Bridging a mechanistic gap from diet to synapses

Human milk, vital for infant nutrition, is a complex blend of nutrients and bioactive components (1). It meets the high energy demands of babies, supporting postnatal growth through proteins, fats, and lactose. While its composition is generally consistent, human milk adjusts macronutrient levels to meet a growing child’s needs. It also provides micronutrients influenced by maternal diet. Besides nutrition, human milk contains nonnutritive bioactive factors promoting infant survival and development. An example is leptin, which aids metabolic programming and protects against later-life obesity. Colostrum and mature milk have bioactive molecules supporting gut function and protecting against infections (1).

It is fascinating that human milk has the optimal composition to fuel rapid brain development in the first 1,000 days, igniting remarkable cognitive progress and programming a lifelong cognitive trajectory. While some beneficial effects of breastfeeding on the construction of cognition in babies have been attributed to crucial nutrients like polyunsaturated fatty acids (PUFAs), human milk oligosaccharides, and sialic acid, the tantalizing impact of other nutritional factors on the brain awaits further exploration.

It is in this context, a report by Paquette and colleagues (2) lays out a strong basis by which to understand the contribution of micronutrients to synaptogenesis. This advancement holds significance in the field of pediatric nutrition, specifically for enhancing infant formulas used in situations where adequate breastfeeding is hindered. They begin by comparing breastmilk composition from samples taken across time from three geographically distinct sites—Mexico City, Shanghai, and Cincinnati—with the goal of identifying micronutrients that are shared, regulated developmentally, and have peak concentrations coinciding with a major increase in cortical synapse density. Myo-inositol (MI) fulfills these criteria.

Inositol has long been recognized as an intriguing component of human milk having multiple potential roles in neonatal development. It is a polyol characterized by a six-carbon ring with each carbon hydroxylated. Among the various isomers, MI is the most common and biologically active (3). It is generally assumed that a regular diet, along with some endogenous synthesis, supplies eukaryotic cells with an adequate amount. Dietary MI primarily originates from phosphatidylinositol (PI) found in plant and animal food sources, and from inositol polyphosphates, particularly phytic acid, which is abundant in many seeds (4). MI serves as a precursor for the production of PI, which is essential in all eukaryotes as a structural component of cell membranes and for its involvement in the synthesis of several other vital lipid molecules, including sphingolipids, ceramides, and glycosyl-PI anchors (3). Independent of this role, free MI acts as an osmolyte, helping cells maintain their osmolarity in hyperosmolar environments like the brain (5). In a healthy human fetus, MI levels estimated using short echo-time magnetic resonance spectroscopy (MRS) decline gradually from a peak at 22 wk through birth (6) and appear to plateau after the second year (7). This is a period that is dominated by synaptogenesis, and the generation of astrocytes and oligodendrocytes, and corresponds roughly to the first few postnatal weeks of development in rodents. In postnatal mammals, high MI levels derive in part from its enrichment in maternal milk early in lactation (2). But how does MI contribute to normal neurodevelopment? (Fig. 1).

Fig. 1.

Nutrients provided in breastmilk aid cognitive development (solid arrow). The impact they have on synapses likely occurs through modifications of lipid bilayer composition, particularly within synapses (boxed region in transmission electron microscope image of glutamatergic synapse). The figure was made in BioRender.com using an image captured on a Hitatchi 7700 microscope in the Microscopy and Advanced Bioimaging Core at Mount Sinai.

Paquette and colleagues focus on the effect of MI on synaptogenesis (2). Their data show that in human iPSCs differentiated to a cortical glutamatergic identity, there is a direct, positive impact of MI on the size and intensity of immunolabeling for Homer, a ubiquitous, excitatory postsynaptic density (PSD) protein. There was no impact on presynaptic terminals as might be expected, perhaps due to the protracted maturation or absence of astrocytes in such preparations. Consistent with this, in more rapidly developing and astrocyte-free cultured rat hippocampal neurons, there was a dose-dependent effect of MI on excitatory synapse density. Thus, MI can enhance synaptogenesis directly in neurons. It would be interesting to know the degree to which MI is required but removing exogenous MI stunted overall neuronal growth and differentiation. Thus, MI can promote synaptogenesis and can do so directly. But does it actually do so? After five weeks of oral MI supplementation in developing mice, immunolabeled clusters of Homer are enlarged in the visual cortex, but no differences in density of either pre- or postsynaptic markers suggest either that accumulation of Homer is selectively greater or that postsynaptic densities enlarge asymmetrically. Either result adds nuance to the assumption that MI simply promotes synaptogenesis, and it will be important in future work to assess synapse strength and impact on visual acuity or cognition. In general, however, the data build a strong case supporting that MI imparts a general synaptogenic effect during development.

A report by Paquette and colleagues lays out a strong basis by which to understand the contribution of micronutrients to synaptogenesis.

The study does not address the route by which MI accesses synapses. At early postnatal ages, exogenously supplied MI probably transits to the brain through leaky intestinal and blood–brain barriers. This could be confirmed by the combined use of hyperpolarized [¹³C₆]-MI and MRS (8), where it would be predicted that access to brain parenchyma from the vasculature would decline as barriers tighten. [¹³C₆]-MI labeling of neuronal cultures and NMR spectroscopy could be used to identify and compare relative levels of phosphoinositides and inositol phosphates over time and assess the impact of increased MI levels. Since phosphoinositides are differentially enriched in organelles, fluorescent lipid biosensors could be used to probe MI-directed changes in intracellular lipid composition or localization using microscopy (9).

Once access to neurons is gained, it will be important to evaluate how MI affects the process of synapse assembly and stabilization. Synapse lipid composition is highly dynamic and regulates function in a manner that is best understood presynaptically, where local, high demand for vesicle recycling places a premium on regulatory events that serve to bend and reorganize lipids rapidly and thereby signal retrieval, trafficking, or fusion. Gene mutations that modify synaptic enzymes important for lipid modification, like Synaptojanin1, impair vesicle endocytosis and trafficking and can lead to serious neurological and developmental brain disorders (10). Synaptic lipid composition also changes more slowly over the course of development with time-dependent increases in canonical lipid raft components and n-3 PUFAs. A model based on such work suggests that lipid-based patterning within synapses could guide the seeding and accumulation of synaptic protein scaffolds (11). This idea gains some support from work in mice that shows enhanced synaptogenesis in the visual cortex by n-3 PUFA supplementation and reduced glutamatergic synapse density in the hippocampus in response to a diet deficient in n-3 PUFA (12, 13). These findings underscore the relevance of lipid composition for normal synapse development, and it will be important to integrate such results with the impact of MI supplementation as outlined by Paquette and colleagues.

The effects of MI given through diet in vivo on synapses were more modest than those observed following direct application in cultured neurons (2). This is likely to reflect (at least in part) contributions of nonneuronal cells in vivo, which were not addressed experimentally. MI was first detected and is enriched in astrocytes, and since astrocytes regulate synapse assembly, neurotransmission, and elimination (14), it is anticipated that they will also mediate effects of MI on synapses. For example, MI could impact inositol 1,4,5-trisphosphate receptor-dependent release of ATP which can trigger synapse pruning in astrocytes (15). In this context, it is interesting that reductions in synapse density that occur in response to a deficient n-3 PUFA diet are mediated by changes in microglial lipid composition that drive microglia to prune excessively (13), and both astrocytes and microglia require PI kinase–directed changes in phospholipid composition to form and internalize phagosomes to effect pruning (3).

In light of Paquette’s findings on the crucial role of MI provided by breast milk in synapse development, it appears important to be mindful of nutritional and medical situations that reduce the bioavailability of MI for the baby. Pharmacological treatment of pregnant/lactating women with the mood stabilizer lithium (LiCl), Western Pattern Dietary profile, which is low in inositol and phytate, and gestational diabetes reduce MI synthesis and its availability in maternal milk (5, 16). As a result, these conditions and MI depletion represent cumulative risk factors for the alteration of brain development and associated cognitive deficits. In mice, a high-fat diet-induced inositol deficit and associated insulin resistance are improved by MI supplementation (17) and some work supports a similar positive impact in humans (16). This suggests that early MI supplementation may benefit brain development of infants from mothers treated with LiCl or suffering from diabetes as suggested by experimental results obtained in animal models (16–18). However, the generally positive effects of increasing MI bioavailability are countered by the abnormally high brain MI levels observed in people with Down syndrome, where MI levels rise due to overexpression of a MI transporter encoded by chromosome 21 (5). Heightened MI levels in the brain are also observed in patients with Alzheimer’s disease and following brain injury where they have been correlated with poorer outcomes (19), suggesting that supplementation in some cases could be detrimental. Such data underscore the need for additional research, and it is encouraging that there is a growing interest in providing mechanistic underpinnings to observational data linking diet and cognitive development. The original results and methods developed by Paquette et al. offer a means by which to unravel the intriguing relationship between MI and synapse formation especially in the adult brain and in different medical conditions.