Original Article

Immunohistochemical Analysis of Human Adenocarcinoma-Associated YH20 6 Detected by a Monoclonal Akira Yachi, Kohzoh Imai, Takao Endo and Yuji Hinoda

A monoclonal antibody YH206(IgM, k) was prepared using human lung adenocarcinoma A549 cell line as an immunogen. The histological distribution of antigenic determinant of the YH206antigen was highly restricted to adenocarcinomas of various organs including the lung, the pancreas and the stomach, but it did not react with 10 cases of colonic adenocarcinomas by the immunoperoxidase technique. It did not stain non-cancerous tissues of 16 different organs, except for the renal tubules and for the secreting glands of the pancreas with which the antibody faintly reacts. It did not react with either granulocytes, lymphocytes or red blood cells. Positive staining was observed in various fetal tissues. Periodic acid and neuraminidase treatments on tissue sections suggested that the chemical nature of the antigenic determi- nant of YH206antigen was carbohydrate in nature which might be maskedby sialic acids. Key Words: Adenocarcinoma-associated antigen, , Monoclonal antibody

There have been some reports concerning MATERIALS AND METHODS monoclonal (MoAbs) to human lung adenocarcinoma associated . None of Monoclonal antibody YH206 these antibodies, however, were found to detect The MoAb YH206 was obtained from the fusion between murine myeloma X63-Ag8.653 corresponding antigen(s) in the sera of lung cancer cells and splenocytes of BALB/c mouse, which patients. Recently, Iguro et al. described that the was immunizedwith humanlung adenocarcinoma new monoclonal antibody, designated as CSLEX- A549 cell line . The method of preparation has 1, could detect the corresponding antigen in the sera of cancer patients including lung cancer. The been described elsewhere . Testing of the mono- antigenic determinant of this antigen was found clonal antibody with rabbit antisera to mouse Ig to be sialylated Lewisx carbohydrate moiety. classes and subclasses (Bionetics, Kensington, MD) in radial showed that the MoAb More recently, we established an interesting mono- YH206isIgM, K. clonal antibody (IgM), designated as YH206, which could detect the circulating antigen(s) in Tissue processing and fixa tion the sera of patients with various cancers (manu- Normal tissues as well as benign and malignant script submitted for publication). In this paper tissues were obtained from patients who under- we would like to describe the immunohistochemi- went surgery. Fetal tissues were obtained from 16- cal analysis of the corresponding antigen (YH206 and 24-week-old fetuses. A portion of each speci- antigen) detected by the MoAbYH206. men was fixed with 10%buffered formaldehyde. Twoconsecutive 5 /im-thick paraffin sections were obtained from each sample. One of the paraffin

From Department of Internal Medicine (Section 1), Sapporo Medical College, Sapporo. Received for publication September 25, 1985. Reprint request to: Akira Yachi, MD,Department of Internal Medicine, Sapporo Medical College, S-l, W-16, Chuo-ku, Sapporo 060,Japan.

JapJMedVol 25,No 2 (May 1986) 127 Yachi et al sections was stained with hematoxylin and eosin as a criterion of positive reactions of the tissue and was studied for its histologic features, while sections with the monoclonal antibody. Unreactive the other section was used as a substrate for tissue sections were tested further with concen- immunoperoxidase analysis. trated (5x) and diluted (10x) solutions of the Immunoperoxidase technique monoclonal antibody to avoid false negative re- A conventional procedure for the immunoper- actions attributable to either a low concentration oxidase technique was employed for the staining of monoclonal antibody or to a prozone-like effect. The specificity of the reaction was control- reactions with the monoclonal antibody. The sec- led by testing tissue sections with myeloma tions were first incubated with 0.6% H2O2 in methanol for 20 min to block endogenous per- protein secreted by the murine myeloma P3-X63- Ag8 cells and with RPMI1640 medium containing oxidase, and were then overlayed with the mono- 10% fetal calf serum. Immunoperoxidase staining clonal antibody for 30 min at room temperature in a moist chamber. After washing three times for ously6).of blood smears was performed as described previ- 5 min in cold phosphate buffered saline (PBS), peroxidase-conjugated rabbit anti-mouse IgM anti- Indirect technique serum (DAKOa/s Immunochemicals, Copenhagen, This technique was performed according to Denmark) was applied for an additional 30 min. standard procedure. FITC-conjugated rabbit anti- After a final wash with cold PBS, the sections were mouse IgM antiserum (DAKOa/s Immunochemi- reacted for 60 sec in a solution containing 40 mg cals, Copenhangen,Denmark)was used as a second 3,3'-diaminobenzidine and 0.01% hydrogen per- antibody. Cells were observed under fluorescent oxide in 100 ml 0.05M Tris-HCl buffer (pH 7.6), microscope (Zeiss, MC65, FRG). after which they were stained with 0.2% Methyl Periodic acid and neuraminidase treatment on Green for 3 to 5 min and were observed with a tissue sections light microscope. One hundred microliters of 1% periodic acid A distinct brown staining of the cells was used solution in PBS(pH 7.2) were placed on each sec- Table 1. Tissue distribution of the antigenic determinant recognized by the monoclonal antibody YH206.

C a n ce r tissu e N o rm al tissu e O rga n T y p e O r g a n A d u l t F e t u s L u n g a d e n o ca rc in o m a + + * (9 /1 3 )* * L u n g - (0 /8 ) + + (1 / 1 ) alv eo lar c el l c ar ci nom a + + (2 / 2 ) C o lo n - (0 /9 ) + (1 / 1 ) e pi d e r m o i d c a r c i no m a F + (1 / 6 1 K id n e y F + (3 /4 ) + + (1 /1 ) lar g e ce ll ca rc in o m a F + (1 / 5 ) L iv e r - (0 /4 ) - (0 / 1 ) sm a ll c ell c ar cin o m a - (0 / 6 ) P a n c re a s F + (3 / 3 ) + (1 /1 ) S m all in te stin e - (0 /4 ) F + (1 / 1 ) Pan cre a s a d e n o ca rc in o m a + + (3 / 4 ) S to m a c h - (0 /8 ) + (1 /1 ) S to m a c h a d e n o ca rc in o m a + (6 / 9 ) A d re n al glan d - (0 /1 ) n .t.* si gne t-r ing c ell ca rci nom a - (0 / 3 ) B ra in - (0 /1 ) n .t . E so p h a g u s - (0 / 2 ) n .t . ch o lan g io ca rc in o m a + (1 / 1 ) L iv e r H e ar t - (0 /1 ) n .t . he pa t oc e ll ul a r c ar ci n om a - (0 / 2 ) L y m p h n o d e - (0 /2 ) n .t . B re a st scirrh o u s ca rc in o m a - (0 / 2 ) S k ele tal m u sc le - (0 /1 ) n .t . S p in a l c o rd - (0 / 1 ) n .t . - (o /i o ) C o lo n a d e n o ca rc in o m a S p le e n - (0 / 2 ) n .t. K id n ey Gr aw it z 's t um or - (0 / 1 ) T h y ro id gla n d - (0 / 1 ) n .t. G all b la d d er a d e n o ca rc in o m a - (o /- i )

*staining intensity: ++; strongly positive, +; positive, F+; faintly positive, -; negative :no. positive/no, tested **no. positive/no. tested ***not tested

128 JapJMedVol 25, No 2 (May 1986) Immunohistology of adenocarcinoma antigen

tion for 10 to 20 min at room temperature. One Tokyo) or 0.1 U/ml endoglycosidase H (Seika- hundred microliters of 1 U/ml neuraminidase gaku-Kogyo, Tokyo) for one hour at 37°C as de- (Sigma, No. N-3001, USA) in 0.01M acetate buffer scribed by Zola et al. solution (pH 5.1) were placed on the section for 12 hours at 37°C. After three washes with PBS, RESULTS each section was stained by immunoperoxidase Staining pattern of cancerous tissues method as described above. MoAb YH206 stained the majority of lung adenocarcinomas and alveolar cell carcinomas, Enzymetreatment on the cultured cells One milliliter of the cultured A549 cells (2 x but did not stain or weakly stained the other 10 /ml) was incubated with a solution of either histplogic types of lung carcinoma. Adenocar- 0.1 U/ml of neuraminidase F (Seikagaku-Kogyo, cinomas of the pancreas and the stomach were Tokyo), 0.5 mg/ml of protease V8 (Miles, USA), also stained, whereas 10 cases of colonic adeno- 2 mg/ml of glycosidase mixed (Seikagaku-Kogyo, carcinomas were negative (Table 1). The predomi-

Fig. 1. Staining pattern of cancerous tissues by peroxidase technique.

secretionApical portionof gastricof lungadenocarcinomaadenocarcinoma (b)(a) wereand apicalstained. portionOriginal andmagnification:intra-luminal a,x90; b,x40.

JapJMedVol 25, No 2 (May 1986) 129 Yachi et al

nant staining pattern obtained in most lung adeno- apical portion and intraluminal secretion were carcinoma specimens wasa labelling of the apical essentially stained (Fig. lb). surfaces of tumor cells (Fig. la). Lung alveolar cell carcinomas (2 cases) showed cytoplasmic staining Staining pattern of noncancerous and fetal tissues In adult lung tissues, any portions including pattern. In gastric adenocarcinoma specimens. epithelial cells of the bronchus or the alveolus

Fig. 2. Staining pattern of noncancerous tissues. Renal tubules were weakly stained. Original magnification: x 1 80.

Fig. 3. The effect of periodic acid treatment on the staining intensity of lung adenocarcinoma tissues. The antibody YH206failed to stain the tissues after they had been treated with 1%periodic acid for 10 min at room temperature (a: before treatment, b : after treatment). Original magnification: x 40.

130 JapJMedVol 25,No 2 (May 1986) Immunohistology of adenocarcinoma antigen

Fig. 4. The effect of neuraminidase treatment on the staining of large cell carcinoma of the lung (a: before treatment, b-: after treatment). Original magnification: x 90.

Fig. 5. The effect of neuraminidase treatment on the staining intensity of noncancerous gastric mucosa (a: before treatment, b: after treatment). Original magnification: x 90. JapJMedVol 25, No 2 (May 1986) 131 Yachi et al and mucinousglands of the bronchus were not formed to determine whether or not sialic acid stained with the MoAb YH206 (Table 1). Among residue involves in the epitope recognized by the 52 tissue specimens of 16 different organs tested, monoclonal antibody. Either abolishment or de- the renal tubules and the secreting glands of the crease of the staining intensity was not observed pancreas were the only portions which were faint- in the tissue sections. On the contrary, as shown in ly stained with the MoAb YH206 (Fig. 2). On the Table 2, majority of the malignant tissues showed other hand, fetal tissues including lung, colon, pan- increased staining intensity after neuraminidase creas and stomach gave strong staining. Finally, treatment. Representative pictures of large cell either granulocytes, lymphocytes or red blood carcinoma of the lung were shown in Fig. 4. The cells did not show any positive staining in the similar tendency was observed in three out of blood smears. eight cases of non-malignant tissues (Table 2, Fig. Periodic acid and neuraminidase teratment on the 5). sec tions Enzymetreatment on the cultured cells Periodic acid and neuraminidase treatments Cultured lung adenocarcinoma A549cells were were done to investigate the chemical nature of the treated with protease V8, trypsin, glycosidase antigenic determinant recognized by the MoAb mixed and endoglycosidase H. Either decrease or YH206. This antibody failed to stain lung adeno- increase of the reactivity with the MoAbYH206 carcinoma tissues after they had been treated with was not observed. Neuraminidase treatment of the 1%periodic acid for 10 min at room temperature cells, however, showed the increased staining in- (Fig. 3). Neuraminidase treatment was also per- tensity whencomparedwith non-treatment cells (data not shown). Table 2. Comparison of the staining pattern detected by the monoclonal antibody YH206 before and after neuraminidase digestion in malignant and non- malignant tissues

Neuraminidase digestion T issu e B efore A fter

M alignan t tissu e (12 /13 )* Lung adenocarcinom a + + ** (2 /2) alveolar cell carcin om a (1/1) epidermoid carcinoma (1/1) large cell carcino m a F+ (1/D sm all cell carcin om a F + (1/D Stomac h adenocarcinoma (1 /1) signet ring cell carcinoma (2 /2) Colon adenocarcinoma (3 /3) H epato cellu lar carcino m a (0 /1)

N o n-m alignan t tissu e (3 /8) L u ng (1/2) S tom ach (2/2) C o lon (0 /3 ) L iver (0/1)

*No. cases that increased the staining intensity/no. cases tested **Staining intensity: +++; very strongly positive, +.+; strongly positive, +; positive, F+; faintly positive, -; negative

132 JapJMedVol 25,No 2 (May 1986) Immunohistology of adenocarcinoma antigen

DISCUSSION was different from that of latter two antigens. The Although the histological distribution of the epitope of CA19-9 was detected in the colonic antigenic determinant recognized by the MoAb carcinoma and in epidermoid and small cell car-

YH206was not specific for cancer tissues, it was fetalcinomalung14).of the Thelung epitope3 , but ofwasCSLEX-1not detectedantigen12)in the highly restricted to cancers, since it weakly stained merely two portions of the non-cancerous tissues was detected in granulocytes, in foveolar cells of tested: the renal tubules and the secreting glands the colonic epithelium, in liver cells, in Kupffer of the pancreas. Among lung carcinoma tissues, cells and in the colonic carcinoma, but was not it essentially stained the adenocarcinomas and detected in cultured A549 cells. Although the alveolar cell carcinomas, latter being histologically histologic distribution of the determinant of our classified as the adenocarcinomas. It did not react YH206antigen is noted to be different from that or weakly reacted with other histologic types of of CA125, CA50 or DU-PAN-2, respectively, fur- lung cancers, and the frequency of the positivity ther chemical analysis needs to be done to draw a was very low (two out of 17 cases). Therefore, this firm conclusion. Another interesting antigen to be MoAb YH206 might be useful for differentiating compared with our antigen would be Thomsen- adenocarcinomas from other histologic types of Friedenreich antigen (T-antigen), since the epitope lung cancer by means of 1) immunoperoxidase of T-antigen was also asialylated form ((Gal(1-3) technique using tumor cells in the pleural effusion GalNac)). This antigen was also known detectable and/or 2) radiolabeled monoclonal antibody in in the sera (Longenecker, B.M.: personal commu- tumor imaging technique. nication). Rahman et al. ' reported that the The YH206antigen was found to be a very high MoAbto T-antigen reacted with colonic carcinoma molecular-weight glycoprotein of more than 330K tissues. Furthermore, it is known that the epitope dalton by western blotting analysis whenthe ex- of T-antigen is exposed on red blood cells after tract of the cultured A549 cells was used as a they are treated with neuraminidase. On the crude antigen (Hinoda et al.: submitted for publi- contrary, the epitope of the YH206antigen was cation). In this study, periodic acid teratment and not detected on red blood cells after neuramini- neuraminidase digestion on the tissue sections dase treatment (unpublished observation). There- suggested that the antigenic determinant consisted fore, it is unlikely that our antibody recognizes of unique carbohydrate chain structures which the epitope of T-antigen. might be masked by sialic acids. The fact that the In collaboration with Dr. S.-I. Hakomori, Fred- antigenic determinant recognized by the YH206 Hutchinson Cancer Institute at Seattle, a complete on the A549 cells was not destroyed by trypsin chemical analysis of the antigenic determinant of treatment and by protease V8 treatment, sup- the YH206 antigen are now under investigation in ported that the determinant of the YH206antigen our laboratory. In any case this YH206 antigen is maybe carbohydrate in nature. of potential use for immunohistologic diagnosis of In the literature, there have been several tumor- cancer patients with adenocarcinomaas well as for associated antigens in the sera which was recently radioimmuno-detection and sero-immunodiagnosis detected by the monoclonal antibodies: CA19-9 , of malignant diseases. CA1259), CA5010), DU-PAN-210 and CSLEX-1- corresponding antigen4' 12. Interestingly, each ACKNOWLEDGEMENTS: We wish to acknowledge the encouragement of Dr. Takeo Wada (President of antigenic determinant of these antigens seems to Sapporo Medical College, Sapporo, Japan). Wealso wish be carbohydrate in nature. The epitope of our to acknowledge the English editing of Mr. H. Tarnoff YH206 antigen appears different from that of (English Instructor, Sapporo Medical College, Sapporo, CA19-9 or CSLEX-1 antigen, since they were Japan). This work was supported by a Grant from the Princess reported to consist of sialylated form of the blood- Takamatsu Cancer Research Fund and by Grants for group sugar chain ' . In addition, the histologic Cancer Research from the Ministry of Education, Science distribution of the epitope of the YH206 antigen and Culture and from the Ministry of Health and Welfare, Japan.

JapJMedVol 25,No 2 (May 1986) 133 Yachi et al

REFERENCES gastroinstestinal cancer is a ganglioside containing sialylated lacto-N-fucopentaose II. J Biol Chem 257: Mazauric T, Mitchell KF, Letchworth GJ, et al: 14365, 1982. Monoclonal antibody-defined human lung cell sur- 9) Bast RC, Klug TL, John ES, et al: A radioimmuno- face protein antigens. Cancer Res 42: 150, 1982. assay using a monoclonal antibody to monitor the Mulshine JL, Cuttita F, Bibro M, et al: Monoclonal course of epithelial ovalian cancer. N Engl J Med antibodies that distinguish non-small cell from small 309: 883, 1983. cell lung cancer.J Immunol 131: 497, 1983. 10) Holmgren J, Lindholm L, Persson B, et al: Detec- Varki NM, Reisfeld RA, Walker LE: Antigen asso- tion by monoclonal antibody of carbohydrate ciated with a human lung adenocarcinoma defined antigen CA50in serum of patients with carcinoma. by monoclonal antibodies. Cancer Res 44: 681, BrMedJ 288: 1497, 1984. 1984. ll) Metzgar RS, Rodriguez N, Finn OJ, et al: Detection Iguro T, Wakisaka A, Terasaki PI, et al: Sialylated of a pancreatic cancer-associated antigen (DU-PAN- Lewisx antigen detected in the sera of cancer pa- 2 antigen) in serum and ascites of patients with tients. Lancet 6: 817, 1984. adenocarcinoma. Proc Natl Acad Sci USA 81: 5242, Giard DJ, Aaronson SA, Todaro GJ, et al: In vitro 1984. cultivation of human tumors: Establishment of cell 12) Fukushima K, Hirota M, Terasaki PI, et al: Char- line derived from a series of solid tumors. J Natl acterization of sialosylated Lewisx as a new tumor- CancerInst 51: 1417, 1973. associated antigen. Cancer Res 44: 5579, 1984. Imai K, Moriya Y, Fujita H, et al: Immunologic 13) Atkinson BF, Ernst CS, Herlyn M, et al: Gastroin- characterization and molecular profile of carcino- testinal cancer-associated antigen in immunoperoxi- embryonic antigen detected by monoclonal anti- dase assay. Cancer Res 42: 4820, 1982. bodies.J Immunol 132: 2992, 1984. 14) Magnani JL, Steplewski Z, Koprowski H, et al: Iden- Zola H, Moore HA, HohmannA, et al: The antigen tification of the gastrointestinal and pancreatic of mature human B cell detected by the monoclonal cancer-associated antigen detected by monoclonal antibody FMC7: Studies on the nature of the antibody 19-9 in the sera of patients as a mucin. antigen and modulation of its expression. J Immunol Cancer Res 43: 5489, 1983. 133: 321, 1984. 15) Rahman AFR, Longenecker BM: A monoclonal anti- Magnani JH, Nilsson B, Brockhaus M, et al: A mono- body specific for the Thomsen-Friedenreich cryptic clonal antibody-defined antigen associated with T antigen.J Immunol 129: 2021, 1982.

134 JapJMedVol 25, No 2 (May 1986)