Search

We examined the expression of the midkine (MK) and alpha-fetoprotein (AFP) genes in 15 paired human specimens obtained from hepatocellular carcinoma (HCC) and the corresponding noncancerous regions of the same patients. A total of 14 HCC but none of the noncancerous specimens were positive for the MK mRNA. In contrast, three HCC specimens and one corresponding noncancerous sample out of the three AFP-positive HCC cases expressed the AFP gene. A 2.3-kb genomic fragment in the regulatory region of the MK gene could activate a fused reporter gene in both AFP-producing and -nonproducing HCC lines, and the MK fragment-mediated transcriptional activity was comparable to the AFP enhancer-linked AFP promoter in AFP-producing cell lines. The AFP-producing but not AFP-nonproducing HCC cell lines that were transfected with the MK promoter-linked herpes simplex virus-thymidine kinase (HSV-TK) gene became susceptible to a prodrug ganciclovir to a similar degree of the HCC transfected with the enhancer-linked AFP promoter-fused HSV-TK gene. These data suggest that the MK promoter can activate a therapeutic gene preferentially in HCC and is as useful as the AFP promoter in clinical settings.

We investigated mechanisms regulating expression of alpha-fetoprotein (AFP) in 3 human hepatoma cell lines, HuH-7, HepG2, and huH-1, producing high, medium, and low levels of AFP, respectively. The silencer, a negative cis-acting element of the AFP gene, was highly activated in huH-1 and HepG2 to repress AFP enhancer activity by 91%, whereas only 26% repression was observed in HuH-7. To account for the difference in AFP production between HepG2 and huH-1, we investigated the roles of two isoforms of the AT motif-binding factor 1 (ATBF1) transcription factor, ATBF1-A and -B. Cotransfection assays showed that the ATBF1 isoforms regulated the AFP gene differently in HepG2 and huH-1. In huH-1 and HuH-7, both ATBF1 isoforms suppressed strongly enhancer activity and slightly promoter activity. In HepG2, on the other hand, ATBF1-A suppressed the enhancer and promoter activities, but surprisingly, ATBF1-B was found to stimulate enhancer activity while showing no effect on the promoter. Levels of ATBF1-A mRNA were similar in all 3 cell lines, whereas the expression ATBF1-B mRNA varied greatly, with the highest level seen in HepG2 followed by huH-1 and HuH-7. These results suggest that, in HepG2, ATBF1-B may have a dominant negative effect to relieve the transcriptional repression caused by its isoform. In support of this view, we found that the N-terminal region specific to the ATBF1-A molecule possessed transcriptional repressor activity. Thus, the use of the ATBF1 variants as well as the silencer may provide a unique mechanism that contributes to the determination of AFP levels in human hepatoma cell lines.

We previously reported that the retroviral vector expressing the herpes simplex virus-thymidine kinase gene under the control of 0.3-kb human alpha-fetoprotein (AFP) gene promoter (AF0.3) provided the cytotoxicity to ganciclovir (GCV) in high-AFP-producing human hepatoma cells but not in low-AFP-producing cells. Therefore, specific enhancement of AFP promoter activity is likely to be required to induce enough cytotoxicity in low-AFP-producing hepatoma cells. In this study, we constructed a hybrid promoter, [HRE]AF, in which a 0.4-kb fragment of human vascular endothelial growth factor 5'-flanking sequences containing hypoxia-responsive element (HRE) was fused to AF0.3 promoter. By means of the reporter gene transfection assay, hypoxia-inducible transcriptions that were mediated by [HRE]AF promoter were detected in low- and non-AFP-producing human hepatoma cells, but not in nonhepatoma cells. When the herpes simplex virus-thymidine kinase gene controlled by [HRE]AF promoter was transduced into hepatoma and nonhepatoma cells by a retroviral vector, the exposure to 1% O2 induced GCV cytotoxicity specifically in the hepatoma cells. Moreover, in nude mice bearing solid tumor xenografts, only the tumors consisting of the virus-infected hepatoma cells gradually disappeared by GCV administration. These results indicate that the hypoxia-inducible enhancer of the human vascular endothelial growth factor gene, which is directly linked to human AFP promoter, involves selective and enhanced tumoricidal activity in gene therapy for hepatocellular carcinoma.

Tissue-specific promoter has been used for cancer-specific suicide gene therapy, but its transcriptional activity is relatively low. For more efficient gene therapy of hepatocellular carcinoma, a simultaneous infection method of two recombinant adenoviruses was developed, in which one carried Cre gene under the control of alpha-fetoprotein promoter and the other a potent expression unit activated by Cre. When the vectors with lacZ reporter gene were introduced systematically into mouse models of disseminated tumors, specific and enhanced gene expression was observed exclusively in hepatocellular carcinomas both in the liver and in the lung. Next, using herpes simplex virus thymidine kinase, the anti-tumor effect was examined. Although in cultured cells, 60-300-fold expression of enzymatic activity and enhanced ganciclovir sensitivity was obtained compared with that of the single recombinant adenovirus directly driven by alpha-fetoprotein promoter, there was no significant anti-tumor effect for subcutaneous tumor on athymic mice. The lack of anti-tumor effect in mice could be explained by insufficient simultaneous transduction of the two vectors in the tumors, since it was found that a high multiplicity of infection was required to activate this system. Some strategies to overcome this dose limitation are needed, at least in the case of hepatocellular carcinoma.

The effects of heparin-binding epidermal growth factor-like growth factor (HB-EGF) on albumin gene expression were analyzed in alpha-fetoprotein (AFP)-producing and AFP-nonproducing human hepatoma cells, HuH-7 and huH-1/cl.2, in which the AFP silencer is inactive and active, respectively. HB-EGF selectively suppressed the AFP enhancer activity, resulting in decreased levels of both albumin and AFP mRNA in HuH-7 cells. In contrast, HB-EGF did not influence the albumin transcripts in huH-1/cl.2 cells or primary culture of rat hepatocytes, although the AFP enhancer activity was reduced by HB-EGF in huH-1/cl.2 cells as well as in HuH-7 cells. These results indicate that the AFP silencer interacts with the AFP enhancer to block its activity regulating both the albumin and AFP promoters, and that HB-EGF downregulates the albumin gene expression only in the absence of the AFP silencer through the repression of the AFP enhancer activity.

Treatment of HuH-7 human hepatocellular carcinoma (HCC) cells with 1-10 mM sodium butyrate (SB) resulted in growth inhibition in a dose-dependent manner. At 3 mM and higher concentrations, SB caused nuclear fragmentation and DNA ladder formation characteristic of apoptosis. In the treated cells, the expression of p21 (WAFI/CIPI) increased and that of alpha-fetoprotein (AFP) decreased. These characteristic changes were also observed with 5 other human HCC cell lines with or without mutation of the p53 gene. The ability of these cells to form colonies in soft agar was suppressed by either pretreating the cells with SB prior to soft agar plating or incubating untreated cells in SB-containing soft agar. Direct injection of SB into tumors developed from HuH-7 cells in nude mice resulted in an increase in the p21 level, a decrease in the tumor size and an increase in the survival time of mice. When the inoculation of HuH-7 cells into nude mice was immediately followed by subcutaneous injection of SB, development of tumors was either significantly delayed or completely suppressed. These results suggest that SB induces cellular differentiation and suppresses growth and tumorigenicity of HCC cells in vitro and in viva by a mechanism independent of p53 but possibly dependent on p21.

Hepatic nuclear factor 1 (HNF-1) regulates genes in a hepatocyte-specific manner. It has been previously reported that the ratio of HNF-1 alpha and HNF-1 beta mRNA is related to histological differentiation hepatocellular carcinoma (HCC). In this study, the expression levels of the HNF-1 alpha and HNF-1 beta proteins were analysed relatively and quantitatively in various histologically differentiated HCC and surrounding non-cancerous tissues, and HNF-1 alpha binding activity for the AT element of the B domain of the human alpha-fetoprotein enhancer was examined. Western blot analysis demonstrated that HNF-1 alpha protein was expressed at a higher level in well-differentiated HCC tissues than in the surrounding non-HCC tissues; on the other hand, the HNF-1 alpha protein was expressed at lower levels in moderately and poorly differentiated HCCs than in the surrounding non-HCC tissues. The levels of HNF-1 beta expression in well-differentiated and poorly differentiated HCCs were similar to and higher than those found in the respective surrounding non-cancerous portions. In binding assays, HNF-1 binding activity was high in well-differentiated HCC and lower in moderately and poorly differentiated HCCs. Most well-differentiated HCC cases showed immunohistochemical expression of HNF-1 alpha. These findings show that poor histological differentiation of HCC correlates with decreases in the level and activity of HNF-1 alpha proteins.

The alpha-fetoprotein (AFP) gene is normally expressed in fetal liver and is transcriptionally silent in adult liver but overexpressed in human hepatocellular carcinoma (HCC). Here, we demonstrate that replication defective recombinant adenoviral vectors, containing the human AFP promoter/enhancer, can be used to express the Escherichia coli cytosine deaminase (CD) gene (AdAFPCD) and the beta-galactosidase gene (AdAF-PlacZ) in AFP-producing HCC cell lines. Expression of the CD gene by adenovirus from the AFP promoter/enhancer (AdAFPCD) induced cells sensitive to 5-fluorocytosine (5FC) in the AFP-producing cells but not in the AFP-nonproducing cells. Transduction by an adenoviral vector harboring an ubiquitous strong promoter and CD gene showed enzymatic activity and 5FC killing in all cell lines. When AdAFPlacZ was injected into the s.c. established hepatoma in vivo, expression of the beta-galactosidase gene was confined to AFP-producing HCC xenografts. Moreover, HCC xenografts regressed by transduction with AdAFPCD and subsequently with 5FC treatment in vivo. These findings suggest that utilization of the AFP promoter/enhancer in an adenoviral vector can confer selective expression of a heterologous suicide gene in hepatocellular carcinoma cells in vitro and in vivo.

Background/aims: Liver-specific protein genes have multiple cis-/trans-acting elements, but those accountable for hepatocytic differentiation are unclear. An AT-rich core sequence (AT motif) is essential as a cis-acting element for the hepatic transcription. Homologous proteins hepatocyte nuclear factor-1 (HNF-1) and variant HNF-1 (vHNF-1) bind to this motif. The ratio of HNF-1 to vHNF-1 mRNA was examined in various liver tissues with respect to their differentiation., Methods: The competitive reverse transcriptional polymerase chain reaction was employed to amplify HNF-1 and vHNF-1 mRNA simultaneously and to examine their expression ratio in total RNA extracted from frozen liver tissues of 37 patients with hepatocellular carcinoma, five patients with hepatoblastoma, and 15 non-neoplastic liver tissues., Results: The ratio of HNF-1 to vHNF-1 mRNA was higher in well-differentiated cases than in poorly-differentiated and undifferentiated cases, except that one poorly-differentiated hepatoblastoma displayed a high ratio. Non-neoplastic liver tissues had low ratios similar to poorly-differentiated hepatocellular carcinoma, the reason for which remained unknown. However, chronic hepatitis and liver cirrhosis cases also demonstrated low ratios, and hence degenerative changes themselves displayed no obvious influence on such ratios. Thus, the gene expression of HNF-1 and vHNF-1 seemed to be differentially regulated in neoplastic and non-neoplastic hepatocytes., Conclusions: These results suggested that the ratio of HNF-1 to vHNF-1 mRNA correlated with histological differentiation of HCC and hepatoblastoma.

We have developed a recombinant replication-defective adenovirus containing human alpha-fetoprotein (AFP) promoter/enhancer to direct cell type-specific expression of the herpes simplex virus thymidine kinase (HSVtk) gene to AFP-producing hepatocellular carcinoma (HCC) cells. After an in vitro infection by a recombinant adenovirus carrying the lacZ gene under the control of human AFP promoter/enhancer (AdAFPlacZ), an expression of the lacZ gene was demonstrated efficiently in AFP-producing HuH-7 and HepG2 cell lines, but not in AFP-nonproducing HLE and HLF cell lines, although lacZ gene expression was demonstrated in all these cell lines when infected with adenovirus vector carrying lacZ gene driven by the beta-actin-based promoter. Expression of the HSVtk gene by adenovirus, from AFP promoter/enhancer (AdAFPtk) induced the cells sensitive to ganciclovir (GCV) in the AFP-producing cell line efficiently, but not in AFP-nonproducing HLF hepatoma cells. An in vitro bystander effect was observed when only 10% of the cells were infected with AdAFPtk. These findings suggest that the AFP promoter/enhancer sequence can provide the tumor-specific activity for the therapeutic gene expression, and that the AdAFPtk vector induces the selective growth inhibition by GCV in the adenovirus-infected human hepatoma cells in vitro. Recombinant adenovirus transfer of the HSVtk gene under the control of tumor-specific promoter followed by GCV may have promise as a targeted in situ treatment for solid neoplasms.

Most patients with hepatocellular carcinoma have an elevated alpha-feto-protein (AFP) level. This high level of AFP expression is transcriptionally controlled by the 5'-flanking sequence of the AFP gene. Using the 5'-flanking sequence as a promoter for the herpes simplex virus thymidine kinase (HSV-TK) gene in an adenoviral vector (Av1AFPTK1), the therapeutic efficacy of adenovirus-mediated HSV-TK gene transduction, followed by ganciclovir (GCV) administration, was studied in tumors in athymic nude mice. Av1AFPTK1 transduction of two cell lines demonstrated HSV-TK enzyme activity only in the AFP-producing cells (HuH7) and not in the AFP nonproducing cells (SK-Hep-1). As expected, only transduced HuH7 cells were killed by GCV treatment. Transduction by an adenoviral vector harboring a Rous sarcoma virus promoter and HSV-TK gene (Av1TK1) showed enzymatic activity and GCV killing in both cell lines. All HuH7 tumors that were transduced with either Av1AFPTK1 or Av1TK1 completely regressed after GCV treatment. On the other hand, there was complete regression of SK-Hep-1 tumors only when treated with Av1TK1 and GCV and not when treated with Av1AFPTK1 and GCV. Thus, cell-specific killing was achieved by adenoviral vector containing AFP promoter for the HSV-TK gene and GCV treatment.

Background/aims: Butyrate, a product of colonic bacterial flora, functions as an antiproliferative agent and induces cell differentiation in a variety of cell types. In the present study, the effects of butyrate on cell growth and expression of alpha-fetoprotein (AFP) and albumin genes in HuH-7 human hepatoma cells were investigated., Methods: The HuH-7 cells were treated with sodium butyrate (0-1 mmol/L), and numbers of viable cells were counted at 24, 48, and 72 hours after treatment. To elucidate the effects of sodium butyrate on AFP and albumin gene expression, Northern blotting and transient chloramphenicol acetyltransferase plasmid transfection experiments were performed., Results: Cell growth was dose dependently inhibited by sodium butyrate. By Northern blot analysis, the level of AFP messenger RNA was reduced by treatment with sodium butyrate, whereas the level of albumin messenger RNA was elevated by this treatment. In transient chloramphenicol acetyltransferase plasmid transfection experiments, sodium butyrate repressed the AFP promoter activity but did not change the AFP enhancer or silencer activities. In contrast, the albumin promoter activity was stimulated by sodium butyrate., Conclusions: These results suggest that butyrate leads to the reciprocal differentiating regulation of AFP and albumin gene expression at the transcriptional level in human hepatoma cells.

Background: Colloid osmotic pressure has been thought to regulate albumin synthesis; however, the exact mechanism remains obscure. In the present study, the effect of colloid osmotic pressure on the albumin and alpha-fetoprotein gene expression in HuH-7 human hepatoma cells was analyzed., Methods: HuH-7 cells were treated with albumin or dextran (mean mol wt, 70,000), and changes in the levels of albumin and alpha-fetoprotein messenger RNA (mRNA) were analyzed by Northern blotting. Furthermore, in transient chloramphenicol acetyltransferase (CAT) plasmid transfection experiments, effects of colloid osmotic pressure on CAT activities were studied., Results: By Northern blot analysis, the levels of both albumin and alpha-fetoprotein mRNA were dose-dependently suppressed by the elevation of colloid osmotic pressure and returned to pretreatment levels 48 hours after the culture medium containing dextran was replaced with a dextran-free fresh medium. In transient CAT plasmid transfection experiments, the increased level of colloid osmotic pressure resulted in the repression of both albumin and alpha-fetoprotein promoter activities. In contrast, alpha-fetoprotein enhancer activity, which possibly regulates not only alpha-fetoprotein but also albumin gene expression, was not affected by changes in colloid osmotic pressure., Conclusions: These results suggest that colloid osmotic pressure regulates both albumin and alpha-fetoprotein gene transcription through the modulation of their promoter activities.

The interaction of IFN-alpha with IL-1 beta or TNF-alpha on hepatitis B surface antigen (HBsAg) expression was analysed in hepatitis B virus (HBV)-DNA integrated PLC/PRF/5 and non-integrated HuH-7 human hepatoma cells. Secretion of HBsAg in PLC/PRF/5 cells was reduced by IFN-alpha, IL-1 beta or TNF-alpha, and synergistically depressed when IFN-alpha was used in combination with IL-1 beta or TNF-alpha. By Northern blot analysis, the levels of HBsAg mRNA were suppressed by IFN-alpha in combination with IL-1 beta or TNF-alpha. In the chloramphenicol acetyltransferase plasmid transfection assay, IFN-alpha in combination with IL-1 beta or TNF-alpha caused a much greater suppression of HBV enhancer activity than IFN-alpha, IL-1 beta or TNF-alpha alone in both hepatoma cells. These findings suggest that the interaction of IFN-alpha with IL-1 beta or TNF-alpha synergistically represses HBV enhancer activity, resulting in depressed expression of HBsAg.

9-deoxy-delta 9,delta 12-13,14-dihydro-prostaglandin D2 (delta 12-PGJ2) is a potent inhibitor of proliferation of tumor cells. In the present study, the effect of delta 12-PGJ2 on the alpha-fetoprotein(AFP) and the albumin gene expression was analyzed in HuH-7 human hepatoma cells. delta 12-PGJ2 inhibited the cell growth and reduced the medium AFP concentrations dose-dependently. To determine whether this decline of AFP depends only on the relative decrease in cell numbers by delta 12-PGJ2, or is in part, due to the decrease in the cellular AFP synthesis by delta 12-PGJ2, Northern blot analysis was performed in this study. By Northern blotting, it was shown that delta 12-PGJ2 caused a marked reduction in the levels of the AFP mRNA and the albumin mRNA. In contrast, the level of the beta-actin mRNA was not changed by delta 12-PGJ2. In the transient chloramphnicol acetyltransferase plasmid transfection experiments, delta 12-PGJ2 did not suppress the AFP enhancer activity, which possibly regulates both the AFP and the albumin gene expression in HuH-7 hepatoma cells, but resulted in the selective repression of the AFP and the albumin promoter activity. These results suggest that delta 12-PGJ2 suppresses not only cell growth but also expression of the AFP gene and the albumin gene at the transcriptional level in human hepatoma cells.

A large percentage of human hepatomas produce alpha-fetoprotein (AFP), but the levels of AFP expression vary greatly among hepatomas. To understand the molecular basis for this variation, we analyzed transcriptional regulatory activities associated with the 5'-flanking region of the AFP gene in two human hepatoma cell lines, HuH-7 and huH-1/cl-2, which produce a high and a low level of AFP, respectively. We found that the low level of AFP production in huH-1/cl-2 is due to the action of at least two silencer regions located between the enhancer and the promoter of the AFP gene. In contrast, no silencer activity is expressed in HuH-7. We identified 5'-CTTCATAACTAATACTT-3' to be a core sequence responsible for the negative regulatory activity. This sequence is repeated four times in a strong, distal silencer region, Sd, whereas one copy is present in a weak, proximal silencer region, Sp. The silencer reduces transcriptional initiation by blocking enhancer activation of the AFP promoter in a position-dependent manner. The silencer functions in the presence of positive transcription factors and may play a key role in developmental repression as well as variable expression of the AFP gene in hepatomas.

Southern blot analysis revealed the presence of an aberrant albumin gene as well as a normal one in a human hepatoma cell line, HuH-7. A genomic sequence carrying this altered gene was isolated and characterized. This clone contains a 3-kbp 3' segment of the albumin gene linked to a non-albumin sequence at intron 11. The non-albumin sequence is assigned to chromosome 4q12-q13 by in situ hybridization. This indicates an interstitial deletion of a chromosomal segment within 4q11-q13 because the albumin gene is mapped there. The truncated albumin gene is detected in an early passage of HuH-7 cells and has been maintained stably in cell culture.

We studied the effects of transfection of the normal c-Ha-ras gene, rasGly-12, and its oncogenic mutant, rasVal-12, on expression of the alpha-fetoprotein (AFP) and albumin genes in a human hepatoma cell line, HuH-7. The mutant and, to a lesser extent, the normal ras gene caused reduction of the AFP mRNA but not the albumin mRNA level in transfected HuH-7 cells. Cotransfection experiments with a rasVal-12 expression plasmid and a chloramphenicol acetyltransferase reporter gene fused to AFP regulatory sequences showed that rasVal-12 suppressed the activity of enhancer and promoter regions containing A + T-rich sequences (AT motif). In contrast, rasVal-12 did not affect the promoter activity of the albumin and human hepatitis B virus pre-S1 genes even though these promoters contain homologous A + T-rich elements. ras transfection appeared to induce phosphorylation of nuclear proteins that interact with the AFP AT motif, since gel mobility analysis revealed the formation of slow-moving complexes which was reversed by phosphatase treatment. However, similar changes in complex formation were observed with the albumin and hepatitis B surface antigen pre-S1 promoters. Therefore, this effect alone cannot explain the specific down regulation of the AFP promoter and enhancer activity. ras-mediated suppression of the AFP gene may reflect the process of developmental gene regulation in which AFP gene transcription is controlled by a G-protein-linked signal transduction cascade triggered by external growth stimuli.

DNase I footprinting and gel mobility shift analysis showed that an HuH-7 hepatoma nuclear protein, termed AFP1, binds specifically to an AT-rich sequence, TGATTAATAATTACA, in domain B of the human alpha-fetoprotein enhancer. No such binding activity was found in HeLa cell nuclei. Transient transfection studies showed that a 54-base-pair region corresponding to the AFP1-binding site could stimulate the simian virus 40 early promoter to express a linked chloramphenicol acetyltransferase gene in an orientation-independent and cell-specific manner. The correlation between the binding of AFP1 and the stimulation of chloramphenicol acetyltransferase gene expression strongly suggests that specific interaction of AFP1 with the AT motif is important for cell-specific transcriptional enhancement. Competition gel mobility shift analysis revealed that similar AT-rich sequences with high affinities to AFP1 were also present in the promoters of the alpha-fetoprotein and albumin genes. These results suggest that AFP1 may function as a common regulatory factor in the transcription of the alpha-fetoprotein and albumin genes.

The expression of the alpha-fetoprotein (AFP) gene in hepatocellular carcinoma (HCC) and liver cirrhosis by in situ hybridization analysis of AFP mRNA in cryostat and paraffin-embedded tissue sections was studied. In HCC sections the majority of tumor cells showed positive hybridization with a 3H-labeled AFP complementary DNA probe. The number of radioactive hybrid grains detected in HCC sections generally paralleled the level of serum AFP in the patient. In liver cirrhosis sections a small number of cells showed positive hybridization. These cells were dispersed in the tissue and morphologically indistinguishable from surrounding hepatocytes. Each of these cells contained a high level of hybridization signals to permit easy identification. In situ hybridization analysis of AFP mRNA may be of use in detecting preneoplastic cells in liver cirrhosis that cannot be defined on the morphologic and immunohistochemical basis.

Detection of RNA transcripts within individual cells by in situ hybridization provides a powerful means for identifying specific cell types actively transcribing specific genes. We have applied this technique in order to analyze expression of the alpha-fetoprotein and albumin genes in a human hepatoma cell line, HuH-7. Using either 3H- or 35S-labeled human alpha-fetoprotein complementary DNA clone as a probe, we found that essentially all HuH-7 cells contained alpha-fetoprotein mRNA, although in various amounts. This correlated well with the presence of alpha-fetoprotein in all cells as detected by the peroxidase-antiperoxidase immunoenzyme method. The intracellular concentration of albumin, on the other hand, was below the level of detection by the peroxidase-antiperoxidase method. Consistent with this observation, we could not detect albumin mRNA with 3H-labeled albumin complementary DNA probes. However, the use of 35S-labeled probes having higher specific activities and higher efficiency of grain development resulted in the detection of albumin mRNA in a small percentage of HuH-7 cells. A variety of parameters involved in the in situ hybridization technique was examined to establish conditions suitable for demonstrating the presence of high- and low-copy numbers of mRNA in various cell and tissue preparations.

Using a titration procedure, we measured the proportion of alpha-fetoprotein (AFP) and albumin mRNA in normal, regenerating, and preneoplastic rat livers. AFP mRNA constitutes approximately 0.006% of the polysomal polyadenylated RNA of normal livers and this proportion increases only slightly before the onset of DNA synthesis in liver regeneration induced by partial hepatectomy or CCl4 injury. In either model of liver regeneration, the proportion of AFP mRNA in polysomal RNA is highest approximately 24 h after the peak of DNA synthesis. The increase in the proportion of AFP mRNA in polysomal RNA is relatively small during liver regeneration (2-4-fold) but is larger (30-50-fold) in preneoplastic livers of rats fed a choline-deficient diet containing 0.1% ethionine. In contrast to those changes in AFP mRNA, albumin mRNA levels remain unchanged during liver regeneration and double in preneoplastic livers. Our results indicate that the concept of "retrodifferentiation" as it applies to liver regeneration and certain types of hepatic neoplasia needs reevaluation.