Detecting spatial memory deficits beyond blindness in tg2576 Alzheimer mice

Neurobiology of Aging 34 (2013) 716 –730 www.elsevier.com/locate/neuaging

Detecting spatial memory deficits beyond blindness in tg2576 Alzheimer mice Nour Yassinea, Anelise Lazarisb, Cornelia Dorner-Ciossekb, Olivier Desprésa, Laurence Meyerc, Michel Maitrec, Ayikoe Guy Mensah-Nyaganc, Jean-Christophe Cassela, Chantal Mathisa,* a

Laboratoire d’Imagerie et de Neurosciences Cognitives, UMR 7237 CNRS, Université de Strasbourg, IFR 37, GDR CNRS 2905, Strasbourg, France b Boehringer-Ingelheim Pharma GmbH and Co., KG, Department of CNS Diseases Research, Biberach an der Riss, Germany c Faculté de Médecine, Université de Strasbourg, Strasbourg, France Received 19 March 2012; received in revised form 30 May 2012; accepted 21 June 2012

Abstract The retinal degeneration Pde6brd1 (rd) mutation can be a major pitfall in behavioral studies using tg2576 mice bred on a B6:SJL genetic background, 1 of the most widely used models of Alzheimer’s disease. After a pilot study in wild type mice, performance of 8- and 16-month-old tg2576 mice were assessed in several behavioral tasks with the challenge of selecting 1 or more task(s) showing robust memory deficits on this genetic background. Water maze acquisition was impossible in rd homozygotes, whereas Y-maze alternation, object recognition, and olfactory discrimination were unaffected by both the transgene and the rd mutation. Spatial memory retention of 8- and 16-month-old tg2576 mice, however, was dramatically affected independently of the rd mutation when mice had to recognize a spatial configuration of objects or to perform the Barnes maze. Thus, the latter tasks appear extremely useful to evaluate spatial memory deficits and to test cognitive therapies in tg2576 mice and other mouse models bred on a background susceptible to visual impairment. © 2013 Elsevier Inc. All rights reserved. Keywords: Alzheimer disease; APP transgenic mice; Object location; Recognition memory; Barnes maze; Spatial memory; Vision; Blind mice; Retinal degeneration; Genetic background; B6SJL mice

1. Introduction Alzheimer’s disease (AD) is associated with a progressive impairment of several aspects of cognition, including spatial disorientation and episodic memory failures, as well as recognition and olfactory memory deficits (Buck et al., 1997; Kalová et al., 2005; Mesholam et al., 1998). The neuropathology of AD is characterized by amyloid plaques, which are extracellular deposits of ␤-amyloid peptide (A␤), and neurofibrillary tangles. Less than 10% of the patients bear the familial form of AD caused by mutations on am* Corresponding author at: Laboratoire d’Imagerie et de Neurosciences Cognitives, UMR 7237 CNRS, Université de Strasbourg, 12 rue Goethe, F-67000 Strasbourg, France. Tel.: ⫹33 368 85 18 76; fax: ⫹33 368 85 19 58. E-mail address: [email protected] (C. Mathis). 0197-4580/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.neurobiolaging.2012.06.016

yloid protein precursor (APP), PS1, or PS2 genes, whereas the remaining majority of the patients are affected by the sporadic form of AD which depends on complex interactions between environmental and genetic risk factors. As most familial AD mutations, the “Swedish” double mutation of the APP gene leads to an abnormal cleavage of APP resulting in increased production of A␤. One of the first APP transgenic models of AD is the tg2576 mouse which overexpresses this mutation (Hsiao et al., 1996). It is now widely used in fundamental AD research and therapeutic development. Tg2576 mice show a rapid increase in cerebral A␤ starting at an age of 6 –7 months; usually, amyloid plaques deposition are found in the hippocampus and the cortex 3 to 6 months later (Kawarabayashi et al., 2001). Tg2576 mice also develop age-dependent spatial reference memory deficits in the Morris water maze, although the age

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

of onset may vary from 3 to 12 months depending on the study (Hsiao et al., 1996; King et al., 1999). Memory deficits were confirmed in a wide variety of spatial tasks including Barnes-maze and spatial alternation tasks, as well as contextual and social memory tasks (Chapman et al., 1999; Deacon et al., 2009; Jacobsen et al., 2006; Middei et al., 2004; Ognibene et al., 2005). Deficits in spatial novelty tasks and preservation of object recognition performance, even at old ages, have been consistently reported (Good et al., 2007; Hale and Good, 2005; Ognibene et al., 2005). The tg2576 model, however, has also generated inconsistent results in several of the aforementioned spatial tasks (Deacon et al., 2008; Holcomb et al., 1999; King et al., 1999). In addition, some of these studies reported high within-group variability in both tg2576 and wild type (wt) littermates, thereby limiting the strength of the behavioral findings and even questioning the validity of this model (Deacon et al., 2008). The presence of the recessive retinal degeneration allele Pde6brd1 (rd) in the B6:SJL genetic background of most available wt/tg2576 cohorts might, at least partly, explain the lack of consistency and behavioral variability. Most rods degenerate within 7 weeks of age in homozygous rd mice, whereas degeneration of the cones is much slower (Bowes et al., 1990; Carter-Dawson et al., 1978). Thus, rd carrying inbred strains, such as the SJL strain, show a dramatic visual deficit by 7–9 weeks of age, which deeply impairs their performance in tasks requiring visual guidance (Brown and Wong, 2007; Clapcote et al., 2005). The challenge of the current work was to identify some tasks demonstrating robust memory deficits that are not affected by the rd mutation in tg2576 mice bred on a B6:SJL genetic background. In a first pilot experiment, we tested a cohort of 11-, 17-, and 21-month-old wt mice in a battery of tasks presumably impaired in tg2576 mice (watermaze, Barnes-maze, Y-maze, and spatial and object recognition tasks). Our aim was to select tasks demonstrating good performance and low variability in a B6:SJL background population made up of a majority of homozygous rd carriers. In a second study, these tasks were used to assess memory performance in a vast cohort of wt versus tg2576 mice of 2 ages (8 and 16 months). After genotyping of all mice, the data were specifically reanalyzed to determine the impact of the rd mutation on the performance of tg2576 and wt mice. 2. Methods 2.1. Animals We used tg2576 female mice which express the transgene coding for the 695-amino acid human APP isoform containing the double Swedish mutation (K670N, M671L; Hsiao et al., 1996) and their wt female littermates. Our breeding colony (Bayer Pharma AG, Wuppertal, Germany) is maintained on a B6:SJL hybrid genetic background by mating hemizygous males with B6SJLF1 females. Our co-

717

horts were composed of sisters from a limited number of litters (3– 4 per age for the wt cohort; 14 –15 per age for the transgenic (tg)/wt cohort). We chose females because they can be maintained in social groups until 2 weeks before the start of the testing period, as opposed to males which are quite aggressive under the same housing conditions. The first cohort consisted of wt mice only, which were 11, 17, and 21 months old at the beginning of the study (n ⫽ 10 –13 per group). The second cohort consisted of tg2576 and wt littermates which were 8 and 16 months old at the beginning of the study (n ⫽ 11–14 per group). We also report here the results of a pilot study on 5-month-old tg2576 female mice (n ⫽ 14) tested in the object exploration paradigm, but not genotyped for rd. Mice were housed individually in transparent cages (22 ⫻ 20 ⫻ 14 cm) under controlled temperature (22 ⫾ 1 °C) and a 12/12 hours light/dark cycle (lights on at 7:00 AM). Food and water was available ad libitum unless indicated otherwise. Paper towels were provided for nesting and a few food pellets were dropped inside the cage (in addition to those present in the cage cover). According to our experience, these conditions contribute to successful reduction of mortality. All experimental procedures were conducted in conformity with the institutional guidelines (council directive 87/848, October 19, 1987, Ministère de l’agriculture et de la Forêt, Service Vétérinaire de la Santé et de la Protection Animale; National Institutes of Health publication, 86 –23, revised 1985). Official permission references for animal experimentation were 67–292 for C.M. and 67–215 for J.-C.C.; N.Y. was under their responsibility. Reference of official permission for holding genetically modified organisms (tg2576) was 5016 CA-II. 2.2. Protocol of the wt cohort pilot study 2.2.1. Locomotor activity test Two weeks after their arrival at the laboratory, the wt cohort was tested in an actimetric device to evaluate general activity during habituation to a novel environment. The mice were placed individually in large transparent cages (42 ⫻ 26 ⫻ 15 cm3) adapted to the testing device. Two infrared light beams were targeted on 2 photocells, 2.5 cm above the cage floor level and 28 cm apart. The number of longitudinal cage crossings was recorded from 11:00 AM to 5:00 PM. 2.2.2. Spatial novelty and object recognition tasks Three days after the activity test, the wt cohort was tested in an object exploration paradigm, which is based on the spontaneous tendency of mice to explore preferentially new objects (object recognition) or objects moved to a new location (spatial novelty). Object recognition and spatial novelty tasks were performed in a Plexiglas open field (52 ⫻ 52 cm) with black walls (40 cm high) and a white floor divided into 25 equal squares by black lines. A striped card was fixed against a wall. The device was illuminated by an indirect halogen light (open field center: 40 lux) and a radio played a background noise 2.30 m from the device

718

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

(open field center: 45 ⫾ 5 dB). We used 8 different objects (size: 10 –17 cm) made of metal, glass, or plastic. They differed in shape and color and each object was available in triplicate (1 for each trial when required). The mice were first submitted to a 10-minute habituation phase for 2 days with 2 different and new objects each day. The next day, the mice were exposed 10 minutes to the spatial configuration of 3 new objects positioned in 3 corners of the open field respectively (approximately 10 cm from the walls). Then, the mice were placed for 3– 4 minutes in their home cage before being subjected to a 10-minute retention trial with a new spatial configuration resulting from the move of 1 object to the opposite, free corner of the open field. We chose to use 3 objects for essentially 2 reasons. First, we wanted to preserve the same overall object configuration after displacement of 1 object, which, using 3 objects, was possible by displacing the object from 1 corner to the opposite corner of the open field. Second, with 3 objects the task was slightly more difficult than with 2, but less than with another usual protocol based on 4, 5, and sometimes even more objects (e.g., Oliveira et al., 2010). The fourth day, the mice received a 10-minute object recognition trial with the same configuration of familiar objects, except that a new object replaced a familiar one. On each trial, mice were placed in the center of the open field. Between each trial, walls, floor, and objects were wiped with 70% ethanol. Exploration was defined as the nose pointing toward the object within 1 cm. Memory performances were evaluated using a discrimination index defined as the time spent exploring the displaced object or the novel object divided by the total time spent exploring all 3 objects in the spatial novelty and the object recognition task, respectively. A preliminary experiment ascertained that mice showed no spontaneous preference for the object to be displaced or for the novel object in comparison with the 2 other objects. 2.2.3. Barnes maze task One week later, the wt cohort was tested in a Barnes maze task. This task became popular to assess spatial memory in mice because of their ability to find an escape through small holes and their poor adaptation to tasks based on water escape responses (Whishaw and Tomie, 1996). The mice were trained to find a target hole among 12 holes equally distributed along the edge of a 1 m circular board. The target hole was connected to a tube leading to the home cage. For a given mouse, the target hole was at a fixed spatial location in the testing room which provided various extra maze visual, auditory, and olfactory cues such as the experimenter or the other mice in 2 racks with 20 –32 cages, furniture, pictures, overhead fluorescent lighting (800 lux at the maze center), and background noise from a radio playing 1.90 m from the maze center (45 ⫾ 5 dB at the maze center). The opaque polyvinyl chloride cylinder used to confine the mice to the center of the maze was lifted at the start of each trial. Between trials, the maze was wiped with

a 70° alcohol solution and rotated in order to disable an odor-based strategy. After a 3-day habituation to the tube, mice were tested on several testing phases designed to evaluate successively their performances on different aspects of spatial learning and memory processes. On Day 1, they received an extensive training of 6 massed trials to test short-term acquisition ability. On Day 2, there was a 120-second probe trial with all holes closed. Preferential exploration of the target hole during the probe trial is considered to reflect good memory retention. Swiss mice trained with this procedure showed a preferential exploration of the target hole (unpublished data). After the probe trial, the mice received 3 additional acquisition trials to evaluate relearning capacities after a 24-hour delay. From Days 3 to 6, the target was moved to the opposite hole of the board, and the mice were trained on a “reversal” task at the rate of 3 trials/day. A second probe trial tested remote spatial memory 11 days after the end of the reversal phase (day 17). On all trials, the latency to enter the target hole and the order of the hole visits (nose pokes) were recorded. For each training trial (acquisition and reversal), we counted the number hole errors defined as visits to nontarget holes per trial. Because the total number of visits may differ among groups on probe trials, retention performances were evaluated in terms of visiting index (number of visits to the hole/total number of visits) of the target hole versus those of 3 equidistant holes (the opposite hole and those located ⫾ 90° apart from the target). In addition, each acquisition trial was classified into 1 of 3 categories of search strategy (adapted from Harrison et al., 2006) reflecting the use of either a direct spatial strategy (defined as direct visit to the target, sometimes preceded by at most 1 adjacent hole visit), a serial strategy (minimum of 2 adjacent hole visits in a serial manner before reaching the target) or a mixed strategy (remaining trials). The data were subsequently analyzed in terms of percentage of trials with a direct spatial strategy or with a serial strategy. 2.2.4. Morris water maze task One week after the Barnes maze task, the wt cohort was tested in the Morris water maze task as described in Moreau et al. (2008). The pool (diameter: 140 cm) filled with 20 °C opaque water was located in a room with many extramaze visual and auditory cues such as the experimenter or the other mice on 2 racks with cages, a computer, a water heater, pictures, an overhead fluorescent light (180 lux at pool center), and a radio playing as a background noise at 2.50 m from the pool center (45 ⫾ 5 dB). The pool surface was virtually divided in 4 quadrants with the 4 starting points defined as north (N), east (E), south (S), and west (W) at their boundaries. Mice were familiarized to the device with a 1-minute swim in 3.5-cm deep water on the first day and a 2-minute free swim on the next day. Training in the reference memory procedure began the following day at the rate of 4 trials per day for 5 consecutive days. On each trial, the mouse had to find a hidden platform (diameter: 10 cm)

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

located in the SW quadrant within 60 seconds. The pseudorandom order of the 4 possible starting points was changed from day to day. Latency and distance to reach the platform were recorded using a video tracking system (Noldus, Wageningen, Netherlands). On the following day, a 60-second probe trial was performed to evaluate the strength and precision of spatial memory by comparing the time spent in the target (SW) quadrant to that spent in the 3 other quadrants. Four days later, the mice were trained in a 4-trial visible platform procedure to check for possible motivation or sensory-motor biases that could have affected training performance. For each of the 4 trials, the position of the visible platform and the starting point were changed in a pseudorandom fashion. 2.2.5. Y-maze alternation task One week later, the wt mice were tested in the Y-maze spontaneous alternation task. The maze consisted of a plexiglas apparatus with 3 identical arms (13 cm ⫻ 4.5 ⫻ 5.5 cm each) in a Y shape. Mice were placed in the center of the maze and left free to explore for 5 minutes. The number of arms visited and the sequence of these visits were recorded. The percentage of spontaneous alternations was calculated by dividing the total number of alternations by the total number of entries minus 2. 2.3. Adaptation of the protocol to the wt/tg2576 cohort study The battery of tasks was modified for the tg2576 and wt cohort according to the outcomes of the wt cohort study. The sequence of behavioral tasks was improved by inserting the Y-maze alternation task between the actography session (3 days after) and the object exploration paradigm (1 week before) in order to follow more closely the principle of testing from the least to the most demanding one. A modified version of the Barnes maze (described below) began 1 week after the object exploration paradigm and was followed, 4 weeks later, by an olfactory discrimination task. 2.3.1. Barnes maze task adapted to tg2576 mice For the tg2576 and wt cohort, the protocol of the Barnes maze task was slightly modified compared with that of the wt cohort. A 48-hour delay (instead of 24 hours) was inserted between the 6 massed acquisition trials (Day 1) and the 3 additional acquisition trials (Day 3) in order to render relearning more challenging. The first probe trial was shifted to the next day (Day 4) in order to improve the retention performances of wt mice. Finally, the second probe trial was scheduled only 7 days after the reversal phase (Day 14) in order to further improve globally the performance of wt mice on this remote memory retention test and, thereby, possibly allowing a better discrimination between age and genotype subgroups. 2.3.2. Olfactory discrimination task Four weeks after the Barnes maze task, tg2576 and wt mice were trained in an olfactory discrimination task

719

adapted from Schellinck et al. (2001). The mice learned to associate 1 odor with the presence of food pellets (durum wheat noodle; 17.3 ⫾ 0.3 mg each) and another odor with the absence of food in a glass pot. Mice were progressively food deprived over 5 days in order to reach 90%– 85% of their ad lib body weight on the first day of training. The mice were trained daily for 3 days with 2 paired trials (with food) and 2 unpaired trials (no food). On paired trials, the mice had to dig in an odor pot filled with bedding to retrieve 6 food pellets lying on a plastic Petri dish (diameter: 5 cm; height: 2 cm) placed in the pot bottom. On unpaired trials, there were no food pellets and the odor was different. The Petri dish with 8 predrilled holes contained an odor stimulus (50 ␮L) on a filter paper. Odor stimuli were phenyl acetate or linalool (Aldrich Chemicals, St. Quentin Fallavier, France) diluted to a concentration of 15% in propylene glycol. On each 5-minute trial, the mice were introduced in a plastic cage (30 ⫻ 20 ⫻ 11 cm) containing an odor pot and consumption of all food pellets was verified on paired trials. The order of paired and unpaired trials was changed daily. Six days later, memory retention was evaluated in a transparent polyvinyl chloride box (45 ⫻ 30 ⫻ 20 cm) divided in 3 equal square units (left, middle, right) interconnected by small doors. During the 2-minute probe trial, the mice were placed in the middle unit with the phenyl acetate pot in the left unit and the linalool pot in the right unit (none contained food pellets). Time spent digging was recorded for each odor pot. Digging was strictly defined as moving the bedding with the paws or the nose. A higher time spent digging in the odor-paired pot versus the odorunpaired pot was considered to reflect good olfactory memory. 2.4. Genotyping The mice of the wt cohort and those of the wt/tg2576 cohort were genotyped for the rd mutation following a polymerase chain reaction (PCR) method adapted from Giménez and Montoliu (2001). Three independent oligonucleotides were used to distinguish between the rd mutant and the wt alleles: RD3 (28-mer, 5=-TGACAATTACTCCTTTTCCCTCAGTCTG-3=, accession number L02109, nucleotides 84 –111), RD4 (28-mer, 5=-GTAAACAGCAAGAGGCTTTATTGGGAAC-3=, accession number L02109, nucleotides 644 – 617), and RD6 (28-mer, 5=- TACCCACCCTTCCTA ATTTTTCTCACGC-3=, accession number L02110, nucleotides 2539 –2512). The RD3/RD6 pair of primers amplified the 0.40 kb PCR product from the wt allele, whereas no PCR product was expected from the rd mutant allele. The RD3/RD4 pair of primers amplified only the 0.55 kb PCR product from the mutant allele. Thus, the presence of a unique 0.40 kb band or a unique 0.55 kb band characterized a wt homozygous mouse or rd homozygous mouse, respectively. Both PCR bands were present in a heterozygous mouse. Briefly, total genomic DNA derived from tail biopsy was used in the following PCR protocol:

720

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

denaturing phase at 94 °C for 30 seconds, annealing phase at 62 °C for 1 minute, synthesis phase at 72 °C for 2 minutes over 35 cycles, and final synthesis phase in same conditions for 10 minutes. PCR reaction was resolved in 1% agarose horizontal submarine gel electrophoresis. We verified that band intensity was similar for RD3/RD4 and RD3/RD6 primer combination for quality control of the PCR assay. 2.5. Statistical analyzes Most data were analyzed by means of analysis of variance with 1 or more factors (age, genotype, and/or rd; Statistica 8.0, StatSoft, Inc., Tulsa, OK, USA) and, when appropriate, with repeated measure factors (hour, day, trial, or odor). Post hoc analyses were performed with the Newman–Keuls (NK) test. Due to the disproportional repartition of unaffected (rd/⫹, ⫹/⫹) mice among genotype/age subgroups, genotype ⫻ age and genotype ⫻ rd interactions were analyzed separately. The effect of the rd factor was tested on groups pooled for age only if there was no significant age effect and no interaction with this factor. The time spent in each quadrant during the water maze probe trial was compared with a 15-second chance value (no quadrant preference) by means of a 1 sample t test. The t test was also used to compare each index performance in the object exploration task to a 0.33 chance ratio which corresponds to 1 third of the total exploration time spent on the displaced object (i.e., no preference for this object).

3.2. Water maze task The performances in the water maze tasks are shown first because of the dramatic effect of the rd/rd genotype. Therefore, these tasks were used only in the wt cohort. Analyses of the distance and latency to reach the platform led to similar results, thus only the distance data are shown here. The wt mice improved modestly their performance in the hidden platform task (from 836 ⫾ 55 cm on day 1 to 600 ⫾ 59 cm on Day 5; day effect: F(4,80) ⫽ 5.35; p ⫽ 0.001, Supplementary Data 1A). There was no significant effect of age and no age ⫻ day interaction. During the probe trial on Day 6, no group spent significantly more than 15 seconds in the goal quadrant (Supplementary Data 1B). When data were reanalyzed in terms of rd status, it became evident that rd/rd mice (n ⫽ 19) were deeply impaired in this task. Their level of performance during acquisition clearly differed from that of the few (n ⫽ 4) unaffected mice (rd effect: F(1,21) ⫽ 7.22, p ⫽ 0.01, Fig. 1A). Moreover, the rd/rd group showed no bias for the goal quadrant contrarily to unaffected group (Fig. 1B). In the visible platform procedure (Supplementary Data 1C and D), the distance performance of the 3 groups remained at a low level (501 ⫾ 68 cm). There was no significant effect of age or trial, or interaction between these factors. The distance swum by the rd/rd mice (537 ⫾ 75 cm) tended to be greater than that of unaffected mice (270 ⫾ 64 cm; rd effect: F(1,21) ⫽ 3.74; p ⫽ 0.06). 3.3. Locomotor activity

3. Results 3.1. Group size and rd genotyping of the cohorts As shown in Table 1, the number of mice could vary from 1 task to another because some mice were unable to perform the task (immobility in the Y-maze, no object exploration, jumping off the Barnes maze, 1 or no hole visit during the probe, swimming difficulties in the water maze) or because the data obtained in some mice were affected by stereotypies (as detailed in the Results section below). Table 1 also shows the number of visually unaffected mice (rd/⫹, ⫹/⫹) for each group.

In both the wt and the wt/tg2576 cohorts, some animals exhibited strong stereotypies (mostly circling, salto jumps, or wall climbing) in their home cage and during the actimetry session. Circling was associated with very high activity scores resulting in high variability of the data. Thus, we eliminated circling animals exhibiting more than 1000 transitions per hour at 1 time point (wt cohort: 3/13 11-monthold wt, 1/10 17-month-old wt and 1/10 21-month-old wt; wt/tg2576 cohort: 1/13 8-month-old wt, 2/15 16-month-old wt, 3/12 8-month-old tg2576, and 0/13 16-month-old tg2576).

Table 1 Group size in the wt cohort and the wt/tg2576 cohort for each task Genotype

Age (mo)

Water maze

Actography

Barnes (acquisition)

Barnes (Probe 1)

Barnes (Probe 2)

Spatial novelty/object recognition

Y-maze

Olfactory task

wt cohort

wt

wt/tg cohort

wt

11 17 21 8 16 8 16

9 (1) 7 (2) 7 (1) X X X X

10 (1) 10 (3) 9 (1) 12 (4) 13 (7) 12 (3) 15 (1)

9 (2) 9 (3) 10 (1) 13 (5) 13 (7) 10 (4) 15 (1)

9 (2) 9 (3) 8 (1) 13 (5) 13 (7) 10 (4) 13 (0)

9 (2) 9 (3) 10 (1) 13 (5) 13 (7) 10 (4) 13 (1)

12 (2)/11 (2) 10 (3) 10 (1)/9 (1) 13 (5) 14 (6)/15 (7) 14 (4) 16 (1)

12 (2) 10 (3) 9 (1) 12 (4) 15 (7) 14 (5) 12 (1)

x x x 11 (4) 13 (6) 12 (4) 12 (1)

tg

For each group, the number of unaffected (rd/⫹ and ⫹/⫹) mice is specified in parentheses. Key: tg, transgenic; wt, wild type; x, not tested.

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

721

Fig. 1. Water maze performances of the wild type (wt) cohort in the hidden platform task. The rd/rd mice showed no improvement of performance during training (A) and no bias for the goal quadrant during the probe test as opposed to unaffected mice (B). Group sizes are indicated in brackets and data are shown as the mean distance swum to the platform and as mean time spent in northeast (NE), northwest (NW), and southeast (SE) quadrants, and the southwest (SW) goal quadrant, respectively. * p ⬍ 0.05, significantly different from 15-second chance value (dotted line).

3.3.1. Wt cohort The mice showed a decrease in activity over the 6-hour habituation phase (hour effect: F(5,125) ⫽ 12.4; p ⬍ 0.001; no effect of age or interaction with this factor; data not shown). On age-pooled data, rd/rd mice were slightly, but not significantly, more active than unaffected mice (data not shown). 3.3.2. Wt/tg2576 cohort Again, the locomotor activity decreased over the 6-hour habituation period (hour effect: F(5,240) ⫽ 10.3; p ⬍ 0.001; Supplementary Data 2A). There was no effect of age, genotype or any interaction between these factors. Therefore, we analyzed the rd mutation effect on both tg2576 and wt mice (Supplementary Data 2B). Again, the rd/rd mice tended to be more active than unaffected mice, but this effect was not significant (rd effect: F(1,48) ⫽ 2.54; p ⫽ 0.12; no genotype effect or interaction between hour, rd, and genotype factors). 3.4. Y-maze spontaneous alternation task 3.4.1. Wt cohort There was a clear effect of age in this task (age effect: F(2,28) ⫽ 4.45; p ⫽ 0.02): the percentage of alternation in the 21-month group (51.1 ⫾ 3.8%) was significantly lower than that in the 11-month group (66.2 ⫾ 3.7%; p ⬍ 0.05), and it did not differ from that of the 17-month group (59.8 ⫾ 3.0%). The total number of arm entries did not differ among the 3 groups. 3.4.2. Wt/tg2576 cohort There was no effect of age, genotype (wt: 63.8 ⫾ 4.4%; tg2576: 58.2 ⫾ 2.9%), or any interaction between these factors on alternation performances. The total number of arm entries did not differ significantly between tg2576 (31.8 ⫾ 3.5) and wt littermates (26.2 ⫾ 1.8). The rd status did not affect the alternation performance (unaffected:

63.6 ⫾ 3.8%, n ⫽ 6; rd/rd: 60.1 ⫾ 3.5%, n ⫽ 25) or the number of arm entries. 3.5. Spatial novelty task 3.5.1. Wt cohort During the acquisition trial, global exploration of the 3 objects was similar among ages and the mice explored similarly the objects to be displaced or nondisplaced. On the retention trial, there was no effect of age on the spatial novelty index (age effect: F(2,29) ⫽ 1.04; p ⫽ 0.37). As illustrated in Fig. 2A, all groups showed a clear preference for the displaced object. As shown in Fig. 2B, rd/rd mice (n ⫽ 26) and unaffected (rd/⫹, ⫹/⫹) mice (n ⫽ 6) showed a similar preference for the displaced object (rd effect: F(1,30) ⫽ 1.38; p ⫽ 0.54; t test in Fig. 2B). 3.5.2. Wt/tg2576 cohort During the acquisition trial, global exploration of the objects was similar among groups and there was no preference for the object to be displaced. On the retention trial, wt and tg mice differed in terms of spatial novelty index (F(1,53) ⫽ 25.70; p ⬍ 0.001; no age effect or interaction with age; each tg group differed from the same age wt group: p ⬍ 0.002, NK test). A preference for the displaced object was only expressed by wt groups (see t test in Fig. 2C). Surprisingly, rd status did not affect spatial novelty performance in wt mice and in tg2576 mice (genotype effect: F(1,53) ⫽ 15.65; p ⬍ 0.001; no effect or interaction with rd: F(1,53) ⬍ 0.03; p ⬎ 0.87; each tg group differed from the same rd genotype wt group: p ⬍ 0.03, NK test; see t test for ability to detect spatial novelty in Fig. 2D). Note that the unaffected tg group included only 4 mice, but their mean performances were very close to those of the 26 rd/rd tg mice.

722

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

Fig. 2. Performance of wild type (wt) and tg2576 mice in the spatial novelty task. The cohort showed a preference for the displaced object (A) whatever the age and (B) independently of the rd status. (C) In the wt/tg2576 cohort, transgenic mice showed a deficit in spatial novelty performance at 8 and 16 months of age. (D) In both wt and tg2576 mice, the level of performance remained unaffected by the rd status. Data are shown as spatial novelty index (time spent on the displaced object divided by the time spent on all objects; ⫾ standard error of the mean) during the 10-minute retention trial. Group sizes are indicated in brackets. * p ⬍ 0.05, ** p ⬍ 0.01, *** p ⬍ 0.001, significantly different from the 0.33 chance value (no preference for the displaced object, t test).

3.5.3. Tg2576 mice at 5 month of age Fourteen naive tg2576 mice were tested in the spatial novelty task at a young age as part of a pilot study. They showed a significant preference for the displaced object (spatial novelty index: 0.52 ⫾ 0.047; t(13) ⫽ 4.08; p ⫽ 0.0013). 3.6. Object recognition task 3.6.1. Wt cohort During the object recognition trial, 2 mice did not explore the objects (1 of the 11-month group and 1 of the 21-month group). The 3 age groups showed a similar preference for the novel object (age effect: F(2,27) ⫽ 1.48; p ⫽ 0.25; t test in Fig. 3A). The rd mutation did not affect the performance of wt mice (n for rd/rd ⫽ 6; n for unaffected ⫽ 24; rd effect: F(2,27) ⫽ 0.008; p ⫽ 0.92, t test in Fig. 3B). 3.6.2. Wt/tg2576 cohort The object recognition index was similar among all groups and they all showed a preference for the novel object (no significant effect of age or genotype, no interaction between these factors; see t test in Fig. 3C). Interestingly, the rd status of wt and tg2576 mice did not affect their

object recognition performance (rd effect: F(1,54) ⫽ 0.11; p ⫽ 0.74; genotype ⫻ rd ⫽ F(1,54) ⫽ 2.41; p ⫽ 0.13, see t test in Fig. 3D). Note that the preference of the unaffected tg group for the novel object did not reach significance: it must be noted that this group included only 5 mice. 3.6.3. Tg2576 mice at 5 months of age The 14 young tg2576 mice showed a significant preference for the novel object (object recognition index: 0.48 ⫾ 0.03; t(13) ⫽ 4.96; p ⬍ 0.001). 3.7. Barnes maze task in the wt cohort Analyses of the latency and the number of errors led to similar results in the wt cohort, thus only the number of errors is illustrated here (Fig. 4A). The mean number of errors improved significantly over the 6 trials of Day 1 (trial effect: F(5,125) ⫽ 20.4; p ⬍ 0.001) and between Day 1 and Day 2 (day effect: F(1,25) ⫽ 19.0; p ⬍ 0.001), with no age effect. Thus, all age groups showed good short-term acquisition and relearning performances. The performances were transiently affected by the new position of the target hole between Day 2 and Day 3 (reversal effect: F(1,25) ⫽ 12.4; p ⫽ 0.002, no effect of age), but their performances im-

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

723

Fig. 3. Performances of wild type (wt) and tg2576 mice in the object recognition task. Mice of the wt cohort showed a preference for the novel object (A) whatever their age and (B) independently of their rd status. (C) In the wt/tg2576 cohort, tg2576 and wt mice explored more the new object than the familiar objects at the ages of 8 and 16 months. (D) The preference for the novel object was not affected by the rd status in wt and tg2576 mice. Data are shown as object recognition index (time spent on the new object divided by the time spent on all objects; ⫾ standard error of the mean) during the 10-minute retention trial. Group sizes are indicated in brackets. * p ⬍ 0.05, ** p ⬍ 0.01, *** p ⬍ 0.001, significantly different from the 0.33 chance value (no preference for the new object, t test).

proved significantly over the 4 days of reversal training (day effect: F(3,75) ⫽ 7.34; p ⬍ 0.001). Surprisingly, the rd/rd status did not affect acquisition and reversal performances in the Barnes maze task (rd effect: F(1,26) ⱕ 2.01; p ⬎ 0.17, no day ⫻ rd interaction; see Fig. 4B). The reversal effect between Day 2 and 3 did not differ significantly between rd/rd mice and unaffected mice (day ⫻ rd interaction: F(1,26) ⫽ 0.50; p ⫽ 0.50). Wt mice had low retention performances during the first probe trial made just before training on Day 2. There was no significant tendency to visit preferentially the target hole (hole effect: F(3.69) ⫽ 1.96; p ⫽ 0.13, no effect of age; data not shown). Therefore, the first probe test of the wt/tg2576 cohort was scheduled 24 hours after completion of the initial acquisition to improve retention performances. On the second probe performed 11 days after the reversal phase (Fig. 4C), the wt cohort had good long-term retention performances as suggested by a higher visiting index for the target hole compared with the opposite hole and those located ⫾ 90° apart (hole effect: F(3,75) ⫽ 18.41; p ⬍ 0.001, no effect of age). Interestingly, the rd/rd mice performed as well as unaffected mice in the second probe test: both groups showed a clear preference for the target hole

(hole effect: F(3,78) ⫽ 13.0; p ⬍ 0.001; hole ⫻ rd interaction: F(3,78) ⫽ 0.07; p ⫽ 0.98; Fig. 4D). 3.8. Barnes maze task in the wt/tg2576 cohort In the wt/tg2576 cohort, one 8-month tg2576 and one 16-month wt were not tested because they jumped off the maze. Globally, short-term acquisition performances greatly improved over the 6 trials of Day 1 (day effect: errors F(5,235) ⫽ 7.03; p ⬍ 0.001, Fig. 5A, latencies: F(5,235) ⫽ 34.6; p ⬍ 0.001, Fig. 5C; no genotype or age effect or interaction with these factors). The performances improved between Day 1 and Day 3 (day effect: errors F(1,47) ⫽ 11.4; p ⫽ 0.001; latencies F(1,47) ⫽ 50.5; p ⬍ 0.001). In terms of errors, there was no genotype or age effect on relearning performances. In terms of latencies, relearning was affected by age, mainly because the old tg2576 mice did not improve their performances (age effect: F(1,47) ⫽ 7.84; p ⫽ 0.007; day ⫻ genotype: F(1,47) ⫽ 11.4; p ⫽ 0.001). The use of spatial strategy increased, except for the 16-month tg2576 mice (day effect: F(1,47) ⫽ 14.8; p ⬍ 0.001; genotype effect: F(1,47) ⫽ 11.1; p ⫽ 0.001; genotype ⫻ age: F(1,47) ⫽ 3.28; p ⫽ 0.07; Fig. 5E). During

724

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

Fig. 4. Performances of the wild type (wt) cohort in the Barnes maze task. There was no significant effect of age (A) or of the rd status (B) on the mean number of errors per trial (⫾ standard error of the mean) made during the acquisition and reversal phases. Eleven days after the reversal phase, the retention performances were not affected by age (C) or rd status (D) during the second probe test. These data are shown as the visiting index (⫾ standard error of the mean) on the target hole, the opposite hole, and those located ⫾ 90° from the target hole. Group size are indicated in brackets. * p ⬍ 0.05, significantly different from the target hole.

acquisition, the use of a serial strategy concerned 20.6% to 28.5% of the trials and remained similar among groups (no day, age, or genotype effect or interaction). The reversal between Days 3 and 4 affected globally latency performances (day effect: F(1,47) ⫽ 10.2; p ⫽ 0.002; no interaction with age or genotype factors), but not error performances. Thereafter, the mice globally improved their performances over the 4 days of reversal training (day effect: errors F(3,141) ⫽ 11.3; p ⬍ 0.001, latencies F(3,141) ⫽ 29.6; p ⬍ 0.001). In terms of errors, the tg2576 mice were less efficient than wt mice (genotype effect: F(1,47) ⫽ 12.3; p ⬍ 0.001; genotype ⫻ day: F(3,141) ⫽ 2.3; p ⫽ 0.07). In terms of latencies, 16-month tg2576 mice were deeply impaired (genotype effect: F(14.86) ⫽ 14.8; p ⬍ 0.001; age effect: F(1,47) ⫽ 17.8; p ⬍ 0.001; genotype ⫻ age: F(1,47) ⫽ 4.47; p ⫽ 0.04). The use of a spatial strategy increased over the 4 days of reversal (day effect: F(5,141) ⫽ 14.1; p ⬍ 0.001), but to a lesser extent in tg2576 mice than in wt mice (genotype effect: F(1,47) ⫽ 11.1; p ⫽ 0.001;

day ⫻ genotype: F(1,47) ⫽ 3.2; p ⫽ 0.07). Age had a global effect on the use of a spatial strategy (age effect: F(1,47) ⫽ 7.65; p ⫽ 0.008; day ⫻ age: F(3,141) ⫽ 2.54; p ⫽ 0.06). Tg2576 mice used a serial strategy more compared with wt mice (genotype effect: F(1,47) ⫽ 5.52; p ⫽ 0.023; wt: 20.2 ⫾ 2.2% and tg2576: 29.2 ⫾ 2.8% of the trials). Only wt mice reduced their use of serial strategy from 28.8 ⫾ 4.8% of the trials on Day 4 to 7.69 ⫾ 2.8% on Day 7 (day effect: F(3,47) ⫽ 3.13; p ⫽ 0.027; day ⫻ genotype: F(3,47) ⫽ 2.61; p ⫽ 0.05; no age effect or interaction with this factor). The effect of the rd mutation was statistically analyzed only in wt mice because the global effect of age was essentially due to the old tg2576 mice. Additional analyses confirmed that there was no effect of age in wt mice. As found in the previous cohort, the rd status did not affect significantly the number of errors and the latency of wt mice to enter the target hole (Fig. 5B and 5D, respectively) during short-term acquisition, relearning and reversal (rd effect:

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

725

Fig. 5. Training performances of the wild type (wt)/tg2576 cohort in the Barnes maze task. (A) Tg2576 mice made more errors than wt mice at both 8 and 16 months of age, and (B) this effect was not affected by the rd status. (C) The oldest tg2576 group took a longer time to enter the target hole throughout training. (D) A delayed latency was also observed in rd/rd tg2576 mice. (E) The oldest tg2576 mice showed a reduced use of spatial strategy compared with the other groups. (F) The rd status had no significant effect on the use of the spatial strategy in wt mice; the rd effect was not statistically analyzed in tg2576 mice due to the age effect. Results are shown as mean values (⫾ standard error of the mean) and group size are indicated in brackets. $ p ⬍ 0.01 significantly different from wt mice of the same age. * p ⬍ 0.05 significantly different from 8-month mice of same genotype. § p ⬍ 0.05 significantly different from wt of same rd status.

F ⬍ 1,88; p ⬎ 0.18; no interaction with day). Surprisingly, rd/rd wt mice used a spatial strategy as frequently as unaffected wt mice (rd effect: acquisition F(1,24) ⫽ 3.19; p ⫽ 0.09, reversal F(1,24) ⫽ 0.007; p ⫽ 0.93; no interaction with day). Visual examination of the learning curves of tg2576 mice suggests that their rd status had minimal effects on error performances, whereas it affected latencies and delayed the use of a spatial strategy in rd/rd tg2576 mice at the beginning of the reversal phase compared with unaffected tg2576 mice and wt mice (Fig. 5F).

On the first probe trial (24 hours after relearning), the data of two 16-month-old tg2576 mice were discarded because they visited no hole during the whole trial. The total number of visits was slightly lower in older animals (age effect: F(1,45) ⫽ 4.94; p ⫽ 0.031; 16-month mice: 13.6 ⫾ 1.5; 8-month mice: 19.1 ⫾ 2.1). There was a global preference for the target hole (hole effect: F(3,135) ⫽ 12.31; p ⬍ 0.001). However, wt mice visited preferentially the target hole, whereas tg2576 mice visited similarly all holes (hole ⫻ genotype: F(3,135) ⫽ 4.62; p ⫽ 0.004; no effect of

726

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

age or interaction with this factor; Fig. 6A). The rd mutation had no effect on the level of retention performance of wt mice and tg2576 mice (rd effect and hole ⫻ rd: F(3,135) ⬍ 0.76; p ⬍ 0.51; Fig. 6B).

On the second probe trial (7 days after the reversal phase), the data of two 16-month-old tg2576 mice were discarded as previously. The total number of visits was also lower in older animals (age effect: F(1,45) ⫽ 4.14; p ⫽ 0.047; 16-month mice: 18.8 ⫾ 2.6; 8-month mice: 26.1 ⫾ 2.6). There was a global preference for the target hole which was mainly due to the performances of the 8-month-old wt mice, whereas 16-month-old wt mice had a moderate bias for the target hole and tg2576 mice visited equally all holes (hole effect: F(3,135) ⫽ 6.29; p ⬍ 0.001; hole ⫻ age: F(3,135) ⫽ 4.73; p ⫽ 0.003; see Fig. 6C). 3.9. Olfactory discrimination task Only the wt/tg2576 cohort was tested in the olfactory discrimination task. This task replaced the water maze task in the behavioral battery given the poor results of rd/rd wt mice. Globally, the time spent digging in the reinforced odor (paired with food) pot was significantly higher than that spent digging in the unreinforced odor pot during the retention session (odor effect: F(1,44) ⫽ 114.96; p ⬍ 0.001; wt: paired 20.7 ⫾ 2.3 and unpaired 2.4 ⫾ 0.9 seconds; tg2576: paired 18.8 ⫾ 1.8 and unpaired 1.3 ⫾ 0.4 seconds). There was no significant effect of genotype, age, or interaction between these factors. As could be expected in this task relying essentially on olfactory discrimination ability, the rd mutation did not affect the performance (no effect of rd or interaction with rd; unaffected: paired 2.2 ⫾ 2.5 and unpaired: 1.1 ⫾ 0.4 seconds; rd/rd: paired 19.1 ⫾ 1.8 and unpaired 2.2 ⫾ 0.7 seconds). 4. Discussion

Fig. 6. Retention performances of wild type (wt)/tg2576 cohort in the Barnes maze probe trials. (A) In the first probe trial (24 hours after relearning), tg2576 mice visited similarly the 4 holes, whereas wt mice visited the target hole more than all other holes whatever their age. (B) In this probe trial, the rd status had no effect on the retention performance of wt and tg2576 mice. (C) In the second probe trial (7 days after the reversal phase), only 8-month-old wt mice showed a clear preference for the target hole. Results are shown as visiting index (⫾ standard error of the mean) on the target hole, the opposite hole, and those located ⫾ 90° from the target hole. Group size are indicated in brackets. $ p ⬍ 0.01 significantly different from wt mice of the same age. * p ⬍ 0.05, significantly different from the target hole.

The tg2576 mice are widely used to model early stages of Alzheimer’s disease. Deficits in spatial navigation tasks (e.g., water-maze, Barnes maze) and other spatially based tasks (Y-maze alternation, spatial novelty) have been described at ages when amyloid deposition is developing (8 months and older), but sometimes also at much younger ages (3–7 months) as reviewed by Stewart et al. (2011). The high variability in the age of onset of these spatial memory deficits might, at least partly, be related to the occurrence of deep visual impairments in a variable proportion of individuals (homozygote for the rd mutation) when the cohort is bred on the B6:SJL genetic background. In the present study, the disproportional representation of rd/rd mice in regard to the 25% expected from Mendel’s law was related to the use of a limited number of litters (most probably fathered by a majority of rd/rd tg2576 males). This provided the unique opportunity to study the behavioral outcome of rd/rd mice in our battery of tasks. The performance of tg2576 mice were deeply impaired in a spatial novelty detection task, as well as in the reversal learning and the probe tests of a Barnes maze task, and these deficits were present in both rd/rd mice and sighted (rd/⫹, ⫹/⫹) mice. We also show that blind mice are able to perform success-

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

fully object recognition and spatial novelty detection tasks, as well as a spatial navigation task in the Barnes maze, although these tasks are often considered to depend essentially on the use of visual information. 4.1. Spatial novelty impairment is not influenced by the rd mutation The present work shows that tg2576 mice are deeply impaired in the spatial novelty task, but not in the object recognition task, at the ages of 8 and 16 months, whatever their rd status. The preserved ability of tg2576 mice to discriminate between familiar and novel objects suggests that object recognition capacities are unaffected even at the latest age tested herein. Thus, the deficit in the spatial novelty task is more likely to reveal a selective deficit in memory for object location. In addition, this is the first study showing that tg2576 mice younger than 7 months (i.e., 5 months of age mice tested in a separate experiment here) performed well in detecting object relocation (Good et al., 2007; Hale and Good, 2005; Ognibene et al., 2005). Thus, the age of onset of their spatial novelty deficits is most probably approximately 6 months of age. Nevertheless, 1 of our most interesting results is that the spatial novelty deficit of tg2576 mice is independent of their rd mutation status. Whether homozygous for the rd mutation or not, wt mice were definitely able to detect the object relocation, whereas tg2576 mice were deeply impaired. The ability to discriminate between 2 objects (novel vs. familiar) and to detect object displacements in virtually blind wt mice may appear surprising because the object recognition task is classically considered as depending on visual perception. Given our criterion for object exploration (less than 1 cm distance), all the mice considered in the present experiments explored the objects at least with their whiskers and possibly with their paws. It is known that vibrissae-dependent exploration of encountered surfaces provide sufficient information to allow object recognition in rodents (Brecht et al., 1997). Therefore, our results suggest that the widely used object recognition paradigm should definitely not be considered as depending solely on visual information, but can be successfully performed by blind rodents. Concerning the spatial novelty task, it must be noted that vision is not an absolute requirement for spatial awareness within the body space when subjects are allowed to explore their environment with other sensory inputs (Jones, 1975). Several studies suggest that blindness does not affect the ability of humans and rodents to elaborate a spatial representation of a new environment (Haber et al., 1993; Kimchi and Terkel, 2001). Hippocampal place cells, whose discharges are strongly related to the location of a subject in a given environment, are thought to be part of an integrated system sustaining cognitive maps necessary for spatial memory and navigation. Interestingly, some studies suggest that blind rodents are able to elaborate cognitive maps and to use them to organize goal-directed behaviors (Farr et al., 2002; Save et

727

al., 1998). In the present spatial novelty task, the blind mice had access to a variety of information (e.g., mainly tactual and kinesthetic, eventually olfactive and auditive) which could have been integrated to establish cognitive maps sufficiently accurate to allow detection of the object relocation. Interestingly, 16-month-old (but not 3-month-old) tg2576 mice show a degraded activity of their place cells which correlates with hippocampal amyloid deposition and spatial memory deficits in a forced-choice T-maze alternation task (Cacucci et al., 2008). Moreover, dendritic spine density and synaptic plasticity is affected in their dentate gyrus at ages as early as 4 – 6 months (Jacobsen et al., 2006). Because the hippocampal region is known to play a crucial role in memory for object location (Bohbot et al., 2002), it is tempting to suggest that such neuroanatomical and functional alterations could play a role in the early deficit of tg2576 mice in the spatial novelty task. 4.2. Spatial memory impairment in the Barnes maze is not influenced by the rd mutation The Barnes maze task used in the present study was specially designed to test several aspects of learning and memory performance in tg2576 mice. These mice showed limited deficits on short-term and initial phases of acquisition on the first 2 days. On the reversal day, changing the position of the target hole to the opposite side of the maze induced moderate perturbations in latency and error performances similarly in wt and tg2576 mice. However, a clear genotype-dependent deficit was observed when considering the whole reversal phase. The difficulty of tg2576 mice to learn the new position of the target hole was associated with a delayed use of a direct spatial strategy, especially in the oldest, as found in APP23 mice in a similar Barnes maze task (Prut et al., 2007). Note that wt, and to a lesser extent tg2576 mice, reached relatively high percentages of spatial strategy trials (wt approximately 90% and tg2576 approximately 50%) by the end of the task, which confirms that the present Barnes maze device is adapted to evaluate spatial memory compared with other devices favoring nonspatial strategies (O’Leary et al., 2011). Thus, our results clearly suggest that tg2576 mice acquired quite slowly the spatial aspects of the task during the reversal phase. Interestingly, the early emergence of reversal-learning deficits in tg2576 mice has been related to increased levels of A␤ in their prefrontal cortex (Zhuo et al., 2008). Although, with extensive reversal training, transgenic animals ended up reaching nearly normal levels of performance, 16-month-old tg2576 mice took a longer time to complete the trials and showed a reduced ability to use a spatial strategy compared with 8-month-old tg2576 mice and wt mice. Similar results were obtained when comparing sighted and visually impaired mice on latency and spatial strategy use (i.e., higher latencies and reduced use of spatial strategy in 16-month-old blind mice). Thus, it is unclear whether the higher latency and the delayed use of a spatial strategy are specific to

728

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

16-month-old tg2576 mice or if it is due to the high proportion of rd/rd subjects within this group. This clearly illustrates the possible impact of the proportion of rd homozygotes on spatial navigation performances of a given tg2576 group and the risk to obtain false positive results related to the rd/rd proportion rather than the transgenic genotype. Nevertheless, it is noteworthy that the deficits of tg2576 mice on errors and memory retention performances were not affected by the rd status and were similar at the 2 ages tested. In fact, the most prominent and robust deficits were obtained on the probe tests. Whether it was performed 24 hours after the short acquisition phase or 1 week after the reversal training, tg2576 mice showed no evidence that they remembered the target hole position. At first, these results may contrast with the lack of deficit reported by other authors (see Stewart et al., 2011), but note that our probe tests were made more difficult on purpose by testing memory retention after a short training and at a remote 7-day delay after extensive training. Taken together, our results strongly suggest that, while spatial learning appears moderately affected in tg2576 mice, long-term spatial memory is dramatically impaired at the ages of 8 and 16 months regardless of the rd status. One of the most surprising results of the present study is the preserved ability of blind wt mice to use a spatial strategy to find the target hole and their excellent performances in the Barnes-maze probe tests. These results contrast with the very low level of performance during training and probe tests in the Morris water maze as shown here and previously in other visually impaired mice, including rd/rd tg2576/wt mice (Brown and Wong, 2007; Clapcote et al., 2005; Garcia et al., 2004). The 2 tasks are based on the similar principle of finding a “hidden” goal to escape from an unsecure/aversive environment. In addition, both tasks were performed in similar conditions according to extramaze cues (visual, auditory, perhaps olfactory) available in the testing room. At least 2 explanations can be proposed here. First, the water maze is known to be much more stressful for mice than dry-land mazes, such as the Barnes maze (Harrison et al., 2009; Whishaw and Tomie, 1996). This could make the former task more difficult to learn than the later for visually deprived mice. However, even the rats, more adapted to the aquatic world, are deeply impaired when only auditory cues are available, whereas a single visual cue is sufficient for them to acquire this task in total darkness (Rossier et al., 2000). Clearly, the visual modality plays a crucial role in rodents to learn the spatial aspects of a water maze task. On the other hand, the absence of visual cues (due to blindness or environmental characteristics) in the Barnes maze task may be compensated by the use of nonspatial strategies (serial strategy) and/or cues provided on other sensory modalities (O’Leary et al., 2011). One explanation for the higher dependency of water maze performance on visual information is provided by the elegant work of Sautter et al. (2008), which shows that propriocep-

tive information is used with much poorer accuracy in a water maze than in a dry-land maze due to movement inertia. This suggests that proprioceptive information does not help much blind mice to navigate in the water maze. Conversely, in the Barnes maze, blind mice could rely on auditory and olfactory cues, as well as these motion-related information to establish spatial maps (Save et al., 1998) and successfully navigate to the target hole. In addition, blindness in humans and animals is associated with improved spatial hearing abilities and capacities of self-positioning from auditory cues (Després et al., 2005; King and Parsons, 1999; Lessard et al., 1998). Thus, our early blind mice most certainly used the auditory cues available with great efficiency to orient their spatial map and navigate through the maze. 4.3. Memory performance not sensitive to the rd mutation is preserved in tg2576 mice In the present study, 3 behavioral tests did not show any difference between wt and tg2576 mice. Our results suggest an age-dependent impairment of Y-maze alternation performance in wt mice, but no genotype effect. It must be noted that such a deficit has been reported quite inconsistently in tg2576 mice, even within the same research groups (e.g., Hsiao et al., 1996; King and Arendash, 2002; Middei et al., 2004; but see King et al., 1999; Deacon et al., 2008). As could be expected from a task which can be performed on the basis of egocentric orientation, the rd status did not affect alternation performance. The rd status is most probably not involved in the discrepancies reported on alternation behavior. Thus, until the reasons for these discrepancies are clearly identified (e.g., maze size, trial duration, nonmnemonic confounding factors, genetic background), this task appears poorly reliable to evaluate memory deficits in tg2576 mice on a B6:SJL background (Rustay et al., 2010). Another task that does not seem adapted to test memory deficits in this mouse line is the olfactory discrimination task. The use of tasks based on olfaction seemed yet justified because olfactory functioning is affected in AD (Mesholam et al., 1998) and 8-month-old tg2576 mice are known to show reduced neurogenesis in the olfactory bulb (Guérin et al., 2009). However, we did not find any deficit in our simple odor discrimination task. It seems that such deficits in tg2576 mice can only be detected during a reversal phase or a more complicated odor span task taxing working memory (Guérin et al., 2009; Young et al., 2009). Finally, habituation to a novel environment appeared globally similar in wt and tg2576 mice, whereas rd/rd mice tended to be more active than sighted mice especially in wt mice. We also reported here the strong impact of stereotypies in both tg2576 and wt mice on activity scores, which led us to eliminate several animals in order to reduce the variability of these data. Nevertheless, it seems possible that both the rd status and the occurrence of stereotypies con-

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

tributed to the high variability of activity scores described in the rare studies monitoring activity rhythms in tg2576 mice (Gorman and Yellon, 2010; Ognibene et al., 2005) and could have masked the consistent hyperactivity induced by the presence of the transgene as shown on a different background (Rustay et al., 2010). 4.4. Conclusions Retinal degeneration can be a major confounder of experimental data in behavioral tests that require integration of visuospatial information. This is particularly true in the context of spatial memory evaluations in tg2576 mice. However, we conclude here that the spatial novelty task is extremely well adapted to test early memory deficits in these mice regardless of their rd mutation status. Moreover, provided minor environmental adaptations (e.g., the presence of extramaze auditory and olfactory cues), the Barnes maze task can also be useful to characterize the spatial navigation deficits of tg2576 mice, which seem to progress between the ages of 8 and 16 months. These findings should help investigators in testing new AD therapies with behavioral paradigms adapted to the B6:SJL genetic background of tg2576 mice, especially because this background appears as 1 of the most suitable for an optimal expression of the APPSWE transgene to model AD (Lassalle et al., 2008). Tg2576 mice on a 129 genetic background may offer an alternative for behavioral studies (Rustay et al., 2010; Westerman et al., 2002). In general, the use of spatial novelty and Barnes maze tasks is highly recommended to test spatial memory in mouse models bred on genetic backgrounds carrying mutations associated with various degrees of visual impairment (e.g., in SLJ, C3H/HeJ, CBA/J, FVB/N backgrounds) and it could sidestep the time-consuming and costly selection of experimental animals free of mutation-related blindness. In addition, the present work shows the remarkably preserved abilities of visually impaired mice to recognize a familiar object, to detect its relocation in a familiar environment, and to navigate spatially within a large dry-land maze to find an escape hole.

Disclosure statement There are no actual or potential conflicts of interest. A.L. and C.D.C. are employees of Boehringer Ingelheim Pharma, GmbH & Co., KG. All experimental procedures were conducted in conformity with the institutional guidelines (council Directive 87/848, October 19, 1987, Ministère de l’agriculture et de la Forêt, Service Vétérinaire de la Santé et de la Protection Animale; NIH publication, 86 –23, revised 1985). Official permission references for animal experimentation were 67–292 for C.M. and 67–215 for J.-C.C.; N.Y. was under their responsibility. Reference of official

729

permission for holding genetically modified organisms (tg2576) was 5016 CA-II. Acknowledgements This work was supported solely by Boehringer Ingelheim Pharma, GmbH (Biberach an der Riss, Germany), University de Strasbourg and Centre National de la Recherche Scientifique. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.neurobiolaging.2012.06.016. References Bohbot, V.D., Jech, R., Ru˚zicka, E., Nadel, L., Kalina, M., Stepánková, K., Bures, J., 2002. Rat spatial memory tasks adapted for humans: characterization in subjects with intact brain and subjects with selective medial temporal lobe thermal lesions. Physiol. Res. 51(Suppl 1):S49 – S65. Bowes, C., Li, T., Danciger, M., Baxter, L.C., Applebury, M.L., Farber, D.B., 1990. Retinal degeneration in the rd mouse is caused by a defect in the beta subunit of rod cGMP-phosphodiesterase. Nature 347, 677– 680. Brecht, M., Preilowski, B., Merzenich, M.M., 1997. Functional architecture of the mystacial vibrissae. Behav. Brain Res. 84, 81–97. Brown, R.E., Wong, A.A., 2007. The influence of visual ability on learning and memory performance in 13 strains of mice. Learn. Mem. 14, 134 –144. Buck, B.H., Black, S.E., Behrmann, M., Caldwell, C., Bronskill, M.J., 1997. Spatial- and object-based attentional deficits in Alzheimer’s disease. Relationship to HMPAO-SPECT measures of parietal perfusion. Brain 120, 1229 –1244. Cacucci, F., Yi, M., Wills, T.J., Chapman, P., O’Keefe, J., 2008. Place cell firing correlates with memory deficits and amyloid plaque burden in Tg2576 Alzheimer mouse model. Proc. Natl. Acad. Sci. U. S. A. 105, 7863–7868. Carter-Dawson, L.D., LaVail, M.M., Sidman, R.L., 1978. Differential effect of the rd mutation on rods and cones in the mouse retina. Invest. Ophthalmol. Vis. Sci. 17, 489 – 498. Chapman, P.F., White, G.L., Jones, M.W., Cooper-Blacketer, D., Marshall, V.J., Irizarry, M., Younkin, L., Good, M.A., Bliss, T.V., Hyman, B.T., Younkin, S.G., Hsiao, K.K., 1999. Impaired synaptic plasticity and learning in aged amyloid precursor protein transgenic mice. Nat. Neurosci. 2, 271–276. Clapcote, S.J., Lazar, N.L., Bechard, A.R., Wood, G.A., Roder, J.C., 2005. NIH Swiss and Black Swiss mice have retinal degeneration and performance deficits in cognitive tests. Comp. Med. 55, 310 –316. Deacon, R.M., Cholerton, L.L., Talbot, K., Nair-Roberts, R.G., Sanderson, D.J., Romberg, C., Koros, E., Bornemann, K.D., Rawlins, J.N., 2008. Age-dependent and -independent behavioral deficits in Tg2576 mice. Behav. Brain Res. 189, 126 –138. Deacon, R.M., Koros, E., Bornemann, K.D., Rawlins, J.N., 2009. Aged Tg2576 mice are impaired on social memory and open field habituation tests. Behav. Brain Res. 197, 466 – 468. Després, O., Candas, V., Dufour, A., 2005. The extent of visual deficit and auditory spatial compensation: evidence from self-positioning from auditory cues. Brain Res. Cogn. Brain Res. 23, 444 – 447. Farr, S.A., Banks, W.A., La Scola, M.E., Morley, J.E., 2002. Blind mice are not impaired in T-maze footshock avoidance acquisition and retention. Physiol. Behav. 76, 531–538.

730

N. Yassine et al. / Neurobiology of Aging 34 (2013) 716 –730

Garcia, M.F., Gordon, M.N., Hutton, M., Lewis, J., McGowan, E., Dickey, C.A., Morgan, D., Arendash, G.W., 2004. The retinal degeneration (rd) gene seriously impairs spatial cognitive performance in normal and Alzheimer’s transgenic mice. Neuroreport 15, 73–77. Giménez, E., Montoliu, L., 2001. A simple polymerase chain reaction assay for genotyping the retinal degeneration mutation (Pdeb(rd1)) in FVB/N-derived transgenic mice. Lab. Anim. 35, 153–156. Good, M.A., Hale, G., Staal, V., 2007. Impaired “episodic-like” object memory in adult APPswe transgenic mice. Behav. Neurosci. 121, 443– 448. Gorman, M.R., Yellon, S., 2010. Lifespan daily locomotor activity rhythms in a mouse model of amyloid-induced neuropathology. Chronobiol. Int. 27, 1159 –1177. Guérin, D., Sacquet, J., Mandairon, N., Jourdan, F., Didier, A., 2009. Early locus coeruleus degeneration and olfactory dysfunctions in Tg2576 mice. Neurobiol. Aging 30, 272–283. Haber, R.N., Haber, L.R., Levin, C.A., Hollyfield, R., 1993. Properties of spatial representations: data from sighted and blind subjects. Percept. Psychophys. 54, 1–13. Hale, G., Good, M., 2005. Impaired visuospatial recognition memory but normal object novelty detection and relative familiarity judgments in adult mice expressing the APPswe Alzheimer’s disease mutation. Behav. Neurosci. 119, 884 – 891. Harrison, F.E., Hosseini, A.H., McDonald, M.P., 2009. Endogenous anxiety and stress responses in water maze and Barnes maze spatial memory tasks. Behav. Brain Res. 198, 247–251. Harrison, F.E., Reiserer, R.S., Tomarken, A.J., McDonald, M.P., 2006. Spatial and nonspatial escape strategies in the Barnes maze. Learn. Mem. 13, 809 – 819. Holcomb, L.A., Gordon, M.N., Jantzen, P., Hsiao, K., Duff, K., Morgan, D., 1999. Behavioral changes in transgenic mice expressing both amyloid precursor protein and presenilin-1 mutations: lack of association with amyloid deposits. Behav. Genet. 29, 177–185. Hsiao, K., Chapman, P., Nilsen, S., Eckman, C., Harigaya, Y., Younkin, S., Yang, F., Cole, G., 1996. Correlative memory deficits, Abeta elevation, and amyloid plaques in transgenic mice. Science 274, 99 –102. Jacobsen, J.S., Wu, C.C., Redwine, J.M., Comery, T.A., Arias, R., Bowlby, M., Martone, R., Morrison, J.H., Pangalos, M.N., Reinhart, P.H., Bloom, F.E., 2006. Early-onset behavioral and synaptic deficits in a mouse model of Alzheimer’s disease. Proc. Natl. Acad. Sci. U. S. A. 103, 5161–5166. Jones, B., 1975. Spatial perception in the blind. Br. J. Psychol. 66, 461– 472. Kalová, E., Vlcek, K., Jarolímová, E., Bures, J., 2005. Allothetic orientation and sequential ordering of places is impaired in early stages of Alzheimer’s disease: corresponding results in real space tests and computer tests. Behav. Brain Res. 159, 175–186. Kawarabayashi, T., Younkin, L.H., Saido, T.C., Shoji, M., Ashe, K.H., Younkin, S.G., 2001. Age-dependent changes in brain, CSF, and plasma amyloid (beta) protein in the Tg2576 transgenic mouse model of Alzheimer’s disease. J. Neurosci. 21, 372–381. Kimchi, T., Terkel, J., 2001. Spatial learning and memory in the blind mole-rat in comparison with the laboratory rat and Levant vole. Anim. Behav. 61, 171–180. King, A.J., Parsons, C.H., 1999. Improved auditory spatial acuity in visually deprived ferrets. Eur. J. Neurosci. 11, 3945–3356. King, D.L., Arendash, G.W., 2002. Behavioral characterization of the Tg2576 transgenic model of Alzheimer’s disease through 19 months. Physiol. Behav. 75, 627– 642. King, D.L., Arendash, G.W., Crawford, F., Sterk, T., Menendez, J., Mullan, M.J., 1999. Progressive and gender-dependent cognitive impairment in the APP(SW) transgenic mouse model for Alzheimer’s disease. Behav. Brain Res. 103, 145–162. Lassalle, J.M., Halley, H., Daumas, S., Verret, L., Francés, B., 2008. Effects of the genetic background on cognitive performances of TG2576 mice. Behav. Brain Res. 191, 104 –110.

Lessard, N., Paré, M., Lepore, F., Lassonde, M., 1998. Early-blind human subjects localize sound sources better than sighted subjects. Nature 395, 278 –280. Mesholam, R.I., Moberg, P.J., Mahr, R.N., Doty, R.L., 1998. Olfaction in neurodegenerative disease: a meta-analysis of olfactory functioning in Alzheimer’s and Parkinson’s diseases. Arch. Neurol. 55, 84 –90. Middei, S., Geracitano, R., Caprioli, A., Mercuri, N., Ammassari-Teule, M., 2004. Preserved fronto-striatal plasticity and enhanced procedural learning in a transgenic mouse model of Alzheimer’s disease overexpressing mutant hAPPswe. Learn. Mem. 11, 447– 452. Moreau, P.H., Cosquer, B., Jeltsch, H., Cassel, J.C., Mathis, C., 2008. Neuroanatomical and behavioral effects of a novel version of the cholinergic immunotoxin mu p75-saporin in mice. Hippocampus 18, 610 – 622. O’Leary, T.P., Savoie, V., Brown, R.E., 2011. Learning, memory and search strategies of inbred mouse strains with different visual abilities in the Barnes maze. Behav. Brain Res. 216, 531–542. Ognibene, E., Middei, S., Daniele, S., Adriani, W., Ghirardi, O., Caprioli, A., Laviola, G., 2005. Aspects of spatial memory and behavioral disinhibition in Tg2576 transgenic mice as a model of Alzheimer’s disease. Behav. Brain Res. 156, 225–232. Oliveira, A.M., Hawk, J.D., Abel, T., Havekes, R., 2010. Post-training reversible inactivation of the hippocampus enhances novel object recognition memory. Learn. Mem. 17, 155–160. Prut, L., Abramowski, D., Krucker, T., Levy, C.L., Roberts, A.J., Staufenbiel, M., Wiessner, C., 2007. Aged APP23 mice show a delay in switching to the use of a strategy in the Barnes maze. Behav. Brain Res. 179, 107–110. Rossier, J., Haeberli, C., Schenk, F., 2000. Auditory cues support place navigation in rats when associated with a visual cue. Behav. Brain Res. 117, 209 –214. Rustay, N.R., Cronin, E.A., Curzon, P., Markosyan, S., Bitner, R.S., Ellis, T.A., Waring, J.F., Decker, M.W., Rueter, L.E., Browman, K.E., 2010. Mice expressing the Swedish APP mutation on a 129 genetic background demonstrate consistent behavioral deficits and pathological markers of Alzheimer’s disease. Brain Res. 1311, 136 –147. Sautter, C.S., Cocchi, L., Schenk, F., 2008. Dynamic visual information plays a critical role for spatial navigation in water but not on solid ground. Behav. Brain Res. 194, 242–245. Save, E., Cressant, A., Thinus-Blanc, C., Poucet, B., 1998. Spatial firing of hippocampal place cells in blind rats. J. Neurosci. 18, 1818 –1826. Schellinck, H.M., Forestell, C.A., LoLordo, V.M., 2001. A simple and reliable test of olfactory learning and memory in mice. Chem. Sens. 26, 663– 672. Stewart, S., Cacucci, F., Lever, C., 2011. Which memory task for my mouse? A systematic review of spatial memory performance in the Tg2576 Alzheimer’s mouse model. J. Alzheimers Dis. 26, 105–126. Westerman, M.A., Cooper-Blacketer, D., Mariash, A., Kotilinek, L., Kawarabayashi, T., Younkin, L.H., Carlson, G.A., Younkin, S.G., Ashe, K.H., 2002. The relationship between Abeta and memory in the Tg2576 mouse model of Alzheimer’s disease. J. Neurosci. 22, 1858 – 1867. Whishaw, I.Q., Tomie, J., 1996. Of mice and mazes: similarities between mice and rats on dry land but not water mazes. Physiol. Behav. 60, 1191–1197. Young, J.W., Sharkey, J., Finlayson, K., 2009. Progressive impairment in olfactory working memory in a mouse model of mild cognitive impairment. Neurobiol. Aging 30, 1430 –1443. Zhuo, J.M., Prakasam, A., Murray, M.E., Zhang, H.Y., Baxter, M.G., Sambamurti, K., Nicolle, M.M., 2008. An increase in Abeta42 in the prefrontal cortex is associated with a reversal-learning impairment in Alzheimer’s disease model Tg2576 APPsw mice. Curr. Alzheimer Res. 5, 385–391.