Chondroitin Synthase 1 Is a Key Molecule in Myeloma Cell-Osteoclast Interactions*

There is a symbiotic relationship between continued growth and proliferation of myeloma cells and the bone destructive process. It has been shown in animal models that blocking bone destruction can result in decreased myeloma tumor burden. Osteoclasts are bone destroying cells found in the bone marrow, and their significance in myeloma is supported by recent findings that osteoclasts alone can support sustained survival and proliferation of purified primary myeloma cells in ex vivo co-cultures. However, molecular mechanisms associated with interactions between myeloma cells and osteoclasts remain unclear. Here, we show that when myeloma plasma cells are co-cultured with osteoclasts, chondroitin synthase 1 (CHSY1) is the most significantly altered soluble, secreted protein present in the conditioned medium. RNA interference experiments with CHSY1 small interfering RNA (siRNA) reduced the amount of CHSY1 in the co-culture conditioned medium, and this was associated with a 6.25-fold increase in apoptotic myeloma cells over control co-cultures. CHSY1 contains a Fringe domain, and Fringe is well known for its regulation of Notch signaling via its DDD motif. And interestingly, Fringe domain in CHSY1 has this DDD motif. Shortly after co-culture with osteoclasts, we found that the Notch2 receptor was activated in myeloma cells but Notch1 was not. Activation of Notch2 was down-regulated by CHSY1 siRNA treatment. Modulating Notch signaling by CHSY1 via its DDD motif provides new insight into mechanisms of the interactions between myeloma cells and their bone marrow microenvironment. Targeting this interaction could shed light on treatment of myeloma, which is currently incurable.

There is a symbiotic relationship between continued growth and proliferation of myeloma cells and the bone destructive process. It has been shown in animal models that blocking bone destruction can result in decreased myeloma tumor burden. Osteoclasts are bone destroying cells found in the bone marrow, and their significance in myeloma is supported by recent findings that osteoclasts alone can support sustained survival and proliferation of purified primary myeloma cells in ex vivo co-cultures. However, molecular mechanisms associated with interactions between myeloma cells and osteoclasts remain unclear. Here, we show that when myeloma plasma cells are co-cultured with osteoclasts, chondroitin synthase 1 (CHSY1) is the most significantly altered soluble, secreted protein present in the conditioned medium. RNA interference experiments with CHSY1 small interfering RNA (siRNA) reduced the amount of CHSY1 in the co-culture conditioned medium, and this was associated with a 6.25-fold increase in apoptotic myeloma cells over control co-cultures. CHSY1 contains a Fringe domain, and Fringe is well known for its regulation of Notch signaling via its DDD motif. And interestingly, Fringe domain in CHSY1 has this DDD motif. Shortly after co-culture with osteoclasts, we found that the Notch2 receptor was activated in myeloma cells but Notch1 was not. Activation of Notch2 was down-regulated by CHSY1 siRNA treatment.

Modulating Notch signaling by CHSY1 via its DDD motif provides new insight into mechanisms of the interactions between myeloma cells and their bone marrow microenvironment. Targeting this interaction could shed light on treatment of myeloma, which is currently incurable.
Multiple myeloma (MM) 1 is a disease characterized by clonal B-cell tumors of slowly proliferating plasma cells within the bone marrow. Despite treatment advances, myeloma is uniformly fatal to the nearly 14,000 people newly diagnosed with the disease in the United States each year. Myeloma is closely associated with lytic bone disease, the most debilitating manifestation of the malignancy, found in 80% of patients (1). The current data on MM disease progression indicate that myeloma cell proliferation is highly dependent upon the state of the bone marrow microenvironment. Furthermore, it has been reported that interactions between myeloma cells and their microenvironment alter the ability of myeloma cells to resist to chemotherapy (2,3).
It appears that continued proliferation of myeloma cells is somehow dependent upon the bone-destructive properties of osteoclasts. This is supported by the clinical observation that patients with smoldering myeloma who were treated only with the bone-resorption inhibitor pamidronate experienced tumor burden reduction (8) and improved survival (9). It has been reported that inhibiting myeloma bone disease by halting bone resorption via treatment with bisphosphonates or inhibitors of receptor activator of nuclear factor B ligand (RANKL) signaling has a profound antimyeloma effect. The data using the severe combined immunodeficient-human (SCID-hu) animal model indicate that osteoclasts, cells responsible for the breakdown of bone, facilitate survival and growth of myeloma cells (4 -7). Furthermore, it has been recently found that sustained survival and proliferation of purified primary myeloma cells in ex vivo cultures can be supported by osteoclasts alone in a contact-dependent manner (10).
While the symbiolic relationaship between myeloma cells and osteoclasts are well demonstrated, their molecular details have not been fully determined. We focused our efforts on examining these molecular mechanisms in an effort to ultimately identify novel therapeutic targets for myeloma. As a first step toward this goal, we cultured myeloma cells with osteoclasts and examined the changes in secreted protein profiles. After co-culturing the cells, we used cytokine arrays to detect changes in specific known proteins and surface enhanced laser desorption/ionization (SELDI) proteomics to examine global changes and also identify proteins undetectable in the cytokine arrays. We identified the most prominent secreted protein as chondroitin synthase 1 (CHSY1), which has a predicted Fringe domain putatively regulating the Notch pathway (11,12). In Drosophila, Fringe activates Notch2 signaling from Delta ligands and simultaneously inhibits Notch1 signaling from Serrate and Jagged ligands (13,14). Recently, Notch signaling has been discovered to be involved in intercellular signaling between myeloma cells and other cells in the bone marrow. We hypothesize that CHSY1 plays a role in myeloma by stimulating Notch2 signaling and inhibiting Notch1 signaling.
Here, we present our results showing that CHSY1, the most prominent secreted protein detected in myeloma cell-osteoclast co-culture conditioned medium, activates Notch2 signaling in myeloma cells. This is particularly significant because of the apparent relationship between Notch signaling and the acquisition of chemotherapy resistance by myeloma cells. Our findings may provide new insight into the biology of myeloma as well as chemotherapy resistance in other cancers.

MATERIALS AND METHODS
Preparation of Myeloma Cells-Samples of myeloma plasma cells (MM PCs) were obtained from heparinized bone marrow aspirates from 50 patients with active myeloma during scheduled clinic visits. Signed Institutional Review Board-approved informed consent forms are kept on record. Bone marrow samples were separated by density centrifugation using Ficoll-Paque (specific gravity: 1.077 g/ml), and the proportion of MM PCs in the light density cell fractions was determined by CD38/ CD45 flow cytometry. MM PCs were isolated using CD138 immunomagnetic bead selection and the autoMACs automated separation system (Miltenyi Biotec, Auburn, CA). Purity of MM PC preparations was determined by CD38/CD45 flow cytometry to be routinely 95% (10).
Preparation of Osteoclasts-Peripheral blood mononuclear cells were obtained from patients with myeloma by leukopheresis. Signed Institutional Review Board-approved informed consent forms are kept on record. The cells were cultured at 37°C in a humidified atmosphere of 5% CO 2 at 2.5 ϫ 10 6 cells/ml in osteoclast medium (␣-minimum essential medium (MEM) supplemented with 10% fetal bovine serum, antibiotics (10 g/ml penicillin, 10 g/ml streptomycin, and 20 g/ml neomycin), RANKL (50 ng/ml], human macrophage colony-stimulating factor (M-CSF; 25 ng/ml), and dexamethasone (10 nM)). After removal of non-adherent cells, 95% of the remaining cells are multinucleated osteoclast (OC) and their precursors. Osteoclasts generated in these culture are multinucleated, TRAP-and vitronectin receptor-positive, and form resorption pits on mineralized discs; details have been described previously (10).
Myeloma Cell-Osteoclast Co-culture-Osteoclast cultures in 6-well and 24-well plates were washed three times with serum-free ␣-MEM medium without RANKL or M-CSF. Purified MM PCs in serum-free ␣-MEM were added to osteoclast cultures in a ratio of ϳ1:1000. After 6 h of co-culture, media and MM PCs were collected and centrifuged, the media frozen at Ϫ80°C for future analysis, and the cells analyzed for viability by trypan blue exclusion and for apoptotic ratio by propidium iodide (PI) flow cytometry.
Apoptosis Assay-The proportion of apoptotic cells was determined using PI flow cytometry. MM PCs were recovered by shaking and washing with phosphate-buffered saline (PBS). Samples were centrifuged for 10 min at 1500 rpm, washed in PBS and centrifuged again, resuspended in 200 l of PBS, fixed in 800 l of 70% ethanol, and stored overnight at 4°C. MM PCs were centrifuged and resuspended in 1 ml of staining solution (RNase A (40 g/ml) and PI (50 g/ml) in PBS) for 1 h at room temperature in the dark. Cells were then analyzed on an Immunocytometry system FACScan flow cytometer (BD Biosciences) using a data collection rate of 50 -80 events/s. Apoptotic sub-G 1 and cell cycle G 2 /M analyses were performed by Cell Quest (BD Biosciences).
Human Cytokine Array Analysis-Aliquots of serum-free media from MM PC cultures (from four different donors), osteoclast cultures, and MM PC-osteoclast co-cultures were collected after 6 h. Media samples were mailed on dry ice to RayBiotech, Inc. (Norcross, GA) for processing on their 79-cytokine array. Briefly, the array membrane was incubated with an experimental sample, a mixture of biotin-labeled antibodies, and then horseradish peroxidase-conjugated streptavidin for detection. Each array has internal positive (biotin-labeled protein) and negative controls (bovine serum albumin).
SELDI ProteinChip Analyses-Soluble proteins in serum-free 6-h conditioned medium were studied using SELDI ProteinChip technology (Ciphergen, Fremont, CA). One-l aliquots were spotted on NP20 Pro-teinChip arrays and air-dried. An energy-absorbing matrix (sinapinic acid, Ciphergen) was added, and protein profiles were analyzed on ProteinChip reader (model PBSII) calibrated with protein standards for mass determination (Ciphergen). Protein mass spectra were generated using 60 laser shots (laser intensity of 230, detector sensitivity of 9 arbitrary units).
Enrichment and Identification of Protein of Interest-The high molecular weight proteins in 6-h conditioned medium were enriched ϳ100fold using Centricon YM-30 centrifugal filters (Millipore, Billerica, MA). Concentrated proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using 12% Ready Gels (Bio-Rad). After staining with colloidal blue staining kit (Invitrogen), a protein band corresponding to the size of the protein of interest, as determined by SELDI, was excised and trypsin-digested using Protein-Chip peptide mapping kit according to manufacturer's instructions (Ciphergen). Mass spectra of digested peptides were generated by SELDI and submitted to the National Center for Biotechnology Information protein data base for identification.
Western Blot Analysis for Notch Activation-For cytosolic protein extraction of myeloma and osteoclast, about 20 millions of myeloma cells and osteoclasts were lysed in 1 ml of a 50 mM Tris buffer, pH 7.4, containing 150 mM NaCl, 0.5% Nonidet P-40, and 0.5% sodium deoxycholate, supplemented with protease inhibitor mixture P 8340 (a mixture of 4-(2-aminoethyl)-benzenesulfonyl fluoride, pepstatin A, E-64, bestain, leupeptin, and aprotinin) (Sigma), after washing three times in PBS buffer, at a dilution of 1:200. The cytosolic fractions were collected by spinning the sample at 3000 rpm in a microcentrifuge for 5 min at 4°C, and the supernatants were collected.
Statistical Analysis-Unless indicated otherwise, all values are expressed as means Ϯ S.E. All experiments were repeated at least three times. Student's paired t test was used.

Evaluation of Cytokines and Chemokines in Myeloma Cell-
Osteoclast Serum-free Co-culture-It has been reported that in co-culture with myeloma cells, osteoclasts support survival and sustained proliferation of freshly purified primary myeloma cells for extended time periods and that this effect requires contact between myeloma cells and osteoclasts (10). Here we present confirmation of these results (Figs. 1 and 2) in serumfree conditions which allow easy collection and subsequent identification of soluble secreted proteins from the media. Coculturing myeloma cells with osteoclasts not only inhibits myeloma cell death ( Fig. 1) but also promotes myeloma cell proliferation. The influence on cell proliferation is evidenced by the 4.5-fold increase in the proportion of MM PCs in G 2 /M phase after 2 days of co-culture with osteoclasts (Fig. 2). Before co-culture, 4% of cells were in G 2 /M; this level increased to 18% after co-culture.
Several cytokines and chemokines have been shown to be involved in growth and proliferation of myeloma cells and the associated destruction of bone, including RANKL, MIP-1␣, interleukin (IL)-1, IL-3, IL-6, TNF-␣, hepatocyte growth factor, and insulin-like growth factor (30 -35). As a first step toward understanding the mechanisms of myeloma cell-osteoclast interactions, we sought to evaluate changes in the cytokine and chemokine milieu in co-culture conditioned media. This experiment serves two purposes. First, by evaluating previously characterized changes in specific proteins (i.e. MIP, IL-3, and IL-6), we can determine whether our experimental system is functional for identifying changes in the profile of secreted proteins during MM PC-osteoclast co-culture. Second, it allows us to analyze changes in levels of other cytokines and chemokines that have not yet been investigated in this context.
We cultured MM PCs and osteoclasts individually and in co-culture and then analyzed the conditioned media using 79cytokine arrays (RayBiotech, Inc.). These arrays use antibodies to 79 cytokines, along with positive and negative controls, to quantitatively identify these known proteins within conditioned medium. We examined four separate co-cultures and compared them with four individual cultures of MM PCs and osteoclasts. We found that within 6 h of initiating co-culture, IL-6, MIP-1␤, MCP-1, GRO, and ENA-78 were consistently up-regulated in all four co-cultures (Fig. 3). IL-7 and IL-10 were up-regulated in two of the four cultures (each with osteoclasts from a different source) (Fig. 3). TNF-␣ and hepatocye growth factor were up-regulated in only one of the four cultures. Our data confirm the previously reported up-regulation of IL-6 and MIP. IL-1, IL-3, and insulin-like growth factor were not detected in co-culture conditioned media, although they have been reported to be involved in myeloma cell-osteoclast interactions. Our data add three cytokines, MCP-1, GRO, and ENA-78, to the list of proteins potentially involved in the interplay between myeloma cells and their microenvironment. Furthermore, it appears that our serum-free co-culture system provides a simple, useful model to study these interactions.
SELDI Analysis of the Secreted Proteome of Myeloma Cell-Osteoclast Co-culture-To identify differences in the secreted proteome of individually cultured versus co-cultured MM PCs and osteoclasts, we used SELDI to analyze conditioned media from these three groups. We found that after 6 h of co-culture, there were about 30 detectable peaks indicating changes before and after co-culture. However, there was one prominent difference between the protein profiles: in co-culture conditioned medium, a 66-kDa protein was up-regulated 4.6-fold over its level in conditioned medium from either MM PCs or osteoclasts cultured alone (Fig. 4).
Because the 66-kDa protein was so highly up-regulated in MM PC-osteoclast co-culture, we focused on determining its identity. We used SDS-PAGE to purify the 66-kDa protein and performed tryptic digests on the sample from the excised band. Based on SELDI spectra, we identified the protein as CHSY1 (p ϭ 1) (Fig. 5). CHSY1 is a type II membrane protein with both ␤1,3-glucuronate transferase and ␤1,4-N-acetylgalactosamine transferase activities (18), as well as a Fringe-like domain (amino acids 87-237; 62.7% homology to Fringe; BLAST analysis, National Center for Biotechnology Information) (19).
CHSY1 Is Critical to MM PC-Osteoclast Interactions-To determine the functional significance of CHSY1 to the myeloma cell-osteoclast interaction, we designed siRNA experiments to examine the effects of inactivating CHSY1 expression. We designed 21-nucleotide CHSY1 siRNAs and transfected them into OC and then cocultured with MM cells. These double-stranded RNA molecules should specifically suppress levels of their target, CHSY1, at both the mRNA and protein levels. We used three different siRNAs to CHSY1 (siRNA1, -2, and -3) in an effort to increase the effectiveness of CHSY1 inactivation. As a negative control designed to identify changes in gene expression that result from the siRNA delivery method, we used negative GAPDH siRNA, which is not complementary to any gene in our target cells.
SELDI analysis of the conditioned media showed markedly reduced levels of CHSY1 in co-cultures of cells carrying transfected siRNA (Fig. 6). To ascertain whether the effect of the siRNA were specific for CHSY1, we used the Ciphergen Pro-teinChip software protein biology system to quantify three protein peaks identified by SELDI (66 kDa (CHSY1), 44 kDa, and 33 kDa). We confirmed that siRNA treatment did not result in a global reduction of secreted proteins. In fact, the 33-kDa protein was up-regulated in cells carrying siRNA (Fig.  7). Therefore, siRNA treatment significantly reduced the level of secreted CHSY1 in a protein-specific manner.
We examined the biological impact of CHSY1 suppression on MM PC proliferation and apoptosis. We reasoned that if CHSY1 is critical to MMPC-osteoclast interactions, then specifically reducing its level should alter the protective effect that osteoclasts impart to myeloma cells (see Figs. 1 and 2). Exposure of MM PC-osteoclast co-cultures to CHSY1 siRNAand and co-culture transfected cells for 2 days resulted in an increased number of apoptotic MM PCs (20% apoptosis without siRNA; 48% with siRNA), decreased MM PC viability (83% viability without siRNA; 53% with siRNA), and a 40% reduction in the number of live MM PCs the siRNA negative control (Figs. 1 and  8). Our data suggest that elevating the level of CHSY1 contributes to the high degree of MM PC growth and proliferation that occurs in co-culture with osteoclasts.
Notch Signaling Is Involved in MM PC-Osteoclast Interactions-CHSY1 has a Fringe-like domain. Fringe regulates Notch1 and -2 signaling via a well characterized glycosyltransferase that modulates the Notch receptors to specific ligands through glycosylation (11, 12, 20 -24). We hypothesize that CHSY1 contributes to myeloma by stimulating Notch2 signaling and inhibiting Notch1 signaling, consistent with the predicted activities of its Fringe-like domain (13,14).
To test our hypothesis, we used Western blot analysis to investigate Notch signaling in MM PCs during co-culture with osteoclasts. Activation of a Notch receptor involves cleavage of a 100-kDa segment, which can be identified by Western blot. Release of the 100-kDa peptide indicates activation of the signaling pathway because it is imported into the nucleus where it binds transcriptional regulators and initiates downstream signaling. Our data show that Notch2 is cleaved in co-cultured MM PCs but not in MM PCs cultured alone (Fig. 9). Activation of Notch1 was not detected (data not shown). Our results demonstrate that Notch2 signaling is activated in MM PCs when co-cultured with osteoclasts but not when cultured alone.
To determine whether Notch2 activation is due to elevated levels of CHSY1, we utilized our siRNA system. We found that siRNA2 showed the strongest reduction in CHSY1 level and siRNA1 showed the weakest (data not shown). Therefore, we examined the effects of both siRNA1 and -2 on activation of Notch2. We did not detect Notch2 activation in co-cultured cells carrying siRNA2, but cells carrying siRNA1 (with less CHSY1 suppression) activated Notch2. These results suggest that Notch2 activation in MM PCs co-cultured with osteoclasts is dependent on the level of CHSY1. Furthermore, the data are consistent with our hypothesis that CHSY1 contributes to myeloma by activating Notch2 signaling and inhibiting Notch1 signaling via its Fringe-like domain.
DDD Motif Is Key to Fringe Modulation through Glycosylation of Notch-Fringe encodes an O-fucose-␤1,3-N-acetylglucosaminyltransferase (␤3GlcNAcT) that adds O-fucose glycans to the Notch EGF repeats. That the biological function of Fringe involves protein glycosylation was at first surprising, and indeed, it has also been suggested that Fringe directly interacts with Notch to antagonize Serrate binding and activation of downstream signaling. However, mutation of conserved residues within the catalytic domain of Fringe proteins eliminate both Fringe modulation of Notch signaling in response to Delta1 and Jagged1 and the biological effects of Dfng on wing development. These studies provide strong support that the biological function of Fringe is dependent on intrinsic glycosyltransferase activity and correlate with studies that support a role for Fringe within the cell, possibly through specific glyco- We quantitated the amounts of three proteins found in 6-h conditioned medium from co-culture, relative to their levels after addition of negative control GAPDH siRNA. Of the three CHSY1 siRNAs tested, siRNA2 and siRNA3 gave consistent effects on myeloma cell death, and siRNA2 is the one with the strongest effect on production of CHSY1. Gray bars represent 66-kDa protein; white bars represent 44-kDa protein, and black bars represent 33-kDa protein. sylation of Notch EGF repeats. Many sugars are transferred to proteins from a sugar nucleotide precursor such as UDP. The hallmark of the catalytic site of many nucleotide diphosphatebinding glycosyltransferase (35,36) is a three-amino acid residue sequence DXD, where D is Asp and X is any amino acid residue. Fringe has a sequence of three Asp residues at positions 236 -238 (called the DDD motif), indicating that it might bind to UDP. Munro and Freeman (37) showed by UV crosslinking experiments that this is the case. Furthermore, when they replaced Asp 236 with an Ala (ADD), they found that the protein had no biological activity. Even when Fringe-ADD was expressed ectopically at high levels, there was no disruption of the dorsoventral margin in the eye or the wing discs. These results suggest that the DDD motif is key to Fringe activity.
To infer function from homology between Fringe-like domain in CHSY1 and Fringe, we have had multiple lines of evidence indicating that the CHSY1 could have the same function as Fringe. The Fringe-like domain in CHSY1 shares 47% amino acid similarity with Fringe, and it also has the key DDD motif; multiple sequence alignment showed that the DDD motif is the only conserved motif shared by all Fringe proteins, from fruit fly to mice and human. The DDD motif was also the only motif that shared with the Manic, Lunatic, and Radical human Fringe, together with CHSY1 (Fig 10). DISCUSSION Cytokines and chemokines are known to be important signaling molecules in myeloma biology. Expression of several of these molecules has been studied in myeloma cells and others have been characterized as playing a role in myelomagenesis. We have identified three other cytokines that appear to be involved in myeloma via interactions between MM PCs and osteoclasts. MCP-1, GRO, and ENA-78 can now be added to the growing list of potential molecules whose regulation is changed in the myelomatous bone marrow microenvironment. Further study of the functions of these proteins in myeloma will contribute to the molecular picture of myeloma cell-bone marrow interactions.
We have confirmed expression of MIP and IL-6 in our serumfree MM PC-osteoclast co-culture system. Involvement of these cytokines in myeloma and osteoclastogenesis has been previously documented (30 -32). Both are produced by myeloma cells and are known to stimulate osteoclast formation; MIP-1␣ directly stimulates osteoclast precursors to differentiate into mature bone-resorbing osteoclasts, and inhibition of MIP-1␣ in murine models of myeloma results in markedly decreased bone destruction and a significant reduction in tumor burden (30 -32). These data suggest that MIP-1␣ is a major mediator of osteoclastic bone destruction in myeloma.
A variety of other cytokines and chemokines have been implicated in the bone destructive process in myeloma, including IL-1, IL-3, TNF-␣, hepatocyte growth factor, and insulin-like growth factor (34). There was some cell culture-dependent variation in the identified cytokines. We believe that the heterogeneity in the cytokines found in the co-culture condition medium reflects differences between the myeloma cells from different patients because we have previously shown that osteoclasts from healthy donors and myeloma patients equally support sustained survival and proliferation of MM PCs from all patients (10). We suggest that interactions between myeloma cells and osteoclasts may reflect heterogeneity in the properties of MM PCs from different patients, possibly due to different disease characteristics.
CHSY1 is the most prominent protein found in conditioned media from our co-cultures. It is present at relatively low levels in both MM PC and osteoclast cultures, and this level is increased 4.6-fold when the two cell types are co-cultured together. This result is highly suggestive that interactions between MM PCs and osteoclasts result in a significant change in regulation of this secreted protein. Because it is well documented that interactions between these cell types result in increased proliferation and apoptotic resistance of MM PCs in addition to increased bone destruction by osteoclasts, we hypothesize that CHSY1 may be a molecular component of these disease properties.
CHSY1 is an enzyme involved in chondroitin sulfate synthesis (18). Chondroitin sulfate and dermatan sulfate are synthesized as glycosaminoglycan side chains of proteoglycans. Chondroitin, chondroitin sulfate, and dermatan sulfate chains play structural roles in cartilage, bone, and other connective tissues, and they also have intriguing functions in growth factor signaling, morphogenesis, and cell division (25)(26)(27). Recent work in Caenorhabditis elegans revealed the novel role for chondroitin in morphogenesis and cell division with the use of RNA interference and squashed vulva mutant analysis (28,29).
We utilized an RNA interference approach to test our hypothesis that the level of CHSY1 in MM PC-osteoclast coculture correlates with growth and proliferation of MM PCs. We confirmed that CHSY1 levels were significantly reduced when cells carrying CHSY1 siRNA were co-cultured with osteoclasts. We found that in these MM PCs there was an increase in the number of apoptotic cells, which supports our hypothesis. Therefore, it appears that high levels of CHSY1 afford MM PCs some degree of apoptotic resistance. This reinforces the notion that CHSY1 is an important component of the bone marrow microenvironment that supports growth of myeloma cells.
The possibility remains, however, that the effect on MM PC viability is not directly due to the decrease in CHSY1. We investigated the levels of two other secreted proteins in this RNA interference experiment. We found that one of these proteins was up-regulated in cells carrying CHSY1 siRNA, while the other was down-regulated. There are several possible explanations for this observation. The first is that CHSY1 siRNA is not specific for CHSY1 and interferes with production of these other proteins also. Another explanation is that altering the level of CHSY1 has an effect on the regulation of these other proteins. If CHSY1 is involved in cell signaling, then regulation of these secreted proteins could be downstream effects of a lack of CHSY1. It is difficult to distinguish these possibilities. In either case, the effect on MM PC viability may be directly or indirectly due to suppression of CHSY1 activity. CHSY1 has two identified functional domains: a galactosyltransferase domain and a Fringe-like domain. Fringe regulation of Notch signaling is well characterized in Drosophila (11,12): Fringe regulates both the Notch1 and Notch2 pathways, inhibiting Notch1 signaling from Serrate/Jagged ligands while potentiating Notch2 signaling from Delta ligands (13,14). Activation of Notch2 signaling is initiated by cleavage of a 100-kDa intracellular domain of Notch that contains a nuclear localization signal. Upon import into the nucleus, the peptide binds to members of the CBF-1/Suppressor of hairless/Lag-1 family of transcriptional regulators, resulting in activation of a number of downstream gene products that effect cell fate. The presence of the Fringe domain lends credibility to the possibil- ity that the change in levels of other secreted proteins seen in the RNA interference experiment is due to regulatory changes due to a decrease in signaling by CHSY1.
Very recent reports show that Notch receptors are expressed on hematopoietic stem cells. These cell-surface receptors interact with their cell-surface ligands on osteoblast-precursor cells, thereby affecting decisions of cell fate and survival in the bone marrow (15). Furthermore, Notch signaling is involved in interactions between myeloma cells and their bone marrow microenvironment. Ligand-induced Notch signaling contributes to myelomagenesis in vivo (15). Notch receptors, Notch1 and Notch2, and the Notch1 ligand Jagged 1 are highly expressed in myeloma cells, but they are undetectable in non-neoplastic cells. Furthermore, it appears that Notch1 activation protects myeloma cells from drug-induced apoptosis, and this protection is associated with up-regulation of p21 and growth inhibition of myeloma cells (16). These results suggest that Notch signaling is important in myeloma biology, and we propose that CHSY1 is a key component in this signaling process.
Here, we have shown that CHSY1 is up-regulated in MM PCs co-cultured with osteoclasts, and we hypothesize that CHSY1 plays a role in myeloma by stimulating Notch2 signaling and inhibiting Notch1 signaling. Our hypothesis is supported by our finding that Notch2 was activated in MM PCs after 6 h of co-culture with osteoclasts, while Notch1 signaling was not detectable. Our Western blot shows that Notch2 activation was decreased when we suppressed CHSY1 activity by siRNA treatment.
Based on our data we propose a model whereby CHSY1 directly modulates Notch signaling via the Fringe-like domain. The Fringe domain may mediate the effects of CHSY1 by either binding to or post-translationally modifying the extracellular domain of Notch. The fact that the Fringe domain showed weak sequence similarity to several bacterial glycosyltransferases raises the possibility that CHSY1 could function by altering the structure of the carbohydrate chains on Notch. Functional studies of CHSY1 and Notch signaling in myeloma cells will elucidate the mechanisms of CHSY1 and its role in myeloma. Targeting CHSY1-Notch signaling could shed light on treatment of MM, a disease that is currently incurable.