The current study was designed to determine the localization of cells expressing putative progenitor markers in normal and OA human articular cartilage. The three selected candidate markers Notch-1, Stro-1, and VCAM-1 have been widely used to identify bone marrow MSC [23
]. Staining patterns for the three markers in normal human articular cartilage were similar with significantly higher staining frequency in the SZ as compared with the MZ and DZ. This is consistent with observations from other laboratories using the same or other stem cell markers [12
]. Using IHC we observed a surprisingly high frequency of cells expressing Notch-1, Stro-1, and VCAM-1 throughout normal human articular cartilage. Using flow cytometry as an alternative method to detect Notch-1 and Stro-1 we observed lower levels of positive cells as compared with IHC. Furthermore, although the percentage of Notch-1 and Stro-1-positive cells was similar by IHC, the flow cytometry results showed much higher expression of Stro-1 as compared with Notch-1. As we demonstrated specificity of the IHC signals, these results suggest that profound changes in the expression of these markers occur upon cell isolation and that the patterns of change are different for each marker. This change could either be the result of a downregulation of protein expression in monolayer culture, indicate a sensitivity to exposure to collagenase digestion, previously demonstrated for numerous surface molecules on human articular chondrocytes [50
] or be because of preferential loss of cells expressing these markers during the isolation process. Enzymatic digestion of cartilage recovers less than 22% of the total number of cells present in the original tissue [51
], indicating that certain subpopulations such as those expressing progenitor markers may be lost.
Given the unexpected high levels of Notch-1, Stro-1, and VCAM-1-positive cells in cartilage, we applied an additional means of identifying stem cells. The Hoechst dye 33342, which defines the so-called SP, was used with freshly isolated cells from human articular cartilage and on flow cytometry we observed that the SP represented only 0.1% of the cells. This frequency is similar to that reported for young bovine cartilage [52
]. However, this is vastly different from the frequency of Notch-1, Stro-1, and VCAM-1-positive cells. The Hoechst dye thus appears to be a more appropriate stem cell marker.
In the present study we did not examine whether the cells expressing Notch-1, Stro-1, or VCAM-1 had multilineage differentiation capacity. Our previous study [10
] and Dowthwaite and colleagues [38
] demonstrated that cartilage cells expressing CD105/CD166 or Notch-1 do indeed have stem cell activities. However, Karlsson and colleagues [53
] recently demonstrated that Notch-1 is not a progenitor marker in cartilage. To reconcile these observations in reflection to this current data set, it is most plausible that a subpopulation of these identified progenitor-positive cells is multi-potent, which is represented by the SP. Further surface molecule characterization of the cartilage SP is required.
Based on the differences in the frequency of Notch-1, Stro-1, or VCAM-1-positive cells versus SP cells, these represent very different cell populations. We propose that the observed high frequency of progenitors in cartilage is a reflection of multiple functions that these progenitor molecules have in the native tissue such as controlling cell fate, proliferation, and apoptosis [30
]. On the other hand, cartilage may contain a very high proportion of progenitor cells due to its avascular quality. Frequency of Notch-1-positive cells among different human tissues ranges from 0 to more than 60% [56
]. The concentration of stem cells in the SZ and on the surface of developing human cartilage is also consistent with a recent report [57
] showing that during postnatal development of rabbit knee joints, the SZ contains stem cells that supply a rapidly dividing, transit-amplifying daughter-cell pool. Following cessation of growth and attainment of joint maturation the stem cell pool in the SZ may provide a reservoir for replenishing cells in the cartilage surface that is the site of biomechanical load and wear. Based on the present results this cellular organization appears also present and maintained in mature human articular cartilage.
This study is the first to analyze changes in the distribution of stem cell markers in OA affected human articular cartilage. High Stro-1 protein expression levels have been observed in OA synovium cell clusters [37
] and the soluble form of VCAM-1 has been implicated in rheumatoid arthritis and OA [58
]. We have previously reported increased expression of Notch-1 in OA cartilage [14
] and a recent study indicates that Notch-1-positive cells and its signaling components, Jagged1 and Hes5, are upregulated in OA and mediate cell proliferation [40
]. The reduction in both Notch-1 and Sox9 in the SZ non-fibrillated OA cartilage is notable because this implies a reduction in progenitor cells and probably normal cartilage ECM production, respectively. This shift may be a consequence of aging and such cell depletion may be an important initiator or a predisposing factor leading to OA development. We have recently demonstrated co-ordination between Notch-1 and Sox9 signaling to either inhibit or promote chondrogenesis [60
]. Imbalance between these pathways may be an inherent feature of OA and a possible therapeutic target.
OA cartilage is characterized by cell cluster formation and abnormal cell differentiation processes with renewed expression of cartilage development related extracellular matrix components [61
]. Genes attributed to dedifferentiated (collagen types I and III, fibronectin) and hypertrophic chondrocytes (collagen type X) are also detected in OA clusters [6
]. Based on the present observations, the cells that compose these clusters are likely to be a result of proliferating chondroprogenitors. Aigner and colleagues [64
] indicated that MZ cells are principally activated in OA tissue and these cells express type IIA procollagen, indicative of the chondroprogenitor phenotype, which is in agreement with our current observations of increased progenitor markers in the same area of OA cartilage. Fukui and colleagues [61
] showed the most profound phenotypic shift as indicated by the expression of type II collagen and fibronectin in OA fibrillated areas where clusters are prominent. Understanding the basis of such aberrant chondrocyte responses and whether resident progenitor cells are involved will be vital for the development of therapies and diagnostic markers to control and prevent OA progression.
Results from the marker staining patterns in OA cartilage show several changes as compared with normal tissue. However, the type of change is also specific for each marker. For example, there is a marked decrease in Notch-1 in the DZ of fibrillated OA cartilage but VCAM-1 is increased by four-fold. These divergent changes further suggest that the selected cell surface receptors are at least in part independently regulated as part of the cell activation process in OA and do not represent suitable stem cell markers in cartilage.