Original Paper
Tumor Biol 1992;13:308-315
a Department of Medical Microbiology and Immunology, b Department of Pathology, c Department of Internal Medicine, 1st Section of Pulmonary Medicine, University of Göteborg, d Department of Thoracic Surgery, Sahlgren’s Hospital, Göteborg, and c Department of Immunology, University of Umeä, Sweden
Coexpression of Ganglioside Antigen Fuc-GM I, Neural-Cell Adhesion Molecule, Carcinoembryonic Antigen, and Carbohydrate Tumor-Associated Antigen CA 50 in Lung Cancer
Key Words Tumor-associated antigens Monoclonal antibodies Immunohistochemical study Pulmonary carcinoma NCAM Ganglioside Fuc-GM 1 CEA CA 50
Abstract With the aid of specific monoclonal antibodies, tumor tissues from 68 patients with lung cancer were examined for their expression of two small cell lung carcinoma (SCLC) antigens, Fuc-GMI (fucosyl GM1; IV2FuclPNeuAc GgOse4) and neu ral-cell adhesion molecule (NCAM), and two broader tumor antigens, carcinoembryonic antigen (CEA) and carbohydrate cancer-associated antigen CA 50. Expression of Fuc-GMI was seen in 75% and NCAM in 78% of the SCLC specimens, but also in 12 and 20% of non-SCLC. Either or both of these anti gens were expressed in more than 90% of SCLC and in 25% of non-SCLC. CEA was found in more than 80% of SCLC and non-SCLC. Expression of CA 50 was seen in 65-68% of nonSCLC and SCLC, showing preference for SCLC and lung ade nocarcinoma. In SCLC, cellular expression of Fuc-GM 1 was generally seen together with NCAM and CA 50, but rarely with CEA. There was considerable inter- and intratumor het erogeneity in the expression of all four antigens. The results suggest that CEA is the antigen of choice for the detection of lung cancer regardless of histotype. In combined analysis of CEA, CA 50, Fuc-GMI and NCAM. two patterns of antigen expression were recognized that appear to discriminate be tween SCLC and non-SCLC tumors, respectively. A consider able fraction of SCLC and non-SCLC tumors, however, exhib ited similar patterns of antigen expression. The biological and clinical significance of these observations remains to be inves tigated.
This study was supported by grants from the Medical Faculty. University of Göteborg, Sweden, the Swedish Cancer Society, and Pharmacia Ltd.
Received: September 17 . 1991 Accepted: May 26, 1992
F.-T. Brc/icka Department of MedicalMicrobiology and Immunology, University of Göteborg Guldhedsgatan 10
© 1992 S. Karger AG. Basel 1010-4283/92/ 0136-0308S2.75/0
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
Fred-Thomas Brezicka a - c Saute Ollingb Bengt Bergmanc Häkan Berggrend Carl-Peter Engström c Sten Hammarström c Jan Holmgrena Sture Larssond Leif Lindholm a
Despite important progress in the therapy of cancer during the last three decades, lung cancer (LC) still has a poor prognosis; e.g., with a 5-year survival rate of 5-10% for small cell LC (SCLC). Since SCLC is highly sensi tive to chemo- and radiation therapy, it is important to accurately identify and discrimi nate this tumor type from non-SCLC. The introduction of the hybridoma technology [1] has allowed for the production of a large num ber of monoclonal antibodies (MAbs) with high and uniform specificity that identify tu mor-associated antigens (TAAgs) expressed preferentially by SCLC tumors. We have pre viously described the production of MAbs FI 2 and FI 5 (both IgG3)that specifically bind to the ganglioside Fuc-GMl. By using these MAbs Fuc-GMl was found to be highly asso ciated with SCLC [2-5]. At the 1st and 2nd Workshops on SCLC Antigens [6, 7], MAbs raised against SCLC were characterized and clustered. MAbs of cluster 1 were all reactive with the neural-cell adhesion molecule (NCAM) [8] and appeared to have the great est potential in discriminating SCLC from non-SCLC. One member of this cluster is represented by MAb 123C3 (IgGi) [8-11], Here, we have examined the expression of four TAAgs in human lung tumor tissues representing the four major histotypes. Be sides the expression of Fuc-GM 1 and NCAM, that of two more generally expressed epithe lial TAAgs, carcinoembryonic antigen (CEA) [ 12] and carbohydrate-associated antigen CA 50 [13-17], was also studied. MAbs F I2 and FI 5, and 123C3 were used to specifically identify Fuc-GMl and NCAM. MAb 11-16 (IgGi) [18-20], produced after immunization with lung tumor-derived CEA, was used for the specific detection of CEA. The reactivity of this MAb panel with lung tumors and its ability to discriminate between SCLC and
non-SCLC was studied. We have also studied the cellular coexpression of Fuc-GMl and NCAM, CEA or CA 50 in SCLC tumors. The results indicate possible implications for fu ture diagnosis, prognosis and therapy of lung cancer.
Materials and Methods MAbs and Reagents The production and characterization of the FucGMl ganglioside-reactive MAbs FI 2 (IgG;,) and FI 5 (IgG^) were described earlier [3-5], Hybridoma culture medium of MAb FI 2 was precipitated with ammo nium sulfate, redissolved in PBS and dialyzed against PBS and then used for immunohistochemical studies. Purified MAb FI 5 was biotinylated [21] or used in native form. Mouse ascites preparations of NCAMrcactive MAb 123C3 (IgGi) [9,11 ] were a generous gift from Dr. K. Moolenaar (The Netherlands Cancer In stitute). Anti-CEA MAb II-16 (IgGi) was raised against purified LC CEA [18-20] and used in purified form. Purified MAb C50 (IgM) [13] defining CA 50 antigens (sialosyl Lea and sialosyllactotctraose [15, 17]), was obtained from Pharmacia CanAg (Göteborg, Sweden). Biotinylated horse anti-mouse antiserum (HAMB) and avidin-TMRITC (AvTMRITC) were purchased from Vector Laboratories Inc. (Burlingame, Calif.), Goat (FAb) anti-mouse IgG-FITC (GAMF) from Cappel Laboratories (Malvern, Pa.), rabbit anti-goat Ig-FITC (RAGF), swine anti-rabbit Ig-FITC (SwARF), rabbit anti-mouse Ig-FITC (RAMF) and ABComplex (K 355) from Dakopatts (Copenhagen, Denmark). The anti bodies and conjugates were diluted in PBS with 5% normal human serum and 0.1 % BSA after titration on control tumor tissues or according to the instructions supplied by the manufacturer. The following concen trations and dilutions were used: FI 2 10-100 pg/ml; F I5 200 pg/ml: biotinylated F I5 (F15B) 120 pg/ml; 123C3 diluted 1/50:11-16 10pg/ml: MAbC50 50-100 pg/ml; GAMF 1/50; RAGF 1/10; SwARF 1/10-1/20: RAMF 1/10. and AvTMRITC 1/50. Stainings using an IgG3MAb(protein-A purified) or irrelevant specificity or dilution buffer only replacing the specific MAbs were done as negative controls at each staining ses sion.
309
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
Introduction
Staining Procedures Tissue sections were fixed in 4% paraformalde hyde in PBS for 30 min at room temperature and rinsed three times with PBS for 5 min. For staining with MAbs 11-16, C50, 123C3, and F12, the ABC method was used. The following steps were performed in sequence: (1) incubation with 5% normal horse serum and 1% BSA in PBS for 15 min; (2) MAb; (3) HAMB. and (4) ABComplex. Each step was per formed for 30 min at room temperature in a humid chamber and followed by rinsing two to three times with PBS for 5 min. Sections were then incubated with 10 mg 3-amino-9-cthylcarbazol and 4 pi 30% HiCL in 6 ml DMSO and 50 ml 0.02 M NaAc (pH 5.5) for 15 min, rinsed twice with PBS for 5 min and once with distilled water, counterstained with Mayers hematoxy lin, and finally mounted with glycerin gelatin. For staining with MAbs FI 2 and FI 5, a double-step immu nofluorescence (IF) method [4] was also use: (1) MAb; (2) RAMF. and (3) SwARF. The ABC staining method was used to assess expression of NCAM, CEA and CA 50, whereas the IF method was used to assess expres sion of Fuc-GM 1. For the simultaneous detection of two separate antigens on individual tumor cells, bioti nylated MAb F I5 (F l5B) was used in conjuction with a double-fluorochrome IF staining method consisting of the following steps: ( 1) MAb (unconjugated); (2) GAMF; (3) RAGF+F15B, and (4) SwARF + AvTMRITC. In the IF methods, each step was fol
310
lowed by rinsing three times with PBS for 5 min, and the slides were finally mounted with 87% glycerin in PBS.
Results Expression of CEA, CA 50, Fuc-GM 1 and NCAM in lung tumors was studied by using highly specific MAbs and sensitive immunohistological staining methods. Expression o f CEA and CA 50 in Lung Tumors The results are summarized in table 1. CEA expression was seen in 82% of nonSCLC and in 85% of SCLC tumors. No pref erence for SQLC, ADLC or LCLC, collec tively denoted non-SCLC, or for SCLC was seen. In SQLC and ADLC, the expression of CEA was generally seen in highly differen tiated areas with various degrees of keratinization or adenoid formations. In these, cen trally located cells or cells lining the glands were most strongly stained. High expression was also occasionally found in less well-differ entiated areas containing conglomerates of positively stained cells, especially in low dif ferentiated non-SCLC. This pattern was also frequently seen in SCLC tumors (fig. la). The intensity and extension of MAb 11-16 staining of LC tumors showed a high degree of intraand intertumor heterogeneity. Also CA 50 was frequently expressed in SCLC (68%; fig. lb) and non-SCLC (65%). A tendency towards preferential expression of CA 50 in SCLC and ADLC tumors was seen. Of these, more than 60% of the cases showed wide (more than 25% of the tumor cells) expression, whereas this degree of expression was seen in only 30% of the SQLC cases. In non-SCLC, the morphological components expressing CA 50 were similar to those ex pressing CEA described above. In less well-
Brezicka/Olling/Bergman/Berggren/ Engstrôm/Hammarstrôm/Holmgren/ Larsson/Lindholm
Tumor Antigen Coexpression in Lung Cancer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
Tumor Tissue Specimens Tumor tissues from 68 patients with LC were col lected at thoracotomy or at autopsy. Twenty-seven of these were resected at thoracotomy and included 14 cases with squamous cell LC (SQLC), 3 with SCLC, 8 with adenocarcinoma (ADLC) and 2 with large cell LC (LCLC). The remaining 41 cases were taken at autopsy and included 12 cases with SQLC, 25 with SCLC, 3 with ADLC, and 1 with LCLC. Tissues were embed ded in mounting medium, snap-frozen in liquid nitro gen and stored at -7 0 °C until used. Cryostat sections (2-4 pm) were cut and put on glass slides. Slides were air-dried and stored at -7 0 °C until used for immuno histochemical staining. Representative sections from each tissue sample were stained with van Gieson's stains or Mayers hematoxylin for histopathological examination and assessment of diagnosis. These were also compared with routinely histopathologically pro cessed biopsies taken at bronchoscopy, mediastinos copy or thoracotomy. The histotype of each lung tumor specimen was assessed by light microscopic cri teria [22],
Table 1. Expression of CEA, CA 50, NCAM and Fuc-GM l in lung tumors Antigen
Score1
Lung tumor histotype SQLC
CEA
CA 50
NCAM
Fuc-GM 1
n
0 1 2 0 1 2 0 1 2 0 1 2
SCLC
ADLC
LCLC
number of cases
%
number of cases
%
number of cases
%
number of cases
3 4 19 9 9 8 21 0 5 25 1 0
12 15 73 35 35 30 81
4 4 20 9 2 17 6 4 18 7 7 14
14 14 71 32 7 61 21 14 64 25 25 50
3 0 8 3 1 7 8 1 2 7 3 1
27
1 1 1 2 1 0 3 0 0 3 0 0
19 96 4
26
28
11
73 27 10 63 73 10 18 64 27 9
3
differentiated areas of non-SCLC, and in SCLC tumors, cell aggregates of various sizes were homogenously stained with MAb C50. Expression o f Fuc-GM 1 and NCAM in Lung Tumors Expression of Fuc-GM 1 and NCAM was highly restricted to the SCLC histotype. Sev enty-five and 78% of the SCLC cases ex pressed these antigens, respectively. The ex pression of both antigens was homogeneous and generally comprised large areas of the tumors (fig. lc, d). Significant expression of Fuc-GM 1 and NCAM was also seen in nonSCLC (10 and 20%, respectively). In nonSCLC, expression of Fuc-GM 1 or NCAM was generally seen in less well-differentiated ar eas.
Combined Expression ofTAAgs in Lung Tumors Expression of both Fuc-GM 1 and NCAM in individual SCLC tumors was observed in 61 % of the cases (table 2), whereas 93% ex pressed at least one of these antigens. At least one of these antigens was also expressed in 25% of the non-SCLC cases, whereas expres sion of both was only seen in 8%. Expression of both CFA and CA 50 was seen in 61 % of the SCLC and in 55% of the non-SCLC tumors (fig. 2). A minority (728%) of SCLC and non-SCLC tumors showed expression of either CEA or CA 50. Expres sion of at least one of these antigens was seen in 91 % of SCLC and in 92% of non-SCLC. The combined expression of Fuc-GM 1, NCAM, CEA and CA 50 observed in lung tumors is summarized in figure 2. Two pat-
311
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
1 Score: 0 = negative reaction; 1 = < 25 % of tumor cells showing positive staining reaction; 2 = > 25 % of tumor cells showing positive staining reaction.
SCLC
non-SCLC
CA 50
Fuc-GM 1 and/or NCAM
Fuc-GMl and/or NCAM
Fig. 2. Combined expression of CEA, Ca 50 and Fuc-GMl and/or NCAM in SCLC and non-SCLC.
Histotype
F+N-
F+N+
F-N+
F-N-
n %
n
n
n
%
%
%
SCLC Non-SCLC
28 40
4 14 2 5
17 61 3 8
5 18 5 12
2 7 30 75
Total
68
6
20 29
10 15
32 47
terns of antigen expression appeared that seem to characterize SCLC and non-SCLC tumors, respectively. Expression of at least one of NCAM or Fuc-GM 1(N/F+) but not CA 50 (N/F+ CA 50") was restricted to SCLC tumors (32%), whereas N/F~ CEA+ was re stricted to non-SCLC (60%). There were large fractions of SCLC (68%) and non-SCLC (32%) tumors that showed overlapping pat terns of antigen expression. Cellular Coexpression ofTAAgs in SCLC Coexpression of two antigens (Fuc-GMl vs. NCAM, CEA and CA 50) was studied on the cellular level in 8 SCLC cases by using a
312
n
9
double-fluorochrome IF method. Fuc-GMl was frequently found to be coexpressed with NCAM (all cases) or CA-50 (7 cases). Expres sion of NCAM was either congruent with that of Fuc-GM 1 (3 cases) or more widely distrib uted including also Fuc-GM 1-expressing cells (5 cases). These patterns were also seen for coexpression of Fuc-GM 1 and CA 50 (1 and 6 cases, respectively; fig le). Fuc-GMl was oc casionally found to be coexpressed with CEA only on small fractions of the tumor cells (4 cases; fig. 10- Generally, CEA and Fuc-GMl expression was complementary (7 cases), i.e., the tumors contained areas expressing either of these antigens.
Brezicka/Olling/Bergman/Berggrcn/ Engstrôm/Hammarstrôm/Holmgren/ Earsson/Lindholm
Tumor Antigen Coexpression in Lung Cancer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
Table 2. Combined expression of Fuc-GMl and NCAM in SCLC and non-SCLC lung tumors. F+ and N+denote expression of FucGMl and NCAM, respectively, and F" and N_ lack of detectable antigen.
In this study we have used well-character ized MAbs for the detection of four TAAgs, CEA, CA 50, NCAM and Fuc-GM 1, in LC. It is of interest that MAb 11-16, which was used for the detection of CEA, was produced after immunization with CEA purified from liver metastases of ADLC [19, 23], whereas almost all other anti-CEA MAbs have been devel oped against colon cancer CEA. Hitherto pub lished results on CEA expression in LC tissues have generally been based on the use of poly clonal antibodies [24, 25], With these, CEA expression has been found in 50-90% of nonSCLC and in less than 30% of SCLC [24, 25], In one study using MAbs, expression was found in more than 90% of non-SCLC and in all 3 SCLC cases tested [26]. By using MAb 11-16, we found CEA expression in more than 80% of SCLC and non-SCLC, suggesting that this antibody is highly sensitive for CEA in lung tumors. Elevated levels of CA 50 in sera from LC patients have been demonstrated by several groups [14, 27, 28]. In the present work, how ever, we report for the first time on the immunohistological expression of CA 50 in LC tis sue. We found that CA 50 was expressed in more than 65% of SCLC and non-SCLC tis sue specimens showing a tendency towards preferential expression in SCLC and ADLC. These results justify the use, not only of CEA, but also of CA 50 as a serological tumor marker in LC. The ganglioside Fuc-GM 1 and NCAM, as defined by specific MAbs, have previously been described as being strongly associated with SCLC [4, 8, 9, 29], Our results are in line with these studies. In the present study, we found considerable expression of these ‘typi cal’ SCLC antigens also in 25% of non-SCLC cases, suggesting that these tumors represent transitions between non-SCLC and SCLC.
By using the panel of all four antigens (FucGM 1, NCAM, CEA and CA 50), two patterns of antigen expression could be recognized showing a high degree of specificity for SCLC (N/F+ CA 50-) and non-SCLC (N/F- CEA+), respectively. A considerable number of SCLC and non-SCLC lung tumors, however, shared similar antigen patterns (fig. 2). The signifi cance of this observation, e.g. with regard to its correlation to proliferative and metastatic potential and sensitivity to chemo- and radia tion therapy remains to be investigated. MAb 11-16 (defining CEA) has been de scribed to be unreactive with any parenchy mal or mesenchymal cellular structures in normal lung and bronchial tissue [20], With MAb C50 (defining CA 50, i.e. sialosyl Lea and sialosyl lactotetraose) reactivity has occa sionally been found with normal bronchial epithelial and glandular cells (unpublished observations), which is consistent with the expression of sialosyl Lea as defined by MAb 19-9 [30-31], Due to the broad reactivity of these MAbs with both SCLC and non-SCLC and their restricted normal lung tissue reac tivity, MAb 11-16 in particular but also MAb C50 may become useful for the immunohisto chemical detection of LC in diagnostic tissue specimens. Double-fluorochrome IF staining of SCLC tissue specimens showed that Fuc-GM 1 was generally coexpressed with NCAM or CA 50, whereas cellular coexpression with CEA ap pears to be rare. These observations, although deduced from a limited number of specimens, support the contention that Fuc-GM 1 is ex pressed on cells with SCLC features. The find ings are also consistent with the view that CEA is mainly expressed on more highly dif ferentiated tumor cells. The intra- and inter tumor heterogeneity of antigen expression ob served might be correlated with the biological properties of different tumor cell populations. This might be especially relevant with regard
313
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
D iscu ssio n
to Fuc-GMl, since expression of this antigen exhibited the greatest degree of intertumoral heterogeneity in the SCLC specimens tested. Heterogeneity in antigen expression and its possible relation to biological and clinical variables warrants further investigation.
Acknowledgements The expert technical assistance of Ms. K. Schohn and Mrs. K. Ahlman is greatfully acknowledged.
1 Köhler G. Milstein C: Continuous cultures of fused cells secreting anti body of predicted specificity. Na ture (Lond) 1975;256:495-497. 2 Nilsson O, Mansson J-E. Brezicka T, Holmgren J, Lindholm L. Soren son S, Yngvason F, Svennerholm L: Fucosyl-GM 1: A ganglioside associ ated with small cell lung carcinoma. Glycoconj J 1984; 1:43-49. 3 Nilsson O. Brezicka FT, Holmgren J. Sorenson S. Svennerholm L, Yng vason F, Lindholm L: Detection of a ganglioside antigen associated with small cell lung carcinomas using monoclonal antibodies directed against fucosyl-GMl. Cancer Res 1986;46:1403-1407. 4 Brezicka F-T, Oiling S, Nilsson O, Bergh J, Holmgren J. Sorenson S, Yngvason F. Lindholm L: Immunohistological detection of fucosylGM l ganglioside in human lung cancer and normal tissues with monoclonal antibodies. Cancer Res 1989;49:1300-1305. 5 Fredman P. Brezicka T, Holmgren J, Lindholm L, Nilsson O, Svenner holm L: Binding specificity of monoclonal antibodies to ganglio side Fuc-GMl. Biochim Biophys Acta 1986;875:316-323. 6 Beverley P, Souhami R, Bobrow L: Proc I st Int Workshop on Small Cell Lung Cancer Antigens. Lung Cancer 1988:4:1-114. 7 Beverley PCL, Bobrow LG, Gilks W, Souhami RL: Proc 2nd Int Workshop on Small Cell Lung Can cer Antigens. Br J Cancer 1991; 63(suppl 14):3—88.
314
8 Patel K. Moore SE, Dickson G, Rossell RJ, Beverly PC, Kemshead JT, Walsh FS: Neural cell adhesion mol ecule (NCAM) is the antigen recog nized by monoclonal antibodies of similar specificity in small-cell lung carcinoma and neuroblastoma. Int J Cancer 1989;44:573-578. 9 Mooi WJ, Wagenaar SS, Schol D, Hilgers J: Monoclonal antibody I23C3 in lung tumor classification immunohistology of 358 resected lung tumours. Mol Cell Probes 1988:2:31-37. 10 Moolenaar CECK. Muller EJ, Schol DJ, Figdor CG, Bock E, BitterSuermann D, Michalides RJAM: Expression of neural cell adhesion molecule-related sialoglycoprotein in small cell lung cancer and neuro blastoma cell lines H69 and CHP212. Cancer Res 1990;50:11021106. 11 Schol DJ, Mooi WJ, van dcrGugten AA, Wagenaar SS, Hilgers J: Mono clonal antibody 123C3, identifying small cell carcinoma phenotype in lung tumours, recognizes mainly, but not exclusively, endocrine and neuron-supporting normal tissues. Int J Cancer 1988;(suppl 2):34—40. 12 Gold P, Freedman SO: Specific carcinoembryonic antigens of the hu man digestive system. J Exp Med 1965;122:467-481. 13 Lindholm L. Holmgren J, Svenner holm L, Fredman P, Nilsson O, Persson B. Myrvold H. Lagergiird T : Monoclonal antibodies against gas trointestinal tumor-associated anti gens isolated as monosialogangliosides. Int Arch Allergy Appl Immu nol 1983;71:178-181.
14 Holmgren J. Lindholm L, Persson B, I.agcrgard T, Nilsson O, Svenner holm L, Rudenstam C-M. Unsgaard B, Yngvason F, Pettersson S, Killandcr AF: Detection by monoclonal antibody of carbohydrate antigen CA 50 and serum of patients with carcinoma. Br Med J 1984:288: 1479-1482. 15 Mansson J-E, Fredman P, Nilsson O. Lindholm L, Holmgren J. Sven nerholm L: Chemical structure of carcinoma ganglioside antigens de fined by monoclonal antibody C-50, and some allied gangliosides of hu man pancreatic adenocarcinoma. Biochim Biophys Acta 1985:834: 110-117. 16 Nilsson O, Lindholm L, Persson B. Fredman P, Mansson J-E, Holmgren J, Svennerholm L: Tissue distribution and concentration of a monoclonal antibody defined tu mor-associated ganglioside antigen. Proc the 7th Int Symp on Glycoconjugates, 1983, pp 852-853. 17 Nilsson O, Mansson J-E. Lindholm L, Holgren J, Svennerholm L: Sialosyllactotetraosylceramidc, a novel ganglioside detected in human carci nomas by a monoclonal antibodv. FEBS Lett 1985;182:398-402. 18 Hedin A, Zoubir F, Lundgren T, Hammarstrom S: Epitope specific ity and cross-reactivity pattern of a large series of monoclonal anti bodies to carcinoembryonic antigen. Mol Immunol 1986;23:1053-1061.
Brezicka/Olling/Bergman/Berggren/ Engström/Hammarström/Holmgren/ Larsson/Lindholm
Tumor Antigen Coexpression in Lung Cancer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
References
24 Kawai T, Torikata C, Suzuki M: 1mmunohistochemical study of pulmo nary adenocarcinoma. Am J Clin Pathol 1988:89:455-462. 25 Sun NCJ, Edgington TD, Carpen tier CL, McAfee W. Terry R, Bate man J: Immunohistochemical local ization of carcinoembryonic antigen (CEA). CEA-S. and nonspecific cross-reacting antigen (NCA) in car cinomas of lung Cancer 1983;52: 1632-1641. 26 Gatter K. Dunnill M, Pulford KAF, Hcryet A, Mason DY: Human lung tumors: A correlation of antigenic profile with histological type. Histopathology 1985;9:805-823. 27 Berthiot G, Maréchal F, Cattan A, Deltour G: Serum levels o f CA-50, CA-19.9, CA-125, neuron specific enolase and carcinoembryonic anti gen in lung cancer and benign dis eases of the lung. Biomed Pharmacother 1989:43:613-620. 28 Macchia V, Mariano A, Calvacanti M, Coppa A. Cecère C. Fraioli G. Elia S, Ferrante G: Tumor markers and lung cancer: Correlation be tween scrum and bronchial secre tion levels of CEA, TPA, CanAg CA-50, NSE and ferritin. Int J Biol Markers 1987;2:151-156.
29 Bcrendsen HH, de Leij L, Poppema S, Postmus PE, Sluiter HJ, The H: Simultaneous standard light micros copy and immunohistology on bronchoscopically procured lung cancer specimens. Eur J Cancer Clin Oncol 1988;24:915-921. 30 Atkinson BF. Ernst CS. Herlyn M, Stcplcwski Z, Sears HF. Koprowski J: Gastrointestinal cancer-associ ated antigen in immunoperoxidase assay. Cancer Res 1982;42:48204823. 31 Arends JW, Verstynen C, Bosnian FT, Hilgers J, Steplewski Z: Distri bution of monoclonal antibody-de fined monosialoganglioside in nor mal and cancerous human tissues: An immunoperoxidase study. Hy bridoma 1983;2:219-229.
315
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
19 Hammerström S, Shively JE. Pax ton RJ, Beatty BG, Larsson Ä, Ghosh R. Bormer O. Buchegger F, Mach J-P, Burtin P, Seguin P, Darbourct B. Degorce F, Sertour J. Jolu JP, Fuks A, Kalthoff H. Schmiegel W, Arndt R. Klöppel G. von Kleist S. Grunert F, Schwartz K. Matsuoka Y, Kuroki M, Wagcner C, Weber T, Yachi A, lmai K. Hishikawa N. Tsujisaki M: Antigenic sites in carcinoembryonic antigen. Cancer Res 1989;49:4852-4858. 20 Zoubir F, Zeromski J. Sikora J, Szmeja J. Hcdin A, Hammarström S: Tumor specificity of monoclonal antibodies to carcinoembryonic an tigen. Tumor Biol 1990:11:5-19. 2 1 Warnke R, Levy R: Detection o f T and B cell antigens with hybridoma monoclonal antibodies. J HistochemCytochem 1980;28:771-776. 22 WHO: Histological Typing of Lung Tumours. International Histological Classification of Tumours. Geneva, WHO, 1981. vol 1. 23 Hedin A, Hammerström S, Larsson A: Specificity and binding proper ties of eight monoclonal antibodies against carcinoembryonic antigen. Mol Immunol 1982:19:1641-1648.
Tumor Biol 1992;13:308-315
a Department of Medical Microbiology and Immunology, b Department of Pathology, c Department of Internal Medicine, 1st Section of Pulmonary Medicine, University of Göteborg, d Department of Thoracic Surgery, Sahlgren’s Hospital, Göteborg, and c Department of Immunology, University of Umeä, Sweden
Coexpression of Ganglioside Antigen Fuc-GM I, Neural-Cell Adhesion Molecule, Carcinoembryonic Antigen, and Carbohydrate Tumor-Associated Antigen CA 50 in Lung Cancer
Key Words Tumor-associated antigens Monoclonal antibodies Immunohistochemical study Pulmonary carcinoma NCAM Ganglioside Fuc-GM 1 CEA CA 50
Abstract With the aid of specific monoclonal antibodies, tumor tissues from 68 patients with lung cancer were examined for their expression of two small cell lung carcinoma (SCLC) antigens, Fuc-GMI (fucosyl GM1; IV2FuclPNeuAc GgOse4) and neu ral-cell adhesion molecule (NCAM), and two broader tumor antigens, carcinoembryonic antigen (CEA) and carbohydrate cancer-associated antigen CA 50. Expression of Fuc-GMI was seen in 75% and NCAM in 78% of the SCLC specimens, but also in 12 and 20% of non-SCLC. Either or both of these anti gens were expressed in more than 90% of SCLC and in 25% of non-SCLC. CEA was found in more than 80% of SCLC and non-SCLC. Expression of CA 50 was seen in 65-68% of nonSCLC and SCLC, showing preference for SCLC and lung ade nocarcinoma. In SCLC, cellular expression of Fuc-GM 1 was generally seen together with NCAM and CA 50, but rarely with CEA. There was considerable inter- and intratumor het erogeneity in the expression of all four antigens. The results suggest that CEA is the antigen of choice for the detection of lung cancer regardless of histotype. In combined analysis of CEA, CA 50, Fuc-GMI and NCAM. two patterns of antigen expression were recognized that appear to discriminate be tween SCLC and non-SCLC tumors, respectively. A consider able fraction of SCLC and non-SCLC tumors, however, exhib ited similar patterns of antigen expression. The biological and clinical significance of these observations remains to be inves tigated.
This study was supported by grants from the Medical Faculty. University of Göteborg, Sweden, the Swedish Cancer Society, and Pharmacia Ltd.
Received: September 17 . 1991 Accepted: May 26, 1992
F.-T. Brc/icka Department of MedicalMicrobiology and Immunology, University of Göteborg Guldhedsgatan 10
© 1992 S. Karger AG. Basel 1010-4283/92/ 0136-0308S2.75/0
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
Fred-Thomas Brezicka a - c Saute Ollingb Bengt Bergmanc Häkan Berggrend Carl-Peter Engström c Sten Hammarström c Jan Holmgrena Sture Larssond Leif Lindholm a
Despite important progress in the therapy of cancer during the last three decades, lung cancer (LC) still has a poor prognosis; e.g., with a 5-year survival rate of 5-10% for small cell LC (SCLC). Since SCLC is highly sensi tive to chemo- and radiation therapy, it is important to accurately identify and discrimi nate this tumor type from non-SCLC. The introduction of the hybridoma technology [1] has allowed for the production of a large num ber of monoclonal antibodies (MAbs) with high and uniform specificity that identify tu mor-associated antigens (TAAgs) expressed preferentially by SCLC tumors. We have pre viously described the production of MAbs FI 2 and FI 5 (both IgG3)that specifically bind to the ganglioside Fuc-GMl. By using these MAbs Fuc-GMl was found to be highly asso ciated with SCLC [2-5]. At the 1st and 2nd Workshops on SCLC Antigens [6, 7], MAbs raised against SCLC were characterized and clustered. MAbs of cluster 1 were all reactive with the neural-cell adhesion molecule (NCAM) [8] and appeared to have the great est potential in discriminating SCLC from non-SCLC. One member of this cluster is represented by MAb 123C3 (IgGi) [8-11], Here, we have examined the expression of four TAAgs in human lung tumor tissues representing the four major histotypes. Be sides the expression of Fuc-GM 1 and NCAM, that of two more generally expressed epithe lial TAAgs, carcinoembryonic antigen (CEA) [ 12] and carbohydrate-associated antigen CA 50 [13-17], was also studied. MAbs F I2 and FI 5, and 123C3 were used to specifically identify Fuc-GMl and NCAM. MAb 11-16 (IgGi) [18-20], produced after immunization with lung tumor-derived CEA, was used for the specific detection of CEA. The reactivity of this MAb panel with lung tumors and its ability to discriminate between SCLC and
non-SCLC was studied. We have also studied the cellular coexpression of Fuc-GMl and NCAM, CEA or CA 50 in SCLC tumors. The results indicate possible implications for fu ture diagnosis, prognosis and therapy of lung cancer.
Materials and Methods MAbs and Reagents The production and characterization of the FucGMl ganglioside-reactive MAbs FI 2 (IgG;,) and FI 5 (IgG^) were described earlier [3-5], Hybridoma culture medium of MAb FI 2 was precipitated with ammo nium sulfate, redissolved in PBS and dialyzed against PBS and then used for immunohistochemical studies. Purified MAb FI 5 was biotinylated [21] or used in native form. Mouse ascites preparations of NCAMrcactive MAb 123C3 (IgGi) [9,11 ] were a generous gift from Dr. K. Moolenaar (The Netherlands Cancer In stitute). Anti-CEA MAb II-16 (IgGi) was raised against purified LC CEA [18-20] and used in purified form. Purified MAb C50 (IgM) [13] defining CA 50 antigens (sialosyl Lea and sialosyllactotctraose [15, 17]), was obtained from Pharmacia CanAg (Göteborg, Sweden). Biotinylated horse anti-mouse antiserum (HAMB) and avidin-TMRITC (AvTMRITC) were purchased from Vector Laboratories Inc. (Burlingame, Calif.), Goat (FAb) anti-mouse IgG-FITC (GAMF) from Cappel Laboratories (Malvern, Pa.), rabbit anti-goat Ig-FITC (RAGF), swine anti-rabbit Ig-FITC (SwARF), rabbit anti-mouse Ig-FITC (RAMF) and ABComplex (K 355) from Dakopatts (Copenhagen, Denmark). The anti bodies and conjugates were diluted in PBS with 5% normal human serum and 0.1 % BSA after titration on control tumor tissues or according to the instructions supplied by the manufacturer. The following concen trations and dilutions were used: FI 2 10-100 pg/ml; F I5 200 pg/ml: biotinylated F I5 (F15B) 120 pg/ml; 123C3 diluted 1/50:11-16 10pg/ml: MAbC50 50-100 pg/ml; GAMF 1/50; RAGF 1/10; SwARF 1/10-1/20: RAMF 1/10. and AvTMRITC 1/50. Stainings using an IgG3MAb(protein-A purified) or irrelevant specificity or dilution buffer only replacing the specific MAbs were done as negative controls at each staining ses sion.
309
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
Introduction
Staining Procedures Tissue sections were fixed in 4% paraformalde hyde in PBS for 30 min at room temperature and rinsed three times with PBS for 5 min. For staining with MAbs 11-16, C50, 123C3, and F12, the ABC method was used. The following steps were performed in sequence: (1) incubation with 5% normal horse serum and 1% BSA in PBS for 15 min; (2) MAb; (3) HAMB. and (4) ABComplex. Each step was per formed for 30 min at room temperature in a humid chamber and followed by rinsing two to three times with PBS for 5 min. Sections were then incubated with 10 mg 3-amino-9-cthylcarbazol and 4 pi 30% HiCL in 6 ml DMSO and 50 ml 0.02 M NaAc (pH 5.5) for 15 min, rinsed twice with PBS for 5 min and once with distilled water, counterstained with Mayers hematoxy lin, and finally mounted with glycerin gelatin. For staining with MAbs FI 2 and FI 5, a double-step immu nofluorescence (IF) method [4] was also use: (1) MAb; (2) RAMF. and (3) SwARF. The ABC staining method was used to assess expression of NCAM, CEA and CA 50, whereas the IF method was used to assess expres sion of Fuc-GM 1. For the simultaneous detection of two separate antigens on individual tumor cells, bioti nylated MAb F I5 (F l5B) was used in conjuction with a double-fluorochrome IF staining method consisting of the following steps: ( 1) MAb (unconjugated); (2) GAMF; (3) RAGF+F15B, and (4) SwARF + AvTMRITC. In the IF methods, each step was fol
310
lowed by rinsing three times with PBS for 5 min, and the slides were finally mounted with 87% glycerin in PBS.
Results Expression of CEA, CA 50, Fuc-GM 1 and NCAM in lung tumors was studied by using highly specific MAbs and sensitive immunohistological staining methods. Expression o f CEA and CA 50 in Lung Tumors The results are summarized in table 1. CEA expression was seen in 82% of nonSCLC and in 85% of SCLC tumors. No pref erence for SQLC, ADLC or LCLC, collec tively denoted non-SCLC, or for SCLC was seen. In SQLC and ADLC, the expression of CEA was generally seen in highly differen tiated areas with various degrees of keratinization or adenoid formations. In these, cen trally located cells or cells lining the glands were most strongly stained. High expression was also occasionally found in less well-differ entiated areas containing conglomerates of positively stained cells, especially in low dif ferentiated non-SCLC. This pattern was also frequently seen in SCLC tumors (fig. la). The intensity and extension of MAb 11-16 staining of LC tumors showed a high degree of intraand intertumor heterogeneity. Also CA 50 was frequently expressed in SCLC (68%; fig. lb) and non-SCLC (65%). A tendency towards preferential expression of CA 50 in SCLC and ADLC tumors was seen. Of these, more than 60% of the cases showed wide (more than 25% of the tumor cells) expression, whereas this degree of expression was seen in only 30% of the SQLC cases. In non-SCLC, the morphological components expressing CA 50 were similar to those ex pressing CEA described above. In less well-
Brezicka/Olling/Bergman/Berggren/ Engstrôm/Hammarstrôm/Holmgren/ Larsson/Lindholm
Tumor Antigen Coexpression in Lung Cancer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
Tumor Tissue Specimens Tumor tissues from 68 patients with LC were col lected at thoracotomy or at autopsy. Twenty-seven of these were resected at thoracotomy and included 14 cases with squamous cell LC (SQLC), 3 with SCLC, 8 with adenocarcinoma (ADLC) and 2 with large cell LC (LCLC). The remaining 41 cases were taken at autopsy and included 12 cases with SQLC, 25 with SCLC, 3 with ADLC, and 1 with LCLC. Tissues were embed ded in mounting medium, snap-frozen in liquid nitro gen and stored at -7 0 °C until used. Cryostat sections (2-4 pm) were cut and put on glass slides. Slides were air-dried and stored at -7 0 °C until used for immuno histochemical staining. Representative sections from each tissue sample were stained with van Gieson's stains or Mayers hematoxylin for histopathological examination and assessment of diagnosis. These were also compared with routinely histopathologically pro cessed biopsies taken at bronchoscopy, mediastinos copy or thoracotomy. The histotype of each lung tumor specimen was assessed by light microscopic cri teria [22],
Table 1. Expression of CEA, CA 50, NCAM and Fuc-GM l in lung tumors Antigen
Score1
Lung tumor histotype SQLC
CEA
CA 50
NCAM
Fuc-GM 1
n
0 1 2 0 1 2 0 1 2 0 1 2
SCLC
ADLC
LCLC
number of cases
%
number of cases
%
number of cases
%
number of cases
3 4 19 9 9 8 21 0 5 25 1 0
12 15 73 35 35 30 81
4 4 20 9 2 17 6 4 18 7 7 14
14 14 71 32 7 61 21 14 64 25 25 50
3 0 8 3 1 7 8 1 2 7 3 1
27
1 1 1 2 1 0 3 0 0 3 0 0
19 96 4
26
28
11
73 27 10 63 73 10 18 64 27 9
3
differentiated areas of non-SCLC, and in SCLC tumors, cell aggregates of various sizes were homogenously stained with MAb C50. Expression o f Fuc-GM 1 and NCAM in Lung Tumors Expression of Fuc-GM 1 and NCAM was highly restricted to the SCLC histotype. Sev enty-five and 78% of the SCLC cases ex pressed these antigens, respectively. The ex pression of both antigens was homogeneous and generally comprised large areas of the tumors (fig. lc, d). Significant expression of Fuc-GM 1 and NCAM was also seen in nonSCLC (10 and 20%, respectively). In nonSCLC, expression of Fuc-GM 1 or NCAM was generally seen in less well-differentiated ar eas.
Combined Expression ofTAAgs in Lung Tumors Expression of both Fuc-GM 1 and NCAM in individual SCLC tumors was observed in 61 % of the cases (table 2), whereas 93% ex pressed at least one of these antigens. At least one of these antigens was also expressed in 25% of the non-SCLC cases, whereas expres sion of both was only seen in 8%. Expression of both CFA and CA 50 was seen in 61 % of the SCLC and in 55% of the non-SCLC tumors (fig. 2). A minority (728%) of SCLC and non-SCLC tumors showed expression of either CEA or CA 50. Expres sion of at least one of these antigens was seen in 91 % of SCLC and in 92% of non-SCLC. The combined expression of Fuc-GM 1, NCAM, CEA and CA 50 observed in lung tumors is summarized in figure 2. Two pat-
311
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
1 Score: 0 = negative reaction; 1 = < 25 % of tumor cells showing positive staining reaction; 2 = > 25 % of tumor cells showing positive staining reaction.
SCLC
non-SCLC
CA 50
Fuc-GM 1 and/or NCAM
Fuc-GMl and/or NCAM
Fig. 2. Combined expression of CEA, Ca 50 and Fuc-GMl and/or NCAM in SCLC and non-SCLC.
Histotype
F+N-
F+N+
F-N+
F-N-
n %
n
n
n
%
%
%
SCLC Non-SCLC
28 40
4 14 2 5
17 61 3 8
5 18 5 12
2 7 30 75
Total
68
6
20 29
10 15
32 47
terns of antigen expression appeared that seem to characterize SCLC and non-SCLC tumors, respectively. Expression of at least one of NCAM or Fuc-GM 1(N/F+) but not CA 50 (N/F+ CA 50") was restricted to SCLC tumors (32%), whereas N/F~ CEA+ was re stricted to non-SCLC (60%). There were large fractions of SCLC (68%) and non-SCLC (32%) tumors that showed overlapping pat terns of antigen expression. Cellular Coexpression ofTAAgs in SCLC Coexpression of two antigens (Fuc-GMl vs. NCAM, CEA and CA 50) was studied on the cellular level in 8 SCLC cases by using a
312
n
9
double-fluorochrome IF method. Fuc-GMl was frequently found to be coexpressed with NCAM (all cases) or CA-50 (7 cases). Expres sion of NCAM was either congruent with that of Fuc-GM 1 (3 cases) or more widely distrib uted including also Fuc-GM 1-expressing cells (5 cases). These patterns were also seen for coexpression of Fuc-GM 1 and CA 50 (1 and 6 cases, respectively; fig le). Fuc-GMl was oc casionally found to be coexpressed with CEA only on small fractions of the tumor cells (4 cases; fig. 10- Generally, CEA and Fuc-GMl expression was complementary (7 cases), i.e., the tumors contained areas expressing either of these antigens.
Brezicka/Olling/Bergman/Berggrcn/ Engstrôm/Hammarstrôm/Holmgren/ Earsson/Lindholm
Tumor Antigen Coexpression in Lung Cancer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
Table 2. Combined expression of Fuc-GMl and NCAM in SCLC and non-SCLC lung tumors. F+ and N+denote expression of FucGMl and NCAM, respectively, and F" and N_ lack of detectable antigen.
In this study we have used well-character ized MAbs for the detection of four TAAgs, CEA, CA 50, NCAM and Fuc-GM 1, in LC. It is of interest that MAb 11-16, which was used for the detection of CEA, was produced after immunization with CEA purified from liver metastases of ADLC [19, 23], whereas almost all other anti-CEA MAbs have been devel oped against colon cancer CEA. Hitherto pub lished results on CEA expression in LC tissues have generally been based on the use of poly clonal antibodies [24, 25], With these, CEA expression has been found in 50-90% of nonSCLC and in less than 30% of SCLC [24, 25], In one study using MAbs, expression was found in more than 90% of non-SCLC and in all 3 SCLC cases tested [26]. By using MAb 11-16, we found CEA expression in more than 80% of SCLC and non-SCLC, suggesting that this antibody is highly sensitive for CEA in lung tumors. Elevated levels of CA 50 in sera from LC patients have been demonstrated by several groups [14, 27, 28]. In the present work, how ever, we report for the first time on the immunohistological expression of CA 50 in LC tis sue. We found that CA 50 was expressed in more than 65% of SCLC and non-SCLC tis sue specimens showing a tendency towards preferential expression in SCLC and ADLC. These results justify the use, not only of CEA, but also of CA 50 as a serological tumor marker in LC. The ganglioside Fuc-GM 1 and NCAM, as defined by specific MAbs, have previously been described as being strongly associated with SCLC [4, 8, 9, 29], Our results are in line with these studies. In the present study, we found considerable expression of these ‘typi cal’ SCLC antigens also in 25% of non-SCLC cases, suggesting that these tumors represent transitions between non-SCLC and SCLC.
By using the panel of all four antigens (FucGM 1, NCAM, CEA and CA 50), two patterns of antigen expression could be recognized showing a high degree of specificity for SCLC (N/F+ CA 50-) and non-SCLC (N/F- CEA+), respectively. A considerable number of SCLC and non-SCLC lung tumors, however, shared similar antigen patterns (fig. 2). The signifi cance of this observation, e.g. with regard to its correlation to proliferative and metastatic potential and sensitivity to chemo- and radia tion therapy remains to be investigated. MAb 11-16 (defining CEA) has been de scribed to be unreactive with any parenchy mal or mesenchymal cellular structures in normal lung and bronchial tissue [20], With MAb C50 (defining CA 50, i.e. sialosyl Lea and sialosyl lactotetraose) reactivity has occa sionally been found with normal bronchial epithelial and glandular cells (unpublished observations), which is consistent with the expression of sialosyl Lea as defined by MAb 19-9 [30-31], Due to the broad reactivity of these MAbs with both SCLC and non-SCLC and their restricted normal lung tissue reac tivity, MAb 11-16 in particular but also MAb C50 may become useful for the immunohisto chemical detection of LC in diagnostic tissue specimens. Double-fluorochrome IF staining of SCLC tissue specimens showed that Fuc-GM 1 was generally coexpressed with NCAM or CA 50, whereas cellular coexpression with CEA ap pears to be rare. These observations, although deduced from a limited number of specimens, support the contention that Fuc-GM 1 is ex pressed on cells with SCLC features. The find ings are also consistent with the view that CEA is mainly expressed on more highly dif ferentiated tumor cells. The intra- and inter tumor heterogeneity of antigen expression ob served might be correlated with the biological properties of different tumor cell populations. This might be especially relevant with regard
313
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
D iscu ssio n
to Fuc-GMl, since expression of this antigen exhibited the greatest degree of intertumoral heterogeneity in the SCLC specimens tested. Heterogeneity in antigen expression and its possible relation to biological and clinical variables warrants further investigation.
Acknowledgements The expert technical assistance of Ms. K. Schohn and Mrs. K. Ahlman is greatfully acknowledged.
1 Köhler G. Milstein C: Continuous cultures of fused cells secreting anti body of predicted specificity. Na ture (Lond) 1975;256:495-497. 2 Nilsson O, Mansson J-E. Brezicka T, Holmgren J, Lindholm L. Soren son S, Yngvason F, Svennerholm L: Fucosyl-GM 1: A ganglioside associ ated with small cell lung carcinoma. Glycoconj J 1984; 1:43-49. 3 Nilsson O. Brezicka FT, Holmgren J. Sorenson S. Svennerholm L, Yng vason F, Lindholm L: Detection of a ganglioside antigen associated with small cell lung carcinomas using monoclonal antibodies directed against fucosyl-GMl. Cancer Res 1986;46:1403-1407. 4 Brezicka F-T, Oiling S, Nilsson O, Bergh J, Holmgren J. Sorenson S, Yngvason F. Lindholm L: Immunohistological detection of fucosylGM l ganglioside in human lung cancer and normal tissues with monoclonal antibodies. Cancer Res 1989;49:1300-1305. 5 Fredman P. Brezicka T, Holmgren J, Lindholm L, Nilsson O, Svenner holm L: Binding specificity of monoclonal antibodies to ganglio side Fuc-GMl. Biochim Biophys Acta 1986;875:316-323. 6 Beverley P, Souhami R, Bobrow L: Proc I st Int Workshop on Small Cell Lung Cancer Antigens. Lung Cancer 1988:4:1-114. 7 Beverley PCL, Bobrow LG, Gilks W, Souhami RL: Proc 2nd Int Workshop on Small Cell Lung Can cer Antigens. Br J Cancer 1991; 63(suppl 14):3—88.
314
8 Patel K. Moore SE, Dickson G, Rossell RJ, Beverly PC, Kemshead JT, Walsh FS: Neural cell adhesion mol ecule (NCAM) is the antigen recog nized by monoclonal antibodies of similar specificity in small-cell lung carcinoma and neuroblastoma. Int J Cancer 1989;44:573-578. 9 Mooi WJ, Wagenaar SS, Schol D, Hilgers J: Monoclonal antibody I23C3 in lung tumor classification immunohistology of 358 resected lung tumours. Mol Cell Probes 1988:2:31-37. 10 Moolenaar CECK. Muller EJ, Schol DJ, Figdor CG, Bock E, BitterSuermann D, Michalides RJAM: Expression of neural cell adhesion molecule-related sialoglycoprotein in small cell lung cancer and neuro blastoma cell lines H69 and CHP212. Cancer Res 1990;50:11021106. 11 Schol DJ, Mooi WJ, van dcrGugten AA, Wagenaar SS, Hilgers J: Mono clonal antibody 123C3, identifying small cell carcinoma phenotype in lung tumours, recognizes mainly, but not exclusively, endocrine and neuron-supporting normal tissues. Int J Cancer 1988;(suppl 2):34—40. 12 Gold P, Freedman SO: Specific carcinoembryonic antigens of the hu man digestive system. J Exp Med 1965;122:467-481. 13 Lindholm L. Holmgren J, Svenner holm L, Fredman P, Nilsson O, Persson B. Myrvold H. Lagergiird T : Monoclonal antibodies against gas trointestinal tumor-associated anti gens isolated as monosialogangliosides. Int Arch Allergy Appl Immu nol 1983;71:178-181.
14 Holmgren J. Lindholm L, Persson B, I.agcrgard T, Nilsson O, Svenner holm L, Rudenstam C-M. Unsgaard B, Yngvason F, Pettersson S, Killandcr AF: Detection by monoclonal antibody of carbohydrate antigen CA 50 and serum of patients with carcinoma. Br Med J 1984:288: 1479-1482. 15 Mansson J-E, Fredman P, Nilsson O. Lindholm L, Holmgren J. Sven nerholm L: Chemical structure of carcinoma ganglioside antigens de fined by monoclonal antibody C-50, and some allied gangliosides of hu man pancreatic adenocarcinoma. Biochim Biophys Acta 1985:834: 110-117. 16 Nilsson O, Lindholm L, Persson B. Fredman P, Mansson J-E, Holmgren J, Svennerholm L: Tissue distribution and concentration of a monoclonal antibody defined tu mor-associated ganglioside antigen. Proc the 7th Int Symp on Glycoconjugates, 1983, pp 852-853. 17 Nilsson O, Mansson J-E. Lindholm L, Holgren J, Svennerholm L: Sialosyllactotetraosylceramidc, a novel ganglioside detected in human carci nomas by a monoclonal antibodv. FEBS Lett 1985;182:398-402. 18 Hedin A, Zoubir F, Lundgren T, Hammarstrom S: Epitope specific ity and cross-reactivity pattern of a large series of monoclonal anti bodies to carcinoembryonic antigen. Mol Immunol 1986;23:1053-1061.
Brezicka/Olling/Bergman/Berggren/ Engström/Hammarström/Holmgren/ Larsson/Lindholm
Tumor Antigen Coexpression in Lung Cancer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
References
24 Kawai T, Torikata C, Suzuki M: 1mmunohistochemical study of pulmo nary adenocarcinoma. Am J Clin Pathol 1988:89:455-462. 25 Sun NCJ, Edgington TD, Carpen tier CL, McAfee W. Terry R, Bate man J: Immunohistochemical local ization of carcinoembryonic antigen (CEA). CEA-S. and nonspecific cross-reacting antigen (NCA) in car cinomas of lung Cancer 1983;52: 1632-1641. 26 Gatter K. Dunnill M, Pulford KAF, Hcryet A, Mason DY: Human lung tumors: A correlation of antigenic profile with histological type. Histopathology 1985;9:805-823. 27 Berthiot G, Maréchal F, Cattan A, Deltour G: Serum levels o f CA-50, CA-19.9, CA-125, neuron specific enolase and carcinoembryonic anti gen in lung cancer and benign dis eases of the lung. Biomed Pharmacother 1989:43:613-620. 28 Macchia V, Mariano A, Calvacanti M, Coppa A. Cecère C. Fraioli G. Elia S, Ferrante G: Tumor markers and lung cancer: Correlation be tween scrum and bronchial secre tion levels of CEA, TPA, CanAg CA-50, NSE and ferritin. Int J Biol Markers 1987;2:151-156.
29 Bcrendsen HH, de Leij L, Poppema S, Postmus PE, Sluiter HJ, The H: Simultaneous standard light micros copy and immunohistology on bronchoscopically procured lung cancer specimens. Eur J Cancer Clin Oncol 1988;24:915-921. 30 Atkinson BF. Ernst CS. Herlyn M, Stcplcwski Z, Sears HF. Koprowski J: Gastrointestinal cancer-associ ated antigen in immunoperoxidase assay. Cancer Res 1982;42:48204823. 31 Arends JW, Verstynen C, Bosnian FT, Hilgers J, Steplewski Z: Distri bution of monoclonal antibody-de fined monosialoganglioside in nor mal and cancerous human tissues: An immunoperoxidase study. Hy bridoma 1983;2:219-229.
315
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:23:26 PM
19 Hammerström S, Shively JE. Pax ton RJ, Beatty BG, Larsson Ä, Ghosh R. Bormer O. Buchegger F, Mach J-P, Burtin P, Seguin P, Darbourct B. Degorce F, Sertour J. Jolu JP, Fuks A, Kalthoff H. Schmiegel W, Arndt R. Klöppel G. von Kleist S. Grunert F, Schwartz K. Matsuoka Y, Kuroki M, Wagcner C, Weber T, Yachi A, lmai K. Hishikawa N. Tsujisaki M: Antigenic sites in carcinoembryonic antigen. Cancer Res 1989;49:4852-4858. 20 Zoubir F, Zeromski J. Sikora J, Szmeja J. Hcdin A, Hammarström S: Tumor specificity of monoclonal antibodies to carcinoembryonic an tigen. Tumor Biol 1990:11:5-19. 2 1 Warnke R, Levy R: Detection o f T and B cell antigens with hybridoma monoclonal antibodies. J HistochemCytochem 1980;28:771-776. 22 WHO: Histological Typing of Lung Tumours. International Histological Classification of Tumours. Geneva, WHO, 1981. vol 1. 23 Hedin A, Hammerström S, Larsson A: Specificity and binding proper ties of eight monoclonal antibodies against carcinoembryonic antigen. Mol Immunol 1982:19:1641-1648.
