All procedures were performed in accordance with European Directives (2010/63UE) and were approved by a local ethical committee (Comité Régional d’Ethique en Matière d’Expérimentation Animale de Strasbourg CEEA 35, reference # 02965.02). The APPSWE (Tg) mouse line carries a transgene coding for the 695-amino acid human amyloid precursor protein (APP) isoform containing the double Swedish mutation (K670N, M671L) [34] which leads to high brain concentrations of Aβ and significant amyloid deposition around the age of 8 months. Female Tg and non-transgenic (NTg) littermates (model 1349-RD1-F, Taconic Bioscience, Germantown, USA) were Rd1 tested to ensure these subjects were not homozygous for the Pde6b^rd1 mutation (i.e., not blind). Thirteen female Tg and NTg mice were assigned to the experiment aiming to define the earliest age of deficits in the PS task. Thirty-five Tg [24 tested and 11 home cage (HC)] and 15 NTg (10 tested and 5 HC) female mice were assigned to the Tg mice anti-Aβ antibody study. Forty-nine male C57BL/6J (B6) mice were used in the aged mice anti-Aβ antibody study: 35 mice were 20-month-old (24 tested and 11 HC) and 14 mice were five-month-old mice (10 tested and 4 HC). Upon arrival, mice were group-housed in standard cages with food and water available ad libitum. Nesting material and a few food pellets were added to the bedding to favor natural behavior expression. A plastic tunnel (4 cm in diameter, 15 cm long) was provided one week before behavioral testing. Two days prior to behavioral testing, mice were familiarized with tube transportation from their HC to their holding cage and vice versa. The same tube was used for all transports between HC, holding cage, and testing apparatus [38]. The housing room was kept at a controlled temperature (23°C ± 1°C) with a 12/12-hour light/dark cycle (lights on at 8.00 AM). Behavioral testing took place between 9.00 AM and 4.00 PM. General health status was checked before the beginning of the experiments and animals lacking vibrissae (barbering) or showing eye abnormalities (e.g., cataract or closed eye) were not tested.

The PS task is based on the spontaneous tendency of mice to explore an object displaced in a familiar context [39]. It resembles the classical object location test except that in our PS test, the object displacement was set at 20 cm, the distance detected by six-month-old B6 mice but not 20-month-old mice, as described in Cès et al. [29]. A longer 35-cm displacement test was integrated into the task as a control for the preserved ability to detect large object displacement. Briefly, mice were first familiarized with the large, dimly lit open field (92.5 cm × 92.5 cm × 35 cm; Ugo Basile, Italy; ~11 lux in the center) for 8–10 min. During the next three days, they underwent a daily two-trial session: i) habituation with no object displacement, ii) testing with long-distance 35 cm displacement, and iii) testing with the critical 20 cm object displacement, with each day involving a different set of objects (see Figure 1A). During each session, mice could explore the two objects located 20 cm apart diagonally in the open field during the acquisition trial. The sessions lasted for 12 min in studies using young Tg and NTg female mice and 15 min in the study using old B6 male mice, as aged mice often explore less than young mice [29]. After a 5-min intermission in their holding cage, mice returned to the open field for a 6-min retention trial with one of the two objects diagonally moved 0, 35, or 20 cm away from the fixed object. Object exploration was only considered when mice engaged in active exploration i.e., when the nose was within two cm facing the object. Discrimination performance was evaluated as the percentage of time spent exploring the displaced object relative to the total time spent exploring the two objects during the first three minutes of the test trial, as the mice tended to ignore the objects during the remaining three minutes. All objects were made of glass or plastic (5–6.5 cm diameter and 4.5–7 cm high). Data from one vehicle-treated aged mouse was excluded due to exaggerated thigmotaxis behavior, which biased performance (0 cm: 0%; 35 cm: 56.9%; 20 cm: 27.1%).

Aβ passive immunization was performed with the m266 antibody (gift from Boehringer Ingelheim, Biberach an der Riss, Germany). The m266 stock solution was at a concentration of 2.2 mg/mL in 20 mM sodium acetate (Sigma-Aldrich), 140 mM sodium chloride (Fisher Chemical, Denmark), and 0.02% Tween 80 (pH 5.5) (P1754, Sigma-Aldrich). Female Tg and NTg mice and male B6 mice received a five mg/kg (0.1 mg/mL, 10 mL/kg) intraperitoneal (IP) injection of m266 or its vehicle (20 mM sodium acetate, 140 mM sodium chloride, and 0.02% Tween 80) once a week during four weeks at 4.30 PM (see Figure 1A). This protocol was adapted from a study on PDAPP mice, an amyloid mouse model [40]. The last injection was given on the day of familiarization to the open-field, three days before PS testing. The day after the first m266 injection, all mice received an IP injection of 5-bromo-2-deoxyuridine [BrdU; B5002, Sigma-Aldrich, Hamburg, Germany; diluted in 0.1 M phosphate buffer (0.081 M Na₂HPO₄, 0.018 M NaH₂PO₄)] (Carl Roth, Karlsruhe, Germany), pH 8.4, 40°C, 100 mg/kg/injection, 0.1 mL/kg/injection twice a day (8.30 AM, 4.30 PM) for four days. The last BrdU injection was made 28 days before testing PS performance.

Ninety minutes after the PS test completion, brains were processed for immunohistochemistry. We collected brains from five NTg, five vehicle-, and six m266-treated Tg female mice, as well as four young male B6 mice, five vehicle-, and six m266-treated aged male B6 mice from the tested groups. For control groups [mice which remained in their home cage (HC groups) during the whole process as controls for the tested groups], we collected brains from five NTg, four vehicle-, and six m266-treated Tg female mice, four young male B6 mice, six vehicle-, and five m266-treated aged male B6 mice. Under deep anesthesia [120 mg/kg pentobarbital (Ceva Santé Animale, Libourne, France), 10 mL/kg, IP], an intracardiac perfusion with phosphate-buffered saline (PBS) [0.01 M PBS, 0.1% heparin (Héparine Choay, CheplaPharm, Arzneimittel GmbH, GR, Germany)] was carried out, followed injection of a fixative solution [4% paraformaldehyde (Carl Roth GmbH + Co KG, Germany) in 0.1 M phosphate buffer, pH 7.4, 4°C]. Brains were postfixed for 24 h in 4% paraformaldehyde and stored in PBS all at 4°C. Fifty µm-thick coronal sections were obtained with a vibratome (Leica Biosystems) and were stored free-floating at –20°C in a cryostorage solution [30% glycerol, 30% ethylene glycol (both from Carl Roth GmbH, Karlsruhe, Germany), 0.026 M phosphate buffer].

NTg and Tg brain sections were PBS-rinsed, and then incubated for 15 min in 2 N hydrochloric acid (Carl Roth GmBH) at 37°C. After washing, they were blocked in 5% normal goat serum in PBS with 0.5% Triton X100 (T9284, Sigma-Aldrich), followed by a 48 hour-incubation at room temperature in a solution combining three antibodies: monoclonal rat anti-BrdU antibody (1/500, ab6326, Abcam) with polyclonal rabbit anti-Egr-1 (1/300, 4153S, Cell signaling, Massachusetts, USA), and polyclonal guinea pig anti-doublecortin (DCX) antibodies (1/5,000, ab2253, Millipore, Temecula, CA, USA). After several washes, sections were incubated in secondary antibody solution: goat anti-rat Alexa 555 with goat anti-mouse Alexa 647 and goat anti-rabbit Alexa 488, or with goat anti-guinea pig Alexa 488 and goat anti-rabbit Alexa 647 (for all: 1/500, Life Technologies, Eugene, OR, USA) 2 h at room temperature. After final washing, sections were mounted in Fluoromount (F4680, Sigma-Aldrich) and stored at 4°C.

BrdU, a thymidine analog incorporated in DNA into dividing cells can be used for birth dating. DCX is commonly recognized as a marker of adult-born neuronal cells within the DG [41–44] allowing the identification of new neuronal progenitors (around 8–15 days post-birth) and new immature neurons (around 15–30 days post-birth) [45, 46]. Thus, in our study, co-labeling of BrdU⁺/DCX⁺ was used to identify immature neurons. Egr-1 is an immediate early gene whose expression is associated with neuronal activity as already shown in adult-born and pre-existing granular cells of the DG [42, 47–52].

Immunohistochemistry for aged mice was carried out with a colorimetric protocol to avoid age-dependent autofluorescence (age-dependent accumulation of lipofuscins). Thus, for Egr-1 staining, brain sections were PBS-rinsed and then incubated for 15 min in 2 N hydrochloric acid (Carl Roth GmbH, Karlsruhe, Germany) at 37°C followed by washing. After washing, sections were incubated in 0.3% hydrogen peroxide (216763, Sigma-Aldrich) in methanol for 30 min at room temperature to quench endogenous peroxidase activity, rinsed several times, blocked in 5% normal goat serum in PBS-0.5% Triton X100 and then incubated in the polyclonal rabbit anti-Egr1 (1/1,000, 4153S, Cell signaling) during 48 h at room temperature. After several washes, sections were incubated with biotinylated goat anti-rabbit antibody (1/500, Vector Laboratories, Burlingame, California) for two hours at room temperature followed by a 45 min incubation with avidin-biotin-peroxidase complex (Vectastain ABC kit; Vector Laboratories, Burlingame, California, USA). Staining was developed at room temperature with 3,3’-diaminobenzindine (DAB kit, Vector Laboratories, Burlingame, California, USA), and the reaction was stopped by rinses in PBS.

Triple immunostainings were analyzed with the Leica Application Suite X Office (Leica microsystems) using confocal images obtained with a Leica TCS SP5 II. Cell counting was carried out in the granule cell layer (GCL), including the subgranular zone (two-cell-thick region along the inner GCL border) of the dorsal DG (called after in the paper just “GCL”), in every fifth section (50 µm thick, 250 µm apart) spanning between –0.94 mm (first section) to –2.06 mm from the bregma. Counting was performed through the entire slice in five sections per mouse. Only well-defined and round/ovoid nuclei were counted. These are indeed more likely to correspond to new GCs since new glial cells migrate mainly into the hilum and rarely into the GCL [53]. The experimenter was blind to both the genotype and treatment. The GCL area was measured for each section with ImageJ software (National Institute of Health, Bethesda, MD), on DG images captured with the 4× objective of an Olympus AHBT3 microscope [54]. GCL volume for each section was obtained by multiplying the total GCL area (left GCL + right GCL) by the section thickness. For each mouse, the density of BrdU or Egr-1 positive cells (BrdU⁺ or Egr-1⁺) was calculated by dividing the total counts of BrdU⁺ or Egr-1⁺ cells by the total volume of GCL screened. The total number of BrdU⁺ cells was calculated by multiplying the density by the GCL reference volume. This reference volume was determined by multiplying the GCL area by the distance between sampled sections (250 μm). The percentage of cells expressing the immature neuron maker (DCX) among the four-week-old BrdU⁺ cell population was calculated as: 100× number of double stained cells/total number of BrdU⁺ cells. The percentage of triple stained cells (BrdU⁺/DCX⁺/Egr-1⁺) was calculated as: 100× number of triple stained cells/total number of BrdU⁺/DCX⁺ neurons.

For Egr-1 staining of B6 sections, digital images of the DG in both hemispheres were captured using a Hamamatsu NanoZoomer S60 Digital slide scanner. Five consecutive coronal sections (each 250 µm apart) were taken from each mouse. The GCL of the DG was delimitated based on the 3rd edition brain atlas of Franklin and Paxinos [55]. The images of the delimitated GCL were converted into 8-bit greyscale and immunopositive cells were counted using ImageJ. Egr-1 positive cells were defined as having immunopositive nuclei (diameter of 4–20 µm, sphericity of above 0.1) stained below a grayscale threshold set directly below the background peak visualized by a pixel intensity histogram. The greyscale threshold was set manually by an experimenter who remained blind to group conditions. Cell density was defined as the total number of Egr-1 positive nuclei divided by the total volume/area across all sections of the GCL.

After cervical dislocation, hippocampal regions from four vehicle- and six m266-treated Tg mice, five young B6 mice, four vehicle and six m266-treated aged B6 mice were quickly removed, frozen in liquid nitrogen, and stored at –80°C. Aβ quantification was performed on the CA3/DG region, as its dentate adult-born neurons are thought to be involved in PS function. Samples were homogenized in 10 volumes of ice-cold guanidine buffer (5 M guanidine-HCl/50 mM Tris–Cl, pH 8.0, G3272 and T3253 respectively, Sigma-Aldrich), mixed for three hours at room temperature and diluted to 1:10 in PBS containing 1% BSA, 0.05% Tween20 (from the ELISA kit) supplemented with protease inhibitor cocktail (P8340, Sigma-Aldrich), phosphatase inhibitor cocktail (PhosStop, 4906845001, Roche Diagnostics GmbH, Mannheim, Germany), and 1 mM phenylmethylsulfonyl fluoride (Sigma-Aldrich). The CA3/DG homogenates were centrifuged at 16,000 g for 20 minutes at 4°C and supernatants were aliquoted. Protein concentration was measured using the Bio-Rad Protein Assay (5000006, Bio-Rad Laboratories, CA, USA). Total levels of human Aβ40 and Aβ42 were determined using the human Aβ(1–40) and Aβ(1–42) high sensitivity ELISA kits (RE59781 and RE59791, respectively) whereas total levels of murine Aβ40 and Aβ42 were determined by using the amyloid-beta (1–40) (Mouse/Rat) and the amyloid-beta (1–42) (Mouse/Rat) ELISA kits (RE45741 and RE45721, respectively). All kits were purchased IBL International (Hambourg, Germany) and used following the manufacturer’s instructions. The signal was normalized to the protein concentration for each sample. All dosages were done in duplicate, and samples with > 10% variation between duplicates were discarded from analysis.

Statistical analysis was completed using GraphPad Prism 7 (GraphPad Inc) and Statistica 13.3 (TIBCO Software Inc). For behavioral data, the main effects were tested with a two-way ANOVA with repeated measures (group × distance). For each distance, groups were compared with a one-way ANOVA when relevant (group). ANOVAs were followed by the post hoc Tukey’s multiple comparisons test (Tukey). In addition, the performance level for each group was compared to the 50% chance level using student’s t test. For immunohistochemical experiments, data sets were analyzed using the non-parametric Kruskal-Wallis (KW), Mann-Whitney (MW), and Wilcoxon (W) tests, as group sizes were relatively small. The Wilcoxon signed rank test (WSR) was used to compare upper and lower blade data. Two-way ANOVAs, with repeated measures when relevant (lower and upper blades), were also performed to test possible interactions (groups, HC, and tested conditions). Whenever outcomes were complementary or differed from non-parametric tests, parametric statistics were reported in the result section. The MW was used for ELISA data. Outliers were identified with the ROUT method in GraphPad prism. Differences were considered significant at P < 0.05. All data are given as mean ± standard error of the mean (SEM) including data analyzed with non-parametric tests instead of median and inter-quartile range.

The behavioral task measures the ability of the mice to detect the move of one of two similar objects. For this experiment, we used the 35-cm displacement as the easy distance and the 20-cm displacement as the difficult/PS distance (Figure 1A). From the age of three months, eight Tg mice and five NTg female mice were tested every month to find the age of onset of PS deficits. At the age of three months, both Tg mice and NTg mice performed above chance level on the easy 35-cm test and the difficult/PS 20-cm test (Figure 1B). When retested at the age of four months, Tg mice were selectively impaired in the detection of the difficult/PS 20-cm displacement while NTg mice maintained high performance on both 20-cm and 35-cm tests (group: F1,11 = 6.95, P = 0.02; test: F1,11 = 10.4, P = 0.008; group × test: F1,11 = 6.1, P = 0.03; Figure 1B). We verified that the two groups similarly explored objects during acquisition trials (Figure S1). Taken together, these results indicate that a specific deficit in the spatial PS test appeared in 4-month-old Tg mice.

We then tested the effect of an immunotherapy treatment using m266 anti-Aβ antibody on PS performance. Four-month-old Tg female mice (n = 12) were treated with m266 (5 mg/kg body weight) or vehicle (n = 12) once per week for four weeks with the last injection given on the day of familiarization to the open field (Figure 2A). Due to an object handling error in the PS test, we discarded one vehicle-treated mouse (final n = 11). Vehicle-treated NTg mice (n = 10) were used as an internal control for the validity of the PS test (20 cm displacement). The performances of the different groups differed depending on the distance tested (group: F2,30 = 9.1, P = 0.0008; test: F1,30 = 15.6, P = 0.0004; group × test: F2,30 = 13, P < 0.0001; Figure 2B). All groups performed similarly and above chance level in the 35-cm test. As expected, PS performance of five-month-old vehicle-treated Tg mice did not differ from chance level, as found in four-month-old Tg. However, m266 treatment was able to statistically prevent this early deficit in Tg mice (F2,30 = 19.1, P < 0.0001) which performed as well as vehicle-treated control NTg mice (Figure 2B). Of note, both Tg groups explored the objects less than NTg mice; m266 did not improve this genotype effect on exploration behavior (group: F2,30 = 7.6, P = 0.002; test: F1,30 = 8.1, P = 0.008; group × test: F2,30 = 0.4, P = 0.7; Figure S2A), suggesting the treatment had no effect on object exploration behavior during this task.

We next tested whether m266 treatment could reverse PS deficits previously described in aged B6 male mice [29]. B6 males were treated with m266 (5 mg/kg body weight, n = 12) or vehicle (n = 12) once per week during four weeks with the last injection administered the first day of behavioral testing (Figure 2A) at 20 months of age (Figure 2C). Young vehicle-treated B6 males (n = 10) served as internal controls to validate the PS test at the age of 4 months. Group performances differed as a function of the distance tested (group: F2,30 = 10.6, P = 0.0003; test: F1,30 = 49.8, P < 0.0001; group × test: F2,30 = 4.6, P = 0.02; Figure 2C). The young mice group outperformed both aged groups in the PS test. Notably, aged vehicle- and m266-treated groups showed similar PS deficits with performance at chance level: chronic m266 treatment did not prevent PS deficits in aged B6 mice. Both aged groups succeeded in detecting the 35-cm object displacement. During the acquisition trials, all groups spent similar time exploring the two objects (group: F2,30 = 0.4, P = 0.7; test: F1,30 = 1.8, P = 0.2; group × test: F2,30 = 0.3, P = 0.7; Figure S2B). Thus, in aged mice, m266 treatment had no effect on initial exploration of the objects and retention performances in both the 35-cm and the PS tests.

The GCL of the dorsal DG is considered part of the neuronal circuit involved in PS. To understand whether PS deficits in Tg and aged mice could result from impaired neuronal activation during the test, we investigated Egr-1 immunostaining in the dorsal GCL. Increase of Egr-1 expression in adult-born and pre-existing granular cells of the DG was already shown after behavioral tasks [47–49, 51]. We also tested whether PS deficiency is associated with impaired adult neurogenesis by staining newly formed neurons with anti-DCX and anti-BrdU antibodies. Thus, triple BrdU/DCX/Egr-1 immunostainings were carried out in young NTg and Tg female mice (Figure 3A). Due to the extremely low levels of newborn neuronal cells in the GCL of aged B6 male mice [29], we only carried out Egr-1 immunostaining in this group and its young counterpart. These immunostainings were performed on brains collected 90 min after completion of the PS test, to allow Egr-1 protein expression, in a subset of NTg, Tg, and B6 groups treated with vehicle or m266, and a subset of NTg, Tg, and B6 groups also treated with m266 or vehicle but left in their HC (non-tested).

In the Tg mice experiment, the density of Egr-1⁺ cells in the GCL was increased about 2-fold in the tested mice (mean = 29.7 ± 3.6 cells/0.001 mm³) compared to the HC mice (mean = 14.2 ± 1.3 cells/0.001 mm³; MW: U = 18, Z = –3.9, P < 0.0001; one m266-treated Tg was excluded as an outlier, Figure 3B and Table S1), showing that neuronal activity has indeed increased in the GCL of mice tested for PS. However, the level of GCL activation clearly differed among the three tested groups (KW: H = 6.3, P = 0.04) with m266-treated Tg mice showing lower levels than vehicle-treated Tg mice (MW: U = 1, Z = 2.4, P = 0.02). The level of activation remained stable among HC groups (KW: H = 0.6, P = 0.7). Thus, a global activation of the GCL was induced by the behavioral task compared to HC controls and this activation was lower in m266-treated Tg mice.

When the upper and lower blade of the DG were quantified separately, we found that the upper blade was globally more active than the lower blade (pooled mice WSR: Z = –4.3, P < 0.0001, Figure 3C, Table S1). This effect was even more pronounced in tested mice than in HC mice (pooled tested mice: upper = 42.2 ± 3.5 versus lower = 22.1 ± 1.7 cells/0.001 mm³, WSR: Z = 3.4, P = 0.0007; pooled HC groups: upper = 17.7 ± 2.8 versus lower = 13.0 ± 1.7 cells/0.001 mm³, WSR: Z = 2.2, P = 0.03; Figure 3C). Interestingly, the lower activation of the DG observed in tested m266-treated Tg mice was associated with lower activation of the upper blade compared to tested vehicle-treated Tg mice (Egr-1⁺ cell density in m266-treated Tg mice: 32.1 ± 3.8, and in vehicle-treated Tg mice: 50.8 ± 5.5; MW: U = 2.0, Z = 2.2, P = 0.03; Figure 3C and Table S1). Thus, the behavioral task activated the upper blade of the GCL in mice, while m266 treatment selectively limited this activation in Tg mice.

In the aging experiment, the density of Egr-1+ cells in the GCL was significantly higher in tested groups than in HC groups (pooled tested mice: 18.2 ± 1.0 cells, pooled HC mice: 6.9 ± 0.2 cells; MW: U = 0, Z = –4.7, P < 0.0001; Figure 3D, Table S1, and Figure S3), a result comparable to the one observed in the Tg experiment. The level of activated cells was similar among the three HC groups (KW: H = 0.2, P = 0.9) and among the three tested groups (KW: H = 3.8; P = 0.15; Figure 3D). The increased number of activated cells in the GCL of tested mice was more pronounced in the upper blade (upper: 20.5 ± 0.8 cells, lower: 16.1 ± 0.7 cells; pooled tested mice W: Z = 3.4, P < 0.001, Table S1) whereas, contrary to the Tg experiment, the lower blade was more active in HC group in B6 mice (pooled groups: W: Z = 3.4, P < 0.001; Figure 3E). For each group, the difference in activation between upper and lower blade was significant only in aged B6 groups (vehicle-treated aged mice W: Z = 2.0, P < 0.05, m266-treated aged mice: W: Z = 2.2, P < 0.05; young mice W: Z = 1.5, P > 0.10). There was no effect of the m266 treatment on task-induced GCL activation in the aged groups in either blade. Thus, the behavioral task induced a similar global activation of the GCL regardless of group and treatment.

About one third of adult-born neurons survive up to one month [56] and neurogenesis is thought to be involved in PS performance. Thus, using BrdU injections, we labeled cells born (BrdU⁺) a few days after passive immunization began (Figure 2A) in order to quantify them one month after PS completion in tested groups or their control (HC) counterparts. This quantification was carried out only in the Tg experiment since aged B6 mice showed dramatically reduced levels of neurogenesis in the GCL [29]. The density of BrdU⁺ cells was stable among groups and conditions (two to three cells/0.001 mm³ GCL; Figure 4A and Table S2). There were significantly more BrdU⁺ cells in the upper blade than in the lower blade regardless of group and condition [pooled mice (mean ± SEM), lower: 2.6 ± 0.2 and upper: 4.0 ± 0.3, W: Z = 4.8, P < 0.00001; pooled tested mice for lower: 2.6 ± 0.3 and upper: 3.9 ± 0.4, W: Z = 3.4, P < 0.001; pooled HC mice for lower: 2.7 ± 0.2 and upper: 4.0 ± 0.3, W: Z = 3.4, P < 0.001, see Table S2]. In addition, there were no significant differences between groups for the lower blade or for the upper blade, regardless of behavioral conditions. Thus, the survival of 4-week-old cells in both DG blades was not affected by the behavioral task, the genotype, or the m266 treatment in Tg mice.

We further evaluated the percentage of four-week-old BrdU⁺ cells expressing DCX protein (Figure 4B). This neuronal marker identifies new neuronal progenitors (around 8–15 days post birth) and new immature neurons (around 15–30 days post birth) [45, 46]. Thus, in our study, co-labeling of BrdU⁺/DCX⁺ was used to identify four-week-old immature neurons. There was no global effect of behavioral conditions on the percentage of new immature neurons in the GCL (pooled HC mice: 22.8 ± 2.3% and pooled tested mice: 23.8 ± 2.7%; MW test: U = 110, Z = –0.4, P = 0.7, Figure 4B and Table S3). However, in HC as in tested groups, the vehicle-treated Tg mice tended to have more new immature neurons in the GCL than vehicle-treated NTg mice and m266-treated Tg mice (tested groups: KW test: H = 5.8, P = 0.056, MW: NTg vs vehicle-treated Tg mice: U = 4, Z = –1.7, P = 0.095, m266-treated Tg vs vehicle-treated Tg: U = 4, Z = 1.9, P = 0.055). We analyzed all three treated groups on pooled behavioral conditions. The population of immature neurons in the GCL was statistically larger in vehicle-treated Tg mice than in other groups (KW test: H = 6.3, P = 0.04, MW: vehicle-treated Tg vs vehicle-treated NTg: U = 16, Z = –2.3, P = 0.02, vehicle-treated Tg vs m266-treated Tg: U = 26, Z = 2, P = 0.05; Figure 4C). No differences were observed between vehicle-treated NTg and m266-treated mice (MW: NTg vs m266-treated Tg2576: U = 50, Z = 0.7, P = 0.5). Taken together, these results suggest that Tg mice displayed newer immature (DCX-positive) neurons than NTg mice, and that the m266 treatment tended to lower the density of new immature neurons to the same level as in NTg mice.

Interestingly, the upper blade displayed a significantly higher percentage of immature new neurons than the lower blade (pooled mice for lower blade: 20.4 ± 2% and upper blade: 25.3 ± 2.0%; W: Z = 3, P = 0.003, see Table S3). However, whatever the behavioral conditions, there only was a significant group effect in the lower blade (KW: upper, H = 4.8, P = 0.09 and lower, H = 7.8, P = 0.02, see Table S3). Indeed, vehicle-treated Tg mice exhibited significantly more new immature neurons in the lower blade compared to m266-treated Tg mice (MW lower blade: Z = 2.5, P = 0.01, MW upper blade: Z = 1.8, P = 0.07) and NTg mice (MW lower blade: Z = –2.1, P = 0.03, MW upper blade: Z = –1.8, P = 0.08). Thus, the m266 treatment of Tg mice lowered the percentage of immature neurons in the lower blade to the same level as in vehicle-treated NTg. These results suggest that vehicle-treated Tg mice displayed more new immature neurons than the other groups and that the m266 treatment restored a percentage of immature neurons similar to that of NTg mice.

Finally, statistical analysis of new neuron activation was not performed due to the very low percentage of new neurons expressing Egr-1 (pooled all groups: % of DCX⁺ neurons: 1.4 ± 0.4%, see Table S4).

After the four-week m266 passive immunization treatment against Aβ, we analyzed the total levels of human Aβ40 and Aβ42 in the CA3/DG region of vehicle-treated and m266-treated Tg female mice (n = 4–5/group, exclusion of one m266-treated Tg which was considered as an outlier), as well as the total levels of murine Aβ40 and Aβ42 in vehicle-treated young and aged B6 male mice (n = 4–6/group), and m266-treated aged B6 male mice (n = 4). In the Tg experiment, the total level of Aβ peptides (Aβ40 + Aβ42) in the CA3/DG region was slightly but not significantly lower in m266-treated Tg mice compared to vehicle-treated mice (m266 Tg, Aβ: 2,127 ± 133 pg/mg of tissue; vehicle Tg, Aβ: 2,736 ± 271 pg/mg of tissue; MW: U = 7, Z = –1.4, P = 0.2, Figure 5A and Table S5). As expected, the level of Aβ40 in young Tg2576 mice was significantly higher than that of Aβ42 whatever the treatment (W: vehicle-treated mice: Z = 2.2, P = 0.03; m266-treated mice: Z = 2.0, P = 0.04, Table S5). The ratio of Aβ42 to Aβ40 remained stable among Tg groups (Figure 5A, Table S5). In the aging experiment, the total level of Aβ peptides remained stable among groups (young B6, Aβ: 486.3 ± 76.8 pg/mg of tissue; m266 old B6, Aβ: 456.7 ± 82.65 pg/mg of tissue; vehicle old B6, Aβ: 380.2 ± 74.5 pg/mg of tissue, Figure 5B and Table S5), with a higher level of Aβ40 compared to Aβ42 (W, vehicle-treated young B6: Z = 2.2, P = 0.03, vehicle-treated aged B6: Z = 2.2, P = 0.03) except in m266-treated aged B6 (Z = 1.8, P = 0.07, see Table S5). The ratio of Aβ42 to Aβ40 in aged mice was not significantly different from the one in young mice (vehicle-treated young B6: 0.3 ± 0.04, vehicle-treated aged B6: 0.4 ± 0.02, m266-treated aged B6: 0.4 ± 0.03, Figure 5B and Table S5). As expected, the total level of human Aβ peptides in Tg mice (mean of vehicle- and treated-mice: 2,459 ± 180.3 pg/mg of tissue) was much higher than that of murine Aβ peptides in young (486.3 ± 76.8 pg/mg of tissue) and aged B6 mice (mean of vehicle- and treated-mice: 410.8 ± 54.0 pg/mg of tissue). Thus, the m266 treatment only had marginal effect on Aβ levels in CA3/DG of 5-month-old Tg mice, and B6 mice.