|Home | About | Journals | Submit | Contact Us | Français|
Dipeptidyl peptidase (DPP) IV activity and/or structure homologues (DASH) are serine proteases implicated in tumourigenesis. We previously found that a DASH protease, fibroblast activation protein (FAP), was involved in osteoclast-induced myeloma growth. Here we further demonstrated expression of various adhesion molecules in osteoclasts cultured alone or cocultured with myeloma cells, and tested the effects of DASH inhibitor, PT-100, on myeloma cell growth, bone disease, osteoclast differentiation and activity, and expression of adhesion molecules in osteoclasts. PT-100 had no direct effects on viability of myeloma cells or mature osteoclasts, but significantly reduced survival of myeloma cells cocultured with osteoclasts. Real-time PCR array for 85 adhesion molecules revealed upregulation of 17 genes in osteoclasts after coculture with myeloma cells. Treatment of myeloma/osteoclast cocultures with PT-100 significantly downregulated 18 of 85 tested genes in osteoclasts, some of which are known to play roles in tumourigenesis and osteoclastogenesis. PT-100 also inhibited osteoclast differentiation and subsequent pit formation. Resorption activity of mature osteoclasts and differentiation of osteoblasts were not affected by PT-100. In primary myelomatous severe combined immunodeficient (SCID)-hu mice PT-100 reduced osteoclast activity, bone resorption and tumour burden. These data demonstrated that DASH proteases are involved in myeloma bone disease and tumour growth.
The successful use of certain antimicroenvironmental therapeutic agents has clinically validated the important role of the microenvironment in the initiation and progression of various malignancies (Joyce, 2005), including multiple myeloma (MM), a clonal malignancy of terminally differentiated plasma cells that typically accumulate in the bone marrow (BM). It has been shown that novel agents, such as thalidomide and bortezomib, that simultaneously target myeloma cells and the BM microenvironment are exceptionally effective in clinically controlling MM (Singhal et al, 1999; Richardson et al, 2003).
Myeloma typically presents with osteolytic bone disease characterized by increased bone destruction that is not followed by evidence of new bone formation (Bataille et al, 1991). This uncoupling of bone remodelling induced by myeloma cells is a result of increased osteoclast activity and reduced osteoblast numbers in the bone area adjacent to tumour cells (Bataille et al, 1991; Roodman, 2004; Edwards et al, 2008).
The close proximity of myeloma cells to osteolytic lesions led to the notion that osteoclasts provide major cellular microenvironmental support for myeloma cells (Epstein & Yaccoby, 2003). We and others have demonstrated the ability of human primary osteoclasts to support long-term survival and proliferation of myeloma cells through physical contact (Abe et al, 2004; Yaccoby et al, 2004, 2006). Osteoclasts have been shown to activate survival signalling pathways in myeloma cells including the p44/p42 MAPK, STAT3, and PI3K/Akt pathways (Hecht et al, 2008) and decreasing JNK activation (Colla et al, 2007). In addition, growth factors and cytokines found in the MM microenvironment, such as interleukin (IL)-6, osteopontin (Abe et al, 2004; Yaccoby et al, 2004), BAFF, and APRIL, (Yaccoby et al, 2008) have also been implicated in osteoclast-induced myeloma cell survival. Despite these findings, this group of signalling molecules does not define the MM phenotype, suggesting that other factors play are involved in survival signalling between osteoclasts and myeloma cells.
To further elucidate the molecular consequences of myeloma cell interactions with osteoclasts, we initially examined global gene expression profiles of osteoclasts following their coculture with primary myeloma cells (Ge et al, 2006). One of the few commonly upregulated genes was fibroblast activation protein (FAP), which was previously implicated as a critical microenvironmental factor in epithelial malignancies (Cheng & Weiner, 2003; Huber et al, 2003).
FAP (also known as seprase) is a type-II integral-membrane glycoprotein that belongs to the serine protease family known as dipeptidyl peptidase-IV activity and/or structure homologs (DASH), which includes the well-studied dipeptidyl peptidase 4 (DPP4 or CD26), as well as DPP2, DPP8, and DPP9, that have been implicated in tumourigenesis (Gorrell et al, 2006). FAP exhibits dual DPP and gelatinase activities (Park et al, 1999). As post-prolyl peptdidase, FAP can cleave amino-terminal dipeptides from polypeptides with penultimate l-prolines, thereby modifying bioactive peptides and changing their functions (Niedermeyer et al, 1998). Recently it was shown that FAP also exists in a soluble, circulating form—α2-antiplasmin cleaving enzyme (APCE) (Lee et al, 2006)—and that α2-antiplasmin and collagen I, -III, and -IV are natural substrates of APCE and FAP (Christiansen et al, 2007; Aggarwal et al, 2008).
Our previous study demonstrated that FAP is not expressed by myeloma cells in myelomatous bones, but it is highly expressed by osteoclasts, osteogenic cells, fibrotic stroma, and some adipocytes and vascular endothelial cells (Ge et al, 2006). Moreover, inhibiting FAP activity in osteoclasts or mesenchymal stem cells (MSCs) by using small-interference (si)RNA reduced the survival of myeloma cells when cocultured with the supportive osteoclasts or MSCs (Ge et al, 2006). These data suggest that FAP is a potential target for MM therapy and that inhibiting FAP activity may interfere with interactions between tumour cells and their BM microenvironment.
In this study, the role of DASH proteases in MM pathogenesis was investigated using the dipeptide boronic-acid DASH inhibitor, PT-100 (l-valine–l-boroproline). PT-100 was screened in a panel of approximately 100 enzymes including intracellular serine proteases (e.g. various DPPs and cathepsins), metalloproteinases, kinases and phosphatases. The only enzymes that were inhibited by PT-100 belonged to the DPP family enzymes, indicating that PT-100 is a selective inhibitor of the DPP family of enzymes. PT-100 inhibits activity of FAP (Cheng et al, 2005) and DPP4 at low nanomolar concentrations. This compound can penetrate the cell membrane and, therefore, inhibit intracellular DPPs, such as DPP2, DPP8, DPP9, and PREP (prolyl endopeptidase), in addition to the cell-surface DPPs such as DPP4 and FAP (Coutts et al, 1996; Shreder et al, 2005).
In mice, PT-100 inhibits tumour growth by interfering with interactions between tumour cells and the microenvironment and by stimulating innate immunity (Adams et al, 2004). PT-100 also promotes hematopoiesis in mice (Jones et al, 2003). Studies suggest that both the anti-tumour and pro-hematopoiesis effects of PT-100 are mainly mediated by inhibition of FAP (Jones et al, 2003; Adams et al, 2004). Because our previous study demonstrated high FAP expression by bone cells in regions of focal myeloma growth (Ge et al, 2006), we also evaluated the effects of FAP and related DPPs in MM-induced bone disease; the possible role of DASH inhibitor in MM-induced osteolysis is demonstrated for the first time.
Myeloma cells were obtained from heparinized BM aspirates taken during scheduled clinic visits from 13 patients with active myeloma. Signed Institutional Review Board–approved informed consent forms from the patients are kept on record. At the time of sample collection, 11 of 13 patients had stage IIIa or IIIb myeloma and had not received prior therapy. Ten of these patients had bone disease, as assessed by standard x-ray radiographs and magnetic resonance imaging (MRI). The circulating level of β2 microglobulin (β2M) ranged between 2·9 and 54·1 mg/l. The BM samples were separated by density centrifugation by using Ficoll-Paque (specific gravity, 1·077 g/ml; Amersham Biosciences Corp., Piscataway, NJ, USA), and the proportion of myeloma plasma cells in the light-density cell fractions was determined by CD38/CD45 flow cytometry. For in vitro studies, myeloma cells were isolated from BM samples by CD138 immunomagnetic bead selection (Miltenyi-Biotec, Auburn, CA, USA) (Yaccoby et al, 2004).
SCID-hu mice were constructed as previously described (Yaccoby et al, 1998; Yaccoby & Epstein, 1999). For each of the five experiments performed, BM cells from patients with MM (3 × 106–10 × 106 cells containing more than 20% plasma cells) were diluted in 100 μl phosphate-buffered saline (PBS) and injected directly into the implanted human bone. Mice were periodically bled from the tail vein, and changes in levels of circulating human light chain immunoglobulin (hIg) of the M-protein isotype were used as an indicator of tumour growth.
Dipeptide boronic-acid DASH inhibitor l-valine-l-boroproline (PT-100) was provided by Point Therapeutics (now Dara Biosciences, Raleigh, NC, USA). For in vivo studies, the agent was diluted in saline. Myelomatous SCID-hu mice were orally administered 100 μl of either saline or PT-100 (10 μg/mouse), twice daily for 4 weeks. For in vitro experiments, PT-100 was diluted in culture medium at the indicated concentrations (0·1–100 μmol/l) and added twice daily.
Levels of human κ and λ light chains were determined by enzyme-linked immunosorbent assay (ELISA) as previously described (Yaccoby et al, 1998; Yaccoby & Epstein, 1999). At the end of each experiment, all samples were analysed in the same assay to preclude interassay variability.
Radiographs taken with an AXR Minishot-100 beryllium source instrument (Associated X-Ray Imaging Corp., Haverhill, MA, USA) used a 10-second exposure at 40 kV. Changes in bone mineral density (BMD) of the implanted bones were determined using PIXImus DEXA (GE Medical Systems LUNAR, Madison, WI, USA) (Yaccoby et al, 2006, 2007).
Human osteoclast precursors and bone-resorbing osteoclasts were prepared from peripheral blood mononuclear cells (PBMCs) from patients with MM and healthy donors, as previously described (Yaccoby et al, 2004). Signed Institutional Review Board–approved informed consent forms from all subjects are kept on record. Briefly, PBMCs were cultured at 2·5 × 106 cells/ml in osteoclast medium containing α minimum essential medium supplemented with 10% foetal bovine serum (FBS), 50 ng/ml receptor activator of NF-kappa B ligand (RANKL) (PeproTech, Rocky Hill, NJ, USA), 25 ng/ml macrophage colony stimulating factor (M-CSF) (PeproTech) and antibiotics for approximately 10 d, at which time they contained large multinucleate osteoclasts exhibiting bone resorption activity (Yaccoby et al, 2004).
Human osteoclast precursors were incubated in osteoclast medium in the absence and presence of PT-100 (1 μmol/l) for 10 d, at which time cells were fixed and stained for tartrate-resistant acid phosphatase (TRAP), and the numbers of multinucleated TRAP-positive osteoclasts were counted (Yaccoby et al, 2004). An Olympus BH2 microscope (Olympus, Melville, NY, USA) was used to obtain images with a SPOT2 digital camera (Diagnostic Instruments Inc., Sterling Heights, MI, USA). Adobe Photoshop 10 (Adobe Systems, San Jose, CA, USA) was used to process the images.
Osteoclast activity was assayed by enumerating resorption pits based on dentine resorption (Yaccoby et al, 2004). Briefly, PBMCs (2·5 × 106 cells/ml) were cultured on dentine slices (Immunodiagnostic Systems Inc., Fountain Hills, AZ, USA) in 96-well plates in osteoclast medium. For testing effects of treatment at early stage of osteoclast formation, dentine slices containing committed osteoclast precursors (approximately 3 d of culture with osteoclast medium) were treated with PT-100 (1 μmol/l) for 10 d. For testing the effect on mature osteoclast activity, bone resorbing multi-nucleated cells were established by culturing the osteoclast precursors on dentine slices for 7–10 more days. The cultures were then treated with PT-100 (1 μmol/l) for 4 d. Dentine slices were treated with 10% bleach solution for 5 min and washed in distilled water. Resorption pits were photographed with a Nikon eclipse 450 microscope (Yaccoby et al, 2004). The ratio of resorption area:total area was quantified by using OsteoMeasure XP (Osteometrics, Atlanta, GA, USA) (Bendre et al, 2003).
Osteoclast precursors were incubated in 96-well plates in osteoclast medium in the absence or presence of PT-100 (1 μmol/l) for 5 d. The amount of p38 and of phosphorylated p38 in osteoclast precursors was quantified with Fast Activated Cell-based ELISA (Active Motif, Rixensart, Belgium) according to the manufacturer's instructions. Chemiluminescence (determined by OD450) was normalized to the number of cells (determined by OD595 analysis of crystal violet staining). NF-κB activity was quantified using an ELISA-based kit (Active Motif) according to the manufacturer's instructions. Cytosolic and nuclear fractions were isolated from osteoclasts using the Nuclear/Cytosol Fractionation Kit (BioVision Research Products, Mountain View, CA, USA) and protein quantified by bicinchoninic acid (BCA)-assay (Pierce, Rockford, IL, USA). Nuclear extracts were incubated in 96-well plates coated with immobilized oligonucleotide that was specific for NF-κB consensus site (50-GGGACTTTCC-30). NF-κB binding to the target oligonucleotide was detected by incubation with primary antibody specific for p65 subunit and visualized with anti-IgG horseradish peroxidase conjugate and developing solution. Level of activity was quantified at 450 nm with a reference wavelength of 655 nm. Background binding was subtracted from the value obtained for binding to the consensus DNA sequence.
Mesenchymal stem cells (MSCs) were prepared as previously described (Ge et al, 2006; Yaccoby et al, 2006). Briefly, human foetal bone fragments (Advanced Bioscience Resources, Alameda, CA, USA) were cultured in MSC medium [DMEM-low glucose (LG) supplemented with 10% FBS and antibiotics]. Half of the medium was replaced every 4–6 d, and adherent cells were allowed to reach 80% confluence before being subcultured with trypsin–EDTA.
To culture differentiated osteoblasts, we incubated MSCs in osteoblast medium (DMEM-LG supplemented with 10% FBS, 100-nmol/l dexamethasone, 10-mmol/l β-glycerophosphate, and 0·05-mmol/l ascorbate) in the absence or presence of indicated doses of PT-100 for 10 d. Differentiation status was determined by measuring alkaline phosphatase (ALP) activity (see below).
ALP activity in osteoblast precursors was measured by using SensoLyte pNPP Alkaline Phosphatase Assay (AnaSpec, San Jose, CA, USA) according to the manufacturer's instructions. Briefly, MSCs cultured for 0, 3, 6, and 10 d in “osteoblast medium”, with or without PT-100, were incubated with 1× Lysis Buffer at 4°C for 10 min with agitation. Cell suspensions were further centrifuged at 2500 g for 10 min at 4°C. Supernatants were collected, and protein concentrations were quantified with BCA assays to assure the uniform protein contents among all samples. Ten microgram of protein extracts were transferred in duplicate to a 96-well plate; pNPP reaction mixture was added to each well, and plates were incubated for 30 min before reactions were stopped by adding stop solution. ALP activity in each sample was determined by measuring OD405 and was compared to an ALP standard curve.
Osteoclasts were washed three times with PBS to detach and remove nonadherent cells. To test the effect of PT-100 on osteoclast-induced myeloma cell survival, primary myeloma cells (0·5 × 106–1 × 106 cells/ml in osteoclast medium) from six patients were cultured with osteoclasts in 24-well plates (1 ml/well) with or without PT-100 (0·1, 1, 10, or 100 μmol/l) for 5–7 d. At the end of each experiment, myeloma cells were collected; cell survival was assessed by trypan blue staining, and overall myeloma cell growth was assessed by MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) assays.
In indicated experiments, myeloma cells (0·5 × 106 cells/ml) or mature osteoclasts were cultured alone in 24-well plates and treated with indicated doses of PT-100 for 72 h. The cells were then collected and subjected to MTT assays.
This PCR array constitutes 89 genes; 84 extracellular matrix and adhesion molecules (ADAMTS1, ADAMTS13, ADAMTS8, CD44, CDH1, CNTN1, COL11A1, COL12A1, COL14A1, COL15A1, COL16A1, COL1A1, COL4A2, COL5A1, COL6A1, COL6A2, COL7A1, COL8A1, VCAN, CTGF, CTNNA1, CTNNB1, CTNND1, CTNND2, ECM1, FN1, HAS1, ICAM1, ITGA1, ITGA2, ITGA3, ITGA4, ITGA5, ITGA6, ITGA7, ITGA8, ITGAL, ITGAM, ITGAV, ITGB1, ITGB2, ITGB3, ITGB4, ITGB5, KAL1, LAMA1, LAMA2, LAMA3, LAMB1, LAMB3, LAMC1, MMP1, MMP10, MMP11, MMP12, MMP13, MMP14, MMP15, MMP16, MMP2, MMP3, MMP7, MMP8, MMP9, NCAM1, PECAM1, SELE, SELL, SELP, SGCE, SPARC, SPG7, SPP1, TGFBI, THBS1, THBS2, THBS3, TIMP1, TIMP2, TIMP3, CLEC3B, TNC, VCAM1, VTN) and 5 house keeping genes (B2M, HPRT1, RPL13A, GAPDH, ACTB). Since alpha9 beta1 have been recently shown to play a critical role in osteoclast formation and function (Rao et al, 2006) we performed individual qRT-PCR for ITGA9 and included it in the overall analysis.
Osteoclasts in 6-well plates were treated overnight with vehicle or PT-100 (1 μmol/l, twice daily) and then cocultured with myeloma cells (0·6 × 106 cells/well) for two additional days in the presence of similar concentrations of vehicle or PT-100. The experiment was repeated using three different sources of osteoclasts. Osteoclasts were then washed three times with PBS to detach and remove nonadherent cells and scraped to extract RNA using the Qiagen Rneasy Minin Kit as described by the manufacturer (Qiagen Inc., Valencia, CA, USA). One microgram of RNA from each sample was reverse-transcribed with the RT2 First Strand Kit (SABiosciences Frederick, MD, USA) and Real-Time PCR array performed according to the manufacturer's instructions using ABI Prism 7300 sequence analyzer (Applied Biosystems, Foster City, CA). The comparative threshold cycle (CT) method was used to calculate the amplification fold using the RT2 Profiler PCR Array Data Analysis Software (SABiosciences Frederick). Paired t- test was then performed for statistical analysis (P values < 0·05 were considered significant for all statistical analyses). Normalizing data with all or each of the five tested house keeping genes yielded similar results.
Bone sections were histochemically stained for TRAP and immunohistochemically stained for osteocalcin as previously described (Yaccoby et al, 2002, 2007). For adhesion molecules and human λ light chain immunohistochemistry, osteoclasts were prepared in 4-well chamber slides (Nalge Nunc International, Naperville, IL, USA). Myeloma cells were cocultured with osteoclasts in the absence and presence of PT-100 for approximately 60 h. Following coculture, myeloma cells were removed and osteoclasts were washed three times with PBS and then fixed with Histochoice (Amresco, Solon, OH, USA) for 20 min. After peroxidase quenching with 3% hydrogen peroxide for 10 min, the slides were incubated with monoclonal antibodies against human control IgG (0·5 μg/ml; R&D Systems, Minneapolis, MN), VCAN, (4 μg/ml; Lifespan Biosciences, Seattle, WA, USA), ITGAL (1:200 dilution; Abcam Inc, Cambridge, MA, USA), CD44 (1:100 dilution; Cell Signaling Technology Inc., Danvers, MA, USA) or COL1A1 (2 μg/ml, Santa Cruz Biotechmology, Santa Cruz, CA, USA) for 60 min, or with anti-human λ light chain (1:500 dilution, Dako, Carpinteria, CA, USA) for 10 min. The assays were completed with the use of the immunoperoxidase kit from Vector Laboratories (Burlingame, CA, USA) and counterstaining with hematoxylin. An Olympus BH2 microscope (Olympus) was used to obtain images with a SPOT 2 digital camera (Diagnostic Instruments Inc.). Adobe Photoshop version 10 (Adobe Systems) was used to process the images.
Myeloma cells were cocultured with osteoclasts in the absence and presence of PT-100 for approximately 60 h. Following coculture, myeloma cells were removed and osteoclasts were washed three times with PBS and cell lysates were obtained (BioVision Research Products). Equal amounts of lysate (40 μg) were separated by electrophoresis on 10% sodium dodecyl sulphate–polyacrylamide gels, and Western blotting was carried out according to the Western Breeze chemiluminescent immunodetection protocol as described by the manufacturer (Invitrogen, Carlsbad, CA, USA). The following primary antibodies were used: MMP-10 (Abcam Inc.), CD44 and beta-actin (Cell Signaling, Boston, MA, USA).
It has been reported that osteoclasts enhance myeloma cell survival, primarily via a mechanism involving cell-to-cell contact (Abe et al, 2004; Yaccoby et al, 2004). We exploited the myeloma cell–osteoclast coculture system (Yaccoby et al, 2004) to examine the effect of PT-100 on primary myeloma cell survival and expression of adhesion molecules in the cocultured osteoclasts. Osteoclasts were plated in a 24-well plates with myeloma cells (0·5 × 106–1 × 106 cells/ml) from a single patient (a total of six patients' samples were used) for 5–7 d with or without increasing concentrations of PT-100 (0·1–100 μmol/l). We have previously demonstrated that MM cells do not firmly adhere to osteoclasts (Yaccoby et al, 2004). The purity of myeloma cells or osteoclasts recovered from cocultured was >95% (Fig 1). We also reported that apoptotic myeloma cells are often phagocytozed by osteoclasts and therefore, measurement of overall metabolic activity (e.g. MTT assay) and/or total number of viable myeloma cells better reflect overall survival of primary myeloma cells in this coculture system. As shown in Fig 2A and B, all tested concentrations of PT-100 significantly inhibited myeloma cell growth (43–72%, P < 0·002 vs. 1 μmol/l). PT-100 had no direct effect on myeloma cells cultured alone (Fig 2C), which is consistent with previous observations that myeloma cells do not express functional DPPs (Ge et al, 2006 and Shmuel Yaccoby (SY), unpublished observations) and suggests that the antimyeloma effects of this agent is entirely mediated through interference with the prevailing myeloma cell–osteoclast interactions.
The activities of DASH proteases in tumourigenesis have been closely implicated with activation of adhesion molecules, such as integrins and metalloproteinases (MMPs) (Kelly, 2005). We reasoned that, mechanistically, DASH proteases affect myeloma–osteoclast interaction through altering expression of adhesion molecules in osteoclasts. To provide mechanistic insight into the role of DASH proteases in osteoclast-induced myeloma cell survival we performed adhesion molecules quantitative PCR arrays on osteoclasts cultured alone and cocultured with myeloma cells, and on cocultured osteoclasts following treatment with PT-100.
A total of 85 extracellular matrix and adhesion molecules genes were analysed by qRT-PCR. Our study revealed that myeloma cells significantly induced upregulation of 17 adhesion molecule genes and resulted in downregulation of three adhesion molecule genes by >2-fold in cocultured osteoclasts. The most upregulated genes (>10-folds) were TIMP3, MMP3, ITGA8, MMP1, LAMB1, SELE and COL11A1 (Fig 3A). CD44, which has been previously implicated with MM pathogenesis (Caers et al, 2006) was also upregulated in cocultured osteoclasts but expression of its ligand, osteopontin (SPP1), was not altered in cocultured osteoclasts, probably due to high baseline expression of this factor in these cells.
Treatment with PT-100 resulted in downregulation of 18 adhesion molecules genes in osteoclasts cocultured with myeloma cells (Fig 3B); eight of 18 genes were upregulated in osteoclasts by myeloma cells (cocultured osteoclasts) including ITGA9, CD44, COL11A1, COL1A1, MMP10, SELE, TIMP3, and LAMA3. Genes whose expression was not altered in cocultured osteoclasts but was downregulated by PT-100 in these cells were MMP8, VCAN, FN1, ITGAL, ITGB5, KAL1, SPARC, COL14A1, ITGAM and ITGAV (Fig 3B).
To further confirm our qRT-PCR Array data we examined changes of selected adhesion molecules at the protein level. Osteoclasts cultured alone or recovered from coculture with MM cells following treatment with vehicle or PT-100 were immunohistochemically stained for VCAN, ITGAL, CD44 and COL1A1, or were protein extracted and the cell lysates tested for CD44 and MMP10 expression by Western Blot (Fig 4). Similar to changes in gene expression, while protein levels of CD44, COL1A1 and MMP10 were increased in osteoclasts following coculture with myeloma cells, levels of all the tested proteins were reduced by PT-100 in cocultured osteoclasts (Fig 4). Our results, therefore, indicate that in MM/osteoclast cocultures, PT-100 alters gene expression and protein levels of various adhesion molecules in osteoclasts.
As IL-6 has been previously suggested to be involved in osteoclast-induced myeloma cell growth we looked for the effect of PT-100 on IL-6 secretion in cocultures. Treating myeloma cell–osteoclast cocultures (n = 5) with the DASH inhibitor did not significantly change the secretion level of this cytokine and the levels in conditioned media of control- and PT-100-treated cocultures were 4·9 ± 1·0 and 4·1 ± 1·4 ng/ml, respectively.
To elucidate the role of DPPs on osteoclast differentiation, osteoclast precursors were cultured for 10 d in osteoclast medium in the absence and presence of PT-100 (0·1–1 μmol/l). Osteoclastogenesis was quantified by enumerating multinucleated TRAP-positive osteoclasts in the cultures. Numbers of mature, TRAP-positive osteoclasts were lower by 44% ± 4% (P < 0·05) and 70% ± 4% (P < 0·01) after treatment with 0·1 μmol/l and 1 μmol/l of PT-100, respectively (Fig 5A and B).
To determine whether PT-100 also alters the bone-resorption activity of osteoclasts, formation of resorption pits was assessed on differentiated or mature osteoclasts. In the first experiment, osteoclast precursors were cultured for 10 d on dentine slices in osteoclast medium in the absence and presence of PT-100 (1 μmol/l). In control vehicle-treated cultures, osteoclast-induced pit formation was 4·93 ± 1·26% of the total area. The area of pit formation was significantly smaller after treatment with PT-100 (P < 0·01 vs. control) (Fig 5C and D). In the second experiment, formation of mature osteoclasts on dentine slices was allowed prior to initiation of treatment with PT-100 (1 μmol/l). In this experiment, PT-100 had no significant effect on bone resorption activity of mature osteoclasts (Fig 5C). Furthermore, this compound had no effect on the survival of mature osteoclasts (Fig 5E). These data demonstrate that PT-100 dramatically and significantly reduces bone resorption activity through inhibition of osteoclastic differentiation.
To shed light on mechanisms associated with inhibition of osteoclast formation by PT-100 we investigated two major signalling pathways that are critical for the induction of osteoclastogenesis by RANKL; NF-κB and P38 MAPK (Li et al, 2003; Asagiri & Takayanagi, 2007). To test the effect of this compound on p38 activity, osteoclast precursors were incubated in osteoclast medium and treated for 5 d with varying concentrations of PT-100 (0·1, 1, 10 μmol/l), or vehicle. Levels of p38 and its phosphorylated form were analysed by ELISA. PT-100 significantly inhibited phosphorylation of p38 in a dose-dependent manner (Fig 6A). To test the effect on NF-κB activity, osteoclast precursors were similarly treated with vehicle, PT-100 (1 μmol/l) or the proteasome inhibitor, bortezomib (2·5 nmol/l), a clinical agent known to inhibit NF-κB activity. Whereas bortezomib significantly reduced NF-κB activity, PT-100 had no effect on NF-κB activity in osteoclast precursors (Fig 6B).
As FAP was shown to be expressed by MSCs and subpopulations of osteoblasts (Ge et al, 2006), we explored whether PT-100 has effects on osteoblast differentiation in vitro. MSCs were cultured in osteoblast medium with or without PT-100 (0·1, 1, 10 μmol/l) or vehicle. Cells were collected on days 3, 6, and 10 of culture and tested for ALP content. PT-100 had no effect on osteoblastogenesis; ALP contents were similarly increased in vehicle- and PT-100-treated groups from 4·8 ± 0·1 ng/μg proteins at day 0 to 24·2 ± 3·2 ng/μg proteins at day 10 of culture. Collectively, these data demonstrate that DASH inhibitor does not influence osteoblast differentiation but has profound effects on osteoclast formation.
We previously demonstrated that FAP is upregulated in myelomatous bones (Ge et al, 2006). In the present study, SCID-hu mice successfully engrafted with myeloma cells from five patients were used to study the effects of PT-100 on myeloma growth and bone disease. All these patients had myeloma bone disease as assessed by standard x-rays and MRI. As previously reported, myeloma growth in the SCID-hu system was restricted to the implanted bone and characterized by increased levels of hIg in mice sera (indicative of tumour growth) and by induction of severe osteolytic bone disease (Yaccoby et al, 1998; Yaccoby & Epstein, 1999). Upon establishment of MM growth, as indicated by hIg level of >10 μg/ml, five hosts engrafted with MM cells from five patients were treated with PT-100, and an additional five matching SCID-hu hosts served as control and were treated with saline for 4 weeks. PT-100 treatment did not induce detectable toxicity in any of the experiments. Whereas in saline-treated hosts, tumour burden increased in all experiments, hosts treated with PT-100 showed marked tumour burden reduction in one experiment, myeloma growth rate inhibition in three experiments, and no antitumour effect in one experiment. Overall, before initiation of treatment, hIg levels were similar among the two tested groups (94·9 and 80·8 μg/ml for saline- and PT-100-treated hosts, respectively; Fig 7A). At the end of the experiments, however, hIg levels were higher in the control group than in PT-100-treated group (420 ± 34 and 142 ± 30 μg/ml for saline- and PT-100-treated hosts, respectively; P < 0·05 for PT-100 vs. control; Fig 7A).
To analyse the effect of PT-100 treatment on myeloma-induced bone disease, BMD of the implanted myelomatous bones was compared before treatment initiation and at the end of each experiment. Overall, during the treatment period, BMD of bones implanted in control hosts decreased by 23·5 ± 4·7%, while it decreased by 3·36 ± 4·6% in PT-100-treated hosts (P < 0·02 vs. control) (Fig 7B). The decrease in bone loss following treatment with PT-100 was also visualized on x-ray radiographs (Fig 7C). We also looked for the effect of PT-100 treatment on BMD of the uninvolved murine femur of SCID-hu hosts. BMD of the murine femur was insignificantly reduced, by −0·5 ± 0·8% and −1·9 ± 0·9% from pretreatment levels in the control- and PT-100-treated hosts, suggesting that PT-100 had no effect on BMD of the uninvolved murine bones.
To analyse the effects of PT-100 on bone cells in vivo, histological sections of the myeloma-engrafted human bones were stained for TRAP and immunohistochemically stained for osteocalcin. The numbers of TRAP-positive, multinucleated osteoclasts and osteocalcin-expressing osteoblasts were enumerated, as previously described (Yaccoby et al, 1998; Yaccoby & Epstein, 1999). Treatment with PT-100 resulted in reduced numbers of multinucleated osteoclasts (P < 0·01) but no significant effects on the numbers of osteoblasts (Fig 7D). These results suggest that the preventive effects of PT-100 on myeloma-induced bone disease in vivo are primarily mediated through inhibition of osteoclastogenesis.
The overall aim of this study was to shed light on the roles of DASH proteases in MM pathogenesis using the specific DASH inhibitor, PT-100 and myeloma cells taken from patients with active MM and bone disease. PT-100 significantly reduced osteoclast-mediated survival of primary myeloma cells in vitro and inhibited MM tumour growth in vivo. This agent showed no evidence of direct cytotoxicity on myeloma cells or mature osteoclasts, suggesting that the attenuation of tumour cell growth was primarily mediated through interference with the interactions of myeloma cells with their immediate cellular BM environment. Indeed, treatment of MM/osteoclast cocultures with PT-100 resulted in downregulation of various adhesion molecules in osteoclasts, some of which reported to play roles in tumourigenesis and osteoclastogenesis. Furthermore, PT-100 inhibited osteoclast formation through suppression of p38 MAPK activity, and subsequently resulted in a significant reduction in bone resorption and MM-associated bone disease.
The anti-myeloma effect of PT-100 was consistent with previous studies demonstrating that FAP is a microenvironmental factor involved in progression and metastasis of epithelial tumours (Cheng et al, 2002; Adams et al, 2004; Huang et al, 2004). FAP overexpression in HEK293 human embryonic kidney cells promotes tumour growth in animal models, and antibodies that neutralize FAP's DPP activity attenuate the tumour growth rate (Cheng et al, 2002). FAP expression is also associated with increased microvessel density, suggesting a potential for the promotion of tumour growth due to increased angiogenesis (Huang et al, 2004). Oral administration of PT-100 slows the growth of syngeneic tumours derived from fibrosarcoma, lymphoma, melanoma, and mastocytoma cell lines (Adams et al, 2004). Phase I and II clinical trials of PT-100 have been performed in patients with stage IV melanoma, non-small-cell lung cancer, or metastatic colon cancer (Nemunaitis et al, 2006; Cunningham, 2007; Narra et al, 2007). In these clinical trials, optimal doses have not been attained, but FAP was partially inhibited and only minimal clinical responses were observed. These clinical studies indicate that inhibition of FAP and other DASH members may not be beneficial in all malignancies. However, such agents may be successfully used against disease stages where the tumour cells are highly dependent on their microenvironment for survival. Because activity of DASH proteins appears to be required for osteoclast formation and for osteoclast-induced myeloma cell survival, inhibition of these serine proteases may benefit unique subsets of patients with MM.
Although previous studies suggested that PT-100 exerts its anti-tumour effects mainly through inhibition of FAP activity (Jones et al, 2003; Adams et al, 2004), our study cannot exclude the possibility that this agent inhibits MM and its associated bone disease through inhibition of other DASH family members. Osteoclasts express high levels of DPP4 and detectable levels of DPP7, DPP8, DPP9, and PREP (unpublished observations), however, only FAP is significantly upregulated in osteoclasts after coculture with myeloma cells (Ge et al, 2006). Some groups have reported that DPP4 is involved in tumourigenesis (Havre et al, 2008), and that DPP8 and DPP9 play important roles in cell adhesion, migration, and apoptosis (Gorrell et al, 2006; Yu et al, 2006).
The molecular mechanisms by which DPPs interact and promote tumourigenesis are largely ambiguous. FAP may form heterodimeric complexes with DPP4 dimers, β1 integrin, and urokinase-type plasminogen activator receptor, which initiate signalling that leads to increased matrix degradation, angiogenesis, and tumour growth, and reduced immune function (Ghersi et al, 2002; Kelly, 2005). FAP, as well as its soluble form, APCE, have been shown to cleave particular collagens (Christiansen et al, 2007; Aggarwal et al, 2008) and to generate Asn-α2 antiplasmin, which efficiently slows clot lysis by plasmin (Lee et al, 2006). Whether these collagen fragments and microscopic fibrin clots in areas adjacent to malignant cells mediate tumour growth is a matter for future investigation.
Our findings, that MM-osteoclast interaction results in alteration in expression of certain adhesion molecules in cocultured osteoclasts and that PT-100 inhibits expression of adhesion molecules in these cells, may provide mechanistic clues for the role of DASH proteases in osteoclast formation and osteoclast-induced myeloma cell survival and growth. The most remarkable changes in cocultured versus cultured osteoclasts were observed in expression of group of metallopeptidases and their inhibitors (MMP1, MMP3, MMP7, MMP10, MMP11, TIMP1, TIMP3 and ADAMTS13), several types of collagens (COL11A1, COL1A1, COLA4A2), laminins (LAMA3, LAMB1) and other adhesion molecules including E-selectin (SELE), CD44 and ITGA9. Some of these factors may interact with other adhesion molecules on myeloma cells (e.g. integrins) to promote myeloma cell survival and invasion. For instance, laminin has been shown to act as a chemoattractant for myeloma cells (Vande et al, 2001) while inhibition of CD44 inhibited myeloma cell survival and metastasis (Caers et al, 2006). Eight of those upregulated genes including MMP10, TIMP3, COL11A1, COL1A1, CD44, SELE, LAMA3 and ITGA9 were significantly downregulated by PT-100 in cocultured osteoclasts. Ten additional factors were not upregulated by myeloma cells in cocultured osteoclasts but their expression was significantly reduced in these cells following treatment of MM/osteoclast cocultures with PT-100. Interestingly, PT-100 had no effect on expression of osteopontin (SPP1) in cocultured osteoclasts or on secretion level of IL-6 in these cocultures. Osteopontin and IL-6 play a partial role in osteoclast-induced myeloma cell survival and growth (Abe et al, 2004; Yaccoby et al, 2004) indicating that additional factors involve in this process. Our results suggest that in MM/osteoclast cocultures, downregulation of various adhesion molecules in cocultured osteoclasts by DASH proteases inhibitor reduces the ability of osteoclasts to support myeloma cell survival and growth.
DASH proteases have not been previously implicated in osteoclastogenesis. Downregulation of certain adhesion molecules in osteoclasts by PT-100 suggests that activity of these proteases impacts osteoclast formation. Certain integrins such as alphaVbeta3 and alpha9 beta1 have been shown to play critical role in osteoclast formation and function (Rao et al, 2006; Nakamura et al, 2007). PT-100 downregulated expression of the integrins ITGA9, ITGAV, ITGB5, ITGAM and ITGAL in osteoclasts by >2-folds (Fig 2). Other integrins such as ITGB1 and ITGTB3 were modestly downregulated by this compound (data not shown). Overall, our findings suggest that additional integrins and adhesion molecules may play a role in osteoclastogenesis. In osteoclast precursors, the DASH proteases inhibitor also dose-dependently inhibits p38 MAPK, which is a downstream kinase critical for osteoclast differentiation (Li et al, 2003). In contrast, NF-κB activity, which is also essential for osteoclast formation was not affected by this compound. Intriguingly, a previous report showed that FAP, although absent in most adult tissues, is occasionally expressed in bone (Rettig et al, 1988). FAP was found to be highly expressed in synoviocytes in rheumatoid arthritis (Bauer et al, 2006), suggesting that DASH proteases may also be involved in other bone manifestations.
In summary, we have shown that DASH proteases are highly expressed by osteoclasts and that inhibition of FAP expression in myeloma cell–osteoclast cocultures attenuated osteoclasts' stimulatory effects on myeloma cell survival (Ge et al, 2006). The specific DASH inhibitor, PT-100, inhibited growth of MM tumour burden and reduced MM-induced bone disease in vivo. Mechanistically, PT-100 downregulated expression of certain adhesion molecules in osteoclasts cocultured with myeloma cells. This agent also significantly inhibited osteoclast formation and subsequent activity without affecting osteoclast viability. Collectively, these data support the idea that certain DASH proteases are involved in the pathogenesis of MM and its associated bone disease.
This work was supported by grants from the National Cancer Institute (CA-93897) (S.Y.) and from the Multiple Myeloma Research Foundation (Senior and Translational Research Awards) (S.Y.). We would like to recognize the efforts of Nisreen Akel. We also wish to thank the faculty, staff, and patients of the Myeloma Institute for Research and Therapy for their support, and the Office of Grants and Scientific Publications at the University of Arkansas for Medical Sciences for editorial assistance during the preparation of this manuscript.
Author's contributions Contribution: A.P. performed in vitro and in vivo experiments, analysed and interpreted data, and wrote the paper. X.L. performed in vitro and in vivo experiments. W.L. constructed SCID-hu mice, monitor in vivo tumour growth and performed immunohistochemistry and histochemistry. S.K. prepared primary myeloma cells for the study and maintain cell cultures. D.G. and L.S. performed BMD analysis and pit formation assays. N.A. helped design the study and interpret data. BB and JS provided patients' materials and interpreted the data. S.Y. designed the study, performed research, conceptualized the work, analysed and interpreted the data, and wrote the paper.
Conflict of interest disclosure: The authors declare no competing financial interests.