Cytochrome P450 interactions are common and consequential in Massachusetts hospital discharges.

Despite the recognition that drug-drug interactions contribute substantially to preventable health-care costs, the prevalence of such interactions related to the cytochrome P450 system in clinical practice remains poorly characterized. This study drew retrospective hospital discharge cohorts from a large health claims data set and a large health system data set. For every hospital discharge, frequency of co-occurrence of substrates and inducers or inhibitors at cytochrome P450 2D6, 2C19, 3A4 and 1A2 were determined. A total of 124 520 individuals in the state of Massachusetts (health claims cohort) and 77 026 individuals in two large academic medical centers (electronic health record (EHR) cohort) were examined. In the claims cohort, 35 157 (28.2%) exhibited at least one CYP450 drug-drug interaction at hospital discharge, whereas in the EHR cohort, 36 750 (47.7%) had at least one interaction. The most commonly affected CYP450 systems were 2C19 and 2D6, with putative interactions observed in at least 10% of individuals at discharge in each cohort. Odds of hospital readmission within 90 days among those discharged with at least one interaction were 10-16% greater, with mean health-care cost $574/month greater over the subsequent year, after adjusting for age, sex, insurance type, total number of medications prescribed, Charlson comorbidity score and presence or absence of a psychiatric diagnosis. These two distinct clinical data types show that CYP450 drug-drug interactions are prevalent and associated with greater probability of early hospital readmission and greater health-care cost, despite the widespread availability and application of drug-drug interaction checking software.

Polygenic loading for major depression is associated with specific medical comorbidity.

Major depressive disorder frequently co-occurs with medical disorders, raising the possibility of shared genetic liability. Recent identification of 15 novel genetic loci associated with depression allows direct investigation of this question. In cohorts of individuals participating in biobanks at two academic medical centers, we calculated polygenic loading for risk loci reported to be associated with depression. We then examined the association between such loading and 50 groups of clinical diagnoses, or topics, drawn from these patients' electronic health records, determined using a novel application of latent Dirichilet allocation. Three topics showed experiment-wide association with the depression liability score; these included diagnostic groups representing greater prevalence of mood and anxiety disorders, greater prevalence of cardiac ischemia, and a decreased prevalence of heart failure. The latter two associations persisted even among individuals with no mood disorder diagnosis. This application of a novel method for grouping related diagnoses in biobanks indicate shared genetic risk for depression and cardiac disease, with a pattern suggesting greater ischemic risk and diminished heart failure risk.

Impact of antidepressant treatment during pregnancy on obstetric outcomes among women previously treated for depression: an observational cohort study.

OBJECTIVE: To examine the impact of pharmacologic treatment for depression on obstetric outcomes in women treated for depression during the 2 years prior to pregnancy. STUDY DESIGN: Observational cohort study among 2859 women treated for depression during the 2 years prior to pregnancy. The primary exposure was any antidepressant treatment during pregnancy. Secondary analyses examined the impact of treatment by period of antidepressant exposure. Multivariable logistic regression models as well as propensity score analysis was utilized. RESULTS: Among 2859 women, 1648 (58%) were treated with antidepressant medication during pregnancy. Women who received antidepressants had no difference in preterm and early-term deliveries, Apgar scores, and small for gestational age (SGA); they had a lower likelihood of breastfeeding (adjusted odds ratio (AOR) 0.69, (95% confidence interval (CI): 0.51 to 0.94)). In secondary analysis, women who used antidepressants all three trimesters who delivered at term were more likely to deliver early term (AOR 1.36, (95% CI: 1.09 to 1.72)). Women who were treated with antidepressants only during the first and second trimesters had a reduced likelihood of SGA (AOR: 0.51 (95% CI: 0.32 to 0.83)). Generally similar results were observed with propensity score analysis. CONCLUSION: Antidepressant exposure during pregnancy does not confer an increased risk of preterm birth nor growth restriction in women recently treated for depression, but also does not appear to markedly improve these outcomes.

Prevalence and implications of cytochrome P450 substrates in Massachusetts hospital discharges.

The cytochrome P450 (CYP450) system of drug-metabolizing enzymes may contribute to individual variation in drug response. We examined prevalence of CYP450 substrates at hospital discharge for patients in two cohorts: insurance claims of Massachusetts residents and the medical records of two academic medical centers. The claims cohort included 47 473 individuals (38.2%) treated with at least one CYP450 2D6, 2C19, 3A4 or 1A2 substrate. The electronic medical records cohort included 45 905 individuals (57.4%) treated with at least one substrate. In adjusted models, substrates of CYP450 2D6 and 2C19 were associated with greater risk for 90-day readmission in both cohorts (odds ratios of 1.104 and 1.128 (P<0.001), respectively). Presence of any CYP450 substrate was associated with increased monthly medical costs (+$397, P<0.003). These analyses of more than 300 000 admissions using two different cohorts and data types indicate that CYP450 substrates are associated with greater readmission rates and greater health-care cost.

Predicting early psychiatric readmission with natural language processing of narrative discharge summaries.

The ability to predict psychiatric readmission would facilitate the development of interventions to reduce this risk, a major driver of psychiatric health-care costs. The symptoms or characteristics of illness course necessary to develop reliable predictors are not available in coded billing data, but may be present in narrative electronic health record (EHR) discharge summaries. We identified a cohort of individuals admitted to a psychiatric inpatient unit between 1994 and 2012 with a principal diagnosis of major depressive disorder, and extracted inpatient psychiatric discharge narrative notes. Using these data, we trained a 75-topic Latent Dirichlet Allocation (LDA) model, a form of natural language processing, which identifies groups of words associated with topics discussed in a document collection. The cohort was randomly split to derive a training (70%) and testing (30%) data set, and we trained separate support vector machine models for baseline clinical features alone, baseline features plus common individual words and the above plus topics identified from the 75-topic LDA model. Of 4687 patients with inpatient discharge summaries, 470 were readmitted within 30 days. The 75-topic LDA model included topics linked to psychiatric symptoms (suicide, severe depression, anxiety, trauma, eating/weight and panic) and major depressive disorder comorbidities (infection, postpartum, brain tumor, diarrhea and pulmonary disease). By including LDA topics, prediction of readmission, as measured by area under receiver-operating characteristic curves in the testing data set, was improved from baseline (area under the curve 0.618) to baseline+1000 words (0.682) to baseline+75 topics (0.784). Inclusion of topics derived from narrative notes allows more accurate discrimination of individuals at high risk for psychiatric readmission in this cohort. Topic modeling and related approaches offer the potential to improve prediction using EHRs, if generalizability can be established in other clinical cohorts.

Absence of evidence for increase in risk for autism or attention-deficit hyperactivity disorder following antidepressant exposure during pregnancy: a replication study.

Multiple studies have examined the risk of prenatal antidepressant exposure and risk for autism spectrum disorder (ASD) or attention-deficit hyperactivity disorder (ADHD), with inconsistent results. Precisely estimating such risk, if any, is of great importance in light of the need to balance such risk with the benefit of depression and anxiety treatment. We developed a method to integrate data from multiple New England health systems, matching offspring and maternal health data in electronic health records to characterize diagnoses and medication exposure. Children with ASD or ADHD were matched 1:3 with children without neurodevelopmental disorders. Association between maternal antidepressant exposure and ASD or ADHD liability was examined using logistic regression, adjusting for potential sociodemographic and psychiatric confounding variables. In new cohorts of 1245 ASD cases and 1701 ADHD cases, along with age-, sex- and socioeconomic status matched controls, neither disorder was significantly associated with prenatal antidepressant exposure in crude or adjusted models (adjusted odds ratio 0.90, 95% confidence interval 0.50-1.54 for ASD; 0.97, 95% confidence interval 0.53-1.69 for ADHD). Pre-pregnancy antidepressant exposure significantly increased risk for both disorders. These results suggest that prior reports of association between prenatal antidepressant exposure and neurodevelopmental disease are likely to represent a false-positive finding, which may arise in part through confounding by indication. They further demonstrate the potential to integrate data across electronic health records studies spanning multiple health systems to enable efficient pharmacovigilance investigation.

Prenatal antidepressant exposure is associated with risk for attention-deficit hyperactivity disorder but not autism spectrum disorder in a large health system.

Previous studies suggested that risk for Autism Spectrum Disorder (ASD) may be increased in children exposed to antidepressants during the prenatal period. The disease specificity of this risk has not been addressed and the possibility of confounding has not been excluded. Children with ASD or attention-deficit hyperactivity disorder (ADHD) delivered in a large New England health-care system were identified from electronic health records (EHR), and each diagnostic group was matched 1:3 with children without ASD or ADHD. All children were linked with maternal health data using birth certificates and EHRs to determine prenatal medication exposures. Multiple logistic regression was used to examine association between prenatal antidepressant exposures and ASD or ADHD risk. A total of 1377 children diagnosed with ASD and 2243 with ADHD were matched with healthy controls. In models adjusted for sociodemographic features, antidepressant exposure prior to and during pregnancy was associated with ASD risk, but risk associated with exposure during pregnancy was no longer significant after controlling for maternal major depression (odds ratio (OR) 1.10 (0.70-1.70)). Conversely, antidepressant exposure during but not prior to pregnancy was associated with ADHD risk, even after adjustment for maternal depression (OR 1.81 (1.22-2.70)). These results suggest that the risk of autism observed with prenatal antidepressant exposure is likely confounded by severity of maternal illness, but further indicate that such exposure may still be associated with ADHD risk. This risk, modest in absolute terms, may still be a result of residual confounding and must be balanced against the substantial consequences of untreated maternal depression.

Using electronic medical records to enable large-scale studies in psychiatry: treatment resistant depression as a model.

BACKGROUND: Electronic medical records (EMR) provide a unique opportunity for efficient, large-scale clinical investigation in psychiatry. However, such studies will require development of tools to define treatment outcome. METHOD: Natural language processing (NLP) was applied to classify notes from 127 504 patients with a billing diagnosis of major depressive disorder, drawn from out-patient psychiatry practices affiliated with multiple, large New England hospitals. Classifications were compared with results using billing data (ICD-9 codes) alone and to a clinical gold standard based on chart review by a panel of senior clinicians. These cross-sectional classifications were then used to define longitudinal treatment outcomes, which were compared with a clinician-rated gold standard. RESULTS: Models incorporating NLP were superior to those relying on billing data alone for classifying current mood state (area under receiver operating characteristic curve of 0.85-0.88 v. 0.54-0.55). When these cross-sectional visits were integrated to define longitudinal outcomes and incorporate treatment data, 15% of the cohort remitted with a single antidepressant treatment, while 13% were identified as failing to remit despite at least two antidepressant trials. Non-remitting patients were more likely to be non-Caucasian (p<0.001). CONCLUSIONS: The application of bioinformatics tools such as NLP should enable accurate and efficient determination of longitudinal outcomes, enabling existing EMR data to be applied to clinical research, including biomarker investigations. Continued development will be required to better address moderators of outcome such as adherence and co-morbidity.

In vitro comparison of conventional film and direct digital imaging in the detection of approximal caries.

OBJECTIVES: To compare the diagnostic accuracy of conventional film, unenhanced direct digital and inversion grayscale direct digital imaging in the detection of approximal caries. METHODS: 150 approximal surfaces of extracted permanent molars and premolars were selected for the study on the basis of varying lesion depth. The teeth were radiographed using Ektaspeed Plus film; digital images were made with a Schick CMOS-APS sensor. 7 examiners evaluated 58 randomized images of each modality. Histological sectioning of the teeth was used to verify the presence and extent of decay. RESULTS: No significant difference was found between the diagnostic accuracies of the three imaging modalities (P=0.226). Analysis of the diagnostic accuracy of the three modalities on lesion depth showed no statistically significant interaction; however, the main effect of the lesion depth was significant (P<0.001, eta(2)=0.936). CONCLUSIONS: The overall diagnostic accuracy of the three modalities in the detection of approximal carious lesions was comparable. All three modalities performed poorly in the detection of enamel lesions.

Structural and functional characterization of B-domain deleted recombinant factor VIII.

A new high-purity recombinant factor VIII preparation has been developed for the treatment of hemophilia A. Structurally, this factor VIII preparation, B-domain deleted recombinant factor VIII (BDDrFVIII), differs from other recombinant and plasma-derived factor VIII preparations in that most of the B-domain has been deleted. To ensure that BDDrFVIII contains the requisite structural and functional features, it has been subjected to detailed biochemical and biophysical characterization in comparison to the plasma-derived form of factor VIII. Laboratory studies have shown that the primary, secondary, and tertiary structures of BDDrFVIII and the posttranslational modifications are similar to those of the [80 + 90]-kd form of plasma-derived factor VIII. In addition, BDDrFVIII has full biologic activity compared with full-length factor VIII preparations.

An analysis of prehospital mortality in an earthquake. Disaster Reanimatology Study Group.

INTRODUCTION: Anecdotal observations about prehospital emergency medical care in major natural and human-made disasters, such as earthquakes, have suggested that some injured victims survive the initial impact, but eventually die because of a delay in the application of life-saving medical therapy. METHODS: A multidisciplinary, retrospective structured interview methodology to investigate injury risk factors, and causes and circumstances of prehospital death after major disasters was developed. In this study, a team of United States researchers and Costa Rican health officials conducted a survey of lay survivors and health care professionals who participated in the emergency medical response to the earthquake in Costa Rica on 22 April 1991. RESULTS: Fifty-four deaths occurred prior to hospitalization (crude death rate = 0.4/1,000 population). Seventeen percent of these deaths (9/54) were of casualties who survived the initial impact but died at the scene or during transport. Twenty-two percent (2/9) were judged preventable if earlier emergency medical care had been available. Most injuries and deaths occurred in victims who were inside wooden buildings (p < .01) as opposed to other building types or were pinned by rubble from building collapse. Autopsies performed on a sample of victims showed crush injury to be the predominant cause of death. CONCLUSIONS: A substantial proportion of earthquake mortality in Costa Rica was protracted. Crush injury was the principal mechanism of injury and cause of death. The rapid institution of enhanced prehospital emergency medical services may be associated with a significant life-saving potential in these events.

Isoenzyme-specific quantitative immunoassays for cytosolic glutathione transferases and measurement of the enzymes in blood plasma from cancer patients and in tumor cell lines.

Enzyme-linked immunoassays (ELISAs) based on the double-antibody sandwich technique have been developed for the quantitative analysis of the major human cytosolic class Pi, Mu and Alpha glutathione transferases (GSTs). The procedures were optimized with respect to antibody concentration for coating of plates as well as other parameters in order to achieve high sensitivity and accuracy. No cross-reactivity was detected between members of the three different classes of GSTs or among the Mu class GSTs M2-2, M3-3 and M4-4 with the ELISA for GST M1-1. The ELISAs have been applied to establish the cytosolic GST profiles of 10 cell lines and to monitor the plasma GST levels in cancer patients. The results revealed that the class Pi GST was the dominant isoenzyme in six (LS 174T, HCT-8, Hu 549 Pat, K-562, U-937 and Hu 549) out of nine tumor cell lines and immortalized hepatocytes (Chang Liver). The isoenzymes A1-1 and M1-1 were determined to be the major GST components in Hep G2 and HeLa cells, respectively. In a clinical study, the majority of the patients with urinary bladder cancer were found to have increased plasma levels of both GST A1-1 and GST P1-1 (10/15), while patients with renal cancer frequently showed increases only in GST P1-1 (5/8). The results demonstrate that the ELISAs are suitable for analyzing GST phenotypes in both normal and tumor cells and in monitoring plasma levels of GSTs in cancer patients.

Cell cycle dependent sensitivity of human melanoma cells to melphalan is correlated with the activity and cellular concentration of glutathione transferases.

Glutathione transferases (GSTs) are enzymes involved in the resistance of tumor cells to bifunctional alkylating cytostatic drugs. We investigated the melphalan sensitivity together with activity and cellular concentration of GST isoenzymes of human melanoma cell line RPMI 8322 in different phases of the cell cycle. By centrifugal elutriation three cell fractions containing different proportions of cells in the G1 phase were isolated. Melphalan sensitivity was estimated by the colony formation assay. The cell fraction with the largest proportion of G1 cells was more sensitive to the drug than the fractions enriched in S and G2 cells. The GST activity of the cell fractions was measured with 1-chloro-2,4-dinitrobenzene (CDNB) as substrate and the concentrations of GST P1-1, GST M1-1 and GST A1-1 were quantitated by use of isoenzyme-specific ELISA. The results show that there were less GST activity and lower GST P1-1 and A1-1 concentrations in the G1 cell enriched fraction, demonstrating a cell cycle dependence of GST expression. Thus, the cell fraction most sensitive to melphalan had the highest proportion of G1 cells and displayed the lowest GST activity, suggesting that the cell cycle dependent sensitivity to melphalan may at least partially depend on the expression of GSTs.

Glutathione analogue sorbents selectively bind glutathione S-transferase isoenzymes.

Novel affinity sorbents for glutathione S-transferases (GSTs) were created by binding glutathione (GSH) analogues to Sepharose 6B. The GSH molecule was modified at the glycine moiety and at the group attached to the sulphur of cysteine. When tested by affinity chromatography in a flow-through microplate format, several of these sorbents selectively bound GST isoenzymes. gamma E-C(Hx)-phi G (glutathione with a hexyl moiety bound to cysteine and phenylglycine substituted for glycine) specifically bound rat GST 7-7, the Pi-class isoenzyme, from liver, kidney and small intestine. gamma E-C(Bz)-beta A (benzyl bound to cysteine and beta-alanine substituted for glycine) was highly selective for rat subunits 3 and 4, which are Mu-class isoenzymes. By allowing purification of the isoenzymes under mild conditions that preserve activity, the novel sorbents should be useful in characterizing the biological roles of GSTs in both normal animal and cancer tissues.

Sensitization of human melanoma cells to the cytotoxic effect of melphalan by the glutathione transferase inhibitor ethacrynic acid.

Glutathione transferases are enzymes implied in the resistance of tumor cells to bifunctional alkylating cytostatic drugs. We have investigated the effect of the glutathione transferase inhibitor by ethacrynic acid on the cytotoxicity of melphalan to a human melanoma cell line (RPMI 8322) with a high level of glutathione transferase activity. Using 1-chloro-2,4-dinitrobenzene as substrate, ethacrynic acid was shown to inhibit the activity of purified human glutathione transferases, with 50% inhibition values of 1, 10, and 15 microM for transferase mu (class mu), transferase epsilon (class alpha) and transferase pi (class pi), respectively, all of which occur in RPMI 8322 cells. Ethacrynic acid at a concentration of 20 microM, which by itself was noncytotoxic, increased the cytotoxicity of melphalan to RPMI 8322 human melanoma cells approximately 2-fold. The induction of DNA interstrand cross-links by 40 microM melphalan was increased 1.4-fold by 30 microM ethacrynic acid. These results indicate that a potentiation of the cytotoxic effect of bifunctional alkylating agents can be achieved by inhibition of glutathione transferase and that the enhanced cytotoxicity may be caused at least in part by increased formation of drug-DNA adducts.

Differences among human tumor cell lines in the expression of glutathione transferases and other glutathione-linked enzymes.

A large number of human tumor cell lines of various origins have been investigated with respect to expression of glutathione-linked enzymes in the cytosol fraction. The amounts of the different enzymes were estimated by use of activity measurements and by silver staining or immunoblot analysis after electrophoresis of cytosol fractions purified by affinity chromatography on S-hexylglutathione Sepharose. Class Pi glutathione transferase was the most abundant enzyme in most tumor cells; the cell lines HepG2 and Raji were exceptions in not expressing significant amounts of this enzyme. HepG2 cells derive from hepatocytes, which normally do not express the class Pi enzyme, whereas Raji cells originate from B-lymphocytes, which normally do express a class Pi glutathione transferase. The highest level of the class Pi transferase, in terms of protein reacting with antibodies as well as enzyme activity, was noted in the colon carcinoma cell line LS174T. Hu549Pat cells, EBV-transformed B-lymphocytes, also expressed high levels of a protein reacting with antibodies specific for class Pi glutathione transferases, but did not display any significant activity with ethacrynic acid, a substrate characteristic for this class. Class Alpha and class Mu glutathione transferases, in cell lines expressing these isoenzymes, were present in significantly lower concentrations than the class Pi enzyme. Most of the tumor cells contained a class Alpha transferase composed of 27.5 kd subunits, which has the physicochemical and immunological properties of the most basic glutathione transferase found in human skin. In several cell lines, a protein was detected with an apparent subunit Mr value of 30 kd that was tentatively identified as an additional class Alpha glutathione transferase not previously described. In addition, other glutathione-linked enzyme activities, namely glutathione peroxidase, glutathione reductase and glyoxalase I, were assayed with specific substrates in the cytosolic fraction of the tumor cells; glyoxalase I could also be estimated semiquantitatively by silver staining of SDS-PAGE cells after affinity chromatography. Like the glutathione transferases, these enzymes displayed distinctly different levels of expression in the various cell lines. Thus, virtually every cell line was found to have a unique pattern of glutathione-linked enzymes, suggesting that the resistance phenotypes of the cells differ accordingly.

[Utero-inguinal hernia]. / Hernia uteroinguinal.

A ten year old boy was operated for left inguinal hernia at age four years. Male gender was confirmed by sex chromatin. Y corpuscle and male genotype in chromosomal studies, and testicular function was tested by normal testosterone serum levels before and after parenteral human chorionic gonadotropin stimulus. On a second operation, two structurally normal testes and was deferens coexisted with Müllerian structures (rudimentary uterus and two Fallopian tubes of normal histological features) were found. A diagnosis of uterus-inguinalis hernia was made. Deficiencies at Müllerian inhibiting factor's secretion, activity or receptors have been postulated to explain this anomaly.

Increase in the amount of glutathione transferase 4-4 in the rat adrenal gland after hypophysectomy and down-regulation by subsequent treatment with adrenocorticotrophic hormone.

The effect of hypophysectomy and subsequent treatment with adrenocorticotropic hormone (adrenocorticotropin, ACTH) on the isoenzymes of glutathione transferase in the rat adrenal gland was investigated. A large increase (approx. 11-fold) in the level of transferase subunit 4 was observed in hypophysectomized animals by immunoblotting. When the activity of glutathione transferase 4-4 was measured in adrenal cytosol using trans-stilbene oxide as a selective substrate, a 15-fold increase was noted. Lack of the pituitary hormone ACTH is apparently related to this increase, since treatment of hypophysectomized animals with ACTH for 2 weeks partially down-regulated subunit 4. Glutathione transferase subunits 3 and 8 in the adrenal were also increased in amount by hypophysectomy, but not at all to the same extent. The activity of glutathione transferase 4-4 was elevated also in the liver and ovary (5 and 1.5 times respectively) after hypophysectomy. These elevated enzyme levels were, however, not affected by ACTH treatment. This down-regulation of glutathione transferases in the rat adrenal by ACTH may be related to the fact that, under normal conditions, this organ is highly susceptible to the toxic effects of various polycyclic hydrocarbons, whereas under circumstances where there is no ACTH production, as in hypophysectomized rats, the adrenal is resistant to these same hydrocarbons.

Glutathione-linked enzymes in normal and tumor cells and their role in resistance against genotoxic agents.

Glutathione is the most abundant low molecular mass thiol in human cells. It is involved in the inactivation of genotoxic electrophilic compounds, and a variety of glutathione-linked enzymes catalyze such detoxication reactions. Within this group, the enzymes occurring in highest intracellular concentrations are the glutathione transferases, which catalyze the detoxication of a broad spectrum of alkylating and oxidizing compounds such as epoxides, reactive alkenes and organic hydroperoxides. Multiple forms of glutathione transferase with distinct substrate specificities exist, and their differential expression in cells contributes to differences in detoxication capacities in tissues. Glyoxalase I catalyzes the inactivation of 2-oxoaldehydes and may also be considered as part of the cellular detoxication system. Characterization of the different enzymes and their differential expression in normal and tumor cells will help to clarify their cellular functions and their significance to human cancer. Clear differences in the occurrence of the various enzyme forms in normal and tumor cells have been demonstrated and variations between different tumors appear to be linked to their degree of resistance to alkylating cytostatic drugs. Modulation of catalytic activities in vitro by administration of enzyme inhibitors may help to overcome this resistance.

Denitrosation of 1,3-bis(2-chloroethyl)-1-nitrosourea by class mu glutathione transferases and its role in cellular resistance in rat brain tumor cells.

1,3-Bis(2-chloroethyl)-1-nitrosourea (BCNU) is known to be detoxified by a denitrosation reaction catalyzed by glutathione-dependent enzymes in rat liver cytosol (R. E. Talcott and V. A. Levin, Drug Metab. Dispos., 11:175-176, 1983). Using a modification of their procedure, we have measured the ability of different purified rat glutathione transferase isoenzymes to denitrosate BCNU. The catalytic efficiencies of the isoenzymes for the denitrosation reaction expressed as the ratio of Vmax to Km were as follows (isoenzyme, Vmax/Km): 1-2, 2.3; 3-3, 12.2; 3-4, 29.2; and 4-4, 26.1. Thus, the class mu isoenzymes containing subunit 4 are by far the best catalysts of the BCNU denitrosation reaction. The class pi transferase 7-7 and class alpha transferases 1-1 and 1-2 demonstrated very weak catalytic activity with BCNU. Determination of the glutathione transferase isoenzyme profiles of 9L rat brain tumor cells and the BCNU-resistant 9L-2 subline by immunoblotting revealed that although the resistant 9L-2 cells contain lower total glutathione transferase activity than 9L cells, they have elevated levels of the class mu transferases. Also, the class pi transferases were found to be down-regulated in 9L-2 as compared with 9L cells. Thus, the increased resistance of 9L-2 cells to BCNU may, in part, be explained by up-regulation of class mu transferase expression with consequent increased capacity for BCNU detoxication. Further support for this hypothesis comes from the fact that pretreatment of 9L-2 cells with the glutathione transferase inhibitors ethacrynic acid or triphenyltin chloride enhanced the cytotoxic effects of BCNU. These results suggest that the class mu transferases play a role in the resistance of brain tumor cells to BCNU.