Erythropoietin is produced by tubular cells of the rat kidney.

The cellular site of erythropoietin (epo) production within the mammalian kidney is still not completely understood. In the present study, we examined the expression of epo mRNA in microdissected rat nephron segments by RT-PCR after induction of epo expression with cobalt chloride. Erythropoietin mRNA was not detected in nephron segments from saline injected rats. In cobalt chloride injected animals, epo mRNA was found in the majority of samples from the cortical region of the nephron, PCT, and CAL. Medullary tubule preparations (MCT and MAL) were mostly negative for epo mRNA, and glomeruli were uniformly negative. The induction of epo transcripts in tubular cells by cobalt chloride was paralleled by stimulation of the major transport enzyme in the kidney, namely, Na-K ATPase in a tubular profile similar to that of induction of epo transcripts. These results support some earlier findings that epo gene expression in response to cobalt salt stimulation of rat kidney occurs in transporting tubular epithelial cells.

Hemin or butyrate increases constitutive erythropoietin formation by mouse erythroleukemia cell lines.

The murine erythroleukemic cell lines NN10 and IW32 secrete erythropoietin (Epo) constitutively. Although both cell lines have Epo receptors (Epo-R), they do not respond to external Epo to undergo erythroid differentiation. In both cell lines, hemoglobin synthesis can be induced by hemin or butyrate. We report here on the effects of these two inducers on Epo production by the same cell lines. Both hemin and butyrate caused a three- to five-fold increase in Epo secretion on a percell basis. The effects were dose- and time-dependent, as was found previously for hemoglobin synthesis. The mechanism of the effect on Epo secretion is posttranscriptional, since the steady-state level of Epo mRNA was not altered by the inducers.

An autocrine role for erythropoietin in mouse hematopoietic cell differentiation.

Erythropoietin (epo) is the primary regulator of the rate of red blood cell formation in mammals. Because it is formed in the kidney and acts on the bone marrow, its action is classically endocrine. We have shown by PCR that marrow cells contain epo mRNA and that antisense oligodeoxynucleotides, to both epo and its receptor, act on multipotent hematopoietic cells to cause a decrease in mixed erythroid:nonerythroid colonies. The antisense oligonucleotides also cause an increase in mixed nonerythroid colonies with no effect on erythroid burst formation. Sense oligonucleotides have no effect. The antisense suppression is not due to adherent cells, cycling late differentiated cells or lymphocytes, and not reversed by exogenous epo. We conclude that normal erythroid differentiation may have an early phase that is dependent on an internal autocrine mechanism involving epo and its receptor.

Co-regulation of heme oxygenase and erythropoietin genes.

The mechanism responsible for the accumulation of heme oxygenase and erythropoietin (epo) transcripts due to cobalt chloride (CoCl2) administration was investigated in rat kidney using a rat heme oxygenase and mouse epo probes. We found an increase of heme oxygenase transcripts in kidney in response to CoCl2. Quantitative evaluation of the heme oxygenase mRNA changes, by scanning densitometry, indicated that the levels of mRNA encoding heme oxygenase were increased by about fiftyfold in rat kidney after administration of CoCl2. That the increase in heme oxygenase mRNA levels resulted from enhanced transcription of the heme oxygenase gene was confirmed by nuclear runoff using isolated rat kidney nuclei after CoCl2 administration. Transcription of the heme oxygenase gene is greatly increased in rat kidney within 1 hr of administration of CoCl2 as evidenced from the levels of 32P-UTP incorporation into the specific transcript. Time course studies showed that stimulation of transcription was increased about fortyfold 3 hr after CoCl2 administration. This stimulation is the most rapid transcriptional response to heavy metals yet described. In addition, Northern blot analysis demonstrated that epo mRNA was first detected 4 hr following CoCl2 administration and reached a maximum at 5 hr. On the other hand, PCR analysis indicated that epo mRNA was increased as early as 1 hr following CoCl2 administration. The fact that CoCl2 caused increased transcription of both the epo and heme oxygenase genes suggests that a common mechanism may be involved in the regulation of these two genes by the heavy metal ion.

Evidence suggesting negative regulation of the erythropoietin gene by ribonucleoprotein.

The promoter regions of the mouse and human erythropoietin genes have regions of identity within 130 base pairs upstream of the cap site, suggesting a cis-acting regulatory role for the conserved sequences. We have used a double-stranded deoxyoligonucleotide corresponding to the -61 to -45 region relative to the start site of transcription of the mouse gene in DNA mobility shift assays. Nuclear extracts from kidneys of both control and cobalt-stimulated mice contain factors that bind to this oligonucleotide in a specific manner. One factor is a 47-kDa protein, whereas the others may be one or more ribonucleoproteins. Under denaturing conditions, four RNA species which show specific binding to the oligonucleotide were observed, suggesting that recognition of the oligonucleotide by ribonucleoprotein is mediated by the RNA component. In nuclear extracts of kidneys from stimulated animals, the amount of the two largest RNA species that bind to the oligonucleotide was reduced relative to that of control, whereas the other RNA species as well as the 47-kDa protein remained relatively unaffected. These results suggest that the ribonucleoprotein containing the down-regulated RNA species may be a negative transcriptional factor and that activation of the erythropoietin gene by cobalt salts may involve, in part, decreased binding of this factor, thus allowing transcription to proceed.

Differential expression of alpha- and beta-globin genes in erythroleukemic cell lines.

The IW32, NN10, and IW201 cell lines are erythroleukemic cell lines isolated from the spleens of mice infected with the Friend virus. IW32 and NN10 cells can be induced toward erythroid differentiation and hemoglobin synthesis by hemin or butyrate. Both cell lines contain some mature alpha- and beta-globin mRNA before induction, and addition of the inducers greatly increases the amount of globin message. Unlike IW32 and NN10 cells, IW201 cells are only partially inducible. Uninduced 201 cells contain a small amount of alpha-globin mRNA but no detectable beta-globin message. After induction, the cells contain markedly increased amounts of alpha-globin mRNA but still do not express the beta-globin gene. Southern blot analysis with 10 restriction enzymes shows that the restriction map of the beta-globin gene in IW201 cells is indistinguishable from that in IW32 and NN10 cells.

Activation of the erythropoietin gene due to the proximity of an expressed gene.

The erythroleukemic cell line IW32 has a normal erythropoietin (epo) gene and one which has undergone rearrangement and amplification. The rearranged epo locus consists of a breakpoint approximately 1 kb upstream from an otherwise structurally unchanged epo gene. DNase I hypersensitivity studies had suggested that the rearranged gene is the transcriptionally active one. To understand this transcriptional activation more fully we have sequenced the upstream regions of both the normal and the rearranged epo genes, the latter encompassing the rearrangement breakpoint. Sequence analysis showed that the new sequence past the breakpoint at the rearranged locus contained all the promoter components needed for a gene in opposite transcriptional orientation to the epo gene. Subsequent Northern blot analysis showed that this gene is transcriptionally active. No homology was found between this gene and any sequence in the data banks. It appears that the transcriptional activation of the rearranged epo gene in IW32 cells has been mediated by a translocation event which has served to bring the epo gene into close juxtaposition to this transcriptionally active gene.

Studies of the constitutive expression of the mouse erythropoietin gene.

The IW32 and NN10 cell lines are erythroleukemic cell lines which secrete erythropoietin (epo) into the culture medium constitutively. IW32 cells have a rearranged and amplified epo gene in addition to the normal gene. NN10 cells contain only the normal epo gene. DNase I hypersensitivity studies suggested that, in IW32 cells, it is the rearranged gene which is transcriptionally active. Sequence analysis of the upstream region of the rearranged epo gene suggests that the rearrangement has served to introduce a transcriptionally active gene close to the epo gene. This juxtaposition may explain the transcriptional activation of the epo gene at the rearranged locus.

Induction of globin mRNA transcription by erythropoietin in differentiating erythroid precursor cells.

The effects of erythropoietin (epo) on the proliferation of late erythroid progenitor cells (CFU-E) and on the formation of hemoglobin and of globin mRNA in these cells are described. CFU-E were isolated from thiamphenicol-pretreated anemic mice by elutriation and Percoll density gradient methods. These CFU-E are restricted in their capacity to proliferate in vitro without added epo. The epo dependence in vitro was not absolute. With no epo in the culture medium the first cell division was unimpaired, whereas the third division was only 1%-2% of the control. In the absence of epo the synthesis of hemoglobin is very low in CFU-E, but is increased significantly after about 5 h of incubation with epo present. In epo deprived cells there was considerable hemoglobin formed at about 14 h, but not earlier. The presence as detected by the Northern blot technique of globin mRNA, isolated from CFU-E, was variable, probably depending on the presence of some more mature erythroid cells. By an extrapolation method we show evidence that pure CFU-E would have virtually no detectable globin mRNA. The production of globin mRNA is rapidly (2 h) induced in cells incubated with epo. We conclude that epo, besides having a mitogenic effect on CFU-E, induces the rapid expression of the globin genes.

Expression of the erythropoietin gene.

Comparison of maps of the human and mouse erythropoietin (EPO) genes shows overall general conservation. A cloned mouse EPO gene was used to study EPO gene expression by the Northern blot method. Ten hours after bleeding to induce stress erythropoiesis in the rat, the only tissue to show detectable message was the kidney. Other preliminary studies on EPO expression in cells transformed by FMuLV that secrete EPO constitutively showed no detectable viral sequences close to the gene regardless of high levels of EPO expression. Further studies on regulation of EPO gene expression are now possible with the probes developed from the cloned gene.

Rearrangement and expression of erythropoietin genes in transformed mouse cells.

The erythroleukemia cell line IW32, derived by transformation with the Friend murine leukemia virus, has been shown previously to produce erythropoietin (EPO) constitutively. Here we demonstrate that, in addition to the normal mouse EPO locus, this cell line has another EPO locus which has undergone rearrangement and amplification. Both loci were cloned, and the rearrangement breakpoint of the second EPO locus was located within a 1.1-kilobase region upstream of an otherwise apparently normal EPO gene. There are no viral sequences present in the immediate vicinity of the rearranged EPO gene. DNase I digestion studies suggest that the rearranged gene is in a region where the chromatin is more sensitive to DNase hydrolysis than is the site of the normal gene. We conclude, tentatively, that the rearranged EPO locus is probably the transcriptionally active one and that either proviral sequences are acting at a distance to activate the EPO gene or the rearrangement itself has served to activate the gene.

Expression of the erythropoietin gene.

Injection of cobalt into rats resulted in erythropoietin (EPO) mRNA accumulation in the kidney. The same response was obtained upon bleeding. No EPO mRNA was detected in the spleen, salivary gland, or thymus following cobalt injection or bleeding. In some animals, but not in others, EPO mRNA was also expressed in the liver in response to cobalt injection. Time course studies showed that message appearance begins sometime between 3 and 6 h after cobalt injection. This correlated very well with the EPO concentration in the circulation; EPO levels in the circulation were the same as those of controls at 3 h but increased to six- to sevenfold that of controls by 6 h after cobalt injection. The mature EPO mRNA in the rat and mouse comigrated with the 18S rRNA, indicating that it is about 1,850 nucleotides in length.

The regulation of heme biosynthesis during erythropoietin-induced erythroid differentiation.

Marrow cells induced toward erythroid differentiation by treatment with erythropoietin respond by increasing the rates of iron uptake and hemoglobin synthesis. Study of the enzymes of heme biosynthesis during erythroid differentiation suggests that induction of heme synthesis in these cells is regulated by synthesis of porphobilinogen deaminase. The activities of delta-aminolevulinic acid synthase, gamma, delta-dioxovaleric acid transaminase, delta-aminolevulinic acid dehydratase, and ferrochelatase were not affected significantly by treatment of suppressed marrow cells with erythropoietin over a period of 4 days, whereas that of porphobilinogen deaminase was increased by as much as 3.5-fold by the 3rd day of incubation. The time course of increase in porphobilinogen deaminase activity was parallel to that of the increase in heme synthesis. Moreover, when porphobilinogen deaminase activity was compared in marrow cells exposed to increased levels of erythropoietin in vivo (hyperplastic marrow) and marrow cells exposed to lowered levels of erythropoietin in vivo (suppressed marrow), the activity in the former case was greater than that in normal cells and for the latter type of cell it was lower than normal. Experiments using actinomycin D and cycloheximide suggest that transcription is required for the erythropoietin-induced porphobilinogen deaminase activity, indicating that induction is probably at the level of de novo synthesis of enzyme.