Galanin acts as a neuroprotective factor to the hippocampus

Abstract

The expression of the neuropeptide galanin is markedly up-regulated in many areas of the central and peripheral nervous system after injury. We have recently demonstrated that peripheral sensory neurons depend on galanin for neurite extension after injury, mediated by activation of the second galanin receptor subtype (GALR2). We therefore hypothesized that galanin might also act in a similar manner in the CNS, reducing cell death in hippocampal models of excitotoxicity. Here we report that galanin acts an endogenous neuroprotective factor to the hippocampus in a number of in vivo and in vitro models of injury. Kainate-induced hippocampal cell death was greater in both the CA1 and CA3 regions of galanin knockout animals than in WT controls. Similarly, exposure to glutamate or staurosporine induced significantly more neuronal cell death in galanin knockout organotypic and dispersed primary hippocampal cultures than in WT controls. Conversely, less cell death was observed in the hippocampus of galanin overexpressing transgenic animals after kainate injection and in organotypic cultures after exposure to staurosporine. Further, exogenous galanin or the previously described high-affinity GALR2 agonist, both reduced cell death when coadministered with glutamate or staurosporine in WT cultures. These results demonstrate that galanin acts an endogenous neuroprotective factor to the hippocampus and imply that a galanin agonist might have therapeutic uses in some forms of brain injury.

The 29-aa neuropeptide galanin (1) is widely expressed in both the central and peripheral nervous system and has strong inhibitory actions on synaptic transmission by reducing the release of a number of classical neurotransmitters (2-7). These inhibitory actions result in a diverse range of physiological effects including, an impairment of working memory (8) and long-term potentiation (9); a reduction in hippocampal excitability with a decreased predisposition to seizure activity (10); and a marked inhibition of nociceptive responses in the intact animal and after nerve injury (11). These neuromodulatory actions of galanin have long been regarded as the principal role played by the peptide in the nervous system. However, there is now a large body of evidence to indicate that injury to many of these neuronal systems markedly induces the expression of galanin at both the mRNA and peptide levels. Examples of such lesion studies include the up-regulation of galanin in (i) the dorsal root ganglion after peripheral nerve axotomy (12), (ii) magnocellular secretory neurons of the hypothalamus after hypophysectomy (13), (iii) the dorsal raphe and thalamus after removal of the frontoparietal cortex (decortication) (14), (iv) the molecular layer of the hippocampus after an entorhinal cortex lesion (15), and (v) the medial septum and vertical limb diagonal-band after a fimbria fornix bundle transection (16). These studies have led a number of investigators to speculate that galanin might play a trophic role in addition to its classical neuromodulatory effects.

To test this hypothesis, we and others have generated transgenic animals bearing loss- or gain-of-function mutations in the galanin gene (17-20). Phenotypic analysis of galanin knockout animals has demonstrated that the peptide acts as a survival factor to subsets of neurons in the developing peripheral and CNS (18, 21). Most recently, we have demonstrated that this trophic role is recapitulated in the adult dorsal root ganglion. Sensory neurons are dependent on galanin for neurite extension after injury, mediated by activation of the second galanin receptor subtype in a PKC-dependent manner (22). We therefore hypothesized that galanin might also act in a similar manner in the CNS, reducing cell death in hippocampal models of apoptosis and excitotoxicity.

Here we show that genetic manipulation of the endogenous levels of galanin or the addition of exogenous galanin peptide modulates hippocampal neuronal survival in a number of in vitro and in vivo models of injury.

Methods

Animals. All animals were fed standard chow and water ad libitum. Animal care and procedures were performed within the United Kingdom Home Office protocols and guidelines.

Galanin knockout mice. Details of the strain and breeding history have been published (23). In brief, mice homozygous for a targeted mutation in the galanin gene were generated by using the E14 cell line. A PGK-Neo cassette in reverse orientation was used to replace exons 1-5, and the mutation was bred to homozygosity and has remained inbred on the 129OlaHsd strain. Age and sex matched WT littermates were used as controls in all experiments.

Galanin-overexpressing mice. Details of the strain and breeding history have been published (17). In brief, galanin overexpressing mice were generated on the CBA/B6 F₁ hybrid background. A mouse 129sv cosmid genomic library was screened and an ≈25-kb region was subcloned, which contained the entire murine galanin-coding region and ≈20 kb of upstream sequence. The transgene was excised by restriction digest and microinjected into fertilized oocytes at 5 ng/μl final concentration. Four galanin overexpressing transgenic lines were generated as described (17), and galanin expression in the hippocampus was assessed by immunocytochemistry (see below). Line 46 was found to have highest levels of galanin expression in the CA1 and CA3 regions of the hippocampus and in the dentate gyrus compared to the three other lines and WT controls (Fig. 1 and data not shown). Line 46 was therefore used for all subsequent experiments.

Fig. 1.

Immunohistochemistry for galanin in WT (A and C) and line 46 galanin overexpressing (OE) animals (B and D)at ×20 (Upper), and CA1 hippocampal subfield at ×40 (Lower). (Bar = 100 μm.) The OE animals have considerably more galanin expression throughout the CA subfields and dentate gyrus. At higher power, it is evident that the majority of galanin overexpression in the CA1 region is in the terminal fields rather than in cell bodies.

Organotypic Hippocampal Cultures. Organotypic cultures were prepared as described (24, 25). In brief, the hippocampi from 5- to 6-day-old pups were rapidly removed under a dissection microscope and sectioned transversely at 400 μm by using a McIlwain tissue chopper (Mickle Laboratory Engineering, Guildford, Surrey, U.K.). The slices were cultured in 95% air and 5% CO₂ at 37°C on a microporous transmembrane biopore membrane (Millipore), in a six-well plate, in 50% MEM with Earle's salts, without L-glutamine, 50% Hanks' balanced salt solution, 25% horse serum, heat inactivated, 5 mg/ml glucose, and 1 ml of glutamine.

Preparation of Primary Neuronal Cultures. Hippocampi from 2- to 3-day-old pups were dissected and placed into 4°C collection buffer prepared with Hanks' balanced salt solution (calcium and magnesium free) (GIBCO/BRL), Hepes [10% (vol/vol)] (ICN), 50 units/ml penicillin (Britannia Pharmaceuticals, Redhill, Surrey, U.K.), 0.05 mg/ml in 100 ml of streptomycin (Sigma), and 0.5% (vol/vol) BSA (ICN). Enzymatic digestion, isolation, and culture of hippocampal neurons were performed as described (26). Cells were counted and plated at 40,000 cells per well onto d-l-polyornithine-coated (Sigma) 96-well plates. After 24 h 5′-fluoro 2′-deoxyuridine (Sigma), (antimitotic agent) was added (10 μg/ml). Cultures were incubated at 37°C with ambient oxygen and 5% CO₂ for 9 days before experimentation. The media was changed after the first 3 days and then every fourth day thereafter.

Immunohistochemistry. Mice were intracardially perfused with 4% paraformaldehyde/PBS. The brains were removed and postfixed for 4 h at room temperature. The brains were then equilibrated in 20% sucrose overnight at 4°C, embedded in OCT-mounting medium, frozen on dry ice, and cryostat-sectioned (30 μM sections). Sections were blocked and permeabilized in 10% normal goat serum/PBS containing 0.2% Triton X-100 for 1 h at room temperature. Sections were then incubated in rabbit polyclonal antibody to galanin (Affinity, Nottingham, U.K.) at 1:1,000 in PBS 0.2% Triton X-100 overnight at room temperature, washed 3 × 10 min in PBS, and incubated in FITC-goat (The Jackson Laboratory) at 1:800 for 3 h at room temperature. After washing, sections were mounted in VECTASHIELD (Vector Laboratories). Images were taken by using a Leica fluorescent microscope with an RT Color Spot camera and spot advance image capture system software (Diagnostic Instruments, Sterling Heights, MI).

Galanin immunohistochemistry was also performed on dispersed hippocampal neurons and organotypic cultures, which were fixed in 4% paraformaldehyde, permeabilized with Triton X-100, and then processed as above.

Staurosporine and Glutamate-Induced Hippocampal Damage. Organotypic hippocampal cultures (14 day) were placed in 0.1% BSA with serum-free media for 16 h before incubation with varying concentration of glutamic acid for 3 h or staurosporine for 9 h. Cultures were washed with serum-free medium and incubated for a further 24 h before imaging. Regional patterns of neuronal injury in the organotypic cultures were observed by performing experiments in the presence of propidium iodide. After membrane injury, the dye enters cells, binds to nucleic acids, and accumulates, rendering the cell brightly fluorescent (27). The CA1 neuronal subfield was clearly visible in a bright field image. Neuronal damage in the area encompassing the CA1 region was assessed by using the density slice function in nih image to establish signal above background. The area of the subfields expressing the exclusion dye propidium iodide was measured and expressed as a percentage of the total area of the subfields as assessed in the bright field image. Furthermore, for consistency in setting the parameters accurately when using the density slice function, the threshold was set against a positive control set of cultures exposed to 10 mM glutamate.

The 9-day primary hippocampal cultures were exposed to staurosporine for 24 h. The viability of neurons was measured by manual counting of both live and dead neurons by using a live/dead kit (Molecular Probes).

Treatments. Organotypic or dispersed primary hippocampal cultures were at various times cultured with or without the addition of the following chemicals: staurosporine (Sigma), L-glutamic acid (Sigma), galanin peptide (Bachem), AR-M1896 [Gal(2-11)Trp-Thr-Leu-Asn-Ser-Ala-Gly-Tyr-Leu-Leu-NH₂] (Astra-Zeneca, Montreal).

Kainate-Induced Hippocampal Injury. Eight-week-old female mice were injected with i.p. kainic acid (Tocris Neuramin, Bristol, U.K.) (20 mg/kg) or vehicle (1 ml/kg PBS). Hippocampal cell death was measured by terminal deoxynucleotidyl transferase-mediated fluorescein-dUTP nick end labeling. Animals were killed at 72 h after injection with kainic acid or vehicle. Mice were intracardially perfused with 4% paraformaldehyde/PBS and the brains rapidly removed and postfixed for 4 h at room temperature. The brains were equilibrated in 20% sucrose overnight at 4°C, embedded in OCT-mounting media, and frozen on dry ice. Sections were cut (16 μm) on a cryostat, thaw mounted onto gelatin-coated slides, and stored at -80°C until use. Apoptosis was evaluated by using an in situ cell detection kit (Boehringer). Every sixth section was collected and blocked with methanol, permeabilized with triton (0.1%) and sodium citrate (0.1%), and then labeled with fluorescein dUTP in a humid box for 1 h at 37°C. The sections were then combined with horse-radish peroxidase, colocalized with diaminobenzidine, and counterstained with haemytoxin. Controls received the same management except the labeling omission of fluorescein dUTP. After washing, sections were mounted in Vectashield (Vector Laboratories). Cells were visualized by using a Leica fluorescent microscope with an RT Color Spot camera and spot advanced image capture system software (Diagnostic Instruments).

Statistical Analysis. Data are presented as the mean + SEM. Student's t test was used to analyze the difference in staurosporine concentrations within groups. ANOVAs or nonparametric Mann-Whitney U post hoc tests were used as appropriate to analyze difference between genotypes and different ligands and/or staurosporine and glutamate points. A P value of <0.05 was considered to be significant.

Results and Discussion

Cell Death in the Hippocampus Depends on the Levels of Endogenous Galanin. The i.p. administration of 20 mg/kg kainic acid was used to induce excitotoxic hippocampal damage as described (28-30). Three days later, brains were harvested and hippocampal cell death assessed by counting the number of terminal deoxynucleotidyl transferase-mediated fluorescein-dUTP nick end labeling-positive cells. The number of apoptotic neurons was significantly greater in both the CA1 and CA3 regions of the galanin knockout animals compared to the WT controls (Fig. 2), an increase of 62.9% and 44.8%, respectively (P < 0.001). Conversely, the degree of cell death was significantly lower in both the CA1 and CA3 regions of the galanin overexpressing animals than in strain matched controls, a decrease of 55.6% and 50.4%, respectively (P < 0.05, Fig. 2).

Fig. 2.

Effects of i.p. administration of 20 mg/kg kainate on hippocampal cell death in vivo. Galanin knockout (KO) animals had significantly more neuronal cell death than strain-matched WT controls, whereas galanin overexpressing (OE) animals had less cell death than strain-matched WT controls. Cell death was measured by counting the numbers of terminal deoxynucleotidyl transferase-mediated fluorescein-dUTP nick end labeling-positive cells in every sixth section of the CA1 and CA3 regions of the hippocampus. Each point represents the mean ± SEM (n = 3). ^**, P < 0.01; ^***, P < 0.001.

To further dissect the neuroprotective role played by galanin in a more tractable in vitro system, we used both primary dispersed and organotypic hippocampal cultures (24). These two techniques are complimentary because the dispersed hippocampal cultures ensure that observed effects are neuron-specific, whereas the organotypic cultures preserve the synaptic and anatomical organization of the neuronal circuitry (24) as well as retaining many of the functional characteristics found in vivo (31). In organotypic hippocampal culture, afferent fibers are cut and thereafter degenerate; however, all neuronal and gill cell types survive, and there is no evidence that the lack of afferent inputs results in the selective loss of a particular cell type (32). Most recently, De Simoni et al. (33) have demonstrated that hippocampal organotypic slices cultures have normal development of synaptic transmission and dendritic length, as well as spine density and proportions of different spine types. Before using these in vitro paradigms to study the role played by galanin in hippocampal neuroprotection, it was important to confirm that prolonged culture did not down-regulate galanin expression. High levels of galanin expression in dispersed WT hippocampal neurons were observed (Fig. 3A), with 51 ± 3% of neurons immunocytochemically positive for galanin. Similarly, robust galanin expression was noted in WT organotypic cultures (Fig. 3B).

Fig. 3.

Immunohistochemistry staining for galanin in WT dispersed primary hippocampal neurons and slice organotypic cultures. (A) One-half of all dispersed neurons are positive for galanin after 9 days in culture (Upper Left), corresponding bright-field image (Upper Right), and overlay (Lower). (B) High levels of galanin staining are noted throughout WT organotypic cultures. Because of the thickness of the slice culture (≈250 μm after 2 wk in culture), the neurons in the periphery of the field are not in focus.

Having confirmed the above, we first studied the effects of staurosporine and glutamate on neuronal cell death in hippocampal cultures (34, 35). Staurosporine at 1 μM and 100 nM caused significant and consistent levels of neurotoxicity in both the WT and galanin knockout cultures. The percentage cell death was significantly higher in galanin knockout animals compared to WT controls at both doses (1 μM, 68 ± 0.5% vs. 38 ± 8%; 100 nM, 65 ± 10% vs. 40 ± 26%; P < 0.05; Fig. 4A). Similarly, we noted a marked and significant excess of cell death in the galanin knockout organotypic cultures after 9 h exposure to 4 mM glutamate, than in WT controls (85 ± 8.6% vs. 61 ± 9.3%, P < 0.05). To ensure that the above effects were neuron-specific, we also studied the effects of staurosporine in dispersed primary hippocampal neurons. Once again, a significant excess of cell death in the galanin knockout cultures was observed, compared to WT controls (P < 0.01) over the range of 1 μM to 10 nM staurosporine (Fig. 4B).

Fig. 4.

Responses of hippocampal cultures after incubation with 10 nM to 1 μM staurosporine (St). (A) Organotypic cultures from galanin knockout (KO) cultures had significantly more neuronal cell death than WT controls at each concentration of staurosporine in the CA1 subfield. Cell death was visualized by propidium iodide staining. Results are expressed as a percentage of the area expressing fluorescence as compared to the untreated “control” cultures. Each point represents the mean ± SEM, n = 4. ^**, P < 0.01; ^***, P < 0.001. (B) Primary hippocampal cultures were incubated for with 10 nM to 1 μM staurosporine. KO cultures had significantly more neuronal cell death than WT cultures at each concentration of staurosporine. Each point represents the mean ± SEM. n = 4. ^**, P < 0.01; ^***, P < 0.001. (Bar = 100 μm.) (C) Organotypic cultures from galanin overexpressing (OE) mice were incubated with 50 nM to 1 μM staurosporine. OE cultures had significantly less neuronal cell death than WT cultures at 100 nM and 50 nM staurosporine. Each point represents the mean ± SEM. n = 4, ^**, P < 0.01, ^***, P < 0.001.

Having demonstrated that an absence of galanin increases the susceptibility to hippocampal cell death, we then extended these studies to the galanin-overexpressing mice. A significant reduction in cell death was observed in the galanin-overexpressing animals after exposure to 50 nM or 100 nM staurosporine, compared to strain-matched WT controls (Fig. 4C). It is unclear why no additional protective effects were observed in the galanin-overexpressing animals at 1 μM staurosporine, whereas an excess of cell death was observed in the galanin knockout cultures at the same dose of staurosporine (Fig. 4A). These data using organotypic and dispersed hippocampal culture systems, confirm the in vivo data that galanin acts as a neuroprotective factor to the hippocampus and that genetic manipulation of galanin expression modulates hippocampal susceptibility to excitotoxic-induced neuronal damage. It should be noted that galanin expression in the hippocampus might be up-regulated after prolonged culture because the levels of the neuropeptide in the hippocampus are known to increase in vivo after axotomy (15). Although an increase in galanin might magnify the observed neuroprotective effects in the WT and overexpressing cultures, the findings that genetic modulation of galanin expression in vivo and the addition of exogenous peptide (see below) have similar effects make this possibility less likely and would not alter the conclusions of this manuscript.

Effect of Exogenous Galanin or Gal(2-11) on Staurosporine and Glutamate-Induced Neuronal Cell Death. To extend the above studies, we then asked whether exogenous galanin would protect WT hippocampal neurons from damage. The coadministration of 100 mM galanin with 100 nM staurosporine provided significant neuroprotection (P < 0.05) in WT organotypic cultures (Fig. 5A). Similarly, galanin was also protective over the dose range 1 μM-10 nM when coadministered with 4 mM glutamate in WT organotypic cultures (Fig. 5B). In keeping with these findings using organotypic cultures, 100 nM galanin also protected WT dispersed primary hippocampal neurons from cell death induced by 10 nM staurosporine (P < 0.05, Fig. 5C).

Fig. 5.

(A) WT organotypic hippocampal cultures were incubated with the indicated concentrations of galanin and AR-M1896 during treatment with 100 nM staurosporine. Galanin significantly reduced staurosporine-induced cell death. Each point represents the mean ± SEM. n = 4; ^**, P < 0.01. (B) WT organotypic hippocampal cultures were incubated with the indicated concentrations of galanin during treatment with 4 mM l-glutamic acid. Galanin significantly reduced glutamate-induced cell death over the range 10 nM-1 μM. Each point represents the mean ± SEM. n = 4; ^*, P < 0.05; ^**, P < 0.01. (C) Dispersed primary hippocampal neurons were incubated with the indicated concentrations of galanin or AR-M1896 during treatment with staurosporine. AR-M1896 and galanin significantly reduced staurosporine-induced cell death in WT cultures. Each point represents the mean ± SEM (n = 4). n = 3; ^**, P < 0.01.

The neuroprotective effects of galanin in the hippocampus are likely to be mediated by activation of one or more of three G protein coupled galanin receptor subtypes, designated GALR1 (36), GALR2 (37-39), and GALR3 (40). Binding of galanin to GALR1 and GALR3 have been shown to inhibit adenylyl cyclase (40-42) by coupling to the inhibitory G_i protein. In contrast, activation of GALR2 stimulates phospholipase C and PKC activity by coupling to G_q/11 (37-39, 43). To date, antisera specific to each receptor subtype have yet to be fully characterized, and the currently available literature is therefore limited to the sites of mRNA synthesis in the hippocampal formation, using in situ hybridization. GALR1 is mainly synthesized in the ventral CA1 and subiculum but not in the dorsal fields or dentate gyrus (44). In contrast, GALR2 is synthesized in the dentate gyrus but not in the CA fields (44). No GALR3 synthesis has been detected in the hippocampus or DG (45).

The lack of receptor subtype-specific antisera and the paucity of galanin ligands that are receptor subtype-specific continue to hamper the analysis of the functional roles played by each receptor. A recent major advance in the field has been the discovery by Hokfelt and coworkers (46), that galanin 2-11 peptide (termed AR-M1896) preferentially binds to GALR2 with a 500-fold specificity, compared to GALR1, and with an almost complete loss of GALR1 activation. There are no published data as to whether AR-M1896 binds, or activates, GALR3. We have previously used AR-M1896 to demonstrate that activation of GALR2 appears to be the principal mechanism by which galanin stimulates neurite outgrowth from adult sensory neurons (22). We therefore also tested the effect of 100 nM AR-M1896 when coadministered with 100 nM staurosporine in organotypic cultures from WT animals. It should be noted that even if AR-M1896 does weakly activate GALR1, this would be most unlikely at 100 nM when the IC₅₀ is 879 nM. AR-M1896 significantly reduced the amount of cell death in WT organotypic cultures to a similar amount observed with equimolar concentrations of galanin (P < 0.05, Fig. 5A). The addition of ARM1896 was also as effective in reducing staurosporine-induced cell death in galanin knockout cultures as that observed in the WT organotypic cultures (data not shown). Dispersed primary hippocampal neurons were also treated with AR-M1896 and staurosporine, demonstrating similar protective effects of the peptide to that observed with full-length galanin (Fig. 5C). No significant effects of galanin or AR-M1896 were noted in the absence of staurosporine in organotypic or primary cultures.

Conclusion

We have demonstrated that galanin acts as an endogenous neuroprotective factor to the hippocampus, and that GALR2 is likely to be the principal receptor subtype that mediates these protective effects. The absence of mRNA for GALR2 in the CA hippocampal fields might initially seem paradoxical if the receptor plays a major role in neuronal protection. It should be noted, however, that GALR2 is synthesized in the medial septum and vertical limb diagonal band neurons of the basal forebrain (44). The GALR2 protein would then be transported to the terminal fields of the CA1 and CA3 in the septohippocampal projection. In support of this, recent data indicate that up to 50% of galanin binding in the hippocampus is GALR2-dependent.† GALR2 signals by activation of PKC and hence the extracellular signal-regulated kinases cascade (43). Preliminary experiments indicate that exogenous galanin also increases the levels of phosphorylated Akt in WT hippocampal organotypic cultures (unpublished data). The hypothesis that, GALR2 activation mediates the inhibition of neuronal cell death in hippocampal neurons would be in keeping with existing studies that indicate that hippocampal protection depends on activation of extracellular signal-regulated kinase (47, 48) and PkB/Akt (49, 50).