Evolutionary developmental biologists try to understand the mechanisms that produce homologous and homoplastic structures, and how variation and novelties originate during development and in evolution. Understanding these principles is essential for studying evolution of a complex biological structure, such as the brain. In this section, I provide a very brief analysis of the causes and principles of evolution, for then applying these ideas to the study of brain evolution in the following section.
1.05.2.1 Causes of Homology, Convergence, and Divergence: Development, Evolution, Epigenesis
Homologous structures can show a high degree of similarity (involving static or conservative evolution) or can differ in form and/or function (involving important evolutionary changes). Nonhomologous structures can reach a high degree of similarity by way of either convergent or parallel evolution (Striedter and Northcutt, 1991). Further, many biological systems exhibit new characters or features with no counterpart in the ancestor (i.e., they have been produced as novelties during evolution) (Striedter, 2005). The reason for all of these examples of static, divergent, convergent, or parallel evolution can be found in development and in the genetic regulatory programs underlying it. Developmental mechanisms and the genetic regulatory networks involved in the formation of complex structures are often highly integrated systems subjected to constraints impeding significant changes, and random variation. Developmental and evolutionary changes are possible due to the modular nature of biological systems, from the level of genetic cascades to the levels of cells and morpho-genetic fields (Raff, 1996; Gilbert et al., 1996). Modules, such as morphogenetic fields, are dynamic, highly stable, and self-regulated systems, able to absorb perturbations produced in the field or outside it (i.e., they are internally constrained). However, some changes in their genetic regulatory programs can alter their development and, if not lethal, can lead to phenotypic variations in the adult. The resultant phenotypes are then selected by external pressures. Genetic regulatory programs can be affected by epigenetic interactions with other modules, such extracellular signaling molecules involved in inductive interactions, and changes in these interactions can lead to phenotypic variations (Raff, 1996). In addition, genetic regulatory programs can be affected by epigenetic environmental features such as temperature, nutrition, or population density, and the phenotypic result can be different under different conditions (Raff, 1996; Hall et al., 2004; see Epigenetic Responses to a Changing Periphery - Wagging the Dog). This takes us to the concept of phenotypic plasticity, according to which each genome is able to produce a range of phenotypes, as a critical adaptive response to different environments (Waddington, 1956; Raff, 1996; Gilbert and Burian, 2003). The issue of developmental and evolutionary constraints, the principle of modularity, and the concept of morphogenetic field are explained in more detail in separate sections.
1.05.2.2 Internal Constraints in Development and Evolution. Epigenetic Constraints
In spite of the rich structural and functional variety of animal body forms found in nature, no more than about 35 different body plans exist, all of which appeared during the Cambrian radiation over a half billion years ago, indicating an astonishing stability of structure once integrated as a complex form (Raff, 1996). Although extensive evolutionary changes have occurred since the Cambrian, the underlying body patterns have been conserved. This means that there are constraints in evolution, which include at least physical/morphological, genetic, developmental, historical, and epigenetic constraints (Raff, 1996; Striedter, 1998, 2005). Once a body plan is assembled, the genetic regulatory programs and developmental mechanisms underlying it become tightly integrated, and significant change is severely constrained (Raff, 1996). The integration of genetic and developmental controls of a body plan during evolution is irreversible, which means that a body plan cannot be transformed into another one without fatally disrupting its ontogeny (Raff, 1996). Once an integrated body plan is established, selection favors improvements/ changes within that body plan, and disfavors new body plans that are unable to compete as well as established ones (Raff, 1996). In addition to genetic constraints due to tight integration of genetic networks making small changes lethal (this does not occur when duplications or alternative links exist), the genomic size is also a source of constraint since it affects properties such as cell size and division rate (Raff, 1996), which affect the morphology. Some simple vertebrates such as lungfish or salamanders have disproportionately huge genomes, much larger than that of humans (Raff, 1996). Larger genomes result in larger cells, and cells containing large genomes have a slower DNA replication, constraining the growth rate of the organism. In frogs and salamanders having a large genome and, as a result, larger cells, this is related to a simplification of brain morphology (Roth et al., 1994).
It appears that the developmental and genetic internal constraints are not equally tight during the whole development, thus allowing for some degree of variation within body plans observed during evolution. Such changes are possible due to the modular nature of development (genetic cascades and networks, transduction/signaling pathways, cells, morphogenetic fields), which allow changes in one module without affecting the rest (Gilbert et al., 1996; Raff, 1996; Gass and Bolker, 2003; Gilbert and Burian, 2003). In vertebrates, echinoderms, and possibly other animal groups, internal developmental constraints are maximal during the phylotypic stage, but are relatively loose during early and late development. The phylotypic stage (called pharyngula in vertebrates) represents the most evo-lutionarily conserved stage of development, in which embryos of different species share a comparable regional/modular organization and comparable expression patterns of developmental regulatory genes. The stability of this stage is due to the existence of tight and complex integration (interactions) between modules, which impeds variation without fatal consequences (Raff, 1996). Curiously, and paradoxically, this highly constrained stage, which later gives rise to a highly conserved body plan, can be achieved by way of very different developmental mechanisms, meaning that early development is loosely constrained. Striedter (1998) explains this paradox of many different developmental mechanisms leading to a conserved phylotypic stage by proposing the existence of attractors, represented by valleys in epigenetic landscapes. These attractors represent the different dynamic states of modules (Raff (1996); these are discussed in the following section). Integration between modules is less strict before and after the phylotypic stage, allowing evolutionary changes. When these changes occur after the phylotypic stage, they produce phenotypic variation in the adult that affect specific modules and their derivatives (mosaic evolution).
1.05.2.3 The Principle of Modularity: Genetic Cascades, Cells, Cell Fields, and Organs
As noted above, evolutionary changes during development are possible due to the modular organization of developing embryos. Developing organisms are made up of partially independent, interacting units or modules at several hierarchical levels (from genetic cascades to cell lineages and cell fields), which are able to change independently of the rest of the body, allowing mosaic evolution in both genes and morphology (Raff, 1996; Gass and Bolker, 2003; Gilbert and Burian, 2003). Modules are thus important units, in both development and evolution, that link genotype to phenotype. Within some limits (due to constraints), modules can become semi-independent entities (dissociation), allowing nonrandom evolutionary variation without fatal consequences for the embryo. Evolutionary variation can occur by changes in timing or location of regulatory gene expression or other processes (including signal transduction interaction between modules), by changes of growth rate or duration within a module, by module duplication followed by divergence (this occurs in the production of serial homologues), or by co-option (Gould, 1977; Gilbert et al., 1996; Raff, 1996; Gass and Bolker, 2003; Gilbert and Burian, 2003).
The correct expression of developmental regulatory genes is essential for body and body part formation. Changes in their expression (timing or location) can produce important morphological alterations, and the occurrence of such changes is considered a major cause for the production of morphological diversity in evolution (Carroll et al., 2001). Developmental regulatory genes, also called master control genes, constitute a small fraction of the genome, and encode transcription factors or signaling proteins, that directly (through direct DNA biding) or indirectly (through signal transduc-tion cascades) regulate the expression of other genes and control key aspects of development and formation of specific body parts (Carroll et al., 2001). The coding region of these genes is extremely well conserved in evolution, but the regulatory region, containing key elements (enhancers) for gene transcription regulation, can change in evolution, producing gene expression changes in timing or location. Such evolutionary variation in gene regulatory regions is possible due to its modular organization, which allows variations such as addition of new enhancers or binding sites for new activators or repressors, that can affect positively or negatively the gene transcription (Gilbert, 2000; Carroll et al., 2001). Evolutionary changes have often occurred at the level of the gene regulatory regions, as well as with the evolution of new activators, repressors, or cofactors acting directly or indirectly on the regulatory region of master control genes (Carroll et al., 2001). Changes can also occur at the level of the downstream targets of master control genes, also in the regulatory region of those target genes, which may produce, for example, a particular transcription factor that can act through a different downstream cascade of genes, producing a different effect and a different pheno-type. The modular nature of the regulatory region of genes, of the genetic cascades, and of the embryo (in different fields) makes possible that particular master control genes can have different effects when acting, in different contexts, in different parts of the body (Carroll et al., 2001). Modular interaction between genetic regulatory cascades is also essential to understand how so much morphological diversity and so many different cell types can be produced from relatively few master control genes. Through such interactions, different master gene combinations defining distinct morphogenetic fields can lead to different body parts and subdivisions (Carroll et al., 2001).
Developmental modules are dynamic and self-regulated entities, that have specific locations in the embryo, and show specific properties (such as specific gene expression patterns) and capacities of interaction with other modules, than can change dynamically during the course of development (Raff, 1996; Gass and Bolker, 2003). Thus, developmental modules are usually transient entities (e.g., cascades of developmental regulatory genes, such as those expressing transcription factors; or morphogenetic fields) that, in general, are not found as such in the adult but produce specific adult phenotypes (e.g., expression of specific structural genes, specific sets of differentiated cells, specific regions or organs). Each state of the module (e.g., each state to pass from being a stem cell, to a neural stem cell, and from here to a forebrain neural stem cell, and so on, to finally become a differentiated cholinergic neuron in the basal telencephalon; or the transient states between a morphogenetic field and a particular adult brain division) can be considered as a 'basin of attraction' or 'attractor', and developmental genes act as regulators to make the transition between different 'basins of attraction' (between transient cell types) (Kauffman, 1987, 1993; Raff, 1996; Striedter, 1998). Of interest, these attractors are the characters or natural units of comparison between different organisms for homology considerations. If the phylogenetic continuity criterion is met (analyzed using a cladistic method), particular attractors in two animals can be considered homologous (see Phylogenetic Character Reconstruction). However, it is important to note that homology at a particular attractor state (e.g., a transient state during development) does not necces-sarily mean complete homology at the final state. For example, a particular morphogenetic field can be homologous in two different animals, but its final product in one of the animals may contain cell populations homologous to similar cells in other animals, plus other nonhomologous cell populations (with no counterpart in other species or the common ancestor, as inferred using a cladistic analysis).
Although developmental modules generally lose some of their attributes after the major developmental events have occurred (e.g., when cells reach a differentiated state), there are cases of modules that either keep their properties (e.g., as stem cells) or are able to recover them in adult animals, and are responsible for cases of regeneration or continuous cell production. This would explain the continuous production of blood cells or epithelial cells in the adult (Gilbert, 2000). This also happens in the brain of different animals, including mammals, where localized foci of adult neurogenesis are found in the subventricular zone leading to the rostral migratory stream (giving rise to cells that migrate tangentially to the olfactory bulb) or in the hippo-campal formation (García-Verdugo et al., 2002; Alvarez-Buylla and García-Verdugo, 2002; Merkle et al., 2004). These examples of developmental modules keeping or recovering their 'generative' properties in the adult are more common in amphibians and fish, but often are also found in reptiles (Gilbert, 2000). For example, amphibians and reptiles are able to regenerate some distal body parts, such as the tail, following amputation. The enigmatic case of 'Wolffian' regeneration in urodele amphibians (regeneration of the eye lens but from a different origin, the iris edge; Gilbert, 2000) may also be a consequence of the persistence of developmental modules with intact generative and self-regulatory properties after the major developmental events have occurred (see also Puelles and Medina, 2002).
1.05.2.4 Morphogenetic Fields as Evolutionary and Developmental Higher-Order Modules Linking Genotype and Phenotype
Among modules, morphogenetic fields constitute higher-order units of embryonic development, that are specified by particular combinations of developmental regulatory genes (master control genes) and that give rise to specific cellular groups and structures found in the adult. They represent the major units establishing the link between genotype to phenotype, and the major units of ontogenetic and phylogenetic change (Gilbert et al., 1996). Further, they constitute major natural units or characters for comparison between species and for homology studies.
In the brain, morphogenetic fields are represented by segments (rhombomeres, prosomeres) and smaller divisions and subdivisions within them (Raff, 1996; Gilbert et al., 1996; Puelles and Rubenstein, 1993, 2003). The evidence for their existence is ample, and includes:
1. data on expression of developmental regulatory genes in different species, allowing visualization of fields as discrete expression domains characterized by specific combinations of genes (Simeone et al., 1992; Bulfone et al., 1993, 1995, 1999; Puelles and Rubenstein, 1993; Puelles et al., 2000; Bachy et al., 2002; Brox et al., 2003, 2004; Lumsden, 2004; see Origins of the Chordate Central Nervous System: Insights from Hemichordates);
2. knockout mutations showing specific alterations of the field and its derivatives, either by lack of formation, malformation, or transformation into an adjacent field (Schneider-Maunoury et al., 1993; McKay et al., 1994; Stoykova et al., 1996; Sussel et al., 1999; Yun et al., 2001, 2003; Bishop et al., 2002, 2003; Muzio et al., 2002);
3. fate map studies showing clonal restriction at least at the level of the ventricular zone (Fraser et al., 1990; Marin and Puelles, 1995; Cambronero and Puelles, 2000; Cobos et al., 2001b; Garcia-Lopez et al., 2004); and
4. in the case of brain segments, transplantation studies showing replication of the field and its derivatives when a specific rhombomere is grafted into a different position of the embryo (Kuratani and Eichele, 1993; Martinez et al., 1995).
During development, fields can be patterned and/ or can reach a new order (including new gene expression) by interaction with other fields. This can occur by way of signaling proteins produced by cells in particular fields or organizer centers (such as Sonic hedgehog, Wnt, BMP, and FGF proteins), that are secreted to the extracellular medium and diffuse to act at a distance in a concentration-dependent manner (Carroll et al., 2001). The response of cells to such inductive proteins depends on both the distance from the source and their expression of appropiate receptors. Examples of such interactions are common in the central nervous system and constitute an important mechanism for patterning (Ericson et al., 1997; Shimamura and Rubenstein, 1997; Ragsdale and Grove, 2001). The internal order of fields can also be affected by signaling from boundaries separating the field from neighbors (Echevarría et al., 2003; Riley et al., 2004).
One important aspect of morphogenetic fields is that, as modular units, they exist in specific locations within the embryo and, for adequate comparison of these fields in different animals, it is important to understand the overall organization plan and major axis (rostrocaudal, dorsoventral) of the embryo or organ/region containing the field. This allows one to know the correct location of the field according to the internal coordinates of the embryo/organ, including its relation to neighbors (topological position; Kuhlenbeck, 1978; Nieuwenhuys, 1974, 1998a, 1998b; Puelles and Medina, 2002). Analyzing the correct topological location of the field avoids wrong comparisons of fields sharing some of the same genetic programs and producing similar cell types, such as the sub-pallium (in the telencephalon) and the ventral thalamus (also called prethalamus; in the diencepha-lon) (Figure 1), both of which express homeobox genes of the Dlx/Distal-less family, and give rise to GABAergic neurons, but are located in completely different positions (Puelles and Rubenstein, 1993; Bulfone et al., 1993; Puelles et al., 2000; Brox et al., 2003). Comparisons of field topological location between animals are useful and make sense when the compared animals, or the organs/regions where the fields are located, share a common organization plan (Kuhlenbeck, 1978; Nieuwenhuys, 1998b; Puelles and Medina, 2002). This is the case with vertebrates, and with the central nervous system of vertebrates (see Basic Nervous System Types: One or Many?, Evolution of the Deuterostome Central Nervous System: An Intercalation of Developmental Patterning Processes with Cellular Specification Processes, Origins of the Chordate Central Nervous System: Insights from Hemichordates).
Another important aspect of morphogenetic fields is that, although highly stable, they are dynamic entities. They exist at a particular place within the embryo, and at a particular time, and their internal features (given by the lower-order modules within them, such as genetic cascades and networks and cells) and interactions with other fields change dynamically during development. At the end, each morphogenetic field gives rise to a particular set of derivatives, but homology of morphogenetic fields in two animals at a particular stage of development does not necessarily mean homology of all their
Figure 1 Schematic lateral views of the embryonic and adult brains of a mouse and a lamprey (a jawless fish close to the origin of vertebrates), showing some of the major divisions and subdivisions as understood nowadays based on available data, including data on expression of developmental regulatory genes (master control genes) during development (Pombal and Puelles, 1999; Murakami etal., 2001; Puelles and Rubenstein, 2003). During development, each division and subdivision is characterized by expression of a specific combination of master control genes, and constitutes a distinct self-regulated morphogenetic field, which gives rise to a particular set of derivatives in the adult. Each morphogenetic field occupies a specific topological position within the brain basic organization plan (or brain archetype). Morphogenetic fields of two animals showing comparable molecular features (similar expression of master control genes) and comparable topological position within the general brain archetype can be considered homologous (if the phylogenetic continuity applies, which can be analyzed using a cladistic method; see A History of Ideas in Evolutionary Neuroscience). For instance, using this analysis, the alar regions of prosomere 2 (p2) of mouse and lamprey, giving rise to the thalamus (Th, or dorsal thalamus), are field-homologous. Also, the basal regions of the mesencephalon (m) of mouse and lamprey are homologous as a field, and the oculomotor neurons (Om, motor neurons of the third cranial nerve) derived from this field are homologous as well. Other cell groups found in adult basal mesencephalon may not be homologous between mouse and lamprey (or between mouse and another vertebrates), since divergent evolution may have led to the appearance of novel cell groups. In the telencephalon (colored in the schematics), two major homologous morphogenetic fields exist in the embryonic brain of both mouse and lamprey, called pallium (blue) and subpallium (orange-red), showing distinct gene expression (Pax6 and Emx1 in the pallium versus Dlx in the subpallium). In addition, mouse (as well as chicks, frogs, and zebra fish) show a smaller expression domain of the homeobox gene Nkx2.1 within the subpallium, allowing subdivisions of this field into two smaller morphogenetic fields, called striatal field (orange) and pallidal field (red), giving rise to the striatum and pallidum, respectively, in the adult. However, lampreys do not show expression of Nkx2.1 in the telencephalon (Murakami etal., 2001; this gene is also previously called TTF-1), suggesting that a pallidal morphogenetic field is absent in lampreys and possibly was absent in the origin of vertebrates. This agrees with the apparent lack of a pallidum in adult lampreys (Pombal et al., 1997a, 1997b). This also suggests that the pallidal morphogenetic field was a novel acquisition during the transtition from jawless to jawed vertebrates. The appearance of a new field does not only imply evolution of a novel histogenetic division, but possibly triggers changes in adjacent or distant fields due to interfield (intermodular) interactions. In the case of the telencephalon, the appearance of a new pallidal field possibly triggered changes in the striatum and even the pallium, part of which was related to the existence of tangential cell migrations from the pallidum to those divisons, involving the acquisition of new cell types (integrated as interneurons of the striatum or pallium, colored in red in schematic) and new modulatory functions on the activity of native striatal and pallial cells. The striatal field also produces part of the interneurons found in the pallium in different vertebrates (colored in orange in the schematic), and this may also be true in lampreys (Pombal and Puelles, unpublished observations). a, alar region (dorsal); b, basal region (ventral); Hb, basal hypothalamus; m, mesencephalon; ob, olfactory bulb; Om, oculomotor neurons; p, pallium; pa, pallidum; Pt, pretectum; PTh, prethalamus or ventral thalamus; ps, secondary prosencephalon; p1, p2, p3, prosomeres 1, 2, or 3; s, subpallium; st, striatum; Th, thalamus or dorsal thalamus.
derivatives at the final stage, because some evolutionary divergence may have occurred during later development, producing novel cell groups and features. Abnormal changes in the internal order of the field or in the interaction with other fields (e.g., due to mutation of master control genes, or alteration in their regulatory region) can lead to field reorganization (to reach a new order) and changes in size, sometimes involving repatterning events. This occurs in the telencephalon after knockout mutation of homeobox patterning genes such as Pax6, Emx2, Gsh2, or Nkx2.1 (Sussel et al., 1999; Yun et al., 2001; Stoykova et al., 2000; Toresson et al., 2000; Bishop et al., 2002, 2003; Muzio et al., 2002). For example, knockout mutation of Nkx2.1 in mouse produces severe malformation and size reduction of the pallidal subdivision (medial ganglionic eminence or MGE), and repatterning of most of its former primordium to give rise to a larger striatal subdivision (lateral ganglionic eminence or LGE) (Sussel et al., 1999). In Emx2/Pax6 double mutant, the pallium is severely malformed and reduced, and its primordium is repatterned to become subpallium (Muzio et al., 2002). These examples indicate the importance of master control genes in the patterning and specification of morphogenetic fields, and show the self-regulatory and dynamic properties of the fields. These dynamic properties of morphogenetic fields are important for homology considerations. Thus, instead of saying that the resultant extra-large striatal subdivision found in Nkx2.1 knockout mice (or at least the part derived from the repatterning of MGE into LGE) is not homologous to the striatal subdivision of wild-type animals because of their different origin, we should consider the LGE of both mutant and wild-type mice as an attractor homologous state (in the sense explained above; see also Raff, 1996; Striedter, 1998), no matter how this was reached.
Was this article helpful?