Glycosylation in Health and Disease

Decisive contributions to the kind of interaction between heparin and AT III, crucial for anticoagulant activity of heparin, were reported by Rosenberg and his group, who were the first to describe the accelerating effect of heparin on neutralizing proteolytic enzymes in haemostasis (Lam et al., 1976). They also identified the monosaccharide residues, critical for anticoagulant activity, by determining the affinity of differently sulfated oligosaccharides to human AT III. A lack of the 3-O-sulfate as well as the absence of the 6-O-sulfate group on the N-acetylated glucosamine resulted in a reduction of binding energy by about 50% (Atha et al., 1985). These results were confirmed later by the finding that heparan sulfate 6-O-sulfotransferase, as a crucial enzyme in the biosynthesis, transfers a sulfate residue to C6 of the GlcNAc unit (see also Chapter 10) in the AT-binding motif (Zhang et al., 2001). Glycan chains once attached to a core protein may be subjected to further modifications initiated by heparanase (see below) that cleaves HS between sulfated glucosamine and glucuronic acid units. The roles the various sulfotransferases and the two known human sulfatases play in tumor suppression, as well as in promoting cell proliferation, angiogenesis, and invasiveness have been reviewed by Knelson et al. (2014).Many patients suffering from cancer exhibit, in addition, an overexpression of the blood coagulation system associated with the risk of venous thromboembolism including deep leg vein thrombosis and pulmonary embolism as the second most frequent cause of death. Hence, prophylaxis with LMW heparins is indispensable. Interestingly, it has been found that anticoagulants like heparins have an antimetastatic and antineoplastic (prevention of maturation and proliferation of neoplasms) effect, predominantly interfering with metastasis formation. The complex process of secondary tumor formation starts with the spreading of single cells from the primary tumor due to loss of cell-cell contacts. Cell migration requires prior degradation of the ECM, separating the tumor from adjoining tissues, and takes place by the action of enzymes such as heparanase, secreted by the mobilized tumor cells. This is followed by intravasation and dissemination to other sites of the body via the lymphatic system or blood stream. Intravascular adhesion mediated by complex mechanisms is then followed by extravasation, proliferation, and tumor angiogenesis (formation of new blood

vessels; see, e.g., Daniel and Abrahamson, 2000; Folkman et al., 1992); the latter event-critical for survival of micrometastases-is achieved by production of growth factors, e.g., the vascular endothelial growth factor VEGF, stimulating among others the generation of new tubular structures sprouting from pre-existing small vessels. Involved in endothelial capillary morphogenesis are the transmembrane HSPGs, known as syndecans (Chapter 2). It was demonstrated by in vitro tumor angiogenesis that downregulation of syndecan-2 impairs angiogenesis in human microvascular endothelial cells so that inhibitors of syndecan-2 may be suited for antiangiogenic therapies (Noguer et al., 2008). Chiodelli et al. (2015) comprehensively reviewed the role of heparin/HS-PGs within the network of interactions (glycomic interactome) associated to angiogenesis.Ibrahim et al. (2013) unraveled mechanisms associated with siRNA-mediated knockdown of Syndecan-1 in human breast cancer stem cells. CSCs are known to possess self-renewal, unlimited proliferative potential, apoptosis resistance, etc., hence displaying increased resistance to therapies compared to the bulk tumor (Ibrahim et al., and literature cited therein). They found that Syndecan-1 knockdown results among others in reduced expression of components of the IL-6 signaling pathway including the transcription factor STAT3 (signal transducer and activator of transcription 3), activated in response to cytokines and growth factors by phosphorylation. Hence, the IL-6/STAT3 pathway may serve as a target for therapies aiming on inhibiting the stemness of CSCs. The small molecule BB1608 was reported to be able to inhibit transcription driven by STAT3 and shown to block cancer relapse and metastasis (Li et al., 2015); “stemness” stands for hypermalignant cancer cells, termed cancer stem cells or stemness-high cancer cells; stemness is assumed to be an important factor for cancer relapse after successful traditional therapy. Syndecans are upregulated in connection with other tumors too, e.g., Syndecan 2 in colorectal cancer cells (Vicente et al., 2013). These and other macromolecules containing carbohydrates are potential cancer biomarkers, a topic treated with the example of

colon cancer by Joo et al. (2014). The role of PGs in breast cancer has been reviewed by Theocharis et al. (2015, 2010). All the different stages the cells segregated from the primary tumor have to pass through to end up in a secondary tumor are basically promising targets for inhibiting cancer metastasis (Fritzsche and Bendas, 2009). In this connection, the above-mentioned heparanase, a HS-degrading endo-b-d-glucuronidase, plays a key role in metastasis and angiogenesis (reviewed, e.g., by Vlodavsky et al., 2006, 1994); a correlation between heparanase activity and metastasis was already described by Nakajima et al. (1984), who also reported that chemically modified heparins used as heparanase inhibitors significantly reduced the number of experimental melanoma lung metastases (Irimura et al., 1986). Hulett et al. (2000) identified the active site residues of heparanase as Glu225 and Glu343; this suggests a general acid catalysis mechanism as used by family A GHs (Chapter 4). Products of heparanase activity are HS-fragments of about 5-7 kDa in size, thus, affecting the ECM barrier properties. Heparanase mediates the induction of VEGF (independent of heparanase activity) as shown by experiments with endogenously added point mutated heparanase, and also the release of the basic fibroblast growth factor (bFGF) in an activity dependent manner; bFGF is present in basement membranes and in the subendothelial matrix of blood vessels and plays a role in wound healing as well as in the development of metastases. As heparanase activity was found to correlate with the metastatic potential of various carcinomas, and on the other hand gene silencing had an anti-cancerous effect, its inhibition provides a promising target for anti-cancer drugs (Vlodavsky et al., 2006 and literature cited therein). Competitive inhibition of heparanase with the consequence of blocking growth factors and/or HS-GAG degradation, and resulting in a reduction of metastasis formation is achieved by non-anticoagulant heparin oligosaccharides, e.g., LMW heparins with defined sulfation patterns being superior to unfractionated heparins in prolonging the survival of cancer patients (Vlodavsky et al. 2006; Marchetti, 2008; Pisano et al., 2014); hence these compounds gain increasing importance as candidates for cancer treatment (Ritchie et al., 2011) and different medications of heparin-based pharmaceuticals including heparin-conjugated nanoparticles for targeted drug delivery have been developed (Lokwani et al., 2014, and literature cited therein); several

of these compounds have meanwhile entered various stages of clinical trials (Sonderson and Iozzo, 2012; Drug Information Portal: https://www.druglib.com/druginfo/ heparin/trials/).Another important target for anti cancer drugs are cell adhesion molecules, among them integrins and selectins (Chapter 2). In case of an inflammation, an adhesion cascade directs lymphocytes along the cuboidal cells lining high endothelial venules by rolling toward the site of infection; this movement is initiated by short-lived weak interactions of L-selectin on the lymphocyte surface with glycans on the endothelium (McEver, 2005). During maturation, tumor cells alter the glycosylation pattern on their surface in a way enabling them to make use of such adhesion mechanisms, too, which requires the expression of the tetrasaccharides sLex and sLea as minimal binding epitopes for selectins. As a consequence, tumor cells bind, e.g., thrombocytes or platelets via P-selectin (also expressed on the endothelium), thereby exerting a protective function against natural killer (NK) cells; in addition, thrombocytes feed tumor cells with growth factors. The immunoglobulin, vascular cell adhesion molecule-1 (VCAM-1), is expressed on activated endothelium and serves as a ligand for very late activation antigen-4 (VLA-4). The integrin VLA-4 is also expressed on malignant melanoma cells so that tumor cells expressing VLA-4 may bind to the activated endothelial. Fritzsche et al. (2008) demonstrated that LMW heparins [e.g., enoxaparin (Clexane®)] inhibited in vitro binding of VLA-4 positive tumor cells by interaction with the integrin (notably, the synthetic anticoagulant heparin analog Fondaparinux lacks this ability). Similarly, heparins bind to P-and L-selectin and inhibit binding of these selectins to their natural ligands, which means that heparins may attenuate carcinoma metastasis, which on the other hand, is facilitated by these adhesion molecules (see, e.g., Stevenson et al., 2007, and for a review Borsig, 2004). 11.2.2 Glycosaminoglycans and HIVAccording to the UNAIDS report of the Joint United Nations program on HIV/AIDS about 35.3 million people were worldwide infected in 2012 with HIV-1 causing the acquired immune deficiency syndrome; it is estimated that about 80% of HIV infections result from sexual

intercourse with an efficiency of transmission being unexpectedly poor in vitro (e.g., Gray et al., 2001). However, it has been found recently that in vivo the 248 to 286 fragment of the prostatic acid phosphatase (PAP248-286) serves as a so-called semen-derived enhancer of viral infections (SEVI). It is assumed that amyloid fibrils formed by these peptides capture HIV particles and promote virion-cell attachment and fusion (Münch et al., 2007; Castellano and Shorter, 2012). Findings of Brender et al. (2009) suggest that an a-helical conformation of PAP248-286 supports its binding to cell membranes and contributes to an infectivity enhancement. Recently, peptides derived from HIV-1 gp120 co-receptor binding domain were also found to form amyloid fibrils and to enhance HIV-1 infection (Tana et al., 2014). The bridging function of SEVI obviously relies on its polycationic nature that reduces the repulsion between the HIV particles and the targeted cells by binding to negatively charged heparan sulfate proteoglycans (see also Chapter 10) present on the cell surface. A mutant PAP248-286 with Arg and Lys residues being exchanged by Ala was still capable of forming fibrils, but lacked the ability to enhance HIV infection as demonstrated by Roan et al. (2009). Thus, inhibitors counteracting the infectivity enhancing properties of SEVI are of great interest as supplements in HIV therapy and prophylaxis. A promising candidate turned out to be the main green tea polyphenol epigallocatechin-3-gallate (Hauber et al., 2009), which also remodels mature b-amyloid fibrils into smaller amorphous protein aggregates with reduced cellular toxicity (Bieschke et al., 2010) or the aminoquinoline surfen, a small molecule antagonist of HSPGs (Roan et al., 2010); surfen was also found to inhibit T cell proliferation in vitro and in vivo (Warford et al., 2014).A distinctive property of the HIV is its capability to evade antibody-mediated neutralization, although its envelope glycoprotein

presents numerous epitopes for receptor binding; this is due to antibody-induced conformational changes as proven by determining thermodynamic parameters of HIV-1 gp120 binding to different antibodies, revealing large negative entropy changes contributing to DG for binding with about 23 kcal/mol on average (Kwong et al., 2002). The envelope gp of HIV is generated as nascent gp 160 before it is cleaved into gp 41 and gp 120 by furin proteases. The HIV genome does not encode enzymes for carbohydrate synthesis-encoded are the sites of O and N-glycosylation (Chapter 2). Hence, the antigen-carbohydrate epitopes are generated exclusively by the enzymes of the infected cell, which means that the HIV virus is coated with immunologically “self” glycans (Scanlan et al., 2007). About half of the molecular weight of gp120 is allotted to its glycocalyx; of the complex glycans, most are of the high-mannose type; in this connection Dunlop et al. (2010) found that mannan polysaccharides produced by α1→3 mannosyltransferase gene (ΔMnn1) deficient S. cerevisiae exhibit a Manα1→2Man motif that is recognized by the broadly neutralizing anti-HIV antibody 2G12. The chemokine receptor CCR5 serves as a co-receptor in connection with the cell entry of HIV. CCR5 is posttranslationally modified by O-glycosylation and by sulfonation of its N-terminal tyrosines by the catalytic action of tyrosylprotein sulfotransferases as reported by Farzan et al. (1999); for entry of the virus into the target cell, the sulfonated CCR5 must bind to the gp120/CD4 complex. On the other hand tyrosine-sulfonated human antibodies bind to the CCR5 region of the gp120/CD4 complex and thereby contribute to neutralizing HIV-1 infection (Choe et al., 2003).Concerning glucosaminoglycans (GAGs), it has been shown already earlier that heparin and heparan sulfate (HS) interact with HIV gp 120 at a site termed the V3 loop, thereby, blocking the interaction with the CD4 receptor of the target cell, thus preventing in principle HIV infection. Binding of heparin and HS is mediated by ionic forces between the polyanionic GAGs and basic amino acid side chains of the neutralizing domain of the envelop protein. The fact that other sulfated polysaccharides such as dermatan and chondroitin sulfate do not show this effect indicates that apart from the polyanionic nature defined, structural characteristics are required for this kind of interaction (Rider, 1997 and literature cited therein). However, Huang et al. (2013) reported that fucosylated chondroitin

sulfate isolated from the sea cucumber Thelenota ananas exhibited strong anti-HIV activities for various HIV-1 strains and inhibited HIV entry into cells. In any case, such anionic polymers, including dextran sulfate (Callahan et al. 1991) and others (Baba et al., 1988; Moulard et al., 2000) belong to the group of candidates for topical microbicides against sexually transmitted pathogens such as HIV (Lackman-Smith et al., 2008, and literature cited therein). Apart from the heparan sulfate binding site within the V3 loop of gp120, Crublet et al. (2008) identified gp120-heparin binding domains (HBDs) in the N-terminus of the V2 loop, the C-terminal domain of the protein, and in the CD4-induced bridging sheet. Three of these HBDs are located in protein areas that undergo structural changes upon binding to CD4 and play a role in co-receptor recognition. In addition, replacement of basic amino acids within the C-terminal domain blocks the above mentioned furin-catalyzed cleavage of gp160 into gp41 and gp120 (Crublet et al., 2008, and literature cited therein). In connection with the high-mannose type glycans on the gp120 surface, lectins have also been discussed as potential topical microbicides. The lectin actinohivin, isolated from the actinomycete Longispora albida has been recently investigated by Tanaka et al. (2009) with respect to its possible function as an anti-HIV drug. Actinohivin exhibited a strong and highly specific affinity with a Kdof 3.4 × 10-8 M toward the many Man8/Man9 units present on the envelope protein. As derived from X-ray studies, this strong affinity is due to multivalent interactions of three sugar-binding pockets with three high-mannose type glycans of gp120 via what is known as the “cluster effect” of lectins. The above mentioned properties of HIV-1, e.g., its extraordinary diversity (the virus with a thousand faces; Korber et al., 2009) and its ability to circumvent adaptive immune responses, are the reasons why a vaccine protecting against HIV-1 infection could not be developed so far despite advancements made with respect to an understanding of HIV-1 pathogenesis, in part derived from investigations into the simian immunodeficiency virus replication (SIV) in rhesus macaques (macaque model). New vaccine strategies based on the importance of HIV-specific CD8+cytotoxic T-lymphocyte responses and including so-called mosaic vaccines have been recently described by Santra et al. (2010) and by Korber et al. (2009), and reviewed by Barouch (2008); for an

excellent introduction into this field see McMichael (2006). For further aspects concerning this topic see, e.g., Goonetilleke and McMichael (2013), and D’Souza and Yang (2015) about adenovirus vectors as HIV-1 vaccines. Guzzo et al. (2013) characterized the novel anti-HIV chemokine, XCL1/lymphotactin, produced primarily by activated CD8+ T cells; of the two conformations adopted by XCL1, interacting directly with the virus envelope, only the alternativelyfolded (all β-sheet) structure of XCL1 is responsible for inhibiting a broad spectrum of HIV-1 isolates. Mechanisms affecting the anti-HIV CD8+ T-cell response in HIV-1 infection have been discussed by Chiozzini et al. (2014). 11.2.3 Neurodegenarative Diseases and GAGsTo the so-called dark sides of heparin and heparan sulfate belongs their involvement in the formation of amyloid (fibrillar proteins) deposits in Alzheimer’s disease, associated with neuronal and

vascular degenerations (Capila and Linhardt, 2002; Bishop et al., 2007; Papy-Garcia et al., 2011; Zhang et al., 2014). Amyloidb-peptides consisting of either 39 or 42 amino acids are the prevailing components in senile plaques and vascular deposits; they are generated from the amyloid precursor protein (APP, an integral membrane protein expressed in many tissues, including particularly the synapses of neurons) by the proteolytic activity of b-and g-secretase. Interestingly, Donmez et al. (2010) reported that an overexpression of the NAD-dependent deacetylase SIRT1 (one of seven known mammalian sirtuins) in the brain suppresses b-amyloid production through activation of the gene ADAM10 encoding a-secretase. The therapeutic potential of upregulating SIRT 1 for Alzheimer’s-disease-type diseases has been discussed by Bonda et al. (2011), and Ng et al. (2015) reviewed recent findings concerning links between SIRT1, aging-associated disorders and lifespan. Only some 30 years ago, HSPGs were reported to be constituents of amyloids (Snow et al., 1987, 1988); meanwhile, GAGs and proteoglycans (Chapter 2) have been found in all kinds of amyloid deposits, e.g., prion protein amyloid plaques of Gerstmann-Straussler syndrome, Creutzfeldt-Jakob disease, and scrapie (Snow and Wight, 1990). Inhibition experiments with different heparins and recombinant prion protein revealed information about the nature of GAG-chain binding: 2-O-sulfate groups turned out to be essential for heparin recognition, rather than 6-O-sulfate residues and three amino acid motifs were identified being capable of binding HS and heparins; one of these motifs exhibited enhanced affinity to the substrates in presence of Cu2+ ions, whereas a peptide of this sequence inhibits binding of the prion protein to heparins (Warner et al., 2002). Of the different subtypes of HSPGs, agrin and glypican-1 (Chapter 2, Section 2.6.2) are expressed in both non-fibrillar (diffuse) and fibrillar (classic) cerebellar senile plaques, syndecans-1 and -3 only in fibrillar ones, and perlecan is absent in all cerebellar senile plaques (van Horssen et al., 2002). The situation is similar for cerebral amyloid angiopathy, were agrin, syndecan-2, and glypican-1 predominate. The fateful role HSPGs play, among others, in Alzheimer’s disease pathology is that they protect amyloids against proteolytic clearance by forming a shield around them. Furthermore, perlecan and agrin promote the conversion of primarily built non-fibrillar to classical

fibrillar amyloid-b-peptides by polymerization, thus contributing to the neurotoxic behavior of amyloid-b (van Horssen et al., 2003, and literature cited therein). This was impressively confirmed by Li and coworkers (2005). They induced inflammation-associated AA (serum amyloid A caused) amyloidosis by s.c. injection of AgNO3in transgenic (tr) mice, overexpressing heparanase, and in a mice control (ctr) group. After 7 days, kidney, liver and spleen were analyzed with respect to heparanase protein activity and amyloid deposition. Both parameters were nearly equal in the spleens of tr and ctr mice; however, in kidney and liver of tr mice, nearly no amyloid deposition was found, whereas it was high in the respective organs of the wild-type animals. An analysis of the size of HS revealed a reduction from an average of about 35 kDa (crt) to ≈3 kDa which led the authors propose a model of HS function in connection with amyloid formation, shown in the scheme of the previous page (modified after Li et al., 2005). This finding, according to which in vivo fragmentation of HS chains by heparanase overexpression renders mice resistant to amyloidosis, suggests that efficient polymerization of amyloid monomers requires a critical HS chain length. As the transgenic mice showed no significant phenotypic alterations compared to the control group individuals, the remaining chain length of HS after its fragmentation by heparanase seems still to be sufficient for sustaining the many regulatory functions of HS, mediated by interaction with growth factors and other biologically active molecules. Major research concerning therapeutic approaches of AD and other neurodegenerative diseases is, among others, on g-secretase inhibitors (Jakob-Roetne and Jacobsen, 2009) and on the development of inhibitors of amyloid fibril formation (Bulic et al., 2009). This research is supported by insights into the supramolecular structures of amyloid fibrils obtained from AFM imaging, X-ray fiber diffraction, and solid-state NMR spectroscopy (Nielsen et al., 2009). New results, shedding light on the toxicity of amyloid-forming proteins mediated by the surface of cell membranes and associated with a damage of the membrane structure, have been reviewed by Butterfield and Lashuel (2010). Based on known atomic structures of segments of amyloid fibrils as templates, Sievers et al. (2011) designed non-natural amino-acid inhibitors of amyloid fibril formation.