Lectins as probes for identification of tumor-associated antigens on urothelial and colonic carcinoma cell lines

Disruption of the FA/BRCA pathway in bladder cancer

Bladder carcinomas frequently show extensive deletions of chromosomes 9p and/or 9q, potentially including the loci of the Fanconi anemia (FA) genes FANCC and FANCG. FA is a rare recessive disease due to defects in anyone of 13 FANC genes manifesting with genetic instability and increased risk of neoplasia. FA cells are hypersensitive towards DNA crosslinking agents such as mitomycin C and cisplatin that are commonly employed in the chemotherapy of bladder cancers. These observations suggest the possibility of disruption of the FA/BRCA DNA repair pathway in bladder tumors. However, mutations in FANCC or FANCG could not be detected in any of 23 bladder carcinoma cell lines and ten surgical tumor specimens by LOH analysis or by FANCD2 immunoblotting assessing proficiency of the pathway. Only a single cell line, BFTC909, proved defective for FANCD2 monoubiquitination and was highly sensitive towards mitomycin C. This increased sensitivity was restored specifically by transfer of the FANCF gene. Sequencing of FANCF in BFTC909 failed to identify mutations, but methylation of cytosine residues in the FANCF promoter region was demonstrated by methylation-specific PCR, HpaII restriction and bisulfite DNA sequencing. Methylation-specific PCR uncovered only a single instance of FANCF promoter hypermethylation in surgical specimens of further 41 bladder carcinomas. These low proportions suggest that in contrast to other types of tumors silencing of FANCF is a rare event in bladder cancer and that an intact FA/BRCA pathway might be advantageous for tumor progression.

Inactivation of the Rb pathway and overexpression of both isoforms of E2F3 are obligate events in bladder tumours with 6p22 amplification

E2F3 and CDKAL1 are candidate genes from the 6p22 region frequently amplified in bladder cancer. Expression of E2F3 isoforms (E2F3a and b) and CDKAL1 were examined and modulated in 6p22-amplified bladder cell lines. Eight lines with amplification showed overexpression of both E2F3 isoforms and CDKAL1. shRNA-mediated knockdown of CDKAL1 had no effect on proliferation. Knockdown of E2F3a or E2F3b alone induced antiproliferative effects, with the most significant effect on proliferation being observed when both isoforms were knocked down together. As E2Fs interact with the Rb tumour suppressor protein, Rb expression was analysed. There was a striking relationship between 6p22.3 amplification, E2F3 overexpression and lack of Rb expression. This was also examined in primary bladder tumours. Array-CGH detected 6p22.3 amplification in 8/91 invasive tumours. Five were studied in more detail. Four showed 13q14.2 loss (including RB1) and expressed no Rb protein. In the fifth, 13q was unaltered but the CDKN2A locus was deleted. This tumour was negative for p16 and positive for Rb protein. As p16 is a negative regulator of the Rb pathway, its loss represents an alternative mechanism for inactivation. Indeed, a phospho-specific Rb antibody showed much Rb protein in a hyperphosphorylated (inactive) form. We conclude that inactivation of the Rb pathway is required in addition to E2F3 overexpression in this subset of bladder tumours.

FGFR3 and Ras gene mutations are mutually exclusive genetic events in urothelial cell carcinoma

Fibroblast growth factor receptor 3 (FGFR3) mutations are frequent in superficial urothelial cell carcinoma (UCC). Ras gene mutations are also found in UCC. As oncogenic activation of both FGFR3 and Ras is predicted to result in stimulation of the mitogen-activated protein kinase (MAPK) pathway, we hypothesized that these might be mutually exclusive events. HRAS mutation has been widely studied in UCC, but all three Ras gene family members have not been screened for mutation in the same sample series. We screened 98 bladder tumours and 31 bladder cell lines for mutations in FGFR3, HRAS, NRAS and KRAS2. FGFR3 mutations were present in 54 tumours (55%) and three cell lines (10%), and Ras gene mutations in 13 tumours (13%) and four cell lines (13%). These included mutations in all three Ras genes; ten in HRAS, four in KRAS2 and four in NRAS and these were not associated with either tumour grade or stage. In no cases were Ras and FGFR3 mutation found together. This mutual exclusion suggests that FGFR3 and Ras gene mutation may represent alternative means to confer the same phenotype on UCC cells. If these events have biological equivalence, Ras mutant invasive UCC may represent a novel subgroup.

Assessment by M-FISH of karyotypic complexity and cytogenetic evolution in bladder cancer in vitro

We carried out multiplex fluorescence in situ hybridization (M-FISH) and follow-up FISH studies on a large series of transitional cell carcinoma (TCC) cell lines and 2 normal urothelium-derived cell lines, several of which have not had karyotypes reported previously. M-FISH analysis, with appropriate follow-up, complements conventional cytogenetic analysis and array CGH studies, allowing a more accurate definition of karyotype. The detailed karyotypic data obtained will assist in choosing suitable cell lines for functional studies and identifies common losses, gains, breakpoints and potential fusion gene sites in TCC. We have shown changes in cell lines RT112 and DSH1 following prolonged culture, and differences in karyotype, between RT112 cultures obtained from different sources. We propose a model for the evolutionary changes leading to these differences. A comparison with the literature found other examples of differences in cell-line karyotypes between different sources. Nevertheless, several karyotypic changes were preserved between different sources of the same cell line and were also seen in more than one cell line. These may be the most important changes and include -8p, +20, 4q-, 10p-, 16p- and breaks in 8p21. We carried out a more detailed follow-up of some regions, which showed involvement of 8p breaks and losses in 15 of 16 TCC cell lines but in neither of the normal urothelium-derived cell lines. Some changes represented distal loss, whereas others were small deletions. Further study of this region is warranted.

Infrequent mutation of TRAIL receptor 2 (TRAIL-R2/DR5) in transitional cell carcinoma of the bladder with 8p21 loss of heterozygosity

Loss of heterozygosity (LOH) on 8p is a frequent event in many cancers and is often associated with more aggressive disease. Tumour necrosis factor-related apoptosis inducing ligand (TRAIL) receptor 2 (TRAIL-R2) also known as TNFRSF10B (tumour necrosis factor receptor (TNFR) super family 10b) or KILLER/DR5, a member of the TNFR family, is a promising candidate tumour suppressor gene at 8p21-22. Mutations in this gene have been identified in non-small cell lung cancer, head and neck cancer, breast cancer and non-Hodgkin's lymphoma. We carried out mutation analysis of TRAIL-R2 in bladder cancer cell lines and in primary bladder tumours. One novel protein truncating mutation was identified in a bladder cancer cell line. Our results suggest that if TRAIL-R2 is the target of LOH events in these cancers, inactivation of the remaining allele is by a mechanism other than mutation.

Mutation analysis of 8p genes POLB and PPP2CB in bladder cancer

The DNA polymerase beta gene (POLB), which encodes a DNA polymerase believed to be involved in short gap-filling DNA synthesis, has been mapped to the proximal region of 8p (8p12-p11), a region commonly deleted in bladder carcinoma and a wide variety of other neoplasms. Also mapped to this region (8p12-p11.2) is the gene encoding the beta isoform of the catalytic subunit of protein phosphatase 2A (PPP2CB), a major serine/threonine phosphatase thought to play a regulatory role in many cellular pathways. The known functions of these proteins make them good candidates for 8p tumor suppressor genes. To test this hypothesis, we assessed a series of bladder tumors and bladder tumor cell lines for sequence variation in POLB and PPP2CB. Single strand conformation polymorphism (SSCP) analysis and direct sequencing of POLB cDNA derived from cell lines and tumors, many with known deletions of proximal 8p, revealed one sequence variant that was shown to represent a normal sequence polymorphism. No tumor-specific sequence variants were identified. The promotor sequence in genomic DNA from tumors with 8p LOH was also screened by SSCP. Four polymorphisms were identified but no tumor-specific mutations were found. PPP2CB was analyzed by SSCP analysis of all 7 coding exons in genomic DNA of bladder tumors and cell lines. Polymorphisms were detected in exons 4 and 5 but no tumor-specific mutations were found. We conclude that these genes are unlikely to be the suppressor genes for bladder cancer targeted by deletions of chromosome arm 8p.

Fluorescence in situ hybridization deletion mapping at 4p16.3 in bladder cancer cell lines refines the localisation of the critical interval to 30 kb

An allelotype analysis of transitional cell carcinoma of the bladder identified loss of heterozygosity (LOH) on chromosome arm 4p in 22% of tumours. In a more detailed LOH study of 178 bladder carcinomas, a 750 kb common region of deletion was identified between the markers D4S43 and D4S127 just telomeric to the Huntington disease locus. To refine this region of deletion at 4p16.3, we have carried out detailed fluorescence in situ hybridisation (FISH) analysis of 12 bladder cancer cell lines by using a chromosome 4 centromeric probe combined with a series of cosmid probes from contigs spanning the 750 kb region of deletion. A common 30 kb region of deletion was identified at 4p16.3 in over one-third of the bladder cancer cell lines analysed. The present study has refined the localisation of the critical region of deletion from 750 kb to approximately 30 kb, providing a precise starting point for positional cloning of the gene(s) involved in bladder cancer from within a very gene-rich region on chromosome band 4p16.3. This study demonstrates that FISH can be used for fine deletion mapping of potential tumour suppressor gene regions. The utilisation of FISH analysis to map chromosomal deletions should facilitate positional cloning of other genes as bacterial artificial chromosome (BAC) and yeast artificial chromosome (YAC) contigs of the human genome are established.

The spectrum of TP53 mutations in bladder carcinoma

The mutational spectrum for the TP53 gene was investigated in a large series of bladder tumors and bladder tumor cell lines. Tumors and cell lines were screened for the presence of TP53 point mutations by single-strand conformational polymorphism analysis followed by direct sequencing. Mutations were detected in 16 of 88 (18%) tumors and 4 of 14 cell lines (28%). In total, twelve missense mutations, one nonsense mutation, three deletions, and two insertions were identified by direct sequencing. Of the thirteen point mutations sequenced, only one was a transition at a CpG site, whereas five G:C-->T:A transversions were found, suggesting a major role for exogenous mutagens in bladder tumorigenesis. Tumors were also examined for loss of heterozygosity (LOH) on chromosome arm 17p. LOH of one or more markers on 17p was detected in 31% of tumors. All eight tumors with a TP53 mutation from patients informative at TP53 had LOH, whereas nine tumors with LOH at TP53 did not have an identified mutation. Three tumors had LOH on 17p at sites distal to the TP53 locus but retained both TP53 alleles, suggesting the involvement of another tumor suppressor gene on 17p in bladder tumorigenesis in some tumors.

Reduced expression of TGF beta is associated with advanced disease in transitional cell carcinoma

The gene structure and expression of the related peptide regulatory factors TGF beta 1 and TGF beta 2 were studied in a panel of seven urothelial carcinoma cell lines and 40 transitional cell carcinomas. The latter comprised 15 grade 1, 18 grade 2 and 5 grade 3 tumours and two cases of carcinoma in situ. Control tissues included ten matched 'field' biopsies and 17 other biopsies including 11 biopsies of macroscopically normal urothelium, two of which were from patients with no history of bladder cancer. No amplification of rearrangements of either TGF beta 1 or TGF beta 2 were detected in any sample. A complex pattern of expression or the two genes was found in the urothelial cell lines. High, but variable levels of the 2.5 kb TGF beta 1 transcript were detected and lower and more variable levels of the three (4.1 kb, 5.1 kb and 6.5 kb) transcripts of TGF beta 2 were detected. Although those cell lines expressing most TGF beta 1 tended to express less TGF beta 2 transcript there was no clear-cut relationship. In comparison, no TGF beta 2 transcript was identified in any primary transitional cell carcinoma or control tissue. Markedly reduced or undetectable levels of TGF beta 1 transcript were detected in 4/15 (26%) grade 1, 5/18 (28%) grade 2 and 3/5 (60%) grade 3 tumours. There was no clear relationship to tumour stage, lymphocytic infiltration or stromal content of the tumours. Clinical review one year after the 2 year period of tumour collection showed that 6/9 (66%) of patients with tumours with reduced levels of transcript had died or had disease which was not controllable by local resection and 3/9 (33%) had developed tumour re-occurrences. In comparison, in the group with normal levels of expression of TGF beta 1, 3/18 (17%) had disease which was not controllable by local means, 9/18 (50%) had tumour re-occurrence and 6/18 (33%) had no evidence of disease. The association of reduced expression of TGF beta 1 and advanced disease was statistically significant P < 0.02 (Fisher's test). Although the sample size is small, we suggest that the loss of expression of TGF beta 1 may be a potential marker of progressive disease or prognosis in transitional cell carcinoma and warrants further study.

Human papillomavirus sequences are not detectable by Southern blotting or general primer-mediated polymerase chain reaction in transitional cell tumours of the bladder

A large series of transitional cell tumours has been screened for the presence of human papillomavirus (HPV) sequences using Southern blotting and general primer-mediated polymerase chain reaction (GP-PCR). The latter technique allows the detection of a broad spectrum of both sequenced and unsequenced HPV types using two primer pairs located in the highly conserved L1 and E1 regions of the HPV genome. No evidence for HPV infection was found in 100 transitional cell tumours, 6 cases of carcinoma in situ, 2 adenocarcinomas and a squamous carcinoma of the bladder and 3 cases of cystitis. Similarly, 12 bladder tumour cell lines were HPV-negative in these assays. Cervical carcinoma cell lines containing from 1-3 to 600 copies of the HPV genome were used as positive controls and were scored positive in all assays by both Southern blotting and GP-PCR. It is concluded that despite the close proximity of the urothelium to the genital mucosa and the resemblance of some bladder tumours to known HPV-induced lesions in other tissues, HPV infection is absent or very uncommon in bladder tumours.

Identification of glycoproteins expressing tumour-associated PNA-binding sites in colorectal carcinomas by SDS-GEL electrophoresis and PNA-labelling

Many tumour-specific antigens in gastrointestinal cancers have carbohydrate immuno-determinants. These epitopes can be identified by lectins and monoclonal antibodies. By using fluorescein-isothiocyanate (FITC)-conjugated peanut agglutinin (PNA) and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) we have investigated glycoproteins carrying altered carbohydrate epitopes in normal and carcinomatous human colorectal mucosa. In normal mucosa PNA stained goblet cell glycoconjugates in the supranuclear (Golgi) distribution. After neuraminidase pretreatment PNA stained actual mucin goblet itself at all levels of the crypts. Colorectal carcinomas displayed a strong and direct binding of PNA to apical cell membranes of carcinomatous cells and intraluminal secretions. Analysis of the glycoproteins by SDS-PAGE and PNA-labelling revealed four carcinoma-associated glycoproteins (26kD, 32kD, 35kD and 50kD). In addition, four glycoproteins (29kD, 30kD, 33kD and 36kD) common to normal and carcinomatous colorectal mucosa could be identified. All of these glycoproteins differed in their molecular weight from those in red cell controls which bind PNA only after desialylation. The study shows that the expression of PNA-binding sites in colorectal carcinomas signifies a cancer-associated carbohydrate alteration. Four carcinoma-associated glycoprotein antigens could be detected by this lectin. The antigens we have identified might be useful in the isolation and purification of more selective reagents for the serologic detection of colorectal cancer.

Radioiodinated peanut lectin: clinical use as a tumour-imaging agent and potential use in assessing renal-tubular function

Peanut lectin (PNA) has been shown to have a high affinity for Thomsen-Friedenreich (T) antigen, which is associated with the membrane of many solid tumour cells. PNA labelled with 131I was used as a tumour-imaging substance in patients with known metastatic cancer. Serial gamma scintiscans were obtained in 17 patients following a single injection of 131I-labelled PNA. Only in 1 patient was this technique able to reveal a known metastasis at analogue imaging. In the remaining patients, no visible uptake of 131I-PNA could be demonstrated at sites of known metastases. PNA is rapidly excreted through the kidneys and localizes in the renal tubules. As a tumour-imaging agent, 131I-PNA appears to be without value, but its renal-excretory characteristics make it a potentially useful agent for the in vivo assessment of renal-tubular disorders.

Specificities and binding properties of 2 monoclonal antibodies against carcinoma cells of the human urinary bladder

Mice were immunized with cultured cells derived from transitional cell carcinoma of the human urinary bladder (TCC). Spleen cells were fused with mouse myeloma cell line Sp2/0-Ag14 and the hybridomas obtained screened for antibody production against a panel of human cells. Two hybridomas were selected for further studies. The antibodies from one of these hybridomas (P7A5-4) could clearly discriminate between malignant and normal cells from the bladder, both when tested with cultured cells and fresh tissue. The P7A5-4 antibodies, however, also reacted with some non-TCC cultured carcinoma and melanoma cells but to a lesser extent. This difference in reactivity was even more pronounced in the fresh tumours tested, thus indicating a quantitative difference in antigen expression between TCC and other cells. From extracts of TCC cells, P7A5-4 bound three polypeptides of mol. wts 92Kd (ConA+), 23 and 17Kd (ConA-). The antibody derived from hybridoma SK4H-12 bound a ConA reactive glycopeptide of 100Kd mol. wt, the expression of which was almost entirely restricted to urothelial cell lines and tissue of TCC origin, as shown by immunocytochemical studies. The finding in this study of new antigens associated with urinary bladder carcinoma, extend the results obtained previously in our laboratory (Koho et al., 1984; Paulie et al., 1984) and further delineate the heterogeneity of tumour-associated antigens in this human tumour system.

Immunohistological evidence, obtained with monoclonal antibodies, of small intestinal brush border hydrolases in human colon cancers and foetal colons

The expression of small intestinal hydrolases associated with the enterocyte brush border membrane was studied in human colon cancers and foetal colons, by means of monoclonal antibodies against human small intestinal sucrase-isomaltase (SI), maltase-glucoamylase (MGA), lactase (L), aminopeptidase N (APN), and dipeptidylpeptidase IV (DPP-IV). The enzymes were visualized by indirect immunofluorescence on cryostat sections of tumors developed in nude mice with 6 human colon carcinoma cell lines (HT-29, Caco-2, SW-480, HRT-18, HCT-8R, and Co-115), of 27 primary colorectal carcinomas from patients, and of human foetal (16 to 20 weeks of gestation) and normal adult small intestines and colons. All 5 monoclonals bound to the brush border of the adult small intestine, but not to that of the adult colon mucosa. Antibodies against SI, APN and DPP-IV also bound to the brush border of the foetal colons, to apical borders in HT-29 and Caco-2 tumors in nude mice, and to brush border-like structures in 7/27 tumors from patients. No binding was observed for MGA and L in either tumors or foetal colons. Binding of anti-SI antibodies to the brush border of the juxta-tumoral mucosal epithelium was observed in 9/11 samples tested. These data indicate that some colon tumors exhibit a typical pattern of enterocytic differentiation which is of foetal type and which involves at least 3 brush border membrane hydrolases. Monoclonal antibodies to small intestinal hydrolases may, therefore, be important tools for identification and characterization of some differentiated colonic tumors.

Monoclonal antibodies to antigens associated with transitional cell carcinoma of the human urinary bladder. I. Determination of the selectivity of six antibodies by cell ELISA and immunofluorescence

Spleen cells from BALB/c mice immunized with cells derived from transitional cell carcinomas (TCC) of the human urinary bladder were fused with mouse myeloma Sp 2/0 Ag14 cells. Monoclonal antibodies from six established hybridomas were investigated for specificity in a cell ELISA and in indirect immunofluorescence against a large panel of fixed intact cells. Three of the antibodies reacted with half or more of the eight bladder tumors and with a few unrelated tumors. They did not react at all with malignant or normal cells of hematopoietic origin. A fourth antibody reacted with seven of eight bladder tumors. It also reacted weakly with a prostatic carcinoma, with five of six malignant or transformed B cell lines, and with a subpopulation of normal lymphocytes, but not with any of the other cells on the test panel. These four antibodies did not react with cells derived from normal urothelium. The results suggest that these antibodies might recognize cell-type-restricted antigens associated with malignancy. Another antibody reacted with almost all urothelium-derived cells. It also reacted with three of three melanomas but not with any other cells on the panel. The sixth antibody reacted with 32 of the 37 cells tested. The spectrum of reactivities displayed by the antibody suggested that it recognizes HLA antigens.

Monoclonal antibodies to antigens associated with transitional cell carcinoma of the human urinary bladder. II. Identification of the cellular target structures by immunoprecipitation and SDS-PAGE analysis

The cellular target structures for six monoclonal antibodies raised against cultured human bladder carcinoma cells (TCC) were investigated. The specificities of these antibodies when tested against a large panel of cells have been described in the companion paper. Radiolabeled cell lysates were precipitated with the different monoclonal antibodies bound to protein A (Staphylococcus aureus) on a matrix of Sepharose beads. The precipitates were separated by sodium dodecyl sulfate- gel electrophoresis (SDS-PAGE) and analyzed by autoradiography. The antibodies 4B5, 7E9, and 14B11 have previously been found to react in a similar way with TCC-targets and some non-TCC tumor cells, but not with normal urothelial cells or cells of hematopoietic origin. When tested with lysates of a TCC-cell line (TCCSuP) a strong 92K band and a weak 23K band were precipitated with any one of these antibodies. These polypeptides were expressed on the cell surface and were not linked by disulfide bonds. Depletion experiments confirmed that the three antibodies recognized the same antigens. Another antibody (4E8) probably directed to a differentiation antigen present on both urothelial and melanoma cells detected two high molecular polypeptides, 190K and 170K. Antibodies from the S2C6 hybridoma, which displayed a distinct dual specificity for TCC- targets and for malignant or transformed cells of B cell origin, precipitated a 50K component from extracts of either TCC- or B cell-derived cell lines. Antibodies produced by the S2A9 hybridoma were shown to bind to a framework epitope of the HLA-A, B, C heavy chain.