Effects of Human Serun Albumin in Some Biological Properties of Rhodium ( II ) Complexes

Cinco complexos de ródio(II) de fórmula geral [Rh2(ponte)4] (ponte = acetato, propionato, butirato, trifluoroacetato e trifluoroacetamidato) tiveram suas afinidades em relação à albumina humana (HSA) determinadas por espectrofotometria, observando-se no caso dos alquilcarboxilatos uma correlação inversa com suas lipossolubilidades. A difusão dos complexos livres ou ligados à proteína para células de Ehrlich in vitro parece primordialmente governada pelo caráter hidrofóbico do complexo. O complexo [Rh2(tfc)4] apresentou afinidade pela proteína (K = 214,1), além de partição celular tanto em ausência (32,1%) como na presença (48,6%) de HSA. Desta forma, o composto HSA: [Rh2(tfc)4] teve sua ação antitumoral investigada em camundongos Balb-c portadores de ascite de Ehrlich, mostrando que a HSA pode ser um reservatório para o complexo de ródio.


Introduction
In the field of metal based cancer chemotherapy, rhodium(II) compounds have been studied as alternatives to the platinum derivatives (cisplatin and carboplatin).Both their antibacterial and antitumoral actions to human and murine cell lines have been a target of research [1][2][3] .Moreover, the biodistribution of these complexes was studied with more detail recently 4 , meeting the current interest in the factors that could affect the availability of these drugs to tumor cells.
The pharmacokinetics and pharmacodynamics of any drug will depend, to a large extent, on the interaction it has with human serum albumin (HSA), the most abundant plasma protein.The hypothesis of HSA acting as a "drug reservoir" for the slow release of cisplatin has already been tested under clinical conditions [5][6][7] , though previous works had questioned this possibility 8 .
The three-dimensional structure of HSA has been published recently and the structural studies of its interactions with a number of ligands have been performed [9][10][11] .Like any protein, HSA has several possible coordination sites for metal ions but two of them deserve special attention, the Cys34 thiol-free residue and the amino-terminal sequence with a histidine in the third position 9 .To our knowledge, HSA has never been co-crystallized with antitumoral platinum-group metallodrugs or with any of the highaffinity metal binding sites occupied, so the evidences for these interactions are mainly spectroscopic.Cisplatin binding to HSA has been the subject of a recent work by Sadler and coworkers 12 .Contrary to previous assumptions, these authors found that Cys34 is not the preferred binding site for cisplatin.A S,N macrochelate, probably involving the surface-exposed Met298 residue, accounted well for the shifts observed in the NMR spectra, and this was speculated to be the main cisplatin binding site of HSA.On the other hand the anti-arthritic gold compound auranofin, triethylphosphine(2,3,4,6-tetra-O-acetylglucopyranosato-S-)gold(I) was found to bind mainly to HSA through formation of a Cys34-S-Au bond 13 .The reaction of rhodium(II) carboxylates with thiols or methionine lead to the disruption of the Rh-Rh bond with formation of square-planar monomeric rhodium(II) species and to bisadducts, respectively, whose electronic absorption spectra have different patterns from both the parent complexes 14 and the albumin-bound rhodium compounds.The permanence of the characteristic Rh-Rh absorption band at ca. 540 nm after the reaction of rhodium dimers with HSA has been understood as evidence that binding to sulfur-donating residues would not be taking place and that the complex remains intact in the reactional medium 15 .Histidine (His) residues were earlier suggested as important coordination sites for these complexes 16 .This interaction may account for the displacement of the Rh-Rh absorption band from ca. 580 nm to 540 nm when rhodium complexes react with HSA 15,17,18 .This is in accordance with the observations made for the reaction of HSA and ruthenium(III) antitumor complexes, where imidazole binding at His residues is usually suggested 19 and verified through the decreased affinity of ruthenated albumin for dyethylpyrocarbonate, a common His modificator 20 .A similar study with rhodium-bound HSA should prove to be valuable in that this putative His binding could be unequivocally demonstrated.
The interactions between a number of rhodium(II) complexes and HSA have already been studied by means of spectroscopic and immunologic techniques 17,18 .A direct dependence has been noted between the liposolubility of the metal complexes and several properties, markedly their denaturing and Trp214 fluorescence quenching abilities 15 , as well as their biological activities 21

Experimental
Reagents HSA fraction V, A1653, was purchased from Sigma.Stock solutions of the protein were prepared in phosphate buffer (pH 7.4) and their concentrations were determined photometrically through the relation HSA concentration (mg mL -1 ) = 1.55A 280 22 .
In all the calculations a M r = 66439 for HSA was assumed 23 .Clinical grade HSA solution (20%, Grifols) was used in the experiments with animals.Rhodium acetate was purchased from Sigma and the other rhodium complexes were synthesized through methods already described 24,25 .The concentrations of the complexes were determined by means of atomic emission spectroscopy with argon plasma (ICP-AES, Spectroflame Modula) calibrated with rhodium atomic absorption standard from Aldrich, using the rhodium emission line at 343.489 nm.

Plasma binding of [Rh 2 (ac) 4 ]
A 20 mL sample of blood was collected from a healthy human volunteer and centrifuged at 3500 rpm for 10 min after its clotting in test tubes.Three plasma samples (1.85 mL) were incubated with 0.5 mL of 1.0 × 10 -3 mol L -1 [Rh 2 (ac) 4 ] for 0, 2 and 24 h in a water bath at 37 o C, being kept under refrigeration.Albumin was fractionated through the salting-out effect with ammonium sulfate, allowing the sedimentation for 24 h at 4 o C. Albumin fractions with 0, 2 and 24 h incubation (henceforward denominated A 0 , A 2 and A 24 ) were resuspended in 2% NaCl and dialyzed against 2 L of distilled water for 24 h at room temperature.
The dialysis bag contents were removed with Pasteur pipettes and aliquots were transferred to a mini-gel SDS-PAGE (Bio-Rad system) to assess protein purity.Protein concentrations were determined by the Lowry method 26 .Rhodium concentrations in the dialyzates were determined by atomic emission spectrometry.

Binding constants
The method originally described by Stephanos 27 was employed for the determination of the binding constants through previously published photometric measurements 15 .It was assumed only one kind of interaction between the albumin and the rhodium complex [Rh 2 (bridge) 4 ], according to the following equation: for which is valid A plot of 1/(A-A 0 ) versus 1/L 0 is linear, and the binding constant K can be estimated from the ratio of the intercept to the slope.A 0 is the initial protein absorbance at the monitored wavelength (~500 nm), and therefore equal to zero.A is the recorded absorption for each increment in the complex concentration (L 0 ).This method has been employed for the determination of K for several drugprotein interactions and specifically to estimate K for the binding of cisplatin to HSA 28 .The titrations were carried out in a Hitachi U3000 spectrophotometer, in the visible range (350 to 750 nm) at 28 o C, as already described 15 .

Partition studies
The partition of the rhodium complexes into tumor cells was determined both in the presence and absence of HSA.In the blank (protein-free) experiments, HSA was replaced by distilled water, with the further steps being identical.Five 0.5 mL HSA samples (3.7 mg mL -1 ) were incubated with 0.1 mL of each ca. 10 -3 mol L -1 rhodium(II) complex solution for 5 min at 37 o C. Triplicate 0.1 mL samples of each mixture were transferred to Eppendorf tubes.Meanwhile a suspension of ascitic Ehrlich cells was prepared.A tumor-bearing mouse was sacrificed at the moment of the experiment. 1 mL of the ascitic fluid was collected, washed twice with 10 mL PBS and centrifuged.The pellet was resuspended in 25 mL PBS and the number of intact live and dead cells was counted.0.9 mL of this suspension was transferred to Eppendorf tubes, and vortexed vigorously for 30 s.After this, the tubes were centrifuged for 5 min at 3000 rpm, and the supernatants were separated and diluted with 2 mL of distilled water.The rhodium concentration in the diluted supernatants was determined by means of the ICP-AES technique.

Antitumor in vivo activity
Thirty Balb-c mice weighting approximately 20g were inoculated with 5 × 10 5 Ehrlich cells, divided in two groups and maintained for 24 h.During this time, 0.0167g of [Rh 2 (tfc) 4 ] was diluted in water (ca.20 mL) and this presolution was dripped over 1.75 mL of HSA 20% to prepare HSA:[Rh 2 (tfc) 4 ].A characteristic color change from blue to reddish was observed, indicating the binding of the complex to the protein.After 10 min, this solution was filtered and transferred to a 25 mL volumetric flask.The final volume was completed with distilled water, and the rhodium concentration was determined by ICP-AES (concentration of [Rh 2 (tfc) 4 ] = 1.12 × 10 -4 mol L -1 ).In these conditions, n Rh /n HSA = 0.5.This assured that all of the complex mole-cules would be bound to the protein.After 24 h, 15 animals were treated i.p. with 0.5 mL HSA solution (control group) and the other 15 received i.p. 0.5 mL of HSA:[Rh 2 (tfc) 4 ], in order to reach a 2.8 µmol kg -1 dose of [Rh 2 (tfc) 4 ].Previous assays showed HSA to be innocuous to the mice.After a further 24 h, all the mice were submitted again to the same treatments.The doses have been chosen considering that 5.6 µmol kg -1 of the drug could lead to a relevant survival percentage, at the expense of important tissue damage 29 .The meaningfulness of the survival patterns of the two groups (displayed in Kaplan-Meier plots) was tested by the Mantel-Haenszel method 30 .A similar procedure was employed to assess the activity of the unbound [Rh 2 (tfc) 4 ] complex (concentration of [Rh 2 (tfc) 4 ] = 1.05 × 10 -4 mol L -1 ).

Results and Discussion
The whole plasma binding experiment (Table 1) showed that albumin is indeed a serological target of the rhodium complexes.The simple salting-out procedure for protein separation afforded >95% HSA, checked electrophoretically.The fact that around 50% saturation is attained with no incubation confirms previous spectroscopic evidences that this reaction is fast 15 .In A 2 , an increase to 70% saturation can be observed, indicating that for biological purposes it is reasonable to assume that rhodium complexes would bind promptly to albumin as soon as they reach the bloodstream.
It is commonly assumed that the main binding site for rhodium(II) complexes in HSA is the histidine (His) residue 16,17 .It was noticed that the stability of rhodium(II) complexes and His was correlated with the lipophilic nature of the metal compounds.Binding constants to the formation of adducts of the kind [Rh 2 (bridge) 4 ] (His) (bridge = ac, pr) were reported as 4.05 and 4.14, respectively, the difference being attributed to the lesser solvation and therefore the increased propionate binding to His 31 .
However, in this work an inverse correlation was observed between the lipophilic nature of the rhodium alkylcarboxylates and the value of K for binding with HSA.(Figure 2 and Table 2).alkylcarboxylates, which suggests an additional stabilization effect nonexistent in the reaction with the free aminoacid.This possibility has indeed been proposed 31 .In fact, one could consider that the overall affinity between rhodium(II) complexes and HSA results from two distinct phenomena: (a) axial coordination with His residues, and (b) hydrophobic interactions between the complex and the protein skeleton.
The affinity values obtained through spectrophotometry reflect specifically the case (a) above.It is suggested that the axial binding of the alkylcarboxylates is weaker for the more lipophilic complexes precisely because of the larger contribution of the hydrophobic interactions (case b).A glance at the scheme proposed in Figure 3 is elucidative.The relative sizes of the arrows indicate the importance of each interactive phenomenon for a given complex.The interaction with histidine residues in the protein molecule seems to be favored for the more solvated Even though exhibiting high hydrosolubility, the His coordination with [Rh 2 (tfa) 4 ] should be of importance, given the higher hardness of the metal center.This effect apparently is lost when CF 3 COO bridging ligands are substituted by the correlated amidate, leading to a better π-donator complex to the axial ligand 32 .
The magnitude of the obtained binding constants is in accordance with the reported results for the interaction of cisplatin with HSA (K = 852) 28 .These affinities are not as high as those described for other drugs (K ~ 10 5 -10 7 ) and thus possibly they are less of a hindrance to release from the protein.So it seems reasonable that the complexes transfer to the target cells by passive diffusion governed by lipophilicity (HSA-bound drugs are not available to cross the cell membrane 33 ).
Turning to the partition results (Table 3), it can be seen that indeed there are different absorptions by the cell suspension correlated to the lipophilicity of the rhodium alkylcarboxylates.In the partition into cells, the complex [Rh 2 (tfa) 4 ] behaved similarly to [Rh 2 (ac) 4 ], even though it is more lipophilic.Considering its already described affinity to HSA 15 , a lesser transference of [Rh 2 (tfa) 4 ] from the protein to the cells would not be surprising.However, the explanation for the relatively small partition of the unbound complex into cells seems to involve other phenomena at the membrane level, which deserve further studies.The complex [Rh 2 (tfc) 4 ] diffused into cells more readily than [Rh 2 (ac) 4 ] (ca.20%) when bound to the protein but this trend is reversed for the free complexes.A more favorable release from the protein could be carried out by the trifluoroacetamidate ligands.
The survival profiles of the mice treated with both HSAbound and unbound [Rh 2 (tfc) 4 ] complex are presented in Figures 4 and 5. [Rh 2 (tfc) 4 ] has been elected to in vivo studies to give continuity to our previous research on the antitumoral action of rhodium(II) amidates 29,34 .tumoral focus caused either by the small amount of inoculated tumor cells or by a therapeutic effect of the drug.The values of the Mantel-Haenszel test (6.8 and 19.4,respectively for HSA:[Rh 2 (tfc) 4 ] and [Rh 2 (tfc) 4 ]) show that the differences in the survival profiles between treated and untreated groups are statistically meaningful (p < 0.01).
These results, together with the above considerations relating to the affinity of the rhodium complexes with HSA and/or intact cells, indicate that albumin can be a drug reservoir for [Rh 2 (tfc) 4 ], and probably for the other rhodium complexes as well, by means of capturing it in the bloodstream after injection and transferring it slowly to the tumor cell by passive diffusion.The possibility for some sort of active transport mechanism through the membrane, however, can not be ruled out with these results.Some of the mice treated with HSA:[Rh 2 (tfc) 4 ] which survived after the period of investigation (ca.30 days) exhibited solid tumors, a physiologic response to the

Figure 2 .
Figure 2.Plot of 1/(A-A 0 ) versus 1/L 0 to the formation of adducts between rhodium(II) complexes and HSA.A is the absorbance at different concentrations L 0 of the complexes.A 0 is the initial absorbance of the albumin at these wavelengths.Throughout this experiment, A 0 = 0.

Figure 3 .
Figure 3. Schematic correlation between lipophilicity of the rhodium(II) alkylcarboxylates (left arrow) and their interactions with HSA.a: axial binding with His residue.b: lipophilic interaction.The relative sizes of the arrows (out of scale) indicate the importance of each kind of interaction for a given complex.It is suggested that a and b are competitive.

Table 1 .
Results of the whole plasma binding test of [Rh 2 (ac) 4 ].

Table 2 .
Affinity results for the binding of rhodium (II) complexes and HSA.