Imipenem in the Rat Brain: A Multidimensional Study on Hippocampal Behavior, GABAergic System, Astrocyte Response, and Neurogenesis

TL;DR

This study examines the effects of imipenem on the rat hippocampus and finds no major behavioral or morphological changes, except for a localized increase in a specific neuronal population.

Contribution

The study uses clinically relevant doses and evaluates both behavioral and morphological effects in naïve rats, addressing gaps in prior animal research.

Findings

Imipenem did not cause significant memory or anxiety changes in rats.

Neurogenesis and astrogliosis were unaffected by imipenem treatment.

A significant increase in calbindin-immunoreactive neurons was observed in the hippocampal CA1 region.

Abstract

Background: After imipenem was introduced in clinical practice, neurologic adverse effects led to recommendations against its use in patients with neurologic conditions. However, these conclusions were drawn without considering pharmacokinetic variations in such patients. Furthermore, animal studies lack the use of clinically relevant doses and supporting morphological studies in both naïve and disease models. Objectives: We aim to study the effects of imipenem in the hippocampus of naïve animals, evaluating potential behavioral and morphological alterations. Methods: Naïve Wistar rats received a 10-day course of intraperitoneal imipenem (40 mg/kg) while controls received a saline injection. After that, they were put through the Morris water maze, elevated plus maze, open-field test, and then euthanized. We analyzed neurogenesis (using doublecortin immunoreactivity), astrogliosis, and the γ-Aminobutyric acid (GABA)ergic system (using parvalbumin (PV), calretinin (CR) and calbindin (CB) immunoreactive (IR) neurons) in the hippocampus. Results: Interestingly, our results showed no significant alterations in both short and long-term memory, nor in anxiety. There were also no significant changes in the neuronal density of doublecortin-immunoreactive (IR) neurons nor in astrogliosis. Furthermore, the areal density of PV- and CR-IR was preserved in all hippocampal subfields. The density of CB-IR neurons also showed no changes in the dentate gyrus, CA3, and subiculum; however, a significant increase was found in the CA1 region. Conclusions: Our results indicate that in naïve individuals, a clinically relevant dose of imipenem does not seem to cause overt behavioral deficits or widespread morphological alterations in the hippocampus. However, a specific increase in the CB-IR neuronal population in the CA1 region highlights a localized cellular alteration/plasticity induced by the imipenem. Hence, pharmacokinetic factors seem to be the potential contributors of imipenem side effects. Further studies should focus on this as a possible cause and focus on individuals with brain diseases.

Figures

Full text

1. Introduction

The introduction of antibiotics into clinical practice was one of the greatest breakthroughs of twentieth-century medicine [1]. However, according to the Global Antimicrobial Resistance and Use Surveillance System report from the World Health Organization (WHO), infections caused by resistant microorganisms are rising, and antimicrobial resistance is now considered one of the major global health problems [2]. Infections caused by antimicrobial-resistant microorganisms have been continuously associated with worse outcomes and higher mortality, posing a serious threat to global health [3,4].

The emergence of multidrug-resistant bacteria and guidelines in antimicrobial stewardship have reinforced the importance of carbapenems in clinical practice and highlighted the need for responsible antibiotic prescription [5,6]. Carbapenems are broad-spectrum antibiotics active against Gram-negative and Gram-positive bacteria [7]. Among them, imipenem, a carbapenem with a broad spectrum, is widely used in severe skin and soft tissue infections, respiratory infections, urinary tract infections, and sepsis [8]. Despite the importance of carbapenems, there is not yet enough information about their side effects on the brain. Some reports indicate potential neurotoxicity, with manifestations ranging from psychosis to convulsions [9]. Studies have shown that an increased serum concentration of beta-lactams is present in antibiotic-treated patients with neurotoxicity symptoms [10].

Though not totally clarified, the neurotoxic mechanisms of carbapenems and other β-lactams are thought to involve inhibition of the γ-Aminobutyric acid (GABA) system [11,12]. Previous animal studies showed an affinity of beta-lactams for the GABA-receptors and their potential competition with benzodiazepines treatment against epileptic convulsions [11]. Reduced inhibition of the GABAergic system enhances the central nervous system excitation and triggers epileptiform discharges [13].

The differences between proconvulsive effects of carbapenems themselves and other β-lactams are associated with the C-2 side chain, rather than the carbapenem core itself [14]. Animal studies from the 1980s revealed that carbapenems differ in their proconvulsive activity, with imipenem being the most proconvulsive [15,16]. However, those experiments used antibiotic doses higher than those applied in clinical practice [15,16]. Furthermore, meta-analysis studies have shown that carbapenem therapy has a higher risk of seizures when compared with other beta-lactams, particularly in patients with predisposing neurological conditions [17].

Given the need for further investigations to elucidate the effects of imipenem on the central nervous system and based on previous research using Wistar rats showing that a 40 mg/kg dose of imipenem can impair hippocampal and cerebellar function in juvenile animals [18], we selected this model to study the drug’s impact on the rat brain. Specifically, we focused on exploring the imipenem impacts on spatial learning and memory, anxiety, and locomotor activity. Furthermore, we evaluated if imipenem effects on neurogenesis using doublecortin (DCX), neuroinflammation using glial fibrillary acidic protein-immunoreactive (GFAP-IR) astrocytes, and the GABAergic system (looking for the calcium-binding proteins (CBPs) parvalbumin (PV), calretinin (CR), and calbindin (CB) [18].

2. Results

2.1. Morris Water Maze

The Morris water maze (MWM) is commonly used to evaluate spatial learning and memory in rodents. To investigate whether imipenem administration affected cognitive performance, animals were tested using both the reference memory and working memory paradigms of the MWM.

2.1.1. MWM—Reference Memory

Long-term spatial learning was examined using the reference memory version of the MWM. As shown in Figure 1A, the mean distance traveled to reach the hidden platform progressively decreased across the 7-day acquisition period. Repeated-measures ANOVA indicated a significant effect of training day, demonstrating learning across sessions (F (3.022, 66.48) = 27.32, p < 0.0001). In contrast, neither a main effect of treatment (F (1, 22) = 0.5072, n.s.) nor a treatment × trial block interaction (F (12, 162) = 1.1, n.s.) was observed. These findings indicate that spatial acquisition was comparable between imipenem-treated rats and controls.

2.1.2. MWM—Reference Memory: Probe Test

Memory retention was assessed during the probe test, with results summarized in Figure 1B,C. Two-way ANOVA revealed a significant effect of quadrant (F (2, 44) = 43.17, p < 0.0001), while no main effect of treatment (F (1, 22) = 0.01715, n.s.) or treatment × quadrant interaction (F (2, 44) = 0.9500, n.s.) was detected (Figure 1C). Animals from all experimental groups spent significantly more time in the target quadrant compared to the opposite quadrant, and no between-group differences were observed in the time spent searching the target area. Similarly, the number of crossings over the former platform location did not differ between groups, as confirmed by an unpaired t-test (n.s.; Figure 1B).

All groups rapidly located the visible platform, and no significant group differences were detected (F (2, 27) = 0.3, n.s.), indicating intact sensorimotor and visual abilities.

2.1.3. MWM—Working Memory

Short-term spatial memory was evaluated using the working memory version of the MWM. Figure 1D presents the mean distance traveled during the information (trial 1) and retention (trial 2) phases. Repeated-measures ANOVA showed a significant effect of trial block (F (1, 22) = 23.79, p < 0.0001), reflecting improved performance from trial 1 to trial 2. However, no main effect of treatment (F (1, 22) = 0.9474, n.s.) or treatment × trial block interaction (F (1, 22) = 0.3591, n.s.) was found, indicating that working memory performance was similar between imipenem-treated and control rats.

2.2. Locomotor Activity

General locomotor activity was evaluated using the open-field and elevated plus-maze tests (Figure 1F,H). In the open-field, the total distance traveled was 2029 ± 61 cm for control animals and 2059 ± 67 cm for imipenem-treated rats. In the elevated plus maze, total distances were 1497 ± 56 cm and 1454 ± 58 cm for control and imipenem-treated groups, respectively. Unpaired t-tests revealed no significant differences between groups in either the open-field (n.s.) or the elevated plus maze (n.s.), indicating that imipenem treatment did not affect locomotor or exploratory activity.

2.3. Open-Field Test

The open-field paradigm was further used to assess exploratory behavior and anxiety-related responses. As illustrated in Figure 1E, two-way ANOVA showed a significant main effect of zone (F (1, 44) = 629.8, p < 0.0001), whereas neither a treatment effect (F (1, 44) = 1.777 × 10^−5^, n.s.) nor a treatment × zone interaction (F (1, 44) = 1.427, n.s.) was observed. Both control and imipenem-treated rats spent more time in the peripheral zone than in the center of the arena (p < 0.0001), with no differences between groups in either zone.

2.4. Elevated Plus Maze

Anxiety-like behavior was further evaluated using the elevated plus maze. ANOVA revealed a significant effect of zone (F (2, 66) = 61.06, p < 0.0001), but no main effect of treatment (F (1, 66) = 9.394 × 10^−5^, n.s.) and no significant treatment × zone interaction (F (2, 66) = 3.470, n.s.). As shown in Figure 1G, both experimental groups spent similar amounts of time in the open arms, closed arms, and central area of the maze.

2.5. Neurogenesis

Estimates of the areal density of doublecortin-immunoreactive (DCX-IR) cells within the subgranular zone of the hippocampal formation are presented in Figure 2A–C. Statistical comparison using an unpaired t-test indicated that imipenem administration did not produce a significant change in DCX-IR neuronal density in the hippocampus (n.s.).

2.6. Astrogliosis

To evaluate whether imipenem treatment elicited astrocytic activation or neuroinflammatory alterations, astrocytes were examined in five distinct hippocampal subregions. The distribution profiles of astrocytic process lengths are illustrated in Figure 2D–H. Analysis of variance demonstrated a significant effect of process-length distribution in all regions analyzed, including the dentate gyrus (F (2.007, 20.07) = 1706, p < 0.0001), dentate hilus (F (1.916, 19.16) = 549.7, p < 0.0001), CA3 (F (1.094, 10.94) = 95.62, p < 0.0001), CA1 (F (1.100, 11.00) = 83.06, p < 0.0001), and subiculum (F (1.419, 12.77) = 201.6, p < 0.0001).

In contrast, no main effect of treatment was detected for astrocytic process length in any hippocampal area, including the dentate gyrus (F (1, 10) = 3.051, n.s.), dentate hilus (F (1, 10) = 2.185, n.s.), CA3 (F (1, 10) = 0.1341, n.s.), CA1 (F (1, 10) = 0.5611, n.s.), or subiculum (F (1, 9) = 0.3328, n.s.). Furthermore, no significant interactions between treatment and process-length distribution were observed in any region analyzed: dentate gyrus (F (2.007, 20.07) = 0.6602, n.s.), dentate hilus (F (1.916, 19.16) = 1.505, n.s.), CA3 (F (1.094, 10.94) = 0.2055, n.s.), CA1 (F (1.100, 11.00) = 0.6455, n.s.), and subiculum (F (1.419, 12.77) = 0.3148, n.s.).

In addition to process length, astrocyte areal density was quantified across the same hippocampal regions (Figure 2J). Unpaired t-test analyses revealed no significant differences between control and imipenem-treated animals in astrocyte density within the dentate gyrus molecular layer (n.s.), dentate hilus (n.s.), CA3 (n.s.), CA1 (n.s.), or subiculum (n.s.).

Astrocyte morphology was further characterized by assessing the number of processes per cell (Figure 2I). No significant effects of imipenem treatment were observed in any hippocampal subregion, including the dentate gyrus molecular layer (n.s.), dentate hilus (n.s.), CA3 (n.s.), CA1 (n.s.), or subiculum (n.s.).

2.7. GABAergic System—PV-, CR- and CB-Immunoreactive (IR) Neurons

2.7.1. PV-IRNeurons

Estimates of PV-IR neuron density within the dentate hilus, CA3, CA1, and subiculum are summarized in Figure 3A–C. Statistical comparisons using unpaired t-tests revealed no significant differences between control and imipenem-treated groups in the dentate hilus (n.s.), CA3 (n.s.), CA1 (n.s.), or subiculum (n.s.).

2.7.2. CR-IR Neurons

The areal density of CR-IR neurons across the dentate hilus, CA3, CA1, and subiculum is shown in Figure 3D–F. Unpaired t-test analyses indicated that imipenem administration did not significantly alter CR-IR neuronal density in any of the hippocampal regions examined, including the dentate hilus (n.s.), CA3 (n.s.), CA1 (n.s.), and subiculum (n.s.).

2.7.3. CB-IR Neurons

Quantitative analysis of CB-IR neurons in the hippocampal formation is presented in Figure 3G–I. Unpaired t-tests showed no significant effect of imipenem treatment on CB-IR neuronal density in the dentate hilus (n.s.), CA3 (n.s.), or subiculum (n.s.). In contrast, a significant increase in CB-IR neuron density was detected in the CA1 region of imipenem-treated animals compared with controls (p < 0.05).

3. Discussion

Our results show that imipenem, at a clinically relevant dose, does not induce significant alterations in spatial learning and memory and anxiety. While the treatment did not cause widespread structural alterations in the hippocampal formation, a specific increase in the density of calbindin-expressing (CB+) neurons was noted in the CA1 region. The absence of differences in the density of DCX-IR neurons, astrocytes, parvalbumin-expressing (PV+) and calretinin-expressing (CR+) GABAergic neurons in the hippocampus, along with the overall lack of behavioral changes, supports the conclusion that imipenem, at a clinically relevant dose, does not exert widespread neurotoxic effects. Nevertheless, these results do not diminish the importance of monitoring the potential neurological side effects of imipenem, particularly in patients with associated risk factors such as renal insufficiency, history of seizures, or central nervous system diseases.

3.1. Locomotor Activity, Anxiety, and Memory

In this study, repeated daily administration of imipenem (40 mg/kg, intraperitoneal for 10 days) in adult male rats did not significantly alter locomotor activity in the open field test nor anxiety-related behavior in the open field or elevated plus maze. Furthermore, performance in the MWM revealed no impairments in spatial learning, reference memory, or working memory. Taken together, these findings indicate that subchronic exposure to imipenem at this dose does not induce detectable alterations in locomotion, anxiety-like behavior, or hippocampal-dependent cognitive function in adult rats.

These results contrast with prior reports showing behavioral effects of imipenem in juvenile rodents. Golchin et al. (2013) demonstrated that imipenem exposure during the developmental stage impaired motor coordination and memory and altered exploratory behavior, with effects influenced by dose and sex [18]. However, it is important to note that they did not also find alterations in learning and memory when young male rats were treated with 40 mg/kg, suggesting that the effect of imipenem in the brain is dose dependent. The absence of comparable effects in the present study highlights the importance of developmental stage: adult animals may be more resilient to neurobehavioral consequences of β-lactam antibiotics than younger cohorts, likely due to the maturation of hippocampal and cerebellar circuits.

Additional support for developmental sensitivity comes from studies using other antibiotics. Ceylani et al. (2018) showed that repeated antibiotic exposure in juvenile mice led to long-lasting increases in anxiety and deficits in memory when tested in adulthood [19]. Similarly, Leclercq et al. (2017) demonstrated that early-life exposure to low-dose penicillin altered gut microbiota composition and brain cytokines, resulting in behavioral changes [20]. These findings suggest that antibiotics may affect cognition and emotion, particularly when administered during critical windows of neurodevelopment. In contrast, adult microbiota and neural systems appear more stable, explaining the absence of behavioral or cognitive changes in our study despite repeated imipenem administration. Clinically, imipenem is associated with neurotoxic side effects such as seizures, particularly at high doses or in vulnerable patients, but cognitive deficits are rarely reported. Toxicological studies in adult rodents also indicate that spontaneous locomotor activity and learning are unaffected even at higher doses than those used here [21]. Our results are therefore consistent with the available preclinical and clinical evidence that imipenem’s neurobehavioral-related effects are limited and context dependent.

Despite the significant behavioral and morphological findings, some limitations of the present study should be acknowledged. First, direct pharmacokinetic parameters, such as imipenem concentrations in the plasma and CSF, were not quantified. However, the observed astrocytic alterations and behavioral deficits provide functional evidence that the drug reached the central nervous system at the selected dosage. Furthermore, the use of a non-invasive protocol was prioritized to ensure the reliability of the behavioral outcomes and to adhere to the 3Rs principles of animal welfare, relying on the drug’s pharmacokinetic profile already validated in the literature [18,22]. Future studies incorporating satellite groups for time-course drug quantification could provide further insights into the correlation between peak plasma levels and the severity of astrocytic response.

In summary, repeated administration of imipenem at 40 mg/kg in adult male rats does not impact locomotion, anxiety-like behavior, or hippocampal-dependent learning and memory. These findings suggest that imipenem is behaviorally neutral under the conditions evaluated, and they contribute to defining its neurobehavioral safety profile. Future studies should address potential effects of higher doses, prolonged exposure, or administration in developmentally or clinically vulnerable populations, as well as the possible contribution of microbiota changes to antibiotic-induced behavioral outcomes.

The absence of significant impairment of anxiety, learning, and memory observed in imipenem-treated animals immediately indicates that there were no major alterations in the hippocampus, a fact that was confirmed by the morphological analysis.

3.2. Neurogenesis

In the present study, we have found that 40 mg/kg imipenem treatment did not induce significant alterations in the adult hippocampal neurogenesis and behavior in a rat model. Our primary findings reveal no significant alteration in the density of doublecortin-expressing (DCX+) cells in the dentate gyrus, a marker for immature neurons and adult hippocampal neurogenesis [23]. Crucially, these structural results were mirrored by the behavioral data, as the treated animals displayed no observed deficits in locomotor activity, anxiety-like behavior, or performance in learning and memory tasks. This composite outcome—the absence of both structural and functional impairment—is significant as it challenges and refines the existing literature on imipenem’s neurotoxicity. Imipenem and its co-administered inhibitor, cilastatin, are known to readily cross the blood–brain barrier and have been implicated in central nervous system adverse effects, most notably seizures, mediated primarily through GABAA receptor antagonism [24]. Furthermore, some studies have previously associated imipenem administration with cognitive deficits in rats [18]. Our findings suggest that the specific regimen employed in this study—40 mg/kg imipenem for 10 days—did not reach a neurotoxic threshold sufficient to compromise the neural circuits underlying these fundamental behaviors. The consistency between the preserved DCX+ cell population and normal behavioral performance is highly informative. Adult neurogenesis is considered a critical process for specific aspects of hippocampal function, including pattern separation and mood regulation [25]. The structural integrity of the DCX+ population corroborates our behavioral observations, suggesting that the rate of neuroblast differentiation and survival was not acutely inhibited. Had subclinical neuroinflammation or oxidative stress (mechanisms previously linked to carbapenem toxicity) occurred to a degree capable of affecting neuronal function, a disruption in the overly sensitive neurogenic process might have been expected. Thus, the stability of the DCX+ population suggests that any minor cellular stress induced by the drug was effectively compensated by endogenous homeostatic mechanisms.

3.3. Astrogliosis and Neuroinflammation

GFAP is a major cytoskeletal protein of astrocytes, often increased in cases of neurotoxicity, be it caused by brain injury or disease [26]. Interestingly, our results show no differences in either the astrocyte density or the number of branches or their length. Previous studies showed that in inflammatory and degenerative models, astrocytes acquire morphological alterations characteristic of a reactive response, the branches increase in number and length [27]. The lack of evidence of neuronal degeneration and a reactive response from the astrocytes suggests that the treatment with imipenem in a 40 mg/kg dose does not cause enough neurotoxic effects with detectable structural repercussions. Additionally, epilepsy models (both kindling and temporal lobe epilepsy) have shown that seizures are associated with an increase in GFAP expression and astrogliosis [28,29]; therefore, our results show that imipenem in the doses used apparently does not cause alterations consistent with acute isolated epileptic phenomena nor long-term effects. The lack of differences in our results also indicates that the interface between the gut biome and neuroinflammation is not altered. Studies with other antibiotics have shown that broad-spectrum antibiotics interfere with the gut biome and reduce neuroinflammation in cases of brain injury [30,31]. In this study, however, the antibiotics used were administered by gavage and can only affect the microbiome and do not interfere with the nervous system by themselves as they are not absorbed in relevant doses through this route. Imipenem is not absorbed efficiently by oral administration and is mostly excreted in the urine and in an inactive form; hence, an intraperitoneal administration as we performed is highly unlikely to have any significant effect on gut microbiome and neuroinflammation [32,33]. Nonetheless, functional alterations [34] can also occur that were not detected by our analysis. Additionally, the sample size (6 animals) and the treatment duration pose a limitation to our study, as long-term exposure could eventually cause alterations as seen with kindling models using pentylenetetrazol [35]. Imipenem interferes with the same receptors as pentylenetetrazol; consequently, alterations caused by a longer period of treatment with imipenem, such as an increase in astrogliosis and in astrocytes with a reactive phenotype, could eventually occur [35].

3.4. Calcium-Binding Proteins—Parvalbumin, Calretinin, and Calbindin

Given the possible interactions of imipenem with the GABAergic system, described as the main mechanism for its neurological adverse effects, we studied its effects on the neuronal density of GABAergic neurons [13]. The GABAergic population is very diverse and heterogeneous and so, in the present study, we focused on the calcium-binding proteins-immunoreactive GABAergic neurons, which represent a large group of the GABAergic population [36], namely PV, CR and CB. Interestingly, our results demonstrate that repeated systemic administration of imipenem did not alter the density of hippocampal neurons expressing the calcium-binding proteins parvalbumin and calretinin, nor the CB+ population in the dentate hilus, CA3, and subiculum. However, a significant increase in calbindin was observed in the CA1 region. These proteins are expressed in distinct populations of GABAergic interneurons and excitatory principal cells, where they play crucial roles in regulating hippocampal excitability, oscillatory activity, and synaptic integration [36,37]. The preservation of most of these neuronal populations suggests that the imipenem treatment regimen used in the present study does not disrupt the widespread inhibitory–excitatory balance in the hippocampus at the cellular level and did not induce relevant global changes in the morphology or function of GABAergic neurons expressing parvalbumin, calretinin, and calbindin in most regions. However, the specific increase the density of CB-IR neurons in CA1 indicates that imipenem induced a relevant, regionally specific, plastic change in this particular population. CB+ neurons contribute to interneuron–interneuron inhibition, dendritic integration, and calcium buffering in principal neurons [38]. The increase in CB-IR neurons in CA1, a region critical for memory formation and excitability, warrants attention and may represent a compensatory or reactive cellular change induced by imipenem in subjects without neurological conditions or other systemic conditions that may alter imipenem concentrations in the blood and brain tissue. Clinical data support our results, as convulsions caused by imipenem are more common in individuals with epilepsy or kidney injury [17,39]. Although the sample size (n = 6 per group) is not large, the low intra-group variability clearly supports the results. The 10 days of exposure to imipenem of our test animals enable us from excluding the possibility of effects caused by prolonged or repetitive exposure to imipenem. The duration of treatment, nonetheless, was chosen to mimic some of the most clinically relevant imipenem administration timelines used for the treatment of infections caused by Gram-negative bacteria [40]. Functional or biochemical impairments resulting from the administration of this drug, such as malfunctions in synaptic efficacy or neurotransmitter release, can also occur and not be detected by our analysis, as they might occur in the absence of morphological changes [41]. It is also important to consider that subtle or region-specific changes in GABAergic neurons may not have been detected by this methodology due to its limitations in sensitivity. It is also noteworthy that the lack of alterations in Wistar rats may not reflect the effects the consequences of the same treatment in humans, as there is a gap in knowledge between the expression of these proteins in humans and rats caused by the lack of uniform characteristics and analysis of samples from the human hippocampus [42].

Positive parvalbumin interneurons are particularly important for network synchrony and gamma oscillations that underlie higher-order cognitive processes such as attention, learning, and memory [43,44]. Alterations in PV+ interneurons have been consistently reported in neuropsychiatric and neurodegenerative disorders, including schizophrenia [45], epilepsy [46], and Alzheimer’s disease [47]. Similarly, CR+ and CB+ neurons contribute to interneuron–interneuron inhibition, dendritic integration, and calcium buffering in principal neurons [38]. Reductions in CB and CR expression have been associated with excitotoxicity, impaired plasticity, and increased neuronal vulnerability in disease states [48,49]. The absence of changes in the PV+ and CR+ populations in our study aligns with our findings showing no imipenem-induced alterations in spatial learning, anxiety-like behaviors, or neurogenesis. The specific increase in the CB+ population in CA1 is, however, a critical finding that warrants further investigation, despite the lack of corresponding behavioral deficits in this study. Together, these results reinforce the conclusion that imipenem does not produce widespread overt neurotoxic or neuromodulator effects under our experimental conditions.

It is well known that β-lactam antibiotics, including imipenem, can be associated with central nervous system side effects such as seizures, especially at high doses or in susceptible patients [50,51]. These adverse effects are thought to arise partly from interference with GABAergic neurotransmission [52]. However, our findings that PV+ and CR+ hippocampal neurons were preserved (alongside the CB+ population in most studied regions) suggest that, at the clinical dose and treatment duration applied here, imipenem does not compromise the overall inhibitory interneuron populations or general hippocampal calcium-buffering mechanisms. The localized increase in CB+ neurons in CA1, however, indicates a specific plastic or reactive cellular change that could be a latent factor for functional deficits under higher doses or longer exposure. This may explain why no overt behavioral or neuroplastic alterations were detected in the present study.

4. Materials and Methods

4.1. Animals and Drug Administrations

Male Wistar rats were obtained from the Institute for Molecular and Cell Biology (Porto, Portugal). Animals were maintained under controlled environmental conditions, including a temperature of 23 ± 1 °C, relative humidity of 45 ± 5%, and a 12 h light/12 h dark cycle (lights on at 07:00 h and off at 19:00 h). Standard laboratory chow and water were provided ad libitum throughout the study.

A total of 24 four-month-old male rats were used. Animals were randomly assigned to either the imipenem-treated group or the control group. Both groups received one daily intraperitoneal injection for 10 consecutive days. Rats in the imipenem group were administered imipenem at a dose of 40 mg/kg, prepared at a concentration of 10 mg/mL, whereas control animals received an equivalent volume of saline solution. The dose of 40 mg/kg of imipenem was selected based on its established neurotoxic profile in rodent models, as previously described by Golchin et al. [18]. This dosage has been shown to achieve clinically relevant systemic exposure and cross the blood–brain barrier effectively in rats. To avoid potential confounding factors such as physiological stress or neuroinflammation associated with invasive sampling—which could interfere with the subsequent behavioral assessments—no direct pharmacokinetic measurements (plasma or CSF) were performed in this cohort [53].

4.2. Behavioral Testing

After treatment, rats were submitted to different behavioral tests. All behavioral tests were performed, with 1-day inter-test intervals, in the following order: Morris water maze (spatial reference memory task, spatial working memory), open-field and elevated plus-maze. Before testing, animals were handled for 5 consecutive days. Tests were performed after at least 30 min of habituation of the rats to the testing room. Tests were done by two experimenters blinded to the treatments performed during the standard light phase, starting at 12:00 h, after at least 30 min of habituation of the rats to the testing room.

4.2.1. Morris Water Maze (MWM)

Spatial Reference Memory Task

Spatial learning and memory were evaluated using a circular water maze consisting of a black pool (180 cm in diameter, 50 cm in depth) filled with water maintained at 21 ± 1 °C. The pool was virtually divided into four equal quadrants and positioned in a room containing distal visual cues. A black circular escape platform (10 cm diameter) was submerged 2 cm below the water surface in one quadrant. Animal trajectories were recorded using a computerized video-tracking system (EthoVision XT 8.5, Noldus, Wageningen, The Netherlands).

During the place learning task, rats were trained to locate and mount the hidden platform. Acquisition consisted of fours trial per day for 7 consecutive days. At the beginning of each trial, animals were released into the pool facing the wall from one of four starting locations, presented in a pseudo-random sequence such that each position was used once per block of four trials. Trials ended when the rat reached the platform or after 60 s had elapsed; animals that failed to locate the platform within this time were gently guided to it and allowed to remain there for 15 s. Between trials, rats were placed in a clean holding cage for 30 s. The platform position remained constant throughout the acquisition phase. Swim path length was recorded for each trial.

Twenty-four hours after the final acquisition trial, memory retention was assessed using a 60 s probe trial in which the platform was removed. The number of crossings over the former platform location and the time spent in the target quadrant were quantified. On the following day, sensorimotor performance was evaluated using a visible platform task. Rats completed one block of four trials with 30 s inter-trial intervals. In this task, a white platform protruding 3 cm above the water surface was used, and its position was changed between trials. The distance traveled to reach the platform was measured and averaged across trials.

Spatial Working Memory

Spatial working memory was evaluated using a delayed match-to-place variant of the Morris water maze, conducted three days after completion of the spatial reference memory task. Animals performed two trials per day over six consecutive days. During the first trial (information trial), the hidden platform was placed in a novel location each day, differing in both quadrant and distance from the pool wall relative to previous sessions. Rats were released into the maze facing the wall from a starting position opposite the platform. If the platform was not located within 60 s, the animal was gently guided to it and allowed to remain there for 15 s.

The second trial (retention trial) was performed with the platform positioned in the same location as in the corresponding information trial, with the starting point again located distal to the platform. All other procedural conditions matched those used in the spatial reference memory task. A 1 min interval separated the information and retention trials.

4.2.2. Open-Field Test

To assess general exploratory locomotion and anxiety-like behaviors, we used an open-field apparatus that consisted of a white acrylic arena (100 cm × 100 cm × 40 cm). The test rat was placed in a corner of the apparatus and tested during 5 min sessions. Distances traveled in the outer zone of the open field, defined as 20 cm from any wall, and in its inner zone, defined as the 60 cm × 60 cm square in the center of the arena, were measured using a computerized video-tracking system (EthoVision XT 8.5, Noldus, The Netherlands). At the end of each session, the number of fecal boli deposited was counted and recorded, and the urine deposited was collected using a filter paper. The difference between the weight (in g) of the paper before and after collecting the urine was considered as a measure of the amount of urine deposited during the session. The floor of the apparatus was thoroughly cleaned and dried between each session.

4.2.3. Elevated Plus-Maze

To further evaluate general exploratory and anxiety-like behaviors, an elevated plus-maze apparatus consisting of a black acrylic cross with 2 opposite open and 2 opposite closed arms (50 cm × 12 cm) joined by a common central square (12 cm × 12 cm) was used. The closed arms were enclosed by 50 cm high walls. The test rat was placed on the central square, facing one of the closed arms, and allowed to explore the apparatus for 5 min. The behavior of the rat was recorded and analyzed using a computerized video-tracking system (EthoVision XT 8.5, Noldus, The Netherlands). The percentages of time spent, and the distances traveled by rats in the open arms, in the closed arms, and in the central square were calculated. At the end of each session, the number of fecal boli and the amount of urine were recorded. The apparatus was then thoroughly cleaned and dried.

4.3. Tissue Collection and Immunocytochemistry

Upon completion of the behavioral experiments, animals (six per experimental group) were randomly selected and deeply anesthetized with sevoflurane (SevoFlo, Abbott Laboratories Ltd., Maidenhead, UK). Rats were then transcardially perfused with 150 mL of 0.1 M phosphate buffer, followed by fixation using 4% paraformaldehyde prepared in phosphate buffer (pH 7.6). Brains were removed, coded to allow blinded processing and analysis, and divided into right and left hemispheres by a midsagittal section. The frontal and occipital poles were discarded, and the remaining tissue blocks containing the hippocampal formation were isolated for subsequent immunohistochemical procedures. To account for previously reported hemispheric asymmetries in the rodent hippocampus, tissue blocks were alternately collected from the right and left hemispheres [54]. Tissue blocks containing the hippocampal formation were post-fixed for 1 h in the same fixative used during perfusion and subsequently cryoprotected overnight in 10% sucrose at 4 °C. Blocks were then mounted on a vibratome and serially cut in the coronal plane into 40 µm-thick sections, which were collected in phosphate-buffered saline (PBS). For each brain, two series of hippocampal sections were selected using a systematic random sampling method. The initial section was randomly chosen from the first 12 collected sections, and additional sections were obtained at regular 480 µm intervals (i.e., every 12th section) along the septotemporal axis of the hippocampal formation. The remaining animals were also deeply anesthetized with sevoflurane and euthanized by decapitation. Immediately following euthanasia, blood samples were collected via cardiac puncture. The samples were centrifuged at 3000 rpm for 10 min at 4 °C, and the resulting plasma was aliquoted and stored at −80 °C for future biochemical analyses.

Immunohistochemical analysis was performed using antibodies directed against DCX, GFAP, PV, CR, and CB. Free-floating sections were washed twice in PBS, incubated for 10 min in 3% hydrogen peroxide to quench endogenous peroxidase activity, and then incubated overnight at 4 °C with the appropriate primary antibody: DCX (Santa Cruz Biotechnology, Dallas, TX, USA, cat. no. sc-271390; 1:500), GFAP (Dako, Carpinteria, CA, USA, cat. no. Z0334; 1:2000), PV (Swant, Geneva, Switzerland, cat. no. PV235; 1:2000), CR (Swant, Switzerland, cat. no. CR7697; 1:2000), or CB (Swant, Switzerland, cat. no. CB38; 1:2000). All primary antibodies were diluted in PBS containing Triton X-100.

After primary antibody incubation, sections were rinsed in PBS and incubated with the corresponding biotinylated secondary antibody (Vector Laboratories, Newark, CA, USA; 1:400). This was followed by incubation with an avidin–biotin peroxidase complex (Vectastain Elite ABC Kit, Vector Laboratories, Newark, CA, USA; 1:800), with both steps performed for at least 1 h at room temperature. Immunoreactivity was visualized using 0.05% diaminobenzidine (Sigma, St. Louis, MO, USA) supplemented with 0.01% hydrogen peroxide. Sections were washed in PBS for a minimum of 15 min between each incubation step. To enhance antibody penetration, 0.5% Triton X-100 was included in all incubation and wash solutions.

Immunostaining specificity was verified by omission of the primary antibody, which resulted in the absence of detectable labeling. All immunohistochemical reactions were conducted simultaneously in 12-well culture plates, with four sections per well, to ensure identical processing conditions across experimental groups. Following staining, sections were mounted on gelatin-coated slides, air-dried, dehydrated through graded ethanol solutions (50%, 70%, 90%, and 100%), and coverslipped using Histomount mounting medium (National Diagnostics, Atlanta, GA, USA).

4.4. Morphometric Analysis

4.4.1. Quantification of Areal Density of DCX-, PV-, CR-, and CB-IR Cells

Estimates for areal densities were obtained by analysis and drawing using a light microscope equipped with a camera lucida, at a final magnification of x160. Cells were considered immunoreactive when they displayed darkly stained perikarya. Immunoreactive cells against DCX were only analyzed in the subgranular layer of the dentate gyrus, while analysis of immunoreactive neurons against PV, CR, and CB was performed in the different regions of the hippocampal formation, dentate gyrus, CA1, CA3, and subiculum. The subgranular layer was consistently delineated at all levels along the septotemporal axis of the hippocampal formation based on cytoarchitectonic criteria, and by using a rat brain atlas; this layer was further defined as an approximately 30 µm thick ribbon of tissue between the granular layer and the hilus, delineation of the dentate gyrus, CA1, CA3, and subiculum followed the same cytoarchitectonic criteria described above [55,56].

The estimates for each cell type were obtained from an average of 12 immunostained sections per rat sampled as described above. The calculation of layer areas was made using drawings obtained from a camera lucida. A transparent sheet bearing a test system composed of a set of regularly spaced points was overlaid on the drawings, and the number of points that fell within the limits of the layer was counted. Given the different sizes of the areas analyzed, different grids were used for DCX and the remaining markers. The area of each layer was estimated by multiplying the number of points that fell within limits by the value of the area per point of the test system (0.0069 mm^2^ for DCX and 0.015625 mm^2^ for the other markers). The cell numbers obtained were divided by the values of the corresponding laminar areas to yield areal density values (number of cells/mm^2^).

4.4.2. Quantification of Astrocyte Morphology

Previously immunostained brain sections were imaged at a final magnification of 20× using a Zeiss Scope A.1 light microscope fitted with an AxioCam MRc5 digital camera. The limits and laminar organization of the dentate gyrus, CA3, CA1, and subiculum were delineated based on cytoarchitectural landmarks along the septotemporal axis of the hippocampal formation, with reference to a standard rat brain atlas [55,56]. Astrocytes located within the CA2 region were grouped with the CA3 area for analytical purposes. Images acquired using AxioVision Rel. 4.8 software were subsequently analyzed in ImageJ 2.16.0/1.54p using a customized macro to quantify GFAP-immunoreactive astrocytes and to measure astrocytic soma size and process length. [57].

4.5. Data Analysis

Behavioral data are presented as mean ± SEM, whereas morphometric measurements are reported as mean ± SD. All statistical analyses and figure preparation were conducted using GraphPad Prism 10 software (GraphPad Software, La Jolla, CA, USA). Repeated-measures ANOVA was applied to evaluate performance variables including average time, distance traveled, and escape latency in the Morris water maze, as well as distance traveled in the spatial working memory task. The effects of treatment and arena zone in the open-field and elevated plus-maze tests were assessed using two-way ANOVA. Outcomes from the remaining behavioral assessments, along with the areal density of DCX-, CB-, CR-, and PV-immunoreactive neurons and the number and morphological characteristics of astrocytes, were analyzed using an independent-samples Student’s t-test, with treatment condition (imipenem versus saline) as the independent factor. Assumptions of data normality and homogeneity of variance were verified prior to analysis using appropriate tests, including the Shapiro–Wilk and Levene’s tests. Statistical significance was defined as p < 0.05 for all analyses.

5. Conclusions

Taken together, our findings show that subchronic administration of imipenem (40 mg/kg, 10 days) in adult rats does not produce detectable alterations at the behavioral or molecular level within the hippocampus. The absence of changes in locomotor activity, anxiety-like behavior, and spatial learning and memory demonstrates that the treatment did not impair fundamental cognitive or affective domains. Consistently, the lack of effects on neurogenesis and neuroinflammation suggests that hippocampal plasticity and glial reactivity were not compromised. Furthermore, while the PV+ and CR+ populations were preserved, the observed increase in CB+ neurons specifically in the CA1 region represents a cellular alteration that warrants careful consideration. This localized effect suggests that while the broader inhibitory and excitatory circuits remained largely intact, a plastic or reactive change occurred within the critical CA1 region.

By converging behavioral outcomes with neurobiological markers, these results support the conclusion that this imipenem treatment paradigm, like that used in clinical practice, does not cause widespread adverse morphological or functional effects on hippocampus in healthy adult rats. The relative preservation of PV+ and CR+ neurons, coupled with the absence of behavioral deficits, strengthens the evidence for the relative neural safety of repeated imipenem exposure under the present conditions. Nevertheless, given that β-lactam antibiotics are known to exert neurotoxic effects under specific circumstances, future studies should extend this work by examining different dosing regimens, prolonged treatments, and vulnerable developmental or aging stages, where this specific cellular increase in CB+ expression in CA1 might emerge as a more significant or functionally relevant deficit.