High laboratory mouse pre-weaning mortality associated with litter overlap, advanced mother age, small and large litters

High and variable pre-weaning mortality is a persistent problem among the main mouse strains used in biomedical research. If a modest 15% mortality rate is assumed across all mouse strains used in the EU, approximately 1 million more pups must be produced yearly to compensate for those which die. A few environmental and social factors have been identified as affecting pup mortality, but optimizing these factors does not cease the problem. This study is the first large study to mine data records from 219,975 pups from two breeding facilities to determine the major risk factors associated with mouse pre-weaning mortality. It was hypothesized that litter overlap (i.e. the presence of older siblings in the cage when new pups are born), a recurrent social configuration in trio-housed mice, is associated with increased newborn mortality, along with high mother age, large litter size, as well as a high number and age of older siblings in the cage. The estimated probability of pup death was two to seven percentage points higher in cages with compared to those without litter overlap. Litter overlap was associated with an increase in percentage of litter losses of 19% and 103%, respectively, in the two breeding facilities. Increased number and age of older siblings, high mother age, small litter size (less than four pups born) and large litter size (over 11 pups born) were associated with increased probability of pup death. Results suggest that common social cage configurations at breeding facilities are dangerous for the survivability of young mouse pups. The underlying mechanisms and strategies to avoid these situations should be further investigated.


39
High pre-weaning mortality of laboratory mice is a major welfare and economic problem affecting 40 mouse breeding at academic and industrial laboratories worldwide. Previous studies report pup 41 mortalities from less than 10% [1] to as high as 49% [2] for C57BL/6 mice, one of the most 42 commonly used mouse strains. Despite the general ongoing effort to reduce the number of animals 43 in research and improve their welfare according to the 3R principle for research [3], high pre-44 weaning mortality rates persist and very little systematic research has been done to identify causes 45 of poor survival. Data from experimental and observational studies conducted by the authors of 46 this work at different breeding facilities in three different countries revealed that 32% of 344 litters 47 (retrospective analysis, Germany[4]), 33% of 55 litters (experimental data, U.K. [5]), and 18% of 48 510 litters (experimental data, Portugal[6]) were completely lost, with the overall mortality varying 49 from 25[6] to 52% in trio-bred mice [5] in the experimental studies. If a modest level of 15% 50 mortality is assumed across all mouse strains, at least 1 million more mice must be produced every 51 year just in the European Union (EU) to compensate for pups that die before they can be used in 52 science (estimate based on the number of mice used yearly in research in the EU; European 53 Commission 2020 [7]). Such losses are contrary to the 3R principle that is now explicit in EU 54 legislation [8] and incur extra breeding costs of €5-8 million yearly. Several environmental, 55 management and behavioural factors have been linked to pup mortality, such as thermal 56 environment of the cage, level of parental care, mother age, litter size, provision of nest material, 57 and cage manipulation [5,[9][10][11][12][13], but manipulating these factors has not as yet eliminated the 58 mortality problem. Recently, we identified the presence of older litter mates in the cage when a 59 new litter is born (litter overlap) as a major factor affecting pup survival [5]. In a study with 55 60 litters of C57BL/6 mice (n=521 pups) housed in trios [5], a 2.3 fold increase was found in litter 61 loss in cages where older littermates were present, compared to trio cages with no older littermates. 62 Litter overlap happens in both trio (two adult females and one male) and pair (one adult female 63 and one male) housing, which are the most common configurations in mouse breeding. Although 64 litter overlap is more frequent in trios due to the presence of two breeding females, the number of 65 pups weaned per litter is not reduced in trios compared to pairs, while trios wean more pups per 66 cage [14]. One possible reason for this is that litter overlap in pair cages affects pup mortality more 67 severely as compared to trio cages. In pair cages, litter overlap occurs when the only female of the 68 cage gives birth before weaning her previous litter. In these cases, the age gap between litters 69 becomes large, which might be especially detrimental to pup survival. 70 Previous research into factors affecting laboratory mouse reproduction used primarily 71 experimental study approaches, where the sample size was small and animal management and data 72 collection differed from standard practice in a breeding facility. With the increasing use of 73 breeding management software, it is now possible to use much larger datasets representing the 74 reality of practical laboratory mouse breeding. In this study, a dataset of 219,975 pups was 75 analysed from two different collaborating breeding facilities in the UK (58,692 and 161,283 pups), 76 by modelling the risk of a newborn mouse dying as a function of the age and number of older 77 littermates, as well as of mother age. It was hypothesized that litter overlap is a recurrent social 78 configuration and that the risk of pup mortality increases with litter overlap, high mother age, large 79 litter size, as well as a high number and age of older siblings in the cage. 80 Enterprises, Rungis, France), resulting in one data line per pup. A total of 11% (C1) and 21% (C2) 104 of the data provided was excluded, mainly due to incongruent data records, implausibly large litters 105 (more than 13 pups, unless confirmed as correct), unreliable information on number and age of 106 older pups in the cage, or missing information. Male pups at C2 euthanized at day 7 pp or later 107 were coded as surviving. Litters with males euthanized before day 7 pp were excluded. 108

Material and methods
Pup mortality before weaning was coded as 0 (survived) or 1 (died) and used as the dependent 109 variable. Environmental and social factors were considered as risk factors for pup death. The risk of pup death was modelled by mixed logistic regression, using the GLIMMIX procedure 118 in SAS (2018 University Edition, SAS Institute Inc., Cary, NC, USA). Multicollinearity among 119 independent variables was checked by using the variance inflation factor (VIF) and regressing 120 each independent variable on the others. As a consequence, Year, Month, and Father Age were 121 excluded from the analysis. 122 Data of C1 and C2 were combined together and two separate models were constructed; one model 123 using data for all pups, including those with and without litter overlap, and another model for pups 124 with litter overlap only. In both models, litter identity was included as a random effect to account 125 for clustering. The models were built by adding one independent variable of interest (Mother Age, 126 Litter Size, Overlap (first model) or Sibling Number and Sibling Age (second model)) at a time in 127 a stepwise process with bidirectional elimination. Independent variables with P ≤ 0.05 were kept 128 in the model. Weekday, Season, and Collaborator were then tested one at a time as confounders, 129 followed by possible interactions and higher order terms. Least-squares means of Weekday, 130 Season, and Overlap were examined and compared among different variable levels. 131

132
The percentage of pups dying before weaning was 39% at C1 and 14% at C2, while the mean 133 number of Litter Size was 7.6 pups born/litter in both collaborators. In 42% of the C1 litters and 134 78% of the C2 litters no pups died. The percentage of litters with at least 90% death rate was 28% 135 at C1 and 9% at C2. Approximately 50% and 57% of the litters were born with the presence of 136 older siblings in the cage (litter overlap) in C1 and C2, respectively. 137 The first model (all pups) contained the variables Collaborator, Season, Weekday, Mother Age, 138 Litter Size and Overlap. The second model (pups with litter overlap) contained variables 139 Collaborator, Season, Weekday, Mother Age, Litter Size, Sibling Number, and Sibling Age. 140 Collaborator interacted significantly with all the independent variables in both models, except 141 Mother Age in the second model. Therefore, results for C1 and C2 are presented separately. Sibling 142 Number interacted significantly with Sibling Age, while Litter Size affected pup death probability 143 in a quadratic fashion in both models. Model details are available in S1 and S2 Tables. 144 The estimated probability of pup death was seven (C1) and two (C2) percentage points higher (P 145 < 0.01) in cages with the presence of older siblings compared to cages without an older litter ( Fig  146   1A and 1B). At C1, 31% of the overlapped and 26% of the non-overlapped litters had a total litter 147 loss (all pups dying, Fig 1C), whereas at C2, the corresponding figures were 12% and 6% (Fig  148   1D). 149 Sibling Age were associated (P < 0.01) with an increase in the probability of pups dying at both 160 C1 and C2 (Figs 2 and 3). 161 Euthanized males after seven days of age were considered as survivals, but this is an assumption 205 and may have led to an underestimation of mortality. Finally, any differences in accuracy of data 206 entry may have affected the results. C1 had a more consistent and early counting of pups than C2. 207 Thus it is possible that C1 has a better accuracy in detecting the number of pups born compared to 208 C2 where pups born could be underestimated, considering that most of deaths happen within 48h 209 pp and dead pups often get cannibalized by the dam, thus not seen by the caretakers. Productivity 210 differences either in pup survivability or in the actual number of pups born due to the distinct 211 mouse sub-strains between C1 and C2 could also underlie the differences in overall mortality 212 found between C1 and C2. 213 The found higher mouse pre-weaning mortality in trios with overlapped litters, i.e. litters born 214 when an older litter was present, compared to non-overlapped litters is in agreement with previous 215 experimental findings. In outbred mice derived from the C57BL/6J, BALB/c, and DBA/1J strains, 216 Schmidt et al. (2015)[17] found pup mortality to increase with increasing age gap between the 217 litters sharing a cage at a specific time. Understanding why being born into a cage with an older 218 litter is so dangerous requires information about events around pup death and the condition of dead 219 and dying pups. Past research have frequently associated pup mortality with infanticide[17,18], 220 assuming that cannibalized pups were killed before they were eaten. However, previous behavior 221 studies conducted by the authors of this study revealed that infanticide precedes less than 15% of 222 the cannibalism events[5,19] and that pups die primarily from other causes than direct killing. 223 Litter asynchrony, which often leads to overlap, is likely to increase unequal competition for access 224 to milk and parental care, trauma caused by trampling and stepping of newborns by the adults or 225 the older siblings, and problems related with increased cage stocking density. The authors found no differences between mother age group on pre-weaning mortality and litter 273 size both at birth and at weaning, but reported that young mothers produced F2 litters with higher 274 expectation of survival and body weight than those of old mothers. 275

276
Pup death probability was higher in either small or large litters (less than four or more than 11 277 pups born). The reduced survivability in small litters is in agreement with previous reports for 278 C57BL/6[5] and F1 hybrid (C57BL/6JIco × CBA/JIco) mice [9]. This may be related to the amount 279 of parental care. Ehret and Bernecker (1986)[31] demonstrated that early pup vocalization, which 280 gradually increases in frequency after birth, is essential to maintain maternal attention at high 281 levels, which leads to improved pup weight gain, as compared to pups from mothers which were 282 unable to hear them. Therefore, it is possible that a small newly born litter does not emit sufficient 283 vocal cues to ensure sufficient maternal care. Rat litters[32] with one single pup were found to 284 perform only about 10% and 5% of the suckling stimuli performed by litters of 10 and 22 pups. 285 As a consequence, the milk yield of dams (estimated based on adjusted measures of the pups' daily 286 weight gain) raising single pups was only -0.4% to 7.0% of those raising 10 pups, which led one-287 pup litters to have the lowest growth rate. More than half of the one-pup litters did not show any 288 weight gain in the first five days pp. From an evolutionary perspective, a small litter is less worth 289 investing in than a larger litter: Maestripieri and Alleva (1991) [33] demonstrated that CD-1 dams 290 of large litters (eight pups) spent more than twice as much time displaying litter defense behaviors 291 against intruder males than dams of small litters (four pups). The increase in pup death probability 292 found in litters of 12 pups and above, on the other hand, may be a result of increased sibling 293 competition for access to milk, as discussed above, and also may represent a ceiling in milk 294 production capacity by the mothers[20,34,35]. 295

296
In both collaborators, there seemed to be a decrease the probability of pup death towards the end 297 of the week, possibly associated with the timing of cage changes, a management routine which 298 affects the mice as well as the accuracy of mortality detection. In C1, which provided records on 299 cage cleaning dates, these closely mirrored the pattern of pup death probability. To reliably count 300 the number of pups, the cage must be opened and animals moved, something that often only 301 happens at cage cleaning when manipulation is unavoidable. Mortality is therefore likely to be 302 more accurately detected for litters born on cage changing days. For example, an eight-pup litter 303 born on a Tuesday with cage cleaning schedule for the same day will be recorded as an eight-pup 304 litter. If two of these pups die in the following 24 hours, this litter's pre-weaning mortality will be 305 recorded as being 25% at weaning. A similar litter born on a Saturday with two pups dying on 306 Sunday, and subsequently cannibalized, will be recorded at the Tuesday cage change as a litter 307 with six pups born with no pre-weaning deaths. 308 Still, the mouse disturbance hypothesis cannot be disregarded. If pup mortality is affected by cage 309 change, the same pattern would be expected in cage change frequency as in pup death probability 310 per Weekday (of birth). Reeb-Whitaker et al. (2000)[1] found a higher pup mortality in cages with 311 weekly changes than those changed once every two weeks. Cage change requires that mice are 312 moved from the dirty to a clean cage, an event that triggers a stress response evidenced by increases 313 in serum corticosterone[36] and general activity [37]. It is possible, therefore, that cage change 314 interferes with parental behavior in breeding cages, which could aggravate pup mortality around 315 those days. 316

317
The present study revealed that high pre-weaning mortality in laboratory mice (C57BL/6) is 318 associated with advanced mother age, litter overlap, the presence of a high number and age of 319 older siblings in the cage, and a small (less than four) or large (more than 11 pups) litter. The