New Extracellular Resistance Mechanism for Cisplatin

The HSQC NMR spectrum of N-cisplatin in cell growth media shows resonances corresponding to the monocarbonato complex, cis-[Pt(NH3)2(CO3)Cl] , 4, and the dicarbonato complex, cis-[Pt(NH3)2(CO3)2] , 5, in addition to cisplatin itself, cis-[Pt(NH3)2Cl2], 1. The presence of Jurkat cells reduces the amount of detectable carbonato species by (2.8 ± 0.7) fmol per cell and has little effect on species 1. Jurkat cells made resistant to cisplatin reduce the amount of detectable carbonato species by (7.9 ± 5.6) fmol per cell and also reduce the amount of 1 by (3.4 ± 0.9) fmol per cell. The amount of detectable carbonato species is also reduced by addition of the drug to medium that has previously been in contact with normal Jurkat cells (cells removed); the reduction is greater when drug is added to medium previously in contact with resistant Jurkat cells (cells removed). This shows that the platinum species are modified by a cell-produced substance that is released to the medium. Since the modified species have been shown not to enter or bind to cells, and since resistant cells modify more than non-resistant cells, the modification constitutes a new extracellular mechanism for cisplatin resistance which merits further attention. 2008 Elsevier Inc. All rights reserved.


Introduction
cis-Diamminedichloroplatinum(II), cisplatin (1; Fig. 1), has been used for decades for the treatment of neck, testicular, and other types of cancer [1][2][3]. A serious shortcoming with this and other drugs for treating cancer is that the tumor ultimately becomes resistant to the agent, forcing the discontinuance of chemotherapy. In the case of cisplatin, resistance mechanisms involving interception of the drug by cellular thiols, repair of platinated genomic DNA, rapid efflux/reduced uptake, and other processes are believed to be responsible for the resistance developed in cisplatin therapy [4][5][6][7][8][9][10]. A common feature of all known resistance mechanisms is that they are associated with the cell itself and not the medium outside the cell.
Earlier we used a sensitive technique, heteronuclear single quantum coherence NMR (HSQC NMR) and 15 Nlabeled cisplatin to study the effects of Jurkat cells on the speciation of platinum in the culture medium outside the cell [11,12]. We found that the platinum species, cis-[Pt(NH 3 ) 2 (CO 3 )Cl] À , 4, Fig. 1, which forms in culture medium from the monoaquo/hydroxo species 2/3, is modified by Jurkat cells in a manner which depends on the number of cells present in the medium. In subsequent studies [13] we showed that each Jurkat cell can modify 2.8 fmol of 4 within $0.6 h and that the modified platinum does not bind to the cell. Because there is only a slow decrease (À1.1 ± 0.4 lM h À1 ) in the amount of unmodified 4 remaining in the medium after 1 h, the cells subsequently lose their ability to modify 4. We suggested that these results are consistent with a hitherto undocumented 0162-0134/$ -see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.jinorgbio.2008.01.008 extracellular detoxification mechanism in which the cells rapidly modify a portion of 4 in the culture medium. Subsequent work by us [14] with the anticancer drug carboplatin, [Pt(NH 3 ) 2 (CBDCA-O,O 0 )], where CBDCA is cyclobutane-1,1-dicarboxylate, showed that Jurkat cells can affect the rate of disappearance of the HSQC NMR peak for carboplatin in a manner which is dependent on the number of cells present in the culture medium. Since the Jurkat cells take up only a small amount of the platinum present in the medium (<1%), the disappearance of the HSQC NMR peak of carboplatin cannot be due to absorption of the drug by the cells. The study suggested that the cells chemically modify carboplatin converting it into to an NMR-silent species.
Suspecting that these observations may indicate a new resistance mechanism for the platinum drugs that operates outside the cell, we made Jurkat cells resistant to cisplatin and used HSQC NMR to determine if and to what extent resistant cells can also affect the distribution of platinum species in the culture medium. In this report we show that resistant Jurkat cells are much more effective than normal Jurkat cells in modifying the monocarbonato complex 4 and the dicarbonato complex 5, and that they can also modify cisplatin, 1. The modifications take place even when the cells are removed from the medium, showing that the platinum-modifying substance is released by the cells to the medium so that the modification takes place outside the cell. These observations are consistent with a new extracellular resistance mechanism for cisplatin.

Cisplatin resistant jurkat cells
Resistant Jurkat cells were prepared by exposing nonresistant Jurkat cells to increasing concentrations of cisplatin (Bedford Laboratories, Bedford, OH), in the concentration range, 1-50 lM, over the course of 5 months. The cells were maintained in suspension culture under a fully humidified atmosphere containing 5% CO 2 at 37°C. The medium was RPMI-1640 supplemented with 10% (v/ v) FBS, 100 lg/mL streptomycin, 100 IU/mL penicillin, and 2.0 mM L-glutamine. For concentrations of cisplatin in the range, 1-10 lM, cells were exposed to drug for $12 h, after which time, the cells were sedimented (200 Â g) for 5 min, the drug-containing medium was removed and the cells were resuspended in fresh medium. These cells were allowed to recover for 3-4 days before exposure to a higher concentration of drug. For drug concentrations greater than 10lM, exposure of cells to drug was 2 h, followed by a 3-4 day recovery period before the next exposure to drug. Viabilities were assessed by light microscopy using a hemacytometer under standard trypan blue staining conditions [15].

1 H-15 N HSQC NMR
The details of the 1 H-15 N HSQC NMR measurements involving 15 N-labeled were published previously [11]. Briefly described, the NMR experiments were two-dimensional, 1 H-15 N, with inverse detection and decoupling during acquisition without spinning the sample. Each NMR experiment involving 65 lM drug in the culture medium was 62 min (48 scans), giving 103 data points in the proton dimension and 64 t 1 values. The time for the first NMR time point (t = 0) was taken as the NMR data collection time plus $15 min for temperature equilibration divided by 2 or $0.63 h after the addition of labeled drug to the medium. The stock solution used in these experiments, which contained $3 mM 15 N-labeled cisplatin in 154 mM NaCl, was allowed to reach equilibrium 24 h before addition of drug to the medium containing the cells. NMR data for a particular number of cells was collected as a function of time for the duration of the 10 h experiment, and plots of the concentrations of species present as a function of time constructed. The NMR experiments involving 400 lM drug were also 10 h in length. In these experiments, spectra were measured every 16  were equilibrated for 24 h prior to their use. The HSQC NMR spectra were collected in a capped tube at 37°C using a Bruker DRX500 AVANCE spectrometer ( 15 N; 50.646 MHz) equipped with a 5-mm triple axis probe. The NMR chemical shifts were referenced externally to 1 M ( 15 NH 4 ) 2 SO 4 in 95%, 5% H 2 O/D 2 O, which was acidified to pH 1 by addition of H 2 SO 4 . The 1 H chemical shifts were referenced to external Me 3 SiCD 2 CD 2 CO 2 Na, 2,2-dimethyl-2-silapentane-5-sulfonate sodium salt, in a 23 mM, pH 7.2, bicarbonate solution. Volume integrations of peaks were obtained using Bruker software.
The samples for HSQC NMR measurement of the amounts of platinum species as a function of the number of resistant Jurkat cells present in the medium were prepared by suspending the desired number of cells in 900 ll of medium containing 95%/5% H 2 O/D 2 O, pH 7.2. Addition of 20 ll of the stock solution of 15 N-labeled cisplatin to the medium gave a final concentration of 65 lM total platinum in the medium in a final volume of 920 ll. The chloride concentration in the final solution, controlled by the RPMI 1640 in the medium, was 105 mM. No NMR peak for NH 3 trans to CO À2 3 of 4 was observed for greater than 5 Â 10 5 resistant Jurkat cells in the NMR experiment.
Experiments showing that the platinum-modifying substance is released by the cells to the medium were done as follows: 1.0 Â 10 7 resistant Jurkat cells were incubated in $1.0 mL of medium for 1 h in a capped Eppendorf tube at 37°C. After 1 h, the cells were sedimented (200 Â g) for 3 min and the 822 lL of medium containing the cellreleased modifying substance was removed. To the recovered medium were added 50 lL of D 2 O, and 128 lL of 3.13 mM 15 N-cisplatin previously equilibrated in either 154 or 105 mM NaCl to give a final concentration of total platinum of 400 lM in a final volume of 1 mL.

Modification of species detected by HSQC NMR
To study the effects of resistant Jurkat cells on cisplatin, the HSQC NMR peak intensities of cisplatin, 1, and its monocarbonato complex, 4, I 1 and I 4 , respectively, Fig. 2, were measured as functions of time after the addition of the drug to culture medium containing different numbers of cells. Fitting these plots of intensity vs. time to exponentials allowed for extrapolation back to t = 0 for each NMR experiment, giving the initial intensity of the species in the presence of the indicated number of cells. With no cells present in the culture medium, I 1 (0) + I 4 (0) represents 65 lM drug, giving the conversion factor C between NMR peak intensity (peak volume) and concentration. As was earlier shown by us [12], the monocarbonato complex 4 has non-equivalent ammonia molecules, Fig. 1, and at neutral pH, the HSQC NMR peak for the ammonia molecule trans to Cl À of 4 lies under the HSQC NMR peak for 1.    (Fig. 3).
In order to determine the number of moles of each platinum species that is modified by a single cell, we multiply the slopes by the reaction volume (920 lL) and divide by 10 6 . We find that one resistant cell modifies (3.4 ± 0.9) fmol of 1 and (7.9 ± 5.6) fmol of 4 in less than $0.6 h (the time at which we make our first NMR measurement). For comparison, a normal Jurkat cell modifies (2.8 ± 0.9) fmol of 4 and only slightly modifies 1, [13].

Release of the platinum-modifying substance to the medium
As indicated in the Materials and methods, these experiments were done by placing normal or resistant Jurkat cells in fresh medium for 1 h, separating the cells from the medium by centrifugation, adding 15 N-cisplatin to the medium and collecting HSQC NMR data on the medium containing the drug. In order to decrease the time required to collect acceptable NMR data, the final concentration of total platinum in these experiments was 400 lM. In addition, stock drug solutions were pre-equilibrated at two different NaCl concentrations, 154 and 105 mM. Typical NMR results for these experiments are shown in Fig. 4, with a, b, and c for medium only, medium exposed to normal Jurkat cells, and medium exposed to resistant Jurkat cells. The true intensities for species 1 and 4 were obtained as I 1 -I 4 and 2I 4 , respectively. The intensity of species 5 was never more than 10% of the intensity of 4. Since 5, a carbonato species like 4, was also attacked by the modifying substance, I 5 was added to the intensity of 4 before obtaining the true intensities.
We expect that intensities for all species decrease with time because of the slow reaction of these compounds with substances in the medium [11,13,14], so it is not possible to obtain the conversion factor between intensities and concentrations. Therefore we have calculated the ratio of the true intensities for 1 and 4 which is the same as the concentration ratio [1]/ [4]. For medium not exposed to cells, the true intensity for 1 is obtained by subtracting the intensity for 4 from that of 1 because the second resonance of 4 lies beneath the resonance of 1. For medium exposed to cells, the changed pH shifts the second resonance of 4 so that it is seen directly and subtraction is not necessary.
If the conversion of 1 to 4 is fast relative to the rate of reaction of these compounds with substances present in culture media (amino acids, nutrients, etc.), this ratio is expected to decrease for the first few hours (the conversion of 1 to 4) and then remain constant, assuming that 1 and 4 react at the same rate with substances in culture media. In Fig. 5 we plot this ratio vs. time for 400 lM cisplatin in (a) culture medium that has not been exposed to cells, (b) in medium exposed to normal Jurkat cells, and (c) in medium exposed to resistant Jurkat cells. The initial decrease in the ratio and its constancy for later times is very clearly shown. Each set of data was fitted to A + Be ÀCt by minimizing the mean-square deviation, as shown. The value of the ratio at (a) Drug in medium not exposed to cells. (b) Drug in medium previously exposed to 3 Â 10 6 normal Jurkat cells for 1 h. (c) Drug in medium previously exposed to 3 Â 10 6 resistant Jurkat cells for 1 h. Spectra taken 2.6 h after addition of drug to the culture medium. Species are as indicated in Fig. 1. The stock solution of drug used in these experiments was equilibrated in 154 mM NaCl. t = 0 (just after the reaction of cisplatin with substances in the medium) is A + B. For culture medium which has not been exposed to cells, the value is 12.1 ± 0.3, for culture medium previously exposed to normal cells, 22.2 ± 1.0 and for medium previously exposed to resistant cells 27.8 ± 1.3. Since the ratio [1]/ [4] is greatest for resistant cells, they are more effective than normal cells in modifying 4 in the culture medium. While this same conclusion was reached in NMR experiments having the cells in contact with the drug (Fig. 3), Fig. 5 shows that the modifying substance is released by the cells and that the modification takes place in the medium. for normal cells, the slopes are $0 (not shown -the actual value is +1.0 ± 2.4 lM per million cells [13]) and À2.4 lM per million cells for species 1 and 4. This means that a single normal cell modifies 2.8 fmol of 4, and a single resistant Jurkat cell modifies 3.4 fmol of 1 and 7.9 fmol of 4. Cisplatin, 1, is more resistant to attack by the modifying substance in the medium than is 4, whether the substance is produced by normal or resistant cells.

Discussion
The conclusion that resistant cells can modify more 4 than normal cells is confirmed in the 400 lM experiments shown in Figs. 4 and 5. However, since the cells were no longer present when the NMR data were collected, it is clear that the platinum-modifying substance is released by the cells to the medium. It is also clear that this modifying reaction is rapid, occurring prior to the first NMR measurement, less than $16 min. Because more 4 than 1 is modified by the cell-released substance, the platinum species re-equilibrate at early times, i.e., 1 converts to 4, producing the exponential decay observed in Fig. 5. At later times, after the platinum species reach equilibrium with each other, the absolute intensities of both decrease slowly with time [11]. Since their ratio remains constant, they appear to react at equal rates with substances in the medium.
Although both the nature of the substance released by cells and the products formed are unknown, some explanations for the loss of the HSQC NMR signals can be ruled out. For example, while the platinum drugs readily react with thiols by ligand displacement, the initial products of these reactions, which have NH 3 trans to S À are easily detected by HSQC NMR [13,16,17]. The fact that no HSQC NMR active product peaks are observed at $À40 ppm (N), strongly suggests that attack by thiols, e.g. GSH, is not responsible for the modification of 1 and 4 in these studies. Another possibility for the loss of peak intensity is that the platinum compounds are binding to a macromolecule, which broadens NMR lines making  15 N-cisplatin in culture medium not exposed to cells, (b) 15 N-cisplatin in medium previously exposed to normal Jurkat cells, (c) 15 N-cisplatin in medium previously exposed to resistant Jurkat cells. Each data set is fitted to A + Be ÀCt by least-squares. signals undetectable. However, earlier studies by Sadler and coworkers showed that HSQC NMR peaks for 15 N cisplatin are easily observed when the drug is bound to the 66 kD protein, human serum albumin [18]. It seems unlikely that the cell releases a high molecular weight protein for the modification of these platinum compounds. Furthermore, the amount of platinum compounds taken up by, or bound to, cells [13] is much less than the decrease in peak intensity seen in the HSQC NMR. Most important, the modification of species 1 and 4 takes place in the absence of cells, so is an extracellular process. Since previous experiments [13] have shown that the modified species, which are HSQC NMR-silent, do not enter cells, this modification is a mechanism for drug resistance. Despite the fact that the molecular mechanisms of resistance to the platinum drugs have been studied for decades, inquiry in this area has largely focused in what happens after the drug enters the cell. However, recent studies suggest that tumors can recognize and reorganize their microenvironment to maximize their survival in the presence of anticancer drugs [18,19]. This ability, which has an important impact on the success of chemotherapy, could be the basis for an extracellular resistance mechanism that modifies toxins before they have the chance to penetrate the membrane and enter the cell.
In this report we use a sensitive NMR technique, HSQC NMR, to show that a substance released by Jurkat cells can modify cisplatin and its mono-and dicarbonato complexes in the culture medium. Since resistant Jurkat cells are able to modify more of the platinum compounds than normal cells, this appears to be an undiscovered resistance mechanism for cisplatin that operates outside the cell. In view of the fact that acquired resistance is one of the major limitations in platinum chemotherapy; this observation can have important consequences for the design of new platinum drugs for treating cancer. We hope that this report will stimulate interest in the role of carbonate in the molecular mechanism of action of the platinum drugs.