Modification of carboplatin by Jurkat cells

Using [H,N] heteronuclear single quantum coherance (HSQC) NMR and N-labeled carboplatin, 1, we show that Jurkat cells affect the rate of disappearance of the HSQC NMR peak in culture medium for this Pt anticancer drug. The decay or disappearance rate constant for 1 in culture medium containing cells is k1 1⁄4 kc1⁄2CO 3 þ km þ kuN , where kc is the rate constant for reaction of 1 with carbonate in the medium, km is the rate constant for reaction of 1 with all other components of the medium, and ku is the rate constant for reaction of 1 with cells having a number density N in the medium. Since Jurkat cells only take up a small amount of the platinum present in the medium (<1%), the observed disappearance of the HSQC NMR peak for 1 cannot be due to uptake of carboplatin by the cells. 2007 Elsevier Inc. All rights reserved.

We here use [ 1 H, 15 N] HSQC NMR and 15 N-labeled 1 to show that the rate of disappearance of carboplatin in the medium is affected not only by substances present in the culture medium but also by an unknown substance(s) released by the cells to the medium. As with previously reported experiments with cisplatin [4], no HSQC NMR active product peaks are observed in the NMR experiment with carboplatin and Jurkat cells. Since Jurkat cells take up only a small fraction of the total amount of platinum present in the medium [3], the disappearance of 1 cannot be due to simple absorption of the compound by the cells. We also demonstrate that the results of the cell/NMR experiments are independent of the nature of the atmosphere above the cells. Rate constants for the disappearance of 1 in the presence of Jurkat cells in a static atmosphere (closed NMR tube) fit well (R 2 = 0.96) with the previously reported rate constant for the disappearance of carboplatin in the presence of 4.5 · 10 7 cells mL À1 kept under standard conditions in a humidified, 5% CO 2 atmosphere in a incubator [2]. . 15 N-labeled carboplatin was prepared as previously described [2]. [ 1 H, 15 N] HSQC NMR spectra were recorded on a Bruker Avance 500 MHz NMR equipped with a 5 mm triple axis probe in the earlier described manner [5]. Peak volumes in arbitrary units, excluding 195 Pt satellites, were calculated using Bruker software. Each kinetic run consisted of ten successive NMR measurements (each 60 min in length, ns = 48) for a total duration of $10 h. The 15 N chemical shift of 1 was referenced externally to 1 M ( 15 NH 4 ) 2 SO 4 in 5% D 2 O which was acidified to pH $1 by the addition of H 2 SO 4 . The 1 H chemical shift of 1 was referenced externally to 3trimethylsilyl propionic acid-d 4 sodium salt (TSP) in a pH 7.15, 23 mM sodium carbonate buffer. Freshly prepared labeled carboplatin in water was added to culture medium (RPMI + 10% FBS, 100 lg/mL streptomycin, 100 IU/mL penicillin, and 2.0 mM S-glutamine) containing either 2.7, 8.2, 21.2 or 27.2 · 10 7 Jurkat cells per mL, in a final volume of 920 ll, and spectra were recorded in 5% D 2 O at 37°C. Jurkat cells used were $97% viable, measured by light microscopy, using a hemacytometer under standard trypan blue conditions [6].

Results and discussion
As is shown in Fig. 2, 15 N-labeled carboplatin exhibits a single HSQC NMR peak due to the two symmetry-equivalent ammonia molecules. The pseudo-first order decay constant, k 1 , for the disappearance of carboplatin in culture medium containing Jurkat cells was determined by fitting a plot of carboplatin concentration (calculated from the HSQC NMR peak intensity) versus time to a line and dividing the slope of the linear fit by its y-intercept. The values obtained for k 1 are given in Table 1, with statistical errors from the linear fits, and are plotted versus cell den-sity in Fig. 3. Also plotted in Fig. 3 are the previously reported k 1 values for the disappearance of carboplatin in   Fig. 3. Observed rate constant of the disappearance of carboplatin obtained from HSQC NMR as a function of number of cells per mL. Open diamonds are k 1 values already reported by us [2]. The best fit line, R 2 = 0.96, is k 1 = (1.94 · 10 À6 + 5.14 · 10 À14 mL cell À1 N) s À1 , so that k u = 5.14 · 10 À14 mL cell À1 s À1 . culture medium (no cells) and in culture medium containing 4.5 · 10 7 cells mL À1 kept under standard conditions in a humidified 5% CO 2 atmosphere in an incubator [2]. In the previously described cell experiment [2], aliquots of cells plus medium were removed from the culture flask, cells were separated from the medium by centrifugation, 5% D 2 O was added to the medium and NMR data was collected for 0.5 h. As can be seen in Fig. 3, all the data points are consistent with the best fit (R 2 = 0.96) line k 1 s = 1.94 · 10 À6 + 5.14 · 10 À14 (mL cell À1 )N, demonstrating that cell studies will yield the same results regardless if the reaction with drug is done in a 5 mm capped NMR tube or in a humidified, 5% CO 2 incubator.
A striking feature of the reaction of 1 in these studies is that no new HSQC NMR peaks associated with products are observed. As we, as well as Sadler and coworkers, have shown [2,[7][8][9], reaction of 1 with nucleophiles lifts the equivalency of the two ammonia molecules to produce new peaks, the positions of which are sensitive to the nature of the ligands trans to the ammonia molecules. Sulfur in the form of a thiol (cysteine) or thioether (methionine) is the most potent nucleophile for Pt 2+ present in culture media. While a sulfur ligand trans to an ammonia molecule would make the latter vulnerable to a substitution reaction due to the trans effect, these complexes, with N trans to S, are sufficiently stable to be easily detected at $À40 ppm (N) using HSQC NMR [4,[7][8][9][10][11]. No products of this type or of any other species were observed in the reaction.
The prospect that the disappearance of the HSQC NMR peak for 1 in the presence of cells is due to absorption of carboplatin by the cells is highly unlikely. As we earlier showed, Jurkat cells under conditions similar to those used here remove platinum species from solution at a rate which is only 16.6 ± 4.2 · 10 À18 mol Pt h À1 cell À1 [3]. This rate of platinum removal by cells is too slow to show an observable effect on the total pool of 1 in the medium above the cells during the time course of these experiments.
In the presence of culture medium without cells, the concentration of carboplatin [1] should obey where k c is the rate constant for reaction of 1 with carbonate in the medium, a second order process, which behaves like a pseudo-first order reaction because [CO 2À 3 ] is constant under the reaction conditions used [2]. The quantity k m is the rate constant for reaction of 1 with other substances in the medium, which is assumed to be a first order process. For example, earlier studies showed that FBS, a suite of proteins needed for cell growth, is also a significant factor causing the disappearance of 1 in culture media [2].
If cells, or some substance released by them, can modify carboplatin so that it does not give an NMR signal, a term Àk u [1]N should be added to the equation for d [1]/dt, where N is the density of cells in the medium. The solution to the resulting equation is an exponential decay, The observed values of k 1 , shown in Fig. 3, depend linearly on cell density. The y-intercept, 1.94 ± 0.76 · 10 À6 s À1 represents k c ½CO 2À 3 þ k m ; from the slope, we obtain k u = 5.14 ± 0.52 · 10 À14 mL cell À1 s À1 . When the density of cells is P8.2 · 10 7 cells mL À1 , the reaction of carboplatin with cells is more important than reaction of drug with carbonate or other substances in the culture medium. Since, as noted, direct removal of 1 by cells is unimportant, the reaction must involve a substance released by the cells.
Earlier studies by us using cisplatin [4,5] showed that Jurkat cells can rapidly modify the mono-carbonato complex, cis-[Pt(NH 3 ) 2 (CO 3 )Cl] À , converting it into a substance that is not capable of binding to the cells and is undetectable using HSQC NMR. Moreover, the amount of modification of the mono-carbonato complex is dependent on the number of cells and recent studies by our group (Centerwall et al., personal communication) show that the platinum-modifying substance is released to the culture medium above the cells. Unlike the relatively slow decay rate observed for carboplatin in the presence of cells in these experiments, the rate of modification of the cisplatin mono-carbonato complex by Jurkat cells is extremely rapid, occurring within $0.6 h of the addition of drug to the culture medium above the cells [4,5]. It seems likely that both drugs are exposed to the same cell-released modifying substance, but since carboplatin has a bidentate chelate ring which blocks nucleophiles from attacking the axial sites of the compound, it is much less susceptible to chemical modification than is the cisplatin mono-carbonato species. Clearly, additional work is needed to determine the nature of the species released by the cells and products formed in the reaction.

Conclusion
Jurkat cells release a substance to the culture medium that reacts with carboplatin, converting it into a substance that, under the conditions of the experiment, is not detectable using HSQC NMR. This disappearance from HSQC NMR is identical to that observed for a mono-carbonato complex of cisplatin except that, due to structural differences between cisplatin and carboplatin, the rate of reaction with carboplatin is much slower than with cisplatin.