Introduction

Metallodrugs with cytotoxic properties have a long history as therapeutic agents in oncology. Ruthenium containing complexes have become one of the best developed representatives as new metallodrugs for the treatment of tumors. They show few side effects and resistance1,2 is unlikely to form against several Ru-based complexes due to the pharmacokinetic and chemical behavior of Ru drugs3,4. Two of the most recent Ru-containing investigational drug candidates subject to investigation are KP1019 (indazolium trans-[tetrachloridobis(1H-indazole)ruthenate(III)])5 and its sodium analogue NKP-1339 (7, sodium trans-[tetrachloridobis(1H-indazole)ruthenate(III)], Supplementary Figure 1)6,7. Due to the identical Ru coordination sphere, [tetrachloridobis(1H-indazole)ruthenate(III)], in KP1019 and 7, it is expected that the main binding partners in vivo and their mode of action is highly similar3,8. Both compounds show remarkable anti-tumor activity in colorectal carcinomas in vivo and a variety of primary explanted human tumors in vitro1,3,8,9. Due to its high water solubility, 7 progressed also to a clinical phase I-IIa study10,11,12. This study is notable for having demonstrated activity of NKP-1339 in cancers unresponsive to prior treatments, such as neuroendocrine tumors of the stomach and small intestine. The current knowledge of the medicinal potential and the mode of action of 7 have been reviewed7.

X-ray absorption near-edge spectroscopy (XANES) provides detailed information about the speciation of metal centers such as oxidation state and coordination environment. X-ray absorption spectroscopy (XAS) has been proven to be successful in the study of metal complexes for cancer therapy in vivo and in vitro13,14,15,16,17,18,19,20. In contrast to the metal K-edge transitions (1 s → np), the metal L-edges transitions (2p → 4d) are very sensitive to changes in the oxidation state and the electronic environment, such as ligand dissociation and association in biological samples. The soft X-rays needed for the L2,3-edges usually have 3–5 times better energy resolution with more intense spectral features21,22, making L-edges superior to K-edges in terms of oxidation state assignment.

The electronic structures of the tumor-inhibiting compounds KP1019 and NAMI-A (imidazolium trans-[tetrachlorido(dimethylsulfoxide)(imidazole)ruthenate(III)])23 have been investigated in their solid state applying Cl K-edge and Ru L3-edge XANES spectroscopy24. Ru K- and L3-edge spectra of NAMI-A in the presence of bovine serum albumin (BSA) have been recorded to investigate the oxidation state and the structure of the formed adducts. The coordination environments of the BSA adduct comprises about 60% N and 40% O ligands coordinated to RuIII20,25. KP1019 was investigated at the Ru K-edge, on the one hand, in citrate saline (CS) buffer (pH 3.5) in the presence of the reducing agent glutathione and, on the other hand, in carbonate buffer (pH 7.4) in a 1:1 mixture with apo-transferrin. XANES analysis revealed replacement of Cl ligands to a different extent in all solutions and suggested a potential interaction with apo-transferrin. KP1019 and 7 were also investigated in tumor and liver tissue with dosages varying between 7.5–40 mg/kg. Ru K-edge spectroscopy did not allow to distinguish between RuIIICl3N2(O/N) and RuIIClN2(O/N)3 as assigned coordination modes in the tissue samples17. A study on human hepatoma cells treated with KP1019 under cell culture conditions found, besides O/N binding from amine/imine and carboxylato groups of proteins, also indications of S donor atoms binding to RuII/III that may originate from Cys residues of cytoplasmic enzymes such as cysteine proteases and thioredoxin reductase19.

To assign the Ru oxidation state and to understand the electronic distribution of the highly promising tumor inhibiting compound 7 Ru L-edge and Cl K-edge spectroscopy has been performed for the first time on tissue samples. Therefore, one mouse with severe combined immunodeficiency genetic disorder (SCID) bearing the human SW480 adenocarcinoma xenograft was treated with 7 for 24 h. Three different tissue types (tumor = T, liver = L and kidney = K) were investigated (T1, K1, L1) using ligand Cl K-edge and Ru L2,3-edge XANES spectroscopy. The XANES investigations of the main detoxification organs in addition to the tumor tissue will address important aspects of the NKP-1339 (RuIIICl4N2) fate in vivo.

Results and Discussion

XANES spectra of the model compounds

The XANES region at the Cl K-edge and the Ru L2,3-edges were investigated for the following ruthenium (chloride) model compounds (Synthetical protocols cited and structures shown in Supplementary Figure 1), which represent possible Ru coordination and oxidation states in vivo: ruthenium(III) acetylacetonate (1, with first coordination shell RuIIIO6, Sigma Aldrich, CAS 14284-93-6, 97%)26,27, hexammineruthenium(III) trichloride (2, RuIIIN6, Sigma Aldrich, CAS 14282-91-8, 99%)28, (nBu4N)2[RuCl3(ox)(NO)] (3, GABU527, RuIIICl3NO2)29, mer,trans-aquatrichloridobis(indazole)ruthenium(III) (4, KASC003, RuIIICl3N2O)30, trans,trans-dichloridotetrakis(indazole)ruthenium(III) chloride (5, GABU129, RuIIICl2N4)31, mer-trichloridotris(indazole)ruthenium(III) (6, GUPL328, RuIIICl3N3)32, sodium trans-[tetrachloridobis(1H-indazole)ruthenate(III)] (7, NKP-1339, RuIIICl4N2)6, mer,trans-trichlorido(dimethylsulfide)bis(indazole)ruthenium(III) (8, FLAN005, RuIIICl3N2S)33, tris(bipyridine)ruthenium(II) chloride (9, RuIIN6, Sigma Aldrich, CAS 50525-27-4, 99.95%)34, hexammineruthenium(II) dichloride (10, RuIIN6, Sigma Aldrich, CAS 15305-72-3, 99.9%)28, mer-trichloridotris(ethylphenylsulfide)ruthenium(III) (11, FLAN006, RuIIICl3S3)35, trans,trans-dichloridotetrakis(indazole)ruthenium(II) (12, GABU128, RuIICl2N4)31 and trans,trans,trans-dichloridobis(dimethylsulfide)bis(indazole)ruthenium(II) (13, FLAN004, RuIICl2N2S2)33. The Cl K-edge and the Ru L3-edge XANES spectra of model compounds 1 to 13 are shown together in Supplementary Figure 2 and the Ru L2-edge spectra in Supplementary Figure 3. The edge energies for the Cl K-edge and Ru L2,3-edges based on the first maximum in the first derivative of the corresponding normalized XANES spectrum are summarized in Table 1.

Table 1 Edge energies and calculated coordination charges of model compounds 1 to 13 (all octahedral), kidney, liver and tumor tissue.

The Cl K-edge energy positions of 1 to 13 (Table 1) span a range of 1.4 eV (2824.2 to 2825.6 eV) with 13 (2 Cl covalently bound to RuII) occupying the highest energy (2825.6 eV) and 5 (2 Cl covalently bound to RuIII) the lowest energy (2824.8 eV) for the compounds containing bound Cl ligands. 10 (2 Cl counter anions to a RuIIN6 coordination sphere) exhibits the lowest edge energy (2824.2 eV). Model compounds with Cl counter anions (2, 9 and 10) show a higher white line intensity at the Cl K-edge and a steeper rise of the edge (Supplementary Figure 2) after normalization compared to model compounds with covalently bound Cl atoms (3 to 8 and 11 to 13). When Cl is covalently bound to an open shell metal, a characteristic pre-edge feature is present in the Cl K-edge spectrum (dashed and dotted spectra in Supplementary Figure 4). This can be attributed to partial mixing of ligand p-orbitals with metal d-orbitals in an otherwise forbidden 1 s → 4d transition36,37. Therefore, the Cl K-edge spectra in 3 to 8 and 11 to 13 exhibit pre-edge features due to the covalent character of the Ru-Cl bonds, whereas they are missing in 2, 9 and 10, where Cl is the counter anion (solid lines in Supplementary Figure 4). Closer inspection of the Cl K-edge pre-edge features reveals distinct differences between RuII and RuIII model compounds. RuII compounds (12 and 13) exhibit one pre-edge transition (dotted lines in Supplementary Figure 4), whereas RuIII complexes (3 to 8, 11) show a shoulder or even small separate peak before the pre-edge transition (dashed lines in Supplementary Figure 4). The shoulder has been attributed to an electronic transition to a singly unoccupied molecular orbital24. Deconvolutions that resulted from fitting of the Cl K-edge pre-edge region of the octahedral Ru–Cl compounds 3 to 8 and 11 to 13 are given in Supplementary Table 1. The table lists the crystallographic Ru-Cl bond distances as well as the pre-edge energy position, amplitude, FWHM (full width at half maximum), the area for each pre-edge transition and the total area of the pre-edge transitions for each compound. The total pre-edge area increases within the same Ru oxidation state as the number of Cl donor ligands grows, since the total charge donated to the Ru ion is larger. The total area for model compounds 5 (2 Cl covalently bound to RuIII), 6 (3 Cl covalently bound to RuIII) and 7 (4 Cl covalently bound to RuIII) is 1.868 ± 0.019, 1.965 ± 0.014 and 2.176 ± 0.009, respectively. As the Ru oxidation state decreases from RuIII to RuII the Ru-Cl bond distance becomes longer resulting in decreased orbital overlap between the ligand and metal and thus a lower total area of the pre-edge. The total area decreases from 1.868 ± 0.019 in 5 (2 Cl covalently bound at a average distance of 2.331 Å to RuIII) to 1.045 ± 0.057 in 12 (2Cl covalently bound at an average distance of 2.418 Å to RuII).

The Ru L3- and L2-edge energies of 1 to 13 span a range of 1.5 eV and 1.4 eV, respectively, as compared to 6.1 eV at the Ru K-edge for an almost identical set of model compounds17. As reported for the Ru K-edges the edge energies at the Ru L2,3-edges increase as the Ru oxidation state changes from RuII to RuIII and within the same Ru oxidation state the defined edges are shifted to higher energies with increasing electronegativity of the first shell atoms (see Table 1). 1 (RuIIIO6) exhibits the highest edge energy (L3: 2842.0 eV L2: 2970.1 eV) and 14 (RuIIS2Cl2N2) the lowest edge energy (L3: 2840.5 eV L2: 2968.7 eV). The RuIII model compounds (1 to 8, 11) show a splitting of the Ru L3-edge (Supplementary Figure 2) which is missing in RuII model compounds (9, 10, 12 and 13), in line with previous observations24,25. The splitting arises from the transformation of the 4d orbitals into t2g and eg orbitals, in the presence of an octahedral RuIII which has a 4d5 electron configuration38. In this case the second maximum in the first derivative was used to determine the edge energy position which better correlates with the observed edge energies of the RuII model compounds.

XANES spectra of the tissue samples

Tumor, liver and kidney tissue sections were collected from a mouse 24 hours after 7 (40 mg/kg) was administrated intravenously. Data collection of the tissue samples at the Cl K-edge and at the Ru L2,3-edges was restricted due to the low Ru concentration and highly abundant Cl in biological tissue. The high Cl concentrations showed Cl K-edge signals that were dominated by Cl contributions, except for the liver sample L1 (vide infra), which contained the lowest Cl to Ru ratio and showed a Cl K-edge pre-edge feature. XANES signals were collected at the Ru L2-edge for the tissue samples L1, T1 and K1. Figure 1 shows the Ru L2-edge and the corresponding first derivative of the tissue samples. The tissue samples show a less pronounced splitting of the edge compared to 7. Unlike the Ru L3-edge, the splitting observed at the Ru L2-edge does not necessarily give indications of the Ru oxidation state and can disappear due to e.g. spin-orbit effects39. The edge position of all tissue samples increased by 0.3 to 0.4 eV (L1: 2969.8 eV; T1 and K1: 2969.9 eV) compared to 7 (2969.5 eV) suggesting RuIII as the oxidation state and a highly similar coordination sphere around RuIII in the tissue samples.

Figure 1: Normalized XANES spectra of the Ru L2-edge of tumor, liver and kidney tissue samples.
figure 1

The first derivative is shown in the inset. Spectra are plotted with an arbitrary vertical shift. 7 measured in boron nitride is shown as a reference. A vertical line is inserted that goes through the maximum of 7.

Coordination charge and 3D XANES plots

Discrete spectral features in the XANES region of the well characterized model compounds 1 to 13 can be plotted together and be used as a basis to assign the oxidation state and coordination mode of Ru centers in tumor, kidney and liver tissue. In Table 1 the coordination charge and the corresponding Ru L2-edge energy are listed for 1 to 13, which was plotted against each other in Fig. 2a17,29. A straight line was regressed with a coefficient of determination R2 = 0.95, demonstrating a linear correlation between the coordination charge and the edge energy positions. The edge energy of 7 in boron nitride (BN) was set as an arbitrary origin. RuIII compounds containing S and RuII compounds are found on the lower energy side, whereas RuIII compounds containing N/O are on the higher energy side. The positioning of the tissue samples show a 25% increase in edge energy (0.3–0.4 eV) when comparing to the entire model compound energy range (1.4 eV) at the Ru L2-edge (Table 1). At the Ru K-edge the same tissue samples showed an 18% increase in edge energy among a very similar set of model compounds17. In terms of Ru L2-edge energy position the tissue samples place themselves between 2 (RuIIIN6) and 3 (RuIIIO2Cl3N).

Figure 2: In vivo Ru oxidation state.
figure 2

(a) Calculated coordination charge ηAR according to the Allred–Rochow scale in comparison to the observed edge energies of the XANES spectra in model compounds 1 to 13 (black circles) and the tissue samples (purple and orange lines). The edge energy position was chosen as the first maximum in the first derivative for all model compounds and tissue samples. The edge energy of 7 in boron nitride (BN) is shown as a red line and was set as an arbitrary origin. (b) Three dimensional XANES plots based on Ru L2-edge maximum height energy (MHE), half height energy (HHE) and energy at the first maximum in the first derivative (1M1D). Model compounds 1 to 13 are shown as black circles and the tissue samples K1, L1 and T1 are shown as colored circles. The plot shows clustering into three distinct regions: RuIII compounds with O ligands (A), RuIII compounds with N/Cl ligands (B) and RuII compounds with N ligands (C). Region D contain compounds that did not fall into a distinctive cluster. (c) The Cl K-edge pre-edge features of 7, 10, 12 and L1. Sample L1 show a pre-edge feature with an initial shoulder before an intense pre-edge transition. The L1 spectrum has been smoothed for improved visual appearance. The original spectrum is shown in Supplementary Figure 5. (d) Proposed elemental composition and oxidation state for the majority of the Ru centers in the tumor, liver and kidney tissue based on the conforming graphs in (a) and (b) and the spectrum in (c). K1 = kidney 1, L1 = liver 1, T1 = tumor 1.

Recently the concept of three dimensional XANES plots40 was successfully introduced. Key XANES parameters present in model compounds and biological samples, such as edge energies of discrete spectral features and white line intensities are plotted against each other and are used for prediction of oxidation states and coordination spheres in unknown samples. The three dimensional plot in Fig. 2b is based on energies taken from the first maximum in the first derivative, the maximum height and the half height given in Table 1 for the Ru L2-edge in order to better separate the Ru oxidation states. 3D-plots with the FWHM and the white line height were also considered in separate graphs but did not generate clusters (the plot including the FWHM is shown in Supplementary Figure 6). The three dimensional plot in Fig. 2b illustrates separation of the model compounds into three different clusters (A-C in Fig. 2b): RuIII compounds with O ligands (A), RuIII compounds with N/Cl ligands (B) and RuII compounds containing six N ligands (C). Compounds 11 to 13 (RuIIIS3Cl3, RuIIN4Cl2 and RuIIS2Cl2N2, respectively) were unable to form a distinctive cluster (region D). The tissue samples fall into the upper end of the B cluster (RuIII compounds with N/Cl ligands) extending towards the A cluster (RuIII compounds with O ligands).

The coordination charge correlation and the 3D XANES plot both suggest the presence of RuIII in the tissue samples, which is in good agreement with the observed Cl K-edge pre-edge features of sample L1. Based on earlier findings24 and the herein investigated model compounds, the shoulder (A in Fig. 2c) before the pre-edge feature (B in Fig. 2c) indicates presence of RuIII. Conclusions of the Ru oxidation state based on the interpretation of the Cl K-edge pre edge features must be drawn with care due to the low quality tissue data (Supplementary Figure 5). However, the existence of a Cl K-edge pre-edge peak indicates that at least one Cl ligand remains covalently bound to Ru (under the assumption that no other Cl-metal open-shell compound is present in L1).

Calculations taking into account the shifts in energy that different ligands and oxidation states induce in the coordination charge correlation confirm that only RuIII is possible for the observed tissue edge energies. The replacement of one Cl ligand by one N ligand in the set of model compounds gives an edge energy shift of approximately 0.1 eV (Table 1). Model compounds 6 (RuIIICl3N3) and 5 (RuIIICl2N4) increase by 0.1 and 0.2 eV, respectively compared to 7 (RuIIICl4N2). The replacement of one N ligand with one O ligand gives no detectable change in the edge energy since both 3 (RuIIIO2Cl3N) and 4 (RuIIIOCl3N2) exhibit Ru L2-edge energies at 2969.8 eV. A change in the oxidation from RuII to RuIII gives an average shift of 0.7 eV when comparing the edge energy differences between compounds 2 (RuIIIN6, 2969.9 eV) and 9 (RuIIN6, 2969.4 eV) and 5 (RuIIIN4Cl2, 2969.7 eV) and 12 (RuIIN4Cl2, 2968.8 eV). Based on these findings, a RuII oxidation state with a high number of O/N ligands such as RuIIO5Cl cannot reach the lowest edge energy observed of at least 2969.8 eV for liver tissue and must therefore be ruled out. However this study cannot rule out S coordinated to Ru(III) in the tissue samples.

The most likely cause for the observed 0.3–0.4 eV shift in all tissue samples compared to solid 7 can be attributed to the replacement of Cl with N and/or O atoms from e.g. amino acids and aqua ligands, respectively (Fig. 2d). The chlorido donors bound to Ru in the metabolites do not necessarily need to come from the parent drug as they can be the result of ligand-exchange reactions with the medium. Ligand exchange has recently been supported by X-ray structure studies41 of the highly similar KP1019 bound to human serum albumin, one of the most likely transport proteins42,43. Two octahedral Ru sites were found with first shell N donor atoms arising from the amino acids (Ru1: His146; Ru2: Lys199 and His242)41.

Conclusions

The complex trans-[tetrachloridobis(1H-indazole)ruthenate-(III)] (NKP-1339) was used to treat a SW480-bearing mouse of which tumor, liver and kidney tissues were investigated by XANES (Cl K-edge and Ru L2,3-edges) and to resolve the Ru oxidation state in target tissue and in the main detoxification organs. The combined XANES analysis from all model compounds and tissue samples suggest a biotransformation of NKP-1339 in all tissues from RuIIIN4Cl2 to RuIIIClx(N/O)6−x (X = 1 or 2). The presence of RuII species at his time point (24 h after treatment) can be ruled out.

Methods

Application scheme mouse experiments

Citrate saline buffer was prepared by adding 0.42 g of anhydrous citric acid (Sigma Aldrich, 77-92-9, ≥99.5%) to a 400 ml sterile normal saline (0.9% w/v NaCl, Sigma Aldrich, CAS 7647-14-5, ≥99.5%) solution. The solution was stirred until the citric acid dissolved completely (solution A). A second solution was prepared by adding 0.59 g of sodium citrate dihydrate (Sigma Aldrich, CAS 6132-04-3, ≥99%) to a 400 ml sterile normal saline (0.9% w/v NaCl) solution and stirring until everything dissolves. 156 ml of the first solution was combined with 44 ml of the second solution. The resulting CS buffer was adjusted to pH 3.5. Sodium trans-[tetrachloridobis(1H-indazole)ruthenate(III)] (NKP-1339) was prepared as previously reported6. CB-17 SCID mice bearing the human xenograft adenocarcinoma cell line SW480 (1 × 106 cells injected subcutaneously into the right flank) were treated on day 21 with NKP-1339. NKP-1339 was dissolved in 5 mM sterile-filtered CS buffer and was applied once intravenously at a dosage of 40 mg/kg. The tissue samples were collected 24 h after the drug administration, flash frozen in liquid nitrogen and kept in −80 °C until the XAS experiments took place 2.5 years later. This batch of samples has also been used to collect the data published in ref. 17 making the performed XAS experiments really complementary. The experiments and methods were done according to the regulations and under the approval of the Ethics Committee for the Care and Use of Laboratory Animals at the Medical University Vienna (proposal number BMWF-66.009/0084-II/3b/2013), the U.S. Public Health Service Policy on Human Care and Use of Laboratory Animals as well as the United Kingdom Coordinating Committee on Cancer Prevention Research’s Guidelines for the Welfare of Animals in Experimental Neoplasia.

XAS sample preparation

Compounds 1 to 13 were diluted in BN (Sigma Aldrich, CAS 10043-11-5, 99.5%) and held in place with conducting (graphite) tape such that the radiation was incident at an angle of 45° or placed in aluminum holders and sealed with Ultralene foil. The Kapton foil was removed on one side prior to the measurement. The BN preparations were prepared for a calculated absorption of about 1 absorbance unit according to standard methods44. The tissue samples were placed in aluminum sample holders and sealed with Kapton foil on both sides.

XAS experimental setup

XAS experiments were performed at beamline B18 at Diamond Light Source, UK. The ring energy was 3.0 GeV and the ring current 200 mA. The beamline was equipped with a Si(111) monochromator. Higher-energy harmonics were rejected using two Ni-coated Si mirrors at 10 mrad incident angle. The absolute energy calibration was performed using ruthenium powder (Sigma Aldrich, CAS 7440-18-8, 99.9%). The edge position was determined over the first maximum in the first derivative and used for the energy calibration. Spectra were collected in total electron yield, fluorescence and transmission mode for the model compounds and fluorescence mode for the tissue samples. A silicon drift detector was used for the measurements in fluorescence mode. XANES was collected from 2700 to 3100 eV in continuous scan mode with a step size of 0.1 eV. On the tissue samples 70 to 90 scans per sample were collected. The spectra of the model compounds are the average of 3 scans. All tissue samples were collected at 77 K in a “6-way cross” vessel cryostat.

XANES analysis

Initially careful radiation damage studies were performed by studying spectra collected at different time intervals. The pre-edge background was removed by a linear approximation in the range of −30 eV to −250 eV before the Cl K-edge. Spectra were normalized by fitting a linear spline function to the post-edge region and normalizing the spline to 1.0 at 2860 eV24. The edge position was determined as the maximum in the first derivative of the spectrum (inflection point of the steeply rising edge). However, at the Ru L3-edge the edge position was determined as the second maximum in the first derivative of the spectrum for all RuIII model compounds due to the splitting observed. For the construction of the 3D XANES plot, additional edge energies were taken at the maximum and half height of the Ru L2-edge (see Table 1). The coordination charge and the construction of the coordination charge versus edge plot was constructed in the same way as previously reported17. The Cl K-edge pre-edge peaks were fit with Fityk using pseudo-Voigt functions. The number of features in the second derivative determined the number of peaks to fit in the pre-edge and edge regions. In the edge region, an arctangent function was fit for the increase in continuum absorption. Fits were done in the energy range 2818−2825 eV. Spectra in Figs 1 and 2c were smoothed for improved visual appearance in ATHENA using the boxcar average algorithm.

Additional Information

How to cite this article: Blazevic, A. et al. Electronic State of Sodium trans-[Tetrachloridobis(1H-indazole)ruthenate(III)] (NKP-1339) in Tumor, Liver and Kidney Tissue of a SW480-bearing Mouse. Sci. Rep. 7, 40966; doi: 10.1038/srep40966 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.