OUP user menu

Microarray Analysis Reveals Differential Gene Expression Patterns in Tumors of the Pineal Region

Michelle Fèvre-Montange PhD, Jacques Champier PhD, Alexandru Szathmari MD, Anne Wierinckx PhD, Carmine Mottolese MD, Jacques Guyotat MD, PhD, Dominique Figarella-Branger MD, PhD, Anne Jouvet MD, PhD, Joël Lachuer PhD
DOI: http://dx.doi.org/10.1097/01.jnen.0000225907.90052.e3 675-684 First published online: 1 July 2006


Several types of tumors are known to originate from the pineal region, among them pineal parenchymal tumors (PPTs) and papillary tumors of the pineal region (PTPRs), probably derived from the subcommissural organ. As a result of their rarity, their histologic diagnosis remains difficult. To identify molecular markers, using CodeLink oligonucleotide arrays, gene expression was studied in 3 PPTs (2 pineocytomas and one pineoblastoma), 2 PTPRs, and one chordoid glioma, another rare tumor of the third ventricle. Because PTPR and chordoid glioma may present ependymal differentiation, gene expression was also analyzed in 4 ependymomas. The gene patterns of the 3 PPTs fell in the same cluster. The pineocytomas showed high expression of TPH, HIOMT, and genes related to phototransduction in the retina (OPN4, RGS16, and CRB3), whereas the pineoblastoma showed high expression of UBEC2, SOX4, TERT, TEP1, PRAME, CD24, POU4F2, and HOXD13. Using reverse transcriptase-polymerase chain reaction on 13 PPTs, we demonstrated that PRAME, CD24, POU4F2, and HOXD13 might be candidates for grading PPT with intermediate differentiation. PTPRs, classified with chordoid glioma and separately from ependymomas, showed high expression of SPEDF, KRT18, and genes encoding proteins reported to be expressed in the subcommissural organ, namely ZFH4, RFX3, TTR, and CGRP. Our results highlight the usefulness of gene expression profiling for classify tumors of the pineal region and identify genes with potential use as diagnostic markers.

Key Words
  • Ependymoma
  • Human pineal parenchymal tumors
  • Microarray
  • Papillary tumors of the pineal region


Several types of tumors are known to originate from the pineal region, including germ cell tumors, gliomas, and pineal parenchymal tumors (PPTs). PPTs, which are derived from the specialized neurosecretory elements of the pineal gland, are divided by the World Health Organization into well-differentiated pineocytoma (PC), poorly differentiated pineoblastoma (PB), and mixed PC-PB or PPT with intermediate differentiation (PPTint) (1, 2). Gliomas of the pineal region are derived from resident pineal glial cells or from other glial cells in the vicinity of the pineal gland. Among gliomas, a papillary tumor probably deriving from specialized ependymocytes of the subcommissural organ (SCO), a small structure close to the pineal gland, has been described in this region and named papillary tumor of the pineal region (PTPR) (3, 4). All these tumors demonstrate strikingly similar imaging features. As a result of their rarity, their histologic diagnosis and classification remain difficult. In the present study, we compared gene expression in PPT with that in PTPR and chordoid glioma (CG), another rare third ventricle tumor with ependymal differentiation (5-7). Based on ultrastructural similarities, the same histogenesis has been proposed for PTPRs and CGs, and these 2 neoplasms may originate from specialized ependymal cells in the SCO (6, 8). As a result of the special anatomic location of CGs, another histogenesis has also been proposed for this neoplasm, namely that it is derived from ependymocytes of the lamina terminalis (7) or third ventricle tanycytes (9). Furthermore, because both PTPR and CG may present ependymal differentiation, we also compared gene expression in these 2 types with that in intracerebral and spinal ependymoma (Ic EP and Sp EP); the latter, similar to PTPR and CG, is more frequently observed in adults. This approach of analyzing expression in various types of periventricular neoplasms might identify molecular markers capable of helping diagnosis and suggest the pathways implicated in malignant transformation.

Materials and Methods

Tumor Specimens

Tumor tissues (3 PC, 4 PB, 6 PPTint, 2 PTPR, one CG, 2 Ic EP, and 2 Sp EP) and 2 normal postmortem pineal glands were obtained from the Neurobiotec Bank (Hôpital Neurologique et Neurochirurgical Pierre Wertheimer, Lyon, France) and from the AP-HM Tissue Bank (Assistance Publique-Hôpitaux de Marseille, France). At surgery, the tumor tissue was divided into 2 fragments, one of which was frozen and stored in liquid nitrogen for RNA extraction, whereas the other was fixed for histopathologic analysis to determine the nature and grading of the tumor. Routine staining was performed using hemalin phloxine saffron.

RNA Extraction

Total RNA was extracted from the different samples using the RNA Plus (Qbiogen, Illkirch, France) procedure based on the method of Chomczynski and Sacchi (10) and then precipitated with ethanol. The quality of the isolated total RNA was evaluated on nanochips using the Agilent 2100 bioanalyzer (Agilent Technologies, Massy, France). RNA from a whole adult human brain (single donor, male, 72 years), provided in 0.1 mM EDTA, pH 8.0, was purchased from Stratagene (Stratagene Europe, Amsterdam, The Netherlands).

Microarray Analysis

This analysis was performed on 2 PC (PC1 and PC2), one PB (PB1), 2 PTPR, one CG, 2 Ic EP, 2 Sp EP, and on normal total brain RNA.

RNA Amplification

Total RNA (2 μg) was amplified and labeled by a round of in vitro transcription using a MessageAmp aRNA kit (Ambion, Cambridgeshire, U.K.) following the manufacturer's protocol. Before amplification, spikes of different concentrations of synthetic mRNA were added to all tubes. These positive controls were used to verify the quality of the process. Amplified RNA (aRNA) was measured with an ultraviolet spectrophotometer and the quality verified on picochips using the Agilent bioanalyzer.

Array Hybridization and Processing

Biotin-labeled aRNA (10 μg) was fragmented with 5 μL of fragmentation buffer (GE Healthcare, Amersham, Saclay, France) in a final volume of 20 μL and the fragmented aRNA added to Amersham hybridization solution (Amersham) (final volume 260 μL) and injected onto CodeLink Uniset human 20K bioarrays containing 19,000 human oligonucleotide gene probes (Amersham). The arrays were hybridized overnight at 37°C at 300 rpm on a rotary mixer in an incubator, washed at 46°C for 1 hour in stringent TNT buffer (100 mM Tris, 150 mM NaCl, 0.02% Tween 20; all from Sigma-Aldrich, Saint Quentin-Fallavier, France), incubated in 3.4 mL of streptavidin-Cy5 (Amersham) solution for 30 minutes, washed 4 times in 240 mL of TNT buffer, rinsed twice in 240 mL of water containing 0.2% Triton X-100, and dried by centrifugation at 600 rpm.

The arrays were scanned with a Genepix 4000B scanner (Axon Instruments, Dipsi Industrie, Chatillon, France) using Genepix software with the laser set at 635 nm, the power at 100%, and the photomultiplier tube voltage at 60%. The scanned image files were analyzed using CodeLink expression software version 4.0, which produces both raw and normalized hybridization signals for each spot on the array.

Microarray Data Analysis

The CodeLink software normalizes the overall raw hybridization signal intensity on each array to the median of the array (median intensity is one after normalization). This study used the normalized signal intensities. Probes containing missing data were eliminated from the list. The threshold of detection was calculated using the normalized signal intensity for the 100 negative controls in the array and spots with signal intensity below the threshold were considered as absent. Quality of processing was evaluated by generating scatterplots of positive signal distribution. The signal intensities were then transformed to logarithm base 2.

A differential expression of at least 1.5 was used to generate the final list of genes of interest. Statistical comparison and filtering were performed using Genespring software 7.0 (Agilent). Two experiments (one with normal total brain and one with PTPR1) were performed in duplicate. Pearson's coefficients between repeat experiments regarding the genes were greater than 0.99.

Real-Time Reverse Transcription and Polymerase Chain Reaction

Oligonucleotide sequences corresponding to the selected gene transcript examined by reverse transcriptase-polymerase chain reaction (RT-PCR) were designed using Primer 3 software (Infobiogen, Villejuif, France) and are available on request from the authors. RNA samples (0.5 μg) from tumors, normal brain, and normal pineal glands were heated for 3 minutes at 70°C and then immediately placed in ice. First-strand DNA was synthesized by incubating 0.5 mM of each dNTP, 10 mM DTT, 40 U of RNA-sin (Promega, Charbonnières-les-Bains, France), 20 μM random hexamers, and 200 U of Moloney murine leukemia virus RT (M-MLV RT; Invitrogen SARL, Cergy Pontoise, France) for 90 minutes at 42°C in a final volume of 20 μL of RT buffer (50 mM Tris-HCl, pH 8.3, 75 mM KCl, and 3 mM MgCl2). The volume was then made to 100 μL with distilled water. Negative controls were performed by replacing the enzyme with water.

PCR was performed on a LightCycler instrument (Roche Diagnostics GmbH, Mannheim, Germany). cDNA samples (2 μL and 0.2 μL) were diluted in glass capillaries to a volume of 20 μL with PCR mix (LightCycler Faststart DNA Master SYBR Green Plus; Roche Diagnostics) containing a final concentration of 4 mM MgCl2 and 0.5 μM 3‘- and 5’-primers. The cDNA was denatured for 8 minutes at 95°C and then amplified by 40 or 50 cycles of 15 seconds at 95°C, 5 seconds at 62°C, and 10 seconds at 72°C. After amplification, the temperature was slowly raised above the melting temperature of the PCR products to measure the fluorescence for the melting curve, demonstrating the purity of the transcripts by their respective melting temperatures. Nonspecific products such as primer dimers could be readily distinguished from the product by their lower melting points. Negative controls without RT were also analyzed. The results were calculated from the crossing point values and expressed as the amount of test gene product relative to the amount of GAPDH product, used as a housekeeping gene, for the same sample. The presence of a single PCR product of the correct size was systematically verified by electrophoresis, the PCR products being electrophoresed for 2 hours at 70 V in 0.5 X Tris-borate-EDTA buffer, pH 8.3, on a 2% agarose gel (Tebu, Le Perray-en-Yvelines, France), and the DNA band visualized using ethidium bromide in the presence of a DNA molecular weight standard (100 bp; Promega) under ultraviolet illumination.


Clinical Data and Histologic Features

The clinical data for the patients are reported in Table 1 and the morphologic features of the tumors are shown in Figure 1. A typical pineocytoma (PC1) contained uniform cells with an amphophilic cytoplasm surrounding a round nucleus. The cells formed large fibrillary pineocytomatous pseudorosettes. The other subtype of pineocytoma (PC2), the pleomorphic variant, contained giant cells with hyperchromatic and strangely shaped nuclei that were sometimes multinucleated. The PB was hypercellular and showed uniform proliferation of small blue cells with a scanty cytoplasm surrounding a round hyperchromatic nucleus. The 2 PTPRs showed diffuse cellular proliferation with large areas with papillary features and were characterized by an epithelial-like growth pattern in which the vessels were covered by layers of tumor cells. Rosettes or tubes were observed together with some perivascular pseudorosettes. The CG consisted of cords and clusters of round epithelioid cells with an abundant eosinophilic cytoplasm distributed within a mucinous stroma containing a lymphoplasmatic infiltrate. The Ic and Sp EPs consisted of piriform cells forming perivascular pseudorosettes or tubes.

View this table:

Light microscopic features of the tumors analyzed (hemalin phloxine saffron staining; original magnification: 400×). (A) Typical pineocytoma with uniform cell proliferation with large fibrillary pineocytomatous pseudorosettes (arrow). (B) Pleomorphic pineocytoma with gangliocytic differentiation with hyperchromatic nuclei. (C) Pineoblastoma: proliferation of tumor cells with a scanty cytoplasm surrounding a hyperchromatic nucleus. (D, E) Papillary tumors of the pineal region. The tumors showed papillary features. The large and columnar neoplastic cells form pseudorosettes around vessels. (F) Chordoid glioma: cords of epithelioid cells embedded in a mucinous stroma. (G, H) Intracerebral (G) or spinal (H) ependymomas. The tumor cells form perivascular pseudorosettes around blood vessels or ependymal tubes.

Gene Expression and Cluster Analysis

The tumor samples expressed between 47.7% and 56.1% of the genes present on the microarray. The hierarchical clusters of the tumors are shown in Figure 2. Hierarchical clustering analysis of the entire set of genes in all tumors separated tumors of the pineal region and the chordoid glioma from ependymomas. The normal brain sample was separate from the tumors (not shown). The 2 PCs presented the same expression signature and the PB had a distinct gene expression profile in the same cluster. The 2 PTPRs were clustered in the same group and were related to the CG profile. The 2 Ic EPs and 2 Sp EPs presented a common signature but were classified in 2 different groups.


Hierarchical cluster analysis of gene expression in papillary tumor of the pineal region, chordoid glioma, and pineal parenchymal tumor (A) or ependymomas (B).

Genes Selectively Expressed in Different Tumors

We selected those genes up- or downregulated in these tumors (Table 2). Six genes were only expressed in PC: OPN4, prkWNK4, SLC22A8, PFKB2, CRB3, and PAX4, coding, respectively, for melanopsin, a protein kinase, an anionic transporter, an enzyme controlling glycolysis, an apical protein of epithelial cells, and a transcription factor. Seven genes were only expressed in PB: HOXD19, PITX2, and POU4F2 encoding transcription factors, Hist1H3D and Hist1H4E encoding histones, DSG1 encoding a protein of intercellular junctions, and TERT encoding an enzyme implicated in telomerase formation. Some genes upregulated in PC were also expressed in PB; these were TPH1 and HIOMT encoding melatonin pathway enzymes, TrkA encoding a neurotrophic factor, RGS16 encoding a regulator of G protein signaling, and BTG1 encoding a protein expressed in the G0/G1 phases of the cell cycle. CHRM3 coding for a muscarinic acetylcholine receptor was upregulated in PC and downregulated in PTPR and ependymomas as compared with normal brain. NRG1 and FHIT, encoding, respectively, a regulator of tyrosine kinase and an oncosuppressor, were present in several tumors but were only overexpressed in PC. Some genes overexpressed in PB were also highly expressed in PC (SOX 4) or in PTPR (PRAME, TEP1, and CDK2). PRAME (preferentially expressed antigen in melanoma) and CD24 (small cell lung carcinoma cluster 4 antigen) were also highly expressed in Ic EP.

View this table:

Five genes were downregulated in PC: SOX9, FAB7, C1QTNF3, FZD7, and SFRP4, encoding, respectively, a transcription factor, a fatty acid transporter, a phosphate transporter, and the Wnt protein receptor and its agonist. SOX9, FAPB7, and FZD7 were also downregulated in PB. Two genes implicated in the Wnt signaling pathway (APC and APC2) were downregulated in PB.

Only one gene, SPDEF, encoding a transcription factor, was only expressed in PTPR. Genes encoding 2 transcription factors (RFX3 and ZFH4), a regulator of cellular death (CLUL1), a hormonal transporter (TTR), a neuropeptide (CGRP), a receptor for retinol-binding protein on the surface of the retinal pigment epithelium (RPE65), and a cytokeratin (KRT18), which were highly expressed in PTPR, were also present in ependymomas.

Seventy-three genes were only expressed in the CG (detailed list of genes not shown). The 3 most highly expressed coded for a RAB-GTP-binding protein (RAB38), a tumor-associated calcium signal transducer (TACSTD2), and a giant phosphoprotein (AHNAK).

Selected genes of interest that were up- or downregulated in EPs are reported in Table 3. Four genes were only expressed in Ic EP and 6 only in Sp EP, whereas 2 genes were not present in Ic EP and 2 not in Sp EP. Wee 1 and MSX1, which were overexpressed in Ic EP, and code for a protein kinase and a transcription factor, were also highly expressed in Sp EP and in PTPR. HOXA5 and MSX2, 2 genes encoding transcription factors, were upregulated in both Sp EP and Ic EP.

View this table:

Real-Time Reverse Transcriptase-Polymerase Chain Reaction Confirmation of Expression Profiles

To assess the microarray data, 15 genes were analyzed in more detail using real-time RT-PCR (Table 4). Thirteen of the transcripts examined demonstrated concordant levels of differential expression in the microarray and RT-PCR analyses. Two transcripts showed no concordance, because RT-PCR failed to detect Wee1 in Sp EPs and did not find high expression of PRL in PPT.

View this table:

Genes found to be differentially expressed between PC and PB by microarray might serve as candidate genes to grade PPT with intermediate differentiation. We therefore used RT-PCR to verify the expression of 4 genes in 3 PC, 4 PB, 6 PPTint, and in 2 normal pineal glands and normal brain (Table 5). The 4 genes (PRAME, CD24, POU4F2, and HOXD13) were absent or showed very low expression in the PCs in the PPTint grade 2 and in normal pineal and normal brain tissues. These genes were upregulated in the PBs and in the PPT grade 3.

View this table:


Gene expression-based classification has been shown to be a useful tool for tumor classification in several types of brain tumors, especially glial tumors (11-14). Our study is the first to use microarray to analyze gene expression in several types of tumors of the pineal region and to identify genes with potential use as diagnostic markers. Clustering analysis based on the entire gene set on the array clearly classifies the PPTs together, the 2 variants of PC (typical and pleomorphic) having very similar profiles. PTPR and CG, 2 tumors probably derived from the circumventricular organs (3, 4, 6, 8), are grouped together and separate from ependymomas. Potentially interesting expression of some genes is identified in each type of tumor. The expression in PC of TPH1 and HIOMT, 2 genes coding for enzymes involved in melatonin synthesis, suggests that neoplastic pinealocytes retain certain specific markers of their normal counterpart, as previously shown in one tumor (15). TPH1 is also expressed in PB, but at a lower level, possibly as a result of the undifferentiated state of the tumor. Several genes that are either selectively expressed or upregulated in PC or PB are related to phototransduction in the retina. OPN4, a gene coding for melanopsin, is expressed in photosensitive mouse retinal cells (16) and in photosensitive chicken pinealocytes (17) but has never been detected in the mammalian pinealocyte. RGS16 may play a role in regulating the kinetics of signaling in the phototransduction cascade (18), whereas CRB3 is essential for the morphogenesis of photoreceptors (19). Finally, POU4F2 is an essential regulator of gene expression in the mouse retinal ganglion cells (20) and is also expressed in the human retina (21). The pineal gland has the function of a photoreceptor organ in lower animals and retains vestiges of a photoreceptor organ in the embryonic period. The reexpression of these genes involved in phototransduction may be related to photosensory differentiation, which can occur in the neoplastic pineal. Furthermore, the expression of several photoreceptor proteins, in particular, rhodopsin, S-antigen, and cellular retinaldehyde-binding protein, has been described in PPT (22-24), and neurosensorial differentiation has been well documented by ultrastructural studies (25). The expression of these photosensory-related proteins in the PPT may provide a useful diagnostic tool for better characterizing the neurosensorial differentiation of these tumors. Two other retinal genes, RPE65, a gene expressed in the retinal pigment epithelium with an important role in retinoid processing and/or retinoid transport in the eye and CLUL1, coding for a cone photoreceptor, the expression of which follows retinal differentiation, are found to be upregulated in PTPRs and EPs. RPE65 transcripts have been detected in transformed kidney cells and in renal tumors in culture (26) and, recently, in nonmelanocytic skin tumors, with the level of expression correlating with the invasiveness of the tumor (27). The role of this gene overexpression in brain tumors remains to be elucidated.

Several genes upregulated in PB have been reported to be strongly expressed in other tumors. UBEC2, coding for a ubiquitous conjugase, is reported to be overexpressed in many different types of cancer and the degree of expression is associated with the degree of tumor differentiation (28). SOX4, coding for a transcription factor involved in neural development, has been reported in medulloblastoma, an embryonic neoplasm morphologically similar to PB (29). Amplification of TERT, coding for telomerase catalytic protein subunit, and its increased mRNA expression have been documented in medulloblastoma and other embryonic tumors and is associated with biologically aggressive tumor behavior (30). In addition, the high levels of telomerase protein component 1 (TEP1) in PB might possibly be linked with abnormal telomerase function in high-grade PPT. The high expression of transcripts encoding histones might be related to the tumoral proliferative state, as previously shown for histone H3 in pediatric brain tumors (31). PRAME is highly expressed in childhood myeloid leukemia (32) and in high-stage neuroblastoma (33) and has been suggested as a target for immunotherapy. CD24 is a cell surface molecule expressed in human glioma with a suggested role in invasion (34). Moreover, PRAME and CD24 were shown to be highly expressed in medulloblastoma using serial analysis of gene expression (35), and our results point out molecular similarities between medulloblastoma and pineoblastoma, 2 embryonic tumors with different locations. POU4F2, highly expressed in PB, has been shown to be a regulatory factor of the growth, behavior, and invasiveness of human neuroblastoma (36). HOXD13, a homeobox gene controlling morphogenesis, is highly expressed in human invasive breast carcinoma (37). As some of these genes are overexpressed in the high-grade tumors (PB) and weakly expressed in lower grades, they could be candidate genes to serve as markers of differentiation. The expression of 4 genes in PPTs of different grades was analyzed by real-time RT-PCR. Our results show that these 4 genes are expressed in PPTint at a level intermediate between those in PC and PB, these levels being concordant with the grades 2 and 3 proposed in our classification for PPT (25). Other genes are downregulated in PPT compared with other tumors or normal brain: 4 of these genes (SFRP4, FZD7, APC, and APCL) are related to the Wnt signaling cascade, which has been shown to be involved in medulloblastoma carcinogenesis (38). The downregulation of APC and APCL, 2 tumor-suppressor genes, confirms our previous data (39).

In PTPRs, one of the most interesting findings is the presence of high levels of genes coding for proteins reported in the SCO. The zinc finger-homeodomain transcription factor (ZFH4) is intensely expressed in SCO cells in the embryonic rat brain, in which it might regulate the differentiation of this structure (40). Abundant amounts of calcitonin gene-related peptide (CGRP or CALCA) are found in SCO ependymocytes in the golden hamster (41). Moreover, PTPRs, like Sp EPs, express high levels of RFX3, which codes for a transcription factor playing an important function in the differentiation of the ependymal cell lineage and shown to be expressed in the mouse SCO (42). Finally, our results also demonstrate the expression of transthyretin (TTR) transcripts in PTPR. The only previously known brain source of TTR, a protein involved in the transport of thyroid hormone, is the choroid plexus, but, recently, the bovine SCO was also reported to synthesize TTR mRNA (43). PTPRs might synthesize TTR, which has been shown by immunohistochemistry to be present in one PTPR (4). Moreover, our study shows the presence of TTR mRNA in PCs. TTR transcripts are abundant in the chick pineal gland (44) but have never been described in human pineal tumors. Positive cytokeratin immunostaining has been reported in PTPR using different broad-spectrum anticytokeratin antibodies (3, 4). The subtype of cytokeratin could be 18CK, because very high levels of transcript encoding this protein are observed in PTPR in our study. Surprisingly, in PTPR, we find selective expression of SPEF (Sam pointed domain-containing ETS transcription factor), a transcription factor only described in the prostate epithelium, which could be implicated in prostate cancer development (45). PTPR are classified in the same family as CG in agreement with a possible similarity between these 2 neoplasms. Nevertheless, only CG expresses certain transcripts such as AHNAK mRNA. This protein is present in most lining epithelia that form physical barriers and in endothelial cells forming specific blood-tissue barriers, but its exact biologic function is unknown. A role in organizing the cytoarchitecture at the plasma membrane has been suggested (46) as well as in the regulation of epithelial cell adhesion and permeability (47). Whether this protein is present in CG and its localization in tumor cells or in vessels remain to be elucidated.

In EPs, 4 genes (IGF2, MSX1, HOXB5, and NF2) were found to be differentially expressed in our study and a previous one (48). Spinal ependymomas, which show high expression of HOXB5, also express high levels of other homeobox genes such as HOXD8, HOXA9, and HOXA5. The high expression of insulin-like growth factor 2 (IGF2) transcripts in Ic EP is in agreement with the positive staining of EPs by anti-IGF2 antibodies (49). The absence of expression of NF2 in Sp EP might correspond to NF2 deletion or mutation, often seen in Sp EP (50). MSX1 and MSX2 were found to be highly expressed in EPs and have been previously reported to be involved in the tumorigenesis of other human malignancies (51). MSX1 is also highly expressed in PTPR, corroborating the strong expression of this gene in the mouse SCO (52). MSX1 might possibly be required for the differentiation or regulation of certain types of mature ependymal cells.

The exact biologic and clinical significance of the different genes reported in our study will only be determined by further studies, especially using an immunohistochemical approach. However, the data presented in this report provide a molecular classification of the tumors of the pineal region and point out candidate genes that could serve as phenotypic or grading markers. It will be interesting to determine the use of this approach by measuring their expression in a larger number of tumor samples. This approach might result in new diagnostic markers for improved management of patients with tumors of the pineal region.


The authors thank Neurobiotec and the neurosurgeons (Prof. R. Deruty, Prof. F. Grisoli, Pr. O. Dutour, Prof. C. Lapras, Dr. G. Lena, Dr. K. Mahla, Dr. I. Pelissou-Guyotat, Prof. G. Perrin, Dr. J. Reymond, Prof. M. Sindou, and Prof. B. Vallee) at the Hôpital Neurocardiologique et Neurochirurgical Pierre Wertheimer in Lyon and the Centre Hospitalier Regional la Timone in Marseille for supplying the tumor samples. The authors also thank Dr. T. Barkas for linguistic help and C. Rey for technical advice in the PCR experiments.


  • Drs. Champier, Szathmari, and Wierinckx contributed equally to this work and Drs. Jouvet and Lachuer codirected this work.

  • This work was supported by INSERM, the Association for Cancer Research (grant ARC 3277), and the Rhône-Alpes region (grant 04 020445 02).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
View Abstract