Search

Background: Glutathione S-transferase is often up-regulated in neoplastic tissues. A single previous study found a loss of expression associated with carcinogenesis of the prostate., Methods: To extend these results, the authors performed immunohistochemical staining for the pi-class of glutathione S-transferase (GSTpi) on 74 archival sequential prostate specimens. The antibody used was derived from rabbits immunized against purified human GSTpi. Paraffin blocks containing both benign tissue and adenocarcinoma were studied., Results: Heterogeneous expression of GSTpi in benign acini was found in 96% of cases, but GSTpi was not expressed in 95% of invasive adenocarcinomas of the prostate, nor was it expressed in any of the foci of high grade prostatic intraepithelial neoplasia. Basal cells of benign acini showed strong, diffuse staining for GSTpi, whereas the secretory luminal epithelium expressed GSTpi weakly and focally., Conclusions: This study confirms the down-regulation of GSTpi in adenocarcinoma of the prostate and shows that the loss of GSTpi expression is a phenotype associated with malignant transformation.

Buthionine sulfoximine (BSO) is a synthetic amino acid that irreversibly inhibits an enzyme, gamma-glutamylcysteine synthetase (gamma-GCS), which is a critical step in glutathione biosynthesis. We isolated three BSO-resistant sublines, KB/BSO1, KB/BSO2, and KB/BSO3, from human epidermoid cancer KB cells. These cell lines showed 10-to 13-fold higher resistance to BSO, respectively, and had collateral sensitivity to cisplatin, ethacrynic acid, and alkylating agents such as melphalan and nitrosourea. Cellular levels of glutathione S-transferase pi (GST-pi) and its mRNA in BSO-resistant cell lines were less than 10% of the parental cells. Nuclear run-on assay showed that the transcriptional activity of GST-pi was decreased in BSO-resistant cells, and transient transfection of GST-pi promoter-chloramphenicol acetyltransferase constructs revealed that the sequences between -130 and -80 base pairs of the 5'-flanking region wer at least partially responsible for the decreased expression of the GST-pi gene. By contrast, gamma-GCS mRNA levels were 3-to 5-fold higher in resistant cell lines than in KB cells, and the gamma-GCS gene was found to be amplified in the BSO-resistant cells lines. GST-pi mRNA levels appeared to be inversely correlated with gamma-GCS mRNA levels in BSO-resistant cells. We further established the transfectants, KB/BSO3-pi1 and KB/ BSO2-pi2, that overexpressed GST-pi, from KB/BSO3, after introducing a GST-pi expression plasmid. These two transfectants had similar levels in gamma-GCS mRNA, drug sensitivity to alkylating agents, and glutathione content at those of KB cells. These findings suggest that the cellular levels of GST-pi and gamma-GCS might be co-regulated in these novel BSO-resistant cells.

The glutathione transferases are involved in intracellular detoxification reactions. One of these, GST pi, is elevated in some breast cancer cells, particularly cells selected for resistance to anticancer agents. We evaluated GST pi expression in 60 human breast tumors by three techniques, immunohistochemistry. Northern hybridization, and Western blot analysis. There was a significant positive correlation between the three methods, with complete concordance seen in 64% of the tumors. There was strong, inverse relationship between GST pi expression and steroid receptor status with all of the techniques utilized. In addition, there was a trend toward higher GST pi expression in poorly differentiated tumors, but no correlation was found between tumor GST pi content and DNA ploidy or %S-phase. GST pi expression was also detected in adjacent benign breast tissue as well as infiltrating lymphocytes; this expression may contribute to GST pi measurements using either Northern hybridization or Western blot analysis. These results suggest that immunohistochemistry is the method of choice for measuring GST pi in breast tumors.

The glutathione S-transferase pi gene (GST pi) is highly expressed in estrogen receptor negative (ER-) but not expressed in ER+ human breast cancer cell lines. To define regulatory mechanisms of GST pi gene expression, we analyzed both the activity of the GST pi promoter and the posttranscriptional fate of GST pi RNA sequences in three ER+ and three ER- breast cancer cell lines. Expression of a transiently transfected CAT reporter gene driven by the GST pi promoter and 2203 nucleotides of 5'-flanking sequences were similar in all six cell lines regardless of ER status. Endogenous GST pi transcription rates in nuclei isolated from ER- cells were quite low despite high steady state levels of cytoplasmic mRNA. Furthermore, the endogenous GST pi gene was transcribed in ER+ nuclei at rates similar to those obtained in ER- nuclei. We determined the stabilities of mRNAs transcribed from the endogenous GST pi gene (ER- cells) and from a stably transfected GST pi cDNA expression vector (ER+ and ER- cells). The endogenous GST pi mRNA was extraordinarily stable in ER- cells. Comparisons between transfected ER+ and ER- cells revealed no significant differences in the stabilities of transfection-derived GST pi mRNA sequences. We conclude that GST pi mRNA stability contributes significantly to the high levels of cytoplasmic mRNA observed in ER- cells, but that the differential expression of GST pi in ER+ versus ER- cells is governed by other posttranscriptional processes.

Increased expression of certain glutathione S-transferase (GST) isoenzymes has frequently been associated with the development of resistance to alkylating agents and other classes of antineoplastic drugs in drug-selected cell lines. The question arises whether this phenomenon is causal or is a stress-induced response associated with drug resistance in these cell lines. We have constructed mammalian expression vectors containing the human GST mu and GST alpha 2 (Ha2) cDNAs and stably transfected them into the human breast cancer cell line MCF-7. Whereas the parental and pSV2neo-transfected cell lines display low GST activity, three individual transfected clones were identified in each group that expressed either GST mu or GST alpha 2. The range of GST activities was similar to those observed in cells selected for anticancer drug resistance. The GST mu specific activities were 56, 150, and 340 mlU/mg, compared with 10 mlU/mg of endogenous GST mu in control lines. Specific activities in GST alpha 2-transfected clones were 17, 28, and 52 mlU/mg, compared with no detectable alpha class GST in control lines. These clonal lines and the parental and pSV2neo-transfected control lines were tested for sensitivity to antineoplastic agents and other cytotoxic compounds. The clones with the highest activity in each group were 1.7-fold (GST alpha 2) to 2.1-fold (GST mu) resistant to the toxic effects of ethacrynic acid, a known substrate for GSTs. However, the GST-transfected cell lines were not resistant to doxorubicin, L-phenylalanine mustard, bis(2-chloroethyl)-1-nitrosourea, cisplatin, chlorambucil, or the GST substrates 1-chloro-2,4-dinitrobenzene or tert-butyl hydroperoxide. Thus, although L-phenylalanine mustard, bis(2-chloroethyl)-1-nitrosourea, chlorambucil, tert-butyl hydroperoxide, and 1-chloro-2,4-dinitrobenzene are known to be metabolized by glutathione-dependent GST-catalyzed reactions, there was no protection against any of these agents in MCF-7 cell lines overexpressing GST mu or GST alpha 2. We conclude that, at the levels of GST obtained in this transfection model system, overexpression of GST mu or GST alpha 2 is not by itself sufficient to confer resistance to these anticancer agents. These studies do not exclude the possibility that GST may be a marker of drug resistance or that other gene products not expressed in MCF-7 cells might cooperate with GST to confer drug resistance.

Mouse monoclonal antibodies (MAb) have been generated against the anionic isozyme of human glutathione S-transferase (GST pi). MAb AGST I can inhibit 50-70% of GST pi enzymatic activity and reacts with a 3-dimensional epitope which includes a putative glutathione binding site on GST pi. A sandwich enzyme-immunoassay established using MAb AGST I and a polyclonal antibody displayed a sensitivity of 0.5 ng/ml. Immunohistochemical analysis of human tissues demonstrated marked increases in GST pi levels in cancers of the brain, cervix, endometrium, colon, rectum and testis and in fibro- and chondrosarcomas.

Studies have suggested that the alpha class glutathione S-transferase (GST) may protect cells from the chemotherapeutic drugs chlorambucil and melphalan. In order to further define the function of human alpha class GST, a complementary DNA which encodes it was ligated into an expression vector under the direction of the human metallothionein-IIA promoter and stably transfected into human MCF-7 breast cancer cells in conjunction with the G418-selectable plasmid pSV2neo. Clonal cell lines were identified which expressed increased levels of GST enzyme activity (2.2- to 5.6-fold). The transfected cell lines also had increased peroxidase activity using cumene hydroperoxide as the substrate (1.9- to 3.8-fold) which is consistent with the intrinsic peroxidase activity of alpha class GSTs. Southern blot analysis indicated that genomic DNA from these cells contained a fragment indistinguishable from the transfected alpha class GST complementary DNA (850 base pairs); Northern blot analysis of total cellular RNA indicated that these cells contained appropriately sized alpha class GST RNA (980 nucleotides); and Western blot analysis indicated that, while MCF-7 cells contained no detectable alpha class GST protein, the transfected cells contained markedly elevated levels of alpha class GST but no detectable mu or pi class GST. These alpha class GST transfected cells had increased resistance to ethacrynic acid (2.1- to 3.0-fold). However, the transfected cells failed to show any increased resistance measured at the drug dosage which inhibited 50% of the colony formation to the chemotherapeutic drugs chlorambucil, melphalan, Adriamycin, or cisplatin under conditions of either continuous or 1-h drug exposure. Neither was there any change in sensitivity to the cytotoxins benzo(a)pyrene, benzo(a)pyrene-trans-7,8-dihydrodiol-9,10-epoxide (anti), or 1-chloro-2,4-dinitrobenzene. These studies indicate that expression of this human alpha class GST by itself in MCF-7 human breast cancer cells does not confer resistance to the chemotherapeutic drugs tested under the conditions used in these studies.

Glutathione S-transferases (GSTs), a family of isoenzymes that play an important role in protecting cells from cytotoxic and carcinogenic agents, can be separated by biochemical and immunologic characteristics into three distinct classes named alpha, mu, and pi. Previous studies have indicated that there is marked heterogeneity in the expression of different GST isoenzymes in different normal and malignant tissues. To better understand the regulation of the human pi class glutathione S-transferase isoenzyme (GST-pi), the tissue distribution of this protein wa studied by an immunohistochemical technique using an anti-GST-pi polyclonal antibody in normal paraffin-embedded human tissues. These studies indicate that there is a broad distribution of GST-pi in normal human tissues and establish a precise localization within the different organs studied. GST-pi was expressed predominantly in normal epithelial cells of the urinary, digestive, and respiratory tracts, suggesting a possible role for GST-pi in detoxication and elimination of toxic substances. Previous studies have indicated that GST-pi and the putative drug efflux pump P-glycoprotein are both overexpressed in multidrug-resistant human breast cancer cells and in xenobiotic resistant preneoplastic rat hyperplastic liver nodules. Results from this study indicate that there are also similarities between the normal tissue distribution GST-pi and that previously reported for mammalian P-glycoprotein, particularly in secretory epithelia. This finding suggests that these two gene products, which have been implicated in the development of resistance to cytotoxic drugs, may be coregulated in normal and malignant cells.

The nucleotide sequence of the mdr1 gene encoding a putative drug efflux pump (P-glycoprotein) is homologous to a class of bacterial membrane-associated transport proteins. These bacterial proteins are part of a multicomponent system that includes soluble periplasmic proteins that bind substrates, channeling them through the membrane in an energy-dependent manner. We have investigated the possibility that a similar multicomponent transport system exists in a multidrug-resistant human MCF-7 breast cancer cell line that was initially selected for resistance to doxorubicin (AdrR MCF-7). AdrR MCF-7 cells overexpress both the mdr1 gene and the pi class isozyme of glutathione S-transferase (GST-pi) (EC 2.5.1.18). The latter is one of several isozymes known to have a ligand-binding function in addition to drug-metabolizing capabilities. Although we have recently shown that transfection of a functional GST-pi expression vector is insufficient to confer resistance to doxorubicin in cells that lack P-glycoprotein expression [Mol. Pharmacol. 36:22-28 (1989)], we examined the possibility that GST-pi interacts with P-glycoprotein to alter multidrug resistance. To do this, we have cloned cDNAs encoding these proteins from AdrR MCF-7 cells, constructed expression vectors containing these two genes, and transfected these vectors sequentially into drug-sensitive MCF-7 cells. The human mdr1 cDNA isolated from AdrR MCF-7 is a variant gene whose sequence differs from that isolated previously from vinblastine-resistant KB cells [Cell 53:519-529 (1989)], resulting in an amino acid substitution of alanine to serine at position 893 (mdr1/893ala). Transfection of eukaryotic expression vectors containing the mdr1 gene isolated from AdrR MCF-7 cells produced a multidrug-resistant phenotype in recipient cells, with a cross-resistance pattern similar to that in the AdrR MCF-7 cells. To determine whether GST-pi expression could augment resistance provided by mdr1, two clones transfected with mdr1, one with high levels (153% of mdr1 RNA in AdR MCF-7 cells) and one with low levels (10% of mdr1 RNA in AdrR MCF-7 cells), were subsequently cotransfected with a GST-pi expression vector and pSVNeo and selected for resistance to G418. Six of these clones contained levels of GST-pi that were 8- to 18-fold greater than GST levels found in mdr1-expressing clones transfected with nonspecific DNA. We found no difference in the degree of resistance to doxorubicin, actinomycin D, and vinblastine between the clones expressing mdr1 only and the clones expressing both mdr1 and GST-pi.(ABSTRACT TRUNCATED AT 400 WORDS)

To investigate the transcriptional regulation of human glutathione S-transferase pi (GST pi) gene expression, we fused the GST pi promoter, including 2203 bp of the 5'-flanking region, exon 1, and most of intron 1, to the chloramphenicol acetyltransferase (CAT)-encoding reporter gene (cat). When transfected into human cell lines, this GST-cat construct (-2203 GST-cat) supported high level cat gene expression. RNase-protection and primer-extension experiments showed that the normal GST pi transcriptional start point (tsp) is utilized, and furthermore, that intron 1 is faithfully removed by splicing from the majority of primary GST-cat transcripts. A series of constructs containing deletions in the GST pi sequences of the -2203 GST-cat vector were prepared to define potential regulatory regions. Transfection of these deletion plasmids revealed that a region between GST pi sequences -80 and -8 is absolutely required for cat expression. Furthermore, transfection of the -2203 GST-cat and deletion vectors into two human cell lines--one line which does not produce endogenous GST pi (HeLa cells) and one which produces high levels of endogenous GST pi (HS 578T cells)--failed to identify sequences that differentially influence the level of transcription in either cell line. A putative TRE (TPA responsive element or AP-1 recognition sequence) strategically situated upstream from the GST pi tsp (-69 to -63) was examined by TPA treatment of HeLa cells transfected with GST-cat DNA. Additionally, the potential interaction of fos and jun proteins with the GST pi promoter was examined by co-transfection of GST-cat constructs with jun and fos expression vectors in F9 cells. Both of these treatments, which are known to enhance transcription of several genes containing 5'-flanking TREs, failed to induce GST-cat expression. These data suggest that the putative TRE sequence in GST pi is unresponsive both to phorbol esters and to these particular transcriptional activating factors of the fos and jun family.

A remarkably diverse family of glutathione S-transferases (GSTs) has evolved in higher organisms. These intracellular enzymes catalyze the nucleophilic attack of glutathione on a variety of hydrophobic, electrophilic xenobiotics often resulting in conjugated or transformed metabolites that are less toxic and more easily excretable. Additionally, some GST isozymes may participate in the repair of oxidative damage to membrane lipids and DNA. Finally, GSTs are high capacity intracellular binding proteins which, independently of their enzymatic activities, may serve in the storage, transport, or sequestration of many hydrophobic compounds. These properties suggest that GSTs may function as important cellular defenses against the cytotoxic effects of carcinogenic and antineoplastic agents. Here we discuss recent evidence that bears upon this notion.

The development of multidrug resistance in MCF-7 human breast cancer cells and the acquisition of broad resistance to xenobiotics in rat hyperplastic nodules are both associated with increased P-glycoprotein (mdr) gene expression as well as changes in activities of intracellular detoxication enzymes; among these changes is a significant increase in the activity of the anionic isozyme of glutathione-S-transferase (GST). We have isolated a cDNA encoding the human anionic glutathione-S-transferase, GST pi-1, from a cDNA library constructed from multidrug-resistant MCF-7 cells. The deduced amino acid sequence of GST pi-1 shows that while the human anionic GST displays 85% nucleotide and amino acid sequence homology to the rat anionic isozyme, it is markedly less related to human basic GST isozymes. We have examined the expression of GST pi and P-glycoprotein in 170 specimens of human tissues and tumors. P-Glycoprotein RNA expression was positive in eight of 23 lymphomas and two of 12 colon tumors; however, many other normal and malignant tissues, including lung, bladder, and breast tumors, had low or undetectable levels of P-glycoprotein RNA expression. In contrast, GST pi was readily detected in a wide variety of normal and malignant tissues. The level of GST pi mRNA expression in normal tissues was heterogeneous, with lowest levels found in liver and the highest levels found in lung, esophagus, and placenta. GST pi was also variably expressed in human tumors, with the lowest relative levels occurring in lymphoma and breast cancer and the highest levels found in lung cancer and head and neck tumors. In addition, comparison of paired specimens from the same patient indicated that GST pi expression was increased in many tumors relative to matched normal tissue.

The complete human genomic glutathione-S-transferase-pi gene (GST-pi) was isolated from a lambda Charon 4A bacteriophage library which was screened by hybridization to a human GST-pi cDNA. We have sequenced 4261 bp which include the entire GST-pi gene as well as over 1200 bp of the 5' and 200 bp of the 3' flanking regions. The GST-pi gene has 7 exons and 6 introns contained within approximately 2.8 kilobases. Primer extension experiments identified four possible transcription start points closely spaced between 29 and 33 nucleotides (nt) 5' to the start of translation. Analysis of the GST-pi promoter region revealed 4 putative transcription regulatory motifs; these sequences include a 'TATA' box 29 bp upstream from the major transcription start point (nt position -29), 2 Sp1 recognition sequences (GGGCGG, nt positions -46 to -41 and -56 to -51), and an AP-1 recognition sequence (TGACTCA, nt positions -69 to -63). The first 200 nt 5' to the start point of transcription contain a G + C-rich region (79%). Additionally, an intriguing A + T-rich region was found between nt positions -505 and -413 which contained 17 AAAAT tandem repeats. Comparison of the GST-pi gene with the homologous rate gene, GST-P, disclosed extensive conservation of genomic organization between the two species.

The cytosolic glutathione S-transferases (GSTs, EC 2.5.1.18) are a superfamily of dimeric isoenzymes which catalyze the conjugation of electrophilic substrates with glutathione. We have isolated from a murine cell line (L929) two mu class GST cDNAs, pmGT10 (1.4 kilobases (kb] and pmGT2 (1.1 kb). Analysis of the deduced amino acid sequences revealed 93% homology between pmGT10 and the rat Yb1 (subunit 3) gene and 95% homology between pmGT2 and the rat Yb2 (subunit 4) gene. Furthermore, the 3'-untranslated regions of pmGT10 display a marked degree of homology to the 3' region of the rat Yb1 gene, while this region of pmGT2 displays marked homology to the corresponding region of the rat Yb2 gene. Using probes specific for each gene, northern blot analysis demonstrated that pmGT10 and pmGT2 hybridized to a 1.3-1.4-kb mRNA species present in L929 cells. These two murine cDNAs were subcloned into bacterial expression vectors, and enzymatically active GSTs encoded by pmGT10 (mGTmu1) or pmGT2 (mGTmu2) were purified from transformed Escherichia coli lysates. Western blot analysis of the individual GSTs produced in E. coli indicated that both genes encode GSTs which reacted with antibodies directed against mu class GST, but not with antibodies directed against alpha or pi class GSTs. The isoelectric points of these purified homodimers are 8.7 and 7.1, respectively, which are remarkably similar to those of the GST homodimers of the homologous rat subunits, 3-3 (or Yb1) and 4-4 (or Yb2), which are 8.4 and 6.9, respectively. Furthermore, the two murine subunits can form a heterodimer having a pI = 8.1 after in vitro dissociation with guanidine HCl, followed by dilution and dialysis to allow reassociation. Both homodimers of mGTmu1 and mGTmu2 and an apparent heterodimer are formed in L929 cells, as demonstrated by the isoelectric focusing profile of GST purified from this cell line. Hybridization of gene-specific probes with RNA from normal mouse tissues revealed that mGTmu1 transcripts were highly expressed in murine kidney, heart, lung, liver, and brain. Lower levels of mGTmu2 transcripts were also detected in kidney, heart, and lung. Expression of mGTmu2 RNA was not detected in murine liver or brain, which is in contrast to the regulation of the homologous rat subunit 4 (Yb2), which is expressed in liver and brain.

GST isozymes are an important part of the normal cellular defense against toxic xenobiotics and carcinogens. These multifunctional proteins can interact with a broad range of substrates in a variety of ways. In particular, GSTs have been implicated in the detoxication of many antineoplastic agents. Elevated levels of certain GST isozymes have been associated with malignant transformation and with experimental drug resistance. Although the role of GST in antineoplastic drug resistance is unclear, recent studies have shown increased activity of GST in many human tumors relative to normal tissues. These findings raise the possibility that the presence of certain GST isozymes may be a marker for malignant transformation in some human tumors, and that GSTs may play a role in de novo and acquired drug resistance. Identifying the factors which regulate the expression of these drug-metabolizing enzymes as well as agents which inhibit their activities may provide new insights into the therapy of tumors clinically refractory to chemotherapy.

Adriamycin-resistant (AdrR) human breast cancer cells have been selected which exhibit cross-resistance to a wide range of anti-cancer drugs. This multidrug-resistant phenotype is associated with increases in the activities of glutathione peroxidase and glutathione transferase. The 45-fold increase in glutathione transferase activity is associated with the appearance of a new anionic isozyme in AdrR cells which is immunologically related to the anionic glutathione transferase present in human placenta. The increase in transferase and the level of drug resistance is relatively stable during passage of AdrR cells in the absence of adriamycin for over 10 months. A similar anionic glutathione transferase isozyme is also found in rat hyperplastic liver nodules, a preneoplastic state resulting from exposure to carcinogens. A rat cDNA which codes for the anionic glutathione transferase in rat hyperplastic nodules hybridizes to a 1.1-kilobase pair mRNA which is overexpressed in the AdrR MCF-7 cells. The anionic transferase has been purified from the AdrR cells and found to have characteristics which distinguish it from other anionic human glutathione transferases, including high levels of intrinsic peroxidase activity. The overexpression of a similar anionic glutathione transferase in human breast cancer cells selected for multidrug resistance and in rat hyperplastic liver nodules, which develop resistance to various hepatotoxins, suggests a possible role for this drug-conjugating enzyme in the mechanism of resistance in both of these states.

The development of multidrug resistance in MCF7 human breast cancer cells is associated with overexpression of P-glycoprotein, changes in activities of several detoxication enzymes, and loss of hormone sensitivity and estrogen receptors (ERs). We have cloned the cDNA for one of the drug-detoxifying enzymes overexpressed in multidrug-resistant MCF7 cells (AdrR MCF7), the anionic isozyme of glutathione S-transferase (GST pi). Hybridization with this GST pi cDNA, GST pi-1, demonstrated that increased GST pi activity in AdrR MCF7 cells is associated with overexpression but not with amplification of the gene. We mapped the GST pi gene to human chromosome 11q13 by in situ hybridization. Since multidrug resistance and GST pi overexpression are associated with the loss of ERs in AdrR MCF7 cells, we examined several other breast cancer cell lines that were not selected for drug resistance. In each of these cell lines we found an inverse association between GST pi expression and ER content. We also examined RNA from 21 primary breast cancers and found a similar association between GST pi expression and ER content in vivo. GST pi mRNA content in 11 ER-positive tumors (less than or equal to 10 fmol/mg of protein) was significantly different from the GST pi content of 10 ER-negative tumors (P = 0.002; Mann-Whitney Wilcoxon test for two independent samples). The finding of similar patterns of expression of a drug-detoxifying enzyme and of ERs in vitro as well as in vivo suggests that ER-negative breast cancer cells may have greater protection against antineoplastic agents conferred by GST pi than ER-positive tumors.

Modified glutathione S-transferase (GST) activity was detected in MCF-7 cell lines selected in vitro for modest levels of resistance (2.6- to 13.7-fold) not only to Adriamycin but also to Novantrone, vincristine (VCR), and etoposide (VP-16). Western blot analysis revealed a qualitative increase in the expression of the pi class GST isozyme and some apparent down regulation of the mu class in these resistant lines relative to the parental line. Overexpression of GST pi protein was also observed after in vitro exposure of MCF-7 cells to fractionated x-irradiation, which resulted in a subline that exhibited 5-fold resistance to VCR and 3-fold resistance to VP-16. Northern blot analysis provided qualitative evidence that these changes in GST pi and mu appeared to be at the transcriptional level in these VCR- and VP-16-resistant MCF-7 sublines. In contrast, similarly selected VCR-resistant, VP-16-resistant, and x-ray-pretreated-resistant sublines, derived from the SuSa cell line established from a human testicular teratoma, did not exhibit significant alterations in GST isozyme profiles compared with their parental cell lines. All parental and resistant SuSa lines that were studied expressed the GST pi isozyme constitutively. Total glutathione content and glutathione peroxidase activities were not modified in parallel with the GST alterations in these human tumor drug resistant sublines. These data suggest that the increased cellular GST activity was associated with selection of resistance to many of the drugs associated with the multidrug resistant phenotype in MCF-7 sublines but that this was not the case for the other human tumor cell line (SuSa) tested.

Increased expression of the glutathione S-transferase (GST; E.C.2.5.1.18) pi class isozyme is associated with both malignant transformation and drug resistance, as well as with decreased estrogen receptor content in breast cancer. In order to further characterize the role of this enzyme in drug resistance, we cloned the cDNA encoding the human isozyme GST pi and developed two eukaryotic expression vectors using this cDNA and either the human metallothionein IIa or cytomegalovirus immediate-early promoters. These GST pi expression vectors were cotransfected with pSV2neo into drug-sensitive MCF-7 human breast cancer cells, which have low amounts of GST activity and which do not express GST pi. The transfected cells were selected for G418 resistance and individual clones were screened for GST activity. Three clones that demonstrated increased GST activity were selected for further study. Immunoprecipitation studies demonstrated that the increase in GST activity in these clones was due to expression of GST pi. Although the total GST activity of the positive clones was increased as much as 15-fold over that in wild-type MCF-7 cells, there was no change in glutathione peroxidase activity, as measured using cumene hydroperoxide as a substrate. Immunoblot studies revealed that the increased GST enzyme produced in the transfected cells was identical in size to endogenous GST pi. Southern blot analysis demonstrated the incorporation of the GST pi expression vector into the genome of the positive clones and Northern blot analysis showed that the transfected genes made a hybrid GST pi RNA that was slightly larger than the endogenous GST pi RNA. Primer extension studies demonstrated that this increase in length corresponded to the added length of the 5' leader sequence of the expression vector. The effect of increased GST pi activity on the sensitivity of the transfected clones to several cytotoxic agents was assessed by colony-forming assay. The transfected clones were slightly more resistant (1.3-4.1-fold) to benzo(a)pyrene and its toxic metabolite benzo(a)pyrene-(anti)-7,8-dihydrodiol-9,10-epoxide, as well as to ethacrynic acid (3.1-to 4.4-fold). Although increased GST pi expression is found in MCF-7 cells selected for doxorubicin resistance, the transfected clones were not consistently more resistant to doxorubicin than control cells. In addition, the transfected cells were not resistant to either melphalan or (cis)-platinum, even though conjugation with glutathione is known to play a role in the detoxification of both of these drugs.(ABSTRACT TRUNCATED AT 400 WORDS)