A firm understanding of the ontogeny of the lateral wall of the nasal capsule (LNC, also termed the paries nasi) elucidates two different phases of paranasal sinus development. Primary pneumatization is the process by which paranasal spaces extend outward from the nasal fossa as recesses (Sperber, 2000). Recesses (the primordial sinuses) develop into “true” sinuses by undergoing secondary pneumatization, or expansion beyond the borders of the nasal capsule, into facial bones (Weiglein, 1999; Maier, 2000; Rossie, 2006; Smith et al., 2011). Using osteological evidence alone, secondary pneumatization can be inferred using cross-age series of mammals (e.g., Rossie, 2006). Paranasal sinuses form by actively excavating bone along paranasal recesses (Witmer, 1999; Smith et al., 2005, 2011). More direct evidence is seen histologically. For example, breakdown of nasal capsule cartilage that delimits a paranasal recess is considered a necessary prelude to secondary pneumatization (Wang et al., 1994; Maier, 2000; Rossie, 2006; Smith et al., 2008). In a previous report, Smith et al. (2008) documented the fate of the nasal capsule cartilage in perinatal tamarins (Saguinus geoffroyi). In newborn S. geoffroyi, the nasal capsule is partially fragmented (anteriorly) while the remainder is in various stages of endochondral replacement by bone (posteriorly). Examination of an age series of tamarins indicates that the breakdown or replacement of nasal capsular cartilage is a relatively rapid process, extending from late fetal to early infant stages (Smith et al., 2008).
To better understand pneumatization as a growth mechanism, longitudinal radiographic analyses (e.g., Shah et al., 2003) and mathematical modeling (e.g., Zollikofer and Weissman, 2008) have been used. Using microanatomical methods, some authors have sought to understand cellular attributes of sinus mucosa (Sato et al., 1998; Wojtowicz et al., 2002) or the wall of sinuses and recesses (Smith et al., 2005, 2008, 2010, 2011). In this study, early (perinatal) spatial relationships and morphology of capsular cartilage is investigated across a broad array of primate species. Variations among species in our sample create natural experiments for elucidating growth processes. For example, our previous reports have shown that the perinatal time period is a dynamic phase of paranasal sinus formation in at least some monkeys, with some species undergoing secondary pneumatic expansion of the maxillary sinus (Smith et al., 2005, 2008, 2010, 2011). Moreover, different sinuses develop on different ontogenetic schedules in monkeys (Rossie, 2006; Smith et al., 2008). Specifically, it is possible to study one sinus during secondary pneumatization (the maxillary sinus) and another prior to secondary pneumatization (the frontal recess) in the same species (Smith et al., 2008). Finally, numerous species of primates fail to form one or both of these sinuses in adults (Rossie, 2006; Tückmantel et al., 2009; Smith et al., 2010). This remarkable diversity is used as a basis for this study, where the morphology, matrix characteristics, and the role of chondro-/osteoclastic activity in the onset of secondary pneumatization are studied.
Based on our earlier observation that different regions of the nasal capsule of S. geoffroyi have different developmental fates (ossification, persistence as cartilage, or apparent resorption) (Smith et al., 2008), here we assess nasal capsule morphology at the perinatal age in a taxonomically broad sample of non-human primates. In addition to traditional histochemical methods, we use osteopontin (OPN) immunohistochemistry and the tartrate-resistant acid phosphatase (TRAP) procedure to provide additional details about the nasal capsule. OPN has been implicated in bone remodeling as well as resorption of mineralized matrix in the epiphyseal plate (Reinholt et al., 1990; Sugiyama et al., 2001; Weizmann et al., 2005). OPN is also expressed in nasal capsule cartilage, particularly in elements of the ethmoid bone (Smith et al., 2008). The exact function of TRAP is unknown, but it may be vital to chondro-/osteoclast function (Hayman et al., 1996; Halleen et al., 1999). The TRAP procedure identifies chondro-/osteoclasts and their precursors (Colnot and Helms, 2001; Yeh and Popowics, 2011). The aims of this study are to: 1) characterize the cartilage matrix and chrondrocytes in the anterior nasal capsule and paranasal regions and 2) establish the presence or absence of cells that resorb bone and cartilage (osteoclasts or chondroclasts) adjacent to the nasal capsule.
MATERIALS AND METHODS
Twenty-seven cadaveric primates were studied (Table 1). All specimens were available as a result of natural deaths from captive populations in zoos or research centers. Most specimens used in this study were fixed in 10% buffered formalin after death. Some were frozen and subsequently fixed in formalin. The sample included a diverse range, seven species of anthropoids and eight species of strepsirrhines (including lemurs and bushbabies). Four of the species of monkeys were available as larger samples (Table 1), collected specifically for the study of early stages of pneumatization (Smith et al., 2010, 2011). All specimens had been prepared for histological observations in previous studies (e.g., Smith et al., 2003, 2005, 2007). For this study, archived, unstained sections of these specimens were used. All specimens had a complete, undamaged nasal capsule on one or both sides.
Table 1. Sample and species
Callithrix jacchus (common marmoset)
Cebuella pygmaea (pygmy marmoset)
Leontopithecus rosalia (lion tamarin)
Saguinus geoffroyi (Geoffroy's tamarin)
Saguinus oedipus (cotton-top tamarin)
Pithecia pithecia (saki monkey)
Saimiri boliviensis (squirrel monkey)
Eulemur macaco (black lemur)
Eulemur mongoz (mongoose lemur)
Lemur catta (ringtail lemur)
Propithecus verrauxi (Verraux's sifaka)
Microcebus murinus (gray mouse lemur)
Mirza coquereli (Coquerel's dwarf lemur)
Galago moholi (lesser bushbaby)
Otolemur crassicaudatus (greater bushbaby)
Histological Preparation of Specimens
Most aspects of specimen preparation were previously described (Smith et al., 2005, 2007, 2011). Briefly, all specimens were decalcified using a sodium citrate–formic acid solution and processed for paraffin embedding. Each paraffin block was serially sectioned at 10–12 μm. Selected sections were mounted on slides for staining with hematoxylin and eosin or other procedures for general micromorphology. Selected sections were mounted on slides or archived (the paraffin section partly pressed to paper and carefully boxed) as a means to save them, unstained, for alternate procedures. In this study, unstained sections were retrieved and used for three procedures. First, sections at selected anteroposterior cross-section levels were stained using an alcian blue procedure, which is used to the stain proteoglycans in developing cartilage, including matrix of the epiphyseal plate (Roach, 1997; Hamrick, 1999).
The sections used to study the nasal capsule were chosen based on the observation that characteristics of the nasal capsule cartilage differ across anteroposterior space (Smith et al., 2008). The developing nasal capsule may be divided into three regions (Fig. 1), which give rise to different structures postnatally (see Smith and Rossie, 2008 for further details). The pars anterior forms the rostral third of the nasal capsule in most primates, most rostrally opening at the anterior nares. Portions of this part of the nasal capsule persist as alar cartilages and small cartilaginous turbinals, but much of the pars anterior disappears (Schaeffer, 1920). In the middle of the nasal capsule, the pars intermedia houses two paranasal cavities, the frontal and maxillary recesses (Fig. 1a,b), and forms the maxilloturbinal as well as some anterior portions of the ethmoid bone. The pars posterior is the posterior-most region that forms the ethmoidal labyrinth (Fig. 1a). The pars posterior and pars intermedia are adjacent to the eye (Fig. 1). Selected sections in the anterior half of the nasal cavity were used for two additional procedures. Sections were selected based on regions of the nasal capsule that were previously demonstrated to exhibit fragmentation of the LNC in New World monkeys (Smith et al., 2005, 2008). In these regions, the cartilage matrix was studied using OPN immunohistochemistry and the TRAP procedure.
Histochemistry and Immunohistochemistry
Alcian blue procedure (pH 2.5) was used to identify cartilage matrix, which can include small remnant in newborn monkeys (Smith et al., 2005), and to characterize the distribution and density of certain hydrophilic macromolecules (proteoglycans) in developing cartilage (Roach, 1997; Hamrick, 1999; Smith et al., 2008).
For OPN immunohistochemistry, selected sections from the anterior and middle nasal capsule regions were obtained from Callithrix jacchus, Leontopithecus rosalia, Saguinus oedipus, and Saimiri boliviensis. Two lemurs (Eulemur spp.) were studied for comparison. The assay was conducted according to Vector protocols (and see Smith et al., 2005). Briefly, labeling of OPN was achieved by incubating slides with OPN antibody (Santa Cruz Biotech, CA) at a 1:250 dilution in 2% donkey serum for 30 min at RT, with a subsequent incubation with donkey anti-goat biotinylated secondary antibody (1:250 in phosphate buffered saline) for 30 min at RT. A Vector ABC kit (Vector Laboratories, CA) was used to stain immunoreactive structures, using diaminobenzidine (DAB) as chromogen. Following treatment with DAB, slides were counterstained with Harris' hematoxylin (Surgipath, IL). OPN staining was considered positive if staining was more intense than that found in comparable tissue on negative control slides that were incubated in the absence of primary antibody.
TRAP assays were conducted on selected sections of the four monkey species studied above, according to Sigma protocol. Slides were deparaffinized and rehydrated through serial alcohols for TRAP staining (acid phosphatase leukocyte TRAP assay, Sigma kit 387A). Slides were then refixed in 3.7% formaldehyde for 3 min at RT. Slides were rinsed for 2 min with dH2O then stained with diazotized fast garnet GBC solution (45 mL dH2O, 0.5 mL Fast Garnet GBC base solution, 0.5 mL sodium nitrile solution, 0.5 mL naphthol AS-BI phosphate solution, 2 mL acetate solution, and 1 mL tartrate solution for 1 hr at 37°C in the dark). Following TRAP staining, slides were rinsed with dH2O and counterstained with Gill's hematoxylin followed by an 8-min wash with running tap water. TRAP staining was considered positive if staining in cells with a reddish-brown color and verified by comparison to similar cells in slides where hematoxylin was used but the TRAP procedure was omitted. Adjacent sections were used when possible, allowing comparison of the same cells as prepared with and without the TRAP procedure.
At birth, the nasal capsule is fragmentary laterally in all species (Figs. 2, 3). More centralized projections, the turbinals, remain at least partially cartilaginous (Fig. 2a,b). Laterally, however, there are gaps in the pars anterior and pars intermedia. Alcian blue staining intensity varies regionally, in many cases being more intense in the septal cartilage than the LNC (e.g., Figs. 2a,c, 3d,f).
The inferior scroll of the pars intermedia that houses the maxillary recess (Figs. 1b, 2a, 3a,c,f) is mostly or completely absent, as a cartilaginous structure, in all primates at birth. In strepsirrhines, the only cartilaginous remnant of this scroll is its medial projection, the maxilloturbinal; even this portion is mostly ossified in most lemurs and all bushbabies. In some smaller strepsirrhines (cheirogaleids), some cartilage of the inferior part of the pars intermedia persists at the base of the maxillary recess. In anthropoids, the pars intermedia is absent in the region that borders the maxillary recess with two exceptions. First, the root of the maxilloturbinal, which is ossified in most of the monkeys, but still cartilaginous in tamarins, forms the lateral margin of the maxillary recess (see also Smith et al., 2008). Second, along the superior and lateral border of the maxillary recess, remnants of the pars intermedia are seen in some tamarins (Saguinus spp.) and S. boliviensis. Remnants are largest in S. boliviensis (Fig. 3a), where it surrounds the inferolateral sides of the recess in some specimens. The matrix of this remnant is lightly stained with alcian blue and is irregular in contour (Fig. 3a).
The region of the frontal recess is nearly or completely devoid of cartilage in some strepsirrhines, such as Propithecusverrauxi (Fig. 2a) and all the galagids studied (e.g., Fig. 2c). In Eulemur spp. (Fig. 2b) and Lemur catta, this recess is bordered medially by cartilage, while the capsular cartilage extends most laterally across the recess in cheirogaleids (Fig. 2d). The frontal recess is small at birth in the newborn monkeys studied. In its nascent region, near the middle meatus, the extent of cartilage is variable. In tamarins, the tectum in limited and the cartilage lateral to it is fragmented (Fig. 3b,c). In Pithecia pithecia (Fig. 3d), C. jacchus (Fig. 3e), and S. boliviensis (Fig. 3f), the nasal tectum is more extensive and continuous with cartilage along the LNC.
Immunohistochemistry and TRAP Procedure
In the specimens studied immunohistochemically, the cartilage of the pars anterior and nasal tectum is OPN (−), with small chondrocytes (Fig. 4a–c). In tamarins, OPN (+) multinucleate cells are observed near the nasal tectum and pars anterior. These cells are closely associated with bone adjacent to the cartilage (Fig. 4a) but are also observed adjacent to or between small islands of cartilage (Fig. 4b,c) or at the margins of the nasal tectum (Figs. 4c, 5). Multinucleate cells at or near the margins of the nasal tectum are TRAP (+) (Fig. 5).
In tamarins, the unossified tip of the maxilloturbinal is OPN (−) (Fig. 6a). The root of the maxilloturbinal, adjacent to the medial margin of the maxillary sinus, is ossified in all primates studied except tamarins. In tamarins, the maxilloturbinal root is cartilaginous. This cartilage is OPN (+) (Fig. 6b) and has hypertrophic chondrocytes and TRAP (+) cells (Fig. 7).
Several authors have suggested that local breakdown of nasal capsule cartilage is an important event that allows the growth of sinuses to occur more rapidly by secondary pneumatization (Wang et al., 1994; Maier, 2000; Smith et al., 2008). The results of this study provide support for this assertion, as those species that form maxillary or frontal sinuses have extensive reductions or local gaps in the LNC at sites that are documented to undergo pneumatization. In this study, we have documented dynamic aspects of nasal capsule morphology in primates that may be significant to pneumatization patterns.
Nasal Capsule Morphology
Although a maxillary paranasal space is nearly ubiquitous in mammals, secondary pneumatization from the primordial sinus space (i.e., the maxillary recess) is not (see Rossie, 2006; Smith et al., 2010, 2011). Some species do not expand the maxillary recess space, proportionally, beyond it natal size and shape (Rossie, 2006; Smith et al., 2010). The extent of secondary pneumatization in strepsirrhines is largely unknown, as the paranasal anatomy has not been well studied in its ontogeny (Table 2). Much of the LNC that surrounds this space in strepsirrhines appears to be absent at birth, especially in larger bodied lemurs. Further study of prenatal specimens is required to establish the stages of cartilage breakdown surrounding the maxillary recess.
Table 2. Species characteristics in the paranasal regions
N, no; Y, yes.
Hershkovitz remarked on variation in frontal pneumatization in the genus Callithrix, with some species having small “cells” communicating with the middle meatus.
Secondary pneumatization of the maxilla has not been investigated in strepsirrhines.
Secondary pneumatization of the maxillary recess in New World monkeys has been studied in many species (Rossie, 2006; Smith et al., 2005, 2010, 2011). This recess undergoes secondary pneumatization early in most New World monkeys, perhaps beginning prenatally in some species (Smith et al., 2008, 2010). In some New World monkeys, this space does not undergo secondary pneumatization (Rossie, 2006). Saimiri spp. are one such exception, in that the maxillary paranasal space does not expand beyond the proportions that are observed in the maxillary recess at birth (Smith et al., 2010). Of all primates studied, perinatal S. boliviensis has the largest remnant of the cartilage surrounding the maxillary recess, although it may degrade shortly after birth. Accordingly, delayed breakdown of cartilage bordering the maxillary recess may signify a reduction or cessation of paranasal expansion during development.
When compared with the maxillary sinus, development of the frontal sinus (in species that eventually form them) is at an earlier stage of expansion in all newborn primates. In species that are known to pneumatize the roof of the middle meatus (in monkeys) or frontal recess (in lemurs), gaps in the portion of the LNC that borders the frontal recess are prevalent. For example, the sifaka (P. verrauxi) has an expansive frontal recess, which extends dorsal to the level of the nasal tectum, and has little cartilage bordering this region (Fig. 2a). The sifaka goes on to develop a large frontal sinus (Tückmantel et al., 2009—Table 2). Some species (e.g., Cheirogaleus medius, Mirza coquereli) have more extensive cartilage in the anterior parts of the LNC at the perinatal age. These same species lack pneumatized frontal bones (Tückmantel et al., 2009). In other species (Lemur, and the bushbabies), the sequence of cartilage breakdown is unclear from the perinatal samples, and earlier stages of development require study.
In the tamarin S. geoffroyi, the nasal capsule consists anteriorly of isolated fragments at the perinatal age (Smith et al., 2008). In this study, we confirm that this is a characteristic of other tamarins, including S. oedipus and L. rosalia, all of which go on to have extensively pneumatized frontal bones (Table 2). In other monkeys, including Cebuella pygmaea, C. jacchus, P. pithecia, and S. boliviensis, the LNC of the pars intermedia and pars anterior are more complete, and there is little or no pneumatization of the frontal bone in these species (Hershkovitz, 1977; Rossie, 2006; Smith et al., 2010; Table 2).
Nasal Capsule Matrix Characteristics
It is well documented that the posterior part of the LNC undergoes endochondral ossification to form the ethmoid bone (e.g., see Schaeffer, 1920; Vidic, 1971a). Smaller portions of the anterior part of the LNC have the same fate, including the semicircular crest (becomes part of the ethmoid) and some more inferior parts of the pars intermedia (gives rise to the lateral wall of the inferior meatus and the maxilloturbinal). These parts of the LNC exhibit characteristics of endochrondral ossification. In the cartilage that forms the lateral wall of the inferior meatus (which also borders the maxillary sinus medially), linear arrays of OPN (+) matrix are seen in perinatal L. rosalia and Saguinus spp., as described for mineralized cartilaginous matrix in ossifying limb bones (Bark-Shalom et al., 1995; Sugiyama et al., 2001; Weizmann et al., 2005). These same cartilages also express linear arrays of alcian blue staining in mineralized matrix, which is also consistent with endochondral ossification (Hirschman and McCabe, 1969; Roach, 1997).
Other parts of the LNC, especially surrounding the lateral limits of the frontal recess, lateral wall of the maxillary sinus, and more anteriorly, have none of these characteristics. Chondrocytes are small and dispersed singly or in isogenous clusters, and the matrix is relatively homogeneous.1 This supports the assertion, previously based on cross-sectional age samples of the tamarin S. geoffroyi, that portions of the LNC undergo a different fate across anteroposterior limits (Smith et al., 2008).
Multinucleate Cells and the Nasal Capsule
Previously, Smith et al. (2008) observed multinucleate cells adjacent to isolated remnants of the anterior part of the LNC as well as within and near sites of endochondral ossification. The latter cells appear in a familiar context as chondroclasts that degrade mineralized cartilage matrix during endochondral ossification (Sugiyama et al., 2001; Weizmann et al., 2005). However, the fate of other parts of the nasal capsule has rarely been discussed. This study, using a broad range of primate species, verifies that chondroclasts are found in two settings throughout the nasal capsule. First, chondroclasts are found near endochondrally ossifying elements of the nasal capsule. In these bones, many chondroclasts are observed in close proximity to hypertrophic chondrocytes (cells which are hypothesized to regulate resorption in long bones) (Sakakura et al., 2005). As no species were studied across age, the sequence of events leading to their recruitment is not clear. Yet, this and other characteristics in the nasal capsule of tamarins suggest some similarities to endochondrally ossifying long bones. For example, some remnants of the posterior part of the LNC have matrix characteristics of endochondrally ossifying bone (see above; and see Smith et al., 2008).
In other locations, chondroclasts are observed in an entirely distinct context. In the anterior part of the LNC of tamarins, cartilage remnants occur in isolated islands. Here, chondroclasts are adjacent to the margins of cartilage that show no features consistent with endochondral ossification. Indeed, many of these islands of cartilage are anterior to the ethmoid bone. The sample studied here revealed some remnants that were partially broken down, and these too have OPN (−) matrix. These distinctions from the locations where endochondral ossification occurs strongly suggest there are two distinct mechanisms of chondro-/osteoclast recruitment in the nasal capsule. Distinct fates of viscerocranial cartilages have been described elsewhere in the facial skeleton. For example, chondroclasts appear to resorb Meckel's cartilage without the involvement of the signaling system described for long bone ossification, such as RANKL (Sakakura et al., 2005).
Findings on chondroclast distribution also bear on the process of pneumatization. Chondroclasts are observed along both cartilaginous and bony surfaces in newborn primates. In some of the species studied, pneumatic expansion of the sinus cavity is most certainly underway at birth, based on mapping of bone cells (Smith et al., 2005, 2011) and cross-age comparisons of sinus size (Smith et al., 2008, 2010). This offers a specific context for cartilage breakdown. Recruitment of chondro-/osteoclasts may be rapid, possibly occurring simultaneously in cartilage and bone in regions of secondary pneumatization.
SUMMARY AND CONCLUSIONS
At birth, matrix properties differ between portions of the nasal capsule that ultimately form elements of the ethmoid bone and regions of the nasal capsule that have no postnatal (descendant) structure. This distinction, regarding expression of OPN and distribution of glycosaminoglycans, applies broadly to primates. However, the extent of cartilage that remains in the pars anterior and pars intermedia varies. To some degree, the extent of cartilage present at birth reflects incipient sites of sinus expansion. Primate species known to form large frontal sinuses have an incomplete cartilaginous LNC at birth. Several species in which the frontal bone does not pneumatize have more complete lateral cartilaginous walls. One interpretation could be that this large remnant delays pneumatization. On the other hand, such remnants may simply reflect the pace of pneumatic expansion (none, in the case of S. boliviensis, Rossie, 2006; Smith et al., 2010).
It should be noted that bushbabies (Otolemur crassicadatus and Galago moholi) have little remaining LNC at birth but appear to have no pneumatization of the frontal bone (Tückmantel et al., 2009). Thus, the timing of breakdown may coincide with the advent of pneumatic expansion into bone is some species, but the LNC is replaced or resorbed regardless of whether secondary pneumatization occurs. We suggest certain findings herein, such as the great number of observations of chondroclasts near anterior nasal capsule fragments in tamarins, may be a manifestation of a specific stage of pneumatization and may occur earlier or later across species. A complete picture of the temporal pattern of pneumatization in primates is lacking, primarily because a broad fetal to postnatal age series is elusive except for very few species (e.g., see Smith et al., 2008).
Finding a non-primate animal model to understand pneumatization, preferably a less rare, faster reproducing taxon would seem a desirable route to further our understanding of pneumatization. Yet unique characteristics of primates make this challenging. One aspect of life history that distinguishes the order Primates from most other mammals is the relatively prolonged gestation length (Martin, 1983). It may be precisely this characteristic that makes the perinatal age a dynamic period in nasal capsule development in primates. Other, more altricial, species have much less ossified and less fragmented nasal capsule in the perinatal period (e.g., rats—Vidic, 1971b, Tupaia glis—Smith, unpublished data) compared to primates. A broader difficulty exists that our understanding of nasal capsule ontogeny and fate in a comparative perspective is poor. This study illustrates that such knowledge may inform our understanding of secondary pneumatization. Focal capsular cartilage breakdown should be regarded as part of the process of secondary pneumatization, with chondroclasts breaking down cartilage as one of the earliest events. Ontogenetic studies of diverse vertebrate taxa (specifically certain mammals and birds) are needed to further our understanding of this unique growth mechanism.
We thank T. Popowics for helpful discussions on the TRAP procedure.
In Smith et al. (2008), an erroneous sentence (p1410) states “The pars anterior and cartilage remnants in the maxillary sinus do exhibit alcian blue reactivity consistent with endochondral ossification.” This sentence should have read that these cartilages do not exhibit alcian blue reactivity consistent with endochondral ossification.