Heparan sulfate proteoglycans in the control of B cell development and the pathogenesis of multiple myeloma



S. T. Pals, Department of Pathology, Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands

Fax: +31 0 20 566 9523

Tel: +31 0 20 566 5635

E-mail: s.t.pals@amc.uva.nl


Heparan sulfate proteoglycans (HSPGs) have essential functions during embryonic development and throughout postnatal life. To exert these functions, they undergo a series of processing reactions by heparan-sulfate-modifying enzymes (HSMEs), which endows them with highly modified heparan sulfate (HS) domains that provide specific docking sites for a large number of bioactive molecules. The development and antigen-dependent differentiation of normal B lymphocytes, as well as the growth and progression of B-lineage malignancies, are orchestrated by an array of growth factors, cytokines and chemokines many of which display HS binding. As discussed in this review, tightly regulated HSPG expression is a requirement for normal B cell maturation, differentiation and function. In addition, the HSPG syndecan-1 functions as a versatile co-receptor for signals from the bone marrow microenvironment, essential for the survival of long-lived plasma cells and multiple myeloma (MM) plasma cells. Targeting of HSMEs or HS chains on MM cells increases their sensitivity to drugs currently used in MM treatment, including bortezomib, lenalidomide or dexamethasone. Taken together, these findings render the HS biosynthetic machinery a promising target for MM treatment.


a proliferation-inducing ligand


B cell activating factor


bone marrow


extracellular matrix


epidermal growth factor




d-glucuronic acid


glucuronyl C5-epimerase


N-acytelated glucosamine


hepatocyte growth factor


heparan sulfate


heparan-sulfate-modifying enzyme


heparan sulfate proteoglycan


heparan sulfate O-sulfotransferase


l-iduronic acid




monoclonal gammopathy of undetermined significance


multiple myeloma




stromal-derived factor-1




transmembrane activator and CAML interactor


tumor necrosis factor


vascular cell adhesion molecule 1


Heparan sulfate proteoglycans (HSPGs) are proteins with covalently attached, unbranched polysaccharide heparan sulfate (HS) chains, which in the unmodified form consist of alternating N-acetylated glucosamine (GlcNAc) and d-glucuronic acid (GlcA) units [1-3]. These macromolecules are expressed in all mammalian tissues as extracellular matrix (ECM) components or as cell-membrane-bound proteins. Three major families of proteoglycan core proteins are the membrane-spanning syndecans (four members) [4], the membrane glycosylphosphatidylinositol-linked glypicans (six members) [5] and the basement membrane or ECM proteoglycans perlecan, agrin and collagen XVIII [6, 7]. Other HSPGs are betaglycan and CD44-HS (CD44v3), a splice variant of the CD44 molecule [8-10]. The different core proteins are expressed in a cell-type-specific and temporal- and spatial-regulated manner, and their expression correlates with different physiological responses in cells [11, 12]. HSPGs can function via both their core protein and the attached HS chains. For example, the syndecan-1 core protein can interact with and regulate the activity of integrin adhesion molecules [13-17]. The cytoplasmic domain contains peptide sequences that can bind cytoskeletal proteins and can serve as substrate for cellular kinases. Thus, syndecans may act as signaling molecules [13, 14, 16, 18]. On the other hand, HSPGs can bind and present proteins via their HS chains, of which the binding capacity and specificity are determined by enzymatic modifications [1, 2, 19-22]. Hence, HSPGs can act as multifunctional scaffolds regulating important biological processes including cell growth, adhesion and migration, tissue morphogenesis, organogenesis and angiogenesis [2, 12].

Although the spatiotemporal and tissue-specific expression of individual proteoglycan core proteins determines where and when HS chains are present, ligand binding depends on the structure of the HS chains [1, 2]. The HS chain modification sequence, which is essential for the binding capacity and specificity of the HS chain, consists of a diverse set of cell-type and tissue-specific chemical modifications, controlled by Golgi-located HS-modifying enzymes (HSMEs). This process is independent of the proteoglycan type. The modification process (Fig. 1A–C) has been reviewed extensively elsewhere in more detail [2, 3, 23-25] and therefore will only be briefly discussed here. HS chain synthesis starts with a chain initiation step consisting of transfer of xylose from UDP-xylose by xylosyltransferase-I and/or -II to specific serine residues within the proteoglycan core protein. This is followed by attachment of two d-galactose residues by galactosyltransferase-I and -II and GlcA by glucuronosyltransferase-I, completing the formation of the core protein linkage tetrasaccharide (Fig. 1C) [1, 2]. The second step, chain elongation or polymerization, can only take place once the tetrasaccharide linker is present. This polymerization reaction is carried out by enzymes from the exostin (EXT) family, with a major role for exostin-1 (EXT1) and EXT2 [26]. The EXT proteins function as a hetero-oligomeric complex to polymerize the HS chain [2, 27, 28]. It should be noted that, whereas reduced function in either co-polymerase results in the bone exostosis phenotype, complete loss of either enzyme results in total loss of HS chains, which is lethal at gastrulation [20, 29]. Subsequently, the unbranched HS chains undergo a final step, which consists of a complex series of processing reactions involving GlcNAc deacetylation and sulfation by the N-deacetylase/N-sulfotransferases (NDST), epimerization of GlcA by glucuronyl C5-epimerase (GLCE), converting it to l-iduronic acid (IdoA), and subsequent O-sulfation at three different positions by heparan sulfate O-sulfotransferases (HS2ST, HS3ST and HS6ST) [1-3]. In contrast to heparin, HS chains are only partially modified and the modified residues are clustered, resulting in a polysaccharide chain having regions that are highly sulfated and flexible alternated with nearly unmodified rigid regions (Fig. 1C). The modified domains provide specific docking sites for a large number of bioactive molecules, including growth factors, chemokines and morphogens. HS glucosaminoglycans preferentially bind to proteins containing the consensus sequences BBXB and BBBXXB, where B is a basic amino acid (e.g. Lys, Arg or His). It should be noted, however, that the validity of the ‘consensus sequence’ concept is debatable. Although the binding site for HS is often defined by positive amino acids within the majority of heparin-binding proteins, it is clear that besides this electrostatic hydrogen bonding also van der Waal interactions and hydrophobic effects contribute to HS binding [25]. Once the HSPGs leave the Golgi, the Golgi-located modifying enzymes can no longer affect the sulfation pattern of the HS chain. However, in addition to the modifications in the Golgi, post-biosynthetic modifications can be made by (extra)cellular sulfatases (SULF-1 and SULF-2) and/or heparanase. These enzymes remove 6-O-sulfate groups or generate smaller HS chain fragments with retained binding capacity, respectively, thereby changing and regulating the functionality of the sugar chain [30-35].

Figure 1.

Schematic representation of HS structure. (A) Unmodified disaccharide consisting of GlcA and GlcNAc, simplified as a blue and red symbol, respectively. (B) Theoretical most highly modified disaccharide consisting of a C5-epimerized IdoA (green symbol) sulfated at the C2-position and GlcNAc sulfated at the N-, C3- and C6-position. (C) An HSPG core protein with an HS chain attached to it, via a linkage tetrasaccharide, polymerized by the HS co-polymerases EXT1 and EXT2. As mentioned in the text, the modifications occur in clusters. Unmodified regions (NA) alternate with highly modified regions (NS). These modified clusters make up binding sites for different proteins, as is shown for antithrombin in this figure. In purple different HSMEs are depicted, including the sites upon which they act. For more details see the text.

Genetic defects in humans and studies with genetically modified animals have unequivocally shown the importance of HSPGs and their correct modification in (embryonic) development, normal physiology and disease [20, 21, 26, 36-39]. Interestingly, numerous studies over the past few years have clearly demonstrated that functions attributed to HSPGs could be relevant for the development, function, differentiation and immunological responses of normal B cells as well as the development and progression into B cell malignancies. In this review, we provide an overview of the current knowledge on the expression and function of HSPGs and HSMEs in B cells, with specific emphasis on normal and malignant multiple myeloma (MM) plasma cells.

HS modification during B cell development and differentiation

During early B cell development in the bone marrow (BM), pro-B cells develop from common lymphoid progenitors into pre-B and immature B cells. These immature B cells complete their maturation after migrating to the periphery, differentiating into mature follicular B cells. Upon challenge with antigen, the mature naïve B cells undergo antigen-specific B cell differentiation, which leads to the formation of antibody-producing plasma cells or memory B cells [40-42]. In mice, the HSPG syndecan-1 (CD138) is expressed on pro/pre-B and immature B cells [43, 44] and the level of syndecan-1 expression on these cells distinguishes different B cell developmental lineages (B-1a/b, B-2) [44]. In addition to syndecan-1, syndecan-4 was identified on murine pre-B cells and maintained until the mature B cell stage [45]. More recently, it was demonstrated that syndecan-4 is predominantly expressed by murine pro-B cells and that its expression is controlled by the B-cell-specific transcription factor PAX-5 [46]. Moreover, it was shown that conditional expression of BLNK/SLP65, a downstream effector of the pre-B cell receptor, markedly reduced the expression of syndecan-4 while it increased the expression of HS-glucosamine 3-O-sulfotransferase-1 (HS3ST1) [47]. These changes in HSPG expression in early B cells coincide with a fundamental change in growth factor requirements, since pro-B cells are dependent on interleukin-7 (IL-7) and stem cell factor for their proliferation and survival [48], while the small pre-B cells become unresponsive to these growth factors [47, 49, 50]. It is conceivable that modification of HS expression alters the HS-dependent growth factor binding specificity and thereby the biological activity of essential factors that control the transition of pro-B to pre-B cells [47]. Consistent with this notion, it was shown that mice deficient for sulfatase modifying factor 1 (SUMF1), and consequently lacking all sulfatase activity, display a block in the maturation of pro- to pre-B cell [51]. In addition, IL-7-mediated proliferation of early murine B cells is enhanced by exogenous HS, either in soluble form or expressed on co-cultured cells [50]. HS has the capacity to bind the λ5 tail of the pre-B cell receptor thereby increasing ERK1/2 phosphorylation and downstream signaling in murine B cell progenitors [52]. Apparently the regulation of the sulfation pattern of HS is essential for the transition and early development of mouse B cells. Interestingly, our recent studies in mice with B cells devoid of the HSME C5-epimerase revealed that lack of chain flexibility in combination with a marked alteration in the sulfation pattern does not affect early B cell development but is important for the transition of immature to mature naïve B cells. C5-epimerase deficiency resulted in a significantly decreased number of mature follicular B cells, which was reflected in a lower basal level of protective antibody production and an impaired antigen-specific response [53]. In line with these findings, a recent glycotranscriptome study showed that the HS biosynthetic machinery is actively regulated during different developmental stages of B cells [54], which suggests that they adjust their HS modification depending on the environment or activation state [53-55].

In contrast to murine B cell development, early human B cell development is IL-7 independent [56]. This is paralleled by a lack of syndecan-1 expression on early human B cells [57]. Whether other HSPGs are expressed on early human B cells remains elusive since no comprehensive studies are available. Most human mature B cell lines (BJA-B, Ramos, Raji; all human Burkitt lymphoma cell lines) lack or show low-level HS expression, as shown by the anti-HS antibody 10E4, recognizing a mixed N-sulfated and N-acetyled epitope [58], or the NAH46 antibody, recognizing an unmodified/non-sulfated epitope [59], and both EXT1 and EXT2 are expressed at low level [59]. In line with these observations, we have demonstrated that HS is hardly detectable on freshly isolated primary human tonsillar B cells. Interestingly, however, in vitro stimulation by CD40 ligation and/or BCR triggering induced strong expression of cell surface HS, signals mimicking early B cell activation through the B cell receptor and follicular T helper cells. This B cell activation enhanced the expression of CD44 splice variants containing exon v3, decorated with HS [60]. B cells expressing cell surface CD44v3-HS were shown to bind hepatocyte growth factor (HGF) and promote HGF-mediated intracellular signaling via its cognate receptor MET [60]. In subsequent studies, Shachar and colleagues revealed that the HGF/c-Met pathway is specifically involved in the homeostasis and survival of mature follicular B cells, presumably involving interaction of HGF with cell surface HSPGs on (activated) follicular B cells [53, 61]. Interestingly, several other cytokines known to be important in the B cell lineage commitment also contain HSPG binding domains, including stromal-derived factor-1 (SDF-1, or CXCL12), CXCL13, IL-4 and IL-6 [21, 62]. The importance and specific role of HSPGs and HS modifications in these settings have not yet been explored. Nevertheless, the available data point toward a function of HS (modification) at several distinct stages during B cell development. Notably, expression array analysis on subsets of human B cells isolated from the tonsil revealed a ‘shutdown’ of nearly all HS biosynthetic enzymes in germinal center B cells, compared with naïve and memory B cells, which displayed distinct but specific expression patterns for the different HSMEs (Reijmers and Pals, unpublished observations). These results are in line with the findings of Duchez et al. in mice [54]. In addition, Bret and colleagues have shown that different B cell subsets have a distinct expression profile of HSME. For example, early plasmablasts and plasma cells expressed high levels of HS2ST and EXT1 compared with memory B cells, whereas HS3ST2 and HS6ST1 expression was higher in memory B cells than in plasmablasts [63].

In summary, recent studies in mice have shown that HSPG expression is regulated during B cell development, and have revealed that in particular the pro-B cell to pre-B cell transition (in mice) is critically dependent on the sulfation pattern of the HS chains controlled by intracellular sulfotransferases and extracellular sulfatases [46, 47, 51, 54]. In addition, recent data from our laboratory obtained in C5-epimerase deficient mice demonstrate that rigid HS chains (leading to an aberrant sulfation pattern) interfere with B cell maturation, resulting in reduced antibody production [53]. Although HS modification during human B cell development has not been explored in full detail, it is clear that human B cells strongly increase their cell surface HS upon antigenic stimulation and CD40 ligation and that they display a specific HSME expression profile depending on the differentiation stage. Ultimately, this may affect cellular function by altering adhesion and migration and the interaction with several important cytokines and growth factors [9, 53, 60, 63]. As mentioned before, antigen-experienced B cells will develop into memory B cells or antibody secreting plasma cells. The latter are characterized by expression of the HSPG syndecan-1, the functions of which will be discussed later. With respect to memory B cells, it should be noted that very little is known about the role of HSPGs in their function and homeostasis.

Expression and function of HSPGs in normal and malignant plasma cells

Function of syndecan-1 in normal plasma cells

The normal BM consists of a heterogeneous population of hematopoietic cells contained within a balanced mixture of mesenchymal stromal cells, adipocytes, endothelial cells and insoluble ECM [64]. In addition, it includes immune cells and cells involved in bone homeostasis such as osteoblasts and osteoclasts [65]. Within this complex tissue, a specialized but poorly defined microenvironment known as the ‘plasma cell niche’ is located. Normal plasma cells within this niche do not show detectable proliferation and lie in close proximity to non-dividing mesenchymal stromal cells, expressing vascular cell adhesion molecule 1 (VCAM-1) and high levels of CXCL12 [66]. CXCL12 is a well-known chemokine attracting plasmablasts to the BM niche and providing survival signals to plasma cells in vitro (Fig. 2) [67]. Three different splice variants of CXCL12 exist in mice and humans of which the recently characterized CXCL12γ variant binds HS with extremely high affinity due to the presence of an extended basic domain [68, 69]. The sites of expression and functional importance of the different CXCL12 isoforms for plasma cell survival and retention within the BM microenvironment remain to be explored. Besides CXCL12, signals from the tumor necrosis factor (TNF) family members, a proliferation-inducing ligand (APRIL) and B cell activating factor (BAFF), and the cytokine IL-6 play a crucial role in plasma cell survival in the BM (Fig. 2) [53, 70, 71]. APRIL is a survival factor for plasma cells and plasmablasts, the proliferative precursor of the plasma cell. In contrast to BAFF, its function has been shown to be dependent on interaction with HSPGs [53, 72-75], which act as a platform for APRIL multimerization, facilitating receptor crosslinking [73]. In mucosal tissues, APRIL was shown to be retained by local HSPGs, binding plasma-cell-expressed syndecan-1 and creating plasma cell niches [76]. Apart from cytokine/growth factor signals, adhesive interactions within the niche are an important factor determining plasmablast/plasma cell survival [67, 77]. Interestingly, CD44 isoforms expressed on murine BM-derived plasma cells as well as human MM cell lines can mediate adhesion to BM stromal cells or ECM components [77, 78]. The possible contribution of the HS-carrying CD44v3 isoform, which is expressed on a number of human MM cell lines [78], has not yet been explored.

Figure 2.

BM plasma cell survival niche. Once B cells have differentiated into a plasma cell, they undergo SDF-1-controlled migration towards the BM. There, BM plasma cells lie in close contact with a stromal cell, which expresses high levels of VCAM-1 and SDF-1. This results in the retention of the BM plasma cell in its niche. In addition, stromal cells express potent survival factors, including BAFF, APRIL and IL-6. Plasma cells express the marker CD138, which is the HSPG syndecan-1. It is unclear whether syndecan-1 plays an indispensible role in facilitating APRIL-mediated survival or SDF-1-induced retention (or migration), and whether syndecan-1 interacts with integrins to establish outside-in signaling and adhesion. For more details see the text.

Expression of the HSPG syndecan-1 is a hallmark of both human and mouse plasma cells [79], but its precise functions remain to be defined. Syndecan-1 deficient mice are viable, develop normally, and are able to produce antigen-specific immunoglobulins [80]; small changes in antibody titers were observed only in an antibody-induced glomerulonephritis model [81]. These findings suggest that normal plasma cells do not require syndecan-1 for their function, although it should be noted that long-lived BM plasma cells have not been studied. Importantly, however, we observed that plasma cells of syndecan-1 deficient mice show normal levels of HS expression, implying redundancy of the syndecan-1 core protein (our own unpublished observation). While the identity of the alternative core protein remains to be established, these data suggest that the HS chains, rather than the syndecan-1 core protein itself, are essential for plasma cell function. Indeed, we showed that C5-epimerase deficiency attenuates survival of plasma cells, while it does not affect the syndecan-1 core protein expression [53].

Function of syndecan-1 in MM plasma cells

MM is a still incurable B-lineage malignancy characterized by uncontrolled expansion of malignant plasma cells in the BM. It develops from a pre-malignant condition termed monoclonal gammopathy of undetermined significance (MGUS). In about 50% of MGUS and MM patients, the clonal plasma cells harbor translocations involving the immunoglobulin heavy chain (IgH) locus and one of the following partner genes: CCND1 (Cyclin D1), MMSET, CCND3 (Cyclin D3), c-MAF or MAFB. Most of the remaining cases of MM are associated with hyperdiploidy (IgH non-translocated MM). The hyperdiploid subgroup is characterized by trisomies of chromosomes 3, 5, 7, 9, 11, 15, 19 and 21 [82]. Genetic abnormalities, including RAS mutations, p16 or p53 loss, or changes involving MYC family genes, have all been identified in clonal plasma cells in association with progression to advanced stage MM [83-85]. Importantly, the BM microenvironment becomes heavily modified in the course of disease progression and plays a crucial role in the biology of the disease. The interaction between MM cells and the BM microenvironment is bidirectional: the MM cells disrupt the homeostasis of the BM, resulting in anemia, aberrant angiogenesis and osteolytic bone disease, while the BM microenvironment supports the growth and survival of the malignant plasma cells through signals mediated by adhesion molecules, cytokines and growth factor [83, 86, 87] (Fig. 3). In this way, activated stroma triggers the paracrine and autocrine production of a variety of cytokines and growth factors by MM cells, of which the most well defined are IL-6, IGF-1, TRANCE, VEGF, WNT, TNFα, CXCL12α, TGFβ, APRIL/BAFF, HGF and HB-EGF (Fig. 3). These factors have been demonstrated to mediate growth, survival, migration and retention of MM cells and to promote BM angiogenesis [65, 88-93]. Thus, although mutations in essential growth control genes underlie MM development, signals emanating from the BM microenvironment are also crucial for driving MM cell growth, survival and migration.

Figure 3.

Interaction of MM cells with the BM microenvironment. A schematic overview of an MM cell interacting with different components of the BM environment. MM cells migrate into the BM, directed by the stromal-cell-produced chemokine SDF-1. Once in the BM, MM cells are provided with an array of autocrine and paracrine soluble survival and proliferation inducing factors. In addition, the MM cell can interact with the ECM, or stromal cells, via several integrins or the HSPG syndecan-1. By producing factors that inhibit osteoblast differentiation and enhancing osteoclast activity, the patient will develop osteolytic bone disease. For more details see the text.

Studies over the past decade have identified syndecan-1 as a key regulator in the interaction of MM cells with the BM microenvironment [20, 36, 89, 90, 94-96]. Specifically, it has become clear that multiple cytokines, chemokines and growth factors in the (MM) plasma cell niche, as well as many ECM proteins, can interact with syndecan-1. Syndecan-1 is the dominant, and generally only, HSPG on the cell surface of MM cells and suppressing its expression results in the loss of HS from the cell surface and negatively affects cell growth and survival [20, 94, 95]. More specifically, several growth factors including HGF and epidermal growth factor (EGF) family members were shown to utilize syndecan-1 as co-receptor, promoting signaling via activation of the phosphatidylinositol 3-kinase (PI3K)/AKT and RAS/mitogen-activated protein kinase (MAPK) pathways and promoting cell survival and proliferation [36, 89, 92]. Two other important MM survival factors in the BM microenvironment are BAFF and APRIL, which induce myeloma survival through activation of nuclear factor-κB (NF-κB), PI3K/AKT and MAPK pathways [97, 98]. Both factors can bind to the TNFR-related receptors BCMA and transmembrane activator and CAML interactor (TACI). In addition, BAFF can bind and activate the BAFF receptor. MM cells express at least one of these receptors, most commonly TACI and/or BCMA [99]. Importantly, in contrast to activation by BAFF, activation of TACI by APRIL requires interaction with syndecan-1 [74, 75, 100] and it was shown that MM cell proliferation and survival in patients with high expression levels of TACI is primarily mediated by APRIL [99, 101]. Interestingly, these patients showed a BM microenvironment-dependent gene-expression signature, suggesting that syndecan-1 is an important co-receptor for APRIL mediating survival of BM microenvironment-dependent myeloma plasma cells [99].

Syndecan-1 supports growth and BM retention of myeloma cells

Several studies in mouse models have addressed the role of syndecan-1/HS in MM growth and survival. Initial studies by the laboratory of Sanderson showed that treatment of subcutaneous transplanted human MM cells with bacterial heparitinase caused a reduction in tumor growth rate. These studies did not discriminate between HSPGs expressed on the MM cell surface or in the tumor microenvironment [96] but in vitro studies by other laboratories suggested that HS expressed by MM cells was crucial for promoting growth and survival signaling [36, 92]. Consistent with this notion, knockdown of syndecan-1 in an MM cell line caused inhibition of subcutaneous MM growth [95]. Strikingly, the tumors that did develop displayed normal levels of cell surface syndecan-1, implying selective pressure favoring outgrowth of syndecan-1 bearing cells [95]. In yet another study overexpression of SULF-1 and -2, which remove 6-O-sulfate groups from HS chains, was found to result in a reduced tumor growth [102].

Recently, advances in bioluminescence imaging allowed for the development of mouse models that more closely resemble the pathobiology of human MM. In these models, human MM cell lines or primary human MM cells, expressing a green fluorescent protein-luciferase fusion protein, are injected into immunodeficient mice (Rag-2−/−γc−/− or SCID). Subsequently, bioluminescence imaging is employed to monitor the engraftment, outgrowth and distribution of the injected MM cells in real time [20, 94, 103-105]. Khotskaya and colleagues applied this novel approach to study dissemination of MM cells in which syndecan-1 was targeted by small hairpin RNAs [94]. Whereas control mice showed MM growth in the femora, spines and ribs, mice that received MM cells in which syndecan-1 was targeted did not display any MM growth in the bones. Upon subcutaneous injection, the syndecan-1 knockdown MM cells also revealed a significantly impaired growth [94]. Consistent with these findings, we observed that L363 MM cells poorly localize to the BM upon doxycycline-induced knockdown of syndecan-1 [20]. In this experiment, we injected a 1 : 1 mixture of dye-labeled [87] L363 MM cells, either expressing syndecan-1 or not. Four hours after injection, twice as many syndecan-1 knockdown cells were found in the circulation, which persisted for at least 24 h, while the majority of MM cells recovered from the BM were syndecan-1 positive, indicating that syndecan-1 is important for MM homing and/or retention in the BM (our own unpublished observations). Taken together, these studies indicate that syndecan-1 is essential for MM growth, survival and localization and suggest that the HS chains of syndecan-1 may play an essential role. In addition to mediating growth factor presentation, the syndecan-1 core protein may act as an adhesion molecule or control adhesion indirectly. However, this function that could be involved in the homing and retention of MM cells to the BM needs further exploration.

The HS chains of syndecan-1 are indispensible for MM growth

The findings discussed above suggest that the HS chains of syndecan-1 are essential for MM growth within the BM microenvironment. To directly assess this role, we generated MM cell lines in which EXT1, an enzyme essential for HS chain synthesis, can be targeted by doxycycline-inducible RNA-interference (RNAi) mediated knockdown [20, 29]. In this system, administration of doxycycline leads to expression of syndecan-1 proteins without HS chains, regardless of whether syndecan-1 is expressed on the cell membrane or shed into the ECM. Importantly, we observed that loss of HS chains severely impairs the growth of the MM cells in vitro. Upon injection into immunodeficient mice, an exponential outgrowth of MM cells in different parts of the bone marrow was observed in mice that had not received doxycycline. By contrast, mice treated with doxycycline from the first day of injection remained completely tumor free. Interestingly, in mice in which tumors were allowed to form before doxycycline treatment was started, doxycycline administration led to growth arrest or even tumor regression [20]. Targeting EXT1 in the MM cells did not affect the expression of the syndecan-1 core protein but did result in enhanced syndecan-1 shedding (our own unpublished data). These results demonstrate a crucial role in MM growth for the HS chains of syndecan-1, and not the core protein per se.

Distinct roles for cell surface and shedded syndecan-1

As discussed above, syndecan-1 at the cell surface can act as a functional co-receptor for cytokine and growth factor signaling. However, it can also be shed into the microenvironment after cleavage by matrix metalloproteinases [36, 92, 99]. This syndecan-1 can interact with multiple cell types and molecules (including collagen and fibronectin) and serve as a reservoir by localizing and concentrating effector molecules, such as growth factors and chemokines, in the vicinity of myeloma cells as well as other cells in the BM. In a paracrine fashion, it may influence the behavior of stromal and endothelial cells affecting angiogenesis and bone homeostasis [106]. An important player in this scenario is heparanase. Heparanase is hardly detectable in normal human tissues; however, in MM and other cancers it is often strongly expressed and this overexpression is correlated with poor prognosis [107, 108]. Heparanase cleaves HS causing the release of fragments consisting of 5–10 disaccharides. These fragments are biologically active and promote the availability and activity of HS binding growth factors that act on both the MM cells and other cells in the tumor microenvironment. Furthermore, heparanase was shown to enhance osteolysis by upregulating the expression of receptor activator of NF-κB ligand (RANKL) [109]. Importantly, heparanase also enhances shedding of syndecan-1 by MM cells, promoting angiogenesis and facilitating tumor migration, invasion and dissemination [94, 107, 108, 110].

Targeting of HS synthesis increases the drug sensitivity of MM

Despite major recent advances in treatment, MM is still an incurable disease with a median survival of 4–6 years. A main cause of treatment failure is that myeloma cells develop drug resistance [111]. In addition, the drug dosages that are required for effective control of tumor growth are high and often limited by serious side-effects [85]. For these reasons, strategies need to be developed that increase the efficacy of current therapies. In this context, recent unpublished studies from our laboratory (Fig. 4), which suggest that targeting the HS synthesis machinery can sensitize MM cells to drug-induced cell death, are of potential interest. In these experiments, we assessed the sensitivity of MM cells to lenalidomide and bortezomib, drugs that are mainstays of current MM treatment, in the presence or absence of EXT1 knockdown (loss of all HS) [85]. As reported previously, loss of only HS reduced cell growth and viability (Fig. 4A,B), while exposure to lenalidomide or bortezomib alone also markedly reduced cell viability (Fig. 4B). Interestingly, targeting HS expression in MM cells exposed to lenalidomide or bortezomib almost completely blocked cell growth (Fig. 4B). These findings indicate that inhibition of HS synthesis, which disables communication with the microenvironment, sensitizes MM cells to drug-induced cell death. Disruption of the MM cell interaction with the tumor microenvironment presumably also underlies the sensitization of myeloma cells to dexamethasone-induced growth inhibition following treatment with SST0001 [112]. This novel heparanase inhibitor prevents degradation of HS and shedding of syndecan-1, interfering with the generation of biologically active HS fragments and preventing the development of a highly growth-factor-enriched tumor microenvironment. In addition, blocking syndecan-1 shedding significantly diminished the angiogenic potential of the syndecan-1 core protein. Interestingly, dexamethasone-resistant MM cells responded poorly to dexamethasone or SST0001 as single treatments, but combination treatment almost completely blocked tumor growth, presumably through dual targeting of the tumor and its microenvironment [112]. Furthermore, by analogy of the use of HS mimetics in hematopoietic stem cell mobilization, these compounds might be employed for mobilization of MM cells from their niches [113]. Thus, targeting HS synthesis, modification or processing may be employed to sensitize MM cells to myeloma drugs, allowing drug dosage reduction and overcoming side-effects and/or the development of drug resistance.

Figure 4.

Doxycycline-induced knockdown of EXT1 results in loss of HS and sensitizes L363 MM cells to drug-induced cell death. L363-shEXT1a cells were incubated with (+dox) or without (−dox) doxycycline for 5 days before each experiment to allow for optimal RNAi-mediated knockdown of EXT1. All experiments were performed in the presence of 10% fetal bovine serum. (A) Expression of cell surface HS on L363-shEXT1a measured by fluorescence-activated cell sorting using antibody 10E4 (Seikagaku America) after doxycycline-induced RNAi-mediated knockdown of EXT1. The expression of untreated samples was normalized to 100%. Bars represent mean ± SEM of three independent experiments. (B) Viable L363-shEXT1a cells after a 5-day incubation period with or without lenalidomide or bortezomib. Untreated samples (no drugs) were normalized to 100%. Bars represent mean ± SEM of two independent experiments, performed in triplicate. (A, B) **< 0.01; ***< 0.001.

Conclusions and perspectives

The synthesis and modification of HS chains consists of a complex series of enzymatic reactions, ultimately leading to an HS-glycosaminoglycan with highly sulfated regions. These regions determine the strength and outcome of HS–ligand interactions [21, 22, 37, 38, 114, 115]. The relevance of HSPGs as co-receptors for many different growth factors and chemokines is now well established and the subject of study in several biomedical fields, including immunology and cancer research. As discussed in this review, B cells express HSPGs and correct modification of HS chains during development is a requirement for B cell maturation, differentiation and function. In addition, normal and malignant plasma cells express high levels of the HSPG syndecan-1, which is essential for the survival of MM plasma cells. Importantly, studies from our own and other laboratories indicate that targeting of (HSMEs or) HS chains on MM cells may increase the sensitivity to drugs currently used in MM treatment including bortezomib, lenalidomide or dexamethasone [84, 85, 111, 116]. This may allow the use of lower drug dosages or shortening of treatment regimens, diminishing or preventing drug-induced side-effects. Similar to conventional therapies, treatments targeting HS synthesis or ligand–HS interaction will inevitably cause side-effects. To minimize these side-effects, it will be essential to identify tumor-specific HS modifications and to define the HS sequences that bind essential growth and survival factors. A paradigm for the existence of such ligand-specific binding sequences is that for antithrombin (Fig. 1C), but during the last two decades data have accumulated identifying specific minimal HS binding sequences for a plethora of growth factors and chemokines [25]. In the context of MM, the finding that HGF binds a preferred HS sequence, clearly different from that of, for example, basic fibroblast growth factor [117, 118], is of great significance, since HGF has been implicated in both myeloma growth and progression. It will be very interesting to explore whether targeting this HS–HGF interaction sequence with, for example, non-coagulant HS mimetics can be employed as a strategy for treating myeloma patients [36, 93].


This work was supported by grants from the Dutch Cancer Society (EU-FP7-OVER-MyR) and the Netherlands Organization for Scientific Research (NWO, ZonMW, Veni grant 916.13.011). The authors would like to thank Chandra Boerlage for graphical support and María José Villalobos for the initial set-up of the experiments. The authors declare no conflict of interest.