For the first time, our research presents a detailed description of the course of pregnancy in KD-fed rat females from the very beginning of pregnancy to the end of lactation period. It became clear that, despite the intake of high-calorie food, in which as much as 80% of energy comes from fat, their average body mass at any of the four time points of pregnancy (up till G20) was not greater than that of the ND-fed females. However, the level of BHB in their peripheral blood was significantly higher from G4 and increased up till postnatal day 21, when it reached the level of 7 mmol/l. According to the literature data (Herrera et al., 1987), ketone bodies circulating in the maternal blood easily cross the placenta. Under the conditions of a normal diet, the fetal development depends on a continuous supply of glucose and appropriate amino acids. Any insufficiency in this supply affects the metabolism in both the mother and the fetus (Barry, 2018).
Our research showed that the glucose level in KD-fed females was lowered only at the last measurement point (G20), i.e. about two days before delivery, so there should rather not lead to any important disturbances. As it is known, long-term ketonemia during pregnancy can evoke metabolic encephalopathies (Al-Mudallal et al., 1995). Sussman (2013) showed that in the early stages of pregnancy, KD intensifies the heart growth (at E13.5) and reduces the growth of the pharynx, cervical spine, pons, midbrain and hypothalamus. On E17.5, she observed a reduction of the heart and thymus, and an enlargement of the cervical spine, pons, midbrain and thalamus. This discrepancies in organ growth may be due to variations in the availability of metabolic "fuel". As revealed by Lust et al. (2003), on the 18th day of gestation in ND-fed females, a significant increase in BHB content was found in the fetal brains.
Our observations showed that KD does not affect the duration of pregnancy, and all the females, regardless of the type of diet applied, gave birth 21–22 days after conception. The number of pups in the litter also did not differ between the animal groups, and there were no changes in proportions of male and female pups. This is likely due to the fact that all pregnant females were still on ND at the time of conception. However, on P2, a significantly reduced body mass was observed in animals born by females treated with KD during their pregnancy. This difference persisted on postnatal days 6. Similar results were obtained by Mark et al. (2011) and Mendes-da-Silva et al. (2014) using a high-fat diet in pregnant females, and Soares et al. (2009) using two KD types with different fat contents. Despite the considerably lowered body mass, we did not observe any morphological anomalies in the pups born by females kept on KD during pregnancy. This early reduction in the body mass was compensated relatively quickly following, as evidenced on P14 by disappearance of differences between offsprings of KDND- and ND-fed females which is the effect of KD replacement with ND. On the contrary, a continuous application of KD resulted in a constant reduction of the body mass during the whole examined period, i.e. up to the weaning time (P21). However, this is only an indicator of compensation in terms of general somatic development, which does not have to correspond with an improvement in the functional neurological state which depends on disorders of the much more restrictive calendar of brain development.
Our present research was focused on the elemental composition of the brain in pregnant females continuously fed with KD. Before, we have already performed a similar study (Chwiej et al., 2015), but only on 30-day-old males who were treated with KD for next 30 days. So far, such studies have not been performed on females. In the present experiment, FTIR microscopy detected significant changes in the biochemical composition of the nervous tissue in KD-treated pregnant females. Generally, in white matter, but also in the hippocampal region, higher ratios of absorbance at 1360–1480 cm− 1 and 1740 cm− 1 were detected. They indicated, respectively, increased levels of lipids and ketone groups-containing compounds thus confirming the changes previously described in males. They also proved that ketone bodies easily passed through the BBB and could accumulate within the brain parenchyma (Barry, 2018). The higher level of lipid bands proved that the increased availability of ketone bodies implied a subsequent increase of the lipid synthesis (Barry, 2018).
An additional Principal Component Analysis of the spectra from particular layers of the hippocampal formation and neocortex detected no interregional differences. It may suggest a generalized nature of the previously described changes. In contrast, our studies on normal animals, that were fed KD for 4 months, showed increases in the number of myelinated fibers passing through the hippocampal formation, striatum, internal capsule and pons (Gzieło et al., 2020). These differences might depend on the length of time during which the diet was applied. When it was shorter, the changes were rather diffuse, and when it was longer, they became more specific to particular brain regions.
To assess how KD, low in proteins and sugars, applied in pregnant females, would affect the postnatal development of their offsprings, a series of neurodevelopmental reflexes was performed in three groups of animals whose mothers (i) were fed KD during pregnancy and lactation (KDs), (ii) were fed KD only during pregnancy and ND from the second postpartum day (KDNDs), (iii) were fed exclusively with ND (NDs). The tests assessed some general parameters of physiological development (like eye opening and posture), reflexes (forelimb and hindlimb grasping, hindlimb placing, body rotation and auditory startle. The tests were successively introduced from P3, when it was not yet possible to observe more complex behavioral parameters. These tests, just like in human newborns, could predict developmental disorders of various origins, such as, for example, cerebral palsy (Nguyen et al., 2017).
All the neurological tests were performed separately on males and females, but the hindlimb grasping and righting tests were the only two of them which detected intersexual differences. The righting test in male pups born by ND-treated females was performed significantly earlier than in their female counterparts originating from KD-treated females. Similarly, the male pups of KDND-fed females were better in the hindlimb grasping tests. No gender-dependent differences were observed in any other test, regardless of the type of food consumed during pregnancy. It appears, therefore, reasonable that until puberty (at about P21), differences between behaviors of males and females may not be expected. Similar conclusions were presented by Sussman et al. (2015).
Interestingly, two important physiological symptoms of maturity, i.e. opening the eyes and maintaining a proper posture, were first observed in the offspring of KDND-fed females, compared to those from females which were always on ND. Similar results were observed by Nguyen (2017), in pups whose mothers received broccoli extract during pregnancy and had induced inflammation. Studies on rodents show that eye opening is correlated with changes in the GABAA receptor subunit in the visual cortex (Heinen et al., 2004). Thus, earlier eyes opening may indicate an intensification of plastic processes in the cortex supposedly evoked by the dietary changes.
In most of the performed tests (except for hindlimb placing), the maintenance of KD throughout pregnancy, as well as during the period including both pregnancy and lactation, delayed the performance of the tests in respective offsprings as compared to the progeny ND-nourished females. The latest performance of the tests was in animals whose mothers were fed KD, both during pregnancy and lactation periods. In ND-fed animals, the energy from the lactate metabolism transformation is obtained only shortly after birth. Then, until the end of lactation period, the ketone bodies are the source of energy for the pups (Prins, 2012), because during the neonatal period their brains are not able to efficiently use glucose as a metabolic substrate (Nehlig, 1999). During this time, not only does the amount of ketones in the body increase, but also the number and activity of their transporters in the BBB (Vannucci and Simpson., 2003).
This is why the immature rat brain is able to take up and accumulate ketone bodies faster and more efficiently than the mature one (Nehlig, 1999). During the period of peak ketone utilization, the brain ability to uptake ß-hydroxybutyrate (ßOHB) is 6-fold greater than that in adulthood (Cremer et al., 1976). So, if the lactating females are additionally fed with KD, these changes may be intensified and lead to metabolic acidosis (Sussman, 2013), which reduces lactation and thus inhibits the developmental progress of the pups. This results, apart from body mass reduction, in definitely delayed muscle development and developmental brain disorders leading to worse performance of neurodevelopmental reflexes.
The prenatal period is critical to the development of the nervous system (Rice and Barone, 2000) and, apart from stress or inflammation, it can be affected by malnutrition (Cusik and Georfieff, 2016). The continuity of prenatal and postnatal development should also be taken into account. Thus, effects of prenatal malnutrition would occur also in brain structures that develop postnatally (Morgane et al., 2002). Following KD consumption during pregnancy, as much as 80% of energy derive from metabolized fats and as much as 70% less protein is provided than in the case of ND. Protein deficiency in the diet of pregnant females, as described by Abey (2019), delayed neurological and postural reflexes. The results of our tests confirm this observation in the offsprings of females receiving KD only during pregnancy or during both pregnancy and lactation. We also observed depleted hair in the offspring of females who were fed with KD while being pregnant and lactating. This could be due to inadequate development of the hair follicles (Guo and Katta, 2017) resulting from protein deficiency. Another reason for the delays in the tests of neurodevelopmental reflexes shown by the offspring of KD-fed females might be a reduction of forebrain vascularization in animals on a protein-deficient diet (Bennis-Taleb et al., 1999) which can be preserved up till adulthood.
During pregnancy, the protein-poor diet may also decrease expression of brain neurotrophins in the rat offspring including that of BDNF which, in turn, affects neuro- or synaptogenesis (Marwarha et al., 2017) resulting in worse performance of neurodevelopmental tests.
The results of seven from of the ten neurological tests applied in the present study clearly indicate that the ketogenic diet withdrawal may lead not only to a general, unspecific restitution of somatic condition, represented here by the compensatory regain of reduced body mass, but may also result in functional restitution of the nervous system, although not to the full extent. However, our study has considered developmental period up to the weaning time, so a further progress in the functional amelioration up till the complete adulthood cannot be excluded. Thus, the first and most important task for further research in this field is to detect, if not the specific mechanisms underlying these changes, then at least their structural correlates.