Urinary Bladder Carcinoma

Transcription

1 Blood Group Precursor T-Antic Expression in Human Urinary Bladder Carcinoma JOHN S. COON, M.D., PH.D., RONALD S. WEINSTEIN, M.D AND JACK L. SUMMERS, M.D. Coon, John S., Weinstein, Ronald S., Summers, Jack L.: Blood group precursor T-antigen expression in human urinary bladder carcinoma. Am J Clin Pathol 77: , Tumor deletion of the ABH blood group antigens (BGAg) heralds an unfavorable prognosis in human bladder cancer. The T antigen (Thomsen-Friedenreich antigen: TAg), a precursor of other BGAg, has previously been found in malignant but not most normal cells, in which the TAg is cryptic but can be unmasked with neuraminidase (NMD). We investigated the prognostic significance of TAg expression in bladder cancer by staining paraffin sections with a T-specific lectin (peanut agglutinin [PNA]) immunoperoxidase technic. Seventy-two cases of low grade, low stage bladder cancer, 21 cases of high grade bladder cancer, and 68 controls were studied. All normals expressed the TAg only after NMD treatment (Cryptic TAg+). The Grade HI cancers, all invasive, either expressed the TAg (TAg+) (67%) or lacked T even after NMD (Cryptic TAg-) (29%), indicating that the T structure was lost rather than masked as in normal tissue. Thirty-nine per cent of 23 Grade I and II cancers which were TAg+ or Cryptic TAg subsequently became invasive (Stage B), compared with 10% of 49 Cryptic TAg+ cancers. For 32 Grade I and II, ABH BGAg negative cancers, 64% of TAg+ or Cryptic TAg cancers became invasive, compared with 17% of cancers which had Cryptic TAg+. Thus, the TAg may be a prognostically useful immunohistochemical tumor marker in bladder cancer, especially for tumors negative for ABH BGAg. (Key words: T antigen; Blood group antigens; Peanut agglutinin; Immunoperoxidase; Urinary bladder carcinoma) NUMEROUS STUDIES have examined the prognostic significance of deletion of the ABH blood group antigens (BGAg) in human carcinoma of the urinary bladder using the red cell adherence test 17 " 9 "" and, more recently, immunoperoxidase assays. 6,19,35,36 It is generally agreed that low grade tumors that express the ABH BGAg become invasive relatively infrequently, whereas those with ABH-antigen deletion do so more frequently., ' 791, ' 17 ', In the clinical setting, the usefulness of tumor BGAg testing in the individual patient is limited by the imperfect correlation between BGAg status and clinical course. The prognosis for patients with ABH BGAg-negative tumors remains particularly difficult to assess. 1 ' 7AU ' 17 ' These Received June 29, 1981; accepted for publication August 3, Supported by the Otho S. A. Sprague Memorial Institute. Address reprint requests to Dr. Coon: Department of Pathology, Rush-Presbyterian-St. Luke's Medical Center, 1753 W. Congress Parkway, Chicago, Illinois /82/0600/0692 $00.90 Department of Pathology, Rush Medical College and Rush- Presbyterian-St. Luke's Medical Center, Chicago, Illinois and Department of Urology, Northeast Ohio Medical College and Akron City Hospital, Akron, Ohio problems have stimulated the search for additional cell surface markers which may be predictive of the clinical course in individual patients with bladder cancer. Although the mechanism of carbohydrate antigen deletion in neoplastic transformation is unclear, there appears to be a general simplification of normal carbohydrate chains on glycolipids and glycoproteins. 10,21,33 This is accompanied by the appearance of precursor or cryptic antigenic specificities (cryptantigens) concomitant with disappearance of normal antigens. 10,21,33 One of the best defined precursor determinants is the T (Thomsen-Freidenreich) antigen, which is expressed in many carcinomas, but not in a number of normal epithelia such as bladder urothelium, unless treated with neuraminidase to unmask the normally cryptic T antigen. 28,29,30,31 In this immunohistochemical study, we examined the expression of the T antigen in normal human urothelium and 93 cases of human urinary bladder transitional cell carcinoma. Expression of the T antigen was limited to the carcinomas and was much more common in high grade carcinomas. When present in low grade carcinomas, the T antigen heralded a relatively poor prognosis. Patient Material Materials and Methods Normal Human Urothelium. Entirely normal human transitional epithelium is difficult to obtain at cytoscopy because of tissue damage which accompanies cystoscopy procedures, 34 or at cystectomy, since patients with carcinoma of the urinary bladder often have widespread field changes. 14 We recently demonstrated that human ureter obtained at autopsy is representative of normal transitional epithelium, and is a suitable control for BGAg studies. 5 Tissue sections used in this study were cut from blocks of well preserved ureters from a series erican Society of Clinical Pathologists 692

2 Vol. 77 No. 6 T ANTIGEN IN HUMAN BLADDER CARCINOMA 693 of 68 adults coming to autopsy at Rush-Presbyterian- St. Luke's Medical Center without genitourinary pathology. 5 Specimens were fixed in neutral formalin and paraffin embedded by standard technics. Adjacent sections were cut for routine histopathology. Urinary Bladder Transitional Cell Carcinomas. The initial diagnostic biopsy specimens from patients with transitional cell carcinoma of the urinary bladder were studied as follows: Grade I or II, Stage 0 or A, 72 cases; Grade III, Stage 0 or A, 6 cases, and Grade III, Stage B or greater, 15 cases. The biopsy procedures were performed at Rush-Presbyterian-St. Luke's Medical Center (44 cases) and at Akron City Hospital, Akron, Ohio (49 cases) between 1962 and All biopsy specimens were fixed in neutral formalin and embedded in paraffin. Follow-up data were available on all patients 3-14 years (average, 5.4 years) after the initial diagnosis of bladder carcinoma. The patients from Rush-Presbyterian-St. Luke's Medical Center were enrolled in protocols of the National Bladder Cancer Project Collaborative Group A. Of the patients with low stage disease at initial biopsy, 14 had s and invasion (defined as Stage B or greater) within four years. An additional 28 patients had one or more noninvasive s. The remaining 30 patients did not have s within the follow-up period. Immunohistologic Methods T- Antigen Detection. We employed the lectin, peanut (Arachis hypogaea) agglutinin (PNA) to detect the T- antigen structure, since PNA exhibits a high degree of specificity for the T antigen. 3,20 For detection of T-antigen expression in tissue sections, we employed a lectinanti-lectin immunoperoxidase technic. Paraffin sections were deparaffinized immediately before staining; endogenous peroxidase was blocked by incubating the sections in 1% H in absolute methanol for 30 minutes. After hydration and washing in phosphate buffered saline (PBS), the sections were incubated in purified PNA (Vector Lab) at a concentration of.001 mg/ml diluted in 10% normal goat serum (NGS) in PBS for two hours. The sections were then washed in PBS, incubated with 5% NGS in PBS for 30 minutes, and subsequently incubated with rabbit anti-pna antibody (Vector Lab.) at a dilution of 1:1000 in 1% NGS in PBS overnight. After washing, the sections were incubated with goat antirabbit IgG (Litton Bionetics), diluted 1:10 in PBS with 1% NGS for 30 minutes, washed again and incubated with rabbit PAP (Litton Bionectics) diluted 1:10 in PBS with 1% NGS for 30 minutes. Sections were washed again and incubated in.05% diaminobenzidine with.01% H in PBS, (ph 7.6). After washing with distilled water, the sections were counterstained with nuclear fast red, dehydrated and mounted with Permount. Whereas the T antigen is ordinarily not detectable on normal erythrocytes or many normal epithelia, it may be unmasked by neuraminidase (NMD) digestion. 28,29,30 ' 31 Sections from specimens in which the T antigen was not detected were digested with neuraminidase, then subjected to immunohistologic staining with PNA for detection of this cryptic T-antigen reactivity. After deparaffinizing, blocking endogenous peroxidase, and hydrating the sections as above, the sections were incubated in neuraminidase Type V (Sigma Chem. Co.) at.02 units/ml in tris saline, ph 5.5, with 10 mm CaCl 2 for one hour, at room temperature. After washing, the sections were subjected to immunoperoxidase staining for the T antigen as above, except that the NGS concentration was increased as follows: PNA was diluted in 50% NGS in PBS; the NGS was increased to 50% in the preincubation step before the anti-pna; and anti- PNA, goat anti-rabbit IgG and rabbit PAP were diluted in 10% NGS. The higher normal goat serum concentrations were used after neuraminidase digestion because of higher background staining observed after digestion. The higher protein concentration did not appear to alter the staining of specimens which had not been subjected to neuraminidase digestion. Controls for the immunoperoxidase methods were, as follows: omitting the PNA from the reaction sequence abolished specific staining; addition to. 1 M galactose to the diluent for PNA completely abolished specific staining. After NMD digestion, red blood cells and vascular endothelium always expressed the T antigen, and thus served as built-in positive controls. Connective tissue and smooth muscle were always negative, thus serving as built-in negative controls. Detection of ABH Blood Group Antigens. The immunoperoxidase procedures for detection of ABH antigens are described elsewhere and are analogous to the method used for detection of the T antigen. 6,36 For the H antigen, purified agglutinin I from Ulex europeus at.004 mg/ml (Vector Lab.) was followed by rabbit antiulex antibody (Vector Lab.) at 1:1000. For antigens A or B, sections were incubated with anti-a or B antibody (Dade, Division of American Hospital Supply) at 1:200 overnight followed by rabbit anti-human IgG (H & L chain specific) (Miles Lab.), and rabbit anti-human IgM (Miles Lab.) at 1:1000. The remainder of the reaction sequence was the same as that used for the T antigen. Controls used in these procedures are described elsewhere. Detection of M and N Blood Group Antigens. The immunoperoxidase procedures for detection of M and N antigens were analogous to those used for detection of the T antigen. Sections were incubated in rabbit anti-

3 A.J.C.P. June 1982 COON ET AL. 694

4 Vol. 77 No. 6 T ANTIGEN IN HUMAN BLADDER CARCINOMA FIG. 1 (upper). Section of Grade III transitional cell carcinoma of the urinary bladder illustrating presence of T antigen on tumor cell membranes by immunoperoxidase method (nuclear fast red counterstain). X FIG. 2 (lower). Grade II transitional cell carcinoma of the urinary bladder stained for T antigen by immunoperoxidase method. Note staining of more superficial layers of tumor, predominantly in the form of intracellular vesicles (nuclear fast red counterstain). X200. INSERT, X1300. M or N antisera (Ortho Diagnostics) at 1:10, 1:100, and 1:1000 followed by goat antirabbit IgG (1:10) and rabbit PAP (1:10). Results Morphology of T-Antigen Expression in Normal and Neoplastic Urothelium Staining for the T antigen was seen in neoplastic, but not normal urothelium. None of the 68 normals showed staining of urothelium or other tissues including erythrocytes and vascular endothelium. Some transitionalcell carcinomas stained for the T antigen, while others did not. Positive tumor cells were distinctly outlined by the dark brown diaminobenzidine H reaction product, indicating presence of the T antigen in the cell membrane (Fig. 1). Staining of the limiting membranes of fusiform intracellular vesicles was also seen, and in some specimens was the predominant pattern (Fig. 2). Background staining over smooth muscle and connective tissue was generally low. The pattern of staining in urothelial tumors was highly variable, ranging from staining of virtually every tumor cell to small foci of staining. Many papillary tumors stained predominantly in the more superficial cell layers (Fig. 2). In areas, single isolated cells showed strongly positive staining for the T antigen (Fig. 3). Tumors were scored as positive if distinct membrane staining could be discerned in at least 10% of the tumor, even if the majority of the tumor cells did not stain. All tissue blocks which gave a negative reaction for the T antigen were tested for the presence of cryptic T antigen by digestion with neuraminidase before staining with PNA-immunoperoxidase. Specimens were considered cryptic T-antigen positive if distinct membrane staining could be discerned in at least 10-20% of the tumor. All specimens of normal urothelium showed intense, uniform cell membrane staining following neuraminidase digestion. Erythrocytes and vascular endothelium were also uniformly strongly positive. Smooth muscle and connective tissue appeared negative, although some amorphous background staining, not associated with cell membranes, was seen in some specimens. The majority of tumors negative for the T antigen were positive for the cryptic T antigen (Fig. 4). Unlike normal urothelium which stained uniformly for cryptic T antigen, tumor staining for the cryptic T antigen was variable, ranging from uniform staining of membranes of all cells to staining of only small nests of cells. Erythrocytes and vascular endothelium stained strongly in all specimens. Significance of T-Antigen Expression in Bladder Carcinoma There were significant differences in the expression of T antigen in low grade (I and II) and high grade (III) transitional-cell carcinomas (Table 1). Twentyeight per cent of low grade tumors were T-antigen positive, and in 94% of those that were T-negative, cryptic T antigen could be detected after neuraminidase treatment. In contrast, 67% of high grade transitional-cell carcinomas were T-antigen positive, and of those that were negative, the cryptic T antigen could be detected in only one case. Thus, high grade tumors which were T-antigen negative did not contain detectable amounts of the T antigen masked by sialic acid, as in normal urothelium and in the majority of low grade tumors. Correlation between ABH-antigen status and histologic grade was also statistically significant (P =.01) but not as strong as that between T-antigen status and histologic grade (P <.01). The correlation between T-antigen and ABH-antigen expression in initial biopsy specimens of low grade carcinomas and the subsequent clinical course for this series of patients is summarized in Table 2. Since all Grade III, Stage 0 tumors in this series subsequently became invasive, they were excluded from this analysis because the antigenic markers did not provide additional predictive information. Patients with Grade I and II tumors that were cryptic T-antigen positive had a significantly lower (10%) incidence of subsequent invasion than patients with T-antigen positive or cryptic T-antigen negative tumors (39%) (P =.006). For this group of patients, ABH antigen expression correlated even better with clinical course than T-antigen expression (P =.0006). The T-antigen status and ABH-antigen status of the tumors were shown to be statistically unrelated, as shown in Table 3. Since expression of ABH and T antigens appeared to be independent variables, we assessed whether their predictive value 4 might be additive in selected groups of patients (Table 4 and 5). This was the case for patients with ABH BGAg negative tumors. For Grade I

5 696 COON AL. A.J.C.P.-June 1982 FIG. 3 (upper, left). Grade II transitional cell carcinoma of the urinary bladder stained for the T antigen. Most of the tumor is negative but individual cells show strong staining (nuclear fast red counterstain). X500. FIG. 4 (lower). Replicate sections of Grade I transitional cell carcinoma of the urinary bladder stained for the T antigen with and without neuraminidase pretreatment. A (left, with neuraminiadase) shows cytoplasmic and membrane staining of tumor and erythrocytes. B (right, without neuraminidase) is negative (nuclear fast red counterstain). X500.

6 Vol. 77. No. 6 T ANTIGEN IN HUMAN BLADDER CARCINOMA Table 1. Relationship of T- and ABH-Antigen Status to Histologic Grade in 93 Bladder Carcinomas Grade I & II Grade III ABH(+) 40 5 ABH(-) Cryptic T( + ) 49 1 T(+) Cryptic T(-) or II, ABH BGAg negative tumors, 64% of T-antigen positive or cryptic T-antigen negative cancers became invasive, compared with 17% of cancers which were cryptic T-antigen positive (P <.01). For low grade ABH BGAg positive tumors, invasive s are infrequent. However,. noninvasive s were twice as frequent in those tumors that were initially T- antigen positive (P <.05). Apparent Absence of M and N Antigens on Normal Urothelium Since the T antigen may be derived from erythrocyte M and N antigens by loss of terminal sialic acid, 28 " 30 we sought to determine whether M and N antigens on normal urothelium were a significant contributor to the observed cryptic T-antigen activity. Although M and/ or N antigens were easily detectable on erythrocytes and vascular endothelium on ureter specimens from each of six patients, definite staining of normal urothelium was not observed in any specimen. Discussion The T antigen was originally described as a very rare erythrocyte antigen in patients whose erythrocytes were agglutinated by normal serum of any blood group 30. The immunodominant structure has been established by inhibition studies to be terminal galactose /? 1-3 N- acetyl galactosamine 26 (Fig. 5). Although originally described as a precursor structure for the M and N blood group antigens which could be unmasked with neuraminidase treatment, it is also the core structure in O-linked oligosaccharides which have other sugar Table 2. T- and ABH-Antigen Status of the Initial Biopsy in Relation to Clinical Course in Low Grade Bladder Cancer* No Noninvasive Invasive ABH(+) ABH(-) Cryptic T(+) T(+) Cryptic T(-) * Antigen results of initial diagnostic biopsies from 72 patients with Grade I or Grade II, Stage 0 or A transitional cell carcinomas of the urinary bladder Table 3. Relationship of T-Antigen Status to ABH- Antigen Status in Bladder Transitional Cell Carcinomas* ABH antigen(+) ABH antigen(-) Cryptic T(+) T(+) Cryptic T(-) 2 7 * Initial diagnostic biopsies from 93 patients with transitional cell carcinoma of the urinary bladder of all grades and stages. The markers are statistically unrelated (P >.05). residues in place of sialic acid. 2 Also many T structures masked by sialic acid are present on glycoproteins which lack the peptide structure required for M and N antigen expression. 24,27 Springer and others have detected the T antigen in neoplastic, but not normal epithelial cells from several organs. 28 " 31 Most of these studies were performed on suspensions of cells derived from relatively small numbers of tumors of any given type. There is only one report on the occurence of T antigen in transitional-cell carcinoma, and no attempt has been made previously to relate T-antigen expression to tumor behavior. 18 Expression of the ABH BGAg, as detected by immunohistochemical methods, has been related to tumor behavior and patient prognosis in transitional-cell carcinoma of the urinary bladder l - 13,6,7 l The correlation between tumor behavior and antigen expression is imperfect, however, limiting the usefulness of the technic for the individual patient. Also, although patients with low grade, ABH positive tumors statistically have a good prognosis, the prognosis for patients presenting with ABH negative tumors is variable. This has stimulated the search for other markers with predictive value. Numerous biochemical studies have shown that there is a simplification of the polysaccharide chains on cell-surface glycolipids and glycoproteins in most malignant cells that have been examined; this results in loss of normal carbohydrate antigens, such as ABH BGAg, and appearance of or increase in various precursor structures. 10,21,33 Study of a possible relationship between expression of the well defined precursor structure, the T antigen, and behavior of human tran- Table 4. T-Antigen Status in Relation to Subsequent Clinical Course in Low Grade, ABH(-) Bladder Transitional Cell Carcinomas Cryptic T( + ) T(+) Cryptic T(-) No Noninvasive Invasive 3 7 2

7 698 COON ET AL. A.J.C.P. -June 1982 Table 5. T-Antigen Status in Relation to Subsequent Clinical Course in Low Grade, ABH(+) Bladder Cancer No Noninvasive Invasive Cryptic T(+) T(+) Cryptic T(-) sitional-cell carcinoma of the urinary bladder was therefore undertaken. Our studies showed that urinary bladder carcinomas could be classified into three categories with respect to their expression of the T antigen: (1) cryptic T-antigen positive, as in normal urothelium; (2) T-antigen positive; (3) T-antigen negative, with and without neuraminidase pretreatment. These categories are illustrated diagrammatically in Figure 5. Several relationships were found between T-antigen status and the morphology and clinical behavior of the bladder tumors. The correlation between T-antigen status and histologic grade was highly significant. Most low grade tumors were cryptic T-antigen positive, whereas high grade tumors were almost all T-antigen positive or cryptic T- antigen negative, reflecting a parallelism between morphological differentiation at the light microscopic level and biochemical differentiation of the cell membrane. This aberration in biochemical differentiation proved to be a significant predictor of biologic behavior in tumors that were indistinguishable histologically. Low grade tumors which were cryptic T-antigen positive had a significantly lower incidence of invasion and than low grade TAg(+) or cryptic TAg(-) tumors. Our data suggested that T-antigen expression is Cryptic T Antigen I NANA T Antigen Gal H= (GalNAc, not directly linked biochemically to ABH antigen expression. Patients whose tumors had both biochemical markers (BGAg(-) and TAg(+) or cryptic TAg(-)) had a significantly worse prognosis than patients whose tumors had only one abnormal marker. Patients whose tumors had neither marker had the best prognosis. Several theoretical points deserve further comment. The identity of the specific molecules that comprised the T antigen in urinary bladder carcinomas and cryptic T antigen detected in normal epithelium is uncertain. Although desialated oligosaccharide components of glycoproteins carrying M- and N-like antigenic structures are the most obvious candidates, we were unable to detect M or N antigens immunohistochemically on normal urothelium, although they were readily demonstrated on erythrocytes and vascular endothelium in tissue sections. Various simplified O-linked oligosaccharide chains on various glycoproteins may be involved. It may be noteworthy that the T-antigen structure has been described within some oligosaccharide chains that have ABH-antigen activity, 23,26 raising the possibility of a direct relationship between the disappearance of the ABH antigens and appearance of the T antigen. However, the lack of correlation between ABH- and T-antigen status in the present study argues against this as an important relationship. The molecular basis of the staining reaction in tumors that were both T-antigen and cryptic T-antigen negative is unclear but may indicate any of several defects in glycosylation of cell surface species that have been described in tumor cells. 15,24 The reason that the carbohydrate antigens detectable on tumors cells are related to the tumor's biologic behavior remains unclear, but this study demonstrates that the phenomenon is not restricted to the ABH antigens. This study illustrates the value of combining predictive markers to more accurately assess the prognosis of a patient with transitional-cell carcinoma of the urinary bladder. Additional markers such as aneuploidy, 32 marker chromosomes 32 and carcinoembryonic antigen 37 appear related to bladder cancer patient prognosis, but have not been used clinically to date. Optimal determination of patient prognosis may require integration of these and other markers into a marker profile for the individual patient which would yield a probability value for prognosis. Acknowledgment. The authors thank Ms. Josefina Fuerte for her expert technical assistance. Cryptic T Antigen Undetectable FIG. 5. Diagrammatic representation of probable structures corresponding to different T-antigen reactivities of bladder carcinomas. Abbreviations: NANA = sialic acid, Gal = D-galactose, Gal NAc = N acetyl-d-galactosamine, Ser = serine, Thr = threonine. References Bergman S, Javadpour N: The cell surface antigen A, B or O (H) as an indicator of malignant potential in Stage A bladder carcinoma. J Urol 1978; 119:49-51 Beyer T, Rearick J, Paulson J, et al: Biosynthesis of mammalian glycoproteins. Glycosylation pathways in the synthesis of the nonreducing terminal sequences. J Biol Chem 1979; 254:

8 Vol. 77 No. 6 T ANTIGEN IN HUMAN BLADDER CARCINOMA Bird G: Lectins, Blood banking, CRC handbook series in clinical laboratory science. Edited by T Greenwalt, Boca Raton, Florida CRC press, 1977, pp Cole P, Morrison A: Basic issues in population screening for cancer. J Nat Cancer Inst 1980; 64: Coon J, Weinsteih RS: Variability in the expression of the O(H) antigen in human transitional epithelium. J Urol 1981; 125: Coon J, Weinstein RS: Detection of ABH tissue isoantigens by immunoperoxidase methods in normal and neoplastic urothelium. Comparison with the red cell adherence method. Am J Clin Pathol, 1981; 76: Decenzq JM, Howard P, Irish CE: Antigenic deletion and prognosis of patients with stage A transitional cell bladder carcinoma. J Urol 1975; 114: Decenzo JM, Leadbetter GW, Jr: The interaction of host immunocompetence and tumor aggressiveness in superficial bladder carcinoma. J Urol 1976; 115: Emmot RC, Javadpour N, Bergman S, et al: Correlation of the cell surface antigens with stage and grade in cancer of the bladder. J Urol 1979; 121: Hakomori S: Structures and organization of cell surface glycolipids dependance on cell growth and malignant transformation. Biochim Bjophys Acta 1975; 417: Jakse G, Hofstadter F: Further experiences with the specific red cell adherence test (SRCA) in bladder cancer. A histological and cytological study. Eur Urol 1978; 4: Johnson JD, Lamm DL: Prediction of bladder tumor invasion with the mixed cell agglutination test. J Urol 1980; 123: Kato T: Detection of A,B and O(H) antigens in normal and neoplastic epithelium of the urinary bladder by the specific red cell adherence test (SRCA). Tohoku J Exp Med 1977; 121: Koss L: Tumors of the urinary bladder. Armed Forces Institute of Pathology. Washington, DC 1975, Krag S: A concanavalin A-resistant Chinese hamster ovary cell line is deficient in the synthesis of [ 3 H]glucosyl oligosaccharidelipid. J Biol Chem 1979; 254: Lange PH, Limas C, Fraley EE: Tissue blood group antigen and prognosis in low stage transitional cell carcinoma of the bladder. J Urol 1978; 119: Limas C, Lange P, Fraley E, et al: A, B, H antigens in transitional cell tumors of the urinary bladder. Correlation with the clinical course. Cancer 1979; 44: Limas C, Lange P: Altered reactivity for A,B,H antigen in transitional cell carcinomas of the urinary bladder. A study of the mechanisms involved. Cancer 1980: 46: Limas C, Coon J, Lange P, Weinstein RS: ABH antigens in urinary bladder carcinomas: detection and clinical applications. Bladder Cancer. AUA Monograph, Vol. I., Edited by WW Bonney and GR Prout, Jr, Baltimore, Williams and Wilkins, 1982, pp Lis H, Sharon N: Lectins: their chemistry and application to immunology. The antigens. Edited by M. Sela, New York, Academic Press 1977, ch Morre D, Kloppel T, Merritt W, et al: Glycolipids as indicators of tumorigenesis. J Supramol Struct 1978; 9: Newman A, Carlton E, Johnson S: Cell surface A, B or O(H) blood group antigens as an indicator of malignant potential in Stage A bladder carcinoma. J Urol 1980; 124: Rauvala H, Finne J: Structural similarity of the terminal carbohydrate sequences of glycoproteins and glycolipids. FEBS Lett 1979; 97: Reitman M, Trowbridge I, Kornfeld S: Mouse lymphoma cell lines resistant to pea lectin are defective in fucose metabolism. J Biol Chem 1980; 255: Richie JP, Bliite RB, Jr, Waisman J: Immunologic indicators of prognosis in bladder cancer: the importance of cell surface antigens. J Urol 1980; 123: Rolih S: Erythrocyte antigens of the MN system and related structures, A seminar on antigens on blood cells and body fluid. Edited by C Bell, Washington, American Association of Blood Banks 1980, ch Schenkel-Brunner H: Blood-group-ABH antigens of human erythrocytes. Quantitative studies on the distribution of H antigenic sites among different classes of membrane components. Eur J Biochem 1980; 104: Springer G, Desai, P: Cross reacting carcinoma-associated antigens with blood group and precursor specificities. Transplant Proc 1977;9: Springer G, Desai P, Yang H, et al: Carcinoma-associated blood group MN precursor antigens against which all humans possess antibodies. Clin Immunol Immunopath 1977; 7; Springer C, Desai P, Murthy M, et al: Precursors of the blood group MN antigens as human carcinoma associated antigens. Transfusion 1979; 19: Stoward P, Spicer S, Miller R: Histochemical reactivity of peanut lectin-horseradish peroxidase conjugate. J Histochem Cytochem 1980; 28: Summers J, Falor W, Ward R: A 10-year analysis of chromosomes in noninvasive papillary carcinoma of the bladder. J Urol 1981; 125: Watanabe K, Hakomori S: Status of blood group carbohydrate chains in ontogenesis and in oncogenesis. J Exp Med 1976; 144: Weinstein RS, Koo C, Pauli B, et al: Epithelial injury by cystoscopy fluids. Semin Oncol 1979; 6: Weinstein RS, Coon J, Summers J, et al: Prognostic value of ABH(O) antigen deletion and marker chromosomes in human urinary bladder carcinoma. (Abstract) Lab Invest 1981; 44:74 A 36. Weinstein RS, Coon J, Alroy J, et al: Tissue-associated blood group antigens in human tumors. Diagnostic Immunohistochemistry. Edited by RD Delellis, New York, Masson Inc., 1981, pp Wiley E, Mendelsohn G, Droller M, Eggleston J: Immunoperoxidase detection of carcinoembryonic antigens and blood group substance in transitional cell carcinomas of the bladder. Submitted for publication.