Comparison of the Antiproliferative Activity of Two Antitumour Ruthenium(III) Complexes With Their Apotransferrin and Transferrin-Bound Forms in a Human Colon Cancer Cell Line

Two ruthenium(III) complexes, namely trans-indazolium[tetrachlorobis(indazole)- ruthenate(III)], HInd[RuInd2Cl4] and trans-imidazolium[tetrachlorobis(imidazole)- ruthenate(III)], HIm[RuIm2Cl4] exhibit high anticancer activity in an autochthonous colorectal carcinoma model in rats. Recently, it has been shown that both complexes bind specifically to human serum apotransferrin and the resulting adducts have been studied through spectroscopic and chromatographic techniques with the ultimate goal of preparing adducts with good selectivity for cancer cells due to the fact that tumour cells express high amounts of transferrin receptors on their cell surface. In order to investigate whether the cellular uptake of the complexes was mediated by apotransferrin or transferrin, we compared the antiproliferative efficacy of HInd[RuInd2Cl4] and HIm[RuIm2Cl4] with its apotransferrin- and transferrin-bound form in the human colon cancer cell line SW707 using the microculture tetrazolium test (MTT). Our results show that especially the transferrin-bound forms exhibit high antiproliferative activity, which exceeds that of the free complex, indicating that this protein can act as a carrier of the ruthenium complexes into the tumor cell.


Introduction
A number of Ru(III) complexes exhibit antitumour activity in animal models (Clarke, 1989). Keppler et al. have recently developed anticancer ruthenium(III) complexes with N-These two complexes show high antitumour activity against an autochthonous colorectal cancer model in rats, a model which simulates the colon cancer of humans very well . Furthermore, both complexes exhibit antiproliferative activity in two human colon cancer cell lines (SW707 and SW948) (Galeano et al., 1992). Considering the observed IC50 concentrations in vitro of 110 tg/ml for Hlnd[Rulnd2C14] and 250 tg/ml for Hlm[Rulm2Cl4] in SW707 with respect to possible tissue levels in vivo, it can be assumed from the dosage active in vivo of 13 mg/ml forHlnd[Rulnd2C14] and 7-10 mg/ml for HIm[Rulm2C14] that levels as high as the IC50 value can be reached for a short period oftime only or by tissue accumulation. Since malignant cells have a high iron requirement and, consequently, a large number of receptors for the iron(III)-transport protein transferfin, it has been suggested that the accumulation of ruthenium(III) complexes in tumours might be mediated by this plasma protein (Clarke, 1989). Indeed, it has been previously reported that transferrin is responsible for the selective delivery of radioactive 67Ga(III) complexes to tumor tissues (Ward and Taylor, 1988). The simaificance of cellular uptake of transferrin with respect to various malignant diseases has been reviewed (Aulbert, 1986). Hence, to obtain a deeper insight of the interactions of antitumour Ru(III) complexes with serum proteins, we have recently investigated the reactions of Hlnd[Rulnd2C14] and HIm[RuIm2C14] with human serum (Kratz et al., 1992 and1993) and especially with apotransferrin (Kratz et al., 1994a): Protein-binding is rapid for HInd[RuInd2C14] (3-4 minutes at T 37 C) but takes about 5 hours for HIm[Rulm2C14] the major amount being bound to human serum albumin and a smaller fraction to human serum transferrin.
Both complexes bind specifically around the iron(III) binding sites of apotransferrin as shown by circular dichroism spectroscopy. More precise information on the binding sites of the complexes has been revealed by soaking experiments of human apolactoferrin crystals with the two Ru(III) complexes. Preliminary results of the subsequent X-ray structure analyses have been reported (Kratz et al. 1994b) demonstrating that both complexes bind L. Messori and M.R. Berger to the unoccupied binding site with the N-hteroyles remaining attached to the ruthenium Due to the protein-binding properties of these omplexes, we wanted to investigate whether their ellular uptake was mediated by transferrin and therefore ompared the antiprolifrative efficacy of HInd[RuInd2Cl4] and HIm[RuIm2C14] with its apotransferrinand transferrin-bound form in the human colon cancer cell line SW707 using the microCulture tetrazolium test (MTT). The results of these experiments are reported in this communiation.

Materials and methods
Substances and equipment. The ruthenium(III) complexes were synthesized as described previously . For the cell culture experiments the pure complexes were dissolved in 0.9% saline solution (c 200 lag/ml). Human serum apotransferrin (98%, crystalline, essentially iron free, MW 80,000) (apoTf hereafter) and human serum transferrin (98%, crystalline, iron saturated, MW 80,000) were purchased from Serva, Heidelberg and Sigma Chemical Company respectively.

Preparation of Hlnd[Rulnd2Cl4] and HIm[Rulm2Cl4] bound to apotransferrin and transferrin
respectively. In the cell culture experiments varying amounts of the complexes (2, 4 and 8 equivalents) were bound to the proteins. The final concentration of the protein/complex solution was e 200 tg of Ru(III)eomplex per ml. In order to prepare 11.0 ml of the protein/complex solution with 2 equivalents of the complex, the following procedure was followed: 5.5 ml of an aqueous solution of the Ru(III) complex with e 0.4 mg/ml (prepared by stirring at 37 C in bidistilled H20 for 5-10 minutes) were added to a 5.5 ml solution of the protein (150.0 mg of apotransferrin or transferrin dissolved in the following buffer: 0.2 M NaC1, 0.008 M NaH2PO4, 0.05 M NaHCO3) and then stirred for 1 hour in the ease of HInd[RuInd2C14] (color of the solution changes from brown to green) and for 12 hours in the ease of HIm[RuIm2C14] (color of the solution changes from orange to dark red). For binding 4 and 8 equivalents of the complex, 75.0 or 37.5 mg of either apotransferrin or transferrin were used.
Previous HPLC studies and ultra,filtration experiments have demonstrated that no free complex is present after the stated time periods (Kratz et al. 1992 and1994). Thus, the protein/complex solutions were directly employed for the cell culture experiments. Apotransferrin or transferrin running as control were dissolved in the same buffer as above and then diluted with bidistilled H20 (150.0, 75.0, or 37.5 mg protein dissolved in 5.5 ml buffer to which 5.5 ml bidistilled. H20 were added).
Cell lines. For the in vitro experiments a human cancer cell line SW707 was used, which was derived from patients with colorectal cancer and was provided by the Tumor Bank of the Institute of Experimental Pathology, German Cancer Research Center (Heidelberg, FRG). SW707 is a cell line derived from welldifferentiated adenocarcinomas of the rectum with a modal chromosome number of 47. Microscopically, SW707 showed no microvesicular bodies. The cell line was routinely checked for their mucus production to exclude the possibility that fibroblasts had overgrown the carcinoma cells. Moreover, they were tested for The manuscript giving details of the X-ray structure analysis is in preparation (Baker et al.) mycoplasma contamination and proved to be negative. The cell lines were grown as monolayer cultures in MEM medium (Gibco) supplemented with 10% heat-inactivated (57 C, 40 .min) fetal calf serum, streptomycin (100 tg/ml), penicillin (100 IU/ml) and L-glutamine (2 tmol/ml; all from Serva, Heidelberg, FRG). The cells were maintained in a humified atmosphere (5% CO2/95% air) at 37 C.
MTT test. Monolayer cell cultures were trypsinised and single-cell suspensions were obtained by repeated pipetting. The percentage of viability was determined by the trypan-blue exclusion test. A final concentration of approximately 2 x 104 cells/ml medium was prepared, and then 1.0 ml was added to each well of a 24-well culture plate. On the day following the plating; of cells, the test compounds were added at the respective concentrations. The MTT assay was performed after the cells had been incubated with the test substances for 3 consecutive days (days 2, 3 and 4). 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; Sigma, FRG) was dissolved in phosphate-buffered saline (Oxoid Ltd., Basingstoke, England) at 5 mg/ml and filtered through a 0.22-tm Millipore filter, after which 100 tl of this MTT solution was added to each well (0.1 ml/ml medium). After incubation of the plates with MTT for 1 h at 37 C, the medium was discarded and 500 tl acid-isopropanol (0.2 ml 0.04N HC1 in 10 ml isopropanol) was added to all wells to stop the enzyme reaction. Within 1 h of this addition the extinction of the colored formazon derivative dissolved in acid-isopropanol was determined on a flow Flow Multiscan MC plate reader at a wavelength of 540 nm (reference wavelength: 690 nm). At a given concentration, the MTT test was performend in 4 wells and the mean value calculated from these experiments is presented in Figures 1 and 2. Standard deviations were below 15% and are not shown in the figure.

Results
The results of our cell culture experiments are shown in Figures 1A and B Figures 2A and B for the HIm[Rulm2C14] series. 2, 4 and 8 equivalents ofthe Ru(III) complex were bound to the proteins in order to establish whether varying amounts of the bound complex had an effect on antiproliferative efficacy. The extinction values are displayed as columns for days 2, 3, and 4, and the tested concentrations were 100, 50 and 25 tg/ml ofthe free or protein-bound complex. Figure 1A illustrates that the transferrin-bound forms of HInd[Rulnd2C14] exhibit a strong inhibitory effect at all of the concentrations tested when compared to the control and exceeds that ofthe free complex at the low dose of 25 tg/ml. By contrast, transferrin alone exhibits a strong stimulatory effect at high concentrations, and this result was to be expected considering that transferrin is an essential growth factor for cell culture.
The apotransferrin-bound form of HInd[RuInd2C14] is not as active as the transferrinbound form as can be seen in Figure lB. Also, the free complex exhibits a strongerinhibitory effect than the apotransferrin-bound forms. Nevertheless, the proteinbound form exhibits quite a strong inhibitory effect at the high concentration of 100 tg/ml compared to the control and exceeds that of pure apotransferrin, which also shows an inhibitory effect compared to the control.  In contrast, the apotransfrrin-bound forms of th scond complex, HIm[RuIm2C14] xhibit a stronger inhibitory ffct than th fre complex, th protein-bound forms being about twic as activ as th fr complex ( Figure 2B). Similar rsults wr obtained for th transfnn-bound form (Figur 2A), th protein-bound form bing two to thr tims mor effective than the free complex. Surprisingly, transferrin on its own did not stimulate cell growth in this experiment prhaps du to th less pronounced growth of th tumor calls which was slower than in the xpdmnts with HInd[RuInd2C14]. In both xpdmnts th amount of complex bound to th proteins (2, 4, or 8 quivalnts) did not hav a significant ffect on antiprolifrativ fficacy.

Discussion
It has ben shown recently that the most promising antitumour Ru(III) complex, HInd[RuInd2C14] which is far less toxic in long trm application than the imidazole analogue, binds within a fw minutes to the srum proteins albumin and transferrin (Kratz t al., 1992). This rsult suggests that apotransferdn or transfrdn could act as a natural carder of th drug to the tumor tissues owing to the high affinity between this mtal transport protein and the larg number oftransfrdn receptors on the sttrfac oftumor cells [Brock, 1985). There is meanwhil a considerable amount of videnc available which demonstrates hat tumours in animals and man exhibit a vry high, "parasitic" uptake of iron(III) when they are fast growing. AulbCrt has reviewed th literature on this phenomenon (Aulbert, 1986) and together with his investigations demonstrated that turnout cells exhibit a high amount of transfrdn rcptors by which they sequester iron. Especially fast growing tumour cells show a much higher uptake ofiron-transferrin than the normal surrounding tissue of th affected organ. For these reasons, it is therefore tmpting to exploit th transferdn cycl as a "natural" rout for th sdectiv ddivry of cytostatic drugs to cancer calls. Transferrin is the iron-carrying protein in th serum of vertebrates. It is a -glycoprotein with a molecular wight of 80 000 and is present in human serum at a concentration of 2.5-3.5 mg/ml. The iron-free transfrdn is called apotransferrin and can bind two equivalents of F(III) in th Nand C-trminal lobes of the protein upon which the protein undergoes a conformational change. When encountering the cell surface, iron-bound transferrin binds to a specific cell surface receptor (transferrin receptor). The next step is intemalisation of the transferrin-receptor complex by endocytosis, followed by removal of iron from transferrin with the aid of intracellular chelates and involves fusion with endosomes, where the pH is lower (pH 5); finally the apotransferrin is returned to the extracellular medium. If the ruthenium complexes are being transported by transferrin or apotransferrin to the tumour cells, they have to be released again to exert their antitumour effect after the protein/complex-adduct has entered the tumour cell. The release should occur at low pHvalues (pH 5) with the aid of chelates. We have previously shown that both Ru(III) complexes can be released tom apotransferrin or transferrin at pH 4-5 in the presence of citric acid or ATP (Kratz et al., 1994), presumably in a form in which the chlorides of the original complex are substituted by citrate and ATP respectively, so that the cellular uptake ofthe lu(IH) complexes could follow the transferrin path.
Our cell culture experiments demonstrate that the transfen-botmd forms of both complexes exhibit high antiproliferative activity which even exceeds that of the free complex. The results for apotransferrin-bound forms, however, show a different picture: the apotransferdn-botmd form of Hlnd[Rulnd2C14] is not as active as the respective free complex, but the bound form of Hlm[Rulm2C14] is. This difference could possibly be explained by the larger change in the protein conformation when HIm[Rulm2C14] binds to apotransferfin due to the smaller N-heterocycle. This change in the conformation is then similar to that that which takes place when iron(III) is bound then, so that the high affinity between the transferrin receptor and transferrin-bound form ofthe complex is retained.
Although we have not demonstrated that the Ru(llI) complexes are released inside the cell, the transferdn-bound forms of both complexes and the apotransferrin-botmd form of Hlm[Rulm2Cl4] retain their antiproliferative effect. Our results therefore shed light on the discrepancies observed between the in vitro IC50 values of HInd[Rulnd2C14] (110 mg/ml) and the dosages active in vivo (13.0 mg/kg). Tissue levels as high as the IC50 value can only be reached for short periods of time at this in vivo dosage or by tissue accumulation which could be achieved by transferrin acting as a carrier ofthe Ru(III) complex.
We are therefore planning animal experiments to strengthen our working hypothesis.