Leukemia (2004) 18, 967–975 & 2004 Nature Publishing Group All rights reserved 0887-6924/04 $25.00 www.nature.com/leu IL-6 regulates CD44 surface expression on human myeloma cells

T Vincent1,2,3 and N Mechti1

1INSERM Unite´ U475, Montpellier Cedex, France; 2UMR-CNRS 5160, Montpellier Cedex, France; and 3Laboratoire d’Immunologie, Hoˆpital St-Eloi, Montpellier Cedex, France

Multiple myeloma (MM) is a progressive B-lineage neoplasia contains 20 : 10 standard exons (1s–10s) encode for the characterized by the accumulation of slow proliferative malig- so-called hematopoietic or standard isoform (CD44s), which is nant plasma cells in the bone marrow compartment where the microenvironment seems to be favorable for their growth and widely distributed and highly expressed on hematopoietic cells: survival. such as syndecan-1 10 variant exons (1v–10v) give rise, by , to and CD44 are thought to play a central role in the survival the formation of CD44 variant isoforms (CD44v) characterized signals provided by these bone marrow survival niches, which by the addition of extracellular membrane-proximal domains in require complex interactions between myeloma cells, extra- various combinations.22 The expression of the CD44v isoform is cellular matrix, stromal cells and soluble factors. In this report, highly restricted and associated with specific processes such as we demonstrate that interleukin-6 (IL-6), the main survival and 23 growth factor for myeloma cells, strongly increases CD44 leukocyte activation or malignant transformation. In multiple expression. In addition, we show that IL-6 modulates CD44 RNA myeloma (MM), both normal and malignant plasma cells alternative splicing and induces the overexpression of all CD44 express CD44s, while some CD44v isoforms are differentially variant exons. Finally, we show that IL-6-induced CD44 cell expressed on bone marrow plasma cells from normal indivi- surface molecules have a functional polarized membrane duals and myeloma patients.24,25 In particular, CD44v9 contain- distribution. As IL-6 secretion induced from bone marrow ing isoforms are associated with progressive disease and a short stromal cells by myeloma cells is partly mediated through overall survival, suggesting that specific expression of some direct cell-to-cell interaction involving CD44 adhesion mole- 24,26,27 cules, our findings suggest that a CD44/IL-6 amplification loop CD44v isoforms are critical for myeloma aggressiveness. plays a crucial role in myeloma cell survival. In addition, CD44v10 and CD44v6 expressions are involved in Leukemia (2004) 18, 967–975. doi:10.1038/sj.leu.2403333 the selective homing of myeloma cells in the bone marrow.28,29 Published online 11 March 2004 More recently, it was documented that abnormally low or high Keywords: myeloma; CD44; ; ; serum levels of hyaluronan (HA), the main ligand for CD44,30 hyaluronan were associated with a poor prognosis in MM.31 In addition, the increase in HA synthesis was described in bone marrow mesenchymental progenitor cells of patients with MM.32 Adhesion molecules can also modulate the secretion and Introduction biological activity of . Indeed, myeloma cell-induced IL-6 production by stromal cells requires cell-to-cell interaction Originating from a postswitch memory B cell or mediated by the cell surface molecules CD56/NCAM, VLA-4, compartment in peripheral lymphoid tissues, malignant myelo- 15,17,18 VLA-5, LFA-1 and CD44. In particular, the expression of ma cells accumulate in the bone marrow where the micro- the CD44v9 isoforms on myeloma cells seems to be required for environment seems to be favorable for their growth and survival. 24 this interaction. As this cellular contact was partly inhibited by 1–4 Interleukin-6 (IL-6), which is mainly produced by the stromal monoclonal against CD44, these data strongly environment, is the major survival and growth factor for suggest the implication of CD44 in the abnormally high levels myeloma cells, both in vitro and in vivo.5–9 Accordingly, of IL-6 found in the stromal microenvironment of MM patients myeloma cells spontaneously express IL-6 (IL-6R), and 17 with active disease. Moreover, some adhesion molecules, both IL-6 and soluble IL-6R serum levels are correlated with the especially proteoglycans with heparan sulfate (HS) side chains disease activity.10,11 Furthermore, IL-6 promotes the prolifera- like syndecan-1 and some CD44 variant isoforms, are able to tion of freshly expanded human myeloma cells in vitro,6 and bind and potentiate the biological activity of heparin binding anti-IL-6 monoclonal antibodies inhibit myeloma 33 34 12,13 such as insulin-like growth factor (IGF-1), FGF, both in vitro and in vivo. Finally, IL-6 prevents of 35 36,37 8 HGF or HB-EGF, which are survival and growth factors myeloma cells induced by serum starvation or dexamethasone, 38 for myeloma cells. Likewise, we have recently demonstrated and Fas.5 However, IL-6 alone is not sufficient to ensure that HA induces survival and proliferation of human myeloma myeloma cell survival. In addition, long-term in vitro culture of 4,39 cells by potentiating the IL-6 autocrine loop. All these data primary plasma cells requires stromal cell feeder layers for their further underline the potential role of CD44 in the physiopathol- survival, demonstrating the central role of bone marrow ogy of MM. microenvironment in myeloma cell survival.14 Indeed, it is Based on these observations and to further elucidate the now well established that the survival and proliferation of interplay between IL-6 and CD44, we analyzed the modulation myeloma cells are dependent on soluble factors and physical of CD44 expression by IL-6 using IL-6-dependent human interactions between malignant cells and both the stromal cells 40 myeloma cell lines. We demonstrate that IL-6 strongly and the bone marrow extracellular matrix (ECM).1,2–4 Different upregulates CD44 cell surface expression and modulates the adhesion molecules such as the family and CD44 have alternative splicing of CD44 mRNA. Furthermore, we show that been involved in these interactions.15–21 The CD44 gene the IL-6-induced CD44 molecules retain their functional polarized membrane expression. Altogether, our findings Correspondence: N Mechti, UMR-CNRS 5160, 240 Avenue Emile provide new insights into the complex interplay between Jeanbrau, 34094 Montpellier Cedex 5, France; Fax: þ 334675218 29; E-mail: [email protected] myeloma cells and their microenvironment providing survival Received 13 October 2003; accepted 30 January 2004; Published niches in the bone marrow, and further underline the relevance online 11 March 2004 of IL-6-targeted therapeutics in MM. CD44 modulation by IL-6 T Vincent and N Mechti 968 Materials and methods RT-PCR analysis

Cell cultures Total RNAs were isolated as previously described43 and cDNAs were prepared from 1 mg of RNA by using Superscript reverse XG–2, XG–6, XG-7, U266 and RPMI8226 human myeloma cells transcriptase (Life Technologies, Pontoise, France) according to were obtained from Dr B Klein (Montpellier, France). The XG the manufacturer’s instructions. For a semiquantitative ap- cell lines, whose survival and growth are completely dependent proach, CD44 cDNAs were amplified by polymerase chain on addition of exogenous IL-6, were cultured in the presence of reaction (PCR), using the forward primer GCACAGACA- 2 ng/ml of recombinant human IL-6 (Sandoz, Vienna, Austria) in GAATCCCTGCTACC located in 5 and the following RPMI 1640 medium supplemented with 10% fetal bovine serum reverse primers located in the different variable exons (v): 2v (FBS). The IL-6-independent U266 and RPMI8226 cell lines (exon 6) CAGCCATTTGTGTTGTTGTG; 3v (exon 7) were cultured in the same medium without IL-6. TGGTGCTGGAGATAAAATCT; 4v (exon 8) CAGT- CATCCTTGTGGTTGTC; 5v (exon 9) TTGTGCTTGTA- GAATGTGGG; 6v (exon 10) CAGCTGTCCCTGTTGTCGAA; 7v (exon 11) CCATCCTTCTTCCTGCTTGA; 8v (exon 12) Antibodies GCGTTGTCATTGAAAGAGGT; 9v (exon 13) TGCTTGATGT- CAGAGTAGAA; 10v (exon 14) CTGATAAGGAACGATTGACA, Fluorescein-conjugated monoclonal anti-CD44 listed 50 to 30 as previously described.44 For global CD44 (CD44-FITC, clone J-173) was purchased from Immunotech, isoforms expression analysis, the cDNAs were amplified using France. Biotinylated monoclonal anti-syndecan-1 antibody the forward primer GCACAGACAGAATCCCTGCTACC located (Mi15-biot) was obtained from Dr B Klein. in exon 5 and reverse primer CCAAGATGATCAGCCATTCTGG located in exon 17. GAPDH used as an invariant control was PCR amplified with the forward (TGAAGGTCGGAGTCAACG- Detection of apoptotic cells GATTTGGT) and reverse: (CATGTGGGCCATGAGGTCCAC- CAC) primers. A total of 23 cycles of PCR were realized in the Apoptotic cells were detected using the fluorescein isothiocya- presence of [32P]-dCTP, and the PCR products were analyzed on nate-labelled annexin V method (FITC-annexin-V, Boehringer a sequencing gel prior to autoradiography. Mannheim). Annexin V has a high affinity for phosphatidylserine present on the outer cytoplasmic membrane of apoptotic cells.41 Cells were washed, labelled with Annexin-V-Fluos according to Statistical analysis the manufacturer’s recommendations and analyzed by flow cytometry. The statistical significance of the data was evaluated by using Student’s t-test for pairs.

Flow cytometry analysis Results

The expression of CD44 molecules on living cells was IL-6 controls CD44 cell surface expression on human quantified by double immunofluorescence staining using myeloma cells CD44-FITC and Mi15-biot monoclonal antibodies (mAbs). A total of 5 Â 105 cells were washed twice with phosphate- To investigate the effect of IL-6 on CD44 membrane expression, buffered saline (PBS) supplemented with 1% (v/v) FBS. The cells XG-2 and XG-6 human myeloma cell lines, whose survival and were resuspended in 50 ml of PBS–1% FBS containing CD44- proliferation are dependent on the addition of exogenous IL-6,40 FITC and Mi15-biot mAbs, and incubated for 45 min at 41C. The were starved of IL-6 and then cultured with or without IL-6 for 3 cells were then washed twice in PBS, and incubated for 45 min days. CD44 cell surface expression was assessed by flow with streptavidin coupled to the peroxidase. After two additional cytometry analysis with a double staining method using anti- washings, the cells were resuspended in 400 ml of PBS and CD44 mAb conjugated to fluorescein (CD44-FITC) and phy- fluorescence analysis was performed with a FACScan fluores- coerythrin-conjugated anti-syndecan-1 mAb (MI15-biot). As cence activated cell sorter (Becton Dickinson), as previously syndecan-1 antigen is rapidly lost by myeloma cells undergoing described.39 The cell preparations were analyzed by size, and apoptosis,45 this approach allowed us to determine the level of 104 cells were evaluated for the percentage of positive cells and CD44 membrane expression on living myeloma cells (synde- their fluorescence intensity. The binding of HA on myeloma can-1 positive cells), and to exclude cells that have been cells was quantified by direct immunofluorescence staining engaged throughout a cell death program (syndecan-1 negative using a fluorescein-conjugated anti-CD44 mAb obtained from cells). Indeed, it is well established that apoptotic processes are Dr M-S Sy.42 associated with a reduction in cell surface expression of some molecules.46,47 A typical experiment performed the XG-6 cell line is presented in Figure 1a. CD44 cell surface expression, Immunofluorescence analysis evaluated by the mean of fluorescence intensity (MFI) on living cells (upper right quadrant in each dot plot), was strongly Cells were collected by centrifugation, resuspended in PBS, reduced in the absence of IL-6 (MFI ¼ 162) vs presence of IL-6 plated on polylysine-coated slides, and then fixed 5 min in PBS (MFI ¼ 567). This experiment was reproduced several times on containing 3.7% formaldehyde. CD44 was detected by direct XG-6 and XG-2 cell lines with average reductions of 71.3% and immunofluorescence using the CD44-FITC mAb (dilution 41.6%, respectively (Figure 1b). The statistical significance of 1:200). Slides were viewed using a Leika microscope and image these experiments is presented in Figure 1b. Time course files were processed with the Adobe Photoshop program. analysis performed on the XG-6 cell line revealed that CD44

Leukemia CD44 modulation by IL-6 T Vincent and N Mechti 969

Figure 1 IL-6 controls CD44 cell surface expression on human myeloma cells. XG-2 and XG-6 IL-6-dependent cell lines starved of IL-6 were cultured with ( þ IL-6) or without (ÀIL-6) for 72 h. CD44 cell surface expression on living cells was assessed by flow cytometry by a double staining method using the CD44-FITC and Mi15-biot antibodies. The cell preparations were analyzed by size and 104 cells were evaluated for the percentage of positive cells and their fluorescence intensity. (a) Flow dot plot analysis of the XG-6 cell line is shown. The abscissa represents the fluorescence intensity of CD44-FITC staining and the ordinate represents Mi15-biot staining. CD44 positive living cells are located in the upper right quadrant in each dot plot. (b) The experiment was reproduced several times on XG-6 and XG-2 cell lines. The histograms represent the relative means of CD44 fluorescence intensity for each cell line in the presence (fill black) or absence (fill gray) of IL-6. The standard deviations were determined for three independent experiments. In the absence of IL-6, CD44 expression was significantly reduced as determined by a Student t-test for pairs (P-values were considered as statistically significant if Po0.05).

expression was promptly downregulated (28% reduction after IL-6 controls CD44 expression at the transcriptional 8 h of depletion) in a time-dependent manner (Figure 2a). This level mechanism was reversible because the level of CD44 expres- sion was completely and rapidly restored when IL-6 was To determine whether the regulation of CD44 cell surface reintroduced into the culture medium (Figure 2b). Interestingly, expression was associated with an alteration of CD44 gene in U266 and XG-7 cell lines, which the survival is, respectively, expression, the level of CD44 mRNA was monitored by entirely and partly IL-6 independent, a less important but semiquantitive RT-PCR analysis. Using glyceraldehyde-3-phos- significant decrease in CD44 expression was observed (27.4% phate dehydrogenase (GAPDH) as an invariant control, equal and 27.3%, respectively), suggesting that the regulation of CD44 amounts of cDNA were used for the analysis of CD44 mRNA. As and myeloma cell survival by IL-6 used distinct pathways (data shown in Figure 4, the amount of CD44 mRNA was strongly not shown). reduced in XG-6 cells starved of IL-6. RPMI8226 myeloma cell To completely ascertain that the reduction of CD44 expres- line, which does not express CD44, was used as a negative sion observed in IL-6-starved myeloma cells did not result from control. These data demonstrated that IL-6 upregulates CD44 early apoptotic processes, we used the ability of human alpha expression at the transcriptional level. interferon (Hu-a2aIFN) to antagonize apoptosis induced in IL-6- dependent myeloma cells by the removal of IL-6.48 XG-6 cells were starved of IL-6 and then cultured 3 days in the presence of IL-6-induced CD44 molecules have a polarized 200 UI/ml of Hu-a2aIFN. The percentage of apoptotic cells and membrane organization the level of CD44 cell surface expression were quantified, as previously described. Although apoptosis induced by IL-6 A proportion of myeloma cells are polarized with a uropod on starvation was strongly antagonized by Hu-a2aIFN (Figure 3a), one end and a lamelipodia on the other end.49 Several cell the reduction of CD44 expression was similar in the presence or surface molecules such as syndecan-1, CD54 and CD44 absence of Hu-a2aIFN (Figure 3b). In addition, no significant localize specifically to the uropod where they promote modification in the expression of CD44 by Hu-a2aIFN was adhesion, intercellular communication, and concentration and observed on cells cultured in the presence of IL-6 (data not presentation of heparin binding cytokines.49 As polarized shown). Altogether, these data clearly demonstrated that the organization of heparan sulfate proteoglycans (HSPGs) seems upregulation of CD44 expression by IL-6 on myeloma cells is to play a major role in cellular communication and signalling, independent of apoptotic processes. we evaluated whether IL-6 could influence the polarized

Leukemia CD44 modulation by IL-6 T Vincent and N Mechti 970 a 100 a 45 90 40 80 35 70 30 60 25 20 50 15 40 10 30 % of apoptotic cells 5 20 % of CD44 expression 0 10 08724 0 time (hours) 04 8 24 48 72 time (hours) b 120 100 b 140 80 120 60 100 40 80 IL-6 20

60 % of CD44 expression 0 40 04872 time (hours)

% of CD44 expression 20 Figure 3 a-interferon antagonizes apoptosis but not CD44 down 0 regulation induced by IL-6 starvation. XG-6 cells were cultured for the 024487296 indicated times (0, 48, 72 h) in the absence (fill black) or presence (fill time (hours) gray) of 200 UI/ml of Hu-a2aIFN. (a) Apoptotic cells were detected by flow cytometry using the annexin V-staining method. The histograms Figure 2 Time-course expression of CD44 on XG-6 cell line. (a) represent the percentage of apoptotic cells determined for each XG-6 cells were cultured without IL-6 and the cell surface expression experimental condition. (b) CD44 cell surface expression was assessed was determined at the indicated times (0, 4, 8, 24, 48, 72 h) as by flow cytometry using the CD44-FITC antibody. The cell prepara- described in Figure 1b. (b) IL-6 was reintroduced in the culture tions were analyzed by size and 104 cells were evaluated for the medium at 48 h. percentage of positive cells and their fluorescence intensity. The histograms represent the relative means of CD44 fluorescence intensity for each experimental condition. membrane organization of CD44 cell surface molecules. To this aim XG-6 cells were starved of IL-6 and then cultured for 3 days in the presence or absence of exogenous IL-6. Cell surface CD44 distribution was analyzed by direct immunofluorescence with a fluorescein-conjugated anti-CD44 antibody, as described in expression on XG-2 and XG-7 cells in similar proportions, Materials and methods. We observed that, in the absence of IL- whereas it failed to increase binding of HA on XG-6 and U266 6, XG-6 cells exhibited a polarized phenotype with a striking cell lines. RPMI8226, which does not express CD44 and does localization of CD44 molecules on membrane protrusions not bind HA-FITC, was used as a negative control (data not (Figure 5). This membrane distribution was not altered during shown). Taken together, these data demonstrated that IL-6 did IL-6-induction of CD44 expression. not influence HA binding on human myeloma cells.

IL-6 does not modulate the binding of HA to CD44 IL-6 modulates CD44 alternative splicing

Although CD44 is widely expressed, only a few cells bind HA In addition to the CD44 standard molecule (CD44s), alternative constitutively.50.In addition, binding of HA to CD44 can be splicing of one or more variable exons (1v–10v) results in the altered by , inside/out signalling and potential creation of a large number of CD44v molecules that extracellular stimuli.51 Previous studies have identified TNFa as are involved in the physiopathology of MM.25–29 We decided to a positive regulator, and IL-4 and IL-13 as negative regulators of investigate whether the expression of the different CD44v HA binding to CD44.50,52,53 To determine whether IL-6 could isoforms could be influenced by IL-6. Using specific primers regulate CD44 affinity to HA, we compared HA binding and for each variant exon, the mRNA of the different CD44v CD44 expression on several myeloma cell lines cultured in the isoforms was analyzed by semiquantitive RT-PCR on XG-6 cells presence or absence of exogenous IL-6. HA binding was cultured in the presence or absence of exogenous IL-6 for 48 h. quantified by direct immunofluorescence staining using fluor- GAPDH was used as an invariant control to ensure that equal escein-conjugated HA (HA-FITC), as previously described.4 As amounts of cDNA were used. As shown in Figure 7, all variant shown in Figure 6, no correlation was observed between the exons were markedly overexpressed in the presence of IL-6. level of CD44 expression and the ability of cells to bind HA. Some isoforms appeared only expressed in the presence of IL-6. Indeed, XG-6 and U266 cell lines that express a high level of Conversely, especially in exons v5 and v8 containing isoforms, CD44 molecules did not bind, or bound weakly to HA-FITC. some CD44 variants were overexpressed in the absence of Conversely, XG-2 and XG-7 cell lines strongly bound HA-FITC. exogenous IL-6. Clearly, IL-6 not only upregulated CD44 gene Interestingly, IL-6 increased both binding of HA and CD44 expression but also strongly modulated the alternative splicing

Leukemia CD44 modulation by IL-6 T Vincent and N Mechti 971 of all variant exons. The RPMI8226 cell line was used as a Discussion negative control (data not shown). Some recent reports point to the existence of long life plasma cells that have the capacity to live for unlimited time if rescued by specific factors provided in survival niches.54 Like their RPMI physiological counterpart, malignant plasma cells specifically XG6 8226 localize in the bone marrow of patients with MM where the IL-6 + - - microenvironment seems to be favorable for their growth and survival.1–4,39 Survival signals provided in these niches require 650 complex interactions between myeloma cells, bone marrow stromal cells, ECM and soluble factors. In particular, IL-6 500 secretion induced from bone marrow stromal cells by myeloma cells is mediated through direct cell-to-cell contacts involving 400 adhesion molecules such as and CD44.15,17,18 In this report, we show that IL-6 strongly induces CD44 cell surface expression on myeloma cells. These data suggest the existence 300 of a CD44/IL-6 amplification loop in which stromal cells induce CD44 expression on myeloma cells through IL-6 and, in turn, myeloma cells induce IL-6 secretion by stromal cells through CD44 molecules. Furthermore, it has been reported that IL-6

200

GAPDH

Figure 4 Quantification of CD44 gene expression in human myeloma cells by semi-quantitative RT-PCR. XG-6 and RPMI8226 cells were cultured for 72 h in the presence ( þ ) or absence (À) of IL-6. Total RNAs were isolated for each experimental condition and cDNAs Figure 5 Polarized organization of CD44 cell surface . XG- were prepared as described in Materials and methods, and amplified 6 cells were cultured for 72 h in the presence or absence of IL-6. CD44 by PCR, using specific CD44 (exon 5 and exon 17) and GAPDH cell surface expression was detected by direct immunofluorescence primers. A total of 23 cycles of PCR were performed in the presence of with CD44-FITC mAb. Slides were viewed using a Leika microscope [32P]-dCTP, and the PCR products were fractionated on a sequencing and image files were processed with the Adobe Photoshop program. gel prior to autoradiography.

XG2 XG6 1200 100 1200 100 1000 cells 80 g 1000 80 800 800 60 60 600 600 40 400 40 400 20 200 20 200 0 0 0 0 % of HA bindin CD44 expression (MFI) % of HA binding cells CD44 expression (MFI) -IL6 +IL6 -IL6 +IL6 XG7 U266

1200 100 1200 100 1000 80 1000 80 800 800 60 60 600 600 40 40 400 400 200 20 200 20 0 0 0 0 % of HA binding cells % of HA binding cells CD44 expression (MFI) -IL6 +IL6 CD44 expression (MFI) -IL6 +IL6

Figure 6 Binding of HA on human myeloma cell. The XG-2, XG-6, XG-7 and U266 cells were cultured for 72 h in the presence ( þ IL-6)or absence (ÀIL-6) of IL-6. The cells were then washed twice with PBS and incubated for 45 min on ice with either HA-FITC or CD44-FITC antibody. Fluorescence analysis was performed with a FACScan fluorescence activated cell sorter (Becton Dickinson). The cell preparations were analyzed by size and 104 cells were evaluated for the percentage of positive cells and their MFI. The histograms represent the means of CD44 fluorescence intensity (fill gray) for each cell line analyzed. The percentage of HA-binding cells is indicated by ~.

Leukemia CD44 modulation by IL-6 T Vincent and N Mechti 972

Figure 7 Analysis of the different CD44 variant isoforms expressed on XG-6 myeloma cells. XG-6 cells were cultured for 72 h in the presence ( þ ) or absence (À) of IL-6. Total RNAs were isolated for each experimental condition and cDNAs were prepared and amplified by PCR, using specific CD44 primers located in the different variant exons (v2–v10), as described in Materials and methods. A total of 23 cycles of PCR were performed in the presence of [32P]-dCTP, and the PCR products were fractionated on a sequencing gel prior to autoradiography. The molecular weights are indicated.

binds soluble HS, suggesting that it can bind HS side chains of stromal cell feeder layers for their survival.14 Recently, some proteoglycans like syndecan-1 and CD44.55–57 Such a glucocorticoid-induced drug resistance of RPMI8226 human binding is proposed to potentiate its biological activity by myeloma cells was correlated with an increased expression of sequestrating and concentrating IL-6 near its site of secretion, the a-isoform of the glucocorticoid receptors, and specific and thus favoring its autocrine or paracrine activity.38 HS modulations in CD23, CD38, CD44 and CD58 cell surface proteoglycans can also act as coreceptors for growth factors by expressions.62 In particular, the prednisone and dexamethasone- the formation of a ternary complex mediated by specific resistant cells strongly overexpressed CD44, suggesting that interactions of the HS side chain with both the growth factor CD44 could contribute in the development of drug resistance. and its receptor, resulting in an increase in receptor dimerization Myeloma cells are characterized by a polarized membrane and cell signalling.35,38,58,59 This process is enhanced by the organization shown to be crucial for both and cell- expression of CD44 variants, especially CD44v3 or CD44v6 to-cell communication. In addition, the migratory potential of isoforms, which contain additional domains that support the myeloma cell lines was correlated to their high number of covalent attachment of side chains.60,61 The polarized cells.49 Interestingly, most HS proteoglycans are fact that IL-6 induces the overexpression of all CD44 variant localized to the uropod of myeloma cells.49 This cell surface exons on myeloma cells provides a higher level of interaction organization leads to the concentration of HBPs to this pole and between soluble factors and adhesion molecules by increasing favors their presentation to target cells. The fact that IL-6- the number of potential heparan binding proteins (HBPs) induced CD44 molecules have a functional polarized mem- binding sites. Other variant exons can play an important role brane organization indicates that IL-6, by increasing the HBPs in the physiopathology of MM. In particular, CD44v10 and binding potential of uropod, might potentiate the biological CD44v6 isoforms have been involved in the adhesive interac- activity of HBPs such as IGF-1,33 FGF,34 HGF35 or HB-EGF,36,37 tion between myeloma cells and the bone marrow endothelial which are survival and growth factors for myeloma cells.38 In and stromal cells, respectively, leading to the selective homing the same way, it has recently been shown that some CD44 of myeloma cells to the bone marrow compartment.28,29 isoforms can act as a cell surface docking receptor for active CD44v9 is involved in IL-6 secretion induced from bone -9, and promote ECM degradation.63,64 marrow stromal cells by adhesion to myeloma cells,25 and its Retention of matrix metalloproteinases on the cell surface by overexpression is associated with progressive diseases and a polarized CD44 molecules leads to the concentration of short overall survival in MM.26,27 These data and our report proteolytic activity at the point of contact between the cells strengthen the idea that IL-6-induced overexpression of CD44 and the ECM, and provides a mechanism for CD44-mediated variant exons might be critical for in vivo myeloma cell growth tumor cell invasion. and survival, in particular in the earlier stage of the disease, HA is thought to play a role in the physiopathology of where long-term in vitro culture of primary plasma cells requires MM.4,31,32,39 Nevertheless, in accordance with previous studies,

Leukemia CD44 modulation by IL-6 T Vincent and N Mechti 973 we did not observe any correlation between CD44 membrane 3 Shain KH, Landowski TH, Dalton WS. The tumor microenviron- expression and the binding of HA. Particularly, XG-6 and U266 ment as a determinant of survival: a possible cell lines, which highly expressed CD44, did not bind HA-FITC mechanism for de novo drug resistance. Curr Opin Oncol 2000; efficiently. In addition, IL-6 did not increase CD44/HA affinity 12: 557–563. 4 Vincent T, Jourdan M, Sy MS, Klein B, Mechti N. significantly, and failed to induce HA binding on XG-6 and induces survival and proliferation of human myeloma cells through U266 cell lines. Accordingly, we have previously shown that an interleukin-6-mediated pathway involving the phosphorylation HA acts as a growth and survival factor for human myeloma of retinoblastoma protein. J Biol Chem 2001; 276: 14728–14736. cells through a CD44-independent pathway.4,39 Indeed, HA is 5 Chauhan D, Kharbanda S, Ogata A, Urashima M, Teoh G, able to potentiate autocrine and/or paracrine activities Robertson M et al. Interleukin-6 inhibits Fas-induced apoptosis very similarly to HSPG side chains. So, it is tempting to and stress-activated protein kinase activation in multiple myeloma cells. Blood 1997; 89: 227–234. speculate that HSPGs and HA act synergistically by concentrat- 6 Klein B, Zhang XG, Jourdan M, Boiron JM, Portier M, Lu ZY et al. ing, protecting and presenting different active molecules. In Interleukin-6 is the central tumor growth factor in vitro and in vivo other pathological models, some growth factors such as the IGF- in multiple myeloma. Eur Cytokine Netw 1990; 1: 193–201. 1 were reported to increase CD44 expression.44,65,66 Interest- 7 Klein B, Zhang XG, Lu ZY, Bataille R. Interleukin-6 in human ingly, in vivo induction of IGF-1 receptor and CD44v6 multiple myeloma. Blood 1995; 85: 863–872. expressions induced on myeloma by the bone marrow micro- 8 Lichtenstein A, Tu Y, Fady C, Vescio R, Berenson J. Interleukin-6 inhibits apoptosis of malignant plasma cells. Cell Immunol 1995; environment have been shown to confer homing and adhesion 28 162: 248–255. to murine myeloma cells. As IGF-1 is a growth and survival 9 Treon SP, Anderson KC. Interleukin-6 in multiple myeloma and 67 33 factor in MM and binds HS, these data strongly suggest that related plasma cell dyscrasias. Curr Opin Hematol 1998; 5: 42–48. other growth factors or cytokines may act synergistically with IL- 10 Barille S, Thabard W, Robillard N, Moreau P, Pineau D, 6 to amplify the CD44/IL-6 loop. Further investigations are Harousseau JL et al. CD130 rather than CD126 expression is needed to clarify this point. associated with disease activity in multiple myeloma. Br J IL-6 has been described as the major survival and growth Haematol 1999; 106: 532–535. 11 Bataille R, Jourdan M, Zhang XG, Klein B. Serum levels of factor for human myeloma cells, both in vitro and in vivo. The interleukin 6, a potent myeloma cell growth factor, as a reflect of demonstration that IL-6 can modulate expression and alternative disease severity in plasma cell dyscrasias. J Clin Invest 1989; 84: splicing of CD44 in malignant plasma cells provides new 2008–2011. insights on the pleiotropic activities of this cytokine and further 12 Bataille R, Barlogie B, Lu ZY, Rossi JF, Lavabre-Bertrand T, Beck T underline the relevance of IL-6-targeted therapeutics in MM. et al. Biologic effects of anti-interleukin-6 murine monoclonal Alternatively, our finding on the existence of a CD44/IL-6 antibody in advanced multiple myeloma. Blood 1995; 86: 685–691. amplification loop may provide the rationale for the develop- 13 Klein B, Wijdenes J, Zhang XG, Jourdan M, Boiron JM, Brochier J ment of therapies targeting CD44. Indeed, the ligation of CD44 et al. Murine anti-interleukin-6 therapy for a with specific mAbs inhibits the proliferation of acute myeloid patient with plasma cell leukemia. Blood 1991; 78: 1198–1204. leukemia cells by stabilizing the cyclin-dependent kinase 14 Degrassi A, Hilbert DM, Rudikoff S, Anderson AO, Potter M, Coon inhibitor p27(Kip1) protein, resulting in inhibition of the cyclin HG. In vitro culture of primary plasmacytomas requires stromal E/Cdk2 kinase activity.68 Accordingly, we have previously cell feeder layers. Proc Natl Acad Sci USA 1993; 90: 2060–2064. reported that a high level of p27(Kip1) protein expression was 15 Uchiyama H, Barut BA, Mohrbacher AF, Chauhan D, Anderson 4 KC. Adhesion of human myeloma-derived cell lines to bone associated with the inhibition of myeloma cells proliferation. marrow stromal cells stimulates interleukin-6 secretion. Blood The fact that some CD44v isoforms are preferentially expressed 1993; 82: 3712–3720. on tumor cells clearly provides a most selective alternative to 16 Chauhan D, Uchiyama H, Akbarali Y, Urashima M, Yamamoto K, target tumor cells. For example, the anti-CD44v6 antibodies Libermann TA et al. Multiple myeloma cell adhesion-induced have been successfully used to block tumor growth and interleukin-6 expression in bone marrow stromal cells involves of an adenocarcinoma in an animal model.69,70 In activation of NF-kappa B. Blood 1996; 87: 1104–1112. 17 Lokhorst HM, Lamme T, de Smet M, Klein S, de Weger RA, van MM, the involvement of CD44v6 and CD44v10 in the homing Oers R et al. Primary tumor cells of myeloma patients induce of myeloma cells in the bone marrow, and the involvement of interleukin-6 secretion in long-term bone marrow cultures. Blood CD44v9 in the CD44/IL-6 amplification loop, provide interest- 1994; 84: 2269–2277. ing targets for the development of new therapeutics. Finally, our 18 Barille S, Collette M, Bataille R, Amiot M. Myeloma cells findings highlight the complex interplay between myeloma upregulate interleukin-6 secretion in osteoblastic cells through cells, bone marrow stromal cells, ECM and soluble factors in cell-to-cell contact but downregulate osteocalcin. Blood 1995; 86: order to provide an optimal microenvironment for myeloma cell 3151–3159. 19 Shain KH, Landowski TH, Dalton WS. Adhesion-mediated growth and survival in bone marrow survival niches. intracellular redistribution of c-Fas-associated death domain-like IL-1-converting enzyme-like inhibitory protein-long confers resis- Acknowledgements tance to CD95-induced apoptosis in hematopoietic cancer cell lines. J Immunol 2002; 168: 2544–2553. 20 Damiano JS, Dalton WS. Integrin-mediated drug resistance in This work was supported by grants from the Association pour la multiple myeloma. Leukemia Lymphoma 2000; 38: 71–81. Recherche contre le Cancer, the Fe´de´ration des Centres de Lutte 21 Damiano JS, Cress AE, Hazlehurst LA, Shtil AA, Dalton WS. Cell contre le Cancer Comite´s de l’He´rault et du Gard and the adhesion mediated drug resistance (CAM-DR): role of integrins and Fondation pour la Recherche Medicale. resistance to apoptosis in human myeloma cell lines. Blood 1999; 93: 1658–1667. 22 Gunthert U. CD44: a multitude of isoforms with diverse functions. References Curr Top Microbiol Immunol 1993; 184: 47–63. 23 Borland G, Ross JA, Guy K. Forms and functions of CD44. 1 Anderson KC, Kyle RA, Dalton WS, Landowski T, Shain K, Jove R Immunology 1998; 93: 139–148. et al. Multiple myeloma: new insights and therapeutic approaches. 24 Van Driel M, Gunthert U, Stauder R, Joling P, Lokhorst HM, Bloem Hematology (Am Soc Hematol Educ Program) 2000, 147–165. AC. CD44 isoforms distinguish between bone marrow plasma cells 2 Bataille R, Harousseau JL. Multiple myeloma. N Engl J Med 1997; from normal individuals and patients with multiple myeloma at 336: 1657–1664. different stages of disease. Leukemia 1998; 12: 1821–1828.

Leukemia CD44 modulation by IL-6 T Vincent and N Mechti 974 25 Van Driel M, Gunthert U, van Kessel AC, Joling P, Stauder R, 44 Fichter M, Hinrichs R, Eissner G, Scheffer B, Classen S, Ueffing M. Lokhorst HM et al. CD44 variant isoforms are involved in plasma Expression of CD44 isoforms in neuroblastoma cells is regulated cell adhesion to bone marrow stromal cells. Leukemia 2002; 16: by PI 3-kinase and protein kinase C. Oncogene 1997; 14: 135–143. 2817–2824. 26 Eisterer W, Bechter O, Hilbe W, van Driel M, Lokhorst HM, Thaler 45 Jourdan M, Ferlin M, Legouffe E, Horvathova M, Liautard J, Rossi JF J et al. CD44 isoforms are differentially regulated in plasma cell et al. The myeloma cell antigen syndecan-1 is lost by apoptotic dyscrasias and CD44v9 represents a new independent prognostic myeloma cells. Br J Haematol 1998; 100: 637–646. parameter in multiple myeloma. Leuk Res 2001; 25: 1051–1057. 46 Homburg CH, de Haas M, von dem Borne AE, Verhoeven AJ, 27 Stauder R, Van Driel M, Schwarzler C, Thaler J, Lokhorst HM, Reutelingsperger CP, Roos D. Human neutrophils lose their surface Kreuser ED et al. Different CD44 splicing patterns define Fc gamma RIII and acquire Annexin V binding sites during prognostic subgroups in multiple myeloma. Blood 1996; 88: apoptosis in vitro. Blood 1995; 85: 532–540. 3101–3108. 47 Greenstein S, Barnard J, Zhou K, Fong M, Hendey B. Fas activation 28 Asosingh K, Gunthert U, Bakkus MH, De Raeve H, Goes E, Van reduces neutrophil adhesion to endothelial cells. J Leukoc Biol Riet I et al. In vivo induction of insulin-like growth factor-I receptor 2000; 68: 715–722. and CD44v6 confers homing and adhesion to murine multiple 48 Ferlin-Bezombes M, Jourdan M, Liautard J, Brochier J, Rossi JF, myeloma cells. Cancer Res 2000; 60: 3096–3104. Klein B. IFN-alpha is a survival factor for human myeloma cells 29 Asosingh K, Gunthert U, De Raeve H, Van Riet I, Van Camp B, and reduces dexamethasone-induced apoptosis. J Immunol 1998; Vanderkerken K. A unique pathway in the homing of murine 161: 2692–2699. multiple myeloma cells: CD44v10 mediates binding to bone 49 Borset M, Hjertner O, Yaccoby S, Epstein J, Sanderson RD. marrow endothelium. Cancer Res 2001; 61: 2862–2865. Syndecan-1 is targeted to the uropods of polarized myeloma cells 30 Aruffo A, Stamenkovic I, Melnick M, Underhill CB, Seed B. CD44 where it promotes adhesion and sequesters heparin-binding is the principal cell surface receptor for hyaluronate. Cell 1990; proteins. Blood 2000; 96: 2528–2536. 61: 1303–1313. 50 Levesque MC, Haynes BF. Cytokine induction of the ability of 31 Dahl IM, Turesson I, Holmberg E, Lilja K. Serum hyaluronan in human monocyte CD44 to bind hyaluronan is mediated primarily patients with multiple myeloma: correlation with survival and Ig by TNF-alpha and is inhibited by IL-4 and IL-13. J Immunol 1997; concentration. Blood 1999; 93: 4144–4148. 159: 6184–6194. 32 Calabro A, Oken MM, Hascall VC, Masellis AM. Characterization 51 Lesley J, Hyman R, Kincade PW. CD44 and its interaction with of hyaluronan synthase expression and hyaluronan synthesis in extracellular matrix. Adv Immunol 1993; 54: 271–335. bone marrow mesenchymal progenitor cells: predominant expres- 52 Maiti A, Maki G, Johnson P. TNF-alpha induction of CD44- sion of HAS1 mRNA and up-regulated hyaluronan synthesis in mediated leukocyte adhesion by sulfation. Science 1998; 282: bone marrow cells derived from multiple myeloma patients. PG- 941–943. 2578-85. Blood 2002; 100: 2578–2585. 53 Levesque MC, Haynes BF. TNFalpha and IL-4 regulation of 33 Arai T, Parker A, Busby Jr W, Clemmons DR. Heparin, heparan hyaluronan binding to monocyte CD44 involves posttranslational sulfate, and dermatan sulfate regulate formation of the insulin-like modification of CD44. Cell Immunol 1999; 193: 209–218. growth factor-I and insulin-like growth factor-binding protein 54 Manz RA, Radbruch A. Plasma cells for a lifetime? Eur J Immunol complexes. J Biol Chem 1994; 269: 20388–20393. 2002; 32: 923–927. 34 Steinfeld R, Van Den Berghe H, David G. Stimulation of fibroblast 55 Ramsden L, Rider CC. Selective and differential binding of growth factor receptor-1 occupancy and signaling by cell surface- interleukin (IL)-1 alpha, IL-1 beta, IL-2 and IL-6 to glycosamino- associated syndecans and . J Cell Biol 1996; 133: glycans. Eur J Immunol 1992; 22: 3027–3031. 405–416. 56 Han ZC, Bellucci S, Shen ZX, Maffrand JP, Pascal M, Petitou M 35 Derksen PW, Keehnen RM, Evers LM, van Oers MH, Spaargaren et al. enhance megakaryocytopoiesis by M, Pals ST. Cell surface syndecan-1 mediates modifying the activities of hematopoietic growth regulators. J Cell hepatocyte growth factor binding and promotes Met signaling in Physiol 1996; 168: 97–104. multiple myeloma. Blood 2002; 99: 1405–1410. 57 Gupta P, McCarthy JB, Verfaillie CM. Stromal fibroblast heparan 36 De Vos J, Couderc G, Tarte K, Jourdan M, Requirand G, Delteil sulfate is required for cytokine-mediated ex vivo maintenance of MC et al. Identifying intercellular signaling expressed in human long-term culture-initiating cells. Blood 1996; 87: malignant plasma cells by using complementary DNA arrays. 3229–3236. Blood 2001; 98: 771–780. 58 van der Voort R, Taher TE, Wielenga VJ, Spaargaren M, Prevo R, 37 Kleeff J, Ishiwata T, Kumbasar A, Friess H, Buchler MW, Lander Smit L et al. Heparan sulfate-modified CD44 promotes hepatocyte AD et al. The cell-surface heparan sulfate proteoglycan glypican-1 growth factor/scatter factor-induced through regulates growth factor action in pancreatic carcinoma cells and is the receptor tyrosine kinase c-Met. J Biol Chem 1999; 274: 6499– overexpressed in human pancreatic cancer. J Clin Invest 1998; 6506. 102: 1662–1673. 59 van der Voort R, Keehnen RM, Beuling EA, Spaargaren M, Pals ST. 38 Turnbull J, Powell A, Guimond S. Heparan sulfate: decoding a Regulation of cytokine signaling by B cell antigen receptor and dynamic multifunctional cell regulator. Trends Cell Biol 2001; 11: CD40-controlled expression of heparan sulfate proteoglycans. 75–82. J Exp Med 2000; 192: 1115–1124. 39 Vincent T, Molina L, Espert L, Mechti N. Hyaluronan, a major non- 60 Barbour AP, Reeder JA, Walsh MD, Fawcett J, Antalis TM, Gotley protein glycosaminoglycan component of the extracellular matrix DC. Expression of the CD44v2-10 isoform confers a metastatic in human bone marrow, mediates dexamethasone resistance in phenotype: importance of the heparan sulfate attachment site multiple myeloma. Br J Haematol 2003; 121: 259–269. CD44v3. Cancer Res 2003; 63: 887–892. 40 Zhang XG, Gaillard JP, Robillard N, Lu ZY, Gu ZJ, Jourdan M et al. 61 Orian-Rousseau V, Chen L, Sleeman JP, Herrlich P, Ponta H. CD44 Reproducible obtaining of human myeloma cell lines as a model is required for two consecutive steps in HGF/c-Met signaling. for tumor stem cell study in human multiple myeloma. Blood Genes Dev 2002; 16: 3074–3086. 1994; 83: 3654–3663. 62 Genty V, Dine G, Dufer J. Phenotypical alterations induced by 41 Vermes I, Haanen C, Steffens-Nakken H, Reutelingsperger C. A glucocorticoids resistance in RPMI 8226 human myeloma cells. novel assay for apoptosis. Flow cytometric detection of phospha- Leuk Res 2004; 28: 307–313. tidylserine expression on early apoptotic cells using fluorescein 63 Yu Q, Stamenkovic I. Localization of matrix metalloproteinase 9 to labelled Annexin V. J Immunol Methods 1995; 184: 39–51. the cell surface provides a mechanism for CD44-mediated tumor 42 Liu D, Sy MS. A cysteine residue located in the transmembrane invasion. Genes Dev 1999; 13: 35–48. domain of CD44 is important in binding of CD44 to hyaluronic 64 Yu Q, Stamenkovic I. Cell surface-localized matrix metalloprotei- acid. J Exp Med 1996; 183: 1987–1994. nase-9 proteolytically activates TGF-beta and promotes tumor 43 Mechti N, Piechaczyk M, Blanchard JM, Marty L, Bonnieu A, invasion and . Genes Dev 2000; 14: 163–176. Jeanteur P et al. Transcriptional and post-transcriptional regulation 65 Hebbard L, Steffen A, Zawadzki V, Fieber C, Howells N, Moll J of c-myc expression during the differentiation of murine erythro- et al. CD44 expression and regulation during mammary gland leukemia Friend cells. Nucleic Acids Res 1986; 14: 9653–9666. development and function. J Cell Sci 2000; 113: 2619–2630.

Leukemia CD44 modulation by IL-6 T Vincent and N Mechti 975 66 Heufelder AE, Bahn RS, Boergen KP, Scriba PC. Detection, leukemia cell proliferation via p27Kip1. Blood 2004; 103: 1059– localization and modulation of hyaluronic acid/CD44 receptor 1068. expression in patients with endocrine orbitopathy. Med Klin 69 Reber S, Matzku S, Gunthert U, Ponta H, Herrlich P, Zoller M. (Munich) 1993; 88: 181–184, 277. Retardation of metastatic tumor growth after immunization with 67 Georgii-Hemming P, Wiklund HJ, Ljunggren O, Nilsson K. Insulin- metastasis-specific monoclonal antibodies. Int J Cancer 1990; 46: like growth factor I is a growth and survival factor in human 919–927. multiple myeloma cell lines. Blood 1996; 88: 2250–2258. 70 Seiter S, Arch R, Reber S, Komitowski D, Hofmann M, Ponta H 68 Gadhoum Z, Leibovitch MP, Qi J, Dumenil D, Durand L, et al. Prevention of tumor metastasis formation by anti-variant Leibovitch S et al. CD44: a new means to inhibit acute myeloid CD44. J Exp Med 1993; 177: 443–455.

Leukemia