OBJECTIVE The objective of this study was to detail the topographical and zonal distribution of α and β subunits of the integrin superfamily in normal and osteoarthritic cartilage.
METHODS Immunohistochemistry utilising antibodies towards α and β subunits was performed on cryostat sections of human articular cartilage from macroscopically normal (n = 6) and osteoarthritic (n = 6) femoral heads. Samples of articular cartilage were obtained from 12 topographically distinct sites from each femoral head. Each section was divided into zones (superficial, middle, deep) and staining scores were recorded.
RESULTS Normal cartilage stained for integrin subunits α1, α5, αV, β1, β4, and β5, but not for α2, α3, α4, α6, β2, β3, and β6. Intact and non-intact residual cartilage from osteoarthritic femoral heads stained for α1, α2, α5, αV, β1, β4, and β5. Staining was occasionally seen for α4 and β2, but not for α3, α6, β3, and β6. There was no topographical variation in the staining for any of the subunits in either normal or osteoarthritic cartilage. The only subunit that displayed a zonal variation was αV; staining for this subunit was most pronounced in the superficial zone compared with the middle and deep zones.
CONCLUSION Chondrocytes in normal and osteoarthritic cartilage express the integrin subunits α1, α5, αV, β1, β4, and β5. Chondrocytes in osteoarthritic cartilage, in addition, express the α2, α4, and β2 subunits. The αv subunit is expressed by more chondrocytes in the superficial zone in comparison with cells in the deeper zones. None of the subunits display topographical variation in expression.
Statistics from Altmetric.com
Interactions between chondrocytes and extracellular matrix in articular cartilage are critical to cell anchorage, matrix biosynthesis, and matrix degradation.1 2 Cell adhesion molecules are essential mediators of these interactions and may play an important part as environmental monitors involved in the regulation of cartilage matrix turnover.3 4 Human articular chondrocytes are known to express a range of cell adhesion molecules including cluster of differentiation (CD) 44 of the hyaluronan binding protein family,5 6 intercellular adhesion molecule (ICAM) -1 of the immunoglobulin superfamily,7 8 and members of the integrin superfamily.9-11 In normal articular cartilage from femoral condyles, tibial plateaus, and taluses the integrin heterodimers α1β1, α5β1, and αVβ5 are expressed strongly and the heterodimers α3β1 and αVβ3 weakly.9-11 The subunits α2 and α6 have been demonstrated on fetal articular chondrocytes.10 12Potential ligands of integrins expressed by articular cartilage chondrocytes such as collagens, fibronectin, vitronectin, thrombospondin, and laminin are prominent components of the pericellular matrix.13 14
In normal human femoral head articular cartilage, thickness and composition vary with topography.15 16 In addition, the morphology of chondrocytes, cellularity, and distribution of extracellular matrix components vary with depth.17-20 In osteoarthritic cartilage, qualitative alteration of the collagen content,21 22 pericellular loss of collagen type II,23 neoexpression of collagen types I and III,24-26 increased expression of collagen type VI,27 IX,28 and X,23 29 and increased levels of fibronectin30 31 have been observed. The αV subunit containing integrins have been immunolocalised preferentially to the superficial zone of normal articular cartilage.9 However, it has not been reported whether the expression of integrins also reflects the topographical heterogeneity and variation in composition of normal articular cartilage, and the expression of integrins in osteoarthritic cartilage has not been detailed.
The objective of this study was, therefore, to detail the topographical and zonal distribution of α and β subunits of the integrin superfamily in normal and osteoarthritic cartilage from human femoral heads.
Samples of macroscopically normal articular cartilage (Collins/McElligott grade 0)32 were obtained at necropsy within 24 hours of death from two women (age 66 and 82 years) and four men (median age 73 years, range 45–88). None of the subjects had a clinical history of arthritis or chronic systemic inflammatory disease.
Samples for cryostat sections were collected from anterior, posterior, lateral, and medial aspects of perifoveal, central, and peripheral areas of femoral heads. A total of 12 triangular (5 × 5 × 5 mm) full thickness samples of articular cartilage were collected from each femoral head, snap frozen in liquid nitrogen, and stored in cryovials at −80 oC. The samples were later cut in 4 μm thick cryostat sections and mounted on SuperFrost/Plus glass slides (Eire Scientific, Portsmouth, NH, USA) followed by fixation with acetone for 10 minutes.
Samples of femoral head articular cartilage were obtained from four women (median age 78 years, range 77–85) and two men (age 68 and 71 years) undergoing replacement surgery for osteoarthritis of the hip. All femoral heads contained areas of denuded bone. The overall Collins/McElligott32 grades of osteoarthritis were III and IV. The samples included a spectrum of histological changes; intact cartilage with all zones preserved, non-intact cartilage showing fibrillation and cluster formation, and non-intact cartilage in which the superficial and middle zones were absent and only part of the deep zone remained. A total of up to 12 samples of residual articular cartilage were collected and processed as above.
All samples from both normal and osteoarthritic femoral heads were obtained from white subjects.
Immunohistochemistry was carried out using a Shandon Sequenza (Life Science International, Basingstoke, UK) to achieve consistency of staining quality and to prevent sections from floating off.
The immunohistochemical staining procedure has been described previously.6 11 In brief: non-specific background staining was blocked by incubating the sections for 20 minutes with 100 μl of normal rabbit serum (DAKO, Glostrup, Denmark) in TRIS buffered saline (TBS) (1:5). The primary antibody (100 μl) was diluted in normal rabbit serum in TBS and allowed to incubate for 30 minutes at room temperature. After washing twice in 2 ml of TBS for five minutes, 100 μl biotinylated rabbit antimouse immunoglobulins (DAKO) diluted 1:400 in normal rabbit serum in TBS was added for 30 minutes, followed by another wash. Antibody binding was visualised with an avidin and biotinylated horseradish peroxidase complex (100 μl) (DAKO), which was diluted according to the DAKO protocol and incubated for 30 minutes, followed by washing and 100 μl of 3-amino-9-ethylcarbazole (Sigma, St Louis, MO, USA) for 20 minutes. After another wash, the sections were counterstained with Mayer’s haematoxylin, washed again, and mounted with coverslips.
Table 1 shows the primary antibodies used in this study.
Negative controls were provided by omitting the primary antibody and also by substituting non-immune mouse immunoglobulins (DAKO) for the primary antibody. No attempt was made to block for endogenous peroxidase as no endogenous peroxidase activity was seen in the negative controls. External positive tissue controls included sections of normal human tissue from skin and tonsils. Internal tissue controls were present in many of the sections; integrin positive cells include osteocytes (α4, α5, αV, β1), osteoblasts (α4, α5, αV, β1), osteoclasts (α2, α5, αV, β1, β3), fibroblasts (α2, α5, β1), and endothelial cells (α1, α2, α3, α4, αV, β1, β2, β3, β4, β5).13 33
To assess if the accessibility of integrin subunit epitopes to antibodies was hindered by extracellular matrix interactions, sections were treated with hyaluronidase (10 mg/ml) (Sigma), collagenase (10 mg/ml) (Sigma), or a combination thereof (hyaluronidase 2 mg/ml; collagenase 10 mg/ml) for 5–30 minutes at 37°C before the standard immunohistochemical staining.
As the samples for cryostat sections were all triangularly shaped in their horizontal dimension, a section of intact cartilage sampled and cut correctly would be rectangular in shape in the vertical dimension. Only sections of intact cartilage that clearly displayed the correct vertical orientation and thus contained an articular surface and calcified cartilage with a tidemark were considered for further examination. Also considered for examination were sections of non-intact residual cartilage from osteoarthritic joints that contained calcified cartilage with a tidemark. Only the deep zone or whatever part of the deep zone present (that is, the zone adjacent to the tidemark) was considered for examination in these sections.
The full section of intact cartilage from articular surface to tidemark was divided into quarters, as described previously.6 The quarter of the full section of cartilage that includes the surface was named the superficial zone. The half of the full section of cartilage below the superficial zone was named the middle zone. The quarter of the full section of cartilage below the middle zone and above the tidemark was named the deep zone. The calcified zone of articular cartilage was not included in the evaluation. Occasionally, the superficial zone was covered by a fibrous pannus, which was not included in the zonal division of the section.
Integrin staining for each zone was scored as: 4 = all chondrocytes positive; 3 = more positive than negative chondrocytes; 2 = positive and negative chondrocytes of approximately equal number; 1 = more negative than positive chondrocytes; 0 = all chondrocytes negative. Positive staining was recognised as a dark red colour associated with the cytoplasmic membrane or cytoplasm, or both. As chondrocytes were either strongly positive or exhibited no staining at all no grading of the positivity was made. Areas of sections that contained folding of the cartilage or exhibited non-specific background staining were not scored. Evaluation was carried out by two investigators and the scores given were agreed upon.
The study was approved by the regional ethics committee as part of a larger project involving articular cartilage and bone. Written information about the project was presented to patients scheduled for surgery and patient consent was obtained in these cases.
The Friedman test was used to test the null hypothesis that there is no systematic topographical or zonal pattern of immunohistochemical scores in normal or osteoarthritic cartilage. Tests for systematic topographical pattern were performed for each zone for all normal femoral heads. Tests were not performed for osteoarthritic femoral heads because of insufficient numbers of available sections and, hence, data to compute. Tests for systematic zonal pattern were performed for the superficial zone versus the deep zone for each normal and osteoarthritic femoral head. The Mann-Whitney U test was used to test the null hypothesis that there is no systematic difference between immunohistochemical scores in normal and osteoarthritic cartilage. The limit of significance was chosen as 0.05.
Forty eight to 70 sections of the possible 72 sections of cartilage from the six macroscopically normal femoral heads stained for the various α and β subunits were available for evaluation (table2). Sections were excluded for evaluation because the tidemark was not included or because of excessive non-specific background staining.
Table 2 presents a summary of results for all integrin subunits for articular cartilage from macroscopically normal femoral heads. There was no systematic pattern in the staining with respect to topographical site for any of the integrin subunits (p > 0.05). Only the αV subunit displayed a systematic pattern in the zonal staining (see below). Occasional β4 positive chondrocytes were seen throughout the zones in approximately one third of the sections.
Table 3 presents the zonal distribution of αV in articular cartilage from macroscopically normal femoral heads. All 12 topographical sites had a decreased number of αV positive chondrocytes in the middle and deep zones compared with the superficial zone (fig 1). This zonal pattern was seen in 48% of sections and this represented a statistically significant difference for the superficial zone versus the deep zone for all femoral heads (p ⩽ 0.05). In only one section did chondrocytes in the superficial zone stain to a lesser degree for αV than cells in the middle and deep zones.
Thirty five to 55 sections of the possible 72 sections of cartilage from the six osteoarthritic femoral heads stained for the various α and β subunits were available for evaluation (table 4). Sections were excluded for evaluation because correct orientation could not be determined, no articular cartilage was present in the section, or because of excessive non-specific background staining.
Table 4 presents a summary of results for all integrin subunits for intact and non-intact residual articular cartilage from osteoarthritic femoral heads. The median score for all topographical sites and zones combined is shown with ranges. There was no overt variation in the staining pattern with respect to topographical site for any of the integrin subunits. Only the αV subunit displayed a systematic pattern in the staining with respect to zone (see below) (fig 1). In contrast with chondrocytes in normal articular cartilage, all sections except one contained occasionally α2 positive chondrocytes (fig 2). Occasional β4 positive chondrocytes were seen throughout the zones in approximately half the sections. In approximately one third of sections occasional chondrocytes stained for α4. In approximately 10% of sections occasional chondrocytes stained for β2.
The differences in immunohistochemical staining scores between normal and osteoarthritic cartilage was not statistically significant for the various α and β subunits (p > 0.05) except for the α2 and α4 subunits (p ⩽ 0.05).
Table 5 presents the zonal distribution of αV in intact and non-intact residual articular cartilage from osteoarthritic femoral heads. The zonal pattern observed for normal cartilage was seen in 63% of sections and was statistically significant for the superficial zone compared with the deep zone for all femoral heads (p ⩽ 0.05). In the remaining sections, no zonal variation was observed.
Chondrocytes in clusters and in areas of excessive surface fibrillation showed a similar pattern of integrin subunit staining as isolated chondrocytes and chondrocytes in other areas of the same section.
The pre-treatment of sections with hyaluronidase, collagenase, or a combination thereof before the standard immunohistochemical staining did not reveal additional integrin subunit staining.
This study confirmed the presence of integrin subunits α1, α5, αV, β1, β4, and β5 and the absence of subunits α2, α3, α4, α6, β2, β3, and β6 in normal human articular cartilage.9 11The α3 subunit was absent in this study, but Salter et al 11 found occasional α3 expression in all zones in femoral condylar and tibial plateaus cartilage. Woods et al 9 could not detect the α3 subunit utilising immunohistochemistry, but these authors9 as well as Yonezawa et al 34 confirmed the presence of this subunit by flow cytometric analysis. In this study, the β3 subunit was also absent, confirming studies performed with immunohistochemistry9 11 and flow cytometric analysis.9 34 However, Woods et al 9 detected the heterodimer αVβ3 on chondrocytes within the most superficial 30 μm of normal articular cartilage. Our detection of the β4 subunit in occasional chondrocytes, but not the α6 subunit of the α6β4 heterodimer may represent a similar instance.
To our knowledge, this is the first study detailing the topographical and zonal distribution of integrins in osteoarthritic cartilage. In residual articular cartilage from osteoarthritic human femoral heads, chondrocytes expressed α1, α5, αV, β1, β4, and β5. In addition, we found neo-expression of α2, α4, and β2 in osteoarthritic cartilage. No staining was noted for α3, α6, β3, and β6. Jobanputra et al 35 did not detect the subunits α2 and β4, but did detect the α3 subunit in articular cartilage from the medial tibial plateaux of patients with osteoarthritis secondary to rheumatoid arthritis. We believe these differences in results reflect the differences in the selected material. In extracted chondrocytes from OA knee joints, the subunits α1–6, αV, and β1 were all detected by flow cytometric analysis.36 The differences in results obtained in studies using immunohistochemistry or flow cytometric analysis may reflect either (1) masking of integrin subunits in cartilage sections eliminated in the process of chondrocyte extraction before flow cytometric analysis or (2) neoexpression of integrin subunits in response to the extraction process.
There was no topographical variation in the expression of subunits in either normal or osteoarthritic cartilage. The only subunit that displayed a zonal variation was αV. Minor variations in the expression of subunits with respect to topography or zones may have been missed because of the rather crude immunoscoring system utilised in this study. Interobserver and intraobserver reproducibilities of immunoscoring are rather low (unpublished data), however, the chosen immunoscoring system was found to be reproducible.
We have not attempted to correlate the integrin staining in residual cartilage from osteoarthritic femoral heads with a histopathological score obtainable through a grading system such as the histological-histochemical grading system for osteoarthritic articular cartilage37 or modifications thereof.38 39Recently, we have demonstrated low intraobserver and interobserver reproducibilities and questionable validity of the original histological-histochemical grading system for osteoarthritic articular cartilage.40
The OA articular cartilage selected for this study has been obtained from femoral heads graded III or IV according to the Collins/McElligott system32 and, hence, represents end stage disease at least in some areas. The OA process in the non-intact residual cartilage of grade III and IV joints may thus be different from the OA process taking place in the non-intact residual cartilage of grade I and II joints. Unfortunately, without a reliable histopathological grading system it is difficult to obtain human articular cartilage representative of the different stages of the OA process. Animal studies may further elucidate the expression of α and β subunits of the integrin superfamily during the different OA stages.
From the integrin subunits detected in this and previous9 11 studies, it is possible to deduce the presence of the following integrin heterodimers in normal and OA human articular cartilage: α1β1, α2β1, α3β1, α4β1, α5β1, αVβ1, αVβ3, α6β4, and αVβ5. The β2 subunit may associate with the αL, αM, and αX subunits. The heterodimers α1β1 and α2β1 are known receptors for collagens and laminin; α3β1 for collagens, fibronectin, and laminin; α4β1 for fibronectin, chondroitin sulphate, and VCAM (vascular cell adhesion molecule)-1; α5β1 for fibronectin; αVβ1 for vitronectin and fibronectin; αVβ3 for collagens, fibronectin, vitronectin, thrombospondin, and laminin; α6β4 for laminin; αVβ5 for vitronectin; αLβ2 for ICAM-1/2/3; αMβ2 for ICAM-1, fibrinogen, factor X, and complement factor iC3b; and αXβ2 for fibrinogen, denatured albumin and complement factor iC3b.13 14 41
It is as yet unclear which extracellular matrix molecules human articular chondrocytes adhere to. Studies of the distribution of possible ligands such as fibronectin31 would suggest potential interaction. Such functional relations based upon codistribution is at best indicative and additional, possibly in vitro, studies are needed to unravel the precise role of integrin subunits in articular cartilage.
The significance of the neoexpression of α2, α4, and β2 in residual cartilage from osteoarthritic femoral heads remains to be elucidated. The α2 subunit is strongly expressed in fetal cartilage10 and the neoexpression in osteoarthritic cartilage of this subunit may represent an example of “fetal phenotype” expression in osteoarthritic cartilage potentially as a functional response to production of collagen type I and III.26
In conclusion, our study demonstrates the expression of α1, α5, αV, β1, β4, and β5 in normal and osteoarthritic cartilage. In osteoarthritic cartilage, neoexpression of α2, α4, and β2 is observed. There is no topographical variation in the expression of subunits in either normal or osteoarthritic cartilage. The only subunit that displays a zonal variation is αV; more chondrocytes express αV in the superficial zone compared with the deep zone.
We gratefully acknowledge Professor G Bendixen and Drs M Brittberg and A Jordan for critical reviews of the manuscript and Associate Professor L Theil Skovgaard for helpful advice regarding statistical methods. Dr B Volck is acknowledged for the sampling of osteoarthritic cartilage. The skilful technical assistance of V Weibull and G Dahl is highly appreciated. The Michaelsen Foundation, the Danish Rheumatism Association, the Danish Medical Research Council, the Danish Biotechnology Program, the Arthritis and Rheumatism Council, and Direktør E Danielsen og Hustrus Fond are acknowledged for their financial support.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.