Cell surface antigens of human bladder tumors: definition of tumor subsets by monoclonal antibodies and correlation with growth characteristics

Normal human urothelium and tumors of urothelial origin were analyzed with a panel of seven mouse monoclonal antibodies that identify surface antigens of cultured bladder cancer cell lines. Three categories of antigens were defined on the basis of differential expression on normal urothelium versus bladder tumors. Om5 (a category 1 antigen) is a highly restricted, differentiation antigen detected in the normal urothelium of 50-60% individuals. No other normal cell type in Om5- or Om5+ individuals expresses Om5. The incidence of Om5 expression in superficial bladder tumors is significantly higher (88%) than in normal urothelium, whereas its expression in invasive or metastatic tumors is far lower (20%), suggesting Om5 gain/loss in bladder tumors. Paired biopsies of normal urothelium and bladder tumors from the same individuals have shown Om5 induction in the superficial bladder tumors of Om5- individuals and Om5 loss in invasive bladder cancers of Om5+ individuals. Category 2 antigens (T43, T138, T23) are not expressed by normal urothelium or most superficial bladder tumors but are detected on a high proportion of invasive or metastatic bladder tumors, indicating that category 2 antigens are associated with late stages of tumor progression. Category 3 antigens (T16, T87, J143) provide lineage markers for normal or neoplastic cells of urothelial origin, being found on normal urothelium and virtually all bladder tumors. Thus, differential expression of category 1 and 2 antigens divide bladder tumors into distinct subsets, and these subsets correlate with pathological and clinical features of the disease.     

Immunohistochemical localization of prostate carcinoma-associated antigens

The immunoperoxidase technique was used to study the localization and distribution of antigens reactive with two monoclonal antibodies, D83.21 and P6.2, produced against cultured prostate tumor cells, in formalin-fixed, paraffin-embedded histological sections of human tissues. Monoclonal D83.21 reacted with 11 of 19 (58%) primary prostate carcinomas and 1 of 6 (17%) metastatic tumors, whereas monoclonal P6.2 reacted with 14 of 19 (68%) primary and 4 of 6 (67%) metastatic prostate tumors. Neither antibody reacted with five primary prostate tumors and one metastatic prostate tumor. In some tumor cells, the antigens recognized by these monoclonals were localized in either the cytoplasm or cell membrane, while in other tumor cells, both diffuse cytoplasmic and membrane or focal staining patterns were observed. In addition to the variable staining patterns, antigenic heterogeneity was also noted within most prostate tumors examined. Two types of staining variability were observed: (a) tumor cells in one area of the tissue section stained positive, but in another area they did not react with the antibody; and (b) both stained and unstained tumor cells were adjacent to each other. These results would suggest that a panel of monoclonals will be required to detect the different subpopulations of prostate tumor cells. Neither antibody reacted with 6 normal or 12 benign prostate tissues, nor any of a variety of other normal human tissues except for staining of the proximal tubules of normal kidneys. The antigen detected by P6.2 demonstrated a wider tissue distribution being found on bladder, breast, lung, and pancreatic tumors, whereas the antigen recognized by D83.21 was restricted to prostate and bladder carcinomas. These antibodies may have clinical applicability for the identification of prostate tumor cells in biopsy specimens and for immunohistopathological classification of prostate carcinomas.     

Targeting of human glioma xenografts in vivo utilizing radiolabeled antibodies

Radiolabeled antibodies provide a potential basis for selective radiotherapy of human gliomas. We have measured tumor targeting by radiolabeled monoclonal and polyclonal antibodies directed against neuroectodermal and tumor-associated antigens in nude mice bearing human glioma xenografts. Monoclonal P96.5, a mouse IgG2a immunoglobulin, defines an epitope of a human melanoma cell surface protein, and specifically binds the U-251 human glioma as measured by immunoperoxidase histochemistry. 111In-radiolabeled P96.5 specifically targets the U-251 human glioma xenograft and yields 87.0 microCuries (microCi) of tumor activity per gram per 100 microCi injected activity compared to 4.5 microCi following administration of radiolabeled irrelevant monoclonal antibody. Calculations of targeting ratios demonstrate deposited dose to be 11.6 times greater with radiolabeled P96.5 administration compared to irrelevant monoclonal antibody. The proportion of tumor dose found in normal organs is less than 10%, further supporting specific targeting of the human glioma xenograft by this antibody. Monoclonal antibody ZME018, which defines a second melanoma-associated antigen, and polyclonal rabbit antiferritin, which defines a tumor-associated antigen, demonstrate positive immunoperoxidase staining of the tumor, but comparatively decreased targeting. When compared to the 111In-radiolabeled antibody, 90Y-radiolabeled P96.5 demonstrates comparable tumor targeting and percentages of tumor dose found in normal organs. To test the therapeutic potential of 90Y-radiolabeled P96.5, tumors and normal sites were implanted with miniature thermoluminescent dosimeters (TLD). Seven days following administration of 100 microCi 90Y-radiolabeled P96.5, average absorbed doses of 3770, 980, 353, and 274 cGy were observed in tumor, liver, contralateral control site, and total body, respectively. Shared cell surface antigens among neuroectodermally derived neoplasms provide a basis for exploration of human glioma radioimmunotherapy.     

Correlation of catabolism and antigen shedding with monoclonal antibody uptake by human colon and lung tumor xenografts

Catabolism of B6.2 and B72.3 monoclonal antibodies by human colon (LS174T) and lung (A549) tumor xenografts appears to be a variable and complex process. 125I-B72.3 and 125I-B6.2 bound to LS174T cells in vitro and localized in tumors in athymic mice. 125I-B6.2 bound to A549 cells in vitro, but did not localize in tumors. To understand these differences, tumors were grown around subcutaneously implanted micropore chambers and tumor fluid was analyzed for the presence of shed tumor antigen or functional antibody following intravenous injection. Up to two orders of magnitude lower functional B6.2 was detected in the A549 tumor fluid than in the LS174T tumor fluid. Also, A549 fluid almost totally (82-94%) inhibited binding of 125I-B6.2 to target cells in vitro due to free B6.2 antigen present in the chamber fluid. The micropore chambers utilized in this study have the potential for expanding our understanding of the way antibodies are metabolized in various types of tumors. Utility of an antibody in vivo cannot be entirely predicted from in vitro binding studies given that tumor related factors such as antibody transport, antibody catabolism and shed antigen may influence localization of antibodies in tumors. Monoclonal antibodies against nonshed antigens may prove to be more appropriate for cancer imaging and therapy.     

Characterization of Mouse Monoclonal Antibody B1.4 Reactive with Human Invasive Bladder Cancer and Some Other Malignant Tumors but Not with Normal Urinary Epithelium

Immunohistochemical analysis by indirect immunoperoxidase staining demonstrated that monoclonal antibody (mAb) B1.4 derived from a mouse immunized with a bladder cancer cell line EJ-1 was reactive with a high proportion of high-grade and invasive bladder tumors, but not with the majority of low-grade and superficial bladder tumors, or normal urinary epithelium. Among 71 primary bladder tumors classified by pathological grading, positive stainings were observed in 1 of 34 tumors (3%) of grade 1, 8 of 20 tumors (40%) of grade 2 and 14 of 17 tumors (82%) of grade 3. When the tumors were classified by pathological staging, positive stainings were observed in only 8 of 54 (15%) superficial tumors of stages Ta and T1, but in 15 of 17 (88%) invasive tumors of stages T2 and T3. mAb B1.4 showed restricted positive stainings with normal tissues including renal glomerulus, vascular endothelium, squamous epithelium of esophagus, glandular epithelium of prostate, and epithelium of pancreatic acinar gland and minute duct, while positive stainings were observed in a range of tumor tissues other than bladder tumor. Mixed hemadsorption assays with a panel of cell cultures showed also that the antigen recognized by mAb B1.4 was expressed on a range of tumor cell lines. These findings suggest that the antigen recognized by mAb B1.4 may appear after malignant transformation, and be an indicator of malignant potential of bladder cancer.     

Identification by monoclonal antibodies of an antigen shed by human bladder cancer cells

The reactivities of two anti-bladder cancer monoclonal antibodies, AN43 and BB369, were characterized. AN43 and BB369 reacted with a majority (greater than 50%) of bladder cancer tissue sections tested by immunoperoxidase staining. When tested against a panel of 27 normal human tissues, AN43 and BB369 reacted only with urothelium and stomach. AN43 and BB369 showed identical binding patterns and competed for binding on bladder cancer cells, suggesting that the two antibodies react with identical or spatially close epitopes. Bound BB369 antibody was rapidly shed from the surface of viable UM-UC-9 human bladder cancer cells. The antigen was found in spent tissue culture medium from the UM-UC-9 human bladder cancer cell line. AN43 and BB369 define a shed bladder tumor-associated antigen with limited distribution on normal tissues. The antigen is different from bladder tumor-associated antigens defined by other monoclonal antibodies and may be useful for the diagnosis and follow-up of patients with bladder cancer.     

Human ovarian cancer xenografts in nude mice: characterization and analysis of antigen expression

We have characterized 13 different human ovarian cancer xenografts grown subcutaneously in nude mice. The tumor lines represented 5 histological subtypes: serous (4), mucinous (4), clear-cell carcinoma (1), carcinosarcoma (1) and undifferentiated (3). The specific histology and the degree of differentiation resembled those of the original patients' tumors and were maintained upon serial transfer. Volume doubling times of the xenografts ranged from 3.5 to 15 days. The xenografts were also analyzed for their antigen expression using 20 monoclonal antibodies (MAbs) reactive with 15 tumor-associated antigens. Immunohistochemical examination of tissue sections showed a positive reaction pattern with MAbs 115D8, 140C1, 139H2, 175C5, HMFG1 and HMFG2, each recognizing episialin, as well as with MAbs AUA1, 358.4.32 and 199-157 in xenografts of the serous, mucinous and clear-cell carcinoma subtype. MAb OC125 was reactive with xenografts of the serous subtypes. Other antibodies, such as 494 and OV-TL3 infrequently demonstrated positive reactions. Reactivity of all MAbs was low in the carcinosarcoma and undifferentiated tumor lines. With the exception of AUA1, 495 and 126E5, all MAbs revealed a heterogeneous staining pattern. MAbs against episialin and OC125 predominantly stained the apical site of the tumor cells. Strongest reactivity with almost all histological subtypes was observed with MAbs 115D8, 140C1, 139H2 and AUA1. In cases where we were able to compare the patients' tumor tissue with the respective xenografts, retention of antigen expression was demonstrated in each instance. Release of tumor-associated antigens was shown for CA125 in 2 serous-tumor lines, for CA15.3 in 1 serous-tumor line, and for CEA in 3 lines of the mucinous subtype. This panel of human tumor xenografts could be a valuable tool to determine the potential usefulness of MAb-guided therapy in ovarian cancer.     

Monoclonal antibodies to six-transmembrane epithelial antigen of the prostate-1 inhibit intercellular communication in vitro and growth of human tumor xenografts in vivo

Six-transmembrane epithelial antigen of the prostate-1 (STEAP-1) is a novel cell surface protein highly expressed in primary prostate cancer, with restricted expression in normal tissues. In this report, we show STEAP-1 expression in prostate metastases to lymph node and bone and in the majority of human lung and bladder carcinomas. We identify STEAP-1 function in mediating the transfer of small molecules between adjacent cells in culture, indicating its potential role in tumor cell intercellular communication. The successful generation of two monoclonal antibodies (mAb) that bind to cell surface STEAP-1 epitopes provided the tools to study STEAP-1 susceptibility to naked antibody therapy. Both mAbs inhibited STEAP-1-induced intercellular communication in a dose-dependent manner. Furthermore, both mAbs significantly inhibited tumor growth in mouse models using patient-derived LAPC-9 prostate cancer xenografts and established UM-UC-3 bladder tumors. These studies validate STEAP-1 as an attractive target for antibody therapy in multiple solid tumors and provide a putative mechanism for mAb-induced tumor growth inhibition.     

Radioimmunoimaging of human bladder tumor xenografts in nude mice by using monoclonal antibodies

Monoclonal antibodies HBJ127 and HBJ8, raised against T24 human bladder cancer cells, predominantly react with the cells in proliferating stages and with a portion of epithelial tumor cells, respectively. To investigate the in vivo localization of these monoclonal antibodies, the antibodies were labeled with radioiodine and indium-111 (111In) and injected into nude mice transplanted with human bladder tumors. The BT-11 bladder tumor had the highest concentration of radioiodinated HBJ127 and HBJ8 monoclonal antibodies, with 11.6 and 14.3% of the injected dose per gram and with a tumor-to-blood ratio of 2.6 and 1.6, respectively, at 4 days after the administration. An irrelevant monoclonal antibody did not show any specific accumulation in the BT-11 tumor. The 111In-labeled HBJ127 antibody was also localized in the tumor with a higher tumor-to-blood ratio than the radioiodinated antibody. The xenografted BT-11 tumor was successfully visualized with the radiolabeled HBJ127 and HBJ8 antibodies by scintigraphy. These monoclonal antibodies and the human bladder tumor xenografts may provide a good model for radioimmunoimaging and possibly therapy.     

Prostate stem cell antigen is overexpressed in human transitional cell carcinoma.

Prostate stem cell antigen (PSCA), a homologue of the Ly-6/Thy-1 family of cell surface antigens, is expressed by a majority of human prostate cancers and is a promising target for prostate cancer immunotherapy. In addition to its expression in normal and malignant prostate, we recently reported that PSCA is expressed at low levels in the transitional epithelium of normal bladder. In the present study, we compared the expression of PSCA in normal and malignant urothelial tissues to assess its potential as an immunotherapeutic target in transitional cell carcinoma (TCC). Immunohistochemical analysis of PSCA protein expression was performed on tissue sections from 32 normal bladder specimens, as well as 11 cases of low-grade transitional cell dysplasia, 21 cases of carcinoma in situ (CIS), 38 superficial transitional cell tumors (STCC, stages T(a)-T(1)), 65 muscle-invasive TCCs (ITCCs, stages T(2)-T(4)), and 7 bladder cancer metastases. The level of PSCA protein expression was scored semiquantitatively by assessing both the intensity and frequency (i.e., percentage of positive tumor cells) of staining. We also examined PSCA mRNA expression in a representative sample of normal and malignant human transitional cell tissues. In normal bladder, PSCA immunostaining was weak and confined almost exclusively to the superficial umbrella cell layer. Staining in CIS and STCC was more intense and uniform than that seen in normal bladder epithelium (P < 0.001), with staining detected in 21 (100%) of 21 cases of CIS and 37 (97%) of 38 superficial tumors. PSCA protein was also detected in 42 (65%) of 65 of muscle-invasive and 4 (57%) of 7 metastatic cancers, with the highest levels of PSCA expression (i.e., moderate-strong staining in >50% of tumor cells) seen in 32% of invasive and 43% of metastatic samples. Higher levels of PSCA expression correlated with increasing tumor grade for both STCCs and ITCCs (P < 0.001). Northern blot analysis confirmed the immunohistochemical data, showing a dramatic increase in PSCA mRNA expression in two of five muscle-invasive transitional cell tumors when compared with normal samples. Confocal microscopy demonstrated that PSCA expression in TCC is confined to the cell surface. These data demonstrate that PSCA is overexpressed in a majority of human TCCs, particularly CIS and superficial tumors, and may be a useful target for bladder cancer diagnosis and therapy.     

Monoclonal antibody against a tumor-associated sialoglycoprotein of superficial papillary bladder tumors and cervical condylomas.

A mouse IgG1 monoclonal antibody (MAb), 19A211, defining a tumor-associated cell-surface antigen of superficial papillary bladder tumors, was generated by immunizing with fresh bladder tumor cells mice neonatally injected with normal human urothelial cells. The reactivity of the antibody with cell lines was examined by indirect immunofluorescence staining and was restricted to 3/14 bladder cancer lines and 3/31 cancer cell lines of non-bladder origin, including HeLa cervical cancer. No normal fibroblast, kidney cells, EBV-lymphocytes, erythrocytes or leukocytes expressed the antigen. Reactivity of MAb 19A211 was well preserved on tissue paraffin sections. Immunoperoxidase staining of normal adult or fetal tissues showed no reactivity except for a patchy or uniform staining of umbrella cells in 6/23 adult and 1/4 fetal urothelium samples. Positive and often heterogeneous staining was observed on 24/38 papillary superficial tumors (Ta) and 4/5 carcinoma in situ bladder lesions but on only 4/20 infiltrating tumors. It was also observed on 5/6 cervical condylomas and one bladder condyloma, but none of 6 penile or vulvar condylomas. All other tumors tested were negative. The antigenic determinant is present on a heterogeneous group of proteins with molecular weights ranging from 90 to 200 kDa. It is sensitive to periodate treatment and to neuraminidase but only partially sensitive to proteases. MAb 19A211 is different from other reported MAbs with similar reactivity to superficial bladder tumors and umbrella cells of normal urothelium. When tested in competition assays, several of these MAbs, but not 19A211, were found to react with Lewis X blood group determinant. Our results suggest that 19A211 may be useful for detection and stratification of bladder tumors.     

Pancreatic cancer-associated new carbohydrate antigen

Antigen T (Thomsen-Friedenreich), a precursor of blood group MN antigens, with the determinant of a carbohydrate (Gal-GalNac), antigen Lewisx (Lex), a trisaccharide found on Type 2 blood group oligosaccharide chains, Sialo-Lex, the derivatives: of Lex and Lewisy (Ley), a difucosylated tetrasaccharide have, behaved as the oncodevelopment tumor-associated antigens in human colon. In the present study, using monoclonal antibodies in immunohistochemical method, the expression and distribution of these antigens in pancreatic cancer and normal pancreatic tissue were observed and compared. It was found that monoclonal antibody AH8-258 derived against antigen T, SSEA-1 and FH-4 derived against short and long chain antigens Lex, FH-6 and IB 9 derived against Sialo-Lex antigen preferentially stained most pancreatic cancer tissues but rarely the normal pancreatic tissues. However, monoclonal antibodies AH-6, KH-1 and CC-1, derived against antigen Ley, stained the majority of tissues regardless of being malignant or normal and were not able to differentiate cancer from the normal tissues. Antigens T, Lex, Sialo-Lex and Ley were also expressed in chronic pancreatitis but the chance of monoclonal antibodies staining in these tissues (18 approximately 36%) was markedly lower than the staining in cancer tissues (54 approximately 77%) except Ley. It is suggested that in human pancreas, haptens T, Lex and Sialo-Lex, the oncodevelopmental cancer-associated antigens, are highly specific markers for malignancy and likely helpful in the early diagnosis of pancreatic cancer.     

Radioimmunolocalisation of bladder tumors xenotransplanted in nude mice

We have previously reported on the derivation of mouse monoclonal antibodies (Mabs), identifying several cell surface antigens selectively associated with cancer of the urinary bladder (TCC) (1-4). Three of these Mabs (4E8, SK4H and 8F4) have now been assessed for their ability to localise TCC-tumor xenografts in nude mice. The biodistribution of 125I-labeled intact Mabs as well as the corresponding Fab and F(ab')2 fragments from two of them were investigated in animals carrying TCC tumors or antigen negative control tumors. Using direct measurements of excised tissues, all three antibodies were seen to accumulate specifically in the TCC tumors, giving tumor to normal tissue ratios of between 3 and 20 depending on the Mab used and the time after injection. Antibody fragments were generally more efficient in their localisation, mainly due to a dramatic reduction in the blood background as compared to intact Ig. One of the antibodies, 4E8, was also employed for external imaging with gamma camera scintigraphy using 111In or 131I as tracers. Excellent visualisation of the tumor sites could be obtained both with Fab fragments and intact antibody within 12-24 hours after injection. As expected, background radioactivity was significantly lower with fragments than with whole molecules. 111In labeled antibodies appeared in all instances to be superior to the corresponding 131I conjugates. In conclusion, the present study indicates that the three anti TCC antibodies may become useful for the in vivo diagnosis of bladder cancer in man.     

Adjuvant intravesicular pharmacotherapy for superficial bladder cancer.

In 1990, bladder cancer, excluding carcinoma in situ, was estimated to contribute 49,000 cases of cancer. In men 75 years old or older, it became the fifth leading cause of cancer deaths. Of patients with bladder cancer, 75%-80% initially present with superficial bladder tumors. Treatment of these tumors has three objectives: 1) to eradicate existing disease, 2) to provide prophylaxis against tumor recurrence, and 3) to avoid deep invasion into the muscle layers of the bladder. Transurethral resection is the primary treatment to eradicate superficial bladder tumors, but 40%-80% of these tumors recur. Because of these high recurrence rates, adjuvant intravesicular pharmacotherapy with cytotoxic and immunomodulatory drugs has gained widespread use. The past two decades of clinical investigations in superficial bladder cancer have provided valuable information on the biology and treatment of the disease. Multivariate analyses have indicated that tumor grade and stage are the most important prognostic variables commonly available to the clinician to identify the patient at greatest risk of developing muscle-invasive or metastatic bladder cancer. These studies have also identified groups at low risk for tumor recurrence and invasive bladder cancer. Randomized trials have shown that recurrence rates are decreased by adjuvant intravesicular pharmacotherapy with a number of drugs: bacillus Calmette-Guérin vaccine (BCG), doxorubicin, ethoglucid (Epodyl), mitomycin-C, teniposide, and thiotepa. However, few studies indicate that adjuvant intravesicular pharmacotherapy can prevent progression to invasive bladder cancer in the high-risk patient with superficial bladder cancer. Additional clinical trials are needed to determine whether such therapy can prevent invasive and metastatic bladder cancer and improve disease-free survival in this group. In addition, the identification of tests (e.g., monoclonal antibody tests, chromosomal analyses, and tumor marker assays) that can help to identify high-risk patients is needed to better develop therapeutic strategies for superficial bladder cancer.     

Influence of cocktails of labeled monoclonal antibodies on the localization of antibodies in human tumor xenografts

In order to evaluate the usefulness of cocktails of labeled monoclonal antibodies (MoAbs) recognizing different antigen molecules to localize human cancer xenografts, we have compared the potential of three MoAbs recognizing representative cancer-associated CA 19-9, 17-1A and CEA antigens when administered alone or in combination. Specific binding of radioiodinated F(ab')2 fragments of these three MoAbs was observed to human colorectal cancer cell lines SW1116, LS180 and Co-3. The percentage of in vitro cell binding of a cocktail of any two MoAbs to cancer cells was equal to the average of those obtained with the two MoAbs alone. The three MoAbs were preferentially localized in tumor tissues xenografted in nude mice. When cocktails of any two MoAbs were used, the obtained tumor-to-normal tissue ratios and percent of injected dose per gram of tumor were between the levels obtained for each MoAb when administered alone, in all three tumors transplanted in nude mice. These data suggest that, although cocktails of labeled MoAbs recognizing different antigens may extend the spectrum of tumor specificities, their use does not improve the tumor localization ability of MoAb-conjugates.     

Preliminary identification of a tumor-associated glycoprotein in bilharzial bladder cancer urine

A tumor-associated antigen was preliminarily identified in urine from bilharzial (squamous-cell carcinoma) bladder cancer patients. Monospecific rabbit antisera were made by immunization with concentrated bladder cancer urine and exhaustive absorption with insoluble normal human serum and urine. The urine tumor-associated antigen was identical to an antigen from 3M KCl bladder tumor extract by immunodiffusion. The antigen in urine was found in nine of 10 bladder cancer patients and was absent from normal urine and serum by immunodiffusion and immunoelectrophoresis. The antigen was a concanavalin-A-binding glycoprotein which was anodal on immunoelectrophoresis. It was stable up to 2 years at -20 degrees C and did not cross-react with carcinoembryonic antigen or with Schistosoma haematobium antigens.     

Antigenic phenotype of radial growth phase melanomas with or without a vertical growth phase portion

The expression of 4 melanoma-associated antigens and of class I and II HLA antigens was investigated in 12 superficial spreading melanomas (SSM) and in 8 SSM with a vertical growth pattern portion (SS + NM) by the use of monoclonal antibodies and an indirect immunoperoxidase procedure. Monoclonal antibodies 225.28, 763.74, CL.203, VF19-LL217, Q5-13, W6-32 and anti-HLA-DR, were used. Each antigen was more frequently expressed by SS + NM on the whole than by SSM and also by the radial growth pattern portions of SS + NM than by SSM. Vertical growth pattern portions of SS + NM were not antigenically similar to radial growth pattern portions in the same tumors. The high frequency of antigen expression in radial growth pattern melanomas seems to be associated with the appearance of a more invasive cell population.     

Human bladder cancer associated antigens: evaluation of antigenicity in TCC tissues of different grades and in normal urothelium

The expression of five antigens, associated with transitional cell carcinoma (TCC) of the urinary bladder on biopsies of tumors or normal urothelium, was studied by immunostaining with the corresponding monoclonal antibodies. Both tissue sections and single cell preparations were investigated with either indirect immunoperoxidase staining or immunofluorescence. All 5 antigens were expressed on the majority (70-90%) of sectioned tumor specimens from 44 TCC patients, and 4 of them were similarly expressed on single cell tumor preparations from 26 additional patients. However, in both types of preparation, the degree of expression of these antigens varied from scattered staining of less than 25% of the tumor cells to homogenous staining of all or almost all cells. This degree of expression varied individually for each of the antigens and was not related to the malignancy grade of the tumors. However, as most of the tumors were of grades II or III, no conclusions regarding the relationship of antigen expression to the aggressiveness of the tumors can be drawn. In any event, all tumors expressed at least one and mostly several of these antigens. Antigen expression on biopsies of normal bladder mucosa from TCC patients or on urothelial biopsies from patients with prostate hyperplasia was also observed on single cell specimens (34 patients) but not on sectioned material (9 patients). However, the frequency of positive specimens was much lower (4-20%). Moreover, the number of cells expressing one or, occasionally, several of the antigens in normal urothelium was small (usually less than 5%). Because of these marked differences in antigen expression between tumors and normal tissue, the results indicate that a combination of 3-5 of the antibodies used in this study may be suitable for diagnostic purposes.     

Influence of Cocktails of Labeled Monoclonal Antibodies on the Localization of Antibodies in Human Tumor Xenografts

In order to evaluate the usefulness of cocktails of labeled monoclonal antibodies (MoAbs) recognizing different antigen molecules to localize human cancer xenografts, we have compared the potential of three MoAbs recognizing representative cancer-associated CA 19–9, 17–1A and CEA antigens when administered alone or in combination. Specific binding of radioiodinated F(ab')₂ fragments of these three MoAbs was observed to human colorectal cancer cell lines SW1116, LS180 and Co-3. The percentage of in vitro cell binding of a cocktail of any two MoAbs to cancer cells was equal to the average of those obtained with the two MoAbs alone. The three MoAbs were preferentially localized in tumor tissues xenografted in nude mice. When cocktails of any two MoAbs were used, the obtained tumor-to-normal tissue ratios and percent of injected dose per gram of tumor were between the levels obtained for each MoAb when administered alone, in all three tumors transplanted in nude mice. These data suggest that, although cocktails of labeled MoAbs recognizing different antigens may extend the spectrum of tumor specificities, their use does not improve the tumor localization ability of MoAb-conjugates.     

A single treatment of yttrium-90-labeled CHX-A"-C6.5 diabody inhibits the growth of established human tumor xenografts in immunodeficient mice

Antitumor diabody molecules are noncovalent single-chain Fv dimers that recapitulate the divalent binding properties of native IgG antibodies. Diabodies are capable of substantial accumulation in tumor xenografts expressing relevant antigens in immunodeficient mouse models. With a Mr of approximately 55,000, diabodies are rapidly cleared from the circulation, resulting in tumor-to-blood ratios that significantly exceed those achieved early after the administration of monoclonal antibodies. We have evaluated the therapeutic potential of the beta-emitting isotope yttrium-90 (t1/2, 64 hours) conjugated to the C6.5K-A diabody that specifically targets the HER2/neu human tumor-associated antigen. We have found that a single intravenous dose of 150 microCi (200 microg) 90Y-CHX-A"-C6.5K-A diabody substantially inhibits the growth rates of established MDA-361/DYT2 human breast tumor xenografts in athymic nude mice. In contrast, 300 microCi (300 microg) 90Y-CHX-A"-C6.5K-A diabody resulted in only a minor delay in the growth of SK-OV-3 human ovarian cancer xenografts. The maximum tolerated dose was also dependent on the tumor xenograft model used. These studies indicate that genetically engineered antitumor diabody molecules can be used as effective vehicles for radioimmunotherapy.