Multiparameter evaluation of the expression in situ of normal and tumor-associated antigens in human colorectal carcinoma.

The immunoreactivity of a panel of monoclonal antibodies (MoAb) was studied in a series of patients with colorectal carcinomas to test the association of antigen expression with other parameters such as histopathologic stage, differentiation, and clinical outcome. Low-level binding to normal tissue and high-level binding to malignant tissue were observed with MoAb defining, respectively, a gastrointestinal cancer antigen (GICA), Leb (distal colon only), A, H type 2 antigen, X-like antigen, and the 200-kilodalton (Kd) protein of carcinoembryonic antigen (CEA). The degree of histologic differentiation correlated with the expression of Lea antigen, A, and Y haptens, whereas a progressive loss of these antigens coincided with loss of differentiation. Two undifferentiated carcinomas expressed only two, H type 2 antigen and a highly glycosylated protein of 20-50 Kd, of the 14 antigens investigated. An interesting, but not significant, association between Leb antigen expression and more extensive disease was found: Whereas 71% of Dukes C tumors were positive for Leb, only 48% of patients with Dukes A and B2 tumors showed the presence of Leb antigen. On the other hand, the presence of B72.3-defined antigen is significantly associated with an earlier stage of disease. Chi-square tests to assess the association of antigen positivity with disease recurrence indicated a significant binding association with tumor recurrence over a broad range of percent positive cells for two MoAb defining different determinants of GICA. Similar associations, but over a narrow range of positive cells, were found for H type 2 antigen and the 200-Kd protein of CEA.

Gastrointestinal cancer-associated antigen in immunoperoxidase assay.

A monosialoganglioside antigen of gastrointestinal adenocarcinomas defined by murine monoclonal antibody was demonstrated by immunoperoxidase (IP) assay in fixed paraffin-embedded tumors in 59% of colonic adenocarcinomas, 86% of pancreatic adenocarcinomas, and 89% of all gastric adenocarcinomas. In all patients with detectable levels of antigen in circulation, the resected tumors also expressed the antigen in IP assay. Six of eight individuals with no detectable levels of antigen in their serum samples expressed the antigen in the tumor tissue. Removal of the sialic acid residue of the antigen abolished the IP reaction. The successful use of the IP assay on fixed tissue to demonstrate the specific sites of gastrointestinal cancer antigen localization in human tumors and normal tissues provides an important tool for the study of developing neoplasia.

Immunoassay for melanoma-associated proteoglycan in the sera of patients using monoclonal and polyclonal antibodies.

A melanoma-associated proteoglycan antigen is expressed by primary cutaneous and ocular melanomas, metastatic melanomas, nevus cells, some astrocytomas, and fetal fibroblasts, and it is shed into culture supernatant by both melanoma and nevus cells. The antigen is also expressed by tumor cells in vivo. Melanoma and nevus cells, but not normal melanocytes, were specifically stained by the immunoperoxidase procedure. The proteoglycan antigen, purified by immunoaffinity chromatography using a monoclonal antibody that specifically detects this antigen, was used to immunize rabbits. The resulting serum was tested by sequential immunoprecipitation and found to react with the same population of molecules detected by the anti-proteoglycan monoclonal antibodies. Furthermore, the reactivity patterns of the rabbit serum and of the monoclonal antibodies with a variety of tumor and normal cells were the same. Based on the these data, we conclude that the entire proteoglycan molecule is a melanoma-associated antigen. The monoclonal antibodies and immunoglobulin from the rabbit serum were tested in a double determinant immunoassay for the detection of antigen in a total of 339 sera from patients with various diseases. Elevated levels of circulating proteoglycan antigen were found in 76% of patients with a high metastatic melanoma tumor burden compared to 2% of healthy donors. A fraction (22%) of patients with light tumor burden or nonmelanoma neoplastic disease also had elevated levels of circulating proteoglycan antigen. The source of the antigen for the latter patients may be collagenous connective tissue which, as judged by immunoperoxidase staining, expresses the antigen in both normal and transformed tissues.

Identification of melanoma-associated antigens using fixed tissue screening of antibodies.

Early culture supernatants from hybridomas that were obtained through fusions of mouse myeloma cells with lymphocytes of melanoma-immunized mice were screened for their reactivity with a paraffin-embedded cell block of a melanoma cell line, using a biotin:avidin immunoperoxidase procedure. Eleven monoclonal antibodies were derived that define several new melanoma-associated antigens. The antigens include a neutral glycolipid, gangliosides, membrane-associated proteins, cytosolic proteins, and strongly secreted proteins. These antibodies, which detect antigens that withstand tissue fixation and embedding procedures, were tested for reactivity in fixed cell lines, as well as in melanoma biopsies. These antibodies may provide powerful tools in diagnostic studies of human malignant melanoma biopsy material.

Immunoperoxidase localization of secretory component in reactive mesothelium and mesotheliomas.

Immunoperoxidase localization of the secretory component (SC) of immunoglobulin (Ig) A was undertaken on paraffin-embedded tissue to investigate its usefulness in differentiating epithelial from mesothelial neoplasms. Epithelial cells, but not mesothelial cells, have been previously reported to contain SC, which may be involved in the intraepithelial transport of secretory IgA. In this study we report, for the first time, focal but unequivocal staining for SC in reactive mesothelium and mesotheliomas (10/16). Reactivity was not associated with any histologic subtype of malignant mesothelioma although staining was usually seen in the "epithelial" type. The significance in terms of possible mesothelial secretory IgA transport is discussed.

An antimelanoma monoclonal antibody and the histopathology of uveal melanomas.

Monoclonal antibody ME491 identifies a cutaneous melanoma-associated antigen in formaldehyde-fixed, paraffin-embedded tissues. This antibody was applied to tissue sections from 79 cases of formaldehyde-fixed, paraffin-embedded uveal melanomas. Sixty-nine (87.3%) of the 79 cases showed staining by monoclonal antibody ME491, thus demonstrating an antigen shared by cutaneous and uveal melanomas. No relationship between the staining pattern and the patient outcome was detected. Twelve (85.7%) of the 14 cases in which balloon cells were present stained with the antibody. The antibody stained the long posterior ciliary nerve in 12 (38.7%) of 31 cases in which the nerve was present in tissue sections. The antibody cross reacted with the retinal pigment epithelium (16.5% of cases) and a variety of normal nonocular tissues.

Immunoreactive secretory component of IgA in human tissues and tumors.

Secretory component (SC) of IgA is known to be produced by many glandular epithelial cells. To assess the usefulness of this antigen as a glandular tumor differentiation marker, the authors evaluated its normal body distribution and tested a variety of glandular neoplasms. The known normal distribution of SC was confirmed and extended to include the prostate. Immunoreactive SC was detected only in epithelial and glandular tumors. Many types of well-differentiated adenocarcinomas, (ovary, prostate, small bowel, pancreas, stomach, biliary) however, contained little or no immunoreactive SC. Therefore, the authors conclude that immunoreactive SC cannot be used as a general differentiation marker in tumors of glandular derivation. A high frequency of antigenic expression was found in adenocarcinomas of the lung (4/10), breast (5/10), and colon (24/27). No correlation between expression and tumor differentiation was observed in pulmonary or mammary tumors. Only in colonic neoplasia was such a relationship detected. SC expression in colonic tumors was not related to mucin content but was associated with the presence of a visible brush border.

Initial clinical evaluation of two murine IgG2a monoclonal antibodies for immunotherapy of gastrointestinal carcinoma.

Eleven patients with advanced gastrointestinal (GI) carcinoma were entered in Phase I initial clinical trials with IgG2a antiGI carcinoma monoclonal antibodies (MAbs) GA733 (five patients) or CO19-9 (six patients). Infusion of MAb GA733 in doses greater than 30 mg was accompanied by mild and short-lasting GI toxicity. Infused MAb GA733 was bound to each patient's tumor tissue in vivo. MAb circulated in the blood for 10-25 days. All patients developed anti-mouse antibodies between 15 and 60 days post infusion. Furthermore, all but one patient raised anti-idiotypic antibodies against MAb GA733. Following administration of 10-600 mg of MAb CO19-9, no immediate or delayed toxicity symptoms were noted. Binding of infused MAb CO19-9 to tumor cells in vivo could not be detected in any of the six patients studied. The MAb circulated in the bloodstream between 5 and 12 days. Human anti-mouse antibody was detected in sera of three patients. None of the eleven patients treated with either MAb had anti-tumor responses in this Phase I clinical trial. The strong binding reactivity of MAb GA733 to tumors in vivo suggests the use of this MAb in cancer patients with less tumor burden to determine the tumoricidal efficacy of this antibody.

Cytologic detection of amyloid in duodenal and ureteral brushings.

Two cases of amyloidosis diagnosed on cytology brush specimens are described. Gastrointestinal involvement in a patient with primary amyloidosis and renal involvement in a patient with multiple myeloma were diagnosed on duodenal and ureteral brush specimens, respectively. Familiarity with the staining characteristics of amyloid should increase it as a consideration in the differential diagnosis of certain specimen types.

Diagnosis of Pneumocystis carinii pneumonia by cytologic evaluation of Papanicolaou-stained bronchial specimens.

To evaluate the sensitivity and specificity of diagnosing Pneumocystis carinii pneumonia (PCP) by Papanicolaou-stained bronchial brushing and wash/lavage specimens obtained by fiberoptic bronchoscopy, the cytologic preparations and clinical records from 58 immunocompromised patients were reviewed. Bronchial brushings and wash/lavage specimens were examined using methenamine silver (Grocott) and Papanicolaou stains. Pneumocystis carinii pneumonia was recognized with Papanicolaou stain by identifying distinctive alveolar casts, which frequently contained collections of encysted sporozoites. Thirty cases of PCP were identified, and Grocott-stained bronchial wash/lavage specimens were positive in 29 instances (97%). Grocott staining of the transbronchial biopsy was positive for PCP in 18 of 22 specimens (82%). Bronchial brushings were insensitive, yielding a positive specimen in only 30% of cases of PCP. Alveolar casts of PCP were identified by Papanicolaou-stained slides of wash/lavage specimens in 83% of cases of Pneumocystis pneumonia. These proteinaceous alveolar casts were not seen in other pulmonary disorders. Encysted sporozoites were found in 56% of cases in which Papanicolaou-stained alveolar casts were identified. We conclude that the diagnosis of PCP can be made rapidly and reliably on the Papanicolaou-stained bronchial wash/lavage or bronchial brush specimens by detecting the characteristic alveolar casts, which contain P. carinii-encysted sporozoites. The presence of encysted sporozoites within alveolar casts is pathognomonic for PCP, and methenamine silver stains can be eliminated in those cases in which encysted sporozoites are identified.

Loss of blood isoantigens in exfoliated cells during the progression of CIN demonstrated by monoclonal antibody staining.

The presence of isoantigens A, B and H in exfoliated cervical epithelial cells was readily demonstrable in cervical smears using monoclonal antibodies and an immunoperoxidase staining technique. The progressive decrease in the percentage of immunoperoxidase-positive dysplastic cells and in the proportion of cytologically normal squamous cells associated with these dysplastic cells reflected the loss of isoantigen expression that paralleled increasingly severe stages of cervical intraepithelial neoplasia. The loss of isoantigen expression in morphologically normal epithelial cells and in dysplastic cells may indicate that a particular patient is at greater risk of disease progression and that close follow-up is warranted.

Pemphigus vulgaris with subclinical involvement of the uterine cervix. Report of a case with persistence of abnormal Papanicolaou smears posthysterectomy.

A case of pemphigus vulgaris is reported in which abnormal Papanicolaou smears continued prior to and after hysterectomy despite a normal pelvic examination. Microscopic foci of cervical pemphigus were found only in a retrospective study of the surgical material. Subclinical involvement of the cervicovaginal area with pemphigus, especially in a patient whose disease is apparently under control with steroids, can be a source of atypical Papanicolaou smears. The importance of careful colposcopic examination and clinical history is stressed for the gynecologist and cytopathologist, respectively.

Detection of murine immunoglobulin in human tissues following therapeutic infusion of monoclonal antibody.

A class switch variant of hybridoma CO19-9 secreting IgG2a antibodies was shown to have the same immunoperoxidase binding pattern in human tissue as the IgG1 antibody secreted by the parental hybridoma. The IP tissue binding of GA73.3 and 17-1A monoclonal antibodies which have been suggested to bind to structurally related antigens were compared; although quite similar in distribution, some differences were noted. GA73.3 bound to 12/12 colon carcinomas compared to 17-1A which reacted with 11/12 tumors. In several cases, the percentage of cells reactive with GA73.3 (90-100%) exceeded those reactive with 17-1A (10-25%). Additionally, the intensity of reactivity for GA73.3 was consistently greater than that seen with 17-1A. The detection of murine antibody bound to human tissues following therapeutic infusion of 19-9 IgG2a or GA73.3 differed. Detection of antibody-antigen complexes was seen less often in patients who had received 19-9 IgG2a (2/6) than in patients who had received GA73.3 (5/5). Additionally, the presence of murine immunoglobulin was seen only in the extracellular mucin of the patients receiving 19-9 IgG2a, whereas strong cellular binding of murine immunoglobulin was noted following infusion with GA73.3.

Comparative study of the binding characteristics of monoclonal antimelanoma antibodies.

The binding specificities of monoclonal antibodies against human cutaneous malignant melanoma were analyzed using radioimmunoassay (RIA), mixed hemadsorption assay (MHA), and peroxidase-antiperoxidase assay on a variety of malignant and nonmalignant cells. Twenty-four of the 30 monoclonal antibodies bound to the majority of melanoma cell lines tested, and only two antibodies did not bind to any of the melanoma lines. Three antibodies bound to melanoma lines only, 13 antibodies reacted also with fetal cells, 21 antibodies bound to at least one carcinoma cell line and nine antibodies reacted with one or more cell lines of leukemic or lymphoid origin. The antibodies could be divided into seven groups based on their binding characteristics. These groups had been established by a panel of monoclonal antimelanoma antibodies produced and characterized at The Wistar Institute.

Monoclonal antibody to a highly glycosylated protein reacts in fixed tissue with melanoma and other tumors.

A highly glycosylated protein with a molecular weight of 30,000 to 60,000 and a protein core of 20,000 daltons has been identified by antimelanoma monoclonal antibodies. The antigenicity of this melanoma-associated glycoprotein (MAG) was not destroyed in fixed paraffin-embedded melanoma tissue, and was present in malignant cells of cutaneous superficial spreading melanomas in skin (31/33) and in half of all metastatic melanomas examined (5/10). The antigen was not expressed by normal melanocytes. The strong reactivity of dysplastic nevi with the anti-MAG antibodies was comparable to that seen in radial growth phase melanoma. Antigen expression was much weaker in compound nevi where reactivity ranged from moderate in the junctional component and the upper dermis to absent at the base of the nevus.

Differential diagnosis between mesotheliomas and metastatic adenocarcinomas using monoclonal antibodies against gastrointestinal carcinoma antigen and stage-specific embryonic antigen.

Monoclonal antibodies made against gastrointestinal carcinoma antigen (GICA) and stage specific embryonic antigen (SSEA) were evaluated for their ability to distinguish normal mesothelial cells present in pleural and peritoneal fluids from adenocarcinoma cells in tissue and cytology specimens. The presence of GICA was documented in a high percentage of adenocarcinomas from the gastrointestinal tract (75/98) and in 52% of pulmonary (15/29) and 29% of ovarian (6/21) adenocarcinomas. GICA was found infrequently in breast carcinoma (1/18) and not in mesotheliomas (0/16). A similar pattern of GICA expression was seen in malignant effusions from adenocarcinomas (18/47) and mesotheliomas (0/6). SSEA was found in a high percentage of adenocarcinomas derived from the gastrointestinal tract (47/56) and the lung (26/29). SSEA was detected in breast carcinoma (8/15) more often than GICA. SSEA was detected rarely in mesotheliomas (1/16). Reactivities for epithelial membrane antigen, keratin, carcinoembryonic antigen, GICA and SSEA in adenocarcinoma and mesotheliomas were compared.

A fetal glycolipid expressed on adenocarcinomas of the colon.

Reactivity of neuraminidase-treated colorectal carcinoma cells with antibodies that detect the X-carbohydrate structure was greater than the reactivity of untreated cells. The same results were obtained with glycolipid extracts of meconium, a colorectal carcinoma cell line, three freshly excised human adenocarcinomas, and normal bronchial mucosa. The glycolipid was either not expressed or expressed in smaller quantities on the corresponding normal colon tissue. Further study showed that the major sialo-X glycolipid has six sugars including a single sialic acid which blocks X-antigenicity. These glycolipids were further analyzed by ion-exchange high-pressure liquid chromatography and thin-layer chromatography. These monosialo-X glycolipid antigens might serve as potential tumor markers.

Characterization of nerve growth factor receptor in neural crest tumors using monoclonal antibodies.

The nerve growth factor (NGF) receptor was characterized by using a new series of anti-receptor monoclonal antibodies (MAbs). These MAbs (i) showed significantly greater reactivity with a melanoma cell line expressing higher levels of NGF receptor, (ii) inhibited the binding of 125I-labeled NGF to its receptor, and (iii) immunoprecipitated both metabolically labeled and 125I-labeled NGF affinity-labeled receptor. These experiments defined the receptor as a 75-kDa cell-surface protein. The NGF receptor was visualized by immunoperoxidase staining in tissue sections of human nevi, melanomas, neurofibromas, a pheochromocytoma, and peripheral nerves. Uniform staining of the cytoplasm suggests that, in addition to cell-surface NGF receptors, there is a population of intracellular receptors.