86 Clin Pathol 1995;48:826-831

Production and characterisation of anantimelanoma monoclonal antibody KBA.62using a new melanoma cell line reactive on

paraffin wax embedded sections

E Cohen-Knafo, T Al Saati, J Aziza, E Ralfkiaer, J Selves, B Gorguet, G Delsol

Department ofPathology and CIGH/CNRS,CHU Purpan,Toulouse,FranceE Cohen-KnafoT Al SaatiJ AzizaJ SelvesB GorguetG Delsol

Department ofPathology,Herlev Hospital andUniversity ofCopenhagen,Herlev,DenmarkE Ralfkiaer

Correspondence to:Dr G DelsolLaboratoire d'AnatomiePathologique,CHU Purpan,Place du Dr. Baylac,31059 Toulouse Cedex,France.

Accepted for publication18 January 1995

AbstractAimns-To generate new monoclonal anti-bodies directed against melanoma as-sociated antigens using a new melanomacell line, KAL.Methods-The melanoma cell line was es-tablished in culture from a lymph nodemetastasis of malignant melanoma.Normal Balb/c mice were immunised withKAL cells. Splenocytes were used forfusion experiments using standard tech-niques. Hybridoma supernatants weretested for antibody binding activity usingan indirect inumunoperoxidase method onfrozen sectionsfromKALtumour cellsxeno-grafted onto nude mice and human ton-sils. KBA.62 was selected because of itsreactivity with melanocytic proliferationson both frozen and paraffin wax sections.Results-On immunoblotting, KBA.62 re-acted with three bands of 140, 135 and 128kD and two weak bands of88 and 73 kD. Innormal human tissues basal melanocytesin the epidermis did not react with thisantibody and only occasional labelling ofendothelial cells was noted. Of the humantumours, KBA.62 reacted strongly anduniformly with the majority ofbenign (21/21) and malignant (75/86) melanocytic pro-liferations. Staining was localised pre-dominantly to the cell membrane withlittle or no cytoplasmic reactivity. Negativestaining was observed in the majority ofhuman non-melanocytic neoplasms, theexceptions being some carcinomas (11/89),particularly the well differentiated squam-ous cell type. This, however, was notthought to present a diagnostic problem.Conclusions-KBA.62 appears to bepotentially useful in ascertaining theimmunomorphological diagnosis of mal-ignant melanoma in routinely processedparaffin wax sections.(J Clin Pathol 1995;48:826-831)

Keywords: KBA.62, melanoma, paraffin wax sections.

Malignant melanoma may sometimes con-stitute a diagnostic problem in surgical path-ology and may mimic other neoplasms, such asundifferentiated carcinoma, sarcoma or largecell lymphoma. The advent of monoclonalantibodies directed against formalin resistantmelanoma associated antigens has provided

a useful tool for the diagnosis of malignantmelanomas even in their amelanotic form.'-3A few antimelanoma antibodies, reactive onparaffin wax sections, are now available. Theusefulness of anti-SlO0 protein and NKI/C-3in the diagnosis of melanocytic proliferationsis limited as these antibodies recognise antigensnot restricted to melanoma. Indeed, theseantibodies react with a broad range of benignand malignant neoplasms24"6 and thus mustbe used in conjunction with other anti-bodies to allow definitive immunohisto-chemical diagnosis.27 As yet, HMB-45 hasproved to be the most specific and useful mono-clonal antibody in routine histopathology. 8

Here, we report the production of a mono-clonal antibody (KBA.62) directed against amelanoma associated antigen which survivesfixation and thus can be used on routinelyprocessed paraffin wax sections. Detailedanalysis of the reactivity of KBA.62 on normaland malignant human tissues is presented.These findings in melanocytic proliferationswere compared with those obtained with HMB-45, which was used as a reference antibody.Several cases were also immunostained withanti-S 100 polyclonal antibody.

MethodsESTABLISHMENT OF THE KAL CELL LINE ANDITS TRANSPLANTATION INTO NUDE MICEThe melanoma cell line (KAL) was establishedin culture from a lymph node metastasis ofmalignant melanoma. A fragment of the freshtissue was minced with a scalpel blade andincubated in RPMI medium supplementedwith 20% fetal calfserum (FCS) and antibioticsat 37°C in a 5% CO2 humidified environment.After 24 hours, only adherent cells were con-served. After two weeks in culture, cells growingin a monolayer were harvested using trypsin/EDTA (0 05/0 02%) and used for further sub-culture. The cell line, KAL, has been nowmaintained in RPMI/10% FCS for more thantwo years. The inoculation of 10-20 x 106 'KALcells subcutaneously into nude mice causedtumour formation in all inoculated mice, withinabout three weeks. The KAL tumours fromheterotransplanted animals were either snapfrozen in liquid nitrogen or fixed in Duboscq-Brasil (ethanol based Bouin's) fluid and pre-pared for later screening of hybridoma su-pernatants.9

826

A novel anti-melanoma antibody reactive on paraffin wax sections

Table I Reactivity ofKBA. 62, HMB-45 and anti-SO00 with melanocytic tumours on paraffin wax embedded sections

KBA. 62 HMB-45 Anti-SlOOTumour type (positiveltested) (positiveltested) (positiveltested)

Naevijunctional 1/1 ND NDintradermal 7/7 3/6* 3/3compound 8/8 2/5* 3/3spitz 2/2 2/2 NDdysplastic 3/3 3/3 NDtotal 21/21 (100%) 10/16 (62%) 6/6 (100%)

Melanomasprimary

superificial spreading 27/29 12/13 13/13nodular 6/7 4/4 4/4acral lentiginous 1/1 1/1 1/1lentigo 2/2 2/2 2/2

metastatic 39/47 30/39 27/29total 75/86 (87%) 49/59 (83%) 47/49 (96%)

Total 96/107 (90%) 59/75 (79%) 53/55 (96%)

ND=not done.* Intradermal naevi and the dermal component of compound naevi were largely non-reactive and only scattered naevic cells inthe papillary dermis are labelled with HMB-45.

PRODUCTION OF MONOCLONAL ANTIBODIES ANDSCREENING METHODSNormal Balb/c mice were immunised by fourintraperitoneal injections of 10-20 x 106 KALcells at two weekly intervals. Three days afterthe last injection, mice were sacrificed and thesplenocytes were used for fusion experimentsusing standard techniques.'0 When hybridomagrowth could be detected, supernatants weretested for antibody binding activity using anindirect immunoperoxidase method on frozensections from both KAL tumour cells andhuman tonsils.9 Of six antibody producing

Table 2 Reactivity ofKBA. 62 with non-melanocytictumours on paraffin wax embedded sections

KBA. 62Tumour type (positiveltested)

Carcinomaskinsquamous cell 5/6*basal cell 2/4

lungsquamous cell 2/7*adenocarcinoma 0/10small cell 0/6

gastrointestinal tractstomach 1/9small intestine 0/4large intestine 0/9

breast 1/8thyroid 0/11pancreas 0/1ovary 0/2kidney 0/4prostate 0/2nasopharynx 0/2thymus 0/1unknown origin 0/3

Haemopoietic neoplasmsnon-Hodgkin's lymphomas 0/28Hodgkin's disease 0/19

Nerve tumours 0/7Vascular tumours 0/9Othersmammary fibroadenoma 0/1carcinoid 0/6pheochromocytoma 0/2adrenocortical carcinoma 0/2nephroblastoma 0/3renal oncocytoma 0/1thymoma 0/2mesothelioma 0/1leiomyoma 0/4histiocytofibroma 0/1liposarcoma 0/1synovialosarcoma 0/1leiomyoblastosarcoma 0/1rhabdomyosarcoma 0/3Ewing's sarcoma 0/1

Total 11/182 (6%)* Particularly well differentiated squamous cell tumour.

hybridomas reacting with melanoma associatedantigens, KBA.62 was selected because of itsreactivity with melanocytic proliferations onboth frozen and paraffin wax sections (seelater). The KBA.62 hybridoma was cloned bylimiting dilution and its isotype was determinedusing immuLnohistochemistry. "

IMMUNOSTAINING OF VARIOUS HUMAN NORMALAND NEOPLASTIC TISSUE SECTIONSSupernatants were produced and used forstudying KBA.62 reactivity on tissue sections ofdifferent tumours (n = 289) and normal tissues(n= 54), from our tissue bank, by the Strept-avidin-biotin peroxidase complex (ABC)method using the Dako StreptABComplex/HRP Duet (Mouse/Rabbit) kit (Dako, Glo-strup, Denmark) without prior trypsinisation,as described elsewhere.'2 Some cases were alsostained using a Streptavidin-biotin alkalinephosphatase method.'3 Improved staining (thatis, enhanced antigen retrieval), particularly on10% formalin fixed specimens, was achievedfollowing pretreatment of tissue sections in amicrowave oven. 14 Only paraffin wax em-bedded tissue sections were used. Tissuesamples were fixed either in Duboscq-Brasilfluid (Toulouse) or in 10% formalin (Cop-enhagen). The human tumours investigated inthis study included: 107 cases of melanocyticneoplasms (table 1), comprising 21 naevi and86 cases of malignant melanoma (the latterincluded 39 primary cutaneous melanomas ofdifferent histological types and 47 cases oflymph node metastases); and 182 cases ofnon-melanocytic tumours of different categories(table 2). The reactivity of KBA.62 with mel-anocytic proliferations was compared with thatof the commercially available HMB-45 mono-clonal antibody (Dako A/S, Copenhagen, Den-mark), which was used as a reference antibody.Most of the cases were also immunostainedwith anti-S 100 polyclonal antibody (Dako A/S) using the peroxidase-antiperoxidase (PAP)immune complex method with prior tryp-sinisation. 5 Unexpected reactivities ofKBA.62antibody were also assessed on paraffin waxsections using the multitumour (sausage) tissueblock method.'6

827

Cohen-Knafo, Al Saati, Aziza, Ralfkiaer, Selves, Gorguet, et al

IMMUNOBLOTTING ANALYSISFor immunoblotting analysis, KAL cells werelysed in lysis buffer containing 1% Nonidet P-40 (Sigma, St Louis, Missouri, USA) in 50mMTris/HCl (pH 8-0), 0-15M NaCl, 0-57mMphenylmethylsulphonyl fluoride, and 0 15,Maprotinin at 40C for 30 minutes. After cent-rifugation, the samples were separated by SDS-polyacrylamide gel electrophoresis (7 5%) ac-cording to Laemmli's method'7 and transferredto nitrocellulose membranes according to themethod of Towbin et al. 8 Nitrocellulose sheetswere incubated with KBA.62 monoclonal anti-body and the bound antibody was targetedwith rabbit anti-mouse and swine anti-rabbitperoxidase conjugates. The peroxidase activitywas revealed using the enhanced chemi-luminescence western blotting detection sys-tem (Amersham, Little Chalfont, UK).

ResultsHYBRIDOMA PRODUCTION AND SELECTION OFKBA.62Supematants from six of 924 hybridomas re-acted with frozen KAL tumour sections butwere negative on human tonsil sections, sug-gesting that a melanoma associated antigen hadbeen detected. Of these six, the KBA.62 clone(IgG2a subclass) was selected for further de-tailed analysis because of its reactivity on bothfrozen and paraffin wax sections. Immuno-blotting analysis performed with KBA.62 pro-duced three bands with apparent molecularweights of 140, 135 and 128 kD and two weakbands of 88 and 73 kD (fig 1).

the staining intensity observed with each anti-body varied from case to case (figs 2E and 2F).Furthermore, some cases were positive withKBA.62 and negative with HMB-45, and viceversa. As expected, virtually all melanocyticproliferations were positive on staining withanti-S 100.

NON-MELANOCYTIC TUMOURSUnexpectedly, KBA.62 reacted with paraffinwax sections of non-melanocytic tumours. Theresults of testing 182 tumour sections are pre-sented in table 2. KBA.62 reacted with someof the carcinomas, of which squamous cellcarcinoma of the skin (five of six cases) and ofthe lung (two of seven cases) predominated.However, staining was mainly present in welldifferentiated keratinising forms and in contrastto melanoma, positive cells were mainly foundat the periphery of malignant lobules.

4.

REACTIVITY OF KBA.62 WITH NEOPLASTICHUMAN TISSUEAs mentioned earlier, paraffin wax sections ofdifferent human tumours from our tissue bank(tables 1 and 2) were investigated and theresults of KBA.62 staining of melanocytic pro-liferations compared with those obtained withHMB-45 and anti-S 100 antibodies.

Melanocytic tumoursThe reactivities ofKBA.62 antibody on paraffinwax embedded tissue sections from patientswith melanocytic tumours are shown intable 1. KBA.62 was positive with the majorityof benign and malignant melanocytic pro-liferations irrespective of their histological type.In most cases, the staining was strong, pre-dominantly in the cell membrane, and involvedall neoplastic cells (fig 2A). Epithelioid melan-oma usually showed stronger and more con-sistent staining than the sarcomatous, spindlecell variant.

Overall, the reactivities of KBA.62 andHMB-45 on melanocytic tumours were com-parable, with the notable exception of thedermal component of the intradermal type andcompound naevi (figs 2C and 2D). In the latterthe intradermal cells were consistently labelledwith KBA.62 (fig 2C) and were weakly positiveor negative with HMB-45 (fig 2D). In addition,

Figure I Western blot analysis of the KAL cell lineimmunostained with KBA. 62 (A) and DBB.42 (B)anti-B monoclonal antibody, which was used as a control.KBA. 62 reacts with three bands of 140, 135 and 128 kDand two weak bands of 88 and 73 kD (A).

828

A novel anti-melanoma antibody reactive on paraffin wax sections 829

..:~~~~~~~3

4*.Xj;. 3.

B~~~~~~~~~~~~~~~~~~~A -

4*4~~~ ~ ~ ~ ~ ~ ~ ~~~4

V~~~~~4.~ ~ Z

La / *^

4-

.4It - ' 31.;~~~ ;s;rf

Cr -:tt3Xs^.*¢^. - ; ;.3 hr

~~~~~~~~~~~~~~~~~' 3

a:, 3t ,.-w 'e ''*:

;.:.. t * .; 4..r4* * +

Figure 2 Lymph node metastasis from malignant melanoma immunostained with KBA. 62 (A) and HMB-45(B). Note the intense membrane labelling of virtually all cells with KBA. 62 monoclonal antibody whereas a diffusecytoplasmic labelling pattern occurs with HMB-45. Compound nevus stained with KBA. 62 (C) and HMB-45(D). KBA. 62 reacts with the entire population of intradermal and junctional nevus cells (C) whereas only junctionalnevus cells and scattered cells in the superficial intradermal component are stained with HMB-45 (D). Note that thebasal cells of the skin also react with KBA. 62 (C). Superficial spreading melanoma stained with KBA. 62 (E) andHMB-45 (F). Ascending intraepidermal and subepidermal tumour cells are strongly labelled with KBA. 62 (E).Staining with HMB-45 (F) is weaker and only a proportion of malignant cells are stained. (Immunoperoxidasetechnique with nuclear counterstaining.)

Cohen-Knafo, Al Saati, Aziza, Ralfkiaer, Selves, Gorguet, et al

REACTIVITY OF KBA.62 WITH NORMAL HUMANTISSUEStaining with KBA.62 was negative with vir-tually all human normal tissues except for theoccasional positive endothelial cell. However,clear staining of hair follicles was noted onall skin biopsy specimens after pretreating thesections in a microwave oven. In contrast tonormal skin, we noted staining of basal cells ofthe uninvolved skin localised above intradermalor compound naevi (fig 2C). Basal melanocytesin the normal skin were negative.

DiscussionThe in situ immunological identification ofhuman malignant proliferations has beengreatly facilitated by the availability of mono-clonal antibodies directed against fixativeresistant molecules." 1920 The KBA.62 mono-clonal antibody was obtained using the KALmelanoma cell line, underlining the value ofnew cell lines in generating novel monoclonalantibodies. "12' The potential diagnostic valueof this monoclonal antibody stems from thefact that it reacts with a melanocyte associatedantigen and can be detected on routinely fixed,paraffin wax embedded tissues. Staining withKBA.62 was positive with all tumours con-taining naevi irrespective of their histo-pathological type (table 1). Approximately 87%of malignant melanomas tested in this studywere positive. KBA.62 reacted with few non-melanocytic neoplasms (6%). However, it didreact with a proportion of cells in well differ-entiated squamous cell carcinomas of the skinand, to a lesser extent, of the lung. Thus,although KBA.62 is not completely specificfor melanocytes, it does distinguish melanomafrom poorly differentiated forms of carcinoma,thus resolving a greater practical problem.

Reactivity of KBA.62 was compared withthat of HMB-45 because of its restricted re-activity with melanocytic tumours and becausethe latter can be used on paraffin wax em-bedded tissue sections. The reactivities ofKBA.62 and HMB-45 suggest that the twoantibodies are different. This was demonstratedby immunoblotting analysis which showed thatKBA.62 reacts with 140, 135, 128, 88, and 73kD polypeptide chains which are different fromthose reported for HMB-45: two bands, one10 kD and the other 100 kD.2223 There are alsoother differences. In normal tissues KBA.62, incontrast to HMB-45, reacts with endothelialcells. In melanocytic proliferations KBA.62 re-acts with two components-that is, junctionaland intradermal, oftypical naevi. We and othershave found that the intradermal naevi and thedermal component of compound naevi reactedto varying degrees with HMB-45, but weremostly negative.32425 In malignant melanomas,despite the overall comparable reactivities ofKBA.62 and HMB-45, some cases were pos-itive with KBA.62 and negative with HMB-45,and vice versa. In addition, KBA.62 usuallystains membranes which is strikingly differentfrom the cytoplasmic staining observed withHMB-45 (figs 2A and 2B). However, the pre-cise immunolocalisation of the antigen re-

cognised by KBA.62 on KAL cells, with andwithout membrane permeabilisation, revealedthat the antigen is probably on the inner part ofthe cell membrane (data not shown). KBA.62reacts with some epithelial tumours, which isalso the case with HMB-45.526 The reactivityof melanocytic neoplasms with anti-S100 iswell documented.127 However, anti-S100 anti-bodies are of limited diagnostic value becauseof their broad spectrum of reactivity with non-melanocytic tumours.45 Thus, the latter markerhas to be used in conjunction with other tissuespecific immunomarkers.

In summary, KBA.62 is a new monoclonalantibody suitable for use on routinely processedparaffin wax sections. This antibody, in as-sociation with other antibodies, may prove to bevaluable for a definitive immunomorphologicaldiagnosis of malignant melanoma in surgicalpathology practice.

We would like to thank Professor S M Chittal, MemorialUniversity, St. John's, New Foundland, Canada, for advice inpreparation of the manuscript. We are also grateful to the staffof the Pathological Anatomy Laboratory, in particular to Ms JBoyes, Ms M A Daussion and Mr M March. This work wassupported by a grant from Conseil Regional de la Region MidiPyrenees and Ligue National Contre le Cancer (Prix de Biarritz).

1 Gatter KC, Pulford KAF, Vanstapel MJ, Roach B, MortimerP, Woolston RE, et al. An immunohistochemical study ofbenign and malignant skin tumours: epithelial aspects.Histopathology 1984;8:209-27.

2 Vennegoor C, Calafat J, Hageman PH, Van Buitenen F.Janssen H, Kolk A, et al. Biochemical characterizationand cellular localization of a formalin-resistant melanoma-associated antigen reacting with monoclonal antibodyNKI/C-3. Int3 Cancer 1985;35:278-95.

3 Gown AM, Hoak D, Gough F, McNutt MA. Monoclonalantibodies specific for melanocytic tumors distiguish sub-populations of melanocytes. Am Jf Pathol 1986;123:195-203.

4 Nakajima T, Kameya T, Watanabe S, Hirota T, Sato Y,Shimosato Y An immunoperoxidase study of S-100 pro-tein distribution in normal and neoplastic tissues. AmJSug Pathol 1982;6:715-27.

5 Herrera GA, Turbat-Herrera EA, Lott RL. S-100 proteinexpression by primary and metastatic adenocarcinomas.AmJ Clin Pathol 1988;89:168-76.

6 Van Duinen S, Buiter DJ, Hageman P, Vennegoor C,Dickersin R, Scheffer E, et al. Immunohistochemical andhistochemical tools in the diagnosis of amelanotic melan-oma. Cancer 1984;53:1566-73.

7 Makie RM, Campell I, Turbitt M. Use of NK1 C3 mono-clonal antibody in the assessment of benign and malignantmelanocytic lesions. J7 Clin Pathol 1984;37:367-72.

8 Herrera GA, Hancock C. Specificity of antibody HMB-45.Arch Pathol Lab Med 1992;116:900-1.

9 Al Saati T, Blancher A, Calvas P, Neulat-Duga I, Delsol G.Production of monoclonal antibodies using spleen cellsfrom nude mice bearing human tumors. Ann Pathol 1987;7:1-8.

10 Kohler G, Milstein C. Derivation of specific antibody pro-ducing tissue culture and tumor lines by cell fusion. EurJF Immunol 1976;6:511-19.

11 Al Saati T, Caspar S, Brousset P, Chittal S, CaveriviereP, Hounieu H, et al. Production of Anti-B monoclonalantibodies (DBA.44, DBB.42, DNA.7, and DND.53)reactive on paraffin embedded tissues with a new B-celllymphoma cell line grafted into athymic nude mice. Blood1989;74:2476-85.

12 Al Saati T, Clamens S, Cohen-Knafo E, Faye JC, Prats H,Coindre JM, et al. Production of monoclonal antibodiesto human estrogen receptor protein (ER) using re-combinant ER (RER). Int Cancer 1993;55:651-4.

13 Lauritzen AF, Hou-Jensen K, Ralfkiaer E. P53 proteinexpression in Hodgkin's disease. APMIS 1993;101:689-94.

14 Shi SR, Key ME, Kalra KL. Antigen retrieval in formalinfixed, paraffin-embedded tissues: An enhancementmethod for immunohistochemical staining based on mi-crowave oven heating of tissue sections. J Histochem Cyto-chem 1991;39:741-8.

15 Laurent G, Delsol G, Reyes F, Abbal M, Mihaesco E.Detection of J chain in lymphomas and related disorders.Clin Exp Immunol 1981;44:620-8.

16 Battifora H. Methods in laboratory investigation. The mul-titumor (sausage) tissue block. Novel method for im-munohistochemical testing. Lab Invest 1986;55:244-8.

17 Laemmli UK. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature 1970;227:680-5.

18 Towbin H, Staehelin T, Gordon J. Electrophoretical transferof proteins from polyacrylamide gels to nitrocellulose

830

A novel anti-melanoma antibody reactive on paraffin wax sections

sheets: procedure and some applications. Proc Natd AcadSci USA 1979;76:4350-4.

19 Gatter KC, Ralfkiaer E, Skinner J, Brown D, Heryet A,Pulford KAF, et al. An immunocytochemical study ofmalignant melanoma and its differential diagnosis fromother malignant tumours. Clin Pathol 1985;38:1353-7.

20 Mason DY, Gatter KC. The role of immunocytochemistryin diagnostic pathology. Clin Pathol 1987;40:1042-54.

21 Delsol G, Blancher A, Al Saati T, Ralfitaer E, Lauritzen A,Brugieres L, et al. Expression of blood group-related Hand Y determinants by anaplastic large cell lymphoma asrecognized by a new monoclonal antibody, BNH.9. BrCancer 1991;64:321-6.

22 Esclamado RM, Gown AM, Vogel AM. Unique proteinsdefined by monoclonal antibodies specific for humanmelanoma. Am J7 Surg 1986;152:376-85.

23 Adema GJ, Boer AJ, Van't Hullenaar R, Denijn M, Ruiter

DJ, Vogel AM, et al. Melanocyte lineage-specific antigensrecognized by monoclonal antibodies NKI-beteb, HMB-50, and HMB-45 are encoded by a single cDNA. AmPathol 1993;143:1579-85.

24 Palazzo J, Duray P. Typical, dysplastic, congenital, and Spitznevi: a comparative immunohistochemical study. HumPathol 1989;20:341-6.

25 Wood WS, Tron VA. Analysis ofHMB-45 immunoreactivityin common and cellular blue nevi. Jf Cutan Pathol 1991;18:261-3.

26 Yates AJ, Banerjee SS, Bishop PW, Graham KE. HMB-45 in nonmelanocytic tumours. Histopathology 1993;23:477-8.

27 Nakajima T, Watanabe S, Sato Y, Kameya T, Shimosato Y,Ishihara K. Immunohistochemical demonstration of S 100protein in malignant melanoma and pigmented nevus,and its diagnostic application. Cancer 1982;50:912-18.

831