Vol.:(0123456789)1 3

Clinical & Experimental Metastasis (2017) 34:479–490 https://doi.org/10.1007/s10585-018-9878-x

RESEARCH PAPER

Prostate-specific membrane antigen (PSMA) expression in breast cancer and its metastases

Mariz Kasoha1 · Clara Unger1 · Erich‑Franz Solomayer1 · Rainer M. Bohle2 · Claudia Zaharia2 · Fadi Khreich3 · Stefan Wagenpfeil4 · Ingolf Juhasz‑Böss1

Received: 30 August 2017 / Accepted: 1 February 2018 / Published online: 10 February 2018 © Springer Science+Business Media B.V., part of Springer Nature 2018

AbstractThe present study was undertaken to investigate the expression of prostate-specific membrane antigen (PSMA) in normal breast tissues, in cancerous breast tissues and in distant metastases from patients with breast cancer. Immunohistochemi-cal analysis was performed to determine PSMA expression and angiogenic activity using anti-PSMA mAb and anti-CD31 mAb respectively. Immunofluorescence staining was applied to confirm the exact co-localization of PSMA and CD31. We observed different patterns of PSMA expression between normal and cancerous tissues. Normal breast tissues showed PSMA expression only in normal glandular cells. However, primary breast tumors and distant metastases showed PSMA expression in tumor cells and in tumor-associated neovasculature. PSMA score group status in primary breast tumors was significantly associated with histologic type and tumor grade (p = 0.026 and p = 0.004 respectively). Distant metastases showed higher PSMA expression in tumor-associated neovasculature comparing with primary tumors. Moreover, brain tumor-associated neovasculture had significantly higher expression of PSMA comparing with bone tumor-associated neovasculture. The localized binding of PSMA mAb to the neovasculature endothelium was confirmed with the double Immunofluorescence staining. 68Ga-PSAM imaging of a patient with metastatic breast cancer showed strong tracer uptake in all known skeletal metastases. To the best of our knowledge, this study is the second one that has assessed PSMA expression in a large number of breast cancer patients. Our findings showed that PSMA is particularly expressed in tumor-associated neovasculature of breast tumors and its distant metastases, thus enhancing the evidence on the potential usefulness of PSMA as a therapeutic vascular target.

Keywords Breast cancer · Prostate-specific membrane antigen (PSMA) · Angiogenesis · Metastases · Neovasculature

Introduction

Prostate-specific membrane antigen (PSMA) is a 84–100 kDa type II transmembrane glycoprotein which is first cloned in the human prostate parenchyma and has a unique 3-part structure, a 707-amino acid external portion, a 24-amino acid transmembrane portion and a 19-amino acid internal portion [1]. While initial observations assumed that the expression of PSMA is limited to prostate cells, it is now well known that it is also expressed to a limited degree in other tissues. For example, PSMA expression is seen in the brush border of the small intestine, in the proximal renal tubules, in the spleen, in the liver, in the salivary glands and in the brain [2]. Moreover, Godron et al. demonstrated for the first time the expression of PSMA in a variety of neovasculature from physiologic regenerative and repara-tive conditions such as vessels in proliferative endometrium

* Mariz Kasoha mariz.kasoha@uks.eu

1 Department of Gynecology, Obstetrics and Reproductive Medicine, University Medical School of Saarland, 66421 Homburg, Saar, Germany

2 Institute of General and Surgical Pathology, University Medical School of Saarland, 66421 Homburg, Saar, Germany

3 Department of Nuclear Medicine, Saarland University, 66421 Homburg, Saar, Germany

4 Institute of Medical Biometry, Epidemiology and Medical Informatics, Saarland University, 66421 Homburg, Saar, Germany

480 Clinical & Experimental Metastasis (2017) 34:479–490

1 3

and granulation tissue [3]. Kinoshita and colleagues found that PSMA is expressed in a variety of normal tissues and this expression is dysregulated in different cancerous tissues [4]. The exact function of PSMA remains unclear, but it is thought to play an important role in nutrients uptake and cell signalling [5]. In addition, PSMA showed to have different enzymatic activity based on its location of expression, (i) folate hydrolase activity at the small intestine where it is responsible for removing the C-terminal glutamates from dietary folate allowing its absorption and transportation into the rest of the body and (ii) N-acetylaspartylglutamate peptidase (NAALADase) activity in the nervous system by which it enhances the production of N-acetylaspartylgluta-mate (NAAG), the most abundant peptide neurotransmitter in the mammalian nervous system [6, 7]. The impact of this enzymatic activities on human prostate tissue and perhaps elsewhere, however, is yet unknown, as does the question regarding the existence of a natural ligand for PSMA.

In the prostate, PSMA is expressed at low levels in benign prostatic epithelium and upregulated several fold in the majority of advanced prostatic malignancies [8, 9]. Consequently, the PSMAʹs interaction with PSMA-specific antibodies or other ligands in normal tissues can be ignored [10, 11]. Additionally, PSMA expression has been shown to correlate with aggressiveness of the prostate cancer (PC), with highest levels expressed in an androgen-deprived state and metastatic disease [5]. Moreover, it has been found that PSMA does have internalization signal that allows inter-nalization of the protein on the cell surface into an endoso-mal compartment [12]. As a result, these observations have made PSMA an attractive molecular target for oncological imaging and radionuclide therapy in the management of PC [13, 14]. Using PSMA PET/CT, it has become apparent that PSMA expression persists in a high percentage of patients with PC in contrast to the expression of blood biomarkers such as prostate-specific antigen (PSA) [15, 16] providing the rationale for the recent introduction of PSMA radioli-gand therapy, with promising initial results [17, 18].

In regard to other cancers, different studies confirmed the expression of PSMA in the neovasculature of numerous solid cancers including pancreatic cancer [19], lung cancer [20], colorectal cancer [21], kidney cancer [22] and blad-der cancer [23]. These studies showed that the endothelial cell expression of PSMA was highly restricted to tumor-associated neovasculature representing candidate target for tumor-specific anti-angiogenic cancer therapy.

Breast cancer (BC) continues to be the second leading cause of cancer-related death among women [24]. Although exciting therapy options have declined the disease incidence and improved the survival rate for BC, optimal regimens remain open questions because of the serious side effects and drug resistance scenarios that can be initiated dur-ing the treatment. Therefore, continued research is highly

recommended to improve patient stratification for the exist-ing therapies and to provide new therapeutic targets. To date, the assessment of PSMA expression in BC and its potential in disease diagnosis and therapy is limited by very little number of studies that included small cohort sizes except for one study of Wernicke et al. [25]. Therefore, we aimed in our study to investigate the expression of PSMA in primary and metastatic breast tumors.

Materials and methods

Patients

This study included a cohort of 72 women with histologi-cally confirmed BC. The study was approved by the local ethics committee of the Medical Association of the Saarland (reference number: 207/10). Written and informed consent was obtained from each patient who was still alive at the beginning of our study.

All specimens were diagnosed according to the WHO classification in effect at the time of initial diagnosis. Demo-graphics and clinicopathologic features of all patients are summarized in Table 1.

Representative formalin-fixed paraffin-embedded blocks were obtained from the Department of Pathology and Neu-ropathology of the University of Saarland, including 70 pri-mary breast tumors of which 39 normal breast tissues and 10 distant metastases were available. In addition, two blocks of two metastases second to breast cancer without their cor-responding primary tumors were included. The metastases cohort consisted of one, six and five cases of lung, bone and brain respectively. Consecutive sections of 4-µm thickness were prepared from the blocks for hematoxylin-eosin (HE) staining, for immunohistochemistry (IHC) analysis and for immunofluorescence (IF) analysis. All HE stained sections were tested by a pathologist to confirm the expected histo-logical characteristics of the investigated block.

Immunohistochemistry (IHC) staining

Immunohistochemical staining was performed to validate PSMA expression. In order to assess the expression of PSMA within the vascular endothelium, we additionally stained for an endothelial cell marker, CD31. In brief, fol-lowing deparaffinization, antigen retrieval was performed on block sections using (Dako retrieval solution S1699, PH = 6). Then, sections were blocked with 3% bovine serum albumin (BSA fraction V, Sigma-Aldrich) for 30 min at room temperature. Incubation with the primary antibodies was done for one hour at 37 °C using PSMA mAb-clone 3E6 and CD31 mAb- clone JC70A from DAKO. Thereaf-ter, staining was applied using DAKO REAL ™ Detection

481Clinical & Experimental Metastasis (2017) 34:479–490

1 3

System, Alkaline phosphatase/RED kit according to manu-facturer instructions. After washing in PBS buffer, slides were counterstained with hematoxylin for 6 min, raised in water, dehydrated in ascending concentrations of ethanol fol-lowed by clearance with xylene and cover slipped. Staining

specificity was checked using negative controls which under-went the entire staining procedure except that the primary antibody was omitted. Prostatic adenocarcinoma specimens and tonsillar tissues served as a positive controls for PSMA and CD31 staining respectively.

Assessment of PSMA expression

Immunohistochemical staining was evaluated by a patholo-gist unaware of the clinical and pathological data. All tissue sections were analyzed under Zeiss microscope (Axioskop 40, Carl Zeiss, Germany) and selected pictures were cap-tured with attached digital camera (AxioCam MRC, Carl Zeiss, Germany) using Axiovision Documentation Rel.4.8 programm.

PSMA expression in tumor cells

Using HE stained slides; we have identified the percentage of tumor cells in each tested slide. PSMA expression was quantified according to Remmele and Stegner method [26]. Staining intensity was scored semi quantitatively as negative (0), weak (1), moderate (2) or strong (3). Extent of staining was scored as percentage of cells stained (0, 0% of cells; 1, < 10% of cells; 2, 10–50% of cells; 3, 51–80% of cells; 4, > 80% of cells). The final immune reactive score (IRS) was determined by multiplying the scores of intensity and extent of staining, ranging from 0 to 12. IRS < 2 was considered as negative staining, 3–4 as weak staining, 6–8 as moderate staining and 9–12 as strong staining.

PSMA expression in tumor-associated neovasculature

Using HE stained slides; we have identified 5 areas (hot-spots) that had the highest concentration of microves-sels. The number of CD31- and PSMA-positively stained microvessels in each of these hot-spots was calculated using CD31 stained slides and PSMA stained slides respectively. Counted vessels were chosen according to Weidner’s method [27]. The percentage of PSMA expression in tumor vessels was calculated using the following equation:

Percentage of PSMA expression in tumor vessels= (Aver-age of vessels numbers stained for PSMA / Average of ves-sels numbers stained for CD31) × 100.

PSMA staining intensity in the neovasculature was graded on a 0–3 scale (0, no staining; 1, weak; 2, moderate; and 3, strong). Then, the final score of PSMA expression in the neovasculature was obtained by multiplying the intensity grade with the percentage of PSMA expression in tumor vessels obtaining a range of 0 to 12 (< 2 was considered as negative staining, 3–4 as weak staining, 6–8 as moderate staining and 9–12 as strong staining).

Table 1 Clinical and pathological characteristics of study patients

a Other: invasive lobular carcinoma (N = 12), invasive ductal and lobular (N = 3), tubular carcinoma (N = 1), and mucinous carcinoma (N = 2)

Characteristics Number Percentage (%)

All cases 72 100Age at diagnosis Median (range); years 60.7 (39.9–86.4)

Menopause Premenopausal 18 25 Postmenopausal 54 75

M Status M0 62 86 M1 10 14

T Stage (unknown: one case) T1 36 50.7 T2 29 40.8 T3 3 4.2 T4 3 4.2

N stage N0 36 50.0 N1 25 34.7 N2 3 4.2 N3 6 8.3 NX 1 1.4 N+ 1 1.4

Histology Invasive ductal 54 75.0 Othera 18 25.0

Receptor status (unknown: one case) ER+ 11 15.5 ER− 60 84.5 PR+ 49 69.0 PR− 22 31.0 Her2Neu+ 9 12.7 Her2Neu− 62 87.3 Triple negative 9 12.7

Ki-67 index (%) (unknown: five cases)< 10% 15 22= 10–14% 24 36> 14% 28 42Grading (unknown: one case) G1 7 9.9 G2 49 69.0 G3 15 21.1

482 Clinical & Experimental Metastasis (2017) 34:479–490

1 3

Immunofluorescence (IF) staining

Double immunofluorescence staining was done to confirm the localized binding of PSMA mAb to the neovasculature endothelium. For this purpose, we have tested all available sections of six different cases including four cases (2 M0 and 2 M1) that showed positive IHC- PSMA staining and two cases (1 M0 and 1 M1) that showed negative IHC-PSMA-staining. The sections were deparaffinized and washed as previously described in the IHC staining method. The anti-gen retrieval was achieved with applying Tris–EDTA buffer (pH 9) (EDTA and Trizma base from Sigma-Aldrich) for 30 min at 38 °C, followed by overnight incubation with the primary antibodies (PSMA mAb-clone 3E6 from DAKO and CD31 rabbit mAb from Abcam) at 4 °C. After washing in phosphate-buffered saline (PBS), the slides were blocked one hour with 3% bovine serum albumin (BSA fraction V, Sigma-Aldrich) for 30 min at room temperature. Secondary anti-mouse (Alexa Fluor®568 goat anti-mouse, Life technol-ogies) and anti-rabbit (Alexa Fluor ® 488 goat anti-rabbit) were applied at room temperature for one hour. The slides were sequentially mounted with DAPI Prolong ® Diamond Antifade Mountant from Life technologies. Images were taken on Olympus BX53 microscope using Olympus XC30 camera and cellSens Standard Olympus program.

Statistical analysis

Descriptive statistics (including median, range, and abso-lute as well as relative frequencies) were used to define the study cohort characterizations. Correlations between parameters were assessed according to the nonparametric Spearman rank correlation and respective statistical test. Student’s t-test and Wilcoxon t-test were used for compar-ing the differences between paired parametric and non-par-ametric continuous data respectively. The Chi square test or Fisherʹs exact test were used to assess the relation of PSMA expression score group to the clinicopathological features as appropriate. P-values were two sided and subject to a local significance level of 5%. Due to the overall nature of the study we did not account for the issue of multiple testing and thus report unadjusted p-values. Overall survival (OS) and metastases-free survival (MFS) survival between the

expression categories were compared using Kaplan–Meier plots and log-rank test. Patients were followed up from the date of surgery with a mean follow-up period of 33 months (Range: 2-296 months). All statistical analyses were per-formed using SPSS version 21 (IBM, Armonk, NY, USA).

Results

PSMA‑IHC staining in primary BC tissues

Seventy formalin-fixed paraffin-embedded blocks of primary breast tumors were available. We have found that PSMA was expressed in tumor cells as well as in tumor -associ-ated endothelial cells. PSMA staining showed cytoplasmic reactivity in tumor cells and membranous with or without cytoplasmic reactivity in the endothelial cells lining tumor microvessels. Staining appeared homogenously in the great majority of the evaluated cases. In case of heterogeneous staining, the predominant pattern was recorded. PSMA expression in tumor cells appeared in 72% (50/70) of tested cases and ranged between weak and moderate. On the other hand, 46% (31/68) of tested cases showed PSMA expression in tumor-associated neovasculature of which 6% had strong staining (IRS = 9–12) (Table 2). A weak but significant cor-relation between PSMA expression in tumor cells and in tumor-associated neovasculture was identified in 68 matched cases (Spearman p = 0.011, R² = 0.053).

As PSMA expression in tumor cells showed different pattern comparing to its expression in tumor-associated neovasculature, we defined specifically the expression pat-tern in each case and found that 18% (12/68) of tested cases had negative PSMA expression both in tumor cells and in tumor-associated neovasculature, 37% (25/68) had posi-tive PSMA expression in tumor cells and negative PSMA expression in tumor-associated neovasculature, 10% (7/68) had negative PSMA expression in tumor cells and positive PSMA expression in tumor-associated neovasculature and 35% (24/68) had positive PSMA expression both in tumor cells and in tumor-associated neovasculature (Fig. 1). PSMA expression was defined as negative when IRS = 0–2 and as positive when IRS = 4–12.

Table 2 Expression pattern of PSMA in 70 primary breast tumor tissues

Data showed as number of cases and percentage

PSMA Staining pattern in primary breast tumors

NegativeIRS: 0–2

WeakIRS: 3–4

ModerateIRS: 6–8

StrongIRS: 9–12

All

PSMA in tumor cells 20 (28%) 25 (36%) 25 (36%) 0 (0%) 70 (100%)PSMA in tumor-associ-

ated neovasculature37 (54%) 14 (21%) 13 (19%) 4 (6%) 68 (100%)

483Clinical & Experimental Metastasis (2017) 34:479–490

1 3

PSMA expression in BC and its relation to the clinicopathological parameters

A summary of the relation of PSMA immunoreactivity to the clinicopathological parameters is presented in Table 3. Study patients were divided in four groups depending on PSMA expression in tumor cells and in tumor-associated neovasculature as mentioned above. Statistically signifi-cant relation was demonstrated between PSMA expression and both tumor histopathological subtype and tumor grad-ing (p = 0.026 and p = 0.004 respectively).

PSMA expression in normal breast tissues

In order to further assess whether PSMA expression is dysregulated in breast tumor tissue, we have identi-fied protein expression pattern in 39 normal breast tis-sues obtained from patients included in the study. PSMA expression in normal glandular breast cells appeared in 67% (26/39) of tested cases, of which 12 cases, 11 cases and 3 cases exhibited weak, moderate and strong staining respectively. In contrast, the vessels of the corresponding benign tissue did not display PSMA immunoreactivity and

Fig. 1 HE and PSMA−/ CD31-IHC-staining in primary breast tumors. Line a: PSMA staining is negative in tumor cells as well as in tumor-associated neovasculature. Line b: PSMA staining is posi-tive in tumor cells but not in tumor-associated neovascular. Line c:

PSMA staining is positive in tumor-associated neovasculature but not in tumor cells. Line d: PSMA staining is positive in tumor cells as well as in tumor-associated neovasculature

484 Clinical & Experimental Metastasis (2017) 34:479–490

1 3

were completely PSMA negative except for one case that showed weak PSMA staining in the vessels. Immunostain-ing for CD31-labeled antibody was present in virtually all vessels in each case (Fig. 2). There was no significant

difference between PSMA-IRS of normal breast cells comparing to tumoral breast cells [IRS median (range): 4 (0–12) vs 4 (0–8) respectively; p = 0.633 (Wilcoxon t-test)].

Table 3 Relation of PSMA immunoreactivity to clinicopathological factors

Expression was coded as (−) when IRS = 0–2 and as (+) when IRS = 4–12NS not significant according to Fisher’s exact test

PSMA expression in primary breast tumors

Tumor cells/tumor-associated neovasculature

Factor Cases -/ - Number of cases

+/- Number of cases

-/+ Number of cases

+/+ Number of cases

P-value

Age < 60.7 34 6 5 13 10 NS ≥ 60.7 34 6 2 12 14

Menopause Premenopausal 16 3 3 6 4 NS Postmenopausal 52 9 4 19 20

Histological subtype Invasive ductal 51 5 5 20 21 0.026 Other 17 7 2 5 3

T status T1 35 7 3 15 10 NS T2, T3, T4 32 5 4 10 13 N status N0 35 9 5 12 9 NS N1, N2, N3 31 3 2 12 14

G status G1, G2 53 12 7 21 13 0.004 G3 15 0 0 4 11

ER receptor ER− 11 0 0 5 6 NS ER+ 57 12 7 20 18 PR receptor PR− 21 4 1 7 9 NS PR+ 47 8 6 18 15

Her2 receptor Her2− 60 11 7 22 20 NS Her2+ 8 1 0 3 4

Hormon status Triple negative 59 12 7 20 20 NS Not triple negative 9 0 0 5 4

Ki-67 Index (%) < 10% 15 5 0 7 3 NS = 10%-14% 21 4 5 5 7 > 14% 28 2 2 11 13

CA-125 > 29U/mL 25 8 2 10 5 NS < 29U/mL 9 1 1 3 4

M status M0 58 11 7 22 18 NS M1 10 1 0 3 6

485Clinical & Experimental Metastasis (2017) 34:479–490

1 3

PSMA expression in distant metastases of BC

We further investigated PSMA expression in 12 different distant metastases secondary to BC including one metastasis in the lung, six metastases in the bone and five metastases in the brain. PSMA in tumor cells was positively expressed in 75% (9/12) of tested cases including 4 cases with weak staining and 5 cases with moderate staining. In addition, PSMA expression in tumor-associated neovasculature was positively expressed in 89% (8/9) of tested cases of which 4 cases had strong staining (IRS = 9–12) (Table 4).

Within 7 matched primary breast tumors and corre-sponding distant metastases, PSMA expression in tumor-associated neovasculature was significantly higher in distant metastases comparing to the primary breast tumors [IRS median (range): 6 (2–12) vs 2 (0–6) respectively; p = 0.049 (Wilcoxon t-test)]. However, PSMA-IRS in tumor cells showed no significant difference between metastases tumors and primary tumors [IRS median (range): 8 (2–8) vs 8 (4–8) respectively; p = 0.318 (Wilcoxon t-test)] (Fig. 3).

Then we have compared PSMA expression pattern between brain metastases and bone metastases (Fig. 4). PSMA-IRS in tumor cells did not differ between both metastases [IRS (average ± STD): 4.6 ± 3.2 vs 4.6 ± 2.4 respectively; p = 0.970 (Student’s t-test)]. However, brain tumor-associated neovasculture had significantly higher expression of PSMA compared with bone tumor-associated neovasculture [IRS (average ± STD): 10.0 ± 2.7 vs 4.0 ± 2.0 respectively; p = 0.014 (Student’s t-test)].

Prognostic significance of PSMA expression in BC

The relationship of OS and MFS with PSMA expression was evaluated. OS was defined from day of surgery until death. MFS was defined from day of surgery until the date of metastasis involvement for patients with metastasis and until the date of the last examination at our department for patients without metastasis. Study patients were divided in 4 groups depending on PSMA expression in tumor cells and in tumor-associated neovasculature as mentioned above. No

Fig. 2 HE and PSMA−/ CD31-IHC-staining in normal breast tissues. Line a: PSMA staining is positive in normal breast cells but not in the associated neovasculature. Line b: PSMA staining is negative both in normal breast cells and in the associated neovasculature

Table 4 Expression pattern of PSMA in 12 distant metastases secondary to BC

Data showed as number of cases and percentage

PSMA Staining pattern in distant metastasis secondary to BC

NegativeIRS: 0–2

WeakIRS: 3–4

ModerateIRS: 6–8

StrongIRS: 9–12

All

PSMA in tumor cells 3 (25%) 4 (33%) 5 (42%) 0 (0%) 12 (100%)PSMA in tumor-associ-

ated neovasculature1 (11%) 2 (22%) 2 (22%) 4 (45%) 9 (100%)

486 Clinical & Experimental Metastasis (2017) 34:479–490

1 3

statistically significant influence of PSMA expression on OS or MFS was observed [Log-Rank (Mantel-Cox): p = 0.259 and p = 0.413 respectively]. Therefore, PSMA expression had no prognostic role within our study collective.

Double IF staining of PSMA and CD31

Double IF staining was done to confirm the localized bind-ing of PSMA mAb to the neovasculature endothelium. The qualitive outcomes of PSMA-IHC staining in the selected tested cases were completely identical with IF results (Fig. 5).

68Ga‑PSMA imaging of metastatic BC

A 79-year-old patient with metastatic BC, which had strong immunohistochemically expression of PSMA in bone metas-tases, was referred to the department of nuclear medicine for 68Ga-PSAM imaging. The patient was subjected too to a Whole-body bone scan with 650 MBq 99Tc-MDP. Stand-ard Staging with CT and MRI showed no lymph node- or visceral metastases. A whole-body PET/CT was done after 60 min. post injection of 87 MBq 68Ga-PSMA-HBED-CC. 68Ga-PSAM-PET-MIP showed a strong tracer uptake in all known skeletal metastases (Fig. 6).

Discussion

Numerous studies have suggested PSMA as an attractive target for the diagnosis and therapy of PC and demonstrated that this therapy is safe and has a low toxicity profile [28, 29]. BC and PC are the two most common invasive cancers in women and men, respectively. Although the breast and the prostate are different in terms of anatomy and physiologi-cal function, tumors that arise from them have remarkable underlying biological similarities. These facts prompted our interest to evaluate PSMA expression and its prognostic sig-nificance in BC.

In the present study, we characterized PSMA expression in 72 patients with BC. IHC-staining was applied on the cor-responding formalin-fixed paraffin-embedded blocks using PSMA mAb-clone 3E6 and CD31 mAb- clone JC70A from

Fig. 3 PSMA expression pattern in 7 matched primary breast tumors and their corresponding distant metastases in tumor cells (a) and in tumor-associated neovasculature (b). Results are presented as median (range)

Fig. 4 PSMA expression pattern in five brain metastases and six bone metastases. O: PSMA-IRS in tumor cells; delta: PSMA-IRS in tumor-associated neovasculature

487Clinical & Experimental Metastasis (2017) 34:479–490

1 3

DAKO. A wide variety of PSMA mAbs has been utilized in different detection methods. Therefore, in order to avoid inconsistent or even contradictory results we have reviewed the common mAbs that were applied for the IHC-detection of PSMA in various tumor entities. Then, we have chosen the most used mAb generally and the one that have been used for PSMA detection in BC particularly [30–32]. Addi-tionally, because PSMA has been reported to be present in the endothelium of tumor vessels, we used a baseline

vascular density score, via CD31staining, to determine the extent of PSMA staining in the neovasculture.

To date, our study is the second one that has assessed PSMA expression in a large cohort of patients with primary and metastatic BC next to Wernicke et al. study (72 and 92 included cases respectively) [25]. In our study, PSMA expression was evaluated according to the final IRS values that were determined by multiplying the scores of intensity and extent of staining, ranging from 0 to 12. Results showed

Fig. 5 IF staining of PSMA and CD31 in primary breast tumor. Line a: PSMA expression is positive in tumor-associated neovasculature and co-localized with CD31. Line b: PSMA expression is negative in tumor-associated neovasculature

Fig. 6 90mTc-DPD planar whole-body shows widespread skeletal metastases (a, b), 63Ga-PSMA-PET-MIP shows a strong tracer uptake in all known skeletal metastases (c)

488 Clinical & Experimental Metastasis (2017) 34:479–490

1 3

that PSMA-immunoreactivity appeared in tumor cells as well as in tumor-associated neovasculature (72% and 46% of tested cases respectively). However, we found it notable that PSMA staining in tumor cells had more extent but less intensity compared to its staining in the tumor-associated neovasculature (data are not shown). PSMA expression in tumor cells was correlated weakly but significantly with its expression in tumor-associated neovasculture. In addi-tion, 39 corresponding normal breast tissues from our study patients were tested and revealed that PSMA-staining was positive in normal glandular breast cells and completely negative in the associated vessels except one case that had low staining score. Our findings are consistent, in part, with Wernicke et al. results which observed that PSMA positiv-ity appeared in tumor-associated vasculature in 90 out of 92 cases. In the remaining two cases, weak staining was also seen in normal tissue-associated vasculature. In contrast to our observation, normal glandular breast tissue and carci-noma cells were PSMA-negative in all studied cases.

PSMA expression in normal and/or tumorous breast has been also investigated in few studies. However, the number of tested specimens was very limited. Chang and coworkers used five different PSMA mAbs and showed that PSMA was positively reacted in the neovasculature but not in tumor cells of 5/6 of breast carcinoma cases. In addition, all 5 mAbs consistently reacted in mammary ductal epi-thelium in 8 cases [32]. In another study by Kinoshita and colleagues, PSMA-expression showed moderate staining in normal breast parenchymal elements in 6 out of 6 cases and weak staining in 1 out of 5 ductal carcinomas [4]. Moreo-ver, 8 breast phylloides tumor, which have been listed as total benign entities in a study by Mhawech-Fauceglia et al. were completely negative for PSMA staining [33]. Finally, PSMA expression was found to be associated with the neo-vasculature of 7 out of 10 infiltrating ductal breast tumors [34]. These studies show conflicting observations concern-ing PSMA expression in normal breast tissue and in breast tumor cells. The reason for this conflict is unclear; however, it might be due to using different anti-PSMA mAbs that recognize different epitopes of the protein or to the differ-ences in the histological components of the tested mammary tissue. On the other hand, our finding that PSMA expression in the neovasculature is largely restricted to the endothelial cells within the pathologically defined tumor area is consist-ent with the previous studies in addition to another recent study that has identified PSMA expression in primary breast tumors and related brain metastases from four patients [29].

Furthermore, our analysis showed that PSMA expression in the primary breast tumors was significantly correlated with histological subtype and tumor grading. In addition, no association was observed between PSMA expression and OS or MFS. These observations were inconsistent with Wernicke et al. findings which showed that increased PSMA

score in the tumor-associated neovasculature was associated with worse OS when was analysed based on less than 50% staining vs more but not when was based on three different PSMA scores. These contradictory observations could be due to the different scoring methodology that has been used in each study. We intended in our study to apply a reference scoring methodology according to Remmele and Stegner [26] in order to be used as a standard for further studies to avoid any inconsistent outcomes that may be resulted from the different scoring methodology applied in data dichotomization.

In regard to PSMA expression in metastases second to BC, the quantification of PSMA staining in 12 different distant metastases from our patient’s cohort ruled out that, similarly to PSMA expression in the primary tumors, PSMA immunoreactivity in tumor-associated neovasculature was stronger in comparing to tumor cells. PSMA expression within our 7 matched primary tumors and their related dis-tant metastases (3 at bone and 4 at brain) was not identical. This is in contrast to the findings of Wernicke et al. and Nomura et al. whereby PSMA expression was identical between primary breast tumor and their related brain metas-tasis in 10 and 4 matched cases respectively. These contra-dicted results can be related to the limited number of tested cases and, once again, to the different scoring methodology that have been used in each study. A notable finding of our study is that PSMA immunoreactivity in metastases second to BC did not demonstrate the same pattern. According to these observations, we suggest that PSMA expression in a distant metastasis second to BC cannot be predicted from its expression in the primary tumor and this expression is differently presented depending on the site of metastases.

The exact function of PSMA is yet unknown. A recent in vitro model by Bradbury and colleagues demonstrated that knockdown of PSMA in two metastatic human BC cell lines resulted in a reduction in the proliferative- and adhe-siveness ability of the cells and a decrease in invasiveness- and migratory capacity. These effects were explained via the disrupted expression profile of different matrix metallo-proteinases including MMP2, MMP3, MMP10 and MMP13 which play a major role in cell behaviors [35]. Regarding the restricted PSMA expression to the tumor-associated neovas-culature, it has been speculated that tumor microenviron-ment may influence the differential expression of PSMA in the neovasculature. Liu and coworkers showed in an in vitro model that PSMA expression on cultured human umbili-cal vein endothelial cells (HUVECs) was specifically lim-ited to growth conditions with tumor-conditioned medium (TCM) derived from human BC cells (MDA-MB-231) in Matrigel [36]. The hypothesis that the cancer cell-HUVEC crosstalk can induce PSMA expression was further tested by Nguyen and colleagues who showed that conditioned media from several human cancer cell lines induced PSMA

489Clinical & Experimental Metastasis (2017) 34:479–490

1 3

expression in HUVECs, in vitro, and these lines induced PSMA, in vivo, in a HUVEC coimplantation mouse model presumably through induction by regulatory factors secreted by tumor cells. In adition, PSMA-expressing HUVECs have the capacity to internalize J591, a mAb that recognize the extracellular domain of PSMA [37]. PSMA expression in tumor-associated neovasculature has been proved in a wide variety of solid tumors including bladder cancer [23], colo-rectal cancer [21], and renal cancer [22, 38] suggesting a plausible role of PSMA in endothelial cell biology. Indeed, it was shown that endothelial cells lacking PSMA show less efficient matrix invasion. This phenotype was suggested to reflect the function of PSMA in a regulatory loop involving ß1 integrin and PAK1, which mediate angiogenesis through cell invasion [39]. Furthermore, Godron and coworkers suggested that the PSMA-mediated folate uptake could be required to regenerate essential cofactors of endothelial nitric oxide synthase which is essential for the angiogenesis [3].

In light of these findings, in recent years, there has been a surge in development of both diagnostics and therapeu-tics that take advantage of the expression and activity of PSMA. These include monoclonal antibodies, RNA apa-tmers, small molecules, peptides, and vaccines [40]. We have shown in our study that 68Ga-PSMA-PET of one of our patients, which had strong immunohistochemically expression of PSMA in bone metastases, presented strong tracer uptake in all known skeletal metastases that have been identified with CT. Sathekge et al. reported a case of a 33-year-old woman with metastatic BC who underwent 68Ga-PSMA and 18F-FDG PET/CT imaging for restaging. Images demonstrated intense skeletal uptake in the axial and appendicular skeleton with liver metastases suggesting that 68Ga-PSMA may provide helpful prognostic information. Furthermore, 68Ga-PSMA-avid metastatic sites may aid in selecting tumors with high PSMA expression for PSMA-directed therapy [41]. In addition, Sathekge et al. showed that the imaging using 68Ga-PSMA-HBED-CC PET in BC confirms the reported considerable variation of PSMA expression on human solid tumors using immunohisto-chemistry [42]. These findings show that PSMA-targeted radionucleotide imaging may pose a novel option for therapy monitoring in advanced BC. In a phase I trial study using the PSMA-targeted docetaxel-containing nanoparticle BIND-014 in patients with advanced solid tumors, one patient with cervical cancer metastatic to lymph nodes showed complete response. Radiographically confirmed partial responses could be detected in a KRAS-mutant lung adenocarcinoma, in an ampullary adenocarcinoma metastatic to the liver, in a case of BC and gastroesophageal cancer [43]. Another phase I study demonstrated that tumor-associated neovascu-lature in multiple advanced metastatic solid cancers could be targeted with an Indium-111-labeled antibody (J591)

binding the extracellular domain of PSMA. No objective tumor regression could be observed whereas toxicity was acceptable [44].

Taken together, our data provide strong evidence that PSMA is expressed in primary breast tumors and their dis-tant metastases; therefore it may be considered as a potential diagnostic and therapeutic target. Nonetheless, we warrant IHC methodological standardization and further validation of the potential usefulness of PSMA as a prognostic marker in a larger prospective serial.

Acknowledgements We thank Barbara Linxweiler from the depart-ment of obstetrics and gynecology, University Medical School of Saar-land for her excellent technical assistance.

References

1. Israeli RS, Powell CT, Fair WR, Heston WD (1993) Molecular cloning of a complementary DNA encoding a prostate-specific membrane antigen. Cancer Res 53: 227–230

2. Silver DA, Pellicer I, Fair WR, Heston WD, Cordon-Cardo C (1997) Prostate-specific membrane antigen expression in normal and malignant human tissues. Clin Cancer Res 3:81–85

3. Godron IO, Tretiakova MS, Noffsinger AE, Hart J, Reuter VE, Al-Ahmadie HA (2008) Prostate-specific membrane antigen expres-sion in regeneration and repair. Mod Pathol 21:1421–1427

4. Kinoshita Y, Kuratsukuri K, Landas S, Imaida K, Peter M, Rovito PM Jr, Wang CY, Haas GP (2006) Expression of prostate-specific membrane antigen in normal and malignant human tissues. World J Surg 30(4): 628–636

5. Rajasekaran AK, Anilkumar G, Christiansen JJ (2005) Is pros-tate-specific membrane antigene a multifunctional protein? Am J Physiol Cell Cell Physiol 288: C975–C981

6. Shafizadeh TB, Halsted CH (2007) Gamma-glutamyl hydrolase, not glutamate carboxypeptidase II, hydrolyzes dietary folate in rat small intestine. J Nutr 137:1149–1153

7. Neale JH, Bzdega T, Wroblewska B (2000) N-Acetylaspartylglu-tamate: the most abundant peptide neurotransmitter in the mam-malian central nervous system. J Neurochem 75(2):443–452

8. Perner S, Hofer MD, Kim R, Shah RB, Li H, Möller P, Hautmann RE, Gschwend JE, Kuefer R, Rubin MA (2007) Prostate-specific membrane antigen expression as a predictor of prostate cancer progression. Hum Pathol 38:696–670

9. Ghosh A, Heston WD (2004) Tumor target prostate specific mem-brane antigen (PSMA) and its regulation in prostate cancer. J Cell Biochem 91:528–539

10. Nanus DM, Milowsky MI, Kostakoglu L, Smith-Jones PM, Valla-bahajosula S, Goldsmith SJ, Bander NH (2003) Clinical use of monoclonal antibody HuJ591 therapy: targeting prostate specific membrane antigen. J Urol 170:S84–S88

11. Troyer JK, Beckett ML, Wright GL Jr (1995) Detection and char-acterization of the prostate-specific membrane antigen (PSMA) in tissue extracts and body fluids. Int J Cancer 62:552–558

12. Rajasekaran SA, Anilkumar G, Oshima E, Bowie JU, Liu H, Hes-ton W, Bander NH, Rajasekaran AK (2003) A novel cytoplasmic tail MXXXL motif mediates the internalization of prostate-spe-cific membrane antigen. Mol Biol Cell 14:4835–4845

13. Pfestroff A, Luster M, Jilg CA, Olbert PJ, Ohlmann CH, Lassmann M, Maecke HR, Ezziddin S, Bodei L (2015) Current status and future perspectives of PSMA-targeted therapy in Europe: oppor-tunity knocks. Eur J Nucl Med Mol Imaging 42:1971–1975

490 Clinical & Experimental Metastasis (2017) 34:479–490

1 3

14. Osborna JR, Akhtar NH, Vallabhajosula S, Anand A, Deh K, Tagawa ST (2013) Prostate-specific membrane antigen-based imaging. Urol Oncol 31:144–154

15. Jadvar H (2015) PSMA PET in prostate cancer. J Nucl Med 56:1131–1132

16. Ceci F, Uprimny C, Nilica B et al (2015) (68)Ga-PSMA PET/CT for restaging recurrent prostate cancer: which factors are associ-ated with PET/CT detection rate? Eur J Nucl Med Mol Imaging 42:1284–1294

17. Scott T, Tagawa MI, Milowsky M, Morris S, Vallabhajosula P, Christos NH, Akhtar et al (2013) Phase II study of lutetium-177-labeled anti-prostate-specific membrane antigen monoclonal antibody J591 for metastatic castration-resistant prostate cancer. Clin Cancer Res 19:5182–5191

18. Weineisen M, Schottelius M, Simecek J et al (2015) 68Ga- and 177Lu-labeled PSMA I&T: optimization of a PSMA-targeted theranostic concept and first proof-of-concept human studies. J Nucl Med 56:1169–1176

19. Ren H, Zhang H, Wang X, Liu J, Yuan Z, Hao J (2014) Prostate-specific membrane antigen as a marker of pancreatic cancer cells. Med Oncol 31(3):857

20. Wang HL, Wang SS, Song WH, Pan Y, Yu HP, Si TG, Liu Y, Cui XN, Gue Z (2015) Expression of prostate-specific membrane antigen in lung cancer cells and tumor neovasculature endothe-lial cells and its clinical significance. PLoS ONE. https ://doi.org/10.1371/journ al.pone.01259 24

21. Haffner MC, Kronberger IE, Ross JS et al (2009) Prostate-specific membrane antigen expression in the neovasculature of gastric and colorectal cancers. Hum Pathol 40:1754–1761

22. Baccala A, Sercia L, Li J, Heston W, Zhou M (2007) Expression of prostate-specific membrane antigen in tumor-associated neo-vasculature of renal neoplasms. Urology 70:385–390

23. Samplaski MK, Heston W, Elson P, Magi-Galluzzi C, Hansel DE (2011) Folate hydrolase (prostate-specific membrane antigen) 1 expression in bladder cancer subtypes and associated tumor neo-vasculature. Mod Pathol 24:1521–1529

24. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM, Forman D, Bray F (2015) Cancer inciedence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer 136:359–386

25. Wernicke AG, Varma S, Greenwood EA, Christos EA, Chao PJ, Liu KS, Bander H, Shin NH SJ (2013) Prostate-specific mem-brane antigen expression in tumor-associated vasculature of breast cancers. Acta Pathol Microbiol Immunol Scand (APMIS) 122:482–489

26. Remmele W, Stegner HE (1987) Recommendation for uniform definition of an immunoreactive score (IRS) for immunohisto-chemical estrogen receptor detection (ER-ICA) in breast cancer tissue. Pathologe 8:138–140

27. Weidner N, Folkman J, Pozza F, Bevilacqua P, Allred EN, Moore DH, Meli S, Gasparini G (1992) Tumor angiogenesis: a new sig-nificant and independent indicator in early-stage breast carcinoma. J Natl Cancer Inst 84:1875–1887

28. Ahmadzadehfar H, Zimbelmann S, Yordanova, A, Fimmers R, Kürpig S, Eppard E, Gaertner FC, Wei X, Hauser S, Essler M (2017) Radioligand therapy of metastatic prostate cancer using 177Lu-PSMA-617 after radiation exposure to 223Ra-dichloride. Oncotarget. https ://doi.org/10.18632 /oncot arget .15698

29. Uwe Haberkorn M, Eder K, Kopka JW, Babich, Eisenhut M (2017) New Strategies in Prostate Cancer: Prostate-Specific Mem-brane Antigen (PSMA) Ligands for Diagnosis and Therapy. Clin Cancer Res. https ://doi.org/10.1158/1078-0432.CCR-15-0820

30. Nomura N, Pastorino S, Jiang P, Lambert G, Crawford JR, Gym-nopoulos M, Piccioni D, Juarez T, Pingle SC, Makale M, Kesari

S (2014) Prostate specificmembraneantigen (PSMA) expression in primary gliomas and breast cancer brain metastases. Cancer Cell Int. https ://doi.org/10.1186/1475-2867-14-26

31. Tykvart J, Navrátil V, Sedlák F, Corey E, Colombatti M, Fracasso G, Koukolík F, Bařinka C, Sácha P, Konvalinka J (2014) Compar-ative analysis of monoclonal antibodies against prostate-specific membrane antigen (PSMA). Prostate 74:1674–1690

32. Chang S, Reuter VE, Heston WD, Bander NH, Grauer LS, Gaudin PB (1999) Five different anti-prostate-specific membrane antigen (PSMA) antibodies confirm PSMA expression in tumor-associ-ated neovasculature. Prostate 59:3192–3198

33. Mhawech-Fauceglia P, Zhang S, Terracciano L, Terracciano L, Sauter G, Chadhuri A, Herrmann FR, Penetrante R (2007) Prostate-specific membrane antigen (PSMA) protein expression in normal and neoplastic tissues and its sensitivity and specific-ity in prostate adenocarcinoma: an immunohistochemical study using mutiple tumour tissue microarray technique. Histopathology 50:472–483

34. Ross JS, Schenkein D, Webb I, Gray G, Deeds J, Meyer R, McDonald A, Sheehan C, Gray K (2004) Expression of prostate specific membrane antigen in the neo-vasculature of non-prostate cancers. J Clin Oncol 22(14-suppl):3110

35. Bradbury R, Jiang WG, Cui YX (2016) MDM2 and PSMA play inhibitory roles in metastatic breast cancer cells through regula-tion of matrix metalloproteinases. Anticancer Res 36:1143–1152

36. Liu T, Jabbes M, Nedrow-Byers JR, Wu LY, Bryan JN, Berkman CE (2011) Detection of prostate-specific membrane antigen on HUVECs in response to breast tumor-conditioned medium. Int J Oncol 38:1349–1355

37. Nguyen DP, Xiong PL, Liu H, Pan S, Leconet W, Navarro V et al (2016) Induction of PSMA and Internalization of an Anti-PSMA mAb in the Vascular Compartment. Mol Cancer Res 14:1045–1053

38. Al-Ahmadie HA, Olgac S, Gregor PD, Tickoo SK, Fine SW, Kondagunta GV, Scher HI, Morris MJ, Russo P, Motzer RJ, Reu-ter VE (2008) Expression of prostate-specific membrane antigen in renal cortical tumors. Mod Pathol 21:727–732

39. Conway RE, Petrovic N, Li Z, Heston W, Wu D, Shapiro LH (2006) Prostate-specific membrane antigen regulates angiogen-esis by modulating integrin signal transduction. Mol Cell Biol 26:5310–5324

40. Evans JC, Malhorta M, Cryan JF, OʹDriscoll CM (2016) The ther-apeutic and diagnostic potential of the prostate specific membrane antigen/glutamate carboxypeptidase II (PSMA/GCPII) in cancer and neurological disease. Br J Pharmacol 173:3041–3079

41. Sathekge M, Modiselle M, Vorster M, Mokgoro N, Nyakale N, Mokaleng B, Ebenhan T (2015) 68Ga-PSMA imaging of meta-static breast cancer. Eur J Nucl Med Mol Imaging 42:1482–1483

42. Sathekge M, Lengana T, Modiselle M, Vorster M, Zeevaat JR, Maes A, Ebenhan T, Van de Wiele C (2017) 68Ga-PSMA-HBED-CC PET imaging in breast carcinoma patients. Eur J Nucl Med Mol Imaging 44:689–694

43. Von Hoff DD, Mita MM, Ramanathan RK, Weiss GJ, Mita AC, LoRusso PM et al (2016) Phase I study of PSMA-targeted docetaxel-containing nanoparticle BIND-014 in patients with advanced solid tumors. Clin Cancer Res 22:3157–3163

44. Milowsky MI, Nanus DM, Kostakoglu L, Sheehan CE, Vallab-hajosula S, Goldsmith SJ, Ross JS, Bander NH (2007) Vascular targeted therapy with anti-prostate specific membrane antigen monoclonal antibody J591 in advanced solid tumors. J Clin Oncol 25:540–547