Immunohistochemical detection of carcinoma in radical prostatectomy specimens following hormone therapy


Tomomi Kusumi, MD, Department of Pathology, Hirosaki University School of Medicine, 5 Zaifu-cho, Hirosaki 036-8562, Japan. Email:


Following hormone therapy, residual carcinoma is frequently difficult to identify on HE-stained prostatectomy specimens. The aim of the present study was therefore to investigate whole-mounted specimens obtained by radical prostatectomy from patients who had undergone hormone therapy. Formalin-fixed and paraffin-embedded specimens were immunostained with prostate secretory cell markers including prostate-specific antigen (PSA), P504S (α-methylacyl-coenzyme A racemase, AMACR), P501S (prostein), and prostate-specific membrane antigen (PSMA). Residual carcinoma was detected in 250 histological slides of 42 patients in a total of 497 slides from 45 patients. In five of 250 slides (2%), carcinoma was not able to be recognized on HE-stained slides, but was found on the immunohistochemistry slides. PSMA had reacted positively beyond a moderate degree in carcinoma from all patients. PSA was positive for carcinoma in most of the patients, while negative or minimal staining was observed in a small number of patients. Carcinoma was positively reactive with P504S and P501S in most of the patients, but was negatively reactive in a few. The Gleason score for a pretreatment needle biopsy correlated with P504S staining of the prostatectomy specimens. P504S and P501S had limited ability to identify degenerated carcinoma. PSMA was the most useful marker to identify carcinoma after hormone therapy.

Hormone (endocrine) therapy is used to treat prostate carcinoma as a curative modality or neoadjuvant/adjuvant therapy with surgery. When hormone therapy is completely effective, carcinoma cells vanish or show vestigial features. In many cases, indeed, neoadjuvant hormone therapy could be a partial cure, and carcinoma cells may survive prostatectomy to a greater or lesser degree. Fibrous stroma shows as a scar in a cured carcinoma area, although fibrosis is frequently seen not only as traces in the prostate gland but also under various conditions such as inflammation and nodular hyperplasia. Residual carcinoma cells after hormone therapy demonstrate destructive and degenerative changes, such as decreased gland diameter, reduced and vacuolized cytoplasm, pyknosis and absence of nuclei, and stromal prominence.1–5 They are frequently difficult to identify and often mistaken for non-carcinoma structures such as atrophic prostate glands, capillaries, plump fibroblasts, inflammatory cells, and macrophages in surgical specimens obtained by prostatectomy on HE-stained histological specimens. Residual carcinoma needs to be exactly identified, although occasionally it has failed to be identified.

Immunohistochemistry will be helpful in identifying residual carcinoma foci. Immunohistochemistry of epithelial cell marker, low-molecular-weight cytokeratin (LMW-CK), clone CAM5.2, is recommended to identify residual carcinoma cells and to differentiate the carcinoma from the non-epithelial structures following hormone therapy.2,4 There are several molecular markers of basal cells including p63 and high-molecular-weight CK (HMW-CK), such as clone 34βE12 and D5/16B4.6,7 Absence of the basal cell layer of the prostate glands was typically noted in prostate carcinoma, and these markers are a useful tool to support the diagnosis of carcinoma in histological specimens. In these markers, CK 34βE12 is suitable and widely used as a marker for prostatic basal cells.

Many molecular markers are now available to detect prostate secretory cells in histological specimens by immunohistochemistry, including prostate-specific antigen (PSA), P504S (α-methylacyl-coenzyme A racemase, AMACR), P501S (prostein), prostate-specific membrane antigen (PSMA), prostate acid phosphatase, precursor forms of PSA (proPSA),8 prostate stem cell antigen,9 NK3 transcription factor related, locus 1 (NKH3.1),10,11 and six-transmembrane epithelial antigen of prostate-1 (STEAP-1).12 In antibodies against those prostate secretory cell markers, PSA is a kallikrein family member with serine protease that is secreted from glandular epithelial cells. Serum PSA level is elevated in patients suffering from prostate cancer, so PSA is one of the most important tools used for diagnosis and observation of the clinical course of prostate cancer.13,14 In immunohistochemistry of prostate tissue, PSA is positively stained in not only carcinoma cells but also non-neoplastic cells. P504S, also known as AMACR, is a mitochondrial and peroxisomal enzyme involved in the β-oxidation of branched chain fatty acids and fatty acid derivatives. P504S is identified as an upregulated expressed gene in prostate carcinoma on complementary DNA subtraction and microarray screening,15 and its protein is positively reactive with the cytoplasm of carcinoma cells and negatively with that of non-neoplastic cells on immunohistochemistry.6,15−21 P501S, also known as prostein, is found in prostate carcinoma using the same method as for P504S,22 and its protein is positively reactive with the cytoplasm of secretory cells in both carcinoma and non-neoplastic glands using immunohistochemistry.23,24 PSMA is a type II membrane protein that is expressed in all forms of prostate tissue, including carcinoma.25–28

The aim of the present study was to estimate the influence of hormone therapy on prostate tissue using immunohistochemistry of prostatic secretory cell markers, and to clarify which is the most valuable marker to identify degenerated carcinoma following hormone therapy.



A total of 45 male patients who had undergone hormone therapy before radical prostatectomy were included in the present study. They received treatment at Hirosaki University Hospital between April 2005 and March 2007, and their mean age was 66.8 years (range, 49–78 years).

In the neoadjuvant hormone therapy group: (i) five patients received luteinizing hormone-releasing hormone (LH-RH) agonist (goserelin acetate or leuprorelin acetate) alone; (ii) two patients received an anti-androgenic agent (bicalutamide) alone; (iii) 11 patients received maximum androgen blockade (MAB, a combination of an LH-RH agonist and an anti-androgenic agent); and (iv) 27 patients received a combination of LH-RH agonist and estramustine phosphate (EMP). The period of the neoadjuvant hormone therapy was between 3 and 17 months, with a median of 6 months.

The serum PSA levels of all patients were measured at initial diagnosis and at termination of neoadjuvant hormone therapy. The mean initial serum PSA level was 25.0 ng/mL (range, 4.4–95.6 ng/mL). The post-treatment PSA levels were all <4.0 ng/mL, with 40 patients having <1.0 ng/mL levels.

With regard to TNM classification for staging of carcinomas of the prostate, clinical T (cT) factors were determined at initial diagnosis by digital rectal examination, and pathological T (ypT) factors were determined after operation on pathology.7 cT1c, cT2a, cT2b and cT3 included eight, three, six and 28 patients, and ypT0, ypT2a, ypT2b, ypT3a, ypT3b and ypT4 included three, seven, 16, eight, nine and two patients, respectively.

The core needle biopsy specimens at diagnosis were examined and the Gleason scores were determined in 39 patients. Gleason score 6, 7, 8, and 9 included one, 17, three, and 18 patients, respectively. Preoperative needle biopsy specimens were not obtained from six patients.

Tissue samples

The radical prostatectomy specimens were fixed in 20% formalin and processed for histology. The distal (apical) and proximal (basal) part of the margins were vertically cut into 4–5 mm slices and the body parts were horizontally cut into 5–6 mm slices. These slices were embedded in paraffin and 6–25 paraffin blocks were prepared for each patient. Each paraffin block was cut into consecutive 4 µm-thick sections. One section was stained with HE, and the remainder of the sections was prepared for special stain and immunohistochemistry. In the HE-stained sections from all patients, the localization and distribution of residual carcinoma cells were carefully examined referring to the previously published reports with regard to histological changes due to neoadjuvant hormone therapy.1−5 Gleason score was not used for post-hormone therapy specimens according to worldwide consensus.29 In some specimens, toluidine blue stain were carried out in order to verify the mast cells.

The paraffin tissue blocks of needle biopsy specimens obtained for diagnosis, which were collected whenever possible, were subjected to immunohistochemistry. In addition as control, 20 specimens from radical prostatectomy without neoadjuvant hormone therapy were also used for immunohistochemistry. These prostatectomy specimens were individually obtained from randomly selected patients whose age and preoperative serum PSA level were approximately consistent with those of patients having neoadjuvant hormone therapy in the present study. We retrospectively utilized routine clinical materials in the present study, and followed the principles of the World Medical Association Declaration of Helsinki.


The antibodies used in the immunohistochemistry included antibodies for PSA (mouse monoclonal, clone ER-RP8, 1:100 dilution), P504S (rabbit monoclonal, clone 13H4, 1:200 dilution), P501S (mouse monoclonal, clone 10E3, 1:100 dilution), PSMA (mouse monoclonal, clone 3E6, 1:100 dilution), pancytokeratin (mouse monoclonal, clone AE1/AE3, 1:50 dilution), LMW-CK (mouse monoclonal, clone CAM5.2, 1:10 dilution), and HMW-CK (mouse monoclonal, clone 34βE12, 1:200 dilution). In these antibodies, CK CAM5.2 was supplied by Becton Dickinson Immunocytometry Systems (San Jose, CA, USA), and all of the others were supplied by Dako (Glostrup, Denmark). In antigen retrieval, a heat-induced procedure was performed using an autoclave at 121°C for 10 min, in a 0.01 mol/L citrate buffer, pH 6, for PSA and CK 34βE12, and in Target Retrieval Solution, pH 9 (Dako), for P504S, P501S and PSMA. A trypsin digestion was performed for CK AE1/AE3 and CK CAM5.2. The staining was carried out using the streptavidin–biotin–peroxidase complex technique with a Histofine kit (Nichirei, Tokyo, Japan) following the manufacturer's instructions. The sections were made to react with the chromogen, 3,3′-diaminobenzidine-tetrahydrochloride (Merck, Darmstadt, Germany), and were then counter-stained with hematoxylin. A specimen with the primary antibodies omitted was used as the negative control.

Immunohistochemistry was assessed on prostate secretory cells of either the carcinoma glands or non-neoplastic glands. The staining intensity in immunohistochemistry was categorized on the following scale based on positive cell rate (Fig. 1): category 1, negative or minimally positive (<10%) cells were observed (Fig. 1c); category 2, focal positive (≥10% and <90%) cells were observed (Fig. 1d,e,g); and category 3, diffuse positive (≥90%) cells were observed (Fig. 1c,f,g).

Figure 1.

Histological features of prostatectomy specimens after hormone therapy. (a) Small glands and clusters of residual carcinoma are densely distributed in the left-lower part (arrowheads) and expanded round glands with basal cell hyperplasia of non-neoplastic prostate glands are situated in the right upper part (arrows). (b) Residual carcinoma cells were arranged in single cells, cords, and small glands (arrowheads), and the non-neoplastic glands had enlarged diameter (arrow). (c) The carcinoma was category 3 (arrowhead) and the non-neoplastic glands was category 1 (arrow) based on prostate-specific antigen (PSA) staining. (d) The carcinoma was category 2 (arrowheads) based on P504S staining. (e) The carcinoma with the feature of perineural invasion was category 2 based on P501S staining (arrowhead). The cytoplasm of mast cell is also reactive for P501S (white arrow). (f) Both the carcinoma (arrowhead) and the non-neoplastic glands (arrows) were category 3 based on prostate-specific membrane antigen (PSMA) staining. (g) The secretory cells of carcinoma were category 2 (arrowhead) and those of non-neoplastic glands were category 3 (arrow) based on low-molecular-weight cytokeratin (LMW-CK) CAM5.2 staining. Some of the basal cell layers of the non-neoplastic glands were also stained with LMW-CK CAM5.2 (white arrowhead). (h) Basal cell layers stained with high-molecular-weight (HMW)-CK 34βE12 are absent in the carcinoma (left side in the panel) and are present with hyperplasia in the non-neoplastic glands (arrow). (a,c,d,fh) are from one patient; (b,e) are from another.

On preliminary immunohistochemistry it was easy and verifiable with reproducible results to differentiate the staining intensity between negative/minimally positive cells and almost all positive cells, with the former and the later assessed as category 1 or 3, respectively. But it was difficult to distinguish weak staining intensity from intermediate staining and determine the exact percentage of positive cells without interobserver variation. For practical purposes, therefore, ‘weak’ and ‘intermediate’ results were lumped together in category 2.

Statistical analysis

To compare parameters, statistical analysis was performed using Spearman's rank correlation coefficient with spss software (version 12.0, SPSS, Chicago, IL, USA). The result was deemed statistically significant for P < 0.05.


In the present study a total of 497 histological slides from 45 patients were examined. Residual carcinoma was detected in 250 slides of 42 patients. In five of those 250 slides (2%), carcinoma was not able to be recognized on HE-stained slides, but was found on the immunohistochemistry slides. In the remaining three of the 45 patients, residual carcinoma was detected on neither HE-stained nor immunohistochemistry slides. In HE-stained slides, effects of neoadjuvant hormone therapy on both carcinoma and non-neoplastic prostate glands were observed in the histological specimens from all 45 patients. The carcinoma glands decreased in diameter, and demonstrated the patterns of small glands, cords, clusters, or individual cells, which were scattered in the fibrous stroma. In carcinoma cells cytoplasm was reduced in volume and vacuolated, and nuclei had pyknotic change (Fig. 1a,b). In the non-neoplastic glands, the diameters were relatively retained or became enlarged with cystic change. Even if diameter was decreased in the non-neoplastic glands, secretory cells had almost disappeared and basal cells were prominent. The prostate secretory cells had reduction in volume, while the basal cells were prominent due to hyperplasia and squamous metaplasia (Fig. 1a,b).

PSA, P504S, P501S, and PSMA were positively reactive with prostate secretory cells. The lack of CK 34βE12-positive basal cell layer was used to ensure differentiation of carcinoma from non-neoplastic glands. Positive staining with P504S was not seen in the non-neoplastic glands (Fig. 1d). PSA and PSMA had a reticulated pattern in the cytoplasm of those cells. P504S and P501S had a granular pattern in the cytoplasm of those cells (Fig. 1d,e). P501S consisted of more rough granules, and was reactive with not only the prostatic cells but also mast cells (Fig. 1e), which had granules showing metachromasia on toluidine blue stain in their cytoplasm (data not shown). Paraffin tissue blocks of the needle biopsy specimens obtained from 25 patients were subjected to immunohistochemistry. The specimens of those needle biopsies and 20 from prostatectomy without hormone therapy produced the same findings that the cytoplasm of carcinoma cells was strongly reactive with PSA, P504S, P501S, and PSMA, with no difference among histological features and Gleason scores, and that the cytoplasm of non-neoplastic secretory cells was strongly reactive with PSA, P501S, and PSMA. In the atrophic non-neoplastic glands, those prostate secretory cell markers demonstrated weak or negative reaction (data not shown).

Total numbers of patients in the immunohistochemistry categories of carcinoma and secretory cells in the non-neoplastic glands are shown in Table 1. PSA was positively reactive with prostate carcinoma and non-neoplastic glands in most of the patients, although one of 42 patients (2%) with regard to carcinoma glands, and three of 45 patients (7%) with regard to non-neoplastic glands had category 1. PSMA was moderately to firmly reactive with carcinoma and non-neoplastic glands of almost all patients, and none of the patients had category 1 in residual carcinoma or non-neoplastic glands. For P504S staining, nine patients (21%) had category 1, and nine patients (21%) had category 3 in residual carcinoma. For P501S staining, six patients (14%) had category 1, and 12 patients (29%) had category 3 in residual carcinoma. CK CAM5.2 was positively reactive with cytoplasm of carcinoma and secretory cells of non-neoplastic glands for all patients (Fig. 1g). According to CK CAM5.2, however, 19 of 42 patients (45%) with regard to carcinoma and nine of 45 patients (20%) with regard to non-neoplastic glands included category 2. CK CAM5.2 reacted with the basal cells without squamous metaplasia as well as secretory cells in the non-neoplastic glands (Fig. 1g). CK AE1/AE3 was stained in the secretory cells and the basal cells even if squamous metaplasia had occurred (data not shown). CK 34βE12 was not reactive with secretory cells in either the carcinoma or non-neoplastic glands. Positive staining of CK 34βE12 was seen in the cytoplasm of basal cells of the non-neoplastic glands (Fig. 1h), in which disruption or unremarkable features were seen when the cytoplasm of basal cells had narrowed.

Table 1.  Immunohistochemistry
 Prostate carcinoma glands (42 patients); n (%)Non-neoplastic prostate glands (45 patients); n (%)
  • In three of 45 patients in total, adenocarcinoma completely disappeared from the prostatectomy specimens.

  • CK, cytokeratin; PSA, prostate-specific antigen; PSMA, prostate-specific membrane antigen.

11 (2)9 (21)6 (14)0 (0)0 (0)3 (7)16 (35)0 (0)0 (0)
215 (36)24 (58)24 (57)5 (12)19 (45)35 (78)25 (56)25 (56)9 (20)
326 (62)9 (21)12 (29)37 (88)23 (55)7 (15)4 (9)20 (44)36 (80)

Differences in histological changes in HE-stained specimens and immunohistological categories were not observed among the four neoadjuvant hormone modalities. The T factors including clinical and pathological were not correlated with immunohistochemistry categories of carcinoma in the present study. Core needle biopsy specimens at diagnosis were obtained from 39 patients, and their Gleason scores were carefully re-examined. Residual carcinoma was not found in one patient with Gleason score 6 and two patients with Gleason score 8. Gleason score 7, 8, and 9 included 17, one, and 18 patients, respectively. Among the prostate secretory cell markers, pretreatment Gleason scores correlated with P504S categories of the surgical specimens after neoadjuvant hormone therapy (P < 0.05, Table 2).

Table 2.  Gleason score in pretreatment needle biopsy vs immunohistochemistry of carcinoma in prostatectomy specimens after hormone therapy
PretreatmentNo. patients*PSA categoryP504S category*P501S categoryPSMA category
Gleason score*†123123123123
  • *

    Pretreatment Gleason scores correlated with P504S categories of the carcinoma after hormone therapy on Spearman's rank correlation coefficient (P < 0.05).

  • Pretreatment needle biopsy specimens were obtained from 36 patients who had residual carcinoma after prostatectomy.

  • PSA, prostate-specific antigen; PSMA, prostate-specific membrane antigen.



Residual carcinoma was mostly found on HE-stained prostatectomy specimens after hormone therapy, but it was occasionally not detected in cases of small glands or single cells infiltrating into the fibrous stroma. Residual carcinoma was confirmed in five of 250 histological slides (2%), in which carcinoma was not detected on HE staining by immunohistochemistry. It was possible to identify these indistinguishable carcinoma foci on immunohistochemistry. There were non-neoplastic glands that contained scant CK 34βE12-positive basal cells in the present study. It is known that normal prostate glands are usually surrounded with CK 34βE12-positive basal cells, but that they occasionally have disrupted or absent basal cells, even if hormone therapy was not performed.30 The negative staining of CK 34βE12 in basal cells did not necessarily have to be induced by hormone therapy. It was difficult to confirm carcinoma on immunohistochemistry using single antibody. It has been noted that CK CAM5.2 was useful to identify the scattered carcinoma cells in prostate tissue after hormone therapy.2,4 Carcinoma cells were category 2 and 3 in the prostatectomy specimens of all patients in the present study, although the basal cell layers of the non-neoplastic glands were also reactive with CK CAM5.2. Because the non-neoplastic glands after hormone therapy consisted mainly of inconspicuous secretory cells and prominent basal cells, positive staining of CK CAM5.2 was more congested than in staining with prostatic secretory cell markers.

Among the prostate secretory cell markers, P504S is useful to distinguish between carcinoma and non-neoplastic glands in prostate tissue and to make a diagnosis using needle biopsy specimens.18 P504S produced strong cytoplasmic granular staining in 100% of prostate carcinomas and negative staining in 88% of non-neoplastic glands.16 Tang et al., however, stated that after hormone therapy the expression of P504S was reduced in 71% of the paired prostate carcinoma cases as compared with pre-hormone therapy.21 Suzue et al. noted that P504S expression was reduced significantly in the majority of prostate carcinomas after hormone therapy.20 Although all examined biopsy specimens of carcinoma at diagnosis had strongly positive P504S staining, 21% of residual carcinomas were category 1 in the prostatectomy specimens after hormone therapy in this study. P504S was not deemed a useful positive marker to identify residual carcinoma. In addition, it was reported by Martens and Keller that P504S did not increase the recognition of prostate carcinoma after radiation therapy as neoadjuvant therapy.31 P504S was not related to Gleason score in prostate carcinoma without hormone therapy in the previous immunohistochemistry study.17,18 In the present study, however, high category of P504S after hormone therapy correlated with high Gleason score before hormone therapy.

P501S is a molecule expressed in normal and carcinoma prostatic tissue, and was immunohistochemically reactive with almost all examined prostatic tissue at the primary and metastatic site according to the literature.23,24 In the present study, however, 14% of carcinomas and 35% of non-neoplastic glands were assessed as scoring category 1 in prostate specimens after hormone therapy, although all prostate glands were positively stained in the needle biopsy specimens before hormone therapy. It was suspected that the expression of P501S was decreased due to hormone therapy, and P501S had limited ability to identify residual carcinoma after hormone therapy. In addition, P501S was also disadvantageous in identifying carcinoma after hormone therapy, because it was reactive with some inflammatory cells, which was not mentioned previously.

Both PSA and PSMA were positively reactive with the majority of carcinomas and non-neoplastic glands in the prostate after hormone therapy in the present study. It has been noted that the expression of PSA was decreased in carcinoma cells after hormone therapy.5,32 Indeed, PSA staining after hormone therapy in the present study was present in 2% and 36% of patients in category 1 and 2, respectively, although strongly positive reaction was present in the control specimens from pretreatment needle biopsy and prostatectomy without hormone therapy. PSMA rather than PSA demonstrated frequent positive staining in prostate glands, and all were category 2 or 3, not category 1. Non-epithelial structures mimicking carcinoma cells, such as capillaries, fibroblasts, inflammatory cells, and macrophages, were negatively stained with PSMA, and were able to be excluded on PSMA-immunostained slides. Although PSMA was reactive for prostate secretory cells of both benign and malignant tissue, carcinoma foci showing small clusters, cords and single cells were more easily identified and distinguished from non-neoplastic epithelial cells than in HE-stained specimens.

On immunohistochemistry in surgical specimens, it was reported by Wright et al. that PSMA-expression was upregulated in prostate carcinoma after hormone therapy.33 In the present study, however, category 3 was not given to all specimens of carcinoma and was only given to fewer than half of the specimens of non-neoplastic glands in PSMA staining, although normal and carcinoma glands were strongly stained in the needle biopsy specimens before hormone therapy. It was, therefore, suggested that expression of PSA and PSMA was decreased by hormone therapy. Other researchers have reported a different result – that PSMA expression was not affected by short-term application of hormone therapy,34 although there was no difference in duration of therapy in the present study. In addition, other reports have stated that PSMA expression correlated with disease recurrence and acted as a predictor of disease progression.26,28 Further examination of influence of hormone therapy on PSMA expression is needed.

It is concluded that detection of residual carcinoma on immunohistochemistry using anti-PSMA antibody can be helpful. The present study demonstrated that the features of residual carcinoma and non-neoplastic glands were the same as those described in the previous reports.1−5 In the neoadjuvant hormone therapy used in the present study, 27 patients received both LH-RH and EMP. EMP combines the properties of hormone therapy and chemotherapy because EMP is a mixture of estradiol and nitrogen mustard. Although the prostate tissues from the patients should be influenced by chemotherapy, histological changes in those tissues were similar to those in other hormone therapies. The present study was retrospective and the duration of hormone therapy was irregular in each modality, but histological differences corresponding to durations of therapy were not distinct. Although neoadjuvant hormone therapy is widely performed and can decrease the rate of positive surgical margin, it did not influence the rate of disease relapse as defined by serum PSA level.35 There are, however, unresolved controversial issues about survival rates of patients who are treated with neoadjuvant hormone therapy, especially those with high-grade prostate cancer.36 It is necessary in neoadjuvant hormone therapy that residual carcinoma is precisely evaluated in order to confirm the presence or absence of disease progression or downstaging, to assess the surgical margin, and to decide which follow-up modality to use.


This work was supported in part by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science.