SEARCH

SEARCH BY CITATION

Keywords:

  • calcified cartilage;
  • coherent anti-Stokes Raman scattering;
  • collagen;
  • second harmonic generation;
  • tidemark;
  • two-photon fluorescence

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion and Conclusions
  7. Acknowledgements
  8. Contributions
  9. Funding source
  10. Conflict of interest
  11. References

Multi-modal multiphoton microscopy was used to investigate tissue microstructure in the zone of calcified cartilage, focussing on the collagen fibre organisation at the tidemark and cement line. Thick, unstained and unfixed sagittal sections were prepared from the equine metacarpophalangeal joint. Second harmonic generation (SHG) provided contrast for collagen, two-photon fluorescence (TPF) for endogenous fluorophores, and coherent anti-Stokes Raman scattering (CARS) allowed the cells to be visualised. The structure of radial and calcified cartilage was found to vary with location across the joint, with the palma regions showing a more ordered parallel arrangement of collagen fibres than the cortical ridge and dorsal regions. These patterns may be associated with regional variations in joint loading. In addition, the cell lacunae had a greater diameter in the dorsal region than in the palmar region. At the cement line some collagen fibres were observed crossing between the calcified cartilage and the subchondral bone. At the tidemark the fibres were parallel and continuous between the radial and calcified cartilage. Beneath early superficial lesions the structure of the tidemark and calcified cartilage was disrupted with discontinuities and gaps in the fibrillar organisation. Cartilage microstructure varies in the deep zones between regions of different loading. The variations in collagen structure observed may be significant to the local mechanical properties of the cartilage and therefore may be important to its mechanical interactions with the subchondral bone. The calcified cartilage is altered even below early superficial lesions and therefore is important in the understanding of the aetiology of osteoarthritis.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion and Conclusions
  7. Acknowledgements
  8. Contributions
  9. Funding source
  10. Conflict of interest
  11. References

Articular cartilage forms a low friction, impact-absorbing layer on the surface of bones in synovial joints. Its main components are a dense network of fine type II collagen fibres and highly hydrated proteoglycans. Three zones at different depths can be identified in cartilage based on differences in the collagen fibre organisation. First there is a superficial zone where the collagen fibres run parallel with the articular surface, then a transitional zone with less ordered collagen which progresses into the deep or radial zone in which the collagen fibres run perpendicular to the articular and bone surfaces (Clarke, 1971; Stockwell, 1979). At the interface with bone there is a region of mineralised cartilage where the collagenous matrix is calcified with crystals of, predominantly, calcium apatite. The boundary between mineralised and unmineralised cartilage forms a distinct line in transverse light microscope sections, known as the tidemark and the boundary with bone is known as the cement line.

The calcified cartilage together with its boundary structures serves a number of important functions. It tethers cartilage to bone, and couples the disparate mechanical properties of cartilage and bone (Mente & Lewis, 1994). It is also a barrier to the exchange of water and solutes such as nutrients and metabolites between the cells of cartilage and the bone microcirculation (Arkill & Winlove, 2008). The structure of this region changes during development and ageing (Clark et al. 1997). This region is also believed to be involved in the development of diseases in both cartilage and bone. Cartilage lesions are often associated with an increase in vascularity and remodelling in the underlying subchondral bone (Radin & Rose, 1986; Thambyah & Broom, 2009) and duplication of the tidemark is regarded as a sign of osteoarthritis (Green et al. 1970; Lane & Bullough, 1980). Some authors suggest that it holds the key to the negative correlation between the incidence of osteoporosis and osteoarthritis (Li & Aspden, 1997; Oegema et al. 1997).

The structure of the zone of calcified cartilage is still, to some extent, incompletely characterised. The well established limitations of light and electron microscopy and histochemistry are compounded by technical problems of preparing the required samples from such a heavily mineralised tissue. In consequence, the structure and biochemical composition of the tidemark, the organisation of collagen, and the viability of cells are still debated, and changes in disease are even less well understood. The present study of this region is facilitated by the use of multiphoton microscopy. Like confocal microscopy this is a laser scanning technique which offers submicron resolution and 3D imaging. Multiphoton processes involve the simultaneous absorption of two or more near infrared (IR) photons and only occur in high electric fields such as at the point of focus of a pulsed laser beam. The advantages of multiphoton microscopy are that the IR excitation wavelengths allow better penetration into scattering media and that it provides intrinsic contrast mechanisms which do not require the sample to be stained. Multiphoton microscopy has already proven to be a powerful tool in cartilage research in studies on the superficial zones of the cartilage (Yeh et al. 2005; Mansfield et al. 2008, 2009). In the present work we apply these methods to the deeper tissue, utilising three different multiphoton imaging modalities combined in a single multimodal microscope. Second harmonic generation (SHG) is used to image the collagen matrix and has previously revealed structural changes in cartilage with disease or damage (Brockbank et al. 2008; Mansfield et al. 2008; Werkmeister et al. 2010). Two-photon fluorescence (TPF) gains contrast from endogenous fluorophores within the tissue (mainly collagen cross-links and elastin in the extra-cellular matrix, and NAD(P)H and flavoproteins within the cells) and has revealed the previously unappreciated presence of a network of elastic fibres in the surface zone (Mansfield et al. 2009). To image the chondrocytes coherent Anti-stokes Raman scattering (CARS) is used, tuned to image the CH2 vibrations in both cellular lipids and in proteins (Cheng et al. 2002).

We applied these techniques to study the osteochondral junction, with particular emphasis on the tide mark and cement line, in the equine metacarpophalangeal joint. The metacarpal joint displays consistent regional variations in superficial cartilage structure that have been correlated with loading patterns (Norrdin et al. 1999; Martinelli et al. 2002; Moger et al. 2007). In this study we investigated regional variations in the osteochondral junction across the distal end of the third metacarpal bone. In sagittal section (see Fig. 1) the articular surface has two radii of curvature. The curvature is greater and the cartilage is thinner on the palmar surface. The anterior (dorsal) cartilage surface articulates with the proximal phalanx and the posterior (palmar) surface articulates with the sesamoid bones. The apex of the joint where the two radii of curvature meet is called the cortical or transverse ridge. The dorsal surface is exposed to mainly shear forces and the plamar surface is exposed to compressive forces exerted by the sesamoid bones, resulting in thickened subchondral bone. It is therefore of interest to enquire whether these different patterns of loading are associated with structural variations in the deepest layers of the tissue. In addition, the horse is a valuable model because specimens can be obtained with a wide age range and frequently display lesions very similar to those of human osteoarthritis (Mackay-Smith, 1962; McIlwraith, 1982). Equine cartilage is thinner than human cartilage (typically 1 mm thickness) but it displays similar zonal structure to that observed in human tissue. In this study we examined focal lesions on the cortical ridge, which are a consistent early degenerative change in the equine metacarpophalangeal joint which occurs in horses that do not yet exhibit clinical joint disease. These lesions may be associated with overuse, as their prevalence is much higher among racehorses (Barr et al. 2009).

image

Figure 1.  A schematic diagram of equine fetlock showing the metacarpophalangeal joint. The enlarged rectangle shows a cross-section of the distal end of the third metacarpal. The joint surface has two different radii of curvature and the junction between these is marked by a cortical ridge, which is a common site of early lesions. The loading of the joint varies across the surface. The palmar surface which articulates with the sesamoids is exposed primarily to compressive loads and the dorsal surface which articulates with the proximal phalanx is exposed to greater shear loads. The thickness of the cartilage also varies, with the thinnest cartilage on the palmar surface and the region with the thickest cartilage is close to the cortical ridge and marked with an arrow.

Download figure to PowerPoint

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion and Conclusions
  7. Acknowledgements
  8. Contributions
  9. Funding source
  10. Conflict of interest
  11. References

Sample preparation

Forelimbs from horses of known age were collected from a local abattoir (Potters, Taunton, UK). In total, 12 samples were used with an age range of 1–15 years. Horses reach skeletal maturity between the ages of 3 and 4 years; where observations differ in the younger horses this is explicitly stated in the results. The joints were opened in the laboratory and examined for signs of apparent macroscopic degeneration; joints with an intact surface that scored zero on the Outerbridge scale (Outerbridge, 1961) were classified as normal. Transverse sagittal sections approximately 500 μm thick were prepared from the distal end of the third metacarpal containing the full depth of cartilage and between 2 and 5 mm of underlying subchondral bone. This was done by first cutting an approximately 1-mm-thick section using a jeweller’s saw and then removing sequential layers using a freezing microtome until a 500-μm section with a smooth cut surface remained. Comparative analysis with fresh cartilage showed that freezing and sectioning did not have an observable effect on collagen organisation or cellular morphology. To avoid surface artefacts, multiphoton images were taken at a focal depth of 10 μm below the cut surface. For en-face imaging of the tidemark region the cartilage together with the subchondral bone was removed from the metacarpal joint using a jeweller’s saw and the superficial cartilage was removed using a scalpel blade to leave only the calcified cartilage and a very thin layer of radial cartilage. To investigate optical effects caused by minerals and whether this contributes to fluorescence, some cartilage sections were demineralised by placing in 8% formic acid for 8 h (Drury & Wallington, 1967).

The structure of the tidemark and cement line and their regional variations were investigated initially in joints which showed no macroscopic degeneration. Variations in collagen fibre organisation and lacunae shapes were investigated in three regions on the distal end of the third metacarpal (palmar, cortical ridge and dorsal) identified in Fig. 1. In addition, samples were taken from two horses (aged 14 and 15 years) displaying early focal lesions on the cortical ridge, where the depth of the lesion did not extend below the transitional zone.

Multiphoton imaging

Imaging was carried out on a custom built multi-modal microscope comprising a modified inverted microscope and confocal laser scanner (IX71 and FV300; Olympus, UK). To optimise the transmission of the near IR light needed for CARS imaging, the standard galvanometer scanning mirrors were replaced with silver galvanometric mirrors and the tube lens was replaced by a MgF2-coated lens. The dichroic mirror within the scan unit was replaced by a silver mirror which gave high reflectivity throughout the visible and NIR (21010; Chroma Technologies). The light was focused onto the sample using a 60X 1.2 NA water immersion objective (UPlanS Apo; Olympus).

SHG and TPF

Both SHG and TPF were excited using the 800-nm output from a mode-locked femto-second Ti:Sapphire laser (Mira 900 D; Coherent) with a pulse width of approximately 100 fs and a 76 MHz repetition rate. The signal was collected in the epi-direction using the objective lens and separated from the laser fundamental using a long pass dichroic mirror (670dcxr; Chroma Technologies). The light was focused on a photomultiplier tube (PMT, R3896; Hamamatsu). Filter combinations (CG-BG-39 and F10-400-5-QBL; CVI laser) and (CG-BG-39 and F70-500-3-PFU; CVI laser) were used to isolate the SHG (at 400 nm) and TPF (centred at 500 nm) light, respectively.

CARS

CARS was excited using the dual wavelength output of an optical parametric oscillator (OPO) (levanter Emerald; APE, Berlin, Germany) pumped with the frequency doubled Nd:Vandium picosecond laser (PicoTrain High-Q Laser; production GmbH). This provided a train of 6-ps pulses at a 76 MHz repetition rate. To image the cells, the difference in wavelengths between the signal and idler outputs were tuned to correspond to the energy of the CH2 symmetric stretch in lipids at 2845 cm−1 (Evans & Xie, 2008). For this the signal and idler wavelengths were 924 and 1254 nm, respectively.

The forward-CARS signal was collected by the air condenser, transmitted by the dichroic mirror and directed onto a red-sensitive PMT (R3896; Hamamatsu). The epi-CARS signal was collected using the objective lens and separated from the pump and Stokes beams by a long-wave pass dichroic mirror (z850rdc-xr; Chroma Technologies) and directed onto a second R3896 PMT at the rear microscope port. The anti-Stokes signal was isolated at each photodetector by a single band-pass filter centred at 750 nm (HQ750/210; Chroma Technologies).

Image analysis

Fourier transform image analysis techniques available in the Fiji version of imagej (http://fiji.sc/wiki/index.php/fiji) were used to parameterise the degree of collagen fibre matrix organisation in the images. Square regions with an area of 25 × 25 μm and containing no cells were selected on the SHG images and the directionality plugin was applied (Tinevez, 2010). This performed a Fourier transform on the image and calculated a weighted sum of the Fourier components at each angle. A Gaussian peak was fitted to these histograms of the components as a function of angle. Both the visibility (peak height – background) of the peak and the width of the peak give an indication of the degree of alignment within the sample. A narrow peak with a high visibility indicates that the majority of collagen fibres are arranged parallel to one another, whereas a broad peak with a low visibility indicates a less ordered configuration.

Statistical analysis

Differences in lacuna width and collagen fibre organisation were observed between the palmar, cortical ridge and dorsal areas on the joint in five horses. The mean and standard deviation of the lacuna width and visibility of the peak calculated by the directionality plugin were calculated for each region in each sample. For all measurements the standard error was small in comparison to the standard deviation within the group. To evaluate whether any differences between regions were significant, t-tests for samples with unequal sizes and unequal variances were performed. The null hypothesis that there was no difference between the regions was rejected if the value for t was greater than the 99% confidence interval. For each sample, over 100 lacunae were measure to determine lacunae variations and over 90 regions of interest were analysed to determine collagen variations, with the sample numbers distributed approximately evenly between the three regions on the joint surface. To avoid any sampling bias, images were taken at set 1-mm intervals across the apex of the third metacarpal and all the lacunae in the images were included in the statistical analysis.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion and Conclusions
  7. Acknowledgements
  8. Contributions
  9. Funding source
  10. Conflict of interest
  11. References

Tidemark region

General appearance of the tidemark

The tidemark delineates the transition from the radial cartilage to the calcified cartilage and it shows up clearly in all the multiphoton imaging modalities employed (as shown in Fig. 2). There is a decrease in signal intensity at the tidemark for both CARS and SHG imaging; however, this is most likely an optical effect due to the changes in refractive index between the calcified and non-calcified cartilage, as it was not present in decalcified tissue (data not shown).

image

Figure 2.  Multi-modal imaging of the tidemark in a 6-year-old horse. (A) Second harmonic generation (SHG) image showing the collagen matrix. The individual collagen fibres are too fine to be resolved individually; however, their arrangement determines the texture of the image. (B) Two-photon fluorescence (TPF) image showing the distribution of endogenous fluorophores and (C) a coherent anti-Stokes Raman scattering (CARS) image taken at the CH2 resonance. (D) Merged image where blue = SHG, green = TPF and red = CARS. The tidemark is evident in all three imaging modalities. In the CARS image the chondrocytes are seen filling their lacunae above the tidemark but below the tidemark the lacunae are empty.

Download figure to PowerPoint

The two-photon fluorescence intensity in the calcified cartilage fluorescence is much higher than in the non-mineralised tissue and therefore the tidemark stood out as a step change in fluorescence intensity. This fluorescence was not removed by decalcification and pure hydroxylapaptite samples did not exhibit fluorescence, indicating that the fluorescence is not due to the mineral crystals (data not shown). In the calcified zone, additional tidemarks, widely reported in the histological literature (Green et al. 1970; Thambyah & Broom, 2009), were also evident as continuous undulating bands of increased fluorescence (these are most pronounced in Fig. 3). The duplicate tidemarks were evident in all samples from animals which had reached skeletal maturity but were not evident in the specimens aged 3 years or less. Fluorescence in the radial zone had a textured appearance, notably different from that in the superficial zone (Mansfield et al. 2009) and in particular no highly fluorescent elastin fibres were evident.

image

Figure 3.  Images in the plane of the tidemark. These images are taken from a z-stack of images taken in a cartilage plug from the cortical ridge of a 6-year-old horse. The stack of images started in the radial tissue and finished in the calcified tissue, with each image being separated by a 1-μm step. The images displayed here were identified as the plane of the tidemark due to a step change in fluorescence intensity between adjacent images as the field of view moved into the calcified tissue. Contrast in the second harmonic generation image (A) is from the collagen matrix and in the two-photon fluorescence image (B) from endogenous fluorophores. The collagen imaging in (A) shows an absence of fibres in the plane of the tidemark and both (A) and (B) show smooth lacuna boundaries.

Download figure to PowerPoint

Figure 4 shows the TPF and SHG at the tidemark imaged en-face arrangement with the sample orientated to ensure that the plane of the tidemark fills the field of view. The images were selected from a z-stack of images parallel to the tidemark on the cortical ridge region. The location of the tidemark in the stack could be identified by an increase in the fluorescent signal between frames in the images as the stack moved from the radial zone to the calcified cartilage. In the SHG images the tidemark was not clearly apparent, indicating that there are no major structural or orientational changes in the collagen at the tidemark. The mottled SHG pattern is indicative of collagen fibres approximately perpendicular to the imaging plane throughout the region. Importantly, there is no evidence for a significant population of collagen fibres parallel to the tidemark, as suggested by some authors (Green et al. 1970; Redler et al. 1975). When imaged at the same power and PMT settings, the SHG from the collagen in the transverse cartilage sections is approximately three times more intense than that from an en-face cartilage sample at the tidemark. This is another indicator that the collagen is predominantly perpendicular to the tidemark, as the SHG from end-on collagen fibres will be very low compared with those parallel to the laser polarisation (Freund et al. 1986). The SHG intensity from collagen in the calcified cartilage and radial zone is over twofold greater than that in the superficial zone (samples both arranged with collagen fibres parallel to the laser polarization); this may be due to larger fibre diameters in the radial zone than in the superficial zone (Eyre, 2002).

image

Figure 4.  The cement line in transverse section. In all images the calcified cartilage is in the top half of the image and the subchondral bone in the bottom half. CC, calcified cartilage; CL, cement line; SB, subchondral bone. (A) SHG image showing differences in collagen organisation between the cartilage and the bone. (B) TPF image of the same field showing the accumulation of fluorescent debris in the lacunae and intense fluorescence at the cement line. (C–E) Higher resolution SHG imaging of collagen fibres at the cement line interface between the calcified cartilage and the subchondral bone. Insets show regions with fibres crossing the cement line at an increased magnification. In (D) there are so many fibres crossing the cement line that it appears obscured in places.

Download figure to PowerPoint

Cells and lacunae at the tidemark

Coherent Raman imaging tuned to the CH2 bond vibrations shows empty lacunae in the zone of calcified cartilage (Fig. 2), whereas on the superficial side of the tidemark chondrocytes are seen filling their lacunae. When the lacunae beneath the tidemark were imaged via coherent Raman, 90% were found to be empty. To avoid any possible confusion with artefactual cell loss caused during the sectioning, images were taken at a depth of 10 μm from the cut surface. The edges of the empty lacunae in the calcified cartilage have an irregular shape which shows up most clearly on the CARS images, indicating a high concentration of CH2 in either lipids or proteins. The fluorescence intensity at the edge of the lacunae is high (clearly shown in Figs 3 and 5) and is greatest close to the subchondral bone.

image

Figure 5.  An example of the histograms derived from directionality analysis on SHG images of two regions (size 25 × 25 μm) of radial zone cartilage, region 1 (from the palmar area) showing highly parallel collagen fibre arrangement and region 2 (from the dorsal area) showing a much less ordered collagen fibre arrangement. The values on the y axis are normalised so that the average value of the Fourier components is equal to 1. The analysis presented in the text was based on 90 regions of interest in each of the five horses used in the directionality study. The visibility (peak height – background) of the peak was used to quantify the variations in collagen fibre organisation.

Download figure to PowerPoint

Cement line

The cement line marks the transition from the zone of calcified cartilage to the subchondral bone. The fluorescence in the underlying subchondral bone is generally higher than that in the cartilage and the cement line is clearly visible in TPF images as a line of increased fluorescence. High resolution SHG imaging in this region (Fig. 5) reveals collagen fibres which cross the cement line. This was evident in seven of the nine samples examined but the distribution was uneven and on a 50-μm-length scale there were regions without apparent collagenous connections between the bone and cartilage. SHG images from the type I collagen in the bone and the type II collagen in the cartilage differ, with the bone containing regions of much brighter SHG and thicker fibres.

Regional variations

Series of images of both the radial and calcified cartilage were taken on sagittal sections of the distal end of the third metacarpal. Samples were taken from five different horses (ages 1, 6, 7, 13 and 14 years) and in each horse three regions, palmar, cortical ridge and dorsal as shown in Fig. 1, were investigated.

Regional variations in collagen structure

SHG imaging showed significant variations in the collagen matrix organisation between the three regions. To parameterise the differences, a fast Fourier transform (FFT) image analysis technique was employed which produced histograms of the Fourier components as a function of angle. Example plots combined with a Gaussian fit to the distributions are shown in Fig. 6 for two regions, one showing a highly ordered parallel collagen fibre structure (from the palmar area) and the other showing very little directionality (from the dorsal area).

image

Figure 6.  Regional variations in the degree of parallel organisation of the collagen fibres. SHG images are from (A) palmar region, (B) cortical ridge region and (C) dorsal region in the 14-year-old horse. In (A) there is a 16° change in the collagen fibre angle across the tidemark. The collagen fibre arrangement appears more ordered and parallel in (A) compared with (B) and (C), and this was confirmed by the Fourier transform analysis.

Download figure to PowerPoint

In all five horses the collagen fibre arrangement was consistently more parallel in the palmar region than in the dorsal and cortical ridge regions (see Fig. 7). To quantify this, 90 regions of interest were analysed in each horse. The mean visibility of the peak and standard deviation were calculated as shown in Table 1 and a t-test (for unequal variances and unequal sample sizes) was performed at a 99% confidence level to see whether the variations were significant. The average ratio of peak height between the palmar region and the cortical ridge and dorsal regions were 1 : 0.59 and 1 : 0.44, respectively. The same patterns in variations in collagen organisation were observed both above and below the tidemark.

image

Figure 7.  Regional variations in cell organisation in a 14-year-old horse. (A) Typical images from the palmar region, (B) typical images from the cortical ridge region and (C) typical images from the dorsal region. In the merged images the auto-fluorescence is green and the SHG from collagen fibres is blue. In the cortical ridge and dorsal regions the width of the lacunae (dimension parallel to the tidemark) is greater than in the palmar region.

Download figure to PowerPoint

Table 1.   Regional variations in the collagen fibre organisation: visibility of the peak calculated from the directionality analysis for each sample.
Horse age in yearsRegions on joint – visibility of peak – in comparison with the palmar regionDifferences between regions significant at 99% confidence level?
PalmarCortical ridgeDorsalPalmar–cortical ridgePalmar–dorsalCortical ridge–dorsal
MeanSDMeanSDMeanSD
 110.2600.5220.1350.2860.137YesYesYes
 610.2110.5970.0930.3670.197YesYesYes
 710.4650.4320.2040.5300.164YesYesNo
1310.3050.7800.2340.4750.169YesYesYes
1410.1850.6170.1810.5490.210YesYesNo

Fourier transforms were used to measure the predominant collagen fibre angle both above and below the tidemark. In the cortical ridge and dorsal regions the collagen fibre angle was unchanged across the tidemark. However, in the palmar region at the site of the thickest cartilage (marked with an arrow in Fig. 1) the collagen fibre angle changed across the tidemark. The difference in collagen fibre angle ranged between 8° and 16° between samples; a sample showing a 16° angle change is shown in Fig. 7A.

Regional variations in cellularity

The width (dimension parallel to tidemark) and height (dimension perpendicular to tidemark) of all the radial zone lacunae were measured in the palmar, cortical ridge and dorsal regions. The mean lacuna length was constant; however, the width varied between regions, being greater in the dorsal and cortical ridge region than in the palmar region (see Fig. 3). The average ratios of lacunae width compared to the palmar region for the cortical ridge region and dorsal region from the five horses are 1 : 1.25 and 1 : 1.34, respectively, and the data are shown in Table 2. The differences between the palmar region and the other regions were significant at a 99% confidence interval in four of the five samples (t-test for unknown variances and different sample sizes). In all regions, cells were aligned within the lacunae perpendicular to the tidemark as described in the Benninghoff model of cartilage structure (Benninghoff, 1925). Histological studies report lacunae one cell in diameter and two to four cells in depth in the radial zone of cartilage (Youn et al. 2006); however, in the dorsal region, many lacunae were greater than one cell in diameter (Fig. 3).

Table 2.   Regional variations in the lacunae diameter for each sample.
Horse age in yearsRegions on joint – lacunae widths in micronsDifferences between regions significant at 99% confidence level?
PalmarCortical ridgeDorsalPalmar–cortical ridgePalmar–dorsalCortical ridge–dorsal
MeanSDMeanSDMeanSD
 17.951.9610.522.7111.372.33YesYesNo
 67.802.229.582.4210.722.86YesYesNo
 77.071.238.551.659.912.31YesYesNo
137.361.628.531.408.033.08YesNoNo
148.591.6911.672.2812.152.92YesYesNo

Early focal lesions

The calcified cartilage underneath focal lesions was investigated in two animals where the lesions did not penetrate below the superficial radial zone. The zone of calcified cartilage appeared normal away from the lesion site; however, underneath the lesions a number of irregularities were observed. Disruptions in the collagen organisation were seen in the calcified cartilage even though the fibrillation in the non-calcified cartilage did not extend much beyond the transitional zone (Fig 8A–C,E). In some images, gaps in the collagen matrix were associated with increased fluorescence (green arrows Fig. 8C). Central to the lesion there was a marked change in the collagen fibre angle (25°) at the tidemark (Fig. 8A). The tidemark was less regular than in the healthy tissue (Fig. 8D) and at some points exhibited increased fluorescence. In the calcified cartilage there were discontinuities in the duplicate tidemarks (Fig. 8C). In one horse the calcified cartilage was thickened beneath the lesion with a depth in excess of 400 μm compared with approximately 170 μm away from the lesion.

image

Figure 8.  Calcified cartilage underlying early focal lesions, from a 15-year-old horse(A–C), and from a 14-year-old horse (D,E). (A) and (B) are from the same region with (A) showing the collagen fibres and (B) showing the endogenous fluorophores. (C) Higher resolution image from the region highlighted in (B) where increased fluorescence can be seen filling the gaps in the collagen matrix (green arrows). In the merged image (D), blue represents collagen imaged through SHG and green represents endogenous fluorophores measured by the TPF. In this region the tidemark appeared irregular. The zone of calcified cartilage is very thick in places (+ 400 μm). Gaps in the collagen matrix were seen beneath the lesion, seen most clearly in (A) and (E).

Download figure to PowerPoint

Discussion and Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion and Conclusions
  7. Acknowledgements
  8. Contributions
  9. Funding source
  10. Conflict of interest
  11. References

Multiphoton microscopy enabled us to investigate several aspects of the structure of the zone of calcified cartilage in thick, unstained sections of fresh tissue, thereby minimising artefacts of sample preparation. High resolution SHG imaging enabled us to explore in detail the organisation of the collagen network. The fibres were continuous across the tidemark and were arranged predominantly perpendicular to it. We found no evidence of a population of fibres parallel to the tidemark and the mechanical arguments for the existence of such a population (Roth et al. 1979) therefore need to be examined. At the bone–cartilage interface it was not possible from the SHG to distinguish types I and II fibres, although the coarser fibres are presumed to be type I. Many fine fibres from the cartilage penetrated across the cement line into the bone. However, there were considerable variations in the density of these fibres on a lateral scale of 50 μm. These fibres are presumably the key components in anchoring cartilage to bone but our observations suggest that the connections are quite focally distributed. This could possibly act to localise the interfacial stress between materials of different mechanical properties in these regions, and elsewhere allow the flexibility to withstand a range of loading conditions.

Chondrocytes were evident in all the lacunae in the non-mineralised tissue investigated with CARS microscopy. However, in the calcified cartilage, 90% of the lacunae were found to be empty. Previous studies have suggested that the chondrocytes in the calcified cartilage die via apoptosis or a similar process (Hashimoto et al. 1998; Aigner et al. 2001; Mobasheri, 2002; Simkin, 2008) and this would help explain why we observe empty lacunae in the mineralised tissue. The empty lacunae also differed from those in the radial zone in that the surrounding collagen had a very irregular boundary instead of the notably smooth one around the intact cells: this may arise from the proteolytic activity of the lysed cell contents.

There was a large TPF signal from the calcified tissue, in agreement with the findings of single photon fluorescence studies (Bachman & Ellis, 1965; Gibson et al. 2001). The TPF images revealed highly fluorescent components in the tidemarks and around the peripheries of the empty lacunae. The latter probably represent an accumulation of fluorescent debris such as flavoproteins released during cellular lysis. It has been noted that a wide range of molecules bind rather tightly to the tidemark (Stockwell, 1979; Lyons et al. 2005) and it may be that its fluorescence arises from accumulation of similar cellular debris rather than intrinsic fluorophores unique to the tidemark. It may be that similar debris also contributes to the low intensity fluorescence present throughout the matrix. Demineralised tissue still exhibited fluorescence, indicating that the fluorescence does not arise from the mineral content of the tissue. Fluorescent collagen cross-links are also thought to be a source of TPF in cartilage, but why these should be much more concentrated in the calcified zone is not clear.

One of the primary aims of the study was to investigate variations in the structure of the calcified zone and deep radial cartilage over the joint surface. The collagen fibre organisation varied with location on the joint. The palmar region showed a highly ordered structure with the majority of fibres aligned perpendicular to the tidemark. In the dorsal and cortical ridge regions the predominant collagen fibre angle was still perpendicular to the tidemark; however, there was a much wider spread of fibre angles, which resulted in a more mesh-like arrangement. The palmar region is subjected to compressive loads and a matrix arranged around columnar collagen fibres is probably well able to support these. However, the dorsal zone and cortical ridge are exposed to shear loading and it is probable that the fibre reinforcement from a 3D network is advantageous under these conditions. An unexpected observation was the change in collagen fibre angle between the radial and calcified cartilage observed in the thick cartilage close to the cortical ridge. Presumably this represents a change in mechanical environment and a question to be investigated in the future is whether this is a compensation measure or a stimulus for degradation at this site. Regional differences were also observed in the morphology of the cellular lacunae, which were much wider in the dorsal and cortical ridge regions than in the palmar region. Mechanical signals are well recognised to be important in determining the synthetic activity of chondrocytes (Grodzinsky et al. 2000). The forces experienced by the cells depend both on the loads applied to the cartilage surface and the mechanical properties of the surrounding matrix. It would be of interest in future experiments to determine the forces experienced by the cells in the different locations and to establish whether the cells have different mechanical properties or transduction processes associated with these morphological differences.

The measurements were too time-consuming to allow us systematically to investigate changes in the calcified cartilage with age. Our small study group contained individuals ranging from 1 to 15 years of age and well documented changes such as the appearance of multiple tidemarks were observed in our population. In two older horses, focal lesions on the cortical ridge were investigated. Lesions in this region are common in equine samples (Mackay-Smith, 1962; McIlwraith, 1982) and there has been speculation that these lesions are associated with the unusual mechanical environment at this site; they are certainly more common in trained horses (Barr et al. 2009). In the present study these lesions provided an opportunity to explore the widely debated association between changes in the superficial cartilage and in the deep zone. We selected very early lesions in which the radial zone of the cartilage appeared normal but, nevertheless, there were disturbances in the calcified zone, most prominently a disruption in the regular organisation of the collagen framework, with large gaps similar to those reported by Boyde & Firth (2008). These were entirely absent in the absence of surface changes, even in animals of advancing years. An interesting question that remains is the cause and effect relationship between the local load-bearing properties and the changes in collagen structure. It needs to be established whether the changes in collagen structure initiate changes in the local loading or whether they have been caused by variations in mechanical load. We also need to establish whether the variations in collagen structure are compensative measures which result in a cartilage structure that is well adapted to the local loading. Alternatively, do these changes result in a failure of the mechanical properties of cartilage which result in micro-damage at the interface, as observed in some older horses? Clearly, the behaviour of the interface both in ageing and in disease requires further investigation, but the methodologies developed in this current work should prove useful tools for this.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion and Conclusions
  7. Acknowledgements
  8. Contributions
  9. Funding source
  10. Conflict of interest
  11. References

We would like to thank Dr Julian Moger for technical support with the multiphoton imaging.

Contributions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion and Conclusions
  7. Acknowledgements
  8. Contributions
  9. Funding source
  10. Conflict of interest
  11. References

Jessica Mansfield was responsible for the collection of the data, data analysis and drafting of the paper. C. Peter Winlove also contributed to the study design and data analysis and document drafting. Both authors were involved in the critical revision process and approval of the final version for submission.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion and Conclusions
  7. Acknowledgements
  8. Contributions
  9. Funding source
  10. Conflict of interest
  11. References