Male gerbils of one month (body weight: 25–30 g) and six months (body weight: 65–75 g) of age were used as young and adult groups. The animals were provided by the Experimental Animal Center of Kangwon National University (Chuncheon, Kangwon, Republic of Korea). They were housed in a conventional state (room temperature, 23±0.5°C; humidity, 55±5%; 12:12 light/dark cycle) with freely accessible pellet feed and water. The protocol of this experiment was approved (approval no. KW-200113-1) by the Institutional Animal Care and Use Committee.

The gerbils used in this research were divided into four groups: i) adult sham group (n=28); ii) tFI-operated adult gerbils (adult ischemia group) (n=56); iii) young sham group (n=28); and iv) young ischemia group (n=56).

As previously described in our published papers (4,19,20), the surgical procedure of tFI in the gerbils was performed as follows. Briefly, the gerbils were adequately anesthetized using mixture of 2.5% isoflurane (Baxtor) in 67% nitrous oxide and 33% oxygen (21). Under the anesthesia, both common carotid arteries, which locate at the neck and supply arterial blood to the brain, were isolated and ligated for 5 min with aneurysm clips. Body temperature was checked using a rectal temperature probe and controlled at normothermic condition (37±0.5°C) using a thermometric blanket before and during tFI. After tFI, body temperature was controlled at normothermia until they awaked. The sham gerbils received the same tFI operation without the occlusion of the arteries.

The gerbils in the sham groups were sacrificed at 1 day (n=7) and 4 days (n=7, data not shown) after tFI, the gerbils in the ischemia groups were sacrificed at 1 day (n=7), 4 days (n=7), 7 days (n=7) and 15 days (n=7) after tFI. The gerbils in all groups were deeply anesthetized with 70 mg/kg pentobarbital sodium, and the dose was increased to 200 mg/kg for euthanasia. The confirmation of death was evaluated with vital signs including heart beats, pupillary response, and respiratory pattern (lack of cardiac activity for 5 min through cardiac palpation, unresponsiveness to light with dilated pupils using light into the eyes of the animal, and lack of spontaneously breathing pattern with shallow and irregular breathing pattern). As previously described (22), hippocampal proteins (50 µg for each sample) were denatured, electrophoretically separated, and transferred to a nitrocellulose membranes. The membrane was blocked with TBS-T buffer containing non-fat dry milk (5%) and incubated for 8 h at 4°C with primary antibodies-rabbit anti-TNF-α or rabbit anti-TNF-R1 (diluted 1:1,000; Abcam). The next day, the membrane was incubated with secondary antibody [peroxidase conjugated goat anti-rabbit IgG (Pierce; Thermo Fisher Scientific)] for 1 h at room temperature. Thereafter, the membranes was analyzed using chemiluminescence system according to the manufacturer (Amersham; GE Healthcare.).

The western blot bands of TNF-α and TNF-R1 were scanned by ChemiDoc Imaging System (Bio-Rad Laboratories, Inc. The protein signal was quantified by scanning densitometry using Scion Image software of Scion Corp. Data were normalized to appropriate references (β-actin for total protein).

The gerbils in the sham groups were sacrificed at 1 day (n=7) and 4 days (n=7, data not shown) after tFI, and the gerbils in the ischemia groups were sacrificed at 1 day (n=7), 4 days (n=7), 7 days (n=7) and 15 days (n=7) after tFI. In our current study, to reduce the numbers of the gerbils, the brain sections of the young and adult sham groups were obtained at 1 and 4 days after sham operation. The brain sections containing the hippocampi were prepared as previously described in our studies (4,19,20). In brief, the gerbils in all groups were deeply anesthetized with 70~90 mg/kg pentobarbital sodium (21) at 1, 4, 7 and 15 days after tFI. Under anesthesia, the gerbils were perfused transcardially with saline, and then with 4% paraformaldehyde solution. The brains were obtained and more fixed in the same fixative for 6 h at room temperature. For cutting the brains, they were infiltrated with 30% sucrose solution to avoid tissue damage from freeze. Finally, the brain tissues were coronally sectioned into 30-µm thickness in a cryostat.

To examine neuronal damage/death (loss) in CA1 after tFI, FJB histofluorescence was done as described previously (23). In short, the tissues (sections) were incubated in 0.06% potassium permanganate and stained with 0.0004% FJB (Histochem) for 10 min using a rotator. Thereafter, the sections were rinsed and incubated in 0.0004% FJB (Histochem) for 20 min After washing, the sections were placed on a slide warmer for the reaction of F-J B. Thereafter, the sections were fully dried, cleared by immersion in xylene and coverslipped with dibutylphthalate polystyrene xylene (DPX; Sigma-Aldrich; Merck KGaA).

To count the FJB positive cells (neurons) in CA1, five sections/gerbil were chosen and analyzed using previously described method (4,19,20). The image of FJB positive cells was taken using fluorescence microscope from Carl Zeiss with blue excitation fluorescence filter between 450–490 nm. The FJB positive cells were captured using image capture software (cellSens Standard; Olympus). The captured FJB positive cells were totally counted in 250 µm² at the middle in CA1. The mean number was calculated using NIH Image 1.59 software (NIH; National Institutes of Health).

According to published methods (4,19,20) with modification, immunohistochemical staining using a single antibody (NeuN, TNF-α and TNF-R1) was performed by streptavidin-biotin-peroxidase method. In brief, the sections were immersed with 0.3% hydrogen peroxide to block endogenous peroxidase activity (20 min at room temperature), and 5% normal horse serum was treated to block unspecific proteins in the tissues (30 min at room temperature). And then, the tissues were incubated with each antibody-mouse anti-NeuN (1:1,100, Chemicon), rabbit anti-TNF-α (1:1,100, Abcam) or rabbit anti-TNF-R1 (1:1,000, Abcam)-for 10 h at 4°C. The tissues were rinsed and incubated with biotinylated horse anti-mouse or anti-rabbit IgG (Vector Laboratories, Inc.) for 2 h at room temperature, and continuously reacted with Elite avidin-biotin enzyme complex (ABC; Vector Laboratories, Inc.) for 1 h at room temperature. The visualization of each immunoreaction was achieved with solution of 3,3′-diaminobenzidine (Vector Laboratories, Inc.).

To count NeuN-immunoreactive cells, and to analyze TNF-α- and TNF-R1-immunoreactive structure, we used published methods (4,19,20) In short, five sections/gerbil with 120-µm interval were chosen. The digital images of NeuN-immunoreactive cells were captured in 250 µm² at the middle of CA1 using light microscope (BX53) from Olympus. For TNF-α- and TNF-R1-immunoreactive structure, their digital images were also captured in all layers using BX53 microscope. TNF-α- and TNF-R1-immunoreactivity was evaluated by relative optical density (ROD) according to published methods (19,20) with some modification. Namely, the color image was converted to 8-bit greyscale images with a rage of 0–255 (from black to white). The background level and the variance of the staining intensity in CA1 was calculated on the 0–255 greyscale. The percentile in the corresponding area was analyzed using image analysis software from Image J (NIH).

To confirm the type of TNF-α- and TNF-R1-immunoreactive cells, double immunofluorescence was performed using rabbit anti-TNF-α (diluted 1:400; Abcam)/mouse anti-GFAP (diluted 1:1,000, Chemicon) or rabbit anti-TNF-R1 (diluted 1:400, Abcam)/mouse anti-GFAP (diluted 1:1,000, Chemicon) for astrocytes. The sections obtained at 7 days after tFI were incubated in the mixture of antisera for 8 h at 4°C. They were briefly washed and reacted with the mixture of both Cy3- or FITC-conjugated donkey anti-rabbit or anti-mouse IgG (diluted 1:200; Jackson ImmunoResearch) for 2 h at room temperature. The double immunoreaction was examined with confocal MS from LSM510 META NLO (Carl Zeiss).

The data presented in this study represent the means ± SEM. The data were statistically analyzed using SPSS 18.0 (SPSS, Inc.). The two-way analysis of variance (ANOVA) with a post hoc Bonferroni's multiple comparison test was done to determine differences among groups. P<0.05 was used for statistical significance.

We examined the difference of tFI-induced DND in CA1 of gerbil hippocampus between adult and young gerbils using neuronal nuclear antigen (NeuN, a marker for neuron) immunohistochemistry and Fluoro-Jade B (FJB, a marker for degenerating or dead neuron) histofluorescence (Fig. 1).

In adult and young sham groups, NeuN immunoreactivity was easily identified in pyramidal neurons (88.0±3.8/250 µm²) located in the stratum pyramidale (SP) (Fig. 1A and 1B). In addition, FJB-positive cells were not detected in SP (Fig. 1A and B). This finding means that no damaged cells were present in the sham groups.

In adult ischemia group, a significant loss of NeuN-immunoreactive CA1 pyramidal neurons (11.3±1.9/250 µm²) and many FJB-positive neurons (78.1±1.7/250 µm²) were observed in SP at 4 days after tFI. In this group, the number of NeuN-immunoreactive neurons was approximately 13% of that of the adult sham group (Fig. 1C and Cc). Thereafter, the numbers of NeuN-immunoreactive neurons and FJB-positive cells at 7 and 15 days after tFI were similar to those at 4 days after tFI (Fig. 1E, Ee, G, Gg, I and J).

In the young ischemia group, the number of NeuN-immunoreactive pyramidal neurons at 4 days after tFI was similar to that in young sham group, although the NeuN immunoreactivity was decreased (Fig. 1D, Dd and I). However, at 7 and 15 days after tFI, the number of NeuN-immunoreactive pyramidal neurons was distinctly decreased (approximately 48.5 and 47.9%, respectively), compared with that in the young sham group (Fig. 1F, H and I). In addition, we found that many FJB-positive pyramidal neurons were observed at 7 and 15 days after tFI, showing that the mean percentage of the neurons was approximately 57 and 55%, respectively, of that in the adult ischemia group, respectively (Fig. 1f, h and J).

To compare the changes of TNF-α and TNF-R1 protein levels in ischemic CA1 between adult and young gerbils, western blot was conducted in CA1 after tFI (Fig. 2).

Protein levels of TNF-α in the adult and young CA1 were significantly changed with time after tFI (Fig. 2). In both groups, TNF-α protein levels were gradually increased from 1 day after tFI, however, TNF-α protein levels were significantly higher in the young group than that in the adult group, showing that relative density in the young was 1.1-fold at 1 day, 1.5-fold at 4 days, 1.2-fold at 7 days and 1.6-fold at 15 days after tFI compared with that in the corresponding adult groups (Fig. 2).

The change pattern of TNF-R1 protein levels in the adult and young was not similar to the changes of TNF-α protein levels, but the changes were significant (Fig. 2). At 1 day after tFI, TNF-R1 protein levels were dramatically increased in both groups (2.1-fold in the adult and 2.1-fold in the young) compared with those in each sham group (Fig. 2). Thereafter, TNF-R1 protein levels in both groups were irregularly altered, showing that relative density in the young groups was 1.7-fold at 4 days, 0.4-fold at 7 days, and 0.5-fold at 15 days, respectively, after tFI compared with that in the corresponding adult groups (Fig. 2).

To investigate the change of TNF-α immunoreactivity in CA1, immunohistochemistry for TNF-α was conducted in adult and young gerbils after tFI (Fig. 3).

Weak TNF-α immunoreactivity was mainly found in CA1 pyramidal neurons in the adult sham group (Fig. 3A and K). In the adult ischemia group, a significant increase of TNF-α immunoreactivity (172.5% of the adult sham group) was found in the pyramidal neurons at 1 day after tFI, compared with that in the adult sham group (Fig. 3C and K). At 4 days after tFI, TNF-α immunoreactivity was hardly observed in the pyramidal neurons; however, at this time, TNF-α-immunoreactivity began to be expressed in non-pyramidal cells located in the stratum oriens (SO) and stratum radiatum (SR) (Fig. 3E and K). Thereafter, TNF-α-immunoreactivity was mainly observed in the non-pyramidal cells, and the relative optical density (ROD) of the TNF-α-immunoreactivity in CA1 was markedly increased (339.1% in the adult and 357.4% in the young) at 7 and 15 days after tFI compared to that in the adult sham group (Fig. 3G, I and K).

In the young sham group, no significant difference in TNF-α immunoreactivity in the pyramidal neurons was found when compared with that in the adult sham group (Fig. 3B and K). In the young ischemia group, TNF-α immunoreactivity in the pyramidal was significantly increased at 1 and 4 days after tFI, and the ROD in CA1 was increased (161.2% in the adult and 156.7% in the young) compared to that in the corresponding adult ischemia group (Fig. 3D, F and K). At 7 and 15 days after tFI, TNF-α immunoreactivity in the pyramidal neurons was still observed, and non-pyramidal cells newly expressed TNF-α immunoreactivity, showing that the ROD in CA1 was increased (117.4% in the adult and 125.8% in the young) compared to that in the corresponding adult ischemia group (Fig. 3H, J and K).

As described above, TNF-α immunoreactivity was shown in non-pyramidal cells at late time after tFI. To identify the cell types of the non-pyramidal cells containing TNF-α immunoreactivity, we performed double immunofluorescence staining for TNF-α/GFAP (a marker for astrocyte), and we found that most of TNF-α-immunoreactive non-pyramidal cells were colocalized with GFAP-immunoreactive astrocytes in the adult and young ischemia groups (Fig. 4).

To examine the change of TNF-R1 immunoreactivity in CA1, immunohistochemistry was done in the adult and young groups (Fig. 5).

In the adult sham group, TNF-R1 immunoreactivity was easily shown in pyramidal neurons (Fig. 5A). In the adult ischemia group, TNF-R1 immunoreactivity was significantly increased (254.9% of the adult sham group) in the pyramidal neurons at 1 day after tFI (Fig. 5C and K). Thereafter, TNF-R1 immunoreactivity in the pyramidal neurons was hardly shown at 4 days after tFI, but TNF-R1 immunoreactivity was newly shown in non-pyramidal cells, and TNF-R1 immunoreactivity in CA1 was 96.4% of the adult sham group (Fig. 5E and K). At 7 and 15 days after tFI, TNF-R1 immunoreactivity was more increased in the non-pyramidal cells, and TNF-R1 immunoreactivity in CA1 was 166.6 and 151.6% of the adult sham group, respectively (Fig. 5G, I and K). These non-pyramidal cells containing TNF-R1 immunoreactivity were identified as GFAP-immunoreactive astrocytes by double immunofluorescence (Fig. 6A, C and E).

In the young sham group, TNF-R1 immunoreactivity was also found in pyramidal neurons, and the immunoreactivity was similar to that in the adult sham group (Fig. 5B and K). In the young ischemia group, TNF-R1 immunoreactivity in the pyramidal neurons was increased only in the pyramidal neurons, showing that the TNF-R1 immunoreactivity was 93.0% at 1 day and 190.9% at 4 days after tFI compared with that in the corresponding adult ischemia group (Fig. 5D, F and K). At 7 and 15 days after tFI, TNF-R1 immunoreactivity in CA1 was markedly decreased (68.1% in the adult and 59.9% in the young) compared to that in the young sham group (Fig. 5H, J and K). At these times, TNF-R1 immunoreactivity was only shown in a few pyramidal neurons; TNF-R1 immunoreactivity was hardly detected in astrocytes by double immunofluorescence (Fig. 6B, D and F).