Maturation and Dysgenesis of the Human Olfactory Bulb

Abstract

The olfactory bulb with its unique architecture was studied for neuronal maturation in human fetuses. Neuroblasts stream into the olfactory bulb from the rostral telencephalon and secondarily migrate radially. The transitory olfactory ventricular recess regresses postnatally. Olfactory is the only sensory system without thalamic projections but incorporates intrinsic thalamic equivalents. The bulb is a repository of progenitor cells. Maturation of the bulb and tract was studied in 18 normal human fetuses of 16–41 weeks gestation; mid‐gestational twins with hydrocephalus; 7 arrhinencephaly/holoprosencephaly; 2 olfactory dysgeneses. Multiple immunoreactivities were performed. Synaptophysin around mitral neurons, in a few synaptic glomeruli and concentric lamination of the outer granular layer, was seen at 16 weeks. Outer granular neurons exhibited NeuN at 16 weeks, only 2/3 were reactive at term. Concentric alternating sheets of granular neurons and their dendrodendritic synapses are seen during maturation. Calretinin reactivity is seen in neurons and neurites, primary olfactory nerve axons, periglomerular cells and neuroepithelial cells surrounding the ventricular recess; reactivity occurs later in synaptic glomeruli than with synaptophysin; not all glomeruli are strongly reactive even at term. Nestin‐ and vimentin‐reactive bipolar progenitor cells were demonstrated at all ages and extend into the olfactory tract. Myelin is demonstrated by Luxol fast blue (LFB) only postnatally. In hydrocephalus, the olfactory recess is dilated. Mitral cell dispersion, disrupted glomeruli, heterotopia and maturational delay are seen in some dysgeneses. Malformations exhibit unique findings. Fusion of hypoplastic bulbs can occur. Abnormal architecture is seen in hemimegalencephaly. More documentation of olfactory dysgenesis is needed in other major brain malformations.

Introduction

The architecture and synaptic circuitry of the olfactory bulb are unlike any other laminar cortex of the brain. Olfactory structures develop early in the fetus, both macroscopically and histologically, but reports using modern immunocytochemical markers of proteins that help define individual neuronal maturation are sparse. Primordial olfactory bulbs appear at 41 days (4.5 weeks) gestation as ventro‐rostral outgrowths of the newly formed telencephalic hemispheres following sagittal cleavage of the prosencephalic vesicle 71, 78. Developmental morphogenesis is unique because none of the intrinsic neurons of the bulb originate there; they stream in from the subventricular neuroepithelium of the main telencephalic hemisphere. The olfactory is the only special sensory system without synaptic relay to neocortex via the thalamus because the olfactory bulb incorporates its own thalamic equivalent, in three components: (i) the core of intrinsic axonless granular interneurons and (ii) periglomerular interneurons, both of which form dendrodendritic synapses, and (iii) the anterior olfactory nucleus. The olfactory tract is much more than a simple white matter fascicle. In addition to axonal projection fibers and a few anterior commissural afferents from the contralateral olfactory bulb, it also includes longitudinal bundles of progenitor cell processes, a caudal extension of granular neurons from the olfactory bulb and a rostral extension of pyramidal neurons of the anterior olfactory nucleus, singly and in clusters. Primary olfactory neurons residing in the nasal epithelium exhibit a turnover and extraordinary regenerative capacity unmatched in any other structure of the fetal or adult brain.

Despite these many unique features, there is little interest by clinical neurologists or neuropathologists in even examining the olfactory system for diagnostic purposes. Neuropathologists usually limit their report of the olfactory bulb to noting its gross presence at autopsy, only rarely submitting it for histological sections even from brains with major malformations. A recent exception is adult brains in which tauopathy is suspected because the olfactory bulb also is affected. Adult neurologists rarely test olfaction during examination of the cranial nerves, again unless dementia is the reason for consultation, and pediatric neurologists rarely ever test smell. Neuroradiologists inconsistently describe the olfactory bulbs in magnetic resonance imaging (MRI) reports. Nevertheless, there are reasons to renew interest in this structure. It is one of the two principal repositories of resident progenitor cells in the mature human brain, with a potential for homologous transplantation for neural regeneration in damaged regions of the central nervous system (CNS). The olfactory system may be altered in some brain malformations and genetic disorders, but reports of olfactory dysgenesis are few.

From an historical perspective, the unique neuroanatomy of the olfactory bulb was recognized by neuroscientists of the mid‐19th century. It was the first structure of the human brain in which the intrinsic organization was first elucidated by silver impregnation, by Camillo Golgi himself in 1875 33 (Figure 1). The contact of peripheral olfactory nerve fibers with processes of mitral cells extending into the glomeruli of the olfactory bulb had been described even earlier by Owsjannikow in 1860 75 and Walter in 1861 121 and was further defined in the late 19th and early 20th centuries by many early developmental neuroanatomists 12, 28, 29, 30, 38, 48, 49, 50, 53, 69, 81, 84, 85, 86, 88, 91. The olfactory bulb served as an important model for demonstrating synaptic architecture of the CNS. Recent histological and Golgi impregnation studies of olfactory bulb architecture confirm the findings of the early investigators and also provide a basis for correlations with clinical disorders of olfaction 72, 116.

Figure 1.

The first depiction of human brain as seen microscopically with silver impregnation is this original camera lucida drawing of the olfactory bulb by Camillo Golgi in 1875 [37]. Note the meticulous detail of the synaptic glomeruli but that they are illustrated as a syncytium rather than as synaptic contacts, corresponding to the belief of Golgi about the architecture of the central nervous system. Periglomerular neurons, which we now know are GABAergic inhibitory interneurons, also are illustrated. The olfactory glomerular surface faces ventrally, but the figure was purposely demonstrated upside‐down so that it could be compared with the cerebral and cerebellar cortices that conventionally are illustrated in this way.

Olfaction is the earliest special sense to evolve: even polyps and jellyfishes perceive noxious odors and chemicals in their surrounding water and retract in response, as noted in 1849 by Thomas Huxley 40. The basic laminar architecture of the olfactory bulb and its afferent and efferent connections are constant in all vertebrates 2, 56, 106. An exception is that mitral cells usually contact several olfactory glomeruli in non‐mammalian vertebrates but project to only one glomerulus in mammals 56. The ultrastructural architecture of the olfactory bulb also is almost identical in all vertebrates 2, 56. The olfactory recess of the rostral lateral ventricle is a transitory fetal structure in mammals, but in some classes of vertebrates, it persists into adult life 106. The olfactory bulb, one of the earliest structures of the vertebrate brain to develop, has retained a constant stability throughout phylogeny.

In 1940, the distinguished neuroanatomist Tryphena Humphrey asserted that the olfactory bulb was mature by 11 weeks gestation 39, reinforced in the classical 1962 neuroanatomy textbook co‐authored by her 20 and by a 1982 study of olfactory epithelium by Pyatkina 82. Their interpretations were based on histology without the advantage of immunocytochemistry at that time. However, even some simple histological features indicate structural immaturity of the olfactory bulb at birth and postnatal maturational changes also are poorly documented. The purposes of the present study are (i) to demonstrate the normal sequence of neuronal maturation in the human fetal and neonatal olfactory bulb and (ii) to offer examples of olfactory bulb dysgenesis.

Structural organization of the olfactory bulb

Because the olfactory bulb is unique in its lamination and architecture, it is useful to precede a description of our findings with an introduction to its normal histology during development and at maturity. The concentric laminar architecture of the mature olfactory bulb is well characterized 5, 51, 76, 77. The histology of the fetal olfactory bulb during development also is documented 39, 72, 116. The cortex of the olfactory bulb consists of six layers parallel to the surface of the bulb, but none corresponds to those of the hexalaminar cerebral neocortex, to the layers of the cerebellar cortex, or even to the looser lamination of the superior colliculus. Table 1 summarizes the structures in each layer of the olfactory bulb.

Table 1.

Lamination of the human olfactory bulb. Abbreviation: GABA = gamma‐aminobutyric acid

Layer	Components	Comments
1	Axons of primary neurons in olfactory epithelium	Acellular; unmyelinated, calretinin‐reactive axons
2	Synaptic glomeruli; periglomerular interneurons	Glomeruli form at 14 weeks on the ventral surface of the bulb facing the cribriform plate, with precise synaptic ratios between olfactory axons and mitral cell dendrites for amplification; periglomerular cells co‐express GABA and dopamine
3	External plexiform layer	Tufted neurons at rostrally end, otherwise cell‐sparse; dendrodendritic synapses: mitral to granule excitatory; granule to mitral inhibitory
4	Mitral neuronal cell layer	Large mitral cells form at 9 weeks.; axons extend into olfactory tract
5	Internal plexiform layer	Neurites; cell‐ and synapse‐sparse
6	Granular layer	GABAergic axonless granular neurons in core of olfactory bulb; form multiple concentric laminae alternating with sheets of dendrodendritic synapses; main reservoir of resident progenitor cells; ventricular recess in fetus, neonate

Materials and Methods

Olfactory bulbs were taken at autopsy from 18 human fetuses and neonates of 16–41 weeks gestation. All had grossly and microscopically normal brains, except for minor lesions such as small germinal matrix hemorrhages. Gestational (ie, post‐conceptional) ages of fetuses were determined by obstetrical data, brain weights and standard fetal anthropometric measurements, including crown–rump length, foot size and other external morphological features. The sexual differential of the fetuses were: 11 male and 7 female. Cause of death in most cases was congenital cardiac disease (ie, hypoplastic left heart, pulmonary artery stenosis, atrioconal anomalies). None had recognizable genetic syndromes. Severe post‐mortem autolysis and artifactual crushing or tearing of the olfactory bulbs during removal of the brain constituted the exclusion criteria. Three postnatal infants of 3–15 months of age also were studied and are an initial part of a larger series of postnatal olfactory bulb maturation being collected to be reported later, but these initial cases provided some basis for late maturational changes such as myelination.

Representative pathological cases included mid‐gestational monozygotic twins with fetal hydrocephalus, seven cases of fetal holoprosencephaly of 13.5–31 weeks gestation, one hemimegalencephaly (HME) and one complex brain malformation with fusion of the olfactory bulbs.

Parents consented to pathological examination of the products of conception in accordance with provincial and hospital standards and with approval of the Conjoint Research Ethics Committee of the University of Calgary and Alberta Health Services.

Laboratory techniques

Olfactory bulbs were removed, together with 2–4 mm of attached olfactory tract, at the time of macroscopic neuropathological examination, before brains were cut. Bulbs were placed in a plastic cassette in preparation for sectioning in the horizontal plane, including the rostral end. In about half of the cases, one of the two bulbs was sectioned transversely. Before removal, the ventral surface that contains most synaptic glomeruli was marked by ink to indicate to the technologist that this was the side to be sectioned. They were fixed in 10% buffered formalin in preparation for paraffin embedding and sectioning at 6 μm.

In addition to hematoxylin–eosin (H&E) stain, Luxol fast blue myelin stain and Bielschowsky silver impregnations were prepared. Sections were prepared for immunoreactivities using antibodies against: (i) synaptophysin (Novocastra Laboratories, Newcastle upon Tyne, UK and distributed through Vison Biosystems, Norwell, MA, USA; NCL‐SYNAP‐299; 1:100 dilution with thermal intensification using an HIER steamer; automated reactivity was performed on Ventana Nexes IHC); (ii) neuronal nuclear antigen (NeuN; Chemicon‐Millipore, Temecula, CA, USA; 1:40 dilution); (iii) calretinin (ie, calbindin‐2) (Zymed, Camarillo, CA, USA; 1:100 dilution); (iv) vimentin (DakoCytomation, Copenhagen, Denmark; 1:400 dilution); (v) nestin (Thermo Scientific, Sacramento, California, USA, MA1‐46100; 1:400 dilution); (vi) glial fibrillary acidic protein (GFAP; DakoCytomation, monoclonal antibody at 1:1000 dilution and polyclonal antibody, 1:4000 dilution); and (vii) Ki67 proliferating cell nuclear antigen (MIB1; MyBioSource MBS684110). (viii) Microtubule‐associated protein‐2 (MAP2) also was examined in a few cases. Simultaneously incubated immunocytochemical controls were performed with post‐mortem term neonatal brain tissue that did not show autolysis. Golgi impregnations and electron microscopy were not performed for this study.

Results

Normal maturation of the fetal olfactory bulb

Histological findings (H&E)

The olfactory recess of the lateral ventricle often was seen within the center of the olfactory tract and extending rostrally to the middle of the bulb of fetuses of all ages examined, including full‐term neonates. It was more often eccentric rather than being at the very center of the core of the olfactory bulb. The canal was widely patent and most prominent in fetuses younger than 18 weeks gestation, after which it became narrowed to a thinner canal. The olfactory recess was ependymal‐lined after mid‐gestation, but in younger fetuses, it was a neuroepithelium including scattered mitotic figures at the luminal surface (Figure 2), similar to the wall of the lateral ventricle in fetuses in the first half of gestation 95, 96. The ependyma was a pseudostratified columnar epithelium and, even in the term neonate, did not become a simple cuboidal epithelium as it does at the lateral ventricles.

Figure 2.

Olfactory ventricular recess in fetuses of (A) 16 weeks; (B,C) 19 weeks; (D) 21 weeks; (E) 24 weeks; and (F) 38 weeks gestation. Erythrocytes within the lumen (F) are due to a previous germinal matrix hemorrhage into the lateral ventricle, but the lateral ventricle was only minimally dilated. Before mid‐gestation, the recess is lined by undifferentiated neuroepithelium with scattered mitoses (*) at the luminal surface and early ependymal differentiation (ep) is seen in part of the luminal surface. C. At 24 weeks, a pseudostratified columnar epithelium is differentiated (E), which later thins to a simple cuboidal epithelium by 36 weeks, similar to the progression of ependymal differentiation of the main lateral ventricles. At term, the recess is still present. Involution of the olfactory ventricular recess occurs in the first postnatal weeks, similar to the spinal central canal (not shown). At mid‐gestation (D), a few desquamated ependymal cells are already present in the lumen. It is thus a transitory fetal structure. A–D,F. hematoxylin–eosin. E. Hematoxylin–eosin/Luxol fast blue.

At the surface of the olfactory bulb was a thin amorphous layer of axons from the olfactory epithelium. Synaptic glomeruli were arranged in a loose row beneath this band of neurites, surrounded by neuropil, but glomeruli were difficult to detect with H&E before 18 weeks. Deeper within the bulb, a single cell or in focal places, two‐cell thick row of large mitral and tufted neurons was identified histologically. Internal to the row of mitral cells were smaller granular neurons with round nuclei and sparse cytoplasm; they were difficult to differentiate with H&E from the tightly packed neuroepithelial cells with sparse cytoplasm that surrounded the olfactory recess. As maturation proceeded, the granular layer became stratified with concentric layers of neurons separated by thin layers of axons and neuropil. This pattern extended caudally into the olfactory tract, with the ratio progressively changing in favor of more axons and fewer granular neurons, but the transition was not abrupt.

Immunocytochemical controls

Control sections of brain from each case, incubated without primary antibody, were negative.

Synaptophysin reactivity

Synaptophysin reactivity (Figure 3) was seen at 16 weeks gestation at the somatic membrane of mitral and tufted neurons, and of a few of the post peripheral granular cells. Reactivity also was seen within the somatic cytoplasm and axons of mitral neurons, an immature feature. Well‐defined synaptic glomeruli were uniformly intensely reactive at mid‐gestation and later; at earlier ages, the histologically defined glomeruli were small and variably reactive for synaptophysin. In the granular layer of the olfactory bulb, synaptophysin reactivity appeared first in the periphery from 16 weeks and last in the central core, but one‐third of deep granular cells lacked associated synaptophysin reactivity even at term. Concentric multiple synaptic layers were seen in the reactive parts of the granular zone at all gestational ages.

Figure 3.

Synaptophysin immunoreactivity to show the sequence of synaptogenesis in the olfactory bulb. A,B. At 16 weeks gestation, there is strong reactivity in the superficial layer within primary olfactory axons, consistent with their immaturity. Initiation of synapse formation also is seen in the mitral cell layer and in some synaptic glomeruli. Reactivity is seen within the somatic cytoplasm of immature mitral cells and a few concentric layers of synapses are seen in the most superficial part of the granular layer (mitr = mitral neurons). C. By 23 weeks gestation, synaptic glomeruli are now reactive but the superficial axons show less reactivity. Mitral cells exhibit synaptic vesicles at the surface membrane of mitral neurons but are no longer seen in their cytoplasm. D. At 40 weeks gestation, the synaptic glomeruli and mitral cell layer exhibit strong reactivity, as do the concentric superficial synaptic layers of granular cells, but the deeper layers in the core show distinctly less intense reactivity because of fewer synapses, indicating incomplete maturation at term.

NeuN

NeuN (Figure 4) exhibited no reactivity of mitral neurons at any age (normal) but the nuclei of the granule cells became intensely reactive as these neurons matured; neuroepithelial cells, by contrast, were non‐reactive. At 20 weeks gestation, periglomerular cells and the peripheral half of the granule cell layer exhibited uniformly reactive granule cell nuclei, whereas the neurons in the inner layer forming the core of the olfactory bulb showed no reactivity. By term, about 30%–40% of granular neurons remained non‐reactive, mainly deep in the core.

Figure 4.

Neuronal nuclear antigen (NeuN). A,B. 23‐week fetus. Scattered neuronal nuclei are marked in the mitral cell layer and in the superficial part of the granular layers, but granular neurons in the deep core of the bulb are non‐reactive. The maturing granular neurons expressing NeuN form concentric layers, whereas the deeper unmarked neurons exhibit no such lamination. C,D. 41‐week term neonate. The proportion of marked granule cells is increased over earlier gestational ages, but is still incomplete. Scattered deep granule cells have reactive nuclei, but others do not. NeuN immunoreactivity is never expressed at any gestational age or at maturity in mitral and tufted neurons, but is expressed in all other neurons of the olfactory bulb. It is a late marker of neuronal differentiation, expressed only in final stages of transition from neuroblast to neuron. By NeuN criteria, the olfactory bulb remains immature at term.

Calretinin (calbindin‐2) (Figure 5)

Figure 5.

A. At 16 weeks gestation, most synaptic glomeruli (g) exhibit no calretinin (calbindin‐2) reactivity (compare with synaptophysin in Figure 3A). Intense reactivity is seen in axons of the primary olfactory nerves, forming the superficial layer of the bulb. Mitral cells do not react at any gestational age and overall reactivity in this layer is sparse. B. At 19 weeks gestation, immature granule cells within the core of the bulb and surrounding the olfactory ventricular recess, are intensely reactive. The superficial layer of primary olfactory axons also is reactive, but most synaptic glomeruli (g) still do not exhibit reactivity, though a few are starting to show weak reactivity. C. At 22 weeks, intensely reactive small periglomerular interneurons (p) surround still non‐reactive synaptic glomeruli (g). D. Horizontal section of ventral surface of the olfactory bulb of an 8‐day‐old infant born at 40 weeks gestation, showing a field of many synaptic glomeruli but variable intensity of immunoreactivity, reflecting immaturity. Primary olfactory axons from neurons in the nasal epithelium are strongly reactive.

Neuroepithelial cells surrounding the olfactory ventricular recess were strongly immunoreactive for calretinin even in the youngest fetuses examined and they remained reactive even after these cells differentiated as granular neurons. From 16 weeks, scattered calretinin‐reactive periglomerular neurons also were seen, very evident at 19 weeks in those glomeruli not yet expressing calretinin. Some glomeruli were reactive for calretinin even at 16 weeks and others that showed synaptophysin reactivity were still calretinin‐negative. After 22 weeks, all glomeruli exhibited calretinin reactivity, but even at term, some synaptic glomeruli were qualitatively more intensely reactive than others, random in distribution. Small interneurons in the mitral cell layer were also calretinin‐reactive from 16 weeks, although mitral and tufted neurons themselves remained non‐reactive throughout gestation. The most superficial layer of the olfactory bulb, composed of axons of primary olfactory epithelial neurons, was reactive early and already strong at 16 weeks, including those in the subarachnoid space before entering the olfactory bulb (Figure 6E). In the olfactory tract, and particularly in the transitional zone between bulb and tract, long neurites interspersed with bipolar neurons were both reactive with calretinin (Figure 6D).

Figure 6.

Vimentin immunoreactivity is a good marker of immature neurons, of progenitor stem cells and of both immature and mature capillary endothelial cells. This intermediate filament protein is well demonstrated in the deep core of the olfactory bulb, particularly around the olfactory ventricular recess, where progenitor cells are uniformly mixed with immature granule cells. Vimentin‐reactive cells and processes are fewer in the more superficial layers of the olfactory bulb. Endothelial cells and the leptomeninges also are normally strongly reactive. This distribution is well demonstrated at (A) 16 weeks, (B) 19 weeks (C) 21 weeks and (D) 38 weeks gestation. Vimentin‐reactive cells and long processes also are demonstrated in the olfactory tract (Figure 8E).

Microtubule‐associated protein‐2 (MAP2)

Although non‐specific for all neurons, but not glial cells, MAP2 is useful in demonstrating mitral cells of the olfactory bulb, particularly since these neurons are never reactive for NeuN and also are non‐reactive for soluble calcium‐binding antigens such as calretinin. Mitral and tufted cells can also be demonstrated by SMI32 (neurofilament protein), neuron‐specific enolase and other neuronal markers not used in the present study.

Vimentin (Figure 6)

Strongly reactive neuroepithelial cells in the zone surrounding the olfactory ventricular recess were demonstrated at all gestational ages. In fetuses of 16 weeks gestation or more, these cells were bipolar with fine, long radial process extending both centripetally and centrifugally, superficially resembling radial glial fibers of the cerebral mantle but less well oriented and more meandering. Similar bipolar cells with shorter fine processes were seen in the periphery of the olfactory bulb, in the mitral cell layer and between synaptic glomeruli. Vimentin reactivity was also present in parenchymal and meningeal endothelial cells of capillaries and arterioles and in meningeal cells and processes at all ages. In the olfactory tracts, bundles of longitudinal slender processes were demonstrated parallel to the bundles of axons, but they were not impregnated by Bielschowsky silver technique and were not associated with synaptophysin, calretinin or GFAP reactivities. They represent processes of progenitor cells and some could be traced to the olfactory bulb in longitudinal sections.

Nestin (Figure 7)

Figure 7.

Nestin immunoreactivity is similar to vimentin in progenitor and endothelial cells. Ependymal cells of the olfactory recess are strongly reactive (normal for ages) (A,B) 23 weeks; (C,D) 32 weeks.

Reactive bipolar cells with fine processes were similar to those demonstrated by vimentin and with the same distribution, representing the same population. The ratio of these cells to differentiated neurons did not change appreciably during maturation by qualitative estimate, although quantitation was not attempted. Endothelial cells were also reactive. The concentration of nestin‐ and vimentin‐reactive bipolar cells in the granular layer core of the olfactory bulb was uniformly mixed between granular neurons. The vimentin‐reactive long bundled processes of the olfactory tract also were reactive with nestin, further confirming their identity as progenitor cell processes.

GFAP

The zone surrounding the olfactory ventricular recess was not gliotic in these controls.

Ki67 proliferating cell nuclear antigen (MIB1)

In the histologically undifferentiated neuroepithelium in the core of the olfactory bulb and especially in the zone surrounding the olfactory ventricular recess, the proliferative index was about 20% (Figure 8). In more differentiated regions and especially in the lamina of mitral cells and synaptic glomeruli, proliferative cells were rare; scattered marked nuclei usually belonged to endothelial cells or glial cells.

Figure 8.

MIB1 in the olfactory bulb of a 21‐week fetus (same case as Figure 2D) demonstrates (A) many scattered marked nuclei of cells capable of proliferation in the deep neuroepithelial core around the olfactory recess, but (B) only rare marked cells in the superficial laminae of the olfactory bulb; many of these are endothelial or glial cells, including an arteriole in the leptomeninges.

Luxol fast blue

Luxol fast blue myelin stain showed no myelination of any axons in the olfactory bulb or tract at any gestational age or in the term newborn. At 15 months postnatally, scattered myelinated axons were seen deep within the olfactory bulb and also in longitudinal bundles within the olfactory tract.

Bielschowsky silver impregnation

Axons of the olfactory tract, including their origin in the olfactory bulb, were impregnated at all ages, despite lack of myelination. Most intrinsic axons within the olfactory bulb are not impregnated in fetal life, however.

Olfactory tract

As previously noted, the olfactory tract is a combination of gray and white matter, not just an axonal fasciculus. Figure 9 illustrates the olfactory tract of a 15‐month‐old postnatal infant with various immunoreactivities, myelin stain and silver impregnation. It also shows the extension of the core granular cells of the olfactory bulb from one end and the anterior olfactory nucleus from the other as the gray matter components of the olfactory tract 20, and the presence of progenitor cells and their long processes. Olfactory tracts at fetal ages were similar except for the lack of myelination.

Figure 9.

Olfactory tract of a 15‐month‐old infant girl with a normal brain who died of dilated cardiomyopathy and intractable congestive heart failure. A. Bielschowsky silver impregnation exhibits the axonal fascicle from mitral neurons in a bundle. Parallel and ventral to this bundle is a pale strip without impregnated fibers. Many scattered granular neurons and other cells are seen within the olfactory tract. B. Luxol fast blue (LFB) shows myelination of most axons that were impregnated in (A). C. Synaptophysin shows many areas of synaptic vesicles within the olfactory tract. The cluster of reactivity at the top edge is a nodule of the anterior olfactory nucleus within the olfactory tract. D. Calretinin‐reactive cells are scattered amongst the granular neurons and axons also are reactive. The nodule of the anterior olfactory nucleus showing intense synaptophysin reactivity (C) exhibits a minority of scattered neurons reactive for calretinin and some reactive axons. E. Vimentin reveals that the longitudinal pale zone contains numerous long processes of progenitor cells and also cellular somata that are (F) non‐reactive for glial fibrillary acidic protein (GFAP, polyclonal antibody). Vimentin also demonstrates capillary endothelium.

Olfactory bulb dysgeneses

The following cases are a few representative dysplasias involving the olfactory bulbs. Explanatory details of histological and immunocytochemical findings are provided in the legends to these figures and not duplicated here.

Fetal hydrocephalus

Figure 10 illustrates one of the pairs of monozygotic female twins of 20 weeks gestation. The olfactory bulb exhibits dilatation of the olfactory recess, proportionate to the ventriculomegaly of the lateral ventricles. The recesses are ependymal‐lined with vimentin and GFAP reactivities of the ependymal, normal in mid‐gestation. Gross sections of the cerebral hemispheres are shown to demonstrate the degree of ventriculomegaly for comparison with the olfactory recess. The brain, including the olfactory bulb, of twin B was nearly identical (not illustrated).

Figure 10.

A. Gross coronal sections of brain of monozygotic twin A female fetus of 20 weeks gestation, showing moderate hydrocephalus associated with a mild aqueductal stenosis. B. The olfactory recess of the lateral ventricle also is proportionately dilated and ependymal‐lined. C. Synaptophysin reactivity in the olfactory bulb exhibits a normal distribution as age‐matched controls (Figure 4). D. Vimentin reactivity in the olfactory bulb shows the expected number and distribution of progenitor cells. The hydrocephalus and olfactory bulb of twin B was nearly identical (not illustrated).

Holoprosencephaly

We previously reported synaptic precociousness in the cerebral cortex and median cyclopean eye of fetuses with alobar holoprosencephaly 101, 107. In none of these cases could we identify olfactory bulbs or tracts either macroscopically or microscopically in the ventral prosencephalic midline, consistent with the well‐known arrhinencephaly in this malformation.

Fusion and dysgenesis of olfactory bulbs

Figure 11 is the left olfactory bulb in a 20‐week gestation fetal brain in which the right bulb was grossly absent and the left was smaller than expected. Microscopically, this left bulb, which appeared macroscopically single, exhibits a bilobar or two fused, dysplastic olfactory bulbs with additional features of maturational delay. The right cerebral hemisphere was smaller than the left and the brain weight was 32.7 g (control 50 g). Right anophthalmia and hypoplasias of the right ear were present. Karyotype was 46XX.

Figure 11.

Olfactory bulb dysgenesis in a 20 weeks F fetus. Gross: absent R olfactory bulb; hypoplastic L bulb; R anophthalmia, absent R optic nerve; malformed R ear and side of mouth; R cerebral hemisphere smaller than L; brain weight 32.7g (control 50 g); fused gyri R > L occipital lobes; cardiac VSD. A. Bilobar or fused, poorly formed and hypoplastic olfactory bulbs with recognizable histological architecture. H&E. B,C. Synaptophysin discloses a paucity of synaptic glomeruli (g) and less activity than expected for gestational age in the mitral cell layer; mitral neurons are not well aligned, many being displaced and disoriented. In the area of fusion of the two olfactory bulbs, continuous immature granular cells form a single sheet without concentric layering. D,E. Vimentin shows the presence of progenitor cells without loss, but they are not normally distributed. The olfactory ventricular (ventr.) Recess is lined by a non‐uniform distribution of ependymal cells, some regions excessively pseudo‐stratified and others with no ependymal cells; vimentin reactivity of the ependyma is normal at this age, however. F. Neuronal nuclear antigen (NeuN) is expressed in fewer neurons than in controls of comparable gestational age (see Figure 4), indicating delayed neuronal maturation. Mitral cells are normally non‐reactive. G. Calretinin shows intense reactivity in the superficial layer of primary olfactory axons and abnormal heterotopia of clustered reactive neurons deep within the molecular zone. Most of the few synaptic glomeruli are non‐reactive.

Dysgenesis of olfactory bulb in HME

Figure 12 illustrates the olfactory bulb of the dysplastic hemisphere in a 2.5‐month‐old boy, born at 35 weeks gestation, with severe HME who died of post‐operative complications following partial hemispherectomy for intractable, pharmacologically refractory epilepsy. Special details of the neuropathological examination of other structures of this brain are described as Case 1 in reference 104. The dysplastic olfactory bulb had an abnormal longitudinal sulcus on its ventral surface.

Figure 12.

Olfactory bulb dysgenesis in a 2.5‐month‐old male infant born at 35 weeks gestation with severe hemimegalencephaly (HME), who died in the post‐operative period following partial hemispherectomy for refractory status epilepticus. The olfactory bulb in the HME hemisphere was enlarged and had an abnormal longitudinal sulcus on the ventral surface. A,B. Synaptophysin demonstrates disorganized general architecture and synaptic glomeruli clustered in a row in the centre of the bulb. These synaptic glomeruli aligned in the centre of the section were adjacent to a deep longitudinal ventral sulcus, hence may not be entirely heterotopic. C,D. Calretinin shows poorly aligned synaptic glomeruli at the tip of the bulb but few were seen on the lateral surfaces. The central area with synaptic glomeruli is strongly reactive, including abnormal clustering of many primary olfactory nerve axons. E. Vimentin‐reactive cells and fibers are neither excessive nor sparse, though poorly organized. Nestin was similar (not shown). F. NeuN shows many fewer reactive granule cells than in controls in this postnatal period. Mitral cells normally are non‐reactive for NeuN.

Olfactory epithelium

Olfactory epithelium is a part of the nasal mucosa that lines the roof of the nasal cavities, the medial part of the nasal septum, the superior and sometimes extending to the middle turbinates 17. Epithelial plugs obstruct the external nares in the first trimester of fetal life but regress between 16 and 24 weeks gestation, as demonstrated as early as 1910 110. Volatile molecules must penetrate an aqueous mucus layer covering the olfactory epithelium to reach the receptor sites on the cilia of primary olfactory dendrites 17. The number of bipolar primary olfactory neurons increases greatly with the formation of the turbinate bones beginning at about 8 weeks gestation, which enables a greatly enlarged surface area of olfactory epithelium in the nasal cavity. An active process of apoptosis occurs simultaneous with the generation of new neurons in the olfactory epithelium 62. Specific olfactory marker proteins are identified in the human olfactory epithelium from 28 weeks gestation 15, 42. In rodents, the zonal organization of the olfactory epithelium is preserved in its somatotopic distribution in the olfactory bulb, as determined by zone‐specific markers 70. Many progenitor cells are present in the olfactory epithelium throughout life, similar to the olfactory bulb 13. Novel DNA microarrays and other genetic studies also recognize human olfactory receptor gene families and classes of neural precursors in the olfactory epithelium 117, 123. The olfactory epithelium with its primary olfactory receptor neurons also does not approach mature scanning ultrastructural morphology in humans until near‐term 44 and transmission electron microscopy also is well described 82.

Olfactory ventricular recess

The outgrowth of the incipient olfactory bulb includes a rostral extension of the primordial frontal horn of the lateral ventricle, the olfactory ventricular recess, a transitory structure that begins to involute in the third trimester but is still present at term. It is not always in the exact middle of the core of the olfactory bulb, but more often is eccentric, although it remains within the granular layer and never protrudes superficial to the mitral cell layer. The olfactory recess initially is lined by primitive neuroepithelium but acquires an ependymal lining beginning at mid‐gestation, arresting mitotic activity at the luminal surface of the neuroepithelium as in the periventricular region of other parts of the lateral ventricles (Figure 2). The olfactory recess ependyma is a pseudostratified columnar epithelium and exhibits reactivity to vimentin and GFAP even in the term neonate, unlike the mature permanent ventricular ependymal of the lateral, third and fourth ventricles in which these reactivities disappear with maturation in the third trimester 95, 96. The neuroepithelium surrounding the olfactory ventricular recess shows proliferative activity (Figure 2). Massive apoptotic cellular death occurs in the murine olfactory bulb, especially among granular neurons, if thiamine deficiency is a condition 34.

Humphrey reported that the olfactory recess begins to involute by growth of its ependymal cells into the lumen starting at 14 weeks and complete by 18 weeks 39, but in our study, we found a well‐formed recess with an ependymal‐lined lumen even in the term neonate, although it is no longer evident within a few weeks postnatally. Remnants of this recess were demonstrated by magnetic resonance imaging (MRI) in 72 of 122 (59%) normal adult subjects 114. These adult remnants are confirmed histologically either as glial‐lined microcysts 114 or as small clusters of residual ependymal cells, as we have observed.

Functional architecture

Unlike the neocortex, the olfactory bulb lacks a molecular zone with Cajal–Retzius neurons and the subplate zone 63. The transitory subpial granular glial layer of Brun of the fetal neocortex also is absent from the olfactory bulbs. Synaptic glomeruli lie superficial to the layer of large mitral cells, unlike the synaptic glomeruli of the internal granular layer of cerebellar cortex. The olfactory groove of the gyrus rectus is among the earliest sulci to develop in the human fetal forebrain, appearing at 16 weeks gestation 14. Its absence at autopsy confirms arrhinencephaly rather than artifactual loss of the olfactory bulb and tract at the time of brain removal.

Axons of the tufted and mitral neurons form the olfactory tract and pass caudally to terminate mainly in the anterior olfactory nucleus, but also directly to other telencephalic and diencephalic regions, particularly the amygdala, entorhinal cortex and hypothalamus. Post‐synaptic axons of anterior olfactory neurons enter the anterior commissure. Some pre‐synaptic olfactory tract axons also cross the midline in the anterior commissure.

Discussion

Our present immunocytochemical study of neuronal maturation in the olfactory bulb clearly shows persistent immaturity in term neonates, by criteria of expression of individual neuronal proteins, synaptogenesis, myelination and persistence of the transitory fetal olfactory ventricular recess. It thus updates the 75‐year‐old premise based on histological architecture that the olfactory bulb matures by 11 weeks gestation 39. However, even histological examination of sections of term neonatal olfactory bulb with H&E stain provides subtle clues that this structure is not yet fully mature: the concentric lamination of the core granular layer is seen only in its outer part; the olfactory ventricular recess is still present in most cases. Clinical olfactory function in the fetus and neonate also does not denote maturity, as with the visual and other special sensory systems. Normal developmental neuroanatomy of the cells of the olfactory bulb must be understood to enable the interpretation of dysgenesis.

Mitral and tufted neurons are similar cells that provide the principal efferent output of the olfactory bulb. Long axons from the single layer or sheet of large mitral neurons descend into the olfactory tract to synapse in various structures, including the anterior olfactory nucleus, amygdala, septum, hypothalamus and entorhinal cortex (part of the parahippocampal gyrus); a few also enter the anterior commissure for reciprocal connections between the olfactory bulbs. There is heterogeneity in the neurochemistry and electrophysiological functions among several classes of mitral neurons 3, 61. During development, mitral neurons are identified as early as 8 weeks gestation, at which time mitral neuronal differentiation and cytoplasmic growth occur. Tufted neurons of layer 3 are similar to mitral neurons, although smaller and occur only in the most rostral regions of the olfactory bulb. They also are glutaminergic and project axons caudally into the olfactory tract. The mitral cell layer is initiated at 9.5 weeks, is well developed by 11 weeks and is more prominent at 18.5 weeks than in the adult 39.

Axonal terminals of primary olfactory neurons entering the olfactory bulb form synaptic glomeruli (glomerulus = Greek; ball of threads) with dendrites of the layered mitral and tufted cells. The glomeruli generally are arranged as a single layer but in places may be two or three glomeruli thick; in superficial horizontal sections of the ventral surface of the olfactory bulb the glomeruli can be seen as a sheet of multiple glomeruli. Synaptic glomeruli exhibit less synaptophysin expression at 16 weeks than at older gestational ages; all are reactive by term. The β‐secretase enzyme BACE‐1, deficient in Alzheimer's disease, is essential for axonal guidance of olfactory sensory neurons and for the formation of synaptic glomeruli during development 83.

Small periglomerular interneurons also occur, often form sheet‐like processes that envelop other small periglomerular neurons 87. Spine‐laden dendrites of periglomerular cells ramify within two or occasionally more glomeruli. Their axons extend across as many as six glomeruli as they contact local interneurons. These periglomerular neurons are heterogeneous in morphology, neurochemistry and physiology; about 10% lack synapses with olfactory neurons 51. They synthesize gamma‐aminobutyric acid (GABA) and dopamine as transmitters and these molecules are co‐localized 27, 47, 73, 74; periglomerular neurons also contain the calcium‐binding proteins calbindin and parvalbumin, as well as calretinin as we here demonstrate. Axons of primary olfactory neurons in the mucosa and dendrites of mitral cells form synapses within the glomeruli. Each glomerulus receives as many as 25 000 axons of olfactory nerve neurons (rabbit) 51. A convergence of about 1000 receptor cell axons upon every second‐order neuron (rabbit) results in major summation or amplification of olfactory stimuli. About 80% of synaptic contacts between neurons are organized as reciprocal pairs: mitral‐to‐granule cell synapses are excitatory, in contrast with granule cell‐to‐mitral neuronal synapses which are inhibitory 51. The external plexiform layer (layer 3) is cell‐sparse and formed largely by primary dendrites of granule cells and secondary dendrites of mitral and tufted neurons. Tufted neurons in this layer are similar to mitral neurons of layer 4 in morphology, although smaller, and also are glutaminergic with similar synaptic connections. The olfactory bulb thus exhibits a mathematical precision and predictability in its intrinsic synaptic relations, reminiscent of the cerebellar cortex. More than 20 known neurotransmitters or modulators have been identified in the olfactory bulb 35. It is not surprising therefore that a large array of genes is expressed during fetal life.

Pre‐ and postnatal neuronal maturation of the olfactory bulb

Immunocytochemical antibodies against specific neuronal proteins not only denote a neuronal lineage of cells and, in some cases are even specific for types of neurons, but also identify maturation of neuroblasts to neurons by their timing of expression 97, 98, 118, 119. Some, such as calretinin, are expressed very early, even in early differentiating neuroepithelial cells, whereas others, such as NeuN, are expressed only late, at the time of terminal maturation of the neurons 97.

Synaptophysin

The early development of synapse formation in the olfactory bulb that we demonstrate here in the human fetus contrasts with the ontogenetically later synaptogenesis of phylogenetically newer structures such as the neocortex, corpus striatum and cerebellar hemispheres in human fetal brain. This determination uses immunoreactivity to synaptophysin, a glycoprotein structural element of the synaptic vesicle membrane, regardless of the neurotransmitter that the vesicle encloses 99, 100, 102, 103, 105, 120. Synaptophysin is a marker of the sequence of synaptogenesis in the fetal brain and is robust molecule that resists post‐mortem autolysis for 5 days or more 105.

Synaptic circuitry in the murine olfactory bulb remains plastic not only in the immature neonatal condition but persists as long as the body is still growing and sexual maturity is achieved 80. Postnatally, the length of mitral cell dendritic branches increases by a factor of 11; the number of glomerular and extra‐glomerular synapses increases by factors of 90 and 170, respectively 80. In addition, olfactory receptors in the mucosa have many de novo genetic mutations. Some neurons may remain dormant for long periods before developing synaptic relations, even though presynaptic and post‐synaptic membranes are in close proximity. An example is the relation between primary olfactory receptor neurons and mitral cells in the synaptic glomeruli of the olfactory bulb 36. Loss of olfaction in the elderly, by contrast, is at least partly due to degenerative disruption of synaptic circuitry in the olfactory bulb 88.

NeuN

Only about 60% of granular neurons exhibit NeuN reactivity. At mid‐gestation, about 25% of olfactory bulbar granular cells exhibit NeuN reactivity at term, the last maturing neurons being deep in the central core. Reactive granular cells are confined to the outer layers, whereas at term, the reactive neurons are also distributed within the inner core. Neural circuitry of the murine olfactory bulb, in terms of dendritic branching of mitral cells and extra‐glomerular synaptogenesis, extends well into the postnatal period and to the time of sexual maturity (10–12 postnatal weeks in the mouse) 80.

NeuN is a late expressed nuclear protein and never expressed at any age, including the adult in certain populations of unrelated neurons: mitral and tufted neurons of the olfactory bulb, Purkinje cells of the cerebellar cortex, inferior olivary and deep cerebellar nuclear neurons, ganglion cells of the peripheral sympathetic chain, photoreceptor cells of the retina. In this study, the lack of expression of NeuN in mitral and tufted neurons at all gestational ages was reconfirmed. Non‐reactivity in these neurons probably is due to reciprocal activity of neuron‐restrictive silencing factor (NRSF; REST), which inhibits NeuN expression in this mixed population of specific neurons 24, 111. NeuN is an epitope of RBFOX3, a novel member of the Rbfox1 family of splicing factors 22, and controls the biogenesis of a subset of microRNAs 43. Granular neurons deep within the bulb became progressively reactive for NeuN, initially in the periphery and latest in the central core, but not all granular neurons were reactive at term, proving evidence that the olfactory bulb is not yet fully mature at birth.

Calretinin (calbindin‐2)

Calretinin is one of several water‐soluble calcium‐binding proteins specific for GABAergic inhibitory interneurons, and such neurons are found throughout the brain 97, 118, 119. Other proteins in this group include parvalbumin and calbindin D28k, each of which has its own specificity. Calretinin identified about 12% of inhibitory interneurons of the cerebral cortex, the other 8% of the total of 20% being reactive for parvalbumin 97. Calretinin is expressed very early in neuronal differentiation and many cells of the ganglionic eminence of the periventricular germinal matrix are already reactive before histological differentiation as neurons 97, 118. In the fetal olfactory bulb, the neuroepithelial cells in the central core and around the olfactory ventricular recess were already strongly reactive as well, before their histological differentiation as granular neurons and long before their reactivity with the late marker NeuN. Scattered non‐granular neurons that were reactive in the mitral cell layer probably correspond to inhibitory interneurons in the cerebral cortex. We demonstrate here that periglomerular GABAergic interneurons are strongly reactive with calretinin, as with previously reported parvalbumin and calbindin D28k reactivities, an anticipated but not previously documented finding. Primary olfactory axons forming layer 1 at the periphery of the olfactory bulb were strongly reactive for calretinin at all ages, but the synaptic glomeruli become reactive only after mid‐gestation and even in the term neonate, the intensity of reactivity in adjacent glomeruli was quite variable. At mid‐gestation, the difference is striking between the intense reactivity of synaptophysin and the weak to absent reactivity of calretinin within synaptic glomeruli. Calretinin reactivity within synaptic glomeruli is in the axons of primary olfactory neurons since mitral cells and their dendrites are non‐reactive. Absence of calretinin in synaptic glomeruli during development implies that few primary axons have penetrated the glomerular parenchyma to establish synapses.

The intense immunoreactivity of both calretinin and synaptophysin in the superficial layer of the olfactory bulb of young fetuses, which consists of axons of primary neurons in the olfactory neuroepithelium without synapses in that layer, followed by diminished intensity in older fetuses, is consistent with the demonstration in other parts of the CNS of maturing white matter. Axonal transport of synaptophysin molecules in immature axons that have not completed their synaptic contacts or assembled these molecules into synaptic vesicles in the axonal terminal is well documented; loss of that axoplasmic reactivity with axonal maturation is exemplified in the cortico‐ponto‐cerebellar pathway of the fetal basis pontis 103.

Myelination

There is little myelination in the forebrain white matter of the term neonate, except for axons of the internal capsule 122. The olfactory tract seems to follow that same pattern of postnatal myelination as with other telencephalic regions, despite being one of the earliest structures to develop both ontogenetically and phylogenetically. The olfactory tract is unmyelinated in fetal life and at term and has a long postnatal myelination cycle: 10% myelinated at 2 months, 50% at 20 months and 90% at 2 years 31. Olfactory contributions form the peripheral rim of fibers of the anterior commissure that myelinate postnatally; at 3–4 months of age, they sharply contrast with the unmyelinated axons from the anterior temporal neocortex that form the core of the anterior commissure, as visualized with LFB myelin stain 122.

The olfactory thalamic equivalent

The olfactory system is the only sensory system of the CNS to lack projections to the thalamus. An intrinsic thalamic equivalent is present within the olfactory bulb and tract, consisting of three components: (i) The inner core of the olfactory bulb consists of concentric alternating layering of granular neurons and synaptic sheets. Granular neurons lack axons and form dendrodendritic synapses. The granular layer extends into the olfactory tract. A ratio of about 50–100 granule cells for each mitral cell is preserved. These small granular interneurons are GABAergic and have no extrinsic connections; they form dendrodendritic synapses with mitral and tufted neurons as well as between themselves within the core granular zone. Lamination in the granular layer, and also the mitral cell layer, is at least partly regulated in fetal mice by the zinc‐finger gene Fex in olfactory neurons 37. (ii) Periglomerular interneurons are another intrinsic cell with no afferent or efferent connections outside the olfactory bulb. They co‐express GABA and dopamine and form dendrodendritic synapses surrounding the synaptic glomeruli 27, 47, 73, 74. (iii) The anterior olfactory nucleus is actually a series of neuronal aggregates, not a single compact cluster, and extends into the olfactory tract with other aggregates just beyond the tract 19. This is the only component with axons projecting to the amygdala, entorhinal cortex and other structures, and is the olfactory equivalent of thalamocortical projections.

Clinical correlates

Preterm neonates reliably respond to olfactory stimuli after 28 weeks gestation 23, 64. Olfactory reflexes are described as a reliable test in the neurological examination of the term neonate 93. They can even be semi‐quantitated in gradients with olfactory sensitivity increasing in the first postnatal days 57. A continuous turnover of amniotic fluid takes place through the fetal nasal passages and the fetus can detect odors in amniotic fluid of strong foods ingested by the mother, such as garlic and onion 23, 66, 67, 108, 109. Neonatal olfactory reflexes may be useful in providing clinical evidence of arrhinencephaly in some malformations (eg, holoprosencephaly) and in certain genetic disorders (eg, Kallmann syndrome) 93.

Olfactory bulb dysgenesis

Neuropathological reports of dysgenesis of the olfactory bulb are sparse, even in autopsies of patients with major cerebral malformations. Post‐mortem descriptions of the olfactory bulb are usually limited to its gross presence or absence, as in holoprosencephaly and Kallmann syndrome.

A case of asymmetrical bi‐hemispheric megalencephaly associated with epidermal nevus syndrome in a 1‐year‐old boy was reported in which the olfactory bulb on the larger side of the brain was described as “enlarged,” but histological examination was not provided 1. This case was almost certainly HME, the most frequent brain malformation associated with epidermal nevus syndrome 26. Robain et al also described a neonate with Jadassohn linear sebaceous nevus of the facial midline and of the right hemithorax, whose brain at autopsy exhibited right HME with particular enlargement of the olfactory bulb on that side but they did not describe the microscopic anatomy of the olfactory bulb 89. This hypertrophic olfactory bulb had an abnormal longitudinal sulcus on its ventral surface, similar to our case. The normal olfactory bulb lacks fissures and sulci. This indentation should not be called a groove because of confusion with the olfactory groove on the gyrus rectus. Enlargement of the olfactory bulb in HME also may be demonstrated during life by MRI 11.

Fused olfactory bulbs with non‐HME asymmetry of the cerebral hemispheres (Figure 11) highlight abnormal development of the olfactory bulb, as can occur in other parts of the brain. Supernumerary olfactory bulbs are described rarely by imaging 55. In tuberous sclerosis complex (TSC), the human olfactory bulb may exhibit hamartomata 54 and, in a murine model of TSC, micronodules containing hypertrophic dendritic trees are observed in the olfactory bulb 25.

Among developmental disorders of the olfactory bulb, agenesis, also known as arrhinencephaly, is the best documented. The incidence is estimated at 700 per 100 000 births 41. Most cases are associated with other malformations, the most frequent being holoprosencephaly in which nearly all cases exhibit this associated feature. In the mildest lobar forms, olfactory bulbs may still be present, although often hypoplastic. Other cerebral malformations that are less constantly associated with arrhinencephaly include septo‐optic‐pituitary dysplasia and agenesis of the corpus callosum 90, 94. Olfactory bulb agenesis also occurs as an isolated defect without other brain anomalies, not as part of Kallmann syndrome. Olfactory bulb agenesis and absent olfactory grooves or sulci can be demonstrated by MRI not only postnatally 10 but also prenatally in fetuses of 30–34 weeks gestation 6.

CHARGE syndrome (coloboma, heart defects, choanal atresia, growth retardation, genital hypoplasia due to hypogonadotropic hypogonadism and ear anomalies with hearing impairment) also includes olfactory bulb agenesis or hypoplasia with anosmia 7, 8, 79. The genetic mutations in the majority of patients with CHARGE syndrome are in the CHD7 (chromodomain helicase DNA‐binding) gene 9, but many patients also have overlapping 22q11.2 deletions of DiGeorge syndrome 16. Deficient gonadotropin‐releasing hormone (GnRH) is demonstrated in Kallmann syndrome and in olfactory bulb agenesis. These numerous neuroendocrine cells normally found in the fetal hypothalamus, particularly the preoptic nuclei, are sparse or absent from the hypothalamus in the arrhinencephaly of CHARGE syndrome, as well as in holoprosencephaly and other cerebral dysplasias with agenesis of the olfactory bulbs 115. A murine genetic model of CHARGE syndrome also exhibits olfactory bulb agenesis 52. Mutations in the CHD7 gene influence neural crest cell development and migration 112. Choanal atresia is a disorder of prosencephalic neural crest.

Resident progenitor stem cells of the olfactory bulb and regeneration

The olfactory bulb has an importance beyond its function in detecting odors. It is one of only two permanent repositories in the mature brain in which resident multi‐potential neural progenitor cells are generated; the other reservoir is the polymorphic layer in the hilus of the hippocampal dentate gyrus. Migratory cells from the periventricular zone of the lateral ventricles of the mouse continue to stream into the olfactory bulb even in adult life to form more neuronal precursor cells, mediated by the chemoattractant activity of the Sonic hedgehog gene 4. Not only olfactory bulb, but also the olfactory mucosal epithelium, is a site of resident stem cells in both the mature rodent and human 13, 21, 32. Axons of primary olfactory neurons in the olfactory mucosa are surrounded by specialized ensheathing cells that differ from ordinary Schwann cells and render transplantation to injured regions of human spinal cord more feasible 60.

A potential for autologous surgical transplant to other regions of the brain in which damage has occurred (eg, infarcts, trauma) or where neurodegeneration has occurred (eg, Parkinson's disease) offers the expectation of neuronal regeneration and repair. This treatment incurs certain risks, however. The autologous transplantation of olfactory mucosa with its progenitor cells to an injured region of spinal cord resulted in the development of a spinal cord tumor at that site 54.

Progenitor cells responsible for the replacement of some neuronal populations within the olfactory bulb reside in the rostral subventricular zone, a remnant of the ganglionic eminence that gives origin to tangentially migrating GABAergic interneurons to the cerebral cortex. A subpopulation of the many generated neuroblasts undergoes restricted chain migration to the olfactory bulb along a subependymal pathway known as the rostral migratory stream 45, 59, 92, 124. This migration terminates in the core of the olfactory bulb without guidance from radial glia or astrocytes 35. The migration is facilitated by a neuron‐specific class III b‐tubulin (TUJ1) and a highly polysialated neural cell adhesion molecule (NCAM) 92. The progenitors of all types of interneurons of the olfactory bulb originate from the telencephalic subventricular zone (medial ganglionic eminence), from whence they migrate rostrally to form the core granular zone of the primordial olfactory bulb 58, 119. Upon reaching the olfactory bulb core, the cells then migrate centrifugally, but specialized radial glial cells and processes similar to those in the main telencephalon are absent. They repopulate the granule cell layer, form the initial population of periglomerular interneurons and continue to contribute these cells even in the adult. Autoradiographic studies in the murine olfactory bulb demonstrated that the tufted cells migrate past the mitral cells to reach the rostral end of the bulb; glioblasts arise from the neuroepithelium around the olfactory ventricular recess 46. The rostral migratory stream is a fetal pathway not easily demonstrated in the mature brain and some olfactory bulb progenitor cells in the adult probably regenerate neurons locally. In aging, mitral neurons and synaptic glomeruli are gradually lost, contributing to hyposmia in elderly adults 65, 68.

Accessory olfactory bulb and vomeronasal system (nervus terminalis)

A small secondary olfactory system also exists in most vertebrates. The histological structure of the accessory bulb is similar to that of the primary olfactory bulb, but usually not as well developed 18, 39. It is a transitory early fetal structure that atrophies by mid‐gestation, as with the septum that becomes the septum pellucidum. Although the accessory olfactory bulb usually is absent in adult humans, a vestigial bulb sometimes is still identified 113. Its afferent nerve twigs constitute the nervus terminalis or cranial nerve 0, and its efferents terminate mainly in the septum and amygdala.

Conclusions

Despite assertions by early observers of histological organization that the olfactory bulb matures in the first trimester, our immunocytochemical studies indicate that maturation is not even complete in term neonates. Except for agenesis, dysplasias of the olfactory bulb are poorly documented, in part because the structure rarely is examined in fetuses and infants, even in the presence of other major brain malformations. The representative examples we present of olfactory dysplasia suggest a need for further systematic post‐mortem study of the olfactory bulb and tract in the fetus and infant.

Conflict of Interest

The authors have no conflicts of interest or financial disclosures to declare.