The Long-Term Behavioural Effects of Maternal Creatine Supplementation in a Spiny Mouse Model of Birth Asphyxia

Birth asphyxia remains a leading cause of morbidity and mortality for newborns globally [1]. Birth asphyxia is defined by an acute episode of insufficient oxygen delivery to tissues from blood, leading to systemic hypoxaemia and hypercapnia. Clinical features that arise after an episode of birth asphyxia include symptoms of depressed levels of consciousness, poor muscle tone, and movement, and in some cases, the development of neonatal hypoxic-ischaemic encephalopathy (HIE) [2]. Survivors of HIE can experience varying degrees of acute and/or long-term neurological impairment, such as difficulties with learning and working memory, and significant neurological and motor deficits, such as cerebral palsy [3]. Even in the era of therapeutic hypothermia, children with HIE have greater neurodevelopmental delays in early childhood and in adolescence [4, 5]. As birth asphyxia events are highly unpredictable and difficult to detect, prevention instead of treatment could prove to be more beneficial in reducing HIE-related morbidity and mortality.

Dietary creatine supplementation during pregnancy is a potential prophylactic treatment strategy for HIE [6]. Creatine, and its phosphorylated form phosphocreatine, are nitrogenous guanidine compounds involved in the spatial and temporal buffering of cellular adenosine triphosphate [7]. Creatine synthesising enzymes and the creatine transporter are expressed within the brain [8], suggesting that creatine is relevant for energy homeostasis for brain function [9]. Increasing cellular creatine and phosphocreatine stores via dietary supplementation can prolong the provision of high-energy phosphate during hypoxia, thereby sustaining redox homeostasis to limit energy failure and support cellular survival [10, 11]. Indeed, in preclinical animal models, antenatal creatine supplementation has been linked to cerebral histopathological and metabolic neuroprotection after hypoxic injury [12‒15]. Whether these cellular protective mechanisms equate to improvements in functional behaviour such as cognition and motor activity following hypoxia requires investigation.

In addition, the long-term neurological effects of antenatal creatine supplementation on normal offspring have not yet been characterized, including the potential effect that antenatal exposure to high creatine concentrations may have on the immature brain and subsequent behaviour. For example, the neurological consequence of increased myelin in white matter tracts observed in late-gestation foetal sheep supplemented with creatine suggests potential influences on motor, sensory, cognitive, and behavioural function [16]. Furthermore, increased excitability and enhanced long-term potentiation of the hippocampus in rat offspring at postnatal day (PND) 14–21 and in adults born to dams supplemented with creatine could suggest enhancements in learning and memory [17, 18]. The use of antenatal creatine supplementation is supported by evidence of maternal, placental, and foetal creatine metabolic changes and needs during pregnancy [19‒22]. As we continue to understand the safety and tolerance of oral creatine in human pregnancy (ACTRN12620001373965), there remains a need for further preclinical translational investigations into the long-term consequences of increased antenatal creatine exposure on the offspring.

The primary aim of the present study was to examine offspring behaviour in a spiny mouse model of birth asphyxia and assess whether maternal creatine supplementation could mitigate any identified deficits linked to acute intrapartum hypoxia. Spiny mice, compared to conventional laboratory rodents, are precocial at birth with brain development at birth similar to that of a term human newborn, evidenced by more mature postnatal hippocampal and cerebral cortex growth, and myelination [23‒25]. As HIE typically results in selective damage to the sensorimotor cortex, basal ganglia and thalamus, regions involved in sensory input and voluntary motor control [26], we conducted behavioural tests that assess exploratory locomotor activity, anxiety and non-spatial memory, and learning behaviour. Creatine crosses the spiny mouse placenta with 5% w/w maternal creatine supplementation from mid-gestation increasing foetal cerebral total creatine concentrations by 4% by term [27]. Thus, a secondary aim of this study was to assess the effects of maternal creatine supplementation and antenatal creatine exposure on the behavioural characteristics of healthy offspring not exposed to birth asphyxia.

The experimentation for this study was approved in advance by Monash University Animal Ethics Committee and was performed in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. The experiments are reported in accordance with the ARRIVE guidelines for reporting animal research. Spiny mice (Acomys dimidiatus) from our in-house breeding colony were time-mated, bred and housed in a temperature-controlled room (25 ± 1°C, humidity 40–50%) and under a 12-h light/dark cycle (07:00 and 19:00) as previously described [28].

Maternal dietary creatine intervention and birth asphyxia methods were conducted as previously detailed [27]. Briefly, from day 20 of gestation (mid-gestation; term is 39 days), pregnant dams were randomly assigned to either the control treatment group that remained on a diet of standard rodent chow (2.16 mg creatine/g), or the experimental treatment group that was fed isocaloric pellets supplemented with 5% w/w creatine monohydrate (32.44 mg creatine/g) (Specialty Feeds, Perth, Australia) (control, n = 33; creatine, n = 15). Dams accessed allocated diet and water ad libitum.

Dams were then further randomly allocated to undergo either a caesarean section (C-section; control) or birth asphyxia (dam numbers: control, n = 18; control asphyxia, n = 15; creatine, n = 8; creatine asphyxia, n = 7). On day 38 of gestation, pregnant dams were killed by cervical dislocation and the abdomen excised. Pups from the C-section group were delivered immediately, and resuscitation was performed by palpitations of the chest using cotton buds. Pups allocated to the birth asphyxia group remained within the uterus in which the uterine horns and cervix were immediately tied with 5-0 surgical silk. The pregnant uterus was then excised and placed in a sterile saline bath at 37°C for 7.5–8 min. The pups were then delivered and resuscitation was performed by palpitations of the chest using cotton buds. Pups were pronounced dead if spontaneous breathing was not achieved after 5 min of attempted resuscitation. For all surviving pups, the placenta was removed 15 min after successful spontaneous breathing and rhythmic breathing was observed.

All surviving pups within a litter were cross-fostered to another dam that had remained on a control diet and had delivered a litter of comparable size within the preceding 12–24 h. Natural pups of the cross-foster dam were removed (killed for use as part of other studies) to ensure adequate lactation and to maximize neonatal care. After ∼35 days of PND age, pups were weaned and housed in same-sex groups of 2–5 animals from the same maternal treatment and intervention groups: control (normal maternal diet and C-section; n = 41), control asphyxia (normal maternal diet and birth asphyxia; n = 28), creatine (maternal creatine diet and C-section; n = 18), or creatine asphyxia (maternal creatine diet and birth asphyxia; n = 15). Pup body weight (g) was recorded on PND 3, 7, 14, 18, 21, 28, and 32 at the same time of day (09:00). Due to technical issues, body weights at PND 41 and 50 were not obtained. Spiny mice pups generated for this study were culled and used to assess multi-organ pathology at varying postnatal ages, as previously published [29‒32]. Therefore, the number of pups that underwent behavioural testing at each postnatal age varied. A summary of the number of pups assigned to each birth type and diet for each behavioural test is provided in online supplementary Table S1 (for all online suppl. material, see https://doi.org/10.1159/000544756). The survival percentage of pups not intentionally culled for other age-related analyses was 100%.

Behavioural testing was conducted in a room separate from the housing room. Testing arenas were thoroughly cleaned with 70% ethanol between tests. All mice were allowed to acclimatize to the testing room for at least 30 min prior to the commencement of each test. All testing was conducted under lighting levels set at 2.8 lx. Animals were tested in a randomized order and conducted by an experimenter (J.T., S.J.E.) blinded to the group allocation. All mouse activity was recorded using an overhead HD digital CCTV video camera and continuously tracked with “Limelight”^® tracking software. All post-test behaviour analysis of activity was conducted using an automated and unbiased CleverSys TopScan^® software (Reston, VA, USA). Neonatal ages were defined as <PND 10 and adolescence/adulthood as >PND 10 as previously defined [33]. Due to technical difficulties and, at times, unreliable tracking software outputs, not all data from each offspring could be recorded. Numbers per group for each properly conducted and recorded behaviour test are displayed in the results section.

Open-field tests were conducted on PND 3, 7, 14, 18, 21, and 28 to assess exploratory locomotor activity and anxiety-like behaviour in mice [34]. Individual mice were placed in the centre of a 40 × 40 cm black open-field apparatus with 50 cm high walls and allowed to explore freely for 10 min. Activity was recorded using an overhead camera located above the apparatus. The open-field apparatus was separated into a central zone defined as the centre 20 × 20 cm and an outer zone, i.e., peripheral zone to the centre. The total distance travelled (mm), total duration of time spent moving (seconds), and average velocity of movement (mm/s) in the apparatus were measured to assess general locomotor activity. Thigmotaxis, or “wall hugging” behaviour, was measured as the percentage of distance travelled within the outer zone over the total distance travelled, and the percentage of time spent moving within the outer zone over the total duration of time spent moving, as a measurement of non-anxious, normal behaviour in spiny mice [33, 35].

Elevated plus maze test was conducted on PND 32 to assess anxiety-like and exploratory behaviour [36]. We have previously concluded that testing anxiety-like behaviour in the elevated plus maze within the neonatal period (PND <14) cannot be accurately assessed and have thus assessed this only at adolescent/adulthood ages [33]. The apparatus consisted of a plus-shaped maze raised 50 cm above the ground, with 4 arms extending from a central 10 × 10 cm platform. The arms were 45 cm in length and 10 cm in width and the closed arms had surrounding walls of 20 cm height. At the start of the test, the mouse was placed in the centre facing an open arm, then allowed to freely explore the maze for 10 min and activity was recorded using an overhead camera located above the apparatus. Distance travelled (mm), and duration of time spent moving (seconds) in total and in the closed and open arms were recorded. The time spent in the closed and open arms expressed as a percentage of the total duration of time spent moving in the maze was also calculated.

Novel object recognition tests were conducted on PND 18, and 41 to assess non-spatial memory and learning behaviour, and novel object exploration [37]. As in the elevated plus maze, we have previously concluded that testing non-spatial memory and learning behaviour in the novel object recognition tests within the neonatal period (PND <14) cannot be confidently assessed and have thus assessed this behaviour only at adolescent/adulthood ages [33]. This test was conducted after a 10-min period of habituation to the arena (same arena used for the open-field test) to reduce the potential effects of anxiety and stress. After ∼30 min back in their home cage, session 1 involved mice being individually placed in the testing arena, in which two identical glass bottles were placed in the open field near the walls of the enclosure to allow for an unobstructed view of the objects. Mice were allowed to freely explore for 10 min in session 1 to allow for a familiarization (learning) period. The mouse was then placed back in its home cage for 1 h. After the inter-trial interval, session 2 involved the same mouse being placed back into the testing arena that now contained one novel object of different shape and colour as well as one familiar object. The animal was allowed to freely explore for 10 min and activity was recorded using an overhead camera located above the arena. The arena and objects were cleaned with 70% ethanol in between trials to eliminate olfactory cues. Time spent exploring any object (total object exploration duration; seconds) was recorded. A “discrimination index” percentage was calculated by subtracting the time the animal spent exploring the familiar object from the time spent exploring the novel object, which was then divided by the total time spent exploring both of the objects. A positive discrimination index indicates a preference to the novel object, and a negative positive discrimination index indicates a preference to the familiar object.

All data are presented as mean ± SD and were assessed for normality using the Shaprio-Wilk test and homogeneity of variance using the Levene Test. All data passed normality and homogeneity unless specified. Data were analysed using a linear mixed-effects model to assess the main effects of age (p_AGE), birth mode (C-section [control] or birth asphyxia; p_BIRTH), diet (control or creatine; p_DIET) with the co-variate effect of sex (male or female; p_SEX) and with cross-foster litter (p_LITTER) as a random effect also included in the model. Interactions between the main effects (p_{AGE × BIRTH,}p_{AGE × DIET,}p_{BIRTH × DIET,}p_{AGE × BIRTH × DIET}) were also established. For significant three-way or two-way interactions, pair-wise comparisons between groups were conducted with a Tukey’s post hoc test. For assessment of the open-field test, data analysis was divided into the neonatal period (PND 3, 7) and adolescent period (PND 14, 18, 21, and 28) as previous examinations demonstrated contrary locomotor activity preferences between PND <10 and PND >10 [33]. For assessment of total object exploration duration changes between session 1 and session 2 in the novel object recognition test, data did not pass normality using the Shaprio-Wilk test and thus a Wilcoxon matched-pairs signed-rank test was conducted on these data. Results were accepted as statistically significant when p < 0.05. All analyses were performed using R (version 4.4.1) and R Studio (version 2024.09.0-375) software. All graphs were generated using GraphPad Prism (GraphPad Prism version 10.2.0 for Windows, GraphPad Software, CA, USA).

Maternal creatine supplementation from mid-gestation was well-tolerated and did not result in any severe adverse events for the mother or her offspring. Body weight gain trajectory increased between PND 3 and 32 with no effect of birth mode (C-section [control] or birth asphyxia) or sex (Fig. 1). Overall, offspring from creatine-supplemented dams were heavier than control non-supplemented offspring (total averages; control diet = 12.9 ± 6.05 g, creatine = 14.00 ± 6.49 g) both in terms of birth mode and between PND 14–32 (p < 0.03).

Exploratory activity in the open-field is a measure of general locomotor activity and anxiety-like behaviour (Fig. 2a). In the neonatal period (PND 3–7), there was no difference in the total duration spent moving by the spiny mice (Fig. 2b); however, spiny mouse pups that underwent birth asphyxia travelled reduced distances (control = 5,621 ± 4,934 mm, asphyxia = 3,844 ± 3,886 mm; F_1,34.10 = 4.44; p_BIRTH = 0.042; Fig. 2c) and at slower average velocities (control = 8.76 ± 8.24 mm/s, asphyxia = 5.89 ± 6.37 mm/s; F_1,32.00 = 4.41; p_{BIRTH =} 0.044; Fig. 2d) in the open-field compared to controls. The degree of exploration in terms of distance (PND 3 = 2,062 ± 1,724 mm, PND 7 = 7,474 ± 4,881 mm; F_1,146.36 = 102.46; p_AGE <0.0001; Fig. 2c) and velocity (PND 3 = 3.4 ± 3.2 mm/s, PND 7 = 11.4 ± 8.48 mm/s; F_1,147.21 = 64.07; p_AGE <0.0001; Fig. 2d) was higher at PND 7 than PND 3. Percentage duration spent moving and distance in the outer zone of the open-field arena (the degree of thigmotaxis) were both affected by age and birth (F_1,161.03 = 7.08; p_{AGE × BIRTH} = 0.009 and F_1,160.08 = 7.09; p_{AGE × BIRTH} = 0.009, respectively) and age and diet (F_1,160.96 = 4.94; p_{AGE × DIET} = 0.028 and F_1,156.00 = 5.00; p_{AGE × DIET} = 0.027 respectively; Fig. 2e, f). Follow-up analysis revealed that asphyxic spiny mice pups at PND 3, irrespective of maternal diet, had reduced percentage durations (C-section control = 93.62 ± 13.97%, asphyxia = 79.28 ± 32.89%, p = 0.015) and distances (C-section control = 89.34 ± 16.13%; asphyxia = 75.31 ± 31.02%, p = 0.017) in the outer zone compared to non-asphyxic spiny mice pups. By PND 7, this deficit was no longer present (Fig. 2e_i, f_i). Furthermore, in PND 3 pups that were antenatally supplemented with creatine, irrespective of mode of birth, the percentage durations (control diet = 83.06 ± 29.37%, creatine = 94.90 ± 11.64%, p = 0.022) and distances (control diet = 78.74 ± 28.04%; creatine = 91.06 ± 14.66%, p = 0.013) were greater in the outer zone than pups born from mothers on the control diet. By PND 7, these differences disappeared with pups exposed to the creatine diet displaying similar percentage durations and distances in the outer zone to pups exposed to the control diet (Fig. 2e_ii, f_ii).

The open-field test was also conducted in offspring at adolescent ages, PND 14–28. The total duration travelled in creatine asphyxia mice at PND 28 was reduced compared to control asphyxia mice at PND 28 (p = 0.037; Fig. 2g). The total distance travelled in adolescent creatine-exposed offspring was not statistically higher than that of controls, irrespective of birth or age (creatine = 27,648 ± 6,641 mm, control diet = 25,631 ± 6,579 mm; F_1,39.2 = 3.41; p_DIET = 0.072; Fig. 2h). There were no significant post hoc differences found for average distance and velocity despite there being a significant three-way interaction (Fig. 2h, i). At PND 14, the percentage duration and distance in the outer zone of the open-field arena was higher compared to all other ages (all p < 0.0001; Fig. 2j, k). There was no effect of sex on any open-field test outcomes.

Exploratory activity in the elevated plus maze was assessed for anxiety-related behaviour (Fig. 3a). The total duration spent moving was less in creatine offspring than control offspring, though this did not reach statistical significance (creatine = 276.3 ± 59.57 s, control diet = 364.9 ± 130.2 s; F_1,30.94 = 3.94; p_DIET = 0.056; Fig. 3b). There was no effect of birth or diet on the total distance travelled or percent of time spent in the closed or open arms (all p > 0.05; Fig. 3c–e). However, maternal creatine-supplemented offspring did spend less time moving within the closed arms compared to control pups (creatine = 138.8 ± 40.90 s, control diet = 191.2 ± 74.53 s; F_1,29.52 = 5.87; p_DIET = 0.022; Fig. 3f), but there was no difference in the total distance travelled in the closed arms (Fig. 3g). Total duration and distance travelled in the open arms was also not different between groups (Fig. 3h, i). There was no effect of birth mode or sex on any elevated plus maze outcomes.

The learning trial of the novel object recognition test measures the amount of time the spiny mouse spends exploring familiar objects in session 1. Then, the time spent exploring the novel object during session 2 is measured to assess non-spatial memory and learning behaviour (Fig. 4a). During the familiarization period (session 1), there was a significant interaction between birth asphyxia and maternal dietary creatine supplementation irrespective of age, with asphyxiated pups born of creatine-supplemented dams spending less time exploring the objects compared to all groups (F_1,58.86 = 21.88, p_{BIRTH × DIET} <0.0001; Fig. 4b; p < 0.0001 compared to all other groups). By session 2, where one of the objects was replaced with a novel object, this interaction between birth asphyxia and maternal dietary creatine supplementation was no longer significant, and there were no differences between groups on the total object exploration duration (Fig. 4c). This result was due to a significant decrease in object exploratory behaviour between sessions in control, control asphyxia, and creatine offspring whereas there was no change in object exploratory behaviour between sessions in creatine asphyxia offspring (Fig. 4d). The preference for the novel object (discrimination index) was lowest in asphyxic pups at PND 18 compared to non-asphyxic pups at PND 18 and pups at PND 41 (F_1,46.08 = 4.09, p_{AGE × BIRTH} = 0.049; Fig. 4e; p < 0.05 compared to all other groups). Maternal creatine did not alter this outcome. There was no effect of sex on any novel object recognition test outcomes.

Previous research has shown that maternal creatine supplementation during pregnancy can reduce short-term cerebral cellular injury in spiny mice offspring 24 h after birth asphyxia [15]. However, the possible therapeutic effects of maternal creatine supplementation on long-term offspring neurological function following acute birth asphyxia were unknown. Furthermore, creatine supplementation during in utero development has been shown to alter offspring neuronal function and morphology in rats and sheep [16‒18], of which the long-term effects have also not been assessed. This is clearly important to understand if maternal dietary creatine supplementation is to be adopted as a prophylactic treatment strategy for newborn HIE.

In this study, acute birth asphyxia of near-term spiny mice pups produced locomotor deficits at PND 3–7 and increased anxiety-like behaviour at PND 3, as measured by the open-field test. By adolescence, besides a reduction in novel object discrimination index, the effects of acute birth asphyxia were minimal. Maternal creatine supplementation during in utero development resulted in offspring displaying reduced anxiety-like behaviour in the open-field test at PND 3 regardless of birth mode, thereby negating the increased anxiety-like behaviour induced by asphyxia. Moreover, irrespective of birth mode, offspring exposed to antenatal creatine demonstrated continued reduced anxiety-like behaviour in the open-field test and elevated plus maze at adolescent ages (PND 28 and PND 32), as well as promoted an increase of body weight. A synergistic effect of maternal creatine and birth asphyxia to reduce object exploratory behaviour in offspring at adolescence/adulthood ages was another unexpected finding. These data demonstrate that antenatal creatine has lasting behavioural effects likely related to alterations in the central processing of emotional behaviour. There was no effect of sex on any behavioural outcomes, irrespective of birth asphyxia, or antenatal creatine supplementation.

Locomotor activity and general exploratory behaviour were assessed using the open-field test. The ambulatory ability of neonatal spiny mice after birth asphyxia was compromised resulting in a reduction in the total distance covered and the average velocity of movement during the test. This aligns with previous spiny mouse birth asphyxia studies, which found neonatal delays in motor coordination and balance at PND 1 and 2, and reduced locomotion at PND 5 and 15 [38, 39], although we did not observe any deficits by PND 14. Newborn spiny mice normally display limited exploratory activity (open-field test), displaying significant thigmotaxis, i.e., preference for the outer zone (duration and distance) [33]. Therefore, reduced movement in the outer zone is abnormal behaviour in newborn spiny mice, potentially reflecting increased anxiety to move from where they were initially placed at the centre of the arena [40]. Birth asphyxia increased this anxiety-like behaviour of reduced movement in the outer zone at PND 3, irrespective of diet. But, pups from the creatine-treated mothers displayed increased movement in the outer zone at PND 3, irrespective of birth mode. This indicates that creatine was able to negate the asphyxia-induced anxiety-like behaviour at PND 3 by overall enhancing passive avoidance, i.e., by successfully avoiding the centre of the field to avoid a spatial location that is anxiogenic [40].

Interestingly, there were synergistic effects of asphyxia and creatine in adolescent/adult spiny mice during the first novel object recognition session, designed to assess exploratory behaviour and novel object exploration. Creatine asphyxia mice displayed reduced total object exploration duration during session one compared to all other groups. There was no change in the total object exploration duration between session one and two in creatine asphyxia mice, suggesting an overall lack of reactivity to the novelty of the object or a more rapid adaptation to novelty. The motivational effort to novelty is connected to dopaminergic neuromodulation [41], thus the reduced object exploration may be due to increased dopaminergic sensitization caused by an interaction of both creatine and asphyxia. Indeed, birth asphyxia induces long-term regional changes in dopaminergic receptor expression and binding within the brain [42]. For example, PND 28 rats display altered dopaminergic-related behaviour which varies depending on the severity of the asphyxia, i.e., shorter asphyxic insults resulted in increased exploratory behaviour, whilst longer asphyxic periods caused a decrease in exploratory behaviour [43]. These behavioural changes were exacerbated with dopamine stimulants [43]. Moreover, creatine supplementation has been shown to induce activation of cerebral dopamine receptors in adult mice [44]. However, no studies have been conducted to determine if the creatine effects on dopaminergic neurotransmission persist after creatine supplementation has ceased.

At a behavioural level, it is difficult to extrapolate the behaviour of reduced interest in a novel object observed in the creatine asphyxic pups to the human condition. This reduced novel object exploration may reflect enhanced learning-related passive avoidance, in this case to a novel object rather than environment (i.e., open field, elevated plus maze), that is intensified with creatine in combination with birth asphyxia [45]. Alternatively, this reduced novel object exploration could reflect “neophobia,” or maladaptation to adapt to novelty or even learning and decision-making [46]. Behavioural testing using goal-directed learning (e.g., positive or negative feedback) could discern whether this combined creatine asphyxia effect is a deficit or enhancement. Furthering our understanding of the underlying cellular effects, especially in regard to dopaminergic neurotransmission, would also be beneficial.

Early neonatal clinical signs of HIE in humans, including reduced tone and lethargy, are highly predictive of poor long-term neurodevelopmental outcomes [47]. Despite the early behavioural deficits in neonatal spiny mice after birth asphyxia, the only behavioural deficits at an adolescence/adulthood age identified in this study was a reduction in the discrimination index in the novel object recognition test at PND 18, which is consistent with previous reports [39]. This may represent a limitation of this study, as we cannot discern if the recovery was due to a relatively mild birth asphyxic injury and/or was the result of the spiny mouse’s unique regenerative properties [48]. The survival rate of the birth asphyxic insult at delivery is ∼63% and ∼95% for creatine-supplemented newborns, as reported elsewhere [27]. Whether a survival effect contributed to the relatively mild behavioural deficits at adolescence/adulthood could not be ascertained in this study. Indeed, the immature and developing brain does possess cellular processes that facilitate functional recovery (sprouting of neurites, synapse formation, and myelination) after cerebral HI injury, as seen in rodents following unilateral carotid artery ligation and exposure to moderate hypoxia [49]. Functional recovery is thus likely to depend on the severity of the injury [50]. The 7.5 min of birth asphyxia induced by submerging the uterus containing pups in saline was adequate to induce histological neuropathology in the brain during the neonatal period [38, 39], but otherwise, as suggested by this study, persisting neuropathology was likely to be minimal and may have resolved by adolescence/adulthood. Future histological assessments are needed to confirm that this is the case.

Another explanation for the lack of asphyxia-mediated behavioural deficits in adolescence/adulthood could be that the spiny mice themselves have greater adaptive neuroplasticity and an increased ability to recover after neurological injury. Indeed, a striking property of adult spiny mice is their ability to regenerate injured tissue, which extends to restoring central nervous system function following a severed spinal cord [51, 52], and wound healing following experimental traumatic brain injury induced by cerebral puncture [53]. On these bases, the spiny mouse may be relevant for preclinical investigations impairments in motor-related and emotion-driven function following birth asphyxia during the acute neonatal period but may not be an ideal animal to model long-term functional deficits or to assess the long-term neuroprotective effects of creatine. There is an interesting further consideration; if the spiny mouse is capable of recovering from a moderate-to-severe hypoxic insult, understanding the unique mechanisms by which this occurs would pave the way for highly innovative approaches for treatment strategies that might prevent neurological impairments in human infants after perinatal HIE.

Previous studies have suggested that antenatal creatine exposure may alter hippocampal development [54, 55]. For example, pregnant rats given 1% creatine in drinking water for 10 days before birth resulted in offspring with faster functional maturation of hippocampal pyramidal neurons at adolescent equivalent age (promoted overgrowth and complex dendritic branching, and increased excitability and neuronal output) [17]. More recently, we showed that 13 days of continuous intravenous creatine infusion in foetal sheep increased hippocampal expression of the choline transporter involved in the maintenance of cholinergic neurotransmission [16]. Moreover, increases in myelination in the major white matter tracts (subcortical and periventricular) were also found in these creatine-supplemented foetal sheep, suggesting effects on wider neural networks relating to cognition such as sensorimotor and visual processing [16, 56]. While it was not possible to concurrently sample tissue and undertake behavioural tests at multiple postnatal ages, this study assessed creatine-mediated effects on behaviour associated with reported hippocampal changes for the first time.

This study found that antenatal creatine supplementation during in utero neurodevelopment, irrespective of birth mode, resulted in adolescent offspring spending less time moving in the closed arms of the elevated plus maze. As observed here and by others previously, spiny mice demonstrate a unique pattern of behaviour with a decline in exploratory behaviour with age, preferring the safety of closed arms and walls of the behavioural apparatus [33, 35]. Increases beyond normal affective behaviour suggest a further age-related reduction in anxiety-like behaviour caused by creatine, though it remains unclear whether these are positive or negative behavioural traits.

Regardless, these results suggest that neural pathways, especially those involved in the emotional processing that drives anxiety-like and exploratory behaviour in offspring, are affected by maternal creatine supplementation. Creatine kinase activity is positively correlated with synchronised neural activity that facilitates information processing [57]. Increases in creatine kinase activity induced by maternal creatine supplementation could enhance innate information processing of sensorimotor or visual inputs, in this case, inclination to the walls of the open-field arena and the closed arms of the elevated plus maze; however, this remains speculative. When considering these results in conjunction with previous histological examinations in multiple species, it appears that antenatal creatine exposure may alter neurodevelopment and function. As such, given the specific behavioural changes identified here for the first time, further behavioural and cellular investigations that assess specific brain regions related to emotional and information processing (e.g., amygdala, ventromedial prefrontal cortex limbic cortex) are needed to confirm whether changes in affective behaviour induced by creatine supplementation during in utero neurodevelopment are aversive or not. Examination of the bioenergetics underlying information and emotional processing are likely to reveal a mechanism for creatine’s effects, as associations between altered brain creatine metabolism and functional cognitive deficits have already been found in preterm infants [58].

Another interesting effect on the offspring from maternal creatine-supplemented dams, regardless of birth asphyxia, was their slightly heavier body weights throughout development (+8.5%; +1.1 g). This is the first indication of a persistent increased weight effect long after cessation of creatine supplementation. This creatine-induced weight gain may result from increased fat-free mass [59] and/or fluid retention, thereby increasing total body water volume [60], at least while tissue creatine loading remains during the washout period. Whether the creatine-induced weight gain is due to increased fat-free mass, and/or increased caloric and fluid intake, which was not recorded, is unknown. However, we have previously shown that at least at PND 1 and 33, there is no effect of creatine on skeletal gastrocnemius muscle cross-sectional area [61]. This effect of maternal creatine diet on postnatal weight gain raises the possibility of its application in infants with poor weight gain after birth.

We have shown that maternal creatine supplementation from mid-gestation can influence offspring behaviour and also increase body weight in spiny mice. Antenatal creatine exposure reduced anxiety-like behaviour, which in spiny mice neonates negated behavioural abnormalities induced by birth asphyxia. This creatine-mediated increase in affective behaviour persists in early adolescence/adulthood (>PND 18) even in the absence of overt behavioural deficits caused by birth asphyxia – whether this is a positive or negative behavioural outcome requires further elucidation. We also show for the first time that the combination of antenatal creatine and birth asphyxia results in adolescent/adult offspring having reduced novel object recognition. This may reflect a maladaptation to novelty or enhanced learning-related passive avoidance perhaps facilitated by creatine’s ability to reduce anxiety-like behaviour. Further investigation into the effects of creatine on other important behavioural parameters, as well as the specific cellular and neural pathways that dictate these behaviours, is still required before definitive conclusions can be made on the long-term safety of maternal creatine supplementation and antenatal creatine exposure.

The authors thank Dr. Domenic A. LaRosa and Dr. Hayley Dickinson for their assistance with initial experimental design and thank the facilities and technical assistance of Monash Animal Research Platform. Schematic diagrams were created with BioRender.com.

The experimentation for this study was approved in advance by Monash University School of Biomedical Sciences Animal Ethics Committee (MMCA2010/26) and was performed in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes.

The authors declare no commercial or financial relationships that could be construed as a potential conflict of interest. S.J.E. does serves as a member of the Scientific Advisory Broad on creatine in health and medicine (AlzChem LLC.). However, the company was not involved in the design or dissemination of this study.

Funding acquisition includes: Jack Brockhoff Foundation Early Career Grant (N.T.T.); National Health & Medical Research Council of Australia 1124493 (S.J.E., D.W.W., R.J.S.); National Health & Medical Research Council of Australia Early Career Fellowship 1125539 (S.J.E.). Funders had no role in the design, data collection, data analysis, and reporting of this study.

N.T.T.: data curation, formal analysis, funding acquisition, investigation, writing – original draft, writing – review and editing, and approval of the final draft. J.T.: data curation, formal analysis, investigation, writing – review and editing, and approval of the final draft. T.Y.: data curation, formal analysis, investigation, methodology, supervision, writing – review and editing, and approval of the final draft. R.J.S.: conceptualization, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, writing – review & editing, approval of the final draft. D.W.W. and S.J.E.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, writing – review & editing, approval of the final draft.

All data are available in the main text or the online supplementary materials. The dataset and code supporting the findings of this study are available in the following GitHub repository: https://github.com/DrNhiThaoTran/Creatine_Long_Term_Behaviour. The repository includes the raw dataset and analysis scripts necessary to replicate the statistical results presented in this study. Further enquiries can be directed to the corresponding author.