QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Goldsmith, P. C. Search for Related Content PubMed PubMed Citation Articles by Goldsmith, P. C.

---

Supplement

Neuroglial Responses to Elevated Glutamate in the Medial Basal Hypothalamus of the Infant Mouse^{1 ,2}

Reproductive Endocrinology Center, Department of Ob/Gyn and Reproductive Sciences, University of California, San Francisco, CA 94143-0556

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS Immunocytochemical staining Glutamate penetration: co... DISCUSSION REFERENCES

 
Elevated plasma glutamate can cause selective loss of neurons in the brains of infant mice. The arcuate nucleus-median eminence region exhibits the greatest sensitivity to glutamate while it undergoes developmental maturation during early postnatal life. To investigate glutamate-induced cellular responses, groups of nursing 7-d-old mice (n = 31–93) were given single subcutaneous injections of 0.1–0.5 mg monosodium glutamate (MSG)/g body wt or an equivalent volume (30–50 µL) of water vehicle (n = 93). Injection of 0.2 mg MSG/g body wt produced a 16-fold rise in plasma glutamate after 15 min (2.10 vs. 0.122 mmol/L control) and was the lowest harmful dose tested. It not only induced injury of small bilateral groups of medial basal hypothalamic neurons at 5 h postinjection, but also enhanced their expression of the N-methyl-D-aspartate (NMDA)R1 glutamate receptor subunit. Higher dosages of 0.3–0.5 mg MSG/g body wt yielded dose-related increases in NMDAR1 staining intensity and larger numbers of damaged neurons within the ventromedial arcuate nucleus. Administration of the live-cell nuclear stain bis-benzimide (0.95 µmol/L) indicated that MSG accessed the entire brain (n = 20) and methylene blue (1.0 g/L) permeated extracellular spaces by 15 min postinjection (n = 19), before cell death was evident (0.75 mmol/L propidium iodide) from co-injected MSG. Immunostaining, which mimicked that for glial fibrillary acidic protein, suggested that glutamate was retained in tanycytes. We conclude that elevated plasma glutamate induces glutamate receptor expression during selective injury of ventromedial arcuate neurons and propose that by sequestering glutamate, tanycytes may amplify local concentrations and promote neuronal damage in infant mice.

KEY WORDS: • monosodium glutamate • arcuate nucleus • median eminence • glutamate receptor • CD1 mouse

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS Immunocytochemical staining Glutamate penetration: co... DISCUSSION REFERENCES

 
Glutamate is a powerful amino acid neurotransmitter that plays a pivotal role in the formation of synapses and neuronal circuitry, long-term potentiation and depression, and both normal learning and addictive behavior (Pelligrini-Giampetro et al. 1997 , Wickelgren 1998 ). On the other hand, excessive glutamate activation contributes to a wide range of neurological insults including ischemia, trauma and epileptic seizures, and to several chronic neurodegenerative disorders in the human brain (Pelligrini-Giampetro et al. 1997 ). Glutamate induces neuronal death principally by increasing the influx of Ca²⁺ through permeable N-methyl-D-aspartate (NMDA)³ and -amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) selective glutamate receptor (GluR) channels (Ankacrona et al. 1995 , Bonfoco et al. 1995 , Brorson et al. 1994 , Choi and Rothman 1990 ), leading to generation of free radicals, activation of proteases, phospholipases and endonucleases, and transcriptional activation of apoptotic programs (Pelligrini-Giampetro et al. 1997 ).

Regions of the brain most susceptible to glutamate exposure are the specialized neurohemal contact zones found at several sites between the cerebral ventricles and the external surface of the brain (Olney 1979 , Weindl and Joynt 1972 ). These highly vascularized areas sit outside the blood-brain barrier, which is formed at their inner surfaces by tight junctions between the modified astroglial cells called tanycytes lining their ventricular boundaries. Most sensitive among them is the median eminence (ME), which receives axon terminals from the nearby arcuate nucleus (ARC) and other hypothalamic neurosecretory neurons. The fenestrated capillary endothelium of the ME renders it totally accessible to plasma-borne amino acids, so that initial glutamate-induced neuronal damage most likely results from circulating levels rather than cerebroventricular pools.

The ARC-ME region of the early postnatal rodent has often been used for studies of monosodium glutamate (MSG) neurotoxicity (Holzwarth-McBride et al. 1976 , Olney 1979 , Reynolds et al. 1976 , Takasaki et al. 1979 ) because of its heightened responsiveness, consistent cytoarchitecture and conspicuous anatomical landmarks (Heywood and Worden 1979 , Olney 1979 , Weindl and Joynt 1972 ). Although the tanycytic framework is already established in neonatal mice, ARC and other neuronal axons continue to invade the ME during the first 25 d of postnatal life (Eurenius and Jarskar 1971 ). Administration of high doses of MSG to immature infants causes no obvious harm to tanycytes or ME axon terminals, but neurons in the ARC nucleus can be severely damaged, with most deteriorating within 6 h (Holzwarth-McBride et al. 1976 , Olney 1979 , Reynolds et al. 1976 ). Because many early studies employed excessive and destructive amounts of MSG, however, little was learned about the events leading to neuronal death. To help understand mechanisms involved in selective neuronal injury during this developmental period, we investigated the effects after exposure to low but deleterious amounts of MSG in the maximally sensitive 7-d-old postnatal mouse (Takasaki et al. 1979 ). The aims of these studies were as follows: 1) to determine peak blood glutamate levels after subcutaneous injection of a minimum-effective glutamate dose-range; 2) to characterize the pattern of arcuate neuronal damage in relation to increasing glutamate exposure; and 3) to evaluate our hypothesis that ventral ARC damage occurs if the ventral ARC-ME tanycytes are exposed to effective glutamate concentrations for a sufficient length of time. A report of some of these studies was published previously (Hu et al. 1998 ).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS Immunocytochemical staining Glutamate penetration: co... DISCUSSION REFERENCES

 
Experimental animals.

The experimental protocols here were reviewed and approved by the Committee on Animal Research at the University of California, San Francisco (Approval #A547–10644), and were performed in compliance with the NIH guidelines (NRC 1985) .

Experienced, lactating female CD1 mice were caged individually with a litter of ten 5-d-old, nursing pups (Charles River, Hollister, CA) under standard conditions. Two days later, the cages were brought to the laboratory between 0930 and 1030 h. The 7-d-old mice were chosen in random order, weighed and injected subcutaneously (s.c.) with graded amounts of aqueous MSG or an equivalent volume (30–50 µL) of water alone (n = 93). The MSG dosages administered (mg MSG/g body wt) were as follows: 0.1 (n = 48), 0.2 (n = 31), 0.3 (n = 93), 0.4 (n = 49), 0.5 (n = 83) and 2.0 (n = 18). Mice were encoded with their treatment and reunited with their mothers in the home cages to continue nursing.

Glutamate administration.

To determine the optimal MSG treatment period, neuronal damage in the ARC was assessed hourly after s.c. injection of 0.0, 0.2. 0.5 or 2.0 mg MSG/g body wt. Reproducible microscopic evidence of ARC cell loss was first observed 5 h after injection, a time course consistent with published reports (Takasaki 1978 , Takasaki et al. 1979 ). Accordingly, mothers were removed from their litters 5 h after treatment and given an overdose of sodium pentobarbital by intraperitoneal injection. Pups were selected in random order and their treatment and gender recorded before rapid decapitation with scissors. Intact brains were removed from the cranial cavity within ~30 s and fixed by immersion in 4% paraformaldehyde plus 0.2% glutaraldehyde in 0.1 mol/L sodium phosphate buffer (pH 7.4). Frontal (coronal) tissue sections through the ARC-ME region were cut at a thickness of 70 µm on an oscillating tissue slicer (OTS 3000–03, Electron Microscopy Sciences, Philadelphia, PA) and stored in fixative at 4°C before routine microscopy or immunocytochemistry.

Plasma glutamate measurements.

Using the same protocol, pups chosen at random were decapitated at 15 min postinjection of 0.0, 0.2, 0.3 and 0.5 mg MSG/g body wt. Trunk blood (50–100 µL) was collected individually in 1.5-mL EDTA-coated microcentrifuge tubes and each plasma fraction frozen and stored at -80°C. Plasma glutamate concentrations were determined using a Beckman Model 6300 Amino Acid analyzer (Beckman Coulter Bioresearch, Fullerton, CA) equipped for fluorometric detection of o-phthalaldehyde–derivatized amino acids as previously described (Hu et al. 1998 ). The glutamate concentration (nmol/mL) in each sample and the mean and SEM for each dose were calculated; the differences were tested using one-factor ANOVA by the Fisher protected least significant difference test and the Scheffé F-test (Hu et al. 1998 ), with P < 0.01 considered significant.

Glutamate penetration: co-injection of vital dyes.

Studies were performed in 7-d-old CD1 mice (n = 40) to assess the effective penetration of MSG into the brain. An aqueous solution of 0.94 µmol/L bis-benzimide (Hoechst 33258, Cat. #H-1398, Molecular Probes, Eugene, OR) and 0.73 µmol/L propidium iodide (Cat. # P-1304, Molecular Probes) was used as the MSG vehicle to ensure exact correlation between the amount of dye and MSG (169.12 Da) delivered. Bis-benzimide (533.9 Da, excitation 352 nm; emission 461 nm), a live-cell nuclear-permeant, blue-fluorescent stain, avidly labels the nucleus of every cell contacted. Propidium iodide (688.4 Da, excitation 535 nm; emission 617 nm), a live-cell nuclear-impermeant, red-fluorescent marker, stains only nuclei in dead cells. Mice were decapitated at 15 or 30 min after s.c. injection of the dye vehicle containing 0.0 (control), 0.2, 0.5 and 2.0 mg MSG/g body wt. Brains were fixed by immersion in buffered aldehyde solution, embedded in paraffin and sectioned at 6-µm thickness through the mid-ARC-ME region in the frontal plane; two to three sections were mounted sequentially on microscope slides.

Experiments in additional infant mice (n = 40) employed treatment with the same MSG dosages (0.0, 0.2, 0.5 and 2.0 mg/g body wt) in an aqueous solution of 1.0 g/L methylene blue (Polysciences, Warrington, PA), a supravital stain for axons and terminals which fluoresces red, and 0.75 mmol/L propidium iodide to label nuclei in dead cells. Because an advancing dye front was not evident, it appeared necessary to use a more concentrated dye solution. Subcutaneous injections of 30 µl Paragon Multiple Stain for Frozen Sections (Cat. #PS1301, Paragon C. & C., New York, NY) failed to gain access to the brain (n = 30) even after 60 min. Because MSG must first enter the vascular compartment to reach the ARC-ME, we performed direct intracardiac injections of 100 µL of undiluted Paragon Multiple Stain alone or together with 0.2, 0.3 or 0.5 mg MSG/g body wt (n = 15). Although mice expired within 2 min postinjection, probably due to cardiac tamponade, brains were immediately removed and immersed in fixative, embedded in paraffin, sectioned at a thickness of 6 µm and mounted as before. Ethanol extraction of Paragon Multiple Stain was avoided by examining paraffin-embedded ARC-ME sections by fluorescence microscopy without the standard deparaffinization.

Immunocytochemical staining.

Vibratome sections were rinsed (3 x 10 min) in Tris-HCl buffer plus 0.9% NaCl (TBS), pH 7.4, before and after sequential pretreatments (30 min each) in TBS solutions containing 50 mmol/L glycine, 50% ethanol, 0.1% saponin plus 0.5% hydrogen peroxide, and 3% normal serum or 1% fish gelatin (G-7765, Sigma Chemical, St. Louis, MO). Sections were incubated in primary antibodies against important ARC-ME phenotypic markers for 48–72 h at 4°C on an orbital shaker. The specificity of the mouse antiglutamate antibody (1:500 Glu-2, #22523, Incstar, Stillwater, MN) has been described previously (Goldsmith et al. 1994 , Goldsmith and Thind 1995 ). To identify NMDA receptors, we used a mouse anti-NMDAR1 immunoglobulin G (IgG) (1:250 #54.1, PharMingen, San Diego, CA); to identify AMPA receptors, we employed a mouse anti-GluR2/4 IgG (1:250 #3A11, PharMingen). The specificity of both antibodies has been described previously (Hu et al. 1998 ). In addition, to label intermediate filaments in (astrocytic) tanycytes, we used a goat anti-glial fibrillary acidic protein (GFAP) antibody (1:50 #C-19, sc-6170, Santa Cruz Biotechnology, Santa Cruz, CA), whose specificity was established by the supplier.

After the samples were rinsed well in TBS, immunofluorescent staining for glutamate was completed using 1:100 goat anti-mouse IgG coupled to CY-2 (excitation peak 490 nm; emission peak 505 nm; Biological Detection Systems, Pittsburgh, PA). Immunostaining for GluR and GFAP was continued with the avidin-biotin-peroxidase method (IgG Elite Kits, Vector, Burlingame, CA) or the peroxidase-antiperoxidase technique as previously described (Goldsmith and Thind 1995 ). Biotinyl tyramide signal amplification (TSA-Indirect Signal Amplification Kit, NEW Life Science Products, Boston, MA) was used if required. Antigenic sites were revealed by incubation in 0.05% 3,3'-diaminobenzidine-4HCl (DAB) and 0.01% hydrogen peroxide, without or with 0.03% NiSO₄ (nickel DAB) as the chromogen. Immunostained tissue sections were routinely dehydrated and mounted permanently on glass slides.

Microscopy and analysis.

Hematoxylin and eosin staining and DAB immunocytochemistry were evaluated at low and high magnifications using a Leica DMRB photomicroscope (Leica, Deerfield, IL). Histochemical and immunofluorescence staining were analyzed using epifluorescence and appropriate filter sets for bis-benzimide (blue: wide band UV excitation 340–380 nm; barrier filter 430 nm), for propidium iodide and Paragon Multiple Stain for Frozen Sections (red: narrow band green excitation 530–560 nm, FITC excluded; barrier filter 580 nm) and for CY-2 (green: narrow band blue excitation 450–490 nm; barrier filter 515 nm). Color images were captured using a DEI-470 CCD camera (Optronics, Goleta, CA) and the Scion Image 1.62a version of NIH Image, and brought to grayscale in Adobe Photoshop 4.0 to compose and label figures.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS Immunocytochemical staining Glutamate penetration: co... DISCUSSION REFERENCES

 
Minimal-effective MSG dose range.

The s.c. injection of water or 0.1 mg MSG/g body wt produced no evidence of ARC-ME injury, whereas light microscopic examination of tissue sections taken 5 h after treatment with 0.2 mg MSG/g dose showed initial injury in small bilateral groups of cells close to the lateral recesses of the third ventricle. Signs of cytological damage included reduced cytoplasmic volume, swollen subcellular organelles, pyknotic nuclei and, in some cases, fragmented and condensed nuclear material (Hu et al. 1998 ). These injured cells were presumably neurons because they were immunonegative for the astrocytic marker GFAP (data not shown). Doses of 0.3–0.5 mg MSG/g body wt produced dose-related increases in the number of damaged neurons within an area expanding outward from the initial site. Even after the 0.5 mg dose, however, injured neurons remained confined to the ventromedial ARC nucleus (Fig. 1A ).

View larger version (64K): [in this window] [in a new window]   Figure 1. Frontal vibratome sections through the midlevel of the arcuate nuclei from 7-d-old mice 5 h after a single subcutaneous injection of 0.5 mg monosodium glutamate (MSG)/g body wt. The median eminence is below the lumen of the third ventricle (center). (A) Hematoxylin and eosin staining shows that pyknotic nuclei of damaged neurons are confined to the ventromedial arcuate nucleus (curved line), the same portion of the nucleus traversed by tanycytes. The asterisk (*) indicates the site of initial neuronal injury. (B) 3,3'-Diaminobenzidine-4HCl (DAB) immunostaining for N-methyl-D-aspartate (NMDA)R1 shows enhanced receptor subunit expression in ventromedial arcuate neurons 5 h after a dose of 0.5 mg MSG/g body wt. Increased NMDAR1 expression is seen only within the ventromedial portion of the arcuate nucleus in which damaged neurons are found. Cells in some of this 70-µm thick tissue section appear out of the focal plane. Magnification: (A) and (B), x200.

Because plasma glutamate concentrations had returned to near normal by 5 h post-treatment (data not shown), they were measured at 15 min after s.c. injection when they are reportedly near maximal (Stegink et al. 1979 ). Water-injected controls (n = 10) had a baseline glutamate level of 0.122 ± 0.10 mmol/L, and single doses of 0.2 (n = 8), 0.3 (n = 10) and 0.5 (n = 10) mg MSG/g body wt raised plasma glutamate concentrations to 2.10 ± 0.236, 4.57 ± 0.492 and 6.29 ± 0.257 mmol/L, respectively. Mean 15-min elevations were significantly different from baseline and from each other (P < 0.01), and constituted increases of 16-, 36- and 51-fold, respectively, above the basal level. A single injection of 0.2 mg MSG/g body wt, which produced a rapid 16-fold increase in the plasma glutamate level to 2.10 mmol/L was the lowest dose tested to initiate injury of specific medial basal hypothalamic neurons in these 7-d-old mice.

	Immunocytochemical staining

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS Immunocytochemical staining Glutamate penetration: co... DISCUSSION REFERENCES

 
Our attempts to immunostain the neurons first affected by elevated plasma glutamate were unsuccessful (data not shown); thus their phenotype has yet to be determined. Because GluR have been implicated in causing neuronal injury, however, we immunostained ARC tissue sections for the principal GluR subunits NMDAR1 and GluR2/4. Evidence of expression of either subunit was barely detectable in vehicle-injected controls or in mice fed 0.1 mg MSG/g body wt (Hu et al. 1998 ). Five hours after treatment with the lowest effective dose of 0.2 mg MSG/g body wt, however, we observed increased NMDAR1 and GluR2/4 expression in the perikaryal cytoplasm of neurons near the lateral recesses of the third ventricle (Hu et al. 1998 ). Administration of 0.3–0.5 mg MSG/g body wt produced a dose-related increase in the numbers of compacted NMDAR1-immunopositive neurons in the ventromedial ARC region (Fig. 1B ), which was not apparent for GluR2/4 (Hu et al. 1998 ). NMDAR1 staining first appeared throughout neuronal cytoplasm, intensified as the cytoplasm condensed and finally dissipated as neurons fragmented. NMDAR1-reactive neurons appeared to spread outward from the initial site of expression in the same manner as the progressive ARC damage (Fig. 1B ).

	Glutamate penetration: co-injection of vital dyes

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS Immunocytochemical staining Glutamate penetration: co... DISCUSSION REFERENCES

 
Results using bis-benzimide showed the live-cell DNA stain penetrated far beyond the ARC-ME region and gained access to the entire brain by 15 min after s.c. injection. Its blue fluorescence appeared in all cell nuclei when used alone (control) (Fig. 2A ) or as the vehicle for graded doses of MSG (Fig. 2B ). No nuclei exhibited red fluorescence from propidium iodide (data not shown), suggesting that cell death had not yet occurred at this early time. Fluorescence microscopy showed that the methylene blue–containing vehicle had permeated extracellular spaces of the medial basal hypothalamus at 15 min postinjection, but had not entered cellular elements (Fig. 3A ). It thereby surrounded and outlined the cell bodies and processes of neurons and glia in the ARC-ME, including the curving shafts of tanycytes. Immunostaining for GFAP confirmed that tanycytic cell bodies lined the floor and walls of the ventral third ventricle (Fig. 3B ). Processes of ventral tanycytes branched repeatedly throughout the ME, and terminal "end feet" formed a thin but discontinuous layer along its basal lamina. In contrast, lateral tanycytes extended long and generally unbranched processes ventrolaterally through the ventromedial ARC nucleus (Fig. 3B ). Results using the Paragon Multiple Stain vehicle alone showed that the stain entered the ME and was taken up by tanycytes as early as 0.5 min after intracardiac injection (Fig. 4A ). Specific immunostaining also showed that glutamate was retained preferentially in tanycyte cell bodies and processes at 5 h after s.c. injection of MSG (Fig. 4B ).

View larger version (78K): [in this window] [in a new window]   Figure 2. Equal fluorescence (blue) of all brain cell nuclei is demonstrated by these examples of the hippocampal dentate gyrus from 7-d-old mice 15 min after subcutaneous injection of (A) an aqueous solution of 0.94 µmol bis-benzimide/L vehicle alone (control), or (B) when combined with 0.2 mg monosodium glutamate (MSG)/g body wt. Nuclear staining throughout the dentate gyrus at the time of peak plasma glutamate concentrations was independent of the dose of MSG tested and failed to provide information about the efficiency of the blood-brain barrier in these 7-d-old mice. Magnification: (A) and (B), x250.

View larger version (73K): [in this window] [in a new window]   Figure 3. Fluorescence micrographs of frontal sections showing the lateral recess of the third ventricle (upper right), the median eminence below, and the ventromedial arcuate nucleus of 7-d-old mice after a single subcutaneous injection of 0.3 mg monosodium glutamate (MSG)/g body wt. (A) Red fluorescence due to the methylene blue (1.0 g/L) vehicle permeates extracellular spaces by 15 min after administration, and outlines cells and processes, including shafts of tanycytes arching through the ventromedial arcuate nucleus. (B) Immunofluorescent staining for glial fibrillary acidic protein (GFAP) with CY-2 distinguishes tanycyte cell bodies lining the ventricular walls, and their processes branching extensively within the median eminence as well as arching downward through the ventromedial arcuate nucleus toward the basal lamina. Magnification: (A) and (B), x375.

View larger version (59K): [in this window] [in a new window]   Figure 4. Fluorescence micrographs of frontal sections through the mid-arcuate nucleus level of 7-d-old mice. The median eminence appears below the lumen of the third ventricle (center). (A) Paragon Multiple Stain in vehicle-injected controls appears to be sequestered by tanycytes as early as 0.5 min after intracardiac injection, i.e., on the first pass through the circulation. (B) In mice examined 5 h after subcutaneous injection of 0.3 mg monosodium glutamate (MSG)/g body wt, immunofluorescent staining with CY-2 indicates that glutamate is preferentially taken up by cytoplasmic foot processes and retained in the cell bodies of tanycytes lining the floor and lateral recesses of the third ventricle. Magnification: (A) and (B), x375.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS Immunocytochemical staining Glutamate penetration: co... DISCUSSION REFERENCES

Our results elucidate earlier work showing MSG-induced ARC lesions in infant mice (Holzwarth-McBride et al. 1976 , Takasaki 1978 , Takasaki et al. 1979 ) by correlating MSG doses with peak plasma glutamate levels and the extent of neuronal damage achieved. ARC neuronal damage in mice does not occur if the plasma glutamate level remains <0.5 mmol/L (Stegink et al. 1974 ). Administration of 0.2 mg MSG/g body wt here produced a 15-min plasma glutamate concentration of 2.10 mmol/L, an increase of > 16-fold above baseline (0.122 mmol/L). Injury with this dose is consistent with previous work showing that ARC neuronal damage in mice appears when the plasma level exceeds baseline by ~20-fold (Bizzi et al. 1977 ).

Interestingly, an eightfold rise in the plasma glutamate concentration to 2.37 mmol/L, a value similar to that observed after our lowest effective MSG dose, was observed previously not to affect hypothalamic glutamate levels (McCall et al. 1979 ). Even a 12-fold increase in plasma glutamate was reported to cause only a slight elevation in the glutamate content of the ARC nucleus (~0.12 µmol/mg protein) (Airoldi et al. 1979 ). Because the blood-brain barrier glutamate carrier is nearly saturated under normal circumstances, it may not allow transport of additional glutamate into the brain (McCall et al. 1979 ). Moreover, because glutamate flux out of the brain is at least seven times that into the brain, even large doses of MSG may fail to produce measurable increases in hypothalamic glutamate concentrations (McCall et al. 1979 ).

A careful analysis across the MSG dose range studied suggested to us that the intensity of glutamate-induced NMDAR1 immunostaining was related to the degree of neuronal injury. Although basal receptor expression in controls was barely detectable, effective doses of 0.2–0.5 mg MSG/g body wt increased NMDAR1 staining intensity in a dose-dependent fashion. Enhanced expression was first seen consistently at the site of initial ARC neuronal injury, and advanced outward into the ventromedial ARC nucleus with escalating amounts of MSG. Weak immunostaining, which coexisted with the initial signs of neuronal damage, became progressively intense in neurons at the same location with higher doses of MSG. This increase in NMDAR1 expression apparently was specific because similar dose-related increases in immunostaining intensity were not observed for other substances present in unaffected ARC neurons (e.g., tyrosine hydroxylase) or with incubation in DAB substrate alone.

An increase in neuronal NMDAR1 expression levels has previously been related to ordinary developmental changes, as seen with certain GluR subunits (Seeburg 1993 ) and postsynaptic metabotropic GluR (McDonald et al. 1993 , Smirnova et al. 1993 ). A correlation between NMDAR1 immunostaining and neuronal injury was also reported in an earlier study in which pretreatment with the NMDA receptor-selective antagonist MK-801 blocked glutamate-induced ARC neurotoxicity more effectively than did the AMPA receptor-selective antagonist NBQX (Lehmann and Jonsson 1992 ). Immunostaining evidence of enhanced NMDAR1 subunit expression therefore may serve not only to identify affected neurons, but the intensity of staining may also reflect their state of neurodegeneration.

Our results showed that the highly membrane-permeant "live cell" dye bis-benzimide gained access to the entire brain by 15 min postinjection in 7-d-old mice. Comparable staining of all brain nuclei at this time of peak plasma glutamate concentrations suggested that bis-benzimide entry into the brain was neither dependent upon nor affected by the amount of coadministered MSG. Consequently, these data provided no substantive evidence about the concomitant entry of glutamate. Several reports have suggested that the sodium introduced by much higher dosages of MSG might elevate plasma osmolarity and compromise function of the anatomical blood-brain barrier present in mice from birth (McCall et al. 1979 , Pardridge 1979 ). Increased glutamate concentrations in minute brain areas might promote neuronal injury in the cerebral cortex, cerebellum and amygdala without affecting overall brain glutamate concentrations (McCall et al. 1979 ). However, no such glutamate penetration and damage occurred as a result of the low dose range of MSG used here.

According to our hypothesis, ventral ARC damage occurs if the ventral ARC-ME tanycytes are exposed to effective glutamate (MSG) concentrations for a sufficient length of time. Our results using Paragon Multiple Stain suggested that tanycytes may begin to sequester blood-borne substances very soon after (0.5 min) intracardiac injection, perhaps on their first pass through the hypophysial-portal circulation. Furthermore, immunostaining evidence (see Fig. 4B ) indicated that elevated glutamate concentrations persist in the tanycytic compartment for 5 h after MSG administration, long after peak plasma levels and any ARC increases have abated (Heywood and Worden 1979 ).

One theory of neurotoxicity is based on the assumption that lesions occur only within the ARC nucleus and other regions thought to be unprotected by an anatomic blood-brain barrier (McCall et al. 1979 ). In the ARC-ME region, the lateral limits of the barrier are represented by the cytoplasmic processes of tanycytes, shown arching ventrolaterally through the middle of the ARC nucleus by our methylene blue staining (Eurenius and Jarskar 1971 , Pardridge 1979 , Weindl and Joynt 1972 ). Therefore, only ventromedial ARC neuronal perikarya and their axon terminals in the ME are vulnerable to elevated concentrations of circulating glutamate because they lie outside the blood-brain barrier. Because a functional blood-brain barrier is already present in newborn mice (McCall et al. 1979 ), other developmental conditions must render ventromedial ARC neurons in infant mice highly susceptible to MSG-induced excitotoxicity, whereas the same neurons in mature adults are relatively refractory.

The tanycytic framework of the ARC-ME region is well established in neonatal mice and rats. However, elongating axons of ARC and other neurons invade and terminate in the rodent ME throughout the first 3–4 wk of postnatal life (Eurenius and Jarskar 1971 ). This period, which corresponds to both reproductive maturation and heightened glutamate responsiveness (Holzwarth-McBride et al. 1976 , Takasaki 1978 ), is likely to include a transient episode of inordinately high tanycyte-to-neurite ratios that confer maximal glutamate sensitivity between postnatal d 5 and 10 (Eurenius and Jarskar 1971 ). Elevated GFAP expression in tanycytes may also contribute to ARC neuronal vulnerability before adulthood. Phenotypically mature tanycytes in neonatal mice display GFAP-containing intermediate filaments known to participate in intracellular uptake and transport processes (de Vitry et al. 1981 ). Tanycytic GFAP expression reaches a peak between d 5 and 10 of postnatal life, the period of greatest glutamate sensitivity (de Vitry et al. 1981 ). Ventromedial ARC neurons and axon terminals may therefore experience unusually high glutamate levels in infant mice due to peak GFAP expression in local tanycytes.

Our studies showed a conspicuous correspondence between the "confluence of tanycytic processes" at the initial site of neuronal injury and the restriction of neuronal damage to the portion of the ventromedial ARC nucleus traversed by tanycytes. Although tanycytic glia are resistant to injury themselves (Levi and Patrizio 1992 ), their excessive number and intimate neuroglial association may jeopardize neighboring ventromedial ARC neurons during early postnatal life. Anatomic disruption of the blood-brain barrier by hyperosmotic plasma need not occur; functional disruption, especially of the efflux mechanism, for example, by reversal of tanycytic glutamate transporters (Attwell et al. 1993 , Kanai and Hediger 1992 , Levi and Raiteri 1993 , Pardridge 1979 ), could be sufficient to increase extraneuronal glutamate concentrations in discrete brain areas (McCall et al. 1979 ) and exacerbate injurious conditions.

In summary, our results show that minimally harmful and higher plasma glutamate levels cause dose-related increases in the number of damaged neurons in the medial basal hypothalamus of infant mice. Ventromedial ARC neurons of 7-d-old mice appear particularly sensitive to elevated concentrations of circulating glutamate for the following reasons: 1) they lie outside the blood-brain barrier; 2) the area is traversed by tanycytes; 3) the tanycytes express high levels of GFAP; and 4) there may be particularly high tanyctye-to-neuron/terminal ratios at this time. Enhanced NMDAR1 receptor expression also coincides with the initial injury of selected neurons and accompanies progressive neurodegeneration. Because glutamate is sequestered by tanycytes and remains elevated for up to 5 h after administration, our results support the hypothesis that glutamate-induced neuronal injury is precipitated by unique neurotanycytic interrelationships in the ventromedial ARC nucleus in early postnatal mice.

	ACKNOWLEDGMENTS

 
The author would especially like to thank John Fernstrom for the plasma glutamate measurements, and Limin Hu and Lyn Garrett for their help with technical aspects of this project.

	FOOTNOTES

 
¹ Presented at the International Symposium on Glutamate, October 12–14, 1998 at the Clinical Center for Rare Diseases Aldo e Cele Daccó, Mario Negri Institute for Pharmacological Research, Bergamo, Italy. The symposium was sponsored jointly by the Baylor College of Medicine, the Center for Nutrition at the University of Pittsburgh School of Medicine, the Monell Chemical Senses Center, the International Union of Food Science and Technology, and the Center for Human Nutrition; financial support was provided by the International Glutamate Technical Committee. The proceedings of the symposium are published as a supplement to The Journal of Nutrition. Editors for the symposium publication were John D. Fernstrom, the University of Pittsburgh School of Medicine, and Silvio Garattini, the Mario Negri Institute for Pharmacological Research.

² Supported by International Glutamate Technical Committee #624–6-01106.

³ AMPA, -amino-3-hydroxy-5-methylisoxazole-4-propionic acid; ARC, arcuate nucleus; DAB, 3,3'-diaminobenzidine-4HCl; GFAP, glial fibrillary acidic protein; GluR, glutamate receptor; IgG, immunoglobulin G; ME, median eminence; MSG, monosodium glutamate; NMDA, N-methyl-D-aspartate; s.c., subcutaneous; TBS, Tris buffered saline.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS Immunocytochemical staining Glutamate penetration: co... DISCUSSION REFERENCES

 

1. Airoldi L., Bizzi A., Salamona M., Garattini S. Attempts to establish the safety margin for neurotoxicity of monosodium glutamate. Filer L. J., Jr Garattini S. Kare M. R. Reynolds A. W. Wurtman R. J. eds. Glutamic Acid: Advances in Biochemistry and Physiology 1979:321-331 Raven Press New York, NY.

2. Ankacrona M., Dypbukt J. M., Bonfoco E., Zhivotovsky B., Orrenius S., Lipton S. A., Nicotera P. Glutamate-induced neuronal death: a succession of necrosis or apoptosis depending on mitochondrial function. Neuron 1995;15:961-973[Medline]

3. Attwell D., Barbour B., Szatkowski M. Nonvesicular release of neurotransmitter. Neuron 1993;11:401-407[Medline]

4. Bizzi A., Veneroni E., Salmona M., Garattini S. Kinetics of monosodium glutamate in relation to its neurotoxicity. Toxicol. Lett. 1977;1:123-130

5. Bonfoco E., Krainc D., Ankacrona M., Nicotera P., Lipton S. A. Apoptosis and necrosis: two distinct events induced respectively by mild and intense insults with N-methyl-D-aspartate or nitric ovide/superoxide in cortical cell cultures. Proc. Natl. Acad. Sci. U.S.A. 1995;92:7162-7166[Abstract/Free Full Text]

6. Brorson J. R., Manzolillo P. A., Miller R. J. Ca²⁺ entry via AMPA/KA receptors and excitotoxicity in cultured cerebellar Purkinje cells. J. Neurosci. 1994;14:187-197[Abstract]

7. Choi D. W., Rothman S. M. The role of glutamate neurotoxicity in hypoxic-ischemic neuronal death. Annu. Rev. Neurosci. 1990;13:171-182[Medline]

8. de Vitry F., Picart R., Jacque C., Tixier-Vidal A. Glial fibrillary acidic protein. A cellular marker of tanycytes in the mouse hypothalamus. Dev. Neurosci. 1981;4:457-460[Medline]

9. Eurenius L., Jarskar R. Electron microscope studies on the development of the external zone of the mouse median eminence. Z. Zellforsch. Mikrosk. Anat. 1971;122:488-502[Medline]

10. Goldsmith P. C., Thind K. K. Morphological basis for neuronal control of GnRH secretion in the monkey. Plant T. M. Lee P. A. eds. The Neurobiology of Puberty 1995;vol. 2:1-14 J. Endocrinology Ltd Bristol, UK.

11. Goldsmith P. C., Thind K. K., Perera A. D., Plant T. M. Glutamate-immunoreactive neurons and their GnRH neuronal interactions in the monkey hypothalamus. Endocrinology 1994;134:858-868[Abstract]

12. Heywood R., Worden A. N. Glutamate toxicity in laboratory animals. Filer L. J., Jr Garattini S. Kare M. R. Reynolds A. W. Wurtman R. J. eds. Glutamic Acid: Advances in Biochemistry and Physiology 1979:203-215 Raven Press New York, NY.

13. Holzwarth-McBride M. A., Hurst E. M., Knigge K. M. Monosodium glutamate induced lesions of the arcuate nucleus. I. Endocrine deficiency and ultrastructure of the median eminence. Anat. Rec. 1976;186:185-205[Medline]

14. Hu L., Fernstrom J. D., Goldsmith P. C. Exogenous glutamate enhances glutamate receptor subunit expression during selective neuronal injury in the ventral arcuate nucleus of postnatal mice. Neuroendocrinology 1998;68:77-88[Medline]

15. Kanai Y., Hediger M. A. Primary structure and functional characterization of a high-affinity glutamate transporter. Nature (Lond.) 1992;360:467-471[Medline]

16. Lehmann A., Jonsson T. MK-801 selectively protects mouse arcuate neurons in vivo against glutamate toxicity. NeuroReport 1992;3:421-424[Medline]

17. Levi G., Patrizio M. Astrocyte heterogeneity: endogenous amino acid levels and release evoked by non-N-methyl-D-aspartate receptor agonists and by potassium-induced swelling in type-1 and type-2 astrocytes. J. Neurochem. 1992;58:1943-1952[Medline]

18. Levi G., Raiteri M. Carrier-mediated release of neurotransmitters. Trends Neurosci 1993;16:415-419[Medline]

19. McCall A., Glaeser B. S., Millington W., Wurtman R. J. Monosodium glutamate neurotoxicity, hyperosmolarity, and blood-brain barrier dysfunction. Neurobehav. Toxicol. 1979;1:279-283[Medline]

20. McDonald J. W., Fix A. S., Tizzano J. P., Schoepp D. D. Seizures and brain injury in neonatal rats induced by 1S,3R-ACPD, a metabotropic glutamate receptor agonist. J. Neurosci. 1993;13:4445-4455[Abstract]

21. National Research Council Guide for the Care and Use of Laboratory Animals. Publication no. 85–23 (rev.) 1985 National Institutes of Health Bethesda, MD.

22. O’hara Y., Iwata S., Ichimure M., Sasaoka M. Effect of administration routes of monosodium glutamate on plasma glutamate levels in infant, weanling and adult mice. J. Toxicol. Sci. 1977;2:281-290

23. Olney J. W. Excitotoxic amino acids: research applications and safety implications. Filer L. J., Jr Garattini S. Kare M. R. Reynolds A. W. Wurtman R. J. eds. Glutamic Acid: Advances in Biochemistry and Physiology 1979:287-331 Raven Press New York, NY.

24. Pardridge W. M. Regulation of amino acid availability to brain: selective control mechanisms for glutamate. Filer L. J., Jr Garattini S. Kare M. R. Reynolds A. W. Wurtman R. J. eds. Glutamic Acid: Advances in Biochemistry and Physiology 1979:125-137 Raven Press New York, NY.

25. Pelligrini-Giampetro D. E., Gorter J. A., Bennett M.V.L., Zukin R. S. The GluR2 (GluR-B) hypothesis: Ca²⁺-permeable AMPA receptors in neurological disorders. Trends Neurosci 1997;20:464-470[Medline]

26. Reynolds W. A., Butler V., Lemkey-Johnston N. Hypothalamic morphology following ingestion of aspartame or MSG in the neonatal rodent and primate: a preliminary report. J. Toxicol. Environ. Health 1976;2:471-480[Medline]

27. Seeburg P. H. The TINS/TiPS Lecture. The molecular biology of mammalian glutamate receptor channels. Trends Neurosci. 1993;16:359-365[Medline]

28. Smirnova T., Laroche S., Errington M. L., Hicks A. A., Bliss T. V., Mallet J. Transsynaptic expression of a presynaptic glutamate receptor during hippocampal long-term potentiation. Science (Washington DC) 1993;262:433-436[Abstract/Free Full Text]

29. Stegink L. D., Reynolds A. W., Filer L. J., Jr, Baker G. L., Daabees T. T., Pitkin R. M. Comparative metabolism of glutamate in the mouse, monkey, and man. Filer L. J., Jr Garattini S. Kare M. R. Reynolds A. W. Wurtman R. J. eds. Glutamic Acid: Advances in Biochemistry and Physiology 1979:85-102 Raven Press New York, NY.

30. Stegink L. D., Shepherd J. A., Brummel M. C., Murray L. M. Toxicity of protein hydrolysate solutions: correlation of glutamate dose and neuronal necrosis to plasma amino acid levels in young mice. Toxicology 1974;2:285-299[Medline]

31. Takasaki Y. Studies on brain lesion by administration of monosodium L-glutamate to mice. I. Brain lesions in infant mice caused by administration of monosodium L-glutamate. Toxicology 1978;9:293-305[Medline]

32. Takasaki Y., Matsuzawa Y., Iwata S., O’hara Y., Yonetani S., Ichimura M. Toxological studies of monosodium L-glutamate in rodents: relationships between routes of administration and neurotoxicity,. Filer L. J., Jr Garattini S. Kare M. R. Reynolds A. W. Wurtman R. J. eds. Glutamic Acid: Advances in Biochemistry and Physiology 1979:255-285 Raven Press New York, NY.

33. Weindl A., Joynt R. J. The circumventricular organs. Brain-Endocrine Interaction. Median Eminence: Structure and Function 1972;vol. I:280-297 Karger Basel, Switzerland. International. Symposium, Munich, 1971

34. Wickelgren I. Teaching the brain to take drugs. Science (Washington DC) 1998;280:2045-2047[Free Full Text]

This article has been cited by other articles:

				E O Farombi and O O Onyema Monosodium glutamate-induced oxidative damage and genotoxicity in the rat: modulatory role of vitamin C, vitamin E and quercetin Human and Experimental Toxicology, May 1, 2006; 25(5): 251 - 259. [Abstract] [PDF]

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Goldsmith, P. C. Search for Related Content PubMed PubMed Citation Articles by Goldsmith, P. C.