The hippocampus plays an important role in human learning and memory and is known to be vulnerable to the effects of stress and disease.1 Within animal models of type 1 diabetes,2 significant hyperglycemia and hypoglycemia exposure has been shown to cause complex molecular and structural changes in the hippocampus.2,3 Within humans with type 1 diabetes, exposure to both extremes of the glycemic spectrum have been inconsistently associated with alterations in hippocampal structure4–9 and worse memory function,8–11 but it is unclear whether these two findings relate to each other. Given that the diagnosis of type 1 diabetes and exposure to glycemic extremes often occur at a time of dynamic neurodevelopment, it has been hypothesized that exposure at early ages may have different effects on the brain than exposure in adulthood, which could explain some of these inconsistent findings. However, data addressing this concept are limited. In this article, we review the existing literature on the effects of hyperglycemia and hypoglycemia in type 1 diabetes, with focus on the structure and function of the hippocampus during both development and adulthood.
Hippocampal Structure, Volume, and Function
The hippocampus is a subcortical, primarily gray matter structure residing bilaterally within the medial temporal lobes. It consists of several regions, including the cornu ammonis (CA) and its subregions (CA1–CA4), the dentate gyrus (DG), and the subiculum, which connects the hippocampus to the parahippocampal gyrus.1 These regions contain intra- and interhippocampal connections through multiple synaptic pathways12 and have been ascribed specialized roles in the processing of information.13–16 Overall, the hippocampus appears to play an important role in laying down and consciously retrieving explicitly learned, novel information.17 Within this complex process, the CA1 has been selectively associated with long-term and autobiographical memory,15,18 the CA2 and CA3 with encoding processes, and the DG with early retrieval and episodic memory formation.15,19
The hippocampus develops heterogeneously and nonlinearly up to age 2520,21 and is a site of sustained neurogenesis throughout the lifespan.20–22 Hippocampal volume is developmentally dynamic, with the posterior and anterior portions increasing and decreasing respectively with age.20The total volume steeply increases until approximately age 4, followed by a gradual increase and reaching a peak in volume around age 10.21 Hippocampal volume remains fairly stable until age 50 in healthy adults, followed by a variable level of decline and a significant decrease in volume by age 80.23 The relationship between developmental changes in hippocampal volume and memory function is complex and still unclear. A meta-analysis of the literature examining hippocampal volume and memory performance in healthy children, adolescents, and young adults revealed that smaller hippocampal volume was associated with better memory performance and that this effect could reflect sufficient developmental pruning leading to overall lower volume.24 In healthy adults, the associations are less clear as both larger25–28 and smaller29,30 volumes have been associated with better explicit memory performance. The dynamic development of the hippocampus could lead to age-selective vulnerability and differential outcomes in patient populations. In adult onset disease states, such as Alzheimer’s disease, smaller volumes are typically related to worse performance.31,32 However, in developmental injury populations, such as prenatal drug exposure, larger volumes havebeen related to worse performance.33 Additionally, adults with early (diagnosed anytime before age 50) and late onset depression have reduced hippocampal volume and memory impairment compared with controls, but those with adult onset are more severely affected than those with earlier onset,34 suggesting that the effects of pathology are also dependent on the stage of development. Hippocampal volumes have been reported to be smaller than controls in adults with schizophrenia,35,36 adult onset epilepsy,37,38 Alzheimer’s disease,39–41 and in both adult42,43 and adolescent onset depression.44 Elevated levels of stress hormones, common in many neuropsychiatric disorders, such as depression, have been associated with smaller hippocampal volume in adults.45 The explanation for hippocampal atrophy from chronic exposure to these hormones appears to be loss of dendrites and synapses within the hippocampus.46 However, larger volumeshave been reported in certain developmental disorders including fragile X syndrome,47 autism,48 and attention-deficit hyperactivity disorder.49 The reasons for hippocampal enlargement are less clear, but a mouse model of fragile X syndrome suggests that longer, immature dendritic spines on hippocampal neurons and reduced pruning could lead to excess volume.50Adult subjects with post-traumatic stress disorder (PTSD) have smaller hippocampal volumes51,52 yet youth and adolescents with PTSD have larger hippocampal volumes compared with controls.53 None of these volumetric differences have been compared with hippocampus-dependent memory performance, leaving the correlation between volume and function in different patient populations still uncertain. Clearly, hippocampal volume is differentially affected by the stage of development in which the pathology occurs. These complex, developmental patterns in the hippocampus are highly relevant to the study of type 1 diabetes, as this disorder is typically diagnosed in childhood and has an impact on blood glucose levels that continues throughout development.
Hippocampal Structure, Volume, and Function in Type 1 Diabetes
The brain relies on a consistent supply of glucose for proper function. Individuals with type 1 diabetes experience both hyperglycemia from disrupted insulin production and hypoglycemic episodes as a complication of insulin treatment. Animal studies have demonstrated that hyperglycemia can affect cell number and anatomy in the hippocampus2,54 and hypoglycemia causes neuronal death, particularly in the DG and CA1 regions.3 Case reports suggest that adults with type 1 diabetes who experienced profound hypoglycemic episodes can result in hippocampal lesions or smaller hippocampal volumes accompanied by cognitive deficits including anterograde amnesia, which is the selective inability to encode new information for later conscious recall.8,9 Severe hypoglycemia can lead to cell death due to an excess release of excitatory neurotransmitter. Due to a high concentration of excitatory neurotransmitter receptors in the hippocampus, this excitotoxic cascade quickly leads to neuronal necrosis55 and without the ability to overcome this cell loss, a reduction in overall volume could result.
Based on these animal studies and case reports, larger group studies have looked for macrostructural differences in the hippocampus of the general type 1 diabetes population, and asked whether hippocampal volumes are related to exposure to glycemic extremes (see Table 1). Lobnig et al.7 examined a small sample of adults with type 1 diabetes and found mild cerebral atrophy and evidence of impaired cognitive performance compared with age- and gender-matched controls, but no differences in hippocampal volume or memory function. Within the type 1 diabetes group, only three individuals had severe hypoglycemic episodes in the past, which significantly limited the power of analyses to detect any potential relationships with hypoglycemia exposure. To our knowledge, no other currently published studies have examined hippocampal volume in adults with type 1 diabetes. To understand the long-term effects of diabetic complications on hippocampal development and function, it will be important to have larger and more varied exposures to hypoglycemia in future studies.
Since type 1 diabetes typically has a childhood onset and some extreme glycemic states occur more frequently in childhood (e.g. severe hyperglycemia onset, severe hypoglycemia), more recent studies have focused on examining the hippocampus in youth with type 1 diabetes (see Figure 1). Hershey et al.4 examined a large sample of youth with type 1 diabetes and compared them with their siblings without diabetes. No between-group differences were found, but hippocampal gray matter volume was larger in those children with type 1 diabetes who experienced three or more severe hypoglycemic episodes in the past compared with all other type 1 diabetes subgroups or controls. This effect was not explained by hyperglycemia exposure, age of onset of type 1 diabetes, age, or gender. In a younger and smaller sample, Aye et al.5 found no significant differences between the type 1 diabetes and control groups, or between type 1 subjects with diabetes with and without hypoglycemia-related seizures in hippocampal volume. However, this study did report a positive correlation between hippocampal volume and age in the control group that was not present in children with type 1 diabetes, suggesting altered hippocampal development in these patients. An additional study by Ho et al.56 found no differences in hippocampal volume between youth with type 1 diabetes who had experienced hypoglycemia-related seizures compared with those who had not. A healthy control group was not included for comparison.
It is unknown how severe hypoglycemia in type 1 diabetes could lead to an enlarged hippocampus in children, but speculation can be made based on the current literature. In animal studies, knocking out a presynaptic scaffolding protein can result in a significantly larger hippocampus.57 While this presynaptic protein may not be relevant to hypoglycemia, this study did find that the enlarged volume was due to uncontrolled neurogenesis and gliosis, seen by an increase in the number of both neurons and glia, as well as reduced cell death and an increase in brain-derived neurotrophic factor (BDNF).57 Brain insults, such as severe hypoglycemia, have been shown to increase expression of BDNF and other growth factors,58 likely as a compensatory mechanism, and levels of these neurotrophic factors fluctuate throughout development with neurogenesis and are particularly high in the hippocampus.59 Moderate, chronic overexpression of BDNF leads to longer, more complex dendrites in hippocampal granule neurons60 that could lead to excess volume without sufficient pruning. Reactive gliosis, neurogenesis, insufficient pruning, and compensatory responses to hypoglycemia could all be involved in hippocampal enlargement and will only be determined by further mechanistic experiments using animal models of diabetes and hypoglycemia. Notably, no relationship was found between measures of hyperglycemia exposure and hippocampal volumes in three of these studies.4,5,7 By contrast, Antenor-Dorsey et al.6 recently used diffusion tensor imaging (DTI) to examine white matter microstructural integrity in adolescents with type 1 diabetes and found that greater exposure to severe hyperglycemic episodes (most involving documented diabetic ketoacidosis) was associated with increased radial diffusivity in the hippocampus, suggesting that anatomical differences other than overall volume could be present in children with type 1 diabetes. Thus, within children with type 1 diabetes, extreme glycemic exposure in the form of severe hypoglycemia or severe hyperglycemia may have a greater impact on the macrostructure of the hippocampus than lower level, chronic exposure. This pattern is in contrast to adults and adolescents with chronic low level hyperglycemia from type 2 diabetes, and no severe glycemic exposure, who tend to have smaller hippocampal volumes compared with controls.61,62 In adults with type 2 diabetes, glycated hemoglobin (HbA1c) levels have been negatively correlated with hippocampal volume;63 however, in an adolescent population, these two factors were not associated which may indicate age-dependent effects from chronic exposure to hyperglycemia on the hippocampus.62 Animal models suggest that chronic hyperglycemia can reduce the complexity of neuronal dendrites and dendritic spines54 and damaging these neuronal processes could lead to a smaller hippocampal volume,46 but developing adolescents might be able to overcome this loss due to the presence of growth factors or other developmentally regulated molecules.
Although type 1 diabetes and exposure to glycemic extremes does not appear to have striking macrostructural effects on the hippocampus in the current human literature, it is possible that hippocampal function could still be compromised. In animals, hyperglycemia is associated with impairment in explicit memory function.64 Limited neuroimaging studies have examined hippocampus-dependent memory tasks in patients with type 1 diabetes, but one functional magnetic resonance imaging (fMRI) study found that acute hypoglycemia is associated with higher brain activation in the hippocampus and other brain areas during working memory tasks, suggesting a decrease in cerebral efficiency, although this was not examined in hippocampus-mediated explicit memory tasks.11 Additionally, in humans, severe hypoglycemic episodes are associated with decreased long-term spatial memory; effects are largest when the severe hypoglycemic episodes were experienced before age five.10 This finding supports the concept that the hippocampus may be differentially vulnerable depending on the developmental stage at which hypoglycemia occurs. This possibility is also supported by the finding that adult rats show more hypoglycemic cell death in the hippocampus than the young,65 but in young rats, long-term potentiation (a presumed mechanism of learning and memory) is severely decreased in the hippocampus with hypoglycemia.66
The hippocampus is an important, complex structure that develops heterogeneously, making it difficult to differentiate the effects of altered glycemic states in childhood on its structure and function. Overall, both hypoglycemic and hyperglycemic episodes have been correlated with structural changes in the hippocampus,4,6 but how these changes relate to memory function is still unclear. Correlations between hippocampal volume and cognitive performance in youth and adults with type 1 diabetes will be necessary and are currently under way. Future cross-sectional studies should continue to include healthy controls and also examine differences within type 1 diabetes based on extreme glycemic exposure at both ends of the spectrum. The age of type 1 diabetes onset and age at which severe glycemic extremes occur also should be considered when examining both hippocampal volumes and memory function in children and adults. Whether structural differences in the hippocampus of children can be overcome with continued development is unknown but it is an important question to address in future, longitudinal experiments. Finally, it is important to mention that glucose is not the only substrate dysregulated in type 1 diabetes that could affect hippocampal volume and function. Insulin also varies significantly in type 1 diabetes patients and the hippocampus has a high density of insulin receptors that can affect memory.67 Continued work in this area is crucial to increase our understanding of the complex interplay between hippocampal development, volume, and function and the fluctuating metabolic status of individuals with type 1 diabetes. Ultimately, this work will improve our ability to predict the neurobiologic and functional risks that these individuals face, providing a rational basis for evaluation, prevention, and treatment.