Vertebrate species have successfully colonized a wide diversity of habitats and ecological niches over the past 400+ million years. Many of the evolutionary innovations characteristic of vertebrates—such as the elaboration of the brain, neural crest, and placodes—involve the head (reviewed by Gans and Northcutt, 1983; Langille and Hall, 1989; Shimeld and Holland, 2000). Indeed, much of the success of vertebrates is due to the evolution and diversification of a distinctly organized “vertebrate” head (Depew et al., 2002aa; Gans and Northcutt, 1983; Northcutt, 2004). The elaboration of this head allowed gnathostomes (jawed vertebrates) to exploit new lifestyles based on active predation rather than sedentary filter feeding (Barghausen and Hopson, 1979; de Beer, 1985; Goodrich, 1958; Halstead, 1968; Moore, 1981; Radinsky, 1987; Romer, 1966). Much of the gnathostome evolutionary radiation was dependent upon the ability to sense and interpret the environment and subsequently act upon this information through utilization of a specialized mode of feeding involving the jaws. While most vertebrate adaptive transitions have involved a role for the skull (Hanken and Thorogood, 1993), it is apodictic that it has been the integrated constituent components of the head, i.e., the brain, sense organs, and skull—that has been at the heart of gnathostome innovations.
Notably, the gnathostome head, reflective of the vertebrate baüplan, is bilaterally symmetric with left and right halves essentially representing mirror images along the midline (Graham et al., 2010). Gnathostomes, however, exhibit two patterns of internal asymmetry: First, the internal organs are asymmetrically distributed within the abdominal cavity; and second, individual organs normally have a tissue-specific asymmetric morphology. With some important exceptions, gnathostome sense organs and skulls typically exhibit bilateral symmetry while the brain is internally asymmetric. Because each side of the medio-lateral axis develops independently (although ostensibly through identical developmental programs), minor left–right (L-R) asymmetries, referred to as fluctuating asymmetry (FA), occur in normal growth and development (Graham et al., 2010; Klingenberg and Nijhout, 1999; Leamy and Klingenberg, 2005; Møller and Swaddle, 1997; Van Valen, 1962). Additionally, overt morphological asymmetries, whether directional (DA) or antisymmetric (AS), of the external head have evolved in various lineages (see below). It is our intention, herein, not to provide a comprehensive review but rather to provide a basic primer on studies of the asymmetric development of the gnathostome head, in particular of the skull.
SYMMETRY AND ASYMMETRY
Whether it be radial, rotational, dihedral, translational, fractal, or bilateral, symmetry is ubiquitous in the living world (Graham et al., 2010; Palmer, 1996, 2009) and the establishment (and breaking) of symmetry are essential aspects of development (Levin, 2005; Palmer, 2004). While, as noted above, gnathostomes are bilaterally arranged organisms, breaking of this bilaterality through the establishment of asymmetric structures during development is typical. Understanding the origins and nature of these asymmetries has been, and continues to be, an active field of investigation, and because the subject has been reviewed in many places and under many umbrellas (e.g., Levin, 2005; Møller and Swaddle, 1997; Palmer, 1996, 2004; Van Valen, 1962), we discuss it only briefly here. With regard to bilaterality, L-R patterning requires the establishment of a midline, and the differential development of the two halves. L-R asymmetry follows from the establishment of antero-posterior (AP) and dorso-ventral (DV) axes within the embryo. Disruptions in AP or DV patterning lead to loss of L-R asymmetry. The positioning of organs is regulated by asymmetrically expressed genes (Whitman and Mercola, 2001). Initial L-R asymmetry is thought to be, in many gnathostomes, initiated by motile cilia beating in the extracellular space at the node (Levin, 2005; Marszalek et al., 1999; Nonaka et al., 1998).
With regard to morphological structures, FA, DA, and AS are three types of asymmetry that are generally recognized (Van Valen, 1962). FA refers to minor non-directional (i.e., random) deviations from perfect symmetry, measured as the differences between corresponding parts on the left (sinistral) and right (dextral) sides of the body. The mean value for L-R differences within a population is zero and conspicuous asymmetry is typically absent. The level of FA has often been used as a measure of the health of a population; with regard to cranial development, FA can be seen, for instance, in the asymmetries of the antlers of cervids and within the human skull (e.g., Ditchkoff and deFreese, 2010; Hoyme, 1994; Mateos et al., 2008; Mills and Peterson, 2013; Fig. 1a). DA and AS refer to two different types of obvious asymmetrical features within populations (Van Valen, 1962). In DA, morphological lateralization is fixed, occurring on the same side in essentially all individuals of a population. The species-specific dextral or sinestral asymmetric placement of the eyes (and the associated skeletal elements) of post-metamorphic flatfish and the coordinated rightward bent of the wrybill plover beak both epitomize DA in the cranium (Fig. 1b; Friedman, 2008; Graf and Baker, 1988; Gregory, 1933; Hubbs and Hubbs, 1945; Okada et al., 2001; Regan, 1910; Schreiber, 2006; 2013). Populations with random lateralization, with discrete asymmetries either to the left or to the right, exhibit AS. The asymmetric jaws of certain fish and the beaks of crossbills both evince AS (Fig. 1c; e.g., Benkman and Lindholm, 1991; Edalaar et al., 2005; Graham et al., 2010; Stewart and Albertson, 2010).
SYMMETRY AND DEVELOPMENT OF THE SKULL
To approach the origins of cranial skeletal asymmetries, it is informative to begin with a brief description of the development of the gnathostome skull. All gnathostome skulls share a conserved structural bauüplan consisting minimally of the chondrocranium (Barghausen and Hopson, 1979; de Beer, 1985; Depew and Compagnucci, 2008; Depew and Simpson, 2006; Goodrich, 1958; Moore, 1981). The chondrocranium has two components: the neurocranium and the splanchnocranium. The latter is comprised of the branchial arch (BA) derived palatoquadrate (PQ) and Meckel's cartilage (MC) cores of the upper and lower jaws, respectively (Fig. 2a). The neurocranium (which, as the name indicates, supports and protects the central and primary peripheral nervous system) is generated from the integration of multiple cartilaginous precursors: the parachordal and trabecular basal plates and the paired sensory capsules (nasal, optic, and otic). The chondrocranium has numerous possible ontogenic fates: endochondral ossification, direct investment by dermal bone, degeneration (resorption), quiescence as cartilage, or transdifferentiation (Barghausen and Hopson, 1979; de Beer, 1985; Depew and Simpson, 2006; Depew et al., 2002ab; Goodrich, 1958). The manifestation of the chondrocranium clearly is a conserved gnathostome feature. The advent of the nascent dentition and, in all but the cartilaginous fish, the dermatocranium (the intramembranous, dermal, elements that surround and develop in close association with the chondrocranium and dorsal brain) brings forth a second phase of skull ontogeny (Fig. 2b). Dermatocranial bones are typically classified topologically: Osteichthyeans are characterized by large numbers of dermal bones, whereas tetrapods are characterized by large-scale reductions in number (known as “Williston's Law”; Gregory, 1935). The syncranium is thus formed of the composite chondrocranium (and its derivatives) and the dermatocranium plus dentition. The skull of the gnathostome adult develops with the refined modeling and remodeling of these cranial elements to fit endpoint demands.
The cranial skeletal tissues of gnathostomes are developmentally derived from two mesenchymal cell populations, cranial neural crest (CNC), and cranial mesoderm (Hall, 1999; Kuratani, 2004, 2005; Le Douarin and Kalcheim, 1999). The CNC arises within the ectoderm of the neural folds, from which it delaminates and migrates into the developing head to target destinations according to their axial origins. Post-migratory CNC mesenchyme encounters distinct epithelia that influence its patterning and differentiation (Fig. 2c). In this regard, several signaling centers play critical roles in proper craniofacial development. One such signaling center, the frontonasal ectodermal zone (FEZ), regulates midfacial width by directing proximodistal outgrowth of the midfacial prominences (Hu and Marcucio, 2009; Hu et al., 2003; Marcucio et al., 2005). While the FEZ sets up aspects of midline patterning, other epithelial signaling systems, depicted in the “Hinge and Caps” model (Fig. 2c), are established bilaterally in the developing skull. This system regulates jaw polarity, modularity, and the functional integration of the lower jaw with the upper jaw, and this latter with the neurocranium (including the nasal and optic capsules; Compagnucci et al.,2013; Depew et al., 2002a,b; 2005; Depew and Compagnucci, 2008; Depew and Simpson, 2006; Fish et al., 2011) (Fig. 2d,e). Thus, bilateral symmetry in the skull is established in large part by the elaboration of paired epithelial signaling centers acting on the subjacent CNC and cephalic mesoderm (where these epithelial centers of patterning information include the placodes around which the primary sensory capsules develop, the FEZ, and those associated with the “Hinge and Caps”).
ADAPTIVE ASYMMETRY OF THE SKULL AND ASSOCIATED SOFT TISSUES
It is in the context of such (and other) patterning mechanisms that selective evolutionary pressures have come to generate adaptive cranial skeletal asymmetries. Indeed, adaptive asymmetries of the skull and associated soft tissues have evolved numerous times. Many of the most notable of these are associated with marine or aquatic gnathostomes. For instance, there are over 700 species (from 15 families) of Pleuronectiform flatfish (e.g., the “Heterosomata” such as the flounder, sole, halibut, turbot, plaice, and tonguefish) that are recognized by their distinct laterally compressed body morphology and unilateral eyes (Gregory, 1933; Hubbs and Hubbs, 1945; Okada et al., 2001; Schreiber, 2006, 2013). Born with bilaterally symmetric skulls, flatfish larvae undergo a dramatic metamorphosis into juvenile (immature adult) forms that exhibit lateralized swim postures, unilateral pigmentation, and extensive craniofacial remodeling—including the migration of one eye to the opposite side of the cranium—and associated nervous system changes (Bao et al., 2011; Friedman, 2008; Graf and Baker, 1983; Gregory, 1933; Fig. 3a). Metamorphosis in flatfish is regulated by thyroid hormone levels and subsequent downstream effectors of patterning and growth. Ostensibly underlying the genetic underpinnings of flatfish asymmetry, disparate flatfish species are typically laterally monomorphic (i.e., display DA) with the adults displaying either exclusively dextral (eyes unilaterally right) or sinistral (eyes unilaterally left) morphology. Asymmetric, post-nodal expression of pitx2 in the habenulae has been correlated with the neuronal architecture associated with the behavioral modifications attendant with this drastic remodeling of the position of the eyes and orbits (Schreiber, 2013; Suzuki et al., 2009). Whether differential pitx2 activity plays a role in the asymmetric cranial skeletal remodeling remains to be fully determined.
Pleuronectiformes are not the only osteichthyans evincing distinct asymmetries. Asymmetry (typically “recorded” as DA) in the jaw morphology and head inclination of certain cichlids (e.g., Julidochromis ornatus, Perissodus microlepis, and Neolamprologus moorii), the Japanese goby (Rhinogobius flumineus), and medaka (Oryzias latipes), as well as the zebrafish (Danio rerio), have all been reported (Hori et al., 2007; Kusche et al., 2012; Mboko et al., 1998; Seki et al., 2000; Stewart and Albertson, 2010). A number of studies on scale-eating fish have suggested that in these fish jaw asymmetry may derive from handed (lateralized) feeding behavior (Hori et al., 2007; Stewart and Albertson, 2010; Van Dooren et al., 2010). Growth (development) of the skull exhibits some degree of plasticity and responsiveness to functional use, and jaw asymmetry is most extreme in individual fish that display the most consistent lateralized feeding patterns (Van Dooren et al., 2010). Further, in contrast to the bimodal nature of jaw laterality in adults, juveniles that have yet to feed on scales show unimodal distribution of jaw sided-ness (Stewart and Albertson, 2010). These data suggest that differences in preference for the left or right flanks of prey influence jaw laterality once fish begin to feed on scales.
Behavior correlated cranial skeletal asymmetry has been well investigated in other marine gnathostomes, in particular with echolocation in odontocete cetaceans (toothed whales; e.g., Cranford et al., 1996, 2008; Fahlkea et al., 2011; Fordyce and Barnes, 1994; Huggenberger et al., 2009; Ketten, 1992; Klima, 1999; Marino, 2004; Marino et al., 2003; MacLeod et al., 2007; Mead, 1975; Miller, 1923; Moore, 1981; Ness, 1967; Racicot and Berta, 2013; Rauschmann et al., 2006; von Baer, 1826). Synapomorphic (shared derived) cranial characters common to all Cetacea include dorso-caudal nares, extensive peribullar and pterygoid sinuses, elongated (or extensively reconfigured) mandibles and maxillae, petrosal bones detached from the skull, and a foreshortened, concave cranial vault (Ketten, 1992). As initially described by Miller (1923), these characteristics reflect the “telescoping” of the cetacean skull which involved the evolutionary revamping of the cranial vault whereby the maxillae of the upper jaw expanded back to the vertex of the skull and covered the reduced frontal bones. Moreover, with this revamping the rostrum elongated and the cranial vault foreshortened, pulling the nares (blow-hole) and narial passages rearward to a superior position behind the eyes; this resulted is a frontally compressed, concave cranium with dorsal nares allowing ventilation with only the most dorsal surface of the head above water as epitomized by the delphinid skull (Fig. 3c–e). Asymmetry in odontodcete skulls, with their dorsal cranial bones shifted posteriorly and to the left, is associated with structural reorganization of the airways and is linked to high-frequency sound production and echolocation. Thus, living odontocetes typically have hypertrophied melons (specialized structures used in echolocation), nasal and pterygoid sinus, and phonic lips used to produce high-frequency biosonar. Odontodcetes generate their echolocation clicks (vocalizations) by a pneumatically driven process in their asymmetric nasal complexes and the asymmetry of the system has been implicated in the echolocation itself. Auditory associated asymmetries are not, however, unique to cetaceans; for instance, certain owls (such as Tengmalm's owl, Aegolius funereus) clearly display extensive DA in the cranial hard and soft tissues linked to their hearing (Fig. 3b; Collett, 1871; Norberg, 1978).
One of the most well recognized adaptive cranial asymmetries in a marine mammal is the left tusk of the male narwhal (Monodon monoceros) that typically extends significantly further forward than the right, which remains vestigial (Fig. 4c; Eales, 1950; Flower and Lydekker, 1891; Nweeia et al., 2012). Occasionally both tusks develop, though notably the chirality (i.e., the orientation of the dental spiraling) of both tusks is identical (sinestral) and is not enantiomorphic (Fig. 3c′; Kingsley and Ramsay, 1988). It is tempting to speculate that this state of identity originates from uniform chirality of the bilateral signaling system generating the tooth germs. Shh, a potent secreted signaling factor, is expressed in dental placodes, where its activity is essential for odontogenesis (Cobourne et al. 2004). Primary cilia, moreover, appear to guide this Shh-regulated activity (Ohazama et al., 2009). It is possible that the chirality of the narwhal tusk is related to the directional movement of the cilia associated with the dental placode; thus the identical chirality of both the right and left tusks, when they do form, may be due to the identical directional movement in both dextral and sinestral tooth germs of the cilia associated with Shh-regulated odontogenesis. In this regard, should sufficient numbers appear in the fossil record, it would be interesting to assay whether bilateral tusks ever appear of the otherwise unilaterally tusked Pliocene delphinid cetacean, Odobenocetops peruvianus, and if so whether the chirality of both tusks is also identical (de Muizon, 1993).
Cephalic asymmetries in gnathostomes are certainly not limited to the skull (or brain). While odontocete cetaceans are generally recognized as asymmetric, it has been believed that mysticete (baleen) whales fail to be so (outside of possible dextral-sinestral pigmentation differences). However, it has recently been recognized that the lunging (feeding) behavior of roqual whales is coordinated by an asymmetrically innervated neomorphic organ positioned at the midlines of the lower jaws of these whales (Pyenson et al., 2012). Additionally, and returning to marine mammals, the male hooded seal (Cystophora cristata) apparently expels its balloon through it sinsitral nostril (Fig. 4f; Bellard and Kovacs, 1995; Brønsted, 1932).
The above examples, and others such as the AS dental formulae in certain bats (Juste and Ibanez, 1993) or the asymmetric dentition of conodonts (Donoghue and Purnell, 1999), demonstrate that the developmental mechanisms patterning the cranial skeleton do not preclude adaptive asymmetries. Perhaps unsurprisingly (considering its central role in craniofacial patterning and development), many of these examples (e.g., the odontocete cetacean blow-hole, the narwhal tusk, unilateral orbits in flatfish, the balloon of the hooded seal) are associated with cranial patterning regulated at the λ-junction (where the maxillary first arch meets the frontonasal processes; Figure 2C). Recent characterization of a “Pbx-Wnt-p63-Irf6” regulatory module (Ferretti et al., 2011) that correlates in complexity with the gradual increase in complexity of the mxBA1–FNP connection at the λ-junction through evolutionary transitions from bony fish to mammals (e.g., with the evolutionary elaboration of the choanae, upper lips, and secondary palate) further highlights the notion of various selective pressures and adaptive responses being centered at this patterning center.
Due to the tremendous radiation and adaptation of gnathostome species, their skulls exhibit a wide array of often amazing phenotypes, including the deviations from bilateral symmetry described above. It has been speculated, however, that the cranial skeletal elements of the dermatocranium and chondrocranium have been under unequal selective pressures. Notably, the morphology of the elements in the chondrocranium is highly conserved throughout the vertebrate lineage. For example, the shape of Meckel's cartilage in birds and mammals is very similar; however, the mandibles of these two groups exhibit divergent morphology. In fact, the mandible in birds forms from the fusion of eight bones, whereas in mice and humans, it essentially forms from only two (the left and right dentary bones). This pattern also holds for other elements of the chondrocranium and dermatocranium. In a more extreme example, the neurocranium of sharks is quite similar in morphology to other vertebrates; however, sharks lack associated dermatocranial elements. These data suggest the hypothesis that the developmental programs regulating dermatocranial formation (perhaps osteoblast and osteoclast activity) are more plastic or less constrained than those regulating chondrocranial formation.
ABNORMAL ASYMMETRY OF THE SKULL
Adaptive asymmetries provide one source of information regarding what is possible in the morphogenesis of the cranium; another source is abnormal variation observed in human craniofacial anomalies or craniofacial asymmetry resulting from experimentally induced abnormal development. Asymmetry in human craniofacial anomalies is most frequently observed in patients with orofacial clefting (OFC), the most common congenital facial defect (Fig. 3g; Gorlin et al., 2001; McIntyre and Mossey, 2010; Neiswanger et al., 2002; Woolf and Gianas, 1976). The etiology of OFC is complex, where both genetic and environmental factors contribute (Bush and Jiang, 2012; Kohli and Kohli, 2012). Other craniofacial anomalies prone to asymmetry include Hemifacial Microsomia (HFM), a maldevelopment of the jaw and ear, and craniosynostosis, the premature fusion of one or more of the sutures in the skull vault (Cohen and MacLean, 2000; Johnson and Wilkie, 2011). Interestingly, unilateral coronal craniosynostosis is estimated to occur 4–7 times as often as bilateral craniosynostosis (Boulet et al., 2008; Wilkie et al., 2010). While this propensity for asymmetry in craniofacial malformation is often noted, it has not typically been a focus of most investigations into their etiology. In this regard, investigations of asymmetry derived from experimentally induced abnormal development have great potential to inform our understanding of developmental mechanisms underlying morphological asymmetry.
Two particularly relevant examples of experimentally induced asymmetry come from mice carrying mutations in the genes encoding for the transcription factor Dlx5 and the secreted signaling factor Fgf8 (Fig. 4a–d). In addition to being expressed in the CNC of the BA and in the otic placode, Dlx5 is expressed in the surface cephalic ectoderm (SCE) of the early embryo (Depew et al., 1999). Dlx5 continues to be expressed in the olfactory pit associated with the λ-junction at critical pharyngeal stages of development. Loss of function of Dlx5 in the mouse leads to DA of the nasal capsule (Fig. 4b) wherein the dextral capsules of 90% of Dlx5−/− mutants are severely hypoplastic (Depew et al., 1999). This is preceded by dextral hypoplasia of the λ-junction and associated frontonasal processes. Such DA in a murine mutant is uncommon, and the significance of this asymmetry remains unclear. One hypothesis, however, relates to chirality of the mouse itself. Early murine embryos undergo a process of “turning” in which the tissues of the future hind limbs and tail change orientation. This process also leads to chirality in the embryo as a whole, with 90% of embryos turning with the same dextral chirality. One consequence of turning is that at a crucial ontogenic patterning timepoint (E10–E10.5 in the mouse), the left hind limb bud typically lies juxtaposed to the right nasal pit and λ-junction (Fig. 4a), raising the possibility that in the mouse, Dlx5 in SCE acts to buffer the dextral λ-junction from potential interactions with the sinestral hind limb bud and whatever might be secreted there.
Like Dlx5, Fgf8 is expressed in the embryonic SCE and λ-junction (Fig. 4c; Griffin et al., 2013). When lacking Fgf8, mice fail to gastrulate, lack mesoderm, and die prior to craniofacial development (Meyers et al., 1998). Fgf8 hypomorphic murine mutants (Fgf8Δ/neo in Fig. 4d) exhibit asymmetric development of the nasal capsules as well as DA in the development of the neurocranial base and optic capsules (Fig. 4d; Griffin et al., 2013). Tissue-specific deletion of Fgf8 in the mouse oral ectoderm, moreover, results in a significant loss of jaw elements (Trumpp et al., 1999). Notably, asymmetry of the jaw is often seen in these mutants, and the left side appears to be more severely affected. DA in the skulls of Fgf8 hypomorphic zebrafish has also been observed (Albertson and Yelick, 2005). In the acerebellar (ace) mutant, where Fgf8 is mutated, craniofacial asymmetry occurs in about two-thirds of the embryos. These zebrafish mutants also display a loss of normal asymmetry in the brain (epithalamus) and viscera, which has been correlated with defective endoderm formation (Albertson and Yellick, 2005).
Notably, the presence of craniofacial asymmetry in zebrafish Fgf8 mutants is associated with the presence of Kupffer's vesicle (KV), the equivalent of the mouse node (Albertson and Yellick, 2005). When insufficient Fgf8 is present, no KV develops, and there is no lateralization of Nodal signaling to the lateral plate mesoderm, causing randomization of the viscera and epithalamus. When Fgf8 levels are high enough for KV development, normal left-sided patterning of the embryo ensures asymmetric development of the viscera and epithalamus. Essential to this normal brain patterning is a slight asymmetry in Fgf8 expression on the left side of the epithalamus (Regan et al. 2009). Potentially, this asymmetry in Fgf8 expression may be sufficient to affect endoderm development and expression of Fgf8 in the pharyngeal pouches, which subsequently influences craniofacial development. Such a mechanism would link developmental mechanisms regulating the normal L-R asymmetric development of the embryo with symmetric craniofacial development.
These data suggest that Fgf8 is an important regulator of proper (a)symmetric development of both the brain and the skull, but that it plays different roles at different times in tissue- and taxa-specific manners. Thus, regulation of Fgf8 may have been a particular target of selection mediating different asymmetries, perhaps through a conserved role in directing cell migration and/or survival. Evidence therefore suggests that the naturally occurring, asymmetric skeletal elements (in particular those associated with the λ-junction and jaws) of the gnathostomes mentioned above are regulated in their normal morphology by Fgf signaling.
ASYMMETRY, DEVELOPMENTAL STABILITY, AND PLASTICITY
The fact that craniofacial anomalies are among the most common human birth defects suggests that craniofacial development is particularly susceptible to perturbation. The propensity for these defects to be asymmetric further highlights the variability in morphogenesis produced by craniofacial developmental pathways. Typically, bilateral asymmetries in normal populations are minor because developmental mechanisms have evolved to buffer environmental and genetic variation (Waddington, 1942). However, the skull appears to be quite susceptible to developmental noise, which generates deviations from the idealized developmental baüplan. This susceptibility to developmental perturbation is not only apparent in measures of FA, but may also contribute to incidences of AS, and perhaps provide some explanation of its developmental origins.
Among gnathostomes as a whole, where AS populations of obvious dextral and sinestral phenotypic traits have been studied, the direction of asymmetry is not inherited (Palmer, 2004, 2009). The absence of a heritable basis to the direction of asymmetry has generated many hypotheses about developmental processes mediating this variation. In particular, Palmer (2009) has suggested that in the case of heart asymmetry, there are two alleles, one for left and one for random. But how is “random” conveyed in the genome? Or is it that some developmental processes are particularly responsive to environmental fluctuation, which then serves as the mediator of randomness in phenotypes?
One mechanism mediating this variation may involve a binary switch mechanism for gene activation deriving from a nonlinear relationship between protein concentration and network activation (Moczek and Nijhout, 2003; Nijhout, 1999). In this model, extreme (high or low) protein concentrations produce robust phenotypes with minimal variation because the network will be either “on” or “off.” However, when protein concentrations are near the threshold for pathway activation, small differences in concentration produce a wide range of phenotypic variation (Hallgrimsson et al., 2009; Lai et al., 2004; Williams et al., 1999; Young et al., 2010). Evidence from mouse genetics suggests that this threshold level is typically below the protein concentration levels produced in heterozygous individuals (50% gene dosage), as most mouse mutant lines are viable and normal in the heterozygous state.
However, not all genes are equally important in maintaining developmental robustness. Gene products known to contribute to robustness are called phenotypic capacitors (Rutherford and Lindquist, 1998). When a phenotypic capacitor is challenged by a harsh environment or mutation, the developmental system becomes less robust and produces greater variation (Levy and Seigal, 2008). One example relevant to craniofacial development is Satb2-mediated variation in jaw length.
Satb2 is a transcription factor expressed in the mesenchyme of the developing upper and lower jaws, where it regulates the size of distal jaw elements (Britanova et al., 2006; Fish et al., 2011). In humans, mutations in SATB2 cause cleft palate (FitzPatrick et al., 2003). Mice lacking Satb2 also die of cleft palate, whereas Satb2+/− heterozygotes exhibit a variable reduction in jaw size that encompasses the entire range of size variation between wild-type and mutant populations (Fish et al., 2011). These data, combined with comparative analyses of Satb2 expression in diverse vertebrate taxa suggest a model where evolutionary modification of the size of the Satb2 expression domain was important to the evolution of distal jaw size (Fish et al., 2011).
Adult Satb2+/− heterozygotes exhibit mild to severe asymmetry of the jaw that is random in laterality (AS; Fish et al., 2011). Asymmetry of this nature is maladaptive, and Satb2+/− individuals would not survive in the wild. Thus, while variation in Satb2 expression may have been subject to selection during evolution, variation in Satb2 during development leads to dysmorphogenesis. Satb2 is a transcription factor that binds to matrix attachment region DNA elements, thereby augmenting the potential for enhancers to act over large distances and regulate multiple genes in concert (Britanova et al., 2006). As such, Satb2 may promote developmental robustness by orchestrating the co-regulation of genes within a particular tissue-specific pathway during normal development, while also enhancing evolvability by having a high threshold for pathway activation, thus making it more susceptible to developmental stress. The AS exhibited by Satb2+/− heterozygotes may therefore simply be an extreme form of FA.
Satb2 is just one of many genes for which craniofacial development has been shown to be dosage sensitive. These phenomena indicate that craniofacial development is very plastic. This is in contrast to heart development, which exhibits declining frequency of reversals in asymmetry, suggesting that its development has become more canalized (Palmer, 2004). The propensity for dysmorphology, therefore, may be an unfortunate consequence of the developmental plasticity underlying the evolvability and adaptability of the craniofacial complex.
The vast divergence in morphology and function of the vertebrate skull can be attributed to its underlying baüplan, which allows for significant evolvability within the overall conserved developmental framework (Depew and Compagnucci, 2008; Depew and Simpson, 2006; Fish et al., 2011). A particularly important component of the developmental pathways underlying the craniofacial baüplan is the λ-junction. The complexity of polygenic interactions involved in integrating signals at the λ-junction have been involved in many important vertebrate evolutionary transitions, as reflected in the elaboration of skeletal elements associated with the dermatocranium. A consequence of this complexity and plasticity in development is the susceptibility to perturbation, which leads to a high incidence of craniofacial malformations.