Inhibitors of Carbohydrate Processing: a New Class of Anticancer Agents1’2

Vol. 1, 935-944, September 1995 Clinical Cancer Research 935

Minireview

Inhibitors of Carbohydrate Processing: A New Class of Anticancer Agents1’2

Paul E Goss,3 Michael A. Baker, clinical applications of this new class of anticancer agents Jeremy P. Carver, and James W. Dennis are emphasized. The Toronto Hospital, Department of Medical Oncology, Faculty of Medicine [P. E. G., M. A. B.] and Department of Medical Genetics Carbohydrate Processing and Malignancy [J. P. C., J. W. D.], University of Toronto, Toronto, Ontario, and Malignant transformation is associated with a variety of Samuel Lunenfeld Research Institute, Mt. Sinai Hospital [J. W. D], structural alterations in the carbohydrate portion of glycopro- Toronto, Ontario M4X 1K9, Canada teins (1-4). Glycoprotein glycosylation (Fig. 1) begins in the lumen of the rough endoplasmic reticulum (5-8) where a subset Abstract of Asn (i.e., N-linked) and SeriThr (i.e., 0-linked) residues on There is a need for anticancer agents with novel mech- newly synthesized proteins are subject to the addition of anisms of action. Recently identified molecular targets for sugars (reviewed in Ref. 6). For N-linked carbohydrates, new anticancer agents include inducers of cell differentia- Glc3Man9GlcNAc2 s preassembled on dolichol PP1 and then tion, cell cycle arrest, and apoptosis, as well as signaling transferred as a unit to Asn-X-Ser/Thr sequences of glycopro- pathways for growth factors and cytokines. teins while they are being synthesized. This initial glycosylating Another unexplored opportunity is presented by the event is required for many glycoproteins to fold into their native ubiquitous intracellular glycoprotein glycosylation pathway. or active conformation. The Glc3Man9GlcNAc2 structures are This complex process, concerned with the addition of sugars then remodeled or processed as the newly synthesized glyco- onto newly synthesized proteins, occurs in the lumen of the proteins are transported through the Golgi on route to the cell rough endoplasmic reticulum and in the Golgi. There are surface. As depicted in Fig. 1, this begins with trimming by estimates of over 200 glycosyltransferase enzymes in this a-glucosidases and a-mannosidases I, leaving Man5GlcNAc2-N pathway, which results in considerable structural diversity which is then substituted by GIcNAc-TI and trimmed by a-man- of carbohydrates found on secreted and transmembrane nosidase II producing GlcNAcMan3GlcNAc2-N. The latter glycoproteins. The specificity of glycosyltransferases for ac- structure is a substrate for the GlcNAc-Ts (i.e., G1cNAc-T-II, ceptors and sugar-nucleotide donors dictates linkage posi- -IV, -V); each enzyme substitutes a distinct position on the tions between sugars, anomeric configuration of linkages, trimannose core to initiate “branches” (6). Cancer cells corn- and monosaccharide composition. Specific carbohydrate monly show increased 31-6-N-acetylglucosamine (G1cNAc)- structures participate in cell-cell and cell-substratum inter- branching at the trimannosyl core of N-linked carbohydrates actions affecting processes such as lymphocyte trafficking, (9-12). For example, increased branching has been noted in immune cell stimulation, embryogenesis, and cancer metas- primary tumors of human carcinoma of the breast, colon, and tasis. skin, and appears also to correlate with disease progression (13, Of the carbohydrate-processing inhibitors presently 14). In normal tissues, the 31-6GlcNAc branched carbohydrate available, the alkaloid swainsonine, a Golgi a-mannosidase structures are restricted to cells capable of invasion including II inhibitor, is the first to have been selected for clinical trophoblasts, endothelial cells, interstitial fibroblasts, and acti- testing based on its anticancer activity, p.o. availability, and vated lymphocytes (14, 15). The key enzyme which initiates the low toxicity in mice. Herein, we review the rationale for 31-6 branching is 31-6GlcNAc TV” (see Fig. 1 and Ref. 9). targeting Golgi carbohydrate processing pathways in the The antenna or branch initiated by G1cNAc-TV is preferred treatment of cancer, and summarize the preclinical and by subsequent enzymes in the pathway for extension with clinical results with swainsonine. Prospects for the develop- polylactosamine (i.e., repeating Gal31-4GlcNAc31-3), Lewis ment of second generation inhibitors with improved speci- antigens, and blood-group sequences (Refs. 16-18 and Fig. 1). ficity for Golgi-processing enzymes are discussed. Potential These sequences are developmentally regulated structures with limited distribution in normal tissues, but they are expressed in human carcinoma and therefore appear to be markers of malignancy (19, 20). Polylactosamine and Lewis antigen expression Received 3/31/95; revised 6/26/95; accepted 7/5/95. on tumor cells may contribute to cell-cell adhesion via selectins 1 This work was supported by research grants from the National Cancer and 3Gal-binding lectins, which are found on the vascular Institute of Canada and Medical Research Council of Canada. J. W. D. endothelium, and thereby enhance invasion and metastasis by is a senior research scientist of the National Cancer Institute of Canada.

2 We dedicate this review to the memory of Dr. Martin L. Breitman, a colleague and good friend who died of cancer February 13, 1994 at the age of 41 years. Among the many accomplishments in Martin’s scien- tific career, we are grateful for his contribution to the development of 4 The abbreviations used are: TV, transferase V; CPI, carbohydrate- this area, and we will greatly miss his inspiration and friendship. processing inhibitor; TIMP, tissue inhibitor of metalloproteinases; NK, 3 To whom requests for reprints should be addressed, at The Toronto natural killer; IL, interleukin; LAK, lymphokine-activated killer; TNF, Hospital, General Division, 200 Elizabeth Street, mlw 2-022, Toronto, tumor necrosis factor; PBL, peripheral blood lymphocytes; L-PHA, Ontario M4X 1K9, Canada.

I r y IlL II other -‘:; termini ; ,, polylactosamlne r hybrid-type complex-type t f t t Lex,1. cpp L k ‘ . .-

-#{248} *T TiLT Y- I Y!!1 Slex Tun cast DMJ SW _____

Tl/’ Ley -*? otMl 7

alternate pathway

GIcNAc Gic A Man 0 Gal #{149} Fuc V SA

Fig. 1 The Golgi N-linked carbohydrate-processing pathway. A large number of sequences, examples of which are shown on the rig/it. can be added to the termini. with preference for addition to the 3 I -6GIcNAc-branched complex-type structures. A minor alternate processing pathway that circumvents the need for a-mannosidase Il is also shown. Enzymes: aG. -glucosida.se: aM!. a-mannosidase I: aM/I, -mannosidase II; TI, TI!, TIV, and TV, respective GlcNAc-Ts. Glucosidase inhibitors:SW. swainsonine; cast, ca.stanospermine; DNJ. 1-deoxynojirimycin: DMJ. 1-deoxymannojirimycin.

Table / Studies on the MDAY-D2 lymphoma

Metastasis”

Phenotype” Tumor (cm3) Spontaneous (s.c.) Experiments (iv.)

Wild-type 4.55 ± 0.40 1(X), 173. 227, 464. 446 >5()0. 500, 5(X), 5(X). 5()0 NeuNGc 2.13 ± 0.49 3. 243. 308. 409. SOt) 312. 4(X). 410. >500, 5(X) a2-#{212}SA-T 1.81 ± 0.39 61. 98, 137, 144 0, 3, 8. 8. 11

GIcNAc-TC 1.57 ± 1.32 0. 0. 0, 0. 1 . I . I 0. 0. 0, 0. 0 UDP-Gal 0.10 ± 0.05 0, 0, 0, 0, 0 0, 0. 0. 0, 0 Swainsonine-treated 1.53 ± 0.50 0. 0. 0, 20. 36 5()-90% decrease

(1 NeuNAc and NeuNGc are two naturally occurring and developmentally regulated forms of sialic acids, N-acetylneuraminic acid and N-glycolylneuraminic acid, respectively. The NeuNGc mutation is due to expression of the enzyme CMP-SA hydroxylase. resulting in glycoconju- gates with NeuNGc rather than NeuNAc. the a2-6SA-T mutation is due to overexpression of a2-6SA transferase. resulting in SA-linked a2-6 rather than a2-3; the GlcNAc-TV mutation is due to loss of this activity, resulting in decreased I-6GlcNAc-branching of N-linked oligosaccharides: and the UDP-Gal mutation is a defect in transport of this sugar nucleotide into the Goli, resulting in loss of Gal and SA from glycoproteins.

I) Metastasis and solid tumor size were measured I 7 days after tumor cells were injected. Summary based on data from multiple experiments.

blood-borne tumor cells (21-23). However, it is important to highly malignant tumor cell lines (27, 28). By examining the note that cancer-associated changes in carbohydrate processing biochemical defects and malignant properties of independent appear to affect not only adhesive phenomena, but also multiple genetic mutations, a direct association has been observed be- other aspects of tumor cell phenotype as discussed below (3). tween expression of 1-6GIcNAc-branched complex-type car- Interestingly, transformation of cells in tissue culture by bohydrates and malignant potential in these mutant cell lines activated oncogenes in the ras-signaling pathway (i.e., T24 (28, 29). Mutations selected in the highly metastatic MDAY-D2 H-ras, v-src, v-fps, middle T of polyoma virus; Refs. 9, 10, 24, lymphoma cell line are listed in Table 1 in order of increasing and 25) as well as c-mvc (26) has been shown to increase effect on the malignant phenotype. Mutants with minor struc- GIcNAc-TV activity. Thus, up-regulation of this and possibly tural changes in sialic acid showed slower solid tumor growth other enzymes in the pathway may represent mandatory steps in but remained metastatic. In contrast. the deficiency in Golgi the phenotypic expression of some malignancies. UDP-Gal transport activity resulted in loss ofGal and sialic acid Somatic mutant cell lines with defects in the N-linked from both 0- and N-linked oligosaccharides, and clearly had the carbohydrate processing pathway have been selected from most marked effect on solid tumor growth and metastasis (28).

Mutants deficient in GlcNAc-TV activity affected only the Swainsonine has been administered to adult mice and rats for N-linked pathway, and this also profoundly suppressed tumor periods of up to 2 months with no mortality (47, 55, 56). One cell metastasis (24, 28). These mutations are causally related to effect is that it causes carbohydrate accumulation (i.e., lysoso- the loss of malignant potential based on several observations. mal storage) in tissues, the least of which occurs in the brain. In Examination of independent isolates with mutations in a corn- a recent report of the distribution of swainsonine when admin- mon gene showed the same phenotype. The mutations have been istered to mice, brain levels were noted to be 5-20-fold lower isolated and characterized several times in independent tumor than those measured in other organs such as the bladder, kidney, cell lines, both mouse and human (24, 28, 30, 31). Finally, or thymus gland (57). revertants of the UDP-Gal transport mutation, i.e., mutant cells Swainsonine is a potent, competitive inhibitor of Golgi that had regained the UDP-Gal transport function, also regained a-mannosidase II (K1 = 40 nM; Refs. 58-60) but not a-man- the malignant phenotype (32, 33). nosidase I, and trimming therefore proceeds as far as More recently, experiments using expression vectors for Man5GlcNAc2 (Fig. 1). Inhibition of a-mannosidase II pre- glycosyltransferases support a direct relationship between 31- vents substitution of the structures by G1cNAc-TII and G1cNAc- 6GlcNAc-branched oligosaccharides and the malignant pheno- TV, thereby diverting the pathway to produce “hybrid-type” type. Immortalized (i.e., premalignant) lung epithelial cells carbohydrates (36). Swainsonine also inhibits lysosomal al-3- transfected with a G1cNAc-TV expression vector showed re- and al-6-mannosidases (IC50 = 70 nM), the degradative path- laxed growth controls, increased cell motility, reduced substra- way for glycoproteins (61-64). This latter activity may produce turn adhesion, and increased incidence of tumorigenesis in nude unwanted clinical side effects and might be eliminated in second mice (34). B16 melanoma cells transfected with G1cNAc-TIII, generation inhibitors as discussed further below. an enzyme that competes with G1cNAc-TV for acceptor sub- As a potential therapeutic agent, swainsonine has several strate, and thereby suppresses the expression of 31-6GlcNAc- apparent advantages, most notably low preclinical toxicity, branched oligosaccharide, has recently been shown to inhibit compared to inhibitors that block processing earlier in the path- lung metastasis (35). These observations suggest that up-regu- way. For example, tunicamycin, which blocks the transfer of lation of GlcNAc-TV can contribute to features of the prema- Glc3Man9GlcNAc2 from dolichol to Asn-X-Ser/Thr, is highly lignant cellular phenotype as well as tumor cell metastasis, and toxic in cell culture and when given to mice (65). Toxicity is may be an important downstream affector of transformation by probably as a result of depletion of the membrane glycoproteins oncogenes in the ras-signaling pathway. due to the inability of some underglycosylated glycoproteins to A number of CPIs isolated from natural sources have been fold properly in the endoplasmic reticulum (66). Castanosper- identified, including tunicamycin and several a-glucosidase and mine, 1 -deoxynojirimycin, and 1-deoxymannojirimycin block a-mannosidase inhibitors (Figs. 1 and 2; Refs. 36 and 37). processing prior to trimming of the oligomannose chains (Fig. Indeed, murine tumor cells cultured in the presence of the 1), and can also adversely affect intracellular movement of dolichol-oligosaccharyltransferase inhibitor, tunicamycin; the glycoproteins (67-69) or the function of some glycoproteins a-glucosidase inhibitors, castanosperrnine, 1-deoxynojirimycin, (68, 70, 71). In contrast, swainsonine does not appear to block or 1,6-epi-cyclophellitol; the Golgi a-mannosidase I and II membrane localization or secretion of glycoproteins (69, 72- inhibitors, 1-deoxymannojirirnycin, mannostatin, or swainso- 74), and shows low toxicity in tissue culture (38, 48, 75) and in nine showed a marked decrease in metastatic potential when mice (47, 55, 56). Swainsonine is also 20-100-fold more potent injected i.v. into mice (38-46). The position in the pathway than castanospermine, 1-deoxynojirimycin, or l-deoxyman- where these inhibitors block N-linked carbohydrate processing nojirimycin as an inhibitor of carbohydrate processing in cell is shown in Fig. 1 . In Fig. 2, the structures of these a-manno- culture (IC50 = 0.1 uM; Refs. 48 and 76). sidase and a-glucosidase inhibitors are shown in an orientation that compares them to the a-linked substrate normally cleaved Antitumor Effects in Mice by the enzymes. Table 1 provides an example of the antitumor Swainsonine has been administered to tumor-bearing mice effect of these inhibitors. Swainsonine-treated tumor cells in a wide range of dosages by either p.o. administration (39, 48, showed much reduced organ colonization, and tumor-bearing 77), i.p. injection (78), or infusion with miniosmotic pumps mice treated with the drug showed reduced solid tumor growth (47). It has been shown to inhibit organ colonization and solid and metastases similar to that observed for the GlcNAc-TV tumor growth rate, and to extend survival of mice following mutants (38, 39, 47-49). surgical removal of primary murine tumor implants (39, 47, 48, 77, 79). Human colorectal carcinoma tumors growing in athy- Swainsonine, A Prototypic CPI for Cancer Therapy mic nude mice were growth rate inhibited by approximately Swainsonine, or 8a3-indolizidine-1a,2a,83-triol, is an in- 50% when mice were provided with drinking water containing dolizidine alkaloid first isolated from the Australian plant 2.5 p.g/rnl swainsonine (estimated to be 0.3-0.5 mg/kg/day; Swainsona canescens (50), and later from North American Ref. 48). Oral administration of swainsonine also inhibited the plants of the genera Astragalus and Oxytropis (51). Ingestion of growth of human melanoma MeWo tumors (47) and human these plants by grazing farm animals is known to result in a breast carcinoma xenografts in nude mice (80). In another series neurological condition called “loco syndrome” (52, 53). Purified of experiments, dose-dependent inhibition of carbohydrate pro- swainsonine alone is now known not to be neurotoxic, but rather cessing and tumor inhibition was observed in athymic nude loco syndrome may be caused by ingestion of a combination of mice bearing either human MeWo melanoma xenografts (47) or alkaloids found in Swainsona canescens or Astragalus (54). metastatic murine mammary carcinoma tumors (81).

OH OH

enzyme HO...’LS1 , HO HO 3 OR HO OR uMannose ctGlucose

HO enzyme HO 1 2 3 enzyme

swainsonine castanospermine

OH OH OH

HOL HO HO deoxynojirimycin

Fig. 2 Glycosidase inhibitors. Mannose and glucose shown in an ca-glycosidic linkage to R are positioned above their analogues swainsonine and 1-deoxymannojirimycin, and castanospermine and 1-deoxynojirimycin, respectively. The structures are vertically aligned and oriented to show their similarity about the forward half of the molecules.

Anticancer Mechanisms: Effects on Tumor Cells cells cultured in the presence of swainsonine show increased Tumor Cell Adhesion to Endothelium. Blood-borne transcription rates for the tissue inhibitor of the metalloprotein- tumor cells escape into secondary tissue by attaching to endo- ase (TIMP-]) gene, while the urokinase plasminogen activator, thelium, either as single cells or by clumping with circulating transin, and actin are not altered (89). A similar up-regulation of host cells, and then extravasating and invading through the liMP is observed in the GlcNAc-TV and UDP-Gal transporter extracellular matrix (82). We and others (21, 83, 84) have glycosylation mutants of MDAY-D2, MeWo, and Chinese ham- observed that 3Gal-binding lectins present on both tumor cells ster ovary cells, suggesting the effect was due to swainsonine’s and endothelial cells contribute to the retention of blood-borne primary action as an inhibitor of carbohydrate processing (89). tumor cells in the microvasculature (83). As noted above, swain- liMP proteins bind to collagenases and inhibit their activity, sonine treatment of tumor cells reduces the number of Gal31- and have been shown to block tumor cell invasion in vivo as 4G1cNAc antennae in N-linked carbohydrates and also reduces well as metastasis in vivo (90, 91). Seftor et al. (87) reported that attachment of the MDAY-D2 tumor cells to endothelial cell both swainsonine and castanospermine block invasion through monolayers in vitro (21). in vivo, swainsonine-treated wild-type extracellular matrix by human tumor cells. In these cell lines, tumor cells show reduced organ retention on initial pass of the the drugs also suppressed expression of collagenase type IV cells through the lung5 (85). As additional evidence of the (87). Therefore, the effects of swainsonine and other CPIs on importance of Gal31-4GlcNAc antennae, mutant lymphoma tumor cell invasion may be largely due to altered TJMP and cells which lack the Golgi UDP-Gal transporter activity are collagenase gene expression, which effectively reduces the hy- deficient in organ colonization in vivo, and attach weakly to drolysis of extracellular matrix. monolayers of vascular endothelial cells in vitro (21). When these mutant cells were treated with purified f31-4-galactosyl- transferase and UDP-Gal to restore 3Gal to the surface, both Oncogenic Glycoproteins organ colonization and adhesion to endothelial cells were par- There is considerable interest in determining which glycop- tially restored. roteins in malignant cells have 31-6GlcNAc-branched carbohy- Tumor Cell Invasion. Swainsonine is a very effective drates and require this posttranslational modification to develop the inhibitor of murine and human tumor cell invasion through invasive and metastatic phenotype. Since most growth factors and extracellular matrix in vitro (86-88). It appears to affect the their receptors are glycoproteins, it is possible that cancer-associ- expression of genes encoding proteins involved in extracellular ated changes in glycosylation of certain ones of these molecules matrix proteolysis. For example, mouse mammary carcinoma may affect signaling pathways in a manner that promotes the malignant phenotype. In this regard, the transmembrane receptor tyrosine kinases such as fins and sea are glycosylated in their extracellular domains, and in addition to constitutively activating

5 5. Yagel and J. W. Dennis, unpublished data. mutations, their transforming activity is dependent on glycosylation

(68, 69). Indeed, castanospermine adminstered p.o. to nude mice sured against human colon carcinoma cells (101). Antibodies to has been shown to reduce the growth ofv-fins-induced tumors (92). IL-2 abolished swainsonine-induced enhancement of LAK cell Swainsonine has been shown to inhibit the growth of T24 H-ras- killing, but lymphocytes showed no increase in IL-2 production transfected NIH 3T3 cells in an anchorage-independent manner in or receptor number. These observations suggest that the a-man- soft agar (93). In general, N-linked carbohydrate processing inhib- nosidase inhibitors increase the sensitivity of LAK cells to IL-2 itors appear to induce subtle alterations in the transformed pheno- activation (101). Infiltrating lymphocytes and monocytes in type. Although H-ras and many other oncogenes are not glyco- tumors are often inactive, but can be activated in vitro by IL-2 proteins, they directly or indirectly regulate the expression of many and other lymphokines (105). Therefore, a-mannosidase inhib- glycoproteins, which in turn are required to create the malignant itors with good bioavailability in tumors might serve to sensitize phenotype (94). LAK cells, NK cells, and macrophages to local lymphokines.

Anticancer Mechanisms: Effects on the Host TNF/IL-1 and Macrophage Activation Immune Stimulation. Swainsonine, which alters glyco- Thioglycollate-elicited peritoneal macrophages are acti- sylation of tumors in situ, is not tumor specific, and therefore vated by culturing the cells in the presence of swainsonine or by would be expected to also have effects on host tissues. Thus, in injecting swainsonine into the peritoneal cavity (106, 107). The addition to its direct action on tumor cells, its effects in mice and macrophages show increased tumoricidal activity, IL- 1 produc- probably in humans include an immune stimulatory activity, tion, induction of protein kinase C activity, and increased cell which may play a significant role in its antitumor action. The surface class II histocompatibility antigen Ia (i.e., HLA-DR in action of swainsonine as an immune modulator has been re- humans, see Phase I trial). viewed by Humphries and Olden (95) and Olden et a!. (96). The Macrophages can be activated by cytokines, notably IL-l following section summarizes these effects. and TNF, and once activated, the cells then produce these NK Cell Activation. The following observations suggest factors. Swainsonine has been shown to enhance the toxicity of that antitumor effects of swainsonine in mice are mediated at the TNF-a toward WEHI 164 tumor cells, and to increase tumori- level of both the tumor cells and the host immune system. cidal activation of human monocytes in vitro (108). These B16F1O melanoma cells grown in the presence of swainsonine observations suggest that swainsonine and IL-lffNF-a may show reduced organ colonization when injected i.v. into un- synergize to enhance direct lymphokine killing of tumor cells as treated mice. Untreated tumor cells injected into mice that have well as by activating host macrophages. In this regard, certain been pretreated with swainsonine also produce fewer metastases functions or signaling by IL-i and TNF-a in vivo may depend (39, 77). However, pretreating mice with swainsonine had little on their oligomannose-binding activity (108, 109). effect when NK cells were eliminated in experiments using either bg/bg mice or mice treated with anti-asialo-GM1 antibod- Protection against Immunosuppressive Agents ies (77). Therefore, in addition to effects on the tumor cells, Cancer patients can exhibit depressed immune responses swainsonine also augments NK cell activity, which participates including delayed-type hypersensitivity, depressed NK cell ac- in the elimination of blood-borne B16F1O tumor cells (79). tivity, and decreased macrophage migration and phagocytosis Glycosylation mutants of Chinese hamster ovary cells and (i 10, reviewed in Ref. 1 1 i). Tumors produce, or induce host MDAY-D2 tumor cells have been shown to be more sensitive tissues to express, immunosuppressive proteins such as trans- than wild-type cells to NK cell lysis in vitro (97). In addition, forming growth factor 3, which is a broadly immunosuppressive purified oligosaccharides, similar in structure to those found in cytokine. In this connection, Kino et a!. (43) have shown that swainsonine-treated cells, stimulate NK cells in vitro (98). This swainsonine restores B-cell response to normal levels in mice suggests that NK cell recognition of hybrid-type carbohydrates treated with cyclophosphamide or in mice with immunosuppres- on target cells may contribute to their activation and/or killing function. In this regard, putative rat and mouse NK receptors sive tumor burdens. have recently been cloned and have been shown to share se- Swainsonine administered to mice i.v. enhanced bone mar- quence homology with the family of carbohydrate-binding pro- row cellularity, engraftment efficiency, and colony forming teins, the C-type lectins (99). units (1 12). Furthermore, swainsonine treatment has been reported to decrease lethality of methotrexate, 5-fluorouracil, cy- IL-2 and LAK Cells clophosphamide, and doxorubicin in nontumor-bearing mice (113). Increased survival of these mice correlated with stimula- Swainsonine is an immune modulator that can stimulate tion of bone marrow proliferation, bone marrow cellularity, and lymphocyte proliferation (100), activate natural antitumor im- engraftment efficiency in the mice. These data suggest that munity (77, 101), and enhance T cell stimulation by antigen swainsonine enhances stem cell proliferation in bone marrow, (102). Although speculative, the basis of immune cell activation possibly due to increased responsiveness of precursor cells to and bone marrow proliferation may be related to the observation endogenous lymphokines. that some lymphokines and growth factors have carbohydrate binding activities (103). This is true for example of IL-i, IL-2, and TNF, all of which have been shown to bind to oligomannose Phase I Trial of Swainsonine in Cancer Patients structures (104). We recently reported the results of our first Phase I trial of It has been shown that incubation of human lymphocytes in swainsonine in patients with advanced malignancies (1 14, 115). the presence of swainsonine enhances LAK cell activity mea- The results are summarized here.

The quantitative and qualitative toxicities of chemically culture; (c) activity in vivo by p.o. administration; (6) low toxicity synthesized swainsonine were studied in patients given a con- in animals with chronic exposure to the drug; (e) antitumor effects tinuous iv. infusion over 5 days. Dose levels were escalated in in preclinical experiments; and (J) activity and acceptable toxicity increments of 100 p.g/kg/day from 50 to 550 jig/kg/day. Nine- in clinical trials. Swainsonine meets criteria b-e and shows prom- teen patients with both advanced solid tumor and hematological ise inf. The side effects of swainsonine seen in patients may be due malignancies were given a total of 31 courses. The maximum to inhibition of lysosomal a-mannosidases. If this is the case, they tolerated dose and, thus, the maximum recommended starting might be eliminated by developing an analogue that retains a-man- dose (maximum tolerated dose-i level) by this route of admin- nosidase II inhibitor activity but lacks activity on lysosomal en- istration were 550 and 450 i.g/kg/day, respectively. However, as zymes. This could be particularly important in patients chronically discussed below, biomarker levels assessed in this study suggest treated with swainsonine. that lower doses of swainsonine may produce the therapeutic In vitro the K for inhibition of a-mannosidase II and the effects desired. IC50 for blocking carbohydrate processing in viable cells is Common side effects in patients treated included peripheral similar, suggesting that swainsonine has optimal cell entry properties (75, 76). Thus, an important consideration in designing edema (n = il/19), mild liver dysfunction in all patients (as- partate aminotransferase up to 4-fold normal), a rise in serum second generation CPIs is to maintain not only the specificity and potency of swainsonine but also the favorable pharmacoki- amylase (n = 8/i9), and decreased serum retinol levels. Acute netics and cell entry properties (76). respiratory distress syndrome, probably precipitated by swain- Based on the preclinical studies described above, an inhib- sonine in one patient with extensive liver metastases and pre- itor of the enzyme GlcNAc-TV should have potent anticancer existing liver dysfunction, resulted in a treatment-related death. activity. Acceptor analogues for GlcNAc-TV that serve as com- Other side effects noted included shortness of breath (n = 3), petitive inhibitors of the enzyme have been developed (116, lethargy (n = 4), nausea (n = 1), and skin rash (n - 1). One 1 17), but entry into the cells of compounds identified to date is patient with head and neck cancer showed >50% shrinkage of poor.6 example is GlcNAc31-2(6-deoxy)Mana1-6Glc3- tumor mass 3 weeks after treatment. Two patients with lym- OCO2(CH2)7CH3, which lacks the OH normally substituted by phangitis carcinomatosis on chest X-ray noted improvement in the enzyme, and has a K. of 70 tiM, similar to the Km for the cough and shortness of breath during the infusion of swainso- 6-hydroxylated acceptor and thus has poor cell entry (117). In nine and for 1 week thereafter. addition, it is an acceptor for 31-4Gal-T which inactivates its Clearance and serum half-life for swainsonine were deter- GlcNAc-TV inhibitory activity. mined to be approximately 2 ml/h - kg and 0.5 days, respectively. It is possible that drug resistance may be encountered for Golgi oligosaccharide processing, the putative anticancer target for compounds designed as CPIs. For example, minor alternate- swainsonine, was inhibited in PBL as evidenced by a decrease in processing pathways exist (1 18-120), one of which transfers an leucoagglutinin (L-PHA) binding after 5 days of treatment. Oligo- unusual form of Man5GlcNAc2 from dolichol to nacient glyco- mannosides in patient urine increased 5-10-fold over the 5 days of proteins in the endoplasmic reticulum (Fig. 1). This circumvents treatment, indicating that tissue lysosomal a-mannosidases were the need for a-mannosidase II in the biosynthesis of complex- also blocked by swainsonine. Urine oligomannoside accumulation type carbohydrates. Thus, it is possible that swainsonine treat- reached steady state at 3 days, approximately 1 day after serum ment of cancer patients may result in selection for tumor cells drug levels reached steady state. The fraction of HLA-DR-positive with enhanced use of an alternate processing pathway. How- cells in PBL increased following 5 days of swainsonine treatment, ever, it may be possible to minimize drug resistance by using an effect similar to that observed for PBL from normal subjects inhibitors acting on different enzymes in the pathway such as cultured with swainsonine. In summary, swainsonine treatment GlcNAc-TI where the alternate pathway will also be blocked. resulted in moderate toxicity in cancer patients with normal pre- On the other hand, the alternate processing pathway allows treatment liver function when administered i.v. at dosages that some normal processing in swainsonine-treated cells, and com- inhibit both Golgi a-mannosidase II and lysosomal a-mannosi- plete blockage of carbohydrate processing may be more toxic in dases. In patients with liver metastases and/or dysfunction, signif- vivo than the selective block produced by swainsonine. icant toxicity occurred resulting in one treatment-related death. We are currently studying chronic p.o. administration of swainsonine in Other Potential Toxicities of CPIs cancer patients selected by criteria similar those applied in our first The serum half-life of many glycoproteins is dependent on study. Patients are receiving swainsonine p.o. (50, 150, 300, and glycosylation and is an important factor in the in vivo bioactivity 600 p.g/kg twice weekly). A dose-dependent trend in hepatic dys- of cytokines such as erythropoietin (121) and peptide hormones function and fatigue have been determined as the dose-limiting (122). For example, serum transferrmn-bearing swainsonine-in- toxicities. L-PHA lectin binding to PBL is significantly supressed duced hybrid-type structures have a serum half-life of 14 h at the i50-pg/kg dose level, and saturation of tissue lysosomal compared to 24 h for asialotransferrmn and 40 h for the normal a-mannosidases occurs between 150 and 300 i.gfkg twice weekly. sialylated biantennary form of transferrin (123). The biological The details of this study will be published later. and clinical consequences of this altered glycoprotein clearance with swainsonine treatment remains to be determined. Considerations in the Design of New CPIs The desired properties of CPIs to be useful as anticancer agents include: (a) specificity for a Golgi-processing enzyme(s);

(b) low K1 for inhibition of the target enzyme in vitro and in cell ‘ 0. Hindsgaul and J. W. Dennis, unpublished results.

Swainsonine does not appear to be mutagenic in cell cul- immune incompetent states, e.g. , AIDS patients, may also ture, and does not cause acute liver cell damage in animals in benefit from the apparent hematorestorative action of swain- sharp contrast to the pyrrolizidine alkaloids, a structurally re- sonine. However, the relative importance of swainsonine’s lated class of compounds. The pyrrolizidine alkaloids are found action on tumor cells versus host immunity is not yet clear, in tansy ragwort (Senecio jacobaea), and their consumption by and may vary for different cancers. In a more general context, rats and chickens results in hepatotoxicity associated with de- results to date suggest that CPIs that block a-mannosidase II pletion of liver retinol stores (124). In chickens fed a diet or enzymes adjacent in the pathway may be useful anticancer supplemented with tansy ragwort, the liver histopathology can agents. If this is indeed the case, inhibitors of G1cNAc-TV as be prevented by a vitamin A supplement, suggesting that its well as other glycosyltransferases may also be developed for depletion is a significant factor in hepatotoxicity (124). Serum the treatment of cancer patients. retinol levels were examined in nine patients in our study, and drug-related depletion was noted in four. In a few patients being Acknowledgments treated on our current p.o. swainsonine study, depletion of We thank Dr. Lou Siminovitch for helpful suggestions. We also serum retinol has not been observed. In subsequent clinical thank Zofia Kryzek and Cathy Bedlington for secretarial assistance. studies of swainsonine, the effect of vitamin A supplementation on hepatotoxicity should be examined. References A subset of patients with the genetic disorder called heredi- 1. Hakomori, S-I. Aberrant glycosylation in tumors and tumor-associ- table erythroblastic multinuclearity associated with positive acidi- ated carbohydrate antigens. Adv. Cancer Res., 52: 257-331, 1989. fled serum shows reduced expression of Golgi a-mannosidase II in 2. Warren, L., Buck, C. A., and Tuszynski, G. P. Glycopeptide changes PBL (125). Glycosylation in lymphocytes of these patients with and malignant transformation: a possible role for carbohydrate in ma- hereditable erythroblastic multinuclearity associated with positive lignant behavior. Biochim. Biophys. Acta., 516: 97-127, 1978. acidified serum resembles that observed in swainsonine-treated 3. Dennis, J. W. Changes in glycosylation with malignant transformation and tumor progression. In: M. Fukuda (ed.), Cell Surface Carbohydrates cells (126), and these patients suffer from chronic anemia. It needs and Cell Development, pp. 161-194, Boca Raton: CRC Press, 1991. to be determined whether similar findings will occur in patients 4. Varki, A. Biological roles of oligosaccharides: all of the theories are treated chronically with swainsonine. correct. Glycobiology, 3: 97-130, 1993. It may be necessary to eliminate certain plant lectins such 5. Kornfeld, R., and Kornfeld, S. Assembly of asparagine-linked oh- as gliadin from the diet of CPI-treated patients to prevent gosaccharides. Annu. Rev. Biochem., 54: 631-664, 1985. gastrointestinal toxicity. Gliadin, a wheat protein mixture, 6. Schachter, H. Biosynthetic controls that determine the branching and causes atrophy of the enterocyte villi in the small intestine, microheterogeneity of protein-bound oligosaccharides. Biochem. Cell. hyperplasia of cryptic cells, and lymphatic infiltration in patients Biol., 64: 163-181, 1986. with celiac disease. Celiac disease can be induced in rats fed 7. Roth, J. Subcellular organization of glycosylation in mammalian cells. Biochim. Biophys. Acta., 7: 405-436, 1987. gliadin plus Astragalus lentiginosus as a source of swainsonine, 8. Rademacher, T. W., Parekh, B., and Dwek, R. A. Glycobiology. although either supplement alone has no effect on intestinal villi Annu. Rev. Biochem., 57: 785-838, 1988. (127). Proteins in gliadin appear to have high mannose-binding 9. Yamashita, K., Tachibana, Y., Ohkura, T., and Kobata, A. Enzymatic lectin activity, and show increased binding to intestinal villi in basis for the structural changes of asparagine-linked sugar chains of rats fed A. lentiginosus (127). membrane glycoproteins of baby hamster kidney cells induced by polyoma transformation. J. Biol. Chem., 260: 3963-3969, 1985. Summary 10. Dennis, J. W., Kosh, IC, Bryce, D-M., and Breitman, M. Oncogenes conferring metastatic potential induce increased branching of Mn-linked Preclinical studies on swainsonine have now been rein- ohigosaccharides in rat2 fibroblasts. Oncogene, 4: 853-860, 1989. forced by the antitumor and biological effects seen in patients 1 1. Miyoshi, E., Nishikawa, A., Ihara, Y., Gu, J., Sugiyama, T., with advanced malignancy. The studies in mice suggest that Hayashi, N. Fusamoto, H., Kamada, T., and Taniguchi, N. N-acetylglu- CPIs in general may be useful in a wide spectrum of malignan- cosaminyltransferase III and V messenger RNA levels in LEC rats cies in both the metastatic and (neo)adjuvant clinical setting. during hepatocarcinogenesis. Cancer Res., 53: 3899-3902, 1993. Therefore, an opportunity exists for developing and testing both 12. Easton, E. W., Blokland, I., Geldof, A. A., Rao, B. R., and van den more potent and less toxic antitumor inhibitors than swainso- Eijnden, D. H. The metastatic potential of rat prostate tumor variant R3327-MatLyLu is correlated with an increased activity of N-acetylglu- nine. L-PHA binding on PBL provides a simple means to cosaminyltransferase III and V. FEBS Lett., 308: 46-49, 1992. determine the activity of new CPIs. Furthermore, the assay 13. Dennis, J. W., and Laferte, S. Oncodevelopmental expression of could readily be applied to sequential biopsies of patients’ -GlcNAc 31-6Man al-6Man 31-6 branching of Asn-hinked ohigosac- tumors in order to correlate L-PHA binding in the tumors to the charides in human breast carcinomas. Cancer Res., 49: 945-950, 1989. biological effects of the inhibitors. 14. Fernandes, B., Sagman, U., Auger, M., Demetrio, M., and Dennis, The anticancer activity of swainsonine in mice has been J. W. 31-6 branched oligosaccharides as a marker of tumor progression shown to be synergistic with IFN-a and IL-2 (39, 48, 79), and in human breast and colon neoplasia. Cancer Res., 51: 718-723, 1991. as such, CPIs may be used in combination with other bio- 15. Lemaire, S., Derappe, C., Michalski, J. C., Aubery, M., and Ned, D. logical agents or existing anticancer therapies. The immune Expression of 31-6-branched N-linked ohigosaccharides is associated with activation in human T4 and T8 cell populations. J. Biol. Chem., stimulatory activity of swainsonine in mice suggests that the 269: 8069-8074, 1994. drug may also be used to enhance recovery of bone marrow 16. van den Eijnden, D. H., Koenderman, A. H. L., and Schiphorst, and immune function in patients following chemotherapy. W. E. C. M. Biosynthesis of blood group I-active polylactosaminogly- Bone marrow transplant patients and patients with other cans. J. Biol. Chem., 263: 12461-12465, 1988.

17. Yousefi, S., Higgins, E., Doaling, Z., Hindsgaul, 0., Pollex-Kruger, 36. Elbein, A. D. Inhibitors of the biosynthesis and processing of N-linked A., and Dennis, J. W. Increased UDP-GlcNAc:Gal 31-3GalNAc-R oligosaccharide chains. Annu. Rev. Biochem., 56: 497-534, 1987. (GhcNAc to GalNAc) 31-6 N-acetylglucosaminyltransferase activity in 37. Fuhrmann, U., Bause, E., and Ploegh, H. Inhibitors of oligosaccha- transformed and metastatic murine tumor cell lines: control of polylac- ride processing. Biochim. Biophys. Acta., 825: 95-1 10, 1985. tosamine-synthesis. J. Biol. Chem., 266: 1772-1783, 1991. 38. Humphries, M. J., Matsumoto, K., White, S. L., and Olden, K. 18. Heffernan, M., Lotan, R., Amos, B., Palcic, M., Takano, Y., and Oligosaccharide modification by swainsonine treatment inhibits pulmo- Dennis, J. W. Branching 31-6N-acetylglucosaminetransferases and nary colonization by B16-F1O murine melanoma cells. Proc. Natl. Acad. polylactosamine expression in mouse F9 teratocarcinoma cells and Sci. USA, 83: 1752-1756, 1986. differentiated counterparts. J. Biol. Chem., 268: 1242-1251, 1993. 39. Dennis, J. W. Effects of swainsonine and polyinosinic-polycytidylic 19. Kim, Y. S., Yuan, M., Itzkowitz, S. H., Sun, Q., Kaizu, T., Palekar, acid on murine tumor cell growth and metastasis. Cancer Res., 46: A., Trump, B. F., and Hakomori, S. Expression of Le” and extended Le” 5131-5136, 1986. blood group-related antigens in human malignant, premalignant, and 40. Pulverer, G., Beuth, J., Ko, H. L., Yassin, A., Oshima, Y., Rosz- nonmalignant colonic tissues. Cancer Res., 46: 5985-5992, 1986. kowski, K., and Uhhenbruck, G. Glycoprotein modifications of sarcoma 20. Itzkowitz, S. H., Yuan, M., Fukushi, Y., Palekar, A. Phelps, P. C., L-1 tumor cells by tunicamycin, swainsonine, bromoconduritol or 1-dc- Shamsuddin, A. M., Trump, B. F., Hakomori, and Kim, Y. S. Lewisx oxynynojirimycin treatment inhibits their metastatic lung colonization in and sialylated Lewis’-related antigen expression in human malignant BALB/c-mice. J. Cancer Res. Clin. Oncol., 114: 217-220, 1988. and non-malignant colonic tissue. Cancer Res., 46: 2627-2632, 1986. 41. Humphries, M. J., Matsumoto, K., White, S. L., and Olden, K. 21. Cornil, I., Kerbel, R. S., and Dennis, J. W. Tumor cell surface 31-4 Inhibition of experimental metastasis by castanospermine in mice: linked galactose binds to lectin(s) on microvascular endothehial cells and blockage of two distinct stages of tumor colonization by oligosaccharide contributes to organ colonization. J. Cell Biol., 111: 773-782, 1990. processing inhibitors. Cancer Res., 46: 5215-5222, 1986. 22. Sawada, R., Lowe, J. B., and Fukuda, M. E-selectin-dependent 42. Irimura, T., Gonzalez, R., and Nicolson, G. L. Effects of tunicamy- adhesion efficiency of colonic carcinoma cells is increased by genetic cm on B16 metastatic melanoma cell surface glycoproteins and blood- manipulation of their cell surface lysosomal membrane glycoprotein- 1 borne arrest and survival properties. Cancer Res., 41: 3411-3418, 1981. expression levels. J. Biol. Chem., 268: 12675-12681, 1993. 43. Kino, T., Inamura, N., Nakahara, K., Kiyoto, S., Goto, T., Terano, 23. Do, K-Y., Smith, D. F., and Cummings, R. D. Lamp-i in CHO cells H., Kohsaka, M., Oaki, H., and Imanaka, H. Effect of swainsonine on is a primary carrier of poly-n-acetyllactosamine chains and is bound mouse immunodeficient system and experimental murine tumor. J. preferentially by a mammalian S-type lectin. Biochem. Biophys. Res. Antibiot. (Tokyo), 38: 936-940, 1985. Commun., 173: 1123-1128, 1990. 44. Olden, K., Mohla, S., Newton, S. A., White, S. L., and Humphries, 24. Dennis. J. W., Laferte, S., Waghorne, C., Breitman, M. L, and Kerbel, M. J. Use of antiadhesive peptide and swainsonine to inhibit metastasis. R. S. 3 1-6 branching of Mn-linked ohigosaccharides is directly associated Ann. N. Y. Aced. Sci., 551: 421-441; discussion 441-442 (157 refs.), with metastasis. Science (Washington DC), 236: 582-585, 1987. 1988. 25. Lu, Y., and Chaney, W. Induction of N-acetylglucosaminyltrans- 45. Spearman, M. A., Damen, J. E., Kolodka, T., Greenberg, A. H., ferase V by elevated expression of activated or proto- Ha-ras oncogenes. Jamieson, J. C., and Wright, J. A. Differential effects of glycoprotein Mol. Cell. Biochem., 122: 85-92, 1993. processing inhibition on experimental metastasis formation by T24-H- ras transformed fibroblasts. Cancer Lett., 57: 7-13, 1991. 26. Hiraizumi, S., Takakasaki, S., Shiroki, K., and Kobata, A. Altered 46. Atsumi, S., Nosaka, C., Ochi, Y., linuma H., and Umezawa, K. protein glycosylation of rat 3Y1 cells induced by activated c-myc gene. Inhibition of experimental metastasis by the a-glucosidase inhibitor, Int. J. Cancer, 48: 305-310, 1991. 1,6-epi-cyclophelhitol. Cancer Res., 53: 4896-4899, 1993. 27. Dennis, J. W., Donaghue, T., Florian, M., and Kerbel, R. S. Ap- 47. Dennis, J. W., Koch, K., Yousefi, S., and VanderElst, I. Growth parent reversion of stable in s’ivo genetic markers detected in tumor cells inhibition of human melanoma tumor xenografts in athymic nude mice from spontaneous metastases. Nature (Lond.), 292: 242-245, 1981. by swainsonine. Cancer Res., 50: 1867-1872, 1990. 28. Dennis, J. W. Different metastatic phenotypes in two genetic classes of 48. Dennis, J. W., Koch, K., and Beckner, D. Inhibition of human HT29 WGA-resistant tumor cell mutants. Cancer Res., 46: 4594-4600, 1986. colon carcinoma growth in vitro and in vivo by swainsonine and human 29. Dennis, J. W., Carver, J. P., and Schachter, H. Asparagine-linked interferon-a2. J. NatI. Cancer Inst., 81: 1028-1033, 1989. ohigosaccharides in murine tumor cells: comparison of WGA-resistant 49. Newton, S. A., White, S. L., Humphries, M. J., and Olden, K. (WGA’) non-metastatic mutant and in a related WGA-sensitive (WGA) Swainsonine inhibition of spontaneous metastasis. J. Natl. Cancer Inst., metastatic line. J. Cell Biol., 99: 1034-1044, 1984. 81: 1024-1028, 1989. 30. Ishikawa, M., Dennis, J. W., and Kerbel, R. S. Isolation and 50. Colegate, S. M., Huxtable, C. R., and Dorhing, P. R. A spectroscopic characterization of spontaneous wheat germ agglutinin-resistant human investigation of swainsonine: an alpha-mannosidase inhibitor isolated melanoma mutants displaying remarkable different metastatic profiles in from Swainsona canescens. Aust. J. Chem., 32: 2257-2264, 1979. nude mice. Cancer Res., 48: 665-670, 1988. 51. Molyneux, R. J., and James, L. F. Loco intoxication: indolizidine 31. Lu, Y., Pelhing, J. C., and Chancy, W. G. Tumor cell surface beta alkaloid of spotted locoweed. Science (Washington DC), 216: 190-191, 1-6 branched ohigosaccharides and lung metastasis. Clin. Exp. Metas- 1981. tasis, 47-54, 1994. 12: 52. Broquist, H. P. The indolizidine alkaloids, slaframine and swainsonine: 32. Dennis, J. W., and Laferte, S. Co-reversion of a lectin-resistant contaminants in animal forages. Ann. Rev. Nutr., 5: 391-409, 1985. mutation and non-metastatic phenotype in murine tumor cells. Int. J. 53. Hagler, W. H., Jr., and Croom, W. J., Jr. Slaframine: occurrence, Cancer, 38: 445-450, 1986. chemistry, and physiological activity. In: P. R., Cheeke (ed), Toxicants 33. Dennis, J. W. Partial reversion of the metastatic phenotypes in a of Plant Origin, pp. 257-279. Boca Raton: CRC Press, 1989. WGA resistant mutant of MDAY-D2 selected with Bandeiraea sim- 54. Bowhin, T. L., and Sunkara, P. S. Swainsonine, an inhibitor of plicifolia seed lectin BShI. J. NatI. Cancer Inst., 74: 1 1 1 1-1 1 16, 1985. glycoprotein processing, enhances mitogen induced interleukin 2 pro- 34. Demetriou, M., Ivan R. Nabi, I. R., Coppohino, M., Dedhar, S., and duction and receptor expression in human lymphocytes. Biochem. Bio- Dennis, J. W. Reduced contact-inhibition and substratum adhesion in epi- phys. Res. Commun., 151: 859-864, 1988. thehial cells expressing G1cNAc-transferase V. J. Cell Biol., in press, 1995. 55. Novikoff, P. M., Touster, 0., Novikoff, A. B., and Tulsiani, D. P. 35. Yoshimura, M., Nishikawa, A., Ihara, Y., Taniguchi, S., and Tan- Effects of swainsonine on rat liver and kidney: biochemical and mor- iguchi, N. Suppression of lung metastasis of melanoma B16F1 by phological studies. J. Cell Biol., 101: 339-349, 1985. N-acetylglucosaminyltransferase III gene transfection. Proc. Am. Assoc. 56. Pritchard, D. H., Huxtable, C. R., Dorling, P. R., and Colegate, Cancer Res., 36: 84, 1995. S. M. The effect of swainsonine on the growth rate of young rats. In:

L. F. James, A. D. Elbein, R. J. Molyneux, and C. D. Warren (eds.), 75. Chotai, K., Jennings, C., Winchester, B., and Dorling, P. The uptake Swainsonine and Related Glycosidase Inhibitors, pp. 360-367. Ames, of swainsonine, a specific inhibitor of alpha-D-mannosidase, into normal IA: Iowa State University Press, 1987. human fibroblasts in culture. J. Cell. Biochem., 21: 107-1 17, 1983. 57. Bowen, D., Adir, J., White, S. L., Bowen, C. D., Matsumoto, K., 76. Dennis, J. W., White, S. L., Freer, A. M., and Dime, D. Carbonoy- and Olden, K. A preliminary pharmacokinetic evaluation of the anti- loxy analogues of the anti-metastatic drug swainsonine: activation in metastatic immunomodulator swainsonine clinical and toxic implica- tumor cells by esterases. Biochem. Pharmacol., 46: 1459-1466, 1993. tions. Anticancer Res., 13: 841-844, 1993. 77. Humphries, M. J., Matsumoto, K., White, S. L., Molyneux, R. J., 58. Elbein, A. D., Solf, R., Dorhing, P. R., and Vosbeck, K. Swainso- and Olden, K. Augmentation of murine natural killer cell activity by nine: an inhibitor of glycoprotein processing. Proc. NatI. Acad. Sci. swainsonine, a new antimetastatic immunomodulator. Cancer Res., 48: USA, 78: 7393-7397, 1981. 1410-1415, 1988. 59. Tulsiani, D. R., Harris, T. M., and Touster, 0. Swainsonine inhibits 78. Hino, M., Nakayama, 0., Tsurumi, Y., Adachi, K., Shibata, T., the biosynthesis of complex glycoproteins by inhibition of Golgi man- Terano, H., Kohsaka, M., Aoki, H., and Imanaka, H. Studies of an nosidase II. J. Biol. Chem., 257: 7936-7939, 1982. immunomodulator, swainsonine. I. Enhancement of immune response 60. Winchester, B., Al Daher, S., Carpenter, N. C., Cenci Di Bello, I., by swainsonine in vitro. J. Antibiot. (Tokyo), 38: 926-935, 1985. Choi, S. S., Fairbanks, A. J., and Fleet, G. W. J. The structural basis of 79. Humphries, M. J., Matsumoto, K., White, S. L., Molyneux, R. J., the inhibition of human a-mannosidases by azafuranose analogues of and Olden, K. An assessment of the effects of swainsonine on survival mannose. Biochem. J., 290: 743-749, 1993. of mice injected with B16-F10 melanoma cells. Clin. Exp. Metastasis, 8: 61. Dorling, P. R., Huxtable, C. R., and Colegate, S. M. Inhibition of 89-102, 1990. lysosomal alpha-mannosidase by swainsonine, an indohizidine alkaloid 80. Mohla, S., Humphries, M. J., White, S. L., Matsumoto, K., Newton, isolated from Swainsona canescens. Biochem. J., 191: 649-651, 1980. S. A., Sampson, C. C., Bowen, D., and Olden, K. Swainsonine: a new antineoplastic immunomodulator. J. Natl. Med. Assoc., 81: 1049-1056, 62. Dc Gasperi, R., Daniel, P. F., and Warren, C. D. A human hysoso- 1989. mal a-mannosidase specific for the core of complex glycans. J. Biol. Chem., 267: 9706-9712, 1992. 81. Korczak, B., Goss, P., Fernandez, B., Baker, M., and Dennis, J. W. Branching N-linked ohigosaccharides in breast cancer. Proceedings of 63. Al Daher, S., Dc Gasperi, R., Daniel, P., Hall, N., Warrern, C. D., the Symposium on Breast Cancer, 5th International Workshop on Breast and Winchester, B. The substrate-specificity of human lysosomal alpha- Cancer Research and Immunology, San Francisco, 1994. D-mannosidase in relation to genetic alpha-mannosidosis. Biochem. J., 82. Nicolson, G. Organ specificity of tumor metastasis: role of prefer- 277: 743-751, 1991. ential adhesion, invasion and growth of malignant cells at specific 64. Cenci Di Bello, I., Fleet, G., Namgoong, 5, K., Tadano, K-I., and secondary sites. Cancer Metastasis Rev., 7: 143-188, 1988. Winchester, B. Structure-activity relationship of swainsonine. Biochem. 83. Raz, A., and Lotan, R. Endogenous galactoside binding lectins: a J., 259: 855-861, 1989. new class of functional tumor cell surface molecules related to metas- 65. Morin, M. J., and Bernacki, R. J. Biochemical effects and thera- tasis. Cancer Metastasis Rev., 6: 433-452, 1987. peutic potential of tunicamycin in murine L1210 leukemia. Cancer Res., 84. Beuth, J., Ko, H. L., Oette, K., Pulverer, G., Roszkowski, K., and 43: 1669-1674, 1983. Uhlenbruck, G. Inhibition of liver metastasis in mice by blocking 66. Gibson, R., Schlesinger, S., and Kornfeld, S. The nonglycosylated hepatocyte lectins with arabinogalactan infusions and D-galactose. J. glycoprotein of vesicular stomatitis virus is temperature-sensitive and Cancer Res. Clin. Oncoh., 113: 151-155, 1987. undergoes intracellular aggregation at elevated temperatures. J. Biol. 85. Schaaf-Lafontaine, N., Balthazart, C., and Hooghe, R. J. Modifica- Chem., 254: 3600-3607, 1979. tion of blood-borne arrest properties of lymphoma cells by inhibitors of 67. Akiyama, S. K., Yamada, S. S., and Yamada, K. M. Analysis of the protein glycosylation suggests the existance of endogenous lectins. role of glycosyhation of the human fibronectin receptor. J. Biol. Chem., Carbohydr. Res., 138: 315-323, 1985. 264: 18011-18018, 1989. 86. Yagel, S., Feinmesser, R., Waghorne, C., Lala, P. K., Breitman, 68. Knight, J., Beug, H., Marshall, J., and Hayman, M. J. Abnormal M. L., and Dennis, J. W. Evidence that 31-6 branched Asn-hinked glycosylation of the env-sea oncogene product inhibits its proteolytic cleav- ohigosaccharides on metastatic tumor cells facilitate invasion of base- age and blocks its transforming ability. Oncogene, 2: 317-326, 1988. ment membranes. Int. J. Cancer, 44: 685-690, 1989. 69. Nichols, E. J., Manger, R., Hakomori, S., Herscovics, A., and 87. Seftor, R. E. B., Seftor, E. A., Grimes, W. J., Liotta, L. A., Rohrschneider, L. R. Transformation by the v-fins oncogene product: Stetler-Stevenson, W. G., Welch, D. R., and Hendrix, M. J. C. Human role of glycosylational processing and cell surface expression. Mol. melanoma cell invasion is inhibited in vitro by swainsonine and deoxy- Cell. Biol., 5: 3467-3475, 1985. mannojirimycin with a concomitant decrease in collagenase IV expres- 70. Duronio, V., Jacobs, S., Romero, P. A., and Herscovics, A. Effects sion. Melanoma Res., 1: 43-54, 1991. of inhibitors of N-linked oligosaccharide processing on the biosynthesis 88. Takano, R., Nose, M., Nishihira, I., and Kyogoku, M. Increase of and function of insulin and insulin-like growth factor-I receptors. J. 31-6-branched ohigosaccharides in human esophageal carcinomas inva- Biol. Chem., 263: 5436-5445, 1988. sive against surrounding tissue in vivo and in vitro. Am. J. Pathol., 137: 71. Montefiori, D. C., Robinson, W. E., Jr., and Mitchell, W. M. Role 1007-1011, 1990. of protein N-glycosylation in pathogenesis of human immunodeficiency 89. Korczak, B., and Dennis, J. W. Inhibition of N-linked ohigosaccha- virus type 1. Proc. NatI. Acad. Sci. USA, 85: 9248-9252, 1988. ride processing in tumor cells is associated with enhanced tissue inhib- 72. Fischer, T., Thoma, B., Scheurich, P., and Pfizenmaier, K. Glyco- itor of metalloproteinases (TIMP) gene expression. Int. J. Cancer, 53: sylation of the human interferon--y receptor. N-linked carbohydrates 634-639, 1993. contribute to structural heterogeneity and are required for ligand bind- 90. Schultz, R. M., Silberman, S., Persky, B., Bajkowski, A. S., and ing. J. Biol. Chem., 265: 1710-1717, 1990. Carmichael, D. F. Inhibition by human recombinant tissue inhibitor of 73. Keegan, A. D., and Conrad, D. H. The murine lymphocyte receptor metalloproteinases of human amnion invasion and lung colonization by for IgE V. biosynthesis, transport, and maturation of the B cell Fc, murine B16-F10 melanoma cells. Cancer Res., 48: 5539-5545, 1988. receptor. J. Immunol., 139: 1199-1205, 1987. 91. Matrisian, L. M. Metalloproteinases and their inhibitors in matrix 74. Franc, J. L., Hovsepian, S., Fayet, G., and Bouchilhoux, S. Inhibi- remodeling. Trends Genet., 6: 121-125, 1990. tion of N-linked oligosaccharide processing does not prevent the secre- 92. Ostrander, G. K., Scribener, N. K., and Rohschneider, L. R. Inhi- tion of thyroglobulin. A study with swainsonine and deoxynojirimycin. bition of v-fins-induced tumor growth in nude mice by castanospermine. Eur. J. Biochem., 57: 225-232, 1986. Cancer Res., 48: 1091-1094, 1988.

93. DeSantis, R., Santer, U. V., and Glick, M. C. NIH 3T3 cells 1 10. Diliman, R., Koziol, J., Zavanelhi, M., Beauregard, J., Halhiburton, transfected with human tumor DNA lose the transformed phenotype B., and Royston, I. Immunocompetence in cancer patients. Assessment when treated with swainsonine. Biochem. Biophys. Res. Commun., 142: by in vitro stimulation tests and quantification of lymphocyte subpopu- 348-353, 1987. lations. Cancer (Phila.), 53: 1484-1491, 1984.

94. Weinberg, R. A. Oncogenes and multistep carcinogenesis. In: R. A. 1 1 1 . Sulitzeanu, D. Immunosuppressive factors in human cancer. Adv. Weinberg (ed), Oncogenes and the Molecular Origins of Cancer, pp. Cancer Res., 60: 247-267, 1993. 307-326. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory 112. White, S. L., Nagai, T., Akiyama, S. K., Reeves, E. J., Grzegor- Press, 1989. zewski, K., and Olden, K. Swainsonine stimulation of the proliferation 95. Humphries, M. J., and Olden, K. Asparagine-hinked ohigosacchar- and colony forming activity of murine bone marrow. Cancer Commun., ides and tumor metastasis. Pharmacol. Ther., 44: 85-105, 1989. 3: 83-91, 1991. 96. Olden, K., Breton, P., Grzegorzewski, K., Yasuda, Y., Gause, B. L., 1 13. Oredipe, 0. A., White, S. L., Grzegorzewski, K., Gause, B. L., Oredipe, 0. A., Newton, S. A., and White, S. L. The potential impor- Cha, J. K., Miles, V. A., and Olden, K. Protective effects of swainsonine tance of swainsonine in therapy for cancers and immunology. Pharma- on murine survival and bone marrow proliferation during cytotoxic col. Ther., 50: 285-290, 1991. chemotherapy. J. Natl. Cancer Inst., 83: 1149-1156, 1991. 97. Dennis, J. W., and Laferte, S. Recognition of asparagine-linked 1 14. Goss, P. E., Baptiste, J., Fernandes, B., Baker, M., and Dennis, oligosaccharides on tumor cells by natural killer cells. Cancer Res., 45: J. W. A study of swainsonine in patients with advanced malignancies. 6034-6040, 1985. Cancer Res., 54: 1450-1457, 1994. 98. Kiyohara, T. K., Dennis, J. W., and Roder, J. C. Double restriction 1 15. Drug Evaluations Annual 1992, pp. 17-38. Chicago: American in NK cell recognition is linked to transmethylation and can be triggered Medical Association, 1992. by asparagine-hinked ohigosaccharides on tumor cells. Cell. Immunol., 106: 223-233, 1987. 1 16. Palcic, M. M., Heerze, L. D., Pierce, M., and Hindsgaul, 0. The use of hydrophobic synthetic glycosides as acceptors in glycosyltrans- 99. Hofer, E., DUchler, M., Fuad, S. A., Houchins, J. P., Yabe, T., and ferase assays. Glycoconj. J., 5: 49-63, 1988. Bach, F. H. Candidate natural killer cell receptors. Immunol. Today, 13: 429-430, 1992. 1 17. Khan, S. H., Crawley, S. C., Kanie, 0., and Hindsgaul, 0. A trisaccharide acceptor analog for N-acetylglucosaminyltransferase V 100. White, S. L., Schweitzer, K., Humphries, M. J., and Olden, K. which binds to the enzyme but sterically precludes the transfer reaction. Stimulation of DNA synthesis in murine lymphocytes by the drug swainsonine: immunomodulatory properties. Biochem. Biophys. Res. J. Biol. Chem., 268: 2468-2473, 1993. Commun., 150: 615-625, 1988. 1 18. Romero, P. A., and Herscovics, A. Transfer of non-glucosylated 101. Yagita, M., and Saksela, E. Swainsonine, an inhibitor of glycop- ohigosaccharides from lipid to protein in mammalian cells. J. Biol. rotein processing, enhances cytotoxicity of large granual lymphocytes. Chem., 261: 15936-15940, 1986. Scand. J. Immunol., 31: 275-282, 1990. 1 19. Lehrman, M. A., and Zeng, Y. Pleiotropic resistance to glycopro- 102. Bowhin, T. L., McKown, B. J., Kang, M. S., and Sunkara, P. 5. tein processing inhibitors in Chinese hamster ovary cells. The role of a Potentiation of human lymphokine-activated killer cell activity by novel mutation in the asparagine-linked glycosylation pathway. J. Biol. swainsonine, an inhibitor of glycoprotein processing. Cancer Res., 49: Chem., 264: 1584-1593, 1989. 4109-4113, 1989. 120. Moore, S. E. H., and Spiro, R. G. Demonstration that golgi 103. Cumming, D. A. Glycosylation of recombinant protein therapeutics: endo-a-D-mannosidase provides a glucosidase-independent pathway for control and functional implications. Glycobiology, 1: 1 15-130, 1991. the formation of complex N-linked oligosaccharides of glycoproteins. J.. Biol. Chem., 265: 13104-13112, 1990. 104. Sherblom, A. P., Sathyamoorthy, N., Decker, J. M., and Much- more, A. V. IL-2, a lectin with specificity for high mannose glycopep- 121. Takeuchi, M., and Kobata, A. Structures and functional roles of the tides. J. Immunol., 143: 939-944, 1989. sugar chains of human erythropoietins. Glycobiology, 1: 337-346, 1991. 105. Colombo, M. P., Modesti, A., Parmiani, G., and Forni, G. Local 122. Fiete, D., Srivastava, V., Hindsgaul, 0., and Baenziger, J. U. A cytokine availability elicits tumor rejection and systemic immunity hepatic reticuloendothehial cell receptor specific for 504- through granulocyte-T-lymphocyte cross-talk. Cancer Res., 52: 4853- 4GalNAc31,4GlcNAc3i, Mana that mediates rapid clearance of 4857, 1992. lutropin. Cell, 67: 1103-1110, 1991. 106. Grzegorzewski, K., Newton, S. A., Akiyama, S. K., Sharrow, S., 123. Hu, W-L., Chindemi, P. A., and Regoeczi, E. In vivo behaviour of Olden, K., and White, S. L. Induction of macrophage tumoricidal rat transferrin bearing a hybrid glycan and its interaction with macro- activity, major histocompatibility complex class II antigen (Ia) expres- phages. Biochem. Cell Biol., 70: 636-641, 1992. sion, and interleukin-1 production by swainsonine. Cancer Commun., 1: 124. Huan, J., Cheeke, P. R., Lowry, R. R., Nakaue, H. S., Snyder, 373-379, 1989. S. P., and Whanger, P. D. Dietary pyrrolizidine (Senecio) alkaloids and 107. Breton, P., Asseffa, A., Grzegorzewski, K., Akiyama, S. K., tissue distribution of copper and vitamin A in broiler chickens. Toxicol. White, S. L., Cha, J. K., and Olden, K. Swainsonine modulation of Lett., 62: 139-153, 1992. protein kinase C activity in murine peritoneal macrophages. Cancer 125. Fukuda, M. N., Masri, K. A., Dell, A., Luzzatto, L., and Moremen, Commun., 2: 333-338, 1990. K. W. Incomplete synthesis of N-glycans in congenital dyserythropoi- 108. Muchmore, A., Decker, J., Shaw, A., and Wingfield, P. Evidence etic anemia type II caused by a defect in the gene encoding a-manno- that high mannose glycopeptides are able to functionally interact with sidase II. Proc. Natl. Acad. Sci. USA, 87: 7443-7447, 1990. recombinant tumor necrosis factor and recombinant interleukin 1. Can- 126. Fukuda, M. N. HEMPAS disease: genetic defect of glycosylation. cer Res., 50: 6285-6290, 1990. Glycobiology, 1: 9-15, 1990. 109. Lucas, R., Magez, S., De Lays, R., Fransen, L., Scheerlinck, J-P., 127. Kottgen, E., Beiswenger, M., James, L. F., and Bauer, C. In vivo Rampelberg, M., Sablon, E., and Dc Baetselier, P. Mapping the lectin- induction of gliadin-mediated enterocyte damage in rats by the manno- like activity of tumor necrosis factor. Science (Washington DC), 263: sidase inhibitor, swainsonine: a possible animal model for celiac dis- 814-817, 1994. ease. Gastroenterology, 95: 100-106, 1988.

P E Goss, M A Baker, J P Carver, et al.

Clin Cancer Res 1995;1:935-944.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/1/9/935

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/1/9/935. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.