Targeting gene expression to hypoxic tumor cells.

Solid tumors with areas of low oxygen tension (hypoxia) have a poor prognosis, as cells in this environment often survive radiation and chemotherapy. In this report we describe how this hypoxic environment can be used to activate heterologous gene expression driven by a hypoxia-responsive element (HRE), which interacts with the transcriptional complex hypoxia-inducible factor-1 (HIF-1). Our results demonstrate that the HIF-1/HRE system of gene regulation is active in hypoxic tumor cells and show the potential of exploiting tumor-specific conditions for the targeted expression of diagnostic or therapeutic genes in cancer therapy.

Oxygen-dependent modulation of erythropoietin mRNA levels in isolated rat kidneys studied by RNase protection.

Using oligonucleotide primers complementary to conserved regions in the mouse erythropoietin (Epo) gene, a portion of the rat Epo gene was amplified by the polymerase chain reaction to produce a probe suitable for assay of rat Epo mRNA by RNAse protection. The assay, which has sufficient sensitivity to measure to Epo mRNA in unstimulated rat kidneys, was used to demonstrate high amplitude in vitro modulation of Epo mRNA levels in response to changes in perfusate flow rate and oxygen tension in isolated kidneys, thus providing clear evidence that all the necessary events linking changes in oxygen delivery to the modulation of Epo mRNA levels can occur intrarenally.

Targeting of HIF-alpha to the von Hippel-Lindau ubiquitylation complex by O2-regulated prolyl hydroxylation.

Hypoxia-inducible factor (HIF) is a transcriptional complex that plays a central role in the regulation of gene expression by oxygen. In oxygenated and iron replete cells, HIF-alpha subunits are rapidly destroyed by a mechanism that involves ubiquitylation by the von Hippel-Lindau tumor suppressor (pVHL) E3 ligase complex. This process is suppressed by hypoxia and iron chelation, allowing transcriptional activation. Here we show that the interaction between human pVHL and a specific domain of the HIF-1alpha subunit is regulated through hydroxylation of a proline residue (HIF-1alpha P564) by an enzyme we have termed HIF-alpha prolyl-hydroxylase (HIF-PH). An absolute requirement for dioxygen as a cosubstrate and iron as cofactor suggests that HIF-PH functions directly as a cellular oxygen sensor.

Activation of the HIF pathway in cancer.

The maintenance of oxygen homeostasis is required both in physiological development and tumour growth. Hypoxia inducible factor (HIF) plays a central role in both processes. Reliable methods for visualising HIF alpha subunits have established that HIF activation occurs in the majority of common cancers. This occurs both by genetic mechanisms and through microenvironmental hypoxia. Activation of the HIF pathway has important effects on patterns of gene expression in tumours.

Mutation of the von Hippel-Lindau gene alters human cardiopulmonary physiology.

Intracellular responses to hypoxia are coordinated by the von Hippel-Lindau--hypoxia-inducible factor (VHL-HIF) transcriptional system. This study investigated the potential role of the VHL-HIF pathway in human systems-level physiology. Patients diagnosed with Chuvash polycythaemia, a rare disorder in which VHL signalling is specifically impaired, were studied during acute hypoxia and hypercapnia. Subjects breathed through a mouthpiece and ventilation was measured while pulmonary vascular tone was assessed echocardiographically. The patients were found to have elevated basal ventilation and pulmonary vascular tone, and ventilatory, pulmonary vasoconstrictive and heart rate responses to acute hypoxia were greatly increased, as were heart rate responses to hypercapnia. The patients also had abnormal pulmonary function on spirometry. This study's findings demonstrate that the VHL-HIF signalling pathway, which is so central to intracellular oxygen sensing, also regulates the organ systems upon which cellular oxygen delivery ultimately depends.

Organ distribution of the three rat endothelin messenger RNAs and the effects of ischemia on renal gene expression.

To determine the organ distribution of production of the three endothelin (ET) isopeptides, we have developed three ribonuclease protection assays specific for the messenger RNAs (mRNAs) of rat ETs 1, 2, and 3.12 organs from adult Sprague-Dawley rats were examined: heart, lung, liver, spleen, kidney, stomach, small intestine, large intestine, testis, muscle, salivary gland, and brain. The mRNA for ET1 was five times more abundant in the lung than in any other organ studied, moderate expression was seen in the large intestine, and lower levels of mRNA were detected in each of the other organs examined. ET2 was expressed at high level in both large and small intestine and at low level in stomach, muscle, and heart, but ET2 mRNA could not be detected elsewhere. ET3 mRNA was found in all organs, particularly in small intestine, lung, kidney, and large intestine. Because of reports suggesting that ETs might be involved in the hypoperfusion and hypofiltration observed in postischemic kidneys, we have also studied levels of mRNA in kidneys that had previously been subjected to 25 or 45 min of clamping of the renal pedicle. At 6 h after 45 min of ischemia, ET1 mRNA increased to a peak of 421 +/- 69% (mean +/- SEM, n = 3) of that in a standard renal RNA preparation. By contrast, ET3 mRNA decreased in the postischemic organ, falling to a value of 19 +/- 2% of standard at the same time point. The effects of ischemia on ET1 and ET3 mRNAs were long-lasting, with elevation of ET1 and depression of ET3 persisting for days. ET2 mRNA remained undetectable throughout. These findings (a) support a role for ET1 in postischemic renal vascular phenomena and (b) demonstrate a situation in which the expression of ET isoforms is clearly subject to differential regulation.

Age-dependent expression of the erythropoietin gene in rat liver and kidneys.

Using RNAse protection, we have made quantitative measurements of erythropoietin (EPO) mRNA in liver and kidneys of developing rats (days 1-54), to determine the relative contribution of both organs to the total EPO mRNA, to monitor changes which occur with development, and to compare the hypoxia-induced accumulation of EPO mRNA with the changes in serum EPO concentrations. To determine whether developmental and organ-specific responsiveness is related to the type of hypoxic stimulus, normobaric hypoxia was compared with exposure to carbon monoxide (functional anemia). Under both stimuli EPO mRNA concentration in liver was maximal on day 7 and declined during development. In contrast, EPO mRNA concentration in kidney increased during development from day 1 when it was 30-65% the hepatic concentration to day 54 when it was 12-fold higher than in liver. When organ weight was considered the liver was found to contain the majority of EPO mRNA in the first three to four weeks of life, and although, in stimulated animals, the hepatic proportion declined from 85-91% on day 1, it remained approximately 33% at day 54 and was similar for the two types of stimuli. When normalized for body weight the sum of renal and hepatic EPO mRNA in animals of a particular age was related linearly to serum hormone concentrations. However, the slope of this regression increased progressively with development, suggesting age-dependent alterations in translational efficiency or EPO metabolism.

Increased plasma viscosity as a reason for inappropriate erythropoietin formation.

The aim of this study was to examine whether altered plasma viscosity could contribute to the inappropriately low production rate of erythropoietin (EPO) observed in patients suffering from hypergammaglobulinemias associated with multiple myeloma or Waldenström's disease. We found that the EPO formation in response to anemia in these patients was inversely related to plasma viscosity. A similar inverse relationship between plasma viscosity and EPO production was seen in rats in which EPO formation had been stimulated by exchange transfusion and the plasma viscosity of which was thereby altered by using exchange solutions of different composition to alter plasma viscosity and thus whole blood viscosity independently from hematocrit. Raising the gammaglobulin concentration to approximately 40 mg/ml plasma in the rats almost totally blunted the rise in serum EPO levels despite a fall of the hematocrit to 20%. Determination of renal EPO mRNA levels by RNase protection revealed that the reductions in serum EPO levels at higher plasma viscosities were paralleled by reductions in renal EPO mRNA levels. Taken together, our findings suggest that plasma viscosity may be a significant inhibitory modulator of anemia-induced EPO formation. The increased plasma viscosity in patients with hypergammaglobulinemias may therefore contribute to the inappropriate EPO production, which is a major reason for the anemia developing in these patients.

HIF-1-dependent regulation of hypoxic induction of the cell death factors BNIP3 and NIX in human tumors.

Solid tumors contain regions of hypoxia, a physiological stress that can activate cell death pathways and, thus, result in the selection of cells resistant to death signals and anticancer therapy. Bcl2/adenovirus EIB 19kD-interacting protein 3 (BNIP3) is a cell death factor that is a member of the Bcl-2 proapoptotic family recently shown to induce necrosis rather than apoptosis. Using cDNA arrays and serial analysis of gene expression, we found that hypoxia induces up-regulation of BNIP3 and its homologue, Nip3-like protein X. Analysis of human carcinoma cell lines showed that they are hypoxically regulated in many tumor types, as well as in endothelial cells and macrophages. Regulation was hypoxia inducible factor-1-dependent, and hypoxia inducible factor-1 expression was suppressed by von Hippel-Lindau protein in normoxic cells. Northern blotting and in situ hybridization analysis has revealed that these factors are highly expressed in human tumors compared with normal tissue and that BNIP3 is up-regulated in perinecrotic regions of the tumor. This study shows that genes regulating cell death can be hypoxically induced and are overexpressed in clinical tumors.

Hypoxia-inducible expression of tumor-associated carbonic anhydrases.

The transcriptional complex hypoxia-inducible factor-1 (HIF-1) has emerged as an important mediator of gene expression patterns in tumors, although the range of responding genes is still incompletely defined. Here we show that the tumor-associated carbonic anhydrases (CAs) are tightly regulated by this system. Both CA9 and CA12 were strongly induced by hypoxia in a range of tumor cell lines. In renal carcinoma cells that are defective for the von Hippel-Lindau (VHL) tumor suppressor, up-regulation of these CAs is associated with loss of regulation by hypoxia, consistent with the critical function of pVHL in the regulation of HIF-1. Further studies of CA9 defined a HIF-1-dependent hypoxia response element in the minimal promoter and demonstrated that tight regulation by the HIF/pVHL system was reflected in the pattern of CA IX expression within tumors. Generalized up-regulation of CA IX in VHL-associated renal cell carcinoma contrasted with focal perinecrotic expression in a variety of non-VHL-associated tumors. In comparison with vascular endothelial growth factor mRNA, expression of CA IX demonstrated a similar, although more tightly circumscribed, pattern of expression around regions of necrosis and showed substantial although incomplete overlap with activation of the hypoxia marker pimonidazole. These studies define a new class of HIF-1-responsive gene, the activation of which has implications for the understanding of hypoxic tumor metabolism and which may provide endogenous markers for tumor hypoxia.

Carbonic anhydrase (CA IX) expression, a potential new intrinsic marker of hypoxia: correlations with tumor oxygen measurements and prognosis in locally advanced carcinoma of the cervix.

There is increasing evidence that hypoxia-regulated gene expression influences tumor aggressiveness, contributing to the poorer outcome of patients with hypoxic tumors. The role of the transcriptional complex hypoxia-inducible factor-1 as an important mediator of hypoxia-regulated gene expression is one of the best documented pathways. Recently, it has emerged that certain tumor-associated carbonic anhydrases (CAs) can be added to the list of known hypoxia-inducible factor-responsive genes. Here we show that the immunohistochemical expression of the tumor-associated CA IX is correlated with the level of hypoxia in human cervical tumors. We performed a prospective study in 68 patients where needle electrodes were used to make direct measurements of tumor oxygenation levels. CA IX expression was evaluated immunohistochemically in pretreatment tumor biopsies. There was a significant positive correlation between the level of tumor hypoxia (HP5) and the extent of CA IX expression. A retrospective study of 130 squamous cell cervical carcinomas demonstrated that a semiquantitative immunohistochemical analysis of CA IX expression in tumor biopsies is a significant and independent prognostic indicator of overall survival and metastasis-free survival after radiation therapy. These studies provide clinical evidence that CA IX expression is up-regulated in hypoxic human cervical tumors and is associated with a poor prognosis. CA IX may act as an intrinsic marker of tumor hypoxia and poor outcome after radiation therapy. The level of CA IX expression may be used to aid in the selection of patients who would benefit most from hypoxia-modification therapies or bio-reductive drugs.

Identification of novel hypoxia dependent and independent target genes of the von Hippel-Lindau (VHL) tumour suppressor by mRNA differential expression profiling.

The von Hippel-Lindau tumour suppressor gene (VHL) targets hypoxia inducible factor (HIF)-alpha subunits for ubiquitin dependent proteolysis. To better understand the role of this and other putative pathways of gene regulation in VHL function we subjected mRNA from VHL defective renal carcinoma cells and transfectants re-expressing a wild type VHL allele to differential expression profiling, and analysed VHL target genes for oxygen regulated expression. Among a group of newly identified VHL target genes the majority but not all were regulated by oxygen, indicating that whilst dysregulation of the HIF system makes a dominant contribution to alterations in transcription, VHL has other influences on patterns of gene expression. Genes newly defined as targets of the VHL/hypoxia pathway (conditionally downregulated by VHL in normoxic cells) include aminopeptidase A, collagen type V, alpha 1, cyclin G2, DEC1/Stra13, endothelin 1, low density lipoprotein receptor-related protein 1, MIC2/CD99, and transglutaminase 2. These genes have a variety of functions relevant to tumour biology. However, not all are connected with the promotion of tumour growth, some being pro-apoptotic or growth inhibitory. We postulate that co-ordinate regulation as part of the HIF pathway may explain this paradox, and that evolution of anti-apoptotic pathways may be required for tumour growth under VHL-dysregulation. Our results indicate that it will be necessary to consider the effects of abnormal activity in integral regulatory pathways, as well as the effects of individual genes to understand the role of abnormal patterns of gene expression in cancer.

The pVHL-associated SCF ubiquitin ligase complex: molecular genetic analysis of elongin B and C, Rbx1 and HIF-1alpha in renal cell carcinoma.

The VHL gene product (pVHL) forms a multimeric complex with the elongin B and C, Cul2 and Rbx1 proteins (VCBCR complex), which is homologous to the SCF family of ubiquitin ligase complexes. The VCBCR complex binds HIF-1alpha and HIF-2alpha, transcription factors critically involved in cellular responses to hypoxia, and targets them for ubiquitin-mediated proteolysis. Germline mutations in the VHL gene cause susceptibility to haemangioblastomas, renal cell carcinoma (RCC), phaeochromocytoma and other tumours. In addition somatic inactivation of the VHL gene occurs in most sporadic clear cell RCC (CC-RCC). However, the absence of somatic VHL inactivation in 30-40% of CC-RCC implies the involvement of other gatekeeper genes in CC-RCC development. We reasoned that in CC-RCC without VHL inactivation, other pVHL-interacting proteins might be defective. To assess the role of elongin B/C, Rbx1 and HIF-1alpha in RCC tumorigenesis we (a) mapped the genes to chromosomes 8q(cen) (elongin C), 16p13.3 (elongin B) and 22q11.2 (Rbx1) by FISH, monochromosomal somatic cell hybrid panel screening and in silico GenBank homology searching; (b) determined the genomic organisation of elongin C (by direct sequencing of PAC clones), Rbx1 and elongin B (by GenBank homology searching); and (c) performed mutation analysis of exons comprising the coding regions of elongins B, C and Rbx1 and the oxygen-dependent degradation domain of HIF-1alpha by SSCP screening and direct sequencing in 35 sporadic clear cell RCC samples without VHL gene inactivation and in 13 individuals with familial non-VHL clear cell RCC. No coding region sequence variations were detected for the elongin B, elongin C or Rbx1 genes. Two amino acid substitutions (Pro582Ser and Ala588Thr) were identified in the oxygen-dependent degradation/pVHL binding domain of HIF-1alpha, however neither substitution was observed exclusively in tumour samples. Association analysis in panels of CC-RCC and non-neoplastic samples using the RFLPs generated by each variant did not reveal allelic frequency differences between RCC patients and controls (P>0.32 by chi-squared analysis). Nevertheless, the significance of these variations and their potential for modulation of HIF-1alpha function merits further investigation in both other tumour types and in non-neoplastic disease. Taken together with our previous Cul2 mutation analysis these data suggest that development of sporadic and familial RCC is not commonly contributed to by genetic events altering the destruction domain of HIF-1alpha, or components of the HIF-alpha destruction complex other than VHL itself. Although (a) activation of HIF could occur through mutation of another region of HIF-a, and (b) epigenetic silencing of elongin B/C, Cul2 or Rbx1 cannot be excluded, these findings suggest that pVHL may represent the sole mutational target through which the VCBR complex is disrupted in CC-RCC. HIF response is activated in CC-RCC tumorigenesis.

Characterisation of functional domains within the mouse erythropoietin 3' enhancer conveying oxygen-regulated responses in different cell lines.

We have analysed sequences within the mouse erythropoietin enhancer which are required for oxygen regulated operation in the erythropoietin producing cell line, HepG2, and in two non-erythropoietin producing cell lines; the lung fibroblastoid cell line a23, and mouse erythroleukaemia (MEL) cells. At least three critical sites were demonstrated within a 96 nucleotide sequence. Oxygen regulated operation was dependent on sites within the first 26 nucleotides. Sequences lying 3' to this region modulated enhancer function but did not themselves convey oxygen regulated operation. In HepG2 cells these 3' sequences co-operated to permit operation of the inducible element at a distance from a promoter, but in MEL cells 3' sequences repressed activity of the inducible element. Though operation of this 3' sequence differed according to the cell type, oxygen regulated operation was dependent on the same two critical sites in the 5' region in both erythropoietin producing and non-erythropoietin producing cells. These findings support the existence of a widespread oxygen sensing system in mammalian cells which is similar to that operating in specific cells to regulate erythropoietin production, and they indicate that the system activates factors with similar DNA sequence specificity in different cells.

Betti reaction enables efficient synthesis of 8-hydroxyquinoline inhibitors of 2-oxoglutarate oxygenases.

There is interest in developing potent, selective, and cell-permeable inhibitors of human ferrous iron and 2-oxoglutarate (2OG) oxygenases for use in functional and target validation studies. The 3-component Betti reaction enables efficient one-step C-7 functionalisation of modified 8-hydroxyquinolines (8HQs) to produce cell-active inhibitors of KDM4 histone demethylases and other 2OG oxygenases; the work exemplifies how a template-based metallo-enzyme inhibitor approach can be used to give biologically active compounds.

Drosophila melanogaster SL2 cells contain a hypoxically inducible DNA binding complex which recognises mammalian HIF-binding sites.

Nuclear extracts from Drosophila SL2 cells were found to contain a hypoxically inducible complex capable of binding to hypoxia response elements from mammalian genes. This complex (HIF-D) resembled mammalian hypoxia inducible factor (HIF-1) in DNA sequence specificity, abrogation of induction by cycloheximide, induction by desferrioxamine and redox sensitivity of DNA binding. However, HIF-D was not induced by cobalt and was less sensitive to phosphatase than HIF-1. Endogenous phosphoglycerate kinase mRNA in SL2 cells showed similar inducible characteristics to HIF-D. These findings are evidence that the mammalian HIF-1 dependent system of oxygen regulated gene expression has a functional homologue in Drosophila.

The prevalence of hyperlipidemia in renal transplant recipients. Associations with immunosuppressive and antihypertensive therapy.

To determine the extent of persisting hyperlipidemia in renal transplant recipients receiving modern maintenance immunosuppressive and antihypertensive therapy we compared plasma levels of total and high-density lipoprotein and triglyceride in 275 renal transplant recipients with stable graft function with age- and sex-matched groups from the local general population (n = 4055). Total cholesterol and triglyceride were higher in transplanted patients in all age groups, but the difference was much more striking in women. Plasma levels of HDL cholesterol were similar or slightly lower in transplanted patients. Association with parameters of graft function, immunosuppressive therapy, and antihypertensive therapy were studied within the transplanted population using multiple regression. Total cholesterol was significantly and independently associated with age, sex, diuretic therapy, and urinary protein. In 127/134 (95%) of patients the diuretic was a loop diuretic. None of the other classes of antihypertensive drug was independently associated with serum cholesterol. The only variables significantly associated with HDL cholesterol were sex and the plasma creatinine. Plasma triglyceride was significantly and independently associated with both diuretic therapy and beta-blocker therapy and with age, urinary protein excretion, and plasma albumin. Plasma cholesterol, HDL cholesterol, and triglyceride levels were almost identical in patients receiving triple therapy (cyclosporine 3-5 mg/kg; prednisolone 7-10 mg o.d.; azathioprine 1-1.5 mg/kg) to those in patients receiving conventional immunosuppression (prednisolone 7-10 mg o.d.; azathioprine 2-2.5 mg/kg). Thus these results do not support the existence of a persisting long-term effect of cyclosporine on plasma cholesterol and triglyceride at these doses of the drug. The more striking abnormality of plasma cholesterol and triglyceride in females is unexplained but might be connected with greater sensitivity to low doses of corticosteroids.

The HIF pathway: implications for patterns of gene expression in cancer.

Regulation of the growth and metabolism of large organisms is tightly constrained by the need for precise oxygen homeostasis. Work on control of the haematopoietic growth factor erythropoietin has led to the recognition of a widespread transcriptional response to hypoxia which provides insights into how this is achieved. The central mediator of this response is a DNA binding complex termed hypoxia inducible factor 1 (HIF-1), which plays a key role in the regulation by oxygen of a large and rapidly growing panel of genes. In cancer, activity of the HIF system is up-regulated both by microenvironmental hypoxia and by genetic changes. The clearest example of genetic activation is seen in the hereditary cancer syndrome von Hippel-Lindau (VHL) disease. In normal cells the product of the VHL tumour suppressor gene targets the regulatory HIF subunits (HIF-1alpha and HIF-2alpha) for oxygen-dependent proteolysis, acting as the substrate recognition component of an E3 ubiquitin ligase. In pVHL defective cells this process is blocked leading to constitutive up-regulation of HIF-1alpha subunits, activation of the HIF complex and overexpression of HIF target genes. Using gene array screens we have defined a large number of VHL-regulated genes. The majority of these show hypoxia-inducible responses, supporting the central involvement of pVHL in gene regulation by oxygen. In addition to known HIF target genes involved in angiogenesis, glucose metabolism and vasomotor control, these new targets include examples with functions in matrix metabolism, apoptosis, carbon dioxide metabolism and secondary cascades of transcriptional control. Thus activation of HIF provides insights into the classical metabolic alterations in cancer cells, and into the mechanisms by which microenvironmental hypoxia might influence tumour behaviour. In the case of VHL disease, this activation can be linked to mutations in a defined tumour suppressor gene. Equally regulation of the HIF-1alpha/pVHL interaction in normal cells should provide insights into the physiological mechanisms operating in cellular oxygen sensing.

Production of erythropoietin by liver cells in vivo and in vitro.

Several lines of evidence indicate that the liver is the predominant production site for EPO during the early stages of development. Moreover, in adults, it may contribute significantly to EPO production in some species. Independent of age hepatocytes are the major cellular sites of EPO gene expression in the liver, but quantitatively less significant expression also occurs in at least one additional type of nonparenchymal cell. Although some indirect evidence suggests that these nonparenchymal cells producing EPO may be Kupffer cells, their identity remains to be clarified. Regarding the mechanisms of the adaptation of hepatic EPO production to changes in oxygen availability in the organism, experiments with isolated perfused livers, mixed liver cell cultures, hepatoma cells, and, more recently, isolated hepatocytes have provided evidence that inherent cellular oxygen-sensing mechanisms exist and that external factors are not essential. The use of hepatoma cells has allowed investigation of these cellular mechanisms of oxygen-dependent gene control, and further studies in nontransformed hepatocytes may complement these studies.