Morphomolecular neuronal phenotypes in the neocortex reflect phylogenetic relationships among certain mammalian orders
Version of Record online: 7 OCT 2005
Copyright © 2005 Wiley-Liss, Inc.
The Anatomical Record Part A: Discoveries in Molecular, Cellular, and Evolutionary Biology
Special Issue: Nature's Experiments in Brain Diversity
Volume 287A, Issue 1, pages 1153–1163, November 2005
How to Cite
Hof, P. R. and Sherwood, C. C. (2005), Morphomolecular neuronal phenotypes in the neocortex reflect phylogenetic relationships among certain mammalian orders. Anat. Rec., 287A: 1153–1163. doi: 10.1002/ar.a.20252
- Issue online: 25 OCT 2005
- Version of Record online: 7 OCT 2005
- Manuscript Accepted: 17 AUG 2005
- Manuscript Received: 16 AUG 2005
- McDonnell Foundation
- Wenner-Gren Foundation for Anthropological Research
- National Science Foundation. Grant Number: BCS-0515484
- brain evolution;
- cerebral cortex;
- pyramidal cells
The cytoarchitecture of the cerebral cortex in mammals has been traditionally investigated using Nissl, Golgi, or myelin stains and there are few comparative studies on the relationships between neuronal morphology and neurochemical specialization. Most available studies on neuronal subtypes identified by their molecular and morphologic characteristics have been performed in species commonly used in laboratory research such as the rat, mouse, cat, and macaque monkey, as well as in autopsic human brain specimens. A number of cellular markers, such as neurotransmitters, structural proteins, and calcium-buffering proteins, display a highly specific distribution in distinct classes of neocortical neurons in a large number of mammalian species. In this article, we present an overview of the morphologic characteristics and distribution of three calcium-binding proteins, parvalbumin, calbindin, and calretinin, and of a component of the neuronal cytoskeleton, nonphosphorylated neurofilament protein in the neocortex of various species, representative of the major subdivisions of mammals. The distribution of these neurochemical markers defined several species- and order-specific patterns that permit assessment of the degree to which neuronal morphomolecular specialization, as well as the regional and laminar distribution of distinct cell types in the neocortex, represents derived or ancestral features. In spite of the remarkable diversity in morphologic and cellular organization that occurred during mammalian neocortical evolution, such patterns identified several associations among taxa that closely match their phylogenetic relationships. © 2005 Wiley-Liss, Inc.
In general terms, neurons in the mammalian neocortex can be divided into two generic classes, pyramidal excitatory cells and inhibitory interneurons. Each class includes many subtypes that can be identified by their size, shape, dendritic and axonal morphology, and connectivity. About a dozen morphologically different subtypes of inhibitory interneurons have been described and there exist an unknown number of pyramidal neuron subpopulations. These neuronal subtypes are known to exhibit a differential distribution among cortical layers and regions, and some of them are also differentially represented among species (for review, see Hof et al., 1999). In addition, neurons can be further classified based on their content of various proteins that can serve as neurochemical markers of identifiable subpopulations. Such neurochemical patterns have also been used to define cytoarchitectural borders or transitions among cortical domains in several mammalian species. The most commonly used markers are dephosphorylated epitopes of the neurofilament protein triplet (NFP) and the calcium-binding proteins parvalbumin (PV), calbindin (CB), and calretinin (CR) (Hof et al., 1999, 2000).
In primates, NFP predominates in a subset of large pyramidal neurons that have an extensive dendritic arborization, a well-defined laminar distribution, and form highly specific long corticocortical projections in macaque monkeys (Campbell et al., 1989, 1991; Hof and Morrison, 1995; Hof et al., 1995b; Nimchinsky et al., 1996; Preuss et al., 1997, 1999; Sherwood et al., 2003a). In the macaque monkey visual cortex, the distribution and density of NFP-containing neurons define quantitatively the regional boundaries of more than 20 visual cortical areas (Hof and Morrison, 1995). Similarly, regionally specific distribution patterns of NFP-immunoreactive pyramidal neurons have been reported in the orbitofrontal, cingulate, retrosplenial, and inferior frontal cortex in macaques, great apes, and humans (Hof and Nimchinsky, 1992; Carmichael and Price, 1994; Hof et al., 1995a; Nimchinsky et al., 1995, 1997; Vogt et al., 2001, 2005; Sherwood et al., 2003a). NFP-immunoreactive pyramidal neurons have also been reported in the neocortex of mouse, rat, hamster, cat, and dog, where they exhibit clear regional and species differences in their distribution and densities (Hof et al., 1996; van der Gucht et al., 2001, 2004, 2005; Kirkcaldie et al., 2002; Boire et al., 2005), and in bottlenose dolphins, where they may identify a major subpopulation of output neurons throughout the neocortex (Hof et al., 1992).
Data from rat and primate neocortex show that PV-, CB-, and CR-immunoreactive neurons represent morphologically nonoverlapping subtypes of GABAergic interneurons (Hendry et al., 1989; Andressen et al., 1993; Condé et al., 1994; DeFelipe, 1997; Gonchar and Burkhalter, 1997; Glezer et al., 1998; Morrison et al., 1998; Hof et al., 1999). CB- and CR- expressing interneurons share many morphologic similarities and are mainly bitufted, bipolar, and double bouquet neurons, as well as a few pyramidal neurons, with minimal overlap among these subpopulations in the rodent and primate neocortex (Rogers, 1992; DeFelipe, 1997; Morrison et al., 1998). PV-immunoreactive neurons are mainly observed in layers II to V and are principally basket and chandelier cells (Blümcke et al., 1990; Van Brederode et al., 1990; Hof and Nimchinsky, 1992; Condé et al., 1994; Nimchinsky et al., 1997). PV has been reported to occur in certain pyramidal neurons in primates as well (Preuss and Kaas, 1996; Sherwood et al., 2004). CB immunoreactivity is found in subpopulations of pyramidal and nonpyramidal cells (DeFelipe et al., 1989; Hof and Morrison, 1991; DeFelipe and Jones, 1992; Hayes and Lewis, 1992; Hof and Nimchinsky, 1992; Condé et al., 1994). Most CB-immunoreactive interneurons are double bouquet cells located in layers II and III (DeFelipe et al., 1989). CB is also present in Martinotti cells in layers V and VI, in some neurogliaform neurons, and Cajal-Retzius cells in layer I (Derer and Derer, 1990). Finally, a population of weakly labeled CB-immunoreactive pyramidal neurons has been reported in layer III in monkeys and human, with clear rostrocaudal density gradients among neocortical areas (Hof and Morrison, 1991; Hayes and Lewis, 1992; Kondo et al., 1994). Most CR-immunoreactive neurons have bitufted, vertically oriented dendrites and a vertically oriented axon, defining a narrow radial domain (Jacobowitz and Winsky, 1991; Hof and Nimchinsky, 1992; Résibois and Rogers, 1992; Condé et al., 1994; Meskenaite, 1997). They are densest in layers II and III and represent the bipolar and double bouquet subtypes. CR-immunoreactive Cajal-Retzius cells are also seen in layer I, and isolated pyramidal neurons containing CR have been reported in several mammalian species (Nimchinsky et al., 1997; Hof et al., 1999). These three calcium-binding proteins are also observed in neurons in a variety of mammalian species, particularly in domesticated carnivores and in cetaceans (Demeulemeester et al., 1991; Glezer et al., 1992, 1993, 1998; Hof et al., 1996, 1999, 2000).
In this article, we describe several aspects of neurochemical specialization in large-brained primates, ungulates, and cetaceans. We compare these features to observations from about 40 species illustrating more than 10 mammalian orders (Hof et al., 1999). We report specific chemoarchitectural characteristics in the context of regional anatomy and cortical microcircuitry, and, based on descriptions of large brained species such as cetaceans, large artiodactyls, and hominoid primates, we discuss how these cellular phenotypes could be used to assess taxonomic affinities among species.
MATERIALS AND METHODS
Specimens were obtained from laboratory animals used in the context of unrelated studies sacrificed for scientific purposes, from animals suffering from a terminal illness and euthanized in zoological facilities for humane reasons, or from animals that died naturally. All euthanasia protocols were reviewed and approved by the relevant institutional animal care and use committees. We had access to over 50 representative species of 12 mammalian orders, including 2 prototherians, 5 marsupials, 1 xenarthran, 1 insectivore, 3 chiropterans, 1 scandentia, 20 primates, 4 rodents, 7 carnivores, 9 artiodactyls, 1 perissodactyl, 12 cetaceans, 1 sirenian, and 1 proboscid [see Hof et al. (1999) for a detailed summary; the elephant is discussed separately in this issue by Hakeem et al. (2005)]. The prototherians, marsupials, the hedgehog, chiropterans, the tree shrew, rodents, the dogs, cats, macaque monkeys, ceboids, and two bottlenose dolphins and one pilot whale were perfused transcardially after injection of a lethal dose of anesthetic with 4% paraformadehyde [the cetaceans were terminally ill and could be perfused through the descending aorta; in this case, the fixative was a mixture of 2.5% glutaraldehyde and 4% paraformaldehyde (Glezer et al., 1993, 1998; Hof et al., 1999)]. All of the other specimens were obtained after the animal had died of natural causes or was sacrificed for humane reasons and fixed by immersion for several weeks in neutral formalin (cetaceans, sea lion, artiodactyls, great apes, manatee, elephant) or 2–3 weeks in 4% paraformaldehyde [humans (Hof et al., 1995a; Nimchinsky et al., 1997)]. The perfusion protocol generally follows the one developed and previously described for macaque monkeys and dogs (Hof and Nimchinsky, 1992; Hof et al., 1996). Following fixation, all specimens were transferred to phosphate-buffered saline (PBS) containing 0.1% sodium azide at 4°C or were immersed in graded sucrose solutions and stored in a cryoprotectant solution at −20°C (Hof et al., 1995b). Some specimens were cryoprotected, frozen on dry ice, and stored in a −80°C freezer. In most cases, the entire brain was available to the authors, except in some hylobatids and great apes, where incomplete specimens were recovered, and in humans, where only one hemisphere was collected for immunohistochemistry.
All specimens were cut in 40–60 μm thick sections on a sliding microtome for large samples or on a cryostat. Sections were mounted every 500–1,000 μm, depending on the species, onto chrom-alum subbed slides and processed for Nissl staining (Fig. 1) and immunohistochemistry. The remaining sections were cryoprotected and stored in serial order at −20°C. For immunohistochemistry, the 40 μm thick free-floating sections were incubated for 48 hr at 4°C with monoclonal antibodies against dephosphorylated epitopes of the heavy and mid molecular weight subunits of NFP (SMI-32; 1:1,500 dilution; Sternberger Monoclonals, Lutherville, MD), PV, or CB, or with a polyclonal antibody against CR (Swant, Bellinzona, Switzerland), at a dilution of 1:5,000, 1:2,000, and 1:3,000, respectively, in PBS containing 0.3% Triton X-100 and 0.5 mg/ml bovine serum albumin. The sections were then processed with species-relevant antimouse or antirabbit secondary antibodies (1:200–1:500 dilution) and the avidin-biotin method using a Vectastain ABC kit (Vector Laboratories, Burlingame, CA) and 3,3′-diaminobenzidine as a chromogen. The immunoreactivity was subsequently intensified in 0.005% osmium tetroxide. In some specimens that had been kept over a long period in formalin, it was necessary to use a microwave pretreatment to visualize reliably the immunoreactivity. This was achieved by washing the sections in 3:1 v/v mixture of methanol and 3% H2O2 for 20 min to quench endogenous peroxidase reaction, and then by microwaving them in citrate buffer pH 6.5 for 3 min at maximum intensity using a commercial microwave oven. After cooling, sections were rinsed in PBS and processed for immunohistochemistry. Although consistent immunohistochemical labeling was generally obtained from these specimens, it should be kept in mind that differences in fixation protocols, duration of fixation, storage conditions, or postmortem delay, which were unavoidable, may have affected the staining patterns in some cases. In view of the size of certain brains, the entire cortex could not be exhaustively sampled. Rather, regions were targeted based on existing knowledge of homologies with monkeys, cats, and rodents, so that a consistent number of areas could be analyzed in most specimens. These included at least the primary and secondary visual cortex, the auditory cortex, and the dorsolateral frontal cortex.
Emergence of a Unique Cellular Specialization in Hominids
The evolution of the neocortex in primates has long been recognized to be the result of great expansion of cortical areas, with a roughly sevenfold increase in cortical volume between strepsirrhines and humans after correcting for body size. In spite of gross morphology and morphometric differences, pyramidal neurons and nonpyramidal interneuron populations have remained remarkably constant across primate species. In fact, the regional distribution of these cell classes has permitted the definition of anatomical boundaries between cortical domains not only in primates but in other mammalian orders as well. A number of studies have pointed to the usefulness of NFP and calcium-binding protein as markers of regional chemoarchitectural features in the primate visual and auditory cortex (Hof and Morrison, 1995; Preuss and Coleman, 2002), as well as in high-order association cortices that have a less distinct cytoarchitecture than primary sensory or motor fields (Carmichael and Price, 1994; Nimchinsky et al., 1997; Vogt et al., 2001, 2005; Sherwood et al., 2003a, 2003b, 2004). In addition, some cortical domains are characterized by the presence of specialized neurons, such as Meynert and Betz cells in the primary visual and primary motor cortices in primates, that exhibit distinct morphology and distribution patterns related to specific projections and functions (Sherwood et al., 2003b, 2004; Rivara et al., 2003). Such specializations are also observed in other species. For example, the primary motor cortex of large carnivores exhibits gigantic NFP-immunoreactive neurons as well as very large multipolar CR-expressing interneurons in layers III and V that are not encountered in other species. Large Meynert-like neurons enriched in NFP have also been reported in the primary visual cortex of the cat (van der Gucht et al., 2001, 2005).
The cingulate cortex and insula of hominids are distinguished by a remarkable cellular specialization, the spindle cells, that are characterized by a vertical, fusiform morphology, very large size, and high levels of NFP immunoreactivity (Fig. 1) (Nimchinsky et al., 1995, 1999). They are prevalent in a restricted sector of the anterior cingulate cortex [areas 25, 24a, and 24b (Vogt et al., 1995)] and are also numerous in the anteroventral agranular insular cortex. These neurons are found exclusively in hominids and have not been reported in any other mammalian species investigated thus far (including other primate species) (Nimchinsky et al., 1999; Hof et al., 2000). In the human and in the bonobo (Pan paniscus) anterior cingulate cortex (Fig. 1A and B), spindle cells occur most often in small clusters and are found exclusively in layer Vb. The common chimpanzee (Pan troglodytes) has abundant spindle cells, but clusters are not as conspicuous. In the gorilla (Gorilla gorilla), the distribution is similar to that in the common chimpanzee, but there are far fewer spindle cells. In the orangutan (Pongo pygmaeus), spindle cells are rare, and they are notably absent in lesser apes (e.g., the white-handed gibbon, Hylobates lar). As a percentage of pyramidal cell numbers in layer V of area 24, spindles cells account for < 1% in orangutan, ∼ 2% in gorilla, ∼ 4% in common chimpanzee, ∼ 5% in bonobo, and ∼ 6% in human, demonstrating their rarity. Stereologic volume estimates indicate that spindle cells are larger than neighboring layer V pyramidal cells and considerably larger than the small fusiform neurons seen in layer VI. They are larger in chimpanzees and human than in gorilla and orangutan, in contrast to pyramidal cells and small fusiform neurons for which no size difference occurs among these species. The volume of spindle cells is strongly correlated with encephalization, which is not the case for the other neuron types (Nimchinsky et al., 1999). These observations suggest the occurrence of functional modifications during the last 15–20 million years in cortical regions that play major roles in the regulation of autonomic function, cognition, self-awareness, emotionality, and vocalization. These spindle cells are a particular type of projection neuron, as they send an axon in the subcortical white matter, although their exact domain(s) of projection cannot be ascertained in the species in which they are present. They may, however, provide well-defined projections, similar to Meynert cells or Betz cells, and their exclusive presence in humans and their closest relatives suggests that the anterior cingulate and insular cortices have been subjected to strong adaptive pressure along the hominid lineage in recent evolution. This view is also supported by the fact that areas 24 and 25 of the anterior cingulate cortex of hominids also contain a population of CR-expressing small pyramidal neurons in layer Va that, much like the spindle cells, are not observed in other species (Hof et al., 2001).
Cetaceans and Artiodactyls Share Several Cytoarchitectural and Neurochemical Features
Extensive paleontologic and molecular evidence indicates that extant cetaceans (whales, dolphins, and porpoises) had a terrestrial origin and are closely related to artiodactyls; particularly among them, hippopotamuses (Thewissen, 1998). In fact, recent genetic data indicate a sister-group relationship between cetaceans and hippopotamuses, placing cetaceans and artiodactyls within a superorder Cetungulata (Milinkovitch, 1995; Buntjer et al., 1997; Gatesy, 1998; Milinkovitch et al., 1998; Nomura et al., 1998). Reports of the cyto- and chemoarchitecture of Odontoceti, particularly of visual and auditory regions, and analysis of neocortical neurons in a few large artiodactyls have indeed revealed commonalities in cortical organization across these species (Morgane et al., 1988, 1990; Glezer et al., 1992, 1993, 1998; Hof et al., 1992; 1999).
The cetacean neocortex is characterized by a general absence of granularity, a thicker and far more cellular layer I than in most terrestrial species, the presence of large, atypical neurons in layer II, and very large pyramidal neurons at the border between layers III and V. This pattern is observed throughout the neocortex with few variations among regions (Morgane et al., 1988), although the local complexity of cellular architecture in the dolphin neocortex is increasingly recognized (see Hof et al., 2005; Manger et al., 1998). Comparable cytoarchitectural patterns have been described in the neocortex of large artiodactyls (Hof et al., 1999). The distribution and morphology of NFP-immunoreactive neurons are comparable in cetaceans and artiodactyls, but differ considerably from that in primates, carnivores (Fig. 2A–C), and rodents. In cetaceans and artiodactyls, NFP is expressed in very large pyramidal neurons located in the deep portion of layer III and in upper layer V (Hof et al., 1992). The later neurons present as clusters of 3–6 neurons regularly spaced throughout the cortical mantle and intensely labeled with prominent apical dendrites extending well into layer I (Fig. 2D–F), with no major regional variability in their densities, whereas in primates and carnivores, large pyramidal NFP-immunoreactive pyramidal cells are present in layers III and V–VI with clear regional patterns of distribution.
Cetaceans and artiodactyls also share generally comparable staining patterns for the three calcium-binding proteins in the neocortex (Glezer et al., 1992, 1993, 1998; Hof et al., 1999), which generally set them apart from other mammals. Figure 3 shows several examples of calcium-binding protein distribution in a monotreme and a marsupial (Fig. 3A–C), a xenarthran (Fig. 3D and E), rodents (Fig. 3F and O), and carnivores (Fig. 3G and I–M) that can be used as a comparison to the specializations seen in a dolphin (Fig. 3H) and a giraffe (Fig. 3N). In cetaceans, PV is present in sparsely distributed large stellate neurons located in layers IIIc/V, and a comparable paucity of labeled neurons is found in artiodactyls. A few small pyramidal neurons in layer III also exhibit PV immunoreactivity in dolphins. In cetaceans and artiodactyls, CB and CR are far more numerous and occur in large fusiform, bipolar or multipolar neurons in layers I, II, and superficial III. CB-containing neurons are much less numerous and less intensely stained than CR-immunoreactive neurons. The CR-containing neurons located in layer I have a morphology quite comparable to that of the bipolar/bitufted CB- or CR-expressing neurons typically seen in layer II of other mammals such as rats, carnivores, and primates (Ballesteros Yáñez et al., 2005), whereas the CR-containing neurons in layers II and III are much larger and more variable in shape than in other species, with a predominance of multipolar and fusiform types (Fig. 3). These neurons have long dendrites that extend into layers I and III (Fig. 3H). Very large CR-immunoreactive neurons are also encountered in layers V and VI, especially in the neocortex of large artiodactyls, such as the giraffe (Giraffa camelopardalis), llama (Lama glama), and camel (Camelus dromedarius), whereas they are less numerous in the pig and in smaller ruminants (Fig. 3N). A few pyramidal-like neurons in layer III are also faintly CR-immunoreactive in dolphins, and the large pyramidal neurons in layer IIIc/V contain low levels of CB. Interestingly, a population of NFP-immunoreactive neurons in layer III of neocortex in anthropoid primates also expresses CB (Kondo et al., 1994). The distribution and typology of calcium-binding proteins in artiodactyls and cetaceans show some commonalities with those in insectivorous bats and hedgehogs but differ from those in rodents and primates (Glezer et al., 1993, 1998; Hof et al., 1999), suggesting that these traits may distinguish among some members of Laurasiatheria and Euarchontoglires (Murphy et al., 2001a, 2001b). Large multipolar PV-containing neurons are observed in the deep layers of the neocortex of hedgehogs as in dolphins, unlike the patterns in rodents, carnivores, and primates, where PV-expressing cells are more numerous in layers III and IV. As in hedgehog, a basal member of the Laurasiatheria, CB and CR are the dominating calcium-binding proteins in cetaceans and artiodactyls. In the cetacean primary visual and auditory cortices, CB- and CR-immunoreactive neurons represent about 40% of the total number of neurons, whereas PV is present only in about 5% of the neurons (Glezer et al., 1993), and these values are probably lower in large artiodactyls. In cetaceans, the proportion of calcium-binding protein-containing neurons is about twice as high as in primates and rodents (Glezer et al., 1993). This may represent a shared trait in whale and ungulate brains. That the cetacean and artiodactyl neocortex is dominated by CB and CR may also constitute an ancestral trait, observed as well in some insectivorous bats and hedgehogs, making it ancestral for just Laurasiatheria, but not for all mammals as insectivores are no longer recognized as a valid monophyletic clade (Murphy et al., 2001a, 2001b); in contrast to PV, these two proteins predominate in phylogenetically older neural systems mammals (Glezer et al., 1993, 1998; Jones, 1998).
Relationships of Neurochemical Phenotype and Phylogenetic Affinities
The variable patterns in distribution of NFP and the three calcium-binding proteins among species can be analyzed in the context of the general cytoarchitecture of the mammalian neocortex. In fact, several sets of cortical organizational patterns emerge from our data and the available literature. Generally, species showing a high degree of morphologic differentiation of neocortical areas and a variable development of layer IV and substantial variation in neuronal size and packing densities across the cortical plate are characterized by a balanced representation of the three calcium-binding proteins and morphological diversity of NFP-immunoreactive pyramidal neurons across cortical regions. In contrast, species characterized by a greater cytoarchitectural monotony throughout the cortical mantle, a poorly defined or lack of layer IV in most regions, and the presence of very large pyramidal cells in all neocortical areas display a predominance of CB- and CR-containing populations in comparison to PV-immunoreactive neurons and rather uniform NFP-containing pyramidal cell morphology. The first type occurs in primates, rodents, carnivores, and, to some extent, megachiropterans, as well as in tree shrews and lagomorphs, whereas the second type is present in cetaceans and ungulates. This fits nicely with a Laurasiatheria vs. Euarchontoglires split (Murphy et al., 2001a, 2001b); allowing for some convergent evolution in the Carnivora, the distribution and typology of calcium-binding proteins in this order show some commonalities to those seen in rodents and primates (Fig. 3) (Glezer et al., 1993, 1998; Hof et al., 1999; Ballesteros Yáñez et al., 2005). In addition, hedgehogs and microchiropterans display a combination of both character complexes, with the second type being generally predominant (Hof et al., 1999). These observations in insectivores need to be confirmed in other species of this paraphyletic group such as representative Afrosoricids and Macroscelids. As a rule, while general patterns are shared by many species, frequently certain patterns are unique to a single taxon or occur at the superfamily level. Thus, the general classes of calcium-binding protein-containing interneurons characterized in the rat, macaque monkey, and human are consistently observed among rodents, lagomorphs, primates, carnivores, as well as to some degree in megachiropterans and tree shrews (Hof et al., 1999; Ballesteros Yáñez et al., 2005). Yet megachiropterans are characterized by the unique absence of neocortical expression of CR in interneurons (while thalamic neurons do express it). However, in prototherians, marsupials, xenarthran, cetaceans, and ungulates, PV, CB, and CR are encountered in diverse types of pyramidal cells, as well as in classes of interneurons unique to these species, with much less consistency than in other mammals (Fig. 3) (Hof et al., 1999). In some cases, these features may even be interpreted as examples of homoplasy, with certain characters occurring in nonrelated groups of mammals, such as the paucity of PV expression observed in the neocortex of cetungulates and marsupials, or the fact that CB and CR predominate in cetaceans, microchiropterans, hedgehogs, and artiodactyls, not only in the neocortex, but also in subcortical systems (Glezer et al., 1993, 1998; Hof et al., 1999). The functional repercussions of these specializations remain, however, unknown.
It is interesting that in canids, felids, and pinnipeds, very large CR-immunoreactive, multipolar neurons occur in layer III of motor cortices (Fig. 3). These neurons are morphologically comparable to the very large CR-containing neurons found in layers III, V, and VI throughout the neocortex of large artiodactyls and cetaceans (Hof et al., 1996, 1999). In carnivores, these neurons are seen only in agranular cortices and the cortex of large ungulates and cetaceans shows a poorly differentiated layer IV only in certain regions (Morgane et al., 1990; Glezer et al., 1993, 1998; Hof et al., 1999). The presence of these particular CR-immunoreactive cells in agranular cortex deserves further investigation and may point to possible phylogenetic affinities between carnivores, ungulates sensu lato (i.e., “Fereungulata”) (see Johnson et al., 1994), carnivores being a sister taxon to the clade containing perissodactyls and cetartiodactyls (Murphy et al., 2001a, 2001b). In this context, the dog neocortex displays several differences in neocortical neurochemical organization compared to anthropoids in spite of many similarities in cortical organization and connectivity between carnivores and primates (Hof et al., 1996, 1999). Canids have a high degree of cellular specialization in primary motor and sensory cortices, contrasting with fairly homogeneous patterns in association cortices. Generally, the distribution patterns and typology of PV- and CB-immunoreactive interneurons in dogs and cats are comparable reflecting their close phylogenetic proximity (Demeulemeester et al., 1991; Hof et al., 1996). Large PV- and CB-containing pyramidal cells are present only in primary motor and visual regions, and high numbers of large PV-immunoreactive basket cells and multipolar CR-containing interneurons occur only in primary motor, somatosensory, and visual areas (Hof et al., 1996). It is not known whether the large PV- and CB-containing pyramidal cells in these species represent subclasses of NFP-enriched neurons and double-labeling studies would be required to answer this question in carnivores and in other taxa. It should be noted in this context that at least some CB-expressing pyramidal cells in primates also display NFP immunoreactivity (Hayes and Lewis, 1992). That the association cortex in the dog appears less differentiated neurochemically than presumably homologous areas in primates supports the concept that cortical regions such as inferior temporal and prefrontal cortex achieved a particular degree of elaboration in primates compared to other mammals (Preuss, 1995). Whether this argument can in fact be generalized awaits further analyses in additional canid and felid species in comparison to primates as well as representative taxa of other orders. Further limitations of our study is that our sample is limited to representative species of Boreoeutheria, permitting assessment of affinities and character specificity only among Laurasiatheria and Euarchontoglires to the exclusion of Afrotheria, for which no detailed regional studies of such patterns are currently available. As such, some of these neurochemical features may relate to event in mammalian brain evolution dating back to about 80–90 million years ago at the time Laurasiatheria (i.e., cetungulates, carnivores, pangolins, bats, hedgehogs, and moles) diverged from Euarchontoglires (i.e., rodents, lagomorphs, primates, tree shrews, and colugos). It is also worth noting that the Australian echidna presents an NFP-expressing cells in layer V in its neocortex (Hassiotis et al., 1994), whereas calcium-binding proteins in this species show a fairly variable morphology and distribution compared to other mammals, including marsupials (Fig. 3) (Hof et al., 1999). Although it would be important to analyze NFP distribution species directly related to echidna such as the platypus and in representative marsupials, the presence of layer V NFP-rich cells may be interpreted has a conservative trait among mammals.
Implications for Cortical Function
Calcium-binding protein-containing interneurons are known to influence the activity of pyramidal neurons in a manner specific to each cell class (Condé et al., 1994; DeFelipe, 1997). Parvalbumin-immunoreactive basket and chandelier cells provide an innervation of the perikaryon and axon initial segment, respectively, whereas CR- and CB-containing bipolar and double bouquet cells target mostly the apical dendritic arbors of pyramidal neurons, and as comparable types of interneurons have been described in many species (Condé et al., 1994; DeFelipe, 1997; Hof et al., 1999; Ballesteros Yáñez et al., 2005), the role of calcium-binding protein in cortical integration is likely to be similar to a large degree at least in rodents, carnivores, and archontans. Yet certain differences exist at the level of particular neuronal subclasses, as recently revealed in a study of CB-expressing interneurons in primates compared to rodents, lagomorphs, carnivores, and artiodactyls (Ballesteros Yáñez et al., 2005). These authors reported that in the nonprimate species they investigated, the vertical bundles of axons of CB-immunoreactive double bouquet cells are not observed except in the visual cortex of carnivores. Even there these axonal “horse tails” are far less prevalent than in the primate neocortex. These data indicate that although CB-expressing cell types comparable to the typical primate double bouquet cell occur in these species, their axonal projections are likely to be organized very differently and may not contribute to the columnar connectivity of the neocortex in the same manner as is observed in primates (Ballesteros Yáñez et al., 2005). Whether this sort of difference in axonal organization of interneurons can be extended to typical PV- and CR-immunoreactive neurons remains to be demonstrated. In any case, the degree to which such functional relationships can be extended to all mammalian orders is difficult to define owing to such differences within a given cell type across species. The relative rarity of PV-immunoreactive neurons in cetaceans and ungulates could be interpreted as an ancestral retention. It also occurs in echolocating bats and in hedgehogs, which have been claimed to have retained many plesiomorphic features (Glezer et al., 1988). This suggests that the inhibitory microcircuitry of the cetacean and ungulate neocortex may be characterized by primitive features involving different cellular interactions compared to other mammalian lineages. In fact, the neocortex of cetaceans and large ungulates appears to contain an inordinate number of cortical columns as revealed by clusters of large NFP-containing layer V pyramidal cells in layer IIIc/V (Glezer et al., 1988; Morgane et al., 1988; Hof et al., 1992). Cortical integration in cetaceans may take place mostly in the highly cellular, comparatively thick layer I that contains approximately 70% of the neocortical synapses in these species (Glezer and Morgane, 1990). This is consistent with the observation that the majority of CB- and CR-containing interneurons are located in layers I and II in cetaceans, the less abundant PV-immunoreactive cells being in the vicinity of layers IIIc/V and VI pyramidal cells. The distribution of PV-immunoreactive neurons indicates that these cells may represent basket cells and that the axons of CB- and CR-immunoreactive interneurons may be located where most of the inputs to the neocortex terminate around the apical dendrites of the deep-layer pyramidal neurons (Glezer et al., 1988; Morgane et al., 1988). In spite of lack of ultrastructural evidence at the synaptic level, it is possible that in cetaceans and large ungulates, CB-, some CR-, and PV-immunoreactive neurons play a very similar role in neocortical microcircuits as in primates and rodents. It is therefore likely that functionally, interspecies differences in neocortical organization notwithstanding, calcium-binding proteins identify classes of interneurons that subserve similar tasks across mammals.
Although the precise function of NFP is not understood, it has a restricted distribution among certain subsets of corticocortical circuits in primates (Hof et al., 1995b, 1996) and is involved in neurodegenerative lesions in subsets of pyramidal neurons particularly vulnerable in dementing disorders in humans (Bussière et al., 2003; Hof and Morrison, 2004). This fact is particularly important considering that NFP-enriched spindle neurons of the hominid cingulate and frontoinsular cortex have been shown to be severely affected in the degenerative process of Alzheimer's disease, suggesting that the neuronal susceptibility that occurs in the human brain in the course of age-related dementing illnesses appeared only recently during primate evolution (Nimchinsky et al., 1995, 1997, 1999; Rapoport, 1999). We have proposed that NFP confers unique neurochemical and morphologic properties to select neuronal subpopulations subserving a range of highly specialized functions in the neocortical connectivity (Hof et al., 1995a, 1995b; Nimchinsky et al., 1996; Bussière et al., 2003; Hof and Morrison, 2004). NFP may thus be present, to some degree, in functionally homologous subsets of cortical output neurons in cetaceans and artiodactyls. The similarities in neurochemical specialization of the cetacean and artiodactyl neocortex parallel the paleontogical and molecular evidence that these species share a relatively recent common ancestor, and that much like primates, the evolution of the species with the largest brains (the delphinids) is a recent event (Marino et al., 2005). Also, it is intriguing that in both cetacean and artiodactyls, compared to other mammals, the calves are born with precocious physical maturity, a crucial factor for survival in the aquatic milieu and, in the case of terrestrial herbivores, to escape predators. It is possible that pedomorphosis occurs in these species. In the case of cetaceans, possible pedomorphic features include the retention in adults of the pontine, mesencephalic and cephalic flexures that are visible only in embryos in other mammals, and the very large size of the brain at birth (Glezer et al., 1998; Hof et al., 1999). Furthermore, that the neocortex of cetaceans and ungulates is dominated by CB- and CR-containing interneurons may also represent a pedomorphic feature because these calcium-binding proteins appear first during development and in rodent, carnivores, and primates and are preferentially distributed in phylogenetically older neural systems (Glezer et al., 1993, 1998; Jones, 1998; Hof et al., 1999).
Although there are major gaps in our knowledge of the evolutionary history of neocortical organization in mammals and of the chemical organization of the cerebral cortex in most species, these observations together indicate that brain organization and neurochemical cellular specialization reflect evolutionary relationships among many mammalian species.
The authors thank W.G.M. Janssen, P.W.H. Lee, and B. Wicinski for expert technical assistance, Drs. J.M. Allman, E.A. Nimchinsky, E. van der Gucht, J.M. Erwin, and M.A. Raghanti for help and discussion. The great ape brain materials were obtained from the Great Ape Aging Project and the Foundation for Comparative and Conservation Biology.
- 1993. Calcium-binding proteins: selective markers of nerve cells. Cell Tissue Res 271: 181–208. , , .
- 2005. Double bouquet cell in the human cerebral cortex and a comparison with other mammals. J Comp Neurol 486: 344–360. , , , , , .
- 1990. Distribution of parvalbumin immunoreactivity in the visual cortex of Old World monkeys and humans. J Comp Neurol 301: 417–432. , , , .
- 2005. Regional analysis of neurofilament protein immunoreactivity in the hamster's cortex. J Chem Neuroanat 29: 193–208. , , , .
- 1997. Artiodactyl interspersed DNA repeats in cetacean genome. J Mol Evol 45: 66–69. , , .
- 2003. Progressive degeneration of nonphosphorylated neurofilament protein-enriched pyramidal neurons predicts cognitive impairment in Alzheimer's disease: a stereologic analysis of prefrontal cortex area 9. J Comp Neurol 463: 281–302. , , , , , .
- 1989. A monoclonal antibody to neurofilament protein (SMI-32) labels a subpopulation of pyramidal neurons in the human and monkey neocortex. J Comp Neurol 282: 191–205. , .
- 1991. A subpopulation of primate corticocortical neurons is distinguished by somatodendritic distribution of neurofilament protein. Brain Res 539: 133–136. , , .
- 1994. Architectonic subdivision of the orbital and medial prefrontal cortex in the macaque monkey. J Comp Neurol 346: 366–402. , .
- 1994. Local circuit neurons immunoreactive for calretinin, calbindin D-28k or parvalbumin in monkey prefrontal cortex: distribution and morphology. J Comp Neurol 341: 95–116. , , , , .
- 1989. Synapses of double bouquet cells in monkey cerebral cortex visualised by calbindin immunoreactivity. Brain Res 503: 49–54. , , .
- 1992. High-resolution light and electron microscopic immunohistochemistry of colocalized GABA and calbindin D-28k in somata of double bouquet cell axons of monkey somatosensory cortex. Eur J Neurosci 4: 46–60. , .
- 1997. Types of neurons, synaptic connections and chemical characteristics of cells immunoreactive for calbindin-D28k, parvalbumin and calretinin in the neocortex. J Chem Neuroanat 14: 1–19. .
- 1991. Calcium binding proteins and neuropeptides as molecular markers of GABAergic interneurons in the cat visual cortex. Exp Brain Res 84: 538–544. , , , , , .
- 1990. Cajal-Retzius cell ontogenesis and death in mouse brain visualized with horseradish peroxidase and electron microscopy. Neuroscience 36: 839–856. , .
- 1998. Molecular evidence for the phylogenetic affinities of Cetacea. In: ThewissenJGM, editor. The emergence of whales: evolutionary patterns in the origin of Cetacea. New York: Plenum. p 63–111. .
- 1988. The “initial” brain concept and its implications for brain evolution in cetacea. Behav Brain Sci 11: 75–116. , , .
- 1990. Ultrastructure of synapses and Golgi analysis of neurons in neocortex of the lateral gyrus (visual cortex) of the dolphin and pilot whale. Brain Res Bull 24: 401–427. , .
- 1992. Calretinin-immunoreactive neurons in the primary visual cortex of dolphin and human brains. Brain Res 595: 181–188. , , .
- 1993. Calcium-binding protein-containing neuronal populations in mammalian visual cortex: a comparative study in whales, insectivores, bats, rodents, and primates. Cereb Cortex 3: 249–272. , , , .
- 1998. Comparative analysis of calcium-binding protein-immunoreactive neuronal populations in the auditory and visual systems of the bottlenose dolphin (Tursiops truncatus) and the macaque monkey (Macaca fascicularis). J Chem Neuroanat 15: 203–237. , , .
- 1997. Three distinct families of GABAergic neurons in rat visual cortex. Cereb Cortex 7: 347–358. , .
- 2005. Brain of the Africa elephant (Loxodonta africana): neuroanatomy from magnetic resonance images. Anat Rec 287A: 1117–1127. , , , , , .
- 2004. Cyto- and chemoarchitecture of the crebral cortex of the Australian echidna (Tachyglossus aculeatus): I, areal organization. J Comp Neurol 475: 493–517. , , .
- 1992. Nonphosphorylated neurofilament protein and calbindin immunoreactivity in layer III pyramidal neurons of human neocortex. Cereb Cortex 2: 56–67. , .
- 1989. Two classes of cortical GABA neurons defined by differential calcium binding protein immunoreactivities. Exp Brain Res 76: 467–472. , , , , , .
- 1991. Neocortical neuronal subpopulations labelled by a monoclonal antibody to calbindin exhibit differential vulnerability in Alzheimer's disease. Exp Neurol 111: 293–301. , .
- 1992. Regional distribution of neurofilament and calcium-binding proteins in the cingulate cortex of the macaque monkey. Cereb Cortex 2: 456–467. , .
- 1992. The primary auditory cortex in cetacean and human brain: a comparative analysis of neurofilament protein-containing pyramidal neurons. Neurosci Lett 146: 91–95. , , , , , .
- 1995. Neurofilament protein defines regional patterns of cortical organization in the macaque monkey visual system: a quantitative immunohistochemical analysis. J Comp Neurol 352: 161–186. , .
- 1995a. The human orbitofrontal cortex: cytoarchitecture and quantitative immunohistochemical parcellation. J Comp Neurol 359: 48–68. , , .
- 1995b. Neurochemical phenotype of corticocortical connections in the macaque monkey: quantitative analysis of a subset of neurofilament protein-immunoreactive projection neurons in frontal, parietal, temporal, and cingulate cortices. J Comp Neurol 362: 109–133. , , .
- 1996. Distribution of neuronal populations containing neurofilament protein and calcium-binding proteins in the canine neocortex: regional analysis and cell typology. J Chem Neuroanat 11: 81–98. , , , .
- 1999. Cellular distribution of the calcium-binding proteins parvalbumin, calbindin, and calretinin in the neocortex of mammals: phylogenetic and developmental patterns. J Chem Neuroanat 16: 77–116. , , , , , , .
- 2000. Neurochemical and cellular specializations in the mammalian neocortex reflect phylogenetic relationships: evidence from cetaceans, artiodactyls, and primates. Brain Behav Evol 55: 300–310. , , , .
- 2001. An unusual population of pyramidal neurons in the anterior cingulate cortex of hominids contains the calcium-binding protein calretinin. Neurosci Lett 307: 139–142. , , , .
- 2004. The aging brain: morphomolecular senescence of cortical circuits. Trends Neurosci 27: 607–613. , .
- 2005. Cortical complexity in cetacean brains. Anat Rec 287A: 1141–1152. , , .
- 1991. Immunocytochemical localization of calretinin in the forebrain of the rat. J Comp Neurol 304: 198–218. , .
- 1994. Phylogeny through brain traits: more characters for the analysis of mammalian evolution. Brain Behav Evol 43: 319–347. , , , .
- 1998. Viewpoint: the core and matrix of thalamic organization. Neuroscience 85: 331–345. .
- 2002. Neurofilament triplet proteins are restricted to a subset of neurons in the rat neocortex. J Chem Neuroanat 24: 163–171. , , , , , .
- 1994. Neurochemical gradient along the monkey occipito-temporal cortical pathway. Neuroreport 5: 613–616. , , , .
- 1998. Modular subdivisions of dolphin insular cortex: does evolutionary history repeat itself? J Cogn Neurosci 10: 153–166. , , , , .
- 2005. Origin and evolution of large brains in toothed whales. Anat Rec 281A: 1247–1255. , , .
- 1997. Calretinin-immunoreactive local circuit neurons in area 17 of the Cynomolgus monkey, Macaca fascicularis. J Comp Neurol 379: 113–132. .
- 1995. Molecular phylogeny of cetaceans prompts revision of morphological transformations. Trends Ecol Evol 10: 328–334. .
- 1998. Cetaceans area highly derived artiodactyls. In: ThewissenJGM, editor. The emergence of whales: evolutionary patterns in the origin of Cetacea. New York: Plenum. p 113–131. , , .
- 1988. Visual cortex of the dolphin: an image analysis study. J Comp Neurol 273: 3–25. , , .
- 1990. Comparative and evolutionary anatomy of the visual cortex of the dolphin. In: JonesEG, PetersA, editors. Comparative structure and evolution of cerebral cortex, pt. 2. New York: Plenum. p 215–262. , , .
- 1998. Neurochemical organization of the primate visual cortex. In: BloomFE, BjörklundA, HökfeltT, editors. The primate nervous system: handbook of chemical neuroanatomy, vol. 14, part 2. New York: Elsevier. p 299–430. , , .
- 2001a. Molecular phylogenetics and the origin of placental mammals. Nature 409: 614–618. , , , , , .
- 2001b. Resolution of the early placental mammal radiation using Bayesian phylogenetics. Science 294: 2348–2351. , , , , , , , , , , .
- 1995. Spindle neurons of the human anterior cingulate cortex. J Comp Neurol 355: 27–37. , , , .
- 1996. Neurochemical, morphologic and laminar characterization of cortical projection neurons in the cingulate motor areas of the macaque monkey. J Comp Neurol 374: 136–160. , , , .
- 1997. Neurofilament and calcium-binding proteins in the human cingulate cortex. J Comp Neurol 384: 597–620. , , , .
- 1999. A neuronal morphologic type unique to humans and great apes. Proc Natl Acad Sci USA 96: 5268–5273. , , , , , .
- 1998. A SINE species from hippopotamus and its distribution among animal species. Mammal Genome 9: 550–555. , , , , .
- 1995. Do rats have prefrontal cortex? the Rose-Woolsey-Akert program reconsidered. J Cogn Neurosci 7: 1–24. .
- 1996. Parvalbumin-like immunoreactivity of layer V pyramidal cells in the motor and somatosensory cortex of adult primates. Brain Res 712: 353–357. , .
- 1997. Multiple divisions of macaque precentral motor cortex identified with neurofilament antibody SMI-32. Brain Res 767: 148–153. , , , .
- 1999. Distinctive compartmental organization of human primary visual cortex. Proc Natl Acad Sci USA 96: 11601–11606. , , .
- 2002. Human-specific organization of primary visual cortex: alternating compartments of dense Cat-301 and calbindin immunoreactivity in layer 4A. Cereb Cortex 12: 671–691. , .
- 1999. How did the human brain evolve? a proposal based on new evidence from in vivo brain imaging during attention and ideation. Brain Res Bull 50: 149–165. .
- 1992. Calretinin in rat brain: an immunohistochemical study. Neuroscience 46: 101–134. , .
- 2003. Stereologic characterization and spatial distribution patterns of Betz cells in the human primary motor cortex. Anat Rec 270A: 137–151. , , , .
- 1992. Immunohistochemical markers in rat cortex: colocalization of calretinin and calbindin-D28k with neuropeptides and GABA. Brain Res 587: 147–157. .
- 2003a. Variability of Broca's area homologue in African great apes: implications for language evolution. Anat Rec 271A: 276–285. , , , , .
- 2003b. Evolution of specialized pyramidal neurons in primate visual and motor cortex. Brain Behav Evol 61: 28–44. , , , , , , , , , .
- 2004. Cortical orofacial motor representation in Old World monkeys, great apes, and humans: II, stereologic analysis of chemoarchitecture. Brain Behav Evol 63: 82–106. , , , .
- 1998. The emergence of whales: evolutionary patterns in the origin of Cetacea. New York: Plenum. .
- 1990. Calcium-binding proteins as markers for subpopulations of GABAergic neurons in monkey striate cortex. J Comp Neurol 298: 1–22. , , .
- 2001. Neurofilament protein: a selective marker for the architectonic parcellation of the visual cortex in adult cat brain. J Comp Neurol 441: 345–368. , , .
- 2004. Characteristic expression of neurofilament protein defines six cortical areas and several subcortical divisions in mouse visual system. Soc Neurosci Abstr 30: 300.7. , , .
- 2005. Light-induced Fos expression in phosphate-activated glutaminase- and neurofilament protein-immunoreactive neurons in cat primary visual cortex. Brain Res 1035: 60–66. , , , .
- 1995. Human cingulate cortex: surface features, flat maps, and cytoarchitecture. J Comp Neurol 359: 490–506. , , , .
- 2001. Cytoarchitecture of human caudomedial cingulate, retrosplenial, and caudal parahippocampal cortices. J Comp Neurol 438: 353–376. , , , .
- 2005. Architecture and neurocytology of monkey cingulate gyrus. J Comp Neurol 485: 218–239. , , , .