Identification of three hydroxyproline O-arabinosyltransferases in Arabidopsis thaliana.

Hydroxyproline (Hyp) O-arabinosylation is a post-translational modification that is prominent in extracellular glycoproteins in plants. Hyp O-arabinosylation is generally found in these glycoproteins in the form of linear oligoarabinoside chains and has a key role in their function by contributing to conformational stability. However, Hyp O-arabinosyltransferase (HPAT), a key enzyme that catalyzes the transfer of the L-arabinose to the hydroxyl group of Hyp residues, has remained undiscovered. Here, we purified and identified Arabidopsis HPAT as a Golgi-localized transmembrane protein that is structurally similar to the glycosyltransferase GT8 family. Loss-of-function mutations in HPAT-encoding genes cause pleiotropic phenotypes that include enhanced hypocotyl elongation, defects in cell wall thickening, early flowering, early senescence and impaired pollen tube growth. Our results indicate essential roles of Hyp O-arabinosylation in both vegetative and reproductive growth in plants.

Expression of the mammalian X chromosome before and after fertilization.

The activity of hypoxanthine-guanine phosphoribosyltransferase in unfertilized mouse ova and in mouse embryos at the two-cell stage is proportional to the number of X chromosomes present during oogenesis. This indicates that the enzyme is X-linked in the mouse and that inactivation of the X chromosome does not occur during oogenesis. However, the genetic dosage effect of the X chromosomes is not present after the increase in hypoxanthine-guanine phosphoribosyltransferase activity in the late morula and the blastocyst stages. These results indicate that the X-linked enzyme lacuts is expressed sometimne after fertilization but before the morula stage.

Non-invasive molecular and functional imaging of cytosine deaminase and uracil phosphoribosyltransferase fused with red fluorescence protein.

INTRODUCTION: Increased expression of cytosine deaminase (CD) and uracil phosphoribosyltransferase (UPRT) may improve the antitumoral effect of 5-fluorouracil (5-FU) and 5-fluorocytosine (5-FC), and thereby enhance the potential of gene-directed enzyme prodrug therapy. For the applicability of gene-directed enzyme prodrug therapy in a clinical setting, it is essential to be able to monitor the transgene expression and function in vivo. Thus, we developed a preclinical tumor model to investigate the feasibility of using magnetic resonance spectroscopy and optical imaging to measure non-invasively CD and UPRT expression and function. MATERIALS AND METHODS: Expression vectors of CD or CD/UPRT fused to monomeric DsRed (mDsRed) were constructed and rat prostate carcinoma (R3327-AT) cell lines stably expressing either CD/mDsRed or CD/UPRT/mDsRed were generated. The expression of the fusion proteins was evaluated by flow cytometry, fluorescence microscopy, and Western blot analysis. The function of the fusion protein was confirmed in vitro by assessing 5-FC and 5-FU cytotoxicity. In vivo fluorine-19 magnetic resonance spectroscopy ((19)F MRS) was used to monitor the conversion of 5-FC to 5-FU in mice bearing the R3327-CD/mDsRed and R3327-CD/UPRT/mDsRed tumor xenografts. RESULTS: Sensitivity to 5-FC and 5-FU was higher in cells stably expressing the CD/UPRT/mDsRed fusion gene than in cells stably expressing CD/mDsRed alone or wild-type cells. Whole tumor (19)F MRS measurements showed rapid conversion of 5-FC to 5-FU within 20 min after 5-FC was administered intravenously in both CD/mDsRed and CD/UPRT/mDsRed tumors with subsequent anabolism to cytotoxic fluoronucleotides (FNucs). CD/UPRT/mDsRed tumor was more efficient in these processes. CONCLUSION: This study demonstrates the utility of these tumor models stably expressing CD or CD/UPRT to non-invasively evaluate the efficacy of the transgene expression/activity by monitoring drug metabolism in vivo using MRS, with potential applications in preclinical and clinical settings.

Proteomic analysis reveals metabolic changes during yeast to hypha transition in Yarrowia lipolytica.

Fungal dimorphism is important for survival in different environments and has been related to virulence. The ascomycete Yarrowia lipolytica can grow as yeast, pseudomycelial or mycelial forms. We have used a Y. lipolytica parental strain and a Deltahoy1 mutant, which is unable to form hypha, to set up a model for dimorphism and to characterize in more depth the yeast to hypha transition by proteomic techniques. A two-dimensional gel electrophoresis (2-DE) based differential expression analysis of Y. lipolytica yeast and hyphal cells was performed, and 45 differentially expressed proteins were detected; nine with decreased expression in hyphal cells were identified. They corresponded to the S. cerevisiae homologues of Imd4p, Pdx3p, Cdc19, Sse1p, Sol3p, Sod2p, Xpt1p, Mdh1p and to the unknown protein YALIOB00924g. Remarkably, most of these proteins are involved in metabolic pathways, with four showing oxidoreductase activity. Furthermore, taking into account that this is the first report of 2-DE analysis of Y. lipolytica protein extracts, 35 more proteins from the 2D map of soluble yeast proteins, which were involved in metabolism, cell rescue, energy and protein synthesis, were identified.

Pyrimidine nucleoside phosphorylase and dihydropyrimidine dihydrogenase activities as predictive factors for the efficacy of doxifluridine together with mitomycin C as adjuvant chemotherapy in primary colorectal cancer.

BACKGROUND/AIMS: Pyrimidine Nucleoside Phosphorylase (PyNPase) converts 5'-deoxy-5-fluorouridine (5'-DFUR, doxifluridine) to 5-fluorouracil (5-FU). While this reaction is taking place Dihydropyrimidine Dihydrogenase (DPD) catalyzes 5-FU to inactive molecules. Mitomycin C (MMC) elevates the PyNPase level in tumor cells. METHODOLOGY: We investigated 17 colorectal cancer patients' PyNPase and DPD activities in tumor and normal tissues using an enzyme-linked immunosorbent assay (ELISA) to assess their clinical significance as indicators for selecting colorectal cancer patients for 5'-DFUR together with MMC as adjuvant chemotherapy. RESULTS: Six of 17 patients developed experienced a recurrence. Tumor DPD activity of the 6 patients who had a recurrence were higher than those of the 11 patients with no recurrence (p = 0.047). On the other hand, there were no significant differences in both the PyNPase and the PyNPase/DPD (P/D) ratio between the group with recurrence and the group without recurrence. For survival analyses, we designed the cut-off value of tumor PyNPase, DPD and P/D ratio as their median value and classified patients into a higher group and a lower group, but there were no significant differences between the groups. CONCLUSIONS: The DPD activity in the tumor may be a useful indicator for selecting patients likely respond to 5'-DFUR together with MMC as adjuvant chemotherapy. If tumor DPD is high, we had better select a different anticancer drug.

Monoclonal antibody and enzymatic profiles of human malignant T-lymphoid cells and derived cell lines.

Recently, four distinct cell lines were established from patients whose malignancies had been defined by immunological and biochemical markers. Each patient had a distinct subtype of a T-cell cancer, and each possessed elevated adenosine deaminase and reduced nucleoside phosphorylase activity. Cell lines cultured in vitro possessed the same basic immunophenotype and biochemical enzyme activity as the patients' original malignant cells. In a direct comparison of the immunophenotype of the cell lines and the patients' malignant cells, full concordance existed for 48 of 52 paired antibody tests performed. However, when compared to the corresponding patient's sample, each cell line showed some minor changes in antigen expression or enzyme level. Antigen loss, de novo antigen expression, or elevated adenosine deaminase levels occurred in the cell lines, and these changes were stable on repeated analysis. While there was good general concordance between the patient's cancer and the established cell line, minor biological differences in the cell lines may reflect cellular maturation or subpopulation selection in vitro.

Detection of NAD:arginine ADPribosyltransferases in animal tissues using 125I-labeled 1-(p-hydroxyphenyl) 2-guanidinoethane as ADPribose acceptor.

125I-labeled 1-(p-hydroxyphenyl) 2-guanidinoethane (N-guanyltyramine), previously used to assay for the bacterial toxin choleragen (Mekalanos, J.J., Collier, R.J. and Romig, W.R. (1979) J. Biol. Chem. 254, 5849-5854) was utilized to identify NAD:arginine ADPribosyltransferases in animal tissues. The use of this radiolabelled ADPribose acceptor, rather than radiolabelled NAD, would bypass the problem posed by the almost ubiquitous presence of enzymes that degrade NAD. With a homogeneous ADPribosyltransferase from turkey erythrocytes, NAD and 125I-labeled guanyltyramine as ADPribose acceptor, formation of ADPribosyl 125I-guanyltyramine was linear with time and enzyme concentration. The product was indistinguishable on both thin-layer and high-performance liquid chromatography from that formed by choleragen. Using 125I-guanyltyramine, ADPribosyltransferase activity was also demonstrated in crude turkey erythrocyte cytosolic and membrane fractions. When rat liver was fractionated, apparent activity was detected primarily in the microsomes. The NAD-dependent product of the microsomal reaction was, however, distinguished from the turkey erythrocyte transferase product by thin-layer and DEAE-Sephadex chromatography; this product had a retention time identical to that of free 125I on high-performance liquid chromatography. In addition to NAD, the microsomal deiodinase activity was supported by NADH, NADP and NADPH. Phenyl boronate selectively bound ADPribosyl 125I-guanyltyramine and other metabolites of 125I-guanyltyramine which were formed by microsomes in a NAD-dependent process. These metabolites were distinguished from ADPribosyl 125I-guanyltyramine by high-performance liquid chromatography. These results indicate that in some cases, for example, turkey erythrocyte cytosolic and membrane fractions, 125I-guanyltyramine can be used to quantify ADPribosyltransferases in crude mixtures, whereas in others, for example, rat liver microsomes, high-performance liquid chromatographic analysis must be used to identify products.

Activation of glutaminase by phosphoribosyl-pyrophosphate and its interference with the assay of phosphoribosylpyrophosphate amidotransferase.

Phosphate-dependent glutaminase (L-glutamine amidohydrolase, EC 3.5.1.2) from rat liver was found to be strongly activated by phosphoribosylpyrophosphate (P-rib-PP), the substrate of amidophosphoribosyltransferase (EC 2.4.2.14). Since the assay of the latter is based on the P-rib-PP-dependent conversion of glutamine to glutamate, the amidotransferase activities determined in crude tissue preparations were found to be too high. The interference of glutaminase, however, could be completely eliminated by its inactivation at 50 degrees C. Amidotransferase was not affected by the heat treatment. Because of the increased rate of the glutamate formation at this temperature, the incubation time of the assay could be significantly reduced.

Purification and tissue distribution of human thymidine phosphorylase; high expression in lymphocytes, reticulocytes and tumors.

Thymidine phosphorylase (dThdPase) is an enzyme involved in pyrimidine nucleoside metabolism, but little is known about its physiological functions. We purified dThdPase from human placenta and used it for antibody preparation. The purified material appears as a single band at 55,000 dalton on sodium dodecylsulfate-polyacrylamide gel electrophoresis. We obtained a specific antibody raised in rabbits that detected a single polypeptide with a molecular weight of 55,000 dalton in the post nuclear homogenates of several human tissues, on immunoblotting. Using the same technique, dThdPase was highly expressed in the liver, lung, spleen, lymph nodes and peripheral lymphocytes. Immunohistochemical staining revealed that macrophage-like cells contained a much higher amount of dThdPase than parenchymal cells in the liver and lung. dThdPase was found to be highly expressed in T- and B-cell-type malignant lymphoma cells, but low in lymphoblastic and myeloblastic leukemia cells. We also found that carcinomas in the stomach, colon and ovary contained higher amounts of this enzyme than non-neoplastic regions of the tissues. These data suggest that dThdPase plays a role in proliferation and/or differentiation of leukocytes and in cancer proliferation.

The locus for human adenine phosphoribosyltransferase on chromosome no. 16.

Evidence for assigning the locus determining the structure of adenine phosphoribosyltransferase (APRT) to human chromosome No. 16 is presented. Hybrids of APRT-deficient mouse cells and of human fibroblasts having normal APRT were isolated by fusing the parental cells with Sendai virus, blocking de novo purine nucleotide synthesis with azaserine and selecting for hybrids that could use exogenous adenine. The hybrid clones that were studied had only APRT activity that was indistinguishable from human APRT with regard to electrophoretic migration and reaction with antibodies against the partially purified human enzyme. No. 16 was the only human chromosome consistently present in all of the clones, and in one clone, it was the only human chromosome detected. Selection against hybrid cells with 2,6-diaminopurine (DAP) yielded DAP-resistant survivors that lacked chromosome No. 16. One hybrid that originally had an intact No. 16 yielded adenine-utilizing subclones that lacked No. 16 but had a new submetacentric chromosome. The distribution of centromere-associated heterochromatin and the fluorescence pattern indicated that this chromosome consisted of a mouse telocentric chromosome and the long arm of No. 16. Cells having the submetacentric chromosome had human APRT. Both the enzyme and the chromosome were absent in DAP-resistant derivatives. These results suggest that the structure of APRT is defined by a locus on the long arm of human chromosome No. 16.

[A change of tumor PyNPase level at intraperitoneal and intravenous administrations of paclitaxel].

A level of PyNPase activity was measured after intraperitoneal (ip) and intravenous (i.v.) administrations of paclitaxel on the animal model. Nude mice received the subcutaneous implantation of WiDr cells. About 3 weeks later, the ip and i.v. administrations of paclitaxel were performed 2 times at 20 mg/kg and 15 mg/kg, respectively. About 1 week later, the mice were sacrificed. The level of PyNPase activity was measured by the ELISA method. The level of PyNPase of ip and i.v. was higher than that of the control group, but the level of PyNPase revealed no significant difference between ip and i.v.. This result suggested that intraperitoneal administration of paclitaxel enhanced an efficiency of 5'-DFUR and capecitabine as much as intravenous administration of paclitaxel.

[Clinicopathological significance of pyrimidine nucleoside phosphorylase (PyNPase) and dihydropyrimidine dehydrogenase (DPD) in advanced colorectal cancer].

We examined clinicopathological characteristics and prognoses of seventy advanced colorectal cancer cases by measuring pyrimidine nucleoside phosphorylase (PyNPase) and dihydropyrimidine dehydrogenase (DPD) in tumor and normal tissue. PyNPase activities in cancerous tissue obtained from resected were 82.7 +/- 41.9 U/mg protein, which were significantly higher than 37.2 +/- 24.0 U/mg protein in normal tissue (p < 0.001). On the other hand, DPD activities in cancerous tissue were significantly lower in normal tissue (p < 0.05). In cases with lymphnode metastases, PyNPase activities of cancerous tissue were significantly higher than that of no lymphnode metastases cases (p < 0.05). In cases with grade 2 side-effects or higher by oral adjuvant chemotherapy, DPD activities in normal tissue were significantly lower than that of other cases (p < 0.05). With regard to Dukes' B and C cases that were resected curatively, PyNPase activities of cancerous tissue of higher group's prognosis were worse than that of the lower group. In the group received 5'-DFUR as adjuvant chemotherapy, non-recurrent survival rate of the group exhibiting higher PyNPase activities was better than that of the lower group.

Transient expression in mammalian cells of the bacterial reporter gene encoding mercuric reductase: effects of various regulatory elements.

The effect of several transcriptional regulatory elements on gene expression in mammalian cells was investigated. As a reporter gene we have used the bacterial gene merA coding for the enzyme mercuric reductase. Several plasmids were constructed with different promoter/enhancer sequences (pSV/E, pSV/L, pMT, pRSV or pAd) at the 5' end and different splicing (small intron of the T antigen of SV40 or the second intron of the rabbit beta-globin gene) and/or polyadenylation signals (AEn, ALn or AR beta Gn) at the 3' end of the merA gene. Expression was measured in five different mammalian cell lines. In COS cells the highest level of expression is obtained with pSV/L and the lowest level with pSV/E. In HeLa, CV-1, Ltk-, and CHO cells merA expression is relatively high, under control of pRSV and pMT and relatively low under control of pSV/L and pAd. The introns studied have a negative effect on the expression of merA. The presence of a polyadenylation signal downstream from the gene is essential for its expression. The three different polyadenylation signals studied give a similar stimulatory effect on the level of expression of the merA gene.

Xylose-specific antibodies as markers of subcompartmentation of terminal glycosylation in the Golgi apparatus of sycamore cells.

Antibodies specific for xylose-containing plant complex N-linked glycans are used for indirect immunolocalization of xylosyltransferase in sycamore cells. The use of high pressure freezing and freeze substitution for sample preparation resulted in very good morphological preservation of the different Golgi cisternae. Xylosyltransferase shows a diffuse distribution all over the Golgi stacks and xylosylation appears to be an early processing event that is initiated in the cis Golgi compartment.

Generation of Arabidopsis thaliana plants with complex N-glycans lacking beta1,2-linked xylose and core alpha1,3-linked fucose.

The plant glycosyltransferases, beta1,2-xylosyltransferase (XylT) and core alpha1,3-fucosyltransferase (FucT), are responsible for the transfer of beta1,2-linked xylose and core alpha1,3-linked fucose residues to glycoprotein N-glycans. These glycan epitopes are not present in humans and thus may cause immunological responses, which represent a limitation for the therapeutic use of recombinant mammalian glycoproteins produced in transgenic plants. Here we report the genetic modification of the N-glycosylation pathway in Arabidopsis thaliana plants. Knockout plants were generated with complete deficiency of XylT and FucT. These plants lack antigenic protein-bound N-glycans and instead synthesise predominantly structures with two terminal betaN-acetylglucosamine residues (GlcNAc(2)Man(3)GlcNAc(2)).

Purine nucleoside phosphorylase, a possible histochemical marker for T-cells in man.

A procedure is described for the histochemical detection of purine nucleoside phosphorylase (PNP) activity in circulating lymphocytes of man. The number of PNP-positive cells, as evaluated on smears of Ficoll--Hypaque purified cells, correlated well with the number of E-rosette-forming cells of the same blood samples of healthy and diseased people with normal or abnormal numbers of E-rosettes. In healthy people, the number of PNP-positive cells was within the range of 70-80% of the total lymphocyte population, whilst the corresponding E-rosette-forming cells were scored between 60-75%. Patients with unusually low or high E-rosettes had equally low or high numbers of PNP-reactive cells. More substantial evidence for the presence of PNP activity in T-cells and not in B cells was gathered from experiments in which PNP activity and surface membrane immunoglobulins (SMIg) were simultaneously demonstrated on the same preparation. These results showed, on the one hand, that the bulk of lymphocytes that are reactive for PNP do not reveal SMIg and, on the other hand, that most Ig-bearing cells were unreactive for PNP.

Immunofluorescence: a sensitive and rapid method for the detection of purine nucleoside phosphorylase in single cells.

Indirect immunofluorescence was used as a rapid, sensitive and specific method for the visualization of the enzyme purine nucleoside phosphorylase in single cells. The enzyme was localized throughout cytoplasm of human lymphoblasts and fibroblasts but not in cell nuclei. This method is valuable for the detection of mutant enzyme protein in cell-mediated immunodeficiencies caused by purine nucleoside phosphorylase deficiency since it does not rely on enzyme activity. It requires only a limited number of cells and can therefore be used for the rapid screening for the presence of cross-reactive protein in immunodeficiency diseases.

Enzymes of pyrimidine biosynthesis in Crithidia luciliae.

Evidence has been obtained for the presence of enzymes of both the de novo and salvage pyrimidine pathways in the protozoan parasite, Crithidia luciliae. Carbamyl phosphate synthetase-II activity could not be unequivocally demonstrated in crude extracts. However, a distinct peak of activity with a molecular weight of approximately 500 000 was observed following chromatography on Sepharose CL-6B. The enzyme preferentially utilised glutamine with respect to ammonia. It was inhibited by UTP and 5-phosphoribosyl-1-diphosphate had a small activating effect. Carbamyl phosphate synthesis by a 'phosphorolytic' citrullinase could not be demonstrated. The ensuing three de novo enzymes could also be separated on Sepharose CL-6B. Approximate molecular weights were estimated: aspartate transcarbamylase (150,000); dihydroorotase (90,000) and dihydroorotate dehydrogenase (70,000). As reported previously, orotate phosphoribosyltransferase and orotidylate decarboxylase were particulate, being associated with the glucosome. Activities of the salvage enzymes, uracil phosphoribosyltransferase, uridine phosphorylase and uridine nucleosidase were observed. All enzymes were cytoplasmic. No uridine kinase activity was detected.

Localization of quinolinic acid metabolizing enzymes in the rat brain. Immunohistochemical studies using antibodies to 3-hydroxyanthranilic acid oxygenase and quinolinic acid phosphoribosyltransferase.

Specific antibodies raised in rabbits against 3-hydroxyanthranilic acid oxygenase (EC 1.13.11.6) and quinolinic acid phosphoribosyltransferase (EC 1.13.11.6) and quinolinic acid phosphoribosyltransferase (EC 2.4.2.19) were used in immunohistochemical studies to map the cellular localization of the quinolinic acid metabolizing enzymes in the adult male rat brain. 3-Hydroxyanthranilic acid oxygenase immunoreactivity was found to be present in glial cells of presumed astroglial identity, as judged by co-localization with glial fibrillary acidic protein. 3-Hydroxyanthranilic acid oxygenase-immunoreactive glial cells were present in all brain regions and within major fiber tracts. The density of 3-hydroxyanthranilic acid oxygenase-immunoreactive glial cells as well as the intensity of staining of these cells differed among brain regions. In general, telencephalic acid diencephalic areas harbored a larger number of 3-hydroxyanthranilic acid oxygenase-positive cells than did mesencephalic regions. In the former regions the caudate nucleus, septum, nucleus accumbens, neocortex and hippocampus were particularly enriched in 3-hydroxyanthranilic acid oxygenase-immunoreactive cells. In the thalamus, regional differences were noted with regard to the intensity of staining among glial cells with high densities of 3-hydroxyanthranilic acid oxygenase cells in the anteroventral, reticular and ventromedial nuclei. In the inferior and superior colliculi, numerous 3-hydroxyanthranilic acid oxygenase-positive glial cells were found in all layers. In the hypothalamus, 3-hydroxyanthranilic acid oxygenase-immunoreactive glial cells were encountered in the zona incerta, the lateral hypothalamic area, the caudal preoptic region and in the dorsomedial nucleus. In the mesencephalon, the substantia nigra contained numerous, moderately stained cells. At caudal levels of the brain-stem, a relatively large number of cells was detected in the nucleus of the solitary tract, the pontine nucleus and in the fascial nerve nucleus, while other nuclei, such as the reticular formation and the area postrema were relatively poor in 3-hydroxyanthranilic acid oxygenase-immunoreactive cells. In addition to staining of glial cells, neuronal cell bodies containing 3-hydroxyanthranilic acid oxygenase immunoreactivity were detected in the main and in the accessory olfactory bulb, as well as in the ventromedial nucleus of the hypothalamus. Quinolinic acid phosphoribosyltransferase immunoreactivity was observed within glial cells and in association with neuronal cell bodies. Some, but not all, quinolinic acid phosphoribosyltransferase positive glial cells contained glial fibrillary acidic protein (Köhl

Kynurenine pathway enzymes in a rat model of chronic epilepsy: immunohistochemical study of activated glial cells.

The kynurenine pathway metabolites quinolinic acid and kynurenic acid have been hypothetically linked to the occurrence of seizure phenomena. The present immunohistochemical study reports the activation of astrocytes containing three enzymes responsible for the metabolism of quinolinic acid and kynurenic acid in a rat model of chronic epilepsy. Rats received 90 min of patterned electrical stimulation through a bipolar electrode stereotaxically positioned in one hippocampus. This treatment induces non-convulsive limbic status epilepticus that leads to chronic, spontaneous, recurrent seizures. One month after the status epilepticus, the rats showed neuronal loss and gliosis in the piriform cortex, thalamus, and hippocampus, particularly on the side contralateral to the stimulation. Astrocytes containing the kynurenic acid biosynthetic enzyme (kynurenine aminotransferase) and the enzymes for the biosynthesis and degradation of quinolinic acid (3-hydroxyanthranilic acid oxygenase and quinolinic acid phosphoribosyltransferase, respectively) became highly hypertrophied in brain areas where neurodegeneration occurred. Detailed qualitative and quantitative analyses were performed in the hippocampus. In CA1 and CA3 regions, the immunostained surface area of reactive astrocytes increased up to five-fold as compared to controls. Enlarged cells containing the three enzymes were mainly observed in the stratum radiatum, whereas the stratum pyramidale, in which neuronal somata degenerated, showed relatively fewer reactive glial cells. Hypertrophied kynurenine aminotransferase- and 3-hydroxyanthranilic acid oxygenase-immunoreactive cells were comparable in their morphology and distribution pattern. In contrast, reactive quinolinic acid phosphoribosyl transferase-positive glial cells displayed diversified sizes and shapes. Some very large quinolinic acid phosphoribosyl transferase-immunoreactive cells were noticed in the molecular layer of the dentate gyrus. In the hippocampus, the number of immunoreactive glial cells increased in parallel to the hypertrophic responses. In addition, pronounced increases in immunoreactivities, associated with hypertrophied astrocytes, occurred around lesioned sites in the thalamus and piriform cortex. These findings indicate that kynurenine metabolites derived from glial cells may play a role in chronic epileptogenesis.