Localisation Ca HMFG2 breast - jcp.bmj.com · Freshly obtained breast tissue from a case of infiltrating duct carcinoma, known to express both Ca and HMFG2antigens, was fixed with

J Clin Pathol 1985;38:512-520

Localisation of Ca and HMFG2 antigens in breasttissue by immunoperoxidase, immunofluorescence,and immunoelectron microscopyURSZULA BECKFORD, CALYPSO BARBATIS,* JE BEESLEY, JANE C LINSELL,tSHIREEN M CHANTLER

From the Wellcome Research Laboratories, Langley Court, Beckenham, Kent, and the Departments of*Pathology and tSurgery, Lewisham Hospital, London SE13

SUMMARY The reactivities of Cal and HMFG2 monoclonal antibodies were compared onparaffin wax embedded breast tissues using indirect immunoperoxidase. The expression of Caantigen, like HMFG2, is not exclusive to malignancy: Ca was present in 41/53 (77%) andHMFG2 in 42/53 (79.2%) non-malignant conditions and both were present in 33/35 (94%)carcinomas. Similar results were obtained when cryostat sections were used. Both antigensshowed striking similarities in their topographical distributions, although quantitative differenceswere seen. Their cellular and sub-cellular localisations were investigated by double labellingimmunofluorescence and immunogold electron microscopy, which showed that the expression ofCa and HMFG2 antigens was closely associated on cell membranes but that the epitopes weredistinct.

Monoclonal Cal antibody, which defines an antigenCa,' has been reported to react selectively withmalignant lesions taken from a variety of tissues.2Subsequent studies have shown a significant degreeof reactivity with non-malignant tissue.3-" Severalmonoclonal antibodies directed against epithelialmembrane antigens react with breast tissue.'2 -' Thestaining pattern reported for one of these, thehuman milk fat globule 2 (HMFG2) antibody,'2appeared similar to our observations with Cal in apreliminary study of breast tissue.3 This study wasundertaken to compare the occurrence, distribution,and staining patterns of Ca and HMFG2 antigens onformalin fixed, paraffin wax embedded and frozentissues derived from malignant and benign lesions ofthe breast and to determine their cellular associationby double label immunofluorescence and immuno-gold electron microscopy.

Material and methods

TISSUESFormalin fixed, paraffin wax embedded breast tis-sues obtained from 4 normal, 49 benign, and 35

Accepted for publication 8 January 1985

malignant lesions were used in the retrospectivestudy. Fresh tissues from 20 benign and 24 malig-nant breast lesions obtained at operation and snapfrozen in liquid nitrogen were used for the prospec-tive study.

METHODSImmunoperoxidaseFormalin fixed, paraffin wax embedded sections(6 um) were dewaxed and rehydrated to waterbefore examination by a modification of the indirectimmunoperoxidase procedure (Table 1).2 Mono-clonal antibodies Cal (IgM, culture fluid, WellcomeDiagnostics Ltd, reconstituted with 1-0 ml of distil-led water) and HMFG2 (IgGl, culture fluid, SewardLaboratories) were diluted 1/10 and 1/5 respecfivelyin phosphate buffered saline (PBS) and 10% bovineserum albumin (BSA) (Miles Laboratories) for use.As controls, monoclonal antibodies R4/23 (IgM,culture fluid, reactive against human dendriticreticulum cells)'5 and WMB 36 (IgM, culture fluid,reactive with a meningococcal antigen)'6 were usedundiluted.

ImmunofluorescenceCryostat sections (6 p.m) were air dried, wrapped in

512

on July 1, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.38.5.512 on 1 M

ay 1985. Dow

nloaded from

http://jcp.bmj.com/

Ca and HMFG2 antigens in breast tissueTable 1 Immunoperoxidase procedure

Dewaxed, rehydrated, formalin fixed, paraffin wax embeddedsections10/1 (vol/vol) methanol: H,02 (30%), 30 min, room temperature(blocks endogenous peroxidase activity)Incubate with monoclonal antibody overnight, 4'CIncubate with conjugate, 30 min, room temperatureIncubate with 500 fig/ml 3,3' diaminobenzidine hydrochlorideSigma) containing 0-015% H 0 in phosphate buffered salinePBS), pH 7-5, 5 min, room temperatureCounterstain with Mayer s haemalum, dehydrate, clear, and mountin DPXSections were thoroughly washed in PBS (3 x 5 min) between eachstep

cellophane, and stored at -20°C. Before use theywere brought to room temperature, unwrapped,fixed in methanol (96%) for 30 s, and air dried.After a wash in PBS (5 min), the sections were incu-bated with the appropriate monoclonal antibodies(overnight, 4°C), washed with PBS (3 x 5 min), andincubated with fluorescein isothiocyanate (FITC)labelled rabbit antimouse Ig (1/20 in PBS, 30 min atroom temperature, prepared in this laboratory).After a final wash in PBS (3 x 5 min), the tissueswere mounted in PBS buffered glycerol containing0-1% p-phenylenediamine'7 before examination byfluorescence microscopy.

Double labelling immunofluorescenceCryostat sections, treated as described above, wereincubated with a mixture of Cal and HMFG2monoclonal antibodies (1/1 vol/vol) overnight at4°C. After a wash in PBS (3 x 5 min), sections werefirst incubated with tetramethylrhodamine (TRITC)labelled rabbit antimouse IgG (1/5 in PBS, 30 min atroom temperature, Cappel Laboratories), washed inPBS (3 x 5 min) and then incubated with FITClabelled sheep antimouse IgM (1/20; in PBS, 30 minat room temperature, prepared in this laboratory).The TRITC conjugate also reacted with mouse IgMand was rendered IgG specific by preabsorption withmouse IgM. After a final wash in PBS (3 x 5 min)the tissues were mounted as described above andexamined by a fluorescence microscope equippedwith incident light illumination: excitation filter BP450-490 barrier filter BP 520-560 for FITC; andexcitation filter BP 546/12 barrier filter LP 590 forTRITC.

Double labelling immunogoldFreshly obtained breast tissue from a case ofinfiltrating duct carcinoma, known to express bothCa and HMFG2 antigens, was fixed with 1% for-maldehyde in 0-1 M cacodylate buffer (1 h at roomtemperature), dehydrated in a graded series ofethanol, and embedded in LR white resin (London

513

Resin Company, Basingstoke). As a primary screen,areas of the tissue reactive with Cal and HMFG2antibodies were first identified by light microscopy.All reagents were diluted with 0-5 M Tris buffer, pH7*4, containing 1% (vol/vol) Tween 20, 1% (wtlvol)ovalbumin, and 0-1% gelatin.'8 Sequential semithin(2 ,um) sections were incubated with Cal (1/10) orHMFG2 (1/10) overnight at 4°C, washed in buffer(2 min), incubated with rabbit antimouse Ig (1/10,30 min at room temperature), washed again in buf-fer (2 min), and finally incubated for 1 h at roomtemperature with an undiluted suspension of 40 nmgold spheres coated with goat antirabbit IgG (Jans-sen Pharmaceutica). The sections were washed withwater (2 min), counterstained with 1% toluidineblue, and examined by light microscopy.

Ultrathin sections (90-100 nm thick) of areasreactive with both Cal and HMFG2 were preparedfor subsequent double label immunoelectron mic-roscopy. One face of the section was incubated withCal monoclonal antibody (1/10 overnight, 4°C) fol-lowed by rabbit antimouse Ig (1/10, 30 min at roomtemperature), and lastly for 1 h with the 20 nm goatantirabbit IgG gold probe (1/20, Janssen Phar-maceutica). The grids were floated on drops of eachreagent and sections were washed with buffer (5 x1.5 min) between all incubations. Care was taken toavoid wetting the reverse side of the section. After afinal wash in water the sections were air dried andcoated with a protective film of carbon and, finally,with a supporting layer of celloidin. The oppositeface of the section was then similarly treated withHMFG2 monoclonal antibody (1/10), rabbit anti-mouse Ig, and finally the 5 nm antirabbit gold probe(1/50, Janssen Pharmaceutica). The sections werecontrasted with osmium tetroxide vapour beforeexamination by electron microscopy. The labellingwas then repeated using the 5 nm gold for Caldetection and the 20 nm gold for HMFG2 detectionto avoid the possibility of labelling artefacts.'9

Results

Localisation ofCa and HMFG2 antigens in formalinfixed, paraffin embedded and frozen breast tissueThe distributions of Ca and HMFG2 antigens innormal and benign breast tissue are shown in Tables2 and 3. Of the 53 tissues studied, 41 (77%) werestained by Cal and 42 (79.2%) by HMFG2.Wherever the two antigens were coexpressed thetissue distribution was similar, although the stainingintensity was variable. Ca and HMFG2 antigenswere not detected in the mammary epithelium oftwo normal breasts. The lactating breast expressedHMFG2 strongly on the luminal surface of the aciniand in the luminal secretion. Ca antigen was present


http://jcp.bmj.com

/J C


ay 1985. Dow

nloaded from

http://jcp.bmj.com/

Table 2 Distribution of Ca and HMFG2 antigens in normal and benign paraffin embedded breast tissue

Condition No Cal HMFG2

Ducts Acini Ducts Acini

LS Cytoplasm LS Cytoplasm LS Cytoplasm LS Cytoplasm

Normal 2 - - - - - - - -Lactating 2 - _ F(1)+ - _ _ F(1)++ -

D(l)++Papilloma 2 F(2) Luminal surtace of papillae or ducts Similar to CalFibroadenoma 17 F(9) + F(1) ++ - - F(10) + + F(1) +

D(1)++Total no tissues = 23 No positive with Cal = 13 (56.5%) No positive with HMFG2 = 14 (60.9%)

LS = luminal surface of the epithelium.F (focal) = positive scattered ducts or acini <10% of the total number.D (diffuse) = positive ducts or acini >50% of the total number.+-+ + + = intensity of staining.(%) = % tissues stained.

Table 3 Distribution of Ca and HMFG2 antigens in sections ofparaffin embedded fibrocystic disease ofthe breast


Epithelium Myoepithelial cells Epithelium Myoepithelial cells

LS Cytoplasm Cytoplasm LS Cytoplasm Cytoplasm

Aprocrinemetaplasia 12 D(12)++-+++ F(4)++ - D(12)++-+++ F(4) +

Adenosis 18 F(17)+-+++ F(2)++ F(3) ++ F(16)+-++ F(2)+ F(3)++Epitheliosis 12 F( 1) + +-+++ - - F( 11) +-+++ -

D(1) D(1)Ducts 30 F(28) +-+ + - - F(28) +-+ + - -Cysts 30 F(28)+-++ - - F(28)+-++ - -Total no tissues = 30 No positive with Cal = 28 (93%) No positive with HMFG2 = 28 (93%)

LS = luminal epithelial surface.F (focal) = scattered positive ductules.D (diffuse) = almost all areas.+-+++ = intensity of staining.

in smaller amounts and with similar distribution toHMFG2. Inter- and intralobular ducts were nega-tive for both antigens.We noted a similar distribution of both antigens in

10/17 fibroadenomas. The expression was invariablyfocal and random, but in one case almost all ductsshowed positive epithelium. In nine tissues the anti-gens were localised on the luminal surface (Fig. la)and in one strong intracytoplasmic staining wasdetected (Fig. lb). In papillomas focal surfacepositivity of the epithelium was noted.The expression of both antigens in 28/30 cases of

fibrocystic disease was either focal or diffuse(Table 3). All cases of apocrine metaplasia showedstrong staining of the entire luminal surface andsecretion and in 4/12 strong focal intracytoplasmicgranular staining was present (Fig. 2). The ductulescomprising foci of adenosis were also positive for Caand HMFG2 antigens, but the number of positiveductules varied from case to case. In two casesstrong intracytoplasmic staining was noted. Bothantigens were expressed, mainly on the surface ofthe epithelium, in areas of epitheliosis.An interesting finding was the identification of

both antigens in the cytoplasm of myoepithelial cellsin three cases of fibrocystic disease. In these areascytoplasmic staining of the adjacent epithelium wasalso noted; hence it is difficult to distinguishbetween expression or absorption.

In no case was staining seen with control mono-clonal antibodies of unrelated specificity and Calstaining was not abolished by pretreating thesections with trypsin or preincubating with proteinrich buffers.2

Table 4 shows the distributions of Ca andHMFG2 antigens in breast carcinomas. The same 33of 35 (94.3%) tissues were stained by both Cal andHMFG2. Most ductal infiltrating carcinomas (18/24), regardless of histological grade or staging,showed surface and cytoplasmic staining of themalignant cells in more than half of the total tumourmass, although the density varied from site to site(Fig. 3). There was striking topographical similarityin the distribution pattern of both antigens withslight variations in intensity. In two tubular car-cinomas both antigens were expressed on the lumi-nal surface of ducts and tubules. There was a distinctquantitative difference in the expression of Ca and

514 Beckford, Barbalis, Beesley, Linsell, Chantler


http://jcp.bmj.com

/J C


ay 1985. Dow

nloaded from

http://jcp.bmj.com/

Ca and HMFG2 antigens in breast tissue

I/7~ .1

f00

Fig. 1 Fibroadenoma (paraffin section) stained with Cal and HRP antimouse Ig showing (a) luminal duct epithelialborder (x80) and (b) intracytoplasmic staining (x300).

HMFG2 antigens in infiltrating lobular carcinoma.HMFG2 antigen was strongly expressed in mostmalignant cells in 4/5 cases; even the intracyto-plasmic lumina were surrounded by a positive rim.By contrast, Ca antigen was expressed in occasionalscattered cells in the four positive cases.

In order to determine whether tissue processinginfluenced the results, the reactivity and stainingpatterns of Ca and HMFG2 antigens were studiedon cryostat sections from freshly collected materialusing immunofluorescence (Table 5). Immuno-fluorescence was selected because immuno-peroxidase gave weak diffuse staining on cryostatsections irrespective of the level of monoclonal anti-body used. There was no detectable difference in theincidence of reactivity with both antibodies whencompared with the formalin fixed material.

Simultaneous detection of Ca and HMFG2 antigensby the double labelled immunofluorescence andimmunogold proceduresCa and HMFG2 antigens had close topographicaldistributions in all tissues examined. In order todetermine the relative cellular localisation withgreater precision, simultaneous staining with Caland HMFG2 antibody was performed on singlecryostat sections using immunofluorescence and onsingle ultrathin sections by the immunogold proce-dure.When separate sections were treated with either

Cal (IgM) or HMFG2 (IgGl) antibody followedsequentially by TRITC antimouse IgG and FITCantimouse IgM, fluorescent staining was observedonly with the homologous conjugate. This confirmsthe specificity of the conjugates and the suitability ofthe microscope filter combinations used. When a

515

Oii.

.11


http://jcp.bmj.com

/J C


ay 1985. Dow

nloaded from

http://jcp.bmj.com/

I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. XI

'00 ~ *'

**~~~~~~~~~~~ ~Fig. 2 Apocrinertnetap)lasia duct1epeinm (parajii

* ~~~~~~~~~~~~~~~controlmnonoclonal atiiibodyv and HRP afthiinoiiAfimminunoglobulint showing staininig oft/ic lIuninal!burl

et. ~~~~~~~~~andin thie cytoplasm. x300.* ~~~~~~

PS

/~~~~~~~~~~~~~~~~~~~~~~~~AA

%

...e........

4~~~~~~

0~~~~~~~~~~~~~~~~~~4


http://jcp.bmj.com

/J C


ay 1985. Dow

nloaded from

http://jcp.bmj.com/

Table 4 Distribution ofCa and HMFG2 antigens in paraffin embedded sections ofbreast carcinoma


Epithelium Epithelium

Surface Cytoplasm Surface Cytoplasm

Ductal 24 F(5)+-+ ++ F(10)++ F(4)++ F( 0)+-+ +D(18)++-+++ D(19)++-+++

Tubular 2 F(1) +++ - -D(1)+++ - D(2)+++

Lobular 5 F(4) occasional scattered cells (+-+ +) F(2)++ F(2)++D(2)+++ D(2)+++

Medullary 3 F(3)+ F(2)+ F(3)++ F(2)++Mucoid 1 - - -Total no tissues = 35 No positive with Cal = 33 (94-3%) No positive with HMFG2 .= 33 (94-3%)

F = luminal surface of scattered ducts (<5% of total number).D = luminal surface or cytoplasm of malignant cells and ducts >50% of the total number.

mixture of the antibodies was applied to a singlesection, both antigens were detected in all eightmalignant and five benign tissues studied. In themajority, both labels were detected in the samegroups of cells, confirming the similar topographicaldistribution of Ca and HMFG2 antigens on a single

Fig. 3 Lobular carcinoma (paraffin secton) ofthe breaststained with Cal and HRP antimouse immunoglobulinshowing cytoplasmic staining and heterogeneity ofantigenexpression. x280.

preparation (Fig. 4). Preincubation of the sectionwith an excess of Cal or of HMFG2 antibody didnot inhibit binding of the other antibody, whichconfirms that the double labelling seen was not dueto sub-optimal levels of antibody in the mixture.The relative sub-cellular localisation of Ca and

HMFG2 antigens in selected ultrathin sections of aninfiltrating duct carcinoma was examined by doublelabelling immunogold electron microscopy. Experi-ments, in which Cal antibody was labelled with the20 nm gold probe and HMFG2 with the 5 nm goldprobe, showed that Ca and HMFG2 were closelyassociated on the membranes of the same cells(Fig. Sa), although HMFG2 was occasionally presentin the absence of Ca (Fig. Sb). The same result wasobtained when the probes were reversed, therebyexcluding the possibility of artefacts due to probesize.

Discussion

This study confirms earlier observations thatimmunohistochemical staining with Cal on tissuesections fails to discriminate between malignant andbenign lesions of the breast.34" The use of freshmaterial failed to improve this differentiation, which

Table 5 Occurrence of Ca and HMFG2 antigens incryostat sections ofbreast tissue detected by indirectimmunofiuorescenceCondition No Reactivity with

Cal HMFG2

Benign lesionsFibrocystic disease 9 7 8Fibroadenoma 1 1 5 7Total 20 12 15% tissues stained 60 75

Malignant lesionsInfiltrating duct carcinoma 24 24 24% tissues stained 100 100

Ca and HMFG2 antigens in breast tissue 517


http://jcp.bmj.com

/J C


ay 1985. Dow

nloaded from

http://jcp.bmj.com/

Beckford, Barbatis, Beesley, Linsell, Chantler

Fig. 4 Fibrocystic disease (cryostat section) double labelled with (a) FITC antimouse IgM (Cal) and (b) TRITCantimouse IgG (HMFG2). Note the occurrence ofstaining in similar groups of cells with both antibodies. x160.

is in agreement with results reported on fine needlebreast aspirates.6 The similarity between resultsobtained with formalin flxed and fresh frozen tissuesexcludes the possibility that poor specificity of Calfor malignant cells is due to artefacts resulting fromcomponent polymerisation during fixation.Although breast carcinomas invariably expressed

intracytoplasmic staining with both antibodies,intracytoplasmic reactivity was noted in the ductalepithelium of fibroadenomas or fibrocystic disease,albeit infrequently on paraffin embedded tissue. Ithas been suggested that intracytoplasmic staining ofbenign epithelium may be associated with lesionswhich progress to malignancy,20 but only retrospec-tive follow up studies will provide conclusive data.Our comparative immunohistochemical studies

with Cal and HMFG2 antibodies on sequential sec-tions showed a striking similarity in the topographi-cal distribution of both antigens, although quantita-tive differences were noted. This similar localisationsuggests that the expression of epitopes recognisedby Cal and HMFG2 antibodies is intimately associ-ated. Synchronous expression of Ca and HMFG2

antigens is not, however, always present.Physicochemical and immunohistochemical

studies indicate several-similarities between Ca andHMFG2 antigens.2' 22 Both are tumour associatedmembrane glycoprotein antigens, which are alsoexpressed in normal, hyperplastic, or benignepithelium.3 1112 Cal and HMFG2 antibodies recog-nise determinants on high molecular weight compo-nents derived from mammary carcinoma cell lines.The antibody binding is blocked by wheat germ lec-tin, which reacts with j3-(1,4)-N-acetylglucosamineand sialic acid residues. The determinants recog-nised by HMFG2 antibody exhibit a greater diver-sity of molecular size (80-400K) than thosedefined by Cal (350-390K), but this difference maybe a reflection of the analytical methods used.2' 22 Inview of the physicochemical similarities and thealmost identical distribution described above, weexamined the relation of the two antigens at lightmicroscopy and ultrastructural level.We have shown that both antigens are similarly

distributed, although they may not be simultane-ously expressed. Double labelling immunofluores-

518


http://jcp.bmj.com

/J C


ay 1985. Dow

nloaded from

http://jcp.bmj.com/

Ca and HMFG2 antigens in breast tissue

cence showed that both antigens were present on thesame group of cells on the luminal surface as well asin the cytoplasm of mammary epithelium. The

41,

~~~~~~~~~~~~¼e

I.4./*

519

finding that both antigens can be detected simul-taneously on the same section and that Cal andHMFG2 binding was not inhibited by preincubation

..

,

*. ,i

I

..

'i

.

.tj,

a.

. _9:'*7

4W.

* iR

*8

*.w

IsV:.

,S;

.1t1.*c.

*fr'

4.S

di.r:..

.-7 1

Fig. 5 Infitrating duct carcinoma double stained by the immunogold procedureusing (a) 20 nm gold probe for Ca and 5 nm for HMFG2 (b) 5 nm gold probe forCa and 20 nm for HMFG2. Note the cellular coexistence of Ca and HMFG2 in (a)and HMFG2 in the absence of Ca in (b).

.t f..,

st,

.4

.411

;...

.-

;9-A. ..,k; f..

.......".00.rl -1-,-

:-, 4;VIN-I'll. i. 1,?: AL'.1 .ik.-, 0

4 'I,.1.rb

. %. r .",f.... .:..

A. f

.AL.VP.,


http://jcp.bmj.com

/J C


ay 1985. Dow

nloaded from

http://jcp.bmj.com/

520

with an excess of the heterologous antibodyindicates that the antigenic epitopes, althoughclosely situated, are not identical and that they aretopographically sufficiently distant to prevent inhibi-tion due to configurational effects.

These conclusions were confirmed and extendedto the cellular level by the simultaneous stainingwith immunogold preparations of defined diameterand examination by electron microscopy. Thesedouble staining results for the first time provide con-clusive evidence that Ca and HMFG2 antigens areclosely, but not exclusively, associated at the cellularlevel, topographically separate, and retain distinctepitope specificity. Their expression may be relatedto the functional activity of individual cells and thisphenomenon needs further investigation.

We thank Dr D Mason (Nuffield Department ofPathology, Oxford) for R4/23; Dr M Mclllmurray(Wellcome Research Laboratories, Beckenham,Kent) for WMB 36; Dr A Price and Professor GSlavin (Northwick Park Hospital) and Dr HW Simp-son (Royal Infirmary, Glasgow) for tissues; thePathology Department (Wellcome ResearchLaboratories) for cutting of sections; and Mrs PThurbin for typing the manuscript.

References

'Ashall F, Bramwell ME, Harris H. A new marker for humancancer cells. 1. The Ca antigen and the Cal antibody. Lancet1982;ii: 1-6.

2 McGee J O' D, Woods JC, Ashall F, Bramwell ME, Harris H. Anew marker for human cancer cells. 2. Immunohistochemicaldetection of the Ca antigen in human tissues with the Calantibody. Lancet 1982;ii:7-10.

3Beckford U, Chantler S. Immunoperoxidase localisation of Caantigen in mammary tumours. In: Peters H ed. Protides ofthebiological fluids Vol 31. Oxford: Pergamon Press, 1984:559-62.

4Clough DGF, Coghill GR, Holley MP. Evaluation of the Calantibody in the diagnosis of invasive breast cancer. J ClinPathol 1984;37: 10-3.

5 Ferguson AM, Fox H. A study of the Ca antigen in epithelialtumours of the ovary. J Clin Pathol 1984;37:6-9.

6 Ghosh AK, Spriggs AI, Boddington MM, Ross HB. Cal anti-body applied to cytological diagnosis of fine needle aspiratesfrom breast lumps. Lancet 1983;ii:287.

Ghosh AK, Spriggs Al, Taylor-Papadimitriou J, Mason DY.Immunocytochemical staining of cells in pleural and peritonealeffusions with a panel of monoclonal antibodies. J Clin Pathol1983;36: 1154-64.

Jha RS, Wickenden C, Anderson MC, Coleman DV. Monoclonal

Beckford, Barbatis, Beesley, Linsell, Chantlerantibodies for the histopathological diagnosis of cervical neo-plasia. Br J Obstet Gynaecol 1984;37:483-8.

Lloyd JM, O'Dowd T, Driver M, Tee DEH. Immunohistochemi-cal detection of Ca antigen in normal, dysplastic and neoplasticsquamous epithelia of the human uterine cervix. J Clin Pathol1984;37: 14-9.

Paradinas FJ, Boxer G, Bagshawe KD. Distribution of the Ca(Oxford) antigen in lung neoplasms and non-neoplastic lungtissues. J Clin Pathol 1984;37: 1-5.

"Simpson HW, Candlish W, Liddle C, McGregor M, Mutch F,Tinkler B. A critical investigation of the Oxford tumourmarker Cal in the histological diagnosis of breast cancer andpre-cancer. Histopathology 1984;8:481-99.

12 Arklie J, Taylor-Papadimitriou J, Bodmer W, Egan M, Millis R.Differentiation antigens expressed by epithelial cells in thelactating breast are also detectable in breast cancers. Int JCancer 1981;28:23-9.

3 Foster CS, Edwards PAW, Dinsdale EA, Neville AM. Mono-clonal antibodies to the human mammary gland. I. Distribu-tion of determinants in non-neoplastic mammary andextramammary tissues. Virchows Arch (Pathol Anat)1982;394: 279-93.

4 Foster CS, Dinsdale EA, Edwards PAW, Neville AM. Mono-clonal antibodies to the human mammary gland. II. Distribu-tion of determinants in breast carcinomas. Virchows Arch(Pathol Anat) 1982;394:295-305.

"Naiem M, Gerdes J, Abdulaziz Z, Stein H, Mason DY. Produc-tion of a monoclonal antibody reactive with human dendriticreticulum cells and its use in the immunohistological analysisof lymphoid tissue. J Clin Pathol 1983;36: 167-75.

6 Chalker SJ, Cook GD, Jones DM, et al. Evaluation of monclonalantibodies as diagnostic reagents for Neisseria meningitidis.Medicine Tropicale 1983;43:81-4.

7 Johnson GD, de C Nogueira Araujo GM. A simple method ofreducing the fading of immunofluorescence during micro-scopy. J Immunol Meth 1981; 43:349-50.

Beesley JE, Day SE, Betts MP, Thorley CM. Immunocytochemi-cal labelling of Bacteroides nodosum pil. using an immunogoldtechnique. J Gen Microbiol 1984; 130:1481-7.

"Beesley JE, Beckford U, Chantler SM. Double immunogoldstaining method for the simultaneous localisation of twotumour associated epithelial cell antigens in breast tissue.Proceedings of Eighth European Congress on Electron Mic-roscopy 1984;3:1599-1600.

20 Heryet JC, Fleming KA, Czerniewski I, Turner ELT,McGee J O'D. Detection of Cal positive cells in normal andnon-neoplastic tissue, and in preinvasive and invasive cancer.Proceedings of the Pathological Society, Edinburgh, July1983. Abstract 20:19.

21 Bramwell ME, Bhavanandan VP, Wiseman G, Harris H.Structure and function of the Ca antigen. Br J Cancer.1983;48: 177-83.

22 Burchell J, Durbin H, Taylor-Papadimitriou J. Complexity ofexpression of antigenic determinants, recognised by mono-clonal antibodies HMFG1 and HMFG2, in normal and malig-nant human mammary epithelial cells. J Immunol1983; 131:508-13.

Requests for reprints to: Dr Shireen M Chantler, TheWellcome Research Laboratories, Langley Court,Beckenham, Kent BR3 3BS, England.


http://jcp.bmj.com

/J C


ay 1985. Dow

nloaded from

http://jcp.bmj.com/

Documents

Localisation Ca HMFG2 breast - jcp.bmj.com · Freshly obtained breast tissue from a case of infiltrating duct carcinoma, known to express both Ca and HMFG2antigens, was fixed with