Liposome-encapsulated peptides protect against experimental allergic encephalitis

Abstract

Multiple sclerosis (MS) is a severe inflammatory and neurodegenerative disease with an autoimmune background. Despite the variety of therapeutics available against MS, the development of novel approaches to its treatment is of high importance in modern pharmaceutics. In this study, experimental autoimmune encephalomyelitis (EAE) in Dark Agouti rats has been treated with immunodominant peptides of the myelin basic protein (MBP) encapsulated in mannosylated small unilamellar vesicles. The results show that liposome-encapsulated MBP_46–62 is the most effective in reducing maximal disease score during the first attack, while MBP_124–139 and MBP_147–170 can completely prevent the development of the exacerbation stage. Both mannosylation of liposomes and encapsulation of peptides are critical for the therapeutic effect, since neither naked peptides nor nonmannosylated liposomes, loaded or empty, have proved effective. The liposome-mediated synergistic effect of the mixture of 3 MBP peptides significantly suppresses the progression of protracted EAE, with the median cumulative disease score being reduced from 22 to 14 points, compared to the placebo group; prevents the production of circulating autoantibodies; down-regulates the synthesis of Th1 cytokines; and induces the production of brain-derived neurotrophic factor in the central nervous system. Thus, the proposed formulation ameliorates EAE, providing for a less severe first attack and rapid recovery from exacerbation, and offers a promising therapeutic modality in MS treatment.—Belogurov, A. A., Jr., Stepanov, A. V., Smirnov, I. V., Melamed, D., Bacon, A., Mamedov, A. E., Boitsov, V. M., Sashchenko, L. P., Ponomarenko, N. A., Sharanova, S. N., Boyko, A. N., Dubina, M. V., Friboulet, A., Genkin, D. D., Gabibov, A. G. Liposome-encapsulated peptides protect against experimental allergic encephalitis.

Multiple sclerosis (MS) is a widespread neurodegenerative disease that affects the central nervous system (CNS) of young adults and leads to neurological disability. Approximately 1 million people all over the world are affected by this autoimmune disease, but its etiology and pathogenesis are still largely obscure. The disease is characterized by severe demyelination, with consequent secondary (bystander) axonal loss, and a major role in this process is played by T and B cells reactive toward components of the myelin membrane (1). The list of potential autoantigens is expanding and currently includes several oligodendrocyte-associated proteins, mainly the myelin basic protein (MBP) and the myelin oligodendrocyte glycoprotein (MOG). Aggressive lymphocytes specific against these proteins, together with macrophages, penetrate the blood–brain barrier (BBB), which results in the formation of inflammatory demyelinating lesions in the CNS. Undoubtedly, T cells orchestrate the massive attack of the immune system on oligodendrocytes. As for B cells, they are currently regarded not only as producers of antibodies (Abs) but also as antigen-presenting and cytokine-producing cells (2). Additional evidence for the involvement of B cells in demyelination and, hence, in MS pathogenesis is provided by identification of catalytic Abs to MBP, which are capable of not only binding but also cleaving the antigen (3). Recent data obtained by our group confirm that the “environmental hypothesis,” based on the existence of autoAbs cross-reactive toward both neuronal and viral antigens, may be relevant to the etiology and pathogenesis of MS (4).

Recent insights into specific features of MS induction and pathogenesis have entailed significant changes in the therapy of this autoimmune abnormality. Today, MS is usually treated with glatiramer acetate (Copaxone), injections of cytokines (IFNβ) and various monoclonal Abs (Natalizumab, Rituximab, etc.), and orally administered low-molecular-weight chemical drugs (for review, see ref. 5). Currently available and prospective therapeutics containing exact MBP sequence fragments or mimicking it can be conditionally divided into 3 groups. Group 1 includes glatiramer acetate (see above), a synthetic random copolymer of 4 aa (Glu, Lys, AL, and Tyr), which may act as a decoy for the immune system and has been approved for treatment of relapsing–remitting MS (RR-MS). This immunomodulator drug induces the population of regulatory Th2 cells to become capable of crossing the BBB and producing anti-inflammatory cytokines IL-4, IL-6, IL-10, and brain-derived neurotrophic factor (BDNF; ref. 6). Group 2 includes so-called altered peptide ligands (APLs) carrying modified (7), mutated (8), or restricted (5) T-cell receptor-binding moieties. APLs interact with T-cell receptors and can partly activate T cells, causing a switch from Th1 to Th2 phenotype, or, in some cases, induce their transformation into anergic cells. The APL derived from the MBP encephalitogenic fragment (MBP_82–98) has proved promising in inhibiting MS progression in patients with the HLA-DR2/DR4 haplotype but, unfortunately, generally failed in phase 3 of the clinical trial (5). Recently, Katsara et al. (9) have shown that a double mutation of MBP_83–99 peptide induces IL-4 responses and antagonizes IFNγ responses. Group 3 includes autoantigen and/or DNA vaccination aimed at inducing nasal or oral tolerance. According to the results of phase 2 trial, injection of BHT-3009, a plasmid encoding the full-length human MBP molecule, triggers significant tolerization of autoAbs toward several myelin antigens (10).

To date, vast clinical, immunological, and biochemical data on MS have been accumulated, and molecular mechanisms of this disease have been discovered in part, but still none of the existing drugs can effectively cure MS or reliably prevent its progression. Moreover, some patients with MS are found to be resistant to all approved types of disease-modifying therapy. Therefore, further improvement of existing approaches and generation of promising novel therapeutics for MS treatment are still a great challenge to both medical and basic sciences. Here, we show that selected immunodominant MBP peptides encapsulated in mannosylated liposomes can suppress ongoing experimental autoimmune encephalomyelitis (EAE) in DA rats and, hence, may be regarded as a prospective therapeutic agent against MS.

MATERIALS AND METHODS

Patients

Blood samples (10 ml) from 35 patients with RR-MS clinically defined according to the 2005 MacDonald criteria were obtained from the Moscow Multiple Sclerosis Center at City Hospital 11 (Moscow, Russia). The patients were between 23 and 61 yr of age (median 32.2 yr), with MS duration ≤ 3 yr. Their Expanded Disability Status Scale (EDSS) scores ranged from 0 to 4 (median 2.0). The EDSS is graded from 0 to 10, with higher grades indicating greater disability (11). None of the patients had received treatment with corticosteroids for 1 mo before blood sampling. All patients signed the informed consent form in accordance with the regulations of the Ministry of Health of the Russian Federation, confirmed by the Ethics Committee of City Hospital 11.

EAE induction in rodents and treatment of Dark Agouti (DA) rats

Experiments with animals were performed at animal facilities of the Assaf Harofeh Medical Center (Zerifin, Israel), in conformity with all rules concerning animal care, and were certified by the Ethics Commission. Animals included in the study fully satisfied the specific pathogen-free criteria. EAE was induced in 3 rodent lines: SJL, C57/BL6, and DA. Female SJL mice aged 6–8 wk were injected with bovine MBP (50 μg/animal) according to the established protocol (3); the antigen was mixed with complete Freund's adjuvant containing 2 mg/ml of Mycobacterium tuberculosis. Female C57BL/6 mice aged 6–8 wk were immunized as described elsewhere (12) with the recombinant MOG extracellular domain (100 μg/animal) mixed with complete Freund's adjuvant containing 0.5 mg/ml of M. tuberculosis. Between d 14 and 28 after immunization, mice with distinct clinical symptoms were euthanized, and their sera were collected for subsequent experiments. Female DA rats aged 8–9 wk (220–250 g) were immunized with a syngeneic spinal cord homogenate (50% w/v in saline) mixed with complete Freund's adjuvant (1:1; Difco, Lawrence, KS, USA) containing 1 mg/ml of M. tuberculosis strain H37 RA (Difco); 100 μl was injected into the hind footpads. Alternatively, the rats were injected intradermally, at the base of the tail, with an inoculum (total volume 200 μl) containing 50 μg of MBP_63–81 (AnaSpec Inc, San Jose, CA, USA) in saline mixed with the equivalent volume of complete Freund's adjuvant. Rats developing symptoms of EAE were subsequently treated with different liposome formulations, Copaxone, or placebo under similar conditions for 6 d. All test substances were administrated by single daily subcutaneous injections. The animals were followed up for 28 d after EAE induction, with the clinical syndromes being evaluated every day. Their disability status was scored as follows: 0, no symptoms; 1, tail weakness; 2, hind-limb weakness or paralysis; 3, hind-limb paralysis (dragging); 4, complete paralysis (inability to move); 5, death.

Histological analysis and staining for cytokines

Spinal cords of the animals were fixed, dehydrated, and embedded in paraffin. Histological sections were stained with hematoxylin–eosin and luxol fast blue and examined for the degrees of gliosis and demyelination. Gliosis was scored as follows: 0, no gliosis; 1, mild gliosis (5–10 cells/focus); 2, moderate gliosis (10–50 cells/focus); 3, severe gliosis (>50 cells/focus). The degree of demyelination was scored on the same scale, from 0 (no demyelination) to 3 (severe demyelination). Immunohistochemical staining for cytokines was performed using Abs to IL-2 (sc-7896; Santa Cruz Biotechnology, Santa Cruz, CA, USA), IFNγ (ab-9657; AbCam, Cambridge, UK), and BDNF (ab-72439, AbCam) according to the manufacturers' protocols. Histological analysis was done in a blinded fashion; the results obtained in each experimental series were compared with controls (preparations from nonimmunized rats and untreated rats with induced EAE).

ELISA and Western blot analysis

Horseradish peroxidase (HRP)-conjugated anti-rat (A9037), anti-mouse (A2554), and anti-human (A0170) Abs were from Sigma-Aldrich (St. Louis, MO, USA). Monoclonal mouse anti-c-myc IgG and anti-MBP F4A3 Abs were raised in our laboratory. In the case of F4A3, goat anti-IgM (mouse μ chain) Ab A8644 (Sigma-Aldrich) diluted 1:4000 was used for detection. ELISA was performed in MaxiSorp microtitration plates (Nunc, Roskilde, Denmark) as previously described (13). Odd-column wells were coated with MBP or Trx-fused MBP peptides (10 μg/ml in 100 mM carbonate/bicarbonate buffer, pH 9.0; 50 μl/well); the plate was sealed with ELISA plate sealer (Costar; R&D Systems, Minneapolis, MN, USA), incubated overnight at 4°C, and then washed 3 times with phosphate-buffered saline (PBS) containing 0.15% Tween-20 (PBS-Tween), 300 μl/well. All wells were then blocked by incubating with 250 μl of 2% bovine serum albumin (BSA; Sigma-Aldrich) in carbonate/bicarbonate buffer (pH 9.0) for 1 h at 37°C and washed with PBS-Tween. Sera were diluted 100- to 50,000-fold in PBS-Tween containing 0.5% BSA and 5 U.S. Pharmacopeia (USP) units of heparin per milliliter (in case of adsorbed MBP), and 50 μl of the diluted sample was added to each well. Rat anti-MBP monoclonal Ab (ab7349, AbCam) was used as a control. The plate was incubated for 1 h at 37°C, washed in 3 portions of PBS-Tween, and treated with corresponding HRP-conjugated Abs at a final dilution of 1:4000, 50 μl/well (1 h at 37°C). After washing in 5 portions of PBS-Tween, 50 μl of tetramethylbenzidine solution was added to each well, and the plate was placed in the dark for 5–15 min. The reaction was stopped by adding 50 μl/well of 10% phosphoric acid. The results were evaluated by measuring OD₄₅₀ values in a Varioscan microplate reader (Thermo Fisher Scientific, Rochester, NY, USA) and visualized in 2 or 3 dimensions using SigmaPlot 11.0 (Systat Software, Inc., Chicago, IL, USA).

Uptake of liposomes by dendritic cells

Dendritic cells from peripheral rat blood were isolated as previously described (14); 5× 10⁴ cells were incubated for 120 min with mannosylated and nonmannosylated liposomes in various concentrations (0–20 mg/ml) loaded with c-myc peptide. Further, cells were washed twice by PBS and lysed by TBS supplemented with 10% glycerol, 0.1% Nonidet P-40, 0.25 mM PMSF, 10 μM β-mercaptoethanol, and 0.2 mM EDTA. Lysates were subjected to Hybond C membrane (GE Healthcare Life Sciences, Piscataway, NJ, USA). Membrane was blocked using 2% BSA in PBS and further stained by anti-c-myc mAb (1:4000) and anti-actin mAb (sc-81178, 1:1000; Santa Cruz Biotechnology); afterward, it was developed by ECL Prime Western blotting Detection Reagent (GE Healthcare Life Sciences) and visualized using VersaDoc MP4000 (Bio-Rad Laboratories, Hercules, CA, USA).

Preparation of liposomes

MBP peptides were synthesized using a solid-phase technique, and their expected masses were confirmed by matrix-assisted laser desorption/ionization–time of flight (MALDI-TOF) mass spectrometry. Small unilamellar vesicles (SUVs) were prepared from egg phosphatidylcholine (PC) and monomannosyl dioleyl glycerol (ManDOG) conjugate (1:100 molar ratio) by high-pressure homogenization, as previously described (15), with certain modifications. Briefly, the lipid mixture (100 mg/ml) in CHCl₃ was dried under vacuum, resuspended in Milli-Q water (Millipore, Billerica, MA, USA) to a final lipid concentration of 50 mg/ml, and homogenized at high pressure (20,000 psi). The homogenate was mixed with peptides (lipid-to-peptide ratio 330:1) and excess of sugar (lactose-to-lipid ratio 3:1) and freeze-dried. At the next step, it was rehydrated under controlled conditions, and the resultant SUV liposomes were washed by centrifugation to remove nonincorporated material. The washed pellets were resuspended in PBS to the required dose volume. Peptide incorporation was estimated by reversed-phase HPLC on a C18 column with a linear gradient of acetonitrile. The z-average diameter and ζ potential of liposomes were measured in a ZetaPlus zetasizer (Brookhaven Instruments, Long Island, NY, USA) at 25°C by diluting 20 μl of the liposome suspension to the required volume with PBS or appropriate medium. Control peptide-free liposomes (vehicle) and liposomes lacking the mannosylated lipid were obtained in the same way but without adding MBP peptides or ManDOG, respectively.

Surface plasmon resonance (SPR) analysis

All measurements were made using a BiaCore T-200 instrument (GE Healthcare Life Sciences). Biotinylated MBP peptides (50 μg/ml; see Fig. 2B) and MBP were immobilized on SA and CM-5 chips, respectively. The procedure followed manufacturer's recommendations. The flow rate of HBS-EP buffer during all measurements was maintained at 10 μl/min. Each serum sample was serially diluted with HBS-EP buffer and tested on both chips at a standard association/dissociation time of 300/300 s. Dissociation constants were calculated using the BiaCore T-200 1.0 evaluation software.

Figure 2.

Characterization of specificity and affinity of polyclonal antibodies from DA rats immunized with MBP_63–81. A) Three MBP fragments recognized by serum autoAbs induced by immunization with MBP_63–81 in DA rats with EAE (top panel). Immunodominant peptides were identified on the basis of ELISA data and theoretical calculations. To verify the results of binding assay, the MBP epitope library was additionally hybridized with monoclonal Abs to c-myc and MBP (F4A3) and with control serum from nonimmunized DA rats (bottom panel). Anti-c-myc mAb proved to bind all members of MBP epitope library due to the presence of the target epitope in all fusion proteins, suggesting that all MBP peptides located directly upstream of the c-myc epitope are exposed and accessible. Monoclonal anti-MBP Ab (clone F4A3, MBP epitope RHGFLPRHR) reacted with peptides 17–41 and 25–54, as expected. Sera of nonimmunized rats did not show any detectable activity. All experiments were performed in triplicate, sd ≤ 5%. B) Quantitative characteristics of autoAb affinity to the identified MBP epitopes according to SPR data. Typical sensograms and effective dissociation constants for MBP and its peptides are shown. Exact epitopes are boldfaced; ND, not determined.

Statistical analysis

Disease scores were compared using the Wilcoxon test (1-way ANOVA for nonparametric statistics) in the SPSS 13.0 program (SPSS Inc., Chicago, IL, USA). Differences were considered significant at P < 0.05.

RESULTS

Recently, we designed an MBP epitope library consisting of fragments of this neuroantigen fused with Trx as a carrier protein and showed that the pattern of autoAbs binding to these epitopes may be regarded as a molecular signature of pathogenic B-cell response in MS (13). In this study, we used the MBP epitope library to reveal and compare the binding patterns of autoAbs contained in serum samples from patients with RR-MS and from rodents with EAE. The results are summarized in Fig. 1. As in our previous studies, the sera of patients with MS were found to contain polyclonal Abs specific to MBP fragments 42–65, 84–100, and 115–170. In three rodent strains used as models of induced EAE (see Materials and Methods), we revealed autoAbs to one immunodominant region (MBP_124–147) in C57/BL6 mice, two regions (MBP_24–44 and MBP_72–139) in SJL mice, and two regions (MBP_40–60 and MBP_107–170) in DA rats (Fig. 1). It can be seen that the last two regions are similar to those in patients with MS, indicating that DA rats with induced EAE are a relevant model of human MS with respect to the pattern of autoAbs toward MBP.

Figure 1.

The model of induced EAE in DA rats is most relevant to MS in terms of anti-MBP autoantibody binding pattern. AutoAbs contained in serum samples from patients with MS and rodents with EAE were tested by ELISA for binding to MBP epitope library. A) Individual binding patterns of autoAbs in samples from patients with MS (gradient color) are combined with those in samples from healthy donors (Hu, n=13) and rodents with EAE: SJL mice (SJL, n=15), DA rats (DA, n=11), and C57/BL6 mice (C57, n=14) (all in cyan, average data). Color scale (bottom left) indicates relative binding strength. TL, carrier protein without MBP peptide. B) Averaged binding patterns of autoAbs contained in test samples. MBP sequence and its division into peptides represented in the epitope library are shown at bottom. Brackets indicate immunodominant peptides (p1–p3) selected for screening their therapeutic efficiency. Letters (A–G) in both plots indicate MBP peptides inducing an increased autoAb response in animals with EAE.

We then searched for an alternative protocol of DA rat immunization to obtain reproducible, high levels of autoAbs not only to the C-terminal part of MBP but also to its encephalitogenic fragment. In previous experiments with fragments of the guinea pig myelin basic protein (GPBP), it was shown that only GPBP_62–84 and slightly GPBP_68–88 were capable to induce EAE in DA rats (16). Accordingly, we performed immunization of DA rats with peptide MBP_63–81 and, analyzing circulating Abs, revealed an enhanced autoimmune response to 3 MBP fragments: MDHARHGFLPRH, QDENPVVHFFKNIV, and IFKLGGRDSRSGSPMARR. Epitopes of polyclonal IgGs were identified according to the level of their binding to the MBP epitope library (determined by ELISA) and subsequent theoretical calculations based on the alignment of overlapping peptides (Fig. 2A).

To obtain quantitative characteristics of epitopes recognition by autoAbs and confirm in vitro their predicted sequence, we used the SPR method to measure the affinity of circulating polyclonal anti-MBP Abs from DA rats toward biotinylated MBP peptides containing specified epitopes (Fig. 2B). Values of effective dissociation constants for the full-size protein as well as for its encephalitogenic and C-terminal fragments were in the nanomolar range, confirming their role as major B-cell epitopes. The binding of autoAbs to the MDHARHGFLPRH peptide was not detected; therefore, this peptide was excluded from the set of biomarkers of EAE progression in DA rats further considered in this study.

Administration of MBP-derived immunodominant peptides encapsulated in mannosylated SUV (mSUV) liposomes significantly ameliorates EAE in DA rats

On the basis of these and our previous data (17), we selected 3 peptides, MBP_46–62 (p1), MBP_124–139 (p2), and MBP_147–170 (p3) (Fig. 1B), for evaluation of their therapeutic potential. Liposomes were prepared from egg PC with (mSUV) or without (SUV) 1% (molar) ManDOG (18). The procedure (Fig. 3A) included the following stages: evaporation of organic solvent from a mixture of lipids (Fig. 3Ai), resulting in the formation of irregular lipid layers (Fig. 3Aii); the first rehydration stage (Fig. 3Aiii), at which multilayer liposome vesicles are formed; high-pressure homogenization to obtain “empty” SUV liposomes (Fig. 3Aiv); freeze-drying of SUV in the presence of peptides, which at this stage are located outside collapsed liposome vesicles(Fig. 3Av); and the second rehydration stage (Fig. 3Avi), resulting in encapsulation of peptides into SUV ∼60–100 nm in diameter that carry mannose residues on their surface.

Figure 3.

Mannosylation of liposomes is critical for their therapeutic efficiency. A) Scheme of encapsulation of MBP peptides into mSUVs by liposomation technique (images created by the Visual Science Company, Moscow, Russia). i) A mixture of lipids (egg PC with 1% ManDOG) in chloroform. ii) Formation of irregular lipid layers during chloroform evaporation. iii) Rehydration, leading to the formation of multilamellar vesicles (MLVs) with an average diameter of 1–5 μM. iv) High-pressure homogenization, resulting in the formation of empty mSUVs. v) Freeze-drying of mSUV and peptide mixture with excess of sugar. Peptides are located outside collapsed liposomes. vi) Repeated rehydration, resulting in encapsulation of peptides into mSUVs with a diameter of ∼60–90 nm and mannose residues on the surface. B) Uptake of mannosylated and nonmannosylated liposomes loaded by c-myc peptide by dendritic cells. Dot-blot is shown on the left, quantitative analysis on the right. C) Mean disease scores of DA rats with EAE treated by the p1 peptide as such (n=8) or encapsulated in nonmannosylated (p1 SUV, n=10) and mannosylated liposomes (p1 mSUV, n=9) liposomes in comparison with the control group (no treatment, n=7). Schematic images of test preparations and their daily doses per rat are shown at top. The statistically significant difference in disease score between p1 SUV and p1 mSUV test groups (the cross-hatched area) is evidence that the presence of ManDOG in liposome composition is important for therapeutic efficiency.

To obtain quantitative data demonstrating uptake of the generated liposomal formulations, we isolated dendritic cells from peripheral rat blood and incubated them with mannosylated and nonmannosylated liposomes in various concentrations loaded with c-myc peptide (Fig. 3B). Our results suggest that dendritic cells uptake mannosylated liposomes more efficiently than nonmannosylated. Further, we tested whether encapsulation of MBP peptides into the liposomes and addition of the mannosylated component to the lipid composition had any beneficial therapeutic effect. To this end, DA rats with EAE induced by MBP_62–84 were subcutaneously injected with the p1 peptide as such or encapsulated in mannosylated or nonmannosylated liposomes (Fig. 3C). The peptide itself showed no detectable therapeutic effect, whereas the liposome-encapsulated formulations significantly inhibited EAE development. Moreover, mannosylated liposomes loaded with p1 were obviously more effective than nonmannosylated liposomes.

To analyze the complete set of selected MBP peptides, we prepared 4 basic formulations in which these three peptides and their mixtures in equivalent proportions were encapsulated in liposomes carrying mannose residues on their surface (Fig. 4A). The resulting liposomes were ∼85 nm in diameter, had a slightly negative charge (from −7.5 to −10.5 mV), and entrapped >90% of the initial peptide amount. To test these mSUVs for therapeutic efficiency, they were subcutaneously injected into the DA rats with induced EAE (Fig. 4B). This treatment was started on d 7 after EAE induction, i.e., at the time of the first clinical manifestations of the disease. Each rat received 6 single daily injections of the test mSUVs. Thereafter, all animal groups were examined to determine the mean disease score (MDS; Fig. 4B) and the median cumulative and maximum disease scores (Fig. 4C). Preparations of the spinal cord stained with hematoxylin-eosin were made to score the degrees of gliosis and demyelination (Fig. 4D). The p1 mSUV group was found to differ significantly from the placebo group (mSUV) in both maximum and cumulative disease scores, while no such difference was observed for the p2 mSUV or p3 mSUV group (Fig. 4C). Nonetheless, MDS profiles (Fig. 4B, left panel) and demyelination scores (Fig. 4D) indicated that final recovery in p2 mSUV and p3 mSUV groups was markedly better than in the p1 mSUV group. Notably, the p123 mSUV formulation proved to significantly ameliorate protracted EAE, lowering the overall disease profile and providing for full recovery from the disease (Fig. 4B, center panel); i.e., this mixture combined beneficial properties of individual peptides (Supplemental Fig. S1). The p123 mSUV group differed significantly from the placebo group in the median maximum disease score (Fig. 4C, center panel) and had the lowest median cumulative disease score (score 14; interquartile range 5.25) among all test groups (Fig. 4C, left panel). With respect to the parameter (delta) characterizing the magnitude of the overall difference in the MDS from the placebo group (Fig. 4C, right panel), the test groups could be arranged in the following ascending series: p1 ≪ p2 mSUV ≈ p3 mSUV < p1 mSUV ≈ Copaxone < p123 mSUV; in the last group, the value of this parameter was the highest (20.3 MDS×d). Copaxone, which was used as a positive control, was similar to the p123 mSUV formulation in the type of EAE suppression profile (Fig. 4B, center panel), and the group treated with this drug significantly differed from the placebo group in the median maximum disease score (Fig. 4C, center panel). As expected, the p1 peptide itself, without liposomes, had no therapeutic effect (Fig. 4B, right panel; C, D).

Figure 4.

Liposome-encapsulated immunodominant MBP peptides suppress experimental autoimmune encephalomyelitis in DA rats. A) Preparations tested for therapeutic effect, compared to placebo (empty mSUV), and their daily doses per animal (for details, see the text). B) Mean ± se disease scores in corresponding groups of DA rats with EAE. Encapsulated peptide p1 mSUV was most effective in reducing the disease score during the first attack, while p2 mSUV and p3 mSUV prevented the development of the exacerbation stage (left panel). Encapsulated peptide mixture p123 mSUV, similar to Copaxone, significantly suppressed protracted EAE (center panel). Peptide p1 without liposomes showed no detectable therapeutic activity (right panel). C) Median cumulative and median maximum disease scores (left and center panels) showed statistically significant differences between test groups and the placebo group (in each case, P value is indicated). Bars on both panels show the interquartile range; NS, nonsignificant. Right panel illustrates statistically significant differences between MDS profiles of the placebo group and individual test groups. *P < 0.05. D) Representative disease profile of an individual rat (black line) combined with the mean disease scores for the placebo group (red) and test groups (green). Bottom panels show typical histological images of the spinal cord in sections stained with hematoxylin–eosin (HE) at different magnifications; corresponding gliosis/demyelination scores are shown.

Liposome-encapsulated immunodominant MBP peptides suppress EAE development by down-regulating Th1 cytokines and inducing BDNF production in the CNS

To clarify the immunological status of test animals, we analyzed them for serum levels of anti-MBP autoAbs and performed staining for cytokines in the CNS (Fig. 5). A significant decrease in the concentration of autoAbs to full-length MBP, compared to the control group, was observed only in rats treated with p1 mSUV, p123 mSUV, and Copaxone (Fig. 5A, left panel). The production of autoAbs to both major MBP epitopes, 81–103 and 146–170, was also down-regulated in these groups of rats (Fig. 5A, center and right panels). The MBP_81–103 fragment was not included in any of liposome formulations, but the concentration of Abs specific to this fragment was reduced to the same extent as that of anti-MBP IgG.

Figure 5.

Liposome-encapsulated immunodominant MBP peptides decrease serum anti-MBP autoantibody titer, down-regulate Th1 CNS cytokine profile, and enhance BDNF expression. A) Concentrations of anti-MBP autoAb in sera of EAE DA rats treated with p1 mSUV, p123 mSUV, and Copaxone in comparison with the placebo group and nonimmunized rats. Error bars = sd. B) Typical histological images of the spinal cord in treated and untreated rats in sections stained with luxol fast blue and immunostained for Th1 cytokines (IL-2, IFNγ) and BDNF. Arrowheads indicate sites of decreased demyelination, down-regulation of Th1 cytokines, and increased BDNF production in rats of test groups, compared to the placebo group.

Staining for Th1 cytokines in the spinal cord of rats treated with p1 mSUV, p123 mSUV, and Copaxone showed that IL-2 and IFNγ were significantly down-regulated in all three groups (Fig. 5B), indicating that the corresponding formulations may be effective as anti-inflammatory drugs in EAE treatment. As followed from the results of luxol fast blue staining, the degree of demyelination in these groups of rats was also reduced. Suppression of neurodegeneration was correlated with enhanced production of BDNF, which was distinctly observed in the CNS of treated rats (Fig. 5B).

DISCUSSION

EAE may be induced in various species by immunization with myelin antigens. Such models have been used in a number of studies on MS, but they are not fully relevant to this disease. A number of therapeutics tested on animals with induced EAE have failed to show any beneficial effect and even caused exacerbation in patients with MS (19). Such cases can be partially explained by the fact that the B-cell response may vary from one EAE model to another as well as depending on differences in the immunization protocol. Therefore, we began our study with careful assessment of EAE models with regard to the spectra of anti-MBP autoAbs in order to verify their immunological relevance to MS and to select appropriate biomarkers for subsequent analysis of response to the applied therapy. Analysis of serum samples from patients with MS showed that they contain polyclonal autoAbs specific to 3 MBP fragments (Fig. 1). It is noteworthy that antibody profiling of cerebrospinal fluid samples from patients with MS provided evidence for the presence of autoAbs against 6 MBP peptides: 124–143, 146–170, 154–170, 10–29, 41–59, and 80–99 (10). This is in line with our data, as 5 out of 6 MBP fragments are overlapped with those discovered by us.

The pattern of anti-MBP autoAbs most similar to that in MS, including enhanced immunoreactivity to the MBP C-terminal peptide, was revealed in DA rats immunized with the spinal cord homogenate. However, taking into account the role of encephalitogenic peptide MBP_81–104 in disease evaluation (20), we used an alternative protocol of DA rat immunization to obtain reproducibly high levels of autoAbs to the encephalitogenic fragment of MBP and also to its C-terminal fragment. Moreover, we refrained from immunization with the spinal cord homogenate in order to make sure that epitope spreading, a hallmark of MS, does take place in EAE pathogenesis. The proposed scheme of immunization with the MBP_63–81 peptide resulted in production of Abs to the required immunodominant epitopes, MBP_81–94 and MBP_153–170. The affinity of these autoAbs toward the full-size protein and its fragments, measured by the SPR method, varied from 8 to 20 nM. It is noteworthy that we failed to detect any significant autoAb activity against the MBP_63–81 peptide used as the antigen, which was evidence for epitope spreading during EAE development. This observation is also in agreement with the original finding that MBP_62–75, being a major encephalitogenic peptide in DA rats, is not immunodominant in the sense defined by Sercarz et al. (21).

Mannosylation of liposomes is critical for their therapeutic effect

Encapsulated myelin autoantigens were previously found to be effective in treating EAE in Lewis rats (22, 23); on the other hand, administration of mannosylated APL M-PLP_139–151 proved to induce peptide-specific tolerance to EAE in SJL mice (7). We considered that a combination of these approaches might result in increased treatment efficiency. A major advantage of the combined approach is that immunodominant peptides are contained in the native form inside the liposome, while their targeted delivery to antigen-presenting cells is accomplished due to surface-exposed mannose residues. These cells carry increased numbers of mannose receptors on their surface, which facilitates endocytosis of mannosylated liposome particles into the cytosol (24). The efficiency of antigen presentation in mannosylated form, compared to native form, may be up to 10,000 times higher (25, 26). The results of our study confirm that exposure of mannose residues on the surface of liposome carriers is critical for their uptake by dendritic cells and therapeutic efficiency: nonmannosylated liposomes have proved to be significantly less effective than mannosylated liposomes loaded with the same p1 peptide.

It should be emphasized that, in general, the composition of liposomes may be critical for their therapeutic effect; moreover, it must be controlled depending on the encapsulated agent (e.g., glucocorticoids or proteins). As shown previously, EAE can be successfully suppressed by injections of glucocorticoid-loaded liposomes, either slightly anionic (−5.0 mV, as in our study; ref. 27) or cationic (28). Interestingly, Cavaletti et al. (29) have recently demonstrated the ability of cationic but not anionic liposomes to target on sites of acute neuroinflammation in animals with experimental EAE. The results of our previous studies suggest, however, that administration of cationic mannose-free liposomes may significantly enhance the antibody response (15) and should be avoided in attempts to induce self-tolerance. Eventually, the choice of the lipid charge is dependent on the encapsulated agent, as in the case of cationic liposomes widely used for delivering negatively charged DNA (30).

MBP peptides encapsulated in mannosylated liposomes have anti-inflammatory, neuroprotective, and regenerative effects

In tests of liposome-encapsulated MBP peptides on DA rats developing EAE, p1 mSUV was the most efficient in reducing general severity of EAE, mainly during the first attack, while p2 mSUV and p3 mSUV effectively prevented the development of the exacerbation stage. To combine these activities, we used a mixture of the three MBP peptides encapsulated in mannosylated liposomes. Indeed, the combined p123 mSUV formulation proved to have the highest therapeutic effect, compared to Copaxone or liposomes loaded with single peptides. Rats treated by p123 mSUV almost completely recovered from EAE within a week after the first clinical manifestations of the disease, whereas untreated rodents remained severely disabled. The median cumulative disease score in this test group decreased by up to 40%, compared to that in the placebo group. Notably, injections of the designed formulations during our experiments were started as soon as EAE was defined symptomatically, in agreement with the mode of clinical application of anti-MS drugs.

Treatment with encapsulated MBP peptides had a variety of immunological consequences for rats with developing EAE. In particular, it provided for a 2-fold decrease in the concentration of circulating Abs to MBP and to its encephalitogenic fragment. Since this fragment was not included in any liposome formulation, we may regard the observed decrease in concentration of Abs as a result of systemic suppression of autoreactive B cells; direct peptide-mediated neutralization of pathogenic IgG in the bloodstream is unlikely in this case. Furthermore, histological analysis for cytokines in the CNS of rats treated with liposome-encapsulated MBP peptides revealed significant down-regulation of Th1 cytokines IL-2 and IFNγ. It is noteworthy in this context that high-avidity myelin-specific CD4⁺ T cells were found to be reactive toward MBP peptides 111–129 and 146–170 (31). These observations, together with our findings (32), provide evidence for significant cross-reactivity of T and B cells toward MBP epitopes. In line with this idea, it appears that the composition of MBP peptides described in the present study may also have an effect on autoreactive T cells.

Suppression of demyelination was correlated with an enhanced expression of BDNF in the spinal cords of treated animals. BDNF is known to induce remyelination, thereby providing for the survival of degenerating neurons. The enhanced BDNF expression in the CNS may indicate that liposome-encapsulated MBP peptides not only have an anti-inflammatory effect but also trigger neuroprotective and regenerative processes. Therefore, the therapeutic effect of proposed encapsulated MBP peptides appears to be similar to that of Copaxone, which is known to stimulate BDNF expression (6).

CONCLUSIONS

In our study, liposomes were administrated by subcutaneous injections, thus, it is unlikely that they by any chance will penetrate the BBB and reach the CNS (e.g., brain). A more realistic scenario is that liposomes act in the periphery, as was proposed for Copaxone (1). Several mechanisms may have a role in disease suppression, including effective harvesting of liposomes with surface-exposed mannose residues by antigen-presenting cells, which are rich in mannose receptors, with subsequent induction of immune suppression, probably via regulatory T cells. The results of this study show that MBP fragments encapsulated in mannosylated liposomes effectively suppress the development of EAE in DA rats, reducing the severity of the first attack and facilitating recovery from the exacerbation stage. They also down-regulate Th1 cytokines, induce BDNF expression in the CNS, and inhibit the production of anti-MBP autoAbs. The observed beneficial effect of MBP fragments on EAE progression in DA rats, together with immunological relevance of this animal model to human MS, suggests a possible innovation in MS treatment.