Firstly, we wondered whether a single icv. injection of Aβ1–42 would affect the encoding and retrieval of spatial and habituation learning and memory in a sex-specific manner. Thus, female and male mice underwent a protocol to measure spatial learning and memory using a Barnes maze or an OLM test, as well as an exploratory habituation protocol using an OF to assess this type of non-associative memory.

Aβ 1–42 impairs spatial learning and memory encoding in both female and male mice

To evaluate spatial memory encoding, we employed the Barnes Maze task, where animals performed 3 trials per day over 4 consecutive days (days 0–3), starting 1 h after icv. injections (performed in day 0, Fig. 1) of either Aβ1–42 or both controls: vehicle and reverse Aβ42–1 peptide. The latency to find the open hole was measured in the first and last trial of each day (Fig. 2A) to assess working memory [43]. No differences between males and females were found within each treatment group (sex effect: F(1,118) = 0.096, p = 0.757). Nevertheless, our data showed a significant treatment (F(2,118) = 18.147, p < 0.001) and time (F(1,118) = 42.742, p < 0.001) effects when comparing the latency to find the escape hole in the first and last trial from each training session, proving that both vehicle (male, n = 14; female, n = 15) and Aβ42–1 (male, n = 7; female, n = 6) groups had a normal spatial working memory, while the Aβ1–42 group (male, n = 14; female, n = 16) showed a deterioration in the formation of this kind of memory, irrespective of sex. Furthermore, the evolution of different parameters measured over the four training days allowed us to evaluate short-term memory, since animals should remember how to perform the task and therefore latencies would decrease over time, even if the target hole was changed. Accordingly, our data showed treatment and time effects for latency (Fig. 2B; F(2,254) = 29.623, p < 0.001 and F(3,254) = 22.861, p < 0.001, respectively), number of errors (Fig. 2C; F(2,241) = 29.358, p < 0.001 and F(3,241) = 3.365, p = 0.019, respectively) and distance traveled (Fig. 2D; F(2,231) = 26.673, p < 0.001 and F(3,231) = 8.508, p < 0.001, respectively). Post-hoc analysis revealed that Aβ1–42 disrupted spatial learning in both female and male animals, since neither latency (F(1,254) = 0.212, p = 0.646), errors (F(1,241) = 1.046, p = 0.307), nor distance (F(1,231) = 0.095, p = 0.758) showed a sex effect. Finally, long-term memory was assessed through two memory tests conducted on days 7 and 11 post-injection, during which all holes were closed (Fig. 2E). Latency to reach the target hole of the last training day (day 3), as well as distance traveled, were quantified. Our results showed a statistically significant difference in latency (Fig. 2F; treatment effect: F(2,113) = 15.894, p < 0.001; sex effect: F(1,113) = 0.002, p = 0.967) and distance traveled (Fig. 2G; treatment effect: F(2,231) = 26.673, p < 0.001; sex effect: F(1,231) = 0.095, p = 0.758). Once again, post-hoc analysis revealed that the differences were due to a worse performance of the Aβ1–42-treated mice.

Fig. 2

Aβ1–42 impairs spatial learning and memory encoding in both female and male mice. A Representative traces of the path traveled during the first and last trial of the last training day (day 3) for Aβ1–42 and controls: vehicle and reverse Aβ42–1 peptide. B–D Escape latency (B; in s), number of errors (C) and distance traveled (D; in cm) during the four training days. Data is expressed as the mean ± SEM of the 3 trials per day. E Overview image of the test phase on the Barnes maze, with all holes closed. F Latency (in s) to reach the target hole of the latest training day for the first time, during the two test sessions. G Distance traveled (in cm) during the two test sessions. H Representative traces of the three possible search strategies: random, serial, and spatial. I Ratio of the use of each search strategy for all the experimental groups during training (Days 0 and 3) and tests (Days 7 and 11) sessions. Stacked bars are normalized so that the sum of the three strategies each day is 100%. N vehicles: males = 14 and females = 15; N Aβ1–42: males = 14 and females = 16; N reverse Aβ42–1 males = 7 and females = 6. Aβ, Amyloid-β; cm, centimeters; s, seconds. *p < 0.05, **p < 0.01, *** p < 0.001 vs. vehicle of the corresponding sex; # p < 0.05, ## p < 0.01 vs. Aβ42–1 of the corresponding sex

Furthermore, navigation strategy during the training and tests sessions was studied for all treatment groups. Although the Barnes maze is considered to encourage allocentric strategies due to the use of distal visual cues [54], different searching strategies, that reflect different levels of learning, could be used by the animals: random, serial and spatial (Fig. 2H). Our data showed that all groups initially employed the random strategy, which decreased across days (Fig. 2I; F(2.85,182.414) = 12.535, p < 0.001, Geisser-Greenhouse’s correction). However, a significant treatment effect was found (F(5.7,182.414) = 2.575, p = 0.022, Geisser-Greenhouse’s correction), while no sex effect was observed (F(2.85,182.414) = 0.037, p = 0.989, Geisser-Greenhouse’s correction). This indicates that both male and female mice treated with Aβ1–42 continued to rely on the random strategy compared to the control groups, which is consistent with the impaired memory observed in the Aβ1–42 group.

Thus, these data suggest an impairment in the encoding of hippocampal-dependent spatial working, short- and long-term memory and learning processes induced by Aβ1–42 in both sexes. In contrast, same concentrations of the control reverse peptide, Aβ42–1, did not differ from the control vehicle, suggesting that the decline induced by Aβ1–42 was therefore specific.

Aβ 1–42 impairs the encoding of exploratory habituation memory in both female and male mice

Afterwards, exploratory habituation memory encoding was assessed using an OF habituation task on days 15 and 16 post-icv. injection (Fig. 3A). Data showed no significant differences in exploration between the groups during the training session (OF1). However, when comparing OF1 and the retrieval session (OF2) (Fig. 3B), significant treatment (F(2,84) = 3.464, p = 0.0358) and time effects (F(1,84) = 34.95, p < 0.0001) were observed, with no sex effect (F(1,84) = 0.0023, p = 0.9614), showing a decrease in exploration in both controls, vehicle (Male, n = 7; Female, n = 12) and Aβ42–1 (Male, n = 7; Female, n = 6) treated animals, regardless of sex, proving that they were able to remember the arena. Conversely, Aβ1–42-treated mice (Male, n = 9; Female, n = 10) performed slightly better without reaching significance, suggesting some deterioration of memory encoding in this group of animals. Moreover, during the retrieval session (OF2), a significant treatment effect was found (Fig. 3B; F(2,42) = 9.342, p = 0.0004) without a sex effect (F(1,42) = 0.0063, p = 0.9369). Post-hoc analyses revealed that both male and female Aβ1–42 mice traveled a longer distance during this session compared to the vehicle and Aβ42–1 control groups, as illustrated in Fig. 3C. Hence, both results show that Aβ1–42 also impairs the ability to generate non-associative habituation memory even 2 weeks after treatment.

Fig. 3

Aβ1–42 administration similarly alters non-associative exploratory habituation memory encoding in both female and male mice. A An OF habituation test was carried out by submitting the animals to the same OF arena twice, on consecutive days 15 and 16 post-icv. injection. B Total distance traveled during the two OF sessions (training -OF1- and retrieval -OF2- sessions). Data is expressed as the percentage (%) of the distance traveled in the training session (OF1). C Examples of mice movement tracked during OF1 and OF2 for the different treatment groups. N vehicles: males = 7 and females = 12; N Aβ1–42: males = 9 and females = 10; N reverse Aβ42–1: males = 7 and females = 6. Aβ, Amyloid-β; icv., intracerebroventricular; OF, open field. * p < 0.05, ** p < 0.01 vs. vehicle of the corresponding sex; # p < 0.05, ## p < 0.01 vs. Aβ42–1 of the corresponding sex; ‡ p < 0.05, ‡‡ p < 0.01, ‡‡‡ p < 0.001 vs. OF1

Aβ 1–42 impairs spatial and habituation memory retrieval in both female and male mice

The above experiments proved the deleterious effect of a single injection of Aβ1–42 on the encoding phase of memory when delivered before learning each task. To investigate the impact on memory retrieval, a second set of experiments was conducted. In this set, Aβ1–42 was injected after the learning phase to evaluate both spatial and exploratory habituation memories using an OLM test and the OF habituation task, respectively (Fig. 4).

Fig. 4

Aβ1–42 administration also alters spatial and non-associative memory retrieval in both female and male mice. A An OLM test was performed, changing the location of one object between the training and each memory test (OLM1 and OLM2). Treatment was administered icv. between OLM1 and OLM2 to evaluate memory retrieval. B Discrimination index during the training, OLM1, and OLM2 sessions. Data is expressed as the mean ± SEM. C An OF habituation test was carried out by submitting the animals to the same OF arena twice, administering the icv. injection between the training and the retrieval sessions. D Total distance traveled during the two OF sessions (training -OF1- and retrieval -OF2- sessions). Data is expressed as the percentage (%) of the distance traveled in the training session (OF1). E Examples of mice movement tracked during OF1 and OF2 for the different treatment groups. N vehicles: males = 8–9 and females = 7; N Aβ1–42: males = 9–10 and females = 7–8; N reverse Aβ42–1: males = 8–9 and females = 6. Aβ, Amyloid-β; icv., intracerebroventricular; OF, open field; OLM, object location memory. *p < 0.05, **p < 0.01, ***p < 0.001 vs. vehicle of the corresponding sex; #p < 0.05, ##p < 0.01 vs. reverse Aβ42–1 of the corresponding sex; ‡‡p < 0.01, ‡‡‡p < 0.001 vs. OF1

Data from the OLM training session showed that all the animals spent equal amount of time exploring both objects (DI ≈ 0 accounts for no preference for a specific object which could have influenced the later results). During OLM1, no differences in the DI due to either sex (F(1,41) = 1.4, p = 0.2436) or treatment (F(2,41) = 1.011, p = 0.3727) were found, indicating that all naïve animals had proper spatial memory encoding. However, during OLM2, conducted 1 h after treatment, two-way ANOVA showed a significant treatment effect (F(2,40) = 23.26, p < 0.0001) regardless of sex (F(1,40) = 0.0003, p = 0.9855). Post-hoc analysis revealed that Aβ1–42 treated mice (Male, n = 10; Female, n = 7) exhibited a lower DI compared to both vehicle (Male, n = 8; Female, n = 7) and Aβ42–1 mice (Male, n = 9; Female, n = 6) of the corresponding sex (Fig. 4A, B).

Furthermore, the OF habituation task (Fig. 4C–E) showed a significant treatment (F(2,82) = 3.277, p = 0.0427) and time effects (F(1,82) = 108.9, p < 0.0001), with no sex effect (F(1,82) = 0.07784, p = 0.7809), when comparing the pre-treatment OF1 with the post-treatment OF2. Resembling earlier outcomes, there was a decrease in exploration in both vehicle (Male, n = 9; Female, n = 7) and reverse Aβ42–1 (Male, n = 8; Female, n = 6) animals, indicating memory retrieval of the arena. In contrast, Aβ1–42-treated mice (Male, n = 9; Female, n = 8) performed slightly better without reaching significance, suggesting some deterioration of memory retrieval for this group. Moreover, during the retrieval session (OF2), a significant treatment effect was found (F(2,41) = 7.647, p = 0.0015), with no sex effect (F(1,41) = 0.1816, p = 0.6722). Post-hoc analyses revealed that both male and female Aβ1–42 mice traveled a longer distance during this session compared to vehicle and reverse Aβ42–1 control groups, as illustrated in Fig. 4D, E.

These results collectively show that Aβ1–42 also impairs both spatial and non-associative habituation memory when administered after learning, thus affecting both encoding and retrieval processes.

Memory impairments were not due to health or locomotor disfunction

Mice underwent a battery of behavioral tests in order to assess general health conditions and confirm that the impairments observed in the encoding and retrieval of hippocampal-dependent spatial and habituation memory were due to specific hippocampal alterations triggered by Aβ1–42 injection. Using the LABORAS® to assess stereotyped behaviors, data showed that all groups (vehicle male, n = 11; vehicle female, n = 12; Aβ1–42 male, n = 11; Aβ1–42 female, n = 12; Aβ42–1 male, n = 7; Aβ42–1 female, n = 6) spent the same amount of time performing the different analyzed behaviors (Fig. 5A): locomotion (treatment effect: F(2,53) = 0.630, p = 0.5334; sex effect: F(1,53) = 0.8804, p = 0.3523), rearing (treatment effect: F(2,53) = 1.204, p = 0.308; sex effect: F(1,53) = 0.6201, p = 0.4345) and grooming (treatment effect: F(2,52) = 1.885, p = 0.162; sex effect: F(1,52) = 3.998, p = 0.051). Climbing behavior exhibited a significant sex effect (F(1,53) = 6.770, p = 0.012), showing that female mice tended to climb more than male, regardless of the treatment (F(2,53) = 1.626, p = 0.2064). All groups equally improved their performance in the rotarod test (Fig. 5B; vehicle male, n = 8; vehicle female, n = 11; Aβ1–42 male, n = 8; Aβ1–42 female, n = 9; Aβ42–1 male, n = 7; Aβ42–1 female, n = 6; time effect: F(5,294) = 3.646, p = 0.003), as no differences in the latency to fall off the rod were found between groups along trials (treatment effect: F(2,258) = 1.552, p = 0.214; sex effect: F(1,258) = 0.415, p = 0.52) nor in the whole session (treatment effect: F(2,43) = 0.36, p = 0.6998; sex effect: F(1,43) = 0.0957, p = 0.7585). Locomotion was also tested in all groups (vehicle male, n = 8; vehicle female, n = 12; Aβ1–42 male, n = 9; Aβ1–42 female, n = 12; Aβ42–1 male, n = 7; Aβ42–1 female, n = 6) using the elevated plus maze. The number of entries into closed (treatment effect: F(2,48) = 0.2235, p = 0.8006; sex effect: F(1,48) = 0.03188, p = 0.859) and total arms (treatment effect: F(2,48) = 0.01929, p = 0.9809; sex effect: F(1,48) = 0.0001, p = 0.9899) did not differ significantly between groups due to either treatment or sex (Fig. 5C), indicating similar locomotor activity. Regarding stress-related behaviors, all animals had the same number of entries (Fig. 5C; treatment effect: F(2,48) = 1.034, p = 0.3633; sex effect: F(1,48) = 0.224, p = 0.6381) and time spent in open arms (treatment effect: F(2,48) = 0.6591, p = 0.5219; sex effect: F(1,48) = 0.02385, p = 0.8779) in the elevated plus maze, and they all had the same immobility time (Fig. 5D; vehicle male, n = 8; vehicle female, n = 12; Aβ1–42 male, n = 9; Aβ1–42 female, n = 12; Aβ42–1 male, n = 7; Aβ42–1 female, n = 6; treatment effect: F(2,48) = 0.05294, p = 0.9485; sex effect: F(1,48) = 0.1535, p = 0.6969) in the tail suspension test, suggesting that the treatment did not increase depression- nor anxiety-like behaviors.

Fig. 5

Aβ1–42 administration does not induce alterations in locomotor activity, anxiety, and depression-like behavior. Behavioral tasks to evaluate general health state were carried out on days 15–17 post-icv. injection. A Stereotyped behaviors were assessed using a LABORAS® system, measuring the time (in s) spent performing each type of activity (locomotion, climbing, rearing, and grooming). B Latency (in s) to fall off the Rotarod during the six trials (left) and the whole session (right). C Number of entries in closed and total arms were used as measure of locomotor activity (left), while anxiety levels were assessed by the percentage (%) of entries and time spent on open arms in an elevated plus maze (right). D Depression-like behavior was assessed by measuring the immobility time (in s) during a single session in the tail suspension test. N vehicles: males = 8–11 and females = 11–12; N Aβ1–42: males = 8–11 and females = 9–12; N reverse Aβ42–1: males = 7 and females = 6). Aβ, Amyloid-β; s, seconds

Thus, this data confirmed that the overall health status and locomotor function were uniform among the different treated groups and, therefore, all learning and memory impairments observed in this work were due to a specific hippocampal disruption caused by Aβ1–42.

Aβ 1–42 inhibits ex vivo LTP similarly in both female and male mice

Given the similar deleterious effects of Aβ1–42 on hippocampal-dependent learning and memory processes in both sexes in the present amyloidosis model, we wondered whether excitability, presynaptic function, and short- and long-term plasticity were affected, since they are the underlying physiological mechanisms of those cognitive capabilities. To pair it with the behavioral tasks, electrophysiological recordings were carried out 1–17 days post-icv. injection of Aβ1–42 or controls: Aβ42–1 and vehicle in a new cohort of mice.

Firstly, I/O curves in all groups (Fig. 6A–C) showed a greater amplitude of both the first (vehicle male, n = 7; vehicle female, n = 7; Aβ1–42 male, n = 8; Aβ1–42 female, n = 5; Aβ42–1 male, n = 5; Aβ42–1 female, n = 4; F(2.264,321.436) = 929.033, p < 0.001, Geisser-Greenhouse’s correction) and the second fEPSP (F(2.255,209.756) = 443.592, p < 0.001, Geisser-Greenhouse’s correction) with increasing intensities. No between-group differences were observed in the amplitude of the 1st fEPSP due to either sex (F(2.264,321.436) = 2.638, p = 0.066, Geisser-Greenhouse’s correction) or treatment (F(4.527,321.436) = 1.732, p = 0.134, Geisser-Greenhouse’s correction). However, a sex-treatment interaction effect was found in the amplitude of the 2nd fEPSP (F(4.511,209.756) = 2.813, p = 0.021, Geisser-Greenhouse’s correction), indicating that male animals injected with Aβ1–42 exhibited higher amplitudes evoked by the second pulse compared to the other sex-matched groups.

Fig. 6

Aβ1–42 inhibits ex vivo hippocampal LTP and induces LTD in both female and male mice. A–C I/O curve with paired fEPSPs collected at increasing stimulus intensities (from 0.075 to 0.4 mA) from control vehicle (A), Aβ1–42 (B) and Aβ42–1 reverse control (C) slices, respectively. Data is expressed as a percentage (%) of the maximum amplitude obtained. D PPF curve with paired fEPSPs collected at interstimulus intervals of 10, 20, 40, 100, 200 and 500 ms. Data is expressed as mean ± SEM amplitude of the second fEPSP as a percentage of the first [(second/first) × 100] for each inter-pulse interval used. E Representative averaged (n = 5) traces of fEPSPs recorded in the CA1 area, collected during the baseline (1) and ≈50 min post-HFS (2) in hippocampal slices from the different groups. F Time course of LTP evoked in the CA1 area after HFS in hippocampal slices from the different groups. Recordings were obtained from day 1 to 17 post-icv. injection. G, H Bars illustrate mean ± SEM fEPSPs amplitude of the last 10 min of the recording, to show acute (G; 24–48 h post-icv. injection) vs. long-term (H; 3–17 days post-icv. injection) effects on LTP. N (slices) vehicles: males = 6–5 and females = 5–5; N Aβ1–42: males = 7–7 and females = 6–7; N reverse Aβ42–1: males = 5–7 and females = 6–7. Aβ, amyloid-β; HFS, High frequency stimulation; LTP, long-term potentiation; mA, milliamperes; ms, milliseconds; min, minutes. ***p < 0.001 vs. vehicle of the corresponding sex; ###p < 0.001 vs. Aβ42–1 of the corresponding sex

Then, we addressed the short-term plasticity phenomenon, PPF. This protocol is also related to neurotransmitter release and, therefore, allowed us to evaluate the presynaptic functionality following Aβ1–42 injection. As shown in Fig. 6D, all groups presented an increased response to the second pulse when the intervals were short (20, 40 and 100 ms), since the ratio between the second and first EPSPs were above 100%, indicating enhanced neurotransmitter release. Nonetheless, neither treatment nor sex caused significant differences at any of the selected intervals (vehicle male, n = 6; vehicle female, n = 6; Aβ1–42 male, n = 7; Aβ1–42 female, n = 5; Aβ42–1 male, n = 3; Aβ42–1 female, n = 3; F(2,83) = 0.297, p = 0.744 and F(1,83) = 3.144, p = 0.08, respectively). This data indicated a normal short-term plasticity and presynaptic vesicle release after the treatment, suggesting that the alterations caused by Aβ1–42 injection may preferentially impact the postsynaptic level.

Finally, we measured the effect of Aβ1–42 injection on long-term synaptic plasticity applying an HFS protocol after a 15-min baseline from day 1 to 17 post-icv. injection (Fig. 6E). Data showed a significant treatment effect (Fig. 6F; vehicle male, n = 11; vehicle female, n = 10; Aβ1–42 male, n = 14; Aβ1–42 female, n = 11; Aβ42–1 male, n = 12; Aβ42–1 female, n = 13; F(2,207) = 87.616, p < 0.001). Post-hoc analysis revealed that the differences were specifically between the Aβ1–42 and the two control groups, regardless of sex (sex effect: F(1,207) = 0.092, p = 0.762). This inhibition of LTP was observed immediately after the HFS and persisted for at least 60 min afterwards. In fact, fEPSPs post-HFS were slightly below the BL for this group, suggesting that a protocol intended to induce LTP instead induced LTD. Importantly, the control reverse peptide did not affect LTP, indicating the specificity of Aβ1–42 peptide´s detrimental effect. Additionally, this study aimed to determine whether the detrimental effect of Aβ1–42 on LTP was an acute or long-term effect of icv. injection. The results demonstrated that Aβ1–42 disrupted LTP both in the short-term (24–48 h post-injection; treatment effect: F(2,96) = 62.30, p < 0.001; Fig. 6G) and long-term (3–17 days post-injection; treatment effect: F(2,105) = 49.94, p < 0.001; Fig. 6H), irrespective of sex. This finding confirms that icv. administration of Aβ1–42 has both acute and long-term effects on synaptic plasticity processes. Notably, even at 17 days post-injection, when Aβ levels were similar between Aβ1–42 treated and control mice based on our western blot results of amyloid peptide clearance (see Additional file 1 for details), the detrimental effects on synaptic plasticity and memory persisted.

Overall, these results indicate that a single Aβ1–42 injection impairs long-term synaptic plasticity, mainly at a postsynaptic level. This effect likely underlies the hippocampal-dependent learning and memory deficits observed in our study in both male and female animals.