A role for αβ T cells in the resistant phase of the Brown Norway rat model of vasculitis

Abstract

Administration of mercuric chloride (HgCl₂) to Brown Norway rats causes Th2 dominated autoimmunity including a caecal vasculitis. Disease peaks 14 days after starting HgCl₂ after which animals immunoregulate spontaneously. In a third phase, if animals are rechallenged with HgCl₂ 6 weeks later they appear resistant, developing only attenuated disease. Previous studies suggested a role for CD8⁺ cells as partial mediators of resistance but no groups had studied the role of αβ T cells, γδ T cells or natural killer (NK) cells in resistance. We used adoptive transfer and in vitro cell depletion to show that αβ T cells are also partially responsible for resistance. Donor animals were treated with HgCl₂ or saline and killed 21 days later. Cells from donor spleens were transferred into recipient animals which were challenged with HgCl₂ and killed 14 days later. Test recipients received spleen cells from HgCl₂-treated donors after in vitro depletion of one subset of cells. Recipients receiving spleen cells from saline-treated donors remained susceptible to HgCl₂-induced vasculitis; those receiving spleen cells from HgCl₂-treated donors were resistant. Animals receiving αβ T-cell-depleted spleen cells from HgCl₂-treated donors showed partial reversal of resistance. Our results suggest a role for αβ T cells in the resistant phase of the Brown Norway rat model of vasculitis.

Introduction

The human vasculitides are autoimmune diseases characterized by leucocytoclastic inflammation of blood vessel walls throughout the body resulting in severe multi-system disease [1]. They are characterized by a relapsing and remitting time course where, following a period of disease activity, a patient's condition may pass spontaneously into a remission phase where they appear refractory to the induction of their autoimmune disease. Traditionally, studies on vasculitis have focused on disease pathogenesis while therapies have consisted of largely non-specific immunosuppression with its associated risks of infection, malignancy and infertility. Relatively little is known of the immunoregulatory events that underlie remission. If we could understand and mimic such events we might be able to design alternative, less toxic therapies for these conditions.

Administration of mercuric chloride (HgCl₂) to Brown Norway (BN) rats causes a Th2-dominated autoimmune syndrome characterized by a huge rise in IgE concentrations [2], a caecal vasculitis [3] and the production of multiple auto-antibodies including those against the glomerular basement membrane (anti-GBM), collagen and the myeloperoxidase antigen of the neutrophil cytoplasm (MPO ANCA) [4–7]. Animals also develop arthritis, mucositis and significant weight loss [8]. Disease peaks 14 days after starting HgCl₂ injections and then remits spontaneously, even if HgCl₂ injections continue [2,4,6,7]. In a third phase, if animals are rechallenged with HgCl₂ 6 weeks after their first challenge they are relatively resistant, showing only attenuated serological changes, arthritis and caecal vasculitis [8–10].

Although comparison of animal models to human disease must be undertaken with caution this syndrome demonstrates raised IgE concentrations, gut vasculitis and MPO ANCA, and thus has some resemblance to human Churg–Strauss syndrome. The system is also of interest as a Th2-biased animal model of autoimmune disease in a field where traditionally autoimmunity has been associated with a Th1 bias, and Th2 activity with the immunoregulation of autoimmunity. The resistant phase of the model may be analogous to remission in human disease, where patients are exposed to stimuli which would normally provoke autoimmunity but have immunoregulated such that they are no longer responsive to such stimuli. We hope that by studying the model's resistant phase, we may gain insights into possible mechanisms of immunoregulation which may be relevant to the control of human disease.

Resistance in the Brown Norway rat model of vasculitis was described originally by Pusey et al., who demonstrated attenuated levels of anti-GBM antibodies after a second mercuric chloride challenge [9]. In later studies Bowman et al. used adoptive transfer to show that if spleen cells from rats killed 21 days after beginning mercuric chloride were transferred into naive recipients who were then challenged with HgCl₂ for a first time, those recipients also developed only attenuated levels of anti-GBM antibodies (no other aspects of autoimmunity were explored at this stage) [11]. Using adoptive transfer in combination with in vitro cell depletion, Bowman et al. went on to show that such immunoregulatory effects were mediated at least partly by a CD8⁺ population with no demonstrable role for a B cell population. They did not, however, study roles for any other cells in the mediation of resistance in the model.

Further support for a role for CD8⁺ cells in resistance was added using HgCl₂ rechallenge in the same animals. Animals underwent thymectomy and CD8⁺ depletion and were then challenged with HgCl₂. Their response to a first HgCl₂ challenge was found to be unaltered compared to control animals challenged with HgCl₂ but not previously CD8⁺ depleted. However, when animals were rechallenged with HgCl₂ 6 weeks later, the anti-GBM titres in the CD8⁺-depleted group were significantly higher than those in control animals, indicating at least partial reversal of resistance [12].

With advances in technology since these original studies and with the recent availability of an antirat γδ T cell antibody V65 [13] to add to those already available against CD8⁺ cells (OX8), natural killer (NK) cells (anti-CD161) and αβ T cells (R73), we felt it might now be possible to investigate the role of other cells in the mediation of resistance in this model. In this study we have re-established the adoptive transfer system originally described by Bowman et al. [11]. In addition, we have used in vitro cell depletion to show a role for the αβ T cell in the mediation of resistance in this model.

Materials and methods

Induction of the animal model of vasculitis

Male BN rats (150–400 g) were obtained from Harlan Olac (Bicester, UK), and were given food and water ad libitum and used in age-matched controls. In all experiments donor animals received five injections with either 1 mg/kg 0·1% HgCl₂ (Sigma, Poole, UK) or an equivalent volume of normal saline over an 8-day period. Approximately 3 weeks after the start of their challenge, donor animals were killed and their spleens harvested. Cells from saline-treated donors were transferred to negative control recipients, and cells from HgCl₂-treated donors to positive control recipients. In addition a number of spleens from the HgCl₂-treated donor animals underwent in vitro cell depletion using magnetic bead cell sorting to remove a particular subset of the spleen cells before being transferred into test recipients via intravenous injection into a tail vein. After 24 h rest, all recipient animals were challenged with five injections of mercuric chloride over an 8-day period. Recipient animals were bled and weighed at regular time-points and were killed at days 14 or 15 and caecal vasculitis scored. In some experiments arthritis was scored between days 12 and 15. Exact experimental protocols looking at the depletion of different cell subsets are shown in Table 1. Numbers of donor cells transferred were chosen following pilot experiments which established a threshold of 0·8 × 10⁸ HgCl₂-treated donor spleen cells for the consistent transfer of resistance to the induction of caecal vasculitis. All experiments involving animals at St George's Hospital Medical School receive local ethical approval prior to commencing work.

Table 1.

Individual experimental protocols for adoptive transfer studies.

Experiment	1	2	3	4	5
Day of adoptive transfer	21	21	21	25	21
No. donor cells per recipient (adjusted in αβ depleted group)	1·5 × 108	1·2 × 108	1·5 × 108	1·5 × 108	2 × 108
Recipient animal groups	Negative controls	Negative controls	Negative controls	Negative controls	Negative controls
	(n = 4)	(n = 3)	(n = 10)	(n = 6)	(n = 4)
	Positive controls	Positive controls	Positive controls	Positive controls	Positive controls
	(n = 7)	(n = 3)	(n = 10)	(n = 6)	(n = 4)
	αβ T cell depleted	NK T cell depleted	αβ T cell depleted	NK T cell depleted	γδ T cell depleted
	(n = 7)	(n = 4)	(n = 11)	(n = 5)	(n = 4)
	CD8+ T cell depleted	αβ T cell depleted		γδ T cell depleted
	(n = 9)	(n = 6)		(n = 6)
Animal kill day	15	14	15	15	14

Negative control recipients received spleen cells from saline treated donors. Positive control recipients received un-manipulated spleen cells from HgCl₂-treated donors.

Monoclonal antibodies

Anti-rat αβ T cell antibody (R73) and antirat γδ T cell antibody (V65) were derived from monoclonal antibody-producing hybridoma cell lines [purchased from European Collection of Animal Cell Cultures (R73) or received as a kind gift from Dr T. Hunig (V65)]. An ammonium sulphate cut was made from tissue culture supernatant and IgG₁ monoclonal antibodies purified by protein A affinity chromatography. The antirat NK cell antibody anti-CD161 was purchased directly from Serotec, Oxford, UK.

In vitro cell depletion using magnetic bead cell sorting

Magnetic bead cell depletion was performed using a Variomacs magnet and CS depletion columns (Miltenyi Biotech, Bergisch Gladbach, Germany) according to the manufacturer's specifications. Briefly, single spleen cell preparations were made and red cells removed by incubation with Boyle's medium prior to suspension in phosphate-buffered saline/1% bovine serum albumin/2 mm ethylenediaminetetra-acetic acid (PBS/BSA/EDTA) at 6 × 10⁷ cells per ml. For depletion of αβ T cell, γδ T cell and NK cell populations, incubation was carried out for 15 min on ice with appropriate IgG₁ monoclonal antibodies (NK cells using the anti-CD161 antibody at 10 µg/ml, γδ T cells using V65 at 5 µg/ml and αβ T cells using R73 at 10 µg/ml). Cells were washed with 10 times their volume of PBS/BSA/EDTA and following centrifugation (300 g for 10 min), the cell pellet resuspended in 80 µl PBS/BSA/EDTA per 10⁷ cells. Rat antimouse IgG₁ microbeads (Miltenyi Biotech) were added at 20 µl per 10⁷ cells and the mixture incubated on ice for 15 min. Following a further wash, cells were resuspended at 10⁸ cells per 500 µl PBS/BSA/EDTA ready for loading on to the depletion column. In depleting CD8⁺ cells, a direct method was used where single spleen cell preparations at a concentration of 10⁷ cells per 80 µl PBS/BSA/EDTA underwent a single 15-min refrigerated incubation with 20 µl per 10⁷ cells of a mouse antirat CD8 microbead (Miltenyi Biotech) before being washed, centrifuged and resuspended as for all other depletion protocols. Cells were then loaded onto the CS depletion column, which had been placed in a Variomacs magnetic field, and negative cell-depleted fractions were collected. Cells were then resuspended in RPMI-1640 (Gibco BRL, Paisley, Scotland, UK) and after standardization of cell numbers, were injected into recipient animals.

Where depletion of γδ T cells, NK cells or CD8⁺ cells was undertaken no allowance was made for the removal of these populations, as they represent a relatively small proportion of spleen lymphocytes (CD8⁺ cells represented approximately 2·7% of spleen lymphocytes in our experiments while NK cells formed 10% and γδ⁺ cells formed 1·5%). However, αβ T cells represented between 35 and 45% of the spleen lymphocyte population in our experiments. Thus a correction step was undertaken to ensure that test animals received donor spleen cell numbers equivalent only to the αβ T cell negative fraction of positive control animals. An intra-experimental predepletion fluorescence activated flow cytometry step was undertaken to determine the exact percentage of αβ T cell-positive spleen cells in any given experiment. Cell numbers returned to recipient animals in the test group were then adjusted to represent only the αβ T cell-negative fraction of those returned to the positive control group.

The efficiency of cell depletion was assessed using flow cytometric analysis of the cell population before and after depletion. For predepletion analysis an aliquot of cells was removed after the initial incubation with the appropriate monoclonal antibody. For post-depletion analysis, an aliquot of the negative fraction of cells was reincubated with the relevant antibody for 15 min on ice at concentrations already given. In each case a further aliquot of cells was incubated with the isotype IgG₁ control antibody MOPC-21 (Sigma, Poole, UK). Washes with PBS/BSA/0·1% azide were performed between each incubation. Cells stained with the control antibody MOPC-21 and those stained for αβ T cells, CD8⁺ T cells and NK cells underwent a single further incubation with 100 µl of fluoroscein isothyocyanate (FITC)-conjugated antimouse IgG₁ antibody (Fab 2) fragment (Sigma, Poole, UK) diluted 1 : 1000 before resuspension in 1% paraformaldehyde pH 7. In the case of the rare γδ T cell population, cells were incubated sequentially on ice for 15 min with (a) V65 at 5 µg/ml, (b) FITC-conjugated antimouse IgG₁ antibody (Fab2) fragment diluted 1 : 1000 and (c) phycoerythropoetin-conjugated anti-CD5 pan T cell marker (Serotec MCA 52PE). An additional incubation with 15% heat-inactivated mouse serum was added between steps (b) and (c) to prevent cross-reactivity between reagents. Final analysis was performed using a Beckton Dickinson system by gating on the lymphocyte population and counting between 10 000 and 20 000 events.

IgE enzyme-linked immunosorbent assay (ELISA)

Total IgE concentrations were measured by ELISA, as described previously [8].

Arthritis scores

Arthritis scores of wrist and ankle joints were recorded by an experienced observer (C.S.V.) blinded to treatment on the following scale [8]: 1, erythema only; 2, swelling confined to wrist or ankle joint; 3, swelling extending distally on to the fore or hind foot; and 4, gross swelling involving the whole of the fore or hind foot. Results were expressed either as an isolated score in each animal for a single final day of an experiment or as the cumulative score for several days between days 12 and 15 of the experiment.

Caecal vasculitis

Caecal vasculitis was scored at necropsy by an experienced observer (D.B.G.O.) blinded to treatment using a previously validated scale [14]. Both serosal and mucosal surfaces were assigned a score between 0 and 4, which were combined to give a final macroscopic score. Directed caecal biopsies were then stained with haematoxylin and eosin and scored by an experienced histopathologist (D.R.T.) blinded to treatment to give a microscopic caecal vasculitis score on a scale of 1–4 as described previously [15].

Statistical analysis

Statistical analysis was performed using the Instat 2 version 2·04a computer program (Graph Pad software). Analyses were performed using two-tailed probability criteria unless stated otherwise. Data were analysed using Mann–Whitney U, Kruskal–Wallis and Fisher's exact methods.

Results

Successful adoptive transfer of resistance to the induction of caecal vasculitis

In all experiments we have been able to reproduce transferable suppression in the BN rat model of vasculitis. We have demonstrated for the first time that adoptive transfer of spleen cells from HgCl₂ treated donor animals into naive recipients results in protection of recipients from the induction of caecal vasculitis upon first challenge with mercuric chloride. Combining normalized data (where vasculitis scores are expressed as a percentage of the maximum vasculitis score seen in that experiment) from positive and negative recipient control groups for experiments 1–5 (Fig. 1), we have shown transferable suppression to be a reproducible effect with recipients of spleen cells from HgCl₂-treated donor animals showing significantly lower macroscopic caecal vasculitis scores than recipients of spleen cells from saline-treated donor animals (P < 0·0001, Mann–Whitney U). In none of the experiments did the transfer of spleen cells from HgCl₂-treated donors result in resistance as measured by IgE serology or arthritis score (data not shown).

Fig. 1.

Normalized data for total macroscopic caecal vasculitis scores in recipient negative and positive control animals in experiments 1–5. Mann–Whitney U, P < 0·0001.

Efficiency of in vitro cell depletion

Efficient cell depletion was achieved in all experiments with greater than 98% removal of all relevant cell populations.

Effects of depletion of individual cell populations on the transfer of resistance to the induction of caecal vasculitis

The effect of αβ T cell depletion

This was studied in experiments 1, 2 and 3, where all three experiments showed that removal of αβ T cells from the HgCl₂-treated spleen cells transferred to recipients resulted in a partial reversal of resistance to the induction of caecal vasculitis. Representative data are shown from experiment 3 (Fig. 2); as with other experiments, animals receiving spleen cells from HgCl₂-treated donors had significantly lower macroscopic caecal vasculitis scores than those receiving spleen cells from saline-treated donors (Fisher's exact P < 0·0001). In addition, macroscopic caecal vasculitis scores in the group receiving donor spleen cells depleted of αβ T cells were significantly higher than those in animals receiving spleen cells from HgCl₂-treated donors that had not been in vitro depleted [Fisher's exact P = 0·035 (one-tailed)]. A similar picture was seen in microscopic caecal vasculitis scores (data not shown). In Fig. 3, we have combined normalized data from control groups and αβ T cell-depleted groups for experiments 1, 2 and 3 by expressing vasculitis scores as a percentage of the maximum score for macroscopic caecal vasculitis seen in that experiment. Data confirm that macroscopic caecal vasculitis scores in the αβ T cell depleted group were significantly higher than those seen in the positive control group given unmanipulated spleen cells from HgCl₂-treated donors (Mann–Whitney U, P = 0·026). Similar results were seen in microscopic caecal vasculitis scores although results only approached statistical significance (median normalized vasculitis score for animals receiving unmanipulated cells from HgCl₂-treated donors = 25% of maximum, median normalized vasculitis score for animals receiving αβ T cell-depleted cells from HgCl₂-treated donors = 62·5% of maximum: Mann–Whitney U, P = 0·062).

Fig. 2.

Total macroscopic caecal vasculitis scores in recipient control and test animals at day 15 in experiment 3 demonstrating partial reversal of resistance after removal of αβ T cells. Fisher's exact test, P = 0·035 (one-tailed) (comparing recipients of cells from HgCl₂-treated animals with recipients of αβ-depleted spleen cells from HgCl₂-treated animals).

Fig. 3.

Combined normalized data for total macroscopic caecal vasculitis scores in positive control and test animals in experiments 1, 2 and 3 demonstrating partial reversal of resistance after removal of αβ T cells. Mann–Whitney U, P < 0·0259.

The effect of CD8⁺ cell depletion

Results from experiment 1 did not show any statistically significant differences following depletion of a CD8⁺ population, but it is likely that this single study was underpowered (data not shown graphically). Macroscopic caecal vasculitis scores showed that while recipients of spleen cells from saline-treated donors (negative control animals) remained susceptible to vasculitis (median score 7), and those receiving spleen cells from HgCl₂-treated donors (positive control animals) were relatively protected (median score 1), those receiving spleen cells from HgCl₂-treated donors that had been CD8⁺-depleted showed partial reversal of such protection (median score 5). Similar results were seen in the microscopic results (median scores 4, 2·5 and 4, respectively).

The effect of NK cell depletion

This was studied in experiments 2 and 4. In experiment 2 (data not shown graphically), while negative control animals were susceptible to macroscopic caecal vasculitis (median score 3) and positive control animals were relatively protected (median score 0), the removal of NK cells from the HgCl₂-treated donor spleen cells did not result in reversal of resistance (median score 1). Similar results were seen in microscopic vaculitis data. In experiment 4, once again no statistically significant reversal of resistance was seen following the removal of NK cells.

The effect of γδ T cell depletion

This was studied in experiments 4 and 5. In experiment 5, once again recipients of spleen cells from saline-treated donors were susceptible to macroscopic caecal vasculitis (median score 7·5) while those receiving spleen cells from HgCl₂-treated donors were relatively protected (median score 1·5). The removal of γδ T cells from spleen cells taken from HgCl₂-treated donors did not result in any reversal of resistance to caecal vasculitis in recipients (median score 0·5). Similar results were seen in the microscopic caecal vasculitis scores (median scores 4, 1·75 and 1·5, respectively). In experiment 4, however, although once again negative control animals remained susceptible to macroscopic caecal vasculitis (median score 4) and positive control animals were protected (median score 0), removal of γδ T cells from the spleens of HgCl₂-treated donor animals did result in a statistically significant partial reversal of such resistance (median score 3·5) (Mann–Whitney U, P = 0·026). Similar results were seen in microscopic scores (median scores 3, 1·25 and 3, respectively), although this did not achieve statistical significance.

Weight loss in recipient animals during mercuric chloride challenge

Combined weight data from experiments 1, 2 and 3 (Fig. 4), all of which contained positive and negative control groups, and an αβ T cell-depleted test group show that positive control animals who received spleen cells from HgCl₂-treated donors and were protected from the induction of caecal vasculitis lost significantly more weight than negative control animals who received spleen cells from saline-treated donors and were not protected from caecal vasculitis (P < 0·05, Kruskal–Wallis). It is also of interest that animals receiving αβ T cell-depleted spleen cells from HgCl₂-treated donors lost an intermediate amount of weight, although results do not differ significantly from either control group.

Fig. 4.

Treatment of donor rat spleens prior to transfer to recipient animals. Weight loss (expressed as a percentage of starting weight) in recipient control and test animals seen during their HgCl₂ challenge in experiments 1, 2 and 3.

Discussion

We have added to the work of Bowman et al. [11] in re-establishing an adoptive transfer system where spleen cells from animals challenged with HgCl₂ beginning 21 days earlier are able to transfer to naive recipients a resistance to the induction of one aspect of the autoimmune syndrome. However, while Bowman et al. were able to demonstrate resistance as shown by lower anti-GBM antibody concentrations in resistant animals, we have demonstrated resistance to the induction of caecal vasculitis.

In addition, we have shown in three separate experiments (1, 2 and 3) that depletion of αβ T cells from the donor cells transferred to recipients' results in partial reversal of resistance to the induction of caecal vasculitis. Failure to achieve complete reversal (also the case in the original Bowman experiments [11]) has several possible explanations. A very small number of undepleted αβ T cells may have multiplied in the recipient into a significant immunoregulatory population perhaps supplemented by previously immature cells expressing only low levels of the αβ T cell receptor and therefore not removed by magnetic bead cell selection. Alternatively, a small residual regulatory population may have been able to induce endogenous immunoregulatory cells in the recipient host animal (infectious tolerance). Perhaps the most likely explanation, however, is that the effect is partial because another cell type is also involved in resistance.

While results on depletion of NK cells do not support any role for these cells in the mediation of the resistance, our results on a possible role for a γδ T cell are more complex and suggest a possible contribution from these cells. Results from experiment 5, where adoptive transfer took place at day 21, do not support any role for γδ T cells in resistance, yet those from experiment 4 (adoptive transfer at day 25) suggest that removal of γδ T cells results in partial reversal of resistance to the induction of caecal vasculitis. One possible explanation for the discrepancy in the results would be that the γδ CD8⁺ T cell population becomes active only in the mediation of resistance at a slightly later time-point compared to the αβ T cell population. It is of interest that in studies by Matheison et al. where adoptive transfer also took place much later (at day 44), the number of donor cells required to achieve resistance was markedly lower than in other studies published where adoptive transfer took place at day 21, clearly suggesting that either the number or the potency of regulatory cells does change with time [12].

In experiment 1, we set out to examine the role of CD8⁺ cell depletion on the transfer of resistance to the induction of caecal vasculitis. Although numbers are too small to be statistically significant, it is of interest that animals receiving the CD8⁺ cell depleted spleen cells also show partial reversal of resistance with results that are similar to those seen in animals receiving the αβ T cell-depleted spleen cells. However, although previous studies on CD8⁺ cell depletion had been the starting point of our work, these results do not demonstrate that our depleted αβ T cell is of a CD8⁺ variety and it is also possible that the resistance mediating cell identified is, in fact, a CD4⁺αβ T cell. In order to address this question in future experiments, we would need to perform a two-stage selection procedure first positively selecting for either the CD4⁺ or the CD8⁺ populations and only then performing the αβ⁺ T cell-depletion step.

Immunoregulatory roles for both αβ and γδ T cells have certainly been previously described in other models of allergy and autoimmunity. Macary et al. have described an αβ⁺CD8⁺ T cell able to down-regulate the IgE response in a model of allergy [16]. When Lister hooded rats are injected intraperitoneally with ovalbumin complexed to aluminium hydroxide, they produce not only a CD4⁺ effector population but also an αβ⁺ CD8⁺ regulatory population that can down-regulate both the magnitude and duration of the IgE response in an effect that is partially interferon-γ mediated. In experimental autoimmune encephalomyelitis, Chen et al. [17] and Miller et al. [18] have demonstrated both transforming growth factor (TGF)-β secreting CD4⁺αβ T cells and TGF-β secreting CD8⁺αβ T cells each capable of suppressing the CD4⁺ major basic protein (MBP)-reactive T cell population that mediates central nervous system damage in this animal model of multiple sclerosis.

More recently immunoregulatory roles for the γδ T cell have also been described. If Brown Norway rats are exposed to ovalbumin via an aerosol inhalation, they develop later resistance to intraperitoneal challenge, no longer developing a significant IgE response [19]. Adoptive transfer studies have shown the effect to be mediated by an interferon (IFN)-γ-producing γδ T cell with as few as 10³γδ T cells needed to transfer the effect. Studies by Peng et al. [20] have described a murine model of lupus where animals deficient in γδ T cells develop a more severe phenotype of disease with a polyclonal expansion of CD4⁺ cells, again suggesting a regulatory role for the γδ T cell.

In our adoptive transfer studies, we have focused particularly on regulation of an autoimmune response seen at a mucosal surface. Mucosal immunoregulation has been studied extensively by Powrie et al. in their work on a mouse model of inflammatory bowel disease. In this model severe combined immunodeficient mice are reconstituted with reciprocal T cell subsets that perform effector and regulatory functions. The regulatory CD45Rb^low CD4⁺CD25⁺ T cell subset is able to prevent induction of disease by a CD45Rb^hi CD4⁺ T cell subset using a mechanism that involves both the secretion of TGF-β[21] and signalling via the cytotoxic T lymphocyte associated antigen 4 (a negative regulator of T cell activation) [22]. Immune regulation was also dependent on interleukin (IL)-10 produced by the regulatory T cell themselves [23]. While regulatory T cells in general are currently enjoying a renewed interest, study of such CD4⁺ CD25⁺ regulatory T cells in particular has shown them to be important in the regulation of both mucosal and systemic autoimmunity as well as in the immunoregulation thought to be important in transplant tolerance [24,25]. Our model is of particular interest as it combines immunoregulation of both a systemic and a mucosal autoimmune response and also because in contrast to most models of autoimmunity it occurs in the setting of Th2 rather than Th1 activation. TGF-β may be a particularly important regulatory cytokine at mucosal surfaces and it would be of interest to study both whether the phenotype of our regulatory cell is also CD4⁺ CD25⁺ and if it is a TGF-β secreting cell.

In contrast to the results of Bowman et al. (who demonstrated resistance as shown by lower anti-GBM serology using transfer of 2 × 10⁸ donor cells per recipient) and Mathieson et al. [12] (who showed transfer of resistance in anti-GBM serology using as few as 2 × 10⁷ donor cells), in the experiments shown, we have not been able to demonstrate the transfer of resistance to a serological part of the autoimmune syndrome as measured by serum IgE concentrations. Differences in the serological results between the three groups may reflect differences in the animal houses (particularly differences in gut pathogen load) or differences in the timing of adoptive transfer, as certainly in the results published by Mathieson et al. [12] transfer from donor animals took place at 44 days after initial HgCl₂ challenge, at which time any resistance-mediating population may be either more numerous or more potent. It is also possible that in a model with such a strong Th2 bias, including an IL-4 driven IgE class switch, it may be more difficult to achieve resistance in IgE rather than IgG serology.

In one previous pilot experiment (data not shown) using adoptive transfer of 4 × 10⁸ donor spleen cells per recipient animal, we demonstrated lower anticollagen antibody and total IgE concentrations at day 14 in animals receiving cells from HgCl₂-treated donors compared to animals receiving cells from saline-treated donors. The cell numbers were, however, prohibitively large to proceed to perform cell depletion studies and the serological effect was not repeated in a subsequent dose titration experiment where a maximum of 2·5 × 10⁸ donor cells per recipient were transferred. We have also not been able to demonstrate transferable resistance to the induction of arthritis in the model. We have, however, for the first time successfully and consistently transferred resistance to the induction of caecal vasculitis.

We have observed repeatedly that resistance to the induction of caecal vasculitis is considerably easier to achieve than either resistance to serological changes or resistance to the induction of arthritis. This may reflect the unique nature of the intestinal mucosa which, by necessity, is biased towards immunoregulation to prevent inappropriate and destructive immune responses to harmless intra-intestinal antigens that are presented every day. The gut associated lymphoid tissue (GALT) contains unique cell phenotypes including a greater proportion of CD8⁺ cells (thought to play an immunoregulatory role) compared to its systemic counterpart [26]. In addition, GALT lymphocytes have unique homing receptors ensuring return of these cells to the gut after systemic surveillance [27]. Thus immunoregulatory cells from a donor animal might target this anatomically discrete area for prevention of vasculitis even where they lack sufficient number and potency to alter the more systemic serological response. To investigate this further donor splenocytes could be labelled with CFSE (carboxy-fluorescein diacetate succinimidyl ester) in order to follow their anatomical fate in recipient animals. The time-scale of vasculitis may also favour its control in the adoptive transfer model. The serological response is reflected in significantly raised IgE levels by day 5 after the start of an HgCl₂ challenge [11], yet late caecal vasculitis is rarely seen prior to 12 days after starting HgCl₂[3], therefore giving greater amounts of time for regulatory cells to multiply, interact with endogenous cells or home to specific anatomical areas.

While we can only speculate on the mechanism of resistance in this model it is of interest, and is contrary to expectation, that resistant recipients protected from the induction of caecal vasculitis actually lost significantly more weight than control animals who developed more severe vasculitis. While the αβ-depleted group do not show weight changes that are statistically significantly different to either of the control groups it is noted none the less that these animals lose an intermediate amount of weight compared to either of the other two groups. Because severity of gut injury cannot explain this, one can speculate that a cytokine perhaps involved in mediation of resistance might also be inducing cachexia in the positive control animals. Other rodent models have shown roles for TNF-α, IL-1 and IL-6 in the induction of cachexia and it would certainly be interesting to study the relative expression of these cytokines in resistant and non-resistant animals [28,29]. In particular one could consider performing intracellular cytokine staining on the αβ T cells transferred across to study whether there was increased expression of any of these cytokines following transfer of the donor cells.

In summary, we have demonstrated an adoptive transfer system allowing resistance to the induction of caecal vasculitis in the Brown Norway/HgCl₂ model of autoimmunity. We have also shown that such resistance can be at least partially reversed by the removal of the αβ T cell population from the spleen cells transferred. At present we can speculate only on the mechanism of resistance, the elucidation of which will require further studies.