CD4+ T cells are believed to play a central role in the initiation and perpetuation of autoimmune diseases such as multiple sclerosis. In the murine model for multiple sclerosis, experimental autoimmune encephalomyelitis, pathogenic T cells exhibit a Th1-like phenotype characterized by heightened expression of proinflammatory cytokines. Systemic administration of “regulatory” cytokines, which serve to counter Th1 effects, has been shown to ameliorate autoimmune responses. However, the inherent problems of nonspecific toxicity limit the usefulness of systemic cytokine delivery as a potential therapy. Therefore, we used the site-specific trafficking properties of autoantigen-reactive CD4+ T cells to develop an adoptive immunotherapy protocol that provided local delivery of a Th1 cytokine antagonist, the p40 subunit of IL-12. In vitro analysis demonstrated that IL-12 p40 suppressed IFN-γ production in developing and effector Th1 populations, indicating its potential to modulate Th1-promoted inflammation. We have previously demonstrated that transduction of myelin basic protein-specific CD4+ T cells with pGC retroviral vectors can result in efficient and stable transgene expression. Therefore, we adoptively transferred myelin basic protein-specific CD4+ T cells transduced to express IL-12 p40 into mice immunized to develop experimental autoimmune encephalomyelitis and demonstrated a significant reduction in clinical disease. In vivo tracking of bioluminescent lymphocytes, transduced to express luciferase, using low-light imaging cameras demonstrated that transduced CD4+ T cells trafficked to the central nervous system, where histological analysis confirmed long-term transgene expression. These studies have demonstrated that retrovirally transduced autoantigen-specific CD4+ T cells inhibited inflammation and promoted immunotherapy of autoimmune disorders.
Inflammatory T cell responses to autoantigens are implicated in a number of autoimmune diseases (1, 2, 3). Systemic administration of anti-inflammatory cytokines such as IL-4, IL-10, and TGF-β has been shown to regulate autoimmune inflammation (4, 5, 6, 7, 8). However, systemic cytokine delivery can also lead to increased risk of infections and malignancies (9, 10). Therefore, recent approaches have explored local delivery of regulatory proteins as an alternative to systemic therapeutic regimens (11, 12, 13). Although adenoviruses have been a common tool for gene delivery, these viruses allow only transient expression in infected cells and have been associated with lethal events in recent clinical trials (14, 15, 16). In contrast, retrovirally mediated gene transfer safely provides stable integration of transgenes (17) with resultant long-term expression. Furthermore, retroviral gene delivery has been used to treat human disease (18) and has avoided the viral protein immunogenicity that occurs with the use of adenoviral gene delivery.
Previously, we have shown that retroviral transduction can preferentially target Ag-specific CD4+ T cells (19). Due to the intrinsic tissue-specific homing properties of autoantigen-specific CD4+ T cells, they should serve as ideal vehicles for site-specific transgene delivery. Indeed, we have demonstrated that autoantigen-specific T cell hybridomas retrovirally transduced to express a regulatory cytokine, IL-4, were capable of trafficking to autoimmune lesions in the CNS to ameliorate experimental autoimmune encephalomyelitis (EAE)4 (20). Disease could be abrogated or exacerbated depending on the cytokine delivered (21). Other studies have used transduced T cell clones instead of hybridomas to deliver immunosuppressive proteins to lesions of EAE (22, 23, 24). However, characteristics of uncontrolled growth and abnormal homing properties render hybridomas and T cell lines a poor option for the therapy of human disease. Therefore, the work presented here explores the use of short-term ex vivo retroviral transduction of primary autoreactive CD4+ T cells for therapeutic transgene delivery.
Th1-type CD4+ T cells are considered to be a predominant contributor to the initiation and persistence of autoimmunity; however, there is conflicting evidence regarding the influence of proinflammatory cytokine expression on autoimmune disease. Mice genetically deficient in proinflammatory mediators such as IFN-γ, TNF, and lymphotoxin are as susceptible to EAE as wild-type mice (25, 26, 27). Furthermore, in some cases, administration of Th1 cytokines can provide protection from autoimmune disease (28, 29, 30). In contrast, IL-12 administration has been shown to promote autoimmune inflammation. IL-12 both accelerated diabetes onset and decreased IL-4 production by islet-infiltrating cells when administered to prediabetic nonobese diabetic mice (31). Ab neutralization of IL-12 was protective for both EAE (32) and collagen-induced arthritis (CIA) in IFN-γR−/− mice (33), and IL-12−/− mice were resistant to EAE or CIA induction (34, 35). Collectively, these studies demonstrate the significant contribution of IL-12 to Th1-directed autoimmune inflammation. Indeed, IL-12 is the predominant cytokine that promotes differentiation of Th1 effector cells (36). Produced by macrophages, dendritic cells, and B cells, IL-12 stimulates IFN-γ production from NK cells and CD4+ and CD8+ T cells. IL-12 exists as a 70- to 75-kDa heterodimer consisting of disulfide-bonded 40-kDa (p40) and 35-kDa (p35) subunits, and both p40 and p35 subunits must be present for biological activity. Interestingly, the IL-12 p40 subunit can bind with high affinity to the IL-12R (37) and has been reported to act as an IL-12R antagonist (38). Antagonism of the IL-12R in vivo by IL-12 p40 may inhibit the development of Th1 effector cells, thus providing an interesting approach to therapy of autoimmune diseases. Thus, rather than administering cytokines that could incidentally promote tissue pathogenesis (39), we chose to regulate IL-12 via CNS-specific delivery of an IL-12R antagonist, the IL-12 p40 subunit. The studies presented here investigate local IL-12 p40 delivery to provide adoptive immunotherapy of EAE using myelin basic protein (MBP)-specific CD4+ T cells retrovirally transduced to express IL-12 p40.
Materials and Methods
PL/J (H-2u) and B10.PL (H-2u) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). B10.PL mice transgenic (Tg) for a Vα4/Vβ8.2 TCR specific for MBP N-acetylated (NAc)1–11 have been described elsewhere (40) and were obtained from Dr. C. C. Whitacre (Ohio State University, Columbus, OH). Mice were used between 6 and 8 wk of age and were maintained according to institutional guidelines under approved protocols in the Stanford Medical Center’s Department of Comparative Medicine (Stanford, CA).
Induction of EAE
MBP peptide NAc1–11 (NAc-ASQKRPSQRHG) was synthesized and HPLC purified at the Protein and Nucleic Acid Facility (Beckman Center, Stanford University). T cell medium (complete RPMI, RPMI-C) consisted of RPMI 1640 (Life Technologies, Gaithersburg MD) supplemented with l-glutamine (2 mM), penicillin (100 U/ml), streptomycin (100 μg/ml), nonessential amino acids (1%), sodium pyruvate (1 mM), HEPES buffer (10 mM), 2-ME (50 μM), and FCS (10%) (HyClone Laboratories, Logan UT). For passive EAE, splenocytes from MBP TCR-Tg mice (5 × 106 cells/ml) were cultured in RPMI-C with NAc1–11 (10 μg/ml) and 100 U/ml IL-12 (BD PharMingen, San Diego CA) and restimulated with NAc1–11, IL-12, and irradiated APCs on days 4 and 8. Cells were maintained at 37°C with 6% CO2. On day 11, 1 × 106 CD4+ MACS-isolated (Miltenyi Biotec, Auburn CA) cells were transferred i.v. to B10.PL recipients. Mice also received 200 ng pertussis toxin (List Biological Laboratories, Campbell, CA) i.p. in 0.2 ml PBS at the time of cell transfer and 48 h later.
Active EAE was induced as previously described (41). Mice were injected s.c. over four sites on the back with a total of 100 μl CFA (containing 200 μg Mycobacterium tuberculosis Jamaica strain) combined with 200 μg guinea pig MBP (Sigma, St. Louis, MO) for B10.PL mice and 100 μg guinea pig MBP for PL/J mice. Pertussis toxin was administered as described above. Animals demonstrated clinical signs (cs) within 7–14 days and were scored as follows: 1, limp tail or waddling gait with tail tonicity; 2, waddling gait with limp tail (ataxia); 2.5, ataxia with partial limb paralysis; 3, full paralysis of one limb; 3.5, full paralysis of one limb with partial paralysis of second limb; 4, full paralysis of two limbs; 4.5, moribundity; and 5, death. Daily cs are averaged for the group and mean incidence; cumulative cs, cs per day, day of onset, highest cs, and the number of relapses ± SEM are described. Mean cumulative cs was calculated by averaging the sum of daily clinical scores for individual animals. The cs per day value was calculated by dividing the cumulative cs by the number of days the animal remained in the study. To determine the mean day of onset, animals not developing EAE were not included in the analysis. To determine the mean highest cs, mice not developing EAE were assigned a value of 0 and included in the analysis. Relapses were defined by a drop in clinical score sustained for at least 2 days followed by an increase in clinical score sustained for at least 2 days.
pGCy (6700 bp) was derived from the previously described Moloney murine leukemia virus-based retroviral vector, pGCIRES, which contains a SrfI polylinker and an encephalomyocarditis internal ribosome entry site (19). The enhanced yellow variant of green fluorescent protein (GFP) (713 bp) was PCR amplified from pEYFP-C1 (CLONTECH Laboratories, Palo Alto, CA) with oligonucleotide primers (sense, 5′-TCG CCA CCA TGG TGA GCA AGG GCG-3′; and antisense, 5′-TCC TCC GGA TCA TTA CTT GTA CAG CTC GTC CAT-3′) and subcloned into pGCIRES at NcoI and BstEI sites, replacing GFP. The murine IL-12 p40 cDNA was obtained from Riken Gene Bank (Tsukuba, Japan) with permission from Dr. H. Hamada (Cancer Chemotherapy Center, Cancer Institute, Tokyo, Japan). The IL-12 p40 gene (1000 bp) was PCR amplified from pMFGmIL-12 p40 with oligonucleotide primers (sense, 5′-GGG TGC ATG CAT GTG TCC TCA GAA GCT AAC C-3′; and antisense, 5′-GCT GCC ATG GCT AGG ATC GGA CCC TGC AGG G-3′) and subcloned into pGCy using SrfI restriction ligation and termed pGCy.p40 (7691 bp). For bioluminescent cell tracking, the GFP/yellow-green luciferase gene fusion (gfp/luc) (42) was isolated from pJW.GFP-yLuc.1 by restriction digestion with HpaI and BglII and was then subcloned into the Moloney murine leukemia virus-based retroviral vector pGCpgk at the SalI and BamHI restriction sites and termed pGCgfp/luc (9200 bp).
The Phoenix retroviral packaging cell lines are derived from a 293 T cell line and have been described elsewhere (43). Ecotropic packaging cells (Phoenix-E) were cultured in complete DMEM (DMEM-C), which contained DMEM (Life Technologies) supplemented with FCS (10%), glutamine (2 mM), penicillin (100 U/ml), and streptomycin (100 μg/ml). Phoenix-E cells were maintained in DMEM-C with diphtheria toxin (1 μg/ml) (Calbiochem, La Jolla, CA) for selection of the ecotropic envelope (env) gene and hygromycin B (140 μg/ml) (Boehringer Mannheim, Indianapolis, IN) for selection of the virion assembly genes (gag-pol), and cells were removed from selection before transfection. The NIH3T3 cell line (ATCC CCL92) (American Type Culture Collection, Manassas, VA) was cultured in DMEM supplemented with bovine serum (10%) (JRH Biosciences, Lenexa KS), l-glutamine (2 mM), penicillin (100 U/ml), and streptomycin (100 μg/ml). MBP and collagen type II (CII)-specific T cell hybridomas were produced as previously described (44). Briefly, spleen cells from either CII-specific or MBP-specific TCR-Tg mice were stimulated with specific peptide Ags (40 μg/ml for CII and 30 μg/ml for MBP NAc1-11). Forty-eight hours after stimulation, CD4+ T cells were purified by MACS using anti-CD4 MicroBeads (Miltenyi Biotec). These cells (1 × 106) were then fused with the BW5147 TCR-αβ-negative T cell line (1 × 106) using 40% polyethylene glycol. Fusions were selected in hypoxanthine-aminopterin-thymidine-supplemented medium for 3 wk, hypoxanthine-thymidine-supplemented medium for 2 wk, and were then cultured in RPMI-C and further selected by FACS sorting for Vβ8.2 TCR-positive cells, yielding ≥98% purity. All cells were maintained at 37°C in 6% CO2.
A total of 2 × 106 Phoenix-E packaging cells were cultured in 12 ml DMEM-C in 100-mm tissue culture dishes (Nalge Nunc International, Roskilde, Denmark). Following overnight incubation, cells were transfected with 10 μg retroviral plasmid DNA (Qiagen, Valencia, CA) or no plasmid DNA (mock) using a modified version of the calcium phosphate precipitation protocol, which is described elsewhere (43). At 8–12 h after transfection, calcium phosphate-containing medium was replaced with DMEM-C, and cultures were maintained at 37°C in 6% CO2 for 24–48 h and then at 32°C in 6% CO2 for 16–24 h. Viral supernatant from transfected cultures was harvested and filtered using a 0.45-μm filter (Nalge Nunc International) and was then stored at −80°C. Virus titers were determined using NIH3T3 lines as previously described (45), and virus stocks with titers >4 × 106 were used.
Splenocytes from MBP TCR-Tg mice (5 × 106 cells/ml) were cultured in RPMI-C with NAc1–11 (10 μg/ml) for 24 h. Then, 2–5 × 106 activated splenocytes or unactivated NIH3T3 cells were cultured in six-well plates (0.5 ml/well) and overlaid with 2–3 ml thawed recombinant retroviral supernatant supplemented with protamine sulfate (8 μg/ml; Sigma). Plates were centrifuged at 2500 rpm at 32°C for 2 h and transferred to 32°C in 6% CO2 for 16 h. Media was exchanged with RPMI-C medium supplemented with 10 U/ml murine rIL-2 (R&D Systems, Minneapolis, MN) and transferred to 37°C in 6% CO2 for an additional 24 h.
Cells were stained with rat anti-mouse CD4-PE (Caltag Laboratories, San Francisco CA), and dead cells were excluded by propidium iodide (PI) (Sigma). Yellow fluorescent protein (YFP) reporter expression was analyzed using the FITC channel and isolation of YFP+CD4+ cells was conducted using a FACStar flow cytometer (Stanford FACS Facility, Stanford University). Sorted samples were then reanalyzed to confirm purity (>95%). Cell surface analysis was performed on a FACScan cytometer (BD Biosciences, Mountain View CA). All data were analyzed using FlowJo (version 3.1.1) flow cytometry software (Tree Star, San Carlos, CA).
Protein production was measured by a standard sandwich ELISA protocol. Cells were Ag stimulated in RPMI-C or X-Vivo 20 serum-free medium (TGF-β) (BioWhittaker, Walkersville MD), and supernatants were harvested at 48 (IL-2, IL-4, IL-5, and IFN-γ) and 72 h (IL-12 p40, IL-10, and TGF-β). Capture Abs were incubated in 96-well Maxisorp ELISA plates (Nalge Nunc International) at 4°C overnight with 1 μg/ml of anti-IL-2, IL-4, or IL-5, 4 μg/ml of anti-IFN-γ, 5 μg/ml of anti-IL-10, 6 μg/ml of anti-IL-12 p40 (C15.6) (BD PharMingen), or 1.5 μg/ml of chicken anti-TGF-β (R&D Systems). After washing and a 1-h room temperature incubation with 5% FCS in PBS or 0.5% enzyme immunoassay-grade gelatin (TGF-β) (Bio-Rad, Hercules, CA), 50 μl sample or standard dilution of murine rIL-2, murine rIL-4, murine IL-5, murine rIFN-γ, murine rIL-10, murine rIL-12 p40 (BD PharMingen), or human rTGF-β (R&D Systems) was incubated overnight at 4°C. TGF-β samples (100 μl) were acidified for 15 min with 4 μl 1N HCl and were neutralized with 4 μl NaOH before analysis. Biotinylated anti-IL-2 (0.5 μg/ml), IL-4 (0.5 μg/ml), IL-5 (1 μg/ml), IFN-γ (0.5 μg/ml), IL-10 (1 μg/ml), or IL-12 p40 (C17.8, 1 μg/ml) (BD PharMingen) was added for 3 h at 4°C. For detection of TGF-β, 0.5 μg/ml mouse anti-TGF-β 1, 2, 3 (Genzyme, Cambridge MA) was added for 2 h, followed by washing, then a 1-h incubation with 2 μg/ml of biotinylated horse anti-mouse IgG (Vector Laboratories, Burlingame CA). For all cytokines, biotinylated Abs were removed, plates were washed, and extravidin (1:1000) (Sigma) was added for 45 min. Wells were developed with 3,3′,5,5′-tetramethyl-benzidine substrate (Sigma), and the reaction was stopped with 1N HCl. Plates were read at 450 nm on a microtiter plate reader (Wallac, Gaithersburg, MD). Cytokine concentrations (nanograms per milliliter) were determined by comparing the OD of samples to the standard curve.
Splenocytes from MBP TCR-Tg mice were cultured in round-bottom 96-well plates (4 × 105 cells/well) with or without NAc1–11 (10 μg/ml) in RPMI-C for 72 h. Wells were pulsed with 1 μCi [3H]thymidine 18 h after culture. Cultures were harvested onto glass-fiber filter mats using a Skatron harvester (Molecular Devices, Sunnyvale, CA) and counted by liquid scintillation on a Wallac Betaplate (Wallac). The means of replicate wells were determined, and results are expressed as mean stimulation index, which was calculated as follows: (mean cpm of cultures with Ag/mean cpm of cultures with medium alone) ± SEM.
Bioluminescent imaging was performed as previously described (46). Briefly, 1 × 106 MBP TCR-Tg splenocytes tranduced with pGCgfp/luc were transferred i.v. to PL/J recipients. Luciferin substrate (126 mg/kg) (Biosynth Technologies, Naperville IL) was injected i.p., and then mice were anesthetized with 250 mg/kg Avertin (Aldrich, Milwaukee WI). Imaging was conducted using an IVIS imaging system equipped with a cooled charge-coupled device camera (Xenogen, Alameda CA) and a Navitar f 0.9 lens (Navitar, Rochester, NY). Mice were imaged in dorsal, ventral, then lateral positions by collecting two images, a grayscale body surface reference image collected under weak illumination and an image of light emission from the animals. The emission images were collected with 5-min integration times, and pseudocolor representation of light intensity (red being the most intense and blue being the least intense) was superimposed over the grayscale body surface reference image. Data acquisition and analyses were performed using the LivingImage (Xenogen) software overlay on the IgorPro image analysis package (WaveMetrics, Seattle, WA). Animals recovered from the anesthetic after ∼30 min under close supervision and were imaged again at the indicated time points.
Tissue samples were harvested and fixed in 4% buffered paraformaldehyde for ≥8 h and were then removed to 30% buffered sucrose for ≥4 h. Frozen samples embedded into OCT compound 4583 (Tissue-Tek, Torrence CA) were then cut into 6-μm sections. Sections were overlaid with PI mounting medium (Vectashield; Vector Laboratories, Burlingame CA), air-dried for 30 min, then examined using fluorescent microscopy for YFP-positive cells (FITC excitation) and PI-positive cells (PE excitation) (Khavari Laboratory, Stanford University).
Experiments with two groups were analyzed by nonparametric ANOVA with Mann Whitney U analysis. Differences in experiments with more than two groups were determined by one-way ANOVA with Tukey-Kramer multiple comparison analysis. All analyses was performed using InStat 2.01 software (GraphPad, San Diego, CA), and values were considered significantly different at p < 0.05.
High levels of IL-12 p40 expression reduces IFN-γ production in transduced MBP-specific T cells
We have recently established an efficient method of transduction of murine CD4+ T cells using a retroviral vector that allowed bicistronic expression of both the therapeutic transgene and a GFP marker (19). GFP has since been further mutagenized to form a YFP (47, 48). YFP has GFP excitation with red-shifted emission, thus providing a brighter marker protein also detected using the FITC channel. Therefore, YFP was used in the therapeutic retroviral vector used for these studies (Fig. 1,A). Splenocytes from MBP TCR-Tg mice were infected with recombinant retrovirus containing the IL-12 p40 transgene, and FACS analysis demonstrated that ∼60% of CD4+ cells were transduced (i.e., YFP positive; Fig. 1,B). Cells transduced with empty vector had similar transduction efficiencies, whereas mock-transduced (no vector) cells had no YFP expression (data not shown). Transduced cells were sorted by FACS into low and high YFP-expressing populations and restimulated with Ag in vitro. Supernatants were analyzed for IL-12 p40 content by ELISA, and data presented in Fig. 1,C demonstrate that IL-12 p40 protein levels correlated with YFP protein expression and are significantly higher in “p40 high”-transduced populations. Supernatants were also analyzed for T cell cytokines. There was no observable change in IL-4, IL-10, IL-5, TGF-β (data not shown), or IL-2 production (Fig. 1,D, ▧). However, IFN-γ levels were markedly reduced with increased levels of IL-12 p40 expression (Fig. 1 D, ▪). These data suggest that IL-12 p40 can inhibit Th1 differentiation (i.e., IFN-γ production) of transduced T cells without hindering T cell viability (i.e., IL-2 production) or promoting Th2 development. However, for therapeutic application, transgene-derived IL-12 p40 must also regulate existing Th1 cells within the autoimmune lesion.
IL-12 p40 inhibits IFN-γ production in developing and effector CD4+ T cell populations
To assess the regulatory effects of IL-12 p40 on Th1 cells, transgene-derived IL-12 p40 was overlaid onto Th1 cell cultures. For robust production of transgene-derived IL-12 p40, NIH3T3 cells were transduced with the IL-12 p40 retroviral vector. Transduction in NIH3T3 cells yielded ≥98% efficiency, and supernatants containing transgene-derived IL-12 p40 were collected and quantified by ELISA. Immunoprecipitation using supernatants from NIH3T3 cells transduced for expression of IL-12 p40 confirmed the presence of both monomeric and dimeric IL-12 p40 species (data not shown). Splenocytes from MBP TCR-Tg mice were then cultured with the immunodominant NAc1–11 MBP epitope and increasing concentrations of transgene-derived IL-12 p40. ELISA analysis of supernatants demonstrated that without IL-12 p40 (0 ng/ml), MBP-specific T cells underwent Th1 differentiation (i.e., IFN-γ production) upon Ag stimulation in vitro (Fig. 2,A, ▪). There were no observable differences in IL-2, IL-4, or IL-10 production (data not shown) or in proliferative responses to Ag following IL-12 p40 addition (Fig. 2 A, ▧). However, IFN-γ production was markedly reduced. These data demonstrate the ability of IL-12 p40 to inhibit Th1 differentiation in developing CD4+ T cells without altering proliferative responses or promoting Th2 cytokine production.
For therapeutic application, effector Th1 cells must also be inhibited by IL-12 p40. Therefore, we added rIL-12 p40 to previously differentiated Th1 cell cultures and measured changes in cytokine production. Splenocytes from MBP TCR-Tg mice were primed with Ag in vitro and then rested for 7 days and restimulated with Ag in the presence of IL-12 p40. Data presented in Fig. 2,B (▧) support what is typically observed with MBP-specific Tg cells, in that effector T cells proliferate less robustly than after primary Ag stimulation (Fig. 2 A). However, it is important to note that proliferation responses are not altered by IL-12 p40 supplementation. In contrast, T cells that produced high amounts of IFN-γ after primary Ag stimulation produced significantly less IFN-γ with the addition of IL-12 p40 at secondary Ag exposure (▪). Subsequent experiments have shown that CD4+ T cells must be exposed to IL-12 p40 within 12 h of Ag exposure to inhibit IFN-γ production (data not shown). These observations demonstrate that IL-12 p40 has a significant inhibitory effect on Th1 effector cells, which suggests that the encephalitogenic potential of Th1 cells may also be suppressed by IL-12 p40.
IL-12 increases encephalitogenicity of TCR-Tg MBP-specific T cells
To demonstrate the role of bioactive IL-12 on encephalitogenicity of MBP-specific Tg cells, splenocytes from MBP TCR-Tg mice were Ag stimulated in vitro with or without IL-12 supplementation. Passive EAE was then initiated by transferring activated cells to syngeneic non-Tg recipients. As shown by data presented in Fig. 3 A, the addition of IL-12 to cultures significantly enhanced the encephalitogenicity of MBP-specific cells. Mean disease incidence (3 of 6 vs 4 of 4) and cumulative clinical score ± SEM (6.8 ± 3.1 vs 24.9 ± 5.4*) increased in mice receiving IL-12-treated cells. In addition, the day of disease onset was earlier (12.3 ± 0.2 vs 10.0 ± 1.2), and the highest clinical score achieved (1.0 ± 0.4 vs 2.8 ± 0.5) was greater upon transfer of 1 × 106 IL-12-treated cells. Of the mice receiving IL-12-treated cells, 3 of 4 reached peak clinical scores (2.5, 3, and 4) by day 12 after transfer, whereas all mice receiving cells not treated with IL-12 demonstrated a very mild EAE clinical course (maximum disease score of 2). It is likely that IL-12 enhanced Th1 differentiation of the Ag-primed cells, because supernatant from IL-12-treated cultures demonstrated a 5-fold increase in IFN-γ production over non-IL-12-treated cells (data not shown). These studies establish the importance of bioactive IL-12 in promoting the encephalitogenic response and suggest that local targeting of the IL-12 pathway could provide protection from EAE.
Autoantigen-specific TCR expression is required for IL-12 p40 therapy
For CNS-targeted delivery of an IL-12R antagonist, MBP-specific T cell hybridomas were transduced to express IL-12 p40 and were adoptively transferred to syngeneic mice immunized to develop EAE. Data presented in Fig. 3 B demonstrate that IL-12 p40-transduced MBP-specific hybridomas significantly suppress disease. Interestingly, adoptive transfer of T cell hybridomas specific for CII and producing equivalent amounts of IL-12 p40 had no effect on EAE in similarly immunized mice but were therapeutic when transferred to mice with CIA (49). The mean cumulative clinical scores ± SEM (12.6 ± 5.4 vs 34.1 ± 4.2 and 34.3 ± 9.1), clinical score per day (0.4 ± 0.2 vs 1.1 ± 0.1 and 1.1 ± 0.3), highest clinical score (1.2 ± 0.5* vs 3.0 ± 0.3 and 3.0 ± 0.4), and the number of relapses (0.0 vs 1.0 ± 0.3 and 0.4 ± 0.2) were also reduced in IL-12 p40-treated mice when compared with nontreated or CII-specific hybridoma-treated mice, respectively. These data suggest that the site-specific TCR expression is required for therapy, presumably providing transduced cell retention by specific Ag recognition in the inflammatory lesion.
IL-12 p40-transduced primary T cells can suppress EAE
Due to the intrinsic hazards of administration of transformed (hybridoma) cells in vivo, it was preferable to establish an IL-12 p40 adoptive immunotherapy protocol using untransformed T cells. Therefore, similar experiments were performed using primary T cells from MBP TCR-Tg mice. MBP-specific CD4+ T cells were transduced to express IL-12 p40 and adoptively transferred to mice immunized for EAE. Following encephalitogenic challenge, B10.PL mice exhibited an initial acute disease followed by intermittent relapses and/or chronic EAE. To determine whether IL-12 p40 could provide therapeutic benefit in vivo, 1 × 106 IL-12 p40-transduced MBP-specific CD4+ T cells were adoptively transferred just before disease onset (day 10 after MBP immunization). Because the disease course is unique for each animal, the mean daily clinical score for each group is shown. Disease was significantly reduced in mice receiving IL-12 p40-transduced cells (Fig. 3,C). Mean cumulative clinical score ± SEM (49.5 ± 9.3 vs 32.8 ± 8.7), clinical score per day (1.4 ± 0.3 vs 0.9 ± 0.2), highest clinical score (3.3 ± 0.4 vs 2.2 ± 0.4), and the number of relapses (0.5 ± 0.2 vs 0.2 ± 0.2) from the day of treatment through 35 days after transfer were also reduced in IL-12 p40-treated mice. It is important to note that the transfer of vector-only-transduced MBP-specific cells did not alter disease when compared with nontreated control mice (data not shown). Furthermore, the amount of IL-12 p40 produced by the 1 × 106 transferred IL-12 p40-transduced T cells (∼40 ng/ml; Fig. 1,C) was approximately equal to levels that inhibited the activation of Th1 cells (50 ng/ml; Fig. 2), as demonstrated by in vitro analysis. These studies have demonstrated the therapeutic potential of IL-12 p40 delivery. However, even though analysis of serum samples 4 and 12 days after cell transfer did not demonstrate detectable systemic levels of IL-12 p40 protein (data not shown), these data alone do not demonstrate that the therapeutic IL-12 p40 transgene is delivered locally to sites of inflammation.
MBP-specific transduced CD4+ T cells traffic to the CNS
To assess the trafficking kinetics of transduced MBP-specific CD4+ T cells, we used in vivo bioluminescent imaging, which allows for sensitive detection of labeled cells in rodent models of human disease (46). This technique used T cell expression of a firefly luciferase gene to track T cell trafficking patterns in vivo. The enzymatic reaction between luciferase and its substrate, luciferin, causes photon emission that can be detected by a cooled charge-coupled device camera and represented as a pseudocolor image of light intensity overlaid onto a grayscale image of an anesthetized mouse. Recipients with white coats minimize light absorption; therefore, syngeneic PL/J mice were used for these studies. Although this mouse strain is susceptible to EAE, under our MBP/CFA/pertussis toxin immunization protocol, PL/J mice typically demonstrated a mild EAE clinical course with late onset and no relapses, which is in contrast to the robust relapsing/chronic EAE in B10.PL mice used to study the therapeutic efficacy of IL-12 p40 (Fig. 3, B and C). MBP-specific CD4+ T cells were transduced with a retroviral vector containing a GFP-luciferase gene fusion (pGCgfp/luc; Fig. 1,A) and were then analyzed and sorted using FACS. Cells expressing the GFP/luc marker protein were transferred (i.v.) to naive or MBP-immunized recipients before the onset of EAE. Upon peritoneal administration of luciferin, luciferase-expressing cells could be detected in the lungs within 5 min of transfer (Fig. 4,A). One day after transfer, mice were reanesthetized, and image analyses demonstrated luciferase-positive cells within the peripheral lymph nodes, spleen, and at the sites of immunization. Luciferase-positive cells were retained only in MBP-immunized mice and were demonstrable in regions consistent with the lumbar and thoracic regions of the spinal cord at approximately the time of disease onset. Interestingly, the time required for CD4+ T cells to traffic to CNS-related sites correlated with the kinetics of therapeutic effects after transfer of IL-12 p40-transduced cells (Fig. 3 C).
A separate experiment was performed in which MBP-specific T cells transduced for luciferase were transferred during EAE cs. Luciferase-positive cells were detected in the peripheral lymphoid tissues for ∼8 days in naive mice (n = 2). However, in MBP/CFA-immunized mice, luciferase-positive cells were present in sites of immunization, and five of six mice with EAE demonstrated luciferase-positive cells in the brain within 3 days of transfer (Fig. 4 B). No luciferase-positive cells were detected in the CNS of naive control mice at any time, and the administration of pertussis toxin alone did not alter transduced cell trafficking when compared with naive control mice (data not shown). These observations suggest that T cell trafficking was accelerated following the breech of the blood-brain barrier. Further analysis demonstrated that luciferase-positive cells were detected as long as 50 days after transfer (data not shown). Therefore, these studies suggest that retrovirally transduced CD4+ T cells traffic to sites of autoimmune inflammation and are capable of long-term persistence.
Retrovirally transduced T cells provide long-term local transgene expression
To confirm that therapeutic transgenes are delivered and expressed locally, histological evaluation was performed on brain and spinal cord sections from EAE mice that were treated with IL-12 p40-transduced cells. Approximately 30 days after adoptive transfer, YFP+ cells were detected in both brain (Fig. 5) and spinal cord sections (data not shown). There was minimal fluorescence emission in organs from mice receiving nontransduced MBP-specific cells (Fig. 5). However, YFP+ cells were detected in sections from spleen and cervical lymph nodes of MBP-immunized mice receiving vector or IL-12 p40-transduced MBP-specific cells (data not shown). It is important to note that vector-only as well as IL-12 p40-transduced cell recipients demonstrated YFP+ cells in the CNS upon histopathological analysis. Therefore, both populations of retrovirally transduced CD4+ T cells are capable of trafficking to the CNS, but only IL-12 p40-transduced cells provided therapeutic effects (Fig. 3 C).
The studies presented here demonstrate that retrovirally transduced CD4+ T cells can operate as effective vehicles for adoptive immunotherapy of autoimmune disease. Following adoptive transfer, MBP-specific CD4+ T cells transduced to express the IL-12 p40 subunit could traffic to the CNS and suppress EAE ( Figs. 3–5). Analysis in vitro demonstrated IL-12 p40 regulation of IFN-γ production (Figs. 1 and 2). Th1 cytokine production from developing and effector Th1 cells was inhibited without altering T cell viability or proliferative responses. These data suggest that IL-12 p40 can act therapeutically by interrupting Th1-mediated inflammatory autoimmune responses.
The differentiation of CD4+ T cells into Th1 cells is greatly influenced by the presence of bioactive IL-12 during Ag recognition (50, 51). Therefore, the expression of the IL-12R plays a critical role in determining the Th1-Th2 balance during the course of an immune response. Unlike the IL-12R β1 chain, the IL-12R β2-chain is not expressed on naive or resting T cells, but is induced after TCR engagement with Ag/MHC class II on APCs (52). As schematically represented in Fig. 6, the T cell up-regulates CD40 ligand (CD40L) surface expression upon activation, which binds to CD40 on the activated APC. Interestingly, CD40-mediated signaling is thought to be more effective in DC activation than engagement of MHC class II molecules (53). This CD40L-CD40 interaction triggers the APC to secrete preformed stores of bioactive IL-12 (p35/p40). IL-12 binding to the high affinity IL-12R, consisting of both the β1 and β2 chains, activates STAT1, STAT3, and STAT4 and results in the production of IFN-γ and Th1 differentiation (54). The p40 subunit of IL-12 has been proposed to bind to the high affinity IL-12R and block engagement of bioactive IL-12 (37), thus inhibiting Th1 commitment. It is known that IFN-γ−/− mice are susceptible to EAE (25); therefore, it is likely that other cytokines also influence Th1 differentiation. IL-4 can regulate the development of Th1 cells by inhibiting expression of the IL-12R β2 chain. However, this inhibition can be overcome by high levels of IFN-γ, even in cells that have begun to differentiate along the Th2 pathway (52, 55). Thus, Th1 differentiation is critically dependent on the cytokine microenvironment during T cell priming, and IL-12 p40 expression may inhibit such pathways. Interestingly, the IL-12 p40 subunit can also associate with a p19 subunit produced by dendritic cells to form the newly characterized cytokine, IL-23 (56). It has been suggested that IL-23 may have a similar biological function to IL-12. Although p19 expression within the CNS has not yet been described, it is known that astrocyte production of IL-12 will promote the development of type 1 T cell cytokine responses and NK cellular immunity (57). Therefore, it remains possible that the therapeutic effects of IL-12 p40 may be attributable to inhibition of both IL-12- and IL-23-mediated signaling during CNS inflammatory responses.
We and others have now shown that transduced primary Ag-specific CD4+ T cells provide many beneficial features for targeted immunotherapy (58). Retrovirally transduced cells were easily selected for dose of transgene based on reporter gene expression and were then safely transferred to recipient animals for local product delivery. Although pGC retroviral vectors use a constitutive promoter, in vitro analysis has demonstrated that transgene expression correlates with the activation state of the CD4+ T cell (G. L. Costa and J. M. Benson, unpublished observations). Therefore, it is reasonable to conclude that transduced T cells require Ag presentation in vivo to produce transgenic proteins. This was evident in bioluminescent cell trafficking studies (Fig. 4,A), in which luciferase production was no longer detected in naive mice once transduced T cells returned to a quiescent state (9 days after transfer). Interestingly, we observed that MBP-specific hybridomas were more efficient in reducing EAE clinical signs when compared with primary-transduced T cells (Fig. 3, B and C), even though each can produce comparable levels of IL-12 p40. It is well established that T cell hybridomas are greatly dysregulated in their growth and cytokine production. Therefore, transduced hybridomas will produce transgene and proliferate at a constant rate. In contrast, transduced primary T cells require Ag presentation for activation, proliferation, and transgene production. Therefore, not only is more time required to establish a “therapeutic” population of transduced primary T cells, but it is also highly likely that primary T cells are more subject to in vivo regulatory processes when compared with T cell hybridomas. Thus, transduced primary T cells offer an effective yet more regulated method for Ag-specific in vivo immunotherapy.
In vitro analysis demonstrated the ability of IL-12 p40 to inhibit both developing and effector Th1 populations (Fig. 2). These observations were also evidenced in vivo by the ability of IL-12 p40 to suppress EAE clinical signs when cells were administered 10 days after MBP immunization (Fig. 3 C), at which time effector Th1 populations most likely already exist within the CNS. Furthermore, T cell expression of IL-12 p40 provided long-term disease suppression, even though recipient mice were actively immunized, which most likely results in the continual repopulation of encephalitogenic T cells. Interestingly, the amelioration of the relapsing/chronic phase of EAE suggests that responses to other CNS Ags could also be suppressed by IL-12 p40 expression. To inhibit T cell IFN-γ production in vitro, we observed that IL-12 p40 must be present in sufficient quantities within 12 h of Ag exposure. Therefore, the encephalitogenicity of T cells stimulated by Ag before sufficient local expression of IL-12 p40 will not be inhibited. Thus, the observed delay between cell transfer and reduced cs may be attributed to the time required for site-specific transduced T cell trafficking, replacement of the proinflammatory environment with IL-12 p40-mediated suppression, and resolution of established inflammation and restoration of motor ability. Preliminary studies suggest that the administration of IL-12 p40-transduced cells during established disease may also suppress EAE; however, further investigation is required (G. L. Costa and J. M. Benson, unpublished observations).
These studies demonstrate that IL-12 p40 did not alter T cell viability or proliferation (Figs. 1 and 2); therefore, adoptively transferred IL-12 p40-transduced CD4+ T cells were expected to retain the ability to traffic to autoimmune sites. This was confirmed by in vivo bioluminescent studies that demonstrated trafficking of luciferase-transduced T cells to the CNS of mice immunized for EAE (Fig. 4). In addition, transduced CD4+ T cells were also present in secondary lymphoid tissue. Yet, if expression of the IL-12 p40 transgene in secondary lymphoid tissue was sufficient for CNS therapy, it would be expected that adoptive transfer of non-antigen-specific CD4+ T cells tranduced for IL-12 p40 expression would also provide therapeutic effect. In contrast, we found that collagen-reactive T cell hybridomas transduced to express IL-12 p40 were not protective when adoptively transferred into recipients immunized for EAE (Fig. 3 B). The lack of protection is most likely due to inadequate retention in sites of autoimmune lesions because subsequent bioluminescent cell trafficking studies have demonstrated that collagen-reactive T cell hybridomas traffic to and persist in peripheral lymph nodes in patterns similar to MBP-specific T cell hybridomas (49). Thus, even though it cannot be conclusively determined whether peripheral expression of the IL-12 p40 transgene influenced disease suppression, it is likely that local expression is required for amelioration of EAE. These experiments support previous studies that describe the requirement of site-specific TCR for adoptive immunotherapy (23, 59) and provide definitive evidence that the IL-12 p40 subunit can effectively block CNS autoimmune Th1 responses in vivo.
Local delivery of an IL-12R antagonist represents a therapeutic strategy for the treatment of organ-specific autoimmune diseases that may be preferable to direct cytokine delivery. Data presented here supports previous observations of IL-12 p40-mediated suppression of diabetes in nonobese diabetic mice (60, 61). However, we have shown that site-specific expression of the IL-12 p40 subunit can be used to target and ablate autoimmune lesions without demonstrable side effects that are associated with elevated levels of regulatory cytokines (21, 39). Although it was not anticipated that high concentrations of receptor antagonist would exacerbate disease, the YFP marker protein allowed cells to be sorted for low or high transgene expression to assess regulatory effects. We demonstrated that IL-12 p40 exhibited therapeutic benefit at levels as low as 40 ng/ml and had no detrimental or toxic effects on T cell populations even when expressed at high levels (100 ng/ml) ( Figs. 1–3). Obtaining autoantigen-specific CD4+ T cells from patients with autoimmune diseases would be technically demanding; however, the work presented here demonstrates the efficacy of adoptive gene therapy using very low numbers of transduced cells. These data suggest that stable expression of antagonist molecules, such as IL-12 p40, using autoreactive CD4+ T cells that traffic preferentially to inflammatory lesions may provide an effective adoptive immunotherapy for organ-specific autoimmune diseases.
This work was supported by Grants AI 39646, AI 36535, AR 6-2227 (to C.G.F.), and TGAI 07290 (to G.L.C.) from the National Institutes of Health. J.M.B. is a Boehringer Ingelheim-sponsored postdoctoral fellow of the National Multiple Sclerosis Society (FG 1340-A-1).
Abbreviations used in this paper: EAE, experimental autoimmune encephalomyelitis; CIA, collagen-induced arthritis; MBP, myelin basic protein; RPMI-C, complete RPMI medium; cs, clinical signs; GFP, green fluorescent protein; YFP, yellow fluorescent protein; Tg, transgenic; DMEM-C, complete DMEM; CII, collagen type II; PI, propidium iodide; CD40L, CD40 ligand; NAc, N-acetylated.