Pretransplant islet culture: a comparison of four serum-free media using a murine model of islet transplantation.

INTRODUCTION: Human islet transplant protocols frequently incorporate a brief period of islet culture before transplantation. The optimal medium for pretransplant islet culture is unknown. METHODS: We compared four serum-free media formulated for human islets: Miami (MM1), Memphis (M-SFM), Edmonton (EDM), and hCell OCZEM-SF/AF (hCell). Islets isolated from a single human pancreas with purity >80% were cultured in 2500-islet-equivalent (IE) fractions using the media listed. After 7 days, each 2500-IE fraction was grafted under the kidney capsule of a streptozocin-diabetic rag1 mouse (n = 4 per group). Mice were evaluated with serum glucose monitoring, stimulated C-peptide release, and glucose tolerance tests. Islet fractions transplanted immediately after isolation (n = 4 mice) served as controls. In vitro islet function was assessed on days 0 and 3 and included insulin release (after static glucose stimulation), total cellular C-peptide content, cell count, and viability. RESULTS: Glucose control was improved in all cohorts of mice after transplant, but only islet grafts cultured in MM1 were statistically indistinguishable from fresh islets. MM1- and hCell-cultured islet grafts showed improved glucose tolerance compared with fresh islets; C-peptide release was similar among the four cohorts. In vitro, only islets cultured in MM1 had similar stimulation index to fresh islets, whereas only hCell- and MM1-cultured islets demonstrated recovery of C-peptide content and insulin release. CONCLUSIONS: Media choice before transplant can influence islet quality, even when culture periods are short. Miami MM1 and hCell media may provide better islet protection than alternative media.

On the possible role of reactive oxygen species in angiogenesis.

Human microvascular endothelial cells grown on a 3-D reconstituted extracellular matrix (Matrigel) spontaneously and rapidly form a capillary network of tubular structures, thus modeling part of the angiogenic cascade. Exposure of the cells at the time of plating onto Matrigel to a brief episode of hypoxia (40-60) min and subsequent reoxygenation, significantly accelerated (up to 3-fold) the rate of tubular morphogenesis, as determined by computer-aided morphometry. This effect was not dependent on activation of PKC or upregulation/release of angiogenic growth factors. Rather, hypoxia/reoxygenation (H/R), but not hypoxia alone, caused the formation of reactive oxygen species (ROS) and the activation of the nuclear transcription factor NF kappa B, both of which were inhibited by ROS-scavengers, such as pyrollidine dithiocarbamate. Tube formation was inhibited, also under normoxic conditions, by diverse ROS antagonists in a dose-dependent fashion. Our results indicate that angiogenesis is accompanied by and/or requires generation of ROS. We hypothesize that in the clinical setting of hypoxia/reoxygenation during ischemic pre-conditioning, enhanced activation of ROS-dependent intracellular signaling may accelerate the rate of neovascularization also in vivo, thus contributing to the alleviation of certain ischemic lesions.

Long-term survival of human central nervous system progenitor cells transplanted into a rat model of Parkinson's disease.

Progenitor cells were isolated from the developing human central nervous system (CNS), induced to divide using a combination of epidermal growth factor and fibroblast growth factor-2, and then transplanted into the striatum of adult rats with unilateral dopaminergic lesions. Large grafts were found at 2 weeks survival which contained many undifferentiated cells, some of which were migrating into the host striatum. However, by 20 weeks survival, only a thin strip of cells remained at the graft core while a large number of migrating astrocytes labeled with a human-specific antibody could be seen throughout the striatum. Fully differentiated graft-derived neurons, also labeled with a human-specific antibody, were seen close to the transplant site in some animals. A number of these neurons expressed tyrosine hydroxylase and were sufficient to partially ameliorate lesion-induced behavioral deficits in two animals. These results show that expanded populations of human CNS progenitor cells maintained in a proliferative state in culture can migrate and differentiate into both neurons and astrocytes following intracerebral grafting. As such these cells may have potential for development as an alternative source of tissue for neural transplantation in degenerative diseases.

Lung fibroblast alpha-smooth muscle actin expression and contractile phenotype in bleomycin-induced pulmonary fibrosis.

The emergence of the myofibroblast phenotype (characterized by alpha-smooth muscle actin expression) in wound healing and in tissues undergoing fibrosis is thought to be responsible for the increased contractility of the affected tissues. In bleomycin-induced pulmonary fibrosis, the myofibroblast is also responsible for the observed increase in collagen gene expression. To evaluate further these phenotypic changes in lung fibroblasts, contractile and other phenotypic properties of fibroblasts isolated from lungs of rats with bleomycin-induced fibrosis were compared with those of normal rats using in vitro models. Pulmonary fibrosis was induced in rats by endotracheal injection on day 0, and 7 and 14 days later the animals were sacrificed and lung fibroblasts isolated. Using immunofluorescence, < 10% of fibroblasts from control animals express alpha-smooth muscle actin when cultured as a monolayer. In contrast, 19% and 21% of cells from day 7 and day 14 bleomycin-treated animals, respectively, expressed this actin and with greater intensity than in control lung cells. This increase in actin expression was associated with enhanced contractility when evaluated using a three-dimensional cell culture model consisting of fibroblast-populated collagen gels. This enhanced contractility was abolished by treatment with antibody to transforming growth factor-beta (TGF-beta), whereas exogenous TGF-beta 1 and serum-stimulated contraction of control lung fibroblasts. TGF-beta 1 gene expression was greater in cells from bleomycin-treated animals than those from control lungs. These results show that cells with the myofibroblast phenotype are more abundant in fibrotic lung, and that these cells possess greater contractile capacity in vitro at least partly by virtue of their enhanced endogenous TGF-beta 1 gene expression.

Growth factor-induced epidermal invasion of the dermis in human skin organ culture: dermal invasion correlated with epithelial cell motility.

We have developed a model of human squamous epithelial cell invasion in human skin organ culture. Epidermal invasion of the dermis occurs in this model when the tissue is exposed to an exogenous source of epithelial cell growth factor. In the present study we sought to determine to what extent growth factor-induced invasion correlates with the ability of the growth factor to induce epithelial cell motility. Histological examination of tissue treated with epidermal growth factor (EGF), hepatocyte growth factor (HGF), insulin-like growth factor-1 (IGF-1) or keratinocyte growth factor (KGF) showed that only HGF and EGF were inducers of invasion while KGF- and IGF-1-treated tissues were histologically similar to untreated controls. In parallel studies, HGF and EGF were found to be potent stimulators of epidermal keratinocyte motility while IGF-1 was less effective and KGF was even less so. None of the growth factors stimulated dermal fibroblast motility while HGF and to a lesser extent IGF-1 (but not EGF or KGF) stimulated motility of dermal vascular endothelial cells. Thus, there is a strong correlation between growth factor capacity to induce epidermal keratinocytes to invade the underlying dermal tissue and to induce motility in the invasive population of cells.

Expression of adhesion molecules in cultured human pulmonary microvascular endothelial cells.

HPMECs were successfully isolated by differential trypsinization from peripheral lung lobes. The cells proliferated rapidly in EGM-MV with 10% FBS and were serially cultivated for more than 20 passages (1:4 split ratio) in vitro. Cells were characterized as endothelial based upon their cobblestone morphology, the presence of factor VIII-related antigen, incorporation of DiI-Ac-LDL, tubule-like structure formation in Matrigel, and positive staining for ACE. Adhesion molecules were tested at passage 3 and passage 12. Cells demonstrated intense staining for PECAM-1 both unstimulated and stimulated with TNF-alpha (20 ng/ml). The adhesion molecules ICAM-1, VCAM-1, ELAM-1, and P-selectin differed in expression on unstimulated cells. ICAM-1 was constitutively expressed on unstimulated cells and the expression was increased by TNF-alpha stimulation (20 hr). In contrast, VCAM-1, ELAM-1, and P-selectin were not detected on unstimulated cells but were detected after stimulation with TNF-alpha. The inducibility of adhesion molecules was different. VCAM-1 (10 hr) and ELAM-1 (4 hr) were expressed more strongly than P-selectin (minutes to 4 hr). The adhesion molecule profile found on passage 12 was the same as on passage 3. CD36 was not detected on both unstimulated and stimulated (4 and 8 hr) cells. The peak of adhesion of HL-60 cells to TNF-alpha activated HPMEC monolayers was around 8 hr. The results indicate that HPMEC can be continuously grown in vitro for many passages without losing their adhesion molecule expression. This expression of adhesion molecules confirms that HPMECs might be a good in vitro model in the understanding of various aspects of pulmonary microvascular endothelial cell function and may be useful as the basis for studies of adhesion molecule targeted therapies of pulmonary inflammatory diseases.

Fibroblast responses to cytokines are maintained during aging.

Impaired wound healing in older individuals may result from a global deficit of fibroblast activity or a decreased response of aged fibroblasts to stimulation by inflammatory cytokines. Twenty-eight fibroblast cell strains were developed from normal human skin aged 3 days to 84 years. Mitogenic response to cytokines, epidermal growth factor, tumor necrosis factor-alpha, platelet-derived growth factor or fetal bovine serum was determined using a 4-hour 3H-thymidine incorporation assay. Synthesis of collagen and noncollagen proteins was determined basally and in response to transforming growth factor-beta using a 6-hour 3H-proline incorporation assay. Neither the mitogenic response to cytokine stimulation (p > 0.3) nor the synthesis of collagen and noncollagen protein after transforming growth factor-beta stimulation (p > 0.4) varied with the age of the cell donor. Individual cell lines' response to cytokine stimulation varied widely, reflecting differences commonly seen in patients' wound-healing abilities. Cellular responses to wound-healing cytokines are preserved as people age. Abnormalities in wound healing in older patients may be the result of altered immune initiation of healing or the cumulative result of concomitant disease.

Hypertrophic scar fibroblasts accelerate collagen gel contraction.

Excessive contraction of hypertrophic scar and subsequent contracture formation are a formidable problem after thermal injury. A comparison between fibroblasts from hypertrophic scar and normal skin was made with the use of fibroblast-populated collagen lattices as a measure of cellular generated contractile forces. Hypertrophic scar and normal skin fibroblasts were mixed with soluble tendon collagen and Dulbecco's modified Eagle medium supplemented with 10% serum, and contraction was measured by serial area measurements. Parallel experiments in the presence of transforming growth factor-beta or anti-transforming growth factor-beta antibody examined the role of this cytokine on lattice contraction. Transforming growth factor-beta activity was measured in an additional set of 10 biopsy specimens. Hypertrophic scar fibroblasts contract lattices at a significantly faster rate than do normal skin fibroblasts. Exogenous transforming growth factor-beta increased lattice contraction by normal skin fibroblasts but had little effect on hypertrophic scar cell-populated lattices. The addition of anti-transforming growth factor-beta antibody decreased lattice contraction by both cell types. Transforming growth factor-beta activity was significantly increased in the hypertrophic scar biopsy specimens. Excessive scar contraction and post-burn scar contracture result from increased contraction forces generated by hypertrophic scar cells. This increased contractility appears to be mediated by increased endogenous presence of transforming growth factor-beta.

Phenotypic differences in cytokine responsiveness of hypertrophic scar versus normal dermal fibroblasts.

The alteration of normal dermal fibroblast function that leads to the development of hypertrophic scar after thermal injury is unknown. To determine functional differences that might explain this process, fibroblasts were cultured from biopsies of post-thermal injury mature hypertrophic scars and patient-matched normal skin. The mitogenic responses of scar cells to fetal bovine serum, epidermal growth factor (EGF), platelet-derived growth factor (PDGF), and tumor necrosis factor alpha (TNF alpha) were determined and compared to normal skin cells. Collagen synthetic rate was also compared in the presence and absence of transforming growth factor beta 1 (TGF beta 1). Whereas both scar and normal cells responded with increased thymidine uptake to serum and cytokines, the stimulation to EGF and serum was significantly lower in scar cells. In contrast, synthesis of collagen, but not of non-collagenous proteins, was increased in scar relative to normal cells, both basally and when stimulated with low doses of TGF beta 1. Additionally, the fraction of protein synthesized as collagen was significantly higher in scar fibroblasts. These results suggest that fibroblasts from hypertrophic scars demonstrate stable phenotypic differences in cytokine responsiveness in comparison to cells from unaffected skin. The increased rate of collagen synthesis and decreased responsiveness to mitogens are consistent with the increased extracellular matrix content and decreased cellularity of hypertrophic scars.

Regulation of rat pulmonary endothelial cell interleukin-6 production by bleomycin: effects of cellular fatty acid composition.

Previous studies have shown upregulation of lung cell interleukin-6 (IL-6) production in bleomycin-induced pulmonary fibrosis. To further elucidate the regulatory mechanisms governing this disease, the effects of bleomycin on the production of the pleiotropic cytokine, IL-6, were investigated in lung endothelial cells. Rat pulmonary artery endothelial cells were treated with bleomycin at doses previously shown to be effective in upregulating cytokine production in these cells, and the conditioned media was collected and assayed for IL-6 activity. The results show that these endothelial cells constitutively produced IL-6 and that bleomycin increased the production in a time- and dose-dependent manner. Feeding rats diets deficient in n-6 fatty acids is known to ameliorate bleomycin-induced lung fibrosis. In order to examine if fatty acids could modulate IL-6 production in vitro, cells were lipid depleted and then supplemented with 18:1n-9, 18:2n-6, or 18:3n-3 fatty acids, and the effects of bleomycin on IL-6 production reexamined. This regimen resulted in significant depletion of arachidonate in the 18:1n-9 and 18:3n-3 supplemented cells, which was associated with significantly reduced IL-6 production relative to the 18:2n-6-supplemented cells, both constitutively and when stimulated with bleomycin. Preincubation with indomethacin did not significantly inhibit the production of IL-6 by all three groups of cells, nor did supplementation with a stable prostacyclin analog increase IL-6 production. These results suggest that endothelial cell IL-6 production is not directly dependent on prostacyclin or other cyclooxygenase metabolites but may require or be upregulated by 18:2n-6 and/or metabolites derived from it.(ABSTRACT TRUNCATED AT 250 WORDS)

Transferrin: a potential source of iron for oxygen free radical-mediated endothelial cell injury.

The ability of transferrin to potentiate oxygen free radical-mediated endothelial cell injury was assessed. 51Cr-labeled endothelial cells derived from rat pulmonary arteries (RPAECs) were incubated with hydrogen peroxide (H2O2) in the presence and absence of holosaturated human transferrin, and the effect of transferrin on H2O2-mediated endothelial cell toxicity was determined. Addition of holosaturated transferrin potentiated H2O2-mediated RPAEC cytotoxicity at concentrations of H2O2 greater than 10 microM, suggesting that transferrin may provide a source of iron for free radical-mediated endothelial cell injury. Free radical-mediated injury is dependent on non-protein-bound iron. The ability of RPAECs to facilitate the release of iron from transferrin was assessed. We determined that RPAECs facilitate the release of transferrin-derived iron by reduction of transferrin-bound ferric iron (Fe3+) to ferrous iron (Fe2+). The reduction and release of transferrin-derived Fe2+ were inhibited by apotransferrin and chloroquine, indicating a dependence on receptor-specific binding of transferrin to the RPAEC cell surface, with subsequent endocytosis, acidification, and reduction of transferrin-bound Fe3+ to Fe2+. The release of transferrin-derived Fe2+ was potentiated by diethyldithiocarbamate, an inhibitor of intracellular superoxide dismutase (SOD). In contrast, exogenous SOD did not alter iron release, suggesting that intracellular superoxide anion (O2-) may play an important role in mediating the reduction and release of transferrin-derived iron. Results of this study suggest that transferrin may provide a source of iron for oxygen free radical-mediated endothelial cell injury and identify a novel mechanism by which endothelial cells may mediate the reduction and release of transferrin-derived iron.

Mechanism of neutrophil-induced xanthine dehydrogenase to xanthine oxidase conversion in endothelial cells: evidence of a role for elastase.

Activated neutrophils cause conversion of xanthine dehydrogenase to its oxidase form (xanthine oxidase) in endothelial cells, the mechanism of which may be related to the cytotoxic effect of activated neutrophils. The elastase inhibitors, elastatinal, alpha 1-antitrypsin, and MeO-Suc-(Ala)2-Pro-Val-CH2Cl, significantly inhibited xanthine dehydrogenase to oxidase conversion by phorbol myristate acetate-stimulated neutrophils without inhibition of neutrophil adherence to the endothelial cell monolayer. The role of elastase in this enzyme conversion process was confirmed by the ability of purified elastase to cause conversion of xanthine dehydrogenase to xanthine oxidase in intact endothelial cells (or cell extracts) without causing cytotoxicity. In contrast, cathepsin G failed to cause conversion. The kinetics of conversion induced by elastase was relatively rapid, being essentially completed by 30 min. Upon removal of elastase, the effect was slowly (greater than 12 h) reversible and could be inhibited by cycloheximide treatment. Exposure of endothelial cells to hypoxia failed to enhance the elastase-induced conversion. Treatment of endothelial cells with Ca2+ ionophores failed to cause conversion of xanthine dehydrogenase to oxidase, suggesting that intracellular Ca(2+)-activated proteases are not sufficient to induce this process. Neutrophil-induced xanthine dehydrogenase to oxidase conversion was inhibited by concomitant treatment with antibodies to CD11b. The results suggest that activated neutrophils induce conversion of xanthine dehydrogenase to oxidase by secretion of elastase in close proximity to the endothelial cells and that this intimate contact between the two cell types enables high local concentrations of elastase to be attained, which are sufficient to cause xanthine dehydrogenase to xanthine oxidase conversion.

Lack of an association between cellular phospholipid triene: tetraene ratio and proliferation of human skin fibroblasts in culture.

Two experiments were performed in an attempt to establish an association between cellular phospholipid triene:tetraene ratio and proliferation of human neonatal skin fibroblasts in culture. In Experiment 1, a low lipid culture medium was developed that caused an accumulation of (n-9) eicosatrienoic acid in the phospholipids of human fibroblasts. This culture medium, when supplemented with a mixture of mitogens, supported growth of human fibroblasts at a level equivalent to that found under conditions of maximal growth using serum supplementation (8% fetal bovine serum). The triene:tetraene ratio of fibroblast phospholipids under the two conditions was 1.88 vs. 0.03, suggesting that the growth of these cells was not adversely affected by a high (greater than 0.4) triene: tetraene ratio. In Experiment 2, cells were cultured in a low lipid, mitogen-supplemented medium with 16:1(n-7), 18:1(n-9), 18:2(n-6) or 20:4(n-6) added as the albumin complex. All the fatty acids permitted an equivalent maximal growth stimulation in the assay system, although having different effects on the phospholipid triene:tetraene ratio. The results suggest that there is a lack of an association between cellular phospholipid triene:tetraene ratio (range, 0.03 to 3.4) and proliferation of human fibroblasts in this culture system.

Effect of cellular growth state on indices of n - 6 polyunsaturated fatty acid status in human fibroblasts.

A 2 x 2 design was employed to examine the effect of cellular growth state and medium serum concentration on potential indices of n - 6 polyunsaturated fatty acid (PUFA) status in human skin fibroblasts. The cells were cultured either as nonmultiplying cell monolayers or as medium-density, log-phase multiplying cells. An interaction of cellular growth state and medium serum concentration influenced the accumulation of 20:3(n - 9), but not 22:3(n - 9), in the cellular phospholipids. The 20:3(n - 9)/20:4(n - 6) ratio was the most sensitive index of n - 6 PUFA status; however, the ratio was significantly affected by cellular growth state. The 22:3(n - 9)/22:4(n - 6) ratio appears to be an index of n - 6 PUFA status in fibroblasts that is not significantly affected by the growth state of cells.

The Na+K+ATPase activity in cultured human fibroblasts with an elevated phospholipid triene:tetraene ratio.

Human skin fibroblasts were cultured at low density for 11 days in MCDB 110, 0.4% fetal bovine serum, a mitogen mixture, and were supplemented with 18:2n-6 or 18:1n-9 as a fatty acid-albumin complex. The cells cultured with the 18:2n-6 supplement had a 20:3n-9/20:4n-6 ratio of 0.29 +/- 0.07; the 18:1n-9 supplemented cells had a ratio of 1.51 +/- 0.27. There was less than 4% difference in total growth of the cell population under the two culture conditions. The cells supplemented with 18:2n-9 had similar levels of protein/cell, K+/mg cell protein and functional Na+K+ATPase activity.

Accumulation of (n-9)-eicosatrienoic and docosatrienoic acids in human fibroblast phospholipids.

An essential fatty acid (EFA) deficiency-like profile of fatty acids has been observed in HF-1 human skin fibroblasts cultured at clonal densities in MCDB 110 and 0.4% fetal bovine serum (FBS). The profile was characterized by an accumulation of 16:1n-7, 18:1n-9, 20:3n-9 and 22:3n-9, a reduction of n-6 fatty acids and a reduction in total polyunsaturated fatty acids (PUFA). The fatty acid composition of sphingomyelin (SPH), phosphatidylcholine (PC), phosphatidylserine (PS), phosphatidylinositol (PI) and phosphatidylethanolamine (PE) was determined and, except for SPH, each displayed an EFA deficiency-like profile. The triene to tetraene ratio (20:3n-9/20:4n-6) ranged from 5.3 in PI to 0.9 in PE. In addition, the highest percentage of 20:3n-9 was present in the PI and the highest percentage of 22:3n-9, in PE. Other human fibroblasts (normal, transformed and at different population doubling number levels [PDL]) were grown under the same conditions and were found to display triene to tetraene ratios (20:3n-9/20:4n-6) in total cellular lipids ranging from 0.7 to 4.5. The accumulation of 20:3n-9 and 22:3n-9 is due primarily to the existence of a basal nutrient medium (MCDB 110) that allows for the rapid clonal growth of human fibroblasts at reduced serum levels (0.4%). This culture procedure can be exploited to further elucidate various aspects of lipid metabolism in human fibroblasts.