Novel gene therapy for rheumatoid arthritis with single local injection: adeno-associated virus-mediated delivery of A20/TNFAIP3

Dear Editor,

Multiple experiments have established TNF alpha-induced protein 3 (A20/TNFAIP3) as a critical regulator associated with rheumatoid arthritis (RA) [1, 2]. The lack of TNF-α-induced protein 3 (A20) promotes the NOD-like receptor protein 3 (NLRP3) inflammasome and induces spontaneous arthritis, while increase of A20 reduces the secretion of IL-1β and favors immunological tolerance [3, 4]. Hence, we investigate the feasibility of recombinant adeno-associated virus 6 (rAAV6)-mediated A20 gene therapy in a collagen-induced arthritis (CIA) model.

We first investigated cytomegalovirus (CMV) promoter-regulated gene delivery (Additional file 1: Fig. S1). rAAV6 (Additional file 1: Fig. S1a, b) was transfected into 293 T cells (Additional file 1: Fig. S2a) to enhance A20 expression (Additional file 1: Fig. S2b). A 10 μl of volume rAAV6-CMV-A20 containing 1 × 10¹² viral genomes (vg) was injected into the left knee, ankle, and tarsal area of CIA mice, while the same dose of rAAV6-CMV-EGFP was injected into the right side (Additional file 1: Fig. S2c). rAAV6 was widely distributed in various cell types (Additional file 1: Fig. S2d) and significantly enhanced A20 expression until 5 weeks after injection (Fig. 1a). Notably, rAAV6-CMV-A20 decreased the clinical arthritis score, paw thickness and total porosity (P < 0.001), increased the bone volume-to-tissue volume ratio (BV/TV, P < 0.001), trabecular number (Tb.N, P = 0.008), and trabecular thickness (Tb.Th, P = 0.002) (Fig. 1b–d, Additional file 1: Fig. S2e–j), and suppressed pannus formation, bone erosion, and cartilage destruction (Fig. 1e). We also found that rAAV6-CMV-A20 significantly suppressed the expression of NLRP3, caspase-1, and IL-1β (Additional file 1: Fig. S2k–m). Thus, the results verified our hypothesis that rAAV6-mediated A20 overexpression is an effective RA therapy.

Fig. 1

rAAV6-CMV-A20 inhibits inflammation in CIA mice. a Immunofluorescence staining of A20 in joints that were injected with rAAV6-CMV-A20 (white arrow: synovium). b Clinical arthritis score of hind paws injected with rAAV6-CMV-EGFP or rAAV6-CMV-A20 (n = 15); Statistical analyses were performed between hind paws injected with rAAV6-CMV-EGFP and rAAV6-CMV-A20 using paired t-tests. c Hind paws of CIA mice injected with rAAV6-CMV-EGFP (right) or rAAV6-CMV-A20 (left) on the day 42nd after primary immunization. d Micro-CT 3D reconstruction image of hind paws. e HE or Safranin-O staining of joint specimens injected with rAAV6-CMV-EGFP or rAAV6-CMV-A20. f Hind paws of CIA mice injected with rAAV6-SP146-EGFP (right) or rAAV6-SP146-A20 (left) on the day 42nd after primary immunization. g Clinical arthritis score of hind paws injected with rAAV6-SP146-EGFP or rAAV6-SP146-A20 (n = 9); Statistical analyses were performed between hind paws injected with rAAV6-SP146-EGFP and rAAV6-SP146-A20 using paired t-tests. HE (h) and Safranin-O (i) staining of joint specimens injected with rAAV6-SP146-EGFP or rAAV6-SP146-A20. j Fluorescence image of RAW264.7 cells transfected with rAAV6-SP146-EGFP. k Western blotting of A20 in RAW264.7 cell lysates following transfection with rAAV6-CMV-A20 or rAAV6-SP146-A20. l Immunofluorescence staining of A20 and F4/80 in joint specimens injected with rAAV6-SP146-A20. CMV cytomegalovirus, NLRP3 NOD-like receptor protein 3

The CMV promoter may attract safety concerns. Therefore, considering that macrophages and fibroblast-like synoviocytes account for the primary components of synovial tissues, especially in RA, we tested which cell type predominantly overexpressed A20. We constructed rAAV6-COL1α-A20 (Additional file 1: Fig. S1c) targeting fibroblast-like synoviocytes and rAAV6-SP146-A20 (Additional file 1: Fig. S1e) targeting macrophages [5]. A 10 μl volume of rAAV6-COL1α-A20 or rAAV6-SP146-A20 (1 × 10¹² vg) was injected into the left knee, ankle, and tarsal area, while the same dose of rAAV6-COL1α-EGFP (Additional file 1: Fig. S1d) or rAAV6-SP146-EGFP (Additional file 1: Fig. S1f) was injected into the same areas on the right side (Additional file 1: Fig. S3a). In contrast to rAAV6-COL1α-EGFP, rAAV6-COL1α-A20 exhibited nearly no effect on suppressing RA symptoms. However, rAAV6-SP146-A20 showed anti-rheumatic therapeutic effects compared with rAAV6-SP146-EGFP. Similar to rAAV6-CMV-A20, rAAV6-SP146-A20 reduced the clinical arthritis score, paw thickness (Fig. 1f, g; Additional file 1: Fig. S3b, c), total porosity (Additional file 1: Fig. S3d), bone erosion activity (Additional file 1: Fig. S3e–g), pannus formation, and cartilage destruction (Fig. 1h, i). rAAV6-SP146-EGFP was well transfected into RAW264.7 cells (Fig. 1j) and exhibited dominant distribution in macrophages (Additional file 1: Fig. S3h). rAAV6-SP146-A20 significantly enhanced A20 expression (Fig. 1k, l) and the collective results further indicated the therapeutic benefits of rAAV6-mediated A20 overexpression. Importantly, macrophages were found to be responsible for the rAAV6-mediated A20 overexpression.

The anti-rheumatic benefits of rAAV6-SP146-A20 were further verified by simultaneous injection of rAAV. CIA mice were injected with rAAV6-SP146-EGFP or rAAV6-SP146-A20 on both sides (1 × 10¹² vg, Additional file 1: Fig. S3i). As expected, rAAV6-SP146-A20 significantly relieved the arthritis symptoms (Additional file 1: Fig. S3j–o, Fig. S4a–c). Furthermore, we tested whether the therapeutic effect was dependent on viral genomes. CIA mice were injected with rAAV6-SP146-EGFP at a dose of 1 × 10¹² vg (EGFP) or rAAV6-SP146-A20 at a dose of 1 × 10⁸ vg (E8), 1 × 10¹⁰ vg (E10), or 1 × 10¹² vg (E12, Additional file 1: Fig. S3p). The results revealed that the protective effect of rAAV6-SP146-A20 was dose-dependent. The E12 group exhibited the lowest clinical arthritis score and increases in paw thickness, bone erosion activity, pannus formation, and cartilage destruction (Additional file 1: Figs. S3q–v, S4d–e). As the treatment dose increased, so did the expression of A20, while NLRP3, caspase-1, and IL-1β were gradually inhibited (Additional file 1: Fig. S4f–g).

Our hypothesis that A20 overexpression is a reasonable strategy for the treatment of RA was initially evidenced. We not only demonstrated the therapeutic effects against RA of ubiquitous CMV promoter-driven delivery of A20, but more importantly, we identified the key role of macrophage-like synovial cells in responding to rAAV6-mediated gene delivery of A20.