Cytokines and insulin induce cationic amino acid transporter (CAT) expression in cardiac myocytes. Regulation of L-arginine transport and no production by CAT-1, CAT-2A, and CAT-2B

Cytokine-dependent production of nitric oxide (NO) by rat cardiac myocytes is a consequence of increased expression of the inducible isoform of nitric oxide synthase (iNOS or NOS2) and, in the presence of insulin, depresses the contractile function of these cells in vivo and in vitro. Experiments reported here show that L-lysine, a competitive antagonist of L-arginine uptake, suppressed NO production (detected as nitrite accumulation) by interleukin (IL)-1beta and interferon (IFN) gamma-pretreated cardiac myocytes by 70%, demonstrating that NO production is dependent on L-arginine uptake. Cardiac myocytes constitutively exhibit a high-affinity L-arginine transport system (Km = 125 microM; Vmax = 44 pmol/2 X 10(5) cells/min). Following a 24-h exposure to IL-1beta and IFNgamma, arginine uptake increases Vmax = 167 pmol/2 X 10(5) cells/min) and a second low-affinity L-arginine transporter activity appears (Km = 1.2 mM). To examine the molecular basis for these cytokine-induced changes in arginine transport, we examined expression of three related arginine transporters previously identified in other cell types. mRNA for the high-affinity cationic amino acid transporter-1 (CAT-1) is expressed in resting myocytes and steady-state levels increase by 10-fold following exposure to IL-1beta and IFNgamma. Only cytokine-pretreated myocytes expressed a second high-affinity L-arginine transporter, CAT-2B, as well as a low-affinity L-arginine transporter, CAT-2A. In addition, insulin, which potentiated cytokine-dependent NO production independent of any change in NOS activity, increased myocyte L-arginine uptake by 2-fold and steady-state levels of CAT-1, but not CAT-2A or CAT-2B mRNA. Thus, NO production by cardiac myocytes exposed to IL-1beta plus IFNgamma appears to be dependent on the coinduction of CAT-1, CAT-2A, and CAT-2B, while insulin independently augments L-arginine transport through CAT- 1.

Nitric oxide donors decrease the function and survival of human pancreatic islets

Nitric oxide (NO) has been proposed as a possible mediator of beta-cell damage in human IDDM. This hypothesis is based on in vitro studies with rodent pancreatic islets. In the present study we examined whether human beta-cells are affected by NO. In view of species differences in beta-cell sensitivity to damaging agents, rat islets were investigated in parallel. Isolated islets were exposed for 90 min to different concentrations of three chemically unrelated NO donors, SIN-1, GSNO or RBS. At the end of this incubation, human insulin release was mostly similar in control and NO-treated islets but, 48 h later, islet retrieval, islet DNA and insulin content, and glucose-induced insulin release were markedly lower in islets exposed to NO donors. Rat islets were already inhibited during the initial 90 min; 48 h later their loss in beta-cell function was similar to that in human islets. Nicotinamide or succinic acid monomethyl ester partially protected against SIN-1 induced islet cell loss, but not against the functional inhibition of human pancreatic islets. Exposure of human or rat islets to RBS was associated with significant DNA strand breakage, as judged by the comet assay (single cell gel electrophoresis) and by ultrastructural signs of cell damage. DNA damage was more severe in rat islet cells exposed to similar amounts of RBS. It is concluded that NO donors can damage human pancreatic islets, an effect paralleled by induction of nuclear DNA strand breaks.

High frequency of p53 gene mutations in primary breast cancers in Japanese women, a low-incidence population

The pattern of acquired mutations in the p53 tumour-suppressor gene is potentially useful for determining factors contributing to carcinogenesis in diverse populations differing in incidence and/or mortality from the disease. We previously reported differences in mutational patterns of the p53 gene in primary breast cancers from Midwest US Caucasian, African-American and Austrian women. Herein, we report 16 mutations in 27 primary breast cancers from Japanese women from Hirosaki, a population with a low incidence of breast cancer. The frequency of 59.3% of p53 mutations is the highest reported in breast cancers from a particular ethnic group thus far. A relatively high number of mutations (7/16) were heterozygous in at least some tumour cell clusters. Intergroup comparisons of the mutational pattern between this population and several other US, European and Japanese populations do not show any statistically significant differences. There were recurrent mutations at two sites, codon 273 (R --> H; three mutations), a common hotspot of mutations in breast and other cancers, and codon 183 (S --> Stop; two mutations), a very rare location for p53 mutations. These mutations were shown to be independent and presumably not in the germ line. The highest frequency of p53 mutations raises the possibility that p53 mutagenesis is a predominant factor for breast cancer development in this low-risk Japanese group, whereas in other cohorts different mechanisms are likely to account for the higher proportion of breast cancer. Further studies are needed to confirm the present observations.

Regulation of interleukin-1-beta-stimulated inducible nitric oxide synthase expression in cultured vascular smooth muscle cells by hemostatic proteins

Experiments were performed to examine the mechanism by which specific hemostatic proteins regulate the release of nitric oxide (NO) from interleukin-1 beta (IL-1beta) stimulated cultured rate aortic smooth muscle cells. Treatment of smooth muscle cells with IL-Beta stimulated inducible nitric oxide synthase (iNOS) mRNA expression, which preceded the release of NO (as measured by the accumulation of nitrite in the culture media). The cytokine-stimulated production of nitrite was blocked by the protein synthesis inhibitor cycloheximide, the transcriptional inhibitor actinomycin D, and the competitive inhibitor of NOS nitro-L-arginine. However, only actinomycin D inhibited IL-1beta-stimulated iNOS mRNA expression, Treatment of smooth muscle cells with IL-1beta in the presence of platelet derived growth factor or thrombin resulted in the inhibition of cytokine-stimulated expression of iNOS mRNA and NO release. The inhibitory effect of thrombin was reversed by hirudin and was mimicked by a 14 amino acid thrombin receptor activating peptide. In contract, the concomitant exposure of smooth muscle cells to IL-1beta-and plasmin resulted resulted in the potentiation of both IL-1beta-stimulated iNOS expression and NO generation. Finally, treatment of smooth muscle cells with IL-1beta in the presence of the hemostatic proteins did not affect the half-life of iNOS mRNA. These results demonstrate that specific protein components of the hemostatic system regulate IL- 1beta-stimulated iNOS mRNA expression in vascular smooth muscle cells. The capacity of hemostatic proteins to modulate the induction of vascular iNOS activity may play an important role in governing the release of NO and regulating thrombogenesis in vivo.

Tumor necrosis factor-alpha and interferon-gamma inhibit insulin secretion and cause DNA damage in unweaned-rat islets. Extent of nitric oxide involvement

Nitric oxide has been implicated as one possible mediator of interleukin-1 beta (IL-1)-induced inhibition of insulin secretion and islet cell damage. The aim of this study was to define the effects of tumor necrosis factor-alpha (TNF) and interferon-gamma (IFN) on nitric oxide production, insulin secretion, and DNA damage in islets from unweaned rats. Treatment of islets with 0.5-500 U/ml of either TNF or IFN on their own inhibited glucose-stimulated insulin secretion in a dose-dependent manner (minimum effective dose 5 U/ml). In combination, the cytokines exerted a pronounced synergistic inhibitory effect on secretion and were equipotent at causing a significant and concentration-dependent increase in culture medium nitrite levels, islet cyclic GMP formation, and DNA damage. Used alone or in combination, TNF and IFN significantly enhanced the activity of inducible nitric oxide synthase as determined by measuring the conversion of 14C-labeled arginine to 14C-labeled citrulline and nitric oxide. Use of arginine-free medium, without or with NG-monomethyl-L-arginine, resulted in inhibition of nitrite formation by 5-1,000 U/ml IFN+TNF and partial restoration of the insulin secretory response to glucose. Treatment of rat islets with increasing doses of TNF+IFN (5, 50, and 500 U/ml) resulted in a progressive increase in DNA damage, as shown by the comet assay, which detects DNA strand breaks in individual islet cells. The DNA damage caused by an intermediate concentration (50 U/ml) of TNF+IFN was comparable to that generated by IL-1 when used at 20 U/ml. We conclude that TNF and IFN induce nitric oxide formation, which partially inhibits glucose-induced insulin secretion and causes significant DNA strand breakage, but that as cytokine concentrations increase, non-nitric-oxide-mediated events predominate.

Comparison of inhibition of glucose-stimulated insulin secretion in rat islets of Langerhans by streptozotocin and methyl and ethyl nitrosoureas and methanesulphonates. Lack of correlation with nitric oxide-releasing or O6-alkylating ability

We have studied inhibition of glucose-stimulated insulin secretion in islets of Langerhans isolated from adult Sprague-Dawley rats and treated with different alkylating agents. Streptozotocin (STZ), N-methyl-N-nitrosourea (MNU), and N-ethyl-N-nitrosourea (ENU) all released nitric oxide, as demonstrated by an increase in medium nitrite and cellular cyclic GMP. Methyl methanesulphonate (MMS) and ethyl methanesulphonate (EMS), which do not possess a nitroso group, did not show evidence of nitric oxide release. All five compounds, however, decreased glucose-stimulated insulin release, suggesting that nitric oxide release was not necessary for the inhibition of secretion. Lack of involvement of nitric oxide was further suggested by the failure of oxyhaemoglobin to reverse STZ and MNU inhibition of insulin secretion. Since ENU was at least as effective as MNU in inhibiting insulin secretion, it appears that alkylation of DNA at the O6 position of guanine may not be involved in this process.

Developmental regulation of the human beta-globin cluster

Expression of the genes of the human beta-globin cluster is subject to stringent developmental regulation. An integral component of this control is competition between the gamma- and beta-promoters for the upstream regulatory sequences of the Locus Control Region (LCR). We have defined two stage selector elements (SSE and SSE-2), located in the proximal gamma-promoter and gamma-genes 5'UTR which mediate the preferential interaction of the gamma-gene with the LCR in the foetal developmental stage. The activity of both elements is dependent upon the binding of foetal and erythroid-specific proteins. We have purified the protein binding the proximal SSE and characterised it as a complex between the ubiquitous transcription factor CP2 and a 40 hD partner protein. Comparative studies between human CP2 and the chicken stage selector protein, NF-E4, demonstrate homology between the protein complexes. These findings demonstrate that competitive silencing of globin genes mediated by developmentally-specific protein complexes is conserved in evolution.

The renal glomerulus of mice lacking s-laminin/laminin beta 2: nephrosis despite molecular compensation by laminin beta 1

S-laminin/laminin beta 2, a homologue of the widely distributed laminin B1/beta 1 chain, is a major component of adult renal glomerular basement membrane (GBM). Immature GBM bears beta 1, which is replaced by beta 2 as development proceeds. In mutant mice that lack beta 2, the GBM remains rich in beta 1, suggesting that a feedback mechanism normally regulates GBM maturation. The beta 2-deficient GBM is structurally intact and contains normal complements of several collagenous and noncollagenous glycoproteins. However, mutant mice develop massive proteinuria due to failure of the glomerular filtration barrier. These results support the idea that laminin beta chains are functionally distinct although they assemble to form similar structures. Laminin beta 2-deficient mice may provide a model for human congenital or idiopathic nephrotic syndromes.

Identification of two novel regulatory elements within the 5'-untranslated region of the human A gamma-globin gene

Interaction between the stage selector element (SSE) in the proximal gamma-globin promoter and hypersensitivity site 2 in the locus control region partly mediates the competitive silencing of the beta-globin promoter in the fetal developmental stage. We have now demonstrated that a second SSE-like element in the 5'-untranslated region of the gamma-gene also contributes to this competitive silencing of the beta-gene. Utilizing transient transfection assays in the fetal erythroid cell line, K562, we have shown that the core enhancer of hypersensitivity site 2 can preferentially interact with the proximal gamma-promoter in the absence of the SSE, completely silencing a linked beta-promoter. Mutation of a 20-base pair sequence of the gamma-gene 5'-untranslated region (UTR) led to derepression of beta-promoter activity. A marked activation of gamma-promoter activity was also observed with this mutation, suggesting the presence of a repressor. Fine mutagenesis dissected these activities to different regions of the 5'-UTR. The stage selector activity was localized to a region centered on nucleotides +13 to +15. Electromobility shift assays utilizing this sequence demonstrated binding of a fetal and erythroid-specific protein. The repressor activity of the 5'-UTR was localized to tandem GATA-like sites, which appear to bind a complex of two proteins, one of which is the erythroid transcription factor GATA-1. These results indicate that the 5'-UTR of the gamma-gene contains sequences that may be important for its transcriptional and developmental regulation.

Novel pattern of P53 mutation in breast cancers from Austrian women

Since mutagens produce an extraordinary diversity of mutational patterns, differential mutational exposures among populations are expected to produce different patterns of mutation. Classical epidemiological methods have been successful in implicating specific mutagens in cancers such as those of lung and skin in which one mutagen predominates. In breast cancer, however, no mutagens have been implicated in an unequivocal manner. In an attempt to facilitate epidemiological studies, we have been studying the pattern of p53 gene mutations in breast cancers from multiple populations with high and low breast cancer incidences. We previously reported that breast cancers from Midwest United States, predominantly rural Caucasian women, have a different pattern of p53 gene mutation from populations of Western European women. Herein, we analyze patterns of p53 mutations from Graz, Austria, another population with a high incidence of breast cancer. Among the 60 Austrian breast cancers analyzed, 14 (23%) have a p53 gene mutation in exons 5-9 or in adjacent splice junctions. Analysis of the patterns of mutation shows differences between the "Western European" profile and the Austrian and Midwest United States groups (P = 0.027 and 0.024, respectively). The Austrian pattern is characterized by a high frequency of A:T-->T:A transversions (P = 0.006). The presence of distinct patterns of mutation among the limited number of analyzed populations of Western European origin supports the idea that differential mutagenic exposure and/or genetic differences contribute to breast cancer mutagenesis among geographically distinct Caucasians of Western European origin.

Hemoglobin switching in man and chicken is mediated by a heteromeric complex between the ubiquitous transcription factor CP2 and a developmentally specific protein

The human stage selector protein (SSP) has been implicated in the developmental regulation of the globin genes. Binding of SSP to the stage selector element (SSE) in the proximal gamma-globin promoter is integral to the competitive silencing of a linked beta-promoter in embryonic/fetal stage erythroleukemia (K562) cells. We now report the biochemical purification of SSP from K562 cell nuclear extract and demonstrate that the ubiquitously expressed transcription factor CP2 is pivotal to, but not sufficient for, SSP binding activity. Although addition of anti-CP2 antiserum disrupts the formation of the SSP-SSE complex in the electrophoretic mobility shift assay (EMSA), recombinant CP2 fails to bind to the SSE. Binding of CP2 to the SSE requires a heterodimeric partner present in K562 cells. We have defined the molecular weight of the partner protein as 40-45 kDa in UV and protein cross-linking experiments. An element analogous to the human SSE has previously been demonstrated in the chicken beta A-gene-promoter. The effects of this element are dependent on the binding of the chicken stage selector protein, NF-E4. Comparative studies between human CP2 and chicken NF-E4 demonstrate homology between the protein complexes. SSP binds to the chicken SSE and formation of this complex is ablated by the addition of anti-CP2 antiserum or a monoclonal antibody to NF-E4. Western analysis of partially purified NF-E4 using anti-CP2 antiserum or the NF-E4 monoclonal antibody both demonstrate a dominant band at 66 kDa. Similarly, the NF-E4 antibody recognizes the 66 kDa human CP2 protein in Western analysis of the SSP-SSE complex.(ABSTRACT TRUNCATED AT 250 WORDS)

Gene transfer into human hemopoietic progenitor cells

Considerable progress is being made in the transfer of genetic material to hematopoietic stem cells. In this chapter we describe how gene transfer is being used to: mark marrow and peripheral blood progenitor cells prior to autologous transplantation, to track their fate on reinfusion and to detect contaminating tumorigenic cells; modulate immunocyte function--important in immunologic disorders and perhaps in cancer therapy; generate tumor vaccines from tumor cells isolated from marrow; correct single gene defects--the 'classical' concept of gene therapy; and finally to modify the drug sensitivity of progenitor cells--enabling them to resist the suppressive effects of cytotoxic drugs during cancer therapy and perhaps providing a mechanism for in vivo selection of gene modified cells.

Cytokines, nitric oxide and insulin secreting cells

Cytokines are a group of regulatory and immunomodulatory proteins involved in a number of physiological processes. Various disease states are believed to involve alteration of normal cytokine activity, including insulin-dependent diabetes mellitus, an autoimmune disease in which insulin secreting beta cells within pancreatic islets of Langerhans are selectively destroyed. Glucose-induced insulin secretion is inhibited by the cytokines interleukin-1 beta (IL-1 beta), interleukin-6 and tumour necrosis factor alpha (TNF) when combined with IL-1 beta in cultured rat islets, by IL-1 beta, TNF and interferon gamma in mouse islets, and by combined treatment of IL-1 beta, TNF and interferon gamma in human islets. Continued cytokine treatment in many cases leads to destruction of some, if not all, islet cells. A key factor in the inhibitory effect of IL-1 beta and TNF in rat islets is the generation of nitric oxide which inactivates enzymes such as aconitase and ribonucleotide reductase by formation of iron-nitrosyl complexes. This in turn may lead to reduced oxidation of glucose and synthesis of ATP and DNA respectively. The causes of cytokine-induced beta cell death are less well defined, but important factors may be nitric oxide-mediated DNA damage, depletion of NAD levels and toxic effects of oxygen free radicals and eicosanoids generated in addition to nitric oxide. Potentially important defence and repair responses induced by IL-1 beta treatment of rat islets are formation of heat shock protein, haem oxygenase, and superoxide dismutase. Other protective responses may be induction of cytokines and cytokine receptor antagonists.(ABSTRACT TRUNCATED AT 250 WORDS)

The regulatory element 3' to the A gamma-globin gene binds to the nuclear matrix and interacts with special A-T-rich binding protein 1 (SATB1), an SAR/MAR-associating region DNA binding protein

A cis-acting DNA regulatory element 3' to the A gamma-globin gene contains eight distinct regions of DNA-protein interaction distributed over 750 bp of DNA. The sequences of two foot-printed regions (sites I and IV) are A-T rich and generate a highly retarded complex on gel shift analysis with nuclear extract from human erythroleukemia (K562) cells. We have purified a 98-kD protein that reproduces this gel shift. Tryptic cleavage and peptide sequence analysis demonstrated that the 98-kD protein is identical to a recently cloned protein, special A-T-rich binding protein 1 (SATB1), that binds selectively to nuclear matrix/scaffold-associated regions of DNA (MARs/SARs). We have shown by functional analysis that the 3' A gamma regulatory element associates with the nuclear matrix. SATB1 mRNA was identified in K562 cells, and reverse transcriptase-polymerase chain reaction (RT-PCR) demonstrated its transcript in several other hematopoietic lines. Antisera to SATB1 caused ablation of the gel shift complex generated by both the crude nuclear extract and the purified 98-kD protein with the site I oligonucleotide. Furthermore, oligonucleotides that bind SATB1 inhibited formation of the site I gel shift complex when added as excess unlabeled competitor. An immunoblot analysis of the site I gel shift complex documented the presence of SATB1. Binding of SATB1 to two sites within the 3' A gamma regulatory element and its MAR/SAR activity suggests that this element may influence gene expression through interaction with the nuclear matrix.

The effect of nitric oxide donors on insulin secretion, cyclic GMP and cyclic AMP in rat islets of Langerhans and the insulin-secreting cell lines HIT-T15 and RINm5F

The aim of this study was to investigate whether short-term treatment with nitric oxide donors could mimic cytokine inhibition of insulin secretion. We tested the nitric oxide generating compounds 3-morpholinosydnonimine (SIN-1), S-nitroso-N-penicillamine (SNAP), S-nitrosoglutathione and hydroxylamine for their ability to inhibit insulin secretion, raise cyclic GMP and lower cyclic AMP levels in isolated rat islets of Langerhans and the insulin-secreting cell lines HIT-T15 and RINm5F. In islets, all nitric oxide donors inhibited glucose-induced insulin secretion and raised cyclic GMP levels. SIN-1 and S-nitrosoglutathione also reduced cyclic AMP, while SNAP and hydroxylamine had no effect. Insulin secretion in HIT-T15 cells was inhibited by SIN-1, SNAP and hydroxylamine and in RINm5F cells by hydroxylamine. Inhibition of HIT-T15 and RINm5F cell insulin secretion was not accompanied by an increase in cyclic GMP levels. The degree of inhibition of insulin secretion was unrelated to the extent of release of nitric oxide by the compounds as measured by nitrite and nitrate production. More effective inhibition by S-nitrosoglutathione and hydroxylamine versus SIN-1 and SNAP may be related to intracellular versus extracellular site of nitric oxide generation.

Detection of p53 antigen expression in cytologic preparations of ovarian carcinomas

We explored the value of p53 immunostaining in detection of malignant cells in cytologic analyses in ovarian cancer patients. Ninety-six percent of 23 tumor touch preparations had nuclear staining with PAb240: 44% had nuclear only, 26% had both nuclear and cytoplasmic staining, and 26% had cytoplasmic staining alone. Eight of 20 cytospins of peritoneal washings from ovarian cancer patients were PAb240 positive, 6 had cytoplasmic, and two had nuclear staining. Cytospins of fluid from 3/15 patients with no malignancy revealed cytoplasmic staining of mesothelial cells. The lack of concordance between staining of primary tumors and cytospins of peritoneal washings and the apparent reactivity in some benign samples may limit the utility of immunohistochemical detection of malignant cells in the peritoneal fluid of ovarian cancer patients.

Differential toxicity of alpha-keto hydroxypyridine iron chelators and desferrioxamine to human haemopoietic precursors in vitro

Compliance with iron chelation therapy improves life expectancy in transfusion-dependent haematological disorders. However, failure of compliance with parenteral desferrioxamine (DF) therapy and the expense incurred makes this drug unavailable for most patients in the developing world. We have been evaluating the orally active iron chelator 1,2-dimethyl-3-hydroxypyrid-4-one (L1) in both preclinical and clinical trials. Five patients have developed reversible agranulocytosis during treatment with this agent. We have now studied the effects of L1, other alpha-ketohydroxypyridines and DF on bone marrow myeloid progenitors using the CFU-GM system. The results show that L1 is less toxic than DF to normal bone marrow myeloid progenitors (ID50:130 mumol/l versus 7.9 mumol/l). The L1 ID50 is within the previously reported range of peak plasma values (80-450 mumol/l). When saturating concentrations of iron were added to the cultures, the mean toxicity of all the chelators was significantly decreased over the range of doses tested, e.g. L1 ID50, 567 mumol/l; DF ID50, > 1000 mumol/l. The toxicity of L1 in vitro was similar for marrows from 3 normal donors and for the recovery marrow from a patient with thalassaemia major who had experienced agranulocytosis. Further studies are required to elucidate the mechanisms of L1-induced agranulocytosis.

Characterization of the third member of the MCAT family of cationic amino acid transporters. Identification of a domain that determines the transport properties of the MCAT proteins

We have identified the third member of a family of cationic amino acid transporters in lipopolysaccharide-stimulated murine macrophages. The deduced amino acid sequence of this transporter is the same as MCAT-2 (mouse cationic amino acid transporter-2), the low affinity transporter expressed in hepatocytes, except for a stretch of 41 amino acids that connects the eighth and ninth membrane-spanning domains. These transporters apparently result from differential splicing of transcripts from a single gene and therefore have been named MCAT-2A (hepatocyte) and MCAT-2B (macrophage). Despite their similarity, MCAT-2B is saturated at one-fifth the arginine concentration, has a lower apparent Vmax, and is more sensitive to trans-stimulation than MCAT-2. Introduction of the unique regions of MCAT-2A and MCAT-2B into the equivalent portion of the related protein, MCAT-1, created chimeric transporters with properties most like the donor of this region. Our findings suggest these 41 amino acids contain a domain that binds the amino acid substrate during its translocation across the membrane.

N-linked glycosylation of the receptor for murine ecotropic retroviruses is altered in virus-infected cells

Pulse-chase labeling identified incomplete maturation of the cationic amino acid transporter that serves as the receptor for ecotropic retroviruses in infected NIH 3T3 fibroblasts. The molecular basis and functional consequences of this change were studied by expressing viral and receptor proteins in Xenopus oocytes. Expression of the ecotropic, but not polytropic, envelope surface protein, gp70, interfered with N-linked glycosylation of the permissive mouse, but not the nonpermissive, human transporter. Incomplete glycosylation may be a direct consequence of gp70 binding since the glycosylated residues, Asn223 and Asn229, are present in the envelope binding domain. Receptors bearing amino acid substitutions in Asn223 and/or Asn229 that prevent glycosylation function normally in virus infection and amino acid transport. However, gp70 expression sufficient to block binding of 125I-gp70 to the receptor in the plasma membrane decreased receptor-mediated arginine uptake by 50%. Taken together, these findings suggest that newly synthesized gp70 may bind to the receptor in the endoplasmic reticulum and prevent normal glycosylation during their transit to the plasma membrane.

Methylation-enhanced binding of Sp1 to the stage selector element of the human gamma-globin gene promoter may regulate development specificity of expression

The human gamma-globin gene promoter contains a stage selector element (SSE) responsible for preferential interaction of the promoter with a powerful erythroid-specific enhancer in the fetal developmental stage (S.M. Jane, P.A. Ney, E.F. Vanin, D.L. Gumucio, and A.W. Nienhuis. EMBO J. 11:2691-2699, 1992). The element binds two proteins, the ubiquitous activator Sp1 and a protein previously known as -50 gamma and now named the stage selector protein (SSP). Binding of the second protein correlates with SSE activity in transient-transfection assays. We now report that a de novo binding site for the SSP is created by the -202(C-->G) mutation that causes hereditary persistence of fetal hemoglobin (HPFH). This site functions in an analogous manner to the SSE in hybrid beta-promoter/reporter gene constructs transfected into K562 cells. In contrast, the wild-type -202 sequence, which fails to bind the SSP, is incapable of activating the beta-gene promoter. Both the -50 and -202 HPFH sites for SSP binding overlap a consensus sequence for the transcriptional regulator Sp1. In addition, both sites contain CpG dinucleotides that are contact bases for SSP. Since the gamma promoter is known to be hypomethylated in fetal cells but fully methylated at CpG residues in adult erythroid cells, we examined the effects of this DNA modification on protein binding to the two regions. Gel mobility shift assays with nuclear extract from K562 cells (which contain both Sp1 and SSP) demonstrate preferential binding of SSP to the SSE and HPFH sites under conditions in which probe was limiting. Methylation of the CpG residues reverses this preference only in the SSE site, with a marked increase in the binding of Sp1 at the expense of the SSP. Purified Sp1 binds with 10-fold higher affinity to the methylated than to the nonmethylated -50 probe but with the same affinity to the -202 HPFH probe. The methylation-induced preferential binding of Sp1 to the SSE at the expense of SSP may be part of the mechanism by which the gamma genes are repressed in normal adult erythroid cells. In cells containing the -202 HPFH mutation, the inability of Sp1 to displace SSP in the methylated state may explain the persistence of gamma-promoter activity and gamma-gene expression observed in adults with this mutation.

Immunotherapy with interleukin 2 after ABMT in AML

Myeloablative chemo (+/- radio) therapy and rescue with ABMT has been used as final consolidation therapy in 18 patients with AML in first remission. In seven (6 autologous, 1 syngeneic) marrow reinfusion was followed by iv IL-2. One patient, who commenced IL-2 4 days after BMT, died from pulmonary oedema due to the capillary leak syndrome. Thereafter, treatment with IL-2 was delayed until the platelet count reached 30 x 10(9)/l. All patients developed reversible hypotension (treated with infusion of colloid), but treatment was otherwise well tolerated. With 21-58 months (median 32 months) from the time of ABMT there has been one relapse (actuarial risk 17%, 95% confidence intervals (CI) 3-31%). The disease-free survival is 71% (95% CI 38-100%). Eleven patients with comparable remission induction and consolidation therapy, and an identical interval between diagnosis and ABMT (5-11 months, median 6 months) received an autograft without immunotherapy. With 24-45 months (median 29 months) follow-up the actuarial disease-free survival is 36% (95% CI 8-64%), the actuarial relapse risk is 54% (95% CI 18-90%). We conclude that immunotherapy given after ABMT to patients with AML in first remission when the platelet count exceeds 30 x 10(9)/l is safe and may induce an immunological environment which results in the elimination of residual leukaemia.

Identification of a low affinity, high capacity transporter of cationic amino acids in mouse liver

The liver regulates the supply of amino acids required for protein synthesis and intermediary metabolism between feeding and fasting in mammals. The flux of amino acids between the liver and other tissues is determined, in part, by the activity of specific carrier proteins. We have identified a carrier of the cationic amino acids arginine, lysine, and ornithine in mouse liver that is closely related to a previously identified transporter with the same substrate specificity expressed in nonhepatic tissues. Uptake studies were performed in Xenopus oocytes injected with cRNA encoding these proteins. The comparison of the two transporters in these studies demonstrated that, unlike the widely-expressed transporter, arginine uptake mediated by the liver carrier is significant only at substrate concentrations that exceed systemic plasma levels and is less dependent on the intracellular concentration of cationic amino acids. These properties enable hepatocytes expressing this carrier to remove excess cationic amino acids from the blood without interfering with their uptake by nonhepatic tissues that express the related transporter.

Envelope-binding domain in the cationic amino acid transporter determines the host range of ecotropic murine retroviruses

Infection of rodent cells by ecotropic type C retroviruses requires the expression of a cationic amino acid transporter composed of multiple membrane-spanning domains. By exchanging portions of cDNAs encoding the permissive mouse and nonpermissive human transporters and examining their abilities to specify virus infection upon expression in human 293 cells, we have identified the amino acid residues in the extracellular loop connecting the fifth and sixth membrane-spanning segments of the mouse transporter that are required for both envelope gp70 binding and infection. These findings strongly suggest that the role of the mouse transporter in determining infection is to provide an envelope-binding site. This role is analogous to those of host membrane proteins composed of a single membrane-spanning domain that serve as binding proteins or receptors for other enveloped viruses such as human immunodeficiency virus, Epstein-Barr virus, and murine and human coronaviruses.

Minus-strand DNA is present within murine type C ecotropic retroviruses prior to infection

Viral minus-strand DNA has been identified within ecotropic murine retroviruses prior to infection. The abundance of minus-strand DNA is inversely proportional to the distance from the primer binding site, suggesting that viral DNA is synthesized by reverse transcriptase with the genomic RNA as template. These findings demonstrate that replication of the retroviral genome is not initiated by infection and may begin after activation of reverse transcriptase by gag-pol cleavage during virus assembly.

Retroviral infection and expression of cationic amino acid transporters in rodent hepatocytes

The susceptibility of rodent hepatocytes to infection by mouse type C retroviruses was examined in vivo and in vitro and compared with the expression of two membrane proteins that function as transporters for the cationic amino acids CAT-1 and CAT-2. CAT-1 expression in rodents determines susceptibility to ecotropic retrovirus infection by serving as the virus receptor. Recently, it has been suggested that CAT-2 may be a receptor for amphotropic murine leukemia virus. In the present study, CAT-1 expression was observed in Hepa1, a cell line derived from a murine hepatoma, and in rat hepatocytes propagated on collagen monolayers in vitro but not in intact or regenerating rat liver in vivo. The expression of CAT-1 correlated with susceptibility to infection by an ecotropic retrovirus encoding beta-galactosidase. CAT-2 expression was observed in hepatocytes in vitro and in vivo, consistent with reports of infection of regenerating and cultured hepatocytes by amphotropic retroviruses. However, introduction of murine CAT-2 into nonpermissive Chinese hamster cells was not sufficient to confer susceptibility to amphotropic retrovirus infection, using a protocol that could easily demonstrate CAT-1-dependent infection by an ecotropic virus. Our data establish CAT-1 as a major determinant of ecotropic retrovirus infection in rodent hepatocytes and suggest that CAT-2 is not a receptor for viruses in the amphotropic subgroup.

Interleukin-1 beta effects on cyclic GMP and cyclic AMP in cultured rat islets of Langerhans-arginine-dependence and relationship to insulin secretion

When islets were cultured with interleukin-1 beta (1 or 100 pmol/l) for 12 h in arginine-containing medium, cyclic GMP levels were increased 1.6- and 4.5-fold respectively. The arginine analogue, N-omega-nitro-L-arginine methyl ester, which blocks nitric oxide formation and partially reverses inhibition of insulin secretion by 100 pmol/l interleukin-1 beta, largely, but not completely, blocked generation of cyclic GMP. Treatment of islets with 100 pmol/l interleukin-1 beta for 12 h significantly decreased islet cyclic AMP generation in the absence of isobutylmethylxanthine (from 13.1 +/- 0.7 to 9.3 +/- 0.8 fmol/micrograms islet protein), this fall was arginine-dependent and may have resulted from an effect on a cyclic AMP phosphodiesterase, since it was masked if isobutylmethylxanthine was present. Isobutylmethylxanthine (0.4 mmol/l) reduced the inhibitory potency of interleukin-1 beta in 15 h slightly but significantly from 80.5 to 59.0%. The morpholinosydnonimine SIN-1, which is a nitric oxide donor, inhibited insulin secretion, raised islet cyclic GMP and lowered cyclic AMP; its effects were similar to those of interleukin-1 beta. However, 6-anilinoquinoline-5,8-quinone, [LY83583 (1-10 mumol/l)], inhibited insulin secretion, and significantly decreased cyclic GMP while 8-bromocyclic GMP stimulated insulin secretion. Both low- and high-dose interleukin-1 beta treatment give a large arginine-dependent and a small, yet significant, arginine-independent increase in cyclic GMP. The inhibitory effect of SIN-1 or interleukin-1 beta on insulin secretion seems to depend to a small extent on decreased islet cyclic AMP, though sustained increases in nitric oxide or depleted islet GTP may directly affect the secretory process.

Considerations in the skeletonization technique of internal thoracic artery dissection

Skeletonization of the internal thoracic artery during myocardial revascularization procedures may provide some degree of protection from sternal wound infection in diabetic, obese, or pulmonary compromised patients when both internal thoracic arteries are used as conduits. Prior descriptions of the technique fail to provide specific details and possible pitfalls of the dissection. A method of skeletonization that has been used at the California Center for Cardiothoracic Surgery in more than 1,000 patients is presented in detail. Potential advantages and disadvantages are discussed.

A stable long-term hepatocyte culture system for studies of physiologic processes: cytokine stimulation of the acute phase response in rat and human hepatocytes

Prior studies on the in vitro hepatic acute phase response have involved either hepatoma cell lines or conventional short-term cultures of primary hepatocytes. No data are available on the response of primary hepatocytes in stable long-term culture systems. In this study, the acute phase response of rat and human hepatocytes in a new long-term culture system was examined in response to interleukin-6 (IL-6), interleukin-1 beta (IL-1 beta), and tumor necrosis factor alpha (TNF-alpha). The cultured cells were sandwiched between two layers of collagen in a (double-gel) configuration which has been shown to preserve both hepatocyte function and morphology over prolonged periods of time. The stability of this culture configuration enabled us to investigate, for the first time, the temporal aspects of the response in addition to the effects of the mediators on protein secretion. Exposure of rat hepatocytes to IL-6 after culture for 16 days resulted in a 2-fold reduction of albumin secretion and a 15-fold increase in the secretion rates of fibrinogen and alpha 2-macroglobulin. In all instances, the peak response occurred at 48 h after IL-6 exposure, and all protein secretion rates returned to pretreatment values within 5 days posttreatment. Changes in the mRNA levels of these proteins in response to IL-6 corresponded with those changes seen with the secreted products, indicating pretranslational regulation. Administration of IL-1 beta to rat hepatocyte produced a similar decline of albumin secretion and a 5-fold increase of fibrinogen secretion, whereas alpha 2-macroglobulin secretion remained undisturbed.(ABSTRACT TRUNCATED AT 250 WORDS)

Roles for the integrin VLA-4 and its counter receptor VCAM-1 in myogenesis

Mammalian myogenesis is biphasic: primary myoblasts fuse to form primary myotubes, then secondary myoblasts align along the primary myotubes and form secondary myotubes, which comprise most of adult muscle. We provide evidence that an integrin (VLA-4) and its counter receptor (VCAM-1) have a role in secondary myogenesis. Both receptors are synthesized by cultured muscle cells: VLA-4 is induced as myotubes form, whereas VCAM-1 is present on myoblasts and myotubes. In vivo, both molecules are expressed at sites of secondary myogenesis, VLA-4 on primary and secondary myotubes, and VCAM-1 on secondary myoblasts and on regions of secondary myotubes apposed to primary myotubes. These patterns suggest that VLA-4-VCAM-1 interactions influence alignment of secondary myoblasts along primary myotubes and/or the fusion of secondary myoblasts. In support of the latter possibility, antibodies to VLA-4 or VCAM-1 inhibit myotube formation in culture.

Molecular cloning and functional expression of an inducible nitric oxide synthase from a murine macrophage cell line

Macrophages activated by exposure to cytokines and/or to endotoxin produce nitric oxide (NO.), a free radical that is a mediator of the host response to infection. Activation induces the expression of nitric oxide synthase, the enzyme that catalyzes formation of NO. from L-arginine and molecular oxygen. We report the cloning of a cDNA encoding the inducible nitric oxide synthase from a murine macrophage cell line, RAW264.7, exposed to interferon-gamma and lipopolysaccharide. Oocytes injected with mRNA transcribed from this cDNA demonstrate arginine-dependent production of nitrite, a stable metabolite of NO.. Nitric production is blocked by the enzyme inhibitor, NG-monomethylarginine, and is independent of calcium/calmodulin. RAW264.7 cells demonstrate rapid accumulation of the nitric oxide synthase-encoding mRNAs upon activation. Comparison of the deduced amino acid sequence to the calcium/calmodulin-dependent nitric oxide synthase previously purified (Bredt, D. S., and Synder, S.H. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 682-685) and cloned (Bredt, D. S., Hwang, P. M., Glatt, C. E., Lowenstein, C., Reed, R. R., and Synder, S. H. (1991) nature 351, 714-718) from rat brain identifies shared binding sites for the cofactors NADPH and flavins in the C-terminal half of both proteins and an additional conserved region near the N terminus that may recognize L-arginine and/or contribute to the active site.

The human cationic amino acid transporter (ATRC1): Physical and genetic mapping to 13q12–q14

The product of the mouse Rec-1 locus is an integral membrane protein that determines susceptibility to infection by murine ecotropic retroviruses. Recently it has been determined that its role in normal cell metabolism is transport of the cationic amino acids, arginine, lysine, and ornithine across the plasma membrane. Southern blot analysis of genomic DNA from a panel of 48 mouse-human somatic cell hybrids assigned the human version of this gene, ATRC1, to chromosome 13. Chromosomal in situ hybridization localized the gene to 13q12-q14. A restriction fragment length polymorphism (RFLP) was detected with TaqI. There were two alleles with frequencies of 0.29 and 0.71. Pairwise linkage analysis established linkage between ATRC1 and ATP1AL1, D13S1, D13S6, D13S10, D13S11, D13S21, D13S22, D13S33, D13S36, and D13S37. Multilocus linkage analysis of five of the loci indicated that the most likely order of loci in this region was D13S11-ATP1AL1-ATRC1-D13S6-D13S33.

Microcarrier culture of hepatocytes in whole plasma for use in liver support bioreactors

Contact activation of plasma clotting may limit the use of some microcarrier types for hepatocyte attachment in a liver assist device. Activation by seven microcarrier types was studied in plasma containing 2-500 units heparin/mL. Clotting was activated by dextran (Cytodex 1 and 2) and collagen-coated (Cytodex 3) microcarriers at 2-25 units/mL (Cytodex 1) and 2-100 units/mL (Cytodex 2 and 3). There was no activation by polystyrene, gelatin, glass or fibronectin-coated polystyrene microcarriers. Compared with culture medium, incubation of HepG2 cells in plasma did not affect cell viability but increased cell number (56.4 versus 65.1 x 10(4) cells; P less than 0.05) and incorporation of [3H]-amino acids into protein (204913 versus 279624 dpm; P less than 0.05). Polystyrene-attached cells demonstrated time-linear protein synthesis, glucose and 7-ethoxycoumarin metabolism. We conclude that polystyrene-attached hepatocytes maintain viability and metabolic activity in plasma and are of potential use in a liver support bioreactor.

Transport of cationic amino acids by the mouse ecotropic retrovirus receptor

Susceptibility of rodent cells to infection by ecotropic murine leukaemia viruses (MuLV) is determined by binding of the virus envelope to a membrane receptor that has multiple membrane-spanning domains. Cells infected by ecotropic MuLV synthesize envelope protein, gp70, which binds to this receptor, thereby preventing additional infections. The consequences of envelope-MuLV receptor binding for the infected host cell have not been directly determined, partly because the cellular function of the MuLV receptor protein is unknown. Here we report a coincidence in the positions of the first eight putative membrane-spanning domains found in the virus receptor and in two related proteins, the arginine and histidine permeases of Saccharomyces cerevisiae (Fig. 1), but not in any other proteins identified by computer-based sequence comparison of the GenBank data base. Xenopus oocytes injected with receptor-encoding messenger RNA show increased uptake of L-arginine, L-lysine and L-ornithine. The transport properties and the expression pattern of the virus receptor behave in ways previously attributed to y+, the principal transporter of cationic L-amino acids in mammalian cells.

Dimethyl sulphoxide induces a reduced growth rate, altered cell morphology and increased epidermal-growth-factor binding in Hep G2 cells

Culture of Hep G2 cells in medium containing 2% (v/v) dimethyl sulphoxide (DMSO) resulted in a slowing of growth and a marked change in morphological appearance. By day 6, cultures containing DMSO had one-third the number of cells compared with parallel control cultures. Measurement of 125I-epidermal-growth-factor (EGF) binding to DMSO-treated cells revealed a striking time-dependent elevation in specific EGF binding to their cell surface. Increased binding was detectable within 24 h of the start of DMSO treatment, reaching, by 6 days, levels almost 25 times greater than those for control cells. Addition of EGF to DMSO-treated cells caused a rapid down-regulation of the EGF receptor, but did not alter their proliferation rate. Slowing of growth by other means, such as serum starvation, growth to confluence or culture in the presence of sodium butyrate, did not affect 125I-EGF binding, indicating a specific effect of DMSO on these cells.

Rostrocaudal variation of fiber type composition in rat intercostal muscles

We used the histochemical stain for ATPase to compare the fiber-type composition of rat internal and external intercostal muscles from thoracic (T) segments 2-5, 8, and 11. At each level, type II fibers were more numerous than type I fibers, type II B fibers were more numerous than II A fibers, and type I fibers were more numerous in external than in internal intercostals. However, fiber type composition varied from segment to segment. For example, the proportion of type II A fibers increased in a rostrocaudal gradient in internal but not external intercostals, and type I fibers were more prevalent at rostral and caudal than at intermediate levels in both internal and external intercostals. These results provide a basis for interpreting previous physiological and molecular studies which have compared intercostal muscles from different segmental levels.

A method for the purification of DNA/protein complexes applied to DNA topoisomerase II cleavage sites

The object of this study was to devise a purification method for DNA/topoisomerase II complexes, with which to examine the enzyme's cleavage site specificity in cellular differentiation. Retinoic acid-induced differentiation involves topoisomerase II-mediated transient changes in DNA supercoiling, but it is not known whether this occurs at specific sites in the genome. Topoisomerase II forms a covalent DNA enzyme complex as it acts, which can be recovered by the sodium dodecyl sulfate (SDS)/KCl precipitation method, but this method fails to recover significantly more DNA from cells induced to differentiate. This may in part reflect the low numbers of retinoic acid-induced protein-linked breaks in DNA and also the method's relative inefficiency for DNA with few attached topoisomerase molecules. This suggested that an additional purification method would be required to enrich sufficiently for cleavage site DNA to address the issue of site specificity. The principle of our method is to couple poly(ethylene glycol) (PEG) to topoisomerase while it is covalently attached to DNA and then to use phase partitioning in an aqueous two-phase system of PEG and phosphate to separate free DNA from DNA bound to PEG-modified topoisomerases (which have high affinities for the phosphate-rich and PEG-rich phases, respectively). The method can be used in conjunction with DNase protection and, unlike the SDS/KCl method, can fractionate short fragments of DNA to which single protein molecules are attached. Using the SDS/KCl precipitation and new method in series, we have recovered protein-linked DNA from HL60 cells induced to differentiate to the granulocyte lineage (by retinoic acid) or to the monocyte/macrophage lineage (by phorbol myristate acetate) and have demonstrated that specific sequences become protein linked, probably to topoisomerase II, during induced differentiation.

Aberrant DNA topoisomerase II activity, radioresistance and inherited susceptibility to cancer

Inherited susceptibility to a wide variety of neoplasias (Li-Fraumeni syndrome), has been shown in studies of one cancer-prone family, to have an intriguing association with an aberrant c-raf-1 gene and inheritance of a radioresistant phenotype in their non-cancerous skin fibroblasts. This association together with observations that DNA topoisomerases, when defective, can introduce errors into DNA and that these enzymes are perturbed in vitro by serine/threonine kinases similar to raf encoded proteins, prompted investigation of DNA topoisomerase activity of the family's fibroblasts. Since radioresistance was transferred to murine cells (NIH-3T3) when the aberrant c-raf-1 gene from this family was transfected, we also examined transformants containing this and other oncogenes. V-raf/c-myc and EJ-ras transformants were examined, the former because the family's skin fibroblasts also have 3-8-fold elevated myc expression (not apparently relevant to radioresistance) and the latter because ras, like raf, conveys radioresistance. The family members' fibroblasts and the three transfected murine lines, showed a similar perturbation of a spermidine and ATP-dependent DNA catenation activity (typical of DNA topoisomerase II). There was a significant positive correlation (r = 0.93; P = 0.0026) between the degree of activation of topoisomerase II and one measure of radioresistance (the Dq value). Relaxation of DNA supercoiling (topoisomerase I activity and other DNA nicking enzymes) was not abnormal. Cytotoxicity assays and evaluation of the influence of topoisomerase II inhibitors on DNA/protein complex formation, corroborated the existence of a qualitative topoisomerase II defect in the family's cells and transfectants. Although the contention that the qualitative topoisomerase II abnormalities observed here may be associated with malfunction is highly speculative, these findings may be relevant to the mechanism of oncogenesis, not only in this family, but with raf and ras type oncogenes.

Human cells infected with retrovirus vectors acquire an endogenous murine provirus

A 5.2-kilobase mouse RNA is expressed in human cells following infection with recombinant retroviruses propagated in mouse NIH 3T3 cells as psi-2 pseudotypes. This RNA is transcribed from a defective mink cell focus-forming provirus and copackaged into virions and integrated into human target cell DNA at a frequency comparable to that of the recombinant retrovirus genome.

Thrombocytopenia--radial aplasia (TAR) syndrome with associated immune thrombocytopenia

We report a 10 year old girl with congenital thrombocytopenia with absent radii (TAR) syndrome. Unusually this patient has shown no abnormal bleeding tendency despite numerous orthopaedic procedures. There is evidence of an associated immune component to the thrombocytopenia which has not been previously documented.

A putative murine ecotropic retrovirus receptor gene encodes a multiple membrane-spanning protein and confers susceptibility to virus infection

Murine type C ecotropic retrovirus infection is initiated by virus envelope binding to a membrane receptor expressed on mouse cells. We have identified a cDNA clone that may encode for this receptor through a strategy combining gene transfer of mouse NIH 3T3 DNA into nonpermissive human EJ cells, selection of EJ clones that have acquired susceptibility to infection by retrovirus vectors containing drug resistance genes, and identification of the putative receptor cDNA clone through linkage to a mouse repetitive DNA sequence. Human EJ cells that express the cDNA acquire a million-fold increase in MuLV infectivity. The predicted 622 amino acid sequence of the putative receptor protein is extremely hydrophobic; 14 potential membrane-spanning domains have been identified. A computer-based search of sequence data banks did not identify a protein with significant similarity to the putative receptor. We conclude that a novel membrane protein determines susceptibility to ecotropic MuLV infection by binding and/or fusion with the virus envelope.

Neurite outgrowth on cryostat sections of innervated and denervated skeletal muscle

To localize factors that guide axons reinnervating skeletal muscle, we cultured ciliary ganglion neurons on cryostat sections of innervated and denervated adult muscle. Neurons extended neurites on sections of muscle (and several other tissues), generally in close apposition to sectioned cell surfaces. Average neurite length was greater on sections of denervated than on sections of innervated muscle, supporting the existence of functionally important differences between innervated and denervated muscle fiber surfaces. Furthermore, outgrowth was greater on sections of denervated muscle cut from endplate-rich regions than on sections from endplate-free regions, suggesting that a neurite outgrowth-promoting factor is concentrated near synapses. Finally, 80% of the neurites that contacted original synaptic sites (which are known to be preferentially reinnervated by regenerating axons in vivo) terminated precisely at those contacts, thereby demonstrating a specific response to components concentrated at endplates. Together, these results support the hypothesis that denervated muscles use cell surface (membrane and matrix) molecules to inform regenerating axons of their state of innervation and proximity to synaptic sites.

The regulated expression of erythropoietin by two human hepatoma cell lines

The development of a cell culture system that produces erythropoietin (Epo) in a regulated manner has been the focus of much effort. We have screened multiple renal and hepatic cell lines (including MDCK, LLC-PK1, BHK, WRL 68, CLCL, A704, CRFK, A498, ACHN, TCMK-1, LLC-MK2, CaKi-2, HepG2, and Hep3B) for either constitutive or regulated expression of Epo. Only the human hepatoma cell lines, Hep3B and HepG2, made significant amounts of Epo as measured both by radioimmunoassay and in vitro bioassay (as much as 330 milliunits per 10(6) cells in 24 hr). The constitutive production of Epo increased dramatically as a function of cell density in both cell lines. At cell densities less than 3.3 X 10(5) cells per cm2, there was little constitutive release of Epo in the medium (less than 30 milliunits per 10(6) cells in 24 hr). With Hep3B cells grown at low cell densities, a mean 18-fold increase in Epo expression was seen in response to hypoxia and a 6-fold increase was observed in response to incubation in medium containing 50 microM cobalt(II) chloride. At similar low cell densities, Epo production in HepG2 cells could be enhanced an average of about 3-fold by stimulation with either hypoxia or cobalt(II) chloride. Upon such stimulation, both cell lines demonstrated markedly elevated levels of Epo mRNA. Hence, both Hep3B and HepG2 cell lines provide an excellent in vitro system in which to study the physiological regulation of Epo expression.