In this paper, we review the scientific literature on maternal behaviour in commensal house mice and laboratory mice. Similar to other altricial species, female mice prepare a nest before parturition. Once the pups are born, nursing is the main part of maternal behaviour, and pups are weaned through a gradual non-aggressive process after about 3 weeks. Mice are social and both males and females show parental behaviour. Female mice giving birth at about the same time form communal nests, where pups are also communally nursed, a phenomenon that may confer benefit in inclusive fitness. However, social living may also be risky with conspecifics being the main predators of pups. A distinct aggressive behaviour pattern shown by pregnant and lactating female is thought to protect nest and pups against such attacks. Maternal aggression is influenced by the presence of pups and by litter size and composition. Communication through external stimuli from the pups contributes to maintaining maternal behaviour, thereby influencing pup growth. Handling of infants and pre-and peri-natal stress affects maternal behaviour. When resources are limited, females may reduce litter size through infanticide; however, the phenomenon of maternal cannibalism under normal laboratory conditions is poorly understood. Many studies included in this review use only standard tests to measure maternal behaviour, and more ethological research would be valuable to understand problems with reproduction in laboratory strains as well as to understand the influence of different housing conditions.

No abstract

We are honored to write the opening article of this focus issue of Lab Animal. This focus issue could not occur at a more important time for biomedical research and the use of animals in science in general. The progressive worsening of success rates in human trials (currently 1 in 9 drugs entering human trials will succeed) 1-4 , combined with the explosion of interest in the reproducibility crisis 5-8 and the recognition that most drugs fail in human trials due to insufficient efficacy 1,2,4 , has led to a growing suspicion that the failure of translation from animal work to human outcomes may in some way reflect issues in animal research itself 5-19 -after all, every drug that fails in humans "worked" in an animal model. Indeed pharmaceutical companies continue to disinvest in internal animal R&D, a trend begun in the last decade, passing on the cost and risk to academia and startups 5,20 . Even this approach is not foolproof as pharmaceutical companies often cannot replicate the results of published work from academia 5,8,13 . Accordingly, there is a growing trend to focus on human, not animal, work for basic discovery 17 .Discarding animal research entirely is not the answer. When properly used, animal models have incredible value, not least the ability to follow biomarkers from birth to disease onset in a year or less (in the case of mice), which is impossible in humans 17,18 . There are patterns and principles that can help us identify models and results that are more or less likely to translate, and there are also easily realized, simple changes in the execution of animal work that will inherently improve translation [16][17][18] . This isn't a new concept; looking back over the last 10-15 years we can see many authors have been candid about the merits, strengths, weaknesses, reproducibility, and translatability of various animal models 1,2,[4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19] . Our goal with this article is to unite the common themes in this broader emerging literature and this special issue.Thus, the central point is that we (i.e., refs. 17,18,26-28,39,41) do not represent a voice in the wilderness, but one voice in a chorus and that this emerging literature 1,2,[4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40]42,43 reflects a nascent discipline which can be codified as the study of how knowledge is gained from animal research. We propose the title "Therioepistemology" This focus issue of Lab Animal coincides with a tipping point in biomedical research. For the first time, the scale of the reproducibility and translatability crisis is widely understood beyond the small cadre of researchers who have been studying it and the pharmaceutical and biotech companies who have been living it. Here we argue that an emerging literature, including the papers in this focus issue, has begun to congeal around a set of recurring themes, which themselves represent a paradigm shift. This paradigm shift can be c...

BackgroundDespite being the most commonly used mammal in biomedical research, problems with perinatal mortality in mice have received little attention and the causes of pup death are still poorly known. Females are often housed alone with their litters and since the lost pups are generally eaten, it is commonly assumed that the mother has killed them. However, more detailed observations than have been reported previously in the literature are required to establish if the cause of death is infanticide. Litter loss can only be prevented efficiently after underlying causes have been carefully investigated and interpreted. The aim of this study was to investigate if females actively kill their pups by observing the behaviour of females and pups in litters that later were lost. We used video recordings of females that lost their entire litter to observe females in detail from parturition until the pups died. In total, 10 C57BL/6 females (wildtype and the knockouts Hfe−/− and β2m−/−) were studied, housed in Makrolon II cages with or without access to a small amount of nesting material.ResultsThree of the females had pups that were never seen moving, and another three females had one or two pups that never moved, indicating that some pups were most likely still-born. In five females with live-born pups, detailed observations from the time when a pup was last seen moving until it died were possible to carry out. We observed females eating dead offspring and interacting with both moving and dead pups. However, we never observed a pup stop moving when manipulated by the female, nor were any wounds seen in the pups. Hence, we found no evidence of infanticide when studying females that had lost their entire litter.ConclusionThese results suggest that other causes than infanticide plays a major role in mouse pup death, and stress the need for more systematic and careful investigations of the causality of litter loss.

Pup mortality is a considerable problem in laboratory mouse breeding and the view that parity influence survival of newborn mice is widespread. Some evidence suggests that maternal behaviour is related to offspring mortality in mice. Parental experience is a factor that can improve maternal behaviour and offspring survival in some mammals. However, few papers report a relationship between parity and pup survival in mice. We investigated the influence of strain and parity on loss of entire litters of C57BL/6 and BALB/c mice using data from a breeding colony. In total, 344 C57BL/6 and 146 BALB/c litters were included. We found a considerable mortality rate for both strains: 32% of C57BL/6 litters and 20% for BALB/c litters were lost. There was a significant difference in survival of the first litter between strains, with 3.6 times higher odds of mortality in C57BL/6 mice (p = 0.0028). Parity or previous parental experience of litter loss did, however, not affect litter loss. The scientific literature does not provide a clear picture of perinatal mortality in laboratory mice. Very few studies report perinatal mortality, and only a handful of papers exist where mortality was systematically studied; this area is thus poorly understood. If perinatal mortality in mice is not recognized and investigated, but instead considered normal when breeding mice, a serious welfare problem might be overlooked.

Injurious home-cage aggression (fighting) in mice affects both animal welfare and scientific validity. It is arguably the most common potentially preventable morbidity in mouse facilities. Existing literature on mouse aggression almost exclusively examines territorial aggression induced by introducing a stimulus mouse into the home-cage of a singly housed mouse (i.e. the resident/intruder test). However, fighting occurring in mice living together in long-term groups under standard laboratory housing conditions has barely been studied. We performed a point-prevalence epidemiological survey of fighting at a research institution with an approximate 60,000 cage census. A subset of cages was sampled over the course of a year and factors potentially influencing home-cage fighting were recorded. Fighting was almost exclusively seen in group-housed male mice. Approximately 14% of group-housed male cages were observed with fighting animals in brief behavioral observations, but only 14% of those cages with fighting had skin injuries observable from cage-side. Thus simple cage-side checks may be missing the majority of fighting mice. Housing system (the combination of cage ventilation and bedding type), genetic background, time of year, cage location on the rack, and rack orientation in the room were significant risk factors predicting fighting. Of these predictors, only bedding type is easily manipulated to mitigate fighting. Cage ventilation and rack orientation often cannot be changed in modern vivaria, as they are baked in by cookie-cutter architectural approaches to facility design. This study emphasizes the need to invest in assessing the welfare costs of new housing and husbandry systems before implementing them.

Efficiency in laboratory mouse breeding is hampered by poor reproductive performance, including the loss of entire litters shortly after birth. However, the underlying mechanisms are not yet fully understood and establishing the cause of death in laboratory mouse pups can be complicated. Newborn mouse pups are generally hidden in nests, dead pups are often eaten by the female, and the widespread practice of leaving periparturient females undisturbed complicates inspection, which may delay the discovery of pup loss. In order to efficiently prevent problems with litter loss, it is important to find key factors for survival. We investigated differences in periparturient behavior between female laboratory mice whose pups survived until weaning and females whose entire litters were lost. Video recordings of 82 primiparous females of the C57BL/6 strain or knockouts with C57BL/6 background were used. The mice were observed from 24 h before until 24 h after parturition and female behaviors coded using a pre-established ethogram. The relationship between behavior and survival was analyzed using logistic models, where litter survival was regressed on the proportion of 30-s observations with at least one occurrence of the behavior. We found that females with surviving litters performed more nest building behavior during the last 24 h before parturition (p = 0.004) and spent less time outside the nest during the entire observation period (p = 0.001). Increased litter survival was also associated with more passive maternal behaviors and the female ignoring still pups less. Females that lost their litters performed more parturition-related behaviors, suggesting prolonged labor. The results indicate that maternal behavior plays a significant role in laboratory mouse pup survival. Complications at parturition also contribute to litter mortality.

The use of animals in research entails a range of societal and ethical issues, and there is widespread consensus that animals are to be kept safe from unnecessary suffering. Therefore, harm done to animals in the name of research has to be carefully regulated and undergo ethical review for approval. Since 2013, this has been enforced within the European Union through Directive 2010/63/EU on the protection of animals used for scientific purposes. However, critics argue that the directive and its implementation by member states do not properly consider all aspects of animal welfare, which risks causing unnecessary animal suffering and decreased public trust in the system. In this pilot study, the ethical review process in Sweden was investigated to determine whether or not the system is in fact flawed, and if so, what may be the underlying cause of this. Through in-depth analysis of 18 applications and decisions of ethical reviews, we found that there are recurring problems within the ethical review process in Sweden. Discrepancies between demands set by legislation and the structure of the application form lead to submitted information being incomplete by design. In turn, this prevents the Animal Ethics Committees from being able to fulfill their task of performing a harm–benefit analysis and ensuring Replacement, Reduction, and Refinement (the 3Rs). Results further showed that a significant number of applications failed to meet legal requirements regarding content. Similarly, no Animal Ethics Committee decision contained any account of evaluation of the 3Rs and a majority failed to include harm–benefit analysis as required by law. Hence, the welfare may be at risk, as well as the fulfilling of the legal requirement of only approving “necessary suffering”. We argue that the results show an unacceptably low level of compliance in the investigated applications with the legal requirement of performing both a harm–benefit analysis and applying the 3Rs within the decision-making process, and that by implication, public insight through transparency is not achieved in these cases. In order to improve the ethical review, the process needs to be restructured, and the legal demands put on both the applicants and the Animal Ethics Committees as such need to be made clear. We further propose a number of improvements, including a revision of the application form. We also encourage future research to further investigate and address issues unearthed by this pilot study.