Mycophenolic acid is a potent inhibitor of the expression of tumour necrosis factor- and tumour necrosis factor-receptor superfamily costimulatory molecules

Abstract

The tumour necrosis factor (TNF) ligands CD154, CD70 and TNF receptors CD134 and CD137 are all involved in allograft rejection. Because these molecules are not present on resting T cells, we investigated whether immunosuppressive drugs could inhibit their induction. Expression was induced in vitro on T cells by phorbol 12-myristate 13-acetate and ionomycin or by allogeneic dendritic cells in the presence or absence of cyclosporin A (CsA), tacrolimus (TAC), rapamycin derivative (SDZ RAD), or mycophenolic acid (MPA), and determined by flow cytometry. To study the effect of in vivo exposure to immunosuppressive drugs on these molecules, immunohistochemistry was performed on human lymph nodes from patients treated with TAC or TAC and MMF. The calcineurin inhibitors (CI) CsA and TAC strongly suppressed the induction of CD70, CD137 and CD154, but not of CD134, upon pharmacological stimulation of T cells in vitro. In allogeneic stimulations only CD137 and CD154 were inhibited by CI. SDZ RAD did not inhibit pharmacological induction, but in allogeneic stimulations all the investigated molecules were partially suppressed. Both in pharmacological and in allogeneic stimulations, MPA inhibited induction of all tested molecules on T cells nearly completely at 4 µg/ml. However, in lymph nodes obtained from patients chronically treated with MMF and TAC no reduction of the expression of these molecules was observed. This was possibly caused by trough levels which were in vivo lower (mean: 2·3 µg/ml) than the concentration giving complete inhibition in vitro. In conclusion, in contrast to CsA, TAC and SDZ RAD, MPA is a potent inhibitor of the induction of TNF- and TNF-receptor superfamily molecules on T cells. To obtain long-term suppression of these molecules in vivo, a plasma trough level of 4 µg/ml is indicated.

Introduction

Naïve T cells require two signals to become activated. The interaction between the T-cell receptor (TCR) and the peptide–major histocompatibility complex (MHC) complex on the antigen-presenting cell (APC) delivers the first signal. The predominant second signal is provided by binding of the costimulatory molecule CD28 to it's ligands CD80 or CD86 on the APC. After initial activation, the T cell, in its turn, stimulates the antigen-presenting capacity of the APC by binding of CD154 (CD40-ligand) on the T cell to CD40 on the APC.¹

Although CD28 and CD154 provide critical costimulatory signals in the initiation of T-cell responses, not all T-cell mediated immune responses are dependent on these molecules. For example, CD28-deficient mice are able to reject skin allografts², CD40- or CD154-deficient mice can mount T-cell responses to certain viral infections³ and blockade of CD154–CD40 interaction is insufficient to inhibit cardiac allograft rejection in a certain donor-recipient mouse strain combination.⁴ Especially CD8 T cells seem to be able to act independently from costimulation via CD28 or CD154.^4–6

Recently several other costimulatory molecule pairs have been identified. The majority of these belong, like CD40 and CD154, to the tumour necrosis factor (TNF)-receptor- and TNF superfamilies. In contrast to CD28, but like CD154, these molecules are not constitutively expressed on T cells, but after activation. The TNF-receptor member CD137 (4-1BB) is expressed on both activated T helper cells and cytotoxic T cells, and it's ligand 4-1BBL on activated APC. Signalling through CD137 preferentially induces CD8 responses^7,8 and can stimulate T cells even in the absence of CD28^9,10 suggesting redundancy in the costimulatory system. CD134 (OX40), another member of the TNF-receptor family, is thought to be important in sustaining, rather than in initiating, CD4⁺ T cell responses.¹¹ Ligation of CD134 by agonistic antibodies can even break an existing state of tolerance in CD4⁺ T-cell compartment.¹² The TNF ligand family member CD70 can be induced on CD8⁺ and CD4⁺ T cells.¹³ Its ligand, CD27, is constitutively expressed on T cells. Interaction between these two molecules provides costimulatory signals to both CD4⁺ and CD8⁺ T cells¹⁴ and is thought to regulate the size of antigen-primed lymphocyte populations.¹⁵ Using CD27^−/− mice, it has been demonstrated that the generation of both CD4⁺ and CD8⁺ T-cell responses against influenza virus is dependent upon the CD27/CD70 pathway.¹⁶

The critical importance of CD154–CD40 and CD28–CD80/CD86 interactions in rejection of organ-allografts has been shown in several rodent and non-human primate models.^17–20 The role of the newly identified costimulatory members of the TNF- and TNF-receptor families in transplant rejection is beginning to be studied. Administration of stimulating anti-CD137 monoclonal antibodies (mAbs) augmented acute graft versus host disease mice that had received bone marrow transplantation²¹ and accelerated rejection of skin and cardiac allografts in mice.^7,22 Blockade of CD134–CD134L interaction significantly ameliorated acute graft versus host disease in mice²³ and prolonged allogeneic islet graft survival in CD28-deficient mice.²⁴ Blockade of the CD27–CD70 pathway resulted in prolonged heart allograft survival in wild-type mice, and in indefinite survival in CD28-deficient recipients.²⁵

Because these TNF- and TNF-receptor members all may contribute to allograft rejection, but are, in contrast to CD28, not constitutively expressed on T cells, we investigated whether immunosuppressive drugs (immunosuppressive drugs) can inhibit their induction. We studied the effects of immunosuppressive drugs which have T cells as their main target, but inhibit T-cell activation by different mechanisms: calcineurin inhibitors (cyclosporin A (CsA), and tacrolimus (TAC)), a target of rapamycin (TOR)-inhibitor (rapamycin derivative (SDZ RAD)), and an inhibitor of purine biosynthesis (mycophenolic acid (MPA), the active metabolite of mycophenolate mofetil (MMF)). In addition to their effect on the induction of CD70, CD134, CD137 and CD154 on T cells in vitro, the influence of exposure to these drugs in vivo was investigated in lymphoid tissues obtained from patients treated with these immunosuppressive drugs.

Materials and methods

Cells and tissues

Peripheral blood mononuclear cells (PBMC) were isolated by Ficoll Hypaque density gradient centrifugation in Leucosep tubes (Greiner Bio-One B.V., Alphen a/d Rijn, the Netherlands) from buffycoat preparations obtained from healthy blood donors (Bloodbank ZWN, Rotterdam, the Netherlands). Three hepatic lymph nodes were obtained during re-transplantation of liver transplant patients treated for 2, 5 and 9 months with TAC and having trough levels of 7·1, 17·8 and 9·9 ng/ml on the day before retransplantation, respectively. Trough levels of TAC in these patients during the last two months before retransplantation ranged between 5·0 and 15·5 ng/ml (mean 10·4). Two inguinal lymph nodes were obtained during re-interventions from kidney transplant recipients treated for 5 months with MMF (2 × 1000, or 2 × 500 mg p.o. daily, respectively) and TAC, and having trough plasma levels of MPA ranging between 1·6 and 3·5 (mean 2·3) mg/l during the last three months before the re-intervention. As controls, five hepatic lymph nodes from multiorgan donors, a hepatic lymph node from a patient with acute hepatitis B virus infection obtained during the liver transplant procedure, and an inguinal lymph node obtained from a kidney transplant recipient during the transplant procedure were used. The lymph nodes were immediately frozen in isopentane (Fluka Chemie, Buchs, Switzerland) cooled in liquid nitrogen and stored at −80°. The study was approved by the Medical Ethical Committee of the Erasmus Medical Centre, and informed consent was obtained from all patients.

Effect of immunosuppressive drugs on the induction of TNF-receptor and TNF- family members on T cells by phorbol 12-myristate 13-acetate (PMA) and ionomycin

PBMC were suspended in a concentration of 10⁶ cells/ml in RPMI-1640 medium (Bio Whittaker Europe, Verviers, Belgium) supplemented with 2 mm l-glutamine, 10% fetal calf serum (Gibco BRL Life Technologies, Breda, the Netherlands), 100 U/ml pencillin and 100 µg/ml streptomycin (Gibco Life Technologies) (together called complete medium) and 50 U/ml interleukin-2 (IL-2; Chiron Proleukin, Amsterdam, the Netherlands). Three ml of cell suspension was transferred into each well of a six-well tissue culture plate (Costar, Schiphol-Rijk, the Netherlands) and immunosuppressive drugs were added to the culture medium. The immunosuppressive drugs tested were: CsA (kindly provided by Novartis Pharma AG, Basel, Switzerland, and dissolved in a 1 : 1 mixture of ethanol with 10% Tween-20 and water); TAC (kindly provided by Fujisawa Holland BV, Houten, the Netherlands, as intravenous infusion fluid); SDZ RAD (kindly provided by Novartis Pharma AG, dissolved in 100% ethanol); and MPA (the active metabolite of MMF, kindly provided by Roche Ltd, Basel, Switzerland, dissolved in dimethylsulphoxide). Cells were incubated with clinically relevant trough and peak levels of the tested immunosuppressive drugs: 50 or 500 ng/ml CsA, 10 or 60 ng/ml TAC, 5 or 20 ng/ml SDZ RAD²⁶ and 2 or 4 µg/ml MPA.²⁷ To control wells, the same amounts of solvents that were used to add immunosuppressive drugs, were supplemented. In the TAC control well, castor oil was added, as this is used to emulsify TAC in the infusion fluid. After 60 min of incubation, the cells were stimulated with 10 ng/ml PMA and 1 µg/ml ionomycin (both from Sigma, St Louis, MO) in a 60-hr culture at 37° with 5% CO₂. In preliminary experiments it was found that at this time-point, induction of all costimulatory molecules tested was maximal. After the culture, expression of costimulatory molecules was determined by flow-cytometry. The experiments were repeated (see Results) with PBMC from different blood donors.

Cultivation of dendritic cells (DCs) from monocytes

Monocytes were isolated from PBMC by Percoll (d = 1·064 g/ml, Amersham Pharmacia Biotec; Uppsala, Sweden) gradient centrifugation and cultured in a concentration of 0·5 × 10⁶ cells/ml in complete medium in the presence of granulocyte–macrophage colony-stimulating factor (GM-CSF; 500 U/ml, Biosource, Camarillo, CA) and IL-4 (250 U/ml, CLB, Amsterdam, the Netherlands). After 5 days freeze-dried Mycobacterium tuberculosis (H-37 RA, Difco Laboratories, Sparks, MD; 20 µg/ml) was added for 48 hr to mature DCs.

Effect of immunosuppressive drugs on the induction of TNF-receptor and TNF-ligand family members on T cells by allogeneic stimulation

PBMC were cocultured with irradiated (3000 Gy) mature monocyte-derived DCs at a ratio of 20 : 1 (10⁶ : 5 × 10⁴/ml) in a six-well tissue culture plate in RPMI supplemented with 10% pooled human serum which was tested to stimulate mixed lymphocyte reactions optimally (gift from Prof Dr F.H.J. Claas, Department of Immunohematology, Leiden University Medical Center, Leiden, the Netherlands) in the presence of IL-2 (50 U/ml) and CD40 mAb 5D12 (kindly provided by Dr M. de Boer, Tanox Pharma, Amsterdam, the Netherlands; 50 ng/ml). CD40 mAb was added to the cultures to prevent CD40–CD154 interaction, which may lead to down-regulation of CD154. immunosuppressive drugs were added in the indicated concentrations, and after 5 days cells were harvested expression of costimulatory molecules was measured by flow-cytometry.

Flow cytometry

Cells were washed with phosphate-buffered saline (PBS), suspended in a concentration of 10⁷ cells/ml in PBS supplemented with 0·3% bovine serum albumin (BSA; Bayer Corporation, Kankabee, IL) and 10% normal human plasma (Bloodbank ZWN), and incubated with CD154 mAb clone 89-76 (Becton Dickinson, San Jose, CA), CD137 mAb clone MCA1612 (Serotec, Breda, the Netherlands), CD134 mAb clone ACT35 (Becton Dickinson Pharmingen), CD70 mAb clone BU69 (Ancell, Lausen, Switzerland), or irrelevant immunoglobulin G1 (IgG1; CLB, Amsterdam, the Netherlands) for 30 min on ice. For detection of mAb binding, cells were incubated phycoerythrin (PE)-conjugated rabbit anti-mouse antibody (Becton Dickinson). In some experiments cells were double-labelled with the costimulatory molecule mAb and with CD3–fluoroscein isothiocyanate (FITC) or CD8–FITC (Becton Dickinson). Before fluorescence-activated cell sorting (FACS) analysis 10 µg/ml propidium iodide (Sigma) was added to exclude dead cells from the analysis. Stained cells were analysed by flow-cytometry on a FACScan (Becton Dickinson) using Cellquest software.

Immunohistochemistry

Immunohistochemistry was performed on 5-µm cryostat sections as described by van Rijen et al.²⁸ with the same costimulatory molecule mAb as used for flow cytometry. Binding of antibodies was visualised by incubation in Fast Blue RR salt/Naphtol AS-BI phosphate solution (all from Sigma-Aldrich Chemie, Steinheim, Germany), and tissues were counterstained with Nuclear Fast Red (Fluka Chemie Zwÿndsecht, the Netherlands). Optimal dilutions of the primary mAb were established in titration experiments on cytospin preparations of PBMC stimulated with PMA and ionomycin. A tonsil section was included in each experiment as positive control tissue, and for each tissue a negative control was performed by replacement of the primary mAb by an isotype-matched irrelevant IgG1 mAb (DAKO). Sections were examined for CD154, CD137, CD134 and CD70 staining only when tonsil tissue showed a characteristic staining pattern, and when the tissue incubated with control antibodies was negative. All stainings were done in duplicate.

Statistics

Pooled data from this study and a previous one²⁸ showed that the percentages of CD154 expression on pharmacologically stimulated T cells from 11 different individuals approximated to a normal distribution. Therefore, the percentage expression found in different experiments on immunosuppressive drug-treated lymphocytes and on lymphocytes incubated with solvents of immunosuppressive drugs only were analysed by a parametric Student's t-test for paired data. P ≤ 0·05 was considered to be statistically significant.

Results

Induction of CD154, CD137, CD134 and CD70 on T-cells by PMA and ionomycin

In non-stimulated PBMC very few cells expressed the investigated costimulatory molecules: 1·2 ± 1·7% CD154⁺ cells, 1·4 ± 1·5% CD137⁺ cells, 5·5 ± 3·6% CD134⁺ cells, and 2·5 ± 1·0% CD70⁺ cells (means ± SD; n = 5). After 6-hr stimulations with PMA and ionomycin, CD154 was expressed on 45 ± 5%, CD137 on 54 ± 12%, CD134 on 59 ± 3%, and CD70 on 24 ± 8% of the cells (n = 3). Double-labelling with costimulatory mAb and CD3 mAb revealed that almost all CD154⁺ cells, CD134⁺ cells, CD70⁺ cells, and CD137⁺ cells in stimulated PBMC were T cells (Fig. 1). Double-labelling of stimulated PBMC with costimulatory mAb and CD8 mAb showed that the investigated molecules were induced on both CD8⁻ and on CD8⁺ T cells (data not shown).

Figure 1.

CD70, CD134, CD137, and CD154 induced on PBMC by PMA and inomycin are expressed on T cells. PBMC were stimulated by PMA and ionomycin and after a 60- h culture, cells were isolated and double-labelled with CD154 mAb (a), CD134 mAb (b), CD70 mAb (c), CD137 mAb (d), or irrelevant IgG1 mAb (e), followed by PE-conjugated rabbit anti-mouse IgG and with CD3–FITC mAb.

Effect of immunosuppressive drugs on induction of CD154, CD137, CD134 and CD70 expression by PMA and ionomyin

To study the effect of immunosuppressive drugs on the induction of costimulatory molecules on T cells, PBMC were incubated with clinically observed trough and peak concentrations of these immunosuppressive drugs and stimulated by PMA and ionomycin. The calcineurin inhibitors CsA and TAC inhibited induction of CD154, CD70 and CD137 significantly at both concentrations applied, but not the expression of CD134 (Fig. 2a, b). SDZ RAD did not suppress any of the investigated molecules significantly (Fig. 2c). In contrast to CI and SDZ RAD, MPA inhibited the induction of all tested TNF-receptor and -ligand family molecules nearly completely at a concentration of 4 µg/ml (Fig. 2d).

Figure 2.

Effects of CsA, TAC, SDZ RAD, and MPA on the induction of expression of TNF- and TNF-receptor superfamily costimulatory molecules by PMA and ionomycin. PBMC were incubated with different concentrations of CsA (a), TAC (b), SDZ RAD (c) or MPA (d), and stimulated with PMA and ionomycin. Expression of costimlatory molecules was determined after 60 hr of stimulation by flow cytometry. To visualize the effect of immunosuppressive drugs more clearly, for each experiment the relative expressions in the presence of immunosuppressive drugs were calculated as percentages of the expression found on control cells which were not incubated with immunosuppressive drugs but with solvents of immunosuppressive drugs only. Depicted are means (± SD) of the percentages relative expression from three (CsA and SDZ RAD) or four (TAC and MPA) independent experiments. For statistical analysis the original percentages expression in the presence and in the absence of immunosuppressive drugs were compared by the student's t-test for paired data. Asterisks indicate significant differences (P ≤ 0·05) compared to expression on cells in the absence of immunosuppressive drugs.

Effect of immunosuppressive drugs on induction of CD154, CD137, CD134 and CD70 expression by allogeneic stimulation

To investigate whether the effects of these immunosuppressive drugs are the same in allogeneic stimulation, PBMC were stimulated with irradiated monocyte-derived mature DCs. The expression of CD154, CD137, CD134 and CD70 on T cells was measured by double-labelling with CD3 mAb after 5 days of incubation. In wells without immunosuppressive drugs CD154 expression was induced on 9·0 ± 4·9%, CD134 on 18·2 ± 0·6%, CD70 on 19·3 ± 2·4%, and CD137 on 11·8 ± 2·4% of the CD3⁺ cells (means ± SD; n = 3). Figure 3 shows that CsA and TAC inhibited induction of CD154 and CD137 significantly, although incompletely, but induction of CD134 and CD70 did not. SDZ RAD inhibited allogeneic induction of all tested TNF superfamily costimulatory molecules partially. However, as in the experiments with pharmacologically stimulated PBMC, 4 µg/ml MPA suppressed induction of costimulatory molecules nearly completely.

Figure 3.

Effect of CsA, TAC, SDZ RAD or MPA on allogeneic induction of TNF family costimulatory molecules by dendritic cells. Monocyte-derived mature DC and PBMC were cultivated at a ratio of 1 : 20, in the presence of CsA, TAC, SDZ RAD or MPA at concentrations equivalent to clinically used peak levels. After 5 days, the expression of CD154, CD134, CD70 and CD137 on CD3⁺ cells was measured by flow-cytometry. To visualize the effects of immunosuppressive drugs more clearly, for each experiment the relative expressions in the presence of immunosuppressive drugs were calculated as percentages of the expression found on control cells which were not incubated with immunosuppressive drugs. Data are expressed as mean (± SD) percentages relative expression compared to control cells obtained in three independent experiments. Asterisks indicate significant differences (P ≤ 0·05) compared to expression on cells not incubated with immunosuppressive drugs.

Effect of immunosuppressive drugs-treatment on expression of CD154, CD137, and CD134 in lymphoid tissue in vivo

To study the effect of exposure to immunosuppressive drugs on expression of CD154, CD137, CD134 in lymphoid tissue in vivo, these molecules were immunohistochemically visualized in cryosections of hepatic lymph nodes from three liver transplant patients treated with TAC, and of inguinal lymph nodes from two kidney transplant recipients treated with MMF plus TAC. Five hepatic lymph nodes from multiorgan donors and two lymph nodes from transplant recipients obtained during the transplant procedure (i.e. before initiation of immunosuppressive drug treatment) were used as controls. Figure 4 shows that lymph nodes from patients treated with MMF plus TAC and control patients showed comparable expressions of CD154, CD137, and CD134. CD154 and CD134 were expressed on single cells in the lymphoid tissue. CD137 showed a characteristic staining pattern on germinal center cells and in intermediate sinuses, as described by Pauly et al.²⁹ which was the same in immunosuppressive drugs-treated and non-immunosuppressive drugs-treated patients. Similar expression patterns were also observed in lymph nodes from patients treated with TAC only. The staining of CD70 was too faint to enable reliable interpretation. Because CD154 and CD134 were expressed on distinct cells, positive cells could be quantified. Table 1 shows that similar numbers of CD154⁺ and CD134⁺ cells were present in lymph nodes from TAC-treated, TAC plus MMF-treated, and control patients.

Figure 4.

Expression of CD154, CD134 and CD137 in lymph nodes from a patient not treated with immunosuppressive drugs (control; a, c, e) and from a patient treated with MMF + TAC (b, d, f). Distinct CD154⁺ cells (a and b) and CD134⁺ cells (c and d) were observed in the lymphoid tissues of both patients. CD137-staining (e and f) was found in intermediate sinuses (arrows) and germinal centres of the follicles. a–d: 400× magnification. e, f: 100× magnification.

Table 1.

Effect of immunosuppressive treatments on numbers of CD154⁺ and CD134⁺ cells in lymph nodes

	Number of positive cells ± SEM per microscopic field
Treatment (number of patients)	CD 154	CD 134
MMF + TAC (n = 2)	12; 2	18; 20
TAC (n = 3)	9 ± 6	19 ± 14
no (n = 7)	10 ± 4	27 ± 14

Numbers of CD154⁺ and CD134⁺ cells were counted in 10 microscopic fields of 400× magnification distributed over two cryostat sections of lymph nodes from patients treated with MMF plus TAC, or TAC only, and from patients without immunosuppressive treatment. Data are depicted as means of the densities of positive cells found in the individual lymph nodes ± SD, except for MMF ± TAC treatment where mean densities counted in sections of the two individual lymph nodes are given separately.

Discussion

In this study the effects of three classes of immunosuppressive drugs on the induction of TNF- and TNF-receptor family molecules on T cells were investigated. We found that MPA was the most potent inhibitor, which at a clinically applied concentration of 4 µg/ml completely inhibited the induction of CD154, CD137, CD134 and CD70 nearly completely, both upon pharmacological and allogeneic stimulation of T cells. Few other data are available on the effect of MPA on expression of these molecules. Gummert et al. have shown that MPA inhibits in vitro the induction of CD134 on rat lymphocytes^30,31 but Smiley et al.³² found no effect of MMF on the induction of CD154 on murine T-cells in vitro. The difference between our observations and those of Smiley et al. may be explained by the their use of the inactive ester MMF instead of MPA, which was probably not hydrolysed during the short culture period used in their experiments.

Two molecular mechanisms may be responsible for the inhibition of the investigated molecules by MPA. MPA, which is a specific inhibitor of inosine monophosphate dehydrogenase (IMPDH), which, in turn is a crucial enzyme in the de novo synthesis of purines, provides depletion of guanine nucleotides in lymphocytes.^30,31,33 This may limit the synthesis of mRNA for these molecules. Second, because guanine nucleotides are necessary for glycosylation of lymphocyte proteins^30,31,33 and because the investigated molecules are glycoproteins, MPA may alter their glycosylation pattern, resulting in a decrease of expression.

In contrast to MPA, SDZ RAD did not significantly suppress the induction of the costimulatory molecules on T cells by PMA and ionomycin. This is in agreement with the absence of inhibition of CD154 expression by rapamycin observed by Smiley et al.³² Rapamycin and it's derivative SDZ RAD are inhibitors of TOR, which is a crucial molecule in the intracellular transduction of stimulatory signals from cytokine receptors and of costimulatory signals from CD28 in T cells.³⁴ Because PMA and ionomycin do not trigger these signal transduction pathways, but instead protein kinase C and intracellular calcium, respectively, it is not surprising that SDZ RAD was not effective in inhibiting the induction of costimulatory molecules in this system. However, SDZ RAD did partially inhibit induction of all tested molecules upon allogeneic stimulation of T cells. In this system, costimulation of T cells via CD28 by B7 molecules on the allogeneic DCs may have contributed to the up-regulation of these molecules. Furthermore, it can not be excluded that the inhibitory effect of SDZ RAD on the induction of costimulatory molecules on T cells in the allogeneic stimulation system was caused by apoptosis induction in the DCs by SDZ RAD.³⁵

CsA and TAC inhibited expression of CD154, CD137 and CD70, but not of CD134. Both are CI, and inhibitors of T-cell activation via the T-cell receptor. It was therefore surprising to find that CD134 expression was not inhibited by these immunosuppressive drugs, because CD134 expression on T cells is induced by T-cell receptor stimulation.³⁶ It also contrasts with the partial inhibition of CD134 induction on rat lymphocytes by CsA reported by Roos et al.³⁷ and Barten et al.³⁸ This difference may be a result of species differences.

To investigate the effect of exposure to immunosuppressive drugs in vivo on expression of the investigated molecules, immunohistochemical quantifications of CD134⁺ and CD154⁺ cells were performed on cryostat sections of lymph nodes of patients treated with TAC and MMF and of individuals not treated with any immunosuppressive drugs. Although from only two patients treated with MMF frozen lymph node tissue was available, the quantitative data on numbers of CD134⁺ and CD154⁺ cells (Table 1) and the qualitative data on expression on CD137 (Fig. 4) suggest that no decrease in the numbers of cells expressing these molecules occurred after several months of exposure to MPA in vivo. This may be explained by the trough levels of MPA in the patients, which were lower as the concentration at which complete inhibition was observed in vitro. The mean trough level in the patients in the three months before the lymph nodes were removed from the body, was 2·3 µg/ml. At this concentration incomplete but significant inhibition of CD134 was found in vitro, while suppressions of CD154 and CD137 were not significant. A second possible cause of the full expression of the investigated molecules in lymph nodes during treatment with MMF may be that desensitization for MPA occurs during chronic treatment, e.g. by induction of IMPDH enzyme activity. In rats, a decrease in the inhibition of induction of several lymphocyte cell surface antigens has been observed after multiple MPA doses.³⁰

Likewise, numbers of CD154⁺ cells and the qualitative expression of CD137 were not lowered in lymph nodes from liver transplant patients treated for several months with TAC, although induction of these molecules by PMA and ionomycin on T-cells was strongly inhibited by TAC in vitro at concentrations similar to the trough levels in the patients. This may be partly explained by the presence of other signals which induce these molecules in vivo, and are insensitive to CI, like the costimulatory signal via CD28. This explanation is supported by the incomplete inhibition of CD137 and CD154 upon allogeneic stimulation with DCs, which express B-7 molecules (Fig. 4). In addition, the absence of inhibition in vivo may be the result of reduced availability of immunosuppressive drugs for lymphocytes in vivo, because of competition by irrelevant binding sites for the lipophilic CI.³⁹ However, we do not prefer this explanation, as in preliminary experiments in which lymphocytes were stimulated by PMA and ionomycin in whole blood, CsA and TAC inhibited CD154 completely at the same concentrations as in the absence of erythrocytes (unpublished results). Finally, the CD154⁺ cells found in lymph nodes may represent cells that have been induced to express these molecules earlier in their lifespan, before immunosuppressive drugs therapy. Skov et al.⁴⁰ have shown that T lymphocytes that have been induced once to express CD154 are able to re-express it during a second stimulation by IL-2 only in a CsA-resistant manner. A first induction of CD154, CD134 and CD137 expression in the transplant patients may have occurred before transplantation when immunosuppressive drug levels are not yet adequate.

In conclusion, MPA is the most potent inhibitor of the induction of CD154, CD137, CD134 and CD70 on T cells of the three classes of immunosuppressive drugs investigated in this study. Its immunosuppressive effect in clinical transplantation may therefore not only be related to inhibition of lymphocyte proliferation, but also to suppression of inducible TNF- and TNF-receptor superfamily molecules on T cells. To obtain long-term suppression of these molecules in vivo, a plasma trough level of 4 µg/ml is indicated.

Abbreviations

CsA

cyclosporin A

TAC

tacrolimus

SDZ RAD

rapamycin derivative

MPA

mycophenolic acid

TCR

T-cell receptor

APC

antigen-presenting cell

MMF

mycophenolate mofetil

CI

calcineurin inhibitors

PMA

phorbol 12- myristate 13-acetate

PBMC

peripheral blood mononuclear cells

DC

dendritic cells

IMPDH

inosine monophosphate dehydrogenase