IL-4 and IL-10 modulate autoimmune haemolytic anaemia in NZB mice

Abstract

New Zealand Black (NZB) mice spontaneously develop autoimmune haemolytic anaemia (AIHA). Here the effect of injecting NZB mice with plasmids encoding IL-4 (pIL-4) or IL-10 (pIL-10) on NZB disease was tested. Both constructs delayed the development of anaemia as judged by increased haematocrit values as compared with controls, but neither altered the IgG1 to IgG2 red blood cell (RBC) bound autoantibody levels. The increased haematocrit value was associated temporally with increased RBC bound IgG in NZB mice treated with pIL-10, but not pIL-4. By contrast, up-regulation of splenic macrophage FcγRIIb2 mRNA was associated temporally with increased haematocrit values in NZB mice given pIL-4. However, no such increase occurred in NZB mice that inhaled a peptide containing a dominant T-cell epitope, although this treatment is known to bias the autoimmune response towards Th2 and to reduce the severity of anaemia. It is considered that IL-4 treatment, in part, ameliorates NZB anaemia by increasing the expression of the inhibitory FcγRIIb2 and thereby reducing the capacity of splenic macrophages to phagocytose autoantibody coated RBC, but that this mechanism does not explain the beneficial effects of the inhaled peptide.

Introduction

There is a pressing need to develop specific immunotherapies for autoimmune diseases, and considerable attention is focused on the potential beneficial effects of deviating the T helper (Th) cytokine balance. However, before such treatments can become a reality for human disease, it is important to understand the mechanisms underlying the action of therapeutic cytokines, particularly in antibody-mediated conditions, where their effects are less well understood.

A number of cytokines including IL-4 and IL-10 are associated with amelioration of autoimmune diseases. Firstly there is ample evidence that manoeuvres to polarize pathogenic Th1 responses towards Th2 can protect against diseases such as experimental allergic encephalomyelitis (EAE) [1], pristane induced arthritis [2] and the spontaneous diabetes of NOD mice [3]. In the last example, direct evidence for the contribution of IL-4 is provided by the finding that the injection of IL-4 [4] or pancreatic expression of IL-4 [5] abrogates the insulitis and diabetes. Similarly, NOD IL-4 deficient mice were resistant to the protective effect of immunization with GAD peptides as compared with wild type NOD mice [3]. However, precisely how IL-4 exerts its effect is still debated.

Evidence exists that IL-10 can also contribute to protection against immune-mediated disease. For example, Powrie and colleagues showed that the inflammatory bowel disease in lymphopenic recipients of CD45RB^high cells can be prevented by cotransfer of CD45RB^low cells, but that this protection can be inhibited with anti-IL10 receptor (R) antibodies [6]. Similarly, the regulatory T-cell responses induced by intranasal administration of the N-terminal peptide of myelin basic protein, which prevent EAE, depend on IL-10 to mediate their effects, as judged by the abrogation of protection by anti-IL-10 treatment [7] and in IL-10 knock out mice [8]. These results are consistent with the accelerated onset and severity of EAE in IL-10 deficient mice [9,10]. Finally administration of IL-10 can suppress collagen-induced arthritis [11], the diabetes of NOD [12] mice as well as EAE [13].

Spontaneous AIHA in the NZB mouse is a classic example of antibody mediated disease in which to dissect the therapeutic effects of changes to the balance of Th cytokines. We have shown that the major target of the pathogenic red blood cell (RBC) autoantibodies in NZB mice is the anion channel protein Band 3 [14], and CD4⁺ T-cells from NZB mice respond to Band 3 [15]. More recently we demonstrated that Band 3 peptide 861–874, which is the predominant sequence recognized by NZB T-cells in vitro [16,17], bears a dominant helper epitope able to modulate the autoimmune haemolytic anaemia in vivo. In particular, inhalation of a soluble analogue (E861,K875) of peptide 861–874 deviated the autoimmune response towards a Th2 profile. There was a marked increase in the ratio of IL-4 to interferon-γ produced by splenic T-cells responding in vitro to either peptide 861–874 or Band 3 together with some increase in IL-10 production. Moreover, in mice that had received such treatment, the proportion of RBC-bound IgG molecules that were of the Th2-associated IgG1 isotype was also increased, and anaemia was less severe [18]. These findings together with those implicating IL-4 and IL-10 in protection from other autoimmune diseases, prompted us to determine if administering such cytokines to NZB could affect the course of their disease, and if so how?

Materials and methods

Mice

NZB (H-2^d) mice were maintained under specific pathogen free conditions in the animal facilities at the University of Bristol. All animal experiments complied with UK Home office regulations, and the ‘Principles of laboratory animal care’ (NIH) were followed throughout.

DNA vaccines

The plasmid pIL-10 and pIL-4 encode the full-length cytokine genes IL-10 and IL-4. pIRES was a parental plasmid used for construction and served as a control plasmid in this study. Plasmid DNA was purified from transformed Escherichia coli strain DH5α by Qiagen Plasmid Giga Kits (Qiagen, Hilden, Germany) according to the manufacturer's instructions and stored at −70°C as pellets. The DNA was reconstituted in sterile saline at a concentration of 2 mg/ml for experimental use.

Immunization

All mice were immunized at 6 weeks of age. In brief, 100 µg of plasmid DNA pIL-10 or pIL-4 in 50 µl of 0·9% sodium chloride was injected intramuscularly in the upper hind limbs of mice using a 29-gauge needle. Mice immunized with pIRES alone served as a control group.

Administration of peptide

A group of six weeks old NZB mice were lightly anaesthetized and given 25 µl of peptide 861–875 (E861, K865) in PBS (4 mg/ml) intranasally on 3 occasions (300 µg in total) at two-day intervals. Age and sex-matched mice receiving PBS acted as a control group.

Haematocrit values

Blood (70 µl) was obtained from the saphenous vein and collected in 1 mm heparinized tubes (Hawksley, West Sussex, UK). The tubes were spun and the haematocrit was determined as the relative height of the RBC column expressed in percent using a Hawksley Micro-haematocrite reader.

Measurement of IgG subclasses bound to RBC

The number of total IgG, IgG1 and IgG2a molecules bound to the surface of RBC was determined using a quantitative and sensitive cellular ELISA based on a direct enzyme-linked antiglobulin test (DELAT) developed in our in our laboratories [19]. Briefly, RBC were washed 6 times in PBS which was warmed to 37°C, and a 10% suspension of washed RBC was fixed in glutaraldehyde. Aliquots (50 µl) of a 2% suspension of fixed RBC were added to the wells of round-bottomed microtiter plates (Corning Inc., Corning, NY, USA) and incubated with sheep antimouse IgG, IgG1 or IgG2a antibodies (The Binding Site, Birmingham, UK) for 1 h at 37°C. After washing three times, the RBC were incubated with alkaline-phosphatase conjugated donkey antisheep antibody (Sigma) for 1 h at 37°C, washed a further three times and allowed to react with phosphatase substrate solution (p-nitrophenyl phosphate, Sigma) for 1 h at 37°C. The supernatant was transferred into flat-bottomed microtiter plates (Corning Inc.) and the absorbance was measured at 405 nm using a microplate reader (Bio-Rad Laboratories Inc., USA).

Macrophage isolation and RNA extraction

Mice were killed, their spleens removed and squeezed through 70 µm mesh. Macrophages were isolated by a cell adhesion method. Splenocytes were suspended in serum free medium (RPMI 1640, Gibco, UK) and incubated in a 6 well cell culture plate (Corning Inc.) for 3 h at 37°C in a 5% CO₂ incubator. Non-adherent cells were removed by washing the plate with RPMI 1640 and total RNA was isolated from adherent cells using TRIZOL Reagent (Invitrogen, Paisley, UK) according to the manufacturer's instructions. In brief, 1 ml of TRIZOL reagent was added directly to adherent cells and then incubated at room temperature for 5 min. Chloroform (0·2 ml) was added, followed by vigorous shaking and incubation at room temperature for 2–3 min. The samples were centrifuged for 15 min at 12 000 g at 4°C for phase separation. To precipitate RNA, 0·5 ml isopropyl alcohol was added to the aqueous phase, the samples were incubated at room temperature for 10 min and centrifuged for 15 min at 12 000 g. The RNA pellet was washed once by 75% ethanol, and briefly dried and then reconstituted in RNase-free water.

Semiquantative reverse transcriptase-polymerase chain reaction (RT-PCR)

Complementary DNA (cDNA) was generated by reverse transcription of 1 µg total RNA using the Promega (Madison, WI, USA) reverse transcription system according to the manufacturer's instructions. The cDNA was amplified in GeneAmp PCR System 9700 (Applied Biosystems, Foster City, CA, USA) using PCR master mix, 2X (Promega) with primers specific for FcγRII [20] sense: 5′-GCTGGAGGAA CAAACTACTGAACAG-3′, antisense: 3′-GCAGCTTCTTC CAGATCAGGAGGA-5′ and β-actin sense: 5′-GTTACCAAC TGGGACGACA-3′ antisense: 5′-TGGCCATCTCCTGCTC GAA-3′. House keeping gene (β-actin) was amplified at different numbers of cycles and the intensities of amplified cDNA were compared when the plateau phase was reached, confirming that same concentration of cDNA was used in each sample. The amplification profile for Fcγ RII involved 30 cycles of denaturation at 94°C for 1 min, annealing at 60°C for 1min and extension at 72°C for 1 min. PCR products were separated by electrophoresis in 2% agarose gels and visualized by ethidium-bromide staining and inspection under UVERSUS light using a dual intensity transilluminator (UVP, Upland, CA, USA).

Statistical analysis

Results were analysed by one-way analysis of variance (anova), and individual groups were compared with the control by Dunnet's multiple comparison test. All statistical calculations were performed using the Prism program, and a value of P < 0·05 was considered significant.

Results

Groups of NZB mice were injected with pIL-4, pIL-10, or control vector pIRES, and the development of anaemia compared. As can be seen from Fig. 1, the mean haematocrit of the mice given pIL-4 or pIL-10 was significantly higher than the control group 12 weeks after plasmid DNA administration. By contrast, at 21 weeks, only mice given pIL-4 had a higher haematocrit. NZB mice were also injected contemporaneously with saline but their haematocrit values (mean 52% ± 1 at 12 weeks after plasmid injection and 49% ± 3 at 21 weeks) were not significantly different from values for pIRES treated mice.

Fig. 1.

Comparison of the haematocrit values of individual NZB mice after they had been injected with pIRES (▪), pIL-10 (▴) or pIL-4 (♦). (a) 12 weeks: pIRES versus pIL-10, P < 0·01 and pIRES versus pIL4, P < 0·05 (One-way anova). (b) 21 weeks: pIL4 versus pIRES P > 0·01, pIL4 versus pIL-10 P > 0·05. The mice were aged 6 weeks at the time of injection.

Estimates were made of the RBC bound autoantibodies. At 6 weeks (Fig. 2) significantly higher levels of RBC bound IgG were found in mice given pIL-10 as compared with controls, but this difference was not maintained at later times. Again the values (mean 0·48 ± 0·6 at 6 weeks after plasmid injection mean 1·2 ± 0·25 at 12 weeks and 1·79 ± 0·26 at 21 weeks) for saline injected NZB mice were not significantly different from values for pIRES treated mice. As there is evidence that NZB derived IgG2a RBC autoantibodies are more pathogenic in mice than IgG1 RBC autoantibodies [21], and IL-4 is associated with the production of IgG1 rather than IgG 2a antibodies, the levels of these autoantibodies was also measured. However, no significant differences in the erythrocyte bound IgG1 to IgG2a ratio between the IL-10 or IL-4 treated groups and control (P > 0·05) was evident (Fig. 3).

Fig. 2.

Erythrocyte bound IgG in mice injected pIRES (▪), pIL-10 (▴) or pIL-4 (♦) after (a) 6 weeks (b) 12 weeks (c) 21 weeks. The erythrocyte bound was significantly higher in the pIL-10 group as compared with the control (P < 0·01) at 6 weeks.

Fig. 3.

Comparison of erythrocyte bound IgG1 to IgG2a ratio in mice injected with pIRES control (n = 8), pIL-10 (n = 9) or pIL-4 (n = 8) after (a) 6 weeks, (b) 12 weeks and (c) 21 weeks. No significant difference was observed between IL-10 or IL-4 and control (P > 0·05).

IL-4 is known to increase the expression of the inhibitory FcRγIIb2 on monocytic cells, and decrease their phagocytic capacity [22,23]. We therefore isolated splenic macrophages from the IL-4 treated and control groups of mice and measured their expression of FcRγII mRNA. It can be seen from Fig. 4 that the band corresponding to FcγRIIb2 mRNA is prominent in splenic macrophages isolated from IL-4 treated mice but is weak or absent in those from either untreated NZB or control plasmid treated mice. However, the intensity of the band corresponding to FcγRIIb1 mRNA appears similar or only slightly stronger in IL-4 treated mice in comparison to the control group or untreated NZB mice.

Fig. 4.

Expression of FcRγII mRNA in macrophages isolated from normal NZB mice or NZB mice treated with pIRES or pIL-4 for 21 weeks. Total RNA was isolated, reverse transcribed and analysed by PCR. β-actin was used to ensure inclusion of equal amounts of target cDNA per reaction. A negative control containing all reagents except cDNA, was included in all PCR reactions.

Since inhalation of the soluble analogue (E861,K875) of peptide 861–874 deviates the autoimmune response in NZB mice towards the production of Th2 cytokines and protects from anaemia, it may be that the expression of the inhibitory FcγRIIb2 on splenic macrophages is increased. To examine this possibility, splenic macrophages were isolated from peptide treated and control groups of mice and their expression of FcγRII mRNA measured. The results (Fig. 5) confirm that the band corresponding to FcRγIIb2 mRNA is weak in splenic macrophages isolated from NZB mice treated with saline, but reveal also that the band remains absent or weak in splenic macrophages from peptide treated mice.

Fig. 5.

Expression of FcRγII mRNA in macrophages isolated from NZB mice three weeks after they inhaled PBS or peptide 861–875. The mice were aged 6 weeks at the time of injection. Total RNA was isolated, reverse transcribed and analysed by PCR. β-actin was used to ensure inclusion of equal amounts of target cDNA per reaction. A negative control containing all reagents except cDNA, was included in all PCR reactions.

Discussion

These results show that development of anaemia is retarded in NZB mice treated with pIL-4 as compared with plasmid treated controls, and that this protection is mediated via a different mechanism than the beneficial effects of the Th2 switch seen after nasal peptide therapy.

It was thought that the therapeutic effect of the IL-4 treatment might be mediated by a change in the isotype of the RBC autoantibodies from the pathogenic IgG2a type to IgG1 [21], as seen after peptide immunotherapy. The results do not satisfy this hypothesis because no such change was observed. Consequently, another possible mechanism was investigated. Smith and his colleagues [24] demonstrated that a haplotype defined by deletions in the promoter of the gene encoding FcγRII is shared by a number of autoimmune strains of mice including those of the NZB strain. Moreover, a reduction of FcγRIIb2 mRNA in peritoneal macrophages from NZB mice as compared with BALB/c mice was associated with an increased ability to engulf antibody-coated sheep RBC [24]. Similarly, macrophages from FcγRII deficient mice exhibit an increase in their capacity to phagocytose antibody-coated pneumococci in vitro as compared with those from wild type mice [25]. Here it was found that up-regulation of splenic macrophage FcγRIIb2 mRNA was associated temporally with increased haemacrit values in NZB mice given pIL-4. Since IL-4 increases FcγRIIb expression on monocytic cells and decrease their phagocytic capacity [22,23], it seems reasonable to suggest that IL-4 reduces anaemia, at least in part, due to this effect.

The development of anaemia was also delayed, albeit for a shorter time period, in NZB mice treated with pIL-10. The amelioration in haematocrit was seen despite being associated temporally with an increase in RBC bound IgG. The latter finding, although surprising in the context of protection, is not without precedent because the administration of IL-10 (but not IL-4) to transgenic mice carrying immunoglobulin genes derived from a hybridoma producing RBC autoantibody in NZB mice increased the number of cells producing these antibodies in the peritoneal cavity [26]. Since B-1 cells are prominent in the peritoneal cavity it was suggested that some of the RBC autoantibody producing cells in conventional NZB mice might be derived from B-1 cells [27], a finding consistent with earlier work that RBC autoantibody production is delayed in NZB xid mice [28]. The current findings that IL-10 increases RBC autoantibody levels in vivo appear to support this hypothesis. Interestingly, IL-10 expression was linked to increased rheumatoid factor production and B-cell activation in rheumatoid arthritis [29], whereas as mentioned earlier IL-10 protects against animal arthritides such as collagen-induced arthritis [11]. How IL-10 affects the haematocrit values remains to be determined.

It was considered that inhalation of the soluble analogue of peptide 861–874 decreased the severity of NZB anaemia by deviating the autoimmune response towards a Th2 profile, with cytokines such as IL-4 inhibiting the production of pathogenic IgG2a RBC autoantibodies and increasing those of the IgG1 isotype [18]. Here an additional possibility is suggested, namely that IL-4 increases the expression of the inhibitory FcγRII on macrophages thereby reducing the capacity to phagocytose autoantibody coated RBC. However, tests on mice that had inhaled the soluble analogue revealed no increase in the expression of splenic macrophage FcγRIIb2 mRNA. It could be argued that insufficient time had elapsed after the treatment to allow any change to take place. We cannot exclude this possibility, but it is known that splenic T cells taken a week earlier from peptide treated mice produce IL-4 in response to both Band 3 and peptide 861–874. Thus, overall, the current results suggests that the mechanism of peptide therapy cannot result from the production of IL-4 alone, but is more likely due to other Th2 cytokines or IL-10 like regulatory cytokines acting together. The study also illustrates that the actions of systemically delivered cytokines do not necessarily replicate the effects of changes in the production of the same cytokines when secreted by Th cells in lymphoid tissues.