Prevention of Organ Allograft Rejection by a Specific Janus Kinase 3 Inhibitor

See allHide authors and affiliations

Science  31 Oct 2003:
Vol. 302, Issue 5646, pp. 875-878
DOI: 10.1126/science.1087061


Because of its requirement for signaling by multiple cytokines, Janus kinase 3 (JAK3) is an excellent target for clinical immunosuppression. We report the development of a specific, orally active inhibitor of JAK3, CP-690,550, that significantly prolongedsurvival in a murine model of heart transplantation and in cynomolgus monkeys receiving kidney transplants. CP-690,550 treatment was not associatedwith hypertension, hyperlipidemia, or lymphoproliferative disease. On the basis of these preclinical results, we believe JAK3 blockade by CP-690,550 has potential for therapeutically desirable immunosuppression in human organ transplantation andin other clinical settings.

In spite of numerous treatment options for organ transplant and autoimmune disease patients (1), there remains a need for effective and safe immunosuppressive agents. The most significant complications of drugs used for transplant patients include nephrotoxicity, neurotoxicity, new-onset posttransplant diabetes mellitus, hyperlipidemia, and hypertension. These side effects occur in part because the molecular targets for all currently used transplant drugs (cyclosporin A, tacrolimus, mycophenolate mofetil, and sirolimus) are ubiquitously expressed (2). In this respect, a molecular target restricted in expression to immune cells could provide immunosuppressive efficacy without the toxicity associated with current therapies. Cytokine receptors, which use the common gamma chain, or γc [interleukin (IL)-2, -4, -7, -9, -15, -21], are critical for the development and homeostasis of immune cells, and patients with mutations in γc suffer from severe combined immunodeficiency (SCID) (3, 4). These receptors all require the cytoplasmic tyrosine kinase JAK3 for signaling, and patients lacking expression of JAK3 also display a SCID phenotype (5, 6). On the basis of the critical but selective role for JAK3 kinase in lymphocyte biology, we searched for inhibitors of this enzyme as potential immunosuppressive therapy.

The Pfizer chemical library was screened for inhibitors of in vitro JAK3 kinase activity, providing the lead compound, CP-352,664. Extensive chemical modification led to CP-690,550 (7) (Fig. 1A). Although CP-690,550 was highly potent for JAK3 inhibition [enzyme inhibitory potency (8) of 1 nM], it was 20- to 100-fold less potent for JAK2 and JAK1, respectively (Table 1). Because JAK2 mediates signaling via many hematopoietic cytokines [e.g., erythropoietin, thrombopoietin, and colony-stimulating factor receptors (3)], potent JAK2 inhibition could result in anemia, thrombocytopenia, and leukopenia in vivo. In addition, CP-690,550 did not have potent activity against 30 other kinases [all median inhibitory concentration (IC50) > 3000 nM]. This included Lck, a key T lymphocyte–signaling molecule downstream of the T cell receptor (9) (Table 1).

Fig. 1.

(A) Structure of JAK3 inhibitors. (B) CP-690,550 inhibits IL-2 phosphorylation of JAK3 and STAT5. Duplicate samples of YT cells were stimulated with media (lane 1) or IL-2 (10 min, lanes 2 and 3) in the presence (+, lane 3) or absence (–, lane 2) of 30 ng/ml CP-690,550. Immunoprecipitated phospho-JAK3 and STAT5 were analyzed by phosphotyrosine antibodies. Parallel samples probed for protein indicated equivalent amounts of JAK3 and STAT5 in all lanes (5).

Table 1.

Potency of CP-690,550 in enzyme assays. Data expressed as the IC50 in nM. All enzymes are purified from insect cells or Escherichia coli.

Enzyme IC50 (nM)
JAK3 1
JAK2 20
JAK1 112
ROCK-II 3,400
Lck 3,870
IRK >10,000
Cdk5 >10,000
Cdk2 >10,000
EGFR >10,000
MAPK2/ERK2 >10,000
JNK1/SAPK1c >10,000
SAPK2a/p38 >10,000
SAPK2b/p38b2 >10,000
SAPK3/p38g >10,000
SAPK4/p38d >10,000
MAPKAP-K1b >10,000
MAPKAP-K2 >10,000
MSK1 >10,000
PRAK >10,000
PKA >10,000
PKCa >10,000
PDK1 >10,000
PKBa >10,000
SGK >10,000
p70S6K >10,000
GSK3b >10,000
AMPK >10,000
CHK1 >10,000
CK2 >10,000
PHK >10,000
CSK >10,000
CDK2/cyclin A >10,000
PI3-K >10,000

Cell-based assays were used to further compare drug effects on signaling via JAK3 (IL-2–induced proliferation of human T cell blasts) with that via JAK2 [granulocyte–macrophage–colony-stimulating factor (GM-CSF)–induced proliferation of HUO3 cells], as well as cellular systems dependent on other kinases. For example, fibroblast proliferation depends on numerous ubiquitous kinases, including the cyclin-dependent kinase family (10). CP-690,550 inhibited IL-2–induced proliferation with 30-fold greater potency than its effects on GM-CSF–induced proliferation (Table 2). Potencies observed for CP-690,550 contrast sharply with data for other compounds cited as being JAK3 inhibitors (1113). We found them all to be ∼1000-fold less potent than CP-690,550. Moreover, in contrast to the reported effects of WHI-P154, CP-690,550 had little effect on T cell receptor signaling and no effect on serum-induced fibroblast proliferation, whereas WHI-P154 was equipotent in these two cellular assays (table S1). CP-690,550 demonstrated potent inhibition in the mixed lymphocyte reaction using murine, monkey, or human cells (tables S2 and S3). Consistent with its mechanism of action, these cellular activities correlated with the ability of CP-690,550 to block IL-2–induced phosphorylation of JAK3 and one of its key substrates, STAT5 (Fig. 1B). Finally, we established a JAK3-dependent cellular assay in the presence of 40 mg/ml human serum albumin to predict exposures required for in vivo activity. The IC90 for CP-690,550 (90% inhibition of activity) in this assay was 160 nM (50 ng/ml) (fig. S1) and was established as an initial target exposure for in vivo efficacy.

Table 2.

Potencies of CP-352,664 and CP-690,550 in cell assays. Data expressed as the IC50 in nM. Human T cell blasts were generated after incubation with phytohaemaglutinnin, and proliferation maintained with addition of IL-2. The human myelomonocytic cell line HUO3 was maintained in culture with human GM-CSF. Effects on T cell receptor signaling were evaluated by stimulating Jurkat T cells with antibodies to CD3 and CD28 and measuring IL-2 production after 24 hours. Nonspecific effects on cellular proliferation were assessed with human foreskin fibroblasts (HFF) cultured in fetal bovine serum. NA, not assayed.

Cell Stimulus CP-352,664 CP-690,550
T cell IL-2 3,694 11
Jurkat αCD3/αCD28 NA 7,840
HFF Calf serum >10,000 >10,000

The in vivo efficacy of CP-690,550 was first studied in a murine model of heterotopic heart transplantation (DBA2 donor heart into C57/BL6 host) (14). Animals treated with vehicle alone rejected their allografts within 12 days (Fig. 2A). In contrast, dosing with CP-690,550 resulted in a dose-dependent increase in survival of transplanted hearts. Although the drug was administered for only 28 days (5), the two higher dose groups shown (n = 6 mice/group) had median survival times (MST) of > 60 days. On the basis of drug exposures measured in all groups, we calculated the EC50 (drug concentration in blood at which 50% of mice will maintain their graft for >28 days) to be ∼60 ng/ml. These data indicate that CP-690,550 can suppress a robust in vivo allogeneic response.

Fig. 2.

CP-690,550 prevents rejection in a murine heterotopic heart transplantation model and production of inflammatory mediators. (A) Efficacy of CP-690,550 in murine heterotopic heart-transplant model. Each line represents a group of six to nine mice. Numbers adjacent to eachsolid line represent the average concentration of drug, ± standard deviation, measured at end of 28-day osmotic pump dosing period. Dashed line, placebo (vehicle). Arrow, time of pump removal and dosing discontinuation. (B) Levels of transcripts in heart (H) and blood (B) at 7 days posttransplant from animals treated with vehicle or CP-690,550. Eachbar represents an individual mouse treated with CP-690,550. Data expressed as relative level of mRNA in drug-treated animals compared to average of all vehicle-treated animals. Value of 1 denotes no change between vehicle- and drug-treated animals. (C) Cellular infiltration into hearts on day 10 posttransplant. Graph depicts increase in cells in the myocardium over a naïve control heart as assessed by nuclear staining of formalin fixed heart sections. n = 3 for CP-690,550 and n = 6 for vehicle.

To explore the molecular mechanisms associated with prevention of allograft rejection, we transplanted a parallel group of animals and dosed them with vehicle (n = 3) or CP-690,550 (n = 4). On day 7 posttransplant, a time at which numerous host cells are known to have infiltrated the graft (15) (Fig. 2C), RNA was prepared from the peripheral blood and transplanted heart of each animal. Transcript levels of 27 genes induced in this model (16) were analyzed by quantitative polymerase chain reaction (PCR). These genes were also induced in purified lymphocytes by IL-2 and/or T cell receptor stimulation and blocked by CP-690,550 (7). Of these 27 genes, CP-690,550 most strongly inhibited transcript expression for Granzyme B, FasL, Rantes, Mig, and IP-10 (Fig. 2B). Although a component of the reduced gene expression in the heart may have been because of fewer cells infiltrating the graft, image analysis of parallel heart tissue taken at 10 days post-transplant indicated that there were still equivalent cellular infiltrates in hearts from CP-690,550–treated animals (∼67% of controls) (Fig. 2C). Many of these genes are also known to be induced by interferon-γ (IFN-γ), a prominent IL-2–induced gene (intracellular staining of IFN-γ in Cynomolgus lymphocytes stimulated with IL-2 is shown in fig. S2). These observations suggest that reductions in chemokines (17) and cytotoxic effector molecules are likely to be a component of the mechanisms by which CP-690,550 promotes allograft survival.

To further establish the immunosuppressive efficacy of CP-690,550, we next used a nonhuman primate (NHP, Macaca fascicularis) model of kidney transplantation (18). Drug vehicle (n = 4) or CP-690,550 (n = 12) was given by oral gavage twice daily, beginning at the time of transplantation. Animals were euthanized at study termination (day 90) or earlier in cases of complications or when graft rejection was manifest as impaired renal function. Drug doses were adjusted three times per week on the basis of measurements of drug trough blood levels (12 hours after the evening dose). Two general dose ranges for CP-690,550 were tested: a high dose targeting drug trough levels of 200 to 400 ng/ml and a lower dose targeting drug trough levels of 50 to 100 ng/ml. As shown in Fig. 3A, vehicle control animals rapidly rejected (MST 6 ± 1 days). CP-690,550 treatment significantly prolonged graft survival as compared to vehicle (MST of 62 ± 6 and 83 ± 6 days for the low- and high-dose groups, P < 0.001 for treated compared with vehicle). We have previously demonstrated with this model that cyclosporin A, the major immunosuppressive drug used in organ transplantation and dosed at clinically relevant drug exposures of 150 to 300 ng/ml trough, results in a MST of 39 days (19). Four of 12 animals dosed with CP-690,550 (two from each dose group) survived to study termination with normal renal function and mild rejection as determined by histopathology (20) (one animal with borderline, one animal with Banff Type IA, and two animals with Banff Type IB rejection). The remaining eight animals were euthanized before study termination for impaired renal function due to acute rejection (n = 5), acute rejection superimposed with polyomavirus nephritis (n = 2), or bleeding (n = 1, euthanized on day 69 for persistent incisional bleeding but with normal graft pathology). Results in the low-dose group confirmed that drug trough levels in the range of the JAK3 cellular IC90 (50 ng/ml) were adequate to prolong allograft survival (Fig. 3B). The doses required to maintain these troughs are shown in fig. S3.

Fig. 3.

CP-690,550 prevents rejection of allogeneic kidneys in NHPs (M. fascicularis). (A) Survival of NHP recipients of life-supporting kidney allografts treated with vehicle (dotted line) or CP-690,550 (low dose, thin line, n = 8; high dose, thick line, n = 4). (B) Drug troughs of transplanted monkeys. Eachline represents one of four animals surviving to study termination. High-dose animals, squares. Low-dose animals, triangles. CP-690,550 was given by oral gavage twice daily beginning at the time of transplantation. (C) Hemoglobin levels in transplanted cynomolgus monkeys. Eachline represents one of four animals surviving to study termination. High-dose animals, squares. Low-dose animals, triangles.

Although the importance of JAK3-dependent cytokines in lymphoid development is well established, the consequence of acutely antagonizing the effects of γc cytokines by inhibiting JAK3 in mature lymphocytes is less clear (21, 22). Decreases in total peripheral blood lymphocytes were noted in the transplant animals, although the variability in these measurements makes conclusions difficult (fig. S4). In naïve, nontransplanted monkeys, treatment with CP-690,550 for 4 weeks had no effects on CD3CD20+ B cells, CD3+CD4+ T cells, or CD3+CD8+ T cells (fig. S5, A to C). In contrast, there were modest decreases seen in CD3CD16+ natural killer (NK) cells (fig. S5D), consistent with a role for JAK3-dependent cytokines in the maintenance and development of NK cells (23). Longer term studies in normal animals are ongoing.

Although efficacy in this series compared favorably with that obtained with established immunosuppressive drugs, CP-690,550 treatment did not produce the dose-limiting side effects associated with current antirejection therapy in humans and NHPs. For all animals treated with CP-690,550, there were no metabolic abnormalities detected (increases in blood lipids or glucose levels) and no evidence of hypertension nor cases of posttransplant lymphoproliferative disease (7). CP-690,550 treatment, however, was associated with dose-related anemia, presumably related to a level of JAK2 inhibition. This was largely restricted to the four animals with the highest drug exposure, who experienced sustained declines in hemoglobin levels. In contrast, the eight animals with ∼fourfold lower drug exposure experienced only minor decreases in hemoglobin levels, with animals surviving 90 days recovering to baseline values (Fig. 3C).

The development of safe and effective inhibitors of tyrosine kinases gained validation with imatinib (Gleevec, Novartis, Basel, Switzerland), which dramatically illustrated the feasibility of this approach for chronic myelogenous leukemia (24). Although a number of studies have described other compounds that inhibit JAK3 (1113), none have shown efficacy in NHPs. More importantly, they do not exhibit the potency and selectivity of CP-690,550 (table S1). Our findings, however, establish that an orally available JAK3 inhibitor produces sufficient immune suppression by itself to prevent organ transplant rejection without inducing many of the side effects observed with current therapies. At a well-tolerated dose of CP-690,550 (Fig. 3, A to C), we have shown 3-month survival of kidney allografts in NHP at drug exposures consistent with in vitro cellular potency and complete recovery of hemoglobin levels. Ongoing further dose reductions in this model, as well as testing in combination with mechanistically distinct immunosuppressive drugs, will suggest an optimal therapeutic strategy for profiling in humans. Immunosuppression as demonstrated here with CP-690,550, without the adverse events observed with current transplant therapy, may represent a major advance in the prevention and treatment of transplant rejection and possibly other immunological conditions.

Supporting Online Material

Materials and Methods

Figs. S1 to S4

Tables S1 to S4

References and Notes

References and Notes

Stay Connected to Science

Navigate This Article