Springer Semin Immunopathol (1993) 14:353-380

SpringerSeminars in Immunopathology 9 Springer-Verlag 1993

The design and development of an immunosuppressive drug, mycophenolate mofetil Anthony C. Allison and Elsie M. Eugui Institute of Immunology and Biological Sciences, Syntex Research, Palo Alto, CA, USA

Introduction Immunosuppressive agents have traditionally been identified in two ways. Some drugs developed for cancer chemotherapy were incidentally found to have immunosuppressive activity (cyclophosphamide, methotrexate, azathioprine), whereas screening of fermentation products revealed the immunosuppressive activity of cyclosporins, FK 506 and rapamycin. Establishing the mode of action of these compounds lagged behind the demonstration of their immunosuppressive activity, and information on this subject is still incomplete. Cyclophosphamide and azathioprine have several active metabolites and mechanisms of action [18, 25]. Many years after the demonstration that cyclosporin A is immunosuppressive [16] the drug was found to bind prolyl cis-trans isomerase and inhibit the activity of the enzyme [29, 84]. All of the currently used immunosuppressive drugs have limitations. Patients with diseases such as rheumatoid arthritis and organ transplant recipients are treated with immunosuppressive agents for years. Genotoxic effects of these drugs and their metabolites can increase the risk of malignancy [54, 57]. Immunosuppressive therapy also increases susceptibility to viral and other infections, and it is desirable that when such a drug is withdrawn immune function can soon recover to control the infections. The effects of cyclophosphamide on lymphocytes are not rapidly reversible, and the drug produces hemorrhagic cystitis, as well as being mutagenic and carcinogenic [12, 22]. The most serious side effect of cyclosporin A is, of course, nephrotoxicity [13]. Methotrexate, which inhibits synthesis of thymidylate as well as de novo purine synthesis [52], has non-selective antiproliferative effects, which are no more potent on lymphocytes than on other cell types, including hemopoietic precursors, fibroblasts and endothelial cells [47, 52]. The toxicities associated with long-term methotrexate therapy include liver damage, pneumonitis and bone marrow suppression [1]. Correspondence to: A. Allison, 2513 Hastings Drive, Belmont, CA 94002, USA

354

A.C. Allison and E. M. Eugui

Hence, there has been a need for a new immunosuppressive drug with reversible antiproliferative effects which are more potent on lymphocytes than on other cell types, including hematopoietic cells, and which is free of hepatotoxicity, nephrotoxicity, mutagenicity and other serious side effects. We decided to approach to design of such a drug on the basis of identifying a metabolic pathway more susceptible to inhibition in human T and B lymphocytes than in other cell types. Our observations on children with inherited defects of purine metabolism provided a lead. Indeed, the belief that this strategy could result in the production of a clinically useful immunosuppressive drug was the principal motivation for us to accept positions in the pharmaceutical industry late in 1981, where we commenced work with a team to accomplish this. Ten years later that confidence was shown to be justified, which illustrates the long time required for drug development. In view of our long-standing interest in the use of drugs in transplantation [42, 89], prevention and treatment of graft rejection was a major objective of the Syntex program on immunosuppressives from its inception and throughout the development of mycophenolate mofetil. It was gratifying when the drug proved to be useful both for the prevention of allograft rejection in humans and for the treatment of ongoing rejection [85]. Purine metabolism in lymphocytes and genetic defects There are two major pathways of purine synthesis (Fig. 1). In the de novo pathway the ribose phosphate portion of purine nucleotides is derived from 5-phosphoribosyl-l-pyrophosphate (PRPP), which is synthesized from ATP and ribose5-phosphate, a product of the pentose phosphate pathway. The purine ring is assembled on ribose phosphate by a series of steps, the first of which is catalyzed by glutamine-PRPP aminotransferase. PRPP is also used by the purine salvage pathways, including the major pathway catalyzed by hypoxanthine-guanine phosphoribosyltransferase (HGPRTase). Production of an adequate level of PRPP is, therefore, essential for synthesis of purine ribonucleotides by either pathway. Ribonucleotide diphosphates (ADP and GDP) are converted by ribonucleotide diphosphate reductase into the corresponding deoxyribonucleotides (dADP and dGDP), which are phosphorylated to produce dATP and dGTP, substrates for D N A polymerase. Activities of the key, rate-limiting enzymes, PRPP synthetase and ribonucleotide reductase, are allosterically regulated by nucleotides (Fig. 1). In bacteria both adenosine and guanosine nucleotides, signalling an abundance of purine nucleotides, inhibit PRPP synthetase [87]. However, we have shown that this is not the case with PRPP synthetase from human lymphocytes [35]: the enzyme is inhibited by adenosine nucleotides (AMP and ADP) but activated by guanosine nucleotides (GMP, GDP and GTP, Fig. 2). Such regulation is a critical factor in the therapeutic strategy discussed in this chapter. The overall catalytic activity of ribonucleotide reductase is decreased by binding of dATP, whereas binding of dGTP stimulates ADP reduction [87]. It follows that an excess of adenosine nucleotides and/or depletion of guanosine nucleotides can decrease the pool of PRPP, and that an excess of dATP and/or depletion of dGTP can inhibit ribonucleotide reductase activity, thereby decreasing the pool of substrates required for D N A polymerase activity. In other words, adequate levels of guanosine and

Mycophenolate mofetil

355 Pathways

of P u r l n e B i o s y n t h e s i s DeNovo Pathway

Glycoprotein Synthesis ~

Guanine

RNA

Ribose-SP + ATP

l

1

Adenosine TP

Ribonucleotide Reductase

MI

pRPP Synthetase

Guanosine TP

a Guanosine MP ,11 PRPP I M P Dehydrogenase (Lesch-Nyhan) (IMPD)

Deoxyguanosine TP

RNA

Inosine MP

~, Adenosine MP ,ll Adenosine Deaminase

~

(ADA)

Mycophenollc Acid

Deoxyg uanosine DP

DNA

Ribonucleotide Reductase

Deoxyadenosine DP

Regulation by nucleotides of rate-limding enzymes Enzyme

Stimulation

Inhibition

PRPP synthetase

GMP, GDP, GTP, dGTP

AMP, ADP dATP

Ribonucleotide red uctase

Oeoxyade•osine TP DNA

Fig. 1. Pathways ofpurine biosynthesis, showing thecentral position ofinosine monophosphate (IMP). Mycophenolic acid inhibits IMP dehydrogenase, thereby depleting GMP, GTP and dGTP. Two rate-limiting enzymes in lymphocytes are activated by guanosine ribonucleotides and dGTP, but inhibited by AMP, ADP and by dATP, respectively

Fig. 2. Two experiments showing effects of added nucleotides on the activity of 5-phosphoribosyl-l-pyrophosphate (PRPP) synthetase from human lymphocytes [35]. The enzyme is activated by guanosine ribonucleotides and inhibited by AMP and ADP

N 0

GMP

cGMP

0

ADP

deoxyguanosine nucleotides are required for proliferative responses of lymphocytes to antigenic and mitogenic stimulation, whereas an excess of adenosine or deoxyadenosine nucleotides would be expected to inhibit proliferation. As shown in Fig. I, IMP is at the branch point in purine nucleotide synthesis, since it can be converted to A M P or GMP. Two enzymes decrease levels of adenosine nucleotides relative to guanosine nucleotides: adenosine deaminase (ADA) and inosine monophosphate dehydrogenase (IMPDH). From the arguments just presented it would be expected that both of these enzymes would be required for lymphocyte proliferation. In 1972 Eloise Giblett and her colleagues [37] showed that children with inherited A D A deficiency have a selective decrease in the numbers and functions of T and B lymphocytes in the presence of normal numbers of neutrophils,

356

A. C. Allison and E. M. Eugui

DeNovo pathway

Salvage pathway

L Lymphocytes

I Fibroblasts

Neurons

Smooth muscle cells

Endothelial cells Intestinal epithelial cells

Fig. 3. Classification of cell types according to the relative importance of de novo purine synthesis and the major purine salvage pathway catalyzed by hypoxanthine-guanine phosphoribosyltransferase (HGPRTase)

1. Effects of genetic defects of enzymes catalyzing purine metabolism on the functions of different cell types in humans

Table

Enzyme defect

T lymphocytes B lymphocytes

Neurons

Reference

Adenosine deaminase Hypoxanthine-guanine phosphoribosyl transferase

$ +

+ +

Giblett et al. [37] Allison et al. [7]

$ +

erythrocytes and platelets and normal brain function. Shortly afterwards we showed that children lacking HGPRTase (Lesch-Nyhan syndrome) have essentially normal numbers and functions of T and B lymphocytes [7] (Table 1). These findings led us to postulate [7, 8] that de novo purine synthesis is crucially important for proliferative responses of human T and B lymphocytes to mitogens, whereas the major salvage pathway catalyzed by HGPRTase is not required for lymphocyte proliferation. Children with the Lesch-Nyhan syndrome have mental deficiency and compulsive self mutilation, showing the importance of the salvage pathway controlled by HGPRTase for brain cells. The level of this enzyme in brain is higher than in any other tissue, whereas the activity of glutamine-PRPP aminotransferase, which catalyzes the comitted step in the de novo pathway, is low in brain [87]. Thus, cell types and tissues can be arranged according to their dependence on the de novo and salvage pathways of purine synthesis (Fig. 3), with lymphocytes at one extreme, brain cells at the other and most cell types, able to use both pathways, occupying an intermediate position. The importance of adequate PRPP levels and of de novo purine synthesis in lymphocytes responding to antigenic and mitogenic stimulation is clear. We found a rapid and sustained increase in PRPP concentrations in human lymphocytes following mitogenic stimulation [48], as well as markedly increased de novo purine synthesis, as shown by incorporation of [14C]glycine into purine nucleotides [50]. In stimulated cells the label was about equally distributed into pools of adenosine nucleotides on the one hand and inosine and guanosine nucleotides on the other, whereas in resting cells synthesis of adenosine nucleotides predominates. Thus, proliferation is accompanied by activation of PRPP synthetase and I M P D H , increasing de novo purine synthesis and channeling it towards GMP. As discussed below, in stimulated lymphocytes the gene for a distinct isoform o f I M P D H is activated, so the total amount of enzyme per cell is increased, in addition to allosteric regulation by nucleotides. Elevation of PRPP concentrations in human lymphocytes following phytohemagglutinin (PHA) stimulation, and suppression of that elevation by adenosine, has also been reported by Peters et al. [74].

Mycophenolate mofetil

357

Inhibiting enzymes of purine metabolism

Coformycin and 2'-deoxycoformycin Because of the immunodeficiency associated with genetic defects of adenosine deaminase it was reasonable to suppose that an inhibitor of that enzyme might have immunosuppressive activity. When such an inhibitor, coformycin, became available, we studied its effects on responses of human peripheral blood lymphocytes to mitogens [49] and of mouse spleen cells to lipopolysaccharide [50]. Coformycin (1 ~tM) inhibited lymphocyte proliferation, and the inhibitory effect of this concentration of the drug, or even lower concentrations, was potentiated by adenosine, suggesting that accumulation of adenosine and/or deoxyadenosine nucleotides was responsible for the antiproliferative effect. Coformycin and T-deoxycoformycin have shown efficacy for treatment of humans with leukemia [77], but side effects preclude their use as immunosuppressive agents.

Mycophenolic acid It was, therefore, necessary to use another approach: depletion of GMP by inhibiting IMPDH. Of several possible inhibitors we selected for detailed study mycophenolic acid (MPA, Fig. 4), a fermentation product of several Penicillium species. MPA was preferred to nucleoside analogues such as mizoribine, bredinin, ribavirin and tiazofurin because the latter must be phosphorylated to inhibit IMPDH, and because of the side effects of this class of compounds. The efficiency

Mycophenolate mofetil

o OH CH3 0 ~ ' ~ ~ ' ~ " I I~OCH 3 CHs

Feces 0~ ~

N/'~O .HCI~

t Intestine

Iesterase Mycophenolic

acid

0 OH CH3 < ~ . ~ . ~ ~

COOH

=,

~r ~OCH3

CH3

~glucur~

aciMY~ OphenOlic glucuronide

/

T I glucur~176 transferase

O 0 ~J~ ~ T

CH3 A ./~L. A

Blood

/

/

J

"OCH3

CH3

Fig. 4. Structure of the morpholinoethylester of mycophenolicacid (mycophenolatemofetil), mycophenolicacid and its glucuronide and sites of the interconversionand excretion

358

A.C. Allison and E. M. Eugui

of nucleoside phosphorylation can vary in different cell types, and the relevance of target cells in addition to lymphocytes (e.g., cells of the monocyte-macrophage lineage, endothelial cells and smooth muscle cells) is described below. Nucleoside analogs frequently have undesirable effects, such as inhibiting DNA repair enzymes and producing chromosome breaks. MPA is not a nucleoside, and is a potent, non-competitive, reversible inhibitor of eukaryotic but not prokaryotic IMP dehydrogenases [32, 90]. Contrary to early reports, MPA does not inhibit GMP synthetase, the enzyme catalysing the conversion of XMP to GMP. Because of the importance of guanosine and deoxyguanosine nucleotides in activating PRPP synthetase and ribonucleotide reductase, respectively, we postulated that depletion of GMP (and consequently GTP and dGTP) by inhibiting IMPDH would have antiproliferative effects; furthermore, since lymphocytes rely on de novo purine synthesis, whereas other cell types do not, antiproliferative effects produced in this way would be more potent on lymphocytes than on other cell types. We established a Syntex Research programme on immunosuppressive drugs in 1982. In that year we found that MPA strongly inhibits responses of human lymphocytes to mitogenic stimulation and in mixed lymphocyte responses (Fig. 4). Soon afterwards we found that MPA inhibits humoral and cell-mediated immune responses in mice. A literature search revealed that MPA and analogs had been studied for anti-tumor effects, mainly by Japanese investigators, some of whom had incidentally observed immunosuppressive activity [64, 71], which had not been followed up. Investigators of the Eli Lilly company had also collaborated with dermatologists to show that orally administered MPA has activity against psoriasis [39]. Patients had received 2 - 7 g of the drug daily for as long as 13 years, with no serious side effects [26]. We were able to obtain samples of serum from three MPA-treated patients and found potent inhibitory activity on mitogenic responses of normal human lymphocytes. These findings, together with the good tolerance of MPA by human subjects, convinced us that we were on the right trail. Early in the research program Dr. Peter Nelson focused synthetic effort on obtaining a derivative of MPA with improved oral bioavailability. Then the morpholinoethyl ester of MPA [mycophenolate mofetil (MM) Fig. 4] was found, unexpectedly, to have improved bioavailability in primates as compared with MPA [61]. The ester is rapidly hydrolyzed to yield MPA, both in human peripheral blood mononuclear cell cultures and in vivo. MM proved suitable for pharmaceutical formulation and was selected for development. Most of our in vivo studies were made with MM, although comparative studies with MPA were often included. MPA is rapidly converted into the glucuronide, which is the principal metabolite.

Isoforms of IMP dehydrogenase Natsumeda et al. [69] isolated two distinct cDNA clones (types I and II) encoding IMPDH from a human spleen cDNA library. The clones encode closely related proteins of 514 residues showing 84% sequence correspondence. Northern hybridization analysis of poly(A) + RNA from normal human leukocytes showed expression of the type I enzyme, whereas leukemic cells and ovarian tumor cells

Mycophenolate mofetil

359

100

_) oo

\\,

[

10

100 1,000 10,000 Concentration of MPA (nm)

Fig. 5. Potent inhibition by mycophenolic acid (MPA) of the proliferation of human peripheral lymphocytes (PBL) responding to stimulation by a T cell mitogen (phytohemagglutinin; PHA), pokeweed mitogen (PWM) and a B cell mitogen (staphylococcal protein A sepharose) [27]. Higher concentrations of MPA are required to inhibit the proliferation of human dermal fibroblasts (FIB) in response to IL-lfl or human umbilical vein endothelial cells (EC) in response to basic fibroblast growth factor (FGF). 9 EC/FGF; o FIB/IL-1; 9 FIB/FGF; A PBL/ PHA; v PBL/PW; o PBL/5PAS

expressed predominantly the type II gene. Immunological studies [55] showed abundant type II I M P D H in leukemic cells. Subsequent studies [68] showed predominant expression of the type I gene in resting human lymphocytes; when the cells were stimulated by P H A (T cells) or transformed by Epstein-Barr virus (B cells) expression of the type I gene was unchanged, whereas the type II gene was strongly expressed. Both isoforms o f the enzyme have been expressed in E. coli and purified [55]. The type II isoform is four to five times more sensitive to MPA than is the type I isoform (Y. Natsumeda, personal communication). Thus, activation of human lymphocytes is accompanied by a marked increase in I M P D H activity, mainly due to the type II isoform. It remains to be seen whether all rapidly dividing cells express type II I M P D H or whether fibroblasts and endothelial cells, for example, which are less sensitive to MPA (Fig. 5), do not. Part o f the lymphocyte-selective activity o f MPA is almost certainly due to the expression in these cells of the type II isoform o f the enzyme. The relative insensitivity o f human neutrophils to MPA-mediated depletion of G T P is probably due, at least in part, to the expression of the type I isoform of the enzyme.

Inhibition of lymphocyte proliferation Concentrations of MPA that are readily attainable therapeutically inhibit the proliferative responses o f human peripheral blood mononuclear cells to P H A (a T cell mitogen), pokeweed mitogen (a T-dependent B cell mitogen) and Staphylococcus protein A sepharose (a B cell mitogen) [27]. Mixed lymphocyte responses are also inhibited by M M and MPA. In all cases the ICso is less than 100 nM, a concentration o f drug having no antiproliferative effect on fibroblasts or endothelial cells (Fig. 5), or on other cell types studied. Thus, the drug has greater in vitro antiproliferative on lymphocytes than on other cell types, as predicted. However, clinically attainable concentrations of MPA inhibit the proliferation of

360

A.C. Allison and E. M. Eugui

Table 2. Effect of mycophenolic acid (MPA) added at the time of initiation of human mixed lymphocyte reaction and 72 h later Time of MPA addition (h) 0 72 0 72

% of inhibition

IC5o nM

1000 nM

100 nM

10 nM

1 nM

98* 90* 94* 95*

81 ** 73** 82** 61"*

25 11 -7 6

4 11 -27 2

30 50 40 70

* P
h u m a n smooth muscle cells, which is relevant to effects on proliferative arteriopathy discussed below. When added to ongoing mixed lymphocyte reactions 72 h after initiation, M P A was still inhibitory, showing that it blocks a late event in lymphocyte responses (Table 2). MPA and M M also strongly inhibit proliferation of all the h u m a n T and B lymphocytic cell lines tested [27]. One of these lines lacks H G PRTase, which is experimentally convenient.

Intracellular pools o f G T P and d G T P

It has been postulated that G-proteins m a y be involved in the transduction of mitogenic signals to h u m a n T lymphocytes [67]; depletion of G T P might, therefore, affect such transduction systems. To ascertain whether the antiproliferative effects o f M P A are due to depletion of G T P or o f d G T P , pools of nucleotides were measured in mitogen-activated h u m a n peripheral blood mononuclear cells and h u m a n T lymphocytic cell lines in the presence or absence of MPA, and in the presence of MPA when Guo, G u a or D G u o were added back to the culture medium (Fig. 6). As expected, MPA depletes G T P and dGTP; G u a or G u o efficiently reverse depletion of G T P and less efficiently restore d G T P ; d G u o efficiently reverses depletion of d G T P and less efficiently restores G T P [9]. In the MPA-treated HGPRTase-deficient T cell line, BUC-7, d G u o restores d G T P but not G T P , as expected from the metabolic pathways. Clinically attainable concentrations of M P A (1 - 10 ktM) significantly deplete G T P in h u m a n lymphocytes and monocytes but not in neutrophils (Fig. 7). This m a y be due to cellular selectivity o f the inhibition of I M P D H or to a low rate of utilization of G T P in neutrophils. The consequences of this difference in cell types are discussed below. The effects of these manipulations on cellular proliferation are shown in Fig. 8 [9]. M P A or M M (1 laM) completely suppressed D N A synthesis in PHA-stimulated peripheral blood cells. Adding back 10 ktM d G u o or 50 IxM G u o restored D N A synthesis to control levels observed in the absence of the drug, as shown by [3H]thymidine (dThd) incorporation. Neither Ado nor dAdo in any concentration tested restored D N A synthesis in MPA-treated cells or had significant effects in the absence of the drug. In HGPRTase-deficient cells treated with M P A d G u o partially restored D N A synthesis, whereas G u a or G u o did not. In cells with no H G P R T a s e higher concentrations of G u o than of d G u o were needed to restore

Mycophenolate mofetil

r

361

1.25

1.25

1.00

1.00

'~O

0.75

0.75 r

0.50

0.50

0.25

0.25

0.00

0.0C

GTP dGTP Fig. 6. Treatment of T lymphocytic cells (CEM) with 1 ~tM MPA for 6 h depletes GTP and dGTP pools. Adding back guanine in the presence of MPA restores GTP and dGTP to levels higher than observed in the absence of MPA (hatched) [9] 120 A ,.I O

100

z

O ~

g

~

80

~0

I= O O ~

2O

Z

0 10-5M

10-6M MYCOPHENOLIC

10-5M ACID

10-5M

10-6M

CONCENTRATION

Fig. 7. Incubation of human peripheral blood lectin-stimulated lymphocytes and monocytes with 10 p.M MPA significantly depletes GTP levels whereas there is no depletion in neutrophils (N. Byars and A.C. Allison, unpublished). [] GTP; 9 ATP D N A synthesis, in keeping with the requirement for conversion to dGTP. Thus, the inhibition of D N A synthesis in MPA-treated lymphocytes is apparently due to depletion of dGTP. The findings show the metabolic selectivity of action o f MPA: if the drug were acting on other enzymes or metabolic functions, or on thymidine transport, it would not have been possible to restore proliferation with G u o or dGuo. Excess dGuo inhibits D N A synthesis in the presence or absence o f MPA, showing that the dGTP pool must be kept within a rather narrow optimal range.

362

A.C. Allison and E. M. Eugui

120

-443 (1 laM)

o 100 vI< 80

o. 0 40

0

I O0

200 pM nucleoslde

300

120" NO RS-61443

~" 100 x

~"

60402o-

1O0

200 pM nucleoside

300

Fig. 8. RS-61443 (MM, 1 laM) in-

hibits to baseline levels DNA synthesis in human PBL stimulated by PHA. Addition of deoxyguanosine, or a higher concentration of guanosine, restores DNA synthesis to levels observed in the absence of MPA. No restoration is observed following addition of adenosine or deoxyadenosine. Excess deoxyguanosine in the presence or absence of MPA decreases DNA synthesis [9]. --m-Guanosine; -e- deoxyguanosine; ~ adenosine; --o- deoxyadenosine

120 2

--~

100

so

.-~ 6o ~.,

411

0 0

10

100

1000

Concentration of Mycophenolic Acid (nM)

Fig. 9. Inhibition by MPA of antibody formation by human peripheral blood lymphocytes polyclonally activated by staphylococcal protein A-sepharose [9]

Cytokine production M P A in c o n c e n t r a t i o n s up to 10 ~tM did n o t inhibit interleukin (IL)-I fl f o r m a t i o n by activated h u m a n m o n o c y t e s [27], Cyclosporin A and F K 506 inhibit early stages o f l y m p h o c y t e s activation, including the p r o d u c t i o n o f IL-2. In contrast, M M or M P A in concentrations up to 1 ~tM had no detectable effect on IL-2 p r o d u c t i o n in mitogen-activated h u m a n peripheral blood lymphocytes [27]. These findings show that early signal transduction systems in T l y m p h o c y t e s are n o t inhibited by M P A , a conclusion confirmed by D a y t o n et al. [21].

Mycophenolate mofetil

363

Antibody formation in vitro Antibody formation by polyclonally activated human B lymphocytes was almost completely inhibited by 100 nM MPA [27] (Fig. 9). In another laboratory therapeutically attainable doses of MPA were found to inhibit secondary responses of human spleen cells to tetanus toxoid; cyclosporin A did not inhibit ongoing antibody responses [40]. The possible relationship of inhibition of antibody formarion to prevention of chronic graft rejection is discussed below. However, the capacity of M M to inhibit antibody formation in human allograft recipients is currently under investigation.

Inhibition of the transfer of fueose and mannose to glycoproteins, including adhesion molecules If depletion of GTP in lymphocytes by MPA does not impair early signal transduction in these cells, the question arises whether it has any other important metabolic consequences. The answer to that question is yes: we have defined one effect which is likely to be important in vivo, and there may be others. MPA-mediated depletion of GTP inhibits the transfer of fucose and mannose to glycoproteins, some of which are adhesion molecules facilitating the attachment of leukocytes to endothelial cells and to target cells. By this mechanism MPA could decrease the recruitment of lymphocytes and monocytes into sites of chronic inflammation, such as synovial tissue of patients with rheumatoid arthritis, and into sites of vascularized organ graft rejection, as well as interactions of lymphocytes with other cell types, thereby inhibiting ongoing rejection. Recruitment of leukocytes into sites of inflammation occurs in several stages. First a relatively weak interaction between leukocytes and endothelial cells allows leukocytes to roll along the walls of postcapillary venules and some other blood vessels rather than being swept through in the circulating blood. The rolling interaction can be converted into a stronger interaction which causes the leukocytes to stick; the surfaces of leukocytes and endothelial cells are deformed sufficiently to allow the leukocytes to squeeze between the endothelial cells and migrate to perivascular connective tissue. Third, the leukocytes in connective tissue respond to chemotactic stimuli by directional migration. The first two stages are mediated by adhesion molecules. Surface expression of one group of adhesion molecules, the selectins, plays a major role in the initial interaction between leukocytes and endothelial cells responsible for rolling; this interaction is a prerequisite for integrin-mediated sticking [59, 60]. Selectins are so termed because of their lectin-like structures and properties. They share amino acid sequences with lectins, and their complementary ligands are fucose-containing oligosa~harides (Table 3). While a great deal of attention has been given to complementary interactions of adhesion molecules in leukocytes and endothelial cells, it is clear that adhesion molecules also participate in the initiation and effector phases of immune responses. Interactions between antigen-presenting cells and lymphocytes require complementary binding of adhesion molecules [11], and the same is true of interactions of effector lymphocytes with target cells [43]. It follows that blocking the interactions between complementary adhesion molecules could exert immuno-

364

A.C. Allison and E. M. Eugui

Table 3. Adhesionmoleculeson leukocytesand complementaryligands on endothelialcells Leukocyte

Endothelial cell

Lacto-N-fucapentaoseIII (CD15) ~(2,3)-Sialyl, -ct(1,3) fucosyllactosaminoglycan L-selectin (LECAM-1, MEL14) VLA-4

P-selectin (GMP-140, CD62) E-selectin (ELAM-1) Sialylated fucosyloligosaccharide VCAM-1

suppressive and anti-inflammatory activity. A striking example is the administration of monoclonal antibodies against ICAM-1 and its complementary ligand LFA-1 to mice that have received cardiac allografts. Treatment with these antibodies for 6 days after transplantation induces donor-strain-specific tolerance to the alloantigens [51]. Glycosylation of proteins and lipids occurs through nucleotide intermediates (Fig. 10). Glucose, galactose and their amines are transferred to dolichol phosphate and then to proteins through uridine-diphospho intermediates, whereas fucose and mannose are transferred through guanosine-diphospho intermediates [56]. We have found that in activated human peripheral blood lymphocytes treatment with MPA significantly decreases the transfer of mannose to dolichol phosphate and to membrane glycoproteins [10]. This was shown by following labelled mannose, by measuring the expression of mannose on the surface of the cells, using a specific lectin for terminal mannose (Fig. 11), and by measuring the terminal mannose content of different membrane glycoproteins. Immunoprecipitation studies showed that one of the lymphocyte glycoproteins affected is VLA-4, the ligand for VCAM-1 on activated endothelial cells [24]. Treatment of either T cells or IL-l-activated endothelial cells with MPA in therapeutically attainable doses (1-10 gM) decreased lymphocyte attachment and, when both cell types were treated with MPA, the attachment was further inhibited (Fig. 12). This effect is presumably due mainly to decreased expression of ligands for selectins on interacting cells. In addition, decreased glycosylation of VLA-4 appears to reduce its affinity for VCAM-1, even though the latter is not a selectin; perhaps the conformation of VLA-4 depends on glycosylation, as known for several glycoproteins [78]. Inhibition of glycosylation by depletion of GTP and/or UTP is theoretically more appealing than attempting to do it by inhibitors of glycosyl transferases. There are several fucosyl and manosyl transferases with different stereospecificities, and it would be difficult to design and use inhibitors of all relevant enzymes. Such mixtures of inhibitors might well have generalized toxicity, whereas the effects of MPA are focused on lymphocytes and monocytes, the cells involved in immunologically driven inflammatory reactions. Use of recombinant adhesion molecules themselves, or of antibodies against them, is also a less attractive option since many are involved [59, 92], and they cannot be administered orally. For the same reasons synthetic carbohydrate ligands are inconvenient. Because transfer of other sugars to glycoproteins requires UDP-sugar intermediates, depletion of UTP would also be expected to decrease glycosylation. Inhibitors of pyrimidine synthesis, such as brequinar, would therefore be expect-

Mycophenolate mofetil

365

Requirement of UTP and GTP for Glycoproteln Synthesis

UDPGleNAc

UTP

9

GTP

9 Dol ( ~ ( ~ GIcNae

GDPMan

9 Ool (~(~) (GIoNac)2 Man

Protein

9

Protein Asn (GIcNac)2 (Man)9

Depletion of UTP and/or GTP inhibits glycoprotein synthesis

Fig. 10. The role of sugar nucleotides and dolichol phosphate in the transfer of N-acetylglucosamine and mannose to asparagine residues of membrane glycoproteins. Depletion of U T P and/ or G T P inhibits glycosylation

Human peripheral blood m o n o n u c l e a r cells: Galanthus nivalis agglutinin

Control

"21

It

El ,

Mycophenolic acid (1 ,um)

T~ b.~ ~"

lea

lea

UlSsCIqOOgt46GT'

["

~ ..... ,

:t-~,1i :

FLI ~ >

r

Fig. 11. Flow cytometric measurements of the binding to h u m a n peripheral blood lymphocytes of a labeled lectin [fluorochrome 1, Galanthus nivalis agglutinin selective for terminal, ~t(1-3)-linked mannose]. The lymphocytes were stimulated for 48 h with concanavalin A in the presence or absence of MPA added to the culture medium for the last 8 h. MPA markedly decreases mannose available on the cells for lectin binding [10]

250 -

200

-

150 -

100 -

50-

O0

10

w i t h o u t IL-1 Endothelial cells ~ ~ IL-1 4 h

100 pm MPA ~--

Fig. 12. Binding of T cells to hum a n umbilical vein cells stimulated by IL-lx (100 ng/ml) is decreased when either the T cells or the endothelial cells are treated with MPA. The inhibition is strongest when both cell types are treated [10]. T cells: [] no MPA; [] 1 IxM MPA; [] 10 I.tM MPA

366

A.C. Allison and E. M. Eugui

ed to have similar effects to those achieved with MPA, and the two drugs should have at least additive effects on glycosylation.

Other possible effects of GTP depletion Other effects of GTP depletion can be postulated, although they have not been established. There are many G-proteins, which vary in their affinity for GTP and GDP and which regulate numerous cellular transduction systems. Some of these may be very sensitive to MPA-mediated GTP and GDP depletion. Inhibition of antibody formation, for example, might be explained by such a mechanism. GTP is also an intermediate in the production of neopterin, which is specific for monocytes activated with interferon-y or in other ways [91]. Pterins are co-factors in the synthesis of nitric oxide, a vasodilator, and are involved in macrophagemediated cell damage in some species of animals.

Effects on proliferation and differentiation of cells of the monocyte-macrophage lineage Cytokines produced by activated macrophages, especially IL-lfl and tumor necrosis factor (TNF)-~ contribute to the pathogenesis of inflammation in several ways [5]. The cytokines induce the production by endothelial cells of adhesion molecules such as ICAM-1 and VCAM-1, which recruit leukocytes. The cytokines also induce in fibroblasts, endothelial cells and other cell types the production of prostaglandins E 2 and I2, which are vasodilators and co-factors in the pathogenesis of edema. Three naturally occurring mechanisms are known to regulate the production and effects of IL-lfl. When IL-1 and TNF-0t circulate, as in septicaemia, they initiate a cascade leading to the release of glucocorticoids, which inhibit IL-lct, IL-lfl and TNF-~ synthesis [5]. A protein binding IL-1 receptors and antagonizing IL-1 activity has been identified; as monocytes differentiate into macrophages they produce less IL-lfl and more IL-1 receptor antagonist, IL-lra [81]. It follows that any drug inhibiting the proliferation of cells of the monocyte-macrophage lineage and accelerating differentiation could have anti-inflammatory activity. Several long-acting anti-rheumatic drugs may exert their effects by this mechanism [6]. MPA, in therapeutically attainable doses (1 IxM), is a potent inhibitor of the proliferation of monocytic lineage cells and an inducer of differentiation, as shown by the acquisition of surface markers (Fc~, receptors, C3 receptors), the production of lysozyme and lysosomal enzymes and expression of IL-lra [6]. This mechanism may decrease IL-1 and IL-6 production and thereby contribute to the efficacy of MM in patients with rheumatoid arthritis. Since activated macrophages are present in allografts during rejection episodes and in chronic rejection, suppression of cytokine production by these cells may also contribute to the efficacy of MM for the prevention and treatment of graft rejection.

Mycophenolate mofetil

367

Lack of effect on neutrophil chemotoxis, superoxide production and microbicidal capacity One of the remarkable features of M M is that even chronic treatment of experimental animals and humans with immunosuppressive doses does not increase susceptibility to infections. We, therefore, ascertained whether clinically attainable concentrations of MPA affect responses of human neutrophils to chemotaxis and their capacity to produce superoxide and kill bacteria [4]. Concentrations o f MPA up to 10 IxM had no demonstrable effect on any of these responses. One reason for lack o f an effect in this system is the fact that 10 ~tM MPA does not deplete G T P in neutrophils (Fig. 7), in contrast to lymphocytes and monocytes.

Effects of MPA in experimental animals

Lymphocyte-selective, reversible antiproliferative effects To ascertain whether the antiproliferative effects o f MPA and M M are also lymphocyte selective in vivo, we injected mice subcutaneously with an antigen (ovalbumin) in adjuvant, which stimulates D N A synthesis in lymph nodes of the drainage chain: while a secondary response to antigen was in progress, the mice were injected intraperitoneally with [3H]dThd. One group of mice was given MPA (100 mg/kg per day orally) while a control group received vehicle. Incorporation of [3H]dThd into D N A was measured in draining lymph nodes, spleen and testis [28]. As shown in Table 4, the orally administered drug strongly inhibited D N A synthesis in the lymph nodes, but had no detectable effect on D N A synthesis in the germinal cells o f the testis or the basal epithelial cells of the small intestine (demonstrated by autoradiography). Effects on the spleen were intermediate, in keeping with the presence in the mouse spleen of hematopoeitic cells as well as lymphocytes. Doses o f M M required to prevent allograft rejection did not affect the production of neutrophils or platelets in any species. In the rat a reversible hypoplastic anemia was observed [65], but this has not been found in any other animal species. These observations show that the antiproliferative effects o f MPA and M M are more potent on lymphocytes than on other cell types in vivo. The cytostatic effects of MPA on lymphocytes are rapidly and fully reversible: peripheral blood mononuclear cells separated from the plasma of patients treated

Table 4. Effect of MPA on [3H]thymidine incorporation into DNA in different tissues [28] Tissue

Control (cpm/sample)

MPA-treated a (cpm/sample)

% Inhibition

P Value

Inguinal lymph b node Spleen b Testis c . .......

8765 2727 7435

528 1070 7043

93.9 60.7 5.3

0.000 0.019 0.726

a 50 m g / k g bid for 2 days b 107 cells c 10 m g tissue

368

A.C. Allison and E. M. Eugui

70 60

~, 5o

._1 o 40

~ 3o m

20 10

to 50

1 to 25

1 to 12.5

Ratio of Target to Effector Cells

Fig. 13. Dose-dependent inhibition of the generation of cytotoxic T lymphocytes in mice treated with MPA [28]

with M M respond normally to mitogenic stimulation, even when the plasma is strongly suppressive. Infections have not been a serious problem in patients treated with MM. However, if virus or other infections do occur, the drug can be withdrawn, and the immunosuppressive effects would be expected to disappear within a few days.

Inhibition of cell-mediated immune responses to allogeneic cells Cytotoxic T lymphocytes are important mediators of allograft rejection [80]. A classical model [17] was used to ascertain the effects of MPA and M M on cytotoxic T lymphocytic responses to allogeneic cells. Tumor cells (P815, H-2 d) were injected intraperitoneally into C57BL mice (H-2 b) to immunize them. Recipient mice were orally dosed with MPA, M M or vehicle and spleens were removed on day 10 or 11 for in vitro cytoxicity studies using 51Cr-labelled P815 target cells. This lysis is genetically restricted and largely due to cytotoxic T cells. As shown in Fig. 13, MPA inhibited in a dose-related fashion the induction of a cytotoxic T lymphocyte response to allogeneic cells [28]. Allogeneic tumor cells in the peritoneal cavity can be quantified and their viability assessed by capacity to grow as colonies in agar. In vehicle-treated animals the tumor cells were found to increase rapidly for the first 4 days and thereafter decrease as the host immune response became effective; by the 8th day no viable tumor cells could be recovered. In MPA-treated mice the tumor cells continued to increase and remain viable until the animals were killed for humane reasons [28]. This is a convenient model because cytotoxic T lymphocytes can be analyzed without complicating effects on the vasculature of allografts. These investigations show that M M and MPA inhibit the generation of cytotoxic T cells and the rejection o f allogeneic cells. Together with our studies on inhibition of antibody formation and leukocyte homing, they provided confirmation of the theoretical justification for use of the drug in organ transplantation.

Inhibition of antibody formation Antibody responses of rats and mice to sheep erythrocytes were analyzed by the Jerne plaque assay. Administration of M M inhibited the formation o f antibodies

Mycophenolate mofetil

369

60005000" 4000-

TT

3000

~,

2OOO"

1000

vehicle

3

9 RS-61443

30 mg/kg/day

Fig. 14. Dose-dependentinhibition of the number of antibodyforming cells in the spleens of rats immunized with sheep red blood cells and treated with MM [28]

in a dose-dependent manner [28]; oral administration 30 mg/kg per day to rats virtually abolished the formation of antibodies against xenogeneic cells (Fig. 14).

Prevention of allograft rejection Collaborations were established with transplantation immunologists to ascertain whether MM can prevent the rejection of tissue and organ allografts. Dr. Kevin Lafferty and his colleagues in the University of Colorado, Denver, studied pancreatic islet allografts in streptozoticin-treated diabetic mice and rats. BALB/c recipients of C57BL/6 islets were orally dosed with MM (80 mg/kg per day) for 30 days; this regime prevented graft rejection and following cessation of therapy allowed indefinite graft survival in 55% of mice. Animals carrying stable islet allografts exhibited a state of donor-specific tolerance, resistant to subsequent challenge with donor spleen cells in the absence of drug [44]. Combined treatment of graft recipients for the first 30 days with cyclosporin A increased the proportion of animals with long-term graft survival following cessation of treatment to 89%, but decreased the proportion of those surviving challenge with donor-strain spleen cells [45]. In this model the initial advantage of combined drug therapy was accompanied by the long-term disadvantage of less stable tolerance induction during the period of therapy. Following preliminary studies in the mouse, Dr. Randall Morris and his co-workers at Stanford investigated effects of MM in Lewis rat recipients of heterotopically grafted BN hearts. Monotherapy with MM (30-40 mg/kg per day) was found to prevent rejection of the grafts; if treatment was discontinued after 50 days, all of the hearts in the animals receiving the higher dose survived in good functional condition indefinitely [65]. When such animals were challenged, in the absence of further immunosuppression, with donor-strain atrial tissue beneath the renal capsule, it continued to beat indefinitely, whereas atrial tissue from a third-party strain (ACI) was promptly rejected [66]. Thus, shortterm treatment with MM could induce a state of donor-specific tolerance. In these animals mixed lymphocyte responses of recipient to donor-strain lymphocytes were significantly lower than in untreated recipients, whereas responses to thirdparty cells were normal.

370

A.C. Allison and E. M. Eugui

The second finding was that when MM and cyclosporin A were used together their effects were at least additive without any demonstrable increase in toxicity [65]. Since the drugs act by different mechanisms this was not surprising. The third finding was that if the first dose of MM was delayed until the 5th day following transplantation, by which time there is a marked mononuclear cell infiltrate into the graft and edema, the heart could still be preserved in good functional condition indefinitely [65]. When azathioprine or cyclosporin was given under comparable conditions it could not prevent rejection. This finding suggested that MM might be efficacious for the treatment of rejection crises. Some immunosuppressive drugs preventing allograft rejection in rodents have failed to do so in large animal models and man. Activity in dogs has proven to be a useful predictor of efficacy in humans. We were, therefore, pleased when studies in the laboratory of Dr. Hans Sollinger, University of Wisconsin, showed that MM was effective in the dog renal allograft model [75]. Monotherapy with MM (40 mg/kg per day) markedly prolonged graft survival, but combined therapy (MM 20 mg/kg per day, cyclosporin A 5 mg/kg per day and methylprednisolone 0.1 mg/kg per day) was more efficacious than either modality alone. There was no nephrotoxicity, hepatotoxicity or bone marrow suppression. The only serious toxicity in the dog was in the small intestine, and may be attributable to the large volume of concentrated bile secreted by the dog; efficient enterohepatic recycling of MPA glucuronide exposes the intestine to high a concentration of the drug. Such intestinal toxicity is not observed in any other species of animal. Later studies showed that MM can reverse acute renal allograft rejection in dogs [76]. While treatment with steroid bolus therapy could only temporarily halt rejection in some dogs, MM reversed rejection and prevented subsequent rejection episodes. These studies in experimental animals showed the efficacy of MM in preventing and treating allograft rejection without limiting toxicity or increased suceptibility to infections. They provided the experimental basis on which studies in human organ transplant recipients were undertaken.

Effects of M M in humans

The pharmacokinetic profile showed that twice daily oral administration of MM provided immunosuppressive doses of MPA throughout the day and night in several experimental animals. Extensive toxicology studies showed that the drug is not mutagenic, does not produce chromosome aberrations and does not have limiting toxicity even when given for long periods to non-human primates in higher doses than are required for human therapy. An open-label study was undertaken in patients with severe rheumatoid arthritis not responsive to cyclooxygenase inhibitors, methotrexate and long-acting anti-rheumatic drugs. Dose escalation studies showed that oral MM in doses up to 1.5 g twice daily was well tolerated, with no nephrotoxicity, hepatotoxicity, myelosuppression or other limiting side effects. The principal side effect is gastrointestinal intolerance, not associated with erosion. Continued treatment with 1 g to 1.5 g MM twice daily induced remissions in two-thirds of rheumatoid arthritis patients [83], which was maintained for more than I year (R. Goldblum, personal communication); infections were not limiting. Some clinical benefit was

Mycophenolate mofetil

371

observed within 2 months of treatment, but the condition of most patients improved with further treatment between the 3rd and 4th month, suggesting that immunosuppressive effects were being reinforced by long-acting anti-rheumatic activity, as predicted. A multi-center, placebo-controlled, double-blind trial of two doses of M M in patients with rheumatoid arthritis is in progress. Several studies of MM in human recipients of organ transplants are in progress [85]. The preliminary findings are encouraging: in a triple regime with cyclosporin and low-dose steroids in renal allograft recipients, M M seems to be superior to azathioprine in preventing early rejection episodes (17% as compared with 60%). MM can reverse ongoing rejection in the majority of renal, cardiac and hepatic allograft recipients even when high-dose steroids and OKT3 have been ineffective. Glucuronide formation The only major metabolite of MPA is the glucuronide (Fig. 3). By the action of a liver glucuronosyl transferase, glucuronic acid is added to the phenolic hydroxyl group of MPA. MPA glucuronide does not inhibit IMPDH. MPA glucuronide is secreted in the bile, especially in the dog, in which the majority of the drug is excreted in feces. In most species of animals, including man, the major execretory product of MPA is urinary glucuronide. MPA glucuronide absorbed in the small intestine can be reconverted by glucuronidase activity into MPA. MPA glucuronide orally administered to mice is as active in suppressing antibody formation as MPA itself (Eugui, unpublished). Enterohepatic recirculation of the drug probably occurs in man, but its quantitative importance in maintaining blood levels of the drug has not been established. The extent to which MPA glucuronide can be re-utilized to exert immunosuppressive activity is unknown. However, human lymphocyte activation by mitogenic or antigenic stimulation is accompanied by expansion of the lysosomal system, including increased/~-glucuronidase activity [3], and activated macrophages secrete/~-glucuronidase [72]. Hence, in sites of graft rejection and inflammation, such as synovia of patients with rheumatoid arthritis, MPA glucuronide is likely to be converted into active MPA. When tested on the promonocytic cell line THP-1, MPA glucuronide has cytostatic activity which cannot be explained by MPA contaminating the added glucuronide, and is presumably due to ~glucuronidase in the cell line. In human T lymphocytic cell lines MPA has much more potent cytostatic effects than the glucuronide, but residual cytostatic activity is likely to be due to release of MPA through ~-glucuronidase activity. Binding to albumin In patients undergoing renal dialysis MPA is not lost. The drug binds with relatively high affinity to human serum albumin: 3 mol MPA/mol HSA (J. Wu, personal communication, 1992). Adding HSA to MPA in culture reduces its capacity to inhibit proliferation of lymphocytes. The importance of protein binding in vivo is still unknown, but this may be one reason why high doses of MM are needed to achieve immunosuppressive activity in humans (1 g twice daily or

372

A.C. Allison and E. M. Eugui

more) when the drug is a potent inhibitor of the proliferation of cultured lymphocytes (IC5o about 100 nM).

Combination therapy with M M and brequinar Brequinar sodium (BQR), a substituted quinoline carboxylic acid, reversibly inhibits dihydroorotate dehydrogenase and thereby blocks de novo pyrimidine synthesis. It has potent antiproliferative effects on lymphocytes and prevents the rejection of heart, liver and kidney allografts in rats [20]. When BQR was administered to humans with cancer, limiting side effects included myelosuppression, especially thrombocytopenia, and mucositis. It would, therefore, be desirable to decrease the dose of BQR administered to human organ graft recipients while maintaining effective immunosuppression. For several reasons MPA and BQR would be expected to have additive effects. BQR inhibits formation of orotate, which is converted into orotidylate by transfer of a phosphoribosyl group from PRPP. MPA depletes PRPP as a secondary effect of depleting guanosine nucleotides (Fig. 1). If two drugs act on sequential steps of a pathway, the effects of the combination are synergistic: a well-known example is the antibacterial effect of trimethoprim sulfamethoxazole. Moreover, UTP is required for the transfer of N-acetylglucosamine to dolichol phosphate, while GTP is required for the next step in the synthesis of glycoproteins, addition ofmannose residues (Fig. 10). For these reasons MM and BQR would be expected to have at least additive effects on lymphocyte glycoprotein synthesis. A collaboration with groups in Madison, Wisconsin, and Kobe, Japan [53] showed that doses of MM and BQR which were well tolerated by rats when used together prevented cardiac allograft rejection. The combination was as effective for reversing ongoing rejection (starting therapy 5 days after transplantation) as high-dose MM alone. Reduced expression of adhesion molecules may contribute to the remarkable efficiency of the drug combination for treatment of ongoing rejection. Combination therapy (cyclosporin A, MM and BQR) prolonged cardiac xenograft survival (hamster to rat) to 37 days, compared with 4 - 6 days in animals receiving monotherapy.

Proliferative arteriopathy associated with chronic rejection Now that acute rejection can be reasonably well controlled, chronic rejection is emerging as the major limitation of long-term allograft survival [31, 73]. Chronic rejection is associated with a proliferative and obliterative arteriopathy, attributed to proliferation of smooth muscle cells and fibroblasts, first observed in small and medium-sized arteries and later throughout the arterial tree of transplanted hearts, kidneys and livers [30]. The lesions in graft recipients are usually generalized and the intimal thickening is concentric, whereas in the general population atherosclerotic lesions tend to be focal and asymmetric and foam cells are more prominent. The pathogenesis of proliferative arteriopathy in grafts is complex: some authorities believe that it is predominantly mediated by T lymphocytes [14], others that antibodies against donor antigens play a pathogenetic role [73]; prob-

Mycophenolate mofetil

40000

373

-- -- ~

CONTROL

@ 30000

20000 10000

i

0.01

I

T

T

0.1 1 10 DRUGCONCENTRATION(/.,'M)

Fig. 15. Effect of mycophenolic acid (MPA) and cyclosporinA (CsA) on the proliferationof human arterial smooth muscle cells in culture (J. Kowalskiand A.C. Allison, unpublished)

ably both humoral and cellular mechanisms are involved. Several growth factors could stimulate proliferation of smooth muscle cells in the neointima. For example, activated macrophages release IL-lfl, IL-6 and platelet-derived growth factor. Activated endothelial cells release basic fibroblast growth factor, IL-6 and endothelins. It is, therefore, a better therapeutic strategy to inhibit the end result (proliferation of arterial smooth muscle cells) than to attempt inhibition of the production or effects of individual cytokines. A major objective of transplantation immunologists is to define a therapy that prevents chronic as well as acute rejection. MPA inhibits not only T lymphocytic responses, and is a better inhibitor of antibody formation than cyclosporin but also, acting over a long period, reduces IL-1 production by macrophages. Perhaps more important, clinically attainable concentrations of MPA (1 to 10 IxM) inhibit the proliferation of human arterial smooth muscle cells in culture (Fig. 15). The dose-response curve is similar to that for fibroblasts and endothelial cells (Fig. 5). Clinically attainable concentrations of cyclosporin and of brequinar do not inhibit smooth muscle cell proliferation. The possibility that MM might be superior to currently used therapies for the prevention of chronic rejection was explored, using rat heterotropic heart allografts [65]. In this model moderate doses of cyclosporin A do not prevent arteriopathy, and low or moderate doses of FK 506 prevent mononuclear cell infiltration of the myocardium but not graft coronary disease. Only a highly toxic dose of azathioprine was able to prevent arteriopathy in grafts examined 1 months after transplantation. Even in long-term recipients of MM the incidence and severity of proliferative arteriopathy was low compared to that in recipients of other drugs [65]. A rat aortic allograft model was also studied [86]. Three months after aortic allografting, the intima showed marked proliferation, which was not seen when syngeneic aortas were grafted. MM significantly decreased proliferation, whereas cyclosporin and brequinar did not. Accelerated proliferative vasculopathy has also been observed in primate cardiac xenografts (cynomolgus hearts into baboon recipients) maintained with cyclosporin, azathioprine and methylprednisolone [63]. When MM was used instead of azathioprine in triple therapy, survival of

374

A.C. Allison and E. M. Eugui

cardiac xenografts was improved and the coronary arteries remained in good condition (R. P. McManus, unpublished, 1992). Thus, MM is so far unique among immunosuppressive drugs in clinical use for allograft recipients, in preventing arterial smooth muscle cell proliferation. Furthermore, the drug more effectively prevents proliferative arteriopathy in rat and primate allografts. What is found in rats and non-human primates may also be applicable to humans. If so, MM maintenance therapy may, by inhibiting proliferative arteriopathy, allow retention of allografts in good condition for long periods.

Risk for lymphoma development Another complication in allograft recipients is the development of lymphomas. This is observed in patients receiving cyclosporin and appears to be aggravated by the use of anti-lymphocyte antibodies. Two factors predisposing to lymphoma development have been identified. One is polyclonal proliferation of populations of B lymphocytes transformed by Epstein-Barr virus. Cyclosporin does not inhibit this process, but does inhibit the T lymphocyte-mediated response that normally restricts the outgrowth of EBV-transformed B cells [15, 43]. Secondarily, chromosome translocations lead to frank malignancy. A well-known example is the c-myc-Ig locus translocation [54], but recently other translocations leading to oncogene activation have been identified [57]. Induction by drugs of chromosome breaks increases the risk of such events. MPA in therapeutically attainable doses inhibits the proliferation of all the EBV-transformed and B lymphoma cell lines tested. Unlike azathioprine and mizoribine, MPA does not produce chromosome breaks. For these reasons the risk of lymphoma development in chronic recipients of MM should be low. If patients have lymphomas in the central nervous system when MM therapy is initiated, the drug may be ineffective in controlling them because of inadequate passage through the blood-brain carrier.

Differences between MM and other drugs inhibiting purine synthesis The most obvious comparisons are with azathioprine and with mizoribine, which has been used as an immunosuppressive in Japan. Both drugs inhibit purine synthesis, but the mechanisms involved are different. Azathioprine and 6-mercaptopurine (6-MP) are converted to the IMP analog 6-thioinosinic acid, which inhibits several enzymes of purine synthesis: IMPDH, PRPP-aminotransferase and adenylosuccinate synthetase [93]. In lymphocytes, 6-MP inhibits proliferation by depletion of adenosine rather than guanosine nucleotides; addition of adenosine restores lymphocyte proliferation [21]. An azathioprine metabolite is incorporated into DNA in the form of thioguanosine [25]. It is not surprising that azathioprine is mutagenic [46]. Inhibition by azathioprine is not reversed by either guanosine or adenosine, so that azathioprine metabolites are not selective inhibitors of I M P D H [21]. In head-to-head comparisons MM is obviously more efficacious than azathioprine even when the latter is used in maximal tolerated

Mycophenolate mofetil

375

doses, e.g., in treatment of ongoing rejection of cardiac allografts in rats [65], and in the prevention of allograft rejection in humans [85]. Mizoribine (bredinin) is an imidazole nucleoside, which when phosphorylated is a selective inhibitor or I M P D H [58]. Suppression of lymphocyte proliferation by mizoribine is reversed by guanosine except when concentrations of the drug are high, which suggests that there is reasonable selectivity for I M P D H [21]. However, the efficiency of phosphorylation of nucleosides can vary in different cell types [79], and the importance of inhibition of I M P D H in monocytes, endothelial cells and smooth muscle cells is obvious. Moreover, phosphorylated nucleosides can inhibit D N A repair mechanisms, which may explain why many chromosome breaks are observed in mizoribine-treated cells [82]. MPA does not require phosphorylation for activity, nor does it produce chromosome breaks. Thus, MPA has effects on the desired cell types and may have a better safety profile for long-term use than mizoribine. Future applications of MM The observations in experimental animal and in human recipients of MM are converging to provide a guide to probable early clinical applications of the drug.

Cyclosporin sparing The nephrotoxicity and other side effects of cyclosporin A are well known. While these may be reduced to some extent using the new cyclosporins, it is unlikely that they will be eliminated altogether. Any combination therapy that allows doses of cyclosporin to be lowered, while preventing allograft rejection, should decrease side effects. In several experimental animal models, including pancreatic islet allografts in the mouse [44, 45], cardiac allografts in the rat [65, 66] and kidney allografts in the dog [75], combined therapy with MM allows doses of cyclosporin to be reduced while preventing allograft rejection. In humans MM combined with cyclosporin and lowdose prednisone decreases the incidence of rejection to 17% in comparison with 60% in the absence of MM [85]. While it has not yet been formally shown that MM has cyclosporin-sparing effects, that is a likely prediction from the data now available.

Treatment of ongoing rejection One of the remarkable properties of MM is its capacity to reverse ongoing rejection. This was demonstrated in several experimental animal models, including cardiac allografts in the rat [65] and renal allografts in the dog [76]. MM has now been used for the treatment of refractory rejection in more than 150 patients, including recipients of kidneys, heart and liver aUografts in a multicenter study. The overall rate of reversion is around 70%. Phase III trials for this indication are ongoing.

376

A.C. Allison and E. M. Eugui

Maintenance therapy of organ graft recipients A long-term application o f M M w o u l d be for maintenance therapy o f allografts, possibly as m o n o t h e r a p y or in c o m b i n a t i o n with a low-dose glucocorticoid. M M has several advantages over cyclosporin for maintenance therapy o f o r g a n graft recipients. In experimental animals allograft rejection can be prevented using M M alone. This has been d e m o n s t r a t e d in several experimental animal models, e.g., pancreatic islet allografts in the m o u s e [44, 45] and cardiac allografts in rats [65, 66]. Liver allografts in three h u m a n s have been retained in g o o d functional c o n d i t i o n using M M with low-dose prednisone but no cyclosporin; in a f o u r t h patient a low m a i n t e n a n c e dose o f cyclosporin was also given [33]. The long-term benefits o f maintenance therapy using M M w i t h o u t cyclosporin, or with a low dose o f cyclosporin, are obvious. First is the reduced nephrotoxicity, hypertension and renal toxicity and other side effects o f cyclosporin. Second is the decreased risk o f proliferative arteriopathy associated with chronic rejection, where M M appears to have unique advantages. Third is the low risk o f l y m p h o m a development. A l t h o u g h the findings in experimental animals are strongly suggestive, careful examination o f patients treated with M M over several years will be required to establish that extrapolation to h u m a n s is valid.

Acknowledgements. Dr. Peter Nelson's chemical syntheses and the collaboration of our immunology staff (Dr. C. Muller, J. Kowalski, A. Mirkovich and S. Almquist) made the basic research program possible. Outside collaboration with Drs. H. Sollinger, K. Lafferty and R. Morris showed the efficacy of the drug in transplantation models. Syntex experts in toxicology, formulation, clinical studies and regulatory affairs were all essential for the later stages of the program. In particular, Dr. Robert Kauffman's supervision of the trials in human organ graft recipients is carrying the program to a successful conclusion.

References 1. Alarc6n GS, Tracy IC, Blackburn WD Jr (1989) Methotrexate in rheumatoid arthritis: toxic effects as the major factor in limiting long-term treatment. Arthritis Rheum 32:671 2. Albeda SM, Buck CA (1990) Integrins and other cell adhesion molecules. FASEB J 4:2868 3. Allison AC (1968) The role of lysosomes in the action of drugs and hormones. Adv Chemother 3:253 4. Allison AC, Ferrante A (1993) The role of G-proteins in human neutrophils: lack of effect of mycophenolic acid on chemotaxis, superoxide production and bacterial killing (in preparation) 5. Allison AC, Lee SW (1989) The mode of action of anti-rheumatic drugs. I. Anti-inflammatory and immunosuppressive effects of glucocorticoids. Prog Drug Res 33:63 6. Allison AC, Waters RV (1993) Long-acting antirheumatic drugs induce differentiation of monocyte lineage cells and change the balance of expression of IL-lfl and the IL-1 receptor antagonist. Agents and Actions (in press) 7. Allison AC, Hovi T, Watts RWE, Webster ADB (1975) Immunological observations on patients with the Lesch-Nyhan syndrome, and on the role of de novo purine synthesis in lymphocyte transformation. Lancet II: 1179 8. Allison AC, Hovi T, Watts RWE, Webster ADB (1977) The role of de novo purine synthesis in lymphocyte transformation. Ciba Found Symp 48:207

Mycophenolate mofetil

377

9. Allison AC, Almquist SJ, Muller CD, Eugui EM (1991) In vitro immunosuppressive effects of mycophenolic acid and an ester prodrug, RS-61443. Transplant Proc 23 [Suppl 2]: 10 10. Allison AC, Kowalski WJ, Muller CJ, Waters RV, Eugui EM (1993) Mycophenolic acid and brequinar, inhibitors of purine and pyrimidine synthesis, block the glycosylation of adhesion molecules. Transplant Proc (in press) 11. Altmann DM, Hogg N, Trowsdale J, Wilkinson D (1989) Cotransfection of ICAM-1 and HLA-DR reconstitutes human antigen-presenting function in mouse L cells. Nature 338:512 12. Baker GL, Kahl LE, Zee BC (1987) Malignancy following treatment of rheumatoid arthritis with cyclophosphamide: a long-term case-control follow-up study. Am J Med 83:1 13. Bennett WM, Pulliam JP (1983) Cyclosporine nephrotoxicity. Ann Intern Med 99:851 14. Billingham ME (1989) Graft coronary disease: the lesions and the patients. Transplant Proc 21:3665 15. Bird AG (1982) Cyclosporin A, lymphomata and Epstein-Barr virus. In: White DJ (ed) Cyclosporin A. Elsevier, Amsterdam, pp 307-315 16. Borel JF, Feurer C, Magn~e C, St~ihelin H (1977) Effects of the new anti-lymphocytic peptide cyclosporin A in animals. Immunology 32:1017 17. Brunner KT, Mauel J, Cerrotini JC, Chapuis B (1988) Quantitative assay of the lytic interaction of murine lymphoma cells on 51Cr-labelled allogeneic target cells in vitro; inhibition by isoantibody and drugs. J Immunol 14:181 18. Clarke L, Waxman DJ (1989) Oxidative metabolism of cyclophosphamide: identification of the hepatic monooxygenase catalysts of drug activation. Cancer Res 49:2344 19. Collier DStJ, Calne RY, Thiru S, Friend PJ, Lim SML, White DJG (1988) FRR900506 (FK 506) in experimental renal allografts in dogs and primates. Transplant Proc 20:226 20. Cramer DV, Chapman FA, Jaffee BD, Jones EA, Knoop M, Hreba-Elias G, Makowka L (1992) The effect of a new immunosuppressive drug, brequinar sodium, on heart, liver and kidney allograft rejection in the rat. Transplantation 53:303 21. Dayton JS, Turka LA, Thompson CB, Mitchell BS (1992) Comparison of the effects of mizoribine with those of azathioprine 6-mercaptopurine and mycophenolic acid on Tlymphocyte proliferation and purine ribonucleotide metabolism. Mol Pharmacol 41:671 22. De Neve W, Valeriote F, Edelstein M, Everett C, BischoffM (1989) In vitro DNA cross-linking by cyclophosphamide: comparison of human chronic lymphocytic leukaemia cells with mouse L1210 leukemia and normal bone marrow cells. Cancer Res 49:3452 23. Eckhoff DE, Landry AS, Eugui E, Allison AC, Sollinger HW, Hullett DA (1993) DAB4s 6IL-2 (IL-2 toxin) and RS-61443 allows prolongation of murine thyroid aUografts. Trans Am Assoc Transplant Surg (in press) 24. Elices M, Osborn L, Takada Y, Crause C, Luhowskyj S, Hemler M, Lobb RR (1990) VCAM-I on activated endothelium interacts with the leukocyte integrin VLA-4 at a site distinct from the fibronectin binding site. Cell 60:577 25. Elion GB (1989) The purine path to chemotherapy. Science 244:41 26. Epinette WW, Parker CM, Jones EL, Griest MC (1987) Mycophenolic acid for psoriasis. A review of pharmacology, long-term efficacy and safety. J Am Acad Dermatol 17:962 27. Eugui EM, Almquist S, Muller CD, Allison AC (1991) Lymphocyte-selective cytostatic and immunosuppressive effects ofmycophenolic acid in vitro: role of deoxyguanosine nucleotide depletion. Scand J Immunol 33:161 28. Eugui EM, Mirkovich A, Allison AC (1991) Lymphocyte-selective antiproliferative and immunosuppressive effects of mycophenolic acid in mice. Scand J Immunol 33:175 29. Fischer G, Wittmann-Liebold K, Lang T, Kiefhaber T, Schmid FX (1969) Cyclosporin and peptidyl-prolyl cis-trans isomerase are probably identical proteins. Nature 337:476 30. Foegh ML (1990) Chronic rejection-graft arteriosclerosis. Transplant Proc 22:119 31. Foster MC, Wenham PW, Rowe PA, Berden RP, Morgan AG, Cotton RE, Blarney RW (1989) The late results of renal transplantation and the importance of chronic rejection as a cause of graft loss. Ann R Coll Surg Engl 71:44 32. Franklin TJ, Cook JM (1969) The inhibition of nucleic acid synthesis by mycophenolic acid. Biochem J 113:515 33. Freise CE, Hebert M, Osorio RW, Nikolai B, Loke JR, Kauffman RS, Ascher NL, Roberts JP (1993) Maintenance immunosuppression with predinisone and RS-61443 alone following liver transplantation. Transplant Proc (in press)

378

A.C. Allison and E. M. Eugui

34. Furst DE, Koeluke R, Burmeister LF, Kohler J, Cargill I (1989) Increasing methotrexate response with increasing dose in the treatment of resistant rheumatoid arthritis. J Rheumatol 16:313 35. Garcia RC, Leoni P, Allison AC (1977) Control ofphosphoribosyl pyrophosphate synthesis in human lymphocytes. Biochem Biophys Res Commun 77:1067 36. Geoffroy JS, Rosen SD (1989) Demonstration that a lectin-like receptor (gp90~t~L)directly mediates adhesion of lymphocytes to high endothelial venules of lymph nodes. J Cell Biol 109:2463 37. Giblett ER, Anderson JE, Cohen F, Meuwissen HJ (1972) Adenosine deaminase deficiency in two patients with severely impaired cellular immunity. Lancet II: 1067 38. Reference deleted 39. Gomez EC, Menendez L, Frost P (1979) Efficacy of mycophenolic acid for the treatment of psoriasis. J Am Acad Dermatol 1:531 40. Grailer A, Nichols J, Hullett D, Sollinger HW, Burlingham WJ (1991) Inhibition of human B cell responses in vitro by RS-61443, cyclosporine A and FK506. Transplant Proc 23:314 41. Green CJ, Allison AC (1978) Extensive prolongation of rabbit kidney allograft survival after short-term cyclosporin A treatment. Lancet I: 1182 42. Green CJ, Allison AC, Precious S (1979) Induction of specific tolerance in rabbits by kidney allografting and a short period of cyclosporin A treatment. Lancet I: 1182 43. Gregory CD, Murray RJ, Edwards CF, Rickinson AD (1988) Down regulation of cell adhesion molecules LFA-3 and ICAM-1 in Epstein-Barr virus-positive Burkitt's lymphoma underlies tumor cell escape from virus-specific T cell surveillance. J Exp Med 167:1811 44. Hao L, Lafferty KJ, Allison AC, Eugui EM (1990) RS-61443 allows islet allografting and specific tolerance induction in adult mice. Transplant Proc 22:1659 45. Hao L, Calcinaro F, Lafferty K J, Allison AC, Eugui EM (1991) Tolerance induction in adult mice: cyclosporine inhibits RS-61443-induced tolerance. Transplant Proc 23 [Suppl 2]: 733 46. Herlia P, Murelli L, Scotti M, Paolini M, Cantelli-Fonti G (1988) Organ-specific activation of azathioprine in mice: role of liver metabolism in mutation induction. Carcinogenesis 9:1011 47. Hirata S, Matsuhara T, Saura R, Tateishi H, Hirohata K (1989) Inhibition of in vitro vascular endothelial cell proliferation and in vivo vascularization by low-dose methotrexate. Arthritis Rheum 32:1065 48. Hovi T, Allison AC, Allsop J (1975) Rapid increase of phosphoribosyl pyrophosphate concentration after mitogenic stimulation of lymphocytes. FEBS Lett 55:291 49. Hovi T, Smyth JF, Allison AC, Williams SC (1976) Role of adenosine deaminase in lymphocyte proliferation. Clin Exp Immunol 23:395 50. Hovi T, Allison AC, Raivio KO, Vaheri A (1977) Purine metabolism and control of cell proliferation. Ciba Found Symp 48:225 51. Isobe M, Yagita H, Okumura K, Ihara A (1992) Specific acceptance of cardiac allograft after treatment with antibodies to ICAM-1 and LFA-1. Science 255:1125 52. Jolivet J, Cowan KH, Curt GA, Clendeninn NJ, Chabner BA (1983) The pharmacology and clinical use of methotrexate. N Engl J Med 309:1094 53. Kuwamura T, Hullett DA, SoUinger HW (1993) Evaluation of RS-61443 and DUP-785 combination on ACI to LEW allogeneic cardiac transplant survival. Transplantation (in press) 54. Klein G, Klein E (1989) How one thing has led to another. Annu Rev Immunol 7:1 55. Konno Y, Natsumeda Y, Nagai M, Yamaji Y, Ohno S, Suzuki K, Weber G (1991) Expression of IMP dehydrogenase types I and II in Escherichia coli and distribution in normal human lymphocytes and leukemic cell lines. J Biol Chem 266:506 56. Kornfeld R, Kornfeld S (1980) Structure ofglycoproteins and their oligosaccharide units. In: Lennarz WJ (ed) Biochemistry glycoproteins and proteoglycans. Plenum press, New York, pp 1-37 57. Korsmeyer SS (1992) Chromosomal translocations in lymphoid malignancies reveal novel protooncogenes. Annu Rev Immunol 10:785 58. Koyama H, Tsuji M (1983) Genetic and biochemical studies on the activation and cytotoxic mechanism of bredinin, a potent inhibitor of purine biosynthesis in mammalian cells. Biochem Pharmacol 32:35

Mycophenolate mofetil

379

59. Laski LA (1992) Selectins: interpreters of cell-specific carbohydrate information during inflammation. Science 258:964 60. Lawrence MB, Springer TA (1991) Leukocytes roll on a selectin at physiologic flow rates: distinction from and prerequisits for adhesion through integrins. Cell 65:859 61. Lee WA, Gu L, Miksztal AR, Chu N, Leung K, Nelson PH (1990) Bioavailability improvement of mycophenolic acid through amino ester derivitization. Pharm Res 7:161 62. Lowe JB, Stoolman LM, Nair RP, Larsen RD, Berbend TL, Marks RM (1990) ELAM-1dependent cell adhesion to vascular endothelium determined by a transfected human fucosyltransferase cDNA. Cell 63:475 63. McManus RP, O'Hair DP, Hunter JB, Komrowski R (1992) Spectrum of vascular changes in primate cardiac xenografts. Transplant Proc 24:619 64. Mitsui A, Suzuki S (1969) Immunosuppressive effect of mycophenolic acid. J Antibiot (Tokyo) 22:358 65. Morris RE, Hoyt EG, Murphy MP, Eugui EM, Allison AC (1990) Mycophenolic acid morpholinoethylester (RS-61443) is a new immunosuppressant that prevents and halts heart allograft rejection by selective inhibition of T and B cell purine synthesis. Transplant Proc 22:1659 66. Morris RE, Wang J, Blum JR, Flavin T, Murphy MP, Almquist SJ, Chu N, Tam YL, Kaloostian M, Allison AC, Eugui EM (1991) Immunosuppressive effects of the morpholinoethyl ester of mycophenolic acid (RS-61443) in rat and nonhuman primate recipients of heart allografts. Transplant Proc 23 [Suppl 2]: 19 67. Mustelin T (1987) GTP dependence of the transduction of mitogenic signals through the T3 complex in T-lymphocytes indicates the involvement of G-protein. FEBS Lett 213:199 68. Nagai M, Natsumeda Y, Weber G (1992) Proliferation-linked regulation of the type II IMP dehydrogenase gene in human normal lymphocytes and HL-60 leukemic cells. Cancer Res 52:258 69. Natsumeda Y, Ohno S, Kawasaki H, Konno Y, Weber G, Suzuki K (1990) Two distinct cDNAs for human IMP debydrogenase. J Biol Chem 265:5292 70. Nelson PH, Eugui EM, Wang CC, Allison AC (1990) Synthesis and immunosuppressive activity of some side-chain variants of mycophenolic acid. J Med Chem 33:833 71. Oshugi Y, Suzuki S, Takagaki Y (1976) Antitumor and immunosuppressive effects of mycophenolic acid derivatives. Cancer Res 36:2923 72. Page RC, Davies P, Allison AC (1978) The macrophage as a secretory cell. Int Rev Cytol 52:119 73. Palmer DC, Tsai CC, Roodman ST, Codd JE, Miller LW, Safarian JE, Williams GA (1985) Heart graft atherosclerosis. Transplantation 39:385 74. Peters GJ, Oosterhof A, Verkamp JH (1982) Effects of adenosine and deoxyadenosine on PHA stimulation of lymphocytes of man, horse and pig. Int J Biochem 14:377 75. Platz KP, Sollinger HW, Hullett DA, Eckhoff DE, Eugui EM, Allison AC (1990) RS-61443, a new, potent immunosuppressive agent. Transplantation 51:27 76. Platz KP, Eckhoff D, Beckstein WO, Suzuki Y, Sollinger HW (1991) RS-61443 for reversal of acute rejection in canine renal allografts. Surgery 110:736 77. Prentice HG, Smyth JF, Ganeshaguru K, Wonke B, Bradstock KF, Janossy G, Goldstone AH, Hoffbrand AV (1980) Remission induction with adenosine deaminase inhibitor 2'-deoxycoformycin in Thy-lymphoblastic leukemia. Lancet II: 170 78. Rademacher TW, Parekh RB, Dwek RA (1988) Glycobiology. Annu Rev Immunol 57:785 79. Richman DD, Kornbluth RS, Carson PA (1987) Failure of dideoxynucleosides to inhibit HIV replication in cultured human macrophages. J Exp Med 166:1144 80. Rosenberg AS, Singer A (1992) Cellular basis of skin graft rejection: an in vivo model of immune-mediated tissue destruction. Annu Rev Immunol 10:333 81. Roux-Lambord P, Modoux C, Dayer JM (1989) Production of interleukin-1 and a specific inhibitor during monocyte-macrophage differentiation: influence of GM-CSF. Cytokine 1:45 82. Sakaguchi K, Tsujino M, Yoshizawa M (1975) Action of bredinin on mammalian cells. Cancer Res 35:1643 83. Schiff MH, Goldblum R, Rees MMC (1990) 2-Morpholino-ethyl mycophenolic acid (MEMPA) in the treatment of refractory rheumatoid arthritis (RA). Arthritis Rheum 33:s155

380

A. C. Allison and E. M. Eugui

84. Schreiber SL (1991) Chemistry and biology of the immunophilinsand their immunosuppressive ligands. Science 251:283 85. Sollinger HW, Deierhoi MH, Belzer FO, Diethelm A, Kauffman RS (1992) RS-61443 - a phase I clinical trial and pilot rescue study. Transplantation 53:428 86. Steele DM, Hullett DA, Bechstein WO, Kowalski J, Smith LS, Kennedy E, Allison AC, Sollinger HW (1993) Effects of immunosuppressive therapy on the rat aortic allograft model. Transplantation Proc 25(1):754 87. Stryer L (1988) Biosynthesis of nucleotides. In: Biochemistry, 3rd edn. W H. Freeman, New York, pp 602-626 88. Tiemeyer M, Swiedler SJ, Ishibara M, Merlland M, Schweingruber H, Hertzer P, Brandley BK (1991) Carbohydrate ligands for endothelial-leukocyte adhesion molecule. 1. Proc Natl Acad Sci USA 88:1138 89. Tutschka PJ, Beschorner WE, Allison AC, Burns WH, Santos GW (1979) Use ofcyclosporin A in allogeneic bone marrow transplantation in the rat. Nature 280:148 90. Verham R, Meek TD, Hedstrom L, Wang CC (1987) Purification, characterization and kinetic analysis of inosine-5'-monophosphate dehydrogenase of Trichornonas foetus. Mol Biochem Parasitol. 24:1 91. Wachter H, Fuchs D, Hausen A, Reibnegger G, Werner FR (1989) Neopterin as a marker for activation of cellular immunity: immunologic basis and clinical application. Adv Clin Chem 27:81 92. Watson SR, Fennie C, Lasky LA (1991) Neutrophil influx into an inflammatory site inhibited by a soluble bearing receptor-Ig chimaera. Nature 349:164 93. Wolberg G (1988) Antipurines and purine metabolism. Handb Exp Pharm 85:517