- Top of page
- Materials and Methods
- LITERATURE CITED
- Supporting Information
Birds have a distally reduced, splinter-like fibula that is shorter than the tibia. In embryonic development, both skeletal elements start out with similar lengths. We examined molecular markers of cartilage differentiation in chicken embryos. We found that the distal end of the fibula expresses Indian hedgehog (IHH), undergoing terminal cartilage differentiation, and almost no Parathyroid-related protein (PTHrP), which is required to develop a proliferative growth plate (epiphysis). Reduction of the distal fibula may be influenced earlier by its close contact with the nearby fibulare, which strongly expresses PTHrP. The epiphysis-like fibulare however then separates from the fibula, which fails to maintain a distal growth plate, and fibular reduction ensues. Experimental downregulation of IHH signaling at a postmorphogenetic stage led to a tibia and fibula of equal length: The fibula is longer than in controls and fused to the fibulare, whereas the tibia is shorter and bent. We propose that the presence of a distal fibular epiphysis may constrain greater growth in the tibia. Accordingly, many Mesozoic birds show a fibula that has lost its distal epiphysis, but remains almost as long as the tibia, suggesting that loss of the fibulare preceded and allowed subsequent evolution of great fibulo–tibial disparity.
The shank (zeugopod) of most tetrapods has two equally long bones—the medial (inner) tibia and the lateral (outer) fibula. In early theropod dinosaurs, which are bird ancestors, both bones were equally long, although the fibula is more narrow and in close contact to the tibia. This condition was still present in the basal bird Archaeopteryx (Ostrom 1973; Mayr et al. 2005). Within the Pygostylia, closer to modern birds, the fibula became shorter than the tibia and splinter-like toward its distal end, no longer reaching the ankle (O'Connor et al. 2011a). In modern birds, the fibula is typically about two-thirds the length of the tibia, but fibulo–tibial proportions show considerable evolutionary variation, with proportionally shorter or longer fibulae in different species (Owen 1866; Streicher and Muller 1992).
Distal fibular reduction occurs after early embryonic patterning of the limb skeleton, at postmorphogenetic stages: bird embryos start out with a fibula as long as the tibia, which contacts the ankle (Heilmann 1926). Fibular reduction can be avoided through experimental manipulations. Placing a barrier between the early cell populations that are precursors of the tibia and fibula generates late chicken embryos with a dinosaur-like fibula, as long as the tibia (Hampe 1958, 1960aa,b; Müller 1989). Dinosaur-like fibulae have also been obtained by adding more mesenchyme to the limb bud (Wolff and Hampe 1954), placing grafts of Sonic Hedgehog (SHH) expressing cells (also known as “ZPA cells,” Archer et al. 1983), and by misexpression of the Hoxd-13 gene in the entire early limb bud (Goff and Tabin 1997). A recurrent explanation is that there is a competition for cells between the two early precursor populations of the tibia and fibula, in which the tibia normally prevails. Adding more cells or separating early precursors with a barrier conceivably allows greater cell allocation to the fibula, which grows to a larger size. An alternative hypothesis is that fibular reduction results from the secondary loss of the growth plate at its distal epiphysis (Archer et al. 1983). A developmental program intrinsic to the fibula was proposed to make its distal end break off, with the resulting fragment becoming the fibulare (the heel bone, also known as the calcaneum, which is found distal to the fibula in all tetrapods). Without a growth plate, the distal fibula would then be incapable of growth. However, embryological descriptions of birds and other tetrapods show the fibulare initiates cartilage formation independently and at an earlier stage, rather than breaking off from the cartilaginous fibula (Shubin and Alberch 1986; Müller and Alberch 1990). It has also been noted that experimental manipulations have a stronger effect on the length of the tibia. As the increase of the fibula is moderate by comparison, this is considered the main effect leading to a dinosaur-like phenotype, with a fibula as long as the tibia (Müller 1989; Goff and Tabin 1997). This has supported the view that fibular reduction is a side effect of competitive dominance and/or increased growth of the tibia, downplaying any role for the distal fibular epiphysis.
It is well known that lengthwise growth at the end of a long bone (the epiphysis) is possible as long as it retains a cartilaginous growth plate. Long bones across tetrapods (including crocodylians, bird's closest living relatives) follow a similar pattern of endochondral cartilage replacement by bone (Johnson 1933; Rieppel 1992, Rieppel 1993a,b,c, 1994; Reno et al. 2007; Mitgutsch et al. 2011; Diaz and Trainor 2015). Maturation of the chondrocytes and then ossification starts at the center of the cartilage (diaphysis) and expands toward both ends (epiphyses). Chondrocyte phenotypes change from rounded immature proliferating cells expressing collagen type 2a (Coll-II) and collagen type 9 (Coll-IX), to become flattened in parallel arrangements, and then terminally hypertrophic cells expressing collagen type 10 (Coll-X). Lengthwise growth is allowed by the persistence of immature, proliferating chondrocytes at the ends of the epiphyses (growth plates; Kuhn et al. 1996; Wilsman et al. 1996; Farnum et al. 2007), and ceases when chondrocyte differentiation and ossification progress into them. This general pattern is conserved despite some differences among clades, such as the presence of secondary epiphyseal ossifications in mammals and lizards (Johnson 1933; Fabrezi et al. 2007).
Recently, important advances have been made in understanding the molecular mechanisms behind these processes (Crombrugghe et al. 2001; Kronenberg 2003; Kobayashi and Kronenberg 2014; Kozhemyakina et al. 2015). The secreted protein Indian hedgehog (IHH), produced by prehypertrophic chondrocytes, was discovered to play an important role in each aspect of endochondral ossification (Vortkamp et al. 1996; Kronenberg et al. 1997; Karp et al. 2000; Kobayashi et al. 2002; Long et al. 2004). It stimulates immature chondrocytes transition to flattened and then hypertrophic chondrocytes, favoring the replacement of cartilage by bone in the diaphysis. IHH also stimulates the production of Parathyroid hormone-related protein (PTHrP) by periarticular chondrocytes, a secreted protein that delays the differentiation of immature chondrocytes in the epiphysis. PTHrP in turn inhibits the production of IHH, generating a negative feedback loop that maintains the coordinated differentiation of immature chondrocytes in the growth plate (Vortkamp et al. 1996; Kronenberg et al. 1997; Karp et al. 2000; see Fig. 4E). Genetic inactivation of PTHrP causes premature ossification and decreased growth of the skeleton (Karaplis et al. 1994; Lanske et al. 1996); Conversely, misexpression of PTHrP in the whole cartilage delays chondrocyte differentiation and impairs ossification (Weir et al. 1996). This PTHrP-IHH feedback system has been comprehensively studied in both chicken and the mouse (e.g., Inada et al. 1999; Long et al. 2001a,b; Ueta et al. 2001; Mak et al. 2006, 2008; Ruiz-Perez et al. 2007). The mouse, like any mammal, is an outgroup to chicken that phylogenetically is maximally distant within the amniotes. This suggests these molecular mechanisms are highly conserved, and can be expected to be present in other reptiles, including crocodylians, which are closer to birds (Reno et al. 2007).
To better understand the development of the fibula, we studied the timing of reduction and ossification in several orders of birds. In the chicken, we studied the molecular development of cartilage maturation and growth plate formation in the tibia and fibula. We also examined the onset of the IHH-PTHrP feedback system in the fibula, including experimental downregulation of IHH signaling. We show that the distal epiphysis of the fibula presents an abnormal molecular profile when compared to other bones, and that a normal growth plate fails to be established. We also examined key specimens of Mesozoic birds documenting the earliest stages in the evolution of fibular reduction. The fossil evidence shows that disruption of the distal epiphysis was an early event, and may have played a key role for subsequent increases in fibular reduction.
- Top of page
- Materials and Methods
- LITERATURE CITED
- Supporting Information
The hypothesis of mesenchymal competition suggests that interactions between the early precursor cell populations of the fibula and the tibia determine a “stunted” condensation of the fibula, and that these interactions may also explain variation among modern birds, some of which show greater fibular reduction than others (Hampe b; Streicher and Muller 1992). A minimum cell number is necessary to form a mesenchymal condensation, and experimental reductions in early limb bud cell number can cause the complete loss of the smaller element in the zeugopod (Wolff and Hampe 1954; Towers et al. 2008). However, the mechanisms that actually achieve fibular reduction in birds only come into effect long after the early patterning of condensations, at late, postmorphogenetic stages. In all bird species we observed, the embryonic fibula remains equal to the tibia near HH31, even when the fibula is drastically reduced, as in the budgerigar (Fig. S2). The hypothesis that the distal end of the fibula breaks off (Archer et al. 1983), although mistaken in structural terms, is correct in its functional implications, because the fibula fails to develop a distal growth plate: the undifferentiated population of cells at the distal fibula become terminally differentiated between HH31 and HH36. In contrast, proper epiphyseal growth plates are established at its proximal end, and at both ends of the tibia. As pointed out above, at HH29 of the chicken, the fibulare becomes closely appressed to the distal fibula. The molecular profile of the fibula suggests that, during this transient contact, the fibulare may act as a functional extension of the fibula or “surrogate epiphysis,” which may allow IHH expression to progress further into the actual distal end of the fibula. As the fibulare then separates from the fibula, the surrogate epiphysis is lost, but cartilage maturation still progresses, impairing the proper development of a growth plate at the distal fibula. This could explain why the bone collar of the fibula is off-center and distally displaced during its early ossification in all bird species observed (Fig. S4).
Cyclopamine downregulates IHH signaling, and in treated legs it may delay the advance of cartilage maturation, as well as impairing the formation of an interzone between the fibula and the fibulare (Mak et al. 2006; Später et al. 2006; Gao et al. 2009; Ray et al. 2015). Fused to the fibulare, the fibula now retains its surrogate epiphysis. Importantly, the fibulare continues to express PTHrP, and no IHH, while maturation is well underway in the tibia and fibula. This is not surprising because in all tetrapods, tarsal elements ossify much later than other limb bones (Rieppel 1992; Caldwell 2002; Fröbisch 2008). The sustained expression of PTHrP at the fibulare may negatively regulate IHH, allowing the maintenance of immature chondrocytes and distal growth at the fibula until later stages. It is worth noting that in all experiments in which the fibula is as long as the tibia, the fibulare is fused to its distal end, despite radically different procedures (mesenchyme grafts, barrier insertions, Shh-expressing cells grafts, Hox misexpression, and Cyclopamine treatment; Hampe 1958, 1960a,b; Archer et al. 1983; Müller 1989; Goff and Tabin 1997). Although different mechanisms somehow impede separation of fibulare and fibula, retention of the “surrogate epiphysis” may explain sustained growth of the fibula in all cases.
During evolution, in the advanced mesotarsal ankle of the basal ornithodira (the earliest bird ancestors to show upright, bipedal locomotion) the fibulare became closely locked into the fibula, with little or no movement between these bones. This greater adult proximity between proximal tarsals and the zeugopod is in fact reflected in early embryonic patterns: The tibiale and the fibulare of birds form very close to the tibia and fibula, in contrast with Alligator and other amniotes in which they are more separate, and movement of the foot involves rotation at their articulation (Fig. S3). The close embryonic proximity of the fibula and fibulare in avian embryos may have been common to all Ornithodira, allowing “surrogation” and loss of the distal fibular epiphysis in those clades in which differentiation of the fibula started before its separation from the fibulare. In fact, besides birds, distal fibular reduction also occurred independently within at least three other lineages of Ornithodira: Alvarezsauridae (Chiappe et al. 2002), Oviraptorosauria (Vickers-Rich et al. 2002), and Pterosauria (Dalla Vecchia 2003; Bonaparte et al. 2010; Fig. 7).
It has been previously suggested that development of a smaller fibula occurs as a mere result of greater growth of the tibia (Müller 1989). Accordingly, it has been proposed that among modern birds, the evolution of a strongly reduced fibula has no adaptive value in itself, but is a developmental by-product of selection for a longer tibia in specialized lifestyles such as wading (Streicher and Muller 1992). However, besides long-legged waders (flamingos, storks, and herons), among modern birds the relatively smallest fibulae (less than 50% the length of the tibia) have also evolved in small, short-legged birds (swifts, passerines, and kingfishers). The presence of strongly reduced fibulae in these short-legged birds could be explained if small body size can also bring about greater fibular reduction (Streicher and Muller 1992). However, the alleged correlation between longer legs and reduced fibulae is also challenged by the fact that long-legged birds such as cursorial ratites (ostriches, emus) and birds of prey (eagles, owls) display some of the relatively largest fibulae (more than 75% the length of the tibia; Baur 1885; Shufeldt 1894; Hampe b; Streicher and Muller 1992).
Our own embryological observations in birds with different degrees of fibula reduction show that roughly adult proportions are already attained in the specific period or “window” between stages HH31-36, following failure of the fibula to maintain a distal growth plate. We infer this is when the tibia is most capable of outgrowing the fibula. Bird species with a greatly reduced fibula display greater growth of the tibia during this embryonic period, regardless of whether they are long- or short-legged as adults. These embryological differences among bird species are hard to interpret in adaptive terms. Thus, although we agree that fibular reduction is probably in itself nonadaptive, it is not easy to identify another directly adaptive trait that leads to it as a by-product. In developmental terms, our new data demonstrate that disruption of the fibular distal epiphysis occurs, which cannot be a mere side effect of greater growth at the tibia. Indeed, loss of the epiphysis may be required for greater growth of the tibia, such that, if the epiphysis is retained, growth in the tibia is significantly decreased.
In the context of our interpretation, we propose a two-step hypothesis for the evolution of fibular reduction in birds. The earliest birds to show fibular reduction have a splinter-like distal end that does not articulate the ankle, suggesting disruption of the growth plate. However, their fibula is still almost as long as the tibia (Fig. 7; Chiappe et al. 1999; Zhou and Zhang 2002; Wang et al. 2014). Increased reduction of the fibula only occurred thereafter and independently in several lineages of Mesozoic birds (including that leading to modern birds, Fig. 7) as an outcome of the evolution of greater growth of the tibia.
Previously, the loss of the distal fibular epiphysis was considered an alternative to greater embryonic growth of the tibia, as mutually exclusive hypotheses. In our view, greater growth in the tibia is important, but only during the embryonic period when the distal fibular epiphysis is prematurely ossifying. Like other recent work, our data stress the importance of postmorphogenetic changes for the origin of evolutionary novelties (Nagashima et al. 2009; Botelho et al. 2014b; Cooper et al. 2014), highlighting the role of spatial relations between skeletal elements at the time of the onset of endochondral ossification.