- Top of page
The first observations of postpartum oxytocin knockout (OTKO) mice found no maternal behavior deficits. However, it is unclear how detailed those observations were. In this study, we compared maternal behavior exhibited by OTKO and wild-type (WT) nullipara toward six 2–4-day-old foster pups during test sessions conducted on 3 successive days. Each day, subjects were placed in a clean cage 30 min prior to introduction of pups which were deposited in a clump adjacent to the middle of a long wall of each test cage. Behavior was measured for 3.5 h after which pups and test subjects were returned to their home cages. On test days 1 and 3, a significantly smaller proportion of OTKO females retrieved pups to a corner of their cage. Also, significantly fewer pups were retrieved to corners by OTKO females. In contrast to most WTs, most OTKO females mothered pups in the center of the cage where they were initially deposited. Pup-licking frequencies were significantly lower in OTKO females. Their self-grooming frequencies also trended toward being lower. Latencies to retrieve and lick pups, latencies to and frequencies of still crouching over pups and proportion of time in nest did not differ between groups. Our findings suggest that OT stimulates a significant proportion of pup-licking in nulliparous mice, a situation similar to lactating rat mothers. Our results also indicate that OT may play a role in the motivation to retrieve pups to a more secure location.
The first papers describing successful generation of oxytocin gene knock-out (OTKO) mice reported that postpartum females were unable to eject milk but displayed maternal behavior that was indistinguishable from wild-type (WT) mice (Nishimori et al. 1996; Young et al. 1996). The seemingly normal maternal behavior in OTKO mice was surprising to these investigators because of considerable evidence that OT facilitated rat, sheep and even wild mouse maternal behavior (Fahrbach et al. 1985; Insel & Harbaugh 1989; Kendrick 2000; McCarthy 1990; McCarthy et al. 1986; Pedersen 1997; Pedersen & Prange 1979; Pedersen et al. 1982; 1985; 1994; Van Leengoed et al. 1987; Wamboldt & Insel 1987). In those species, however, OT had chiefly been implicated in the initial postpartum activation of maternal behavior which involved a rapid reversal of avoidant, aggressive or infanticidal responses to newborns exhibited by nulliparous females. In contrast to rats, sheep and wild mice, nulliparous females of most strains of laboratory mice are not at all hostile to newborns but rather rapidly and avidly exhibit all components of species-typical maternal behavior. The aversion seen in wild nulliparous mice appears to have been selected out during many generations of captive breeding. Considering the lack of qualitative change in behavioral responses to newborns in parturient laboratory mice, it is not surprising that OTKO mice show no gross deficits in postpartum maternal behavior.
Recent studies employing sensitive quantitative behavior measures have found that intracerebroventricular (ICV) administration of an OT antagonist in rat mothers that had been nursing pups for several days significantly lowered their frequencies of pup-licking and upright nursing (Champagne et al. 2001; Pedersen & Boccia 2002; 2003). These findings indicate that central OT continues to stimulate about 20–40% of these components of rat maternal behavior after the early postpartum period. Maternal behavior measurements in OTKO mouse studies were not clearly described (Nishimori et al. 1996) and may have been insufficiently sensitive to detect differences of the magnitude seen in OT antagonist-treated rat mothers. In the current study, we test in nulliparous females the hypothesis that OTKO mice exhibit lower frequencies of some components of maternal behavior, in particular pup-licking and upright nursing.
- Top of page
All females that retrieved all six pups from the site where they were initially deposited moved pups to and grouped them in a corner of the cage where they established a nest. A few females retrieved some pups to a corner but failed to retrieve all pups to that site. Other females did not retrieve pups from the site in which they were initially deposited (adjacent to the middle of the front long wall of the cage) but rather established a nest there. Some of the latter females retrieved some pups that were apart from the main clump of pups to this nest site.
The pattern of pup retrieval and location of the nest differed significantly between OTKO and WT females. On test days 1 and 3, a significantly greater proportion of WTs compared with OTKOs retrieved all six pups to a corner (day 1 χ2 = 9.333, df = 1, P = 0.002; day 3 χ2 = 7.337, df = 1, P = 0.007; Fig. 1a) and retrieved at least one pup to a corner on test day 3 (χ2 = 2.397, df = 1, P = 0.028; data not shown). On test day 2, a higher proportion of WT compared with OTKO females retrieved all six pups to a corner, but this comparison was not significantly different. There was a main effect of genotype on the number of pups retrieved to a corner, such that WTs retrieved significantly more pups than OTKO mice (F1,25 = 8.886, P = 0.006, Fig. 1b). Post-tests revealed that on test days 1 and 3, the mean number of pups retrieved was significantly higher in WTs compared with OTKOs (Day 1 t = 2.760, df = 12, P = 0.017; Day 3 t = 2.397, df = 18, P = 0.028). A similar pattern was seen on day 2, but the difference between genotypes was not significant. There was a significant effect of genotype on the number of pups that were not moved from the site in which they were initially deposited on test day 1 but not on days 2 or 3 (t = 2.351, df = 25, P = 0.027; see Fig. 1c). Wild types left fewer pups in the deposition site than did OTKOs on test day 1.
Figure 1. Comparisons of oxytocin gene knock-out (OTKO) and wild-type (WT) nulliparous females on test days 1, 2 and 3 on proportion that retrieved all six pups to a cage corner(a), mean (±SEM) number of pups retrieved to a cage corner(b) and mean (±SEM) number of pups left where they were initially deposited(c) during the 30 min after the introduction of pups.*significant at P < 0.05.
Download figure to PowerPoint
Latencies to retrieve, lick and crouch over pups were only compared among animals that exhibited the respective behaviors during the first 30 min after the introduction of pups (Table 1). The proportion of subjects that exhibited these behaviors during the initial 30-min period only differed significantly between genotypes for retrieval of the first pup and retrieval of the last pup on test day 3 (14/14 WT females vs. 10/14 OTKO females for both, χ2 = 4.667, df = 25, P = 0.03). There was a significant effect of test day for latency to retrieve the first pup (F2,30 = 20.044, P < 0.001), with the latencies decreasing across the 3 test days. On test day 1, there were trends for WTs to retrieve the first and last pups sooner than the OTKOs (t = 1.934, df = 19, P = 0.07; t = 1.872, df = 17, P = 0.08, respectively); no group differences were found on test days 2 and 3. There were significant effects of test day, but no genotype or interaction effects, for latencies to pup lick (F2,50 = 7.828, P = 0.001) and to still crouch (F2,50 = 21.406, P < 0.001). Pup-lick latency was significantly shorter on test day 3 than days 1 or 2 for both WT and OTKO animals. Latency to still crouch declined significantly across the 3 days of testing for both WT and OTKO animals.
Table 1. Mean latencies ( ± SEM) in seconds to display maternal behaviors during the first 30 min after the introduction of pups across the 3 days of testing in WT and OTKO mice
| ||Day 1|| ||Day 2|| ||Day 3|| |
|Behavior||WT (n = 14)||KO (n = 14)||WT (n = 13)||KO (n = 14)||WT (n = 14)||KO (n = 14)|
|Retrieve first pup||120.73 ± 18.7||331.3 ± 143.7||72.9 ± 22.8||127.1 ± 59.6||78.4 ± 47.5||69.8 ± 44.5|
| ||n = 11||n = 10||n = 12||n = 10||n = 14||n = 10|
|Retrieve last pup||233.7 ± 31.8||347.4 ± 58.5||120.6 ± 29.5||172.6 ± 34.4||223.2 ± 78.8||171.5 ± 51.2|
| ||n = 11||n = 8||n = 11||n = 9||n = 14||n = 10|
|Pup lick||68.4 ± 11.4||84.2 ± 43.2||57.3 ± 11.8||96.0 ± 43.2||23.2 ± 7.2||13.6 ± 2.2|
| ||n = 14||n = 14||n = 12||n = 14||n = 14||n = 14|
|Still crouch||1035.4 ± 114.4||877.1 ± 91.8||580.1 ± 105.0||724.2 ± 161.2||443.9 ± 108.4||417.2 ± 76.6|
| ||n = 13||n = 13||n = 12||n = 11||n = 14||n = 13|
There was a significant main effect of genotype (F1,25 = 7.573, P = 0.011) and test day (F2,50 = 19.881, P < 0.001) but no interactions on pup-licking frequency (see Fig. 2a). There was also a trend for main effect of genotype (F1,25 = 3.735, P = 0.06) and significant effect of test day (F2,50 = 3.129, P = 0.05) but no interaction on self-grooming frequency (see Fig. 2b). Wild types exhibited more pup-licking and self-grooming than did OTKO mice. There were no main effects or interactions for frequencies of still crouch and female out of the nest.
Figure 2. Comparisons of knock-out (OTKO) and wild-type (WT) nulliparous females on test days 1, 2 and 3 on mean (±SEM) frequencies of pup-licking(a) and self-grooming(b) during 3-h observation periods. Over all test days, OTKOs licked pups at significantly lower frequencies (F1,25 = 7.573, P = 0.011) and trended toward grooming themselves at significantly lower frequencies (F1,25 = 3.735, P = 0.06).
Download figure to PowerPoint
In summary, nulliparous OTKO mice were less likely than nulliparous WT mice to retrieve pups from the initial deposition site to a corner of the test cage as indicated by significant differences in the proportion of each group that retrieved all six pups to a corner, the number of pups retrieved to a corner and the number of pups that were left in the initial deposition site. Perhaps, the most important finding in this study is the significantly lower pup-licking frequency in OTKO females. There was also a strong trend toward less self-grooming in OTKOs. Latencies to retrieve and lick pups as well as to adopt a still crouch over pups were not significantly different between genotypes and declined significantly across the 3 test days.
- Top of page
We found that nulliparous OTKO mice, just like nulliparous C57BL/6J WT mice and nullipara of many laboratory strains of mice, rapidly exhibit species-typical maternal behavior when given newborn pups. Quantitative behavior measurements, however, reveal a number of differences between OTKO and WT nullipara in pup retrieval behavior as well as in latencies of onset and frequencies of some categories of maternal and other behaviors. Some of our findings indicate that pup retrieval and pup-licking deficits may be constitutive in OTKO female mice. On test days 1 and 3, significantly lower percentages of OTKO females retrieved all six test pups to corner nest sites remote from the place where pups were initially deposited. The numbers of pups OTKOs retrieved to corner nest sites were also significantly lower on those test days. Although not reaching significance, a lower percentage of OTKOs retrieved all six pups, and fewer pups were retrieved by OTKOs on test day 2 as well. Furthermore, OTKO pup-licking frequencies were significantly lower across all test days. Nearly significant differences between OTKO and WT females in latencies to retrieve the first and the last pup appear to be transient, as they were found only on test day 1.
It is difficult to compare our findings of significant differences between OTKO and WT nullipara in some aspects of maternal responses to foster newborns with the report from Nishimori et al. (1996), the only other study to examine OTKO mouse maternal behavior, which found no differences between postpartum homozygote and heterozygote mice in pup retrieval, time in nest and time grooming pups. They made no comparisons with postpartum WT mothers. Unfortunately, they also provided no description of their maternal behavior tests. The behavior measurement methods used by Nishimori et al. (1996) may not have been sufficiently sensitive to reveal significant group differences in pup-licking frequencies in the 20–30% range, as we have found in the current study. Their test conditions may not have been conducive to identifying differences in pup retrieval (see Discussion below of how the way in which test pups are presented may affect behavior outcomes).
Differences in origin of the OTKO may also have contributed to the contrasting findings in the current study and that of Nishimori et al. (1996). So far, three separate lines of OTKO mice have been generated, two with a C57BL/6J background [Nishimori et al. 1996; Young et al. 1996[the latter used in the current study)] and the other a Black Swiss background (Gross et al. 1998; Robinson et al. 2002). The behavioral phenotypes resulting from the inactivation of some other genes have varied depending upon the background strain (Dobkin et al. 2000; Paradee et al. 1999; Toth 2003). Behavior phenotype may also be influenced by the construct used to inactivate a gene even within the same background strain. Both similar and contrasting behavioral phenotypes have been reported among the three lines of OTKO mice. While male anxiety is lower in both C57BL/6J background lines of OTKO mice (Mantella et al. 2003; Winslow et al. 2000), male aggression is reduced in one line (DeVries et al. 1997) but increased in the other (Winslow et al. 2000). Deficits in social memory are exhibited by one of the C57BL/6J background lines (Ferguson et al. 2000) as well as the Black Swiss background line (Choleris et al. 2003). Mothering may be a behavioral domain that is affected differently among the lines of OTKO mice.
The significant differences in pup retrieval we found between OTKO and WT females reflects contrasting patterns of response to encountering test pups deposited adjacent to the center of a long wall of the test cage, a location well away from the corners of the cage. Wild-type females responded more frequently by retrieving all pups to a corner, a pattern typical of rodent mothers. This may be the result of an instinct to initially move pups to a less exposed and safer nesting site. Oxytocin gene knock-out females, on the other hand, more frequently initiated and maintained their maternal responses to newborns in the vicinity of where the pups were initially deposited. They appear to be less motivated to move pups to a corner nest site. These contrasting response patterns between OTKOs and WTs persisted across the 3 test days and therefore may be a stable difference. This observation suggests that OT may play a role in retrieval or the motivation to move pups to a more ‘secure’ location. Future studies will put this hypothesis to the test by comparing OTKO and WT responses when pups are introduced in a more dispersed pattern and in a larger cage. Our results raise the possibility of a broader deficit of nesting behavior in OTKO females. Unfortunately, we did not observe whether OTKOs were less likely to establish nests in the corners of their home cages. Also, because we provided only a thin layer of bedding in test cages to prevent females from piling wood chips high enough to obstruct a clear view of their behavior, we did not quantify nest building during maternal behavior tests.
Early studies in rats unanimously found that ICV administration of OT antagonist or antiserum or lesioning the hypothalamic paraventricular nucleus, the origin of most OT projections within the brain, profoundly delayed the postpartum and ovarian steroid-induced onset of all components of maternal behavior (Fahrbach et al. 1985; Insel & Harbaugh 1989; Pedersen et al. 1985; Van Leengoed et al. 1987). Similar manipulations in lactating rat mothers that had several days of postpartum mothering experience failed to eliminate any components of maternal behavior leading to the conclusion that central OT played an important role in the initial onset but not the maintenance of established postpartum rat maternal behavior (Fahrbach et al. 1985; Insel & Harbaugh 1989; Numan & Corodimas 1985). However, recent investigations employing more quantitative behavior measurement methods have found that ICV-infused OT antagonist in nursing rat mothers several days postpartum significantly diminished the frequencies of pup-licking and upright posturing over pups by about 20–40% (Champagne et al. 2001; Pedersen & Boccia 2003). The results of our current comparison between OTKO and WT mice indicate that OT stimulates a similar percentage of pup-licking in highly spontaneous maternal nulliparous mice. Oxytocin enhancement of pup-licking is an important part of a larger role of OT in the intergenerational transmission of similar levels of maternal behavior and possibly acute stress responses in rats (Champagne et al. 2001; Pedersen & Boccia 2002). It will be of great interest to determine whether mice also exhibit intergenerational transmission of mothering and stress responses and whether OT stimulation of pup-licking is involved.
The lack of difference between OTKO and WT nullipara in the frequency of still crouch is not consistent with our prior finding that ICV-administered OT antagonist significantly decreased the frequency of still upright nursing in lactating rats (Pedersen & Boccia 2003). This negative finding could be an artifact of the rather brief duration (≥4 second) of still crouch required in the current study to score the behavior. If we compared frequencies of longer bouts or total duration of still crouching, differences in OTKO nullipara similar to those seen in OT antagonist-treated rat mothers may become apparent. Ventral trunk stimulation by pups can induce quiescent upright nursing in nulliparous female rats, but, because of their underdeveloped nipples, suckling stimulation is inadequate to sustain the long bouts seen in lactating females (Lonstein et al. 1999). Similar ineffectiveness of nipple stimulation in nulliparous mice would diminish the likelihood of observing still crouch differences between WT and OTKO females. Genotype differences could be more apparent in the postpartum period during which nipple stimulation-induced central OT release may be more involved in initiating and sustaining still crouching bouts. It will also be important to determine whether the lower pup-licking frequencies and altered pup retrieval pattern we found in OTKO nullipara are exhibited during the quite different physiological conditions of the postpartum period.
While we hypothesize that their lack of central OT activity reduces pup-licking and pup retrieval to corner nests in OTKO females, several other factors may have contributed to differences in maternal behavior between OTKO and WT mice. Oxytocin gene knock-out females were reared by heterozygote mothers (OT+/OT–), while WTs were reared by WT mothers. Heterozygote mothers may bestow less maternal care than WT mothers, and this difference in mothering may be transmitted to their offspring just as high vs. low frequencies of rat maternal behavior are transmitted from mothers to daughters (Francis et al. 1999). The odor of male OTKO mouse urine is aversive to female mice (Kavaliers et al. 2004). Compared with WT females, odors or other sensory cues from OTKO females may be unpleasant to pups resulting in alterations in their behavior which make them less attractive to OTKO females. Cross-fostering studies found that the amount of maternal behavior exhibited by rat mothers was altered when they were given pups born to females of other strains (Cierpial et al. 1990; Moore et al. 1997). Hence, the magnitude of genetic difference between mothers and the pups they rear can influence their maternal behavior. The only study of this type we could locate in mice found no effect on maternal behavior of cross-fostering pups between two strains (Carlier et al. 1983). Among the animals used in the current study, the OTKO females differ more genetically from the C57BL/6J test pups than do the WT females. It is possible that the greater genetic difference between OTKO females and test pups may diminish their maternal behavior by non-specific mechanisms other than OT deficiency.
Female OTKO mice exhibit greater anxiety and fear in novel situations and greater hypothalamic pituitary–adrenal axis activation by acute stressors than female WT mice (Amico et al. 2004; Mantella et al. 2003). In the current study, nulliparous female responses to newborn pups were first tested 30 min after transfer to clean cages with fresh bedding. This was not a novel experience, as used cages are replaced with clean cages with fresh bedding twice weekly in our animal facilities. It is possible, however, that the maternal behavior differences we found in nulliparous OTKO mice may be the result of greater sensitivity to being placed in a clean cage, even though they had experienced transfer to identical clean cages many times prior to behavior testing. The trends toward longer latencies to retrieve the first and last pups on test day 1 in OTKO mice are consistent with a greater initial inhibitory effect of transfer to a clean cage. On the other hand, the lack of group differences in latencies to lick pups or still crouch and the lack of group differences in frequencies with which females still crouched or were in nest indicate that OTKO subjects were not inhibited in making physical contact with and maintaining proximity to pups. The OTKO females' significantly less frequent pup retrieval to corner nest sites suggests they were less anxious about mothering pups in a more exposed position. The significantly reduced pup-licking and retrieval of pups to a corner that we found in OTKO females were of similar magnitude on test days 1 and 3, even though, on the latter test day, OTKOs were more familiar with the test procedure and the test cages. If greater response to novelty in OTKO females contributed to these differences, they would most likely have decreased across test days. An inverse relationship between anxiety and pup-licking frequencies has been observed in rats (Boccia & Pedersen 2001; Francis et al. 1999). It is of great interest that OTKO female mice, which exhibit greater anxiety in tests like the elevated plus maze (Mantella et al. 2003), also lick pups at lower frequencies than WT females. It will be important to determine whether anxiety and pup-licking frequencies are negatively correlated in mice as they are in rats.