Acute HNs are a feature of pre-radiographic hand OA and these gelatinous synovial cysts are thought to be a forerunner of new bone formation
[1–4]. The nature of the gelatinous material (being rich in hyaluronan
[2, 14]) and the new bone formation that occurs in OA at these locations lead us to the hypothesis that MSCs may become entrapped from the joint SF in such material in the earliest stages of OA. Herein, we show the presence of HN resident cells which met functional, phenotypic, and transcriptional profiles of MSCs that are closely related to SF-MSCs
Although HNs are a common form of generalized nodal OA, clinical presentation at the acute phase, prior to ossification, is somewhat rare; only three cases have appeared in our clinic in the past five years
[1, 3]. Consequently, only a small number of such gelatinous cysts could be sampled. Despite these low sample numbers, the functional, phenotypic, and transcriptional data presented here are consistent with these cells being MSCs. There was, however, a surprising difference in the capacity of the HN-MSCs to differentiate toward the adipogenic lineage. Functional assays showed only isolated cells with the accumulation of micro-vesicles in each donor, despite the expression of the adipogenic transcription factors CEBPA (CCAAT/enhancer-binding protein) and PPRAG. A lack of adiopogenesis has been linked to ‘in vitro’ aging of MSCs, whereby the ability of cells to differentiate is reduced as the cells approach senescence
[25, 26]. Owing to the prolonged culture expansion period needed to expand HN-MSCs to sufficient numbers to complete our functional, phenotypic, and expression assays, our cells may have been approaching senescence, which has reduced their capacity to differentiate toward the adipogenic lineage.
Differential transcriptional analysis has previously been shown to reliably distinguish MSCs from FBs and subdivide MSCs from their tissue of origin by identifying commonly expressed genes when cells are grown under the same conditions
[27–30]. Here, we used a similar experimental approach and confirm our previous findings
, indicating mesenchymal lineage-specific expression of genes on our custom array (Table
1). We verified that FZD1 (frizzled family receptor 1), IGFBP3 (insulin-like growth factor-binding protein 3), PAPSS2 (3′-phosphoadenosine 5′-phosphosulfate synthase 2), and VEGFC (vascular endothelial growth factor C) are differentially expressed between MSCs and FBs
 and also newly identify genes differentially expressed between MSCs and FBs such as GDF5 (growth differentiation factor-5) and HAPLN1 (hyaluronan and proteoglycan link protein 1), which are involved in skeletal development, cartilage formation, and maintenance
[24, 31]. Further to defining the HN cells as MSCs, our transcriptional analysis indicates that these MSCs share a gene-specific expression profile more comparable to SF-MSCs. We compared HN- and SF-MSCs not only with normal ICBM-MSCs but also with BM from OA patients. For ethical reasons, OA BM could not be collected from the iliac crest of these patients. Rather, MSCs were derived from two sources: (a) the medullary canal of the femur and (b) trabecular bone of femoral head from OA patients all undergoing hip arthroplasty. We have previously described these MSCs as functional and phenotypic indistinct from their ICBM counterparts
[15, 16]. Such analysis allowed joint/BM-specific differences to be identified rather than differences relating to OA. Among genes expressed by HN- and SF-MSCs (but downregulated in BM-MSCs) were the recently identified marker of synovium MSCs, SFPR4
, and several genes known to be associated with cartilage formation and maintenance, such as ACAN, BMP4, PRELP, and TNFIP6 (Figure
4 and Table
[22, 24, 31, 32]. To fully confirm the tissue origin of MSCs in vivo and to establish their migration patterns following injury or in disease, lineage tracing experiments in animal models are required
. Nevertheless, our transcript data as well as previous findings
[27–30] indicate that broad tissue or niche-specific transcriptional profiles are maintained in culture and as such can be informative on the MSC ‘in vivo’ functionality in a given environment.
With both normal human and animal joints having resident SF-MSCs
, it is possible that HN-MSCs are entrapped from SF extruded through sites of weakness in the joint capsule
 and that, like the knee, small joints contain a resident SF-MSC population. Previous gene profile analysis indicates that SF-MSCs in turn originate from the adjacent synovium
[8, 9], and our data suggest a similar origin for HN-MSCs based on transcriptional analysis. Our data cannot exclude the possibility that HN-MSCs are BM-derived but transcriptionally ‘altered’ following their entrance into the joint environment; this is, however, unlikely as no damage to bone and hence the entry route for these MSCs can be identified at this early stage of OA. We therefore conclude that, similar to SF-MSCs, HN-MSCs are most likely of synovial origin. Although HNs are forerunners of new bone formation, no significant osteogenic bias was seen in any HN-MSC cultures (either by qPCR or differentiation assay); this may be explained by the fact that, in this study, HN material was collected early in disease, prior to ossification. As such, our data suggest that these MSCs are unlikely to have intrinsic osteogenic bias, perhaps with the aberrant new bone formation seen in HNs a result of subsequent alterations in joint biomechanics.