Some 2000 years ago, the Roman philosopher Seneca the Younger proposed that a human embryo is an adult in miniature, and thus the task of development is simply to grow bigger. This idea of preformation was so appealing that it was widely believed for centuries. Even with the development of the microscope, the appeal of preformation proved so strong that biologists claimed to see microscopic horses in horse semen.

By the mid-1800s, preformation began to lose ground as people realized that embryos look nothing like the adults they become. In fact, it was obvious that embryos of different species more closely resemble one another than their respective parents. The top row of Figure 8-2 shows the striking similarity among embryos of species as diverse as salamanders, chickens, and humans, each shown in fetal form in the bottom row.

As embryos, all vertebrate species have a similar-looking primitive head, a region with bumps or folds, and a tail. Only as an embryo develops does it acquire the distinctive characteristics of its species. The similarity of young embryos is so great that many nineteenth-century biologists saw it as evidence for Darwin’s view that all vertebrates arose from a common ancestor millions of years ago.

249

The embryonic nervous systems of vertebrates are as similar structurally as their bodies are. Figure 8-3 details the three-chambered brain of a young vertebrate embryo: forebrain, midbrain, and hindbrain. The remaining neural tube forms the spinal cord. How do these three regions develop? We can trace the events as the embryo matures.

When a sperm fertilizes an egg, the resulting human zygote consists of just a single cell. But this cell soon begins to divide. By the fifteenth day after fertilization, the emerging embryo resembles a fried egg (Figure 8-4), a structure formed by several sheets of cells with a raised area in the middle called the embryonic disc—essentially the primitive body.

By day 21, 3 weeks after conception, primitive neural tissue, the neural plate, occupies part of the outermost layer of embryonic cells. First the neural plate folds to form the neural groove, detailed in Figure 8-5. The neural groove then curls to form the neural tube, much as you can curl a flat sheet of paper to make a cylinder.

250

The cells that form the neural tube can be regarded as the nursery for the rest of the central nervous system. The open region in the tube’s center remains open and matures into the brain’s ventricles and the spinal canal. The micrographs in Figure 8-6 show the neural tube closing in a mouse embryo.

The human body and nervous system change rapidly in the next 3 weeks (Figure 8-7). By 7 weeks (49 days), the embryo begins to resemble a miniature person. The brain looks distinctly human by about 100 days after conception, but it does not begin to form gyri and sulci until about 7 months. By the end of the ninth month, the fetal brain has the gross appearance of the adult human brain, but its cellular structure is different.

The neural tube is the brain’s nursery. Neural stem cells lining it have an extensive capacity for self-renewal. When a stem cell divides, it produces two stem cells; one dies and the other lives to divide again. This process repeats again and again throughout life. In an adult human, neural stem cells line the ventricles, forming the subventricular zone.

If lining the ventricles were all that stem cells did throughout the decades of a human life, they would seem very odd cells to possess. But neural stem cells have a function beyond self-renewal: they give rise to progenitor cells (precursor cells), which also can divide. As shown in Figure 8-8, progenitor cells eventually produce nondividing cells known as neuroblasts and glioblasts. In turn, neuroblasts and glioblasts mature into neurons and glia. Neural stem cells, then, are multipotent: they give rise to all the many specialized cell types in the CNS.

Sam Weiss and his colleagues (1996) discovered that stem cells remain capable of producing neurons and glia not just into early adulthood but even in an aging brain. This important discovery implies that neurons that die in an adult brain should be replaceable. But neuroscientists do not yet know how to instruct stem cells to replace them.

One possibility is to make use of signals that the brain typically uses to control stem cell production in adults. For example, the level of the neuropeptide prolactin increases when female mice are pregnant and stimulates the fetal brain to produce more neurons (Shingo et al., 2003). These naturally occurring hormonal signals have been shown to replace lost neurons in brain-injured laboratory animals.

251

How does a stem cell know to become a neuron rather than a skin cell? In each cell, certain genes are expressed (turned on) by a signal, and those genes then produce a particular cell type. Gene expression means that a formerly dormant gene is activated so that the cell makes a specific protein. You can easily imagine that certain proteins produce skin cells, whereas other proteins produce neurons.

The specific signals for gene expression are largely unknown but probably are chemical, and they form the basis of epigenetics. A common epigenetic mechanism that suppresses gene expression during development is gene methylation, or DNA methylation. Here a methyl group (CH₃) attaches to the nucleotide base cytosine lying next to guanine on the DNA sequence. It is relatively simple to quantify gene methylation in different phenotypes, reflecting either an increase or a decrease in overall gene expression.

Methylation alters gene expression dramatically during development. Prenatal stress can reduce gene methylation by 10 percent. This means that prenatally stressed infants express 2000 more genes (of the more than 20,000 in the human genome) than unstressed infants (Mychasiuk et al., 2011). Other epigenetic mechanisms, such as histone modification and mRNA modification, can regulate gene expression, but these mechanisms are more difficult to quantify.

Thus, the chemical environment of a brain cell is different from that of a skin cell: different genes in these cells are activated, producing different proteins and different cell types. The chemical environments needed to trigger cellular differentiation could be produced by the activity of neighboring cells or by chemicals, such as hormones, that are transported in the bloodstream.

252

The differentiation of stem cells into neurons must require a series of gene-activating signals. A chemical signal must induce stem cells to produce progenitor cells; another chemical signal must induce the progenitor cells to produce either neuroblasts or glioblasts. Finally, a chemical signal—perhaps a set of signals—must induce the genes to make a particular type of neuron.

Compounds that signal cells to develop in particular ways are neurotrophic factors (-trophic means nourishing). By removing stem cells from an animal’s brain and placing those cells in solutions that keep them alive, researchers can study how neurotrophic factors function. One compound, epidermal growth factor (EGF), when added to the stem cell culture stimulates production of progenitor cells. Another compound, basic fibroblast growth factor (bFGF, or FGF-2), stimulates progenitor cells to produce neuroblasts.

At this point, the destiny of a given neuroblast is undetermined. The blast can become any type of neuron if it receives the right chemical signals. The body relies on a general-purpose neuron that matures into a specific cell type in a particular location when exposed to certain neurotrophic factors.

This flexibility makes brain development simpler than it would be if each different cell type, as well as the number of cells of each type, had to be specified precisely in an organism’s genes. In the same way, building a house from all-purpose two-by-fours that can be cut to any length as needed is easier than specifying in a blueprint a precise number of precut pieces of lumber that can be used only in a certain location.

The human brain requires approximately 10 billion (10¹⁰) cells to form just the cortex that blankets a single hemisphere. This means it must produce about 250,000 neurons per minute at the peak of prenatal brain development. But as Table 8-1 shows, this rapid formation of neurons (neurogenesis) and glia (gliogenesis) is just the first step in brain growth. These new cells must travel to the correct destination (migration), they must differentiate into the right type of neuron or glial cell, and the neurons then must grow dendrites and axons and form synapses.

The brain also prunes unnecessary cells and connections, sculpting itself according to the particular person’s experiences and needs. We consider these stages in brain development next, focusing on cortical development, because neuroscientists know more about development of the cortex than of any other area of the human brain. The principles derived from our examination of the cortex, however, apply to neural growth and development in other brain regions as well.

Neuronal Generation, Migration, and Differentiation

Figure 8-9 shows that neurogenesis is largely complete after about 5 months of gestation. (Some growth continues until about 5 years of age.) An important exception is the hippocampus, where new neurons continue to develop throughout life.

Until after full-term birth, however, the fetal brain is especially delicate and extremely vulnerable to injury, teratogens (chemicals that cause malformations), and trauma. Apparently, the developing brain can more easily cope with injury earlier, during neurogenesis, than it can during the later stages of cell migration or cell differentiation, when cell maturation begins (see Table 8-1). One reason may be that once neurogenesis has slowed, it is very hard to start it up again. If neurogenesis is still progressing at a high rate, more neurons can be made to replace injured ones, or perhaps existing neurons can be allocated differently. The absence of neurogenesis in adulthood, other than in the hippocampus, also explains why adult brain tumors arise from glial cells, which are generated throughout adulthood, rather than from neurons. In contrast, brain tumors in young children are sometimes neuronal, reflecting some lingering neurogenesis.

253

Cell migration begins shortly after the first neurons are generated and continues for about 6 weeks in the cerebral cortex (and throughout life in the hippocampus). Cell differentiation, in which neuroblasts become specific types of neurons, follows migration. Cell differentiation is essentially complete at birth, although neuron maturation, which includes the growth of dendrites, axons, and synapses, goes on for years and in some parts of the brain may continue throughout adulthood.

The cortex is organized into layers distinctly different from one another in their cellular makeup. How does this arrangement of differentiated areas develop? Neuroscientist Pasko Rakic and his colleagues (e.g., Geschwind & Rakic, 2013) have been finding answers to this question for more than four decades. Apparently, the subventricular zone contains a primitive cortical map that predisposes cells formed in a certain ventricular region to migrate to a certain cortical location. One subventricular region may produce cells destined to migrate to the visual cortex; another might produce cells destined to migrate to the frontal lobes, for example.

How do the migrating cells know where to find these different parts of the cortex? They follow a path made by radial glial cells. A glial fiber from each of these path-making cells extends from the subventricular zone to the cortical surface, as illustrated in Figure 8-10A. The close-up views in Figure 8-10B and C show that neural cells from a given subventricular region need only follow the glial road to end up in the correct location.

254

As the brain grows, the glial fibers stretch but still go to the same place. Figure 8-10B also shows a cell moving across the radial glial fibers. Although most cortical neurons follow the radial glial fibers, a small number appear to migrate by seeking some type of chemical signal. Researchers do not yet know why these cells function differently.

Cortical layers develop from the inside out, much like adding layers to a tennis ball. The neurons of innermost layer VI migrate to their locations first, followed by those destined for layer V and so on, as successive waves of neurons pass earlier-arriving neurons to assume progressively more exterior positions in the cortex. Cortex formation is a bit like building a house from the ground up until you reach the roof. The materials needed to build higher floors must pass through lower floors to get to their destinations.

To facilitate house construction, each new story has a blueprint-specified dimension, such as 10 feet high. How do neurons determine how thick a cortical layer should be? This is a tough question, especially when you consider that the cortical layers are not all the same thickness.

Local environmental signals—chemicals produced by other cells—likely influence the way cells form layers in the cortex. These intercellular signals progressively restrict the choice of traits a cell can express, as illustrated in Figure 8-11. Thus, the emergence of distinct cell types in the brain results not from the unfolding of a specific genetic program but rather from the interaction of genetic instructions, timing, and signals from other cells in the local environment.

Neuronal Maturation

After neurons migrate to their destination and differentiate, they begin to mature by (1) growing dendrites to provide surface area for synapses with other cells and (2) extending their axons to appropriate targets to initiate synapse formation.

Two events take place in dendrite development: dendritic arborization (branching) and the growth of dendritic spines. As illustrated in Figure 8-12, dendrites in newborn babies begin as individual processes protruding from the cell body. In the first 2 years of life, dendrites develop increasingly complex extensions that look much like leafless tree branches: they undergo arborization. The dendritic branches then begin to form spines, where most synapses on dendrites are located.

Although dendritic development begins prenatally in humans, it continues for a long time after birth, as Figure 8-12 shows. Dendritic growth proceeds at a slow rate, on the order of microns (µm, millionths of a meter) per day. Contrast this with the development of axons, which grow on the order of a millimeter per day, about a thousand times faster.

255

The disparate developmental rates of axons and dendrites are important, because the faster-growing axon can reach its target cell before the cell’s dendrites are completely formed. Thus the axon may play a role in dendritic differentiation and ultimately in neuron function—for example, as part of the brain’s visual, motor, or language circuitry. Abnormalities in the maturation rate can produce abnormalities in patterns of neural connectivity, as explained in Clinical Focus 8-2, Autism Spectrum Disorder.

256

Axon-appropriate connections may be millimeters or even a meter away in the developing brain, and the axon must find its way through complex cellular terrain to make them. Axon connections present a significant engineering problem for the developing brain. Such a task could not possibly be specified in a rigid genetic program. Rather, genetic–environmental interaction is at work again, as various molecules that attract or repel the approaching axon tip guide the formation of axonic connections.

Santiago Ramón y Cajal was the first scientist to describe this developmental process a century ago. He called the growing tips of axons growth cones. Figure 8-13A shows that as growth cones extend, they send out shoots, analogous to fingers reaching out to find a pen on a cluttered desk. When one shoot, a filopod (pl. filopodia), reaches an appropriate target, the others follow.

Growth cones are responsive to cues from two types of molecules (Figure 8-13B):

Although Ramón y Cajal predicted them more than 100 years ago, tropic molecules have proved difficult to find. So far, only one group, netrins (from Sanskrit for to guide), has been identified in the brain. Given the enormous number of brain connections and the great complexity in wiring them, many other types of tropic molecules undoubtedly will be found.

Synaptic Development

The number of synapses in the human cerebral cortex is staggering, on the order of 10¹⁴, or 100,000 trillion. A genetic program that assigns each synapse a specific location could not possibly determine each spot for this huge number. As with all stages of brain development, only the general outlines of neuronal connections in the brain are likely to be genetically predetermined. The vast array of specific synaptic contacts is then guided into place by a variety of local environmental cues and signals.

A human fetus displays simple synaptic contacts in the fifth gestational month. By the seventh gestational month, synaptic development on the deepest cortical neurons is extensive. After birth, synapse numbers increase rapidly. In the visual cortex, synaptic density almost doubles between ages 2 months and 4 months and then continues to increase until age 1 year.

Cell Death and Synaptic Pruning

257

To carve statues, sculptors begin with blocks of stone and chisel away the unwanted pieces. The brain does something similar during cell death and synaptic pruning. The chisel in the brain could be a genetic signal, experience, reproductive hormones, stress, even SES. The effect of these chisels can be seen in changes in cortical thickness over time, as illustrated in Figure 8-14, an atlas of brain images. The cortex actually becomes measurably thinner in a caudal–rostral (back-to-front) gradient, a process that is probably due both to synaptic pruning and to white matter expansion. This expansion stretches the cortex, leading to increased surface area, as illustrated in Research Focus 8-1.

The graph in Figure 8-15 plots this rise and fall in synaptic density. Pasko Rakic (1974) estimated that at the peak of synapse loss, a person may lose as many as 100,000 per second. Synapse elimination is extensive. Peter Huttenlocher (1994) estimated it at 42 percent in the human cortex. We can only wonder what the behavioral consequence of this rapid synaptic loss might be. It is probably no coincidence that children, especially toddlers and adolescents, seem to change moods and behaviors quickly.

How does the brain eliminate excess neurons? The simplest explanation is competition, sometimes referred to as neural Darwinism. Charles Darwin believed that one key to evolution is the variation it produces in the traits possessed by a species. Those whose traits are best suited to the local environment are most likely to survive. From a Darwinian perspective, then, more animals are born than can survive to adulthood, and environmental pressures weed out the less-fit ones. Similar pressures cause neural Darwinism.

What exactly causes this cellular weeding out in the brain? It turns out that when neurons form synapses, they become somewhat dependent on their targets for survival. In fact, deprived of synaptic targets, neurons eventually die. They die because target cells produce neurotrophic (nourishing) factors absorbed by the axon terminals that function to regulate neuronal survival. Nerve growth factor (NGF), for example, is made by cortical cells and absorbed by cholinergic neurons in the basal forebrain.

If many neurons compete for a limited amount of a neurotrophic factor, only some can survive. The death of neurons deprived of a neurotrophic factor is different from the cell death caused by injury or disease. When neurons are deprived of a neurotrophic factor, certain genes seem to be expressed, resulting in a message for the cell to die. This programmed process is called apoptosis.

Apoptosis accounts for the death of overabundant neurons, but it does not account for the synaptic pruning from cells that survive. In 1976, French neurobiologist Jean-Pierre Changeux proposed a theory for synapse loss that also is based on competition (Changeux & Danchin, 1976). According to Changeux, synapses persist into adulthood only if they have become members of functional neural networks. If not, they are eventually eliminated from the brain. We can speculate that environmental factors such as hormones, drugs, and experience would influence active neural circuit formation and thus influence synapse stabilization and pruning.

In addition to outright errors in synapse formation that give rise to synaptic pruning, subtler changes in neural circuits may trigger the same process. One such change accounts for the findings of Janet Werker and Richard Tees (1992), who studied the ability of infants to discriminate speech sounds taken from widely disparate languages, such as English, Hindi (from India), and Salish (a Native American language). Their results show that young infants can discriminate speech sounds of different languages without previous experience, but their ability to do so declines in the first year of life. An explanation for this declining ability is that synapses encoding speech sounds not typically encountered in an infant’s daily environment are not active simultaneously with other speech-related synapses. As a result, they are eliminated.

258

Synaptic pruning may also allow the brain to adapt more flexibly to environmental demands. Human culture is probably the most diverse and complex environment with which any animal must cope. Perhaps the flexibility in cortical organization achieved by the mechanism of selective synaptic pruning is a necessary precondition for successful development in a cultural environment.

Synaptic pruning may also be a precursor related to different perceptions that people develop about the world. Consider, for example, the obvious differences in Eastern and Western philosophies about life, religion, and culture. Given the cultural differences to which people in the East and West are exposed as their brain develops, imagine how different their individual perceptions and cognitions may be. Considered together as a species, however, we humans are far more alike than we are different.

An important and unique characteristic common to all humans is language. As illustrated in Figure 8-14, the cortex generally thins from age 5 to 20. The sole exception: major language regions of the cortex actually show an increase in gray matter. Figure 8-16 contrasts the thinning of other cortical regions with the thickening of language-related regions (O’Hare & Sowell, 2008). A different pattern of development for brain regions critical in language processing makes sense, given language’s unique role in cognition and the long learning time.

The imaging atlas in Figure 8-14 confirms that the frontal lobe is the last brain region to mature. Since the atlas was compiled, neuroscientists have confirmed that frontal lobe maturation extends far beyond its age 20 boundary, including in the dorsolateral prefrontal cortex. The DLPFC, which comprises Brodmann areas 9 and 46, makes reciprocal connections with posterior parietal cortex and the superior temporal sulcus: it selects behavior and movement with respect to temporal memory. Zdravko Petanjek and colleagues (2011) analyzed synaptic spine density in the DLPFC in a large sample of human brains ranging in age at death from newborn to age 91 years.

The analysis confirms that dendritic spine density, a good measure of the number of excitatory synapses, is two to three times greater in children than in adults and that spine density begins to decrease during puberty. The analysis also shows that dendritic spines continue to be eliminated well beyond age 20, stabilizing at the adult level around age 30. Two important correlates attend slow frontal lobe development:

Astrocytes and oligodendrocytes begin to develop after most neurogenesis is complete and continue to develop throughout life. CNS axons can function before they are myelinated by oligodendria, but healthy adult function is attained only after myelination is complete. Consequently, myelination is a useful rough index of cerebral maturation.

In the early 1920s, Paul Flechsig noticed that cortical myelination begins just after birth and continues until at least 18 years of age. He also noticed that some cortical regions were myelinated by age 3 to 4 years, whereas others showed virtually no myelination at that time. Figure 8-18 shows one of Flechsig’s cortical maps with areas shaded according to earlier or later myelination.

Flechsig hypothesized that the earliest-myelinating areas control simple movements or sensory analyses, whereas the latest-myelinating areas control the highest mental functions. MRI analyses of myelin development in the cortex show that white matter thickness largely does correspond to the progress of myelination, confirming Flechsig’s ideas. Myelination continues until at least 20 years of age, as illustrated in Figure 8-19, which contrasts total brain volume, gray matter volume, and white matter volume during brain development in females and males.

260

Neurobiology of Development

Before you continue, check your understanding.

Question 1

Question 2

Question 3

Question 4

Question 5

Answers appear in the Self Test section of the book.