J Clin Pathol 1990;43:239-242

Immunohistological study of oestrogen receptorsin breast carcinomas that are biochemicallyreceptor negative

S Shousha, T Stamp, K R James, J Alaghband-Zadeh

AbstractThe reliability of an immunohistologicalmethod, applied to paraffin wax sections,was assessed for determination of oes-trogen receptor content of biochemicallyoestrogen receptor negative breast car-cinomata. Sixty consecutive tumourswith oestrogen receptor concentrations ofless than 10 fmol/mg cytosol protein, asestimated by dextran-coated charcoalbiochemical assay, were examined.Paraffin wax sections were treated withDNAse before applying a peroxidase-anti-peroxidase method using ER-ICAmonoclonal antibodies. Fifty one cases(85%) were negative, six (10%) weaklypositive, and three (5%) were moderatelypositive. No strongly positive cases wereseen.

It is suggested that cases with weaklypositive staining, especially when local-ised to a small area, should be regarded asnegative. On the other hand, as the threemoderately stained cases included twosmall tubular carcinomas and an invasiveductal carcinoma with high progesteronereceptor concentrations, it is more likelythat the biochemiccal assay in these casesrepresented false negative results due tosampling error or inclusion of fibrous orother non-neoplastic tissue in the assayedsamples. It is concluded that theimmunohistological method used here isfairly reliable and would be especiallyvaluable for determination of oestrogenreceptor content in small, mammo-graphically detected tumours from whichno tissue would be available for bio-chemical assay or frozen section examin-ation.

Department ofHistopathology,Charing Cross andWestminster MedicalSchool, LondonS ShoushaT StampK R JamesDepartment ofChemical PathologyJ Alaghband-ZadehCorrespondence to:Dr S Shousha, Departmentof Histopathology, CharingCross Hospital, FulhamPalace Road, LondonW6 8RF, England.Accepted for publication19 October 1989

The recent introduction of immunohistologicalmethods for the detection of oestrogen recep-tors in paraffin wax sections provides a muchneeded facility for examination of routinelyprocessed histological material. 1-3 Recently, weassessed the use of one of these methods in a

study of 19 cases of mucinous breast carcin-oma, a tumour known to be rich in oestrogenreceptor content. We found the results consis-tent with those obtained by the standard dex-tran-coated charcoal biochemical assay in thecases examined by both methods.4 The currentstudy was aimed at assessing the same immuno-histological method in a large series of breastcarcinomas found to be oestrogen receptornegative by the biochemical assay.

MethodsThe study included 60 consecutive cases ofbreast carcinoma with oestrogen receptor con-centrations of less than 10 fmol/mg cytosolprotein, as estimated by the dextran-coatedcharcoal biochemical assay.5 All tumours wereremoved surgically at Charing Cross Hospitalbetween 1979-1982. The original haematoxy-lin and eosin stained sections were reviewedand a representative section of each case wasselected. Sections 5 ,um thick were cut from thecorresponding formalin fixed, paraffin waxembedded blocks of tumour tissue and used forthe immunohistological study.As previously described,4 the sections were

dewaxed, hydrated through graded alcohols,and rinsed in TRIS buffer (pH 7 6). Endogen-ous peroxidase activity was blocked by using0-50o hydrogen peroxide in methanol for 10minutes. The method described by Shintakuand Said for the demonstration of oestrogenreceptors in paraffin wax sections' was thenapplied. The sections were covered withDNAse I (Sigma D5025) solution in a moistchamber for two hours. The concentration ofthe DNAse solution was decided by titratingagainst a known oestrogen receptor positivecontrol case, using doubling dilutions from40 mg to 5 mg per ml of 0-01M magnesiumsulphate in TRIS buffer. It is essential to dothis for each batch of DNAse as the enzymeactivity may vary considerably.6 In this studythe enzyme activity was 2200 Kunitz units permg solid, and a dilution of20 mg per ml (44 000Kunitz units per ml) was found to be optimal.The sections were then rinsed in TRIS

buffered saline (TBS), covered for 30 minuteswith normal goat serum, followed by overnightincubation in a moist chamber with severaldrops of ER-ICA monoclonal antibody(Abbott Laboratories England). After rinsingin TBS the sections were successively coveredfor two hours each with goat anti-rat immuno-globulin and rat peroxidase-anti-peroxidasecomplex. This was followed by rinsing in TBS,incubation in diaminobenzidine for 10 min-utes, washing in tap water and light counter-staining with haematoxylin. A control slidefrom each case was treated similarly except forreplacement of the specific antiserum with thenegative control antibody supplied with theER-ICA kit.The staining results were assessed

semiquantitatively according to the percentageof cells stained and the intensity of the staining,as previously described.4 A scale of 1-3 wasused for each of these two components. For thepercentage of stained cells, 1 referred to stain-

239

on 2 July 2018 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.43.3.239 on 1 M

arch 1990. Dow

nloaded from

Shousha, Stamp, James, Alaghband-Zadeh

Table I Distribution of histological types of breastcarcinoma with oestrogen receptor (number (°o))concentrations of < 10 fmollmg cytosol protein comparedwith carcinomas with concentrations of > IOfmol

< lOfmol IO fmolHistological type (n = 60) (n = 190)

Invasive ductal 48 (80 0) 146 (76 8)Medullary 7 (11-7) 2(1 1)Tubular 2 (3 3) 2 (1-1)Invasive lobular 1 (1-7) 37 (19 5)Mucinous 1 (1 7) 1 (0 5)Squamous 1 (17) 0Papillary 0 2(1-1)

ing of less than 33%O of tumour cells, 2 for 33660o, and 3 for the presence of more than 66%positively stained cells. For staining intensity,1 indicated weak, 2 moderate, and 3 strongstaining. The two factors were then multipliedby each other and the final result expressed asfollows: - = negative (no staining);+ = weakly positive (for a total score of 1-3);+ + = moderately positive (score 4-6);+ + + = strongly positive (score of > 6). Thetotal score, when mentioned, was followedbetween parentheses by the two numbersrepresenting its components. For example, amoderately positive case with a score of 6(3 x 2) means that more than two thirds of itsneoplastic cells showed oestrogen receptorstaining of a moderate intensity, while a casewith a score of 6 (2 x 3) means that about halfits tumour cells were strongly stained.The significance of the results was assessed

by the fourfold tables ofthe X2test. Probabilitieswere calculated with one degree of freedom.

ResultsThe 60 tumours included 48 (800o) invasiveductal carcinomas, seven (11700) medullary(three typical and four atypical), two (3 3o0)tubular, one (1 700) each of invasive lobular,pure mucinous, and pure squamous cell carcin-oma. Table 1 compares the distribution ofhistological types in these cases with that of 190cases of breast carcinoma with oestrogen recep-tor concentrations of 10 fmol/mg cytosolprotein or more, seen during the same period(1979-1982). The two main differences be-tween the two groups were a significantly muchhigher incidence of medullary carcinoma(p < 0 001) and a low incidence of invasivelobular carcinoma (p < 0 001) in the oestrogenreceptor negative group.The immunohistological brown positive

staining, when present, was restricted to thenuclei of neoplastic cells and some non-neo-

Table 2 Immunohistological oestrogen receptor stainingresults according to histological type of tumour

Histological type - + + + + + +

Invasive ductal (n = 48) 42 5 1 0Medullary (n = 7) 6 1 0 0Tubular (n = 2) 0 0 2 0Invasive lobular (n = 1) 1 0 0 0Mucinous (n = 1) 1 0 0 0Squamous (n = 1) 1 0 0 0Total(n = 60) 51 (85"%) 6(10%o) 3(50o) 0

*1'~~~~~~~S

~~~~~~~I4~~ ~ ~ ~ ~ ~~~4

#A .0

W..~~ ~ ~ ~ ~ ~ ~ -.

ik its ,.

4;* -w , A ; ,

0'0

a *. * ; 5. ; 4

* .5 a : --' S

4.d

*,AXj ¢

Figure 1 Intraduct cribriform carcinoma showingmoderately positive staining for oestrogen receptors.(Immunoperoxidase. )

plastic epithelial cells. No cytoplasmic,stromal, or background staining was seen.

Fifty one tumours (850*) were immunohis-tologically oestrogen receptor negative, sixweakly positive, and three moderately positive.No strongly positive cases were seen. Thedistribution of the results according to his-tological type is shown in table 2.The weakly positive cases included five

invasive ductal carcinomas and an atypicalmedullary carcinoma. The total score was 1(I1 x 1 ) in four of these tumours and 3 (3 x 1 )in two. Three out of the six weakly positivecases contained foci of intraductal carcinomaand these were all positively stained for oes-trogen receptor (I00°/). In contrast, only one(70°,") out of 14 negatively stained invasiveductal carcinoma with intraductal elementsshowed positive staining of these elements.Most of the positively stained intraductalelements were of the solid and cribriform type(figure 1), and most of the negative ones were ofthe comedo type.The three tumours with moderately positive

oestrogen receptor staining included the onlytwo pure tubular carcinomas included in thisseries and a well differentiated invasive ductalcarcinoma. For the tubular carcinomas thetotal oestrogen receptor score was 4 (2 x 2) inone and 6 (3 x 2) in the other. The twotumours were of a small size, each measuringI cm in maximum dimension, and their neo-plastic glands were widely separated by abun-dant fibrous tissue (figure 2). The second ofthese tumours had foci of solid intraductalcarcinoma which were positively stained. Themoderately stained ductal carcinoma had ascore of 4 (2 x 2) and contained positivelystained solid intraductal elements. It had aprogesterone receptor concentration of80 fmol/mg cytosol protein (assayed bio-chemically). No record of the size of thistumour was available.

240

on 2 July 2018 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.43.3.239 on 1 M

arch 1990. Dow

nloaded from

Immunohistological study of oestrogen receptors in breast carcinomas that are biochemically receptor negative

A.-~~~~~

'+ ,d<I . '

Figure 2 Tubular carcinoma of the breastubules show moderate stainingfor oestrogwhile the normal ducts, seen here crossing tthe field, are oestrogen receptor negative.(Immunoperoxidase.)

Although the invasive elementslobular carcinoma included in thi;oestrogen receptor negative, the ttained foci of lobular carcinoma iiwere faintly stained.The results of progesterone r(

chemical assay were available fornine immunohistologically oestrolweakly or moderately positive tumall 51 oestrogen receptor negatiFive (620.) of the eight oestrogpositive carcinomas, including twmoderate and three out of six we

cases, were progesterone recep

(> 10 fmol/mg cytosol protein; ra

mean 30), compared with eight (11-46; mean 25) of the oestrognegative tumours. The differeincidence of progesterone recept4between the two groups was

(p < 0 01). When the results of(n = 6) and moderately (n = 2) rreceptor positive tumours were se

mean progesterone receptor concthe first group (11 fmol) was mucdthat of the second group (45 fmoltical analysis was carried out be(small size of the groups.

DiscussionIn this study we concentrated o

receptor negative breast carcinombiochemically, for two reasons. Fiithe usefulness and reliability of tIhistological method applied tcprocessed wax paraffin sections, fcmination of the oestrogen receptosuch tumours; and secondly, topossibility that in some tumours t

-. oestrogen receptor biochemical assay may notbe the result of real absence of receptors, butthe result of the presence of a "diluting" factorlike normal tissue in a small sized tumour or

-> abundant fibrous tissue stroma in larger ones.The results of the study show complete

* ii- agreement between the chemical and his-- tological methods in 85O4 of cases. This is

similar to the results obtained by applying theimmunohistological method to frozen sectionsof the tumours.78 In a further 10O of our cases

;._-t^0t in which the staining was only weakly positivethe amount of receptors shown was probablyinsignificant, especially where the total scorewas 1, reflecting a weak focal staining of a small

w/'- part of the tumour, and even in cases with ascore of 3 (3 x 1) where a more diffuse faintstaining was seen. It may be more prudent,from a practical point of view, to regard thesecases as negative, as suggested by the bio-chemical assay. This may not be the case with ascore of 3 (1 x 3) where an intense stainingwould be present in a small part of the tumour,

s 4- but we did not encounter such cases in this-t. Malignant series.en receptors, The situation with the three cases whichrhe middle of showed a moderately positive degree of stain-

ing is different. Here the staining was moreintense and most cells stained. In at least twocases the tumours were of a small size and the

s in the only neoplastic glands were widely separated froms study were each other by abundant fibrous tissue. Bio-tumour con- chemical assay for progesterone receptor wasn situ which high (80 fmol) in one tumour, low positive

(1 1 fmol) in the second, and was not assayed ineceptor bio- the third tumour because of the unavailabilityeight out of of tissue as a result of the small size of thegen receptor lesion. All these factors would support con-ours and for sidering these cases oestrogen receptor positiveve tumours. and the biochemical assay as representing ayen receptor false negative result. A similar discrepancyo out of two between oestrogen receptor biochemical assayakly stained and immunohistological result has beenItor positive observed previously in small tumours.78ange: 1 1-80; The pattern of oestrogen receptor staining of(160°o; range in situ carcinoma noted in this study wasren receptor basically similar to that previously reported bynce in the us in a series of mostly oestrogen receptoror positivity positive mucinous carcinomas.4 In particular,

significant the staining of in situ elements tended to bethe weakly similar to that of the associated invasive lesion.

progesterone This was the case in all but two of the studiedparated, the tumours which contained in situ malignant:entration in elements. The two exceptions were an oes-h lower than trogen receptor negative invasive ductal1). No statis- tumour with positive solid and cribriformcause of the intraductal carcinoma and an oestrogen recep-

tor negative invasive lobular tumour withpositive foci of in situ lobular carcinoma. Ingeneral, most of the solid, cribriform, andpapillary intraductal carcinomata showed vary-

,n oestrogen ing degrees of positivity, while intraductala, estimated carcinoma of the comedo type was mostlyrst, to assess negative.he immuno- It is concluded that the immunohistologicalD routinely method used in this study for the demonstra-)r the deter- tion of oestrogen receptor in routinely proces-tr content of sed paraffin wax sections can be reliably usedassess the for the determination of the oestrogen receptor

the negative content of negative as well as positive tumours.

241

on 2 July 2018 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.43.3.239 on 1 M

arch 1990. Dow

nloaded from

Shousha, Stamp, James, Alaghband-Zadeh

In particular, negative staining in this and inour previous study,4 was only seen in bio-chemically oestrogen receptor negativetumours. It seems, however, that a minority ofbiochemically oestrogen receptor negativetumours may show positive staining. Thestaining in most of these cases is weak, and forpractical purposes these tumours may have tobe considered negative, especially if the faintstaining is localised to a small area. In smallmoderately stained tumours the resultsobtained with this method are probably moreaccurate than those obtained by biochemicalassay, because of the possibility of samplingerrors. The most critical step of the method isthe determination of the suitable concentrationof DNAse that can be used and this has to bedecided by titration of each new batch of theenzyme. The method would be particularlyvaluable for the assessment of the oestrogenreceptor content of the increasing numbers ofsmall tumours likely to be encountered inclinical practice with the introduction of theNational Breast Screening Programme. Inmany of these small tumours, or suspectedtumours, no tissue will be available for bio-chemical assay and frozen section examinationwill be contraindicated. The whole lesion has tobe examined thoroughly by paraffin wax sec-tions, some ofwhich can be used for assessment

of oestrogen receptor content by this method, ifand when it is needed.

This study was supported by a grant from the Peel MedicalResearch Trust.

1 Shintaku P, Said JW. Detection of estrogen receptors withmonoclonal antibodies in routinely processed formalin-fixed paraffin sections of breast carcinoma. Use ofDNAsepretreatment to enhance sensitivity of the reaction. Am JClin Pathol 1987;87:161-7.

2 Andersen J, Orntoft TF, Poulsen HS. Immunohisto-chemical demonstration of estrogen receptors (ER) informalin-fixed, paraffin-embedded human breast cancertissue by use of a monoclonal antibody to ER. J HistochemCytochem 1988;36:1553-60.

3 Cheng L, Binder SW, Fu YS, Lewin KJ. Demonstration ofestrogen receptors by monoclonal antibody in formalin-fixed breast tumors. Lab Invest 1988;58:346-53.

4 Shousha S, Coady AT, Stamp T, James KR, Alaghband-Zadeh J. Oestrogen receptors in mucinous carcinoma ofthe breast: an immunohistological study using paraffinwax sections. J Clin Pathol 1989;42:902-5.

5 King RJB, Redgrave S, Hayward JL, Millis RR, RubensRD. The measurement of receptors for oestradiol andprogesterone in human breast tumours. In: RJB King, ed.Steroid receptor assays in human breast tumours. Meth-odological and clinical aspects. Cardiff: Alpha Omega,1979:55-72.

6 Said J, Shintaku IP. Detection of estrogen receptors withmonoclonal antibodies in paraffin sections. Am J ClinPathol 1988;90:120.

7 Cudahy TJ, Boeryd BR, Franlund BK, Nordenskjold BA. Acomparison of three different methods for the determina-tion of estrogen receptors in human breast cancer. Am JClin Pathol 1988;90:583-90.

8 Hiort 0, Kwan PWL, DeLellis RA. Immunohistochemistryof estrogen receptor protein in paraffin sections. Effects ofenzymatic pretreatment and cobalt chloride intensifica-tion. Am J Clin Pathol 1988;90:559-63.

242

on 2 July 2018 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.43.3.239 on 1 M

arch 1990. Dow

nloaded from