Experimental Biology and Medicine

In summary, we derived an experimental system that allows us to dissect the function of GATA-1 in red cell development at a genetic level. We have established the essential nature of GATA-1 during both primitive and definitive erythropoiesis. By ablating the expression of the endogenous GATA-1 gene, we are in a position to introduce a variety of constructs that harbor subtle modifications in flanking or protein-coding sequences. We can now study regulatory regions and functional domains of the protein in the context of a true erythroid environment, experiments that have not been possible heretofore. Although the assay involves the dramatic loss of red cell production, it should be possible to define important regulatory domains that can then be assayed using less stringent systems, such as cell-free extracts for in vitro transcription. The ideal situation would be analyses conducted in GATA-1^- erythroid cells. However, these cells have been impossible to generate given the requirement of GATA-1 for Epo receptor expression and red cell viability (C. Simon and S. Orkin, unpublished observations). It may be possible to produce such cells by first expressing the Epo receptor under the influence of a constitutive promoter and then targeting the GATA-1 gene. If GATA-1^- red cells were available, the analyses would involve the actual transcription of or chromatin structure surrounding the globin genes. Structure-function studies of the GATA-1 protein could be greatly simplified and a larger number of mutants studied. However, the ES cell system can be used as an alternative until targeted erythroleukemia cells become available. Other applications involve the introduction of other GATA-binding protein family members to determine whether they rescue the mutation. If they cannot, chimeric proteins can be tested to identify which amino acids distinguish the different family members.

We feel that these experiments are vital to understanding the function of GATA-1 during erythroid ontogeny. How does GATA-1 regulate red cell genes like globin or the Epo receptor? Once we identify the functional domains of the GATA-binding proteins, we hope to learn what proteins GATA-1 binds to in the basic transcription machinery or in chromatin. Is GATA-1 necessary for globin gene switching? GATA-1 may be modified differently during development so that the locus control region can interact with different globin promoters. We may find that one region of the protein is required for embryonic expression and another for adult globin gene expression.

The mutant ES cell system provides a novel approach to study directly the activity of a transcription factor in development in vivo and in vitro. It appears to recapitulate all that occurs in the developing mouse and we hope that this method can be successfully applied to other proteins important to red cell development.

The author wishes to thank her postdoctoral advisor, Dr. Stuart Orkin, for his support and guidance during the last 4 years and her and Dr. Orkin's collaborators on all experiments concerning mutagenesis of GATA-1, Dr. Frank Costantini and Larysa Pevny (Columbia University). Stuart Orkin, Merlin Crossley, and Gerd Blobel reviewed the manuscript prior to submission. M. C. S. is a fellow of the Medical Foundation, Inc.

Posttransfusion graft-versus-host disease is a lethal adverse effect of blood transfusions affecting the skin, liver, gastrointestinal tract, and, in addition, the lymphohematopoietic systems. It is a disease that has been described for nearly three decades with well over 400 cases known worldwide; however, the immunopathogenesis has not been fully described. By using murine models, combined with human case reports, the immune mechanism and risk factors are outlined. The models and case reports prove a histocompatibility disparity between donor and recipient. The dose of lymphocytes in the products, types of T lymphocyte subsets in the products, and degree of immune suppression in the host are all factors necessary in the immunopathogenesis of posttransfusion graft-versus-host disease. The mechanism is that of acute, lethal, suppressive immune dysregulation in the host (recipient).

The purpose of the experiments reported herein was to investigate the relative importance of new hormone synthesis to basal and prostaglandin E₂-stimulated rates of relaxin release. A relaxin-reverse hemolytic plaque assay was used to monitor relaxin release from individual large luteal cells (LLC) in which new protein synthesis was inhibited by cycloheximide or actinomycin D. These treatments significantly decreased the rate of relaxin release. In addition, cycloheximide reduced the total fraction of LLC possessing the ability to form plaques by about 10%, suggesting complete suppression of relaxin from this subset of cells. Exposure of inhibitor-treated LLC to prostaglandin E₂ (a relaxin stimulatory secretagogue) enhanced relaxin release, and restored suppressed LLC back into the secretory population. Taken overall, these results demonstrate that the majority of relaxin-releasing LLC exploit a mixture of newly synthesized and older, stored hormone to achieve basal secretion. A minority of relaxin-releasing LLC, however, appear to depend wholly on newly synthesized hormone for basal secretion. The differential activity (and interaction) of these pathways in individual LLC may provide a potential explanation for the markedly heterogenous manner of hormone release observed in this (and other) cell types, and for the action of relaxin secretagogues.

Normal physiological clearance of gut-derived endotoxin lipopolysaccharide [LPS] has been described previously; initially, there is uptake by Kupffer cells (KC), then release of modified LPS, followed by hepatocyte uptake. Previous work in our laboratories indicated that LPS is structurally modified with loss of carbohydrate prior to its release by KC. In this study, we functionally characterize KC modified LPS. KC-modified ¹²⁵I-LPS was prepared from primary rat KC. Escherichia coli 0127:B8 native ¹²⁵I-LPS or KC-modified ¹²⁵I-LPS (40 ng) was incubated for 1 hr with 1 × 10E6 primary hepatocytes. The binding of KC-modified LPS was 4.33-fold higher than native LPS (P = 0.0024). Binding analysis studies were conducted to determine the region of KC-modified LPS responsible for enhanced hepatocyte binding. KC-modified Salmonella minnesota LPS was competed with 100-fold excess native or mutant (Ra, Rc, Rd, or Re) strains of LPS or Lipid A with no decrease to hepatocyte binding. S. minnesota-native ¹²⁵I-LPS was compared with KC-modified ¹²⁵I-LPS in a study to assess induction of tumor necrosis factor (TNF)-γ by rat peritoneal macrophages. Native or KC-modified ¹²⁵I-LPS (100 ng) was presented to 1 × 10E7 peritoneal macrophages for 6 hr. TNF-α was measured in supernatants using the WEHI-164 cytotoxicity assay. Native LPS induced 5.7-fold higher TNF-α levels than KC-modified LPS (P < 0.0001). The above data suggest that structural alterations in KC-modified LPS are accompanied by functional alterations resulting in enhanced hepatocyte binding and decreased TNF-α release. The latter result implies that an early step in LPS detoxification occurs in the KC in which LPS is modified to prevent elicitation of biologically active cytokines.

Experimental results were obtained that demonstrate the importance of the relative concentrations of Zn ligands and their affinities for Zn to the rate of Zn transport across a biological membrane. The transport rate into cultured endothelial cells became saturated near a Zn concentration of 30 μM in the presence or absence of 14% serum. However, the maximum transport rate identified by the saturation plateau was nearly twice as fast from serum-free medium. The addition of histidine or picolinic acid to the medium with serum resulted in the coincidental shift of Zn from high molecular weight ligands to low molecular weight ligands and increased the overall transport rate. In serum-free medium, adding histidine or picolinic acid slowed the transport rate. This indicates that the rate of Zn transport is influenced by the ligand to which it is associated and that altering the relative proportions of specific ligands influences the Zn transport rate. The rate of Zn transport decreased in a stepwise fashion as the albumin to Zn ratio increased from 0 to 4:1, with further increases having little effect. This suggests that albumin has a special role as modulator of Zn transport into endothelial cells. These studies underscore the importance of controlling the relative concentrations of Zn and its ligands in Zn transport kinetic research and suggest that varying their concentrations in a physiological range may be a method of regulating the distribution of Zn into specific cells and tissues.

Platelet thromboxane (TX) production was examined in response to dietary copper. Groups of eight rats were fed copper-deficient, -marginal, and -adequate diets providing 0.5, 1.7, and 7.5 μg Cu/g, respectively, with controlled dietary Se and vitamin E. Platelets were purified and washed by centrifugation. Separate platelet samples from each rat were challenged with 10 μg/ml of collagen and 1 unit/ml (27.3 nM) of thrombin in Tyrode's buffer, 2.0 mM Ca²⁺. Platelet copper-dependent superoxide dismutase (CuSOD) activity showed a significant depression with reduced diet copper, but platelet glutathione peroxidase activity was unaffected. Challenged platelet TX production showed a significant 1.5- to 2.5-fold increase in response to both dietary copper deficiency and marginality, with highly significant negative correlations between challenged platelet TX production and platelet CuSOD activity and between TX production and copper status (liver copper). Endogenous (unchallenged) platelet lipid hydroperoxide concentrations, measured as free fatty acid hydroperoxides by a glutathione-disulfide-specific glutathione reductase recycling assay, showed a nonsignificant 47-67% increase in copper deficiency. Pooled data showed a significant 71% increase in platelet lipid hydroperoxides in copper deficiency. Platelet TX production showed a significant correlation with endogenous lipid hydroperoxides. The results suggest that dietary copper insufficiency increases platelet TX synthesis through changes in CuSOD in a dose-responsive (diet copper and platelet CuSOD activity) manner, and that platelet TX synthesis is influenced by lipid hydroperoxides (peroxide tone).

The survival of density-arrested quiescent murine BALB/c-3T3 cells in serum-free Dulbecco's medium requires the presence of cell growth factors or second messenger agonists. The protein synthesis inhibitor anisomycin blocks the survival-mediating action of the basic fibroblast growth factor (bFGF) and of 12-O-tetradecanoylphorbol 13-acetate (TPA), but has little or no effect on the protective action of platelet-derived growth factor or 8-bromoadenosine 3′:5′-cyclic monophosphate (Br-cAMP). The effects of anisomycin are concentration dependent in the range from 2.5 to 25 μM and show that the survival-enhancing abilities of bFGF and TPA critically require protein synthesis, whereas those of platelet-derived growth factor and Br-cAMP do not. The survival-mediating action of bFGF and TPA can also be blocked with the RNA synthesis inhibitors actinomycin D and 5,6-dichloro-1-β-d-ribofuranosylbenzimidazole (DRB), whereas the action of platelet-derived growth factor and Br-cAMP is largely resistant. Results on the time course of action of DRB, a selective inhibitor of the synthesis of mRNA precursor molecules, suggest that the RNA required for the survival-enhancing action of bFGF and TPA is present in cells at the time of serum withdrawal and addition of the survival factor and has a half-life greater than 3 h. The new evidence provides further support for the hypothesis that protection of serum-deprived, density-arrested BALB/c-3T3 fibroblasts against death can be achieved either via pathways that entail the synthesis of protein and RNA (e.g., via diacylglycerol-protein kinase C) or via pathways that do not involve de novo biosynthesis (e.g., via cAMP-protein kinase A).

A new low shear stress, low turbulence microcarrier culture system has been developed at NASA's Johnson Space Center that permits large-scale three-dimensional tissue culture. Tissue culture bioreactors called rotating-wall vessels were used in conjunction with multicellular cocultivation to develop a unique in vitro tissue-modeling system. Normal small intestine epithelium and mesenchymal cells were cocultivated on Cytodex-3 microcarriers and were initiated in two phases. Normal small intestine mesenchymal cells were inoculated into the rotating-wall vessel at 2 × 10⁵ cells/ml and allowed to attach and proliferate for 2 to 3 days. Normal small intestine epithelium was then added at an innoculum of 2 × 10⁵ cells/ml and cultivation continued for 30 to 40 days. These cocultures attained cell numbers of 4-6 × 10⁶ cells/ml and differentiated to form tissue-like masses of 0.4-0.5 cm with minimal necrosis. The masses displayed apical brush borders, differentiated epithelial cells, cellular polarity, extracellular matrix, and basal lamina. Verification of mesenchymal and epithelial cell expression was determined by immunocytochemistry and scanning electron microscopy. These data suggest that the rotating-wall vessel affords a new tissue culture model for investigation of growth, regulatory, and differentiation processes within normal tissues.

The effects of epidermal growth factor (EGF) and 17β-estradiol (E₂) on the expression of prolactin (PRL), the transcription factor Pit-1/GHF-1 (Pit-1), and on dopamine D₂ receptor mRNA in GH₃ cells were analyzed by immunocytochemistry, in situ hybridization, and Northern analysis in a defined serum-free cell culture medium. Radioimmunoassay was used to determine PRL secretion. Both EGF and E₂ stimulated PRL mRNA and PRL secretion, although the effects of EGF were more rapid than those of E₂. Pit-1 mRNA levels were not significantly changed in spite of the 2- to 8-fold increases in PRL mRNA levels and significant increases in PRL secretion. Analysis of dopamine D2 receptor mRNA by in situ hybridization and Northern hybridization detected expression of dopamine receptor in GH₃ cells, but the receptor mRNA levels were not modified by EGF or E₂ treatment in complete serum or in serum-free media. These observations suggest that EGF and E₂ modulate PRL mRNA levels and PRL secretion significantly in vitro, while the mRNA levels of Pit-1 do not change significantly in GH₃ cells cultured in a defined culture medium.

The nonobese diabetic (NOD) mouse is a model of Type I (insulin-dependent) diabetes. It develops autoimmune pancreatic β-cell lesions characterized by lymphocytic infiltration and β-cell destruction. The incidences of diabetes for male and female NOD mice in our colony were 24% and 73%, respectively. In this study, we investigated the effect of neonatal manipulation of the sex hormone profile on the incidence of diabetes in male and female NOD mice. One day after birth, male mice were castrated and female mice were either ovariectomized, given testosterone, or ovariectomized and given testosterone. The mice were maintained for 140 days and blood samples were collected biweekly starting at 42 days old. Diabetes was determined by three consecutive blood glucose levels > 200 mg/dl. Neonatal gonadectomy increased the incidence of diabetes in males but decreased it in females. Females treated with testosterone also had a decreased incidence of diabetes, whereas ovariectomy plus testosterone increased the incidence to 100%. Castration decreased the body weight in males and increased body weight in females. Testosterone treatment with or without ovariectomy also increased body weight. From these studies, we concluded that neonatal hormonal imprinting has a significant influence on the incidence of diabetes in the NOD mouse.

We measured specific activity of ornithine decarboxylase (ODC) and contents of putrescine and of the polyamines (spermidine and spermine) in isolated villus and crypt enterocytes from the jejunum of adolescent streptozotocin-diabetic and weight-matched control rats and diabetic and control rats treated with difluoromethyl ornithine (DFMO) 10 days after induction of diabetes. Consistent with previous observations by others of elevated ODC activity and contents of putrescine and of the polyamines in the intestinal epithelium undergoing hyperplasia, our studies showed elevated ODC activity and contents of putrescine and spermidine, but not of spermine, in the hyperplastic intestinal epithelium of diabetic rats. As in previous studies, suppression of ODC activity by DFMO prevented not only the jejunal epithelial hyperplasia in the diabetic rats, but also retarded jejunal epithelial growth in the control rats. DFMO administration lowered ODC activity by over 80% in both diabetic and control rat enterocytes and prevented the rise in enterocyte contents of putrescine and spermidine in the diabetic rat. The observation that, in both diabetic and control rats, treatment with DFMO lowered spermidine content in the crypt enterocytes but had no similar consistent effect on contents of putrescine or spermine suggested that spermidine could have been responsible for the intestinal epithelial hyperplasia in the diabetic rats and for the normal growth of the intestinal epithelium in control rats.

Ascorbic acid is thought to contribute to protection against the potentially damaging effects of radiation, oxygen toxicity, and abrasion in the eye. The anterior surface of the cornea is particularly subject to insult from each threat. We considered the possibility that the lacrimal gland of pigs has transport and/or metabolic capability to sequester the reduced or oxidized form of ascorbic acid and prepare it for secretion in the tears. Slices of fresh lacrimal gland were incubated in a physiologic buffer and exposed to ≤ 12 μM ¹⁴C-labeled ascorbic acid or dehydro-l-ascorbic acid over 40-min incubation periods. Dehydro-l-ascorbic acid was taken up to a greater extent than the reduced compound. ¹⁴C-Label recovered from the tissue was at least 75% in the form of ascorbic acid after incubation with either substrate. Uptake of both the reduced and oxidized substrates proceeded to a tissue to medium ratio in excess of unity; the former was prevented by the presence of nonlabeled ascorbate in the bathing medium, but was unaffected by the removal of Na⁺ from the bath. The uptake of both substrates was less after inhibition of cellular metabolic energy. The lacrimal gland in this diurnal animal species has transport and metabolic capabilities that could serve in secretion of ascorbic acid into tears. This might help to protect the corneal epithelium against various forms of damage.

Infusion of norepinephrine (NE) into humans and experimental animals induces diuresis by mechanisms that are not completely understood. Two series of experiments were performed to determine whether changes in plasma levels of vasopressin or changes in mean arterial pressure (MAP) are important factors in NE-induced diuresis in conscious dogs. First, plasma vasopressin (PAVP) levels were measured in normal and cardiac-denervated conscious dogs during a 30-min intravenous infusion of NE (0.5 μg · kg^-1 · min^-1). When NE was administered to normal dogs, urine flow increased from 0.3 ± 0.1 to 0.9 ± 0.4 ml/min. PAVP did not decrease, in spite of increases in mean arterial pressure (from 103 ± 4 to 123 ± 6 mm Hg) and left atrial pressure (from 5.2 ± 0.9 to 8.6 ± 1.4 mm Hg). The same dose of NE infused into cardiac-denervated dogs significantly increased urine flow (from 0.2 ± 0.1 to 0.8 ± 0.3 ml/min) and MAP (from 107 ± 5 to 147 ± 10 mm Hg), and decreased PAVP (from 1.8 ± 0.3 pg/ml to 1.2 ± 0.3 pg/ml). In the second series of experiments, NE was infused into cardiac-denervated dogs for 40 min. During the final 20 min of NE infusion, nitroprusside was infused to offset the pressor effect of NE by returning MAP to the initial control level. Urine flow increased during the first 20 min in which NE alone was given; however, when MAP was returned to the control level by nitroprusside infusion, urine flow also returned to the control level. PAVP increased from a control value of 3.6 ± 0.6 to 18.9 ± 3.8 pg/ml 15 min after the NP infusion had begun. We conclude that a decrease in PAVP is not required to elicit diuresis during NE infusion in normal conscious dogs and that the pressor effect of NE appears to play a major role in NE-induced diuresis.

Two studies were conducted to determine whether hepatic iron overload in rats fed fructose plays a role in the exacerbation of the signs associated with copper deficiency. When fed the adequate iron diet (50 μg Fe/g), copper-deficient rats fed either fructose or starch exhibited hepatic iron overload of similar magnitude. However, only livers of copper-deficient rats fed fructose exhibited the presence of high peaks associated with an iron compound detected by electron spin resonance. In addition, only copper-deficient rats fed fructose developed anemia, pancreatic atrophy, and heart hypertrophy with histopathologic changes, and they died prematurely of heart-related abnormalities. Lowering dietary iron from 50 μg/g to 30 μg/g was not sufficient to protect the animals against the pathologic consequences of copper deficiency. In contrast, the consumption of a fructose diet inadequate in both copper (0.6 μg/g) and iron (17 μg/g) resulted in the reduction of hepatic iron, which in turn caused the amelioration of the deficiency, compared with rats fed the adequate iron (50 μg/g) diet. None of these rats developed pancreatic atrophy, none exhibited myocardial lesions, and none died of the deficiency. Electron spin resonance spectra of their livers did not show the presence of free radicals. The data suggest that hepatic iron overload plays a role in the exacerbation of copper deficiency only when fructose diets are consumed.

Two chemically similar nonsteroidal anti-inflammatory drugs, orpanoxin and F-1067, had almost identical potencies and efficacies as anti-inflammatory (rat paw edema) and analgesic (mouse writhing) agents, but differed markedly in gastrotoxicity. Orpanoxin alone aggravated stomach lesions in rats subjected to pylorus ligation and failed to protect stomachs of rats challenged with indomethacin. The compounds did not differ in their in vitro enzyme inhibition effects, both failing to inhibit 5- and 15-lipoxygenase and both inhibiting prostaglandin synthetase. Extraction of prostagiandins from the gastric mucosa of pylorus-ligated rats revealed, however, that the safer F-1067 depleted prostaglandin 6-keto-F_1α less and increased prostaglandin E₂ much more than did orpanoxin. A possible causality is suggested.

The polyunsaturated ω-6 fatty acid, arachidonic acid ([AA] 20:4n-6), is both the key of the immunoregulatory substances, prostaglandins, and leukotrienes, and an essential component of immune cell membrane phospholipids, providing stability and flexibility to ensure cellular function. To explore possible effects of the physiological burden of viral replication in chronic viral infections on AA availability, plasma total esterified fatty acid (FA) proportions were measured in the feline leukemia (FeLV) model. Plasma FA profiles of 12 specific-pathogen-free cats with chronic infections with Rickard strain feline leukemia virus (FeLV-R) were compared with 12 age- and sex-matched uninfected specific-pathogen-free cats at 4 months after infection. A significant decrease from normal of average AA proportion was found in FeLV-R-infected cat plasma, while other major FA (palmitic, stearic, and oleic and ω-3 FA normally remained present until near death. Since plasma FA content rapidly affects circulating immune cell membrane composition and since FeLV infection also targets immune cells, we compared FA profiles of feline T4-thymic lymphoma 3201 cell membranes that were infected with virulent FeLV-R or less virulent FeLV-A, at 20 days after viral inoculation with sham-inoculated uninfected 3201 cells. Significantly altered FA proportions and ratios of saturated to unsaturated FA found in infected cell membranes were similar to plasma FA changes and paralleled the virulence of the FeLV inoculum. Altered postinfection FA proportions may impart serious functional defects to the immune cells of chronic FeLV-infected cats, contributing to the inability of their immune systems to eliminate FeLV by depleted plasma AA stores and modified cell membrane composition. Decreased AA availability may be an important factor in the cachexia and fatal outcome of FeLV infection.

The effects of single and repeated short-term (4 hr) obstruction of pancreaticobiliary duct (PBDO), with or without exocrine stimulation (intraductal hypertension) by cerulein infusion (0.2 μg/kg.hr), on the exocrine pancreas were evaluated in the rat. Single blockage of pancreaticobiliary duct for 4 hr caused a significant rise in serum amylase levels, pancreatic water content, and redistribution of lysosomal enzyme, cathepsin B from the lysosomal fraction to the zymogen fraction, which was considered to mean the colocalization of lysosomal enzymes with pancreatic digestive enzymes in the same subcellular compartment in acinar cells. In addition, the accelerated lysosomal and mitochondrial fragility was observed in the single pancreaticobiliary-duct-obstructed animals. Moreover, the repeated PBDO for 4 hr (2 hr in each obstruction and 1 hr of free flowing of pancreaticobiliary juice between two obstructions) caused more marked changes in almost the all parameters, and the repeated PBDO with intraductal hypertension caused an activation of trypsinogen in the pancreas, making more marked changes in almost the all parameters than the repeated PBDO only group. These results indicate that the present model of repeated PBDO with exocrine stimulation seems to be a pertinent model for gallstone pancreatitis in humans, and that redistribution of lysosomal enzymes and subcellular organellar fragility seem to play an important role in the pathogenesis of pancreatic injuries induced by PBDO, particularly by repeated PBDO with exocrine stimulation, probably via activation of trypsinogen to trypsin by lysosomal enzyme, cathepsin B.