Behavioral analysis of rats subjected to brain-derived neurotrophic factor (BDNF) knockdown (KD) in the dorsal dentate gyrus (dDG). Lentiviral vector (LV)-shBDNF (knockdown) or LV-shSCR (control) constructs were injected into the dDG of mature rats. The battery of behavioral tests was initiated 3 weeks after recovery from the surgery and the following parameters were measured: sucrose preference (a), activity in a modified forced swim test (FST) (b), home-cage locomotion (c), exploration of a novel environment (d), latency to reach the platform in the water maze task (e) and percent of time spent at the target quadrant during the probe trial in the water maze task (f). Sucrose preference was assessed over a 10-day period using a 0.2% sucrose solution (t(43)=3.526, P=0.01). A modified FST was videotaped, and activity was scored and analyzed using our locally developed software which involves a computer attached joystick³⁵ (t(38)=2.183, P<0.05). Locomotion was automatically measured over four consecutive nights in the home cage of rats (t(41)=2.855, P<0.01). The total distance traveled and number of rearings were automatically quantified in the test for exploration in a novel environment. The water maze task consisted of five training sessions over 5 days, and a test day 3 days later. A two-way repeated measures analysis of variance (ANOVA) revealed a significant Group X session interaction (F(1,26)=4.835, P<0.05). In the water maze task, the latency to find the platform was significantly lower in control rats than in BDNF KD rats at days 2, 3 and 4 (t(26)>2.115, P<0.05) but not at days 1 and 5. The probe trial was conducted 3 days after the training sessions, and the percentage of time spent at the target quadrant was measured during a 1-min test. The level of a random distribution between quadrants was 25%. Values are mean±s.e.m. (n=20 to 23 per group for all tests except for the water maze in which n=14 per group; ^*P<0.05, ^**P≤0.01).

The behavioral effect of brain-derived neurotrophic factor (BDNF) knockdown in the ventral subiculum (vSUB). Lentiviral vector (LV)-shBDNF (knockdown) or LV-shSCR (control) constructs were injected into the vSUB of mature rats. The battery of behavioral tests was initiated 3 weeks after recovery from the surgery and the following parameters were measured: sucrose preference (a), activity in a modified forced swim test (FST) (b), home-cage locomotion (c), exploration of a novel environment (d), latency to reach the platform in the water maze task (e) and percent of time spent at the target quadrant during the probe trial in the water maze task (f), as described in Figure 2. Values are mean±s.e.m. (^* P<0.05, n=11 per group). A significant difference between the experimental groups was observed only in the sucrose preference test (t(20)=2.418, P<0.05).

Validation of LV-shBDNF infection and brain-derived neurotrophic factor (BDNF) knockdown (KD) in vitro and in vivo. BDNF protein expression was measured in a C6 rat glioma cell line and in vivo in response to infection with lentiviral vector (LV) expressing either green fluorescent protein (GFP) only (LV-GFP; infection control), scrambled shRNA (LV-shSCR; shRNA control) or shRNA complementary to the coding axon of the rat BDNF gene (LV-shBDNF; active sequence). Panel a presents BDNF levels measured in a C6 cell medium 48 h after infection and normalized per 10⁶ cell number. Data are presented as percentages of BDNF secreted from noninfected control cells. One-way analysis of variance (ANOVA) reveals a significant main effect of treatment (F(3, 20)=374.756, P<0.001). Differences between the various LV treatments were assessed using Fisher post hoc analysis (^***P<0.001; n=6 per treatment). Panel b presents BDNF levels in the dentate gyrus (DG) measured by immunohistochemistry and epifluorescence microscopy. The intensity per unit area was measured for each slice. BDNF levels are expressed as a percentage of control, by measuring the intensity in slices taken from infected rats relative to control slices taken from the noninfected rats. One-way ANOVA revealed a significant main effect of treatment (F(3, 17)=29.926, P<0.01). Differences between control and infected brains were assessed using Fisher post hoc analysis (^**P<0.01; n=5 to 6 per group). Panel c presents BDNF levels in the dDG measured by ELISA and normalized per total protein. BDNF levels are expressed as a percentage of control. One-way ANOVA revealed a significant main effect of treatment (F(3, 14)=5.868, P<0.01). Differences between control and infected brains were assessed using Fisher post hoc analysis (^**P<0.01; n=4 or 5 per group). Panel d presents representative micrographs of hippocampal slices including the dDG from the rats injected with LV-shSCR (control) or LV-shBDNF (BDNF KD). The micrographs in the left present the infection spread based on the GFP expression (green) within the sections. BDNF protein expression was visualized by immunohistochemical staining (middle lane). Note the significant reduction in BDNF expression induced by the LV-shBDNF microinjection. Cell nuclei were visualized using Hoechst staining (lower lane). The micrographs on the right show a higher magnification of an infected hippocampal region. The upper lane shows GFP expression as an infection marker, whereas the middle lane shows BDNF protein staining. The lower lane shows superimposition of the upper and middle lanes to highlight cells expressing both GFP and BDNF. The arrows highlight that while the rat brain infected with LV-shSCR show BDNF expression within infected (GFP) cells, none of the infected cells in the LV-shBDNF rat brain show BDNF expression. Scale bar=200 μm at the left micrographs and 20 μm at the right micrographs.

Behavioral analysis of rats subjected to brain-derived neurotrophic factor (BDNF) knockdown (KD) in the CA3. Lentiviral vector (LV)-shBDNF (knockdown) or LV-shSCR (control) constructs were injected into the CA3 of mature rats. The battery of behavioral tests was initiated 3 weeks after recovery from the surgery and the following parameters were measured: sucrose preference (a), activity in a modified forced swim test (FST) (b), home-cage locomotion (c), exploration of a novel environment (d), latency to reach the platform in the water maze task (e) and percent of time spent at the target quadrant during the probe trial in the water maze task (f), as described in Figure 2. Values are mean±s.e.m. (n=8 per group).

The effect of brain-derived neurotrophic factor (BDNF) knockdown on hippocampal neurogenesis. Neurogenesis in the subgranular zone (SGZ) was measured 2 months after the induction of BDNF knockdown (KD) in the dorsal dentate gyrus (dDG), CA3 or ventral subiculum (vSUB) of mature rats. Rats were injected intraperitoneally with 50 mg kg⁻¹ bromodeoxyuridine (BrdU) at 12 h intervals for 2 days. Twenty-four hours or a week later rats were deeply anesthetized and transcardially perfused. Immunohistochemistry was performed to detect neuronal dividing cells by counting the cells that incorporated BrdU into their genome, and express the immature neuronal marker doublecortin X (DCX), as shown in the small micrographs. Representative micrographs of the dDG from rats injected with lentiviral vector (LV)-shSCR (Control) or LV-shBDNF (BDNF KD) into their dDG are presented in panel a. Proliferation was measured by counting BrdU-stained cells (upper lane), immature neurons were visualized by DCX (middle lane), and differentiation was measured by counting double-labeled BrdU and DCX cells (lower lane). Scale bar=200 μm. Panel b shows a quantitative analysis of proliferation comparing control and BDNF KD rats after LV infections in the dDG. This panel presents data from rats that were euthanized 24 h after BrdU injections. Panels c,d and e show a quantitative analysis of proliferation and differentiation comparing control and BDNF KD rats after LV infections in the dDG, CA3 or vSUB, respectively. These rats were injected with BrdU a week before euthanasia. A significant effect of BDNF KD was observed only in the dDG group and only in the number of differentiated neuronal cells (BrdU + DCX) (t(11)=5.672, P<0.001), but not in total proliferation (BrdU). Values are mean±s.e.m. (^*** P<0.001, n=6–9 per group).