Changes in aldehyde dehydrogenase during rat urinary bladder carcinogenesis

Transcription

1 Carclnogenesisvol.10no.ll pp , 1989 Changes in aldehyde dehydrogenase during rat urinary bladder carcinogenesis Peggy Campbell 1, Charles C.Irving 2 and Ronald Lindahl 1^ 'Biochemistry Program, Department of Biology, The University of Alabama, Tuscaloosa, AL 3548 and 2 Veterans Administration Medical Center, 1030 Jefferson Avenue, Memphis, TN 38104, USA 'To whom correspondence should be sent at: Department of Biochemistry and Molecular Biology, University of South Dakota School of Medicine, Vermillion, SD 5069, USA We have reported that normal rat urinary bladder possesses significant amounts of an aldehyde dehydrogenase (class 3 ALDH) expressed during hepatocarcinogenesis, but not detectable in normal liver. Changes in expression of both liver and bladder ALDH during JV-butyl-AK4-hydroxybutyl) nitrosamine (BBN)-induced bladder carcinogenesis were studied. Hie ALDH phenotype was determined at intervals over 42 weeks by histochemical analysis, total ALDH activity assays and gel electrophoresis using propionaldehyde and NAD (P-NAD) which characterizes class 1 and 2 ALDH, or benzaldehyde and NADP (B-NADP) to determine class 3 ALDH. By total activity assays and gel electrophoresis, there was a significant decrease in bladder class 3 ALDH activity during weeks Histochemical analysis clearly demonstrates changes in ALDH early in neoplastic development. Intense staining with B-NADP in regions of hyperplasia was first detectable at week 10. Staining in hyperplastic regions was accompanied by a significant decrease in ADLH in neighboring, apparently normal urothelium. As the urothelium became more abnormal, class 3 ALDH activity increased. By week 25, the bladder class 3 ALDH activity of BBN-treated animals was 2 times greater than the control group class 3 ALDH activity. Histochemkally, all papillomas and carcinomas examined possessed class 3 ALDH. However, staining was heterogeneous within the lesions. Bladder neoplasm class 3 ALDH specific activity was greater than control group class 3 ALDH activity in 0% of papillomas and carcinomas. These results suggest events may be occurring in bladder similar to those in liver which alter expression of aldehyde dehydrogenase during carcinogenesis. Introduction We have recently shown that normal rat urinary bladder possesses significant aldehyde dehydrogenase (aldehyde: NAD + oxidoreductase, EC ALDH*) activity (1). Urinary bladder ALDH snares many properties with a tumor-associated (T-ADLH) aldehyde dehydrogenase appearing during rat hepatocarcinogenesis (1). The tumor-associated ALDH has recently been shown to represent a separate class (class 3) of mammalian Abbreviations: ALDH, aldehyde dehydrogenase; T-ALDH, tumor-specific ALDH; BBN, Aif-buryl-A'-<4-hydroxybutyI)nitrosamine; DEN, diethylnitrosamine; DMN, dimethylnitrosamine; BCPN, W-butyl-AH3-carboxypropyl)iutrosamine; BFPN, A r -butyl-a'-<3-formylpropyl)nitrosamine; GGT, gamma-glutamyltranspeptidase. IRL Press ALDHs distinct from cytosolic (class 1) and mitochondrial (class 2) ALDHs (2,3). The class 3 enzymes from various species, including rat and human, appear to be functionally identical (3-5). Like the class 3 ALDH of hepatocarcinogenesis, bladder ALDH is cytosolic, prefers NADP + as coenzyme, and preferentially oxidizes benzaldehyde-like aromatic aldehydes. The electrophoretic mobility of bladder ALDH is identical to that of the class 3 ALDH. Antibodies to class 3 ALDH cross-react with complete identity with bladder ALDH. Thus, these two enzymes appear to be identical (1). We have reported that histochemically urinary bladder ALDH is limited to the very active urothelium and to die inner and outer smootii muscle layers (1). In contrast to the class 3 ALDH, normal rat liver ALDHs preferentially use NAD + as coenzyme, are localized to mitochondria and microsomes, oxidize small aliphatic aldehydes, and are immunochemically distinct (6 11). Normal rat liver lacks T-ALDH activity and no T-ALDH mrna is detectable in normal liver (12). During hepatocarcinogenesis, the class 3 ALDH is first detectable by histochemistry as enzyme-altered foci at a time in hepatocarcinogenesis when other analytical medwds cannot detect significant changes in the ALDH phenotype (13). Class 3 ALDH is also detectable in hyperplastic nodules as well as carcinomas (13). The class 3 ALDH has appeared during rat hepatocarcinogenesis using at least five different protocols (-11). The observation that normal rat urinary bladder possesses an enzyme activity very similar to the one expressed during liver carcinogenesis indicates that expression of the class 3 ALDH gene can be altered during carcinogenesis. Urinary bladder is a useful model system for studying the expression of the class 3 ALDH during bladder carcinogenesis because the pre-neoplastic changes that occur in bladder are well defined (14). There are not, however, as manyreliablebiochemical markers for tumorigenesis in bladder as there are for other tissues such as liver and colon (13). Only one enzyme, NADH diaphorase, has been shown to be helpful in the early identification of pre-neoplastic urinary bladder cells (15). We have shown that diethylnitrosamine (DEN) is very effective in inducing die T-ALDH phenotype during rat hepatocarcinogenesis (). Therefore, to study changes in ALDH during bladder carcinogenesis, we have employed the protocol of Irving et al. (16) which utilizes A/-butyl-A/-(4-hydroxybutyl)nitrosamine (BBN). This protocol is most suitable for the current study for a variety of reasons. The metabolic activation sequence of BBN is well denned (1,18). BBN is a complete carcinogen when administered over several weeks and has carcinogenic potential essentially limited to the bladder. It should offer the greatest probability of inducing changes similar to those induced by DEN during hepatocarcinogenesis. The purpose of this study was to determine if changes in ALDH activity occur during bladder carcinogenesis and to determine if there are events occurring in bladder similar to those of liver which alter the expression of the class 3 ALDH gene during carcinogenesis. 2081

2 P.Campbell, C.C.Irving and R.LJndahl Materials and methods Chemicals Aldehydes were from Aldrich Chemical Co. (Milwaukee, WI). Propionaldehyde and benzaldehyde were redistilled before use. Dimethylsulfoxide, NAD +, NADP +, EDTA, 2-mercaptoethanol, collagenase, trypsin, polyvinylalcohol, nitrobiuetetrazouum, phenazine methosulfate and Triton X-100 were from Sigma Chemical Co. (St Louis, MO). Hematoxylin and eosin were from Fisher Scientific (Fair Lawn, NJ), OCT compound was from Lab-Tek Products, Inc. (Naperville, IL), BBN was from CTC Organics (Atlanta, GA) and Af-butyl-^J-carboxypropyl)nhrosamine (BCPN) was synthesized by the method of Irving (1). Bladder carcinogenesis protocol The urinary bladder carcinogenesis protocol of Irving et al. was employed (16). Male Wistar rats (Charles River, Wilmington, MA) weighing g were used. The animals were randomly separated into two groups: group 1 (100 rats), 0.025% BBN in the drinking water ad libinun for 15 weeks, then BBN-free drinking water for the duration of the protocol; and group 2 (42 rats), untreated drinking water ad libitum throughout. Both groups were maintained on Purina 2082 Rodent Laboratory Chow no ad libitum for the duration of the experiment. At 5-week intervals, animals were killed by suffocation as a result of CO2 sublimation in a large container. The bladders were inflated with 0.5 ml 50% (v/v) OCT cryogenic tissue-embedding compound in H2O. The livers and inflated bladders were removed and rinsed in cold 60 mm phosphate buffer, ph 8.5, containing 1 mm EDTA and 1 mm 2-mercaptoethanol. They were examined for any gross abnormalities. Each inflated bladder and a piece of liver were embedded in OCT compound for histochemical analysis. The remainder of the liver was quick frozen in a solid CO2-acetone bath and stored at -80 C. Tissue preparation All procedures were performed at 0-4 C. Livers were prepared as 33% homogenates in the 60 mm phosphate buffer system, ph 8.5. Homogenates were made to 1% with Triton X-100 (final concentration v/v) and incubated at 0 C for 30 min. The solubilized homogenates were centrifiiged at g for 30 min, and the supematants used for ALDH determinations. After cutting representative sections from each bladder, the remaining tissue was thawed on ice, removed from OCT and rinsed in the 60 mm phosphate buffer. To free urothelial cells from the underlying stroma, the bladders and, when size Fig. 1. Histochemical localization of ALDH during urinary bladder carcinogenesis. (A) Activity pattern with P-NAD in liver of BBN-treated animals; (B) activity pattern with B-NADP in BBN-treated liver; (C) activity pattern with B-NADP in control urinary bladder; (D) activity pattern with B-NADP in 10 week BBN-treated bladder showing papillary hyperplasia; (E) activity pattern with B-NADP of large papilloma appearing in BBN-treated urinary bladder at week 42. (A,B) x 18; ( C - E ) x 250.

3 Changes in aldehyde dehydrogeoase O 253O Week of Experiment Fig. 2. Changes in liver ALDH during BBN-induced urinary bladder carcinogenesis., P-NAD: control; +, B-NADP: control; *, P-NAD: 0.025% BBN; +, B-NADP: 0.025% BBN. Each point is the average of at least three determinations. allowed, individual gross lesions were placed in 200 ji\ of 0.25% trypsin and 0.30% collagenase as suggested by Roszell et al. (19) in the 60 mm phosphate buffer. The resulting epithelial cell populations were gently homogenized and the homogenates centrifuged at g in a microcentnfuge for 5 min. The resulting supernatants were used for bladder ALDH determinations. ALDH determination ALDH activity was assayed at 25 C by monitoring the change in A^ caused by NADH or NADPH production during die oxidation of aldehyde substrate as described (3). Propionaldehyde and NAD + were used to determine basal (class 1 and 2) liver ALDH activity and benzaldehyde and NADP + were used to determine class 3 ALDH activity. Protein concentrations were determined by the method of Lowry el al. (20). Purification of class 3 ALDH Class 3 ALDH was purified from rat cornea using the method of Undahl et al. (21). Electrophorctic analyses SDS-PAGE was performed as described by Laemmli (22). Electrophoretic transfer of proteins was by the method of Burnette (23). Anti-class 3 ALDH antibodies were those produced by Harper et al. (24). Non-denaturing PAGE was done as described previously (25). Histochemistry For each inflated bladder or piece of liver, 5 fim frozen sections were cut at 20 C. Three approximately serial sections were prepared per slide and allowed to dry at room temperature. For liver, the serial sections were stained for ALDH with propionaldehyde-nad + and benzaldehyde-nadp + at the substrate and coenzyme concentrations used previously (13), and/or -glutamyltranspeptidase (GGT) (13). For bladder, the three sections were stained with propionaldehyde NAD + benzaldehyde NADP + and no substrate or no coenzyme respectively. The histochemical localization of ALDH activity was performed in the same manner as liver with some modifications. The final aldehyde substrate concentrations were increased from 10 to 100 mm. Twenty-two per cent polyvinylalcohol was added to 60 mm phosphate buffer to decrease the diffusion of the reduced formazan into the staining medium as suggested by Chieco et al. (26). Sections of bladder were incubated for 40 min in the dark. Histology Histopathological examinations were performed on hematoxylin and eosin-stained serial cryostat sections. The classification of lesions observed followed the established criteria for liver (2) or bladder (14). Results No significant differences were observed in liver or body weight gain between the two groups during the experiment. The liver/body ratio in animals maintained on water containing BBN for 15 weeks was not significantly different from the ratio of the control animals. It was not possible to weigh bladders as they were injected and inflated in situ. ALDH activity of liver The livers of animals receiving 0.025% BBN were grossly identical to the livers of control animals throughout the 42-week experiment. No foci, neoplastic nodules or other lesions were observed in H&E-stained liver sections from BBN-treated or control animals. Histochemically, no significant qualitative changes in liver ALDH were observed during the experiment. The staining pattern in BBN-treated livers was identical to that of the control livers (Figure 1A and B). There were no GGT-positive foci observed at any time throughout the experiment. Every 5 weeks, when the animals were killed, liver ALDH specific activity was determined (Figure 2). There was no significant change in normal liver or class 3 ALDH activity in liver throughout the experiment. Two additional methods were used to determine the ALDH phenotype of liver. No evidence for class 3 ALDH was seen using non-denaturing gel electrophoresis followed by staining for ALDH activity using propionaldehyde and NAD + or benzaldehyde and NADP + in either BBN-treated or control livers. Western blot analysis using antibodies to class 3 ALDH also indicated no class 3 ALDH in either experimental or control livers (data not shown). ALDH activity in urinary bladder Areas of hyperplasia were first seen at 10 weeks in the urothelium of BBN-treated animals. Throughout the remainder of the experiment a wide spectrum of urothelial morphologies was observed. These included normal urothelium, simple hyperplasia, microscopic papillary growths, grossly visible papillomas and invasive carcinomas (Figure 1). Histochemically, the urothelium in control rats stained readily for ALDH activity using propionaldehyde NAD + and benzaldehyde NADP + (Figure 1C). The staining patterns in the 2083

4 P.CampbeU, C.C.Irving and R.LJndahl! o a a) Week of Experiment Fig. 3. Changes in urinary bladder activity during BBN-induced bladder carcinogenesis., P-NAD: control; +, B-NADP: control;, P-NAD: 0.025% BBN; +, B-NADP: 0.025% BBN. Each point is the average of the values of the determinations as shown in Table I. Table I. Correlation between various probes for class 3 ALDH during BBN-induced bladder carcinogenesis Week of experiment Control No. of animals Histochemistry 5 Class 3 ALDH sp. act. (mlu/mg of protein) Histological classification Hyperplasia Papilloma Carcinoma *+ for control bladder histochemistry indicates uniform staining of entire bladder urothelium., / +, or + for BBN-treated bladders indicates focal or lesionspecific staining superimposed on the ALDH-negative background observed at weeks See text for details. b Simple and papillary hyperplasia. c Average control group bladder class 3 ALDH specific activity. control animals were consistent for the duration of the experiment. At week 10, the urothelium of BBN-treated animals showed areas of intense focal staining for class 3 ALDH in regions associated with simple or papillary hyperplasia (Figure ID). In these areas, the staining occurred 3 4 times faster than in control bladder urothelium. Along with an increase in staining in hyperplastic regions, there was a significant decrease in stain intensity in the neighboring, apparently normal urothelium. This difference in intensity was often quite dramatic (Figure ID). Increased ALDH activity concomitant with increased simple hyperplasia and papillary hyperplasia continued throughout the experiment. In small papillomas, the ALDH activity detected histochemically was comparable to that of areas of hyperplasia. However, by week 25, as the papillomas increased in size, marked heterogeneity in ALDH staining was seen both within a single lesion as well as between different lesions. Some areas within a single large papilloma or carcinoma possessed very 2084 strong ALDH activity, whereas other areas had reduced ALDH activity, occasionally less than normal urothelium (Figure IE). H&E-stained serial cryostat sections indicated that for some, but not all papillomas and carcinomas, the heterogeneity in ALDH activity was due to a large ALDH-negative stromal contribution to the tumor. Urinary bladder ALDH-specific activity was also determined whenever the animals were killed (Figure 3). There was no significant change in class 1, 2 or 3 ALDH activity of control bladders throughout the experiment (Figure 3). In BBN-treated animals, the bladder class 3 ALDH specific activity reflected the changes observed histochemically (Figure 3 and Table I). In the first 15 weeks, class 3 ALDH specific activity was significantly lower in the urothelium of BBN-treated animals than the ALDH activity of control bladders. However, by week 21, class 3 specific activity in the urothelium of BBN-treated animals exceeded bladder class 3 ALDH activity of control animals. As

5 Changes in aldehyde dehydrogenase Table D. Effect on BCPN on class 3 ALDH Treatment1 Class 3 ALDH sp. act. Control-P/NAD Control -B/NADP BCPN-P/NAD BCPN -B/NADP "Class 3 ALDH pre-incubated 5 min in assay buffer containing 2 mm BCPN in the presence of the substrate or coenzyme and the reaction started by the addition of coenzyme or substrate as necessary. than class 3 ALDH activity of the control group in 0% of papillomas and carcinomas, while 50% of the papillomas and carcinomas had greater specific activity than the neighboring host urothelium (Figure 4). There was no correlation between the size or histological classification of a lesion and its ALDH activity. Papilloma and carcinoma class 1 and 2 ALDH activities were not significantly different from host or control group activities (data not shown). A single isozyme of ALDH was detected in BBN-treated and control bladders using non-denaturing gel electrophoresis. Western blot analysis using antibodies to the class 3 ALDH indicated that this enzyme is class 3 ALDH (Figure 5). All papillomas and carcinomas examined cross-reacted with antibodies to the class 3 ALDH (data not shown). Discussion The results reported here indicate that significant changes in ALDH occur during urinary bladder carcinogenesis. This is the first report of transformation-associated changes in ALDH in an organ other than liver. Interestingly, the progression of changes in ALDH expression in urinary bladder is different from that seen in liver. In previous studies of class 3 ALDH expression during rat hepatocarcinogenesis, class 3 ALDH, which is not detectable in normal liver, appeared to be tumor specific (-11). By histochemistry, well defined focal populations of hepatocytes expressing the phenotype were identified long before the development of overt neoplasms (,8,11). The phenotypic change was limited to the neoplasm with the neighboring host tissue not possessing the class 3 ALDH phenotype ( 11). I a i I O 34 3b % Animal Nnmh«r Fig. 4. Distribution of class 3 ALDH activity during BBN-induced bladder carcinogenesis. Activity is considered significantly elevated or reduced if it is 2 SD from the appropriate mean control bladder class 3 ALDH activity (using B-NADP) of 24.0 (horizontal line). Host, grossly normal appearing bladder; lesions I, 2, grossly observed papillomas or carcinomas large enough to be separated from host bladder and individually analyzed for ALDH activity the urothelium became more abnormal, class 3 ALDH specific activity continued to increase. By week 25, class 3 ALDH activity was 2 times greater than bladder class 1 and 2 ALDH activity (measured using propionaldehyde NAD+), which was not significantly different from controls (Figure 3 and Table I). Irving (14) recently reported that BBN is oxidized by rat liver class 1 or 2 aldehyde dehydrogenase to BCPN via N-butyl-A/(3-formylpropyl)nitrosamine (BFPN), which contains a primary aldehyde group. In this activation scheme, BCPN is the major metabolic product of BBN that reaches the bladder. To determine if the decrease in ALDH activity observed between weeks 5 and 15 was due to direct inhibition of ALDH enzyme by BCPN, the effect of BCPN on purified class 3 ALDH activity was determined (Table II). There was no significant inhibition of ALDH by BCPN. When papillomas and/or carcinomas were large enough, they were prepared for ALDH specific activity determination. Class 3 ALDH specific activity in BBN-treated animals was greater

6 P.Campbell, C.C.Irvlng and R.Lindahl In bladder, class 3 ALDH activity is normally present. However, histochemical analysis clearly demonstrates changes in bladder class 3 ALDH activity before such changes are detectable by total activity assays. ALDH activity decreases dramatically after initiation except in areas containing putative pre-neoplastic bladder cells, which then appear as ALDH-positive areas of hyperplasia. Thus, focal populations of putative initiated cells, hyperplastic regions and ALDH-positive neoplasms appear against an ALDH-negative background. The decrease in the bladder ALDH in normal urothelium is quite dramatic, suggesting class 3 ALDH is the major isozyme present in normal urinary bladder. As hyperplastic areas progress to papillomas and carcinomas, class 3 ALDH specific activity increases to ~ 2 times the basal class 3 ALDH activity. A similar increase in activity is also seen in liver neoplasms, but in liver the increase is as much as 10-fold (-11). Although areas of hyperplasia display homogeneous ALDH staining patterns, larger bladder papillomas and carcinomas show marked heterogeneity. This is consistent with earlier studies of liver ALDH ( 11) and a bladder carcinogenesis study using NADH-diaphorase (15). For some bladder lesions, histochemical heterogeneity in enzyme expression appears to be the result of the progression of the neoplasms to lesions containing subpopulations of urothelial cells which have variable altered ALDH gene expression. For other lesions, the heterogeneity may be due in part to the sometimes pronounced stromal contribution of certain bladder neoplasms. The histochemically observed ALDH heterogeneity is also reflected in the marked variability in class 3 ALDH specific activity noted in different neoplastic lesions within a single bladder and the variability noted between various host bladders. Since the urothelium of lesions was separated from the stroma by trypsin treatment during sample preparation, it appears likely that the specific activity differences seen are indicative of the presence of sub-populations of urothelial cells with differing levels of ALDH gene expression. This is also consistent with earlier reports for liver ALDH ( 11). The heterogeneity in host bladder class 3 ALDH is likely due to the fact that many bladders with very large lesions also possessed numerous smaller neoplastic and pre-neoplastic lesions. Therefore, the bladder tissue remaining after removal of large lesions from bladders containing 2086 That there was no change in the ALDH phenotype of liver during bladder carcinogenesis suggests that expression of class 3 ALDH during bladder tumorigenesis is a result of bladderspecific changes in ALDH gene expression. As is the case during hepatocarcinogenesis, changes in bladder class 3 ALDH activity appear to be transformation-associated, stable and limited to the target organ. Moreover, since no changes in liver ALDH activity were detected in the present study and since class 3 ALDH is not detectable in normal rat liver at either the protein or mrna levels (12), it appears that the class 3 ALDH is not involved in the production of the proximate carcinogen BCPN. Since BCPN is the metabolite of BBN active in urinary bladder, and since class 3 ALDH was not directly inhibited by BCPN, the decrease in class 3 ALDH activity appears to be due to a BCPN-induced alteration in class 3 ALDH gene expression. Whether the effect is at the transcriptional or translational level is currently unknown. We have recently suggested that there may be a role for class 3 ALDH in the oxidation of aldehydes generated by lipid metabolism including lipid peroxidation (29). We proposed that class 3 ALDH may function in this role in both physiological and pathophysiological conditions such as tumorigenesis. This may explain the presence of class 3 ALDH in normal urinary bladder. One of the results of lipid metabolism is the formation of cytotoxic aldehydes which are stable, long-lived and highly diffusable. Even though much of the metabolism of these aldehydes takes place in the liver (29), some of the more stable forms may be transported to and stored in the urinary bladder prior to excretion. Class 3 ALDH is localized to the bladder epithelium and is therefore present in cells which line the lumen of the bladder. The normal high activity of class 3 ALDH in bladder may be in response to the need to oxidize these aldehydes. There may also be a role for ALDH in protecting initiated cells from the cytotoxic effects of aldehydes produced by carcinogen and/or promoter metabolism or from alterations in endogenous aldehyde metabolism caused by initiation or early promotion events. For example, for many nitrosamines, including BBN, DEN and DMN, aldehydes are produced during their metabolism (30,31). The acetaldehyde produced from DEN and butyl-n- Fig. 5. Western blot of ALDH at various times during bladder carcinogenesis. Bladder ALDH (lanes 1-18, 15 fig) and purified class 3 ALDH (lanes 19, 2.5 ng) were analyzed by Western blot analysis following SDS-PAGE (20). Anti-class 3 ALDH was the first antibody (21). The second antibody was a goat anti-rabbit IgG peroxidase conjugate. Lanes 1-4, BBN-treated bladders; lanes 5-1 8, lesions removed from BBNtreated bladders. numerous neoplasms is not completely uninvolved, ALDHnormal urothelium, but may still contain many regions with elevated or reduced class 3 ALDH activity. The early reduction in ALDH activity, resulting in the absence of background staining, suggests that class 3 ALDH could be a particularly useful histochemical marker for urinary bladder carcinogenesis. The only other reliable early marker for urinary bladder in the transformation process is NADH-diaphorase. However, for diaphorase there is much more background staining (15). In the present study, some bladder sections were stained for GGT, but no activity was seen during the 42-week experiment, confirming an earlier report (15). Other markers that have been used for urinary bladder with limited success include alkaline phosphatase and succinic dehydrogenase (28). Among all these, class 3 ALDH is the only enzyme marker for bladder that shows the great contrast between normal urothelium and regions of early hyperplasia. However, before ALDH can be considered a biochemical marker of bladder carcinogenesis, it must be demonstrated that the changes observed are not related simply to increased urothelial cell proliferation. For liver, it has been demonstrated that class 3 ALDH expression during tumorigenesis is not related to increased hepatocyte proliferation due to liver regeneration following partial hepatectomy (11) or to re-expression of a fetal liver ALDH (25).

7 Changes in aldehyde dehydrogenase (3-formylpropyl)nitrosamine produced from BBN are excellent substrates for the class 1 and 2 ALDHs (30,1). The formaldehyde produced by DMN is the substrate for a formaldehydespecific ALDH (32). Benzaldehyde from JV-nitroso-iV-methylbenzylamine oxidation would be a preferred substrate for class 3 ALDH (33). Further, among the mammalian ALDHs, only the class 3 enzyme and a phenobarbital-responsive form are inducible by xenobiotics (34). This suggests they play a role in the oxidation of aldehydes produced by xenobiotic metabolism or effects. If certain pre-neoplastic or early neoplastic cells can more effectively deal with any aldehydes generated because of increased ALDH activity, due to response-modification for example (35), they may have a selective growth advantage over neighboring cells. Subsequent clonal expansion of these cells would result in an ALDH-expressing neoplasm. Acknowledgements This work was supported by Grant no. CA to R.L. from the National Cancer Institute. References 1. Lindahl.R. (1986) Identification of hepatocarcinogenesis-associated aldehyde dehydrogenase in normal rat urinary bladder. Cancer Res., 46, HempeU., Harper.K. and Lindahl.R. (1988) Inducible (class 3) aldehyde dehydrogenase from rat hepatocellular carcinoma and 2,3,,8-tetrachlorodibenzo-p-dioxin-treated liver: distant relationship to the class 1 and 2 enzymes from mammalian liver cytosol/mitochondria. Biochemistry, 28, Anonymous (1989) Nomenclature of mammalian aldehyde dehydrogenases. In Weiner,H and Flynn.G.T. (eds), Enzymology and Molecular Biology of Carbonyl Metabolism II. Alan R.Liss, New York, p. xix-xxi. 4. Meier-Tackmann.D., Agarwel.D.P, Saha,N. and Goedde.H W. (1984) Aldehyde dehydrogenase isozymes in stomach autopsy specimens from Germans and Chinese. Enzyme, 32, Santisteban.I., Povey.S., West.L.F., Parrington,J.M. and Hopkinson.D.A. (1985) Chromosome assignment, biochemical and immunological studies on a human aldehyde dehydrogenase, ALDH3. Ann. Hum. Genet., 49, Lindahl.R. and Evces,S. (1984) Rat hver aldehyde dehydrogenase. I. Isolation and characterization of four high Km normal liver isozymes. J. Biol. Chem., 259, Lindahl.R. and Evces.S. (198) Changes in aldehyde dehydrogenase activity during diethylnitrosamine-initiated rat hepatocarcinogenesis. Carcinogenesis, 8, Evces.S. and Lindahl,R. (1986) Changes in aldehyde dehydrogenase occurring during rat hepatocarcinogenesis induced by ethionine combined with dietary choline deficiency. Cancer Res., 46, Wischusen,S.M., Evces,S. and Lindahl.R. (1983) Changes in aldehyde dehydrogenase activity during diethylnitrosamine- or 2-acetylaminofluoreneinitiated rat hepatocarcinogenesis. Cancer Res., 43, Lindahl.R., Evces.S. and Sheng.W.L. (1982) Expression of the tumor ALDH phenotype during 2-acetylaminofluorene-induced rat hepatocarcinogenesis. Cancer Res.. 42, Jones,D.E., Evces.S. and Lindahl.R. (1984) Expression of tumor-associated aldehyde dehydrogenase during rat hepatocarcinogenesis using the resistant hepatocyte model. Carcinogenesis, 5, Lin,K.-H., Brennan.M.D. and Lindahl.R. (1988) Expression of tumorassociated aldehyde dehydrogenase gene in rat hepatoma cell lines. Cancer Res.. 48, Lindahl.R., Clark,R. and Evces,S. (1983) Histochemical localization of aldehyde dehydrogenase during rat hepatocarcinogenesis. Cancer Res., 43, Akaza,H., Murphy,W. and Soloway,M. (1984) Bladder cancer induced by non-carcinogenic substances. J. Urol., 131, Vanderlaan.M., Fong,S. and King.E. (1982) Histochemistry of NADH diaphorase and gamma-glutamyltranspeptidase in rat bladder tumors. Carcinogenesis, 3, Irving.C, Tice.A. and Murphy.W. (199) Inhibition of W-butyl-AH4-hydroxybutyl)nitrosamine-induccd urinary bladder cancer in rats by administration of disulfiram in the diet. Cancer Res., 39, Irving.C.C. (1988) Oxidation of N-butyl-A'-(3-formylpropyl)nitrosamine to A'-butyl-A'-(3-carboxypropyl)nitrosarnine in rat liver and inhibition by disulfiram. Carcinogenesis, 9, Irving.C.C. and Daniel,D.S. (198) Influence of disulfiram on the metabolism of the urinary bladder carcinogen (V-butyl-/V-{4-hydroxybutyl)nitrosamine in the rat. Carcinogenesis, 8, RoszelM.A., Douglas.C.J. and Irving.C.C. (19) Polyamine-stimulated growth of cultured rat urinary bladder epithelial cells. Cancer Res., 3, Lowry.O.H., Rosebrough.N.J., Farr.A.L. and Randall.R.J. (1951) Protein measurement with the Folin reagent. J. Biol. Chem., 193, Lindahl.R., Baggett.D.W. and Winters,A.L. (1985) Characterization of aldehyde dehydrogenase from HTC rat hepatoma cells. Biochim. Biophys. Ada, 843, Laemmh.U.K. (190) Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Namre, 22, Bumette.W.N. (1981) 'Western blotting': electrophoreuc transfer of proteins from sodium dodecyl sulfate-polyacrylamide gels to unmodified nitrocellulose and radiographic detection with antibody and radioiodinated protein A. Anal. Biochem., 112, Harper.K., Jones,D.E, Brennan.M.D. and Lindahl.R. (1988) Characterization of a functional recombinant rat liver aldehyde dehydrogenise: expression as a non-fusion protein in E.coli. Biochem. Biophys. Res. Commun., 152, Lindahl.R (19) Aldehyde dehydrogenase in 2-acetyl-aminofluorene-induced rat hepatomas. Ontogeny and evidence that the new isocnzymes are not due to normal gene derepression. Biochem. J., 164, Chieco,P., Normanni.P., Moslen.M.T. and Maltoru.C (1986) Quantitative histochemistry of benzaldehyde dehydrogenase in hepatocellular carcinomas of vinyl chloride-treated rats J. Histochem. Cytochem., 34, Stewart.H.L. (1980) Histologic typing of liver tumors of the rat. J. Natl. Cancer Inst., 64, Kunze,E. and Schauer,A. (191) Enzymhistochemische und autoradiographische untersuchungen an dibutylnitrosamin-induzierten harnblasenpapillomen der ratt. Z Krebsforsch., 5, Marselos.M. and Lindahl.R. (1988) Substrate preference of a cytosolic aldehyde dehydrogenase inducible in rat liver by treatment with 3-methylcholanthrene. Toxicol. Appl. Pharmacol., 95, Magour,S. and Nievel,J.G. (191) Effects of inducers of drug metabolizing enzymes on diethylnitrosamine metabolism and toxicity. Biochem. J., 123, Crafstrom,R.C, Curren,R.D., Yang.L L. and Harris.C.C. (1985) Genotoxicity of formaldehyde in cultured human bronchial fibroblasts. Science, 228, Uotila.L and Koivusalo.M. (194) Formaldehyde dehydrogenase from human liver. Purification properties and evidence for the formation of glutathione thiol esters by the enzyme. J. Biol. Chem., 249, Schweinsberg.F., Danecki.S., GrotzkeJ., Von Karsa,L. and Burkle,V. (1986) Modifying effects of disulfiram on DNA adduct formation and persistence of benzaldehyde in A'-nitroso-A'-methylbenzylamine-induced carcinogenesis. J. Cancer Res. Clin. Oncol., 112, Simpson,V.J., Baker.R. and Deitnch.R.A. (1985) Inducible aldehyde dehydrogenases in rat liver cytosol. Toxicol. Appl. Pharmacol., 9, Cerrutti.P.A. (1988) Response modification creates promotability in multistage carcinogenesis. Carcinogenesis, 9, Received on May 16, 1989: revised on August 14, 1989; accepted on August 22,

8