Article Text
PDF
Abstract
Objectives Rheumatoid arthritis (RA) is an autoimmune disease characterised by aggressive fibroblast-like synoviocytes (FLSs). Very few RA patients-derived FLSs (RA-FLSs)-specific surface signatures have been identified, and there is currently no approved targeted therapy for RA-FLSs. This study aimed to screen therapeutic aptamers with cell-targeting and cytotoxic properties against RA-FLSs and to uncover the molecular targets and mechanism of action of the screened aptamers.
Methods We developed a cell-specific and cytotoxic systematic evolution of ligands by exponential enrichment (CSCT-SELEX) method to screen the therapeutic aptamers without prior knowledge of the surface signatures of RA-FLSs. The molecular targets and mechanisms of action of the screened aptamers were determined by pull-down assays and RNA sequencing. The therapeutic efficacy of the screened aptamers was examined in arthritic mouse models.
Results We obtained an aptamer SAPT8 that selectively recognised and killed RA-FLSs. The molecular target of SAPT8 was nucleolin (NCL), a shuttling protein overexpressed on the surface and involved in the tumor-like transformation of RA-FLSs. Mechanistically, SAPT8 interacted with the surface NCL and was internalised to achieve lysosomal degradation of NCL, leading to the upregulation of proapoptotic p53 and downregulation of antiapoptotic B-cell lymphoma 2 (Bcl-2) in RA-FLSs. When administrated systemically to arthritic mice, SAPT8 accumulated in the inflamed FLSs of joints. SAPT8 monotherapy or its combination with tumour necrosis factor (TNF)-targeted biologics was shown to relieve arthritis in mouse models.
Conclusions CSCT-SELEX could be a promising strategy for developing cell-targeting and cytotoxic aptamers. SAPT8 aptamer selectively ablates RA-FLSs via modulating NCL-p53/Bcl-2 signalling, representing a potential alternative or complementary therapy for RA.
- Rheumatoid Arthritis
- Fibroblasts
- Inflammation
- Therapeutics
- Arthritis, Experimental
Data availability statement
Data are available in a public, open access repository. Data are available on reasonable request. Data are available in a public, open access repository. The RNA sequencing data have been uploaded to Gene Expression Omnibus platform with the accession number GSE268214.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Fibroblast-like synoviocytes (FLSs)-directed therapies have long been suggested as an alternative or complementary approach to the current immunosuppressive agents for rheumatoid arthritis (RA).
Cadherin-11 (CDH11) has been identified as a surface target for developing selective anti-FLS therapy, while antibodies against CDH11 fail to display therapeutic effects for patients with RA.
A primary obstacle is that molecular events driving pathological transformation of FLSs in RA are poorly understood, and few FLSs-specific surface targets have been identified.
WHAT THIS STUDY ADDS
This study presents a cell-specific and cytotoxic systematic evolution of ligands by exponential enrichment (CSCT-SELEX) technology to screen aptamers with cell-targeting and cytotoxic properties without prior knowledge of cell surface signatures.
The SAPT8 aptamer generated by CSCT-SELEX selectively ablates inflamed FLSs and relieves arthritis symptoms in mouse models after monotherapy or combined therapy with an anti-tumour necrosis factor (TNF) agent.
Nucleolin (NCL) is identified as the specific target of SAPT8 on the surface of inflamed FLSs, and SAPT8 induces lysosomal degradation of NCL to modulate p53/B-cell lymphoma 2 signalling.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
CSCT-SELEX is a promising technology for identifying therapeutic aptamers and surface molecular targets for disease cells.
The SAPT8 aptamer targeting RA patients-derived FLSs (RA-FLSs) represents an alternative or complementary therapy to the current immunosuppressive agents for RA.
NCL offers potential as a molecular target on the surface of RA-FLSs for developing FLSs-directed therapies for RA.
Introduction
Rheumatoid arthritis (RA) is a systemic, autoimmune, peripheral polyarthritis that primarily involves the small joints of feet and hands.1 It is characterised by chronic inflammation, synovial hyperplasia, pannus formation and destruction of bone and cartilage.2 Currently, the aetiology of RA remains unclear, and numerous genetic and environmental factors have been reported to be associated with an increased risk of RA.3 Pathophysiology of RA involves the overactivation of immune cells such as T cells, B cells and macrophages and the release of large numbers of inflammatory mediators, such as interleukin 6 (IL-6), IL-1β, tumour necrosis factor-α (TNF-α) and matrix metalloproteinases (MMPs).4–8 Significant progress in RA management has been made in the last decades, since the introduction of diverse synthetic and biologic disease-modifying antirheumatic drugs (DMARDs).9 Despite this, still a considerable proportion of patients with RA do not respond adequately to the currently available therapies,10 and there is an urgent need for new treatment options for RA.
The commonly used DMARDs are immunosuppressive agents that prevent the excessive activation of immune response in RA.11 However, it is now widely considered that an immunosuppressive strategy is not sufficient to completely halt RA progression because non-immune cells have also been suggested to play critical roles in RA development besides the innate and adaptive immune systems, particularly the fibroblast-like synoviocytes (FLSs).10 FLSs, also known as synovial fibroblasts, are highly specialised mesenchymal cells that populate the intimal lining of synovium.12 The normal function of FLSs is to maintain the homeostasis of the diarthrodial joints by producing extracellular matrix components and synovial fluid, thereby lubricating and nourishing cartilage surfaces.11 In RA, however, FLSs are activated to undergo phenotypic transformation into tumour-like cells13 and produce many inflammatory mediators, leading to growing synovial pannus that invade adjacent cartilage and bone.14 Thus, RA patients-derived FLSs (RA-FLSs)-directed therapies have long been suggested as a potentially alternative or complementary approach to the current DMARDs-based immune-directed therapies in RA.12–14
Depressingly, no clinically available therapy selectively ablates RA-FLSs until now.10 So far, only cadherin-11 (CDH11), a major cell surface adhesion molecule that acts as a critical regulator of synovial lining formation,15 has been extensively tested as a potential target for developing selective anti-RA-FLSs therapy,16 even though a monoclonal antibody against CDH11 fails to display sufficient efficacy in treating RA in a clinical trial.10 Recently, another emerging RA-FLSs-specific cell surface molecular target is receptor tyrosine phosphatase sigma (PTPRS), while efforts to target it remain in their infancy.16 17 A primary obstacle to developing selective anti-RA-FLSs therapies is that molecular events that drive the tumour-like transformation of RA-FLSs are poorly understood, and very few RA-FLSs-specific cell surface molecular signatures have been identified.18 We wonder whether it is possible to design a high-throughput screening strategy to develop selective anti-RA-FLS therapies without prior knowledge of RA-FLSs-specific cell surface molecular signatures.
Keeping the above idea in mind, we think of aptamer technology.19 Aptamers are single-stranded DNA (ssDNA) or RNA, which can fold into unique three-dimensional structures and specifically recognise a variety of targets, including organic molecules, peptides, proteins and even whole cells.20–22 They can be screened by an in vitro high-throughput procedure called the systematic evolution of ligands by exponential enrichment (SELEX).23 Besides being widely employed as navigation tools in nanotechnology,24 aptamers are emerging as a new class of nucleic acid drugs with affinity and specificity, rivalling that of antibodies.20 When carried out against whole cells, the SELEX is specified as cell-SELEX,25 a counterselection strategy to screen aptamers that only interact with target cells but not control cells.20 The unique advantage of cell-SELEX is that the process can generate aptamers without any prior knowledge of specific molecular signatures expressed on the cell surface of target cells and also enables the identification of previously unknown cell surface molecular signatures.26 This advantage of cell-SELEX seems to fit in with our idea of developing selective anti-RA-FLS therapies without prior knowledge of RA-FLSs-specific cell surface molecular signatures. However, in most cases, cell-SELEX only generates cell-targeting aptamers,27 but barely produces therapeutic aptamers exhibiting functional impacts on target cells.28 Thus, cell-SELEX could not guarantee the identification of desired aptamers that selectively ablate RA-FLSs.
In our study, we develop a cell-specific and cytotoxic SELEX (CSCT-SELEX) strategy to screen aptamers with both cell-targeting and cytotoxic properties without any prior knowledge of molecular signatures specifically expressed on the cell surface of target cells. By CSCT-SELEX, we obtained two aptamers (SAPT4 and SAPT8) that display specific binding with RA-FLSs and proapoptotic potential against RA-FLSs. We further show that SAPT4 and SAPT8 also specifically recognise and kill a human synovial cell line SW982 and FLSs from arthritic rat and mouse models in vitro. Mechanistically, we failed to identify the target of SAPT4, but we discovered that the molecular target of SAPT8 is nucleolin (NCL), which is specifically expressed on the cell surface of RA-FLSs. NCL has been reported to act as an RNA-binding protein that controls gene expression by regulating either mRNA stability or translation.27 Interestingly, NCL has been shown to shuttle from the nucleus to the cell membrane to facilitate resistance to apoptosis and the growth of tumours.29 However, NCL is rarely studied in RA. Here, we reveal the pathogenic role of NCL in the tumour-like transformation of RA-FLSs. We find that SAPT8 recognises and ablates RA-FLSs via targeting and degrading NCL in a lysosome-dependent manner. SAPT8 modulates the expression of NCL target genes, including proapoptotic p53 and antiapoptotic B-cell lymphoma 2 (Bcl-2),30 leading to inhibited tumour-like phenotypes of RA-FLSs. Our in vivo study shows the obvious accumulation of SAPT8 in FLSs of arthritic mice and the desired therapeutic effects of SAPT8 monotherapy or SAPT8 in combination therapy with a TNF-targeted biological DMARD for treating arthritis.
In summary, this study provides a new type of SELEX, that is, CSCT-SELEX, which would be a promising strategy for screening aptamers possessing both cell-targeting and cytotoxic properties without prior knowledge of cell surface molecular signatures. By CSCT-SELEX, we obtain an aptamer SAPT8 that selectively ablates RA-FLSs via modulating NCL-p53/Bcl-2 signalling, representing a potential alternative or complementary therapy to the current DMARDs for RA.
Results
Development of CSCT-SELEX to generate RA-FLSs-specific and cytotoxic aptamers
We developed a new SELEX strategy, called cell-specific and cytotoxic SELEX (CSCT-SELEX), to screen cytotoxic aptamers specifically targeting RA-FLSs. The CSCT-SELEX was a tandem process consisting of a cell-specific SELEX (CS-SELEX) module followed by a cytotoxic SELEX (CT-SELEX) module (figure 1A). Briefly, CS-SELEX was performed using a random ssDNA library and included multiple rounds of positive selection and negative selection, which was similar with that of the traditional cell-SELEX.31 32 RA-FLSs, characterised by high expression of fibroblastic biomarkers THY1 and vimentin33 34 as well as absence of a macrophage marker CD6835 (online supplemental figure S1A,B), were employed as target cells in the positive selection (figure 1A), and human monocytic THP-1 cells and osteoblastic hFOB 1.19 cells were used as control cells in the negative selection (figure 1A). Via the CS-SELEX, ssDNA sequences specifically recognising RA-FLSs could be enriched, and interfering sequences bound with the common cell surface molecular signatures between target cells and control cells could be removed (figure 1A). Then, the CT-SELEX was initiated using the ssDNA enriched pool that exhibited the strongest binding ability with RA-FLSs in the CS-SELEX module (figure 1A). The CT-SELEX module was an iterative process via treating RA-FLSs with the enriched ssDNA pools and collecting apoptotic cells rather than live cells for isolating proapoptotic aptamers (figure 1A).
Supplemental material
Generation of rheumatoid arthritis patients-derived fibroblast-like synoviocytes (RA-FLSs)-specific and cytotoxic aptamers by cell-specific and cytotoxic systematic evolution of ligands by exponential enrichment (CSCT-SELEX). (A) Workflow of the CSCT-SELEX. Briefly, the CSCT-SELEX was a tandem process consisting of a cell-specific SELEX (CS-SELEX) module followed by a cytotoxic SELEX (CT-SELEX) module. The CS-SELEX module included multiple rounds of positive selection against target cells (RA-FLSs) and negative selection against control cells (a human monocytic cell line THP-1 and a human osteoblastic cell line hFOB 1.19) using an initial random single-stranded DNA (ssDNA) library. After the CS-SELEX, the CT-SELEX module was conducted using the enriched ssDNA pool that exhibited the strongest binding ability with RA-FLSs in the CS-SELEX module. The CT-SELEX module was an iterative process via treating RA-FLSs with the enriched ssDNA pools and collecting apoptotic cells rather than live cells for isolating proapoptotic aptamers. (B) Mean fluorescence intensity (MFI) of RA-FLSs, THP-1 and hFOB 1.19 cells after incubation with the enriched ssDNA pools from the CS-SELEX module, as determined by a microplate reader. The cells were incubated with 200 nM Cy3-labelled ssDNA library (Lib) or the enriched ssDNA pools from the 4th round (R4), the 8th round (R8) and the 12th round (R12) of CS-SELEX for 3 hours. (C) Percentage of apoptotic cells of RA-FLSs after treatment with the enriched ssDNA pools from the CT-SELEX module, as determined by the annexin V-coated magnetic beads. The RA-FLSs were daily treated with 400 nM ssDNA enrichment pools from the first round (R1′), the third round (R3′), the sixth round (R6′) and the ninth round (R9′) of CT-SELEX for 3 days. (D) Binding of RA-FLSs with 200 nM Cy3-labelled negative control (NC) or the most enriched eight aptamer candidates (SAPT1-8) from CSCT-SELEX after incubation for 3 hours, as determined by flow cytometry. (E) Apoptosis of RA-FLSs after daily treatment with phosphate buffered saline (PBS), 400 nM NC or SAPT1-8 for 4 days, as determined by flow cytometry. (F) Binding of RA-FLSs with 200 nM Cy3-labelled SAPT4 or SAPT8 at different time points (0, 1, 3 and 6 hours). (G) Binding of RA-FLSs with Cy3-labelled SAPT4 or SAPT8 at different concentrations (0, 50, 100, 200 or 400 nM) for 3 hours. (H) Binding curves of SAPT4 or SAPT8 with RA-FLSs. The cells were incubated with increasing concentrations (conc.) of Cy3-labelled SAPT4 or SAPT8 for 3 hours. The equilibrium dissociation constants (Kd) were calculated by nonlinear fitting analysis. Data from three (B–G) or six (H) independent biological replicates were represented as mean±standard deviation (SD) and statistical significance was calculated by the one-way analysis of variance with a post hoc test. *p<0.05, ***p<0.001, ****p<0.0001. (I) Secondary structures of SAPT4 and SAPT8 predicted by the RNAstructure software.
During CS-SELEX, we monitored the binding ability of the enriched ssDNA pools with target cells (RA-FLSs) and control cells (THP-1 and hFOB 1.19 cells). After incubation with Cy3-labelled enriched pools from increasing rounds of CS-SELEX, the mean fluorescence intensity (MFI) of RA-FLSs was gradually enhanced, whereas there was no obvious improvement in MFI of THP-1 and hFOB 1.19 cells (figure 1B). We also examined the proapoptotic capacity of the enriched pools from the CT-SELEX on RA-FLSs. Apoptotic cells of RA-FLSs were increased after incubation with the enriched pools from the increasing rounds of the CT-SELEX (figure 1C). After the whole CSCT-SELEX, we sequenced the enriched ssDNA and chose the most enriched eight aptamer candidates (SAPT1-8) for further characteristics (online supplemental table S1). Among the candidates, SAPT4 and SAPT8 not only showed good binding with RA-FLSs (figure 1D) but also had proapoptotic capacity against RA-FLSs (figure 1E and online supplemental figure S1C). Notably, the binding ability and proapoptotic capacity of SAPT8 against RA-FLSs were better than those of SAPT4 (figure 1D,E and online supplemental figure S1C). Both SAPT4 and SAPT8 displayed time-dependent and concentration-dependent binding with RA-FLSs (figure 1F,G). The equilibrium dissociation constants (Kd) of SAPT4 and SAPT8 with RA-FLSs were 101.7±29.6 and 71.2±15.0 nM, respectively (figure 1H). Neither SAPT4 nor SAPT8 could bind with THP-1 and hFOB 1.19 cells (online supplemental figure S1D). We observed no obvious binding between SAPT4 or SAPT8 with a human normal skin fibroblast cell line BJ (online supplemental figure S1D). We obtained minimally activated FLSs from individuals who suffered from an anterior cruciate ligament injury (ACLI), that is, ACLI-FLSs. SAPT8 only showed marginal binding with ACLI-FLSs (online supplemental figure S1D). Neither SAPT4 nor SAPT8 induced apoptosis of THP-1 cells, hFOB 1.19 cells, BJ cells and ACLI-FLSs (online supplemental figure S1E). Both SAPT4 and SAPT8 showed good serum stability for at least 24 hours (online supplemental figure S1F). The predicted secondary structures of SAPT4 and SAPT8 are shown in figure 1I.
SAPT4 and SAPT8 inhibit tumour-like phenotypes of RA-FLSs and SW982 in vitro
Besides apoptosis resistance, RA-FLSs also acquire other tumour-like phenotypes, such as hyperproliferation and increased invasion and migration.36 37 We examined whether SAPT4 and SAPT8 had inhibitory effects on these phenotypes of RA-FLSs in vitro. We incubated RA-FLSs with phosphate buffered saline (PBS), a negative control (NC) sequence, SAPT4 or SAPT8 and performed Cell Counting Kit-8 (CCK-8), colony formation, transwell and wound healing assays. CCK-8 and colony formation assays demonstrated that both SAPT4 and SAPT8 inhibited cell viability and proliferation of RA-FLSs when compared with PBS or NC (figure 2A,B). Transwell assays showed that both SAPT4 and SAPT8 reduced invasion and migration of RA-FLSs (figure 2C,D). Wound healing assay confirmed the inhibitory effects of both SAPT4 and SAPT8 on RA-FLSs migration (figure 2E). RA-FLSs are also activated to produce a range of inflammatory mediators, such as IL-1β and MMPs.38 When treated with SAPT4 or SAPT8, RA-FLSs had lower expression of IL-1β and MMP3 than the cells treated with NC or PBS (figure 2F). Notably, the inhibitory effects of SAPT8 on these phenotypes of RA-FLSs were superior to those of SAPT4 (figure 2A–F). We also tested the in vitro binding ability and therapeutic effects of SAPT4 and SAPT8 against a human synovial cell line SW982. We observed that both SAPT4 and SAPT8 bound with SW982 cells (online supplemental figure S2A), induced apoptosis and decreased viability, proliferation, invasion and migration of SW982 cells in vitro (online supplemental figure S2B–G). Consistently, SAPT8 also displayed better therapeutic effects on SW982 cells than SAPT4 (online supplemental figure S2A–G).
Effects of SAPT4 and SAPT8 on rheumatoid arthritis patients-derived fibroblast-like synoviocytes (RA-FLSs) in vitro. (A) Viability of RA-FLSs after daily treatment with phosphate buffered saline (PBS) or 400 nM negative control (NC), SAPT4 or SAPT8 for 5 days as determined by Cell Counting Kit-8 assays. (B) Colony formation of RA-FLSs after daily treatment with PBS or 400 nM NC, SAPT4 or SAPT8 for 13 days. (C) Invasion of RA-FLSs after daily treatment with PBS or 400 nM NC, SAPT4 or SAPT8 for 2 days, as determined by transwell assays. Scale bars=200 µm. (D) Migration of RA-FLSs after daily treatment with PBS or 400 nM NC, SAPT4 or SAPT8 for 2 days, as determined by transwell assays. Scale bars=200 µm. (E) Wound healing assays of RA-FLSs after treatment with PBS or 400 nM NC, SAPT4 or SAPT8 for 36 hours. Scale bars=200 µm. (F) Protein expression of interleukin (IL)-1β and MMP3 in RA-FLSs after treatment with PBS or 400 nM NC, SAPT4 or SAPT8 for 24 hours, as determined by western blotting. Data from six independent biological replicates were represented as mean±standard deviation (SD) and statistical significance was calculated by the two-way analysis of variance (ANOVA) (A) or the one-way ANOVA (B–E) with a post hoc test. *p<0.05, **p < 0.01, ***p<0.001, ****p<0.0001. OD, optical density.
SAPT4 and SAPT8 suppress aggressive phenotypes of FLSs from arthritic rats and mice in vitro
We established rat and mouse models with collagen-induced arthritis (CIA) and isolated CIA rat FLSs (CIA-RFLSs) and CIA mouse FLSs (CIA-MFLSs), respectively.39 40 CIA-RFLSs characterised by excessive expression of both THY1 and vimentin and absence of CD68 (online supplemental figure S3A) were incubated with PBS, NC, SAPT4 or SAPT8. SAPT4 or SAPT8 had strong binding ability with CIA-RFLSs (online supplemental figure S3B). Dramatic apoptosis was observed in CIA-RFLSs after treatment with SAPT4 or SAPT8 (online supplemental figure S3C). The viability, proliferation, invasion and migration of CIA-RFLSs were suppressed by SAPT4 or SAPT8 (online supplemental figure SD–H). There were better binding ability and therapeutic effects of SAPT8 than SAPT4 against CIA-RFLSs (online supplemental figure S3B–H). In addition, SAPT4 or SAPT8 also showed good binding ability with CIA-MFLSs that displayed positive expression of both THY1 and vimentin and negative expression of CD68 (online supplemental figure S4A,B), while they could not recognise non-CIA-MFLSs, including healthy mouse FLSs (H-MFLSs), a mouse osteoblast-like cell line MC3T3-E1, a mouse macrophage-like cell line RAW 264.7, a mouse fibroblastic cell line L929 and a mouse chondrogenic cell line ATDC5 (online supplemental figure S4B,C). No apoptosis was observed for H-MFLSs incubated with SAPT4 or SAPT8 (online supplemental figure S4D), whereas apoptosis of CIA-MFLSs was induced after treatment with SAPT4 or SAPT8 when compared with PBS or NC (online supplemental figure S4E). Cell viability, proliferation, invasion and migration of CIA-MFLSs were decreased by SAPT4 or SAPT8 (online supplemental figure S4F–J). SAPT8 had stronger binding ability with CIA-MFLSs and better inhibitory effects on tumour-like phenotypes of CIA-MFLSs (online supplemental figure S4B, E–J).
NCL is identified as the cell surface target of SAPT8 on RA-FLSs
To identify the cell surface targets of SAPT4 and SAPT8, we conducted a pull-down assay using streptavidin-coated beads after incubating membrane proteins of RA-FLSs with PBS or biotin-labelled NC, SAPT4 or SAPT8. The beads, along with the captured proteins, were eluted and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Silver staining demonstrated that SAPT8 specifically interacted with a protein that had a molecular weight between 100 and 130 kDa, whereas none of any specific protein was captured by SAPT4, when compared with NC or PBS (figure 3A). Next, we performed liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis to identify the SAPT8-captured protein. Among the 100 candidate proteins identified by LC-MS/MS (online supplemental table S2), NCL attracted our attention, as it had been reported to have a predicted molecular weight of 76.6 kDa and an observed molecular weight of 100–110 kDa,41 42 which were consistent with our silver staining and LC-MS/MS results. We further verified that NCL could be captured by SAPT8 in a pull-down assay followed by western blotting (figure 3B). We performed a dot blotting assay and showed a direct interaction between SAPT8 and recombinant human NCL (rhNCL) (figure 3C). Previously, no study revealed the presence of NCL on the surface of RA-FLSs. We isolated membrane proteins of hFOB 1.19 cells and RA-FLSs and observed the high level of NCL on surface of RA-FLSs rather than on surface of hFOB 1.19 cells (figure 3D). We performed gene silencing of NCL in RA-FLSs and found that there was decreased binding of SAPT8 with NCL-silenced RA-FLSs (figure 3E), suggesting that NCL was the cell surface target of SAPT8 on RA-FLSs.
Identification of the target proteins of SAPT8. (A) Pull-down assays of SAPT4-interacting or SAPT8-interacting proteins. The membrane proteins of rheumatoid arthritis patients-derived fibroblast-like synoviocytes (RA-FLSs) were incubated with phosphate buffered saline (PBS) or 1 µM biotin-labelled negative control (NC), SAPT4 or SAPT8 for 6 hours before the addition of streptavidin-coated beads. The NC-interacting, SAPT4-interacting or SAPT8-interacting proteins were separated on sodium dodecyl sulfate-polyacrylamide gel electrophoresis, followed by silver staining. The arrow showed the protein band of interest, which was collected for liquid chromatography-tandem mass spectrometry analysis. (B) The interaction between SAPT8 and nucleolin (NCL) in RA-FLSs detected by pull-down assay followed by western blotting. (C) Analysis of direct interaction between SAPT8 and NCL by dot blotting. 20 ng recombinant human NCL (rhNCL) was immobilised on a nitrocellulose membrane and then incubated with 1 µM biotin-labelled NC or SAPT8 before the addition of horseradish peroxidase (HRP)-conjugated streptavidin. (D) Expression of NCL on the surface of RA-FLSs or hFOB 1.19 cells. (E) MFI of RA-FLSs transfected with siRNA targeting NCL (si-NCL) or negative control siRNA (si-NC) after incubation with 200 nM NC or SAPT8 for 3 hour. (F) Relative mRNA level of NCL in synovial tissues from knee joints of individuals with anterior cruciate ligament injury (ACLI) (n=3) or patients with RA (n=4). (G) Immunofluorescence staining of NCL (green) on sections of the synovial tissues. Cell nuclei were counterstained with 4’,6-diamidino-2-phenylindole (DAPI) (blue). Scale bars=50 µm. (H) Proliferation of RA-FLSs after transfection with si-NC or si-NCL for 48 hours, as measured by 5-ethynyl-2′-deoxyuridine (EdU) staining. Scale bars=100 µm. (I) Apoptosis of RA-FLSs after transfection with si-NC or si-NCL, as measured by terminal deoxynucleotidyl transferase 2'-deoxyuridine 5'-triphosphate (dUTP) nick end labelling (TUNEL) staining. Scale bars=100 µm. (J) Invasion of RA-FLSs after transfection with si-NC or si-NCL. Scale bars=200 µm. (K) Migration of RA-FLSs after transfection with si-NC or si-NCL. Scale bars=200 µm. (L) Wound healing assays of RA-FLSs after transfection with si-NC or si-NCL. Scale bars=200 µm. (M) Relative mRNA levels of NCL, IL-1β and MMP3 in RA-FLSs after transfection with si-NC or si-NCL. (N) Protein levels of NCL, IL-1β and MMP3 in RA-FLSs after transfection with si-NC or si-NCL. (O) Volcano plot of differentially expressed genes between RA-FLSs infected with lentivirus vector expressing shRNA targeting NCL (sh-NCL) and RA-FLSs infected with lentivirus vector expressing negative control shRNA (sh-NC), as determined by RNA sequencing. Red: upregulated genes, blue: downregulated genes, grey: non-differentially expressed genes. (P) Kyoto Encyclopedia of Genes and Genomes pathway analysis of the downregulated genes in NCL-silenced RA-FLSs. (Q) Gene ontology (GO) enrichment analysis of molecular function (MF), cellular components (CC) and biological processes (BP) for the downregulated genes in NCL-silenced RA-FLSs. (R) Gene set enrichment analysis of inflammation and survival-related signalling in NCL-silenced RA-FLSs. Data from three independent biological replicates were represented as mean±standard deviation (SD) and statistical significance was calculated by the one-way analysis of variance with a post hoc test (E) or the two-tailed Student’s t-test (F–M). **p<0.01, ***p<0.001, ****p<0.0001.
NCL plays a pathological role in the tumour-like transformation of RA-FLSs
Until now, NCL has been found to be a multifunctional shuttling protein present in the nucleus, cytoplasm and on the surface of tumour cells.43 The cell surface NCL contributes to resistance to apoptosis and the growth of tumour cells.43 To date, no study demonstrates the pathological role of NCL in RA-FLSs. We examined the expression of NCL in synovial tissues of patients with RA and synovial tissues of individuals with ACLI. There was higher mRNA and protein expression of NCL in synovial tissues of patients with RA when compared with those of individuals with ACLI (figure 3F,G). We manipulated NCL expression in RA-FLSs in vitro. The siRNA targeting NCL (si-NCL) significantly inhibited proliferation and promoted apoptosis of RA-FLSs when compared with the NC siRNA (si-NC), as determined by 5-ethynyl-2′-deoxyuridine (EdU) staining and terminal deoxynucleotidyl transferase 2'-deoxyuridine 5'-triphosphate (dUTP) nick end labelling (TUNEL) staining, respectively (figure 3H,I). Invasion and migration of RA-FLSs were decreased by the si-NCL rather than the si-NC (figure 3L). The si-NCL reduced the mRNA and protein expression of IL-1β and MMP3 (figure 3M,N). We constructed lentiviral vectors expressing short hairpin RNA targeting NCL (sh-NCL) or NC shRNA (sh-NC) and performed RNA sequencing in RA-FLSs infected with lentiviral particles. The volcano plot showed significant downregulation of NCL in RA-FLSs treated with sh-NCL and numerous differentially expressed genes (DEGs) between NCL-silenced RA-FLSs and RA-FLSs transduced with sh-NC (figure 3O). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis revealed that sh-NCL downregulated a series of immune-associated pathways in RA-FLSs, such as IL-17, TNF, nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) and janus kinase-signal transducers and activators of transcription (JAK-STAT) (figure 3P). Gene Ontology (GO) enrichment analysis demonstrated that biological events including cell proliferation, migration and inflammatory response were suppressed in RA-FLSs treated with sh-NCL (figure 3Q). Gene set enrichment analysis (GSEA) confirmed that the downregulated genes in RA-FLSs treated with sh-NCL were mainly involved in inflammation and survival of RA-FLSs (figure 3R). We also overexpressed NCL in RA-FLSs. The overexpression of NCL enhanced proliferation and suppressed apoptosis of RA-FLSs (online supplemental figure S5A,B). Invasion and migration of RA-FLSs and expression of IL-1β and MMP3 in RA-FLSs were boosted after overexpression of NCL (online supplemental figure S5C–G). These results suggested that NCL played a critical role in the activation and tumour-like transformation of RA-FLSs.
SAPT8 interacts with NCL and modulates NCL-p53/Bcl-2 signalling in RA-FLSs
To explore how SAPT8 inhibits activation and tumour-like transformation of RA-FLSs via targeting NCL, we examined the mRNA and protein levels of NCL in RA-FLSs after treatment with PBS, NC or SAPT8. Compared with PBS or NC, SAPT8 had no effect on mRNA expression of NCL (figure 4A), but SAPT8 significantly decreased the protein level of NCL (figure 4B). Bcl-2 and p53 are the two most studied target genes of NCL that are involved in resistance to apoptosis and growth of tumour cells.30 NCL interacts with the p53 mRNA 5′ untranslated region (5′-UTR) and prevents its translation, whereas NCL stabilises Bcl-2 mRNA by binding to the 3′-UTR and protecting it from nuclease degradation, thus allowing tumour cells to avoid apoptosis.27 30 We showed that SAPT8 increased the expression of p53 and reduced the level of Bcl-2 (figure 4B). We examined the subcellular distribution of SAPT8 in RA-FLSs and observed both the cell surface and intracellular localisation of SAPT8, suggesting that SAPT8 could enter RA-FLSs via targeting cell surface NCL (figure 4C). To determine the endocytic pathways of SAPT8, we incubated RA-FLSs with SAPT8 with or without the presence of an inhibitor of macropinocytosis (amiloride hydrochloride), an inhibitor of clathrin-mediated endocytosis (chlorpromazine (CPZ)) or an inhibitor of caveolae-dependent endocytosis (genistein (GNT)).44 Both CPZ and GNT decreased cellular uptake of SAPT8 (figure 4D), demonstrating that SAPT8 entered RA-FLSs via both clathrin-dependent and caveolae-dependent endocytic pathways. It is generally accepted that endocytosis via caveolae or clathrin directs the receptor-ligand complex to lysosomal and/or proteasomal degradation.45 46 We found that SAPT8 was colocalised with a lysosome tracker in RA-FLSs (figure 4E). We examined whether SAPT8 modulated the levels of NCL as well as its target genes Bcl-2 and p53 in a proteasome-dependent or lysosome-dependent manner. We treated the RA-FLSs with SAPT8 with or without the presence of a proteasome inhibitor MG132 or lysosomal inhibitors bafilomycin A1 (Baf A1) and chloroquine (CQ).47 MG132 had no effect on SAPT8-induced alteration of NCL, Bcl-2 and p53, whereas Baf A1/CQ reversed SAPT8-induced degradation of NCL, a decrease of Bcl-2 and an increase of p53 in RA-FLSs (figure 4F), suggesting that SAPT8 specifically induced lysosomal degradation of NCL and subsequently altered the expression of p53 and Bcl-2 in RA-FLSs.
Mechanism of action of SAPT8. (A) Relative mRNA level of nucleolin (NCL) in rheumatoid arthritis patients-derived fibroblast-like synoviocytes (RA-FLSs) after treatment with phosphate buffered saline (PBS) or 400 nM negative control (NC) or SAPT8 for 24 hours. (B) Protein expression of NCL, B-cell lymphoma 2 (Bcl-2) and p53 in RA-FLSs after treatment with PBS, 400 nM NC or SAPT8 for 24 hours. (C) Subcellular distribution of SAPT8 in RA-FLSs. RA-FLSs were treated with 200 nM Cy3-labelled NC or SAPT8 for 6 hours. (D) Binding of SAPT8 with RA-FLSs with or without the presence of different inhibitors of endocytic pathways. RA-FLSs were incubated with amiloride hydrochloride (ALH, an inhibitor of macropinocytosis), chlorpromazine (CPZ, an inhibitor of clathrin-mediated endocytosis) or genistein (GNT, an inhibitor of caveolae-mediated endocytosis) for 1 hour and then treated with PBS or 200 nM Cy3-labelled NC or SAPT8 for 3 hours. (E) Colocalisation of SAPT8 with the lysosome in RA-FLSs. RA-FLSs were incubated with 200 nM Cy3-labelled NC or SAPT8 (red) for 6 hours before the addition of a lysosome tracker (green). Scale bars=100 µm. (F) Protein expression of NCL, Bcl-2 and p53 in RA-FLSs after treatment with 400 nM NC or SAPT8 for 24 hours, with or without the presence of a proteasome inhibitor MG132 or lysosomal inhibitors chloroquine and bafilomycin A1 (CQ/Baf A1). (G) Binding of SATP8 with RA-FLSs with or without the presence of AS1411. RA-FLSs were incubated with 200 nM Cy3-labelled NC, SAPT8, AS1411 or SAPT8 with the presence of 800 nM AS1411 for 3 hours. (H) Protein expression of NCL, Bcl-2 and p53 in RA-FLSs after treatment with 400 nM NC or SAPT8 for 24 hours, with or without the presence of 800 nM AS1411. (I) Colocalisation of SAPT8 with lysosome in RA-FLSs with or without the presence of 800 nM AS1411. (J) Proliferation of RA-FLSs after treatment with 400 nM NC or SAPT8 for 4 days, with or without the presence of 800 nM AS1411, as measured by 5-ethynyl-2′-deoxyuridine (EdU) staining. Scale bars=100 µm. (K) Apoptosis of RA-FLSs after treatment with 400 nM NC or SAPT8 for 4 days, with or without the presence of 800 nM AS1411, as measured by terminal deoxynucleotidyl transferase 2'-deoxyuridine 5'-triphosphate (dUTP) nick end labelling (TUNEL) staining. Scale bars=100 µm. (L) Invasion of RA-FLSs after treatment with 400 nM NC or SAPT8 for 48 hours, with or without the presence of 800 nM AS1411. Scale bars=200 µm. (M) Migration of RA-FLSs after treatment with 400 nM NC or SAPT8 for 48 hours, with or without the presence of 800 nM AS1411. Scale bars=200 µm. Fluorescence images were captured by a confocal microscope. Data from three independent biological replicates were represented as mean±standard deviation (SD) and statistical significance was calculated by the one-way analysis of variance with a post hoc test (A and J–M) or the two-tailed Student’s t-test (E,I). ns, no significance, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.
To make sure that SAPT8-induced lysosomal degradation of NCL and alteration of p53/Bcl-2 signalling was the dominant mechanism of action for its inhibitory effects on RA-FLSs, we performed NCL gene silencing in RA-FLSs and then incubated these cells with NC and SAPT8. SAPT8 could not induce apoptosis of the NCL-silenced RA-FLSs and was also unable to inhibit proliferation, invasion and migration of the RA-FLSs (online supplemental figure S6A–D), suggesting that SAPT8 exerted therapeutic effects on RA-FLSs in an NCL-dependent manner. Besides the SAPT8 identified in our study, AS1411 has already been well-known as an aptamer that specifically recognises NCL on the surface of tumour cells.48 49 We found that AS1411 also recognised RA-FLSs and competed with SAPT8 for binding with RA-FLSs (figure 4G). Unlike the role of SAPT8 in mediating the lysosomal degradation of NCL, AS1411 mainly inhibits activation (phosphorylation) of NCL rather than affects the expression of NCL.48 Moreover, it is worth noting that even though AS1411 at nanomolar concentrations could recognise NCL, micromolar concentrations of AS1411 are typically required to inhibit NCL phosphorylation and display therapeutic effects.50–52 We confirmed that the protein levels of NCL, p53 and Bcl-2 were not changed in RA-FLSs after treatment with nanomolar AS1411, while nanomolar AS1411 rescued SAPT8-induced degradation of NCL and alteration of Bcl-2 and p53 (figure 4H). Cellular uptake and lysosomal localisation of SAPT8 were also decreased after the addition of nanomolar AS1411 (figure 4I). Consistently, we also demonstrated that nanomolar AS1411 had no effect on tumour-like phenotypes of RA-FLSs (figure 4J–M), whereas it reversed SAPT8-induced suppression of the tumour-like phenotypes of RA-FLSs (figure 4J–M). In addition, we examined whether AS1411 at micromolar concentrations could display therapeutic effects on RA-FLSs. Our results showed that AS1411 at a concentration of 5 µM reduced the level of phosphorylated NCL (online supplemental figure S7A) and inhibited the tumour-like phenotypes of RA-FLSs (online supplemental figure S7B–E). In addition to AS1411, iSN04 is another newly identified aptamer targeting NCL.42 53 We showed that iSN04 also could bind with RA-FLSs (online supplemental figure S8A). Like AS1411, iSN04 did not induce the degradation of NCL but decreased the phosphorylation of NCL and inhibited the phenotypes of RA-FLSs (online supplemental figure S8B–F). However, iSN04 was less potent than AS1411 as we demonstrated that iSN04 only exhibited therapeutic effects at a concentration of 10 µM (online supplemental figure S8B–F), which was much higher than the effective concentration of AS1411 (5 µM).
SAPT8 is selectively accumulated in FLSs of CIA mice in vivo
SAPT8, on the one hand, displayed stronger binding ability with RA-FLSs, SW982, CIA-MFLSs and CIA-RFLSs, as well as better inhibitory effects on the tumour-like transformation of these FLSs than SAPT4; on the other hand, the mechanism of action of SAPT8 was also clearly interpreted. We thus chose SAPT8 for in vivo studies. To examine the tissue and cellular accumulation of SAPT8, CIA mice were administrated with Cy3-labelled NC or SAPT8 by tail-vein injection, and the in vivo and ex vivo fluorescence distribution was evaluated. The whole-body biophotonic imaging demonstrated stronger fluorescence intensity of SAPT8 in the inflamed paws, whereas only a few fluorescence signals were observed in the inflamed paws of mice injected with NC (figure 5A). Then, the mice were sacrificed, and major organs (heart, liver, spleen, lung and kidney), as well as synovial tissues and hind paws, were harvested and visualised ex vivo. Both NC and SAPT8 displayed no obvious fluorescence signals in the heart, liver, spleen, lung and kidney (figure 5B). The fluorescence signals of SAPT8 in synovial tissues and hind paws were much stronger when compared with those of NC (figure 5B). We performed immunofluorescent staining to detect THY1-positive or vimentin-positive FLSs at the inflamed paws and investigated whether there was selective cellular accumulation of SAPT8 within CIA-MFLSs in vivo. Confocal imaging showed that fluorescence signals were highly aggregated in the CIA-MFLSs of mice injected with SAPT8, whereas signals were hardly observed in the CIA-MFLSs of mice injected with NC (figure 5C,D).
Tissue and cellular distribution of SAPT8 in collagen-induced arthritis (CIA) mice. (A) The global fluorescence signal of Cy3-labelled negative control (NC) or SAPT8 in mice with CIA as determined by the in vivo imaging system. CIA mice were treated with 5 mg/kg Cy3-labelled NC or SAPT8 by tail-vein injection for 24 hours. (B) Ex vivo fluorescence of the major organs (heart, liver, spleen, lung and kidney), synovial tissues and hind paws of the CIA mice treated with Cy3-labelled NC or SAPT8 for 24 hours. n=4 mice per treatment group. Data were represented as mean±standard deviation (SD) and statistical significance was calculated with the two-tailed Student’s t-test. **p<0.01, ***p<0.001, ****p<0.0001. (C,D) Fluorescence images showing the cellular distribution of Cy3-labelled NC or SAPT8 (red) on sections of the inflamed hind paws of CIA mice as determined by confocal microscopy. Immunofluorescence staining of THY1 (green, (C) or vimentin (green, (D) indicated the fibroblast-like synoviocytes in the inflamed hind paws. Cell nuclei were counterstained with 4’,6-diamidino-2-phenylindole (DAPI) (blue). Scale bars=50 µm.
SAPT8 monotherapy or SAPT8 combined with anti-TNF etanercept attenuates CIA in mice
CIA mice were intravenously injected with PBS, NC, SAPT8, biologic anti-TNF etanercept, a combination of NC with etanercept or a combination of SAPT8 with etanercept two times per week (figure 6A). After the treatment, arthritic scoring showed that SAPT8 dramatically impeded the progression of CIA when compared with NC or PBS (figure 6B). Microcomputerised tomography (μCT) scans demonstrated that CIA mice treated with SAPT8 showed reduced bone erosion and a lower ratio of bone surface to bone volume (BV) of ankle joints than the mice treated with NC or PBS (figure 6C,D). Histological analysis by H&E and safranin O/fast green (SO/FG) staining showed that SAPT8 caused a notable inhibition of synovial hyperplasia, bone erosion and cartilage destruction (figure 6E and online supplemental figure S9A). Quantitation of synovitis score, bone erosion and cartilage erosion on H&E-stained and SO/FG-stained sections consistently demonstrated the good therapeutic effects of SAPT8 in CIA mice (figure 6F–H). Immunofluorescence staining demonstrated that SAPT8 significantly increased p53 level and decreased expression of NCL, Bcl-2, IL-6, IL-1β, MMP3 and cyclooxygenase-2 (COX-2, a key mediator of inflammatory pain54 55) (figure 6I and online supplemental figure S9B–G). TUNEL assay showed that SAPT8 induced apoptosis of synovial tissues (figure 6J). Furthermore, we evaluated the therapeutic efficacy of the combination of SAPT8 with etanercept in the CIA mice. The combination of SAPT8 with etanercept more significantly relieved CIA when compared with the monotherapy using SAPT8 or etanercept or a combination of NC with etanercept (figure 6B–J and online supplemental figure S9A–G). In addition, we examined the hepatotoxicity of SAPT8 monotherapy and SAPT8 in combined therapy with etanercept. Blood biochemical assays demonstrated that the CIA mice treated with SAPT8 or a combination of SAPT8 with Etanercept had no obvious changes in liver function parameters including alanine aminotransferase (ALT), aspartate aminotransferase (AST), albumin (ALB) and total protein (TP), when compared with other treatment groups (online supplemental figure S9H).
Therapeutic effects of SAPT8 with or without the presence of an anti-tumour necrosis factor (anti-TNF) etanercept in collagen-induced arthritis (CIA) mice. (A) Flowchart for experimental design in CIA mice. Briefly, mice were immunised with bovine type II collagen for 28 days to establish CIA. Then, the CIA mice were systemically administered with phosphate buffered saline (PBS), negative control (NC), SAPT8, etanercept, a combination of NC with etanercept (NC+etanercept) or a combination of SAPT8 with etanercept (SAPT8+etanercept) two times per week for 39 days. NC or SAPT8 was intravenously injected at a dosage of 30 mg/kg, and etanercept was intraperitoneally administrated at a dosage of 5 mg/kg. (B) Arthritis scoring of paws from non-immunised (NI) mice or CIA mice in different treatment groups. (C) Representative microcomputerised tomography (μCT) images of the hind paws (dorsal and lateral view) from mice in each treatment group. (D) The ratio of bone surface to bone volume (BS/BV) of ankle joints as measured by μCT. (E) Representative images of H&E staining and SO/FG staining on sections of the hind paws (lateral view) from mice in each treatment group. Scale bars=500 µm. Quantification of synovitis score (F), bone erosion (G) and cartilage erosion (H). (I) Immunofluorescence staining of NCL on sections of the hind paws. Scale bars=50 µm. (J) Apoptosis of synovial tissues on sections of the hind paws as determined by terminal deoxynucleotidyl transferase 2'-deoxyuridine 5'-triphosphate (dUTP) nick end labelling (TUNEL) assay. Scale bars=50 µm. n=5 mice per treatment group. Data were represented as mean±standard deviation (SD) and statistical significance was calculated by the two-way analysis of variance (ANOVA) (B) or the one-way ANOVA (D and F–H) with a post hoc test. *p<0.05, **p<0.01.
SAPT8 recognises osteoarthritis (OA)-FLSs and relieves destabilisation of the medial meniscus (DMM)-induced OA in mice
Unlike RA, OA, characterised by cartilage deterioration, osteophyte formation and synovial pannus, is not an autoimmune disease. However, it is the most common degenerative joint disease.56 Pathogenesis of OA is complex and also not fully understood.57 Increasing evidence indicates that FLSs are activated in all stages of OA and contribute to the progression of OA.58 It has been proposed that FLSs-targeted therapeutic strategies in OA may possibly prevent cartilage breakdown while alleviating other symptoms.59 We showed that mRNA and protein expression of NCL in synovial tissues from OA patients was higher when compared with the synovial tissues of individuals with an ACLI (online supplemental figure S10A–B). OA-FLSs were isolated and characterised by high expression of THY1 and vimentin as well as the absence of CD68 (online supplemental figure S10C). SAPT8 also recognised OA-FLSs with good affinity (online supplemental figure S10D). SAPT8 modulated NCL-p53/Bcl-2 signalling and demonstrated inhibitory effects on the aggressive phenotype of OA-FLSs in vitro (online supplemental figure S10E–I). We conducted the surgical DMM to induce OA in mice and administrated the mice with PBS, NC or SAPT8 by intra-articular injection two times per week (online supplemental figure S11A). μCT scans showed that DMM mice treated with SAPT8 had less osteophyte formation and lower BV of calcified meniscus and synovium than the mice treated with PBS or NC (online supplemental figure S11B–D). H&E and SO/FG staining of joint sections demonstrated that synovial hyperplasia, bone erosion and cartilage erosion were significantly reduced in DMM mice treated with SAPT8 (online supplemental figure S11E–F). Quantitative analysis of synovitis score, Osteoarthritis Research Society International Score, and cartilage area consistently suggested the protective effects of SAPT8 on DMM-induced OA mice (online supplemental figure S11G–I). There were decreased NCL expression and increased apoptotic cells in synovial tissues of DMM mice treated with SAPT8 (online supplemental figure S11J and K). The level of p53 was higher, but levels of Bcl-2, IL-6, IL-1β, MMP3 and COX-2 were lower in synovial tissues of DMM mice treated with SAPT8 than those in synovial tissues of DMM mice treated with PBS or NC (online supplemental figure S12A–F). Blood biochemical assays demonstrated no obvious change in liver function parameters, including ALT, AST, ALB and TP, in OA mice treated with SAPT8 when compared with the mice treated with PBS or NC (online supplemental figure S12G).
Discussion
RA remains an unmet clinical challenge despite a variety of immunomodulatory therapies are available.18 During RA, FLSs, as the dominant non-immune cells of synovial tissues, are activated and suggested to be not just ‘passive responders’ but ‘imprinted aggressors’ in the joint microenvironment.12 RA-FLSs acquire a series of aggressive phenotypes, which are like those of the tumour cells, and transform from a friend to a foe.60 RA-FLSs play critical roles in many pathogenic events of RA. They break down the extracellular matrix by producing MMPs, directly invade and digest articular cartilage, promote bone erosion and enhance inflammation through the secretion of inflammatory mediators.61 In addition to the DMARDs-based immune therapies, exploring RA-FLSs targeted therapies has been highlighted for RA treatment.11 However, therapies targeting RA-FLSs have been hampered because of the lack of RA-FLSs-specific cell surface molecular signatures.16 In our study, we made a breakthrough in developing selective anti-RA-FLSs therapies without prior knowledge of RA-FLSs-specific surface signatures and identified a previously undiscovered molecular target for developing FLSs-directed therapies.
The ideal anti-FLS therapies have been proposed to be cell-specific and capable of interrupting one or more tumour-like phenotypes of RA-FLSs.18 To date, therapeutics have been developed to target the only two available cell surface molecular targets, CDH11 and PTPRS, to block adhesion and migration and invasion of RA-FLSs, respectively.18 However, no therapies specifically inducing apoptosis of RA-FLSs have been reported. Here, we addressed this issue using nucleic acid aptamers. We developed a novel CSCT-SELEX method for de novo screening of cytotoxic aptamers without prior knowledge of RA-FLSs-specific cell surface molecular signatures. Compared with the conventional cell-SELEX that only ensured the identification of cell-targeting aptamers, the CSCT-SELEX was designed to juggle the cell specificity and cytotoxicity of the aptamers against target cells via a CS-SELEX module in tandem with a CT-SELEX module. Monocytes are circulating leukocytes important in immune defence and stay for a relatively long time in circulation.62 63 We chose monocytes as one type of control cell in our CSCT-SELEX to eliminate the non-specific binding between aptamers and monocytes after in vivo administration of the selected aptamers in CIA mice. Additionally, osteoblasts are key cells involved in bone formation and joint repair.64 We chose osteoblasts as another type of control cell to minimise the influence of aptamers on osteoblasts in the joint microenvironment of CIA mice. Further, both were cell lines that could be easily obtained. Initially, our intention was to use normal joint FLSs from healthy individuals as non-RA-FLSs for the negative selection of the CSCT-SELEX to ensure the selectivity of the screened aptamers for RA-FLSs. This choice was predicated on the understanding that normal joint FLSs are the progenitors from which the aggressive RA-FLSs derive. However, CTCX-SELEX involved iterative cycles of selection, amplification and incubation to isolate aptamers. Each cycle consisted of repeated washing and eluting steps and required a tremendous number of target or NC cells. Regrettably, we encountered difficulties in acquiring sufficient human normal joint samples and thus were unable to proceed with using these cells as NCs. During the revision of our work, some insightful feedback from reviewers has led us to recognise that, besides the normal joint FLSs, a broader range of readily accessible non-RA fibroblasts, such as those from skin punch biopsies, immortalised fibroblast lines or iPSC-derived fibroblasts, could also serve as alternative NC cells for CSCT-SELEX due to their fibroblastic characteristics similar to RA-FLSs. Although these non-RA fibroblasts were not employed as NC cells during our CSCT-SELEX, we did investigate the binding specificity of our selected aptamers to both a human normal skin fibroblastic cell line (BJ) and a mouse fibroblastic cell line (L929) as per your suggestion. The results indicated that our aptamers did not bind to BJ or L929 cells, demonstrating their high selectivity for RA-FLSs. This selectivity is likely due to the specific expression of the target molecules of the aptamers on the surface of RA-FLSs, rather than on the non-RA fibroblasts, such as NCL for the SAPT8 aptamer.65 66
Although both SAPT4 and SAPT8 met our expectations as cell-specific and cytotoxic aptamers against RA-FLSs, we noticed that SAPT8 displayed advantages over SAPT4 regardless of the binding affinity and the proapoptotic capacity. We speculated that they might recognise different molecular signatures on the surface of the RA-FLSs. Then, we isolated membrane proteins of RA-FLSs and tried to identify the molecular targets of both SAPT4 and SAPT8 by pull-down assays. Interestingly, no protein was captured by SAPT4, while a specific protein with an observed molecular weight between 100 and 130 kDa was captured by SAPT8. Thus, we continued to identify the protein target of SAPT8 by mass spectrometry and interpreted the mechanism of action of SAPT8. Regarding the mechanism of action of SAPT4, we did not obtain a clue from the pull-down assay. However, it has been reported that the potential molecular targets of aptamers include not only proteins but also lipids or saccharides on the surface of the target cells.67 68 We raised a bold hypothesis that the target molecules of SAPT4 might be lipids or saccharides, which could not be identified by our pull-down assays. Nevertheless, we did attempt to prove this hypothesis and interpret the mechanism of action of SAPT4 since the characterisation of lipids or saccharides was not our expertise. By LC-MS/MS analysis and experimental verification, the protein target of SAPT8 was proven to be NCL, which has been reported to be a ubiquitous expressed nucleolar protein primarily involved in the synthesis and maturation of ribosomes,66 but it also serves as a specific shuttling protein between nucleus and membrane of various tumour cells.43 Previously, the surface expression of NCL on RA-FLSs was not reported and the pathologic role of NCL in RA-FLSs was also not characterised. Here, we revealed the presence of NCL on the surface of RA-FLSs and demonstrated that NCL conferred resistance to apoptosis and promoted migration, invasion and proliferation of RA-FLSs. These results suggested that NCL was a new surface molecular target for developing FLSs-directed therapies.
Mechanistically, we found that SAPT8 not only interacted with NCL but also decreased the protein level rather than the mRNA level of NCL in RA-FLSs. The most studied target genes of NCL are proapoptotic p53 and antiapoptotic Bcl-2.30 NCL interacts with p53 mRNA 5′-UTR and prevents its translation, whereas NCL stabilises Bcl-2 mRNA by binding to the 3′-UTR and protecting it from nuclease degradation.27 30 Both p53 and Bcl-2 are suggested to be key regulators involved in the tumour-like transformation of RA-FLSs.27 30 We demonstrated that SAPT8 targeted NCL to upregulate p53 expression and downregulate Bcl-2 level, providing a reasonable explanation for the proapoptotic capacity of SAPT8 against inflamed FLSs. Furthermore, it was interesting to see that SAPT8 could be internalised via both clathrin-mediated and caveolae-mediated endocytic pathways and enter the lysosome of RA-FLSs, suggesting that SAPT8 might carry NCL into the lysosome for degradation. Consistently, inhibition of lysosomal activity of RA-FLSs diminished the SAPT8-induced degradation of NCL protein, and blockage of SAPT8-NCL interaction by AS1411 reversed the SAPT8-induced suppression of tumour-like phenotypes of RA-FLSs, validating that SAPT8 acted as a lysosomal degrader of NCL to modulate NCL-p53/Bcl-2 signalling and display anti-RA-FLSs potential. As NCL shuttles between the nucleus, the cytoplasm and the cell surface of RA-FLSs, we speculated that, once the SAPT8-mediated elimination of surface NCL happens, intracellular NCL would be retranslocated onto the surface of RA-FLSs and excessive free SAPT8 in the extracellular environment would then bind with new NCL on the surface of RA-FLSs and induce a new cycle of NCL degradation until the exhaustion of extracellular SAPT8. This possibly explained the SAPT8-induced overall degradation of NCL in RA-FLSs as observed in this study. Besides our SAPT8, AS1411 and iSN04 were previously reported to be aptamers targeting NCL.42 43 53 However, unlike the SAPT8, we demonstrated that neither AS1411 nor iSN04 triggered the degradation of NCL, while they decreased the NCL phosphorylation in RA-FLSs. Based on all our results, we could easily find that SAPT8 at nanomolar concentration was enough to inhibit the aggressive phenotypes of RA-FLSs, but micromolar AS1411 or iSN04 were required to exhibit therapeutic effects on RA-FLSs. Moreover, we also observed that iSN04 was less potent than AS1411 since iSN04 only showed effects at a concentration of 10 µM, whereas the effective concentration of AS1411 was 5 µM. The underlying reasons for the varied mechanisms of action and differential therapeutic efficacies among SAPT8, AS1411 and iSN04 remain to be elucidated. As NCL is an established cell surface molecular target for a variety of tumours and AS1411 targeting NCL has been tested for antitumour therapy,43 we propose that SAPT8 may also be used for antitumour therapy, which needs to be further confirmed in future studies. The aptamer-induced lysosomal degradation of target proteins was not merely seen in our study. A previous study identified an aptamer that was shown to target and facilitate the lysosomal degradation of an oncoprotein ErbB-2, thus retarding the tumorigenic growth of gastric cancer.69
When applied SAPT8 in CIA mice, we observed the selective accumulation of SAPT8 in FLSs both at tissue and cellular levels, which was consistent with the good binding ability and specificity of SAPT8 with CIA-MFLSs in vitro. After periodic administration, SAPT8 monotherapy effectively decreased NCL expression and induced cell apoptosis of synovial tissues, leading to inhibited synovial hyperplasia, bone erosion and cartilage destruction in CIA mice. Moreover, combined therapy using SAPT8 and an anti-TNF biological DMARD synergistically alleviates the disease progression of CIA in mice. These results supported the currently popular theory that RA-FLSs-directed therapies could be a potentially alternative or complementary approach to the current DMARDs-based immune-directed therapies.12–14 Besides RA, OA is also characterised by synovial hyperplasia, bone erosion and cartilage destruction, even though it has a very different aetiology from RA.70 OA is a degenerative joint disease with a still-evolving concept of pathophysiology from being viewed as a cartilage-limited destruction to a multifactorial chronic disease that affects the whole joint.71 FLSs have been recognised to play a critical role in the pathophysiology of OA and are suggested to be cellular targets for OA treatment.72 In our study, we found that OA patients also had enhanced expression of NCL in synovial tissues, and intra-articular injection of SAPT8 reduced NCL expression, induced cell apoptosis of synovial tissues and displayed protective effects on DMM-induced OA mice. These studies demonstrated the broad prospects of SAPT8 in treating a variety of joint diseases featuring synovial hyperplasia or synovitis.
As SAPT8 was a cytotoxic aptamer selected for specifically targeting RA-FLSs, it was necessary to evaluate the side effects of SAPT8 on normal tissues in vivo. In CIA mice, we showed that there was no obvious accumulation of SAPT8 in the heart, liver, spleen and lung and less distribution of SAPT8 in the kidney. Liver function parameters, such as ALT, AST, ALB and TP, were in normal ranges in CIA mice systemically treated with SAPT8 monotherapy or SAPT8 in a combined therapy with an anti-TNF biological DMARD. Local administration of SAPT8 also showed a good safety profile in DMM-induced OA mice. These could be attributed to the excellent specificity and binding affinity of SAPT8 with inflamed FLSs and reflect the huge advantage of our proposed CSCT-SELEX strategy in developing cytotoxic aptamers specifically killing target cells but without damaging normal cells. This advantage of CSCT-SELEX will also benefit the screening of new types of cytotoxic drugs for treating other prevalent diseases, especially tumours. Traditional cytotoxic drugs, also known as chemotherapeutics, are medications that drive cells into apoptosis, thus causing the arrest of cancers or other non-malignant diseases, such as RA, multiple sclerosis and lupus.73 74 Most of them are small molecular drugs that interfere with DNA synthesis, or produce chemical damage to DNA, leading to cell death. However, they are not specific to diseased cells and kill all dividing cells including healthy cells, leading to severe side effects, such as nausea and vomiting, and a narrow therapeutic window.75 76 The current leading-edge solution is to conjugate the cytotoxic drugs with monoclonal antibodies against the specific antigens on the surface of tumour cells, that is, antibody-drug conjugates (ADCs), to achieve the tumour-selective delivery of the cytotoxic drugs.77 However, ADCs have limited solid tumour permeability due to their large molecular weights, and the complexity of payload pharmacokinetics and insufficient release can also affect the efficacy of ADCs.78 The CSCT-SELEX strategy developed in our study could directly confer the cell-targeting ability to the cytotoxic aptamers, which may address the limitation of ADCs.
There are also a few limitations of our study. Even though we proved that NCL was expressed on the surface of RA-FLSs and cell surface NCL contributed to the tumour-like phenotypes of RA-FLSs, we still know nothing about how NCL shuttles from the nucleus to the cell surface of RA-FLSs. The same question was also an unaddressed challenge in tumours although the surface expression of NCL on tumours has been discovered for decades.30 Second, the efficacy of nucleic acid therapies can be limited by unwanted nuclease-mediated degradation. Nucleic acid drugs are always designed to enhance their pharmacokinetic properties in vivo via chemical modifications, such as PS linkage, 2′-OMe, 2′-MOE, 2′-F and LNA.79 In our study, as we saw the excellent stability of SAPT8 in a serum degradation test in vitro, we directly performed the proof-of-concept study to determine the therapeutic efficacy of unmodified SAPT8 in arthritic animal models. In the future, SAPT8 should be extensively modified before commercialization and entering clinical trials. Although we observed that SAPT8 decreased the expression of COX-2, a key mediator involved in inflammatory pain, in arthritic mice at the end of SAPT8 aptamer treatment, it should be worth exploring whether the induction of apoptosis of RA-FLSs by SAPT8 could potentially lead to temporary increased pain in future studies.
In conclusion, CSCT-SELEX is a promising strategy for generating cell-specific and cytotoxic aptamers. SAPT8, selected by CSCT-SELEX, could be an aptamer drug selectively killing RA-FLSs, which may be a potential alternative or complementary approach to the current DMARDs-based immune-directed therapies. NCL is a newly identified molecular target on the surface of RA-FLSs and will pave the way for developing diverse FLSs-directed therapies for treating arthritis.
Materials and methods
Cell culture
RA-FLSs, OA-FLSs and ACLI-FLSs were isolated from synovial tissues of patients with RA, OA and ACLI, respectively.80 81 H-MFLSs, CIA-MFLSs and CIA-RFLSs were isolated from synovial tissues of healthy mice, CIA mice and CIA rats, respectively.39 The human synovial cell line SW982 (HTB-93), human monocytic cell line THP-1 (TIB-202), human osteoblastic cell line hFOB 1.19 (CRL-3602), human normal skin fibroblast cell line BJ (CRL-2522), mouse osteoblast-like cell line MC3T3-E1 (CRL-2593), mouse macrophage-like cell line RAW 264.7 (TIB-71), mouse fibroblast cell line L929 (CCL-1) and human embryonic kidney cell line 293 T cells (CRL-3216) were obtained from American Type Culture Collection. The mouse chondrogenic cell line ATDC5 (CVCL_3894) was obtained from the Type Culture Collection of the Chinese Academy of Sciences. RA-FLSs, OA-FLSs and ACLI-FLSs were cultured in Dulbecco’s Modified Eagle Medium (DMEM, Corning, USA) supplemented with 20% fetal bovine serum (FBS). H-MFLSs, CIA-MFLSs, CIA-RFLSs, BJ cells, RAW 264.7 cells and 293 T cells were cultured in DMEM medium supplemented with 10% FBS. SW982 cells were cultured in Leibovitz’s L-15 medium (Corning, USA) supplemented with 10% FBS. THP-1 cells were cultured in RPMI-1640 medium supplemented with 10% FBS. hFOB 1.19 cells were cultured in DMEM/Ham’s F12 (Thermo Fisher, USA) medium supplemented with 10% FBS, 2.5 mM L-glutamine, and 300 µg/mL G418 (Beyotime, China). MC3T3-E1 cells were cultured in Alpha MEM medium (Corning, USA) supplemented with 10% FBS. L929 cells were cultured in EMEM medium (Corning, USA) supplemented with 10% FBS. ATDC5 cells were cultured in DMEM/F12 supplemented with 10% FBS. All media were supplemented with 1% penicillin–streptomycin. All cells were negative for mycoplasma and maintained at a 37°C humid atmosphere with 5% CO2.
Random ssDNA library and primers
The initial ssDNA library contained a central randomised sequence of 35 nucleotides flanked by two 20-nt primer hybridisation sites: 5′-TGAGAATATGTAGACGATCC-(35N)-CGGAGCTTCAAGATGATCTG-3′, the forward primer: 5′-TGAGAATATGTAGACGATCC-3′, the reverse primer: 5′-CAGATCATCTTGAAGCTCCG-3′. The ssDNA library, primers and all ssDNA aptamer candidates were synthesised by Sangon Biotech (Shanghai, China).
CSCT-SELEX procedure
The CSCT-SELEX consisted of a CS-SELEX module followed by a CT-SELEX module. The CS-SELEX was based on the traditional cell-SELEX with some modifications.25 During CS-SELEX, RA-FLSs were employed as target cells in the positive selection, and THP-1 and hFOB 1.19 cells were used as control cells in the negative selection. We used the FLSs from patients who shared some similar features (such as gender and age) and comparable levels of indices for the severity of RA (such as CRP and ESR), and we characterised them with equivalent expression of molecular markers, including THY1 and vimentin, and used them at uniform passages in a specific experiment. For the first round of CS-SELEX, a 10 nmol random ssDNA library was dissolved in 400 µL binding buffer containing 4.5 g/L glucose, 0.1 mg/mL yeast tRNA, 5 mM MgCl2 and 1 mg/mL bovine serum albumin (BSA) in Dulbecco’s PBS (DPBS). After denaturing at 95°C for 5 min and rapid cooling on ice for 10 min, the ssDNA library was incubated with 1×106 RA-FLSs for 3 hours at 37°C. Then, the cells were centrifuged and washed with 500 µL washing buffer containing 4.5 g/L glucose and 5 mM MgCl2 in DPBS to remove unbound sequences. The bound sequences were recovered by heating the cell-DNA complex at 95°C and PCR amplified to create a new pool. PCR was conducted using unlabelled forward primers and biotin-labelled reverse primers (5 min at 95°C, 10–20 cycles of 30 s at 95°C, 30 s at 55°C and 30 s at 72°C, followed by 5 min at 72°C). After denaturing in an alkaline condition (0.1 M NaOH), the sense ssDNA strand was separated from the biotinylated antisense ssDNA strand by Streptavidin Sepharose High-Performance Beads (GE Healthcare, UK), desalted and lyophilised for the next round of selection. From the third round of CS-SELEX, negative selection was carried out to filter out sequences that might bind to the molecules existing on the surface of both target cells and control cells prior to the positive selection. 200 pmol enriched ssDNA pool dissolved in 400 µL binding buffer was incubated with 5×106 control cells (THP-1 and hFOB 1.19 cells) for 3 hours at 37°C. The unbound ssDNA was collected and used for positive selection with RA-FLSs. To acquire aptamers with high affinity and specificity, the wash strength was enhanced gradually by increasing the volume of washing buffer (from 0.5 to 5 mL) and the number of washes (from three to five). The progress of the CS-SELEX was monitored by a microplate reader (PerkinElmer, EnSpire, USA). Briefly, the enriched ssDNA pools from the 4th round (R4), the 8th round (R8) and the 12th round (R12) were PCR-amplified using Cy3-labelled forward and biotin-labelled reverse primers before obtaining the sense ssDNA strand by alkaline denaturation and subsequent streptavidin-coated magnetic beads separation. The enriched ssDNA pools with Cy3 labelling were then incubated with 5×105 RA-FLSs, THP-1 cells or hFOB 1.19 cells for 3 hours at 37°C before determining the MFI of the cells using the microplate reader (PerkinElmer, EnSpire, USA).
After 12 rounds of CS-SELEX, CT-SELEX was started using the enriched ssDNA pool that exhibited the highest binding ability with RA-FLSs from the CS-SELEX module as the initial pool. 1000 nM enriched ssDNA pool was heated at 95°C for 5 min, chilled on ice for 10 min and then incubated with 5×106 RA-FLSs for 3 days. During the incubation, the medium containing 1000 nM enriched ssDNA pool was changed every day. After the incubation, apoptotic RA-FLSs were collected using an Annexin V MicroBead Kit (Miltenyi Biotec, USA) and resuspended in 400 µL binding buffer. The apoptotic cell-associated sequences were recovered by heating the apoptotic cell-DNA complex at 95°C. PCR was conducted using unlabelled forward primers and biotin-labelled reverse primers (5 min at 95°C, 10–20 cycles of 30 s at 95°C, 30 s at 55°C and 30 s at 72°C, followed by 5 min at 72°C). After denaturing in an alkaline condition (0.1 M NaOH), the sense ssDNA was separated from the biotinylated antisense ssDNA strand by Streptavidin Sepharose High-Performance Beads (GE Healthcare, UK), desalted and lyophilised for the next round of selection. The progress of the CT-SELEX was monitored by the annexin V-coated magnetic beads. Briefly, 5×106 RA-FLSs were daily treated with 400 nM enriched ssDNA pools from the first round (R1′), the third round (R3′), the sixth round (R6′) and the ninth round (R9′) of CT-SELEX for 3 days. The percentage of apoptotic RA-FLSs was determined by the Annexin V MicroBead Kit (Miltenyi Biotec, USA). After the CT-SELEX, the enriched ssDNA pool with the highest proapoptotic ability was PCR-amplified using unlabelled primers and cloned into Escherichia coli by using the TA cloning kit (Invitrogen, USA). Cloned sequences were sequenced to identify individual aptamer candidates by Sangon Biotech (Shanghai, China).
Binding assays by flow cytometry
RA-FLSs (6×104 per well), ACLI-FLSs (6×104 per well), CIA-MFLSs (5×104 per well), CIA-RFLSs (3×104 per well), SW982 cells (4×104 per well), hFOB 1.19 cells (5×104 per well), MC3T3-E1 cells (4×104 per well), RAW 264.7 cells (6×104 per well), L929 cells (4×104 per well), ATDC5 cells (3×104 per well), H-MFLSs (5.5×104 per well), BJ cells (4×104 per well) or OA-FLSs (5×104 per well) were seeded in six-well plates and allowed to adhere for 24 hours. THP-1 cells (6×104 per well) were maintained in six-well plates for 24 hours. After different treatments in each study, the adherent cells were digested with 0.25% trypsin–EDTA (Thermo Fisher, USA) and washed with PBS. Fluorescence signals were detected by a flow cytometer (BD, FACSCanto SORP, USA) and data were analysed by FlowJo software. To measure the equilibrium dissociation constant (Kd), RA-FLSs (6×104 per well) were treated with increasing concentrations of Cy3-labelled aptamers. The Kd value of the interaction between aptamers and RA-FLSs was determined by fitting the dependence of fluorescence intensity using Prism GraphPad V.9 software with the equation Y=BmaxX/(Kd+X), where X was the aptamer concentration; Y was MFI of X; and Bmax was maximum MFI.82
Serum stability assay
Aptamers were incubated with 50% FBS in the medium for 0, 1, 3, 6, 12 and 24 hours at 37°C. Subsequently, the samples were mixed with DNA loading dye and subjected to non-denaturing PAGE for 2 hours at 100 V. The gel was stained with 4S Gelblue (Sangon Biotech, China) for 15 min at room temperature and then visualised under ultraviolet light.83
Apoptosis analysis by flow cytometry
RA-FLSs (3×104 per well), CIA-MFLSs (2×104 per well), CIA-RFLSs (2×104 per well), SW982 cells (3×104 per well), BJ cells (4×104 per well) or OA-FLSs (3.5×104 per well) were seeded in six-well plates and allowed to adhere for 24 hours. After different treatments in each study, the cells were digested with 0.25% trypsin–EDTA and washed with PBS. An Annexin V-FITC Apoptosis Detection Kit (Beyotime, China) was used to stain apoptotic cells. Flow cytometry was performed to detect apoptotic cells and data were analysed by FlowJo software.84
TUNEL assay
After different treatments in each study, RA-FLSs and joint tissue sections were fixed in 4% formalin for 10 min and incubated with Cy3-labelled dUTP and terminal deoxynucleotidyl transferase enzyme mixture from a One Step TUNEL Apoptosis Assay Kit (Beyotime, China) for 1 hour at room temperature. Cell nuclei were counterstained with DAPI (Thermo Fisher, USA), and apoptotic cells were imaged using a confocal fluorescence microscope (Zeiss, LSM980, DE).85
CCK-8 assay
RA-FLSs (2×103 per well), CIA-MFLSs (1×103 per well), CIA-RFLSs (1×103 per well) or SW982 cells (1×103 per well) were seeded in 96-well plates and allowed to adhere for 24 hours. After different treatments in each study, proliferation of the cells was measured using CCK-8 according to the manufacturer’s instructions (MedChemExpress, MCE, USA). The cells were washed with PBS, and 10 µL CCK-8 solution and 90 µL complete medium were added to each well. Plates were detected by a microplate reader at 450 nm.86
Colony formation assay
RA-FLSs (2×103 per well), CIA-MFLSs (800 per well), CIA-RFLSs (500 per well), SW982 cells (1×103 per well) or OA-FLSs (3×103 per well) were seeded in six-well plates and allowed to adhere for 24 hours. After different treatments in each study, the cells were fixed in 4% formalin for 20 min and stained with crystal violet (Aladdin, China) for 30 min at room temperature.80 Digital images of the colonies were obtained using a camera. The number of colonies was counted using ImageJ software.
Transwell migration and invasion assays
To examine cell migration, 200 µL serum-free medium containing RA-FLSs (8×103 per well), CIA-MFLSs (6×103 per well), CIA-RFLSs (8×103 per well), SW982 cells (1×104 per well) or OA-FLSs (8.5×103 per well) were directly seeded in 24-well transwell inserts with a pore size of 8.0 µm (Corning, New York, USA). To determine cell invasion, the upper chambers were coated with a solution of matrigel at a concentration of 1.25 mg/mL (20 µL per well) before cell seeding. Then, 600 µL medium containing 20% FBS was added to the lower chamber. After different treatments in each study, non-migrated cells were removed by cotton swabs, and cells that migrated to the lower face of the upper chamber were fixed in 4% formalin for 20 min and stained with crystal violet for 30 min at room temperature.87 The chambers were washed with PBS and air dried. Photographs were obtained by an inverted microscope (NanoZoomer S60, Japan), and the number of migrated cells was quantitated using ImageJ software.87
Wound healing assay
RA-FLSs, CIA-MFLSs, CIA-RFLSs or SW982 cells were seeded in six-well plates and cultured until cells became more than 90% confluent. Subsequently, the cells were scratched uniformly with 200 µL sterile pipette tips, and floating cells and cell debris were washed off with culture medium. After different treatments in each study, wound healing was captured by the inverted microscope, and the area of the wound gap was calculated and recorded using ImageJ software.77
Western blotting
RA-FLSs or OA-FLSs, after different treatments in each study, were lysed in lysis buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% Triton X-100, 2 mM EDTA and 10% glycerol) containing protease inhibitor cocktails (Sigma, USA). The total protein concentration of each cell sample was evaluated by a bicinchoninic acid assay. The proteins in each cell sample were separated by SDS-PAGE and transferred to a polyvinylidene fluoride membrane (Millipore, Massachusetts, USA) using a transfer apparatus (Bio-Rad, Trans-Blot Turbo, USA). After blocking, the membranes were reacted with primary antibodies at 4°C overnight. Subsequently, the membranes were washed and reacted with the corresponding horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 hour at room temperature. The membranes were probed by using an Enhanced Chemiluminescence ECL Kit (ABclonal, China) and visualised with a chemiluminescence imaging system (Tanon, Multi5200, China).83 The primary antibodies were as follows: anti-NCL antibody (14574, Cell Signaling Technology, CST, USA), anti-Bcl-2 antibody (ab182858, Abcam, UK), anti-p53 antibody (ab26, Abcam, UK), anti-IL-1β antibody (ab254360, Abcam, UK), anti-MMP3 antibody (ab52915, Abcam, UK) and anti-GAPDH antibody (AC002, Abclonal, China).
Pull-down assay by aptamers
Membrane proteins of RA-FLSs were extracted using a Minute Plasma Membrane Protein Isolation Kit (Invent Biotechnologies, USA) and incubated with biotin-labelled aptamers at 4°C for 6 hours. The Streptavidin Sepharose High-Performance Beads (GE Healthcare, UK) were added to the mixture of membrane proteins and aptamers and incubated overnight at 4°C. The beads were washed and heated to elute the captured proteins. The captured proteins were separated by SDS-PAGE gel. After silver staining of the gel using a Pierce Silver Stain for Mass Spectrometry Kit (Thermo, USA), the specific band of protein captured by aptamer was excised for digestion and analysed by LC-MS/MS in BGI Genomics. A MASCOT database search was used to assign possible protein candidates to the MS results.88
Dot blotting assay
The rhNCL protein (MCE, USA) was diluted into 10 µg/mL and spotted (2 µL/dot) onto a nitrocellulose membrane. The membrane was dried naturally, blocked with 5% non-fat dry milk for 2 hours at room temperature and incubated with biotin-labelled aptamer overnight at 4°C. Subsequently, the membrane was washed, incubated with HRP-labelled streptavidin (MCE, USA) for 1 hour at room temperature and probed by using an Enhanced Chemiluminescence ECL Kit (ABclonal, China) and visualised with a chemiluminescence imaging system (Tanon, Multi5200, China).82
Real-time PCR
Total RNA was extracted using the TransZol Up Plus kit (TransGen, China). cDNA was synthesised using the PrimeScript RT Master Mix (Takara, Japan). Real-time PCR was performed on a CFX 96 Touch Real-Time PCR System (Bio-Rad, USA), with SYBR Green qPCR Master Mix from TransGen Biotech. Data were normalised to the expression levels of the housekeeping gene GAPDH and expressed as mean±SEM. Results were calculated as fold change compared with the expression level of each gene in control samples using the comparative threshold cycle (Ct) method with the formula 2(−∧∧Ct).83 Primer sequences are listed in online supplemental table S3.
Gene silencing by siRNA
RA-FLSs were seeded in six-well plates and allowed to adhere for 24 hours. According to the manufacturer’s instructions, the cells were transfected at a final concentration of 20 nM siRNA using Lipofectamine RNAiMAX Reagent (Thermo Fisher, USA). The sequences of siRNA are listed in online supplemental table S4.
Lentivirus production and transduction
Lentivirus production and transduction were performed as previously described.89 Briefly, 1 mg of desired plasmids, 0.75 mg psPAX2 and 0.25 mg pMD2.G were cotransfected into 293 T cells using Effectene transfection reagent (Qiagen, Germany). The culture medium was replenished at 12 hours post transfection. Subsequently, the supernatant was collected at 48 and 72 hours. For transduction, RA-FLSs (5×104 per well) were seeded in six-well plates and allowed to adhere for 24 hours. The lentivirus solutions with 8 µg/mL polybrene (Sigma-Aldrich, USA) were added to the cells. The culture medium was changed at 36 hours post transfection, and antibiotic selection was started at 72 hours post transfection with 2 µg/mL puromycin (Beyotime, China) in the medium. The targeted gene expression on RA-FLSs was measured at 4 days after adding antibiotic selection. The information on plasmids used in this study is listed in online supplemental table S5.
EdU proliferation assay
Cell proliferation was determined using a BeyoClick EdU Cell Proliferation Kit with Alexa Fluor 488 (Beyotime, China). Briefly, RA-FLSs (1×104 cells per well) were seeded in confocal plates and allowed to adhere for 24 hours. After different treatments in each study, 50 µL EdU was added into the cells and incubated at 37°C. The cells were fixed in 4% formalin for 10 min, permeabilized with 0.2% Triton X-100 for 10 min, and blocked with 5% BSA in PBS for 1 hour. A staining cocktail including 430 µL click reaction buffer, 50 µL click additive solution, 1 µL Alexa Fluor 488 and 20 µL CuSO4 were added into the cells for 30 min. Finally, the cells were counterstained with DAPI, and images were acquired using a confocal fluorescence microscope.90
RNA sequencing analysis
RNA sequencing was performed in Sangon Biotech (Shanghai, China). Raw sequencing reads were subjected to a comprehensive quality check using FastQC (V.0.11.9). Cleaned reads were aligned to a reference genome. Postalignment data management, including sorting of alignment files, index construction and alignment efficiency statistics, was performed using SAMtools91 (V.1.10). Gene expression levels were determined using featureCounts92 (V.2.0.1), which assigned reads to genes or transcripts to generate counts for each feature. Normalisation of gene expression across sh-NCL and sh-NC datasets was executed using the DESeq V.2 package.93 DEGs between sh-NCL and sh-NC groups were identified based on criteria of p value<0.05 and |log2FC|>1. DEGs visualisation was achieved using volcano plots (ggplot V.2 package) and heatmaps (pheatmap package). DEGs were subjected to GO and KEGG pathway enrichment analyses via the clusterProfiler R package.94 95 Only pathways with a p value less than 0.05 were considered for significant enrichment. Differences in biological processes between si-NCL and si-NC groups were explored using GSEA.96 Analysis was based on the ‘c2.cp.all.v2022.1.Hs.symbols.gmt’ gene set from MSigDB, with pathways holding a false discovery rate<0.25 regarded as significantly enriched.
Collection of human samples
Synovial tissues of RA and OA patients were collected during joint replacement surgery and synovial tissues of ACLI patients were collected during ACL reconstruction surgery. The collected synovial tissues were preserved in liquid nitrogen. The synovial tissues were detected for positive expression of fibroblastic biomarkers THY1 and vimentin,33 34 as well as the absence of a macrophage marker CD68.35 The clinical characteristics of patients are shown in online supplemental table S6.
CIA rat model
Male SD rats (170–190 g of weight) were purchased from the Chinese University of Hong Kong and kept with free access to food and water under standard temperature conditions (22°C) and a 12 hours light/dark cycle. The CIA rat model was established as follows.97 Briefly, bovine type II collagen (Chondrex, USA) was emulsified in an equal volume of Incomplete Freund’s adjuvant (Chondrex, USA). The rats were immunised subcutaneously at the base of the tail with 200 µL emulsion containing 200 µg collagen. Booster immunisation was administered on day 7 with a 100 µL injection of the same emulsion as the first time. Arthritis severity was evaluated by clinical arthritic scores, which were performed by two independent, blinded observers. Scoring was performed with a 0–4 scale, where 0=no swelling or erythema; 1=slight swelling and/or erythema; 2=low-to-moderate oedema; 3=pronounced oedema with limited joint usage; and 4=excess oedema with joint rigidity. Each paw was graded, and the maximum possible score was 16 for each rat. A rat with a score of one or above was regarded as arthritic.
CIA mouse model
Male DBA/1 mice aged 8–10 weeks were from the Gempharmatech and kept with free access to food and water under standard temperature conditions (22°C) and a 12 hours light/dark cycle. CIA mouse model was established as follows.98 Briefly, bovine type II collagen (Chondrex, USA) was emulsified in an equal volume of Complete Freund’s adjuvant (Chondrex, USA). The mice were immunised with a single subcutaneous injection at the base of the tail with 100 µL emulsion containing 100 µg collagen and 2 mg/mL Mycobacterium tuberculosis. Arthritis severity was evaluated by clinical arthritic scores, which were performed by two independent, blinded observers. Scoring was performed with a 0–4 scale, where 0=normal; 1=redness and/or swelling in one joint; 2=redness and/or swelling in more than one joint; 3=redness and/or swelling in the entire paw; and 4=deformity and/or ankylosis. Each paw was graded, and the maximum possible score was 16 for each mouse. A mouse with a score of one or above was regarded as arthritic.
DMM mouse model
Male C57BL/6J mice aged 6–8 weeks were purchased from the Gempharmatech and kept with free access to food and water under standard temperature conditions (22°C) and a 12 hours light/dark cycle. DMM surgery was conducted in the right knees in C57BL/6 mice for the establishment of the OA model.99 Briefly, a mouse was anaesthetised using a 2.5% avertin with a dosage of 250 mg/kg body weight via intraperitoneal injection. The medial meniscotibial ligament of the right knee joint of each mouse was transected. The sham-operated controls underwent the same surgery except that the meniscotibial ligament was not cut.
In vivo fluorescence imaging
Cy3-labelled NC or SATP8 was intravenously given to the CIA mice via the tail vein. An in vivo imaging system Spectrum imaging system (PerkinElmer, USA) was used to examine the fluorescence of NC and SAPT8. The quantification of fluorescence was analysed via Living Image V.4.4 software (Caliper Life Sciences, USA).100
μCT analysis
The joints were fixed in 10% formalin (Solarbio, China) and then transferred to 70% ethanol. Scanning was performed at a Skyscan scanner 1276 high-resolution µCT scanner (Bruker, Germany) with a voxel size of 10 µm, at 60 kVp/100 µAmp. The images were reconstructed to generate three-dimensional (3D) images using NRecon software (Bruker, Germany) and Date Viewer (Bruker, Germany). The μCT data were processed using CTAn software (Bruker, Germany) and images were generated using CTVOX software (Bruker, Germany). Bone erosion was quantified as previously described.99 Scanned images were processed using the same thresholds to keep the 3D structural rendering of each sample.
Histological analysis
The joints were fixed, decalcified, dehydrated and embedded in paraffin. A series of sections (5 µm thick) was stained with H&E and SO/FG (Solarbio, China) using standard protocols according to the manufacturer’s instructions. The specimens were independently viewed and analysed by two pathologists under a light microscope (NanoZoomer S60, Japan). The histopathological scoring was performed as previously described.16 99
Immunofluorescent staining
Tissue sections were subjected to deparaffinization using xylene, rehydrated and antigen retrieval. The sections were permeabilised with 0.2% Triton X-100 for 10 min, blocked with 5% BSA in PBS for 1 hour and incubated with primary antibodies anti-NCL (14574, CST, USA), anti-Bcl-2 antibody (ab182858, Abcam, UK), anti-p53 antibody (ab26, Abcam, UK), anti-IL-6 (ab233551, Abcam, UK), anti-IL-1β (ab254360, Abcam, UK), anti-MMP3 (ab52915, Abcam, UK) and anti-COX2 (ab179800, Abcam, UK) at 4°C overnight. After washing, the sections were incubated with anti-mouse Alexa Fluor 488 (ab150113, Abcam, UK) or anti-rabbit Alexa Fluor 488 (ab150077, Abcam, UK) for 1 hour at room temperature. Finally, the sections were counterstained with DAPI, and images were acquired using a confocal fluorescence microscope.99
Serum biochemical assays
Blood samples were collected from mice after treatment via hepatic portal vein puncture. Serum was collected via centrifugation of blood samples in 12 000×g at 4°C for 30 min. Serum biochemical parameters, including ALT, AST, ALB and TP, were analysed by an MS-480 Automatic Biochemistry Analyzer (Medicalsystem Biotechnology, China).101
Statistical analysis
All numerical data are expressed as the mean±SD. Statistical differences among multiple independent groups were analysed by the one-way analysis of variance (ANOVA) with a post hoc test. Statistical differences between two independent groups were determined by the two-tailed Student’s t-test. Statistical differences between the groups that were defined by two categorical factors were analysed by the two-way ANOVA with a post hoc test. All statistical analyses were performed with GraphPad Prism V.9 software. P<0.05 was considered statistically significant. We chose the representative images based on the average level of the data for each group. For the in vivo experiments, the sample size was predetermined by a power calculation. The mice were grouped randomly and blindly by researchers. The mice in poor body condition before the experiments were excluded.
Data availability statement
Data are available in a public, open access repository. Data are available on reasonable request. Data are available in a public, open access repository. The RNA sequencing data have been uploaded to Gene Expression Omnibus platform with the accession number GSE268214.
Ethics statements
Patient consent for publication
Ethics approval
All the clinical procedures were approved by the committees of clinical ethics in the Shanghai Guanghua Hospital of Integrative Medicine, Shanghai University of Traditional Chinese Medicine. We obtained informed consent from the participants. The protocols of animal experiments were approved by the Institutional Animal Care and Use Committee of the Southern University of Science and Technology (SUSTech-JY202302087, SUSTech-JY20241003) and the Committees of Animal Ethics and Experimental Safety of the Hong Kong Baptist University (REC/21-22/0151). We complied with all relevant ethical regulations for animal testing and research.
Acknowledgments
The authors acknowledge the assistance of the Southern University of Science and Technology Core Research Facilities, the Microscope and Imaging Center and the Experimental Animal Center of the Southern University of Science and Technology. Schematic diagram of CTCS-SELEX was created with BioRender.com.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Handling editor Josef S Smolen
FQ and DX contributed equally.
Contributors CL, AL and XF are guarantors for all content. CL, AL, and XF supervised and revised the manuscript. FQ and DX performed the major research and wrote the paper with equal contributions. HC, ZW, JH, CC, YL and XY provided the technical support and professional expertise. All authors contributed to the article and approved the submitted version.
Funding This work is supported by the National Natural Science Foundation Council of China (82172386 and 81922081 to CL and 82100943 to XF), the 2020 Guangdong Provincial Science and Technology Innovation Strategy Special Fund (Guangdong-Hong Kong-Macau Joint Lab) (2020B1212030006 to AL), the Guangdong Basic and Applied Basic Research Foundation (2022A1515012164 to CL), the Science, Technology, and Innovation Commission of Shenzhen (JCYJ20210324104201005 to CL), the Shenzhen Medical Research Fund (A2303061 to XF), the Hong Kong General Research Fund (12102722 to AL), the Hong Kong RGC Theme-based Research Scheme (T12-201/20-R to AL) and Shenzhen LingGene Biotech.
Competing interests Shenzhen LingGene Biotech has a patent application related to this work.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.