Vasoactive intestinal peptide induces CD4+,CD25+ T regulatory cells with therapeutic effect in collagen-induced arthritis

Authors


Abstract

Objective

CD4+,CD25+ T regulatory cells (Treg) control the immune response to a variety of antigens, including self antigens, and may offer opportunities to intervene in the course of autoimmune diseases. Several models support the idea of the peripheral generation of CD4+,CD25+ Treg from CD4+,CD25− T cells, but little is known about the endogenous factors and mechanisms controlling the peripheral expansion of CD4+,CD25+ Treg. We undertook this study to investigate the capacity of the vasoactive intestinal peptide (VIP), an immunosuppressive antiarthritic neuropeptide, to induce functional Treg in vivo during the development of collagen-induced arthritis (CIA).

Methods

We measured the number of CD4+,CD25+ Treg following VIP administration to CIA mice, and we characterized their phenotype and their ability to suppress activation of autoreactive T cells. We determined the capacity of VIP to induce Treg in vitro as well as the use of Treg in the treatment of CIA, measuring the clinical evolution and the inflammatory and autoimmune components of the disease.

Results

The administration of VIP to arthritic mice resulted in the expansion of CD4+,CD25+,Foxp3+ Treg in the periphery and joints, which inhibited autoreactive T cell activation/expansion. VIP induced more efficient suppressors on a per-cell basis. The VIP-generated CD4+,CD25+ Treg transfer suppressed and significantly ameliorated the progression of the disease.

Conclusion

These results demonstrate the involvement of the generation of Treg in the therapeutic effect of VIP on CIA. The generation of highly efficient Treg by VIP ex vivo could be used as an attractive therapeutic tool in the future, avoiding the administration of the peptide to patients with rheumatoid arthritis.

The induction of antigen-specific tolerance is essential to maintain immune homeostasis, to control autoreactive T cells, and to prevent the onset of autoimmune diseases. To a large degree, thymic selection prevents the release of functional autoreactive T cells. However, potential autoreactive T cells persist in the periphery of healthy individuals and retain the capacity to initiate autoimmune disease. Thus, peripheral regulatory mechanisms are required to protect against self-directed immune responses. Recently, the crucial role of CD4+,CD25+ T regulatory cells (Treg) in the control of self-reactive T cells and in the induction of peripheral tolerance has been demonstrated in both mice and humans (1, 2). Failures in the function of the Treg compartment can therefore be responsible for the development of autoimmune diseases, and enhancing its function may represent a treatment strategy.

Two major types of Treg have been characterized in the CD4+ population (i.e., the naturally occurring, thymus-generated Treg and the peripherally induced, interleukin-10 [IL-10]– or transforming growth factor β [TGFβ]–secreting Treg [1–3]). The CD4+,CD25+,Foxp3+ Treg generated in thymus migrate and are maintained in the periphery. The number of CD4+,CD25+ Treg in the periphery does not decrease with age, although these cells are anergic and prone to apoptosis, and their site of origin, the thymus, undergoes age-related involution. This suggests that CD4+,CD25+ Treg can be generated peripherally. Indeed, several experimental models support the idea of peripheral generation of CD4+,CD25+ Treg from CD4+,CD25− T cells (4). The endogenous factors and mechanisms controlling the peripheral expansion of CD4+,CD25+ Treg are mostly unknown.

Vasoactive intestinal peptide (VIP) is a potent Th2-produced immunosuppressive agent that has proved to be protective in several models of autoimmune diseases such as collagen-induced arthritis (CIA), inflammatory bowel disease, uveoretinitis (5–7), and experimental autoimmune encephalomyelitis (8). Until now, the mechanisms described for VIP immunosuppressive activity included the deactivation of macrophages, dendritic cells, and microglia and the promotion of Th2 effector differentiation and survival (9). Because CD4+,CD25+ Treg have been found to play a pivotal role in the regulation of rheumatoid arthritis (RA) and CIA (10–13), in the present study we investigated whether VIP might exert its protective effect against CIA by increasing the generation/activation of the Treg compartment, and we also investigated whether VIP-induced Treg could be used as a cell-based therapy for CIA.

MATERIALS AND METHODS

Animals and peptides.

Eight-week-old male DBA/1 mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Bovine type II collagen (CII) was purchased from Sigma (St. Louis, MO). VIP was purchased from Calbiochem (Laufelfingen, Switzerland).

Arthritis induction and treatment.

To induce CIA, DBA/1 mice were injected subcutaneously (SC) in the base of the tail with 200 μg of CII emulsified in Freund's complete adjuvant (CFA) containing 200 μg of Mycobacterium tuberculosis H37Ra (Difco, Detroit, MI). On day 21 after primary immunization, mice were boosted SC with 100 μg of CII in CFA. VIP therapy consisted of the intraperitoneal (IP) administration of 2 nmoles (6.6 μg/mouse/day) on 5 consecutive days after the secondary immunization. Controls were injected IP with medium. For adoptively transferred CIA, draining lymph node (DLN) cells (5 × 106) were purified 14 days after immunization and stimulated in vitro with 20 μg/ml of CII for 72 hours. Whole LN cells or isolated CD4+, CD4+,CD25+, or CD4+,CD25− T cells were then injected intravenously (IV) into naive DBA/1 mice. In some experiments, VIP-treated CIA mice received IV injections of anti–CTLA-4 monoclonal antibody (mAb), neutralizing anti–IL-4 polyclonal antibody, neutralizing anti–IL-10 polyclonal antibody, neutralizing anti-TGFβ mAb, or preimmune rat IgG used as control Ig (500 μg antibody/mouse) on alternate days up to 10 days after secondary immunization.

Mice were evaluated by 2 independent, blinded examiners every other day and monitored for signs of arthritis onset using 2 clinical parameters, paw swelling and clinical score. Paw swelling was assessed by measuring the thickness of the affected hind paws with 0–10-mm calipers. Clinical arthritis was assessed by using the following system: grade 0 = no swelling; grade 1 = slight swelling and erythema; grade 2 = moderate swelling and edema; grade 3 = extreme swelling and pronounced edema; grade 4 = joint rigidity. Each limb was graded, giving a maximum possible score of 16 per animal. Animal experimental protocols were reviewed and approved by the Ethical Committee of the Spanish Council of Scientific Research.

For CD4+,CD25+ T cell depletion, DBA/1 mice were treated IV with 1 mg anti-CD25 antibody (clone PC61) 3 days before immunization with CII. CD4+,CD25+ T cell depletion was >96% in the spleen at 72 hours and at 7 days, as determined by flow cytometry.

Tissue collection and cell isolation.

At various time points after immunization, spleen, DLN, and synovial membrane were removed. Single-cell suspensions were obtained from spleen and DLN. Synovial cells were isolated by enzymatic digestion of the synovium of the knee joints as previously described (5). For the isolation of different T cell populations (CD4+, CD4+,CD25+, CD4+,CD25−), spleen, DLN, and synovial cells were labeled with phycoerythrin (PE)–conjugated anti-CD25 and peridin chlorophyll protein (PerCP)–conjugated anti-CD4 antibodies as described below, and the different populations were gated and sorted using a FACSCalibur flow cytometer (Becton Dickinson, San Diego, CA). Antigen-presenting cells (APCs) were prepared by immunomagnetic T cell depletion of DBA/1 spleen cells using microbead-conjugated anti-CD8 and anti-CD4 mAb (Miltenyi Biotec, Bergisch Gladbach, Germany), followed by treatment with 50 μg/ml mitomycin C (Sigma).

Cell cultures and assessment of T cell autoreactive response.

DLN cells were recovered from the DBA/1 mice at the peak of clinical CIA (day 35 postimmunization). Cells (106/ml) were stimulated in complete medium (RPMI 1640 containing 10% fetal calf serum, 2 mML-glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin) with different concentrations of CII for 48 hours (for cytokine determination) or for 72 hours (for proliferative response). Cell proliferation was evaluated by using a cell proliferation assay (bromodeoxyuridine) from Roche Diagnostics (Mannheim, Germany). Cytokine content in culture supernatants was determined by specific sandwich enzyme-linked immunosorbent assays (ELISAs) (5).

To determine the suppressive capacity of regulatory CD4+ cells, autoreactive CD4+ cells (4 × 105/well) isolated from DLN of CIA mice were stimulated with spleen APCs (105/well) and CII (20 μg/ml) in the absence or presence of DLN CD4+ cells from untreated or VIP-treated CIA mice (5 × 104/well). Some cultures were performed in the presence of blocking anti–IL-10 (10 μg/ml), anti-TGFβ1 (40 μg/ml), and/or anti–CTLA-4 (10 μg/ml) mAb, or in the presence of control IgG mAb (30 μg/ml), or of IL-2 (100 units/ml). To determine the cell-contact dependence of the regulatory response, we placed autoreactive CD4+ cells isolated from DLN of CIA mice (responder CD4+ [rCD4] cells) (5 × 104) with spleen APCs (105) and CII (20 μg/ml) in the bottom well of a transwell system, and we placed DLN regulatory CD4+ cells (2 × 104) isolated from VIP-treated CIA mice with spleen APCs (105) and CII (20 μg/ml) in the upper transwell chambers. After 72 hours, we measured the proliferative response of the autoreactive CD4+ cells from the bottom well. To determine the in vitro generation of CD4+,CD25+ cells, CD4+ or CD4+,CD25− cells (0.5 × 106/ml) isolated from spleen and DLN of CIA mice at the peak of the disease were stimulated with anti-CD3 (5 μg/ml) plus anti-CD28 (1 μg/ml) antibodies or with CII (20 μg/ml) plus APCs (105/ml) in the absence or presence of VIP (10−8M) for different periods of time, and the percentage of CD4+,CD25+,CTLA-4+ cells and the expression of Foxp3 messenger RNA (mRNA) were determined as described below.

Flow cytometric analysis.

Synovial and DLN cells incubated with various mAb (PE–conjugated anti-CD25, fluorescein isothiocyanate [FITC]–conjugated anti-CD62L, FITC-conjugated anti-CD69, FITC-conjugated anti–glucocorticoid-induced tumor necrosis factor [TNF] receptor [anti-GITR], FITC-conjugated anti-CD45RB, PerCP-conjugated anti-CD4), (2.5 μg/ml final concentration) were fixed in 1% paraformaldehyde and analyzed on a FACSCalibur flow cytometer. We used isotype-matched antibodies as controls, and IgG block (Sigma) to avoid nonspecific binding to Fc receptors. For analysis of intracellular CTLA-4, synovial and DLN cells were stained with PerCP-conjugated anti-CD4 and FITC-conjugated anti-CD25 mAb, fixed with Cytofix/Cytoperm solution (BD PharMingen, San Diego, CA), incubated with PE-conjugated anti–CTLA-4 mAb diluted in 0.5% saponin, and analyzed by flow cytometry.

Messenger RNA analysis.

Total RNA was isolated from CD4+ T cells (106) or from homogenized joints using Ultraspec RNA reagent (Biotecx, Houston, TX). Real-time polymerase chain reaction (PCR) was used to determine Foxp3 and neuropilin-1 (Nrp1) mRNA expression in isolated DLN CD4+ cells. Two micrograms of total RNA was reverse transcribed with oligo(dT) primers and Moloney murine leukemia virus reverse transcriptase polymerase (Invitrogen, Carlsbad, CA). Quantitative real-time reverse transcriptase–PCR was performed in an ABI PRISM cycler (Applied Biosystems, Foster City, CA) using a SYBR Green PCR kit (Applied Biosystems). A threshold was set in the linear part of the amplification curve, and the number of cycles needed to reach the threshold was calculated for each gene. Relative mRNA levels were determined by using standard curves for each individual gene and further normalization to hypoxanthine guanine phosphoribosyltransferase (HPRT). Melting curves established the purity of the amplified band. Primer sequences were as follows: for Nrp1, 5′-GCCTGCTTTCTTCTCTTGGTTTCA-3′ (forward) and 5′-GCTCTGGGCACTGGGCTACA-3′ (reverse); for Foxp3, 5′-CTGGCGAAGGGCTCGGTAGTCCT-3′ (forward) and 5′-CTCCCAGAGCCCATGGCAGAAGT-3′ (reverse); for HPRT, 5′-TGGAAAGAATGTCTTGATTGTTGAA-3′ (forward) and 5′-AGCTTGCAACCTTAACCATTTTG-3′ (reverse). Cytokine mRNA expression was determined from 5 μg RNA isolated from joints by RNase protection assay using the Riboquant MultiProbe RNase Protection Assay System according to the instructions of the manufacturer (BD PharMingen).

Tracing adoptively transferred cells.

To trace CII-specific autoreactive T cells in vivo, spleen and DLN cells recovered on day 28 postimmunization from immunized DBA/1 mice were stimulated in vitro with 20 μg/ml CII for 72 hours, then CD4+ T cells were isolated as described above and incubated (2 × 107/ml) with 10 mM 5-carboxyfluorescein diacetate succinimidyl ester (CFSE; Molecular Probes, Eugene, OR) at 37°C for 20 minutes. CFSE-labeled cells were injected IP (2 × 107/mouse) into naive DBA/1 mice. Synovial and spleen cells were isolated 10 days posttransfer, and the presence of CFSE-labeled CD4+ cells was determined by flow cytometry. Mitotic events were determined as described (14).

Measurement of serum anti-CII antibody levels.

Serum samples were collected at the peak of the disease (day 35 postimmunization), and the levels of anti-CII IgG and IgG2a were measured by ELISA as described (5).

Statistical analysis.

The nonparametric Mann-Whitney U test was used to determine the statistical significance of clinical and cell culture results.

RESULTS

Induction by VIP of emergence of CD4+,CD25+ Treg during CIA.

As previously described (5), VIP administration prevented the progression of CIA (Figure 1A). The therapeutic effect of VIP on CIA was previously associated with a striking reduction of the two deleterious components of the disease (i.e., the autoimmune and inflammatory responses). VIP decreased the presence of autoreactive Th1 cells in the joint and in the DLN (5). In addition, VIP strongly reduced the inflammatory response during CIA progression by down-regulating the production of several inflammation mediators in the joints (5). Several studies have indicated that Treg confer significant protection against CIA by decreasing the activation and joint homing of autoreactive Th1 cells (3, 12, 15). Because the VIP also inhibited events in the inflammatory phase of CIA following the activation of antigen-specific CD4+ Th1 cells, the possibility exists that VIP induces Treg with suppressive activity during the progression of the disease. In this regard, in contrast to CD4+ T cells isolated from untreated animals, CD4+ T cells from VIP-treated CIA mice did not transfer the disease to naive mice (Figure 1B). However, when these cells were depleted of CD4+,CD25+ cells prior to transfer, they were able to transfer the disease (Figure 1B). This suggests that VIP might induce the generation and/or activation of CD4+,CD25+ Treg in vivo. Therefore, we investigated whether VIP induces Treg during CIA.

Figure 1.

Induction by vasoactive intestinal peptide (VIP) of generation of CD4+,CD25+ T regulatory cells (Treg) in mice with collagen-induced arthritis (CIA). A and B, CD4+ T cells from VIP-treated CIA mice do not transfer disease. A, CIA mice were treated with medium (control) or with VIP (2 nmoles/day) for 5 days, and clinical score was measured daily. Values are the mean ± SD. Numbers in parentheses represent percentage of disease incidence in each group (10 mice per group). B, Draining lymph node (DLN) cells isolated from CIA mice treated with medium (control) or with VIP (2 nmoles/day) were activated in vitro with type II collagen and sorted into CD4+ or CD4+,CD25− cell populations. Cells (2 × 107) were then transferred into naive DBA/1 mice. Clinical score was measured in recipient mice. Values are the mean ± SD. Numbers in parentheses represent percentage of disease incidence in each group (10–12 mice per group). C and D, VIP induces the emergence of CD4+,CD25+ Treg during CIA. C, DLN and synovial (joint) cells isolated from untreated (control) or VIP-treated CIA mice at the peak of disease were stained for CD4 and CD25 and different markers (CD62L, CD69, CD45RB, or CTLA-4). Percentages and numbers of CD4+,CD25− and CD4+,CD25+ cells in DLN and joints were determined by flow cytometry analysis. Gated CD4+,CD25+ and CD4+,CD25− populations were additionally analyzed for CD62L, CD69, CD45RB, or CTLA-4 expression (histograms). Similar profiles were observed for synovial cells. Black bars = CD4+,CD25+ T cells; white bars = CD4+,CD25− T cells. D, Foxp3, neuropilin-1 (Nrp1), and glucocorticoid-induced tumor necrosis factor receptor (GITR) expression was analyzed in DLN CD4+ cells isolated from untreated (control) or VIP-treated arthritic mice by real-time polymerase chain reaction (Foxp3 and Nrp1) or by flow cytometry (GITR). Values in C and D are the mean and SD of 3 experiments performed in duplicate. ∗ = P < 0.001 versus control.

VIP-treated CIA mice had significantly higher percentages and numbers of CD4+,CD25+ cells in both DLN and synovium compared with control CIA mice (Figure 1C). VIP-induced CD4+,CD25+ cells exhibited an activated Treg phenotype (16–18) (i.e., CD45RBlow,CD62Lhigh,CD69high,CTLA-4high) (Figure 1C). CD4+,CD25+ cells isolated from VIP-treated mice also expressed higher levels of CTLA-4, a key player in Treg function (Figure 1C). Although Treg constitutively express CD25 and CTLA-4, these receptors are also expressed on activated effector T cells. Several other markers have been recently identified in Treg (e.g., Nrp1, the transcription factor Foxp3, and GITR) (16–19). We found that CD4+ cells isolated from DLN of VIP-treated CIA mice expressed higher levels of Foxp3, Nrp1, and GITR than those isolated from DLN of control CIA mice (Figure 1D). These data suggest that VIP promotes the generation of activated Treg during CIA.

When stimulated, Treg suppress the proliferation and IL-2 production of antigen-specific effector T cells. Several mechanisms have been identified for Treg function, such as surface CTLA-4 and TGFβ expression, costimulatory blockade, and release of IL-10 and/or TGFβ (1–4). Naturally occurring CD4+,CD25+ Treg exert their suppressive activity primarily through direct cellular contact, whereas peripheral T cell suppressors act primarily through cytokines (1–3). To determine whether CD4+ T cells isolated from VIP-treated CIA mice function as suppressive Treg, we cocultured CD4+ cells isolated from DLN of arthritic mice treated with medium (CD4control cells) or VIP (CD4VIP cells) with rCD4 cells in the presence of APCs and antigen (CII). CD4control cells did not suppress the proliferation of rCD4 cells, and they slightly up-regulated IL-2 and interferon-γ (IFNγ) production in response to antigen (Figure 2A). In contrast, CD4VIP cells suppressed the proliferation of autoreactive rCD4 cells (Figure 2A). The suppression increased with the number of CD4VIP cells, being effective even at a CD4VIP cell:rCD4 cell ratio as low as 1:8 (Figure 2B). CD4VIP cells also inhibited IL-2 and IFNγ production, while increasing the levels of the regulatory cytokines IL-10 and TGFβ (Figure 2A).

Figure 2.

Induction by VIP of functional Treg in CIA mice. A, Autoreactive CD4+ cells isolated from CIA mice were stimulated with spleen antigen-presenting cells (APCs) and type II collagen (CII) in the absence (none) or presence of DLN CD4+ T cells isolated from untreated (CD4 control) or VIP-treated (CD4 VIP) arthritic mice and tested for cell proliferation and cytokine production. As controls, cells cultured with an irrelevant antigen (ovalbumin) did not produce cytokines. ∗ = P < 0.001 versus CD4 control. B, Autoreactive CD4+ cells (4 × 105) isolated from CIA mice were mixed with increasing numbers of DLN CD4+ cells (regulatory CD4) from untreated (control) or VIP-treated CIA mice (ratios from 1:64 to 1:1) and stimulated with CII and splenic APCs. The proliferative response was measured. ∗ = P < 0.001 versus CD4 control. C, Coculture experiments were performed as in B in the absence (none) or presence of blocking anti–interleukin-10 (αIL-10), anti–transforming growth factor β (αTGF), and/or anti–CTLA-4 (αCTLA-4) monoclonal antibodies (mAb), or in the presence of recombinant murine IL-2. Addition of a control IgG mAb did not show any blocking effect on the suppressive activity of the regulatory CD4 cells. In addition, cell contact between autoreactive and regulatory CD4 cells was prevented using a transwell system. ∗ = P < 0.001 versus CD4 VIP samples incubated without any blocking antibody. Each result is the mean and SD of 3 experiments performed in duplicate. IFNγ = interferon-γ (see Figure 1 for other definitions).

We addressed the question of whether CD4VIP cells inhibited autoreactive T cells through direct cellular contact and/or soluble factors. When CD4VIP cells and autoreactive CD4+ cells were separated in transwell experiments, the suppressive activity was partially abolished (Figure 2C), indicating that both direct contact and soluble factors mediate the inhibitory effect. In regular cocultures, addition of anti-TGFβ, anti–IL-10, or anti–CTLA-4 antibodies modestly reversed CD4VIP cell–mediated inhibition. However, blocking both IL-10 and TGFβ had a more pronounced effect, and addition of all 3 antibodies (anti–IL-10, anti-TGFβ, anti–CTLA-4) reversed the inhibitory effect completely (Figure 2C). In addition, as previously described for Treg, exogenous IL-2 overcame the suppressive activity (Figure 2C). These results demonstrate that VIP administration during arthritis development induces the generation and/or activation of Treg that efficiently suppress autoreactive CD4+ T cells.

Induction by VIP of generation of peripheral CD4+,CD25+ Treg from CD4+,CD25− T cells.

CD4+,CD25+ Treg can be generated peripherally from CD4+,CD25− T cells (4). To determine whether the VIP-induced increase in CD4+,CD25+ Treg during CIA was due to the expansion of the existing naturally occurring CD4+,CD25+ Treg or to newly generated Treg from CD4+,CD25− T cells, we depleted the CIA mice of CD4+,CD25+ T cells by treatment with anti-CD25 antibody before VIP inoculation. As previously described (13), depletion of CD25+ T cells prior to CIA induction resulted in a more severe disease than in controls, with an earlier onset and higher clinical scores (Figure 3A). In contrast, CD25+ depletion did not affect the beneficial effect of VIP (Figure 3A). At the time of maximum clinical score in the control group (day 35), the CIA mice depleted of CD25+ cells and treated with VIP possessed almost the same number of splenic CD4+,CD25+ T cells as the VIP-treated nondepleted mice (10% and 12%, respectively) (Figure 3A, right panel). This contrasted with findings in the control arthritic mice not treated with VIP (2.5% CD4+,CD25+ cells).

Figure 3.

Induction by VIP of generation of peripheral CD4+,CD25+ Treg from CD4+,CD25− T cells. A, Left, DBA/1 mice were injected with anti-CD25 antibodies (αCD25 Ab) to eliminate CD25+ T cells. Ten days later, the mice were immunized with type II collagen (CII). Treatment with VIP was started on day 22 postimmunization. Values are the mean of 6 mice per group. Right, Percentage of CD4+,CD25+ T cells in spleen on day 28. Values are the mean and SD of 6 mice per group. ∗ = P < 0.001 versus control. B, DLN CD4+,CD25− cells isolated from CIA mice at the peak of the disease were cultured with medium (none) or stimulated with anti-CD3 (αCD3; 5 μg/ml) plus anti-CD28 (αCD28; 1 μg/ml) or with CII (20 μg/ml) in the absence or presence of VIP (10−8M). After different periods of time, the percentages of CD4+,CD25+ cells were determined by flow cytometry. The expression of Foxp3 mRNA was determined by real-time polymerase chain reaction after 2 days of culture. The expression of CTLA-4 in gated CD4+,CD25+ cells was analyzed by flow cytometry after 4 days of culture. Numbers shown in each quadrant represent percentages of positive cells. Results are the mean and SD of 4 experiments performed in duplicate. ∗ = P < 0.001 versus samples treated with anti–CD3/CD28 or CII alone. See Figure 1 for other definitions.

These experiments suggest that VIP could induce the peripheral generation of CD4+,CD25+,Foxp3+ Treg from the CD4+,CD25− compartment. To further confirm this hypothesis, CD4+,CD25− cells isolated from DLN of CIA mice at the peak of the disease were stimulated in vitro with anti–CD3/CD28 or with CII in the absence or presence of VIP. The incubation with VIP significantly increased the percentage of CD4+,CD25+ cells and the levels of Foxp3 in the cultures (Figure 3B). VIP-induced CD4+,CD25+ Treg expressed higher amounts of CTLA-4 in comparison with control CD4+,CD25+ cells stimulated with antigen alone (Figure 3B). Interestingly, VIP did not generate CD4+,CD25+ Treg in the absence of any stimulation (data not shown).

Involvement of Treg in the therapeutic effect of VIP in CIA.

Because VIP seems to induce in vivo generation of IL-10/TGFβ–producing CD4+,CD25+,CTLA-4+ Treg in CIA mice, we further examined the role of these cells in the therapeutic effect of VIP in arthritis. In vivo blockade experiments showed that treatment with antibodies against CTLA-4, IL-10, or TGFβ, major mediators of Treg function, significantly decreased the therapeutic effect of VIP, and treatment with all 3 antibodies partially reversed the effect of VIP in arthritis (Figure 4A), suggesting the involvement of Treg. However, in vivo blockade of the Th2 cytokine IL-4 did not significantly affect the beneficial effect of VIP in CIA (Figure 4A).

Figure 4.

Involvement of Treg in the therapeutic effect of VIP in CIA. A, CIA mice were treated with medium (control) or with VIP (2 nmoles/day) for 5 days and with control Ig, anti–interleukin-4 (αIL-4), anti–IL-10 (αIL-10), anti–transforming growth factor β (αTGF), or anti–CTLA-4 (αCTLA-4) monoclonal antibodies (mAb), or a combination of the last 3 mAb, on alternate days starting with the first VIP administration (10 mice per group). Arthritis clinical score was measured 15 days after initiation of treatment. ∗ = P < 0.001 versus VIP-treated CIA animals without any blocking antibodies. Data represent the mean and SD of 5 mice per group. B, DLN cells from CIA mice were stimulated in vitro with type II collagen (CII) as described in Materials and Methods, followed by isolation of CD4+ cells and labeling with 5-carboxyfluorescein diacetate succinimidyl ester (CFSE). The CFSE-labeled CD4+ cells (2 × 107) were transferred into naive hosts together with DLN CD4+ T cells (5 × 106) from untreated (control) or VIP-treated CIA mice. The data represent the mean clinical score for each group. Numbers in parentheses represent percentage of disease incidence (n = 8 mice per group). C, The relative presence, proliferative state, and number of mitotic events of the autoreactive CFSE-stained CD4+ cells recovered from the spleen and synovium of recipient mice, following cotransfer with DLN CD4+ cells from untreated (control) or VIP-treated CIA mice, were determined on day 10 after transfer. Data represent the mean and SD of 5 mice per group. ∗ = P < 0.001 versus control. D, The CFSE-labeled CD4+ T cells (2 × 107) were cotransferred into naive hosts with DLN CD4+ cells from untreated CIA mice (control; 5 × 106 cells) or with unfractionated (total LN), CD4−, or CD4+,CD25− cells (5 × 106) or CD4+,CD25+ cells (106) isolated from DLN of VIP-treated CIA mice (VIP). The maximal clinical score was determined. Data represent the mean and SD of 5 mice per group. ∗ = P < 0.001 versus CD4 control. E, The CFSE-labeled CD4+ T cells (2 × 107) were transferred into naive hosts together with DLN CD4+ cells from untreated CIA mice (control; 5 × 106 cells) or with DLN CD4+ or CD4+,CD25− cells (5 × 106) or CD4+,CD25+ cells (106) from VIP-treated CIA mice (VIP). After transfer, DLN cells isolated at the peak of the disease were stimulated with CII (10 μg/ml), and the cell proliferative response and the cytokine levels in supernatants were determined. Cells cultured with an irrelevant antigen (ovalbumin; 10 μg/ml) or medium did not induce proliferation (A450 nm <0.110) or any cytokine (<0.1 ng interferon-γ [IFNγ]/ml). Data represent the mean and SD of 5 mice per group. ∗ = P < 0.001 versus CD4 control. See Figure 1 for other definitions.

We next tested in vivo the function of VIP-generated Treg in the adoptive transfer RA model by administering CD4control and CD4VIP T cells together with activated CII-specific, CFSE-labeled rCD4 cells to naive recipients. In contrast to CD4control T cells, the administration of CD4VIP T cells prevented the adoptive transfer of CIA by autoreactive rCD4 cells (Figure 4B). The CD4VIP cell–induced clinical improvement correlated with the reduction in the number and the proliferation of autoreactive CD4+ cells (CFSE labeled) in DLN and joints (Figure 4C). This suggests that the VIP-induced Treg suppressed the activation and proliferation of autoreactive CD4+ cells. Removal of CD4+ or CD25+ T cells abrogated or significantly reduced the protective action of the LN cells isolated from VIP-treated mice, whereas cotransfer of isolated CD4+,CD25+VIP cells prevented CIA in the recipients (Figure 4D). The therapeutic effect of CD4VIP cells and CD4+,CD25+VIP cells was associated with the down-regulation of the autoimmune component of the disease, because DLN T cells from CD4+VIP cell– and CD4+,CD25+VIP cell–treated mice showed weak proliferation and IFNγ production in response to the autoantigen (Figure 4E).

Amelioration of disease in CIA by VIP-induced Treg.

Since VIP induces Treg in vivo, we assessed the ability of these Treg to affect CIA after disease onset. CIA mice were treated 2 days after the onset of the disease with CD4+ T cells or with purified CD25+ and CD25− T cells obtained from arthritic mice treated with medium (control) or VIP. Mice inoculated with CD4VIP cells had a much improved clinical score, and depletion of CD25+ cells from CD4VIP cells significantly abolished their therapeutic effect (Figure 5A). CD4+,CD25+control cells had a slight beneficial effect, while the therapeutic effect of CD4+,CD25+VIP cells was highly significant (Figure 5A). The expression of proinflammatory cytokines (TNFα, IFNγ, and IL-12p40) in the joints of mice inoculated with CD4VIP cells was significantly reduced compared with that in mice that received CD4control cells or no treatment (Figure 5B). We also evaluated the proliferative capacity and the cytokine profile of spleen cells restimulated with CII ex vivo. Spleen cells from mice inoculated with CD4VIP cells exhibited reduced proliferation and did not secrete Th1-type cytokines (i.e., IL-2 and IFNγ) (Figure 5C). In addition, mice treated with CD4VIP cells had decreased titers of serum anti-CII antibodies, especially of the Th1-mediated IgG2a isotype (Figure 5D). These results indicate that VIP-induced Treg are able to suppress the proliferation of and cytokine production by autoreactive T cells in CIA, leading to a significant clinical improvement even after disease onset.

Figure 5.

Prevention of CIA by in vivo VIP-induced Treg. CD4+ cells (5 × 106), CD4+,CD25− cells (5 × 106), or CD4+,CD25+ cells (106) were isolated from DLN of untreated (control) or VIP-treated CIA mice at the peak of the disease and injected intravenously into arthritic mice 2 days after disease onset (8 animals per group). CIA animals without any treatment (none) were used as arthritic controls. A, Arthritis clinical score was evaluated 10 days after transfer. B, Messenger RNA expression for different cytokines was evaluated by MultiProbe RNase protection assay analysis in the affected joints at the peak of the disease. C, Proliferation and cytokine production in response to the autoantigen type II collagen (CII) were evaluated in spleen cell cultures. Cells cultured with an irrelevant antigen (ovalbumin; 10 μg/ml) or medium did not induce proliferation (A450 nm <0.110) or any cytokine (<0.1 ng interferon-γ [IFNγ]/ml). Cells stimulated with anti-CD3 monoclonal antibody (5 μg/ml) were used for assessment of nonspecific stimulation (mean ± SD A450 nm was 2.12 ± 0.24, 2.07 ± 0.14, and 2.06 ± 0.24 for none, CD4 control, and CD4 VIP, respectively). D, Levels of anti-CII antibodies (Abs) were determined in the serum collected 10 days after transfer. Data represent the mean and SD of 5 mice per group. TNFα = tumor necrosis factor α; IL-12 = interleukin-12 (see Figure 1 for other definitions).

We next investigated whether a similar therapeutic effect on CIA could be observed with CII-specific CD4+,CD25+ Treg generated with VIP in vitro. To address this issue, spleen CD4+ cells isolated from CIA mice were cultured with APCs and CII for 4 days in the absence (control) or presence of VIP. At the end of culture, CD4+, CD4+,CD25−, and CD4+,CD25+ cells were sorted and injected IV into mice with established arthritis (arthritis score >4). Whereas CD4+ cells isolated from untreated controls did not show any therapeutic effect, CD4+ cells obtained from VIP-treated cultures inhibited the progression of the disease and significantly ameliorated disease severity (Figure 6A, left panel), and depletion of CD25+ cells abolished the therapeutic effect of these cells (Figure 6A, middle panel). CD4+,CD25+ cells from VIP-treated samples were more efficient than those obtained from control cultures at ameliorating the disease on a per-cell basis (Figure 6A, right panel). The effects of these cells on clinical severity were correlated with their effects on the titers of anti-CII antibodies (Figure 6B). The presence and proliferative state of the transferred T cells was ascertained by fluorescence-activated cell sorting analysis for CFSE. CFSE-labeled CD4+ cells from both control and VIP-treated cultures were present in the spleens of recipients 10 days after inoculation (Figure 6C). Although both populations proliferated, transferred CFSE-labeled CD4+ cells from VIP-treated samples proliferated to a lesser degree than those transferred from control cultures (Figure 6C).

Figure 6.

Amelioration of established arthritis by VIP-induced Treg in vitro. Spleen CD4+ T cells isolated from CIA mice at the peak of the disease were cultured with T cell–depleted spleen cells isolated from naive DBA/1 mice (acting as antigen-presenting cells) and stimulated with type II collagen (CII; 20 μg/ml) in the absence (control) or presence of VIP (10−8M). After 4 days, CD4+ cells (107), CD4+,CD25− cells (107), or CD4+,CD25+ cells (2 × 106) were isolated from the culture, labeled with 5-carboxyfluorescein diacetate succinimidyl ester (CFSE), and transferred intravenously into mice with established arthritis (arrows). CIA animals treated with medium alone (none) were used as controls. A, Arthritis clinical score was evaluated daily after cell transfer. Data represent the mean of 8 mice per group. B, Anti-CII antibodies were determined in the serum obtained 10 days after transfer. Data represent the mean and SD of 8 mice per group. C, The relative presence and proliferative state of the CFSE-stained CD4+ cells recovered from the spleens of recipient mice were determined by flow cytometry 10 days after the transfer (8 animals per group). See Figure 1 for other definitions.

DISCUSSION

RA is a systemic inflammatory disease, presumably of autoimmune origin. Due to its pathologic, immunologic, and clinical similarities to human RA, CIA is a commonly used model for studying RA and testing potential therapeutic agents. Both CIA and RA are considered archetypal CD4+ Th1 cell–mediated autoimmune diseases in which Th1 cells reactive to components of the joint infiltrate the synovium, release proinflammatory cytokines and chemokines, and promote macrophage infiltration and activation (20). Inflammation mediators such as cytokines and free radicals, produced by infiltrating inflammatory cells, play a critical role in cartilage and bone damage, contributing to joint destruction (20). The fact that the inflammatory process in RA is chronic suggests that immune regulation in the joint is disturbed. This disturbance is probably caused by an excessive inflammatory response together with a deficiency in the mechanisms that control the immune response. Although available therapies based on immunosuppressive agents inhibit the inflammatory component of RA and either reduce the relapse rate or delay disease onset, they do not suppress progressive clinical disability.

The autoimmune-protective actions of VIP have been demonstrated in a variety of contexts, by reducing pathologic Th1 responses and deactivating dendritic cells and macrophages (5–7, 9). The present study demonstrates that VIP induces the generation and/or activation of efficient CD4+,CD25+ Treg during CIA, and that Treg play a pivotal role in the therapeutic effect of VIP on CIA. CD4+,CD25+ Treg have been reported to have a critical function in the regulation of autoimmune diseases, including RA and CIA (1–3, 10–14). Consistent with the findings of others (10, 13), our results demonstrate that depletion of CD4+,CD25+ cells before CIA induction hastens the onset of severe disease, and that the transfer of CD4+,CD25+ cells into depleted mice or mice with established CIA ameliorates disease progression. Of physiologic relevance is the fact that administration of VIP to CIA mice induced the appearance of CD4+,CD25+ cells with a Treg phenotype in the draining lymph nodes and joints. The Treg induced by VIP in the periphery apparently could migrate to the joints and induce suppression of the self-reactive response locally.

CD4+,CD25+ Treg have been characterized by high expression of the transcription repressor Foxp3, high surface expression of GITR, Nrp1, CD103, CD62L, and CD69, and low expression of CD45RB (16–19). The CD4+,CD25+ population from VIP-treated CIA mice showed a decrease in CD45RB, and increases in the expression of all the other markers, compared with the CD4+ T cells from antigen-inoculated mice.

In addition to expanding the CD4+,CD25+ population, VIP also induced more efficient Treg, in terms of both cytokine secretion and suppressive activity. The VIP-induced CD4+ Treg produce high levels of IL-10 and TGFβ. Also, on a per-cell basis, the VIP-induced CD4+ Treg are very strong suppressors of responder autoreactive T cell proliferation, particularly at low T regulatory cell:autoreactive T cell ratios. Although VIP-induced CD4+,CD25+ Treg secrete IL-10 and TGFβ, they also inhibit autoreactive T cell proliferation through direct cellular contact. This distinguishes the VIP-induced Treg from the classic Tr1/Th3(Tr2) regulatory CD4+ T cells, the suppressive mechanism of which is cytokine dependent (21–23), and from the recently reported CD25+ cell–cell contact–dependent and cytokine-independent suppressors recruited from the peripheral CD25− population by CD4+,CD25+ T cells stimulated with IL-2 and TGFβ (24), suggesting that the Treg population induced by VIP during CIA resembles different populations of Treg. In this regard, we have recently reported that VIP induces the differentiation of tolerogenic dendritic cells with the capacity to generate Tr1-like cells with regulatory action (25). In fact, the administration of VIP-induced dendritic cells and/or Tr1-like cells inhibited the progression of CIA (26).

A characteristic marker of Treg is the constitutive expression of CTLA-4, which has been reported to be a negative regulatory factor critical for the induction and function of Treg (27). Consistent with these reports, VIP-induced Treg expressed high levels of CTLA-4, explaining the partial dependence of the regulatory activity of these cells on cell–cell contact. In addition, CTLA-4 plays a role in the suppressive mechanism of VIP-induced CD4+,CD25+ Treg, since suppression is abrogated by treatment with anti–CTLA-4 antibodies. Although CTLA-4 is expressed at high levels in Treg, its role in the development and/or function of Treg is not clear. A mechanism involving induction of indoleamine 2,3-dioxygenase following CTLA-4 binding to CD80/CD86 has been recently proposed (28).

In a previous report describing the therapeutic action of VIP on CIA (5), investigators in our group suggested that this effect was mediated by down-regulation of both components of the disease, the synovial inflammatory response and the Th1-mediated autoimmune response, in which the generation of Th2 cells could be crucial. The induction of Treg by VIP adds a new player to this scenario. In this context, it is important to determine the involvement of the CD4+,CD25+ population in the therapeutic effect of VIP. Some evidence indicates that these Treg play a major role. Interestingly, VIP also inhibited events in the inflammatory phase of CIA following the activation/differentiation of antigen-specific effector Th1 cells (5). In addition, whereas T cell proliferation in response to the autoantigen was almost abolished in VIP-treated animals, levels of Th2 cytokines produced by these low-proliferating cells were significantly increased (5). Both observations could be explained by the fact that Treg confer significant protection against CIA, by promoting protective Th2 responses and decreasing the joint homing of autoreactive cells (3, 12, 15).

In addition, the present study demonstrates that unfractionated CD4+ cells isolated from VIP-treated mice have a significant therapeutic effect on CIA. The CD4+,CD25+ cell population represents ∼23% of these cells. CD4+,CD25+VIP cells had a therapeutic effect similar to that of 5-fold numbers of CD4VIP cells, and the therapeutic effect of CD4VIP cells was almost abrogated when they were depleted of CD4+,CD25+ cells (Figure 5). However, although the CD4+,CD25+ population bears the major suppressive capacity within the CD4+ T cells induced by VIP, the CD4+,CD25− population, presumably containing Tr1-like cells, also showed a partial therapeutic effect on CIA. Furthermore, in vivo blockade of the Treg mediators CTLA-4, TGFβ, and IL-10, but not of the Th2-type cytokine IL-4, significantly reversed the therapeutic effect of VIP in CIA, supporting the notion of a major role of Treg versus Th2 cells in the VIP action. Therefore, the generation of Treg by VIP could explain the selective inhibition of Th1 immune responses once T cells have completed differentiation into Th1 effector cells, as evidenced by the therapeutic effect of delayed administration of VIP in established CIA (5).

Importantly, we have recently found that Th1 effectors are more susceptible than Th2 cells to suppression by VIP-induced Treg (25). However, because VIP directly affects Th1/Th2 balance and the inflammatory response through multiple mechanisms (5–7, 9), a Treg-independent effect of VIP in the differentiation of Th1 cells and in synovial inflammation in CIA cannot be ruled out. Along these lines, VIP has been shown to inhibit the production of inflammation mediators by synovial cells isolated from arthritic mice within a time course that excludes the involvement of newly generated Treg (5).

The mechanisms involved in the generation/activation of Treg by VIP during CIA are not fully understood. Whether VIP acts directly on T cells to induce the generation or expansion of CD4+,CD25+ Treg remains to be established. However, the present study provides evidence that they can be peripherally generated from the CD4+,CD25− compartment, because VIP prevented CIA progression in CD25-depleted mice and restored the number of CD4+,CD25+ Treg. In addition, VIP is able to convert in vitro activated antigen-primed CD4+,CD25− cells to very efficient CD4+,CD25+,Foxp3+,CTLA-4+ cells. The VIP-induced presence of Treg in the draining lymph nodes and joints of arthritic mice could simply be a consequence of recruitment of Treg to these sites through the effect on the production of certain synovial/APC/T cell–related chemokines. However, whereas treatment of CIA mice with VIP increased the expression in the synovium of CCR4 and CCR8, 2 chemokine receptors specifically expressed in Treg (29, 30), reflecting an increased presence of Treg, it failed to increase the expression of the ligands for CCR4 and CCR8 (CCL17 and CCL1) (data not shown), which is inconsistent with this possibility.

Considerable effort has recently been focused on the use of antigen-specific Treg generated ex vivo for the treatment of several autoimmune diseases. We have found that treatment with VIP-induced Treg suppresses the systemic CII-specific T and B cell responses and the synovial inflammatory response and prevents the progression of the disease. Morgan and coworkers have previously reported the effective treatment of CIA by adoptive transfer of CD4+,CD25+ Treg (10, 13), although the therapeutic effect was mainly associated with local regulation of the inflammatory response in the joint, but not with regulation of the systemic T and B cell responses. These differences could be attributed to the fact that, in comparison with conventional CD4+,CD25+ Treg, VIP-induced Treg are a mixture of cells expressing higher amounts of the mediators involved in their suppressive action, such as CTLA-4, IL-10, and TGFβ, making them very efficient at suppressing autoreactive Th1 cells and subsequent B cell responses. The generation of highly efficient Treg by VIP ex vivo could be used as an attractive therapeutic tool in the future, avoiding the administration of the peptide to the patient. These observations provide a powerful rationale for the assessment of the efficacy of VIP as a novel approach to the treatment of RA.

Ancillary