Breast CancerResearch and Treatment16: 243-251, 1990. © 1990KluwerAcademic Publishers. Printedin the Netherlands. Report

Colocalization of estrogen and progesterone receptors with an estrogenregulated heat shock protein in paraffin sections of human breast and endometrial cancer tissue Daniel R. Ciocca 1, Arturo O. Stati 1 and Marfa M. Amprino de Castro 2 1 Unidad de Histo-Patologia Enddcrina, Laboratorio de Reproducci6n y Lactancia (LARLAC), Centro Regional de Investigaciones Cientificas y Tecnoldgicas ( CRICYT) ; 2Servicio de Anatomfa Patoldgica, Hospital Central, Mendoza 5500, Argentina

Key words: estrogen receptor, progesterone receptor, estrogen-regulated protein, immunocytochemistry, breast cancer, endometrial cancer, heat shock protein Summary

We have studied by immunocytochemistry and monoclonal antibodies the presence and localization of estrogen receptors, progesterone receptors, and a 24-kD estrogen-regulated heat shock protein in biopsies from breast and endometrial cancer patients. Three different tissue processing protocols were used to colocalize the antigens in the same tissue sections: a) frozen sections, b) formalin fixation with routine paraffin embedding, and c) picric acid-formaldehyde (PAF) fixation with a rapid embedding in paraffin. Frozen sections showed good receptor staining but poor 24-kD protein immunoreactivity, while routine paraffin sections,(with or without DNase pretreatment) were inadequate to reveal the nuclear receptor proteins at the same level seen in frozen sections. On the other hand, all three proteins could be detected satisfactorily in PAF-fixed paraffin-embedded tissue. Using this procedure we were able to visualize 24-kD protein and estrogen receptor or progesterone receptor in individual cells in paraffin sections. The study revealed that in all of the estrogen receptor positive breast and endometrial tumor samples, almost 90% of the cells expressing the cytoplasmic 24-kD protein contained estrogen receptor in the cell nucleus. In contrast, 24-kD immunoreactive cells did not express progesterone receptors in almost 40% of the progesterone receptor positive tumor samples.

Introduction

Several laboratories are investigating additional markers of hormone sensitivity for more accurate identification of hormone dependent tumors, and among these markers there are estrogen-regulated messengers or proteins, estrogen-induced enzymes, estrogen-modulated secretory growth factors, and estrogen-regulated genes [1-3]. One of the estrogen-regulated proteins under study is that

characterised by a molecular weight of 24,000 daltons (24-kD) which is regulated by estrogen in the MCF-7 human breast cancer cell line [4]. This protein has also been termed 28-kD in previous publications since this has been the reported molecular weight after its purification by monoclonal antibody affinity chromatography [5]. This protein has recently been identified as a heat shock protein [6]. Using monoclonal antibodies and a competitive enzyme-linked immunoassay, elevated expression

Address for offprints: D.R. Ciocca, LARLAC, Casilla de Correo 855, 5500Mendoza, Argentina

244

DR Ciocca et al.

of 24-kD protein was found to correlate well with presence of estrogen and progesterone receptors in human breast tumor cytosols [7]. However, no relationship was shown recently between expression of 24-kD protein and presence of estrogen receptors in human breast cancer biopsies using immunohistochemistry [8]. In addition, using double immunoperoxidase cytochemistry, the presence of 24-kD protein correlated with that of estrogen receptor in human breast cancer, but not on a cell-bycell basis [9]. Therefore, in order to clarify this subject, in the present report we have studied by immunocytochemistry the presence of 24-kD protein, estrogen receptor (ER), and progesterone receptor (PR) in human breast and endometrial adenocarcinomas. We have used a fixation and embedding protocol which allowed us to visualize 24-kD protein and E R or PR in individual cells in paraffin sections.

Materials and methods

Tissue processing Forty-six primary carcinoma samples from 34 women with breast cancer and from 12 women with endometrial cancer were studied. The first group of tumors was composed of 12 breast and 12 endometrial adenocarcinomas. These samples were obtained immediately after surgery, cut into small pieces of 2-3 mm thickness, and divided into three fractions: a) quick-frozen for cryostat sectioning; b) fixed in neutral buffered formalin for 12-18 hr at room temperature and processed for routine paraffin embedding; and c) fixed in picric acid-formaldehyde (PAF) solution [10] during 30 min at room temperature. The tissue samples fixed in PAF solution were washed in the buffer used to prepare this fixative [10] (2 changes of 5 min each) and dehydrated at room temperature as follows: 70% ethanol (2 changes of 10min each), 96% ethanol (2 changes of 10 min each), 100% ethanol (2 changes of 10 min each), and xylene (2 changes of 10 min each). Then, the samples were passed by three Histowax (R. Jung GmbH, Heidelberg, FRG) series (of 15 min each) and embedded in Histowax. In

order to increase the number of tumor samples, a second group of 22 breast carcinomas was added to the study. Since the first group showed that PAF fixation was adequate to reveal the 3 antigens under study, the second group of tumors was processed for fixation only in PAF solution with the rapid embedding in paraffin.

Immunological reagents The mouse monoclonal antibody (Cll) against the 24-kD protein prepared from the human breast tumor cell line MCF-7 [5], was kindly provided by Dr. W.L. McGuire (The University of Texas Health Science Center at San Antonio, Texas); the rat monoclonal antibody (H222) against E R prepared from MCF-7 cells [11] was kindly provided by Dr. E. De Sombre (The University of Chicago, Illinois); and the mouse monoclonal antibody (mPRI) against PR prepared from human uterine tissue [12] was purchased from Transbio SARL (Paris, France). As second antibody we used antimouse IgG (whole molecule), biotin conjugate, developed in goat (Sigma Chemical Company, St. Louis, Missouri), and anti-rat IgG, biotin conjugate, developed in sheep (Amersham International, UK). Extravidin peroxidase conjugate (Sigma) was used for binding to the second antibodies. We also used the ER-ICA kit (Abbott Laboratories, North Chicago, Illinois).

Immunocytochemical procedures For the study of 24-kD protein, ER, and PR in cryostat sections, 8/zm sections were thaw-mounted onto 0.1% poly-L-lysine (w/v) (Sigma)-coated slides and immediately fixed in PAF solution [10] for 3 min at room temperature. After rinsing in phosphate buffer, the sections were incubated with 3% normal sheep or goat serum for 30 min. Then, the sections were incubated overnight at 4° C with the primary antibodies: H222, Cll, mPRI (10/zg/ ml), or with normal rat and mouse serum (negative control slides), in a moist incubation chamber. Then, the sections were incubated successively

Colocalization of estrogen and progesterone receptors with biotinylated anti-rat or anti-mouse antibodies diluted 1 : 50 for 45 min at room temperature, followed by Extravidin-peroxidase complex diluted 1 : 100 for 45 min. Each incubation was followed by two 10min washings in PBS. Finally, the sections were incubated with diaminobenzidine (DAB) tetrahydrochloride (Sigma)-hydrogen peroxide reagents prepared as reported previously [13] for 15 min, dehydrated, cleared in xylene, and mounted with Permount. Some of them were previously counterstained with hematoxylin. The paraffin-embedded sections (5-6/xm thickness) of the tissue fixed in PAF solution were also mounted on poly-L-lysine-coated slides. Serial sections and mirror sections were used for the precise demonstration of the same cells in two consecutive paraffin sections. The mirror sections were obtained by mounting two consecutive sections on two slides with the cut surfaces facing each other upward; in this manner, the same surface of the sectioned cells was ready to be stained by two different monoclonal antibodies. In addition, since 24-kD protein is a cytoplasmic antigen while ER and PR are nuclear antigens, once a section immunostained for ER or PR was studied and photographed, it was placed in xylene overnight to detach the coverslip and then it was immunostained again but to reveal 24-kD protein. Since we used two different chromogens, the receptors gave a brown reaction product while the 24-kD protein gave a red reaction. The immunostaining was done as reported previously [13]. Briefly, deparaffinized slides were incubated overnight with the primary antibodies: H222 (25/xg/ml) or mPRI (10/xg/ml), at room temperature in moist chambers. For ER we obtained better staining using the biotinylated anti-rat antibody at i : 20 and the Extravidin-peroxidase complex at 1 : 50 dilution; while for PR we used the biotinylated anti-mouse antibody at 1 : 50 and the Extravidin-peroxidase complex at 1 : 100. In these cases each incubation time was of 45 min at room temperature. Receptors were visualized with DAB-H202 with an incubation time of 15 min. Once the slides were mounted, they were studied and photographed under a Zeiss IM35 microscope using differential interference contrast optics whenever required. The immunostaining for 24-

245

kD protein was done by incubating the tissue with C l l (5/~g/ml) antibody overnight at room temperature [14], or with preabsorbed C l l antibody (negative controls). In this case the biotinylated antimouse antibody was used at 1 : 100 dilution and the Extravidin-peroxidase complex was used at 1 : 180 dilution, in both cases for 45 min at room temperature. The 24-kD protein was visualized incubating with DAB-H20 z for 5min, or with AEC (aminoethylcarbazole)-H202 for 15 min.

Routinely processed formalin-fixed paraffin sections (6/xm thickness) mounted onto poly-L-lysinecoated slides, with or without treatment with DNase (Sigma) [15], were used to visualize the receptor proteins and the 24-kD protein. The sections were covered with DNase prepared as published [15] and then incubated for 30 min, 45 min, 60min, 90min, or 2hr in humidity chambers at room temperature before immunostaining. The immunocytochemical procedure to detect ER, PR, and 24-kD protein in these sections was as described above for the PAF-fixed paraffin sections.

Positive controls Each immunocytochemical run included positive controls for ER, PR, and 24-kD protein. The presence of ER or PR in these samples was confirmed using a one point dextran-coated charcoal assay performed as described elsewhere [16], while the presence of the 24-kD protein was confirmed by the Western blot technique [16]. Previous studies have shown an excellent agreement when the presence of ER and PR was evaluated by immunocytochemistry and by conventional steroid-binding assays [17, 18].

Quantitative and statistical analysis The quantitative analysis of the immunostained tumor cells was assessed by visually counting the percentage of reacting nuclei (for ER and PR) and the percentage of reacting cells containing nuclei (for 24-kD protein); at least 1,000 cells were counted in each sample under the × 100 objective. In

246

D R Ciocca et al.

order to compare the immunostaining in frozen versus paraffin sections, and in the paraffin serial sections, we used a systematic random sampling method similar to that employed in the selection of tissue blocks for electron microscopy [19]. Therefore, sample fields representative of the different areas of the tumor were evaluated; this was useful for those cases with a heterogeneous immunostaining distribution. In the mirror sections of the paraffin-embedded material, the same tumor areas and cells were evaluated for 24-kD versus ER and for 24-kD versus PR. In addition, we were able to evaluate the cells with double immunostaining: for nuclear ER or PR and for cytoplasmic 24-kD protein. A regression analysis was performed when the percentage of ER or PR positive cells seen in frozen sections was compared with the percentage of ER or PR positive cells observed in PAF-fixed paraffin-embedded sections (Fig. 2). In this case we applied the equation:

log q

087 log P Sps + 0.24

The student's t test for proportions was applied when appropriate (Table 2, Fig. 4).

Results

A summary of the immunostaining results for ER, PR, and 24-kD protein according to the different tissue processing conditions is shown in Table 1. Frozen sections were suitable to reveal the receptor proteins in the cell nuclei, but 24-kD protein identification was difficult owing to diffusion of the protein and to the weakness of the staining. On the other hand, formalin-fixed paraffin sections, with or without pretreatment with DNase [15], did not show the nuclear receptor proteins with a percentage similar to that seen in frozen sections. In most of the cases the percentage of ER-containing cells was lower and less dependable in formalin-fixed paraffin sections than in frozen sections. Figure 1 illustrates mirror sections of DNase-pretreated sections with areas where 24-kD protein appeared in cells with ER positive nuclei. However, this figure also shows areas where 24-kD protein was seen in cells lacking ER. The immunostaining pattern for ER improved with shorter incubations with DNase, but the results were still not comparable to those seen in frozen sections. Moreover, staining for PR could not be evaluated in formalin-fixed paraffin sections since the very low number of PR positive cells seen in untreated sections disappeared when the sections were pretreated with DNase. On the other hand, the percentage of ER and PR

Table 1. Effect of fixation on E R , PR, and 24-kD protein immunostaining in h u m a n breast and endometrial carcinomas PAF-fixed frozen sections"

PAF-fixed paraffin sections a

Formalin-fixed paraffin sections

Formalin-fixed paraffin sections b

ER

+ nuclear

+ nuclear, %c correlated with frozen sections

+ nuclear, % not correlated with frozen sections

PR

+ nuclear

24-kD

+ cytoplasmic but also diffuse b e t w e e n the t u m o r cells

+ nuclear, % correlated with frozen sections + cytoplasmic, % similar to that found in formalin-fixed paraffin sections

+ nuclear, % not correlated with frozen sections + cytoplasmic

+ nuclear, % correlated with frozen sections in s o m e samples - no staining

PAF: picric acid-formaldehyde solution [10]. b Pretreated with D N a s e I, as described in Shintaku and Said [15]. c %: percentage of i m m u n o s t a i n e d t u m o r cells. a

+ cytoplasmic

Colocalization o f estrogen and progesterone receptors

247

Fig. 1. Formalin-fixed paraffin sections from an endometrial adenocarcinoma. Mirror sections (a versus b; c versus d) were mounted to compare ER and 24-kD immunostaining, a, numerous cells with immunoreactive nuclei for ER can he seen (arrows) in this section treated with DNase before incubation with H222 antibody, b, note the presence of 24-kD protein in the apical cytoplasm of the cells (arrow), compare with a. c, very few immunostained nuclei for ER can be seen in this section treated with DNase before incubation with H222 antibody, d, note the presence of numerous glands displaying immunostaining for 24-kD protein, some of them lacking ER (compare with c). Biotin-Extravidin-peroxidase immunostaining without counterstaining, a and b: × 450; c and d: x 350. positive nuclei observed in frozen sections correlated significantly (p -< 0.001) with that obtained from the same samples processed for PAF-fixed paraffin sections (Fig. 2). It is important to mention here that to obtain consistent results in the detection of the receptor proteins in PAF-fixed paraffin sections, the protocol described above in Materials and Methods should be followed with preciseness. For instance, we failed when the E R - I C A kit was used to immunostain PAF-fixed paraffin sections. In the present report we used a higher concentration of the primary antibodies and the very sensitive biotin-Extravidin technique to reveal the receptor proteins. Subsequently, we investigated in tissue samples fixed in P A F solution and e m b e d d e d in paraffin the presence and localization of E R , 24-kD protein,

and PR. These proteins were first evaluated in serial sections. Fig. 3a, b shows a breast carcinoma with immunoreactive E R and 24-kD protein. The quantitative analysis showing the percentages of E R , 24-kD protein, and P R immunostained cells in breast and endometrial carcinomas m a y be seen in Table 2. The results of this evaluation show that in 100% of the samples the percentage of cells expressing the 24-kD protein was similar to or lower than the percentage of E R immunostained cells. In almost 70% of the E R positive cases, the percentage of t u m o r cells expressing E R was similar to that expressing the 24-kD protein, while in the other 30% of the cases there were m o r e t u m o r cells expressing E R than 24-kD protein. On the other hand, Table 2 shows that only in 23% of the P R positive cases was the percentage of cells express-

248

D R Ciocca et al.

100

reactive cells was similar or lower than the percentage of PR immunoreactive cells (Table 2).

0

9O

•

A

a0

8o

z

70

0 0 0000

00000 Q

Discussion

o%

G)

O

N 5O v

O

O

~ 40

Oo

O

~ 30 •

,5 20 N P ,, 10

O

e-Breast Cancer O-Endometrial Cancer

o o o I

I

1

I

I

I

I

I

I

I

30 40 50 60 70 80 90 I00 PAF Fixed Parafin Sections (% Stained Nuclei )

!0 20

Fig. 2. Correlation b e t w e e n percentages of E R and P R i m m u nostained cells obtained in frozen sections and percentages of E R and P R i m m u n o s t a i n e d ceils obtained in PAF-fixed paraffin sections. T h e s e samples were obtained from patients with breast and endometrial adenocarcinomas, n = 48, y = 0.87x +0.24; R 2= 0.85.

ing PR similar to that expressing the 24-kD protein, while in 40% of the cases there were more tumor cells expressing 24-kD protein than PR. The detailed analysis of the serial sections revealed that the tumor areas with immunoreactive 24-kD cells always had ER, while the presence of PR immunoreactive cells in these same areas was variable. The colocalization study of 24-kD protein and ER or PR was carried out in mirror sections (Fig. 3c, d), and in double immunostained paraffin sections (Fig. 3e-g), of breast and endometrial adenocarcinomas. This evaluation was performed on a cell by cell basis, recording the percentage of 24-kD immunoreactive cells with and without the receptor proteins. This quantitative analysis revealed that in 100% of the ER positive breast and endometrial tumors almost 90% of the cells expressing 24-kD protein contained ER in the cell nucleus (Fig. 4). The colocalization study also revealed that from 82% to 94% of the cells expressing 24-kD protein contained PR in the cell nucleus, but this was obvious only in that 60% of the PR positive samples where the percentage of 24-kD immuno-

This study shows that frozen sections were not appropriate to study the 24-kD protein by immunocytochemistry. This result is not surprising since one of the problems created in frozen sections is that soluble antigens may migrate and may be lost during the freezing process [20], which occurs beTable 2. Immunocytochemical evaluation of E R , 24-kD protein, and P R in breast and endometrial tumors fixed in P A F solution and e m b e d d e d in paraffin Case

% Stained cells a ER

24-kD

PR

Endometrial carcinomas 1 74 73 74 2 71 69 63 3 47 45 53 4 38 27 17 5 75 63 18 8 81 > 60 56 9 30 > 9 < 55 10 83 > 30 < 52 11 77 > 33 < 70 12 82 > 42 > 12 Breast carcinomas 1 4 0 3 2 0 0 0 3 0 0 0 4 0 0 0 5 5 0 0 6 0 0 0 7 0 0 0 8 0 0 0 9 76 68 73 10 66 69 58

a Expressed

Case

% Stained cells a ER

Breast 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

24-kD

carcinomas 60 > 46 94 > 47 46 > 23 77 > 18 75 > 27 62 > 38 64 > 23 56 > 5 73 > 15 77 64 56 54 42 37 37 35 43 50 85 77 72 77 56 50 78 75 73 68 63 60 75 68 57 60 86 88 37 40

> > > > < < < < < < < < > > > > > > > > > > >

PR

40 12 12 5 10 76 35 60 47 82 87 89 77 13 26 0 3 56 5 8 8 0 0 0

as percentage of i m m u n o s t a i n e d cells (reacting nuclei for E R and PR, reacting cells containing nuclei for 24-kD protein). bThe symbol > or < is used w h e n the differences between the percentages of i m m u n o s t a i n e d cells are statistically significant (p -< 0.01).

Colocalization of estrogen and progesterone receptors

249

Fig. 3. PAF-fixed paraffin sections from breast and endometrial adenocarcinomas. Serial sections (a and b) and mirror sections (c versus d) were used to reveal the receptor proteins and the 24-kD protein in adjacent sections. Some sections were double immunostained (e-g). a and b, serial sections of a breast tumor showing ER immunoreactive cells (a) and 24-kD immunoreactive cells (b). Biotin-Extravidin-peroxidase technique without counterstaining, x 350. c and d, mirror sections of an endometrial carcinoma showing E R immunoreactive cells (c) and 24-kD immunostained cells (d). The arrows point to ER-containing cells shown with more detail in e. Biotin-Extravidin-peroxidase technique without counterstaining, x 350. e, this section from an endometrial carcinoma was double immunostained, first to reveal nuclear E R (c, arrows), and then to show cytoplasmic 24-kD protein. Biotin-Extravidin-peroxidase technique using DAB and A E C as chromogens, x 900. f, section from a breast carcinoma with double immunostaining: for E R (nuclear staining) and for 24-kD protein (cytoplasmic staining). In this case the cells have large nuclei. Biotin-Extravidin-peroxidase technique without counterstaining. × 900. g, section from an endometrial carcinoma showing a large cell immunostained for PR (nuclear) and for 24-kD protein (cytoplasmic). Biotin-Extravidin-peroxidase technique without counterstaining, x 1,000.

250

D R Ciocca et al.

Breast Cancer

D+,ER+~

24-K 24-KD+,ER-I E--I Endometrial Cancer

24-KD+,ER+L_.~

24-KD+,ER

i

I0 30 50 7O % IrnmunoreQctive Cells.

i

9O

Fig. 4. Colocalization of 24-kD protein and E R in samples from breast and endometrial carcinomas. The difference between the percentages of 24-kD +, E R + cells versus 24-kD +, E R cells is statistically significant (p -< 0.01) in both tumor types.

fore fixation. The presence of the protein in these samples was confirmed by the Western blot technique in the positive controls and by immunocytochemistry in the paraffin sections of tissues fixed without freezing (formalin and PAF fixation). Therefore, our finding may explain the lack of relationship between staining for E R and 24-kD protein reported by Horne et al. [8], who utilized frozen sections to study these proteins in breast tumor biopsy samples. The use of frozen sections has been recommended for detecting E R and PR by immunocytochemistry with the antibodies employed in the present study [11, 12]. Thus, in order to colocalize the 24-kD protein, ER, and PR, we needed to develop a fixation and embedding procedure avoiding freezing of the samples but sensitive enough to retain ER and PR immunoreactivity. Our results show that all three proteins can be detected satisfactorily in PAF-fixed paraffin-embedded tissue. This protocol gave percentages of E R and PR positive cells comparable to those seen in frozen sections, when the primary antibodies were used at higher concentrations and applying the sensitive biotin-Extravidin immunocytochemical technique. The use of paraffin embedded material was of value for detailed morphological interpretation of the immunoreactivities allowing colocalization of the

24-kD protein with E R or PR in adjacent and in the same tissue section. Our results show that the presence of the 24-kD protein correlates significantly with the presence of ER on a cell-by-cell basis in both breast and endometrial adenocarcinomas. In addition, the immunocytochemical study revealed the high heterogeneity in content and distribution of ER, PR, and 24-kD protein in the tumor samples. We found that in 40% of the PR positive samples the PR content was significantly lower than that of 24-kD protein, and in these cases the colocalization study revealed that the cells with 24-kD protein may not express PR. The presence of the 24-kD protein correlated with the presence of PR only in those cases where the percentage of PR immunoreactive cells was similar or higher than the percentage of E R positive cells (60% of the PR positive samples). Therefore, the present study indicates that 24-kD protein is present in individual cells carrying E R and that when the tumor samples have a relatively high content of PR, the cells with 24-kD protein contain PR as well. This is consistent with a previous report where the presence of 24-kD protein was correlated with that of ER and PR in breast tumor cytosols, with the correlation coming primarily through E R [7]. A similar result was obtained in patients with endometrial adenocarcinomas without squamous components [16]. In this study the presence of 24kD protein correlated significantly with that of ER in endometrial tumor cytosols, and when the evulation was performed in tumor sections (frozen sections for ER and PR and routine paraffin sections for 24-kD protein) the presence of the 24-kD protein correlated mainly with that of E R (p-< 0.005) as compared with that of PR (p - 0.05). Finally, we found a low percentage of 24-kD positive cells which did not express ER or PR. This may be due to: a) methodological failures to reveal the receptor proteins (amount under the detection limit); or b) that in these cells the presence of the 24-kD protein is independent of the presence of the receptor proteins.

Colocalization of estrogen and progesterone receptors

Acknowledgments This w o r k was s u p p o r t e d by a G r a n t f r o m t h e N a-

11.

tional Research Council ( C O N I C E T ) of Argentina. 12.

References 1. Lippman ME, Dickson RB, Kasid A, Gelmann E, Davidson N, McNamara M, Huff K, Bronzert D, Bates S, Swain S, Knable C: Autocrine and paracrine growth regulation of human breast cancer. J Steroid Biochem 24: 147-154, 1986 2. Jeltsch JM, Roberts M, Schatz C, Garnier JM, Brown AMC, Chambon P: Structure of the human oestrogenresponsive gene pS2. Nucleic Acids Res 15: 1401-1413, 1987 3. Ciocca DR: Tumores hormonodependientes: conceptos actuales. Medicina (Bs As) 48: 87-94, 1988 4. Edwards DP, Adams DJ, McGuire WL: Estradiol stimulates synthesis of a major intracellular protein in a human breast cancer cell line (MFC-7). Breast Cancer Res Treat 1: 209-223, 1981 5. Adams D J, Hajj H, Bitar KG, Edwards DP, McGuire WL: Purification of an estrogen-regulated breast cancer protein by monoclonal antibody affinity chromatography. Endocrinology 113: 415-417, 1983. 6. Fuqua SAW, Blum-Salingaros M, McGuire WL: Induction of the estrogen-regulated '24K' protein by heat shock. Cancer Res 49: 41264129, 1989. 7. Adams D J, McGuire WL: Quantitative enzyme-linked immunosorbent assay for the estrogen-regulated Mr 24,000 protein in human breast tumors: correlation with estrogen and progesterone receptors. Cancer Res 45: 2445-2449, 1985. 8. Horne GM, Angus B, Wright C, Needham G, Nicholson S, Harris AL, Innes B, Horne CHW: Relationship between oestrogen receptor, epidermal growth factor receptor, ERD5, and P24 oestrogen regulated protein in human breast cancer. J Pathol 155: 143-150, 1988. 9. Skovgaard Poulsen H, Andersen J, Skovbon H: Estrogen regulated proteins in human breast cancer: immunocytochemical detection in formalin fixed, paraffin-embedded surgical specimens; their interrelationship and their predictive values in patients response to endocrine therapy. Proc AACR 29: A985, 1988. 10. Stefanini M, De Martino C, Zamboni L: Fixation of ejac-

13.

14.

15.

16.

17.

18.

19.

20.

251

ulated spermatozoa for electron microscopy. Nature 216: 173-174, 1967. Press MF, Greene GL: An immunocytochemical method for demonstrating estrogen receptor in human uterus using monoclonal antibodies to human estrophilin. Lab Invest 50: 480--486, 1984. Perrot-Applanat M, Groyer-Picard MT, Lorenzo F, Jolivet A, Hai MTV, Pallud C, Spyratos F, Milgrom E: Immunocytochemical study with monoclonal antibodies to progesterone receptor in human breast tumors. Cancer Res 47: 2652-2661, 1987. Ciocca DR, Bjercke ILl: Immunocytochemical techniques using monoclonal antibodies. In Methods in Enzymology, Vol. 121. Immunochemical Techniques. Part I. Hybridoma Technology and Monoclonal Antibodies. Edited by JJ Langone, H Van Vunakis. Academic Press, Miami, 1986, p. 562-579. Ciocca DR, Adams DJ, Bjercke RJ, Edwards DP, Chamness GC, McGuire WL: Monoclonal antibody storage conditions, and concentration effects on immunohistochemical specificity. J Histochem Cytochem 31: 691-696, 1983. Shintaku IP, Said JW: Detection of estrogen receptors with monoclonal antibodies in routinely processed formalinfixed paraffin sections of breast carcinoma. Use of DNase pretreatment to enhance sensitivity of the reaction. Am J Clin Pathol 87: 161-167, 1987. Ciocca DR, Puy LA, Fasoli LC: Study of estrogen receptor, progesterone receptor, and the estrogen-regulated Mr 24,000 protein in patients with carcinomas of the endometrium and cervix. Cancer Res 49: 4298-4303, 1989. De Sombre ER, Thorpe SM, Rose C, Blough RR, Andersen KW, Rasmussen BB, King WJ: Prognostic usefulness of estrogen receptor immunocytochemical assays for human breast cancer. Cancer Res 46 (Suppl): 4256s-4264s, 1986. Perrot-Applanat M, Groyer-Picard MT, Lorenzo F, Jolivet A, Vu Hai MT, Pallud C, Spyratos F, Milgrom E: Immunocytochemical study with monoclonal antibodies to progesterone receptor in human breast tumors. Cancer Res 47: 2652-2661, 1987. Weibel ER: Stereological principles for morphometry in electron microscopic cytology. Internat Rev Cyto126: 235302, 1969. Taylor CR: Immunomicroscopy: A Diagnostic Tool for the Surgical Pathologist. WB Saunders Company, Philadelphia, 1986, p. 43-69.