Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

In mammalian cells, glutathione (GSH) is the major source of available nucleophilic thiol equivalents. Thiol homeostasis is carefully maintained, and thiol:disulfide exchange reactions are important to the functional status of many proteins (Thomas and Sies, 1992). Catalytic conjugations of glutathione to low molecular weight acceptors frequently involve glutathione S-transferases (GSTs) (Boyland and Chasseaud, 1969). It is now clear that the GST gene family is extensive, with distinct isozymes expressing different functional properties. The family is characterized by a promiscuous substrate specificity with low "catalytic efficiency", characteristics integral to the evolution of GSTs as detoxifiers of a broad spectrum of endogenous and environmental chemicals. GSTP1-1 is the most prevalent isozyme in nonhepatic tissues, and increased expression of GSTP1-1 has been extensively linked to drug resistance and the malignant phenotype of many solid tumors (Tew, 1994). Although GST catalysis will result in glutathione S-conjugate formation of drugs with an electrophilic center, there are numerous examples of GSTP1-1 overexpressing drug-resistant cell lines where the selecting drug is not a substrate and no conjugate is formed (Tew, 1994). In addition, a functional link between GSTP1-1 overexpression and the transformed phenotype remains elusive. To this end, our recent description of the protein:protein interactions between GSTP1-1 and c-Jun NH₂-terminal kinase (JNK) (Adler et al., 1999) has provided a framework for contemplating a role for this protein in regulation of stress response, proliferation, and apoptosis (Davis et al., 2001). In addition, a novel GSTP1-1 inhibitor, TLK199 [-glutamyl-S-(benzyl)cysteinyl-R-phenyl glycine diethyl ester], was synthesized (Lyttle et al., 1994) with the goal of potentiating the efficacy of anticancer drugs in tumors with a GSTP1-1 overexpressing resistant phenotype (Morgan et al., 1996). The drug is a peptidomimetic of GSH, esterified to enhance cellular uptake and designed to bind to the "G-site" of GSTP1-1. The inhibition constant (K_i) for GSTP1-1 catalytic activity (chlorodinitrobenzene as substrate) is 400 nM. This demonstrates significant specificity for the -family, since the K_i for the GST and -µ families range from approximately 20 to 75 µM. Independent of catalysis inhibition, TLK199 also disrupts the protein:protein interaction site(s) between GSTP1-1 and JNK1 (Adler et al., 1999; Wang et al., 2001). Interference with the ligand binding properties of GSTP1-1 may prove to be a critical pharmacological property for the drug in mediating its myeloproliferative effects.

Many drugs produce reactive oxygen species (ROS) as direct or indirect by-products of their metabolism. These can change redox conditions and trigger cellular responses through a number of different pathways. The nature and extent of the ROS insult can determine the threshold of the cellular response, manifest as proliferation, stress response and damage repair, or apoptosis (Adler et al., 1999). How these optional pathways are regulated is not known. However, links between thiol active molecules such as GST and stress-activated protein kinases such as JNK and thioredoxin and apoptosis signaling kinase have been established (Saitoh et al., 1998; Adler et al., 1999). In an unstressed cellular environment, JNK and apoptosis signaling kinase are kept in an inactive mode by the presence of these ligand-binding proteins. Under conditions of oxidative stress or in the presence of inhibitors, GSTP1-1 dissociates from JNK (Adler et al., 1999), activating the catalytic kinase activity and phosphorylating c-Jun. This process can activate kinase cascades involving the numerous sequential downstream kinases. These results are evidence that GSTP1-1 can have a nonenzymatic regulatory role in controlling cellular response to external stimuli (Davis et al., 2001).

There are indications that GSH and associated enzymes play a role in control of cellular immunity. For example, GSH levels in antigen-presenting cells determine whether a Th1 or Th2 pattern of response predominates (Peterson et al., 1998). In human immunodeficiency virus patients, GSH levels in antigen-presenting cells may be important in determining disease progression (Herzenberg et al., 1997). By extending the GST:JNK link, recent results have shown that JNK expression may affect immune response/myeloproliferation. T cells from a JNK1 knockout mouse hyperproliferated, exhibited decreased activation, and induced cell death and preferentially differentiated into Th2 cells (Dong et al., 1998). Despite the redundancy accorded to this system by the expression of JNK2, it appears reasonable to conclude that JNK1 signaling pathways may play a role in T-cell receptor initiated T-cell proliferation and differentiation.

In this present study, we provide evidence from pharmacologic and genetic approaches that GSTP1-1 contributes to control of cell proliferation through a noncatalytic, ligand binding activity with signaling kinases. These results imply that drug targeting of the GST isozyme can be used as a therapeutic approach to stimulate myeloproliferation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Lines. Human myeloid leukemic cells (HL60) were chronically exposed to TLK199, and drug-resistant cells were selected by passage. Cells were initially exposed to 2.5 µM (1/10th IC₅₀), and the drug concentration was increased in increments of 5 µM after 3 to 4 passages. Cells resistant to 50 µM TLK199 were cloned from single cells and selected for investigation.

Mouse Embryo Fibroblasts (MEFs). The development of GST^/ mouse was described earlier (Henderson et al., 1998). Removal of the murine GST gene cluster completely abolished the coding sequences of P1, leaving 5 exons for the coding region of P2. We established MEF cell lines from GST^+/+ and GST^/ mice. Timed pregnant mice were euthanized by cervical dislocation and the uterus aseptically removed for dissection of the embryos. Embryos were harvested at day 11. Tissue was finely chopped, rinsed, and the pieces seeded onto the culture surface in a medium containing serum. Cultures were kept at 37°C for 18 to 24 h. Once the tissue pieces adhered, the medium was replaced until a substantial outgrowth of cells was observed, at which point the cells were passaged.

Transfection of GSTP1-1. The influence of GSTP1-1 on proliferative rate was considered by implementing retroviral gene transfer into MEF cells from GST^/ mice. Briefly, stable lipofectamine-mediated HEK293 packaging cell transfectants were made with a full-length cDNA for GSTP1-1 in the retroviral expression vector pLPCX containing a puromycin resistance selection marker. Several clones were selected and grown in drug-free medium for 48 h. High viral titer supernatants were recovered from the packaging cells, filtered, and added to the MEF cells. Efficiency of transfection was assessed by cotransfection of a green fluorescent protein vector into parallel cultures and by immunostaining. The average transfection efficiency was 65% (proportion of cells), and the transfected cells expressed levels of GST equivalent to approximately 50% of the wild type (WT) MEF cells. Whole cell extracts were used for verification of expression patterns for the transfected cDNAs in the appropriate cell lines. Blots were probed with antisera against actin (Amersham, Piscataway, NJ) to verify equivalent loading.

Cell Doubling Experiments. A master mix was prepared containing the appropriate number of cells to aliquot 7 flasks of 100,000 cells per 10 ml of RPMI media. Cell doubling rates for GST^+/+ and GST^/ cells were determined by counting cells every 24 h for 7 days using a Coulter counter (Beckman Coulter, Fullerton, CA). Final counts were averaged from at least five experiments.

Drugs, Antibodies, and Enzyme Assays. The GSTP1-1 inhibitor TLK199 (Fig. 1) was obtained from Telik (San Francisco, CA). Antibodies to GST isozymes were obtained from Biotrin (Dublin, Ireland). Anti-active phospho-JNK antibody was obtained from Promega (Madison, WI), and the anti-JNK1 (clone C-17), the anti-JNK2 (clone N-18), the anti-ERK2 (clone C14-G), and the anti-p-ERK (clone E-4) antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA).

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Structure of TLK199, a glutathione peptidomimetic and specific inhibitor of GST. De-esterification releases the active form of the drug, TLK117.

Protein Analysis. Cells for UV irradiation were exposed to UV-C (60 J/m²) in a minimum amount of phosphate-buffered saline. Protein concentrations were determined using the Bradford reagent (Bio-Rad Laboratories, Hercules, CA). Whole cell extracts (100 µg) were separated on 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by transfer overnight at 30 V onto a polyvinylidene difluoride membrane. Blots were blocked in 10% nonfat dried milk in TBS-T (50 mM Tris, pH 7.5, 2.3% NaCl, 0.2% Tween 20) and washed between steps in TBS-T. All subsequent antibody incubations were conducted in blocking solution. Proteins were detected using a chemiluminescence detection kit (NEN Life Science Products, Boston, MA). Band intensity was analyzed using NIH Image software (version 1.57).

Apoptosis. Cells were exposed to either TLK199 (0-50 µM), UV-C (60 J/m²), or a combination of the two. Following an incubation period of 8 h or 24 h at 37°C and 5% CO₂, cells were washed in phosphate-buffered saline, resuspended in ethidium bromide (75 µM), and subjected to fluorescence-activated cell sorter analysis using a Beckman FACSCAN (Beckman Coulter, Fullerton, CA). The percentage of apoptotic cells was calculated manually using the Mac Cycle AV program (Phoenix Flow System, San Diego, CA) and statistical significance determined by Student's unpaired t test.

Animal Studies. Mice (129 GST^+/+ and GST^/) were maintained in a barrier facility with ad libitum access to food and water. For all bone marrow experiments, mice of 6 to 12 weeks of age were sacrificed with CO₂, the pelt clipped, and hind limbs exposed. Both femurs were removed and cells collected in Iscove's modified Dulbecco's culture media containing 2% fetal bovine serum. Red blood cells were lysed with 0.165 M NH₄Cl. Cells (1 ml) were plated (5 × 10⁴ cells/ml) in Methocult M3434 methylcellulose medium (Stemcell Technologies, Vancouver, BC) into 35-mm culture dishes, treated with TLK199 (0.2% dimethyl sulfoxide in H₂O) or vehicle, and incubated at 37°C and 5% CO₂. In vivo, mice (GST^+/+ and GST^/) were treated with TLK199 (50 or 75 mg/kg; 0.2% dimethyl sulfoxide in water) or vehicle control. Seventy-two hours postinjection, mice were sacrificed with CO₂ and cervical dislocation and the bone marrow removed as described previously. Cells were plated (5 × 10⁴ cells/ml) in Methocult M3434. Colony-forming units (CFUs) were scored (colony >50 cells) 7 to 10 days after plating and expressed as a percentage of control. Statistical significance was determined by Student's paired t test.

Detection of Superoxide Anion Levels. Superoxide anions were detected using the LumiMax Superoxide detection kit (Stratagene, La Jolla, CA). Briefly, cells were collected from bone marrow (129 GST^+/+ or GST^/ mice) and resuspended in superoxide assay medium (5 × 10⁶ cells/ml). Cells (100 µl) were added to 100 µl of reagent mix containing 200 µM luminol and 250 µM enhancer, ±- phorbol-12-myristate-13-acetate (PMA, 0.4 µM; Sigma, St. Louis, MO) and/or TLK199 (0-100 µM) and assayed for superoxide anions using a luminometer over 60 min. Statistical significance was determined by Student's unpaired t test.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

GST Status and Cell Proliferation. MEF cultures from GST^/ and GST^+/+ (11-day-old embryos) demonstrated consistently distinct growth patterns. GST^/ cells doubled faster than GST^+/+ cells (26.2 h versus 33.6 h, respectively; Table 1) and reached confluence at higher cell density. These values were calculated from at least five separate MEF cultures and were consistent in cultures established from 13- or 15-day-old embryos. Introduction of GSTP1-1 cDNA into the GST^/ MEF cells through retroviral transfection extended the doubling time to 30.4 h. Transfection efficiency was estimated at ~65% with ~50% GST protein levels, compared with the GST^+/+ cells (data not shown). These results are consistent with the rationale that GSTP1-1 plays a direct role in control of proliferation. Selection of immortalized cell cultures was achieved by continued growth of cultures beyond 40 passages. These cells had doubling times in the 12- to 14-h range, but there was no difference between GST^+/+ and GST^/ cells (Table 1).

The basal (unstimulated) white blood cell counts in GST^/ mice were found to be significantly higher than in GST^+/+ animals (Table 2). In vivo treatment of GST^+/+ mice with TLK199 (75 mg/kg i.p.) resulted in a significant 2-fold increase in white blood cells 3 days after treatment, due to an increase in circulating lymphocytes (Table 2). Analysis of blood from GST^/ mice showed no increase in white blood cell count above control following TLK199 treatment. Whereas basal proliferation for the GST^/ cells was the highest of all groups, this result implies that the target of TLK199 may be required for the pharmacological myeloproliferative effects to occur.

View this table: [in this window] [in a new window]   TABLE 2 Peripheral white blood cell counts in GST+/+ and GST/ mice treated with TLK199

Direct effects of TLK199 on mouse bone marrow progenitor cells were demonstrated either by in vivo or ex vivo treatment. Drug treatment induced a proliferative response (approximately 2-fold above vehicle control) in cells (or animals) wild type for GST (Table 3). Bone marrow cells null for GST did not respond to TLK199 treatment. TLK199 enhanced the number of cells from the granulocyte-macrophage (GM) lineage with an increased number of GM-CFUs detected either by in vivo or ex vivo experiments.

The Effect of TLK199 on Superoxide Production. To consider a possible role for ROS in the myeloproliferative effects of TLK199, bone marrow cells from GST^+/+ and GST^/ animals were treated with combinations of PMA and TLK199. As a single agent, TLK199 did not stimulate superoxide anion production above control levels. In contrast, TLK199 inhibited the PMA-initiated production of superoxide anions in a dose-dependent fashion (Fig. 2, a and b). For example, 50 µM TLK199 caused a significant (P < 0.05) reduction in superoxide detection measured from 5 to 60 min, reducing levels at 5 min from 174,800 RFU to 48,029 RFU. The GST phenotype influenced the production of superoxide anions in bone marrow cells following PMA stimulation (Fig. 2c). Bone marrow cells isolated from GST^/ mice produced less superoxide anions compared with GST^+/+ (134,600 RFU, compared with 20,650 RFU).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Effect of TLK199 (, 0 µM; , 10 µM; , 50 µM; , 100 µM) on superoxide anion detection. (a) In GST+/+ and (b) GST/ (open symbols), bone marrow cells stimulated with PMA (0.4 µM). TLK199 inhibited the detection of superoxide anions in a dose-dependent manner. P < 0.01 for 50 and 100 µM when compared with no drug (c) reveals the time course of PMA response in GST+/+ ( no PMA;  + PMA) and GST/ bone marrow cells (, no PMA; , + PMA). Note that  and  overlap exactly in the figure. In PMA-treated cells ( and ), the response of GST+/+ cells is significantly different from GST/ (P < 0.05).

GST Status and Kinase Activity. Regulatory kinase pathways, especially those involving JNK and ERK, are frequently linked with cell proliferative status (Pedram et al., 1998; Rausch and Marshall, 1999). Because of our earlier studies ascribing a role for GST in regulating these kinases (Adler et al., 1999; Yin et al., 2000; Davis et al., 2001), we analyzed ERK activity under conditions where GST expression was genetically or pharmacologically manipulated. Figure 3 shows that, whereas protein levels of ERK1/ERK2 were similar in early passage GST^+/+ and GST^/ MEF cells (lanes 3 and 4), enhanced kinase activity was reflected either by ERK phosphorylation (~3.4-fold) or phosphorylation of the downstream substrate ELK-1 (~2.3-fold). In immortalized MEF cultures (passage 105, lanes 5 and 6) quantitative kinase activities still remained, albeit at a reduced level (1.3-fold p-ERK; 2-fold p-ELK-1).

View larger version (36K): [in this window] [in a new window]   Fig. 3.   Immunoblot analysis of basal levels of protein, phosphorylation, and activity of p44/42 MAP kinases (ERK1/2) in MEF cells and selected resistant cells. Lanes (1) HL60 (WT), (2) HL60/TLK199 (3) GST+/+ passage 1, (4) GST/ passage 1, (5) GST+/+ passage 105, and (6) GST/ passage 105. One representative set of data is shown. The mean ± S.D. (3 experiments) for the ratios of each are: total ERK, HL60/TLK199:WT, 0.87 ± 0.24; GST/:GST+/+ (passage 1), 1.02 ± 0.18; GST/:GST+/+ (passage 105), 1.16 ± 0.08; p-ERK, HL60/TLK199:WT, 2.25 ± 0.40; GST/:GST+/+ (passage 1), 3.43 ± 0.14; GST/:GST+/+ (passage 105), 1.3 ± 0.13; p-ELK-1, HL60/TLK199:WT, 2.20 ± 0.41; GST/:GST+/+ (passage 1), 2.26 ± 0.25; GST/:GST+/+ (passage 105), 2.03 ± 0.38.

View larger version (44K): [in this window] [in a new window]   Fig. 4.   Selection and growth of HL60 WT or cells grown in the constant presence of 50 µM TLK199 influence expression and enzyme activity (JNK phosphorylation status)measured by immunoblot with or without exposure to 60 J/m2 UV. Results shown are representative of three independent experiments. Quantitation was achieved using NIH Image software. JNK1 protein levels in HL60/TLK199 cells are increased 2.6 ± 1.1-fold over WT. Exposure to UV does not alter this ratio. JNK1 enzyme activity is below detection threshold until stimulated by UV, whereupon the ratio for TLK199 resistant (Res) cells is 3.0 ± 1.2.

The Effect of TLK199 on Apoptosis. Wild type HL60 cells were induced to undergo apoptosis by 8- and 24-h treatments with 10 and 50 µM TLK199 (Fig. 5). Additive apoptotic effects were seen when HL60 cells were co-treated with 50 µM TLK199 and UV. In contrast, the pattern of apoptosis was significantly different in the HL60/TLK199 cell line. For these experiments, TLK199 was removed from the resistant cell line for one passage prior to additional exposure. Figure 5 shows that overall levels of induced apoptosis were significantly lower, even for single modality UV light exposure. Whereas the HL60/TLK199 cells did undergo apoptosis following combined UV and TLK199 treatment, a significantly lower level (15.1% ± 1.9, compared with 54.6% ± 6.5 for HL60 WT at 24 h) was apparent.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Effect of TLK199 (10 and 50 µM) and UV (60 J/m2) treatment on apoptosis in HL60 and HL60/TLK199 cells. Samples were taken 8 and 24 h after treatment. For each of the treatments and combinations, the resistant cells have significantly lower levels of apoptosis, compared with the sensitive cells (P < 0.05).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

It has been known for many years that GSTP1-1 is expressed at high levels in many solid tumors and in cells made resistant to a wide range of anticancer drugs. This is true even when the selecting drug is not a substrate for thioether product formation through GST catalysis (Tew, 1994). In addition, whereas GSTP1-1 is essentially absent from adult rat liver, high levels are expressed in preneoplastic foci in liver (Imai et al., 1997). In light of our recent data implicating GST in regulation of JNK and other kinases, such observations have prompted an assessment of the role that GST may play in influencing mitogenic pathways with a mechanism unrelated to its catalytic properties. The present study has used two model systems to establish whether different expression levels of GSTP1-1 influence cell proliferation. The GST^/ mouse was established previously (Henderson et al., 1998) using gene-targeting technology to delete the murine GST gene cluster (P1 and P2). The initial phenotype described an increased sensitivity to two-step carcinogenesis induction by polycyclic aromatic hydrocarbons and 12-O-tetradecanoylphorbol-13-acetate, resulting in a 3-fold increase in the incidence of skin papillomas 20 weeks following exposure. Although implying a detoxification function, these data did not distinguish between catalytic or ligand binding functions for the protein.

The present data show that GST^/ mice have higher levels of circulating white blood cells when compared with wild type animals, predominantly observed as an increase in circulating lymphocytes. These elevated counts do not seem to be in response to infection, since the animals are free of apparent pathology or common veterinary disease. Previous reports have linked GSH and associated enzymes with regulation of cellular immunity (Herzenberg et al., 1997; Kim et al., 1998). Thus, the animal phenotype would be consistent with the involvement of GSTP1-1 in control of myeloproliferation. Furthermore, the fact that only GST^+/+ mice showed a myeloproliferative response when treated with TLK199, the selective GST inhibitor, provides additional support for the importance of GST to the proliferative response. These data are consistent with the results showing that TLK199 caused an increase in GM-CFU only in bone marrow from GST^+/+ mice. This suggests that the presence of, and subsequent inhibitor binding to, GST is critical to the proliferative effects of the drug, and confirm that manipulation of GST is a prerequisite for the stimulation of myeloid progenitor proliferation. In a series of rodent studies, TLK199 impacted positively on the nadir of neutrophil counts and enhanced the recovery period following a myelotoxic dose of 5-fluorouracil. Both in terms of quantitative pharmacologic response and time to recovery, the TLK199 effect was essentially equivalent to granulyte-colony stimulating factor [unpublished data and Dr. Rey Gomez (Telik, San Francisco) personal communication].

Further support for a direct role for GST in control of proliferation is afforded by the MEF cultures established from the GST^+/+ and GST^/ mice. The GST^/ cells have a faster doubling time than the GST^+/+. Retroviral-mediated transfection of GSTP1-1 into GST^/ MEF cells partially abrogated the doubling time difference. The fact that proliferative rate was not slowed to the equivalent of the GST^+/+ cells may be explained by the transfection efficiency (~65%), which would be reflected by an overall "averaging" of heterogeneity in the cell populations. Immortalization of the MEF cultures resulted in a significantly enhanced doubling time for both cultures. Preliminary indications suggest that changes in p53 expression and mutations may account for the enhanced rates, and that this may abrogate the difference between GST^+/+ and GST^/ (data not shown).

Insight into how GSTP1-1 may participate in regulation of proliferation may be gained from our recent reports showing that the protein is a negative regulator of JNK catalytic activity (Adler et al., 1999) and also influences other downstream kinase pathways (Yin et al., 2000). The functional ligand binding properties of GSTP1-1 are exemplified by fluorescence resonance energy transfer data showing a K_d of ~200 nM for the interaction with purified full-length JNK1 (Wang et al., 2001). There is now significant evidence that kinase pathways, particularly involving ERK and JNK, have prominent roles in regulation of proliferative pathways in a wide range of cell types (Pedram et al., 1998; Ishikawa and Kitamura, 1999; Rausch and Marshall, 1999). The increase in ERK1/ERK2 activity in GST^/ MEF would be consistent with the absence of GST-mediated regulatory control producing elevated kinase activity and increased proliferation rates. Although it could be argued that elevation in ERK1/ERK2 and JNK activities is a result rather than a cause of proliferation, the specific nature of the defined GST null phenotype must causally link these pathways in some mechanistic fashion. Other small molecule thiol active agents, such as the aminothiol WR2721, have also been found to cause myeloproliferation (List and Gerner, 2000), and it may be through a mechanism such as thiol modification in either GST (inhibiting its negative regulation of JNK) or other proteins important to proliferation that these agents share a common mechanism.

We established another model system to take advantage of the adaptive changes brought about by the process of chronic exposure and selection in TLK199. HL60 cells are of leukemic myeloid lineage. The HL60/TLK199 cell line grows routinely in 50 µM of drug, a concentration that induces apoptosis in the parent HL60 cell line. Such a model is valuable in determining possible adaptive responses that may indicate the molecular targets of the drug. Indeed, some of the changes have been directly associated with detoxification and enhanced drug efflux (e.g., increased expression of -glutamylcysteine synthetase and multidrug resistance-associated protein) (O'Brien et al., 1999). The present data suggest that changes in stress kinase response pathways are also an adaptation to chronic TLK199 exposure. This adaptive response extends our earlier in vitro data showing that TLK199 can cause a dissociation of the GSTP1-1:JNK1 interaction (Adler et al., 1999). In effect, the value of the chronic drug exposure lies in the extrapolation of the in vitro observations into a cellular system. Although correlative in nature, the cells do provide a functional in vivo link with TLK199 treatment and JNK expression.

Apoptosis induced by UV was observed in HL60 WT, as well as in HL60 cells resistant to TLK199. However, the levels of induced apoptosis were significantly lower in the resistant cell line presenting higher ERK1/2 and JNK1 activities. These data are consistent with the numerous studies on the roles of MAP kinases in apoptosis (Cross et al., 2000). Indeed, whereas the ERK family is reported to have antiapoptotic properties, JNK family members have been described as proapoptotic proteins. Moreover, it was demonstrated that the induction of ERK activity could protect HL60 cells against apoptosis induced by anisomycin, a potent inducer of JNK activity (Stadheim and Kucera, 1998). Then, following the concept that the balance between pro- and anti-apoptotic factors determines the fate of a cell, the increased activity of ERK in HL60 cells resistant to TLK199 may partially protect these cells against apoptosis induced by UV, despite the increased JNK1 activity. The pharmacological inhibition of GST by non-cytotoxic concentrations of TLK199 did not modify significantly the level of UV-induced apoptosis in either WT or resistant HL60 cells. In addition, in WT cells, a cytotoxic concentration of TLK 199 and UV had only an additive effect on apoptosis. For example, even where the differences in response appear most marked (8 h in sensitive cells), 52% apoptosis caused by the combination of 50 µM TLK199 and 60 J/m² UV is not significantly different from the additive values of TLK199 and UV alone (17 + 21 = 38%; P > 0.05). Thus, whereas the forced overexpression of GST protects NIH3T3 cells against apoptosis induced by H₂O₂ by modifying the activation of MAP kinase pathways (Yin et al., 2000), the pharmacological inhibition of this enzyme may appear to be insufficient to alter the level of cell death observed in our cell lines after UV treatment. This discrepancy could be explained by the fact that, although UV irradiation causes ROS, ultimately cytotoxic effects could involve other mechanisms unrelated to ROS generation.

Paradoxically, a number of studies have emphasized the role of low-level ROS in governing cell proliferation in response to growth factor stimulation (Smith et al., 2000). Extremely low concentrations of anticancer drugs, such as Adriamycin, produced oxidative stress that can actually stimulate tumor cell proliferation. Thus, in clonogenic survival assays, proliferative indices will frequently exceed 100% prior to achieving therapeutic cell kill. For TLK199, we considered the possibility that the altered kinase activities were the consequence of drug-induced ROS. However, in bone marrow cells mitogenically stimulated by PMA, TLK199 appeared to act as a scavenger of ROS, such as superoxide anions. In the GST^/ marrow samples, PMA produced lower levels of superoxide, providing support to the principle that the proliferative effects were not a function of simple changes in ROS levels.

In summary, we provide evidence that GST is involved directly in controlling cellular mitogenic pathways that influence proliferation. The reported role of GST as a negative repressor of JNK (and other kinases) could significantly impact on kinase cascades that influence stress response, proliferation, or apoptosis pathways. Conceivably, the extent and duration of kinase stimulation may determine cellular fate (Davis et al., 2001). Evolutionary conservation and ubiquitous expression of GST at variable levels in proliferative tissues reinforces the principle that this protein has an important role as a ligand-binding protein involved in kinase-mediated stress and proliferation pathways. Pharmacological manipulation of GSTP1-1 with a GSH peptidomimetic agent can be achieved with a concomitant impact upon mitogenic response. This approach may have a useful therapeutic application where a nongrowth factor-mediated stimulation of myeloproliferation can be achieved.