HPC-1/syntaxin 1A and syntaxin 1B play distinct roles in neuronal survival

Authors


  • Read the Editorial Highlight for this article on page 469.

Abstract

Two types of syntaxin 1 isoforms, HPC-1/syntaxin 1A (STX1A) and syntaxin 1B (STX1B), are thought to have similar functions in exocytosis of synaptic vesicles. STX1A−/− mice which we generated previously develop normally, possibly because of compensation by STX1B. We produced STX1B−/− mice using targeted gene disruption and investigated their phenotypes. STX1B−/− mice were born alive, but died before postnatal day 14, unlike STX1A−/− mice. Morphologically, brain development in STX1B−/− mice was impaired. In hippocampal neuronal culture, the cell viability of STX1B−/− neurons was lower than that of WT or STX1A−/− neurons after 9 days. Interestingly, STX1B−/− neurons survived on WT or STX1A−/− glial feeder layers as well as WT neurons. However, STX1B−/− glial feeder layers were less effective at promoting survival of STX1B−/− neurons. Conditioned medium from WT or STX1A−/− glial cells had a similar effect on survival, but that from STX1B−/− did not promote survival. Furthermore, brain-derived neurotrophic factor (BDNF) or neurotrophin-3 supported survival of STX1B−/− neurons. BDNF localization in STX1B−/− glial cells was disrupted, and BDNF secretion from STX1B−/− glial cells was impaired. These results suggest that STX1A and STX1B may play distinct roles in supporting neuronal survival by glia.

image

Syntaxin 1A (STX1A) and syntaxin 1B (STX1B) are thought to have similar functions as SNARE proteins. However, we found that STX1A and STX1B play distinct roles in neuronal survival using STX1A−/− mice and STX1B−/− mice. STX1B was important for neuronal survival, possibly by regulating the secretion of neurotrophic factors, such as BDNF, from glial cells.

Abbreviations used
BDNF

brain-derived neurotrophic factor

DMEM

Dulbecco's modified Eagle's medium

FBS

fetal bovine serum

G3PDH

glyceraldehyde 3-phosphate dehydrogenase

GCM

glial conditioned medium

LDH

lactate dehydrogenase

MAP2

microtubule-associated protein 2

NT-3

neutotrophin-3

PLL

poly-l-lysine

HPC-1/syntaxin 1A (STX1A) is abundantly expressed in most neurons and is localized on the neuronal plasma membrane (Bennett et al. 1992; Inoue et al. 1992). STX1A is thought to be a neuronal soluble N-ethylmaleimide-sensitive fusion protein attachment protein receptor (SNARE) (Rizo and Südhof 2002; Südhof and Rothman 2009) and forms a complex called the SNARE complex with other proteins such as 25 kDa synaptosomal-associated protein (SNAP25) and vesicle-associated membrane protein (VAMP2)/synaptobrevin. This complex is considered essential for synaptic vesicle exocytosis. In addition to STX1A, syntaxin 1B (STX1B), which is a STX1 isoform, is also localized on the neuronal plasma membrane. These two isoforms are highly homologous and are coexpressed with similar distributions in individual neurons (Bennett et al. 1992, 1993; Inoue et al. 1992). Therefore, both STX1A and STX1B may be functionally similar to neuronal SNAREs.

In addition to synaptic vesicle exocytosis, STX1 may be required for development and survival of neurons, because botulinum neurotoxin, which cleaves proteins required for synaptic vesicle exocytosis including STX1, is well known to cause neuronal degeneration in vitro and in vivo (Igarashi et al. 1996; Williamson and Neale 1998; Berliocchi et al. 2005; Zhao et al. 2010; Peng et al. 2013). Furthermore, STX1 may be essential for embryonic viability (Schulze et al. 1995; Schulze and Bellen 1996; Burgess et al. 1997), and the absence of STX1 causes cell lethality in Drosophila (Stowers and Schwarz 1999). On the other hand, tetanus neurotoxin-mediated cleavage of VAMP2, which is localized on synaptic vesicle membranes, does not affect neuronal survival in culture (Habig et al. 1986; Osen-Sand et al. 1996). Previously, we reported that STX1A gene knockout mice appear to develop normally (Fujiwara et al. 2006), possibly because of compensation for the loss of STX1A by STX1B.

In this study, we generated gene knockout mice for STX1B and examined the role of STX1 isoforms on neuronal development and neuronal survival using STX1A−/− or STX1B−/− mice. STX1B−/− mice died within 14 days after birth and showed abnormal brain development, unlike STX1A−/− mice. In dissociated culture, the cell viability of STX1B−/− neurons was lower than that of STX1A−/− neurons, but survival of STX1B−/− neurons was supported by WT and STX1A−/− glial cells as well as glial conditioned medium (GCM) from WT and STX1A−/− glial cells. Some trophic factors were also effective at promoting survival. We also observed that the STX1 isoform expressed in glial cells was STX1B. In STX1B−/− glial cells, the localization of brain-derived neurotrophic factor (BDNF) was disrupted, and BDNF protein accumulated in the cytoplasm. BDNF secretion from STX1B−/− glial cells was impaired, but BDNF secretion was partially recovered by expression of recombinant STX1B in STX1B−/− glial cells. Thus, we showed that STX1A and STX1B play distinct roles in neuronal survival that is partly owing to secretion of neurotrophic factors, such as BDNF and neutotrophin-3 (NT-3) from glial cells.

Materials and methods

Generation of STX1B knockout mice

A BAC clone (37L7) containing the mouse STX1B gene was purchased from Invitrogen (Carlsbad, CA, USA). A 13-kb fragment containing exons 4–10 was subcloned into pBluescript. The region from exon 9 to exon 10 that encoded the H3 and transmembrane domains was replaced with a neomycin resistance gene, and the diphtheria toxin gene (DT-A) was attached to the 3′ end of the construct (Fig. 1a). The linearized targeting vector was transfected into Embryonic Stem (ES) cells, which were selected in culture medium containing G418. G418-resistant ES cell colonies were screened with PCR and Southern blot analysis. The knockout ES cell clones were injected into blastocysts and implanted into pseudopregnant ICR females (Charles River Lab., Yokohama, Japan). The resulting chimeric mice were bred to C57Bl/6J mice (Charles River Lab.) to generate heterozygous mutant mice, and the genotypes of the offspring were determined with PCR analysis. Finally, heterozygotes were backcrossed to C57Bl/6J mice for four to five generations (Fujiwara et al. 2009).

Figure 1.

Generation of STX1B knockout mice. (a) Schematic showing the WT mouse syntaxin 1B locus, targeting vector construct, and targeted allele. The solid bars indicate STX1B exons 5–10. The targeting vector contains exons 5–10 of the gene. The region from exons 9–10 was replaced with a neomycin expression cassette (Neo). A diphtheria toxin expression cassette (DT-A) was attached to the 3′ end of the construct for negative selection. The probe for Southern blot analysis is indicated. (b) Viability of STX1B−/− mice. STX1B−/− mice began to die at P7, and almost all STX1B−/− mice died by P14 (WT = 88, STX1B+/− = 132, STX1B−/− = 52). (c) Western blotting analysis of STX1B, STX1A in STX1B−/− mice. Whole-brain homogenate was prepared from P7 mice. 14D8 is the STX1A-specific monoclonal antibody. Polyclonal STX1 antibody is reactive to both STX1A and STX1B. Expression of STX1B protein was absent in STX1B−/− mice. Expression of STX1A was not changed by deletion of STX1B. Arrow indicates STX1A, and arrowhead represents STX1B.

Western blotting analysis

Whole-brain homogenate was obtained from postnatal day 7 (P7) mice, and glial cell lysates were prepared from glial feeder layers cultured for 14 days (see 'Neuronal cultures'). These were used for western blotting as described previously (Fujiwara et al. 2001). The HPC-1/STX1A-specific monoclonal antibody 14D8 (Kushima et al. 1997), the polyclonal antibody against STX1B, and the polyclonal STX1 antibody were obtained as described previously (Iwahashi et al. 2003). The monoclonal antibody against Munc18 was purchased from Transduction Laboratories (Lexington, KY, USA). The monoclonal antibody against BDNF was purchased from PROSPEC (East Brunswick, NJ, USA). The polyclonal antibody against glial fibrillary acidic protein and the monoclonal antibody against α-tubulin were purchased from Sigma (St. Louis, MO, USA). The monoclonal antibody against NeuN was purchased from Millipore (Billerica, MA, USA). The intensity of band was measured by the LAS4000 system (GE Healthcare, Buckinghamshire, UK) and analyzed using NIH ImageJ software (Bethesda, MD, USA).

Histological studies

P7 mouse brains were fixed with 4% paraformaldehyde/phosphate-buffered saline (PBS) and cryoprotected with 20% sucrose/PBS. Cryostat sections (18 μm) were collected on gelatin-coated slides. The sections were stained with 0.1% cresyl violet solution and examined under a light microscope. For the immunohistochemical study, the sections were incubated with anti-NeuN (Millipore) or anti-calbindin antibody (Swant, Marly, Switzerland) in PBS containing 5% normal goat serum and 0.1% Triton X-100 (NGS/PBS-T). Subsequently, sections were incubated with biotinylated anti-rabbit IgG, and then with avidin-biotin-peroxidase complex (Vector Laboratories, Burlingame, CA, USA). The immunoreactive cells were visualized by incubating in 3,3′-Diaminobenzidine (Sigma) and examined microscopically.

Neuronal cultures

Hippocampal cultured neurons were prepared as described previously (Mishima et al. 2002). In brief, the hippocampus from P0 to P2 mice (WT, STX1A−/− or STX1B−/−) was treated with trypsin (5 mg/mL for 15 min at 37°C; Sigma) and dissociated by pipetting through plastic tips. Dissociated neurons were plated on poly-l-lysine (PLL; Sigma)-coated glass coverslips or glial feeder layers prepared from WT, STX1A−/− or STX1B−/− mice at low density (1–2 × 104 cells/cm2) in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS), and the medium was changed to DMEM supplemented with 2% B-27 after 1 day. GCM was prepared as follows: Glial cells from the cortex or hippocampus of WT, STX1A−/− or STX1B−/− mice at P0-2 were cultured in DMEM containing 10% FBS for 14 days. Then, the culture medium was changed to FBS-free DMEM and collected after 14 days as GCM. To study the effect of GCM on survival, GCM was added to the culture medium beginning at 2 DIV, and half the volume was changed every 3 days. The neurotrophic factors or growth factors, such as BDNF, NT-3, insulin, glial cell line–derived neurotrophic factor or basic fibroblast growth factor were purchased from Sigma, and nerve growth factor was purchased from TOYOBO (Osaka, Japan). Anti-BDNF for inhibiting BDNF biological activity was purchased from Biosensis (New South Wales, Australia). Anti-NT-3 for inhibiting NT-3 activity was purchased from Peprotech (Rocky Hill, NJ, USA).

Immunocytochemical staining

Dissociated cultured neurons were fixed with 4% paraformaldehyde/PBS for 30 min. After treatment with PBS-T for 15 min, the cells were incubated in NGS/PBS-T for 1 h to block non-specific binding. After washing with PBS, the cells were incubated sequentially with anti-microtubule-associated protein 2 (MAP2) monoclonal antibody (Sigma)/anti-synaptophysin polyclonal antibody (TOYOBO) or anti-BDNF monoclonal antibody (PROSPEC)/anti-glial fibrillary acidic protein polyclonal antibody (Sigma) in NGS/PBS-T (at 4°C overnight), followed by incubation with Alexa 488- and Alexa 555-conjugated secondary antibody (at 4°C for 1 h; Life Technologies, Carlsbad, CA, USA). The immunoreactive cells were examined with fluoromicroscopy (Olympus, Tokyo, Japan) or confocal microscopy (Carl Zeiss, Jena, Germany).

Analysis of cell viability

Cells were cultured in 24-well microplates. Phase-contrast images of cultured neurons were obtained at each DIV, and MAP2-positive cells were counted per field of 0.137 mm2 as described (Heeroma et al. 2004). Cell viability was determined as the percent of the number of neurons at 1 DIV. Lactate dehydrogenase (LDH) activity was measured with an LDH cytotoxicity detection kit (TAKARA, Shiga, Japan) as recommended. LDH activity was determined as the percent of absorbance in 1 DIV medium prepared from neuronal cultures treated with 1% Triton X-100/PBS.

RT-PCR analysis

Total RNA was extracted from cultured WT, STX1A−/− or STX1B−/− glial cells at 14 DIV with TRIzol Reagent (Invitrogen). First-strand cDNA was synthesized from RNA using Oligo-dT primers, random primers, and reverse transcriptase (TAKARA). Aliquots of cDNA (corresponding to 500 ng RNA) were amplified in 20 μL of PCR cocktail containing each specific primer and ExTaq DNA polymerase (TAKARA). Glyceraldehyde 3-phosphate dehydrogenase (G3PDH) transcripts were used as a positive internal control. Using the optimal conditions for amplification with each primer pair, the target cDNAs of BDNF and G3PDH were amplified for 30 and 25 cycles, respectively. The following primer pairs were used: for BDNF, 5′-AGGTGAGAAGAGTGATGACCATCC-3′ (sense) and 5′-CAACATAAATCCACTATCTTCCCC-3′ (antisense); and for G3PDH, 5′-TGAAGGTCGGTGTGAACGGATTTG-GC-3′ (sense), 5′-CATGTAGGCCATGAGGTCCACCAC-3′ (antisense) (Toyomoto et al. 2004). The expected product sizes were 777 bp for BDNF and 983 bp for G3PDH. The PCR products were electrophoresed on an agarose gel, stained with ethidium bromide, and photographed under UV illumination with the LAS4000 system (GE Healthcare).

Enzyme immunoassay of BDNF and transfection

Glial cells from WT, STX1A−/− or STX1B−/− mice were seeded in 24-well microplates. Cells were cultured in DMEM containing 10% FBS for 14 days, and the culture medium was changed to FBS-free DMEM (cont.) or HK medium (40 mM KCl) to induce the secretion of BDNF (Abiru et al. 1998). Each medium was collected after 3 days. Recovery of BDNF in collected medium was achieved using Microcon YM-10 centrifugal filter device (Millipore), and BDNF concentration was measured with BDNF Emax ImmunoAssay Systems (Promega, Madison, WI, USA detection range; 7.8–500 pg/mL) as described previously (Toyomoto et al. 2004) with slight modification. In brief, in microtiter plates coated with anti-BDNF monoclonal antibody collected medium or BDNF standard solution were incubated for 2 h. Microtiter plate was incubated sequentially with biotinylated anti-BDNF polyclonal antibody for 2 h, followed by incubation with streptavidin-linked horseradish peroxidase for 1 h. After treatment with 3,3′,5,5′-tetramethylbenzidine substrate, absorbance was measured at 450 nm. BDNF concentration was shown as the percent of WT control. Expression vector of recombinant STX1B was created by subcloning of STX1B cDNA into pRc/cytomegalovirus promoter expression vector (Invitrogen). STX1B−/− glial cells were transfected with the expression vector using FuGENER 6 Transfection Reagent (Promega) as recommended. After 4 days from transfection, the culture medium was changed to HK medium to induce the secretion of BDNF, and BDNF concentration was measured described above.

Animals

All experimental protocols were approved by the institutional Animal Care and Use Committee of Kyorin University School of Medicine.

Statistics

All data are presented as the means ± SEM. Statistical significance was determined using one-way anova and two-tailed Student's t-tests. Differences were considered significant at p < 0.05.

Results

Syntaxin 1B−/− mice were not viable in adulthood, unlike STX1A−/− mice

A targeting vector containing the replaced exon 9 and a portion of exon 10 of STX1B with a Neor cassette (Fig. 1a) was used to generate mutant mice with targeted disruption of STX1B (Fujiwara et al. 2009). Homozygous mutant (STX1B−/−) mice were born alive, although the genotype ratio of the pups was slightly lower than the expected Mendelian ratio from intercrossing of the heterozygous mutant (STX1B+/−) mice (+/+ : +/−:−/− = 1 : 1.50 : 0.59; = 191). Neonatal STX1B−/− mice were smaller than WT littermates, and abnormal motor coordination was observed at P7 in STX1B−/− mice. STX1B−/− mice began to die at P7, and almost all STX1B−/− mice died by P14 (Fig. 1b), unlike STX1A−/− mice. Thus, we analyzed motor function using P7 mice. The righting reflex test, which evaluates motor function, is a score of the time needed for the mouse to turn over after being placed on its back on a flat surface. WT and STX1B+/ mice quickly turned over after their body was inverted (Figure S1). However, STX1B−/− mice could not turn their body over within 180 s, indicating that the absence of STX1B caused severe disruption of motor function.

To confirm that STX1B−/− mice completely lacked expression of STX1B, we conducted western blotting analysis using brain extracts prepared from P7 mice. As shown in Fig. 1c, STX1B−/− mice showed a total absence of STX1B protein, and expression in STX1B+/ mice was reduced by about 50% compared to WT mice. We examined whether the loss of STX1B expression affected the expression of STX1 isoform, STX1A in the brain. The amount of STX1A was not changed by deletion of STX1B.

We previously reported that STX1A−/− mice developed normally and that morphological abnormalities were not observed in the brain (Fujiwara et al. 2006). Thus, we examined the morphology of the brain of STX1B−/− mice. The size of the cerebellum in STX1B−/− mice appeared to be smaller than that in WT littermates. As shown in Fig. 2a, Nissl staining showed that the molecular layer in STX1B−/− mice at P7 was thinner than that in WT littermates (WT 100.0 ± 2.0 = 24, STX1B+/− 101.6 ± 2.3 = 32, STX1B−/− 69.6 ± 1.9 = 32, the percent of WT). Immunostaining for calbindin, a marker of Purkinje cells, revealed that the dendrites of Purkinje cells in STX1B−/− mice were less arborized compared with those in WT littermates (Fig. 2b). In the hippocampus, Nissl staining showed a normal gross structural appearance in STX1B−/− mice compared to WT littermates or STX1A−/− mice (Fig. 2c). However, immunostaining for NeuN revealed that the number of NeuN-positive cells in STX1B−/− mice was decreased compared with WT littermates or STX1A−/− mice (Fig. 2d, WT 30.7 ± 1.3 = 11, STX1A−/− 32.8 ± 1.2 = 9, STX1B−/− 16.6 ± 0.9 = 12, the number of NeuN-positive cells per 100 × 100 μm area in CA1). These results suggested that deletion of STX1B affected the development of the brain unlike deletion of STX1A.

Figure 2.

STX1B−/− mice exhibit disrupted brain morphology. (a, b) Sagittal sections of P7 cerebellum from WT, STX1B+/− or STX1B−/− mice were stained with cresyl violet or immunostained for calbindin, a marker of Purkinje cells. (a′) is a higher magnification image of the boxed region in (a). The molecular layer in STX1B−/− mice was thinner than that in WT or STX1B+/− littermates. The dendrites of Purkinje cells in STX1B−/− mice were less arborized compared with WT or STX1B+/− littermates. Hippocampal sections from WT, STX1A−/−, or STX1B−/− mice were visualized with Nissl staining (c) or immunostaining for NeuN (d). The number of NeuN-positive cells in STX1B−/− mice was decreased compared with WT littermates or STX1A−/− mice. Bar indicates 50 μm.

STX1B was essential for neuronal survival

We examined survival of STX1B−/− hippocampal neurons from neonatal brains in culture. Hippocampal neurons were plated on PLL-coated glass coverslips at low density (1–2 × 104 cells/cm2). Trypan Blue staining, which indicated the viability of WT, STX1A−/− or STX1B−/− neurons in vitro, was similar to staining before plating (data not shown). Next, the number of neurons attached to the substrate was counted 24 h after plating, and > 95% of WT, STX1A−/− or STX1B−/− neurons had attached (data not shown), suggesting equal initial cell viability among all genotypes.

The number of neurons as identified by immunocytochemical staining for MAP2 was counted at each DIV, and cell viability was determined as the percent of 1 DIV neurons. Up to 7 DIV, no difference was found in cell survival among the genotypes. However, cell survival of STX1B−/− neurons was lower than that of WT or STX1A−/− neurons beginning at 9 DIV (Fig. 3a). Similarly, cell death, which was determined as LDH activity in culture medium that leaked from dead cells, was increased in the STX1B−/− neuronal culture beginning at 9 DIV compared with the WT neuronal culture (Fig. 3b). These results indicated that STX1B−/− neurons from the neonatal hippocampus were as viable as WT or STX1A−/− neurons for the first 7 DIV. However, lower survival of STX1B−/− neurons beginning at 9 DIV suggested a role for STX1B in neuronal survival, unlike STX1A.

Figure 3.

Cell viability of dissociated cultured neurons on poly-l-lysine (PLL). (a) Hippocampal neurons were prepared from neonatal WT, STX1A−/− or STX1B−/− mice and cultured on PLL-coated coverglass. Viable neurons (MAP2-positive cells) were counted at each DIV. Cell viability of STX1B−/− neurons was lower than that of WT or STX1A−/− neurons beginning at 9 DIV (WT = 8, STX1A−/− = 4, STX1B−/− = 8). (b) Cell viability was determined as lactate dehydrogenase (LDH) activity that had leaked from dead cells. Culture medium at each DIV was collected, and LDH activity was measured. Cell death in STX1B−/− neuronal cultures was increased more than that in WT or STX1A−/− cultures beginning at 9 DIV (WT = 4, STX1A−/− = 3, STX1B−/− = 4). **p < 0.01 compared to the WT.

We examined the morphology of cultured neurons. At 7 or 14 DIV, WT, STX1A−/− or STX1B−/− neurons were fixed and immunostained for synaptophysin as a pre-synaptic marker and for MAP2 (Fig. 4a). The number of synaptophysin-positive puncta on the branches (per 10 μm length of the process) was counted, but no difference was observed among genotypes at 7 and 14 DIV (Fig. 4b). Furthermore, we examined the number of main branches from the soma. No difference was found among the genotypes at 7 and 14 DIV (Fig. 4c). These results suggested that STX1B−/− neurons died rapidly between 7 and 9 DIV, but the surviving STX1B−/− neurons developed normally compared to WT or STX1A−/− neurons.

Figure 4.

Morphological analysis of dissociated cultured neurons. (a) Hippocampal neurons from neonatal WT, STX1A−/− or STX1B−/− mice were cultured for 14 days. The neurons were fixed and immunostained for synaptophysin (red) and MAP2 (green). Bar indicates 20 μm. (b) The number of synaptophysin-positive puncta on branches that were 15–25 μm long from the center of the soma was counted (per 10 μm length of the process). No difference was observed among genotypes at 7 and 14 DIV (WT = 21, STX1A−/− = 22, STX1B−/− = 20). (c) The number of main branches from the soma in hippocampal neurons was counted. No difference was found among the genotypes at 7 and 14 DIV (WT = 21, STX1A−/− = 22, STX1B−/− = 20).

Culture on a glial cell layer or in GCM supported survival of STX1B−/− neurons

Glial cells provide most of the neuronal support in the brain (Barres 1991). Therefore, WT glial cells may support neuronal survival of STX1B−/− neurons. WT or STX1B−/− neurons were cultured on a WT glial cell layer. The cell viability of WT neurons was not affected by the WT glial cell layer at any DIV (Fig. 5a). No additive effect of the WT glial cell layer was observed, possibly because of maximal neuronal survival in these culture conditions including optimal culture medium or cell density (Banker and Cowan 1977; Hartikka and Hefti 1988; Brewer et al. 1993). On the other hand, the viability of STX1B−/− neurons beginning at 9 DIV was restored on the WT glial cell layer (Fig. 5b). This effect on neuronal survival may be because of factors released from glial cells. To study this possibility, WT or STX1B−/− neurons were cultured in GCM, which was prepared from WT glial culture for 14 DIV and then added to the culture medium at half volume. WT neurons were not affected by GCM from the WT glial culture (Fig. 5c), but GCM from the WT glial culture supported the survival of STX1B−/− neurons beginning at 9 DIV (Fig. 5d). To investigate if the effect of the glial cell layer on neuronal survival was different among genotypes, WT or STX1B−/− neurons were cultured on the STX1A−/− or STX1B−/− glial cell layer. Survival of WT neurons was not affected by the STX1A−/− or STX1B−/− glial cell layer (Fig. 5a). GCM from STX1A−/− or STX1B−/− glial culture had no additive effect on neuronal survival (Fig. 5c). On the other hand, the STX1B−/− glial feeder layer did not support the survival of STX1B−/− neurons at 9 or 11 DIV, although STX1B−/− neurons survived normally on the STX1A−/− glial cell layer (Fig. 5b). Moreover, GCM from the STX1B−/− glial culture did not affect STX1B−/− neuronal survival, although GCM from WT or STX1A−/− glial culture had a similar effect on survival (Fig. 5d). These results suggested that at least a portion of the glial function that supports neuronal survival was absent in STX1B−/− mice.

Figure 5.

Cell viability of cultured neurons on a glial feeder layer or in glial conditioned medium (GCM). Hippocampal neurons were prepared from neonatal WT or STX1B−/− mice and cultured on a glial feeder layer (a, b) or in GCM (c, d). (a) WT neurons were cultured on a glial feeder layer derived from WT, STX1A−/− or STX1B−/− mice. The cell viability of WT neurons was not different among genotypes (WT = 3, STX1A−/− = 3, STX1B−/− = 5). (b) STX1B−/− neurons were cultured on a glial feeder layer derived from WT, STX1A−/− or STX1B−/− mice. STX1B−/− neurons survived normally on WT or STX1A−/− glial feeder layers, whereas the STX1B−/− glial feeder layer had a weak effect on survival beginning at 9 DIV (WT = 4, STX1A−/− = 3, STX1B−/− = 5). (c) WT neurons were cultured in GCM prepared from a WT, STX1A−/− or STX1B−/− glial culture, and half the volume of medium was replaced with this GCM every 3 days. The cell viability of WT neurons was not different among genotypes (WT = 3, STX1A−/− = 3, STX1B−/− = 4). (d) STX1B−/− neurons were cultured in GCM prepared from a WT, STX1A−/− or STX1B−/− glial culture. GCM from the WT or STX1A−/− glial culture had similar effect on survival, but that from the STX1B−/− glial culture did not (WT = 4, STX1A−/− = 3, STX1B−/− = 5). *p < 0.05, **p < 0.01 compared to the WT.

Neurotrophic factors supported survival of STX1B−/− neurons

Neurotrophic factors or growth factors support neuronal survival, and some of those factors are released from glial cells as well as neurons (Rudge et al. 1992; Condorelli et al. 1994). We examined whether addition of neurotrophic factors or growth factors to the culture medium supported neuronal survival of STX1B−/− neurons. Various types of factors such as BDNF, nerve growth factor, NT-3, insulin, glial cell line–derived neurotrophic factor or basic fibroblast growth factor were added to the culture medium at 2 DIV. These factors had no effect on WT neurons at 9 or 11 DIV (Fig. 6a). However, among these factors, BDNF and NT-3 supported survival of STX1B−/− neurons (Fig. 6b). To test if the prominent BDNF or NT-3 effect was specific, anti-BDNF or anti-NT-3 antibody was applied. These antibodies inhibited the effect of BDNF or NT-3 on survival of STX1B−/− neurons (Fig. 6b). Thus, application of BDNF or NT-3 supported survival of STX1B−/− neurons.

Figure 6.

Survival of STX1B−/− neurons was supported by neurotrophic factors. Hippocampal neurons from neonatal WT (a) or STX1B−/− mice (b) were cultured in the presence of neurotrophic factors [14.5 ng/mL brain-derived neurotrophic factor (BDNF), 100 ng/mL nerve growth factor (NGF), 30 ng/mL neutotrophin-3 (NT-3)] or growth factors [100 nM insulin, 10 ng/mL glial cell line–derived neurotrophic factor (GDNF), 10 ng/mL basic fibroblast growth factor (bFGF)]. Neurotrophic factors were added to the culture medium beginning at 2 DIV. Viable neurons were counted at each DIV. BDNF or NT-3 promoted survival of STX1B−/− neurons (BDNF = 5, NT-3 = 4, other factors = 3 each). Anti-BDNF antibody inhibited this effect of BDNF. +α-BDNF indicates simultaneous application of BDNF and anti-BDNF antibody (+α-BDNF, = 3). +α-NT-3 indicates simultaneous application of NT-3 and anti-NT-3 antibody (+α-NT-3 = 2). *p < 0.05, **p < 0.01 compared to the control.

Expression of STX1 isoforms and Munc18-1 in glial cells

We examined the expression of STX1 isoforms in glial cells. Western blotting analysis revealed that the expression of STX1B was predominant compared to that of STX1A in glial cells (Fig. 7A), consistent with a previous report (Paco et al. 2009). As in STX1B−/− neurons, lack of Munc18-1, which tightly binds to STX1, resulted in cell-autonomous degeneration of neurons (Heeroma et al. 2004). Therefore, we examined the expression of Munc18-1 in glial cells. The expression of Munc18-1 was decreased in STX1B−/− glial cells, unlike in STX1A−/− glial cells (Fig. 7a, WT 100.0 ± 7.8 = 3, STX1A−/− 104.7 ± 12.0 = 3, STX1B−/− 24.8 ± 6.5 = 3, percent of WT).

Figure 7.

Expression of STX1 isoforms, Munc18-1, and brain-derived neurotrophic factor (BDNF) in glial cells. (a) Western blotting analysis of STX1B, STX1A and Munc18-1 in glial cell lysates. The expression of STX1B protein was predominant compared to that of STX1A in WT or STX1A−/− glial cells, but not in STX1B−/− glial cells. The expression of Munc18-1 protein was lower in STX1B−/− glial cells than in WT or STX1A−/− glial cells. (b) Confocal microscopic images of immunostaining for BDNF (green) or glial fibrillary acidic protein (GFAP) (red) in WT, STX1A−/− or STX1B−/− glial cells. BDNF-positive puncta were observed in all genotypes, but in STX1B−/− glial cells, BDNF was localized around the nucleus. Bar indicates 10 μm. (c) The number of BDNF-positive puncta in the cytoplasm that were located 20–40 μm or 40–60 μm from the center of the nucleus was counted (20 × 20 μm area). In STX1B−/− glial cells, the number of BDNF-positive puncta in the periphery of cells was lower than that in WT or STX1A−/− glial cells (WT = 8, STX1A−/− = 7, STX1B−/− = 8). **p < 0.001. (d) Western blotting analysis of glial cell lysates and whole-brain lysates. The amount of BDNF in STX1B−/− glial cells was higher than that in WT or STX1A−/− glial cells, while that in brain was not different among genotypes. (e) RT-PCR analysis of BDNF expression in glial cells. BDNF mRNA levels were not different among genotypes.

Expression of BDNF in glial cells

We next examined the expression of BDNF in STX1B−/− glial cells. Immunocytochemical studies showed that BDNF-positive puncta were present in all genotypes, but in STX1B−/− glial cells, BDNF was localized around the nucleus (Fig. 7b). Furthermore, in STX1B−/− glial cells, the number of BDNF-positive puncta at the periphery of the cell was lower than that in WT or STX1A−/− glial cells (Fig. 7c). Western blotting analysis using glial cell lysates showed that the amount of BDNF in STX1B−/− glial cells was higher than that in WT or STX1A−/− glial cells (Fig. 7d, WT 100.0 ± 6.3 = 3, STX1A−/− 93.5 ± 5.8 = 3, STX1B−/− 158.1 ± 9.1 = 3, percent of WT). However, the BDNF mRNA level in glial cells was not different among genotypes (Fig. 7e). On the other hand, the amount of BDNF in whole brain was not different among genotypes (Fig. 7d). These results indicated that the BDNF localization pattern in glial cells was disrupted in STX1B−/− mice, such that more BDNF-positive puncta were located around the nuclear regions compared with WT or STX1A−/− glial cells. Moreover, abundant BDNF protein accumulated in the cytoplasm of STX1B−/− glial cells, although the BDNF mRNA level was normal, suggesting that BDNF secretion from glial cells may be impaired in STX1B−/− glial cells.

BDNF concentration and BDNF secretion from glial cells

BDNF secretion was induced with high concentration of extracellular potassium (Abiru et al. 1998). To maximally stimulate secretion, glial cells were cultured in high potassium (HK) medium (40 mM KCl) for 3 days, and BDNF secretion was measured (Fig. 8a). In control medium, no difference was found among genotypes (WT 10.2 ± 1.9 = 9, STX1A−/− 9.0 ± 2.4 = 9, STX1B−/− 8.6 ± 1.8 pg/mL = 9). However, we found that HK stimulation enhanced BDNF secretion from WT or STX1A−/− glial cells, but not from STX1B−/− glial cells. BDNF secretion was unchanged by glutamate stimulation in all genotypes (data not shown). Furthermore, we studied reconstitution experiment by expression of recombinant STX1B in STX1B−/− glial cells. As shown in Fig. 8b, BDNF secretion with HK stimulation was partially recovered in STX1B−/− glial cells with transiently expressing recombinant STX1B. Amount of expressed STX1B was about 34% compared to WT control (data not shown). These results suggest that STX1B, but not STX1A, is related to the BDNF secretionary process from glial cells.

Figure 8.

Brain-derived neurotrophic factor (BDNF) secretion from glial cells. (a) Glial cells from WT, STX1A−/− or STX1B−/− mice were treated with fetal bovine serum (FBS)-free Dulbecco's modified Eagle's medium (DMEM) (cont.) or HK medium (40 mM KCl) to induce BDNF secretion. After 3 days, BDNF concentration in collected medium was measured by ELISA assay (WT = 9, STX1A−/− = 9, STX1B−/− = 9). BDNF concentration in control medium was not different among genotypes. HK stimulation enhanced BDNF secretion from WT or STX1A−/− glial cells, but not from STX1B−/− glial cells. (b) Expression vector of recombinant STX1B was transfected into STX1B−/− glial cells, and BDNF concentration in control or HK medium was measured by ELISA assay (WT + mock = 9, STX1B−/− + STX1B = 7). BDNF secretion was partially recovered by expression of recombinant STX1B. **p < 0.01 compared to each control.

Discussion

Both STX1A and STX1B are thought to function similarly as neuronal t-SNAREs (Gerber et al. 2008; Zhou et al. 2013) and may mediate the secretion of neurotransmitters (glutamate or GABA) or neuromodulators (5-HT, norepinephrine, or dopamine). We showed here that STX1B−/− mice died by P14, whereas STX1A−/− mice developed normally (Fujiwara et al. 2006). Moreover, brain development, such as in the cerebellum and hippocampus, was morphologically impaired in STX1B−/− mice (Fig. 2). These results indicate that STX1A and STX1B play distinct roles in neuronal development. We also found that the cell viability of STX1B−/− neurons was lower than that of WT or STX1A−/− neurons in dissociated culture on PLL (Fig. 3). This result suggests that STX1B is necessary for neuronal development and neuronal survival, unlike STX1A.

During brain development, a surplus of neurons and synapses are formed, and many excess neurons and synapses are eliminated. Synaptic activity is a major determinant of the elimination and regulation of neuronal survival (Lichtman and Colman 2000; Katz and Crowley 2002). For example, inhibiting synaptic transmission by motor neurons selectively eliminates the inactive synapses (Buffelli et al. 2003). Similarly, prevention of neuronal elimination by genetically or pharmacologically blocking neurotransmission is related to the survival of neurons (Houenou et al. 1990; Banks et al. 2001). Secretion of trophic factors induced by synaptic activity is thought to underlie regulation of elimination and neuronal survival (Katz and Shatz 1996; Sanes and Lichtman 1999). We previously reported that STX1A−/− mice exhibit normal fast synaptic transmission (Fujiwara et al. 2006). On the other hand, basal fast synaptic transmission is impaired in STX1B−/− mice (Mishima et al. 2014). Therefore, abnormal synaptic activity may be an important factor causing impaired brain development and neuronal survival in STX1B−/− mice, unlike in STX1A−/− mice.

The cell viability of STX1B−/− neurons cultured on PLL was lower beginning at 9 DIV compared to WT or STX1A−/− neurons (Fig. 3). It was reported by morphological observation that in our culture condition the majority of dissociated hippocampal cells alive at 7 DIV were excitatory pyramidal neurons (Kaech and Banker 2006; Beaudoin et al. 2012). Therefore, we think that most survived neurons were pyramidal neurons, and other types of hippocampal cells, such as granule cells and interneurons, might be few, if present. However, cellular perturbations are likely in STX1B−/− neurons between 7 and 9 DIV, because STX1B−/− neurons that survived after 9 DIV were normal in appearance (Fig. 4). For example, transformation from an immature to a mature synaptic state proceeds between 7 and 9 DIV in hippocampal neurons (Barsarsky et al. 1994), and this maturation process may be impaired in STX1B−/− neurons. BDNF is an important factor in the formation, maturation, and plasticity of synapses (Gottmann et al. 2009). We also showed that application of BDNF supported survival of STX1B−/− neurons (Fig. 6). Therefore, loss of synaptic maturation, which is normally induced by BDNF, may be related to the lower cell viability of STX1B−/− neurons. Glial cells provide most of the neuronal support in the brain (Barres 1991). In fact, STX1B−/− neurons survived on a WT glial feeder layer (Fig. 5). Furthermore, GCM prepared from WT glial cultures also supported the survival of STX1B−/− neurons (Fig. 5). On the other hand, STX1B−/− glial feeder layers and GCM were not effective in promoting survival of STX1B−/− neurons (Fig. 5). These results suggest that some factors derived from normal glial cells may be important for survival of STX1B−/− neurons, and secretion of important factors may be reduced in STX1B−/− glial cells. We found that BDNF, and to a lesser extent NT-3, were effective in supporting survival of STX1B−/− neurons (Fig. 6). Moreover, abundant BDNF protein accumulated in the cytoplasm in STX1B−/− glial cells, unlike in STX1A−/− glial cells, whereas the BDNF mRNA level in glial cells was not different among genotypes (Fig. 7). The pattern of BDNF localization in glial cells was also disrupted. BDNF secretion from STX1B−/− glial cells was impaired (Fig. 8a), but BDNF secretion was partially recovered by expression of recombinant STX1B in STX1B−/− glial cells (Fig. 8b). Furthermore, expression of STX1B in glial cells was predominant compared to that of STX1A (Fig. 7). These results suggest that STX1B, but not STX1A, is closely involved in the mechanism of BDNF secretion from glial cells. GCM from glial cells was less effective on neuronal survival than glial feeder layer among all genotypes (Fig. 5). Also, glial cells support neuronal survival by direct interaction between neurons and glial cells via cell adhesive proteins and extracellular matrix (Cestelli et al. 1992; Schemalenbach and Müller 1993). Therefore, neurotrophic factors, such as BDNF and NT-3, could support the neuronal survival of STX1B−/− neurons together with direct interaction. It should be clarified in future study.

STX1 forms a SNARE complex by interacting with other SNARE proteins, such as SNAP25 and VAMP2. SNAP25−/− and VAMP2−/− mice die immediately after birth because of the absence of evoked synaptic transmission (Schoch et al. 2001; Washbourne et al. 2002). Hippocampal neurons from SNAP25−/− mice gradually degenerate in vitro (Washbourne et al. 2002; Delgado-Martinez et al. 2007). Recently, we have reported that the expression of other SNARE proteins in whole brain, such as SNAP25, VAMP2, and synaptotagmin1, were not affected by loss of STX1B, while the amount of Munc18-1 was decreased depending on the STX1B gene dosage (Mishima et al. 2014). Munc18-1−/− mice die at birth because of a complete absence of basal and evoked neural transmission (Verhage et al. 2000). The absence of Munc18-1 results in cell-autonomous degeneration of neurons, but application of BDNF induces prolonged neuronal survival of Munc18-1−/− neurons (Heeroma et al. 2004). We found that the expression of Munc18-1 was decreased in STX1B−/− glial cells (Fig. 7a). Therefore, the decrease in Munc18-1 in STX1B−/− glial cells may impair formation of the SNARE complex and reduced neuronal viability, causing neuronal death. Further study is necessary to make clear this possibility.

In summary, we found that STX1A and STX1B play distinct roles in neuronal development and neuronal survival. STX1B was essential for neuronal development and neuronal survival, possibly by regulating the secretion of neurotrophic factors from glial cells.

Acknowledgments and conflict of interest disclosure

This study was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan, to T.F. (no. 22500338); a Grant-in-Aid for Scientific Research (B) to K.A. (no. 19300133, 24300142) from MEXT, Japan; and a Grant-in-Aid from the Promotion and Mutual Aid Cooperation for Private Schools of Japan to K.A. All experiments were conducted in compliance with the ARRIVE guidelines. The authors have no conflict of interest to declare.

Ancillary