Int. J. Cancer: 121, 2434–2442 (2007) ' 2007 Wiley-Liss, Inc.
Erythropoietin gene expression in renal carcinoma is considerably more frequent than paraneoplastic polycythemia .. Michael S. Wiesener1,2*, Philine Munchenhagen3, Markus Gl€ aser3, Bettina A. Sobottka3, Karl X. Knaup1,2, Katrin Jozefowski1, Jan Steffen J€ urgensen3, Jan Roigas4, Christina Warnecke2, Hermann-Josef Gr€ one5, 6 2 2 Patrick H. Maxwell , Carsten Willam and Kai-Uwe Eckardt 1 Interdisciplinary Centre for Clinical Research (IZKF), Friedrich-Alexander University of Erlangen-Nuremburg, Erlangen, Germany 2 Department of Nephrology and Hypertension, Friedrich-Alexander University of Erlangen, Erlangen-Nuremburg, Germany 3 Department of Nephrology and Medical Intensive Care, Charit e of Humboldt-University Berlin, Berlin, Germany 4 Department of Urology, Charit e of Humboldt-University Berlin, Berlin, Germany 5 Department of Cellular and Molecular Pathology, German Cancer Research Centre (DKFZ), University of Heidelberg, Heidelberg, Germany 6 Renal Section, Hammersmith Campus, Imperial College of London, London, United Kingdom Signalling by erythropoietin (EPO) is increasingly recognised as a relevant mechanism in tumour biology, potentially leading to enhanced proliferation, angiogenesis and therapy resistance. Paraneoplastic polycythemia by cancerous overproduction of EPO is a rare event, but most frequently seen in patients with renal cell carcinoma (RCC). The majority of clear cell RCC displays a strong activation of the transcription factor regulating EPO, the Hypoxia-inducible Factor (HIF). Therefore, it is unclear why only a small minority of patients develop polycythemia. We studied 70 RCC for EPO gene and HIFa isoform expression. 34% of all RCC showed expression of EPO mRNA in RNase protection assays, which were almost exclusively of the clear cell type. Only 1 patient presented with polycythemia. In situ hybridisation revealed that expression of EPO was in the tumour cells. Expression of EPO mRNA was always associated with activation of HIF, which could involve HIF-1a and/or HIF-2a. The frequency of EPO gene expression in RCC is therefore much higher than the prevalence of polycythemia. Furthermore, activation of HIF appears necessary for EPO gene expression in RCC, but is clearly not the only determinant. Further to the reported expression of EPO receptors in tumour tissues, the finding of widespread expression of EPO in RCC supports the recent notion of an involvement of this system in paracrine or autocrine effects of tumour cells. ' 2007 Wiley-Liss, Inc. Key words: HIF-2a; EPAS-1; erythrocytosis; oxygen; transcription factor
Erythropoietin (EPO) is the single most important stimulatory factor for erythropoiesis, controlling red cell mass mainly by inhibiting apoptosis of erythrocyte progenitors.1,2 The physiological stimulus for EPO gene expression is reduced delivery of oxygen to the kidney. The molecular events linking changes of reduced oxygenation to increased EPO expression are quite well understood. The transcription factor Hypoxia-inducible Factor 1 (HIF1) is activated in hypoxia and has been identified in vitro as a binding protein complex to an enhancer element 30 to the EPO gene.3 HIF-1 is a heterodimer, consisting of 2 subunits, a constitutive b-subunit and an oxygen dependent a-subunit. The latter is strictly controlled by pericellular oxygen tensions, being rapidly destroyed in the presence of molecular oxygen. With decreasing oxygen availability the HIFa isoform is stabilised, accumulates in the nucleus and transactivates its target genes. To date 2 oxygen dependent HIFa isoforms have been identified, HIF-1a and HIF2a, which are very similarly regulated in vitro4 but differentially expressed in vivo.5–7 Interestingly, there does seem to be a target gene specificity in that HIF-2a seems to be responsible for EPO induction in the kidney, which has been indicated in vitro and in vivo.6,8–11 In the presence of oxygen, HIFa is targeted by a specific E3 ubiquitin ligase, with the von Hippel Lindau (VHL) tumour suppressor protein as recognition component. Binding of the E3 ligase leads to ubiquitination and rapid proteasomal destruction. A prerequisite for binding is the hydroxylation of defined proline residues by a new family of oxygen and iron dePublication of the International Union Against Cancer
pendent proline hydroxylases (for review see Ref. 12). HIF is an important regulator of the hypoxic adaptation, leading not only to increased erythropoiesis but also regulating critical processes, such as glycolysis, angiogenesis, pH-regulation and cell cycle. As such, transactivation by HIF is believed to result in adverse outcome in cancerous disease (for review see Ref. 13). HIF is now known to operate in all mammalian cell types. In contrast, the physiologically relevant production of EPO for hematopoiesis is restricted to a small number of specific cells. In the adult mammal these are mainly the peritubular fibroblasts of the renal cortex. It is possible that hepatic Ito cells and hepatocytes also contribute to hematopoiesis, yet the proportion of this effect is hard to estimate.14–17 Elegant studies with EPO transgenic mice have defined regulatory regions of EPO gene expression 50 of the transcriptional start site, controlling cell specific expression of EPO in general, as well as renal and hepatic expression.18–21 These studies have shown that there is an element between 20.4 and 26 kb upstream of the transcriptional start site, which represses EPO expression in most tissues (but not kidney). Between 26 and 214 kb there is an element which permits or controls renal expression and beyond 214 kb an element which suppresses expression in hepatocytes. To date, the mechanisms by which these repressive elements exert their control are unknown. This could be of clinical interest since recombinant EPO (rEPO) is an expensive treatment in patients with renal insufficiency and cancer. We reasoned that understanding paraneoplastic erythrocytosis—in which normal repressive mechanisms are presumably reversed—could shed light on this. Of the different tumours associated with paraneoplastic erythrocytosis, renal cell carcinoma (RCC) is the most common. RCC is believed to arise from proximal renal tubular epithelium, which does not normally express a detectable level of EPO even when oxygen delivery to the kidney is reduced and HIF is activated. The possibility that the non-cancerous stroma/adjacent renal interstitium are the source of hormone secretion has effectively been
This article contains supplementary material available via the Internet at http://www.interscience.wiley.com/jpages/0020-7136/suppmat. Abbreviations: EPO, erythropoietin; Hb, haemoglobin concentration; HIF, hypoxia-inducible factor; RCC, renal cell carcinoma; rEPO, recombinant EPO; RPA, RNase protection assay; RT-PCR, reverse transcriptionpolymerase chain reaction; VHL, von Hippel Lindau. Grant sponsor: Deutsche Forschungsgemeinschaft; Grant number: SFB 423; Grant sponsors: IZKF Erlangen, EU (Framework 6 Integrated Project EUROXY). *Correspondence to: Nikolaus-Fiebiger-Zentrum f€ur Molekulare Medizin, Friedrich-Alexander Universit€at Erlangen-N€urnberg, Gl€uckstrasse 6, 91074 Erlangen, Germany. Fax: 149-9131-8539340. E-mail:
[email protected] Received 1 September 2006; Accepted after revision 21 May 2007 DOI 10.1002/ijc.22961 Published online 19 July 2007 in Wiley InterScience (www.interscience. wiley.com).
ACTIVATION OF HIF AND EPO IN RENAL CELL CARCINOMA
excluded as follows. First, in several cases of paraneoplastic polycythemia caused by RCC, it has convincingly been shown that the tumour cells are the site of EPO overproduction.22–24 Second, this has been confirmed in cell lines derived from primary RCC25–29. Third, when these cell lines were transplanted as subcutaneous xenografts the recipient mice developed polycythemia.26,27,30,31 Importantly, RCC’s often have biallelic inactivation of the VHL gene. As a result, HIFa subunits are stabilised and HIF is constitutively active.32 This situation applies in 70% of clear cell RCC, which accounts for 85% of all RCC. VHL defective tumours show homogeneous overexpression of HIFa subunits and their target genes, including a range of genes whose products promote angiogenesis, proliferation, and metastasis.33,34 Besides RCC, there are other links between mutations of the VHL gene and polycythemia. First, paraneoplastic erythrocytosis has been reported with other tumours, which are associated with loss of VHL function, including haemangioblastoma and pheochromocytoma.35 Second, several mutations of the VHL gene have been identified, which cause familial and congenital polycythemia,36–41 without any tumour formation. In these individuals it is not yet known whether excess EPO is produced by ‘professionalÕ EPO producing cells, or whether it is derepressed in other cell types. Taken together, these observations led us to speculate that there may be a connection between VHL function and the normal physiological tissue restriction of EPO expression. Possibly confirming the notion of a special role of VHL for regulation of EPO, a number of studies have previously demonstrated a very high frequency of coexpression of EPO and its receptor in tumours of the VHL-syndrome: endolymphatic sac tumours, pheochromocytomas, renal cysts and RCC.42,43 Very recently this has also been demonstrated in sporadic clear cell RCC, of which the majority also bear a VHL inactivation.44 In general, expression of EPO and its receptor in tumours and tumour cell lines has attracted much attention recently. Numerous studies have demonstrated expression of both related proteins in a large array of different tumours, which has been reviewed in depth recently.45 A growing body of evidence suggests that signalling through these tumour EPO receptors, which could for instance occur under anaemia therapy with rEPO, may lead to enhanced aggressiveness of tumour cells. This area of research is highly controversial, as cellular effects could be counterbalanced by improved oxygen transport and oxygenation of the tumour tissues.45,46 However, 2 recent randomised and placebo controlled reports have demonstrated worse outcome in terms of progression-free survival in the rEPO treated groups of breast and head and neck cancer.47,48 Equally to the effects of exogenous rEPO, intrinsic expression and secretion of EPO from the tumour tissue could theoretically lead to autocrine and/or paracrine effects. To the best of our knowledge, a systematic analysis of EPO gene expression in a large series of sporadic RCC has so far not been performed. Furthermore, to understand the underlying mechanisms of EPO expression, it would be of importance to study the expression of HIF-1a and HIF-2a in parallel. We have therefore collected clinical samples of RCC, generated numerous cell lines derived from primary RCC and analysed these for HIF and EPO expression. Material and methods Tumour collection Tissue specimens were collected from total nephrectomy in patients with RCC. All patients gave their informed consent. The study was approved by the local ethics committees. Samples of healthy kidney tissue distant from the tumour and tumour-tissue were snap frozen in the operating theatre, immediately after surgical removal and stored at –80°C until analysis. Preoperative haemoglobin concentrations (Hb) and routine pathological findings were recorded, according to the international classification.49 In 37 patients deep frozen samples of preoperative serum was available for analysis.
2435
Cell culture As standard for HIFa protein or EPO mRNA Hep3B or HepG2 cells were used for immunoblotting or RNAse protection. Cells were grown under standard conditions (DMEN, 10% fetal calf serum, 100 U/ml penicillin, 0.1 mg/ml streptomycin [antibiotics] from PAA, Coelbe, Germany) to subconfluence and exposed to either normoxia or hypoxia (1% O2, 5% CO2, 94% N2) for 4 hr (protein extraction) or 16 hr (RNA extraction). Hypoxic incubations were performed in the ‘invivo400Õ hypoxic workbench (Ruskinn Technology, West Yorkshire, UK). RCC4 cells were described previously.32 Haemagglutinin (HA) tagged VHL cDNA, wild-type and with site directed mutagensis (598 C > T), were transfected into RCC4. Selection of stable clones was performed with G418, 1 mg/ml (PAA, Coelbe, Germany). RNA extraction and RNase protection assay Tissues were weighed and homogenised into RNA Bee (Biogenesis, Poole, U.K.) in a weight/volume ratio of 1:10 with an electrical homogeniser (Ultra-Turrax, IKA, Staufen, Germany). Extraction of total RNA was performed according to the instructions of the manufacturer. RNase protection assays (RPA) were essentially performed as previously described.33 The following radiolabeled probes and amounts of total RNA were used: EPO protected fragment 150 bp, accession number X02157, nucleotides 358–391, 50 lg, Carbonic anhydrase 9 (CA9) 146 bp, accession number Z54349, nucleotides 3,631–3,777, 20 lg and U6 small nuclear RNA 106 bp, accession number X01366, nucleotides 1–107, 1 lg. Images were visualised by autoradiography. Protein extraction and Immunoblotting Tissues and cells were homogenised into protein extraction buffer (7 M urea, 10% glycerol, 10 mM Tris-HCl (pH 6.8), 1% SDS, 5 mM DTT, 0.5 mM 4-(2-aminoethyl)benzenesulfonyl fluoride and 1 mg/l each of aprotinin, pepstatin and leupeptin) as described in detail.4,33 Immunoblotting was performed as described previously.4 For HA detection clone 12CA5 (Roche, Mannheim, Germany) was used at 0.5 lg/ml. For HIF-1a clone 54 at 1 lg/ml (Transduction Laboratories, Lexington, KY) and for HIF-2a clone 190b4 at a dilution of 1:500 or ab-199 (1:1,000, Novus Biologicals, Littleton, CO) were used, respectively. Secondary anti-mouse or anti-rabbit HRP-conjugated antibodies (DAKO, Ely, U.K.) were used at a dilution of 1:2,000. Detection was performed with enhanced chemiluminesense (SuperSignal West Dura, Pierce, Rockford, IL) and autoradiography. Each membrane was finally stained with Comassie blue to assess for equivalence of transfer and loading. Tumour extraction for generation of cell-lines For generation of primary cell lines, fresh tumour tissue was collected in the operating theatre and transported in sterile PBS at 4°C. The tissues were mechanically minced and digested at 37°C for 120 min in digestion buffer (100 U/ml DNase1, 50 U/ml collagenase III, 150 U/ml collagenase IV, 200 U/ml hyaluronidase in final volume of 20 ml PBS). Homogenates were then centrifuged and erytrocyte lysis performed with hypotonic buffer. Finally the homogenate was again centrifuged and the upper layer of cells were seeded into tissue culture flasks into Leibovitz Medium, supplemented with 10% FCS, antibiotics (see earlier), 1% MEM Vitamins, 80IE Insulin, 2.5 mg/l Transferrin and 1 g/l sodiumbicarbonate (all from PAA, Coelbe, Germany). Approximately 2 weeks after cell extraction, cells were transferred to RPMI, 10% FCS and antibiotics. The epithelial origin of most of these clones was confirmed by detecting cytokeratin 1850,51 expression in immunoblots (data not shown). In situ hybridisation Eight micrometer cryosections were cut and treated as described. Sense and antisense probes were labelled with 35S-UTP
2436
WIESENER ET AL. TABLE I – HIFa PROTEIN EXPRESSION IN CLEAR CELL AND OTHER TYPES OF RCC
Total (n 5 70) Clear cell RCC (n 5 53) Other (n 5 17)
HIF-1a positive
HIF-2a positive
Only HIF-1a
Only HIF-2a
No HIFa
56 (80%) 47 (89%) 9 (53%)
52 (74%) 42 (79%) 10 (59%)
7 2
2 3
2 5
Numbers shown are from results of immunoblotting for HIF-1a and HIF-2a protein in the investigated tumours. Percentage values were rounded to the nearest full number. TABLE II – EPO GENE EXPRESSION IN CLEAR CELL AND OTHER TYPES OF RCC
Total tumours (n 5 64) Clear cell RCC (n 5 52) Other (n 5 12)
EPO positive
EPO negative
24 (38%) 23 (44%) 1
40 (63%) 29 (56%) 11
Numbers shown are results from RNAse protection assays from the investigated tumours. Percentage values were rounded to the nearest full number.
FIGURE 1 – HIFa protein status and target gene expression in nephrectomy samples of individual patients. (a) Immunoblots for HIF1a and HIF-2a from whole cell extracts of tumour (T, all clear cell RCC) and adjacent healthy kidney (k) of the same patient, showing positive signals of varying extent in all specimens. Protein lysates from Hep3B tissue cultures exposed to either normoxia (N) or hypoxia (H, 1% O2 for 4 hr) were used as standard. From each sample 100 lg of protein were resolved on SDS-PAGE. (b) RNAse protection assay for carbonic anhydrase 9 (CA9) and EPO for the identical patients as in (a). Signals for CA9 indicate strong activation of the HIF pathway in each sample of clear cell RCC. Nevertheless, EPO is expressed only in part of these tumours, at varying degrees. U6 small nuclear RNA (U6sn) was used as internal control. Hep3B cells were used as controls, with hypoxic stimulation (H, 1% O2) for 16 hr.
(Amersham Biosciences, Freiburg, Germany) for full length human EPO as templates. In-situ hybridisation was then performed as described previously.52 ELISA of serum EPO Serum EPO concentrations were measured from 25 ll patient samples using a commercial kit, according to the manufacturer’s instructions (medac GmbH, Hamburg, Germany). Reagents Unless otherwise indicated chemicals were purchased from Sigma (St. Louis, MO). Results We have collected and analysed a total of 70 RCC from 3 different centres in Germany. 53 of these were of the clear cell (76%) and 17 of various other histopathological types (24%). This is the expected ratio, which is in accordance with previous reports.53
First, we analysed the tumours for HIF activation. We saw HIF1a overexpression in 80% of all RCC. HIF-1a accumulation was more frequent in clear cell RCC than in other types of RCC (89% vs. 53%, respectively). This was as expected from previous work of our own, and other groups.33,34 HIF-2a protein showed a similar frequency of overexpression, being found in 74% of all RCC, 79% of clear cell RCC and 59% of other RCC (Table I). Figure 1a shows immunoblots of tissues from the tumour and adjacent kidney for HIF-1a and HIF-2a of selected patients with a clear cell RCC. Strong expression of both HIFa isoforms can be seen in tumour tissue, whereas there is little or no signal in nontumourous renal tissue. The abundance of HIFa expression was generally lower or absent in other types of RCC, when compared to clear cell RCC (data not shown and33). These findings are compatible with genetic, homogenous stabilisation of HIFa isoforms in the majority of clear cell RCC, where VHL has been inactivated. Less frequent and less pronounced HIFa expression in other tumours is most likely due to microenvironmental stimulation, namely intratumoral regional hypoxia. Much attention has recently been directed towards the HIFa specific expression and function. Cumulative evidence may indicate a prominent role of HIF-2a over HIF-1a for tumour progression, in particular for RCC.54–57 In our material, most tumours showed overexpression of both HIFa isoforms (n 5 49), whereas a minority of cases displayed either HIF-1a or HIF-2a overexpression (n 5 14), or no HIFa expression at all (n 5 7). Interestingly, 2 clear cell RCC showed no detectable HIFa expression at all. Overall, HIFa isoform expression showed a higher frequency in clear cell than non clear cell RCC and there was a tendency towards more frequent expression of HIF-1a than HIF-2a, in particular in clear cell RCC (Table I). It is clear from these data that the HIF overexpression alone in renal tumours is not sufficient to cause polycythemia. Next we directly assayed EPO and another gene, CA9, using ribonuclease protection assays. We selected CA9 because it is a specific HIF-1a target gene, which is highly expressed in RCC.56 We therefore compared expression of EPO and CA9 in the renal and tumerous tissues of each patient material displayed. Figure 1b shows the expression of both HIF target genes in the identical patient samples, for which HIFa protein expression is depicted in Figure 1a. Functional activation of the HIF pathway can be verified by the strong expression of CA9 in every single tumour, similarly to the coordinated upregulation of VEGF and GLUT1, which we described previously.33 In sharp contrast, most tumours do not show any expression of EPO, although there is an obvious activation of its transcriptional machinery. In total, EPO mRNA could be found in 34% of all tumours assayed. With 1 exception (an oncocytoma) all these EPO positive tumours were clear cell RCC, 40% of clear cell RCC in total (Table II). Figure 2 depicts a larger
ACTIVATION OF HIF AND EPO IN RENAL CELL CARCINOMA
2437
FIGURE 2 – EPO gene expression in 40% of clear cell RCC. RNAse protection assay for CA9 and EPO in a large number of clear cell RCC, and 2 nonclear cell RCC. The assay shows very strong expression of CA9 in most of clear cell RCC. In contrast, EPO is expressed in 40% of clear cell RCC, at greatly varying levels. Only very few tumours show high and medium level of expression, whereas the majority of positive tumours were only modestly positive. The 2 nonclear cell RCC tumours show no or very weak induction of CA9 and no EPO expression, which is representative for most tumours of this group. U6sn and extracts of Hep3B cells (normoxia N, 16 hr hypoxia, H) were used for controls.
number of RCC analysed for EPO and CA-9 expression, showing that the expression level of EPO mRNA in individual tumours was different. A few tumours showed a very strong level of expression, whereas most others displayed a modest or weak expression level. Comparison to the pattern of HIF isoform expression showed that all EPO expressing tumours showed overexpression of at least 1 HIFa isoform, and most showed expression of both. Interestingly, we found 2 EPO positive tumours, which showed expression of a single isoform only, 1 with HIF-1a and the other HIF-2a. This would be consistent with the idea that HIF activation is necessary for EPO expression, and suggests that both HIF-1a and HIF-2a are capable of driving EPO expression in this setting. Making the reverse comparison, the majority of HIFa-positive tumours did not show EPO expression, although they did show strong CA9 expression. Thus, activation of the HIF pathway is obviously not sufficient to release EPO gene expression, which may point to other intrinsic, to date undefined differences in individual tumours. There is uncertainty about the functional significance of EPO expression/secretion by RCCs for hematopoiesis. In contrast to the rare occurrence of polycythemia, our analysis shows that up to 40% of tumours express EPO. We therefore wanted to know whether the EPO status of the tumour had influences on the preoperative Hb concentrations of the patients. Figure 3 shows the analysis for the patients with EPO negative and EPO positive tumour tissues. Although the difference was not statistically significant (p 5 0.06) there was a trend towards higher Hb concentrations in the patients with a tumour expressing EPO mRNA. Additionally, we aimed to evaluate the effect of EPO expression in the tumour on circulating EPO concentrations. In 37 cases we were able to retrieve serum from frozen samples and performed ELISA for EPO measurement. Forming 2 groups for patients expressing, or not expressing EPO mRNA in their RCC, the mean concentration of serum EPO was almost doubled in the former (15.88 IU/l 6 5.8 SEM), than in the latter (8.43 IU/l 6 1.3 SEM). Nevertheless, using the Mann–Whitney U Test this difference did not reach statistical significance. This robust, nonparametric test was chosen because distributional assumptions were in doubt. To further investigate the molecular mechanisms underlying EPO expression by some RCCs, we aimed to generate EPO producing cell lines from these tumours. Seeing that 40% of tumours show expression of EPO and given that EPO secreting cell-lines have been successfully established in RCC patients with polycythemia (see Introduction), we reasoned that this should be technically feasible. To do this, we needed fresh material, so recruited further patients. From these we successfully generated and expanded 24 RCC cell cultures sufficiently to prepare protein and RNA extracts. However, not all of these continued to grow
FIGURE 3 – The existence of EPO gene expression is weakly associated with Hb concentrations. Preoperative Hb levels of 101 patients were compared to the EPO status of the tumour (negative or positive) as determined in RNAse protection assays, irrespective of quantitative differences in mRNA expression. The Hb concentration (Hb, mean 6 SD) was 12.88 g/dl (61.83) and 13.57 g/dl (62.51) for patients without and with EPO expression in tumours, respectively. The Mann– Whitney test for comparison of means was used because distributional assumptions were in doubt. The difference approached but did not reach statistical significance in this cohort (p 5 0.06). The boxplots illustrate the distribution of Hb concentrations in both groups. Boxes extend from the lower to the upper quartile, with the median (bold bar) dividing the box. ‘WhiskersÕ range from the least to the greatest values. The male:female ratio was comparable in both groups. [Color figure can be viewed in the online issue, which is available at www. interscience.wiley.com.]
indefinitely. Most of the cell lines originated from clear cell RCC. 16 of these cell cultures showed constitutive expression for HIF1a and/or HIF-2a, indicating that the original tumour bears a VHL inactivation (Table III). Figure 4 shows immunoblots for both HIFa subunits from selected cell lines, with inducible, constitutive and variable expression pattern of the HIFa subunits. In accordance with the data from the tumour tissue described earlier, most of these cell lines showed expression of both HIFa subunits with some showing expression of one or the other HIFa subunit. Interestingly, from the 16 constitutive HIF-expressing lines 4 each showed expression of only HIF-1a or HIF-2a (Table III). When assaying for HIF target gene expression, CA9 expression strongly correlated with the HIF status of the cell line, being either constitutive or inducible by hypoxia. Strikingly, none of these lines expressed EPO, even when incubated in hypoxia to activate HIF (Fig. 5a). In the cases where tumour tissue was primarily intended for generation of a cell line, substantial amount of the tumour was needed for the extraction procedure. Therefore analysis for EPO expression of the tumour was restricted to cases where a large amount of tissue was provided. Figure 5b shows the samples of 2 patients in whom analysis of both tissues and cell line was possible, and in whom EPO expression is clearly present in the tumour but not in the derived cell culture. Several hypotheses could be considered in the light of this negative result. One could be that the tumour cells dedifferentiate under the culture conditions or lack some critical cytokine. Another could be that EPO expression in the tumour derives from interstitial/stromal cells, as in the normal kidney, and would not
2438
WIESENER ET AL. TABLE III – HIF PROTEIN STATUS OF NEWLY GENERATED RCC LINES Constitutive regulation (n 5 16)
HIF-1a1-2a 8
Only HIF-1a 4
Inducible regulation (n 5 8)
Only HIF-2a 4
HIF-1a12a 2
Only HIF-1a 5
Only HIF-2a 1
From 24 different primary tumour tissues cell lines were successfully generated and passaged for at least 4 weeks prior to analysis. Cells were exposed to normoxia or hypoxia (1% O2, 4 hr) and whole cell extracts immunoblotted for both HIFa subunits. The table differentiates into inducible and constitutive expression of HIFa, where the latter most likely arises from inactivation of VHL.
FIGURE 4 – RCC derived cell lines show a heterogeneous pattern of HIFa expression. Immunoblots for HIF-1a or HIF-2a from different cell lines derived from primary tumour tissues. Cells were exposed to normoxia (N) or hypoxia (H, 4 hr 1% O2). The figure shows that the different lines vary in their HIFa expression, where some express only one, others both HIFa isoforms. Furthermore, some of the lines show inducible HIFa expression (arrow pointing upward), whereas others demonstrated constitutive expression (horizontal arrow), indicating inactivation of VHL. Hep3B cells were used as standards.
continue to grow under the culture conditions. Against this, others have reported that tumour cells and not the stroma express the EPO gene (see Introduction). To address this in our material we performed in situ hybridization for EPO on selected tissues of healthy kidney, EPO negative, weakly positive and strongly positive tumours (n 5 3 of each). Figure 6 shows representative results for these tissues. Only background signal was seen in healthy kidney (Fig. 6a) and EPO negative tumours (Fig. 6b), as well as in the tumours that showed only weak EPO expression in RPA, where the sensitivity was presumably not sufficient for reliable detection. Strongly EPO expressing tumours showed a strong positive labelling (Figs. 6c and 6d), which clearly mapped to the tumour cells. Successful generation of EPO expressing cell lines have so far all been undertaken from tumours of polycythemic patients. However, none of the patients from whom we developed cell cultures were polycythemic. One hypothesis would be that specific VHL mutations might cause a gain of function, in addition to their activation of HIF, resulting in derepression of EPO by renal epithelial cells. Unfortunately, there is as yet no study of the VHL inactivation events in polycythemic patients with RCC, which might capture such mutations. However, a prominent VHL mutation which is connected to familial polycythemia, and leads to EPO overexpression, is the ‘ChuvashÕ mutation (598 C > T missense mutation resulting in Arg200Trp in VHL41). We therefore tested whether the presence of mutant VHL proteins bearing this mutation might derepress EPO production in a RCC cell line. RCC4 cells, which have a null background for VHL and EPO were stably transfected with a HA tagged cDNA baring the VHL ‘ChuvashÕ mutation. Numerous clones were analysed, most of which exhibited detectable levels of Chuvash-VHL, as shown by immunoblotting against the HA-tag (Suppl. Fig. 1a). The transgene expression led to a varying extent of HIF-1a repression in normoxia across the different clones, with restored inducibility, consistent with previous evidence that the mutant protein retains some ability to capture and regulate HIFa subunits (Suppl. Fig. 1b).
FIGURE 5 – RCC derived cell lines loose the ability to express EPO. (a) RNAse protection assays for EPO and CA9 from different cell lines derived from primary tumour tissues. Cells were exposed to normoxia (N) or hypoxia (H, 16 hr 1% O2). Whereas most RCC lines show expression of CA9, none of these express EPO. Hypoxic stimulation does not release EPO gene expression in these cultures. (b) RNAse protection assays for EPO and CA9 for the samples of 2 patients. Whereas EPO was expressed in the tumour tissue (T), but not the adjacent kidney (k), the derived cell line did not show any expression. Interestingly, although both tumour tissues were positive for EPO, the cells from patient CL18 showed an inducible pattern for CA9, indicating a fully functional VHL status in terms of HIFa degradation.
The clones that showed detectable levels of transgene expression were taken through to evaluation of target gene expression by RNase Protection (Suppl. Fig 1c). Again, CA9 expression was consistently influenced and inducibility restored to a different extent in most clones, displaying the indirect transcriptional effects of the stably integrated transgene. Nevertheless, none of the clones showed any expression of the EPO gene. Therefore, the Arg200Trp mutation of VHL is not able to derepress EPO expression in this cell line.
Discussion The syndrome of paraneoplastic polycythemia has long been considered a typical feature of RCC. Nevertheless, this condition
ACTIVATION OF HIF AND EPO IN RENAL CELL CARCINOMA
2439
FIGURE 6 – Tumour cells are the site of EPO expression in a subset of RCC. In situ hybridisation for EPO in different tissues. Whereas no specific signal was detectable in normal kidney (a) or weakly expressing RCC (b), as determined by RPA, strongly expressing RCC (c, d) showed positive labelling for EPO. Detection of EPO was clearly mapped to the tumour cells in RCC. Large images represent the dark field microscopy, whereas the small images show the identical tissue section in bright field microscopy, counter stained with hematoxylin. Corresponding images are connected by the black bar.
is rare, being reported in less than 5% in these patients.58 The great majority of RCCs are caused by inactivation of the tumour suppressor gene VHL. This leads to stabilisation and activation of HIFa subunits with strong upregulation of its target genes in the tumour cells, throughout the tissue.33,34 Genes like VEGF, GLUT1 and CA9 are coordinately overexpressed together with activation of HIF in the majority of RCCs. EPO is a further classical target gene of HIF, but in contrast to the other genes characterised by a physiologically restricted tissue and cell specific expression. VHL-associated RCC is of clear cell histopathological phenotype. These tumours are believed to arise from renal tubular epithelial cells,53 which are normally not able to express the EPO gene. Therefore, a tumour-related event obviously takes place in RCC, which leads to release of EPO expression. Our study set out to determine the frequency of this molecular alteration and to integrate HIF and EPO expression with Hb concentrations and polycythemia in a large number of RCC patients. Overexpression of the EPO gene occurs frequently in CCRCC A major finding of our study is that EPO gene expression in RCC is far more frequent than the incidence of polycythemia suggests, being observed in 40% of cases. Of note, these figures are based on the results of robust RPAs, which are less sensitive than reverse transcription—polymerase chain reaction (RT-PCR) analysis, but are completely specific and have been shown to reliably detect expression levels that are relevant in terms of the erythropoietic effects of EPO production. Using the same type of assay in experimental animals EPO gene expression has previously been demonstrated in liver and kidneys, its main production sites, and was also detected in much smaller amounts in some other organs, which are
unlikely to contribute significantly to serum EPO production.2 Importantly, when analysed as EPO mRNA per microgram total RNA the level of EPO expression in these tumours was grossly different. Few tumours showed strong expression, comparable to that in severely hypoxic rodent kidney, whereas in most others only a moderate or barely detectable level was found, suggesting that influences on EPO gene expression are not uniform between different tumours. It is possible that with more sensitive techniques, such as RT-PCR, an even higher prevalence of expression of the EPO gene as compared to adjacent kidney would have been detectable, but such a finding would not alter the implications. Theoretically the differences in EPO expression could result from different origins of expression, with tumour cells in some cases and the nonmalignant stromal cells in others. In the latter case influences from the microenvironment, namely hypoxia, could be the driving force and not genetic influences, thus also potentially explaining greater heterogeneity. Therefore we attempted to detect EPO in these tissues on a cellular level by in situ hybridisation. Unfortunately the sensitivity of the method was only sufficient for reliable detection in the small group of tumours with very high levels of EPO expression. In these tissues the tumour cells were confirmed to be the site of EPO expression, as previously reported in 2 case reports of patients with polycythemia.22,23 Along the same lines, a number of recent publications have reported immunostaining for EPO in tumour cells of kidney cancer.59–61 However, when stimulated EPO secretion is dependent on de novo synthesis, as it is not stored intracellularly.62 Furthermore, EPO is primarily regulated at its RNA level, equally by transcriptional mechanisms and mRNA stability.2 We therefore preferred to specifically detect the mRNA transcripts by in situ hybridisation, although sensitivity may have been a confounding factor.
2440
WIESENER ET AL.
EPO gene expression is weakly associated with serum EPO concentrations and Hb levels The frequency of EPO gene expression raises the question as to why RCCs are only associated with polycythemia in rare cases. Accordingly, in our series only 1 patient had polycythemia, which we reported previously.63 The serum EPO concentrations were higher in the group of patients expressing EPO mRNA in their tumour. In the whole study population slightly higher preoperative Hb concentrations were observed in those patients in whom a positive signal for EPO mRNA was detectable in the tumour. However, both parameters approached but failed to reach statistical significance, which is probably caused by the large variance and too small patient numbers. In large these data indicate that the tumerous tissue can contribute to hematopoiesis. In addition to the level of EPO mRNA expression, it is conceivable that the mere size of the lesion could influence the effect by increasing amounts of secreting tissue. Nevertheless, the tumour size did not correlate with the Hb concentrations in our series (data not shown). When considering the relationship between EPO expression and Hb concentrations in patients with RCC it also needs to be taken into account that tumour related conditions, such as decreased iron availability and inflammation (i.e., causes of the anaemia of cancer), could counterbalance elevated EPO blood levels, at least to some extent. We are not aware of reports comparing the occurrence and severity of anaemia from patients with an RCC with other types of malignant diseases, which might show a difference. Stabilization of HIF is a prerequisite for, but insufficient for EPO gene expression in RCC All but one of the tumours overexpressing EPO were clear cell RCC, which could already point to the involvement of VHL inactivation and/or HIF stabilisation, which is characterisitc for this type of RCC. Without exception the EPO positive tumours also showed overexpression of HIFa, which is therefore apparently a conditio sine qua non for EPO gene expression. The significant differences in EPO gene expression between tumours, however, were not related to the level of HIF accumulation. In this respect EPO was found to differ markedly from other HIF target genes, for which the difference in expression levels was lower and a clear association was observed with HIF. One of the possible factors determining different levels of EPO gene expression could have been differential expression of the 2 HIF isoforms HIF-1a and HIF-2a. In previous studies we were able to demonstrate that EPO induction in cell lines is dependent on HIF-2a, not HIF-1a.8 This has been confirmed in vivo in HIF2a knock-out animals.9–11 Furthermore, the cells physiologically producing EPO in liver and kidneys also express HIF-2a rather than HIF-1a.5,6 In our analysis of the tumours both the tumour tissues and the derived cell lines showed approximately equal distribution of expression for one or the other HIFa isoform, or both. Most of the EPO positive tumours displayed overexpression of both HIFa subunits, thereby precluding any conclusions on isoform specificity. Nevertheless, in 2 EPO positive tumours we were only able to detect HIF-1a expression. Therefore, at least in these 2 cases the ‘physiologicalÕ target gene specificity of HIFa subunits seems to have been overcome. Recently the target gene specificity of a number of HIF genes has been investigated in a variety of tumour cell lines. Interestingly, some genes also showed irregular dependency on either HIF-1a or HIF-2a, in particular in RCC cells.56 Irrespective of the mechanisms responsible for this, it is clear that the different level of EPO gene expression cannot be explained on the basis of differential HIF isoform expression. Although most frequently seen in RCC, the syndrome of paraneoplastic polycythemia due to EPO overexpression has also been reported in single cases in a great number of different types of tumours, such as uterus, stomach, testes, liver and brain derived.64–68 The molecular mechanism of this uncoupled EPO secretion in these cases has not been investigated. Whether VHL mutations could contribute in this background remains specula-
tive. However, recent reports could suggest a critical importance of the type of VHL mutation, rather than a mere functional inactivation for the development of polycythemia. Different germline VHL mutations have been identified as being associated with congenital polycythemia without being tumorigenic, including the ‘ChuvashÕ mutation (R200W) and 9 further mutations. Of note, these mutations are exclusively point mutations, which lead to a single amino acid exchange (P192A, L188V, H191D, D126Y, V130L, L163P, G144R, L188V, Y175C36–40). Furthermore most of these mutations affect the a/b interphase of the VHL molecule and some are reported to occur heterozygously with the wild type allele (D126Y, G144R, Y175C, G104V). Others show a homozygous or heterozygous compound defect. These findings could imply a ‘gain of functionÕ situation, for which reason we undertook the stable transfection of the ‘ChuvashÕ mutation into RCC4. At least in this background EPO expression was not released, leaving the question open as to whether even further events are necessary. Irrespective of what the essential mechanism is in addition to HIF stabilization that drives EPO gene expression in RCC, it is intriguing that this mechanism was lost during the development of cell lines from RCC. At least in some cases we were able to demonstrate EPO expression in the tumour, whereas the derived cell line did not show any expression, despite persistent dysregulation of VHL and continuous upregulation of HIF in an oxygen-independent fashion. To date, despite extensive attempts it has also not been possible to cultivate EPO producing cells from healthy kidney, most likely because of dedifferentiation processes in tissue culture.2 This may also have occurred in our cell lines. Why our results differ from those of other investigators who have previously demonstrated the successful generation of EPO producing cell lines from RCC, remains unclear. It must be stressed however, that in contrary to these reports, none of the patients from which we generated cell lines were polycythemic. This could imply that there is a clear mechanistic difference rather than just a quantitative difference between these populations.
Potential future implications The interest in EPO regulation has weaned since the availability of recombinant human EPO for replacement therapy in patients with renal and non-renal anaemias. However, EPO therapy is associated with high costs and worldwide the potential benefits of this therapy go far beyond its affordability. The present study demonstrates that EPO gene expression can occur rather frequently under certain pathologic circumstances in an oxygen-independent fashion in a cell type that does not normally produce EPO. Unraveling the underlying mechanisms may offer novel therapeutic options to stimulate endogenous EPO formation. Furthermore, the EPO secretion of RCCs may also be important for endogenous tumour behaviour. Numerous studies have demonstrated expression of the EPO receptor in tumour tissues of different origins and derived cell lines, which has lead to concerns about the effects of treatment with rEPO for cancer anaemia (for review see Ref. 45). In RCC expression of the EPO receptor has been demonstrated previously.42,44,69 Stimulation of the tumour cells with EPO exerted powerful mitogenic effects.69 Taking these observations, in combination with the findings presented in our study, one could speculate that paracrine or autocrine signalling of EPO may be a relevant tumour growth promoting mechanism in RCC. We propose therefore that EPO gene expression be evaluated as a prognostic marker of RCC and that blocking strategies are worth to be explored. Acknowledgements 8 novel primary RCC cell lines were kindly provided by Uwe Lohmann and Ingo Schmidt. Riboprobes and RCC4 were a gift from Peter Ratcliffe.
ACTIVATION OF HIF AND EPO IN RENAL CELL CARCINOMA
2441
References 1. 2. 3. 4.
5.
6.
7.
8.
9.
10. 11. 12. 13. 14.
15.
16.
17. 18. 19.
20.
21.
22. 23.
Fisher JW. Erythropoietin: physiology and pharmacology update. Exp Biol Med (Maywood) 2003;228:1–14. Fandrey J. Oxygen-dependent and tissue-specific regulation of erythropoietin gene expression. Am J Physiol Regul Integr Comp Physiol 2004;286:R977–R988. Wang GL, Semenza GL. Purification and characterization of hypoxiainducible factor 1. J Biol Chem 1995;270:1230–7. Wiesener MS, Turley H, Allen WE, Willam C, Eckardt KU, Talks KL, Wood SM, Gatter KC, Harris AL, Pugh CW, Ratcliffe PJ, Maxwell PH. Induction of endothelial PAS domain protein-1 by hypoxia: characterization and comparison with hypoxia-inducible factor-1a. Blood 1998;92:2260–8. Wiesener MS, Jurgensen JS, Rosenberger C, Scholze CK, Horstrup JH, Warnecke C, Mandriota S, Bechmann I, Frei UA, Pugh CW, Ratcliffe PJ, Bachmann S, et al. Widespread hypoxia-inducible expression of HIF-2a in distinct cell populations of different organs. FASEB J 2003;17:271–3. Rosenberger C, Mandriota S, Jurgensen JS, Wiesener MS, Horstrup JH, Frei U, Ratcliffe PJ, Maxwell PH, Bachmann S, Eckardt KU. Expression of hypoxia-inducible factor-1a and -2a in hypoxic and ischemic rat kidneys. J Am Soc Nephrol 2002;13:1721–32. Stroka DM, Burkhardt T, Desbaillets I, Wenger RH, Neil DA, Bauer C, Gassmann M, Candinas D. HIF-1 is expressed in normoxic tissue and displays an organ-specific regulation under systemic hypoxia. FASEB J 2001;15:2445–53. Warnecke C, Zaborowska Z, Kurreck J, Erdmann VA, Frei U, Wiesener M, Eckardt KU. Differentiating the functional role of hypoxia-inducible factor (HIF)-1a and HIF-2a (EPAS-1) by the use of RNA interference: erythropoietin is a HIF-2a target gene in Hep3B and Kelly cells. FASEB J 2004;18:1462–4. Scortegagna M, Ding K, Zhang Q, Oktay Y, Bennett MJ, Bennett M, Shelton JM, Richardson JA, Moe O, Garcia JA. HIF-2a regulates murine hematopoietic development in an erythropoietin-dependent manner. Blood 2005;105:3133–40. Scortegagna M, Morris MA, Oktay Y, Bennett M, Garcia JA. The HIF family member EPAS1/HIF-2a is required for normal hematopoiesis in mice. Blood 2003;102:1634–40. Gruber M, Hu CJ, Johnson RS, Brown EJ, Keith B, Simon MC. Acute postnatal ablation of Hif-2a results in anemia. Proc Natl Acad Sci USA 2007;104:2301–6. Schofield CJ, Ratcliffe PJ. Oxygen sensing by HIF hydroxylases. Nat Rev Mol Cell Biol 2004;5:343–54. Pouyssegur J, Dayan F, Mazure NM. Hypoxia signalling in cancer and approaches to enforce tumour regression. Nature 2006;441:437– 43. Bachmann S, Le Hir M, Eckardt KU. Co-localization of erythropoietin mRNA and ecto-50 -nucleotidase immunoreactivity in peritubular cells of rat renal cortex indicates that fibroblasts produce erythropoietin. J Histochem Cytochem 1993;41:335–41. Maxwell PH, Osmond MK, Pugh CW, Heryet A, Nicholls LG, Tan CC, Doe BG, Ferguson DJ, Johnson MH, Ratcliffe PJ. Identification of the renal erythropoietin-producing cells using transgenic mice. Kidney Int 1993;44:1149–62. Maxwell PH, Ferguson DJ, Osmond MK, Pugh CW, Heryet A, Doe BG, Johnson MH, Ratcliffe PJ. Expression of a homologously recombined erythopoietin-SV40 T antigen fusion gene in mouse liver: evidence for erythropoietin production by Ito cells. Blood 1994;84: 1823–30. Eckardt KU, Pugh CW, Meier M, Tan CC, Ratcliffe PJ, Kurtz A Production of erythropoietin by liver cells in vivo and in vitro. Ann NY Acad. Sci 1994;718:50–60. discussion 61–3. Semenza GL, Traystman MD, Gearhart JD, Antonarakis SE. Polycythemia in transgenic mice expressing the human erythropoietin gene. Proc Natl Acad Sci USA 1989;86:2301–5. Semenza GL, Koury ST, Nejfelt MK, Gearhart JD, Antonarakis SE. Cell-type-specific and hypoxia-inducible expression of the human erythropoietin gene in transgenic mice. Proc Natl Acad Sci USA 1991;88:8725–9. Semenza GL, Dureza RC, Traystman MD, Gearhart JD, Antonarakis SE. Human erythropoietin gene expression in transgenic mice: multiple transcription initiation sites and cis-acting regulatory elements. Mol Cell Biol 1990;10:930–8. Madan A, Lin C, Hatch SL, II, Curtin PT. Regulated basal, inducible, and tissue-specific human erythropoietin gene expression in transgenic mice requires multiple cis DNA sequences. Blood 1995;85: 2735–41. Noguchi Y, Goto T, Yufu Y, Uike N, Hasegawa Y, Fukuda T, Jimi A, Funakoshi A. Gene expression of erythropoietin in renal cell carcinoma. Intern Med 1999;38:991–4. Da Silva JL, Lacombe C, Bruneval P, Casadevall N, Leporrier M, Camilleri JP, Bariety J, Tambourin P, Varet B. Tumor cells are the
24. 25.
26. 27. 28. 29. 30.
31. 32.
33.
34.
35. 36.
37. 38. 39.
40.
41.
42.
43.
44. 45.
site of erythropoietin synthesis in human renal cancers associated with polycythemia. Blood 1990;75:577–82. Clark D, Kersting R, Rojiani AM. Erythropoietin immunolocalization in renal cell carcinoma. Mod. Pathol 1998;11:24–8. Hagiwara M, Chen IL, McGonigle R, Beckman B, Kasten FH, Fisher JW. Erythropoietin production in a primary culture of human renal carcinoma cells maintained in nude mice. Blood 1984;63: 828–35. Katsuoka Y, McGonigle R, Rege AB, Beckman B, Fisher JW. Erythropoietin production in human renal carcinoma cells passaged in nude mice and in tissue culture. Gann 1983;74:534–41. Okabe T, Urabe A, Kato T, Chiba S, Takaku F. Production of erythropoietin-like activity by human renal and hepatic carcinomas in cell culture. Cancer 1985;55:1918–23. Sytkowski AJ, Richie JP, Bicknell KA. New human renal carcinoma cell line established from a patient with erythrocytosis. Cancer Res 1983;43:1415–9. Sherwood JB, Shouval D. Continuous production of erythropoietin by an established human renal carcinoma cell line: development of the cell line. Proc Natl Acad Sci USA 1986;83:165–9. Toyama K, Fujiyama N, Suzuki H, Chen TP, Tamaoki N, Ueyama Y. Erythropoietin levels in the course of a patient with erythropoietinproducing renal cell carcinoma and transplantation of this tumor in nude mice. Blood 1979;54:245–53. Shouval D, Anton M, Galun E, Sherwood JB. Erythropoietin-induced polycythemia in athymic mice following transplantation of a human renal carcinoma cell line. Cancer Res 1988;48:3430–4. Maxwell PH, Wiesener MS, Chang GW, Clifford SC, Vaux EC, Cockman ME, Wykoff CC, Pugh CW, Maher ER, Ratcliffe PJ. The tumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature 1999;399:271–5. Wiesener MS, Munchenhagen PM, Berger I, Morgan NV, Roigas J, Schwiertz A, Jurgensen JS, Gruber G, Maxwell PH, Loning SA, Frei U, Maher ER, et al. Constitutive activation of hypoxia-inducible genes related to overexpression of hypoxia-inducible factor-1a in clear cell renal carcinomas. Cancer Res 2001;61:5215–22. Turner KJ, Moore JW, Jones A, Taylor CF, Cuthbert-Heavens D, Han C, Leek RD, Gatter KC, Maxwell PH, Ratcliffe PJ, Cranston D, Harris AL. Expression of hypoxia-inducible factors in human renal cancer: relationship to angiogenesis and to the von Hippel-Lindau gene mutation. Cancer Res 2002;62:2957–61. Prchal JT Pathogenetic mechanisms of polycythemia vera and congenital polycythemic disorders. Semin Hematol 2001;38(Suppl 2):10–20. Randi ML, Murgia A, Putti MC, Martella M, Casarin A, Opocher G, Fabris F. Low frequency of VHL gene mutations in young individuals with polycythemia and high serum erythropoietin. Haematologica 2005;90:689–91. Pastore YD, Jelinek J, Ang S, Guan Y, Liu E, Jedlickova K, Krishnamurti L, Prchal JT. Mutations in the VHL gene in sporadic apparently congenital polycythemia. Blood 2003;101:1591–5. Pastore Y, Jedlickova K, Guan Y, Liu E, Fahner J, Hasle H, Prchal JF, Prchal JT. Mutations of von Hippel-Lindau tumor-suppressor gene and congenital polycythemia. Am J Hum Genet 2003;73:412–9. Bento MC, Chang KT, Guan Y, Liu E, Caldas G, Gatti RA, Prchal JT. Congenital polycythemia with homozygous and heterozygous mutations of von Hippel-Lindau gene: five new Caucasian patients. Haematologica 2005;90:128–9. Cario H, Schwarz K, Jorch N, Kyank U, Petrides PE, Schneider DT, Uhle R, Debatin KM, Kohne E. Mutations in the von Hippel-Lindau (VHL) tumor suppressor gene and VHL-haplotype analysis in patients with presumable congenital erythrocytosis. Haematologica 2005;90: 19–24. Ang SO, Chen H, Hirota K, Gordeuk VR, Jelinek J, Guan Y, Liu E, Sergueeva AI, Miasnikova GY, Mole D, Maxwell PH, Stockton DW, et al. Disruption of oxygen homeostasis underlies congenital Chuvash polycythemia. Nat Genet 2002;32:614–21. Lee YS, Vortmeyer AO, Lubensky IA, Vogel TW, Ikejiri B, Ferlicot S, Benoit G, Giraud S, Oldfield EH, Linehan WM, Teh BT, Richard S, et al. Coexpression of erythropoietin and erythropoietin receptor in von Hippel-Lindau disease-associated renal cysts and renal cell carcinoma. Clin Cancer Res 2005;11:1059–64. Vogel TW, Vortmeyer AO, Lubensky IA, Lee YS, Furuta M, Ikejiri B, Kim HJ, Lonser RR, Oldfield EH, Zhuang Z. Coexpression of erythropoietin and its receptor in endolymphatic sac tumors. J. Neurosurg 2005;103:284–8. Gong K, Zhang N, Zhang Z, Na Y. Coexpression of erythopoietin and erythopoietin receptor in sporadic clear cell renal cell carcinoma. Cancer Biol Ther 2006;5:582–5. Hardee ME, Arcasoy MO, Blackwell KL, Kirkpatrick JP, Dewhirst MW. Erythropoietin biology in cancer. Clin Cancer Res 2006;12: 332–9.
2442
WIESENER ET AL.
46. Vaupel P, Mayer A, Briest S, Hockel M. Hypoxia in breast cancer: role of blood flow, oxygen diffusion distances, and anemia in the development of oxygen depletion. Adv Exp Med Biol 2005;566:333–42. 47. Henke M, Laszig R, Rube C, Schafer U, Haase KD, Schilcher B, Mose S, Beer KT, Burger U, Dougherty C, Frommhold H. Erythropoietin to treat head and neck cancer patients with anaemia undergoing radiotherapy: randomised, double-blind, placebo-controlled trial. Lancet 2003;362:1255–60. 48. Leyland-Jones B, Semiglazov V, Pawlicki M, Pienkowski T, Tjulandin S, Manikhas G, Makhson A, Roth A, Dodwell D, Baselga J, Biakhov M, Valuckas K, et al. Maintaining normal hemoglobin levels with epoetin alfa in mainly nonanemic patients with metastatic breast cancer receiving first-line chemotherapy: a survival study. J Clin Oncol 2005;23:5960–72. 49. Kovacs G, Akhtar M, Beckwith BJ, Bugert P, Cooper CS, Delahunt B, Eble JN, Fleming S, Ljungberg B, Medeiros LJ, Moch H, Reuter VE, et al. The Heidelberg classification of renal cell tumours. J Pathol 1997;183:131–3. 50. Skinnider BF, Folpe AL, Hennigar RA, Lim SD, Cohen C, Tamboli P, Young A, de Peralta-Venturina M, Amin MB. Distribution of cytokeratins and vimentin in adult renal neoplasms and normal renal tissue: potential utility of a cytokeratin antibody panel in the differential diagnosis of renal tumors. Am J Surg Pathol 2005;29:747–54. 51. Chu PG, Weiss LM. Keratin expression in human tissues and neoplasms. Histopathology 2002;40:403–39. 52. Maxwell PH, Dachs GU, Gleadle JM, Nicholls LG, Harris AL, Stratford IJ, Hankinson O, Pugh CW, Ratcliffe PJ. Hypoxia-inducible factor-1 modulates gene expression in solid tumors and influences both angiogenesis and tumor growth. Proc Natl Acad Sci USA 1997;94: 8104–9. 53. Maynard MA, Ohh M. Von Hippel-Lindau tumor suppressor protein and hypoxia-inducible factor in kidney cancer. Am J Nephrol 2004; 24:1–13. 54. Maranchie JK, Vasselli JR, Riss J, Bonifacino JS, Linehan WM, Klausner RD. The contribution of VHL substrate binding and HIF1-a to the phenotype of VHL loss in renal cell carcinoma. Cancer Cell 2002;1:247–55. 55. Kondo K, Klco J, Nakamura E, Lechpammer M, Kaelin WG, Jr. Inhibition of HIF is necessary for tumor suppression by the von HippelLindau protein. Cancer Cell 2002;1:237–46. 56. Raval RR, Lau KW, Tran MG, Sowter HM, Mandriota SJ, Li JL, Pugh CW, Maxwell PH, Harris AL, Ratcliffe PJ. Contrasting properties of hypoxia-inducible factor 1 (HIF-1) and HIF-2 in von Hippel-
57. 58. 59. 60.
61.
62. 63.
64.
65. 66. 67. 68. 69.
Lindau-associated renal cell carcinoma. Mol Cell Biol 2005;25: 5675–86. Yan Q, Bartz S, Mao M, Li L, Kaelin WG, Jr. The HIF2{a} N-terminal and C-terminal transactivation domains cooperate to promote renal tumorigenesis in vivo. Mol Cell Biol 2007;27:2092–102. Motzer RJ, Bander NH, Nanus DM. Renal-cell carcinoma. N Engl J Med 1996;335:865–75. Acs G, Acs P, Beckwith SM, Pitts RL, Clements E, Wong K, Verma A. Erythropoietin and erythropoietin receptor expression in human cancer. Cancer Res 2001;61:3561–5. Dagnon K, Pacary E, Commo F, Antoine M, Bernaudin M, Bernaudin JF, Callard P. Expression of erythropoietin and erythropoietin receptor in non-small cell lung carcinomas. Clin Cancer Res 2005;11: 993–9. Vogel TW, Brouwers FM, Lubensky IA, Vortmeyer AO, Weil RJ, Walther MM, Oldfield EH, Linehan WM, Pacak K, Zhuang Z. Differential expression of erythropoietin and its receptor in von hippel-lindau-associated and multiple endocrine neoplasia type 2-associated pheochromocytomas. J Clin Endocrinol Metab 2005;90:3747–51. Jelkmann W. Erythropoietin: structure, control of production, and function. Physiol Rev 1992;72:449–89. Wiesener MS, Seyfarth M, Warnecke C, Jurgensen JS, Rosenberger C, Morgan NV, Maher ER, Frei U, Eckardt KU. Paraneoplastic erythrocytosis associated with an inactivating point mutation of the von Hippel-Lindau gene in a renal cell carcinoma. Blood 2002;99:3562–5. Suzuki M, Takamizawa S, Nomaguchi K, Suzu S, Yamada M, Igarashi T, Sato I. Erythropoietin synthesis by tumour tissues in a patient with uterine myoma and erythrocytosis. Br J Haematol 2001;113: 49–51. Matsuo M, Koga S, Kanetake H, Ohtani H, Higami Y, Shimokawa I, Watanabe J EPO-producing gastric carcinoma in a hemodialysis patient. Am J Kidney Dis 2003;42:E3–E4. Kaito K, Otsubo H, Usui N, Kobayashi M. Secondary polycythemia as a paraneoplastic syndrome of testicular seminoma. Ann Hematol 2004;83:55–7. Cheng TI, Tsou MH, Yang PS, Sung SM, Chuang VP, Sung JL. Dermatomyositis and erythrocytosis associated with hepatocellular carcinoma. J Gastroenterol Hepatol 2002;17:1239–40. Wang LY, Shih LY, Chen SH, Liu HC, Chai IJ, Liang DC. Neuroblastoma with expression of erythropoietin resulting in erythrocytosis. J Pediatr Hematol Oncol 2003;25:649–50. Westenfelder C. Unexpected renal actions of erythropoietin. Exp Nephrol 2002;10:294–8.
