- Email: [email protected]
Molecular orientation of collagen in intact planar connective tissues under biaxial stretch
Acta BIOMATERIALIA Acta Biomaterialia 1 (2005) 45–54 www.actamat-journals.com
Molecular orientation of collagen in intact planar connective tissues under biaxial stretch Jun Liao a, Lin Yang b, Jonathan Grashow a, Michael S. Sacks a
a,*
Engineered Tissue Mechanics Laboratory, Department of Bioengineering and the McGowan Institute for Regenerative Medicine, University of Pittsburgh, Room 234, 100 Technology Drive, Pittsburgh, PA 15219, United States b National Synchrotron Light Source (NSLS), Brookhaven National Laboratory, Upton, New York 11973, United States Received 17 August 2004; received in revised form 8 September 2004; accepted 8 September 2004
Abstract Understanding of the mechanical behavior of collagenous tissues at different size scales is necessary to understand their physiological function as well as to guide their use as heterograft biomaterials. We conducted a first investigation of the kinematics of collagen at the molecular and fiber levels under biaxial stretch in an intact planar collagenous tissue. A synchrotron small angle X-ray scattering (SAXS) technique combined with a custom biaxial stretching apparatus was used. Collagen fiber behavior under biaxial stretch was then studied with the same specimens using small angle light scattering (SALS) under identical biaxial stretch states. Both native and glutaraldehyde modified bovine pericardium were investigated to explore the effects of chemical modification to collagen. Results indicated that collagen fiber and molecular orientation did not change under equibiaxial strain, but were observed to profoundly change under uniaxial stretch. Interestingly, collagen molecular strain initiated only after 15% global tissue strain, potentially due to fiber-level reorganization occurring prior to collagen molecule loading. Glutaraldehyde treatment also did not affect collagen molecular strain behavior, indicating that chemical fixation does not alter intrinsic collagen molecular stiffness. No detectable changes in the angular distribution and D-period strain were found after 80 min of stress relaxation. It can be speculated that other mechanisms may be responsible for the reduction in stress with time under biaxial stretch. The results of this first study suggest that collagen fiber/molecular kinematics under biaxial stretch are more complex than under uniaxial deformation, and warrant future studies. 2004 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Collagen orientation; Small angle X-ray scattering; Small angle light scattering; Stress relaxation; Bovine pericardium
1. Introduction Collagen type I is the most abundant structural protein in the connective tissues. In order to bear tensile loads in an efficient manner type I collagen consists of a complex, structural hierarchy at various size scales [1,2]. The fundamental structural unit of collagen is
*
Corresponding author. Tel.: +1 412 235 5146; fax: +1 412 235 5160. E-mail address: [email protected] (M.S. Sacks).
the collagen molecule (diameter / 1.5 nm), which consists of triple helix amino acid chains [3]. The deposition of heavy metal stains in electron micrographs at the ends of the molecules produce dark bands commonly seen in collagen fibrils. The length of each collagen molecule is 4.4 times that of the period of the striation length D observed in electron micrographs. Along the molecular axis a gap of 0.6D is left between the ends of successive molecules. Through intrafibrillar bonding collagen molecules aggregate into collagen fibrils (/ 50 nm) in the quarter stagger form, with a D-period ranging from 64 to 68 nm [4,5]. Fibrils then form collagen fibers
1742-7061/$ - see front matter 2004 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2004.09.007
46
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
(/ 1 lm) via interfibrillar interactions mediated by proteoglycans [6]. The mechanical properties of collagenous connective tissues are largely determined by the unique structure of their constituent collagens. When the structure of these collagens are altered as a result of disease or chemical modification, tissue mechanical properties will be profoundly affected. It is thus important to understand the structure-function relationship in both native and biologically derived collagenous tissues. Moreover, constitutive (stress–strain) modeling of connective tissues is required in many physiological, surgical, tissue engineering, and medical device applications [5,7]. To predict the mechanical behavior of connective tissue, structural and compositional information are needed [8,9]. Clearly, knowledge of collagen kinematics in the loading process would directly benefit the structural based modeling. The mechanics of planar soft tissues are complex; loading induces straightening of highly crimped fibers, fiber rotation and stretch, and intra- and inter-molecular deformations. For example, structural models of skin have demonstrated that it is extreme extensibility (>50%) can be explained by reorientation of the collagen fibers alone [10,11]. We have developed a method that combines small angle light scattering (SALS) technique and biaxial stretch for measuring fiber kinematics in two dimensional soft collagenous tissues [9,12,13]. SALS provided a straightforward and rapid method to quantify the fiber orientation [14–16]. Incorporation of planar biaxial stretch and information of fiber angular distribution have provided much insight into the local fiber kinematics. For example it was found that fiber kinematics are highly tissue specific and fiber deformation is generally non-affine [9,13]. However, SALS can only provide orientation information at the fiber level. It is still unclear how fiber rotations and fiber uncrimping relate to the underlying collagen molecular rotations and stretch. It is well known that the X-ray scattering is an ideal tool to characterize the ultrastructure of collagen, such as molecule packing, fibril spacing, and orientation. Small angle X-ray scattering (SAXS) has been widely used to investigating the collagen structure [17–21]. Molecular strain in the intact collagen fibril under uniaxially stretched tendon/ligament have been widely reported [2,22–27]. In these studies tension-induced increases in the D-period were observed, with D-period strain proportional to the average strain of the specimen. A limitation of the above studies is that they are focused on uniaxially aligned tissues. In reality, most connective tissues are loaded in two and three dimensions and can have complex multi-dimensional fiber structures. Under multi-dimensional loading states collagen fiber deformations are more complex, and can include coupled effects of large rotations and non-affine deformations.
The above facts underscore the importance of understanding collagen molecule deformation and reorientation in intact issues under multi-dimensional strain states. Moreover, such understanding can lead to an improved understanding of the degradative processes of current biologically derived heterograft collagenous tissues. For example, molecular-level information can aid the development of novel tissue chemical treatment methods by providing a means to assess intrinsic collagen damage, independent of other damage mechanisms such as layer delamination. The objective of the current study was to initiate our investigations of collagen kinematics at the molecular level within an intact tissue under biaxial stretch using SAXS. We also compared the results of SAXS and SALS on the same tissue specimens in an attempt to link multi-axial kinematics at both the fiber and molecular levels. Both uniaxial and equibiaxial strain states were applied; the later used to restrict fiber rotations [7] so that the effects of fiber straightening alone could be studied.
2. Methods 2.1. SAXS biaxial stretch system A biaxial stretching device was custom made for the synchrotron X-ray beamline (X21 SAXS endstation) in NSLS of Brookhaven National Laboratory (Fig. 1). Two pairs of orthogonally positioned stepper motors (Parker automation, Stepper Motors VS23) provided motion to the four lead screws. Each pair was mounted in opposition and each motor was driven by a microstepping drive board (Parker Automation, Microstepping drive E-AC). The four microstepping drives were connected to motor controller card (Galil Motion Control, DMC-1800) via an interconnection module (Galil Motion Control, ICM-1900). The water-tight sample chamber was constructed using an aluminum alloy. Inside the sample chamber, there were two pairs of orthogonally positioned tissue holders with holes for the sutures. Each holder was mounted on a lead screw and a guide shaft. The front side of sample chamber was covered by transparent polyvinyl chloride (PVC) glass for strain monitoring purposes. A hollow stage right behind the specimen area was raised from the rear wall of the sample chamber. Another PVC glass plate covered the hollow stage. Two holes of 1 mm and 3 mm diameter were made on the front and rear PVC glass, respectively. Kapton film, used as X-ray windows, was pasted over the two holes with epoxy glue. A CCD camera (National Instrument, IMAQ CCD) perpendicular to the beamline was used to track four centrally located graphite markers on tissue with the help of a 45 mirror, in which a centered hole
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
Fig. 1. (a) A schematic illustration of SAXS biaxial stretch system, (b) SAXS biaxial stretch setup in X21 station of NSLS in BNL. Labels: (1) Specimen chamber transparent to both X-ray and light; (2) Guide shaft; (3) Lead screw; (4) Tissue holder; (5) Concave stage; (6) Motor; (7) Suture; (8) Tissue in PBS solution; (9) 45 mirror; (10) CCD camera; and (11) X-ray detector.
(/ = 3 mm) was made to allow the X-ray passing through. A Labview (National Instrument, Version 6) program was written to control the motion of motor and track the tissue strain at the same time. 2.2. Sample preparation Native bovine pericardium (BP) and 0.625% glutaraldehyde in phosphate buffered saline treated bovine pericardium (GLBP) specimens were used as the representative planar collagenous tissues. Tissue specimens (n = 5 for each group) were cut into 25 mm square samples. Four graphite markers (0.3 mm diameter) were affixed to the center of each specimen with cyanoacrylate adhesive in a 5 · 5 mm square pattern. The specimen was then mounted on the tissue holders with polyester suture (size 0) via four loops on each side. In the experiment, tissue was immersed in phosphate buffered saline at room temperature. 2.3. SAXS measurements SAXS data was collected at beamline X21 at NSLS ˚ X-rays. The sample-to-detector distance with 1.55 A was 1.33 m and the beam size at the sample was
47
0.4 mm. The diffraction patterns are recorded with a Mar CCD detector. The D-period of the collagen fibrils was extracted from the reciprocal spacing between the Bragg peaks produced by these structures. Specimens were mounted in the unstretched state, then stretched just enough until there was no observable sagging. This state was used as reference state for strain calculations. Both equibiaxial stretch and strip biaxial stretch (i.e. where the stretch along one axis is increased while the orthogonal axis is held at zero strain) were applied. In the equibiaxial stretch studies, specimens were stretched using a sequence of 5% equibiaxial strain increments. At each strain level, SAXS patterns were recorded. In the strip stretch studies, specimens were stretched in one direction while maintaining the other direction at 0% strain. A sequence of strip strains with a 5% increment was applied and SAXS pattern recorded. The accuracy of the stretch measurements was within ±0.001. For both BP and GLBP, the data was collected by averaging two reading (120 s for each reading). Scattering patterns were then analyzed with custom written program. In order to correctly calculate the averaged scattering profile and angular intensity distributions, masks (i.e. a fan or a ring shape) were used to block off unusable portions of the intensity data. A fan centered at beam center were used for quantifying the scattering profile and a ring enclosing fifth diffraction maximum were used for quantifying angular distribution. Details of the SAXS technique are given in the Appendix A. 2.4. SALS based fiber-level kinematic measurements A SALS system combined with a custom made biaxial device was used to determine the collagen fiber orientation in the same GLBP after SAXS testing. A detailed description of SALS protocol and effectiveness on fiber architecture evaluation has been previously presented [12,13,15]. Briefly, a monochromatic laser Helium–Neon laser light (k = 632.8) is passed through a tissue specimen. Since the wavelength of light is within an order of magnitude of the diameter of collagen and elastic fibers, small angle scattering pattern represents information of fiber orientation. The SALS-biaxial stretch system is similar with the system used in SAXS [12]. The biaxial stretching device consisted of two pairs of lead screws driven by two computer-controlled stepper motors. Specimen mounting method and testing protocol were as same as SAXS. A 0.8 mm laser beam size was used to assure that the area tested by SAXS was included within the light beam envelope. At each strain level, SALS pattern were recorded and analyzed using custom software. To facilitate comparisons with the SAXS data, statistical distribution function of the angular distribution of collagen fibers or molecules, R(h), was defined as [9]
48
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
IðhÞ RðhÞ ¼ Ph¼p=2 h¼p=2 IðhÞDh
ð1Þ
where I(h) represents the angular intensity distribution.
3. Results 3.1. SAXS D-period strain measurements In SAXS patterns of BP and GLBP we found that in the initial 0% strain state, diffractions demonstrated a relatively weak molecular orientation (Fig. 2a and d). For BP, we found that both stretch protocols induced a smooth change of molecular orientation, with only one major orientation direction observed (Fig. 2b and c). Interestingly, for GLBP we observed multiple major molecular orientation peaks, probably a result of multiple fiber populations in this specimen (Fig. 2e and f). Representative scattering profiles (intensity-q spectra line) along the orientation peak for GLBP demonstrated a change in the D-period with applied strip stretch (Fig. 3). D-period changes for both BP and GLBP along the strip stretch direction (90) and along the 120 direction for the same sample were quantified (Table 1, Fig. 4). Interestingly, we found that the BP D-period did not increase until the 20% global strip strain reached (Fig. 4). In contrast, GLBP D-period began to increase from the 10% global strip strain, which was quicker than BP (Fig.
Fig. 3. GLBP D-period diffraction spectra under different strip stretches, indicated on the right hand side of the plot, and shifted upward for clarity. Peak values of diffraction spectra were found to change with the increase of strip strain for strains >10%, indicating the molecular strain did not occur until at least 10% global tissue strain.
4). To estimate the local molecular-level loading to collagen in relation to the global tissue strains, we computed the D-period strain/global tissue strain ratio.
Fig. 2. Typical SAXS patterns of native BP and glutaraldehyde fixed bovine pericardium (GLBP) under biaxial and strip stretch, (a–c) BP in the initial state, an equibiaxial strain of 20%, and a maximum strip strain of 30%, respectively, (d–f) GLBP in the initial state, equibiaxial strain of 10%, and under a maximum strip strain of 25%, respectively. Note for BP there was only one observable orientation peak, while for the GLBP four peaks could be discerned.
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
49
Table 1 ˚ ) of native bovine pericardium and glutaldehyde treated bovine pericardium with applied global strip biaxial stretch Representative D-period (A
BP (strip stretch direction, 90) GLBP (strip stretch direction, 90) GLBP (120 direction)
0%
5%
10%
15%
20%
25%
30%
654.50 654.50 654.50
654.50 654.50 654.50
654.50 654.50 654.50
654.50 675.61 661.39
654.50 682.95 668.42
668.42 690.46 668.42
675.61 – –
Strain of fibril D-period (%)
Note: GLBP did not reach 30% global strip strain.
6
BP along vertical stretch direction
5 4 3 2
Slope = 0.32%/%
1 0 0
5
10
15
20
25
30
35
Strain of fibril D-period (%)
Global tissue strain along stretch direction (%) 6
GLBP along vertical stretch direction
5 4 3
Slope = 0.35%/%
2 1 0 0
5
10
15
20
25
30
35
Global strain along stretch direction (%) Fig. 4. Relationship between D-period strain and global strip strain for are representative specimen. Top curve showed D-period change in BP measured along the vertical strip stretch direction. Bottom curve showed D-period changes in the same GLBP measured along the vertical strip stretch direction. Variations of the five specimens tested were small and were within ±0.1% of the mean D-period fibril strain at each global tissue strain tested.
The relative change of the BP D-period was 0.32%/% and 0.35%/% for GLBP, indicating little differences between these two specimens. 3.2. Effects of biaxial stretch on collagen molecule rotational kinematics With increasing equibiaxial strain of up to 20%, collagen molecules were observed to exhibit an increased degree of orientation (Fig. 5). The angular molecular distribution was relatively smooth and major orientation exhibited a peak at 125, and included a secondary peak at 100 (Fig. 5a). In comparison, when a strip stretch was applied in the vertical (90 orientation)
Fig. 5. Representative angular distribution of D-period diffraction intensity (a) when BP was under different equibiaxial strains, (b) when BP was under different strip strains. Strip stretch was applied along the vertical (i.e. 90) direction. SAXS intensity distribution at different global strain level was shifted upwards for demonstration purpose. The dotted lines indicate the angular locations of the main orientation peaks at maximum stretch.
direction, a peak occurred at 105, with a secondary peak at 80 (Fig. 5b). Moreover, the uniaxial strip stretch distribution was noticeably narrower (Fig. 5b) compared to the biaxial stretch distribution (Fig. 5a). When chemically treated similar reorientation behavior was observed for the GLBP specimens (Fig. 6). In the initial state, the GLBP specimen exhibited a weak orientation peak at 130. Under equibiaxial strain two major orientation peaks at 40 and 130 formed (Fig. 6a). Upon further stretch to 10% strain these orientation peaks split into three peaks at 50, 105, and 135 (Fig. 6a). Under vertical (i.e. 90 orientation) strip stretch, the orientation peak splitting was more prominent with four peaks occurring at 60, 100, 115, and 130 at 25% strip strain (Fig. 6b). 3.3. Collagen fiber kinematics SALS results for the same GLBP specimen exhibited a major orientation peak at 130 (Fig. 7a). This was
50
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
When the equibiaxial strain increased to 10%, peak intensity was found still around 130. Under strip (90 orientation) stretch fiber orientation peak gradually switched from 130 to 90 (Fig. 7b), indicating fibers gradually aligned toward the strip stretch direction. This observation is consistent with the molecular orientation range of 40–140 shown by SAXS (Fig. 6b). 3.4. Collagen orientation and D-period in stress relaxation
Fig. 6. Angular distribution of D-period diffraction intensity (a) when GLBP was under different equibiaxial strains, (b) when GLBP was under different strip strains. Strip stretch was applied along the vertical direction. SAXS intensity distribution at different global strain level was shifted upwards for demonstration purpose. The dotted lines indicate the angular locations of the main orientation peaks at maximum stretch.
Fig. 7. Representative angular distribution of scattered light intensity (a) when GLBP was under different equibiaxial strains, (b) when GLBP was under different strip strains. Strip biaxial stretch was applied along the vertical direction. SALS intensity distribution at different global strain level was shifted upwards for demonstration purpose. The dotted lines indicate the angular location of the main orientation peak at minimum and maximum stretch.
consistent with collagen molecular orientation (130) observed by SAXS under equibiaxial strain (Fig. 6a).
To investigate time-course changes in orientation under a fixed kinematic state, we investigated changes in molecular orientation during biaxial stretch relaxation. A native BP specimen was loaded and held at a 30% strip strain level. D-period diffraction patterns were recorded every 4 min during the stress relaxation process over a total time period of 80 min. We observed that the angular distribution of D-period diffraction did not change in the stress relaxation process (Fig. 8), suggesting that the collagen molecular orientation was constant. Moreover, quantification of diffraction spectra showed that D-period also kept constant in the relaxation process (Fig. 8b). When the specimen was unloaded the major orientation at 30% strip strain disappeared (Fig. 8a) and D-period fell back to initial value (Fig. 8b).
Fig. 8. (a) For BP the angular distribution of the D-period diffraction intensity was found not to change with stress relaxation process under 30% fixed global strip strain. However, the intensity distribution changed back to its initial state when the tissue was unloaded, (b) D-period strain at 30% fixed global strip strain was also found not to change in the time course of stress relaxation (240 s to 80 min), suggesting that the relaxation of the load was not due to collagen.
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
51
4. Discussion 4.1. Collagen molecular and fiber orientations The present study represents a first look at the molecular and fiber level kinematical behavior of a planar collagenous tissue under biaxial stretch. The molecular level orientation changes revealed by SAXS were generally consistent with the fiber level orientation changes revealed by SALS. These results are consistent with those found by Kronick and Buechler [15], who found close agreement in the collagen molecular and fiber orientations using SALS and SAXS for calf skin under uniaxial loading conditions. However, in the present study we found that the kinematics of collagen molecules in intact planar connective tissues were more complex than in previous uniaxial studies. This is most easily seen in the SAXS molecular orientation responses between the equibiaxial and strip biaxial strain states (Figs. 5 and 6, respectively). Firstly, we noted that in the unstrained state the tissue the molecules had only a very modest degree of alignment (Figs. 5a and 6a). Upon equibiaxial stretch, molecular orientation developed prominent peaks. This can be explained that in the initial unloaded state the collagen fibers are very undulated (i.e. crimped) [9,28,29] which would have a ‘‘randomizing’’ effect on the overall orientation. As the tissue is stretched equibiaxially, the fibers are kinematically restricted and cannot rotate [9]. Thus, the fibers can only straighten and lose their initial crimp. This results in an improved molecular alignment with increasing equibiaxial stretch (Figs. 5a and 6a). Note too that under strip stretch (where the vertical axis is stretched but the horizontal axis is kept at zero strain), similar qualitative observations occurred (Figs. 5b and 6b). However, in this case the resulting distributions were narrower, indicative of a great degree of molecular alignment. It is also interesting to note that in both the native (Fig. 4a) and chemically modified tissues (Fig. 4b) there was a delay in molecular stretch with tissue strain. This phenomena has been observed previously in tendon [25,26,30]. This delay in molecular strain is generally attributed to the initial fiber reorientation that occurs in the toe region of the stress–strain curve, where the collagen fibers are subjected to negligible loading. Thus, no molecular-level strain occurs until the shortest collagen fibers become straightened and begin to bear load. A collagen crimp period of 25–30 lm has been observed for BP collagen fibers (e.g. Fig. 9), with an estimated amplitude of 10 lm. This explanation is thus consistent with the substantial degree of crimp found for native BP [31]. The slope of the D-period strain to global tissue strain would be 1.0%/% if all the tissue strain was uniformly transmitted to the molecular level. In the present
Fig. 9. Polarized photomicrograph of a BP specimen, displaying a collagen crimp period of 30 lm, crimp amplitude was estimated to be 15 lm. These structural results indicate that BP has a substantial degree of crimp which may be responsible for the differences found between fiber and molecular level stain behavior.
study this ratio was 0.33%/% (Fig. 4) was indicative that not all the global tissue strain is translated to the molecular level. The reason for this lower value is most likely due to the fact that, even under equibiaxial strain, not all the collagen fibers are loaded to the same local strain level as the tissue is loaded. That is, there is a distribution of collagen fiber straightening strain that has been theorized to produce the non-linear tissue stress– strain curve [9]. Thus, there will be an equivalent distribution of fiber strains, resulting in an average D-period strain to global tissue strain slope less than 1%/%. It is also possible that inter-fibrillar slippage may contribute to the lower slope. Clearly, additional studies are required to elucidate this phenomenon. 4.2. Effects of exogenous crosslinks After glutaraldehyde fixation collagen is crosslinked by aldehyde functionalities which stiffened the tissue and presumably limit the freedom of fibril realignment. Yet, we did not observe any differences in the D-period strain/global tissue strain ratio (Fig. 4). The relative change of the BP D-period was 0.32%/% and 0.35%/% for GLBP. These results suggest that chemical modification did not measurably affect collagen molecular stiffness. We have observed that chemical fixation main affects the low strain region of the stress–strain curve under biaxial stretch [32,33]. The results of the present study further suggest that the formation of exogenous crosslinks affects larger scale structures, and not the collagen molecules themselves. 4.3. Collagen molecular behavior in stress relaxation In the present study we found that in the time course 4–80 min no detectable changes in the angular distribution and D-period strain occurred during stress relaxation (Fig. 8). This observation is consistent with similar
52
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
studies by Purslow et al. on rat skin [24]. Moreover, according to Mosler et al. [34] when a constant 6.5% grip-to-grip strain was applied to rat tail tendon collagen fibers an initial increase in D-period occurs. However, this value decreases back to the original 67 nm value at a faster rate compared to the rate of stressrelaxation in the same specimen [34,35]. Closer inspection of MoslerÕs data indicated that the D-period decreased in the first 200 s, but did not change from 200 to 400 s. In our study the time resolution is 240 s, which is the time required to obtain the SAXS image. We found no detectable changes in planar connective D-period in over the time course of 240 s to 80 min in the stress relaxation. Our results were also consistent with PurslowÕs [24] conclusion that stress relaxation behavior is not caused by time-dependent reorientation of collagen fibers. Other mechanisms for the relaxation in collagenous tissues include (i) the matrix surrounding the collagen fibers and (ii) time-dependent effects at the interface between fiber and matrix [24]. One might overlook the contribution of matrix based on concept that significant forces are borne by the collagen fibers [24]. However, with the importance of interfibrillar interactions mediated by proteoglycans gradually becoming better appreciated [36,37] additional studies are needed to elucidate the influence of interfibrillar matrix. 4.4. Study limitations and future directions The primary limitation of the current study is that the present SAXS biaxial apparatus does not allow for simultaneous force measurements. Without force information, it was not possible to determine the actual molecular stiffness. However, we felt that a feasibility study to determine if there were indeed additional effects of encountered in biaxial loading on the collagen fiber kinematics was warranted before undertaking the additional significant cost and time to implement force measurements. Given the intriguing results of the present study, we are currently embarking on modifications to the biaxial test system to perform simultaneous force measurements. This will allow us to determine critical mechanical parameters for collagen, such as the modulus, under biaxial stretch. Moreover, more sophisticated time dependent experiments (such as creep) will be able to be preformed. In our experimental design, length of X-ray beam passing through both water and tissue is 4 mm. This path length requires a relatively long time (240 for both BP and GLBP) to obtain the scattering pattern, which inevitably influenced our ability to get higher time-resolution when analyzing the stress relaxation. Finally, the current specimens were randomly chosen from the pericardial sac. This was done intentionally to allow for an unbiased examination of collagen fiber kinematics.
However, given the natural variability in collagen structure of the pericardial sac [38,39] future studies should include presorting [40,41]. 4.5. Summary We conducted a first study of the kinematics of collagen at the molecular and fiber levels under biaxial stretch in an intact planar collagenous tissue. Our key observations include: • Collagen molecular strain in BP did not increase immediately with global tissue strain, but only began to increase only after 10–20% tissue strain, suggesting that substantial fiber-level reorganization occurs prior to collagen molecular loading under biaxial stretch. • The resulting SAXS intensity distributions demonstrated greater alignment under strip biaxial stretch compared to the equibiaxial stretch state. • No detectable changes in the angular distribution and D-period strain were found in the up to 80 min of stress relaxation. These final results suggest that viscoelastic mechanisms in collagenous tissues are not due to molecular-level changes, but results from changes at larger structural scales. • Chemical fixation indicated no change in molecular stiffness (we believe this is the first observation of this). Clearly, additional studies will be required to elucidate the observed phenomenon.
Acknowledgment This work is supported by NIH grant HL63026 and an Established Investigator Award of the American Heart Associate to MSS. Research carried out (in whole or in part) at the National Synchrotron Light Source, Brookhaven National Laboratory, which is supported by the US Department of Energy, Division of Materials Sciences and Division of Chemical Sciences, under Contract No. DE-AC02-98CH10886. The authors would like to thank Daniel Hildebrand and Thomas Gilbert for invaluable discussions.
Appendix A A SAXS pattern from a group of parallel aligned collagen fibrils (along vertical direction) consist of two orthogonal diffraction series: (i) In the direction that collagen fibril aligned, axial D-period works as a diffraction spacing and generates meridional series of up to 30 relatively sharp diffraction centered on the fibril axis and
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
the beam center; (ii) In the direction perpendicular to collagen fibrils, a more diffuse diffraction (Equator series) can be observed and the first peak corresponds to an intermolecular packing distance of 1.5 nm [23,24, 42]. If the diffraction is caused by an angularly dispersed population of collagen fibrils, angular dispersion of diffraction series will be observed. Since meridional series is caused by fibril axial spacing. The high intensity on the angular dispersion arc represents the major fibril orientation (i.e. collagen molecule orientation) [24]. The diffraction maxima occurring at angles 2h to the X-ray beam (Bragg reflections) is given by equation: 2 sin(h) = kn/d, where n is the order of the reflection, k, is X-ray wavelength, and d is the spacing. Scattering angle 2h and spacing d were reciprocally related. Therefore, using specific calibration, values of d can be determined from diffraction pattern. Scattering vector ˚ 1), defined as q = 4p sin(h)/k, was used in this q (unit, A study. Therefore, we have q = 2p/d. Value of 2p/d is the distance between intensity peaks and can be accurately calculated using linear fitting (peaks from 3rd to 9th order used in this study). Spacing d can then be calculated by equation d = 2p/q.
References [1] Kastelic J, Baer E. Deformation of tendon collagen. In: Vincient JF, Currey JD, editors. The Mechanical Properties of Biological Materials. Great Britain: Society for Experimental Biology; 1980. p. 397–433. [2] Fratzl P, Misof K, Zizak I, Rapp G, Amenitsch H, Bernstorff S. Fibrillar structure and mechanical properties of collagen. J Struct Biol 1998;122:119–22. [3] Chapman JA, Hulmes DJS. Electron microscopy of the collagen fibril. In: Ruggeri A, Motto PM, editors. Ultrastructure of the connective tissue matrix. Boston: Martinus Nijhoff; 1984. p. 1–33. [4] Hodge AJ, Petruska JA. Recent studies with the electron microscope on ordered aggregates of the tropocollagen molecule. Aspects of protein chemistry. London: Academic Press; 1963. [5] Fung YC. Biomechanics: Mechanical properties of living tissues. New York: Springer Verlag; 1993. [6] Kastelic J, Galeski A, Baer E. The multicomposite structure of tendon. Connect Tissue Res 1978;6:11–23. [7] Sacks MS. Biaxial mechanical evaluation of planar biological materials. J Elasticity 2000;61:199–246. [8] Wren TA, Carter DR. A microstructural model for the tensile constitutive and failure behavior of soft skeletal connective tissues. J Biomech Eng 1998;120:55–61. [9] Sacks MS. Incorporation of experimentally derived fiber orientation into a structural constitutive model for planar collagenous tissues. J Biomech Eng 2003;125:280–7. [10] Lanir Y. Biaxial stress–strain relationship in the skin. Israel J Technol 1979;17:78–85. [11] Lanir Y. A structural theory for the homogeneous biaxial stress– strain relationships in flat collageneous tissues. J Biomech 1979;12:423–36. [12] Billiar KL, Sacks MS. A method to quantify the fiber kinematics of planar tissues under biaxial stretch. J Biomech 1997;30:753–6.
53
[13] Sacks MS, Gloeckner DC. Quantification of the fiber architecture and biaxial mechanical behavior of porcine intestinal submucosa. J Biomed Mater Res 1999;46:1–10. [14] Chien JCW, Chang EP. Small-angle light scattering of reconstituted collagen. Macromolecules 1972;5:610–7. [15] Kronick PL, Buechler PR. Fiber orientation in calfskin by laser light scattering or X-ray diffraction and quantitative relation to mechanical properties. J Amer Leather Chem Assoc 1986;81: 221–9. [16] Sacks MS, Smith DB, Hiester ED. A small angle light scattering device for planar connective tissue microstructural analysis. Ann Biomed Eng 1997;25:678–89. [17] Farkasjahnke MVS. Small-angle X-ray diffraction studies on rattail tendon. Acta Physiol Acad Sci 1965;28:1–17. [18] Bigi A, Incerti A, Leonardi L, Miccoli G, Re G, Roveri N. Role of the orientation of the collagen fibers on the mechanical properties of the carotid wall. Boll Soc Ital Biol Sper 1980; 56:380–4. [19] Suarez G, Oronsky AL, Bordas J, Koch MH. Synchrotron radiation X-ray scattering in the early stages of in vitro collagen fibril formation. Proc Natl Acad Sci USA 1985;82: 4693–6. [20] Aspden RM, Bornstein NH, Hukins DW. Collagen organisation in the interspinous ligament and its relationship to tissue function. J Anat 1987;155:141–51. [21] Fratzl P, Daxer A. Structural transformation of collagen fibrils in corneal stroma during drying. An X-ray scattering study. Biophys J 1993;64:1210–4. [22] Misof K, Rapp G, Fratzl P. A new molecular model for collagen elasticity based on synchrotron X-ray scattering evidence. Biophys J 1997;72:1376–81. [23] Wess TJ, Purslow PP, Kielty CM. X-ray diffraction studies of fibrillin-rich microfibrils: effects of tissue extension on axial and lateral packing. J Struct Biol 1998;122:123–7. [24] Purslow PP, Wess TJ, Hukins DW. Collagen orientation and molecular spacing during creep and stress-relaxation in soft connective tissues. J Exp Biol 1998;201(Pt l):135–42. [25] Sasaki N, Odajima S. Stress–strain curve and YoungÕs modulus of a collagen molecule as determined by the X-ray diffraction technique. J Biomech 1996;29:655–8. [26] Sasaki N, Odajima S. Elongation mechanism of collagen fibrils and force–strain relations of tendon at each level of structural hierarchy. J Biomech 1996;29:1131–6. [27] Sasaki N, Shukunami N, Matsushima N, Izumi Y. Time resolved X-ray diffraction from tendon collagen during creep using synchrotron radiation. J Biomech 1999;32:285–92. [28] Fentie IH, Alien DJ, Schenck MH, Didio LJA. Comparative electron microscopic study of bovine, porcine, and human parietal pericardium, as materials for cardiac valve bioprostheses. J Submicrosc Cytol 1986;18:53–65. [29] Shabetai R. The pericardium. New York: Grune and Stratton; 1981. [30] Gathercole LJ, Keller A. X-ray diffraction effects related to superstructure in rat tail tendon collagen. Biochim Biophys Acta 1978;535:253–71. [31] Sun W, Sacks M, Fulchiero G, Lovekamp J, Vyavahare N, Scott M. Response of heterograft heart valve biomaterials to moderate cyclic loading. J Biomed Mater Res 2004;69A:658–69. [32] Sacks MS. A structural constitutive model for chemically treated planar connective tissues under biaxial loading. Comput Mech 2000;26:243–9. [33] Sun W, Sacks MS, Sellaro TL, Slaughter WS, Scott MJ. Biaxial mechanical response of bioprosthetic heart valve biomaterials to high in-plane shear. J Biomech Eng 2003;125:372–80. [34] Mosler E, Folkhard W, Knorzer E, Nemetschek-Gansler H, Nemetschek T, Koch MH. Stress-induced molecular rearrangement in tendon collagen. J Mol Biol 1985;182:589–96.
54
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
[35] Folkard W, Geercken W, Knorzer E, Mosler E, NemetschekGansler H, Nemetschek T, et al. Structural dynamic of native tendon collagen. J Mol Biol 1987;193:405–7. [36] Silver FH, Kato YP, Ohno M, Wasserman AJ. Analysis of mammalian connective tissue: relationship between hierarchical structures and mechanical properties. J Long Term Eff Med Implants 1992;2:165–98. [37] Elliott DM, Robinson PS, Gimbel JA, Sarver JJ, Abboud JA, lozzo RV, et al. Effect of altered matrix proteins on quasilinear viscoelastic properties in transgenic mouse tail tendons. Ann Biomed Eng 2003;31:599–605. [38] Hiester ED, Sacks MS. Optimal bovine pericardial tissue selection sites. II. Cartographic analysis. J Biomed Mater Res 1998;39: 215–21.
[39] Hiester ED, Sacks MS. Optimal bovine pericardial tissue selection sites. I. Fiber architecture and tissue thickness measurements. J Biomed Mater Res 1998;39:207–14. [40] Sacks MS, Chuong CJ. Orthotropic mechanical properties of chemically treated bovine pericardium. Ann Biomed Eng 1998;26: 892–902. [41] Waldman SD, Sacks MS, Lee JM. Boundary conditions during biaxial testing of planar connective tissues: Part II: Fiber orientation. J Mater Sci Lett 2002;21:1215–21. [42] Cameron GJ, Alberts IL, Laing JH, Wess TJ. Structure of type I and type III heterotypic collagen fibrils: an X-ray diffraction study. J Struct Biol 2002;137:15–22.
Molecular orientation of collagen in intact planar connective tissues under biaxial stretch Jun Liao a, Lin Yang b, Jonathan Grashow a, Michael S. Sacks a
a,*
Engineered Tissue Mechanics Laboratory, Department of Bioengineering and the McGowan Institute for Regenerative Medicine, University of Pittsburgh, Room 234, 100 Technology Drive, Pittsburgh, PA 15219, United States b National Synchrotron Light Source (NSLS), Brookhaven National Laboratory, Upton, New York 11973, United States Received 17 August 2004; received in revised form 8 September 2004; accepted 8 September 2004
Abstract Understanding of the mechanical behavior of collagenous tissues at different size scales is necessary to understand their physiological function as well as to guide their use as heterograft biomaterials. We conducted a first investigation of the kinematics of collagen at the molecular and fiber levels under biaxial stretch in an intact planar collagenous tissue. A synchrotron small angle X-ray scattering (SAXS) technique combined with a custom biaxial stretching apparatus was used. Collagen fiber behavior under biaxial stretch was then studied with the same specimens using small angle light scattering (SALS) under identical biaxial stretch states. Both native and glutaraldehyde modified bovine pericardium were investigated to explore the effects of chemical modification to collagen. Results indicated that collagen fiber and molecular orientation did not change under equibiaxial strain, but were observed to profoundly change under uniaxial stretch. Interestingly, collagen molecular strain initiated only after 15% global tissue strain, potentially due to fiber-level reorganization occurring prior to collagen molecule loading. Glutaraldehyde treatment also did not affect collagen molecular strain behavior, indicating that chemical fixation does not alter intrinsic collagen molecular stiffness. No detectable changes in the angular distribution and D-period strain were found after 80 min of stress relaxation. It can be speculated that other mechanisms may be responsible for the reduction in stress with time under biaxial stretch. The results of this first study suggest that collagen fiber/molecular kinematics under biaxial stretch are more complex than under uniaxial deformation, and warrant future studies. 2004 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Collagen orientation; Small angle X-ray scattering; Small angle light scattering; Stress relaxation; Bovine pericardium
1. Introduction Collagen type I is the most abundant structural protein in the connective tissues. In order to bear tensile loads in an efficient manner type I collagen consists of a complex, structural hierarchy at various size scales [1,2]. The fundamental structural unit of collagen is
*
Corresponding author. Tel.: +1 412 235 5146; fax: +1 412 235 5160. E-mail address: [email protected] (M.S. Sacks).
the collagen molecule (diameter / 1.5 nm), which consists of triple helix amino acid chains [3]. The deposition of heavy metal stains in electron micrographs at the ends of the molecules produce dark bands commonly seen in collagen fibrils. The length of each collagen molecule is 4.4 times that of the period of the striation length D observed in electron micrographs. Along the molecular axis a gap of 0.6D is left between the ends of successive molecules. Through intrafibrillar bonding collagen molecules aggregate into collagen fibrils (/ 50 nm) in the quarter stagger form, with a D-period ranging from 64 to 68 nm [4,5]. Fibrils then form collagen fibers
1742-7061/$ - see front matter 2004 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2004.09.007
46
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
(/ 1 lm) via interfibrillar interactions mediated by proteoglycans [6]. The mechanical properties of collagenous connective tissues are largely determined by the unique structure of their constituent collagens. When the structure of these collagens are altered as a result of disease or chemical modification, tissue mechanical properties will be profoundly affected. It is thus important to understand the structure-function relationship in both native and biologically derived collagenous tissues. Moreover, constitutive (stress–strain) modeling of connective tissues is required in many physiological, surgical, tissue engineering, and medical device applications [5,7]. To predict the mechanical behavior of connective tissue, structural and compositional information are needed [8,9]. Clearly, knowledge of collagen kinematics in the loading process would directly benefit the structural based modeling. The mechanics of planar soft tissues are complex; loading induces straightening of highly crimped fibers, fiber rotation and stretch, and intra- and inter-molecular deformations. For example, structural models of skin have demonstrated that it is extreme extensibility (>50%) can be explained by reorientation of the collagen fibers alone [10,11]. We have developed a method that combines small angle light scattering (SALS) technique and biaxial stretch for measuring fiber kinematics in two dimensional soft collagenous tissues [9,12,13]. SALS provided a straightforward and rapid method to quantify the fiber orientation [14–16]. Incorporation of planar biaxial stretch and information of fiber angular distribution have provided much insight into the local fiber kinematics. For example it was found that fiber kinematics are highly tissue specific and fiber deformation is generally non-affine [9,13]. However, SALS can only provide orientation information at the fiber level. It is still unclear how fiber rotations and fiber uncrimping relate to the underlying collagen molecular rotations and stretch. It is well known that the X-ray scattering is an ideal tool to characterize the ultrastructure of collagen, such as molecule packing, fibril spacing, and orientation. Small angle X-ray scattering (SAXS) has been widely used to investigating the collagen structure [17–21]. Molecular strain in the intact collagen fibril under uniaxially stretched tendon/ligament have been widely reported [2,22–27]. In these studies tension-induced increases in the D-period were observed, with D-period strain proportional to the average strain of the specimen. A limitation of the above studies is that they are focused on uniaxially aligned tissues. In reality, most connective tissues are loaded in two and three dimensions and can have complex multi-dimensional fiber structures. Under multi-dimensional loading states collagen fiber deformations are more complex, and can include coupled effects of large rotations and non-affine deformations.
The above facts underscore the importance of understanding collagen molecule deformation and reorientation in intact issues under multi-dimensional strain states. Moreover, such understanding can lead to an improved understanding of the degradative processes of current biologically derived heterograft collagenous tissues. For example, molecular-level information can aid the development of novel tissue chemical treatment methods by providing a means to assess intrinsic collagen damage, independent of other damage mechanisms such as layer delamination. The objective of the current study was to initiate our investigations of collagen kinematics at the molecular level within an intact tissue under biaxial stretch using SAXS. We also compared the results of SAXS and SALS on the same tissue specimens in an attempt to link multi-axial kinematics at both the fiber and molecular levels. Both uniaxial and equibiaxial strain states were applied; the later used to restrict fiber rotations [7] so that the effects of fiber straightening alone could be studied.
2. Methods 2.1. SAXS biaxial stretch system A biaxial stretching device was custom made for the synchrotron X-ray beamline (X21 SAXS endstation) in NSLS of Brookhaven National Laboratory (Fig. 1). Two pairs of orthogonally positioned stepper motors (Parker automation, Stepper Motors VS23) provided motion to the four lead screws. Each pair was mounted in opposition and each motor was driven by a microstepping drive board (Parker Automation, Microstepping drive E-AC). The four microstepping drives were connected to motor controller card (Galil Motion Control, DMC-1800) via an interconnection module (Galil Motion Control, ICM-1900). The water-tight sample chamber was constructed using an aluminum alloy. Inside the sample chamber, there were two pairs of orthogonally positioned tissue holders with holes for the sutures. Each holder was mounted on a lead screw and a guide shaft. The front side of sample chamber was covered by transparent polyvinyl chloride (PVC) glass for strain monitoring purposes. A hollow stage right behind the specimen area was raised from the rear wall of the sample chamber. Another PVC glass plate covered the hollow stage. Two holes of 1 mm and 3 mm diameter were made on the front and rear PVC glass, respectively. Kapton film, used as X-ray windows, was pasted over the two holes with epoxy glue. A CCD camera (National Instrument, IMAQ CCD) perpendicular to the beamline was used to track four centrally located graphite markers on tissue with the help of a 45 mirror, in which a centered hole
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
Fig. 1. (a) A schematic illustration of SAXS biaxial stretch system, (b) SAXS biaxial stretch setup in X21 station of NSLS in BNL. Labels: (1) Specimen chamber transparent to both X-ray and light; (2) Guide shaft; (3) Lead screw; (4) Tissue holder; (5) Concave stage; (6) Motor; (7) Suture; (8) Tissue in PBS solution; (9) 45 mirror; (10) CCD camera; and (11) X-ray detector.
(/ = 3 mm) was made to allow the X-ray passing through. A Labview (National Instrument, Version 6) program was written to control the motion of motor and track the tissue strain at the same time. 2.2. Sample preparation Native bovine pericardium (BP) and 0.625% glutaraldehyde in phosphate buffered saline treated bovine pericardium (GLBP) specimens were used as the representative planar collagenous tissues. Tissue specimens (n = 5 for each group) were cut into 25 mm square samples. Four graphite markers (0.3 mm diameter) were affixed to the center of each specimen with cyanoacrylate adhesive in a 5 · 5 mm square pattern. The specimen was then mounted on the tissue holders with polyester suture (size 0) via four loops on each side. In the experiment, tissue was immersed in phosphate buffered saline at room temperature. 2.3. SAXS measurements SAXS data was collected at beamline X21 at NSLS ˚ X-rays. The sample-to-detector distance with 1.55 A was 1.33 m and the beam size at the sample was
47
0.4 mm. The diffraction patterns are recorded with a Mar CCD detector. The D-period of the collagen fibrils was extracted from the reciprocal spacing between the Bragg peaks produced by these structures. Specimens were mounted in the unstretched state, then stretched just enough until there was no observable sagging. This state was used as reference state for strain calculations. Both equibiaxial stretch and strip biaxial stretch (i.e. where the stretch along one axis is increased while the orthogonal axis is held at zero strain) were applied. In the equibiaxial stretch studies, specimens were stretched using a sequence of 5% equibiaxial strain increments. At each strain level, SAXS patterns were recorded. In the strip stretch studies, specimens were stretched in one direction while maintaining the other direction at 0% strain. A sequence of strip strains with a 5% increment was applied and SAXS pattern recorded. The accuracy of the stretch measurements was within ±0.001. For both BP and GLBP, the data was collected by averaging two reading (120 s for each reading). Scattering patterns were then analyzed with custom written program. In order to correctly calculate the averaged scattering profile and angular intensity distributions, masks (i.e. a fan or a ring shape) were used to block off unusable portions of the intensity data. A fan centered at beam center were used for quantifying the scattering profile and a ring enclosing fifth diffraction maximum were used for quantifying angular distribution. Details of the SAXS technique are given in the Appendix A. 2.4. SALS based fiber-level kinematic measurements A SALS system combined with a custom made biaxial device was used to determine the collagen fiber orientation in the same GLBP after SAXS testing. A detailed description of SALS protocol and effectiveness on fiber architecture evaluation has been previously presented [12,13,15]. Briefly, a monochromatic laser Helium–Neon laser light (k = 632.8) is passed through a tissue specimen. Since the wavelength of light is within an order of magnitude of the diameter of collagen and elastic fibers, small angle scattering pattern represents information of fiber orientation. The SALS-biaxial stretch system is similar with the system used in SAXS [12]. The biaxial stretching device consisted of two pairs of lead screws driven by two computer-controlled stepper motors. Specimen mounting method and testing protocol were as same as SAXS. A 0.8 mm laser beam size was used to assure that the area tested by SAXS was included within the light beam envelope. At each strain level, SALS pattern were recorded and analyzed using custom software. To facilitate comparisons with the SAXS data, statistical distribution function of the angular distribution of collagen fibers or molecules, R(h), was defined as [9]
48
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
IðhÞ RðhÞ ¼ Ph¼p=2 h¼p=2 IðhÞDh
ð1Þ
where I(h) represents the angular intensity distribution.
3. Results 3.1. SAXS D-period strain measurements In SAXS patterns of BP and GLBP we found that in the initial 0% strain state, diffractions demonstrated a relatively weak molecular orientation (Fig. 2a and d). For BP, we found that both stretch protocols induced a smooth change of molecular orientation, with only one major orientation direction observed (Fig. 2b and c). Interestingly, for GLBP we observed multiple major molecular orientation peaks, probably a result of multiple fiber populations in this specimen (Fig. 2e and f). Representative scattering profiles (intensity-q spectra line) along the orientation peak for GLBP demonstrated a change in the D-period with applied strip stretch (Fig. 3). D-period changes for both BP and GLBP along the strip stretch direction (90) and along the 120 direction for the same sample were quantified (Table 1, Fig. 4). Interestingly, we found that the BP D-period did not increase until the 20% global strip strain reached (Fig. 4). In contrast, GLBP D-period began to increase from the 10% global strip strain, which was quicker than BP (Fig.
Fig. 3. GLBP D-period diffraction spectra under different strip stretches, indicated on the right hand side of the plot, and shifted upward for clarity. Peak values of diffraction spectra were found to change with the increase of strip strain for strains >10%, indicating the molecular strain did not occur until at least 10% global tissue strain.
4). To estimate the local molecular-level loading to collagen in relation to the global tissue strains, we computed the D-period strain/global tissue strain ratio.
Fig. 2. Typical SAXS patterns of native BP and glutaraldehyde fixed bovine pericardium (GLBP) under biaxial and strip stretch, (a–c) BP in the initial state, an equibiaxial strain of 20%, and a maximum strip strain of 30%, respectively, (d–f) GLBP in the initial state, equibiaxial strain of 10%, and under a maximum strip strain of 25%, respectively. Note for BP there was only one observable orientation peak, while for the GLBP four peaks could be discerned.
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
49
Table 1 ˚ ) of native bovine pericardium and glutaldehyde treated bovine pericardium with applied global strip biaxial stretch Representative D-period (A
BP (strip stretch direction, 90) GLBP (strip stretch direction, 90) GLBP (120 direction)
0%
5%
10%
15%
20%
25%
30%
654.50 654.50 654.50
654.50 654.50 654.50
654.50 654.50 654.50
654.50 675.61 661.39
654.50 682.95 668.42
668.42 690.46 668.42
675.61 – –
Strain of fibril D-period (%)
Note: GLBP did not reach 30% global strip strain.
6
BP along vertical stretch direction
5 4 3 2
Slope = 0.32%/%
1 0 0
5
10
15
20
25
30
35
Strain of fibril D-period (%)
Global tissue strain along stretch direction (%) 6
GLBP along vertical stretch direction
5 4 3
Slope = 0.35%/%
2 1 0 0
5
10
15
20
25
30
35
Global strain along stretch direction (%) Fig. 4. Relationship between D-period strain and global strip strain for are representative specimen. Top curve showed D-period change in BP measured along the vertical strip stretch direction. Bottom curve showed D-period changes in the same GLBP measured along the vertical strip stretch direction. Variations of the five specimens tested were small and were within ±0.1% of the mean D-period fibril strain at each global tissue strain tested.
The relative change of the BP D-period was 0.32%/% and 0.35%/% for GLBP, indicating little differences between these two specimens. 3.2. Effects of biaxial stretch on collagen molecule rotational kinematics With increasing equibiaxial strain of up to 20%, collagen molecules were observed to exhibit an increased degree of orientation (Fig. 5). The angular molecular distribution was relatively smooth and major orientation exhibited a peak at 125, and included a secondary peak at 100 (Fig. 5a). In comparison, when a strip stretch was applied in the vertical (90 orientation)
Fig. 5. Representative angular distribution of D-period diffraction intensity (a) when BP was under different equibiaxial strains, (b) when BP was under different strip strains. Strip stretch was applied along the vertical (i.e. 90) direction. SAXS intensity distribution at different global strain level was shifted upwards for demonstration purpose. The dotted lines indicate the angular locations of the main orientation peaks at maximum stretch.
direction, a peak occurred at 105, with a secondary peak at 80 (Fig. 5b). Moreover, the uniaxial strip stretch distribution was noticeably narrower (Fig. 5b) compared to the biaxial stretch distribution (Fig. 5a). When chemically treated similar reorientation behavior was observed for the GLBP specimens (Fig. 6). In the initial state, the GLBP specimen exhibited a weak orientation peak at 130. Under equibiaxial strain two major orientation peaks at 40 and 130 formed (Fig. 6a). Upon further stretch to 10% strain these orientation peaks split into three peaks at 50, 105, and 135 (Fig. 6a). Under vertical (i.e. 90 orientation) strip stretch, the orientation peak splitting was more prominent with four peaks occurring at 60, 100, 115, and 130 at 25% strip strain (Fig. 6b). 3.3. Collagen fiber kinematics SALS results for the same GLBP specimen exhibited a major orientation peak at 130 (Fig. 7a). This was
50
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
When the equibiaxial strain increased to 10%, peak intensity was found still around 130. Under strip (90 orientation) stretch fiber orientation peak gradually switched from 130 to 90 (Fig. 7b), indicating fibers gradually aligned toward the strip stretch direction. This observation is consistent with the molecular orientation range of 40–140 shown by SAXS (Fig. 6b). 3.4. Collagen orientation and D-period in stress relaxation
Fig. 6. Angular distribution of D-period diffraction intensity (a) when GLBP was under different equibiaxial strains, (b) when GLBP was under different strip strains. Strip stretch was applied along the vertical direction. SAXS intensity distribution at different global strain level was shifted upwards for demonstration purpose. The dotted lines indicate the angular locations of the main orientation peaks at maximum stretch.
Fig. 7. Representative angular distribution of scattered light intensity (a) when GLBP was under different equibiaxial strains, (b) when GLBP was under different strip strains. Strip biaxial stretch was applied along the vertical direction. SALS intensity distribution at different global strain level was shifted upwards for demonstration purpose. The dotted lines indicate the angular location of the main orientation peak at minimum and maximum stretch.
consistent with collagen molecular orientation (130) observed by SAXS under equibiaxial strain (Fig. 6a).
To investigate time-course changes in orientation under a fixed kinematic state, we investigated changes in molecular orientation during biaxial stretch relaxation. A native BP specimen was loaded and held at a 30% strip strain level. D-period diffraction patterns were recorded every 4 min during the stress relaxation process over a total time period of 80 min. We observed that the angular distribution of D-period diffraction did not change in the stress relaxation process (Fig. 8), suggesting that the collagen molecular orientation was constant. Moreover, quantification of diffraction spectra showed that D-period also kept constant in the relaxation process (Fig. 8b). When the specimen was unloaded the major orientation at 30% strip strain disappeared (Fig. 8a) and D-period fell back to initial value (Fig. 8b).
Fig. 8. (a) For BP the angular distribution of the D-period diffraction intensity was found not to change with stress relaxation process under 30% fixed global strip strain. However, the intensity distribution changed back to its initial state when the tissue was unloaded, (b) D-period strain at 30% fixed global strip strain was also found not to change in the time course of stress relaxation (240 s to 80 min), suggesting that the relaxation of the load was not due to collagen.
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
51
4. Discussion 4.1. Collagen molecular and fiber orientations The present study represents a first look at the molecular and fiber level kinematical behavior of a planar collagenous tissue under biaxial stretch. The molecular level orientation changes revealed by SAXS were generally consistent with the fiber level orientation changes revealed by SALS. These results are consistent with those found by Kronick and Buechler [15], who found close agreement in the collagen molecular and fiber orientations using SALS and SAXS for calf skin under uniaxial loading conditions. However, in the present study we found that the kinematics of collagen molecules in intact planar connective tissues were more complex than in previous uniaxial studies. This is most easily seen in the SAXS molecular orientation responses between the equibiaxial and strip biaxial strain states (Figs. 5 and 6, respectively). Firstly, we noted that in the unstrained state the tissue the molecules had only a very modest degree of alignment (Figs. 5a and 6a). Upon equibiaxial stretch, molecular orientation developed prominent peaks. This can be explained that in the initial unloaded state the collagen fibers are very undulated (i.e. crimped) [9,28,29] which would have a ‘‘randomizing’’ effect on the overall orientation. As the tissue is stretched equibiaxially, the fibers are kinematically restricted and cannot rotate [9]. Thus, the fibers can only straighten and lose their initial crimp. This results in an improved molecular alignment with increasing equibiaxial stretch (Figs. 5a and 6a). Note too that under strip stretch (where the vertical axis is stretched but the horizontal axis is kept at zero strain), similar qualitative observations occurred (Figs. 5b and 6b). However, in this case the resulting distributions were narrower, indicative of a great degree of molecular alignment. It is also interesting to note that in both the native (Fig. 4a) and chemically modified tissues (Fig. 4b) there was a delay in molecular stretch with tissue strain. This phenomena has been observed previously in tendon [25,26,30]. This delay in molecular strain is generally attributed to the initial fiber reorientation that occurs in the toe region of the stress–strain curve, where the collagen fibers are subjected to negligible loading. Thus, no molecular-level strain occurs until the shortest collagen fibers become straightened and begin to bear load. A collagen crimp period of 25–30 lm has been observed for BP collagen fibers (e.g. Fig. 9), with an estimated amplitude of 10 lm. This explanation is thus consistent with the substantial degree of crimp found for native BP [31]. The slope of the D-period strain to global tissue strain would be 1.0%/% if all the tissue strain was uniformly transmitted to the molecular level. In the present
Fig. 9. Polarized photomicrograph of a BP specimen, displaying a collagen crimp period of 30 lm, crimp amplitude was estimated to be 15 lm. These structural results indicate that BP has a substantial degree of crimp which may be responsible for the differences found between fiber and molecular level stain behavior.
study this ratio was 0.33%/% (Fig. 4) was indicative that not all the global tissue strain is translated to the molecular level. The reason for this lower value is most likely due to the fact that, even under equibiaxial strain, not all the collagen fibers are loaded to the same local strain level as the tissue is loaded. That is, there is a distribution of collagen fiber straightening strain that has been theorized to produce the non-linear tissue stress– strain curve [9]. Thus, there will be an equivalent distribution of fiber strains, resulting in an average D-period strain to global tissue strain slope less than 1%/%. It is also possible that inter-fibrillar slippage may contribute to the lower slope. Clearly, additional studies are required to elucidate this phenomenon. 4.2. Effects of exogenous crosslinks After glutaraldehyde fixation collagen is crosslinked by aldehyde functionalities which stiffened the tissue and presumably limit the freedom of fibril realignment. Yet, we did not observe any differences in the D-period strain/global tissue strain ratio (Fig. 4). The relative change of the BP D-period was 0.32%/% and 0.35%/% for GLBP. These results suggest that chemical modification did not measurably affect collagen molecular stiffness. We have observed that chemical fixation main affects the low strain region of the stress–strain curve under biaxial stretch [32,33]. The results of the present study further suggest that the formation of exogenous crosslinks affects larger scale structures, and not the collagen molecules themselves. 4.3. Collagen molecular behavior in stress relaxation In the present study we found that in the time course 4–80 min no detectable changes in the angular distribution and D-period strain occurred during stress relaxation (Fig. 8). This observation is consistent with similar
52
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
studies by Purslow et al. on rat skin [24]. Moreover, according to Mosler et al. [34] when a constant 6.5% grip-to-grip strain was applied to rat tail tendon collagen fibers an initial increase in D-period occurs. However, this value decreases back to the original 67 nm value at a faster rate compared to the rate of stressrelaxation in the same specimen [34,35]. Closer inspection of MoslerÕs data indicated that the D-period decreased in the first 200 s, but did not change from 200 to 400 s. In our study the time resolution is 240 s, which is the time required to obtain the SAXS image. We found no detectable changes in planar connective D-period in over the time course of 240 s to 80 min in the stress relaxation. Our results were also consistent with PurslowÕs [24] conclusion that stress relaxation behavior is not caused by time-dependent reorientation of collagen fibers. Other mechanisms for the relaxation in collagenous tissues include (i) the matrix surrounding the collagen fibers and (ii) time-dependent effects at the interface between fiber and matrix [24]. One might overlook the contribution of matrix based on concept that significant forces are borne by the collagen fibers [24]. However, with the importance of interfibrillar interactions mediated by proteoglycans gradually becoming better appreciated [36,37] additional studies are needed to elucidate the influence of interfibrillar matrix. 4.4. Study limitations and future directions The primary limitation of the current study is that the present SAXS biaxial apparatus does not allow for simultaneous force measurements. Without force information, it was not possible to determine the actual molecular stiffness. However, we felt that a feasibility study to determine if there were indeed additional effects of encountered in biaxial loading on the collagen fiber kinematics was warranted before undertaking the additional significant cost and time to implement force measurements. Given the intriguing results of the present study, we are currently embarking on modifications to the biaxial test system to perform simultaneous force measurements. This will allow us to determine critical mechanical parameters for collagen, such as the modulus, under biaxial stretch. Moreover, more sophisticated time dependent experiments (such as creep) will be able to be preformed. In our experimental design, length of X-ray beam passing through both water and tissue is 4 mm. This path length requires a relatively long time (240 for both BP and GLBP) to obtain the scattering pattern, which inevitably influenced our ability to get higher time-resolution when analyzing the stress relaxation. Finally, the current specimens were randomly chosen from the pericardial sac. This was done intentionally to allow for an unbiased examination of collagen fiber kinematics.
However, given the natural variability in collagen structure of the pericardial sac [38,39] future studies should include presorting [40,41]. 4.5. Summary We conducted a first study of the kinematics of collagen at the molecular and fiber levels under biaxial stretch in an intact planar collagenous tissue. Our key observations include: • Collagen molecular strain in BP did not increase immediately with global tissue strain, but only began to increase only after 10–20% tissue strain, suggesting that substantial fiber-level reorganization occurs prior to collagen molecular loading under biaxial stretch. • The resulting SAXS intensity distributions demonstrated greater alignment under strip biaxial stretch compared to the equibiaxial stretch state. • No detectable changes in the angular distribution and D-period strain were found in the up to 80 min of stress relaxation. These final results suggest that viscoelastic mechanisms in collagenous tissues are not due to molecular-level changes, but results from changes at larger structural scales. • Chemical fixation indicated no change in molecular stiffness (we believe this is the first observation of this). Clearly, additional studies will be required to elucidate the observed phenomenon.
Acknowledgment This work is supported by NIH grant HL63026 and an Established Investigator Award of the American Heart Associate to MSS. Research carried out (in whole or in part) at the National Synchrotron Light Source, Brookhaven National Laboratory, which is supported by the US Department of Energy, Division of Materials Sciences and Division of Chemical Sciences, under Contract No. DE-AC02-98CH10886. The authors would like to thank Daniel Hildebrand and Thomas Gilbert for invaluable discussions.
Appendix A A SAXS pattern from a group of parallel aligned collagen fibrils (along vertical direction) consist of two orthogonal diffraction series: (i) In the direction that collagen fibril aligned, axial D-period works as a diffraction spacing and generates meridional series of up to 30 relatively sharp diffraction centered on the fibril axis and
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
the beam center; (ii) In the direction perpendicular to collagen fibrils, a more diffuse diffraction (Equator series) can be observed and the first peak corresponds to an intermolecular packing distance of 1.5 nm [23,24, 42]. If the diffraction is caused by an angularly dispersed population of collagen fibrils, angular dispersion of diffraction series will be observed. Since meridional series is caused by fibril axial spacing. The high intensity on the angular dispersion arc represents the major fibril orientation (i.e. collagen molecule orientation) [24]. The diffraction maxima occurring at angles 2h to the X-ray beam (Bragg reflections) is given by equation: 2 sin(h) = kn/d, where n is the order of the reflection, k, is X-ray wavelength, and d is the spacing. Scattering angle 2h and spacing d were reciprocally related. Therefore, using specific calibration, values of d can be determined from diffraction pattern. Scattering vector ˚ 1), defined as q = 4p sin(h)/k, was used in this q (unit, A study. Therefore, we have q = 2p/d. Value of 2p/d is the distance between intensity peaks and can be accurately calculated using linear fitting (peaks from 3rd to 9th order used in this study). Spacing d can then be calculated by equation d = 2p/q.
References [1] Kastelic J, Baer E. Deformation of tendon collagen. In: Vincient JF, Currey JD, editors. The Mechanical Properties of Biological Materials. Great Britain: Society for Experimental Biology; 1980. p. 397–433. [2] Fratzl P, Misof K, Zizak I, Rapp G, Amenitsch H, Bernstorff S. Fibrillar structure and mechanical properties of collagen. J Struct Biol 1998;122:119–22. [3] Chapman JA, Hulmes DJS. Electron microscopy of the collagen fibril. In: Ruggeri A, Motto PM, editors. Ultrastructure of the connective tissue matrix. Boston: Martinus Nijhoff; 1984. p. 1–33. [4] Hodge AJ, Petruska JA. Recent studies with the electron microscope on ordered aggregates of the tropocollagen molecule. Aspects of protein chemistry. London: Academic Press; 1963. [5] Fung YC. Biomechanics: Mechanical properties of living tissues. New York: Springer Verlag; 1993. [6] Kastelic J, Galeski A, Baer E. The multicomposite structure of tendon. Connect Tissue Res 1978;6:11–23. [7] Sacks MS. Biaxial mechanical evaluation of planar biological materials. J Elasticity 2000;61:199–246. [8] Wren TA, Carter DR. A microstructural model for the tensile constitutive and failure behavior of soft skeletal connective tissues. J Biomech Eng 1998;120:55–61. [9] Sacks MS. Incorporation of experimentally derived fiber orientation into a structural constitutive model for planar collagenous tissues. J Biomech Eng 2003;125:280–7. [10] Lanir Y. Biaxial stress–strain relationship in the skin. Israel J Technol 1979;17:78–85. [11] Lanir Y. A structural theory for the homogeneous biaxial stress– strain relationships in flat collageneous tissues. J Biomech 1979;12:423–36. [12] Billiar KL, Sacks MS. A method to quantify the fiber kinematics of planar tissues under biaxial stretch. J Biomech 1997;30:753–6.
53
[13] Sacks MS, Gloeckner DC. Quantification of the fiber architecture and biaxial mechanical behavior of porcine intestinal submucosa. J Biomed Mater Res 1999;46:1–10. [14] Chien JCW, Chang EP. Small-angle light scattering of reconstituted collagen. Macromolecules 1972;5:610–7. [15] Kronick PL, Buechler PR. Fiber orientation in calfskin by laser light scattering or X-ray diffraction and quantitative relation to mechanical properties. J Amer Leather Chem Assoc 1986;81: 221–9. [16] Sacks MS, Smith DB, Hiester ED. A small angle light scattering device for planar connective tissue microstructural analysis. Ann Biomed Eng 1997;25:678–89. [17] Farkasjahnke MVS. Small-angle X-ray diffraction studies on rattail tendon. Acta Physiol Acad Sci 1965;28:1–17. [18] Bigi A, Incerti A, Leonardi L, Miccoli G, Re G, Roveri N. Role of the orientation of the collagen fibers on the mechanical properties of the carotid wall. Boll Soc Ital Biol Sper 1980; 56:380–4. [19] Suarez G, Oronsky AL, Bordas J, Koch MH. Synchrotron radiation X-ray scattering in the early stages of in vitro collagen fibril formation. Proc Natl Acad Sci USA 1985;82: 4693–6. [20] Aspden RM, Bornstein NH, Hukins DW. Collagen organisation in the interspinous ligament and its relationship to tissue function. J Anat 1987;155:141–51. [21] Fratzl P, Daxer A. Structural transformation of collagen fibrils in corneal stroma during drying. An X-ray scattering study. Biophys J 1993;64:1210–4. [22] Misof K, Rapp G, Fratzl P. A new molecular model for collagen elasticity based on synchrotron X-ray scattering evidence. Biophys J 1997;72:1376–81. [23] Wess TJ, Purslow PP, Kielty CM. X-ray diffraction studies of fibrillin-rich microfibrils: effects of tissue extension on axial and lateral packing. J Struct Biol 1998;122:123–7. [24] Purslow PP, Wess TJ, Hukins DW. Collagen orientation and molecular spacing during creep and stress-relaxation in soft connective tissues. J Exp Biol 1998;201(Pt l):135–42. [25] Sasaki N, Odajima S. Stress–strain curve and YoungÕs modulus of a collagen molecule as determined by the X-ray diffraction technique. J Biomech 1996;29:655–8. [26] Sasaki N, Odajima S. Elongation mechanism of collagen fibrils and force–strain relations of tendon at each level of structural hierarchy. J Biomech 1996;29:1131–6. [27] Sasaki N, Shukunami N, Matsushima N, Izumi Y. Time resolved X-ray diffraction from tendon collagen during creep using synchrotron radiation. J Biomech 1999;32:285–92. [28] Fentie IH, Alien DJ, Schenck MH, Didio LJA. Comparative electron microscopic study of bovine, porcine, and human parietal pericardium, as materials for cardiac valve bioprostheses. J Submicrosc Cytol 1986;18:53–65. [29] Shabetai R. The pericardium. New York: Grune and Stratton; 1981. [30] Gathercole LJ, Keller A. X-ray diffraction effects related to superstructure in rat tail tendon collagen. Biochim Biophys Acta 1978;535:253–71. [31] Sun W, Sacks M, Fulchiero G, Lovekamp J, Vyavahare N, Scott M. Response of heterograft heart valve biomaterials to moderate cyclic loading. J Biomed Mater Res 2004;69A:658–69. [32] Sacks MS. A structural constitutive model for chemically treated planar connective tissues under biaxial loading. Comput Mech 2000;26:243–9. [33] Sun W, Sacks MS, Sellaro TL, Slaughter WS, Scott MJ. Biaxial mechanical response of bioprosthetic heart valve biomaterials to high in-plane shear. J Biomech Eng 2003;125:372–80. [34] Mosler E, Folkhard W, Knorzer E, Nemetschek-Gansler H, Nemetschek T, Koch MH. Stress-induced molecular rearrangement in tendon collagen. J Mol Biol 1985;182:589–96.
54
J. Liao et al. / Acta Biomaterialia 1 (2005) 45–54
[35] Folkard W, Geercken W, Knorzer E, Mosler E, NemetschekGansler H, Nemetschek T, et al. Structural dynamic of native tendon collagen. J Mol Biol 1987;193:405–7. [36] Silver FH, Kato YP, Ohno M, Wasserman AJ. Analysis of mammalian connective tissue: relationship between hierarchical structures and mechanical properties. J Long Term Eff Med Implants 1992;2:165–98. [37] Elliott DM, Robinson PS, Gimbel JA, Sarver JJ, Abboud JA, lozzo RV, et al. Effect of altered matrix proteins on quasilinear viscoelastic properties in transgenic mouse tail tendons. Ann Biomed Eng 2003;31:599–605. [38] Hiester ED, Sacks MS. Optimal bovine pericardial tissue selection sites. II. Cartographic analysis. J Biomed Mater Res 1998;39: 215–21.
[39] Hiester ED, Sacks MS. Optimal bovine pericardial tissue selection sites. I. Fiber architecture and tissue thickness measurements. J Biomed Mater Res 1998;39:207–14. [40] Sacks MS, Chuong CJ. Orthotropic mechanical properties of chemically treated bovine pericardium. Ann Biomed Eng 1998;26: 892–902. [41] Waldman SD, Sacks MS, Lee JM. Boundary conditions during biaxial testing of planar connective tissues: Part II: Fiber orientation. J Mater Sci Lett 2002;21:1215–21. [42] Cameron GJ, Alberts IL, Laing JH, Wess TJ. Structure of type I and type III heterotypic collagen fibrils: an X-ray diffraction study. J Struct Biol 2002;137:15–22.