Homeobox genes in normal hematopoiesis and leukemia.
Explore the source record for details and available documents.
Biomedical subjects
Publications and source records attributed to C Largman.
Explore the source record for details and available documents.
Transketolase (EC 2.2.1.1) is the enzyme that, together with aldolase, forms a reversible link between the glycolytic and pentose phosphate pathways. We have cloned and sequenced the transketolase gene from yeast (Saccharomyces cerevisiae). This is the first transketolase gene of the pentose phosphate shunt to be sequenced from any source. The molecular mass of the proposed translated protein is 73,976 daltons, in good agreement with the observed molecular mass of about 75,000 daltons. The 5'-nontranslated region of the gene is similar to other yeast genes. There is no evidence of 5'-splice junctions or branch points in the sequence. The 3'-nontranslated region contains the polyadenylation signal (AATAAA), 80 base pairs downstream from the termination codon. A high degree of homology is found between yeast transketolase and dihydroxyacetone synthase (formaldehyde transketolase) from the yeast Hansenula polymorpha. The overall sequence identity between these two proteins is 37%, with four regions of much greater similarity. The regions from amino acid residues 98-131, 157-182, 410-433, and 474-489 have sequence identities of 74%, 66%, 83%, and 82%, respectively. One of these regions (157-182) includes a possible thiamin pyrophosphate (TPP) binding domain, and another (410-433) may contain the catalytic domain.
We have previously reported that certain genes of the HOX2 cluster of homeobox genes on human chromosome 17 are specifically expressed in human leukemic cell lines with erythroid potential, suggesting that these genes are involved in hematopoietic differentiation. We now show that the expression of the HOX 2.2 gene decreases during erythropoietin-induced differentiation of the erythroid cell line MB02. In order to study the role of the HOX 2.2 homeobox gene in hematopoiesis, vectors producing sense or antisense transcripts were introduced into K562 and HEL cells, pluripotent lines with erythroid and myeloid features. Overexpression of HOX 2.2 is associated with loss of erythroid features in both lines and an increase in certain myelomonocytic markers in K562 cells. Expression of antisense HOX 2.2 is associated with an increase in erythroid features in HEL cells and a mild decrease in myeloid characteristics in K562 cells. Overexpression of the adjacent HOX 2.1 gene in K562 cells does not produce similar phenotype changes. These data demonstrate that modulation of a specific HOX 2 homeobox gene can change the phenotype of somatic cells and suggest that certain HOX 2 genes play a role in blood cell differentiation.
The expression of the genes in the human HOX2 locus has been studied during differentiation of two human neuroblastoma (SH-SY5Y and Kelly), a human glioblastoma (251-MG), and the murine F9 embryonal carcinoma cell lines. Cells were differentiated with retinoic acid (RA), or with RA together with dibutyral cyclic AMP (db-cAMP) and nerve growth factor (NGF) in order to assess the changes in the expression patterns of these homeobox genes during neuronal differentiation. We show that the genes of the HOX2 locus are expressed in a complex transcription pattern that varies with cell type. The two uninduced neuroblastoma cell lines show a similar pattern of expression for a number of HOX2 genes although the levels of expression are different for individual cell lines. The embryonal carcinoma cell line F9 expresses low levels of several HOX2 genes which is restricted to the 5' region of the HOX2 cluster. The glioblastoma cell line, 251-MG expresses almost all of the genes of the HOX2 locus. Differentiation of these cells modulates the expression of the HOX2 genes in a manner that is dependent upon the cell type as well as the differentiation factor. Differentiation affects both the level of HOX2 gene expression and the distribution of transcript sizes. In conclusion, our analysis reveals a complex pattern of expression for the genes of the HOX2 locus in neuronal and glial cells and suggests that the cell-specific expression of these genes may be correlated with the phenotypic differences that are observed between different neuronal and glial cell populations within the nervous system.
We have previously reported that certain members of the HOX 1 and HOX 2 clusters of class 1 homeobox-containing genes showed lineage-restricted patterns of expression in a small series of human hematopoietic cell lines. We now report on the expression patterns of the entire HOX 2 cluster, consisting of nine homeobox genes, in a broad survey of leukemic cell lines of different phenotypes. The most striking observation is that all but one of the HOX 2 genes are consistently expressed in cells with erythroid character and/or potential, but, with rare exception, not in cells with myelomonocytic or T- or B-lymphoid phenotype. By contrast, several genes of the HOX 1 and 3 loci are not expressed in erythroid lines. Within erythroid cell lines, many of the HOX 2 genes are expressed as multiple transcripts. Expression of some HOX 2 genes is detectable in normal human marrow. These data show that in human hematopoietic cell lines HOX 2 homeobox gene expression is largely restricted to cells of erythroid phenotype and suggest that these genes play a role in erythropoiesis.
As part of a survey of the expression of homeobox-containing genes in human hematopoietic cells, we identified a novel gene (PL1) expressed only in cells of the myelomonocytic lineage (Shen et al., Proc. Natl. Acad. Sci, USA 86, 8536, 1989). On Northern gel analysis, major transcripts of 3.0 and 2.2 kb length are observed. Alternatively spliced homeobox-containing cDNAs, corresponding to the major transcripts, have been cloned from two myeloid leukemia cell libraries. The two cDNAs share the homeodomain and 3' flanking region but have unique 5' flanking regions. The longer transcript, would encode a 496 amino acid homeobox-containing protein, while the shorter message would encode a 94 amino acid homeobox-containing protein lacking the extended amino-terminal region. These two transcripts are differentially expressed in human leukemia cell lines. The larger transcript is exclusively expressed in cells with myelomonocytic features, while the smaller transcript is expressed in a variety of hematopoietic cell types. PL mRNA is also detectable in normal human bone marrow by RNAse protection. Neither transcript is expressed in uninduced teratocarcinoma cells or in the adult human tissues surveyed. The homeodomain is identical to the genomic sequence for Hox 1H, a newly identified member of the Hox 1 locus (Acampora et al. Nucl. Acids Res. 17, 10385, 1989). The PL1 gene was localized to chromosome 7 using chromosome specific blots and sublocalized to region pI4-21 by in situ hybridization of chromosomal spreads, confirming its location within the Hox 1 complex.
The HOX 2.2 homeobox gene is expressed in human hematopoietic cell lines with erythroid features (W.-F. Shen, et al, Proc. Natl. Acad. Sci. 86, 8536-8540, 1989). Both human and murine Hox 2.2 genes contain a single 1 kb intron which interrupts the sequence encoding the proposed homeobox protein. Four human erythroleukemia cell lines express the spliced, homeobox-coding transcript as the major form of message, and variable low amounts of unspliced HOX 2.2 mRNAs. Murine embryonic tissues and adult kidney and uterus contain approximately equal amounts of transcripts containing this intron and mRNAs from which the intron has been excised. The spliced transcript encodes a 224 amino acid homeobox protein, while the unspliced transcript would potentially encode a 140 residue protein containing the same N-terminal sequence but lacking the homeodomain.
Plasma levels of amylase, lipase, and immunoreactive cationic trypsinogen (ICT) were monitored in conscious rats to study the effects of cholecystokinin octapeptide (CCK-8) plus secretin administration, cephalic stimuli, and food intake. Stepwise increasing doses of CCK-8 (1, 5, 15, 30 Ivy dog units: IDU/kg/h) caused significant dose-related increases in plasma levels of each enzyme in a similar manner as those previously observed for response patterns of CCK-induced exocrine protein secretion in the same species. ICT showed the greatest response to CCK-8 with a maximal concentration 30 times above basal levels resulting in the steepest slope of the dose-response curve. Plasma lipase showed a maximal response that was 5 times above the basal level, while the plasma amylase level was increased only by 50% at the maximal response. Computed ED50 of CCK-8 for each enzyme confirmed this relative sensitivity of the response: 3.0, 8.7, 11.0 IDU/kg/h for ICT, lipase, and amylase, respectively. Plasma levels of amylase and lipase did not change significantly in response to the intake of either a liquid diet or fiber pellets containing no caloric value. Plasma ICT levels, however, were elevated significantly by 23 and 53% at the time when the liquid diet or fiber pellets were given and when cephalic stimulation appeared maximally induced. The increased levels declined thereafter and were no longer significantly different from the basal levels for postprandial 2 h.(ABSTRACT TRUNCATED AT 250 WORDS)
Effects of chronic alcohol intake on secretion and synthesis of amylase and cationic trypsinogen (CT) were studied with pancreatic acini isolated from male Sprague-Dawley rats fed a Lieber-DeCarli ethanol or control diet for 30 days. Pancreatic acini were incubated in a media containing increasing concentrations of cholecystokinin octapeptide (CCK-8: 0-1000pM) followed by addition of [3H]leucine. Amylase and CT secreted in the media and those labeled in acini were quantitated. Basal and CCK-stimulated secretion of CT was not different in alcoholic and control groups. On the other hand, a dose-response curve of CCK-stimulated amylase secretion from alcoholic acini was markedly reduced with both basal and maximal secretion decreased to only 40% of controls. Basal incorporation of [3H]leucine into amylase was reduced by 70% in alcoholic acini compared to controls while that into CT was not different in the two groups. CCK-8 exhibited a biphasic effect on [3H]leucine incorporation into both enzymes in alcoholic acini: low concentrations of CCK-8 (less than 100pM) increased the incorporation whereas high concentrations (greater than 100pM) decreased it. However, in control acini, CCK-8 induced progressive suppression of the incorporation into these enzymes, the pattern of which was similar to that previously observed in fasted rats (Am. J. Physiol. 241:G116-G112, 1981). This difference in the response pattern resulted in significantly higher rates of CCK-8 induced incorporation into CT in alcoholic acini. These results suggest that the differences observed may possibly be attributable not only to alcohol intake but also to the differences in carbohydrate intake and in temporal patterns of diet consumption.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
A specific radioimmunoassay has been developed for human pancreatic cationic trypsin. The assay has been employed for the determination of immunoreactive forms of pancreatic cationic trypsin in blood. The trypsin employed as radioiodinated tracer in the assay was inactivated with tosyl-L-lysine chloromethyl ketone (TLCK) to prevent binding of the tracer to the serum inhibitors while maintaining its immunoreactivity. The average normal serum level determined was 26 ng/ml, with a range of 12--41 ng/ml. Eight of nine patients with acute pancreatic inflammation had at least a 15-fold elevation of total serum immunoreactive cationic trypsin. Cationic trypsinogen and cationic trypsin bound to alpha1-antitrypsin cross-react strongly in the radioimmunoassay. Thus it is possible to measure these potential molecular forms of cationic trypsin in serum. When normal human serum was fractionated on Sephadex G-200, all of the immunoreactive material eluted as a single peak of approximately 23,000 mol wt. No cationic trypsin could be detected in association with alpha1-antitrypsin or alpha2-macroglobulin. The 23,000-mol-wt peak was definitively shown to contain trypsinogen by affinity chromatography and by activation with human enteropeptidase. The identification of cationic trypsinogen in blood implies that the zymogen is secreted into the circulation by the pancreas rather than entering the bloodstream via absorption from the intestine.
The molecular forms of immunoreactive pancreatic cationic trypsin in sera of patients with acute pancreatic inflammation have been characterized using a radioimmunoassay technique that is capable of detecting trypsinogen as well as trypsin bound to alpha 1-antitrypsin. Trypsin bound to alpha 2-macroglobulin is not immunoreactive under normal assay conditions. However, alpha 2-macroglobulin-bound trypsin can be detected after gel filtration of serum on Bio-Gel A-0.5 m and acid treatment of column fractions. The average serum level of immunoreactive cationic trypsin from 20 patients with acute pancreatic inflammation was 1,590 ng/ml. An average normal value of 26 ng/ml has been obtained previously. Serum samples from 14 patients with pancreatic inflammation were chromatographed under conditions that resolve trypsinogen, alpha 1-antitrypsin-bound trypsin, and alpha 2-macroglobulin-bound trypsin. In each case, the major portion of the immunoreactive material eluted at a position corresponding to free trypsinogen, while a minor fraction of the immunoreactive material appeared to be trypsin bound to alpha 1-antitrypsin. The zymogen nature of the major peak was confirmed in one case by activation with human enteropeptidase. In 11 of 14 patients, acid treatment of the alpha 2-macroglobulin peak yielded immunoreactive trypsin.
A specific radioimmunoassay for human pancreatic anionic trypsin has been developed. The trypsin employed as radioiodinated tracer in the assay was inactivated with tosyl-L-lysine chloromethyl ketone in order to prevent binding of the tracer to the serum inhibitors alpha1-antitrypsin and alpha2-macroglobulin. A normal serum level of immunoreactive anionic trypsin of 5.45 ng/ml was determined. The results of experiments in which serum was fractionated by Sephadex G-200 gel filtration suggest that essentially all of the immunoreactive material in normal human serum is trypsinogen. This finding implies that a small fraction of the zymogens synthesized in the pancreas are released directly into the circulation.
Human cationic trypsinogen is activated by human enteropeptidase much more readily than bovine trypsinogen, the ratios kcat/Km being 330 and 11 mM-1S-1, respectively. Conversely, porcine enteropeptidase activates bovine trypsinogen much more rapidly (kcat/Km = 630 mM-1S-1) than human cationic trypsinogen (kcat/Km = 2.4 mM-1S-1). The primary structure of the activation region of human cationic trypsinogen has been investigated in an attempt to elucidate the basis for these findings. The sequence of the first 12 residues at the NH2-terminus of human cationic trypsinogen has been shown to be Asp-Lys-Ile-Val-Gly-Gly-Tyr-Asn-Cys-Glu-Glu-Asn. Furthermore, the activation peptide derived from human cationic trypsinogen has been isolated and shown to be the dipeptide Asp-Lys. This result is in contrast to the Val-(Asp)4-Lys activation peptide from bovine trypsinogen and demonstrates that human cationic trypsinogen does not contain the (Asp)4 sequence present in many other mammalian trypsinogens. It is proposed that the high degree of specificity for activation of human cationic trypsinogen by human enteropeptidase is due to the preferential recognition of the novel activation peptide sequence in the human zymogen. Thus, these two functionally related proteins, cationic trypsinogen and enteropeptidase, may have evolved in a parallel manner in the human lineage.
Human pancreatic cationic trypsinogen has been purified to homogenity from an acetone powder of pancreatic tissue. After an initial ion exchange chromatography step on sulfopropyl (SP)-Sephadex at pH 2.6, cationic trypsinogen was separated from the majority of trypsin activity by passage through an affinity column of lima bean trypsin inhibitor-agarose at high ionic strength. The zymogen was then further purified by affinity chromatography on the same material at low ionic strength. Highly purified trypsinogen was resolved from containing chymotrypsinogen by ion exchange chromatography on SP-Sephadex at pH 6.0. The purified zymogen was shown to be homogeneous by polyacrylamide gel electrophoresis at pH 2.1 and at pH 4.3 as well as by discontinuous sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The autoactivation of human trypsinogen was investigated at pH 5.6 and at pH 8.0. The rate of autoactivation of the human zymogen is rapid at pH 5.6 and is maximal in approximately 1 mM Ca2+. These results are in marked contrast to those previously reported for autoactivation of bovine trypsinogen, which is extremely slow at pH 5.6 and which shows a dependence on at least 50 mM Ca2+ for maximum rate of activation (MacDonald, M. R., AND Kunitz, M. (1941) J. Gen. Physiol. 25, 53-73).