Tumor-associated Aldehyde Dehydrogenase Research Articles

Abstract 1991: Vulnerability of platinum-resistant ovarian cancer to FAK inhibition

Abstract Platinum (CP)-resistant ovarian cancer (OC) has few effective treatment options. Adaptive chemotherapy resistance occurs in part through cancer stem cell (CSC) generation. Wnt/beta-catenin signaling is a driver of CSC survival via induction of gene expression, including aldehyde dehydrogenase (ALDH) enzymes. The gene for focal adhesion kinase (FAK) is commonly amplified in advance-stage OC and this is associated with decreased patient survival. How FAK is activated in OC and whether this is connected to CSC survival is unknown. Here, we find that FAK Y397 phosphorylation (a marker of FAK activation) is increased in non-necrotic Pax8-positive OC tumor tissue after neo-adjuvant chemotherapy compared to paired pre-treatment tumor biopsy samples. FAK activation occurs after CP plus paclitaxel treatment of xenograft tumors, within 60 min of CP treatment of OC cells, and FAK Y397 levels are constitutively-elevated in OC cells pre-adapted to exhibit elevated CP resistance. Pharmacological FAK inhibition (VS-4718, 0.1 µM) selectively prevented CP-resistant tumorsphere formation in vitro and combined with CP to promote cell apoptosis. Combinatorial VS-4718, CP, and paclitaxel chemotherapy exhibited additive inhibitory effects in preventing CP-resistant tumor growth in mice. VS-4718 monotherapy of tumor-bearing mice reduced tumor-associated ALDH activity and resulted in an 85-fold reduction in secondary tumors formed in limiting dilution assays. CRISPR-mediated FAK knockout in OVCAR3 cells combined with stable re-expression of FAK wildtype or a kinase-inactive FAK (K454R) mutant revealed that intrinsic FAK activity was essential for beta-catenin activation, ALDH-1A1 expression, and OVCAR3 tumorsphere growth. As activated constructs of beta-catenin but not YAP1 rescued FAK KO OVCAR3 phenotypes, these studies provide important insights into a FAK signaling linkage to beta-catenin in promoting CSC survival and adaptive resistance to CP chemotherapy. These studies provide the foundational support for a Phase I-II clinical trial for treatment of recurrent platinum-resistant ovarian cancer (NCT03287271) termed ROCKIF: Re-sensitization of carboplatin-resistant Ovarian Cancer by Kinase Inhibition of FAK. As FAK is activated in tumor cells surviving carboplatin-paclitaxel chemotherapy, co-targeting of this FAK/beta-catenin adaptive resistance pathway may expose a vulnerability of CP-resistant tumors. Citation Format: Carlos J. Diaz Osterman, Lisa M. Bean, Florian J. Sulzmaier, Kristin N. Taylor, Shulin A. Jiang, Isabelle Tancioni, Kristen Anderson, Christine Jean, Xiao Lei Chen, Elizabeth G. Kleinschmidt, Vihren N. Kolev, David T. Weaver, Jonathan A. Pachter, Denise C. Connolly, Alfredo Molinolo, David D. Schlaepfer. Vulnerability of platinum-resistant ovarian cancer to FAK inhibition [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 1991.

Abstract 3812: FAK inhibition resensitizes platinum-resistant serous ovarian cancer

Abstract Platinum and taxol administration is standard of care chemotherapy for serous ovarian cancer. Tumor recurrence occurs in a high percentage of patients, and this is directly related to poor overall survival. Ovarian cancer stem cell (CSC) resistance to chemotherapy treatment can give rise to tumor recurrence. Focal adhesion kinase (FAK), an intracellular tyrosine kinase, has been linked to mesothelioma and breast CSC survival. Here, we find that FAK activation is elevated in platinum (CP)-resistant ovarian cancer cells and that FAK tyrosine phosphorylation is increased after CP treatment of CP-sensitive ovarian cancer cells. Nanomolar levels of FAK inhibitor (VS-4718) selectively blocked CP-resistant ovarian carcinoma methylcellulose colony growth via cell cycle inhibition but not apoptosis. Oral VS-4718 administration to mice reduces CP-resistant orthotopic tumor burden with a concomitant decrease in tumor-associated aldehyde dehydrogenase (ALDH) activity, a marker of ovarian CSCs. Residual ovarian tumor cells from VS-4718-treated mice exhibit reduced ALDH activity and secondary tumor initiating capacity. CRISPR-mediated FAK knockout or VS-4718 treated ovarian carcinoma cells exhibit diminished Oct-4 transcription factor and ALDH-1A1 CSC-associated protein marker expression. Co-administration of VS-4718 with CP-taxol chemotherapy reduced CP-resistant tumor burden and exhibited additive inhibitory effects on ovarian carcinoma spheroid growth. As we find that CP activates FAK and that FAK activity sustains ovarian carcinoma CSC phenotypes, our results support the testing of FAK inhibitors in combination with CP to prevent recurrent and chemo-resistant ovarian cancer. Citation Format: Lisa M. Bean, Florian J. Sulzmaier, Isabelle Tancioni, Sean Uryu, Christine Jean, Christine Lawson, Xiao Lei Chen, Elizabeth G. Kleinschmidt, Kristen M. Anderson, Edward A. Cordasco, Joshua Axelrod, Vihren N. Kolev, Jonathan A. Pachter, Dwayne G. Stupack, David D. Schlaepfer. FAK inhibition resensitizes platinum-resistant serous ovarian cancer. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 3812.

Primary structure of human hepatocellular carcinoma-associated aldehyde dehydrogenase

Tumor-associated aldehyde dehydrogenase (T-ALDH) is strongly expressed in hepatocellular carcinoma (HCC) but undetectable in normal liver. In the present study, this enzyme from human HCC, HCC T-ALDH, was purified and the partial amino acid sequences (384 residues) determined by direct protein sequencing matched the amino acid sequence (453 residues) deduced from cloned HCC T-ALDH cDNAs with an open reading frame. The coding sequences of HCC T-ALDH cDNA, human stomach ALDH3A1 cDNA [Hsu et al., J. Biol. Chem. 267 (1992) 3030–3037] and human squamous cell carcinoma (SCC) T-ALDH cDNA (Schuuring et al., GenBank I.D. M74542) matched one another except for discrepancies at four positions, with consequent P12R, I27F and S134A substitutions. R and A were found in HCC and SCC T-ALDHs, whereas P and S were present in stomach ALDH3A1. To confirm that these discrepancies would have general occurrence, coding sequences of HCC T-ALDH cDNAs from six patients and stomach ALDH3A1 cDNAs from two individuals were examined and all were found to encode ALDH3A1 having R, I and A at protein positions 12, 27 and 134, respectively, indicating HCC T-ALDH to be variant ALDH3A1 which is common in human stomach tissues.

Aryl hydrocarbon receptor-associated genes in rat liver: Regional coinduction of aldehyde dehydrogenase 3 and glutathione transferase Ya

The tumor-associated aldehyde dehydrogenase 3 (ALDH3) and the glutathione transferase (GST)Ya form are coded by members of the Ah (aryl hydrocarbon) battery group of genes activated in the liver by polycyclic hydrocarbons such as 2,3,7,8-tetrachlorodibenzo- p-dioxin (TCDD). The physiological role of the Ah receptor (AHR), its gene-activating mechanism and its endogeneous ligands are still poorly clarified. We had previously observed that 3-methylcholanthrene (3MC) and β-naphthoflavone (βNF) induced the AHR-associated CYP1A1/1A2 pair in different liver regions, an effect not explained by the acinar distribution of the AHR protein. Here, we investigated AHR-associated regional induction by comparing the expression patterns of ALDH3 and GSTYa. Analysis of samples from periportal and perivenous cell lysates from 3MC-treated animals revealed that ALDH3 mRNA, protein and benzaldehyde-NADP associated activity were all confined to the perivenous region. In contrast, such regio-specific induction was not seen after β-NF induction. Immunohistochemically, a peculiar mono- or oligocellular induction pattern of ALDH3 was seen, consistently surrounding terminal hepatic veins after 3MC but mainly in the midzonal region after βNF. A ligand-specific difference in regional induction of GSTYa1 mRNA was also observed: The constitutive perivenous dominance was preserved after 3MC while induction by βNF was mainly periportal. A 3MC-βNF difference was also seen by immunohistochemistry and at the GSTYa protein level, in contrast to that of the AHR-unassociated GSTYb protein. However, experiments with hepatocytes isolated from the periportal or perivenous region to replicate these inducer-specific induction responses in vitro were unsuccessful. These data demonstrate that the different acinar induction patterns by 3MC and βNF previously observed for CYP1A1 and CYP1A2 are seen also for two other Ah battery genes, GSTYa1 and ALDH3, but in a modified, gene-specific form. We hypothesize that unknown protein(s) operating in vivo and modifying the Ah-mediated response at the common XRE element located upstream of these genes is affected zonespecifically by 3MC and βNF.

Hepatocyte expression of tumor associated aldehyde dehydrogenase (ALDH-3) and p21 Ras following diethylnitrosamine (DEN) initiation and chronic exposure to di(2-ethylhexyl)phthalate (DHEP).

Phthalate esters such as di(2-ethylhexyl)phthalate (DEHP) either promote or inhibit rat liver tumorigenesis depending on the carcinogenesis protocol. In this study, we examined the expression of two histochemical markers, the tumor associated isozyme of aldehyde dehydrogenase (ALDH-3) and the oncoprotein p21 Ras, in the livers of male F344 rats. The rats were initiated with DEN and further treated with either DEHP (a known inhibitor of hepatocarcinogenesis), phenobarbital (PB, a known promoter of hepatocarcinogenesis), or a combination of DEHP and PB. The studies were designed to examine the expression of these markers in both normal appearing liver and hepatic hyperplastic and neoplastic lesions and to correlate the early expression of the markers at 26 weeks in the normal appearing liver to later tumor incidence at 52 weeks. The expression of each marker was detected by immunohistochemical methods on formalin-fixed paraffin embedded sections of normal appearing liver or liver lesions. We found that ALDH-3 and p21 expression were significantly enhanced in rats receiving PB after DEN initiation at 26 weeks and that the incidence of hepatocellular carcinomas was likewise increased compared to control or DEN only treated animals. DEN initiation followed by a combination of PB and either 0.1 or 0.5% DEHP significantly reduced ALDH-3 but not p21 Ras expression at 26 weeks compared to DEN plus PB only. These treatment regimens also reduced the incidence of hepatocellular carcinomas at 52 weeks. DEN followed by any of the three doses of DEHP without PB resulted in ALDH-3 expression similar to DEN alone. However, p21 Ras expression was significantly increased after these treatments. For all treatment groups, both the early (26 weeks) expression of p21 Ras and ALDH-3 correlated with hepatocellular carcinoma incidence at 52 weeks. However, the correlation between hepatocellular carcinoma and ALDH-3 expression was better than p21 Ras or the other markers we have studied. We concluded that ALDH-3 expression is significantly downregulated after DEHP treatment, and that expression of the isozyme correlated with later hepatocarcinoma incidence and may indicate a significant relationship between ALDH-3 expression and hepatocarcinogenesis during DEHP treatment.

Quantitation of ultraviolet light-absorbing fractions of the cornea.

This study attempts to measure the quantitative contribution of major chemical fractions of the whole bovine cornea to ultraviolet (UV) absorption between 240 and 300 nm, with special attention to the biologically significant range of 290-300 nm. The cornea was divided into water-insoluble, nonprotein small water-soluble, water-soluble protein, and lipid-soluble fractions. The insoluble fraction (largely collagen) was solubilized by enzymatic digestion. Solutions of individual fractions equivalent to a constant mass of fresh cornea were scanned for absorption from 240 to 300 nm. The sum of the absorbances of the individual fractions closely approximated the absorbance of whole corneas at all wavelengths examined. The extinction coefficients of the lipid and water-soluble fractions were several times greater than that of the insoluble fraction throughout the studied spectrum. Yet, because of its large mass (75% of cornea dry weight), the insoluble fraction accounted for 40-50% of UV absorbance between 240 and 280 nm. However, in the range of 290-300 nm, the water-soluble plus lipid-soluble fractions accounted for 60-65% of the total absorption, with the water-soluble proteins alone accounting for 40-45% of the total. The soluble proteins comprised only approximately 17% of the cornea's dry weight. The special contribution of the water-soluble proteins to absorption was attributed to their relatively high tryptophan content (approximately 1.6% by weight). A 54-kDa protein, identified by others as tryptophan-rich, tumor-associated aldehyde dehydrogenase, accounted for approximately 30% of the total soluble protein mass.(ABSTRACT TRUNCATED AT 250 WORDS)

The role of hyperplastic nodules in dichloroacetic acid-induced hepatocarcinogenesis in B6C3F1 male mice

Dichloroacetic acid (DCA) has recently been shown to increase significantly the incidence of hepatic adenomas (HAs) and hepatocarcinomas (HCs) in male B6C3F1 mice. Although little is known about the mechanism of DCA carcinogenesis, chronic ingestion of the compound in drinking water induces primarily hyperplastic nodules (HNs) prior to the appearance of HAs and HCs. Given the putative preneoplastic potential of the HNs, we undertook this study to determine the role of the HNs in the progression of DCA-induced hepatocarcinogenesis. This role was assessed by detecting the expression of five different tumor markers: p21 ras, p39 c-jun, phosphotyrosine, tumor-associated aldehyde dehydrogenase and alpha-fetoprotein, all known from previous studies to be expressed more often in neoplastic liver lesions than in normal liver. Tumor marker expression was detected by immunohistochemical methods using formalin-fixed, paraffin-embedded sections of normal B6C3F1 mouse liver, and DCA-induced HNs, HAs and HCs. The results demonstrated that, except for the c-jun marker, HNs expressed the markers significantly less often than either HAs or HCs. Equal expression of c-jun occurred in any of the three lesion types. Although these results could be used to argue that no relationship existed between HNs and later-appearing HAs and HCs, those HNs that were marker positive contained small nests of marker-positive hepatocytes among a field of normally appearing unstained hepatocytes. No similar nests of marker-positive cells were detected in any area of normal liver outside the HNs. Also very few altered hepatic foci (AF) were detected with these markers or with hematoxylin and eosin, or with histochemical stains for ATPase or glucose-6-phosphatase deficiencies. These results suggested that these nests within some HNs were areas of transformed, or neoplastic hepatocytes. Phenotypic heterogeneity analysis, in which the number of tumor markers co-expressed by any given lesion was examined, confirmed a significantly greater percentage of HAs and HCs expressing multiple markers than HNs. Those HNs that expressed multiple markers, however, expressed at the same frequency as HAs and HCs and the expression was confined to the same nests of cells. Taken together, these data suggest that these nests of marker-positive cells within the HNs were neoplastic and could develop into later-appearing HAs and/or HCs. The absence of marker expression in normal liver and limited expression in the few AF indicates that the HNs may be the only significant preneoplastic lesion in DCA-induced hepatocarcinogenesis.

Aldehyde dehydrogenase heterogeneity in rat hepatic cells

In normal rat liver, aldehyde dehydrogenase (Aldehyde: NAD+ oxidoreductase, EC 1.2.1.3; ALDH) is found primarily in mitochondrial and microsomal fractions. During hepatocarcinogenesis, an additional tumor-associated aldehyde dehydrogenase (T-ALDH) is detectable in the cytosol of preneoplastic and neoplastic cells. We report here differences in the ALDH distribution pattern in different rat hepatoma cell lines compared to normal rat hepatocytes. Of the four basal ALDH enzymes, one mitochondrial ALDH and one microsomal ALDH account for 96% of total ALDH molecules detectable with our probes in normal hepatocytes. The other two mitochondrial and microsomal ALDH enzymes are only detectable in the appropriate subcellular fraction from large populations of cells. The tumor-associated ALDH is not detectable in normal hepatocytes. In addition to varying amounts of T-ALDH in the six different rat hepatoma cell lines examined, differences in the amounts of mitochondrial and microsomal ALDHs also occur in both high and low T-ALDH activity hepatoma cell lines. Each of five ALDH enzymes examined has a characteristic half-life varying from 45 min to 95 h.

Effects of hepatocarinogenic initiators on aldehyde dehydrogenase gene expression in cultured rat hepatic cells

The effects of certain in vivo inducers of tumor-associated aldehyde dehydrogenase (aldehyde:NAD+ oxidoreductase, EC 1.2.1.3; ALDH) activity on the expression of tumor-associated ALDH (T-ALDH) in vitro have been investigated using cultured rat hepatocytes and hepatoma cell lines. Two distinct groups of T-ALDH inducers have been identified. Three hepatocarcinogenic initiators 2-acetylaminofluorene, diethylnitrosamine and ethionine, which cause changes in T-ALDH in vivo, do not induce T-ALDH activity in cultured rat hepatocytes or hepatoma cell lines following either short-term or long-term exposures. In contrast, polycyclic aromatic hydrocarbons, such as 3-methylcholanthrene, benzo[a]pyrene and 7,12-dimethylbenz[a]anthracene, induce an immediate increase of T-ALDH activity in both cultured rat hepatocytes and hepatoma cell lines. Synthesis and degradation rates of T-ALDH mRNA and protein have also been determined. The synthesis of T-ALDH protein is coupled with the increased synthesis of T-ALDH mRNA when the T-ALDH gene is constitutively expressed or activated by an inducer. Both T-ALDH mRNA (t1/2 = 25 - 34 h) and protein (t1/2 = 88 - 95 h) in high T-ALDH activity cell lines or low-activity cell lines treated with an inducer are relatively stable. Combined with previous studies, the results suggest that at least two different mechanisms are involved in T-ALDH gene expression; events occurring during initiation as well as during promotion appear to be involved in the genetically stable changes in T-ALDH gene expression which occur in vivo. The results also indicate that the lack of T-ALDH activity in normal hepatocytes or low-activity hepatoma cell lines is due to repression of the T-ALDH gene rather than to the differential stability of T-ALDH mRNA or protein.

Characterization of rat cornea aldehyde dehydrogenase

Aldehyde dehydrogenase has been purified from rat cornea in a single step. The enzyme is a class 3 aldehyde dehydrogenase. Cornea aldehyde dehydrogenase is a 100-kDa dimer composed of 51-kDa subunits, prefers NADP + as coenzyme, and preferentially oxidizes benzaldehyde-like aromatic aldehydes as well as medium chain length (4–9 carbons) aliphatic aldehydes. The substrate and coenzyme specificity, immunochemical properties, effect of disulfiram, pH profile, and isoelectric point of cornea aldehyde dehydrogenase are identical to those of tumor-associated aldehyde dehydrogenase, the prototype class 3 enzyme. The substrate and coenzyme preferences are consistent with a role for cornea aldehyde dehydrogenase in the oxidation of a variety of aldehydes generated by lipid metabolism, including lipid peroxidation.

Substrate preference of a cytosolic aldehyde dehydrogenase inducible in rat liver by treatment with 3-methylcholanthrene

The substrate preference of an aldehyde dehydrogenase induced in rat liver cytosol by 3-methylcholanthrene was examined. This enzyme, T-ALDH, is identical to the aldehyde dehydrogenase inducible in rat liver by 2,3,7,8-tetrachloro-dibenzo- p-dioxin and the tumor-associated aldehyde dehydrogenase found in rat hepatocellular neoplasms. With either NAD or NADP as coenzyme, the preferred substrates were the aliphatic aldehydes n-hexanal, n-nonanal, and isobutyraldehyde and the aromatic aldehydes 2,5-dihydroxybenzaldehyde, benzaldehyde, and 3-hydroxybenzaldehyde. The results indicate that T-ALDH may play a role in oxidizing a variety of aldehydes produced in physiological lipid metabolism. On the contrary, this isozyme does not seem to participate in the oxidation of small aliphatic aldehydes generated during lipid peroxidation. Similarly, no significant activity could be detected when the enzyme was tested with aldehydes produced in carbohydrate, amino acid, polyamine, steroid, and vitamin metabolism.

Cloning and complete nucleotide sequence of a full-length cDNA encoding a catalytically functional tumor-associated aldehyde dehydrogenase.

To study the mechanism(s) controlling expression of the tumor-associated aldehyde dehydrogenase (tumor ALDH), which appears during rat hepatocarcinogenesis, cDNAs encoding this isozyme were cloned and identified with an antibody probe. Poly(A)-containing RNA from HTC rat hepatoma cells, which have been shown to possess high levels of tumor ALDH, was used as template to synthesize double-stranded cDNA. The cDNA was methylated to protect internal sites. Two different synthetic DNA linkers were added sequentially to the cDNA to insure correct orientation for expression from the lac promoter of pUC8. A library of 100,000 independent members carrying inserts greater than 1 kilobase was obtained. From this library, two apparently identical tumor ALDH clones, differing only in size, were identified with an indirect immunological probe. The larger of the cDNA clones identified, pTALDH, was chosen for further study. Interestingly, since tumor ALDH is a dimeric enzyme, pTALDH directs synthesis of a functional tumor ALDH in the bacterial cell. The cDNA sequence has been confirmed by comparison to the amino acid sequence of tumor ALDH purified from HTC cells.

Changes in aldehyde dehydrogenase activity during diethylnitrosamine-initiated rat hepatocarcinogenesis.

Diethylnitrosamine following partial hepatectomy followed by phenobarbital promotion was used to study changes in aldehyde dehydrogenase (ALDH) activity during rat hepatocarcinogenesis. Over a period of 350 days, animals were killed at intervals and the ALDH phenotype of normal liver and any lesions was characterized by histochemical analysis, total activity assays and gel electrophoresis using propionaldehyde and NAD+ to detect normal liver ALDH activities, and benzaldehyde and NADP+ for tumor-associated ALDH. In contrast to previously tested protocols, no significant changes in ALDH activity were demonstrable by histochemistry or total activity assays in preneoplastic livers. However, nine of 16 (56%) of the hepatocellular carcinomas examined expressed the tumor-associated ALDH phenotype. The present results are integrated with previous observations as a hypothesis explaining the roles of initiation and promotion in expression of the tumor-associated aldehyde dehydrogenase phenotype.

Characteristics and aldehyde dehydrogenase activity of four rat hepatoma cell lines produced by diethylnitrosamine-phenobarbital treatment.

Recent studies in our laboratory have shown that five established rat hepatoma cell lines provide a wide spectrum of tumor-associated aldehyde dehydrogenase (ALDH) activity representative of the range of activities of this enzyme seen in primary rat hepatocellular carcinomas. Four newly established rat hepatoma cell lines, RLT-2M, RLT-3C, RLT-9F, and RLT-5G, were derived from a primary hepatocellular carcinoma. The primary tumor was induced by a single injection of diethylnitrosamine (15 microM/g body weight) to a 1-d-old female S-D rat followed at weaning by chronic phenobarbital treatment. RLT-2M was established from outgrowths of minced tumor pieces. RLT-3C, RLT-9F, and RLT-5G were cloned from RLT-2M by the serial endpoint dilution. All four lines have been maintained in culture for over 100 passages. The ALDH phenotype in both the primary tumor and the four new cell lines was determined by total activity assay, gel electrophoresis, and histochemistry. By total activity assay, the primary tumor did not possess significant tumor-ALDH activity. In contrast, the four new cell lines expressed tumor-ALDH activity. However, they differed in their basal ALDH activities and in ALDH inducibility by 3-methylcholanthrene, benzo(a)pyrene, and phenobarbital. Additionally, significant decreases in tumor-ALDH activity occurred when cells from each line were passaged in vivo. The four lines have been characterized by light and electron microscopic morphology, tumorigenicity, chromosome number, doubling time, and colony formation efficiency in soft agar.

Expression of tumor-associated aldehyde dehydrogenase during rat hepatocarcinogenesis using the resistant hepatocyte model.

The resistant hepatocyte model was used to study expression of tumor-associated aldehyde dehydrogenase (ALDH) activity during the course of rat hepatocarcinogenesis. The hepatic ALDH phenotype was determined at intervals over 280 days by histochemical analysis, total ALDH activity assays and gel electrophoresis, using propionaldehyde and NAD (P/NAD) to characterize normal liver ALDH activity or benzaldehyde and NADP (B/NADP) to determine tumor-associated ALDH activity. By total activity assays and gel electrophoresis, no significant changes in ALDH activity occurred until day 70. However, histochemical analysis clearly demonstrated changes in ALDH activity early in neoplastic development. Intense focal hepatocyte staining with P/NAD and/or B/NADP was first detectable at day 28. The number of P/NAD-positive foci increased until day 35 then declined until day 70. The number of B/NADP-positive foci also increased until day 35, but then remained relatively constant for the remainder of the experiment. GGT activity of serial sections indicated that early ALDH-positive lesions represent a small subpopulation (9%) of all GGT-positive foci. However, by day 168 a significant portion (80%) of persistent GGT-positive neoplastic nodules were also B/NADP-positive histochemically. In addition, virtually all hepatocellular carcinomas (96%) generated by this protocol possessed significantly elevated levels of tumor-associated ALDH by histochemical analysis, total ALDH activity and gel electrophoresis. These results indicate that early appearing ALDH-positive lesions may define one early subpopulation of all initiated cells that have a high probability of progressing to the ultimate neoplasm.