Relative percentage and zonal distribution of mesenchymal progenitor cells in human osteoarthritic and normal cartilage

Abstract

Introduction: Mesenchymal stem cells (MSC) are highly attractive for use in cartilage regeneration. To date, MSC are usually recruited from subchondral bone marrow using microfracture. Recent data suggest that isolated cells from adult human articular cartilage, which express the combination of the cell-surface markers CD105 and CD166, are multi-potent mesenchymal progenitor cells (MPC) with characteristics similar to MSC. MPC within the cartilage matrix, the target of tissue regeneration, may provide the basis for in situ regeneration of focal cartilage defects. However, there is only limited information concerning the presence/abundance of CD105+/CD166+ MPC in human articular cartilage. The present study therefore assessed the relative percentage and particularly the zonal distribution of cartilage MPC using the markers CD105/CD166.

Methods: Specimens of human osteoarthritic (OA; n = 11) and normal (n = 3) cartilage were used for either cell isolation or immunohistochemistry. Due to low numbers, isolated cells were expanded for 2 weeks and then analyzed by flow cytometry (FACS) or immunofluorescence in chamber slides for the expression of CD105 and CD166. Following immunomagnetic separation of CD166+/- OA cells, multi-lineage differentiation assays were performed. Also, the zonal distribution of CD166+ cells within the matrix of OA and normal cartilage was analyzed by immunohistochemistry.

Results: FACS analysis showed that 16.7 ± 2.1% (mean ± SEM) of OA and 15.3 ± 2.3 of normal chondrocytes (n.s.) were CD105+/CD166+ and thus carried the established MPC marker combination. Similarly, 13.2% ± 0.9% and 11.7 ± 2.1 of CD105+/CD166+cells, respectively, were identified by immunofluorescence in adherent OA and normal chondrocytes. The CD166+ enriched OA cells showed a stronger induction of the chondrogenic phenotype in differentiation assays than the CD166+ depleted cell population, underlining the chondrogenic potential of the MPC. Strikingly, CD166+ cells in OA and normal articular cartilage sections (22.1 ± 1.7% and 23.6% ± 1.4%, respectively; n.s.) were almost exclusively located in the superficial and middle zone.

Conclusions: The present results underline the suitability of CD166 as a biomarker to identify and, in particular, localize and/or enrich resident MPC with a high chondrogenic potential in human articular cartilage. The percentage of MPC in both OA and normal cartilage is substantially higher than previously reported, suggesting a yet unexplored reserve capacity for regeneration.

Figure 1

Image of a representative osteoarthritis cartilage/bone specimen obtained during a total joint replacement surgery. Locations used for the harvesting of cartilage are indicated by red lines, and excluded regions are encircled with black dotted lines.

Figure 2

FACS analysis of CD105 and CD166 expressions on isolated normal and osteoarthritis (OA) chondrocytes and enrichment of CD166⁺ OA cells after immunomagnetic separation. Enzymatically isolated, non-separated OA (n = 11 patients) and normal (n = 3) chondrocytes were subjected to FACS analysis after initial in vitro culture (a). After magnetic separation of CD166⁺ cells, both the positive (b) and the negative (c) fractions were studied for the expressions of CD105 and CD166. The bar charts show the mean ± standard error of the mean (SEM) for all analyzed patients; for each section, a representative example of the FACS measurement is included as a dot plot diagram, in which CD105⁺/CD166⁺ cells are located in the upper right quadrant. In addition, the relative expressions of CD105 and CD166 (colored lines) and the corresponding isotype controls (filled graphs) are shown in histograms. FACS, fluorescence-activated cell sorting; FITC, fluorescein isothiocyanate; RPE, R-phycoerythrin.

Figure 3

Immunohistochemical staining of normal and osteoarthritis (OA) chondrocytes after isolation and initial adherence. Representative images of an OA chondrocyte sample are shown in (a-d); staining results for normal chondrocytes were identical (not shown). Nuclei were visualized with DAPI (4',6-diamidino-2-phenylindole) (a). Cell membranes were double-stained for CD105 by using an Alexa Fluor 488-coupled secondary antibody and for CD166 by using an Alexa Fluor 594-coupled secondary antibody and analyzed by fluorescence microscopy. Co-expression of CD105 and CD166 was identified after computed superimposition of fluorescence signals from both channels. Almost all cells are CD105⁺ (b), whereas only a few cells express CD166 (c); but in all cases, this is accompanied by the co-expression of CD105 (d). In addition, the results of the quantitative analysis of normal (e) and OA (f) chondrocytes are shown. Magnifications: 100 ×, 400 × (insets). SEM, standard error of the mean.

Figure 4

Microscopic images and corresponding FACS analysis of chondrocytes before and after immunomagnetic separation. After incubation with bead-coupled antibodies to the surface antigen CD166, both membrane-bound and free magnetic anti-CD166 particles are visible, as are non-labeled chondrocytes (a1). In the positive fraction after separation, only magnetobead-covered CD166⁺ cells and free magnetobeads are present (b1). In the negative fraction, only CD166^- cells without magnetobeads are present; also, no free magnetobeads are present (c1). These results were confirmed by FACS analysis, demonstrating an enrichment of CD166⁺ cells from 10% before (a2) to approximately 87% after (b2) separation, whereas the negative fraction contained only a very low percentage of CD166⁺ cells (c2). FACS, fluorescence-activated cell sorting; FITC, fluorescein isothiocyanate.

Figure 5

Cell differentiation studies with CD166⁺-enriched and -depleted chondrocytes. Isolated osteoarthritis chondrocytes were immunomagnetically enriched (a,c,e) or depleted (b,d,f) for CD166⁺ cells and subjected to adipogenic, osteogenic, and chondrogenic differentiation. CD166⁺-enriched as well as CD166⁺-depleted cell populations differentiate to adipocytes after culture in adipogenic medium, as detected by oil red O staining (a,b). After cultivation with osteogenic medium, both cell populations demonstrate differentiation toward osteoblasts observable as positive alkaline phosphatase activity within the cells (c,d). However, this effect is more pronounced in cells previously enriched for CD166⁺ chondrocytes (c). Pellet cultures of cells that underwent chondrogenic differentiation are visualized by Alcian blue staining of newly deposited proteoglycans (e,f). Exclusively, the population with CD166⁺-enriched cells shows a clear chondrogenic phenotype (e).

Figure 6

Distribution of CD166⁺ chondrocytes within the osteoarthritis (OA) joint cartilage matrix. CD166⁺ cells, characterized by an intense brown staining of the cell membrane and pericellular matrix, are localized primarily in the superficial and middle cartilage zones of OA cartilage. In contrast, almost no CD166⁺ cells are in the deep and calcified cartilage zones. The corresponding isotype control of the immunohistochemical mesenchymal progenitor cell staining shows an absence of a positive signal, supporting the specificity of the above-named CD166 staining. Magnifications: 100 × (large micrograph), 400 × (insets). Results of semiquantitative analysis of cartilage samples from six patients with OA are expressed as mean ± standard error of the mean (SEM).

Figure 7

Distribution of CD166⁺ chondrocytes within the normal joint cartilage matrix. CD166⁺ cells within normal cartilage matrix show a distribution pattern similar to that of osteoarthritic cartilage (shown in Figure 6), and the absence of a positive staining signal in the corresponding isotype control proves the specificity of the CD166 staining. Magnifications: 100 × (large micrograph), 400 × (insets). Results of semiquantitative analysis of cartilage samples from three normal donors are expressed as mean ± standard error of the mean (SEM).