How a radial focal incision influences the internal shear distribution in articular cartilage with respect to its zonally differentiated microanatomy

Abstract

Articular surface fibrillation and the loss of both transverse interconnectivity and zonal differentiation are indicators of articular cartilage (AC) degeneration. However, exactly how these structural features affect the load-redistributing properties of cartilage is still poorly understood. This study investigated how a single radial incision made to varying depths with respect to the primary zones of AC influenced its deformation response to compression. Three depths of incision were applied to cartilage-on-bone tissue blocks: one not exceeding the transition zone; one into the mid-radial zone; and one down to the calcified cartilage. Also included were non-incised controls. All samples were compressed to a near-equilibrium strain using a flat-faced indenter that incorporated a central relief channel within which the incision could be positioned lengthwise along the channel axis. Employing fixation under load followed by decalcification, the structural responses of the cartilage-on-bone samples were investigated. The study provides an analysis of the micro-morphological response that is characteristic of a completely normal cartilage-on-bone system but which contains a defined degree of disruption induced by the focal radial incision. The resulting loss of transverse continuity of the cartilage with respect to its zonally differentiated structure is shown to lead to an altered pattern of internal matrix shear whose intensity varies with incision depth.

Introduction

The ability of articular cartilage (AC) to redistribute stress away from the directly loaded region has been demonstrated previously in several experiments that captured the tissue response under load (Kääb et al. 1998, 2000; Glaser & Putz, 2002; Thambyah & Broom, 2006). These previous studies imaged the lateral bulging of the tissue under compression, such that the articular surface formed a smooth and curved transition between the directly loaded and adjacent non-directly loaded regions. In addition, it was also shown that a significant shear discontinuity formed primarily between the tangential layer and the mid to deep matrix in the directly loaded region and beyond, such that fibrils and cells were sheared laterally when the tissue was subject to direct compression.

This inter-zonal matrix shear was attributed to the strain-limiting tangential zone interacting with a less stiff radial zone which is, in turn, integrated with the subchondral bone via the zone of calcification (Thambyah & Broom, 2006, 2007). Other experiments and modelling studies have further demonstrated the variation in mechanical properties across the matrix zones associated with the depth-dependent structural inhomogeneity in cartilage (Schinagl et al. 1997; Wilson et al. 2006; Buckley et al. 2008; Shirazi & Shirazi-Adl, 2008; Halonen et al. 2013; Thambyah & Broom, 2013; Hosseini et al. 2014). Studies have also shown that when the tangential zone is artificially removed and the remaining matrix then compressed, the shear discontinuity disappears, thus indicating that there is no inter-zonal matrix shear (Glaser & Putz, 2002; Bevill et al. 2010) Further, in this modified state the ability of the matrix to redistribute stress away from the directly loaded regions is compromised.

Previous studies have also shown that in the healthy matrix the transverse interconnectivity in the collagen network of the radial zone (Broom & Silyn-Roberts, 1989; Chen & Broom, 1998, 1999; Broom et al. 2001) plays an important role in this mechanism of stress redistribution away from the directly loaded region (Thambyah & Broom, 2007; Bevill et al. 2010; Thambyah et al. 2011; Nickien et al. 2013). In mildly degenerate cartilage where the general matrix has undergone fibrillar destructuring and lost some of this transverse inter-connectivity, the lateral bulging of the tissue when compressed is significantly reduced and so too the intensity of inter-zonal matrix shear (Thambyah & Broom, 2007; Thambyah et al. 2011).

By coupling this shear visualisation with a specialised indentation technique that incorporates a central relief channel into which the non-directly loaded cartilage is allowed to bulge (Thambyah et al. 2009), the shear discontinuity in the boundary region between the directly loaded region and bulge is greatly accentuated and thus becomes a visual indicator of subtle alterations in the cartilage matrix response to loading due to either degeneration or other modes of structural disruption. It was found that in progressing from mild, to moderate, to severe levels of degeneration, the shear bands varied in extent in the channel bulge region from minor to non-existent, respectively (Thambyah et al. 2011). However, in the previous study the severely degenerate samples also exhibited a loss of structural integrity of the surface layer. It was therefore difficult to delineate easily the relative contributions to load redistribution arising from alterations in general matrix properties such as fibril interconnectivity and those arising from focal disruption of the tangential or superficial zone of the otherwise intact tissue.

Thus, the aim of this new study was to compare the compressive deformation response of completely normal cartilage-on-bone samples that ranged from completely intact with those containing a single radial cut varying in depth with respect to the zonally differentiated structure of the AC. Specifically, the aim was to investigate how such a focal disruption of varying severity might influence the internal shear response of the normal cartilage matrix at the microstructural level, the possibility being that intense shear deformation in the matrix has the potential to contribute to its eventual structural disruption.

Materials and methods

Samples and preparation

Bovine patellae were collected from freshly slain ∼2–3-year-old male animals, and stored frozen at −20 °C. Prior to experimentation, each patella was thawed in cold running water and the cartilage surface stained with Indian ink to confirm visually the absence of any surface fibrillation, this being indicative of a normal cartilage matrix in the bovine patella model (Hargrave-Thomas et al. 2013). A single block of cartilage-on-bone approximately 1.5 × 1.5 × 1.5 cm in size was sawn from the distal-lateral quadrant of each patella, and then equilibrated in 0.15 m saline at 4 °C for at least half an hour before being embedded in a custom-built holder using quick-setting dental plaster.

Using a thin blade, the sample was artificially disrupted with a radial incision commencing at the articular surface and penetrating to a defined depth that was controlled by varying the extent of protrusion of the blade from its holder, the aim being to achieve cuts ranging from superficial right down to the calcified cartilage. For each sample, an additional reference incision of equal depth was made in the region of cartilage distant from the intended site of indentation, repeatability of the incision depth having been confirmed previously using a separate group of non-test samples. The articular surface was then re-stained with Indian ink to reveal the incision site en face, after which a channel indenter of 8 × 7 mm with a 1-mm channel in its centre was positioned on the cartilage surface with the incision line located lengthwise and as central as possible within the channel (Fig. 1).

Fig. 1.

Macro view of a channel-indented cartilage-on-bone sample, including the channel indenter footprint. Note: h₁, h₂ = heights of the uncompressed regions; h₃, h₄ = heights of the directly compressed regions; h_b = height of the bulge region; h_r = distance from the tidemark to the reference incision tip.

Loading protocol

Each sample was maintained immersed in 0.15 m saline both prior to and during indentation. A compressive displacement of 0.6 mm per minute was applied to the indenter until a stress of 4 MPa was reached, after which this stress was held for 3 h while displacement was recorded, at the end of which the position of the indenter was maintained constant. The creep stress of 4 MPa was chosen both because it lies within the physiological range of stress for the human knee (Thambyah et al. 2005) and because it allows for comparison with other studies conducted in the authors’ laboratory (Thambyah et al. 2009, 2011; Bevill et al. 2010; Thambyah & Broom, 2013). The saline was drained and the sample fixed in situ in 10% formalin overnight. The sample was then removed from its holder and the bone trimmed away to leave only a thin supporting subchondral layer (2–3 mm thick), which was then decalcified in formic acid for 3–4 days. Using a scalpel, the sample was cut in half, creating a cross-sectional plane perpendicular to the channel axis from which 30-μm-thick cryosections were obtained.

Microstructural analysis

While maintained in a fully hydrated unstained state, the cryosections were cover-slipped and imaged at low magnification with brightfield optics to obtain the macroscopic level of deformation (Fig. 1), and then analysed microstructurally using differential interference contrast optics. AC thickness was quantified from measurements taken both outside the indenter imprint [(h₁ + h₂)/2] and in the directly loaded region [(h₃ + h₄)/2], and the average axial compressive strain determined. Also determined from the macro-level images was the distance between the tidemark and the incision tip both in the bulge (h_b) and in the uncompressed region (h_r).

Determination of incision depth with respect to zonal differentiation

Using the lines of chondron continuity as indicators of the primary fibrillar direction (Kääb et al. 2003), the tangential zone was defined as the region in which the chondrons were aligned tangential to the articular surface, the radial zone similarly defined as the region in which the chondrons were aligned radially. The transition zone was accordingly defined as the region in between the tangential and radial zones in which the chondrons tended to have a somewhat ambiguous alignment (Fig. 2). Although the blending nature of these inter-zonal boundaries made the depth measurements relatively imprecise, it is still possible to discuss the approximate depth of any incision with respect to this zonal differentiation. The incision depths were established from the reference incisions in the optical cryosections and divided into three categories post hoc: those not exceeding the transition zone and defined as ‘shallow’ (Fig. 2A); those entering the radial zone and defined as ‘mid-depth’ (Fig. 2B); and those extending right down into the zone of calcified cartilage and defined as ‘full-depth’ (Fig. 2C).

Fig. 2.

Images showing: (A) shallow incision not exceeding the transition zone; (B) mid-depth incision in radial zone; and (C) full-depth incision down to the subchondral bone. Scale bar: 100 μm.

Sample numbers

A total of 18 samples containing incisions of varying depths were indented as per the procedure described above. Six samples were classified as shallow, six as mid-depth and six as full-depth. An additional six samples without incisions were indented and utilised as controls. All values reported in the Results section are group or overall averages ± standard error of the mean. IBM® SPSS® Statistics version 21 was used to compare values between groups with different incision depths. An independent-samples Kruskal–Wallis test was performed to test the null hypothesis that the distribution of the variables across all of the groups was the same. The significance level used was 0.05. However, the statistical analysis was used only to support the visual microstructural data as the sample size was small and, consequently, the power of the test relatively low.

Results

General observations

All strain vs. time graphs showed a typical creep response of the tissue with a near-equilibrium strain being achieved after 3 h. The average axial strain was 0.51 ± 0.012 irrespective of incision depth, including the controls, and its distribution did not vary significantly between groups (P = 0.665). Thus, whether or not an incision to any depth was present in the bulge region had no demonstrable influence on the macro-level compressive response of the tissue in the directly loaded region.

The incision boundary in those samples with a shallow incision appeared as a straight line from the incision tip up to the articular surface (see dotted line in Fig. 3B). In contrast, this same boundary in the mid-depth and full-depth incision samples was curvilinear (Fig. 3C,D), with the latter having the highest degree of curvature.

Fig. 3.

Images showing the bulge region of: (A) the control; (B) shallow; (C) mid-depth; and (D) full-depth incision samples. The oblique-counter-oblique shear band pattern appearing both at the top and in the deep matrix of the bulge of the control and shallow incision samples are indicated with black arrows in (A) and (B). The distinct shear discontinuity extending into the bulge is indicated with white arrows in all images. The dashed white lines emphasise the straight cut boundary in the shallow incision sample (B) vs. its curvilinear form in both the mid-depth and full-depth incision samples (C and D, respectively). Scale bar: 500 μm.

The distance radially from the tidemark to the incision tip was compared between the undisturbed control site (h_r) and the bulge region (h_b) for the shallow and mid-depth incision samples (i.e. h_r − h_b in Fig. 1). The difference h_r − h_b between sample groups was significant (P = 0.026), with the average values being 366 ± 97 μm and −80 ± 130 μm for samples with a shallow vs. mid-depth incision, respectively. The positive values indicate that the matrix immediately beneath the incision has been compressed during indentation (Fig. 3B). Conversely, the negative values obtained for the mid-depth incision indicate that with this same degree of compression there has been an upwards extension of the general matrix (Fig. 3C).

Internal matrix shear development

Internal matrix shear can be inferred from both the presence of the characteristic oblique-counter-oblique bands in the bulge region and the shear discontinuity formed in the directly compressed region and extending into the bulge (Thambyah et al. 2009). The shear bands were observed only in the control and shallow incision samples, and were less extensive in the latter and were confined to the upper-mid and deep radial zones (see black arrows in Fig. 3A,B). Progressively higher magnification imaging of these shear bands revealed a well-defined structural response arising from a repeating and coordinated in-phase collapse of the radially aligned fibrosity (Fig. 4A–C), which was also mirrored in the morphological distortion of the chondrocyte columns (Fig. 4B). The shear bands in the deep radial zone (see lower black arrows in Fig. 3A,B), although less clear, seem also to have arisen from this same in-phase collapse (Fig. 5).

Fig. 4.

Higher magnification images of the shear bands in the bulge region of the control sample shown in Fig. 3A. Progressively higher magnification views of the boxed region in (A) are shown in (B) and (C). The crimp in the dark shear bands is clearly imaged along with the alternating direction of the fibrosity within the crimped shear bands. Scale bar: 10 μm.

Fig. 5.

(A) Faintly resolved alternating oblique-counter-oblique shear bands in the deep radial zone adjacent to the tidemark below the bulge region of the control sample in Fig. 3A; (B) an enlarged view of the boxed region in (A). Crimp formation is faintly visible in the ellipsed region marked in (B). The tidemark is indicated with the dotted line in (A). Scale bar: 50 μm.

In all samples the shear discontinuity extended from the directly compressed matrix into the bulge region (indicated by white arrows in Fig. 3). In the shallow incision samples and controls, this shear discontinuity became less distinct as it merged with the region containing the oblique shear bands (Fig. 6A,B). In the mid- and full-depth samples, the shear discontinuity faded just before it reached the incision (Fig. 6C,D). The shear discontinuity in the bulge region appeared sharper with increasing incision depth (c.f. Fig. 6C,D), which is consistent with there being a more acute change in direction of the lines of chondrocyte continuity (see enlargements in Fig. 6E and F, respectively).

Fig. 6.

Images showing the deformation fields revealed by the lines of chondrocyte continuity in the bulge region of (A) control, (B) shallow, (C) mid-, (D) full-depth incision samples. (B) Arrow 1 indicates the straight incision boundary; arrow 2 indicates the tangential strain-limiting layer; R_T indicates the region of transition matrix below the tangential layer; circled region indicates faint shear band development in upper bulge region. (E and F) Enlargements of boxed regions in (C and D), respectively. Scale bar: 100 μm.

Discussion

In this study, a channel indentation technique has been utilised as an experimental tool to investigate how the morphological response to compression of cartilage-on-bone is modified by a focal discontinuity created by a radial incision. With this incision positioned near centrally within the relief channel and aligned parallel to its axis, the cartilage matrix in this region is free to bulge upwards to an extent dictated by its own structural integrity. By varying the incision depth it has been possible to investigate the altered patterns of shear arising in the cartilage matrix from the intrinsic transverse constraint provided by each of its primary structural zones.

It is important to note that the channel indentation technique, by its very nature, does not capture the physiological response of the cartilage-on-bone system in vivo. However, the technique does provide a means of investigating the deformation behaviour of a fully intact healthy matrix (see Fig. 3A) and exploring how specific modes of departure from structural normality, whether induced artificially or by degeneration, can lead to an altered morphological response. Earlier studies have shown that neither shear bands nor a compression-induced shear discontinuity in the bulge region can be generated if the cartilage matrix is either severely destructured, i.e. where there is a substantial loss of transverse interconnectivity in the radial zone (Thambyah et al. 2011), or when a normal matrix has had its tangential zone artificially removed (Bevill et al. 2010).

The previously reported shear bands that form within the bulge region (Thambyah et al. 2009, 2011; Bevill et al. 2010) are assumed to arise from a complex mode of deformation produced by an indirect, downwards compression of the upwards-swelling matrix into the relief channel. This compression is a direct result of draw-down either by the intact tangential zone or by the partly intact transition zone (Fig. 3A,B), and is evident from the regular, repeating manner in which the interconnected, radially aligned fibrils collectively collapse via kink formation within the alternating shear bands (Fig. 4). The fact that these bands have an oblique-counter-oblique morphology is a direct consequence of the axisymmetric pull-down from both transverse directions, and this is termed a coherent response arising from a largely intact, zonally differentiated cartilage matrix.

The oblique shear bands were observed only in the intact samples and in those containing a shallow incision (Fig. 3), and this indicates that a component of transverse strain-limitation is essential for their development. Further, the constraint provided by the tidemark creates a similar but reduced shear response in the deep cartilage matrix (see lower black arrows in Figs 3A,B and 5). It is hypothesized that the primary transverse constraint is provided by those fibrils in the uppermost tangential layer being aligned parallel to the articular surface, as reflected in the control sample response (Fig. 3A). A lesser degree of constraint is also provided by the transition zone as indicated by the reduced amount of shear band formation in the shallow incision samples (Fig. 3B). This is consistent with the observations of Glaser & Putz (2002), who reported that the transition zone contains load-redistributing properties similar to the tangential zone but effective to a lesser extent. However, whether or not the transition zone then contributes to the shear band development when there is a fully intact tangential zone remains unresolved in the present study.

Because the transition zone contains fibrils that arch upwards into the tangential zone in both transverse directions (Benninghoff, 1925; McCall, 1968; Weiss et al. 1968), the matrix with a shallow incision is still able to provide axisymmetric pull-down but with less intensity than in the intact matrix (c.f. Fig. 6A,B). The straight cut boundary of the shallow incision (see arrow 1 in Fig. 6B) indicates that the matrix immediately adjacent to it is largely undisturbed. Such a straight incision boundary can only arise if there is little relative shear between the uppermost strain-limiting layer (see arrow 2 in Fig. 6B) and the transition matrix immediately below (see region marked R_T). Thus, there is little indirect force transmission from the directly indented cartilage into the region of matrix immediately above the incision tip and adjacent to this boundary. Hence, this region can be considered as ‘stress-shielded’ by the transition matrix immediately below the incision tip (see region marked with a circle in Fig. 6B). It can be argued that this transition region is able to develop a degree of transverse re-alignment of the fibrils sufficient to enable it to carry the bulk of the pull-down forces acting almost axisymmetrically from the incision tip. This then leads to some shear band formation in the upper radial zone (see circled region in Fig. 6B).

As a structural reference point, the schematic in Fig. 7 shows how the fibrillar architecture is differentiated across the three primary zones of the uncompressed AC. Based on the microscopic observations, the schematics in Fig. 8 illustrate the different fibrillar responses to compression following the incorporation of an incision to the three specified depths.

Fig. 7.

Simplified schematic illustrating the zonally differentiated fibrillar orientations in the uncompressed AC. The outlined region identifies that part of the cartilage matrix from which the schematics in Fig. 8 are derived.

Fig. 8.

Schematics showing fibrillar responses as inferred from the actual tissue responses imaged in Fig. 6. The red arrows indicate the relative magnitudes of the transverse forces that constrain lateral movement. (A) In the compressed control sample these forces are largest due to the strain-limiting properties of the intact tangential layer. (B) The lesser constraint is provided by the intact portion of the remaining transition zone. (C) The only lateral constraint remaining is that provided by fibril interconnectivity in the radial zone. (D) All constraint is now absent due to the full-depth radial incision.

Figure 8A illustrates how the intact tangential layer is able to translate the pull-down (and shear) from the adjacent directly compressed regions into an indirect compression of the bulging matrix, creating the axisymmetric shear bands. A similar mechanism produces a reduced amount of shear in the shallow incision sample (Figure 8B). By contrast, in the mid-depth incision sample (schematic Fig. 8C) the forces from the adjacent directly compressed regions act almost independently on each side of the incision. Its force trajectories are spread in a curvilinear manner over almost the entire region of matrix above the incision tip. There is now a significant degree of shear between the uppermost strain-limiting layer and its underlying transition zone, hence the distinct non-linear form of the incision boundary (see also dashed lines in Fig. 3C). The tangential and transition zones are pulled towards the directly compressed region, and because of the intrinsic inter-zone structural continuity the radial zone is similarly sheared laterally as is apparent from the lines of chondrocyte continuity (Fig. 6C).

Because the mid-depth incision has completely destroyed the continuity of both the tangential and transition zones, the pull-down forces are diminished to the point where there is now insufficient indirect transmission of compression into the bulge region and thus no shear bands are generated (see also Fig. 3C). This lack of indirect compression also leads to a decreased resistance to the upwards bulging of cartilage matrix in the channel. This can be seen from the decreased value of the parameter h_r − h_b when comparing mid-depth and shallow incision samples (see Results section).

Further, the curvature of the incision boundary of the mid-depth incision samples is lower than that of the full-depth incision samples because the matrix in the former is still constrained to a limited degree by the transverse interconnectivity of the fibrils in the remaining intact portion of the radial zone. This is indicated in schematic Fig. 8C by the reduced size of the transverse force vectors at the incision tip relative to those shown at the shallow incision tip (schematic Fig. 8B). There is, of course, a complete absence of these vectors in the full-depth incision sample (schematic Fig. 8D).

All samples, whether intact or incised, exhibited a gross shear boundary in the radial zone matrix extending from the edges of each of the directly compressed regions into the bulge (Thambyah & Broom, 2006). Both the abruptness and extent of this boundary varied with incision depth (see Fig. 6). However, rather than reflecting a localised structural discontinuity, this boundary is generated by the juxtaposition of two distinct deformation fields within the same radial zone structure: one deformation field arises from the lateral movement of the matrix in the directly compressed region; the other from the upwards bulging of the matrix into the relief channel. A similar phenomenon of upwards bulging was also demonstrated in an earlier study by Bevill et al. (2010) using samples that had their entire tangential layer removed and were similarly compressed. These authors showed that with decreasing osmolarity (i.e. increased swelling pressure) the bulge height normalised to the uncompressed thickness of the surface-removed samples actually increased, thus confirming that the bulge is not simply a residue of undisturbed matrix left in the channel by the compressed regions on either side.

Because this gross shear boundary represents a localised collapse of the collagenous network in the mid-zone matrix, the acuteness of shear as indicated by the distortion in the lines of chondrocyte continuity is an indicator of the difficulty or ease with which this collapse can occur. The increasing acuteness of shear in progressing from the intact to the shallow-incision samples, and then to the mid- and full-depth incisions (Fig. 6A–D) suggests that resistance of the radial matrix to collapse relies on it having a component of transverse continuity along which force transmission can take place. Note especially that in the full-depth incision samples transverse matrix continuity over the full cartilage depth has been destroyed, hence these samples exhibited the most acute or sharpest shear boundary (Fig. 6D and enlargement in 6F). By contrast, with the shallow incisions there is still a sufficient component of transverse fibrillar alignment in the remaining thickness of intact transition zone to provide for some transverse force transmission. The mid-zone incision response clearly demonstrates an intermediate behaviour (Fig. 6C,E).

The combined presence of the oblique/counter-oblique shear bands and the gross shear boundaries as revealed in the channel indentation experiment therefore provide a baseline micro-morphological response that characterises a completely normal cartilage-on-bone system. This new study demonstrates that in addition to the previously demonstrated effects of both matrix destructuring (Thambyah et al. 2011) and complete articular surface removal (Bevill et al. 2010), a focal radial incision that disrupts the normal transverse continuity of the cartilage also leads to an altered internal pattern of matrix shear whose intensity varies with incision depth.

Potential clinical implications

This study has shown that with a shallow focal incision (i.e. to a depth not exceeding the transition zone) and at a substantial compressive stress, the cartilage matrix still contains sufficient strain-limiting abilities to distribute loads away from the directly compressed area and thereby provide a degree of protection for the underlying matrix. It implies that a significant depth of surface fibrillation is required during in vivo cartilage degeneration to effect major differences in the mechanical microenvironment. In the clinical context, the current study highlights the importance of transverse connectivity in the cartilage matrix, especially in the uppermost layers, and this is of potential relevance when considering the suturing of cartilage grafts. Finally, the work presented in this paper stresses the importance both of recreating the superficial zone in tissue-engineered cartilage, and developing computational models with appropriate depth-varying anisotropic material properties.

Authorship

Mieke Nickien: concept/design, acquisition of data, data analysis/interpretation, drafting of the manuscript. Ashvin Thambyah: concept/design, acquisition of data, data analysis/interpretation, critical revision of the manuscript. Neil Broom: concept/design, acquisition of data, data analysis/interpretation, critical revision of the manuscript.