The cell surface expression of CD9, a glycoprotein of the tetraspanin family influencing several processes including cell motility and metastasis, inversely correlates with progression in several solid tumors. In the present work, we studied the expression and role of CD9 in multiple myeloma (MM) biology using the 5T33MM mouse model. The 5T33MMvitro cells were found to be CD9 negative. Injection of these cells in mice caused upregulation of CD9 expression, while reculturing them resulted in downregulation of CD9. Coculturing of CD9-negative 5T33MMvitro cells with BM endothelial cells (BMECs) resulted in a partial retrieval of CD9. Laser microdissection followed by real-time polymerase chain reaction and immunohistochemistry performed on bone sections of 5T33MMvivo diseased mice demonstrated strong local expression of CD9 on MM cells in contact with BMEC compared to MM cells further away. These findings were also confirmed by immunohistochemistry in MM patients. Neutralizing anti-CD9 antibodies inhibited transendothelial invasion of CD9-expressing human MM5.1 and murine 5T33MMvivo cells. In conclusion, we provide evidence that CD9 expression by the MM cells is upregulated in vivo by close interaction of the cells with BMEC and that CD9 is involved in transendothelial invasion, thus possibly mediating homing and/or spreading of the MM cells.
Multiple myeloma (MM) is an incurable plasma cell malignancy characterized by an uncontrolled expansion and accumulation of monoclonal plasma cells (PCs) in the bone marrow (BM), secretion of paraprotein in the serum, development of osteolytic bone lesions and angiogenesis in the BM. The malignant cells depend on the BM microenvironment for their survival and growth and are therefore selectively localized in this tissue. Only in late disease stages, myeloma cells become independent of their microenvironment, resulting in extramedullary growth.1, 2, 3, 4 Given that the origin of the myeloma cells is post-germinal, their specific localization in the BM during the main course of the disease evolution implies migration from the vascular to the extravascular compartment of the BM.5 This migration is referred to as homing. A model described by Butcher and Picker6 represents the trafficking and homing of lymphocytes as a multistep process involving reversible rolling on the blood vessel wall, activation-dependent arrest and subsequent transendothelial migration. In analogy to the homing of lymphocytes, one can assume that the homing of myeloma cells is also such a multistep process.
CD9 is a cell surface glycoprotein belonging to the tetraspanin family. Tetraspanins constitute a family of over 30 membrane glycoproteins with four hydrophobic transmembrane domains. Most tetraspanins, except for CD53 and CD37, are broadly expressed. It has been suggested that they serve as adaptor molecules and they interact with numerous proteins including other members of the tetraspanin family, different types of receptors, proteins with Immunoglobulin (Ig) domains, integrins and major histocompatibility complex molecules. Most of the natural ligands are still unknown.7, 8, 9, 10, 11 It has been reported that tetraspanins can influence several biological and pathological processes, including cell motility and metastasis, adhesion, cell proliferation, differentiation, fusion, cellular signaling, cytoskeletal reorganization and virus infection.11, 12 Moreover, the expression of many members of the family is altered in different types of carcinomas.13, 14 In particular, the inverse correlation between the CD9 expression levels and tumor progression and metastasis in several solid tumors, including melanoma,15 breast,16 lung,17 colon18 and ovarian cancer,19 suggests that low CD9 expression might be one of the steps favoring cancer progression. Moreover, transfection of CD9 cDNA resulted in the suppression of motility and metastasis, suggesting that CD9 downregulation might be associated with highly metastatic behavior of the tumor cells.20, 21 However, recent reports are questioning this hypothesis: in human osteosarcoma22 and in breast cancer,23 CD9 expression levels did not provide useful prognostic information. Moreover, a study showed that in cervical carcinomas, CD9 is indeed globally downregulated in most invasive cervical carcinomas, but re-expressed in cells close to vessels and in the process of transendothelial transition.24 These findings indicate that the role of CD9 in cancer progression is very complex.
Until now, little is known about the expression and role of CD9 and tetraspanins in general in the MM cell biology and disease progression. In the early 1990s, Ikeyama et al.20 demonstrated suppression of cell motility after introduction of CD9 cDNA in the CD9-negative, highly motile, Epstein–Barr virus (EBV)-transformed human cell line ARH77. Recently, surface expression of several tetraspanin antigens (among which CD9) was investigated in a small number of newly diagnosed, untreated monoclonal gammopathy of undetermined significance (MGUS) (n=5) and MM (n=5) patients. Clonal PC from MM patients showed significant reduced CD9 expression as compared to PC from MGUS patients and to their normal counterparts.25 Screening of human myeloma cell lines (n=6) for tetraspanin expression by flow cytometry revealed the predominant absence of CD9, CD81 and CD82. This absence was characterized by reduced steady-state mRNA levels and methylation of the promoter regions. Re-expression of these tetraspanins both at mRNA and protein level following de-methylation supported functional contribution of the hypermethylation to regulating transcription of CD9 protein.26
In the present study, we used the 5T33MM experimental mouse model and the human myeloma cell line MM5.1 to investigate the expression and role of CD9 in MM. The 5T33MM mouse model is a suitable model to study the development of MM. The 5TMM cell lines originated spontaneously in elderly C57Bl/KaLwRij mice and have since been propagated in syngeneic, young mice by transplantation of diseased BM cells isolated from MM-bearing mice.27 By this method, several exclusively in vivo growing MM cell lines were obtained. The clinical and molecular characteristics of these models are very similar to those of the human disease.4, 28
The cellular expression of CD9 was analyzed both before and after in vivo passage followed by cultivating with or without BM endothelial cells (BMECs). Differential expression in the myeloma cells near BMEC or positioned at a distance from the BMEC was assessed in situ with laser microdissection (LMD) combined with quantitative real-time polymerase chain reaction (PCR) analysis. In addition, immunostainings for CD9 were performed on bone sections from both 5T33MM mice and untreated newly diagnosed MM patients. Finally, the possible role of CD9 in transendothelial invasion of the myeloma cells was evaluated.
Materials and methods
C57Bl/KaLwRij mice were purchased from Harlan CPB (Horst, The Netherlands). Mice were used at 8–10 weeks of age and housed and treated following the conditions approved by the Ethical Committee for Animal Experiments, VUB (License no. LA1230281).
The 5T33MMvv cell line originated from elderly C57Bl/KaLwRij mice that spontaneously developed MM. The cells have since been propagated into syngeneic, young mice by transplantation of diseased BM cells. Mice were either killed when clear signs of paralysis of the hind legs were observed (terminally diseased mice, 28 days) or at day 13 after intravenous (i.v.) injection of the tumor cells, a time point when they are first detectable by flow cytometry (intermediate stage, tumor load between 5 and 20%).29, 30 Isolation and purification of the myeloma cells in the BM was performed as described previously.31
The murine 5T33MMvt cell line is a clonally identical, but in vitro stroma-independent growing variant of the 5T33MMvv cell line. Cells were cultured and maintained in Rosewell Park Memorial Institute (RPMI)-1640 medium (BioWhittaker, Verviers, Belgium) supplemented with 10% fetal calf serum (FCS) (Fetal clone I; Hyclone, Logan, UT, USA), 1% natriumpyruvate, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM L-glutamine and 1% minimum essential medium (supplements from BioWhittaker). These cells are also able to induce MM, when i.v. injected in syngeneic, young mice. The take time of this 5T33MMvtvv model is 10 weeks. Mice were either killed when a serum paraprotein concentration of 10 mg/ml was reached (terminally diseased mice, 10 weeks) or 4–5 weeks after i.v. injection of the myeloma cells when they are detectable by flow cytometry (early stages, tumor load between 5 and 15%). Isolation and purification of the myeloma cells in the BM was performed as described previously.31
The murine BM sinusoidal endothelial cell line STR4 was originally established by transfecting primary endothelial cell cultures with SV40 and was cultured in RPMI-1640 medium supplemented as described above.32
The well-characterized human stroma-dependent MM5.1 and stroma-independent MM5.2 MM cell lines were selected for our experiments. Cells were kept in culture as described previously.33
The human BM endothelial cell line 4LHBMEC (kindly provided by Dr A Dräger, Department of Hematology, Free University Amsterdam, The Netherlands) was cultured in 10% FCS-199 medium (International Medical, Brussels, Belgium) with 10 ng/ml endothelial cell growth factor (ECGF) (Roche, Brussels, Belgium) until a confluent adherent cell layer was obtained. Cells were recovered for further use after trypsinization.
Cocultures of BM endothelial cells with 5T33MMvt and vtvvvt cells
STR4 cells (0.5 × 106) were seeded in 25 cm2 flasks (BD Falcon, Mountain View, CA, USA) in complete growth medium. After 24 h, cells were washed twice with serum-free medium and cultured with 5T33MMvt or 5T33MMvtvvvt cells (1 × 106 cells) in 6 ml RPMI-1640 medium with 5% fetal clone I (FCI) and supplements for 5 days.
CD9 cell membrane expression on the 5T33MM myeloma cells during disease progression was detected by a double-staining procedure. The rat anti-mouse antibody (KMC8) was used as primary antibody (BD, Mountain View, CA, USA) and fluorescein isothiocyanate (FITC)-conjugated goat anti-rat Ig G monoclonal antibody was used as a second step (Oxford Biotechnology, Oxfordshire, UK). An isotype-matched irrelevant antibody was used as control. The proportion of tumor cells was determined by staining the cells with a tumor-specific anti-idiotype monoclonal antibody as described previously.34 Cells were analyzed with a FACSCanto flow cytometer (BD) using CellQuest and FACS Diva software.
LMD and pressure catapulting microscopy
The hind limbs were dissected from 5T33MMvv terminally diseased mice and directly freeze embedded in OCT compound (Optimal Cutting Temperature; Tissue-TEK from Sakura Finetek, Zoeterwoude, The Netherlands) by cooling in liquid nitrogen. To prevent RNA degradation within the bones, the time between the death of the mouse and freezing the bones down to −80°C was kept below 5 min. Five-micrometer-thick sections were cut on the Cryo-microtome (Bright, OTF5000, Cambridgeshire, UK) combined with a CryoJane system (Instrumedics, St Louis, MO, USA). The sections were fixed in acetone for 3 min at 4°C and subsequently incubated for 90 s at 4°C with FITC-labeled rat anti-mouse CD31 (BD) antibodies diluted 1:5 in phosphate-buffered saline (PBS). Finally, the sections were quickly rinsed in PBS, dehydrated in series of ethanol and kept in 99% ethanol during the selection of cells for dissection. The slide was then positioned in a computer-assisted laser microdissecting and pressure-catapulting microscope (PALM laser technology, Bernried, Germany) equipped for fluorescence microscopy. Three areas were selected: CD31-positive BMECs, myeloma cells adjacent to the BMEC and myeloma cells away from the BMEC. As preliminary immunohistochemical evaluations of adjacent cryosections using the CD31 binding antibody and a tumor-specific anti-idiotype monoclonal antibody revealed the marrow to be packed with tumor cells, we assumed and confirmed by quantitative PCR (Q-PCR) that the dissected cells adjacent and away from the EC were essentially myeloma cells. The selected and dissected areas where sequentially catapulted into a cap containing 35 μl extraction buffer (Picopure, Arcturus, Mountain View, CA, USA) and the RNA was purified from the dissected cells as described by the manufacturer (Arcturus).
The quality of the RNA was assessed using the Experion system from Biorad (Hercules, CA, USA). Its principle is based on the quantification of the 18S/28S ratios. Ratios close to 2 were defined as high-quality RNA. Only specimens with such ratios were included in our analysis.
Quantitative real-time PCR
The RNA obtained from the laser microdissections was converted into cDNA using the Iscript first-strand synthesis system (Biorad) using random hexamer primers. Reference RNA samples with 1, 5 and 25 μg of RNA purified from full sections of diseased 5T33MM bones were converted into cDNA together with the laser dissected samples. Real-time Q-PCR was performed using the ABI PRISM 7700 Sequence Detector (Applied Biosystems, Foster City, CA, USA). The samples were amplified in 25 μl reactions using a Taqman assay-on-demand for CD31 (Mm00476702 and Mm01242575), CD9 (Mm01182917), GUS (β-glucuronidase; Mm00446953) from Applied Biosystems and a 5T33MM-specific primer (forward primer: 5′-IndexTermAAG-TTC-AAG-GGC-AAG-GCA-ACA-3′ and reverse primer: 5′-IndexTermCAG-GCT-GCT-GAG-CTG-CAT-G-3′) and probe (5′-IndexTermTGC-AGA-CAA-ATC-CTC-CAG-CAC-TGC-C-3′) set located between the CDR2 and CDR3 region. The design was based on the previously published sequence of the immunoglobulin heavy-chain sequence expressed by the 5T33MM tumor cells.35 The following conditions were used: 50°C for 2 min and 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. The standard curve method was used to quantify the gene expression of CD31, CD9 and 5T33MM idiotype relative to the endogenous reference gene GUS.
Cultured cells (5 × 106) undergoing gene expression analysis were lysed in 250 μl of Trizol (Invitrogen, Carlsbad, CA, USA) and maintained at −80°C. Total RNA was extracted using the Rneasy Mini kit (Qiagen, Valencia, CA, USA) as described by the manufacturer. The concentration and purity of RNA were determined by spectrophotometric measurement (Gene Quant II; Pharmacia Biotech, Cambridge, UK). Two microgram of total RNA was converted into cDNA by the superscript first-strand synthesis system (Invitrogen) using random hexamers as primers. Real-time Q-PCR was performed as described for the laser microdissected samples; however, amplifying in a 50 μl reaction. Each sample was amplified in triplicate. The relative standard curve method was used to quantitate the relative CD9 mRNA expression in myeloma cells. Relative standard curves (50, 10, 2, 0.2 ng) were prepared for CD9 and β-glucoronidase (GUS) using cDNA from 5TMM diseased BM.
The hind limbs from end-stage 5T33MM diseased mice were dissected and fixed in Zn-fixative, decalcified in 15% ethylenediaminetetraacetic acid (EDTA) and embedded in paraffin as described previously.36 The human bone biopsies from untreated newly diagnosed patients were fixed in formalin and decalcified in formic acid before paraffin embedment. The use of human bone biopsies was approved by the local ethical committee. Series of adjacent sections were cut and subjected to hematoxylin and eosin staining or processed for single/double-immunostaining as follows: sections were deparaffinized in xylene, washed in 99% ethanol, treated with 0.45% hydrogen peroxide in 99% ethanol to inactivate endogenous peroxides and rehydrated. Subsequently, the sections were demasked in Tris/EDTA buffer (pH 9.0) for 10–14 h at 60°C followed by blocking with 0.5% casein in Tris-buffered saline and with a biotin-blocking system (DakoCytomation, Glostrup, Denmark). The single-immunostaining on mouse bones was performed by incubating the sections with a rat anti-mouse CD9 antibody (BD), which was detected by a biotin-labeled rabbit anti-rat IgG antibody (DakoCytomation) followed by a step in horse radish peroxidase (HRP)-conjugated streptavidin (DakoCytomation). The detection was amplified by precipitation of biotin-conjugated tyramide, followed by a reincubation with HRP-conjugated streptavidin. The binding was visualized by diaminobenzidine tetrahydrochloride (DAB)+ precipitation.
The double-immunostaining on human tissues was performed in two steps: First, the sections were incubated with a mouse anti-human CD9 (1:40, clone 72F6, Novocastra, Newcastle, Upon Tyne, UK), which was detected with a polymer of HRP-conjugated goat anti-mouse IgG antibodies (EnVision, DakoCytomation) and visualized by DAB+ (DakoCytomation). Secondly, the sections were incubated with an FITC-labeled mouse anti-human CD34 class II (1:100, clone QBEnd10, Dako), which was detected with an alkaline phosphatase (AP)-conjugated sheep anti-FITC FAB fragment (1:20, Roche) and visualized by Fuchsin+ (DakoCytomation). Finally, the sections were stained with hematoxylin and aqua mounted.
Transendothelial invasion assay
To assess the effect of the myeloma cell-BMEC interaction on the myeloma cell invasiveness, invasion was evaluated in the presence of a BM endothelial cell layer.
The transendothelial invasion assays for the MM5.1 and 5T33MMvv myeloma cells were conducted in a Transwell migration system with a 8-μm pore size polyethylene (PET) membrane (Elscolab, Kruibeke, Belgium). Before cell culture, the upper face of the filters was coated with cold Matrigel (BD). Next, BMEC (3 × 104) were plated onto the Matrigel-coated PET membranes in the upper cell culture insert chamber and cultured until confluence (48 h for STR-4 cells and 24 h for 4LHBMEC cells). Myeloma cells (105 for the MM5.1 cells and 2 × 105 for 5T33MMvv cells) suspended in 200 μl of serum free RPMI-1640 medium were seeded in the upper compartment and incubated for 24 h. As chemoattractant, we used RPMI-1640 medium with 20% FCI (300 μl) in the lower compartment of the invasion chamber. Cells that invaded the BMEC monolayer and Matrigel were recovered from the lower compartment and counted by flow cytometry. A known number of sphero blank calibration beads were added as an internal standard (BD) for quantification of the migrated myeloma cells. The migrated myeloma cells were discriminated from potentially migrated EC by their different forward/sideward scatter characteristics.
To examine the role of CD9 in transendothelial invasion, the myeloma cells (murine and human) were pre-treated with either the neutralizing anti-mouse CD9 monoclonal antibody KMC8 (10 μg/ml, BD, Mountain View, CA, USA)37, 38, 39, 40 or the neutralizing anti-human CD9 monoclonal antibody ALB6 (10 and 20 μg/ml; Immunotech, Marseille, France)41 or the isotype-matched control antibody for 30 min. Antibodies were also added to the lower compartment. Results are presented as the percentage cell invasion compared with control invasion in the absence of monoclonal antibody. In some experiments, CD9 expression on the 5T33MMvv cells was investigated by flow cytometric analysis before and after transendothelial invasion.
For statistical analysis of the in vitro tests the Mann–Whitney U-test was used. For the LMD experiments, we used the paired Student's t-test. P⩽0.05 was considered significant.
CD9 expression in the 5T33MM mouse model
To investigate CD9 expression in MM, we used the murine model 5T33MM. Terminally diseased mice were killed and by a double-flow cytometric staining procedure with CD9 and tumor-specific anti-idiotype monoclonal antibodies, we demonstrated CD9 expression on the 5T33MMvv (in vivo growing) cell line at low levels (about 20% of the myeloma cells, range 18–32%) as shown by Figure 1a. To investigate CD9 expression at the early stages of disease development, 5T33MMvv cells obtained from terminally diseased mice were injected in naive mice and CD9 expression was ascertained in the intermediate stage by flow cytometry. As shown by Figure 1a, CD9 expression was detected on the majority of the myeloma cells (70–75%).
We next determined CD9 expression on the stroma-independent 5T33MMvt cell line. In contrast to the stroma-dependent variant (5T33MMvv), this cell line was completely negative for CD9 (Figure 1b). However, when these 5T33MMvt cells were injected in young, naive mice (in vivo passage) and isolated in the quiescent/intermediate stage, CD9 expression on the obtained myeloma cells (now termed 5T33MMvtvv) reappeared (Figure 1b). To confirm these data on RNA level by real-time Q-PCR, myeloma cells from terminally diseased 5T33MMvtvv and 5T33MMvv mice were isolated. In both 5T33MMvv and 5T33MMvtvv cells a distinct CD9 expression compared to the 5T33MMvt control cells (as shown in Figure 1c) was observed. There was an equal amount of CD9 mRNA detected in 5T33MMvtvv and 5T33MMvv cells of terminally diseased mice. These data indicate that the BM microenvironment most likely induced expression of CD9 on the 5T33MMvt myeloma cells (in vivo upregulation). Reculturing of 5T33MMvtvv cells for 14 days resulted in the downregulation of CD9 (Figure 1b and c), further confirming that the CD9 upregulation is dependent on the BM microenvironment.
BMEC dependency of CD9 expression on the myeloma cells
The BM microenvironment is a complex structure of various extracellular components and many cell types, including fibroblasts, endothelial cells, adipocytes, osteoblasts, osteoclasts, inflammatory cells, macrophages and hematopoietic progenitors. As the BMECs are the first cell type the myeloma cells encounter when homing to the BM, we investigated the potential involvement of BMEC in the CD9 upregulation on the myeloma cells. By real-time Q-PCR and flow cytometric analysis, we compared CD9 expression of the 5T33MMvtvvvt cells cultured with or without BMEC. Coculturing was performed 14 days after the isolation of the cells, a time point at which downregulation of CD9 had already started, as shown in Figure 1b. After 5 days of coculturing with the BMEC, CD9 expression was distinctly upregulated compared to control cells (Figure 2a). Moreover, we also observed a modest CD9 expression on the completely CD9-negative 5T33MMvt cells after 5 days of coculture with BMEC (Figure 2b). These data demonstrated that BMEC are at least one of the cell types engaged in the BM microenvironmental modulated CD9 upregulation.
Quantitative expression of CD9 in 5T33MMvv BM tissue specimens
To confirm the above data in situ, LMD and laser pressure catapulting were performed on sections of bones from terminally diseased 5T33MMvv mice, followed by real-time Q-PCR. In preliminary experiments, immunostainings with the 5T33MM idiotype-specific antibody and an antibody against CD31, a marker for BMEC,36 revealed that the BM is packed with myeloma cells and infiltrated with BMEC, as described, previously.29 For LMD, BMECs were stained with CD31 immunofluorescence, and three separate zones were laser dissected as follows (Figure 3a): (i) CD31-positive BMEC (EC), (ii) myeloma cells in contact with BMEC (called ‘close’ in Figure 3) and (iii) the myeloma cells more than five cell layers away from the BMEC (called ‘away’ in Figure 3). In preliminary experiments, the RNA quality was checked after the cutting and staining steps, and proved to be high (data not shown). Real-time Q-PCR revealed a predominant CD9 expression in myeloma cells in contact with the BMEC (close) compared to more distant positioned myeloma cells (away) as shown in Figure 3b. This differential expression was consistent within the four mice analyzed (Figure 3), and also consistent within different LMD preparations from the same mice (data not shown). The nature of the microdissected cells was verified by real-time Q-PCR by amplifying the 5T33MM idiotype-specific Ig heavy-chain gene and the BMEC marker CD31 (Figure 3b). The 5T33MM idiotype gene was expressed at equal levels in the myeloma cells microdissected close to and away from BMEC, demonstrating an equal purity of myeloma cells in the preparations irrespective of their proximity to BMEC. The level of CD31 expression in the myeloma cell preparations was not affected by how close they were to BMEC when microdissected. Thus, there is no evidence for increased contamination of myeloma cells by BMEC, when they are microdissected from sites close to BMEC. Furthermore, these myeloma cell preparations showed on average a 10 times lower expression of CD31, compared to the BMEC preparations. To rule out the possibility that this differential CD9 expression might be owing to difference in oxygen tension, 5T33MMvt cells were incubated under hypoxic conditions (1% oxygen) for 24 h at 37°C. Subsequent real-time Q-PCR showed no difference in CD9 expression in cells cultured in hypoxic conditions compared to normoxia (20% oxygen), whereas the vascular epidermal growth factor gene, known to be upregulated at hypoxic conditions42 was (data not shown). To rule out the possibility that the differential expression might be owing to RNA contamination of pericytes, the presence of pericytes between the BMEC and myeloma cells was investigated. By actin staining we confirmed the absence of pericytes (data not shown).
Immunolocalization of CD9 in BM biopsies from 5T33MMvv diseased mice and human myeloma patients
To confirm at the positioning level the in situ data obtained by real-time Q-PCR analysis of laser-microdissected cells, we performed immunohistochemical stainings for CD9. The immunostainings on bone sections from the end-stage 5T33MMvv diseased mice, revealed a distinct immunolocalization of CD9 in some of the myeloma cells next to the BMEC, whereas more distant myeloma cells were negative for CD9, as shown in two different 5T33MMvv mice (Figure 4a and b). The BMEC were also observed positive for CD9 (Figure 4a and b).
To assess the relevance of CD9 to myeloma patients, we performed immunostainings for CD9 in combination with the detection of CD34 class II (classical BMEC marker) on sections of 30 different BM biopsies from newly diagnosed myeloma patients. Here we observed CD9 immunoreactivity in myeloma cells in 70% (21 of 30) of the biopsies. If considering only the biopsies with interstitial distribution of MM cells, 80% of them (20 of 25) were partially positive for CD9. In the latter, 85% (17 of 20) showed the CD9-positive MM cells positioned mainly in the neighborhood of the CD34+ BMEC, whereas MM cells more distant from the BMECs were negative for CD9 (Figure 4c and d). This observation is consistent with observations from 5T33MM mice. Only two of six biopsies having a nodular MM cell distribution and two of nine biopsies having a packed MM marrow showed a local CD9 expression in a few MM cells. Interestingly, the latter had all a direct contact with BMEC (Figure 4e). The majority of the MM cells in nodular and packed distributions are negative for CD9 (Figure 4f). In contrast with the immunostainings on the 5T33MMvv diseased mice, the BMECs in the human biopsies were always negative for CD9. Megakaryocytes, from which CD9 originally was cloned,43 were used as positive controls for CD9 in the sections where MM cells and BMECs were negative.
Role of CD9 in transendothelial cell invasion of murine 5T33MMvv and human MM5.1 myeloma cells
To investigate the role of CD9 in transendothelial invasion of the myeloma cells, transendothelial invasion assays were performed. By flow cytometric analysis, we demonstrated that about 75% of the transendothelial-invaded 5T33MMvv myeloma cells expressed CD9 compared to only 20% CD9 expression before transendothelial invasion (Figure 5a). These data suggest the involvement of CD9 in this process.
To further investigate the role of CD9 in transendothelial invasion of the myeloma cells, transendothelial invasion assays were performed in which CD9-expressing murine 5T33MMvv cells were incubated with the monoclonal anti-CD9-neutralizing antibody KMC8. The invasion in the negative control was between 5 and 8%. Maximal invasion was 20–25%. As shown in Figure 5b, KMC8 significantly inhibited transendothelial invasion of the myeloma cells by about 47% at a concentration of 10 μg/ml (range 43–51%) as compared with the invasion in the absence of monoclonal antibody or in the presence of irrelevant isotype antibody. No further inhibition was observed when increasing the concentration of KMC8 to 15 or 20 μg/ml. The data indicate that CD9 indeed is involved in transendothelial invasion.
Next, to confirm these data for the human condition the human stroma-dependent MM5.1 and stroma-independent MM5.2 myeloma cells were screened by flow cytometric analysis for CD9 expression, as these cell lines resemble the murine situation (5T33MMvv versus 5T33MMvt) as closely as possible. Flow cytometric analysis showed strong CD9 expression on the MM5.1 myeloma cells, whereas the MM5.2 cells were CD9 negative (data not shown). The involvement of CD9 in transendothelial invasion was also confirmed when using the MM5.1 cell line and the human monoclonal anti-CD9-neutralizing antibody ALB6. The invasion in the negative control was between 2 and 5%. Maximal invasion was 15–20%. ALB6 significantly inhibited transendothelial invasion of the myeloma cells by about 29% at 10 μg/ml (range 25–31%) and 48% at 20 μg/ml (range 42–54%) as compared with the invasion in the absence of monoclonal antibody or in the presence of irrelevant isotype antibody (Figure 5c). Again, these data indicate involvement of CD9 in transendothelial invasion.
Despite intensive research and emerging therapies in the past few years, MM still remains an incurable malignancy. One of the characteristics is the predominant localization of the myeloma cells in the BM. This restricted presence is most likely owing to the combination of selective homing and subsequent selective survival of the myeloma cells in the BM.28 Homing of the myeloma cells is thus an essential event in the disease development. This process is mediated by chemokines, adhesion molecules and proteases.6 CD9 has been described to be involved in the regulation of cellular processes considered relevant for malignant conversion like migration and invasion.7, 12, 24 Accordingly, CD9 expression levels were found to correlate inversely with tumor progression in several malignant diseases.15, 16, 17, 18, 19 Based on these reports, it was assumed that low levels of CD9 expression could thus be one of the promoting steps of cancer progression and might thus have high prognostic value. However, recent reports questioned the reliability of CD9 expression in assessing prognosis of breast cancer and human osteosarcoma.22, 23 Moreover, recently it was demonstrated by Sauer et al.24 that in cervical carcinoma CD9 expression is downregulated during tumor progression to promote expansion of the malignant cells, but on the other hand is also locally re-expressed to adjust to microenvironmental requirements. They propose that cells with low CD9 expression are responsible for tumor growth, but only those cells able to upregulate CD9 can enter the blood circulation during further tumor progression. The 5T33MM model is a good model to test this hypothesis, as by in vivo transplantation, the expression of CD9 during the course of the disease can be followed.29
In the present study, we found by flow cytometric analysis a low CD9 expression on the surface of 5T33MMvv myeloma cells of terminally diseased mice, whereas the stroma-independent variant 5T33MMvt cell line did not express CD9. To analyze whether the BM microenvironment was engaged in the regulation of this differential CD9 expression, 5T33MMvt cells were injected in naive mice, isolated in the quiescent/intermediate stage (5T33MMvtvv cells) and were now found to express CD9. As reported previously, three phases can be distinguished during 5TMM tumorigenesis: a quiescent stage of slow tumor progression, an intermediate stage of moderate tumor progression and an end stage with accelerated progression.44 However, after in vitro culturing of the 5T33MMvtvv cells, CD9 expression was again downregulated, suggesting that tumor–stroma interactions are involved in the CD9 upregulation in vivo. These data were confirmed by the upregulation of CD9 after injecting low CD9-expressing 5T33MMvv end-stage myeloma cells in naive mice. These data clearly demonstrated that in the early disease stages, the BM microenvironment induces expression of CD9 in the 5T33MM cells.
Based on the fact that the BMEC are the first cell type the myeloma cells come across when homing to the BM and based on the recent publication reporting CD9 re-expression on carcinoma cells in process of invading blood and lymphatic vessels,24 the involvement of BMEC in the CD9 expression on the myeloma cells was assessed. BMECs were able to partially retrieve CD9 expression on the 5T33MMvtvvvt and 5T33MMvt cells. The difference between both cell lines concerning the extent of the retrieval of CD9 expression might be explained by difference in hypermethylation of the CD9 promoter. Indeed, variegated methylation patterns, even in clonally derived cell populations, as well as environment and stage-dependent levels of methylation have been reported.26, 45 In human cells, flow cytometric analysis showed very low CD9 expression on the stroma-independent growing MM5.2 myeloma cells (data not shown). However, after coculturing with the human BMEC 4LHBMEC cell line CD9 expression was again partially retrieved (data not shown). These data suggest that BMEC are, at least partially, responsible for the in vivo upregulation of CD9. Furthermore, using real-time Q-PCR on myeloma cells isolated in situ by LMD, we demonstrated reproducibly a predominant CD9 expression on the myeloma cells in contact with the BMEC in comparison to those myeloma cells more distant from BMEC. In contrast, control genes were expressed at similar levels at these respective areas. To our knowledge, we are the first to use LMD to assess microenvironmental effects on myeloma cells. Immunohistochemistry allowed confirming of these findings at the protein level and, furthermore, showed that not all MM cells in contact with BMEC are CD9 positive. This differential expression may relate to different contact times with BMEC, and is compatible with a transient upregulation of CD9. The immunostainings in the human biopsies showed the relevance of these observations to patients, and documented even more so the importance of the myeloma cell microenvironment for CD9 expression. There is indeed a big contrast between the high CD9 expression of myeloma cells in the interstitial configuration, which favors heterotypic cell interactions, and the low CD9 expression when they are in nodular or packed configuration, where heterotypic interactions are rare. In the latter configurations, the importance of the microenvironment is further documented by the fact that myeloma cells expressing CD9 are rare and only seen at the vicinity of endothelial cells.
Our group has already reported BMEC-dependent upregulation of MMP931 and insulin like growth factor (IGF)-1R46 in the 5T33MMvt cells, which are involved in the homing of the myeloma cells. As CD9 has been described to be involved in the cell adhesion, migration and invasion, we investigated the role of CD9 in transendothelial invasion. The majority of the transendothelial invading murine myeloma cells expressed CD9, suggesting that CD9 indeed might be involved in this process (either by upregulation of expression or by selective migration). Neutralizing anti-CD9 monoclonal antibodies significantly inhibited transendothelial invasion of the CD9 expressing human MM5.1 and murine 5T33MMvv cell line, again confirming CD9 involvement in homing. Recently, Longo et al.47 also provided evidence for a role of CD9 in tumor–endothelial cell interaction and vascular dissemination in melanoma by in vitro studies of transendothelial migration, which was significantly blocked by treatment with monoclonal anti-CD9 antibodies.
Our data thus confirm and extend the work reported by Sauer et al.24 We thus postulate that as myeloma cells encounter BMEC, CD9 becomes upregulated making them able to pass the endothelial barrier, either to enter the extravascular compartment of the BM or to enter the vasculature, necessary for the spreading of the myeloma cells to anatomically distant BM sites or extramedullary sites at the end stage of the disease. There is a significant increased microvessel density (MVD) in myeloma-invaded BM compared to non-invaded BM.36 During myeloma progression, the increase in tumor load is however more extensive than the increased microvessel density in the BM. Moreover, the size and perimeter of the microvessels decreases during myeloma progression.30 Both observations result in an increase in the proportion of myeloma cells to BMEC during myeloma progression. As such, as disease progresses, the percentage of myeloma cells in contact with BMEC declines, leading to a reduced number of myeloma cells that express CD9. The recent observation that clonal PC from MM patients show significantly reduced CD9 expression as compared to PC from MGUS patients confirm our data.25
In conclusion, our results demonstrate an upregulation of CD9 expression in the myeloma cells in early disease stage in vivo by the interaction of the myeloma cells with the BMEC. Moreover, in terminally diseased 5T33MMvv mice and newly diagnosed patients, we demonstrate a strong local expression of CD9 on myeloma cells in contact with BMEC. We also provide evidence that CD9 is involved in transendothelial invasion, thus possibly mediating homing and/or spreading of the myeloma cells.
Caligaris-Cappio F, Bergui L, Gregoretti MG, Gaidano G, Gaboli M, Schena M et al. Role of bone marrow stromal cells in the growth of human multiple myeloma. Blood 1991; 77: 2688–2693.
Lokhorst HM, Lamme T, de Smet M, Klein S, de Weger RA, van Oers R et al. Primary tumor cells of myeloma patients induce interleukin-6 secretion in long-term bone marrow cultures. Blood 1994; 84: 2269–2277.
Kuehl WM, Bergsagel PL . Multiple myeloma: evolving genetic events and host interactions. Nat Rev Cancer 2002; 2: 175–187.
Menu E, Asosingh K, Van Riet I, Croucher P, Van Camp B, Vanderkerken K . Myeloma cells (5TMM) and their interactions with the marrow environment. Blood Cells Mol Dis 2004; 33: 111–119.
Bakkus MHC, Heirman C, Van Riet I, Van Camp B, Thielemans K . Evidence that multiple myeloma Ig heavy chain VDJ genes contain somatic mutations but show no intraclonal variation. Blood 1992; 80: 2326–2335.
Butcher EC, Picker LJ . Lymphocyte homing and homeostasis. Science 1996; 272: 60–66.
Boucheix C, Rubinstein E . Tetraspanins. Cell Mol Life Sci 2001; 58: 1189–1205.
Berditchevski F . Complexes of tetraspanins with integrins: more than meets the eyes. J Cell Sci 2001; 114: 4143–4151.
Maecker HT, Todd SC, Levy S . The tetraspanin superfamily: molecular facilitators. FASEB J 1997; 11: 428–442.
Hemler ME . Integrin associated proteins. Curr Opin Cell Biol 1998; 10: 578–585.
Martin F, Roth DM, Jans DA, Pouton CW, Partrigde LJ, Monl PN et al. Tetraspanins in viral infections: a fundamental role in viral biology? J Virol 2005; 79: 10839–10851.
Hemler ME . Specific tetraspanin functions. J Cell Biol 2001; 155: 1103–1107.
Boucheix C, Duc GH, Jasmin C, Rubinstein E . Tetraspanins and malignancy. Expert Rev Mol Med 2001, 1–17.
Wright MD, Moseley GW, van Spriel AB . Tetraspanin microdomains in immune cell signalling and malignant disease. Tissue Antigens 2004; 64: 533–542.
Si Z, Hersey P . Expression of the neuroglandular antigen and analogues in melanoma. CD9 expression appears inversely correlated to metastatic potential of melanoma. Int J Cancer 1993; 54: 37–43.
Miyake M, Nakano K, Leki Y, Adachi M, Huang CL, Itoi S et al. Motility related protein 1 (MRP-1/CD9) expression: inverse correlation with metastases in breast cancer. Cancer Res 1995; 55: 4127–4131.
Adachi M, Taki T, Konishi T, Huang CI, Higashiyama M, Miyake M . Novel staging protocol for non-small-cell lung cancers according to MRP-1/CD9 and KAI1/CD82 gene expression. J Clin Oncol 1998; 16: 1397–1406.
Mori M, Mimori K, Shiraishi T, Haraguchi M, Ueo H, Barnard GF et al. Motility related protein 1 (MRP1/CD9) expression in colon cancer. Clin Cancer Res 1998; 4: 1507–1510.
Houle CD, Ding XY, Foley JF, Afshari CA, Barrett JC, Davis BJ . Loss of expression and altered localization of KAI1 and CD9 protein are associated with epithelial ovarian progression. Gynecol Oncol 2002; 86: 69–78.
Ikeyama S, Koyama M, Yamaoko M, Sasada R, Miyake M . Suppression of cell motility and metastasis by transfection with human motility-related protein (MRP-1/CD9) DNA. J Exp Med 1993; 177: 1231–1237.
Miyake M, Inufusa H, Adachi M, Ishida H, Hashida H, Tokuhara T et al. Suppression of pulmonary metastasis using adenovirally motility related protein-1 (MRP-1/CD9) gene delivery. Oncogene 2000; 19: 5221–5226.
Kubista B, Erovic BM, Klinger H, Sulzbacher I, Trieb K . CD9 expression is not a prognostic factor in human osteosarcoma. Cancer Lett 2004; 209: 105–110.
Jamil F, Peston D, Shousha S . CD9 immunohistochemical staining of breast carcinoma: unlikely to provide useful prognostic information for routine use. Histopathology 2001; 39: 572–577.
Sauer G, Windisch J, Kurzeder V, Heilmann V, Kreienberg R, Deissler H . Progression of cervical carcinoma is associated with down-regulation of CD9 but strong re-expression on sites of transendothelial invasion. Clin Cancer Res 2003; 9: 6426–6431.
Barrena S, Almeida J, Yunta M, Lopez A, Fernandez-Mosteirin N, Giralt M et al. Aberrant expression of tetraspanin molecules in B-cell chronic lymphoproliferative disorders and its correlation with normal B-cell maturation. Leukemia 2005; 19: 1376–1383.
Drucker L, Tohami T, Tartakover-Matalon S, Zismanov V, Shapiro H, Radnay J et al. Promoter hypermethylation of tetraspanin members contributes to their silencing in myeloma cell lines. Carcinogenesis 2006; 27: 197–204.
Radl J, De Glopper E, Schuit HRE . Idiopathic paraproteinemia II. Transplantation of the paraprotein-producing clone from old to young C57BL/KaLwRij mice. J Immunol 1979; 122: 609–613.
Vanderkerken K, Asosingh K, Croucher P, Van Camp B . Multiple myeloma biology: lessons from the 5TMM models. Immunol Rev 2003; 194: 196–206.
Vanderkerken K, De Raeve H, Goes E, Van Meirvenne S, Radl J, Van Riet I et al. Organ involvement and phenotypic adhesion profile of 5T2 and 5T33 myeloma cells in the C57BL/KaLwRij mouse. Br J Cancer 1997; 76: 451–460.
De Raeve H, Asosingh K, Wisse E, Van Camp B, Van Marck E, Vanderkerken K . Part of the multiple myeloma-associated microvessels is functionally connected to the systemic circulation: a study in the murine 5T33MM model. Virchows Archiv 2004; 445: 389–395.
Van Valckenborgh E, Bakkus M, Munaut C, Noel A, St Pierre Y, Asosingh K et al. Upregulation of matrix metalloproteinase-9 (MMP-9) in murine 5T33 multiple myeloma cells by interaction with bone marrow endothelial cell. Int J Cancer 2002; 101: 512–518.
Imai K, Kobayashi M, Wang J, Ohiro Y, Hamada J, Cho Y et al. Selective transendothelial migration of hematopietic progenitor cells: a role in homing of progenitor cells. Blood 1999; 93: 149–156.
Van Riet I, De Greef C, Aharchi F, Woischwill C, De Waele M, Bakkus M et al. Establishment and characterization of a human stroma-dependent myeloma cell line (MM5.1) and its stroma-independent variant (MM5.2). Leukemia 1997; 11: 284–293.
Menu E, Asosingh K, Indraccolo S, De Raeve H, Van Riet I, Van Valckenborgh E et al. The involvement of stromal derived factor 1alpha in homing and progression of multiple myeloma in the 5TMM model. Haematologica 2006; 91: 605–612.
Zhu D, van Arkel C, King CA, van Meirvenne S, de Greef C, Thielemans K et al. Immunoglobulin VH gene sequence analysis of spontaneous murine immunoglobulin-secreting B-cell tumors with clinical features of human disease. Immunology 1998; 93: 162–170.
Van Valckenborgh E, De Raeve H, Devy L, Blacher S, Munaut C, Noel A et al. Murine 5T multiple myeloma cells induce angiogenesis in vitro and in vivo. Br J Cancer 2002; 86: 796–802.
Oritani K, Wu X, Medina K, Hudson J, Miyake K, Gimble JM et al. Antibody ligation of CD9 modifies production of myeloid cells in long-term cultures. Blood 1996; 87: 2252–2261.
Oka M, Tagoku K, Russell TL, Nakano Y, Hamazaki T, Meyer EM et al. CD9 is associated with leukaemia inhibitory factor-mediated maintenance of embryonic stem cells. Mol Biol Cell 2002; 13: 1274–1281.
Tanio Y, Yamazaki H, Kunisada T, Miyake K, Hayashi SI . CD9 molecule expressed on stromal cells is involved in osteoclastogenesis. Exp Hematol 1999; 27: 853–859.
Stojanovic M, Germain M, Nguyen M, Shore GC . BAP13 and its caspase cleavage product regulate cell surface expression of the tetraspanins and integrin-mediated cell survival. J Biol Chem 2005; 280: 30018–30024.
Baudoux B, Castanares-Zapatero D, Leclerq-Smekens M, Berna N, Poumay Y . The tetraspanin CD9 associates with the integrin á6â4 in cultured human epidermal keratinocytes and is involved in cell motility. Eur J Cell Biol 2000; 79: 41–51.
Dery MA, Michaud MD, Richard DE . Hypoxia-inducible factor 1: regulation by hypoxic and non-hypoxic activators. Int J Biochem Cell Biol 2005; 37: 535–540.
Bouchel C, Benoit P, Frachet P, Billard M, Worthington RE, Gagnon J et al. Molecular cloning of the CD9 antigen. J Biol Chem 1991; 266: 117–122.
Asosingh K, De Raeve H, Van Riet I, Van Camp B, Vanderkerken K . Multiple myeloma tumor progression in the 5T2MM murine model is a multistage and dynamic process of differentiation, proliferation, invasion and apoptosis. Blood 2003; 101: 3136–3141.
Turker M . Gene silencing in mammalian cells and the spread of DNA methylation. Oncogene 2002; 21: 5388–5393.
Asosingh K, Günthert U, Bakkus MHC, De Raeve H, Goes E, Van Riet I et al. In vivo induction of insulin-like growth factor-I receptor and CD44v6 confers homing and adhesion to murine multiple myeloma cells. Cancer Res 2000; 60: 3096–3104.
Longo N, Yáñes-Mó M, Mittelbrunn M, de la Rosa G, Munoz ML, Sánchez-Madrid F et al. Regulatory role of tetraspanin CD9 in tumor–endothelial cell interaction during transendothelial invasion of melanoma cells. Blood 2001; 98: 3717–3726.
We thank T Herløv Jensen, N Arras, A Willems and C Seynaeve for expert technical assistance, Professor F Gorus (AZ VUB, Brussels) for serum paraprotein analysis, Dr M de Ridder (Kankeronderzoek, VUB, Brussels) for performing the hypoxia experiments and the Hematology and the Pathology departments of Vejle hospital for providing the human biopsies for the immunohistochemical study. The work was financially supported by the Fonds voor Wetenschappelijk Onderzoek Vlaanderen (FWO-Vl), the Stichting tegen Kanker and the Onderzoeksraad Vrije Universiteit Brussel (OZR-VUB). K Vanderkerken is a postdoctoral fellow of FWO-Vl.
About this article
Cite this article
De Bruyne, E., Andersen, T., De Raeve, H. et al. Endothelial cell-driven regulation of CD9 or motility-related protein-1 expression in multiple myeloma cells within the murine 5T33MM model and myeloma patients. Leukemia 20, 1870–1879 (2006). https://doi.org/10.1038/sj.leu.2404343
- multiple myeloma
- endothelial cells
Identification of key candidate genes and pathways in multiple myeloma by integrated bioinformatics analysis
Journal of Cellular Physiology (2019)
Cancer Biomarkers (2018)
International Journal of Molecular Sciences (2017)
Tetraspanin 7 (TSPAN7) expression is upregulated in multiple myeloma patients and inhibits myeloma tumour development in vivo
Experimental Cell Research (2015)
In vivo treatment with epigenetic modulating agents induces transcriptional alterations associated with prognosis and immunomodulation in multiple myeloma