Immunoregulatory soluble CTLA-4 modifies effector T-cell responses in systemic lupus erythematosus

Background The inhibitory CTLA-4 molecule is a crucial regulator of immune responses and a target for therapeutic intervention in both autoimmunity and cancer. In particular, CTLA-4 is important in controlling antigen-specific immunity, including responses to autoantigens associated with autoimmune disease. Here, we investigate cytokine responses to a range of lupus-associated autoantigens and assess whether the alternatively spliced isoform of CTLA-4, soluble CTLA-4 (sCTLA-4), contributes to immune regulation of autoantigen-specific immunity in systemic lupus erythematosus (SLE). Methods The cell culture supernatant production of sCTLA-4 as well as the cytokines IL-10, IFN-γ, and IL-17 from peripheral blood mononuclear cells (PBMC) from lupus patients and age- and sex-matched healthy volunteer donors were measured in response to previously identified histone and small nuclear ribonucleoprotein (snRNP) autoantigen-derived peptides (H391-105, H471-93, and U170K131-151) by ELISA. We also examined the functional contribution of sCTLA-4 to immune regulation in the context of these autoantigenic peptides following blockade of sCTLA-4 with a selective anti-sCTLA-4 monoclonal antibody, JMW-3B3. Results We identified responses to autoantigenic peptides, which revealed qualitative differences in cytokine (IL-10, IL-17, and IFN-γ) profiles between SLE patients and healthy donors. PBMC from healthy donors responded to each of the lupus peptides by secreting IFN-γ and IL-17, but PBMC from SLE patients produced IL-10. Although we did not observe differences in the levels of serum or PBMC culture supernatant sCTLA-4 in either cohort, blockade of sCTLA-4 in PBMC cultures responding to antigen enhanced the cytokine profiles associated with each group. Conclusion The results show that lupus autoantigen-derived peptides display varied immunogenicity in lupus versus healthy volunteer donors, while sCTLA-4 acts to regulate the T-cell activity independently of response profile. Electronic supplementary material The online version of this article (doi:10.1186/s13075-016-1075-1) contains supplementary material, which is available to authorized users.


Background
Systemic lupus erythematosus (SLE) is a multisystem autoimmune disorder characterised by a range of clinical manifestations including arthritis, nephritis, central nervous system disease, alopecia, and photosensitivity. The disease is most prominent among women in their child-bearing years and, despite recent improvements in therapeutic options, many patients experience poor treatment outcomes [1]. At a molecular level, dysregulated apoptosis, inefficient clearance of cellular debris, and enhanced cytokine activity, notably interleukin (IL)-10 and interferon (IFN)-α, all appear to be important factors contributing to the inflammatory disease process [2].
Chronically activated autoantigen-specific T cells that promote production of high-affinity autoantibodies specific for nucleic acids/nucleic acid binding proteins are also factors in SLE [3][4][5][6]. These autoreactive T cells are driven primarily by recognition of self-peptides derived primarily from the same nucleic acid binding proteins that represent the main targets of the autoimmune response in lupus, e.g. histones [7]. Although these peptide autoantigens can drive pathological processes in lupus, interestingly they can also induce tolerance by influencing increased regulatory T-cell (Treg) activity, or by stimulating production of anti-inflammatory cytokines such as IL-10. The administration of self-peptides to promote tolerogenic responses in lupus has already been successful in murine models of lupus with one such peptide, Lupozor™, progressing through to clinical trials as a specific immunotherapy for the treatment of lupus [8,9]. Peptides as therapeutics are thus very interesting because they are relatively cheap to produce and purify compared with biologics and can be more easily administered [10].
Analysis of cytokine responses in lupus patients driven by autoantigenic peptides has demonstrated a certain amount of heterogeneity in terms of the cytokine profiles stimulated by particular peptides, suggesting that some peptides may preferentially drive regulatory immune responses, e.g. by producing IL-10, and others a more typical effector T-cell cytokine response, e.g. IFN-γ [3], but while these response differences were identified in individual lupus patients it is unclear as to whether any particular profile is solely typical of lupus disease or if healthy donors also raise similar response patterns or any response at all to these lupus-associated peptides. Finally, the recent emergence of T helper (Th)17 T cells and their relevance to SLE raises the question of whether these peptides might drive IL-17 production [11].
Cytotoxic T lymphocyte antigen 4 (CTLA-4) is an inhibitory cell-surface receptor central to immune regulation and is predominantly expressed by Treg and activated CD4 + T-effector cells [12,13]. CTLA-4 has become one of the most clinically exploited immune molecules of the last decade. Notably, augmenting CTLA-4 function by infusing the recombinant soluble form of CTLA-4 called CTLA4-Ig (or abatacept) selectively inhibits T-cell co-stimulation and is beneficial for treating some autoimmune diseases, notably rheumatoid arthritis [14]. CTLA4-Ig is currently approved for the treatment of rheumatoid arthritis, but in SLE its benefits have been more difficult to establish, even though its immunosuppressive effects were first clearly established in lupus mouse models [15]. A recent phase II clinical trial showed no overall patient benefit for this therapy, despite anecdotal reports that it benefited individual patients [16].
While natural CTLA-4 is generally considered as a cell-surface receptor, it can also be produced as a naturally secreted soluble molecule [17][18][19]. This natural soluble CTLA-4 (sCTLA-4) lacks the transmembrane domain of the full-length CTLA-4 receptor and carries a different C-terminal amino acid sequence, but crucially retains the capacity to engage with B7 ligands [18]. Thus, this secretable isoform has the potential to influence immune responses including those associated with autoimmune disease. Little is known of any regulatory role for natural sCTLA-4 in this respect although serum levels of sCTLA-4 are higher in SLE patient cohorts and other autoimmune diseases compared with healthy individuals [20,21]. Furthermore, it is not known if sCTLA-4 might regulate autoantigenic T cells stimulated by self-peptides in lupus.

Donors and sample preparation
Blood samples were collected by venepuncture into heparan sulphate-containing vacutainers from lupus patients attending either the Rheumatology or Vasculitis department at Aberdeen Royal Infirmary (see Table 1 for demographic summary), and age-and sex-matched healthy volunteer donors with informed consent. Only patients REC 10/S1401/20). PBMC were prepared by Lymphoprep 1.077 (Axis Shield, Dundee, UK) density gradient centrifugation and cultured essentially as previously described [24] in RPMI 1640 medium (Invitrogen, Paisley, UK) supplemented with 5 % autologous human serum in an atmosphere of 37°C, 5 % CO 2 ; 1 × 10 6 PBMC were cultured for 5 days in 1-ml wells unless otherwise stated.

T-cell proliferative responses and antibody blockade
Cell proliferation was measured by 3 H thymidine incorporation in duplicate samples using a 1450 Microbeta liquid scintillation counter (LKB Wallac, Turku, Finland). Results are presented as mean counts per minute (CPM). Anti-sCTLA-4 blockade responses were performed with the JMW-3B3 anti-sCTLA-4 mAb, developed in-house at the University of Aberdeen. JMW-3B3 is an IgG1λ antibody that selectively binds sCTLA-4 and has previously been demonstrated to enhance antigen-specific immune responses compared with an IgG1 isotype control [25]. Enhanced responses were similar to initial adsorption experiments with AS-33P anti-CTLA-4 mAb (Antibody Solutions, Sunnyvale, CA, USA). Fresh antibody preparations were tested for the presence of endotoxin and stored without preservative at -20°C prior to use.

Soluble CTLA-4 ELISA
The selective ELISA for human sCTLA-4 used the anti-CTLA-4 murine mAb clone BNI3 (2 μg/ml) as a capture reagent, and biotinylated JMW-3B3 [25] as the sCTLA-4-specific detection reagent according to the protocol described for the cytokine ELISA above. Affinity-purified sCTLA-4 was used to construct standard curves.

Data analysis
For greater clarity in the main figures, and for comparison of peptide-or control antigen-induced responses, all data are presented as Stimulation Indices (SI), i.e. the rates of stimulated versus resting cell responses. The full datasets are also presented in Additional file 1 together with a comprehensive statistical analysis (Prism GraphPad). SI values above 2 (i.e. double resting cell levels) were considered positive.

Results
Soluble CTLA-4 levels in serum and PBMC cell culture supernatants of lupus patients Previous reports identified relatively high serum levels of sCTLA-4 in SLE patients [20,21], but that work used antibodies capable of binding either native sCTLA-4, exocytosed full-length CTLA-4, or fragments of the extracellular domain of CTLA-4 cleaved from the cell surface. To ensure only sCTLA-4 and not cleaved products were detected, we used an antibody selective for a unique sequence of sCTLA-4 to analyse cell culture supernatant and serum levels of sCTLA-4 [25].
First, we measured levels of T helper-associated effector cytokines and sCTLA-4 in cell culture supernatants from resting PBMC, as well as sera from SLE patients and healthy donors and age-and sex-matched healthy volunteer donors (Fig. 1). As expected, median supernatant levels of IFN-γ and IFN-α in patient PBMC resting cell cultures were significantly higher than those of healthy volunteer donors (IFN-γ: 173.8 pg/ml versus 77.0 pg/ml, p < 0.05; IFN-α: 988.5 pg/ml versus 94.8 pg/ml, p < 0.0001; Mann-Whitney U test) but no significant difference in cellular proliferation (180 versus 158 CPM) or the cytokines IL-10 (668.9 pg/ml versus 702.3 pg/ml), IL-17 (87.7 versus 44.4 pg/ml), or sCTLA-4 (292.1 versus 552.5 pg/ml) was observed (Fig. 1). Levels of sCTLA-4 in supernatants from lupus patient PBMC varied widely between individuals (range 0-4288 pg/ml; Fig. 1). However, cell culture sCTLA-4 levels from the healthy donor cohort also varied widely (range 0-4027 pg/ml; Fig. 1). A similar pattern was observed in serum sCTLA-4 levels from SLE patients (range 0-6326 pg/ml, median 1044.0 pg/ml; Fig. 2) compared with sera from healthy donors (range 0-4421 pg/ ml, median 792.4 pg/ml). In this analysis, in contrast to previous studies [20], despite an overall increase in serum levels of sCTLA-4, there was no significant difference between the healthy and patient donor cohort.

Effect of sCTLA-4 on responses to peptide autoantigens in PBMC from lupus patients or healthy volunteer donors
Previously, sCTLA-4 was demonstrated to regulate cytokine responses in PBMC and T cells, but its regulatory effects in autoimmune disease have not been examined. Strong stimuli (e.g. anti-CD3 mAb) can temporarily suppress sCTLA-4 expression and production but, in healthy PBMC, more physiological stimuli (e.g. recall antigens) actually maintain sCTLA-4 levels and may even increase production [25]. To examine sCTLA-4 production in lupus, we investigated five peptide autoantigens identified in previous studies for their capacity to influence sCTLA-4 production.
Three of the five peptides examined (H3 Fig. 3a and Additional file 1). Each of these peptides also induced lower but significant increases in IL-10 production compared with resting cells with a mean SI of 5 or less depending on the peptide and peptide dose. These healthy donors were, in effect, mounting a response characteristic of a mixed Th1/Th17 T-cell response to these lupus peptide autoantigens.
In contrast to healthy donors, SLE patient responses were characterised by increased production of IL-10 to the peptides rather than increased IFN-γ and IL-17. Despite production of immunosuppressive IL-10, PBMC from the SLE patient cohort significantly increased levels of cell proliferation to the peptides compared with resting cells, although levels were significantly lower than those observed in healthy donors (p < 0.01, Mann Whitney). Thus, this initial assessment of anti-peptide cytokine responses revealed a clear differentiation in cytokine profile between healthy and SLE patient donors.
We concurrently examined sCTLA-4 supernatant levels and, despite individual donors mounting large sCTLA-4 increases to the peptides, as a population both patient and healthy donor levels of sCTLA-4 were not significantly increased or decreased (Additional file 1).
To determine whether lupus patients would respond in the same way as healthy donors regarding sCTLA-4 production by PBMC stimulated by antigens rather than autoantigens, we examined a panel of stimulants, including PPD recall antigen, SEB superantigen, and agonist anti-CD3 mAb (Fig. 3b). Mean SI levels of sCTLA-4 were generally higher upon stimulation with PPD recall antigen in both patient and healthy donor cohorts and SEB in the lupus patient cohort, although levels were not statistically significant compared with resting PBMC (Wilcoxon). Further analysis revealed that the mean increase in SI primarily came from just a few individual donors that responded to PPD by secreting increased amounts of sCTLA-4 in an antigen-dependent manner. Reduction of sCTLA-4 by stimulation of PBMC with high doses of anti-CD3 mAb has been observed [17,19] (Fig. 3b and Additional file 1) and was similar here [25].

Effect of anti-sCTLA-4 antibody blockade on autoantigenspecific immune responses in lupus patients
Previously, we demonstrated in healthy donors that naturally secreted sCTLA-4 contributes to the regulation of antigen-specific effector T-cell responses by regulatory T cells, and that blocking its activity with a soluble CTLA-4 selective antibody can enhance T-effector response intensity compared with an IgG1 isotype control [25]. Little is known, however, whether amplification of such responses is intact in the context of autoimmunity. We therefore investigated the effects of sCTLA-4 blockade in lupus patients or healthy volunteer donors. Blockade of sCTLA-4 with the anti-sCTLA-4 mAb JMW-3B3 reduced supernatant levels of detectable sCTLA-4 by more than 92 % (n = 24 donors; p < 0.0001, Wilcoxon).
We first assessed whether sCTLA-4 inhibition in cell cultures stimulated with lupus peptide autoantigens would affect any aspect of effector T-cell activity. Inhibition of sCTLA-4 in healthy donor cell cultures stimulated with U170K 131-151 , H3 91-105 , and H4 71-93 had mixed effects on effector cytokine production and cell proliferation specific to each peptide, and were generally also individual to particular donors (summarised in Table 2 for both healthy and lupus donors). Blockade of sCTLA-4 in the healthy donor cohort generally suppressed or did not alter cell proliferation and cytokine production ( Fig. 4 and summarised in Table 2; p < 0.05, Wilcoxon), while the SLE cohort following sCTLA-4 inhibition significantly enhanced the prevailing IL-10 cytokine phenotype stimulated by the peptides. Overall, our assessment of the effects of sCTLA-4 blockade on anti-peptide PBMC responses from healthy and lupus donors indicates that PBMC from the lupus cohort were more sensitive to changes in IL-10 levels following removal of the suppressive effects of sCTLA-4 suggesting that, nevertheless, sCTLA-4 suppresses effector T-cell responses independent of the profile of cytokine(s) it is producing.
An interesting corollary to the investigation of sCTLA-4 blockade on peptides was the same assessment of the effects on recall responses in lupus patients compared with healthy donors, where healthy donor PBMC were  Table 2), but there were no significant changes in levels of IL-17 or IL-10 following blockade of sCTLA-4. In contrast, sCTLA-4 blockade in resting PBMC from SLE patients raised both cell proliferation and IL-10 levels ( Fig. 5 and Table 2), suggesting that PBMC in lupus patients seem primed to secrete higher levels of IL-10.

Discussion
In this study we conducted a comprehensive review of PBMC responses to peptide autoantigens associated with SLE, and for the first time demonstrated qualitative differences in the responsive cytokine profile to autoantigenic peptides between SLE patients and healthy donors, together with an analysis of a role for sCTLA-4. In our study, of five autoantigenic peptides previously identified to contain T-cell stimulatory epitopes [3,22,23], three (H3 91-105 , H4 71-93 , and U170K 131-151 ) were universally immunogenic for both SLE patients and healthy volunteer donors, allowing us to observe that lupus patients respond very differently to peptide autoantigens compared with healthy donors. This difference also extended to the immunoregulatory function of sCTLA-4, because removing its regulatory function accentuated differences between healthy and patient donor responses.
Soluble CTLA-4 is derived from an alternatively spliced isoform of full-length CTLA-4 and is secreted by several cell types including Treg [25], monocytes [27], melanoma cell lines [28], and pituitary gland cells [29], but its effects remain largely unexplored despite the clinical relevance of CTLA-4 both in autoimmune disease and cancer. High sCTLA-4 serum levels have previously been identified in lupus patients, and low efficiency of production was mooted to be associated with a susceptibility CT60 SNP genotype of CTLA-4 [30,31].
The question here was could lupus-associated peptide autoantigens induce increased levels of sCTLA-4 in either patients or healthy donors? The answer was yes, but only in particular individuals. The question of why some individuals have the capacity to increase sCTLA-4 in response to antigens points to a particular primed T-cell subset with that capacity. Identifying methods for isolating and expanding such T-cell populations could offer a way forward for future understanding of immune regulation in SLE.
Blockade of sCTLA-4 revealed significant response differences between PBMC from healthy donors and SLE patients with only IL-10 levels increasing following PPD or peptide stimulation in the SLE patient group, suggesting that any response enhancement by blocking sCTLA-4 function is likely to reflect the predominant T-cell phenotype and the associated cytokine response, in this case IL-10. This points to a more general sCTLA-4 effect in which it may regulate the activity of antigen-specific effector responses but does not drive, in the short term at least, a skewing towards a particular cytokine profile produced by those activated T cells. In effect, sCTLA-4 most probably exerts its immunosuppressive effect by simply blocking B7/CD28 interactions, thus raising the threshold before which successful T-cell co-stimulation can occur.
Regarding the use of recombinant CTLA-4 products, e.g. CTLA4-Ig, in SLE, a key question is whether natural sCTLA-4 is functionally comparable. Previous studies identified sCTLA-4 to engage with B7 molecules and to effectively compete with CTLA4-Ig for binding [18,32], but there are no hard data on relative binding affinities of sCTLA-4 with the full-length CTLA-4 receptor for B7 ligands. Our initial analysis of affinity-purified and recombinant sCTLA-4 structure indicates that it forms functional dimers mediated by a cysteine residue in the C-terminal amino acid sequence unique to sCTLA-4 (manuscript in preparation). This makes it more likely that affinity of sCTLA-4 for B7.1/B7.2 is comparable with that of both full-length CTLA-4 receptor and recombinant CTLA4-Ig. It is likely, therefore, that sCTLA-4 is functionally relevant to immune responses, but how serum levels in patients might impact on any CTLA-4based therapy is unknown. Perhaps the most surprising aspect of this study was the identification of a clear difference in the profile of cytokines produced in response to the peptide autoantigens between healthy or lupus patient PBMC. Healthy donors responded to the peptides by proliferating and producing higher levels of both IFN-γ and IL-17, whereas PBMC from lupus patients also proliferated but secreted IL-10. This supports the notion that, in lupus, T-cell responses are acting to support autoreactive B-cell responses and the production of high-affinity IgG autoantibodies [33][34][35][36]. While IL-10 is generally considered an anti-inflammatory cytokine, serum levels are often raised in lupus patients where it can act as a potent B-cell growth factor, driving autoantibody production, class switching, and plasmablast differentiation [33,35,37]. Thus, IL-10 may be important for suppressing tissue damage from activated pro-inflammatory macrophages and neutrophil responses to immune complex deposits, but can also contribute indirectly to the development of pathogenic immune complexes by raising high-affinity anti-nucleic acid and nucleoprotein autoantibody levels, a process remote from the inflammatory lesion [33]. It is of great interest, therefore, that responsive peptide autoantigen-specific T cells are a source of this IL-10.
This study provides evidence of a role for sCTLA-4 in immune regulation, yet clearly we need to understand more about the mechanistic processes which drive its production in response to antigen. One key issue that remains to be addressed is what is the primary stimulus of sCTLA-4 production? It is likely that Fig. 4 Effect of anti-sCTLA-4 mAb blockade on immune responses to H4 71-93 peptide isolated from PBMC of healthy donors (left panels) or SLE patients (right panels). PBMC were incubated with H4 71-93 (20 μg/ml) in the presence of JMW-3B3 anti-sCTLA-4 mAb or an IgG1 isotype antibody (both 10 μg/ml) for 5 days before analysis of cellular proliferation and levels of cell culture supernatant cytokines (IFN-γ, IL-17, IL-10, and sCTLA-4; n = 45 per group; *p < 0.05, **p < 0.01, ***p < 0.001, non-parametric Mann-Whitney U test; median values are shown). CPM counts per minute, IFN interferon, IL interluekin, ns not significant, sCTLA-4 soluble cytotoxic T lymphocyte antigen 4, SLE systemic lupus erythematosus production of sCTLA-4 in response to antigen relies both on antigen-specific T-cell responses but also the environmental milieu. We have recently observed that TGF-β2 is an important stimulatory factor of sCTLA-4 and it is now important to identify other stimulatory or suppressive factors that regulate the production of sCTLA-4.

Conclusions
SLE patients present different cytokine response patterns to lupus-associated autoantigenic peptides compared to healthy donors, secreting significantly higher levels of IL-10. The natural immunosuppressant, sCTLA-4, supresses antigen-specific cytokine responses and antibody blockade of its activity enhances cytokine responses in healthy and lupus patient donors, independent of which cytokine is being secreted. Some individual responses indicated that peptides can induce increased production of sCTLA-4 but this effect was not significant at a population level for lupus peptide autoantigens.
Immunostimulatory peptide autoantigens including those investigated here may, therefore, be pivotal in driving the disease process, raising the prospect that targeting antigenspecific T cells to actually produce less IL-10, and allowing the immune response to progress to full resolution, could be a viable route to reduce immune complex-mediated pathology in lupus.

Additional file
Additional file 1: Full analysis of individual patient and donor peptide response datasets used to generate the summary data presented in  Effect of anti-sCTLA-4 mAb blockade on PBMC immune responses to PPD recall antigen. PBMC were incubated with the stimulants in the presence of JMW-3B3 anti-sCTLA-4 mAb or an IgG1 isotype antibody (both 10 μg/ml) for 5 days before analysis of cellular proliferation and levels of cell culture supernatant cytokines (IFN-γ, IL-17, IL-10, and sCTLA-4; n = 45 per group; *p < 0.05, **p < 0.01, ***p < 0.001, non-parametric Mann-Whitney U test; median values are shown). CPM counts per minute, IFN interferon, IL interluekin, ns not significant, PPD purified protein derivative, sCTLA-4 soluble cytotoxic T lymphocyte antigen 4