Studies with 1,2-dithiole-3-thione as a chemoprotector of hydroquinone-induced toxicity to DBA/2-derived bone marrow stromal cells.

Stromal cells from DBA/2 mouse bone marrow have been shown to be susceptible to cytotoxicity induced by several redox-active metabolites of benzene, including hydroquinone (HQ). Treatment with HQ also alters the composition of stromal cell populations by preferentially killing stromal macrophages compared to stromal fibroblasts. This cytotoxicity can be prevented by 1,2-dithiole-3-thione (DTT) as a result of the induction of quinone reductase (QR), a quinone-processing enzyme, and glutathione. The inductive activities of DTT protected stromal cells against HQ-induced cytotoxicity and against HQ-induced impairment of stromal cell ability to support myelopoiesis. In vivo feeding of DTT to DBA/2 mice increased QR activity within the bone marrow compartment and protected bone marrow stromal cells isolated from the DTT-fed animals from ex vivo HQ challenge. Thus, the inducibility of cellular defense mechanisms and xenobiotic-processing enzymes by chemoprotective agents such as DTT may be a useful strategy for protecting against chemically induced bone marrow toxicities.

The purposes of the current study were to determine if DTT protects DBA/2derived stromal elements against the cytotoxic effects of HQ and if DTT also protects stromal cell functional activity by supporting myelopoiesis (Fig. 1). Additionally, in vivo induction of QR activity within the bone marrow compartment was studied by feeding DTT to mice.  (23) and pooled in either PBS or DMEM (no additions). The procedure used to establish primary adherent stromal cell cultures was a modification of the method of Zipori and Bol (24). For adherent cell culture, we pooled cell suspensions from two or more animals and diluted them with DMEM supplemented with 15% FBS, 2 mM glutamine, 50 jiM 2-mercaptoethanol, 100 ,ug/ml L-asparagin, 100 IU/ml penicillin, and 100 jig/ml streptomycin, and then plated the suspensions on 100-mm tissue culture dishes, six plates per two animals. Twenty-four hours later, we gently aspirated media and replaced it. At 48 hr we washed cultures twice with PBS to remove unattached cells and debris. Cultures were maintained in 100-mm tissue culture dishes, changing media every 5-6 days, until they were used in microtiter plate and other assays 10-18 days after isolation. For all toxicity assays, we trypsinized cells and replated them in 96-well microtiter plates at a concentration of 3 x 104 cells/well. Chemical Protection against Hydroquinone-induced Toxicity in Primary Stromal Cells. We plated cells into 96well plates at 3 x I04 cells/well and treated them with 75 jIM DTT 24 hr later. After another 24 hr, we replaced the chemoprotector with HQ, and 24 hr later (a total of 72 hr after plating), we assayed cells for survival with crystal violet staining as previously described (11). Peak induction of QR activity occurs between 16 and 24 hr after a single dose of DTT, and it has been determined that 75 jM DTT induces essentially the maximal QR activities observed in these cells after DTT treatment (12). We report survival as percent survival of treated: control groups, with control cells representing 100% [(Abs610 treated cells/Abs6l0 control cells) x100].
In studies to determine if DTT could protect against HQ inhibition of stromal cell support of myelopoiesis, we pretreated primary cultures for 24 hr with 75 pM DTT, removed DTT and replaced it with 20 piM HQ, and conditioned and concentrated the medium for CFU-G/M (colony-forming units of granulocytes and monocytes) assays as described below.

Assessment of Preferential KiUling by
Hydroquinone of Macrophages versus Fibroblasts in DBA12-derived Primary Stromal Cels. Six replicate TC 60 plates per n (n refers to one discreet experiment/observation) were seeded with 2 x 106 2-week-old stromal cells. One 96-well microtiter plate per n was also seeded with 3 x 104 cells/well. We treated three plates/n and three rows of wells/n 24 hr later (day 2) with 75 jM DTT. On day 3, we treated one plate or microtiter plate row with 0, 35, or 50 jiM hydroquinone (these doses of HQ are the LC25 and LC50 doses for DBA/2 primary stromal cells).
Finally, on day 4, 72 hr after plating, we collected cells in the TC60 plates by scraping and centrifugation and stained them for esterase activity using two different substrates, ct-naphthylacetate (nonspecific) and cx-naphthylbutyrate (Sigma diagnostic kits 90-Al and 181-B). We assessed population differentials by determining the percent esterase-positive cells (macrophages) using standard light microscopy. Cells in the microtiter plates were stained with 0.4% crystal violet to determine survival (11).
Assay of Dicoumarol-Inhibitable Quinone Reductase Activity. Quinone reductase (QR) activity was assayed by modification of the microtiter plate procedure developed by Prochaska and Santamaria (25). In this procedure, QR activity is assessed by measuring the dicoumarol-inhibitable NADPH-dependent menadiol-mediated reduction of 3-(4,5dimethylthiazo-2-yl)-2,5-diphenyltetrazolium bromide (MTT) to the blue formazan dye. For this assay we trypsinized stromal cells, spun them at 1000 rpm for 10 min, discarded the supernatant, and lysed the cells by incubation at 370C for 10 min with 0.8% digitonin in 2 mM EDTA, pH 7.8. We assessed dicoumarol-inhibitable QR activity by adding the MTT reaction mixture, described below, to digitoninlysed stromal cells in the presence and absence of 60 gv dicoumarol. We recorded the spectrophotometric kinetic measurement of change in absorbance/min at 610 nm and calculated QR activity using the extinction coefficient for MTT (11,300 M'1 cm' ) and expressed this value as nmol MTT reduced/min/mg protein.
Protein was quantified by a modified Lowry assay (26) or by the Bio-Rad protein assay (cat. no. 500-0006), which is based on the Bradford assay (27). The MTT reaction mixture consisted of 25 mM Tris-Cl, 0.67 mg/ml bovine serum albumin, 0.01% Tween-20, 5 gM flavin adenine dinucleotide, 1 mM glucose-6phosphate, 30  penicillin, and 100 pg/ml streptomycin. We then added cells to the agar mixture in no more than 100-p1 volumes to produce nucleated cell concentrations ranging from 4 to 20 x 10 cells/ml. Next we added 1 ml of the cell-agar suspension to gridded TC35 dishes containing 100 p1 of medium containing colony-stimulating factor. The colony-stimulating activity (CSA) was derived from medium that had been conditioned (12 days) by primary bone marrow stromal cells in the presence or absence of hydroquinone and/or DTT, as described below. After plating the cell-agar mixture, we allowed the agar to solidify at room temperature and incubated cultures at 370C in a humidified, 5% CO2 atmosphere for 8-12 days. Colonies consisting of at least 50 cells were scored from four replicate plates per test group using an inverted phase-contrast microscope. Hydroquinone Inhibition of Stromal Cell Ability to Support Myelopoiesis in Soft Agar. The effect of noncytotoxic doses of hydroquinone on the ability of DBA/2-derived stromal cells to support myelopoiesis was evaluated by treating 10-14-day cultures with 15 IM HQ and allowing the treated cells to condition the media for 12 days. We then collected this conditioned media and concentrated it for use in the CFU-G/M assay described above. Because the HQ was not removed during the media-conditioning period, we also determined CFU-G/Ms with conditioned media from untreated cells and added 20 IM HQ at the time of the colony-forming assay to control for any effect that residual HQ may have had on myelopoiesis (data not shown). Stromal cell conditioned media was concentrated approximately 10-fold in CentriCells (Polysciences, Inc.) with a 30,000 molecular weight cut-off point. We centrifuged the CentriCells at 2000g in a swingingbucket rotor for 30 min. After centrifugation, we determined the volume of the least concentrated sample and added DMEM with all additives to the other samples to bring them up to this volume so that all samples had an equal concentration. Individual CentriCells were used a maximum of 3 times, and concentrated volumes on any one run varied by ± 10%.
In Vivo Feeding of DTT. We acclimated DBA/2 mice (9-10-week-old males) and maintained them on an antioxidantfree powdered diet (AIN-76 purified diet with additional menadione and without ethoxyquin, TEKLAD DIETS) for 1 week. The test group was switched to diet containing 0.1% DTT, a concentration which was well tolerated by the animals based on appearance, weight, and activity levels.
After 6 days on test or control diet, animals were humanely euthanized by cervical dis-location, and bone marrow and livers were removed for enzyme analysis. We determined QR activity in whole bone marrow and in 24-hr primary cultures of bone marrow stromal cells. Livers were used as a positive control for QR and glutathione-Stransferase(GST) induction by DTT (data not shown).
In experiments to determine if in vivo feeding of 0.1% DTT could protect primary bone marrow cells against ex vivo challenge by hydroquinone, we treated animals as described immediately above and flushed bone marrow cells from the femurs using two mice per n. Equal numbers of nucleated cells were then plated into four TC1 00 dishes per n in the presence (two plates) or absence (two plates) of 50 P.M HQ, the LC50 dose for DBA/2 primary stromal cells. By this protocol each n had two test plates and two control plates to serve as its own control. Twenty-four hours later, we washed plates four times with PBS to remove dead and unattached cells. We then collected surviving, attached cells by scraping the plates with a rubber scraper and assessed survival by counting the number of surviving cells using a Coulter counter. Data are expressed as percent of survival for controls.
Statistics. Computations and statistics were performed using Lotus and Statpak software on an IBM personal computer. We used Students t-test (two-tailed) and one-way analysis of variance (ANOVA); values were considered significantly different ifp < 0.05 or if the F ratio had a significance < 0.05.

Results
Based on more extensive dose-response studies, the LC50 for hydroquinone was determined to be 49 ± 6 P.M for 3 x 104 stromal cells/well for DBA/2 (11). Twenty-four hours of pretreatment of stromal cells with 75 ,uM DTT protected against HQ-induced cytotoxicity, even at 95 pM HQ, a concentration that killed all cells in non-DTT-treated controls (12).
When DBA/2-derived stromal cells were exposed to LC25 and LC50 concentrations of HQ and the proportion of macrophages versus fibroblasts to untreated cells were compared, the proportion of macrophages in the HQ-treated cells dropped from approximately 60% to about 40%, indicating that the macrophages were more sensitive to HQ than the fibroblasts (Table  1). This is consistent with results obtained by Thomas and co-workers with cells derived from B6C3F1 mice (6). At the LC25 of HQ, at which 25% of the cells are killed, a 20% decrease in the macrophage population means that virtually all the cells that were killed were macrophages. In the same experiment, another group of cells was pretreated for 24 hr with 75 pM DTT before exposure to HQ, and the percentage of macrophages versus fibroblasts was determined. DTT completely protected the cells from the cytotoxic effects of HQ, and the proportion of macrophages to fibroblasts was identical to control populations ( Table 1). As previously reported, QR activity approximately doubles, and GSH concentration increases by about one-third after pretreatment with 75 pM DTT (12). The increases in both GSH concentration and QR activity would therefore appear to underlie the chemoprotective actions of DTT. Mechanistically, GSH would interact directly with the electrophilic benzoquinone, whereas increased QR activity would reduce benzoquinone to the less chemically reactive hydroquinone. The studies described thus far used only cell death as the endpoint for toxicity to HQ. Because the stromal element of the bone marrow is involved in the regulation and maintenance of hematopoiesis, chemical toxicity could also be expressed by impairment of regulatory function without causing cell death. This form of HQ-induced toxicity has recently been reported in long-term stromal cultures derived from the B6C3F1 mouse (28). Consequently, we were interested to see if DTT could protect primary stromal cells from HQ-induced toxicity using this functional endpoint as an assessment of toxicity. Stromal cell support of myelopoiesis can be assayed in two ways: by plating cells in semisolid agar directly onto adhered stromal cells or by plating cells in agar onto media that has been conditioned by stromal cells. Conditioned media has to be   concentrated to express enough colonystimulating activity to support myelopoiesis. We used the second method to obtain the stromal cell-derived colonystimulating factors necessary for the support of myelopoiesis because conditioned media has the additional advantage of allowing us to treat the stromal cells without concurrently treating the naive bone marrow cells used in the colony-forming assay.
The data presented in Table 2  if the in vivo administration of DTT could induce QR activity in the bone marrow, and if so, could in vivo feeding of DTT protect bone marrow stromal cells from ex vivo challenge with HQ. It was necessary to remove the bone marrow cells from the animals for HQ challenge rather than challenge the animals in vivo to demonstrate definitively that relevant biochemical changes had occurred in the target organ. Demonstration of DTT protection against in vivo HQ or benzene challenge would require extensive pharmacokinetic analysis to establish if protection resides solely at the level of the bone marrow as opposed to (or in addition to) other sites such as the liver, as has been shown with oral benzo-[a]pyrene-induced bone marrow toxicity (29). In vivo feeding of DTT does induce QR activity in rat liver (30).
DBA/2 mice were fed 0.1% DTT mixed in the diet for 6 days. This dose of DTT was well tolerated as assessed by weight gain, activity levels, and overall appearance, which were comparable to control animals. At the end of the test period, we euthanized animals and cul-tured bone marrow cells for assessment of QR activity or treatment. Livers were also removed, perfused, and immediately frozen in liquid nitrogen to be used as a positive control, and 2-to 3-fold inductions of QR activity were observed (data not shown). Induction of hepatic QR activity was expected due to the fact that in vivo induction by a number of phase II inducers has been demonstrated in the DBA/2 mouse (22). As shown in Table 3, there was an increase in QR activity in endogenous bone marrow preparations as well as in 24hr stromal cultures. DTT feeding increased QR activity in whole bone marrow preparations from 17 to 22 nmol/min/mg protein, and from 30 to 46 nmol/min/mg protein in 24-hr stromal cells (Table 3). In keeping with the induction of QR activity within the bone marrow compartment, in vivo feeding of DTT protected primary stromal cells against a subsequent in vitro challenge with 50 jiM HQ during the first 24 hr in culture (Table 4).

Discussion
The concept of chemoprotection, that is, protection from the toxicity of one chemical by the administration of another chemical, was first observed more than 50 years ago in rodents when skin tumors were produced by local applicatlion of carcinogens (17,18). Mechanistically, chemoprotection has been associated with the induction of three classes of detoxification enzymes: enzymes involved in ascorbic acid biosynthesis, the microsomal P450 class of enzymes, and soluble, cytosolic enzymes known collectively as phase II detoxification enzymes (17,31). Extensive research since that time has demonstrated that induction of phase II enzymes as well as Significantly different from respective control by Student's t-test, p<0.05. glutathione is a useful strategy for enhancing the clearance and/or detoxification of chemically reactive intermediates (14,18,32). Quinone reductase (QR) has also been shown to be coordinately induced with other electrophile-processing phase II enzymes such as glutathione-S-transferases (14,17,18). Recently, the protection of hepatoma cells against the toxicity of a number of redox-active xenobiotics has been linked to the in vitro induction of QR (19). A class of chemicals currently under extensive investigation as chemoprotective agents are the dithiolethiones (33). Bone marrow is a target organ for toxicities induced by a spectrum of chemicals (34) including the environmental pollutants benzene and benzo[a]pyrene (BaP) (35). The stromal cell component from bone marrow is particularly susceptible to toxicity induced by several redox-active metabolites of the hematotoxin benzene, such as benzoquinone and HQ (12,28,36). Consequently, it was of great interest when recent studies in our laboratory demonstrated that bone marrow-derived stromal cells could also be protected against hydroquinone-induced cytotoxicity by pretreating cells with the phase II enzyme inducer DTT (12). This observation prompted us to further examine whether the inducing activity of DTT in bone marrow stroma also protects against HQ-induced modulation of stromal-dependent myelopoiesis. In this study, we have also examined whether in vivo feeding of DTT induces QR activity within the stromal compartment and as such protects against the in vitro toxicity of HQ.
As shown in Table 1, HQ was toxic to DBA/2-derived bone marrow stroma and demonstrated preferential killing of stromal macrophages. Primary cultures of bone marrow stroma are not a pure population of one type of cell; these cultures consist of a 60:40 mixture of resident macrophages and fibroblastoid cells, respectively (24,37). Primary stromal macrophages were more sensitive to HQ than the fibroblastoid stromal cells, as illustrated by the shift in the ratio of macrophages to fibroblasts from 60:40 to 40:60 after LC25 or LC50 doses of HQ ( Table 1). Pretreatment of the stromal macrophages and fibroblasts with DTT prevented this shift in ratio ( Table 1).
The data described above complement previous studies performed in our laboratory comparing QR and GST activities and GSH concentration between cell types within the bone marrow stroma and between strains of mice with differential susceptibility to HQ-induced cytotoxicity.
These previous studies demonstrated that basal QR activity was lower in the stromal macrophage versus the stromal fibroblast, with activity for the mixed population falling between the values for the individual cell types (12). The lower QR activity in the more HQ-sensitive cell type mimicked the difference in QR activity previously observed between whole stromal populations from strains of mice with differential sensitivity to HQ-induced toxicity (12). These data also agree with previous data evaluating the effect of DTT on stromal cell QR and GST activities and cytosolic GSH concentration. Treatment of primary cells in vitro with 75 jM DTT resulted in a 2-fold increase in QR activity in DBA/2 stroma and about a one-third increase in cytosolic GSH concentration (12). Unlike other tissues (22,32), DTT treatment of primary bone marrow stromal cells had no inductive effect on GST activity. Thus, the increase of both GSH concentration and QR activity appear to be important in the chemoprotection against HQ-induced toxicity provided by DTT. This hypothesis was substantiated by the effects of dicoumarol, an inhibitor of QR activity (31), on HQ toxicity and through protection by DTT. Dicoumarol potentiated HQ toxicity and interfered with DTT protection against HQ-induced toxicity in stromal cells (12,38). Likewise, depletion of GSH in stromal cells by buthionine sulfoximine potentiated HQ-induced toxicity (39).
Because enhanced stromal cell survival was used as an index of chemoprotection from HQ-induced toxicity, we thought it was important to also examine a noncytotoxic functional endpoint of toxicity induced by HQ. A potential target for nonlethal toxicity is the ability of stromal cells to support myelopoiesis, which is an in vivo function of the stromal cells. It has been previously demonstrated that purified populations of bone marrow stromal fibroblasts can support myelopoiesis to only 50% of the number of colonies that are supported by mixed populations of stromal fibroblasts and stromal macrophages.
Purified stromal macrophages cannot support significant myelopoiesis (28). The data presented in Table 2 demonstrate that noncytotoxic concentrations of HQ impaired the ability of stromal cells to support myelo-poiesis by approximately 40%. This indicated again that the stromal macrophage was the preferential target of noncytotoxic concentrations of HQ, with the amount of observed myelopoiesis being supported primarily by stromal fibroblasts. This is in agreement with the previous data demonstrating preferential killing of stromal macrophages over stromal fibroblasts by HQ ( Table 1). Pretreatment of cells by DTT protected the primary stromal cells from both forms of HQ-induced toxicity (Tables 1 and 2).
Chemoprotection of primary bone marrow stromal cells by DTT was also observed after in vivo feeding of DTT. Using QR as a biomarker of the inducing effect of DTT within the bone marrow, a 1.5-fold increase in QR activity was observed in 24-hr stromal cultures derived from animals that had received 0.1% DTT in the diet for 6 days (Table 3). This inductive activity of DTT was reflected in protection against the cytotoxic effects of 50 FM HQ in stromal cells derived from DTT-treated animals compared to stromal cells derived from animals that did not receive DTTin vivo (Table 4). More importantly, based on the data presented in Table 2, a protective effect against HQinduced alterations in myelopoiesis would also be expected. Thus, the inducibility of cellular defense mechanisms and xenobiotic-processing enzymes by chemoprotective agents such as DTT may prove to be a useful strategy in preventing chemically induced cell dysfunction and death in the bone marrow (Fig. 1). In this regard, one of the advantages of an agent like DTT is that it is a monofunctional inducer and as such does not require interaction with the Ah receptor for its inducing activity (22). The availability of chemicals that function as inducers primarily of phase II enzymes independent of the Ah receptor is important from the perspective that genetic differences in the Ah locus are relevant to humans as well as to different strains of mice (21