Hypoglycemia occurs in the neonatal period, during the metabolic transition to the extrauterine environment as maternal glucose delivery to the infant ceases after birth, resulting in a fall in neonatal blood glucose levels which will later normalise. Although it is more common in preterm or growth restricted or macrosomic infants, or those born to a mother with poorly controlled diabetes, it may also occur in infants not considered to be at risk, they stabilize their blood sugar concentration by day 4 [1–4]. However, many controversies persist, including the definition of hypoglycemia, whether current screening guidelines are adequate and what devices should be used to measure and monitor blood glucose concentrations [5]. The pathophysiology of neonatal hypoglycemia remains speculative in a majority of cases [6].

The main deleterious effect of hypoglycemia is neuroglycopenia, or brain energy insufficiency occurring in the central nervous system [5, 7, 8]. This is unsurprising as 60% of the glucose available for the body is utilized by the brain as metabolic substrate required to generate adenosine triphosphate (ATP) for its normal physiologic functions. With this high energy demand, the brain needs continuous delivery of glucose. With the existing linear relationship between plasma glucose levels and brain glucose concentration, the brain can survive only for a few minutes after glucose deprivation [9]. When blood glucose concentration reaches approximately 42 mg/dL, the brain glucose level drops to zero due to its continuous consumption by brain cells.

In newborn rats, brain damage is due to recurrent hypoglycemia [10]. In humans, hypoglycemia compromises brain structure [7] and function [11, 12] and, if severe, white matter abnormalities are observed on brain magnetic resonance imaging [13]. The parietal and occipital lobes are the most commonly affected regions [7, 11, 12, 14]. The hypoglycemia-induced cerebral injury may be sub-clinical or presents as neurologic dysfunctions, such as irritability, neurological deficits, seizure, coma and, ultimately, neuronal death with permanent long-term neurodevelopmental impairment [8, 15–17]. These complications are less frequent when blood glucose concentration is maintained above 47 mg/dL [18]. Even early transient newborn hypoglycemia, thought initially to be benign, is associated with lower achievement test scores at 10 years of age [19].

The pathophysiology of hypoglycemic brain injury, the role of its severity, age of occurrence or duration are not well known. The developing brain of a newborn is exquisitely sensitive to substrate reduction since it has minimal high-energy phosphates reserves, and hypoglycemia also inhibits glycolysis and glycogenolysis, making it even more vulnerable to injury. Mitochondria generate cellular energy by recycling low energy adenosine diphosphate (ADP) to high energy ATP which is the source of energy for use and storage in the cells. A modest but nearly constant amount of intracellular ATP provides the mitochondria with regular energy input. When consumed in metabolic processes, ATP is converted either to ADP or to adenosine monophosphate (AMP).

ATP is synthesized mainly in mitochondria where the oxidative phosphorylation performs coupling between oxygen consumption and phosphorylation of ADP. However, mitochondrial fragmentation occurs within minutes of glucose starvation [20]. This results in disruption of their structure, reduction of their number by upregulating the expression of mitophagy markers, and the production of reactive oxygen species (ROS) [21]. The mitochondrial membrane potential is also impaired and its energy metabolism is suppressed resulting in impairment in mitochondrial dynamics [22]. Nonetheless, cellular ATP content is not the best representative of cellular energy status because it is, instead, the balance between the three nucleotides (ATP, ADP, and AMP) which is essential for cellular homeostasis [23]. This is better represented by the cellular adenylate energy charge, abbreviated as AEC which is expressed by the equation [ATP] + 0.5 [ADP]/[ATP] + [ADP] + [AMP] [24]. While its value varies within a scale between 0 and 1, it is preserved in a narrow range between 0.7 and 0.95 in the cells of both prokaryotes and eukaryotes [25]. The energy of the system is fully charged when all adenylate is in the form of ATP and fully discharged when only AMP is present. AMP-activated protein kinase (AMPK), a sensor of cellular energy change, is activated by increased AMP/ATP and/or ADP/ATP ratios which occur during metabolic stresses that either interfere with the generation of ATP (e.g., glucose deprivation, hypoxia, ischemia) or accelerate its consumption (e.g., muscle contraction) [26, 27]. AMPK maintains the energy balance by decreasing the ATP-consuming processes while the metabolic pathways such as glucose and fatty transport, fatty acid oxidation, autophagy, mitochondrial synthesis and oxidative metabolism are increased to preserve ATP during energy deficiency [27, 28]. Hypoglycemia causes functional brain failure due to decrease in AEC which represents brain fuel that alters brain function and behavior. Hence, it is a marker of cell failure. However, if managed early, this can be reversed by increasing plasma glucose concentrations. Nevertheless, prolonged hypoglycemia causes activation of neuronal glutamate receptors, glutamate release, production of ROS, activation of poly (ADP-ribose) polymerase, neuronal zinc release and transition in mitochondrial permeability resulting in hypoglycemia induced neuronal death. The low AEC is therefore both a marker of neuronal cell failure and also a direct cause of early and also, indirectly, of permanent neuronal dysfunction in neuroglycopenia [23].

Although the brain contains small amounts of glycogen [29], its complete oxidation would sustain basal respiration for only a very short time and it would be quickly depleted without the prior availability of external glucose. This limited glycolytic capacity of the neurons makes their energetic needs highly dependent on mitochondrial oxidative phosphorylation [30, 31]. Although the normal metabolic adaptation to extrauterine life also involves the use of ketone bodies as alternative substrates for cerebral oxidative metabolism, the newborn infant, however, has a limited ability to mobilize these nutrients [6]. Lactate is also a potential substitute substrate for energy metabolism [32], but it is insufficient to support metabolism as an alternate substrate to replace glucose [33].

Hypoglycemia is also more prevalent soon after birth and its prevalence decrease over the following days. Postnatal age influences hypoglycemia-induced neuronal injury in the rat brain and cellular respiration and ATP consumption have been shown to be lower in young pups than older mice, with this being exacerbated in the absence of glucose [34].

We, therefore, developed a neonatal murine model, with insulin-induced hypoglycemia, to analyze by liquid chromatography with tandem mass spectrometry (LC-MS/MS), the in-vitro effects of hypoglycemia on brain nucleotides (AMP, ADP, and ATP) as well on AEC, throughout the first ten days of life. The cellular AEC was considered a surrogate biomarker for the hypoglycemia-induced brain injury. The study hypothesis was that hypoglycemia-induced brain insult results from depletion of AEC and these effects vary throughout the initial ten days after birth. Our aims were to establish in the first ten days after birth, the following facts: (1) the baseline blood glucose concentration, as well as both the dose of insulin required to induce hypoglycemia and the time required to accomplish this after the initial dose, (2) the concomitant changes in AEC and adenine nucleotide levels in the brain with hypoglycemia, and (3) which of the ATP level or AEC better reflects the effects of hypoglycemia in the brain.

Both the weight and the baseline blood glucose level of all 60 pups, before sacrifice, progressively increased with postnatal age (Table S1 and Figure S1).

Our results confirm that, physiologically, blood glucose levels are low soon after birth and increase progressively in the first week of life. Furthermore, with increasing age of the pups, the development of hypoglycemia required a higher number of successive insulin injections and a higher final insulin dose. We have also established that hypoglycemia causes significant reduction in the cellular AEC by at least 50% in the occipital and parietal lobes resulting in cellular energy failure. None of the adenosine nucleotides accurately reflected cellular energy except in the occipital tissue where the ATP level closely mirrored the cellular AEC from day 1 to day 10.

These findings are consistent with recent reports demonstrating decreased cellular respiration and ATP consumption in the brain of younger mice pups [34]. The decrease in AEC that we demonstrated in parietal and occipital tissues is in keeping with the predominance of parietal and occipital lobes injury demonstrated by imaging in neonatal hypoglycemia [7]. As the glucose content in the brain is directly regulated by blood glucose levels, already low physiologically in the neonatal period, the administration of insulin will lead directly to a further decrease in brain glucose content, resulting in reduction in brain AEC. Mitochondrial respiration, the principal source of energy in the neurons, produces ATP. Insulin-induced hypoglycemia decreases mitochondrial respiratory chain efficiency. Furthermore, it has been demonstrated that recurrent hypoglycemia changes the ratio of mitochondrial respiratory chain complex I subunits, which also increases free radical production and cell death [37]. It is therefore likely that the extensive loss of AEC in hypoglycemic pups, in both the parietal and occipital tissues is caused by acute neuronal decompensation with extensive deterioration of cellular energy status caused by hypoglycemia.

In the parietal tissue, the high AEC of 1.0 remained constant over 10 days in the controls, reflecting the absence of energy loss, with the adenylate pool consisting only of ATP. In contrast, the hypoglycemic pups showed lower level of cellular AEC, indicating the presence of ATP, ADP, and AMP, with a comparatively higher or equal level of AMP, reflecting the underlying metabolic stress caused by hypoglycemia [28, 38].

In the occipital tissue, the initially high AEC of 0.7 to 1.0 in the first 5 days in the controls, with the adenylate pool consisting mainly of ATP, indicates, therefore, the absence of metabolic stress. In contrast, the AEC quickly decreased to very low levels in the hypoglycemic group in close parallel to ATP, with an adenylate pool consisting mainly of AMP and ADP, explained by the underlying metabolic stress caused by hypoglycemia [28, 38]. The reason for the vulnerability of the occipital lobe to hypoglycemia remains unclear. Positron emission tomography has demonstrated that glucose demand in neonatal visual areas is relatively low. Perhaps the thicker the fourth lamina of the visual cortex is thicker with more neurons and synapses, thus requiring more glucose. Furthermore, the proliferating occipital axonal growth and synaptogenesis in the neonatal period require more glucose as substrate. However, as the occipital lobes are supplied with blood (and glucose) from the posterior cerebral artery, their demands are in direct competition with the other metabolically active areas such as the brainstem, cerebellum and part of the thalamus which share the same blood supply [39]. The high rate of energy loss between day 2 and 7, more in the occipital tissue and in hypoglycemic pups, probably results from decreased glucose utilization secondary to reduced blood glucose supply. Another reason for the very low occipital AEC at day 10 might be the rapid maturation of the occipital lobe, with migrating visual fibres that are very sensitive to changes in glucose concentrations.