Glycans As Novel Therapeutic Targets Cancer Intelligence Acquired

Assigned Reading:

The Sweet and Sour of Cancer: Glycans As Novel Therapeutic Targets

Cancer Intelligence Acquired (CIA): tumor glycosylation and sialylation codes dismantling antitumor defense

Oncology meets immunology: The Cancer Immunity Cycle

REVIEWS

THE SWEET AND SOUR OF CANCER: GLYCANS AS NOVEL THERAPEUTIC TARGETS

Mark M. Fuster* and Jeffrey D. Esko‡ Abstract | A growing body of evidence supports crucial roles for glycans at various pathophysiological steps of tumour progression. Glycans regulate tumour proliferation, invasion, haematogenous metastasis and angiogenesis, and increased understanding of these roles sets the stage for developing pharmaceutical agents that target these molecules. Such novel agents might be used alone or in combination with operative and/or chemoradiation strategies for treating cancer.

GLYCOCONJUGATE Several glycans, on both the tumour surface and host membrane bound MUCINS; the glycosaminoglycans, A molecule in which one or elements, have now been identified as mediating key which are glycans present as free polysaccharides more glycan units are covalently pathophysiological events during the various steps of (such as hyaluronan) or as part of PROTEOGLYCANS (such linked to a non-carbohydrate tumour progression. Tumour progression involves a as heparan sulphate and chondroitin sulphate); the entity. range of unique alterations in intracellular and inter- glycosphingolipids, which consist of oligosaccharides NLINKED GLYCANS cellular signalling. These serve to promote dysregu- glycosidically linked to ceramide; the glycosylphos- Glycans covalently linked to an lation of the cell cycle and facilitate proliferation; to phatidylinositol (GPI)-linked proteins, which are asparagine residue of a promote the emergence of a subset of invasive cells proteins that bear a glycan chain linked to phosphati- polypeptide chain in the consensus sequence that dissociate from the tumour and digest and migrate dylinositol; and nuclear and cytoplasmic proteins, –Asn–X–Ser/Thr. through host extracellular matrix (ECM) and base- which bear the monosaccharide O-linked N-acetylglu- ment membranes; to summon an endothelial-lined cosamine (O-GlcNAc) linked to serine, often at sites GLYCOPROTEIN neovascular network from nearby host endothelial that are normally phosphorylated1 (FIG. 1). Most classes A protein with one or more cells (angiogenesis); to endow disseminating tumour of glycan exist as membrane-bound glycoconjugates covalently bound glycans. cells with cell-surface characteristics that promote (for example, in the GLYCOCALYX) or as secreted mol- *Department of Medicine, adhesive interactions with platelets, leukocytes and ecules, which can become integral parts of the ECM. Division of Pulmonary and blood or lymphatic vascular endothelial cells; and to These locations place glycans in a position to medi- Critical Care Medicine, University of California, facilitate the evasion of innate immunity. Ultimately, a ate cell adhesion and motility, as well as intracellular 2 San Diego, La Jolla, ‘survivor’ subset of cells must invade, neovascularize, signalling events . California, 92093-0687, disseminate, extravasate and proliferate at a new tissue In the tumour environment, changes in glycosylation USA. location to become a pathological metastasis. allow neoplastic cells to usurp many of the events that ‡ Department of Cellular Glycans are covalent assemblies of sugars (oligosac- occur in development (for example, receptor activation, and Molecular Medicine, Glycobiology Research and charides and polysaccharides) that exist in either free cell adhesion and cell motility), which allows tumour Training Center, University form or in covalent complexes with proteins or lipids cells to invade and spread throughout the organism3. of California, San Diego, GLYCOCONJUGATES. There are several main families of Malignant transformation is often accompanied by La Jolla, California, glycoconjugates: the Asn-linked NLINKED oligosac- the expression of oncofetal antigens — epitopes that 92093-0687, USA. Correspondence to J.D.E. charides of many GLYCOPROTEINS; the Ser- or Thr-linked are expressed on embryonic tissues and tumour cells, e-mail: [email protected] OLINKED oligosaccharides that are present on many and only in a few cell types in the adult. Many of the doi:10.1038/nrc1649 glycoproteins and that predominate on secreted and first-identified tumour-specific antibodies were directed

REVIEWS

Summary as mannose-binding proteins or galactose-binding lectins (galectins), and many of these can associate • Tumours aberrantly express various glycans. Glycans regulate many different with tumour-cell-associated glycans. The interac- aspects of tumour progression, including proliferation, invasion, angiogenesis tions of lectins with tumour-cell glycans facilitate all and metastasis. aspects of tumour progression. • The proliferation of tumour cells is potentiated by the ability of glycoproteins and This article will mainly focus on examples in which glycosphingolipids to directly activate growth-factor receptor tyrosine kinases and genetic or pharmacological data confirm a correlation by the ability of proteoglycans to function as co-receptors for soluble tumour between a specific pattern of glycan expression and growth factors. tumour progression. We have organized this review • The overexpression of specific glycosyltransferases by tumour cells promotes the according to the main stages of tumour progression: formation of tumour glycans that facilitate invasion. proliferation, invasion, angiogenesis, metastasis and • Carcinomas commonly overexpress O-linked glycans in the form of cell-surface immunity (FIG. 2). We refer readers interested in more and secreted mucins that present ligands for adhesion receptors, such as the details to several excellent reviews on specific glycan selectins, which promote the ability of tumour cells to interact with host platelets, classes and cancer3,7–13. leukocytes and endothelial cells. These interactions facilitate haematogenous metastasis of tumour cells. Proliferation of tumour cells • Glycosphingolipids, in the form of gangliosides, are overexpressed by a range of Tumour cells arise from mutations in proto-oncogenes tumours, and their shedding into the bloodstream might impair host immunity to and tumour suppressors that normally control the cell some tumours. cycle and affect DNA repair14. Tumours often express • During tumour proliferation and invasion, heparan-sulphate proteoglycans a unique repertoire of glycans. These changes might (HSPGs) that are present on the surface of tumour cells function as co-receptors to reflect selection of cell clones that have distinct glycan stabilize growth-factor receptor signalling complexes. Secreted HSPGs that are compositions, but correlating these changes to pro- present in the extracellular matrix store growth factors that can be mobilized by the liferation itself has been difficult. Nevertheless, a few action of tumour heparanases. A similar mechanism that involves endothelial- examples illustrate how alterations in protein or lipid associated HSPGs and endothelial growth factors facilitates vascular sprouting glycosylation can stimulate cell growth TABLE 1. during tumour angiogenesis. • Some glycans can be measured in the bloodstream, and their use as markers of Effects on tumour growth-factor receptors. N-linked disease burden can be used to screen for specific cancers as well as track response glycans have important roles in protein folding, quality to therapy. control and half-life; cell–cell recognition and adhe- 15 • Experiments in which glycan function is genetically altered in cell-culture systems sion; and signalling (FIG. 1). Because N-glycosylation or mouse tumour models validate their potential as targets for anticancer therapy. of glycoproteins is so prevalent, the changes in • A few glycan-based targeting strategies are currently being tested in clinical trials. N-glycan biosynthesis can have pleiotropic effects on As we learn more about the roles of glycans in tumour progression, new targets will many systems, making it difficult to correlate tumour continue to emerge for drug design. growth with a change in glycosylation of a specific protein. For example, N-glycosylation of the insulin- like growth factor 1 receptor (IGF1R) is required for against carbohydrate oncofetal antigens presented on IGF1R phosphorylation, cell-surface translocation, tumour glycoproteins and glycosphingolipids3,4. In and the subsequent growth and survival of melanoma some cases, the underexpression, truncation or altered and sarcoma cells16. Treating cells with N-glycosylation branching patterns of certain glycans correlate with cell inhibitors in vitro can inhibit the survival of tumours growth. Similar alterations on tumours endow them that depend on signalling by IGF1R. with enhanced proliferative capacity, which could reflect Extensive literature exists about the roles of O-linked the outcome of ‘Darwinian selection’ of rapidly grow- glycans and mucins in proliferation (FIG. 1). Mucins are ing cells that can endure survival pressures imposed by glycoproteins that contain numerous O-glycans in clus- the host. A massive potential for glycan diversity exists, tered domains along the core protein. The MUC gene but a relatively limited array of glycans correlates with superfamily (reviewed in REF. 12) is a family of mucin OLINKED GLYCANS invasion and metastatic potential across a wide range core proteins that are expressed in a tissue-specific Glycans glycosidically linked to of tumours. manner and serve as markers for epithelial cells. Most the hydroxyl group of the Initial insight into the unique repertoire of gly- carcinomas express mucins, either as transmembrane amino acids serine, threonine, tyrosine or hydroxylysine. cans expressed on tumour cells emerged from the proteins on the cell surface or as secreted proteins. increased ability of tumours to bind a range of plant MUC expression often goes awry in the tumour. For MUCINS LECTINS5. Lectins exhibit protein folds that define example, mammary tumours overexpress MUC4 on Large glycoproteins with a high families of carbohydrate-binding proteins that can the cell surface17, whereas it is not expressed on the content of serine, threonine and bind in a specific ‘lock-and-key’ fashion6. Various normal mammary epithelium. MUC4 contains an epi- proline residues, and numerous O-linked glycans, often endogenous animal lectins also exist, and these facili- dermal growth factor (EGF)-like motif on its extracel- occurring in clusters on the tate fundamental processes such as quality control lular (juxtamembrane) domain that directly interacts polypeptide. Tumour mucins of secreted proteins, cell–cell recognition, cell adhe- with ERBB2 (a member of the EGFRECEPTOR FAMILY), are often decorated by unique sion and motility, and pathogen–host recognition. initiating phosphorylation of the receptor tyrosine small glycans such as Tn (O-linked GalNAc) or sialyl Tn Many lectins exist on the surface of immune cells and kinase in the absence of more typical ERBB ligands 12 antigens (sialic acid-capped endothelial cells that line the vasculature, as ECM (such as EGF) . ERBB2 is frequently overexpressed (or O-linked GalNAc). proteins and as soluble adhesion molecules such mutated to the activated state) in breast carcinomas.

REVIEWS

e Hyaluronan

f Proteoglycans

P S 4S 4S 4S 4S Heparan sulphate Ser-O

6S 6S6S 6S 6S Chondroitin sulphate O-Ser NS NS 2SNS 2SNS NS 3S b O-linked a N-linked glycans glycans d GPI-anchored proteins sLeA/X

c Glycosphingolipids

Etn GD3 GD2 P

GM3 GM1 STn GnTV Tn NH2 Inositol O O O N P Ser/Thr Ser Ser Asn

PROTEOGLYCAN A protein with one or more covalently attached glycosaminoglycan chains, such as heparan sulphate or Cytoplasm chondroitin sulphate (and dermatan sulphate). In the O tumour environment, a range of Ser heparan-sulphate proteoglycans g O-linked GIcNAc expressed by both tumour cells as well as endothelial cells affect growth-factor signalling. Glc Man GalNAc IdoA Xyl Gal GlcNAc GlcA Fuc Sia GLYCOCALYX The cell-coat structure Figure 1 | Important glycans involved in tumour progression. Unique glycans are involved in promoting the progression of consisting of glycans and various carcinomas. a | N-linked glycans on glycoproteins are covalently bound to Asn residues. Typical branched structures glycoconjugates surrounding contain two or more ‘antennae’. The enzyme N-acetylglucosaninyltransferase V (GnTV) generates a specific antenna on some animal cells. This is seen as an glycoproteins and has been implicated in tumour invasion. b | O-linked glycans are found covalently linked to Ser or Thr electron-dense layer by electron residues on glycoproteins and mucins. SLeX/A (sialyl Lewis X or A) are carbohydrate determinants composed of four sugars in microscopy. specific linkage to one another, and are commonly overexpressed on tumour-cell mucins. The determinant forms part of a ligand for the selectin class of adhesion receptors involved in tumour-cell aggregation with leukocytes and platelets, and LECTIN adhesion of tumour cells to endothelial cells. Tn and STn are tumour antigens that consist of truncated O-linked chains. Their A protein (other than an anti- accumulation in many tumours correlates with invasion. c | Glycosphingolipids consist of the lipid ceramide linked to one or carbohydrate antibody) that more sugars. Certain sialic-acid-containing glycosphingolipids, called gangliosides (for example, GM , GM , GD and GD ), specifically recognizes and 1 3 2 3 have been correlated with tumour growth. d | Glycosylphosphatidylinositol (GPI)-linked proteins are anchored in the outer binds to glycans without leaflet of the plasma membrane by a glycan covalently linked to phosphatidylinositol. Glycosaminoglycans can occur as free catalysing a modification of the chains (hyaluronan; e) or as covalent complexes with proteoglycan core proteins (heparan sulphate, chondroitin sulphate and glycan. dermatan sulphate, a type of chondroitin-sulphate-containing iduronic acid (IdoA)). f | Proteoglycans participate in growth- EGFRECEPTOR FAMILY factor activation and cell adhesion. g | Various cytoplasmic and nuclear proteins contain O-linked N-acetylglucosamine Epidermal growth factor (O-GlcNAc). Some glycoconjugates can be tethered to the plasma membrane as depicted or secreted into the extracellular receptors have important roles matrix. In some cases, hybrid molecules exist, containing more than one type of glycan. Sugars are represented by coloured in initiating the signalling that geometric symbols. Glc, glucose; Gal, galactose; Man, mannose; GlcNAc, N-acetylglucosamine; GalNAc, directs the behaviour of N-acetylgalactosamine; GlcA, glucuronic acid; Fuc, fucose; Xyl, xylose; Sia, sialic acid. epithelial cells and tumours of epithelial origin. The four members of the family are also Recent experiments show that the overexpression protein folding or exposure of the EGF-like domains known as ERBB receptor tyrosine kinases (ERBB1–4), of MUC4 in melanoma cells induces rapid cellular on the MUC4 backbone in a manner that promotes and share structural and growth and the suppression of tumour apoptosis in interaction with ERBB2. functional similarities. human melanoma-bearing nude mice18. Furthermore, Many tumours overexpress glycosphingolip- MUC4-overexpressing cells can autophosphorylate ids, especially GANGLIOSIDES, which are ‘capped’ with GANGLIOSIDES 18 Anionic glycosphingolipids ERBB2, which contributes to inhibition of apoptosis . sialic acid (a negatively charged acidic sugar) on the containing one or more residues The specific role of O-glycosylation in this process outermost tip of the glycan (FIG. 1). In normal cells, of sialic acid. is unknown. The glycan moiety might affect MUC4 gangliosides are often associated with various receptor

REVIEWS

Transformed complexes in LIPID RAFTS (FIG. 3). In particular, the gan- cell glioside G is necessary for ERBB2 and ERBB3 recep- a Tumour-cell M1 proliferation tors to form growth-factor-responsive heterodimers Primary tumour on lipid rafts20, thereby facilitating ERBB signalling. The ERBB receptors provide an example in which the surrounding milieu of glycans (in this case glycosphingolipids) and a specific protein–glycoprotein (in this case MUC4) interaction can modulate the function of a crucial growth receptor. Basement membrane Glycans as co-receptors for soluble tumour growth fac- b Tumour-cell tors. Proteoglycans contain a core protein and one or dissociation more covalently attached glycosaminoglycan (GAG) and invasion side chains (FIG. 1). Unlike N- or O-linked glycans, GAGs are composed of repeating disaccharide units that are composed of an amino sugar (GlcNAc in heparan sulphate or GalNAc in chondroitin sulphate; see FIG. 1) linked to a uronic acid (GlcA or IdoA). Host Negatively charged sulphate groups at discrete posi- lymphocyte tions along heparan-sulphate chains facilitate interactions with basic amino-acid residues on a range of Platelets protein ligands involved in cell–matrix interactions (for example, laminin, fibronectin and thrombospon- c Tumour-cell din), inflammation (for example, the P and L SELECTINS) adhesion and metastasis and growth (for example, the fibroblast growth factors (FGFs), vascular endothelial growth factor (VEGF), transforming growth factor-β (TGFβ) and interleukin- 8 (IL-8))21–23. GAGs are able to facilitate the formation of ligand–receptor complexes, lowering the effective Extracellular concentration of ligand required for receptor activa- d Angiogenesis matrix tion. In this way, GAGs act as ‘co-receptors’. GAGs also facilitate the storage of ligands for future mobilization and the protection of ligands from degradation21,22. These functions are usurped in the tumour environment, allowing tumour-released growth factors to exert autocrine effects as well as paracrine effects on surrounding host cells, such as endothelial cells that also bear growth-factor receptors. The proliferation of tumour cells depends on Metastatic tumour heparan-sulphate proteoglycans (HSPGs). For example, glypican-1 (a GPI-anchored HSPG; FIG. 1) is over- Figure 2 | Stages of tumour progression. Tumour proliferation (a) is crucial at early stages of produced by pancreatic cancer cells, and mediates progression. b | During invasion, tumour cells gain the capacity to degrade, and migrate mitogenic responses by tumour cells to basic fibroblast through, basement membranes and extracellular matrix. c | During the dissemination of tumour growth factor (FGF2) and heparin-binding EGF-like cells through the bloodstream, they aggregate with host cells such as platelets and growth factor (HB-EGF) by facilitating the formation of lymphocytes and eventually lodge in the small vessels of distant organs. d | Tumour ligand–receptor complexes24. This interaction occurs at angiogenesis is required for pathological growth of the primary cancer and its metastases. Glycans have roles in each of these stages of tumour progression. See FIG. 3 for more details. specific sulphate-modified domains along cell-surface heparan-sulphate chains22. In pancreatic, breast, ovarian and hepatocellular cancers, tumour cells regulate tyrosine kinases (such as EGF or insulin receptors) and genes that modulate sulfation of cell-surface HSPGs LIPID RAFTS Microdomains in the plasma modulate their phosphorylation. Evidence indicates in a manner that increases their binding capacity for membrane that are enriched in that the interaction between gangliosides and receptor growth-factors and the activation of receptor tyrosine sphingolipids, cholesterol and tyrosine kinases serves important growth-promot- kinases25. Other examples are provided in TABLE 2. GPI-linked proteins. They ing functions. For example, the gangliosides G or In some cells, HSPGs can have the opposite function as signalling platforms M3 through their ability to GD3 — which are commonly overexpressed in lung effect, acting as tumour suppressors. For example, concentrate signalling proteins, cancers, melanomas and neurogenic tumours — are in Simpson–Golabi–Behmel syndrome, deletions or resulting in increased output able to regulate growth signalling through interactions point mutations within the glypican-3 gene cause a from receptors that require with membrane-associated receptor tyrosine kinases congenital overgrowth syndrome, and patients have cross-activating interactions and 3,8,19 26 increasing local concentrations or protein kinase C . At a higher level of molecu- an increased risk of developing certain malignancies . of other downstream signalling lar organization, gangliosides on breast cancer cell It is unclear how glypican-3 modulates growth, but components. membranes regulate the formation of EGF receptor it appears to modify epithelial responses to growth

REVIEWS

Table 1 | Examples of glycan families involved in tumour progression Glycan involved Proposed major function(s) Possible therapeutic targeting Examples of References neoplasms Growth and proliferation N-glycans Suppress apoptosis; growth-factor Alkaloid inhibitors of N-linked Breast, melanoma, 16,18 signalling processing Ewing’s sarcoma O-glycans Mucin (MUC4)-mediated activation of Immunotherapy targeting MUC4 Breast 18 ERBB2 receptors (similar to other mucin-targeting immunotherapy) O-glycans Suppress apoptosis (possibly through Galectin-3 inhibitors (β-galactosides) Colon, pancreatic 150 galectin-3 binding to tumour O- glycans expressing terminal galactose) Glycosphingolipids Control of signalling through lipid rafts Ceramide glycosylation inhibitors; Breast 20 ganglioside-targeted vaccines Heparan-sulphate Coreceptors for tumour growth Heparin derivatives as Pancreatic, ovarian, 24,25 proteoglycans factors heparan-sulphate competitors; renal, hepatic sulphotransferase inhibitors Hyaluronan Signaling through hyaluronan Hyaluronan oligomers; adenoviral Colon, breast 13,30 receptors (for example, CD44) delivery of hyaluronan-binding protein genes O-GlcNAc Modify oncogene phosphorylation O-GlcNAc transferase inhibitors Pancreatic 31 Tumour invasion N-glycans Alter E-cadherin-dependent tumour Alkaloid inhibitors of N-glycan Breast, colon 41,74 adhesion processing N-glycans Tumour repulsion (for example, Sialyltransferase inhibitors Neuroblastoma, 43 polysialylation) lung (small cell) O-glycans Stimulate migration; potentiate Vaccines (for example, conjugated Breast, gastric, 47 migration of tumour cells through sialyl Tn) ovarian inhibition of cell–cell contacts (for example, sialyl Tn on mucins)

Glycosphingolipids Tumour repulsion (for example, GM3) Glycosphingolipid inhibitors; Melanoma, 35,36, ganglioside-targeted vaccines neuroblastoma, 134,151 breast Heparan-sulphate Matrix growth factor storage Heparin fragments and analogues; Breast, colon, 68,69,72 proteoglycans (heparanase substrate) sulphotransferase inhibitors; xylosides; hepatic, lymphoma, antisense RNA to perlecan melanoma Chondroitin-sulphate Modulate tumour–matrix attachment Xylosides Melanoma, glioma, 61–63 proteoglycans lung Hyaluronan Coordinate tumour growth signalling Target tumour hyaluronan receptors Breast 28 with cytoskeletal events during (for example, gene silencing of CD44) migration Tumour metastasis O-glycans Facilitate tumour adhesion during Disaccharide primers of glycosylation Colon 101,112,118 haematogenous metastasis (SLeX, (reduce tumour SLeX); competition by SLeA and other selectin ligands); intravenous heparin N-linked and O-linked Promote tumour aggregation Galectin-3 inhibitors (β-galactosides) Melanoma 150 glycans (galectin-3 binding) Glycosphingolipids Tumour adhesion (sulphated selectin Disaccharide primers; competition with Colon 101,118 ligands) heparin Tumour angiogenesis N-glycans Promote migration of endothelia Alkaloid inhibitors of N-linked Prostate 152 glycosylation Heparan-sulphate Co-receptor for growth factors; matrix Heparin fragments and analogues; Colon, renal, 9,71 proteoglycans growth factor storage; co-receptor for sulphotransferase inhibitors; xylosides; melanoma, breast matrix proteins antisense RNA to perlecan Tumour immunity Glycosphingolipids Immune ‘silencing’ (ganglioside Ganglioside vaccines Melanoma, 35,36,134 shedding) neuroblastoma, breast O-GlcNAc, O-linked N-acetylglucosamine; SLe, sialyl Lewis.

REVIEWS

and differentiation factors such as bone morphogenic targeting approaches might be considered for novel proteins27. Therefore, HSPGs might have important drug development. MUC4 represents a possible future roles in maintaining epithelial differentiation and target for cancer immunotherapy using anti-mucin suppressing progression to malignancy. Similarly, monoclonal antibodies or mucin peptide-based vac- heterozygous mutations in EXT1 and EXT2, which cines. Other tumour-associated mucins are now being encode the heparan-sulphate copolymerase, result in targeted in clinical trials. MUC16 (which carries the benign osteochondromas in hereditary multiple exos- CA125 tumour antigen) is currently a target for anti- toses, a paediatric bone disorder in which patients body-based therapy (in Phase II testing). Peptide-based are predisposed to developing malignant chond- vaccine therapy for targeting MUC1 is in Phase III test- rosarcomas. The signalling pathway that underlies ing for ovarian and breast carcinomas12,32,33. In addition, this disorder is unknown, but might involve Indian small monosaccaride and disaccharide decoys — Hedgehog (IHH) or FGF18 (known to bind heparan glycan substrates designed to divert (and thereby alter) sulphate), which coordinate cartilage growth and endogenous glycan synthesis — might also be useful bone deposition. Whether HSPGs act as tumour pro- for blocking tumour growth through their ability to moters or suppressors probably depends on the array alter O-glycans expressed on tumour mucins BOX 1. of growth factors and the cell-type-specific regulation The use of these mucin-targeting approaches might of heparan-sulphate formation. be limited because secreted- and membrane-bound mucins have important roles in the homeostasis of Growth-regulatory effects of plasma membrane-, normal epithelial barriers12. Nevertheless, safety has cytosolic- and nuclear-assembled glycans. The glycans been demonstrated in Phase I studies of MUC1 pep- of glycoproteins, proteoglycans and glycosphingolip- tide-based vaccine34. Tumour gangliosides are also ids are assembled as they pass through the endoplas- proving to be appealing targets for tumour-growth mic reticulum and Golgi, but a few examples exist inhibition through vaccine approaches35,36. in which glycosylation occurs in the plasma mem- Interference with the co-receptor activity of HSPGs brane, the cytoplasm or the nucleus. Hyaluronan, a on tumour cells represents a novel therapeutic strategy large, anionic glycosaminoglycan, is polymerized at for altering tumour proliferation. This might be achieved, the plasma membrane and secreted into the ECM. for example, through the inhibition of sulphotransferases Tumour matrices are especially rich in hyaluronan. that are responsible for sulphating heparan-sulphate Interactions between hyaluronan and its main recep- chains during biosynthesis, or through the design of tor CD44 (a transmembrane glycoprotein) as well as competitive blocking agents such as heparan-sulphate CD168, activate signal-transduction pathways that mimetics (for example, heparin-based derivatives that promote cytoskeletal changes involved in cell motil- have structural similarities to heparan sulphate). One ity and growth28. Studies of tumours transfected with interesting approach is to combine an inhibitor of HSPG an anti-hyaluronan synthase cDNA have shown synthesis with difluoromethylornithine, an inhibitor that hyaluronan has an important role in tumour of polyamine formation37,38. Polyamines are aliphatic proliferation and survival29. Receptor activation by cations that are synthesized by all cells during urea tumour hyaluronan glycans promotes signalling and metabolism and can be taken up from the extracellular proliferation through the mitogen-activated protein environment. Polyamines function as tumour growth kinase pathway and the phosphatidylinositol 3-kinase factors through their ability to modulate gene expression (PI3K)–AKT survival pathway. Interestingly, hyaluro- and signal transduction37. As polyamine uptake depends nan oligomers compete with endogenous polymeric in part on HSPGs, the combination of the two drugs is hyaluronan and block signalling responses in breast sufficient to block tumour growth through inhibition tumour cells30. of tumour polyamine uptake and synthesis38. Because Many soluble oncogene products and tumour- HSPGs appear to serve important roles in multiple steps suppressor proteins contain O-GlcNAc (FIG. 1). The during tumour progression, the targeting of HSPGs is addition of O-GlcNAc to specific serine residues on discussed in more detail below. the MYC proto-oncogene protein by O-GlcNAc trans- Little is known about possible approaches to target ferase promotes mitogenesis in a manner similar to that hyaluronan in human cancer, although one promising promoted by phosphorylation of the same amino acids finding demonstrated the ability of hyaluronan oli- SELECTINS 31 39 C-type calcium-dependent by protein kinases . In another example, O-GlcNAc gomers to inhibit the growth of melanoma in mice . lectin expressed by cells in the modification of the tumour-suppressor protein p53 Treatment with either hyaluronan oligomers or the ade- vasculature and bloodstream. appears to mask or block its ability to bind to crucial noviral administration of hyaluronan-binding protein The three known selectins are regions of DNA, and this also promotes growth31. These decoys provide other possible therapeutic avenues13. L-selectin/CD62L (expressed by complementary effects of modification by O-GlcNAc most leukocytes), E-selectin/ CD62E (expressed by cytokine- on major tumour cell-cycle effector proteins make Tumour invasion activated endothelial cells) and tumour O-GlcNAc transferase an appealing target for During invasion, tumour cells detach from one P-selectin/CD62P (expressed by future cancer therapy. another and from the ECM and migrate through activated endothelial cells and neighbouring tissue. This requires the remodelling platelets). Important ligands for the selectins include glycans Therapeutic targeting of glycans that affect tumour of cell-surface adhesion receptors and ligands, and containing sialyl Lewis X and proliferation. Given the broad spectrum of glycans that the secretion of proteolytic enzymes and glycosidases sialyl Lewis A. affect tumour proliferation, a range of glycan-specific (which catalyse the hydrolysis of glycosidic bonds in

REVIEWS

a Tumour-cell b Tumour-cell dissociation and invasion proliferation

Etn P

Plasma membrane

NH2 ERBB2–ERBB3

Inositol GM STn-rich mucin P 1 ST6GaINAcI

Lipid raft

Sphingomyelin Signalling event Cholesterol Proliferation Phospholipid

c Tumour-cell adhesion and metastasis d Angiogenesis Resting platelet VEGF

Leukocyte VEGF Activated platelet VEGF

P L Proteoglycan

P VEGF VEGF Adenocarcinoma P Heparan sulphate P Endothelial P cell E Activated endothelium Signalling event Mitogenesis and sprouting

Figure 3 | Glycans participate in major pathophysiological events during tumour progression. Tumour proliferation is crucial during early stages of progression and after metastasis, when secondary tumours form at sites distant from the primary tumour. a | Various growth-factor receptors are modulated by glycans, for example, through oligomerization of their respective receptor tyrosine kinase receptors in lipid rafts mediated by gangliosides. The figure illustrates the ability of

the ganglioside GM1 to promote ERBB2–ERBB3 receptor heterodimerization. Glycosylphosphatidylinositol-anchored proteins also co-localize in rafts. b | Tumour invasion occurs when cells gain the capacity to degrade and migrate through basement membranes and extracellular matrix. Invasion is also associated with the expression of sialylated glycans that promote dissociation of tumour cells. STn is one such class of sialylated glycan overexpressed on tumour cells. c | Dissemination of tumour cells through the bloodstream is facilitated by host elements such as platelets and lymphocytes that promote embolization and arrest of tumour at distant endothelial sites. Selectins are important adhesion receptors expressed on activated platelets (P-selectin), leukocytes (L-selectin) and endothelial cells (E-selectin) that bind to specific glycan receptors containing sialyl Lewis X (SLeX) or SLeA (expressed in the respective tumour ligands labelled ‘P,’ ‘L’, and ‘E’ in the figure). d | Tumour angiogenesis is required for pathological growth of the primary cancer and its metastases. Angiogenesis is the sprouting of blood-vessel endothelial cells in response to pro-angiogenic factors such as vascular endothelial growth factor and fibroblast growth factor (released from the tumour cells and from the host stroma). Heparan sulphate on the endothelial-cell surface facilitates growth-factor binding and activation of endothelial receptor tyrosine kinases.

REVIEWS

glycans) to degrade ECM components. Glycans are potentiates tumour invasiveness. Recent studies have involved at each of these stages TABLE 1. Tumours shown that a common defect in many tumours is loss also generate glycoconjugates that appear to serve a of a chaperone called COSMC, which is required for repulsive role, physically facilitating the detachment maturation of the galactosyltransferase that acts on of invasive cells from the primary tumour. the core GalNAc residue of O-linked glycans46. Lack of COSMC results in the accumulation of αGalNAc- Cell–cell interactions. Malignant cells commonly Ser linkages (Tn antigen), which in turn allows the exhibit increased expression of complex β1,6-branched addition of sialic acid, thereby forming STn antigen. N-linked glycans on their cell surface (FIG. 1), caused Transfection of breast cancer cells with cDNA that by an increase in the expression of the enzyme encodes the human sialyltransferase responsible for N-acetylglucosaminyltransferase V (GnTV). In the STn biosynthesis (ST6GalNAc-I) reduces cell–cell Golgi, this enzyme transfers a GlcNAc residue onto interactions and increases the migration potential growing N-linked glycans so that subsequent gly- of transfected cells47. STn antigen is also associated cosylation results in ‘multi-antennary’ chains15. The with increased metastatic potential, and therefore presence of such complex β1,6-branched N-glycans poor prognosis, in colorectal, gastric and ovarian on tumour-cell E-cadherin — an adhesion molecule carcinomas TABLE 1. that normally mediates cell aggregation through homotypic interactions — reduces tumour cell–cell Cell–matrix interactions. Invasion by tumour cells adhesion. Therefore, increased expression of this involves the alteration of cell–matrix interactions, enzyme promotes cell detachment and invasion3,8,40,41. which are mediated by adhesion molecules present Interestingly, E-cadherin is often downregulated by on the tumour cells that bind to ECM components. invasive cancers42. Therefore, the coordinated control Integrins represent a particularly important class of of E-cadherin and GnTV expression can affect the cell-surface adhesion receptors that mediate attachment invasive characteristics of a tumour. The GnTV modi- to important ECM protein ligands, such as fibronectin fication also affects other molecules that are important and laminin. Increased GnTV expression is a hallmark in modulating tumour adhesion, including the integrin of many malignant cells, and results in increased β1,6 β α β family of adhesion receptors, as discussed below. branching on the 1 subunit of tumour 5 1 integrins, Tumour cells tend to produce increased levels of and this disrupts the ability of integrins to cluster on the glycoconjugates containing sialic acid, an acidic sugar tumour-cell membrane48. Altered integrin clustering, in that imparts a negative charge to the glycan chain turn, reduces the formation of tumour-cell focal adhe- (FIG. 1). The formation of polysialic acid also occurs sions, and this increases tumour motility through the and it is overexpressed on some tumour cells7,43. ECM as well as invasion across basement membranes. Enhanced sialic acid expression might promote cell Focal adhesions are supramolecular complexes that link detachment from the tumour mass through charge integrin-mediated ECM adhesion to the cytoskeleton repulsion, which physically inhibits cell–cell appo- as well as initiate signalling pathways that affect cell sition. Polysialylation is often associated with the migration, mechanosensing, and proliferation during increased invasive potential of tumour cells in both cell motility49. cultured cell lines as well as clinical tumours, and its Whereas the disruption of tumour-cell focal expression correlates with poor prognosis7,43,44. adhesions can promote migration, in some cases Sialic acid capping of terminal galactose residues the formation of focal adhesions promotes invasion on N-linked glycans by the enzyme ST6Gal-I (which is by facilitating tumour-cell spreading on the ECM. upregulated in human breast and colon malignancies) During this process, HSPGs on the surface of tumour has also been reported to alter tumour cell–cell interac- cells work together with integrins to form focal adhe- tions in a manner that promotes invasion. Transfection sions7,23,50–56. Syndecans are a class of cell-surface- of breast cancer cells with this enzyme increases cell bound HSPGs that have particularly important roles migration and reduces cell–cell adhesion, whereas in facilitating these contacts50–52,57,58. Syndecan-4 is transfection with antisense ST6Gal-I RNA enhances frequently upregulated in a range of malignancies, homotypic cell–cell adhesion45. Tumour sialic acids including hepatocellular carcinomas and malignant could also potentiate invasiveness by receptor-depend- mesotheliomas. It binds to fibronectin and laminin ent processes or by facilitating interactions between and enhances the function of β1 integrins during cell tumour sialic acids and matrix proteins (such as lam- spreading on matrix52,58. Syndecan-1 is overexpressed inin or fibronectin), but additional studies are needed by a range of human tumours58, including pancreatic, to determine if these processes actually occur. gastric and breast carcinomas. Its functional cou- α β In tumours, sialic acids are also found on O-linked pling with v 3 integrins on breast carcinoma cells α β 59 sugars in structures that are rare in normal tissues. results in v 3-dependent spreading and migration . A well-known example involves the sialyl Tn antigen Interestingly, in some tumours the downregulation of (STn), a disaccharide commonly expressed on tumour HSPG on cells can promote anchorage-independent mucin backbones that consists of the O-linked core growth57 and increase invasive potential58,60. Therefore, sugar αGalNAc ‘capped’ by a sialic acid residue. This the activity of HSPGs is context dependent. It is also glycan is commonly overexpressed in the mucin-rich possible that coordinated changes in the expression of surfaces of cancer cells (FIG. 1), and its expression specific HSPGs over time might promote sequential

REVIEWS

adhesion (which is necessary for tumour-cell spread- such as perlecan56,68,69 TABLE 2. Indeed, the release ing) followed by disengagement (which is necessary for of heparanase by a range of tumours correlates with motility) during tumour invasion. metastatic potential70,73. Other glycans that might have key roles in tumour–ECM interactions during invasion include Therapeutic targeting of glycans during tumour inva- chondroitin-sulphate proteoglycans (CSPGs) and sion. From the above discussion, therapeutic target- hyaluronan (FIG. 1). Chondroitin sulphate is an impor- ing of three major glycosylation pathways should be tant constituent of certain tissue matrices (for exam- considered as anti-invasion approaches: blockade of ple, cartilage). Recent studies on brain neoplasms and tumour N-glycosylation (specifically focusing on the melanomas show that CSPGs on both the tumour-cell inhibition of GnTV); blockade of sialylation pathways surface and the ECM facilitate tumour invasion by in tumours (that is, those that generate polysialic acid enhancing integrin-mediated cell adhesion, motil- or STn); and targeting approaches aimed at HSPGs ity and intracellular signalling13,61–63. Inhibition of and CSPGs. The inhibition of tumour hyaluronan the motility and invasion of melanoma cells across might also prove effective. type 1 collagen has been achieved by treating cells Inhibition of GnTV promises to be an appealing with xylosides. Xylosides consist of xylose attached therapeutic target74. However, there are no specific to a hydrophobic aglycone and resemble xylosylated inhibitors of GnTV currently under therapeutic proteoglycan core proteins. Xylosides are able to testing. Swainsonine, a competitive inhibitor of uncouple chondroitin-sulphate formation from pro- N-glycan processing in the Golgi (and which par- teoglycan formation by ‘decoying’ chondroitin-sul- tially blocks the action of GnTV), reduced invasion phate polymerization away from endogenous CSPG with low toxicity in Phase I human cancer trials75, core proteins63. The main hyaluronan receptor on and is currently in Phase II testing. The disadvan- tumour cells, CD44, seems to have a prominent role tage of these types of inhibitor is that they affect all in mediating the binding of tumours to the ECM64,65. N-glycan biosynthesis, potentially leading to many Interactions between matrix hyaluronan and CD44 side effects. Better knowledge of the crystal struc- also appear to facilitate the tight coordination of ture of GnTV might facilitate the design of specific tumour-growth signals with cytoskeletal events in inhibitors76. The isoenzymes ST8SiaII and ST8SiaIV tumour migration28. Additionally, proteoglycan forms catalyse the synthesis of polysialic acid in mamma- of CD44 that contain chondroitin- and heparan- lian cells. For example, the inhibition of ST8SiaII sulphate chains facilitate the binding of tumour to using synthetic sialic acid precursors has recently fibronectin66,67, thereby affecting the organization of been shown77, but the use of these agents in vivo has tumour matrix while anchoring tumour cells along a not been established. scaffold during invasion. Targeting proteoglycans has also been considered. Administration of clinical heparin inhibits tumour Matrix degradation and the release of sequestered heparanase activity and invasion by tumour cells growth factors. HSPGs secreted into the ECM by in vivo55,78, an effect that probably results from the the tumour have the capacity to bind and sequester ability of heparin to compete with matrix heparan large amounts of important heparin-binding growth sulphate as a substrate for tumour heparanase. factors68,69 TABLE 1. This sequestration protects Furthermore, specific alterations to the length and growth factors from denaturation or proteolysis and sulphate-substitution pattern of heparin have been thereby augments their activity. Moreover, tumour shown to improve its efficacy55,56,69. Competitive cells that secrete HSPG-digestive enzymes (glycosi- antagonists of CSPGs might also be considered as dases) are able to release and use the glycan-bound possible candidates. Cell-surface CSPGs are over- factors during matrix invasion70. Perlecan, a secreted expressed by certain tumours such as melanoma79 HSPG that makes up normal basement membranes, and breast neoplasms80. This observation has been has a particularly important role in matrix growth- exploited in the laboratory to improve tumour- factor sequestration71. Perlecan is overexpressed in specific drug delivery (using novel chondroitin- several carcinomas, including melanoma, breast and sulphate-binding cationic liposomes loaded with colon malignancies, and transfection of metastatic chemotherapeutic agents) and to reduce systemic melanoma cells with perlecan antisense oligonucle- drug toxicity. Cisplatin (a commonly used chemo- otides significantly reduces their invasive potential72. therapeutic agent) delivered in this way to tumour- Heparanase is an endoglycosidase that partially depo- bearing mice resulted in reduced tumour growth, lymerizes heparan-suphate chains, and is secreted by reduced drug toxicity, and improved survival com- invasive normal cells (for example, cytotrophoblasts, pared with control mice treated with systemic drug81. which are the specialized epithelial cells in the pla- This last example illustrates an alternative approach centa that have an important role in implantation) and to the direct targeting of glycans that promote tumour malignant cells (including carcinoma and lymphoma invasion, and demonstrates how one might exploit cells). During invasion by tumour cells, heparanases the tumour’s overexpression of certain glycans (for released by migrating tumour cells can liberate seques- example, proteoglycans or other glycoconjugates) to tered growth factors at an invading tumour front by improve tumour-specific drug targeting and reduce degrading the heparan-sulphate chains of HSPGs drug toxicity.

REVIEWS

Table 2 | Potentiation of tumour growth by HSPGs and heparin-binding growth factors HSPG alterations on Examples of tumours Mechanism of tumour growth References tumour cells potentiation by HSPG FGF2 Downregulation of SULF1 Ovarian, breast, hepatocellular, Increase heparan sulphate binding 25 (endosulphatase) pancreatic, colon, renal sites for FGF2 Increased expression of Pancreatic Facilitate FGF2 binding to FGF 24,153 glypican-1 (a GPI-anchored receptors HSPG) HB-EGF Downregulation of SULF1 Pancreatic, breast, Increase heparan sulphate binding 25 (endosulfatase) hepatocellular, colon, ovarian, sites for HB-EGF renal Increased tumour Breast, pancreatic, ovarian, Facilitate HB-EGF binding to EGF 154 glypican-1 expression bladder receptors HGF Downregulation of SULF1 Ovarian, breast, hepatocellular, Increase heparan sulphate binding 25 (endosulphatase) pancreatic, squamous cell sites for HGF Increased glypican-1 Breast, pancreatic, Facilitate HGF binding to HGF 24,153 expression receptors (MET) HGF binding to syndecan-1 Multiple myeloma Facilitate HGF binding and 155 (a membrane-bound HSPG) activation of MET VEGF Role of specific HSPGs Breast, ovarian, pancreatic, Possibly stabilize ligand–receptor 156 unknown prostate ternary complexes; increase matrix growth factor storage Increased expression of Melanoma Possible VEGF reservoir, but role as 72 perlecan (secreted HSPG) tumour-cell effector not clear PDGF Increased expression of Mesothelioma PDGF might stimulate increased 157 syndecans and perlecan by production of matrix proteoglycans PDGF Role of specific HSPGs Glioblastoma, prostate, Unknown 158 unknown gastrointestinal stromal tumours FGF2, fibroblast growth factor 2; GPI, glycosylphosphatidylinositol; HB-EGF, heparin-binding epidermal growth factor; HGF, hepatocyte growth factor; HSPG, heparan-sulphate proteoglycan; PDGF, platelet-derived growth factor; SULF1, sulfatase 1; VEGF, vascular endothelial growth factor.

Tumour angiogenesis Specific HSPGs might have especially crucial roles Solid tumours and their metastases must acquire a vas- in tumour angiogenesis. Genetic targeting of the culature through the process of tumour angiogenesis basement-membrane proteoglycan perlecan blocks in order to achieve diameters of greater than 2 mm82. carcinoma growth and angiogenesis in several in vivo Direct genetic evidence exists that supports a role for models84,85. Additionally, the release of heparanase from HSPGs in tumour angiogenesis9,71. The neovasculature endothelial cells might promote the cleavage of matrix that forms during angiogenesis consists of microvas- HSPG, mobilizing pro-angiogenic factors that support cular endothelia, which express 10–15 times the level the growing tumour neovasculature56,69,86, making of HSPG found in macrovascular endothelia83. Genetic heparanase an attractive target for anti-angiogenesis studies show that endothelial growth and migration are therapy, as discussed below. Indeed, the importance of stimulated by several pro-angiogenic factors, including mammalian heparanase in vascular remodelling was FGF2, VEGF, hepatocyte growth factor (HGF), IL-8, recently demonstrated by the overexpression of human platelet-derived growth factor (PDGF), TGFβ and heparanase in transgenic mice87. tumour-necrosis factor-α7,71, all of which bind heparan sulphate. Furthermore, the endothelial-specific deletion Therapeutic targeting of glycans in tumour angiogen- of the important biosynthetic enzyme N-deacetylase/N- esis. Therapy directed against tumour angiogenesis sulphotransferase (NDST1), which is involved in the has the advantage that it might induce less drug resist- sulphation of nascent heparan-sulphate chains, dimin- ance than that induced by standard chemotherapy ishes tumour angiogenesis in mouse models (M.M.F. directed against the tumour, because endothelial cells et al., unpublished observations). are genetically stable compared with tumour cells88. A

REVIEWS

Box 1 | Strategies to alter expression of cell-surface glycans on tumour cells Treatment of tumour cells with peracetylated disaccharides can alter the assembly of cell-surface glycans (see figure part a). The figure shows the disaccharide GlcNAcβ1,3Galβ, which is peracetylated and covalently linked to a hydrophobic aglycone (non-carbohydrate ‘R’ moiety in figure) — these modifications render the disaccharide membrane-permeable. Once inside the tumour cell, the disaccharide is deacetylated and treated as a substrate by one or more glycosyltransferases (which catalyse the transfer of a sugar from a sugar nucleotide donor to a substrate). Assembly of oligosaccharides on the disaccharide ‘primer’ decoys or diverts glycosylation away from endogenous glycoproteins. This results in decreased expression of specific glycans on the cell surface. In the example, the disaccharide inhibits the expression of sialyl Lewis X (SLeX) on cell-surface mucins, which blocks tumour dissemination in mice112. Other examples of similar decoys include monosaccharide primers such as α-d-N-acetylgalactosaminides (for example, GalNAcα-benzyl)146, which targets O-linked glycan biosynthesis, and β-d-xylosides that target heparan-sulphate and chondroitin- sulphate assembly38,147. Some analogues of sugars that contain fluorine instead of a specific hydroxyl group directly inhibit glycosylation148. Part b of the figure depicts another use of glycosides. Tumour cells express sialic acids that are normally tolerated by the immune system, and antitumour immunity can be generated in principle by introducing unnatural sialic acid precursors to tumours, so that their metabolic incorporation leads to the generation of unnatural sialic-acid epitopes. In this example, a mannosamine analogue is fed to tumour cells, where it is converted to a sialic acid derivative. This results in the formation of a cytidine monophosphate-sialic acid derivative that is then used by one or more sialyltransferases to produce a new cell-surface glycoconjugate containing the sialic acid analogue. By adjusting the R-group in the added mannosamine129,143,144, a range of unnatural chemical structures can be inserted biosynthetically on the cell surface, rendering the cell more immunoreactive or chemically reactive. In this way, it might be possible to boost complement-mediated lysis of the targeted tumour by vaccinating a host with the unnatural sialic acid epitope before administering the analogue. Owing to their high expression of sialic acids, tumour cells might be preferentially targeted by this approach149.

a Reducing adhesive potential Tumour cell expressing pro-adhesive Reduced SLeX at tumour glycan, SLeX on surface glycoproteins cell surface

Metabolic decoy (peracetylated disaccharide)

OAc OAc OAc

O O AcO AcO O O - R AcHN AcO

Altered adhesion potential

b Boosting tumour reactivity Unnatural sialic acid biosynthetic precursor (peracetylated mannosamine Tolerated sialic acid (Sia) analogue) Induced neo-antigen (Sia-R) self antigen O on cell-surface glycoconjugates AcO R OH OH HN O OO– OO– AcO OH OH AcO OAc HO HO O O O O HN HN O HO O HO Tumour cell R Engineered cell

Altered immunoreactivity Altered chemical reactivity

REVIEWS

few groups have achieved glycan-targeted inhibition of (for example, in lymph nodes) promotes the transmi- angiogenesis through introduction of modified heparin gration of leukocytes to infected and/or inflammatory fragments or of sulphated oligosaccharide heparan- sites and platelets to sites of vascular damage. Unlike sulphate mimetics that specifically inhibit hepara- their normal cell counterparts, tumour cells frequently nase89–91. Both types of agent can also interfere with overexpress SLeX (for example, on lung and colon ade- the ability of heparan sulphate to facilitate endothelial nocarcinomas) or SLeA (for example, on colon, gastric growth-factor signalling by interfering with the for- and pancreatic carcinomas) on glycoproteins or gly- mation of growth factor–heparan sulphate–receptor cosphingolipids on the surface of tumour cells3,100–102. ternary complexes. For example, sulphonic-acid poly- The expression of these ligands is inversely correlated mers abrogate the formation of the FGF2–HSPG–FGF- with patient survival103–105. receptor complexes and potently inhibit angiogenesis In contrast to the important role of O-glycans in vivo91,92. Ongoing research to model the carbohy- and glycosphingolipids during the circulatory phase drate-binding motifs of known glycan-binding growth of metastasis, there is little evidence that other gly- factors or to probe the binding potential of specific cans potentiate this process. One exception might be growth factors with libraries of enzymatically treated endothelial HSPGs, which have the ability to bind heparin fragments93 might yield promising anti-ang- several chemokines and facilitate the establishment iogenic compounds. The use of low-molecular-weight of chemokine gradients across vascular endothelial heparins should also be explored, as there is some layers159. The latter appear to potentiate extravasation evidence of improved cancer outcomes related to their and metastasis of a range of human tumour cells that ability to block angiogenesis94. commonly overexpress chemokine receptors such as In contrast to heparan sulphate, very little genetic CXCR4 or CXCR7 REFS 106,107. Furthermore, other evidence to date implicates other glycan classes in lectins such as siglecs (that might modulate binding of either pathological angiogenesis (which occurs during innate immune cells to sialic-acid-containing tumour tumour growth) or physiological angiogenesis. Various glycans) and galectins (which bind to galactose-con- cell-surface proteins such as integrins are targets for anti- taining glycans) might also have important roles in angiogenic therapy, so one might imagine that the gly- effecting circulatory tumour dissemination108–110. cosylation of these proteins would affect angiogenesis95. Treatment of endothelial cells with deoxymannojirimy- Targeting glycans in circulatory tumour dissemina- cin, a plant-derived alkaloid, prevents the synthesis of tion. The mechanism by which tumour-cell aggregates endothelial hybrid and complex-type N-linked oligosac- facilitate metastasis has not been established. The charides and inhibits the formation of capillary tubes coating of tumour cells by platelets and lymphocytes in vitro96, but whether this effect was due to alteration of might serve to stabilize tumour emboli composed of integrin activity is unknown. Further studies are needed activated platelets and leukocytes. The cellular cloak to establish the importance of various glycan classes in of platelets and leukocytes might facilitate the survival tumour endothelial growth and remodelling. (including protection from innate immunity), microvascular arrest and eventual growth of metastases in Tumour dissemination through the circulation the microvasculature of distant organs100,101,111–113. Tumour dissemination occurs via the microvascula- Agents that can interfere with selectin–carbohydrate ture, after cells invade the fine capillaries and lymphat- interactions have been considered for the treatment of ics within the tumour. After entry into the vasculature, metastatic disease. These interactions can be targeted tumour cells can absorb blood-borne factors, form through the administration of neutralizing anti-selec- large aggregates with platelets and leukocytes, and tin antibodies or small-molecule mimetics of the SLeX eventually lodge in the small vessels of distant organs. or SLeA selectin ligands114–117, but the applicability of The importance of aggregation of tumour cells with such experimental competitive therapies to clinical cellular blood components for METASTATIC SEEDING is now tumour biology and metastasis remains challenging. well established97,98. Some novel approaches of this type involve the com- Over the past two decades, an important role has petition by heparin118 or synthetic negatively charged been established for O-glycans in promoting adhesion polymers119, and the alteration of the tumour ligands METASTATIC SEEDING between tumour cells, platelets and leukocytes by way themselves through metabolic approaches112,120,121 The colonization of an organ or BOX 1 tissue by metastatic tumour of receptors known as selectins. Selectins normally . Selectins are also expressed in lymphatic ves- cells. mediate adhesion between platelets (which express sels, and might be involved in lymphatic spread of P-selectin), leukocytes (L-selectin) or endothelium carcinoma122,123. Therefore, targeting the glycan ligands LEWIS TYPE BLOOD GROUP (E- and P-selectins) that bear the glycosylated ligands. for the selectins might serve to limit tumour metastasis ANTIGENS The ligands typically consist of clustered arrangements through both blood and lymphatic routes. Obviously, A structurally similar set of fucose-containing (α1-3- of oligosaccharides that bear sialic acid, fucose and these strategies might be limited by undesirable fucosylated) oligosaccharides sulphate, presented at the ‘tips’ of predominantly effects on normal processes mediated by the selectin found on normal epithelia and O-linked glycans99. One class of carbohydrate ligand adhesion system (for example, normal leukocyte and blood cells, a few of which (for consists of the LEWIS TYPE BLOOD GROUP ANTIGENS, sialyl platelet functions). However, the dire consequences example, the sialyl Lewis X or X A sialyl Lewis Y antigens) are Lewis X (SLe ) and sialyl Lewis A (SLe ). The presence of metastatic disease warrants further development of overexpressed on the surface of of such ligands on blood cells (such as monocytes or this class of therapeutic agents, which might work best certain epithelial tumour cells. neutrophils) or on certain vascular endothelial cells in combination with conventional chemotherapy.

REVIEWS

Glycans in tumour immunity of immunogenic gangliosides, and often shed them Relatively little is known about the mechanisms by into the bloodstream3,8. Paradoxically, this can lead which tumour glycans affect or alter host immunity. to immune silencing, which probably involves both Glycosphingolipids might have an important role in the inhibition of co-stimulatory molecule synthesis certain tumours. Some classes of ganglioside produced as well as the arrest of dendritic-cell maturation, by certain tumours can lead to immune silencing resulting in the inability of dendritic cells to generate though mechanisms discussed below. It is also pos- effective antitumour T-cell immune responses130,131. sible that certain glycan patterns, such as O-GlcNAc- This observation indicates that inhibitors of the gan- modified peptides presented on the tumour surface124, glioside assembly process might prove effective in could stimulate cytotoxic T-cell-mediated responses these types of tumour132,133. by the host. Regardless of the mechanisms by which Although glycosphingolipids are relatively poor tumour glycosylation influences host immunity, our immunogens, certain glycosphingolipids might be increasing knowledge of tumour glycan expression manipulated to generate both passive immunity (by the means that it is now possible to exploit tumour glycans infusion of antibodies to glycans) and active immu- as a means to augment antitumour immunity. nity (by eliciting a host response to a glycan) against

tumours. Immunization with purified GD2 or GM2 Augmenting immunity against tumour mucin gly- gangliosides (FIG. 1) has been carried out, initially with cans. A tumour’s ability to generate or overexpress some success in animal models and preliminary human unique mucins is now being harnessed in an attempt melanoma trials. This approach is showing promise to improve or augment the immune system’s ability to in several current clinical trials against melanoma 134 recognize and destroy tumours. Methods to destroy (using GD2–KLH ), neuroblastoma (using anti-GD2 35 tumour cells using monoclonal antibodies or vaccines small immunoproteins ), breast (using NeuGcGM3 that target the tumour mucins MUC1 or MUC16, or proteoliposomes36 that contain the N-glycolyl form of the mucin-associated O-linked glycan STn, have been sialic acid) and prostate carcinomas (using a conjugate introduced as strategies to inhibit tumour prolifera- between KLH and Globo-H hexasaccharide135, a glycan tion or invasion. Indeed, carcinoma mucins aberrantly expressed on human prostate and breast cancer gly- glycosylated with either Tn (αGalNAc) or STn often colipids). Finally, liposomal drug-delivery techniques decorate the tumour surface, creating clustered sites can take advantage of ganglioside overexpression by for antibody attachment, thereby improving their certain tumours. For instance, MYB antisense oli- activity as tumour immunogens. Because glycans godeoxynucleotides (for inhibiting the synthesis of are weak immunogens, vaccines of this type are typi- MYB) have been delivered to human neuroblastomas cally prepared by conjugating the glycan to a carrier by encapsulation in cationic liposomes that have been protein (for example, keyhole limpet haemocyanin covalently coupled to monoclonal antibodies against 136 (KLH)), and then injecting the compound into the the GD2 ganglioside . patient together with an adjuvant that boosts T-cell responses. Vaccines that contain these epitopes are Measurement of serum tumour glycans currently being used in breast cancer patients to aug- Serum measurement of certain glycans on the sur- ment antitumour immune responses, and a positive face of tumour cells is currently used to facilitate survival effect has been noted in patients treated diagnosis, track tumour recurrence or tumour bur- with the STn–KLH conjugate ‘Theratope’ (currently den or provide a surrogate measure for therapeutic in Phase III clinical trials125). Ongoing clinical trials response. These glycans might be regarded as part are examining the use of this agent to treat ovar- of a larger array of ‘metastatic codes’ that a tumour’s ian and colorectal cancer patients126,127. In addition, glycan profile (or ‘glycotype’) might represent. For peptide mimetics of tumour carbohydrates, such as example, the serological markers CA125, CA19-9 SLeX, SLeA or SLY, have also been shown to stimulate and CA15-3 are mucin glycoconjugates that are tumour immunity, although these studies have only commonly overexpressed by ovarian, pancreatic been carried out in animal models128. and breast adenocarcinomas, respectively, and In another promising strategy, antitumour immu- their serum levels correlate with tumour burden nity has been generated by making unique alterations and prognosis137–139. In ovarian cancer, the tumour to tumour-cell-surface sialic acids (which are normally antigen CA125 is associated with a large mucin- tolerated by the immune system). This might be accom- like glycoprotein called MUC16 that interacts with plished by introducing unnatural sialic acid precursors galactose-binding lectins (galectins) that are secreted to tumours BOX 1, whereby metabolic incorporation into the tumour matrix. One of the ways in which leads to the generation of unnatural sialic acid epitopes, galectins might regulate tumour progression is forming the basis for the induction of novel antitumour through their ability to facilitate cell adhesion and immune responses129. promote the binding of tumour cells to laminin and fibronectin110. Their overexpression in ovarian Augmenting immunity against tumour glycosphin- cancer might facilitate tumour–matrix interactions golipids. Some mammary carcinomas, neurob- during invasion of ovarian cancer cells140. CA19-9 lastomas, sarcomas, melanomas, small-cell lung is the epitope that interacts with SLeA on pancreatic carcinomas and lymphomas express very high levels carcinoma mucins141, and its expression facilitates

REVIEWS

Table 3 | Current clinical targets of glycans in cancer Therapeutic agent Mechanism of action Clinical status References Targeting N-glycans GD0039 (Swainsonine) Blocks Golgi Phase II (renal cancer) 11 α-mannosidase Targeting O-glycans Theratope Conjugated vaccine targets Phase III (breast cancer) 125 the STn mucin epitope Disaccharide primers Alter SLeX (selectin ligand Preclinical 112 (peracetylated facilitating haematogenous GlcNAcβ1,3Gal-naphthalen- metastasis) emethanol) Heparin Inhibits P-selectin and On the market 118 L-selectin binding to SLeX ligand Targeting heparan-sulphate proteoglycans Heparin Inhibits heparanase and On the market 78 interactions between growth factor and heparan sulphate PI88 Inhibits FGF release via Phase I and II (melanoma) 11 heparanase inhibition Phase II (multiple myeloma) Targeting gangliosides

GMK vaccine Conjugated GM2 vaccine Phase III (melanoma) 11

NeuGc GM3/VSSP NeuGcGM3 proteoliposomes Phase I (breast) 36

Anti-GD2 SIP Anti-GD2 small Preclinical 35 immunoproteins

GD2 lactone–KLH Conjugated GD2 vaccine Phase I (melanoma) 134 Targeting galectin-binding glycans GCS-100 (citrus pectin Blocks galectin-3 binding; Phase II (pancreatic and colon 150 derivative) inhibits tumour apoptosis; cancers) facilitates tumour aggregation

selectin-mediated adhesion during haematogenous therapeutic strategies can target more than one class metastasis. In breast cancer, the tumour antigen CA15- of glycan–protein interaction. For example, heparin 3 is expressed on MUC1, which is aberrantly expressed will block selectin–SLeX interactions and also block in more than 90% of breast carcinomas and appears invasion and angiogenesis by inhibiting heparanases to promote invasion142. Therefore, the glycans CA125, and growth-factor interactions with their receptors. CA19-9 and CA15-3 are examples of molecules that Alternatively, metabolic decoys might alter tumour-cell not only serve as tumour markers for diagnosis, but interactions that are mediated by selectins and affect the also appear to serve important pathophysiological immunogenicity of carcinoma mucins BOX 1. Altering roles in cancer progression. For now, although these N-glycosylation of tumour glycoproteins might confer and other glycan diagnostic markers are used clinically additional biological effects by affecting the folding as sensitive markers for recurrence of disease following of several crucial membrane proteins145. Finally, one initial treatment, they might also be used to facilitate should also consider the potential synergistic effects of the timing of glycan-based therapies in future cancer altering tumour glycans in combination with existing treatment programmes. radiotherapeutic and chemotherapeutic therapies or anti-angiogenic therapies. An added advantage of this Implications and future directions approach is that it might also overcome the problems A few important patterns emerge from the above dis- of tumour heterogeneity and drug resistance, which are cussion. First, a given glycan might act at different stages concerns that are inherent in any specific targeting of of tumour progression, so targeting that glycan might tumour cells. have broad effects (FIG. 3). Second, at any given stage of Ultimately, the choice and timing of any future gly- progression, a specific glycan-targeting strategy might can-based therapy against cancer should be guided by alter several glycan-dependent interactions. For exam- both serological assays for glycan markers as well as ple, altering selectin interactions would affect platelet, novel biopsy information. Refinement of technologies lymphocyte and endothelial adhesion. Third, some to analyse clinical biopsy material (yielding a glycotype

REVIEWS

for the tumour) in addition to serum glycans might pharmacological approaches might then result assist with the rational choice and/or early application from profiling drug design according to genetically of glycan-based immunomodulators. Glycan analysis validated targets. of serum or biopsy material could also facilitate the As we continue to uncover clues about pathogenic preoperative administration of tumour-glycan inhibi- mechanisms in tumour progression that glycans tors to reduce the risk of operation-associated dissem- facilitate, or in some cases, govern, the integration ination of microscopic tumour deposits. The recent of glycan-targeted therapy with existing cancer application of microarray technology might provide treatment protocols might have significant impact insights into the array of enzymes and glycoproteins on disease outcomes. Progress so far on targeting associated with the grade and stage of a given tumour. complex carbohydrates in cancer (summarized in This, in turn, might indicate appropriate tumour and TABLE 3) might represent the ‘tip of an iceberg’ of patient-specific panels of inhibitors to test. Valuable therapeutic potential that awaits future discovery.

1. Wells, L., Vosseller, K. & Hart, G. W. Glycosylation of 20. Nagy, P. et al. Lipid rafts and the local density of ErbB 35. Bestagno, M., Occhino, M., Corrias, M. V., Burrone, O. & nucleocytoplasmic proteins: signal transduction and proteins influence the biological role of homo- and Pistoia, V. Recombinant antibodies in the immunotherapy O-GlcNAc. Science 291, 2376–2378 (2001). heteroassociations of ErbB2. J. Cell Sci. 115, of neuroblastoma: perspectives of new developments. 2. Marth, J. Glycosylation changes in ontogeny and cell 4251–4262 (2002). Cancer Lett. 197, 193–198 (2003). activation. in Essentials of Glycobiology (eds Varki, A. et al.) 21. Esko, J. D. in Essentials of Glycobiology (eds Varki, A. et al.) 36. Carr, A. et al. Immunotherapy of advanced breast cancer 515–536 (Cold Spring Harbor Laboratory Press, New 145–159 (Cold Spring Harbor Laboratory Press, New with a heterophilic ganglioside (NeuGcGM3) cancer York, 1999). York, 1999). vaccine. J. Clin. Oncol. 21, 1015–1021 (2003). 3. Varki, A. in Essentials of Glycobiology (eds Varki, A. et al. ) 22. Esko, J. D. & Lindahl, U. Molecular diversity of heparan 37. Thomas, T. & Thomas, T. J. Polyamine metabolism and 537–549 (Cold Spring Harbor Laboratory Press, New sulfate. J. Clin. Invest. 108, 169–173 (2001). cancer. J. Cell Mol. Med. 7, 113–126 (2003). York, 1999). 23. Bernfield, M. et al. Functions of cell surface heparan 38. Belting, M. et al. Tumor attenuation by combined heparan 4. Feizi, T. Carbohydrate antigens in human cancer. Cancer sulfate proteoglycans. Annu. Rev. Biochem. 68, 729–777 sulfate and polyamine depletion. Proc. Natl Acad. Sci. Surv. 4, 245–269 (1985). (1999). USA 99, 371–376 (2002). 5. Raedler, A. & Schreiber, S. Analysis of differentiation and 24. Kleeff, J. et al. The cell-surface heparan sulfate 39. Zeng, C., Toole, B. P., Kinney, S. D., Kuo, J. W. & transformation of cells by lectins. Crit. Rev. Clin. Lab. Sci. proteoglycan glypican-1 regulates growth factor action in Stamenkovic, I. Inhibition of tumor growth in vivo by 26, 153–193 (1988). pancreatic carcinoma cells and is overexpressed in human hyaluronan oligomers. Int. J. Cancer 77, 396–401 (1998). 6. Weis, W. I. & Drickamer, K. Structural basis of lectin- pancreatic cancer. J. Clin. Invest. 102, 1662–1673 (1998). Disruption of the interaction between host tissue carbohydrate recognition. Annu. Rev. Biochem. 65, 25. Lai, J. et al. Loss of HSulf-1 up-regulates heparin-binding stromal hyaluronan and the hyaluronan receptor 441–473 (1996). growth factor signaling in cancer. J. Biol. Chem. 278, CD44 on tumour cells can be achieved by treatment Describes the structural basis for sugar recognition 23107–23117 (2003). with hyaluronan oligomers. This study illustrates and adhesion by lectins expressed in bacteria, Demonstration of the importance of HSPGs in how infusion with such oligomers can block growth animals and plants. tumour proliferation. The cell-surface-associated of experimental melanomas in vivo. 7. Bogenrieder, T. & Herlyn, M. Axis of evil: molecular enzyme SULF1 diminishes HSPG sulphation. 40. Dennis, J. W., Pawling, J., Cheung, P., Partridge, E. & mechanisms of cancer metastasis. Oncogene 22, Downregulation of SULF1 by tumours increases Demetriou, M. UDP-N-acetylglucosamine:α-6-D- 6524–6536 (2003). growth-factor signalling and resistance to mannoside β1, 6 N-acetylglucosaminyltransferase V 8. Hakomori, S. Tumor malignancy defined by aberrant apoptosis, representing a novel mechanism by (Mgat5) deficient mice. Biochim. Biophys. Acta. 1573, glycosylation and sphingo(glyco)lipid metabolism. Cancer which cancer cells can enhance growth. 414–422 (2002). Res. 56, 5309–5318 (1996). 26. DeBaun, M. R., Ess, J. & Saunders, S. Simpson Golabi 41. Yoshimura, M., Ihara, Y., Matsuzawa, Y. & Taniguchi, N. 9. Iozzo, R. V. & San Antonio, J. D. Heparan sulfate Behmel syndrome: progress toward understanding the Aberrant glycosylation of E-cadherin enhances cell–cell proteoglycans: heavy hitters in the angiogenesis arena. molecular basis for overgrowth, malformation, and cancer binding to suppress metastasis. J. Biol. Chem. 271, J. Clin. Invest. 108, 349–355 (2001). predisposition. Mol. Genet. Metab. 72, 279–286 (2001). 13811–13815 (1996). 10. Liu, D., Shriver, Z., Qi, Y., Venkataraman, G. & 27. Filmus, J. Glypicans in growth control and cancer. 42. Hirohashi, S. & Kanai, Y. Cell adhesion system and Sasisekharan, R. Dynamic regulation of tumor growth and Glycobiology 11, 19R–23R (2001). human cancer morphogenesis. Cancer Sci. 94, 575–581 metastasis by heparan sulfate glycosaminoglycans. 28. Turley, E. A., Noble, P. W. & Bourguignon, L. Y. Signaling (2003). Semin. Thromb. Hemost. 28, 67–78 (2002). properties of hyaluronan receptors. J. Biol. Chem. 277, 43. Seidenfaden, R., Krauter, A., Schertzinger, F., Gerardy- 11. Maeder, T. Sweet medicines. Sci. Am. 287, 40–47 (2002). 4589–4592 (2002). Schahn, R. & Hildebrandt, H. Polysialic acid directs tumor 12. Hollingsworth, M. A. & Swanson, B. J. Mucins in cancer: 29. Simpson, M. A., Wilson, C. M. & McCarthy, J. B. Inhibition cell growth by controlling heterophilic neural cell adhesion protection and control of the cell surface. Nature Rev. of prostate tumor cell hyaluronan synthesis impairs molecule interactions. Mol. Cell. Biol. 23, 5908–5918 (2003). Cancer 4, 45–60 (2004). subcutaneous growth and vascularization in This study examines the role of polysialic acid on 13. Toole, B. P. Hyaluronan: from extracellular glue to immunocompromised mice. Am. J. Pathol. 161, 849–857 neural cell adhesion molecule (NCAM) in pericellular cue. Nature Rev. Cancer 4, 528–539 (2004). (2002). neuroblastoma cells, and demonstrates the ability of 14. Felsher, D. W. Reversibility of oncogene-induced cancer. 30. Misra, S., Ghatak, S., Zoltan-Jones, A. & Toole, B. P. polysialic acid to control tumour-cell growth and Curr. Opin. Genet. Dev. 14, 37–42 (2004). Regulation of multidrug resistance in cancer cells by differentiation by interfering with NCAM signalling at 15. Marth, J. in Essentials of Glycobiology (eds Varki, A. et al. ) 85– hyaluronan. J. Biol. Chem. 278, 25285–25288 (2003). cell–cell contacts. 100 (Cold Spring Harbor Laboratory Press, New York, 1999). 31. Chou, T. Y. & Hart, G. W. O-linked N-acetylglucosamine 44. Angata, K. & Fukuda, M. Polysialyltransferases: major 16. Girnita, L. et al. Inhibition of N-linked glycosylation down- and cancer: messages from the glycosylation of c-Myc. players in polysialic acid synthesis on the neural cell regulates insulin-like growth factor-1 receptor at the cell Adv. Exp. Med. Biol. 491, 413–418 (2001). adhesion molecule. Biochimie 85, 195–206 (2003). surface and kills Ewing’s sarcoma cells: therapeutic Discusses the importance of modification by 45. Lin, S., Kemmner, W., Grigull, S. & Schlag, P. M. Cell implications. Anticancer Drug Des. 15, 67–72 (2000). O-GlcNAc of key cellular oncogenes or tumour- surface α 2, 6 sialylation affects adhesion of breast 17. Bharathan, S., Moriarty, J., Moody, C. E. & Sherblom, A. P. suppressor proteins. O-GlcNAc transferase carcinoma cells. Exp. Cell Res. 276, 101–110 (2002). Effect of tunicamycin on sialomucin and natural killer glycosylates tumour-suppressor proteins at amino- 46. Ju, T. & Cummings, R. D. A unique molecular chaperone susceptibility of rat mammary tumor ascites cells. Cancer acid residues that either promote oncogene activity Cosmc required for activity of the mammalian core 1β Res. 50, 5250–5256 (1990). or inhibit tumour-suppressor functions, highlighting 3-galactosyltransferase. Proc. Natl Acad. Sci. USA 99, 18. Komatsu, M., Jepson, S., Arango, M. E., Carothers the importance of O-GlcNAc in tumorigenesis and 16613–16618 (2002). Carraway, C. A. & Carraway, K. L. Muc4/sialomucin tumour proliferation. 47. Julien, S. et al. Expression of sialyl-Tn antigen in breast complex, an intramembrane modulator of ErbB2/HER2/ 32. Noujaim, A. A., Schultes, B. C., Baum, R. P. & cancer cells transfected with the human CMP-Neu5Ac: Neu, potentiates primary tumor growth and suppresses Madiyalakan, R. Induction of CA125-specific B and T cell GalNAc α2, 6-sialyltransferase (ST6GalNac I) cDNA. apoptosis in a xenotransplanted tumor. Oncogene 20, responses in patients injected with MAb-B43.13 — Glycoconj. J. 18, 883–893 (2001). 461–470 (2001). evidence for antibody-mediated antigen-processing and 48. Guo, H. B., Lee, I., Kamar, M., Akiyama, S. K. & Pierce, M. Demonstrates how genetic upregulation of MUC4 in presentation of CA125 in vivo. Cancer Biother. Aberrant N-glycosylation of β1 integrin causes reduced tumour A375 melanoma cells inhibits apoptosis of Radiopharm. 16, 187–203 (2001). α5β1 integrin clustering and stimulates cell migration. tumour cells and promotes tumour growth in vivo. 33. Musselli, C. et al. Reevaluation of the cellular immune Cancer Res. 62, 6837–6845 (2002). MUC4 modulates phosphorylation of the receptor response in breast cancer patients vaccinated with MUC1. 49. Wozniak, M. A., Modzelewska, K., Kwong, L. & Keely, P. J. tyrosine kinase in the presence and absence of the Int. J. Cancer 97, 660–667 (2002). Focal adhesion regulation of cell behavior. Biochim. major EGF receptor ligand, heregulin. 34. Ramanathan, R. K. et al. Phase I study of a MUC1 vaccine Biophys. Acta 1692, 103–119 (2004). 19. Hakomori, S. Bifunctional role of glycosphingolipids. composed of different doses of MUC1 peptide with 50. Tumova, S., Woods, A. & Couchman, J. R. Heparan Modulators for transmembrane signaling and mediators SB-AS2 adjuvant in resected and locally advanced sulfate proteoglycans on the cell surface: versatile for cellular interactions. J. Biol. Chem. 265, 18713–18716 pancreatic cancer. Cancer Immunol. Immunother. 54, coordinators of cellular functions. Int. J. Biochem. Cell Biol. (1990). 254–264 (2005). 32, 269–288 (2000).

REVIEWS

51. Woods, A., McCarthy, J. B., Furcht, L. T. & Couchman, J. R. examining mammary tumour growth and metastasis 100. Kim, Y. J., Borsig, L., Varki, N. M. & Varki, A. P-selectin A synthetic peptide from the COOH-terminal heparin- in mice bearing a deletion in the enzyme. deficiency attenuates tumor growth and metastasis. Proc. binding domain of fibronectin promotes focal adhesion 75. Goss, P. E., Baptiste, J., Fernandes, B., Baker, M. & Natl Acad. Sci. USA 95, 9325–9330 (1998). formation. Mol. Biol. Cell 4, 605–613 (1993). Dennis, J. W. A phase I study of swainsonine in patients Describes the observation that P-selectin is 52. Saoncella, S. et al. Syndecan-4 signals cooperatively with with advanced malignancies. Cancer Res. 54, 1450–1457 important for the formation of tumour-platelet integrins in a Rho-dependent manner in the assembly of (1994). emboli during metastasis, and facilitates focal adhesions and actin stress fibers. Proc. Natl Acad. 76. Schachter, H. The joys of HexNAc. The synthesis and haematogenous dissemination of mucin-producing Sci. USA 96, 2805–2810 (1999). function of N- and O-glycan branches. Glycoconj. J. 17, carcinoma cells. 53. Kusano, Y., Yoshitomi, Y., Munesue, S., Okayama, M. & 465–483 (2000). 101. Borsig, L., Wong, R., Hynes, R. O., Varki, N. M. & Varki, A. Oguri, K. Cooperation of syndecan-2 and syndecan-4 77. Horstkorte, R. et al. Selective inhibition of Synergistic effects of L- and P-selectin in facilitating tumor among cell surface heparan sulfate proteoglycans in the polysialyltransferase ST8SiaII by unnatural sialic acids. metastasis can involve non-mucin ligands and implicate actin cytoskeletal organization of Lewis lung carcinoma Exp. Cell Res. 298, 268–274 (2004). leukocytes as enhancers of metastasis. Proc. Natl Acad. cells. J. Biochem. (Tokyo) 135, 129–137 (2004). 78. Kakkar, A. K. An expanding role for antithrombotic therapy Sci. USA 99, 2193–2198 (2002). 54. Culp, L. A. et al. Heparan sulfate proteoglycans of Ras- in cancer patients. Cancer Treat. Rev. 29 (Suppl. 2), 23–26 102. Varki, A. Selectin ligands: will the real ones please stand transformed 3T3 or neuroblastoma cells. Differing (2003). up? J. Clin. Invest. 100, S31– S35 (1997). functions in adhesion on fibronectin. Ann. NY Acad. Sci. 79. Bumol, T. F. & Reisfeld, R. A. Unique glycoprotein- 103. Nakamori, S. et al. Involvement of carbohydrate antigen 556, 194–216 (1989). proteoglycan complex defined by monoclonal antibody on sialyl Lewis(x) in colorectal cancer metastasis. Dis. Colon 55. Vlodavsky, I. et al. Inhibition of tumor metastasis by human melanoma cells. Proc. Natl Acad. Sci. USA 79, Rectum. 40, 420–431 (1997). heparanase inhibiting species of heparin. Invasion 1245–1249 (1982). 104. Baldus, S. E. et al. Histopathological subtypes and Metastasis 14, 290–302 (1994). 80. Olsen, E. B., Trier, K., Eldov, K. & Ammitzboll, T. prognosis of gastric cancer are correlated with the 56. Vlodavsky, I. & Friedmann, Y. Molecular properties and Glycosaminoglycans in human breast cancer. Acta Obstet. expression of mucin-associated sialylated antigens: involvement of heparanase in cancer metastasis and Gynecol. Scand. 67, 539–542 (1988). Sialosyl-Lewis(a), Sialosyl-Lewis(x) and sialosyl-Tn. Tumour angiogenesis. J. Clin. Invest. 108, 341–347 (2001). 81. Lee, C. M. et al. Novel chondroitin sulfate-binding cationic Biol. 19, 445–453 (1998). Reviews the many roles of heparanase in the tumour liposomes loaded with cisplatin efficiently suppress the 105. Nakamori, S. et al. Molecular mechanism involved in environment, and its importance in tumour local growth and liver metastasis of tumor cells in vivo. increased expression of sialyl Lewis antigens in ductal progression. Cancer Res. 62, 4282–4288 (2002). carcinoma of the pancreas. J. Exp. Clin. Cancer Res. 18, 57. Kato, M., Saunders, S., Nguyen, H. & Bernfield, M. Loss of 82. Folkman, J. in The Molecular Basis of Cancer (eds 425–432 (1999). cell surface syndecan-1 causes epithelia to transform into Mendelsohn, J., Howley P. M., Israel M. A. & Liotta L. A.) 106. Muller, A. et al. Involvement of chemokine receptors in anchorage-independent mesenchyme-like cells. Mol. Biol. 206–232 (W. B. Saunders, Philadelphia, 1995). breast cancer metastasis. Nature 410, 50–56 (2001). Cell 6, 559–576 (1995). 83. Marcum, J. A. & Rosenberg, R. D. Heparinlike molecules 107. Almofti, A. et al. The clinicopathological significance of the 58. Beauvais, D. M. & Rapraeger, A. C. Syndecans in tumor with anticoagulant activity are synthesized by cultured expression of CXCR4 protein in oral squamous cell cell adhesion and signaling. Reprod. Biol. Endocrinol. 2, 3 endothelial cells. Biochem. Biophys. Res. Commun. 126, carcinoma. Int. J. Oncol. 25, 65–71 (2004). (2004). 365–372 (1985). 108. Crocker, P. R. Siglecs: sialic-acid-binding immunoglobulin- 59. Beauvais, D. M., Burbach, B. J. & Rapraeger, A. C. The 84. Zhou, Z. et al. Impaired angiogenesis, delayed wound like lectins in cell-cell interactions and signalling. Curr. α β syndecan-1 ectodomain regulates v 3 integrin activity in healing and retarded tumor growth in perlecan heparan Opin. Struct. Biol. 12, 609–615 (2002). human mammary carcinoma cells. J. Cell Biol. 167, 171– sulfate-deficient mice. Cancer Res. 64, 4699–4702 109. Brinkman-Van der Linden, E. C. & Varki, A. New aspects of 181 (2004). (2004). siglec binding specificities, including the significance of 60. Sanderson, R. D. Heparan sulfate proteoglycans in 85. Sharma, B. et al. Antisense targeting of perlecan blocks fucosylation and of the sialyl-Tn epitope. Sialic acid- invasion and metastasis. Semin. Cell Dev. Biol. 12, 89–98 tumor growth and angiogenesis in vivo. J. Clin. Invest. binding immunoglobulin superfamily lectins. J. Biol. Chem. (2001). 102, 1599–1608 (1998). 275, 8625–8632 (2000). 61. Iida, J., Meijne, A. M., Knutson, J. R., Furcht, L. T. & 86. Folkman, J. & Shing, Y. Control of angiogenesis by heparin 110. Liu, F. T. & Rabinovich, G. A. Galectins as modulators of McCarthy, J. B. Cell surface chondroitin sulfate and other sulfated polysaccharides. Adv. Exp. Med. Biol. tumour progression. Nature Rev. Cancer 5, 29–41 proteoglycans in tumor cell adhesion, motility and invasion. 313, 355–364 (1992). (2005). Semin. Cancer Biol. 7, 155–162 (1996). 87. Zcharia, E. et al. Transgenic expression of mammalian 111. Pearlstein, E., Salk, P. L., Yogeeswaran, G. & Karpatkin, S. 62. Nutt, C. L., Zerillo, C. A., Kelly, G. M. & Hockfield, S. Brain heparanase uncovers physiological functions of heparan Correlation between spontaneous metastatic potential, enriched hyaluronan binding (BEHAB)/brevican increases sulfate in tissue morphogenesis, vascularization, and platelet-aggregating activity of cell surface extracts, and aggressiveness of CNS-1 gliomas in Lewis rats. Cancer feeding behavior. FASEB J. 18, 252–263 (2004). cell surface sialylation in 10 metastatic-variant derivatives Res. 61, 7056–7059 (2001). 88. Boehm, T., Folkman, J., Browder, T. & O’Reilly, M. S. of a rat renal sarcoma cell line. Proc. Natl Acad. Sci. USA 63. Faassen, A. E. et al. A cell surface chondroitin sulfate Antiangiogenic therapy of experimental cancer does not 77, 4336–4339 (1980). proteoglycan, immunologically related to CD44, is involved induce acquired drug resistance. Nature 390, 404–407 112. Fuster, M. M., Brown, J. R., Wang, L. & Esko, J. D. A in type I collagen-mediated melanoma cell motility and (1997). disaccharide precursor of sialyl Lewis X inhibits metastatic invasion. J. Cell Biol. 116, 521–531 (1992). 89. Lapierre, F. et al. Chemical modifications of heparin that potential of tumor cells. Cancer Res. 63, 2775–2781 64. Gunthert, U. et al. A new variant of glycoprotein CD44 diminish its anticoagulant but preserve its heparanase- (2003). confers metastatic potential to rat carcinoma cells. Cell 65, inhibitory, angiostatic, anti-tumor and anti-metastatic 113. Al-Mehdi, A. B. et al. Intravascular origin of metastasis 13–24 (1991). properties. Glycobiology 6, 355–366 (1996). from the proliferation of endothelium-attached tumor cells: 65. Aruffo, A., Stamenkovic, I., Melnick, M., Underhill, C. B. & 90. Parish, C. R., Freeman, C., Brown, K. J., Francis, D. J. a new model for metastasis. Nature Med. 6, 100–102 Seed, B. CD44 is the principal cell surface receptor for & Cowden, W. B. Identification of sulfated (2000). hyaluronate. Cell 61, 1303–1313 (1990). oligosaccharide-based inhibitors of tumor growth and 114. Winn, R. K., Liggitt, D., Vedder, N. B., Paulson, J. C. & 66. Jalkanen, S. & Jalkanen, M. Lymphocyte CD44 binds the metastasis using novel in vitro assays for angiogenesis Harlan, J. M. Anti-P-selectin monoclonal antibody COOH-terminal heparin-binding domain of fibronectin. and heparanase activity. Cancer Res. 59, 3433–3441 attenuates reperfusion injury to the rabbit ear. J. Clin. J. Cell Biol. 116, 817–825 (1992). (1999). Invest. 92, 2042–2047 (1993). 67. Wolff, E. A. et al. Generation of artificial proteoglycans 91. Presta, M. et al. Heparin derivatives as angiogenesis 115. Fukuda, M. N. et al. A peptide mimic of E-selectin ligand containing glycosaminoglycan-modified CD44. inhibitors. Curr. Pharm. Des. 9, 553–566 (2003). inhibits sialyl Lewis X-dependent lung colonization of tumor Demonstration of the interaction between rantes and 92. Liekens, S. et al. Modulation of fibroblast growth factor-2 cells. Cancer Res. 60, 450–456 (2000). chondroitin sulfate. J. Biol. Chem. 274, 2518–2524 receptor binding, signaling, and mitogenic activity by 116. Kaila, N. & Thomas, B. E. t. Design and synthesis of sialyl (1999). heparin-mimicking polysulfonated compounds. Mol. Lewis(x) mimics as E- and P-selectin inhibitors. Med. Res. 68. Vlodavsky, I. et al. Mammalian heparanase: gene cloning, Pharmacol. 56, 204–213 (1999). Rev. 22, 566–601 (2002). expression and function in tumor progression and 93. Jemth, P. et al. Biosynthetic oligosaccharide libraries for 117. Leppanen, A. et al. A novel glycosulfopeptide binds to metastasis. Nature Med. 5, 793–802 (1999). identification of protein-binding heparan sulfate motifs. P-selectin and inhibits leukocyte adhesion to P-selectin. 69. Goldshmidt, O. et al. Cell surface expression and secretion Exploring the structural diversity by screening for fibroblast J. Biol. Chem. 274, 24838–24848 (1999). of heparanase markedly promote tumor angiogenesis and growth factor (FGF)1 and FGF2 binding. J. Biol. Chem. 118. Varki, N. M. & Varki, A. Heparin inhibition of selectin- metastasis. Proc. Natl Acad. Sci. USA 99, 10031–10036 277, 30567–30573 (2002). mediated interactions during the haematogenous phase of (2002). 94. Pisano, C. et al. Undersulfated, low-molecular-weight carcinoma metastasis: rationale for clinical studies in 70. Vlodavsky, I. et al. Molecular properties and involvement of glycol-split heparin as an antiangiogenic VEGF antagonist. humans. Semin. Thromb. Hemost. 28, 53–66 (2002). heparanase in cancer progression and normal Glycobiology 15, 1C–6C (2005). 119. Orner, B. P., Derda, R., Lewis, R. L., Thomson, J. A. & development. Biochimie 83, 831–839 (2001). 95. Hwang, R. & Varner, J. The role of integrins in tumor Kiessling, L. L. Arrays for the combinatorial exploration of 71. Jiang, X. & Couchman, J. R. Perlecan and tumor angiogenesis. Hematol. Oncol. Clin. North Am. 18, 991– cell adhesion. J. Am. Chem. Soc. 126, 10808–10809 angiogenesis. J. Histochem. Cytochem. 51, 1393–1410 1006, vii (2004). (2004). (2003). 96. Nguyen, M., Folkman, J. & Bischoff, J. 120. Brown, J. R., Fuster, M. M., Whisenant, T. & Esko, J. D. 72. Adatia, R. et al. Suppression of invasive behavior of 1-Deoxymannojirimycin inhibits capillary tube formation Expression patterns of α 2, 3-sialyltransferases and α 1, 3- melanoma cells by stable expression of anti-sense in vitro. Analysis of N-linked oligosaccharides in bovine fucosyltransferases determine the mode of sialyl Lewis X perlecan cDNA. Ann. Oncol. 8, 1257–1261 (1997). capillary endothelial cells. J. Biol. Chem. 267, 26157– inhibition by disaccharide decoys. J. Biol. Chem. 278, 73. Zcharia, E. et al. Molecular properties and involvement of 26165 (1992). 23352–23359 (2003). heparanase in cancer progression and mammary gland 97. Krause, T. & Turner, G. A. Are selectins involved in 121. Sarkar, A. K., Rostand, K. S., Jain, R. K., Matta, K. L. & morphogenesis. J. Mammary Gland Biol. Neoplasia 6, metastasis? Clin. Exp. Metastasis 17, 183–192 (1999). Esko, J. D. Fucosylation of disaccharide precursors of 311–322 (2001). 98. Borsig, L. et al. Pictures in molecular medicine: three- sialyl LewisX inhibit selectin-mediated cell adhesion. 74. Granovsky, M. et al. Suppression of tumor growth and dimensional visualization of intravascular tumor cells in J. Biol. Chem. 272, 25608–25616 (1997). metastasis in Mgat5-deficient mice. Nature Med. 6, mice. Trends Mol. Med. 7, 377 (2001). Demonstrates the ability of disaccharide decoys of 306–312 (2000). 99. Cummings, R. D. Selectins. in Essentials of Glycobiology glycosylation to alter expression of the glycan Demonstrates the importance of N-acetylglucosami- (eds Varki, A. et al.) 391–415 (Cold Spring Harbor selectin ligand SLeX on the surface of lymphoma nyltransferase V (GnTV) in tumour progression by Laboratory Press, New York, 1999). cells. This is a novel method to reduce the levels of

REVIEWS

selectin ligands on a range of carcinomas, thereby 135. Slovin, S. F. et al. Carbohydrate vaccines in cancer: 153. Kleeff, J. et al. Stable transfection of a glypican-1 altering the metastatic potential of tumour cells. immunogenicity of a fully synthetic globo H antisense construct decreases tumorigenicity in PANC-1 122. Qian, F., Hanahan, D. & Weissman, I. L. L-selectin can hexasaccharide conjugate in man. Proc. Natl Acad. Sci. pancreatic carcinoma cells. Pancreas 19, 281–288 (1999). facilitate metastasis to lymph nodes in a transgenic mouse USA 96, 5710–5715 (1999). 154. Miyamoto, S. et al. Heparin-binding EGF-like growth factor model of carcinogenesis. Proc. Natl Acad. Sci. USA 98, 136. Pagnan, G. et al. Delivery of c-myb antisense is a promising target for ovarian cancer therapy. Cancer 3976–3981 (2001). oligodeoxynucleotides to human neuroblastoma cells via Res. 64, 5720–5727 (2004). 123. Renkonen, J., Paavonen, T. & Renkonen, R. Endothelial disialoganglioside GD(2)-targeted immunoliposomes: 155. Derksen, P. W. et al. Cell surface proteoglycan syndecan-1 and epithelial expression of sialyl Lewis(x) and sialyl antitumor effects. J. Natl Cancer Inst. 92, 253–261 mediates hepatocyte growth factor binding and promotes Lewis(a) in lesions of breast carcinoma. Int. J. Cancer 74, (2000). Met signaling in multiple myeloma. Blood 99, 1405–1410 296–300 (1997). 137. Nakagoe, T. et al. Difference in prognostic value between (2002). 124. Glithero, A. et al. Crystal structures of two H-2Db/ sialyl Lewis(a) and sialyl Lewis(x) antigen levels in the 156. Mercurio, A. M., Bachelder, R. E., Bates, R. C. & Chung, J. α β glycopeptide complexes suggest a molecular basis for preoperative serum of gastric cancer patients. J. Clin. Autocrine signaling in carcinoma: VEGF and the 6 4 CTL cross-reactivity. Immunity 10, 63–74 (1999). Gastroenterol. 34, 408–415 (2002). integrin. Semin. Cancer Biol. 14, 115–122 (2004). 125. Ibrahim, N. K. & Murray, J. L. Clinical Development of the 138. Nakagoe, T. et al. Preoperative serum level of CA19–9 157. Syrokou, A., Tzanakakis, G. N., Hjerpe, A. & Karamanos, N. K. STn–KLH Vaccine (Theratope(R)). Clin. Breast Cancer 3 predicts recurrence after curative surgery in node-negative Proteoglycans in human malignant mesothelioma. (Suppl. 4), 139– 143 (2003). colorectal cancer patients. Hepatogastroenterology 50, Stimulation of their synthesis induced by epidermal, insulin The authors describe the development of a unique 696–699 (2003). and platelet-derived growth factors involves receptors with tumour vaccine by conjugating a synthetic STn 139. Satoh, H. et al. Predictive value of preoperative serum tyrosine kinase activity. Biochimie 81, 733–744 (1999). epitope to a high-molecular-weight carrier, KLH. The sialyl Lewis X-i antigen levels in non-small cell lung cancer. 158. George, D. Targeting PDGF receptors in cancer— STn–KLH vaccine generates a humoral and cellular Anticancer Res. 18, 2865–2868 (1998). rationales and proof of concept clinical trials. Adv. Exp. immune response, and its safety was confirmed in 140. Seelenmeyer, C., Wegehingel, S., Lechner, J. & Nickel, W. Med. Biol. 532, 141–151 (2003). the ongoing Phase III trial described in the article. The cancer antigen CA125 represents a novel counter 159. Wang, L., Fuster, M. M., Sriranardo, P. & Esko, J. D. 126. Holmberg, L. A., Oparin, D. V., Gooley, T. & Sandmaier, B. M. receptor for galectin-1. J. Cell Sci. 116, 1305–1318 (2003). Endothelial deficiency of heparan sulphate impairs L-selectin The role of cancer vaccines following autologous stem cell 141. Hayashi, N. et al. Association between expression levels of and chemokine mediated neutrophil trafficking during rescue in breast and ovarian cancer patients: experience CA 19–9 and N-acetylglucosamine-β;1, 3- inflammatory responses. Nature Immunol. (in the press). with the sTn–KLH vaccine (Theratope(R)). Clin. Breast galactosyltransferase 5 gene in human pancreatic cancer Cancer 3 (Suppl. 4), 144–151 (2003). tissue. Pathobiology 71, 26–34 (2004). Acknowledgements: 127. de Kleijn, E. M. & Punt, C. J. Biological therapy of 142. Schroeder, J. A., Adriance, M. C., Thompson, M. C., We would like to thank A. Varki for many helpful comments. This colorectal cancer. Eur. J. Cancer 38, 1016–1022 (2002). Camenisch, T. D. & Gendler, S. J. MUC1 alters β-catenin- work was supported by a Research Career Development Award 128. Monzavi-Karbassi, B., Cunto-Amesty, G., Luo, P. & dependent tumor formation and promotes cellular from the US Department of Veteran’s Affairs (to M.F.) and National Kieber-Emmons, T. Use of surrogate antigens as invasion. Oncogene 22, 1324–1332 (2003). Institutes of Health (to J.D.E.). vaccines against cancer. Hybrid Hybridomics 21, 103– 143. Luchansky, S. J., Goon, S. & Bertozzi, C. R. Expanding the 109 (2002). diversity of unnatural cell-surface sialic acids. Competing interests statement 129. Lemieux, G. A. & Bertozzi, C. R. Modulating cell surface Chembiochem. 5, 371–374 (2004). The authors declare no competing financial interests. immunoreactivity by metabolic induction of unnatural 144. Saxon, E. & Bertozzi, C. R. Cell surface engineering by a carbohydrate antigens. Chem. Biol. 8, 265–275 (2001). modified Staudinger reaction. Science 287, 2007–2010 Sialic acid is often expressed by tumour-associated (2000). Online links oligosaccharide antigens; the authors explored 145. Dwek, R. A., Butters, T. D., Platt, F. M. & Zitzmann, N. whether intercepting the sialic acid biosynthetic Targeting glycosylation as a therapeutic approach. Nature DATABASES pathway with unnatural precursors might modulate Rev. Drug Discov. 1, 65–75 (2002). The following terms in this article are linked online to: the immunogenicity of targeted tumour cells. 146. Brown, J. R., Fuster, M. M. & Esko, J. D. in Carbohydrate Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. Antibodies generated against the unnatural sialic Based Drug Discovery, Vol. 2 (ed. Wong, C.-H.) 883–898 fcgi?db=gene acid directed complement-mediated lysis of tumour (Wiley VCH, Weinheim, 2003). CD44 | COSMC | E-cadherin | EGF | ERBB2 | FGF2 | glypican-1 | cells. 147. Fritz, T. A., Lugemwa, F. N., Sarkar, A. K. & Esko, J. D. glypican-3 | HB-EGF | IGF1R | IL-8 | MUC1 | MUC16 | MUC4 | 130. Wolfl, M., Batten, W. Y., Posovszky, C., Bernhard, H. & Biosynthesis of heparan sulfate on β-D-xylosides depends TGFβ | VEGF Berthold, F. Gangliosides inhibit the development from on aglycone structure. J. Biol. Chem. 269, 300–307 National Cancer Institute: http://www.cancer.gov/ monocytes to dendritic cells. Clin. Exp. Immunol. 130, (1994). breast cancer | melanoma | ovarian cancer | pancreatic cancer 441–448 (2002). 148. Dimitroff, C. J., Sharma, A. & Bernacki, R. J. Cancer 131. Caldwell, S. et al. Mechanisms of ganglioside inhibition of metastasis: a search for therapeutic inhibition. Cancer FURTHER INFORMATION APC function. J. Immunol. 171, 1676–1683 (2003). Invest. 16, 279–290 (1998). Complex Carbohydrate Research Center: http://www.ccrc. 132. Freeze, H. H., Sampath, D. & Varki, A. α- and β-xylosides 149. Prescher, J. A., Dube, D. H. & Bertozzi, C. R. Chemical uga.edu/ alter glycolipid synthesis in human melanoma and Chinese remodelling of cell surfaces in living animals. Nature 430, Consortium for Functional Glycomics: http://www. hamster ovary cells. J. Biol. Chem. 268, 1618–1627 873–877 (2004). functionalglycomics.org/static/consortium/ (1993). 150. Takenaka, Y., Fukumori, T. & Raz, A. Galectin-3 and Glycobiology journal: http://glycob.oupjournals.org 133. Weiss, M., Hettmer, S., Smith, P. & Ladisch, S. Inhibition of metastasis. Glycoconj. J. 19, 543–549 (2004). Glycobiology Research and Training Center: http://grtc. melanoma tumor growth by a novel inhibitor of 151. Kojima, N. & Hakomori, S. Cell adhesion, spreading, and ucsd.edu glucosylceramide synthase. Cancer Res. 63, 3654–3658 motility of GM3-expressing cells based on glycolipid- National Cancer Institute Glycobiology Resources: http:// (2003). glycolipid interaction. J. Biol. Chem. 266, 17552–17558 glycores.ncifcrf.gov 134. Ragupathi, G. et al. Consistent antibody response against (1991). Oxford Glycobiology Institute: http://www.bioch.ox.ac.uk/ ganglioside GD2 induced in patients with melanoma by a 152. Pili, R. et al. The α-glucosidase I inhibitor castanospermine glycob/ GD2 lactone-keyhole limpet hemocyanin conjugate alters endothelial cell glycosylation, prevents angiogenesis, Society for Glycobiology: http://www.glycobiology.org/pub/ vaccine plus immunological adjuvant QS-21. Clin. Cancer and inhibits tumor growth. Cancer Res. 55, 2920–2926 handler.asp Res. 9, 5214–5220 (2003). (1995). Access to this interactive links box is free online.

Cancer Intelligence Acquired (CIA): tumor glycosylation and sialylation codes dismantling antitumor defense

Cell. Mol. Life Sci. (2015) 72:1231–1248 DOI 10.1007/s00018-014-1799-5 Cellular and Molecular Life Sciences

REVIEW

Cancer intelligence acquired (CIA): tumor glycosylation and sialylation codes dismantling antitumor defense

Kayluz Frias Boligan • Circe Mesa • Luis Enrique Fernandez • Stephan von Gunten

Received: 30 September 2014 / Revised: 27 November 2014 / Accepted: 1 December 2014 / Published online: 7 December 2014 Ó Springer Basel 2014

Abstract Aberrant glycosylation is a key feature of lectins, such as siglecs, selectins, C-type lectins and malignant transformation and reﬂects epigenetic and galectins, may lead to novel treatment strategies, not only in genetic anomalies among the multitude of molecules cancer, but also in autoimmune disease or transplantation. involved in glycan biosynthesis. Although glycan biosynthesis is not template bound, altered tumor glycosylation is Keywords Xeno-autosialylation Cancer glycosylation not random, but associated with common glycosylation Cancer immunoediting Sialoglycans Altered branching patterns. Evidence suggests that acquisition of distinct Tumor immunosuppression glycosylation patterns evolves from a ‘microevolutionary’ process conferring advantages in terms of tumor growth, Abbreviations tumor dissemination, and immune escape. Such glycosyl- C2GnT Core 2 b1,6-N- ation modiﬁcations also involve xeno- and hypersialylation. acetylglucosaminyltransferase Xeno-autoantigens such as Neu5Gc-gangliosides provide CEACAM-1 CEA-related cell adhesion molecule-1 potential targets for immunotherapy. Hypersialylation may CLR C-type lectin receptors display ‘enhanced self’ to escape immunosurveillance and CRD Carbohydrate-recognition domains involves several not mutually exclusive inhibitory pathways DTDST Diastrophic dysplasia sulfate transporter that all rely on protein–glycan interactions. A better FUT Fucosyltransferase understanding of tumor ‘glycan codes’ as deciphered by ITIM Immunoreceptor tyrosine-based inhibition motif LOH Loss of heterozygosity MGAT3 Mannosyl (beta-1,4-)-glycoprotein beta- K. F. Boligan S. von Gunten (&) 1,4-N-acetylglucosaminyltransferase Institute of Pharmacology, University of Bern, MGAT5 Mannosyl (alpha-1,6-)-glycoprotein beta- Friedbu¨hlstrasse 49, 3010 Bern, Switzerland 1,6-N-acetylglucosaminyltransferase e-mail: [email protected] MICA MHC class I-related chain A K. F. Boligan NGcGM3 N-Glycolyl GM3 ganglioside e-mail: [email protected] NAcGM3 N-Acetyl GM3 ganglioside C. Mesa NSCLC Non-small cell lung cancer Immunobiology Division, Center of Molecular Immunology, ppGalNAcT Polypeptide N-acetylgalactosyl transferase 216 St and 15th Ave., Atabey, Playa, P.O. Box 16040, PSGL-1 P-selectin glycoprotein ligand-1 11600 Havana, Cuba e-mail: [email protected] SAMPs Self-associated molecular patterns Siglec Sialic acid-binding immunoglobulin-like L. E. Fernandez lectin Innovation Division, Center of Molecular Immunology, sT Sialyl-T antigen 216 St and 15th Ave., Atabey, Playa, P.O. Box 16040, 11600 Havana, Cuba ST6Gal 1 Beta-galactoside a-2,6-sialyltransferase 1 e-mail: [email protected] sTn Sialyl-Tn antigen 123 1232 K. F. Boligan et al.

T Thomsen-Friedenreich linkage of two CRDs in tandem by a single polypeptide TACA Tumor-associated carbohydrate antigen chain (‘‘tandem-repeat’’ galectins; Galectin-4, -6, -8, -9, VSSP Very small-sized proteoliposomes and -12), or oligomerization induced by a non-lectin N-terminal region (Galectin-3). In analogy to lattices formed by antibodies and multivalent antigens, galectins can form ordered arrays of complexes [14], and may confer Introduction innate antibody-like immunity [11]. Several glycoconjugates constitute danger-associated The ‘‘glycome’’ has been defined as the complete set of molecular patterns (DAMPs; e.g. hyaluronan fragments, glycans and glycoconjugates (covalent complexes with certain proteoglycans) or pathogen-associated molecular proteins or lipids) that are made by a cell or organism under patterns (PAMPs; e.g. lipo-oligosaccharides, peptidogly- specific conditions [1]. The glycome depends on biosyn- cans). Other glycans, including sialic acid-containing thetic pathways that involve the concerted action of glycans (sialoglycans), which are dominant on cells of the glycosyltransferases, glycosidases and other glycan-modi- deuterostome lineage of animals [15] and are recognized fying enzymes that are encoded by 250–500 glycogenes by intrinsic inhibitory receptors of immune cells, could be representing 1–2 % of the total human genome [2]. Despite regarded as ‘self-associated molecular patterns (SAMPs)’ the conservation of glycogenes, major intra- and interspe- [16]. Indeed, hypersialylation of cancer cells may display cies variations of the glycome exist [3, 4], whereby ‘enhanced self’ or ‘super-self’ to exploit natural inhibitory glycosylation patterns are often tissue- or cell lineage- pathways for immune evasion [17]. specific [5, 6]. Thus, glycan biosynthesis represents an Interestingly, certain glycosylation phenotypes occur organized, highly regulated process, even if the generation more frequently in cancer than others (Fig. 1), suggest- of glycans, unlike proteins, does not rely on a template- ing that these changes reflect the outcome of ‘Darwinian based mechanism, but instead on a machinery of different selection’ [9] in terms of the capacity to grow, dissem- enzymes that elongate, branch or trim a specific substrate. inate, and to cope up with diverse microenvironmental By its complex nature, involving multiple players, glycan and immunological survival pressure. Glycan-binding biosynthesis is particularly sensitive to epigenetic and proteins, as sensors of glycosylation patterns, play a key environmental factors (Fig. 1). Yet, given the involvement role in these processes. We discuss common tumor- of glycans in myriad biological processes, tight control of associated glycosylation patterns and highlight the con- glycan biosynthesis and repertoire is required to maintain cept that acquired epigenetic or genetic alterations of the tissue homeostasis and health [7]. An altered cellular gly- glycosylation code displayed by cancer cells, may cosylation profile is not only indicative for a pathological engage or deafen different protein–glycan sensor systems process, but may have functional consequences, and to misinform and subvert actions of the immune system. influence the pathogenesis or progression of disease, A special focus of this article is on recent insights including cancer. Indeed, aberrant glycosylation is a char- related to tumor-associated modifications of the sialome, acteristic feature of carcinogenesis that influences tumor which are among the most prominent features of altered immunity, angiogenesis and multiple steps of tumor pro- tumor glycosylation. gression, such as tumor growth and proliferation, migration, invasion, and metastasis [8, 9]. As a further biochemical dimension to nucleic acids and Tumor glycosylation and sialylation patterns proteins [10], the glycome encodes biological information, which is deciphered by glycan-binding proteins (lectins) Essentially all tumor cells exhibit altered carbohydrate and antibodies [11, 12]. Glycan-binding proteins include composition and structure compared to their progenitors, membrane-associated molecules, such as selectins, siglecs and distinct glycosylation patterns have been associated and most C-type lectins, or soluble proteins such as with oncogenic transformation, which may result from galectins. Latter are evolutionary highly conserved [13], altered glycan biosynthesis, neosynthesis of onco-fetal and by means of carbohydrate-recognition domains antigens or organizational changes in density or crypticity (CRDs) that recognize b-galactose, they are able to at surfaces [18]. Furthermore, certain glycoproteins and establish bivalent or multivalent cell–cell and cell–matrix lipids, such as mucins and gangliosides, are overproduced interactions, and to crosslink transmembrane proteins, by cancer cells, and are eventually secreted or ‘shed’ as eventually resulting in a cascade of transmembrane sig- soluble glycoconjugates [19–22]. Indeed, most serum naling events [14]. The bi- or multivalency of galectins markers currently in use are glycoconjugates that harbor involves the formation of dimers (‘‘prototype’’ galectins; tumor-associated carbohydrate antigens (TACA) [23]. Galectin-1, -2, -5, -7, -10, -11, -13, -14, and -15), the Common cancer glycans include sialyl Lewis x (sLex),

123 Tumor glycosylation and sialylation codes 1233

Fig. 1 Characteristic patterns of tumor surface glycosylation as a and genetic changes. Environmental factors and immune pressure consequence of epigenetically and genetically modified glycan may lead to characteristic tumor-associated cell surface glycosylation biosynthesis. Glycan synthesis is not template bound but involves patterns (see main text), including hypersialylation, truncation, altered the concerted action of glycosyltransferases, glycosidases and glycan- branching, and even xenoglycosylation, which may confer a survival modifying enzymes (e.g. sulfotransferases). The glycan machinery is advantage to face the diverse challenges imposed by the host frequently modified upon malignant transformation due to epigenetic sialyl Lewis a (sLea), sialyl Tn (sTn), Thomsen-Frieden- residues of a carrier protein, whereby the addition of an reich (T), Lewis y (Ley), Globo H, polisialic acid (PSA), N-acetylgalactosamine to the Ser or Thr is mediated by a GD2, GD3, fucosyl GM1, GM2 [19]. Advantages of glycan polypeptide N-acetylgalactosyl transferase (ppGalNAcT), antigens may be that unlike proteins a glycan is not a direct of which at least twelve isoforms are present in mammals. product of genes, and may reflect multiple genetic and O-linked glycans are very heterogeneous, with up to eight epigenetic anomalies and even genetic silencing (incom- different core structures (Core 1–8) [6, 25]. O-linked plete synthesis). glycans have been found to be relevant to mucin structure Glycans range from highly branched and complex and to determine their functions in tumor progression, structures (e.g. N-linked and O-linked glycans, and gly- including cellular growth, differentiation, invasion and colipids) to linear glycans (e.g. proteoglycans). They immune surveillance (reviewed in [20, 22]). include N-glycans, O-glycans, and glycolipids, and are Sialic acids are a family of nine-carbon backbone typically composed of a central core, which is constitutive monosaccharides that are typically found at the outermost in most cell types, an intermediate backbone (e.g. end of glycoconjugates, including N-acetylneuraminic acid repeating poly-N-acetyllactosamine units), and subulti- (Neu5Ac) and the non-human N-glycolylneuraminic acid mate and terminal sugars at ‘outer’ positions [6]. The (Neu5Gc), which are the most common sialic acid struc- biosynthesis of N-linked glycans involves the en bloc tures found on mammalian cells [26]. Sialic acid- transfer of a common 14-sugar glycan to an asparagine containing carbohydrate structures (sialoglycans) and other residue of a protein that is being synthesized and trans- terminal or subultimate sugars, often determine the func- located through the endoplasmatic reticulum (ER) tion(s) or recognition properties of glycoconjugates and are membrane, and the subsequent enzymatic remodeling of frequently altered upon malignant transformation [6, 9, 19]. the glycan structure in the ER and Golgi [24]. O-linked In the following, common patterns or tumor glycosyla- glycans are attached to serine (Ser) or threonine (Thr) tion are discussed.

123 1234 K. F. Boligan et al.

Increased branching of glycans expression was also observed in the highly metastatic derivative Skov3-ip cells of the ovarian adenocarcinoma A common glycosylation aberration in malignancies con- cell line Skov3, which together may contribute to increased stitutes the overexpression of complex b-1,6-branched N- b-1,6-branching of N-linked glycans of metastatic cells linked glycans on the cell surface. This modification results [41]. MGAT3 and the bisecting GlcNAc have also been from the enhanced transcription of Mgat5, in which prod- reported to reduce galectin–lattice dependent growth factor uct enzyme N-acetyl-glucosaminyltransferase V (GnT-V; signaling leading to retarded tumor progression [42]. GlcNAcT-V; MGAT5) catalyzes the transfer of GlcNAc Taken together, the competition between MGAT5 and residues to growing N-glycans, thus leading to ‘multi-an- MGAT3 with altered expression of b-1,6-branched tennary’ chains [27]. Increased expression of b-1,6- N-linked glycans or bisected plays a fundamental role in branched N-linked glycans and MGAT5 is critically tumor growth and metastasis [32, 33]. The reduction of involved in tumor growth and metastasis [28–30], which bisecting N-glycans and the concomitant increase of b-1,6- seems to involve modified cell signaling and adhesion branched N-linked glycans, as a consequence of altered functions as a consequence of altered N-glycosylation of transcription of enzymes involved in N-glycan synthesis, cell surface receptors [31–33]. Altered N-glycosylation may lead to hypersialylation and modulate interactions activity of MGAT5 has been shown to modulate cell with glycan-binding proteins that are relevant to tumor apoptosis [34] and to promote cancer metastasis by progression and immune escape. affecting the stability or function of relevant proteins [32], including cadherins, integrins, or matriptase [35, 36]. An High-density glycosylation sites increase of b-1,6-branched N-linked glycans on integrins has been reported to promote cell-extracellular matrix Focal enrichment of carbohydrate structures on the cell (ECM) interactions and tumor dissemination by enhanced membrane may lead to different effects compared to the adhesion to fibronectin and basement membrane laminin same glycans distributed at low density [18]. The focal [37]. In addition, altered expression of MGAT5 in tumors density of glycolipids is influenced by coexisting mem- not only benefits tumor metastasis but also impairs a cor- brane components and may affect oncogenetic events, rect antitumor immune response, as it has been shown to tumor immunosurveillance or the recognition of tumor- alter cytokine-mediated leukocyte signaling [31] and associated antigens by diagnostic antibodies [18, 21]. cytokine secretion by splenocytes and proliferation of Gangliosides are sialic acid-containing glycosphingolipids CD4? T cells [38]. Increased branching of glycans creates (GSLs) that contain a hydrophilic glycan head group linked addition sites for terminal sialic acid residues leading to an to a ceramide anchor. Together with other membrane increase in global sialylation [19], which may modulate sphingolipids and cholesterol, gangliosides can segregate interactions with sialic acid-specific-binding proteins, and form dynamic nanoscale ‘‘clusters’’, so-called lipid including selectins and siglecs (discussed below). Inter- rafts; more highly unsaturated components, such as glyc- estingly, MGAT5 itself also acts as an inducer of erophospholipids, provide the membrane with flexibility angiogenesis through a mechanism independent of glyco- [21]. Through lateral carbohydrate–protein or carbohy- sylation that involves the release of fibroblast growth drate–carbohydrate interactions gangliosides regulate factor-2 from cell surface or ECM heparin sulfate proteo- signaling proteins in cis, such as epidermal growth factor glycan [39]. receptor (EGFR) or vascular endothelial growth factor Mgat3 encodes N-acetylglucosaminyltransferase III receptor (VEGFR), or even in trans, to modulate cellular (GnT-III; GlcNAcT-III; MGAT3), a glycosyltransferase differentiation, growth, and processes related to oncogen- that transfers a ‘‘bisecting’’ N-acetylglucosamine residue to esis, metastasis or tumor immunity [21]. The association of a b-linked mannose in the N-glycan core. The presence of gangliosides with cholesterol and other constituents of lipid the bisecting N-acetylglucosamine inhibits the action of an rafts can lead to conformational changes in the glycolipid a-mannosidase, which is involved in the early biosynthetic head group [21], and thereby affect carbohydrate-related steps of complex N-glycans, including 1,6-branched functions of gangliosides. Ganglioside clustering may be of N-linked glycans [24, 30]. MGAT3 activity with formation special importance for the function of organized structural of bisecting N-glycans may therefore regulate the synthesis domains, such as the immunological synapse formed of branched N-glycans. In cancer, epigenetic control on the between immunological effector cells and the targeted expression of MGAT3 seems to be responsible for the tumor cell [43]. Several antibodies recognizing tumor occurrence of bisected N-glycans [40]. Overexpression of antigens, such as on melanoma or Burkitt lymphoma, were MGAT3 with concomitant decrease in b-1,6-branches found to bind to gangliosides organized in clusters on suppresses the occurrence of experimental induced metas- tumor cells, but failed to bind when their density was tasis [30]. Decreased MGAT3 and increased MGAT5 below a certain threshold, suggesting altered density and 123 Tumor glycosylation and sialylation codes 1235 structural characteristics of glycosylation sites of malignant Thr) antigens, and their sialylated derivatives sTn and sT. cells compared to their healthy parental cells [18, 44]. The appearance of the Tn and sTn structures results from Glycoproteins can also contribute to high-density gly- the lack of b1,3-galactosyltransferase (C1b3Gal-T) activ- cosylation sites. Mucins harbor multiple serine and ity, which most probably originates from mutations in threonine O-glycosylation sites in a tandem-repeat domain Cosmc [48]. Dysfunction of Cosmc, a chaperon responsible of identical or highly similar sequences, and may exhibit for folding and stability of b1,3-galactosyltransferase (T altered density of tandem-repeat glycosylation in cancer synthase), prevents the b1,3-galactose extension of the compared to normal cells [20]. Carcinoma mucins, aber- simplest O-glycan Tn antigen, resulting in its accumula- rantly glycosylated with either Tn or sTn, frequently form tion. Similarly, decreased activity of the enzyme b1,6- clusters on the tumor surface [20]. Such high-density gly- GlcNAc transferase (C2GnT1) prevents further extension cosylation sites may lead to aberrant functions during the of the T antigen (core 1) to form O-glycan core 2, resulting pathogenesis of cancer, affect immune responses, or, in in the accumulation of T or sT antigens in malignant tis- analogy to clustered gangliosides, influence antibody sues. The sialylation of Tn or T antigens, yielding sTn and attachment [9, 20]. sT antigens, prevents further glycan extension, which is often promoted by enhanced sialyltransferase expression in Aberrant O-glycan synthesis resulting in truncated cancer, and has been associated with increased tumor or incomplete glycans growth, dissemination [8], and maturation and activity of dendritic cells, causing insufficient stimulation of T cells A block of glycan synthesis often results in incomplete [49]. Mechanisms other than incomplete synthesis may glycan structures, with or without accumulation of pre- contribute to enhanced formation of Tn, T, sTn or sT cursors [18]. As a consequence, malignant cells tend to antigens in cancer, including enhanced availability of the express glycans with simpler structures than those found in substrate UDP-galactose or UDP-galactose transporter [8]. normal cells [23]. In certain cases, the accumulation of Truncated antigen expression in cancer is also evidenced specific incomplete glycans is detectable by diagnostic by alterations in the ABO antigens. Reduced levels of A antibodies, such as sialyl Lewis A (Lea), the glycan epitope and B antigens have been reported in leukemia, bladder of the tumor marker CA 19-9 for digestive organs, and oral cancer, among others [50, 51]. Specific allelic loss, including pancreas and biliary tract [23]. Disialyl Lea, in oral carcinomas, is in part responsible for the dysregu- which is expressed on non-transformed cells, differs lation in A and B antigens expression. Additionally, in structurally by the presence of one extra sialic acid bladder cancer loss of heterozygosity (LOH) and micro- attached to the C-6 position of bGlcNAc. Reduced tran- satellite instability is related to ABO modifications due to a scription of the sialyltransferase responsible for the a2-6 chromosomal deletion at 9q34. Moreover, DNA methyla- sialylation at this position has been shown to be signifi- tion of the CpG islands in the proximal promoter of the cantly decreased in cancer due to epigenetic silencing, A/B encoding gene corresponds to the loss of blood anti- resulting in loss of the more complex disialyl Lea, and gens [50]. concomitant acquisition of sialyl Lea expression [45]. Importantly, the incomplete synthesis of glycans may In analogy, sialyl Lex is utilized as a tumor marker. In result in the appearance of tumor-associated carbohydrate contrast, its 6-sulfated derivative, sialyl 6-sulfo Lex,is antigens of simpler structure that may exhibit altered preferentially expressed in non-transformed cells [46]. binding capacity to glycan-binding proteins, including the Reduced expression of molecules responsible for sialyl Lex sialic acid-binding receptors selectins and siglecs [23]. sulfation, including 6-sulfotransferase, PAPS synthase, and sulfate transporter was observed in cancer, whereby Hypersialylation reduced transcription of the sulfate transporter diastrophic dysplasia sulfate transporter (DTDST) seems to have a Hypersialylation, in one of the most common glycosylation dominant role for the reduction of sialyl 6-sulfo Lex and changes in cancer and its aptitude to promote tumor growth gain of non-sulfated sialyl Lex [47]. Silencing of DTDST in was already recognized in the late 1960s [27, 52]. The malignancies appeared to be epigenetically regulated, enhanced density of sialylated epitopes on tumor cells whereby histone modification rather than DNA methylation compared to normal tissue has led to the establishment of turned out to be responsible for controlling the expression novel tumor markers, such as CA19-9, SLX, CSLEX or of this gene. NCC-ST-439, which have been successfully utilized in Incomplete synthesis at early steps of O-linked glycan cancer diagnosis [18, 23]. Upregulation of sialyltransfe- generation is linked to the appearance of short truncated rases is one dominant mechanism underlying glycans in cancer, such as the Tn (GalNAc-a1-O-Ser/Thr) hypersialylation in cancer, the transcription of which is and T (Thomsen-Friedenreich; Galb1-3GalNAc-a1-O-Ser/ controlled by the proto-oncogenes, Ras and c-Myc, and can 123 1236 K. F. Boligan et al. be induced by hypoxia [23, 52]. Furthermore, increased are eventually induced by dietary-Neu5Gc uptake via substrate availability, as a consequence of overexpression commensal bacteria [73, 74]. Evidence suggests that of genes involved in sialic acid biosynthesis or transport Neu5Gc-containing epitopes incorporated in human tissues molecules, may lead to hypersialylation of cancer cells [23, from dietary sources (primarily red meats) may act as 53]. Conversely, reduced degradation of sialoglycans due ‘xeno-autoantigens’ and in interaction with Neu5Gc-spe- to declined expression of endogenous sialidases, enzymes cific ‘xeno-autoantibodies’ may stimulate chronic that cleave sialic acids from glycans, has been proposed as inflammation and promote cancer, or other diseases, such a potential mechanism for tumor hypersialylation [54]. In as atherosclerotic vascular disease [73, 74]. This concept addition, several forms of altered tumor glycosylation, has recently been coined as ‘xenosialitis hypothesis’ [73]. including enhanced glycan branching and the accumulation Interestingly, anti-Neu5Gc antibodies exhibit dualistic of truncated sialic acid acceptor molecules, contribute to effects on cancer growth in animal models, whereby lower global tumor cell hypersialylation. Importantly, the clus- antibody concentrations stimulate, and higher concentra- tering of sialoglycans in functional microdomains, such as tions inhibit tumor growth [72, 75, 76]. Murine tumors lipid rafts [21] or the immunological synapse [43, 55], may expressing human-like levels of Neu5Gc exhibited accel- significantly influence processes, such as tumor growth, erated growth in Cmah-/- mice with a human-like dissemination or immunosurveillance. deficiency of Neu5Gc, which was associated with the induction of anti-Neu5Gc antibodies and increased infil- Expression of xenoglycans tration of inflammatory cells [76]. Similarly, passive transfer of anti-Neu5Gc antibodies promoted tumor growth The predominant sialic acids on most mammalian cells are at lower concentrations, while inhibitory effects were N-glycolylneuraminic acid (Neu5Gc) and N-acetylneu- observed at higher concentrations [76]. This concentration- raminic acid (Neu5Ac). These two sialic acids differ by dependent dualistic effect is mediated by remarkably nar- one oxygen atom which is added via hydroxylation of the row ranges of antibodies and is also observed for other cytidine-50-monophosphate-Neu5Ac (CMP-Neu5Ac) into tumor-directed antibodies than anti-Neu5Gc, suggesting CMP-Neu5Gc. This reaction is catalyzed by the enzyme that these inverse hormesis effects represent a general CMP-Neu5Ac hydroxylase (cmah) and is decisive in the characteristic of tumor-directed antibodies [75]. This levels of Neu5Gc glycoconjugates produced by a cell [56– biphasic tumor growth response may involve cancer pro- 58]. In humans, the presence of the Neu5Gc variant is moting effects of M2-polarized tumor-associated practically undetectable in normal cells, due to the inacti- macrophages at lower concentrations, and NK cell antitu- vating deletion in the cmah gene [59]. Despite the genetic mor responses at higher, tumor-inhibiting concentrations incapability of human cells to synthesize Neu5Gc, [75]. numerous reports have demonstrated the expression of this sialic acid variant on glycoproteins, but especially on gangliosides, by cells from various human malignancies, Immunological implications of tumor glycosylation including colon and breast cancer [60, 61], melanoma [62], or retinoblastoma [63]. Protein–glycan interactions have been reported to modulate In humans, the absence, or very rare expression of the activity of multiple arms of cellular and humoral glycoconjugates carrying Neu5Gc sialic acid in normal immunity including immune cells [77, 78], antibodies [79], cells indicates that this sialic acid variant needs to be and complement [80]. An abundance of glycan-binding incorporated from external sources, such as human diet proteins, i.e. lectins, exists that recognizes specific glycans [64–66]. However, it is assimilated preferentially in cancer and endows immune cells the capacity to decipher the cells with respect to normal cells. Two causes have been ‘‘Glyco-Code’’ of healthy or diseased host cells and associated to this differential expression. The increased pathogens to induce or support immune responses, or to metabolic rate of tumor cells is generally accepted as one maintain immune homeostasis [81, 82]. Although certain of the causes [67] and is also relevant to the hypoxic families of lectins are defined by binding to a specific condition inside the tumor mass [68]. It has been demon- terminal carbohydrate moiety, e.g. to sialic acid for ‘sialic strated that during hypoxia the transcription of Sialin, a acid-binding immunoglobulin-like lectins’ (siglecs), the sialic acid transporter, on cancer cells is induced [69]. This binding specificity of an individual family member often transporter mediates the incorporation of external sialic depends on additional parameters including carbohydrate acid in vitro from the culture medium and, presumably, linkage, type of subterminal structures, and chemical also facilitates this phenomenon in vivo [69] (Fig. 2). modifications (e.g. sulfation) [83]. Several studies reported the occurrence of xenoantibodies Altered tumor glycosylation promotes or diminishes to Neu5Gc in the serum of healthy individuals [70–72], that interactions with various lectin families including the 123 Tumor glycosylation and sialylation codes 1237

Fig. 2 Incorporation of Neu5Gc-containing xenoglycans in human cancer. Neu5Gc is rare in human tissues due to the deletion of the enzyme that mediates hydroxylation of CMP-Neu5Ac into CMP-Neu5Gc (CMAH), and animal dietary sources may constitute the main supply of Neu5Gc. Under hypoxia, commonly found in human solid tumors, the expression of a lysosomal sialic acid transporter, Sialin, is enhanced. This leads to increased disponibility of free cytosolic Neu5Gc derived from lysosomal digested glycoproteins or -lipids, eventually resulting in overexpression of Neu5Gc- containing xenoglycans on the cancer cell surface

galectins that recognize the disaccharide N-acetyllactos- protects from NK cell-mediated cytotoxicity [55]. Pre- amine [Galb(1–4)-GlcNAc; LacNAc], the heterogenous treatment of peripheral blood lymphocytes (PBL) with C-type lectin receptors (CLRs), which predominantly rec- sialidase, to unmask Siglec-7 receptors bound in cis to ognize asialylated glycans (e.g. mannose-, fucose-, ligand on the same cell surface, reduced cytotoxic galactose-, or GalNAc-containing carbohydrates), and the responses towards cell lines expressing Siglec-7 ganglio- sialic acid-binding selectins and siglecs. Notably, the side ligands in vitro, such as GD3 synthase-transfected atypical presence of sialic acids on cognate ligands, as P815 cells overexpressing GD3 [43] or ACHN cells- often occurring in cancer, can reduce or abrogate recog- expressing DSGb5 [86]. More recently, we showed that nition by lectins with specificity for asialylated glycans, sialidase treatment of NK cells is dispensable for siglec- such as the galectins [84]. mediated downregulation of NK cell responses by certain Modified lectin–glycan interactions have been shown to types of cancer that highly express Siglec-7 or -9 ligands affect both innate and adaptive immunity in cancer, even- [85]. Enzymatic removal of Siglec-7 or -9 ligands on tumor tually resulting in tumor-associated inflammation or cells by sialidase or interference with blocking antibodies undesired immune escape [8, 9, 81]. Thus, the outcome of to Siglec-7 or -9 both dramatically enhanced the NK cell- an immune response is often critically influenced by the mediated killing of target cells, not only of NK cell-sen- specific glycosylation profile of the malignant cell, which sitive K562 cells, but also of presumably NK cell-resistant may involve the concerted and mutually non-exclusive tumor cells. Furthermore, the recovery of peritoneally action of lectins with different glycan-binding preferences. injected K562 and HeLa cells in humanized NOD-SCID- - - For instance, several different mechanisms involving gly- cc / (huNSG) mice with NK cells-expressing human can-lectin interactions have been found to influence natural Siglec-7 was significantly lower for desialylated tumor killer (NK) cell-mediated tumor immunosurveillance cells compared to Siglec-7/-9 ligands-expressing cells [85]. (Fig. 3). Removal of sialic acids on tumor cells has also been Recent studies indicate that overexpression of sialogly- shown to enhance binding of the activating NK cell can ligands for the inhibitory NK cell receptors Siglec-7 receptor NKG2D to ligands on the tumor cell and to and -9 on tumor cells impairs NK cell antitumor defense enhance NKG2D-dependent NK cell cytotoxicity, and it mechanisms. While Siglec-7 is expressed on all human NK was proposed that non-specific charge repulsion of sialic cells, Siglec-9 defines a subset of cytotoxic NK cells with acid residues in their vicinity, or on NKG2D ligands enhanced chemotactic potential that is reduced in the themselves, might be responsible for this effect [87]. It is peripheral blood of cancer patients [85]. Coating of tumor possible that tumor hypersialylation synergistically down- cells by passive membrane insertion with synthetic sialy- regulates NK cell antitumor activity by attenuation of lated glycopolymers that serve as ligands for Siglec-7 activatory NKG2D receptor–ligand interactions and 123 1238 K. F. Boligan et al.

Engagement of inhibitory Siglec-7 and -9

Cancer cell Siglec-7/-9

NK cell

Abrogation of NKG2D activation

NKG2D MICA

- - -

Gal3

Platelet-mediated TGF-β release

TGF-β receptor Mucins

β TGF- Platelet

Fig. 3 The multiplicity of putative protein–glycan interactions that interaction of activatory NK cell receptor NKG2D with its ligand determines NK cell tumor immunosurveillance. Overexpression of MICA on the cancer cell, as a result of enhanced MICA sialylation, speciﬁc sialic acid (purple diamonds)-containing carbohydrates masking by sialoglycans, or Galectin-3 (Gal3) binding to glycosylated (sialoglycans) leading to suppressed NK cell reactivity by engage- MICA. TGF-b release of P-selectin-sialoglycan bound platelets, as ment of their inhibitory Siglec-7 and -9 receptors. Interference of the found in tumor emboli, may result in suppressed NK cell activity concomitant engagement of inhibitory siglec receptors on Platelets contribute to tumor dissemination by formation NK cells. of tumor microemboli, whereby platelet adhesion to tumor Another glycosylation-related mechanism that leads to cells is primarily mediated by P-selectin and its interaction tumor resistance towards NK cell lysis was reported by with sLex/a-carrying structures on tumor cells [89]. For Tsuboi et al. [88] and involves the presence of poly-N- instance P-selectin mediates the adhesion of platelets to acetyllactosamine on the NKG2D-binding site of its ligand neuroblastoma and non-small cell lung cancer cells [90]. MHC class I-related chain A (MICA) on tumor cells Similarly sulfated ceramides in the colon carcinoma MC- overexpressing the glycosyltransferase C2GnT. In this 38 cells that serve as ligands for P-selectin foster the study, soluble lectin Galectin-3 was found to bind to poly- adhesion of activated platelets to tumor cells [91]. N-acetyllactosamine on MICA, thereby inhibiting the P-selectin mediated adhesion of platelets not only enhances NKG2D-MICA interaction with concomitant reduction of metastatic spread and intravascular tumor-cell survival [8, NK cell function, including IFN-c and granzyme B secre- 92], but appears to shield circulating tumor cells from NK tion and tumor lysis. Expression of C2GnT was found to be cell lysis [93–95]. Recently, it has been shown that platelet- higher in metastatic bladder cells, and based on patient derived transforming growth factor b (TGF-b) impairs data; Tsuboi et al. concluded that the expression of C2GnT NK cell antitumor immunity by downregulation of the NK might be a more useful prognostic indicator than patho- cell activating receptor NKG2D with concomitant reduc- logical grade and stage in bladder tumors. tion of NK cell degranulation, cytokine production and

123 Tumor glycosylation and sialylation codes 1239 cytotoxicity [95]. The enhanced expression of P-selectin multiple biological systems and conditions, including sialoglycan ligands on tumor cells correlates with cancer immunity and cancer [80]. Several classes of sialic acid- progression [8], which may be attributed to events that binding proteins, including siglecs, selectins and comple- involve both endothelial and platelet P-selectin that toge- ment factor H, have been shown to be critically involved ther promote metastatic tissue colonization and immune in tumor progression and antitumor immunity, often as a escape. functional consequence of enhanced or altered sialylation Altered glycosylation patterns may not only protect patterns of cancer cells. For instance, hypersialylation has tumor cells from innate immunity, but may shape emerging been shown to endow tumor cells the capacity to evade adaptive immune responses by the display of specific complement-mediated attack by binding of Factor H, glycan ligands to lectins on antigen-presenting cells which negatively regulates the alternative complement (APCs) [78, 96]. Dendritic cells (DCs) express myriad pathway [100–102]. While the role of siglecs and selectins glycan-binding receptors, including siglecs and various in NK cell-mediated antitumor defense has been discussed members of the heterogenous C-type lectin family, such as above, further tumor-related aspects of these receptors are DC-SIGN, BDCA2, DCIR and MICL, which modulate discussed next. toll-like receptor (TLR) signaling, or receptors that directly influence gene expression, eventually by involvement of Siglecs in cancer nuclear factor-jB (NF-jB) as in the case of Dectin 1 [82]. Depending on the tumor ‘glycosylation signature’ different Siglecs constitute a family of cell surface receptors with an lectins will be engaged that trigger DC programs that extracellular domain composed of a sialoglycan-binding potentiate or suppress antitumor responses [81]. For N-terminal V-set domain, a variable number of C2-set Ig instance, while conjugation of specific glycan ligands domains, a transmembrane region and a cytoplasmic tail, might lead to enhanced presentation of tumor-associated latter of which in most Siglec members contain one or antigens to CD4? T cells or cross-presentation to CD8? T more membrane-proximal immunoreceptor tyrosine-based cells [97, 98], Lewis glycans expressed on carcinoembry- inhibitory motif (ITIM) and a membrane-distal ITIM-like onic antigen (CEA) or CEA-related cell adhesion motif [83, 103–105]. The siglec family comprises an evo- molecule-1 (CEACAM-1), in contrast, have been shown to lutionary conserved group, including sialoadhesin (Siglec- negatively affect DC function, which may result in 1), CD22 (Siglec-2), myelin-associated glycoprotein impaired antitumor responses [99]. (MAG; Siglec-4) and Siglec-15, a group of rapidly evolv- Galectins can be released by tumor cells or immune ing siglecs including CD33 and the CD33 (Siglec-3)- cells and have caught special attention in terms of their related siglecs (CD33rSiglecs; Siglec-5 to -14, and -16) dual role in antitumor defense or immune escape. Different [105–107]. The species-related differences of CD33rSig- members of this family have shown to modulate the sur- lecs require special attention for the design of experimental vival and function of both myeloid and lymphoid in vivo models [85, 108]. Besides adhesion, endocytosis leukocytes, resulting either in pro- or anti-inflammatory and pathogen internalization, siglec functions include immune responses [14]. Tumor secretion of galectins is a immunoregulatory processes [103, 105, 109], such as mechanism adopted by a wide range of malignancies to inhibition of cellular activity and proliferation, or regula- promote an immunosuppressive environment [81]. Besides tion of leukocyte survival [110–114]. The siglec-based impairing NK cell responses to tumor cells (see above), immunoregulatory system seems to include naturally galectins have also been shown to affect adaptive immune occurring and anti-idiotypic antibodies to siglecs [115– responses by attenuating DC responses, limiting T cell 117], which may contribute to the anti-inflammatory survival and favoring anergy or exhaustion of tumor-spe- effects of high-dose intravenous immunoglobulin (IVIG) cific T cells and the expansion of regulatory T cells (Tregs) [79, 114, 118–120]. It has been proposed that by their [8, 14, 81]. capacity to discriminate between PAMP and ‘SAMP’ or DAMP, siglecs may play a role in the protection of self tissue as well as the repression of tissue damage-induced The role of sialic acid-binding proteins in cancer immune responses [16, 121]. By hypersialylation, as display of ‘enhanced self’, tumors seem to exploit inhibitory Until the 1980s, the functions of sialic acids were pri- siglecs for immune evasion. marily associated with providing negative charge and Recently, we demonstrated overexpression of Siglec-7 hydrophilicity to vertebrate cell surfaces, masking of and -9 ligands on melanoma cells in histological skin glycan ligands, or serving as attachment sites of pathogens sections and on primary CLL and AML cells from patients and toxins. More recent research revealed a functional role [85]. In a subsequent, independent study, the upregulation of their interaction with sialic acid-binding proteins in of Siglec-9 ligands in sections of colorectal, breast, 123 1240 K. F. Boligan et al. prostate, ovarian, and non-small-cell lung cancer (NSCLC) While the exact sialoglycan ligands of siglecs expressed was reported [122]. In contrast, no overexpression of Sig- on distinct types of tumors still remain elusive, a number of lec-7/-9 ligands was found in basal cell carcinoma (BCC), mucins have been shown to bind to members of this lectin squamous cell carcinoma (SCC), and cutaneous T cell family and to eventually regulate leukocyte functions as lymphoma (CTCL). Furthermore, Siglec-7/-9 ligand cell membrane bound or soluble factor [125–127], expression was heterogeneous in a broad range of different including binding of the tumor marker CA125/MUC16 that cell lines. These findings point to some degree of vari- interacts with Siglec-2, -3, -7, -9 and -10 [128, 129]. ability in terms of siglec ligands expression in tumors of Notably, siglec interactions with their mucin counterre- different tissue-specific origin, which has implications in ceptors might influence tumor progression in two directions terms of tumor immune evasion strategies or potential [130], i.e. transduction of inhibitory signals in immune therapeutic approaches involving siglecs or their ligands. cells by siglecs, or by induction or modulation of mucin- Siglec-7/-9 ligand expression may also occur in certain mediated signaling pathways [130, 131]. Vascular adhesion non-malignant tissues, and it has been hypothesized that protein-1 (VAP-1) was identified as a binding partner of loss of immunosuppressive glycans during colonic carci- Siglec-9 and -10, which might be exploited for imaging of nogenesis enhances inflammatory mediator production vasculature at sites of inflammation and cancer [132, 133]. [123]. Siglecs themselves given their function, their cell-specific Recent evidence suggests that overexpression of native expression on leukocyte subsets and their endocytotic siglec ligands by tumor cells dampens innate immune capacities, exhibit properties that might be exploited for responses mediated by NK cells [85] and neutrophils [122]. cell-directed therapy. Various approaches targeting siglecs Laübli et al. [122] reported that engagement of human are currently tested in pre-clinical and clinical studies for Siglec-9 in vitro and of related murine Siglec-E in vivo the treatment of leukemia and inflammatory disorders [105, inhibits the tumoricidal activity of neutrophils. Interest- 134–136]. ingly, although tumors developed later in Siglec-E null (SigE-/-) mice in a 3-methylcholanthrene (MCA) model Selectins in malignant disease of carcinogenesis, they grew faster and larger, once they appeared, and exhibited increased infiltration of tumor Selectins compose a family of C-type lectins expressed on growth-promoting M2 macrophages. In line with the pro- the surface of leukocytes and endothelial cells, known to angiogenic function of M2 macrophages, increased mediate cellular adhesion processes in immune responses angiogenesis was also observed in tumors of SigE-/- mice. or in cancer [23, 137, 138]. Selectins are type-I receptors The authors hypothesized that in their model, the expres- with an extracellular part comprising a C-type lectin sion of Siglec-E ligands might efficiently suppress innate domain, an EGF domain and a variable number of short immunity during early tumorigenesis or metastasis, how- consensus repeat domains, as well as a transmembrane and ever, at a later stage the expression of Siglec-E ligands a short cytoplasmic tail of maximally 35 aa that can might induce polarization to M1 macrophages, but even- interact with cytoskeleton components and is responsible tually, because the tumor is already at a larger size, this of targeting selectins to different cellular compartments does not lead to rejection. Furthermore, these investigators [139]. This receptor family is constituted of three mem- found an association of Siglec-9 polymorphism with early bers: E-selectin expressed on endothelial cells, L-selectin survival of NSCLC patients [122]. Together, these data on leukocytes and P-selectin on platelets and endothelial suggest a dual function of myelomonocytic cells in cancer cells. In contrast to L-selectins that are constitutively progression and that targeting siglec receptor–ligand expressed on all myeloid cells and on subsets of blood- interactions might be beneficial in the correct temporal and borne lymphocytes, E- and P-selectin expression on situative context. In a recent study, transgenic mice endothelium or platelets is mainly induced following cel- expressing a soluble form of human Siglec-9 (sSiglec-9) lular activation. P-selectin is stored in the Weibel–Palade showed longer survival and less advanced tumor progres- bodies of endothelial cells or a-granules of platelets, and sion compared to non-transgenic mice following can be rapidly redistributed to the cell membrane following intraperitoneally transplants of mouse mammary tumor cellular activation. Selectins can also be found in soluble cells-expressing human MUC1 [124]. Although, as pro- form in the circulation mainly associated with damage of posed by the authors, sSiglec-9 in this model might involve the endothelium, cell activation and certain pathological competitive inhibition of MUC1 binding to certain recep- conditions [140–142]. tors, the identity and tissue expression of latter remains Selectins interact by means of their ligand-binding unclear, especially given that Siglec-9 is not naturally domains with sulfated glycoconjugates on a broad range of expressed in rodents. glycoproteins and -lipids, including mucins, heparin, and

123 Tumor glycosylation and sialylation codes 1241 heparan sulfate [143]. The minimal recognition motif for Xeno-autosialylation: Neu5Gc gangliosides in cancer selectins is Lewis structures, specifically sLea and sLex, the generation of which involves the a(1,3)-fucosyltransfe- In humans the repertoire of sialoglycans, the sialome [26], rases IV or VII (FUT4 and FUT7), a2,3-sialyltransferases, undergoes dramatic modifications upon malignant trans- b1,4-galactosyltransferases and N-acetyl-glucosaminyl- formation that not only relates to hypersialylation of transferases [137]. Carcinoma mucins carrying sLex/a endogenous Neu5Ac glycans, but also to neoexpression of moieties represent major selectin ligands on cancer cells; Neu5Gc-containing xenoglycans. To date the role of other selectin ligands on malignant cells include CD24, endogenous Neu5Ac-gangliosides in tumor biology and CD44, death receptor 3, ESL-1 and PSGL-1 [137]. The immunosuppression has received considerable attention. acquisition of selectin ligands seems to permit cancer cells Overexpression of this kind of gangliosides can provoke a to engage in similar mechanisms of adhesion and extrava- reduction on tumor cell motility and a propensity to sation as used by leukocytes [23, 137, 144]. Furthermore, undergo apoptosis [149–151]. Also, it is well established selectins seem to promote metastasis by shaping the meta- that N-acetyl GM3 ganglioside (NAcGM3) interacts with static niche [137]. Indeed, mice that are deficient in FUT7 the epidermal growth factor receptor (EGFR) inhibiting the and display a minimal amount of selectin ligands exhibit EGF-induced receptor activation [152]. Additionally, var- reduced metastasis formation in comparison with their ious Neu5Ac-gangliosides are known to promote a wild-type counterparts [145]. Attenuation of tumor metas- suppressive tumor microenvironment [21, 43, 153, 154]. In tasis was observed in mice deficient in L- and P-selectins, or contrast, the biological consequences of aberrant expres- mice treated with heparin, as a primary consequence of sion of xenogeneic Neu5Gc-gangliosides for tumor biology P-selectin inhibition [146]. Similarly, formation of sponta- have not been studied in depth. Experimental evidence in neous distant metastases of the human OH-1 small cell lung this sense is derived from murine tumor cell lines in which cancer (SCLC) cell line xenografted into E-/P-selectin- the cmah gene has been knocked-out and the Neu5Gc sialic deficient mice was significantly reduced [147]. By intravital acid has been incorporated exogenously [155]. In this microscopy of murine mesenterial vasculature the authors sense, Gabri et al. [156] demonstrated that mouse B16 found that the OH-1 cells mimic the selectin-dependent melanoma and F3II mammary carcinoma cells induced a rolling behavior of leukocytes along vessel walls. higher number of lung metastasis and increased tumor The different selectins, E-, L- and P-selectin, seem to progression after in vitro incorporation of Neu5Gc-rich synergistically act in the tissue colonization by mediating mucin on the membrane. Similarly, GM3-expressing 3LL- heterotypic interactions between cancer cells, leukocytes D122 Lewis lung carcinoma cells produced more experi- and endothelial cells within the metastatic microenvironmental lung tumor metastasis, if they were pre-incubated ment [8, 137]. Tumor embolus formation during with Neu5Gc [157]. More recently, it has been published intravascular circulation or after microvascular arrest of that silencing the cmah gene in N-glycolyl GM3 ganglio- tumor cells in distant organs involves P-selectin-mediated side (NGcGM3)-expressing L1210 mouse lymphocytic cancer cell-platelet interactions [89]. These events lead to leukemia B cells caused a marked expression of NAcGM3 local activation of endothelial cells recruitment and acti- and an impaired tumor development in vivo [155]. These vation of leukocytes that depend on both L-selectin and observations contrast with a report showing that B16 selectin ligands (e.g. PSGL-1) on leukocytes, latter of melanoma cells transfected with the cmah gene and thus which may bind either to vascular selectins, P- and/or expressing NGcGM3 displayed a reduced tumor growth as E-selectins, or through binding to L-selectin on already compared with the parental cell line [158]. Other evidence adherent leukocytes [145, 148]. Recruited leukocytes, supporting the concept that Neu5Gc expression promotes including monocytes and descendant macrophages, seem to tumor progression is the correlation of NGcGM3 expres- increase vascular permeability and transendothelial sion with the stage of malignancy, with enhanced migration of cancer cells, or may eventually contribute to expression in metastatic lesions [159]. Furthermore, xe- local immunosuppression [137]. Given that selectins are nosialitis induced by anti-Neu5Gc antibodies may promote predominantly implicated in hematogenous metastasis, cancer progression by enhancing tumor-related inflamma- selectin-targeting therapeutics might be beneficial in the tion, as discussed above and reviewed in [73, 74]. Yet, prevention of cancer cell dissemination eventually after other work points to a immunosuppressive capacity of both surgical removal of the primary tumor. Notably, unfrac- Neu5Ac and Neu5Gc-gangliosides, which exhibited similar tioned and certain low molecular weight heparins exhibit capacity to downmodulate CD4 molecules on the T lym- both anticoagulant and selectin blocking activity and are phocyte surface or to inhibit dendritic cells differentiation currently evaluated in clinical trials administered in addi- and maturation in vitro [160, 161]. Taken together, current tion to adjuvant systemic treatment after surgery of evidence indicates that preferential expression of the localized tumors [137]. Neu5Gc-glycoconjugates on tumor cells could provide a 123 1242 K. F. Boligan et al. mechanism utilized by cancer to promote tumor progres- cell death of NGcGM3-expressing X63 murine myeloma sion, metastasis and to shape immune responses (reviewed target cells [167]. Another recent randomized, multicenter, in [73]). placebo-controlled clinical trial in advanced NSCLC Targeting xenogenic TACA for cancer immunotherapy patients indicated a clinically relevant improvement in appears as an attractive option, including tumor-associated overall survival and progression-free survival for racotu- sialoglycans that aberrantly express Neu5Gc (discussed momab-alu-treated patients in comparison with the placebo above). group [168]. Noteworthy, the capacity of induced serum The ‘non-human’ glycolipid NGcGM3 was reported in IgM anti-NGcGM3 antibodies to bind or destroy at least the 1990s to be present in human breast tumors while it 30 % of antigen positive murine leukemia L1210 cells was remained undetectable in normal breast tissue [61]. Sub- clearly associated with longer overall survival [168]. These sequently, several other studies have been conducted to studies provide compelling evidence that points to address its expression in other types of malignancies [62, NGcGM3 as a clinically meaningful therapeutic target 162–164]. [159, 169]. Recent studies extend or reaffirm the presence of As an alternative approach, a cancer vaccine has been NGcGM3 ganglioside in NSCLC, digestive system developed by combining chemically unmodified NGcGM3 tumors, primary lymphoid tumors, pediatric nervous ganglioside with the hydrophobic outer membrane protein system tumors, sarcomas, thyroid carcinomas, and oral complex derived from Neisseria meningitides [170]. Very cavity melanomas [62, 162–165]. Noteworthy a sig- small-sized proteoliposomes (VSSP) with nanometric nificantly elevated expression of this target ranged diameters were obtained and significant humoral immune from 33 % of studied cases in esophagus cancer up to responses in mice and monkeys against the highly tolerated 100 % of stomach, large intestine and pancreatic GM3 and NGcGM3 were induced by vaccination [170]. A tumors [163]. controlled Phase II clinical trial of the NGcGM3/VSSP As previously stated, one of the more striking features of vaccine was conducted in 79 patients with metastatic breast this glycolipid is its tumor-restricted character. While cancer [171]. In the intent to treat analysis, there was a almost 50 % of NSCLC samples are highly positive for trend toward a survival advantage for the vaccine group NGcGM3, no Neu5Gc-containing ganglioside expression and this effect was significant for patients with non-visceral was detected in normal lung tissues [165]. Blanco et al. metastasis [171]. Similarly patients with metastatic cuta- [164] extended the same finding to normal tissues of skin, neous melanoma were included in a Phase I/II clinical trial esophagus, stomach, large intestine, liver, pancreas, testis, and treated with a new formulation of the NGcGM3/VSSP prostate, kidney, urinary bladder, brain, cerebellum, spinal vaccine by the subcutaneous route [172]. Objective cord, peripheral nerves, thymus, tonsils, lymph nodes, responses or stable disease was observed in 38.46 % of spleen, heart, arteries, veins, muscles, pituitary and thyroid patients with a global median overall survival of glands. Only a weak reaction of a specific anti- NGcGM3 20.20 months [172]. monoclonal antibody (14F7 mAb) with mucous cells from While these results are rather encouraging, a more firm small intestine samples was observed in one-third of the clinical validation of this particular glycan NGcGM3 as a cases [164]. In addition, ten oral melanocytic nevi evi- target of cancer immunotherapy for different malignancies, denced no reactivity with 14F7 mAb [62]. may be expected soon from ongoing Phase III clinical Racotumomab (formerly known as 1E10; VaxiraÒ)isan studies evaluating Vaxira and NGcGM3/VSSP vaccines (to anti-idiotypic antibody obtained from mice immunized be published). with the murine IgM P3 (coupled to KLH in adjuvant), specifically reacting with the tumor-associated ganglioside NGcGM3 and other Neu5Gc-containing gangliosides. Conclusions Anti-idiotypic racotumomab acts as a molecular mimicry of these parental gangliosides. More than a decade after the Cells exhibit differential glycosylation patterns that are racotumomab was first described by Vazquez et al. [166], differentiation and cell type-specific [5], but is subject to recent clinical trials show promising results of a vaccine significant modification under distinct physiological or formulation of aluminium hydroxide-precipitated racotu- pathological conditions [7, 23]. Glycosylation changes, as momab. In a study of advanced non-small cell lung cancer a consequence of epigenetic or genetic events, are a (NSCLC), the racotumomab vaccine was shown to be hallmark of tumor development. Since the first demon- immunogenic in 16 out of 20 NSCLC patients, raising stration of altered glycosylation in transformed cells by detectable serum levels of specific IgM and IgG antibody Meezan et al. in 1969 [173], many tumor-associated car- responses against NGcGM3 ganglioside. Hyperimmune bohydrate antigens have been diagnostically utilized as patient sera were able to induce complement-independent tumor markers. Although the functional consequences of 123 Tumor glycosylation and sialylation codes 1243 altered tumor glycosylation remain to be further explored, 9. Fuster MM, Esko JD (2005) The sweet and sour of cancer: several promising therapeutic strategies targeting glycans, glycans as novel therapeutic targets. Nat Rev Cancer 5(7):526–542. doi:10.1038/nrc1649 glycan-binding proteins, or protein–glycan interactions 10. Gabius HJ (2000) Biological information transfer beyond the have been proposed or are under clinical evaluation and genetic code: the sugar code. Naturwissenschaften 87(3): have been extensively reviewed elsewhere [9, 19]. Glycan- 108–121 based therapeutics include, among others, tumor vaccines 11. Stowell SR, Arthur CM, McBride R, Berger O, Razi N, Heim- burg-Molinaro J, Rodrigues LC, Gourdine JP, Noll AJ, von based on carbohydrate-mimetic peptides [174] or anti- Gunten S, Smith DF, Knirel YA, Paulson JC, Cummings RD idiotypic antibodies (e.g. racotumomab), natural or syn- (2014) Microbial glycan microarrays define key features of host- thetic glycans to block glycan-binding proteins or glycan- microbial interactions. Nat Chem Biol 10(6):470–476. doi:10. modifying enzymes [9, 19, 92, 175], as well as antibodies 1038/nchembio.1525 12. von Gunten S, Smith DF, Cummings RD, Riedel S, Miescher S, or glycan-coated nanoparticles for cell-directed therapy Schaub A, Hamilton RG, Bochner BS (2009) Intravenous [105, 135]. immunoglobulin contains a broad repertoire of anticarbohydrate Evidence suggests that tumor cells acquire in a antibodies that is not restricted to the IgG2 subclass. J Allergy ‘microevolutionary’ process [9, 27], a glycosylation profile Clin Immunol 123(6):1268–1276. doi:10.1016/j.jaci.2009.03. 013 that provides them advantages in terms of tumor growth, 13. Houzelstein D, Goncalves IR, Fadden AJ, Sidhu SS, Cooper dissemination and immune escape. A better understanding DN, Drickamer K, Leffler H, Poirier F (2004) Phylogenetic of the immunomodulatory glycosylation signals and codes analysis of the vertebrate galectin family. Mol Biol Evol that are sent out by tumor cells, and received and deci- 21(7):1177–1187. doi:10.1093/molbev/msh082 14. Liu FT, Rabinovich GA (2005) Galectins as modulators of phered by detection systems of different arms of the tumour progression. Nat Rev Cancer 5(1):29–41. doi:10.1038/ immune system, may lead to novel and improved thera- nrc1527 peutic strategies, not only for cancer, but also for 15. Varki A, Schauer R (2009) Sialic acids. In: Varki A, Cummings autoimmune disease and transplantation. RD, Esko JD et al (eds) Essentials of glycobiology, 2nd edn. Cold Spring Harbor, New York 16. Varki A (2011) Since there are PAMPs and DAMPs, there must Acknowledgments Research by SVG is supported by the Swiss be SAMPs? Glycan ‘‘self-associated molecular patterns’’ dam- National Science Foundation (Grant No. 310030_135734), the Bul- pen innate immunity, but pathogens can mimic them. garian-Swiss Research Program (BSRP) No. IZEBZO_142967, Swiss Glycobiology 21(9):1121–1124 Cancer Research (KFS-3248-08-2013) and CSL Behring AG, Bern, 17. Macauley MS, Paulson JC (2014) Immunology: glyco-engi- Switzerland. The authors thank Aldona von Gunten for full illustra- neering ‘super-self’. Nat Chem Biol 10(1):7–8. doi:10.1038/ tion of Fig. 3 and assistance with Fig. 2. nchembio.1415 18. Hakomori S (1984) Tumor-associated carbohydrate antigens. Annu Rev Immunol 2:103–126. doi:10.1146/annurev.iy.02. 040184.000535 References 19. Dube DH, Bertozzi CR (2005) Glycans in cancer and inflammation–potential for therapeutics and diagnostics. Nat Rev Drug 1. Hart GW, Copeland RJ (2010) Glycomics hits the big time. Cell Discov 4(6):477–488. doi:10.1038/nrd1751 143(5):672–676. doi:10.1016/j.cell.2010.11.008 20. Hollingsworth MA, Swanson BJ (2004) Mucins in cancer: 2. Schachter H, Freeze HH (2009) Glycosylation diseases: quo protection and control of the cell surface. Nat Rev Cancer vadis? Biochim Biophys Acta 1792(9):925–930. doi:10.1016/j. 4(1):45–60. doi:10.1038/nrc1251 bbadis.2008.11.002 21. Krengel U, Bousquet PA (2014) Molecular recognition of gan- 3. Varki A (2006) Nothing in glycobiology makes sense, except in gliosides and their potential for cancer immunotherapies. Front the light of evolution. Cell 126(5):841–845. doi:10.1016/j.cell. Immunol 5:325. doi:10.3389/fimmu.2014.00325 2006.08.022 22. Kufe DW (2009) Mucins in cancer: function, prognosis and 4. Gagneux P, Varki A (1999) Evolutionary considerations in therapy. Nat Rev Cancer 9(12):874–885. doi:10.1038/nrc2761 relating oligosaccharide diversity to biological function. Gly- 23. Kannagi R, Sakuma K, Miyazaki K, Lim KT, Yusa A, Yin J, cobiology 9(8):747–755 Izawa M (2010) Altered expression of glycan genes in cancers 5. North SJ, von Gunten S, Antonopoulos A, Trollope A, Mac- induced by epigenetic silencing and tumor hypoxia: clues in the Glashan DW Jr, Jang-Lee J, Dell A, Metcalfe DD, Kirshenbaum ongoing search for new tumor markers. Cancer Sci AS, Bochner BS, Haslam SM (2012) Glycomic analysis of 101(3):586–593. doi:10.1111/j.1349-7006.2009.01455.x human mast cells, eosinophils and basophils. Glycobiology 24. Stanley P, Schachter H, Taniguchi N (2009) N-glycans. In: 22(1):12–22. doi:10.1093/glycob/cwr089 Varki, Cummings RD, Esko JD et al (eds) Essentials of glyco- 6. Stanley P, Cummings RD (2009) Structures common to differ- biology, 2nd edn. Cold Spring Harbor, New York ent glycans. In: Varki A, Cummings RD, Esko JD et al (eds) 25. Brockhausen I, Schachter H, Stanley P (2009) O-GalNAc gly- Essentials of glycobiology, 2nd edn. Cold Spring Harbor, New cans. In: Varki A, Cummings RD, Esko JD et al (eds) Essentials York of glycobiology, 2nd edn. Cold Spring Harbor, New York 7. Ohtsubo K, Marth JD (2006) Glycosylation in cellular mecha- 26. Cohen M, Varki A (2010) The sialome—far more than the sum nisms of health and disease. Cell 126(5):855–867. doi:10.1016/j. of its parts. OMICS 14(4):455–464. doi:10.1089/omi.2009.0148 cell.2006.08.019 27. Varki A, Kannagi R, Toole BP (2009) Glycosylation changes in 8. Haüselmann I, Borsig L (2014) Altered tumor-cell glycosylation cancer. In: Varki A, Cummings RD, Esko JD et al (eds) promotes metastasis. Front Oncol 4:28. doi:10.3389/fonc.2014. Essentials of Glycobiology, 2nd edn. Cold Spring Harbor, New 00028 York 123 1244 K. F. Boligan et al.

28. Dennis JW, Laferte S, Waghorne C, Breitman ML, Kerbel RS 43. Nicoll G, Avril T, Lock K, Furukawa K, Bovin N, Crocker PR (1987) Beta 1-6 branching of Asn-linked oligosaccharides is (2003) Ganglioside GD3 expression on target cells can modulate directly associated with metastasis. Science 236(4801):582–585 NK cell cytotoxicity via siglec-7-dependent and -independent 29. Granovsky M, Fata J, Pawling J, Muller WJ, Khokha R, Dennis mechanisms. Eur J Immunol 33(6):1642–1648. doi:10.1002/eji. JW (2000) Suppression of tumor growth and metastasis in 200323693 Mgat5-deficient mice. Nat Med 6(3):306–312. doi:10.1038/ 44. Nores GA, Dohi T, Taniguchi M, Hakomori S (1987) Density- 73163 dependent recognition of cell surface GM3 by a certain anti- 30. Yoshimura M, Nishikawa A, Ihara Y, Taniguchi S, Taniguchi N melanoma antibody, and GM3 lactone as a possible immunogen: (1995) Suppression of lung metastasis of B16 mouse melanoma requirements for tumor-associated antigen and immunogen. by N-acetylglucosaminyltransferase III gene transfection. Proc J Immunol 139(9):3171–3176 Natl Acad Sci USA 92(19):8754–8758 45. Miyazaki K, Ohmori K, Izawa M, Koike T, Kumamoto K, 31. Partridge EA, Le Roy C, Di Guglielmo GM, Pawling J, Cheung Furukawa K, Ando T, Kiso M, Yamaji T, Hashimoto Y, Suzuki P, Granovsky M, Nabi IR, Wrana JL, Dennis JW (2004) Reg- A, Yoshida A, Takeuchi M, Kannagi R (2004) Loss of disialyl ulation of cytokine receptors by Golgi N-glycan processing and Lewis(a), the ligand for lymphocyte inhibitory receptor sialic endocytosis. Science 306(5693):120–124. doi:10.1126/science. acid-binding immunoglobulin-like lectin-7 (Siglec-7) associated 1102109 with increased sialyl Lewis(a) expression on human colon 32. Zhao YY, Takahashi M, Gu JG, Miyoshi E, Matsumoto A, cancers. Cancer Res 64(13):4498–4505. doi:10.1158/0008-5472. Kitazume S, Taniguchi N (2008) Functional roles of N-glycans CAN-03-3614 in cell signaling and cell adhesion in cancer. Cancer Sci 46. Izawa M, Kumamoto K, Mitsuoka C, Kanamori C, Kanamori A, 99(7):1304–1310. doi:10.1111/j.1349-7006.2008.00839.x Ohmori K, Ishida H, Nakamura S, Kurata-Miura K, Sasaki K, 33. Taniguchi N, Miyoshi E, Ko JH, Ikeda Y, Ihara Y (1999) Nishi T, Kannagi R (2000) Expression of sialyl 6-sulfo Lewis X Implication of N-acetylglucosaminyltransferases III and V in is inversely correlated with conventional sialyl Lewis X cancer: gene regulation and signaling mechanism. Biochim expression in human colorectal cancer. Cancer Res 60(5): Biophys Acta 1455(2–3):287–300 1410–1416 34. Inamori K, Gu J, Ohira M, Kawasaki A, Nakamura Y, Nakagawa 47. Yusa A, Miyazaki K, Kimura N, Izawa M, Kannagi R (2010) T, Kondo A, Miyoshi E, Nakagawara A, Taniguchi N (2006) Epigenetic silencing of the sulfate transporter gene DTDST High expression of N-acetylglucosaminyltransferase V in induces sialyl Lewisx expression and accelerates proliferation of favorable neuroblastomas: involvement of its effect on apoptosis. colon cancer cells. Cancer Res 70(10):4064–4073. doi:10.1158/ FEBS Lett 580(2):627–632. doi:10.1016/j.febslet.2005.12.089 0008-5472.CAN-09-2383 35. Ihara S, Miyoshi E, Ko JH, Murata K, Nakahara S, Honke K, 48. Ju T, Lanneau GS, Gautam T, Wang Y, Xia B, Stowell SR, Dickson RB, Lin CY, Taniguchi N (2002) Prometastatic effect Willard MT, Wang W, Xia JY, Zuna RE, Laszik Z, Benbrook of N-acetylglucosaminyltransferase V is due to modification and DM, Hanigan MH, Cummings RD (2008) Human tumor anti- stabilization of active matriptase by adding beta 1-6 GlcNAc gens Tn and sialyl Tn arise from mutations in Cosmc. Cancer branching. J Biol Chem 277(19):16960–16967. doi:10.1074/jbc. Res 68(6):1636–1646. doi:10.1158/0008-5472.CAN-07-2345 M200673200 49. Carrascal MA, Severino PF, Guadalupe Cabral M, Silva M, 36. Ihara S, Miyoshi E, Nakahara S, Sakiyama H, Ihara H, Akinaga Ferreira JA, Calais F, Quinto H, Pen C, Ligeiro D, Santos LL, A, Honke K, Dickson RB, Lin CY, Taniguchi N (2004) Addition Dall’Olio F, Videira PA (2014) Sialyl Tn-expressing bladder of beta1-6 GlcNAc branching to the oligosaccharide attached to cancer cells induce a tolerogenic phenotype in innate and Asn 772 in the serine protease domain of matriptase plays a adaptive immune cells. Mol Oncol 8(3):753–765. doi:10.1016/j. pivotal role in its stability and resistance against trypsin. Gly- molonc.2014.02.008 cobiology 14(2):139–146. doi:10.1093/glycob/cwh013 50. Dabelsteen E, Gao S (2005) ABO blood-group antigens in oral 37. Rambaruth ND, Dwek MV (2011) Cell surface glycan-lectin cancer. J Dent Res 84(1):21–28 interactions in tumor metastasis. Acta Histochem 113(6): 51. Bianco T, Farmer BJ, Sage RE, Dobrovic A (2001) Loss of red 591–600. doi:10.1016/j.acthis.2011.03.001 cell A, B, and H antigens is frequent in myeloid malignancies. 38. Li D, Li Y, Wu X, Li Q, Yu J, Gen J, Zhang XL (2008) Blood 97(11):3633–3639 Knockdown of Mgat5 inhibits breast cancer cell growth with 52. Bu¨ll C, Stoel MA, den Brok MH, Adema GJ (2014) Sialic acids activation of CD4? T cells and macrophages. J Immunol sweeten a tumor’s life. Cancer Res 74(12):3199–3204. doi:10. 180(5):3158–3165 1158/0008-5472.CAN-14-0728 39. Taniguchi N, Ihara S, Saito T, Miyoshi E, Ikeda Y, Honke K (2001) 53. Almaraz RT, Tian Y, Bhattarcharya R, Tan E, Chen SH, Dallas Implication of GnT-V in cancer metastasis: a glycomic approach MR, Chen L, Zhang Z, Zhang H, Konstantopoulos K, Yarema KJ for identification of a target protein and its unique function as an (2012) Metabolic flux increases glycoprotein sialylation: impli- angiogenic cofactor. Glycoconj J 18(11–12):859–865 cations for cell adhesion and cancer metastasis. Mol Cell 40. Anugraham M, Jacob F, Nixdorf S, Everest-Dass AV, Hein- Proteomics 11(7):M112 017558. doi:10.1074/mcp.M112.017558 zelmann-Schwarz V, Packer NH (2014) Specific glycosylation 54. Miyagi T, Takahashi K, Hata K, Shiozaki K, Yamaguchi K of membrane proteins in epithelial ovarian cancer cell lines: (2012) Sialidase significance for cancer progression. Glycoconj glycan structures reflect gene expression and DNA methylation J 29(8–9):567–577. doi:10.1007/s10719-012-9394-1 status. Mol Cell Proteomics. doi:10.1074/mcp.M113.037085 55. Hudak JE, Canham SM, Bertozzi CR (2014) Glycocalyx engi- 41. Zhang X, Wang Y, Qian Y, Wu X, Zhang Z, Liu X, Zhao R, neering reveals a Siglec-based mechanism for NK cell Zhou L, Ruan Y, Xu J, Liu H, Ren S, Xu C, Gu J (2014) immunoevasion. Nat Chem Biol 10(1):69–75. doi:10.1038/ Discovery of specific metastasis-related N-glycan alterations in nchembio.1388 epithelial ovarian cancer based on quantitative glycomics. PLoS 56. Muchmore EA, Milewski M, Varki A, Diaz S (1989) Biosyn- ONE 9(2):e87978. doi:10.1371/journal.pone.0087978 thesis of N-glycolylneuraminic acid. The primary site of 42. Miwa HE, Song Y, Alvarez R, Cummings RD, Stanley P (2012) hydroxylation of N-acetylneuraminic acid is the cytosolic sugar The bisecting GlcNAc in cell growth control and tumor pro- nucleotide pool. J Biol Chem 264(34):20216–20223 gression. Glycoconj J 29(8–9):609–618. doi:10.1007/s10719- 57. Shaw L, Schauer R (1988) The biosynthesis of N-glycoloyl- 012-9373-6 neuraminic acid occurs by hydroxylation of the CMP-glycoside

123 Tumor glycosylation and sialylation codes 1245

of N-acetylneuraminic acid. Biol Chem Hoppe Seyler K, Chen X, Varki A, Schwab R (2011) Human xeno-autoanti- 369(6):477–486 bodies against a non-human sialic acid serve as novel serum 58. Shaw L, Schauer R (1989) Detection of CMP-N-acetylneuram- biomarkers and immunotherapeutics in cancer. Cancer Res inic acid hydroxylase activity in fractionated mouse liver. 71(9):3352–3363. doi:10.1158/0008-5472.CAN-10-4102 Biochem J 263(2):355–363 73. Samraj AN, Laubli H, Varki N, Varki A (2014) Involvement of 59. Malykh YN, Schauer R, Shaw L (2001) N-Glycolylneuraminic a non-human sialic Acid in human cancer. Front Oncol 4:33. acid in human tumours. Biochimie 83(7):623–634 doi:10.3389/fonc.2014.00033 60. Higashi H, Hirabayashi Y, Fukui Y, Naiki M, Matsumoto M, 74. Amon R, Reuven EM, Leviatan Ben-Arye S, Padler-Karavani V Ueda S, Kato S (1985) Characterization of N-glycolylneurami- (2014) Glycans in immune recognition and response. Carbohydr nic acid-containing gangliosides as tumor-associated Res 389:115–122. doi:10.1016/j.carres.2014.02.004 Hanganutziu-Deicher antigen in human colon cancer. Cancer 75. Pearce OM, Laubli H, Verhagen A, Secrest P, Zhang J, Varki Res 45(8):3796–3802 NM, Crocker PR, Bui JD, Varki A (2014) Inverse hormesis of 61. Marquina G, Waki H, Fernandez LE, Kon K, Carr A, Valiente cancer growth mediated by narrow ranges of tumor-directed O, Perez R, Ando S (1996) Gangliosides expressed in human antibodies. Proc Natl Acad Sci USA 111(16):5998–6003. breast cancer. Cancer Res 56(22):5165–5171 doi:10.1073/pnas.1209067111 62. Zhong Y, Wu Y, Li C, Tang J, Wang X, Ren G, Carr A, Perez R, 76. Hedlund M, Padler-Karavani V, Varki NM, Varki A (2008) Guo W (2012) N-Glycolyl GM3 ganglioside immunoexpression Evidence for a human-specific mechanism for diet and antibody- in oral mucosal melanomas of Chinese. Oral Dis 18(8):741–747. mediated inflammation in carcinoma progression. Proc Natl doi:10.1111/j.1601-0825.2012.01939.x Acad Sci USA 105(48):18936–18941. doi:10.1073/pnas. 63. Higashi H, Sasabe T, Fukui Y, Maru M, Kato S (1988) Detec- 0803943105 tion of gangliosides as N-glycolylneuraminic acid-specific 77. Marth JD, Grewal PK (2008) Mammalian glycosylation in tumor-associated Hanganutziu-Deicher antigen in human reti- immunity. Nat Rev Immunol 8(11):874–887. doi:10.1038/ noblastoma cells. Jpn J Cancer Res 79(8):952–956 nri2417 64. Banda K, Gregg CJ, Chow R, Varki NM, Varki A (2012) 78. van Kooyk Y, Rabinovich GA (2008) Protein-glycan interac- Metabolism of vertebrate amino sugars with N-glycolyl groups: tions in the control of innate and adaptive immune responses. mechanisms underlying gastrointestinal incorporation of the Nat Immunol 9(6):593–601. doi:10.1038/ni.f.203 non-human sialic acid xeno-autoantigen N-glycolylneuraminic 79. von Gunten S, Shoenfeld Y, Blank M, Branch DR, Vassilev T, acid. J Biol Chem 287(34):28852–28864. doi:10.1074/jbc. Ka¨sermann F, Bayry J, Kaveri S, Simon HU (2014) IVIG M112.364182 pluripotency and the concept of Fc-sialylation: challenges to 65. Bardor M, Nguyen DH, Diaz S, Varki A (2005) Mechanism of the scientist. Nat Rev Immunol 14(5):349. doi:10.1038/ uptake and incorporation of the non-human sialic acid N-glyc- nri3401-c1 olylneuraminic acid into human cells. J Biol Chem 80. Varki A (2007) Glycan-based interactions involving vertebrate 280(6):4228–4237. doi:10.1074/jbc.M412040200 sialic-acid-recognizing proteins. Nature 446(7139):1023–1029. 66. Tangvoranuntakul P, Gagneux P, Diaz S, Bardor M, Varki N, doi:10.1038/nature05816 Varki A, Muchmore E (2003) Human uptake and incorporation 81. Rabinovich GA, Croci DO (2012) Regulatory circuits mediated of an immunogenic nonhuman dietary sialic acid. Proc Natl by lectin-glycan interactions in autoimmunity and cancer. Acad Sci USA 100(21):12045–12050. doi:10.1073/pnas. Immunity 36(3):322–335. doi:10.1016/j.immuni.2012.03.004 2131556100 82. Geijtenbeek TB, Gringhuis SI (2009) Signalling through C-type 67. Hedlund M, Tangvoranuntakul P, Takematsu H, Long JM, lectin receptors: shaping immune responses. Nat Rev Immunol Housley GD, Kozutsumi Y, Suzuki A, Wynshaw-Boris A, Ryan 9(7):465–479. doi:10.1038/nri2569 AF, Gallo RL, Varki N, Varki A (2007) N-glycolylneuraminic 83. von Gunten S, Bochner BS (2008) Basic and clinical immu- acid deficiency in mice: implications for human biology and nology of Siglecs. Ann N Y Acad Sci 1143:61–82. doi:10.1196/ evolution. Mol Cell Biol 27(12):4340–4346. doi:10.1128/MCB. annals.1443.011 00379-07 84. Zhuo Y, Bellis SL (2011) Emerging role of alpha2,6-sialic acid 68. Yin J, Miyazaki K, Shaner RL, Merrill AH Jr, Kannagi R (2010) as a negative regulator of galectin binding and function. J Biol Altered sphingolipid metabolism induced by tumor hypoxia— Chem 286(8):5935–5941. doi:10.1074/jbc.R110.191429 new vistas in glycolipid tumor markers. FEBS Lett 85. Jandus C, Boligan KF, Chijioke O, Liu H, Dahlhaus M, Dem- 584(9):1872–1878. doi:10.1016/j.febslet.2009.11.019 oulins T, Schneider C, Wehrli M, Hunger RE, Baerlocher GM, 69. Yin J, Hashimoto A, Izawa M, Miyazaki K, Chen GY, Tak- Simon HU, Romero P, Munz C, von Gunten S (2014) Interac- ematsu H, Kozutsumi Y, Suzuki A, Furuhata K, Cheng FL, Lin tions between Siglec-7/9 receptors and ligands influence NK CH, Sato C, Kitajima K, Kannagi R (2006) Hypoxic culture cell-dependent tumor immunosurveillance. J Clin Invest induces expression of sialin, a sialic acid transporter, and can- 124(4):1810–1820. doi:10.1172/JCI65899 cer-associated gangliosides containing non-human sialic acid on 86. Kawasaki Y, Ito A, Withers DA, Taima T, Kakoi N, Saito S, human cancer cells. Cancer Res 66(6):2937–2945. doi:10.1158/ Arai Y (2010) Ganglioside DSGb5, preferred ligand for Siglec- 0008-5472.CAN-05-2615 7, inhibits NK cell cytotoxicity against renal cell carcinoma 70. Zhu A, Hurst R (2002) Anti-N-glycolylneuraminic acid anti- cells. Glycobiology 20(11):1373–1379. doi:10.1093/glycob/ bodies identified in healthy human serum. Xenotransplantation cwq116 9(6):376–381 87. Cohen M, Elkabets M, Perlmutter M, Porgador A, Voronov E, 71. Padler-Karavani V, Yu H, Cao H, Chokhawala H, Karp F, Varki Apte RN, Lichtenstein RG (2010) Sialylation of 3-methylcho- N, Chen X, Varki A (2008) Diversity in specificity, abundance, lanthrene-induced fibrosarcoma determines antitumor immune and composition of anti-Neu5Gc antibodies in normal humans: responses during immunoediting. J Immunol 185(10): potential implications for disease. Glycobiology 18(10): 5869–5878. doi:10.4049/jimmunol.1001635 818–830. doi:10.1093/glycob/cwn072 88. Tsuboi S, Sutoh M, Hatakeyama S, Hiraoka N, Habuchi T, 72. Padler-Karavani V, Hurtado-Ziola N, Pu M, Yu H, Huang S, Horikawa Y, Hashimoto Y, Yoneyama T, Mori K, Koie T, Muthana S, Chokhawala HA, Cao H, Secrest P, Friedmann- Nakamura T, Saitoh H, Yamaya K, Funyu T, Fukuda M, Ohy- Morvinski D, Singer O, Ghaderi D, Verma IM, Liu YT, Messer ama C (2011) A novel strategy for evasion of NK cell immunity

123 1246 K. F. Boligan et al.

by tumours expressing core2 O-glycans. EMBO J 30(15): 104. Crocker PR, Redelinghuys P (2008) Siglecs as positive and 3173–3185. doi:10.1038/emboj.2011.215 negative regulators of the immune system. Biochem Soc Trans 89. Kim YJ, Borsig L, Varki NM, Varki A (1998) P-selectin defi- 36(Pt 6):1467–1471. doi:10.1042/BST0361467 ciency attenuates tumor growth and metastasis. Proc Natl Acad 105. Jandus C, Simon HU, von Gunten S (2011) Targeting siglecs—a Sci USA 95(16):9325–9330 novel pharmacological strategy for immuno- and glycotherapy. 90. Stone JP, Wagner DD (1993) P-selectin mediates adhesion of Biochem Pharmacol 82(4):323–332. doi:10.1016/j.bcp.2011.05. platelets to neuroblastoma and small cell lung cancer. J Clin 018 Invest 92(2):804–813. doi:10.1172/JCI116654 106. Cao H, Crocker PR (2011) Evolution of CD33-related siglecs: 91. Garcia J, Callewaert N, Borsig L (2007) P-selectin mediates regulating host immune functions and escaping pathogen metastatic progression through binding to sulfatides on tumor exploitation? Immunology 132(1):18–26. doi:10.1111/j.1365- cells. Glycobiology 17(2):185–196. doi:10.1093/glycob/cwl059 2567.2010.03368.x 92. Borsig L, Wong R, Feramisco J, Nadeau DR, Varki NM, Varki 107. Macauley MS, Crocker PR, Paulson JC (2014) Siglec-mediated A (2001) Heparin and cancer revisited: mechanistic connections regulation of immune cell function in disease. Nat Rev Immunol involving platelets, P-selectin, carcinoma mucins, and tumor 14(10):653–666. doi:10.1038/nri3737 metastasis. Proc Natl Acad Sci USA 98(6):3352–3357. doi:10. 108. von Gunten S, Simon HU (2006) Sialic acid binding immuno- 1073/pnas.061615598 globulin-like lectins may regulate innate immune responses by 93. Nieswandt B, Hafner M, Echtenacher B, Mannel DN (1999) modulating the life span of granulocytes. FASEB J 20(6): Lysis of tumor cells by natural killer cells in mice is impeded by 601–605. doi:10.1096/fj.05-5401hyp platelets. Cancer Res 59(6):1295–1300 109. Crocker PR, McMillan SJ, Richards HE (2012) CD33-related sig- 94. Palumbo JS, Talmage KE, Massari JV, La Jeunesse CM, Flick lecs as potential modulators of inflammatory responses. Ann N Y MJ, Kombrinck KW, Jirouskova M, Degen JL (2005) Platelets Acad Sci 1253:102–111. doi:10.1111/j.1749-6632.2011.06449.x and fibrin(ogen) increase metastatic potential by impeding nat- 110. Nutku E, Aizawa H, Hudson SA, Bochner BS (2003) Ligation of ural killer cell-mediated elimination of tumor cells. Blood Siglec-8: a selective mechanism for induction of human eosin- 105(1):178–185. doi:10.1182/blood-2004-06-2272 ophil apoptosis. Blood 101(12):5014–5020. doi:10.1182/blood- 95. Kopp HG, Placke T, Salih HR (2009) Platelet-derived trans- 2002-10-3058 forming growth factor-beta down-regulates NKG2D thereby 111. von Gunten S, Yousefi S, Seitz M, Jakob SM, Schaffner T, Seger inhibiting natural killer cell antitumor reactivity. Cancer Res R, Takala J, Villiger PM, Simon HU (2005) Siglec-9 transduces 69(19):7775–7783. doi:10.1158/0008-5472.CAN-09-2123 apoptotic and nonapoptotic death signals into neutrophils 96. Crespo HJ, Lau JT, Videira PA (2013) Dendritic cells: a spot on depending on the proinflammatory cytokine environment. Blood sialic acid. Front Immunol 4:491. doi:10.3389/fimmu.2013. 106(4):1423–1431. doi:10.1182/blood-2004-10-4112 00491 112. Mitsuki M, Nara K, Yamaji T, Enomoto A, Kanno M, Yam- 97. Napoletano C, Rughetti A, Agervig Tarp MP, Coleman J, aguchi Y, Yamada A, Waguri S, Hashimoto Y (2010) Siglec-7 Bennett EP, Picco G, Sale P, Denda-Nagai K, Irimura T, Mandel mediates nonapoptotic cell death independently of its immuno- U, Clausen H, Frati L, Taylor-Papadimitriou J, Burchell J, Nuti receptor tyrosine-based inhibitory motifs in monocytic cell M (2007) Tumor-associated Tn-MUC1 glycoform is internal- line U937. Glycobiology 20(3):395–402. doi:10.1093/glycob/ ized through the macrophage galactose-type C-type lectin and cwp195 delivered to the HLA class I and II compartments in dendritic 113. von Gunten S, Jakob SM, Geering B, Takala J, Simon HU cells. Cancer Res 67(17):8358–8367. doi:10.1158/0008-5472. (2009) Different patterns of Siglec-9-mediated neutrophil death CAN-07-1035 responses in septic shock. Shock 32(4):386–392. doi:10.1097/ 98. Singh SK, Streng-Ouwehand I, Litjens M, Kalay H, Saeland E, SHK.0b013e3181a1bc98 van Kooyk Y (2011) Tumour-associated glycan modifications of 114. von Gunten S, Simon HU (2007) Autophagic-like cell death in antigen enhance MGL2 dependent uptake and MHC class I neutrophils induced by autoantibodies. Autophagy 3(1):67–68 restricted CD8 T cell responses. Int J Cancer 128(6):1371–1383. 115. von Gunten S, Schaub A, Vogel M, Stadler BM, Miescher S, doi:10.1002/ijc.25458 Simon HU (2006) Immunologic and functional evidence for 99. Nonaka M, Ma BY, Murai R, Nakamura N, Baba M, Kawasaki anti-Siglec-9 autoantibodies in intravenous immunoglobulin N, Hodohara K, Asano S, Kawasaki T (2008) Glycosylation- preparations. Blood 108(13):4255–4259. doi:10.1182/blood- dependent interactions of C-type lectin DC-SIGN with colo- 2006-05-021568 rectal tumor-associated Lewis glycans impair the function and 116. von Gunten S, Vogel M, Schaub A, Stadler BM, Miescher S, differentiation of monocyte-derived dendritic cells. J Immunol Crocker PR, Simon HU (2007) Intravenous immunoglobulin 180(5):3347–3356 preparations contain anti-Siglec-8 autoantibodies. J Allergy Clin 100. Fedarko NS, Fohr B, Robey PG, Young MF, Fisher LW (2000) Immunol 119(4):1005–1011. doi:10.1016/j.jaci.2007.01.023 Factor H binding to bone sialoprotein and osteopontin enables 117. Schaub A, von Gunten S, Vogel M, Wymann S, Ruegsegger M, tumor cell evasion of complement-mediated attack. J Biol Chem Stadler BM, Spycher M, Simon HU, Miescher S (2011) Dimeric 275(22):16666–16672. doi:10.1074/jbc.M001123200 IVIG contains natural anti-Siglec-9 autoantibodies and their 101. Shi WX, Chammas R, Varki NM, Powell L, Varki A (1996) anti-idiotypes. Allergy 66(8):1030–1037. doi:10.1111/j.1398- Sialic acid 9-O-acetylation on murine erythroleukemia cells 9995.2011.02579.x affects complement activation, binding to I-type lectins, and 118. von Gunten S, Simon HU (2012) Granulocyte death regulation tissue homing. J Biol Chem 271(49):31526–31532 by naturally occurring autoantibodies. Adv Exp Med Biol 102. Ajona D, Castano Z, Garayoa M, Zudaire E, Pajares MJ, Mar- 750:157–172. doi:10.1007/978-1-4614-3461-0_12 tinez A, Cuttitta F, Montuenga LM, Pio R (2004) Expression of 119. von Gunten S, Simon HU (2010) Cell death modulation by complement factor H by lung cancer cells: effects on the acti- intravenous immunoglobulin. J Clin Immunol 30(Suppl 1):S24– vation of the alternative pathway of complement. Cancer Res S30. doi:10.1007/s10875-010-9411-8 64(17):6310–6318. doi:10.1158/0008-5472.CAN-03-2328 120. von Gunten S, Wehrli M, Simon HU (2013) Cell death in 103. Crocker PR, Paulson JC, Varki A (2007) Siglecs and their roles immune thrombocytopenia: novel insights and perspectives. in the immune system. Nat Rev Immunol 7(4):255–266. doi:10. Semin Hematol 50(Suppl 1):S109–S115. doi:10.1053/j. 1038/nri2056 seminhematol.2013.03.016

123 Tumor glycosylation and sialylation codes 1247

121. Chen GY, Tang J, Zheng P, Liu Y (2009) CD24 and Siglec-10 134. Kiwamoto T, Kawasaki N, Paulson JC, Bochner BS (2012) selectively repress tissue damage-induced immune responses. Siglec-8 as a drugable target to treat eosinophil and mast cell- Science 323(5922):1722–1725. doi:10.1126/science.1168988 associated conditions. Pharmacol Ther 135(3):327–336. doi:10. 122. Laübli H, Pearce OM, Schwarz F, Siddiqui SS, Deng L, 1016/j.pharmthera.2012.06.005 Stanczak MA, Deng L, Verhagen A, Secrest P, Lusk C, Sch- 135. O’Reilly MK, Paulson JC (2009) Siglecs as targets for therapy wartz AG, Varki NM, Bui JD, Varki A (2014) Engagement of in immune-cell-mediated disease. Trends Pharmacol Sci myelomonocytic Siglecs by tumor-associated ligands modulates 30(5):240–248. doi:10.1016/j.tips.2009.02.005 the innate immune response to cancer. Proc Natl Acad Sci USA. 136. Chen WC, Sigal DS, Saven A, Paulson JC (2012) Targeting B doi:10.1073/pnas.1409580111 lymphoma with nanoparticles bearing glycan ligands of CD22. 123. Miyazaki K, Sakuma K, Kawamura YI, Izawa M, Ohmori K, Leuk Lymphoma 53(2):208–210. doi:10.3109/10428194.2011. Mitsuki M, Yamaji T, Hashimoto Y, Suzuki A, Saito Y, Dohi T, 604755 Kannagi R (2012) Colonic epithelial cells express specific 137. Laübli H, Borsig L (2010) Selectins promote tumor metastasis. ligands for mucosal macrophage immunosuppressive receptors Semin Cancer Biol 20(3):169–177. doi:10.1016/j.semcancer. siglec-7 and -9. J Immunol 188(9):4690–4700. doi:10.4049/ 2010.04.005 jimmunol.1100605 138. Girard JP, Moussion C, Forster R (2012) HEVs, lymphatics and 124. Tomioka Y, Morimatsu M, Nishijima K, Usui T, Yamamoto S, homeostatic immune cell trafficking in lymph nodes. Nat Rev Suyama H, Ozaki K, Ito T, Ono E (2014) A soluble form of Immunol 12(11):762–773. doi:10.1038/nri3298 Siglec-9 provides an antitumor benefit against mammary tumor 139. Blagoveshchenskaya AD, Norcott JP, Cutler DF (1998) Lyso- cells expressing MUC1 in transgenic mice. Biochem Biophys somal targeting of P-selectin is mediated by a novel sequence Res Commun 450(1):532–537. doi:10.1016/j.bbrc.2014.06.009 within its cytoplasmic tail. J Biol Chem 273(5):2729–2737 125. Ohta M, Ishida A, Toda M, Akita K, Inoue M, Yamashita K, 140. Sanada H, Midorikawa S, Yatabe J, Yatabe MS, Katoh T, Baba Watanabe M, Murata T, Usui T, Nakada H (2010) Immuno- T, Hashimoto S, Watanabe T (2005) Elevation of serum soluble modulation of monocyte-derived dendritic cells through ligation E- and P-selectin in patients with hypertension is reversed by of tumor-produced mucins to Siglec-9. Biochem Biophys Res benidipine, a long-acting calcium channel blocker. Hypertens Commun 402(4):663–669. doi:10.1016/j.bbrc.2010.10.079 Res 28(11):871–878. doi:10.1291/hypres.28.871 126. Toda M, Akita K, Inoue M, Taketani S, Nakada H (2008) Down- 141. Kisucka J, Chauhan AK, Zhao BQ, Patten IS, Yesilaltay A, modulation of B cell signal transduction by ligation of mucins to Krieger M, Wagner DD (2009) Elevated levels of soluble CD22. Biochem Biophys Res Commun 372(1):45–50. doi:10. P-selectin in mice alter blood–brain barrier function, exacerbate 1016/j.bbrc.2008.04.175 stroke, and promote atherosclerosis. Blood 113(23):6015–6022. 127. Toda M, Hisano R, Yurugi H, Akita K, Maruyama K, Inoue M, doi:10.1182/blood-2008-10-186650 Adachi T, Tsubata T, Nakada H (2009) Ligation of tumour-pro- 142. Ay C, Simanek R, Vormittag R, Dunkler D, Alguel G, Koder S, duced mucins to CD22 dramatically impairs splenic marginal zone Kornek G, Marosi C, Wagner O, Zielinski C, Pabinger I (2008) B-cells. Biochem J 417(3):673–683. doi:10.1042/BJ20081241 High plasma levels of soluble P-selectin are predictive of venous 128. Mitic N, Milutinovic B, Jankovic M (2012) Assessment of sialic thromboembolism in cancer patients: results from the Vienna acid diversity in cancer- and non-cancer related CA125 antigen Cancer and Thrombosis Study (CATS). Blood 112(7):2703–2708. using sialic acid-binding Ig-like lectins (Siglecs). Dis Markers doi:10.1182/blood-2008-02-142422 32(3):187–194. doi:10.3233/DMA-2011-0872 143. Kawashima H, Fukuda M (2012) Sulfated glycans control 129. Belisle JA, Horibata S, Jennifer GA, Petrie S, Kapur A, Andre S, lymphocyte homing. Ann N Y Acad Sci 1253:112–121. doi:10. Gabius HJ, Rancourt C, Connor J, Paulson JC, Patankar MS 1111/j.1749-6632.2011.06356.x (2010) Identification of Siglec-9 as the receptor for MUC16 on 144. Krause T, Turner GA (1999) Are selectins involved in metas- human NK cells, B cells, and monocytes. Mol Cancer 9:118. tasis? Clin Exp Metastasis 17(3):183–192 doi:10.1186/1476-4598-9-118 145. Hoos A, Protsyuk D, Borsig L (2014) Metastatic growth pro- 130. Sabit I, Hashimoto N, Matsumoto Y, Yamaji T, Furukawa K, gression caused by PSGL-1-mediated recruitment of monocytes Furukawa K (2013) Binding of a sialic acid-recognizing lectin to metastatic sites. Cancer Res 74(3):695–704. doi:10.1158/ Siglec-9 modulates adhesion dynamics of cancer cells via cal- 0008-5472.CAN-13-0946 pain-mediated protein degradation. J Biol Chem 288(49): 146. Borsig L, Wong R, Hynes RO, Varki NM, Varki A (2002) 35417–35427. doi:10.1074/jbc.M113.513192 Synergistic effects of L- and P-selectin in facilitating tumor 131. Tanida S, Akita K, Ishida A, Mori Y, Toda M, Inoue M, Ohta M, metastasis can involve non-mucin ligands and implicate leuko- Yashiro M, Sawada T, Hirakawa K, Nakada H (2013) Binding cytes as enhancers of metastasis. Proc Natl Acad Sci USA of the sialic acid-binding lectin, Siglec-9, to the membrane 99(4):2193–2198. doi:10.1073/pnas.261704098 mucin, MUC1, induces recruitment of beta-catenin and sub- 147. Heidemann F, Schildt A, Schmid K, Bruns OT, Riecken K, Jung sequent cell growth. J Biol Chem 288(44):31842–31852. doi:10. C, Ittrich H, Wicklein D, Reimer R, Fehse B, Heeren J, Luers G, 1074/jbc.M113.471318 Schumacher U, Heine M (2014) Selectins mediate small cell 132. Aalto K, Autio A, Kiss EA, Elima K, Nymalm Y, Veres TZ, lung cancer systemic metastasis. PLoS ONE 9(4):e92327. Marttila-Ichihara F, Elovaara H, Saanijoki T, Crocker PR, doi:10.1371/journal.pone.0092327 Maksimow M, Bligt E, Salminen TA, Salmi M, Roivainen A, 148. Laübli H, Stevenson JL, Varki A, Varki NM, Borsig L (2006) Jalkanen S (2011) Siglec-9 is a novel leukocyte ligand for L-selectin facilitation of metastasis involves temporal induction vascular adhesion protein-1 and can be used in PET imaging of of Fut7-dependent ligands at sites of tumor cell arrest. Cancer inflammation and cancer. Blood 118(13):3725–3733. doi:10. Res 66(3):1536–1542. doi:10.1158/0008-5472.CAN-05-3121 1182/blood-2010-09-311076 149. Prinetti A, Aureli M, Illuzzi G, Prioni S, Nocco V, Scandroglio 133. Kivi E, Elima K, Aalto K, Nymalm Y, Auvinen K, Koivunen E, F, Gagliano N, Tredici G, Rodriguez-Menendez V, Chigorno V, Otto DM, Crocker PR, Salminen TA, Salmi M, Jalkanen S Sonnino S (2010) GM3 synthase overexpression results in (2009) Human Siglec-10 can bind to vascular adhesion protein-1 reduced cell motility and in caveolin-1 upregulation in human and serves as its substrate. Blood 114(26):5385–5392. doi:10. ovarian carcinoma cells. Glycobiology 20(1):62–77. doi:10. 1182/blood-2009-04-219253 1093/glycob/cwp143

123 1248 K. F. Boligan et al.

150. Prinetti A, Cao T, Illuzzi G, Prioni S, Aureli M, Gagliano N, digestive system. ISRN Gastroenterol 2011:645641. doi:10. Tredici G, Rodriguez-Menendez V, Chigorno V, Sonnino S 5402/2011/645641 (2011) A glycosphingolipid/caveolin-1 signaling complex 164. Blanco RQ, Cedenõ M, Rengifo CE, Fro´meta M, Rıós M, inhibits motility of human ovarian carcinoma cells. J Biol Chem Rengifo E, Carr A (2013) Tissue reactivity of the 14F7 Mab 286(47):40900–40910. doi:10.1074/jbc.M111.286146 raised against N-glycolyl GM3 ganglioside in tumors of neuro- 151. Tringali C, Lupo B, Cirillo F, Papini N, Anastasia L, Lamorte G, ectodermal, mesodermal, and epithelial origin. J Biomarkers Colombi P, Bresciani R, Monti E, Tettamanti G, Venerando B 2013:9. doi:10.1155/2013/602417 (2009) Silencing of membrane-associated sialidase Neu3 165. Hayashi N, Chiba H, Kuronuma K, Go S, Hasegawa Y, Ta- diminishes apoptosis resistance and triggers megakaryocytic kahashi M, Gasa S, Watanabe A, Hasegawa T, Kuroki Y, differentiation of chronic myeloid leukemic cells K562 through Inokuchi J, Takahashi H (2013) Detection of N-glycosylated the increase of ganglioside GM3. Cell Death Differ gangliosides in non-small-cell lung cancer using GMR8 mono- 16(1):164–174. doi:10.1038/cdd.2008.141 clonal antibody. Cancer Sci 104(1):43–47. doi:10.1111/cas. 152. Handa K, Hakomori SI (2012) Carbohydrate to carbohydrate 12027 interaction in development process and cancer progression. 166. Vazquez AM, Perez A, Hernandez AM, Macias A, Alfonso M, Glycoconj J 29(8–9):627–637. doi:10.1007/s10719-012-9380-7 Bombino G, Perez R (1998) Syngeneic anti-idiotypic mono- 153. Potapenko M, Shurin GV, de Leon J (2007) Gangliosides as clonal antibodies to an anti-NeuGc-containing ganglioside immunomodulators. Adv Exp Med Biol 601:195–203 monoclonal antibody. Hybridoma 17(6):527–534 154. Varki A, Freeze HH (2009) Glycans in acquired human diseases. 167. Hernandez AM, Toledo D, Martinez D, Grinan T, Brito V, In: Varki A, Cummings RD, Esko JD et al (eds) Essentials of Macias A, Alfonso S, Rondon T, Suarez E, Vazquez AM, Perez glycobiology, 2nd edn. Cold Spring Harbor, New York R (2008) Characterization of the antibody response against 155. Casadesus AV, Fernandez-Marrero Y, Clavell M, Gomez JA, NeuGcGM3 ganglioside elicited in non-small cell lung cancer Hernandez T, Moreno E, Lopez-Requena A (2013) A shift from patients immunized with an anti-idiotype antibody. J Immunol N-glycolyl- to N-acetyl-sialic acid in the GM3 ganglioside 181(9):6625–6634 impairs tumor development in mouse lymphocytic leukemia 168. Alfonso S, Valdes-Zayas A, Santiesteban ER, Flores YI, Areces cells. Glycoconj J 30(7):687–699. doi:10.1007/s10719-013- F, Hernandez M, Viada C, Mendoza IC, Guerra PP, Garcia E, 9473-y Ortiz RA, de la Torre AV, Cepeda M, Perez K, Chong E, 156. Gabri MR, Otero LL, Gomez DE, Alonso DF (2009) Exogenous Hernandez AM, Toledo D, Gonzalez Z, Mazorra Z, Crombet T, incorporation of Neugc-rich mucin augments N-glycolyl sialic Perez R, Vazquez AM, Macias AE (2014) A randomized, acid content and promotes malignant phenotype in mouse tumor multicenter, placebo-controlled clinical trial of racotumomab- cell lines. J Exp Clin Cancer Res 28:146. doi:10.1186/1756- alum vaccine as switch maintenance therapy in advanced non- 9966-28-146 small-cell-lung cancer patients. Clin Cancer Res. doi:10.1158/ 157. Segatori VI, Vazquez AM, Gomez DE, Gabri MR, Alonso DF 1078-0432.CCR-13-1674 (2012) Preclinical evaluation of racotumomab, an anti-idiotype 169. Fernandez LE, Gabri MR, Guthmann MD, Gomez RE, Gold S, monoclonal antibody to N-glycolyl-containing gangliosides, Fainboim L, Gomez DE, Alonso DF (2010) NGcGM3 gangli- with or without chemotherapy in a mouse model of non-small oside: a privileged target for cancer vaccines. Clin Dev Immunol cell lung cancer. Front Oncol 2:160. doi:10.3389/fonc.2012. 2010:814397. doi:10.1155/2010/814397 00160 170. Estevez F, Carr A, Solorzano L, Valiente O, Mesa C, Barroso O, 158. Segatori VI, Otero LL, Fernandez LE, Gomez DE, Alonso DF, Sierra GV, Fernandez LE (1999) Enhancement of the immune Gabri MR (2012) Antitumor protection by NGcGM3/VSSP response to poorly immunogenic gangliosides after incorpora- vaccine against transfected B16 mouse melanoma cells over- tion into very small size proteoliposomes (VSSP). Vaccine expressing N-glycolylated gangliosides. Vivo 26(4):609–617 18(1–2):190–197 159. Labrada M, Clavell M, Bebelagua Y, Leon J, Alonso DF, Gabri 171. Mulens V, de la Torre A, Marinello P, Rodriguez R, Cardoso J, MR, Veloso RC, Verez V, Fernandez LE (2010) Direct vali- Diaz R, O’Farrill M, Macias A, Viada C, Saurez G, Carr A, dation of NGcGM3 ganglioside as a new target for cancer Crombet T, Mazorra Z, Perez R, Fernandez LE (2010) Immu- immunotherapy. Expert Opin Biol Ther 10(2):153–162. doi:10. nogenicity and safety of a NeuGcGM3 based cancer vaccine: 1517/14712590903443084 results from a controlled study in metastatic breast cancer 160. de Leon J, Fernandez A, Clavell M, Labrada M, Bebelagua Y, patients. Hum Vaccin 6(9). doi:10.4161.hv.6.9.12571 Mesa C, Fernandez LE (2008) Differential influence of the 172. Osorio M, Gracia E, Reigosa E, Hernandez J, de la Torre A, tumour-specific non-human sialic acid containing GM3 gangli- Saurez G, Perez K, Viada C, Cepeda M, Carr A, Avila Y, oside on CD4? CD25- effector and naturally occurring CD4? Rodriguez M, Fernandez LE (2012) Effect of vaccination with CD25? regulatory T cells function. Int Immunol 20(4): N-glycolyl GM3/VSSP vaccine by subcutaneous injection in 591–600. doi:10.1093/intimm/dxn018 patients with advanced cutaneous melanoma. Cancer Manag Res 161. de Leon J, Fernandez A, Mesa C, Clavel M, Fernandez LE 4:341–345. doi:10.2147/CMAR.S22617 (2006) Role of tumour-associated N-glycolylated variant of 173. Meezan E, Wu HC, Black PH, Robbins PW (1969) Comparative GM3 ganglioside in cancer progression: effect over CD4 studies on the carbohydrate-containing membrane components expression on T cells. Cancer Immunol Immunother 55(4): of normal and virus-transformed mouse fibroblasts. II. Separa- 443–450. doi:10.1007/s00262-005-0041-6 tion of glycoproteins and glycopeptides by sephadex 162. Blanco R, Blanco D, Quintana Y, Escobar X, Rengifo CE, chromatography. Biochemistry 8(6):2518–2524 Osorio M, Gutierrez Z, Lamadrid J, Cedeno M, Frometa M, Carr 174. Kieber-Emmons T, Saha S, Pashov A, Monzavi-Karbassi B, A, Rengifo E (2013) Immunoreactivity of the 14F7 Mab raised Murali R (2014) Carbohydrate-mimetic peptides for pan anti- against N-Glycolyl GM3 ganglioside in primary lymphoid tumor responses. Front Immunol 5:308. doi:10.3389/fimmu. tumors and lymph node metastasis. Pathol Res Int 2013:920972. 2014.00308 doi:10.1155/2013/920972 175. Shirota K, Kato Y, Irimura T, Kondo H, Sugiyama Y (2001) 163. Blanco R, Rengifo E, Cedeno M, Rengifo CE, Alonso DF, Carr Anti-metastatic effect of the sialyl Lewis-X analog GSC-150 on A (2011) Immunoreactivity of the 14F7 Mab raised against N- the human colon carcinoma derived cell line KM12-HX in the glycolyl GM3 ganglioside in epithelial malignant tumors from mouse. Biol Pharm Bull 24(3):316–319

123

Oncology meets immunology: The Cancer Immunity Cycle

Immunity Review

Oncology Meets Immunology: The Cancer-Immunity Cycle

Daniel S. Chen1,3 and Ira Mellman2,3,* 1Stanford Medical Oncology, Stanford University School of Medicine, Stanford, CA 94305, USA 2Department of Biochemistry & Biophysics, University of California, San Francisco School of Medicine, San Francisco, CA 94143, USA 3Genentech, 1 DNA Way, South San Francisco, CA 94080, USA *Correspondence: [email protected] http://dx.doi.org/10.1016/j.immuni.2013.07.012

The genetic and cellular alterations that define cancer provide the immune system with the means to generate T cell responses that recognize and eradicate cancer cells. However, elimination of cancer by T cells is only one step in the Cancer-Immunity Cycle, which manages the delicate balance between the recognition of nonself and the prevention of autoimmunity. Identification of cancer cell T cell inhibitory signals, including PD-L1, has prompted the development of a new class of cancer immunotherapy that specifically hinders immune effector inhibition, reinvigorating and potentially expanding preexisting anticancer immune responses. The presence of suppressive factors in the tumor microenvironment may explain the limited activity observed with previous immune-based therapies and why these therapies may be more effective in combination with agents that target other steps of the cycle. Emerging clinical data suggest that cancer immunotherapy is likely to become a key part of the clinical management of cancer.

Introduction 2013). Factors in the tumor microenvironment can act to modu- The development of cancer immunotherapy has reached an late the existing activated antitumor T cell immune response, important inflection point in the history of cancer therapy acting as an immune rheostat or ‘‘immunostat.’’ This class of (reviewed in Mellman et al., 2011). Durable monotherapy re- molecules, including PD-L1:PD-1 (reviewed in Chen et al., sponses are consistently being reported for a broad range of 2012; Topalian et al., 2012a), emphasizes that the immune human cancers with several different agents (Hamid et al., response in cancer reflects a series of carefully regulated events 2013a; Herbst et al., 2013; Hodi et al., 2010; Topalian et al., that may be optimally addressed not singly but as a group. The 2012b), providing a compelling argument that cancer immuno- challenge now is to use this new understanding to develop therapy is active in a range of indications beyond melanoma, a new drugs and implement clinical strategies. disease often thought to be atypically immunogenic (Jacobs The articles contained in this issue each address key aspects et al., 2012). In addition to encouraging activity, many of the can- of how the immune response can control or be manipulated to cer immunotherapy approaches report safety profiles that are enhance anticancer immunity (Galon et al., 2013; Kalos and milder and more manageable than traditional or targeted (i.e., June, 2013; Motz and Coukos, 2013; Palucka and Banchereau, oncogene-centric) cancer therapies. 2013; van den Boorn and Hartmann, 2013; Zitvogel et al., Cancer is characterized by the accumulation of a variable 2013). Here, we will integrate this information and consider number of genetic alterations and the loss of normal cellular reg- how it might best be used in clinical development. ulatory processes (Tian et al., 2011). These events have long The Cancer-Immunity Cycle been known to result in the expression of neoantigens, differen- For an anticancer immune response to lead to effective killing of tiation antigens, or cancer testis antigens, which can lead to pre- cancer cells, a series of stepwise events must be initiated and sentation of peptides bound to major histocompatibility class I allowed to proceed and expand iteratively. We refer to these (MHCI) molecules on the surface of cancer cells, distinguishing steps as the Cancer-Immunity Cycle (Figure 1). In the first step, them from their normal counterparts. Since the work of Boon neoantigens created by oncogenesis are released and captured and colleagues, we have known that these cancer-specific pep- by dendritic cells (DCs) for processing (step 1). In order for this tide-MHCI complexes can be recognized by CD8+ T cells pro- step to yield an anticancer T cell response, it must be accompa- duced spontaneously in cancer patients (Boon et al., 1994). nied by signals that specify immunity lest peripheral tolerance to However, even when T cell responses occurred, they rarely pro- the tumor antigens be induced. Such immunogenic signals might vided protective immunity nor could they be mobilized to provide include proinflammatory cytokines and factors released by dying a basis for therapy. tumor cells or by the gut microbiota (Figure 2, Table 1). Next, DCs As demonstrated by elegant analyses of cancer in mice, the present the captured antigens on MHCI and MHCII molecules to continued deletion of cancer cells expressing T cell targets T cells (step 2), resulting in the priming and activation of effector (immune editing) may enable cancers to evolve to avoid attack T cell responses against the cancer-specific antigens (step 3) (Dunn et al., 2002). Despite these findings, recent results from that are viewed as foreign or against which central tolerance human cancer have demonstrated that overcoming negative has been incomplete. The nature of the immune response is regulators to T cell responses in lymphoid organs (checkpoints) determined at this stage, with a critical balance representing and in the tumor bed (immunostat function) are likely to explain the ratio of T effector cells versus T regulatory cells being key the failure of immune protection in many patients (Mullard, to the final outcome. Finally, the activated effector T cells traffic

Immunity 39, July 25, 2013 ª2013 Elsevier Inc. 1 Immunity Review

Figure 1. The Cancer-Immunity Cycle The generation of immunity to cancer is a cyclic process that can be self propagating, leading to an accumulation of immune-stimulatory factors that in principle should amplify and broaden T cell responses. The cycle is also characterized by inhibitory factors that lead to immune regulatory feedback mechanisms, which can halt the development or limit the immunity. This cycle can be divided into seven major steps, starting with the release of antigens from the cancer cell and ending with the killing of cancer cells. Each step is described above, with the primary cell types involved and the anatomic location of the activity listed. Ab- breviations are as follows: APCs, antigen presenting cells; CTLs, cytotoxic T lymphocytes. to (step 4) and infiltrate the tumor bed (step 5), specifically recog- dampen or arrest the antitumor immune response, the most nize and bind to cancer cells through the interaction between its effective approaches will involve selectively targeting the rate- T cell receptor (TCR) and its cognate antigen bound to MHCI limiting step in any given patient. Amplifying the entire cycle (step 6), and kill their target cancer cell (step 7). Killing of the can- may provide anticancer activity but at the potential cost of cer cell releases additional tumor-associated antigens (step 1 unwanted damage to normal cells and tissues. Many recent clin- again) to increase the breadth and depth of the response in sub- ical results suggest that a common rate-limiting step is the im- sequent revolutions of the cycle. In cancer patients, the Cancer- munostat function, immunosuppression that occurs in the tumor Immunity Cycle does not perform optimally. Tumor antigens may microenvironment (Predina et al., 2013; Wang et al., 2013). not be detected, DCs and T cells may treat antigens as self rather Initiating Anticancer Immunity: Antigen Release, than foreign thereby creating T regulatory cell responses rather Presentation, and T Cell Priming than effector responses, T cells may not properly home to Attempts to activate or introduce cancer antigen-specific T cells, tumors, may be inhibited from infiltrating the tumor, or (most as well as stimulate the proliferation of these cells over the last 20 importantly) factors in the tumor microenvironment might sup- years, have led to mostly no, minimal or modest appreciable press those effector cells that are produced (reviewed by Motz anticancer immune responses. The majority of these efforts and Coukos, 2013). involved the use of therapeutic vaccines because vaccines can The goal of cancer immunotherapy is to initiate or reinitiate a be easy to deploy and have historically represented an approach self-sustaining cycle of cancer immunity, enabling it to amplify that has brought enormous medical benefit (reviewed by Pal- and propagate, but not so much as to generate unrestrained ucka and Banchereau, 2013). Yet, cancer vaccines were limited autoimmune inflammatory responses. Cancer immunotherapies on two accounts. First, until recently, there was a general lack of must therefore be carefully configured to overcome the negative understanding of how to immunize human patients to achieve feedback mechanisms. Although checkpoints and inhibitors are potent cytotoxic T cell responses. This limitation reflects built into each step that oppose continued amplification and can continued uncertainties concerning the identities of antigens to

2 Immunity 39, July 25, 2013 ª2013 Elsevier Inc. Immunity Review

TNF-α

IFN-α

IFN-γ

TGF-β

Figure 2. Stimulatory and Inhibitory Factors in the Cancer-Immunity Cycle Each step of the Cancer-Immunity Cycle requires the coordination of numerous factors, both stimulatory and inhibitory in nature. Stimulatory factors shown in green promote immunity, whereas inhibitors shown in red help keep the process in check and reduce immune activity and/or prevent autoimmunity. Immune checkpoint proteins, such as CTLA4, can inhibit the development of an active immune response by acting primarily at the level of T cell development and proliferation (step 3). We distinguish these from immune rheostat (‘‘immunostat’’) factors, such as PD-L1, can have an inhibitory function that primarily acts to modulate active immune responses in the tumor bed (step 7). Examples of such factors and the primary steps at which they can act are shown. Abbreviations are as follows: IL, interleukin; TNF, tumor necrosis factor; IFN, interferon; CDN, cyclic dinucleotide; ATP, adenosine triphosphate; HMGB1, high-mobility group protein B1; TLR, Toll-like receptor; HVEM, herpes virus entry mediator; GITR, glucocorticoid-induced TNFR family-related gene; CTLA4, cytotoxic T-lympocyte antigen-4; PD-L1, programmed death-ligand 1; CXCL/CCL, chemokine motif ligands; LFA1, lymphocyte function-associated antigen-1; ICAM1, intracellular adhesion molecule 1; VEGF, vascular endothelial growth factor; IDO, indoleamine 2,3-dioxygenase; TGF, transforming growth factor; BTLA, B- and T-lymphocyte attenuator; VISTA, V-domain Ig suppressor of T cell activation; LAG-3, lymphocyte-activation gene 3 protein; MIC, MHC class I polypeptide-related sequence protein; TIM-3, T cell immunoglobulin domain and mucin domain-3. Although not illustrated, it is important to note that intratumoral T regulatory cells, macrophages, and myeloid-derived suppressor cells are key sources of many of these inhibitory factors. See text and Table 1 for details. use, their mode of delivery, the types of adjuvants required, and generate sufficiently robust T cell responses in all patients. More- the proximal characteristics of the desired T cell response over, a single antigenic target, especially one not derived from a (Palucka and Banchereau, 2013). Second, the presence of the protein that is an inherent oncogenic driver, seems more likely to immunostat in the tumor microenvironment may dampen or enable resistance by antigenic drift (immune editing) than a disable antitumor immune responses before clinically relevant multivalent vaccine (Palucka and Banchereau, 2013). tumor kill can occur. Thus, as long as these negative signals Deciding how to configure multivalent vaccines is itself a daunt- are in place, the prospects for vaccine-based approaches ing challenge. It may be insufficient to rely entirely on sequencing used alone are likely to be limited. the expressed tumor genome looking for point mutations, trans- Although vaccination can accelerate the anticancer immunity location fusions, or C-T antigens. Not only might this vary from in the context of treatments that suppress negative regulators patient to patient or even from cell to cell within a single patient’s (Palucka and Banchereau, 2013), a number of significant chal- tumor, expression at the messenger RNA or protein level does not lenges need to be overcome. First is the identification of the assure that predicted antigenic peptides will be generated and appropriate tumor antigens to include in any vaccine. A large, expressed as peptide-MHCI complexes, especially in the face monovalent antigen trial (using the C-T antigen MAGE-A3) is of the allelic complexity in the MHC. Several groups are actively currently under way (Kruit et al., 2013; Vansteenkiste et al., approaching this problem by using a combination of informatics 2013), yet it is not clear that any one candidate will necessarily and mass spectroscopy of peptides eluted from MHCI molecules

Immunity 39, July 25, 2013 ª2013 Elsevier Inc. 3 Immunity Review

Table 1. Cancer-Immunity Cycle: Examples of Positive and Negative Regulators at Each Step Steps (+) Stimulators (À) Inhibitors Other Considerations Example References 1. Release of cancer Immunogenic or necrotic Tolergenic or apoptotic Tumor-associated Ferguson et al., 2011 antigens cell death cell death neoantigens and cancer testis antigens 2. Cancer antigen d Proinflammatory cytokines IL-10, IL-4, IL-13 Dendritic cell maturity Lippitz, 2013; presentation (e.g., TNF-a, IL1, IFN-a) Mellman et al., 2011 d Immune cell factors: CD40L/CD40 d Endogenous adjuvants released from dying tumors: CDN (STING ligand), ATP, HMGB1 d Gut microbiome products: TLR ligands 3. Priming and activation CD28:B7.1, CD137 (4-1BB)/ CTLA4:B7.1, PD-L1:PD-1, Central tolerance, T cell Franciszkiewicz et al., 2012; CD137L, OX40:OX40L, PD-L1:B7.1, prostaglandins repertoire, T regulatory Lippitz, 2013; CD27:CD70, HVEM, GITR, cells Riella et al., 2012; IL-2, IL-12 So et al., 2006 4. Trafficking of T cells CX3CL1, CXCL9, CXCL10, Franciszkiewicz et al., 2012; to tumors CCL5 Peng et al., 2012 5. Infiltration of T cells LFA1:ICAM1, selectins VEGF, endothelin B receptor Franciszkiewicz et al., 2013 into tumors 6. Recognition of cancer T cell receptor Reduced peptide-MHC Mellman et al., 2011 cells by T cells expression on cancer cells 7. Killing of cancer cells IFN-g, T cell granule PD-L1:PD-1, PD-L1:B7.1, T regulatory cells, Chen et al., 2012; content TIM-3:phospholipids, BTLA, myeloid-derived Greaves and Gribben, 2013; VISTA, LAG-3, IDO, Arginase, suppressor cells, M2 Mellman et al., 2011; MICA:MICB, B7-H4, TGFb macrophages, hypoxia Topalian et al., 2012a on both cell lines and primary tumors (Kasuga, 2013; Rammensee Work on vaccines should continue in a systematic fashion with et al., 2002; Segal et al., 2008). In principle, this information can be human studies, because animal models are unlikely to validate used to guide the formulation of multivalent vaccines, although it the best path forward (Davis, 2012). It is also unlikely that the does not necessarily address the problem of how to identify MHC best vaccine approaches will differentiate themselves any time class II epitopes that may be required to provide CD4 T cell help soon, given the lack of direct comparisons in clinical studies. that might be needed to produce protective CD8 responses. The This represents a substantial logistical challenge to incorpo- use of intact proteins as an immunogen may help mitigate this rating vaccination as part of a drug development plan. Not only issue. Moreover, it has thus far proved difficult to identify MHCI- are such trials long and expensive, but they also represent only bound peptides that harbor point mutations that could selectively one of many potential combinations that are competing to be target T cell responses to cancer cells, which is unfortunate given evaluated in conjunction with other immunotherapies (Vanne- that targeting somatic mutations should reduce the chances of man and Dranoff, 2012). generating autoimmunity or the need to overcome central toler- Therapeutic vaccination is not the only approach to acceler- ance (Mellman et al., 2011). Even assuming the correct antigens ating and expanding the production of T cell immunity. Because are in hand, how best to deliver them to patients remains a critical anticancer T cells can be produced spontaneously, there is a unknown. Peptides in emulsified vehicles represent a common growing appreciation that the tumor itself represents a type of and accessible approach, but in the absence of compelling pos- endogenous vaccine. Accessing the naturally occurring source itive controls for any vaccine, it is impossible to know whether it is of cancer-associated antigens avoids problems associated an effective approach. Other strategies include direct targeting to with selection and delivery (Heo et al., 2013; van den Boorn DCs, adoptive transfer of antigen-loaded DCs or tumor cells, re- and Hartmann, 2013). This approach is also convenient, but combinant viral vectors, and bacterial vectors (especially Listeria; achieving it requires detailed knowledge around whether stan- reviewed in Kalos and June, 2013; Palucka and Banchereau, dard of care chemotherapy or targeted therapies are compatible 2013). Work must continue evaluating each of these looking for with immunotherapies. Some therapies are thought to cause pharmacodynamic read-outs of CD8 T cell responses. With the tumor cell death in a fashion that promotes immunity (reviewed clinical success of anti-PD-L1 and anti-PD-1 antibodies, it should in Zitvogel et al., 2013). However, it is unclear whether this effect now be possible to evaluate different vaccines, adjuvants, and can be accurately predicted and will, in any event, require empir- delivery approaches in combination and therefore under condi- ical study. Chemotherapy, radiation therapy, and targeted ther- tions that should enhance the ability to judge their relative apies must also be evaluated for their effects on the immune efficacies with common clinical read-outs, such as partial or system. Although it is assumed that many might be antagonistic, complete responses in established tumors. there are some reports that others might promote T cell activity

4 Immunity 39, July 25, 2013 ª2013 Elsevier Inc. Immunity Review

Table 2. Inhibitors of PD-L1 or PD-1 Currently Being Developed in Oncology Interaction

Therapeutic Lead Company Antibody Type Afﬁnity/Kd Inhibited Development Anti-PD-L1 MPDL3280A Herbst et al., 2013. Genentech/Roche Engineered IgG1 0.4 nM PD-L1:PD-1 Broad (lung pivotal) (no ADCC) PD-L1:B7.1 MEDI-4736 Stewart et al., 2011. AstraZeneca Modiﬁed IgG1 Not available PD-L1:PD-1 Phase I (no ADCC) PD-L1:B7.1 Anti-PD-1 Nivolumab Brahmer et al., 2010. Bristol-Myers IgG4 2.6 nM PD-L1:PD-1 Broad (lung, melanoma, Squibb PD-L2:PD-1 RCC pivotal) Lambrolizumab Patnaik et al., 2012. Merck & Co IgG4 (humanized) 29 pM PD-L1:PD-1 Broad (melanoma pivotal) PD-L2:PD-1 Pidilizumab Rotem-Yehudar et al., 2009; CureTech IgG1 (humanized) Not available Broad Westin et al., 2012. AMP-224 Smothers et al., 2013. GlaxoSmithKline PD-L2 IgG1 Fc Not available PD-L1:PD-1 Phase I fusion PD-L2:PD-1

(Demaria et al., 2005; Duraiswamy et al., 2013; Hiniker et al., T cell expansion combined with the fundamental importance of 2012; Ott et al., 2013; Postow et al., 2012; Stagg et al., 2011; Zit- CTLA4 as a checkpoint may underlie the significant immune- vogel et al., 2013). related toxicities seen in patients treated with ipilimumab (Hodi Bypassing Vaccination by Adoptive T Cell Therapy et al., 2010). Another exciting development is that the initial demonstrations Nevertheless, the ability of this ‘‘lever’’ to create durable clin- that genetically modified autologous T cells could be reinfused ical responses in some patients has triggered a great deal of into patients to yield substantial clinical benefit, at least in certain effort to seek other immune modulators; modulators that can B cell malignancies (Grupp et al., 2013; reviewed in Kalos and achieve what ipilimumab can, but in a more selective and June, 2013). The most well developed of these is the use of controllable fashion that will provide the potential for greater ef- ‘‘CARs,’’ or chimeric antigen receptors, in which a patient’s ficacy and frequency of response, with less autoimmune-related T cells are transfected with a construct encoding an antibody toxicity. In addition, the combination of ipilimumab with agents against a tumor surface antigen (typically CD19) fused to T cell that modulate complimentary steps on the Cancer-Immunity Cy- signaling domains (Kochenderfer and Rosenberg, 2013). Similar cle are already underway (Karan and Van Veldhuizen, 2012; Ma- approaches are under investigation with recombinant T cell dan et al., 2012), and preliminary results from combinations that receptors (reviewed in Kalos and June, 2013). The procedure inhibit tumor immunosuppression appear very promising in avoids the need for immunization and may even overcome enhancing both antitumor immune responses and autoimmune mechanisms of immune suppression by overwhelming the sys- toxicity (see below). tem through infusion of large quantities of the modified T cells. Immunostat Blockade: PD-L1 and PD-1 This can force the revolution of the Cancer-Immunity Cycle, The identification of PD-L1 as a distal immune modulator enhancing the accumulation of stimulatory immune factors, expressed in 20%–50% of human cancer (Herbst et al., 2013) and potentially promotes eventual self-propagation of the cycle. has led to the development of a number of cancer immunother- The potential limitations here, which are yet to be fully deter- apies that target the interactions between PD-L1:PD-1, PD- mined, include whether the approach can be extended to can- L1:B7.1, and PD-L2:PD-1 (Table 2; reviewed in Chen et al., cers beyond hematologic malignancies, whether the delivery of 2012; Topalian et al., 2012a). Anti-PD-L1 and anti-PD-1 mono- large numbers of monospecific T cells will cause resistance therapy response rates have been presented for over 750 pa- due to antigenic drift, and whether the toxicity issues already tients (ranging from 13% to 38%) treated across a broad range identified can be safely managed. of human cancer types. Agents tested as monotherapy include T Cell Priming and Activation MPDL3280A (anti-PDL1; Genentech/Roche; Cho et al., 2013; Whether tumor antigens are delivered exogenously or are Hamid et al., 2013b; Herbst et al., 2013; Powderly et al., 2013; captured and presented by DCs endogenously, another strategy Spigel et al., 2013; Tabernero et al., 2013), nivolumab (anti-PD- for intervening in the Cancer-Immunity Cycle involves the control 1; Bristol Myers Squibb; Brahmer et al., 2013; Drake et al., of T cell activation. This is the presumed primary mechanism of 2013; Sznol et al., 2013; Topalian et al., 2013), and lambrolizu- action of anti-CTLA4 antibodies, such as ipilimumab, which mab (anti-PD-1; Merck; Hamid et al., 2013a). Antitumor activity blocks the interaction of the major negative regulator of T cells of the PD-L1- and PD-1-targeted therapies that utilize an engi- (CTLA4) with its ligands B7.1 and B7.2 (CD80 and CD86; Qureshi neered immunoglobulin G1 (IgG1) (MPDL3280A; modified to et al., 2011). Thus, during antigen presentation in lymphoid eliminate ADCC by altering FcgR binding; Herbst et al., 2013) organs (or in the periphery), the expansion of T cell responses or IgG4 antibody (nivolumab and lambrolizumab; expected to is disinhibited, thereby promoting the production of autoreactive reduce ADCC; Isaacs et al., 1996) backbone have resulted in T cells, including tumor-specific T cells. The lack of selectivity in particularly rapid antitumor activity, with some responses

Immunity 39, July 25, 2013 ª2013 Elsevier Inc. 5 Immunity Review

observed within days of starting treatment. These data suggest proaches should not be necessary. Furthermore, the potential that, for many human cancers, the Cancer-Immunity Cycle is for biomarker-driven patient selection to optimize treatment intact up to the point of tumor cell killing by T cells, which can benefit appears promising and might distinguish patients most be potently restrained by PD-L1. Once the PD-L1:PD-1 interac- likely to have strong benefit from the inhibition of PD-L1:PD-1 tion is blocked, preexisting anticancer T cells can have their as monotherapy opposed to those that may most likely require effector function rapidly restored. This is consistent with the pro- a different or combinatorial approach (Powderly et al., 2013; posed mechanism of action of this negative regulator, where PD- Topalian et al., 2013). These results also emphasize the likely L1 (expressed either on tumor cells or on tumor-infiltrating im- importance of immunosuppression in the natural history of can- mune cells) binding to PD-1 on activated effector T cells causes cer. Unfortunately, as the clinical trial data to date confirm, the the recruitment of the phosphatase SHP-2 and subsequent inac- majority of patients will not respond or will respond only incom- tivation of the PI3 kinase-signaling cascade (Chemnitz et al., pletely to PD-L1 or PD-1 inhibitors. Because multiple other 2004; Parry et al., 2005). These events block the secretion or pro- mechanisms of immunosuppression are known that may work duction of cytotoxic mediators required for killing. However, this together or in parallel with PD-L1:PD-1-mediated inhibition, block appears to be rapidly reversible once the inhibition is lifted. there is a need to pursue other potential agents that exhibit the Most importantly, the PD-L1 and PD-1 antagonists have same profile of rapid, substantial responses. For example, exhibited significant response rates, and largely unprecedented many tumors are characterized by significant infiltration by T reg- durable responses. In melanoma, the anti-PD-1 antibody nivolu- ulatory cells, and targeting these may prove to be a fruitful mab has reported an overall response rate (ORR) of 31% (33/ approach (Jacobs et al., 2012). It is possible that even ipilimu- 107) and a duration of response of 18.4 to 117.0+ weeks (Sznol mab works, at least in part, by causing T regulatory cell (Treg) et al., 2013), whereas lambrolizumab reported an ORR of 38% depletion. and duration of response of 1.9 to 10.8 months (Hamid et al., Combination Immune Therapies 2013a). Across a broad range of human cancers, which included It is reasonable to suspect that immunotherapy approaches, lung, colon, head and neck, and gastric cancers in addition to from vaccines to CARs, would be more effective when given in melanoma and renal cell carcinoma, the anti-PD-L1 antibody combination with a PD-L1 or PD-1 inhibitor (Goding et al., MPDL3280A had an ORR of 21% (29% in melanoma, 22% in 2013; West et al., 2013). By disabling the immune inhibition in lung cancer) with 26 of 29 responses ongoing at the time of the the tumor microenvironment, approaches that target earlier report (time from starting treatment ranged from 3 to 15+ steps in the Cancer-Immunity Cycle (steps 1–6) are more likely months) (Herbst et al., 2013; Hamid et al., 2013b; Spigel et al., to lead to cancer cell killing. Conversely, PD-L1 or PD-1 inhibition 2013). Additionally, the safety profile of these agents suggests may not be sufficient for optimal antitumor activity in some that while many cancer types express PD-L1 to inhibit anticancer patients, particularly those that do not demonstrate evidence immune responses, most patients do not have underlying auto- of tumor immune cell infiltration (Gajewski et al., 2011, 2013; Ga- immunity inhibited only by PD-L1 expression (Francisco et al., jewski, 2012). PD-L1- and PD-1-targeted therapies suggest that 2010). Grade 3-4 treatment-related adverse events were noted in patients whose tumors express PD-L1, the proximal steps of to occur in 13% to 21% of patients treated (Hamid et al., the Cancer-Immunity Cycle are intact and may not require further 2013a; Herbst et al., 2013; Sznol et al., 2013). The majority of re- enhancement. These patients are most commonly the patients ported cases have been readily manageable with supportive who exhibit rapid and durable response to PD-L1 or PD-1 inhibi- care or by immune suppression with steroid administration. tion. However, although some PD-L1-negative tumors still This is in stark contrast to the safety profiles of most therapies respond to PD-L1 or PD-1 monotherapy, the majority of tumors that target more proximal steps in the Cancer-Immunity Cycle do not (Powderly et al., 2013; Grosso et al., 2013). This outcome (e.g., anti-CTLA4; Hodi et al., 2010) and might hint at the benefits can be indicative of patients who have a defect in steps 1 to 6 of of specifically targeting the properties of cancer that inhibit the the Cancer-Immunity Cycle and may be most commonly a immune response rather than nonspecific activation of the im- defect in cancer antigen-specific T cell activation or infiltration mune system. Although it is still relatively early in the develop- of T cells into tumors (Powderly et al., 2013). However, more ment of these inhibitors (phase II/III trials are underway), the data from human tumors are likely to be necessary to further fact that three different antibodies have yielded such results elucidate what critical breaks in the cycle are most prominent greatly increases the confidence in a positive outcome. in different human cancers. From a drug development and clinical care perspective, the One approach, combining a CTLA4 targeted therapy (ipilimu- activity observed with anti-PD-L1 or anti-PD-1 is clear. Robust mab) with a PD-1-targeted inhibitor (nivolumab), appears to single-agent activity was observed rapidly and for extended du- enhance the immune activity in patients over either therapy alone rations without identified off-target toxicity (Topalian et al., 2013). in an early phase study (Wolchok et al., 2013). Anti-CTLA4 can This situation is distinct from the majority of other agents under lead to enhanced priming and activation of antigen-specific investigation (or approved) in oncology, except for a select group T cells and potentially clearance of regulatory T cells from the of small-molecule inhibitors that target driver oncogenic translo- tumor microenvironment (Table 1). The blocking of PD-L1 or cations or mutations (e.g., imatinib for BCR-Abl [Lin et al., 2013], PD-1 can remove the inhibition of cancer cell killing by T cells crizotinib for ALK translocations [Rothschild and Gautschi, (Figure 3). By inhibiting two targets that affect two steps in the 2013], vemurafenib [Huang et al., 2013], and dabrafenib [Huang Cancer-Immunity Cycle, rapid and deep responses (by modified et al., 2013] for the V600E BRAF mutation and erlotinib for mutant WHO criteria) were observed in patients with melanoma (ORR: EGF receptor [Bulgaru et al., 2003]). Therefore, extended trials 40% [21/52]; CR: 10% [5/52]). Immune-related toxicities were looking for incremental effects or complex combination ap- also enhanced in their magnitude, frequency, and onset (53%

6 Immunity 39, July 25, 2013 ª2013 Elsevier Inc. Immunity Review

IFN-α

Figure 3. Therapies that Might Affect the Cancer-Immunity Cycle The numerous factors that come into play in the Cancer-Immunity Cycle provide a wide range of potential therapeutic targets. This ﬁgure highlights examples of some of the therapies currently under preclinical or clinical evaluation. Key highlights include that vaccines can primarily promote cycle step 2, anti-CTLA4 can primarily promote cycle step 3, and anti-PD-L1 or anti-PD-1 antibodies can primarily promote cycle step 7. Although not developed as immunotherapies, chemotherapy, radiation therapy, and targeted therapies can primarily promote cycle step 1, and inhibitors of VEGF can potentially promote T cell inﬁltration into tumors—cycle step 5. Abbreviations are as follows: GM-CSF, granulocyte macrophage colony-stimulating factor; CARs, chimeric antigen receptors.

Grade 3-4 treatment-related toxicities). Although many of these class, inhibition of VEGF by the anti-VEGF antibody bevacizu- were serious and required treatment, therapy discontinuation, mab appears to be a promising candidate based on hints from or hospitalization, the durable partial and complete responses the preclinical and clinical literature (Motz and Coukos, 2013; in melanoma may warrant this approach in some patients. In Hodi et al., 2010). Similarly, B-Raf inhibitors (vemurafenib) may particular, combination therapy appeared to most dramatically also enhance T cell infiltration into tumors (Liu et al., 2013). Of benefit patients who were less likely to benefit from PD-L1 or course, with increased activity due to combinations comes the PD-1 inhibition alone, because their tumors were PD-L1-nega- increased chance for additive or synergistic toxicity. This further tive (6/13 PD-L1-positive and 9/22 PD-L1-negative patients re- highlights the importance of selecting therapeutic targets that sponded to combination therapy; Wolchok et al., 2013). The are specific to the ability of a tumor to escape immune eradica- addition of a CTLA4-targeted therapy may be completing the tion over targets that may also play an important role in mediating defect in the Cancer-Immunity Cycle for patients who are PD- immune homeostasis and preventing autoimmunity. L1-negative. Further studies of preipilimumab and on ipilimumab treatment tumor samples are warranted to better understand this Concluding Remarks effect. The objective of understanding the inherent immune biology Other combinations that merit serious consideration include related to cancer is to better define strategies to harness the anti-PD-L1 or anti-PD-1 with vaccination, especially if it be- human immune response against cancer to achieve durable re- comes possible to monitor a patient’s T cell profile to distinguish sponses and/or complete eradication of cancer in patients individuals who have generated suboptimal T cell responses to safely. Multiple approaches to cancer therapy exist, and few their cancers (Duraiswamy et al., 2013; Ge et al., 2013). In addi- are as complicated as immune-based therapy. Multiple numbers tion, combinations with agents that will enhance T cell trafficking of systemic factors can effect or contribute to the success or fail- and infiltration into the tumor bed should be explored vigorously, ure of immune therapy and lends to this complexity. Results may because the entry step may be important in some patients. In this be confounded by many currently unmeasured variables,

Immunity 39, July 25, 2013 ª2013 Elsevier Inc. 7 Immunity Review

including any given patient’s gut microbiome, their diet, and Corbie` re, V., Chapiro, J., Stroobant, V., Ma, W., Lurquin, C., Lethe´ , B., van whether they contract an underlying infection (Brestoff and Artis, Baren, N., Van den Eynde, B.J., Boon, T., and Coulie, P.G. (2011). Antigen spreading contributes to MAGE vaccination-induced regression of melanoma 2013; James et al., 2012, Rothman and Paterson, 2013; Xu et al., metastases. Cancer Res. 71, 1253–1262. 2012). Yet, as complicated and incompletely understood human Davis, M.M. (2012). Immunology taught by humans. Sci. Transl. Med. 4, immunology may be, the immune response to cancer may be 117fs2. less complicated and less protean than the biology of cancer cells themselves. Demaria, S., Kawashima, N., Yang, A.M., Devitt, M.L., Babb, J.S., Allison, J.P., and Formenti, S.C. (2005). Immune-mediated inhibition of metastases after The early data from clinical studies of immune-based thera- treatment with local radiation and CTLA-4 blockade in a mouse model of pies suggest durable activity that few cancer therapies can breast cancer. Clin. Cancer Res. 11, 728–734. approximate. The immune response can be rapid, durable, and Drake, C.G., McDermott, D.F., Sznol, M., Choueiri, T.K., Kluger, H.M., Powd- adaptable. Once activated, it has the potential to be self-propa- erly, J.D., Smith, D.C., Sankar, V., Gutierrez, A.A., Wigginton, J.M., et al. gating (for example, see Hamid et al., 2013a). These characteris- (2013). Survival, safety, and response duration results of nivolumab (Anti- PD-1; BMS-936558; ONO-4538) in a phase I trial in patients with previously tics may preempt the development of secondary resistance seen treated metastatic renal cell carcinoma (mRCC): Long-term patient follow- with most cancer therapies. In fact, with each revolution of the up. J. Clin. Oncol. 31(suppl), 4514. cycle, not only can an accumulation of immune stimulatory fac- Dunn, G.P., Bruce, A.T., Ikeda, H., Old, L.J., and Schreiber, R.D. (2002). Can- tors occur (Powderly et al., 2013), reinforcing the ongoing anti- cer immunoediting: from immunosurveillance to tumor escape. Nat. Immunol. tumor immune response, but it can also stimulate the generation 3, 991–998. of new antitumor immune responses by promoting antigen Duraiswamy, J., Kaluza, K.M., Freeman, G.J., and Coukos, G. (2013). Dual spreading (for example, see Corbie` re et al., 2011). As much as Blockade of PD-1 and CTLA-4 Combined with Tumor Vaccine Effectively Re- immune responses can completely and safely eradicate viral stores T-Cell Rejection Function in Tumors. Cancer Res. 73, 3591–3603. infections, the goal of cancer immunotherapy should remain Ferguson, T.A., Choi, J., and Green, D.R. (2011). Armed response: how dying the complete and safe eradication of cancer from individual cells influence T-cell functions. Immunol. Rev. 241, 77–88. patients. Meeting this objective may require only monotherapy Francisco, L.M., Sage, P.T., and Sharpe, A.H. (2010). The PD-1 pathway in approaches in some patients, whereas others may require com- tolerance and autoimmunity. Immunol. Rev. 236, 219–242. bined therapies. By understanding the biology present in specific Franciszkiewicz, K., Boissonnas, A., Boutet, M., Combadie` re, C., and Mami- patients, immune-related biomarkers may allow us to map out Chouaib, F. (2012). Role of chemokines and chemokine receptors in shaping the Cancer-Immunity Cycle for individual patients and enable the effector phase of the antitumor immune response. Cancer Res. 72, tailoring of specific immune therapies or combinations of im- 6325–6332. mune therapies. Franciszkiewicz, K., Le Floc’h, A., Boutet, M., Vergnon, I., Schmitt, A., and Mami-Chouaib, F. (2013). CD103 or LFA-1 engagement at the immune synapse between cytotoxic T cells and tumor cells promotes maturation and reg- REFERENCES ulates T-cell effector functions. Cancer Res. 73, 617–628.

Boon, T., Cerottini, J.C., Van den Eynde, B., van der Bruggen, P., and Van Pel, Gajewski, T.F. (2012). Cancer immunotherapy. Mol. Oncol. 6, 242–250. A. (1994). Tumor antigens recognized by T lymphocytes. Annu. Rev. Immunol. 12, 337–365. Gajewski, T.F., Fuertes, M., Spaapen, R., Zheng, Y., and Kline, J. (2011). Mo- lecular proﬁling to identify relevant immune resistance mechanisms in the Brahmer, J.R., Drake, C.G., Wollner, I., Powderly, J.D., Picus, J., Sharfman, tumor microenvironment. Curr. Opin. Immunol. 23, 286–292. W.H., Stankevich, E., Pons, A., Salay, T.M., McMiller, T.L., et al. (2010). Phase I study of single-agent anti-programmed death-1 (MDX-1106) in refractory Gajewski, T.F., Woo, S.R., Zha, Y., Spaapen, R., Zheng, Y., Corrales, L., and solid tumors: safety, clinical activity, pharmacodynamics, and immunologic Spranger, S. (2013). Cancer immunotherapy strategies based on overcoming correlates. J. Clin. Oncol. 28, 3167–3175. barriers within the tumor microenvironment. Curr. Opin. Immunol. 25, 268–276. Brahmer, J.R., Horn, L., Antonia, S.J., Spigel, D.R., Gandhi, L., Sequist, L.V., Sankar, V., Ahlers, C.M., Wigginton, J.M., Kollia, G., et al. (2013). Survival Galon, J., Angell, H.K., Bedognetti, D., and Marincola, F. (2013). The contin- and long-term follow-up of the phase I trial of nivolumab (Anti-PD-1; BMS- uum of cancer immunosurveillance: prognostic, predictive, and mechanistic 936558; ONO-4538) in patients (pts) with previously treated advanced non- signatures. Immunity 39, this issue, 11–26. small cell lung cancer (NSCLC). J. Clin. Oncol. 31(suppl), 8030. Ge, Y., Xi, H., Ju, S., and Zhang, X. (2013). Blockade of PD-1/PD-L1 immune Brestoff, J.R., and Artis, D. (2013). Commensal bacteria at the interface of host checkpoint during DC vaccination induces potent protective immunity against metabolism and the immune system. Nat. Immunol. 14, 676–684. breast cancer in hu-SCID mice. Cancer Lett. 336, 253–259.

Bulgaru, A.M., Mani, S., Goel, S., and Perez-Soler, R. (2003). Erlotinib (Tar- Goding, S.R., Wilson, K.A., Xie, Y., Harris, K.M., Baxi, A., Akpinarli, A., Fulton, ceva): a promising drug targeting epidermal growth factor receptor tyrosine A., Tamada, K., Strome, S.E., and Antony, P.A. (2013). Restoring immune func- kinase. Expert Rev. Anticancer Ther. 3, 269–279. tion of tumor-speciﬁc CD4+ T cells during recurrence of melanoma. J. Immunol. 190, 4899–4909. Chemnitz, J.M., Parry, R.V., Nichols, K.E., June, C.H., and Riley, J.L. (2004). SHP-1 and SHP-2 associate with immunoreceptor tyrosine-based switch Greaves, P., and Gribben, J.G. (2013). The role of B7 family molecules in he- motif of programmed death 1 upon primary human T cell stimulation, but matologic malignancy. Blood 121, 734–744. only receptor ligation prevents T cell activation. J. Immunol. 173, 945–954. Grosso, J., Horak, C.E., Inzunza, D., Cardona, D.M., Simon, J.S., Gupta, A.K., Chen, D.S., Irving, B.A., and Hodi, F.S. (2012). Molecular pathways: next-gen- Sankar, V., Park, J.-S., Kollia, G., Taube, J.M., et al. (2013). Association of tu- eration immunotherapy—inhibiting programmed death-ligand 1 and pro- mor PD-L1 expression and immune biomarkers with clinical activity in patients grammed death-1. Clin. Cancer Res. 18, 6580–6587. (pts) with advanced solid tumors treated with nivolumab (anti-PD-1; BMS- 936558; ONO-4538). J. Clin. Oncol. 31(suppl), 3016. Cho, D.C., Sosman, J.A., Sznol, M., Gordon, M.S., Hollebecque, A., Hamid, O., McDermott, D.F., Delord, J.-P., Rhee, I.P., Mokatrin, A., et al. (2013). Clinical Grupp, S.A., Kalos, M., Barrett, D., Aplenc, R., Porter, D.L., Rheingold, S.R., activity, safety, and biomarkers of MPDL3280A, an engineered PD-L1 anti- Teachey, D.T., Chew, A., Hauck, B., Wright, J.F., et al. (2013). Chimeric antigen body in patients with metastatic renal cell carcinoma (mRCC). J. Clin. Oncol. receptor-modiﬁed T cells for acute lymphoid leukemia. N. Engl. J. Med. 368, 31(suppl), 4505. 1509–1518.

8 Immunity 39, July 25, 2013 ª2013 Elsevier Inc. Immunity Review

Hamid, O., Robert, C., Daud, A., Hodi, F.S., Hwu, W.J., Kefford, R., Wolchok, astatic castration-resistant prostate cancer: a phase 1 dose-escalation trial. J.D., Hersey, P., Joseph, R.W., Weber, J.S., et al. (2013a). Safety and tumor Lancet Oncol. 13, 501–508. responses with lambrolizumab (anti-PD-1) in melanoma. N. Engl. J. Med. Pub- lished online June 2, 2013. http://dx.doi.org/10.1056/NEJMoa1305133. Mellman, I., Coukos, G., and Dranoff, G. (2011). Cancer immunotherapy comes of age. Nature 480, 480–489. Hamid, O., Sosman, J.A., Lawrence, D.P., Sullivan, R.J., Ibrahim, N., Kluger, H.M., Boasberg, P.D., Flaherty, K., Hwu, P., Ballinger, M., et al. (2013b). Clin- Motz, G.T., and Coukos, G. (2013). Deciphering and Reversing Tumor Immune ical activity, safety, and biomarkers of MPDL3280A, an engineered PD-L1 anti- Suppression. Immunity 39, this issue, 61–73. body in patients with locally advanced or metastatic melanoma (mM). J. Clin. Oncol. 31(suppl), 9010. Mullard, A. (2013). New checkpoint inhibitors ride the immunotherapy tsunami. Nat. Rev. Drug Discov. 12, 489–492. Heo, J., Reid, T., Ruo, L., Breitbach, C.J., Rose, S., Bloomston, M., Cho, M., Lim, H.Y., Chung, H.C., Kim, C.W., et al. (2013). Randomized dose-ﬁnding clin- Ott, P.A., Henry, T., Baranda, S.J., Frleta, D., Manches, O., Bogunovic, D., and ical trial of oncolytic immunotherapeutic vaccinia JX-594 in liver cancer. Nat. Bhardwaj, N. (2013). Inhibition of both BRAF and MEK in BRAF(V600E) mutant Med. 19, 329–336. melanoma restores compromised dendritic cell (DC) function while having differential direct effects on DC properties. Cancer Immunol. Immunother. 62, Herbst, R.S., Gordon, M.S., Fine, G.D., Sosman, J.A., Soria, J.-C., Hamid, O., 811–822. Powderly, J.D., Burris, H.A., Mokatrin, A., Kowanetz, M., et al. (2013). A study of MPDL3280A, an engineered PD-L1 antibody in patients with locally Palucka, K., and Banchereau, J. (2013). Dendritic-cell-based therapeutic can- advanced or metastatic tumors. J. Clin. Oncol. 31(suppl), 3000. cer vaccines. Immunity 39, this issue, 38–48.

Hiniker, S.M., Chen, D.S., Reddy, S., Chang, D.T., Jones, J.C., Mollick, J.A., Parry, R.V., Chemnitz, J.M., Frauwirth, K.A., Lanfranco, A.R., Braunstein, I., Swetter, S.M., and Knox, S.J. (2012). A systemic complete response of meta- Kobayashi, S.V., Linsley, P.S., Thompson, C.B., and Riley, J.L. (2005). static melanoma to local radiation and immunotherapy. Transl Oncol 5, CTLA-4 and PD-1 receptors inhibit T-cell activation by distinct mechanisms. 404–407. Mol. Cell. Biol. 25, 9543–9553.

Hodi, F.S., O’Day, S.J., McDermott, D.F., Weber, R.W., Sosman, J.A., Haanen, Patnaik, A., Kang, S.P., Tolcher, A.W., Rasco, D.W., Papadopoulos, K.P., J.B., Gonzalez, R., Robert, C., Schadendorf, D., Hassel, J.C., et al. (2010). Beeram, M., Drengler, R., Chen, C., Smith, L., Perez, C., et al. (2012). Phase Improved survival with ipilimumab in patients with metastatic melanoma. I study of MK-3475 (anti-PD-1 monoclonal antibody) in patients with advanced N. Engl. J. Med. 363, 711–723. solid tumors. J. Clin. Oncol. 30(suppl), 2512.

Huang, T., Karsy, M., Zhuge, J., Zhong, M., and Liu, D. (2013). B-Raf and the Peng, W., Liu, C., Xu, C., Lou, Y., Chen, J., Yang, Y., Yagita, H., Overwijk, inhibitors: from bench to bedside. J Hematol Oncol 6, 30–38. W.W., Lize´ e, G., Radvanyi, L., and Hwu, P. (2012). PD-1 blockade enhances T-cell migration to tumors by elevating IFN-g inducible chemokines. Cancer Isaacs, J.D., Wing, M.G., Greenwood, J.D., Hazleman, B.L., Hale, G., and Res. 72, 5209–5218. Waldmann, H. (1996). A therapeutic human IgG4 monoclonal antibody that de- pletes target cells in humans. Clin. Exp. Immunol. 106, 427–433. Postow, M.A., Callahan, M.K., Barker, C.A., Yamada, Y., Yuan, J., Kitano, S., Mu, Z., Rasalan, T., Adamow, M., Ritter, E., et al. (2012). Immunologic corre- Jacobs, J.F., Nierkens, S., Figdor, C.G., de Vries, I.J., and Adema, G.J. (2012). lates of the abscopal effect in a patient with melanoma. N. Engl. J. Med. Regulatory T cells in melanoma: the ﬁnal hurdle towards effective immuno- 366, 925–931. therapy? Lancet Oncol. 13, e32–e42. Powderly, J.D., Koeppen, H., Hodi, F.S., Sosman, J.A., Gettlinger, S.N., Desai, James, B.R., Tomanek-Chalkley, A., Askeland, E.J., Kucaba, T., Grifﬁth, T.S., R., Tabernero, J., Soria, J.-C., Hamid, O., Fine, G.D., et al. (2013). Biomarkers and Norian, L.A. (2012). Diet-induced obesity alters dendritic cell function in and associations with the clinical activity of PD-L1 blockade in a MPDL3280A the presence and absence of tumor growth. J. Immunol. 189, 1311–1321. study. J. Clin. Oncol. 31(suppl), 3001.

Kalos, M., and June, C.H. (2013). Adoptive T cell transfer for cancer immuno- Predina, J., Eruslanov, E., Judy, B., Kapoor, V., Cheng, G., Wang, L.C., Sun, J., therapy in the era of synthetic biology. Immunity 39, this issue, 49–60. Moon, E.K., Fridlender, Z.G., Albelda, S., and Singhal, S. (2013). Changes in the local tumor microenvironment in recurrent cancers may explain the failure Karan, D., and Van Veldhuizen, P. (2012). Combination immunotherapy with of vaccines after surgery. Proc. Natl. Acad. Sci. USA 110, E415–E424. prostate GVAX and ipilimumab: safety and toxicity. Immunotherapy 4, 577–580. Qureshi, O.S., Zheng, Y., Nakamura, K., Attridge, K., Manzotti, C., Schmidt, E.M., Baker, J., Jeffery, L.E., Kaur, S., Briggs, Z., et al. (2011). Trans-endocy- Kasuga, K. (2013). Comprehensive analysis of MHC ligands in clinical material tosis of CD80 and CD86: a molecular basis for the cell-extrinsic function of by immunoafﬁnity-mass spectrometry. Methods Mol. Biol. 1023, 203–218. CTLA-4. Science 332, 600–603.

Kochenderfer, J.N., and Rosenberg, S.A. (2013). Treating B-cell cancer with Rammensee, H.G., Weinschenk, T., Gouttefangeas, C., and Stevanovic, S. T cells expressing anti-CD19 chimeric antigen receptors. Nat Rev Clin Oncol (2002). Towards patient-speciﬁc tumor antigen selection for vaccination. Im- 10, 267–276. munol. Rev. 188, 164–176.

Kruit, W.H., Suciu, S., Dreno, B., Mortier, L., Robert, C., Chiarion-Sileni, V., Riella, L.V., Paterson, A.M., Sharpe, A.H., and Chandraker, A. (2012). Role of Maio, M., Testori, A., Dorval, T., Grob, J.J., et al. (2013). Selection of immunos- the PD-1 pathway in the immune response. Am. J. Transplant. 12, 2575–2587. timulant AS15 for active immunization with MAGE-A3 protein: results of a randomized phase II study of the European Organisation for Research and Rotem-Yehudar, R., Rodionov, G., and Landes, S. (2009). CT-011, a human- Treatment of Cancer Melanoma Group in metastatic melanoma. J. Clin. Oncol. ized monoclonal antibody, interacts with the PD-1 receptor and modulates 31, 2413–2420. survival and trafﬁcking signals in effector/memory T lymphocytes. Proc. Am. Assoc. Cancer Res., Apr 18–22 Abstract 2410. Lin, Y.L., Meng, Y., Jiang, W., and Roux, B. (2013). Explaining why Gleevec is a speciﬁc and potent inhibitor of Abl kinase. Proc. Natl. Acad. Sci. USA 110, Rothman, J., and Paterson, Y. (2013). Live-attenuated Listeria-based immuno- 1664–1669. therapy. Expert Rev. Vaccines 12, 493–504.

Lippitz, B.E. (2013). Cytokine patterns in patients with cancer: a systematic re- Rothschild, S.I., and Gautschi, O. (2013). Crizotinib in the Treatment of Non- view. Lancet Oncol. 14, e218–e228. Small-Cell Lung Cancer. Clin. Lung Cancer. Published online June 19, 2013. http://dx.doi.org/10.1016/j.cllc.2013.04.006. Liu, C., Peng, W., Xu, C., Lou, Y., Zhang, M., Wargo, J.A., Chen, J.Q., Li, H.S., Watowich, S.S., Yang, Y., et al. (2013). BRAF inhibition increases tumor inﬁltra- Segal, N.H., Parsons, D.W., Peggs, K.S., Velculescu, V., Kinzler, K.W., Vogel- tion by T cells and enhances the antitumor activity of adoptive immunotherapy stein, B., and Allison, J.P. (2008). Epitope landscape in breast and colorectal in mice. Clin. Cancer Res. 19, 393–403. cancer. Cancer Res. 68, 889–892.

Madan, R.A., Mohebtash, M., Arlen, P.M., Vergati, M., Rauckhorst, M., Stein- Smothers, F., Hoos, A., Langermann, S., Marshall, S., May, R., and Fleming, M. berg, S.M., Tsang, K.Y., Poole, D.J., Parnes, H.L., Wright, J.J., et al. (2012). (2013). AMP-224, A fusion Protein That Targets PD-1. Ann. Oncol. 24(suppl), Ipilimumab and a poxviral vaccine targeting prostate-speciﬁc antigen in met- i7–i17.

Immunity 39, July 25, 2013 ª2013 Elsevier Inc. 9 Immunity Review

So, T., Lee, S.W., and Croft, M. (2006). Tumor necrosis factor/tumor necrosis Topalian, S.L., Sznol, M., Brahmer, J.R., McDermott, D.F., Smith, D.C., Get- factor receptor family members that positively regulate immunity. Int. J. Hem- tinger, S.N., Taube, J.M., Drake, C.G., Pardoll, D.M., Powderly, J.D., et al. atol. 83, 1–11. (2013). Nivolumab (anti-PD-1; BMS-936558; ONO-4538) in patients with advanced solid tumors: Survival and long-term safety in a phase I trial. Spigel, D.R., Gettinger, S.N., Horn, L., Herbst, R.S., Gandhi, L., Gordon, M.S., J. Clin. Oncol. 31(suppl), 3002. Cruz, C., Conkling, P., Cassier, P.A., Antonia, S.J., et al. (2013). Clinical activity, safety and biomarkers of MPDL3280A, an engineered PD-L1 antibody in van den Boorn, J.G., and Hartmann, G. (2013). Turning tumors into vaccines: patients with locally advanced or metastatic non-small cell lung cancer Co-opting the Innate Immune System. Immunity 39, this issue, 27–37. (NSCLC). J. Clin. Oncol. 31(suppl), 8008. Vanneman, M., and Dranoff, G. (2012). Combining immunotherapy and tar- Stagg, J., Loi, S., Divisekera, U., Ngiow, S.F., Duret, H., Yagita, H., Teng, M.W., geted therapies in cancer treatment. Nat. Rev. Cancer 12, 237–251. and Smyth, M.J. (2011). Anti-ErbB-2 mAb therapy requires type I and II inter- ferons and synergizes with anti-PD-1 or anti-CD137 mAb therapy. Proc. Natl. Vansteenkiste, J., Zielinski, M., Linder, A., Dahabreh, J., Gonzalez, E.E., Mali- Acad. Sci. USA 108, 7142–7147. nowski, W., Lopez-Brea, M., Vanakesa, T., Jassem, J., Kalofonos, H., et al. (2013). Adjuvant MAGE-A3 immunotherapy in resected non-small-cell lung Stewart, R.A., Morrow, M., Chodorge, M., Marcus, D., Boyle, M., Mulgrew, K., cancer: Phase II randomized study results. J. Clin. Oncol. 31, 2396–2403. Hammond, S., Wang, S., Alimzhanov, M., Robbins, P., et al. (2011). Delivering effective blockade of immunosupression to enhance tumour rejection: Mono- Wang, L., Qian, J., Lu, Y., Li, H., Bao, H., He, D., Liu, Z., Zheng, Y., He, J., Li, Y., clonal antibody discovery and preclinical development. Cancer Res. et al. (2013). Immune evasion evasion of mantle cell lymphoma: expression of 71(Suppl ), LB-158. B7-H1 leads to inhibited T-cell response to and killing of tumor cells. Haema- tologica. Published online March 18, 2013. http://dx.doi.org/10.3324/haema- Sznol, M., Kluger, H.M., Hodi, F.S., McDermott, D.F., Carvajal, R.D., Law- tol.2012.071340. rence, D.P., Topalian, S.L., Atkins, M.B., Powderly, J.D., Sharfman, W.H., et al. (2013). Survival and long-term follow-up of safety and response in pa- West, E.E., Jin, H.T., Rasheed, A.U., Penaloza-Macmaster, P., Ha, S.J., Tan, tients (pts) with advanced melanoma (MEL) in a phase I trial of nivolumab W.G., Youngblood, B., Freeman, G.J., Smith, K.A., and Ahmed, R. (2013). (anti-PD-1; BMS-936558; ONO-4538). J. Clin. Oncol. 31(suppl), CRA9006. PD-L1 blockade synergizes with IL-2 therapy in reinvigorating exhausted T cells. J. Clin. Invest. 123, 2604–2615. Tabernero, J., Powderly, J.D., Hamid, O., Gordon, M.S., Fisher, G.A., Braiteh, F.S., Garbo, L.E., Fine, G.D., Kowanetz, M., McCall, B., et al. (2013). Clinical Westin, J., Chu, F., Fayad, L., Kwak, L.W., Fowler, N.H., Romaguera, J., Hage- activity, safety, and biomarkers of MPDL3280A, an engineered PD-L1 anti- meister, F.B., Fanale, M., Samaniego, F., Allen, R., et al. (2012). Phase II Safety body in patients with locally advanced or metastatic CRC, gastric cancer and Efﬁcacy Study of CT-011, a Humanized anti-PD-1 Monoclonal Antibody, (GC), SCCHN, or other tumors. J. Clin. Oncol. 31(suppl), 3622. in Combination with Rituximab in Patients with Relapsed Follicular Lymphoma. Blood 120, 793. Tian, T., Olson, S., Whitacre, J.M., and Harding, A. (2011). The origins of cancer robustness and evolvability. Integr Biol (Camb) 3, 17–30. Wolchok, J.D., Kluger, H., Callahan, M.K., Postow, M.A., Rizvi, N.A., Lesokhin, A.M., Segal, N.H., Ariyan, C.E., Gordon, R.A., Reed, K., et al. (2013). Nivolu- Topalian, S.L., Drake, C.G., and Pardoll, D.M. (2012a). Targeting the PD-1/B7- mab plus ipilimumab in advanced melanoma. N. Engl. J. Med. H1(PD-L1) pathway to activate anti-tumor immunity. Curr. Opin. Immunol. 24, 207–212. Xu, X., Ye, L., Araki, K., and Ahmed, R. (2012). mTOR, linking metabolism and immunity. Semin. Immunol. 24, 429–435. Topalian, S.L., Hodi, F.S., Brahmer, J.R., Gettinger, S.N., Smith, D.C., McDer- mott, D.F., Powderly, J.D., Carvajal, R.D., Sosman, J.A., Atkins, M.B., et al. Zitvogel, L., Galluzzi, L., Smyth, M.J., and Kroemer, G. (2013). Mechanism of (2012b). Safety, activity, and immune correlates of anti-PD-1 antibody in can- action of conventional and targeted anticancer therapies: reinstating immuno- cer. N. Engl. J. Med. 366, 2443–2454. surveillance. Immunity 39, this issue, 74–88.

10 Immunity 39, July 25, 2013 ª2013 Elsevier Inc.