EFFECTS OF LEFLUNOMIDE AND OTHER IMMUNOSUPPRESSIVE AGENTS ON T CELL PROLIFERATION IN VITRO¹

Abstract

Leflunomide and its active metabolite, A771726, are structurally unrelated to immunosuppressive agents currently under investigation. Previous in vitro studies have revealed that leflunomide primarily inhibits interleukin-2-stimulated T cell proliferation. In the current study, we have extended our previous work and demonstrate that leflunomide prevents T cell progression induced by phytohemagglutinin into the S phase of the cell cycle. To discriminate further the action on T cells of leflunomide from other immunosuppressive agents, we performed kinetic studies where leflunomide was added either after the initiation of mixed lymphocyte cultures (MLC) or after interleukin-2 stimulation of CTLL-4 cell proliferation. These studies revealed that leflunomide acted comparably to rapamycin, but was distinct from brequinar sodium in the MLC, and from cyclosporine and mycophenolic acid in both MLC and CTLL-4. Although previous biochemical studies indicated that leflunomide can inhibit src-family tyrosine kinase activity, more recent studies have suggested that leflunomide can also inhibit pyrimidine synthesis. Our data demonstrate that the ability of leflunomide (25-100 μM) to inhibit MLC and CTLL-4 cell proliferation is partially antagonized by uridine (25-100μM), and support the hypothesis that leflunomide inhibits pyrimidine synthesis in T cells. Unique molecular mechanisms of immunosuppression suggest that drug combinations may result in synergistic immunosuppression. Our in vitro studies revealed synergistic inhibition of T cell proliferation with the combinations of leflunomide with cyclosporine or with rapamycin. We have extended those studies to quantitate inhibition of MLC by the combinations of leflunomide and brequinar sodium or mycophenolic acid.

The anti-inflammatory and immunosuppressive characteristics of leflunomide were first described using two rodent models of autoimmune disease, adjuvant arthritis and allergic encephalomyelitis ⁽¹⁾. Since those studies, additional reports have confirmed that leflunomide is effective in controlling a number of autoimmune diseases in rodents, including systemic lupus erythematosus, proteoglycan-induced polyarthritis, organ-specific nephritic diseases, and autoimmune uveitis^(2-9). Based on those observations, the effects of leflunomide have been tested in patients with rheumatoid arthritis and where no clinically significant toxicity was detected at doses where clinical and immunological improvements had been detected ⁽¹⁾.

The first studies of leflunomide in transplantation experiments were with mice undergoing chronic graft-versus-host disease ⁽¹⁰⁾. Since then, leflunomide has been reported to prevent and reverse ongoing allograft rejection of hearts, kidneys, and skin in rodents, dogs, and cynomolgus monkeys ^(11-14). Additional studies have also demonstrated the ability of leflunomide to significantly delay the rejection of concordant cardiac xenografts⁽¹⁵⁾ and, most recently, to halt the progression of chronic rejection in rat allografts ^{(16, 17)}. These studies suggest that leflunomide controls both humoral and cellular immune responses in vivo ^{(17, 18)}.

Leflunomide and its active metabolite, A771726, are structurally unrelated to other known immunosuppressive agents ⁽¹⁷⁾. In previous in vitro studies, we reported that leflunomide inhibits T cell proliferation primarily at the level of inhibiting interleukin(IL^*)-2-driven T cell proliferation⁽¹⁹⁾.⁶ We here extend those in vitro studies to further define the effects of leflunomide on mixed lymphocyte cultures (MLC) and IL-2-driven cell proliferation, and have used the median effect analysis to define the in vitro interactions of leflunomide with cyclosporine (CsA), rapamycin (Rapa), brequinar sodium (BQN), or mycophenolic acid (MPA).

MATERIALS AND METHODS

Reagents. Leflunomide (A771726) and BQN were gifts from Dr. Robert R. Bartlett (Hoechst AG, Weisbaden, Germany), Rapa was provided by Dr. Suren Sehgal (Wyeth-Ayerst Research, Princeton, NJ), and IL-2 was from Chiron Corp. (Emeryville, CA). MPA and CsA were purchased from Sigma (St. Louis, MO) and Sandoz (East Hanover, NJ), respectively.

Mixed lymphocyte cultures (MLC). Human peripheral blood mononuclear cells (PBMC) were isolated from heparinized blood of normal healthy volunteers, and MLC were performed as described previously⁽¹⁹⁾. The percentage of inhibition of[³H]thymidine incorporation by PBMC after exposure to a single drug or a combination of drugs was calculated using the formula:Equation

Positive controls were counts per minute (cpm) from wells without drugs, and negative controls were cpm from wells containing only responder cells.

Progression of T cells into the S phase of cell cycle. Synchronously activated T cells were prepared as described previously^{(20, 21)}. Briefly, PBMC were incubated for 72 hr with phytohemagglutinin (PHA) (1 μg/ml, Sigma), then washed and incubated with IL-2 (100 Cetus U/ml) for 10 days. The washed cells were rested by overnight culture in cRPMI/10% FBS, then stimlulated with phorbol 12,13-dibutyrate (50 ng/ml, Sigma) for 2 hr. The phorbol 12,13-dibutyrate-treated cells were then washed and rested overnight in cRPMI/10% FBS. Entry of these synchronously activated T cells into S phase was marked by incorporation of [³H]thymidine by these cells when stimulated with 100 Cetus U/ml IL-2. Cells were pulsed for 4 hr and incorporation of[³H]thymidine was determined as described for MLC.

IL-2-stimulated CTLL-4 proliferation assay. The IL-2-dependent murine T cell clone, CTLL-4 (a generous gift from Dr. W. Paul, NIH), was an IL-4-dependent subclone of CTLL-2. CTLL-4 cells were maintained as suspension cultures in 50 Cetus U/ml IL-2 in RPMI 1640 supplemented with 7% FBS, 2 mM L-glutamine, 5×10^-5 M β-mercaptoethanol, 10 μM MEM nonessential amino acids, and 0.1 mM sodium pyruvate. CTLL-4 cells were starved of IL-2 for 6 to 18 hr before use in a proliferation assay. Starved CTLL-4 cells were then incubated at 10⁴/well with 100 Cetus U/ml IL-2 for 20 to 24 hr in the presence of the indicated concentrations of immunosuppressive agents. Cell proliferation was quantitated as described above for MLC.

Cell cycle analyses. T cells were enriched from PBMC by passage through a nylon wool column, then cocultured for 72 hr with PHA (1 μg/ml; Sigma). For cell cycle analyses, 5×10⁵ cells were resuspended in 100 μl of PBS then incubated with 0.5 ml of DNA staining solution (New Concept Scientific, Mississaugo, Ontario, Canada) for 30 min in the dark. The cells were then analyzed by flow cytometry on the Epics-Profile (Coulter, Hialeah, FL) and the stages in the cell cycle were determined using the MULTICYCLE program (Phoenix Flow Systems).

Analyses of drug interactions. In vitro drug interactions were analyzed, as proposed by Kahan et al. ⁽²²⁾, using the median effect principle of Chou ⁽²³⁾, where synergism, additivity, or antagonism of combined drugs was determined using the combination index. The combination index was calculated using a computer software package developed by Chou and Chou (²⁴, Dose-effect analysis with microcomputers; Biosoft, Cambridge, UK).

RESULTS

Inhibition of cell cycle progression. Previous observations that leflunomide inhibits T cell proliferation prompted us to define where in the cell cycle T cells were arrested. Cell cycle analyses were performed using a commercially prepared propidium iodide reagent for staining nuclear DNA and flow cytometry. A computer program was then used to determine the percentage of T cells in the G₀/G₁, S, and G₂/M phases. T cells were stimulated with PHA to induce entry into cell cycle (Table 1). This entry into the S and G₂/M cell cycle was completely inhibited by leflunomide, which indicates that leflunomide, similar to other immunosuppressive agents such as CsA, FK506, Rapa, and BQN, retained T cells predominantly in either the G₀/G₁ phase or the early S phase of the cell cycle (reviewed in ²⁵; A.S-F. Chong, unpublished data).

Leflunomide inhibits IL-2-stimulated T cell progression in the S phase of the cell cycle. The next series of experiments was performed to confirm that leflunomide was able to specifically inhibit IL-2-driven T cell progression into the S phase of the cell cycle. Synchronously activated T cells were prepared as described previously ^{(20, 21)}, and were restimulated with IL-2 to enter the S phase. Entry into S phase was indicated by the acquired ability to incorporate[³H]thymidine, and reproducibly occurred approximately 9 hr after addition of IL-2 (Fig. 1A). Leflunomide (100-150 μM) was able to inhibit T cell progression into the S phase(Fig. 1B). Delayed addition of leflunomide reduced its ability to inhibit progression into the S phase, although significant inhibition was observed even when leflunomide was added 12 hr after addition of IL-2. It should be noted that higher doses of leflunomide (100-150 μM) were required for inhibition of T cell cycle progression compared with inhibition of MLC, and that inhibition of progression into the S phase was only partial, even at concentrations of 150 μM leflunomide. The explanation for this observation is not clear, but may reflect the shortened assay period, thus reducing the length of time T cells are incubated with leflunomide.

Kinetics of inhibition of IL-2-stimulated T cell proliferation. Site of action is anticipated to approximately correlate with kinetics based on the rationale that drugs acting early in the cell cycle, e.g., CsA, need to be added early in the stimulation assay, but drugs that work later in the cell cycle, e.g., purine and pyrimidine synthesis inhibitors, are still effective when added late in the assay ^{(26, 27)}. Proliferation of CTLL-4 cells is dependent on the presence of IL-2, and leflunomide, at doses of 25-100 μM, inhibits the proliferation of these cells. Delayed addition of leflunomide resulted in a gradual decrease its antiproliferative activity, and 50% of maximal inhibition was observed when leflunomide was added 12-16 hr after addition of IL-2 (Fig. 2A).

We also performed similar kinetic experiments with Rapa, BQN, and MPA(Fig. 2B). Leflunomide, Rapa, and BQN displayed similar kinetics, in that 50% of maximal inhibition was observed when leflunomide was added 14-18 hr after addition of IL-2. In contrast to these immunosuppressive agents, >50% of maximal inhibition by MPA was observed when added as late as 20 hr after addition of IL-2. These experiments suggest that leflunomide acts earlier in the cell cycle than MPA, but the kinetics of its inhibition of IL-2-stimulated CTLL-4 cell proliferation is indistinguishable from Rapa or BQN.

Kinetics of inhibition of MLC. To further discriminate the action of leflunomide on T cells from other immunosuppressive agents, we compared their relative antiproliferative efficacy when added after the initiation of an MLC. Human PBMC were stimulated with gamma-irradiated allogeneic PBMC for 6 days, and different concentrations of leflunomide were added either at the initiation of the MLC or 1-5 days after initiation of the MLC. Doses of leflunomide ranging from 1 to 100 μM were effective in inhibiting T cell proliferation when added as late as 2 days into the MLC, and it was 50% as effective when added 3-4 days after initiation of the MLC(Fig. 3A).

We then performed similar kinetic experiments with immunosuppressive agents that act either early or late during T cell activation. CsA was maximally effective when added on day 0-1 (Fig. 3B), whereas Rapa, comparable to leflunomide, was maximally effective when added on day 0-2(Fig. 1B). In contrast, the pyrimidine synthesis inhibitor, BQN, was maximally effective when added as late as day 4, whereas MPA was effective when added on day 5 when the MLC was pulsed with[³H]thymidine. These kinetic experiments support our previous conclusion that the site of action of leflunomide resembles that of Rapa in that both drugs inhibit at similar stages of IL-2-stimulated T cell proliferation.

Ability of uridine and cytidine to antagonize the antiproliferative effects of leflunomide. Recent studies have indicated that the majority of the antiproliferative effects of leflunomide can be reversed by the addition of uridine ^{(18, 26)},⁷ which suggests that leflunomide may function as an inhibitor of pyrimidine synthesis. Therefore, we tested the ability of the pyrimidines, uridine and cytidine, to antagonize the antiproliferative activity of leflunomide in an MLC. Both uridine and cytidine were able to reverse the ability of leflunomide to inhibit an MLC (Fig. 4, A and B). Uridine, 25-100μM, was able to partially reverse the antiproliferative effects of leflunomide (25-100 μM), while cytidine was somewhat more potent than uridine and completely antagonized the antiproliferative effects of 25-50μM leflunomide.

Only uridine was able to antagonize the antiproliferative effect of leflunomide in the IL-2-dependent CTLL-4 cells (Fig. 5, A and B). Uridine, 25-100 μM, alone inhibited CTLL-4 proliferation by 20%; these doses of uridine were able to partially reverse the antiproliferative effects of 100 μM leflunomide and completely reverse the antiproliferative effects of 25-50 μM leflunomide. In contrast, comparable doses of cytidine had no significant effect on the antiproliferative effects of leflunomide, consistent with murine CTLL-4 cells not expressing cytidine deaminase necessary for the conversion of cytidine into uridine.

Drug interactions. In vivo studies have indicated a potent synergistic interaction between leflunomide and CsA in control of allograft, xenograft, and chronic rejection. We now have quantitated the in vitro antiproliferative potencies of leflunomide when used in combination with either CsA, Rapa, BQN, or MPA (Fig. 6), and drug interactions were analyzed using the median effect principle⁽²⁴⁾. As reported previously, the combination of leflunomide with CsA (leflunomide to CsA ratios of 2,000:1, 1,000:1, and 500:1) was synergistic at several drug mixture ratios over the most relevant effect range of 75-100% inhibition. Within the same effect range, the combination of leflunomide and Rapa produced additive to synergistic effects at both drug ratios tested (leflunomide to Rapa ratios of 50,000:1 and 25,000:1). The combination of leflunomide and BQN also resulted in synergistic effects at several relevant drug mixture ratios (leflunomide to BQN ratios of 16:1, 8:1, and 4:1). In contrast, leflunomide produced additive to slightly antagonistic effects with MPA at several relevant drug mixture ratios(leflunomide to MPA ratios of 50:1, 100:1, and 200:1). While in vitro studies do not necessarily predict in vivo drug interactions, these in vitro studies do, at minimum, indicate that leflunomide in combination with CsA, Rapa, or BQN does not result in reduced inhibition compared with each agent by itself.

DISCUSSION

This report further defines the mechanism of immunosuppression of leflunomide on T cells. Direct cell cycle analyses reveal that leflunomide inhibits PHA-stimulated T cell entry into the S phase of the cell cycle, observations that are consistent with our previous reports that leflunomide inhibits IL-2-driven T cell proliferation⁽¹⁹⁾.⁶ These observations that leflunomide inhibits T cell transition into the S phase were confirmed using T lymphoblasts that had been synchronized early in the G1 phase of the cell cycle. The addition of IL-2 induced these cells to reproducibly enter the S phase within 9 hr, as assayed by their ability to incorporate[³H]thymidine. Leflunomide was able to inhibit this entry into the S phase, and delayed addition of leflunomide relative to IL-2 gradually reduced the ability of leflunomide to inhibit progression in the S phase.

The ability of leflunomide to inhibit T cell proliferation when added after T cell stimulation was confirmed with CTLL-4 cells stimulated with IL-2. The observation that CTLL-4 cells are more sensitive to inhibition by leflunomide compared with human PBL cells is consistent with previous reports that murine cells are more sensitive to the antiproliferative effects of leflunomide⁽¹⁾. Delayed addition of leflunomide reduced its antiproliferative activity, and 50% of maximal inhibition was observed when leflunomide was added 12-16 hr into the 22-hr assay. Leflunomide, Rapa, and BQN displayed similar kinetics, whereas MPA exhibited >50% inhibition when added as late as 20 hr into the assay. The difference between the kinetics of MPA and BQN is of interest because the mode of immunosuppression by both MPA and BQN is postulated to be inhibition of de novo nucleotide synthesis, yet MPA retained the ability to inhibit CTLL-4 proliferation better than BQN when added late in the assay. These observations are consistent with the recent report that BQN may have an additional effect that inhibits early in the cell cycle by inhibiting the transcription of IL-2 mRNA and IL-2R expression in activated T cells ⁽²⁸⁾.

The ability of leflunomide to inhibit T cell proliferation when added after T cell stimulation was further confirmed in a 6-day MLC proliferation assay. Leflunomide was effective in inhibiting T cell proliferation when added as late as 2 days after initiation of MLC and was 50% effective when added 3-4 days after initiation of MLC. Such kinetics were similar to those observed with Rapa, and are consistent with our previous conclusion that leflunomide resembles Rapa in that both drugs inhibit IL-2-stimulated signals that drive T cell proliferation. In contrast to leflunomide and Rapa, the nucleoside synthesis inhibitor BQN was equally effective when added between days 0 and 4 after initiation of MLC, whereas MPA was equally effective when added as late as day 5 of the MLC.

We have reported previously that leflunomide is able to inhibit the early tyrosine phosphorylation events associated with TCR/CD3 aggregation and IL-2 binding to the IL-2 receptor ⁽²⁹⁾.⁶ However, preliminary observations indicate that late addition of leflunomide no longer inhibited intracellular tyrosine phosphorylation, but that cell proliferation was still inhibited (A.S-F. Chong, unpublished data). These observations suggested that leflunomide may have other effects on T cells besides inhibition of tyrosine phosphorylation. Recent reports have suggested that much of the antiproliferative effect of leflunomide can be reversed by uridine ^{(18, 26)}.⁷ We here provide additional data in support of those observations and demonstrate that uridine and cytidine can reverse, at least in part, the ability of leflunomide to inhibit MLC. In addition, we also report that only uridine, not cytidine, can partially reverse the antiproliferative effects of leflunomide on CTLL-4 cells. The different effects of cytidine on the MLC and CTLL-4 cells may be related to the absence of cytidine deaminase, which converts cytidine to uridine, in some murine cell lines (43). The uridine reversal suggests that a major effect of low doses of leflunomide on T cells is inhibition of pyrimidine synthesis. Thus, as with BQN⁽³⁰⁾, the effect of leflunomide on T cells is complex; at lower concentrations and in the absence of exogenous pyrimidine, inhibition of pyrimidine synthesis may be the primary mode of immunosuppression; however, at high doses (≥100 μM) and in the presence of exogenous pyrimidines(10-25 μM), the ability to inhibit the early tyrosine kinase-dependent signaling events may also contribute to the potent immunosuppressive effects of leflunomide.

It has been proposed that drugs with different mechanisms of immunosuppression may function synergistically when used together^{(22, 31)}. We have, in fact, observed that the combination of leflunomide and CsA was synergistic in controlling xenograft⁽¹⁵⁾ and chronic rejection ⁽¹⁷⁾. We have performed in vitro experiments with leflunomide in combination with CsA, Rapa, BQN, or MPA. Drug interactions were analyzed by the median effect analyses ^{(23, 24)}. As reported previously, the combinations of leflunomide with CsA or with Rapa were, in general, synergistic or additive at several drug mixture ratios. We here report that the combination of leflunomide with BQN was similarly synergistic or additive, while the combination of leflunomide with MPA was additive or slightly antagonistic depending on drug ratios. Although in vitro studies do not necessarily predict in vivo outcomes, such studies have provided insights into molecular targets, as illustrated by the interaction of FK506 and Rapa^(32-35). Thus, the observation that selective combinations of leflunomide and MPA are antagonistic needs to be further investigated, particularly with respect to their effects on total intracellular purine and pyrimidine pools.

Our studies presented here suggest that leflunomide inhibits T cell activation by inhibiting entry into the S phase of the cell cycle. Although kinetics of inhibition suggest that leflunomide is most similar to Rapa, the observation that much of its antiproliferative effects can be reversed by uridine suggests that leflunomide inhibits pyrimidine synthesis. These observations, together with those of inhibition of early tyrosine phosphorylation events, illustrate a multiplicity of effects of leflunomide on T cells. The relative contributions of these activities to the potent immunosuppression by leflunomide in vivo remain to be elucidated.

Acknowledgments. We are grateful for the gifts of leflunomide and brequinar sodium from Dr. Robert R. Bartlett (Hoechst AG Werk Kalle-Albert, Weisbaden, Germany), and rapamycin from Dr. Suren Sehgal (Wyeth-Ayerst Research, Princeton, NJ). IL-2 was a generous gift from the Chiron Corp.(Emeryville, CA).

Figure 1: Ability of leflunomide to inhibit T cell progression into the S phase of the cell cycle. T cells synchronized at early G₁ reproducibly enter the S phase at approximately 9 hr after addition of IL-2 (100 Cetus U/ml) (A). Leflunomide inhibits T cell progression in the S phase (B). T cells were synchronized by treating PBMC with PHA (1 μg/ml) for 72 hr and with IL-2 (100 Cetus U/ml) for 10 days, starving, and then treating with phorbol 12,13-dibutyrate (50 ng/ml) for 2 hr. Cells were then washed and rested for 16 hr. These cells were resuspended in cRPMI/2% FBS, aliquoted at 10⁴ cells/well, and restimulated with IL-2 (100 Cetus U/ml). Leflunomide (Lef; 100 and 150 μM) was added at the indicated times after addition of IL-2 (B), and 1 μCi of [³H]thymidine/well was added at the indicated times (A) or at 12 hr after IL-2 (B). Data were provided as cpm from 1 representative experiment of 5 independent experiments. Counts for CTLL-4 proliferation at 12 hr for positive controls ranged from 25,000 cpm.

Figure 2: Effect of delayed addition of leflunomide (Lef) on its ability to inhibit CTLL-4 cell proliferation (A). Effect of delayed addition of Lef, compared with Rapa, BQN, or MPA, on its ability to inhibit CTLL-4 cell proliferation (B). Immunosuppressive agents were added to CTLL-4 cells(10⁴/well) at the indicated concentrations with IL-2 (100 Cetus U/ml), or at the indicated times thereafter. Cells were pulsed after 20 hr for 2 hr with 1 μCi/well [³H]thymidine. Proliferation assays were performed as described in Materials and Methods. Data represent mean% inhibition ± SEM from 1 of 3 representative experiments (A), or% maximum inhibition ± SEM calculated from 3 to 4 separate experiments(B). Maximum inhibition refers to inhibition by each drug when added at time 0. Counts for positive, uninhibited controls ranged from 11,000 to 102,000 cpm and maximum inhibition by drugs ranged from 60%.

Figure 3: Effect of delayed addition of leflunomide (Lef) on its ability to inhibit MLC (A). Effect of delayed addition of Lef, compared with Rapa, BQN, or MPA, on its ability to inhibit MLC (B). Responder PBMC(2×10⁵/well) and irradiated stimulator cells(2×10⁵/well) were cultured in cRPMI/5% pooled human serum. Immunosuppressive agents were added at the indicated concentrations with irradiated stimulators, or at the indicated times thereafter. Cells were pulsed on day 5 for 16 hr with 1 μCi/well [³H]thymidine. Proliferation assays were performed as described in Materials and Methods. Data represent mean% inhibition ± SEM from 1 of 3 representative experiments (A), or% maximum inhibition ± SEM calculated from 3 to 4 separate experiments (B). Maximum inhibition refers to inhibition by each drug when added at time 0. Counts for positive (uninhibited) controls ranged from 10,000 to 72,000 cpm and maximum inhibition by drugs was approximately 60%.

Figure 4: Uridine (A) and cytidine (B) can reverse the antiproliferative effects of leflunomide (Lef) in MLC. Uridine, cytidine, or leflunomide was added at the indicated concentrations at the start of MLC, and the proliferation assays were performed as described in Materials and Methods. Data represent mean% inhibition ± SEM calculated from 3 separate experiments. Counts for positive (uninhibited) controls ranged from 10,000 to 72,000 cpm.

Figure 5: Uridine (A), but not cytidine (B), can reverse the antiproliferative effects of leflunomide (Lef) on IL-2-driven CTLL-4 cells. Uridine, cytidine, or leflunomide was added at the indicated concentrations, prior to addition of IL-2 (100 Cetus U/ml), and the proliferation assays were performed as described in Materials and Methods. Data represent mean% inhibition ± SEM calculated from 3 separate experiments. Counts for positive (uninhibited) controls ranged from 11,000 to 102,000 cpm.

Figure 6: Inhibition of MLC by combination of leflunomide (Lef) with CsA(A), Rapa (B), BQN (C), or MPA (D). The indicated concentrations of these drugs were added at the start of MLC, and the proliferation assays were performed as described in Materials and Methods. Data represent mean% inhibition ± SEM calculated from 3 to 4 separate experiments.