There is compelling evidence that alterations of glutamatergic neurotransmission are implicated in neuropsychiatric disorders (Sanacora et al., 2012; Uno and Coyle, 2019). In particular, the incorporation of the L-alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate (AMPA) glutamate receptor subunit 1 (GluA1) into synaptic AMPA receptors is considered to be a key mechanism of synaptic structural as well as functional plasticity (Sprengel et al., 1998; Hayashi et al., 2000; Shi et al., 2001; Malinow and Malenka, 2002; Kopec et al., 2007; Lee et al., 2010; Huganir and Nicoll, 2013). In humans the AMPA receptor subunit GluA1 is encoded by the GRIA1 gene, a locus with genome-wide association to schizophrenia (Ripke et al., 2013; Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2014). Gria1 deficient mice, harboring the global loss of GluA1, mimic multiple features of neuropsychiatric disease states related to glutamatergic dysfunction, among others, schizophrenia and schizoaffective disorders, attention deficit hyperactivity disorders (ADHD), bipolar disorders, and mood disorders (Wiedholz et al., 2008; Fitzgerald et al., 2010; Ben Abdallah et al., 2011; Barkus et al., 2012; Sanacora et al., 2012; Vogt et al., 2014). Data from adult mice suggest that the global loss of GluA1 induces a complex behavioral phenotype, comprising spatial working memory deficits with an intact reference memory (Zamanillo et al., 1999; Reisel et al., 2002; Schmitt et al., 2003; Bannerman et al., 2004), novelty-induced hyperexcitability (Wiedholz et al., 2008) and hyperlocomotion (Vekovischeva et al., 2001; Schmitt et al., 2003; Wiedholz et al., 2008), as well as a depression-associated phenotype (Chourbaji et al., 2008; Austen et al., 2017). While the behavioral phenotype of adult GluA1 deficient mice has been studied extensively, there is still limited information on the course and the onset of behavioral modifications during post-weaning development. Here, we aimed to profile aspects of the behavioral phenotype of young Gria1 knockout mice as a genetic loss-of-function model, exposing female and male Gria1^–/– mice to a home-cage based testing battery during post-weaning development. The applied behavioral and biochemical composite measure scheme was derived from an extensive dataset of consortium data for the purpose of severity assessment in different adult rodent models (Bleich et al., 2020). Our group has further figured out the parameters rendering the highest informative value based on a bioinformatic approach (Keubler et al., 2020; van Dijk et al., 2020), and we subsequently have evaluated their versatile applicability in adolescent C57BL/6J wild-type mice (Reiber et al., 2022). Despite the aspect of ethical responsibility that researchers bear, animal welfare assessments are mandatory according to Directive 2010/63/EU. For implementation of the Directive 2010/63/EU, animal experiments have to be designed to reduce pain, suffering, distress or lasting harm to a minimum (Smith et al., 2018). Unfortunately, there are still relevant gaps in knowledge on evidence-based, validated methods to determine and grade levels of distress and to decipher the multidimensional aspects of life-time severity (Keubler et al., 2020). Particularly in neuroscience research, genetic mouse models can experience peaks of distress during the course of post-weaning development (Sukoff Rizzo and Crawley, 2017). GluA1 deficient mice of the strain B6N.129-Gria1^{TM 1Rsp} represent a loss-of-function model of neuropsychiatric relevance, which has been classified to experience “no harm burden.” However, following the guidelines provided by the German Centre for the Protection of Laboratory Animals (Bf3R), the entire developmental phase of the mice between weaning in late infancy and young adulthood is not necessarily involved when investigating severity grading of genetically modified lines (EU, 2010; Bf3R, 2016; Reiber et al., 2022). Ethologically-based characterization, comprising cross-model, cross-species and cross-age validations, allow for conclusions about face validity (Percie du Sert et al., 2020), adequate refinement measures and welfare-based model prioritization (van Dijk et al., 2020). Our findings aim to shed light on subtle levels of distress associated with genetic deficiency, here Gria1^–/–, and provide further information for conclusions on cumulative severity grading and recommendations for refinement of genetic mouse lines.

In line with our aim, we have identified behavioral and biochemical alterations in the Gria1 knockout mouse model during the course of post-weaning development. Furthermore, we could demonstrate validity and feasibility of the newly designed composite measure scheme for the use in young mice of a loss-of-function genetic model of neuropsychiatric relevance. In this context, it is of importance that home-cage based approaches attracted increasing attention since they provide a non-intrusive, continuous, observer-independent, and, to a large extent, objective tool to uncover spontaneous activities (Pernold et al., 2019; Voikar and Gaburro, 2020). Investigations under standardized housing conditions reinforce reproducibility and external, inter-facility validity (Percie du Sert et al., 2020), and make practical implementation in the majority of conventional animal facilities possible. However, as applied here, the frequent presentation of novel objects in the home-cage, frequent handling of the young mice and the sequential testing situation represent confounding factors, which can essentially impact many of the readouts.

Among the parameters analyzed in the familiar environment of the mice, we examined the preference for saccharin, which we have previously introduced as a behavioral assay for severity assessment in laboratory rodents (e.g., Boldt et al., 2021). The innate preference for sweetness displayed by laboratory rodents can be considered to determine depression-related phenotypical traits (Crawley, 2006). The reduced preference for a sweet solution may indicate a reduced ability or inability to experience pleasure, which represents a key symptom of depression (Klein et al., 2015). Our previously generated baseline data from young wild-type mice (Reiber et al., 2022) suggest that the saccharin preference test can be applied in young mice from a prepubescent age onward. We have previously detected increased baseline preferences for saccharin in prepubescent and pubescent female mice, and a decline in sweetness preference becoming evident with reaching sexual maturity, which corresponds to sweetness preference in humans as reported in a scoping review by Venditti et al. (2020). In contrast, an earlier study conducted in rabbits reported a reduced sweetness preference displayed by younger animals (Reiber et al., 2022). Moreover, another more recent study on sucrose preference in young mice has successfully examined relevant strain differences in adolescent mice (Eltokhi et al., 2021). While there is no information available on sweetness preference in adolescent Gria1^–/– mice, respective data from adult Gria1^–/– mice draw an ambiguous picture so far: Austen and colleagues (Austen et al., 2017) reported that Gria1^–/– mice aged 11–30 weeks showed reduced licking rates, but, in line with our findings in early- and late-adolescent deficient mice, normal sucrose consumption levels. Moreover, Strickland et al. (2021) suggest that GluA1 is not implicated in hedonic value, based on investigations in approximately 17- to 34-week-old mice. In young mice, higher intake of a sweet solution can also be considered as a mechanism to compensate energy loss or higher metabolic needs, especially in light of reduced post-weaning body weights indicating signs of developmental delay.

Toward examination of spontaneous activities, monitoring of burrowing activity and non-maternal nest building behavior have been introduced as approaches reflecting complex behaviors of laboratory rodents in their everyday environment (Deacon, 2006a,b). Since mice show high levels of intrinsic motivation for these “non-essential” activities, they can be viewed as a “positive indicator of well-being” (Jirkof, 2014). Inversely, observed aberrations may therefore indicate even slight signs of discomfort or a reduced ability to experience pleasure. Since burrowing and nesting both represent complex tasks, requiring persistent, goal-oriented, concentrated execution, they can provide a more in-depth understanding of normal behavioral function (Jirkof, 2014). Moreover, nest building has been suggested as a tool to evaluate sensory gating impairments in mice (Kraeuter et al., 2019). Sensory and motor gating deficits have been described as a clinically relevant phenotypical feature in adult Gria1^–/– mice (Bannerman et al., 2004). A study analyzing nest building and burrowing performance in adult Gria1 knockout mice revealed impaired nest building activity in male Gria1^–/– mice as compared to male wild-type littermates, while burrowing performance was not affected by the Gria1 genotype (Bannerman et al., 2004). In addition, Pedersen et al. (2014) have reported limited burrowing performance in an adult sub-chronic phencyclidine mouse model of schizophrenia.

Considering nest building activity in young mice, there are two relevant studies examining the ability of young mice to build nests which show diverging results: An earlier report from Moy et al. (2004) found relevant levels of nest building activity in mice aged 3–4 weeks, while Eltokhi et al. (2020) detected no relevant nesting behavior in adolescent mice. However, a range of interfering factors have to be considered when interpreting nesting data, among others, body weight, strain and housing conditions (Bult and Lynch, 1997; Martin et al., 2016; Robinson-Junker et al., 2017; Schwabe et al., 2020). Considering burrowing activity, earlier studies indicated relevant burrowing activity in adolescent mice (Hart et al., 2012; McLinden et al., 2012; Eltokhi et al., 2020). Here, we observed a significant relation between GluA1 deficiency and nest building performance in adolescent female and male mice. In addition, we found a reduction of overnight burrowing performance in sexually mature adolescent mice significantly related to the GluA1 deficiency. Limited nesting and burrowing performance observed in Gria1^–/– mice may be interpreted as a consequence of novelty-induced hyperactivity associated with the genetic deficiency (Sanderson et al., 2008; Aitta-Aho et al., 2019). In addition, recent evidence suggests increased activity of the hippocampus in GluA1 deficient mice (Bannerman et al., 2014) as a cause of poorer adaptation and impaired short-term memory (Taylor et al., 2011; Barkus et al., 2014; Aitta-Aho et al., 2019). When exposed to the new stimuli of nesting and burrowing material, hyperactive mice may face difficulties with paying attention to a specific task and targeted action.

Home-cage monitoring of activity provides information throughout the circadian rhythm (Pernold et al., 2019). Considering voluntary wheel running activity, an innate and self-rewarding behavior displayed by laboratory rodents, Häger et al. (2018) have demonstrated a high sensitivity for severity grading in mice. We have previously detected relevant levels of wheel running activity in young female and male wild-type mice from prepubescent age onward (Reiber et al., 2022). Considering wheel running in GluA1 deficient mice, there exist two studies analyzing the performance of adult Gria1^–/– mice (Maksimovic et al., 2014; Ang et al., 2021). In line with our findings in adolescent Gria1^–/– mice, Maksimovic et al. (2014) found no significant genotype-related difference during adulthood considering the distance run in the wheel. Ang et al. (2021) reported moderately reduced levels of overall wheel running activity in Gria1^–/– mice with a significant activity reduction during the dark phase, suggesting relevant circadian rhythm impairments. Interestingly, a correlation between saccharin preference – which likewise represents a highly self-rewarding behavior – and voluntary wheel running activity has been described for wild-type mice, which reduced their intake of sucrose when simultaneously offered a freely accessible running wheel (Maksimovic et al., 2014).

Considering bench-top assessment, GluA1 deficient adult mice have been observed to show normal locomotor activity during continuous light/dark phase monitoring (Fitzgerald et al., 2010; Procaccini et al., 2011). However, as recently reported by Ang et al. (2021), Gria1^–/– mice displayed significantly reduced locomotor activity in the dark phase and increased locomotor activity in the light phase as compared to the wild-type control group, indicating relevant circadian rhythm abnormalities related to the genetic deficiency. Thus, although hyperlocomotion and increased resting durations in the zone ‘feeding’ observed in adolescent Gria1^–/– mice may be linked to lower post-weaning body weights, the activity levels measured are likely to be biased due to the selected short observation interval. In addition, relevant strain differences have been reported by Loos and colleagues (Loos et al., 2014), and there exists a number of external factors interfering with home-cage activity levels such as the presentation of novel objects and acclimatization to the cages.

When exposed to an open field arena in early and late adolescence, GluA1 deficient mice of both sexes displayed significant hyperlocomotion during both tests. This corresponds to earlier findings in adult Gria1^–/– mice, which displayed pronounced hyperlocomotion after short-term exposure to a novel environment (Bannerman et al., 2004; Wiedholz et al., 2008; Fitzgerald et al., 2010; Procaccini et al., 2011; Aitta-Aho et al., 2019; Ang et al., 2021). After exposure to a novel home-cage environment, hyperlocomotion was found to decrease to baseline levels within 5–6 h (Procaccini et al., 2011). Notably, we observed pronounced thigmotaxis in female prepubescent Gria1^–/– mice. Thigmotaxis, in particular the tendency to remain close to the arena walls, was introduced by Simon et al. (1994) to grade levels of anxiety in mice. Since we applied a dimmed lighting level with low lux values during the open field test, increased durations in the wall zone can also point to deficits in goal-directed exploration of the entire arena.

Analysis from classical neurobehavioral scoring in mice as originally described by Irwin (1968) draws an ambiguous picture. Although alterations of handling-associated parameters were detected, we did not see an increase of sum scores associated with the genetic deficiency. Since Irwin scoring primarily provides methodological guidance on the standardized investigation of behavioral changes in preclinical drug assessment (Irwin, 1968; Moscardo et al., 2007; Lynch et al., 2011), it may show relatively poor results for severity grading in mouse lines burdened mild to moderately. Nevertheless, when repeatedly applied, Irwin scoring may provide a complementary tool to detect and track conspicuous changes of the autonomous nervous system over the developmental course.

The assessment of adrenocortical activity may depict slight signs of discomfort on a non-invasive basis in laboratory rodents (Touma et al., 2004; Palme, 2019). Nascent evidence from investigations in young wild-type mice suggests that determination of fecal corticosterone metabolites may provide sensitive information on preceding distress in laboratory mice during post-weaning development (Kolbe et al., 2015; Reiber et al., 2022). Increased levels of adrenocortical activity detected in prepubescent female Gria1^–/– mice and in sexually mature Gria1^–/– mice may indicate higher levels of distress experienced by the deficient mice in their home-cage environment. However, higher levels of FCMs can be biased because of higher metabolic needs, e.g., as a consequence of hyperlocomotion, and do not directly allow for conclusions about discomfort (Koolhaas et al., 2011). Regarding the first time point of sampling during prepubescence, it has to be considered that increased adrenocortical levels may be influenced by disturbing influences such as weaning and new group constellation as demonstrated by Kolbe et al. (2015).

Considering post-weaning weight assessment, we could demonstrate reduced body weights significantly related to the GluA1 deficiency. The reduction of body weight at certain postnatal days can represent signs of developmental delay. As stated by Talbot et al. (2020), body weight loss can provide a sensitive criterion in the context of severity assessment, once combined with further model-specific criteria in a behavioral approach.

In summary, we could demonstrate behavioral modifications during post-weaning development in mice with global GluA1 depletion. The behavioral readouts argue against a persistent or long-lasting level of distress in adolescent GluA1 deficient mice. The findings rather support mild temporarily-limited behavioral impairments, relevant to evidence-based welfare assessment. In particular, reduced burrowing behavior in late adolescence and reduced nesting performance in early and late adolescence indicate behavioral aberrations in the mice carrying the genetic deficiency, which can reflect a transient, mild level of burden. Activity patters, however, draw an ambiguous picture so far: while open field data point to hyperactivity and transient thigmotactic behavior, data from continuous home-cage based wheel running do not support an overall increase of activity. Moreover, activity patterns should be interpreted in light of the integrity of circadian rhythm, and confounding factors such as novelty of experimental setting and direct animal-observer-interaction should be carefully considered, this also applies to home-cage based testing batteries. In addition, sex-specific differences emphasize the relevance for the inclusion of both sexes for behavioral welfare-assessments in genetic mouse lines. Taken together, the transient, mild behavioral impairments support a classification of the overall harm burden of the line as “mild.” This laboratory-specific suggestion should be regarded as a recommendation, and animal husbandry-related factors should not be ignored. Moreover, further investigations in genetically modified mouse lines of neuropsychiatric relevance are necessary in order to confirm the robustness and generalizability of the candidate parameter selection.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. In addition, datasets are available on the DFG FOR2591 Online Repository. Available online at: https://for.severity-assessment.de/.

The animal study was reviewed and approved by Government of Upper Bavaria, Maximilianstr. 39, 80538 Munich, Germany.

MR: study design, conduction of the experiments, data analysis and statistics, writing of the manuscript. HS: data analysis. RS: colony breeding and study design. PG: study design. RP: analysis of fecal corticosterone metabolites. HP: study concept and design, supervision, and writing of the manuscript. All authors: reviewed the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.