The immunosuppressive drug mycophenolate mofetil impairs the adhesion capacity of gastrointestinal tumour cells

Abstract

Immunosuppression correlates with the development and recurrence of cancer. Mycophenolate mofetil (MMF) has been shown to reduce adhesion molecule expression and leucocyte recruitment into the donor organ. We have hypothesized that MMF might also prevent receptor-dependent tumour dissemination. Therefore, we have investigated the effects of MMF on tumour cell adhesion to human umbilical vein endothelial cells (HUVEC) and compared them with the effects on T cell–endothelial cell interactions. Influence of MMF on cellular adhesion to HUVEC was analysed using isolated CD4⁺ and CD8⁺ T cells, or WiDr colon adenocarcinoma cells as the model tumour. HUVEC receptors ICAM-1, VCAM-1, E-selectin and P-selectin were detected by flow cytometry, Western blot or Northern blot analysis. Binding activity of T cells or WiDr cells in the presence of MMF were measured using immobilized receptor globulin chimeras. MMF potently blocked both T cell and WiDr cell binding to endothelium by 80%. Surface expression of the endothelial cell receptors was reduced by MMF in a dose-dependent manner. E-selectin mRNA was concurrently reduced with a maximum effect at 1 µm. Interestingly, MMF acted differently on T cells and WiDr cells. Maximum efficacy of MMF was reached at 10 and 1 µm, respectively. Furthermore, MMF specifically suppressed T cell attachment to ICAM-1, VCAM-1 and P-selectin. In contrast, MMF prevented WiDr cell attachment to E-selectin. In conclusion, our data reveal distinct effects of MMF on both T cell adhesion and tumour cell adhesion to endothelial cells. This suggests that MMF not only interferes with the invasion of alloactivated T cells, but might also be of value in managing post-transplantation malignancy.

INTRODUCTION

Mycophenolate mofetil (MMF, CellCept® Roche Bioscience, Grenzach-Wyhlen, Germany) is a recently developed immunosuppressive drug. MMF effects are based on the inhibition of inosine monophosphate dehydrogenase (IMPDH) and prevention of guanosine monophosphate synthesis from inosine monophosphate, a rate-limiting step in the purine biosynthesis in lymphocytes. Consequently, MMF blocks the proliferation and clonal expansion of T and B lymphocytes and prevents the generation of cytotoxic T cells, as well as other effector T cells [1].

Other mechanisms may also contribute to the efficacy of MMF in preventing allograft rejection. By depleting guanosine nucleotides, MMF suppresses glycosylation and the expression of some adhesion molecules, thereby decreasing recruitment of lymphocytes and monocytes into sites of inflammation and graft rejection [1]. Immunoprecipitation studies have shown that one of the glycoproteins affected is the lymphocytic VLA-4 receptor, the ligand for VCAM-1 on activated endothelial cells. Further experiments have revealed inhibition of LFA-1, the counter-receptor of ICAM-1, after MMF administration [2,3]. Other reports demonstrate down-regulation of ICAM-1, VCAM-1, E-selectin and P-selectin [4,5].

One of the most undesirable complications of an effective immunosuppressive therapy is the recurrence of neoplastic tumours and the development of de novo cancer [6]. Five of seven patients with bronchioloalveolar carcinoma who received a lung transplant had recurrent bronchioloalveolar carcinoma within the donor lungs [7]. High rates of recurrent tumours have also been observed in patients who undergo liver transplantation for unresectable cholangiocarcinoma or cholangiohepatoma [8]. Other reports have revealed that immunosuppression is a high risk factor for developing de novo malignancies after liver or renal transplantation [9,10]. Cancer has therefore become a major cause of death in patients otherwise treated successfully by organ transplantation. Based on the knowledge that MMF blocks adhesion molecule expression, we postulate that MMF might not only suppress leucocyte recruitment to the donor graft, but may also prevent adhesion receptor-dependent tumour dissemination. To explore how MMF may serve as a metastasis-blocking agent, we investigated its effects on the adhesion of WiDr colon adenocarcinoma cells to human endothelial cells in vitro. The influence of MMF on CD4⁺ and CD8⁺ T cell adhesion was also measured, to allow clear interpretation of the data. Comparative analysis demonstrates that although MMF suppresses T cell and WiDr cell binding to endothelium, it acts differentially on both cell types: (1) WiDr cells are more sensitive to MMF than T cells and (2) MMF selectively blocks T cell attachment to ICAM-1, VCAM-1 and P-selectin, whereas MMF prevents WiDr cell attachment to E-selectin and P-selectin. The results indicate that MMF may have some effect against tumour metastasis and this may, to some extent, counterbalance concerns arising from its immunosuppressive properties.

MATERIALS AND METHODS

Cell cultures

Human umbilical vein endothelial cells (HUVEC) were extracted from human umbilical cords using chymotrypsin (α-chymotrase; Hasenclever, Bonn, Germany) to loosen the cells from the umbilical veins. After harvesting, cells were plated in 75 cm² tissue culture flasks and cultivated in M199 medium, supplemented with 10% fetal calf serum, 10% human serum, 20 µg/ml endothelial cell growth factor, 0·1% heparin, 100 µg/ml gentamycin and 20 mm HEPES-buffer (pH 7·4). Subcultures from passages 2–4 were selected for the experiments and were used either unstimulated or stimulated with 100 U/ml IL-1 (4-h prestimulus). Peripheral blood mononuclear cells were isolated by Ficoll-Hypaque centrifugation and resuspended in HUVEC medium. Peripheral blood lymphocytes (PBL) were prepared from peripheral blood mononuclear cells by adhering monocytes/macrophages in 75 cm² tissue culture flasks overnight at 37°C and 5% CO₂. CD4⁺ and CD8⁺ T lymphocytes were selected positively using high-gradient magnetic columns with anti-CD4 or anti-CD8 monoclonal antibodies (MACS CD4/CD8-microbeads; Miltenyi Biotec GmbH, Bergisch Gladbach, Germany). The human colon adenocarcinoma cell line WiDr was obtained from the tumour cell bank of the Johannes Gutenberg-University, Mainz, Germany, and grown in RPMI-1640 medium supplemented with 10% fetal calf serum, 100 ng/ml gentamycin and 20 mm HEPES-buffer (pH 7·4).

Adhesion assay

Round cover slips were treated with 3-aminopropyl-triethoxy-silan (2%; Sigma, München, Germany) − acetone solution for 60 min (20°C) to allow firm adhesion of HUVEC and placed into six-well multi-plates (Falcon Primaria; Becton Dickinson, Heidelberg, Germany). HUVEC subcultures were transferred to prepared multi-plates. When confluency was reached, 0·5 × 10⁶ CD4⁺ T, CD8⁺ T or WiDr cells/well were added carefully to the HUVEC monolayer for various periods of time. Subsequently, non-adherent cells were washed off using warm (37°C) M199 medium. The remaining cells were fixed with 1% glutaraldehyde (Merck, Darmstadt, Germany), counted in five different fields (5 × 0·25 mm²) using a phase contrast microscope (20 × objective), and mean cellular adhesion was calculated.

Cell proliferation

Proliferative activity of tumour cells and HUVEC were estimated by the PicoGreen assay: several time-points after plating the cells in six-well multi-plates, culture medium was removed and cells were digested with papain (0·125 mg protein/ml) for 20 h at 60°C [11]. Thereafter, the fluorescent dye PicoGreen (MoBiTec, Goettingen, Germany), which shows high specificity for dsDNA, was added (1 : 200 dilution) for 10 min at 20°C. Fluorescence intensity was determined using a computer-controlled fluorescence reader (Cytofluor 2300 plate scanner; Millipore, Eschborn, Germany) at λ_ex = 485 nm and λ_em = 530 nm.

MMF

To investigate the effects of MMF on cellular adhesion dynamics, the following experiments were performed: (a) MMF was added to CD4⁺ T, CD8⁺ T or WiDr cells which were then pipetted onto HUVEC already treated with MMF. (b) MMF was added to CD4⁺ T, CD8⁺ T or WiDr cells which were then pipetted onto HUVEC which had not been treated with MMF. (c) CD4⁺ T, CD8⁺ T or WiDr cells, not treated with MMF were pipetted onto HUVEC which had been treated with MMF. Cellular adhesion was measured in all experiments and the MMF concentration ranged from 0 (control) to 10 µm. To confirm that MMF effects observed were caused by the active metabolite mycophenolic acid (MPA), adhesion experiments were repeated using MPA in various concentrations and the results compared. Furthermore, concentrations of MPA in the cell culture supernatant were determined after exposure of HUVEC or T cells with MMF for 3 h by reverse phase high-performance liquid chromatography (HPLC).

Cell binding to immobilized receptor protein chimeras

To evaluate the binding capacity of CD4⁺ T, CD8⁺ T and WiDr cells to isolated endothelial adhesion proteins in the presence of MMF, chimeric receptor globulins were constructed as described previously [12]. Proteins containing the extracellular domain of E-selectin, ICAM-1, VCAM-1 or P-selectin (fused to the hinge region of human IgG1) were expressed in COS-7 cells. COS-7 cells were transfected with 3 µg plasmid-DNA using the DEAE/dextran method. Seven days after transfection, supernatants were collected and stored at −20°C. The concentration of the receptor globulin chimeras was determined by the enzyme-linked immunosorbent assay (ELISA) using a monoclonal rat antihuman IgG antibody conjugated to peroxidase [12].

Round culture dishes (Falcon Primaria, Becton Dickinson) were incubated with a spot of 50 µl goat-antihuman IgG (Sigma) at a concentration of 10 mg/ml in 50 mm Tris, pH 9·5, for 90 min (20°C). Dishes were washed three times with PBS (Seromed, Berlin, Germany) and blocked with 1% bovine serum albumin (Merck) overnight at 4°C. The dishes were incubated subsequently with 1 ml of cell culture supernatant, containing 5 µg/ml E-selectin, ICAM-1, VCAM-1 or P-selectin IgG fusion protein for 30 min at 20°C. Dishes were then washed three times and T cells or tumour cells were resuspended at a density of 3 × 10⁶ cells/ml in binding buffer containing 50 mm HEPES (pH 7·4), 100 mm NaCl, 1 mg/ml bovine serum albumin, 2 mm MgCl₂, 1 mm CaCl₂, 3 mm MnCl₂, 0·02% NaN₃ and 0·2 mm PMSF (Sigma) and transferred to the culture dishes for 30 min. The buffer composition was shown to be non-toxic to tumour cells and T cells, as assessed by the trypan blue dye exclusion test, propidium iodide dsDNA-intercalation and quantitative fluorescence analysis of enzyme-catalysed fluorescein-diacetate metabolism. After 30 min incubation, non-adherent cells were washed off and the remaining cells were counted using a phase contrast microscope (× 20 objective). Five observation fields were chosen at random in each dish and the mean value of the number of adherent cells per field was calculated.

Analysis of endothelial adhesion receptors − FACS

MMF was added to HUVEC, which were stimulated with 100 U/ml IL-1 or prostaglandin E₂ (PGE₂; 10⁻⁶m for 5–10 min to induce surface P-selectin expression). HUVEC were washed in blocking solution (PBS, 0·5% BSA) and then incubated for 60 min at 4°C with FITC-conjugated (goat-antimouse IgG) anti-ICAM-1 (clone BBIg-I₁), anti VCAM-1 (clone BBIg-V₁), anti-E-selectin (clone BBIg-E₄; all from Biermann, Bad Nauheim, Germany) or anti-P-selectin (clone CLB-Thromb/6; Immunotech, Hamburg, Germany) monoclonal antibodies. Surface expression of the respective molecules was then detected using a FACScan (Becton Dickinson, Heidelberg, Germany) and expressed as relative fluorescence units (RFU).

Western blot

Total P-selectin content in HUVEC were evaluated by Western blot analysis: HUVEC lysates were applied to a 7% polyacrylamide gel under denaturing conditions and electrophoresed for 90 min at 60 V. The gel was then transferred to nitrocellulose membranes. After blocking, they were incubated overnight with sheep-antihuman P-selectin-antibodies (sheep IgG antibodies; 1 : 500 dilution; R&D Systems). Horseradish peroxidase (HRP)-conjugated rabbit-antisheep IgG (1 : 5000 dilution; Upstate Biotechnology, Lake Placid, NY, USA) served as the secondary antibody. The filter was exposed to an X-ray-film (Hyperfilm^TM EC^TM; Amersham, Braunschweig, Germany) and proteins visualized by enhanced chemiluminescence (Amersham). Human platelets, stimulated with ADP disodium dihydrate (20 µm, 10 min) were used as positive controls.

Northern blot

E-selectin was evaluated further by Northern blot. Total RNA was prepared from HUVEC (RNeasy Kit; Qiagen, Hilden, Germany) and 20 µg of total RNA were subjected to electrophoresis in 1% agarose gel containing 2% formaldehyde. The separated RNA was transferred to nylon membranes (Polytrap 296 PE; Schleicher & Schuell, Dassel, Germany) and hybridized overnight at 42°C. For hybridization, the plasmid pCDM8, including a human E-selectin cDNA clone, was cut using EcoRI restriction enzyme. This 790 bp fragment, which represents the 5′ region of the E-selectin cDNA, was ³²P-labelled by using a random priming kit (Rediprime^TM II; Amersham Pharmacia Biotech, Braunschweig, Germany). The membrane was washed at 42°C (2 × 15 min with 2 × SSC, 0·1% SDS and at least 30 min with 0·2 × SSC, 0·1% SDS) and the radioactivity visualized by the automatic detector system BAS 1500 (Raytest; Fujifilm, Straubenhardt, Germany). Densitometric analysis was performed on an ImageQuant densitometric scanner. E-selectin mRNA levels were normalized to that of GAPDH by dividing the E-selectin densitometric units by the corresponding GAPDH units. The ratio in the IL-1-activated experimental sample was defined as 100%.

Statistical analysis

Fifity per cent effect inhibition (ID₅₀) was calculated using the modified HILL equation:

where N describes a slope factor indicating the steepness of the concentration/response curve. Statistical significance was investigated by the Wilcoxon signed rank test showing two-sided probabilities and using normal approximation. Differences were considered statistically significant at a P-value less than 0·05.

RESULTS

Time-course experiments

To define the ascending phase of cell adhesion, detailed kinetic studies have been carried out (Fig. 1a–c, insets). The number of adherent CD4⁺ and CD8⁺ T cells increased after 1 h (CD4⁺: 58·4 cells/mm²; CD8⁺: 63·2 cells/mm²; interassay variation <20%) and reached a plateau after 4 h (CD4⁺: 352·0 cells/mm²; CD8⁺: 365·6 cells/mm²; interassay variation <25%). No difference could be found between the adhesion capacity of isolated CD4⁺ and CD8⁺ T cells. WiDr cells attached to HUVEC more rapidly and in greater number than T cells. After 30 min, 446·8 cells/mm² were counted (interassay variation <15%). The plateau phase was already attained after 2 h (840·4 cells/mm²; interassay variation <25%). The PicoGreen assay revealed no proliferative activity of tumour cells or HUVEC during this time (P < 0·05). The subsequent blocking studies using MMF were carried out after 60 min incubation, which is the ascending phase of both WiDr cell and T cell adhesion.

Fig. 1.

Blocking of CD4⁺ T (a), CD8⁺ T (b) or WiDr (c) cell adhesion by MMF (0·001–10 µm). HUVEC were activated by IL-1 (100 U/ml), and T cells/WiDr cells were added for 60 min. Controls (no MMF exposure) were set at 100%. Either T cells/WiDr cells [squares; depicted as ‘CD4 incubated’ (a), ‘CD8 incubated’ (b), or ‘WiDr cells incubated’ (c)] or HUVEC (triangles, depicted as ‘HUVEC incubated’) were incubated with MMF or both T cells/WiDr cells and HUVEC were incubated with MMF [circles; depicted as ‘CD4⁺ HUVEC incubated’ (a), ‘CD8 + HUVEC incubated’ (b), or ‘WiDr⁺ HUVEC incubated’ (c)]. Each point represents the mean of four experiments. SD was usually about 20%. The figure insets show time-dependent adhesion of CD4⁺ T (a), CD8⁺ T (b), or WiDr cells (c) to HUVEC, without MMF application. The adhesion values are given in cells/mm² (y-axis). x-axis shows time-course, ranging from 0·5 to 8 h (a,b) or from 0·5 to 4 h (c).

MMF down-regulates cell adhesion with different efficacy

MMF reduced T cell adhesion to IL-1 activated HUVEC in a dose-dependent manner, with similar effects on CD4⁺ and CD8⁺ T cells (Fig. 1a,b). Marked suppression was observed at 10 µm MMF. Incubation of both T cells and HUVEC with MMF led to 80% reduction. Separate incubation of T cells or HUVEC with this compound also evoked significant suppression, demonstrating that MMF acts on both cell types. HPLC analysis revealed that MMF was hydrolysed into MPA, not only in T cells but also in HUVEC (Table 1). Similar to MMF, MPA reduced T cell adhesion to IL-1 activated HUVEC in a dose-dependent manner.

Table 1.

Metabolization of MMF

MMF added to cell cultures	MPA detected in the supernatant*
HUVEC + 100 µg/ml MMF	23·13 ± 0·88 µg/ml
HUVEC + 10 µg/ml MMF	2·74 ± 0·81 µg/ml
HUVEC + 1 µg/ml MMF	0·20 ± 0·02 µg/ml
CD4+ T cells + 10 µg/ml MMF	1·53 ± 0·17 µg/ml
CD8+ T cells ± 10 µg/ml MMF	1·49 ± 0·32 µg/ml

^*Mean values ± s.d. from four experiments.

MMF also down-regulated WiDr tumour cell binding to activated HUVEC. However, when both tumour cells and HUVEC were treated with MMF, the maximum effect was reached at a MMF concentration of 0·1–1 µm. Separate incubation of WiDr cells or HUVEC also evoked reduced tumour cell binding with a maximum efficacy at 0·1 µm or 1 µm, respectively (Fig. 1c).

MMF suppresses endothelial adhesion receptors

MMF significantly reduced surface expression of ICAM-1, VCAM-1, E-selectin and P-selectin (Fig. 2) when given in concentrations ranging from 0·01 to 10 µm (compared to untreated controls). E-selectin mRNA was analysed by Northern blot. A faint 3·8-kb band of E-selectin mRNA was visualized before the stimulation of HUVEC, which was increased 4 h after stimulation of HUVEC with IL-1. After exposing the cells to MMF, E-selectin mRNA was reduced in a dose-dependent manner, with a maximum effect at 1 µm(Fig. 3). The amount of control GAPDH mRNA was similar between MMF treated and non-treated HUVEC. Total P-selectin content was assessed additionally by Western blot. Analysis of PGE₂ activated HUVEC and platelets, which served as controls, revealed two distinct bands at 40 and 50 kDa. In strong contrast to surface expression levels, the intensity of the 50 kDa protein band did not change when HUVEC were incubated with MMF at various concentrations. The intensity of the 40 kDa protein band was reduced in the presence of 10 µm MMF (Fig. 4).

Fig. 2.

Fluorescence analysis (relative fluorescence units, RFU) of endothelial cell ICAM-1, VCAM-1, E-selectin and P-selectin surface expression. For maximal stimulation, HUVEC were incubated with IL-1 (100 U/ml) for 12 h to up-regulate ICAM-1, and to induce VCAM-1 and E-selectin. HUVEC were incubated with PGE₂ (10⁻⁶m) for 10 min to induce P-selectin expression. FL-1H (log) channel histogram analysis; 1 × 10⁴ cells/scan. Control values were set at 100%. Mean ± s.d. of three experiments.

Fig. 3.

MMF suppresses IL-1 induced E-selectin mRNA expression. Total RNA was isolated and E-selectin mRNA levels were determined by Northern blot analysis. GAPDH was used to assess equivalent RNA loading C⁻= unstimulated control cells, C⁺= IL-1 stimulated control cells. Results of densitometry are shown in the graph below and are expressed as the ratio of E-selectin : GAPDH mRNA, relative to the control (IL-1 activation without MMF), which was assigned a value of 100%. The figure shows one representative experiment from three.

Fig. 4.

Western blot analysis of P-selectin in endothelial cells. HUVEC were stimulated with 10⁻⁶m PGE₂ and incubated additionally with different concentrations of MMF. Throm = control analysis of platelet P-selectin protein, C^-= unstimulated HUVEC without MMF, C⁺ = PGE₂-stimulated HUVEC without MMF. The figure shows one of four separate experiments.

MMF differentially blocks binding of WiDr cells and T cells to immobilized adhesion proteins

Pilot studies demonstrated that CD4⁺ and CD8⁺ T cells bound well to immobilized E-selectin (mean: 280 cells/mm²; interassay variation <30%), P-selectin (mean: 720 cells/mm²; interassay variation <20%), ICAM-1 (170 cells/mm²; interassay variation <30%) and VCAM-1 (mean: 640 cells/mm²; interassay variation <30%). Binding to uncoated plates or to IgG was <20 cells/mm². In good accordance, T cells expressed the adhesion ligands sLeX (RFU_mean: CD4⁺: 10·04; CD8⁺: 28·28), PSGL-1 (RFU_mean: CD4⁺: 173·60; CD8⁺: 161·51), LFA-1 (RFU_mean: CD4⁺: 41·10; CD8⁺: 53·28) and VLA-4 (RFU_mean: CD4⁺: 94·08; CD8⁺: 101·86). Antibodies to each of the immobilized adhesion proteins totally blocked T cell attachment to the respective counter-receptors (>95% reduction compared to background binding). In contrast, WiDr cells did not express LFA-1 and VLA-4 and did not bind to immobilized ICAM-1 and VCAM-1. However, they strongly expressed sLeX on their membrane (113·4 ± 17·6 RFU) and attached to immobilized E-selectin and P-selectin IgG protein chimeras (>1000 cells/mm²). Binding to uncoated plates or to IgG was <30 cells/mm². Monoclonal antibodies directed against E- or P-selectin abolished tumour cell attachment (>95% reduction compared to background binding).

MMF differentially influenced the interaction of the specific adhesion receptors with their respective ligands. MMF diminished adhesion of CD4⁺ and CD8⁺ T cells to immobilized ICAM-1, VCAM-1 and P-selectin, but not to immobilized E-selectin (Figs 5a.b). On the other hand, MMF blocked binding of WiDr cells to immobilized P-selectin as well as to immobilized E-selectin (Fig. 5c). Maximum efficacy was reached in the presence of 10 µm MMF (blocking of T cell attachment) versus 0·1 µm MMF (blocking of WiDr cell attachment).

Fig. 5.

Adhesion of CD4⁺ T (a), CD8⁺ T (b) or WiDr (c) cells to immobilized adhesion receptor globulin chimeras. Purified T cells or WiDr cells were added to culture dishes coated with the extracellular domain of ICAM-1, VCAM-1, E-selectin or P-selectin. The immobilized adhesion proteins were coupled to goat-antihuman IgG. Cells were pretreated with different concentrations of MMF. Untreated cells served as controls (100% value). Adhesion was measured after 30 min incubation and after removal of non-adherent cells, according to the protocol given in Materials and methods. Each point represents the mean ± s.d. of four experiments; s.d. was usually about 20%.

DISCUSSION

The phenomenon that MMF reduces endothelial adhesion receptor expression is not only of great relevance to the interaction of alloactivated T cells with the donor vessel wall but also opens the possibility that MMF might also interfere with the process of tumour cell attachment to endothelial cells. Indeed, our adhesion experiments showed that MMF influences both binding of T cells as well as binding of WiDr cells to endothelium. Surprisingly, tumour cells responded differently to MMF than did T cells; 10 µm MMF was necessary to evoke maximum effects on CD4⁺ and CD8⁺ T cells, whereas 1 µm MMF was sufficient to evoke maximum effects on WiDr cells. These differences might be caused by different metabolic activity of tumour versus T cells. Franchetti and Grifantini have shown that IMPDH type II increases significantly in cancer cells [13]. Possibly, less MMF might be necessary to interfere with IMPDH turnover in rapidly proliferating cells.

Aside from the above-mentioned possibility, we should be aware that T cell binding depends on all endothelial receptors investigated, whereas WiDr cells contact endothelium via E- and P-selectin exclusively. Down-regulation of each of these receptors might therefore evoke different responses in T cells versus WiDr cells. Loss of selectins might be accompanied by a total blockade of WiDr cell binding, whereas T cells might still be able to attach to endothelial cells via ICAM-1 or VCAM-1.

Interestingly, MMF effects on E-selectin surface expression correlated positively with its effects on E-selectin mRNA when applied at 0·01–1 µm, but not when applied at 10 µm. At 10 µm, the E-selectin surface level was reduced drastically compared to the E-selectin mRNA level, which reached nearly control values. There is no clear explanation for this phenomenon. Hypothetically, high MMF doses may down-regulate transcription of mRNA but prolong the half-life of the remaining mRNA, as suspected by Hauser et al. [14]. More unspecifically, ‘overloading’ the cells with MMF might lead to wrong occupancy or displacement of MPA from its binding sites and, hence, resulting in a decreased effect on protein biosynthesis. Membrane-bound E-selectin may then be shed independently of mRNA modulation.

A similar mode of action might be the reason why MMF maximally blocks WiDr cell adhesion when given in concentrations ranging from 0·1 to 1 µm, but is less potent when applied at 10 µm. Interestingly, the same characteristic has been observed when analysing the binding behaviour of WiDr cells to immobilized selectin proteins. Based on this, we speculate that high-dose MMF might diminish its antitumour properties, probably by reducing its influence on sLeX or PSGL-1, the counterparts of P- and E-selectin receptors. Considering the transplant situation, accumulation of MMF beyond a critical concentration might possibly attenuate its effects on de novo malignancy or cancer recurrence.

Our data demonstrate that MMF reduces the binding capacity of lymphocytic LFA-1, VLA-4 and PSGL-1 to the counter-receptors ICAM-1, VCAM-1 and P-selectin, but not sLeX interaction with E-selectin. In strong contrast, MMF significantly prevented not only sLeX triggered WiDr cell binding to P-selectin but also to E-selectin. Data suggest the critical role of sialyl Lewis X carbohydrate antigens and E-selectin in tumour metastasis: colonization of carbohydrate-bearing tumour cells was abolished completely in E-selectin knock-out mice [15]. Blocking of E-selectin by monoclonal antibodies abolished binding of Colo-205 colon tumour cells to activated endothelium [16]. In addition, preventing sLeX–E-selectin interaction by peptides mimicking carbohydrate ligands inhibits the adhesion of HL-60 and B16 melanoma cells to E-selectin in vitro and blocks lung colonization of B16 cells in vivo [17]. A similar experimental in vivo model has demonstrated the inability of HT-29 adenocarcinoma cells to metastasize to the liver after stable transfection of antisense sequences directed at the human Lewis alpha (1,3/1,4) fucosyltransferase gene [18]. A recently published study including 64 colorectal cancer patients demonstrated that cimetidine blocks the expression of E-selectin on vascular endothelium and inhibits the adhesion of cancer cells to the endothelium, especially when expressing high levels of sialyl Lewis antigens X and A [19].

Currently, it is thought that the interaction of sLeX-expressing tumour cells with E-selectin is a necessary prerequisite to induce tyrosine phosphorylation in a number of proteins. It is also a prerequisite for the expression of cellular properties correlated to the metastatic phenotype, i.e. cytoskeletal reorganization, collagenase secretion and motile behaviour [20]. Under these circumstances, MMF might be useful for the management of certain post-transplantation malignancies. The reason why MMF acts exclusively on sLeX-driven tumour cell but not T cell attachment to E-selectin is not clear. Possibly, both cell types differentially regulate sLeX synthesis and expression. MMF might then interfere specifically with the glycosylation pathway used by tumour cells. It has been observed recently that the nucleotide sugar transporter for UDP-galactose is increased significantly in colon cancer tissues compared with non-malignant mucosal tissues. This transporter is involved in the enhanced expression of sLeX carbohydrates [21]. Nakamori et al. [22] have presented evidence that glycosyltransferase genes, which are normally involved in the synthetic pathways of sialyl Lewis antigens, are not responsible for the molecular mechanisms of increased expression of sLeX antigens in ductal carcinomas of the pancreas. In good accordance, the alpha fucosyltransferase FucT-VII, which is one of the key glycosyltransferases involved in the biosynthesis of the sLeX antigen on human leucocytes [23,24], is expressed at a very low level only in non-cancerous colorectal tissues and not up-regulated in cancerous colorectal tissues [25]. Investigation of the gene expression of fucosyl-transferases in breast cancer revealed no FucT-VII expression. However, a correlation was found between the concentration of sLeX and the amount of FucT-VI mRNA [26].

In summary, MMF acts on both endothelial cells and tumour cells, respectively. Experiments in monoculture systems demonstrate distinct suppressive activities on receptor expression and binding activity of adhesion ligands. However, considering the coculture adhesion assays, MMF's effects on cell adhesion might not exclusively be attributed to the cell type initially incubated with this compound. Rather, MPA might be released from the MMF treated cells into the supernatant and subsequently influence the ‘non-treated’ population to some degree.

When finally interpreting the effects of MMF on tumour cells, it should be taken into consideration that endothelial adhesion molecules are also important in host defense mechanisms [27]. From a clinical viewpoint, further studies must be undertaken which evaluate the tumour recurrence rate in MMF versus non-MMF treated transplant patients.