Indoleamine 2,3 dioxygenase contributes to transferable tolerance in rat red blood cell inducible model of experimental autoimmune haemolytic anaemia

Abstract

Autoimmune haemolytic anaemia (AIHA) is caused by autoantibodies against red blood cell (RBC) surface antigens that render RBC susceptible to Fc-mediated phagocytosis and complement-mediated lysis. Experimental AIHA can be induced by injection of rat RBC to naive mice, but a lymphocyte-mediated regulatory mechanism eventually suppresses the production of autoantibodies specific for mouse RBC. Critically, this tolerogenic response can be transferred to naive mice by splenocytes from the rat RBC-immunized mouse. Here we investigate whether indoleamine 2,3 dioxygenase (IDO) or the initiators of IDO cascade, including the cytotoxic T lymphocyte antigen (CTLA)-4 receptor and its soluble isoform, contribute to this tolerogenic mechanism. Splenocytes from experimental AIHA mice were transferred adoptively to naive mice under the cover of anti-CTLA-4, anti-soluble CTLA-4 antibodies or IDO inhibitor 1-methyl tryptophan (1-MT). Recipient mice were immunized with rat RBC and levels of antibody against self-RBC and rat-RBC were monitored. Our results indicate that transfer of tolerance to naive recipients is dependent upon IDO-mediated immunosuppression, as mice receiving previously tolerized splenocytes under the cover of 1-MT were refractory to tolerance and developed haemolytic disease upon further challenge with rat RBC. Initiators of IDO activity, CTLA-4 or soluble CTLA-4 did not mediate this tolerogenic process but, on their blockade, boosted antigen-specific effector immune responses.

Introduction

Autoimmune haemolytic anaemia (AIHA) is an autoimmune disease characterized by the presence of autoantibodies that bind to red blood cell (RBC) surface antigens 1,2. This results in a reduced lifespan and premature destruction of RBC through phagocytosis by splenic macrophages or complement-mediated lysis 1. Typically, the antibodies are of immunoglobulin (Ig)G isotype and react optimally at 37°C, hence warm-type antibodies as opposed to IgM cold autoantibodies, which are more reactive at 0°C and cause cold agglutinin disease 3. Although the aetiology of the disease is unknown, warm IgG antibody-mediated AIHA provides a well-characterized disease model with which to study the loss of immunological tolerance to self-antigens 3.

The New Zealand black (NZB) mouse strain has been considered a good animal model for the study of spontaneous AIHA. Mice from this strain are genetically predisposed to developing autoantibodies against self-erythrocytes leading to the development of haemolytic anaemia by 6 months of age 4,5. In otherwise normal mice, autoantibodies specific for self-RBC can be induced by weekly intraperitoneal (i.p.) injection of rat RBC to naive mice, a model developed by Playfair and Marshall-Clarke in 1973 6. A cross-reactive epitope, contained within the phylogenetically conserved erythrocyte anion channel, AE-1, also known as Band 3 7, precipitates pathogenic autoantibody responses, resulting in loss of tolerance to self-RBC surface antigens 8. Immunized mice develop autoantibodies specific for both rat and self-RBC, but this process also stimulates a lymphocyte-mediated regulatory mechanism that eventually suppresses the production of those autoantibodies specific for self-RBC while retaining the antibody response to rat RBC 9. This exquisitely specific tolerogenic response can be transferred adoptively to naive mice by i.p. injection of splenic cells from rat RBC immunized mice. Naive recipients of these RBC tolerant lymphocytes are then rendered resistant to developing autoantibodies specific for mouse RBC, following immunization with rat RBC, but retain xenogeneic rat RBC immunity 9.

In this model system, the mechanism by which splenic lymphocytes ‘transfer’ tolerance to the recipients remains unknown, although a role for CD4⁺CD25⁺ regulatory T cells has been suggested 10. Further, analysis of in-vitro assays where CD8⁺ T cells from young NZB mice were co-cultured with splenic cells from old, actively autoimmune NZB mice indicated that these CD8⁺ T cells were capable of suppressing autoantibody responses, evidence that they may also have a role in regulating disease 11. While identification of the T cell subset(s) that mediate and transmit tolerogenic responses in AIHA continues, it is also important to address, at a molecular level, mechanisms that underpin this important immunological feature.

Indoleamine 2,3-dioxygenase (IDO) is a tryptophan-depleting enzyme, expressed predominantly by macrophages and dendritic cells (DCs) that can have profound regulatory effects on T cell-mediated effector responses 12,13. It has been shown that treatment of pregnant female mice crossed with an allogeneic major histocompatibilty complex (MHC) mismatched strain, with the IDO inhibitor 1-methyl tryptophan (1-MT), could break the tolerance that protected the fetus from the maternal immune system, and acute disruption of IDO activity in this model either proved catastrophic to the fetus or enhanced disease pathology markedly 14. In a murine model of experimental autoimmune encephalomyelitis (EAE), blockade of IDO also induced exacerbation of clinical and histological disease parameters, suggesting that IDO contributes to the regulation of T cell activity associated with this animal model of multiple sclerosis 15. Recently, IDO has also been shown to engage in intracellular signalling events that allow self-amplification and maintenance of a stably regulatory phenotype in plasmacytoid DCs, a regulatory function of IDO that is mechanistically independent of its enzymatic activity 16. Based on these studies indicating that IDO may contribute to immunological tolerance, we asked if, in this model of AIHA, blockade of IDO would abolish the tolerance conferred by rat RBC immune splenocytes and allow the haemolytic disease to emerge.

One of the most effective initiators of IDO cascade is engagement of the T cell co-stimulation inhibitor, cytotoxic T lymphocyte antigen (CTLA)-4, whether membrane-bound as a receptor 17 or in recombinant soluble form (CTLA-4-Ig), with its ligands B7·1/B7·2 on macrophages and DCs 18. Therefore, we further assessed whether CTLA-4 or the alternatively spliced native soluble isoform of CTLA-4 (sCTLA-4) played a role in the tolerogenic process in this experimental model of transferable tolerance. Although antibody-mediated blockade of CTLA-4 or the soluble isoform did not abrogate tolerance conferred by immune splenocytes from previously tolerized animals, recipient mice treated with IDO inhibitor 1-MT did develop AIHA, showing that recipients devoid of this critical immunosuppressive enzyme activity are susceptible to the loss of immunological tolerance to self-antigens.

Materials and methods

Sprague–Dawley rats and BALB/C mice were provided by, and housed in, the Medical Research Facility, University of Aberdeen. The procedures adopted conformed to the regulations of Animal Scientific Procedure Act (UK), 1986. All work was carried out with UK Home Office project licence approval.

Induction of AIHA in BALB/C mice

Rat blood was drawn into a 10-ml syringe via cardiac puncture from an anaesthetized rat into a lithium heparin tube (BD Vacutainer Systems, Oxford, UK), transferred into a 50-ml centrifuge tube, filled with sterile phosphate-buffered saline (PBS) and centrifuged at 1800 g for 5 min at room temperature. The supernatant was discarded, RBC were washed twice with sterile PBS and cell number adjusted to a concentration of 1 × 10⁹ cells/ml. Normal 8–10-week-old BALB/C mice were injected i.p. with rat RBC (2 × 10⁸ cells in 200 μl sterile PBS) on four occasions at 7-day intervals. Control mice were given 200 μl sterile PBS i.p. Each experimental group contained eight mice.

Measure of autoantibodies by direct enzyme-linked anti-globulin test (DELAT)

Blood samples (10 μl) from immunized mice were collected via tail-nicking and stored in 50 μl Alsever's solution (Sigma-Aldrich, Poole, UK). Levels of RBC-bound IgG were determined by DELAT, as described previously 19. Blood was transferred to a round-bottomed 96-well plate, washed six times with PBS and fixed with 2% glutaraldehyde for 30 min at room temperature (RT). Cells were washed twice, suspended in 400 μl PBS and stored at −20°C pending use. Stored glutaraldehyde-fixed RBC were washed with 0·2% bovine serum albumin (BSA)/PBS before incubation with goat anti-mouse IgG (Sigma-Aldrich). Bound goat IgG was detected with alkaline-phosphatase conjugated anti-goat IgG (Sigma-Aldrich). Cells were again washed six times and allowed to react with 100 μl phosphatase substrate solution (p-nitrophenyl phosphatase; Sigma-Aldrich). Optical density (OD) of supernatant analytes was measured at 450 nm (Multiskan MS microplate photometer; Life and Laboratory Sciences, Basingstoke, UK). Glutaraldehyde-fixed RBCs from NZB mice that develop AIHA spontaneously were used as a positive control. Mouse antibody responses to rat RBC were measured by an indirect enzyme-linked anti-globulin test (IELAT) in which rat RBC were co-incubated with mouse serum prior to fixing with glutaraldehyde.

Preparation of splenic cell suspension and adoptive transfer of splenocytes

Spleens from immunized mice were removed and disrupted gently through a cell strainer to recover a splenocyte suspension. Whole splenocytes were washed twice with Hanks's balanced salt solution (HBSS) and the cell number was adjusted to 5 × 10⁷ cells/ml. Each naive recipient mouse was infused i.p. with 200 μl of splenic cell suspension in HBSS (1 × 10⁷ splenocytes) and these mice were subsequently challenged i.p. with 2 × 10⁸ rat RBC (as above) on four occasions at 7-day intervals.

Preparation of RBC ghosts

RBC were subjected to hypotonic lysis to deplete cell contents using ice-cold 20 mM Tris, pH 7·6, following centrifugation at 30 000 g for 30 min at 4°C to isolate the RBC membrane fraction.

Cell enzyme-linked immunosorbent assay (ELISA) and proliferation assay

Splenocytes, prepared as above, were resuspended in RPMI-1640 medium containing 1% autologous decomplemented serum, 100 U/100 μg/ml penicillin/streptomycin, 1% L-glutamine and 1 mM 2-mercaptoethanol. The cell number was adjusted to 2 × 10⁶ cells/ml. Cells were stimulated with increasing amounts of autologous or rat RBC ghosts and incubated at 37°C, 5% CO₂ for 5 days. Cell culture supernatant levels of interferon (IFN)-γ, (anti-IFN-γ clones AN-18 and R4-6A2), interleukin (IL)-10 (anti-IL-10 clones JES5-2A5 and SXC-1) and IL-17 (anti-IL-17 clones TC11-18H10·1 and TC11-8H4·1; BD Biosciences, Oxford, UK) were measured by ELISA, as described previously 20. Cell proliferation in cultures was measured by [³H]-thymidine incorporation in triplicate samples using a 1450 Microbeta liquid scintillation counter (LKB Wallac, Turku, Finland). Results are presented as mean counts per minute (cpm) ± standard deviation (s.d.).

Administration of 1-methyl tryptophan (1-MT)

1-MT (Sigma-Aldrich) was co-administered to mice transferred adoptively with splenocytes at a dose of 1 mg/ml in drinking water, commencing 3 days prior to the transfer of splenocytes to naive mice and maintained for 21 days. 1-MT was reconstituted in 0·1 M NaOH (with ultrapure water), pH 8·9 to give a final concentration of 1 mg/ml. In order to render the solution palatable an artificial sweetener, aspartame (Sigma-Aldrich), was dissolved (0·25 μg/ml) in reconstituted 1-MT water. Experimental mice received 1-MT water while control mice were given aspartame water alone (0·25 μg/ml) at pH 8·9 (experimental plan shown in Fig. 4a). The weights of mice and amounts of water consumed were monitored to ensure that no adverse effects were occurring. Each mouse drank 4–5 ml water per day, implying that each mouse consumed 90–100 mg 1-MT during the course of treatment. There was no profound weight gain or loss (23 ± 1 g) during the treatment in both the experimental and control groups.

Figure 4.

Blockade of indoleamine 2,3 dioxygenase (IDO) via 1-methyl tryptophan (1-MT) abrogates transferable tolerance. Outline of experimental plan for clarity (a). Group A mice are recipients of splenocytes (1 × 10⁷) from naive mice, while both groups B and C mice were transferred adoptively with splenocytes (1 × 10⁷) from previously rat red blood cell (RBC)-immunized and tolerized mice, and given a placebo (aspartame) or 1-MT, respectively, in drinking water (b). All animals in the experimental groups were challenged with weekly rat RBC (2 × 10⁸) intraperitoneal (i.p.) immunizations on four occasions. Presence of RBC-bound immunoglobulin (Ig)G autoantibody was measured by direct enzyme-linked anti-globulin test (DELAT) after final rat RBC immunization. RBC from New Zealand black (NZB) mice and naive mice were used as positive and negative controls, respectively (four mice in each group).

Treatment of mice with anti-CTLA-4 and anti-sCTLA-4 monoclonal antibody (mAb)

Mice were administered 100 μg of pan-specific anti-CTLA-4 mAb (clone 9H10; Insight Biotechnology, Wembley, UK) or anti-sCTLA-4 mAb developed in-house (clone JMW-3B3) weekly for 3 weeks i.p. (each mouse receiving 100 μg antibody per week). Control mice were given 200 μl PBS vehicle.

Statistical analysis

Differences in DELAT and IELAT scores were analysed using Student's t-test. A P-value < 0·05 was considered significant at 95% confidence interval.

Results

Kinetics of the autoantibody response in rat RBC-induced AIHA in BALB/C mice

AIHA is characterized by the production of IgG autoantibodies that bind self-RBC. We therefore monitored the presence of IgG molecules bound to the surface of murine RBC by using a quantitative and sensitive cellular ELISA based on DELAT to map the anti-RBC immune response following immunization of BALB/C mice with xenogeneic rat RBC. RBC from NZB mice, which develop AIHA spontaneously after birth and have autoantibodies bound to their RBC, were used as a positive control.

BALB/C mice immunized with rat RBC had a DELAT score significantly higher than that of control PBS vehicle-immunized mice when measured at week 5 after the first rat RBC immunization, indicating the presence of autoantibodies bound to mouse RBC (Fig. 1a, P = 0·008, n = 8). As predicted, the autoantibody response receded gradually after week 13, the DELAT score being comparable to that of the PBS control mouse group at week 15 (Fig. 1b). Serum from rat RBC-immunized, but not PBS-treated mice, also contained anti-rat RBC antibody as detected by IELAT (Fig. 1c). The course of the active immune response at different weeks in both PBS-treated and rat RBC-immunized mice is shown in Fig. 1d. The data provided confidence that this murine mouse model of inducible AIHA was established and a regulatory mechanism that subtly regulates autoimmune responses against autologous RBC was detectable.

Figure 1.

Anti-red blood cell (RBC) antibody response in rat RBC-immunized mice. Groups of eight BALB/C mice were immunized with 2 × 10⁸ rat RBC in 200 μl phosphate-buffered saline (PBS) per week for 4 weeks; control mice received PBS (200 μl). RBC from rat RBC-immunized or PBS-treated BALB/C mice at weeks 5 (a) and 15 (b) were assessed for the levels of autoantibodies by direct enzyme-linked anti-globulin test (DELAT). RBCs from New Zealand black (NZB) mice were used as a positive control. Mouse antibody response to rat RBC was measured by indirect enzyme-linked anti-globulin test (IELAT) (c). The course of the active immune response in the two groups of mice is shown (d). At week 16, splenocytes (1 × 10⁷) from rat RBC-immunized or PBS-treated mice were transferred adoptively by intraperitoneal (i.p.) injection into naive BALB/C mice. These mice were immunized further with weekly i.p. injections of 2 × 10⁸ rat RBC in 200 μl PBS (for 4 weeks) and assessed for the presence of self-RBC-bound autoantibodies by DELAT (e) and rat RBC by IELAT (f). Serum from a naive BALB/C mouse was used as a negative control.

On week 16, splenocytes from tolerant mice were transferred adoptively into naive mice and the engrafted mice challenged with weekly injections of rat RBC (2 × 10⁸ cells in 200 μl sterile PBS) on four occasions. When mice were assessed at week 3 for the presence of anti-RBC autoantibody, recipients of RBC-tolerant splenocytes were resistant to developing anti-RBC autoantibodies (Fig. 1e, P = 0·019, n = 4) but retained the capacity to produce antibodies against rat RBC (Fig. 1f, P < 0·0001). However, mice that received splenocytes from PBS-treated mice, and were therefore effectively naive, produced RBC autoantibodies following immunization with rat RBC. This shows that transferred splenic lymphocytes from donor mice that were tolerized to rat RBC antigens curtailed the autoantibody response in recipient mice.

Blockade of pan-CTLA-4 or spliced variant soluble CTLA-4 does not affect transferable tolerance

Previous work has shown that adoptive transfer of purified CD4⁺CD25⁺ but not CD4⁺CD25^– T cells from immunized mice prevented the induction of AIHA, highlighting the importance of CD4⁺CD25⁺ regulatory T cells (T_reg) for the control and induction of AIHA 10. The co-stimulatory inhibitor CTLA-4 is expressed constitutively by T_reg 21, and in an adoptive transfer model of diabetes and colitis, antigen-specific T_reg lacking CTLA-4 were unable to regulate disease 22,23. While CTLA-4 is perceived usually as a cell surface receptor on T cells, it can also be secreted by T cells 24 and monocytes 25 as an alternatively spliced soluble product (sCTLA-4). Interestingly, selective ablation of sCTLA-4 in T_reg also failed to inhibit autoimmune disease induced by transfer of CD4⁺CD25RB^hi cells into non-obese diabetic severe combined immunodeficient (NOD.SCID) mice and accelerated the onset of autoimmune diabetes 26. We therefore investigated whether total ‘pan’ blockade of CTLA-4 (receptor and sCTLA-4 isoforms of CTLA-4) or selective blockade of sCTLA-4 had a role to play in acquired tolerance to AIHA following adoptive transfer of previously tolerized splenocytes and subsequent immunization with rat RBC. To test this, mice transferred adoptively with tolerized splenocytes were administered with 100 μg of pan-specific anti-CTLA-4 mAb (clone: 9H10) or the sCTLA-4 selective JMW-3B3 mAb as outlined in Fig. 2a. When levels of mouse RBC-bound autoantibody were assessed at week 3 (Fig. 2b, upper graph) and again at week 8 following first rat RBC immunization (lower graph), DELAT scores were similar in recipients of both anti-CTLA-4 and anti-sCTLA-4 mAbs, and comparable to PBS controls. Normally, autoantibody production, if any, can be detected prominently within a few days of rat RBC immunization, but in this study follow-up until week 8 did not reveal any evidence of autoantibody production, implying that blockade of CTLA-4 or sCTLA-4 does not disrupt transfer of the tolerance process mediated by splenic lymphocytes from previously tolerized mice. Also, immunized animals in all groups were effectively producing xenoantibodies against rat RBC as measured by IELAT (Fig. 2c), indicating that their ability to mount an effective antigen-specific immune response is not compromised either by blockade of pan-CTLA-4 or sCTLA-4 alone.

Figure 2.

In-vivo administration of anti-cytotoxic T lymphocyte antigen (CTLA)-4 or anti-soluble sCTLA-4 (sCTLA-4) monoclonal antibodies (mAb) does not prevent transfer of anti-mouse red blood cell (RBC) tolerance. Three groups of BALB/C mice (four mice in each group) transferred adoptively with tolerized splenocytes were administered anti-CTLA-4, anti-sCTLA-4 mAb (both 100 μg/mouse) or phosphate-buffered saline (PBS) vehicle alone on three occasions, and challenged with rat RBC, as shown (a). At weeks 3 (upper graph) and 8 (lower graph), mouse RBC were assessed for the presence of anti-mouse RBC autoantibodies by direct enzyme-linked anti-globulin test (DELAT) (b) and presence of anti-rat RBC antibody in the serum of mice by indirect enzyme-linked anti-globulin test (IELAT) (c).

Antibody-mediated blockade of pan-CTLA-4/sCTLA-4 amplifies antigen-specific immune responses

Antibody-mediated blockade of pan-CTLA-4 or sCTLA-4 did not break tolerance, as determined by a lack of autoantibody response to mouse RBC in this transferable tolerance model. We then asked how splenic cells from these antibody-mice would respond and compare to rat erythrocyte ghosts likely to contain a wide range of extracellular RBC antigens. Splenic cell cultures from mice after 3 weeks of final rat RBC immunization as outlined in Fig. 2a were incubated with increasing amounts of rat erythrocyte ghosts. Additionally, they were stimulated with autologous mouse RBC ghosts to assess if any responses were also being targeted to self-antigens present on the mouse RBC membrane.

The proliferative and IFN-γ response of splenocyte cultures responding to rat RBC ghosts from mice treated with either pan-specific CTLA-4 mAb 9H10 or sCTLA-4 selective mAb, JMW-3B3, were higher than that of cells from PBS vehicle-treated mice (Fig. 3a). Interestingly, IFN-γ production from the anti-sCTLA-4-treated group was approximately three–fivefold that of the anti-CTLA-4 mAb-treated group at higher doses of RBC ghosts, indicating that blockade of sCTLA-4 enhances responses against rat RBC even though it does not break tolerance. This is consistent with our human studies, in which the intensity of antigen-specific responses are regulated by endogenous levels of sCTLA-4 27. Unlike IFN-γ, we did not detect changes in levels of IL-10 and IL-17 in splenic cell cultures from mice that were administered with either anti-CTLA-4 or anti-sCTLA-4 mAbs. This may correlate with the finding that murine AIHA is T helper type 1 (Th1)-mediated, with IFN-γ prominently driving the autoimmune pathology 28.

Figure 3.

Effect of anti-cytotoxic T lymphocyte antigen (CTLA)-4 monoclonal antibodies (mAbs) on effector response to rat red blood cell (RBC) and autologous RBC ghosts. Splenic cell cultures from mice after 3 weeks of final rat RBC immunization as outlined in Fig. 2a were stimulated in vitro with increasing concentrations of rat RBC ghosts (a) or autologous RBC ghosts (b) and proliferative responses and levels of interferon (IFN)-γ, interleukin (IL)-10 and IL-17 measured on day 5 of cell culture.

Furthermore, all these responses were targeted to rat RBC membrane antigens but not to autologous ghosts (Fig. 3b), suggesting that, at the T cell level, mice were tolerized to autologous antigen but retained the ability to mount immune responses to rat RBC antigens.

Transfer of tolerance to haemolytic disease can be inhibited by the IDO inhibitor 1-MT

Studies have demonstrated that CTLA-4 is capable of inducing the tolerogenic IDO-mediated tryptophan catabolism pathway by binding B7 ligands on DCs and macrophages 17. While native sCTLA-4 may not play a direct role in tolerance induction, we have observed its capacity to induce IDO activity (manuscript in preparation) in a manner similar to that observed previously for CTLA-4-Ig 18; therefore, we investigated whether IDO has any role to play in the tolerance process associated with the rat RBC-inducible AIHA model. Blockade of IDO in mice that had received tolerized splenocytes and were thus expected to be tolerant upon immunization with rat RBC was achieved by feeding them with aspartame-flavoured water containing 1-MT, a tryptophan homologue that disrupts the catabolic activity of IDO (experimental plan in Fig. 4a). As expected, naive mice in group A (Fig. 4b), receiving rat RBC for the first time following adoptive transfer of naive splenocytes, developed mouse RBC autoantibodies, while mice in group B receiving splenocytes from rat RBC-tolerant mice were resistant to rat RBC immunization. However, tolerance to AIHA was abrogated significantly in these mice when allowed to imbibe water containing 1-MT (group C, P = 0·019) compared with mice imbibing aspartame-flavoured water alone (group B). These data indicate for the first time that the transfer of tolerance to naive recipients is at least partially dependent upon IDO-mediated immunosuppression, as mice receiving tolerized splenocytes under the cover of 1-MT were refractory to tolerance and developed autoantibodies to self-RBC.

Discussion

We have re-established Playfair and Marshall-Clarke's model of murine AIHA successfully and shown that tolerance acquired to rat RBC-induced haemolytic disease can be transferred subsequently to naive recipients: a model of transferable or infectious tolerance. Mice recover from disease due to a lymphocyte-mediated tolerance process with exquisite specificity, which suppresses only responses to self-epitopes while allowing responses to xenogenic rat-specific determinants to be boosted selectively. This protection can be transferred subsequently by injecting splenocytes from recovered mice into naive recipients, indicating that tolerant lymphocytes can, in some way, educate naive unprimed lymphocytic processes. Transfer of tolerance to naive recipients is, possibly among many other factors, dependent upon IDO-mediated immunosuppression, as mice receiving previously tolerized splenocytes under the cover of an IDO inhibitor, 1-MT, were refractory to tolerance and developed haemolytic disease upon further challenge with rat RBC.

Initiators of IDO activity, CTLA-4 17,18 or sCTLA-4 did not mediate this tolerogenic process, even though blockade boosted antigen-specific effector immune responses leading to an increase in antigen-specific cell culture supernatant levels of IFN-γ and cell proliferation. Several models of autoimmune disease, including those that are antibody-mediated, are driven by helper responses associated with the Th1 subset, and it was reported that the CD4⁺ Th response to RBC autoantigen in patients with AIHA 29 or in NZB mice 30 was associated with IFN-γ. Acute IFN-γ secretion by splenocytes from rat RBC-immunized mice following sCTLA-4 blockade was approximately three–fivefold higher than that from groups receiving pan-specific antibody, raising the question of whether functional blockade of the CTLA-4 receptor is required to enhance antigen-specific effector immune responses. This is supported further by recent findings that loss of sCTLA-4 alone compromises immune regulation, and therefore the function of the soluble isoform may be distinct from and not redundant with the CTLA-4 receptor 26. However, no abolition of the RBC autoantibody resistance was received from the adoptive transfer of previously tolerized splenocytes on administration of either antibodies. This points to the existence of other unidentified initiators of the IDO cascade, and while CTLA-4 isoforms may be able to induce IDO expression, the regulation and control of this important tryptophan catabolic enzyme system seems to be more complex than believed previously.

Data from Mqadmi et al. have shown that transfer of purified CD4⁺CD25⁺ but not CD4⁺CD25^– T cells from immunized mice prevented the induction of AIHA on subsequent immunization of naive mice with rat RBC 10. The immunosuppressive enzyme IDO can have a multi-tiered role to play at this cellular level. IDO catalyses oxidative cleavage of the indole ring of the essential amino acid tryptophan, producing immunoregulatory metabolites known collectively as kynurenines 18. Tryptophan deficiency and kynurenine excess, resulting from IDO activity, both have an immunomodulatory effect, in particular to drive pathogenic autoreactive Th1 cells to become prone to apoptosis 31. Triggering of the IDO cascade also leads to de-novo regulatory conversion of effector T cells 32,33, while kynurenine, the first breakdown product in IDO-mediated tryptophan degradation, is also known to activate the Aryl hydrocarbon receptor (AHR), leading to AHR-dependent generation of T_regs 34. Blocking IDO activity with the pharmacological inhibitor 1-MT abrogates significantly T_reg generation and suppressor function 33,35. This may lead to activation and accumulation of autoreactive T cells which, in turn, provide help to B cells to produce autoantibodies against self-RBC and break tolerance to RBC surface glycoproteins. In addition to its enzymatic tryptophan-catabolizing activity, IDO may be using its recently described non-enzymatic function to stabilize and maintain regulatory phenotype of dendritic cells, particularly in non-inflammatory contexts 16.

The mechanism by which 1-MT reverses amplification of T_reg number and function is not exactly known, but in conditions where breakdown of tolerance to tumour antigens is a keenly sought goal, 1-MT has been considered for use as a therapeutic. Mice that are administered with 1-MT before injection of tumour cells develop tumours later than control mice and have enhanced CTL activity against tumour cells and reduced tumour growth 36. These findings from animal studies have been extended to therapeutic intervention in humans; 1-MT is being used currently in clinical trials in patients with relapsed or refractory solid tumours to boost effector immune response, where IDO-expressing tumour cells create an immunosuppressive milieu 37.

In summary, this work extends previous findings 14 that demonstrated the importance of IDO-mediated immunosuppression in maintaining tolerance to self. Tolerance to putative self-antigens may be less secure in recipients who do not have functional IDO enzymatic activity. Conversely, 1-MT may also hold the promise of breaking tolerance to tumour antigens in non-solid tumours, where IDO is used as a route to immune escape.

Disclosure

F.J.W., L.N.D. and R.N.B. have applied for a patent to use anti-sCTLA-4 monoclonal antibody JMW-3B3 as a therapeutic.