Abstract
Small interfering RNA (siRNA)–based therapies allow targeted correction of molecular defects in distinct cell populations. Although efficient in multiple cell populations, dendritic cells (DCs) seem to resist siRNA delivery. Using fluorescence labeling and radiolabeling, we show that cholesterol modification enables siRNA uptake by DCs in vitro and in vivo. Delivery of cholesterol-modified p40 siRNA selectively abolished p40 transcription and suppressed TLR-triggered p40 production by DCs. During immunization with peptide in CFA, cholesterol-modified p40 siRNA generated p40-deficient, IL-10–producing DCs that prevented IL-17/Th17 and IFN-γ/Th1 responses. Only cholesterol-modified p40-siRNA established protective immunity against experimental autoimmune encephalomyelitis and suppressed IFN-γ and IL-17 expression by CNS-infiltrating mononuclear cells without inducing regulatory T cells. Because cholesterol-modified siRNA can thus modify selected DC functions in vivo, it is intriguing for targeted immune therapy of allergic, autoimmune, or neoplastic diseases.
Introduction
The RNA interference (RNAi) technology is established as efficient treatment strategy for various gene-targeted therapies. Because RNAi delivery to target cells has to be effective and safe, nonviral delivery systems and biochemical modifications are coming into major focus for targeting gene expression in vivo. The in vivo application of small interfering RNA (siRNA) requires structural modifications to improve serum stability and target tissue delivery. For instance, siRNA encapsulated by lipid nanoparticles facilitates siRNA delivery to hepatocytes followed by sequence-specific knockdown. Recently published clinical trials demonstrated first results on the efficacy of RNAi therapy in humans to correct metabolic disorders (1, 2). In contrast, the in vivo use of RNAi technologies for the therapy of inflammatory autoimmune diseases still faces major hurdles. For the successful delivery of RNAi in autoimmunity, the constructs should preferentially target immune cells without inducing major off-target effects. In addition, the constructs should not be immunogenic to minimize the risk for activating TLR and to avoid the induction of proinflammatory responses. One possibility to minimize off-target effects is the selection of a target gene preferentially expressed by immune cells. Therefore, we decided to focus on sequences that silence p40 (Il12b). As part of IL-12 and IL-23, p40 is mainly produced by dendritic cells (DCs) (3). DCs producing IL-12 and IL-23 are of central importance for the activation of autoreactive Th cells toward pathogenic Th1 and Th17 cells (4). Both of these Th cell subsets are crucially involved in the pathogenesis of numerous autoimmune diseases such as multiple sclerosis or psoriasis (5, 6). Notably, the spontaneous uptake of siRNA by DCs is very limited and typically requires the use of transfection reagents (7), which cannot be applied in vivo. Also, the use of naked siRNA is not ideal, because naked siRNA is less stable and does not penetrate the cell membrane easily. Therefore, we decided to use siRNA with a chemical modification. Cholesterol modification of siRNA has been introduced to deliver specific RNAi to liver and other scavenger receptor–expressing tissue (8–12). Because DCs express scavenger receptors (13), we aimed to generate cholesterol-modified siRNAs (chol-siRNA) specific for p40. To demonstrate that p40-specific chol-siRNAs are functional in DCs in vivo, we used these constructs in the mouse model of experimental autoimmune encephalomyelitis (EAE), where the key role of p40-expressing DCs is well established (14, 15).
We first identified p40-specific sequences and analyzed the uptake of p40-targeting chol-siRNA by DCs in vitro and in vivo. Furthermore, we studied the effects of p40-specific chol-siRNA on the DC phenotype, and demonstrate its potential as an RNAi-based therapeutic to induce type II DCs and to silence autoimmune inflammation in the model of EAE.
Materials and Methods
Mice
SJL mice, 6–8 wk of age, were purchased from Janvier and maintained in the animal care facilities of the University Medical Center Tübingen. Animal experiments were approved by the Institutional Animal Care and Use Committee of the Regierungspräsidium Tübingen (HT01/08 and HT10/13).
Experimental autoimmune encephalomyelitis
EAE was induced by s.c. immunization of female SJL mice with 37.5 μg PLP139–151 peptide (EMC Microcollections) in CFA (Difco) and i.p. injection of 200 ng pertussis toxin (Calbiochem) (5). The clinical EAE score was followed and rated by the following scale: 0 = no disease; 1 = limp tail; 2 = hind-limb weakness or partial paralysis; 3 = complete hind-limb paralysis; 4 = forelimb and hind-limb paralysis; 5 = moribund state (5).
RNA isolation and gene expression
Total RNA was purified from cultured cells or from ex vivo–isolated cells and reverse transcribed into cDNA using commercially available kits (Biozym). Relative gene expression levels were determined by quantitative real-time PCR using TaqMan probes (TM; TIB MolBiol) for ßactin, Il12p40, Il23p19, Il4, Il6, Il10, Il17, Ifnγ, and Foxp3 and the LightCycler480 system (Roche). The relative expression of the indicated genes was calculated relative to the expression of β-actin. Control conditions were set as 1.0 as indicated. Actb: forward (for), 5′-ACCCACACTgTgCCCATCTA, reverse (rev), 5′-gCCACAggATTCCATACCCA; TM: FAM-CATCCTgCgTCTggACCTggC-BBQ; Il4: for, 5′-CCAAACgTCCTCACAgCAAC-3′, rev, 5′-gCATCgAAAAgCCCgAAAg-3′; TM: FAM-AgAACACCACAgAgAgTgAgCTCgTCTgTA-BBQ; Il6: for, 5′-CggAggCTTAATTACACATgTTCTC-3′, rev, 5′-ggTAgCTATggTACTCCAgAAgACCA-3′; TM: FAM-ACgATgATgCACTTgCAgAAAACAATCTgA-BBQ; Il10: for, 5′-AgCTggACAACATACTgCTAAC-3′, rev, 5′-CTCTTATTTTCACAggggAGAA-3′; TM: FAM-CgCCTCAgCCgCATCCTGAgggTC-BBQ; Il17: for, 5′-AgggTgACgTggAACggT-3′, rev, 5′-gAgAgCTTCATCTgTgTCTCTgATgC-3′; TM: FAM-TggACACgCTgAgCTTTgAgggATgAT-BBQ; Ifnγ: for, 5′-CgAAAAAggATgCATTCATgAgTA-3′, rev, 5′-gCTggTggACCACTCggA-3′; ΤM: FAM-TgCCAAgTTTgAggTCAACAACCCA-BBQ; Il12p40: for, 5′-gCTCAgAgTCTCgCCTCCTT-3′, rev, 5′-gAgCTggAgAAAgACgTTTATgTTg-3′; TM: FAM-ACATggAgTCATAggCTCTggAAAgACCC-BBQ; Il23p19: for, 5′-CACCAgCgggACATATgAATC-3′, rev, 5′-CAgAACTggCTgTTgTCCTTgA-3′; TM: FAM-CACTggATACggggCACATTATTTTTAgTC-BBQ; Foxp3 for, 5′-gCAATAgTTCCTTCCCAgAgTTCTT-3′, rev, 5′-CAAAgCACTTgTgCAggCTC-3′; TM: FAM-TTTCTgAAgTAggCgAACATgCgAgTAA-BBQ.
Cytokine analysis and flow cytometry
Commercially available ELISA kits were used for the quantification of the cytokines IL-12p40, IL-12p70, IL-23, IL-10, and IL-6, from serum or cell culture supernatants. Cells were isolated from draining lymph nodes on day 3 or 7 as indicated or from the CNS on day 16 after immunization and analyzed for cytokine production by quantitative real-time PCR or flow cytometry. Intracellular cytokine staining was performed after stimulating cells with PMA and ionomycin (Sigma-Aldrich) in the presence of GolgiStop (BD Biosciences) for 4 h (5). Cells were fixed with 2% formaldehyde, permeabilized with saponin-containing buffer, and stained with fluorochrome-labeled Abs directed against CD4 (Gk1.5; Biolegend) and IFN-γ (XMG1.2; eBioscience), IL-4 (11B11; eBioscience), IL-17A (TC11-18H10; BD Biosciences), IL-2 (JES6-5H4; eBioscience), IL-10 (JES5-16E3; eBioscience), TNF (MP6-XT22; eBioscience), or Foxp3 (FJK-16s; eBioscience). Fluorochrome-labeled anti-CD11c (HL3; BD Biosciences) Abs were used for staining DCs. Cells were analyzed by flow cytometry (LSRII flow cytometer; BD Biosciences), and collected data were analyzed by FCS Express software (De Novo Software).
LPS activation of DCs in vivo and in vitro
Mice received PBS or 50 ng cholesterol-modified p40 siRNA (chol-p40-siRNA) in PBS i.v. on 2 consecutive days and 5 μg LPS i.p. (Sigma-Aldrich) on day 3. To analyze the indicated serum cytokine production by ELISA (R&D Systems or BD Biosciences), we sacrificed mice after 6 h. For ex vivo LPS activation of DCs, mice received either PBS alone or 50 ng chol-p40-siRNA or cholesterol-modified control siRNA (chol-luc-siRNA) in PBS on 2 consecutive days before isolating DCs from spleen and lymph nodes by using CD11c beads and magnetic cell sorting (MACS). Purified CD11c+ DCs were activated with LPS (100 ng/ml). Bone marrow–derived DCs were generated as described previously (5), and immature DCs were incubated with the indicated siRNA constructs for 2 h before stimulation with LPS (100 ng/ml).
In vivo siRNA treatment
The siRNAs specific for p40 were designed using the software EMBOSS (Pasteur Institute) and cholesterol-modified as described previously (8, 9) All siRNA constructs were purchased from Thermo Fisher Scientific. Stocks were prepared in the manufacturer’s buffer, diluted with NaCl (0.5 μg/ml), and injected via the tail vein (100 μl/mouse/d). In EAE experiments, mice were injected on 4 consecutive days with the indicated siRNA and immunized on day 3 for EAE. Alternatively, mice were immunized for EAE and injected on 3 consecutive days with the indicated siRNA at present clinical disease (score ≥ 1). To analyze the biodistribution of the siRNA constructs, we labeled p40-siRNA or chol-p40-siRNA with 5′-[33P] (Hartmann Analytic). Two hours after i.v. injection of 50 ng [33P]-siRNA, blood and tissue samples were isolated and measured in a scintillation counter MicroBeta TriLux (Perkin Elmer).
Immunofluorescence
In vitro–generated DCs or CD11c+ DCs isolated from lymph nodes and spleen were seeded in eight-well culture slides (BD) and allowed to attach overnight. Then DCs were incubated with FAM-p40-siRNA or FAM-chol-p40-siRNA (Thermo Fisher Scientific) for 2 h at 37°C. Cells were fixed in cold ethanol/acetone and stained with DAPI and Alexa Fluor 488 phalloidin (Invitrogen). Raftlike lipid microdomains were chemically disrupted by depleting cholesterol with 7.5 mM methyl-β-cyclodextrin (MBCD; Sigma) for 1 h before incubation with the siRNA constructs. For the analysis of the CNS, isolated brain tissue was embedded in Jung Tissue Tek freezing medium (Leica Microsystems). Five-micrometer sections were fixed in cold ethanol/acetone and stained with anti-CD3 (Dako), DAPI, and FluoroMyelin stain (Invitrogen). Stained cells or tissue sections were examined with an Axiovert 200 microscope (Zeiss). Images were processed using an OptiGrid system and VisiView Imaging software (Visitron Systems).
T cell proliferation assay
Mice were injected with the indicated siRNA on 2 consecutive days and immunized for EAE on day 3. To determine the Ag-specific proliferation of in vivo–activated T cells, we isolated CD4+ cells from draining lymph nodes by magnetic cell sorting. Equal cell numbers of CD4+ T cells were cocultured with PLP-pulsed DCs. On day 3, [3H]thymidine (0.25 μCi/well; PerkinElmer) was added for 16 h before harvest. Incorporation of radioactivity was determined by scintillation counting.
Statistical analyses
Statistical analyses were performed by paired or unpaired t test or by ANOVA, followed by Dunnett’s multiple comparison test using GraphPad Prism 5 software. The p values <0.05 were considered significant.
Results
Construction and identification of a chol-siRNA specific for IL-12/IL-23p40
We first constructed a series of siRNA sequences for specific silencing of the IL-12/IL-23 subunit p40 (Supplemental Fig. 1A). To evaluate the functional activity of each sequence, we transfected immature in vitro–generated DCs with different siRNA constructs by electroporation. After electroporation, DCs were activated with LPS (16). Only two of the siRNA constructs tested, siRNA4 and a previously published sequence (17) (siRNA6), reproducibly diminished LPS-triggered IL-12p70 protein production down to 10–20% of control DCs (Supplemental Fig. 1B). To test the potential biological activity of siRNA4 and siRNA6 in vivo, we cholesterol-modified the constructs (chol-p40-siRNA) as described previously (8, 9) (Supplemental Fig. 1C).
Following an application protocol that allows efficient prevention of EAE using naked siRNA against T-bet (18, 19), we tested whether this chol-p40-siRNA could interfere with IL-12 production by APCs. We injected mice i.v. with chol-p40-siRNA or PBS, challenged the mice with LPS, and determined serum concentrations of p40, IL-10, and IL-6 after 6 h. In PBS-treated mice, all three cytokines were readily detectable in the sera, albeit IL-10 at very low levels (Fig. 1A). Sera from mice treated with chol-p40-siRNA contained significantly reduced concentrations of p40, whereas IL-10 protein was increased. IL-6 levels remained unaffected by chol-p40-siRNA treatment, indicating selectivity and specificity of the chol-p40-siRNA (Fig. 1A). These in vivo results were surprising, as LPS induces rapid IL-12 production by DCs, a cell population that is thought to resist transfection with siRNA under in vitro and in vivo conditions (7, 20, 21).
Chol-p40-siRNA prevents IL-12p40 production after LPS activation. (A) Mice received chol-p40-siRNA or PBS and challenged with 5 μg LPS for 6 h. Serum cytokine concentrations of IL-12p40, IL-10, and IL-6 were determined by ELISA. n = 3–6/group, *p < 0.05, **p < 0.01. (B) Effect of p40-siRNA and chol-p40-siRNA on p40 and Il6 expression by LPS-activated DCs. n = 3–5 independent experiments, *p < 0.01. Data were normalized to β-actin, and expression of untreated DCs before LPS activation was set as 1.0. (C) DCs were incubated with FAM-p40-siRNA or FAM-chol-p40-siRNA for 2 h without using any transfection reagents. Cells were stained with DAPI (blue) and phalloidin (red) for immunofluorescence analysis (original magnification ×400). (D and E) DCs were treated as in (C), and siRNA uptake was determined by flow cytometry. Histogram plots from single experiments are depicted in (D); pooled data are shown in (E). n = 3–5. *p < 0.05. (F–H) DCs were preincubated in medium alone or with MBCD, treated with FAM-chol-p40-siRNA, and analyzed as in (C) and (D). Immunofluorescence (original magnification ×400) is shown in (F), flow cytometry from single experiments are depicted in (G), and pooled data in (H). n = 3. *p < 0.01. (I) DCs were treated as in (F), stimulated with LPS, and analyzed for Il12p40 expression. n = 3–5 independent experiments. Data were normalized to β-actin, and expression of untreated DCs before LPS activation was set as 1.0. All data are shown as mean ± SEM. *p < 0.05.
Chol-p40-siRNA prevents IL-12p40 production after LPS activation. (A) Mice received chol-p40-siRNA or PBS and challenged with 5 μg LPS for 6 h. Serum cytokine concentrations of IL-12p40, IL-10, and IL-6 were determined by ELISA. n = 3–6/group, *p < 0.05, **p < 0.01. (B) Effect of p40-siRNA and chol-p40-siRNA on p40 and Il6 expression by LPS-activated DCs. n = 3–5 independent experiments, *p < 0.01. Data were normalized to β-actin, and expression of untreated DCs before LPS activation was set as 1.0. (C) DCs were incubated with FAM-p40-siRNA or FAM-chol-p40-siRNA for 2 h without using any transfection reagents. Cells were stained with DAPI (blue) and phalloidin (red) for immunofluorescence analysis (original magnification ×400). (D and E) DCs were treated as in (C), and siRNA uptake was determined by flow cytometry. Histogram plots from single experiments are depicted in (D); pooled data are shown in (E). n = 3–5. *p < 0.05. (F–H) DCs were preincubated in medium alone or with MBCD, treated with FAM-chol-p40-siRNA, and analyzed as in (C) and (D). Immunofluorescence (original magnification ×400) is shown in (F), flow cytometry from single experiments are depicted in (G), and pooled data in (H). n = 3. *p < 0.01. (I) DCs were treated as in (F), stimulated with LPS, and analyzed for Il12p40 expression. n = 3–5 independent experiments. Data were normalized to β-actin, and expression of untreated DCs before LPS activation was set as 1.0. All data are shown as mean ± SEM. *p < 0.05.
The data suggest that, in contrast with the reported resistance of APCs to spontaneously take up naked siRNA under in vitro conditions (22), DCs are capable of incorporating biologically efficient amounts of chol-p40-siRNA in vivo. We therefore tested the spontaneous uptake of chol-p40-siRNA and nonmodified p40-siRNA by DCs.
Spontaneous uptake of chol-p40-siRNA by DCs in vitro and ex vivo
We generated a FAM-labeled chol-p40-siRNA (FAM-chol-p40-siRNA) and first performed functional analysis. Immature DCs were incubated with medium alone, or with medium and either nonmodified p40-siRNA or chol-p40-siRNA, and activated with LPS for 1 h (Fig. 1B). Whereas in the absence of LPS activation p40 and ll6 were undetectable at mRNA level, LPS stimulation readily induced their expression. Only chol-p40-siRNA significantly inhibited LPS-mediated p40 expression, whereas nonmodified p40-siRNA did not affect p40 mRNA (Fig. 1B). In line with the in vivo data, the level of Il6 remained unaffected by either siRNA construct, demonstrating the specificity of the siRNA sequences for p40 (Fig. 1B). To show that the functional difference correlated with the uptake of the chol-p40-siRNA by DCs, we performed immunofluorescence microscopy and flow cytometry. Both techniques demonstrated strong intracellular fluorescence in DCs incubated with FAM-chol-p40-siRNA (Fig. 1C–E). In sharp contrast, fluorescence was almost absent when DCs were incubated with FAM-labeled nonmodified p40-siRNA (Fig. 1C–E). This confirms previous data that DCs are not capable of incorporating significant amounts of nonmodified siRNA. Conversely, the data show that the cholesterol modification obviously enables DCs to take up functionally active siRNA.
DCs can incorporate lipoproteins such as cholesterol through pinocytosis/endocytosis. To test whether these pathways were responsible for the chol-p40-siRNA uptake by DCs, we used established inhibitors during siRNA incubation of DCs (11). Indeed, the pinocytosis inhibitor MBCD strongly diminished the uptake of FAM-chol-p40-siRNA as shown by immunofluorescence microscopy (Fig. 1F) and flow cytometry (Fig. 1G, 1H). As control, the phagocytosis inhibitor cytochalasin D did not affect the uptake of the FAM-chol-p40-siRNA (data not shown). Importantly, functional analysis showed that MBCD also fully restored p40 expression of LPS-stimulated DCs treated with FAM-chol-p40-siRNA (Fig. 1I).
In vivo regulation of p40 expression by DCs through chol-p40-siRNA
To test whether freshly isolated DCs are also capable of taking up chol-p40-siRNA, we purified CD11c+ DCs from healthy mice and incubated them with either nonmodified FAM-p40-siRNA or with cholesterol-modified FAM-chol-p40-siRNA. Ex vivo–derived CD11c+ DCs spontaneously incorporated only the cholesterol-modified constructs, as illustrated by immunofluorescence (Fig. 2A). To test whether chol-p40-siRNA really affected DC functioning in vivo, we treated mice with either PBS, chol-p40-siRNA, or with a chol-luc-siRNA on 2 consecutive days before isolating CD11c+ DCs. We stimulated these freshly isolated CD11c+ DCs with LPS. Upon LPS stimulation, DCs isolated from PBS-treated mice expressed high levels of p40 mRNA and produced large amounts of IL-12 and IL-23 protein. In vivo treatment with chol-p40-siRNA abolished the capacity of DCs to express p40 and to produce IL-12 and IL-23 protein, but significantly increased the Il10 mRNA expression and IL-10 protein production. These effects were specific, as control siRNA affected neither p40 and Il10 expression nor IL-12, IL-23, and IL-10 protein production (Fig. 2B, 2C). We conclude that CD11c+ DCs efficiently incorporated biologically significant amounts of chol-p40-siRNA that severely impaired p40 expression, but enhanced IL-10 expression, under in vitro and in vivo conditions. These data underline the specificity of the chol-p40-siRNA and are consistent with results from Tyk2-, STAT4-, STAT1-, or IL-12Rβ1–deficient mice, where diminished IL-12 or IL-23 production is associated with enhanced IL-10 production after LPS activation in vivo (5). Other data confirm the mutually inverse regulation of IL-10 expression and of IL-12 expression by DCs in vitro and in vivo (23, 24).
IL-12p40 expression is suppressed in CD11c+ DCs by in vivo treatment with chol-p40-siRNA. (A) Freshly isolated CD11c+ DCs were treated as in Fig. 1C. Cells were stained with DAPI (blue) and phalloidin (red) for immunofluorescence analysis (original magnification ×400). (B) Mice were injected with chol-luc-siRNA or chol-p40-siRNA. CD11c+ DCs were isolated and activated ex vivo with LPS for 1 (B) to 18 h (C). Expression of the indicated genes was determined by quantitative RT-PCR (B) and by ELISA (C). n = 3 independent experiments (data on IL-23 and IL-10 protein were generated in separate experiments). Bars represent mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. (B) Data were normalized to β-actin, and expression of CD11c+ DCs isolated from PBS-treated mice after LPS activation was set as 1.0.
IL-12p40 expression is suppressed in CD11c+ DCs by in vivo treatment with chol-p40-siRNA. (A) Freshly isolated CD11c+ DCs were treated as in Fig. 1C. Cells were stained with DAPI (blue) and phalloidin (red) for immunofluorescence analysis (original magnification ×400). (B) Mice were injected with chol-luc-siRNA or chol-p40-siRNA. CD11c+ DCs were isolated and activated ex vivo with LPS for 1 (B) to 18 h (C). Expression of the indicated genes was determined by quantitative RT-PCR (B) and by ELISA (C). n = 3 independent experiments (data on IL-23 and IL-10 protein were generated in separate experiments). Bars represent mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. (B) Data were normalized to β-actin, and expression of CD11c+ DCs isolated from PBS-treated mice after LPS activation was set as 1.0.
Chol-p40-siRNA suppresses the development of Th1 and Th17 cells
As DC is the key APC that primes naive T cells and determines the differentiation fate of CD4+ T cells during this priming period, we investigated whether the treatment with chol-p40-siRNA interferes with the polarization of naive CD4+ T cells in vivo. To address this question, we injected SJL mice with chol-p40-siRNA or p40-siRNA and immunized the mice for active EAE, a protocol that strongly induces Th1 and Th17 cells (25). The immunization for EAE induced the DC cytokines p40, Il6, and Il10 on day 3 and the Th1 and Th17 cytokines on day 7 in draining lymph node cells. On day 3 after immunization, treatment with chol-p40-siRNA significantly suppressed p40 expression but induced Il10 mRNA without affecting Il6 expression (Fig. 3A). On day 3, nonmodified p40-siRNA also impaired p40 expression under conditions of in vivo immunization, although significantly less than chol-p40-siRNA (p = 0.0013; Fig. 3A). Importantly, this slight impairment had no further consequences on the EAE-specific immune response and is best explained by a discrete unspecific siRNA uptake (see later) as observed with other naked siRNA (26). On day 7 after immunization, chol-p40-siRNA treatment significantly reduced the expression of Ifnγ and Il17, as well as the generation of Th1 and Th17 cells, respectively (Fig. 3B, 3C). The impairment of p40 expression by the nonmodified p40-siRNA had no significant impact on the development of Th1 and Th17 responses (Fig. 3B, 3C). Foxp3 expression and TNF production by T cells remained unaffected by chol-p40-siRNA and by p40-siRNA treatment.
Chol-p40-siRNA accumulates in lymphoid tissue and impairs Th1 and Th17 development in vivo. (A–C) Mice were treated with PBS, p40-siRNA, or chol-p40-siRNA and immunized for EAE. Expression of Il12p40, Il10, and Il6 in draining lymph nodes was determined on day 3 after immunization (A). Data of three independent experiments (n = 9) are shown. Data were normalized to β-actin, and cytokine expression of control mice was set as 1.0. On day 7 after immunization, the expression of Il17, Ifnγ, and Foxp3 mRNA (B), as well as IL-17, IFN-γ, and TNF protein by CD4+ T cells (C), was determined. Before flow-cytometric analysis, cells were stimulated with PMA/ionomycin. Data of three independent experiments (n = 12) are shown. (D) Mice received a single i.v. injection of [33P]-p40-siRNA or [33P]-chol-p40-siRNA. After 2 h the indicated tissues were isolated and analyzed for [33P] by liquid scintillation counting. (E) Mice were treated as in (D). The uptake of 33P-labeled siRNA by isolated CD11c+ or CD11c− cell fractions from spleen and lymph nodes was determined by liquid scintillation counting. Bars in (D) and (E) represent mean ± SEM of four independent experiments, n = 8. *p < 0.05, **p < 0.01, ***p < 0.001.
Chol-p40-siRNA accumulates in lymphoid tissue and impairs Th1 and Th17 development in vivo. (A–C) Mice were treated with PBS, p40-siRNA, or chol-p40-siRNA and immunized for EAE. Expression of Il12p40, Il10, and Il6 in draining lymph nodes was determined on day 3 after immunization (A). Data of three independent experiments (n = 9) are shown. Data were normalized to β-actin, and cytokine expression of control mice was set as 1.0. On day 7 after immunization, the expression of Il17, Ifnγ, and Foxp3 mRNA (B), as well as IL-17, IFN-γ, and TNF protein by CD4+ T cells (C), was determined. Before flow-cytometric analysis, cells were stimulated with PMA/ionomycin. Data of three independent experiments (n = 12) are shown. (D) Mice received a single i.v. injection of [33P]-p40-siRNA or [33P]-chol-p40-siRNA. After 2 h the indicated tissues were isolated and analyzed for [33P] by liquid scintillation counting. (E) Mice were treated as in (D). The uptake of 33P-labeled siRNA by isolated CD11c+ or CD11c− cell fractions from spleen and lymph nodes was determined by liquid scintillation counting. Bars in (D) and (E) represent mean ± SEM of four independent experiments, n = 8. *p < 0.05, **p < 0.01, ***p < 0.001.
Because chol-p40-siRNA significantly affected Th1/Th17 development in draining lymph nodes in vivo, we asked whether the siRNA construct was detectable in lymphoid organs. Therefore, we labeled both the chol-p40-siRNA and the nonmodified p40-siRNA with [33P]. Mice were injected with either [33P]-chol-p40-siRNA or nonmodified [33P]-p40-siRNA. After 2 h, total tissue or isolated cell populations were analyzed for [33P] uptake. Both p40 constructs were detectable in the indicated organs (Fig. 3D). Yet, [33P]-chol-p40-siRNA accumulated selectively at significantly higher levels in lymphoid tissues. Otherwise the organ biodistribution of the p40-siRNA constructs was similar to that of other i.v. injected siRNA constructs (9, 27). More importantly, separate analysis of CD11c+ cells and of CD11c− cells showed only a significant enrichment of [33P]-chol-p40-siRNA in DCs of lymphoid tissue. In sharp contrast, CD11c+ cells showed no preferential uptake of the nonmodified [33P]-p40-siRNA (Fig. 3E). This explains the modest inhibition of p40 by p40-siRNA in vivo (Fig. 3A) and the strong, biologically significant impact that had exclusively the chol-p40-siRNA on p40, Th1, and Th17 cells (Fig. 3A–C). So far, only CpG-conjugated siRNA constructs or siRNA encapsulated in polysaccharides have been reported to target macrophages in vivo. Yet, these constructs are not suitable for the generation of type II DCs (21, 28). In vivo treatment with chol-p40-siRNA provided the unique possibility to induce type II DCs that severely impaired the generation of autoreactive Th1 and Th17 cells (Fig. 3B, 3C). Therefore, we speculated that chol-p40-siRNA might protect against EAE.
Only chol-p40-siRNA protect from severe EAE, as nonspecific chol-siRNA and nonmodified p40-siRNA do not show clinical benefit
To study the effects of chol-p40-siRNA on EAE, we immunized SJL mice for EAE induction and treated the mice with either PBS, chol-p40-siRNA, or chol-luc-siRNA. PBS-treated mice and chol-luc-siRNA–treated mice developed severe EAE and paraplegia within 20 d (grade 2–3; Fig. 4A). In sharp contrast, chol-p40-siRNA–treated mice developed only very mild clinical signs of EAE, manifesting as moderate weakness of the tail mobility (grade 1; Fig. 4B). Because chol-luc-siRNA did not attenuate the PLP139–151–induced EAE, the protection provided by the chol-p40-siRNA was highly specific. To test the importance of the cholesterol modification for the therapeutic effects of the p40-siRNA, we additionally compared the impact of nonmodified p40-siRNA on EAE. Nonmodified p40-siRNA did not affect the severe course of EAE (Fig. 4C), despite the slight suppression of p40 mRNA expression on day 3 of the immunization (Fig. 3A). To address the underlying mechanisms of disease protection, we analyzed the phenotype of CNS-infiltrating mononuclear cells during EAE. Immunofluorescence analysis of the CNS of mice on day 16 after immunization showed strong inflammation as determined by CD3 staining and severe myelin loss shown with FluoroMyelin staining in mice treated with either PBS or nonmodified p40-siRNA. In contrast, mice immunized in the presence of chol-p40-siRNA had mild, nondestructive inflammation and very discrete myelin loss (Supplemental Fig. 2). The presence of CNS-infiltrating cells in chol-p40-siRNA–treated mice suggested that other mechanisms than inhibition of migration operatively protected against EAE. None of the siRNA constructs had any significant impact on the proliferation of PLP139–151–reactive T cells as determined by [3H]thymidine incorporation (Supplemental Fig. 3A). Analyzing the cytokine expression pattern of these cells, we observed major differences among PBS-, p40-siRNA–, or chol-p40-siRNA–treated mice. The expression of the lineage-defining cytokines IL-17A and IFN-γ in the CNS was significantly suppressed in mice receiving chol-p40-siRNA as compared with PBS-treated and p40-siRNA–treated controls, and IL-4 expression was selectively induced in chol-p40-siRNA–treated mice (Fig. 4D, Supplemental Fig. 3B–D). Similarly, only the chol-p40-siRNA treatment suppressed the absolute numbers of CNS-infiltrating Th1 and Th17 cells, whereas it increased the infiltration of the CNS by Th2 cells (Supplemental Fig. 3E). Notably, chol-p40-siRNA did not affect Foxp3 expression within the CNS (Fig. 4D, Supplemental Fig. 3C), confirming that silencing p40 during immunization protected from EAE without inducing regulatory T cells. In a setting where we treated mice with established neuroinflammation (EAE score ≥ 1) with either PBS, chol-luc-siRNA, p40-siRNA, or chol-p40-siRNA, only the cholesterol-modified p40-siRNA attenuated the disease slightly, but significantly (Supplemental Fig. 4A–C).
Vaccination with chol-p40-siRNA in vivo protects from EAE. (A–C) Mice treated with PBS or the indicated siRNA constructs were immunized for EAE and followed for clinical signs. Dots represent mean EAE score ± SEM from two (A and C) or four (B) independent experiments with five to nine mice per group. (D) Mice were treated and immunized as in (B) and (C). On day 16 after immunization, shortly before the maximal intensity of encephalomyelitis was achieved, CNS-infiltrating mononuclear cells were quantified. n = 9. Expression of the indicated genes was determined by quantitative RT-PCR. Data of two independent experiments are shown (n = 15). Data were normalized to β-actin, and gene expression of CNS-infiltrating cells from the PBS-treated group was set as 1.0. All data are shown as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001.
Vaccination with chol-p40-siRNA in vivo protects from EAE. (A–C) Mice treated with PBS or the indicated siRNA constructs were immunized for EAE and followed for clinical signs. Dots represent mean EAE score ± SEM from two (A and C) or four (B) independent experiments with five to nine mice per group. (D) Mice were treated and immunized as in (B) and (C). On day 16 after immunization, shortly before the maximal intensity of encephalomyelitis was achieved, CNS-infiltrating mononuclear cells were quantified. n = 9. Expression of the indicated genes was determined by quantitative RT-PCR. Data of two independent experiments are shown (n = 15). Data were normalized to β-actin, and gene expression of CNS-infiltrating cells from the PBS-treated group was set as 1.0. All data are shown as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001.
Discussion
In vivo, siRNA allows selective targeting and manipulation of metabolic functions in a variety of cells such as microglia, endothelia, tumor cells, or hepatocytes (8, 9, 28). In sharp contrast, in vitro data suggest that macrophages and DCs largely resist siRNA uptake. To circumvent this problem, we followed different strategies. For instance, siRNA was encapsulated with liposomes, lipoplexes, polymers, and peptides, or conjugated to Ab complexes or CpG as siRNA carriers (28–30). Yet, none of these approaches seems to be suitable for treatment of inflammatory autoimmune diseases, because they cause major problems related to direct activation of APCs (28, 31), pharmacokinetics (21), toxicities, and target cell delivery. We show that in vitro and in vivo chol-siRNA allow specific silencing of genes and relatively precise targeting of cells where lipoprotein-bound conjugates can be internalized by receptor-mediated processes with predictable pharmacodynamics (8, 9). So far, only one report suggested using chol-siRNA for silencing gene expression by DCs, confirming the feasibility of our approach (32). Pinocytosis/endocytosis of chol-siRNA depends on the binding to lipoprotein receptors and transmembrane proteins like scavenger receptor class B type I or CD36 (9). CD11c+ DCs typically express such receptors (33) and are therefore ideal targets for chol-siRNA constructs (9). By using two methodologies, FAM-labeling and [33P]-radiolabeling, we could demonstrate for the first time, to our knowledge, the efficient uptake of chol-siRNA by DCs in vitro and in vivo. Although CpG-conjugated siRNA is an elegant alternative approach to target DCs, the use of this strategy is not feasible in the setting of inflammatory autoimmune diseases like EAE. In contrast, by using chol-p40-siRNA, we could show that cholesterol modification of siRNA had no stimulatory effects on the DC phenotype or their maturation status in vitro and in vivo (data not shown). An additional advantage of chol-siRNA is the prolonged stability compared with nonmodified constructs (8). As reported, cholesterol modifications help to reduce the sensitivity of siRNA to serum nucleases. This explains that only the cholesterol-modified p40-siRNA, but not the nonmodified p40-siRNA construct, effectively suppressed Th1/Th17 responses and improved EAE. Interestingly, chol-p40-siRNA treatment not only suppressed p40 expression but simultaneously promoted IL-10 production, a phenomenon reported previously in mice deficient in p40 signaling (34, 35). Thus, chol-p40-siRNA induced a type II phenotype in DCs with low IL-12, low IL-23, and high IL-10 levels. Such type II DCs can prime T cells to suppress inflammatory autoimmune diseases (5, 36). Modulating immune responses by specific targeting of p40 with chol-p40-siRNA seems to be advantageous over approaches using Abs against p40, because siRNA may directly educate DCs to prime nonpathogenic T cells (5, 36). Moreover, p40 as target gene/sequence limits unwanted off-target effects. In agreement with the early peak of IL-12 and IL-23 expression after EAE immunization (37) and the previous experience with p40 Ab treatments in MS and EAE (38, 39), chol-p40-siRNA treatment also is most potent during early disease states. Thus, delivering chol-p40-siRNA after clinically established EAE had only mild effects. Continuous application of chol-p40-siRNA may be required to enhance the positive effects and to reach levels observed in the early therapeutic setting. Our findings should allow the development of targeted therapies with protective chol-p40-siRNA for a larger spectrum of organ-specific inflammatory autoimmune diseases in experimental mice and also in humans.
Acknowledgements
We thank W. Hoetzenecker for discussions. We thank E. Müller-Hermelink and J. Holstein for excellent technical assistance.
Footnotes
This work was supported by the Deutsche Forschungsgemeinschaft (DFG) Sonderforschungsbereich 685 (to K.G., A.S.Y., and M.R.) and DFG RO 764/14-1 (to M.R.), Wilhelm-Sander Stiftung Grant 2012.056.1/2 (to M.R.), Deutsche Krebshilfe Grant 110664 (to M.R.), Eberhard Karls University Tübingen Interdisziplinäres Zentrum für Klinische Forschung Verbundprojekt 1 (to K.G., C.K., and M.R.), and by Bundesministerium für Bildung und Forschung Grant 0315079 (to K.G.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
References
Disclosures
The authors have no financial conflicts of interest.