Morphofunctional analysis of fibroblast-like synoviocytes in human rheumatoid arthritis and mouse collagen-induced arthritis

Fibroblast-like synoviocytes (FLS) play a prominent role in rheumatoid synovitis and degradation of the extracellular matrix through the production of inflammatory cytokines and metalloproteinases (MMPs). Since animal models are frequently used for elucidating the disease mechanism and therapeutic development, it is relevant to study the ultrastructural characteristics and functional responses in human and mouse FLS. The objective of the study was to analyze ultrastructural characteristics, Interleukin-6 (IL-6) and Metalloproteinase-3 (MMP-3) production and the activation of intracellular pathways in Fibroblast like synoviocytes (FLS) cultures obtained from patients with rheumatoid arthritis (RA) and from mice with collagen-induced arthritis (CIA). FLSs were obtained from RA patients (RA-FLSs) (n = 8) and mice with CIA (CIA-FLSs) (n = 4). Morphology was assessed by transmission and scanning electron microscopy. IL-6 and MMP-3 production was measured by ELISA, and activation of intracellular signaling pathways (NF-κB and MAPK: p-ERK1/2, p-P38 and p-JNK) was measured by Western blotting in cultures of RA-FLSs and CIA-FLSs stimulated with tumor necrosis factor-alpha (TNF-α) and IL-1β. RA-FLS and CIA-FLS cultures exhibited rich cytoplasm, rough endoplasmic reticula and prominent and well-developed Golgi complexes. Transmission electron microscopy demonstrated the presence of lamellar bodies, which are cytoplasmic structures related to surfactant production, in FLSs from both sources. Increased levels of pinocytosis and numbers of pinocytotic vesicles were observed in RA-FLSs (p < 0.05). Basal production of MMP-3 and IL-6 was present in RA-FLSs and CIA-FLSs. Regarding the production of MMP-3 and IL-6 and the activation of signaling pathways, the present study demonstrated a lower response to IL-1β by CIA-FLSs than by RA-FLSs. This study provides a comprehensive understanding of the biology of RA-FLS and CIA-FLS. The differences and similarities in ultrastructural morphology and important inflammatory cytokines shown, contribute to future in vitro studies using RA-FLS and CIA-FLS, in addition, they indicate that the adoption of CIA-FLS for studies should take careful and be well designed, since they do not completely resemble human diseases.


Background
Rheumatoid arthritis (RA) is the most common inflammatory rheumatic disease. Destructive damage of the joints, resulting in physical disability, is the ultimate result of bone and cartilage damage mediated by the production of chemokines and matrix components by synoviocytes [1]. The etiopathogenesis of RA involves an intricate network of cellular interactions in the synovial membrane characterized by the proliferation of cells in the synovial lining, resulting in hyperplasia, pannus formation, and tissue destruction [1,2].
Fibroblast-like synoviocytes (FLSs) are mesenchymalderived cells responsible for providing support, nourishment and lubrication to the joint tissue. FLSs one of the most prominent cells in inflamed tissue in RA [2]. In the chronic inflammatory milieu of the rheumatoid synovium, these cells become autonomous, hyperplastic, and invasive and produce large amounts of proinflammatory cytokines, chemokines, and matrix-degrading enzymes [2]. Among these products, metalloproteinases (MMPs) are of fundamental importance in the pathophysiology of RA by promoting bone and cartilage degradation [3,4] Stromelysin (MMP-3) is one of the most important factors in RA [5,6]. In addition, FLSs produce interleukin 6 (IL-6), an important cytokine in the inflammatory cascade, leading to the interaction of FLS and other immune cells in the synovium and exacerbating inflammatory processes [7][8][9][10]. Previous studies of FLSs from patients with RA indicated that both IL-6 and MMP production increased after IL-1 and TNF-α stimulation and were dependent on the NF-κB and MAPK pathways (ERK, JNK and p38), ultimately contributing to the pathogenesis of RA [11][12][13][14][15][16].
Animal models have contributed to improved understanding of human disease and provide a useful tool for therapeutic testing. However, the therapies for humans' diseases have developed, they are more precise and specifically targeted, and it becomes gradually important to understand the limitations of generalizing data from mice to humans. Discrepant results from animal and human studies are worrisome, and efforts to evaluate morphological and functional similarities between animal and human cells are required. One of the most commonly used experimental models to study RA is murine collagen-induced arthritis (CIA). This animal model is characterized by peripheral symmetrical joint involvement that shares many histopathological features with human RA, including synovitis, pannus formation, cartilage and bone erosion [17,18]. Furthermore, as in patients with RA, IL-6 and MMP-3 also play key roles in this model [19][20][21]. However, we could only find one study showing the effect of IL-1β and TNF-α in FLSs in CIA [20].
Primary FLS cultures from CIA and RA patients have been used as an in vitro exchange model to study the responses of these cells to different stimuli and pharmacological drugs. There have been no studies that exploring the morphological ultrastructures and responses of cultures of FLSs derived from patients with RA (RA-FLS) and mice with CIA (CIA-FLS). Our objective was to investigate RA-FLS and CIA-FLS cultures regarding their ultrastructural characteristics as well as their functional responses, such as the signaling pathways activated and the production of IL-6 and MMP-3 after stimulation with TNF-α and IL-1β.

Sample collection and culture of RA patient cells
Synovial fluid was obtained from eight (n = 8) RA patients according to the American College of Rheumatology (ACR) criteria [22]. The mean age of the patients was 52.62 years (range 37-66 years). Samples of synovial fluid were collected and centrifuged at 1200 revolutions per minute (rpm) for 10 min. The pellet was resuspended in complete medium (DMEM high glucose plus 10% fetal bovine serum [FBS, Gibco, Life Technologies, USA], 1% streptomycin(S)/penicillin(P) and amphotericin B [Gibco, Life Technologies, USA] and 1% nonessential amino acids [Gibco, Life Technologies, USA]) and kept at 37 °C and 5% CO 2 in cell culture flasks. The culture medium was exchanged every three days until the cells were frozen. FLSs were used for experiments after five passages. The Ethics Committee approved the protocol (CAAE Human Research Ethics Committee: 08387918.6.0000.5149). Informed consent was obtained from all participants in the study.

Sample collection and culture of mouse collagen-induced arthritis cells
Four (n = 4) male DBA1/J mice (8-12 weeks, average weight of 20 g) were used to obtain CIA-FLSs. The animals were immunized with 100 μL of emulsions containing equal parts of Freund's complete adjuvant containing 5 mg/ml heat-killed Mycobacterium tuberculosis antigen (strain H37Ra; Difco, Lawrence, USA) and 2 mg/ml type II collagen (Chondrex; Washington, USA) in 10 mM acetic acid on day 0, the administration was by intradermal injection at the base of the tail. The booster was administered after 18 days, consisting of equal parts of Freund's incomplete adjuvant and 2 mg/ml type II collagen in 10 mM acetic acid. Clinical evaluation was performed, the degree of swelling and erythema of the paws, using a standardized method of arthritis scoring, in which 0 = normal, 1 = mild swelling and erythema, 2 = moderate swelling and erythema, 3 = severe swelling and erythema plus loss of function in 2 paws, and 4 = total loss of function in a minimum of 3 paws. The onset of the disease, characterized by the development of erythema and/or swelling of the paw. After ten days, the animals were sacrificed by cervical dislocation. The synovial tissue was removed from ankle, knee and elbow joints and processed with a 1 mg/ml collagenase solution I (Merck, Saint Louis, USA) for one hour. After centrifugation, the pellet was resuspended in complete medium and cultured in the same way described for human samples until the cells were frozen. FLSs were used for experiments after five passages. All procedures using mice were in accordance with the National Institutes of Health Guide for the Care and Use of Animals. The Animal Use Ethics Committee (Protocol 293/2018) approved the animal study. Cells were labeled with antibodies for 20 min at 4 °C in the dark, washed, resuspended in 1 × PBS and analyzed on a FACSCanto II flow cytometer (BD Biosciences). A minimum of 100,000 events were acquired for each sample, and the acquisitions were processed using Diva software (BD Biosciences). FlowJo software (Tree Star) was used to analyze the data.

Electron microscopy
Morphological analysis was performed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). For SEM, RA-FLSs and CIA-FLSs were cultured without stimulation on polyl-lysine-treated coverslips. After cell fixation, the media was aspirated, and the cells were washed with 1 × saline phosphate buffer (SPB). The cultures were then immersed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer. After 120 min, the fixation solution was aspirated and replaced with 0.1 M cacodylate buffer. The samples were washed in aldehyde three times for 10 min each in 0.05 M cacodylate buffer and immersed in 1% osmium tetraoxide (OsO4) solution in 0.05 M cacodylate buffer for one hour at room temperature. The samples were washed in distilled water and dehydrated with increasing concentrations of alcohol (30,50,70, 90 and 100%) and acetone for 10 min each. The dehydrated samples were dried until the acetone was completely removed. The samples were visualized by SEM (FEG-Quanta 200 FEI). Cell shape, the emission of cytoplasmic projections, and the presence of filopodia and lamellipodia were analyzed in 20 RA-FLSs and 20 CIA-FLSs.
For TEM analysis of unstimulated RA-FLS and CIA-FLS, the preparation included aspiration of DMEM followed by washes with SPB. The cultures were then fixed for 120 min with 2.5% glutaraldehyde in 0.1 M cacodylate buffer at 4 °C. The fixed samples were scraped off with a spatula, transferred to sterile 50 ml tubes and centrifuged at 3000 rpm for 5 min. The pellets were fixed in 2% osmium tetroxide (Sigma-Aldrich, Sao Paulo, Brazil) in 0.1 M cacodylate buffer for 2 h at room temperature. After being washed three times with distilled water for 15 min each, the samples were incubated with 2% uranyl acetate (EMS) for 24 h at room temperature in the dark. The samples were then washed with distilled water as described above and dehydrated in an increasing series of alcohol (35,50,70,85,95, and 100%) and acetone for 20 min each. After dehydration, infiltration was followed by incubation in a 1:2, 1:1 and 1:2 mixture of acetone-Epon resin (EMBed Resin 812, EMS) and pure resin for 3 h. Subsequently, the cells were incubated in the same resin in BEEM capsules (Ted Pella, California, USA) and polymerized at 60 °C for 48 h. After polymerization of the resin, 300 nm semifine sections were obtained from the surface of the blocks with the aid of glass razors and stained with toluidine-sodium borate blue. Ultrathin sections (~ 60 nm) were obtained on a Leica EM UC6 ultramicrotome (Leica Microsystems GmbH, Wetzlar, Germany) using a diamond razor. The sections were mounted on 200 mesh copper (Ted Pella), stained with lead citrate (Merck KGaA, Darmstadt, Germany) for 5 min and analyzed under a transmission electron microscope (Tecnai G2-12 Spirit Biotwin Thermo Fischer Scientific/FEI 2006, Eindhoven, the Netherlands) at 120 kV. Cells were visualized by MET (Tecnai G2-12 Spirit Biotwin FEI -120 kV).
To investigate the morphological characteristics of synoviocytes, images of 20 cells per group revealing the entire cellular profile, including the nucleus, and subcellular details were obtained at 4200-fold, 16,500-fold and 20,500-fold magnifications. The basic ultrastructural analyses focused on organelles involved with the classic cell synthesis and secretion pathway: the nucleus (N), rough endoplasmic reticulum (RER), Golgi apparatus (GA) and mitochondria (MT). The morphometric analyses were focused on organelles involved in cellular pinocytosis (pinocytosis (P) and pinocytic vesicles (PV)), lamellar bodies (LBs) and MT. The morphometric analysis was based on the relative quantity of each organelle per number of cells.

Statistical analysis
The Mann-Whitney U test nonparametric analysis was used to compare MMP-3 and IL-6 concentrations between the TNF-α-and IL-1β-stimulated groups and to compare the amounts of pinocytosis, pinocytic vesicles, mitochondria, and lamellar bodies between RA-FLSs and CIA-FLSs. GraphPad Prisma version 6.01 was used for data analysis and graph production. We considered evidence of significant effects at p-values < 0.05.

Scanning electron microscopy
RA-FLS and CIA-FLS cultures exhibited populations of fusiform FLSs with cytoplasmic expansions forming branches. Both cell types were anchored to the substrate with numerous cytoplasmic projections at the lamellipodia ends, and the protein structure of the actin cytoskeleton protruded at the mobile end of the cell [23]. Lamellipodia propelled entire cell structures through the substrate (Fig. 2).
The cytoplasmic projections that extended beyond the lamellipodia border on migrating cells were often anchored to neighboring cells (Fig. 2).

Transmission electron microscopy
Evaluation of RA-FLSs and CIA-FLSs by Transmission Electron Microscopy (TEM) revealed subcellular characteristics of intense cellular activity, represented by a large and euchromatic nucleus with a thin layer of heterochromatin, a prominent and well-developed rough endoplasmic reticulum and Golgi apparatus, and the presence of several mitochondria close to these organelles (Fig. 3). As shown in Fig. 4, we observed a high concentration of intermediate filaments, which are filamentous protein structures that make up the cytoskeleton of cells and assist in cell morphology. Ultrastructural differences were observed between RA-FLSs and CIA-FLSs in the degree of pinocytosis (number of plasma membrane invaginations) and the number of pinocytotic vesicles, which were related to the process of cellular pinocytosis. Both structures were more frequently seen in RA-FLSs than in CIA-FLSs (p < 0.05) (Fig. 4). No significant difference was observed in the number of mitochondria (data not shown).
Lamellar bodies were observed in RA-FLS and CIA-FLS cultures (Fig. 3). Lamellar bodies consist of cavitary structures with circular walls in unique juxtaposition to FLSs in the synovial environment; these structures are secreted by exocytosis in the synovial fluid and secrete hyaluronic acid and surfactant (proteins and lipids) [24]. Lamellar bodies at various stages of maturation were identified in the same cell (Fig. 3C). No significant difference was observed in the number of lamellar bodies between RA-FLSs and CIA-FLSs (data not shown). Adjacent FLSs were observed to be in close contact through their cytoplasmic projections, corresponding to the electron-dense regions at the communication/adherent junction sites. Figure 5 shows the activation of MAPK and p-NF-κB signaling in RA-FLSs and CIA-FLSs after stimulation with TNF-α and IL-1β. In RA-FLSs and CIA-FLSs, TNF-α showed a trend to activate p-ERK1/2 (2.25-fold and 2.44-fold, respectively), p-NF-κB (3.91-fold and 1.18fold, respectively) and p-P38 (1.48-fold and 4.89-fold, respectively), and TNF-α significantly activated p-JNK (8.53-fold) only in RA-FLS. On the other hand, IL-1β significantly activated p-P38, p-NFκB, p-ERK1/2 and p-JNK (56.06-fold, 35.96-fold, 11.96-fold and 142.31fold, respectively) in RA-FLSs. In CIA-FLS, IL-1β did not show considerable effect (Fig. 5).

Discussion
Many in vitro studies use cells from experimental models to study the etiopathogenesis of RA [25]. FLSs from experimental models are frequently used in the search of therapeutic targets and to study the effects of new drugs, and it is well established that FLSs play important roles in cartilage and bone destruction in the joints of patients with RA [2]. Previous study has shown that histopathological and pathological characteristics in the inflamed tissues of mice with CIA and patients with RA are similar [26]. However, the present study showed ultrastructural differences, differences in the production of important mediators of inflammation in RA, and different signaling pathways were activated between human-and mousederived FLSs in vitro. Our group showed previously the similarity of FLS derived from synovial tissue and FLS derived from synovial fluid, which explains the use of RA-FLS from synovial fluid and CIA-FLS from tissue fluid in this study [27]. Furthermore, it was a limitation of the study has only FLS from experimental models in an established phase of the disease, however, this places these cells at the same time as RA-FLS, in which the disease is already established and not in the early stage.
In accordance with our data, we found important morphological differences between RA-FLS and CIA-FLS. Our data demonstrate for the first time a higher degree of pinocytosis and more pinocytotic vesicles in human cells than in mouse cells. Pinocytosis, a type of endocytosis, plays a critical role in cell transport, endocytosis, signal transduction, and cell proliferation [28][29][30]. When endocytosis occurs, the cell extends and folds around the extracellular material, forming a pocket and then creating pinocytic vesicles that are absorbed [31,32]. Therefore, these morphological differences could indicate increased functional activity in RA-FLSs.
Despite these differences, many similarities between RA-FLS and CIA-FLS morphology were observed, including the presence of lamellar bodies, fusiform shape, the presence of lamellipodia, large euchromatic nucleus, prominent and well-developed rough endoplasmic reticulum and Golgi apparatus, and the presence of several mitochondria nearby to these organelles [24].
Upregulation of IL-6 and MMP-3 after in vitro stimulation with TNF-α and IL-1β in RA-FLS has been previously demonstrated [20,33,34]. This stimulation was mediated by activation of the MAPK pathways (ERK, p38 and JNK), as well as the transcription factor NF-κB, and was supported by the findings of other studies [5,14,15,[35][36][37][38][39][40][41][42]. Unlike RA-FLSs, in CIA-FLS, after IL-1β stimulation, there was a lack of activation of the MAPK pathways and NF-κB and, consequently, the absence of an increase in the production of IL-6 and MMP-3. Although IL-1β is presented on the inflamed synovium in CIA [43,44], few studies have demonstrated the expression of the IL-1β receptor (IL1r1) in CIA-FLSs [42]. The low expression of the IL-1β receptor in CIA-FLSs could explain the lack of activation of downstream signaling pathways and the lack of an increase in MMP-3 and IL-6 production [45]. However, IL-6 production has been observed after IL-1β stimulation in CIA-FLSs [45], but the researchers did not analyze the activation of the signaling pathways that was reported in our work. Our results suggest that the low activation and nonactivation of signaling pathways (p-ERK, p-p38, p-JNK and NF-κB transcription factor) was consistent with the lack of production of MMP-3 and IL-6 [46]. Furthermore, TNF-α plays a major role in the pathogenesis of RA, as well as in CIA [2,44,47,48]. Both groups of cells responded to TNF-α stimulation by activation of the p-ERK, p-p38 and p-NF-κB signaling pathways. These activations mediate an increase in IL-6 production, but only JNK was not phosphorylated in CIA-FLSs. Despite studies showing the activation of this pathway in CIA-FLSs [43], we demonstrated that MMP-3 production was insignificant, which was consistent with the lack of activation of p-JNK [38,41]. It is possible that MMP-3 does not play as important a role in CIA as it does in RA patients [6,36,49]. Therefore, further studies will need to show the differences and similarities between RA-FLS and CIA-FLS.

Conclusion
Experimental models, play an important role in the study of disease mechanisms and the search for possible treatments. It is hard to draw conclusions about the significance of differences between mouse and human biology, however, despite the benefits these models bring, they must be interpreted with care since they do not completely resemble human diseases. Regarding the ultrastructural morphology and important inflammatory cytokines involved in the pathophysiology of RA responses, our study shows where some differences and similarities occur between RA-FLS and CIA-FLS, this could help in future studies using in vitro FLS, and also demonstrate that a well interpretation must be applied in using CIA-FLSs to design in vitro studies related to etiopathogenesis and new therapeutic targets in RA.