A new marker for human cancer cells. 3. Immunocytochemical detection of malignant cells in serous fluids with the Ca1 antibody

Diagnostic Interest of the Monoclonal Antibody Ca1 in Malignant Pleural Effusion

The Ca1 antibody was used in an immunohistochemical procedure on smears of cells from 40 patients with malignant pleural effusion. The control group consisted of 25 benign pleural effusions with a high percentage of reactive mesothelial cells. The Ca1 Me Ab was positive in 19 (79%) of the 24 pleural effusions with positive malignant cytology. In all the benign cases the Ca1 Mc Ab was negative (100% specificity). The Ca1 Mc Ab detected malignant mesothelial cells in two cases and was negative with reactive mesothelial cells and other nucleated cells present in the pleural effusion. We conclude that the Ca1 antibody offers a useful diagnostic method for malignant pleural effusions, when the morphological interpretation is doubtful.

CD 34 immunotyping of blasts in myelodysplasia

We studied the expression of the hematopoietic progenitor cell antigen CD 34 in six patients with refractory anemia with excess of blasts (RAEB), five patients with RAEB in transformation (RAEB-T), and seven patients with chronic myelomonocytic leukemia (CMML). Immunocytochemical labeling of bone marrow cells was performed by an indirect immunoperoxidase method with preservation of morphological details. The cells were stained with May-Grünwald-Giemsa, photographed, destained, and immunolabeled by the immunoperoxidase technique. We found 1.5 +/- 0.5% blasts and 0.8 +/- 0.4% CD 34+ blasts in normal bone marrow. The CD 34 positivity of blasts was 53 +/- 9%. The patients with RAEB showed 1.7 +/- 1.4% CD 34+ blasts. The CD 34 positivity of blasts (11.8 +/- 5.6%) was lower than in normal bone marrow. The patients with RAEB-T had a higher percentage of CD 34+ blasts (7.3 +/- 3.4) and a higher CD 34 positivity of blasts (28.2 +/- 14.6%) than patients with RAEB. The CMML patients showed a percentage of CD 34+ blasts and a CD 34 positivity of blasts in the range of RAEB. We found an increase of promonocytes (PMC) in 5/7 patients. In some patients the PMC were CD 34 positive. Our results indicate that the increase of blasts in REAB is related to CD 34-negative blasts. With progression to RAEB-T the percentage of CD 34-positive blasts increased. Some of the CMML patients also showed a population of CD 34-positive PMC. A clone of undifferentiated CD 34-positive cells is characteristic for patients with these types of myelodysplasia.

Immunotyping of blasts in refractory anaemia with excess of blasts

A study of immunological markers was performed in 16 patients with newly diagnosed refractory anaemia with excess of blasts (RAEB) and RAEB in transformation (RAEB-T) and in 12 other patients with acute myeloid leukaemia evolving from RAEB or RAEB-T. Immunocytochemical investigation of bone marrow blasts was done using a modified indirect immunoperoxidase technique. This method permitted accurate morphological identification of blasts and other cells in bone marrow. The monoclonal antibodies used in RAEB and RAEB-T samples were anti-CD34, -c-kit, -HLA-DR and -CD13. The range of CD34 expression of blasts in RAEB samples was 1-14% (mean 6.2%) and in RAEB-T samples 29-48% (mean 35.5%). CD34 positivity was detected in 3-94% (mean 47.4%) of the bone marrow blasts in acute myeloid leukaemia evolving from RAEB and RAEB-T. Expression of c-kit was demonstrated only in a low percentage of blast cells in RAEB, RAEB-T and acute myeloid leukaemia following myelodysplasia. A high percentage (> 30%) of blasts in most patients with RAEB, RAEB-T and acute myeloid leukaemia was HLA-DR and CD13 positive. We observed the transformation from RAEB to acute myeloid leukaemia in three patients. The proportion of CD34 positive blasts increased to 25% and 32% in two patients. The third patient showed an unchanged percentage of CD34 positivity of blasts. These findings indicate that the CD34 positivity of blasts increases with the progression of myelodysplasia to RAEB-T and acute myeloid leukaemia demonstrating the instability of the clonal defect in myelodysplasia.

Immunophenotyping of mesothelial cells and carcinoma cells with monoclonal antibodies to cytokeratins, vimentin, CEA and EMA improves the cytodiagnosis of serous effusions

This paper presents an immunocytochemical study performed on cytocentrifuged deposits from 109 peritoneal and pleural effusions including 20 transudates, 43 malignant metastatic effusions and 46 effusions containing atypical cells, unidentifiable as reactive mesothelial or malignant epithelial cells on the classical morphological criteria. A panel of four monoclonal antibodies (MAb) was used, including KL1 directed to cytokeratins (KER), V9 to vimentin (VIM), NEO 723 to carcinoembryonic antigen (CEA) and E29 to epithelial membrane antigen (EMA). In most transudates the reactive mesothelial cells coexpressed VIM and KER with a ring-like pattern for the latter proteins. In contrast, they were unreactive to anti-CEA and weakly and inconsistently reactive to anti-EMA. In malignant effusions, most carcinoma cells coexpressed EMA, CEA and KER with a predominant diffuse cytoplasmic pattern for the latter. Only a few malignant epithelial cells from five metastatic adenocarcinomas weakly expressed VIM. When used on the 46 effusions with unidentifiable cells, the panel of MAb allowed reactive mesothelial cells and malignant epithelial cells to be distinguished from each other in 39 of 46 cases (85%).

Use of monoclonal antibody 17.13 in the diagnosis of human oral squamous cell carcinoma

We assessed the ability of monoclonal antibody (MAb) 17.13 to react with human oral squamous cell carcinoma (OSCC) and other oral tissues. MAb 17.13 was reacted with acetone-fixed, serial frozen sections of 21 OSCC, 7 fibromas, 2 squamous papillomas, 1 melanosis and 3 normal oral mucosa, by means of an avidin-biotin-immunoperoxidase assay. Twenty-one of 21 tumors (100% sensitivity) reacted strongly with MAb 17.13, with a homogeneous staining pattern. Individual tumor cells could be clearly seen, improving the detection of microinvasion or borderline lesions. The epithelium of benign lesions and normal oral mucosa showed staining of the basal cells only. We conclude MAb 17.13 could be used as an accurate diagnostic tool for OSCC.

Production and characterization of a new monoclonal antibody to colorectal carcinoma

This study describes a new murine monoclonal antibody (MoAb) 5C1 raised against human colorectal carcinoma, which gave a differential reaction on formalin-fixed sections of the gastrointestinal tract. The MoAb 5C1 of immunoglobulin M (IgM) isotype reacted with both the cytoplasm and membrane of all normal colonic epithelia, and with all benign colonic polyps and all premalignant colonic lesions. However, there was a decreased expression of the 5C1 antigen in most cases of colonic malignancy and it was this feature that makes MoAb 5C1 unique. The distribution of the 5C1 epitope in normal gastrointestinal tract is limited to a few epithelial cells in the mid-portion of the small intestine but this distribution increased progressively down the digestive tract until it was found on greater than 90% of normal epithelial cells (in membrane and cytoplasm) of the colon. In addition, the 5C1 epitope was present on mucin secreting cells from normal organs of the gastrointestinal, reproductive and pulmonary tract and benign and malignant tissues of the colon. On Western blots, MoAb 5C1 was found to detect a heterogeneous population of molecules with molecular weights greater than 100 kDa with the strongest staining bands found between 230 and 300 kDa. MoAb 5C1 does not detect carcino-embryonic antigens (CEA), human milk fat globules (HMFG), human lymphocyte antigens (HLA) or ABO blood group antigens. The combination of its presence in mucin secreting cells and its broad molecular weight bands suggest that the antigen detected is a mucin.

Serial measurements of Ca antigen in sera from patients with advanced malignant breast disease

Ca antigen levels in serum samples from three groups of patients were assayed. From a survey of 173 patients with various malignancies, elevated levels were found most consistently in patients with metastatic breast cancer. Spearman rank correlation values of Ca and CEA values on individual serum samples, 0.3009, (n = 194), or individual and serial samples, 0.2406, (n = 264) from a total 194 patients with metastatic breast cancer showed that correlation between Ca and CEA values was poor. For a group of 20 patients within the 194, from whom fortnightly serial samples were available, serum levels for 10/20, measured retrospectively, corroborated clinical observations on the course of their disease, although only 4/20 showed marked elevations during active disease. No correlations between expression of the tumour marker and histological type of the primary tumour, age of the patient, site of recurrence nor aberrant elevation in response to cytotoxic drug could be found to explain the non-correspondence of marker behaviour and disease status in the remaining 10 patients. The indications from this small study are that serial Ca antigen serum measurement could be misleading in 50% of patients with metastatic breast cancer, and that the assay is unsuitable for follow-up of patients with breast cancer.

The value of immunocytochemical staining of lymph node aspirates in diagnostic cytology

This study investigates the applicability of immunocytochemical techniques towards improving the cytological diagnosis of lymph node disorders. Cytocentrifuge preparations of fine needle aspirates were examined using an indirect immunoperoxidase method and the alkaline phosphatase-antialkaline phosphatase method. 36 reactive lymph nodes were evaluated. 23 patients showed T cell predominance as assessed by the presence of CD3+ cells. 13 patients showed an excess of CD24+ cells over CD3+ cells. In all patients the B cell population showed no light chain class restriction. 48 aspirates from patients with B-non-Hodgkin's lymphoma of low malignancy were investigated. The majority of the cells in each of these cases expressed CD24 and 47 cases were monoclonal with respect to their light chain determinants. In 37/48 cases the antibody OKT9 reacted with less than 15% of cells. 18 aspirates were obtained from patients with B-non-Hodgkin's lymphoma of high malignancy. 17 patients showed light chain class restriction and a high percentage of CD24+ (greater than 80%) and OKT9+ (greater than 40%) cells. 8 malignancies were considered to be of T-cell origin. A low percentage (less than 15%) of Ig+CD24+ cells with a high number of CD3+ and/or CD4+ cells suggests a T-cell lymphoma. The majority of neoplastic cells in peripheral T-cell lymphomas (excluding mycosis fungoides) expressed OKT9 and CD3 expression was found to be variable. In 15 cases of Hodgkin's disease, there were 11 correct cytologic diagnoses. A significant number of BerH2 (CD30)+ cells having the morphology of Sternberg-Reed cells supports this diagnosis. 25 aspirates were obtained from patients with metastatic malignant tumours. Marker studies in eight cases helped in distinguishing metastatic malignant tumours from malignant lymphomas. Our findings indicate that the immunocytochemical technique is applicable to cytological material and that the simplicity of the procedure merits application to routine diagnostic cytology.

Use of antimesothelial cell antibody and computer assisted quantitative analysis for distinguishing between reactive and neoplastic serosal tissues

Nine malignant mesotheliomas and 12 specimens of benign reactive mesothelial hyperplasia were examined using a specific antimesothelial cell antibody. Immunostaining intensity was subsequently estimated by means of image analysis. The mean and standard deviation of integrated optical density in the malignant mesotheliomas differed significantly from those in the mesothelial hyperplasias. In all mesotheliomas but one the integrated optical density was greater than in reactive mesothelial hyperplasia. No significant difference in optical density was observed between the two groups; the standard deviation was significantly higher in the reactive lesions. This technique may be adopted to complement the traditional morphological assessment of primary lesions of the serosal cavities.

Pros and cons of gynecologic tumor markers

The use of a tumor marker not only depends upon its sensitivity and specificity, but also upon its ability to influence decisions between alternative plans for patient management. Use of beta human chorionic gonadotropin (hCG) for monitoring gestational trophoblastic neoplasia has set the standard to which other assays must be compared. Beta hCG and alphafetoprotein have provided useful markers for ovarian germ cell tumors. Recently, a monoclonal antibody-based immunoassay for CA 125 antigen has been used to monitor the treatment of epithelial ovarian carcinomas. Persistent elevation of CA 125 in serum has generally reflected persistence of disease at second look surveillance procedures. CA 125 levels can, however, return to within normal limits and residual disease can be found at laparoscopy or laparotomy. CA 125 shows promise for distinguishing benign from malignant pelvic masses. Trials are currently underway to evaluate the potential of CA 125 in combination with other markers to facilitate earlier detection of occult ovarian cancer.

Evaluation of the Ca 1 monoclonal antibody in distinguishing oral squamous cell carcinoma from non-malignant or premalignant oral lesions

The Ca antigen has been reported to be present on the surface of malignant but not, with few exceptions, non-malignant cells. We investigated the potential usefulness of the monoclonal Ca 1 antibody in differentiating oral squamous cell carcinoma from non-malignant or premalignant oral neoplasms. Paraffin-embedded sections from 33 biopsy specimens of 12 hyperplastic and 21 neoplastic oral lesions were examined by immunohistochemical staining. Seven of the 33 specimens showed positive staining for Ca antigen. Fifteen of 21 specimens of oral squamous cell carcinoma were negative for Ca antigen, and one case of focal keratosis was positive. The results indicate that the use of the Ca 1 antibody to distinguish oral squamous cell carcinoma from non-malignant or premalignant oral lesions is highly unreliable.

[Prospective comparison of the diagnostic value of cytology and immunocytology in pleural effusion studied by thoracoscopy and biopsy]

Two issues have been elaborated: the value of immunocytochemistry in the diagnosis of pleural effusions, and the reactivity of the investigated antibodies with different classes of cells in pleural effusions. Effusions of unknown origin from 38 patients were investigated using thoracoscopy, pleural biopsies, conventional cytology, and immunocytochemistry. The following antibodies were used: those monoclonal against various leukocyte antigens, macrophage antigens, epithelial membrane antigen (EMA), various cytoskeleton antigens, and melanoma antigens; those polyclonal against CEA and ferritin. All of the techniques used showed 18 patients (48%) as having a tumor-cell negative effusion. A pleural tumor with a malignant effusion showed in 13 patients (34%); in 12 of these immunocytochemistry also revealed tumor cells. Seven patients (18%) had a tumor of the pleura with a tumor-cell negative effusion; in 2 of these immunocytochemistry revealed a tumor-cell positive effusion. There was no difference with regard to the number of NK cells in patients with inflammation of the pleura and negative cytology and patients with tumor of the pleura and malignant effusion (3% vs 4.5%). Tumor cells were mainly stained by EMA, cytokeratin, and CEA. CEA was the only antibody to be tumor-cell specific, while EMA and cytokeratin were expressed by mesothelial cells also. The antibody against ferritin was a significant marker for mesothelial cells.

Applications of immunocytochemistry to clinical cytology

This article reviews the recent studies reporting the applications of immunocytochemistry to diagnostic problems in clinical cytology. A series of studies with monoclonal antibody (MAb) B72.3 is discussed in detail. MAb B72.3, reactive with a high molecular weight, glycoprotein, tumor-associated antigen, designated TAG-72, has been shown previously to be reactive with formalin-fixed, paraffin-embedded tissue sections of adenocarcinomas of the ovary, colon, and breast, but not a variety of normal adult tissues. It has demonstrated utility as an immunocytochemical adjunct to diagnose carcinoma in cell block and cytocentrifuge preparations of human serous effusions, with selective reactivity for tumor cells (particularly adenocarcinoma) over reactive mesothelium. Using the avidin-biotin complex (ABC) method of immunoperoxidase staining and formalin-fixed, paraffin-embedded cell suspensions, MAb B72.3 detected tumor cells in effusions from the majority of patients with adenocarcinoma of the breast. No reactivity was demonstrated in any cell type in benign effusions. In contrast, MAb B72.3 showed no reactivity to leukemic or lymphomatous effusions, or to mesothelial cells from malignant effusions. MAb B72.3 also detected tumor cells in effusion specimens from most of the patients with "non-small cell" carcinoma of the lung and with carcinoma of the ovary. MAb B72.3 was also used with fine-needle aspiration biopsies (FNABs) and corresponding surgically excised tumors to determine cellular reactivity. Positive staining with MAb B72.3 was observed in needle aspirates of the great majority of "non-small cell" carcinomas of the lung, adenocarcinomas of the breast, adenocarcinomas of the colon, and carcinomas from other body sites. In contrast, small cell carcinomas of the lung, malignant melanomas, lymphomas, sarcomas, and glial tumors stained negatively with the antibody. Most benign lesions from the breast, lung, pancreas, parotid, and thyroid also showed no staining. In many patients, tumor-bearing tissue had also been resected and was available for comparative examination with MAb B723. In more than 90% of these patients, the staining patterns of tumor cells in the aspirates were found to be predictive of the patterns of antibody reactivity in the comparable surgically resected tumors. From these studies it is concluded that MAb B72.3 defines a tumor-associated antigen that is expressed in neoplastic cells versus benign cells, is most selectively expressed in carcinomas, and may be used as a novel adjunct for the diagnosis of neoplasms in effusions and in FNABs.

Use of immunocytochemical staining to identify cells in peritoneal fluid and washings at laparoscopy and laparotomy

Specimens of peritoneal fluid or peritoneal washings from a series of 106 patients who had had laparotomy or laparoscopy for gynaecological complaints were studied "blind" by conventional cytology and immunocytochemical staining. The antibodies used were Ca 1 or Ca 2, anti-CEA, and HMFG-2 or E29. All these are directed against epithelial antigens and are expressed on most malignant epithelial cells and weakly or not at all on mesothelial cells. It was hoped that these reactions would confirm diagnoses made by conventional cytology and possibly show malignant cells which had not already been identified. Of 28 patients with malignant disease (chosen to exclude any with frank ascites), eight gave positive immunochemical reactions, only four having been reported positive from conventional examination. Of 77 patients without malignant disease, HMFG-2 or E29 gave positive reactions in seven, Ca 1 or Ca 2 in two, and anti-CEA in two (reactions with plasma cells being disregarded). Some misleading reactions were probably due to endometrial cells. It was concluded that the antibodies used in this study are not sufficiently specific or sensitive to allow immunocytochemical staining to replace conventional cytological diagnosis but are a useful supplementary aid.

Detection of Ca antigen in sera from normal individuals and patients with benign and malignant breast disease

Two assay procedures, an inhibition radioimmunoassay (Inhibition-RIA) and an immunoradiometric assay (IRMA), were established for the detection of circulating tumour-associated Ca antigen. There was a good correlation between results (r = 0.987) but the Inhibition-RIA was selected for extended tests on human sera from patients with breast disease because of its greater ease and economy in use. Circulating Ca antigen was not exclusive to malignancy and the level failed to discriminate between patients with primary carcinoma and those with benign disease. Ca antigen was present in sera of 100 healthy individuals (median 7.1 micrograms ml-1, range 1.8-24.4 micrograms ml)-1, 39 patients with benign disease (median 9.9 micrograms ml-1, range 2.5- greater than 100 micrograms ml-1) and in 67 patients with primary carcinoma (median 11.0 micrograms ml-1, range 3.8- greater than 100 micrograms ml-1). Elevated Ca antigen levels were found in 50% of patients with metastatic spread (median 30.7 micrograms ml-1, range 8.2- greater than 100 micrograms ml-1) and in some patients with primary disease but further studies are needed to determine the prognostic significance. Immunochemical studies confirmed that Ca antigen is a normal serum product but its function is unclear.

Ca2 and Ca3. New monoclonal antibodies evaluated as tumor markers in serous effusions

Ca2 and Ca3 are new monoclonal antibodies of IgG1 class, directed against the Ca antigen, a mucus-type glycoprotein expressed on the surface of a wide range of malignant human cells and certain specialized normal epithelia. These antibodies were produced by immunization with purified preparations of the Ca antigen. They were tested to assess their value in the diagnosis of malignant effusions. Immuno-alkaline-phosphatase staining was used. Smears of pleural and peritoneal effusions were chosen to show: undoubted malignant cells of various types; and mesothelial cells in effusions from cases in which cancer was not in question. The Ca2 antibody, at 1 in 20 dilution of the culture supernatant, was the most specific, giving no reactions with benign mesothelial cells from any of the 35 cases tested. Malignant cells were clearly stained in 35 of 40 cases of carcinoma or mesothelioma. The staining was negative in two cases of oat cell bronchial carcinoma, and in three of four cases of carcinoma of the colon. Ca3 gave similar, but somewhat stronger, reactions with carcinoma cells, but was less specific, reacting weakly with mesothelial cells in 8 of 35 benign effusions. Because the false-negative reactions given by the Ca series of antibodies are to some extent complementary to those given by monoclonal antibodies directed against the carcinoembryonic antigen (CEA), a combination of Ca2 and anti-CEA is recommended as a most useful addition to the normal cytologic examination of effusions.

Flow cytometric identification of cancer cells in effusions with Ca1 monoclonal antibody

The fluorescence of cells from 42 pleural and peritoneal effusions stained with Ca1 monoclonal antibody (Ca1MA) was studied by flow cytometry. In 14 of 17 malignant effusions a significantly higher intensity of fluorescence was observed in samples exposed to Ca1MA when compared with controls. There was no increase of fluorescence intensity in 25 benign effusions. The method failed in three malignant effusions: one due to endometrial carcinoma and two to malignant lymphoma. The sensitivity of the method was tested in experimental samples with a known percentage of malignant cells. The positive fluorescence with Ca1MA was detected in samples containing 0.1% of carcinoma cells. Flow cytometry with Ca1MA can be a relatively simple method of identification of malignant cells in effusions.

Expression of Ca antigen on human urinary bladder tumors

Flow cytometric measurements of Ca antigen (Ca) visualized by immunofluorescence with Ca1 monoclonal antibody (Ca1) were performed on cells from 26 urinary bladder carcinomas of various histologic grades and DNA ploidy and on 6 samples of morphologically normal urothelium. In normal urothelium and in the 14 diploid tumors the proportion of positively staining cells was low and the differences were not statistically significant. A significant increase in the proportion of fluorescent cells was observed in all but one of the 12 aneuploid tumors. Among the 11 tumors with an increased proportion of positive cells, 7 were invasive. Data recorded in this study suggest that the quantitation of Ca1 fluorescence corresponding to the presence of Ca may prove to be helpful in the evaluation of clinical behavior of bladder tumors.

Immunocytochemical reaction of Ca1 and HMFG2 monoclonal antibodies with cells from serous effusions

The Ca1 antibody was used in an alkaline phosphatase immunocytochemical method on cells obtained from 150 specimens of pleural and ascitic fluids. The results were compared with the routine cytology report based on the light microscopical appearances. The Ca1 antibody identified tumour cells in 51 of 57 specimens with malignant cells. The exceptions were four small cell carcinomas, one malignant lymphoma, and one adenocarcinoma. A further seven specimens reported as containing atypical cells but without conclusive evidence of malignancy were Ca1 positive. The Ca1 antibody did not give a positive reaction with benign mesothelial cells. Similar results were obtained with the HMFG2 antibody and malignant cells, but in eight of 18 benign effusions it reacted with mesothelial cells.

Use of monoclonal antibody panel to identify malignant cells in cerebrospinal fluid

A panel of monoclonal antibodies was systematically applied to cerebrospinal fluid from 17 patients with suspected neoplastic meningitis and the results were compared with those obtained from routine cytological preparations. The antibody panel consisted of markers for neuroectodermal tissue ( UJ13A ), epithelial cytokeratin ( LE61 ), leucocytes ( 2D1 ), and neoplastic neuroblasts ( UJ181 .4). Additional antibodies were used to refine diagnosis when indicated. Cerebrospinal fluid samples from 12 patients with non-neoplastic conditions were used as controls. The use of monoclonal antibodies gave a positive diagnosis in 16/17 cases and the cells were accurately categorised as carcinoma (5/6 cases), neuroectodermal tumour (8/8 cases), and lymphoma (3/3 cases). In the 14 cases examined by routine cytology, malignant cells were reported in 10 cases and accurately categorised in only 3/14 cases. Immunocytological testing of cerebrospinal fluid with an antibody panel has greatly increased the accuracy with which malignant cells can be identified and categorised.

Monoclonal antibodies for the histopathological diagnosis of cervical neoplasia

Five monoclonal antibodies (Ca1, HMFG 1 and 2, 8.30.3 and 77.1) were used to study the distribution of antibody binding sites in cervical tissue with a view to identifying a marker which would distinguish between benign and malignant cervical epithelium. Both benign tissue (mature and immature metaplastic squamous epithelium, congenital transformation zone and glandular epithelium) and neoplastic tissue (cervical intraepithelial neoplasia, 1, 2 and 3 and invasive squamous cell carcinoma) were stained by these antibodies. Although immature metaplastic epithelium stained strongly with all the antibodies, the intensity and distribution of staining in general did not distinguish between benign and neoplastic conditions. All five antibodies, raised against three different antigens, stained cervical tissue in a similar way and thus were unsuitable for use as specific tumour markers in equivocal cases. Further studies on other tumour markers are indicated in view of the potential value of this approach.

Distribution of the Ca (Oxford) antigen in lung neoplasms and non-neoplastic lung tissues

The Ca (Oxford) antigen was originally isolated from a malignant neoplasm and with few exceptions was reported to discriminate between malignant and non-malignant neoplasms or normal tissues. Using the Ca 1 antibody we have studied the Ca distribution in 54 lung neoplasms and adjacent non-neoplastic lung tissue. Staining of tumours was very focal and the proportion of positive cells varied from about 50% for adenocarcinomas to less than 1% for oat cell carcinomas, which were often negative. Focal cytoplasmic staining can be seen in all neoplasms, whereas membrane staining is mainly seen in their areas of glandular and squamous differentiation. We found consistently strong membrane staining of alveolar type II pneumocytes in non-neoplastic lung. This staining may be useful in differentiating type II cells from alveolar macrophages which only occasionally showed granular cytoplasmic staining, probably due to phagocytosed Ca. Mucin from tumours and bronchi did not stain but there was consistent staining of alveolar serous exudate suggesting extracellular location of Ca.

Immunocytochemical staining with monoclonal antibodies in cytologically "negative" serous effusions from patients with malignant disease

In order to assess the practical value of immunocytochemical techniques in routine cytological diagnosis, a panel of three monoclonal antibodies (anti-CEA, Ca 1 and HMFG-2) were used to search for malignant cells in 53 samples of pleural or peritoneal fluid from 41 selected patients. All these patients suffered from malignant disease, but neoplastic cells had not been reported on conventional cytological examination. In 12 of the 41 cases immunocytochemical staining revealed previously unrecognised malignant cells. This represents an increase in diagnostic accuracy of approximately 20% and suggests that immunocytochemical labelling should be used routinely for the examination of cytologically negative samples from patients with suspected malignant disease.

Immunocytochemical staining of cells in pleural and peritoneal effusions with a panel of monoclonal antibodies

A panel of seven monoclonal antibodies was applied to smears of cell deposit from 70 pleural and peritoneal fluids, using an immunoalkaline phosphatase (IAP) procedure. The cases were chosen to show typical cytological patterns, both benign and malignant, and in this way the diagnostic value of the method could be assessed. The antibodies used were 2D1 (anti-leucocyte), Ca 1, HMFG-2 (anti-milk fat globule membrane), LE61 and M73 (both anti-intermediate filament antibodies), anti-CEA, and K92 (anti-keratin). The anti-leucocyte antibody was found useful for distinguishing lymphoma from carcinoma. Anti-CEA gave positive reactions in 80% of carcinoma cases and was not seen to react with any other cell types. Ca 1 was positive with some cells in 95% of carcinoma cases, but mesothelial cells reacted with it in two cases. A strong reaction with the anti-milk fat globule membrane antibody was very constant in carcinoma but was also seen in mesothelial cells in 30% of benign effusions. The anti-keratin reacted with malignant cells in only a small proportion of cases. The antibodies against epithelial intermediate filaments reacted equally strongly with benign mesothelial cells and carcinoma cells, but gave negative reactions with lymphoma cells. It is concluded that a suitably chosen panel of monoclonal antibodies can be of great value in identifying neoplastic cells in serous effusions.

Structure and function of the Ca antigen

The Ca antigen, which can be detected in a wide range of malignant human tumours by means of the Cal antibody, is a glycoprotein of the mucin type. At least 95% of the carbohydrate is 0-glycosidically linked to the polypeptide which contains high proportions of glycine, serine and glutamic acid. The carbohydrate has a very simple structure: it is composed almost entirely of tetra- tri- and disaccharides having the general formula (NeuNac)n leads to [Gal leads to GalNac] alpha leads to, where n = 0, 1 or 2. In many malignant cell lines, the antigen is produced constitutively in vitro; but in one that has been examined, its synthesis can be induced by high concentrations of lactate. Evidence is presented for the view that a primary function of this glycoprotein is to shield the cells that produce it from hydrogen ion concentrations outside of the physiological range. The presence of the Ca antigen in malignant tumours may thus be a reflection of metabolic conditions that are known to be characteristics of such tumours.