEA (14 mg, 0.11 mmol) in DMSO(1 mL). The mixture was stirred for 24 h at room temperature and

protected from light. The product was then precipitated by addingdiethyl ether/acetone (10 mL, 7/3 (v/v)) which was collected aftercentrifugation of the suspension at 8300g for 5 min. The residue wasdried, then rinsed with CH3OH (3 mL). The product was collected as ayellow powder. Yield: 110 mg (82%). M.p. > 300 °C. TLC: silica iso-propanol/ammonia (5 M) (7/3), Rf 0.1. IR (KBr): v = 3410 (br), 2986,2920, 2866, 1605 (br), 1389, 1323, 1277, 1242, 1184, 1084,1018 cm−1. Calculated for C48H56EuN15O14 ⋅3H2O ⋅2CH3SOCH3: C,43.70; H, 5.22; N, 14.70. Found: C, 43.61; H, 5.36; N, 14.55. 1H NMR(DMSO-d6, 500 MHz): δ 41.66, 38.83, 33.59, 32.11, 30.63, 18.92 (br),15.35 (br), 13.01, 12.27, 10.76, 10.60, 9.26 to 1.23 (m),−1.32,−3.59(d), −4.54, −6.57, −7.59 (d), −10.46, −13.86 (d), −15.78,−17.39 (m), −19.72 (br), −23.48. ESI-MS−: m/z = 1218.8 (M-1)−.HPLC: t = 4.1 min, 4.4 min.

2.2.4. 2,2′,2″-(10-(2-(4-(2-(2-(tert-butoxycarbonylamino)ethylamino)-2-oxoethyl)-2-oxo-1,2-dihydroquinolin-7-ylamino)-2-oxoethyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetic acid.Tb (Tb.4)

Compound Tb.4 was prepared using the protocol described tosynthesise Eu.3, except that Tb.2 (see ESI for synthesis) (193 mg,0.25 mmol) was used to react with tert-butyl 2-aminoethylcarbamate(85 mg, 0.53 mmol) instead of Eu.1. Yield: 187 mg (83%).M.p. > 300 °C. IR (KBr) v = 3433 (br), 2982, 2920, 2874, 1609 (br),1389, 1323, 1281, 1246, 1165, 1084, 1018 cm−1. 1H NMR (D2O,500 MHz): δ 260.40, 252.39, 226.19, 136.21, 127.27, 117.41, 82.59 50to −50 (br), −52.08, −63.48, −66.56, −73.65, −95.84, −99.23,−102.00, −113.40, −128.81, −144.22, −351.92, −369.80,−400.00, −409.86. ESI-MS+: m/z= 903.5 (M + 1)+.

2.2.5. 2,2′,2″-(10-(2-(4-(2-(2-aminoethylamino)-2-oxoethyl)-2-oxo-1,2-dihydroquinolin-7-ylamino)-2-oxoethyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetic acid.Tb (Tb.6)

Compound Tb.6 was prepared using the protocol described for thesynthesis of Eu.5 except that Tb.4 (207 mg, 0.23 mmol) was stirred in a

Fig. 1. The chemical structures of folic acid (FA), methotrexate (MTX) and four lanthanide complexes; Eu-FA, Tb-FA, Eu-MTX and Tb-MTX (all lanthanide complexes displayed as γ-isomer).

Z. Du et al. Journal of Inorganic Biochemistry 178 (2018) 32–42

34

solution of DCM/TFA (5 mL, 1/1 (v/v)) overnight instead of Eu.3. Theproduct was isolated as a yellow crystals. Yield: 133 mg (71%).M.p. > 300 °C. IR (KBr): v = 3433 (br), 2990, 2916, 2878, 1620 (br),1400, 1319, 1281, 1242, 1169, 1084, 1022 cm−1. 1H NMR (D2O,500 MHz): δ 259.78, 249.3, 233.28, 223.11, 135.90, 126.04, 116.79,82.90, 50 to −50 (br), −52.39, −63.48, −67.49, −73.04, −95.53,−99.54, −101.39, −113.40, −128.20, −143.30, −351.61,−369.49, −398.46, −408.62. ESI-MS+: m/z= 803.4 (M + 1)+.

2.2.6. Tb-FACompound Tb-FA was prepared using the procedure described for

Eu-FA except that Tb.6 (95 mg, 0.12 mmol) was used instead of Eu.5 toreact with folic acid (52 mg, 0.12 mmol), HOBt (17 mg, 0.13 mmol),BOP (57 mg, 0.13 mol) and DIPEA (46 mg, 0.36 mmol) in DMSO(5 mL). The final product was collected as a yellow powder. Yield:95 mg (65%). M.p. > 300 °C. TLC: silica isopropanol/ammonia (5 M)(7/3), Rf 0.1. IR (KBr): v = 3422 (br), 2990, 2920, 2874, 1605 (br),1400, 1319, 1242, 1184, 1126, 1084, 1003 cm−1. Calculated forC48H56TbN15O14 ⋅3H2O ⋅2CH3SOCH3: C, 43.48; H, 5.19; N, 14.63.Found: C, 43.68; H, 5.31; N, 14.64. 1H NMR (DMSO-d6, 500 MHz): δ284.74, 232.97, 155.62, 145.76, 50 to −50 (br), −73.04, −102.62,−117.10, −156.86, −183.67, −345.45, −363.63. ESI-MS−: m/z = 1224.7 (M-1)−. HPLC: t = 4.1 min, 4.4 min.

2.2.7. Eu-MTXMethotrexate (54 mg, 0.12 mmol) was dissolved in DMSO (4 mL) and

DIPEA (31 mg, 0.26 mmol) added, followed by the sequentially additionof HOBt (18 mg, 0.13 mmol) and BOP (56 mg, 0.13 mmol). The reactionmixture was stirred for about 5 min, then was mixed with Eu.5 (96 mg,0.3 mmol) together with DIPEA (16 mg, 0.12 mmol) in DMSO (5 mL).The mixture was stirred for 24 h at room temperature in the dark. Theproduct was then precipitated by the addition of ethyl ether/acetone(10 mL, 7/3 (v/v)) which was collected by centrifugation (8300 g,5 min). The residue was dried and washed by CH3OH (3 mL). The pro-duct was collected as a yellow powder. Yield: 95 mg (64%).M.p. > 300 °C. TLC: silica isopropanol/ammonia (5 M) (7/3), Rf 0.1. IR(KBr): v= 3426 (br), 2990, 2920, 1605 (br), 1385, 1319, 1242, 1200,1084, 1022 cm−1. Calculated for C49H59EuN16O13⋅H2O ⋅5CH3SOCH3: C,43.19; H, 5.59; N, 13.66. Found: C, 43.26; H, 5.59; N, 13.89. 1H NMR(DMSO-d6, 500 MHz): δ 41.54, 38.95, 33.65, 32.30, 30.63, 18.98, 15.53,12.33, 10.79, 8.60-0 (m), −1.17, −3.45 (d), −4.47, −6.56, −7.55,−10.42, −13.84, −17.29 (m), −19.82. ESMS−: m/z 1231.8 (M-1)−.HPLC: t = 4.3 min, 4.6 min.

2.2.8. Tb-MTXTb-MTX was prepared through the same method that was used to

prepare Eu-MTX except Tb.6 (70 mg, 0.087 mmol) was used to reactwith methotrexate (40 mg, 0.087 mmol), HOBt (14 mg, 0.10 mmol), BOP(44 mg, 0.10 mmol) in DMSO (5 mL). Yield: 72 mg (67%). TLC: silicaisopropanol/ammonia (5 M) (7/3), Rf 0.1. M.p. > 300 °C. IR (KBr):v= 3422 (br), 2990, 2916, 2882, 1605 (br), 1385, 1319, 1242, 1200,1084, 1022 cm−1. Calculated for C49H59Tb.N16O13⋅H2O ⋅4CH3SOCH3:C, 43.62; H, 5.46; N, 14.28. Found: C, 43.57; H, 5.50; N, 14.09. 1H NMR(DMSO-d6, 500 MHz): δ 295.84, 208.32, 167.64, 153.16, 143.60, 93.07,50 to −50 (br), −72.73, −102.31, −118.03, −156.55, −183.36,−347.30, −365.79, −429.27. ESMS−: m/z 1238.0 (M-1)−. HPLC:t = 4.3 min, 4.6 min.

2.3. Spectroscopic methods

All pH measurements were conducted using an Orion Ross pHmeter. Deionised water; that had been purified with the MiliQ-Reagentsystem to produce water with a specific resistance of> 18.2 MΩ cm−1,then boiled for 30 min to remove CO2 and cooled under a drying tubefilled with soda lime; was used to prepare all aqueous solutions. Allsolutions were prepared freshly prior to measurement. Samples of the

complexes were first dissolved in NaH2PO4.2H2O (20 mM) solution andthen further diluted with water (I = 0.01 (NaCl)). Optical spectroscopyexperiments were recorded in 100% water at constant ionic strength(I= 0.01 (NaCl)) using a Varian CARY 50 UV–Vis spectrophotometeror a Varian Cary Eclipse spectrophotometer at room temperature. TheUV–Visible absorbance spectra were recorded over the wavelengthrange of 250–450 nm with a scan rate of 600 nm/min. Baseline cor-rection measurements were used for all spectra with a blank containing0.01 M NaCl. For the fluorescence emission spectra data was obtainedbetween approximately 400–600 nm with both excitation slit andemission slit widths at 10 nm. The concentrations of the samples werethe same as that used for the UV–Vis absorbance measurements. Theexcitation wavelength was 355 nm. The Ln(III) emission spectra wererecorded on a Varian Cary Eclipse spectrophotometer. The data wasobtained between approximately 550–750 nm for Eu(III) complexesand 450–650 nm for Tb(III) complexes with both the excitation andemission slit widths at 10 nm. The concentrations of the samples werethe same as that used for the UV–Vis absorbance measurements. Theexcitation wavelength was 355 nm. Luminescent lifetime measure-ments were performed on a Varian Cary Eclipse spectrophotometerwith a sample dissolved in a minimal amount of NaH2PO4·2H2O(20 mM) or of NaD2PO4 (20 mM) solution and then further diluted to aconcentration of 10−5 M (Eu(III) complexes) or 0.5 × 10−5 (Tb(III)complexes) in either H2O (3 mL) or D2O (3 mL). The measurement ofemission was at 615 nm for Eu(III) complexes and 545 nm for Tb(III)complexes. The PMT voltage was set at 550 V.

2.4. Epi-fluorescence imaging

2.4.1. Cell cultureCells used in the assay, including HeLa (human cervical cancer

cells), 293 T (human embryonic kidney cells), MDA-MB-231 (humanbreast cancer cells), U251 (human glioma cells), RAW264.7 (murinemacrophage cells), A549 (human lung epithelial cells), PC-3 (humanprostate cancer cells), U-2 OS (human osteosarcoma cells), Cal-27(human tongue squamous carcinoma cells), MLO-Y4 (human normalosteocyte cells), ST2 (mouse mesenchymal stem cells), MC3T3-E1(mouse pre-osteoblast cells), HL7702 (human normal liver cells) werecommercially available from the Shanghai Cell Center. The passagenumbers for each cell line used for the experiments ranged from 3 to 10.Cells were seeded at the density of 4 × 105 cells/well in a six-wellpolystyrene plate. Cells were cultured in the Dulbecco's Modified EagleMedium (DMEM; Hyclone, Logan, UT, USA) with 10% fetal bovineserum (FBS; Gibco, Thermo Fisher Scientific Inc., Waltham, MA, USA)and 1% penicillin/streptomycin. Cultures were maintained at 37 °C in afully humidified atmosphere of 5% CO2 in air. The culture medium waschanged every 3 days.

2.4.2. Cell fluorescence detectionFor microscopy, cells seeded on coverslips were incubated with the

complex (Eu-FA, Tb-FA, Eu-MTX, Tb-MTX; 50 μM) dissolved in freshmedium at 37 or 4 °C in a 12-well plate for 24 h or 48 h. Then the cellswere washed with PBS at least three times and immediately fixed bypre-cooled methanol (−20 °C) for 10 min, washed with PBS for 5 mintwice, and mounted onto slides with a drop of glycerinum for detection.Cells were examined by microscopy using the appropriate excitationand emission filters, which include the narrow band UV excitation(exciter filter BP360-370, dichroic beamsplitter DM400, barrier filterBA 420-460), narrow band interference filter blue excitation (exciterfilter BP470-495, dichroic beamsplitter DM505 and barrier filterBA510-550), narrow band interference filter green excitation (exciterfilter BP540-550, dichroic beamsplitter DM570 and barrier filterBA575-725). Images were captured by fluorescence microscopy(Olympus BX53, Tokyo, Japan) with an image size of1600 × 1200 pixels (200× magnification at 1.40 NA).

Z. Du et al. Journal of Inorganic Biochemistry 178 (2018) 32–42

35

2.5. Anti-proliferative methods

The tested chemicals were dissolved in DMSO or PBS and diluted tothe desired concentration with culture medium before using. The celllines were maintained in RPMI-1640 medium supplemented with 10%(v/v) heat-inactivated fetal bovine serum (FBS) and incubated at 37 °Cin a humidified incubator with 5% CO2. Cell proliferation was de-termined by the MTT (3-[4,5-dimethyl-2-thiazolyl]-2,5-diphenyl-2H-tetrazolium bromide) or MTS methods. Briefly, cells were seeded in a96-well plate (104 cells per well). After 4 h incubation, chemicals weresubsequently added to wells to achieve final concentration of 200, 150,100, 50, and 10 μM. Control wells were prepared by addition of culturemedium. Treated cells were then incubated for 48 h. Once completionof incubation, 1% of 0.5 mg/mL MTT solution was added to each welland incubated for an additional 4 h. Formazan formed from MTT wasextracted by adding 100 μL of DMSO and mixed for another 15 min.The optical density was measured using an enzyme-linked im-munosorbent assay (ELISA) reader (Model 680, BIO-RAD) at 570 nm.

For MTS method HeLa cells were seeded at 1 × 105 cells/mL in 96well plates. Cells were then incubated with FA, MTX, Eu-FA, Eu-MTXor Eu.1 at 200, 150, 100, 50 or 10 μM in cell culture media containing1% DSMO. Control cells were also incubated with cell culture mediawith 1% DMSO. Following a 48 h incubation, cells were wash 2 timesfor 5 min with PBS before cell culture media was replaced. Cells werethen incubated with 10% v/v MTS reagents (BioVision) for 2 h to assesscell viability. Absorbance was read on an EnSpire® multimode platereader (Perkin Elmer®). Results were calculated as the relative absor-bance of control cells incubated in cell culture containing 1% DMSO.The growth inhibition rate was calculated as [(ODc − ODt) /(ODc − ODz)] × 100%. In this formula, ODc represents the OD valuesof the control group, ODt represents the OD values of the treatinggroups, and ODz represents the OD values of the zero-setting groups.Three independent experiments with triplicated samples were per-formed to achieve the cytotoxic results, and the IC50 values were cal-culated according to inhibition ratios.

3. Results and discussion

3.1. Synthesis of the probes

The synthesis of the final complexes was a multistep procedure,involving the initial formation of Eu.1 or Tb.2. Eu.1 and Tb.2 wereformed using a ring closing reaction to generate the carbostyril an-tenna, attachment to the cyclen based ligand and Ln(III) complexation(synthesis of Eu.1 is described elsewhere [46], see ESI for synthesis ofTb.1). This was followed by the attachment of a reactive linker inter-mediate to generate Eu.3 or Tb.4, then removal of linker protectinggroup and finally coupling to either FA or MTX to yield Eu-FA, Tb-FA,Eu-MTX or Tb-MTX, respectively, Scheme 1. To ensure sufficientflexibility in the targeting moiety for optimal cellular uptake and totake account of our previous findings [44], a small ethane linker groupwas included in the synthesis to provide a ‘spacer’ unit between therigid carbostyril antenna and the FA or MTX residue. The carbostyrilantenna was chosen as it is known to be very effective at populating theexcited states of both Eu(III) and Tb(III) ions [47,48]. Furthermore, ithas two reactive sites, which allows it to act as conduit between theligand and the targeting group. In addition, it is relatively small in sizewhich ensures that the Ln complexes will have a suitable molecularweight/volume, which is very important for biological applications[49].

Conjugation of the linker, tert-butyl 2-aminoethylcarbamate, to ei-ther Eu.1 or Tb.1 was achieved via an amide coupling reaction with 1-hydroxybenzotriazole (HOBt), (benzotriazol-1-yloxy)tris(dimethyla-mino)phosphonium hexafluorphosphate (BOP), N,N-diisopropylethyla-mine (DIPEA) in dimethyl sulfoxide (DMSO). No change in the broad IRυC = O peak for either complex was observed following conjugation,

confirming that the attachment of the linker did not affect the Ln(III)coordination. To generate the free amine site for consequent FA or MTXcoupling, the tert-butyl group on the ethyl linker was removed by stir-ring Eu.3 or Tb.4 in trifluoroacetic acid (TFA)/dichloromethane (DCM)(1/1, v/v), followed by purification using a Sephadex column yieldingeither Eu.5 or Tb.6, respectively. The synthesis of the four complexes,Eu-FA, Tb-FA, Eu-MTX and Tb-MTX was achieved by coupling the freeamine site of Eu.5 or Tb.6 to folic acid or MTX in the presence of HOBt,BOP and DIPEA in DMSO; yields for this final step ranged from 60% to80% (see ESI Figs. S4–5 for representative for 1H NMR spectra of Eu-FAand Eu-MTX). No bis-coupling of FA or MTX to either Eu.5 or Tb.6 wasdetected by MS analysis for any of the final complexes. Elementalanalysis of Eu-FA, Tb-FA, Eu-MTX and Tb-MTX were in close ap-proximation with the calculated values and indicated that all complexesexisted in the presence of small number of solvent molecules. Thepresence of the solvent molecules are not unexpected in lanthanidecomplexes and have often been reported [50,51]. The HPLC analysis ofall four targeted complexes showed two peaks of relatively equal area(see ESI Figs. S6–9 for HPLC traces of Eu-FA, Tb-FA, Eu-MTX and Tb-MTX). This suggested that coupling occurred at equal rates at the twocarboxylate groups (α or γ) for both FA/MTX (Fig. 2). This was ex-pected based on literature precedence [52,53] and our own previouswork [44], where the reduced steric hindrance at the γ-COOH wouldresult in the formation of the γ-isomer as the major product. The rela-tively equal formation rates of the two isomers suggests that the in-clusion of the carbostyril antenna may alter the steric hindrance withinthe molecule or act in a synergistic manner, and hence, the formation ofa predominant isomer was less affected by the space limitation. There isa significant number of conflicting reports within the literature withregards to isomer selectivity, with a large proportion of work on FAconjugates involving mixed isomeric solutions [10,40–43], hence forthis work we used the Eu(III) and Tb(III) labelled FA and MTX deri-vatives as mixed isomeric solutions for all further characterisation.

3.2. Photophysical characterisation

3.2.1. Lifetimes and q valuesThe emission lifetimes of the Eu-FA, Tb-FA, Eu-MTX and Tb-MTX

complexes were determined to provide information that was relevantfor time-gated measurements and cell imaging. The lifetimes also allowthe q value (the number of metal bound waters) to be quantified foreach complex, which helped to confirm the coordination environmentaround the lanthanide ion. The inner sphere hydration state (q) (i.e. thenumber of metal bound water molecules) was defined by measuring theexcited state lifetimes of the Eu-FA, Tb-FA, Eu-MTX and Tb-MTXcomplexes, in H2O and D2O, respectively. The q values were calculatedusing q = 1.11 (kH2O-kD2O-0.31) [54] for Eu(III) complexes and q= 4.2(kH2O-kD2O) [55] for Tb(III) complexes (Table 1). The tau (τ) valuesfound in Table 1 are the reciprocal of k used in the two equations andwere determined using the Cary Eclipse Win FLR lifetime program. Aq ~ 1 was found for all of the imaging agents, indicating that bothcomplexes possessed one water molecule in the inner coordinationsphere, supporting the formation of an 8 coordination complex with theligand.

3.2.2. UV–Visible, fluorescence and lanthanide ion emission spectroscopyUnder UV light excitation all four complexes were found to be

highly emissive, with Eu-FA and Eu-MTX appearing as bright red dueto the sensitised emission of europium while as anticipated Tb-FA andTb-MTX both appeared as bright green due to the sensitised terbiumemission. The two Tb(III) complexes were significantly more emissivethan the corresponding Eu(III) complexes, and this is attributed to themore closely aligned triplet state energies between carbostyril and Tb(III) ion compared to the Eu(III) ion [56]. These differences in emissioncan be easily observed using a benchtop UV light (see ESI Fig. S10 for acomparative image of Eu-FA vs. Tb-FA). As a result of this difference in

Z. Du et al. Journal of Inorganic Biochemistry 178 (2018) 32–42

36

emission, a higher concentration of the europium complexes wastherefore used to determine the absorbance and emission properties,when compared to the terbium complexes.

The photo-physical behaviour of each of the four Eu-FA, Tb-FA, Eu-MTX and Tb-MTX complexes was investigated as a function of pH. Ascellular pH varies significantly (from pH 7 in the cytoplasm to pH < 4in the lysosome), it was important to understand the photo-physicalproperties of each of the complexes prior to imaging in cells.Understanding the influence of pH may also give insight into the in-fluence that the metal ion has on the overall functionality of the com-plex. The absorption spectra of Tb-FA and Eu-MTX is shown as afunction of pH in Fig. 3 (see ESI Fig. S11 for Eu-FA and Tb-MTX). TheUV–Visible spectrum of each complex at acidic pH exhibits an intensiveband at around 290 nm. This was attributed to the pi-pi* transitionswithin the structure of each complex and agrees with previously re-ported absorptions of the p-aminobenzoate structure of FA [57]. Anumber of smaller emission bands were also observed between 330 and350 nm, which are typical absorption bands from the carbostyril con-jugate structure [48,58,59]. A less intensive n-pi* band was also ob-served at ~385 nm. On increasing pH, the pi-pi* absorption bands at~290 nm increased in molar absorptivity and underwent a bath-ochromic shift, the extent of which varied for each complex, indicatinga small increase in electron delocalisation. The two characteristic bandsfrom the carbostyril antenna, however, only underwent an increase inmolar absorptivity upon increasing pH; and no wavelength shift was

observed. The molar absorptivity of the n-pi* band at ~385 nm in-creased upon the addition of base to varying degrees for each complex.Absorbance versus pH plots (inset Fig. 3a and b for Tb-FA and Eu-MTX,respectively; see ESI Fig. S11 for Eu-FA and Tb-MTX) show that themolar absorptivity for each complex underwent small changes betweenpH 3 and pH 7. The pKa values of FA are: 2.4 (N1H+), 3.4 (α-COOH),4.8 (γ-COOH) and 8.0 (N3H/CO) [60]. From HPLC studies conducted, itis known that all the complexes exist as a roughly 1:1 ratio of α to γisomers. Therefore, the change in molar absorptivity below pH 7 maybe associated with deprotonation of either the α-COOH or γ-COOHgroup, the N3H/CO group (pKa reduced due to presence of Ln(III) ion)or a combination of these groups.

The fluorescence spectra for the titration of each of the complexeswas recorded upon incremental addition of base (NaOH) at an excita-tion wavelength of 355 nm (see Fig. 3c and d for Tb-FA and Eu-MTX,respectively; see ESI Fig. S12 for Eu-FA and Tb-MTX). A large structure-

Scheme 1. Synthesis of the four Ln(III) complexes; Eu-FA, Tb-FA, Eu-MTX and Tb-MTX.

Fig. 2. Structures of the two isomeric forms of Eu-FA or Tb-FAcomplexes.

Table 1Luminescence lifetimes and inner sphere hydration numbers (q).

Probe τ (H2O)/ms τ (D2O)/ms q

Eu-FA 0.722 ± 0.0 1.611 ± 0.1 0.8Tb-FA 0.936 ± 0.0 1.169 ± 0.0 0.9Eu-MTX 0.643 ± 0.0 1.943 ± 0.0 1.1Tb-MTX 0.921 ± 0.1 1.116 ± 0.1 0.8

Z. Du et al. Journal of Inorganic Biochemistry 178 (2018) 32–42

37

less band was observed at ~450 nm for each complex. A slight shoulderat 490 nm can be observed in the spectra of Tb-FA, Eu-MTX and Tb-MTX, but not in Eu-FA. The fluorescence emission of both Eu-FA andTb-FA was only slightly altered upon increasing pH. A slight increase inemission intensity was observed as the pH increased from pH 3 to pH 6,where the emissions plateaued, and the Tb-FA complex showed thelargest change in emission. The slight emission enhancement at acidicpH may be assigned to the protonation effect of either or a combinationof the α-COOH or γ-COOH from the FA (the pKa values of the twocarboxylates are 3.4 (α-COOH) and 4.8 (γ-COOH)) [60]. In contrast toEu-FA and Tb-FA, the MTX complexes Eu-MTX and Tb-MTX showed asignificant increase in fluorescence emission intensity at higher pH(Fig. 3c and d). This difference in fluorescence emission response be-tween the FA and MTX compounds may be attributed to the deproto-nation of the amine group at the 4-position of the MTX pteridine ringupon the addition of base. This effect was not significantly noticeable inthe UV–Visible spectra for either of the MTX compounds and suggeststhat the changes in the fluorescence emission may be due to a singletstate effect in either the carbostyril antenna or the aromatic structuresof MTX in both Eu-MTX and Tb-MTX.

The Eu(III) emission spectra of Eu-FA and Eu-MTX recorded atacidic pH displayed four characteristic emission bands from the deac-tivation 5D0 level to 7FJ (J = 1, 2, 3 and 4) (Fig. 4a and c). A plot of thehypersensitive ΔJ = 2 band as function of pH (Fig. 4a(inset) and c(inset)) showed that the Eu(III) emission intensity between pH 3 andpH 6 remained relatively constant for both the Eu-FA and Eu-MTXcomplexes, respectively. A slight reduction in Eu(III) emission intensitycan be observed between pH 6–8 for Eu-FA that is not observed for Eu-MTX. Above pH 8 a significant reduction in Eu(III) emission is observedfor both Eu(III) complexes, which is most likely caused by the depro-tonation of the metal bound water (Eu-FA and Eu-MTX were defined tohave one metal bound water molecule, see Table 1). However, a com-binatorial effect from deportation of the pteridine ring N3H/CO and theamide adjacent to the cyclen [61] can not be ruled out. The emissionspectra from the Tb(III) ions of Tb-FA and Tb-MTX demonstrated fourbands from the deactivation 5D4 level to 7FJ (J = 6, 5, 4 and 3) with aΔJ = 5 band showing the highest emission intensity (Fig. 4b and d).While the emission of Eu-FA and Eu-MTX was essentially unchangedbetween pH 3 and pH 8 (N.B. Eu-FA had a slight decrease between pH 6

and pH 8), the emission from Tb-FA and Tb-MTX decreased con-tinuously upon increasing pH, which was most notable above pH 9(Fig. 4b and d). The different emission response observed betweenpH 3–8 is presumably an effect of the ligand energy levels versus the Tb(III) energy levels undergoing a change. The quenching of the Tb(III)emission after pH 8 was attributed to the deprotonation of the metalbound water (both Tb-FA and Tb-MTX were calculated to have onemetal bound water molecule, see Table 1). Importantly, high emissionwas recorded over biologically relevant pH ranges of 4–8 for each of thecomplexes in water, which suggests that these complexes are suitablefor cell imaging.

3.3. Evaluation of cellular uptake and cytotoxicity

To assess the influence of the lanthanide ion and also to explore thedifferences between FA and MTX targeting, all four complexes wereevaluated for cellular uptake in a panel of: FR-negative non-malignantcells (MLO-Y4osteocyte cells, ST2 human mesenchymal stem cells,MC3T3 pre-osteoblasts cells, and HL7702 hepatocytes); and malignantcancer cell lines which were either FR-positive (HeLa cervical cancercells, 293 t embryonic kidney cancer cells, MDA-MB-231 breast cancercells, U251 glioma cells, U2OS osteosarcoma cells and RAW264.7 au-toimmune leukaemia cells) or FR-negative (A549 lung cancer cells, PC-3 prostate cancer cells, and CAL-27 human tongue squamous cell car-cinoma cells). The emission from each complex was easily recorded incells using either a blue (360–370), green (470–495) or red (540–550)laser filter. This choice of wavelength is important for cell biologists asit ensures that these complexes are able to match commonly availablelasers; a key parameter in the design of optical imaging agents. Fig. 5shows representative images taken from control (MLO-Y4), FR-negativemalignant (PC-3 and CAL-27) and FR positive malignant (HeLa) celllines following 24 h incubation with each of the complexes (see ESI forcomprehensive range of images in all cell lines evaluated, Figs.S13–S25).

For non-malignant cell lines, no obvious fluorescence enhancementwas observed following 24 h incubation at all wavelengths (360–370,470–495 and 540–550 nm) using any of the four Ln(III) complexeswhen compared with untreated cells. Furthermore, for two of the FR-negative cell lines evaluated, A549 (lung cancer) and PC-3 (prostate

Fig. 3. Changes in absorbance and fluorescence emission for complexes Tb-FA (0.5 × 10−5 M) and Eu-MTX (5 × 10−5 M) as a function of pH; a) UV–Visible spectra of compound Tb-FA(inset shows changes at 280 nm), b) UV–Visible spectra of compound Eu-MTX (inset shows changes at 280 nm), c) fluorescene spectra of compound Tb-FA (inset shows changes at455 nm, λex 355 nm), d) fluorescene spectra of compound Eu-MTX (inset shows changes at 455 nm, λex 355 nm).

Z. Du et al. Journal of Inorganic Biochemistry 178 (2018) 32–42

38

Fig. 4. Changes in lanthanide emission for each complex as a function of pH; λex 355 nm; a) Eu(III) emission from Eu-FA (5 × 10−5 M) with inset shows changes at 616 nm, b) Tb(III)emission from Tb-FA (0.5 × 10−5 M) with inset shows changes at 545 nm, c) Eu(III) emission from Eu-MTX (5 × 10−5 M) with inset shows changes at 616 nm, d) Tb(III) emission fromTb-MTX (0.5 × 10−5 M) with inset shows changes at 545 nm.

Fig. 5. Fluorescence images of cells fol-lowing 24 h incubation either in the absence(labelled control) or in the presence of oneof Eu-FA, Tb-FA, Eu-MTX or Tb-MTX takenfrom control (MLO-Y4), FR-negative malig-nant (PC-3), FR positive malignant (HeLa)and FR-negative cell lines CAL-27 (humantongue squamous cell carcinoma) cell lines.

Z. Du et al. Journal of Inorganic Biochemistry 178 (2018) 32–42

39

cancer), no emission increase was observed when incubated with any ofthe complexes as compared with untreated cells. However, all four Ln(III) complexes showed strong fluorescence emission in malignant FR-positive cells following 24 h incubation at all wavelengths as comparedwith untreated cells. This indicates that all four of the complexes weretaken up in the six FR positive cell lines evaluated. These findings de-monstrate that, as expected, all four Ln(III) complexes showed selectivecellular uptake for FR-positive tumour cells, compared with either thenon-malignant cells or the A549 or PC-3 (FR-negative) tumour cells.This data suggests that all four complexes enter the cell via the FR andthus validated our strategy for designing potent FA/MTX-tethered lu-minescent Ln(III) complexes as tumour cell-oriented imaging agents. Nodifference in the use of FA or MTX as targeting group or Eu(III) vs. Tb(III) as emissive metal, could be observed in emission intensity fol-lowing incubation in the malignant FR-positive cells, non-malignantcontrols or the malignant FR-negative A549 or PC-3 cell lines.

Interestingly, we did observe a difference in uptake between the FAand MTX tethered complexes, as assessed by changes in fluorescenceemission, in the FR-negative cell lines CAL-27 (Fig. 5 last column).Specifically, more emission was detected from CAL-27 cells incubatedwith the MTX-tethered complexes, Eu-MTX and Tb-MTX, comparedwith either untreated cells or the FA-tethered counterparts, Eu-FA orTb-FA. Negligible difference in fluorescence emission was detectedbetween untreated cells and those incubated with either Eu-FA or Tb-FA. As CAL-27 cells express only very low quantities of FR, this resultsuggests that the antifolate analogues (Eu-MTX and Tb-MTX) mayenter the CAL-27 cells via an alternate uptake mechanism. A noticeabledifference in cell counts can also be observed in Fig. 5 in CAL-27 cellstreated with either Eu-FA or Tb-FA analogues versus those treated withthe therapeutic MTX analogues, Eu-MTX and Tb-MTX. This suggeststhat MTX even when tethered to the lanthanide chelate may have re-tained a portion of its toxic therapeutic activity.

Given the anti-folate MTX has been widely used in clinical treat-ments, albeit with severe side effects due to its high toxicity in healthycells, we anticipated that the MTX-tethered Ln complexes, Eu-MTX andTb-MTX, might exert anti-neoplastic efficacy when they enter tumourcells, in addition to the above-mentioned tumour-oriented imaging ef-fects. We therefore initially observed the cellular morphology afterextended incubation times with all four Ln(III) complexes in the FR-positive HeLa cell line, as shown in Fig. 6. Importantly, significant celldeath was observed for cells incubated with either of the MTX-tetheredcomplexes, Eu-MTX or Tb-MTX, following a 48 h incubation. Cell deathwas also observed for the FA-modified complexes, Eu-FA and Tb-FA,but to a reduced degree compared with that of MTX-modified com-plexes. This may be attributable to the Ln(III) complexes having some

cytotoxicity in their own right. The increased cytotoxicity of MTX-modified complexes might be due to a synergistic effect of Ln(III) ion/chelate in close proximity to the MTX moiety, or the release of MTXfollowing degradation of the amide bond after a long incubation time.

To further understand the cytotoxic effects of these lanthanidecomplexes we investigated a panel of tumour cells (MDA-MB-231,HeLa, PC-3, CAL-27) and non-malignant cell (MLO-Y4) lines. The datain Table 2, illustrated that both the MTX-tethered Ln complexes had acytotoxic effect when compared to the FA-derived compounds. This wasconsistent with a role for the MTX moiety in these cytotoxic effects. Thecytotoxic effect of the four complexes was most noticeable on the twoFR + ve cell lines, MDA-MB-231 and HeLa, whereas the overall effecton proliferation for either PC-3 or MLO-Y4 cells as compared with theeffect of MTX alone was relatively limited. The MTX analogues showeda greater cytotoxic effect in the CAL-27 cells as compared with the FAanalogues. This correlates with the quantity of fluorescence detectedfrom each of the cell lines following 24 h incubation. Interestingly, bothEu-FA and Eu-MTX were found to have greater cytotoxicity than the Tb(III) analogues in the FR + ve cell lines, suggesting that the presence ofthe Eu(III) ion has had an influence. Moreover, Eu-MTX was found tohave a greater effect on cytotoxicity in both the FR + ve cell lines thanMTX alone. The differences in cytotoxicity between the ‘isostructural’Eu(III) and Tb(III) complexes was not expected and it is note-worthythat this difference is only observable in the FR + ve cell lines. Thecause of this difference in cytotoxicity is not clear, however the var-iance in pH sensitivity of the Eu(III) vs. Tb(III) analogues may hold aclue. These results are of course preliminary and additional con-firmatory studies will be required to explore this difference in influenceon cell survival. It is important to note that all four complexes werefound to be highly stable in acidic and basic aqueous environments withno loss of metal ion from the chelate observed.

To further explore the relationship between the Eu(III) ion and thetargeting moiety (FA or MTX) on activity, an MTS cell proliferationassay was performed on HeLa cells using Eu.1, Eu-FA and Eu-MTX, freeFA and free MTX (see ESI Fig. S26). This data shows that Eu.1 and Eu-FA have effects comparable to free FA. Eu-MTX was found to be morecytotoxic than free MTX in solution, in line with the anti-proliferativeassay, suggesting that there may be a synergistic effect between the Eu(III) ion/chelate and MTX.

4. Conclusion

Recent efforts to develop selective and sensitive methods for cancerimaging have focused on targeted imaging agents to detect cell surfacereceptors that have increased expression in cancer cells. This strategy

Fig. 6. Transmitted light images of HeLa cells following 24 and 48 h incubations with Eu-FA, Tb-FA, Eu-MTX and Tb-MTX.

Z. Du et al. Journal of Inorganic Biochemistry 178 (2018) 32–42

40

depends on binding affinity, receptor specificity and cancer specificreceptor expression and the development of highly emissive targetedimaging agents. We have successfully generated four targeted lumi-nescent lanthanide complexes that are highly stable under a range ofconditions. We have also shown that folate and MTX tethered com-plexes are not efficiently internalised into non-malignant cells.However, all four complexes were found to be internalised by FR + vecancer cell lines, confirming that both MTX and FA can be used astargeting groups for the folate receptor. As expected the MTX tetheredcomplexes were found to be more toxic than the FA, with the MTXcomplexes offering the potential to image the site of therapeutic ac-tivity. Interestingly the Eu(III) complexes were found to be slightlymore toxic than corresponding Tb(III) complexes. This set of complexesoffers cell biologists the opportunity to explore MTX toxicity in con-junction with FA action by incubating cells with both Eu-MTX and Tb-FA, as the emissions from these complexes can be easily resolved.Furthermore, the difference between the FA and MTX labelled com-plexes in the oral adenosquamous cell line (CAL-27) suggests that it isviable to explore differences in folate receptors using folate and anti-folate imaging. The ability to visualise and monitor the selective ac-cumulation of cytotoxic agents within a cancer cells will aid sig-nificantly in the development of new therapeutics.

Abbreviations

α alphaA549 human lung epithelial cellsAr aromaticβ betaBOP (benzotriazol-1-yloxy)tripyrrolidino-phosphoniumBSA bovine serum albuminca. approximatelyCAL-27 oral adenosquamous carcinoma cellsCH3CN acetonitrileCI computed tomographyconc. concentratedcyclen 1,4,7,10-tetraazacyclododecaned doubletδ chemical shiftDCM dichloromethane° degrees celsiusDIPEA N,N-DiisopropylethylamineDMEM Dulbecco's Modified Eagle MediumDMF N,N-DimethylformamideDMSO dimethylsulfoxideDNA deoxyribonucleic acidESI-MS electrospray ionization mass spectrometryELISA enzyme-linked immunosorbent assayEu(III) europium

Acknowledgment

This work was financially supported by the China-Australia Centrefor Health Sciences Research Program (No. 2015GJ07), the Science and

Technology Major Projects of Shandong Province (No 2016GSF201175to X. Li; Major Key Technology, No. 2015ZDJS04001 to F.S. Wang), andZD was supported by a University of South Australia President's post-graduate scholarship.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.jinorgbio.2017.10.003.

References

[1] M. Fenech, Mutat. Res. 475 (2001) 57.[2] S.J. Duthie, Br. Med. Bull. 55 (1999) 578–592.[3] S.J. Weinstein, T.J. Hartman, R. Stolzenberg-Solomon, P. Pietinen, M.J. Barrett,

P.R. Taylor, J. Virtamo, D. Albanes, Cancer Epidemiol. Biomark. Prev. 12 (2003)1271–1272.

[4] E. Stokstad, J. Koch, Physiol. Rev. 47 (1967) 83–116.[5] M.A. Sanjoaquin, N. Allen, E. Couto, A.W. Roddam, T.J. Key, Int. J. Cancer 113

(2005) 825–828.[6] S.M. Zhang, W.C. Willett, J. Selhub, D.J. Hunter, E.L. Giovannucci, M.D. Holmes,

G.A. Colditz, S.E. Hankinson, J. Natl. Cancer Inst. 95 (2003) 373–380.[7] L.E. Voorrips, R.A. Goldbohm, H.A. Brants, G.A. van Poppel, F. Sturmans,

R.J. Hermus, P.A. van den Brandt, Cancer Epidemiol. Biomark. Prev. 9 (2000)357–365.

[8] L.H. Matherly, Z. Hou, Vitam. Horm. 79 (2008), http://dx.doi.org/10.1016/S0083-6729(1008)00405-00406.

[9] R. Zhao, I.D. Goldman, Curr. Opin. Pharmacol. 13 (2013) 875–880.[10] C. Leamon, J. Reddy, Adv. Drug Deliv. Rev. 56 (2004) 1127–1141.[11] S. Sabharanjak, S. Mayor, Adv. Drug Deliv. Rev. 56 (2004) 1099–1109.[12] B. Kamen, M.T. Wang, A.J. Streckfuss, X. Peryea, R. Anderson, J. Biol. Chem. 263

(1988) 13602–13609.[13] Z. Hou, L.H. Matherly, Curr. Top. Membr. 73 (2014) 175–204.[14] A. Quintana, E. Raczka, L. Piehler, I. Lee, A. Myc, I. Majoros, A. Patri, T. Thomas,

J. Mulé, J. Baker, Pharm. Res. 19 (2002) 1310–1316.[15] M. Salazar, M. Ratnam, Cancer Metastasis Rev. 26 (2007) 141–152.[16] W. Moon, Y. Lin, T. O'Loughlin, Y. Tang, D. Kim, R. Weissleder, C. Tung, Bioconjug.

Chem. 14 (2003) 539–545.[17] G.L. Zwicke, G.A. Mansoori, C.J. Jeffery, Nanotechnol. Rev. 3 (2012), http://dx.doi.

org/10.3402/nano.v3403i3400.18496.[18] J.A. Ledermann, S. Canevari, T. Thigpen, Ann. Oncol. 26 (2015) 2034–2043.[19] Y. Lu, E. Sega, C.P. Leamon, P.S. Low, Adv. Drug Deliv. Rev. 56 (2004) 1161–1176.[20] V.B. Junyaprasert, S. Dhanahiranpruk, J. Suksiriworapong, K. Sripha,

P. Moongkarndi, Colloids Surf. B: Biointerfaces 136 (2015) 383–393.[21] R. Pignatello, G. Spampinato, V. Sorrenti, C. Di Giacomo, L. Vicari, J.J. McGuire,

C.A. Russell, G. Puglisi, I. Toth, Eur. J. Pharm. Sci. 10 (2000) 237–245.[22] L.K.A. Rahman, S.R. Chhabra, Med. Res. Rev. 8 (1988) 95–155.[23] H. Kantarjian, D. Thomas, S. O'Brien, J. Cortes, F. Giles, S. Jeha, C.E. Bueso-Ramos,

S. Pierce, J. Shan, C. Koller, M. Beran, M. Keating, E.J. Freireich, Cancer 101 (2004)2788–2801.

[24] G. Bonadonna, P. Valagussa, A. Moliterni, M. Zambetti, C. Brambilla, New Engl. J.Med. 332 (1995) 901–906.

[25] E. Guardiola, F. Peyrade, L. Chaigneau, D. Cupissol, X. Tchiknavorian, E. Bompas,A. Madroszyk, P. Ronchin, M. Schneider, J.P. Bleuze, J.Y. Blay, X. Pivot, Eur. J.Cancer 40 (2004) 2071–2076.

[26] P. Cipriani, P. Ruscitti, F. Carubbi, V. Liakouli, R. Giacomelli, Expert. Rev. Clin.Immunol. 10 (2014) 1519–1530.

[27] N. Hagner, M. Joerger, Cancer Manag. Res. 2 (2010) 293–301.[28] E.I. Sega, P.S. Low, Cancer Metastasis Rev. 27 (2008) 655–664.[29] M. Srinivasarao, C.V. Galliford, P.S. Low, Nat. Rev. Drug Discov. 14 (2015)

203–219.[30] G.M. van Dam, G. Themelis, L.M.A. Crane, N.J. Harlaar, R.G. Pleijhuis, W. Kelder,

A. Sarantopoulos, J.S. de Jong, H.J.G. Arts, A.G.J. van der Zee, J. Bart, P.S. Low,V. Ntziachristos, Nat. Med. 17 (2011) 1315–1319.

[31] M.D. Kennedy, K.N. Jallad, D.H. Thompson, D. Ben-Amotz, P.S. Low, J. Biomed. Op.8 (2003) 636–641.

Table 2The cytotoxic effects of Ln-tethered complexes.

MDA-MB-231a (IC50, μM) HeLaa (IC50, μM) PC-3a (IC50, μM) CAL-27a (IC50, μM) MLO-Y4a (IC50, μM)

Eu-FA 3.59 ± 1.33⁎ 1.67 ± 0.52⁎ 9.60 ± 2.52⁎ 8.97 ± 3.23⁎ 12.86 ± 4.61⁎

Tb-FA 5.24 ± 1.27⁎ 2.25 ± 0.86⁎ 10.04 ± 3.35⁎ 9.19 ± 4.09⁎ 12.79 ± 3.56⁎

Eu-MTX 0.26 ± 0.06⁎ 0.58 ± 0.79⁎ 7.33 ± 2.49⁎ 3.61 ± 1.45⁎ 9.07 ± 3.25⁎

Tb-MTX 1.90 ± 0.32⁎ 1.64 ± 0.79⁎ 7.85 ± 3.03⁎ 3.88 ± 1.21⁎ 9.22 ± 4.17⁎

MTX 1.60 ± 0.22 1.91 ± 0.17⁎ 2.82 ± 0.31 2.42 ± 0.14 2.91 ± 0.25

a Results are presented as the mean ± SD for at least three experiments for each group.⁎ p < 0.05 compared to the control (MTX).

Z. Du et al. Journal of Inorganic Biochemistry 178 (2018) 32–42

41

[32] M.P. Coogan, V. Fernandez-Moreira, Chem. Commun. 50 (2014) 384–399.[33] J.-C.G. Bünzli, Chem. Rev. 110 (2010) 2729–2755.[34] J.-C.G. Bünzli, Acc. Chem. Res. 39 (2006) 53–61.[35] E.G. Moore, A.P.S. Samuel, K.N. Raymond, Acc. Chem. Res. 42 (2009) 542–552.[36] R. Poole, G. Bobba, M. Cann, J. Frias, D. Parker, R. Peacock, Org. Biomol. Chem. 3

(2005) 1013–1024.[37] D.G. Smith, B.K. McMahon, R. Pal, D. Parker, Chem. Commun. 48 (2012)

8520–8522.[38] E.J. New, D. Parker, Org. Biomol. Chem. 7 (2009) 851–855.[39] C.P. Montgomery, B.S. Murray, E.J. New, R. Pal, D. Parker, Acc. Chem. Res. 42

(2009) 925–937.[40] S. Wang, P.S. Low, J. Control. Release 53 (1998) 39–48.[41] S. Wang, J. Luo, D.A. Lantrip, D.J. Waters, C.J. Mathias, M.A. Green, P.L. Fuchs,

P.S. Low, Bioconjug. Chem. 8 (1997) 673–679.[42] J.W. Lee, J.Y. Lu, P. Low, P. Fuchs, Bioorg. Med. Chem. 10 (2002) 2397–2414.[43] C. Müller, A. Hohn, P. Schubiger, R. Schibli, Eur. J. Nucl. Med. Mol. Imaging 33

(2006) 1007–1016.[44] Z. Du, G.N. Borlace, R.D. Brooks, R.N. Butler, D.A. Brooks, S.E. Plush, Inorg. Chim.

Acta 410 (2014) 11–19.[45] S. Quici, A. Casoni, F. Foschi, L. Armelao, G. Bottaro, R. Seraglia, C. Bolzati,

N. Salvarese, D. Carpanese, A. Rosato, J. Med. Chem. 58 (2015) 2003–2014.[46] S.N.A. Jenie, Z. Du, S.J.P. McInnes, P. Ung, B. Graham, S.E. Plush, N.H. Voelcker, J.

Mater. Chem. B 2 (2014) 7694–7703.

[47] A. Mustaev, M. Kozlov, S. Marras, L. Krasnoperov, L. Wirpsza, S. Pillai, WO PatentWO/2011/088,193, 2011.

[48] L.N. Krasnoperov, S.A.E. Marras, M. Kozlov, L. Wirpsza, A. Mustaev, Bioconjug.Chem. 21 (2010) 319–327.

[49] M.J. Waring, Bioorg. Med. Chem. Lett. 19 (2009) 2844–2851.[50] A. Harte, P. Jensen, S. Plush, P. Kruger, T. Gunnlaugsson, Inorg. Chem. 45 (2006)

9465–9474.[51] T. Gunnlaugsson, D.F. Brougham, A.-M. Fanning, M. Nieuwenhuyzen, J.E. O'Brien,

R. Viguier, Org. Lett. 6 (2004) 4805–4808.[52] O. Aronov, A.T. Horowitz, A. Gabizon, D. Gibson, Bioconjug. Chem. 14 (2003)

563–574.[53] S.F. Atkinson, T. Bettinger, L.W. Seymour, J.-P. Behr, C.M. Ward, J. Biol. Chem. 276

(2001) 27930–27935.[54] R.M. Supkowski, W.D. Horrocks, Inorg. Chim. Acta 340 (2002) 44–48.[55] W.D. Horrocks, D.R. Sudnick, J. Am. Chem. Soc. 101 (1979) 334–340.[56] D. Kovacs, X. Lu, L.S. Mészáros, M. Ott, J. Andres, K.E. Borbas, J. Am. Chem. Soc.

139 (2017) 5756–5767.[57] C. Lorente, A.H. Thomas, Acc. Chem. Res. 39 (2006) 395–402.[58] P. Ge, P.R. Selvin, Bioconjug. Chem. 15 (2004) 1088–1094.[59] J. Chen, P.R. Selvin, J. Photochem. Photobiol. A 135 (2000) 27–32.[60] Z. Szakács, B. Noszál, Electrophoresis 27 (2006) 3399–3409.[61] C.P. McCoy, F. Stomeo, S.E. Plush, T. Gunnlaugsson, Chem. Mater. 18 (2006)

4336–4343.

Z. Du et al. Journal of Inorganic Biochemistry 178 (2018) 32–42

42

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具