2571

Immunohistochemical Expression of the Estrogen Receptor-Related Antigen (ER-D5) in Human Intracranial Tumors Humayun Khalid, M.B.B.S.,*Akio Yasunaga, M.D., Ph.D.,* Masao Kishikawa, M.D., Ph.D.,t and Shobu Shibata, M.D., Ph.D.*

Background. Expression of the estrogen receptor-re- lated antigen (ER-D5) has been reported in some normal and neoplastic tissues. The authors evaluated the expres- sion of ER-D5 in 143 intracranial tumors of different his- tologic types.

Methods. Formalin fixed, paraffin embedded tumor sections were stained with the monoclonal D5 antibody by avidin-biotin complex immunohistochemistry.

Results. Eighty-eight (62%) of the 143 brain tumors showed positive ER-D5 immunoreactivity. ER-D5 expres- sion was observed in 9/30 low grade astrocytomas, in 61 13 anaplastic astrocytomas, in 16/27 glioblastomas, in 2/ 5 ependymomas, in 5/8 medulloblastomas, in 10/15 me- ningiomas, in 20/23 schwannomas, in 11/11 hemangi- oblastomas, in 9/9 germ cell tumors, in 0/2 oligodendro- gliomas, and in 17/28 pediatric and childhood brain tu- mors. The mean percentage of ER-D5-positive cells varied in different tumor types, was lowest in the meningothe- liomatous meningiomas, and was highest in the heman- gioblastomas. ER-D5 immunoreactivity was also ob- served in the microvascular endothelial proliferations and in tumor blood vessels. ER-D5 expression in tumors was not related to the overall tumor grades, but a statisti- cally significant higher percentage of ER-D5-positive cells was noted in the glioblastomas compared with the low grade astrocytomas (P < 0.05) and in the combined high grade tumors compared with the low grade tumors ( P < 0.005) if vascular-origin tumor hemangioblastomas are considered a separate entity from other brain tumors.

Conclusion. The current study suggests that the ER- D5 antigen may participate in the growth of the intracra-

~~ ~~

From the Department of *Neurosurgery, and Department of t Pathology, Scientific Data Center for the Atomic Bomb Disaster, Na- gasaki University School of Medicine, Nagasaki, Japan.

The authors thank Ms. R. Yamasaki and Ms. Y. Komatsu for their technical assistance.

Address for reprints: Masao Kishikawa, M.D., Ph.D., Depart- ment of Pathology, Scientific Data Center for the Atomic Bomb Di- saster, Nagasaki University School of Medicine, 1-12-4 Sakamoto, Nagasaki 852, Japan.

Received August 18, 1994; revision received January 5, 1995; accepted January 30, 1995.

nial tumors and tumor angiogenesis. ER-D5 in embryonal and germ cell brain tumors suggests that ER-D5 may be a developmentally regulated protein. Cancer 1995; 75: 2571-8.

Key words: intracranial tumors, estrogen receptor-re- lated antigen, immunohistochemistry, embryonal tumor, germ cell tumor, vascular tumor, endothelial prolifera- tion.

The monoclonal D5 antibody already has been pre- pared' against the partially affinity-purified cytosolic estrogen receptor preparation from human myome- trium, and this antibody recognizes the 29 KDa (p29) cytosolic serine phosphoprotein termed estrogen recep- tor-related antigen (ER-D5). The ER-D5 under some ac- tivating conditions may form complex with the cyto- solic estrogen receptor.',2 The expression of ER-D5 has been reported in normal tissue, such as placenta3 and skin,4 as well as in breast, gastric, colorectal, pancreatic, and prostatic cancers; in malignant melanoma5-I3; and in human intracranial meningioma~.'~ A close associa- tion between ER-D5 and estrogen receptor has been demonstrated in some tumor^^,^,'" but not in oth- ers,6,14.15 In our previous study, we demonstrated the ER-D5 in 34 estrogen receptor-negative meningio- mas.14

The exact role of ER-D5 in cellular growth is un- known, but some physiologic and biochemical charac- teristics of ER-D5 have been reported in the literature. Phosphorylated p29 has guanosine triphosphate and, to a lesser extent, adenosine triphosphate binding

Guanosine triphosphate, phosphatase, and possibly phosphatase inhibitor molybdate acts on p29, endogenous nucleotide could then be released by dena- turation or condition of the elevated temperature.16 ER- D5 acts at the postreceptor stage and after binding to guanosine triphosphate may play a significant role in nucleotide metabolism.2 A close identity between ER-

2572 CANCER May 25,2995, Volume 75, No. 10

Table 1. Clinicopathology of the 143 Intracranial Tumors

Patient’s age (yr) Tumors from

Histology Range Median Male Female Total

Low grade astrocytoma 1-70 32 20 10 30 Anaplastic astrocytoma 15-70 62 11 2 13 Glioblastoma 17-79 54 18 9 27 Oligodendroglioma 41-42 41 2 0 2 Benign ependymoma 26-52 38 2 1 3 Anaplastic ependymoma 1-47 24 0 2 2

Benign meningioma 10-6 1 56 1 11 12 Anaplastic meningioma 59-76 65 0 3 3

Hemangioblastoma 25-82 45 8 3 11

Medulloblastoma 1-22 8 3 5 8

Schwannoma 22-71 68 11 12 23

Germ cell tumor 5-24 15 7 2 9 Total 83 60 143

D5 and heat shock protein-27 has been rep~rted.’~,’’ Meldelsohn et aL17 demonstrated that D5 antibody se- lectively recognizes thrombin-induced 29 KDa protein as well as unphosphorylated 27 KDa protein, and using this antibody, the authors isolated and sequenced a hu- man cDNA clone encoding a protein identical to the mammalian 27 KDa heat shock protein. The N-terminal sequence of p29 has homology with murine cytomega- loviral protein p89, one of the first proteins expressed in the cell after infection or reactivation from latency.”

In the current study we tried to demonstrate the im- munohistochemical expression of ER-D5 antigen in a large number of human intracranial tumors of different types. In the current study, 62% of the brain tumors showed ER-D5 positive immunoreactivity, and the staining characteristics are described.

Materials and Methods

Tumor Samples

One hundred forty-three brain tumors included in this study were operated on in our department at the Naga- saki University Hospital, Nagasaki, Japan, during the period from January 1987 to April 1994, and resected tumor samples were fixed in formalin and embedded in paraffin. The 5-pm tumor sections were taken for his- tology (hematoxylin and eosin stain), and consecutive sections were stained immunohistochemically with D5 antibody. Clinicopathologic features of patients are shown in Table I . The histologic classification of tumors was done based on criteria of the World Health Organi- zation’s Histologic Typing of the Central Nervous Sys- tem Tumors.” Histologically, there were 62 high grade tumors (13 anaplastic astrocytomas, 27 glioblastomas, 2

anaplastic ependymomas, 8 medulloblastomas, 3 ana- plastic meningiomas and 9 germ cell tumors) and 81 low grade tumors (30 low grade astrocytomas, 2 oligo- dendrogliomas, 3 benign ependymomas, 12 benign me- ningiomas, 23 schwannomas, and 11 hemangioblasto- mas). Twenty-eight of the 143 tumors were from pedi- atric patients. Two hemangiomas of the scalp and two dysgerminomas of the ovary were also immunostained.

Immunohistochemical Staining of ER-D5

Deparaffinized tumor sections were placed in methanol containing 0.3% hydrogen peroxide for 30 minutes and washed in 0.05 M phosphate-buffered saline (pH. 7.4) for 15 minutes. Tissue nonspecific activity was blocked by normal goat serum for 30 minutes, and sufficient pri- mary antibody monoclonal mouse anti ER-D5 (D5 anti- body; Amersham International, Buckinghamshire, U.K.) was applied and incubated overnight at 4OC. Bio- tinylated goat anti-mouse immunoglobulin was applied for 1 hour at room temperature, followed by avidin-bi- otin complex, which was applied for 1 hour (both anti- bodies were from Vector Laboratories, Burlingame, CA). After each incubation, sections were washed in phosphate-buffered saline for 15 minutes. The sub- strate chromogen was 3,3’-diamino-benzidine-4 hydro- chloride (Dojindo, Kumamoto, Japan), in which the sec- tions were incubated for about 30 minutes. The sections were then counterstained in Mayer‘s hematoxylin, de- hydrated, and mounted. Breast cancer tissue served as a positive control. An appropriate negative control was obtained by omitting the primary antibody, and com- parison was made with that of the positive control. The light microscopic visual assessment of ER-D5 positive cells was obtained as the percentage of total tumor cells

ER-D5 in the Intracranial Tumors/Khalid e t al . 2573

in a given section. When at least 5% of the total tumor cells in that section were mildly stained by the primary antibody, the section was considered to be ER-D5 posi- tive. A statistical analysis was performed using the Stu- dent's t test (mean values, standard deviations, and P values).

Results

cytoplasmic location. The positivity rate was 30% in the low grade astrocytomas and 55% in the high grade astrocytomas (anaplastic astrocytomas, 46%; glioblas- tomas, 59%). The small and large tumor cells, especially in glioblastomas, and cells in the pseudopalisading ar- eas, microvascular endothelial proliferations, and tu- mor blood vessels were usually ER-D5 positive. The mean percentage of ER-D5-positive cells was in the

Eighty-eight (62%) of the 143 brain tumors showed positive ER-D5 immunoreactivity with variable staining intensities and percentages of immunostained cells. ER- D5 staining patterns are shown in Figures 1 and 2. The positivity rate and mean percentage of ER-D5-positive cells varied in histologic type (Table 2). The ER-D5 po- sitivity rate was 63% (39/62) in the high grade tumors and 60% (49/81) in the low grade tumors. The order of increasing in frequency of ER-D5 positivity rate was oligodendrogliomas (OYO), low grade astrocytomas, be- nign ependymomas, anaplastic astrocytomas, anaplas- tic ependymomas, glioblastomas, medulloblastomas, meningiomas, schwannomas, and hemangioblastomas and germ cell tumors (100%). For mean percentage of ER-D5-positive cells, the order was oligodendrogliomas (O%), meningotheliomatous meningiomas, fibroblastic meningiomas, schwannomas, low grade astrocytomas, medulloblastomas, anaplastic astrocytomas, anaplastic meningiomas, benign ependymomas, glioblastomas, transitional meningiomas, anaplastic ependymomas, and germ cell tumors and hemangioblastomas (60.90 k 21.54%). Seventeen (61%) of the 28 tumors from the pediatric patients were ER-D5 positive. Normal brain, arachnoid, and dura mater adjacent to the tumor tissue seen in some sections were ER-D5 negative. The mean (k standard deviation) percentage of ER-D5-positive cells overall in high grade tumors was 30.12 k 29.05% and in low grade tumors was 22.10 k 26.79%. There was no statistically significant difference between the two groups (P > 0.1). If a hemangioblastoma is consid- ered a vascular-origin tumor and a separate entity from the rest of the brain tumors, the ER-D5 positivity rate in the low grade tumors stands at 56%, and the mean percentage of ER-D5-positive cells was statistically sig- nificantly higher in the high grade compared with the low grade tumors (30.12 f 29.05% vs. 16.0 k 22.35%, P < 0.005). The ER-D5 positivity rate was 63% in tu- mors from males and 60% in tumors from females. The mean age of the patients was 44.93 k 22.49 years for those with ER-D5 positive tumors and 42.24 -t 22.40 years for those with negative tumors.

gliobla&omas, 31.48 k 31.50% (range, 0-959'0); in the anaplastic astrocytomas, 21.15 f 27.13% (range, 0- 80%); and in the low grade astrocytomas, 15.33 -t 30.01% (range, 0-80%). There was a statistically sig- nificant higher percentage of ER-D5 positive cells in the glioblastomas compared with the low grade astrocyto- mas (P < 0.05). The positivity rate was in males, 51%; and in females, 29%. The mean age of patients was with ER-D5 positive tumors, 46.52 f 21.74 years; and with negative tumors, 42.95 k 23.78 years.

Meningiomas. Ten (67%) of the 15 meningiomas (two of three meningotheliomatous, two of three fi- broblastic, four of six transitional, and two of three an- aplastic) were ER-D5 positive. The moderate to intense immunostaining was seen in the meningothelial com- ponents and in transitional whorl formations, whereas fibroblastic components were mildly stained. In case of anaplastic meningiomas, tumor cells near the necrotic areas and tumor blood vessels showed positive ER-D5 immunoreactivity. The percentage of ER-D5 positive cells in meningiomas ranged from 0% to 60%; the mean percentage was in the benign meningiomas, 21.25 f 22.58% (meningotheliomatous, 6.67 -t 5.77%; fibro- blastic, 13.33 k 11.55%; transitional, 32.50 * 27.16%), and in the anaplastic meningiomas, 21.67 f 18.93%. ER-D5 expression in meningiomas had no relation to the patient's age and sex.

Schwannomas. Twenty (87%) of the 23 schwan- nomas were ER-D5 positive. The positivity rate was in the Antony A types, 50%; and in the Antony B types and in the mixed types, 100%. Few tumor blood vessels also showed the positive ER-D5 immunoreactivity. The percentage of ER-D5 positive cell ranged from 0% to 30%. The mean percentage was 13.91 _t 7.68% (in the A types, 10.0 k 12.65%; in the B types, 16.0 * 4.18; and in the mixed types, 14.58 k 4.50%).

Ependymomas. One of the two anaplastic ependy- momas with 70% tumor cells and one third of the be- nign ependymomas with 85% tumor cells showed pos- itive ER-D5 immunoreactivity, and the immunostain was seen in the isolated tumor cells, ependymal ro- settes, proliferative endothelial cells in anaplastic epen-

EX-05 Expression in the Different Brain Tumors dyrnoma, and in tumor blood vessels. Hemangioblastomas. All (100%) of the 11 heman-

gioblastomas were ER-D5 positive. Hemangioblasto- mas of reticular and cellular varieties, endothelial cells

Astrocytomas. Thirty-one (44%) of the 70 astrocy- tomas showed positive ER-D5 immunoreactivity at the

2574 CANCER May 25, 2995, Volume 75, No. 10

Figure 1. The representative sections of the brain tumors immunostained with monoclonal D5 antibody against ER-D5. Cytoplasmic ER-D5 is shown in the glioblastoma-pseudopalisading area (top left) and proliferative endovascular component (middle left), in transitional meningioma (bottom left), in schwannoma (top right), in ependymoma (middle right), and in hemangioblastoma (bottom right) (avidin-biotin complex, original magnification X400).

lining the capillary sinuses, intravascular cords, reti- d i n fibers, intervascular stromas, and tumor blood vessels usually showed positive ER-D5 immunoreactiv- ity. The mean percentage of ER-D5-positive cells was 60.90 * 21.54% (range, 20-90%). Both hemangiomas of the scalp were ER-D5 negative.

Medulloblastomas. Five (63%) of the eight medullo-

blastomas were positive for ER-D5, and the immunostain was seen either in the nucleus or in the cytoplasm and was heterogeneous. ER-D5 immunoreactivity was seen in the typical tissue pattern of medulloblastomas, such as rosette formations, vascular stromas, and microvascular endothe- lial proliferations. The mean percentage of ER-D5-positive cells was 20 * 21.88% (range, 0-60%).

ER-D5 in the Intracranial TumorslKhaZid et al . 2575

Figure 2. The representative sections of the brain tumors immunostained with monoclonal D5 antibody. ER-D5 antigen is shown in the nucleus of tumor cells and cytoplasm of endothelial proliferation in a medulloblastoma (top left), in the cytoplasm of germinoma (middle left) and endodermal sinus tumor (bottom left), and in the nucleus of an embryonal carcinoma (top right). There was no ER-D5 immunoreactivity in negative controls in glioblastoma (middle right). ER-D5 immunoreactivity is shown in a section of dysgerminoma of the ovary (bottom right) (avidin-biotin complex, original magnification X400).

Germ Cell Tumors. All (100%) of the nine germ ity was seen in the cytoplasm, except in the embryonal cell tumors (four germinomas, three embryonal carci- carcinomas, in which ER-D5 was seen either in the cy- nomas, one endodermal sinus tumor, and one imma- toplasm or in the nucleus. Blood vessels in two germi- ture teratoma) were positive for ER-D5. One of the two nomas also showed ER-D5 immunoreactivity. The dysgerminomas of the ovary showed moderately mean percentage of ER-D5-positive cells was 44.38 t stained ER-D5 immunoreactivity. The immunoreactiv- 30.64% (range, 15-85%).

2576 CANCER May 15,1995, Volume 75, No. 10

Table 2. The Expression of ER-D5 in the 143 Intracranial Tumors

Total Positive cells Histology tumors ER-D5t (%) (YO)*

Low grade astrocytoma 30 9 (30) 15.33 f 30.01 Anaplastic astrocytoma 13 6 (46) 21.15 f 27.13 Glioblastoma 27 16 (59) 31.48 k 31.50t

Benign ependymoma 3 1(33) 28.33 Anaplastic ependymoma 2 1(50) 35.0 Medulloblastoma 8 5 (63) 20.00 f 21.88 Benign meningioma 12 8 (67) 21.25 f 22.58 Anaplastic meningioma 3 2 (67) 21.67 f 18.93 Schwannoma 23 20 (87) 13.91 f 7.68 Hemangioblastoma 11 11 (100) 60.90 f 21.54 Germ cell tumor 9 9 1100) 44.38 * 30.64

Oligodendroglioma 2 0

* Mean k SD, or average. t Glioblastomas vs. low grade astrocvtomas, P < 0.05 (Student t test)

All (2/2) oligodendrogliomas were ER-D5 nega- tive.

Discussion

After our previous demonstration of ER-D5 in human meningioma~,'~ in the current study we have shown the immunohistochemical expression of ER-D5 in 62% of the 143 brain tumors of different histologic types. We noted not only the positivity rates but also the percent- age of ER-D5-positive tumor cells and areas of distribu- tion in sections of brain tumors, in which ER-D5 expres- sions were remarkably demonstrated in the cells and tissue pattern typical of different types of brain tumors. In the case of high grade astrocytomas, especially in glioblastomas, small and large tumor cells, multinucle- ated tumor giant cells, cells in the pseudopalisading ar- eas, and microvascular endothelial proliferations were usually ER-D5 positive. The expression of ER-D5 in brain tumors was cytoplasmic, except in medulloblasto- mas and embryonal carcinomas, in which ER-D5 im- munostain was seen either in the cytoplasm or in the nucleus. Similar cytoplasmic and nuclear location of ER-D5 has been reported in the syncytiotrophoblast of human placental villi.3 In the current study, we did not find any correlation between ER-D5 expression in brain tumors and patient's age and sex; our result agreed with the reported studes of tumors other than in the brain.5-8 In the case of breast and colon cancers, it has been re- ported that moderately differentiated tumors had higher ER-D5 values, although the authors did not found any correlation between the ER-D5 expression and tumor histologic In our study, ER-D5 ex- pression in brain tumors was not related to the tumor histologic grades, except in the astrocytic tumors, in

which a significantly higher positivity rate and percent- age of ER-D5-positive cells were seen in the glioblasto- mas compared with the low grade astrocytomas. The overall positivity rate and percentage of ER-D5-positive cells varied in different histologic types of brain tumors. In the case of high grade tumors, both positivity rates and mean percentage of ER-D5-positive cells were highest in the germ cell tumors, and the positivity rate was lowest in the anaplastic astrocytomas. The percent- age of ER-D5-positive cells was lowest in the medullo- blastomas. In the case of low grade tumors, both positi- vity rate and mean percentage of ER-D5-positive cells were highest in the hemangioblastomas. If vascular-or- igin tumor hemangioblastomas are excluded from the statistical analysis, a statistically significant higher per- centage of ER-D5-positive cells was observed in the overall high grade compared with low grade brain tu- mors. In the current study, normal brain tissue and me- ninges were ER-D5 negative, which agrees with our previous report.14 Our results suggest that ER-D5 anti- gen may play a significant role in the growth of human intracranial tumors.

In the current study, all germ cell tumors and most of the embryonal tumor medulloblastomas and pediat- ric brain tumors were positive for ER-D5. To the best of our knowledge, there is no previous report of ER-D5 antigen in the germ cell tumors, be they intracranial or not. The ER-D5 has been reported in the human term p la~en ta ,~ and in our study one of the two dysgermino- mas of the human ovary was positive for the ER-D5. Our findings suggest that ER-D5 may be a developmen- tally regulated protein. The ER-D5 in the tumors of vas- cular origin has not reported previously. We have seen the expression of ER-D5 in all hemangioblastomas, a tumor type of vascular origin, and in other brain tumors

ER-D5 in the Intracranial Tumors/Khalid et al. 2577

with high vascularity, in which ER-D5 immunoreactiv- ity was seen in most of the microvascular endothelial proliferations and tumor blood vessels. In the current study, we were unable to stain sufficient numbers of vascular tumor from other locations; the only two hem- angiomas in our study were ER-DS negative. Pre- viously, one of us (H.K.) demonstrated ER-D5 antigen in the hemangiopericytic meningi~rna. '~ Other authors have demonstrated ER-D5 antigen in the smooth mus- cle of tumor blood vessels in female sex organ^,^ in co- lorectal cancers," and in the vascular endothelium of breast cancers." Our study may suggest that certain brain tumors secrete ER-D5, and, probably by interact- ing with tumor angiogenic factors, influence the tumor angiogenesis.

We have demonstrated that a large number of hu- man intracranial tumors expressed the ER-D5 antigen, but the exact role of ER-D5 in the growth of brain tu- mors remains unknown. In a separate experiment from our laboratory, we have seen the expression of ER-D5 in the estrogen receptor-progesterone receptor negative human glioblastoma cell lines A172 and T98G. The ER- D5 participates at the very early stage of growth of the human glioblastoma cells in the in vitro condition, and the growth could be suppressed completely by treating the cell lines with the antibody against ER-D5 (manu- script submitted for publication. Khalid H, Yamashita H, Shibata S, 1994.). Originally, D5 antibody was ob- tained during the preparation of antibodies against par- tially affinity purified cytosolic estradiol receptor prepa- ration from human myometrium.' Subsequently, some authors have described ER-D5 as a substitute of estro- gen receptor identification or considered it to be quali- tatively and quantitatively related to estrogen recep- tors,5-8, 16.23,24 but other a ~ t h o r s ' ~ , ' ~ did not agree with the above hypothesis because of wide differences be- tween these two proteins.',6 In our previous study, we saw the expression of ER-D5 in the large number of es- trogen receptor-negative progesterone receptor-posi- tive intracranial meningioma~.'~ In the current study we were not concerned with finding a correlation between ER-D5 and estrogen receptor in the brain tumors, but all schwannomas studied were estrogen receptor negative (Unpublished observation. Khalid H, Shibata S, Kishi- kawa M, Yasunaga A, Tsutsumi K. Immunohistochem- ical study of ER-D5 antigen, estrogen and progesterone receptors in human schwannoma. Proceedings of the 52nd General Meeting of Japanese Society of Neurosur- gery, 1993 Sept 20-22; Tokyo.). Additional studies are needed to see whether a correlation exists between ER- D5 and estrogen receptor in brain tumors.

The clinical significance of ER-D5 in brain tumors needs to be studied. It has been demonstrated that ta- moxifen therapy increased the remarkable high sur-

vival rate in patients with ER-D5 positive pancreatic and breast cancers, compared with the negative tu- mors. 11,16 Other series, however, have reported that ER- D5-negative gastric cancer patients had significantly higher survival rates compared with patients with ER- D5-positive and there was no overall effect of tamoxifen on survival of patienk8 Thus, the reported data are conflicting. Another author" demonstrated that ER-D5 has no value as a prognosis marker. We are now studying the clinical outcome of patients with ER- D5-positive and -negative brain tumors.

In conclusion, the current study suggests that most human intracranial tumors express the ER-DS antigen, and ER-D5 may participate in the growth of brain tu- mors and tumor angiogenesis. The expressions of ER- D5 in the embryonal and germ cell tumors suggest that ER-DS may be a developmentally regulated protein. Further study is needed.

References

1 .

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

Coffer AI, Lewis KM, Brokas AJ, King RJB. Monoclonal antibod- ies against a component related to soluble estrogen receptor. Cancer Res 1985;45:3686-93. Coffer AI, King RJB. Characterization of p29, an estrogen recep- tor-associated tumor marker. ] Steroid Biocheni 1988;31:745-50. Cano A, Rivera J. Demonstration of p29, an estrogen receptor- associated tumor marker in human term placenta. Placenta

Viac J, Su H, Reano A, Kanitakis J, Chardonnet Y, Thivolet J. Distribution of an estrogen receptor-related protein (p29) in nor- mal skin and in cultured human keratinocytes. ] Dermatology

Coffer AI, Spiller GH, Lewis KM, King RIB. Immunoradiometric studies with monoclonal antibody against a component related to human estrogen receptor. Cancer Res 1985;45:3694-8. Cano A, Coffer Al, Adatia R, Millis RR, Rubens RD, King RJB. Histochemical studies with an estrogen receptor-related protein in human breast tumors. Cancer Res 1986;46:6475-80. King RJB, Coffer AI, Gilbert J, Lewis K, Nash R, Millis R, et al. Histochemical studies with a monoclonal antibody raised against a partially purified soluble estradiol receptor preparation from human myometrium. Cancer Res 1985;45:5728-33. Harrison JD, Morris DL, Ellis 10, Jones JA, Jackson 1. The effect of tamoxifen and estrogen receptor status on survival in gastric carcinoma. Cancer 1989;64:1007-10. Harrison JD, Jones JA, Ellis 10, Morris DL, Oestrogen receptor D5 antibody is an independent negative factor in gastric cancer. Br Surg 1991; 78:334-6. Takeda H, Yamakawa M, Takahashi T, Imai Y, Ishikawa M. Im- munohistochemical study with an estrogen receptor-related pro- tein (ER-D5) in human colorectal cancer. Cancer 1992; 69:907-

1991; 121153-60.

1989; 16:98-102.

IL. Hiromi T, Morita S, Takeda I, Mori J, Majima K, Matsuda H, et al. Hormone therapy of tamoxifen in resected carcinoma of the pancreas (in Japanese). Nippon Geka Gakkai Zasshi 1993;94:730- 5. Ohishi M, Miyake K, Koshikawa T, Asai J, Murase T. Determi- nation of ER-D5 (estrogen receptor related antigen) in prostatic

2578 CANCER M a y 25,1995, Volume 75, No. 10

cancer and its significance (in Japanese). Hiiiyokika Kiyo 1992; 38:

Paridaens DA, Alexander RA, Hungerford JL, MoCartney AC. Oestrogen receptors in conjuntival malignant melanoma: immu- nohistochemical study using formalin fixed paraffin wax sec- tions. 1 Cliti Pafliol 1991;44:840-3. Khalid H. Immunohistochemical study of estrogen receptor-re- lated antigen, progesterone and estrogen receptors in human in- tracranial meningiomas. Cancer 1994; 74:679-85.

15. Giri DD, Dangerfield VJM, Lonsdale R, Rogers K, Underwood ICE. Immunohistology of estrogen receptor and D5 antigen in breast cancer: correlation with estrogen receptor content of ad- jacent cryostat sections assayed by radioligand binding and en- zyme immunoassay. 1 Cliti Pathol 1987;40:734-40.

16. King RJB, Finley JR, Coffer AI, Millis RR, Rubens RD. Character- ization and biological relevance of a 29-KDa, estrogen receptor- related protein. ] Steroid Biochenz 1987;27:471-5.

789-96. 13.

14.

17. Mendelsohn ME, Zhu Y, O’Neill 5. The 29-KDa proteins phos- phorylated in thrombin-activated human platelets are forms of the estrogen receptor-related 27-KDa heat shock protein. PYOC Nut/ Acad Sci USA 1991; 88:11212-6. Ciocca DR, Leque LH. Immunological evidence for the identity between the hsp27 estrogen regulated heat shock protein and p29 estrogen receptor-associated protein in breast and endome- trial cancer. Breast Caiicev Res Treat 1991; 20:33-42. Hayward JR, Coffer AI, King RJB. Immunoaffinity purification and characterization of p29-an estrogen receptor related pro- tein. ] Steroid Biochenz Molec B id 1990;37:513-9.

20. Ziilch KJ. International Histological Classification of Tumours. No. 21. World Health Organization; Geneva, 1979.

21. Burlimann J, Gebhard S, Gomez F. Oestrogen receptor, proges- terone receptor, pS2, ER-D5, HSP27 and cathepsin Din invasive ductal breast carcinomas. Histopathology 1993; 23:239-48.

18.

19.