Sutures are fibrous connective tissue articulations between intramembranous craniofacial bones. Sutures are composed of fibroblastic cells with their matrices in the center and osteogenic cells in the periphery producing a matrix that is mineralized during skeletal growth. Whether oscillatory forces stimulate sutural growth is unknown. In the present work, we applied static and cyclic forces with the same peak magnitude of 5N to the maxilla in growing rabbits and quantified (1) acute in vivo sutural bone strain responses and (2) chronic growth responses in the premaxillomaxillary suture (PMS) and nasofrontal suture (NFS). Bone strain recordings showed that the waveforms of static force and 1-Hz cyclic force were expressed as corresponding static and cyclic sutural strain patterns in both the PMS and NFS, with the mean peak PMS strain (−1451 ± 137 με for the cyclic and −1572 ± 138 με for the static) approximately 10-fold higher than the mean peak NFS strain (124 ± 9 με for the cyclic and 134 ± 9 με for the static). Strain polarity was the opposite: compressive for the PMS but tensile for the NFS. However, on application of repetitive 5N cyclic and static forces in vivo for 10 minutes/day over 12 days, cyclic loading induced significantly greater sutural widths for the compressed PMS (95.1 ± 8.3 μm) than sham control (69.8 ± 8.2 μm) and static loading (58.9 ± 2.8 μm; p < 0.01). Interestingly, the same trend was true for the NFS under tensile strain: significantly greater sutural width for cyclic loading (267.4 ± 64.2 μm) than sham control (196.0 ± 10.1 μm) and static loading (169.9 ± 11.4 μm). Cell counting in 110 × 110 μm grids laid over sutures disclosed significantly more sutural cells on repetitive cyclic loading than sham control and static loading (p < 0.05) for both the PMS and NFS. Fluorescent labeling of newly formed sutural bone demonstrated more osteogenesis on cyclic loading in comparison with sham control and static loading. Thus, the oscillatory component of cyclic force or more precisely the resulting cyclic strain experienced in sutures is a potent stimulus for sutural growth. The increased sutural growth by cyclic mechanical strain in the tensed NFS and compressed PMS suggests that both microscale tension and compression induce anabolic sutural growth response.
SINCE THE RECOGNITION that bone architecture is affected by mechanical stresses, a myriad of studies have been performed in attempts to define what parameters of mechanical forces are determinants of anabolic bone responses. Although the precise relationship between mechanical stimuli and bone apposition is not clearly understood, several models have yielded insight into the mechanical modulation of periosteal and endocortical bone apposition. In avian and sheep ulna models subjected to four-point loading, periosteal and endocortical bone appositions are increased by short daily doses of mechanical stimuli as few as 36 cycles per day.(1,2) In a rat tibia model subjected to four-point loading up to 1200 cycles per day, periosteal and endocortical bone apposition rates vary as a function of strain rates.(3,4) Both macroscopic architecture and mineral apposition rates of the ulna in growing rats are affected by dynamic strain magnitudes and strain rates of axial compressive loading.(5,6)
A survey of the literature reveals that our knowledge of mechanical modulation of bone apposition has been gained almost exclusively from studies of the appendicular skeleton. Little is known about whether the craniofacial skeleton is responsive to mechanical stimuli in a similar fashion to the appendicular skeleton. The craniofacial skeleton differs from appendicular bones by not only its origin from neural crest cells, but also its irregular shape and the unique presence of sutures. Sutures are fibrous connective articulations between craniofacial bones and are composed of fibroblastic cells, with their matrices in the center, and osteogenic cells in the periphery, producing a matrix that is mineralized during skeletal growth. Sutures are present only in skull bones and experience mechanical strain in mastication(7–9) and on simulated orthopedic loading.(10,11) There has been a lack of knowledge of whether sutural osteogenesis is induced by precise doses of exogenous forces. Sutural growth consists not only of osteogenesis along the sutural bone edge, but also fibrogenesis that competes with sutural osteogenesis to maintain the existence of sutures.(12)
Mechanical modulation of sutural growth is of importance to the field of craniofacial orthopedics. It is commonly accepted that tension leads to bone formation and compression leads to bone resorption.(13,14) Albeit with little support of experimental evidence, this paradigm accounts for several clinical observations in the fields of distraction osteogenesis and orthodontic tooth movement.(12) However, an opposite notion is recognized in orthopedic medicine: compression leads to bone formation, whereas tension leads to bone resorption.(15,16) Compressive forces applied to long bones induce bone apposition on both periosteal and endocortical surfaces.(1,4,5) It is repeatedly observed in sports medicine that many activities inducing compressive loading on appendicular bones lead to increased bone density and architectural geometry.(17,18) Conversely, space flights or regimens of unloading lead to reduced bone mass.(19) In consideration of these contrasting tension-compression paradigms and the paucity of experimental evidence of force-induced sutural growth in craniofacial bones, we made an assumption in the present study that an overall compressive force with the same peak magnitude and two different waveforms, static and sine-wave cyclic, would induce net sutural bone apposition and resorption. Accordingly, we applied 5N cyclic and static forces to the rabbit maxilla for 10 minutes/day over 12 days. Our data show increasing sutural growth as evidenced by significantly greater sutural widths, greater sutural cell count, and increased osteogenesis as evidenced by fluorescence-labeled new bone formation on repetitive application of cyclic loading, in comparison with sham control and static loading. Interestingly, sutural growth increased in both the premaxillomaxillary suture under compressive strain, and the nasofrontal suture under tensile strain. Thus, small doses of cyclic tensile and compressive strains in two facial sutures increase sutural growth responses.
MATERIALS AND METHODS
For the chronic experiment, a total of 19 male, 6-week-old New Zealand White rabbits with a mean body weight of 1.1 kg (range, 0.9-1.3 kg) were randomly allocated to sham control (n = 5), static loading (n = 7), and cyclic loading (n = 7) groups. The rabbits were housed in a temperature-controlled room (23–25°C) and given standard amounts of food and water. The present animal protocol was approved by the Animal Care Committee of the University of Illinois at Chicago.
Chronic delivery of mechanical stimuli
General anesthesia was induced for all procedures by intramuscular injection of a cocktail containing 90% ketamine (100 mg/ml; Aveco, Fort Dodge, IA, USA) and 10% xylazine (20 mg/ml; Mobay, Shawnee, KS, USA). The rabbit was placed in a supine position in a custom-made resin body holder that rigidly supported the occipital region of the skull. The premaxilla was secured tightly to the resin body holder by restraining the oral commissure with stainless steel wires wrapped in a plastic sheath. A previously fabricated acrylic mold was used to secure the two maxillary incisors to the handset of a computerized servohydraulic system (858 Minibionix II; MTS, Eden Prairie, MN, USA).
Compressive forces with the same peak magnitude of 5N were preprogrammed and delivered through the acrylic mold placed on the maxillary incisors (Fig. 1) for 10 minutes/day over 12 days with two different waveforms: static forces with a frequency of 0 Hz and sine-wave cyclic forces with a frequency 1 Hz, leading to 600 cycles per day (cf., Fig. 2). After daily loading episodes, the rabbits were returned to cages and allowed food and water intake. Calcein green (15 mg/kg; Sigma, St. Louis, MO, USA) was injected intraperitoneally on day 2. After the last episode of loading on day 12, the rabbits were killed by pentobarbital overdose.
Strain gauge installation and recording
In a separate acute experiment under general anesthesia, five rabbits were placed in a supine position on a surgical table. Two facial sutures, the premaxillomaxillary suture (PMS) and the nasofrontal suture (NFS), were surgically exposed with technical details provided elsewhere.(11) Uniaxial strain gauges and three-element strain rosettes were installed across the inferior portion of the PMS and the bulk of the NFS in directions parallel to the direction of exogenous force. Despite the miniature size of strain gauges and strain rosettes for engineering applications, they covered the bulk of the inferior portion of the PMS and the entire NFS, providing global information of sutural mechanical strain. Installation of strain gauges (EA-06-062AQ-350; Measurements Group, Raleigh, NC, USA) and strain rosettes (WK-06-030WR-120; Measurements Group) followed procedures described in detail elsewhere.(10,20,21) All strain gauges/rosettes were excited with 1000 mV DC in 1/4-Wheatstone bridge circuits, and the output signals were conditioned and digitally recorded with computer data acquisition (Model 6000; Measurements Group).
After death, the premaxillomaxillary and nasofrontal sutures were isolated and dissected with at least 5 mm of surrounding bone. The sutural specimens on the left half of the skull were trimmed, dehydrated and demineralized in 50% formic acid and 20% sodium citrate, and embedded in paraffin. Sequential 8-μm sections were cut in the parasagittal plane and stained with hematoxylin and eosin (H&E). The right-side sutural specimens were trimmed and dehydrated in graded ethanol and acetone, and further prepared for undecalcified embedding using 85% methyl methacrylate (MMA) and 15% dibutyl phthalate. The polymerized MMA-specimen blocks were trimmed with a band saw. Sequential undemineralized 15-μm sections were cut in the parasagittal plane using a microtome (Leica Instruments, Nussloch, Germany).
The premaxillomaxillary and nasofrontal sutures were quantitatively assessed by computerized histomorphometric analysis (ImagePro and Nikon Eclipse E800; Nikon Corp., Melville, NY, USA). Standardized grids (1175 × 880 μm2) were constructed and laid over sutural histological specimens under a 4× objective. The linear sutural width was measured by constructing one circle bisecting the geometric width of the suture in the center of each grid block. Each circle's diameter was equal to the width of the suture within each grid block. The total number of sutural cells, regardless of cell type, was manually tagged and automatically counted in additional uniform grids (110 × 110 μm2) under a 10× objective. The average total numbers of sutural cells in six randomly selected grid blocks per sutural specimen were calculated. Newly mineralized bone along sutural edges labeled with calcein in undemineralized sections was imaged under a fluorescence microscope.
Data analysis and statistics
Means and SEM were calculated for peak bone strain from each rabbit. After confirmation of normal data distribution with Shapiro-Wilk tests, peak sutural strain in response to static and cyclic loadings per suture was compared with Student t-tests. ANOVA with Bonferroni tests was applied to compare the sutural widths and sutural cell count among sham control, static loading, and cyclic loading samples. Statistical significance was indicated when p values were <0.05.
Exogenous compressive forces with the same peak magnitude of 5N but with static and cyclic waveforms (Figs. 2A and 2B, respectively) were delivered to the maxilla and measured with a load cell attached to the computerized servohydraulic system. Both static and cyclic forces were compressive and of equal peak magnitude at 5N. However, sine-wave cyclic force demonstrated oscillations in force magnitude at 1 Hz, totaling 20 cycles in the 20-s representative time course (Fig. 2B), whereas static force showed no substantial oscillation in magnitude over time (Fig. 2A).
Sutural bone strain
During recordings with both uniaxial strain gauges and three-element strain rosettes, the principal bone strain of the PMS was compressive, whereas the principal bone strain of the NFS was tensile. After confirmation that the peak sutural strain lacked significant differences with strain recordings between uniaxial gauges and rosettes (data not shown), the following data were acquired from sutural strain recordings with uniaxial strain gauges. In response to exogenous static and cyclic forces with the same peak magnitude of 5N (Figs. 2A and 2B), the peak static and cyclic strains of the PMS were −1451 ± 137 με (mean ± SE; n = 7) for cyclic loading and −1572 ± 138 με (n = 7) for static loading, lacking significant differences. Bone strain patterns induced by static and cyclic forces, however, differed drastically for both the PMS and NFS. Both the patterns and rates of static and cyclic bone strain of the PMS reflected those of static and cyclic forces (Figs. 2C and 2D). For instance, cyclic strain of the PMS oscillated at 1 Hz (Fig. 2D) as the frequency of the exogenous cyclic force (Fig. 2B). In response to the same 5N exogenous force, the NFS exhibited mean peak bone strain of 124 ± 9 με (n = 7) for cyclic loading and 134 ± 9 με (n = 7) for static loading, also lacking significant differences. The mean NFS sutural strain was approximately 10-fold smaller than the mean PMS sutural strain (Figs. 2C-2F). Static and cyclic strain waveforms of the NFS (Figs. 2E and 2F), despite their further distance from exogenous load, also had different responses to static and cyclic forces, respectively. Strain polarity was the opposite: compressive (negative) for the PMS (Figs. 2C and 2D) but tensile (positive) for the NFS (Figs. 2E and 2F). Clearly, strain patterns of the PMS and NFS were modulated by corresponding waveforms of static and sine-wave cyclic forces by comparing sutural strain traces with force waveforms in Figs. 2A and 2B.
Representative photomicrographs of H&E-stained sections of the PMS and the NFS exhibited a high degree of sutural separation evoked by cyclic sutural strain (Figs. 3C and 3F) in comparison with both static strain (Figs. 3B and 3E) and sham control (Figs. 3A and 3D). The average sutural width of the compressed PMS treated with cyclic loading (95.1 ± 8.3 μm; n = 7) was significantly higher than static loading (58.9 ± 2.8 μm; n = 7) and sham control (69.8 ± 8.2 μm; n = 5; Fig. 4A). Interestingly, the average sutural width of the tensed NFS treated with cyclic loading (267.4 ± 64.2 μm; n = 7) was also significantly higher than static loading (169.9 ± 11.4 μm; n = 7) and sham control (196.0 ± 10.1 μm; n = 7; Fig. 4B). There was a lack of significant difference in sutural widths of both the PMS and NFS between static loading and sham controls (Figs. 4A and 4B).
Sutural cell count
The average counts of sutural cells within standardized grids for both the PMS and NFS treated with cyclic loading were significantly greater than both static loading and sham controls (Figs. 5A and 5B). The average sutural cell count of the compressed PMS treated with cyclic loading (50.5 ± 1.7) was significantly higher than both static loading (38.4 ± 1.2; n = 42) and sham control (41.7 ± 3.6; Fig. 5A). Interestingly, the average sutural cell count of the tensed NFS treated with cyclic loading (57.9 ± 3.3) was also significantly higher than static loading (50.0 ± 2.0) and sham control (45.8 ± 4.6; Fig. 5B). There was a lack of significant differences in sutural cell counts for both the PMS and NFS between static loading and sham controls (Figs. 5A and 5B).
Fluorescence labeling of newly mineralized bone
Fluorescence labeling by incorporation of calcein in newly formed bone minerals qualitatively demonstrated new bone formation. Compressive cyclic loading of the PMS (Fig. 6C) induced marked new bone formation as evidenced by calcein incorporation in comparison with both static loading (Fig. 6B) and sham control (Fig. 6A). Tensile cyclic loading of the NFS (Fig. 6F) evoked marked new bone formation in comparison with both static loading (Fig. E) and sham control (Fig. 6D).
The 5N compressive force delivered to the rabbit maxilla for 10 minutes/day over 12 days induces compressive strain of the PMS and tensile strain of the NFS. However, growth responses occurred in both sutures. These contrasting strain patterns of the PMS and NFS inevitably need to be discussed in the context of the accepted notion of tension leading to bone formation and compression leading to bone resorption in the fields of distraction osteogenesis and orthodontics,(12) and yet an opposite paradigm recognized in orthopedics.(15,16) The common association of tension or compression with bone growth refers to Newton-scale forces and tissue-borne strain. Convex and concave surfaces of both long bones and cranial bones likely experience tensile and compressive microstrains, respectively, potentially accountable for their separate formation and resorption processes.(6,10,11,15,16) However, Frost's flexural neutralization theory(15) was not designed to account for sutural bone growth. Sutural osteogenesis is likely modulated by microscale shear stresses inducible by either Newton-scale tension or compression.(12) Cellular growth in bone likely is a function of certain parameters of tissue-borne mechanical stresses such as fluid flow and the resulting deformation on cell membranes or cytoskeleton.(22–27) Even if tensile microstrain is successfully distinguished from compressive microstrain and delivered to a suture, collagen fibers in a three-dimensional mesh may become taut and thus compress the cells that reside within.
The present work shows the net effect of delivering either tensile or compressive cyclic mechanical strain up to about 1500 με to craniofacial sutures to be an acceleration of sutural growth, if growth is defined as increases in number and/or mass.(28) Sutural growth includes increases in cell numbers and masses of matrices of both fibroblastic and osteoblastic cell lineages. Increasing osteogenesis is exemplified by newly formed sutural bone labeled with vital stain in association with cyclic bone strain. Increasing fibrogenesis is evidenced by significantly increased sutural width and sutural cell count also in association with cyclic sutural strain. Sutures are geometrically complex structures. Quantification of sutural strain in the present work was simplified by both uniform axial loading direction and acquisition of global sutural strain information with strain gauges and strain rosettes that provided nearly complete coverage of the sutural course.
While it might have been reasonable to anticipate that the two bone edges of the NFS under tensile loading were to widen, the increased sutural width of the PMS under compressive loading is surprising. Because of a lack of prior work using precise application of mechanical stimuli to induce sutural growth, the following speculations are provided to account for the increased PMS sutural width. First, compressive loading of the PMS may have physically removed bony interdigitations (cf. Fig. 1B). Greater sutural space may result from removal of microfractured bone and subsequent replacement by sutural fibrous tissues, leading to increased PMS sutural width. This possibility, however, is against a consideration that the present compressive loading was perpendicular, instead of at an angle, to bony interdigitations of the PMS. Furthermore, there was a lack of microscopic evidence of bone microfracture in histological specimens, although it cannot be ruled out that microfracture may have taken place in the course of mechanical loading. Nonetheless, shear forces would have been more effective in physically microfracturing sutural interdigitations. Second, hypothetical osteoclastogenesis and bone resorption induced by cyclic strain in the PMS could have induced greater sutural width. Bone resorption can occur radially at a rate of 12 μm/day.(29,30) Thus, the maximum amount of bone resorption that could have taken place in sutures over the present 12-day loading duration is sufficient to account for the increased sutural width. However, this assumption is valid only if a large number of mature osteoclasts were present before mechanical loading. The total loading duration of 12 days in the present work seems to be too short for substantial osteoclastogenesis to occur, not to mention the possibility that mechanical strain could have inhibited osteoclastogenesis.(31) Third, sutures can perhaps be viewed as a “confined chamber” containing fibroblastic cells in the center, osteogenic cells in the periphery between sutural bone edges (cf. Fig. 3), and covered by dense fibrous connective tissue periosteum. This “confined chamber” seems to constitute an environment sensitive to mechanical stimulation. It is probable that mechanically induced fibrogenesis, including fibroblast proliferation and matrix production, accounts for at least some of the increased sutural width and is indirectly supported by the present increases in sutural cell count upon cyclic strain and the observation that the increased sutural space is occupied by cells and their extracellular matrices.
Strain amplitude is relevant to a potential threshold of anabolic bone responses. Despite a 10-fold difference between the PMS strain (∼1500 με) and the NFS strain (∼140 με), growth responses were accelerated in both sutures on cyclic loading. Sutural anabolic responses to a peak cyclic strain of 140 με of the NFS is interesting because this peak strain is within the disuse window proposed in Frost's Mechanostat model.(32) In most experimental studies involving long bones, strain amplitudes capable of inducing periosteal and endocortical bone appositions are above 1000 με.(6,33,34) Although bone strain of the skull could be above 1000 με,(8,35) sutural strain is higher than its adjacent bone strain. Thus, small pulsatile strain of the NFS may indeed be sufficient to activate anabolic responses of sutural cells. Given the model of viewing sutures as “confined chambers” potentially sensitive to mechanical stimulation introduced above, it is conceptually plausible that sutural cells can be activated with small deformation (strain).
The present increases in cell count provide a quantitative measure of the total number of sutural cells without identification of their lineages. Both sutural osteogenic cells and fibroblastic cells are likely derived from the common progenitor of mesenchymal cells. In vitro, both fibroblastic and osteoblastic cells are sensitive to mechanical stimuli.(36–38) The present increases in total sutural cell count on cyclic loading may include increased proliferation and/or decreased apoptosis of any of these cell lineages. The possibility that inflammatory cells contribute to the total sutural cell count is unlikely because of at least two considerations. First, there is a lack of significant differences in total cell count between static loading and sham control. Despite the same loading duration of 10 minutes/day, static force is at its peak at all times and much longer than cyclic force.(39) Thus, static loading would have induced greater inflammatory responses than cyclic loading. Second, daily loading for 10 minutes represents only 0.05% of total daily time. Although inflammatory responses in sutures on mechanical loading are inevitable, they most likely are transient instead of long lasting.
The present data must be interpreted with the following caveats, in addition to those already discussed above. Use of only one vital stain (calcein) did not provide the opportunity to calculate bone apposition rates. Sutures are exceedingly complex structures for histomorphometric analysis in comparison with transverse sections of long bones. In the present work, standardized grids and circles were used to provide some measure of sutural geometry so that quantitative analysis could be performed. These grids and circles may have the advantage over previously used techniques such as implantation of two amalgam markers across the suture, because two amalgam markers may not account for regional variations in sutural morphology. The present cell counting, although repeatable on different occasions, reflects a gross cellular profile of sutures instead of a true distinction between cell proliferation and apoptosis. Cell labeling with bromodeoxyuridine (BrdU) is being used in additional experiments for subsequent determination of the number proliferating cells in the S phase of the cell cycle. Even using BrdU will not allow differentiation between cells of osteoblastic lineage and cells of fibroblastic lineage in sutures.
This study has allowed a glimpse into the potential of oscillatory mechanical stimuli in modulating sutural growth. These data may have eventual implications in craniofacial orthopedics. From a therapeutic standpoint, only static forces are conventionally used to modulate sutural growth in craniofacial orthopedics including orthodontics.(14,15) There is the potential that small doses of oscillatory mechanical stimuli can effectively modulate sutural growth by either accelerating sutural osteogenesis or initiating net sutural bone resorption for different therapeutic goals.
We thank Drs Walter Greaves and Robert Scapino for their constructive comments on an earlier version of the manuscript. We thank two anonymous reviewers for their suggestions that have substantially improved the quality of the manuscript. Kevin O'Grady, Lauren Weitzman, James Nudera, and Verna Brown are gratefully acknowledged for their technical assistance. This research was supported by USPHS Research Grants DE13964 and DE13088 from the National Institute of Dental and Craniofacial Research.