Electronic Supplementary Material for Davis, Brakora, and Lee, “Evolution of ruminant headgear: a review.”
EXPANDED REVIEW OF HISTOGENESIS OF RUMINANT HEADGEAR
Here, we enter into a much more detailed review of the state of knowledge of the development of ruminant headgear. Figure citations refer to those in the main paper.
Antlers
Antlers are unique among mammalian appendages in their ability to completely and periodically regenerate in adults; this fact has prompted more study of antlers than of any other headgear type [1-2], and the process of regeneration has been reviewed extensively [2-4]. For comparison with other ruminant headgear types, ourreview summarises the state of knowledge of the histogenesis and morphogenesis of primary (first-year) antlers found in red deer (Cervus elaphus) [2, 4-6], fallow deer (Dama dama) [7], white-tailed deer (Odocoileus virginianus) [8], and reindeer (Rangifer tarandus) [9].
An antler is bony projection from the lateral crest of the frontal bone and consists of a core of cancellous (spongy) bone surrounded by a sleeve of compact (dense) bone. Externally during growth and prior to rutting, an antler is covered by skin and subcutaneous loose connective tissue (SLCT). Both internal and external components of an antler experience coordinated changes during growth, so a review of antler growth must treat the bone and skin as an integrated structure.
Growth is divided into three stages. First, the stage of intramembranous ossification produces the initial bony outgrowth from the lateral crest to form a palpable bump (< 10 mm in height) called a pedicle[4-5]. This pedicle serves as the base for the primary as well as subsequent generations of antlers. As the name of this stage suggests, periosteal deposition of cancellous bone occurs by intramembranous ossification (i.e., ossification directly from a membrane, in this case the periosteum). During this stage, the developing pedicle is covered by the same type of skin and SLCT that covers the rest of the skull[4, 6]. The histology of the skin consists of hair follicles, arrector pilli muscles, sweat glands, and mono-lobed sebaceous glands. Although richly vascularised, the SLCT is generally unremarkable [6].
Second, the stage of transitional ossification furthers pedicle elongation to the point where it is clearly visible (approximately 25–40 mm in height). As the name of the stage suggests, there is a transformation in the mode of ossification demonstrating that the mode need not be strictly intramembranous or endochondral as described in introductory textbooks. In fact, similar transformations occur in the mandibular condyle [e.g., 10], bone fracture repair [e.g., 11], and solitary osteochondroma formation [12]. In the growing pedicle, the apical-most periosteum partially transforms into perichondrium. Consequently, osteoblasts as well as chondroblasts form in the apical region of the growing pedicle. Apical osteoblasts deposit bony spicules (via intramembranous ossification) that encase clusters of chondroblasts (now chondrocytes) to form bony trabeculae with cores of cartilage[4-5]. Rapid apical growth and the differential ability of the vascular supply to keep pace with pedicle growth is the likely explanation for the clustering of cartilage within bone. Where vascular formation is unable to keep pace with rapid pedicle growth, cartilage is deposited. Conversely, where vascular formation is able to maintain pace with pedicle growth, bone is deposited[5]. In addition to this transitional ossification, typical intramembranous ossification occurs along the sides of the pedicle and progressively thickens the peripheral bony sleeve. Marked changes also occur in the skin and underlying SLCT. The apical-most skin of the pedicle thickens, sweat glands accumulate, and the sebaceous glands enlarge. Compression caused by rapid elongation of the pedicle somewhat flattens the normal undulatory interface between the epidermis and dermis (i.e., the rete apparatus) and compacts the underlying SLCT [6]. The coordinated transformation of the skin/SLCT and periosteum/perichondrium suggests the presence of molecular signalling between different tissue types (i.e., heterotypic signalling). In fact, pedicle elongation (and antler formation) can be completely arrested if an impermeable membrane is inserted between the SLCT and the periosteum before the transitional stage of ossification begins [13].
Third, the stage of endochondral ossification involves the completion of the pedicle and the establishment of a base from which rapid growth of an antler proceeds. Rapid apical growth exceeds the ability of the rich vascular network to keep pace and causes the apical-most periosteum to completely transform into perichondrium. Elongated columns of cartilage, separated by highly vascularised spaces, are deposited on top of the older osseocartilaginous tissue. As more cartilage is deposited, the oldest and deepest cartilaginous tissues undergo chondroclasia (i.e., the removal of mineralised cartilage by phagocytic cells) and are remodelled into bony tissue much in the same way that endochondral ossification proceeds in long bones. The point at which growth transitions from pedicle to antler is not obvious when looking at the internal components because endochondral ossification characterises both the late apical growth of a pedicle and the entire apical growth of an antler. However, antler skin and pedicle skin differ substantially. Unlike the skin covering a pedicle, the skin covering an antler contains hair follicles that lack arrector pilli muscles and are connected to extremely large bi- to multi-lobar sebaceous glands. This velvet, as the new skin is called, lacks sweat glands and has thickened to the point where the once-undulatory rete apparatus is completely flattened. In addition, the underlying SLCT is flattened into a thin layer, merging almost completely with the periochondrium.
Induction of pedicle and antler formation is the result of a complex interplay of signalling molecules. As introduced above, the pedicle originatesas an intramembranous outgrowth from the lateral crest [5, 14]. The periosteum of the lateral crest is unusual because it contains glycogen-rich, embryonic-like cells [15] that appear predetermined to form lateral crests; transplantation of that periosteum elsewhere on the body induces an ectopic crest-pedicle-antler complex [15]. In addition, the formation of a lateral crest occurs in the absence of any apparentsignalling induction from the skin; the presence of an impermeable membrane between the periosteum and skin does not prevent its formation[13].
Initiation of pedicle growth from each lateral crest closely coincides with male puberty in most cervids (two exceptions are Hydropotes, which is antler-less, and Rangifer, in which both sexes develop pedicles and antlers at puberty [16]) and an increase in circulating levels of testosterone [17-18]. However, bone growth is more directly regulated by oestradiol, which is converted from testosterone by osteoblasts [19]. Oestradiol suppresses the RANKL-RANK system of bone resorption [20] and is clearly important in antler development: oestradiol is concentrated in antlerogenic and neighbouring tissues [21], promotes pedicle growth in males [22], induces premature mineralisation of antlers and shedding of velvet [23], and regulates the antler cycle in female reindeer [24].
In addition to androgenic signals, signalling between the apical periosteum and the skin is required to initiate and modulate the growth of pedicles (and antlers)[13]. Presumably, the close contact between the apical periosteumand skin promotes transit of molecules that are essential to the initiation ofpedicle formation and antlerogenesis[13]. For example, pedicle formation and antlerogenesis is completely arrested if an impermeable membrane is placed between the apical periosteum and the skin before the stage of transitional ossification. Once the stage of transitional ossification has begun, however, the impermeable membrane no longer prevents the completion of a pedicle and antlerogenesis, although longitudinal growth is retarded [13].
Termination of endochondral ossification and longitudinal growth coincides with a large pulse in circulating androgens and the rutting season [4]. Velvet (but not pedicle skin) and most of the antlerogenic tissue are shed, which exposes the bare bone of the antler. The superficial bone tissue likely dies; however, deeper bone tissue remains alive, and continued intramembranous ossification forms lamellar bone within antlers [25]. That a naked antler still contains living bony tissue is likely the reason the antler is not immediately cast; osteoblasts within the antler convert circulating testosterone into oestradiol, which locally suppresses bone resorption [20]. The inhibitory effect of oestradiol on bone resorption weakens with the seasonal decline in testosterone [23] or by the death of osteoblasts in the antler. Consequently, a large number of osteoclasts are recruited. Interestingly, their resorptive activity is not systemic but is limited to the pedicle-antler junction, leading to antler casting. How this focused bone resorption is directed is currently unknown, and future studies to understand the mechanism will be valuable in controlling degenerative bone diseases (e.g., osteoporosis) in other species.
Horns
Bovid horns are composed of a scabbard-like keratinous sheath covering a bony horncore, neither of which are shed [26-27]. The bony horncore joins seamlessly to the frontal bone via a constricted neck at the base of the horncore. What little is known of early horn development is obscured by inconsistent identification and naming of primordial horn structures. Further confusion has resulted from the inclusion of animals with scurs, also called“loose horns”: incompletely developed horns that have a solid bony core but only a soft tissue connection to the skull. Scurs may arise genetically or via pathology, and they exist in a range of severities. In this review, we wish to establish language to distinguish between three structures of the early horn. In ontogenetic order, these are 1) the soft-tissue anlage, which precedes the development of visible horns, 2) the os cornu, a loose, palpable nodule, and 3) the horncore bud, the macroscopic bony bump seamlessly fused to the frontal bone. Future revisions to terminology are likely as additional data become available. More detailed reviews may be found in Janis and Scott [26] and Dove [27].
Before birth, above the presumptive horn sites on the frontal bone, the epidermal and dermal components of the future hornacquire their potentials and irreversibly differentiate, evidently precluding post-natal signalling between these tissues [27].The anlage is positioned within, and differentiates from, both the dermis and SLCT above the periosteum [27], and it retains its connective tissue character. The anlage seems to be the primary inducer of horncore growth, although the mechanism(s) of induction are, to our knowledge, entirely unknown.
Despite its name, the relationship of the os cornu to the adult horncore is not at all straightforward. Whether it arises from the anlage, or becomes the horncore bud, or has a separate role (if any) is unclear. Several authors conclude that the os cornu does not exist as a discrete structure (if at all) during normal horn development. Rather, they posit that the os cornu only appears in animals heterozygous for genes controlling horn presence [i.e. in scurred animals; 27]), animals in poor health [28-29], or from manipulation [30], including surgery [27]. Nevertheless, many authors proceed to describe early horn development in terms of the os cornu.
The os cornu is generally thought to be a palpable nodule that begins in the supra-periosteal tissue. It has been reported to be made of dermis and/or SLCT [27, 31-32] or cartilage [33-34], and may ossify independently in the soft tissue (forming a scur), or after attaching to the frontal bone (forming a normal horn)[27]. Janis and Scott [26] persuasively argue for intramembranous ossification throughout horn development, and Durst [31] and Brandt [32] showed that there is no cartilaginous preformation of the horncore, in contrast to earlier reports [24, 25]. Other reports of cartilage in or around the horncore or frontal bone [33-35] are unexplained in this framework.
Summarising Dove [27], the os cornu does not ossify prior to fusion to the frontal bone. If there is no SLCT between the os cornu and the periosteum, it fuses through the periosteum to the presumptive horn site. Removing or doubling the periosteum has no effect. Whether embedding and fusion of the os cornu to the frontal bone are simultaneous is unknown, although macroscopic evidence of fusion is very rare. Ganey et al. [36]depicted a bone disc embedded in the frontal bone, separated by a thin layer of connective tissue in a two-thirds term bovine foetus; its genotype for horns was not reported. Janis and Scott [26, 31] stated that the first step involves a connection between the SLCT and the osteoid of the superficial frontal bone, followed by thickening of the osteoid soon after birth. In scurs, it appears that the union between the os cornu and the frontal bone is weak or absent; evidently the os cornu cannot penetrate the supra-periosteal SLCT and ossifies in place, becoming a scur [27]. However, the fate of the os cornu in normal vs. scurred animals has not been directly studied.
After induction by the anlage and/or fusion of the os cornu to the frontal bone, the horncore bud begins to develop. By the time the horncore bud is observable, its microstructure differs from the frontal [31]; Dove [27]concluded that the dermal portion of the os cornu becomes the tip of the horncore bud and that the SLCT portion of the os cornu forms the eventual neck of the horn. (The neck has also been called the pedicle; however, we prefer a distinct term to aid discussion and to avoid assumptions of homology with the cervid pedicle.)
Further growth of the horncore bone tissue is appositional at both the tip and the surface [26, 31, 37-38]. As horncore growth slows, deposition of compact bone proceeds simultaneously from surface to lumen and base to tip, often continuing into adulthood. Alternatively, depending on the clade, sex, and age of an animal, the horncore may be invaded by the frontal sinus [39], and the horncores of some taxa may be very thin-walled. Horncores remain living organs throughout life and are actively remodelled to accommodate sheath shape and physiological demands [40]. No evidence of signalling or other interaction between the frontal bones and the horncore periosteum has been reported for any stage of development; unlike antlers, the frontalsappear to neither hinder nor accelerate the production of any horn part [27].
Nevertheless, there is wide disagreement whether the bovid horncore is apophyseal (i.e., a direct outgrowth), epiphyseal (i.e., separated at least initially by non-bony tissue), or a combination, with respect to the frontal bone [17, 26-27, 31, 35, 41-43]. Hypotheses about os cornu homology are problematic because of imprecise definitions of the os cornu and the dependence of its structure and fate on genotype and environmental conditions. Resolving these hypotheses will require establishing the relationships among the anlage, os cornu, horn bud, neck, and horncore in normal horns.
The skin over the presumptive horn site on the frontal bone likely gains the capacity to form horn tissue (i.e., the keratin sheath) before birth, although normal sheath growth is dependent upon the presence of the os cornu[27]. Horn tissue is continuously produced by the epithelium covering the horncore [26, 31], with newer layers pushing the older layers distally like a stack of keratin cones (Fig. 2b); this layeringis more easily detected in temperate species with strong seasonal fluctuations in sheath growth rates [17, 44]. Nevertheless, the sheath tip is usually much thicker proximodistally than the walls are mediolaterally[17], suggesting more rapid production of keratin at the tip. Conventional wisdom that the horn sheath growsonly from the proximal end of the horn is therefore unfounded [26]. The horn tissue of juveniles is softer and more fibrous [45] and may “exfoliate” before adulthood [31, 38, 43, 45-47], exposing harder and more completely keratinised horn tissue. Distinctive horn shapes are thought to arise by modulating zones of keratin production in the skin surrounding the horn [1, 48], but experimental studies are lacking.
The development of polled (hornless) breeds of domesticated bovids has enabled partial identification of the genetic basis of horns. In cattle, the presence or absence of horns (or scurs) is genetically determined by three or four genes and (in scurs) by sex; alleles for horns are recessive [49-50], and heterozygotes may be scurred [21]. In goats, genetic hornlessness is tightly linked to intersexuality and sterility [51]. In cattle, horn-forming tissues have down-regulated genes coding for cadherin junction elements (i.e., cell membrane structures used in cellular adhesion) and epidermal development. Compared to genetically hornless animals, those with scurs have higher expression of genes involved in extracellular matrix remodelling [52].
Although no comprehensive picture of signalling pathways triggering or regulating horn development is available for any one species, evidence so far indicates that the sensitivity of horn-producing soft tissues to various exogenous and endogenous molecules changes rapidly after birth. For example, horn development can be fully blocked at Day 2 in calves, but only partially blocked at Day 4 using the same substance[53]. In adult mouflons (Ovis gmelini, a temperate caprine),summer increases in prolactin concentration are positively correlated with horn growth in adults [44, 54]; similarly, in domestic sheep, increased melatonin secretion (from shorter photoperiod in the winter) suppresses prolactin[55-56] and increases gonadotropin secretion [57]. Yet in mouflon lambs, horn growth is insensitive to melatonin [58], and in subadults,melatonin concentration does not correlate with prolactin concentration[44, 54]. Plasma testosterone concentrations are inversely correlated with horn growth both seasonally and with age in male mouflons[54, 59]. Across the bovid family, though, the male phenotype is associated with increased expression of horns (earlier and faster growth, greater size and symmetry, etc.) [60]. Castration experiments in cattle show that testosterone is important for development of normal male horns, although castrated males do not have female-like horns [61]. All this suggests that the relationship among various hormones and horn growth is complex and changes with maturation, but too little is known to infer common signalling pathways among species.