Altered peptide ligands of myelin basic protein (MBP_87–99) conjugated to reduced mannan modulate immune responses in mice

Abstract

Mutations of peptides to generate altered peptide ligands, capable of switching immune responses from T helper 1 (Th1) to T helper 2 (Th2), are promising candidates for the immunotherapy of autoimmune diseases such as multiple sclerosis (MS). We synthesized two mutant peptides from myelin basic protein 87–99 (MBP_87–99), an immunodominant peptide epitope identified in MS. Mutations of residues K⁹¹ and P⁹⁶, known to be critical T-cell receptor (TCR) contact sites, resulted in the mutant peptides [R⁹¹, A⁹⁶]MBP_87–99 and [A⁹¹, A⁹⁶]MBP_87–99. Immunization of mice with these altered peptide ligands emulsified in complete Freund’s adjuvant induced both interferon-γ (IFN-γ) and interleukin-4 (IL-4) responses compared with only IFN-γ responses induced to the native MBP_87–99 peptide. It was of interest that [R⁹¹, A⁹⁶]MBP_87–99 conjugated to reduced mannan induced 70% less IFN-γ compared with the native MBP_87–99 peptide. However, [A⁹¹, A⁹⁶]MBP_87–99 conjugated to reduced mannan did not induce IFN-γ-secreting T cells, but elicited very high levels of interleukin-4 (IL-4). Furthermore, antibodies generated to [A⁹¹, A⁹⁶]MBP_87–99 peptide conjugated to reduced mannan did not cross-react with the native MBP_87–99 peptide. By molecular modelling of the mutant peptides in complex with major histocompatibility complex (MHC) class II, I-A^s, novel interactions were noted. It is clear that the double-mutant peptide analogue [A⁹¹, A⁹⁶]MBP_87–99 conjugated to reduced mannan is able to divert immune responses from Th1 to Th2 and is a promising mutant peptide analogue for use in studies investigating potential treatments for MS.

Introduction

Myelin basic protein (MBP)^a is a major autoantigen in the autoimmune disease, multiple sclerosis (MS).¹ Consequently, T cells specific for the MBP_87–99 epitope (VHFFKNIVTPRTP) have been detected in the blood or cerebrospinal fluid of MS patients^2–5 and are related to the induction of MS in humans.^6,7 Immunization with the MBP_87–99 epitope induces experimental autoimmune encephalomyelitis (EAE) in SJL/J mice^8,9 and in Lewis rats.¹⁰

Studies have shown that amino acids F⁹⁰, N⁹², I⁹³ and V⁹⁴ interact with major histocompatibility complex (MHC), whilst amino acids K⁹¹, T⁹⁵ and P⁹⁶ interact with the T-cell receptor (TCR).^10–12 Single alanine substitutions at positions 91, 95 or 96 of MBP_87–99 were able to antagonize T-cell responses in vitro, with [A⁹⁵]MBP_87–99 and [A⁹⁶]MBP_87–99 being the most effective.¹⁰ However, in Lewis rats, [A⁹¹]MBP_87–99 was able to prevent and reverse the clinical signs of EAE whilst [A⁹⁵]MBP_87–99 and [A⁹⁶]MBP_87–99 did not.¹⁰ In addition, administration of the [A⁹¹]MBP_87–99 peptide analogue reduced the production of tumour necrosis factor-α (TNF-α) and interferon-γ (IFN-γ), cytokines responsible for inflammation.¹⁰

In another study, a single alanine substitution at position 96 to [A⁹⁶]MBP_87–99 prevented and reversed EAE. It was shown that [A⁹⁶]MBP_87–99 bound weakly to I-A^s and induced a lower response to the L10C1 clone, a T-cell line specific for the native MBP_87–99 peptide, which induces severe EAE.¹³ Single mutant analogues ([A⁹¹]MBP_87–99, [A⁹²]MBP_87–99 and [A⁹³]MBP_87–99) were not beneficial for the treatment of EAE in (PL/JxSJL/J)F1 mice.¹³ The mechanism of tolerance of the [A⁹¹]MBP_87–99 peptide analogue showed that in vitro it acts as a partial agonist for the L10C1 clone, and its therapeutic effect for EAE cannot be attributed to MHC or TCR antagonism.¹³ Furthermore, EAE induced from T cells specific to the MBP_1–11 epitope could be ameliorated when mice were injected with [A⁹⁶]MBP_87–99 and this effect was related to interleukin (IL)-4 secretion.¹³ Neutralization of IL-4 using an antibody to IL-4 was shown to block the therapeutic effect of the [A⁹⁶]MBP_87–99 peptide analogue.¹³ Moreover, the [A⁹⁶]MBP_87–99 peptide increased the IL-4 : TNF-α ratio.¹³ IL-4 has been found to be a potent inhibitor of TNF-α and is responsible for the inhibition of EAE.^14,15 Additionally, IL-10 and IL-13 can inhibit TNF-α production and EAE.

The single-mutant [A⁹¹]MBP_87–99 and [A⁹⁷]MBP_87–99 peptide analogues could block the development of EAE in SJL/J mice, ameliorate the symptoms of EAE and reduce the proliferation of a T-cell clone specific for the MBP_87–99 epitope by 70%. The non-encephalitogenic [A⁹¹]MBP_87–99 peptide analogue could also increase the IL-4 : IFN-γ or the IL-4 : IL-2 ratios.¹⁶ Conversely, the superagonist [A⁹⁷]MBP_87–99 peptide caused deletion of MBP_87–99-responding cells. Thus, a single substitution at different positions of the MBP_87–99 epitope plays an important role in the modulation of the immune response and could inhibit EAE.

We designed and synthesized linear MBP_87–99 peptide and double-mutant analogues containing substitutions of critical TCR contact residues.^17–20 Positions K⁹¹ and P⁹⁶ were replaced with alanine/alanine ([A⁹¹, A⁹⁶]MBP_87–99) or with arginine/alanine ([R⁹¹, A⁹⁶]MBP_87–99), respectively. The K⁹¹ and P⁹⁶ residues have also been predicted to be TCR contact residues when bound to H2 I-A^s.²¹ Molecular modelling studies with I-A^s suggested that the side-chains of the mutated residues at positions 91 and 96 were exposed to make contact with the TCR,^11,12 which is similar to HLA-DR2. In addition, similarly to MS, EAE susceptibility is dependent on the MHC background of the mouse, and different peptides are immunogenic and induce EAE in different mouse strains. The SJL/J mouse strain (H2^s haplotype) is commonly used for EAE, as numerous histopathological, clinical and immunological features resemble those of human MS. SJL/J mice do not express the H2-E-α chain, and therefore the only functional MHC class II molecule in this strain is H2-A^s (I-A^s). In the SJL/J mouse strain, residues from the encephalitogenic epitope MBP_81–100 have been shown to bind with high affinity. The minimum epitopes required for binding are represented by the peptides MBP_83–99 and MBP_87–99.²¹ Thus, the MBP_87–99 peptide is a potential candidate for the design of peptide analogues, which could be used to alter T-cell responses in the SJL/J mouse model.

In our previous studies, native linear MBP_87–99 peptide induced EAE in Lewis rats, which was inhibited by the linear double-mutant analogue [R⁹¹, A⁹⁶]MBP_87–99. In addition, [R⁹¹, A⁹⁶]MBP_87–99 increased the T helper 2(Th2) : T helper 1 (Th1) cytokine ratio in blood from MS patients in vitro and suppressed proliferation of the CD4⁺ T-cell line from MS patients.²² Single or double mutants of the longer MBP_83–99 peptide epitope conjugated to reduced mannan could divert immune responses from Th1 to Th2 in SJL/J mice.^{1,11,12,23,24} Reduced mannan targets C-type lectin receptors, including the mannose receptor on dendritic cells, and generates Th2 responses.^25–32

Herein, we examined cytokine induction by two double-mutant peptide analogues of the minimal MBP_87–99 peptide ([R⁹¹, A⁹⁶]MBP_87–99 and [A⁹¹, A⁹⁶]MBP_87–99) compared with the native peptide, when emulsified in complete Freund’s adjuvant (CFA) or when conjugated to reduced mannan. The [A⁹¹, A⁹⁶]MBP_87–99 peptide analogue conjugated to a suitable carrier (reduced mannan) did not induce IFN-γ-secreting T cells, but elicited very high levels of IL-4, and thus the diversion of the cytokine profile. As determined by structural analysis, the [A⁹¹, A⁹⁶]MBP_87–99 peptide showed the greatest differences in intermolecular hydrogen-bond interactions compared with [R⁹¹, A⁹⁶]MBP_87–99 and native MBP_87–99 peptides.

Materials and methods

Solid-phase peptide synthesis of peptide analogues

Peptides MBP_87–99 (VHFFKNIVTPRTP), [R⁹¹, A⁹⁶]MBP_87–99 (VHFFRNIVTARTP) and [A⁹¹, A⁹⁶]MBP_87–99 (VHFFANIVTARTP) were prepared on 2-chlorotrityl chloride resin (CLTR-Cl) using Fmoc/tBu methodology.^22,33–36. Preparative high-performance liquid chromatography (HPLC) for MBP_87–99, [R⁹¹, A⁹⁶]MBP_87–99 and [A⁹¹, A⁹⁶]MBP_87–99 peptide analogues was performed using a Lichrosorb RP-18 reverse-phase semipreparative column with 7 μm packing material. The peptides were > 95% pure, as analyzed by analytical reverse-phase (RP)-HPLC and electrospray ionization-mass spectrometry (ESI-MS) (Scheme 1).

Scheme 1.

(a) Analytical reverse-phase high-performance liquid chromatography (RP-HPLC) of purified linear [A⁹¹, A⁹⁶]MBP_87–99 analogue after its purification by semipreparative RP-HPLC and lyophilization. Column: Nucleosil C18, 250 × 4·6 mm, 5 μm. T_R: 17·3 min. Conditions: gradient 5% (B)–100% (B) in 35 min; flow rate 1 ml/min. [Eluents (A): Solution trifluoroacetic acid (TFA) in H₂O 0·08% (v/v), (B): Solution TFA in acetyl-nitrile (AcN) 0·08% (v/v)]. (b) Electrospray ionization-mass spectrometry (ESI-MS) of linear [A⁹¹, A⁹⁶]MBP_87–99 analogue. M⁺ : 1473·31, M+2H⁺/2 : 737·29.

Conjugation of reduced mannan to MBP_87–99 peptides via a keyhole limpet haemocyanin linker

MBP_87–99, [R⁹¹, A⁹⁶]MBP_87–99 or [A⁹¹, A⁹⁶]MBP_87–99 peptides were conjugated to keyhole limpet haemocyanin (KLH) via glutaraldehyde, which acts as a linker between mannan and peptide.²⁷ Mannan (14 mg from Saccharomyces cerevisiae; Sigma, Melbourne, Vic., Austraila) was dissolved in 1 ml of sodium phosphate buffer (pH 6·0), then 100 μl of 0·1 m sodium periodate (dissolved in pH 6·0 phosphate buffer) was added and the mixture was incubated at 4° for 1 hr in the dark. Ethanediol (10 μl) was added to the mixture, which was then incubated for a further 30 min at 4°. The resultant mixture (oxidized mannan) was passed through a PD-10 column (Sephadex G-25 M column; Amersham Biosciences, Melbourne, Vic., Austraila) pre-equilibrated in phosphate buffer (pH 9·0) and 2 ml of solution comprising oxidized mannan was collected. One milligram each of MBP_87–99–KLH, [A⁹¹, A⁹⁶]MBP_87–99–KLH or [R⁹¹, A⁹⁶]MBP_87–99–KLH peptides were added to 2 ml of an oxidized mannan solution and incubated overnight at room temperature in the dark. Conjugation occurs via Schiff base formation between free amino groups of KLH and oxidized mannan. Reduced mannan–KLH–MBP_87–99, reduced mannan–KLH–[R⁹¹, A⁹⁶]MBP_87–99 or reduced mannan–KLH–[A⁹¹, A⁹⁶]MBP_87–99 complexes were prepared by adding 1·0 mg of sodium borohydride for 6–8 hr at room temperature in the dark and were used without further purification, as previously described.^12,23,30,31 MBP peptide analogues were previously characterized by capillary electropheresis for conjugation to mannan³⁷ and by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE), followed by staining with Coomassie Brilliant Blue, Silver stain or Schiff’s reagent (data not shown). Peptides were 100% conjugated to reduced mannan.

Mice and immunizations

Female, 6–8-week-old SJL/J mice, used in this study, were purchased from the Walter and Eliza Hall Institute (Vic., Australia) and housed at the Biological Research Laboratory at Burnet Institute (Austin campus, Heidelberg, Australia) in accordance with the guidelines of the National Health and Medical Research Council (NIMRC) of Australia. For EAE experiments, mice were purchased from the Hellenic Pasteur Institute (Athens, Greece) and housed in the P3 facility of the B’ Neurological department of the AHEPA University Hospital, Aristotle University Medical School, in accordance with the National Institute of Health (NIH) guidelines. MBP_87–99, [R⁹¹, A⁹⁶]MBP_87–99 or [A⁹¹, A⁹⁶]MBP_87–99 peptides were dissolved in phosphate-buffered saline (PBS) and emulsified in an equal volume of CFA (Sigma). SJL/J mice were given one subcutaneous injection containing 50 μg of peptide. For immunizations with peptide–KLH–reduced mannan, SJL/J mice were immunized twice on days 0 and 14, intradermally (at the base of the tail), with 50 μg of MBP_87–99–KLH–reduced mannan, [R⁹¹, A⁹⁶]MBP_87–99–KLH–reduced mannan or [A⁹¹, A⁹⁶]MBP_87–99–KLH–reduced mannan conjugates. For EAE experiments, seven mice per group ([A⁹¹, A⁹⁶]MBP_87–99–KLH–reduced mannan, reduced mannan alone or the control PBS group) were given three intradermal injections in the base of the tail.

Enzyme-linked immunospot assay

The enzyme-linked immunospot (ELISpot) assay detects specific T-cell responses to antigens by measuring the secretion of specific cytokines from individual cells. Spleen cells from immunized SJL/J mice were isolated 28 days after immunization with CFA or 14 days after the last immunization with reduced mannan-peptides and were assessed using ELISpot assays for IFN-γ and IL-4 secretion by T cells. The IFN-γ ELISpot assay was performed on a MultiScreen-IP Filter Plate (MAIP S4510) with hydrophobic poly(vinylidene difluoride) (PVDF) filters (Millipore, Melbourne, Vic., Australia), while IL-4 ELISpot assays were performed on a MultiScreen-HA Filter Plate (MAHA S4510) with mixed cellulose-ester filters (Millipore). MAIP S4510 plates were prewetted with 50 μl of 70% ethanol, washed five times with 200 μl of sterile PBS and coated with 70 μl of 5 μg/ml of anti-IFN-γ capture antibody, AN18 (Mabtech, Melbourne, Vic., Australia) in PBS and incubated overnight at 4°. Seventy microlitres of 5 μg/ml anti-IL-4 capture antibody (Mabtech) was added directly to MAHA S4510 plates and incubated overnight at 4° without treatment with 70% ethanol. Following five washes with PBS, plates were blocked by the addition of 200 μl of culture medium (supplemented with 2·5% fetal calf serum) and incubated for 2 hr at 37°. The blocking medium was discarded and 10 μg/ml of MBP_87–99, [R⁹¹, A⁹⁶]MBP_87–99 or [A⁹¹, A⁹⁶]MBP_87–99 recall peptides were added into each defined well. Concanavalin A (ConA) (1·0 μg/ml) was used as an internal positive control and no peptide (cells alone) was used as a negative control. Triplicate wells were set up for each condition. Spleen cells (0·5 million cells) in 100 μl of culture medium were seeded into each well and incubated at 37° for 18 hr for IFN-γ production or at 37° for 24 hr for IL-4 production. The plates were washed five times with PBS/0·05% Tween 20, five times with PBS and then incubated for 2 hr at room temperature with anti-mouse IFN-γ or IL-4 monoclonal antibody–biotin. Plates were washed and streptavidin–alkaline phosphate (ALP) conjugate was added at 1·0 μg/ml and incubated for a further 2 hr at room temperature. Spots of activity were detected using a colorimetric AP kit (Bio-Rad, Hercules, CA) and counted using an AID ELISpot plate reader (Autoimmun Diagnostika GmbH, Heidelberg, Germany). Data are presented as mean spot-forming units (SFU) per 0·5 million cells ± standard error of the mean (SEM).

Enzyme-linked immunosorbent assay

Blood was collected and sera isolated from mice before and 4 weeks after immunization with CFA, or 14 days after the last immunization with reduced mannan conjugates. MBP_87–99 or [A⁹¹, A⁹⁶]MBP_87–99 peptides conjugated to bovine serum albumin (BSA) were coated onto polyvinyl chloride (PVC) microtiter plates at 10 μg/ml in 0·2 m NaHCO₃ buffer, pH 9·6, overnight at 4°. Non-specific binding was blocked with 2% BSA for 1·0 hr at room temperature. After washing (0·05% Tween-20/PBS), serial dilutions of sera were added and incubated for a further 2 hr at room temperature. The plates were washed and bound antibody was detected using horseradish peroxidase (HRP)-conjugated sheep anti-mouse IgG (1 : 1000 dilution in PBS) (Amersham, Melbourne, Vic., Australia) and developed using 2,2’-azino-di(3-ethylbenzthiazoline)6-sulfonic acid (ABTS) (Sigma, UK). Absorbance at 405 nm was recorded using a Fluostar Optima microplate reader (BMG Labtech, Offenburg, Germany).

EAE induction and clinical evaluation

EAE was induced in SJL mice after two immunizations. On the day of EAE induction, a further immunization with the conjugates was performed. EAE was induced with a subcutaneous injection of 150 μg of proteolipid protein (PLP) (Day 0) emulsified in 100 μl of an emulsion composed of 50 μl of PBS and 50 μl of CFA containing 2 mg/ml of Mycobacterium tuberculosis H37RA. One-hundred microlitres of the emulsion was injected into each hind-flank of each animal. In addition, mice were injected intraperitoneally with 200 ng of pertussis toxin (Sigma) diluted in 0·5 ml of filtered PBS (day 0) and with another 100 ng of pertussis toxin diluted in 0·5 ml of filtered PBS on day 2 post-induction. Mice were clinically evaluated and weighed daily. The clinical status of each mouse was graded using the following scale: 0, normal; 1, flail tail; 2, tail paralysis; 3, hind limb weakness sufficient to impair righting; 4, paraplegia; 5, paraplegia with forelimb paresis or plegia; 6, death from EAE.

Statistical analysis

Mean values were compared using the Student’s two-tailed t-test for all immunological analyses. A P-value threshold of < 0·01 indicates a statistically significant difference. For EAE experiments, statistical analysis of the data were performed using the spss 11·5 software (Gainesville, FL). Values are expressed as mean ± standard deviation (SD).

Molecular modeling

Molecular modelling was carried out using the HyperChem modelling package (version 7·52; HyperCube Inc. Gainesville, FL) as previously described.¹² The optimized potentials for liquid simulations (OPLS) force field was used for molecular mechanics geometry optimization. The MHC–peptide complexes were generated based on the crystal structure of the I-A^u complex with MBP_1–11 peptide [protein data bank (PDB) code 1K2D]. This template was chosen upon the consideration of nine relevant crystal structures, based on the combination of sequence identity to the target, crystal structure, resolution, and the degree of disruption to the peptide interaction residues, upon mutation.^11,12 Alignment of the MBP peptide within the MHC cleft was carried out based on the analysis of all possible MBP positions, and the following preferred binding register was deduced.^11,12

To produce mutated MBP peptides (see below), the lysine residue at position 91 and the proline residue at position 96 were mutated to A⁹¹ and A⁹⁶ ([A⁹¹, A⁹⁶]MBP_87–96) and R⁹¹ and A⁹⁶ ([R⁹¹, A⁹⁶]MBP_87–96), respectively. The complexes were then optimized and the intermolecular interactions in the complexes were studied using the program LigPlot.³⁸

1K2D	(G	G)	A1	S	Q	Y	R	P	S	Q8
MBP	V87	H	F	F	K	N	I	V	T	P96

Results

[R⁹¹, A⁹⁶]MBP_87–99 peptide conjugated to reduced mannan decreases IFN-γ production and generates high levels of IL-4

The ability of the native MBP_87–99 peptide and the double mutant analogue [R⁹¹, A⁹⁶]MBP_87–99 to induce immune responses in SJL/J mice was assessed using IFN-γ and IL-4 ELISpot assays. High levels of IFN-γ were generated to MBP_87–99 peptide, which were decreased by 30% (P< 0·01) when mice were immunized with [R⁹¹, A⁹⁶]MBP_87–99 peptide emulsified in CFA (Fig. 1a). However, the levels of IFN-γ were further decreased (70%) when [R⁹¹, A⁹⁶]MBP_87–99 peptide was conjugated to reduced mannan (P < 0·01) (Fig. 1a). Furthermore, immunization of mice with the native MBP_87–99 peptide either emulsified in CFA or conjugated to reduced mannan did not induce IL-4 cytokine-secreting T cells (Fig. 1b). It was of interest, however, that immunization of mice with [R⁹¹, A⁹⁶]MBP_87–99 peptide emulsified in CFA induced low levels of IL-4 (significantly above background, P < 0·01) and very high IL-4 levels when conjugated to reduced mannan (P < 0·01) (Fig. 1b). It is clear that [R⁹¹, A⁹⁶]MBP_87–99 significantly decreases IFN-γ levels and generates high levels of IL-4 when conjugated to reduced mannan.

Figure 1.

(a) Interferon-γ (IFN-γ; upper panel) and (b) interleukin-4 (IL-4; lower panel) responses in SJL/J mice immunized with either MBP_87–99 or [R⁹¹, A⁹⁶]MBP_87–99 peptides. Immunization with 50 μg of peptide emulsified in complete Freund’s adjuvant (CFA) (black bar) or conjugated to reduced mannan (white bar). Negative (background levels) are indicated as grey bars. IFN-γ or IL-4 responses are shown as spot-forming units (SFU) per 0·5 million cells ± standard error of mean. The results shown are representative of two experiments with three mice per group. MBP, myelin basic protein. **(P< 0·01).

[A⁹¹, A⁹⁶]MBP_87–99 peptide conjugated to reduced mannan diverts immune responses from Th1 (IFN-γ) to Th2 (IL-4)

The ability of the linear MBP_87–99 or [A⁹¹, A⁹⁶]MBP_87–99 peptides either emulsified in CFA or conjugated to reduced mannan, to induce T-cell responses after one or two immunizations, was assessed for IFN-γ and IL-4 secretion using ELISpot analysis. Mice immunized with MBP_87–99 peptide (either emulsified in CFA or conjugated to reduced mannan) produced high levels of IFN-γ-secreting T cells (P < 0·01) (Fig. 2a,b). The cytokine IL-4 was not induced after immunization with MBP_87–99 peptide prepared in either conjugate (Fig. 2a,b). Mice immunized with linear [A⁹¹, A⁹⁶]MBP_87–99 emulsified in CFA produced high levels of IL-4 (P< 0·01) (Fig. 3); however, high levels of IFN-γ were also produced (P < 0·01) (Fig. 2a). Interestingly, when [A⁹¹, A⁹⁶]MBP_87–99 peptide was conjugated to reduced mannan, high levels of IL-4 were induced (Fig. 3b) and no IFN-γ was detected (P < 0·01) (Fig. 2b). Overall, [A⁹¹, A⁹⁶]MBP_87–99 emulsified in CFA generated both Th1 and Th2 responses; however, the use of reduced mannan conjugated to [A⁹¹, A⁹⁶]MBP_87–99 was able to divert immune responses from Th1 to Th2. Peptide was omitted to serve as a negative control and ConA was used as an internal positive control, which consistently induced > 1000 SFU/0·5 million cells for both IFN-γ and IL-4 (data not shown).

Figure 2.

Interferon-γ (IFN-γ) responses in SJL/J mice immunized with either MBP_87–99 or [A⁹¹, A⁹⁶]MBP_87–99 peptides. (a) Immunization with 50 μg of peptide emulsified in complete Freund’s adjuvant (CFA) and (b) immunization with 50 μg of peptide conjugated to reduced mannan. IFN-γ responses are shown as spot-forming units (SFU) per 0·5 million cells ± standard error of mean. The results shown are representative of two experiments with three mice per group. MBP, myelin basic protein.

Figure 3.

Interleukin-4 (IL-4) responses in SJL/J mice immunized with either MBP_87–99 or [A⁹¹, A⁹⁶]MBP_87–99 peptides. (a) Immunization with 50 μg of peptide emulsified in complete Freund’s adjuvant (CFA) and (b) immunization with 50 μg of peptide conjugated to reduced mannan. IL-4 responses are shown as spot-forming units (SFU) per 0·5 million cells ± standard error of mean. The results shown are representative of two experiments with three mice per group. MBP, myelin basic protein.

T cells from SJL/J mice immunized with [A⁹¹, A⁹⁶]MBP_87–99 emulsified in CFA or conjugated to reduced mannan do not cross-stimulate with the native MBP_87–99 peptide

After immunization with the linear mutant [A⁹¹, A⁹⁶]MBP_87–99 peptide, either emulsified in CFA or conjugated to reduced mannan, T cells were examined (using an ELISpot assay) to determine whether they cross-stimulated the native MBP_87–99 peptide. T cells from mice immunized with linear [A⁹¹, A⁹⁶]MBP_87–99 peptide emulsified in CFA (Fig. 4a) or conjugated to reduced mannan (Fig. 4b) did not cross-stimulate with the native MBP_87–99 peptide (Fig. 4a,b). Peptide was omitted to serve as a negative control and ConA was used as an internal positive control, which consistently induced > 1000 SFU/0·5 million cells for both IFN-γ and IL-4 (data not shown). In addition, T cells from [A⁹¹, A⁹⁶]MBP_87–99 did not cross-stimulate with [R⁹¹, A⁹⁶]MBP_87–99 peptide, which acted as an additional negative (background) control (Fig. 4).

Figure 4.

Interferon-γ (IFN-γ) production by T cells from SJL/J mice immunized with the mutant peptide [A⁹¹, A⁹⁶]MBP_87–99 (a) emulsified in complete Freund’s adjuvant (CFA) or (b) conjugated to reduced mannan. Recall peptides were MBP_87–99, [A⁹¹, A⁹⁶]MBP_87–99 or [R⁹¹, A⁹⁶]MBP_87–99, or no peptide (negative) (x-axis). IFN-γ production is shown as spot-forming units (SFU)/0·5 million cells ± standard error of mean. The results shown are representative of two experiments with three mice per group. MBP, myelin basic protein.

Antibody responses to MBP_87–99 or [A⁹¹, A⁹⁶]MBP_87–99 peptides emulsified in CFA or conjugated to reduced mannan

The production of total IgG responses in mice immunized with MBP_87–99 or [A⁹¹, A⁹⁶]MBP_87–99 peptide analogues, either emulsified in CFA (Fig. 5a) or conjugated to reduced mannan (Fig. 5b), were measured using enzyme-linked immunosorbent assays (ELISAs). No IgG was generated in mice immunized with [A⁹¹, A⁹⁶]MBP_87–99 peptide emulsified in CFA, and only very low levels of IgG were generated to the native MBP_87–99 (Fig. 5a). However, high antibody levels were induced to both MBP_87–99 and [A⁹¹, A⁹⁶]MBP_87–99 when they were conjugated to reduced mannan (titre > 1 : 16 400) (Fig. 5b). In previous studies using MBP_83–99 modified peptides or MUC1 peptides conjugated to reduced mannan, the antibody isotype induced was of the IgG1 subtype. Conjugation of peptides to oxidized mannan induced antibodies of the IgG2a subtype. Hence, reduced mannan is a strong inducer of antibodies of the IgG1 subtype.^{11,12,23,25–32,39}

Figure 5.

SJL/J mice immunized with (a) MBP_87–99 or [A⁹¹, A⁹⁶]MBP_87–99 peptide analogues emulsified in complete Freund’s adjuvant (CFA), or, (b) MBP_87–99 or [A⁹¹, A⁹⁶]MBP_87–99 peptide analogues conjugated to reduced mannan. Total immunoglobulin G (IgG) levels were measured by enzyme-linked immunosorbent assay (ELISA) coating with each respective peptides conjugated to bovine serum albumin (BSA) or keyhole limpet haemocyanin (KLH). (c,d) Cross-reactive IgG levels were measured coating with native MBP_87–99 peptide and using sera from SJL/J mice immunized with MBP_87–99 or [A⁹¹, A⁹⁶]MBP_87–99 peptides (c) emulsified in CFA or (d) conjugated to reduced mannan. Error bars depict the standard error of the mean. The results shown are representative of two experiments with three mice per group. MBP, myelin basic protein.

Because higher levels of antibodies were generated to linear [A⁹¹, A⁹⁶]MBP_87–99 peptide when conjugated to reduced mannan, it was investigated whether sera from these mice cross-reacted with the native MBP_87–99 peptide. Sera from mice immunized with MBP_87–99 reacted with MBP_87–99 peptide (positive control), while antibodies from mice immunized with linear [A⁹¹, A⁹⁶]MBP_87–99 conjugated to reduced mannan did not cross-react with the native MBP_87–99 peptide at 1 : 500 and 1 : 1000 dilutions of sera (Fig. 5d). Moreover, mice immunized with [A⁹¹, A⁹⁶]MBP_87–99 emulsified in CFA also did not cross-react with the native MBP_87–99 peptide (Fig. 5c). Overall, while IgG antibodies were noted for [A⁹¹, A⁹⁶]MBP_87–99 conjugated to reduced mannan, they did not cross-react with the native MBP_87–99 peptide.

The altered responses induced by the [A⁹¹, A⁹⁶]MBP_87-99–KLH–reduced mannan complex do not protect mice against EAE via a bystander effect

All mice were clinically followed up to day 27 post-immunization in order to study both the first relapse and the remission phases of the disease. The clinical course of acute EAE is presented in Fig. 6. The mean daily clinical scores and body weights (data not shown) did not statistically differ at any time-point (P < 0·05). The mean maximal score for the PBS control group was 2·0 ± 1·4, that for the reduced mannan alone group was 2·2 ± 1·0 and that for the [A⁹¹, A⁹⁶]MBP_87–99–-KLH–-reduced mannan group was 1·7 ± 1·0, indicating a similar severity of disease in control and immunized mice. In addition, disease onset was not statistically different (Kaplan–Meier survival analysis, log rank P> 0·05) between the three groups (data not shown), indicating that immunization did not affect the onset of EAE. Overall, our data indicate that immunization with [A⁹¹, A⁹⁶]MBP_87–99–KLH–reduced mannan did not have any effect on the clinical course of PLP-induced EAE in SJL mice. These results imply that the [A⁹¹, A⁹⁶]MBP_87–99–KLH–reduced mannan-induced Th2 responses had no bystander effects on the PLP-induced immune responses in the SJL mice. Of note, we recently demonstrated that immunization of C57BL/6 mice with myelin oligodendrocyte glycoprotein_35–55–reduced mannan and induction of EAE with MOG_35–55 peptide protected mice against EAE (Tseveleki et al., submitted for publication). Likewise, immunization of PLP_139–151–reduced mannan and induction of EAE in SJL mice using PLP_139–151 peptide also protected against EAE symptoms. In addition, immunization of Lewis rats with MBP_83–99-–reduced mannan conjugates, and induction of EAE using an irrelevant peptide MBP_74–85 did not protect animals against EAE (Tseveleki et al., submitted for publication). In that study it was shown that the most likely mechanism of immunity and protection against EAE was via the induction of T-cell tolerance and not via a bystander effect. As previously demonstrated, native MBP_87–99 peptide induced EAE in Lewis rats, which was inhibited by the double mutant analogue [R⁹¹, A⁹⁶]MBP_87–99 [22]. In addition, [R⁹¹, A⁹⁶]MBP_87–99 increased the Th2/Th1 cytokine ratio in blood from MS patients in vitro and suppressed proliferation of the CD4⁺ T-cell line from MS patients.²² We are currently analyzing the mechanism of action of [A⁹¹, A⁹⁶]MBP_87–99–KLH–reduced mannan peptide conjugates in human T cells from MS patients (manuscript in preparation), which will lead to further information on the mechanism of action and on its in vivo effects. Furthermore, we are testing the in vivo effects of [A⁹¹, A⁹⁶]MBP_87–99-–KLH–reduced mannan peptide conjugates on EAE in Lewis rats using the native MBP_87–99 peptide to induce EAE.^40,41 A study by Weissert et al.,⁴² demonstrated that a single amino acid substitution at position 79 of the peptide MBP_68–85 dramatically altered protection against EAE in Lewis rats and hence was able to modulate disease. They showed that protection was not caused by a bystander effect and the mechanism of protection was not clear.

Figure 6.

SJL/J mice were immunized twice with [A⁹¹, A⁹⁶]MBP_87–99–keyhole limpet haemocyanin (KLH)–reduced mannan. On day 0, mice were immunized again and experimental autoimmune encephalomyelitis (EAE) was induced using PLP_139–151 peptide. The clinical score was measured until day 27.

Interactions of [A⁹¹, A⁹⁶]MBP_87–99 and [R⁹¹, A⁹⁶]MBP_87–99 peptides in complex with MHC class II, I-A^s

In the modelled H2 I-A^s complex of the MBP peptide, the residues V⁸⁷, F⁹⁰, N⁹² and T⁹⁵ are anchored in MHC pockets P1, P4, P6 and P9, respectively.^11,12. The residue K⁹¹ is pointing up from the MHC groove. The models indicate that mutations of the K⁹¹ residue do not cause major disruptions to the structures or to the intermolecular interactions between the peptide and the MHC cleft (Fig. 7). The effect of mutation at position 91 has been studied by analyzing intermolecular interactions in the mutated complexes. These are summarized in Tables 1 and 2. The intermolecular interactions in the modelled complex with ([R⁹¹, A⁹⁶]MBP_87–96) have been described previously.¹² The analogue ([A⁹¹, A⁹⁶]MBP_87–96) shows similar trends, as described below (Fig. 7).

Figure 7.

Models of major histocompatibility complex (MHC) class II I-A^s with bound myelin basic protein (MBP) peptide ligands. MHC-binding grooves are shown as ribbons (α-chain, pale green; β-chain, grey) with the bound peptides as stick representations: (a) MBP wild-type (K⁹¹) peptide in cyan; (b) [R⁹¹, A⁹⁶]MBP_87–99 mutant analogue in magenta; (c) [A⁹¹, A⁹⁶]MBP_87–99 mutant analogue in orange. (d) Peptides are shown as overlays with the corresponding MBP residue positions (87–96) indicated.

Table 1.

Hydrogen-bonding interactions in the modelled complexes of MBP_87–96 peptide and mutant analogues with H2 I-A^s

Peptide		MHC
Residue	Atom	Residue	Atom	MBP87–96	[A91,A96]MBP87–96	[R91,A96]MBP87–96
V87	N	S53(A)1	O	+	+	+
H88	N	N82(B)2	OD1	+	−	+
H88	NE2	E74(B)	OE2	+	+	+
H88	O	N82(B)	ND2	+	+	+
H88	NE2	T77(B)	OG1	−	+	−
F90	N	Y9(A)	O	+	−	+
K91/A91/R91	N	E74(B)	OE1	+	−	+
K91	NZ	Q70(B)	OE1	+	n/a3	+4
N92	N	N62(A)	OD1	+	+	+
N92	ND2	N62(A)	O	+	+	+
I93	N	Y30(B)	OH	+	+	+
V94	N	Y67(B)	OH	+	−	+
V94	O	Y67(B)	OH	+	−	+
T95	O	Y68(A)	OH	+	−	+
P96/A96	N	D57(B)	OD1	−	+	−

¹α-chain,

²β-chain,

³not applicable,

⁴the NH1 atom of R⁹¹ in the peptide makes this contact.

MBP, myelin basic protein.

Table 2.

Van der Waals interactions in the modelled complexes of MBP_87–96 peptide and mutant analogues with H2 I-A^s

Peptide residue	MHC residue	MBP87–96	[A91,A96]MBP87–96	[R91,A96]MBP87–96
V87	Y9(A)1	+	−	+
V87	F54(A)	+	+	+
V87	H81(B)2	+	−	−
V87	N82(B)	+	−	+
H88	E74(B)	+	+	+
H88	T77(B)	+	+	+
H88	V78(B)	+	+	+
F89	Y9(A)	+	−	+
F89	F54(A)	+	+	+
F89	G58(A)	+	+	+
F89	V78(B)	−	+	−
F90	F11(B)	+	+	+
F90	G13(B)	+	+	+
F90	E14(B)	+	+	+
F90	C15(B)	+	+	+
F90	V78(B)	+	+	−
F90	C79(B)	−	+	−
K91	Q70(B)	+	n/a3	n/a
R91/A91	F11(B)	n/a	+	+
R91	Y67(B)	n/a	n/a	+
N92	T65(A)	+	+	+
N92	F11(B)	+	+	+
I93	T65(A)	+	+	+
I93	Y61(B)	+	+	+
I93	Y67(B)	+	+	+
V94	T65(A)	+	−	+
V94	Y61(B)	−	+	−
V94	Y68(A)	+	+	+
T95	D57(B)	+	+	+
T95	T69(A)	−	+	+
T95	Y61(B)	−	+	−
P96	Y60(B)	+	n/a	n/a

¹α-chain,

²β-chain,

³not applicable.

MHC, major histocompatibility complex.

With respect to hydrogen-bonding interactions, the most pronounced effect includes the loss of contact made by the nitrogen atoms of K⁹¹ and F⁹⁰ (MBP). Also, [A⁹¹, A⁹⁶]MBP_87–96 does not show the contacts to the MHC residues N⁸²(B), Y⁶⁷(B) and Y⁶⁸(A) that are observed in the complex with the wild-type peptide. However, compensatory interactions are observed to the MHC residues T⁷⁷(B) and D⁵⁷(B). It must be noted that [A⁹¹, A⁹⁶]MBP_87–96 displays the greatest deviation from the intermolecular hydrogen-bond interactions of the wild-type complex, compared with all other mutants studied previously.^11,12. Interestingly, the differences in the van der Waals interactions made by the native peptide versus [A⁹¹, A⁹⁶]MBP_87–96 were less pronounced than observed with other analogues.^11,12 In this case, the loss of contact made by K⁹¹ is once more observed. To balance, the analogue exhibits the A⁹¹ (MBP) interactions with F¹¹(B) (MHC). Other losses of contacts by [A⁹¹, A⁹⁶]MBP_87–96 include V⁸⁷ to Y⁹(A), H⁸¹(B) and N⁸²(B), F⁸⁹ to Y⁹(A), and V⁹⁴ to T⁶⁵(A). All of these are compensated for by the interactions observed in the mutant complex, but not exhibited by the native MBP peptide itself: V⁸⁷ to F⁵⁴(A), F⁸⁹ to V⁷⁸(B), and V⁹⁴ to Y⁶¹(B), respectively.

Discussion

Activation of CD4⁺ T cells is initiated by the interaction between the TCR and a peptide antigen that is presented by MHC class II molecules, and the engagement of costimulatory molecules of antigen-presenting cells.⁴³ This process is followed by T-cell proliferation, stimulation of reactive T cells specific to the antigen and the secretion of relevant cytokines. Many studies have shown that peptides with mutations at critical TCR contact residues result in altered T-cell function.^44–46 In particular, altered peptide ligands (or mutant peptides) have been found to shift the balance of immune responses from Th1 to Th2.^44,47,48 Th1 responses (IFN-γ) involve pro-inflammatory cytokines that mediate autoimmune diseases, and Th2 responses (IL-4, IL-10) reduce IFN-γ secretion and other inflammatory cytokines, preventing autoimmunity.^15,49,50

The development of safe and effective vaccines and immunotherapeutic approaches against MS are being studied actively. To date, a number of clinical trials, investigating altered peptide ligands, have been undertaken in MS patients.^51,52 Even though the peptides were demonstrated to induce appropriate responses in preclinical studies, these clinical trials were discontinued because of adverse reactions.^51,52 In some patients, unanticipated cross-reactions were stimulated by the peptide analogues against the native peptide/protein.^51,52 Thus, further preclinical testing is required and new modified peptides need to be designed in order to develop an effective vaccine for MS. In the cancer setting, it was demonstrated that the mutation of the HLA-A2-derived peptide (from carcinoembryonic antigen) at position 6 generated a superagonist. It induced CD8⁺ T cells that cross-reacted with high concentrations of the native peptide but which, however, did not recognize carcinoembryonic antigen-expressing cancer cells.⁵³ Thus, there is a need for extensive analysis of tumour cross-recognition before any clinical use can be permitted of altered peptide ligands as vaccines. Therefore, we synthesized two linear mutant peptides, based on the short immunodominant epitope, MBP_87–99, with mutations at positions K⁹¹ and P⁹⁶ (K⁹¹ was modified to R⁹¹ or A⁹¹ and P⁹⁶ was modified to A⁹⁶). The mutant analogues were injected into SJL/J mice in order to examine their ability to shift immune responses from Th1 to Th2, and to investigate whether T cells and antibodies cross-reacted with the native peptide. An adjuvant (CFA) or a suitable carrier (reduced mannan) was used.

Immunization of SJL/J mice with native MBP_87–99 peptide emulsified in CFA generated high levels of IFN-γ, whilst the double-mutant analogue [R⁹¹, A⁹⁶]MBP_87–99 decreased IFN-γ secretion by 30%. These findings are similar to those previously published for the longer double-mutant peptide [R⁹¹, A⁹⁶]MBP_83–99.¹¹ It is of interest that the levels of IFN-γ were further decreased (70%) when [R⁹¹, A⁹⁶]MBP_87–99 peptide was conjugated to the carrier, reduced mannan. Furthermore, immunization of mice with the native MBP_87–99 peptide, which was either emulsified in CFA or conjugated to reduced mannan, did not induce IL-4 cytokine-secreting T cells. Thus, substitution of K⁹¹ and P⁹⁶ with R⁹¹ and A⁹⁶, respectively, could decrease IFN-γ levels, generate high levels of IL-4 when [R⁹¹, A⁹⁶]MBP_87–99 peptide was conjugated to reduced mannan, but there was still secretion of IFN-γ. By contrast, immunization with the double-mutant analogue [A⁹¹, A⁹⁶]MBP_87–99 conjugated to reduced mannan induced high levels of IL-4 and no IFN-γ was detected. Substitution of residues 91 and 96 with Ala ([A⁹¹, A⁹⁶]MBP_87–99) has the ability to generate IL-4; however, with the appropriate carrier, it is able to divert immune responses from IFN-γ to IL-4. Overall, the use of adjuvant (CFA) generates both Th1 and Th2 responses and does not seem to be beneficial, but the use of an appropriate carrier (such as reduced mannan) conjugated to [A⁹¹, A⁹⁶]MBP_87–99 was able to divert immune responses from Th1 to Th2. Most importantly, T cells secreting IFN-γ and IgG generated to the double mutant [A⁹¹, A⁹⁶]MBP_87–99 peptide did not cross-react with the native MBP_87–99 peptide. It is clear that the double mutant [A⁹¹, A⁹⁶]MBP_87–99 peptide analogue is a promising candidate for further studies for use in the immunotherapy of MS. Likewise, we recently demonstrated, using single amino acid mutations of the longer peptide MBP_83–99, that a single amino acid change of K⁹¹ to Y⁹¹ ([Y⁹¹]MBP_83–99), when conjugated to reduced mannan, also diverted immune responses from Th1 to Th2,¹² similarly to the [A⁹¹, A⁹⁶]MBP_87–99 peptide conjugated to reduced mannan. It is clear that the peptide sequence and length, and the type of adjuvant/carrier used, are important for eliciting the desired immune response to a peptide-based vaccine. The mechanism by which reduced mannan–MBP_87–99 analogues switch immune responses is currently under investigation and preliminary studies suggest that tolerance is induced and responses are modulated by IL-17 and regulatory T cells (Treg cells) after immunization (manuscript in preparation). Furthermore, using MUC1 peptides conjugated to mannan, the addition of other adjuvants, such as muramyl dipeptide (MDP), glutaminyl-muramyl dipeptide (GMDP), aluminium hydroxide and adjuprime had no effect on the immune responses induced. Hence, mannan is an effective carrier for immune induction.⁵⁴

Modelling of the ([A⁹¹, A⁹⁶]MBP₈₇₋₉₆) and ([R⁹¹, A⁹⁶]MBP_87–96) analogues, in complex with H2 I-A^s, revealed that the side-chains of the mutated residues are exposed to make contacts with the TCR and the mutations do not cause major disruptions to interactions between the peptides and the MHC cleft. It was noted that the backbone of [R⁹¹, A⁹⁶]MBP_87–96, [A⁹¹, A⁹⁶]MBP_87–96 and native MBP_87–96 peptides overlapped very closely and only minor conformational changes were observed in the amino acids at positions 91 and 96 (Fig. 1d). Not unexpectedly, certain specific contacts observed in the complex with the native peptide were lost in the complexes with the mutants (Tables 1 and 2). These were compensated for by novel interactions with MHC residues. Most notably, the ([A⁹¹, A⁹⁶]MBP_87–96) analogue displayed the greatest reduction of intermolecular hydrogen bond interactions with respect to the native complex, when compared with all other mutants previously studied.^11,12 This is not surprising given the extent of side-chain modification when an alanine residue is used instead of arginine and proline.

Peptides causing antagonism have been shown to have fewer hydrogen bond contacts between the peptide side-chains and the CDR3 loops of the TCR.⁵⁵ Loss of hydrogen bond contact can cause agonist or super-agonist peptides (hyper-stimulatory altered peptide ligands) to become antagonists. For example, a single amino acid mutation in vesicular stomatitis virus peptide (VSV8) (RGYVYQGL to RGYVYEGL) leads to antagonism of T-cell hybridomas specific to native VSV8. The crystal structure of this altered peptide ligand with H-2K^b demonstrated that a minor peptide modification induced a large biological effect.⁵⁶ The TCR, which recognizes VSV8 (RGYVYQGL) peptide and its altered peptide ligand (RGYVYEGL), was mutated by a single amino acid at the CDR3β loop, and this was able to modulate the TCR-antagonistic properties of an altered peptide ligand.⁵⁷ These examples validate the results of our modelling studies with respect to MBP mutant peptides, namely that even subtle conformational changes are sufficient to have profound biological effects.

Glossary

Abbreviations:

CFA

complete Freund’s adjuvant

ConA

concanavalin A

EAE

experimental autoimmune encephalomyelitis

ELISpot

enzyme-linked immunospot

HPLC

high-performance liquid chromatography

IFN-γ

interferon-γ

IL

interleukin

KLH

keyhole limpet haemocyanin

MBP

myelin basic protein

MHC

major histocompatibility complex

MS

multiple sclerosis

PBS

phosphate-buffered saline

SEM

standard error of the mean

SFU

spot-forming units

TCR

T-cell receptor

Th1

T helper 1

Th2

T helper 2

TNF-α

tumour necrosis factor-α

Disclosures

The authors have nothing to disclose.