OUP user menu

Expression of Hydroxyindole-O-Methyltransferase Enzyme in the Human Central Nervous System and in Pineal Parenchymal Cell Tumors

Takahiro Fukuda MD, PhD, Nobutake Akiyama PhD, Masahiro Ikegami MD, PhD, Hitoshi Takahashi MD, Atsushi Sasaki MD, PhD, Hidehiro Oka MD, Takashi Komori MD, Yuko Tanaka MD, Youichi Nakazato MD, Jiro Akimoto MD, Masahiko Tanaka MD, Yoshikazu Okada MD, Saburo Saito MD, PhD
DOI: http://dx.doi.org/10.1097/NEN.0b013e3181db7d3c 498-510 First published online: 1 May 2010

Abstract

Pineal parenchymal tumor (PPT) cells usually show immunoreactivity for synaptophysin, neuron-specific enolase, neurofilament protein, class III β-tubulin, tau protein, PGP9.5, chromogranin, serotonin, retinal S-antigen, and rhodopsin, but these markers are not specific for PPTs. Melatonin is produced and secreted mainly bypineal parenchymal cells; hydroxyindole-O-methyltransferase (HIOMT) catalyzes the final reaction in melatonin biosynthesis. We hypothesized that HIOMT could serve as a tumor marker of PPTs, and we investigated HIOMT localization and HIOMT expression in samples of normal human tissue and in PPTs, primitive neuroectodermal tumors, and medulloblastomas. In normal tissue, HIOMT was expressed in retinal cells, pineal parenchymal cells, neurons of the Edinger-Westphal nucleus, microglia, macrophages, thyroid follicular epithelium, principal and oxyphil cells of parathyroid gland, adrenal cortical cells, hepatic parenchymal cells, renal tubule epithelium, and enteroendocrine cells of stomach and duodenum. The HIOMT was also expressed in all 46 PPTs studied. The proportions of HIOMT-immunoreactive cells successively decreased in the following tumors: pineocytoma, pineal parenchymal tumor of intermediate differentiation, and pineoblastoma. A few HIOMT-immunoreactive cells were observed in one of 6 primitive neuroectodermal tumors and 23 of 42 medulloblastomas. These results indicate that HIOMT immunohistochemistry may be useful for the diagnosis of PPTs and be a prognostic factor in PPTs.

Key Words
  • Edinger-Westphal nucleus
  • Hydroxyindole-O-methyltransferase
  • Microglia
  • Pineal gland
  • Pineal parenchymal cell tumors
  • Retinal S-antigen

Introduction

Pineal parenchymal tumors (PPTs), derived from pineocytes, are rare neoplasms and account for less than 1% of primary central nervous system (CNS) tumors (1). The revised World Health Organization classification of CNS tumors divides PPTs into pineocytoma (PC), pineoblastoma (PB), and PPT of intermediate differentiation (PPTID) (2). The 5-year survival rates of PC, PPTID, and PB are 86% to 100%, 39% to 74%, and 58%, respectively (2). Histologically, a high mitotic index and necrosis are associated with a poorer outcome, whereas positive immunostaining for neurofilaments is associated with a better survival (1). The distinction of high-grade PPTs from other embryonal tumors in CNS has relied heavily on knowledge of the primary site of the tumor and the extent of pineocytomatous differentiation, such as club-shaped argyrophilic processes, pineocytomatous rosettes, and photoreceptors (2). The PPTs show positive immunostaining for S100 protein, neuron-specific enolase, synaptophysin, neurofilaments, class III β-tubulin, tau protein, PGP9.5, retinal S-antigen (SAG), chromogranin A, serotonin, and α-B crystallin (1, 3-8). These proteins are useful for the identification of neural and neural crest cells. There is, however, no specific marker available for differentiating high-grade PPTs from other embryonal tumors (2).

Melatonin is a powerful antioxidant molecule involved in the protection of nuclear and mitochondrial DNA and in the regulation of circadian seasonal rhythms and immune function (9-13). It is produced and secreted predominantly by the pineal gland. Tryptophan, the precursor of melatonin, is metabolized into 5-hydroxytryptophan by tryptophan-hydroxylase; 5-hydroxytryptophan is then metabolized by aromatic amino acid decarboxylase into N-acetylserotonin. N-acetylserotonin then undergoes modification by arylalkylamine-N-acetyltransferase. The result of this modification is metabolized into melatonin by hydroxyindole-O-methyltransferase (HIOMT) (14).

The expression of messenger RNAs (mRNA) for tryptophan-hydroxylase, arylalkylamine-N-acetyltransferase, and HIOMT has been previously detected in tissues or cell cultures derived from PPTs by microarray, real-time reverse transcription and polymerase chain reaction, in situ hybridization, and Northern blot analyses (15-17). Furthermore, HIOMT activity has been detected in PPTs (18-20). The evidence of high mRNA expression in PPTs for these enzymes suggests that one or more of them could serve as diagnostic markers and tools for understanding the biology of PPTs. In the present study, we focused on HIOMT and analyzed normal pineal glands, other human tissues, PPTs, and embryonal tumors in CNS by immunohistochemistry (IHC) for evidence of melatonin synthesis. We further correlated HIOMT expression with the histological differentiation of PPTs. We also compared HIOMT expression with that of retinal SAG, a major soluble photoreceptor protein that is involved in desensitization of the photoactivated transduction cascade (21-23), in CNS PPTs and embryonal tumors.

Materials and Methods

Generation of Human Full-Length HIOMT- and Retinal SAG-Specific Antibodies

Using primers (AGATCTACCATGGGATCCTCAGAGGAC and TTTCTCGAGTTTCCTGGCTAAAATGGC for HIOMT, GGATCCACCATGGCAGCCAGCGGGAAGA and TTTGTCGACCTCATCAACGTCATTCTTGT for SAG), full-length HIOMT (GenBank Accession No. NM_004043.2) or SAG (GenBank Accession No. NM_000541.4) cDNA was amplified from human retinal total RNA (Clontech, Mountain View, CA) and cloned into plasmid pET-30a (Novagen, Madison, WI). Using this vector, HIOMT (or SAG) protein with His6-tag was synthesized in the Rapid Translation System RTS500 (Escherichia coli HY kit; Roche, Denver, CO). The His6-fused HIOMT (or SAG) protein was purified using a Ni-NTA agarose column (QIAGEN, Hilden, Germany), and anti-HIOMT (or SAG) protein polyclonal antibodies were raised in 2-month-old male BALB/C mice (Sankyo Laboratories, Tokyo, Japan) or 4-month-old male Japanese White rabbits (Sankyo Laboratories) against the human HIOMT (or SAG) protein. The HIOMT (or SAG) protein 100 µg with Freund complete adjuvant (Pierce Biotechnology, Inc, Rockford, IL) was injected intraperitoneally in mice and subcutaneously in rabbits. Thereafter, HIOMT (or SAG) protein 100 µg with Freund incomplete adjuvant (Pierce Biotechnology) was injected 6 times at an interval of 2 weeks. Animal sera were screened for antibody against the full-length HIOMT (or SAG) protein by an ELISA. For immunization, we used HIOMT isoform 3 (identifier: P46597-3, 373 amino acids) (24, 25). Affinity-purified anti-HIOMT (or SAG) antibody (IgG, final concentration 1 mg/mL) was furthermore purified by Melon Gel IgG Purification Kit (Pierce Biotechnology).

Transfection

Using primers (AGATCTACCATGGGATCCTCAGAGGAC and AAACTCGAGGTGCTTTGACGTTAGTTC for HIOMT, GGATCCACCATGGCAGCCAGCGGGAAGA and AAAGTCGACCAAACAAGCTTTATTTCTCGGA for SAG), full-length HIOMT (or SAG) cDNA was amplified from human retinal total RNA (Clontech) and cloned into pcDNA3.1(+) (Invitrogen, Carlsbad, CA). Using the FreeStyle 293 expression system (Invitrogen), the vector was transfected to the cell line. For the negative control, we also made pcDNA3.1(+)-enhanced green fluorescent protein (EGFP) vector-transfected cells.

Western Blotting

Transfected cell extracts were prepared by the addition of lysis buffer (20 mmol/L HEPES, pH 7.0, 1 mmol/L EDTA, 1 mmol/L dithiothreitol, EDTA-free protease inhibitor cocktail [Roche], 120 mmol/L NaCl, and 0.5% Triton). To the supernatant of 8 mg/mL protein, 1 volume of 2× sample buffer (Sigma-Aldrich, St Louis, MO) was added. Protein samples (20 µg) were resolved on a 12.5% sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred to nitrocellulose transfer membranes (Whatman Schleicher & Schuell, Dassel, Germany). Membranes were blocked and incubated overnight at 4°C with rabbit and mouse anti-HIOMT (or SAG) antibodies (1:100,000-500,000). Bound antibody was visualized by incubation with horseradish peroxidase-conjugated secondary antibody (1:10,000; Millipore, Bedford, MA) and subsequent enhanced chemiluminescence (ECL plus system; Amersham Biosciences, Piscataway, NJ) followed by exposure to Hyperfilm-ECL.

Immunocytochemical Staining of Cultured Cells

Immunocytochemistry was performed on both pcDNA3.1(+)-HIOMT vector-transfected cells or pcDNA3.1(+)-SAG vector-transfected cells and pcDNA3.1(+)-EGFP vector-transfected cells on collagen I-coated cover slides. After fixation with 4% paraformaldehyde in 0.01 mol/L PBS, slides were blocked with 3% H2O2 in methanol. Anti-HIOMT (or SAG) antibody was used at a 1:250,000 dilution and incubated overnight at 4°C. Biotinylated anti-mouse or anti-rabbit IgG (H + L) antibody (Vector Laboratories, Burlingame, CA) in PBS was applied for 1 hour, and slides were serially stained with avidin-biotin complex using ABC elite kit (Vector Laboratories) and 3,3′-diaminobenzidine as a chromogen. Counterstaining for cell nuclei was performed with Mayer hematoxylin. pcDNA3.1(+)-EGFP vector-transfected cells were used as negative controls.

Human Tissue

We analyzed the cerebrum, cerebellum, brainstem, spinal cord, dorsal root ganglion, and pineal gland of 5 autopsy cases (postmortem interval, 2-4 hours) without neurological disease or neuropathologic abnormalities. Control tissue without pathological abnormalities consisted of pituitary, thyroid, parathyroid, adrenal, lung, heart, liver, spleen, pancreas, stomach, duodenum, small intestine, colon, kidney, prostate, testis, ovary, and bone marrow, and was obtained from surgical resection specimens (4-6 cases per organ). Tumor specimens consisted of 46 PPTs (8 PCs, 25 PPTIDs, 13 PBs), 6 CNS primitive neuroectodermal tumors (PNETs), and 42 medulloblastomas (MBs). Clinical data are summarized in Tables 1 and 2. Experiments were approved by the ethics committee of the Jikei University School of Medicine (Permission No. 20-201 5491).

View this table:
TABLE 1.
View this table:
TABLE 2.

Immunohistochemical Analysis

After fixation with 10% neutral phosphate-buffered formalin, 3- or 6-µm-thick paraffin-embedded tissue sections or 20-µm-thick linear slicer sections (PRO7; Dosaka, Kyoto, Japan) were cut. Sections were stained with hematoxylin and eosin. Tumor grading was established by light microscopy according to the World Health Organization classification (2) and Jouvet classification (1). Immunohistochemistry and double-immunofluorescence staining were performed, as described (26). Anti-HIOMT and anti-SAG antibodies were used at dilutions of 1:5000 to 50,000. In negative controls, the primary antibodies were omitted. We also evaluated the expression of synaptophysin (SVP38, 1:100, Sigma-Aldrich; or SY38, 1:50, PROGEN Biotecknik, Heidelberg, Germany) and neurofilament (2F11 [68 kd and 200 kd], 1:1000, Dako, Glostrup, Denmark; or cocktail of NR4 [68 kd], 1:400, Sigma-Aldrich; NN18 [160 kd], 1:100, Sigma-Aldrich; and SMI311 [160 kd and 200 kd], 1:1000, Covance, Emeryville, CA) in PPTs. Histochemical identification of microglia was performed using biotinylated Ricinus communis agglutinin I (RCA-I, 1:10,000; Vector Laboratories). The expression of HIOMT in microglia was examined by double-fluorescence staining using rabbit anti-HIOMT antibody and biotinylated RCA-I. Indocarbocyanine (Cy3)-conjugated anti-rabbit immunoglobulin (Jackson Immuno Research, West Grove, PA) and Alexa488-conjugated streptavidin (Invitrogen) were used for fluorescence labeling. Colocalization of HIOMT and SAG in the pineal gland, PPTs, PNETs, and MBs, was investigated using double-immunofluorescence staining with rabbit anti-HIOMT antibody and mouse anti-SAG antibody, Alexa488-conjugated anti-rabbit immunoglobulin (Invitrogen), and Cy3-conjugated anti-mouse immunoglobulin (Jackson ImmunoResearch). In the PPTs, we counted HIOMT-, SAG-, or human Ki-67 (MIB-1, dilution 1:500; Zymed Laboratories, San Francisco, CA)-immunoreactive cells per 800 to 1,200 tumor cells and calculated the percentages of immunoreactive cells. Immunoreactive cells per 800 to 1,200 tumor cells were also counted in the pineal glands, PNETs, and MBs.

Statistical Analysis

Data are expressed as mean ± SEM. The percentages of HIOMT- and Ki-67-immunoreactive cells among the groups of Jouvet prognostic 4-grade scheme (1) were analyzed using 1-way analysis of variance followed by a post hoc Tukey-Kramer test. All statistical analyses were conducted using JMP software (version 7.0.2; SAS Institute Inc, Cary, NC). A level of p < 0.05 was considered significant.

Results

Specificity of Affinity-Purified Antibodies

On Western blot, rabbit and mouse anti-HIOMT antibodies recognized a 38-kd band in the protein extracts from pcDNA3.1(+)-HIOMT-transfected cells (Fig. 1, Lanes a and c). There was no band in extracts from pcDNA3.1(+)-EGFP-transfected cells (Fig. 1, Lanes b and d). Rabbit and mouse anti-SAG antibodies also recognized a 45-kd BAND in the protein extracts from pcDNA3.1(+)-SAG-transfected cells (Fig. 1, Lanes e and g) but not in extracts from pcDNA3.1(+)-EGFP-transfected cells (Fig. 1, Lanes f and h). Immunocytochemistry with rabbit and mouse anti-HIOMT (or anti-SAG) antibodies revealed the cytoplasmic localization of HIOMT (or SAG) in pcDNA3.1(+)-HIOMT (or SAG)-transfected cells (Figs. 1a, c, e, g) but not in pcDNA3.1(+)-EGFP-transfected cells (Figs. 1b, d, f, h).

FIGURE 1.

Immunocytochemistry and Western blot of affinity-purified anti-hydroxyindole-O-methyltransferase (HIOMT) and anti-retinal S-antigen (SAG) antibodies. Rabbit (a) and mouse (c) anti-HIOMT antibodies show cytoplasmic localization of HIOMT in the pcDNA3.1(+)-HIOMT-transfected FreeStyle 293 cells but not in the pcDNA3.1(+)-EGFP-transfected cells (b, d). Rabbit (Lane a) and mouse (Lane c) anti-HIOMT antibodies recognize a band of 38 kd in the pcDNA3.1(+)-HIOMT-transfected FreeStyle 293 cell lysate but not in the pcDNA3.1(+)-EGFP-transfected cell lysate (Lanes b and d). Rabbit (e) and mouse (g) anti-HIOMT antibodies show cytoplasmic localization of SAG in the pcDNA3.1(+)-SAG-transfected FreeStyle 293 cells but not in the pcDNA3.1(+)-EGFP-transfected cells (f, h). Rabbit (Lane e) and mouse (Lane g) anti-HIOMT antibodies recognize a band of 38 kd in the pcDNA3.1(+)-HIOMT-transfected FreeStyle 293 cell lysate but not in the pcDNA3.1(+)-EGFP-transfected cell lysate (Lanes f and g). Scale bars = 20 μm.

Localization of HIOMT and Retinal SAG in Human Tissues

By IHC, HIOMT was diffusely distributed in the cytoplasm of pineal parenchymal cells (Fig. 2e), epithelial cells of the choroid plexus close to the pineal gland (Fig. 2b), and renal tubule epithelial cells (Fig. 2l). The HIOMT-positive granules were found spread diffusely in the cytoplasm of retinal cells (Fig. 2a), microglia (Fig. 2c), macrophages, and lymphocytes (spleen; Fig. 2k, gastrointestinal tracts, lung, etc), thyroid follicular epithelium (Fig. 2f), principal and oxyphil cells of the parathyroid gland (Fig. 2g), adrenal cortex cells (Fig. 2h), hepatic parenchymal cells (Fig. 2i), and enteroendocrine cells of the stomach (Fig. 2j) and duodenum. The HIOMT-immunoreactive granules localized at the perinuclear cytoplasm of neurons in the Edinger-Westphal nucleus (Fig. 2d) and at the apical portion of the thyroid follicular cells. The SAG was diffusely distributed in the cytoplasm of pineal parenchymal cells (Fig. 2m), retinal cells (Fig. 2n), adenohypophyseal parenchymal cells (Fig. 2o), vascular smooth muscle cells, and renal tubule epithelial cells (Fig. 2p). Double-immunofluorescence labeling revealed that HIOMT-positive SAG-positive cells made up 5.8% of parenchymal cells; HIOMT-positive SAG-negative cells, 92.4%; and HIOMT-negative SAG-positive cells, 1.8%.

FIGURE 2.

Hydroxyindole-O-methyltransferase (HIOMT) and retinal S-antigen (SAG) expression in non-neoplastic normal tissues. (a-p) The HIOMT is expressed in retinal cells (a), choroid plexus epithelial cells (b), microglia (c), neurons in the nonpreganglionic Edinger-Westphal nucleus (d), pineal parenchymal cells (e), thyroid follicular epithelium (f), principal and oxyphil cells of parathyroid gland (g), adrenal cortex cells (h), hepatic parenchymal cells (i), enteroendocrine cells of stomach (j), macrophages and lymphocytes in spleen (k), and renal tubule epithelial cells (l). The SAG is expressed in pineal parenchymal cells (m), retinal cells (n), adenohypophyseal parenchymal cells (o), and renal tubule epithelial cells (p). (q-v) Double-immunofluorescence labeling of HIOMT and SAG reveals HIOMT-positive and SAG-positive cells, HIOMT-positive and SAG-negative cells, and HIOMT-negative and SAG-positive cells in the pineal gland (q, HIOMT; r, merged; s, SAG). Ricinus communis agglutinin I (RCA-I)-reactive perivascular microglia demonstrates HIOMT immunoreactivity (t, RCA-I; u, merged; v, HIOMT). Scale bars = (a-s) 20 μm; (t-v) 10 μm.

Tumor Pathology

Pineocytomas typically showed uniform cells with amphophilic cytoplasm surrounding a round or indented nucleus (Fig. 3a). The cells formed pineocytomatous rosettes. Mitotic activity, endothelial cell hyperplasia, and necrosis were not present. The PPTIDs had 4 different morphological subtypes: a dense lobular arrangement of cells with endocrine-like vessels (Fig. 3d) (lobulated subtype; Cases 14, 26, and 29); more diffuse proliferation mimicking oligodendroglioma or neurocytoma (diffuse subtype; Cases 9, 15, 17, 20-24, 31, and 33); lobulated and/or diffuse areas associated with other areas containing pineocytomatous rosettes corresponded to transitional forms (transitional subtype; Cases 11, 16, 27, 28, 30, and 32); and a distinctly biphasic pattern including areas of typical PC (Fig. 3b) and PB (Figs. 3c, e) (mixed subtype; Cases 10, 12, 13, 18, and 25). From 0 to 5 mitotic figures were seen in 23 of 25 PPTIDs, and necrosis was present in 7 of 25 cases and endothelial cell hyperplasia in 12 cases. The PBs were densely cellular and consisted of uniform small blue cells with scant cytoplasm surrounding round and highly chromatic nuclei (Fig. 3f). Mitotic figures were present in all PBs, with numbers ranging from 3 to 34 per 10 high-power field. Necrosis was present in 9 of 13 cases, and endothelial cell hyperplasia was present in 10.

FIGURE 3.

Light microscopy and immunohistochemistry of pineal parenchymal tumors. (a) Typical pineocytoma (PC), uniform cell proliferation with fibrillary pineocytomatous rosettes (Case 1). (b-e) Pineal parenchymal tumors of intermediate differentiation (PPTID). There is a distinctly biphasic pattern including typical pineocytomatous areas (b) and pineoblastomatous areas (c, Case 12; b, Case 33). Dense lobular arrangement of cells with an endocrine-like vascularity in PPTID (e, Case 26). (f) Pineoblastoma (PB) showing uniform proliferation of small blue cells with scant cytoplasm surrounding round and hyperchromatic nuclei (Case 34). (g, j, m, p, s) Ki-67 immunolabeling in a PC (g, Case 1), a PPTID (j, large tumor cells in pineocytomatous region; m, Case 26; p, small tumor cells of pineoblastomatous region; Case 33), and a PB (s, Case 34). (h, k, n, q, t) Synaptophysin immunolabeling in a PC (h, Case 1), PPTIDs (k, large tumor cells of pineocytomatous region [as in b]; n, Case 26, small tumor cells of pineoblastomatous region [q, Case 33]), and a PB (t, Case 34). (i, l, o, r, u) Neurofilament immunolabeling in a PC (i, Case 1), PPTIDs (l, large tumor cells of pineocytomatous region [as in b]; o, Case 26), small tumor cells of pineoblastomatous region (r, Case 33), and a PB (u, Case 34). (a-f) hematoxylin and eosin; (g-u) immunoperoxidase with hematoxylin. Scale bars = 20 μm.

Immunohistochemistry

Immunohistochemistry data are summarized in Tables 1 and 2. All PPTs had HIOMT-immunoreactive cells. The HIOMT was diffusely distributed in the tumor cell cytoplasm in PCs (Fig. 4a), and there were small HIOMT-positive granules in the tumor cell cytoplasm in PB (Fig. 4f). Both cytoplasmic patterns were observed in PPTIDs. The diffuse distribution of HIOMT-positive granules was predominantly in the relatively large tumor cells in pineocytomatous areas (Fig. 4b), and there was a granular distribution in the relatively small tumor cells of the pineoblastomatous areas (Fig. 4d). Large tumor cells had HIOMT-positive cytoplasm in the lobulated subtype of PPTID (Fig. 4e). The SAG was identified in 4 of 8 PCs, 7 of 25 PPTIDs, and 3 of 13 PBs. The SAG-positive tumor cells were diffusely distributed in PPTs (Fig. 4c). The percentages of SAG-positive cells were from 0.4% to 24.8%. Double-immunofluorescence labeling revealed that PPTs had HIOMT-positive SAG-negative cells and sparse numbers of HIOMT-negative SAG-positive cells. Intense positive staining for synaptophysin was seen in PCs (Fig. 3h) and PPTIDs (Figs. 3k, n) and was also present in tumor cell cytoplasm in the immature pineoblastomatous areas of PPTID but at a lower intensity (Fig. 3q). Synaptophysin immunolabeling was found in all PBs (Fig. 3t) but at a lower intensity compared with other PPTs. Neurofilament was present in all PCs (Fig. 3i), 13 of 25 PPTIDs (Fig. 3l), and 5 of 13 PBs. No or very low neurofilament immunoreactivity was observed in some PPTIDs (Figs. 3o, r) and PBs (Fig. 3u). There were no differences in immunoreactivity between anti-neurofilament antibody (2F11) and the monoclonal antibody cocktail.

FIGURE 4.

Hydroxyindole-O-methyltransferase (HIOMT) and SAG expression in pineal parenchymal tumors, primitive neuroectodermal tumors (PNETs), and medulloblastomas (MBs). (a) The HIOMT is diffusely distributed in the cytoplasm of pineocytoma (Case 1). (f) The HIOMT-immunoreactive granules are found in the cytoplasm of pineoblastoma (Case 34). (b,d, e) In pineal parenchymal tumors of intermediate differentiation (PPTIDs), there is diffuse cytoplasmic distribution predominantly in the relatively large tumor cells of pineocytomatous region (b, Case 12; there is granular distribution in the relatively small tumor cells of pineoblastomatous region (d, Case 12). Large tumor cells demonstrate HIOMT-immunoreactive cytoplasm in the PPTID with a dense lobular arrangement of cells and endocrine-type vascularity (e, Case 26). (c) The SAG-immunoreactive tumor cells are diffusely distributed in the pineoblastoma (Case 47). (g, h) In 1 of 3 PNETs studied, there are HIOMT-immunoreactive tumor cells (g, Case 47) and SAG-immunoreactive tumor cells (h, Case 47). (i-k) Double-immunofluorescence labeling of HIOMT and SAG reveals that they tend to colocalize in the same tumor cells in MBs (i, HIOMT; j, merged; k, SAG; Case 56). Scale bars = 20 μm.

According to Jouvet prognostic 4-grade scheme, 8 PCs were Grade I, 13 PBs were Grade IV, and 25 PPTIDs were divided into 2 Grades: 9 were Grade II consisting of strong immunolabeling for neurofilaments and fewer than 6 mitoses; 16 were Grade III with either 6 or more than 6 mitoses or fewer than 6 mitoses but without immunostaining for neurofilaments (Table 1). The percentages of HIOMT-positive cells significantly decreased as follows (Fig. 5): 81.8% ± 5.6% in Grade I, 53.3% ± 6.1% in Grade II, 26.3% ± 5.5% in Grade III, and 6.2% ± 1.5% in Grade IV, respectively. The PPT cases were 30.4% SAG-positive overall (Grade I, 50.0% [4/8 cases]; Grade II, 22.2% [2/9 cases]; Grade III, 31.3% [5/16 cases]; Grade IV, 23.1% [3/13 cases]). The mean percentage of SAG-positive cells in PPTs was 5.0% ± 1.7% (Grade I, 1.4% ± 0.3%; Grade II, 6.1% ± 1.9%; Grade III, 8.2% ± 4.5%; Grade IV, 3.9% ± 2.3%). The percentages of Ki-67-positive cells were 1.1% ± 0.3% in Grade I (Fig. 3g), 8.6% ± 2.1% in Grade II (Fig. 3m), 19.7% ± 3.6% in Grade III (Figs. 3j, p), and 50.1% ± 5.9% in Grade IV (Fig. 3s) PPTs, respectively. The Ki-67 labeling index of PBs was significantly high compared with those of PC and PPTID (Jouvet Grades II and III).

FIGURE 5.

Correlation of the percentages of hydroxyindole-O-methyltransferase (HIOMT)-immunoreactive cells in pineal parenchymal tumors with Jouvet prognostic grade scheme. The mean percentages of HIOMT-immunoreactive cells were, in approximate order of significance decreased 81.8% ± 5.6% in Grade I, 53.3% ± 6.1% in Grade II, 26.3% ± 5.5% in Grade III, and 6.2% ± 1.5% in Grade IV PPTs.

In 1 of 6 PNETs and 23 of 42 MBs, there were both HIOMT-positive and SAG-positive tumor cells. Percentages of HIOMT-positive and SAG-positive cells in 24 patients of PNET/MB were 3.2% ± 0.9% and 5.8% ± 1.7%, respectively. Double-immunofluorescence labeling revealed that MB/PNETs had few HIOMT-positive SAG-negative cells and multifocal clusters of HIOMT-positive SAG-positive cells (Fig. 4).

Discussion

Our IHC results confirmed previously reported HIOMT expression in the pineal gland but less so in the retina, nervous system, intestine, liver, kidney, muscle, and testis (26, 27) and further demonstrated expression in specific cell types. Thus, the data support results showing HIOMT gene expression in humans and animals: melatonin-producing cells are found in the gastrointestinal tract, liver, kidney, adrenal glands, thyroid gland, and lymphatic system (28-32).

Microglia are primarily CNS-immune effector cells that have major roles in the defense and repair of CNS tissues in response to injury (33). In addition to phagocytic activity, they release a variety of cytotoxic substances (e.g. H2O2, NO, proteases, glutamate, and aspartate) and extracellular signaling molecules (e.g. interferon-γ and tumor necrosis factor) that maintain CNS homeostasis. They express major histocompatibility complex class I/II proteins, and when activated by interferon-γ, they become antigen-presenting cells (34). The detection of HIOMT mRNA in T lymphocytes, B lymphocytes, and natural killer cells has also suggested that melatonin should act directly as a paracrine and/or autocrine agent in the human immunoregulatory system (35, 36). In view of the origin of microglia and their relation to the immune system, their expression of HIOMT is not surprising.

The HIOMT-positive neurons were generally found in the CNS but were also found in the Edinger-Westphal nucleus, the cholinergic parasympathetic nucleus that projects to the ciliary ganglion and regulates pupil function. Choline acetyltransferase-negative nonpreganglionic neurons containing urocortin 1, cocaine- and amphetamine-regulated transcript, cholecystokinin, neuropeptide B, substance P, and Period 2 are present in the dorsomedial area of the Edinger-Westphal nucleus and do not innervate the ciliary ganglion (37-43). Urocortin 1 neurons in the nonpreganglionic Edinger-Westphal nucleus (npEW) are involved in stress adaptation processes and are recruited by various acute stressors. Urocortin 1 mRNA expression is upregulated by acute pain and stress (40, 44-46). The neuropeptides cocaine- and amphetamine-regulated transcript (47), cholecystokinin (48, 49), neuropeptide B (50), and substance P (51) are also thought to be associated with stress, anxiety, appetite, and mood. These data suggest that the npEW plays a central role in stress adaptation (46). Period 2 is an important gene involved in the regulation of the major circadian clock in the mammalian CNS, the suprachiasmatic nucleus, and the npEW (43, 52). Further investigations are necessary to delineate the interactions between melatonin and these neuropeptides in npEW neurons to understand stress adaptation and the circadian clock in humans.

The pineal gland is virtually ubiquitous throughout the vertebrate animal kingdom. In nonmammalian vertebrates, it functions as a photoreceptive third eye and an endocrine organ. In mammals, it serves as an endocrine organ regulated by light entering the body via the eyes. In the normal human pineal gland, the photoreceptor protein SAG is found in approximately 5% to 10% of pineocytes (53); we found them in 7.6% in this study. However, three quarters of SAG-positive cells had HIOMT in their cytoplasm. Melatonin (54-56) and retinal SAG (57, 58) have the common mechanism of their actions in the interaction of G protein-coupled receptor signaling. More studies are necessary to understand the relationship of melatonin and retinal SAG in the control of G protein-coupled receptor signaling.

All cases of PPTs contained HIOMT-immunoreactive melatonin-producing cells. This is consistent with previous reports on HIOMT activity in an experimental pineal tumor induced by a human papovavirus (59), in a biopsy specimen of a human PC (20), postmortem analysis of HIOMT in PC (60), and the presence of serotonin in PC (61). These data suggested to us that HIOMT expression would be useful to evaluate the histological features in PPTs. Indeed, we found that the mean percentages of HIOMT-positive cells decreased with increasing Jouvet grade (1), suggesting that HIOMT IHC may be a prognostic indicator in PPTs. The antibodies we developed and used in this study allowed demonstration of HIOMT in formalin-fixed paraffin-embedded tissues, thereby allowing assessment on a large number of routinely processed archival samples and potential wide application.

The pineal gland functions as a photoreceptor organ in lower animals and retains vestiges of photoreceptor organ in the embryonic period. The expression of several photoreceptor proteins, in particular, rhodopsin, retinal SAG, and cellular retinaldehyde-binding protein, has been described in PPTs (16,61-63). Thus, SAG IHC should be useful for making a diagnosis of PPTs. We found that the mean percentages of SAG-immunoreactive cells in PPTs with SAG-immunoreactive cells were low, but absent SAG immunoreactivity could not rule out the diagnosis of PPT. On the other hand, 24 of 48 PNETs/MBs had SAG-immunoreactive cells, in agreement with previous reports on retinal SAG in 27% to 50% of MBs (64-70). Immunohistochemistry studies of rhodopsin, rod-opsin, and interphotoreceptor retinoid-binding protein and electron microscopy studies have also shown photoreceptor cell differentiation in MBs (68, 70, 71). In the pineal gland (Figs. 2q-s) and embryonal tumors (Figs. 4i-k), HIOMT-positive SAG-immunonegative cells would have pineocytic differentiation; HIOMT-negative and retinal SAG-positive cells would have the differentiation of pineocytes with photoreceptor protein.

Based on their expression of HIOMT and SAG, PBs could be grouped as PPT, but PBs resemble other embryonal tumors of the CNS (PNETs and MBs) in many respects. There is increasing evidence that PNETs and MBs might derive from progenitor cells in the subependymal matrix zone and/or the external granule cell layer of the cerebellum (72-75). The PB, PPTID, and PC may also originate from multipotent neural stemlike cells that lead to pineocytes (5,76). Low Ki-67 labeling indices of PCs and PPTIDs compared with those of the PBs and the significantly different expression of HIOMT suggest that the brain cancer stem cells of origin might be different for different PPT types.

The distinction of high-grade PPTs from other PNETs has relied heavily on specific knowledge of the primary site of the lesion and evidence of pineocytomatous differentiation, such as club-shaped argyrophilic processes, pineocytomatous rosettes, and photoreceptor expression. When the primary site is clinically obscure and without histological findings suggesting pineocytomatous differentiation, a precise diagnosis may not be possible. Our data suggest that HIOMT and retinal SAG IHC might support the differential diagnosis of these neoplasms (Fig. 6). The finding that all PPTs had HIOMT-positive cells suggests that no immunoreactivity of HIOMT might indicate PNET/MBs rather than a high-grade PPT. The HIOMT-positive and SAG-immunonegative cells were observed in PPTs, but seldom in PNET/MBs, suggesting that finding HIOMT-positive SAG-immunonegative cells would argue in favor of a PPT. The HIOMT-positive and SAG-positive tumor cells tended to be sparse in PPTs and were in clusters in PNET/MBs. Photoreceptor protein differentiation might occur randomly in each tumor cell of PPTs, whereas monoclonal proliferation of pineocytic tumor cells with photoreceptor protein might exist multifocally in PNET/MBs. Multifocal clustered HIOMT-positive and SAG-positive tumor cells might, therefore, support the diagnosis of PNET/MB.

FIGURE 6.

Differential diagnoses of pineal parenchymal tumors (PPTs), primitive neuroectodermal tumors (PNET), and medulloblastomas (MBs) based on the results of this study.

Acknowledgments

The authors thank Kumiko Iwabuchi, Naoko Takabayashi, and Youko Natake for excellent technical assistance.

References

  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.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
View Abstract