(Pentamethylcyclopentadienato)rhodium Complexes for Delivery of the Curcumin Anticancer Drug

A Journal of Accepted Article Title:Rhodium Pentamethylcyclopentadienato Complexes for Delivery of the Curcumin Anti-Cancer Drug Authors:Jack Markham, Jun Liang, Aviva Levina, Rachel Mak, Bernt Johannessen, Peter Kappen, Chris J Glover, Barry Lai, Stefan Vogt, and Peter Andrew Lay This manuscript has been accepted after peer review and appears as an Accepted Article online prior to editing, proofing, and formal publication of the final Version of Record (VoR). This work is currently citable by using the Digital Object Identifier (DOI) given below. The VoR will be published online in Early View as soon as possible and may be different to this Accepted Article as a result of editing. Readers should obtain the VoR from the journal website shown below when it is published to ensure accuracy of information. The authors are responsible for the content of this Accepted Article. To be cited as: Eur. J. Inorg. Chem. 10.1002/ejic.201601331 Link to VoR: http://dx.doi.org/10.1002/ejic.201601331 European Journal of Inorganic Chemistry 10.1002/ejic.201601331 FULL PAPER Rhodium Pentamethylcyclopentadienato Complexes for Delivery of the Curcumin Anti-Cancer Drug Jack Markham,[a] Jun Liang,[a] Aviva Levina,[a] Rachel Mak,[a] Bernt Johannessen,[b] Peter Kappen,[b] Chris J. Glover,[b] Barry Lai,[c] Stefan Vogt,[c] and Peter A. Lay*[a] Abstract: [RhIII(*Cp)Cl(X,Y)]n+ complexes (X,Y = Cl, PTA, n = 0 (2); amphiphilic nature due to the lipophilic arene ring and the X,Y = en, n = 1 (3, Cl- salt; 4, PF6 - salt); X,Y = acac, n = 0 (5); X,Y auxiliary ligands creating a more hydrophilic metal center can be = cur, n = 0 (6), where *Cp = pentamethylcyclopentadienato, curH = tailored for optimal solubility and hydrolysis/aquation kinetics. curcumin; PTA = 1,3,5-triaza-7-phosphatricyclo[3.3.1.1]decane; en = These complexes include: those with the general structure, 1,2-ethanediamine; acac = acetylacetonato = 2,4-pentanedionato(1-)) [RuX(η6-arene)(N,N’)]+ (X = Cl- or I-, (N,N’) = 1,2-ethanediamine were synthesized from [Rh(*Cp)(µ-Cl)Cl]2 (1). While 2-5 were inactive or N-ethyl-1,2-ethanediamine),[1] and RAPTA complexes with against human epithelial A549 lung cancer cells in assays of three monodentate ligands and a general structure, [RuX2(η6- cytotoxicity, and anti-metastatic and pro-apoptotic behaviours, 6 had arene)Y] (X = Cl-, Y = PTA in the case of RAPTA-C) (Figure 1).[6] similar a cytotoxic activity as curH over 72 hr, but at 24 hr in real- Some Ru(II) arene complexes show similar cytotoxicity as time cell migration assays, it was less active, showing slow release carboplatin, but have slightly lower activity than cisplatin against of curH. All complexes underwent ligand-exchange reactions with A2780 human ovarian cancer cell line.[7] There is no in-vivo or biomolecules and cells within the timeframes of the assays (X-ray in- vitro cross-resistance with cisplatin, which is likely due to absorption spectroscopy). Intracellular elemental distributions (X-ray fluorescence microscopy) showed that 6 effectively delivered curH to cells, where it was released. Other elemental distributions and caspase activities were consistent with pre-apoptotic activities. As such 6 is a promising delivery agent for bioactive ligands, such as curH. However, curcumin itself showed a previously unrecognised ability to promote migration at sub-toxic concentrations, which is a concern for its widespread use as a nutritional supplement and as a potential drug. This aspect warrants further research. Introduction Since the discovery and subsequent clinical success of cisplatin, interest has been ongoing on new metallodrugs, especially for cancer chemotherapy.[1,2] Other Pt drugs, i.e., oxaliplatin and carboplatin[3] are also in widespread clinical use.[1] Investigations of complexes of neighboring metals in the periodic table have led to very promising Ru drugs.[1] The most clinically advanced of these complexes are KP1019 and its Na+ salt, NKP1339, which are potent cytotoxins with low side-effects.[4] In early trials, NAMI-A was a promising drug with high anti-metastatic activity,[5] that reduced the spread and growth of new cancers. Another class of Ru anti-cancer drugs currently is Ru(II) arene complexes. Their highly flexible structural permutations and [a] J. Markham, J. Liang, A. Levina, R. Mak, P. A. Lay* School of Chemistry The University of Sydney NSW 2006, Australia E-mail: peter.lay@sydney.edu.au [b] B. Johannessen, P. Kappen, C. J. Glover Australian Synchrotron 800 Blackburn Rd, Clayton VIC 3168, Australia [c] B. Lai, S. Vogt Advanced Photon Source, Building 401 Argonne National Laboratory 9700 South Cass Ave, Lemont IL 60439, USA Supporting information for this article is given via a link at the end of the document. Figure 1. Representative Ru(II)-arene and Ru(III) N-heterocyclic anticancer drugs currently under development. This article is protected by copyright. All rights reserved. European Journal of Inorganic Chemistry 10.1002/ejic.201601331 FULL PAPER their different mechanisms of action.[8] Some examples of Ru(II) arene complexes are RM175, [Ru(acac)Cl(η6-p-cymene)], and RAPTA-C (Figure 1). A derivative of RM175, ONCO4417 (in which the Cl- counterion is substituted with PF6 -),[9] is also in development. RAPTA-C exhibits pronounced anti-metastatic properties, possibly as a result of its ability to bind to proteins,[10] such as cathespin b[11] (a cysteine peptidase responsible for the degradation of the ECM in metastasis). The promise of Ru(II)-arene drugs drove the exploration of drugs of metals adjacent to Pt and Ru on the periodic table. The very substitution inert d6 Rh(III) complexes were considered less promising candidates, but recently many Rh(III) complexes have exhibited anticancer activities.[12] However, the high degree of nephrotoxicity exhibited by most Rh complexes is an issue.[1] Much research has focused on dinuclear Rh(II) complexes and their cytotoxicities,[13] but the focus has since shifted to monomeric Rh(III) complexes.[12] The cancer inhibitory properties of Rh(III) chloride were first reported in 1953.[14] Later studies led to the development of mer-[RhCl3(NH3)3], which had strong anti-proliferative effects and prolonged the lifespan of mice with ascetic sarcoma 180.[15] Recently Rh(III)-*Cp complexes were synthesized by Dyson and colleagues[16] and screened as Rh-*Cp analogues of RAPTA-C but there was no significant improvement in efficacy compared with the isoelectronic Ru(II) complex. Replacement of three aqua ligands in [RhIII(OH2)6]3+ by an arene or *Cp ligand has a significant labilizing effect on the remaining ligands.[17] This labilizing effect is particularly pronounced for *Cp,[17] which is a significant step in increasing the reactivity of highly inert Rh(III)- halido bonds in spontaneous dissociatively activated aquation reactions.[18] After the first aquation reaction, the aqua ligand can undergo partial hydrolysis at physiological pH values, to further labilize other ligands via conjugate base mechanisms.[18b] In the current study, *Cp complexes, 2-6, were prepared from the dimeric precursor, 1 (Figure 2) and were screened for anti- cancer properties on the A549[19] non-small cell lung carcinoma Figure 2. Molecular structures of the Rh complexes investigated in this study, cell line. Complex 6, contains the biomolecule, curcumin, which and their numerical designations. has been widely researched as an anti-cancer drug.[20] Complex 6 was tested alongside curcumin at identical dosages, to differentiate the effect of the dissociated ligand from the effect of Complex 6 was unknown at the time of preparation, but was the complex. X-ray absorption spectroscopy (XAS) was since reported by Dyson, et al.[23] The synthesis for 6 was based conducted to follow the changes in speciation with biomolecules on those used for Ru(II)-arene-curcumin complexes.[24] The first as was previously deployed using XAS techniques,[21] since the step was chromatographic separation of pure curcumin from rates and modes of reactivity are important determinants of dimethoxy and bisdemethoxy derivatives in the commercial whether drugs are cytotoxic or anti-metastatic.[22] sample. In the synthesis of a Ru(II) complex,[25] the separation was performed after the synthesis, but for 6, curcumin was purified beforehand to reduce the waste of high-cost Rh. Results and Discussion Chromatographic purification of curcumin was performed via silica gel column using 99:1 DCM:MeOH as the mobile phase. Synthesis The synthesis 6 was similar to that used for Ru(II) complexes[24] with longer reaction times (48 hr instead of 24 hr) because of the slower substitution kinetics of Rh(III) relative to Ru(II).[18,26] The Complexes 2-7 were prepared in moderate to high yields (24– 1H and 13C NMR[23] spectra showed the expected resonances 94%) from 1. All compounds were fully characterized by FTIR, and 1H integration values were in agreement with a 1:1 UV-Vis and NMR spectroscopies and mass spectrometry, and *Cp:curcumin stoichiometry with no minor peaks to indicate the purities were confirmed by elemental analysis. All were in close presence of bisdemethoxy- and dimethoxy-curcumin complexes. agreement with literature data (ESI). This article is protected by copyright. All rights reserved. European Journal of Inorganic Chemistry 10.1002/ejic.201601331 FULL PAPER The IR spectra of the complex showed the typical bands due to compromised by apoptosis. Most likely, the complex is exploiting the ν(O-H) and ν(C=C, C=O) modes (3506, 1626, 1601 cm-1, the same transport mechanisms as neat curcumin, or is able to respectively, for the ligand) that shift to lower wavenumbers enter the cell due to the increased lipophilicity of the aromatic (3216, 1619, and 1591 cm-1) as a consequence of metal binding curcumin ligand. The fact that the Rh content in cells treated through both carbonyl groups of the acac moiety within curcumin. with 6 was an order of magnitude higher than any of the other Rh complexes shows that the bulk of the Rh is transported with Biological Assays curcumin bound. Complex 4 has the lowest uptake, which is attributed to the bulky, polar PTA ligand retarding cell uptake Different times and concentrations were chosen deliberately in due to poor interaction with the cell membrane. different assays to accentuate the features being examined and thus enable different aspects of the biological activities to be probed, as described under individual assays. MTT Assays In MTT cytotoxicity assays[27] on A549 lung cancer cells, both curcumin and 6 possessed equivalent IC50 values (curcumin = 32 ± 1 µM, 6 = 31 ± 2 µM, 72 hr to enable complete dissociation of the curcumin ligand). Thus the cytotoxicity of 6 was attributed to dissociation of curcumin from 6 with no synergistic effects of the Rh fragment. This was consistent with no cytotoxicity for 5, which possesses an identical co-ordination sphere with an acac ligand that underwent substitution on a similar timescale with similar products. Complexes, 2, 3 and 4 were not cytotoxic below the maximum experimental concentration (200 µM). This shows that this class of Rh complexes could be effective low- toxicity delivery agents for the slow release of cytotoxins. Table 1. IC 50 values calculated from MTT assay results for each Rh complex Figure 3. Comparison of cellular uptakes for Rh-*Cp complexes (10 µM) over and curcumin after 72 hr. a 4-hr treatment period at 37 ºC. Error bars represent standard deviations over two AAS or ICPMS measurements for three replicate cell treatments, giving a total of six replicate measurements. A control treatment (C) using untreated Complex IC50 (μM)[a] R2 Value A549 cells has been included for comparison. The * shows a P- value < 0.05; ** indicates a P-value < 0.01, compared to controls. 2 > 200 N/A Caspase-3/7 Apoptosis Assay 3 > 200 N/A Shorter reaction times (20 hr) than the MTT assay were used to 4 > 200 N/A maximize the number of intact cells. Concentrations near the IC50 values were chosen for 6 and curcumin (25 µM) and a much 5 > 200 N/A higher concentration (100 µM) for 2-5, was chosen to test for any signs of pre-apoptosis in the non-cytotoxic drugs. 6 31±2 0.995 Caspases 3 and 7 are activated by apoptotic processes. During Curcumin 31±1 0.996 apoptosis, mitochondrial potential is lost to enable release of caspases that cleave other cellular proteins and result in cell [a] Values of >200 μM indicate no cytotoxicity under the treatment conditions. death.[28] In these assays, 6 was highly effective in inducing Figures S1-6 (ESI) give plots of cell viability for each treatment are available. apoptosis at 25 μM with a large fraction of apoptotic/dead cells (85 ± 1%) but was somewhat less active than curcumin (Figure 4) Metal Uptake While these apoptosis assays indicate that the curcumin species It was important to ascertain whether the MTT assays arose from are more cytotoxic than indicated by the MTT assays, different different cell permeabilities. Sub-toxic concentrations and short cytotoxicity assays cannot be compared numerically because incubation times (10 µM, and 4 hr) were used so that the assays they measure different biochemical aspects. For instance, the were not complicated by binding to cell debris and dead cells. process of treating adherent cells with trypsin is likely to increase apoptosis, especially for cells that are already compromised by Even at these short reaction times and low concentrations, 2–5 the uptake of a drug. were sufficiently permeable to be cytotoxic, thus low cellular uptake was not the reason lack of activities. However, their Together with the permeability results, the results are uptakes were much lower than the active complex 6 (Figure 3). consistent with the cytotoxicity of 6 being due to intracellular The sub-toxic concentration and short reaction makes it highly dissociation of curcumin. Other complexes (2-5 at 100 µM) were unlikely that this was due to cell membrane integrity being ineffective at inducing apoptosis. Complexes 2 and 5 did induce This article is protected by copyright. All rights reserved. European Journal of Inorganic Chemistry 10.1002/ejic.201601331 FULL PAPER Figure 4. Caspase-3/7 apoptotic assays showing the proportions of apoptotic and dead cells within populations treated with the Rh-*Cp complexes, cisplatin and curcumin over 20 hr. Treatment concentrations were 100 µM for 2-5, 25 μM for 6 and curcumin, and 50 μM for cisplatin. Stock solutions of Rh-*Cp complexes and curcumin were prepared in DMSO, and resulted in a final concentration 1% DMSO. Cisplatin was prepared immediately before use in PBS. Error bars represent standard error values between three cell populations. apoptosis in a small fraction of their populations (6.3±0.2% and 8±1%, respectively) but given the high concentrations (100 μM) they are not cytotoxins. Cisplatin (50 µM) was included as a positive control in this assay using a concentration slightly below its IC50 value for A549 cells[29] (64 μM). This positive control showed that both curcumin and 6 were more effective in inducing apoptosis on A549 cells than was cisplatin. Migration Assays Assays on curcumin and 6 were undertaken at sub-toxic concentrations (10 µM) and shorter incubation times (40 hr) than the MTT assays, to minimize cytotoxicity complications. Epithelial Growth Factor (EGF) was included (100 ng mL-1) to stimulate cell migration and to increase the physiological Figure 5. Migration assays for 10 and 20 µM curcumin and 6 and controls. relevance in modelling metastasis, since in-vivo metastases are Relative wound density was calculated by (initial wound area/wound area)-1. stimulated by EGF upon proteolysis.[30] The first control (EGF, Error bars represent standard error between replicate treatments. DMF) contained identical concentrations of EGF (100 ng mL-1) which was outside the experimental error margin. Treatment and DMF (2 % v/v) as those included in other treatments, with the DMF originating from the stock solutions. The second control with 6 induced a similar effect; however, compared to curcumin, (EGF, -DMF), contained EGF (100 ng/mL) but no DMF to the effect was significantly diminished, which was attributed to the slow release of curcumin. evaluate the effect of the DMF on cell migration which was negligible, since both the first two controls had near-identical Higher concentrations (20 µM) near the curcumin IC50 value migration profiles. As such, a third control (-EGF, -DMF), with (~30 µM) caused some inhibition of cell migration (Figure 5). The neither DMF nor EGF and was not included in the plots. rate of wound healing was well below that of the controls that Unexpectedly, low curcumin concentrations (10 µM) stimulated included EGF; however, it is above the EGF-depleted control. cell migration, i.e., highest rate of wound healing (Figure 5), As with the 10 µM treatment, a similar effect was exhibited by 6, with a rate of wound healing slightly below that of EGF- This article is protected by copyright. All rights reserved. European Journal of Inorganic Chemistry 10.1002/ejic.201601331 FULL PAPER containing controls. The fact that both the migration-enhancing and the migration-inhibiting effects are reduced for 6, in comparison to curcumin, indicates that complexation to the Rh- *Cp complex results in the slow release of curcumin in vitro. The lack of cytotoxicity of 2, 3, and 5 enabled much high concentrations (100 µM) to be assayed to determine whether Rh*Cp complexes might inhibit migration. Again, there was a surprising slight increase in wound healing for the Rh complexes compared to the positive controls (Figure S7, ESI), but these concentrations are not relevant to the use of 6 as a cytotoxin. K-edge Rh X-ray Absorption Spectroscopy (XAS) XAS was performed on all complexes, 1-6, and a crown thiaether model complex, [Rh(*Cp)(9S3)](PF6)2] (9S3 = 1,4,7- trithiacyclononane, 7) diluted by 1:2500 (w/w) with BN (Figures S7-12, ESI). Linear combination fitting (LCF) of K-edge Rh XAS (calibrated to 23.2220 keV)[31] from reactions of 2 or 6 in HEPES buffer, cell culture medium (Advanced DMEM), protein solutions (FBS, HSA, BTf, e.g., Figure 6) and A549 cells, used literature methods.[21] Examples of XAS fits from reactions of 6 are in Figures 6 and 7 and the rest are in the ESI (Tables S1-S2, Figures S13-18). In HEPES buffer, the major XAS contributions were from complexes with O-donor-rich coordination spheres (5 and 6). This most likely indicated some ligand exchange of curcumin by aqua/hydroxido ligands. The XAS contribution from the chlorido-bridged dimer indicated some degree of Cl- binding, most likely due to the high [Cl-] (150 mM) in HEPES. The inclusion of a small weight of XAS from the S-donor complex, 7, is probably due to chlorido species that are not as well matched with the current XAS model complexes, given the absence of S- containing components of HEPES. In the HEPES + BSA solution, XAS from O coordination decreased relative to that from the S-coordinated complex (7) compared to spectra in HEPES. This most likely indicated binding of Rh to BSA thiol and/or methionine groups. The remaining O-binding could result Figure 6. Comparison of LCF plots for 6 (100 μM) in DMEM with and without from either protein binding or aquation. The treatment in HEPES FBS (10 %) (4 hr, 37°C). Reduced χ2 = 8.9 × 10-5 for DMEM, 8.0 × 10-5 for + BTf solution gave similar results to those with HEPES + BSA, DMEM + FBS. Best fits for both samples were very similar, with an increase in S and N donor weights in the FBS sample, indicating Rh-protein complexation. but with the addition of some N-donor character (3), which suggested further protein binding, as no other source of N- donors was present in solution. In Ru complexes the binding to serum albumin increased anti-metastatic activities,[22] but this is not the case for the Rh adducts (see later). XAS fitting for 6 in DMEM indicated some binding to N- and S- donors occurred, as shown by the presence XAS from complexes 3 and 7 in the best fits. DMEM contains a high concentration of thiol- and amine-containing amino acids for cell nutrition (L-methionine and L-cysteine for S-donation, all amino acids possess terminal amines). The high percentage of O donors (6) contributing to the best fit of the XAS was likely a result of coordination of oxygen atoms from amino acids, proteins, small molecules, such as citrate, and perhaps some aquation products that had not undergone further substitution. weight of XAS contributions from the N-donor complex to the fit, The best fits for the DMEM + FBS treatment conditions showed relative to that from the DMEM treatment, implied additional similar results to the DMEM-only treatment. The increased binding to FBS proteins. This article is protected by copyright. All rights reserved. European Journal of Inorganic Chemistry 10.1002/ejic.201601331 FULL PAPER Figure 9. LCF weights of Rh-*Cp model complexes for 2 (100 µM) treated in cell culture medium (4 hr, 37˚C, DMEM) and HEPES buffer and proteins (10% FBS, 100 µM BSA, 100 µM BTf). Also included is the bulk cell (A549) sample Figure 7. LCF weights of Rh-*Cp model complexes for 6 (100 µM) treated treated with 2 (200 µM, 24 hr). LCF was performed with a fitting range of -100 under biologically simulating conditions (4 hr, 37˚C) in cell culture medium to 400 eV, relative to the sample’s specific edge energy (E0). Reduced χ2 (DMEM) and HEPES buffer with various proteins (10% FBS, 100 µM BSA, 100 values: HEPES, 3.5 × 10−4; HEPES = BSA, 1.1 × 10-4; HEPES + BTf, 8.9 × µM BTf). Also included is the bulk cell (A549) sample treated with 6 (200 µM, 10−5; DMEM, 8.2 × 10-5; DMEM + FBS, 9.9 ×10-5; A549 cells, 2.2 ×10-3. 24 hr). LCF was performed with a fitting range of -100 to 400 eV, relative to the sample’s specific edge energy (E0). Reduced χ2 values: HEPES, 8.7 × S-donors dominate. These donors must be mainly derived from 10−5; HEPES = BSA, 7.9 × 10-4; HEPES + BTf, 2.8 × 10-4; DMEM, 8.9 × 10-5; intracellular biomolecules, since if they were from extracellular DMEM + FBS, 8.0 ×10-5; A549 cells, 5.1 ×10-3. reactions, this would yield similar Rh cellular contents for 6 and the rest of the complexes, which is not the case. LCF fits to the XAS from reaction products of 2 are in Figures 8 and 9 and the ESI (Tables S1-S2, Figures S19-24). In the XAS fits of 2 treated with HEPES, the minor weight of the XAS from 2 showed that the PTA ligand was easily substituted, most likely by Cl-, which was present in high concentration (150 mM). This was supported by the high fit weight of the XAS from the dimer complex, with a Rh*CpCl3 coordination sphere. In the HEPES + BSA treatment, the best fit had a high weight of the XAS from S- donor groups (7) and a smaller weight from the XAS of N-donor groups (3) which showed formation of Rh-BSA adducts. In the HEPES + BTf treatment (Figure 8), the XAS fits were similar; strong Rh-BTf binding was indicated by the high weight of XAS from the S-donor model complex. A small portion of the Rh-Cl binding was also apparent, implying a lower proportion of protein binding than in the HEPES + BSA treatment were. A Figure 8. LCF fitting plot for 2 (100 µM) in HEPES with BTf (100 µM, 4 hr, small increase in S and N complexation was apparent from the 37˚C). Reduced χ2 = 8.9 × 10−5. The residual oscillations within the red box XAS fits for the FBS-containing solution, compared to the DMEM indicate the presence of a Rh-*Cp coordination environment not represented solution, which indicated the formation of FBS protein adducts. by any available model complexes. XAS from 2 with A549 cells had high contributions of XAS from The best fits to the XAS from A549 bulk cells showed extensive O- and S-donating model complexes (6 and 7), which provided curcumin dissociation and protein/biomolecule complexation evidence of Rh-protein adduct formation. However, the XAS had with S-donor ligands. The higher levels of noise, even after ten a higher level of noise, which made the fitting results somewhat scans, did not allow minor components to be deduced with less reliable than those obtained from the other samples. In certainty, but this does not alter the major finding that binding to none of the reactions was there evidence for *Cp ligand loss. This article is protected by copyright. All rights reserved. European Journal of Inorganic Chemistry 10.1002/ejic.201601331 FULL PAPER Figure 10. Ca and Zn elemental maps of A549 cells after treatment with (A) 6 (30 μM, 24 hr), (B) 5 (30 μM, 24 hr), (C) curcumin (30 μM, 24 hr) and (D) DMSO (1% v/v, 24 hr). Maps were taken using a 10.4 keV excitation energy in fly scan mode, with a step size of 0.5 μm, a beam size of 0.4 μm and a dwell time of 300 ms and processed using the MAPS software (v 1.70). X-ray Fluorescence Microscopy (XFM) Experiments were performed at: 10.4 keV (Figures 10 and S25- 49, ESI) to enhance the signals from the lighter elements; and above the Rh K-edge (23.7 keV, Figures 11 and S50-59, ESI) to obtain Rh maps and correlate them with the other elements. The time of incubation was chosen so as to maximize the Rh content, Figure 11. S, K and Rh elemental maps for A549 cells treated with 6 (30 μM, while maintaining enough viable cells at the IC50 values. 24 hr) (A) and 5 (30 μM, 24 hr) (B). Images were taken using a 23.7 keV In the elemental maps collected at 10.4 keV, the control cells and excitation energy in step scan mode (step size, 1 μm; beam size, 0.4 μm; a dwell time 10 s) and processed using the MAPS software (v 1.70). cells treated with 5 had the same levels of intracellular Ca This article is protected by copyright. All rights reserved. European Journal of Inorganic Chemistry 10.1002/ejic.201601331 FULL PAPER densities (4.1 ± 0.2 and 3.9 ± 0.2, respectively, ratios of cell:background densities, Figure 12). Cells treated with curcumin have significantly lower intracellular Ca density (2.4 ± 0.3), and cells treated with the less potent 6 shown increased intracellular Ca density (4.8 ± 0.3), compared to that found in the control cells. Acting as a secondary messenger, intracellular Ca is extensively involved in various biochemical processes that are critical to tumor progression.[32] Cells undergoing early stage apoptosis have significantly increased cytosolic Ca levels.[33] Therefore, the increased intracellular Ca densities observed from 6-treated cells, suggested that its anticancer activity was likely to be due to apoptosis. However, the spatial and temporal distributions of intracellular free Ca2+ are strictly regulated by various membrane-bonded protein complexes and modulation of their activities via chemical or physical means were investigated in various pre-clinical studies.[32b,34] The reduction of intracellular Ca levels in the curcumin-treated cells suggested the biological activities of curcumin might arise from interfering with intracellular Ca homeostasis.[35] In addition to Ca, intracellular Zn levels decreased in cells treated with curcumin only, or Rh complexes, as compared to those amounts found in the control cells. The specifics of the relationship between cytosolic Zn concentrations and apoptosis are still unclear. While high levels of Zn can induce apoptosis in some cancers, Zn can be a protective agent against apoptosis in Figure 12. Elemental densities of Ca and Zn in A549 cells after treatment with others.[36] In the case of Bcl-2 induced apoptosis there is down- 5, 6 and curcumin. Densities are expressed as the ratio between the mean cell density and the mean background density (both in μg.cm-2). * Indicates a P- regulation of Zn during the apoptotic process, consistent with the value < 0.05; ** indicates a P-value < 0.01 compared to the control cells results we obtained.[37] However, this is not necessarily observed in early stages of apoptosis, which can somewhat paradoxically other pharmacokinetic characteristics reduces its efficacy. [20] The ability of the Rh(III)-*Cp vehicle to deliver curcumin into cancer be associated with large increases of intracellular Zn cells, where the drug is released, makes it a very promising concentration and Zn is also liberated from proteins and other cell components.[36b,38] This is a possible explanation for the high candidate for selective delivery of such drugs. In this respect, this class of Rh anti-cancer drugs appear to differ substantially levels of Zn observed in the cells treated with 6. from Ru analogues, since unlike these Rh-protein adducts, XFM at 23.7 keV revealed a difference in Rh content between which show no activities, Ru-protein adducts are strongly cells from cells treated with 5 or 6. Corroborating the metal involved in the anti-metastatic activities[22] and are also likely to uptake studies on bulk cells, the cells treated with 6 had a higher play a role in cytotoxicity through intracellular binding to proteins Rh-content than those treated with 5 (6 = 1.16±0.01, 5 = with thiol containing groups at their active sites.[39] 1.06±0.01, P < 0.05, Figure S60, ESI), although the difference is However, the migration studies suggest that curcumin-based much smaller than that observed in the bulk metal uptake drugs need to be considered with some caution as they increase studies. This may be due to the much smaller sample size for the the rate of migration at low concentrations. Thus low (non- studies with single cells and/or a selection of a healthier sub- cytotoxic) concentrations of curcumin may make cancer cells group of cells for imaging for 6. There does not appear to be any more aggressive and increase metastases and, hence, higher specific localization of Rh within the cell for either treatment. doses are required to induce apoptosis. Despite extensive There was co-localization of the Rh and S maps, which was studies, this surprising result has not been observed previously consistent with the XAS results that indicated that 6 reacts with cancers cells and warrants further investigation, as it has intracellular with thiol donors to facilitate the release of curcumin from the RhIII carrier. many implications in the widespread consumption of curcumin supplements and its use as an anti-cancer drug. This result is consist, however, with the recent observation that low levels of curcumin increases the proliferation and migration of non- Conclusions cancerous olfactory ensheathing cells.[40] While most of the Rh complexes exhibited little biological Irrespective of whether curcumin is or is not a viable anti-cancer activity, the slow release of the ligands and the lack of toxicity drug, the Rh*Cp delivery system has considerable potential for make them suitable drug delivery systems for cytotoxins that are delivering other anti-cancer drugs. too toxic to be delivered directly, or have poor pharmacokinetics. In this case, curcumin has generated considerable interest as a potential cytotoxin for cancer chemotherapy, but its solubility and This article is protected by copyright. All rights reserved. European Journal of Inorganic Chemistry 10.1002/ejic.201601331 FULL PAPER Experimental Section First, A549 cells were sub-cultured and suspended in DMEM (as above) then diluted to 3 x104 cells mL-1. A 96-well Falcon tissue culture plate was prepared with 100 μL of the cell suspension in each well. The plates Synthesis were then incubated at 310 K for 24 hr as cells adhered and reached confluence. The old medium was removed and replaced with medium Synthesis of complexes 1-5 and 7. The synthetic procedures and containing the relevant Rh complex. Cells were treated with a wide range characterization data for 1–5 and 7 are based on those in the of concentrations for each complex, with a maximum concentration of literature[16,41] and are contained within the ESI. 200 μM. The concentration was serially diluted in half with each sample, giving a final (lowest) concentration of 0.00153 μM (1.53 nM). Control [Rh(*Cp)Cl(cur)] (6) wells containing only the solvent used for stock solution preparation. The plates were then incubated for a further 72 hr. After 72 h incubation the At the time of preparation, 6 had not been reported and its synthesis was spent medium was replaced with 100 µL fresh medium plus 25 µL MTT based on that for a Ru(II)-p-cymene complex of curcumin.[24] A somewhat solution (5 mg/mL in PBS), followed by an additional incubation of 4 h. different synthesis was subsequently reported by Dyson, et al.[23] and The MTT-containing medium was removed, and the formazan crystals their characterization data are supplied below for comparison. produced by the viable cells were solubilised in DMSO (100 µL per well). Curcumin was purified by chromatography a silica column (Ajax The plates were read at 490 nm (λmax value of formazan,[42] acquisition Finechem, 40 – 63 µm) using a 99:1 DCM:MeOH mixture as an eluent. time 1.0 s) using a Perkin-Elmer Victor3 Multilabel Plate Reader, after Pure curcumin (119 mg, 0.32 mmol) was dissolved in MeOH (25 mL) and shaking to ensure complete formazan dissolution. The MTT assay data the bright orange solution was flushed with Ar. The inert Ar atmosphere were processed using OriginLab’s OriginPro 2015 software. Data were was maintained until the removal of the solvent via rotary evaporation. fitted using a standard dose-response model[43] to determine half The addition of NaOMe (17.2 mg, 0.32 mmol) afforded a sudden colour maximal inhibitory concentration (IC50) values. For experiments that did change to dark red. After 1 hr of stirring, complex 1 (100 mg, 0.16 mmol) not show a response in cell viability, IC50 values are listed as above 200 was added. The resulted solution was stirred for 48 hr during which time μM. Each assay was performed on six separate 96-well plates, each with the solution colour became significantly lighter. The solvent was removed three replicates of each concentration with a total of 18 measurements via rotary evaporation (337 mBar, 333 K) and the yellow solid was per concentration of each complex. dissolved in DCM and filtered to remove NaCl. The solvent was again removed by rotary evaporation, and the crude product was purified via Metal Uptake recrystallizing from hot MeOH and the purity of resulted product (50 mg, 24% yield) was confirmed by the absence of additional resonances Atomic absorption spectrophotometry (AAS) measurements were taken corresponding to free curcumin ligand in its 1H NMR spectrum. an Agilent Technologies 200 Series AA, equipped with a GTA 120 Graphite Tube Atomizer and a Rh lamp at 343.5 nm at a slit width of 0.5 1H-NMR (300 MHz, CDCl3) δ 7.58 (d, J = 15.3 Hz, 2H), 7.06 (d, J = 13.8 nm. Results were made from a calibration curve using dilutions of a 50 Hz, 4H, Ar-H), 6.90 (d, J = 7.4 Hz, 2H, Ar-H ), 6.51 (d, J = 16.1 Hz, 2H ppb standard (AccuStandard, USA), with a minimum concentration of 5 C(3,3’)H ), 5.78 (s, 2H, Ar-OH), 5.43 (s, 1H, -CH-), 3.94 (s, 4H, O-CH3), ppb. All data were processed using Microsoft Excel.[44] ICP-MS was 1.70 (s, 15H, Ar-CH3). 13C NMR (75 MHz, CDCl3) δ 178.48, 146.98, performed using a Perkin Elmer Nexion 350 X (plasma gas and nebulizer 146.71,138.21, 128.64, 126.50, 122.31, 114.65, 109.22, 92.14, 92.01, gas flow rates, 0.96 L min-1; Pt sample and skimmer cones; RF power, 55.96, 8.58; UV-Vis, λmax (nm), ε (M-1 cm-1): 404 (8.58 x 107) ESI-MS: m/z 1500 W; isotopes monitored, 103Rh and 193Ir (internal standard, Choice = 605.11 [C31H35ClO6Rh]+ FTIR (cm-1): 3216 wb, 1619 m, 1591 m, 1504 s, Analytical, 99.99%); dwell time, 90 ms AMU-1). Standards were prepared 1403 m, 1369 m, 1277 s, 1238 s, 1160 s, 1119 s, 1030 m, 990 m, 966 m, using a Hamilton diluter. 842 m, 803 m, 726 w, 554 w, 472 m. Anal. Calc. C31H34ClO6Rh·2H2O C, 58.09; H, 5.66; Cl, 5.24; Found. C, 57.74; H, 5.51; Cl, 5.30 %. Cultured A549 cells were seeded in six-well cell culture plates at a density of 105 cells mL-1 and allowed to adhere overnight. Rh complexes Literature:[23] 1H NMR (CDCl3, 293 K):δ7.57 (d, 2H, C(4,4′) H of cur, 3J were dissolved in serum-free DMEM FBS-deficient cell culture medium at trans = 16 Hz), 7.05 (m, 4H,C(6,6′)H and C(10,10′)H of cur), 6.90 (d, 2H, a concentration of 10 μM. Old medium was removed and the cell C(9,9′)H of cur, 3J arom H–H = 9 Hz), 6.51 (d, 2H, C(3,3′)H of cur, 3J monolayer was washed with PBS (3 mL) before the addition of the Rh- trans = 16 Hz), 5.78 (s, 1H,C(1)H of cur), 3.94 (s, 6H, OCH3 of cur, 1.69 containing medium (3 mL per well). The cells were then incubated (37°C, (s, 15H, CH3 *Cp ). 13C-NMR (CDCl3, 293 K):δ178.7 (s, C(2, 2′) vO of 5% CO2) for 4 hr. After incubation, the Rh-containing medium was cur), 147.2 (s, C(8,8′) of cur), 146.8 (s, C(7,7′) of cur), 138.4 (s, C(4,4′) of removed and the monolayer was washed thoroughly with PBS (3 x 3 mL). cur), 128.8 (s, C(3, 3′) of curc), 126.7 (s, C(5, 5′) of cur), 122.5 (s, Cells were detached via trypsinization to form a suspension and cell C(10,10′) of cur), 114.8 (s, C(9,9′) of cur), 109.4 (s, C(6, 6′) of cur), 102.2 numbers were counted were performed. Cell suspensions were then (s, C(1) of cur), 94.7 (d, C *Cp , J(103Rh–13C) = 6.9 Hz), 56.1 (s, OCH3 of transferred to 1.5 mL centrifuge tubes and centrifuged (2000 rpm, 3 min) cur), 8.8 (s, CH3*Cp ). IR (cm−1): 3563 w, 3157 m, 1622 m, 1590 m,1505 to form a cell pellet. The supernatant was removed and cells were s, 472 m, 454 w, 242 s. ESI-MS m/z = 605 [Rh(*Cp)(cur)]+. digested in concentrated HNO3 (100 μL, 15.8 M) over 24 hr. AAS and ICPMS-determined concentrations of Rh in nitric acid digests were Cell Culture Methods divided by the number of cells present (measured using the Countess haemocytometer immediately after treatment). The measured Rh Human non-small-cell lung carcinoma (A549) cells (ATCC) were cultured concentrations in cell lysate were normalized against the cell numbers in monolayers. Advanced DMEM was used as the cell culture medium. and expressed as ng Rh (or nmol) per 105 cells. For each metal complex Cells were grown in a humidified atmosphere of 5% CO2/95% air at 310 treatment, three replicate wells were prepared. For each well, two AAS K and sub-cultured every 2-3 d to approximately 90% confluence and one ICPMS measurement were taken. MTT Assays Caspase-3/7 Apoptotic Assay This article is protected by copyright. All rights reserved. European Journal of Inorganic Chemistry 10.1002/ejic.201601331 FULL PAPER Caspase assays on A549 cells were performed on a Merck Millipore Sample Preparation. Mixtures of Rh complexes with BN (w/w ratio of MUSE™ Cell Analyzer flow cytometer (ESI Figure S61). Cells were 1:2500) were used as standards. This ratio was determined by scanning cultured, as above, before being treated with Rh complexes for a period of mixtures with Rh compounds at Rh:BN ratios of 1:100, 1:1000, 1:2500 of 24 hr under incubation, as well as relevant controls (curcumin, cisplatin and 1:10000. The 1:2500 ratio had adequate signal-to-noise, as well as and DMSO). Rh-*Cp stock solutions were prepared in DMSO low levels of self-absorption (< 1% jump in the ratio of absorbance to immediately before use and resulted in a 1% v/v DMSO concentration fluorescence over the edge energy). upon treatment. A cell suspension (50 μL, 1 x 106 cells mL-1) of treated cells was isolated and treated with MUSE™ Caspase-3/7 Reagent (5 μL, Complexes 6 and 2 were chosen for XAS studies (100 μM, 4 hr, 310 K) 1:8, stock solution in PBS). After mixing, the suspension was incubated with proteins, and biological fluids: (1) HEPES-buffered saline, (2) (30 min, 310 K, 5% CO2), followed by the addition of MUSE™ Caspase HEPES-buffered saline + bovine Tf (100 μM), (3) HEPES-buffered saline 7-AAD working solution (150 μL). After mixing and further incubation (5 + BSA (100 μM), (4) DMEM and (5) DMEM + FCS (10%). After treatment, min) the samples were loaded into the flow cytometer using a 1.5 mL all solutions were snap-frozen in solid CO2 before being freeze-dried for plastic centrifuge tube and measurements were taken. Cells were divided up to 96 hr to remove all water. The dried samples were mixed in their into four populations based on their fluorescence intensity corresponding sample tubes using a spatula to ensure homogeneity before their XAS to the two reagents. Cell viability corresponded with increased 7-AAD were taken. No dilutions of the solid samples with BN were necessary. reagent (Merck Millipore) fluorescence, and apoptosis with increased Caspase-3/7 reagent (Merck Millipore) fluorescence intensity. A549 human adenocarcinoma cells were grown and maintained in A- DMEM in the same manner described previously. A 75 cm2 flask of Migration Assay confluent cells was prepared and the A-DMEM was replaced by A-DMEM containing high concentrations (200 μM) of either 2 or 6. The flask was A549 cells were cultured as above. A cell suspension (2 x 104 cells mL-1) returned to the incubator for 24 hr. After incubation, the Rh-containing was prepared and aliquots (100 μL) were transferred into 60 wells of a medium was removed, and the cell monolayer was rinsed with PBS and 96-well pre-coated Image Lock plate (Essen Bioscience).[45] The plate trypsinized. Cells were then centrifuged (2000 rpm, 3 min), the was incubated at 310 K (5% CO2/95% air) for 96 hr to allow cells to supernatant was decanted and the solid cell pellet was flash-frozen using adhere and reach confluence. The medium was replaced with DMEM frozen CO2. Once frozen, the bulk cell samples were freeze dried and containing 0.5% FCS approximately 5 hr before treatment. Wounds were stored in the desiccator before transport to the synchrotron. made in the cell monolayer using a Essen 96-Well Wound Maker. The wells were quickly washed with FCS supplied A-DMEM (50 μL, 0.5% Data Collection and Analysis FCS) to remove the detached cells, then filled with more A-DMEM (75 μL). Treatment solutions were prepared in the same medium at 4x the Rhodium K-edge spectra were recorded at the XAS beamline at the desired final concentration from stock DMF solutions (final concentration Australian Synchrotron.[47] Small amounts of sample (ca. 20 mg) were 2% v/v). A stock solution of EGF in sterilized (MilliQ) H2O was added to loaded into a plastic sample holder as a tightly packed pellet and secured give a final concentration of 100 ng mL-1 in all but the EGF control on both faces using Kapton tape. Sample holders were inserted into a column. Aliquots (25 μL) of the stock solutions were added to the closed-cycle He-cooled cryostat that maintained a temperature ≤ 10 K. corresponding wells. The 96-well plate was transferred to the Incucyte Data were recorded in fluorescence mode, using a 100-element Ge ZOOM Live Cell Imaging system where they were kept under incubation fluorescence detector (Eurisys). Beamline energies were controlled using (310 K, 5% CO2) and imaged and automatically processed to determine a Si(311) crystal monochromator. A foil of metallic Rh was used the size and density of the wound every 2 hr for a total of 38 hr. Assays throughout the experiments as an internal reference, and reference data were monitored remotely using the Incucyte ZOOM software. Changes were recorded simultaneously with the sample data. Repeat are expressed as a fraction of original wound area. Image processing measurements were taken at a minimum distance of 0.2 mm apart, to parameters were adjusted manually until the automatically determined reduce any potential photo-damage caused by overexposure to the wound area matched the actual area of the wound satisfactorily. sample in one position that may cause changes in sample composition or Rh oxidation state,[48] although no such effects were observed for repeat Several wells were excluded from analysis due to results with either zero scans at the same position. For model compounds, three scans were migration, or a rate of migration significantly different to those in identical taken (scan time ca. 1 hr) with an energy range of 23.02–24.10 keV; k = treatment wells; however, each treatment condition had at least five 16 Å-1, step sizes, 9 eV up to 23.20 keV, 0.4 eV from 23.20–23.27 keV replicates. Data from wells with identical treatment conditions were and 0.035 Å-1 in k space above 23.27 keV). Calibration was achieved by merged, and relative wound density measurements were plotted against setting the first peak of the first derivative of the Rh foil spectrum at time. Each treatment (with the exception of two controls) was supplied 23.2220 keV.[31] Calibration and merging of all XAFS data were achieved with epithelial growth factor (EGF, 100 ng mL-1, Sigma-Aldrich) in order to using the Demeter suite[49] (ver. 0.9.21). Normalisation and linear stimulate cell migration. This is highly relevant to studies of metastasis, combination fitting (LCF) analysis were also performed using Demeter.[48] as cancer cells are stimulated by biological EGF when they initially A fitting range that extended from 100 eV below the sample edge energy extravade from their point of origin into the extracellular matrix, and more to 400 eV above was used for LCF, as this range captured the majority of aggressively migratory cancers have been shown to greatly over-express spectral features but excluded some recurring pre-edge glitches and the epithelial growth factor receptor (EGFR).[46] Curcumin and 6 were glitches observed above 24 keV. LCF results that included negative applied at concentrations significantly below the IC50 values determined weights for any model compound were discarded and the best fits previously (i.e., 10 and 20 μM). The remaining Rh complexes were produced the lowest error (χ2) value. Spectra were plotted using Origin applied at high concentrations (100 μM). The first control contained DMF 2015 and LCF analysis weights were graphed using Microsoft Excel.[44] at the same concentration test solutions (0.1 %) as well as the same K-edge X-ray fluorescence spectra of all Rh model complexes exhibited EGF concentration used. The second control had no DMF, but did an absorption edge at approximately 23,220 eV with strong post-edge contain EGF, and the third control had neither EGF nor DMF. peaks in the vicinity of 23,240, 23,300 and 23,370 eV. Significant differences were observed for each model complex, with peak positions X-ray Absorption Spectroscopy (XAS) migrating up to 10 eV due to differences in co-ordination environment. This article is protected by copyright. All rights reserved. European Journal of Inorganic Chemistry 10.1002/ejic.201601331 FULL PAPER X-ray Fluorescence Microscopy (XFM) pixels above a certain threshold intensity until the XFM images were congruent with the DIC images. Both XFM elemental maps and DIC Sample Preparation. Silicon nitride windows (Silson, North Hampton, images are included within the ESI for this paper. UK) (Si3N4) were coated with matrigel (Corning, NY, USA)[50] to simulate the extracellular matrix (ECM) biological environment. To achieve this, a solution of Matrigel in FBS supplied DMEM (1 mg mL-1) was prepared Acknowledgements and heated to 310 K. Si3N4 windows (3 x 3 mm2) were then dipped in the solution and then transferred to a six-well cell culture plate and incubated The authors are grateful for support from the Australian with cells (24 hr). A suspension of cultured A549 cells was applied to each window (1 mL, 105 cells mL-1) and left overnight to enable the cells Research Council (DP140100176) and the salary of Dr. Aviva to adhere.[50] Solutions of curcumin, 5, and 6 in A-DMEM were prepared Levina (DP130103566, DP160104172) and for the ion cyclotron (30 μM), as well as a control treatment of DMSO in A-DMEM (1% v/v). mass spectrometer (LE0668439). We also acknowledge support These solutions were applied to the cell monolayers and incubated for from the Australian Synchrotron for use of the XAS beamline another 24 hr. After treatment was completed, the windows were each and ISAP funding for travel to the Argonne National Laboratory. rinsed by dipping five times in DMEM followed by PBS to remove the Rh- This research used resources of the Advanced Photon Source, containing medium. Windows were then rapidly dipped in cold (-20°C) MeOH five times to fix the cells.[51] The windows were then stored in a a U.S. Department of Energy (DOE) Office of Science User desiccator until transport to the beamline. Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357. Data Collection and Analysis. The specimen windows were placed We are grateful for assistance from Dr. Nicholas Proschogo in onto a kinematic specimen holder suitable for both optical and X-ray mass spectroscopy and Dr Ian Luck in NMR spectroscopy. The fluorescence microscopy. The specimen was first examined under a light authors acknowledge the facilities, and the scientific and microscope (Leica DM6000 B) and selected areas of interest were technical assistance, of the Australian Microscopy & located relative to reference points of the silicon nitride membrane using Microanalysis Research Facility at University of Sydney, high resolution motorized x,y stage of the microscope. Coordinates were recorded and were used to precisely locate the target area once the particularly Minh Huynh. The authors acknowledge the support specimen was transferred to the X-ray microprobe. received from the Bosch Institute's Molecular Biology Facility, and the expert help of Facility staff, especially Sheng Hua. Scanning XFN was performed at the 2-ID-D beamline of the Advanced Photon Source at Argonne National Laboratory.[52] The source was a 3.3- Keywords: rhodium pentamethylcyclopentadienato complexes cm-period undulator, and X-rays were monochromatized by a double- •curcumin • cytotoxicity • X-ray absorption spectroscopy • X-ray crystal Si(111) monochromator (Kohzu). The X-ray energy of 23.7 keV was chosen to excite the Kα emission lines for elements up to Rh. In fluorescence microscopy order to increase the focusing efficiency at such high energy, five Fresnel zone plates were aligned in intermediate field,[53] which focused the X-ray [1] F. Trudu, F. Amato, P. Vanhara, T. Pivetta, E. M. Pena-Mendez, J. beam to a spot size of ~ 0.4 μm on the sample. Alternatively, the X-ray Havel, J. Appl. Biomed. 2015, 13, 79-103. energy of 10.0 keV was chosen to excite the Kα emission lines for [2] J. Graham, M. Muhsin, P. Kirkpatrick, Nat. Rev. Drug Discov. 2004, 3, elements up to Zn. A single Fresnel zone plate with 160 μm diameter and 11-12. 70 nm fines zones was used to focus the X-ray beam. The focal length is [3] F. A. Blommaert, H. C. M. van Dijk-Knijnenburg, F. J. Dijt, L. den 91.3 mm at this energy. The specimen was placed in a He environment Engelse, R. A. Baan, F. Berends, A. M. J. Fichtinger-Schepman, and mounted on a x,y translation stage at 75º to the incident beam. An Biochemistry 1995, 34, 8474-8480. energy-dispersive silicon drift detector (Vortex-EM) was used to collect X- [4] (a) M. Pongratz, P. Schluga, M. A. Jakupec, V. B. Arion, C. G. ray fluorescence spectra from the sample while it was being scanned Hartinger, G. N. Allmaier, B. K. Keppler, J. Analyt. Atom. Spectrom. across the focus spot. 2004, 19, 46-?; (b) C. G. Hartinger, S. Zorbas-Seifried, M. A. Jakupec, B. Kynast, H. Zorbas, B. K. Keppler, J. Inorg. Biochem. 2006, 100, 891- Maps at 10.4 keV were recorded using the fly-scan mode, with a 0.5 μm 904; (c) R. T. L.S. Flocke, M.A. Jakupec, B.K. Keppler, Invest. New step-size and a 300 ms dwell time per pixel. Maps recorded at 23.7 keV Drugs 2016, 34, 261-268. used a 1 μm step-size, with dwell times ranging from 10–20 s and were [5] S. Leijen, S. A. Burgers, P. Baas, D. Pluim, M. Tibben, E. van performed in step mode. Werkhoven, E. Alessio, G. Sava, J. H. Beijnen, J. H. M. Schellens, Invest. New Drugs 2015, 33, 201-214. MAPS[54] fitted spectra at each pixel to remove overlaps between [6] W. H. Ang, E. Daldini, C. Scolaro, R. Scopelliti, L. Juillerat-Jeannerat, P. adjacent emission lines in order to achieve more accurate quantification. J. Dyson, Inorg. Chem. 2006, 45, 9006-9013. Conversion of elemental fluorescence intensities to areal densities in μg [7] R. E. Morris, R. E. Aird, P. del Socorro Murdoch, H. Chen, J. cm-2 was performed by comparing X-ray fluorescence intensities with Cummings, N. D. Hughes, S. Parsons, A. Parkin, G. Boyd, D. I. Jodrell, those from thin film standards NBS-1832 and NBS-1833 (NIST). P. J. Sadler, J. Med. Chem. 2001, 44, 3616-3621. [8] R. E. Aird, J. Cummings, A. A. Ritchie, M. Muir, R. E. Morris, H. Chen, P. J. Sadler, D. I. Jodrell, Br. J. Cancer 2002, 86, 1652-1657. Regions of interest (ROIs) were determined using the MAPS software [9] E. S. Antonarakis, A. Emadi, Cancer Chemother. Pharmacol. 2010, 66, and validated via differential interference contrast (DIC) microscopy in 1-9. order to separate the images into cell and background regions. Both [10] A. Dorcier, P. J. Dyson, C. Gossens, U. Rothlisberger, R. Scopelliti, I. background and cellular regions had their light-element concentration Tavernelli, Organometallics 2005, 24, 2114-2123. quantified as an average for the ROI. The results are presented as the [11] A. Casini, C. Gabbiani, F. Sorrentino, M. P. Rigobello, A. Bindoli, T. J. average ratio of cell to background elemental concentration for all cells in Geldbach, A. Marrone, N. Re, C. G. Hartinger, P. J. Dyson, L. Messori, identical treatments. Hotspots from dust were removed by removing J. Med. Chem. 2008, 51, 6773-6781. This article is protected by copyright. All rights reserved. European Journal of Inorganic Chemistry 10.1002/ejic.201601331 FULL PAPER [12] Y. Geldmacher, M. Oleszak, W. S. Sheldrick, Inorg. Chim. Acta 2012, [30] J. T. Price, E. W. Thompson, Expert Opin. Therap.Targets 2002, 6, 393, 84-102. 217-233. [13] N. Katsaros, A. Anagnostopoulou, Critic. Rev. Oncol. Hematol. 2002, [31] S. Kraft, J. Stümpel, P. Becker, U. Kuetgens Rev. Sci. Inst. 1996, 67, 42, 297-308. 681 [14] N. Carcmichael, A. Taylor, Biochem. Inst. Studies 5, Cancer Studies 2. [32] (a) A. Rimessi, C. Giorgi, P. Pinton, R. Rizzuto, Biochim. Biophys. Acta 1953, 36, 36 - 79. Bioenerget. 2008, 1777, 808-816; (b) N. Prevarskaya, H. Ouadid- [15] M. J. Cleare, P.C.Hydes, Met. Ions Biol. Syst. 1980, 11, 1-62. Ahidouch, R. Skryma, Y. Shuba, Philos. Trans. Roy. Soc. Lond. B: Biol. [16] A. Dorcier, W. H. Ang, S. Bolaño, L. Gonsalvi, L. Juillerat-Jeannerat, G. Sci. 2014, 369, Art No 20130097; (c) J. Parkash, K. Asotra, Life Sci. Laurenczy, M. Peruzzini, A. D. Phillips, F. Zanobini, P. J. Dyson, 2010, 87, 587-595. Organometallics 2006, 25, 4090-4096. [33] R. Rizzuto, P. Pinton, D. Ferrari, M. Chami, G. Szabadkai, P. J. [17] L. Dadci, H. Elias, U. Frey, A. Hornig, U. Koelle, A. E. Merbach, H. Magalhaes, F. Di Virgilio, T. Pozzan, Oncogene 2003, 22, 8619-8627. Paulus, J. S. Schneider, Inorg. Chem. 1995, 34, 306-315. [34] E. C. Kohn, L. A. Liotta, Cancer Res. 1995, 55, 1856-1862. [18] (a) P. A. Lay Coord. Chem. Rev. 1991, 110, 213-231. (b) Lay, P. A. [35] (a) M. H. Khosropanah, A. Dinarvand, A. Nezhadhosseini, A. Haghighi, Comments Inorg. Chem. 1991, 11, 235-284. S. Hashemi, F. Nirouzad, S. Khatamsaz, M. Entezari, M. Hashemi, H. [19] D. J. Giard, S. A. Aaronson, G. J. Todaro, P. Arnstein, J. H. Kersey, H. Dehghani, Iran. J. Pharm. Res. 2016, 15, 231-239; (b) X. D. Xu, D. Dosik, W. P. Parks, J. Nat. Cancer Inst. 1973, 51, 1417-1423. Chen, B. Ye, F. M. Zhong, G. Chen, Int. J. Mol. Med. 2015, 35, 1610- [20] (a) N. G. Vallianou, A. Evangelopoulos, N. Schizas, C. Kazazis, 1616. Anticancer Res. 2015, 35, 645-651; (b) D. K. Agrawal, P. K. Mishra, [36] (a) R. B. Franklin, L. C. Costello, J. Cell. Biochem. 2009, 106, 750-757; Med. Res. Rev. 2010, 30, 818-860; (c) Z.-Y. Du, Y.-F. Jiang, Z.-K. Tang, (b) A. Q. Truong-Tran, J. Carter, R. E. Ruffin, P. D. Zalewski, BioMetals R.-Q. Mo, G.-H. Xue, Y.-J. Lu, X. Zheng, C.-Z. Dong, K. Zhang, Biosci. 2001, 14, 315-330. Biotechnol. Biochem. 2011, 75, 2351-2358; (d) V. P. Menon, A. R. [37] N. Fujita, A. Nagahashi, K. Nagashima, S. Rokudai, T. Tsuruo, Sudheer, Molecular Targets and Therapeutic Uses of Curcumin in Oncogene 1998, 17, 1295-1304. Health and Disease 2007, 595, 105-125.; (e) J. S. Jurenka, Altern. Med. [38] P. D. Zalewski, I. J. Forbes, R. F. Seamark, R. Borlinghaus, W. H. Betts, Rev. 2009, 14, 141-153; (f) C. V. Rao, A. Rivenson, B. Simi, B. S. S. F. Lincoln, A. D. Ward, Chem. Biol. 1994, 1, 153-161. Reddy, Cancer Res. 1995, 55, 259-266; (f) A. K. Renfrew, N. S. Bryce, [39] D. Chatterjee, A. Mitra, A. Levina, P. A. Lay, Chem. Commun. 2008, T. W. Hambley, Chem. Soc. 2013, 4, 3731-3739; (g) A. K. Renfrew, 2864-2866. Metallomics 2015, 6, 1324-1335. [40] J. T. Velasquez, M. E. Watts, M. Todorovic, L. Nazareth, E. Pastrana, J. [21] (a) A. Levina, J. B. Aitken, Y. Y. Gwee, Z. J. Lim, M. M. Liu, A. M. Diaz-Nido, F. Lim, J. A. K. Ekberg, R. J. Quinn, J. A. St John, Plos One Singharay, P. F. Wong, P. A. Lay, Chem. Eur. J. 2013, 19, 3609-3619; 2014, 9, e111787. (b) J. B. Aitken, A. Levina, P. A. Lay, Curr. Top Med. Chem., 2011, 11, [41] (a) G. J. Grant, J. P. Lee, M. L. Helm, D. G. Van Derveer, W. T. 553-571; (c) A. Levina, H. H. Harris, P. A. Lay, J. Am. Chem. Soc. 2007, Pennington, J. L. Harris, L. F. Mehne, D. W. Klinger, J. Organomet. 129, 1065-1075; 2007, 129, 9832; (d) A. Levina, A. I. McLeod, J. Chem. 2005, 690, 629-639; (b) W. Rigby, H. B. Lee, P. M. Bailey, J. A. Seuring, P. A. Lay, J. Inorg. Biochem. 2007, 101, 1586-1593; (e) A. McCleverty, P. M. Maitlis, J. Chem. Soc., Dalton Trans. 1979, 387-394; Nguyen, I. Mulyani, A. Levina, P. A. Lay, Inorg. Chem. 2008, 47, 4299- (c) M. A. Scharwitz, I. Ott, Y. Geldmacher, R. Gust, W. S. Sheldrick, J. 4309; (f) A. Levina, A. I. McLeod, A. Pulte, J. B. Aitken, P. A. Lay, Inorg. Organomet. Chem. 2008, 693, 2299-2309; (d) G. Garcia, G. Sanchez, I. Chem. 2015, 54, 6707-6718; (g) A. Levina, A. I. McLeod, S. J. Romero, I. Solano, M. D. Santana, G. Lopez, J. Organomet. Chem. Gasparini, A. Nguyen, W. G. M. De Silva, J. B. Aitken, H. H. Harris, C. 1991, 408, 241-246; (e) B. L. Booth, R. N. Haszeldine, M. Hill, J. Chem. Glover, B. Johannessen, P. A. Lay, Inorg. Chem. 2015, 54, 7753-7766 Soc. A 1969, 1299-1303; (f) J. G. Hernandez, C. Bolm, Chem. (h) L. E. Wu, A. Levina, H. H. Harris, Z. Cai, B. Lai; S. Vogt, D. E. Commun. 2015, 51, 12582-12584. James, P. A. Lay, P. A. Angew. Chem., Int. Ed., 2016, 55, 1742–1745; [42] H. Senoz, Hacettepe J.Biol. Chem. 2012, 40, 293-301. (i) G. K. Gransbury, P. Kappen, C. J. Glover, J. N. Hughes, A. Levina, P. [43] F. Bretz, J. C. Pinheiro, M. Branson, Biometrics 2005, 61, 738-748. A. Lay, I. F. Musgrave, H. H. Harris, Metallomics, 2016, 8, 762-773; (j) [44] Microsoft Corporation, ver. 16.0.6701.1041 64-bit ed. 2016. H. A. O’Riley, A. Levina, J. B. Aitken, P. A. Lay, Inorg. Chim. Acta, 2017, [45] Essen BioScience, 2016 454, 128-138; (k) A. Levina, D. C. Crans, P. A. Lay, Coord. Chem. Rev., [46] T. Sasaki, K. Hiroki, Y. Yamashita, BioMed. Res. Int. 2013, 2013, 1-8. 2017, in press. [47] C. Glover, J. McKinlay, M. Clift, B. Barg, J. Boldeman, M. Ridgway, G. [22] (a) A. Levina, A. Mitra, P. A. Lay, Metallomics 2009, 1, 458-470; (b) M. Foran, R. Garrett, P. Lay, A. Broadbent, AIP Conf. Proc. 2007, 882, Liu, Z. Lim, Y. Y. Gwee, A. Levina, P. A. Lay, Angew. Chem., Int. Ed, 884-886. 2010, 49, 1661-1664. [48] G. N. George, I. J. Pickering, M. J. Pushie, K. Nienaber, M. J. Hackett, I. [23] R. Pettinari, F. Marchetti, C. Pettinari, F. Condello, A. Petrini, R. Ascone, B. Hedman, K.O. Hodgson, J. B. Aitken, A. Levina, C. Glover, Scopelliti, T. Riedel, P. J. Dyson, Dalton Trans. 2015, 44, 20523-20531. P. A. Lay, J. Synchrotron Radiat. 2012, 19, 875-886. [24] L. Bonfili, R. Pettinari, M. Cuccioloni, V. Cecarini, M. Mozzicafreddo, M. [49] B. R. M. Newville, J. Synchrotron Rad. 2005, 12, 537-541. Angeletti, G. Lupidi, F. Marchetti, C. Pettinari, A. M. Eleuteri, [50] H. K. Kleinman, G. R. Martin, Semin. Cancer Biol. 2005, 15, 378-386. ChemMedChem 2012, 7, 2010-2020. [51] E. A. Carter, B. S.; Rayner, A. I. McLeod, L. E. Wu, C. P. Marshall, A. [25] R. Pettinari, F. Marchetti, F. Condello, C. Pettinari, G. Lupidi, R. Levina, J. B. Aitken, P. K. Witting, B. Lai, Z. Cai, S. Vogt, Y.-C. Lee, C.-I. Scopelliti, S. Mukhopadhyay, T. Riedel, P. J. Dyson, Organometallics Chen, M. J. Tobin, H. H. Harris, P. A. Lay, Mol Biosyst. 2010, 6, 1316- 2014, 33, 3709-3715. 1322 [26] C. Bartel, A. E. Egger, M. A. Jakupec, P. Heffeter, M. Galanski, W. [52] Z. Cai, B. Lai, W. Yun, P. Ilinski, D. Legnini, J. Maser, W. Rodrigues, Berger, B. K. Keppler, J. Biol. Inorg. Chem. 2011, 16, 1205-1215. AIP Conf. Proc. 2000, 507, 472-477. [27] T. Riss, in “Is Your MTT Assay Really the Best Choice?” URL [53] S.-C. Gleber, M. Wojcik, J. Liu, C. Roehrig, M. Cummings, J. Vila- (http://au.promega.com/resources/pubhub/is-your-mtt-assay-really- Comamala, K. Li, B. Lai, D. Shu, S. Vogt, Opt. Exp. 2014, 22, 28142- thebest-choice/), 2014 accessed 20/09/15 28153. [28] D. R. McIlwain, T. Berger, T. W. Mak, Cold Spring Harbor Persp. Biol. [54] S. Vogt, J. Phys. IV 2003, 104, 635-638. 2013, 5, 28-56. [29] P. Zhang, W. Y. Gao, S. Turner, B. S. Ducatman, Mol. Cancer 2003, 2, 1476-4598. This article is protected by copyright. All rights reserved. European Journal of Inorganic Chemistry 10.1002/ejic.201601331 FULL PAPER FULL PAPER Anti-Cancer Drug Jack Markham, Jun Liang, Aviva Levina, [RhIII(*Cp)(curcuminato)Cl] C u r c u m i n Apoptosis Rachel Mak, Bernt Johannessen, Peter + [Rh(*Cp)L ]n+ Kappen, Chris J. Glover, Barry Lai, 3 Stefan Vogt, and Peter A. Lay* Page No. – Page No. The [RhIII(*Cp)(curcuminato)Cl] (*Cp = pentamethylcyclopentadienato) is readily taken up by the A549 lung cancer cell line and enables the slow intracellular Rhodium Pentamethylcyclopenta- release of curcumin that causes cytotoxicity via apoptosis. This makes the complex dienato Complexes for Delivery of the an effective drug delivery system for the poorly bioavailable curcumin drug. Curcumin Anti-Cancer Drug This article is protected by copyright. All rights reserved.