Oxytocin Signal and Social Behaviour: Comparison among Adult and Infant Oxytocin, Oxytocin Receptor and CD38 Gene Knockout Mice

Authors

  • H. Higashida,

    1. Department of Biophysical Genetics, Kanazawa University Graduate School of Medicine, Kanazawa, Japan.
    2. Research Center for Child Mental Development, Kanazawa University, Kanazawa, Japan.
    3. Core Research for Evolutional Science and Technology, Tokyo, Japan.
    Search for more papers by this author
  • O. Lopatina,

    1. Department of Biophysical Genetics, Kanazawa University Graduate School of Medicine, Kanazawa, Japan.
    2. Core Research for Evolutional Science and Technology, Tokyo, Japan.
    3. Department of Biochemistry, Medical, Pharmaceutical and Toxicological Chemistry, Krasnoyarsk State Medical University, Krasnoyarsk, Russia.
    Search for more papers by this author
  • T. Yoshihara,

    1. Research Center for Child Mental Development, Kanazawa University, Kanazawa, Japan.
    Search for more papers by this author
  • Y. A. Pichugina,

    1. Department of Biophysical Genetics, Kanazawa University Graduate School of Medicine, Kanazawa, Japan.
    2. Department of Psychiatry, Krasnoyarsk State Medical University, Krasnoyarsk, Russia.
    Search for more papers by this author
  • A. A. Soumarokov,

    1. Department of Psychiatry, Krasnoyarsk State Medical University, Krasnoyarsk, Russia.
    Search for more papers by this author
  • T. Munesue,

    1. Research Center for Child Mental Development, Kanazawa University, Kanazawa, Japan.
    2. Department of Psychaitry and Neurobiology, Kanazawa University Graduate School of Medicine, Kanazawa, Japan.
    Search for more papers by this author
  • Y. Minabe,

    1. Research Center for Child Mental Development, Kanazawa University, Kanazawa, Japan.
    2. Department of Psychaitry and Neurobiology, Kanazawa University Graduate School of Medicine, Kanazawa, Japan.
    Search for more papers by this author
  • M. Kikuchi,

    1. Department of Psychaitry and Neurobiology, Kanazawa University Graduate School of Medicine, Kanazawa, Japan.
    Search for more papers by this author
  • Y. Ono,

    1. Department of Psychaitry and Neurobiology, Kanazawa University Graduate School of Medicine, Kanazawa, Japan.
    Search for more papers by this author
  • N. Korshunova,

    1. Department of Biophysical Genetics, Kanazawa University Graduate School of Medicine, Kanazawa, Japan.
    Search for more papers by this author
  • A. B. Salmina

    1. Department of Biochemistry, Medical, Pharmaceutical and Toxicological Chemistry, Krasnoyarsk State Medical University, Krasnoyarsk, Russia.
    Search for more papers by this author

Correspondence to: Haruhiro Higashida, Department of Biophysical Genetics, Kanazawa University Graduate School of Medicine, 13-1 Takara-machi, Kanazawa 920-8640, Japan (e-mail: haruhiro@med.kanazawa-u.ac.jp).

Abstract

Oxytocin in the hypothalamus is the biological basis of social recognition, trust, love and bonding. Previously, we showed that CD38, a proliferation marker in leukaemia cells, plays an important role in the hypothalamus in the process of oxytocin release in adult mice. Disruption of Cd38 (Cd38−/−) elicited impairment of maternal behaviour and male social recognition in adult mice, similar to the behaviour observed in Oxt and oxytocin receptor (Oxtr) gene knockout (Oxt−/− and Oxtr−/−, respectively) mice. Locomotor activity induced by separation from the dam was higher and the number of ultrasonic vocalisation calls was lower in Cd38−/− than Cd38 +/+ pups. However, these behavioural changes were much milder than those observed in Oxt−/− and Oxtr−/− mice, indicating less impairment of social behaviour in Cd38−/− pups. These phenotypes appeared to be caused by the high plasma oxytocin levels during development from the neonatal period to 3-week-old juvenile mice. ADP-ribosyl cyclase activity was markedly lower in the knockout mice from birth, suggesting that weaning for mice is a critical time window of plasma oxytocin differentiation. Breastfeeding was an important exogenous source of plasma oxytocin regulation before weaning as a result of the presence of oxytocin in milk and the dam’s mammary glands. The dissimilarity between Cd38−/− infant behaviour and those of Oxt−/− or Oxtr−/− mice can be explained partly by this exogenous source of oxytocin. These results suggest that secretion of oxytocin into the brain in a CD38-dependent manner may play an important role in the development of social behaviour.

Oxytocin, a nonapeptide involved in reproduction, is synthesised in the paraventricular nucleus and supraoptic nucleus of the hypothalamus, and travels down neuronal axons to the posterior pituitary (1–4). Oxytocin is secreted into the general circulation from the nerve endings of the neurohypophysis and into the brain from dendrites. It is well known that oxytocin is linked to complex social behaviour (5–9) (Fig. 1a). In humans, intranasal oxytocin may promote trust (10), gaze (11) or face recognition (12), and infusion of oxytocin can increase generosity (13). In rodents, oxytocin is highly involved in social interaction, social recognition, pair bonding and maternal behaviour (14–21). In addition, animal studies have shown that increased levels of oxytocin in the early postnatal period may affect behaviour and last into adulthood (8, 15, 22), and that s.c. administration of low doses of oxytocin facilitates social recognition (21). Two types of mice with oxytocin (Oxt) or oxytocin receptor (Oxtr) gene knockout (Oxt−/− or Oxtr−/−) show profound social amnesia (14, 18, 22–26) (Fig. 1b). Social amnesia can be fully rescued by injection of oxytocin into the medial amygdale in Oxt−/− mice. Impairment of social behaviour is clearly observed even in pups. These observations suggest that oxytocin plays an important role in social behaviour by stimulation of oxytocin receptors during brain development throughout the juvenile to adult stages.

Figure 1.

 Regulatory mechanisms affecting oxytocin activity in the context of social behavior. (a) Regulation at the level of central and peripheral oxytocin (Oxt) secretion and action. (b) Disruption of components of the CD38/oxytocin/oxytocin receptor (Oxtr) system affects social memory. (c) Treatment strategy to improve oxytocin-mediated regulation of social behavior.

Recently, we reported that CD38 is required for regulation of social behaviour by oxytocin in mice (27, 28). The transmembrane glycoprotein, CD38, detected in leucocytes is expressed on many neuronal cells, although its function in the brain remains unclear (29, 30). CD38 possesses ADP-ribosyl cyclase activity, which produces cyclic ADP-ribose (cADPR) from an abundant substrate in the brain, β-NAD+ (29). cADPR is a potential intracellular second messenger and a cofactor, together with Ca2+, for Ca2+ mobilisation from ryanodine-sensitive Ca2+ pools (30, 31). Adult mice with a null mutation in CD38 showed deficiencies in social behaviour, such as mate recognition and memory in males and pup retrieval in females, as a result of an abnormality of central and peripheral oxytocin secretion (27, 28). We also showed that decreased formation of cyclic ADP-ribose (cADPR) results in dysfunction of Ca2+-induced Ca2+-release for oxytocin secretion in hypothalamic oxytocin neurones (27). In the present study, we present an overview of the relationship between social behaviour and oxytocin levels in CD38 knockout mice, and compare the results among three different genotypes: Oxt−/−, Oxtr−/− and Cd38−/−.

Growth of knockout mice

Both Cd38+/+ and Cd38−/− mice grew well and gained weight (27, 28). There were no significant sex-related differences in body weight, indicating that these mice fed well from the infant stage through the dam’s milk and by solid food after weaning to the adult stage. Cd38−/− mice showed no deficits in lactation/milk ejection (Table 1).

Table 1.   Comparison of Phenotypes in Three Knockout Mouse Strains.
 Oxt−/−*Oxtr−/−**Cd38−/−***
  1. References: *(22, 26); **(24); ***(27, 28).

  2. +, not impaired; −, impaired; +/−, partially impaired under stressful condition; ND, not determined.

Fertility+++
Uterine labour+++
Nurturing
 Milk ejection+
 Lactating action+++
 Pup retrieval++/−+/−
 Crouching over+++
Growth+
Social memory
Locomotor activityNDHighHigh
Mam’s anxietyND+
Male aggressionND+

Nishimori et al. (23) reported that mice lacking Oxt are both viable and fertile. oxytocin-deficient females have no obvious defects in fertility or reproduction, including gestation and parturition. However, all offspring die shortly after birth because of the dam’s inability to nurse. However, postpartum injection of oxytocin into the Oxt-deficient dams restores milk ejection and rescues the offspring (23). Similarly, Oxtr−/− mice are viable and have no obvious defects in fertility or reproductive behaviour, although dams exhibit normal parturition followed by defects in lactation and maternal nurturing (24, 25). These results indicate that the oxytocin/oxytocin receptor/CD38 signalling pathway is not essential for normal parturition in mice, although milk ejection requires oxytocin/oxytocin receptor (Table 1).

Stress-induced behaviour in infant mice

We have examined isolation-induced locomotor behaviour and ultrasonic vocalisation (USV) of infant male mice (24, 26) on postnatal day 7. Cd38−/− males showed significantly higher levels of locomotor activity during the first 3 min after separation from the dam, when they were examined individually in the grid-crossing test in an observation chamber (28). Next, activity of pups was measured simultaneously with four pups together for a much longer period, in which there was no physical contact but some type of interaction, probably with USV as discussed below. Higher locomotor activity in Cd38−/− pups was observed that persisted up to 18 min in the test, although the difference was not marked compared to the other mice (28).

Both Cd38−/− and Cd38+/+ pups emitted calls on isolation. The properties of USV were similar in that the frequency was approximately 70 KHz and duration was approximately 60 ms in both genotypes. However, the calls per 2-min session were less frequent in Cd38−/− infants than wild-type controls, with an average reduction of 38% (Table 2). These results agreed well with previous observations in Oxt−/− and Oxtr−/− mice (14, 23, 24, 26). However, the degree of disruption of infant behaviour appeared to be much milder in the case of Cd38−/−, in comparison with the two oxytocin-related knockout mouse strains (Table 2). These observations suggested that Cd38−/− pups retain the ability to interact socially with others to a greater extent than Oxt−/−and Oxtr−/− pups.

Table 2.   Comparison of Infant Ultrasonic Vocalisation and Locomotor Activity in Three Knockout Mouse Strains.
 Knockout strain
Oxt−/−*Oxtr−/−**Cd38−/−***
  1. Refs: *(22, 26); **24; ***(27, 28). n.d., not determined.

Call−80%−95%−38%
Locomotorn.d.+500%+161%

Maternal nurturing behaviour in female mice under stress

Female Cd38−/− mice displayed disrupted maternal behaviour after separation of dam and pups (27). We observed the dams’ behaviour with their newborn pups in home cages with the pups temporarily placed outside the nest. Normal dams retrieved five pups tested precisely and very quickly (with the average latency of 43 ± 3 s) to the same small area of the nest (27), whereas Cd38−/− dams took a significantly longer time to begin retrieval (the average latency of 76 ± 8 s; P < 0.01), moved around continuously and sniffed as if their interest was drawn to many things other than their pups. They often dropped the pups during retrieval as if they did not remember the way to the nest, and this resulted in the pups becoming scattered in different places. However, Cd38−/− dams fed the pups sufficiently for them to grow to the same weight as the controls, as noted above. These results indicated clear abnormalities in maternal nurturing behaviour in Cd38−/− postpartum mice under stressful conditions, such as separation. This imperfect and neglect-like nurturing behaviour partially resembles observations in human cases, because a child’s or pup’s basic needs (i.e. staying in a warm and comfortable house or nest) are not adequately met (32, 33).

Pregnant Oxt−/− dams clean their offspring after delivery and keep them in the nest (23). The dams quickly retrieve offspring that are moved outside the nest. However, milk was not observed in the stomachs of offspring born to Oxt−/− females. Dams encouraged latching on behaviour by newborn offspring to the nipples, but failed to elicit milk release. Because they have milk in their mammary glands, i.p. injection of oxytocin produced milk ejection.

Oxtr−/− dams, however, build nests and spend the majority of this period crouching over their pups (24). However, pups of Oxtr−/− females are often found scattered around the cage. The dams’ responses to pups placed in different corners of the cage were monitored. Oxtr−/− dams showed a significantly longer latency to begin retrieval and required a longer time for complete retrieval of all the pups. The impairment of retrieval observed in Oxtr−/− dams was similar to that in Cd38−/− females.

Social amnesia in male mice

Normal mice that experienced repeated pairings with the same female showed a significant decline in the time spent investigating the female upon subsequent presentations of the same animal (14, 22). Cd38+/+ young adults showed this phenotype, which was not the result of a loss of interest but instead the retained memory of the paired female. Therefore, they did not need to investigate further, but instantaneously recognised the paired mate. By contrast, Cd38−/− males showed sustained high levels of investigation at each encounter with the same female and the same level of investigation. This behaviour was a result of males’ amnesia of conspecifics (14).

Because Cd38−/− mice did not have deficits in either olfactory-guided foraging or habituation to a nonsocial olfactory stimulus, as determined from the preference ratio of consumption of isovaleric acid solution (which was a noxious or offensive odour) in their drinking water, the impairment of social memory did not depend on deficits in main olfactory bulb function (27). Although the function of CD38 in social behaviour could be very specific to the particular neural circuitry involved, the data presented to date do not exclude a more general cognitive dysfunction. Therefore, we examined whether the CD38 mutants are able to learn the shock experience in the passive avoidance test. The results obtained indicated that they can indeed learn. We concluded that Cd38−/− males with persistent interest during repeated presentations fail to develop social memory. The abnormality found in Cd38−/− resembles a memory deficit found in Oxt−/− and Oxtr−/− mice (14, 24) (Table 2).

Plasma oxytocin levels in adult mice

To determine the link between CD38 and oxytocin, we measured plasma and cerebrospinal fluid (CSF) oxytocin levels (Fig. 2a, b) and found that, in comparison with wild-type, Cd38−/− mice had reduced plasma oxytocin levels, but elevated levels in the hypothalamus and pituitary (27). These observations indicated that although oxytocin is produced and packaged into vesicles in the hypothalamic neurones and posterior pituitary nerve endings in Cd38−/− mice, it is not released into the brain and bloodstream. Therefore, oxytocin does not function. Indeed, the behavioural phenotype of Cd38−/− mice could be normalised even by a single s.c. oxytocin injection, because oxytocin is transported into the brain, probably through the blood–brain barrier, as reported by Jin et al. (27) (Fig. 2c, d). We also used a genetic approach by infusion of a virus carrying the human CD38 gene into the third ventricle of knockout mice. This procedure resulted in normalisation of the plasma and CSF oxytocin levels and thereby of normalised social memory, indicating that the mechanisms underlying social behaviour require CD38-dependent oxytocin secretion.

Figure 2.

 Oxytocin (OXT) and vasopressin (AVP) levels in cerebrospinal fluid (CSF). Cerebrospinal oxytocin (a) or vasopressin (b) concentrations in wild-type (Cd38+/+), knockout (Cd38−/−) and Cd38−/− mice re-expressing hCD38 by injection of lentiviral vector. Cerebrospinal oxytocin levels after s.c. administration of oxytocin (100 ng/ml) or saline (0.2 ml) in Cd38 +/+ (c) and Cd38−/− (d) mice. Data are the mean ± SEM. n = 3–6. *P < 0.01 from Cd38+/+ and Cd38−/− mice re-expressing hCD38.

Plasma oxytocin levels in infant mice

As mentioned above, the degree of disruption of infant behaviour appeared to be much milder in the case of Cd38−/−, and this speculation prompted us to measure plasma oxytocin levels to determine whether the oxytocin level in pups is decreased as adults. We measured the plasma oxytocin levels in five different phases of development (from 1 week to 2 months) in both genotypes (Fig. 3a). Surprisingly, the plasma oxytocin concentration in Cd38−/− pups at 1–3 weeks of age was not decreased, but was similar to that in Cd38+/+ mice of the same age. However, as expected, at 2 months of age (young adult) after weaning, we found a significantly lower plasma concentration of oxytocin in Cd38−/− than in Cd38+/+ mice (28).

Figure 3.

 Plasma concentrations of oxytocin (a) and ADP-ribosyl cyclase activity in the hypothalamus (b) during development in Cd38−/− (circle) and Cd38+/+ (triangle) mice. Bar indicates the weaning period. SEM are within symbols. n = 5–11. *P < 0.01 from Cd38−/−.

Critical period of plasma oxytocin levels in juvenile to adult stages

The above decrease in oxytocin concentration after weaning during development observed only in Cd38−/− mice suggests an important critical period to distinguish different plasma oxytocin switching from the juvenile stage to the adult stage. Because breast milk is the only food source in lactating pups, we speculated that oxytocin is inevitably taken in from the breast milk. To confirm this, we performed quantitative analysis to determine whether oxytocin is present in the mammary glands of lactating dams and milk in pups. Milk curd was found in the stomachs of the offspring born to Cd38−/− females (Fig. 4), which was quite different from that in those born to Oxt−/− and Oxtr−/− females (23, 24). Oxytocin was abundant in the mammary gland tissue and breast milk in lactating dams of both genotypes, with no significant difference between Cd38+/+ and Cd38−/− animals (28). In the three stages of growth and development, the plasma oxytocin appears to be controlled from different sources. (i) Foetal stage: the foetus can obtain oxytocin from the placenta, which is linked with the mother, because oxytocin is transported to the foetus through the blood-placental-barrier (34). (ii) Infant stage (breastfeeding stage): the dams provides oxytocin to the pups from breast milk until weaning. Because milk is the only food source for pups during lactation, breastfeeding is the sole exogenous origin of plasma oxytocin. (iii) Adult stage (weaning stage): during this period, the plasma oxytocin is derived entirely from its own synthesis and secretion in the absence of exogenous supply. Oxytocin levels remained high in wild-type mice during all developmental stages. By contrast, oxytocin levels decreased significantly after weaning in CD38 knockout mice. These observations can explain why the Cd38−/− mice begin to show different pathological phenotypes at each stage.

Figure 4.

 Oxytocin levels in breast milk (a) and mammary gland (b) in Cd38+/+ and Cd38−/− lactating mice. n = 4–6. No significant difference.

ADP-ribosyl cyclase activity

In the central nervous system, ADP-ribosyl cyclase activity corresponding to CD38 was detected as early as embryonic day 15 in mouse development (35), and the endogenous brain cADPR content is higher in the developing brain and declines in adults (35). ADP-ribosyl cyclase activity was measured in the hypothalamus (Fig. 3b) and posterior pituitary (data not shown) in CD38 wild-type and knockout mice during the first 2 months of life.

In the hypothalamus, ADP-ribosyl cyclase activity in 1-week-old mice was 6% that in 2-month-old Cd38+/+ mice. Its activity in Cd38+/+ mice was significantly higher than that in age-matched knockout mice. From the second week of life, Cd38+/+ mice showed significantly higher levels of ADP-ribosyl cyclase activity in the hypothalamus, and the difference in comparison with Cd38−/− mice increased markedly (27, 28). In Cd38−/− mice, ADP-ribosyl cyclase activity in the hypothalamus decreased slightly with age.

ADP-ribosyl cyclase activity in the pituitary was lower than that in the hypothalamus in both Cd38+/+ and Cd38−/− mice (data not shown). Cd38−/− mice showed little or no increase in ADP-ribosyl cyclase activity in the pituitary during development.

The role of CD38 in regulation of oxytocin secretion through cADPR-mediated intracellular calcium signalling has been demonstrated in adult mice. Here, we found lower and similar levels of ADP-ribosyl cyclase activity were found in the hypothalamus and pituitary of both 1-week-old Cd38+/+ and Cd38−/− pups. On the basis of these observations, we speculated that the level of ADP-ribosyl cyclase activity is also relatively low at the foetal stage, suggesting that maintenance of plasma oxytocin relies mainly on the exogenous source from the placenta and breast milk. Because CD38 knockout pups suddenly lose their exogenous oxytocin supply after weaning and the intrinsic activity of ADP-ribosyl cyclase remains at a low level, the resulting relative shortage of endogenous oxytocin release during development of the adult stage thus affects the social recognition behaviour in adult animals.

Implications for developmental disorders

A series of recent studies suggested that oxytocin may be related to autism (3–8, 36–39). It has been reported that plasma oxytocin levels in autistic children are lower than those in age-matched normal controls, although the precise deviation is very small (35). Infusion of oxytocin reduces repetitive behaviours in adults with autistic and Asperger’s disorders (37). However, these studies focused largely on autism in older children and adults. There have been only a few studies in infants during breastfeeding and the early postnatal period; Tanoue and Oda (40) reported that children in their control group breast-fed significantly longer than autistic infants, and Fries et al. (41) provided evidence that oxytocin is critical for regulating social behaviour during early experience in infants. Although the precise mechanism is unclear, our data obtained in mice suggest that a lack of adequate exogenous oxytocin during the infant period would affect the normal development of the brain in genetically susceptible infants, thereby increasing the risk of autism.

Oxytocin treatment as a compensatory method has been proposed (42), and its use has begun in several hospitals, including Kanazawa and Krasnoyarsk State Medical University Hospitals (43). In some cases, we have observed improvements in social behaviour, such as increased eye contact and positive communication (43), as suggested based on human studies reported previously (10–13, 44, 45).

Recently, Ebstein et al. (46) suggested some interesting links between autism spectrum disorders (ASD) and CD38 in humans. Genetic polymorphisms in the vasopressin–oxytocin pathway, notably the arginine vasopressin receptor 1a, the oxytocin receptor, neurophysin I and II, and CD38, contribute to deficits in socialisation skills in ASD patients. They presented the first evidence that CD38 expression in lymphoblastoid cells derived from subjects diagnosed with autism can be correlated with social skill phenotype. Very recently, Ebstein’s group and our group found associations between two single nucleotide polymorphisms in CD38 and ASD subjects (43, Lerer E, Levi S, Israel S, Salomon S, Yirmiya N, Ebstein RP, manuscript in preparation) Further clinical trials of oxytocin treatment using such biomarkers in CD38 in patients with ASD are warranted.

Conclusions

Although the mechanism of by which milk oxytocin is transported into the brain through the blood-brain-barrier is unclear, it is undeniable that breastfeeding is not only important for infant survival, but also for infant brain development. Taken together with the observations of previous integrated studies (3–8, 20, 30, 47–50), we concluded that different sources of oxytocin appear to impact brain development at different stages of growth, and thus the maintenance of a high oxytocin level is secured by these two sources for development and social behaviour.

Acknowledgements

This work was supported by the bilateral grant 08-04-91209 from the Russian Foundation for Basic Research and the Japan Society for the Promotion of Science (RFBR-JSPS) (A.B.S., H.H.).

Ancillary