Cyclosporin A and tacrolimus reduce T-cell polyfunctionality but not interferon-γ responses directed at cytomegalovirus

Abstract

Cytomegalovirus (CMV) -specific immunity is often estimated by the number of in vitro CMV antigen-inducible interferon-γ-positive (IFN-γ⁺) T cells. However, recent work indicates that simultaneous production of IFN-γ, tumour necrosis factor-α (TNF-α) and interleukin-2 (IL-2) (referred to as ‘polyfunctionality’) is more relevant for anti-viral protection. Here, we compared polyfunctionality of CMV-specific T cells (pp65 and IE-1 proteins) in 23 solid-organ transplant patients and seven healthy controls by flow cytometry. The proportions of TNF-α⁺/IFN-γ⁺/IL-2 cells among the activated cells were significantly reduced in transplant patients but not the frequencies of IFN-γ⁺ CD8⁺ T cells. Immunosuppression reduces polyfunctionality, which reflects the increased infection risk in this patient group.

Introduction

In healthy individuals, CD4⁺ and CD8⁺ T cells restrain many infectious pathogens but in transplant patients these mechanisms are weakened by the immunosuppressive medication required to prevent graft rejection. Several studies have shown an association between diminished T-cell responses and cytomegalovirus (CMV) disease,^1,2 and the success of adoptive T-cell transfer in preventing CMV disease after bone marrow transplantation confirms these observations.^3,4 Prevention and treatment of CMV reactivation and disease significantly contribute to the high cost of transplantation.⁵

There is currently no clinical test for assessing the degree of immunosuppression, either in general or with respect to a specific pathogen. As regards CMV, most published studies in transplant patients are focused on the detection of CMV-specific T cells based on interferon-γ (IFN-γ) production (intracellular staining) or MHC-multimer staining and quantitative changes of the identified populations in relation to clinical events. Because CMV is large and complex, published studies are generally focused on one or two CMV proteins, usually pp65, sometimes also IE-1. However, there is controversy about how measuring the frequencies of CMV-specific IFN-γ-producing T cells will help to determine CMV-specific immunity. Several studies have linked increasing frequencies of CMV-specific T cells to decreasing rates of CMV detection or CMV-related complications after bone marrow or solid organ transplantation.^6,7

As T-cell polyfunctionality has been proposed to be important for protection from viral diseases, this study was designed to assess the effect of post-transplantation immunosuppression on T-cell polyfunctionality. Multi-parameter flow cytometry permits the assessment of response size and‘quality’ (functional composition).⁸ The use of a ‘qualitative’ approach is supported by results in HIV-positive patients suggesting that progression to AIDS correlates with the loss of HIV-specific CD8⁺ T cells with several simultaneous functions.⁹ Here, we considered the CMV-specific production of IFN-γ, tumour necrosis factor-α (TNF-α), interleukin-2 (IL-2) and degranulation of CD8⁺ and CD8⁻ T-cells at the same time in 23 heart and heart–lung transplant patients and seven healthy controls in response to pp65 and IE-1. This allows us to detect potential differences in functional profiles relating to different CMV specificities.

Materials and methods

Patient and donor recruitment

All heart (n = 16) and lung (n = 7) transplant recipients (eight women, 15 men; mean age 51·2 years, minimum 18 years, maximum: 64 years) were recruited at the German Heart Centre (DHZB) Berlin. All had been CMV-seropositive (IgG) before transplantation. Fourteen patients received a graft from a CMV-positive donor. Immunosuppression consisted of cyclosporin A (22/23 patients), tacrolimus (1/23), everolimus (7/23), mycophenolate mofetil (8/23) and corticosteroids (23/23). Seventeen patients had PCR-proven CMV reactivation and two suffered from clinical disease (duodenitis).

Healthy volunteers (three women, four men) known to have T-cell responses to CMV pp65 or IE-1 (n = 7) included hospital personnel and medical students. No significant differences between the groups existed in terms of gender distribution. The control group was on average younger than the transplant group (mean age: 34·1 years, minimum: 19 years, maximum: 55 years). The study was covered by the Charité Ethics Committee and in agreement with the declaration of Helsinki.

Peripheral blood mononuclear cell preparation

Blood was drawn into vacutainers (BD, Heidelberg, Germany) containing sodium citrate for anticoagulation. Peripheral blood mononuclear cells were separated using density centrifugation (Ficoll-Paque; Pharmacia, Uppsala, Sweden), suspended in supplemented RPMI-1640 medium [containing 2 mm l-glutamine, 10% (volume/volume) heat-inactivated fetal calf serum (FCS) and 100 IU/ml penicillin/streptomycin] and pre-incubated overnight at 37°.

Antibodies

Fluorochrome-conjugated antibodies were obtained from the following companies: CD3-PacificBlue, CD45-peridinin chlorophyll protein (PerCP), TNF-α-PeCy7, IL-2-phycoerythrin (PE), CD8-allophycocyanin-Cy7 (APCCy7), CD107a/b-FITC, CD8-PerCP, Perforin-PE, GranzymeA-FITC and GranzymeB-Alexa700 were from BD Biosciences (San Jose, CA); CD28-Texas red-PE was from Beckmann Coulter (Fullerton, CA); and IFN-γ-APC was from IQ Products (Groningen, the Netherlands).

Peptides

Lyophilized peptide pools (15mers with an 11 amino acid overlap) representing the pp65 or IE-1 protein of CMV (Swiss-Prot Accession nos. P06725 and P13202) were purchased from JPT (Berlin, Germany) and diluted in DMSO (1 μg of each peptide per test) and used at a total volume of 4 μl. CMV specific epitopes were synthesized as free acids with > 95% purity (JPT) and used at a concentration of 1 μg/test.

Cell stimulation and staining

Cytokine production and degranulation were assessed in parallel as described previously.^10,11 Four hundred microlitres of peripheral blood mononuclear cell suspension (5 × 10⁶ cells/ml) were stimulated with pp65 or IE-1 peptide pools dissolved in DMSO (Perbio Science, Bonn, Germany) in the presence of monensin (Golgistop, 1 μl/ml; BD Biosciences) and anti-human CD107a/b-FITC for 2 hr at 37°. Stimulation with staphylococcus enterotoxin B (Sigma-Aldrich, Taufkirchen, Germany) was used as a positive control, DMSO (equivalent to the amount added with peptide pools) was added to the unstimulated samples (negative control). After the addition of Brefeldin A (10 μg/ml; Sigma), samples were incubated for another 4 hr and then washed (PBS containing 0·5% bovine serum albumin and 0·1% sodium azide) and stained with surface antibodies for 30 min at 4°. After washing, lysis and permeabilization (Perm 2 and Lysis; BD Biosciences, according to manufacturer’s instructions) cells were stained intracellularly (30 min, 4°). Following staining, the cells were washed, fixed (PBS with 0·5% paraformaldehyde) and stored on melting ice until sample acquisition.

Data acquisition and analysis

All samples were measured on an LSRII flow cytometer (BD). FlowJo software (Treestar, Ashland, OR) was used for data analysis. Cell doublets were excluded using forward scatter height versus forward scatter area. Leucocytes were gated using CD45 expression versus side scatter area. Lymphocytes were gated in a side scatter versus forward scatter scatter gate and further divided into CD8⁺ and CD8⁻ T cells according to CD3 and CD8 expression. CD8⁻ T cells (representing mainly T helper cells) were also analysed, although they were not the main focus of this work.

The frequency of cells expressing a certain marker was calculated in relation to the number of cells in the relevant subset. Unstimulated samples were used as negative controls. spss 18.0 software was used for statistical analysis and P-values were corrected for multiple testing (Bonferroni-correction).

Results

The overall frequencies of inducible CMV-specific CD8⁺ T cells are not significantly changed by immunosuppression

For the purpose of this study heart and heart–lung recipients were generally treated as one group (transplant patients). This study was focused on CD8⁺ T cells but pp65-specific CD8⁻ T cells were also explored. However, IE-1-specific CD8⁻ T cells were detected infrequently and the numbers were small, so this subset was not analysed further.¹² The frequencies of inducible pp65-specific or IE1-specific CD8⁺ T cells or pp65-specific CD8⁻ T cells were subject to large inter-individual variation. A trend towards smaller frequencies of IFN-γ-producing, TNF-α-producing or IL-2-producing IE1-specific CD8⁺ T cells in transplant patients was observed, but this was not true for pp65-specific CD8⁺ or CD8⁻ T-cell responses. None of the observed differences was statistically significant (Fig. 1a). No difference was observed between patients who had received a CMV⁺ or a CMV^– graft (not shown).

Figure 1.

The size and composition of the T-cell response to pp65 and IE-1 is marked by large inter-individual variation in both groups. (a) Intracellular interferon-γ (IFN-γ), tumour necrosis factor-α (TNF-α), interleukin-2 (IL-2) and degranulation were used to enumerate CD8 T cells specific for the cytomegalovirus (CMV) proteins pp65 or IE-1. Relative values are given in % of all CD8⁺ or CD8⁻ T cells. Values for ‘at least one function’ combine all cells that express any of the measured markers. Diagrams show median (bars) and 95% confidence interval (error bars). (b) T cells producing IFN-γ and TNF-α at the same time are drastically reduced after transplantation. Subsets producing IFN-γ/TNF-α/IL-2 (+/− degranulation), or IFN-γ/TNF-α are markedly decreased after transplantation. Bar charts: bars show medians, error bars show 95% confidence intervals. Bonferroni correction with a factor of 4 was applied and the significance level was set to 0.0125. Only differences significant at this level are indicated. (c) Fifteen distinct subsets are derived from Boolean combination of activation markers. Polyfunctional subsets are decreased and monofunctional subsets are increased after transplantation. The most striking changes occurred with respect to CD8⁺ and CD8⁻ T cells exclusively displaying degranulation. Bonferroni correction with a factor of 15 was applied and the significance level was set to 0.003. Only differences significant at this level are indicated.

Degranulating cells were much more frequent than cytokine-producing cells in transplant patients but not in controls

Interferon-γ is a frequently used read-out for T-cell activation in the transplant setting; the median frequencies of CD8⁺ and CD8⁻ T cells exhibiting ‘at least one marker’/IFN-γ-positive cells in % of the reference subset (either all CD8⁺ or CD8⁻ T cells) were as follows, CD8⁺/pp65: transplant group 1·05/0·25, control 0·35/0·26; CD8⁺/IE-1: transplant group 0·58/0·14, control 0·70/0·52; CD8⁻/pp65: transplant group 0·34/0·14, control 0·43/0·18. Of interest, the differences in frequency between degranulating and IFN-γ-producing cells were significant in transplant recipients but not in controls (Fig. 1a). The same was true for the frequencies of degranulating compared with TNF-α-producing or IL-2-producing cells. With respect to pp65-specific CD8⁺ T cells all the same differences were also significant in heart recipients analysed separately. The lung recipients were a smaller group and not all of the same differences (though suggested by the data) were significant, in particular the differences with respect to pp65-specific CD8⁻ T cells did not reach statistical significance (not shown). Of note, frequencies of IFN-γ⁺ T cells were significantly higher than IL-2⁺ T cells within the CD8⁺ subset of transplant patients for both antigens tested (P = 0·0006 for pp65 and P = 0·005 for IE1). Differences for the pp65 CD8⁻ T cells were non-significant (P = 0·144). In summary, the data clearly demonstrated that degranulation of CD8⁺ T cells was the dominant function found under immunosuppression.

The immune response ‘flavour’ changes with immunosuppression

We next explored the proportions of T cells expressing each of the tested activation markers as a percentage of all activated cells (i.e. expressing at least one of the markers). This clearly confirmed that degranulation became increasingly dominant after transplantation, with a median of 92% of CD8⁺ pp65-specific T cells and 85% of IE-specific CD8⁺ T cells expressing this marker (alone or in combination) after transplantation compared with 84% and 71%, respectively, in controls (not shown).

However, because of their likely protective role, we were primarily interested in the effect of immunosuppression on the T cells producing IFN-γ, TNF-α and IL-2 simultaneously, or any two of them.⁹ For this purpose, the analyses shown in Fig. 1(b) disregard degranulation and focus on IFN-γ, TNF-α and IL-2 alone. They show that the most dominant CMV-specific CD8⁺ subset (as defined by these functions) in healthy donors produces just IFN-γ and TNF-α, while the subset producing all three measured cytokines is the only other sizeable subset. Both are strongly reduced in transplant patients. A similar distribution was observed for pp65-specific CD8⁻ T cells.

When studying each of 15 non-overlapping functional subsets individually (Boolean gating) it became apparent that T cells exhibiting degranulation as a single function were dramatically increased in transplant patients (Fig. 1c). As all patients received calcineurin inhibitors (but only one-third each received everolimus or mycophenolate mofetil), we attempted to reproduce this effect in vitro by incubating donor-derived cells overnight with the calcineurin inhibitors cyclosporin A or tacrolimus before stimulation, because these were the most likely drugs to cause this change. This resulted in a dose-dependent reduction of polyfunctionality; the subsets producing IFN-γ, TNF-α and IL-2, or IFN-γ and TNF-α decreased (Fig. 2a,b) whereas subsets displaying only single functions emerged and increased (Fig. 2c,d). Dot plots in Fig. 2(e) show a dose-dependent decrease in TNF-α, IFN-γ and IL-2 production, but little effect on degranulation.

Figure 2.

Calcineurin inhibitors abolish polyfunctionality in ex vivo overnight incubated cytomegalovirus (CMV) -specific T cells. (a–d) Ex vivo pre-incubation with the calcineurin inhibitors, cyclosporin A and tacrolimus (FK506), reduces the proportion of polyfunctional subsets by reducing interleukin-2 (IL-2), tumour necrosis factor-α (TNF-α) and interferon-γ (IFN-γ) production (a, b) and at the same time increases the proportion of cells with just one function, predominantly degranulation (c, d). Bars show medians, error bars show 95% confidence intervals. Donor cells (three different donors for each experiment shown) were incubated with the indicated concentrations of immunosuppressive agents over night before stimulation with pp65. The effects are dose dependent, and at near physiological concentrations, i.e. 200 ng/ml of cyclosporin A and 10 ng/ml of tacrolimus (FK506) were comparable to those observed in transplant patients on these medications. (e) Dot plots show degranulation versus TNF-α (top panel), IFN-γ (middle panel) and IL-2-production (bottom panel) in CD8 T cells of a healthy donor (one representative experiment of three is shown). x and y-axes show mean fluorescence intensity.

Discussion

Our results show that immunosuppression induces marked changes in the CMV-specific T-cell response after heart and lung transplantation. These are reflected in response quality (i.e. the functional response profile) rather than quantity (i.e. the number of inducible cells). The most obvious effects were reduction of IL-2 and TNF-α production, IFN-γ seemed somewhat less affected and degranulation not at all. This predominantly translated into the generation of T-cell subsets with one single function, most frequently degranulation, at the expense of subsets displaying IFN-γ, TNF-α and IL-2 at the same time. Degranulation was the most inclusive marker of total response size but not the most informative with regard to the effect of immunosuppression.

The importance of polyfunctional T cells in controlling infectious pathogens has been widely accepted¹³ and because anti-infectious immunity is weakened under immunosuppression, it was a logical step to explore if this is reflected by reduced polyfunctionality of the CMV-specific T-cell response. Interestingly, the overall frequencies of pp65 or IE-1 inducible IFN-γ⁺ CD8⁺ T cells were higher in healthy donors than in heart and lung transplant patients. As would be expected in a human population, there were large variations in the frequencies of these T cells in each group (see 95% CI intervals in Fig. 1a).

It was interesting to note, however, that in transplant patients most IFN-γ producing T cells had no other function, whereas in healthy donors they also produced TNF-α and degranulated. To explore this observation further, the number of cells displaying at least one of the measured activation markers was established (‘all activated cells’). Cells exhibiting a specific profile were expressed as a proportion of ‘all activated cells’. This approach has proven extremely useful for measuring response quality in a number of studies.^13,14 In our study, transplant patients had generally fewer ‘polyfunctional’ T cells than healthy controls, but much higher numbers of cells displaying only degranulation. Overnight incubation of peripheral blood mononuclear cells with cyclosporin A or tacrolimus also produced cells only exhibiting degranulation, along with smaller numbers of single cytokine producers, suggesting that these agents may be directly responsible for the effect observed in vivo. The effects of everolimus and mycophenolate mofetil were not analysed in the same way because the effect in question was sufficiently reproduced with calcineurin inhibitors.

The relative reduction of T-cell subsets producing IFN-γ and TNF-α, with or without simultaneous IL-2 production in transplant patients compared with healthy donors was obvious and highly significant, and could be reproduced in vitro by overnight incubation with cyclosporin A or tacrolimus. We believe this is one direct correlate of immunosuppression (and most likely failing defences), because exactly these subsets have been linked to protection after vaccination.⁹

Funding sources

The work was funded in part through Charité– Universitätsmedizin Berlin, Germany, and Brighton and Sussex Medical School, Brighton, UK.

Disclosures

H.D.V. and F.K. are inventors on a patent relating to the use of protein spanning peptide mixes and epitope mapping by flow cytometry.