Suppression of inflammation and structural damage in experimental arthritis through molecular targeted therapy with PPI-2458

Authors


Abstract

Objective

To determine the disease-modifying activity and mechanism of action of the orally available methionine aminopeptidase type 2 inhibitor, [(1R)-1-carbamoyl-2-methyl-propyl]-carbamic acid-(3R,4S,5S,6R)-5-methoxy-4-[(2R,3R)-2-methyl-3-(3-methyl-but-2-enyl)-oxiranyl]-1-oxa-spiro [2.5] oct-6-yl ester (PPI-2458), in a rat model of peptidoglycan–polysaccharide (PG-PS)–induced arthritis.

Methods

Arthritis was induced in rats by administration of PG-PS, causing tarsal joint swelling and histopathologic changes characteristic of rheumatoid arthritis (RA). PPI-2458, a potent irreversible methionine aminopeptidase type 2 inhibitor, was administered orally every other day at 1, 5, or 10 mg/kg.

Results

In an in vitro osteoclastogenesis model, PPI-2458 potently inhibited osteoclast differentiation and bone resorption. In the rat PG-PS arthritis model, PPI-2458 afforded significant protection against established disease after therapeutic dosing. This in vivo activity of PPI-2458 was linked to the inhibition of methionine aminopeptidase type 2. Histopathologic assessment of affected joints showed improvement in processes of inflammation, bone resorption, and cartilage erosion, associated with significant improvement in all clinical indices. The protective effects of PPI-2458 against bone destruction in vivo, including the structural preservation of affected hind joints, correlated with improvements in bone histomorphometric markers, as determined by microfocal computed tomography and a significant decrease in systemic C-telopeptide of type I collagen, suggesting decreased osteoclast activity in vivo. Moreover, PPI-2458 prevented cartilage erosion as shown by a significant decrease in systemic cartilage oligomeric matrix protein.

Conclusion

The findings of this study suggest that PPI-2458 exerts disease-modifying activity in experimental arthritis through its direct inhibition of several pathophysiologic processes of this disease. These results provide a rationale for assessing the potential of PPI-2458 as a novel RA therapy.

Rheumatoid arthritis (RA), which affects ∼1% of the US population, is a systemic, chronic inflammatory autoimmune disease with a multifactorial, poorly defined etiology. The progression of RA is driven by an interdependent network of closely connected pathogenetic mechanisms, which ultimately lead to chronically inflamed joints and structural joint damage (1). The disease is mediated by pathologic neovascularization and the secretion of proinflammatory mediators, such as tumor necrosis factor α (TNFα), interleukin-1 (IL-1), and IL-17, from activated endothelial cells. Cells of the synovial intima (macrophage-like synoviocytes and fibroblast-like synoviocytes) and activated T cells, which are recruited to the inflamed joint, also contribute to synovial hyperplasia and the growth of the destructive pannus (2–4). The major structural damage in RA is associated with periarticular bone erosions and juxtaarticular osteoporosis (5). Bone erosion is mediated by osteoclasts, highly specialized multinucleated cells which are derived from hematopoietic precursors through a sequence of processes involving proliferation, differentiation, fusion, and activation (6, 7). In the inflamed joint, activated synoviocytes and T cells secrete RANKL, which, in synergy with TNFα, IL-1, and other osteoclastogenic mediators, stimulates pathologic bone resorption through osteoclast recruitment, differentiation, and activation (5, 7–10).

Over the past decade, treatment regimens for RA patients have shifted, and clinical strategies involving the “therapeutic pyramid” have been supplanted by a new paradigm which emphasizes the early use of disease-modifying antirheumatic drugs (DMARDs) (3, 11). A DMARD is used, either as monotherapy or in combination with other DMARDs, in patients with the potential for progressive disease (3, 11). Intensive translational research in RA has led to the identification and validation of novel molecular targets, such as TNFα, RANK, and RANKL, and the elucidation of their molecular mechanisms and functions in the pathology of this disease (7, 12). These studies have directly contributed to the successful clinical development of novel DMARDs, such as the anti-TNFα agents infliximab, adalimumab, and etanercept (13, 14). In a significant number of patients, however, RA has not responded or has become refractory to currently available DMARDs, or treatment has had to be interrupted due to intolerable side effects. Hence, the development of innovative therapies that target novel effectors and unique pathways in RA, and at the same time provide a safe alternative to existing DMARDs, remains a high priority in the management of this disease (11).

The fumagillin analog [(1R)-1-carbamoyl-2-methyl-propyl]-carbamic acid-(3R,4S,5S,6R)-5-methoxy-4-[(2R,3R)-2-methyl-3-(3-methyl-but-2-enyl)-oxiranyl]-1-oxa-spiro [2.5] oct-6-yl ester (PPI-2458) is an orally available, irreversible inhibitor of methionine aminopeptidase type 2, the molecular target of this class of molecules (15–19). Methionine aminopeptidase type 2 is a cotranslational regulator of protein synthesis involved in N-terminal protein processing, and recent functional studies have demonstrated that this enzyme is an important molecular regulator of mammalian cell growth (20–23). PPI-2458 is a potent inhibitor of the in vitro proliferation of activated human umbilical vein endothelial cells (HUVECs) and human fibroblast-like synoviocytes (FLS) from RA patients, both cell types that directly contribute to the pathogenesis of RA (18, 23). This in vitro growth inhibition was directly proportional to the amount of methionine aminopeptidase type 2 enzyme inhibited in these cells, which suggested that the inhibition of methionine aminopeptidase type 2 function is a critical step in the growth inhibition of PPI-2458–sensitive HUVECs and human FLS from RA patients (18). Moreover, these pharmacologic properties of PPI-2458 observed in vitro appear to translate into significant protection against disease in several rat models of arthritis (18, 24, 25).

The principal objective of this study was to show that the significant protection against disease afforded by PPI-2458 in an experimental model of arthritis was linked in vivo to inhibition of the molecular target, methionine aminopeptidase type 2. We sought to demonstrate that this disease-modifying activity was further associated with significant improvements in all pathophysiologic processes of inflammation and joint destruction, including protection against osteoclast-mediated bone resorption, preservation of joint architecture and integrity, and prevention of the cartilage erosion that occurs in this model of RA.

MATERIALS AND METHODS

Reagents.

PPI-2458 was synthesized at Praecis Pharmaceuticals. For in vitro studies, a 10-mM stock solution in ethanol was prepared. For in vivo administration, PPI-2458 was dissolved in 11% 2-hydroxypropyl-β-cyclodextran (Cargill, Minneapolis, MN). Peptidoglycan–polysaccharide (PG-PS) was obtained from Lee Biomolecular Laboratories (Grayson, GA), dexamethasone (DEX) (4 mg/ml in phosphate buffered saline [PBS]) from Henry Schein (Melville, NY), RANKL from R&D Systems (Minneapolis, MN), and E-64 from Sigma (St. Louis, MO).

Osteoclast differentiation assays.

Primary human osteoclast precursors (OCPs; Cambrex, Walkersville, MD) were seeded at 10,000 cells/well (50,000 cells/ml) in OCP growth medium (Cambrex). The cells were cultured for 7 days with either macrophage colony-stimulating factor (M-CSF; 33 ng/ml) alone, M-CSF (33 ng/ml) and RANKL (33 ng/ml), or with both cytokines and different concentrations of PPI-2458. Osteoclast differentiation was determined by staining for the osteoclast marker tartrate-resistant acid phosphatase (TRAP), using a leukocyte acid phosphatase kit. Briefly, after 7 days in culture, the cells were rinsed once with PBS, fixed with 37% formaldehyde in acetone-citrate buffer for 1 minute, and stained for development of red color, according to the recommendations of the manufacturer (Sigma).

Rat model of PG-PS–induced arthritis.

Female Lewis rats (7–8 weeks old) were obtained from Charles River (Wilmington, MA). PG-PS (25 mg/kg) was injected intraperitoneally on day 1, and responding animals were randomized into treatment groups on day 14. Vehicle, DEX (1 mg/kg), or PPI-2458 (1, 5, or 10 mg/kg) was administered orally, every other day. Paw swelling was monitored using a plethysmometer (Stoelting, Woodale, IL), according to instrument specifications. The volumes of the 2 hind paws were measured and averaged on days 1, 4, 6, 8, 10, 13, 15, 17, 20, 22, 23, 27, 29, and 31. Ten animals were assigned to each group, except the vehicle group (n = 4) and a group of animals that received 10 mg/kg PPI-2458 but no PG-PS (n = 4). All animal studies were approved by the Praecis Pharmaceuticals Institutional Animal Care and Use Committee.

Histologic assessment of PG-PS arthritis.

The histopathologic evaluation was performed on the left and right hind joints of randomly selected animals from each study group, by an independent histopathologist who had no knowledge of specific interventions. After completion of the treatment, the left and right hind ankles were removed, fixed in 10% buffered formalin, decalcified in 5% formic acid, embedded in paraffin, sectioned, and stained with hematoxylin and eosin for histologic evaluation. A joint histology scoring system, which grades the severity of 4 histopathologic processes (cell infiltration, pannus formation, cartilage erosion, and bone resorption), was used to quantify hind joint involvement (26). The total score was the sum of the scores assigned for each parameter (0 = normal, 1 = minimal, 2 = mild, 3 = moderate, 4 = marked); the maximum possible total score was 16 per ankle and 32 per animal.

Methionine aminopeptidase type 2 pharmacodynamic assay.

The methionine aminopeptidase type 2 assay measures the amount of uninhibited methionine aminopeptidase type 2 in cells or tissue that has not been derivatized by prior treatment with PPI-2458 (18, 23). Briefly, white blood cells (WBCs) from animals of each study group were pooled, and cell lysates were prepared as previously described (18, 23). WBC protein (10–20 μg) was incubated with a biotinylated analog of PPI-2458, and the biotinylated methionine aminopeptidase type 2–inhibitor complex was captured on a plate with immobilized streptavidin (Pierce, Rockford, IL). The complex was detected with the methionine aminopeptidase type 2 antibody CM33 (0.5 μg/ml) (Zymed, South San Francisco, CA), followed by horseradish peroxidase–conjugated goat anti-rabbit IgG secondary antibody (Amersham, Pittsburgh, PA). The amount of uninhibited methionine aminopeptidase type 2 was determined by measuring absorption at 450 nm using a Multiskan plate spectrophotometer (Labsystems, Helsinki, Finland). The detection limit of this assay was 0.47 ng methionine aminopeptidase type 2 protein/mg WBC protein.

Enzyme-linked immunosorbent assays (ELISAs) for cartilage and bone biochemical markers.

The amount of cartilage oligomeric matrix protein (COMP) in serum was measured with a competitive enzyme immunoassay, according to the recommendations of the manufacturer (MD BioSciences, St. Paul, MN). The detection limit of this ELISA is 0.2 units/liter. Helical peptide (amino acids 620–633) from the α1-chain of bone-specific human C-telopeptide of type I collagen was measured either in cell culture supernatants of primary human OCPs cultured on OsteoAssay plates (Cambrex) as described above, or in urine with a competitive enzyme immunoassay (Quidel, San Diego, CA). The detection limit of this ELISA is 8 μg/liter. All measurements of C-telopeptide of type I collagen in urine were corrected for urinary creatinine excretion for each sample to account for potential differences in renal clearance rates among the different study groups. Urinary creatinine was measured with a colorimetric assay (Quidel).

Microfocal computed tomography (micro-CT).

All specimens were scanned on an AG μCT 40 system (Scanco Medical,Wayne, PA). Images were obtained with an isotropic voxel resolution of 20μ. A matrix size of 1,024 × 1,024 with 1,000 projections was used for all scans. A total of 1,836 slices were scanned for each specimen (the number of slices scanned was determined by the length of the scan needed to cover the entire ankle including the distal tibia). The scan time per specimen was ∼2.6 hours. The images were then volume rendered using 2 different fixed thresholds, 255 and 140. The total bone volume and bone mineral density were calculated over the same regions of all specimens. Micro-CT was performed at Scanco.

RESULTS

PPI-2458–induced inhibition of osteoclast differentiation and bone resorption in vitro.

We developed an in vitro osteoclastogenesis model to examine the effects of PPI-2458 on osteoclast differentiation and bone resorption, based on the ability of primary human OCPs to recapitulate essential aspects of osteoclast differentiation and activity in vitro. Primary human OCPs were cultured for 7 days in the presence of M-CSF and RANKL, and treated with either vehicle or increasing concentrations of PPI-2458. Cells cultured with only M-CSF and RANKL differentiated into large, multinucleated osteoclasts, as demonstrated by the appearance of numerous TRAP-stained cells. Cells cultured with M-CSF, RANKL, and PPI-2458 at a concentration of 1 nM yielded almost no TRAP-positive multinucleated osteoclasts, while PPI-2458 at a concentration of 0.1 nM had little detectable effect on the ability of cells to differentiate into TRAP-positive, mature osteoclasts (Figure 1A). The ability of PPI-2458 to inhibit osteoclast differentiation was fully reversible (results not shown), consistent with previously reported findings in human FLS from RA patients (18), and was dependent on rates of metabolic turnover of the methionine aminopeptidase type 2 enzyme and normal turnover of affected cell types. Furthermore, incubation of these cells with PPI-2458 at concentrations of up to 100 nM did not induce cytotoxicity (data not shown).

Figure 1.

Inhibition, by PPI-2458, of osteoclast differentiation and bone resorption in vitro. A, Morphology of primary human osteoclast precursors (OCPs; 10,000 cells/well) cultured for 7 days with 33 ng/ml macrophage colony-stimulating factor (M-CSF) and 33 ng/ml RANKL, and vehicle or PPI-2458 at concentrations of 0.1 nM or 1 nM. Numerous multinucleated tartrate-resistant acid phosphatase (TRAP)–staining cells were seen with M-CSF/RANKL treatment. TRAP staining was inhibited by treatment with PPI-2458 at 1 nM, but not at 0.1 nM. Results are representative of 3 independent experiments (original magnification × 4). B, Amount of C-telopeptide of type I collagen (CTX-I) in cell culture supernatants. Primary human OCPs (10,000 cells/well) were seeded into wells coated with a thin layer of human bone particles (duplicate wells per experimental condition) and cultured with vehicle alone (first unlabeled bar), 33 ng/ml M-CSF and 33 ng/ml RANKL plus vehicle (second unlabeled bar), or 33 ng/ml M-CSF and 33 ng/ml RANKL plus 100 nM E-64 or PPI-2458 at concentrations of 0.01, 0.1, 1, or 10 nM. After 7 days, cell culture supernatants were collected and the amount of CTX-I was measured by enzyme-linked immunosorbent assay. The detection limit in this assay was 8 μg/liter. Values are the mean and SEM and are representative of 2 experiments with similar results.

To determine whether the observed inhibition of osteoclast differentiation by PPI-2458 resulted in inhibition of bone resorption in vitro, we cultured primary human osteoclasts on a thin layer of human bone particles in the presence of M-CSF and RANKL, with vehicle or increasing concentrations of PPI-2458. The nonspecific cysteine proteinase inhibitor E-64 (100 nM), a known inhibitor of bone resorption in vitro, was used as a positive control. After 7 days, the culture supernatant was collected and the amount of bone-specific C-telopeptide of type I collagen was measured by ELISA. PPI-2458 potently inhibited the bone-resorbing activity of human osteoclasts in a dose-dependent manner (50% inhibition concentration ≤0.1 nM), and the degree of inhibition at 1 nM and 10 nM was comparable with the inhibitory activity of E-64 at 100 nM (Figure 1B).

Association of potent antiinflammatory activity of PPI-2458 with inhibition of methionine aminopeptidase type 2 function in the rat arthritis model.

The potent in vitro inhibition of PPI-2458—sensitive HUVECs and human FLS from RA patients has previously been shown to be directly proportional to the amount of methionine aminopeptidase type 2 enzyme inhibited in these cells (18). We investigated whether the therapeutic effects of PPI-2458 on established disease in the rat PG-PS arthritis model were linked in vivo to inhibition of the molecular target, methionine aminopeptidase type 2. The progression of disease in this model follows a biphasic mode, with an early acute, predominantly neutrophil-driven phase which persists to day 6 or 7, followed by a chronic, T cell–dependent phase (evident at approximately day 12) (27).

Therapeutic dosing of animals administered vehicle or PPI-2458 (1, 5, or 10 mg/kg orally, every other day) started on day 15, after the chronic destructive phase of the disease was established, and terminated on day 31. PPI-2458 at all 3 doses resulted in significant amelioration of joint swelling and inflammation, as measured by swelling of the hind paws, when compared with vehicle-treated animals (Figure 2A).

Figure 2.

In vivo antiinflammatory activity of PPI-2458 and methionine aminopeptidase type 2 inhibition. A, Paw swelling in rats administered vehicle, dexamethasone (DEX; 1 mg/kg), or PPI-2458 (1, 5, or 10 mg/kg). Treatment started on day 15, after the onset of chronic disease. Values are the mean ± SEM paw volume (n = 10 rats per group, except for naive rats treated with vehicle [n = 4]) and are representative of at least 6 independent experiments. ∗ = P < 0.0001 versus rats treated with peptidoglycan–polysaccharide (PG-PS) and vehicle, by one-way analysis of variance followed by Dunnett's multiple comparison test. B, Amount of methionine aminopeptidase type 2 inhibited in white blood cell (WBC) lysates, determined by methionine aminopeptidase type 2 pharmacodynamic assay at the completion of the treatment. Values are the mean percent amount of methionine aminopeptidase type 2 compared with that in vehicle-treated rats with PG-PS–induced arthritis (which corresponded to 41 ng methionine aminopeptidase type 2/mg of WBC protein, arbitrarily set at 100% for comparison across groups) (n = 10 rats per group, except for naive rats treated with vehicle [n = 4]).

We measured the amount of uninhibited methionine aminopeptidase type 2 in WBCs of animals from all study groups at the conclusion of the treatment protocol, using the methionine aminopeptidase type 2 pharmacodynamic assay (18, 23). In animals administered PPI-2458 at a dose of 1 mg/kg, ≥60% of the methionine aminopeptidase type 2 in WBCs was inhibited relative to the vehicle-treated PG-PS arthritis group, while ≥95% of methionine aminopeptidase type 2 was inhibited with PPI-2458 at doses of 5 and 10 mg/kg (Figure 2B). Notably, ≥90% methionine aminopeptidase type 2 inhibition was also observed after the administration of DEX. No methionine aminopeptidase type 2 inhibition was observed in naive animals treated with DEX for 12 days.

Decrease in the severity of indices of inflammation and joint destruction in rats with experimental arthritis treated with PPI-2458.

The PG-PS animal model of arthritis is characterized by aggressive synovitis, extensive pannus formation, cartilage degradation, and focal bone erosion. We investigated whether the protective activity of PPI-2458 was mediated through a decrease in the severity of all of these clinical indices, or whether the activity of PPI-2458 affected only specific pathogenetic processes. Therapeutic dosing with PPI-2458 at doses of 1, 5, and 10 mg/kg resulted in significant improvements in the total arthritis score, compared with vehicle-treated animals (mean ± SEM 7.00 ± 0.58, 7.60 ± 0.67, and 5.83 ± 0.71, respectively, versus 13.85 ± 0.75) (Table 1). Moreover, this significant improvement in the total arthritis score was reflected in significantly improved scores for all clinical indices, with the highest level of protection (∼80%) observed for cartilage erosion with PPI-2458 at a dose of 10 mg/kg (Table 1). These results demonstrate that protection against arthritis in this model was mediated through a significant decrease in all indices of inflammatory and destructive processes.

Table 1. Joint histology scores by treatment group*
Treatment groupCell infiltrationPannus formationCartilage erosionBone resorptionArthritis (total)
  • *

    Values are the mean ± SEM joint histology score on a 5-point scale, where 0 = normal, 1 = minimal, 2 = mild, 3 = moderate, and 4 = marked. For each group, the left and right hind ankles of 5 rats (n = 10 joints) were scored, except for the group of naive rats treated with vehicle (n = 4 joints). Results are representative of at least 3 independent experiments. DEX = dexamethasone.

  • P < 0.01 versus rats treated with peptidoglycan–polysaccharide (PG-PS) plus vehicle, by one-way analysis of variance followed by Dunnett's multiple comparison test.

  • P < 0.0001 versus rats treated with PG-PS plus vehicle, by one-way analysis of variance followed by Dunnett's multiple comparison test.

Vehicle alone00000
PG-PS plus vehicle3.45 ± 0.263.65 ± 0.163.15 ± 0.193.60 ± 0.1713.85 ± 0.75
PG-PS plus DEX (1 mg/kg)0.25 ± 0.080.55 ± 0.080.60 ± 0.140.95 ± 0.132.35 ± 0.23
PG-PS plus PPI-2458 (1 mg/kg)1.85 ± 0.152.00 ± 0.141.10 ± 0.122.05 ± 0.247.00 ± 0.58
PG-PS plus PPI-2458 (5 mg/kg)2.15 ± 0.182.40 ± 0.201.25 ± 0.181.80 ± 0.207.60 ± 0.67
PG-PS plus PPI-2458 (10 mg/kg)1.38 ± 0.311.72 ± 0.200.72 ± 0.142.00 ± 0.165.83 ± 0.71

Preservation of the structural integrity of affected joints by PPI-2458 treatment.

We further investigated the effect of PPI-2458 treatment on the structural preservation of hind joints in rats with established disease. Three-dimensional–rendered micro-CT imaging, which allows for the noninvasive visualization of pathologic joint changes, demonstrated major structural damage and degenerative changes in the joint architecture of vehicle-treated animals (Figure 3). Therapeutic dosing with PPI-2458 (10 mg/kg) resulted in significant protection against bone destruction and preserved the architecture of affected hind joints. Moreover, quantitative micro-CT analysis demonstrated that treatment with PPI-2458 led to preservation of bone volume and protection against loss of bone mineral density, 2 important indices of bone homeostasis, although the observed improvements were not statistically significant (Table 2).

Figure 3.

Three-dimensional (3-D)–rendered microfocal computed tomography (micro-CT) images of hind paws from representative rats from each of 4 study groups, showing preservation of joint architecture by PPI-2458. After completion of the treatment protocol, hind paws from rats were excised, and joints were visualized using 3-D–rendered micro-CT to detect pathologic bone changes. All images were obtained using a Scanco Medical AG μCT 40 system with an isotropic voxel resolution of 20μ. See Figure 2 for other definitions.

Table 2. Total bone volume and bone mineral density (BMD), by treatment group*
GroupBone volume, voxelsBMD, mg hydroxyapatite/cm3
  • *

    Values are the mean ± SEM. The n values are the number of joints examined. See Table 1 for other definitions.

Vehicle alone (n = 2)206.7 ± 8.931,161 ± 1.15
PG-PS plus vehicle (n = 3)162.6 ± 9.911,044 ± 17.88
PG-PS plus DEX (1 mg/kg) (n = 3)240.7 ± 12.681,125 ± 2.59
PG-PS plus PPI-2458 (10 mg/kg) (n = 5)213.8 ± 5.231,087 ± 17.44

PPI-2458–induced inhibition of bone resorption and cartilage erosion in vivo in the PG-PS arthritis model.

Bone resorption represents the best surrogate marker of joint destruction. To assess the activity of PPI-2458 in protection against ongoing in vivo bone resorption during the RA process in this disease model, we measured the amount of C-telopeptide of type I collagen in urine, normalized for urinary creatinine excretion, at the conclusion of the treatment. Therapeutic administration of PPI-2458 resulted in significantly decreased systemic urinary levels of C-telopeptide of type I collagen compared with vehicle-treated animals (P < 0.0001 at doses of 1 and 5 mg/kg, P < 0.00001 at 10 mg/kg) (Figure 4A). These results confirmed that PPI-2458 inhibited bone resorption in this model, consistent with the clinical assessment of this parameter (Table 1).

Figure 4.

Inhibition, by PPI-2458, of bone resorption and cartilage erosion in the rat model of PG-PS–induced arthritis. A, Levels of C-telopeptide of type I collagen (CTX-I) in urine, measured by enzyme-linked immunosorbent assay (ELISA) and corrected for urinary creatinine excretion. On the day rats were killed, urine was collected from naive rats treated with vehicle or with PPI-2458 (10 mg/kg), and from rats with PG-PS–induced arthritis treated with either vehicle, DEX (1 mg/kg), or PPI-2458 (1, 5, or 10 mg/kg). B, Amount of cartilage oligomeric matrix protein (COMP) in serum, measured by ELISA. At completion of the treatment, serum was prepared from peripheral blood of naive rats treated with vehicle or with PPI-2458 (10 mg/kg), and from rats with PG-PS–induced arthritis treated with vehicle, DEX (1 mg/kg), or PPI-2458 (1, 5, or 10 mg/kg). Values are the mean and SEM and are representative of 2 experiments with similar results (in both assays, n = 10 rats per group, except for naive rats treated with vehicle or PPI-2458 [n = 4 rats per group]). ∗ = P < 0.0001; ∗∗ = P < 0.00001, versus rats treated with PG-PS and vehicle, by one-way analysis of variance followed by Dunnett's multiple comparison test. See Figure 2 for other definitions.

COMP is a major component of the extracellular matrix of the musculoskeletal system that mediates chondrocyte attachment through interactions with integrins (28). The amount of COMP present in the serum of animals treated with vehicle or PPI-2458 was measured at the conclusion of the treatment. A significant decrease in systemic levels of this marker, even below the level of serum COMP measured in naive animals treated with vehicle, was detected after treatment with PPI-2458 at all doses (all P < 0.00001) (Figure 4B), consistent with the clinical assessment of protection against cartilage erosion (Table 1).

DISCUSSION

In the chronically inflamed joints in RA, synovial hyperplasia, along with local neovascularization, ultimately results in the destruction of articular bone and cartilage (2–4). Angiogenesis has been recognized as a major contributing factor in the pathologic process in RA, and the concept of antiangiogenesis has recently emerged as a new strategy for the treatment of disease states involving proliferative synovitis, particularly RA (29). Earlier studies of TNP-470, a member of the fumagillin class of irreversible methionine aminopeptidase type 2 inhibitors like PPI-2458, have shown that it provides protection against experimental arthritis, thus validating angiogenesis as a therapeutic target in RA (30–32). TNP-470 was originally developed as an anticancer agent, but its clinical development was discontinued, primarily due to dose-limiting toxicities to the central nervous system and an unfavorable pharmacokinetic profile (33–35).

PPI-2458 was designed to overcome the clinical deficiencies of TNP-470, while at the same time maintaining the potent antiangiogenic and antiproliferative activity of this class of molecules. A previous study of PPI-2458 demonstrated a significant reduction in adverse effects on the central nervous system, as compared with TNP-470, and showed that it was effective when administered orally (18). Moreover, PPI-2458 potently inhibited the in vitro proliferation of activated endothelial cells and synoviocytes, both cell types which directly contribute to the pathogenesis of RA, through a mechanism that was proportional to the amount of methionine aminopeptidase type 2 enzyme inhibited in these cells (18, 23). These pharmacologic properties of PPI-2458 suggest that therapeutic benefit in RA could be derived by targeting several inflammatory processes, including synovial hyperplasia and angiogenesis; such treatment could lead to a decrease in the production of proinflammatory mediators, an inhibition of leukocyte adhesion and migration through decreased endothelial cell surface area, and a diminished supply of nutrients to the rapidly proliferating and erosive synovium.

While most current treatment strategies in RA have targeted the inflammatory cascade, the discovery of the osteoprotegerin (OPG)/RANKL/RANK axis and its critical function in bone resorption has not only led to an improved understanding of the molecular mechanisms involved in bone resorption, but has also guided the clinical development of novel therapies, such as the anti-RANKL antibody AMG-162 (denosumab) and bisphosphonate drugs, which specifically target the pathologic bone loss in RA (36, 37). The destruction of articular bone in diseased joints, a hallmark of RA, is caused by excessive activity of bone-resorbing osteoclasts. Due to this critical role of osteoclasts in the pathologic process in RA, we investigated the activity of PPI-2458 on the differentiation of human OCPs and the ability of this agent to prevent bone resorption in an in vitro model of osteoclastogenesis. We demonstrated a dose-dependent inhibition of both processes. This further suggested that PPI-2458 may have direct beneficial effects on bone homeostasis in RA.

The potent in vitro growth inhibition of activated endothelial cells and synoviocytes through methionine aminopeptidase type 2 inhibition prompted us to further investigate whether the therapeutic effects of PPI-2458 on established disease in rat experimental arthritis were linked to methionine aminopeptidase type 2 inhibition in vivo. Therapeutic dosing with PPI-2458 administered orally at 1, 5, or 10 mg/kg significantly reversed joint swelling and inflammation in PG-PS—induced arthritis, and this activity of PPI-2458 was associated with the inhibition of methionine aminopeptidase type 2 in circulating WBCs. These results demonstrate the effectiveness of this orally administered agent and further suggest that the amount of methionine aminopeptidase type 2 enzyme inhibited in WBCs could serve as a pharmacodynamic marker of the biologic activity of PPI-2458 in vivo, which could potentially be useful in the clinical setting.

Notably, the amount of free methionine aminopeptidase type 2 in WBCs of PG-PS–treated rats was significantly elevated over that in naive animals, suggesting a mechanism of inducible methionine aminopeptidase type 2 expression in this compartment in inflammatory disease. Another unexpected finding was that the treatment of rats with PG-PS arthritis with DEX also decreased the amount of uninhibited methionine aminopeptidase type 2 by ≥90%, a reduction that was not observed in naive animals after treatment with DEX for 12 days with different treatment schedules. These results suggest a novel mechanism of glucocorticoid activity in protection against experimental arthritis, possibly linked to a decrease in free methionine aminopeptidase type 2 enzyme, similar to the activity of PPI-2458.

The development and progression of disease in the PG-PS arthritis model closely resemble the pathogenetic processes observed in humans with RA. We used an established scoring system to measure whether the observed protective activity of PPI-2458 decreased the severity of all indices of joint inflammation and joint destruction, or whether this agent selectively targeted specific pathogenetic processes. Therapeutic administration of PPI-2458, even at the lowest dose (1 mg/kg), not only significantly reduced the total arthritis score, but this broad protection also translated into significant decreases in the severity of all indices tested, suggesting that PPI-2458 exhibits disease-modifying activity in this animal model of arthritis.

The most critical issue in the treatment of RA is the prevention of disability due to joint destruction caused by excessive bone resorption and cartilage erosion. The recent elucidation of the OPG/RANKL/RANK axis has broadened our understanding of the molecular mechanisms of osteoclastogenesis and bone resorption (5–7). Since histologic assessment of arthritic joints from rats with PG-PS–induced arthritis treated with PPI-2458 demonstrated significant protection against bone resorption in vivo, we investigated further the potential mechanism by which PPI-2458 afforded this protection.

The systemic levels of biochemical markers generated during the destruction of bone in RA, such as C-telopeptide of type I collagen breakdown products released into the peripheral blood, have been established as useful markers for monitoring tissue involvement in the development and progression of this disease, and to assess therapeutic interventions (38, 39). Our results show that PPI-2458 significantly decreased the level of urinary C-telopeptide of type I collagen, consistent with our findings that PPI-2458 inhibited osteoclast-mediated bone resorption in vitro, and suggest that this marked effect on bone resorption in vivo is mediated through a direct effect on osteoclasts. Alternatively, the potent inhibition of inflammatory processes in vivo by PPI-2458 could be mediated through a mechanism involving the decrease of mediators of inflammation and osteoclastogenic mediators, such as TNFα and RANKL.

Since progressive destruction of articular joints is the radiographic hallmark of RA, we examined whether the observed in vivo protection against bone resorption afforded after treatment with PPI-2458 would correspond to changes in the architectural structure of arthritic joints. Micro-CT analysis of arthritic hind paws demonstrated a dramatic deterioration of joint integrity in vehicle-treated rats with PG-PS–induced arthritis, while therapeutic dosing with PPI-2458 prevented structural joint damage, as evidenced by the suppression of focal bone erosions at the joints and the maintenance of bone integrity. Moreover, the protection of joint integrity was further evidenced by improvements in bone morphometric indices, including the preservation of local bone volume in affected joints and a decreased loss of bone mineral density, a process commonly observed in rats with experimental arthritis (40, 41). These results suggest that the marked protection against bone destruction afforded by PPI-2458 in this model of arthritis was mediated through a direct effect on osteoclasts, or that these protective effects were primarily regulated through the modulation of inflammatory processes in arthritic joints, a pathogenetic environment known to promote the recruitment, differentiation, and activation of osteoclasts.

Novel antiresorptive agents that specifically target the activity of osteoclasts, such as the anti-RANKL antibody AMG-162 (denosumab) and drugs of the bisphosphonate class such as zoledronic acid, are currently the subjects of clinical trials for the treatment of RA (42, 43). Results from preclinical studies and clinical trials, however, have shown that these drugs provided minimal protection against chronic synovial inflammation ongoing in arthritic joints. Hence, it has been suggested that the clinical use of antiresorptive drugs in RA will require combination therapy with potent antiinflammatory DMARDs to target the chronic inflammatory state in this disease (42, 43). In contrast, the demonstrated disease-modifying activity of PPI-2458 on inflammatory and destructive processes in RA suggests the potential use of this orally available agent as a monotherapy.

Cartilage erosion is another important pathophysiologic process involved in the destruction of arthritic joints. Levels of circulating COMP have been validated for use as a biomarker of cartilage erosion in RA, and to assess the efficacy of therapeutic interventions (44, 45). Therapeutic administration of PPI-2458 at all doses significantly reduced detectable levels of serum COMP to below the baseline level of COMP measured in serum of naive animals treated with vehicle. These results demonstrated that PPI-2458 potently inhibited accelerated cartilage turnover and cartilage surface erosion in articular joints, consistent with the histopathologic assessment of this parameter.

The protective activity of PPI-2458 against cartilage erosion could potentially result from direct effects on chondrocytes. However, the above-suggested alternative mechanism to account for the marked protection against bone destruction in vivo, involving the potent inhibition of inflammatory processes, could also play an important role in the observed cartilage protection, by preventing the expression and activation of cartilage-degrading enzymes, such as matrix metalloproteinases. In contrast to PPI-2458, antiresorptive therapies that specifically target osteoclasts have shown only minor benefits for cartilage preservation, most likely through protection of the integrity of the subchondral bone, thus preventing cartilage erosion from below the bone (43, 46).

In summary, PPI-2458 exhibited disease-modifying activity in a rat model of arthritis, as demonstrated by the marked suppression of the chronic inflammatory and destructive processes observed in this model. This potent in vivo activity was linked to the inhibition of the molecular target of PPI-2458, methionine aminopeptidase type 2, in peripheral WBCs. Moreover, the marked improvements in the structural preservation of arthritic joints, through protection against bone resorption and cartilage erosion, observed after treatment with PPI-2458 suggest a novel mechanism of disease-modifying activity not previously associated with methionine aminopeptidase type 2 inhibition. These findings provide a rationale for assessing the therapeutic potential of PPI-2458 in RA in clinical trials.

AUTHOR CONTRIBUTIONS

Dr. Hannig had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Hannig, Bernier, Hoyt, Doyle, Westlin.

Acquisition of data. Hannig, Bernier, Hoyt, Doyle, Clark, Karp, Lorusso, Westlin.

Analysis and interpretation of data. Hannig, Bernier, Hoyt, Doyle, Clark, Karp, Lorusso, Westlin.

Manuscript preparation. Hannig, Hoyt, Westlin.

Statistical analysis. Doyle.

Acknowledgements

The authors thank Dr. Rasesh Kapadia (Scanco USA, Wayne, PA) for micro-CT analysis and excellent advice.

Ancillary