Gene Therapy Induces Antigen-Specific Tolerance in Experimental Collagen-Induced Arthritis

Abstract

Here, we investigate induction of immunological tolerance by lentiviral based gene therapy in a mouse model of rheumatoid arthritis, collagen II-induced arthritis (CIA). Targeting the expression of the collagen type II (CII) to antigen presenting cells (APCs) induced antigen-specific tolerance, where only 5% of the mice developed arthritis as compared with 95% of the control mice. In the CII-tolerized mice, the proportion of Tregs as well as mRNA expression of SOCS1 (suppressors of cytokine signaling 1) increased at day 3 after CII immunization. Transfer of B cells or non-B cell APC, as well as T cells, from tolerized to naïve mice all mediated a certain degree of tolerance. Thus, sustainable tolerance is established very early during the course of arthritis and is mediated by both B and non-B cells as APCs. This novel approach for inducing tolerance to disease specific antigens can be used for studying tolerance mechanisms, not only in CIA but also in other autoimmune diseases.

Fig 1

Lentiviral constructs and model design (A) LNT-Ctrl; lentiviral control construct that encodes the MHC class II–associated invariant chain peptide (CLIP) stabilizing Invariant chain (Ii) driven by the spleen focus-forming virus (SFFV) promoter. LNT-CII; lentiviral construct that encodes the rat collagen epitope (aa259-270, GIAGFKGEQGPK) fused into Ii. LTR; long term repeat, cPPT; central polypurine tract, WPRE; Woodchuck Post Transcriptional Element. (B) Cartoon describing the design of the experiment. Hematopoietic stem cells (HSCs) from DBA/1 mice were transduced with lentiviral particles injected into lethally irradiated recipient mice. CIA was induced according to standard protocol at least 10 weeks after HSC transplantation. (C) Naïve splenocytes were stimulated with naked, hydroxylated or glycosylated peptides and co-cultured with T cell hybridomas specific to each post-translational modification of the CII-peptide, i.e. Hcq4 recognizes the naked peptide, HdBr1 the hydroxylated and Hcq3 the glycosylated CII-peptide. The response was measured as IL-2 production by ELISA and performed three times. (D) IL-2 production in response to co-culture of the T cell hybridomas with bone marrow cells, peritoneal macrophages or spleen cells from HSC transplanted mice (n = 6/group) and performed twice. (E) Co-cultures of the T cell hybridomas with B cells (CD19⁺) or non-B cells (CD19^-) from LNT-Ctrl and LNT-CII HSC transplanted mice (n = 3/group) and performed twice. Statistical analysis was performed by Student’s t-test, mean±SEM.

Fig 2. Development of arthritis and B cell responses in the novel model of tolerance to collagen-induced arthritis.

Severity (A) as well as incidence of arthritis (B) were determined by macroscopical examination (LNT-Ctrl n = 25 (day 0–42), n = 17 (day 43–49) LNT-CII n = 24 (day 0–42), n = 11 (day 43–49)). The experiments have been repeated five times and the graph shows the pooled results from three experiments (C) Histopathological examination of synovitis and bone/cartilage erosivity at termination of two pooled experiments (days 42–49). (D) Histological photos from an LNT-Ctrl and an LNT-CII mouse at day 42 after CII immunization, B = bone, C = cartilage, S = synovium. Scale bar 100 μm. (E) Development of CII specific IgG antibodies measured by ELISA in serum at indicated time points during CIA (n = 3-6/group). (F) Serum levels of CII-specific IgG and subclasses IgG1, IgG2a and IgG2b at days 42–49. Experiments in Fig E and F have been performed at least twice. Statistical analysis was performed by linear regression in (A, D), logistic regression in (B) and Mann-Whitney U-test in (C and E), mean±SEM.

Fig 3. Adoptive transfer of T cells, B cells and APCs.

(A) Design of experiment. Adoptive transfer of sorted T cells, B cells and APCs from LNT-Ctrl (n = 5) and LNT-CII (n = 6) mice respectively were transferred into naive recipients 2 days before CII immunization. (B) Severity of arthritis after immunization in mice receiving T cells (LNT-Ctrl n = 6, LNT-CII n = 6). (C) B cells (LNT-Ctrl n = 7, LNT-CII n = 7) or (D) APCs (LNT-Ctrl n = 6, LNT-CII n = 4). (E) Histopathological examination of synovitis and bone/cartilage erosivity. The experiments have been performed once. Statistical analysis was performed using linear regression for comparing development of arthritis, Mann-Whitney U-test for comparing histopathological score, mean±SEM.

Fig 4. Involvement of T cells in tolerance induction.

(A) Capacity of Tregs cells to down-regulate effector T cell responses. Sorted Tregs cells from LNT-Ctrl or LNT-CII mice were co-cultured with LNT-Ctrl T effector cells in ratios 1:1, 1:5 and 1:10 for three days and proliferation measured by radioactive thymidine incorporation (counts per minutes). Y-axis—percentage of proliferation compared to the negative control containing sorted T effector cells only (n = 6/group). The experiment has been repeated three times. (B) CII-specific IFN-γ production from stimulated spleen cells obtained 14 days after CII immunization, n = 4–5 and repeated once. Statistical analysis was performed using Two-way ANOVA for evaluation of proliferative responses in (A) and Mann-Whitney U-test in (B), mean±SEM.

Fig 5. qPCR array and SOCS1 association with LNT-Ctrl vs LNT-CII at day 3 after CII immunization.

(A) OPLS-DA scatter dot plot showing the separation of gene expression in tolerized or non-tolerized mice and (B) OPLS-DA column loading plot that depicts the association between LNT-CII and LNT-Ctrl mice with the expression of different genes. X-variables represented with a positive bar are positively associated with LNT-CII mice, whereas variables in the opposite direction are inversely related to this group of mice. The larger the bar and smaller the error bar, the stronger and more certain is the contribution to the model. The final OPLS-DA loading plots are based on parameters with VIP values ≥1.3. R2Y indicates how well the variation of Y is explained, whereas Q2 indicates how well Y can be predicted. Univariate analysis (Student’s t-test) of (C) SOCS1 (D) of IL-10 (E) IFN-γ (F) IL-6, n = 2–4 animals/group, mean ± SEM. Actb and Gadph were selected as housekeeping genes and relative quantification was calculated from the sample of a naïve DBA/1 mouse.