PSA - Oxford University Press

Glycobiology vol. 13 no. 6 pp. 457±470, 2003 DOI: 10.1093/glycob/cwg041

Altered glycosylation pattern allows the distinction between prostate-specific antigen (PSA) from normal and tumor origins

Rosa Peracaula2,3, Gl oria Tabares2, Louise Royle3, David J. Harvey3, Raymond A. Dwek3, Pauline M. Rudd1,3, and Rafael de Llorens1,2 2 Unitat de BioquõÂmica i Biologia Molecular, Departament de Biologia, Universitat de Girona, Campus de Montilivi s/n. 17071, Girona, Spain; and 3Glycobiology Institute, Department of Biochemistry, Oxford University, Oxford OX1 3QU, United Kingdom

Received on October 23, 2002; revised on November 29, 2002; accepted on December 12, 2002

Prostate-specific antigen (PSA) is a glycoprotein secreted by prostate epithelial cells. PSA is currently used as a marker of prostate carcinoma because high levels of PSA are indicative of a tumor situation. However, PSA tests still suffer from a lack of specificity to distinguish between benign prostate hyperplasia and prostate cancer. To determine whether PSA glycosylation could provide a means of differentiating between PSA from normal and tumor origins, N-glycan characterization of PSA from seminal fluid and prostate cancer cells (LNCaP cell line) by sequencing analysis and mass spectrometry was carried out. Glycans from normal PSA (that correspond to low and high pI PSA fractions) were sialylated biantennary complex structures, half of them being disialylated in the low pI PSA fraction and mostly monosialylated in the high pI PSA. PSA from LNCaP cells was purified to homogeneity, and its glycan analysis showed a significantly different pattern, especially in the outer ends of the biantennary complex structures. In contrast to normal PSA glycans, which were sialylated, LNCaP PSA oligosaccharides were all neutral and contained a higher fucose content. In 10±15% of the structures fucose was linked a1-2 to galactose, forming the H2 epitope absent in normal PSA. GalNAc was increased in LNCaP glycans to 65%, whereas in normal PSA it was only present in 25% of the structures. These carbohydrate differences allow a distinction to be made between PSA from normal and tumor origins and suggest a valuable biochemical tool for diagnosis and follow-up purposes. Key words: N-glycosylation/LNCaP cells/prostate cancer/ prostate-specific antigen/tumor marker Introduction Prostate-specific antigen (PSA), a member of the human kallikrein gene family of serine-proteases (Watt et al., 1

To whom correspondence should be addressed; e-mail: [email protected] or [email protected]

1986), is secreted almost exclusively by epithelial prostate cells. PSA is currently used as a marker for the detection and monitoring of prostate carcinoma (PCa) because serum PSA elevations, mostly produced as a result of disruption of prostate basement membrane by cancer cells, are indicative of PCa (Brawer, 1999; Laguna and Alivizatos, 2000; Milford Ward et al., 2001). However, PSA cannot differentiate between cell changes caused by cancer and those caused by benign changes in the prostate, such as benign prostate hyperplasia (BPH). As a result, PSA tests have a high rate of false positives, which can mean repeat needle biopsies and unnecessary surgery (Laguna and Alivizatos, 2000; Stamey et al., 2002; Bonn, 2002). Several approaches have been developed to improve the specificity of PSA tests (see Milford Ward et al., 2001, and references cited herein) to avoid diagnosing false positives as prostate cancer. The most common approach is to measure the ratio of free PSA (fPSA)/total PSA (tPSA) (free PSA plus PSA complexed to a-1-antichymotrypsin), which is lower in prostate cancer; this procedure is used in clinic diagnoses, although it still gives false positives. Other alternatives are PSA density (ratio of PSA concentration to prostate volume), the detection of pro-PSA and kallikrein 2 (Stephan et al., 2000; Peter et al., 2001), or PSA velocity (change in PSA concentration over time), but unfortunately none of these approaches has clearly shown a substantial improvement in the ability to distinguish between BPH and PCa. A deeper knowledge of the PSA molecule is required to examine whether some of its biochemical characteristics, like glycosylation, could allow the distinction of PSA from normal and tumor origins and, therefore, be useful in developing new prostate cancer markers. PSA is a glycoprotein with a single N-oligosaccharide chain attached to Asn-45. The structure of PSA glycans from the seminal fluid of healthy donors have been partially characterized by 1 H-nuclear magnetic resonance (NMR) (Belanger et al., 1995) and oligosaccharide sequencing (Okada et al., 2001). PSA glycans have been described as sialylated complex biantennary glycans, mostly core fucosylated and with a minor presence of GalNAcs in the antennae, the proportion of which seemed to be increased in the highest pI PSA fractions (Okada et al., 2001). However, the type and extent of sialylation of these glycans were not determined. Because oncogenesis leads to a remarkable alteration of cellular glycosylation, tumor-secreted glycoproteins may reflect the altered glycosylation pattern of cancer cells and are likely to be good candidates for tumor markers (Lis and Sharon, 1993; Kim and Varki, 1997; Dennis et al., 1999; Orntoft and Vestergaard, 1999). In that context, we have recently described that human pancreatic RNase 1, a

Glycobiology vol. 13 no. 6 # Oxford University Press 2003; all rights reserved.

457

R. Peracaula et al.

glycoprotein secreted mostly by pancreatic cells, has completely different oligosaccharide chains when produced by pancreatic tumor cells (FernaÂndez-Salas et al., 2000; Peracaula et al., 2003), suggesting a new possibility of using serum RNase 1 as a tumor marker of pancreatic adenocarcinoma. These results prompted us to investigate the glycosylation pattern of other tumor-secreted proteins, such as PSA. To establish whether the glycosylation of PSA, which is secreted either by normal and tumor prostate cells, could be altered in prostate tumors, we analyzed and compared the glycosylation pattern of PSA purified from seminal fluid of healthy donors with PSA produced by prostate tumor cells from a human prostate tumor cell line, LNCaP. Human prostate carcinoma cell lines are very useful models for characterizing markers for PCa and for elucidating the mechanisms of these markers. Few prostate tumor cell lines have been established; among them, the LNCaP cell line (established from a prostate cancer metastasis in the lymph node) (Horoszewicz et al., 1980) is the most commonly used model for the study of PCa because it retains some of the most prominent differentiated features of the human prostate cell, including the synthesis of the prostatespecific proteins such as PSA (Kumar et al., 2000). PSA glycans from LNCaP cells had been partially characterized by oligosaccharide polyacrylamide gel electrophoresis (PAGE) (Prakash and Robbins, 2000); tumor PSA appears to contain glycans with some triantennary structures apart from the biantennary structures that were also present in normal PSA. Chromatofocusing analysis (Wu et al., 1998) showed that LNCaP PSA had higher pI than normal PSA, indicative of a minor content of sialic acid in tumor PSA. To confirm and clearly establish the glycosylation differences between PSA from normal prostate and from LNCaP cells, a detailed analysis of the oligosaccharides of PSA of normal and tumor origin was carried out, including highperformance liquid chromatography (HPLC), exoglycosidase digestions analysis, and mass spectrometry (MS). Our results show significant differences between PSA from the normal and tumor situations. In particular, major differences, which could also be detected by specific antibodies against carbohydrate epitopes, were the complete absence of sialic acid and the presence of higher fucose and GalNAc content in LNCaP PSA. Results Characterization of oligosaccharides of PSA from normal seminal fluid PSA from seminal fluid is composed of two main enzymatically active isoforms that differ in their isoelectric point (pI) and that are separated during the purification procedure by anion-exchange chromatography. The major isoform comprises about 90±95% of overall PSA and has a pI of 6.9 (low pI PSA), and the minor one (high pI PSA) has a pI of 7.2 (Zhang et al., 1995). The difference in pI is attributed to their oligosaccharide moieties, in particular to their different levels of sialylation, as treatment of both PSA isoforms with sialidase raises the pI of both to 7.7. 458

Fig. 1. (a) Coomassie blue staining of SDS±PAGE gels of the PSA fractions from normal seminal fluid. Lanes 1 and 12, molecular markers; lanes 3±9, low pI PSA; lanes 10±11, high pI PSA. The PSA bands that were excised from the gel are boxed. (b) NP-HPLC profile of 2AB-labeled N-linked oligosaccharides obtained by in-gel digestion of low and high pI PSA fractions.

Overall profiles of PSA glycans (high and low pI PSA) from normal seminal fluid by NP-HPLC analysis Glycosylated PSA samples (low and high pI) were electrophoresed in sodium dodecyl sulfate (SDS)±PAGE gels under reducing conditions and used for glycan analysis (Figure 1a). PSA oligosaccharides were released from gel bands by in situ digestion with PNGase F and fluorescently labeled with 2-aminobenzamide (2AB). Normal-phase (NP) HPLC of PSA-labeled oligosaccharides showed that they contained glycans with glucose unit (GU) values ranging from 7.0 to 8.6 for low pI PSA and from 6.2 to 7.9 for high pI PSA (Figure 1b). The structural determination of the glycans from PSA fractions was carried out by a combination of HPLC analysis, exoglycosidase sequencing, and mass spectrometry (MS). Sequential oligosaccharide digestions of PSA glycans (low pI PSA) from normal seminal fluid First, NP-HPLC analysis of an aliquot of the glycan pool from low pI PSA before and after digestion with exoglycosidase arrays was carried out (Figure 2). Treatment of the PSA glycans with Arthrobacter ureafaciens sialidase (ABS) (broad specificity for a-linked sialic acids) indicated that nearly all the structures contained sialic acid as most of the peaks decreased in GU after sialidase

Tumor PSA carries altered sugar chains

Fig. 2. Sequential exoglycosidase digestions of N-linked oligosaccharides obtained by in-gel digestion with PNGase F of the low pI PSA. One aliquot was analyzed directly by HPLC (a), and the remaining were treated with arrays of exoglycosidase enzymes prior to HPLC analysis as indicated in each panel (b±g). See text for enzyme abbreviations and specificities. Arrows indicate the shifts of the glycans digested by the subsequent enzyme array. In a and b, the sialylated peaks are marked with one or two stars to indicate mono- or disialylated glycans, respectively. Symbol representation of glycans for this and later figures are: GlcNAc, closed square; mannose, open circle; GalNAc, closed diamond; galactose, open diamond; fucose, open diamond with a dot inside; sialic acid, closed star; beta linkage, solid line; alpha linkage, dotted line; unknown linkage, tildes; the linkage itself is indicated by the angle linking adjacent residues thus: 1-4-linkage, horizontal line; 1-3-linkage, angled (down to left) line; 1-6-linkage, angled (down to right) line; 1-2 linkage, vertical line.

459

R. Peracaula et al.

digestion (Figure 2c). To establish the linkage of these sialic acids, digestion with Streptococcus pneumoniae neuraminidase (NANI) (specificity for a2-3 sialic acids) was carried out (Figure 2b). In this case, only a subset of the sialylated structures were digested, and the profile following NANI digestion showed significant differences from that following ABS digestion, indicating that sialic acids were linked both a2-6 and a2-3 in PSA glycans. Analysis of sialylation of low pI PSA glycans from normal seminal fluid To determine the extent and type of sialylation of each of the individual PSA glycans, weak anion exchange chromatography (WAX) was carried out (Figure 3). The column was calibrated with N-glycans (neutral, monosialylated, disialylated, and trisialylated) from bovine fetuin (Figure 3a). 2AB-labeled N-glycan pools from PSA untreated (Figure 3b) or treated with NANI (Figure 3c) were separated according to charge. An increase in the proportion of monosialylated glycans was detected after digestion by NANI, indicating that some of the PSA disialylated glycans contained one sialic acid linked a2-3. Three pools were collected from each chromatographic WAX run: 0±5 min, contained neutral oligosaccharides; 7±9 min, contained monosialylated; and 10±12 min, contained disialylated structures. Subsequent analysis of each pool by NP-HPLC chromatography (data not shown) revealed which glycan structures were neutral or contained one or two sialic acids (Figure 2a and Table I). Determination of the glycan structures of low pI PSA from normal seminal fluid Digestion with ABS (Figure 2c) shows the presence of four major structures FcA2Gal2, A2Gal2, FcA2Gal1GalNAc1, and A2Gal1GalNAc1 (see Table II for abbreviations). Further exoglycosidase digestions (Figure 2d±g) confirmed the structure of these glycans. Digestion with S. pneumoniae galactosidase (SPG, specificity for galactose linked b1-4) (data not shown) in the

enzyme array produced the same results as when using bovine testes b-galactosidase (BTG, specificity for galactose linked b1-3/4 4 6) (Figure 2d), indicating that all the galactose residues are linked b1-4 to GlcNAc. Treatment with S. pneumoniae b-N-acetylhexosaminidase (SPH, specificity for GlcNAc linked b1-2 4 4 to mannose) resulted in the formation of FcGlcNAc2Man3 from the FcA2 structure, GlcNAc2Man3 from the A2 structure, and FcA1GalNAc1 and A1GalNAc1 from the glycans that contained GalNAc (Figure 2e). Most of the structures (83%) contained one fucose a1-6 linked to the core GlcNAc. The core fucose was digested when a further treatment with bovine kidney fucosidase (BKF, broad specificity for a-linked fucoses) was carried out (Figure 2f), resulting in the formation of two structures: the pentasaccharide core, GlcNAc2Man3, and A1GalNAc1 in a proportion of 75% to 25%, respectively. Finally, a further treatment with jack bean b-N-acetylhexosaminidase (JBH), which digests terminal GalNAc and GlcNAc (Figure 2g), digested the A1GalNAc1 to GlcNAc2Man3. MS-ESI-LC analysis of low pI PSA glycans from normal seminal fluid Electrospray ionization (ESI) mass spectra of the glycans from low pI PSA contained three major ions of masses 1099.9, 1245.4, and 1265.9 with compositions Hex5HexNAc4Fuc1Neu5Ac1, Hex5HexNAc4Fuc1Neu5Ac2, and Hex4HexNAc5Fuc1Neu5Ac2, respectively (Table I). These masses agree with the assignments derived from the digestion data. In conclusion, the major glycan structures of low pI PSA were complex biantennary structures containing mono- or disialylated glycans with sialic acid linked a2-6 and/or a2-3. About 25% of glycans contained GalNAc. Fucose was present linked a1-6 to the core GlcNAc in 83% of the structures.

Table I. Main PSA glycans from normal seminal plasma

GU

Hex

HexNAc

Fuc

Sialic acid

8.03

1099.9

5

4

1

1

8.55

1245.4

5

4

1

2

8.18

1265.9

4

5

1

2

1

460

Structure

Monoisotopic mass of 2AB derivative. Symbol representation of glycans is: Neu5Ac, closed star; GlcNAc, closed square; mannose, open circle; galactose, open diamond; GalNAc, closed diamond; fucose, diamond with a dot inside; beta linkage, solid line; alpha linkage, dotted line; unknown linkage, tilde; 1-4 linkage, horizontal line; 1-2 linkage, vertical line; 1-3-linkage, angled (down to left) line; 1-6 linkage, angled (down to right) line.

Fig. 3. WAX profiles of standard N-linked glycans from bovine fetuin (a); N-linked oligosaccharides released by in gel-digestion with PNGase F of low pI PSA glycans untreated (b) or treated with NANI (c).

Composition

ESI-MS1 ([M2H]2+)


Sequential oligosaccharide digestions of PSA glycans (high pI PSA) from normal seminal fluid The structural determination of the glycans from high pI PSA was carried out by NP-HPLC analysis of an aliquot of the glycan pool before and after digestion with exoglycosidase arrays (Figure 4). Treatment with ABS (Figure 4b) digested most of the glycan peaks to neutral glycans. According to the GU values of the glycans and their chromatographic shift

after ABS digestion, they correspond to structures with one sialic acid. Further treatment with BTG (Figure 4c) indicated the presence of the following neutral glycans: FcA2, FcA2GalNAc1, A2, and FcA1. These structures were corroborated with subsequent digestions with BKF and SPH (Figure 4d, e). High pI PSA glycans appear to be mostly monosialylated, which explains the higher pI of this PSA fraction. Some of the structures to which sialic acids are attached correspond

Fig. 4. Sequential exoglycosidase digestions of N-linked oligosaccharides obtained by in-gel digestion with PNGase F of the high pI PSA. One aliquot was analyzed directly by HPLC (a) and the remaining were treated with arrays of exoglycosidase enzymes prior to HPLC analysis as indicated in each panel (b±e). See text for enzyme abbreviations and specificities. In a, the sialylated peaks are marked with one star to indicate monosialylated glycans.

461

R. Peracaula et al.

to oligosacharides with incomplete antennae, that is, FcA2Gal1, FcA1Gal1, and FcA2GalNAc1. The proportion of structures containing GalNAc is around 20%, similar to that found for the low pI PSA glycans. In the same way, fucose was only detected linked a1-6 to GlcNAc. Culture of LNCaP cells in different conditions LNCaP cells were grown in presence and absence of 10% fetal bovine serum (FBS). In the latter case, cells were stimulated with dihydrotestosterone (DHT) to increase PSA expression. The conditioned media of LNCaP cells grown in 10% FBS produced 1.5 mg/ml PSA, determined by sandwich enzyme-linked immunosorbent assay (ELISA), a similar concentration to the one described by Prakash and Robbins (2000) when they stimulated these cells with DHT. In contrast, PSA concentration from LNCaP cells grown without FBS was only 0.6 mg/ml, a lower concentration explained by the androgen-dependent behavior of LNCaP cells for both growing and secreting PSA (Corey et al., 1998; Langeler et al., 1993; VaõÈsaÈnen et al., 1999). When these LNCaP cells growing in the absence of FBS were supplemented with 125 nM DHT, PSA concentration rose to 3.5 mg/ml. Purification of PSA from the tumor cell line LNCaP A new method was set up to purify to homogeneity PSA from LNCaP-conditioned media. The purification procedure was carried out slightly differently when cells were growing in the absence or presence of FBS. In the first case, only two chromatographic steps were required to purify PSA; in the latter case two extra chromatographic steps were added to obtain pure PSA separated from the other contaminating proteins contained in FBS. When purifying PSA from LNCaP media without FBS, the first step was an affinity-chromatography using a Cibacron-Blue 3GA column. This column has been reported to bind to several proteins such as enzymes with known affinities to nucleotide cofactors. Previous assays have indicated that this resin binds strongly to PSA, which was eluted by high ionic strength. A typical elution profile is shown in Figure 5a. Positive fractions that contained PSA were detected by ELISA sandwich assay that has a sensitivity of 1±5 ng/ml. These fractions were then pooled and rechromatographed in a reverse-phase column C-4 (Figure 5b). The PSA fraction eluted at 40% of acetonitrile. PSA purity after these two chromatographies was checked by SDS electrophoresis silver-stained gel and western blot (Figure 6), where a major band of about 32 kDa corresponded to the purified protein. A minor band of a slightly lower molecular weight was also detected. One hundred fifty milliliters of LNCaP medium without FBS and stimulated with DHT yielded 50 mg of pure PSA, whereas 150 ml LNCaP medium without FBS and hormone yielded 10±15 mg protein. Due to the high protein concentration in LNCaP media with FBS, two more chromatographic steps were required between the Cibacron-blue column and reverse-phase chromatography. A BioGel P-60 column enabled the separation of bovine serum albumin (BSA, molecular mass 66 kDa) contained in FBS from PSA (molecular mass of 31 kDa). 462

Fig. 5. Profiles of the purification of tumor PSA from LNCaP media cultured without FBS but with 125 nM DHT. (a) Cibacron-blue chromatography: PSA started to elute with the saline gradient (25 mM Tris-HCl/2 M NaCl, pH 7.5) and ended after 1 column volume (CV) of buffer C (25 mM Tris-HCl/1 M NaSCN, pH 7.5). (b) Reverse-phase chromatography: PSA eluted between 39±40% of buffer F (0.1% trifluoroacetic acid in acetonitrile).

Fig. 6. (a) Analysis by silver-stained SDS±PAGE of the different PSA fractions obtained during the purification of tumor PSA from LNCaP cell media cultured without FBS and with 125 nM DHT. Each lane (1±3) contained 100 ng of total protein of pooled fractions after each chromatographic step. M: molecular weight markers. 1: LNCaP cell culture media. 2: PSA positive fractions after Cibacron-blue. 3: PSA positive fractions after reverse-phase C4. (b) The same fractions (lanes 1±3) were analyzed by western blotting with a rabbit polyclonal antibody against PSA.


Then, another Cibacron-blue column purification was carried out to improve the purity of the sample obtained from the gel filtration column before the last chromatographic step. In this case, from 500 ml of LNCaP medium with 10% FBS, about 40±50 mg of pure PSA was obtained. N-terminal sequencing analysis of LNCaP PSA N-terminal sequence analysis of purified PSA from LNCaP media was performed to further characterize the protein. Most PSA molecules contained the mature protein, with a regular NH2 terminus. Pro-PSA forms (zymogen forms) were also identified in about one third of the molecules: a more abundant ÿ5 (LILSR) form and a minor ÿ7 (APLILSR) form.

Characterization of oligosaccharides of PSA from LNCaP tumor cells Overall profiles of PSA glycans from LNCaP tumor cells by NP-HPLC analysis. Oligosaccharides from the electrophoretic bands of PSA purified from LNCaP cells grown in different conditionsÐthat is, with FBS, without FBS, with presence and absence of DHT (Figure 7a,b)Ð were released by in situ digestions with PNGase F, fluorescently labeled with 2AB and subjected to NP-HPLC and matrix-assisted laser desorption ionization and time-offlight MS (MALDI-TOF MS). The NP-HPLC chromatography showed similar profiles for all the PSA fractions analyzed: eight chromatographic peaks with the same GU units, although in slightly different proportions (Figure 7c). The minor band of PSA isolated when LNCaP cells were grown in absence of FBS contained lower GU glycans (only peaks 1, 3, 4, 5, and 6 were significant) in relation to the other PSA fractions. Sequential oligosaccharide digestions of PSA glycans from LNCaP tumor cells. NP-HPLC analysis of an aliquot of the glycan pools from PSA fractions before and after digestion with exoglycosidase arrays was carried out (Figures 8 and 9). Treatment with ABS did not digest any glycans from PSA fractions, consistent with the MALDI-TOF data (Figure 10) that indicated the absence of sialic acid in PSA from LNCaP cells. Treatment with almond meal a-fucosidase (AMF, specificity for outer arm fucose linked a1-3/4 to GlcNAc) also did not digest any glycans. However, BKF, which has a broad specificity for a-fucoses, did digest the glycans, indicating the presence of fucoses linked a1-6 and/ or a1-2 (Figure 8b). Some of the glycan peaks (structures 4b, 5b, and 7) showed shifts higher than 0.4 GU, indicating that more than core fucose was present in some structures

Fig. 7. Coomassie blue staining of SDS±PAGE gels of the tumor PSA fractions from LNCaP cells. (a) Lanes 1±2, purified PSA from LNCaP cells grown in absence of serum and DHT; lanes 4±7, purified PSA from LNCaP cells grown in absence of serum and supplemented with DHT; and lane 9, molecular markers. (b) Lanes 1±5, purified PSA from LNCaP cells grown in 10% FBS and lane 7, molecular markers. The PSA bands that were excised from the gel are boxed. (c) NP-HPLC profile of 2AB-labeled N-linked oligosaccharides obtained by in-gel digestion of PSA fractions from LNCaP cells. The different glycan peaks are numbered from 1 to 8.

Fig. 8. Exoglycosidase digestion of N-linked oligosaccharides obtained by in-gel digestion with PNGase F of the tumor PSA from LNCaP cells with BKF. One aliquot was analyzed directly by HPLC (a) and the remaining was treated with BKF prior to HPLC analysis (b). Solid arrows indicate the digestion of one fucose and dotted arrows the digestion of two fucoses.

463

R. Peracaula et al.

Fig. 9. Sequential exoglycosidase digestions of N-linked oligosaccharides obtained by in-gel digestion with PNGase F of the tumor PSA from LNCaP cells. One aliquot was analyzed directly by HPLC (a), and the remaining were treated with arrays of exoglycosidase enzymes prior to HPLC analysis as indicated in each panel (b±e). See text for enzyme abbreviations and specificities.

(Figure 8b). To characterize the linkage of these fucoses and the structure of major PSA glycans, the following digestions were carried out. BTG (Figure 9b) digested some of the glycan structures (2, 4a, 5a, and 7) to FcA2, FcA2GalNAc1, and FcA2Gal1F1(1±2). Peak 3, which corresponded to FcA2GalNAc2, 464

was not digested because GalNAc can only be removed by JBH. Peaks 4b (FcA2Gal1F1[1±2]) and 5b (FcA2Gal1 GalNAc1F1[1±2]) were also not digested because their outer galactose residues had fucose a1-2 linked to them. Further digestion with SPH (Figure 9c) resulted in the formation of FcGlcNAc2Man3 from FcA2, FcA1GalNAc1


been fully determined and could provide valuable information to distinguish between PSA from normal and tumor origins.

Fig. 10. Positive ion reflectron MALDI-TOF mass spectrum of glycans from PSA from LNCaP cells before 2AB labeling. The spectrum has been processed using the maximum entropy deconvolution algorithm in the Micromass Mass-Lynx software. The numbers in parentheses refer to the peak numbers in Figure 7c and Table II. Key to ionic compositions for this figure: H hexose; N HexNAc; F deoxyhexose (fucose).

from FcA2GalNAc1, and FcA1Gal1F1(1±2) from FcA2Gal1F1(1±2). Peaks 3 (FcA2GalNAc2) and 5b (FcA2Gal1GalNAc1F1[1±2]) were undigested because GalNAc needs JBH to be digested. The proportion of glycan structures containing GalNAc was 65%. Next, digestion with BKF (Figure 9d) indicated that all glycans contained core fucose because all shifted at least 0.4 GU. Some glycans had extra fucoses linked a1-2 to galactose (FcA1Gal1F1[1±2] and FcA2GalNAc1Gal1F1 [1±2]) and were digested to GlcNAc2Man3 and A1GalNAc1, respectively. Finally, a further JBH treatment (Figure 9e) digested glycans with GalNAcs to the pentasaccharide core structure, GlcNAc2Man3. MALDI analysis of PSA glycans from LNCaP tumor cells MALDI analysis of glycans from tumor PSA is shown in Figure 10 and Table II. The major ions corroborated the structures described by exoglycosidase arrays analysis. Summary of PSA glycans from LNCaP tumor cells PSA glycans from LNCaP were all neutral and fucosylated. Major glycans are listed in Table II. Fucosylation was found to be core fucose in nearly all glycan structures, and about 15% contained outer-arm fucose linked a1-2 to galactose residues, which gives rise to the H2 epitope (Fuca1-2Galb14GlcNAc, Figure 8) (Oriol, 1995). The proportion of glycan structures that contain GalNAc was 65%. Discussion The biochemical properties of PSA from the prostate cancer cell line LNCaP and from normal seminal fluid have not yet

PSA from LNCaP cells contains zymogen forms About two-thirds of the PSA from LNCaP cells consisted of normal NH2 terminus PSA, and about one-third contained two zymogen forms. VaõÈsaÈnen et al. (1999) described higher amounts of the pro-PSA forms in LNCaP media that could be explained by the absence of human kallikrein 2 in their PSA preparation. Human kallikrein 2 is likely to activate any pro-PSA form into mature PSA (VaõÈsaÈnen et al., 1999; LoÈvgren et al., 1999). Different proforms of PSA in serum of prostate cancer patients have been identified as a possible diagnostic for distinguishing PCa from BPH, because proPSA forms were more abundant in PCa tissues (Mikolajczyk et al., 2000). However, the detection of this pro-PSA form in sera and its diagnostic value still remain uncertain (Peter et al., 2001). Major glycosylation differences in PSA from normal and tumor origins The glycosylation analysis of PSA from normal seminal fluid and from the prostate cell line LNCaP revealed major differences in their glycan structures. Significant changes mostly affected the outer ends of the oligosaccharide chains. PSA glycans from normal and tumor sources were both complex biantennary structures that differed in their content of GalNAc, sialic acid, and fucose, giving rise to distinct carbohydrate epitopes. Changes in GalNAc content An increase in levels of GalNAc (from 25% to 65%) was found in PSA from LNCaP cells. GalNAc is not commonly found in N-glycans of vertebrates. However, it has been identified in some glycoproteins from the pituitary gland and from other vertebrate sources, such as bovine milk, rat prolactin, or kidney epithelial cells (Manzella et al., 1996; Varki et al., 1999). b-Linked GalNAc has also been reported in melanoma tissues (Chan et al., 1991; Kuo et al., 1998); we have recently described it in some of the tumor N-glycan structures from pancreatic adenocarcinoma cells Capan-1 (Peracaula et al., 2003), suggesting that the expression of this carbohydrate is related to a malignant transformation. Differential expression in sialic acid: sialylated glycans are absent in tumor PSA One of the most interesting differences between PSA from normal and tumor origin was the content of sialic acid. In contrast to LNCaP glycans that did not contain sialic acid in their structures, sialic acid was detected in nearly all glycans from the most abundant PSA fraction (low pI PSA) from seminal fluid. Belanger et al. (1995) had already described by 1 H-NMR that the major glycoform was a disialylated biantennary complex structure, with a core fucose, FcA2Gal2Neu5Ac2(2±6). More recently, Okada et al. (2001) characterized the N-glycans of the two PSA isoforms (low and high pI PSA) from 465

R. Peracaula et al.

Table II. PSA glycans from LNCaP tumor cells Composition ID

GU

Mass MALDI-TOF-MS1

Hex

HexNAc

Fuc

Sialic acid

1

6.39

1688.6

3

5

1

0

2

6.64

1647.6

4

4

1

0

3

6.92

1891.7

3

6

1

0

4a

7.18

1850.7

4

5

1

0

4b

7.18

1793.7

4

4

2

0

5a

7.55

1809.7

5

4

1

0

5b

7.55

1996.7

4

5

2

0

7

7.87

1955.7

5

4

2

0

1485.5

3

4

1

0

1704.6

4

5

0

0

2012.7

5

5

1

0

2037.8

3

6

2

0

2069.7

5

6

0

0

2101.8

5

4

3

0

2215.8

5

6

1

0

Structure

MALDI calculated mass of [M Na] . All experimental values were within 0.2 mass units of calculated value. Structures are named as follows: A, number of antennae; Gal, number of galactoses; GalNAc, number of GalNAc; Fc, fucose linked a1-6 to core GlcNAc; F, number of outer-arm fucoses, with the fucose linkages in brackets. Symbol representation of glycans is: GlcNAc, closed square; mannose, open circle; galactose, open diamond; GalNAc, closed diamond; fucose, diamond with a dot inside; beta linkage, solid line; alpha linkage, dotted line; unknown linkage, tilde; 1-4 linkage, horizontal line; 1-2 linkage, vertical line; 1-3-linkage, angled (down to left) line; 1-6 linkage, angled (down to right) line. 1

seminal fluid but without analyzing their type and extent of sialylation. Here we carried out a detailed analysis of the sialic acids. Interestingly, in addition to sialic acid± linked a2-6, we also detected sialic acid±linked a2-3 in some of the disialylated structures of low pI PSA, indicating the activity of more than one sialyltransferase on PSA (Varki et al., 1999). The different pI of both PSA isoforms is indicative of their content of sialic acid. Low pI PSA had about 50% of each, mono- and disialylated structures, whereas high pI PSA only contained monosialylated structures. LNCaP PSA was reported to have a higher pI than normal PSA from seminal plasma (Huber et al., 1995; Wu et al., 1998; Herrala et al., 1998). These differences are consistent with the lack of sialic acid we described for LNCaP PSA 466

glycans. Similar differences in PSA pIs have been reported between serum PSA of benign hyperplasia and prostate tumor patients, with higher pIs in the latter case (Huber et al., 1995), suggesting that the glycosylation changes described in PSA from prostate tumor cells may be reflected in serum PSA. Differential expression in fucose: H2 antigen is present only in tumor PSA Fucose content was also altered between PSA from normal seminal fluid and prostate tumor cells. Most normal PSA glycans (83%) contained core fucose linked a1-6 to GlcNAc. This proportion was increased in PSA LNCaP glycans where nearly all structures were core fucosylated.


Moreover, 10±15% of the glycan structures from LNCaP PSA had anotherfucoselinked a1-2 to anouter-arm galactose, giving rise to the H2 epitope (Fuca1-2Galb1-4GlcNAc) (Oriol, 1995). The presence of the H2 epitope agrees with the reported expression of a1-2-fucosyltranferase in LNCaP cells (Marker et al., 2001; Chandrasekaran et al., 2002). In these studies, Marker et al. (2001) suggest that the activity of a1-2-fucosyltranferase may contribute to pathological prostatic growth. Thus, as it had been described for the lack of sialic acid in prostate tumor cells, the H2 epitope could be expressed in the tumor situation and be used for tumor PSA identification. Changes in glycosylation, in particular fucosylation and sialylation, have been reported in prostate tumor tissues. By immunohystochemistry studies, carbohydrate antigens of Lewis class II, especially Lewisy , known to be minimal or absent in benign secretory epithelial cells, are highly expressed in tumor tissues (Martensson et al., 1995; Zhang et al., 1997; Culig et al., 1998). In addition, the up-regulation of sialyl Lewisx correlates with poor prognosis of the tumor (metastasic prostate cancer) (Jorgensen et al., 1995). Lectin studies on tumor tissues have also revealed a different glycosylation pattern on their N- and O-linked glycoproteins (Arenas et al., 1999). A strong expression of fucose residues detected by the lectins Ulex europaeus agglutinin and Aleuria aurantia agglutinin staining was described for N-linked glycoproteins of PCa tissues. Some of these changes could be reflected in prostate tumor±secreted glycoproteins such as PSA, which does present a higher fucose content, and could be used for identifying tumor PSA. In conclusion, the need to identify PSA when it originates from tumor cells to distinguish between benign and malign prostatic pathologies has prompted us to analyze the glycosylation pattern of normal and tumor PSA. The differences reported, especially those referring to the content of sialic acid and fucose, suggest a possible way to discriminate between both situations, although further investigations are needed to reveal whether these glycosylation differences could be reflected in serum PSA and their usefulness for diagnosis purposes. Materials and methods Sandwich ELISA for the detection and quantification of PSA Biotinylated mouse monoclonal antibodies M-30 against fPSA or M-36 against tPSA from Roche Diagnostics (Basel, Switzerland) were diluted at 4 mg/ml and 2 mg/ml, respectively, in ELISA buffer (phosphate buffered saline [PBS], 1% BSA, 0.05% Tween-20) and bound to streptavidincoated microplates for 30 min at room temperature. After washing the plates with saline (0.9% NaCl solution)-Tween 0.05%, 20 ml of sample were tested over a final volume of 120 ml ELISA buffer for 1 h at room temperature. Plates were then washed and the second antibody, peroxidase conjugated mouse MAb M66 anti PSA (Roche) diluted 1:6000 in ELISA buffer, was added and allowed to stand for 1 h at room temperature. Colorimetric detection was carried out using BluePeroxidase substrate soluble (Roche), and the absorbance was read at a wavelength of 450 nm

with a reference of 620 nm. The detection range of this assay is 0±25 ng/ml for fPSA and tPSA using a PSA standard from the Cobas Core kit (Roche). PSA from human seminal fluid (low pI PSA and high pI PSA) were purchased from Lee Scientific (St. Louis, MO). Electrophoresis and western blotting Electrophoresis in 12% SDS±polyacrylamide mini-gels was performed at room temperature according to the method of Laemmli (1970). The gels were silver stained following the method of Blum et al. (1987). Molecular mass standards were obtained from Invitrogen (Carlsbad, CA). All samples were reduced with 5% 2-mercaptoethanol before analysis. Analysis of PSA was performed following standard western blot protocols. PSA samples were electrophoresed in a 12% SDS±PAGE gel and transferred to a polyvinylidene difluoride membrane (Millipore, Bedford, MA) at a constant voltage of 30 V, overnight at 4 C, in 192 mM glycine/ Tris 25 mM/methanol 20%. Filters were blocked in 3% (w/v) nonfat milk and 0.1% Tween in Tris-buffered saline and incubated for 1 h with a rabbit polyclonal antibody against PSA (Dako, Glostrup, Denmark) diluted 1:2000 in blocking buffer. Secondary antibody, peroxidase-conjugated goat anti-rabbit IgG (Pierce, Rockford, IL) was added at 1/10,000 and incubated for 1 h. Detection was performed using a Super Signal West Dura chemiluminescence kit (Pierce). Cell culture conditions LNCaP cells (ATCC CRL-1740) were a generous gift from Dr. A. Berenguer of the Department of Urology (Hospital Universitario de Getafe, Madrid, Spain). LNCaP cells (106 ) at passage 70±75 were cultured and maintained in RPMI 1640 with glutamine (Gibco, Paisley, Scotland, United Kingdom) supplemented with 10% FBS (Linus [Cultek], Madrid, Spain) and 50 mg/ml gentamicin (Gibco). All cell cultures were incubated at 37 C in a humidified atmosphere (95%) and 5% CO2 for 2 or 3 days. The supernatant was removed, and the cells were washed with PBS. LNCaP cells were grown in one of the following media for 2 or 3 days: (1) RPMI 1640 with glutamine supplemented with 10% FBS; (2) RPMI 1640 with glutamine without FBS and supplemented with 125 nM DHT (Sigma, St. Louis, MO); or (3) RPMI 1640 with glutamine without FBS and without DHT. The different media were collected, centrifuged, and stored at ÿ20 C. Chromatographic purification of PSA from LNCaP cells medium grown with FBS Cell culture media was filtered through a 0.22-mm membrane (Gelman-Pall, Ann Arbor, MI, ) and concentrated 10 times with a tangential filtration device by using a 5000-Da cut-off polysulfone membrane (Millipore). The concentrated medium was dialyzed against buffer A: 25 mM Tris-HCl/20 mM NaCl, pH 7.5, by using 3500 Da cut-off tubing (Spectra-Por, Rancho DomõÂnguez, CA, ), and was loaded into a Cibacron-Blue 3GA (Sigma) affinity column 467

R. Peracaula et al.

at a flow rate of 0.2 ml/min that had been preequilibrated using the same buffer. The unbound protein was eluted with 4 column volumes (CVs) of buffer A at a flow rate of 0.5 ml/min. PSA fractions were eluted with a 25 mM±2 M NaCl linear gradient (4 CVs), followed by 3 CVs of buffer B: 25 mM Tris-HCl/2 M NaCl, pH 7.5, and 3 CVs of buffer C: 25 mM Tris-HCl/1 M NaSCN, pH 7.5. Five- to ten-milliliter fractions were collected after the gradient start. The fractions collected were tested by sandwich ELISA, and PSA-containing fractions were pooled and freeze-dried. The next chromatographic step was a gel filtration chromatography in Biogel P60 (BioRad, Hercules, CA). The buffer used for sample dilution, column equilibration, and protein elution is buffer D: 50 mM Tris-HCl/200 mM NaCl, pH 7.5. The flow rate was maintained at 0.3 ml/min. Two-milliliter fractions were collected and assayed for PSA presence by sandwich ELISA. Positive fractions for PSA were pooled, dialyzed against water, and freeze-dried. The next chromatographic step was Cibacron-Blue column 1 ml (Amersham Pharmacia, Little Chalfont, United Kingdom) in an HPLC system (AKTA, Amersham Pharmacia), using the same buffers (A, B, and C) and protocol as previously described. Finally, a reversed-phase chromatography column (214 TP-RP C-4; Vydac, Hesperia, CA, ) was performed in an HPLC system. The sample was dissolved in 0.5 ml Milli-Qgrade water. The starting and column equilibration buffer was 90% buffer E: 0.1% trifluoroacetic acid in Milli-Qgrade water/10% buffer F: 0.1% trifluoroacetic acid in acetonitrile. The flow rate was 0.5 ml/min. The following linear step gradient was used: 10±25 min, 10±25% eluent F; 25±75 min, 25±50% eluent F; 75±85 min, 50±100% eluent F. Fractions were tested for PSA by sandwich ELISA, and the positive ones were pooled and freeze-dried. Chromatographic purification of PSA from LNCaP cells medium grown without FBS A two-step chromatographic purification protocol was carried out after filtering the conditioned media through a 0.22-mm membrane. First, PSA was purified by Cibacronblue column and then by a reversed-phase chromatography column, following the same protocols as described. Amino-terminal sequence analyses were performed by automated Edman degradation on a Beckman LF3000 Protein Sequencer at the Proteomics facility of the Institut de Biomedicina i Biotecnologia (Universitat Aut onoma de Barcelona, Spain). Release and purification of N-linked oligosaccharides N-glycans were released from purified PSA by in situ digestion of the protein in SDS±PAGE gel bands with PNGase F, (Roche) as described earlier (KuÈster et al., 1997). Briefly, purified PSA fractions were separated by electrophoresis under reducing conditions and visualized by Coomassie staining. Bands containing the glycoprotein were excised from the gel, reduced, alkylated, and treated with PNGase F to release the N-linked glycans. Wide range molecular markers were from Sigma. 468

Fluorescent labeling of the reducing terminus of oligosaccharides Oligosaccharides were fluorescently labelled with 2AB by reductive amination (Bigge et al., 1995) using the Oxford GlycoSciences (Abingdon, Oxford, United Kingdom) Signal labeling kit. HPLC NP-HPLC was performed using a TSK-Gel Amide-80 4.6 mm 250 mm column (Tosoh Biosep, Montgomeryville, PA) on a 2690 Alliance separations module (Waters, Milford, CT) equipped with a Waters temperature control module and a Waters 474 fluorescence detector. Solvent A was 50 mM formic acid adjusted to pH 4.4 with ammonia solution. Solvent B was acetonitrile. The column temperature was set to 30 C. Gradient conditions were a linear gradient of 20±58% A, over 152 min at a flow rate of 0.4 ml/min. Samples were injected in 80% acetonitrile. Fluorescence was measured at 420 nm with excitation at 330 nm. The system was calibrated using an external standard of hydrolyzed and 2AB-labeled glucose oligomers to create a dextran ladder, as described previously (Guile et al., 1996). WAX HPLC (Guile et al., 1994) was performed using a Vydac 301VHP575 7.5 50 mm column (Anachem, Luton, Bedfordshire, United Kingdom) according to the modified methodology described by Zamze et al. (1998). Briefly, solvent A was 0.5 M formic acid, adjusted to pH 9 with ammonia solution. Solvent B was 10% (v/v) methanol in water. Gradient conditions were a linear gradient of 0±5% A over 12 min at a flow rate of 1 ml/min, followed by 5±21% A over 13 min, then 21±50% A over 25 min, 80±100% A over 5 min, then 5 min at 100% A. Samples were injected in water. Simultaneous oligosaccharide sequencing by exoglycosidase digestions All enzymes were purchased from Glyko (Novato, CA). The 2AB-labeled oligosaccharides were digested in a volume of 10 ml for 18 h at 37 C in 50 mM sodium acetate buffer, pH 5.5, using arrays of the following enzymes: ABS (EC 3.2.1.18), 1 U/ml; NANI (EC 3.2.1.18), 1 U/ml; AMF (EC 3.2.1.111), 3 mU/ml; BKF (EC 3.2.1.51), 1 U/ml; SPG (EC 3.2.1.23), 0.1 U/ml; BTG (EC 3.2.1.23), 1 U/ml; SPH (EC 3.2.1.30), 120 mU/ml; and JBH (EC 3.2.1.30), 50 U/ mL. After incubation, enzymes were removed by filtration through a protein binding nitrocellulose membrane (Pro-Spin 45mm CN filters, Radley and Co., Essex, United Kingdom), and oligosaccharides were analyzed by NP-HPLC. MALDI-TOF MS Positive ion MALDI-TOF mass spectra were recorded with a Micromass TofSpec 2E reflectron-TOF mass spectrometer (Micromass, Manchester, United Kingdom) fitted with delayed extraction and a nitrogen laser (337 nm). The acceleration voltage was 20 kV; the pulse voltage was 3200 V; the delay for the delayed extraction ion source was 500 ns. Samples were prepared by adding 0.5 ml of an aqueous solution of the sample to the matrix solution (0.3 ml of a


saturated solution of 2,5-dihydroxybenzoic acid in acetonitrile) on the stainless steel target plate and allowing it to dry at room temperature. The sample/matrix mixture was then recrystallized from ethanol (Harvey, 1993). HPLC-ESI-MS ESI LC/MS data were obtained with a Waters CapLC HPLC system interfaced with a Micromass hybrid quadrupole time-of-flight mass spectrometer fitted with a Z-spray electrospray ion source and operated in positive ion mode. A 1 150 mm microbore NP-HPLC column was packed with stationary phase material from a GlycoSep N column (Oxford GlycoSciences). The operating conditions for the mass spectrometer were: source temperature 100 C, desolvation temperature 120 C,desolvation gas flow 200 L/h, capillary voltage 3000 V, cone voltage 30 V, TOF survey scan time 1 s, mass range m/z 50±2300; TOF MS/MS scan time 1 s, survey scan 950±1600 with detection mass range m/z 50±3500, mass selection resolution about 3 Da. Acknowledgments We thank Dr. Wolfgang Hoesel for critical reading of the manuscript and A. Ferruelo for advising in the culture of LNCaP cells. R.P. gratefully thanks EMBO (European Molecular Biology Organization) for a postdoctoral short-term fellowship. G.T. is a recipient of a predoctoral fellowship from Fundaci o Dr. J. Trueta (Girona), Roche Diagnostics. This work was supported in part by the Spanish Ministerio de Educaci on y Cultura (grant SAF 98-0086 and BIO 98-0362), by Generalitat de Catalunya (grant SGR2001-295) awarded to R. de L., and by the Biotechnology and Biological Sciences Research Council. Abbreviations 2AB, 2-aminobenzamide; ABS, Arthrobacter ureafaciens sialidase; AMF, almond meal a-fucosidase; BKF, bovine kidney fucosidase; BPH, benign prostate hyperplasia; BSA, bovine serum albumin; BTG, bovine testes b-galactosidase; CV, column volume; DHT, dihydrotestosterone; ELISA, enzyme-linked immunosorbent assay; ESI, electrospray ionization; FBS, fetal bovine serum; GU, glucose units; HPLC, high-performance liquid chromatography; JBH, jack bean b-N-acetylhexosaminidase; MALDI, matrix-assisted laser desorption-ionization; MS, mass spectrometry; NMR, nuclear magnetic resonance; NP, normal-phase; NANI, Streptococcus pneumoniae sialidase recombinant in E. coli; PAGE, polyacrylamide gel electrophoresis; PBS, phosphate buffered saline; PCa, prostate carcinoma; PSA, prostate-specific antigen; SDS, sodium dodecyl sulfate; SPG, Streptococcus pneumoniae galactosidase; SPH, S. pneumoniae hexosaminidase; TOF, timeof-flight; WAX, weak anion exchange. References Arenas, M.I., Romo, E., de Gaspar, I., de Bethencourt, F.R., SanchezChapado, M., Fraile, B., and Paniagua, R. (1999) A lectin

histochemistry comparative study in human normal prostate, benign prostatic hyperplasia, and prostatic carcinoma. Glycoconj. J., 16, 375±382. Belanger, A., van Halbeek, H., Graves, H.C.B., Garndbois, K., Stamey, T.A., Huang, L., Poppe, I., and Labrie, F. (1995) Molecular mass and carbohydrate structure of prostate specific antigen: studies for establishment of an international PSA standard. Prostate, 27, 187±197. Bigge, J.C., Patel, T.P., Bruce, J.A., Goulding, P.N., Charles, S.M., and Parekh, R.B. (1995) Nonselective and efficient fluorescent labeling of glycans using 2-aminobenzamide and anthranilic acid. Anal. Biochem., 230, 229±238. Blum, H., Beier, H., and Gross, H.J. (1987) Improved silver staining of plant proteins, RNA and DNA in polyacrylamide gels. Electrophoresis, 8, 93±96. Bonn, D. (2002) Predictive value of PSA in prostate cancer is doubtful. Lancet Oncol., 3, 130. Brawer, M.K. (1999) Prostate-specific antigen: current status. CA Cancer J. Clin., 49, 264±281. Chan, A.L., Morris, H.R., Panico, M., Etienne, A.T., Rogers, M.E., Gaffney, P., Creighton-Kempsford, L., and Dell, A. (1991) A novel sialylated N-acetylgalactosamine-containing oligosaccharide is the major complex-type structure present in Bowes melanoma tissue plasminogen activator. Glycobiology, 1, 173±185. Chandrasekaran, E.V., Chawda, R., Locke, R.D., Piskorz, C.F., and Matta, K.L. (2002) Biosynthesis of the carbohydrate antigenic determinants, globo H, blood group H, and Lewis b: a role for prostate cancer cell a1,2-L-fucosyltransferase. Glycobiology, 12, 153±162. Corey, E., Brown, L.G., Corey, M.J., Buhler, K.R., and Vessella, R.L. (1998) LNCaP produces both putative zymogen and inactive, free form of prostate-specific antigen. Prostate, 35, 135±143. Culig, Z., Hittmair, A., Hobisch, A., Bartsch, G., Klocker, H., Pai, L.H., and Pastan, I. (1998) Expression of Lewis carbohydrate antigens in metastatic lesion from human prostatic carcinoma. Prostate, 26, 162±167. Dennis, J.W., Granovsky, M., and Warren, C.E. (1999) Protein glycosylation in development and disease. BioEssays, 21, 412±421. FernaÂndez-Salas, E., Peracaula, R., Frazier, M.L., and de Llorens, R. (2000) Ribonucleases expressed by human pancreatic adenocarcinoma cell lines. Eur. J. Biochem., 267, 1484±1494. Guile, G.R., Wong, S.Y., and Dwek, R.A. (1994) Analytical and preparative separation of anionic oligosaccharides by weak anionexchange high-performance liquid chromatography on an inert polymer column. Anal. Biochem., 222, 231±235. Guile, G.R., Rudd, P.M., Wing, D.R., Prime, S.B., and Dwek, R.A. (1996) A rapid high-resolution high-performance liquid chromatographic method for separating glycan mixtures and analyzing oligosaccharide profiles. Anal. Biochem., 240, 210±226. Harvey, D.J. (1993) Quantitative aspects of the matrix-assisted laser desorption mass spectrometry of complex oligosaccharides. Rapid Commun. Mass Spectrom., 7, 614±619. Herrala, A., Kurkela, R., Vihinen, M., Kalkkinen, N., and Vihko, P. (1998) Androgen-sensitive human prostate cancer cells, LNCaP, produce both N-terminally mature and truncated prostate-specific antigen isoforms. Eur. J. Biochem., 255, 329±335. Horoszewicz, J.S., Leong, S.S., Chu, T.M., Wajsman, Z.L., Friedman, M., Papsidero, L., Kim, U., Chai, L. S., Kakati, S., Arya, S.K., and Sandberg, A.A. (1980) The LNCaP cell line: a new model for studies on human prostatic carcinoma. Prog. Clin. Biol. Res., 37, 115±132. Huber, P.R., Schmid, H.P., Mattarelli, G., Strittmatter, B., van Steenbrugge, G.J., and Maurer, A. (1995) Serum free prostate specific antigen: isoenzymes in benign hyperplasia and cancer of the prostate. Prostate, 27, 212±219. Jorgensen, T., Berner, A., Kaalhus, O., Tveter, K.J., Danielsen, H., and Bryne, M. (1995) Up-regulation of the oligosachharide sialyl-Lewisx : a new prognostic parameter in metastatic prostate cancer. Cancer Res., 55, 1817±1819. Kim, Y.J. and Varki, A. (1997) Perspectives on the significance of altered glycosylation of glycoproteins in cancer. Glycoconj. J., 14, 569±576. Kumar, A., Mikolajczyk, S.D., Hill, T.M., Millar, L.S., and Saedi, M.S. (2000) Different proportions of various prostate-specific antigen (PSA)

469

R. Peracaula et al.

and human kallikrein 2 (hK2) forms are present in noninduced and androgen-induced LNCaP cells. Prostate, 44, 248±254. Kuo, C.T., Bostick, P.J., Irie, R.F., Morton, D.L., Conrad, A.J., and Hoon, D.S. (1998) Assessment of messenger RNA of b1-4-Nacetylgalactosaminyl transferase as a molecular marker for metastatic melanoma. Clin. Cancer Res., 4, 411±418. KuÈster, B., Wheeler, S.F., Hunter, A.P., Dwek R.A., and Harvey, D.J. (1997) Sequencing of N-linked oligosaccharides directly from protein gels: in-gel deglycosylation followed by matrix-assisted laser desorption/ionization mass spectrometry and normal-phase highperformance liquid chromatography. Anal. Biochem., 250, 82±101. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680±685. Laguna, P. and Alivizatos, G. (2000) Prostate specific antigen and benign prostatic hyperplasia. Curr. Opin. Urol., 10, 3±8. Langeler, E.G., van Uffelen, C.J., Blankenstein, M.A., van Steenbrugge, G.J., and Mulder, E. (1993) Effect of culture conditions on androgen sensitivity of the human prostatic cancer cell line LNCaP. Prostate, 23, 213±223. Lis, H. and Sharon, N. (1993) Protein glycosylation: structural and functional aspects. Eur. J. Biochem., 218, 1±27. LoÈvgren, J., Airas, K., and Lilja, H. (1999) Enzymatic action of human glandular kallikrein 2 (hK2). Substrate specificity and regulation by Zn2 and extracellular protease inhibitors. Eur. J. Biochem., 262, 781±789. Manzella, S.M., Hooper, L., and Baenziger, J. (1996) Oligosaccharides containing b1,4-linked N-acetylgalactosamine, a paradigm for protein specific glycosylation. J. Biol. Chem., 271, 12117±12120. Marker, P.C., Stephan, J.-P., Lee, J., Bald, L., Mather, J.P., and Cunha, G.R. (2001) Fucosyltransferase 1 and H-type complex carbohydrates modulate epithelial cell proliferation during prostatic branching morphogenesis. Dev. Biol., 233, 98±105. Martensson, S., Bigler, S.A., Brown, M., Lange, P.H., Brawer M.K., and Hakomori, S. (1995) Sialyl-Lewisx and related carbohydrate antigens in the prostate. Human Pathol., 26, 735±739. Mikolajczyk, S.D., Millar, L.S., Wang, T.J., Rittenhouse, H.G., Marks, L.S., Song, W., Wheeler, T.M., and Slawin, K.M. (2000) A precursor form of prostate-specific antigen is more highly elevated in prostate cancer compared with benign transition zone prostate tissue. Cancer Res., 60, 756±759. Milford Ward, A., Catto, J.W.F., and Hamdy, F.C. (2001) Prostate specific antigen: biology, biochemistry and available commercial assays. Ann. Clin. Biochem., 38, 633±651. Okada, T., Sato, Y., Kobayashi, N., Sumida, K., Satomura, S., Matsuura S., Takasaki, M., and Endo, T., (2001) Structural characteristics of the N-glycans of two isoforms of prostate-specific antigens purified from human seminal fluid. Biochim Biophys. Acta, 1525, 149±160. Oriol, R. (1995) ABO, Hh, Lewis and secretion: serology, genetics and tissue distribution. In Cartron, J.P. and Rouger, P. (eds), Blood cell

470

biochemistry, volume 6: molecular basis of major human blood group antigens. Plenum Press, New York, pp. 37±73. Orntoft, T.F. and Vestergaard, E.M. (1999) Clinical aspects of altered glycosylation of glycoproteins in cancer. Electrophoresis, 20, 362±371. Peracaula, R., Royle, L., Tabares, G., MallorquõÂ-FernaÂndez, G., Barrabes, S., Harvey, D.J., Dwek, R.A., Rudd, P.M., and de Llorens, R. (2003) Glycosylation of human pancreatic ribonuclease: differences between normal and tumour states. Glycobiology, 13, 227±244. Peter, J., Unverzagt, C., Krogh, T.N., Vorm, O., and Hoesel, W. (2001) Identification of precursor forms of free prostate-specific antigen in serum of prostate cancer patients by immunosorption and mass spectrometry. Cancer Res., 61, 957±962. Prakash, S. and Robbins, P.W. (2000) Glycotyping of prostate specific antigen. Glycobiology, 10, 173±176. Stamey, T.A., Johnstone, I.M., McNeal, J.E., Lu, A.Y., and Yemoto, C.M. (2002) Preoperative serum prostate specific antigen levels between 2 and 22 ng./ml. correlate poorly with post-radical prostatectomy cancer morphology: prostate specific antigen cure rates appear constant between 2 and 9 ng./ml. J. Urol., 167, 103±111. Stephan, C., Jung, K., Lein, M., Sinha, P., Schnorr, D., Loening, S.A. (2000) Molecular forms of prostate-specific antigen and human kallikrein 2 as promising tools for early diagnosis of prostate cancer. Cancer Epidemiol. Biomarkers Prev., 9, 1133±1147. VaõÈsaÈnen, V., LoÈvgren, J., Hellman, J., Piironen, T., Lilja, H., and Pettersson, K. (1999) Characterization and processing of prostate specific antigen (hK3) and human glandular kallilrein (hK2) secreted by LNCaP cells. Prostate Cancer Prostatic Dis., 2, 91±97. Varki, A., Cummings, R., Esko, J., Freeze, H., Hart, G., and Marth, J. (1999) Essentials of glycobiology. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Watt, K.W., Lee, P.J., M'Timkulu, T., Chan, W.P., and Loor, R. (1986) Human prostate specific antigen: structural and functional similarity with serine proteases. Proc. Natl. Acad. Sci. USA, 83, 3166±3170. Wu, J.T., Bradley, W.L., Grace, H.L., and Wu, L.L. (1998) Production of miligram concentrations of free prostate specific antigen (fPSA) from LNCaP cell culture: difference between fPSA from LNCaP cell and seminal plasma. J. Clin. Lab. Anal., 12, 6±13. Zamze, S., Harvey, D.J., Chen, Y.J., Guile, G.R., Dwek, R.A., and Wing, D.R. (1998) Sialylated N-glycans in adult rat brain tissueÐa widespread distribution of disialylated antennae in complex and hybrid structures. Eur. J. Biochem., 258, 243±270. Zhang, S., Zhang, H.S., Cordon-Cardo, C., Reuter, V.E., Sighal, A.K., Lloyd, K.O., and Livingston, P.O. (1997) Selection of tumor antigens as targets for immune attack using immunohistochemistry: II. Blood group related antigens. Int. J. Cancer, 73, 50±56. Zhang, W.-M., Leinonen, J., Kalkkinen, N., Dowell, B., and Stenman, U.-H. (1995) Purification and characterization of different molecular forms of prostate-specific antigen in human seminal fluid. Clin. Chem., 41, 1567±1573.