Amplification and Expression of Splice Variants of the Gene Encoding the P450 Cytochrome 25-Hydroxyvitamin D 3 1,Α -Hydroxylase (CYP 27B1) in Human Malignant Glioma

868 Vol. 7, 868–875, April 2001 Clinical Cancer Research

Amplification and Expression of Splice Variants of the Gene

Encoding the P450 Cytochrome 25-Hydroxyvitamin D3 1,␣-Hydroxylase (CYP 27B1) in Human Malignant Glioma

Ruth Maria Maas, Katrin Reus, Britta Diesel, forme. The endogenous expression of the 25-hydroxyvitamin ␣ Wolf-Ingo Steudel, Wolfgang Feiden, D3 1, -hydroxylase gene and the appearance of alternative Ulrike Fischer, and Eckart Meese1 splice variants reveal a new feature of the molecular pathogenesis of glioblastoma and may represent a new target for Institut fu¨r Humangenetik, Theoretische Medizin [R. M. M., K. R., glioma therapy. B. D., U. F., E. M.], Neurochirurgische Universita¨tsklinik [W-I. S.], and Abteilung fu¨r Neuropathologie [W. F.], Universita¨t des Saarlandes, 66421 Homburg/Saar, Germany INTRODUCTION GBM2 is the most malignant and frequent intracranial tumor (1). Despite intensive effort in surgery, radiotherapy, and ABSTRACT chemotherapy, only minimal therapeutic advances toward a Purpose: Recently, we reported the isolation of six novel long-term survival for glioblastoma patients have been achieved genes termed glioma-amplified sequences (GASs) from the in recent years. Many genetic alterations such as amplifications, glioblastoma cell line TX3868 using microdissected medi- deletions, and translocations are known to occur within the cells ated cDNA capture (U. Fischer et al., Hum. Mol. Genet., 5: during tumor progression. Cytogenetic studies demonstrated 595–600, 1996). The aim of this study was to further char- that the most frequent aberrations in GBM are losses of chro- acterize the gene GAS89. mosome 10, 19, and 22, as well as gains on chromosome 7 Experimental Design: To determine the amplification attributable to the amplification of the EGFR gene in the chro- frequency, we performed comparative PCR studies and mosomal region 7p11–12 (2). EGFR is amplified in ϳ40% of Southern blot hybridization experiments. To identify full- glioblastomas resulting in increased expression of the corre- length clones of GAS89 we screened a HybriZAP library. sponding protein (3). Another, but less frequently amplified Reverse transcription-PCR was performed to isolate splice region, is chromosome 12q13–14 with an incidence of ϳ15% of variants and to determine expression levels. malignant gliomas (4). Recently, we were able to identify sev- Results: We identified for the gene GAS89 an amplifi- eral cDNAs from a homogeneously staining region at chromo- cation frequency of 25% in 28 examined glioblastoma mul- some 12q13–15 using microdissection-mediated cDNA capture tiforme samples. Screening a HybriZAP library, we isolated (5). One of these genes, GAS41, could be identified as a poten- an incomplete gene sequence showing identity with the gene tial transcription factor and was amplified in 23% of GBM (6). for 25-hydroxyvitamin D 1,␣-hydroxylase. Different full- 3 In this study, we isolated and characterized the correspond- length clones were then isolated using PCR primers chosen ing cDNA of another microdissected cDNA fragment (GAS89). ؅ ؅ from the 3 - and 5 -untranslated regions. As determined by Sequence comparison shows identity of the GAS89 cDNA with sequencing, the clones represent various splice variants of the gene encoding 25-hydroxyvitamin D 1,␣-hydroxylase the 25-hydroxyvitamin D 1,␣-hydroxylase gene. The clones 3 3 (CYP27B1), that catalyzes a key step in vitamin D metabolism. encode truncated proteins but also one potentially functional 3 Precursors of vitamin D are metabolized in the skin under the enzyme variant. Reverse transcription-PCR studies revealed 3 influence of UV-light and are then hydroxylated in the liver to overexpression of several variants in glioblastoma samples 25-hydroxyvitamin D . The further hydroxylation to the active with GAS89 amplification in comparison with normal brain 3 form of vitamin D , calcitriol or 1,25(OH) D , is performed by RNA and glioblastoma without GAS89 amplification. 3 2 3 the enzyme 25-hydroxyvitamin D 1,␣-hydroxylase in the kid- Conclusions: This is the first report of gene amplifica- 3 ney. The cDNA for this enzyme was cloned from rat (7) and tion for 25-hydroxyvitamin D 1,␣-hydroxylase and the ap- 3 human (8, 9) material. Using fluorescence in situ hybridization pearance of mRNA splice variants in glioblastoma multi- analysis, the gene was localized to chromosome 12q13–14, and the genomic DNA fragment was organized in 9 exons and 8 introns (10). The biological active form of vitamin D , calcitriol, is a Received 11/27/00; revised 1/23/01; accepted 1/23/01. 3 The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 To whom requests for reprints should be addressed, at Department of 2 The abbreviations used are: GBM, glioblastoma multiforme; Human Genetics, Building. 60, Medical School, University of Saarland, GAS, glioma-amplified sequence; RT-PCR, reverse transcription- ␣ 66421 Homburg/Saar, Germany. Phone: 49-6841-166038; Fax: 49- PCR; 1,25(OH)2D3, 1,25-dihydroxyvitamin D3;1, -hydroxylase, 25- ␣ 6841-166186; E-mail: [email protected]. hydroxyvitamin D3 1, -hydroxylase.

secosteroid hormone best known for its role in calcium and bone phocytes was isolated according to standard protocols (13). To metabolism. In addition, calcitriol provides antiproliferative and screen E. coli transformants, plasmid DNA from bacterial cul- differentiating effects through binding to the vitamin D receptor tures was isolated using alkaline lysis without a further purifi- belonging to the steroid/thyroid/retinoic acid receptor family cation step (14). A column purification was added for plasmids that functions as a ligand-dependent transcription factor (11). that were sequenced (Plasmid Mini kit; QIAgen, Hilden, Ger- Herein we report the endogenous expression of the 25- many). ␣ hydroxyvitamin D3 1, -hydroxylase in human malignant gli- Comparative PCR. Comparative PCR was performed as oma that potentially could result in the production of biologi- described previously (15). The aim of this method is to deter- cally active calcitriol. mine the number of copies of a specific gene in tumor DNA samples versus normal blood DNA. Concentration of DNA samples were determined by absorbance measurement and ad- MATERIALS AND METHODS justed to 5 ng/␮l. Three dilutions from each test sample (0.5, cDNA Library Construction and Screening. cDNA 1.0, and 2.0 ng/␮l) were used as templates in different reactions was synthesized from poly(A) mRNA of the glioblastoma cell to obtain the same concentrations in normal and tumor DNA. line TX3868 using oligo (dT) primers. Second-strand synthesis Primers for the gene Mucin 2 (MUC2A: 5Ј-CATTCTCAAC- was performed according to the manufacturer’s instructions GACAACCCCT-3Ј; MUC2Z: 5Ј-GCAAGAGAT-GTTACT- (Stratagene). After EcoRI adapter ligation and digestion with GCC-3Ј), which maps at 11p15.5 and for clone 9-1 plus (9-1 XhoI, the cDNA fragments were size fractionated with Sepha- forward: 5Ј-GCTTTATTGCCCTTGTCTATG-3Ј; 9-1 reverse: Ͼ rose CL-2B columns, and cDNAs 0.6 kb in length were 5Ј-TCCTATCTGCTT-CTACCATCC-3Ј) located at 14q were ligated into the HybriZAP vector. Packaging was done using the used for calibration purposes. Both genes represent single copy Gigapack III Gold packaging extract from Stratagene. The li- genes. When each dilution of the tumor and blood DNA was brary was amplified once prior to cDNA library screening. adjusted to the same concentration, a final PCR with the Transfection of the cDNA library into Escherichia coli- GAS89-specific primers H89 forward and H89 reverse was XL1-blue MRFЈ was performed using standard protocols. performed to detect amplification of the GAS89 gene. The PCR The corresponding phage clones were transferred onto nylon reactions were carried out in a thermal cycler (PTC100; MJ membranes (Boehringer Mannheim, Mannheim, Germany). Research) with the following cycling conditions for the MUC Plaque hybridization was done using the digoxygenin-labeled primers: 26 cycles of 1 min denaturation at 94°C, 45 s annealing insert of plasmid pGAS89 according to the manufacturer’s protocol (DIG Nonradioactive Labeling and Detection kit, at 60°C, 45 s extension at 72°C. The PCR for the 9-1 primers Boehringer Mannheim). was run for 26 cycles at 94°C for 45 s, 58°C for 45 s, and 72°C Ј Sequence Analysis. Sequencing was performed accord- for 45 s. The GAS89 specific primers H89 forward (5 -TGCAG- Ј Ј ing to the manufacturer’s instructions using the Amersham CATCAATGAACACTAT-3 ) and H89 reverse (5 -GGCCCT- Ј Thermo Sequenase-labeled primer cycle sequencing kit with TCTGATCATGTATGC-3 ) were deduced from the originally 7-deaza-dGTP. Plasmid inserts were sequenced with an auto- isolated 350-bp fragment of pGAS89 that was localized in this Ј mated sequencer (Licor 4000 L; MWG Biotech, Ebersberg, study to the 3 -untranslated region of the gene by sequence Germany). The obtained sequences were edited using the comparison. PCR conditions for the H89 primer were: 27 cycles SEQUENCHER 3.0 program (GeneCodes; Ann Arbor, MI). of 1 min at 94°C, 45 s at 56°C, and 45 s at 72°C. Homology search was done with the BLASTN and BLASTX RT-PCR. Prior to utilization, the RNA that served as a algorithms (12). template for RT-PCR was DNaseI treated. The existence of Tumors and Tissues. Tumor samples were kindly pro- genomic DNA contaminants was then excluded by Alu-PCR. vided by the Department of Neurosurgery, Homburg, Univer- First-strand cDNA was synthezised with the PROSTAR First sita¨t des Saarlandes (Homburg, Germany). The samples had strand RT-PCR kit (Stratagene) using 5 ␮g of total RNA, been snap-frozen in liquid nitrogen immediately after surgical oligo (dT)-Primer and Moloney murine leukemia virus- excision and stored at Ϫ75°C until usage. Cell line TX3868 was reverse transcriptase in a volume of 50 ␮l and stored at established and cultured as described previously (5). All of the Ϫ20°C. Five ␮l of the first-strand cDNA were used in the tumor samples originated from grade IV glioblastoma, as deter- following PCR. To obtain the entire cDNA fragments of the mined by the Department of Neuropathology. Normal brain 1,␣-hydroxylase from TX3868 cDNA, the primer 5Ј-end RNA was obtained from Clontech, Palo Alto, California. Nor- (5Ј-TATGATGCTCAGGAGAAGCG-3Ј) was chosen from mal kidney RNA was a gift from Dr. H-P. Sattler, Department the 5Ј-untranslated region of the known cDNA. Primer H89 of Human Genetics. forward binds to the antisense strand in the 3Ј-untranslated DNA and RNA Isolation. RNA isolation was carried out region and was, therefore, used as the 3Ј-primer for this according to the manufacturer’s instructions (Stratagene). In purpose. PCR was carried out with the Expand High Fidelity brief, frozen tissue was homogenized, proteins and DNA were PCR system (Boehringer Mannheim) for 33 cycles as fol- extracted using phenol/chloroform (pH 5.3–5.7), and the re- lows: 94°C for 1 min, 56°C for 45 s, and 72°C for 1 min and maining RNA was precipitated with isopropanol. The RNA was 30 s. PCR products were excised from the gel and extracted quantified spectrophotometrically, and its integrity was con- with the QIAquick Gel Extraction kit (QIAgen). Cloning into trolled by agarose gel electrophoresis in 3-(N-morpholino) pro- the pGEM-T Easy Vector System (Promega, Mannheim, Ger- pane sulfonic acid buffer. many) was performed according to the manufacturer’s in- Genomic DNA from frozen tissue samples and blood lym- structions. To analyze the 1,␣-hydroxylase splice variants,

primers 5Ј-gap (5Ј-TGAACAACGTAGTCTGCGACCTT-3Ј) and 5, and insertion of intron 2 at the 5Ј-end and was, therefore, and 3Ј-gap (5Ј-TGACACAGAGTGACCAGCGTATTT-3Ј) termed hydroxylase-variant 1 (Hyd-V1). Isolation of this clone were chosen from exon 3 and exon 8. After 34 cycles (1 min was the first evidence for aberrantly spliced variants of the gene at 94°C, 1 min at 60°C, 1 min 72°C) of amplification with for 1,␣-hydroxylase in the TX3868 glioblastoma cell line. Taq-DNA Polymerase (Pharmacia, Freiburg, Germany), the Amplification Analysis for the 1,␣-Hydroxylase Gene in resulting bands were cloned using the pMOSBlue blunt- Glioblastoma Samples. To determine the frequency of gene ended cloning kit (Amersham, Braunschweig, Germany). amplification, tumor samples were analyzed using comparative Southern Blot Analysis. PCR products were separated PCR with pGAS89-specific primers. In 7 of 28 tested GBM on 1%-agarose gels and then transferred to a Gene-Screen samples (25%), an amplification of the 1,␣-hydroxylase gene Nylon membrane in 20ϫ SSC. Labeling and detection of the was detected compared with peripheral blood DNA. One rep- probe was performed with the DIG Non-Radioactive Labeling resentative example of a strong amplification (H549) and three and Detection kit (Boehringer Mannheim) as recommended by samples with weak GAS89 amplifications (H556, H346, and the manufacturer. The membrane was hybridized with the H246) are shown in Fig. 1A. Additionally to comparative PCR, 1.6-kb insert from clone pGAS89-b in 5ϫ SSC, 1% blocking amplification analysis was performed by Southern blot hybrid- reagent, 0.1% lauroylsarcosin, and 0.02% SDS overnight at ization using the insert of pGAS89-b as the probe. Only tumors 42°C. Washing was done twice for 5 min in 2ϫ SSC/0.1% SDS with high amplification levels (H91, H118, and H385) revealed at room temperature and twice for 15 min in 0.1ϫ SSC/0.1% by comparative PCR show strong signals (7.65 kb) in the SDS at 55°C. hybridization experiments as shown in Fig. 1B. DNA (5 ␮g) from peripheral blood lymphocytes and from Identification and Cloning of Splice Variants of the tumor samples was digested with EcoRI, separated on a 0.8% 1,␣-Hydroxylase Gene. To obtain the entire cDNA from the agarose gel and blotted onto a nylon membrane (GeneScreen). glioblastoma cell line, we used primers deduced from the 5Ј- Ј Southern hybridization was carried out in a 500-mM phosphate coding region and the 3 -noncoding region of the known cDNA buffer (pH 7.2) as described previously (6). Insert DNA (50 ng) for 1,␣-hydroxylase. RT-PCR resulted in two major products was labeled with [␣-32P]dATP using the Random-primed DNA that were excised from the gel and subcloned into pGEM-Teasy. labeling kit (Boehringer Mannheim). Restriction analysis of the resulting transformants revealed three different inserts. Sequence analysis of the subcloned cDNAs and comparison with the known sequence for 1,␣-hydroxylase indi- RESULTS cated that these clones represented additional splice variants and pGAS89-b Represents a Splice Variant of the Human were named hydroxylase-variant 2, 3, and 4 (Hyd-V2, -V3, and ␣ Gene for 25-Hydroxyvitamin D3 1, -Hydroxylase. Plasmid -V4; Fig. 2). Variant Hyd-V2 lacks exons 4 and 5 resulting in a clone pGAS89 with an insert of 350 bp was previously isolated truncated protein of 207 amino acids in contrast to 508 amino from a homogeneously staining region of the glioblastoma cell acids encoded by the full-length 1,␣-hydroxylase gene. This is line TX3868 and subsequently localized on chromosome attributable to a termination codon that occurs after the first 29 12q13–15 (5). To obtain the full-length cDNA, we screened a bp in Exon 6 and is caused by a frameshift. The resulting protein HybriZAP (Stratagene) cDNA library established from glioblas- contains neither the ferredoxin nor the haem-binding site of the toma cell line TX3868 using the insert of plasmid pGAS89 as P450 protein and, therefore, probably represents an inactive the probe. From 50,000 plaque-forming units that were exam- variant of the enzyme. ined, one positive phage clone was identified. The clone was The second variant, Hyd-V3, contains only a part of exon reconfirmed by PCR with pGAS89 specific primers and then in 8. Exon 8 contains an additional alternative 5Ј-splice site, AG, vivo excised into a plasmid clone pGAS89-b with an insert at the appropriate position obviously recognized in variant Hyd- length of 1.6 kb. Sequence analysis and homology search using V3. The partial deletion leads to a frameshift in the remaining the BLASTX and BLASTN algorithms revealed identity with part of exon 8, which results in a premature termination signal. parts of the sequence for the human cDNA for 25-hydroxyvi- The encoded protein consists of 414 amino acids and contains ␣ ␣ tamin D3 1, -hydroxylase, which was recently localized on the ferredoxin but lacks the haem-binding site of the 1, - chromosome 12q13–15 by somatic cell hybrid mapping (8, 9, hydroxylase. It is unlikely that this variant could produce an ␣ 10). The 25-hydroxyvitamin D3 1 -hydroxylase, a mitochon- active enzyme. drial cytochrome P450 enzyme, catalyzes the hydroxylation of Variant Hyd-V4 is complete with regard to the exons of the Ј 25-hydroxyvitamin D3 to 1,25(OH)2D3, the active form of known cDNA but additionally contains the 3 -region of intron 2 ␣ Ј vitamin D3. The 1, -hydroxylase gene is located in the ampli- using an alternative 5 -splice signal-AG in intron 2. Translation fication unit 12q13–14 between MDM2 and CDK4/SAS accord- of the Hyd-V4 cDNA would result in an in-frame insertion of 27 ing to the National Center for Biotechnology Information Map- amino acids to generate a protein product of 535 amino acids View program. containing both the ferredoxin- and the haem-binding site. This The insert of the primary isolated clone pGAS89 was protein product may represent an active variant with regard to identical with a sequence in the 3Ј-untranslated region of the the 1,␣-hydroxylase specific properties. 1,␣-hydroxylase gene. The 5Ј-region of the isolated clone Expression of Splice Variants in Several Normal pGAS89-b was incomplete probably resulting from an incom- Tissues and Glioblastoma. To examine the correlation be- plete reverse transcription reaction during the cDNA library tween amplification and the appearance of splice variants of the construction. In contrast to the known sequence for the 1,␣- 1,␣-hydroxylase gene, we performed RT-PCR with RNA from hydroxylase cDNA, pGAS89-b exhibited deletions of exons 4 normal tissues and tumor samples. We used five glioblastoma

Fig. 1 Amplification analysis of GAS89 in GBM. A, comparative PCR for tumors H549, H546, H485, H545, H556, H346, H412, H246, and H366. Three different concentrations of tumor and peripheral blood (pb) DNA (0.5, 1, and 2 ng) were amplified by PCR using primers specific for GAS89 (lower panel) and Muc2 or 9-1 plus as a single-copy control (upper panel). Ar- rows, PCR product sizes. GAS89 shows a strong amplification in tumor H549 and is weakly amplified in tumors H556, H346, and H246 in comparison with peripheral blood DNA. The “control” sample contained no template DNA. B, Southern blot hybridization using the insert of pGAS89-b as the probe and gel loading control. GAS89 was found to be amplified in tumors H91, H118, and H385 compared with peripheral blood DNA, whereas tumors H77 and H30 show no amplification.

samples with amplification and nine glioblastoma samples with- pGAS89-b as the probe to confirm specificity (Fig. 3). The out amplification of the 1,␣-hydroxylase gene as determined by strongest signal in kidney was that of 763 attributed to the comparative PCR. Normal tissues were normal brain and kidney normal cDNA containing both exons 4 and 5, as expected. tissues in which the 1␣-hydroxylase is primarily active. Because Variants Hyd-V2 and Hyd-V1 that probably encode truncated attempts to amplify the entire cDNA only result in weak bands, proteins were also expressed but at a lower level. This propor- we used primers chosen from exons 3 and 8 restricting our tion was reverted in normal brain in which the expression level contemplation at the lack of exon 4 and 5. The variants sum- of variants Hyd-V2/V1 was stronger than the expression of the marized in Fig. 2B should generate two products (763 bp and normal 1,␣-hydroxylase cDNA or variant Hyd-V4, respectively. 389 bp) that cannot be ascribed to single templates. Both Similar expression patterns were detected in tumor samples Hyd-V1 and Hyd-V2 can produce the signal of 389 bp, whereas without amplification of GAS89. In tumors carrying 1,␣- the product of 763 bp can be generated by both Hyd-4 and the hydroxylase gene amplifications and in TX3868, the glioblas- normal cDNA of the1,␣-hydroxylase gene. Variant Hyd-V3 toma cell line from which GAS89 was isolated, the expression cannot be amplified with the primers used in this study because level of the normal 1,␣-hydroxylase cDNA/Hyd-V4 was higher of the deletion of the binding site of the reverse primer at the than in normal brain. The cDNA fragment of 763 bp was barely beginning of exon 8. visible in the normal brain sample. An additional product of Four representative tumor samples, cell line TX3868, and about 1000 bp appeared in most tumors with amplification and normal tissues were blotted and hybridized with the insert of in TX3868. Cloning and sequencing of this fragment revealed a

Fig. 2 A, sequence analysis of variants Hyd-V2, -V3, and -V4 and comparison with the known sequence of 1␣-hydroxylase. Arrows, the exons and intron 2. B, structures of the normal and the alternatively spliced 1,␣-hydroxylase mRNAs. Variant Hyd-V1 corresponds to the insert of pGAS89-b and is incomplete at the 5Ј-end. Heavy bars (in the normal 1,␣-hydroxylase mRNA), the coding regions for the ferredoxin-binding site in exon 6 and for the haem-binding site in exon 8. Arrows, premature stop codons.

Fig. 2. Continued

DISCUSSION In this study, we describe the isolation of the 25-hydroxyvi- ␣ tamin D3 1, -hydroxylase cDNA from an amplified region of chromosome 12q13–15 derived from the glioblastoma cell line TX3868. Chromosome 12q13–15 is frequently amplified in human sarcomas and brain tumors and several genes of potential significance for neoplastic growth are located in this genomic segment such as Gli, SAS, CDK4, and MDM2 (16). Amplifica- tion mapping of chromosome 12q13–15 in malignant glioma reveals the highest amplification level (about 15%) for the two loci CDK4/SAS and MDM2 suggesting these three genes as likely targets of the amplification unit (4). A refined mapping of this region in 24 primary gliomas using seven known loci located between CDK4 and MDM2 shows no continuity in this region of the amplicon (17). However, it is still possible that all amplicons share another amplified region between MDM2 and CDK4/SAS which might contain one or more as yet unidentified genes. One of the candidate target genes on chromosome Fig. 3 Expression analysis of the 1,␣-hydroxylase gene in tumor sam- 12q13–15 is GAS41, probably encoding a transcription factor. ples and normal tissues (kidney and normal brain). A, RT-PCR stand- GAS41 is amplified in 23% of glioblastoma, in 50% of astro- ardization of DNaseI-treated tumor and normal tissue RNA with glyc- cytoma grade II and grade III, and even in 80% of analyzed eraldehyde-3-phosphate dehydrogenase specific primers. B, Southern blot hybridization of RT-PCR with the insert of pGAS89-b as the probe. pilocytic astrocytoma (6). The relatively high amplification fre- Primers were chosen from exon 3 and exon 8 of the 1,␣-hydroxylase quency (25%) detected in this study for the 1,␣-hydroxylase gene. H385 and H91 were glioblastomas with strong GAS89 amplifi- gene suggests that this gene may be a further likely target gene cation, TX3868 is the cell line from which GAS89 was primarily of the amplification unit 12q13–15. To our knowledge, this is isolated; H282 and H322 were glioblastomas without GAS89 amplification. Arrows, variants of the 1,␣-hydroxylase cDNA. the first report concerning the chromosomal alteration of the 1,␣-hydroxylase gene in malignant tissue. Another gene from vitamin D metabolism, the 24-hydroxylase, could recently be detected as a likely target oncogene of the amplification unit 20q13.2 in breast cancer tumors and cell new variant Hyd-V5 containing intron 5. This caused a frame- lines using array CGH (18). The 24-hydroxylase inactivates the shift resulting in a premature stop codon soon within exon 6. biologically active 1,25(OH) D by a further hydroxylation step Summarizing the RT-PCR results, we assume there is a corre- 2 3 so that overexpression attributable to amplification may abro- lation between the amplification of the 1,␣-hydroxylase gene in gate vitamin D3-mediated growth control. This known function GBM and the expression level of several splice variants and the supports its candidacy for being an oncogene although 24- ␣ normal cDNA of the 1, -hydroxylase gene. hydroxylase has not been previously implicated in breast cancer.

It is more difficult to define the role of the 1,␣-hydroxylase potentially useful agent in the therapy of human malignant

gene in the pathogenesis of malignant glioma because this glioma (27). However, the clinical use of 1,25(OH)2D3 is im- enzyme catalyzes the synthesis of biologically active vitamin paired by its potent hypercalcemic effects. One possibility of

D3. Amplification and overexpression of this gene should exert reducing hypercalcemia is the use of less calcemic synthetic antiproliferative and prodifferentiating effects according to the vitamin D3 analogues which have also been shown to induce in vitro studies. Analysis of numerous normal and cancer cell cell death of rat glioma cells (29). Pending further mechanistic Ϫ9 Ϫ7 lines demonstrates that at high concentrations (10 to 10 M) studies, our findings raise the possibility that the supplementa-

1,25(OH)2D3 inhibits the growth of tumor cells in vitro, and it tion of patients’ treatment with 25(OH)D3 may promote the has also been shown that 1,25(OH)2D3 has beneficial effects in local synthesis of 1,25(OH)2D3 by glioma cells without several in vivo models of various types of cancer (19, 20). If the the systemic side effects of hypercalcemia as discussed for the overexpression of the 1,␣-hydroxylase gene results in an over- chemopreventive therapy of invasive prostate cancer (30).

production of 1,25(OH)2D3, the accumulation of this metabolite should have beneficial effects as shown by the in vitro and in ACKNOWLEDGMENTS vivo studies. The question about the cellular and systemic con- sequences of the endogenous expression of the 1,␣-hydroxylase We thank Doris Hemmer for contribution to the amplification analysis and Daniela Scherer for technical assistance. We also thank in malignant tissues remains to be clarified. Brenda Glass for critically reading the manuscript. Furthermore, we were able to demonstrate a correlation between amplification of the gene and its overexpression in human malignant glioma. The main site of expression of the REFERENCES 1,␣-hydroxylase gene is in the kidney, and the major part of 1. Kleihues, P., Burger, P. C., Collins, V. P., Newcomb, E. W., Ohgaki, H., and Cavenee, W.K. Glioblastoma. In: P. Kleihues and W. K. circulating 1,25(OH)2D3 is made by the proximal tubular cells of the renal cortex (21). By using RT-PCR, it was possible to Cavenee (eds.), Tumors of the nervous system, pp. 29–39. Lyon: IARC Press, 2000. detect 1,␣-hydroxylase expression in keratinocytes, brain, and 2. Van de Kelft, E. Molecular pathogenesis of astrocytoma and glio- testis (8). There are also reports concerning extrarenal sites of blastoma multiforme. Acta. Neurochir. (Wien), 139: 589–599, 1997. ␣ 1, -hydroxylase activity, which could be the cause of hypercal- 3. Eley, G., Frederick, L., Wang, X. Y., Smith, D. I., and James, C. D. cemia associated with lymphoma and certain solid tumors (22, 3Ј end structure and rearrangements of EGFR in glioblastomas. Genes 23). Recently it was found that the extrarenal 1,␣-hydroxylase Chromosomes Cancer, 23: 248–254, 1998. activity in a human non-small cell lung carcinoma is attributable 4. Reifenberger, G., Reifenberger, J., Ichimura, K., Meltzer, P. S., and to the same gene as in the kidney (24). Collins, V. P. Amplification of multiple genes from chromosomal region 12q13–14 in human malignant gliomas: preliminary mapping of We reported the cloning and expression of several alterna- the amplicons shows preferential involvement of CDK4, SAS, and tively spliced 1,␣-hydroxylase cDNAs from the glioblastoma MDM2. Cancer Res., 54: 4299–4303, 1994. cell line TX3868, including transcripts encoding truncated pro- 5. Fischer, U., Gracia, U., Elkahloun, A., Trent, J. M., Meese, E., and teins. Differences in the expression levels of variants Hyd-V1, Meltzer, P. S. Isolation of genes amplified in human cancers by micro- Hyd-V2, Hyd-V4, and Hyd-V5 and the normal cDNA were dissection mediated cDNA capture. Hum. Mol. Genet., 5: 595–600, found in kidney, normal brain, and glioblastoma with and with- 1996. out amplification of the 1,␣-hydroxylase gene. Variant Hyd-V3 6. Fischer, U., Heckel, D., Michel, A., Janka, M., Hulsebos, T., and Meese, E. Cloning of a novel transcription factor-like gene amplified in could not be detected in this RT-PCR study because of the human glioma including astrocytoma grade I. Hum. Mol. Genet., 6: primer location. However it was obvious that the expression 1817–1822, 1997. pattern in glioma with amplification is different from that in 7. Shinki, T., Shimada, H., Wakino, S., Anazawa, H., Hayashi, M., normal brain, where the variant without exons 4 and 5 encoding Saruta, T., DeLuca, H. F., and Suda, T. Cloning and expression of rat a truncated protein is dominant. Alternative splicing is a fre- 25-hydroxyvitamin D3–1␣-hydroxylase cDNA. Proc. Natl. Acad. Sci. USA, 94: 12920–12925, 1997. quent feature in the expression of many P450 genes, is consid- 8. Fu, G. K., Lin, D., Zhang, M. Y., Bikle, D. D., Shackleton, C. H., ered as an important factor in regulating the enzyme level, and Miller, W. L., and Portale, A. A. Cloning of human 25-hydroxyvitamin may be the cause for tissue-specific variation (25). Until now, D-1 ␣-hydroxylase and mutations causing vitamin D-dependent rickets there has been no indication of the existence of alternatively type 1. Mol. Endocrinol., 11: 1961–1970, 1997. spliced mRNAs of the 1,␣-hydroxylase gene and their tissue- 9. Monkawa, T., Yoshida, T., Wakino, S., Shinki, T., Anazawa, H., specific preference. The effect of the expression of alternative Deluca, H. F., Suda, T., Hayashi, M., and Saruta, T. Molecular cloning transcripts on the 1,␣-hydroxylase activity level in glioma has to of cDNA and genomic DNA for human 25-hydroxyvitamin D3 1 ␣-hydroxylase. Biochem. Biophys. Res. Commun., 239: 527–533, 1997. be investigated in future experiments. ␣ 10. Fu, G. K., Portale, A. A., and Miller, W. L. Complete structure of The overexpression of the 1, -hydroxylase gene in glioma the human gene for the vitamin D 1␣-hydroxylase, P450c1␣. DNA Cell. could result in a higher level of 1,25(OH)2D3. A prerequisite for Biol., 16: 1499–1507, 1997. the antiproliferative action of 1,25(OH)2D3 is the availability of 11. Christakos, S., Raval-Pandya, M., Wernyj, R. P., and Yang, W. the vitamin D receptor. Previous works demonstrated the pres- Genomic mechanisms involved in the pleiotropic actions of 1,25-dihy- ence of the vitamin D receptor in normal brain, in which its droxyvitamin D3. Biochem. J., 316: 361–371, 1996. physiological function remains unknown, and in human anaplas- 12. Altschul, S. F., Gish, W., Miller, W., Myers, E.W., and Lipman, D. J. Basic local alignment search tool. J. Mol. Biol., 215: 403–410, tic astrocytoma and glioblastoma (26). Furthermore the known 1990. antiproliferative and prodifferentiating effects of 1,25(OH)2D3 13. Muellenbach, R., Lagoda, P. J., and Welter, C. An efficient salt- have been observed on rat and human glioblastoma cell lines chloroform extraction of DNA from blood and tissues. Trends Genet., 5: (27, 28). These results suggest that 1,25(OH)2D3 may be a 391, 1989.

14. Birnboim, H. C., and Doly, J. A rapid alkaline extraction procedure 23. Mawer, E. B., Hayes, M. E., Heys, S. E., Davies, M., White, A., for screening recombinant plasmid DNA. Nucleic Acids Res., 7: 1513– Stewart, M. F., and Smith, G. N. Constitutive synthesis of 1,25-dihy- 1523, 1979. droxyvitamin D3 by a human small cell lung cancer cell line. J. Clin. 15. Brass, N., Heckel, D., and Meese, E. Comparative PCR: an improved Endocrinol. Metab., 79: 554–560, 1994, 1994. method to detect gene amplification. Biotechniques, 24: 22–24, 1998. 24. Jones, G., Ramshaw, H., Zhang, A., Cook, R., Byford, V., White, J., 16. Elkahloun, A. G., Bittner, M., Hoskins, K., Gemmill, R., and and Petkovich, M. Expression and activity of vitamin D-metabolizing Meltzer, P. S. Molecular cytogenetic characterization and physical map- cytochrome P450s (CYP1␣ and CYP24) in human non-small cell lung ping of 12q13–15 amplification in human cancers. Genes Chromosomes carcinomas. Endocrinology, 140: 3303–3310, 1999. Cancer, 17: 205–214, 1996. 25. Ding, S., Lake, B. G., Friedberg, T., and Wolf, C. R. Expression and 17. Reifenberger, G., Ichimura, K., Reifenberger, J., Elkahloun, A. G., alternative splicing of the cytochrome P-450 CYP2A7. Biochem. J., Meltzer, P. S., and Collins, V. P. Refined mapping of 12q13–q15 306: 161–166, 1995. amplicons in human malignant gliomas suggests CDK4/SAS and 26. Magrassi, L., Bono, F., Milanesi, G., and Butti, G. Vitamin D MDM2 as independent amplification targets. Cancer Res., 56: 5141– receptor expression in human brain tumors. J. Neurosurg. Sci., 36: 5145, 1996. 27–30, 1992. 18. Albertson, D. G., Ylstra, B., Segraves, R., Collins, C., Dairkee, 27. Magrassi, L., Butti, G., Pezzotta, S., Infuso, L., and Milanesi, G. S. H., Kowbel, D., Kuo, W. L., Gray, J. W., and Pinkel, D. Quantitative Effects of vitamin D and retinoic acid on human glioblastoma cell lines. mapping of amplicon structure by array CGH identifies CYP24 as a Acta. Neurochir. (Wien), 133: 184–190, 1995. candidate oncogene. Nat. Genet., 25: 144–146, 2000. 28. Naveilhan, P., Berger, F., Haddad, K., Barbot, N., Benabid, A. L., 19. van Leeuwen, J. P. T. M., and Pols, H. A. P. Vitamin D: anticancer Brachet, P., and Wion, D. Induction of glioma cell death by 1,25(OH)2 and differentiation. In: D. Feldmann, F. H. Glorieux, and J. W. Pike vitamin D3: towards an endocrine therapy of brain tumors? J. Neurosci. (eds.), Vitamin D, pp. 1089–1099. London: Academic Press, 1997. Res., 37: 271–277, 1994. 20. Studzinski, G. P., and Moore, D. C. Sunlight—can it prevent as 29. Baudet, C., Chevalier, G., Naveilhan, P., Binderup, L., Brachet, P., well as cause cancer? Cancer Res., 55: 4014–4022, 1995. and Wion, D. Cytotoxic effects of 1 ␣,25-dihydroxyvitamin D3 and 21. Brunette, M. G., Chan, M., Ferriere, C., and Roberts, K. D. Site of synthetic vitamin D3 analogues on a glioma cell line. Cancer Lett., 100: 1,25(OH)2 vitamin D3 synthesis in the kidney. Nature (Lond.), 276: 3–10, 1996. 287–289, 1978. 30. Schwartz, G. G., Whitlatch, L. W., Chen, T. C., Lokeshwar, B. L., 22. Davies, M., Hayes, M. E., Yin, J. A., Berry, J. L., and Mawer, E. B. and Holick, M. F. Human prostate cells synthesize 1,25-dihydroxyvita- Abnormal synthesis of 1,25-dihydroxyvitamin D in patients with ma- min D3 from 25-hydroxyvitamin D3. Cancer Epidemiol. Biomark. lignant lymphoma. J. Clin. Endocrinol. Metab., 78: 1202–1207, 1994. Prev., 7: 391–395, 1998.

Downloaded from clincancerres.aacrjournals.org on September 29, 2021. © 2001 American Association for Cancer Research. Amplification and Expression of Splice Variants of the Gene Encoding the P450 Cytochrome 25-Hydroxyvitamin D 3 1,α -Hydroxylase (CYP 27B1) in Human Malignant Glioma

Ruth Maria Maas, Katrin Reus, Britta Diesel, et al.

Clin Cancer Res 2001;7:868-875.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/7/4/868

Cited articles This article cites 26 articles, 5 of which you can access for free at: http://clincancerres.aacrjournals.org/content/7/4/868.full#ref-list-1

Citing articles This article has been cited by 6 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/7/4/868.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/7/4/868. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.