Search

Structural and functional integrity of organ function profoundly depends on a regular oxygen and glucose supply. Any disturbance of this supply becomes life threatening and may result in severe loss of organ function. Particular reductions in oxygen availability (hypoxia) caused by respiratory or blood circulation irregularities cannot be tolerated for longer periods due to an insufficient energy supply by anaerobic glycolysis. Complex cellular oxygen sensing systems have evolved to tightly regulate oxygen homeostasis. In response to variations in oxygen partial pressure (PO2), these systems induce adaptive and protective mechanisms to avoid or at least minimize tissue damage. These various responses might be based on a range of oxygen sensing signal cascades including an isoform of the neutrophil NADPH oxidase, different electron carrier units of the mitochondrial chain such as a specialized mitochondrial, low PO2affinity cytochrome coxidase (aa3) and a subfamily of 2-oxoglutarate dependent dioxygenases termed HIF (hypoxia inducible factor) prolyl-hydroxylase and HIF asparaginyl hydroxylase called factor-inhibiting HIF (FIH-1). Thus, specific oxygen sensing cascades involving reactive oxygen species as second messengers may by means of their different oxygen sensitivities, cell-specific and subcellular localization help to tailor various adaptive responses according to differences in tissue oxygen availability.

Hypoxia in tumors is generally associated with chemoresistance and radioresistance. However, the correlation between the heterodimeric hypoxia‐inducible factor‐1 (HIF‐1) and the multidrug resistance transporter P‐glycoprotein (P‐gp) has not been investigated. Herein, we demonstrate that with increasing size of DU‐145 prostate multicellular tumor spheroids the pericellular oxygen pressure and the generation of reactive oxygen species decreased, whereas the α‐subunit of HIF‐1 (HIF‐1α) and P‐gp were up‐regulated. Furthermore, P‐gp was up‐regulated under experimental physiological hypoxia and chemical hypoxia induced by either cobalt chloride or desferrioxamine. The pro‐oxidants H2O2and buthionine sulfoximine down‐regulated HIF‐1α and P‐gp, whereas up‐regulation was achieved with the radical scavengers dehydroascorbate, N‐acetylcysteine, and vitamin E. The correlation of HIF‐1α and P‐gp expression was validated by the use of hepatoma tumor spheroids that were either wild type (Hepa1) or mutant (Hepa1C4) for aryl hydrocarbon receptor nuclear translocator (ARNT), i.e., HIF‐1β. Chemical hypoxia robustly increased HIF‐1α as well as P‐gp expression in Hepa1 tumor spheroids, whereas no changes were observed in Hepa1C4 spheroids. Hence, our data demonstrate that expression of P‐gp in multicellular tumor spheroids is under the control of HIF‐1.

Two human neuroblastoma (NB) cell lines, SH-SY5Y and Kelly, were found to express the gene for erythropoietin (EPO) in an oxygen (O2)-dependent manner. However, NB cells had maximal production of EPO with lower partial pressure of O2 values than the well-characterized hepatoma cell line HepG2. This maximal EPO expression was preceded by accumulation of the O2-sensitive α subunit of the heterodimeric transcription-factor complex hypoxia-inducible factor 1 (HIF-1). Western blot analysis revealed that the amount of the β subunit of HIF-1, identical to aryl hydrocarbon receptor nuclear translocator 1 (ARNT1), and the homolog ARNT2 increased in nuclear extracts from SH-SY5Y cells exposed to anoxia. In neuronal cells, ARNT1 and ARNT2 can form a heterodimer with HIF-1α, generating a functional HIF-1 complex. Using the hypoxia response element of the human EPO enhancer, we conducted electrophoretic mobility shift assays that showed accumulation and binding of HIF-1 complexes containing both ARNT1 and ARNT2 in NB cells. In addition to the HIF-1 complex, hepatocyte nuclear factor 4α (HNF4α) was found to be indispensable for hypoxia-induced EPO gene expression in hepatoma cells. Western blot analysis and polymerase chain reaction assessment showed that NB cells express neither HNF4α nor the splicing variant HNF4α7 and thus express EPO in an HNF4α-independent manner. Together, SH-SY5Y and Kelly cells may provide a new in vitro model for studying the mechanism of tissue-specific, hypoxia-inducible EPO gene expression.

Two human neuroblastoma (NB) cell lines, SH-SY5Y and Kelly, were found to express the gene for erythropoietin (EPO) in an oxygen (O2)-dependent manner. However, NB cells had maximal production of EPO with lower partial pressure of O2values than the well-characterized hepatoma cell line HepG2. This maximal EPO expression was preceded by accumulation of the O2-sensitive α subunit of the heterodimeric transcription-factor complex hypoxia-inducible factor 1 (HIF-1). Western blot analysis revealed that the amount of the β subunit of HIF-1, identical to aryl hydrocarbon receptor nuclear translocator 1 (ARNT1), and the homolog ARNT2 increased in nuclear extracts from SH-SY5Y cells exposed to anoxia. In neuronal cells, ARNT1 and ARNT2 can form a heterodimer with HIF-1α, generating a functional HIF-1 complex. Using the hypoxia response element of the human EPO enhancer, we conducted electrophoretic mobility shift assays that showed accumulation and binding of HIF-1 complexes containing both ARNT1 and ARNT2 in NB cells. In addition to the HIF-1 complex, hepatocyte nuclear factor 4α (HNF4α) was found to be indispensable for hypoxia-induced EPO gene expression in hepatoma cells. Western blot analysis and polymerase chain reaction assessment showed that NB cells express neither HNF4α nor the splicing variant HNF4α7 and thus express EPO in an HNF4α-independent manner. Together, SH-SY5Y and Kelly cells may provide a new in vitro model for studying the mechanism of tissue-specific, hypoxia-inducible EPO gene expression.

Tumor vascularization is the rate‐limiting step for the progression of cancer. Differential steps of tumor‐induced angiogenesis were studied by a novel in vitroconfrontation culture of avascular multi‐cellular prostate tumor spheroids and embryoid bodies grown from pluripotent embryonic stem (ES) cells. Vascularization in embryoid bodies started on day 5 of cell culture and was paralleled by down‐regulation of hypoxia‐inducible factor 1α (HIF‐1α) and vascular endothelial growth factor (VEGF). In parallel, a dissipation of gradients in the pericellular oxygen pressure was observed as measured by O2‐sensitive microelectrodes. After 24–48 h of confrontation culture, cells positive for platelet endothelial cell adhesion molecule (PECAM‐1) became visible in the contact region between the embryoid body and the tumor spheroid and sprouted within the confrontation cultures during subsequent days. Tumor‐induced angiogenesis resulted in growth stimulation of tumor spheroids, disappearance of central necrosis and a reduction of the pericellular oxygen pressure. Furthermore, tumor vascularization resulted in elevated levels of HIF‐1α, VEGF, heat shock protein 27 (HSP27), and P‐glycoprotein. Tumor‐induced angiogenesis may augment the oxygen consumption in tumors resulting in an increased expression of hypoxia‐related, proangiogenic genes as well as of HSP27 and P‐glycoprotein, which are involved in a multidrug resistance phenotype.—Wartenberg, M., Dönmez, F., Ling, F. C., Acker, H., Hescheler, J., Sauer, H. Tumor‐induced angiogenesis studied in confrontation cultures of multicellular tumor spheroids and embryoid bodies grown from pluripotent embryonic stem cells. FASEB J.15, 995–1005 (2001)

Changes in ambient Po2need to be sensed to allow long-term adaptation of cellular functions via the regulation of gene activity. The generation of OH• from H2O2in an iron-dependent perinuclear Fenton reaction for triggering gene expression could be the key event in the O2signaling pathway.

H2O2 mimicked the action of periportal pO2 in the modulation by O2 of the glucagon-dependent activation of the phosphoenolpyruvate carboxykinase (PCK) gene and the insulin-dependent activation of the glucokinase (GK) gene. H2O2 can be converted in the presence of Fe2+ in a Fenton reaction into hydroxyl anions and hydroxyl radicals (•OH). The hydroxyl radicals are highly reactive and might interfere locally with transcription factors. It was the aim of the present study to investigate the role of and to localize such a Fenton reaction. Hepatocytes cultured for 24 h were treated under conditions mimicking periportal or perivenous pO2 with glucagon or insulin plus the iron chelator desferrioxamine (DSF) or the hydroxyl radical scavenger dimethylthiourea (DMTU) to inhibit the Fenton reaction. PCK mRNA was induced by glucagon maximally under conditions of periportal pO2 and half-maximally under venous pO2. GK mRNA was induced by insulin with reciprocal modulation by O2. DSF and DMTU reduced the induction of PCK mRNA to about half-maximal and increased the induction of GK mRNA to maximal under both O2 tensions. Hydroxyl radical formation was maximal under arterial pO2. Perivenous pO2, DSF and DMTU each decreased the formation of •OH to about 70% of control. The Fenton reaction could be localized in a perinuclear space by confocal laser microscopy and three-dimensional reconstruction techniques. In the same compartment, iron could be detected by electron-probe X-ray microanalysis. Thus a local Fenton reaction is involved in the O2 signalling, which modulated the glucagon- and insulin-dependent PCK gene and GK gene activation.

Little is known about the factors influencing the hemoglobin switch in vertebrates during development. Inasmuch as the mammalian conceptus is exposed to changing oxygen tensions in utero, we examined the effect of different oxygen concentrations on β‐globin switching. We used an in vitromodel of mouse embryogenesis based on the differentiation of blastocyst‐derived embryonic stem cells to embryoid bodies (EBs). Cultivation of EBs at increasing oxygen concentrations (starting at 1% O2) did not influence the temporal expression pattern of embryonic (βH1) globin compared to the normoxic controls (20% O2). In contrast, when compared to normoxically grown EBs, expression of fetal/adult (βmaj) globin in EBs cultured at varying oxygen concentrations was delayed by about 2 days and persisted throughout differentiation. Quantitation of hemoglobin in EBs using a 2,7‐diaminofluorene‐based colorimetric assay revealed the appearence of hemoglobin in two waves, an early and a late one. This observation was verified by spectrophotometric analysis of hemoglobin within single EBs. These two waves might reflect the switch of erythropoiesis from yolk sac to fetal liver. Reduced oxygenation is known to activate the hypoxia‐inducible factor‐1 (HIF‐1), which in turn specifically induces expression of a variety of genes among them erythropoietin (EPO). Although EBs increased EPO expression upon hypoxic exposure, the altered β‐globin appearance was not related to EPO levels as determined in EBs overexpressing EPO. Since mRNA from both mouse HIF‐1α isoforms was detected in all EBs tested at different differentiation stages, we propose that HIF‐1 modulates β‐globin expression during development.—Bichet, S., Wenger, R. H., Camenisch, G., Rolfs, A., Ehleben, W., Porwol, T., Acker, H., Fandrey, J., Bauer, C., Gassmann, M. Oxygen tension modulates β‐globin switching in embryoid bodies. FASEB J.13, 285–295 (1999)

In order to elucidate the components of the oxygen sensory complex in HepG2 cells which regulates the production of erythropoietin, we have microinjected recombinant variants of the human small GTP-binding protein hRac1 and measured their effects on the production of reactive oxygen species (ROS) by the dihydrorhodamine-123 technique. The dominant-negative mutant hRac1(T17N) inhibits the NADH-stimulated production of ROS in HepG2 cells, whereas the constitutively activated hRac1(G12V) leads to an increase in intracellular ROS concentration. Reverse transcriptase PCR analysis showed that the hRac1, but not the hRac2, gene is expressed in HepG2 cells. These results demonstrate that hRac1, and not hRac2, is involved in the regulation of ROS production in HepG2 cells and suggest that hRac1 specifically functions in the non-phagocytic NAD(P)H oxidase complex.

Abstract:  Small intensely fluorescent (SIF) cells are paraganglionic cells derived from sympathicoblasts which may serve as interneurons, endo-/paracrine cells or arterial chemoreceptors within sympathetic ganglia. Like paraganglionic cells of other locations, e.g., carotid body glomus cells, they are responsive to hypoxia. Recent studies on glomus cells and other hypoxia-sensing cells suggested the involvement of a b 558 -type cytochrome and intracellular generation of H2O2 in the process of oxygen sensing. In the present study, we demonstrate the occurrence of the small subunit of cytochrome b 558 , p22phox, in SIF cells of guinea-pig sympathetic ganglia by immunohistochemistry using two different antisera. H2O2 production was monitored in explanted intact superior cervical ganglia of 2-day-old rats by confocal laser scanning analysis of rhodamine 123 fluorescence generated due to oxidation of dihydrorhodamine 123 by H2O2. Using this technique, SIF cell clusters appeared as sites of highest H2O2 production within the ganglia. Thus, SIF cells exhibit two key features of an oxidase system generating reactive oxygen species. This may be involved in the proposed chain of events in oxygen sensing, but alternative cellular functions of this system have also to be considered.

Multicell-mediated drug resistance is a major impediment for the effectiveness of chemotherapeutic approaches and has been shown to be a feature of many solid tumors. We used confocal laser scanning microscopy to evaluate the depth distribution of the fluorescent cytostatic drug doxorubicin (Dox) in two size classes of multicellular cancer spheroids (MCS) (Ø150 ± 50 μm and 350 ± 50 μm). MCS (Ø150 ± 50 μm) solely consist of proliferating cells, whereas in MCS (Ø350 ± 50 μm) peripheral proliferating cell layers are followed in the depth of the tissue by drug resistant quiescent cell areas. A technique was developed which allows noninvasively to trace fluorescence distributions down to a depth of approximately 180 μm in living MCS. This was achieved by confocal radial recordings of the mean Dox fluorescence in 600 μm² regions of interest (ROI), equidistantly spaced (10 μm) from the center of MCS towards their periphery. The resulting fluorescence intensitiy profiles were subsequently corrected for absorbtion and light scattering in the depth of the tissue by a convenient algorithm. A 10 min incubation of MCS (Ø150 ± 50 μm) with Dox (10 μM) led to a peripheral accumulation, after 2 h Dox was homogeneously distributed within the whole MCS. In contrast, after Dox treatment of MCS (Ø350 ± 50 μm) for 2 h, the drug was accumulated within the peripheral proliferating cell rim of 78 ± 8 μm, whereas deeper, quiescent cell layers remained unstained. When MCS were incubated with verapamil, cyclosporin A, orthovanadate, and quinidine, which are known to reverse P-glycoprotein (Pgp)-mediated multidrug resistance (MDR), Dox accumulated also in deeper cell layers. Genistein and indometacin which reverse multidrug resistance mediated by the multidrug resistance-associated protein (MRP) were without effects. The optical probe technique proved to be well suited to study MDR in a living three dimensional tissue context. Cytometry 31:137–145, 1998. © 1998 Wiley-Liss, Inc.

Local in the carotid body after perfusion stop depends on the oxygen consumption of the tissue and the apparent O2 solubility coefficient α* of the blood. Oxygen consumption of the carotid body tissue can be determined by measuring the local with needle electrodes after Krebs-Henseleit perfusion. Assuming the same for the blood-perfused carotid body the actual decrease can be used to estimate the hemoglobin content of the tissue. The influence of hemoglobin is described by the factor α*/α. Control values measured in the Krebs-Henseleit-perfused carotid body yielded values of α*/α = 0.8 −3. In a series of 11 blood-perfused carotid body preparations with 105 perfusion stops the quotient α*/α changed with arterial values higher than 100 Torr α*/α was between 0.8 and 2. This value thus was in the same range as the control values. Below 100 Torr higher values up to 11 were observed. Since a blood-perfused tissue (15 g % Hb) would have an α*/α of about 128, we conclude that the carotid body is mainly perfused by plasma, and that with decreasing arterial more red cells flowed through the carotid body.

Summary After perfusion stop a decrease has been measured in the carotid body tissue by means of elastically suspended needle electrodes. With the formula it is possible to calculate the oxygen consumption from this decrease. Our experiments have shown that the oxygen consumption of the non-perfused carotid body of the cat depends linearly on local . The measured in situ (Krebs-Henseleit perfusion) is with 6 Torr/s at a tissue of 100 Torr (37° C) two times higher than the values measured in vitro (Krebs-Henseleit-perfusion). The in vivo corresponds to an oxygen consumption of 1.2 ml O2/100 g min at a of 100 Torr at 37° C. The dependence of the oxygen consumption of the carotid body might be caused by a special oxidase which may be involved in the chemoreceptive process.

Isolated rat hepatocytes maintained in primary culture on gas permeable membrane for 20 h form monolayers and establish at their cell borders a network of canaliculi (approximate diameter 3.5 μm). In the presence of the known choleretic bile acid dehydrocholate, dilation of canaliculi occurs. When nonfluorescent carboxyfluorescein diacetate ester is added to the culture medium, fluorescent carboxyfluorescein appears in the intracanalicular space. In the dilated state, fluid containing the fluorescent compound could be collected from the canaliculi by puncture with a micropipette. The intracanalicular space shows a negative electrical potential difference of 31 mV in reference to the bath solution and is 13.5 mV more positive with reference to recordings from the cytosol of cultured rat hepatocytes. Cultured rat hepatocytes grown on gas permeable membrane are energetically stable over 3 d. On Day 4, ATP levels increase markedly, whereas Na+−K+-ATPase activity declines. Ionic composition of hepatocytes, as measured by electronprobe element analysis on cryosection samples, does not change markedly during monolayer formation. With formation of bile canaliculi, the activity of alkaline phosphatase rapidly increases within 24 h and is stable for the next 3 d. Within that time the activity of γ-glutamyltranspeptidase, however, increases steadily, reaching a 1.6-fold higher activity than freshly isolated hepatocytes. Bile acids appear in the culture supernatant after 1 d. When unconjugated [14C]cholic acid is added to the cultures the supernatant contains also [14C]tauro- and [14C]glycocholic acid, indicating the preservation of conjugation capacity in these cultures. Total bile acid concentrations in the supernatant increase from 5 to 26 μMon Day 4. The cultures do not secrete α-fetoprotein. Monolayer cultures of hepatocytes in the presence of choleretic bile acids seem to be a suitable model system to collect and to analyze the composition of primary bile. In conjunction with the electrical parameters, it is possible to describe directly properties of bile secretion at the canalicular pole of the intact hepatocyte.

Summary: A new method was tested for studies of penetration of substances into tumorlike tissue. The penetration of the ions K+, Cl−, and Ca2+ through several layers of tumor cells was demonstrated by using double barrelled, ion sensitive microelectrodes with extra thin tip diameters. Spheroids consisting of human glioma, U-118 MG, and human thyroid cancer, HTh-7, cells were used as models of tumor tissue. A microelectrode was inserted into the center of a spheroid. Thereafter, the concentration of the test substance was increased in the surrounding medium. The change in concentration inside the spheroid was recorded and the penetration pattern evaluated. All three types of tested ions penetrated easily through the spheroids. The K+ ions penetrated most efficiently, and the Ca2+ ions showed the slowest penetration. The Ca2+ ions penetrated somewhat more slowly in the U-118 MG spheroids (which had rather small extracellular spaces) than in the HTh-7 spheroids (which had larger extracellular spaces). Ion sensitive electrodes, which are easily available, were used in this study only to demonstrate the principle. We hope that the method described can be used for penetration studies of various substances. For example, all substances that can be detected by enzyme microelectrodes could be studied. The main advantage of the method is that the complete penetration pattern can be studied as a function of time in individual spheroids. Previously described methods require histological procedures for each analyzed penetration time.

Cellular retention and processing of a radiolabelled monoclonal anti-prostate antibody were evaluated after binding to prostatic adenocarcinoma DU 145 cells. An endocytosis assay revealed that the rate of release of radioactivity from the cells had an initially rapid phase within the first hours after antibody incubation, which was presumably due to release of monovalently bound antibodies. This was followed by a slower phase, with the possible release of intact bivalently bound antibodies and excretion of degraded internalized antibodies. The relative amount of released radioactivity of high molecular mass was high, indicating that the major part of the antibodies were released without being internalized and degraded. However, when only the radiolabelled antibody that remained cell-associated after 2 h and longer was considered, a substantial part was found to be internalized and radioactive degradation products were excreted. About 30% of the initially cell-associated radioactivity still remained associated with the cells after 48 h, indicating a rather slow antibody processing, which is favourable if the antibody is to be used for targeted radiotherapy. The retention of cell-associated radioiodine was very similar irrespective of whether the antibodies were radiolabelled with the direct chloramine-T method or the indirect ATE (succinimidyl based reagent) method. Since the ATE method can be used to form stable antibody constructs with the therapeutically relevant alpha-emitting radionuclide, astatine-211, this was an interesting finding that will be further evaluated in the future.

Light‐absorption spectra and afferent chemoreceptor discharge were simultaneously recorded on superfused rat carotid bodies (CBs) under the influence of cytochrome a3‐CuB ligands (O2, CN–, CO) in order to identify the primary mitochondrial cytochrome c oxidase (CCO) oxygen sensor. Spectra could be described on the basis of weighted light‐absorption spectra of cytochrome b558 of the NAD(P)H oxidase and mitochondrial cytochromes b and c, CCO, cytochrome a3, and an unusual cytochrome a peaking at 592 nm. Discharge signals were deconvoluted into phasic and tonic activity for comparing different CB responses. The spectral weight of cytochrome a592 decreased significantly starting at high PO2 (100 mm Hg) and low sodium cyanide (CN–, 10 μM) accompanied by increasing phasic peak discharge. Combined CO‐hypoxia or CO‐CN–application inhibited photolysis of CO‐stimulated chemoreceptor discharge, revealing photometrically cytochrome a592 as central in oxygen sensing. Control spectra in tissue from sympathetic and nodose ganglia did not show any cytochrome a592 contribution. According to these results, cytochrome a592 is assumed as a unique component of CB CCO, revealing in contrast to other cytochromes an apparent low PO2 and high CN–affinity, probably due to a shortcut of electron flow within CCO between CuA and cytochrome a3‐CuB.

APCO2 electrode working on the principle of electrical conductivity is described. The calibration curve can be linearized according to the formula . This linearity has been tested in thePCO2 range of 0.93–9.33 kPa (7–70 Torr). For the experiments electrodes are used which have conductivity values of about 50 nS and drifts of maximally 5%/h at aPCO2 of 5.33 kPa (40 Torr). The response time (T90) is about 20 s. The temperature sensitivity is 2.4 nS/1 K between 298K–310K. The standard error of the measurements is σ=0.33 nS. With these electrodes tissuePCO2 can be measured on the surface of various organs.