Combined effect of buthionine sulfoximine and cyclophosphamide upon murine tumours and bone marrow

Lowering Water's Octane Liquid carbon dioxide (CO2) is a powerful solvent used for purposes such as extracting the peanut flavor from peanuts and decaffeinating coffee. Now, two scientist-entrepreneurs in Berkeley, California, say that it may also be the best way to remove the possibly carcinogenic fuel additive methyl tertiary butyl ether (MTBE) from groundwater. Marc Sims, a chemical engineer, and his partner Jim Robinson, a molecular biologist, developed a device called PoroCrit that uses thin, microporous polypropylene tubes to expose the polluted liquid to pressurized liquid CO2. Originally, the device was designed to extract food flavorings. Then, says Sims, "We realized just how similar MTBE is to all the flavor compounds that we were extracting." PoroCrit works well on MTBE, says Sims, because the pollutant is about 100 times more soluble in CO2 than it is in water. The membranous tubes in the device create over 50 m2 of surface area through which the MTBE is drawn off through the micropores by the CO2. The end result is cleaner, slightly carbonated water. Other water pollutants such as gasoline, benzene, and chlorinated solvents, which are also highly soluble in carbon dioxide, may also be removed from water by the device. Originally introduced in 1979 as a way to boost the octane in gasoline, MTBE came into widespread use as a fuel additive because of its apparent ability to protect the public health by reducing automobile carbon monoxide emissions. In 1990, the Clean Air Act was amended to require the use of cleaner-burning fuels in areas with high carbon monoxide levels (those in nonattainment for National Ambient Air Quality Standards) in winter months. Oxygenated gasoline programs, including the use of MTBE, became the most popular means of meeting the new requirement. MTBE is currently found in about 25% of the gasoline used in the United States. In 1996, however, it was discovered that the additive had found its way into the groundwater in Santa Monica, California, prompting the city to shut down half its water supply wells. Other studies found traces of MTBE in 5% of the wells across the United States. Scientists suspect that in most cases the chemical is released into the environment by leaking fuel storage tanks and is washed into wells by rainwater, which readily dissolves the chemical. In December 1997, the EPA issued a health advisory alerting people to the possible danger of MTBE in water. The health effects of ingesting MTBE in the concentrations being found in drinking water are not known, but at high concentrations the chemical has been shown to cause cancer in animals. Even if the chemical does not pose a serious health risk, its strong taste and smell can seriously deteriorate the quality of the water in which it is found. Once MTBE gets into water, it becomes very difficult to remove. MTBE is extremely soluble in water-about 30 times more soluble than benzene-and very resistant to biodegradation. Because it does not readily adsorb to soil particles (unlike other fuel constituents), it tends to travel with groundwater plumes, as fast as the water travels. These characteristics ofMTBE seriously hamper the effectiveness of traditional groundwater remediation techniques on the pollutant. Granular activated carbon filters, for example, do not work at all. Up until now, the best remediation technology for

Cellular sulfhydryls, especially glutathione (GSH), have been found to protect cells via detoxification of alkylating agents such as nitrogen mustard (Arrick & Nathan, 1984). In fact, the relationship between the sensitivity of tumour cells to nitrogen mustard and their sulfhydryl contents has been demonstrated by several investigators (Hirono, 1961;Calcutt & Connors, 1963;Ball et al., 1966;Goldberg, 1969;Morita, 1973;Begleiter et al., 1983). Additionally, the increased sensitivity of cells to melphalan following either the incubation of cells in cysteine-deficient medium or exposure to a non-protein sulfhydryl (NPSH) depleting agent, such as diethyl maleate (DEM) or buthionie sulfoximine (BSO), has been reported by several groups (Suzukake et al., 1982;Taylor et al., 1982;Green et al., 1984).
Cyclophosphamide (CYC) is a widely used alkylating agent. According to the generally accepted mechanism for the generation of active metabolites, CYC becomes phosphoramide mustard, one of the nitrogen mustards, via oxidation to 4-hydroxycyclophosphamide in liver. Additionally, in other pathways for 4-hydroxycyclophosphamide, a formation of sulfhydryl reactive intermediates has been suggested (Ludlum, 1977;Cates & Li, 1982). Therefore, the pretreatment of tumours with an NPSH depleting agent may potentiate the response of tumours to CYC. In fact, it has been reported that an increase in the response of murine tumours to CYC was obtained by a single treatment of BSO (Tomoshefsky et al., 1985). The combination of BSO and CYC may have therapeutic benefit because the changes of GSH contents in Correspondence: K. Ono.
Received 1 April 1986; and in revised form, 30 June 1986. tissues following BSO treatments differ greatly from tissue to tissue (Griffith & Meister, 1979;Minchinton et al., 1984;Yu & Brown, 1984) and an optimal administration schedule of BSO may decrease the GSH content in tumours, but not in target organs of CYC.
One of the important roles of chemotherapy in cancer treatments is to eradicate micrometastases of tumours. In this study, we have examined the combined effect of multiple BSO pre-treatments and CYC on artificial lung micrometastases of murine tumours and bone marrow toxicity in mice.

Materials and methods
Mice and tumours Eight week old C3H/He male mice obtained from the animal centre of Kyoto University were used. They were caged in groups of 10 or less at a constant temperature with food and drinking water available ad libitum.
A fibrosarcoma, NFSa which arose spontaneously in a C3H mouse (Ando et al., 1979), were provided by Dr. K. Ando (National Institute of Radiological Science, Chiba, Japan). The 17th generation of tumour was cryopreserved. When required, the contents of an ampoule were thawed and transplanted into a number of recipient mice. The outgrowing tumours were excised and minced. Tumour cell suspensions were then prepared enzymatically (0.2% trypsin, 0.02% pancreatin).
After cell counting, _104 viable cells in 10 p1 were inoculated s.c. into both thighs of the new recipient mice. Tumour from this 19th transplant generation reached a volume of 500 mm3 14 days after inocu-() The Macmillan Press Ltd., 1986. lation. At this tumour volume, changes in NPSH content of tumours after BSO treatment were studied. For the study on the effect of cytotoxic treatments, artificial lung micrometastases were produced by injecting the viable tumour cells via the tail vein of the recipient mice. Ten days later, these micrometastases became macroscopic tumour nodules on the surface of the lung. In order to increase the formation rate of the tumour nodules in the lung, mice were treated with 150mg kg-1 CYC 24h previously and HIR (heavily irradiated, 100Gy) tumour cells were mixed with the viable cells (Ando et al., 1984). A total of 2 x 106 tumour cells in 0.5ml were injected. The number of tumour nodules was directly proportional to the number of the viable tumour cells injected, and tumour nodules were enumerated up to 100 per lung. The formation rate of the tumour nodules in the lung of the control was -5%. Therefore, to construct the dose survival curve of the lung micrometastases, the mixture of HIR and the appropriate number of viable tumour cells which were expected to produce about 50 nodules after CYC treatment with or without BSO were injected via the tail vein. In 2 to 5 separate experiments, 10 to 25 mice in total were used for each point.
Drug treatment BSO was prepared as described by Griffith and Meister (1979), and dissolved in sterile physiological saline (400mM). For measurement of NPSH contents in the tumours after BSO treatment, 5mmolkg-' BSO was injected s.c. into the posterior neck of mice 2 or 4 times every 12 h. Twelve hours after final injection of BSO, the tumours were excised and their NPSH contents were measured. For study of the combined effect of multiple BSO pretreatments and CYC toward lung micrometastases, 5 mmol kg-1 BSO was injected every 12 h in the manner described above. The BSO treatment was 48 h after i.v. injection of the tumour cells when BSO was administered twice every 12 h. When given four times, BSO was started 24 h after tumour cell injection. CYC was dissolved in physiological saline and was administered i.p. 12 h after the final BSO treatment, 72 h after tumour cell injection, in either BSO treatment schedule. To study bone marrow toxicity of the drugs, CYC was administered i.p. to non-tumour-bearing mice in combination with or without 4 injections of 5 mmol kg 1 BSO on the schedule described above.
Non-protein suithydryl (glutathione, cysteine) measurement This assay has been described in detail elsewhere (Komuro et al., 1985). Briefly, 10% homogenates of tumours were prepared in 0.02 M ethylene diamine tetra acetic acid-disodium (EDTA) on ice. Aliquots (2.5 ml) of the homogenates were diluted with 2 ml H20 and 0.5 ml 50% trichloroacetic acid. The mixture was shaken vigorously and kept at 0°C for 15 min. The suspensions were centrifuged at 2000 g and 4°C for 10 min. Two ml of the supernatant were added to 4 ml of 0.4 M Tris buffer (pH 8.9), and 0.1 ml of 0.01 M DTNB was added and shaken. After filtering through a 0.22 gm millipore filter (Millipore, Bedford, MA, USA) absorbance at 412 nm of the sample solution was recorded on a spectrophotometer (UVIDEC 610 Spectrophotometer, Jasco, Tokyo, Japan) for the measurement of total non-protein sulfhydryl (NPSH). The sample was subjected to high-performance liquid chromatography (Hitachi Model 655 Solvent Delivery System, Tokyo, Japan) for measurement of GSH and cysteine content.

Measurement of lung micrometastases and bone marrow response
Lungs were removed 10 days after treatment and fixed in Bouin's solution, and tumour nodules on the surface of the lungs were counted by eye. Surviving fractions of the lung micrometastases were obtained by dividing the formation rates of the tumour nodules at the CYC doses administered by that of the control.
Drug effects on normal bone marrow were determined by the spleen colony assay (Till & McCulloch, 1961). Femurs from both legs of mice without tumours were removed 24 h after CYC treatment, and the marrow was washed out with 1 ml of cold physiological saline. Bone marrow suspensions from 3 mice were combined, nucleated cells counted and an adequate number of cells were then injected i.v. into 7 mice which had received wholebody gamma-irradiation (9 Gy) 24 h previously. Spleens were removed 8 days later, fixed in Bouin's solution and the numbers of colonies counted.

Results
Non-protein sulfJhydryl (glutathione, cysteine) depletion by BSO Figure 1 shows the changes of NPSH, GSH, and cysteine contents in NFSa tumours after repeated administration of 5 mmol kg-' of BSO. NPSH content in untreated tumours was 2.590mmolkg-1 which involves 2.067 mmolkg-' of GSH and 0.301 mmol kg-' of cysteine. Two BSO treatments given at 12h intervals reduced the NPSH, GSH and cysteine contents of tumours to 36.9, 24.0 and 84.7% of the untreated levels, respectively, 12 h after the second BSO treatment. Following four BSO treatments, the contents of NPSH, GSH and cysteine in tumours decreased to 14.1, 1.78, and 42.9% of the untreated levels, respectively, 12 h after the fourth BSO treatment.
Effects on lung micrometastases of NFSa tumour The response of lung micrometastases which were treated with CYC 12 h after tumour cell injection was described by an exponential survival curve with a small shoulder ( Figure 2). When CYC treatment was 72 h after tumour cell injection, the dose survival curve was an exponential curve with an almost equal slope, but a large shoulder compared with the former (Figure 2). The cytotoxic effect of CYC on lung micrometastases was markedly enhanced when CYC was given to mice after two or four BSO treatments (Figure 3). Enhancement ratios (ERs) ranging from 2.41 (SF = 5%) to 2.73 (SF = 80%) were obtained when four BSO treatments were given before CYC administration. In the case of two BSO treatments, they ranged from 1.75 (SF=5%) to 1.83 (SF=40%). The average colony formation rate was 5.0% throughout these experiments and was not affected by BSO treatments.

CFU-S sensitivity studies
The results of dose response studies of CFU-S to CYC in combination with or without BSO treatments are presented in Figure 4.    (Ono et al., 1986). The repeated BSO treatment depleted the tumour GSH to very low levels, according to the number of BSO administrations (Figure 1), and did not exhibit any untoward effects such as weight loss or death of a \ mice. Therefore, repeated administration of BSO r' \alone on an optimal schedule appears to be a sufficient and safe method to deplete tumour GSH to very low levels. The survival curves of 12 and 72 h old lung micrometastases have almost the same slopes (Figure 2). This finding may indicate that tumour cells in the micrometastases of both age groups have similar sensitivities to CYC and similar proliferation status, since CYC has been known to --' ' ' ' ' ' exhibit smaller cytotoxicity towards noncycling 100 200 300 tumour cells compared with cycling cells (Begg et Dose of CYC (mg kg-1) al., 1985). The survival curve of 72h old micro-I The surviving fractions of CFU-S of bone metastases has a wider shoulder compared with cells treated with CYC alone (0) or in that of 12h old micrometastases (Figure 2). This ition with four pre-treatments of BSO may be due to the proliferation of cells in the lung kg-1 (Q). Vertical lines represent s.d. of two prior to treatment leading to a greater cell number experiments.
In our experiments, the yield of tumour nodules affect the cytotoxicity of CYC on bone in the lung was not affected by BSO treatment. CFU-S. Rojas et al. (1984) reported that BSO treatment did not inhibit the growth of tumours. However, Midander and Revesz (1984) reported that BSO at in high doeses and long exposure time showed cytotoxicity and growth inhibitory effects toward V79 nges in GSH contents in tumours of mice cells. Pronounced cytotoxic effects of prolonged O treatment have been examined by several exposure to high doses of BSO were observed in tors. Minchinton et al. (1984) reported other cell lines (Shrieve & Harris, 1986). In in vitro n GSH depletion down to 38% in the systems, BSO at low doses and short exposure time tumour and 57% in SAFA tumour, has not exhibited cytotoxicity . ely, 8 or 12 h after treatment with It has been reported that BSO administered to mice g-I of D, L,-BSO. Rojas et al. (1984) is completely excreted or metabolized within 24 h he same lines of tumours repeatedly with a (Griffith, 1982). Therefore, the discrepancy between se of 0.5 or 1.0mmolkg-1, but maximum the results may be caused by the different BSO pletion was to 37.8% of the control values. exposure times.
Brown (1984) have reported that GSH BSO increased the sensitivity of the lung microin EMT6 and KHT tumours treated with metastases of NFSa tumour cells to CYC (Figure 3). doses, higher than 0.33mmolkg-1, Further, this increase appears to correlate inversely d to minimum levels 6 to 8 h after the with the GSH level at the time of CYC treatment, it, and complete recovery in GSH contents because four BSO treatments decreased GSH ieved by 24h. They also reported that the content of NFSa tumours to a lower level ntents in five different tumour lines treated compared with two BSO treatments. Tomashefsky ily doses of 1 or 3 mmol kg-I of BSO et al. (1985) have reported a similar phenomenon, d, respectively, to 20-50%, and 19-40% of showing that the growth delay time of MBT-2 rol values. In order to get further GSH tumour, a transplantable murine bladder tumour, i they combined DEM (300mgkg-1) with after CYC treatment correlated inversely with GSH iulting in a decrease to 8% of the control contents. Somfai-Relle et al. (1984) have also our studies, 5 mmol kg-I of D, L,-BSO was reported that the sensitivity of L1210 leukaemia cells to L-PAM becomes higher with longer BSO treatments (i.e., at lower GSH levels). In our studies, the ERs of BSO treatments were large at high survival levels (i.e., at small CYC doses). This finding may be applicable to cancer therapy, because small doses of CYC are administered repeatedly in clinical cancer chemotherapy.
The sensitivity of bone marrow cells to CYC did not increase after BSO treatment (Figure 4). Moreover, BSO did not affect the yield of nucleated cells per femur (data not presented). It is not possible by available methods to measure the GSH content of bone marrow stem cells. However, GSH content in whole bone marrow cells decreased to minimum levels and completely recovered, 4 and 12h respectively after the treatment with 5 mmol kg-1 of D, L,-BSO (Komuro et al., unpublished). Therefore, it seems probable that the treatment with BSO at an interval of 12 h did not decrease the GSH content in bone marrow stem cells to levels low enough to increase the sensitivity to CYC.
In conclusion, our results suggest that multiple BSO treatment on an appropriate administration schedule can increase the sensitivity of tumours to CYC without changing that of bone marrow stem cells. However, more toxicity studies on other critical normal tissues (e.g. gut) are needed before clinical application can be considered.