Immunohistochemical Evaluation of Seven Monoclonal Antibodies for Difierentiation of Pleural Mesothelioma From Lung Adenocarcinoma Patricia R. Wirth, BS," Jacques Legier, MD,tr$ and George L. Wright, Jr, PhD"

A panel of seven monoclonal antibodies including anti-vimentin, anti-keratin markers AE1/AE3 and EAB902, human milk fat globule (HMFG-21, €372.3, anti- carcinoembryonic antigen (CEA), and anti-Leu-M1 were used for an immunoperoxidase staining assay to determine their value in the differentiation of pleural mesothelioma from lung adenocarcinoma. Anti-vimentin positively identified 86% of the mesotheliomas and none of the adenocarcinomas. AEI/AE3, EAB902, and B72.3 reacted with a high percentage of both mesothelioma and adenocarcinoma specimens. With HMFG-2, both membrane and cytoplasmic staining was observed in 92% of the adenocarcinomas and in 14% of the mesotheliomas, whereas 26"/0 of the mesotheliomas only exhibited membrane staining. Eighty percent of the adenocarcinomas and 8% of the mesothelioma tissues stained with anti-Leu-MI. Anti-CEA did not react with any of the 50 mesotheliomas tested but did react with 95% of the lung adenocarcinomas tested. From this study, it was concluded that anti-CEA and anti-Leu-M1 were the most effective of the seven tumor markers evaluated; and that 100% of the pleural mesothelioma tissues could be correctly differentiated from lung adenocarcinomas using a panel consisting of anti-vimentin, HMFG-2, anti-CEA and anti-Leu-M1 monoclonal antibodies. Cancer 67:655-662,1991.

HE HISTOPATHOLOGIC DIFFERENTIATION of malig- T nant epithelial mesothelioma from adenocarcinoma is frequently difficult. A correct diagnosis is important to ensure that appropriate therapy is instituted. Histochem- ical and electron microscopic studies are used to aid in diagnosis but do not always give a conclusive answer. Im- munohistochemistry is now used as another tool in the differential diagnosis of mesothelioma from adenocarci- noma. The immunoperoxidase staining technique has been used to identify the presence or absence of a variety

From the *Department of Microbiology and Immunology and the t Department of Pathology, Eastern Virginia Medical School, Norfolk, Virginia, and the $Department of Pathology, Rwerside Regional Medical Center Hospital, Newport News, Virginia.

Supported in part by a grant from the Elsa U. Pardee Foundation, Midland, Michigan.

Address for reprints: George L. Wright, Jr., PhD, Department of Mi- crobiology and Immunology, Eastern Virginia Medical School, Lewis Hall, 700 Olney Road, P.O. Box 1980, Norfolk, VA 23501.

Accepted for publication April 9, 1990.

of antigenic substances in these tumors. Among those studied are vimentin, carcinoembryonic antigen (CEA),5-13 Leu-M 1 , 6 , 7 3 ' 4 Human Milk Fat Globule (HMFG-2),6-11.15 B72.3,9-16 and a variety of kera-

e results reported from several of these studies, however, have conflicted with data generated in our laboratory, as well as in laboratories of other in- ves t iga to r~ .~ ,~*~~ ' I 4

In this study, we compared the immunohistochemical staining patterns of seven monoclonal antibodies (MoAbs) on 50 pleural mesothelioma and 20 to 25 lung adenocar- cinoma specimens. By using all MoAbs, titering each of these antibodies on both mesothelioma and adenocarci- noma specimens to determine optimal staining intensity, and by screening the MoAbs on a large sample of the same histologic type of tumor tissue, we hoped to deter- mine which marker, or combination of markers, would work best in differentiating pleural mesothelioma from lung adenocarcinoma.

tins. 1-3.5-7.9,11-13,17 Th

655

656 CANCER February I 199 1 Vol. 67

Materials and Methods

Specimens

Fifty cases of pleural mesothelioma (36 epithelial, eight sarcomatous, and six biphasic) and 20-25 cases of lung adenocarcinoma were examined from the files of the Pa- thology Department of Riverside Regional Medical Center Hospital (Newport News, VA). The specimens were ob- tained from surgery as well as from autopsy. The diagnoses of most of the mesothelioma specimens used in this study were confirmed by special histochemical stains to detect the presence or absence of mucin, Leu-MI, CEA, and keratin. In cases in which there was enough tumor tissue, electron microscopic studies were also performed to study the morphology and identify the diagnostic features of mesothelioma. When available, lung tissue from a site other than that of the tumor was also examined for the presence of asbestos fibers. The tissues were formalin- fixed, routinely processed in ethanol, and embedded in paraffin wax. Sections were cut (4 ym) and stained with Mayer’s hematoxylin and eosin. Further serial sections were stained with the panel of MoAbs in an avidin-biotin immunoperoxidase assay (see below).

Anti-vimentin: This IgG 1 MoAb was purchased from DAKO, Santa Barbara, CA, and is specific for the 57Kd intermediate filaments. Anti-vimentin was supplied as di- alyzed culture supernatant and was used at a concentra- tion of 9.7 pg/ml.

AEI/AE3; This antibody is an anti-keratin pool with a 20: 1 mixture of AEI to AE3. AE1 is specific for most acidic (Type I) cytokeratins (CK) (50K [CK14] and 56.5K [CKIO]) and identifies the epidermal basal layer and al- most all epithelia. AE3 is specific for most basic (Type 11) cytokeratins (58K [CK5] and 65-67K triplet [CK1,2]) and identifies the entire epidermis and all epithelia. Both AE 1 and AE3 are IgGl isotypes. This antibody was supplied as purified ascites (BMB, San Diego, CA) and was used at a concentration of 10 pg/ml.

EZ4B902: This antibody is an anti-keratin panepithelial marker specific for the 54Kd (CK7) protein. It will dis- tinguish squamous from nonsquamous epithelium and has an IgM isotype. This antibody was supplied as ascitic fluid (Enzo Biochem Inc., New York, NY) and was used at a 1:2,500 dilution.

HMFG-2: This is an IgGl immunoglobulin directed against a carbohydrate found on the human milk fat glob- ule membrane and was purchased from Unipath, Bedford, England. The membrane component to which this anti- body binds is a large molecular weight glycoprotein im- munologically related to epithelial membrane antigen

(EMA). This antibody was supplied as culture supernatant and was used at a 1: 100 dilution.

B72.3: This antibody was generated against a mem- brane-enriched fraction of human metastatic breast car- cinoma and has been used in the identification of ade- nocarcinomas of the lung, breast, colon, and ovary (Szpak et af.).I6 The antigen to which this antibody binds has been shown to be a tumor-associated glycoprotein com- plex of high molecular weight, TAG-72. This IgGl anti- body was supplied as crude ascitic fluid and was used at a concentration of 8.2 &200 pl/slide. (B72.3 was a gift from Dr. Jeffrey Schlom, The National lnstitutes of Health, Bethesda, MD)

Anti-CEA: This antibody is specific for human carci- noembryonic antigen, i.e., does not bind normal colon antigen (NCAg) (Zymed Laboratories, San Francisco, CA). Anti-CEA is an IgGl antibody that was supplied purified from ascitic fluid and was used at a concentration of 5 &ml.

Anti-Leu-M1: This is an IgM immunoglobulin specific for the human myelomonocytic antigen Leu-M 1 /CD 15. The antibody was supplied as purified ascites (Becton Dickinson, San Jose, CA) and was used at a concentration of 25 yg/ml.

Immunoperoxidase Procediire

Immunohistochemical staining with the seven MoAbs just described was performed according to a modified av- idin-biotin method using the ABC Vectastain kit (Vector Labs, Burlingame, CA). Whenever possible, the seven MoAbs were evaluated on serial sections of each tissue in order that the reactivity of the various antibodies could be compared on the same area of the tissue specimens. Tissue slides were heated overnight in a 60 C oven, de- paraffinized in two changes of xylene, and rehydrated through a series of graded alcohols. Endogeneous perox- idase activity was blocked by a 30-minute incubation in a 0.43% hydrogen peroxide-methanol solution and washed in phosphate-buffered saline. Sections stained for the anti- keratin MoAbs AE 1 /AE3 and EAB902 were pretreated with 0.05% pronase for 5 minutes and 10 minutes, re- spectively. Nonspecific binding of MoAbs was blocked in a 20-minute incubation with 10% normal horse serum. Sections were then incubated for 1 hour with MoAbs (overnight for B72.3 and 30 minutes for anti-CEA), washed, and then incubated 30 minutes with biotinylated horse anti-mouse IgG (H+L) immunoglobulin. After washing, the sections were incubated for 45 minutes with an avidin biotinylated peroxidase complex and washed again. The binding of the MoAbs to the tissue antigens was visualized after a 5-minute incubation with 0.05% 3,3’diaminobenzidine tetrahydrochloride (Sigma Isopac #D90 15) in PBS with 0.1% H202. Sections were counter-

No. 3 DIFFERENTIATION OF MESOTHELIOMA FROM ADENOCARCINOMA - Wirth et al. 657

stained with Mayer’s hematoxylin and 0.3% ammonia water and then coverslipped. Positive tissue controls were performed with each MoAb used. Either culture super- natant from a mouse myeloma cell line, which secretes an IgG 1 antibody, or purified mouse IgM ( 1 Fg/ml) was used as a negative antibody control. The percentage of cells staining and the staining intensity of all sections were scored independently by two individuals.

Results

The immunoperoxidase staining results of the meso- thelioma and adenocarcinoma tissues with the panel of seven MoAbs are summarized in Table 1.

Anti-vimentin: The tumor cells present in the 20 lung adenocarcinoma specimens tested were negative for vi- mentin (Fig. lA, Table 1). Anti-vimentin also stained normal cells of mesenchymal origin in all tissue specimens tested, often several times stronger than the tumor cells, making interpretation tedious and difficult (Figs. 1 A and 1B). One hundred percent of the tumor cells in the sar-

TABLE 1. Immunostaining Results of Formalin-Fixed Lung Tumor Tissues

Mesotheiioma Adenocarcinoma

Monoclonal NO. pas/ No. pos/ antibody no. test (% POS) no. test (W pos)

Anti-vimentin E S B

E S B

E S B

E S B

B72.3 E S B

E S B

E S B

AEI/AE3

EAB902

HMFG-2

Anti-CEA

Anti-Leu-M 1

29/36 818 616

0120 (0)

20120 (100)

24/25 (96)

15/20 (75)

19/20 (95)

E: epithelial pattern of mesothelioma; S: sarcomatous pattern of me- sothelioma: 9 biphasic-mixed epithelial and sarcomatous patterns of mesothelioma.

* 1/16 tumor secretions positive. tumor cells negative.

comatous (8/8) and biphasic (6/6) mesotheliomas stained intensely (Fig. 1 B). In the biphasic mesothelioma speci- mens, most of the positively staining cells were of the sarcomatous type, whereas only a few of the epithelial cells were stained. Eighty percent (29/36) of the epithelial mesothelioma specimens were positive, with staining in- tensity of the tumor cells ranging from pale to dark. How- ever, only a few tumor cells (less than 10%) within each of these specimens were stained, with most of the tumor cells being negative. The staining observed with MoAb anti-vimentin was localized within the cytoplasm of both the tumor and mesenchymal cells.

Anti-keratin: Monoclonal antibody AE I /AE3 stained most of the tumor cells in all 20 lung adenocarcinoma tissues tested and 90% (45/50) of the pleural mesothelioma specimens with varying degrees of staining intensity within the cytoplasm (Table 1, Figs. 1C and 1D). Monoclonal antibody EAB902 stained 88% of the tumor cells in both the lung adenocarcinoma ( 14/16) and pleural mesothe- lioma (44/50) specimens, also with varying degrees of staining intensity within the cytoplasm. EAB902, how- ever, did stain the pleural mesothelioma cells more in- tensely than the lung adenocarcinoma cells.

Anti-CEA: Ten percent to 100% of the tumor cells in 19/20 (95%) of the lung adenocarcinoma specimens showed positive cytoplasmic staining with the anti-CEA MoAb (Table 1). These positive tumor cells exhibited a range in staining intensity from dark to pale (Fig. 1E). All 50 of the mesothelioma tissues were negative for this marker (Fig. IF).

HMFG-2: This antibody stained 96% (24/25) of the lung adenocarcinoma tissues, as well as 38% (19/50) of the pleural mesothelioma specimens (47% or 17/36 of the epithelial type stained). Of the 24 positive lung adeno- carcinoma tissues, 23 exhibited both cytoplasmic and membrane staining, and one exhibited only membrane staining. Five percent to 90% of the tumor cells had a range in staining intensity from pale to dark, with most of the tumor cells exhibiting moderate staining intensity (Fig. 1G). In the pleural mesothelioma specimens, most of the staining was associated with the cell membrane only. There was, however, cytoplasmic staining observed in seven of the 17 epithelial mesothelioma specimens, which also displayed membrane staining (Fig. 1 H).

B72.3: Fifteen of the 20, or 75%, of the lung adeno- carcinoma tissues tested were positive for the TAG-72 antigen (Table I) . Intense cytoplasmic and membrane staining was observed in 10% to 70% of the tumor cells in these specimens (Fig. 1 I). In contrast, 1 % to 15% of the tumor cells in 48% of the pleural mesothelioma specimens showed pale cytoplasmic staining with an occasional darkly stained tumor cell (Fig. I J).

Anti-Leu-MI: This antibody stained 80% (16/20) of the lung adenocarcinoma tissues tested (Table 1). Twenty

658 CANCER February 1 199 1 Vol. 67

FIGS. IA-IF. (A) Lung adenocarcinoma section stained with anti-vimentin. Note staining of normal mesenchymal cells (m) and lack of staining in tumor cells (T) (X400). (B) Two biphasic mesotheliomas stained with anti-vimentin. Note staining of both normal mesenchymal cells (arrows) and tumor cells (X400). (C) Lung adenocarcinoma stained with anti-keratin AE I/AE3. The neoplastic cells show intense cytoplasmic staining (X400). (D) Epithelial mesothelioma stained with anti-keratin AE 1/AE3. Most of the mesothelioma cells also show intense cytoplasmic, while others display focal or pennuclear staining (X400). (E) Lung adenocarcinoma reacted with anti-CEA. Tumor cells exhibit pale to dark cytoplasmic staining (X400). (F) Epithelial mesothelioma section showing an example of weak background but not specific staining of the normal and tumor cells with anti-CEA (X400). The brown-stained cells (peroxidase-DAB reaction) in each figure are the locations of specific MoAb binding. The blue counterstain is hematoxylin.

No. 3 DIFFERENTIATION OF MESOTHELIOMA FROM ADENOCARCINOMA - Wirth et a/. 659

FIGS. 1G-IL. (G) Lung adenocarcinoma stained with HMFG-2 showing intense cytoplasmic staining of the tumor cells (X400). (H) Epithelial mesothelioma showing both intense HMFG-2 membrane staining (arrows) and intense to pale cytoplasmic staining of the neoplastic cells (X400). (I) Lung adenocarcinoma stained with B72.3. Note the intense membrane (arrows) and cytoplasmic staining of the tumor cells (X400). (J) Epithelial mesothelioma stained with B72.3. Note that the majority of the positive-stained tumor cells exhibit pale cytoplasmic staining (X400). Scattered dark staining neoplastic cells (arrows) were occasionally observed (see insert, X200). (K) Lung adenocarcinoma stained with anti-Leu-M 1. Tumor cells display both varying degrees of cytoplasmic, membrane and focal staining (X400). (L) Biphasic mesothelioma stained with anti-Leu-M 1. Pale to dark cytoplasmic and focal staining of the tumor cells is present (X400). The brown-stained cells (peroxidasc-DAB reaction) in each figure are the locations of specific MoAb binding. The blue counterstain is hematoxylin.

660 CANCER February 1 199 1 Vol. 61

percent to 90% of the tumor cells in 15/16 (75%) positive specimens had a range in staining intensity from pale to dark. Both cytoplasmic and membrane staining were ob- served (Fig. 1K). The one remaining positive specimen exhibited a different staining pattern in which the tumor secretions were strongly positive, but the tumor cells were negative. This one specimen was scored as positive and was included in the group of lung adenocarcinoma spec- imens determined to be positive for this marker. Eight percent of the mesothelioma specimens (3/36 epithelial and I /6 biphasic) stained with the anti-Leu-M 1 antibody. Granules within 2% of the tumor cells in two of the three positive epithelial mesothelioma tissues showed pale staining. Moderate to dark cytoplasmic staining was ob- served in 80% of the tumor cells in one of the three positive epithelial specimens and in the one positive biphasic me- sothelioma specimen (Fig. 1 L).

Discussion

Many laboratories have performed immunohisto- chemical studies using various combinations of mono- clonal and polyclonal antibodies in an attempt to differ- entiate between mesothelioma and adenocarcinoma. The results from these studies, and those of the study described in this report, have generated divergent results. Others have also noticed these contradictory reports and ex- pressed concern about them.l7 We have therefore made a careful attempt to address the issues associated with the interpretation of the reactivity of seven of these antibodies. In most cases, serial sections of each tissue examined were used to precisely document the reactivity of the different antibodies to the same tissue areas, and color photographs representative of each antibody reaction have been in- cluded to clearly illustrate the staining patterns and prob- lems of interpretation.

Vimentin appeared to be the only marker of the seven we tested to positively identify pleural mesothelioma. Our results are in agreement with others who also found anti- vimentin to be positive for mesothelioma and negative for adenocarcinoma. Churg, however, recommended that, to make optimal use of vimentin staining for the diagnosis of mesothelioma, fixatives other than formalin be used because vimentin is not believed to retain its an- tigenicity after formalin fixation. Formalin fixation did not seem to be a problem in this study because tumor cells were found to be stained in 100% of the sarcomatous and biphasic mesothelioma specimens and 80% of the epithelial mesothelioma specimens.

Although vimentin appears to be a positive marker for mesothelioma, a major problem with its use is that it strongly stains normal mesenchymal cells, making inter- pretation difficult, especially when poorly differentiated tumors are intimately intermixed with stromal cells.

Anti-keratin MoAbs AE I/AE3 and EAB902 were un- able to differentiate between mesothelioma and adeno- carcinoma. Positive staining was observed in a high per- centage (88%- 100%) of the mesothelioma and adenocar- cinoma specimens tested. These results compare favorably with those of others who also were unable to differentiate between mesothelioma and adenocarcinoma by showing that keratin was present in both mesotheliomas and ad- enocar~inornas.'.~.~-~,~ ' Other investigator^,'^.'^ using a polyclonal rabbit anti-human keratin, have shown strong positive staining of tumor cells present in mesothelioma tissues and weak or no staining of lung adenocarcinoma tumor cells, and have suggested that this antibody could be used for distinguishing between these two tumor types. This polyclonal antibody was generated against cytoker- atins purified from human stratum corneum varying from 4 1 Kd to 65Kd. Unfortunately, these investigators did not report which specific classes of keratins were identified with this antibody in the mesothelioma and adenocarci- nomas tested. With increasing evidence that different classes of cytokeratins are deleted or overexpressed in car- cinomas, additional studies are needed.to verify if specific classes of cytokeratins are expressed in pleural mesothe- liomas and not in adenocarcinomas.

Two patterns of HMFG-2 staining were observed in our study: cytoplasmic and membrane. Several investi- gators have also reported similar observations. However, there is disagreement about how to interpret these staining patterns.'-' ' All laboratories6-' reported strong positive cytoplasmic staining in the adenocarcinomas tested, with most specimens also exhibiting cell membrane staining. Membrane staining of epithelial mesotheliomas was re- ported by each investigator (14%-57%), with some re- porting weak to moderate cytoplasmic staining in a small number of cases as The manner in which these findings were reported differed among laboratories. Bat- tifor and Kopinski' reported a study in which only a few of the epithelial mesotheliomas tested displayed cell membrane staining, with none of the specimens exhibiting cytoplasmic staining, and concluded that cytoplasmic staining should be present for a result to be considered positive.' Others have reported epithelial mesotheliomas as being positive for HMFG-2 regardless of where the staining was observed.'.'' We recorded 17 of the 36 epi- thelial mesotheliomas as being positive for HMFG-2, with seven of 17 exhibiting weak to moderate cytoplasmic staining, as well as membrane staining. Membrane stain- ing alone was observed in ten of 17. By recording either membrane or cytoplasmic staining as positive for this marker, we found that HMFG-2 may be useful as a neg- ative marker for epithelial mesothelioma when used in conjunction with other markers.

The staining results of B72.3 reported in this study are similar to those of Szpak d a1.I' In this study, intense

No. 3 DIFFERENTIATION OF MESOTHELIOMA FROM ADENOCARCINOMA - Wirth el al. 66 1

staining was observed in at least 10% or more of the tumor cells in 75% of the adenocarcinomas examined. Szpak et a1.,I6 on the other hand, found 86% of the adenocarci- nomas to be positive for B72.3. We observed pale staining in 1 % to 15% of the tumor cells in 24/50 mesotheliomas tested and reported these specimens as being positive for this marker. Szpak et ~ 1 . ’ ~ interpreted a positive reaction to be one in which 10% or more of the tumor cells were positive for B72.3. Therefore, the positive staining they found in 1% to 10% of the tumor cells in nine of 19 me- sothelioma specimens was reported as being negative for B72.3. The discrepancy between these reports lies not with the immunohistochemical staining observations, but with how these results were interpreted and reported.

Although anti-CEA is a negative marker for mesothe- lioma, it was clearly the best marker in this panel for distinguishing between pleural mesothelioma and lung adenocarcinoma. None of the 50 mesothelioma specimens tested were reactive with anti-CEA. Of the 20 lung ade- nocarcinoma tissues tested, CEA was detected in 19. These results agree with those of recent studies reported by other investigators (Table 2).

Most of the lung adenocarcinoma tissues tested were positive when stained with anti-Leu-MI (80%). In con- trast, only 8% of the mesothelioma tissues were positive for this marker. In addition, the staining patterns of these two tumor types differed. The adenocarcinoma tissues exhibited both cytoplasmic as well as membrane staining, whereas two of the four positive mesothelioma tissues showed only cytoplasmic staining and two exhibited pos- itive staining granules within the tumor cells. Our results agree with those of other investigators who have also re- ported a high percentage of adenocarcinoma specimens to be positive with anti-Leu-M 1. These investigators, however, have reported that pleural mesothelioma is neg- ative for this tissue

The conflicting immunohistochemical reports can be

attributed to several factors, the first of which is the method used for preparation of the tissue. Frozen and fixed tissues can result in different staining patterns, and, if the tissue was fixed, the type of fixative used may also have an effect on the stability of the antigen. For example, it may alter or destroy antigenic sites. Second, some lab- oratories used polyclonal antibodies while others used MoAbs. Third, the type of immunoperoxidase staining procedure used varied among investigators. Most labo- ratories tailor their staining procedure to the antibodies they are assaying. Incubation times as well as antibody and reagent concentrations may also vary. Also, some immunoperoxidase staining procedures are considered to be more sensitive than others. Finally, one of the strongest reasons for these divergent results are the variable inter- pretations of the staining reaction.

Both technique and the analysis of staining results must be standardized before immunohistochemical staining can be useful in the differential diagnosis of malignant me- sothelioma and lung adenocarcinoma. There is also a need for positive rather than negative markers for identifying mesothelioma so that pathologists will not have to rely on “subtraction diagnosis.”

None of the MoAbs in our panel was completely specific for mesothelioma or adenocarcinoma when used alone. Although anti-vimentin was unable to identify all of the mesothelioma specimens tested, it was shown to be a pos- itive marker for this neoplasm. Neither anti-keratin marker or B72.3 was helpful in distinguishing mesothe- lioma from adenocarcinoma. The staining reactivity of HMFG-2 positively identified most of the adenocarci- noma specimens from the mesothelioma specimens. Re- gardless of the staining patterns accepted as positive (i.e., cytoplasmic vs. membrane), a small number of mesothe- lioma specimens did express this marker and, therefore, HMFG-2 should be used in conjunction with other markers. Anti-CEA was the most effective of the MoAbs

TABLE 2. Results of Previous lmmunohistochemical Studies of Adenocarcinomas and Mesotheliomas Using Anti-carcinoembryonic Antibodies

Mesothelioma Lung adenocarcinoma

Authors Tissue Antibody No. pos/no. test (% POS) No. pos/no. test (% Pod

Sheibani* FR MoAb 36/50 (72) 0128 (0) Szpak16 FX PC 19/20 (95) 0120 (0)

Battifora7 FX PC 24/24 (100) 2/12 (17) Marshall’ FX PC 6J7 (86) 0115 (0) Conon” FX PC 20120 (100) 9/20 (45) Said l4 FX PC 616 (100) 218 (25) Tron ’ I FX PC 17/20 ( 8 5 ) 3/20 (15) PfaltzI7 FX PC 22/22 (100) 2/47 (4) Otis” FX PC 13/14 (93) 119 (1 1) Strickler l o FX PC NT 016 (0)

Ghosh’’ FX MoAb 314 (75) NT

FR: frozen; FX: fixed; MoAb: monoclonal antibody; PC polyclonal antibody; NT: not tested.

662 CANCER February 1 199 1 Vol. 67

screened in this study. Although anti-CEA did not stain any of the mesothelioma specimens tested, 95% of the lung adenocarcinoma specimens could be correctly iden- tified. The reactivity of anti-Leu-MI was similar to that of anti-CEA, but was not as sensitive or specific. The one adenocarcinoma tissue that was anti-CEA negative was positively identified by anti-Leu-M 1.

In conclusion, a MoAb panel consisting of anti-vimen- tin, HMFG-2, anti-CEA, and anti-Leu-M 1 was found to be clinically useful for the differential diagnosis of pleural mesothelioma. By using this panel, 100% of the mesoth- eliomas tested in this study could be correctly identified and differentiated from lung adenocarcinomas. The lack of a specific marker for mesothelioma continues to make diagnosis by immunoperoxidase alone difficult. Our lab- oratory is now in the process of generating MoAbs to identify a marker specific for pleural mesothelioma to dif- ferentiate this tumor type from lung adenocarcinoma.

REFERENCES

I . Churg A. Immunohistochemical staining for vimentin and keratin in malignant mesothelioma. AIM J Sitrg Puthol 1985; 9:360-365.

2. Blobel GA. Moll R. Franke WW, Kayser KW, Could VE. The intermediate filament cytoskeleton of malignant mesotheliomas and its diagnostic significance. A m J Puzhol 1985; I2 1:235-247.

3. Mullink H, Henzen-Logmans SC, Alons-van Kordelaar JJM, Tad- ema TM, Meijer CJLM. Simultaneous immunoenzyme staining of vi- mentin and cytokeratins with monoclonal antibodies as an aid in the differential diagnosis of malignant mesothelioma from pulmonary ad- enocarcinoma. Virchows .4rch [B] 1986; 52:55-65.

4. Jasani B, Edwards RE, Thomas ND, Gibbs AR. The use of vimentin antibodies in the diagnosis of malignant mesothelioma. Virchows Arch [A] 1985; 406:441-448.

5. Strickler JG, Herndier BG, Rouse RV. Immunohistochemical staining in malignant mesotheliomas. A m J Clin Puthnl 1987; 88:6 10- 614.

6. Battifora H. Kopinski MI. Distinction of mesothelioma from ad- enocarcinoma: An immunohistochemical approach. Cuncrr 1985; 55:

7. Sheibani K, Battifora H, Burke JS. Antigenic phenotype of malig- nant mesotheliomas and pulmonary adenocarcinomas: An immunohis- tologic analysis demonstrating the value of Leu M I antigen. Am J Pathol 1986; 123912-219.

8. Marshall RJ. Herbert A, Braye SG, Jones DB. Use of antibodies to carcinoembryonic antigen and human milk fat globule to distinguish carcinoma, mesothelioma. and reactive mesothelioma. J Clin Pathol 1984; 37: 1215-122 1.

9. Otis CN, Carter D, Cole S, Battifora H. Immunohistochemical evaluation of pleural mesothelioma and pulmonary adenocarcinoma: A bi-institutional study of 47 cases. Am J Swg Pathol 1987; 11:445-456.

10. Tron V. Wright JL, Churg A. Carcinoembryonic antigen and milk- fat globule protein staining of malignant mesothelioma and adenocar- cinoma of the lung. Arch Paihol Lab Med 1987; 1 I1:291-293.

11. Ghosh AK, Gatter KC, Dunnill MS, Mason DY. Immunohis- tological staining of reactive mesothelium, mesothelioma. and lung car- cinoma with a panel of monoclonal antibodies. J Clin Pathol 1987; 40:

12. Corson JM. Pinkus GS. Mesothelioma: profile of keratin proteins and carcinoembryonic antigen. An immunoperoxidase study of 20 cases and comparison with pulmonary adenocarcinomas. Am J Puthol 1982;

13. Said JW. Nash G, Tepper G, Banks-Schlegel S. Keratin proteins and carcinoembryonic antigen in lung carcinoma: An immunoperoxidase study of fifty-four cases, with ultrastructural correlations. Hum

14. Sheibani K, Battifora H, Burke JS, Rappaport H. Leu-M1 antigen in human neoplasms: An immunohistologic study of 400 cases. A m J Sirrg Puthol 1986: 10:227-236.

15. Ghosh AK, Sprigs AI. Taylor-Papadimitriou J, Mason DY. Im- munocytochemical staining of cells in pleural and peritoneal effusions with a panel of monoclonal antibodies. J Clin Puttiol 1983; 36:1154- 1 164.

16. Szpak CA. Johnston WW. Roggli V et al. The diagnostic distinc- tion between malignant mesothelioma of the pleural and adenocarcinoma of the lung as defined by a monoclonal antibody (B72.3). Am J Pulhol 1986; 122252-260.

17. Pfaltz M, Odermatt B, Christen B, Ruttner JR. Immunohisto- chemistry in the diagnosis of malignant mesothelioma. Virchons Arch [A] 1987; 41 1:387-393.

1679-1685.

19-25.

108:80-88.

P~tholl983: 14:70-76.