Characterizing heterogeneity in the response of synovial mesenchymal progenitor cells to synovial macrophages in normal individuals and patients with osteoarthritis

The 2001 discovery of resident mesenchymal progenitor cells within the synovial membrane (sMPCs) of the joint presented an opportunity to understand the limited self-renewal and healing capacity of articular cartilage [1, 2]. Mesenchymal progenitors have the potential to differentiate into bone, fat, and importantly cartilage. However, in diseases such as osteoarthritis (OA), this chondrogenic phenotype appears to be modified for yet unknown reasons [3, 4]. This observation is of great interest given the potential cell-based therapies that can be derived using resident sMPCs. Interestingly, it has also been observed that inflammation of encompassing synovium (synovitis) and corresponding intimal hyperplasia, subintimal macrophage infiltration and increased vascularity are not observed without involvement of cartilage disturbance in OA – thereby implicating the synovium in disease pathogenesis [5‐9]. This raises particular questions of whether components of the synovial inflammatory response are contributing to the loss of sMPC chondrogenic phenotype.

Macrophages are immune cells derived from the monocyte lineage that have been demonstrated to play roles in nearly all aspects of the immune system, tissue remodeling and wound healing [10]. In recent years, the regulation and function of macrophage sub-types has become quite complex, but most agree that there are at least three distinct subtypes; M0 (or inactive), M1 (pro-inflammatory) and M2 (anti-inflammatory) [11]. A number of studies have also demonstrated sub-sets within each of these subtypes that can be distinguished by cell surface receptors and protein secretion. Macrophages are the primary inflammatory custodians of the synovial tissue and can partake in pro- (M1) and anti- (M2) inflammatory activities [12]. To date however, there has been limited characterization of the synovial macrophage; the components of its inflammatory milieu (across patient populations and disease severity), and if these cells/properties regulate/modify sMPC function and/or potential. Early findings from Bondeson et al. [6] attempting to deplete digested OA synovium cultures of macrophages (using anti-CD14 conjugated beads) found reduced levels of cytokines known to degrade cartilage, such as TNFα, IL-1β, IL-6, IL-8, MMP-1 and MMP-3. However, despite such findings, this study did not assess the sMPC population, or changes to its differentiation capacity (if any) following macrophage depletion. A recent study by Fahy and colleagues in 2014 explored this issue, exposing sMPCs to conditioned media (CM) from either digested OA synovium or peripheral blood mononuclear cell-derived M1/M2 macrophages [7]. The authors found that an anti-inflammatory supernatant was more conducive to sMPC chondrogenesis compared to a pro-inflammatory supernatant. Such findings encouraged the present study, with specific interest to assess variations across the patient population.

While the sMPC-macrophage relationship remains undefined in OA, macrophages have demonstrated both promoting and hindering roles on MPC activity in other tissues. For example, tumorigenic MPCs were found to be activated by macrophages (and their inflammatory cytokines) in the tumor site of gastric cancers [13]. Meanwhile, macrophages were also found to be critical regulators of MPC osteogenic differention via STAT3 signalling [14]. Additionally, M2-phenotype macrophages and their associated cytokines supported MPC-graft survival in the infarct site of damaged heart muscles compared to an M1-phenotype macrophage and associated cytokines [15]. Such studies are noteworthy, as they help to illustrate the dynamic MPC-macrophages relationship across normal and diseases states. The goal of the present study was to study endogenous synovial macrophages and their interactions with sMPC chondrogenesis in normal individuals and OA patients using a novel explant system. By manipulating the presence and/or phenotype of synovial macrophages, we hypothesized that the innate chondrogenic capacity of OA sMPCs could be attenuated. Our findings support this hypothesis in patient-specific manner, and also demonstrate a systemic shift in the inflammatory phenotype of synovial tissue itself in response to cytokine stimulation and/or macrophage depletion.

The aim of the present study was to elucidate the relationship between synovial macrophages and synovial MPCs. This was prompted by literature findings that macrophages are drivers of synovitis and cartilage loss in OA, and are thus likely candidates to regulate sMPC chondrogenic differentiation [6, 7, 16, 21]. Other studies have hypothesized and reported that an M1-macrophage phenotype and its mediators have anti-chondrogenic effects [6, 7]. Despite these findings, the present study is the first of its kind (to our knowledge) to show a heterogeneous and patient-specific response following attenuation of macrophages state (via cytokine stimulation or depletion) on sMPC chondrogenesis.

Using a synovial explant model that was to be subjected to pro- and anti-inflammatory cytokine stimulation and/or macrophage depletion, we first confirmed the presence of endogenous macrophages using flow cytometry and immunofluorescent staining (Table 2 and Additional file 1: Figure S1). Post 21-day sMPC pellet-culture chondrogenesis, qRT-PCR was used to quantify the differentiation process. Pooled sMPC chondrogenic gene expression data from 8 OA patients and 4 normal individuals revealed a non-treatment specific up-regulatory response in the levels of Acan, Col2a, and Sox9 expression (Fig. 2) when the explant (not the purified sMPCs) had been treated with IFNγ, TNFα, IL-4 or IL-10 in the presence or absence of clodronate. The nature of these findings prompted closer inspection of the data on an individual patient-specific basis.

Assessment of the qRT-PCR findings of all 8 OA patients and 4 normal individuals revealed a heterogeneous response to all treatments in the OA patient derived sMPCs, in which macrophages promoted, inhibited, or had no apparent effect on sMPC chondrogenesis. For example, in OA Patient 1, sMPC Col2a expression was increased following TNFα treatment compared to TNFα treatment alongside macrophage depletion (clodronate), while, in contrast, sMPCs from OA Patient 4 demonstrated an increased in Col2a expression in the presence of TNFα and clodronate. Patient 3 and 5 demonstrated a near global decrease in chondrogenic gene expression across all treatment groups, while sMPCs from patients 1, 7 and 8 demonstrate a wide variety of responses to all treatment groups (Fig. 2) apparently irrespective of macrophage presence. Overall, these varying treatment responses illustrate the heterogeneity between the sMPCs of OA patients and reinforce the idea that OA is a spectrum disease with numerous avenues (injury, metabolic, idiopathic) that can lead to a clinical diagnosis [22]. Taken together, it is difficult to interpret the significant differences observed in the pooled data because of the wide and varied response observed within the 8 OA patient samples. This could speak to differences in age, sex and/or severity of disease in the patients, but without a larger sample size it would be inappropriate to stratify and comment on these variables.

The changes in the normal and OA explant inflammatory secretome profiles of during the 12-day treatment phase provide complementary data to the qRT-PCR analysis. These findings shed light on the response of normal and OA synovial tissue to cytokine stimulation (IFNγ, TNFα, IL-4 or IL-10) in the presence or absence of macrophage depletion (clodronate). Interestingly, in normal synovial samples, no profile level changes (all 41 analytes compared together) were observed within or between treatment groups, however, at the individual protein level, it was observed that MCP-1/CCL2 was differentially expressed in under all treatment effects. Our group has previously implicated MCP-1/CCL2 in the loss of chondrogenic capacity of synovial MPCs [21]. In that study we were able to demonstrate that MCP-1/CCL2 doesn’t impede the proliferation MPCs, but only decreased their chondrogenic capacity. Since MCP-1/CCL2 was the only protein that appeared to be regulated by cytokine stimulation with macrophage depletion, coupled with the observation that only a very few treatments reduced the chondrogenic marker expression in normal sMPCs; suggests that the normal synovial explants may have the ability the ‘buffer’ the sMPCs against cytokine stimulation and this may not dependent on the presence of synovial macrophages. Additionally, these results seem to suggest that cytokine stimulation enhances the chondrogenic differentiation of the synovial MPCs. While numerous previous studies have demonstrated that pro-inflammatory cytokines (e.g. TNFα, IL-1β) inhibit the chondrogenic capacity of stem cells [23, 24], it is important to recognize that in the current study, the purified MPCs were never exposed directly to the cytokines, instead the intact synovial biopsies were exposed the cytokines, the cytokines were then removed and MPCs were then derived, purified and differentiated. This could suggest, that while certain cytokines can regulate the capacity of stem cells/MPCs, that when the cells are still within their niche, interaction with the inflammatory micro-environment and other tissue resident cells can lead to distinct outcomes.

The analysis of the OA patient inflammatory secretomes, demonstrated distinct responses compared to the normal explants. Specifically, a number of cytokines and chemokines were differentially expressed based on cytokine stimulation and the state of macrophages in the samples. While there was no single protein that was found across all treatment groups, MCP-3, MIP-1α, and VEGF were all observed in half of the treatment groups. Interestingly, IL-10 was observed in the IFNγ treatment group, while IL-4 was observed to be differentially regulated in the presence of IL-10. Overall, it would appear that OA synovium may no longer have the ability to ‘buffer’ the cytokine stimulation and that effect appears to be regulated by the presence of synovial macrophages. These findings lend support to the idea that the macrophages may have a role to propagate and sustain cytokines being secreted within the synovium, as proposed in the cytokine theory of OA pathogenesis [25‐28]. When the inflammatory secretomes of all the samples were analyzed by dendrogram clustering analysis, it was observed that the normal samples always grouped together regardless of treatment group, or if clodronate was present or absent (Fig. 4). Interestingly, the OA samples appeared to group together based on if clodronate was present or absent, however, the same samples didn’t always group together under each treatment group (IFNγ, TNFα, IL-4 or IL-10). For one specific example; samples from patient 1 and 8 without clodronate group together in the presence of TNFα, IL-4 and IL-10, but don’t group together after treatment with IFNγ. This suggests that explants from different patients respond differently to specific cytokines and that this is at least partially dependent on the presence of macrophages in that synovial explant. While a sample size of eight is not large enough to make any definitive conclusions, this could be part of the reason that OA patients demonstrate a wide response to anti-cytokine treatments [29, 30]

Overall, inflammatory secretome data from OA and normal patients reveal a pathology-dependent response, where normal patients did not display observable levels of any analytes (except for MCP-1/CCL2) following cytokine treatments in the presence or absence of macrophages, while OA patient synovium reacted both to the cytokine stimulation and macrophage depletion. Such findings from the normal patients maybe reflective of state of synovial homeostasis in their joints that is ‘buffered’ to changes in the microenvironment. These findings are consistent with previous studies, which have suggested that a healthy synovial joint can maintain a balanced microenvironment [31]. It would be of interest to find exactly when the normal synovium loses the ability to ‘buffer’ cytokine stimulation with the onset and progression of OA. This would most likely require a pre-radiographic and pre-symptomatic patient population that would be difficult to characterize. Additionally, since the normal synovial ‘buffering’ ability did not appear to be solely dependent on the presence of macrophages, it would be of interest to elucidate exactly which cell population(s) are either: providing this capacity in normal tissues, or alternatively, are responsible for the dramatic changes in inflammatory response observed in OA tissues.

This study was not without limitations. The most salient of which is the number of enrolled patients (8 OA and 4 normal). Additionally, despite stringent inclusion/exclusion criteria, the spectrum of disease severity amongst the OA samples was uncontrolled for in this study, and was likely reflected (to some extent) in the degree of heterogeneity observed in the qRT-PCR data and Luminex. Future studies will seek to increase patient enrolment such that early, mid, and late stage OA cohorts could be assessed separately. Beyond this, our study did not assess how the differentially regulated analytes in the OA explant supernatant in response to the cytokine treatments would have ultimately affected the chondrogenic phenotype of sMPCs thereafter. These may serve as novel targets to study, as they may be implicated in the sMPC-macrophage relationship.

Using a synovial explant model, we have shown that the chondrogenic capacity of sMPCs can be regulated (positively or negatively) by modifying the endogenous synovial macrophage population (via cytokine stimulation and/or depletion). This study has also revealed that the inflammatory phenotype of the OA synovium itself can be responsive to changes in the macrophage population. These findings lend support to the idea that macrophages are intricately involved in sMPC chondrogenesis and are therefore an important cell type for further investigation, and future therapy development.