Regulation of allograft survival by inhibitory FcgammaRIIb signaling

Abstract

Fc gamma receptors (FcγR) provide important immunoregulation. Targeting inhibitory FcγRIIb may therefore prolong allograft survival, but its role in transplantation has not been addressed. FcγRIIb signaling was examined in murine models of acute or chronic cardiac allograft rejection by transplanting recipients that either lacked FcγRIIb expression (FcγRIIb^−/− ) or over-expressed FcγRIIb on B cells (BTG). Acute heart allograft rejection occurred at the same tempo in FcγRIIb^−/− C57Bl/6 (B6) recipients as WT recipients, with similar IgG alloantibody responses. In contrast, chronic rejection of MHC class II-mismatched bm12 cardiac allografts was accelerated in FcγRIIb^−/− mice, with development of more severe transplant arteriopathy and markedly augmented effector autoantibody production. Autoantibody production was inhibited, and rejection delayed, in BTG recipients. Similarly, whereas MHC class I-mismatched B6.K^d hearts survived indefinitely and remained disease-free in B6 mice, much stronger alloantibody responses and progressive graft arteriopathy developed in FcγRIIb^−/− recipients. Notably, FcγRIIb-mediated inhibition of B6.K^d heart graft rejection was abrogated by increasing T cell help through transfer of additional H2.K^d-specific CD4 T cells.

Thus, inhibitory FcγRIIb signaling regulates chronic but not acute rejection, most likely because the supra-optimal helper CD4 T cell response in acute rejection overcomes FcγRIIb-mediated inhibition of the effector B cell population. Immunomodulation of FcγRIIb in clinical transplantation may hold potential for inhibiting progression of transplant arteriopathy and prolonging heart transplant survival.

Introduction

Transplantation represents the best treatment option for most patients with end-stage organ failure. Nevertheless, despite advances in immunosuppression, the majority of organ grafts eventually fail, resulting in death or need for re-transplantation. Although multifactorial, transplant failure is principally a culmination of adaptive alloimmune recognition and thus further improvements in graft survival are likely to require development of additional strategies that focus on inhibiting the alloimmune response. One potential approach to achieve this inhibition is to target immunomodulatory ligands that are widely expressed on immune-system cells, and of these, receptors for the IgG Fc region (FcγRs) hold particular appeal.

There are four different classes of FcγRs (1, 2), encompassing distinct receptors that differ in their respective affinity for IgG subclass and in whether their ligation triggers activating or inhibitory intracellular signaling cascades (reviewed in (1-4)). Binding to an activating Fcγ receptor can trigger a variety of effector functions including phagocytosis (5), antibody-dependent cell-mediated cytotoxicity (6), and release of inflammatory mediators. The balance of expression of inhibitory to activating FcγRs on macrophages and DCs dictates their excitatory state, and thus provides important immunoregulatory function for controlling cellular immunity. Perturbations in this balance, for example through genetic variation in the affinity of activating FcγRs, are associated with autoimmune disease in humans, predominantly SLE (1, 7, 8).

The immunomodulatory properties of the FcγR family are particularly relevant to humoral immunity, because signaling through the generally low affinity receptors is dependent upon recognition of immune complexes (which enable simultaneous binding to multiple antibody Fc regions), and because FcγRIIb (CD32B), the sole inhibitory Fc receptor in both mice and humans, is the only FcγR expressed on B cells. FcγRIIb is composed of two Ig-like extracellular domains, one transmembrane domain, and an intracytoplasmic domain that contains a single immunoreceptor tyrosine-based inhibitory motif. The binding of IgG-antigen complexes to FcγRIIb inhibits the activation signal delivered through activating Fcγ receptors or the BCR, and diminishes B lymphocyte stimulator receptor signaling and upregulation (9). Loss of FcγRIIb expression therefore enhances T-dependent and T-independent antibody responses (10, 11). Human and murine studies have accordingly confirmed that FcγRIIb plays an important role in many disease processes, including autoimmunity (7, 12), bacterial sepsis (10), parasitic infection (8), and cancer (13).

Although acute antibody-mediated allograft rejection has been recognized for decades (14), the contribution of humoral alloimmunity to chronic allograft failure, the major cause of graft loss (15), has only become appreciated recently. It is now clear that donor-specific antibodies, whether present before transplantation, or that develop afterwards, are associated with failure of both kidney (16, 17) and heart (18, 19) allografts. Chronic graft loss is characterized by the development of transplant arteriopathy (TA), which causes progressive ischemic parenchymal obliteration and replacement fibrosis (20). Several animal studies have demonstrated that alloantibody contributes to the development of TA via complement-dependent and -independent mechanisms (21-23).

Analysis of the role of FcγRIIb in rejection of solid-organ allografts may therefore inform development of novel therapies that block progression of TA and prolong allograft survival, but has yet to be performed. Herein, we report studies using murine models of acute and chronic cardiac allograft rejection that incorporate recipients that either over-express FcγRIIb on select leukocyte lineages or are genetically-deficient in FcγRIIb; thus enabling detailed examination of the role of inhibitory FcγRIIb receptors in regulating the alloimmune response. We demonstrate that FcγRIIb signaling influences chronic, but not acute, murine cardiac allograft rejection, and propose this difference relates to differences in the level of CD4 T cell help available. These findings highlight the potential for targeting FcγRIIb as immunomodulation in clinical transplantation.

Materials and Methods

Animals

C57Bl/6 (H-2^b), BALB/c (H-2^d), and CBA/Ca (H-2^k) mice were obtained from Charles River Laboratories (Margate, UK). FcγRIIb^−/− mice (11), mice over-expressing FcγRIIb on either B cells or macrophages (24), B6.C-H-2^bm12 (bm12) mice (25), TCR75 mice (26), and B6.K^d mice (kindly gifted by Dr R.P. Bucy, University of Alabama, Alabama) (27) were bred in-house. TCR75 mice were bred on a RAG1^−/− background to ensure monoclonality. FcγRIIb^−/− mice (kindly gifted by Prof J. V. Ravetch, The Rocker University, New York) were generated from heterozygous breeders; in some experiments FcγRIIb^+/+ LMCs were used. All transgenic strains were on a B6 background. Animals were bred and maintained in specific-pathogen-free conditions. All experiments were approved by the UK Home Office under the Animals (Scientific Procedures) Act 1986.

Heterotopic cardiac transplantation

Donor hearts were transplanted intra-abdominally into recipient mice, as described previously (28). Briefly, the donor aortic arch was anastomosed end-to-side with the recipient abdominal aorta, and the donor pulmonary artery was anastomosed to the recipient inferior vena cava in the same manner. Rejection was defined as the complete cessation of palpable myocardial contraction, and was confirmed at explant. Grafts were excised and transected; one half was fixed in 10% formal saline, the other was stored at −80°C.

Histopathology and Immunohistochemistry

Formalin-fixed hearts were paraffin-mounted and stained using Weigert’s Elastin van Gieson (EVG) method to delineate the internal elastic lamina (IEL). TA was assessed by morphometric analysis of EVG⁺ vessels using Cell^R digital imaging software (Olympus, Tokyo, Japan). Percentage cross-sectional area luminal stenosis was calculated as ((area within IEL - area of lumen)/area within IEL) x 100. All EVG⁺ vessels in each section were assessed, with a minimum of 8 vessels per section examined. 7μm cryostat sections of OCT embedded hearts were examined for complement C4d deposition as described previously (29).

Anti-K^d alloantibody measurement by ELISA

Alloantibodies against H-2K^d were detected and quantified using refolded recombinant H-2K^d as the target protein on ELISA. Recombinant soluble H-2K^d was produced using the method previously described for rat MHC class I (30). In brief, pET-22b+ expression plasmids containing the DNA sequences encoding either amino acids 1-280 of the H-2K^d heavy chain (α1, α2 and α3 extracellular domains) or the murine β2-microglobulin (kindly gifted by Professor P. Lehner, University of Cambridge, UK) were transformed into E. coli BL21 (DE3) strain bacteria (Novagen, Merck, UK) and grown in LB broth (Invitrogen, Paisley, UK). Recombinant H-2K^d heavy chain or β2-microglobulin was extracted from inclusion bodies released from E. coli pellets by chemical lysis. Soluble H-2K^d molecules were generated by refolding the purified heavy chain and β2-microglobulin around a synthetic peptide (TYQRTRALV) (ISL, Paignton, UK) using Garboczi’s dilution method (31). Finally, fast protein liquid chromatography purification of the refold mixture was performed (AKTA FPLC, Amersham Biosciences, Buckinghamshire, UK) and the appropriate fraction was collected, pooled, filter sterilized, and stored in aliquots at 4°C.

H-2K^d was bound to Immulon 4 HBX ELISA plates (Thermo, Milford, MA, USA) at 5 μg/ml in NaCO-NaHCO buffer (pH 9.6) and unbound protein-binding sites were blocked with 1% Marvel (Premier International Foods, Spalding, UK). Serial tripling dilutions of test sera were added and bound antibodies detected by incubation with biotinylated anti-mouse Ig antibodies and ExtrAvidin peroxidase conjugate (Sigma, St Louis, MO, USA). TMB Microwell Peroxidase substrate (KPL, Gaithersburg, MD, USA) was then applied, the reaction stopped by addition of 0.2 M H₂SO4 and absorbance at 450 nm was measured on a FLUOstar OPTIMA plate reader (BMG LABTECH GmbH, Offenburg, Germany). An absorbance versus dilution curve was plotted, the AUC calculated, and expressed as percentage of a positive control (pooled hyperimmune serum). Naive FcγRIIb and WT B6 sera had AUC <9% of the positive control.

Alloantibody measurement by flow cytometry

Thymuses were harvested from the donor strain and single cell suspensions were prepared. Non-specific antibody binding was blocked by incubating thymocytes with anti-CD16/CD32 FcγR block (BD Pharmingen, Franklin Lakes, NJ, USA). Serial tripling dilutions of heat-inactivated serum samples were added to each well and incubated. Serum alloantibody was detected with FITC-conjugated antibodies directed against pan IgG (STAR70, Serotec, Oxford, UK), and cells were analyzed on FACSCalibur^TM or FACSCanto II^TM flow cytometers with CellQuest or FACSDiva software (all BD Biosciences, San Jose, CA, USA). Results were expressed as geometric mean channel fluorescence against serum dilution.

Autoantibody determination

Antinuclear autoantibody responses were determined by HEp-2 indirect immunofluorescence (The Binding Site Ltd, Birmingham, UK). Test sera were incubated on slides coated with HEp-2 cells and bound autoantibody was detected with STAR70 anti-mouse IgG:FITC mAb (Serotec, Oxford, UK). For each sample, four random photomicrographs were taken at x20 magnification, and the intensity of staining was determined using MetaMorph software (Molecular Devices, Sunnyvale, CA, USA). The relative fluorescence was then derived by comparison to a standard curve obtained for each assay by serial dilution of a pooled positive control serum assigned an arbitrary value of 1000. Positive control serum was generated by challenging TCR KO B6 mice with 10⁶ bm12 CD4 T cells i.v. and collecting serum at day 50. Fluorescence values of sera from naive FcγRIIb and WT B6 mice were less than 50.

CD4 T cell preparation

Splenocytes from TCR75 mice were prepared and the CD4 T cell percentage measured by flow cytometry, following 7-aminoactinomycin D dead cell exclusion, by labeling with allophycocyanin conjugated rat-anti mouse CD4 (Clone RM4-5, BD Pharmingen,San Diego, CA, USA). Appropriate doses of live TCR75 CD4 T cells were then administered by i.v. injection.

CD8 T cell ELISPOT

CD8 T cells from the spleens of recipient mice were purified using anti-CD8a (Ly-2) coated magnetic microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) as per the manufacturer’s instructions, mixed with irradiated BALB/c stimulator splenocytes, and added to Multiscreen™ HTS filtration system plates (Millipore Corporation, Billerica, MA, USA) that had been coated with anti-mouse IFN-γ (BD Pharmingen, Franklin Lakes, NJ, USA) in carbonate-bicarbonate buffer. Six wells were used per experimental animal. Plates were incubated at 37 °C and 5% CO₂ for 20 hours. After washing, each well underwent further incubation and washing cycles with biotinylated rat anti-mouse IFN-γ (BD Pharmingen, Franklin Lakes, NJ, USA) and then ExtrAvidin™ peroxidase (Sigma-Aldrich, Gillingham, UK).Five milligrams of 3-Amino 9-ethyl-carbazole was dissolved in 0.5ml N, N Dimethylforamide, added to 10ml 0.05M sodium acetate buffer (pH5.0), filtered and 10μl 30% H₂O₂ (all Sigma-Aldrich, Gillingham, UK) was added to the resulting solution just prior to addition to the plate. After incubation at room temperature the plate was washed with distilled water and the backing was removed. After drying overnight the plates were read (Autoimmun Diagnostika GmbH, Straβberg, Germany), and data were expressed as spot counts per 10⁶ responder CD8 T cells for each well.

Statistical analysis

Graft survival was depicted using Kaplan-Meier analysis and groups compared using the log-rank test. Serial ELISA and HEp-2 data were analyzed by calculating the area under the curve for each animal; groups were then compared using Student’s two-tailed t test. Likewise, Student’s t test was used to compare mean TA scores and ELISPOT counts. P values < 0.05 were considered statistically significant. Statistical analyses were performed using GraphPad Prism (GraphPad Software, San Diego, CA, USA).

Online supplemental material

Supplemental Figure 1 compares the alloantibody response and allograft survival in wild-type and FcγRIIb^−/− B6 recipients of CBA/Ca heart allografts.

Results

FcγRIIb signaling does not influence acute cardiac allograft rejection

The role of FcγRIIb signaling in inhibiting the immune response to vascularized allografts was first studied using a well-characterized murine model of acute heart graft rejection. Completely MHC-mismatched BALB/c (H-2^d) heart grafts provoke strong early IgG and CD8 T cell responses in C57Bl/6 (B6) recipients and are rejected acutely, with a median survival time (MST) of 7-8 days (32). Prior publications have highlighted augmented cellular and humoral immunity in B6 mice genetically deficient in FcγRIIb (10, 11), but despite this, kinetics of BALB/c heart graft rejection in FcγRIIb^−/− recipients were similar to those in WT B6 recipients (MST 8 days vs 7 days; p=0.17, Figure 1A). Similarly, B6 recipients with transgenic overexpression of FcγRIIb on B cells (BTG mice (24), described in greater detail below) did not reject BALB/c heart grafts more slowly than either non-transgenic littermate control (BTNG LMC) or WT B6 recipients (p = 0.45 and p = 0.97, respectively, Figure 1A).

Figure 1. The impact of FcγRIIb signaling on murine acute cardiac allograft rejection.

(A) BALB/c heart graft survival in FcγRIIb^−/− (FcGKO, n = 12), B6 (n = 7), BTG (n=6, see figure 3 legend for details), and BNTG (n = 5) recipients. (B) Serum IgG anti-K^d alloantibody responses in FcγRIIb^−/− (FcGKO, n = 12), B6 (n = 7), BTG (n=6) and BNTG (n=5) recipients of BALB/c heart allografts. Data is represented as mean ± SD. There were no statistically significant differences (log-rank analysis) between the FcγRIIb^−/− and B6 curves (p = 0.25), and BTG and BNTG curves (p = 0.32). (C) Flow cytometric analysis of serum IgG alloantibodies to BALB/c thymocyte targets in FcγRIIb^−/− (FcGKO, n = 3) and B6 (n = 4) recipients of BALB/c heart grafts 42 days after transplantation. Data is expressed as mean (±SD) geometric mean channel fluorescence; the positive standard was the same serum as used for the ELISA analysis, with serum from naive FcγRIIb^−/− and B6 mice used as negative controls. (D) IFN-γ responses of splenic CD8 T cells harvested from B6 (n = 3) and FcγRIIb^−/− (FcGKO; n = 6) recipients of BALB/c heart grafts 14 days after transplantation, with responses from naïve mice depicted as control. Each point represents the mean of six wells from one animal, with data expressed as spot counts per 10⁶ responder cells. Bars represent mean (±SD) counts.

Surprisingly, IgG alloantibody responses against donor MHC class I H-2K^d antigen were not augmented in FcγRIIb^−/− recipients (Figure 1B), even at late time points when the formation of alloantibody/alloantigen complexes would be expected to deliver more pronounced inhibitory FcγRIIb signaling (10, 11, 24). This unanticipated finding was not a peculiarity of the response against MHC class I alloantigens, because the global alloantibody response in the FcγRIIb^−/− recipients, as assayed by flow cytometric detection of binding to BALB/c targets, was comparable to WT recipients (Figure 1C). Anti-K^d alloantibody responses were also similar in BTG and BNTG recipients and not different from WT recipients (Figure 1B). Finally, cellular alloimmune responses in the recipient mice were determined by assaying development of CD8 T cell cytotoxicity against BALB/c targets. Responses in FcγRIIb^−/− and WT B6 recipients were comparable (Figure 1D).

The inability of FcγRIIb signaling to regulate acute heart allograft rejection was not restricted to BALB/c donors, as there were no differences between FcγRIIb^−/− and WT B6 recipients in either graft survival or alloantibody responses upon challenge with CBA/Ca (H-2^k) heart grafts (Supplementary Figures 1A, B, and C).

FcγRIIb^−/− signaling modifies chronic cardiac allograft rejection

FcγRIIb^−/− signaling in autoantibody-mediated allograft arteriopathy

Although FcγRIIb signaling did not influence acute cardiac allograft rejection, we thought it likely that rejection occurred before the immune complexes responsible for mediating inhibitory FcγR signaling had formed and thus hypothesized that chronic humoral responses underpinning allograft arteriopathy would be more susceptible to such immune modulation. This was examined using a well-characterized mouse model of chronic heart allograft rejection in which the donor bm12 strain differs from the recipient B6 by only three amino acids on the MHC class II I-A β chain (25). Transplanted hearts develop marked neointimal proliferation resembling human TA and reject more than 4 weeks after transplantation (33). We have recently shown that, due to the lack of a B cell epitope, B6 recipients do not mount an alloantibody response to bm12 heart allografts, but that transplantation instead triggers anti-nuclear autoantibody, which contributes to the progression of TA and is dependent upon help provided by donor CD4 T cells for its development (29, 34).

Contrary to our findings in the acute rejection models, FcγRIIb^−/− mice rejected bm12 heart grafts more rapidly than LMCs or WT recipients (Figure 2A). Histological analysis confirmed that accelerated rejection was associated with development of more severe allograft arteriopathy (Figures 2B and C). Similarly, bm12 heart allografts provoked an augmented autoantibody response in FcγRIIb^−/− mice; a difference that was apparent even at early time points (Figure 2D).

Figure 2. Graft survival, allograft vasculopathy, and autoantibody responses in FcγRIIb^−/− and B6 recipients of MHC class II-mismatched bm12 heart allografts.

(A) Bm12 cardiac allograft survival in FcγRIIb^−/− (FcGKO, n=8), littermate control (LMC, n=4), and wild-type (WT) B6 recipients (n=17), with rejection significantly more rapid in FcγRIIb^−/− recipients that in LMC and WT B6 recipients (p = 0.03 and p < 0.01, respectively, log-rank analysis). (B) Representative EVG-stained paraffin sections of bm12 heart allografts explanted from FcγRIIb^−/− (left) and LMC (right) recipients 50 days after transplantation (x20 magnification), depicting marked neointimal proliferation and obliteration of the vascular lumen in the allograft from the FcγRIIb^−/− recipient. The bars represent a 200 μm scale. (C) Morphometric analysis of arterial luminal stenosis in bm12 heart allografts explanted 50 days after transplantation from FcγRIIb^−/− (FcGKO, n=5) and LMC (n = 4) mice. Each point represents mean percentage vascular luminal stenosis from a single explanted heart, with mean (±SD) indicated by line bars (FcGKO vs LMC, p = 0.02, Student’s t test). (D) Autoantibody production of FcγRIIb^−/− (FcGKO, n = 5) and LMC mice (n = 4) following challenge with bm12 heart allografts, with individual time points representing mean ± SD (FcGKO vs LMC, p < 0.01, Student’s t test area under the curve analysis).

We next sought to determine if over-expression of FcγRIIb would inhibit the development of autoantibody and TA, by using, as recipients, B6 mice in which FcγRIIb was transgenically overexpressed in either the macrophage or B cell lineages (24). This approach conferred the additional advantage of controlling for the potential confounding effect of integration of the lupus susceptibility locus from the 129 strain that persists despite back-crossing onto the B6 background (35-37). BTG mice express approximately twice as many FcγRIIb receptors on the majority of B cells while 10-20% of B cells have a 10-fold increase in expression. Other cell types do not express the transgene (24). Macrophage transgenic mice (MPTG) have modestly increased expression of FcγRIIb on peritoneal, bone-marrow and splenic macrophages, and have normal numbers of macrophages (24).

Bm12 heart grafts transplanted into BTG or MPTG recipients were rejected more slowly than B cell non-transgenic (BNTG) (Figure 3A) or macrophage non-transgenic (MPNTG) LMCs (Figure 3B, BTG v BNTG p = 0.17, MPTG v MPNTG p = 0.05). Surprisingly, non-transgenic BNTG and MPNTG littermate control recipients rejected bm12 heart allografts more rapidly than WT control recipients (cf Figure 2a), possibly reflecting slight genetic differences between different colonies of C57Bl/6 mice (38) used in backcrossing the macrophage and B cell transgenic mice. In comparison to BNTG recipients, the autoantibody response to bm12 heart grafting in the BTG group was diminished, although this failed to reach statistical significance (Figure 3C). MPNTG mice tended to produce less autoantibody than MPTG mice at early time points after challenge with bm12 heart grafts (Figure 3D), but this difference was not statistically significant and there was marked variation in the responses between individual animals at later time points. Histopathological analysis revealed significant arteriopathy within all heart allografts (mean +/− SD vascular stenosis 57 +/− 22 %), but the analysis was limited to those hearts still beating at excision at day 100, preventing meaningful comparison between the transgenic and non-transgenic littermate control groups.

Figure 3. Bm12 heart graft survival and autoantibody responses in B6 recipients over-expressing FcγRIIb either on B cells (BTG) or macrophages (MPTG).

(A) Bm12 heart graft survival in BTG (n = 8) and littermate control (BNTG (n = 9)) recipients. (B) Bm12 heart graft survival in MPTG (n = 9) and littermate control (MPNTG (n = 8)) recipients. (C) Autoantibody production (mean ± SD) in BTG (n = 8) and BNTG (n = 9) recipient of bm12 heart allografts (BNTG vs. BTG, p = 0.07; Student’s t test analysis of area under the curve). (D) Mean (± SD) autoantibody production of MTG (n = 8) and MNTG (n = 7) mice after transplantation of bm12 heart grafts (MPTNG vs. MPTG, p = 0.93).

FcγRIIb^−/− signaling in alloantibody mediated allograft arteriopathy

Because FcγRIIb deficiency on the B6 background augments humoral immunity and susceptibility to autoimmunity (35, 37), it is perhaps unsurprising that the autoantibody-mediated allograft arteriopathy that characterizes bm12 heart graft rejection in B6 hosts was exacerbated in FcγRIIb^−/− recipients. Therefore to ensure that the findings held wider relevance for chronic allograft rejection in which alternative mechanisms of rejection were active, a second model of chronic rejection was utilized. B6.K^d express MHC class I H-2K^d as a transgene and Honjo et al have demonstrated that B6 recipients reject MHC class I-mismatched B6.K^d heart grafts with an MST of 78 days (27).

We found that B6 recipients were unable to reject B6.K^d heart grafts within 100 days after transplantation. In contrast, heart grafts in FcγRIIb^−/− recipients were rejected slowly (MST 89 days, Figure 4A). Notably, whereas WT recipients of B6.K^d heart grafts mounted only a weak IgG anti-K^d alloantibody response, FcγRIIb^−/− recipients developed a much stronger response from day 28 onwards (Figure 4B). Neither WT nor FcγRIIb^−/− recipients of B6.K^d heart grafts mounted autoantibody responses (data not shown), presumably because graft-versus-host recognition of the disparate MHC class I alloantigen by donor CD4 T cells does not occur. B6.K^d heart grafts harvested 100 days after transplantation had developed more severe arteriopathy in FcγRIIb^−/− recipients (Figure 4C), but this was not statistically significant (p = 0.12), and immunohistochemical analysis confirmed complement C4d deposition, whereas deposition in B6.K^d hearts excised from WT recipients was negligible (Figure 4D). Because B6.K^d hearts provoked only negligible responses in WT B6 mice, transplantation into BTG and MPTG recipients was not performed. The results from this second model provide further support that FcγRIIb signaling influences adaptive immune responses underpinning chronic, but not acute, allograft rejection.

Figure 4. Graft survival, alloantibody responses and allograft vasculopathy in FcγRIIb^−/− and B6 recipients of MHC class I-mismatched B6.K^d heart allografts.

(A) B6.K^d cardiac allograft survival in FcγRIIb^−/− (FcGKO, n = 5), and WT B6 recipients (n = 15) (FcGKO vs B6, p < 0.01, log rank test). (B) Serum IgG anti-K^d alloantibody responses (mean ±SD) to B6.K^d heart grafts in FcγRIIb^−/− (FcGKO, n = 5) and B6 (n= 15) recipients (FcGKO vs B6, P< 0.01, Student’s t test area under the curve analysis). (C) Arterial luminal stenosis within B6.K^d allografts transplanted into FcγRIIb^−/− (n = 5) or B6 (n = 15) recipients was assessed 100 days after transplantation, as described in Figure 2 (FcGKO vs B6, p = 0.12, Student’s t test). (D) Photomicrographs depicting representative (of at least three hearts per group) immunohistochemical analysis of arterial complement C4d deposition (brown staining) within heart allografts excised 100 days after transplantation from (i) B6 heart allograft transplanted in B6 recipient (ii) B6.K^d heart allograft transplanted in B6 recipient (iii) ) B6.K^d heart allograft transplanted in FcγRIIb^−/− recipient. The bars represent a 200 μm scale.

The magnitude of T cell help determines the role of FcγRIIb in allograft rejection

The failure of FcγRIIb signaling to either attenuate rejection or inhibit alloantibody production in the models of acute allograft rejection is surprising; the inhibitory role of FcγRIIb in the humoral responses to model protein antigens is well documented. We hypothesized that the complex nature of cellular alloimmunity, in which alloantigen is uniquely recognized by two distinct pathways (39, 40), may prime a helper CD4 T cell response so effectively that it is not susceptible to inhibitory FcγRIIb signaling. This was examined by challenging FcγRIIb^−/− and WT B6 mice with 10⁷ BALB/c splenocytes intravenously, which in the WT recipients provoked a substantially weaker IgG anti-K^d alloantibody response than occurred following heart transplantation (c.f. Figures 1B and 5A).

Figure 5. CD4 T cell help modulates the inhibitory impact of FcγRIIb signaling.

A) Serum IgG anti H-2K^d alloantibody responses (mean ± SD) in FcγRIIb^−/− (FcGKO, n = 4) and wild-type B6 mice (n = 5) following i.v. challenge with 10⁷ BALB/c splenocytes (B6 vs FcGKO, p < 0.01, Student’s t test). (B) B6.K^d cardiac allograft survival in FcγRIIb^−/− (FcGKO) and WT B6 recipients administered 10³, 10⁴, or 10⁵ TCR75 CD4 T cells i.v. 1-2 days after transplantation. Graft survival in those hearts functioning at cull on day 42 is depicted as 42 days; the bar indicates median survival. (C, D & E) Serum IgG anti H-2K^d alloantibody responses to B6.K^d heart allografts in: (C) FcγRIIb^−/− (FcGKO, n = 6) and WT B6 mice (n = 6) adoptively-transferred with 10³ TCR75 CD4 T cells (FcGKO vs B6, p < 0.01, Student’s t test); (D) FcγRIIb^−/− (FcGKO, n = 4) and WT B6 mice (n = 5) adoptively transferred with 10⁴ TCR75 CD4 T cells (FcGKO vs B6, p = 0.02, Student’s t test); (E) FcγRIIb^−/− (FcGKO, n = 4) and WT B6 mice (n = 5) adoptively transferred with 10⁵ TCR75 CD4 T cells (p = 0.86). (F) Scatter plot showing arterial luminal stenosis within B6.K^d allografts harvested 42 after transplantation into FcγRIIb^−/− (FcGKO) or B6 recipients adoptively transferred with 10³, 10⁴, or 10⁵ TCR75 CD4 T cells. Each point represents mean percentage vascular luminal stenosis from a single explanted heart, with mean and SD indicated by line bars.

Unlike the response to a heart graft, the weaker allogeneic response provoked by splenocyte injection was susceptible to inhibitory FcγRIIb signaling, in that the IgG anti-K^d alloantibody response in FcγRIIb^−/− mice to i.v. splenocytes was substantially greater than observed in similarly-challenged WT B6 mice (Figure 5A).

To formally examine whether the inhibitory impact of FcγRIIb signaling is lost with increasing levels of CD4 T cell help, a series of experiments was performed in which FcγRIIb^−/− or WT B6 recipients of B6.K^d cardiac allografts were adoptively transferred at transplantation with varying numbers (10³, 10⁴ or 10⁵) of an additional population of helper CD4 T cells. As a helper population, monoclonal populations of CD4 T cells from RAG1^−/− TCR75 mice were transferred. TCR75 T cells recognize the dominant H-2K^d peptide presented in the context of self-I-A^b (26), and we have recently shown that they act as T follicular helper (T_FH) cells to provide help for germinal center alloantibody responses (41). Kinetics of cardiac allograft rejection, anti-K^d alloantibody responses, and allograft arteriopathy were then determined. Compared to recipients that received no additional CD4 T cells, transfer of TCR75 CD4 T cells accelerated B6.K^d heart graft rejection in both FcγRIIb^−/− and WT B6 recipients, even when only 10³ cells were transferred (c.f. Figures 5B and 4A), confirming that the adoptively-transferred cells influence the alloimmune response to the B6.K^d graft. Notably, the difference in graft survival between FcγRIIb^−/− and B6 mice was maintained when low numbers (10³) of TCR75 T cells were transferred (MST 20 days vs >42 days, p<0.01), but with increasing numbers of TCR75 T cells, the rejection times became similar (MST 10 days, Figure 5B). These findings were reflected in IgG anti-K^d alloantibody production, with the response in FcγRIIb^−/− recipient mice receiving 10³ or 10⁴ TCR75 CD4 T cells greater than that observed in the comparable WT B6 recipients (Figures 5C and 5D), but similar in mice receiving 10⁵ cells (Figure 5E). The severity of allograft arteriopathy also increased with the number of TCR75 CD4 T cells transferred (Figure 5F), but again, the degree of luminal stenosis was only different in the FcγRIIb^−/− and B6 groups that received 10³ TCR75 T cells. These experiments therefore demonstrate that increasing levels of CD4 T cell help abrogate the impact of inhibitory FcγR signaling, and may explain the ability of FcγRIIb to modulate chronic, but not acute cardiac allograft rejection.

Discussion

The family of Fc receptors for IgG provides important immunoregulatory control of adaptive immunity (2). Our study assesses, for the first time, the contribution of FcγRIIb signaling to rejection of cardiac allografts. In accordance with current understanding of FcγRIIb as the principal Fc receptor for regulating humoral immunity, we demonstrate that loss of this inhibitory signal augments chronic allo and autoantibody effector responses, with a corresponding increase in the severity of TA and more rapid graft rejection. The lack of impact of FcγRIIb signaling on alloantibody production and on the kinetics of rejection in the models of acute allograft rejection was unexpected, and subsequent experiments highlighted a novel mechanism whereby the inhibitory impact of signaling through FcγRIIb is dependent on the level of CD4 T cell help available.

In FcγRIIb deficient recipients, the augmented alloantibody and autoantibody responses to, respectively, the MHC class I- and MHC class II-mismatched heart grafts probably explain the increased severity of arteriopathy and accelerated graft rejection. Murine models have clearly highlighted an association with alloantibody and the development of allograft arteriopathy (22, 23), although the causative link between the two remains unclear. The contribution of autoantibody to allograft arteriopathy is more controversial (42), but we have recently described IgG and complement deposition on the endothelium of MHC class II-mismatched heart grafts, despite lack of demonstrable alloantibody (29). However, loss of FcγRIIb signaling probably also amplifies a number of additional effector mechanisms that contribute to progression of TA. Without tonic FcγRIIb signaling, DCs up-regulate expression of co-stimulatory ligands and acquire an activated, pro-inflammatory phenotype (43-45). Similarly, FcγRIIb inhibits B cell antigen processing and presentation (46) and consequently, self-restricted ‘indirect-pathway’ CD4 T cell responses to processed alloantigen (40), which may autonomously effect development of allograft arteriopathy (47), are likely augmented. FcγRIIb deficient B6 recipients of B6.K^d heart grafts probably also mount more aggressive cytotoxic cellular responses, because FcγRIIb expression on DCs influences CD8 T cell differentiation (48). Our observation that heart graft rejection was delayed in the two mutant strains overexpressing FcγRIIb on either macrophages alone or B lymphocytes alone is thus informative; confirming a role in allograft rejection for FcγRIIb mediated through the B cell, while simultaneously highlighting the potential contribution of signaling through other cell types. Such additional signaling likely delays allograft rejection through predominantly inhibiting cellular alloimmunity; notably the autoantibody response in the macrophage transgenic recipients was similar to that in the non-transgenic littermate controls, and our previous work, demonstrating, for example, that bm12 heart allografts still develop arteriopathy when transplanted into B-cell deficient recipients (29), has confirmed that conventional cellular responses also contribute to bm12 heart allograft rejection. Future studies will examine the role of inhibitory FcγRIIb signaling in regulating cellular alloimmunity more formally, by comparing development of transplant arteriopathy in WT and FcγRIIb^−/− recipients of CD4-T cell depleted bm12 heart allografts that do not provoke an autoantibody response (34).

Despite the profound inhibition of chronic humoral allo- and auto-responses, FcγRIIb signaling made no discernible impact in the models of acute heart allograft rejection, in that WT and FcγRIIb recipients mounted similar cellular and humoral alloimmune responses, with grafts rejected equally rapidly in both groups. As explanation for this unexpected result, we reasoned that the immunological stimulus provided by a heart graft was so strong as to provoke supra-optimal helper T cell activation resistant to inhibitory FcγRIIb signaling. Subsequent experiments with adoptive transfer of allopeptide-specific CD4 T cells supported this hypothesis, by highlighting that inhibitory signaling through FcγRIIb is modified by availability of CD4 T cell help. This has not been described previously, but notably IL-4, one of the key cytokines released by helper T cells, can ameliorate FcγRIIb-mediated B cell suppression in vitro (49).

The observation that CD4 T cell responses modulate FcγRIIb signaling raises an interesting question regarding the point at which the helper and inhibitory signals intersect. One of the difficulties in addressing this is that previous publications have implicated FcγRIIb in controlling many stages of the B cell response: pre-B cell repertoire selection (50); follicular seeding of immature B cells (51); the initial peri-follicular response at the T/B border (52, 53); the germinal center (24, 54); and plasma cell survival (55). Because CD4 T cell survival signaling is reliant upon recognition of processed target antigen following BCR-mediated internalization, it most likely influences FcγRIIb inhibition at either the peri-follicular or germinal center stage. Of these, modulation of the germinal center response is more compelling, as it provides an elegant compatibility with the proposed model whereby centrocyte selection upon encounter with immune-complexes on follicular DCs is determined by the relative input of positive signaling from BCR ligation against inhibitory FcγRIIb signaling. Key to selection is competition for help (56), provided by limiting numbers of highly-specialized T_FH (57, 58) that are dependent upon interaction with B cells for migration into the follicle (59). Thus increasing numbers of T_FH may establish a less competitive environment and overcome the selection disadvantage conferred by dominant FcγRIIb signaling on individual centrocytes. Notably, bm12 heart graft rejection is characterized by development of long-lasting and numerous germinal centers (34) and we have recently highlighted that adoptively-transferred TCR75 T cells acquire a T_FH phenotype in response to challenge with H-2^d heart grafts (41). This hypothesis that T cell modulation of inhibitory FcγRIIb signaling is mediated exclusively by T_FH CD4 T cells could be tested by transfer of monoclonal populations of helper CD4 T cells that do not express SAP (signaling lymphocyte activation molecule-associated protein) and are consequently unable to acquire a T_FH phenotype (60).

The mouse and human Fc gamma receptor systems differ subtly, but given the general similarities, particularly the presence of only one inhibitory receptor in both species, our findings in the mouse may be relevant to clinical transplant practice. Specifically, polymorphisms in the putative promoter and regulatory regions of Fcgr2b are linked in mice to downregulation of FcγRIIb expression and development of spontaneous autoimmunity (61, 62). Similarly in humans, a single nucleotide polymorphism that results in substitution of threonine for isoleucine at position 232 within the transmembrane domain of FcγRIIb (referred to as FcγRIIb^T232) results in exclusion from the lipid rafts and subsequent impairment of FcγRIIb-mediated inhibition (63). This polymorphism is more prevalent within African and Southeast Asian populations (12, 64), and is associated with both with survival benefit to malaria, and, when expressed homozygously, SLE (1, 8). Our study infers that graft survival in transplant patients with the FcγRIIb^T232 polymorphism may be poorer, which if confirmed would merit consideration of more aggressive immunosuppressant administration in this cohort.

Our results, however, raise the more exciting prospect that novel strategies aimed at augmenting inhibitory FcγRIIb signaling could improve long-term transplant outcomes. In support, in autoimmune-prone mice, reconstitution with syngeneic bone marrow that had been transduced with FcγRIIb-expressing retrovirus (65) or transfer of FcγRIIb-overexpressing immature DCs (66) resulted in abrogation of autoantibody and amelioration of autoimmune disease. An immediate problem with targeting FcγRIIb clinically is avoiding simultaneous, and perhaps overriding, activating signaling from binding to other FcγRs. This might be solved by administering particular IgG isotypes, such as IgG1, that have a low activation to inhibitory ratio (67), or by engineering the Fc domain of therapeutic antibodies to bind FcγRIIb almost exclusively (68). Finally, antibody directed against human FcγRIIb has now been developed and its recent use as one limb of a bispecific antibody (69) to ameliorate development of murine collagen induced arthritis suggests that such strategies may also be successful in improving allograft survival.

Abbreviations

B6

C57Bl/6

BNTG

B cell non-transgenic

BTG

B cell transgenic

EVG

elastic van Gieson

FcγR

Fcgamma receptor

IEL

internal elastic lamina

LMC

littermate control

MPNTG

macrophage non-transgenic

MPTG

macrophage transgenic

MST

median survival time

TA

transplant arteriopathy

T_FH

T follicular helper