Cartilage Structure The structure of articular cartilage has been described historically in a series of zones according to either chondrocyte morphology or collagen organization. The classic description of 'arcades' of collagen fibres by Benninghoff (1) and the relationship of this to the more easily visible cells by Collins (2) has been supported by numerous studies over many years. We have used X-ray diffraction (3) and polarized light microscopy (4) to quantify the distribution of collagen fibrils within the matrix. In our first studies, a focussed x-ray beam, 200 µm across, was passed through a block of tissue about 1 mm square in cross-section, and a diffraction pattern recorded from each location. An orientation distribution function can be measured from each diffraction pattern (Figure 1) that describes the probability of finding a fibril at a given angle to a chosen direction. By passing the beam through the tissue in different directions and at various depths from the articular surface, we showed that there is a preferred orientation, which changes with depth in the tissue, largely as described qualitatively by Benninghoff. The distribution of orientations about that preferred direction is broad. Figure 2 is a schematic diagram of a block of tissue that was measured showing how the fibril orientations lie within it.
Figure 1. Orientation distribution functions obtained from x-ray diffraction patterns recorded at interval of 200 µm from the articular surface to the subchondral bone. 0° represents the normal to the articular surface. These data show clearly the change in preferred orientation in passing from the surface to the deep zone and evidence for a bimodal distribution in the transition zone. These data have been used in subsequent models to estimate stresses within the fibrils and show how they can provide reinforcing to the tissue.
|
 |
| Collagen is a birefringent material, i.e. the speed of light passing through it depends on the relative orientation of the electric vector of the light and the axis of the fibre. Because these birefringent fibrils have a preferred orientation, the tissue also is birefringent, so that the passage of polarised light through a thin section will depend on how the direction of polarisation is oriented with respect to the tissue structure. Figure 3 shows how this can give rise to some rather beautiful colours that reflect the collagen organisation when a full-wave plate is inserted in the path of the beam. This may look pretty but hides the fact that quantitative information can be derived from such images by measuring the birefringence. Often overlooked is that this can be done without the addition of stains, and it is a puzzle why some users still insist on staining the tissue with the attendant risks of artefacts and loss of quantitative data. However, there are two contributions to the birefringence; the intrinsic birefringence of an oriented array of birefringent fibres, and the form birefringence arising form the mismatch between the refractive indices of the fibres and the surrounding material. The former can be analysed mathematically and leads to an expression relating birefringence to an orientation distribution function. The second is less tractable and is best dealt with by immersing the tissue section in a medium with a refractive index matching that of collagen. This effectively removes the form birefringence. Experiments showed that glycerol is a suitable medium with a refractive index of 1.47 (4) . |
Figure 2. Schematic drawing of the orientations of collagen fibrils in human patellar articular cartilage showing that the distribution observed, and hence the thickness of the zones, is not necessarily the same when viewed from different directions. |
|
The articular surface
The surface of any material has unique properties arising from the energy associated with it. This makes the study of surfaces surprisingly difficult as they tend to accumulate a molecular layer by adsorption from the surroundings. Determining then what is part of the tissue structure becomes a debatable issue. In any case, such an adsorbed layer must be there in vivo and maybe the issue is slightly academic, though interesting. Early Anatomists believed that a membrane had to be associated with any tissue surface in the body, but as microscopes improved and the magnitude of the forces involved became appreciated it became apparent that no membrane existed at the articular surface of cartilage (5) . As that idea died away the introduction of phase contrast microscopy lead to another observational artefact that still stubbornly resists removal. A bright line was observed at the articular surface and given the name Lamina Splendens . Some simple experiments and rather more complicated theory showed this, however, to be an artefact of the technique (6) . What was observed was certainly not any part of the structure of the tissue. (see footnote )
Figure 3. Polarised light micrograph of a thin section of canine articular cartilage. Inserting a full-wave plate produces the striking colouration. The blue layer at the articular surface represents collagen fibrils with a preferred orientation parallel with the surface. A narrow transition zone separates this from the deep zone (yellow) in which the fibres are preferentially oriented perpendicular to the surface, and to the subchondral bone. |
|
Structure and function
The measured orientation distributions were later used to estimate the stresses that may be found in the collagen fibrils to determine whether a fibre-composite approach might give some insight into the structure-function relationships (7;8) . The first model was based on the concept of a pressure vessel. A series of such pressure vessels were envisaged within the tissue enabling estimates to be made of the stress with the collagen fibrils. This model provides a mechanism for how the swelling pressure developed within the hydrophilic gel could be taken up by the fibrils, which, like ropes, are strong in tension (9). This still leaves the question, however, of how is that stress transferred to the fibres? Possible mechanisms must lie somewhere between two extremes; either a number, probably small, of covalent cross-links formed between the surrounding matrix and specific sites on surface collagen molecules, or a large number of weak, non-specific interactions along the length of the fibril. Evidence for specific cross-links is still lacking, though collagen type IX has been proposed as a possibility. Exploring the other extreme, is it possible for weak interactions to transfer enough stress to the fibrils? Modelling some of the different mechanisms for stress transfer at the fibril surface showed that if we assume there is, on average, only one interaction per 67 nm D-period n the collagen fibril, that interaction need only have a strength of about 10 - 100 pN ( 1pN = 10 -12 N). Clearly, more interactions would reduce these figures further. This is comparable with the strength of a hydrogen or a van der Waals bond and indicates that fixed covalent links may not be required (8) .
There are various approaches to understanding the mechanical properties of articular cartilage. Arguably the one most used is that popularised by Van Mow and colleagues over many years based on a theory of mixtures and often known by its original name of biphasic theory, though several more phases have since been proposed. In its simplest form the tissue is considered as a porous solid phase in which water is free to move. The solid matrix and thermodynamic equilibrium of fluid loss under load yields an equilibrium modulus and frictional drag of water through the pores provides the dynamic properties. As a model it is very similar to consolidation theory developed by Biot for soils and treated in more detail by Neil Broom. While providing equations with adjustable parameters that can be fitted to experimental data, neither of these provide a clear link between tissue composition and structure and mechanical behaviour. Not least is the question of what constitutes the 'solid matrix' in a material that is 70% water, of which only a few percent is lost during loading. Another approach, though less well-developed is to consider the tissue as a viscoelastic material in which water acts as a 'plasticiser', i.e. it makes the matrix deformable - dry cartilage is rigid. Viscoelastic materials have mechanical properties that depend strongly on the rate at which loads are applied. The modulus of cartilage is often reported to be in the region 0.1 - 10 MPa. This considerable range reflects whether the equilibrium modulus is being quoted, i.e. that recorded after prolonged static loading and therefore very low, or a higher value recorded during conventional materials testing in which load is increased steadily and the slope of the stress-strain relationship is found. Studies have shown that as the loading rate increases so does the modulus. Neil Broom's group, using a pendulum device, found that the modulus increased linearly with strain rate up to about 1% s -1 and then reached a steady value. The modulus found though, was still quite low because the applied stresses were low. Because the stress-strain relationship is J-shaped, the modulus, given by the slope of the curve, also increases with stress. We have found values of in excess of 100 MPa during impact loading, though again there is evidence for a loss of strain-rate dependency at high strain-rates. These results indicate that cartilage becomes elastic at high strain rates.
Footnote
Following the publication of this paper, I received a letter from Professor Michael MacConaill:
Curiously, though Professor MacConaill never took the lamina splendens seriously others did. So much so that it still appears in textbooks and papers, to the extent that pictures show it being peeled back off the surface by hypothetical forceps. Anyone who has attempted that will know it cannot be done. Many have tried hard to find its composition. As mentioned above, the surface will be somewhat different to the bulk tissue and is worthy of study as it will influence friction and lubrication. Whether a structural lamina exists is still open to question. It should never, however, be called the Lamina splendens . |
 |
| |
References (1) Benninghoff A. Form un Bau der Gelenkknorpel in ihren Beziehungen zur Funktion. II. Der Aufbau des Gelenkknorpels in seinen Bezeihungen zur Funktion. Zeit Zellforsch und Mikroskop Anat 2 : 783-862, 1925.
(2) Collins DH. The pathology of articular and spinal deseases. London : Edward Arnold, 1949.
(3) Aspden RM, Hukins DWL. Collagen organisation in articular cartilage determined by X-ray diffraction and its relationship to tissue function. Proc R Soc Lond B212 : 299-304, 1981.
(4) Yarker YE, Aspden RM, Hukins DWL. Birefringence of articular cartilage and the distribution of collagen fibre orientations. Connect Tissue Res 11 : 207-213, 1983.
(5) MacConaill MA. The movements of bones and joints 4. The mechanical structure of articular cartilage. J Bone Joint Surg 33B : 251-257, 1951.
(6) Aspden RM, Hukins DWL. The lamina splendens of articular cartilage is an artefact of phase contrast microscopy. Proc R Soc Lond B206 : 109-113, 1979.
(7) Aspden RM, Hukins DWL. Stress in collagen fibrils of articular cartilage calculated from their measured orientations. Matrix 9 : 486-488, 1989.
(8) Aspden RM. Fibre reinforcing by collagen in cartilage and soft connective tissues. Proc R Soc Lond B-258 : 195-200, 1994.
(9) Hukins DWL, Aspden RM. Composition and properties of connective tissues. Trends Biochem Sci 10 : 260-264, 1985.
|
