Introduction

Survival of mammalian newborns relies on the interaction they must ensure with the mother in order to find the nipples and suck. Olfaction plays a decisive role in this process since maternal odors efficiently guide the newborns towards the source of milk (nipples, teat, breast). Odor learning often (but not always, see below) contributes to this crucial attention and active behavioral response of neonates to maternal odor cues. It is for instance critical for rodent pups, who only rely on olfaction and somatosensory inputs to interact with the mother at birth1,2. Importantly, maternal odor learning is associated with a general state of arousal, i.e. of particular sensitivity to environmental stimuli. In humans for instance, labor contractions occurring during delivery provoke neonates’ arousal and facilitate odor learning: infants exposed to an odorant immediately after birth oriented preferentially to this odor several days later, whereas those exposed to the odorant 12 h post-partum did not3. In human neonates delivered by caesarean section and exposed to an odorant shortly after birth, only those exposed to labor contractions before delivery displayed preference for the odorant4.

Interestingly, when maternal odors are learned, they are generally learned pre- or postnatally through associative processes5,6. For instance, rat and mouse pups learn to associate maternal odors (corresponding to conditioned stimuli, CS) with maternal thermo-tactile cues (corresponding to unconditioned stimuli, US), displaying then approach behavior towards these odors initially inactive or weakly active on their behavior7,8. This conditioned response is analogous to the unconditioned one initially elicited by thermo-tactile cues. Most of the US involved in associative neonatal learning (such as milk delivery or stroking in rodents) trigger neonates’ arousal, and this learning is effective immediately after birth: rat pups can associate a neutral odor (CS) with an appetitive gustatory stimulus (milk, US) during the first two postnatal hours9. A similar odor learning has been described in humans, as 1-day-old infants can associate a neutral odor (CS) and a tactile stimulation (stroking, US) and then display a positive motor response to the odor10.

The noradrenergic system is known for driving arousal and is therefore crucial for learning11. For instance, human neonates aroused by labor contractions before delivery by cesarean section have higher plasma noradrenaline levels at birth and display better odor learning performances than those not exposed to contractions4. In associative odor conditioning, the arousal induced by the reinforcer (US) is also mediated by noradrenaline release. Indeed, systemic blockade of β-adrenoceptors, the main receptors for noradrenaline, impairs acquisition of stroking-induced appetitive odor learning in rat and mouse pups12,13,14. The neural circuits on which noradrenaline is acting to underlie this learning have been well described in rodents. In particular, US induces noradrenaline release from the locus coeruleus, a small brainstem nucleus15,16, into olfactory structures such as the olfactory bulb (OB), the brain region receiving direct inputs from olfactory sensory neurons and olfactory receptors, and the anterior piriform cortex (aPC)17, the main olfactory cortex receiving information directly from the OB18. In the OB, activation of β-adrenoceptors enhances the activation of excitatory mitral cells by directly activating them, or indirectly by inhibiting the surrounding interneurons19,20. Activation of β-adrenoceptors concomitantly to odor presentation (CS) in either the OB or the aPC is necessary and sufficient for neonatal odor learning21,22.

The newborn rabbit provides a very relevant model to further study neonatal odor learning. Like other mammalian newborns, rabbit pups must rapidly find nipples and ingest milk to survive. This vital challenge is intensified by the once-per-day only, and dramatically brief (3–4 min), visit of the mother to the nest23. In this context, missing one sucking occasion, i.e., one nursing episode, threatens pups’ survival and missing two causes death24. As altricial mammals, they do not see and hear before the second postnatal week25 and substantially depend on olfaction to interact with, and adapt to, the environment. Unique among mammalian species, in the current state of knowledge, lactating rabbit females emit a monomolecular signal in their milk – an aldehyde (2-methylbut-2-enal) – carrying the properties of a pheromone and named the mammary pheromone (MP)26,27,28. The MP presents two complementary functions. First, it acts spontaneously, i.e., independently of any perinatal learning, as a behavioral releaser triggering a strong arousal and the typical nipple-searching behavior in newborns (rapid and stereotyped orocephalic movements), allowing them to find the nipples in less than 15 seconds after the entrance of the mother in the nest29. Second, the MP acts as a reinforcer US promoting appetitive learning of odorants in an exceptionally efficient way: a single and short (5 min-long) simultaneous pairing between a neutral odor CS and the MP is sufficient for the CS to subsequently trigger the typical orocephalic movements in newborn rabbits30. CS responding is optimal 24 h after the CS-MP pairing and persists for several days, demonstrating consolidation of a long-term CS memory31. This remarkably brief, selective and robust associative pheromone-induced memory offers considerable opportunity to examine the brain circuits early engaged by odor learning.

However, only scarce data are available regarding the neural mechanisms of this pheromone-induced odor learning. To date, brain activation patterns induced by MP processing32,33 and odor learning34,35,36 have been evidenced by expression of the immediate early gene c-Fos. For instance, exposure at postnatal day 4 to MP-learned odorant or to the MP itself are associated with an increased c-Fos expression (compared to control conditions) in the OB and aPC. However, no study has yet examined the neurotransmitters involved in odor learning in rabbit pups.

In the present study, we investigated the role of the noradrenergic system in MP-induced odor conditioning in newborn rabbits, combining behavioral, pharmacological, and molecular approaches. We hypothesized that the MP might recruit noradrenaline release in a similar way that a thermo-tactile or milk US does in rodent pups, producing odor learning in presence of a new odorant. Thus, we examined the behavioral consequences of an intraperitoneal injection of propranolol, a β1- and β2-adrenoceptors antagonist known to cross the blood-brain barrier when injected systemically, on MP-induced conditioning, as it has previously been shown to block acquisition of appetitive odor learning in rodent pups12,13,14. Then, to better understand the effect of Propranolol on MP-induced conditioning at the neurobiological level, we quantified the c-Fos gene expression in OB, aPC and hippocampus following Propranolol injection and/or odor learning. We focused our analysis on the OB and aPC because of their known noradrenaline modulation in rodent newborn odor learning1,22, but also on the hippocampus, which is associated with adult odor memory37. Finally, since several learning tasks induced changes in neurotransmitter receptor expression in rats and birds38,39,40,41, we evaluated a potential impact of Propranolol and/or odor learning on β1- and β2-adrenoceptor genes expression in the same brain regions.

Results

The noradrenergic system is required for pheromone-induced odor learning in pups

In order to determine whether the noradrenergic system was involved in neonatal odor learning promoted by the mammary pheromone (MP), rabbit pups were injected intraperitoneally on day 2 with either the β-adrenoceptors antagonist propranolol at 10 mg/kg (n = 8) or 20 mg/kg (n = 8), or a saline solution (Vehicle, n = 7), 30 min before exposure to odorant A+MP mixture. Pups were tested on day 3 for their responsiveness to odorant A, MP and odorant B, a neutral odorant never paired with the MP, used as a negative control (Figure 1A). As expected, pups injected with saline responded highly and similarly to A and MP, but not to B (A vs. B vs. MP: Q = 14, P < 0.005; B vs. A or B vs. MP: χ2 = 5.1, P < 0.05). The conditioning was thus effective and selective, i.e. triggering a response to odorant A only (Figure 1B). In contrast, pups injected with propranolol at 10 mg/kg or 20 mg/kg responded only to MP, without responding to A or B (A vs. B vs. MP: Q > 14, P < 0.005; A vs. MP: χ2 ≥ 5.1, P < 0.05; B vs. MP: χ2 = 6.1, P < 0.05). Importantly, responses to A in both groups injected with propranolol differed from Vehicle-treated group (between group comparisons: χ2 > 11, P < 0.001) indicating that the conditioning was blocked. Moreover, the proportions of responses to the odorants were similar after 10 or 20 mg/kg propranolol injection (between group comparisons: χ2 < 0.05, P = 1) suggesting that these doses were equally effective in blocking odor learning.

Fig. 1
figure 1

Propranolol blocks acquisition, but not memory consolidation or retrieval, of pheromone-induced odor conditioning in rabbit pups. (A) Rabbit pups were intraperitoneally injected with saline (vehicle) or 10 or 20 mg/kg propranolol on day 2 then conditioned 30 min later to odorant A by association with the mammary pheromone (MP). (B) The histogram represents the proportion of pups responding behaviorally to odorants A, B (novel odor representing negative control) and to MP (positive control) on day 3. (C) Pups were intraperitoneally injected immediately after MP-conditioning to odorant A with saline (vehicle) or 10 mg/kg propranolol on day 2 and (D) tested for their responsiveness to A, B and MP on day 3. (E) Pups were conditioned to odorant A with the MP on day 2 then intraperitoneally injected 24 hours later (day 3) with 10 mg/kg propranolol 30 min before retrieval test (F) evaluating their responsiveness to odorants A, B and MP. *p < 0.05: different from response to MP within the same group (χ2 test of McNemar). #p < 0.05: different from response to A in Vehicle group (χ2 test of Pearson).

Full size image

To evaluate whether the noradrenergic system played a role in the consolidation process of reinforced memory, propranolol was injected immediately after exposure to A+MP mixture (Figure 1C). In this case the injection of propranolol at 10 mg/kg was ineffective (Figure 1D): saline and propranolol-injected pups (n= 6 and 8, respectively) responded highly to A and MP, but not to B (A vs. B vs. MP: Q ≥ 14, P < 0.005; B vs. A or B vs. MP: χ2 ≥ 5.1, P < 0.05), with the same efficiency in both groups (between group comparisons: χ2 < 0.05, P = 1). Thus, activation of β-adrenoceptors is necessary for MP-induced odor learning but not for post-acquisition process, i.e. consolidation, in rabbit pups.

To determine whether the noradrenergic system had an effect on retrieval processes of MP-conditioned odor memory, propranolol was injected 30 min before the test. Pups (n = 7) were conditioned on day 2 to A by pairing it with MP, then injected 24 h after the conditioning, i.e. on day 3, with 10 mg/kg propranolol (since this concentration was sufficient to block the response to A when injected before conditioning; Figure 1B) and tested 30 min after injection (Figure 1E). The pups responded highly and similarly to A and MP without responding to B (A vs. B vs. MP: Q = 12.3, P < 0.005; B vs. A: χ2 = 4.2, P < 0.05; B vs. MP: χ2 = 5.1, P < 0.05; Figure 1F), demonstrating that the conditioning was effective and specific to odorant A. This result demonstrates that propranolol had no effect on retrieval and therefore on odor perception, stressing that the effect of propranolol injected before conditioning (Figure 1A-B) was due to a specific impairment of MP-induced olfactory learning.

Propranolol impedes odor learning-induced activation of the olfactory bulb and the anterior piriform cortex, but not of the hippocampus

To further understand the neurobiological basis supporting MP-induced odor learning and its modulation by the noradrenergic system, we evaluated the effect of conditioning and β-adrenoceptors blockade on neuronal activation of three brain structures important for olfactory learning, namely the olfactory bulb (OB), the anterior piriform cortex (aPC) and the hippocampus. For this purpose, rabbit pups were injected intraperitoneally on day 2 with 10 mg/kg propranolol or a saline solution (Figure 2A). Some of the pups were left undisturbed in the nest for 65 min before being euthanized (control groups; saline: n = 6; propranolol, n = 6) whereas other were conditioned to A+MP (for 5 min) 30 min after the injection and were euthanized 30 min after conditioning (A+MP groups; saline: n = 10; propranolol, n = 10). Brains were then collected and quantifications of c-Fos gene expression, as a proxy of neuronal activation, were performed in the OB, the aPC and the hippocampus.

Fig. 2
figure 2

Conditioning-induced brain activation is blocked by propranolol in the olfactory bulb and anterior piriform cortex, but not in the hippocampus. Rabbit pups were intraperitoneally injected on day 2 with saline or with 10 mg/kg propranolol 30 min before being conditioned to odorant A + MP (A,B) or before being exposed to MP alone (C,D). Control pups were injected but left undisturbed. All pups were euthanized 65 min after the injection. (B, D) Histograms represent the relative quantification (% of saline-injected control group) of c-Fos mRNA in the olfactory bulb (left), the anterior piriform cortex (middle) and the hippocampus (right). All data are presented as mean ± SEM.(B) (olfactory bulb and piriform cortex): *p < 0.05, ****p < 0.0001; difference between saline-conditioned group and the other 3 groups (two-way ANOVA, significant drug x conditioning interaction, followed by post-hoc Tukey’s HSD test). (B) (hippocampus), (D) (olfactory bulb and piriform cortex): *p < 0.05, **p < 0.01, ****p < 0.0001; difference between control groups and A+MP (B) or MP (D) groups (two-way ANOVA, significant exposure effect).

Full size image

After exposure to A+MP, two-way ANOVA revealed in both the OB and the aPC no drug effect [comparison Saline vs Propranolol; OB: F(1,28) = 3.4, p = 0.07; aPC: F(1,22) = 1, p = 0.3], a significant effect of conditioning [comparison control vs A+MP; OB,: F(1,28) = 70.8, p < 0.0001; aPC: F(1,22) = 6.1, p = 0.02] and, more importantly, a significant interaction between conditioning and drug effects [OB: F(1,28) = 4.6, p = 0.04; aPC: F(1,22) = 4.1, p = 0.05] on c-Fos expression. This indicates that propranolol effect is influenced by conditioning. Post-hoc analyses indicated that c-Fos mRNA was significantly higher in the OB of conditioned pups injected with saline compared to control pups injected with either saline or propranolol (p < 0.0001; Figure 2B). c-Fos mRNA was also significantly higher in the aPC of conditioned pups injected with saline compared to control pups injected with saline (p = 0.02; Figure 2B). These results indicate that OB and aPC were both activated during MP-induced odor learning. Importantly, propranolol injection did not by itself modify c-Fos expression in control pups (OB: p = 1; aPC: p= 0.9), but significantly decreased it in the OB of conditioned pups (p = 0.01; Figure 2B). Moreover, in the OB, propranolol did not completely suppress the conditioning effect as propranolol-injected conditioned pups showed higher c-Fos expression than saline- or propranolol-injected control pups (p < 0.05), indicating that blocking the noradrenergic system attenuated (but did not abolish) odor learning-induced OB activation. In the aPC, no difference was found between conditioned pups injected with propranolol and both control pups (p > 0.7), indicating that propranolol abolished the learning-induced aPC activation. In the hippocampus, two-way ANOVA revealed a significant effect of conditioning [F(1,21) = 9.2, p = 0.006] on c-Fos mRNA level, revealing a higher c-Fos expression in conditioned pups compared to control pups (Figure 2B, right). However, no effect of drug or drug x conditioning interaction [F(1,21) < 0.1, p > 0.7] was found, demonstrating that propranolol did not modify basal or learning-induced hippocampal activation. Therefore, the blocking effect of propranolol on pheromone-induced odor learning in rabbit pups was associated with a decrease of neural activation in the OB and the aPC, but not in the hippocampus.

We then assessed whether the higher neuronal activation after A+MP exposure and its noradrenergic modulation are specific to conditioning and not related to PM exposure only. Rabbit pups were injected intraperitoneally on day 2 with 10 mg/kg propranolol or a saline solution, then 30 min later they were exposed to MP alone (for 5 min) and were euthanized 30 min later (MP groups; Saline: n = 7; Propranolol, n = 8; Figure 2C). Control pups were left undisturbed in the nest after injection for 65 min before being euthanized (Saline: n = 5; Propranolol, n = 5). After exposure to MP alone, only a significant effect of MP exposure on c-Fos mRNA level was evidenced in the OB and the aPC [OB: F(1,21) = 58.2, p < 0.0001; aPC: F(1,21) = 6.2, p = 0.02] revealing a higher c-Fos expression in MP exposed pups compared to control pups (Figure 2D). However, no effect of drug or drug x MP interaction [OB: F(1,21) < 3.6, p > 0.07; aPC: F(1,21) < 0.35, p > 0.5] was found, demonstrating that propranolol did not modify MP-induced OB or aPC activation. Interestingly, MP did not affect hippocampal activation as evidence by an absence of effect of conditioning, drug or drug x conditioning interaction [F(1,20) < 0.1, p > 0.7]. Moreover, exposure to odorant A alone or A + B mixture did not affect c-Fos expression (Control: n= 5; Exposed to A or A+B: n= 10; Supplementary Figure 1). Altogether, these results indicate that activation of β-adrenoceptors is specifically involved in conditioning-induced OB and aPC activation.

The effects of odor learning and propranolol do not involve a change in β-adrenoceptors expression

We also evaluated if the effects of odor conditioning and propranolol on brain activations were associated with a modification of β-adrenoceptors genes expression in our target brain structures (control groups; Saline: n = 4–6; Propranolol, n = 6; A+MP groups; Saline: n = 7–10; Propranolol, n = 7–10; Figure 3A). β1-adrenoceptor (ARDB1) mRNA level was not affected by conditioning, drug or drug x conditioning interaction in either the OB [F(1,28) < 2.3, p > 0.1], the aPC [F(1,20) < 2.2, p > 0.1] or the hippocampus [F(1,21) < 1.6, p > 0.2; Figure 3B]. For β2-adrenoceptor (ARDB2), no effect of conditioning, drug or interaction were found in the OB [F(1,28) < 2.7, p > 0.1] and the aPC [F(1,20) < 1.7, p > 0.2; Figure 3C]. In the hippocampus, β2-adrenoceptor mRNA level was decreased by conditioning [F(1,21) = 18.2, p < 0.0005; Figure 3C] without any effect of drug or interaction [F(1,21) < 2, p > 0.1].

Fig. 3
figure 3

Neither MP-induced conditioning nor propranolol induces a general change in β-adrenoceptors expression. (A) Rabbit pups were intraperitoneally injected on day 2 with saline or with 10 mg/kg propranolol, then conditioned to odorant A + MP or not (control) before being euthanized 65 min after the injection. (B, C) Histograms represent the relative quantification (% of saline-injected control group) of Adrb1 (B) and Adrb2 (C) mRNA in the olfactory bulb (left), the anterior piriform cortex (middle) and the hippocampus (right). All data are presented as mean ± SEM. ***p < 0.0005: difference between control and A+MP groups (two-way ANOVA, significant conditioning effect).

Full size image

The absence of effect of odor conditioning and propranolol on β1 or β2-adrenoceptors expression in OB and aPC indicates that the effect of propranolol on conditioning-induced c-Fos expression in these brain structures (Figure 2B) was not related to a change of β-adrenoceptors expression.

Discussion

The present study aimed to identify the role of the noradrenergic system in the establishment of neonatal odor learning in an ecologically and ethologically relevant model, the newborn rabbit. This species provides an opportunity to decipher the neural mechanisms that support a remarkable form of single-trial appetitive learning functional early in life, i.e. odor learning promoted by the mammary pheromone (MP) naturally emitted by rabbit mothers, and taking place during the daily nursing session typical of this species.

According to literature on newborn humans and rodents, the noradrenergic system appears to play a crucial role in neonatal odor learning4,12,13. Notably, previous studies in rat and mouse pups have shown that the β-adrenoceptor antagonist propranolol blocks associative odor learning reinforced by tactile stimulation used as a US12,13,14. Strikingly, we demonstrated here that an intraperitoneal injection of 10 or 20 mg/kg propranolol blocked a form of odor-odor (CS-US) associative conditioning, i.e. the MP-induced conditioning to a neutral odorant, leading to long-term memory in 2-day-old rabbit neonates. Altogether, these results highlight the crucial role of noradrenaline (NA) in the learning system of newborn mammals.

In order to decipher which steps of memory formation - acquisition or post-acquisition (i.e. consolidation) - is affected by β-adrenoceptors blockade, we evaluated the effect of propranolol injection before and immediately after MP-induced conditioning. Propranolol blocked odor conditioning only when injected before odorant + MP exposure, stressing an effect on odor acquisition rather than consolidation. Interestingly propranolol did not affect memory retrieval of MP-conditioned pups. These results discard any perceptual, motor, motivational or attentional deficits induced by propranolol, and are consistent with previous studies in adult rabbits showing that propranolol did not alter odor detection42. Our results indicate that propranolol induced a specific deficit in odor learning. This corroborates previous studies demonstrating that propranolol disturbs memory acquisition in diverse appetitive and aversive learning tasks in both humans and non-human animal models (e.g.43,44,45,46,47,).

Analyzing gene expression, we found that odor conditioning in rabbit neonates enhanced c-Fos expression – revealing neural activation – in the olfactory bulb (OB), the anterior piriform cortex (aPC) and the hippocampus compared to basal conditions. These results highlight that these brain regions are involved in MP-induced learning of novel odors in newborn rabbits, as previously indicated for OB and aPC using immunohistochemical studies for Fos protein expression34,35,36. Remarkably, propranolol had no effect on basal or MP-induced c-Fos expression in OB, aPC and hippocampus of rabbit pups, but it did block conditioning-induced c-Fos expression in OB and aPC, without affecting hippocampus. This suggests that propranolol impairs MP-induced odor learning at least in part by blocking neural activation in the OB and aPC. In contrast, propranolol effects on learning do not appear to depend on the hippocampus.

The results mentioned above are consistent with rodent literature, which demonstrates that neonatal odor learning is dependent on β-adrenoceptors activation on mitral cells and interneurons in the OB19,20,21 and the aPC22. In order to determine whether noradrenergic modulation of OB and aPC is necessary for learning in rabbit pups, it would now be valuable to infuse locally propranolol directly into OB or aPC. To date, such intracerebral infusions have never been conducted in rabbit neonates, whose brains mostly develop perinatally48.

Previous studies indicate that learning can induce changes in neurotransmitter receptors expression in rats and birds38,39,40,41. We therefore evaluated whether the MP promotes learning by modifying β-adrenoceptors expression and if propranolol impairs learning through a change in β-adrenoceptors expression. However, the present results showed that there was no effect of conditioning or propranolol on β-adrenoceptors expression in OB and aPC, therefore disproving the assumption that changes in β-adrenoceptors expression in these structures could be responsible for the behavioral and neural (c-Fos in OB and aPC) effects of propranolol on conditioning. Therefore, we propose that the noradrenergic system mediates MP-induced odor learning by modulation of OB and aPC activity through local β-adrenoceptors signaling but without modifying β-adrenoceptors expression.

In rodents, newborn odor learning is known to be supported by a unique neural circuit differing from the adult one. Notably, the hippocampus is still developing and not sufficiently functional to participate in odor learning in neonates1,49. Conversely, we showed here that MP-induced conditioning, but not exposure to MP alone, enhanced hippocampal activation. This suggests that the hippocampus is, at least partly, functional in newborn rabbits, despite being not (yet) sensitive to β-adrenoceptors modulation. This singularity is not clear but could be related to the ecology of the rabbit, which has a longer gestation than rodents potentially allowing some brain areas, such as the hippocampus, to appear more mature postnatally.

Overall, the present and previous results suggest that reinforcers from different sensory modalities, such as the MP in newborn rabbits and thermo-tactile stimulation in newborn rodents, may operate via similar mechanisms to promote the learning of new and initially neutral odors in newborn mammals, which is crucial for their immediate and progressive adaptation to their environment. In rabbit pups, the MP would stimulate norepinephrine release from the locus coeruleus, activating β1- and β2-adrenoceptors in the OB and aPC. This NA modulation occurring alongside neutral odor presentation would favor odor learning. Interestingly, neonates (including rabbit pups) also display some forms of incidental odor learning without initially triggering strong arousal, such as sensory preconditioning50,51. Therefore, it could be assumed that this incidental odor learning does not depend on the noradrenergic system. However, the effects of propranolol on higher-order conditioning remains to be evaluated in young mammals. The newborn rabbit is a highly suitable model for such exploration.

Methods

Animals and housing conditions

New Zealand x Californian adult male and female European rabbits (Oryctolagus cuniculus) were housed in individual cages at the breeding unit of Lyon Neurosciences Research Center. A nest-box was added on the outside of the females’ cages three days before parturition. Day of parturition was considered as day 0. Then, females had access to their nest every day between 11:30 a.m. and 01:00 p.m., which allowed them to nurse according to the daily rhythm of the species, and to equalized pups’ nursing experience23. Animals were kept under constant 12:12 light:dark cycle (light on at 07:00 a.m.) with ambient air temperature maintained between 16–21°C. Water and pelleted food were provided ad libitum. In total, 116 pups (from 24 litters) were used for the experiments (44 for pharmacology and behavior and 72 for pharmacology and biochemistry). All experiments were performed in accordance with ethical rules enforced by French law and approved by the ethical committees of Lyon 1 University (CEEA-42 and CEEA-55) and the French Ministry of Higher Education and Research under APAFIS number #27874–2020110416356847 v2. The present study was performed in accordance with ARRIVE guidelines (https://arriveguidelines.org).

Odorants

The mammary pheromone (MP; 2-methylbut-2-enal; Sigma-Aldrich, Saint-Quentin Fallavier, France) was used as the unconditioned stimulus in odor pairing (day2), and as control stimulus eliciting the typical orocephalic movements in pups in behavioral testing (day 3). Ethyl-isobutyrate (odorant A; Sigma-Aldrich) was used as a conditioned stimulus in odor pairing and testing, and ethyl maltol (odorant B; Sigma-Aldrich) as a control stimulus in testing. The odorants and mixtures of odorants (50/50 v/v ratio) were prepared in distilled water at 10−5 g.mL−1 (concentration previously shown as efficient in MP-induced conditioning31). These odorants were chosen because they have previously been used to study olfactory memory in newborn rabbits and are known to be easily learned by them27.

Pharmacological treatment and odor conditioning

All experiments occurred before nursing to minimize the impact of satiation on pups’ responsiveness and to equalize their motivation52. A maximum of 5 pups from a same litter were used per group to reduce litter effect. Between 9:15 and 10:30 a.m. of postnatal day 2, pups were transferred in a box from their nest to an experimental room. They were individually marked with low-odor ink and weighted. To test the role of the noradrenergic in the memory formation of odor learning promoted by the mammary pheromone (MP), in a first experiment, pups were intraperitoneally injected (0.1 mL/50g) with two different doses of propranolol (10 or 20 mg/kg; Sigma-Aldrich) diluted in 0.9% saline solution. Propranolol is a β1- and β2-adrenoceptors antagonist known to cross the blood-brain barrier when injected systemically53. Doses were chosen based on the literature in rodent pups12,13,14,21. Control pups received 0.1 mL/50g of 0.9% saline solution (vehicle). Pups returned to their nest for 30 min, then odor conditioning was conducted following validated procedures (e.g.,31): pups were transferred to the experimental room and exposed to a cotton pad scented with 8 mL of odorant A+MP mixture and held 1 cm above the pups for 5 min. Two minutes after the end of conditioning, pups returned to their nest and the box was cleaned with alcohol and water. In a second experiment, to determine whether propranolol had an effect on consolidation processes, some pups were conditioned to odorant A on day 2 and immediately injected with either propranolol (10 mg/kg) or vehicle. They were tested on day 3 following the same procedures (see below). In a third set of experiment, to determine whether propranolol had an effect on retrieval processes or odorant perception, some pups were conditioned to odorant A on day 2 and were injected with propranolol (10 mg/kg) 30 minutes before being tested on day 3. In all those experiments, pups were weighted after nursing on days 2 and 3 to monitor any effect of propranolol injection on milk intake.

Behavioral testing

Pups were tested 24 h after conditioning, i.e. on day 3, in an oral activation test classically used in newborn rabbits27. Each pup was individually immobilized in one gloved hand of the experimenter, the head of the pup being left free. The odor stimulus was presented for 10 sec with a glass rod placed 0.5 cm in front of the nares. The dependent variable was the proportion of pups exhibiting a positive response (i.e., successful conditioning) defined by typical head-searching movements (stretching towards the rod and vigorous vertical and horizontal scanning movements) usually followed by oral grasping of the rod27. Non-responding pups showed no response except sniffing. Each pup from a litter was successively tested for its responsiveness to odorants A, B and to the MP with an inter-trial interval of 60 sec. Stimuli presentation order was counterbalanced except for the MP which was always presented in last (positive control).

RNA extraction and cDNA synthesis

Rabbit pups were deeply anesthetized with pentobarbital (Exagon,100mg/kg,i.p.) and decapitated for brain extraction 65 min after intraperitoneal injection (Saline or Propranolol). Control pups were not exposed to any odorant after the Saline or Propranolol injection. The other pups were exposed for 5 min to either odorant A + MP (conditioning), to MP alone or to neutral odorant (A or A + B) alone 30 min after intraperitoneal injection. They were euthanized 30 min after odor exposure, i.e. 65 min after injection.

During brain extraction, the OB was immediately separated from the rest of the brain, then OB and brain were frozen in liquid nitrogen and stored at −80°C (Conditioning: cont sal n=6, cont prop n=6; A+MP sal n= 10; A+MP prop n= 10; MP alone: cont sal n=5, cont prop n=5; MP sal n= 7; MP prop n= 8; neutral odorant: cont n= 5, exposed n= 10). The aPC and hippocampus were bilaterally punched from frozen brains using a 1.0 mm Rapid-Core Punch (Ted Pella, Inc., Redding CA, USA), thanks to a recent brain atlas of the 4-day-old rabbit54. Due to technical problems, some brains were not punched for aPC (n = 6: A+MP sal n= 3, A+MP prop n= 3) and hippocampus (n = 8: cont sal n=2, A+MP sal n= 3, A+MP prop n= 2; MP prop n=1), respectively. Total RNA was extracted from the OB, aPC and hippocampus with TRIzol® reagent (Invitrogen, Cergy Pontoise, France). Brain tissue was homogenized in 500 µL of Trizol with a Tissue Lyser (Qiagen, Hilden, Germany) during 1 min at 30 Hz, then 100 µL of chloroform was added to separate the organic and watery phases. After 10-min centrifugation at 10 000 g and 4 °C, the upper water phase containing total RNAs was collected and the organic phase containing DNA and proteins was stored at −20°C. RNAs were precipitated by adding 12.5 µL of glycogen (20 mg/mL) and 200 µL of isopropanol. After 10-min centrifugations at 10 000 g and 4 °C, RNA pellet was successively washed and centrifuged with ethanol (70 or 100 %) or sodium acetate (3 M). Finally, RNAs were dissolved in 20 µL of sterile water. RNA quantity and purity were determined using a Nanodrop spectrophotometer (Thermo Scientific, Waltham, USA) and RNA quality was assessed on RNA 6000 Nano Kit with a 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). 1 µg of total RNA was reverse transcribed with Superscript IV (Invitrogen) in 20 µL following manufacturer’s instructions. Due to low RNA quantity, 2 aPC samples of control group injected with saline were not used for adrb1 and adrb2 PCR.

Gene expression analysis by qPCR

10 µL final quantitative Polymerase Chain Reaction (qPCR) was performed with a LightCycler 480 (Roche Diagnostics, Rotkreuz, Switzerland) using ONEgreen qPCR fast premix (Ozyme, Saint-Cyr-l’Ecole, France). For each target and reference gene, primer couples were designed using PrimerExpress software (Applied Biosystem, Waltham, USA). 100% amplification efficiency of the qPCR reaction was determined based on the slope of a cDNA serial dilution curve (0.6 to 5 ng/µL) obtained for each primer pair, run in duplicate. PCR was realized following the manufacturer’s protocol in a final volume of 10 µL containing 8 µL of 2X ONEgreen premix and 300 nM of each primer, and 2 µL of each cDNA sample (2.5 ng/µL) or sterile water (negative control) in duplicate for each sample. Amplified products specificity was verified by melting curve analysis. Relative quantification analysis was performed on GenEx software (MultiD Analyses AB, Göteborg, Sweden). Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) housekeeping gene was determined as a reference gene since its expression level was constant in all experimental conditions.

Statistical analyses

When the same pups were tested to different stimuli, proportions of responding pups were compared by the Cochran’s Q test (global comparison) then the χ2 test of McNemar (comparison between 2 conditions). When different groups of pups were tested to the same stimulus, proportions of responding pups were compared between groups using the χ2 test of Pearson. For RT-qPCR analyses, data were analyzed using GraphPad Prism software (San Diego, CA, USA). Gene expression comparisons between groups were performed using two-way ANOVA to examine drug effect (Saline vs. Propranolol injection), exposure effect (control vs. A+MP or control vs MP) and drug x exposure interaction. Post-hoc Tukey’s HSD tests were conducted when interaction was significant. Differences were considered significant if p < 0.05.